From 9447082ae666fbf3adbe9c9152117daa31a8b737 Mon Sep 17 00:00:00 2001 From: Sakari Ailus Date: Wed, 16 Nov 2016 12:21:19 -0200 Subject: [PATCH 001/343] [media] smiapp: Implement power-on and power-off sequences without runtime PM Power on the sensor when the module is loaded and power it off when it is removed. Signed-off-by: Sakari Ailus Signed-off-by: Mauro Carvalho Chehab --- drivers/media/i2c/smiapp/smiapp-core.c | 29 +++++++++----------------- 1 file changed, 10 insertions(+), 19 deletions(-) diff --git a/drivers/media/i2c/smiapp/smiapp-core.c b/drivers/media/i2c/smiapp/smiapp-core.c index 59872b31f832..620f8ce8185e 100644 --- a/drivers/media/i2c/smiapp/smiapp-core.c +++ b/drivers/media/i2c/smiapp/smiapp-core.c @@ -2741,8 +2741,6 @@ static const struct v4l2_subdev_internal_ops smiapp_internal_ops = { * I2C Driver */ -#ifdef CONFIG_PM - static int smiapp_suspend(struct device *dev) { struct i2c_client *client = to_i2c_client(dev); @@ -2783,13 +2781,6 @@ static int smiapp_resume(struct device *dev) return rval; } -#else - -#define smiapp_suspend NULL -#define smiapp_resume NULL - -#endif /* CONFIG_PM */ - static struct smiapp_hwconfig *smiapp_get_hwconfig(struct device *dev) { struct smiapp_hwconfig *hwcfg; @@ -2913,13 +2904,9 @@ static int smiapp_probe(struct i2c_client *client, if (IS_ERR(sensor->xshutdown)) return PTR_ERR(sensor->xshutdown); - pm_runtime_enable(&client->dev); - - rval = pm_runtime_get_sync(&client->dev); - if (rval < 0) { - rval = -ENODEV; - goto out_power_off; - } + rval = smiapp_power_on(&client->dev); + if (rval < 0) + return rval; rval = smiapp_identify_module(sensor); if (rval) { @@ -3100,6 +3087,9 @@ static int smiapp_probe(struct i2c_client *client, if (rval < 0) goto out_media_entity_cleanup; + pm_runtime_set_active(&client->dev); + pm_runtime_get_noresume(&client->dev); + pm_runtime_enable(&client->dev); pm_runtime_set_autosuspend_delay(&client->dev, 1000); pm_runtime_use_autosuspend(&client->dev); pm_runtime_put_autosuspend(&client->dev); @@ -3113,8 +3103,7 @@ out_cleanup: smiapp_cleanup(sensor); out_power_off: - pm_runtime_put(&client->dev); - pm_runtime_disable(&client->dev); + smiapp_power_off(&client->dev); return rval; } @@ -3127,8 +3116,10 @@ static int smiapp_remove(struct i2c_client *client) v4l2_async_unregister_subdev(subdev); - pm_runtime_suspend(&client->dev); pm_runtime_disable(&client->dev); + if (!pm_runtime_status_suspended(&client->dev)) + smiapp_power_off(&client->dev); + pm_runtime_set_suspended(&client->dev); for (i = 0; i < sensor->ssds_used; i++) { v4l2_device_unregister_subdev(&sensor->ssds[i].sd); From 4bfb934b0067b7f6a24470682c5f7254fd4d8282 Mon Sep 17 00:00:00 2001 From: Sakari Ailus Date: Sat, 19 Nov 2016 19:50:10 -0200 Subject: [PATCH 002/343] [media] smiapp: Make suspend and resume functions __maybe_unused The smiapp_suspend() and smiapp_resume() functions will end up being unused if CONFIG_PM is enabled but CONFIG_PM_SLEEP is disabled, causing a compiler warning from both of the function definitions. Fix this by marking the functions with __maybe_unused. Suggested-by: Arnd Bergmann Signed-off-by: Sakari Ailus Signed-off-by: Mauro Carvalho Chehab --- drivers/media/i2c/smiapp/smiapp-core.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/media/i2c/smiapp/smiapp-core.c b/drivers/media/i2c/smiapp/smiapp-core.c index 620f8ce8185e..f4e92bdfe192 100644 --- a/drivers/media/i2c/smiapp/smiapp-core.c +++ b/drivers/media/i2c/smiapp/smiapp-core.c @@ -2741,7 +2741,7 @@ static const struct v4l2_subdev_internal_ops smiapp_internal_ops = { * I2C Driver */ -static int smiapp_suspend(struct device *dev) +static int __maybe_unused smiapp_suspend(struct device *dev) { struct i2c_client *client = to_i2c_client(dev); struct v4l2_subdev *subdev = i2c_get_clientdata(client); @@ -2766,7 +2766,7 @@ static int smiapp_suspend(struct device *dev) return 0; } -static int smiapp_resume(struct device *dev) +static int __maybe_unused smiapp_resume(struct device *dev) { struct i2c_client *client = to_i2c_client(dev); struct v4l2_subdev *subdev = i2c_get_clientdata(client); From f3854973f196baad5be6b62d8f5ea24b0346b63f Mon Sep 17 00:00:00 2001 From: Hans Verkuil Date: Tue, 6 Dec 2016 11:17:12 -0200 Subject: [PATCH 003/343] [media] cec: fix report_current_latency In the (very) small print of the REPORT_CURRENT_LATENCY message there is a line that says that the last byte of the message (audio out delay) is only present if the 'audio out compensated' value is 3. I missed this, and so if this message was sent with a total length of 6 (i.e. without the audio out delay byte), then it was rejected by the framework since a minimum length of 7 was expected. Fix this minimum length check and update the wrappers in cec-funcs.h to do the right thing based on the message length. Signed-off-by: Hans Verkuil Signed-off-by: Mauro Carvalho Chehab --- drivers/media/cec/cec-adap.c | 2 +- include/uapi/linux/cec-funcs.h | 10 +++++++--- 2 files changed, 8 insertions(+), 4 deletions(-) diff --git a/drivers/media/cec/cec-adap.c b/drivers/media/cec/cec-adap.c index 0ea4efb3de66..f15f6ffd75b2 100644 --- a/drivers/media/cec/cec-adap.c +++ b/drivers/media/cec/cec-adap.c @@ -851,7 +851,7 @@ static const u8 cec_msg_size[256] = { [CEC_MSG_REQUEST_ARC_TERMINATION] = 2 | DIRECTED, [CEC_MSG_TERMINATE_ARC] = 2 | DIRECTED, [CEC_MSG_REQUEST_CURRENT_LATENCY] = 4 | BCAST, - [CEC_MSG_REPORT_CURRENT_LATENCY] = 7 | BCAST, + [CEC_MSG_REPORT_CURRENT_LATENCY] = 6 | BCAST, [CEC_MSG_CDC_MESSAGE] = 2 | BCAST, }; diff --git a/include/uapi/linux/cec-funcs.h b/include/uapi/linux/cec-funcs.h index 3cbc327801d6..c451eec42a83 100644 --- a/include/uapi/linux/cec-funcs.h +++ b/include/uapi/linux/cec-funcs.h @@ -1665,14 +1665,15 @@ static inline void cec_msg_report_current_latency(struct cec_msg *msg, __u8 audio_out_compensated, __u8 audio_out_delay) { - msg->len = 7; + msg->len = 6; msg->msg[0] |= 0xf; /* broadcast */ msg->msg[1] = CEC_MSG_REPORT_CURRENT_LATENCY; msg->msg[2] = phys_addr >> 8; msg->msg[3] = phys_addr & 0xff; msg->msg[4] = video_latency; msg->msg[5] = (low_latency_mode << 2) | audio_out_compensated; - msg->msg[6] = audio_out_delay; + if (audio_out_compensated == 3) + msg->msg[msg->len++] = audio_out_delay; } static inline void cec_ops_report_current_latency(const struct cec_msg *msg, @@ -1686,7 +1687,10 @@ static inline void cec_ops_report_current_latency(const struct cec_msg *msg, *video_latency = msg->msg[4]; *low_latency_mode = (msg->msg[5] >> 2) & 1; *audio_out_compensated = msg->msg[5] & 3; - *audio_out_delay = msg->msg[6]; + if (*audio_out_compensated == 3 && msg->len >= 7) + *audio_out_delay = msg->msg[6]; + else + *audio_out_delay = 0; } static inline void cec_msg_request_current_latency(struct cec_msg *msg, From 120476123646ba3619c90db7bcbc6f8eea53c990 Mon Sep 17 00:00:00 2001 From: Hans Verkuil Date: Fri, 9 Dec 2016 11:14:32 -0200 Subject: [PATCH 004/343] [media] cec: when canceling a message, don't overwrite old status info When a pending message was canceled (e.g. due to a timeout), then the old tx_status info was overwritten instead of ORed. The same happened with the tx_error_cnt field. So just modify them instead of overwriting them. Signed-off-by: Hans Verkuil Signed-off-by: Mauro Carvalho Chehab --- drivers/media/cec/cec-adap.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/media/cec/cec-adap.c b/drivers/media/cec/cec-adap.c index f15f6ffd75b2..3191c0c5eae1 100644 --- a/drivers/media/cec/cec-adap.c +++ b/drivers/media/cec/cec-adap.c @@ -288,10 +288,10 @@ static void cec_data_cancel(struct cec_data *data) /* Mark it as an error */ data->msg.tx_ts = ktime_get_ns(); - data->msg.tx_status = CEC_TX_STATUS_ERROR | - CEC_TX_STATUS_MAX_RETRIES; + data->msg.tx_status |= CEC_TX_STATUS_ERROR | + CEC_TX_STATUS_MAX_RETRIES; + data->msg.tx_error_cnt++; data->attempts = 0; - data->msg.tx_error_cnt = 1; /* Queue transmitted message for monitoring purposes */ cec_queue_msg_monitor(data->adap, &data->msg, 1); From a24f56d47930492c94ef6875bf45adf7607ca1a4 Mon Sep 17 00:00:00 2001 From: Hans Verkuil Date: Fri, 9 Dec 2016 11:28:19 -0200 Subject: [PATCH 005/343] [media] cec: CEC_MSG_GIVE_FEATURES should abort for CEC version < 2 This is a 2.0 only message, so it should return Feature Abort if the adapter is configured for CEC version 1.4. Right now it does nothing, which means that the sender will time out. Signed-off-by: Hans Verkuil Signed-off-by: Mauro Carvalho Chehab --- drivers/media/cec/cec-adap.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/media/cec/cec-adap.c b/drivers/media/cec/cec-adap.c index 3191c0c5eae1..c05956fdd500 100644 --- a/drivers/media/cec/cec-adap.c +++ b/drivers/media/cec/cec-adap.c @@ -1777,9 +1777,9 @@ static int cec_receive_notify(struct cec_adapter *adap, struct cec_msg *msg, } case CEC_MSG_GIVE_FEATURES: - if (adap->log_addrs.cec_version >= CEC_OP_CEC_VERSION_2_0) - return cec_report_features(adap, la_idx); - return 0; + if (adap->log_addrs.cec_version < CEC_OP_CEC_VERSION_2_0) + return cec_feature_abort(adap, msg); + return cec_report_features(adap, la_idx); default: /* From 7af26f889eb67db272021a939f7d4a57e96dd961 Mon Sep 17 00:00:00 2001 From: Hans Verkuil Date: Fri, 9 Dec 2016 11:54:06 -0200 Subject: [PATCH 006/343] [media] cec: update log_addr[] before finishing configuration The loop that sets the unused logical addresses to INVALID should be done before 'configured' is set to true. This ensures that cec_log_addrs is consistent before it will be used. Signed-off-by: Hans Verkuil Signed-off-by: Mauro Carvalho Chehab --- drivers/media/cec/cec-adap.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/media/cec/cec-adap.c b/drivers/media/cec/cec-adap.c index c05956fdd500..f3fef487f28d 100644 --- a/drivers/media/cec/cec-adap.c +++ b/drivers/media/cec/cec-adap.c @@ -1250,6 +1250,8 @@ configured: for (i = 1; i < las->num_log_addrs; i++) las->log_addr[i] = CEC_LOG_ADDR_INVALID; } + for (i = las->num_log_addrs; i < CEC_MAX_LOG_ADDRS; i++) + las->log_addr[i] = CEC_LOG_ADDR_INVALID; adap->is_configured = true; adap->is_configuring = false; cec_post_state_event(adap); @@ -1268,8 +1270,6 @@ configured: cec_report_features(adap, i); cec_report_phys_addr(adap, i); } - for (i = las->num_log_addrs; i < CEC_MAX_LOG_ADDRS; i++) - las->log_addr[i] = CEC_LOG_ADDR_INVALID; mutex_lock(&adap->lock); adap->kthread_config = NULL; mutex_unlock(&adap->lock); From 52bc30fda9622f492427d484bd4dd8ee42cc4667 Mon Sep 17 00:00:00 2001 From: Hans Verkuil Date: Fri, 9 Dec 2016 11:36:03 -0200 Subject: [PATCH 007/343] [media] cec: replace cec_report_features by cec_fill_msg_report_features The fill function just fills in the cec_msg struct, it doesn't transmit the message. This is now done explicitly. This makes it possible to switch to transmitting this message with adap->lock held. Signed-off-by: Hans Verkuil Signed-off-by: Mauro Carvalho Chehab --- drivers/media/cec/cec-adap.c | 48 +++++++++++++++++++----------------- 1 file changed, 26 insertions(+), 22 deletions(-) diff --git a/drivers/media/cec/cec-adap.c b/drivers/media/cec/cec-adap.c index f3fef487f28d..2b668510ca36 100644 --- a/drivers/media/cec/cec-adap.c +++ b/drivers/media/cec/cec-adap.c @@ -30,8 +30,10 @@ #include "cec-priv.h" -static int cec_report_features(struct cec_adapter *adap, unsigned int la_idx); static int cec_report_phys_addr(struct cec_adapter *adap, unsigned int la_idx); +static void cec_fill_msg_report_features(struct cec_adapter *adap, + struct cec_msg *msg, + unsigned int la_idx); /* * 400 ms is the time it takes for one 16 byte message to be @@ -1258,16 +1260,21 @@ configured: mutex_unlock(&adap->lock); for (i = 0; i < las->num_log_addrs; i++) { + struct cec_msg msg = {}; + if (las->log_addr[i] == CEC_LOG_ADDR_INVALID || (las->flags & CEC_LOG_ADDRS_FL_CDC_ONLY)) continue; - /* - * Report Features must come first according - * to CEC 2.0 - */ - if (las->log_addr[i] != CEC_LOG_ADDR_UNREGISTERED) - cec_report_features(adap, i); + msg.msg[0] = (las->log_addr[i] << 4) | 0x0f; + + /* Report Features must come first according to CEC 2.0 */ + if (las->log_addr[i] != CEC_LOG_ADDR_UNREGISTERED && + adap->log_addrs.cec_version >= CEC_OP_CEC_VERSION_2_0) { + cec_fill_msg_report_features(adap, &msg, i); + cec_transmit_msg(adap, &msg, false); + } + cec_report_phys_addr(adap, i); } mutex_lock(&adap->lock); @@ -1526,36 +1533,32 @@ EXPORT_SYMBOL_GPL(cec_s_log_addrs); /* High-level core CEC message handling */ -/* Transmit the Report Features message */ -static int cec_report_features(struct cec_adapter *adap, unsigned int la_idx) +/* Fill in the Report Features message */ +static void cec_fill_msg_report_features(struct cec_adapter *adap, + struct cec_msg *msg, + unsigned int la_idx) { - struct cec_msg msg = { }; const struct cec_log_addrs *las = &adap->log_addrs; const u8 *features = las->features[la_idx]; bool op_is_dev_features = false; unsigned int idx; - /* This is 2.0 and up only */ - if (adap->log_addrs.cec_version < CEC_OP_CEC_VERSION_2_0) - return 0; - /* Report Features */ - msg.msg[0] = (las->log_addr[la_idx] << 4) | 0x0f; - msg.len = 4; - msg.msg[1] = CEC_MSG_REPORT_FEATURES; - msg.msg[2] = adap->log_addrs.cec_version; - msg.msg[3] = las->all_device_types[la_idx]; + msg->msg[0] = (las->log_addr[la_idx] << 4) | 0x0f; + msg->len = 4; + msg->msg[1] = CEC_MSG_REPORT_FEATURES; + msg->msg[2] = adap->log_addrs.cec_version; + msg->msg[3] = las->all_device_types[la_idx]; /* Write RC Profiles first, then Device Features */ for (idx = 0; idx < ARRAY_SIZE(las->features[0]); idx++) { - msg.msg[msg.len++] = features[idx]; + msg->msg[msg->len++] = features[idx]; if ((features[idx] & CEC_OP_FEAT_EXT) == 0) { if (op_is_dev_features) break; op_is_dev_features = true; } } - return cec_transmit_msg(adap, &msg, false); } /* Transmit the Report Physical Address message */ @@ -1779,7 +1782,8 @@ static int cec_receive_notify(struct cec_adapter *adap, struct cec_msg *msg, case CEC_MSG_GIVE_FEATURES: if (adap->log_addrs.cec_version < CEC_OP_CEC_VERSION_2_0) return cec_feature_abort(adap, msg); - return cec_report_features(adap, la_idx); + cec_fill_msg_report_features(adap, &tx_cec_msg, la_idx); + return cec_transmit_msg(adap, &tx_cec_msg, false); default: /* From d3d64bc7408f1ff0b0ff8354056e2a48eda5886d Mon Sep 17 00:00:00 2001 From: Hans Verkuil Date: Fri, 9 Dec 2016 11:48:34 -0200 Subject: [PATCH 008/343] [media] cec: move cec_report_phys_addr into cec_config_thread_func It's only a small function and this makes it easier to switch to transmitting the message with adap->lock held in the next patch. Signed-off-by: Hans Verkuil Signed-off-by: Mauro Carvalho Chehab --- drivers/media/cec/cec-adap.c | 25 +++++++------------------ 1 file changed, 7 insertions(+), 18 deletions(-) diff --git a/drivers/media/cec/cec-adap.c b/drivers/media/cec/cec-adap.c index 2b668510ca36..f3d495654a53 100644 --- a/drivers/media/cec/cec-adap.c +++ b/drivers/media/cec/cec-adap.c @@ -30,7 +30,6 @@ #include "cec-priv.h" -static int cec_report_phys_addr(struct cec_adapter *adap, unsigned int la_idx); static void cec_fill_msg_report_features(struct cec_adapter *adap, struct cec_msg *msg, unsigned int la_idx); @@ -1275,7 +1274,13 @@ configured: cec_transmit_msg(adap, &msg, false); } - cec_report_phys_addr(adap, i); + /* Report Physical Address */ + cec_msg_report_physical_addr(&msg, adap->phys_addr, + las->primary_device_type[i]); + dprintk(2, "config: la %d pa %x.%x.%x.%x\n", + las->log_addr[i], + cec_phys_addr_exp(adap->phys_addr)); + cec_transmit_msg(adap, &msg, false); } mutex_lock(&adap->lock); adap->kthread_config = NULL; @@ -1561,22 +1566,6 @@ static void cec_fill_msg_report_features(struct cec_adapter *adap, } } -/* Transmit the Report Physical Address message */ -static int cec_report_phys_addr(struct cec_adapter *adap, unsigned int la_idx) -{ - const struct cec_log_addrs *las = &adap->log_addrs; - struct cec_msg msg = { }; - - /* Report Physical Address */ - msg.msg[0] = (las->log_addr[la_idx] << 4) | 0x0f; - cec_msg_report_physical_addr(&msg, adap->phys_addr, - las->primary_device_type[la_idx]); - dprintk(2, "config: la %d pa %x.%x.%x.%x\n", - las->log_addr[la_idx], - cec_phys_addr_exp(adap->phys_addr)); - return cec_transmit_msg(adap, &msg, false); -} - /* Transmit the Feature Abort message */ static int cec_feature_abort_reason(struct cec_adapter *adap, struct cec_msg *msg, u8 reason) From f60f35609f89ef4fee73776bc1ef697923251995 Mon Sep 17 00:00:00 2001 From: Hans Verkuil Date: Fri, 9 Dec 2016 12:00:49 -0200 Subject: [PATCH 009/343] [media] cec: fix race between configuring and unconfiguring This race was discovered by running cec-compliance -A with the cec module debug parameter set to 2: suddenly the test would fail. It turns out that this happens when the test configures the adapter in non-blocking mode, then it waits for the CEC_EVENT_STATE_CHANGE event and once the event is received it unconfigures the adapter. What happened was that the unconfigure was executed while the configure was still transmitting the Report Features and Report Physical Address messages. This messed up the internal state of the cec_adapter. The fix is to transmit those messages with the adap->lock mutex held (this will just queue them up in the internal transmit queue, and not actually transmit anything yet). Only unlock the mutex once everything is done. The main thread will dequeue the messages from the internal transmit queue and transmit them one by one, unless an unconfigure was done, and in that case any messages are just dropped. Signed-off-by: Hans Verkuil Signed-off-by: Mauro Carvalho Chehab --- drivers/media/cec/cec-adap.c | 18 +++++++++++++----- 1 file changed, 13 insertions(+), 5 deletions(-) diff --git a/drivers/media/cec/cec-adap.c b/drivers/media/cec/cec-adap.c index f3d495654a53..ebb5e391b800 100644 --- a/drivers/media/cec/cec-adap.c +++ b/drivers/media/cec/cec-adap.c @@ -1256,8 +1256,17 @@ configured: adap->is_configured = true; adap->is_configuring = false; cec_post_state_event(adap); - mutex_unlock(&adap->lock); + /* + * Now post the Report Features and Report Physical Address broadcast + * messages. Note that these are non-blocking transmits, meaning that + * they are just queued up and once adap->lock is unlocked the main + * thread will kick in and start transmitting these. + * + * If after this function is done (but before one or more of these + * messages are actually transmitted) the CEC adapter is unconfigured, + * then any remaining messages will be dropped by the main thread. + */ for (i = 0; i < las->num_log_addrs; i++) { struct cec_msg msg = {}; @@ -1271,7 +1280,7 @@ configured: if (las->log_addr[i] != CEC_LOG_ADDR_UNREGISTERED && adap->log_addrs.cec_version >= CEC_OP_CEC_VERSION_2_0) { cec_fill_msg_report_features(adap, &msg, i); - cec_transmit_msg(adap, &msg, false); + cec_transmit_msg_fh(adap, &msg, NULL, false); } /* Report Physical Address */ @@ -1280,12 +1289,11 @@ configured: dprintk(2, "config: la %d pa %x.%x.%x.%x\n", las->log_addr[i], cec_phys_addr_exp(adap->phys_addr)); - cec_transmit_msg(adap, &msg, false); + cec_transmit_msg_fh(adap, &msg, NULL, false); } - mutex_lock(&adap->lock); adap->kthread_config = NULL; - mutex_unlock(&adap->lock); complete(&adap->config_completion); + mutex_unlock(&adap->lock); return 0; unconfigure: From 78ccbf9ff89bd7a20d36be039cb3eab71081648c Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Sun, 11 Sep 2016 10:31:28 -0300 Subject: [PATCH 010/343] [media] media/cobalt: use pci_irq_allocate_vectors Simply the interrupt setup by using the new PCI layer helpers. Despite using pci_enable_msi_range, this driver was only requesting a single MSI vector anyway. Signed-off-by: Christoph Hellwig Signed-off-by: Hans Verkuil Signed-off-by: Mauro Carvalho Chehab --- drivers/media/pci/cobalt/cobalt-driver.c | 8 ++------ drivers/media/pci/cobalt/cobalt-driver.h | 2 -- 2 files changed, 2 insertions(+), 8 deletions(-) diff --git a/drivers/media/pci/cobalt/cobalt-driver.c b/drivers/media/pci/cobalt/cobalt-driver.c index 979634000597..d5c911c09e2b 100644 --- a/drivers/media/pci/cobalt/cobalt-driver.c +++ b/drivers/media/pci/cobalt/cobalt-driver.c @@ -308,9 +308,7 @@ static void cobalt_pci_iounmap(struct cobalt *cobalt, struct pci_dev *pci_dev) static void cobalt_free_msi(struct cobalt *cobalt, struct pci_dev *pci_dev) { free_irq(pci_dev->irq, (void *)cobalt); - - if (cobalt->msi_enabled) - pci_disable_msi(pci_dev); + pci_free_irq_vectors(pci_dev); } static int cobalt_setup_pci(struct cobalt *cobalt, struct pci_dev *pci_dev, @@ -387,14 +385,12 @@ static int cobalt_setup_pci(struct cobalt *cobalt, struct pci_dev *pci_dev, from being generated. */ cobalt_set_interrupt(cobalt, false); - if (pci_enable_msi_range(pci_dev, 1, 1) < 1) { + if (pci_alloc_irq_vectors(pci_dev, 1, 1, PCI_IRQ_MSI) < 1) { cobalt_err("Could not enable MSI\n"); - cobalt->msi_enabled = false; ret = -EIO; goto err_release; } msi_config_show(cobalt, pci_dev); - cobalt->msi_enabled = true; /* Register IRQ */ if (request_irq(pci_dev->irq, cobalt_irq_handler, IRQF_SHARED, diff --git a/drivers/media/pci/cobalt/cobalt-driver.h b/drivers/media/pci/cobalt/cobalt-driver.h index ed00dc9d9399..00f773ec359a 100644 --- a/drivers/media/pci/cobalt/cobalt-driver.h +++ b/drivers/media/pci/cobalt/cobalt-driver.h @@ -287,8 +287,6 @@ struct cobalt { u32 irq_none; u32 irq_full_fifo; - bool msi_enabled; - /* omnitek dma */ int dma_channels; int first_fifo_channel; From 48775cb73c2e26b7ca9d679875a6e570c8b8e124 Mon Sep 17 00:00:00 2001 From: Max Kellermann Date: Thu, 15 Dec 2016 19:51:07 -0200 Subject: [PATCH 011/343] [media] pctv452e: move buffer to heap, no mutex commit 73d5c5c864f4 ("[media] pctv452e: don't do DMA on stack") caused a NULL pointer dereference which occurs when dvb_usb_init() calls dvb_usb_device_power_ctrl() for the first time, before the frontend has been attached. It also caused a recursive deadlock because tt3650_ci_msg_locked() has already locked the mutex. So, partially revert it, but move the buffer to the heap (DMA capable), not to the stack (may not be DMA capable). Instead of sharing one buffer which needs mutex protection, do a new heap allocation for each call. Fixes: commit 73d5c5c864f4 ("[media] pctv452e: don't do DMA on stack") Cc: stable@vger.kernel.org # For Kernel 4.9 Signed-off-by: Max Kellermann Signed-off-by: Mauro Carvalho Chehab --- drivers/media/usb/dvb-usb/pctv452e.c | 133 +++++++++++++++------------ 1 file changed, 72 insertions(+), 61 deletions(-) diff --git a/drivers/media/usb/dvb-usb/pctv452e.c b/drivers/media/usb/dvb-usb/pctv452e.c index 07fa08be9e99..d54ebe7e0215 100644 --- a/drivers/media/usb/dvb-usb/pctv452e.c +++ b/drivers/media/usb/dvb-usb/pctv452e.c @@ -97,14 +97,13 @@ struct pctv452e_state { u8 c; /* transaction counter, wraps around... */ u8 initialized; /* set to 1 if 0x15 has been sent */ u16 last_rc_key; - - unsigned char data[80]; }; static int tt3650_ci_msg(struct dvb_usb_device *d, u8 cmd, u8 *data, unsigned int write_len, unsigned int read_len) { struct pctv452e_state *state = (struct pctv452e_state *)d->priv; + u8 *buf; u8 id; unsigned int rlen; int ret; @@ -114,36 +113,39 @@ static int tt3650_ci_msg(struct dvb_usb_device *d, u8 cmd, u8 *data, return -EIO; } - mutex_lock(&state->ca_mutex); + buf = kmalloc(64, GFP_KERNEL); + if (!buf) + return -ENOMEM; + id = state->c++; - state->data[0] = SYNC_BYTE_OUT; - state->data[1] = id; - state->data[2] = cmd; - state->data[3] = write_len; + buf[0] = SYNC_BYTE_OUT; + buf[1] = id; + buf[2] = cmd; + buf[3] = write_len; - memcpy(state->data + 4, data, write_len); + memcpy(buf + 4, data, write_len); rlen = (read_len > 0) ? 64 : 0; - ret = dvb_usb_generic_rw(d, state->data, 4 + write_len, - state->data, rlen, /* delay_ms */ 0); + ret = dvb_usb_generic_rw(d, buf, 4 + write_len, + buf, rlen, /* delay_ms */ 0); if (0 != ret) goto failed; ret = -EIO; - if (SYNC_BYTE_IN != state->data[0] || id != state->data[1]) + if (SYNC_BYTE_IN != buf[0] || id != buf[1]) goto failed; - memcpy(data, state->data + 4, read_len); + memcpy(data, buf + 4, read_len); - mutex_unlock(&state->ca_mutex); + kfree(buf); return 0; failed: err("CI error %d; %02X %02X %02X -> %*ph.", - ret, SYNC_BYTE_OUT, id, cmd, 3, state->data); + ret, SYNC_BYTE_OUT, id, cmd, 3, buf); - mutex_unlock(&state->ca_mutex); + kfree(buf); return ret; } @@ -410,53 +412,57 @@ static int pctv452e_i2c_msg(struct dvb_usb_device *d, u8 addr, u8 *rcv_buf, u8 rcv_len) { struct pctv452e_state *state = (struct pctv452e_state *)d->priv; + u8 *buf; u8 id; int ret; - mutex_lock(&state->ca_mutex); + buf = kmalloc(64, GFP_KERNEL); + if (!buf) + return -ENOMEM; + id = state->c++; ret = -EINVAL; if (snd_len > 64 - 7 || rcv_len > 64 - 7) goto failed; - state->data[0] = SYNC_BYTE_OUT; - state->data[1] = id; - state->data[2] = PCTV_CMD_I2C; - state->data[3] = snd_len + 3; - state->data[4] = addr << 1; - state->data[5] = snd_len; - state->data[6] = rcv_len; + buf[0] = SYNC_BYTE_OUT; + buf[1] = id; + buf[2] = PCTV_CMD_I2C; + buf[3] = snd_len + 3; + buf[4] = addr << 1; + buf[5] = snd_len; + buf[6] = rcv_len; - memcpy(state->data + 7, snd_buf, snd_len); + memcpy(buf + 7, snd_buf, snd_len); - ret = dvb_usb_generic_rw(d, state->data, 7 + snd_len, - state->data, /* rcv_len */ 64, + ret = dvb_usb_generic_rw(d, buf, 7 + snd_len, + buf, /* rcv_len */ 64, /* delay_ms */ 0); if (ret < 0) goto failed; /* TT USB protocol error. */ ret = -EIO; - if (SYNC_BYTE_IN != state->data[0] || id != state->data[1]) + if (SYNC_BYTE_IN != buf[0] || id != buf[1]) goto failed; /* I2C device didn't respond as expected. */ ret = -EREMOTEIO; - if (state->data[5] < snd_len || state->data[6] < rcv_len) + if (buf[5] < snd_len || buf[6] < rcv_len) goto failed; - memcpy(rcv_buf, state->data + 7, rcv_len); - mutex_unlock(&state->ca_mutex); + memcpy(rcv_buf, buf + 7, rcv_len); + kfree(buf); return rcv_len; failed: err("I2C error %d; %02X %02X %02X %02X %02X -> %*ph", ret, SYNC_BYTE_OUT, id, addr << 1, snd_len, rcv_len, - 7, state->data); + 7, buf); - mutex_unlock(&state->ca_mutex); + kfree(buf); return ret; } @@ -505,7 +511,7 @@ static u32 pctv452e_i2c_func(struct i2c_adapter *adapter) static int pctv452e_power_ctrl(struct dvb_usb_device *d, int i) { struct pctv452e_state *state = (struct pctv452e_state *)d->priv; - u8 *rx; + u8 *b0, *rx; int ret; info("%s: %d\n", __func__, i); @@ -516,11 +522,12 @@ static int pctv452e_power_ctrl(struct dvb_usb_device *d, int i) if (state->initialized) return 0; - rx = kmalloc(PCTV_ANSWER_LEN, GFP_KERNEL); - if (!rx) + b0 = kmalloc(5 + PCTV_ANSWER_LEN, GFP_KERNEL); + if (!b0) return -ENOMEM; - mutex_lock(&state->ca_mutex); + rx = b0 + 5; + /* hmm where shoud this should go? */ ret = usb_set_interface(d->udev, 0, ISOC_INTERFACE_ALTERNATIVE); if (ret != 0) @@ -528,66 +535,70 @@ static int pctv452e_power_ctrl(struct dvb_usb_device *d, int i) __func__, ret); /* this is a one-time initialization, dont know where to put */ - state->data[0] = 0xaa; - state->data[1] = state->c++; - state->data[2] = PCTV_CMD_RESET; - state->data[3] = 1; - state->data[4] = 0; + b0[0] = 0xaa; + b0[1] = state->c++; + b0[2] = PCTV_CMD_RESET; + b0[3] = 1; + b0[4] = 0; /* reset board */ - ret = dvb_usb_generic_rw(d, state->data, 5, rx, PCTV_ANSWER_LEN, 0); + ret = dvb_usb_generic_rw(d, b0, 5, rx, PCTV_ANSWER_LEN, 0); if (ret) goto ret; - state->data[1] = state->c++; - state->data[4] = 1; + b0[1] = state->c++; + b0[4] = 1; /* reset board (again?) */ - ret = dvb_usb_generic_rw(d, state->data, 5, rx, PCTV_ANSWER_LEN, 0); + ret = dvb_usb_generic_rw(d, b0, 5, rx, PCTV_ANSWER_LEN, 0); if (ret) goto ret; state->initialized = 1; ret: - mutex_unlock(&state->ca_mutex); - kfree(rx); + kfree(b0); return ret; } static int pctv452e_rc_query(struct dvb_usb_device *d) { struct pctv452e_state *state = (struct pctv452e_state *)d->priv; + u8 *b, *rx; int ret, i; u8 id; - mutex_lock(&state->ca_mutex); + b = kmalloc(CMD_BUFFER_SIZE + PCTV_ANSWER_LEN, GFP_KERNEL); + if (!b) + return -ENOMEM; + + rx = b + CMD_BUFFER_SIZE; + id = state->c++; /* prepare command header */ - state->data[0] = SYNC_BYTE_OUT; - state->data[1] = id; - state->data[2] = PCTV_CMD_IR; - state->data[3] = 0; + b[0] = SYNC_BYTE_OUT; + b[1] = id; + b[2] = PCTV_CMD_IR; + b[3] = 0; /* send ir request */ - ret = dvb_usb_generic_rw(d, state->data, 4, - state->data, PCTV_ANSWER_LEN, 0); + ret = dvb_usb_generic_rw(d, b, 4, rx, PCTV_ANSWER_LEN, 0); if (ret != 0) goto ret; if (debug > 3) { - info("%s: read: %2d: %*ph: ", __func__, ret, 3, state->data); - for (i = 0; (i < state->data[3]) && ((i + 3) < PCTV_ANSWER_LEN); i++) - info(" %02x", state->data[i + 3]); + info("%s: read: %2d: %*ph: ", __func__, ret, 3, rx); + for (i = 0; (i < rx[3]) && ((i+3) < PCTV_ANSWER_LEN); i++) + info(" %02x", rx[i+3]); info("\n"); } - if ((state->data[3] == 9) && (state->data[12] & 0x01)) { + if ((rx[3] == 9) && (rx[12] & 0x01)) { /* got a "press" event */ - state->last_rc_key = RC_SCANCODE_RC5(state->data[7], state->data[6]); + state->last_rc_key = RC_SCANCODE_RC5(rx[7], rx[6]); if (debug > 2) info("%s: cmd=0x%02x sys=0x%02x\n", - __func__, state->data[6], state->data[7]); + __func__, rx[6], rx[7]); rc_keydown(d->rc_dev, RC_TYPE_RC5, state->last_rc_key, 0); } else if (state->last_rc_key) { @@ -595,7 +606,7 @@ static int pctv452e_rc_query(struct dvb_usb_device *d) state->last_rc_key = 0; } ret: - mutex_unlock(&state->ca_mutex); + kfree(b); return ret; } From aff808e813fc2d311137754165cf53d4ee6ddcc2 Mon Sep 17 00:00:00 2001 From: Laurent Pinchart Date: Fri, 9 Dec 2016 09:47:17 -0200 Subject: [PATCH 012/343] [media] v4l: tvp5150: Reset device at probe time, not in get/set format handlers The tvp5150 doesn't support format setting through the subdev pad API and thus implements the set format handler as a get format operation. The single handler, tvp5150_fill_fmt(), resets the device by calling tvp5150_reset(). This causes malfunction as the device can be reset at will, possibly from userspace when the subdev userspace API is enabled. The reset call was added in commit ec2c4f3f93cb ("[media] media: tvp5150: Add mbus_fmt callbacks"), probably as an attempt to set the device to a known state before detecting the current TV standard. However, the get format handler doesn't access the hardware to get the TV standard since commit 963ddc63e20d ("[media] media: tvp5150: Add cropping support"). There is thus no need to reset the device when getting the format. However, removing the tvp5150_reset() from the get/set format handlers results in the function not being called at all if the bridge driver doesn't use the .reset() operation. The operation is nowadays abused and shouldn't be used, so shouldn't expect bridge drivers to call it. To make sure the device is properly initialize, move the reset call from the format handlers to the probe function. Cc: stable@vger.kernel.org # For Kernel 4.5 and upper Signed-off-by: Laurent Pinchart Signed-off-by: Mauro Carvalho Chehab --- drivers/media/i2c/tvp5150.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/drivers/media/i2c/tvp5150.c b/drivers/media/i2c/tvp5150.c index 3a0fe8cc64e9..a30bfcb4eec6 100644 --- a/drivers/media/i2c/tvp5150.c +++ b/drivers/media/i2c/tvp5150.c @@ -861,8 +861,6 @@ static int tvp5150_fill_fmt(struct v4l2_subdev *sd, f = &format->format; - tvp5150_reset(sd, 0); - f->width = decoder->rect.width; f->height = decoder->rect.height / 2; @@ -1524,7 +1522,6 @@ static int tvp5150_probe(struct i2c_client *c, res = core->hdl.error; goto err; } - v4l2_ctrl_handler_setup(&core->hdl); /* Default is no cropping */ core->rect.top = 0; @@ -1535,6 +1532,8 @@ static int tvp5150_probe(struct i2c_client *c, core->rect.left = 0; core->rect.width = TVP5150_H_MAX; + tvp5150_reset(sd, 0); /* Calls v4l2_ctrl_handler_setup() */ + res = v4l2_async_register_subdev(sd); if (res < 0) goto err; From b4b2de386bbb6589d81596999d4a924928dc119b Mon Sep 17 00:00:00 2001 From: Laurent Pinchart Date: Fri, 9 Dec 2016 09:47:18 -0200 Subject: [PATCH 013/343] [media] v4l: tvp5150: Fix comment regarding output pin muxing The FID/GLCO/VLK/HVLK and INTREQ/GPCL/VBLK pins are muxed differently depending on whether the input is an S-Video or composite signal. The comment that explains the logic doesn't reflect the code. It appears that the comment is incorrect, as disabling the output data bus in composite mode makes no sense. Update the comment to match the code. While at it define macros for the MISC_CTL register bits, the code is too confusing with numerical values. Cc: stable@vger.kernel.org # For Kernel 4.5 and upper Signed-off-by: Laurent Pinchart Signed-off-by: Mauro Carvalho Chehab --- drivers/media/i2c/tvp5150.c | 24 +++++++++++++++++------- drivers/media/i2c/tvp5150_reg.h | 9 +++++++++ 2 files changed, 26 insertions(+), 7 deletions(-) diff --git a/drivers/media/i2c/tvp5150.c b/drivers/media/i2c/tvp5150.c index a30bfcb4eec6..8852fa8c957b 100644 --- a/drivers/media/i2c/tvp5150.c +++ b/drivers/media/i2c/tvp5150.c @@ -291,8 +291,12 @@ static void tvp5150_selmux(struct v4l2_subdev *sd) tvp5150_write(sd, TVP5150_OP_MODE_CTL, opmode); tvp5150_write(sd, TVP5150_VD_IN_SRC_SEL_1, input); - /* Svideo should enable YCrCb output and disable GPCL output - * For Composite and TV, it should be the reverse + /* + * Setup the FID/GLCO/VLK/HVLK and INTREQ/GPCL/VBLK output signals. For + * S-Video we output the vertical lock (VLK) signal on FID/GLCO/VLK/HVLK + * and set INTREQ/GPCL/VBLK to logic 0. For composite we output the + * field indicator (FID) signal on FID/GLCO/VLK/HVLK and set + * INTREQ/GPCL/VBLK to logic 1. */ val = tvp5150_read(sd, TVP5150_MISC_CTL); if (val < 0) { @@ -301,9 +305,9 @@ static void tvp5150_selmux(struct v4l2_subdev *sd) } if (decoder->input == TVP5150_SVIDEO) - val = (val & ~0x40) | 0x10; + val = (val & ~TVP5150_MISC_CTL_GPCL) | TVP5150_MISC_CTL_HVLK; else - val = (val & ~0x10) | 0x40; + val = (val & ~TVP5150_MISC_CTL_HVLK) | TVP5150_MISC_CTL_GPCL; tvp5150_write(sd, TVP5150_MISC_CTL, val); }; @@ -455,7 +459,12 @@ static const struct i2c_reg_value tvp5150_init_enable[] = { },{ /* Automatic offset and AGC enabled */ TVP5150_ANAL_CHL_CTL, 0x15 },{ /* Activate YCrCb output 0x9 or 0xd ? */ - TVP5150_MISC_CTL, 0x6f + TVP5150_MISC_CTL, TVP5150_MISC_CTL_GPCL | + TVP5150_MISC_CTL_INTREQ_OE | + TVP5150_MISC_CTL_YCBCR_OE | + TVP5150_MISC_CTL_SYNC_OE | + TVP5150_MISC_CTL_VBLANK | + TVP5150_MISC_CTL_CLOCK_OE, },{ /* Activates video std autodetection for all standards */ TVP5150_AUTOSW_MSK, 0x0 },{ /* Default format: 0x47. For 4:2:2: 0x40 */ @@ -1050,11 +1059,12 @@ static int tvp5150_s_stream(struct v4l2_subdev *sd, int enable) { struct tvp5150 *decoder = to_tvp5150(sd); /* Output format: 8-bit ITU-R BT.656 with embedded syncs */ - int val = 0x09; + int val = TVP5150_MISC_CTL_YCBCR_OE | TVP5150_MISC_CTL_CLOCK_OE; /* Output format: 8-bit 4:2:2 YUV with discrete sync */ if (decoder->mbus_type == V4L2_MBUS_PARALLEL) - val = 0x0d; + val = TVP5150_MISC_CTL_YCBCR_OE | TVP5150_MISC_CTL_SYNC_OE + | TVP5150_MISC_CTL_CLOCK_OE; /* Initializes TVP5150 to its default values */ /* # set PCLK (27MHz) */ diff --git a/drivers/media/i2c/tvp5150_reg.h b/drivers/media/i2c/tvp5150_reg.h index 25a994944918..30a48c28d05a 100644 --- a/drivers/media/i2c/tvp5150_reg.h +++ b/drivers/media/i2c/tvp5150_reg.h @@ -9,6 +9,15 @@ #define TVP5150_ANAL_CHL_CTL 0x01 /* Analog channel controls */ #define TVP5150_OP_MODE_CTL 0x02 /* Operation mode controls */ #define TVP5150_MISC_CTL 0x03 /* Miscellaneous controls */ +#define TVP5150_MISC_CTL_VBLK_GPCL BIT(7) +#define TVP5150_MISC_CTL_GPCL BIT(6) +#define TVP5150_MISC_CTL_INTREQ_OE BIT(5) +#define TVP5150_MISC_CTL_HVLK BIT(4) +#define TVP5150_MISC_CTL_YCBCR_OE BIT(3) +#define TVP5150_MISC_CTL_SYNC_OE BIT(2) +#define TVP5150_MISC_CTL_VBLANK BIT(1) +#define TVP5150_MISC_CTL_CLOCK_OE BIT(0) + #define TVP5150_AUTOSW_MSK 0x04 /* Autoswitch mask: TVP5150A / TVP5150AM */ /* Reserved 05h */ From 79d6205a3f741c9fb89cfc47dfa0eddb1526726d Mon Sep 17 00:00:00 2001 From: Laurent Pinchart Date: Fri, 9 Dec 2016 09:47:19 -0200 Subject: [PATCH 014/343] [media] v4l: tvp5150: Don't override output pinmuxing at stream on/off time The s_stream() handler incorrectly writes the whole MISC_CTL register to enable or disable the outputs, overriding the output pinmuxing configuration. Fix it to only touch the output enable bits. The CONF_SHARED_PIN register is also written by the same function, resulting in muxing the INTREQ signal instead of the VBLK/GPCL signal on the INTREQ/GPCL/VBLK pin. As the driver doesn't support interrupts this is obviously incorrect, and breaks operation on other devices. Fix it by removing the write. Cc: stable@vger.kernel.org # For Kernel 4.5 and upper Signed-off-by: Laurent Pinchart Signed-off-by: Mauro Carvalho Chehab --- drivers/media/i2c/tvp5150.c | 31 ++++++++++++++++++------------- 1 file changed, 18 insertions(+), 13 deletions(-) diff --git a/drivers/media/i2c/tvp5150.c b/drivers/media/i2c/tvp5150.c index 8852fa8c957b..48646a7f3fb0 100644 --- a/drivers/media/i2c/tvp5150.c +++ b/drivers/media/i2c/tvp5150.c @@ -1058,22 +1058,27 @@ static const struct media_entity_operations tvp5150_sd_media_ops = { static int tvp5150_s_stream(struct v4l2_subdev *sd, int enable) { struct tvp5150 *decoder = to_tvp5150(sd); - /* Output format: 8-bit ITU-R BT.656 with embedded syncs */ - int val = TVP5150_MISC_CTL_YCBCR_OE | TVP5150_MISC_CTL_CLOCK_OE; + int val; - /* Output format: 8-bit 4:2:2 YUV with discrete sync */ - if (decoder->mbus_type == V4L2_MBUS_PARALLEL) - val = TVP5150_MISC_CTL_YCBCR_OE | TVP5150_MISC_CTL_SYNC_OE - | TVP5150_MISC_CTL_CLOCK_OE; + /* Enable or disable the video output signals. */ + val = tvp5150_read(sd, TVP5150_MISC_CTL); + if (val < 0) + return val; - /* Initializes TVP5150 to its default values */ - /* # set PCLK (27MHz) */ - tvp5150_write(sd, TVP5150_CONF_SHARED_PIN, 0x00); + val &= ~(TVP5150_MISC_CTL_YCBCR_OE | TVP5150_MISC_CTL_SYNC_OE | + TVP5150_MISC_CTL_CLOCK_OE); - if (enable) - tvp5150_write(sd, TVP5150_MISC_CTL, val); - else - tvp5150_write(sd, TVP5150_MISC_CTL, 0x00); + if (enable) { + /* + * Enable the YCbCr and clock outputs. In discrete sync mode + * (non-BT.656) additionally enable the the sync outputs. + */ + val |= TVP5150_MISC_CTL_YCBCR_OE | TVP5150_MISC_CTL_CLOCK_OE; + if (decoder->mbus_type == V4L2_MBUS_PARALLEL) + val |= TVP5150_MISC_CTL_SYNC_OE; + } + + tvp5150_write(sd, TVP5150_MISC_CTL, val); return 0; } From 4dd19196c5539c377beaa9850fac30c18318c7a1 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Fri, 9 Dec 2016 09:36:29 -0200 Subject: [PATCH 015/343] [media] dvb: avoid warning in dvb_net With gcc-5 or higher on x86, we can get a bogus warning in the dvb-net code: drivers/media/dvb-core/dvb_net.c: In function 'dvb_net_ule': arch/x86/include/asm/string_32.h:78:22: error: '*((void *)&dest_addr+4)' may be used uninitialized in this function [-Werror=maybe-uninitialized] The problem here is that gcc doesn't track all of the conditions to prove it can't end up copying uninitialized data. This changes the logic around so we zero out the destination address earlier when we determine that it is not set here. This allows the compiler to figure it out. Signed-off-by: Arnd Bergmann Signed-off-by: Mauro Carvalho Chehab --- drivers/media/dvb-core/dvb_net.c | 15 +++++---------- 1 file changed, 5 insertions(+), 10 deletions(-) diff --git a/drivers/media/dvb-core/dvb_net.c b/drivers/media/dvb-core/dvb_net.c index bd833b0824c6..eb60cb1442f2 100644 --- a/drivers/media/dvb-core/dvb_net.c +++ b/drivers/media/dvb-core/dvb_net.c @@ -719,6 +719,9 @@ static void dvb_net_ule_check_crc(struct dvb_net_ule_handle *h, skb_copy_from_linear_data(h->priv->ule_skb, dest_addr, ETH_ALEN); skb_pull(h->priv->ule_skb, ETH_ALEN); + } else { + /* dest_addr buffer is only valid if h->priv->ule_dbit == 0 */ + eth_zero_addr(dest_addr); } /* Handle ULE Extension Headers. */ @@ -750,16 +753,8 @@ static void dvb_net_ule_check_crc(struct dvb_net_ule_handle *h, if (!h->priv->ule_bridged) { skb_push(h->priv->ule_skb, ETH_HLEN); h->ethh = (struct ethhdr *)h->priv->ule_skb->data; - if (!h->priv->ule_dbit) { - /* - * dest_addr buffer is only valid if - * h->priv->ule_dbit == 0 - */ - memcpy(h->ethh->h_dest, dest_addr, ETH_ALEN); - eth_zero_addr(h->ethh->h_source); - } else /* zeroize source and dest */ - memset(h->ethh, 0, ETH_ALEN * 2); - + memcpy(h->ethh->h_dest, dest_addr, ETH_ALEN); + eth_zero_addr(h->ethh->h_source); h->ethh->h_proto = htons(h->priv->ule_sndu_type); } /* else: skb is in correct state; nothing to do. */ From c739c0a7c3c2472d7562b8f802cdce44d2597c8b Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Fri, 9 Dec 2016 09:41:29 -0200 Subject: [PATCH 016/343] [media] s5k4ecgx: select CRC32 helper A rare randconfig build failure shows up in this driver when the CRC32 helper is not there: drivers/media/built-in.o: In function `s5k4ecgx_s_power': s5k4ecgx.c:(.text+0x9eb4): undefined reference to `crc32_le' This adds the 'select' that all other users of this function have. Fixes: 8b99312b7214 ("[media] Add v4l2 subdev driver for S5K4ECGX sensor") Signed-off-by: Arnd Bergmann Signed-off-by: Mauro Carvalho Chehab --- drivers/media/i2c/Kconfig | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/media/i2c/Kconfig b/drivers/media/i2c/Kconfig index b31fa6fae009..b979ea148251 100644 --- a/drivers/media/i2c/Kconfig +++ b/drivers/media/i2c/Kconfig @@ -655,6 +655,7 @@ config VIDEO_S5K6A3 config VIDEO_S5K4ECGX tristate "Samsung S5K4ECGX sensor support" depends on I2C && VIDEO_V4L2 && VIDEO_V4L2_SUBDEV_API + select CRC32 ---help--- This is a V4L2 sensor-level driver for Samsung S5K4ECGX 5M camera sensor with an embedded SoC image signal processor. From 4ea33ef0f9e95b69db9131d7afd98563713e81b0 Mon Sep 17 00:00:00 2001 From: Sven Eckelmann Date: Tue, 27 Dec 2016 08:51:17 +0100 Subject: [PATCH 017/343] batman-adv: Decrease hardif refcnt on fragmentation send error An error before the hardif is found has to free the skb. But every error after that has to free the skb + put the hard interface. Fixes: 8def0be82dd1 ("batman-adv: Consume skb in batadv_frag_send_packet") Signed-off-by: Sven Eckelmann Signed-off-by: Simon Wunderlich --- net/batman-adv/fragmentation.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/net/batman-adv/fragmentation.c b/net/batman-adv/fragmentation.c index 9c561e683f4b..0854ebd8613e 100644 --- a/net/batman-adv/fragmentation.c +++ b/net/batman-adv/fragmentation.c @@ -474,7 +474,7 @@ int batadv_frag_send_packet(struct sk_buff *skb, primary_if = batadv_primary_if_get_selected(bat_priv); if (!primary_if) { ret = -EINVAL; - goto put_primary_if; + goto free_skb; } /* Create one header to be copied to all fragments */ @@ -502,7 +502,7 @@ int batadv_frag_send_packet(struct sk_buff *skb, skb_fragment = batadv_frag_create(skb, &frag_header, mtu); if (!skb_fragment) { ret = -ENOMEM; - goto free_skb; + goto put_primary_if; } batadv_inc_counter(bat_priv, BATADV_CNT_FRAG_TX); @@ -511,7 +511,7 @@ int batadv_frag_send_packet(struct sk_buff *skb, ret = batadv_send_unicast_skb(skb_fragment, neigh_node); if (ret != NET_XMIT_SUCCESS) { ret = NET_XMIT_DROP; - goto free_skb; + goto put_primary_if; } frag_header.no++; @@ -519,7 +519,7 @@ int batadv_frag_send_packet(struct sk_buff *skb, /* The initial check in this function should cover this case */ if (frag_header.no == BATADV_FRAG_MAX_FRAGMENTS - 1) { ret = -EINVAL; - goto free_skb; + goto put_primary_if; } } @@ -527,7 +527,7 @@ int batadv_frag_send_packet(struct sk_buff *skb, if (batadv_skb_head_push(skb, header_size) < 0 || pskb_expand_head(skb, header_size + ETH_HLEN, 0, GFP_ATOMIC) < 0) { ret = -ENOMEM; - goto free_skb; + goto put_primary_if; } memcpy(skb->data, &frag_header, header_size); From fd7c99142d77dc4a851879a66715abf12a3193fb Mon Sep 17 00:00:00 2001 From: Dave Martin Date: Fri, 6 Jan 2017 17:54:51 +0000 Subject: [PATCH 018/343] tile/ptrace: Preserve previous registers for short regset write Ensure that if userspace supplies insufficient data to PTRACE_SETREGSET to fill all the registers, the thread's old registers are preserved. Cc: stable@vger.kernel.org Signed-off-by: Dave Martin Signed-off-by: Chris Metcalf --- arch/tile/kernel/ptrace.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/tile/kernel/ptrace.c b/arch/tile/kernel/ptrace.c index d89b7011667c..e279572824b1 100644 --- a/arch/tile/kernel/ptrace.c +++ b/arch/tile/kernel/ptrace.c @@ -111,7 +111,7 @@ static int tile_gpr_set(struct task_struct *target, const void *kbuf, const void __user *ubuf) { int ret; - struct pt_regs regs; + struct pt_regs regs = *task_pt_regs(target); ret = user_regset_copyin(&pos, &count, &kbuf, &ubuf, ®s, 0, sizeof(regs)); From 3e70c5d6ea510e38f612d07fa0fd7487277b7087 Mon Sep 17 00:00:00 2001 From: Nicolas Iooss Date: Mon, 26 Dec 2016 14:52:23 +0100 Subject: [PATCH 019/343] drm/i915/gvt: verify functions types in new_mmio_info() The current prototype of new_mmio_info() uses void* for parameters read and write, which are functions with precise calling conventions (argument types and return type). Write down these conventions in new_mmio_info() definition. This has been reported by the following warnings when clang is used to build the kernel: drivers/gpu/drm/i915/gvt/handlers.c:124:21: error: pointer type mismatch ('void *' and 'int (*)(struct intel_vgpu *, unsigned int, void *, unsigned int)') [-Werror,-Wpointer-type-mismatch] info->read = read ? read : intel_vgpu_default_mmio_read; ^ ~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~ drivers/gpu/drm/i915/gvt/handlers.c:125:23: error: pointer type mismatch ('void *' and 'int (*)(struct intel_vgpu *, unsigned int, void *, unsigned int)') [-Werror,-Wpointer-type-mismatch] info->write = write ? write : intel_vgpu_default_mmio_write; ^ ~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ This allows the compiler to detect that sbi_ctl_mmio_write() returns a "bool" value instead of an expected "int" one. Fix this. Signed-off-by: Nicolas Iooss Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/handlers.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/gvt/handlers.c b/drivers/gpu/drm/i915/gvt/handlers.c index 522809710312..052e57124c0a 100644 --- a/drivers/gpu/drm/i915/gvt/handlers.c +++ b/drivers/gpu/drm/i915/gvt/handlers.c @@ -93,7 +93,8 @@ static void write_vreg(struct intel_vgpu *vgpu, unsigned int offset, static int new_mmio_info(struct intel_gvt *gvt, u32 offset, u32 flags, u32 size, u32 addr_mask, u32 ro_mask, u32 device, - void *read, void *write) + int (*read)(struct intel_vgpu *, unsigned int, void *, unsigned int), + int (*write)(struct intel_vgpu *, unsigned int, void *, unsigned int)) { struct intel_gvt_mmio_info *info, *p; u32 start, end, i; @@ -974,7 +975,7 @@ static int sbi_data_mmio_read(struct intel_vgpu *vgpu, unsigned int offset, return 0; } -static bool sbi_ctl_mmio_write(struct intel_vgpu *vgpu, unsigned int offset, +static int sbi_ctl_mmio_write(struct intel_vgpu *vgpu, unsigned int offset, void *p_data, unsigned int bytes) { u32 data; From a12010534d0984f91bc5bdcf9e27bd55e20d82da Mon Sep 17 00:00:00 2001 From: Changbin Du Date: Tue, 27 Dec 2016 13:24:52 +0800 Subject: [PATCH 020/343] drm/i915/gvt: fix error handing of tlb_control emulation Return ealier for a invalid access, else it would false set tlb flag for RCS. Signed-off-by: Changbin Du Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/handlers.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/i915/gvt/handlers.c b/drivers/gpu/drm/i915/gvt/handlers.c index 052e57124c0a..e06d5f37bb92 100644 --- a/drivers/gpu/drm/i915/gvt/handlers.c +++ b/drivers/gpu/drm/i915/gvt/handlers.c @@ -1367,7 +1367,6 @@ static int ring_mode_mmio_write(struct intel_vgpu *vgpu, unsigned int offset, static int gvt_reg_tlb_control_handler(struct intel_vgpu *vgpu, unsigned int offset, void *p_data, unsigned int bytes) { - int rc = 0; unsigned int id = 0; write_vreg(vgpu, offset, p_data, bytes); @@ -1390,12 +1389,11 @@ static int gvt_reg_tlb_control_handler(struct intel_vgpu *vgpu, id = VECS; break; default: - rc = -EINVAL; - break; + return -EINVAL; } set_bit(id, (void *)vgpu->tlb_handle_pending); - return rc; + return 0; } static int ring_reset_ctl_write(struct intel_vgpu *vgpu, From 39762ad437f1149b904e6baeaf28824da34a89c1 Mon Sep 17 00:00:00 2001 From: Changbin Du Date: Tue, 27 Dec 2016 13:25:06 +0800 Subject: [PATCH 021/343] drm/i915/gvt: fix return value in mul_force_wake_write All mmio handlers should return a negetive value for failure, not 1. Signed-off-by: Changbin Du Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/handlers.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/gvt/handlers.c b/drivers/gpu/drm/i915/gvt/handlers.c index e06d5f37bb92..8cbaf1c83720 100644 --- a/drivers/gpu/drm/i915/gvt/handlers.c +++ b/drivers/gpu/drm/i915/gvt/handlers.c @@ -220,7 +220,7 @@ static int mul_force_wake_write(struct intel_vgpu *vgpu, default: /*should not hit here*/ gvt_err("invalid forcewake offset 0x%x\n", offset); - return 1; + return -EINVAL; } } else { ack_reg_offset = FORCEWAKE_ACK_HSW_REG; From 905a5035ebe79e89712cda0bed1887c73aa8e9bb Mon Sep 17 00:00:00 2001 From: Changbin Du Date: Fri, 30 Dec 2016 14:10:53 +0800 Subject: [PATCH 022/343] drm/i915/gvt: always use readq and writeq The readq and writeq are already offered by drm_os_linux.h. So we can use them directly whithout dectecting their presence. This patch removed the duplicated code. Signed-off-by: Changbin Du Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/gtt.c | 15 ++------------- 1 file changed, 2 insertions(+), 13 deletions(-) diff --git a/drivers/gpu/drm/i915/gvt/gtt.c b/drivers/gpu/drm/i915/gvt/gtt.c index 6c5fdf5b2ce2..a32e59de0eff 100644 --- a/drivers/gpu/drm/i915/gvt/gtt.c +++ b/drivers/gpu/drm/i915/gvt/gtt.c @@ -240,15 +240,8 @@ static inline int get_pse_type(int type) static u64 read_pte64(struct drm_i915_private *dev_priv, unsigned long index) { void __iomem *addr = (gen8_pte_t __iomem *)dev_priv->ggtt.gsm + index; - u64 pte; -#ifdef readq - pte = readq(addr); -#else - pte = ioread32(addr); - pte |= (u64)ioread32(addr + 4) << 32; -#endif - return pte; + return readq(addr); } static void write_pte64(struct drm_i915_private *dev_priv, @@ -256,12 +249,8 @@ static void write_pte64(struct drm_i915_private *dev_priv, { void __iomem *addr = (gen8_pte_t __iomem *)dev_priv->ggtt.gsm + index; -#ifdef writeq writeq(pte, addr); -#else - iowrite32((u32)pte, addr); - iowrite32(pte >> 32, addr + 4); -#endif + I915_WRITE(GFX_FLSH_CNTL_GEN6, GFX_FLSH_CNTL_EN); POSTING_READ(GFX_FLSH_CNTL_GEN6); } From 901a14b721feef1b37cfe6362ee103e135133677 Mon Sep 17 00:00:00 2001 From: Pei Zhang Date: Wed, 4 Jan 2017 22:32:23 +0800 Subject: [PATCH 023/343] drm/i915/gvt: print correct value for untracked mmio In function intel_vgpu_emulate_mmio_read, the untracked mmio register is dumped through kernel log, but the register value is not correct. This patch fixes this issue. V2: fix the fromat warning from checkpatch.pl. Signed-off-by: Pei Zhang Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/mmio.c | 31 ++++++++++++++++--------------- 1 file changed, 16 insertions(+), 15 deletions(-) diff --git a/drivers/gpu/drm/i915/gvt/mmio.c b/drivers/gpu/drm/i915/gvt/mmio.c index 09c9450a1946..e60701397ac2 100644 --- a/drivers/gpu/drm/i915/gvt/mmio.c +++ b/drivers/gpu/drm/i915/gvt/mmio.c @@ -125,25 +125,12 @@ int intel_vgpu_emulate_mmio_read(struct intel_vgpu *vgpu, uint64_t pa, if (WARN_ON(!reg_is_mmio(gvt, offset + bytes - 1))) goto err; - mmio = intel_gvt_find_mmio_info(gvt, rounddown(offset, 4)); - if (!mmio && !vgpu->mmio.disable_warn_untrack) { - gvt_err("vgpu%d: read untracked MMIO %x len %d val %x\n", - vgpu->id, offset, bytes, *(u32 *)p_data); - - if (offset == 0x206c) { - gvt_err("------------------------------------------\n"); - gvt_err("vgpu%d: likely triggers a gfx reset\n", - vgpu->id); - gvt_err("------------------------------------------\n"); - vgpu->mmio.disable_warn_untrack = true; - } - } - if (!intel_gvt_mmio_is_unalign(gvt, offset)) { if (WARN_ON(!IS_ALIGNED(offset, bytes))) goto err; } + mmio = intel_gvt_find_mmio_info(gvt, rounddown(offset, 4)); if (mmio) { if (!intel_gvt_mmio_is_unalign(gvt, mmio->offset)) { if (WARN_ON(offset + bytes > mmio->offset + mmio->size)) @@ -152,9 +139,23 @@ int intel_vgpu_emulate_mmio_read(struct intel_vgpu *vgpu, uint64_t pa, goto err; } ret = mmio->read(vgpu, offset, p_data, bytes); - } else + } else { ret = intel_vgpu_default_mmio_read(vgpu, offset, p_data, bytes); + if (!vgpu->mmio.disable_warn_untrack) { + gvt_err("vgpu%d: read untracked MMIO %x(%dB) val %x\n", + vgpu->id, offset, bytes, *(u32 *)p_data); + + if (offset == 0x206c) { + gvt_err("------------------------------------------\n"); + gvt_err("vgpu%d: likely triggers a gfx reset\n", + vgpu->id); + gvt_err("------------------------------------------\n"); + vgpu->mmio.disable_warn_untrack = true; + } + } + } + if (ret) goto err; From 888530b57f88f2bc856f181479df732c9622fa22 Mon Sep 17 00:00:00 2001 From: Zhenyu Wang Date: Thu, 5 Jan 2017 10:26:13 +0800 Subject: [PATCH 024/343] drm/i915/gvt: adjust high memory size for default vGPU type Previous high mem size initialized for vGPU type was too small which caused failure for some VMs. This trys to take minimal value of 384MB for each VM and enlarge default high mem size to make guest driver happy. Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/vgpu.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/gvt/vgpu.c b/drivers/gpu/drm/i915/gvt/vgpu.c index 536d2b9d5777..398abb98dd0a 100644 --- a/drivers/gpu/drm/i915/gvt/vgpu.c +++ b/drivers/gpu/drm/i915/gvt/vgpu.c @@ -177,7 +177,7 @@ int intel_gvt_init_vgpu_types(struct intel_gvt *gvt) if (low_avail / min_low == 0) break; gvt->types[i].low_gm_size = min_low; - gvt->types[i].high_gm_size = 3 * gvt->types[i].low_gm_size; + gvt->types[i].high_gm_size = max((min_low<<3), MB_TO_BYTES(384U)); gvt->types[i].fence = 4; gvt->types[i].max_instance = low_avail / min_low; gvt->types[i].avail_instance = gvt->types[i].max_instance; @@ -217,7 +217,7 @@ static void intel_gvt_update_vgpu_types(struct intel_gvt *gvt) */ low_gm_avail = MB_TO_BYTES(256) - HOST_LOW_GM_SIZE - gvt->gm.vgpu_allocated_low_gm_size; - high_gm_avail = MB_TO_BYTES(256) * 3 - HOST_HIGH_GM_SIZE - + high_gm_avail = MB_TO_BYTES(256) * 8UL - HOST_HIGH_GM_SIZE - gvt->gm.vgpu_allocated_high_gm_size; fence_avail = gvt_fence_sz(gvt) - HOST_FENCE - gvt->fence.vgpu_allocated_fence_num; From 2fcdb66364ee467d69228a3d2ea074498c177211 Mon Sep 17 00:00:00 2001 From: Zhenyu Wang Date: Thu, 5 Jan 2017 10:26:24 +0800 Subject: [PATCH 025/343] drm/i915/gvt: remove duplicated definition Remove duplicated definition for resource size in aperture_gm.c which are already defined in gvt.h. Need only one to take effect. Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/aperture_gm.c | 7 ------- 1 file changed, 7 deletions(-) diff --git a/drivers/gpu/drm/i915/gvt/aperture_gm.c b/drivers/gpu/drm/i915/gvt/aperture_gm.c index 0d41ebc4aea6..65200313515c 100644 --- a/drivers/gpu/drm/i915/gvt/aperture_gm.c +++ b/drivers/gpu/drm/i915/gvt/aperture_gm.c @@ -37,13 +37,6 @@ #include "i915_drv.h" #include "gvt.h" -#define MB_TO_BYTES(mb) ((mb) << 20ULL) -#define BYTES_TO_MB(b) ((b) >> 20ULL) - -#define HOST_LOW_GM_SIZE MB_TO_BYTES(128) -#define HOST_HIGH_GM_SIZE MB_TO_BYTES(384) -#define HOST_FENCE 4 - static int alloc_gm(struct intel_vgpu *vgpu, bool high_gm) { struct intel_gvt *gvt = vgpu->gvt; From 2e51ef32b0d66fcd5fe45c437cf7c6aef8350746 Mon Sep 17 00:00:00 2001 From: Changbin Du Date: Thu, 5 Jan 2017 13:28:05 +0800 Subject: [PATCH 026/343] drm/i915/gvt: fix use after free for workload In the function workload_thread(), we invoke complete_current_workload() to cleanup the just processed workload (workload will be freed there). So we cannot access workload->req after that. This patch move complete_current_workload() afterward. Signed-off-by: Changbin Du Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/scheduler.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/gvt/scheduler.c b/drivers/gpu/drm/i915/gvt/scheduler.c index 4db242250235..c694dd039f3b 100644 --- a/drivers/gpu/drm/i915/gvt/scheduler.c +++ b/drivers/gpu/drm/i915/gvt/scheduler.c @@ -459,11 +459,11 @@ complete: gvt_dbg_sched("will complete workload %p\n, status: %d\n", workload, workload->status); - complete_current_workload(gvt, ring_id); - if (workload->req) i915_gem_request_put(fetch_and_zero(&workload->req)); + complete_current_workload(gvt, ring_id); + if (need_force_wake) intel_uncore_forcewake_put(gvt->dev_priv, FORCEWAKE_ALL); From 440a9b9fae37dfd7e4c7d76db34fada57f9afd92 Mon Sep 17 00:00:00 2001 From: Changbin Du Date: Thu, 5 Jan 2017 16:49:03 +0800 Subject: [PATCH 027/343] drm/i915/gvt: dec vgpu->running_workload_num after the workload is really done The vgpu->running_workload_num is used to determine whether a vgpu has any workload running or not. So we should make sure the workload is really done before we dec running_workload_num. Function complete_current_workload is not the right place to do it, since this function is still processing the workload. This patch move the dec op afterward. v2: move dec op before wake_up(&scheduler->workload_complete_wq) (Min He) Signed-off-by: Changbin Du Reviewed-by: Min He Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/scheduler.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/i915/gvt/scheduler.c b/drivers/gpu/drm/i915/gvt/scheduler.c index c694dd039f3b..e91885dffeff 100644 --- a/drivers/gpu/drm/i915/gvt/scheduler.c +++ b/drivers/gpu/drm/i915/gvt/scheduler.c @@ -350,13 +350,15 @@ static void complete_current_workload(struct intel_gvt *gvt, int ring_id) { struct intel_gvt_workload_scheduler *scheduler = &gvt->scheduler; struct intel_vgpu_workload *workload; + struct intel_vgpu *vgpu; int event; mutex_lock(&gvt->lock); workload = scheduler->current_workload[ring_id]; + vgpu = workload->vgpu; - if (!workload->status && !workload->vgpu->resetting) { + if (!workload->status && !vgpu->resetting) { wait_event(workload->shadow_ctx_status_wq, !atomic_read(&workload->shadow_ctx_active)); @@ -364,8 +366,7 @@ static void complete_current_workload(struct intel_gvt *gvt, int ring_id) for_each_set_bit(event, workload->pending_events, INTEL_GVT_EVENT_MAX) - intel_vgpu_trigger_virtual_event(workload->vgpu, - event); + intel_vgpu_trigger_virtual_event(vgpu, event); } gvt_dbg_sched("ring id %d complete workload %p status %d\n", @@ -373,11 +374,10 @@ static void complete_current_workload(struct intel_gvt *gvt, int ring_id) scheduler->current_workload[ring_id] = NULL; - atomic_dec(&workload->vgpu->running_workload_num); - list_del_init(&workload->list); workload->complete(workload); + atomic_dec(&vgpu->running_workload_num); wake_up(&scheduler->workload_complete_wq); mutex_unlock(&gvt->lock); } From 59c0573dfbd5f66e3aa54c2ce0bebcb0953d4db4 Mon Sep 17 00:00:00 2001 From: Jike Song Date: Fri, 6 Jan 2017 15:16:21 +0800 Subject: [PATCH 028/343] drm/i915/gvt: init/destroy vgpu_idr properly An idr should be initialized before use and destroyed afterwards. Signed-off-by: Jike Song Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/gvt.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/gvt/gvt.c b/drivers/gpu/drm/i915/gvt/gvt.c index 398877c3d2fd..e6bf5c533fbe 100644 --- a/drivers/gpu/drm/i915/gvt/gvt.c +++ b/drivers/gpu/drm/i915/gvt/gvt.c @@ -201,6 +201,8 @@ void intel_gvt_clean_device(struct drm_i915_private *dev_priv) intel_gvt_hypervisor_host_exit(&dev_priv->drm.pdev->dev, gvt); intel_gvt_clean_vgpu_types(gvt); + idr_destroy(&gvt->vgpu_idr); + kfree(dev_priv->gvt); dev_priv->gvt = NULL; } @@ -237,6 +239,8 @@ int intel_gvt_init_device(struct drm_i915_private *dev_priv) gvt_dbg_core("init gvt device\n"); + idr_init(&gvt->vgpu_idr); + mutex_init(&gvt->lock); gvt->dev_priv = dev_priv; @@ -244,7 +248,7 @@ int intel_gvt_init_device(struct drm_i915_private *dev_priv) ret = intel_gvt_setup_mmio_info(gvt); if (ret) - return ret; + goto out_clean_idr; ret = intel_gvt_load_firmware(gvt); if (ret) @@ -313,6 +317,8 @@ out_free_firmware: intel_gvt_free_firmware(gvt); out_clean_mmio_info: intel_gvt_clean_mmio_info(gvt); +out_clean_idr: + idr_destroy(&gvt->vgpu_idr); kfree(gvt); return ret; } From 4e5378918b5b96e6b93fcadf1ab84a8486ca60a1 Mon Sep 17 00:00:00 2001 From: Jike Song Date: Fri, 6 Jan 2017 15:16:22 +0800 Subject: [PATCH 029/343] drm/i915/gvt: destroy the allocated idr on vgpu creating failures Once idr_alloc gets called data is allocated within the idr list, if any error occurs afterwards, we should undo that by idr_remove on the error path. Signed-off-by: Jike Song Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/vgpu.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/gvt/vgpu.c b/drivers/gpu/drm/i915/gvt/vgpu.c index 398abb98dd0a..f0e86123e45b 100644 --- a/drivers/gpu/drm/i915/gvt/vgpu.c +++ b/drivers/gpu/drm/i915/gvt/vgpu.c @@ -304,7 +304,7 @@ static struct intel_vgpu *__intel_gvt_create_vgpu(struct intel_gvt *gvt, ret = setup_vgpu_mmio(vgpu); if (ret) - goto out_free_vgpu; + goto out_clean_idr; ret = intel_vgpu_alloc_resource(vgpu, param); if (ret) @@ -355,6 +355,8 @@ out_clean_vgpu_resource: intel_vgpu_free_resource(vgpu); out_clean_vgpu_mmio: clean_vgpu_mmio(vgpu); +out_clean_idr: + idr_remove(&gvt->vgpu_idr, vgpu->id); out_free_vgpu: vfree(vgpu); mutex_unlock(&gvt->lock); From 03551e971f6e52c8dedd5741bf48631e65675759 Mon Sep 17 00:00:00 2001 From: Jike Song Date: Fri, 6 Jan 2017 15:16:23 +0800 Subject: [PATCH 030/343] drm/i915/gvt: cleanup opregion memory allocation code According to the spec, ACPI OpRegion must be placed at a physical address below 4G. That is, for a vGPU it must be associated with a GPA below 4G, but on host side, it doesn't matter where the backing pages actually are. So when allocating pages from host, the GFP_DMA32 flag is unnecessary. Also the allocation is from a sleepable context, so GFP_ATOMIC is also unnecessary. This patch also removes INTEL_GVT_OPREGION_PORDER and use get_order() instead. Signed-off-by: Jike Song Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/opregion.c | 8 ++++---- drivers/gpu/drm/i915/gvt/reg.h | 3 +-- 2 files changed, 5 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/i915/gvt/opregion.c b/drivers/gpu/drm/i915/gvt/opregion.c index 81cd921770c6..d9fb41ab7119 100644 --- a/drivers/gpu/drm/i915/gvt/opregion.c +++ b/drivers/gpu/drm/i915/gvt/opregion.c @@ -36,9 +36,9 @@ static int init_vgpu_opregion(struct intel_vgpu *vgpu, u32 gpa) vgpu->id)) return -EINVAL; - vgpu_opregion(vgpu)->va = (void *)__get_free_pages(GFP_ATOMIC | - GFP_DMA32 | __GFP_ZERO, - INTEL_GVT_OPREGION_PORDER); + vgpu_opregion(vgpu)->va = (void *)__get_free_pages(GFP_KERNEL | + __GFP_ZERO, + get_order(INTEL_GVT_OPREGION_SIZE)); if (!vgpu_opregion(vgpu)->va) return -ENOMEM; @@ -97,7 +97,7 @@ void intel_vgpu_clean_opregion(struct intel_vgpu *vgpu) if (intel_gvt_host.hypervisor_type == INTEL_GVT_HYPERVISOR_XEN) { map_vgpu_opregion(vgpu, false); free_pages((unsigned long)vgpu_opregion(vgpu)->va, - INTEL_GVT_OPREGION_PORDER); + get_order(INTEL_GVT_OPREGION_SIZE)); vgpu_opregion(vgpu)->va = NULL; } diff --git a/drivers/gpu/drm/i915/gvt/reg.h b/drivers/gpu/drm/i915/gvt/reg.h index 0dfe789d8f02..fbd023a16f18 100644 --- a/drivers/gpu/drm/i915/gvt/reg.h +++ b/drivers/gpu/drm/i915/gvt/reg.h @@ -50,8 +50,7 @@ #define INTEL_GVT_OPREGION_PARM 0x204 #define INTEL_GVT_OPREGION_PAGES 2 -#define INTEL_GVT_OPREGION_PORDER 1 -#define INTEL_GVT_OPREGION_SIZE (2 * 4096) +#define INTEL_GVT_OPREGION_SIZE (INTEL_GVT_OPREGION_PAGES * PAGE_SIZE) #define VGT_SPRSTRIDE(pipe) _PIPE(pipe, _SPRA_STRIDE, _PLANE_STRIDE_2_B) From 5753394b64a07dd502cb288a5fd52e71fb01fc5d Mon Sep 17 00:00:00 2001 From: Jike Song Date: Fri, 6 Jan 2017 15:16:20 +0800 Subject: [PATCH 031/343] drm/i915/gvt/kvmgt: return meaningful error for vgpu creating failure The vgpu_create() routine we called returns meaningful errors to indicate failures, so we'd better to pass it to our caller, the mdev framework, whereby the sysfs is able to tell userspace what happened. Signed-off-by: Jike Song Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/kvmgt.c | 14 ++++++++++---- 1 file changed, 10 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/i915/gvt/kvmgt.c b/drivers/gpu/drm/i915/gvt/kvmgt.c index faaae07ae487..0c9234a87a20 100644 --- a/drivers/gpu/drm/i915/gvt/kvmgt.c +++ b/drivers/gpu/drm/i915/gvt/kvmgt.c @@ -398,6 +398,7 @@ static int intel_vgpu_create(struct kobject *kobj, struct mdev_device *mdev) struct intel_vgpu_type *type; struct device *pdev; void *gvt; + int ret; pdev = mdev_parent_dev(mdev); gvt = kdev_to_i915(pdev)->gvt; @@ -406,13 +407,15 @@ static int intel_vgpu_create(struct kobject *kobj, struct mdev_device *mdev) if (!type) { gvt_err("failed to find type %s to create\n", kobject_name(kobj)); - return -EINVAL; + ret = -EINVAL; + goto out; } vgpu = intel_gvt_ops->vgpu_create(gvt, type); if (IS_ERR_OR_NULL(vgpu)) { - gvt_err("create intel vgpu failed\n"); - return -EINVAL; + ret = vgpu == NULL ? -EFAULT : PTR_ERR(vgpu); + gvt_err("failed to create intel vgpu: %d\n", ret); + goto out; } INIT_WORK(&vgpu->vdev.release_work, intel_vgpu_release_work); @@ -422,7 +425,10 @@ static int intel_vgpu_create(struct kobject *kobj, struct mdev_device *mdev) gvt_dbg_core("intel_vgpu_create succeeded for mdev: %s\n", dev_name(mdev_dev(mdev))); - return 0; + ret = 0; + +out: + return ret; } static int intel_vgpu_remove(struct mdev_device *mdev) From f0a8b49c03d22a511a601dc54b2a3425a41e35fa Mon Sep 17 00:00:00 2001 From: Marek Szyprowski Date: Fri, 30 Dec 2016 10:57:46 +0100 Subject: [PATCH 032/343] drm/bridge: analogix dp: Fix runtime PM state on driver bind Analogix_dp_bind() can be called from component framework, which doesn't guarantee proper runtime PM state of the device during bind operation, so ensure that device is runtime active before doing any register access. This ensures that the power domain, to which DP module belongs, is turned on. While at it, also fix the unbalanced call to phy_power_on() in analogix_dp_bind() function. This patch solves the following kernel oops on Samsung Exynos5250 Snow board: Unhandled fault: imprecise external abort (0x406) at 0x00000000 pgd = c0004000 [00000000] *pgd=00000000 Internal error: : 406 [#1] PREEMPT SMP ARM Modules linked in: CPU: 0 PID: 75 Comm: kworker/0:2 Not tainted 4.9.0 #1046 Hardware name: SAMSUNG EXYNOS (Flattened Device Tree) Workqueue: events deferred_probe_work_func task: ee272300 task.stack: ee312000 PC is at analogix_dp_enable_sw_function+0x18/0x2c LR is at analogix_dp_init_dp+0x2c/0x50 ... [] (analogix_dp_enable_sw_function) from [] (analogix_dp_init_dp+0x2c/0x50) [] (analogix_dp_init_dp) from [] (analogix_dp_bind+0x184/0x42c) [] (analogix_dp_bind) from [] (component_bind_all+0xf0/0x218) [] (component_bind_all) from [] (exynos_drm_load+0x134/0x200) [] (exynos_drm_load) from [] (drm_dev_register+0xa0/0xd0) [] (drm_dev_register) from [] (drm_platform_init+0x58/0xb0) [] (drm_platform_init) from [] (try_to_bring_up_master+0x14c/0x188) [] (try_to_bring_up_master) from [] (component_add+0x88/0x138) [] (component_add) from [] (platform_drv_probe+0x50/0xb0) [] (platform_drv_probe) from [] (driver_probe_device+0x1f0/0x2a8) [] (driver_probe_device) from [] (bus_for_each_drv+0x44/0x8c) [] (bus_for_each_drv) from [] (__device_attach+0x9c/0x100) [] (__device_attach) from [] (bus_probe_device+0x84/0x8c) [] (bus_probe_device) from [] (deferred_probe_work_func+0x60/0x8c) [] (deferred_probe_work_func) from [] (process_one_work+0x120/0x318) [] (process_one_work) from [] (process_scheduled_works+0x28/0x38) [] (process_scheduled_works) from [] (worker_thread+0x204/0x4ac) [] (worker_thread) from [] (kthread+0xd8/0xf4) [] (kthread) from [] (ret_from_fork+0x14/0x3c) Code: e59035f0 e5935018 f57ff04f e3c55001 (f57ff04e) ---[ end trace 3d1d0d87796de344 ]--- Reviewed-by: Sean Paul Signed-off-by: Marek Szyprowski Signed-off-by: Archit Taneja Link: http://patchwork.freedesktop.org/patch/msgid/1483091866-1088-1-git-send-email-m.szyprowski@samsung.com --- drivers/gpu/drm/bridge/analogix/analogix_dp_core.c | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c b/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c index eb9bf8786c24..18eefdcbf1ba 100644 --- a/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c +++ b/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c @@ -1382,6 +1382,7 @@ int analogix_dp_bind(struct device *dev, struct drm_device *drm_dev, pm_runtime_enable(dev); + pm_runtime_get_sync(dev); phy_power_on(dp->phy); analogix_dp_init_dp(dp); @@ -1414,9 +1415,15 @@ int analogix_dp_bind(struct device *dev, struct drm_device *drm_dev, goto err_disable_pm_runtime; } + phy_power_off(dp->phy); + pm_runtime_put(dev); + return 0; err_disable_pm_runtime: + + phy_power_off(dp->phy); + pm_runtime_put(dev); pm_runtime_disable(dev); return ret; From 9631739f8196ec80b5d9bf955f79b711490c0205 Mon Sep 17 00:00:00 2001 From: Jike Song Date: Mon, 9 Jan 2017 15:38:38 +0800 Subject: [PATCH 033/343] drm/i915/gvt: cleanup GFP flags In gvt, almost all memory allocations are in sleepable contexts. It's fault-prone to use GFP_ATOMIC everywhere. Replace it with GFP_KERNEL wherever possible. Signed-off-by: Jike Song Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/gtt.c | 39 +++++++++++++--------------------- 1 file changed, 15 insertions(+), 24 deletions(-) diff --git a/drivers/gpu/drm/i915/gvt/gtt.c b/drivers/gpu/drm/i915/gvt/gtt.c index a32e59de0eff..3cf0df0bb391 100644 --- a/drivers/gpu/drm/i915/gvt/gtt.c +++ b/drivers/gpu/drm/i915/gvt/gtt.c @@ -1369,8 +1369,7 @@ static int gen8_mm_alloc_page_table(struct intel_vgpu_mm *mm) info->gtt_entry_size; mem = kzalloc(mm->has_shadow_page_table ? mm->page_table_entry_size * 2 - : mm->page_table_entry_size, - GFP_ATOMIC); + : mm->page_table_entry_size, GFP_KERNEL); if (!mem) return -ENOMEM; mm->virtual_page_table = mem; @@ -1521,7 +1520,7 @@ struct intel_vgpu_mm *intel_vgpu_create_mm(struct intel_vgpu *vgpu, struct intel_vgpu_mm *mm; int ret; - mm = kzalloc(sizeof(*mm), GFP_ATOMIC); + mm = kzalloc(sizeof(*mm), GFP_KERNEL); if (!mm) { ret = -ENOMEM; goto fail; @@ -1875,30 +1874,27 @@ static int alloc_scratch_pages(struct intel_vgpu *vgpu, struct intel_gvt_gtt_pte_ops *ops = vgpu->gvt->gtt.pte_ops; int page_entry_num = GTT_PAGE_SIZE >> vgpu->gvt->device_info.gtt_entry_size_shift; - struct page *scratch_pt; + void *scratch_pt; unsigned long mfn; int i; - void *p; if (WARN_ON(type < GTT_TYPE_PPGTT_PTE_PT || type >= GTT_TYPE_MAX)) return -EINVAL; - scratch_pt = alloc_page(GFP_KERNEL | GFP_ATOMIC | __GFP_ZERO); + scratch_pt = (void *)get_zeroed_page(GFP_KERNEL); if (!scratch_pt) { gvt_err("fail to allocate scratch page\n"); return -ENOMEM; } - p = kmap_atomic(scratch_pt); - mfn = intel_gvt_hypervisor_virt_to_mfn(p); + mfn = intel_gvt_hypervisor_virt_to_mfn(scratch_pt); if (mfn == INTEL_GVT_INVALID_ADDR) { - gvt_err("fail to translate vaddr:0x%llx\n", (u64)p); - kunmap_atomic(p); - __free_page(scratch_pt); + gvt_err("fail to translate vaddr:0x%lx\n", (unsigned long)scratch_pt); + free_page((unsigned long)scratch_pt); return -EFAULT; } gtt->scratch_pt[type].page_mfn = mfn; - gtt->scratch_pt[type].page = scratch_pt; + gtt->scratch_pt[type].page = virt_to_page(scratch_pt); gvt_dbg_mm("vgpu%d create scratch_pt: type %d mfn=0x%lx\n", vgpu->id, type, mfn); @@ -1907,7 +1903,7 @@ static int alloc_scratch_pages(struct intel_vgpu *vgpu, * scratch_pt[type] indicate the scratch pt/scratch page used by the * 'type' pt. * e.g. scratch_pt[GTT_TYPE_PPGTT_PDE_PT] is used by - * GTT_TYPE_PPGTT_PDE_PT level pt, that means this scatch_pt it self + * GTT_TYPE_PPGTT_PDE_PT level pt, that means this scratch_pt it self * is GTT_TYPE_PPGTT_PTE_PT, and full filled by scratch page mfn. */ if (type > GTT_TYPE_PPGTT_PTE_PT && type < GTT_TYPE_MAX) { @@ -1925,11 +1921,9 @@ static int alloc_scratch_pages(struct intel_vgpu *vgpu, se.val64 |= PPAT_CACHED_INDEX; for (i = 0; i < page_entry_num; i++) - ops->set_entry(p, &se, i, false, 0, vgpu); + ops->set_entry(scratch_pt, &se, i, false, 0, vgpu); } - kunmap_atomic(p); - return 0; } @@ -2197,7 +2191,7 @@ int intel_vgpu_g2v_destroy_ppgtt_mm(struct intel_vgpu *vgpu, int intel_gvt_init_gtt(struct intel_gvt *gvt) { int ret; - void *page_addr; + void *page; gvt_dbg_core("init gtt\n"); @@ -2210,17 +2204,14 @@ int intel_gvt_init_gtt(struct intel_gvt *gvt) return -ENODEV; } - gvt->gtt.scratch_ggtt_page = - alloc_page(GFP_KERNEL | GFP_ATOMIC | __GFP_ZERO); - if (!gvt->gtt.scratch_ggtt_page) { + page = (void *)get_zeroed_page(GFP_KERNEL); + if (!page) { gvt_err("fail to allocate scratch ggtt page\n"); return -ENOMEM; } + gvt->gtt.scratch_ggtt_page = virt_to_page(page); - page_addr = page_address(gvt->gtt.scratch_ggtt_page); - - gvt->gtt.scratch_ggtt_mfn = - intel_gvt_hypervisor_virt_to_mfn(page_addr); + gvt->gtt.scratch_ggtt_mfn = intel_gvt_hypervisor_virt_to_mfn(page); if (gvt->gtt.scratch_ggtt_mfn == INTEL_GVT_INVALID_ADDR) { gvt_err("fail to translate scratch ggtt page\n"); __free_page(gvt->gtt.scratch_ggtt_page); From cc31d43b4154ad5a7d8aa5543255a93b7e89edc2 Mon Sep 17 00:00:00 2001 From: Pau Espin Pedrol Date: Fri, 6 Jan 2017 20:33:27 +0100 Subject: [PATCH 034/343] netfilter: use fwmark_reflect in nf_send_reset Otherwise, RST packets generated by ipt_REJECT always have mark 0 when the routing is checked later in the same code path. Fixes: e110861f8609 ("net: add a sysctl to reflect the fwmark on replies") Cc: Lorenzo Colitti Signed-off-by: Pau Espin Pedrol Signed-off-by: Pablo Neira Ayuso --- net/ipv4/netfilter/nf_reject_ipv4.c | 2 ++ net/ipv6/netfilter/nf_reject_ipv6.c | 3 +++ 2 files changed, 5 insertions(+) diff --git a/net/ipv4/netfilter/nf_reject_ipv4.c b/net/ipv4/netfilter/nf_reject_ipv4.c index fd8220213afc..146d86105183 100644 --- a/net/ipv4/netfilter/nf_reject_ipv4.c +++ b/net/ipv4/netfilter/nf_reject_ipv4.c @@ -126,6 +126,8 @@ void nf_send_reset(struct net *net, struct sk_buff *oldskb, int hook) /* ip_route_me_harder expects skb->dst to be set */ skb_dst_set_noref(nskb, skb_dst(oldskb)); + nskb->mark = IP4_REPLY_MARK(net, oldskb->mark); + skb_reserve(nskb, LL_MAX_HEADER); niph = nf_reject_iphdr_put(nskb, oldskb, IPPROTO_TCP, ip4_dst_hoplimit(skb_dst(nskb))); diff --git a/net/ipv6/netfilter/nf_reject_ipv6.c b/net/ipv6/netfilter/nf_reject_ipv6.c index 10090400c72f..eedee5d108d9 100644 --- a/net/ipv6/netfilter/nf_reject_ipv6.c +++ b/net/ipv6/netfilter/nf_reject_ipv6.c @@ -157,6 +157,7 @@ void nf_send_reset6(struct net *net, struct sk_buff *oldskb, int hook) fl6.fl6_sport = otcph->dest; fl6.fl6_dport = otcph->source; fl6.flowi6_oif = l3mdev_master_ifindex(skb_dst(oldskb)->dev); + fl6.flowi6_mark = IP6_REPLY_MARK(net, oldskb->mark); security_skb_classify_flow(oldskb, flowi6_to_flowi(&fl6)); dst = ip6_route_output(net, NULL, &fl6); if (dst->error) { @@ -180,6 +181,8 @@ void nf_send_reset6(struct net *net, struct sk_buff *oldskb, int hook) skb_dst_set(nskb, dst); + nskb->mark = fl6.flowi6_mark; + skb_reserve(nskb, hh_len + dst->header_len); ip6h = nf_reject_ip6hdr_put(nskb, oldskb, IPPROTO_TCP, ip6_dst_hoplimit(dst)); From bf99b4ded5f8a4767dbb9d180626f06c51f9881f Mon Sep 17 00:00:00 2001 From: Pau Espin Pedrol Date: Fri, 6 Jan 2017 20:33:28 +0100 Subject: [PATCH 035/343] tcp: fix mark propagation with fwmark_reflect enabled Otherwise, RST packets generated by the TCP stack for non-existing sockets always have mark 0. The mark from the original packet is assigned to the netns_ipv4/6 socket used to send the response so that it can get copied into the response skb when the socket sends it. Fixes: e110861f8609 ("net: add a sysctl to reflect the fwmark on replies") Cc: Lorenzo Colitti Signed-off-by: Pau Espin Pedrol Signed-off-by: Pablo Neira Ayuso --- net/ipv4/ip_output.c | 1 + net/ipv6/tcp_ipv6.c | 1 + 2 files changed, 2 insertions(+) diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c index fac275c48108..b67719f45953 100644 --- a/net/ipv4/ip_output.c +++ b/net/ipv4/ip_output.c @@ -1629,6 +1629,7 @@ void ip_send_unicast_reply(struct sock *sk, struct sk_buff *skb, sk->sk_protocol = ip_hdr(skb)->protocol; sk->sk_bound_dev_if = arg->bound_dev_if; sk->sk_sndbuf = sysctl_wmem_default; + sk->sk_mark = fl4.flowi4_mark; err = ip_append_data(sk, &fl4, ip_reply_glue_bits, arg->iov->iov_base, len, 0, &ipc, &rt, MSG_DONTWAIT); if (unlikely(err)) { diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index 73bc8fc68acd..2b20622a5824 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -840,6 +840,7 @@ static void tcp_v6_send_response(const struct sock *sk, struct sk_buff *skb, u32 dst = ip6_dst_lookup_flow(ctl_sk, &fl6, NULL); if (!IS_ERR(dst)) { skb_dst_set(buff, dst); + ctl_sk->sk_mark = fl6.flowi6_mark; ip6_xmit(ctl_sk, buff, &fl6, NULL, tclass); TCP_INC_STATS(net, TCP_MIB_OUTSEGS); if (rst) From 68f458eec7069d618a6c884ca007426e0cea411b Mon Sep 17 00:00:00 2001 From: Peter Ujfalusi Date: Mon, 9 Jan 2017 16:31:58 +0200 Subject: [PATCH 036/343] drm: Schedule the output_poll_work with 1s delay if we have delayed event Instead of scheduling the work to handle the initial delayed event, use 1s delay. This delay should not be needed, but Optimus/nouveau will fail in a mysterious way if the delayed event is handled as soon as possible like it is done in drm_helper_probe_single_connector_modes() in case the poll was enabled before. Reverting 339fd36238dd would give back the 10 sec (!) delay to handle the delayed event. Adding 1sec delay to the poll_work is enough to work around the issue in Optimus setups and gives shorter response on handling the initial delayed event. Fixes: 339fd36238dd ("drm: drm_probe_helper: Fix output_poll_work scheduling") Cc: stable@vger.kernel.org # v4.9 Signed-off-by: Peter Ujfalusi [danvet: Add FIXME to the comment to make it stick out more.] Signed-off-by: Daniel Vetter Link: http://patchwork.freedesktop.org/patch/msgid/20170109143158.21917-1-peter.ujfalusi@ti.com --- drivers/gpu/drm/drm_probe_helper.c | 12 +++++++++++- 1 file changed, 11 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/drm_probe_helper.c b/drivers/gpu/drm/drm_probe_helper.c index ac953f037be7..cf8f0128c161 100644 --- a/drivers/gpu/drm/drm_probe_helper.c +++ b/drivers/gpu/drm/drm_probe_helper.c @@ -143,8 +143,18 @@ void drm_kms_helper_poll_enable_locked(struct drm_device *dev) } if (dev->mode_config.delayed_event) { + /* + * FIXME: + * + * Use short (1s) delay to handle the initial delayed event. + * This delay should not be needed, but Optimus/nouveau will + * fail in a mysterious way if the delayed event is handled as + * soon as possible like it is done in + * drm_helper_probe_single_connector_modes() in case the poll + * was enabled before. + */ poll = true; - delay = 0; + delay = HZ; } if (poll) From 71d3f6ef7f5af38dea2975ec5715c88bae92e92d Mon Sep 17 00:00:00 2001 From: Gerd Hoffmann Date: Mon, 28 Nov 2016 08:52:20 +0100 Subject: [PATCH 037/343] drm/virtio: fix framebuffer sparse warning virtio uses normal ram as backing storage for the framebuffer, so we should assign the address to new screen_buffer (added by commit 17a7b0b4d9749f80d365d7baff5dec2f54b0e992) instead of screen_base. Reported-by: Michael S. Tsirkin Signed-off-by: Gerd Hoffmann --- drivers/gpu/drm/virtio/virtgpu_fb.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/virtio/virtgpu_fb.c b/drivers/gpu/drm/virtio/virtgpu_fb.c index dd21f950e129..cde9f3758106 100644 --- a/drivers/gpu/drm/virtio/virtgpu_fb.c +++ b/drivers/gpu/drm/virtio/virtgpu_fb.c @@ -331,7 +331,7 @@ static int virtio_gpufb_create(struct drm_fb_helper *helper, info->fbops = &virtio_gpufb_ops; info->pixmap.flags = FB_PIXMAP_SYSTEM; - info->screen_base = obj->vmap; + info->screen_buffer = obj->vmap; info->screen_size = obj->gem_base.size; drm_fb_helper_fill_fix(info, fb->pitches[0], fb->depth); drm_fb_helper_fill_var(info, &vfbdev->helper, From 0c19f97f12bbb1c2370cb62e31d0f749642937ee Mon Sep 17 00:00:00 2001 From: Gerd Hoffmann Date: Mon, 21 Nov 2016 19:00:30 +0100 Subject: [PATCH 038/343] drm: update MAINTAINERS for qemu drivers (bochs, cirrus, qxl, virtio-gpu) Changes: * add myself as maintainer, so patches land in my inbox. * add virtualization@lists.linux-foundation.org mailing list. * add drm-qemu git repo. * flip bochs and qxl status to "Maintained". Signed-off-by: Gerd Hoffmann --- MAINTAINERS | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) diff --git a/MAINTAINERS b/MAINTAINERS index 5f0420a0da5b..15eadc8c58c9 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -4106,11 +4106,16 @@ F: drivers/gpu/drm/bridge/ DRM DRIVER FOR BOCHS VIRTUAL GPU M: Gerd Hoffmann -S: Odd Fixes +L: virtualization@lists.linux-foundation.org +T: git git://git.kraxel.org/linux drm-qemu +S: Maintained F: drivers/gpu/drm/bochs/ DRM DRIVER FOR QEMU'S CIRRUS DEVICE M: Dave Airlie +M: Gerd Hoffmann +L: virtualization@lists.linux-foundation.org +T: git git://git.kraxel.org/linux drm-qemu S: Odd Fixes F: drivers/gpu/drm/cirrus/ @@ -4304,7 +4309,10 @@ F: Documentation/devicetree/bindings/display/renesas,du.txt DRM DRIVER FOR QXL VIRTUAL GPU M: Dave Airlie -S: Odd Fixes +M: Gerd Hoffmann +L: virtualization@lists.linux-foundation.org +T: git git://git.kraxel.org/linux drm-qemu +S: Maintained F: drivers/gpu/drm/qxl/ F: include/uapi/drm/qxl_drm.h @@ -13085,6 +13093,7 @@ M: David Airlie M: Gerd Hoffmann L: dri-devel@lists.freedesktop.org L: virtualization@lists.linux-foundation.org +T: git git://git.kraxel.org/linux drm-qemu S: Maintained F: drivers/gpu/drm/virtio/ F: include/uapi/linux/virtio_gpu.h From af3076e67c31ceb3e314933dd61cb68a1d5120cf Mon Sep 17 00:00:00 2001 From: Gerd Hoffmann Date: Mon, 21 Nov 2016 19:32:06 +0100 Subject: [PATCH 039/343] drm: flip cirrus driver status to "obsolete". Also update Kconfig help text, explaining things: Cirrus is obsolete, the hardware was designed in the 90ies and can't keep up with todays needs. More background: https://www.kraxel.org/blog/2014/10/qemu-using-cirrus-considered-harmful/ Better alternatives are: - stdvga (DRM_BOCHS, qemu -vga std, default in qemu 2.2+) - qxl (DRM_QXL, qemu -vga qxl, works best with spice) - virtio (VIRTIO_GPU), qemu -vga virtio) Signed-off-by: Gerd Hoffmann --- MAINTAINERS | 3 ++- drivers/gpu/drm/cirrus/Kconfig | 9 +++++++++ 2 files changed, 11 insertions(+), 1 deletion(-) diff --git a/MAINTAINERS b/MAINTAINERS index 15eadc8c58c9..b8e2bbe621a3 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -4116,7 +4116,8 @@ M: Dave Airlie M: Gerd Hoffmann L: virtualization@lists.linux-foundation.org T: git git://git.kraxel.org/linux drm-qemu -S: Odd Fixes +S: Obsolete +W: https://www.kraxel.org/blog/2014/10/qemu-using-cirrus-considered-harmful/ F: drivers/gpu/drm/cirrus/ RADEON and AMDGPU DRM DRIVERS diff --git a/drivers/gpu/drm/cirrus/Kconfig b/drivers/gpu/drm/cirrus/Kconfig index 04b3c161dfae..7f4cc6e172ab 100644 --- a/drivers/gpu/drm/cirrus/Kconfig +++ b/drivers/gpu/drm/cirrus/Kconfig @@ -7,3 +7,12 @@ config DRM_CIRRUS_QEMU This is a KMS driver for emulated cirrus device in qemu. It is *NOT* intended for real cirrus devices. This requires the modesetting userspace X.org driver. + + Cirrus is obsolete, the hardware was designed in the 90ies + and can't keep up with todays needs. More background: + https://www.kraxel.org/blog/2014/10/qemu-using-cirrus-considered-harmful/ + + Better alternatives are: + - stdvga (DRM_BOCHS, qemu -vga std, default in qemu 2.2+) + - qxl (DRM_QXL, qemu -vga qxl, works best with spice) + - virtio (DRM_VIRTIO_GPU), qemu -vga virtio) From 4fe7c2962e110dfd58e61888514726aac419562f Mon Sep 17 00:00:00 2001 From: Steve Wise Date: Thu, 22 Dec 2016 07:04:59 -0800 Subject: [PATCH 040/343] iw_cxgb4: refactor sq/rq drain logic With the addition of the IB/Core drain API, iw_cxgb4 supported drain by watching the CQs when the QP was out of RTS and signalling "drain complete" when the last CQE is polled. This, however, doesn't fully support the drain semantics. Namely, the drain logic is supposed to signal "drain complete" only when the application has _processed_ the last CQE, not just removed them from the CQ. Thus a small timing hole exists that can cause touch after free type bugs in applications using the drain API (nvmf, iSER, for example). So iw_cxgb4 needs a better solution. The iWARP Verbs spec mandates that "_at some point_ after the QP is moved to ERROR", the iWARP driver MUST synchronously fail post_send and post_recv calls. iw_cxgb4 was currently not allowing any posts once the QP is in ERROR. This was in part due to the fact that the HW queues for the QP in ERROR state are disabled at this point, so there wasn't much else to do but fail the post operation synchronously. This restriction is what drove the first drain implementation in iw_cxgb4 that has the above mentioned flaw. This patch changes iw_cxgb4 to allow post_send and post_recv WRs after the QP is moved to ERROR state for kernel mode users, thus still adhering to the Verbs spec for user mode users, but allowing flush WRs for kernel users. Since the HW queues are disabled, we just synthesize a CQE for this post, queue it to the SW CQ, and then call the CQ event handler. This enables proper drain operations for the various storage applications. Signed-off-by: Steve Wise Signed-off-by: Doug Ledford --- drivers/infiniband/hw/cxgb4/cq.c | 21 +++-- drivers/infiniband/hw/cxgb4/iw_cxgb4.h | 6 +- drivers/infiniband/hw/cxgb4/provider.c | 2 - drivers/infiniband/hw/cxgb4/qp.c | 112 +++++++++++++++---------- drivers/infiniband/hw/cxgb4/t4.h | 2 + 5 files changed, 85 insertions(+), 58 deletions(-) diff --git a/drivers/infiniband/hw/cxgb4/cq.c b/drivers/infiniband/hw/cxgb4/cq.c index 19c6477af19f..bec82a600d77 100644 --- a/drivers/infiniband/hw/cxgb4/cq.c +++ b/drivers/infiniband/hw/cxgb4/cq.c @@ -504,6 +504,15 @@ static int poll_cq(struct t4_wq *wq, struct t4_cq *cq, struct t4_cqe *cqe, goto skip_cqe; } + /* + * Special cqe for drain WR completions... + */ + if (CQE_OPCODE(hw_cqe) == C4IW_DRAIN_OPCODE) { + *cookie = CQE_DRAIN_COOKIE(hw_cqe); + *cqe = *hw_cqe; + goto skip_cqe; + } + /* * Gotta tweak READ completions: * 1) the cqe doesn't contain the sq_wptr from the wr. @@ -753,6 +762,9 @@ static int c4iw_poll_cq_one(struct c4iw_cq *chp, struct ib_wc *wc) c4iw_invalidate_mr(qhp->rhp, CQE_WRID_FR_STAG(&cqe)); break; + case C4IW_DRAIN_OPCODE: + wc->opcode = IB_WC_SEND; + break; default: printk(KERN_ERR MOD "Unexpected opcode %d " "in the CQE received for QPID=0x%0x\n", @@ -817,15 +829,8 @@ static int c4iw_poll_cq_one(struct c4iw_cq *chp, struct ib_wc *wc) } } out: - if (wq) { - if (unlikely(qhp->attr.state != C4IW_QP_STATE_RTS)) { - if (t4_sq_empty(wq)) - complete(&qhp->sq_drained); - if (t4_rq_empty(wq)) - complete(&qhp->rq_drained); - } + if (wq) spin_unlock(&qhp->lock); - } return ret; } diff --git a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h index 4788e1a46fde..7b1e465b2a5e 100644 --- a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h +++ b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h @@ -480,8 +480,6 @@ struct c4iw_qp { wait_queue_head_t wait; struct timer_list timer; int sq_sig_all; - struct completion rq_drained; - struct completion sq_drained; }; static inline struct c4iw_qp *to_c4iw_qp(struct ib_qp *ibqp) @@ -615,6 +613,8 @@ static inline int to_ib_qp_state(int c4iw_qp_state) return IB_QPS_ERR; } +#define C4IW_DRAIN_OPCODE FW_RI_SGE_EC_CR_RETURN + static inline u32 c4iw_ib_to_tpt_access(int a) { return (a & IB_ACCESS_REMOTE_WRITE ? FW_RI_MEM_ACCESS_REM_WRITE : 0) | @@ -997,8 +997,6 @@ extern int c4iw_wr_log; extern int db_fc_threshold; extern int db_coalescing_threshold; extern int use_dsgl; -void c4iw_drain_rq(struct ib_qp *qp); -void c4iw_drain_sq(struct ib_qp *qp); void c4iw_invalidate_mr(struct c4iw_dev *rhp, u32 rkey); #endif diff --git a/drivers/infiniband/hw/cxgb4/provider.c b/drivers/infiniband/hw/cxgb4/provider.c index 49b51b7e0fd7..c156413515b1 100644 --- a/drivers/infiniband/hw/cxgb4/provider.c +++ b/drivers/infiniband/hw/cxgb4/provider.c @@ -607,8 +607,6 @@ int c4iw_register_device(struct c4iw_dev *dev) dev->ibdev.uverbs_abi_ver = C4IW_UVERBS_ABI_VERSION; dev->ibdev.get_port_immutable = c4iw_port_immutable; dev->ibdev.get_dev_fw_str = get_dev_fw_str; - dev->ibdev.drain_sq = c4iw_drain_sq; - dev->ibdev.drain_rq = c4iw_drain_rq; dev->ibdev.iwcm = kmalloc(sizeof(struct iw_cm_verbs), GFP_KERNEL); if (!dev->ibdev.iwcm) diff --git a/drivers/infiniband/hw/cxgb4/qp.c b/drivers/infiniband/hw/cxgb4/qp.c index cda5542e13a2..31ab4512f827 100644 --- a/drivers/infiniband/hw/cxgb4/qp.c +++ b/drivers/infiniband/hw/cxgb4/qp.c @@ -776,6 +776,64 @@ static int ring_kernel_rq_db(struct c4iw_qp *qhp, u16 inc) return 0; } +static void complete_sq_drain_wr(struct c4iw_qp *qhp, struct ib_send_wr *wr) +{ + struct t4_cqe cqe = {}; + struct c4iw_cq *schp; + unsigned long flag; + struct t4_cq *cq; + + schp = to_c4iw_cq(qhp->ibqp.send_cq); + cq = &schp->cq; + + cqe.u.drain_cookie = wr->wr_id; + cqe.header = cpu_to_be32(CQE_STATUS_V(T4_ERR_SWFLUSH) | + CQE_OPCODE_V(C4IW_DRAIN_OPCODE) | + CQE_TYPE_V(1) | + CQE_SWCQE_V(1) | + CQE_QPID_V(qhp->wq.sq.qid)); + + spin_lock_irqsave(&schp->lock, flag); + cqe.bits_type_ts = cpu_to_be64(CQE_GENBIT_V((u64)cq->gen)); + cq->sw_queue[cq->sw_pidx] = cqe; + t4_swcq_produce(cq); + spin_unlock_irqrestore(&schp->lock, flag); + + spin_lock_irqsave(&schp->comp_handler_lock, flag); + (*schp->ibcq.comp_handler)(&schp->ibcq, + schp->ibcq.cq_context); + spin_unlock_irqrestore(&schp->comp_handler_lock, flag); +} + +static void complete_rq_drain_wr(struct c4iw_qp *qhp, struct ib_recv_wr *wr) +{ + struct t4_cqe cqe = {}; + struct c4iw_cq *rchp; + unsigned long flag; + struct t4_cq *cq; + + rchp = to_c4iw_cq(qhp->ibqp.recv_cq); + cq = &rchp->cq; + + cqe.u.drain_cookie = wr->wr_id; + cqe.header = cpu_to_be32(CQE_STATUS_V(T4_ERR_SWFLUSH) | + CQE_OPCODE_V(C4IW_DRAIN_OPCODE) | + CQE_TYPE_V(0) | + CQE_SWCQE_V(1) | + CQE_QPID_V(qhp->wq.sq.qid)); + + spin_lock_irqsave(&rchp->lock, flag); + cqe.bits_type_ts = cpu_to_be64(CQE_GENBIT_V((u64)cq->gen)); + cq->sw_queue[cq->sw_pidx] = cqe; + t4_swcq_produce(cq); + spin_unlock_irqrestore(&rchp->lock, flag); + + spin_lock_irqsave(&rchp->comp_handler_lock, flag); + (*rchp->ibcq.comp_handler)(&rchp->ibcq, + rchp->ibcq.cq_context); + spin_unlock_irqrestore(&rchp->comp_handler_lock, flag); +} + int c4iw_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr, struct ib_send_wr **bad_wr) { @@ -794,8 +852,8 @@ int c4iw_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr, spin_lock_irqsave(&qhp->lock, flag); if (t4_wq_in_error(&qhp->wq)) { spin_unlock_irqrestore(&qhp->lock, flag); - *bad_wr = wr; - return -EINVAL; + complete_sq_drain_wr(qhp, wr); + return err; } num_wrs = t4_sq_avail(&qhp->wq); if (num_wrs == 0) { @@ -937,8 +995,8 @@ int c4iw_post_receive(struct ib_qp *ibqp, struct ib_recv_wr *wr, spin_lock_irqsave(&qhp->lock, flag); if (t4_wq_in_error(&qhp->wq)) { spin_unlock_irqrestore(&qhp->lock, flag); - *bad_wr = wr; - return -EINVAL; + complete_rq_drain_wr(qhp, wr); + return err; } num_wrs = t4_rq_avail(&qhp->wq); if (num_wrs == 0) { @@ -1550,7 +1608,12 @@ int c4iw_modify_qp(struct c4iw_dev *rhp, struct c4iw_qp *qhp, } break; case C4IW_QP_STATE_CLOSING: - if (!internal) { + + /* + * Allow kernel users to move to ERROR for qp draining. + */ + if (!internal && (qhp->ibqp.uobject || attrs->next_state != + C4IW_QP_STATE_ERROR)) { ret = -EINVAL; goto out; } @@ -1763,8 +1826,6 @@ struct ib_qp *c4iw_create_qp(struct ib_pd *pd, struct ib_qp_init_attr *attrs, qhp->attr.max_ird = 0; qhp->sq_sig_all = attrs->sq_sig_type == IB_SIGNAL_ALL_WR; spin_lock_init(&qhp->lock); - init_completion(&qhp->sq_drained); - init_completion(&qhp->rq_drained); mutex_init(&qhp->mutex); init_waitqueue_head(&qhp->wait); kref_init(&qhp->kref); @@ -1958,40 +2019,3 @@ int c4iw_ib_query_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr, init_attr->sq_sig_type = qhp->sq_sig_all ? IB_SIGNAL_ALL_WR : 0; return 0; } - -static void move_qp_to_err(struct c4iw_qp *qp) -{ - struct c4iw_qp_attributes attrs = { .next_state = C4IW_QP_STATE_ERROR }; - - (void)c4iw_modify_qp(qp->rhp, qp, C4IW_QP_ATTR_NEXT_STATE, &attrs, 1); -} - -void c4iw_drain_sq(struct ib_qp *ibqp) -{ - struct c4iw_qp *qp = to_c4iw_qp(ibqp); - unsigned long flag; - bool need_to_wait; - - move_qp_to_err(qp); - spin_lock_irqsave(&qp->lock, flag); - need_to_wait = !t4_sq_empty(&qp->wq); - spin_unlock_irqrestore(&qp->lock, flag); - - if (need_to_wait) - wait_for_completion(&qp->sq_drained); -} - -void c4iw_drain_rq(struct ib_qp *ibqp) -{ - struct c4iw_qp *qp = to_c4iw_qp(ibqp); - unsigned long flag; - bool need_to_wait; - - move_qp_to_err(qp); - spin_lock_irqsave(&qp->lock, flag); - need_to_wait = !t4_rq_empty(&qp->wq); - spin_unlock_irqrestore(&qp->lock, flag); - - if (need_to_wait) - wait_for_completion(&qp->rq_drained); -} diff --git a/drivers/infiniband/hw/cxgb4/t4.h b/drivers/infiniband/hw/cxgb4/t4.h index 862381aa83c8..640d22148a3e 100644 --- a/drivers/infiniband/hw/cxgb4/t4.h +++ b/drivers/infiniband/hw/cxgb4/t4.h @@ -179,6 +179,7 @@ struct t4_cqe { __be32 wrid_hi; __be32 wrid_low; } gen; + u64 drain_cookie; } u; __be64 reserved; __be64 bits_type_ts; @@ -238,6 +239,7 @@ struct t4_cqe { /* generic accessor macros */ #define CQE_WRID_HI(x) (be32_to_cpu((x)->u.gen.wrid_hi)) #define CQE_WRID_LOW(x) (be32_to_cpu((x)->u.gen.wrid_low)) +#define CQE_DRAIN_COOKIE(x) ((x)->u.drain_cookie) /* macros for flit 3 of the cqe */ #define CQE_GENBIT_S 63 From c12a67fec8d99bb554e8d4e99120d418f1a39c87 Mon Sep 17 00:00:00 2001 From: Steve Wise Date: Thu, 22 Dec 2016 07:40:36 -0800 Subject: [PATCH 041/343] iw_cxgb4: free EQ queue memory on last deref Commit ad61a4c7a9b7 ("iw_cxgb4: don't block in destroy_qp awaiting the last deref") introduced a bug where the RDMA QP EQ queue memory (and QIDs) are possibly freed before the underlying connection has been fully shutdown. The result being a possible DMA read issued by HW after the queue memory has been unmapped and freed. This results in possible WR corruption in the worst case, system bus errors if an IOMMU is in use, and SGE "bad WR" errors reported in the very least. The fix is to defer unmap/free of queue memory and QID resources until the QP struct has been fully dereferenced. To do this, the c4iw_ucontext must also be kept around until the last QP that references it is fully freed. In addition, since the last QP deref can happen in an IRQ disabled context, we need a new workqueue thread to do the final unmap/free of the EQ queue memory. Fixes: ad61a4c7a9b7 ("iw_cxgb4: don't block in destroy_qp awaiting the last deref") Cc: stable@vger.kernel.org Signed-off-by: Steve Wise Signed-off-by: Doug Ledford --- drivers/infiniband/hw/cxgb4/device.c | 9 +++++++ drivers/infiniband/hw/cxgb4/iw_cxgb4.h | 18 +++++++++++++ drivers/infiniband/hw/cxgb4/provider.c | 20 ++++++++++++--- drivers/infiniband/hw/cxgb4/qp.c | 35 +++++++++++++++++++------- 4 files changed, 69 insertions(+), 13 deletions(-) diff --git a/drivers/infiniband/hw/cxgb4/device.c b/drivers/infiniband/hw/cxgb4/device.c index 516b0ae6dc3f..40c0e7b9fc6e 100644 --- a/drivers/infiniband/hw/cxgb4/device.c +++ b/drivers/infiniband/hw/cxgb4/device.c @@ -846,9 +846,17 @@ static int c4iw_rdev_open(struct c4iw_rdev *rdev) } } + rdev->free_workq = create_singlethread_workqueue("iw_cxgb4_free"); + if (!rdev->free_workq) { + err = -ENOMEM; + goto err_free_status_page; + } + rdev->status_page->db_off = 0; return 0; +err_free_status_page: + free_page((unsigned long)rdev->status_page); destroy_ocqp_pool: c4iw_ocqp_pool_destroy(rdev); destroy_rqtpool: @@ -862,6 +870,7 @@ destroy_resource: static void c4iw_rdev_close(struct c4iw_rdev *rdev) { + destroy_workqueue(rdev->free_workq); kfree(rdev->wr_log); free_page((unsigned long)rdev->status_page); c4iw_pblpool_destroy(rdev); diff --git a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h index 7b1e465b2a5e..8cd4d054a87e 100644 --- a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h +++ b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h @@ -45,6 +45,7 @@ #include #include #include +#include #include @@ -107,6 +108,7 @@ struct c4iw_dev_ucontext { struct list_head qpids; struct list_head cqids; struct mutex lock; + struct kref kref; }; enum c4iw_rdev_flags { @@ -183,6 +185,7 @@ struct c4iw_rdev { atomic_t wr_log_idx; struct wr_log_entry *wr_log; int wr_log_size; + struct workqueue_struct *free_workq; }; static inline int c4iw_fatal_error(struct c4iw_rdev *rdev) @@ -480,6 +483,8 @@ struct c4iw_qp { wait_queue_head_t wait; struct timer_list timer; int sq_sig_all; + struct work_struct free_work; + struct c4iw_ucontext *ucontext; }; static inline struct c4iw_qp *to_c4iw_qp(struct ib_qp *ibqp) @@ -493,6 +498,7 @@ struct c4iw_ucontext { u32 key; spinlock_t mmap_lock; struct list_head mmaps; + struct kref kref; }; static inline struct c4iw_ucontext *to_c4iw_ucontext(struct ib_ucontext *c) @@ -500,6 +506,18 @@ static inline struct c4iw_ucontext *to_c4iw_ucontext(struct ib_ucontext *c) return container_of(c, struct c4iw_ucontext, ibucontext); } +void _c4iw_free_ucontext(struct kref *kref); + +static inline void c4iw_put_ucontext(struct c4iw_ucontext *ucontext) +{ + kref_put(&ucontext->kref, _c4iw_free_ucontext); +} + +static inline void c4iw_get_ucontext(struct c4iw_ucontext *ucontext) +{ + kref_get(&ucontext->kref); +} + struct c4iw_mm_entry { struct list_head entry; u64 addr; diff --git a/drivers/infiniband/hw/cxgb4/provider.c b/drivers/infiniband/hw/cxgb4/provider.c index c156413515b1..fa64f5d93b11 100644 --- a/drivers/infiniband/hw/cxgb4/provider.c +++ b/drivers/infiniband/hw/cxgb4/provider.c @@ -93,17 +93,28 @@ static int c4iw_process_mad(struct ib_device *ibdev, int mad_flags, return -ENOSYS; } -static int c4iw_dealloc_ucontext(struct ib_ucontext *context) +void _c4iw_free_ucontext(struct kref *kref) { - struct c4iw_dev *rhp = to_c4iw_dev(context->device); - struct c4iw_ucontext *ucontext = to_c4iw_ucontext(context); + struct c4iw_ucontext *ucontext; + struct c4iw_dev *rhp; struct c4iw_mm_entry *mm, *tmp; - PDBG("%s context %p\n", __func__, context); + ucontext = container_of(kref, struct c4iw_ucontext, kref); + rhp = to_c4iw_dev(ucontext->ibucontext.device); + + PDBG("%s ucontext %p\n", __func__, ucontext); list_for_each_entry_safe(mm, tmp, &ucontext->mmaps, entry) kfree(mm); c4iw_release_dev_ucontext(&rhp->rdev, &ucontext->uctx); kfree(ucontext); +} + +static int c4iw_dealloc_ucontext(struct ib_ucontext *context) +{ + struct c4iw_ucontext *ucontext = to_c4iw_ucontext(context); + + PDBG("%s context %p\n", __func__, context); + c4iw_put_ucontext(ucontext); return 0; } @@ -127,6 +138,7 @@ static struct ib_ucontext *c4iw_alloc_ucontext(struct ib_device *ibdev, c4iw_init_dev_ucontext(&rhp->rdev, &context->uctx); INIT_LIST_HEAD(&context->mmaps); spin_lock_init(&context->mmap_lock); + kref_init(&context->kref); if (udata->outlen < sizeof(uresp) - sizeof(uresp.reserved)) { if (!warned++) diff --git a/drivers/infiniband/hw/cxgb4/qp.c b/drivers/infiniband/hw/cxgb4/qp.c index 31ab4512f827..04c1c382dedb 100644 --- a/drivers/infiniband/hw/cxgb4/qp.c +++ b/drivers/infiniband/hw/cxgb4/qp.c @@ -715,13 +715,32 @@ static int build_inv_stag(union t4_wr *wqe, struct ib_send_wr *wr, u8 *len16) return 0; } -static void _free_qp(struct kref *kref) +static void free_qp_work(struct work_struct *work) +{ + struct c4iw_ucontext *ucontext; + struct c4iw_qp *qhp; + struct c4iw_dev *rhp; + + qhp = container_of(work, struct c4iw_qp, free_work); + ucontext = qhp->ucontext; + rhp = qhp->rhp; + + PDBG("%s qhp %p ucontext %p\n", __func__, qhp, ucontext); + destroy_qp(&rhp->rdev, &qhp->wq, + ucontext ? &ucontext->uctx : &rhp->rdev.uctx); + + if (ucontext) + c4iw_put_ucontext(ucontext); + kfree(qhp); +} + +static void queue_qp_free(struct kref *kref) { struct c4iw_qp *qhp; qhp = container_of(kref, struct c4iw_qp, kref); PDBG("%s qhp %p\n", __func__, qhp); - kfree(qhp); + queue_work(qhp->rhp->rdev.free_workq, &qhp->free_work); } void c4iw_qp_add_ref(struct ib_qp *qp) @@ -733,7 +752,7 @@ void c4iw_qp_add_ref(struct ib_qp *qp) void c4iw_qp_rem_ref(struct ib_qp *qp) { PDBG("%s ib_qp %p\n", __func__, qp); - kref_put(&to_c4iw_qp(qp)->kref, _free_qp); + kref_put(&to_c4iw_qp(qp)->kref, queue_qp_free); } static void add_to_fc_list(struct list_head *head, struct list_head *entry) @@ -1706,7 +1725,6 @@ int c4iw_destroy_qp(struct ib_qp *ib_qp) struct c4iw_dev *rhp; struct c4iw_qp *qhp; struct c4iw_qp_attributes attrs; - struct c4iw_ucontext *ucontext; qhp = to_c4iw_qp(ib_qp); rhp = qhp->rhp; @@ -1726,11 +1744,6 @@ int c4iw_destroy_qp(struct ib_qp *ib_qp) spin_unlock_irq(&rhp->lock); free_ird(rhp, qhp->attr.max_ird); - ucontext = ib_qp->uobject ? - to_c4iw_ucontext(ib_qp->uobject->context) : NULL; - destroy_qp(&rhp->rdev, &qhp->wq, - ucontext ? &ucontext->uctx : &rhp->rdev.uctx); - c4iw_qp_rem_ref(ib_qp); PDBG("%s ib_qp %p qpid 0x%0x\n", __func__, ib_qp, qhp->wq.sq.qid); @@ -1829,6 +1842,7 @@ struct ib_qp *c4iw_create_qp(struct ib_pd *pd, struct ib_qp_init_attr *attrs, mutex_init(&qhp->mutex); init_waitqueue_head(&qhp->wait); kref_init(&qhp->kref); + INIT_WORK(&qhp->free_work, free_qp_work); ret = insert_handle(rhp, &rhp->qpidr, qhp, qhp->wq.sq.qid); if (ret) @@ -1915,6 +1929,9 @@ struct ib_qp *c4iw_create_qp(struct ib_pd *pd, struct ib_qp_init_attr *attrs, ma_sync_key_mm->len = PAGE_SIZE; insert_mmap(ucontext, ma_sync_key_mm); } + + c4iw_get_ucontext(ucontext); + qhp->ucontext = ucontext; } qhp->ibqp.qp_num = qhp->wq.sq.qid; init_timer(&(qhp->timer)); From 3bcf96e0183f5c863657cb6ae9adad307a0f6071 Mon Sep 17 00:00:00 2001 From: Steve Wise Date: Thu, 22 Dec 2016 07:40:37 -0800 Subject: [PATCH 042/343] iw_cxgb4: do not send RX_DATA_ACK CPLs after close/abort Function rx_data(), which handles ingress CPL_RX_DATA messages, was always sending an RX_DATA_ACK with the goal of updating the credits. However, if the RDMA connection is moved out of FPDU mode abruptly, then it is possible for iw_cxgb4 to process queued RX_DATA CPLs after HW has aborted the connection. These CPLs should not trigger RX_DATA_ACKS. If they do, HW can see a READ after DELETE of the DB_LE hash entry for the tid and post a LE_DB HashTblMemCrcError. Signed-off-by: Steve Wise Signed-off-by: Doug Ledford --- drivers/infiniband/hw/cxgb4/cm.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/drivers/infiniband/hw/cxgb4/cm.c b/drivers/infiniband/hw/cxgb4/cm.c index f1510cc76d2d..9398143d7c5e 100644 --- a/drivers/infiniband/hw/cxgb4/cm.c +++ b/drivers/infiniband/hw/cxgb4/cm.c @@ -1804,20 +1804,21 @@ static int rx_data(struct c4iw_dev *dev, struct sk_buff *skb) skb_trim(skb, dlen); mutex_lock(&ep->com.mutex); - /* update RX credits */ - update_rx_credits(ep, dlen); - switch (ep->com.state) { case MPA_REQ_SENT: + update_rx_credits(ep, dlen); ep->rcv_seq += dlen; disconnect = process_mpa_reply(ep, skb); break; case MPA_REQ_WAIT: + update_rx_credits(ep, dlen); ep->rcv_seq += dlen; disconnect = process_mpa_request(ep, skb); break; case FPDU_MODE: { struct c4iw_qp_attributes attrs; + + update_rx_credits(ep, dlen); BUG_ON(!ep->com.qp); if (status) pr_err("%s Unexpected streaming data." \ From 3546fb0cdac25a79c89d87020566fab52b92867d Mon Sep 17 00:00:00 2001 From: Lucas Stach Date: Mon, 12 Dec 2016 16:15:17 +0100 Subject: [PATCH 043/343] drm/etnaviv: trick drm_mm into giving out a low IOVA After rollover of the IOVA space, we want to get a low IOVA address, otherwise the the games we play by remembering the last IOVA are pointless. When we search for a free hole with DRM_MM_SEARCH_DEFAULT, drm_mm will pop the next entry from the free holes stack, which will likely be a high IOVA. By using DRM_MM_SEARCH_BELOW we can trick drm_mm into reversing the search and provide us with a low IOVA. Signed-off-by: Lucas Stach Reviewed-by: Wladimir van der Laan --- drivers/gpu/drm/etnaviv/etnaviv_mmu.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/etnaviv/etnaviv_mmu.c b/drivers/gpu/drm/etnaviv/etnaviv_mmu.c index 169ac96e8f08..fe0e85b41310 100644 --- a/drivers/gpu/drm/etnaviv/etnaviv_mmu.c +++ b/drivers/gpu/drm/etnaviv/etnaviv_mmu.c @@ -116,9 +116,14 @@ static int etnaviv_iommu_find_iova(struct etnaviv_iommu *mmu, struct list_head list; bool found; + /* + * XXX: The DRM_MM_SEARCH_BELOW is really a hack to trick + * drm_mm into giving out a low IOVA after address space + * rollover. This needs a proper fix. + */ ret = drm_mm_insert_node_in_range(&mmu->mm, node, size, 0, mmu->last_iova, ~0UL, - DRM_MM_SEARCH_DEFAULT); + mmu->last_iova ? DRM_MM_SEARCH_DEFAULT : DRM_MM_SEARCH_BELOW); if (ret != -ENOSPC) break; From ecc8995363ee6231b32dad61c955b371b79cc4cf Mon Sep 17 00:00:00 2001 From: Mika Westerberg Date: Tue, 10 Jan 2017 17:31:56 +0300 Subject: [PATCH 044/343] pinctrl: broxton: Use correct PADCFGLOCK offset PADCFGLOCK (and PADCFGLOCK_TX) offset in Broxton actually starts at 0x060 and not 0x090 as used in the driver. Fix it to use the correct offset. Signed-off-by: Mika Westerberg Reviewed-by: Andy Shevchenko Signed-off-by: Linus Walleij --- drivers/pinctrl/intel/pinctrl-broxton.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/pinctrl/intel/pinctrl-broxton.c b/drivers/pinctrl/intel/pinctrl-broxton.c index 59cb7a6fc5be..901b356b09d7 100644 --- a/drivers/pinctrl/intel/pinctrl-broxton.c +++ b/drivers/pinctrl/intel/pinctrl-broxton.c @@ -19,7 +19,7 @@ #define BXT_PAD_OWN 0x020 #define BXT_HOSTSW_OWN 0x080 -#define BXT_PADCFGLOCK 0x090 +#define BXT_PADCFGLOCK 0x060 #define BXT_GPI_IE 0x110 #define BXT_COMMUNITY(s, e) \ From 17fab473693e8357a9aa6fee4fbed6c13a34bd81 Mon Sep 17 00:00:00 2001 From: Andy Shevchenko Date: Mon, 2 Jan 2017 14:07:22 +0200 Subject: [PATCH 045/343] pinctrl: intel: Set pin direction properly There are two bits in the PADCFG0 register to configure direction, one per TX/RX buffers. For now we wrongly assume that the GPIO is always requested before it is being used, which is not true when the GPIO is used through irqchip. In this case the GPIO is never requested and we never enable RX buffer for it. Fix this by setting both bits accordingly. Reported-by: Jarkko Nikula Signed-off-by: Andy Shevchenko Signed-off-by: Linus Walleij --- drivers/pinctrl/intel/pinctrl-intel.c | 30 +++++++++++++++++---------- 1 file changed, 19 insertions(+), 11 deletions(-) diff --git a/drivers/pinctrl/intel/pinctrl-intel.c b/drivers/pinctrl/intel/pinctrl-intel.c index 1e139672f1af..6df35dcb29ae 100644 --- a/drivers/pinctrl/intel/pinctrl-intel.c +++ b/drivers/pinctrl/intel/pinctrl-intel.c @@ -353,6 +353,21 @@ static int intel_pinmux_set_mux(struct pinctrl_dev *pctldev, unsigned function, return 0; } +static void __intel_gpio_set_direction(void __iomem *padcfg0, bool input) +{ + u32 value; + + value = readl(padcfg0); + if (input) { + value &= ~PADCFG0_GPIORXDIS; + value |= PADCFG0_GPIOTXDIS; + } else { + value &= ~PADCFG0_GPIOTXDIS; + value |= PADCFG0_GPIORXDIS; + } + writel(value, padcfg0); +} + static int intel_gpio_request_enable(struct pinctrl_dev *pctldev, struct pinctrl_gpio_range *range, unsigned pin) @@ -375,11 +390,11 @@ static int intel_gpio_request_enable(struct pinctrl_dev *pctldev, /* Disable SCI/SMI/NMI generation */ value &= ~(PADCFG0_GPIROUTIOXAPIC | PADCFG0_GPIROUTSCI); value &= ~(PADCFG0_GPIROUTSMI | PADCFG0_GPIROUTNMI); - /* Disable TX buffer and enable RX (this will be input) */ - value &= ~PADCFG0_GPIORXDIS; - value |= PADCFG0_GPIOTXDIS; writel(value, padcfg0); + /* Disable TX buffer and enable RX (this will be input) */ + __intel_gpio_set_direction(padcfg0, true); + raw_spin_unlock_irqrestore(&pctrl->lock, flags); return 0; @@ -392,18 +407,11 @@ static int intel_gpio_set_direction(struct pinctrl_dev *pctldev, struct intel_pinctrl *pctrl = pinctrl_dev_get_drvdata(pctldev); void __iomem *padcfg0; unsigned long flags; - u32 value; raw_spin_lock_irqsave(&pctrl->lock, flags); padcfg0 = intel_get_padcfg(pctrl, pin, PADCFG0); - - value = readl(padcfg0); - if (input) - value |= PADCFG0_GPIOTXDIS; - else - value &= ~PADCFG0_GPIOTXDIS; - writel(value, padcfg0); + __intel_gpio_set_direction(padcfg0, input); raw_spin_unlock_irqrestore(&pctrl->lock, flags); From 04ff5a095d662e0879f0eb04b9247e092210aeff Mon Sep 17 00:00:00 2001 From: Andy Shevchenko Date: Tue, 10 Jan 2017 16:38:52 +0200 Subject: [PATCH 046/343] pinctrl: baytrail: Rectify debounce support The commit 658b476c742f ("pinctrl: baytrail: Add debounce configuration") implements debounce for Baytrail pin control, but seems wasn't tested properly. The register which keeps debounce value is separated from the configuration one. Writing wrong values to the latter will guarantee wrong behaviour of the driver and even might break something physically. Besides above there is missed case how to disable it, which is actually done through the bit in configuration register. Rectify implementation here by using proper register for debounce value. Fixes: 658b476c742f ("pinctrl: baytrail: Add debounce configuration") Cc: Cristina Ciocan Signed-off-by: Andy Shevchenko Acked-by: Mika Westerberg Signed-off-by: Linus Walleij --- drivers/pinctrl/intel/pinctrl-baytrail.c | 28 ++++++++++++++---------- 1 file changed, 17 insertions(+), 11 deletions(-) diff --git a/drivers/pinctrl/intel/pinctrl-baytrail.c b/drivers/pinctrl/intel/pinctrl-baytrail.c index 37300634b7d2..67e92699b84e 100644 --- a/drivers/pinctrl/intel/pinctrl-baytrail.c +++ b/drivers/pinctrl/intel/pinctrl-baytrail.c @@ -1092,6 +1092,7 @@ static int byt_pin_config_get(struct pinctrl_dev *pctl_dev, unsigned int offset, enum pin_config_param param = pinconf_to_config_param(*config); void __iomem *conf_reg = byt_gpio_reg(vg, offset, BYT_CONF0_REG); void __iomem *val_reg = byt_gpio_reg(vg, offset, BYT_VAL_REG); + void __iomem *db_reg = byt_gpio_reg(vg, offset, BYT_DEBOUNCE_REG); unsigned long flags; u32 conf, pull, val, debounce; u16 arg = 0; @@ -1128,7 +1129,7 @@ static int byt_pin_config_get(struct pinctrl_dev *pctl_dev, unsigned int offset, return -EINVAL; raw_spin_lock_irqsave(&vg->lock, flags); - debounce = readl(byt_gpio_reg(vg, offset, BYT_DEBOUNCE_REG)); + debounce = readl(db_reg); raw_spin_unlock_irqrestore(&vg->lock, flags); switch (debounce & BYT_DEBOUNCE_PULSE_MASK) { @@ -1176,6 +1177,7 @@ static int byt_pin_config_set(struct pinctrl_dev *pctl_dev, unsigned int param, arg; void __iomem *conf_reg = byt_gpio_reg(vg, offset, BYT_CONF0_REG); void __iomem *val_reg = byt_gpio_reg(vg, offset, BYT_VAL_REG); + void __iomem *db_reg = byt_gpio_reg(vg, offset, BYT_DEBOUNCE_REG); unsigned long flags; u32 conf, val, debounce; int i, ret = 0; @@ -1238,36 +1240,40 @@ static int byt_pin_config_set(struct pinctrl_dev *pctl_dev, break; case PIN_CONFIG_INPUT_DEBOUNCE: - debounce = readl(byt_gpio_reg(vg, offset, - BYT_DEBOUNCE_REG)); - conf &= ~BYT_DEBOUNCE_PULSE_MASK; + debounce = readl(db_reg); + debounce &= ~BYT_DEBOUNCE_PULSE_MASK; switch (arg) { + case 0: + conf &= BYT_DEBOUNCE_EN; + break; case 375: - conf |= BYT_DEBOUNCE_PULSE_375US; + debounce |= BYT_DEBOUNCE_PULSE_375US; break; case 750: - conf |= BYT_DEBOUNCE_PULSE_750US; + debounce |= BYT_DEBOUNCE_PULSE_750US; break; case 1500: - conf |= BYT_DEBOUNCE_PULSE_1500US; + debounce |= BYT_DEBOUNCE_PULSE_1500US; break; case 3000: - conf |= BYT_DEBOUNCE_PULSE_3MS; + debounce |= BYT_DEBOUNCE_PULSE_3MS; break; case 6000: - conf |= BYT_DEBOUNCE_PULSE_6MS; + debounce |= BYT_DEBOUNCE_PULSE_6MS; break; case 12000: - conf |= BYT_DEBOUNCE_PULSE_12MS; + debounce |= BYT_DEBOUNCE_PULSE_12MS; break; case 24000: - conf |= BYT_DEBOUNCE_PULSE_24MS; + debounce |= BYT_DEBOUNCE_PULSE_24MS; break; default: ret = -EINVAL; } + if (!ret) + writel(debounce, db_reg); break; default: ret = -ENOTSUPP; From fdf35a6b22247746a7053fc764d04218a9306f82 Mon Sep 17 00:00:00 2001 From: Takashi Iwai Date: Mon, 9 Jan 2017 15:56:14 +0100 Subject: [PATCH 047/343] drm: Fix broken VT switch with video=1366x768 option MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit I noticed that the VT switch doesn't work any longer with a Dell laptop with 1366x768 eDP when the machine is connected with a DP monitor. It behaves as if VT were switched, but the graphics remain frozen. Actually the keyboard works, so I could switch back to VT7 again. I tried to track down the problem, and encountered a long story until we reach to this error: - The machine is booted with video=1366x768 option (the distro installer seems to add it as default). - Recently, drm_helper_probe_single_connector_modes() deals with cmdline modes, and it tries to create a new mode when no matching mode is found. - The drm_mode_create_from_cmdline_mode() creates a mode based on either CVT of GFT according to the given cmdline mode; in our case, it's 1366x768. - Since both CVT and GFT can't express the width 1366 due to alignment, the resultant mode becomes 1368x768, slightly larger than the given size. - Later on, the atomic commit is performed, and in drm_atomic_check_only(), the size of each plane is checked. - The size check of 1366x768 fails due to the above, and eventually the whole VT switch fails. Back in the history, we've had a manual fix-up of 1368x768 in various places via c09dedb7a50e ("drm/edid: Add a workaround for 1366x768 HD panel"), but they have been all in drm_edid.c at probing the modes from EDID. For addressing the problem above, we need a similar hack to the mode newly created from cmdline, manually adjusting the width when the expected size is 1366 while we get 1368 instead. Fixes: eaf99c749d43 ("drm: Perform cmdline mode parsing during...") Cc: Signed-off-by: Takashi Iwai Link: http://patchwork.freedesktop.org/patch/msgid/20170109145614.29454-1-tiwai@suse.de Reviewed-by: Ville Syrjälä Signed-off-by: Ville Syrjälä --- drivers/gpu/drm/drm_modes.c | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/drivers/gpu/drm/drm_modes.c b/drivers/gpu/drm/drm_modes.c index ac6a35212501..e6b19bc9021a 100644 --- a/drivers/gpu/drm/drm_modes.c +++ b/drivers/gpu/drm/drm_modes.c @@ -1460,6 +1460,13 @@ drm_mode_create_from_cmdline_mode(struct drm_device *dev, return NULL; mode->type |= DRM_MODE_TYPE_USERDEF; + /* fix up 1368x768: GFT/CVT can't express 1366 width due to alignment */ + if (cmd->xres == 1366 && mode->hdisplay == 1368) { + mode->hdisplay = 1366; + mode->hsync_start--; + mode->hsync_end--; + drm_mode_set_name(mode); + } drm_mode_set_crtcinfo(mode, CRTC_INTERLACE_HALVE_V); return mode; } From 49c03096263871a68c9dea3e86b7d1e163d2fba8 Mon Sep 17 00:00:00 2001 From: Andy Shevchenko Date: Wed, 11 Jan 2017 15:49:00 +0200 Subject: [PATCH 048/343] pinctrl: baytrail: Do not add all GPIOs to IRQ domain When DIRECT_IRQ_EN is set, the pin is routed directly to the IO-APIC bypassing the GPIO driver completely. However, the mask register is still used to determine if the pin is supposed to generate IRQ or not. So with commit 3ae02c14d964 the IRQ core masks all IRQs (because of handle_bad_irq()) the pin connected to the touchscreen gets masked as well and hence no interrupts. To make this all work as expected we do not add those GPIOs to the IRQ domain that can actually propagate interrupts. Fixes: 3ae02c14d964 ("pinctrl: intel: set default handler to be handle_bad_irq()") Reported-by: Robert R. Howell Suggested-by: Mika Westerberg Signed-off-by: Andy Shevchenko Acked-by: Mika Westerberg Signed-off-by: Linus Walleij --- drivers/pinctrl/intel/pinctrl-baytrail.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) diff --git a/drivers/pinctrl/intel/pinctrl-baytrail.c b/drivers/pinctrl/intel/pinctrl-baytrail.c index 67e92699b84e..c123488266ce 100644 --- a/drivers/pinctrl/intel/pinctrl-baytrail.c +++ b/drivers/pinctrl/intel/pinctrl-baytrail.c @@ -1623,6 +1623,8 @@ static void byt_gpio_irq_handler(struct irq_desc *desc) static void byt_gpio_irq_init_hw(struct byt_gpio *vg) { + struct gpio_chip *gc = &vg->chip; + struct device *dev = &vg->pdev->dev; void __iomem *reg; u32 base, value; int i; @@ -1644,10 +1646,12 @@ static void byt_gpio_irq_init_hw(struct byt_gpio *vg) } value = readl(reg); - if ((value & BYT_PIN_MUX) == byt_get_gpio_mux(vg, i) && - !(value & BYT_DIRECT_IRQ_EN)) { + if (value & BYT_DIRECT_IRQ_EN) { + clear_bit(i, gc->irq_valid_mask); + dev_dbg(dev, "excluding GPIO %d from IRQ domain\n", i); + } else if ((value & BYT_PIN_MUX) == byt_get_gpio_mux(vg, i)) { byt_gpio_clear_triggering(vg, i); - dev_dbg(&vg->pdev->dev, "disabling GPIO %d\n", i); + dev_dbg(dev, "disabling GPIO %d\n", i); } } @@ -1686,6 +1690,7 @@ static int byt_gpio_probe(struct byt_gpio *vg) gc->can_sleep = false; gc->parent = &vg->pdev->dev; gc->ngpio = vg->soc_data->npins; + gc->irq_need_valid_mask = true; #ifdef CONFIG_PM_SLEEP vg->saved_context = devm_kcalloc(&vg->pdev->dev, gc->ngpio, From 581d3c2025632f838fb08e5160dab752b3a1f527 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Wed, 11 Jan 2017 15:36:56 +0100 Subject: [PATCH 049/343] pinctrl: amd: avoid maybe-uninitalized warning Since gpio_dev->hwbank_num is now a variable, the compiler cannot figure out if pin_num is initialized at all: drivers/pinctrl/pinctrl-amd.c: In function 'amd_gpio_dbg_show': drivers/pinctrl/pinctrl-amd.c:210:3: warning: 'pin_num' may be used uninitialized in this function [-Wmaybe-uninitialized] for (; i < pin_num; i++) { ^~~ drivers/pinctrl/pinctrl-amd.c:172:21: warning: 'i' may be used uninitialized in this function [-Wmaybe-uninitialized] This adds a 'default' statement to make that case well-defined. Fixes: 3bfd44306c65 ("pinctrl: amd: Add support for additional GPIO") Signed-off-by: Arnd Bergmann Signed-off-by: Linus Walleij --- drivers/pinctrl/pinctrl-amd.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/pinctrl/pinctrl-amd.c b/drivers/pinctrl/pinctrl-amd.c index c9a146948192..537b52055756 100644 --- a/drivers/pinctrl/pinctrl-amd.c +++ b/drivers/pinctrl/pinctrl-amd.c @@ -202,6 +202,8 @@ static void amd_gpio_dbg_show(struct seq_file *s, struct gpio_chip *gc) i = 128; pin_num = AMD_GPIO_PINS_BANK2 + i; break; + default: + return; } for (; i < pin_num; i++) { From e88893fea17996018b2d68a22e677ea04f3baadf Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Thu, 5 Jan 2017 15:59:40 +0000 Subject: [PATCH 050/343] drm/i915: Clear ret before unbinding in i915_gem_evict_something() Missed when rebasing patches, I failed to set ret to zero before starting the unbind loop (which depends upon ret being zero). Reported-by: Matthew Auld Fixes: 9332f3b1b99a ("drm/i915: Combine loops within i915_gem_evict_something") Signed-off-by: Chris Wilson Cc: Matthew Auld Link: http://patchwork.freedesktop.org/patch/msgid/20170105155940.10033-1-chris@chris-wilson.co.uk Reviewed-by: Matthew Auld Cc: # v4.9+ (cherry picked from commit 121dfbb2a2ef1c5f49e15c38ccc47ff0beb59446) Signed-off-by: Jani Nikula --- drivers/gpu/drm/i915/i915_gem_evict.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/gpu/drm/i915/i915_gem_evict.c b/drivers/gpu/drm/i915/i915_gem_evict.c index bd08814b015c..d534a316a16e 100644 --- a/drivers/gpu/drm/i915/i915_gem_evict.c +++ b/drivers/gpu/drm/i915/i915_gem_evict.c @@ -199,6 +199,7 @@ found: } /* Unbinding will emit any required flushes */ + ret = 0; while (!list_empty(&eviction_list)) { vma = list_first_entry(&eviction_list, struct i915_vma, From e4621b73b6b472fe2b434b4f0f76b8f33ee26a73 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Fri, 6 Jan 2017 15:22:38 +0000 Subject: [PATCH 051/343] drm/i915: Fix phys pwrite for struct_mutex-less operation Since commit fe115628d567 ("drm/i915: Implement pwrite without struct-mutex") the lowlevel pwrite calls are now called without the protection of struct_mutex, but pwrite_phys was still asserting that it held the struct_mutex and later tried to drop and relock it. Fixes: fe115628d567 ("drm/i915: Implement pwrite without struct-mutex") Signed-off-by: Chris Wilson Cc: Joonas Lahtinen Cc: Link: http://patchwork.freedesktop.org/patch/msgid/20170106152240.5793-1-chris@chris-wilson.co.uk Reviewed-by: Tvrtko Ursulin (cherry picked from commit 10466d2a59b23aa6d5ecd5310296c8cdb6458dac) Signed-off-by: Jani Nikula --- drivers/gpu/drm/i915/i915_gem.c | 34 ++++----------------------------- 1 file changed, 4 insertions(+), 30 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 3dd7fc662859..4b23a7814713 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -595,47 +595,21 @@ i915_gem_phys_pwrite(struct drm_i915_gem_object *obj, struct drm_i915_gem_pwrite *args, struct drm_file *file) { - struct drm_device *dev = obj->base.dev; void *vaddr = obj->phys_handle->vaddr + args->offset; char __user *user_data = u64_to_user_ptr(args->data_ptr); - int ret; /* We manually control the domain here and pretend that it * remains coherent i.e. in the GTT domain, like shmem_pwrite. */ - lockdep_assert_held(&obj->base.dev->struct_mutex); - ret = i915_gem_object_wait(obj, - I915_WAIT_INTERRUPTIBLE | - I915_WAIT_LOCKED | - I915_WAIT_ALL, - MAX_SCHEDULE_TIMEOUT, - to_rps_client(file)); - if (ret) - return ret; - intel_fb_obj_invalidate(obj, ORIGIN_CPU); - if (__copy_from_user_inatomic_nocache(vaddr, user_data, args->size)) { - unsigned long unwritten; - - /* The physical object once assigned is fixed for the lifetime - * of the obj, so we can safely drop the lock and continue - * to access vaddr. - */ - mutex_unlock(&dev->struct_mutex); - unwritten = copy_from_user(vaddr, user_data, args->size); - mutex_lock(&dev->struct_mutex); - if (unwritten) { - ret = -EFAULT; - goto out; - } - } + if (copy_from_user(vaddr, user_data, args->size)) + return -EFAULT; drm_clflush_virt_range(vaddr, args->size); - i915_gem_chipset_flush(to_i915(dev)); + i915_gem_chipset_flush(to_i915(obj->base.dev)); -out: intel_fb_obj_flush(obj, false, ORIGIN_CPU); - return ret; + return 0; } void *i915_gem_object_alloc(struct drm_device *dev) From 3846fd9b86001bea171943cc3bb9222cb6da6b42 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Wed, 11 Jan 2017 10:01:17 +0100 Subject: [PATCH 052/343] drm/probe-helpers: Drop locking from poll_enable It was only needed to protect the connector_list walking, see commit 8c4ccc4ab6f64e859d4ff8d7c02c2ed2e956e07f Author: Daniel Vetter Date: Thu Jul 9 23:44:26 2015 +0200 drm/probe-helper: Grab mode_config.mutex in poll_init/enable Unfortunately the commit message of that patch fails to mention that the new locking check was for the connector_list. But that requirement disappeared in commit c36a3254f7857f1ad9badbe3578ccc92be541a8e Author: Daniel Vetter Date: Thu Dec 15 16:58:43 2016 +0100 drm: Convert all helpers to drm_connector_list_iter and so we can drop this again. This fixes a locking inversion on nouveau, where the rpm code needs to re-enable. But in other places the rpm_get() calls are nested within the big modeset locks. While at it, also improve the kerneldoc for these two functions a notch. v2: Update the kerneldoc even more to explain that these functions can't be called concurrently, or bad things happen (Chris). Cc: Dave Airlie Reviewed-by: Chris Wilson Cc: Chris Wilson Tested-by: Lyude Reviewed-by: Lyude Signed-off-by: Daniel Vetter Link: http://patchwork.freedesktop.org/patch/msgid/20170111090117.5134-1-daniel.vetter@ffwll.ch --- drivers/gpu/drm/drm_probe_helper.c | 51 +++++++++++----------------- drivers/gpu/drm/i915/intel_hotplug.c | 4 +-- include/drm/drm_crtc_helper.h | 1 - 3 files changed, 22 insertions(+), 34 deletions(-) diff --git a/drivers/gpu/drm/drm_probe_helper.c b/drivers/gpu/drm/drm_probe_helper.c index cf8f0128c161..b452a7ccd84b 100644 --- a/drivers/gpu/drm/drm_probe_helper.c +++ b/drivers/gpu/drm/drm_probe_helper.c @@ -115,24 +115,27 @@ static int drm_helper_probe_add_cmdline_mode(struct drm_connector *connector) #define DRM_OUTPUT_POLL_PERIOD (10*HZ) /** - * drm_kms_helper_poll_enable_locked - re-enable output polling. + * drm_kms_helper_poll_enable - re-enable output polling. * @dev: drm_device * - * This function re-enables the output polling work without - * locking the mode_config mutex. + * This function re-enables the output polling work, after it has been + * temporarily disabled using drm_kms_helper_poll_disable(), for example over + * suspend/resume. * - * This is like drm_kms_helper_poll_enable() however it is to be - * called from a context where the mode_config mutex is locked - * already. + * Drivers can call this helper from their device resume implementation. It is + * an error to call this when the output polling support has not yet been set + * up. + * + * Note that calls to enable and disable polling must be strictly ordered, which + * is automatically the case when they're only call from suspend/resume + * callbacks. */ -void drm_kms_helper_poll_enable_locked(struct drm_device *dev) +void drm_kms_helper_poll_enable(struct drm_device *dev) { bool poll = false; struct drm_connector *connector; unsigned long delay = DRM_OUTPUT_POLL_PERIOD; - WARN_ON(!mutex_is_locked(&dev->mode_config.mutex)); - if (!dev->mode_config.poll_enabled || !drm_kms_helper_poll) return; @@ -160,7 +163,7 @@ void drm_kms_helper_poll_enable_locked(struct drm_device *dev) if (poll) schedule_delayed_work(&dev->mode_config.output_poll_work, delay); } -EXPORT_SYMBOL(drm_kms_helper_poll_enable_locked); +EXPORT_SYMBOL(drm_kms_helper_poll_enable); static enum drm_connector_status drm_connector_detect(struct drm_connector *connector, bool force) @@ -287,7 +290,7 @@ int drm_helper_probe_single_connector_modes(struct drm_connector *connector, /* Re-enable polling in case the global poll config changed. */ if (drm_kms_helper_poll != dev->mode_config.poll_running) - drm_kms_helper_poll_enable_locked(dev); + drm_kms_helper_poll_enable(dev); dev->mode_config.poll_running = drm_kms_helper_poll; @@ -479,8 +482,12 @@ out: * This function disables the output polling work. * * Drivers can call this helper from their device suspend implementation. It is - * not an error to call this even when output polling isn't enabled or arlready - * disabled. + * not an error to call this even when output polling isn't enabled or already + * disabled. Polling is re-enabled by calling drm_kms_helper_poll_enable(). + * + * Note that calls to enable and disable polling must be strictly ordered, which + * is automatically the case when they're only call from suspend/resume + * callbacks. */ void drm_kms_helper_poll_disable(struct drm_device *dev) { @@ -490,24 +497,6 @@ void drm_kms_helper_poll_disable(struct drm_device *dev) } EXPORT_SYMBOL(drm_kms_helper_poll_disable); -/** - * drm_kms_helper_poll_enable - re-enable output polling. - * @dev: drm_device - * - * This function re-enables the output polling work. - * - * Drivers can call this helper from their device resume implementation. It is - * an error to call this when the output polling support has not yet been set - * up. - */ -void drm_kms_helper_poll_enable(struct drm_device *dev) -{ - mutex_lock(&dev->mode_config.mutex); - drm_kms_helper_poll_enable_locked(dev); - mutex_unlock(&dev->mode_config.mutex); -} -EXPORT_SYMBOL(drm_kms_helper_poll_enable); - /** * drm_kms_helper_poll_init - initialize and enable output polling * @dev: drm_device diff --git a/drivers/gpu/drm/i915/intel_hotplug.c b/drivers/gpu/drm/i915/intel_hotplug.c index 3d546c019de0..b62e3f8ad415 100644 --- a/drivers/gpu/drm/i915/intel_hotplug.c +++ b/drivers/gpu/drm/i915/intel_hotplug.c @@ -180,7 +180,7 @@ static void intel_hpd_irq_storm_disable(struct drm_i915_private *dev_priv) /* Enable polling and queue hotplug re-enabling. */ if (hpd_disabled) { - drm_kms_helper_poll_enable_locked(dev); + drm_kms_helper_poll_enable(dev); mod_delayed_work(system_wq, &dev_priv->hotplug.reenable_work, msecs_to_jiffies(HPD_STORM_REENABLE_DELAY)); } @@ -511,7 +511,7 @@ static void i915_hpd_poll_init_work(struct work_struct *work) } if (enabled) - drm_kms_helper_poll_enable_locked(dev); + drm_kms_helper_poll_enable(dev); mutex_unlock(&dev->mode_config.mutex); diff --git a/include/drm/drm_crtc_helper.h b/include/drm/drm_crtc_helper.h index 982c299e435a..d026f5017c33 100644 --- a/include/drm/drm_crtc_helper.h +++ b/include/drm/drm_crtc_helper.h @@ -73,6 +73,5 @@ extern void drm_kms_helper_hotplug_event(struct drm_device *dev); extern void drm_kms_helper_poll_disable(struct drm_device *dev); extern void drm_kms_helper_poll_enable(struct drm_device *dev); -extern void drm_kms_helper_poll_enable_locked(struct drm_device *dev); #endif From ef736d394e85b1bf1fd65ba5e5257b85f6c82325 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Tue, 20 Dec 2016 16:35:50 -0500 Subject: [PATCH 053/343] drm/radeon/si: load special ucode for certain MC configs Special MC ucode is required for these memory configurations. Acked-by: Edward O'Callaghan Signed-off-by: Alex Deucher --- drivers/gpu/drm/radeon/si.c | 12 +++++++++++- 1 file changed, 11 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c index e8a38d296855..32d55220a2d3 100644 --- a/drivers/gpu/drm/radeon/si.c +++ b/drivers/gpu/drm/radeon/si.c @@ -115,6 +115,8 @@ MODULE_FIRMWARE("radeon/hainan_rlc.bin"); MODULE_FIRMWARE("radeon/hainan_smc.bin"); MODULE_FIRMWARE("radeon/hainan_k_smc.bin"); +MODULE_FIRMWARE("radeon/si58_mc.bin"); + static u32 si_get_cu_active_bitmap(struct radeon_device *rdev, u32 se, u32 sh); static void si_pcie_gen3_enable(struct radeon_device *rdev); static void si_program_aspm(struct radeon_device *rdev); @@ -1650,6 +1652,7 @@ static int si_init_microcode(struct radeon_device *rdev) int err; int new_fw = 0; bool new_smc = false; + bool si58_fw = false; DRM_DEBUG("\n"); @@ -1742,6 +1745,10 @@ static int si_init_microcode(struct radeon_device *rdev) default: BUG(); } + /* this memory configuration requires special firmware */ + if (((RREG32(MC_SEQ_MISC0) & 0xff000000) >> 24) == 0x58) + si58_fw = true; + DRM_INFO("Loading %s Microcode\n", new_chip_name); snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", new_chip_name); @@ -1845,7 +1852,10 @@ static int si_init_microcode(struct radeon_device *rdev) } } - snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", new_chip_name); + if (si58_fw) + snprintf(fw_name, sizeof(fw_name), "radeon/si58_mc.bin"); + else + snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", new_chip_name); err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev); if (err) { snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc2.bin", chip_name); From f1d877be65d36806c581c32b4687d4acefa55960 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Tue, 20 Dec 2016 16:40:48 -0500 Subject: [PATCH 054/343] drm/amdgpu/si: load special ucode for certain MC configs Special MC ucode is required for these memory configurations. Acked-by: Edward O'Callaghan Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c index 45a573e63d4a..920960089f27 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c @@ -44,6 +44,7 @@ MODULE_FIRMWARE("radeon/tahiti_mc.bin"); MODULE_FIRMWARE("radeon/pitcairn_mc.bin"); MODULE_FIRMWARE("radeon/verde_mc.bin"); MODULE_FIRMWARE("radeon/oland_mc.bin"); +MODULE_FIRMWARE("radeon/si58_mc.bin"); #define MC_SEQ_MISC0__MT__MASK 0xf0000000 #define MC_SEQ_MISC0__MT__GDDR1 0x10000000 @@ -113,6 +114,7 @@ static int gmc_v6_0_init_microcode(struct amdgpu_device *adev) const char *chip_name; char fw_name[30]; int err; + bool is_58_fw = false; DRM_DEBUG("\n"); @@ -135,7 +137,14 @@ static int gmc_v6_0_init_microcode(struct amdgpu_device *adev) default: BUG(); } - snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name); + /* this memory configuration requires special firmware */ + if (((RREG32(mmMC_SEQ_MISC0) & 0xff000000) >> 24) == 0x58) + is_58_fw = true; + + if (is_58_fw) + snprintf(fw_name, sizeof(fw_name), "radeon/si58_mc.bin"); + else + snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name); err = request_firmware(&adev->mc.fw, fw_name, adev->dev); if (err) goto out; From 89d5595a6f53eba4d274c1d577d649db47620601 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Thu, 5 Jan 2017 13:04:47 -0500 Subject: [PATCH 055/343] drm/amdgpu: drop oland quirks Fixed by the new 58 MC firmware. Acked-by: Edward O'Callaghan Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/si_dpm.c | 11 ----------- 1 file changed, 11 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/si_dpm.c b/drivers/gpu/drm/amd/amdgpu/si_dpm.c index 10bedfac27b8..34cb7ea17caf 100644 --- a/drivers/gpu/drm/amd/amdgpu/si_dpm.c +++ b/drivers/gpu/drm/amd/amdgpu/si_dpm.c @@ -3487,17 +3487,6 @@ static void si_apply_state_adjust_rules(struct amdgpu_device *adev, (adev->pdev->device == 0x6817) || (adev->pdev->device == 0x6806)) max_mclk = 120000; - } else if (adev->asic_type == CHIP_OLAND) { - if ((adev->pdev->revision == 0xC7) || - (adev->pdev->revision == 0x80) || - (adev->pdev->revision == 0x81) || - (adev->pdev->revision == 0x83) || - (adev->pdev->revision == 0x87) || - (adev->pdev->device == 0x6604) || - (adev->pdev->device == 0x6605)) { - max_sclk = 75000; - max_mclk = 80000; - } } else if (adev->asic_type == CHIP_HAINAN) { if ((adev->pdev->revision == 0x81) || (adev->pdev->revision == 0x83) || From 5cc6f520ace3aa0086747e08417c2627374af1d7 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Thu, 5 Jan 2017 13:05:52 -0500 Subject: [PATCH 056/343] drm/amdgpu: drop the mclk quirk for hainan fixed by the new 58 mc firmware. Acked-by: Edward O'Callaghan Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/si_dpm.c | 1 - 1 file changed, 1 deletion(-) diff --git a/drivers/gpu/drm/amd/amdgpu/si_dpm.c b/drivers/gpu/drm/amd/amdgpu/si_dpm.c index 34cb7ea17caf..e9242d6fbf84 100644 --- a/drivers/gpu/drm/amd/amdgpu/si_dpm.c +++ b/drivers/gpu/drm/amd/amdgpu/si_dpm.c @@ -3495,7 +3495,6 @@ static void si_apply_state_adjust_rules(struct amdgpu_device *adev, (adev->pdev->device == 0x6665) || (adev->pdev->device == 0x6667)) { max_sclk = 75000; - max_mclk = 80000; } } /* Apply dpm quirks */ From 3a69adfe5617ceba04ad3cff0f9ccad470503fb2 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Thu, 5 Jan 2017 13:07:14 -0500 Subject: [PATCH 057/343] drm/radeon: drop oland quirks Fixed by the new 58 MC firmware. Acked-by: Edward O'Callaghan Signed-off-by: Alex Deucher --- drivers/gpu/drm/radeon/si_dpm.c | 11 ----------- 1 file changed, 11 deletions(-) diff --git a/drivers/gpu/drm/radeon/si_dpm.c b/drivers/gpu/drm/radeon/si_dpm.c index 13ba73fd9b68..668060fa68c1 100644 --- a/drivers/gpu/drm/radeon/si_dpm.c +++ b/drivers/gpu/drm/radeon/si_dpm.c @@ -3008,17 +3008,6 @@ static void si_apply_state_adjust_rules(struct radeon_device *rdev, (rdev->pdev->device == 0x6817) || (rdev->pdev->device == 0x6806)) max_mclk = 120000; - } else if (rdev->family == CHIP_OLAND) { - if ((rdev->pdev->revision == 0xC7) || - (rdev->pdev->revision == 0x80) || - (rdev->pdev->revision == 0x81) || - (rdev->pdev->revision == 0x83) || - (rdev->pdev->revision == 0x87) || - (rdev->pdev->device == 0x6604) || - (rdev->pdev->device == 0x6605)) { - max_sclk = 75000; - max_mclk = 80000; - } } else if (rdev->family == CHIP_HAINAN) { if ((rdev->pdev->revision == 0x81) || (rdev->pdev->revision == 0x83) || From a628392cf03e0eef21b345afbb192cbade041741 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Thu, 5 Jan 2017 13:08:01 -0500 Subject: [PATCH 058/343] drm/radeon: drop the mclk quirk for hainan fixed by the new 58 mc firmware. Acked-by: Edward O'Callaghan Signed-off-by: Alex Deucher --- drivers/gpu/drm/radeon/si_dpm.c | 1 - 1 file changed, 1 deletion(-) diff --git a/drivers/gpu/drm/radeon/si_dpm.c b/drivers/gpu/drm/radeon/si_dpm.c index 668060fa68c1..2944916f7102 100644 --- a/drivers/gpu/drm/radeon/si_dpm.c +++ b/drivers/gpu/drm/radeon/si_dpm.c @@ -3016,7 +3016,6 @@ static void si_apply_state_adjust_rules(struct radeon_device *rdev, (rdev->pdev->device == 0x6665) || (rdev->pdev->device == 0x6667)) { max_sclk = 75000; - max_mclk = 80000; } } /* Apply dpm quirks */ From 3731d12dce83d47b357753ffc450ce03f1b49688 Mon Sep 17 00:00:00 2001 From: Rex Zhu Date: Tue, 10 Jan 2017 19:26:49 +0800 Subject: [PATCH 059/343] drm/amd/powerplay: fix vce cg logic error on CZ/St. can fix Bug 191281: vce ib test failed. when vce idle, set vce clock gate, so the clock in vce domain will be disabled. when need to encode, disable vce clock gate, enable the clocks to vce engine. Signed-off-by: Rex Zhu Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/powerplay/hwmgr/cz_clockpowergating.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/cz_clockpowergating.c b/drivers/gpu/drm/amd/powerplay/hwmgr/cz_clockpowergating.c index b0c63c5f54c9..6bb79c94cb9f 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/cz_clockpowergating.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/cz_clockpowergating.c @@ -200,7 +200,7 @@ int cz_dpm_powergate_vce(struct pp_hwmgr *hwmgr, bool bgate) cgs_set_clockgating_state( hwmgr->device, AMD_IP_BLOCK_TYPE_VCE, - AMD_CG_STATE_UNGATE); + AMD_CG_STATE_GATE); cgs_set_powergating_state( hwmgr->device, AMD_IP_BLOCK_TYPE_VCE, @@ -218,7 +218,7 @@ int cz_dpm_powergate_vce(struct pp_hwmgr *hwmgr, bool bgate) cgs_set_clockgating_state( hwmgr->device, AMD_IP_BLOCK_TYPE_VCE, - AMD_PG_STATE_GATE); + AMD_PG_STATE_UNGATE); cz_dpm_update_vce_dpm(hwmgr); cz_enable_disable_vce_dpm(hwmgr, true); return 0; From a844764751275e0e5d381958e3c7e6e0fe739e25 Mon Sep 17 00:00:00 2001 From: Flora Cui Date: Tue, 10 Jan 2017 10:50:30 +0800 Subject: [PATCH 060/343] drm/amdgpu: fix vm_fault_stop on gfx6 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Flora Cui Reviewed-by: Christian König Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c | 23 +++++++++-------------- 1 file changed, 9 insertions(+), 14 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c index 920960089f27..e2b0b1646f99 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c @@ -472,19 +472,11 @@ static int gmc_v6_0_gart_enable(struct amdgpu_device *adev) WREG32(mmVM_CONTEXT1_CNTL, VM_CONTEXT1_CNTL__ENABLE_CONTEXT_MASK | (1UL << VM_CONTEXT1_CNTL__PAGE_TABLE_DEPTH__SHIFT) | - ((amdgpu_vm_block_size - 9) << VM_CONTEXT1_CNTL__PAGE_TABLE_BLOCK_SIZE__SHIFT) | - VM_CONTEXT1_CNTL__RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK | - VM_CONTEXT1_CNTL__RANGE_PROTECTION_FAULT_ENABLE_DEFAULT_MASK | - VM_CONTEXT1_CNTL__DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK | - VM_CONTEXT1_CNTL__DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT_MASK | - VM_CONTEXT1_CNTL__PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK | - VM_CONTEXT1_CNTL__PDE0_PROTECTION_FAULT_ENABLE_DEFAULT_MASK | - VM_CONTEXT1_CNTL__VALID_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK | - VM_CONTEXT1_CNTL__VALID_PROTECTION_FAULT_ENABLE_DEFAULT_MASK | - VM_CONTEXT1_CNTL__READ_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK | - VM_CONTEXT1_CNTL__READ_PROTECTION_FAULT_ENABLE_DEFAULT_MASK | - VM_CONTEXT1_CNTL__WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK | - VM_CONTEXT1_CNTL__WRITE_PROTECTION_FAULT_ENABLE_DEFAULT_MASK); + ((amdgpu_vm_block_size - 9) << VM_CONTEXT1_CNTL__PAGE_TABLE_BLOCK_SIZE__SHIFT)); + if (amdgpu_vm_fault_stop == AMDGPU_VM_FAULT_STOP_ALWAYS) + gmc_v6_0_set_fault_enable_default(adev, false); + else + gmc_v6_0_set_fault_enable_default(adev, true); gmc_v6_0_gart_flush_gpu_tlb(adev, 0); dev_info(adev->dev, "PCIE GART of %uM enabled (table at 0x%016llX).\n", @@ -763,7 +755,10 @@ static int gmc_v6_0_late_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; - return amdgpu_irq_get(adev, &adev->mc.vm_fault, 0); + if (amdgpu_vm_fault_stop != AMDGPU_VM_FAULT_STOP_ALWAYS) + return amdgpu_irq_get(adev, &adev->mc.vm_fault, 0); + else + return 0; } static int gmc_v6_0_sw_init(void *handle) From ab8db87b8256e13a62f10af1d32f5fc233c398cc Mon Sep 17 00:00:00 2001 From: Rex Zhu Date: Tue, 10 Jan 2017 15:47:50 +0800 Subject: [PATCH 061/343] drm/amd/powerplay: refine vce dpm update code on Cz. Program HardMin based on the vce_arbiter.ecclk if ecclk is 0, disable ECLK DPM 0. Otherwise VCE could hang if switching SCLK from DPM 0 to 6/7 Signed-off-by: Rex Zhu Acked-by: Alex Deucher Signed-off-by: Alex Deucher --- .../gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c | 24 ++++++++++++------- 1 file changed, 16 insertions(+), 8 deletions(-) diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c index 4b14f259a147..0fb4e8c8f5e1 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c @@ -1402,14 +1402,22 @@ int cz_dpm_update_vce_dpm(struct pp_hwmgr *hwmgr) cz_hwmgr->vce_dpm.hard_min_clk, PPSMC_MSG_SetEclkHardMin)); } else { - /*EPR# 419220 -HW limitation to to */ - cz_hwmgr->vce_dpm.hard_min_clk = hwmgr->vce_arbiter.ecclk; - smum_send_msg_to_smc_with_parameter(hwmgr->smumgr, - PPSMC_MSG_SetEclkHardMin, - cz_get_eclk_level(hwmgr, - cz_hwmgr->vce_dpm.hard_min_clk, - PPSMC_MSG_SetEclkHardMin)); - + /*Program HardMin based on the vce_arbiter.ecclk */ + if (hwmgr->vce_arbiter.ecclk == 0) { + smum_send_msg_to_smc_with_parameter(hwmgr->smumgr, + PPSMC_MSG_SetEclkHardMin, 0); + /* disable ECLK DPM 0. Otherwise VCE could hang if + * switching SCLK from DPM 0 to 6/7 */ + smum_send_msg_to_smc_with_parameter(hwmgr->smumgr, + PPSMC_MSG_SetEclkSoftMin, 1); + } else { + cz_hwmgr->vce_dpm.hard_min_clk = hwmgr->vce_arbiter.ecclk; + smum_send_msg_to_smc_with_parameter(hwmgr->smumgr, + PPSMC_MSG_SetEclkHardMin, + cz_get_eclk_level(hwmgr, + cz_hwmgr->vce_dpm.hard_min_clk, + PPSMC_MSG_SetEclkHardMin)); + } } return 0; } From d22a48bf7302ef064295749fa79cd47093c5a000 Mon Sep 17 00:00:00 2001 From: Changbin Du Date: Fri, 13 Jan 2017 11:15:56 +0800 Subject: [PATCH 062/343] drm/i915/gvt: introudce intel_vgpu_reset_resource() to reset vgpu resource state This patch introudces a new function intel_vgpu_reset_resource() to reset allocated vgpu resources by intel_vgpu_alloc_resource(). So far we only need clear the fence registers. The function _clear_vgpu_fence() will reset both virtual and physical fence registers to 0. Signed-off-by: Changbin Du Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/aperture_gm.c | 29 ++++++++++++++++++++++++-- drivers/gpu/drm/i915/gvt/gvt.h | 1 + 2 files changed, 28 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/gvt/aperture_gm.c b/drivers/gpu/drm/i915/gvt/aperture_gm.c index 65200313515c..f7bce8603958 100644 --- a/drivers/gpu/drm/i915/gvt/aperture_gm.c +++ b/drivers/gpu/drm/i915/gvt/aperture_gm.c @@ -158,6 +158,14 @@ void intel_vgpu_write_fence(struct intel_vgpu *vgpu, POSTING_READ(fence_reg_lo); } +static void _clear_vgpu_fence(struct intel_vgpu *vgpu) +{ + int i; + + for (i = 0; i < vgpu_fence_sz(vgpu); i++) + intel_vgpu_write_fence(vgpu, i, 0); +} + static void free_vgpu_fence(struct intel_vgpu *vgpu) { struct intel_gvt *gvt = vgpu->gvt; @@ -171,9 +179,9 @@ static void free_vgpu_fence(struct intel_vgpu *vgpu) intel_runtime_pm_get(dev_priv); mutex_lock(&dev_priv->drm.struct_mutex); + _clear_vgpu_fence(vgpu); for (i = 0; i < vgpu_fence_sz(vgpu); i++) { reg = vgpu->fence.regs[i]; - intel_vgpu_write_fence(vgpu, i, 0); list_add_tail(®->link, &dev_priv->mm.fence_list); } @@ -201,13 +209,14 @@ static int alloc_vgpu_fence(struct intel_vgpu *vgpu) continue; list_del(pos); vgpu->fence.regs[i] = reg; - intel_vgpu_write_fence(vgpu, i, 0); if (++i == vgpu_fence_sz(vgpu)) break; } if (i != vgpu_fence_sz(vgpu)) goto out_free_fence; + _clear_vgpu_fence(vgpu); + mutex_unlock(&dev_priv->drm.struct_mutex); intel_runtime_pm_put(dev_priv); return 0; @@ -306,6 +315,22 @@ void intel_vgpu_free_resource(struct intel_vgpu *vgpu) free_resource(vgpu); } +/** + * intel_vgpu_reset_resource - reset resource state owned by a vGPU + * @vgpu: a vGPU + * + * This function is used to reset resource state owned by a vGPU. + * + */ +void intel_vgpu_reset_resource(struct intel_vgpu *vgpu) +{ + struct drm_i915_private *dev_priv = vgpu->gvt->dev_priv; + + intel_runtime_pm_get(dev_priv); + _clear_vgpu_fence(vgpu); + intel_runtime_pm_put(dev_priv); +} + /** * intel_alloc_vgpu_resource - allocate HW resource for a vGPU * @vgpu: vGPU diff --git a/drivers/gpu/drm/i915/gvt/gvt.h b/drivers/gpu/drm/i915/gvt/gvt.h index 0af17016f33f..455281d07d75 100644 --- a/drivers/gpu/drm/i915/gvt/gvt.h +++ b/drivers/gpu/drm/i915/gvt/gvt.h @@ -323,6 +323,7 @@ struct intel_vgpu_creation_params { int intel_vgpu_alloc_resource(struct intel_vgpu *vgpu, struct intel_vgpu_creation_params *param); +void intel_vgpu_reset_resource(struct intel_vgpu *vgpu); void intel_vgpu_free_resource(struct intel_vgpu *vgpu); void intel_vgpu_write_fence(struct intel_vgpu *vgpu, u32 fence, u64 value); From b611581b375ce28536ab50be9cd507bb6092fb1e Mon Sep 17 00:00:00 2001 From: Changbin Du Date: Fri, 13 Jan 2017 11:15:57 +0800 Subject: [PATCH 063/343] drm/i915/gvt: introuduce intel_vgpu_reset_gtt() to reset gtt This patch introduces a new function intel_vgpu_reset_gtt() to reset the all GTT related status, including GGTT, PPGTT, scratch page. This function can free all shadowed PPGTT, clear all GGTT entry, and clear scratch page to all zero. After this, we can ensure no gtt related information can be leakaged from one VM to anothor one when assign vgpu instance across different VMs (not simultaneously). Signed-off-by: Changbin Du Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/gtt.c | 27 +++++++++++++++++++++++++++ drivers/gpu/drm/i915/gvt/gtt.h | 1 + 2 files changed, 28 insertions(+) diff --git a/drivers/gpu/drm/i915/gvt/gtt.c b/drivers/gpu/drm/i915/gvt/gtt.c index 3cf0df0bb391..47dec4acf7ff 100644 --- a/drivers/gpu/drm/i915/gvt/gtt.c +++ b/drivers/gpu/drm/i915/gvt/gtt.c @@ -2277,3 +2277,30 @@ void intel_vgpu_reset_ggtt(struct intel_vgpu *vgpu) for (offset = 0; offset < num_entries; offset++) ops->set_entry(NULL, &e, index + offset, false, 0, vgpu); } + +/** + * intel_vgpu_reset_gtt - reset the all GTT related status + * @vgpu: a vGPU + * @dmlr: true for vGPU Device Model Level Reset, false for GT Reset + * + * This function is called from vfio core to reset reset all + * GTT related status, including GGTT, PPGTT, scratch page. + * + */ +void intel_vgpu_reset_gtt(struct intel_vgpu *vgpu, bool dmlr) +{ + int i; + + ppgtt_free_all_shadow_page(vgpu); + if (!dmlr) + return; + + intel_vgpu_reset_ggtt(vgpu); + + /* clear scratch page for security */ + for (i = GTT_TYPE_PPGTT_PTE_PT; i < GTT_TYPE_MAX; i++) { + if (vgpu->gtt.scratch_pt[i].page != NULL) + memset(page_address(vgpu->gtt.scratch_pt[i].page), + 0, PAGE_SIZE); + } +} diff --git a/drivers/gpu/drm/i915/gvt/gtt.h b/drivers/gpu/drm/i915/gvt/gtt.h index b315ab3593ec..f88eb5e89bea 100644 --- a/drivers/gpu/drm/i915/gvt/gtt.h +++ b/drivers/gpu/drm/i915/gvt/gtt.h @@ -208,6 +208,7 @@ extern void intel_vgpu_clean_gtt(struct intel_vgpu *vgpu); void intel_vgpu_reset_ggtt(struct intel_vgpu *vgpu); extern int intel_gvt_init_gtt(struct intel_gvt *gvt); +extern void intel_vgpu_reset_gtt(struct intel_vgpu *vgpu, bool dmlr); extern void intel_gvt_clean_gtt(struct intel_gvt *gvt); extern struct intel_vgpu_mm *intel_gvt_find_ppgtt_mm(struct intel_vgpu *vgpu, From 536fc234074b09adae1763d8fb5b2d947847ad1d Mon Sep 17 00:00:00 2001 From: Changbin Du Date: Fri, 13 Jan 2017 11:15:58 +0800 Subject: [PATCH 064/343] drm/i915/gvt: move cfg space inititation function to cfg_space.c Move the configuration space inititation function setup_vgpu_cfg_space() in vgpu.c to dedicated source file cfg_space.c, and rename the function as intel_vgpu_init_cfg_space(). Signed-off-by: Changbin Du Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/cfg_space.c | 49 ++++++++++++++++++++++++++++ drivers/gpu/drm/i915/gvt/gvt.h | 2 ++ drivers/gpu/drm/i915/gvt/vgpu.c | 44 +------------------------ 3 files changed, 52 insertions(+), 43 deletions(-) diff --git a/drivers/gpu/drm/i915/gvt/cfg_space.c b/drivers/gpu/drm/i915/gvt/cfg_space.c index 711c31c8d8b4..82f449d59884 100644 --- a/drivers/gpu/drm/i915/gvt/cfg_space.c +++ b/drivers/gpu/drm/i915/gvt/cfg_space.c @@ -282,3 +282,52 @@ int intel_vgpu_emulate_cfg_write(struct intel_vgpu *vgpu, unsigned int offset, } return 0; } + +/** + * intel_vgpu_init_cfg_space - init vGPU configuration space when create vGPU + * + * @vgpu: a vGPU + * @primary: is the vGPU presented as primary + * + */ +void intel_vgpu_init_cfg_space(struct intel_vgpu *vgpu, + bool primary) +{ + struct intel_gvt *gvt = vgpu->gvt; + const struct intel_gvt_device_info *info = &gvt->device_info; + u16 *gmch_ctl; + int i; + + memcpy(vgpu_cfg_space(vgpu), gvt->firmware.cfg_space, + info->cfg_space_size); + + if (!primary) { + vgpu_cfg_space(vgpu)[PCI_CLASS_DEVICE] = + INTEL_GVT_PCI_CLASS_VGA_OTHER; + vgpu_cfg_space(vgpu)[PCI_CLASS_PROG] = + INTEL_GVT_PCI_CLASS_VGA_OTHER; + } + + /* Show guest that there isn't any stolen memory.*/ + gmch_ctl = (u16 *)(vgpu_cfg_space(vgpu) + INTEL_GVT_PCI_GMCH_CONTROL); + *gmch_ctl &= ~(BDW_GMCH_GMS_MASK << BDW_GMCH_GMS_SHIFT); + + intel_vgpu_write_pci_bar(vgpu, PCI_BASE_ADDRESS_2, + gvt_aperture_pa_base(gvt), true); + + vgpu_cfg_space(vgpu)[PCI_COMMAND] &= ~(PCI_COMMAND_IO + | PCI_COMMAND_MEMORY + | PCI_COMMAND_MASTER); + /* + * Clear the bar upper 32bit and let guest to assign the new value + */ + memset(vgpu_cfg_space(vgpu) + PCI_BASE_ADDRESS_1, 0, 4); + memset(vgpu_cfg_space(vgpu) + PCI_BASE_ADDRESS_3, 0, 4); + memset(vgpu_cfg_space(vgpu) + INTEL_GVT_PCI_OPREGION, 0, 4); + + for (i = 0; i < INTEL_GVT_MAX_BAR_NUM; i++) { + vgpu->cfg_space.bar[i].size = pci_resource_len( + gvt->dev_priv->drm.pdev, i * 2); + vgpu->cfg_space.bar[i].tracked = false; + } +} diff --git a/drivers/gpu/drm/i915/gvt/gvt.h b/drivers/gpu/drm/i915/gvt/gvt.h index 455281d07d75..ab505552d313 100644 --- a/drivers/gpu/drm/i915/gvt/gvt.h +++ b/drivers/gpu/drm/i915/gvt/gvt.h @@ -412,6 +412,8 @@ int intel_gvt_ggtt_index_g2h(struct intel_vgpu *vgpu, unsigned long g_index, int intel_gvt_ggtt_h2g_index(struct intel_vgpu *vgpu, unsigned long h_index, unsigned long *g_index); +void intel_vgpu_init_cfg_space(struct intel_vgpu *vgpu, + bool primary); int intel_vgpu_emulate_cfg_read(struct intel_vgpu *vgpu, unsigned int offset, void *p_data, unsigned int bytes); diff --git a/drivers/gpu/drm/i915/gvt/vgpu.c b/drivers/gpu/drm/i915/gvt/vgpu.c index f0e86123e45b..35c274177da8 100644 --- a/drivers/gpu/drm/i915/gvt/vgpu.c +++ b/drivers/gpu/drm/i915/gvt/vgpu.c @@ -66,48 +66,6 @@ int setup_vgpu_mmio(struct intel_vgpu *vgpu) return 0; } -static void setup_vgpu_cfg_space(struct intel_vgpu *vgpu, - struct intel_vgpu_creation_params *param) -{ - struct intel_gvt *gvt = vgpu->gvt; - const struct intel_gvt_device_info *info = &gvt->device_info; - u16 *gmch_ctl; - int i; - - memcpy(vgpu_cfg_space(vgpu), gvt->firmware.cfg_space, - info->cfg_space_size); - - if (!param->primary) { - vgpu_cfg_space(vgpu)[PCI_CLASS_DEVICE] = - INTEL_GVT_PCI_CLASS_VGA_OTHER; - vgpu_cfg_space(vgpu)[PCI_CLASS_PROG] = - INTEL_GVT_PCI_CLASS_VGA_OTHER; - } - - /* Show guest that there isn't any stolen memory.*/ - gmch_ctl = (u16 *)(vgpu_cfg_space(vgpu) + INTEL_GVT_PCI_GMCH_CONTROL); - *gmch_ctl &= ~(BDW_GMCH_GMS_MASK << BDW_GMCH_GMS_SHIFT); - - intel_vgpu_write_pci_bar(vgpu, PCI_BASE_ADDRESS_2, - gvt_aperture_pa_base(gvt), true); - - vgpu_cfg_space(vgpu)[PCI_COMMAND] &= ~(PCI_COMMAND_IO - | PCI_COMMAND_MEMORY - | PCI_COMMAND_MASTER); - /* - * Clear the bar upper 32bit and let guest to assign the new value - */ - memset(vgpu_cfg_space(vgpu) + PCI_BASE_ADDRESS_1, 0, 4); - memset(vgpu_cfg_space(vgpu) + PCI_BASE_ADDRESS_3, 0, 4); - memset(vgpu_cfg_space(vgpu) + INTEL_GVT_PCI_OPREGION, 0, 4); - - for (i = 0; i < INTEL_GVT_MAX_BAR_NUM; i++) { - vgpu->cfg_space.bar[i].size = pci_resource_len( - gvt->dev_priv->drm.pdev, i * 2); - vgpu->cfg_space.bar[i].tracked = false; - } -} - void populate_pvinfo_page(struct intel_vgpu *vgpu) { /* setup the ballooning information */ @@ -300,7 +258,7 @@ static struct intel_vgpu *__intel_gvt_create_vgpu(struct intel_gvt *gvt, vgpu->gvt = gvt; bitmap_zero(vgpu->tlb_handle_pending, I915_NUM_ENGINES); - setup_vgpu_cfg_space(vgpu, param); + intel_vgpu_init_cfg_space(vgpu, param->primary); ret = setup_vgpu_mmio(vgpu); if (ret) From c64ff6c774413fdbffd7f0f3ef5b04127d461cf4 Mon Sep 17 00:00:00 2001 From: Changbin Du Date: Fri, 13 Jan 2017 11:15:59 +0800 Subject: [PATCH 065/343] drm/i915/gvt: introduce intel_vgpu_reset_cfg_space to reset configuration space This patch introduces a new function intel_vgpu_reset_cfg_space() to reset vGPU configuration space. This function will unmap gttmmio and aperture if they are mapped before. Then entire cfg space will be restored to default values. Currently we only do such reset when vGPU is not owned by any VM so we simply restore entire cfg space to default value, not following the PCIe FLR spec that some fields should remain unchanged. Signed-off-by: Changbin Du Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/cfg_space.c | 25 +++++++++++++++++++++++++ drivers/gpu/drm/i915/gvt/gvt.h | 2 ++ 2 files changed, 27 insertions(+) diff --git a/drivers/gpu/drm/i915/gvt/cfg_space.c b/drivers/gpu/drm/i915/gvt/cfg_space.c index 82f449d59884..4a6a2ed65732 100644 --- a/drivers/gpu/drm/i915/gvt/cfg_space.c +++ b/drivers/gpu/drm/i915/gvt/cfg_space.c @@ -331,3 +331,28 @@ void intel_vgpu_init_cfg_space(struct intel_vgpu *vgpu, vgpu->cfg_space.bar[i].tracked = false; } } + +/** + * intel_vgpu_reset_cfg_space - reset vGPU configuration space + * + * @vgpu: a vGPU + * + */ +void intel_vgpu_reset_cfg_space(struct intel_vgpu *vgpu) +{ + u8 cmd = vgpu_cfg_space(vgpu)[PCI_COMMAND]; + bool primary = vgpu_cfg_space(vgpu)[PCI_CLASS_DEVICE] != + INTEL_GVT_PCI_CLASS_VGA_OTHER; + + if (cmd & PCI_COMMAND_MEMORY) { + trap_gttmmio(vgpu, false); + map_aperture(vgpu, false); + } + + /** + * Currently we only do such reset when vGPU is not + * owned by any VM, so we simply restore entire cfg + * space to default value. + */ + intel_vgpu_init_cfg_space(vgpu, primary); +} diff --git a/drivers/gpu/drm/i915/gvt/gvt.h b/drivers/gpu/drm/i915/gvt/gvt.h index ab505552d313..9c97d7d36449 100644 --- a/drivers/gpu/drm/i915/gvt/gvt.h +++ b/drivers/gpu/drm/i915/gvt/gvt.h @@ -414,6 +414,8 @@ int intel_gvt_ggtt_h2g_index(struct intel_vgpu *vgpu, unsigned long h_index, void intel_vgpu_init_cfg_space(struct intel_vgpu *vgpu, bool primary); +void intel_vgpu_reset_cfg_space(struct intel_vgpu *vgpu); + int intel_vgpu_emulate_cfg_read(struct intel_vgpu *vgpu, unsigned int offset, void *p_data, unsigned int bytes); From cdcc43479c9b929940a1955d2e7bae696d2b9496 Mon Sep 17 00:00:00 2001 From: Changbin Du Date: Fri, 13 Jan 2017 11:16:00 +0800 Subject: [PATCH 066/343] drm/i915/gvt: move mmio init/clean function to mmio.c Move the mmio space inititation function setup_vgpu_mmio() and cleanup function clean_vgpu_mmio() in vgpu.c to dedicated source file mmio.c, and rename them as intel_vgpu_init_mmio() and intel_vgpu_clean_mmio() respectively. Signed-off-by: Changbin Du Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/gvt.h | 1 - drivers/gpu/drm/i915/gvt/handlers.c | 2 +- drivers/gpu/drm/i915/gvt/mmio.c | 42 +++++++++++++++++++++++++++++ drivers/gpu/drm/i915/gvt/mmio.h | 3 +++ drivers/gpu/drm/i915/gvt/vgpu.c | 37 +++---------------------- 5 files changed, 49 insertions(+), 36 deletions(-) diff --git a/drivers/gpu/drm/i915/gvt/gvt.h b/drivers/gpu/drm/i915/gvt/gvt.h index 9c97d7d36449..9832e7f2a662 100644 --- a/drivers/gpu/drm/i915/gvt/gvt.h +++ b/drivers/gpu/drm/i915/gvt/gvt.h @@ -429,7 +429,6 @@ void intel_vgpu_clean_opregion(struct intel_vgpu *vgpu); int intel_vgpu_init_opregion(struct intel_vgpu *vgpu, u32 gpa); int intel_vgpu_emulate_opregion_request(struct intel_vgpu *vgpu, u32 swsci); -int setup_vgpu_mmio(struct intel_vgpu *vgpu); void populate_pvinfo_page(struct intel_vgpu *vgpu); struct intel_gvt_ops { diff --git a/drivers/gpu/drm/i915/gvt/handlers.c b/drivers/gpu/drm/i915/gvt/handlers.c index 8cbaf1c83720..fd96898db71d 100644 --- a/drivers/gpu/drm/i915/gvt/handlers.c +++ b/drivers/gpu/drm/i915/gvt/handlers.c @@ -257,7 +257,7 @@ static int handle_device_reset(struct intel_vgpu *vgpu, unsigned int offset, mutex_unlock(&vgpu->gvt->lock); intel_vgpu_clean_gtt(vgpu); mutex_lock(&vgpu->gvt->lock); - setup_vgpu_mmio(vgpu); + intel_vgpu_init_mmio(vgpu); populate_pvinfo_page(vgpu); intel_vgpu_init_gtt(vgpu); } diff --git a/drivers/gpu/drm/i915/gvt/mmio.c b/drivers/gpu/drm/i915/gvt/mmio.c index e60701397ac2..f7da735b7919 100644 --- a/drivers/gpu/drm/i915/gvt/mmio.c +++ b/drivers/gpu/drm/i915/gvt/mmio.c @@ -303,3 +303,45 @@ err: mutex_unlock(&gvt->lock); return ret; } + +/** + * intel_vgpu_init_mmio - init MMIO space + * @vgpu: a vGPU + * + * Returns: + * Zero on success, negative error code if failed + */ +int intel_vgpu_init_mmio(struct intel_vgpu *vgpu) +{ + const struct intel_gvt_device_info *info = &vgpu->gvt->device_info; + + if (vgpu->mmio.vreg) + memset(vgpu->mmio.vreg, 0, info->mmio_size * 2); + else { + vgpu->mmio.vreg = vzalloc(info->mmio_size * 2); + if (!vgpu->mmio.vreg) + return -ENOMEM; + } + vgpu->mmio.sreg = vgpu->mmio.vreg + info->mmio_size; + + memcpy(vgpu->mmio.vreg, vgpu->gvt->firmware.mmio, info->mmio_size); + memcpy(vgpu->mmio.sreg, vgpu->gvt->firmware.mmio, info->mmio_size); + + vgpu_vreg(vgpu, GEN6_GT_THREAD_STATUS_REG) = 0; + + /* set the bit 0:2(Core C-State ) to C0 */ + vgpu_vreg(vgpu, GEN6_GT_CORE_STATUS) = 0; + + return 0; +} + +/** + * intel_vgpu_clean_mmio - clean MMIO space + * @vgpu: a vGPU + * + */ +void intel_vgpu_clean_mmio(struct intel_vgpu *vgpu) +{ + vfree(vgpu->mmio.vreg); + vgpu->mmio.vreg = vgpu->mmio.sreg = NULL; +} diff --git a/drivers/gpu/drm/i915/gvt/mmio.h b/drivers/gpu/drm/i915/gvt/mmio.h index 87d5b5e366a3..1b74ac9650d5 100644 --- a/drivers/gpu/drm/i915/gvt/mmio.h +++ b/drivers/gpu/drm/i915/gvt/mmio.h @@ -86,6 +86,9 @@ struct intel_gvt_mmio_info *intel_gvt_find_mmio_info(struct intel_gvt *gvt, *offset; \ }) +int intel_vgpu_init_mmio(struct intel_vgpu *vgpu); +void intel_vgpu_clean_mmio(struct intel_vgpu *vgpu); + int intel_vgpu_gpa_to_mmio_offset(struct intel_vgpu *vgpu, u64 gpa); int intel_vgpu_emulate_mmio_read(struct intel_vgpu *vgpu, u64 pa, diff --git a/drivers/gpu/drm/i915/gvt/vgpu.c b/drivers/gpu/drm/i915/gvt/vgpu.c index 35c274177da8..958c3ed1c6c6 100644 --- a/drivers/gpu/drm/i915/gvt/vgpu.c +++ b/drivers/gpu/drm/i915/gvt/vgpu.c @@ -35,37 +35,6 @@ #include "gvt.h" #include "i915_pvinfo.h" -static void clean_vgpu_mmio(struct intel_vgpu *vgpu) -{ - vfree(vgpu->mmio.vreg); - vgpu->mmio.vreg = vgpu->mmio.sreg = NULL; -} - -int setup_vgpu_mmio(struct intel_vgpu *vgpu) -{ - struct intel_gvt *gvt = vgpu->gvt; - const struct intel_gvt_device_info *info = &gvt->device_info; - - if (vgpu->mmio.vreg) - memset(vgpu->mmio.vreg, 0, info->mmio_size * 2); - else { - vgpu->mmio.vreg = vzalloc(info->mmio_size * 2); - if (!vgpu->mmio.vreg) - return -ENOMEM; - } - - vgpu->mmio.sreg = vgpu->mmio.vreg + info->mmio_size; - - memcpy(vgpu->mmio.vreg, gvt->firmware.mmio, info->mmio_size); - memcpy(vgpu->mmio.sreg, gvt->firmware.mmio, info->mmio_size); - - vgpu_vreg(vgpu, GEN6_GT_THREAD_STATUS_REG) = 0; - - /* set the bit 0:2(Core C-State ) to C0 */ - vgpu_vreg(vgpu, GEN6_GT_CORE_STATUS) = 0; - return 0; -} - void populate_pvinfo_page(struct intel_vgpu *vgpu) { /* setup the ballooning information */ @@ -226,7 +195,7 @@ void intel_gvt_destroy_vgpu(struct intel_vgpu *vgpu) intel_vgpu_clean_gtt(vgpu); intel_gvt_hypervisor_detach_vgpu(vgpu); intel_vgpu_free_resource(vgpu); - clean_vgpu_mmio(vgpu); + intel_vgpu_clean_mmio(vgpu); vfree(vgpu); intel_gvt_update_vgpu_types(gvt); @@ -260,7 +229,7 @@ static struct intel_vgpu *__intel_gvt_create_vgpu(struct intel_gvt *gvt, intel_vgpu_init_cfg_space(vgpu, param->primary); - ret = setup_vgpu_mmio(vgpu); + ret = intel_vgpu_init_mmio(vgpu); if (ret) goto out_clean_idr; @@ -312,7 +281,7 @@ out_detach_hypervisor_vgpu: out_clean_vgpu_resource: intel_vgpu_free_resource(vgpu); out_clean_vgpu_mmio: - clean_vgpu_mmio(vgpu); + intel_vgpu_clean_mmio(vgpu); out_clean_idr: idr_remove(&gvt->vgpu_idr, vgpu->id); out_free_vgpu: From 97d58f7dd0ff12e5fddeffb40aed845daa628149 Mon Sep 17 00:00:00 2001 From: Changbin Du Date: Fri, 13 Jan 2017 11:16:01 +0800 Subject: [PATCH 067/343] drm/i915/gvt: introduce intel_vgpu_reset_mmio() to reset mmio space This patch introduces a new function intel_vgpu_reset_mmio() to reset vGPU MMIO space (virtual registers of the vGPU). The default values are loaded as firmware during gvt inititiation. Signed-off-by: Changbin Du Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/mmio.c | 39 +++++++++++++++++++++------------ drivers/gpu/drm/i915/gvt/mmio.h | 1 + 2 files changed, 26 insertions(+), 14 deletions(-) diff --git a/drivers/gpu/drm/i915/gvt/mmio.c b/drivers/gpu/drm/i915/gvt/mmio.c index f7da735b7919..4df078bc5d04 100644 --- a/drivers/gpu/drm/i915/gvt/mmio.c +++ b/drivers/gpu/drm/i915/gvt/mmio.c @@ -304,6 +304,26 @@ err: return ret; } + +/** + * intel_vgpu_reset_mmio - reset virtual MMIO space + * @vgpu: a vGPU + * + */ +void intel_vgpu_reset_mmio(struct intel_vgpu *vgpu) +{ + struct intel_gvt *gvt = vgpu->gvt; + const struct intel_gvt_device_info *info = &gvt->device_info; + + memcpy(vgpu->mmio.vreg, gvt->firmware.mmio, info->mmio_size); + memcpy(vgpu->mmio.sreg, gvt->firmware.mmio, info->mmio_size); + + vgpu_vreg(vgpu, GEN6_GT_THREAD_STATUS_REG) = 0; + + /* set the bit 0:2(Core C-State ) to C0 */ + vgpu_vreg(vgpu, GEN6_GT_CORE_STATUS) = 0; +} + /** * intel_vgpu_init_mmio - init MMIO space * @vgpu: a vGPU @@ -315,22 +335,13 @@ int intel_vgpu_init_mmio(struct intel_vgpu *vgpu) { const struct intel_gvt_device_info *info = &vgpu->gvt->device_info; - if (vgpu->mmio.vreg) - memset(vgpu->mmio.vreg, 0, info->mmio_size * 2); - else { - vgpu->mmio.vreg = vzalloc(info->mmio_size * 2); - if (!vgpu->mmio.vreg) - return -ENOMEM; - } + vgpu->mmio.vreg = vzalloc(info->mmio_size * 2); + if (!vgpu->mmio.vreg) + return -ENOMEM; + vgpu->mmio.sreg = vgpu->mmio.vreg + info->mmio_size; - memcpy(vgpu->mmio.vreg, vgpu->gvt->firmware.mmio, info->mmio_size); - memcpy(vgpu->mmio.sreg, vgpu->gvt->firmware.mmio, info->mmio_size); - - vgpu_vreg(vgpu, GEN6_GT_THREAD_STATUS_REG) = 0; - - /* set the bit 0:2(Core C-State ) to C0 */ - vgpu_vreg(vgpu, GEN6_GT_CORE_STATUS) = 0; + intel_vgpu_reset_mmio(vgpu); return 0; } diff --git a/drivers/gpu/drm/i915/gvt/mmio.h b/drivers/gpu/drm/i915/gvt/mmio.h index 1b74ac9650d5..3bc620f56f35 100644 --- a/drivers/gpu/drm/i915/gvt/mmio.h +++ b/drivers/gpu/drm/i915/gvt/mmio.h @@ -87,6 +87,7 @@ struct intel_gvt_mmio_info *intel_gvt_find_mmio_info(struct intel_gvt *gvt, }) int intel_vgpu_init_mmio(struct intel_vgpu *vgpu); +void intel_vgpu_reset_mmio(struct intel_vgpu *vgpu); void intel_vgpu_clean_mmio(struct intel_vgpu *vgpu); int intel_vgpu_gpa_to_mmio_offset(struct intel_vgpu *vgpu, u64 gpa); From cfe65f4037cedb911a840ebcf6dafc5b69e535b4 Mon Sep 17 00:00:00 2001 From: Changbin Du Date: Fri, 13 Jan 2017 11:16:02 +0800 Subject: [PATCH 068/343] drm/i915/gvt: fix vGPU instance reuse issues by vGPU reset function Our function tests found several issues related to reusing vGPU instance. They are qemu reboot failure, guest tdr after reboot, host hang when reboot guest. All these issues are caused by dirty status inherited from last VM. This patch fix all these issues by resetting a virtual GPU before VM use it. The reset logical is put into a low level function _intel_gvt_reset_vgpu(), which supports Device Model Level Reset, Full GT Reset and Per-Engine Reset. vGPU Device Model Level Reset (DMLR) simulates the PCI reset to reset the whole vGPU to default state as when it is created, including GTT, execlist, scratch pages, cfg space, mmio space, pvinfo page, scheduler and fence registers. The ultimate goal of vGPU DMLR is that reuse a vGPU instance by different virtual machines. When we reassign a vGPU to a virtual machine we must issue such reset first. Full GT Reset and Per-Engine GT Reset are soft reset flow for GPU engines (Render, Blitter, Video, Video Enhancement). It is defined by GPU Spec. Unlike the FLR, GT reset only reset particular resource of a vGPU per the reset request. Guest driver can issue a GT reset by programming the virtual GDRST register to reset specific virtual GPU engine or all engines. Since vGPU DMLR and GT reset can share some code so we implement both these two into one single function intel_gvt_reset_vgpu_locked(). The parameter dmlr is to identify if we will do FLR or GT reset. The parameter engine_mask is to specific the engines that need to be resetted. If value ALL_ENGINES is given for engine_mask, it means the caller requests a full gt reset that we will reset all virtual GPU engines. Signed-off-by: Changbin Du Reviewed-by: Jike Song Reviewed-by: Kevin Tian Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/gvt.h | 2 + drivers/gpu/drm/i915/gvt/vgpu.c | 73 ++++++++++++++++++++++++++++++++- 2 files changed, 74 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/gvt/gvt.h b/drivers/gpu/drm/i915/gvt/gvt.h index 9832e7f2a662..e227caf5859e 100644 --- a/drivers/gpu/drm/i915/gvt/gvt.h +++ b/drivers/gpu/drm/i915/gvt/gvt.h @@ -376,6 +376,8 @@ void intel_gvt_clean_vgpu_types(struct intel_gvt *gvt); struct intel_vgpu *intel_gvt_create_vgpu(struct intel_gvt *gvt, struct intel_vgpu_type *type); void intel_gvt_destroy_vgpu(struct intel_vgpu *vgpu); +void intel_gvt_reset_vgpu_locked(struct intel_vgpu *vgpu, bool dmlr, + unsigned int engine_mask); void intel_gvt_reset_vgpu(struct intel_vgpu *vgpu); diff --git a/drivers/gpu/drm/i915/gvt/vgpu.c b/drivers/gpu/drm/i915/gvt/vgpu.c index 958c3ed1c6c6..7295bc8e12fb 100644 --- a/drivers/gpu/drm/i915/gvt/vgpu.c +++ b/drivers/gpu/drm/i915/gvt/vgpu.c @@ -327,7 +327,75 @@ struct intel_vgpu *intel_gvt_create_vgpu(struct intel_gvt *gvt, } /** - * intel_gvt_reset_vgpu - reset a virtual GPU + * intel_gvt_reset_vgpu_locked - reset a virtual GPU by DMLR or GT reset + * @vgpu: virtual GPU + * @dmlr: vGPU Device Model Level Reset or GT Reset + * @engine_mask: engines to reset for GT reset + * + * This function is called when user wants to reset a virtual GPU through + * device model reset or GT reset. The caller should hold the gvt lock. + * + * vGPU Device Model Level Reset (DMLR) simulates the PCI level reset to reset + * the whole vGPU to default state as when it is created. This vGPU function + * is required both for functionary and security concerns.The ultimate goal + * of vGPU FLR is that reuse a vGPU instance by virtual machines. When we + * assign a vGPU to a virtual machine we must isse such reset first. + * + * Full GT Reset and Per-Engine GT Reset are soft reset flow for GPU engines + * (Render, Blitter, Video, Video Enhancement). It is defined by GPU Spec. + * Unlike the FLR, GT reset only reset particular resource of a vGPU per + * the reset request. Guest driver can issue a GT reset by programming the + * virtual GDRST register to reset specific virtual GPU engine or all + * engines. + * + * The parameter dev_level is to identify if we will do DMLR or GT reset. + * The parameter engine_mask is to specific the engines that need to be + * resetted. If value ALL_ENGINES is given for engine_mask, it means + * the caller requests a full GT reset that we will reset all virtual + * GPU engines. For FLR, engine_mask is ignored. + */ +void intel_gvt_reset_vgpu_locked(struct intel_vgpu *vgpu, bool dmlr, + unsigned int engine_mask) +{ + struct intel_gvt *gvt = vgpu->gvt; + struct intel_gvt_workload_scheduler *scheduler = &gvt->scheduler; + + gvt_dbg_core("------------------------------------------\n"); + gvt_dbg_core("resseting vgpu%d, dmlr %d, engine_mask %08x\n", + vgpu->id, dmlr, engine_mask); + vgpu->resetting = true; + + intel_vgpu_stop_schedule(vgpu); + /* + * The current_vgpu will set to NULL after stopping the + * scheduler when the reset is triggered by current vgpu. + */ + if (scheduler->current_vgpu == NULL) { + mutex_unlock(&gvt->lock); + intel_gvt_wait_vgpu_idle(vgpu); + mutex_lock(&gvt->lock); + } + + intel_vgpu_reset_execlist(vgpu, dmlr ? ALL_ENGINES : engine_mask); + + /* full GPU reset or device model level reset */ + if (engine_mask == ALL_ENGINES || dmlr) { + intel_vgpu_reset_gtt(vgpu, dmlr); + intel_vgpu_reset_resource(vgpu); + intel_vgpu_reset_mmio(vgpu); + populate_pvinfo_page(vgpu); + + if (dmlr) + intel_vgpu_reset_cfg_space(vgpu); + } + + vgpu->resetting = false; + gvt_dbg_core("reset vgpu%d done\n", vgpu->id); + gvt_dbg_core("------------------------------------------\n"); +} + +/** + * intel_gvt_reset_vgpu - reset a virtual GPU (Function Level) * @vgpu: virtual GPU * * This function is called when user wants to reset a virtual GPU. @@ -335,4 +403,7 @@ struct intel_vgpu *intel_gvt_create_vgpu(struct intel_gvt *gvt, */ void intel_gvt_reset_vgpu(struct intel_vgpu *vgpu) { + mutex_lock(&vgpu->gvt->lock); + intel_gvt_reset_vgpu_locked(vgpu, true, 0); + mutex_unlock(&vgpu->gvt->lock); } From c34eaa8d0f9d9ae26a4a6af7bc3aca57310cf483 Mon Sep 17 00:00:00 2001 From: Changbin Du Date: Fri, 13 Jan 2017 11:16:03 +0800 Subject: [PATCH 069/343] drm/i915/gvt: rewrite gt reset handler using new function intel_gvt_reset_vgpu_locked GT reset and FLR share some operations and they are both implemented in our new function intel_gvt_reset_vgpu_locked(). This patch rewrite the gt reset handler using this new function. Besides, this new implementation fixed the old issue in GT reset. The old implementation reset GGTT entries which is illegal. We only clear GGTT entries at PCI level reset. Signed-off-by: Changbin Du Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/handlers.c | 90 ++++++++++------------------- 1 file changed, 29 insertions(+), 61 deletions(-) diff --git a/drivers/gpu/drm/i915/gvt/handlers.c b/drivers/gpu/drm/i915/gvt/handlers.c index fd96898db71d..ab2ea157da4c 100644 --- a/drivers/gpu/drm/i915/gvt/handlers.c +++ b/drivers/gpu/drm/i915/gvt/handlers.c @@ -231,77 +231,45 @@ static int mul_force_wake_write(struct intel_vgpu *vgpu, return 0; } -static int handle_device_reset(struct intel_vgpu *vgpu, unsigned int offset, - void *p_data, unsigned int bytes, unsigned long bitmap) -{ - struct intel_gvt_workload_scheduler *scheduler = - &vgpu->gvt->scheduler; - - vgpu->resetting = true; - - intel_vgpu_stop_schedule(vgpu); - /* - * The current_vgpu will set to NULL after stopping the - * scheduler when the reset is triggered by current vgpu. - */ - if (scheduler->current_vgpu == NULL) { - mutex_unlock(&vgpu->gvt->lock); - intel_gvt_wait_vgpu_idle(vgpu); - mutex_lock(&vgpu->gvt->lock); - } - - intel_vgpu_reset_execlist(vgpu, bitmap); - - /* full GPU reset */ - if (bitmap == 0xff) { - mutex_unlock(&vgpu->gvt->lock); - intel_vgpu_clean_gtt(vgpu); - mutex_lock(&vgpu->gvt->lock); - intel_vgpu_init_mmio(vgpu); - populate_pvinfo_page(vgpu); - intel_vgpu_init_gtt(vgpu); - } - - vgpu->resetting = false; - - return 0; -} - static int gdrst_mmio_write(struct intel_vgpu *vgpu, unsigned int offset, - void *p_data, unsigned int bytes) + void *p_data, unsigned int bytes) { + unsigned int engine_mask = 0; u32 data; - u64 bitmap = 0; write_vreg(vgpu, offset, p_data, bytes); data = vgpu_vreg(vgpu, offset); if (data & GEN6_GRDOM_FULL) { gvt_dbg_mmio("vgpu%d: request full GPU reset\n", vgpu->id); - bitmap = 0xff; + engine_mask = ALL_ENGINES; + } else { + if (data & GEN6_GRDOM_RENDER) { + gvt_dbg_mmio("vgpu%d: request RCS reset\n", vgpu->id); + engine_mask |= (1 << RCS); + } + if (data & GEN6_GRDOM_MEDIA) { + gvt_dbg_mmio("vgpu%d: request VCS reset\n", vgpu->id); + engine_mask |= (1 << VCS); + } + if (data & GEN6_GRDOM_BLT) { + gvt_dbg_mmio("vgpu%d: request BCS Reset\n", vgpu->id); + engine_mask |= (1 << BCS); + } + if (data & GEN6_GRDOM_VECS) { + gvt_dbg_mmio("vgpu%d: request VECS Reset\n", vgpu->id); + engine_mask |= (1 << VECS); + } + if (data & GEN8_GRDOM_MEDIA2) { + gvt_dbg_mmio("vgpu%d: request VCS2 Reset\n", vgpu->id); + if (HAS_BSD2(vgpu->gvt->dev_priv)) + engine_mask |= (1 << VCS2); + } } - if (data & GEN6_GRDOM_RENDER) { - gvt_dbg_mmio("vgpu%d: request RCS reset\n", vgpu->id); - bitmap |= (1 << RCS); - } - if (data & GEN6_GRDOM_MEDIA) { - gvt_dbg_mmio("vgpu%d: request VCS reset\n", vgpu->id); - bitmap |= (1 << VCS); - } - if (data & GEN6_GRDOM_BLT) { - gvt_dbg_mmio("vgpu%d: request BCS Reset\n", vgpu->id); - bitmap |= (1 << BCS); - } - if (data & GEN6_GRDOM_VECS) { - gvt_dbg_mmio("vgpu%d: request VECS Reset\n", vgpu->id); - bitmap |= (1 << VECS); - } - if (data & GEN8_GRDOM_MEDIA2) { - gvt_dbg_mmio("vgpu%d: request VCS2 Reset\n", vgpu->id); - if (HAS_BSD2(vgpu->gvt->dev_priv)) - bitmap |= (1 << VCS2); - } - return handle_device_reset(vgpu, offset, p_data, bytes, bitmap); + + intel_gvt_reset_vgpu_locked(vgpu, false, engine_mask); + + return 0; } static int gmbus_mmio_read(struct intel_vgpu *vgpu, unsigned int offset, From f65a7c9cb3770ed4d3e7c57c66d7032689081b5e Mon Sep 17 00:00:00 2001 From: Andrzej Hajda Date: Mon, 9 Jan 2017 15:33:02 +0100 Subject: [PATCH 070/343] drm/exynos/decon5433: update shadow registers iff there are active windows Improper usage of DECON_UPDATE register leads to subtle errors. If it set in decon_commit when there are no active windows it results in slow registry updates - all subsequent shadow registry updates takes more than full vblank. On the other side if it is not set when there are active windows it results in garbage on the screen after suspend/resume of FB console. The patch hopefully fixes it. Signed-off-by: Andrzej Hajda Signed-off-by: Inki Dae --- drivers/gpu/drm/exynos/exynos5433_drm_decon.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/exynos/exynos5433_drm_decon.c b/drivers/gpu/drm/exynos/exynos5433_drm_decon.c index 6ca1f3117fe8..ef7fcb5f044b 100644 --- a/drivers/gpu/drm/exynos/exynos5433_drm_decon.c +++ b/drivers/gpu/drm/exynos/exynos5433_drm_decon.c @@ -188,8 +188,6 @@ static void decon_commit(struct exynos_drm_crtc *crtc) /* enable output and display signal */ decon_set_bits(ctx, DECON_VIDCON0, VIDCON0_ENVID | VIDCON0_ENVID_F, ~0); - - decon_set_bits(ctx, DECON_UPDATE, STANDALONE_UPDATE_F, ~0); } static void decon_win_set_pixfmt(struct decon_context *ctx, unsigned int win, @@ -340,8 +338,9 @@ static void decon_atomic_flush(struct exynos_drm_crtc *crtc) for (i = ctx->first_win; i < WINDOWS_NR; i++) decon_shadow_protect_win(ctx, i, false); - /* standalone update */ - decon_set_bits(ctx, DECON_UPDATE, STANDALONE_UPDATE_F, ~0); + /* update iff there are active windows */ + if (crtc->base.state->plane_mask) + decon_set_bits(ctx, DECON_UPDATE, STANDALONE_UPDATE_F, ~0); if (ctx->out_type & IFTYPE_I80) set_bit(BIT_WIN_UPDATED, &ctx->flags); From 821b40b79db7dedbfe15ab330dfd181e661a533f Mon Sep 17 00:00:00 2001 From: Andrzej Hajda Date: Fri, 13 Jan 2017 10:20:58 +0100 Subject: [PATCH 071/343] drm/exynos/decon5433: set STANDALONE_UPDATE_F also if planes are disabled STANDALONE_UPDATE_F should be set if something changed in plane configurations, including plane disable. The patch fixes page-faults bugs, caused by decon still using framebuffers of disabled planes. v2: fixed clear-bit code (Thx Marek) v3: use test_and_clear_bit (Thx Joonyoung) Signed-off-by: Andrzej Hajda Tested-by: Joonyoung Shim Signed-off-by: Inki Dae --- drivers/gpu/drm/exynos/exynos5433_drm_decon.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/exynos/exynos5433_drm_decon.c b/drivers/gpu/drm/exynos/exynos5433_drm_decon.c index ef7fcb5f044b..09e8cc36948e 100644 --- a/drivers/gpu/drm/exynos/exynos5433_drm_decon.c +++ b/drivers/gpu/drm/exynos/exynos5433_drm_decon.c @@ -46,7 +46,8 @@ enum decon_flag_bits { BIT_CLKS_ENABLED, BIT_IRQS_ENABLED, BIT_WIN_UPDATED, - BIT_SUSPENDED + BIT_SUSPENDED, + BIT_REQUEST_UPDATE }; struct decon_context { @@ -313,6 +314,7 @@ static void decon_update_plane(struct exynos_drm_crtc *crtc, /* window enable */ decon_set_bits(ctx, DECON_WINCONx(win), WINCONx_ENWIN_F, ~0); + set_bit(BIT_REQUEST_UPDATE, &ctx->flags); } static void decon_disable_plane(struct exynos_drm_crtc *crtc, @@ -325,6 +327,7 @@ static void decon_disable_plane(struct exynos_drm_crtc *crtc, return; decon_set_bits(ctx, DECON_WINCONx(win), WINCONx_ENWIN_F, 0); + set_bit(BIT_REQUEST_UPDATE, &ctx->flags); } static void decon_atomic_flush(struct exynos_drm_crtc *crtc) @@ -338,8 +341,7 @@ static void decon_atomic_flush(struct exynos_drm_crtc *crtc) for (i = ctx->first_win; i < WINDOWS_NR; i++) decon_shadow_protect_win(ctx, i, false); - /* update iff there are active windows */ - if (crtc->base.state->plane_mask) + if (test_and_clear_bit(BIT_REQUEST_UPDATE, &ctx->flags)) decon_set_bits(ctx, DECON_UPDATE, STANDALONE_UPDATE_F, ~0); if (ctx->out_type & IFTYPE_I80) From c57a94ffd0105d58ab104fe383148c5eda5aa033 Mon Sep 17 00:00:00 2001 From: Rob Clark Date: Wed, 11 Jan 2017 10:16:12 -0500 Subject: [PATCH 072/343] drm/msm/mdp5: rip out plane->pending tracking It would race between userspace thread and commit worker. Ie. vblank irq would trigger event and userspace could begin the next atomic update, before the commit worker had a chance to clear the pending flag. If we do end up needing something to prevent userspace from trying another pageflip before getting vblank event, it should probably be implemented as a pending_planes bitmask, similar to pending_crtcs. See start_atomic() and end_atomic(). Signed-off-by: Rob Clark --- drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.c | 6 ------ drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.h | 4 ---- drivers/gpu/drm/msm/mdp/mdp5/mdp5_plane.c | 22 ---------------------- 3 files changed, 32 deletions(-) diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.c index 5f6cd8745dbc..c396d459a9d0 100644 --- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.c +++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.c @@ -119,13 +119,7 @@ static void mdp5_prepare_commit(struct msm_kms *kms, struct drm_atomic_state *st static void mdp5_complete_commit(struct msm_kms *kms, struct drm_atomic_state *state) { - int i; struct mdp5_kms *mdp5_kms = to_mdp5_kms(to_mdp_kms(kms)); - struct drm_plane *plane; - struct drm_plane_state *plane_state; - - for_each_plane_in_state(state, plane, plane_state, i) - mdp5_plane_complete_commit(plane, plane_state); if (mdp5_kms->smp) mdp5_smp_complete_commit(mdp5_kms->smp, &mdp5_kms->state->smp); diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.h b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.h index 17b0cc101171..cdfc63d90c7b 100644 --- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.h +++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.h @@ -104,8 +104,6 @@ struct mdp5_plane_state { /* assigned by crtc blender */ enum mdp_mixer_stage_id stage; - - bool pending : 1; }; #define to_mdp5_plane_state(x) \ container_of(x, struct mdp5_plane_state, base) @@ -232,8 +230,6 @@ int mdp5_irq_domain_init(struct mdp5_kms *mdp5_kms); void mdp5_irq_domain_fini(struct mdp5_kms *mdp5_kms); uint32_t mdp5_plane_get_flush(struct drm_plane *plane); -void mdp5_plane_complete_commit(struct drm_plane *plane, - struct drm_plane_state *state); enum mdp5_pipe mdp5_plane_pipe(struct drm_plane *plane); struct drm_plane *mdp5_plane_init(struct drm_device *dev, bool primary); diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_plane.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_plane.c index c099da7bc212..25d9d0a97156 100644 --- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_plane.c +++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_plane.c @@ -179,7 +179,6 @@ mdp5_plane_atomic_print_state(struct drm_printer *p, drm_printf(p, "\tzpos=%u\n", pstate->zpos); drm_printf(p, "\talpha=%u\n", pstate->alpha); drm_printf(p, "\tstage=%s\n", stage2name(pstate->stage)); - drm_printf(p, "\tpending=%u\n", pstate->pending); } static void mdp5_plane_reset(struct drm_plane *plane) @@ -220,8 +219,6 @@ mdp5_plane_duplicate_state(struct drm_plane *plane) if (mdp5_state && mdp5_state->base.fb) drm_framebuffer_reference(mdp5_state->base.fb); - mdp5_state->pending = false; - return &mdp5_state->base; } @@ -288,13 +285,6 @@ static int mdp5_plane_atomic_check(struct drm_plane *plane, DBG("%s: check (%d -> %d)", plane->name, plane_enabled(old_state), plane_enabled(state)); - /* We don't allow faster-than-vblank updates.. if we did add this - * some day, we would need to disallow in cases where hwpipe - * changes - */ - if (WARN_ON(to_mdp5_plane_state(old_state)->pending)) - return -EBUSY; - max_width = config->hw->lm.max_width << 16; max_height = config->hw->lm.max_height << 16; @@ -370,12 +360,9 @@ static void mdp5_plane_atomic_update(struct drm_plane *plane, struct drm_plane_state *old_state) { struct drm_plane_state *state = plane->state; - struct mdp5_plane_state *mdp5_state = to_mdp5_plane_state(state); DBG("%s: update", plane->name); - mdp5_state->pending = true; - if (plane_enabled(state)) { int ret; @@ -851,15 +838,6 @@ uint32_t mdp5_plane_get_flush(struct drm_plane *plane) return pstate->hwpipe->flush_mask; } -/* called after vsync in thread context */ -void mdp5_plane_complete_commit(struct drm_plane *plane, - struct drm_plane_state *state) -{ - struct mdp5_plane_state *pstate = to_mdp5_plane_state(plane->state); - - pstate->pending = false; -} - /* initialize plane */ struct drm_plane *mdp5_plane_init(struct drm_device *dev, bool primary) { From de85d2b35ac74f6be769573d4a8708c823219900 Mon Sep 17 00:00:00 2001 From: Rob Clark Date: Thu, 12 Jan 2017 17:41:44 -0500 Subject: [PATCH 073/343] drm/msm: fix potential null ptr issue in non-iommu case Fixes: 9cb07b099fb ("drm/msm: support multiple address spaces") Reported-by: Riku Voipio Signed-off-by: Rob Clark --- drivers/gpu/drm/msm/adreno/adreno_gpu.c | 5 ++--- drivers/gpu/drm/msm/msm_gem.c | 2 ++ 2 files changed, 4 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/msm/adreno/adreno_gpu.c b/drivers/gpu/drm/msm/adreno/adreno_gpu.c index 14ff87686a36..686a580c711a 100644 --- a/drivers/gpu/drm/msm/adreno/adreno_gpu.c +++ b/drivers/gpu/drm/msm/adreno/adreno_gpu.c @@ -345,7 +345,6 @@ int adreno_gpu_init(struct drm_device *drm, struct platform_device *pdev, { struct adreno_platform_config *config = pdev->dev.platform_data; struct msm_gpu *gpu = &adreno_gpu->base; - struct msm_mmu *mmu; int ret; adreno_gpu->funcs = funcs; @@ -385,8 +384,8 @@ int adreno_gpu_init(struct drm_device *drm, struct platform_device *pdev, return ret; } - mmu = gpu->aspace->mmu; - if (mmu) { + if (gpu->aspace && gpu->aspace->mmu) { + struct msm_mmu *mmu = gpu->aspace->mmu; ret = mmu->funcs->attach(mmu, iommu_ports, ARRAY_SIZE(iommu_ports)); if (ret) diff --git a/drivers/gpu/drm/msm/msm_gem.c b/drivers/gpu/drm/msm/msm_gem.c index d8bc59c7e261..8098677a3916 100644 --- a/drivers/gpu/drm/msm/msm_gem.c +++ b/drivers/gpu/drm/msm/msm_gem.c @@ -294,6 +294,8 @@ put_iova(struct drm_gem_object *obj) WARN_ON(!mutex_is_locked(&dev->struct_mutex)); for (id = 0; id < ARRAY_SIZE(msm_obj->domain); id++) { + if (!priv->aspace[id]) + continue; msm_gem_unmap_vma(priv->aspace[id], &msm_obj->domain[id], msm_obj->sgt); } From 69bcc0b7140c30de552aa3ef08322295862e8e2f Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Michel=20D=C3=A4nzer?= Date: Wed, 11 Jan 2017 18:27:25 +0900 Subject: [PATCH 074/343] Revert "drm/amdgpu: Only update the CUR_SIZE register when necessary" This reverts commits 7c83d7abc9997cf1efac2c0ce384b5e8453ee870 and a1f49cc179ce6b7b7758ae3ff5cdb138d0ee0f56. They caused the HW cursor to disappear under various circumstances in the wild. I wasn't able to reproduce any of them, and I'm not sure what's going on. But those changes aren't a big deal anyway, so let's just revert for now. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=191291 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=99143 Acked-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/dce_v10_0.c | 22 +++++++--------------- drivers/gpu/drm/amd/amdgpu/dce_v11_0.c | 22 +++++++--------------- drivers/gpu/drm/amd/amdgpu/dce_v6_0.c | 24 +++++++++--------------- drivers/gpu/drm/amd/amdgpu/dce_v8_0.c | 22 +++++++--------------- 4 files changed, 30 insertions(+), 60 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c index 9999dc71b998..ccb5e02e7b20 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c @@ -2512,6 +2512,8 @@ static int dce_v10_0_cursor_move_locked(struct drm_crtc *crtc, WREG32(mmCUR_POSITION + amdgpu_crtc->crtc_offset, (x << 16) | y); WREG32(mmCUR_HOT_SPOT + amdgpu_crtc->crtc_offset, (xorigin << 16) | yorigin); + WREG32(mmCUR_SIZE + amdgpu_crtc->crtc_offset, + ((amdgpu_crtc->cursor_width - 1) << 16) | (amdgpu_crtc->cursor_height - 1)); return 0; } @@ -2537,7 +2539,6 @@ static int dce_v10_0_crtc_cursor_set2(struct drm_crtc *crtc, int32_t hot_y) { struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); - struct amdgpu_device *adev = crtc->dev->dev_private; struct drm_gem_object *obj; struct amdgpu_bo *aobj; int ret; @@ -2578,7 +2579,9 @@ static int dce_v10_0_crtc_cursor_set2(struct drm_crtc *crtc, dce_v10_0_lock_cursor(crtc, true); - if (hot_x != amdgpu_crtc->cursor_hot_x || + if (width != amdgpu_crtc->cursor_width || + height != amdgpu_crtc->cursor_height || + hot_x != amdgpu_crtc->cursor_hot_x || hot_y != amdgpu_crtc->cursor_hot_y) { int x, y; @@ -2587,16 +2590,10 @@ static int dce_v10_0_crtc_cursor_set2(struct drm_crtc *crtc, dce_v10_0_cursor_move_locked(crtc, x, y); - amdgpu_crtc->cursor_hot_x = hot_x; - amdgpu_crtc->cursor_hot_y = hot_y; - } - - if (width != amdgpu_crtc->cursor_width || - height != amdgpu_crtc->cursor_height) { - WREG32(mmCUR_SIZE + amdgpu_crtc->crtc_offset, - (width - 1) << 16 | (height - 1)); amdgpu_crtc->cursor_width = width; amdgpu_crtc->cursor_height = height; + amdgpu_crtc->cursor_hot_x = hot_x; + amdgpu_crtc->cursor_hot_y = hot_y; } dce_v10_0_show_cursor(crtc); @@ -2620,7 +2617,6 @@ unpin: static void dce_v10_0_cursor_reset(struct drm_crtc *crtc) { struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); - struct amdgpu_device *adev = crtc->dev->dev_private; if (amdgpu_crtc->cursor_bo) { dce_v10_0_lock_cursor(crtc, true); @@ -2628,10 +2624,6 @@ static void dce_v10_0_cursor_reset(struct drm_crtc *crtc) dce_v10_0_cursor_move_locked(crtc, amdgpu_crtc->cursor_x, amdgpu_crtc->cursor_y); - WREG32(mmCUR_SIZE + amdgpu_crtc->crtc_offset, - (amdgpu_crtc->cursor_width - 1) << 16 | - (amdgpu_crtc->cursor_height - 1)); - dce_v10_0_show_cursor(crtc); dce_v10_0_lock_cursor(crtc, false); diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c index 2006abbbfb62..a7af5b33a5e3 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c @@ -2532,6 +2532,8 @@ static int dce_v11_0_cursor_move_locked(struct drm_crtc *crtc, WREG32(mmCUR_POSITION + amdgpu_crtc->crtc_offset, (x << 16) | y); WREG32(mmCUR_HOT_SPOT + amdgpu_crtc->crtc_offset, (xorigin << 16) | yorigin); + WREG32(mmCUR_SIZE + amdgpu_crtc->crtc_offset, + ((amdgpu_crtc->cursor_width - 1) << 16) | (amdgpu_crtc->cursor_height - 1)); return 0; } @@ -2557,7 +2559,6 @@ static int dce_v11_0_crtc_cursor_set2(struct drm_crtc *crtc, int32_t hot_y) { struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); - struct amdgpu_device *adev = crtc->dev->dev_private; struct drm_gem_object *obj; struct amdgpu_bo *aobj; int ret; @@ -2598,7 +2599,9 @@ static int dce_v11_0_crtc_cursor_set2(struct drm_crtc *crtc, dce_v11_0_lock_cursor(crtc, true); - if (hot_x != amdgpu_crtc->cursor_hot_x || + if (width != amdgpu_crtc->cursor_width || + height != amdgpu_crtc->cursor_height || + hot_x != amdgpu_crtc->cursor_hot_x || hot_y != amdgpu_crtc->cursor_hot_y) { int x, y; @@ -2607,16 +2610,10 @@ static int dce_v11_0_crtc_cursor_set2(struct drm_crtc *crtc, dce_v11_0_cursor_move_locked(crtc, x, y); - amdgpu_crtc->cursor_hot_x = hot_x; - amdgpu_crtc->cursor_hot_y = hot_y; - } - - if (width != amdgpu_crtc->cursor_width || - height != amdgpu_crtc->cursor_height) { - WREG32(mmCUR_SIZE + amdgpu_crtc->crtc_offset, - (width - 1) << 16 | (height - 1)); amdgpu_crtc->cursor_width = width; amdgpu_crtc->cursor_height = height; + amdgpu_crtc->cursor_hot_x = hot_x; + amdgpu_crtc->cursor_hot_y = hot_y; } dce_v11_0_show_cursor(crtc); @@ -2640,7 +2637,6 @@ unpin: static void dce_v11_0_cursor_reset(struct drm_crtc *crtc) { struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); - struct amdgpu_device *adev = crtc->dev->dev_private; if (amdgpu_crtc->cursor_bo) { dce_v11_0_lock_cursor(crtc, true); @@ -2648,10 +2644,6 @@ static void dce_v11_0_cursor_reset(struct drm_crtc *crtc) dce_v11_0_cursor_move_locked(crtc, amdgpu_crtc->cursor_x, amdgpu_crtc->cursor_y); - WREG32(mmCUR_SIZE + amdgpu_crtc->crtc_offset, - (amdgpu_crtc->cursor_width - 1) << 16 | - (amdgpu_crtc->cursor_height - 1)); - dce_v11_0_show_cursor(crtc); dce_v11_0_lock_cursor(crtc, false); diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c index b4e4ec630e8c..39df6a50637f 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c @@ -1859,6 +1859,8 @@ static int dce_v6_0_cursor_move_locked(struct drm_crtc *crtc, struct amdgpu_device *adev = crtc->dev->dev_private; int xorigin = 0, yorigin = 0; + int w = amdgpu_crtc->cursor_width; + amdgpu_crtc->cursor_x = x; amdgpu_crtc->cursor_y = y; @@ -1878,6 +1880,8 @@ static int dce_v6_0_cursor_move_locked(struct drm_crtc *crtc, WREG32(mmCUR_POSITION + amdgpu_crtc->crtc_offset, (x << 16) | y); WREG32(mmCUR_HOT_SPOT + amdgpu_crtc->crtc_offset, (xorigin << 16) | yorigin); + WREG32(mmCUR_SIZE + amdgpu_crtc->crtc_offset, + ((w - 1) << 16) | (amdgpu_crtc->cursor_height - 1)); return 0; } @@ -1903,7 +1907,6 @@ static int dce_v6_0_crtc_cursor_set2(struct drm_crtc *crtc, int32_t hot_y) { struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); - struct amdgpu_device *adev = crtc->dev->dev_private; struct drm_gem_object *obj; struct amdgpu_bo *aobj; int ret; @@ -1944,7 +1947,9 @@ static int dce_v6_0_crtc_cursor_set2(struct drm_crtc *crtc, dce_v6_0_lock_cursor(crtc, true); - if (hot_x != amdgpu_crtc->cursor_hot_x || + if (width != amdgpu_crtc->cursor_width || + height != amdgpu_crtc->cursor_height || + hot_x != amdgpu_crtc->cursor_hot_x || hot_y != amdgpu_crtc->cursor_hot_y) { int x, y; @@ -1953,16 +1958,10 @@ static int dce_v6_0_crtc_cursor_set2(struct drm_crtc *crtc, dce_v6_0_cursor_move_locked(crtc, x, y); - amdgpu_crtc->cursor_hot_x = hot_x; - amdgpu_crtc->cursor_hot_y = hot_y; - } - - if (width != amdgpu_crtc->cursor_width || - height != amdgpu_crtc->cursor_height) { - WREG32(mmCUR_SIZE + amdgpu_crtc->crtc_offset, - (width - 1) << 16 | (height - 1)); amdgpu_crtc->cursor_width = width; amdgpu_crtc->cursor_height = height; + amdgpu_crtc->cursor_hot_x = hot_x; + amdgpu_crtc->cursor_hot_y = hot_y; } dce_v6_0_show_cursor(crtc); @@ -1986,7 +1985,6 @@ unpin: static void dce_v6_0_cursor_reset(struct drm_crtc *crtc) { struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); - struct amdgpu_device *adev = crtc->dev->dev_private; if (amdgpu_crtc->cursor_bo) { dce_v6_0_lock_cursor(crtc, true); @@ -1994,10 +1992,6 @@ static void dce_v6_0_cursor_reset(struct drm_crtc *crtc) dce_v6_0_cursor_move_locked(crtc, amdgpu_crtc->cursor_x, amdgpu_crtc->cursor_y); - WREG32(mmCUR_SIZE + amdgpu_crtc->crtc_offset, - (amdgpu_crtc->cursor_width - 1) << 16 | - (amdgpu_crtc->cursor_height - 1)); - dce_v6_0_show_cursor(crtc); dce_v6_0_lock_cursor(crtc, false); } diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c index 584abe834a3c..28102bb1704d 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c @@ -2363,6 +2363,8 @@ static int dce_v8_0_cursor_move_locked(struct drm_crtc *crtc, WREG32(mmCUR_POSITION + amdgpu_crtc->crtc_offset, (x << 16) | y); WREG32(mmCUR_HOT_SPOT + amdgpu_crtc->crtc_offset, (xorigin << 16) | yorigin); + WREG32(mmCUR_SIZE + amdgpu_crtc->crtc_offset, + ((amdgpu_crtc->cursor_width - 1) << 16) | (amdgpu_crtc->cursor_height - 1)); return 0; } @@ -2388,7 +2390,6 @@ static int dce_v8_0_crtc_cursor_set2(struct drm_crtc *crtc, int32_t hot_y) { struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); - struct amdgpu_device *adev = crtc->dev->dev_private; struct drm_gem_object *obj; struct amdgpu_bo *aobj; int ret; @@ -2429,7 +2430,9 @@ static int dce_v8_0_crtc_cursor_set2(struct drm_crtc *crtc, dce_v8_0_lock_cursor(crtc, true); - if (hot_x != amdgpu_crtc->cursor_hot_x || + if (width != amdgpu_crtc->cursor_width || + height != amdgpu_crtc->cursor_height || + hot_x != amdgpu_crtc->cursor_hot_x || hot_y != amdgpu_crtc->cursor_hot_y) { int x, y; @@ -2438,16 +2441,10 @@ static int dce_v8_0_crtc_cursor_set2(struct drm_crtc *crtc, dce_v8_0_cursor_move_locked(crtc, x, y); - amdgpu_crtc->cursor_hot_x = hot_x; - amdgpu_crtc->cursor_hot_y = hot_y; - } - - if (width != amdgpu_crtc->cursor_width || - height != amdgpu_crtc->cursor_height) { - WREG32(mmCUR_SIZE + amdgpu_crtc->crtc_offset, - (width - 1) << 16 | (height - 1)); amdgpu_crtc->cursor_width = width; amdgpu_crtc->cursor_height = height; + amdgpu_crtc->cursor_hot_x = hot_x; + amdgpu_crtc->cursor_hot_y = hot_y; } dce_v8_0_show_cursor(crtc); @@ -2471,7 +2468,6 @@ unpin: static void dce_v8_0_cursor_reset(struct drm_crtc *crtc) { struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); - struct amdgpu_device *adev = crtc->dev->dev_private; if (amdgpu_crtc->cursor_bo) { dce_v8_0_lock_cursor(crtc, true); @@ -2479,10 +2475,6 @@ static void dce_v8_0_cursor_reset(struct drm_crtc *crtc) dce_v8_0_cursor_move_locked(crtc, amdgpu_crtc->cursor_x, amdgpu_crtc->cursor_y); - WREG32(mmCUR_SIZE + amdgpu_crtc->crtc_offset, - (amdgpu_crtc->cursor_width - 1) << 16 | - (amdgpu_crtc->cursor_height - 1)); - dce_v8_0_show_cursor(crtc); dce_v8_0_lock_cursor(crtc, false); From d7f5762c5e532dfe8247ce1bc60d97af27ff8d00 Mon Sep 17 00:00:00 2001 From: Alexander Alemayhu Date: Wed, 4 Jan 2017 23:16:41 +0100 Subject: [PATCH 075/343] netfilter: nf_tables: fix spelling mistakes o s/numerice/numeric o s/opertaor/operator Signed-off-by: Alexander Alemayhu Signed-off-by: Pablo Neira Ayuso --- include/uapi/linux/netfilter/nf_tables.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/include/uapi/linux/netfilter/nf_tables.h b/include/uapi/linux/netfilter/nf_tables.h index 881d49e94569..e3f27e09eb2b 100644 --- a/include/uapi/linux/netfilter/nf_tables.h +++ b/include/uapi/linux/netfilter/nf_tables.h @@ -235,7 +235,7 @@ enum nft_rule_compat_flags { /** * enum nft_rule_compat_attributes - nf_tables rule compat attributes * - * @NFTA_RULE_COMPAT_PROTO: numerice value of handled protocol (NLA_U32) + * @NFTA_RULE_COMPAT_PROTO: numeric value of handled protocol (NLA_U32) * @NFTA_RULE_COMPAT_FLAGS: bitmask of enum nft_rule_compat_flags (NLA_U32) */ enum nft_rule_compat_attributes { @@ -499,7 +499,7 @@ enum nft_bitwise_attributes { * enum nft_byteorder_ops - nf_tables byteorder operators * * @NFT_BYTEORDER_NTOH: network to host operator - * @NFT_BYTEORDER_HTON: host to network opertaor + * @NFT_BYTEORDER_HTON: host to network operator */ enum nft_byteorder_ops { NFT_BYTEORDER_NTOH, From 6443ebc3fdd6f3c766d9442c18be274b3d736050 Mon Sep 17 00:00:00 2001 From: Liping Zhang Date: Sat, 7 Jan 2017 20:49:18 +0800 Subject: [PATCH 076/343] netfilter: rpfilter: fix incorrect loopback packet judgment Currently, we check the existing rtable in PREROUTING hook, if RTCF_LOCAL is set, we assume that the packet is loopback. But this assumption is incorrect, for example, a packet encapsulated in ipsec transport mode was received and routed to local, after decapsulation, it would be delivered to local again, and the rtable was not dropped, so RTCF_LOCAL check would trigger. But actually, the packet was not loopback. So for these normal loopback packets, we can check whether the in device is IFF_LOOPBACK or not. For these locally generated broadcast/multicast, we can check whether the skb->pkt_type is PACKET_LOOPBACK or not. Finally, there's a subtle difference between nft fib expr and xtables rpfilter extension, user can add the following nft rule to do strict rpfilter check: # nft add rule x y meta iif eth0 fib saddr . iif oif != eth0 drop So when the packet is loopback, it's better to store the in device instead of the LOOPBACK_IFINDEX, otherwise, after adding the above nft rule, locally generated broad/multicast packets will be dropped incorrectly. Fixes: f83a7ea2075c ("netfilter: xt_rpfilter: skip locally generated broadcast/multicast, too") Fixes: f6d0cbcf09c5 ("netfilter: nf_tables: add fib expression") Signed-off-by: Liping Zhang Signed-off-by: Pablo Neira Ayuso --- include/net/netfilter/nft_fib.h | 6 ++++++ net/ipv4/netfilter/ipt_rpfilter.c | 8 ++++---- net/ipv4/netfilter/nft_fib_ipv4.c | 15 +++++---------- net/ipv6/netfilter/ip6t_rpfilter.c | 8 ++++---- net/ipv6/netfilter/nft_fib_ipv6.c | 13 ++++--------- 5 files changed, 23 insertions(+), 27 deletions(-) diff --git a/include/net/netfilter/nft_fib.h b/include/net/netfilter/nft_fib.h index cbedda077db2..5ceb2205e4e3 100644 --- a/include/net/netfilter/nft_fib.h +++ b/include/net/netfilter/nft_fib.h @@ -9,6 +9,12 @@ struct nft_fib { extern const struct nla_policy nft_fib_policy[]; +static inline bool +nft_fib_is_loopback(const struct sk_buff *skb, const struct net_device *in) +{ + return skb->pkt_type == PACKET_LOOPBACK || in->flags & IFF_LOOPBACK; +} + int nft_fib_dump(struct sk_buff *skb, const struct nft_expr *expr); int nft_fib_init(const struct nft_ctx *ctx, const struct nft_expr *expr, const struct nlattr * const tb[]); diff --git a/net/ipv4/netfilter/ipt_rpfilter.c b/net/ipv4/netfilter/ipt_rpfilter.c index f273098e48fd..37fb9552e858 100644 --- a/net/ipv4/netfilter/ipt_rpfilter.c +++ b/net/ipv4/netfilter/ipt_rpfilter.c @@ -63,10 +63,10 @@ static bool rpfilter_lookup_reverse(struct net *net, struct flowi4 *fl4, return dev_match || flags & XT_RPFILTER_LOOSE; } -static bool rpfilter_is_local(const struct sk_buff *skb) +static bool +rpfilter_is_loopback(const struct sk_buff *skb, const struct net_device *in) { - const struct rtable *rt = skb_rtable(skb); - return rt && (rt->rt_flags & RTCF_LOCAL); + return skb->pkt_type == PACKET_LOOPBACK || in->flags & IFF_LOOPBACK; } static bool rpfilter_mt(const struct sk_buff *skb, struct xt_action_param *par) @@ -79,7 +79,7 @@ static bool rpfilter_mt(const struct sk_buff *skb, struct xt_action_param *par) info = par->matchinfo; invert = info->flags & XT_RPFILTER_INVERT; - if (rpfilter_is_local(skb)) + if (rpfilter_is_loopback(skb, xt_in(par))) return true ^ invert; iph = ip_hdr(skb); diff --git a/net/ipv4/netfilter/nft_fib_ipv4.c b/net/ipv4/netfilter/nft_fib_ipv4.c index 965b1a161369..2981291910dd 100644 --- a/net/ipv4/netfilter/nft_fib_ipv4.c +++ b/net/ipv4/netfilter/nft_fib_ipv4.c @@ -26,13 +26,6 @@ static __be32 get_saddr(__be32 addr) return addr; } -static bool fib4_is_local(const struct sk_buff *skb) -{ - const struct rtable *rt = skb_rtable(skb); - - return rt && (rt->rt_flags & RTCF_LOCAL); -} - #define DSCP_BITS 0xfc void nft_fib4_eval_type(const struct nft_expr *expr, struct nft_regs *regs, @@ -95,8 +88,10 @@ void nft_fib4_eval(const struct nft_expr *expr, struct nft_regs *regs, else oif = NULL; - if (nft_hook(pkt) == NF_INET_PRE_ROUTING && fib4_is_local(pkt->skb)) { - nft_fib_store_result(dest, priv->result, pkt, LOOPBACK_IFINDEX); + if (nft_hook(pkt) == NF_INET_PRE_ROUTING && + nft_fib_is_loopback(pkt->skb, nft_in(pkt))) { + nft_fib_store_result(dest, priv->result, pkt, + nft_in(pkt)->ifindex); return; } @@ -131,7 +126,7 @@ void nft_fib4_eval(const struct nft_expr *expr, struct nft_regs *regs, switch (res.type) { case RTN_UNICAST: break; - case RTN_LOCAL: /* should not appear here, see fib4_is_local() above */ + case RTN_LOCAL: /* Should not see RTN_LOCAL here */ return; default: break; diff --git a/net/ipv6/netfilter/ip6t_rpfilter.c b/net/ipv6/netfilter/ip6t_rpfilter.c index d5263dc364a9..b12e61b7b16c 100644 --- a/net/ipv6/netfilter/ip6t_rpfilter.c +++ b/net/ipv6/netfilter/ip6t_rpfilter.c @@ -72,10 +72,10 @@ static bool rpfilter_lookup_reverse6(struct net *net, const struct sk_buff *skb, return ret; } -static bool rpfilter_is_local(const struct sk_buff *skb) +static bool +rpfilter_is_loopback(const struct sk_buff *skb, const struct net_device *in) { - const struct rt6_info *rt = (const void *) skb_dst(skb); - return rt && (rt->rt6i_flags & RTF_LOCAL); + return skb->pkt_type == PACKET_LOOPBACK || in->flags & IFF_LOOPBACK; } static bool rpfilter_mt(const struct sk_buff *skb, struct xt_action_param *par) @@ -85,7 +85,7 @@ static bool rpfilter_mt(const struct sk_buff *skb, struct xt_action_param *par) struct ipv6hdr *iph; bool invert = info->flags & XT_RPFILTER_INVERT; - if (rpfilter_is_local(skb)) + if (rpfilter_is_loopback(skb, xt_in(par))) return true ^ invert; iph = ipv6_hdr(skb); diff --git a/net/ipv6/netfilter/nft_fib_ipv6.c b/net/ipv6/netfilter/nft_fib_ipv6.c index c947aad8bcc6..765facf03d45 100644 --- a/net/ipv6/netfilter/nft_fib_ipv6.c +++ b/net/ipv6/netfilter/nft_fib_ipv6.c @@ -18,13 +18,6 @@ #include #include -static bool fib6_is_local(const struct sk_buff *skb) -{ - const struct rt6_info *rt = (const void *)skb_dst(skb); - - return rt && (rt->rt6i_flags & RTF_LOCAL); -} - static int get_ifindex(const struct net_device *dev) { return dev ? dev->ifindex : 0; @@ -164,8 +157,10 @@ void nft_fib6_eval(const struct nft_expr *expr, struct nft_regs *regs, lookup_flags = nft_fib6_flowi_init(&fl6, priv, pkt, oif); - if (nft_hook(pkt) == NF_INET_PRE_ROUTING && fib6_is_local(pkt->skb)) { - nft_fib_store_result(dest, priv->result, pkt, LOOPBACK_IFINDEX); + if (nft_hook(pkt) == NF_INET_PRE_ROUTING && + nft_fib_is_loopback(pkt->skb, nft_in(pkt))) { + nft_fib_store_result(dest, priv->result, pkt, + nft_in(pkt)->ifindex); return; } From d21e540b4dd74a26df7a66ebab75c693a4a6a861 Mon Sep 17 00:00:00 2001 From: Liping Zhang Date: Sat, 7 Jan 2017 20:51:50 +0800 Subject: [PATCH 077/343] netfilter: nf_tables: fix possible oops when dumping stateful objects When dumping nft stateful objects, if NFTA_OBJ_TABLE and NFTA_OBJ_TYPE attributes are not specified either, filter will become NULL, so oops will happen(actually nft utility will always set NFTA_OBJ_TABLE attr, so I write a test program to make this happen): BUG: unable to handle kernel NULL pointer dereference at (null) IP: nf_tables_dump_obj+0x17c/0x330 [nf_tables] [...] Call Trace: ? nf_tables_dump_obj+0x5/0x330 [nf_tables] ? __kmalloc_reserve.isra.35+0x31/0x90 ? __alloc_skb+0x5b/0x1e0 netlink_dump+0x124/0x2a0 __netlink_dump_start+0x161/0x190 nf_tables_getobj+0xe8/0x280 [nf_tables] Fixes: a9fea2a3c3cf ("netfilter: nf_tables: allow to filter stateful object dumps by type") Signed-off-by: Liping Zhang Signed-off-by: Pablo Neira Ayuso --- net/netfilter/nf_tables_api.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c index 0db5f9782265..091d2dcc63b2 100644 --- a/net/netfilter/nf_tables_api.c +++ b/net/netfilter/nf_tables_api.c @@ -4262,10 +4262,11 @@ static int nf_tables_dump_obj(struct sk_buff *skb, struct netlink_callback *cb) if (idx > s_idx) memset(&cb->args[1], 0, sizeof(cb->args) - sizeof(cb->args[0])); - if (filter->table[0] && + if (filter && filter->table[0] && strcmp(filter->table, table->name)) goto cont; - if (filter->type != NFT_OBJECT_UNSPEC && + if (filter && + filter->type != NFT_OBJECT_UNSPEC && obj->type->type != filter->type) goto cont; From e4670b058af64639ec1aef4db845c39bfdfff7c4 Mon Sep 17 00:00:00 2001 From: William Breathitt Gray Date: Mon, 9 Jan 2017 17:24:18 -0500 Subject: [PATCH 078/343] netfilter: Fix typo in NF_CONNTRACK Kconfig option description The NF_CONNTRACK Kconfig option description makes an incorrect reference to the "meta" expression where the "ct" expression would be correct.This patch fixes the respective typographical error. Fixes: d497c6352736 ("netfilter: add help information to new nf_tables Kconfig options") Signed-off-by: William Breathitt Gray Signed-off-by: Pablo Neira Ayuso --- net/netfilter/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/netfilter/Kconfig b/net/netfilter/Kconfig index 63729b489c2c..bbc45f8a7b2d 100644 --- a/net/netfilter/Kconfig +++ b/net/netfilter/Kconfig @@ -494,7 +494,7 @@ config NFT_CT depends on NF_CONNTRACK tristate "Netfilter nf_tables conntrack module" help - This option adds the "meta" expression that you can use to match + This option adds the "ct" expression that you can use to match connection tracking information such as the flow state. config NFT_SET_RBTREE From 7e9081c5aac73b8a0bc22e0b3e7a12c3e9cf5256 Mon Sep 17 00:00:00 2001 From: Gustavo Padovan Date: Fri, 13 Jan 2017 12:22:09 -0200 Subject: [PATCH 079/343] drm/fence: fix memory overwrite when setting out_fence fd Currently if the userspace declares a int variable to store the out_fence fd and pass it to OUT_FENCE_PTR the kernel will overwrite the 32 bits above the int variable on 64 bits systems. Fix this by making the internal storage of out_fence in the kernel a s32 pointer. Reported-by: Chad Versace Signed-off-by: Gustavo Padovan Fixes: beaf5af48034 ("drm/fence: add out-fences support") Cc: Daniel Vetter Cc: Rafael Antognolli Cc: Laurent Pinchart Acked-by: Laurent Pinchart Reviewed-and-Tested-by: Chad Versace Link: http://patchwork.freedesktop.org/patch/msgid/1484317329-9293-1-git-send-email-gustavo@padovan.org --- drivers/gpu/drm/drm_atomic.c | 12 ++++++------ include/drm/drm_atomic.h | 2 +- include/drm/drm_mode_config.h | 2 +- 3 files changed, 8 insertions(+), 8 deletions(-) diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c index 60697482b94c..50f5cf7b69d1 100644 --- a/drivers/gpu/drm/drm_atomic.c +++ b/drivers/gpu/drm/drm_atomic.c @@ -291,15 +291,15 @@ drm_atomic_get_crtc_state(struct drm_atomic_state *state, EXPORT_SYMBOL(drm_atomic_get_crtc_state); static void set_out_fence_for_crtc(struct drm_atomic_state *state, - struct drm_crtc *crtc, s64 __user *fence_ptr) + struct drm_crtc *crtc, s32 __user *fence_ptr) { state->crtcs[drm_crtc_index(crtc)].out_fence_ptr = fence_ptr; } -static s64 __user *get_out_fence_for_crtc(struct drm_atomic_state *state, +static s32 __user *get_out_fence_for_crtc(struct drm_atomic_state *state, struct drm_crtc *crtc) { - s64 __user *fence_ptr; + s32 __user *fence_ptr; fence_ptr = state->crtcs[drm_crtc_index(crtc)].out_fence_ptr; state->crtcs[drm_crtc_index(crtc)].out_fence_ptr = NULL; @@ -512,7 +512,7 @@ int drm_atomic_crtc_set_property(struct drm_crtc *crtc, state->color_mgmt_changed |= replaced; return ret; } else if (property == config->prop_out_fence_ptr) { - s64 __user *fence_ptr = u64_to_user_ptr(val); + s32 __user *fence_ptr = u64_to_user_ptr(val); if (!fence_ptr) return 0; @@ -1915,7 +1915,7 @@ EXPORT_SYMBOL(drm_atomic_clean_old_fb); */ struct drm_out_fence_state { - s64 __user *out_fence_ptr; + s32 __user *out_fence_ptr; struct sync_file *sync_file; int fd; }; @@ -1952,7 +1952,7 @@ static int prepare_crtc_signaling(struct drm_device *dev, return 0; for_each_crtc_in_state(state, crtc, crtc_state, i) { - u64 __user *fence_ptr; + s32 __user *fence_ptr; fence_ptr = get_out_fence_for_crtc(crtc_state->state, crtc); diff --git a/include/drm/drm_atomic.h b/include/drm/drm_atomic.h index d6d241f63b9f..56814e8ae7ea 100644 --- a/include/drm/drm_atomic.h +++ b/include/drm/drm_atomic.h @@ -144,7 +144,7 @@ struct __drm_crtcs_state { struct drm_crtc *ptr; struct drm_crtc_state *state; struct drm_crtc_commit *commit; - s64 __user *out_fence_ptr; + s32 __user *out_fence_ptr; }; struct __drm_connnectors_state { diff --git a/include/drm/drm_mode_config.h b/include/drm/drm_mode_config.h index bf9991b20611..137432386310 100644 --- a/include/drm/drm_mode_config.h +++ b/include/drm/drm_mode_config.h @@ -488,7 +488,7 @@ struct drm_mode_config { /** * @prop_out_fence_ptr: Sync File fd pointer representing the * outgoing fences for a CRTC. Userspace should provide a pointer to a - * value of type s64, and then cast that pointer to u64. + * value of type s32, and then cast that pointer to u64. */ struct drm_property *prop_out_fence_ptr; /** From 58c744da9dcc82a4b55a18e05149ae0e32624d11 Mon Sep 17 00:00:00 2001 From: Tina Zhang Date: Tue, 17 Jan 2017 14:33:16 +0800 Subject: [PATCH 080/343] drm/i915/gvt: Enable the shadow batch buffer Shadow batch buffer is used to shadow the privileged batch buffer which is submitted by vGPU's workload. This patch is used to unmark this functionality. Signed-off-by: Tina Zhang Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/cmd_parser.c | 4 ---- 1 file changed, 4 deletions(-) diff --git a/drivers/gpu/drm/i915/gvt/cmd_parser.c b/drivers/gpu/drm/i915/gvt/cmd_parser.c index d26a092c70e8..e4563984cb1e 100644 --- a/drivers/gpu/drm/i915/gvt/cmd_parser.c +++ b/drivers/gpu/drm/i915/gvt/cmd_parser.c @@ -481,7 +481,6 @@ struct parser_exec_state { (s->vgpu->gvt->device_info.gmadr_bytes_in_cmd >> 2) static unsigned long bypass_scan_mask = 0; -static bool bypass_batch_buffer_scan = true; /* ring ALL, type = 0 */ static struct sub_op_bits sub_op_mi[] = { @@ -1525,9 +1524,6 @@ static int batch_buffer_needs_scan(struct parser_exec_state *s) { struct intel_gvt *gvt = s->vgpu->gvt; - if (bypass_batch_buffer_scan) - return 0; - if (IS_BROADWELL(gvt->dev_priv) || IS_SKYLAKE(gvt->dev_priv)) { /* BDW decides privilege based on address space */ if (cmd_val(s, 0) & (1 << 8)) From 62f0a11e2339e1ba154600d1f49ef5d5d84eaae4 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Fri, 6 Jan 2017 19:58:16 +0000 Subject: [PATCH 081/343] drm/i915/gvt: Fix relocation of shadow bb set_gma_to_bb_cmd() is completely bogus - it is (incorrectly) applying the rules to read a GTT offset from a command as opposed to writing the GTT offset. And to cap it all set_gma_to_bb_cmd() is called within a list iterator of the most strange construction. Fixes: be1da7070aea ("drm/i915/gvt: vGPU command scanner") Signed-off-by: Chris Wilson Cc: Zhenyu Wang Cc: Zhi Wang Cc: Yulei Zhang Cc: # v4.10-rc1+ Tested-by: Tina Zhang Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/execlist.c | 64 ++++++++-------------------- drivers/gpu/drm/i915/gvt/scheduler.h | 2 +- 2 files changed, 19 insertions(+), 47 deletions(-) diff --git a/drivers/gpu/drm/i915/gvt/execlist.c b/drivers/gpu/drm/i915/gvt/execlist.c index f32bb6f6495c..fb852c51d00e 100644 --- a/drivers/gpu/drm/i915/gvt/execlist.c +++ b/drivers/gpu/drm/i915/gvt/execlist.c @@ -364,58 +364,30 @@ static void free_workload(struct intel_vgpu_workload *workload) #define get_desc_from_elsp_dwords(ed, i) \ ((struct execlist_ctx_descriptor_format *)&((ed)->data[i * 2])) - -#define BATCH_BUFFER_ADDR_MASK ((1UL << 32) - (1U << 2)) -#define BATCH_BUFFER_ADDR_HIGH_MASK ((1UL << 16) - (1U)) -static int set_gma_to_bb_cmd(struct intel_shadow_bb_entry *entry_obj, - unsigned long add, int gmadr_bytes) -{ - if (WARN_ON(gmadr_bytes != 4 && gmadr_bytes != 8)) - return -1; - - *((u32 *)(entry_obj->bb_start_cmd_va + (1 << 2))) = add & - BATCH_BUFFER_ADDR_MASK; - if (gmadr_bytes == 8) { - *((u32 *)(entry_obj->bb_start_cmd_va + (2 << 2))) = - add & BATCH_BUFFER_ADDR_HIGH_MASK; - } - - return 0; -} - static void prepare_shadow_batch_buffer(struct intel_vgpu_workload *workload) { - int gmadr_bytes = workload->vgpu->gvt->device_info.gmadr_bytes_in_cmd; + const int gmadr_bytes = workload->vgpu->gvt->device_info.gmadr_bytes_in_cmd; + struct intel_shadow_bb_entry *entry_obj; /* pin the gem object to ggtt */ - if (!list_empty(&workload->shadow_bb)) { - struct intel_shadow_bb_entry *entry_obj = - list_first_entry(&workload->shadow_bb, - struct intel_shadow_bb_entry, - list); - struct intel_shadow_bb_entry *temp; + list_for_each_entry(entry_obj, &workload->shadow_bb, list) { + struct i915_vma *vma; - list_for_each_entry_safe(entry_obj, temp, &workload->shadow_bb, - list) { - struct i915_vma *vma; - - vma = i915_gem_object_ggtt_pin(entry_obj->obj, NULL, 0, - 4, 0); - if (IS_ERR(vma)) { - gvt_err("Cannot pin\n"); - return; - } - - /* FIXME: we are not tracking our pinned VMA leaving it - * up to the core to fix up the stray pin_count upon - * free. - */ - - /* update the relocate gma with shadow batch buffer*/ - set_gma_to_bb_cmd(entry_obj, - i915_ggtt_offset(vma), - gmadr_bytes); + vma = i915_gem_object_ggtt_pin(entry_obj->obj, NULL, 0, 4, 0); + if (IS_ERR(vma)) { + gvt_err("Cannot pin\n"); + return; } + + /* FIXME: we are not tracking our pinned VMA leaving it + * up to the core to fix up the stray pin_count upon + * free. + */ + + /* update the relocate gma with shadow batch buffer*/ + entry_obj->bb_start_cmd_va[1] = i915_ggtt_offset(vma); + if (gmadr_bytes == 8) + entry_obj->bb_start_cmd_va[2] = 0; } } diff --git a/drivers/gpu/drm/i915/gvt/scheduler.h b/drivers/gpu/drm/i915/gvt/scheduler.h index 3b30c28bff51..2833dfa8c9ae 100644 --- a/drivers/gpu/drm/i915/gvt/scheduler.h +++ b/drivers/gpu/drm/i915/gvt/scheduler.h @@ -113,7 +113,7 @@ struct intel_shadow_bb_entry { struct drm_i915_gem_object *obj; void *va; unsigned long len; - void *bb_start_cmd_va; + u32 *bb_start_cmd_va; }; #define workload_q_head(vgpu, ring_id) \ From 7622b25543665567d8830a63210385b7d705924b Mon Sep 17 00:00:00 2001 From: Eric Anholt Date: Mon, 10 Oct 2016 09:44:06 -0700 Subject: [PATCH 082/343] drm/vc4: Fix memory leak of the CRTC state. The underscores variant frees the pointers inside, while the no-underscores variant calls underscores and then frees the struct. Signed-off-by: Eric Anholt Fixes: d8dbf44f13b9 ("drm/vc4: Make the CRTCs cooperate on allocating display lists.") Cc: stable@vger.kernel.org --- drivers/gpu/drm/vc4/vc4_crtc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/vc4/vc4_crtc.c b/drivers/gpu/drm/vc4/vc4_crtc.c index a0fd3e66bc4b..7aadce1f7e7a 100644 --- a/drivers/gpu/drm/vc4/vc4_crtc.c +++ b/drivers/gpu/drm/vc4/vc4_crtc.c @@ -839,7 +839,7 @@ static void vc4_crtc_destroy_state(struct drm_crtc *crtc, } - __drm_atomic_helper_crtc_destroy_state(state); + drm_atomic_helper_crtc_destroy_state(crtc, state); } static const struct drm_crtc_funcs vc4_crtc_funcs = { From 21ccc32496b2f63228f5232b3ac0e426e8fb3c31 Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Fri, 13 Jan 2017 10:49:00 +0300 Subject: [PATCH 083/343] drm/vc4: fix a bounds check We accidentally return success even if vc4_full_res_bounds_check() fails. Fixes: d5b1a78a772f ("drm/vc4: Add support for drawing 3D frames.") Signed-off-by: Dan Carpenter Reviewed-by: Eric Engestrom Reviewed-by: Eric Anholt --- drivers/gpu/drm/vc4/vc4_render_cl.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/vc4/vc4_render_cl.c b/drivers/gpu/drm/vc4/vc4_render_cl.c index 08886a309757..5cdd003605f5 100644 --- a/drivers/gpu/drm/vc4/vc4_render_cl.c +++ b/drivers/gpu/drm/vc4/vc4_render_cl.c @@ -461,7 +461,7 @@ static int vc4_rcl_surface_setup(struct vc4_exec_info *exec, } ret = vc4_full_res_bounds_check(exec, *obj, surf); - if (!ret) + if (ret) return ret; return 0; From 0f2ff82e11c86c05d051cae32b58226392d33bbf Mon Sep 17 00:00:00 2001 From: Eric Anholt Date: Tue, 17 Jan 2017 21:42:53 +1100 Subject: [PATCH 084/343] drm/vc4: Fix an integer overflow in temporary allocation layout. We copy the unvalidated ioctl arguments from the user into kernel temporary memory to run the validation from, to avoid a race where the user updates the unvalidate contents in between validating them and copying them into the validated BO. However, in setting up the layout of the kernel side, we failed to check one of the additions (the roundup() for shader_rec_offset) against integer overflow, allowing a nearly MAX_UINT value of bin_cl_size to cause us to under-allocate the temporary space that we then copy_from_user into. Reported-by: Murray McAllister Signed-off-by: Eric Anholt Fixes: d5b1a78a772f ("drm/vc4: Add support for drawing 3D frames.") --- drivers/gpu/drm/vc4/vc4_gem.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/vc4/vc4_gem.c b/drivers/gpu/drm/vc4/vc4_gem.c index db920771bfb5..c5fe3554858e 100644 --- a/drivers/gpu/drm/vc4/vc4_gem.c +++ b/drivers/gpu/drm/vc4/vc4_gem.c @@ -594,7 +594,8 @@ vc4_get_bcl(struct drm_device *dev, struct vc4_exec_info *exec) args->shader_rec_count); struct vc4_bo *bo; - if (uniforms_offset < shader_rec_offset || + if (shader_rec_offset < args->bin_cl_size || + uniforms_offset < shader_rec_offset || exec_size < uniforms_offset || args->shader_rec_count >= (UINT_MAX / sizeof(struct vc4_shader_state)) || From 6b8ac63847bc2f958dd93c09edc941a0118992d9 Mon Sep 17 00:00:00 2001 From: Eric Anholt Date: Tue, 17 Jan 2017 21:58:06 +1100 Subject: [PATCH 085/343] drm/vc4: Return -EINVAL on the overflow checks failing. By failing to set the errno, we'd continue on to trying to set up the RCL, and then oops on trying to dereference the tile_bo that binning validation should have set up. Reported-by: Ingo Molnar Signed-off-by: Eric Anholt Fixes: d5b1a78a772f ("drm/vc4: Add support for drawing 3D frames.") --- drivers/gpu/drm/vc4/vc4_gem.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/gpu/drm/vc4/vc4_gem.c b/drivers/gpu/drm/vc4/vc4_gem.c index c5fe3554858e..ab3016982466 100644 --- a/drivers/gpu/drm/vc4/vc4_gem.c +++ b/drivers/gpu/drm/vc4/vc4_gem.c @@ -601,6 +601,7 @@ vc4_get_bcl(struct drm_device *dev, struct vc4_exec_info *exec) sizeof(struct vc4_shader_state)) || temp_size < exec_size) { DRM_ERROR("overflow in exec arguments\n"); + ret = -EINVAL; goto fail; } From 501db511397fd6efff3aa5b4e8de415b55559550 Mon Sep 17 00:00:00 2001 From: Rolf Neugebauer Date: Tue, 17 Jan 2017 18:13:51 +0000 Subject: [PATCH 086/343] virtio: don't set VIRTIO_NET_HDR_F_DATA_VALID on xmit This patch part reverts fd2a0437dc33 and e858fae2b0b8 which introduced a subtle change in how the virtio_net flags are derived from the SKBs ip_summed field. With the above commits, the flags are set to VIRTIO_NET_HDR_F_DATA_VALID when ip_summed == CHECKSUM_UNNECESSARY, thus treating it differently to ip_summed == CHECKSUM_NONE, which should be the same. Further, the virtio spec 1.0 / CS04 explicitly says that VIRTIO_NET_HDR_F_DATA_VALID must not be set by the driver. Fixes: fd2a0437dc33 ("virtio_net: introduce virtio_net_hdr_{from,to}_skb") Fixes: e858fae2b0b8 (" virtio_net: use common code for virtio_net_hdr and skb GSO conversion") Signed-off-by: Rolf Neugebauer Acked-by: Michael S. Tsirkin Signed-off-by: David S. Miller --- include/linux/virtio_net.h | 2 -- 1 file changed, 2 deletions(-) diff --git a/include/linux/virtio_net.h b/include/linux/virtio_net.h index 66204007d7ac..56436472ccc7 100644 --- a/include/linux/virtio_net.h +++ b/include/linux/virtio_net.h @@ -91,8 +91,6 @@ static inline int virtio_net_hdr_from_skb(const struct sk_buff *skb, skb_checksum_start_offset(skb)); hdr->csum_offset = __cpu_to_virtio16(little_endian, skb->csum_offset); - } else if (skb->ip_summed == CHECKSUM_UNNECESSARY) { - hdr->flags = VIRTIO_NET_HDR_F_DATA_VALID; } /* else everything is zero */ return 0; From a249708bc2aa1fe3ddf15dfac22bee519d15996b Mon Sep 17 00:00:00 2001 From: Julia Lawall Date: Tue, 17 Jan 2017 12:23:21 +0100 Subject: [PATCH 087/343] stmmac: add missing of_node_put The function stmmac_dt_phy provides several possibilities for initializing plat->mdio_node, all of which have the effect of increasing the reference count of the assigned value. This field is not updated elsewhere, so the value is live until the end of the lifetime of plat (devm_allocated), just after the end of stmmac_remove_config_dt. Thus, add an of_node_put on plat->mdio_node in stmmac_remove_config_dt. It is possible that the field mdio_node is never initialized, but of_node_put is NULL-safe, so it is also safe to call of_node_put in that case. Signed-off-by: Julia Lawall Acked-by: Alexandre TORGUE Signed-off-by: David S. Miller --- drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c index 082cd48db6a7..36942f5a6a53 100644 --- a/drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c +++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c @@ -351,6 +351,7 @@ void stmmac_remove_config_dt(struct platform_device *pdev, if (of_phy_is_fixed_link(np)) of_phy_deregister_fixed_link(np); of_node_put(plat->phy_node); + of_node_put(plat->mdio_node); } #else struct plat_stmmacenet_data * From d5ff72d9af73bc3cbaa3edb541333a851f8c7295 Mon Sep 17 00:00:00 2001 From: Lance Richardson Date: Mon, 16 Jan 2017 18:37:58 -0500 Subject: [PATCH 088/343] vxlan: fix byte order of vxlan-gpe port number vxlan->cfg.dst_port is in network byte order, so an htons() is needed here. Also reduced comment length to stay closer to 80 column width (still slightly over, however). Fixes: e1e5314de08b ("vxlan: implement GPE") Signed-off-by: Lance Richardson Acked-by: Jiri Benc Signed-off-by: David S. Miller --- drivers/net/vxlan.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c index ca7196c40060..8a79cfcac889 100644 --- a/drivers/net/vxlan.c +++ b/drivers/net/vxlan.c @@ -2890,7 +2890,7 @@ static int vxlan_dev_configure(struct net *src_net, struct net_device *dev, memcpy(&vxlan->cfg, conf, sizeof(*conf)); if (!vxlan->cfg.dst_port) { if (conf->flags & VXLAN_F_GPE) - vxlan->cfg.dst_port = 4790; /* IANA assigned VXLAN-GPE port */ + vxlan->cfg.dst_port = htons(4790); /* IANA VXLAN-GPE port */ else vxlan->cfg.dst_port = default_port; } From e05208ded1905e500cd5b369d624b071951c68b9 Mon Sep 17 00:00:00 2001 From: Rex Zhu Date: Tue, 10 Jan 2017 20:00:40 +0800 Subject: [PATCH 089/343] drm/amdgpu: fix bug set incorrect value to vce register MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Set the proper bits for clockgating setup. Signed-off-by: Rex Zhu Acked-by: Christian König Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/vce_v3_0.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c index 5fb0b7f5c065..b621bde8c240 100644 --- a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c @@ -175,7 +175,7 @@ static void vce_v3_0_set_vce_sw_clock_gating(struct amdgpu_device *adev, WREG32(mmVCE_UENC_CLOCK_GATING_2, data); data = RREG32(mmVCE_UENC_REG_CLOCK_GATING); - data &= ~0xffc00000; + data &= ~0x3ff; WREG32(mmVCE_UENC_REG_CLOCK_GATING, data); data = RREG32(mmVCE_UENC_DMA_DCLK_CTRL); From 50a1ebc70a2803deb7811fc73fb55d70e353bc34 Mon Sep 17 00:00:00 2001 From: Rex Zhu Date: Tue, 10 Jan 2017 20:03:59 +0800 Subject: [PATCH 090/343] drm/amdgpu: fix program vce instance logic error. MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit need to clear bit31-29 in GRBM_GFX_INDEX, then the program can be valid. Signed-off-by: Rex Zhu Acked-by: Christian König Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/vce_v3_0.c | 25 ++++++++++++++++--------- 1 file changed, 16 insertions(+), 9 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c index b621bde8c240..37ca685e5a9a 100644 --- a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c @@ -43,9 +43,13 @@ #define GRBM_GFX_INDEX__VCE_INSTANCE__SHIFT 0x04 #define GRBM_GFX_INDEX__VCE_INSTANCE_MASK 0x10 +#define GRBM_GFX_INDEX__VCE_ALL_PIPE 0x07 + #define mmVCE_LMI_VCPU_CACHE_40BIT_BAR0 0x8616 #define mmVCE_LMI_VCPU_CACHE_40BIT_BAR1 0x8617 #define mmVCE_LMI_VCPU_CACHE_40BIT_BAR2 0x8618 +#define mmGRBM_GFX_INDEX_DEFAULT 0xE0000000 + #define VCE_STATUS_VCPU_REPORT_FW_LOADED_MASK 0x02 #define VCE_V3_0_FW_SIZE (384 * 1024) @@ -54,6 +58,9 @@ #define FW_52_8_3 ((52 << 24) | (8 << 16) | (3 << 8)) +#define GET_VCE_INSTANCE(i) ((i) << GRBM_GFX_INDEX__VCE_INSTANCE__SHIFT \ + | GRBM_GFX_INDEX__VCE_ALL_PIPE) + static void vce_v3_0_mc_resume(struct amdgpu_device *adev, int idx); static void vce_v3_0_set_ring_funcs(struct amdgpu_device *adev); static void vce_v3_0_set_irq_funcs(struct amdgpu_device *adev); @@ -249,7 +256,7 @@ static int vce_v3_0_start(struct amdgpu_device *adev) if (adev->vce.harvest_config & (1 << idx)) continue; - WREG32_FIELD(GRBM_GFX_INDEX, VCE_INSTANCE, idx); + WREG32(mmGRBM_GFX_INDEX, GET_VCE_INSTANCE(idx)); vce_v3_0_mc_resume(adev, idx); WREG32_FIELD(VCE_STATUS, JOB_BUSY, 1); @@ -273,7 +280,7 @@ static int vce_v3_0_start(struct amdgpu_device *adev) } } - WREG32_FIELD(GRBM_GFX_INDEX, VCE_INSTANCE, 0); + WREG32(mmGRBM_GFX_INDEX, mmGRBM_GFX_INDEX_DEFAULT); mutex_unlock(&adev->grbm_idx_mutex); return 0; @@ -288,7 +295,7 @@ static int vce_v3_0_stop(struct amdgpu_device *adev) if (adev->vce.harvest_config & (1 << idx)) continue; - WREG32_FIELD(GRBM_GFX_INDEX, VCE_INSTANCE, idx); + WREG32(mmGRBM_GFX_INDEX, GET_VCE_INSTANCE(idx)); if (adev->asic_type >= CHIP_STONEY) WREG32_P(mmVCE_VCPU_CNTL, 0, ~0x200001); @@ -306,7 +313,7 @@ static int vce_v3_0_stop(struct amdgpu_device *adev) vce_v3_0_set_vce_sw_clock_gating(adev, false); } - WREG32_FIELD(GRBM_GFX_INDEX, VCE_INSTANCE, 0); + WREG32(mmGRBM_GFX_INDEX, mmGRBM_GFX_INDEX_DEFAULT); mutex_unlock(&adev->grbm_idx_mutex); return 0; @@ -586,17 +593,17 @@ static bool vce_v3_0_check_soft_reset(void *handle) * VCE team suggest use bit 3--bit 6 for busy status check */ mutex_lock(&adev->grbm_idx_mutex); - WREG32_FIELD(GRBM_GFX_INDEX, INSTANCE_INDEX, 0); + WREG32(mmGRBM_GFX_INDEX, GET_VCE_INSTANCE(0)); if (RREG32(mmVCE_STATUS) & AMDGPU_VCE_STATUS_BUSY_MASK) { srbm_soft_reset = REG_SET_FIELD(srbm_soft_reset, SRBM_SOFT_RESET, SOFT_RESET_VCE0, 1); srbm_soft_reset = REG_SET_FIELD(srbm_soft_reset, SRBM_SOFT_RESET, SOFT_RESET_VCE1, 1); } - WREG32_FIELD(GRBM_GFX_INDEX, INSTANCE_INDEX, 0x10); + WREG32(mmGRBM_GFX_INDEX, GET_VCE_INSTANCE(1)); if (RREG32(mmVCE_STATUS) & AMDGPU_VCE_STATUS_BUSY_MASK) { srbm_soft_reset = REG_SET_FIELD(srbm_soft_reset, SRBM_SOFT_RESET, SOFT_RESET_VCE0, 1); srbm_soft_reset = REG_SET_FIELD(srbm_soft_reset, SRBM_SOFT_RESET, SOFT_RESET_VCE1, 1); } - WREG32_FIELD(GRBM_GFX_INDEX, INSTANCE_INDEX, 0); + WREG32(mmGRBM_GFX_INDEX, GET_VCE_INSTANCE(0)); mutex_unlock(&adev->grbm_idx_mutex); if (srbm_soft_reset) { @@ -734,7 +741,7 @@ static int vce_v3_0_set_clockgating_state(void *handle, if (adev->vce.harvest_config & (1 << i)) continue; - WREG32_FIELD(GRBM_GFX_INDEX, VCE_INSTANCE, i); + WREG32(mmGRBM_GFX_INDEX, GET_VCE_INSTANCE(i)); if (enable) { /* initialize VCE_CLOCK_GATING_A: Clock ON/OFF delay */ @@ -753,7 +760,7 @@ static int vce_v3_0_set_clockgating_state(void *handle, vce_v3_0_set_vce_sw_clock_gating(adev, enable); } - WREG32_FIELD(GRBM_GFX_INDEX, VCE_INSTANCE, 0); + WREG32(mmGRBM_GFX_INDEX, mmGRBM_GFX_INDEX_DEFAULT); mutex_unlock(&adev->grbm_idx_mutex); return 0; From ca581e45335c6aa45e5b27999bc13bdefb7e84d9 Mon Sep 17 00:00:00 2001 From: Rex Zhu Date: Thu, 12 Jan 2017 21:48:26 +0800 Subject: [PATCH 091/343] drm/amdgpu: change clock gating mode for uvd_v4. use sw cg when decode. and hw cg when idle. fixes: https://bugs.freedesktop.org/show_bug.cgi?id=99313 https://bugzilla.kernel.org/show_bug.cgi?id=192161 Signed-off-by: Rex Zhu Ack-by: Tom St Denis Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c | 42 +++++++-------------------- 1 file changed, 10 insertions(+), 32 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c index 96444e4d862a..7fb9137dd89b 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c @@ -40,13 +40,14 @@ #include "smu/smu_7_0_1_sh_mask.h" static void uvd_v4_2_mc_resume(struct amdgpu_device *adev); -static void uvd_v4_2_init_cg(struct amdgpu_device *adev); static void uvd_v4_2_set_ring_funcs(struct amdgpu_device *adev); static void uvd_v4_2_set_irq_funcs(struct amdgpu_device *adev); static int uvd_v4_2_start(struct amdgpu_device *adev); static void uvd_v4_2_stop(struct amdgpu_device *adev); static int uvd_v4_2_set_clockgating_state(void *handle, enum amd_clockgating_state state); +static void uvd_v4_2_set_dcm(struct amdgpu_device *adev, + bool sw_mode); /** * uvd_v4_2_ring_get_rptr - get read pointer * @@ -140,7 +141,8 @@ static int uvd_v4_2_sw_fini(void *handle) return r; } - +static void uvd_v4_2_enable_mgcg(struct amdgpu_device *adev, + bool enable); /** * uvd_v4_2_hw_init - start and test UVD block * @@ -155,8 +157,7 @@ static int uvd_v4_2_hw_init(void *handle) uint32_t tmp; int r; - uvd_v4_2_init_cg(adev); - uvd_v4_2_set_clockgating_state(adev, AMD_CG_STATE_GATE); + uvd_v4_2_enable_mgcg(adev, true); amdgpu_asic_set_uvd_clocks(adev, 10000, 10000); r = uvd_v4_2_start(adev); if (r) @@ -266,11 +267,13 @@ static int uvd_v4_2_start(struct amdgpu_device *adev) struct amdgpu_ring *ring = &adev->uvd.ring; uint32_t rb_bufsz; int i, j, r; - /* disable byte swapping */ u32 lmi_swap_cntl = 0; u32 mp_swap_cntl = 0; + WREG32(mmUVD_CGC_GATE, 0); + uvd_v4_2_set_dcm(adev, true); + uvd_v4_2_mc_resume(adev); /* disable interupt */ @@ -406,6 +409,8 @@ static void uvd_v4_2_stop(struct amdgpu_device *adev) /* Unstall UMC and register bus */ WREG32_P(mmUVD_LMI_CTRL2, 0, ~(1 << 8)); + + uvd_v4_2_set_dcm(adev, false); } /** @@ -619,19 +624,6 @@ static void uvd_v4_2_set_dcm(struct amdgpu_device *adev, WREG32_UVD_CTX(ixUVD_CGC_CTRL2, tmp2); } -static void uvd_v4_2_init_cg(struct amdgpu_device *adev) -{ - bool hw_mode = true; - - if (hw_mode) { - uvd_v4_2_set_dcm(adev, false); - } else { - u32 tmp = RREG32(mmUVD_CGC_CTRL); - tmp &= ~UVD_CGC_CTRL__DYN_CLOCK_MODE_MASK; - WREG32(mmUVD_CGC_CTRL, tmp); - } -} - static bool uvd_v4_2_is_idle(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; @@ -685,17 +677,6 @@ static int uvd_v4_2_process_interrupt(struct amdgpu_device *adev, static int uvd_v4_2_set_clockgating_state(void *handle, enum amd_clockgating_state state) { - bool gate = false; - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - - if (!(adev->cg_flags & AMD_CG_SUPPORT_UVD_MGCG)) - return 0; - - if (state == AMD_CG_STATE_GATE) - gate = true; - - uvd_v4_2_enable_mgcg(adev, gate); - return 0; } @@ -711,9 +692,6 @@ static int uvd_v4_2_set_powergating_state(void *handle, */ struct amdgpu_device *adev = (struct amdgpu_device *)handle; - if (!(adev->pg_flags & AMD_PG_SUPPORT_UVD)) - return 0; - if (state == AMD_PG_STATE_GATE) { uvd_v4_2_stop(adev); return 0; From 4e6e98b1e48c9474aed7ce03025ec319b941e26e Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Tue, 17 Jan 2017 15:02:13 -0500 Subject: [PATCH 092/343] drm/radeon: add support for new hainan variants New hainan parts require updated smc firmware. Cc: Sonny Jiang Signed-off-by: Alex Deucher --- drivers/gpu/drm/radeon/si.c | 13 +++++++++---- 1 file changed, 9 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c index 32d55220a2d3..414776811e71 100644 --- a/drivers/gpu/drm/radeon/si.c +++ b/drivers/gpu/drm/radeon/si.c @@ -114,6 +114,7 @@ MODULE_FIRMWARE("radeon/hainan_mc.bin"); MODULE_FIRMWARE("radeon/hainan_rlc.bin"); MODULE_FIRMWARE("radeon/hainan_smc.bin"); MODULE_FIRMWARE("radeon/hainan_k_smc.bin"); +MODULE_FIRMWARE("radeon/banks_k_2_smc.bin"); MODULE_FIRMWARE("radeon/si58_mc.bin"); @@ -1653,6 +1654,7 @@ static int si_init_microcode(struct radeon_device *rdev) int new_fw = 0; bool new_smc = false; bool si58_fw = false; + bool banks2_fw = false; DRM_DEBUG("\n"); @@ -1730,10 +1732,11 @@ static int si_init_microcode(struct radeon_device *rdev) ((rdev->pdev->device == 0x6660) || (rdev->pdev->device == 0x6663) || (rdev->pdev->device == 0x6665) || - (rdev->pdev->device == 0x6667))) || - ((rdev->pdev->revision == 0xc3) && - (rdev->pdev->device == 0x6665))) + (rdev->pdev->device == 0x6667)))) new_smc = true; + else if ((rdev->pdev->revision == 0xc3) && + (rdev->pdev->device == 0x6665)) + banks2_fw = true; new_chip_name = "hainan"; pfp_req_size = SI_PFP_UCODE_SIZE * 4; me_req_size = SI_PM4_UCODE_SIZE * 4; @@ -1886,7 +1889,9 @@ static int si_init_microcode(struct radeon_device *rdev) } } - if (new_smc) + if (banks2_fw) + snprintf(fw_name, sizeof(fw_name), "radeon/banks_k_2_smc.bin"); + else if (new_smc) snprintf(fw_name, sizeof(fw_name), "radeon/%s_k_smc.bin", new_chip_name); else snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", new_chip_name); From 17324b6add82d6c0bf119f1d1944baef392a4e39 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Tue, 17 Jan 2017 15:06:58 -0500 Subject: [PATCH 093/343] drm/amdgpu: add support for new hainan variants New hainan parts require updated smc firmware. Cc: Sonny Jiang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/si_dpm.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/si_dpm.c b/drivers/gpu/drm/amd/amdgpu/si_dpm.c index e9242d6fbf84..6e150db8f380 100644 --- a/drivers/gpu/drm/amd/amdgpu/si_dpm.c +++ b/drivers/gpu/drm/amd/amdgpu/si_dpm.c @@ -64,6 +64,7 @@ MODULE_FIRMWARE("radeon/oland_smc.bin"); MODULE_FIRMWARE("radeon/oland_k_smc.bin"); MODULE_FIRMWARE("radeon/hainan_smc.bin"); MODULE_FIRMWARE("radeon/hainan_k_smc.bin"); +MODULE_FIRMWARE("radeon/banks_k_2_smc.bin"); union power_info { struct _ATOM_POWERPLAY_INFO info; @@ -7701,10 +7702,11 @@ static int si_dpm_init_microcode(struct amdgpu_device *adev) ((adev->pdev->device == 0x6660) || (adev->pdev->device == 0x6663) || (adev->pdev->device == 0x6665) || - (adev->pdev->device == 0x6667))) || - ((adev->pdev->revision == 0xc3) && - (adev->pdev->device == 0x6665))) + (adev->pdev->device == 0x6667)))) chip_name = "hainan_k"; + else if ((adev->pdev->revision == 0xc3) && + (adev->pdev->device == 0x6665)) + chip_name = "banks_k_2"; else chip_name = "hainan"; break; From 3fbfadce6012e7bb384b2e9ad47869d5177f7209 Mon Sep 17 00:00:00 2001 From: Martin KaFai Lau Date: Mon, 16 Jan 2017 22:17:29 -0800 Subject: [PATCH 094/343] bpf: Fix test_lru_sanity5() in test_lru_map.c test_lru_sanity5() fails when the number of online cpus is fewer than the number of possible cpus. It can be reproduced with qemu by using cmd args "--smp cpus=2,maxcpus=8". The problem is the loop in test_lru_sanity5() is testing 'i' which is incorrect. This patch: 1. Make sched_next_online() always return -1 if it cannot find a next cpu to schedule the process. 2. In test_lru_sanity5(), the parent process does sched_setaffinity() first (through sched_next_online()) and the forked process will inherit it according to the 'man sched_setaffinity'. Fixes: 5db58faf989f ("bpf: Add tests for the LRU bpf_htab") Reported-by: Daniel Borkmann Signed-off-by: Martin KaFai Lau Acked-by: Daniel Borkmann Acked-by: Alexei Starovoitov Signed-off-by: David S. Miller --- tools/testing/selftests/bpf/test_lru_map.c | 53 +++++++++++----------- 1 file changed, 27 insertions(+), 26 deletions(-) diff --git a/tools/testing/selftests/bpf/test_lru_map.c b/tools/testing/selftests/bpf/test_lru_map.c index b13fed534d76..9f7bd1915c21 100644 --- a/tools/testing/selftests/bpf/test_lru_map.c +++ b/tools/testing/selftests/bpf/test_lru_map.c @@ -67,21 +67,23 @@ static int map_equal(int lru_map, int expected) return map_subset(lru_map, expected) && map_subset(expected, lru_map); } -static int sched_next_online(int pid, int next_to_try) +static int sched_next_online(int pid, int *next_to_try) { cpu_set_t cpuset; + int next = *next_to_try; + int ret = -1; - if (next_to_try == nr_cpus) - return -1; - - while (next_to_try < nr_cpus) { + while (next < nr_cpus) { CPU_ZERO(&cpuset); - CPU_SET(next_to_try++, &cpuset); - if (!sched_setaffinity(pid, sizeof(cpuset), &cpuset)) + CPU_SET(next++, &cpuset); + if (!sched_setaffinity(pid, sizeof(cpuset), &cpuset)) { + ret = 0; break; + } } - return next_to_try; + *next_to_try = next; + return ret; } /* Size of the LRU amp is 2 @@ -96,11 +98,12 @@ static void test_lru_sanity0(int map_type, int map_flags) { unsigned long long key, value[nr_cpus]; int lru_map_fd, expected_map_fd; + int next_cpu = 0; printf("%s (map_type:%d map_flags:0x%X): ", __func__, map_type, map_flags); - assert(sched_next_online(0, 0) != -1); + assert(sched_next_online(0, &next_cpu) != -1); if (map_flags & BPF_F_NO_COMMON_LRU) lru_map_fd = create_map(map_type, map_flags, 2 * nr_cpus); @@ -183,6 +186,7 @@ static void test_lru_sanity1(int map_type, int map_flags, unsigned int tgt_free) int lru_map_fd, expected_map_fd; unsigned int batch_size; unsigned int map_size; + int next_cpu = 0; if (map_flags & BPF_F_NO_COMMON_LRU) /* Ther percpu lru list (i.e each cpu has its own LRU @@ -196,7 +200,7 @@ static void test_lru_sanity1(int map_type, int map_flags, unsigned int tgt_free) printf("%s (map_type:%d map_flags:0x%X): ", __func__, map_type, map_flags); - assert(sched_next_online(0, 0) != -1); + assert(sched_next_online(0, &next_cpu) != -1); batch_size = tgt_free / 2; assert(batch_size * 2 == tgt_free); @@ -262,6 +266,7 @@ static void test_lru_sanity2(int map_type, int map_flags, unsigned int tgt_free) int lru_map_fd, expected_map_fd; unsigned int batch_size; unsigned int map_size; + int next_cpu = 0; if (map_flags & BPF_F_NO_COMMON_LRU) /* Ther percpu lru list (i.e each cpu has its own LRU @@ -275,7 +280,7 @@ static void test_lru_sanity2(int map_type, int map_flags, unsigned int tgt_free) printf("%s (map_type:%d map_flags:0x%X): ", __func__, map_type, map_flags); - assert(sched_next_online(0, 0) != -1); + assert(sched_next_online(0, &next_cpu) != -1); batch_size = tgt_free / 2; assert(batch_size * 2 == tgt_free); @@ -370,11 +375,12 @@ static void test_lru_sanity3(int map_type, int map_flags, unsigned int tgt_free) int lru_map_fd, expected_map_fd; unsigned int batch_size; unsigned int map_size; + int next_cpu = 0; printf("%s (map_type:%d map_flags:0x%X): ", __func__, map_type, map_flags); - assert(sched_next_online(0, 0) != -1); + assert(sched_next_online(0, &next_cpu) != -1); batch_size = tgt_free / 2; assert(batch_size * 2 == tgt_free); @@ -430,11 +436,12 @@ static void test_lru_sanity4(int map_type, int map_flags, unsigned int tgt_free) int lru_map_fd, expected_map_fd; unsigned long long key, value[nr_cpus]; unsigned long long end_key; + int next_cpu = 0; printf("%s (map_type:%d map_flags:0x%X): ", __func__, map_type, map_flags); - assert(sched_next_online(0, 0) != -1); + assert(sched_next_online(0, &next_cpu) != -1); if (map_flags & BPF_F_NO_COMMON_LRU) lru_map_fd = create_map(map_type, map_flags, @@ -502,9 +509,8 @@ static void do_test_lru_sanity5(unsigned long long last_key, int map_fd) static void test_lru_sanity5(int map_type, int map_flags) { unsigned long long key, value[nr_cpus]; - int next_sched_cpu = 0; + int next_cpu = 0; int map_fd; - int i; if (map_flags & BPF_F_NO_COMMON_LRU) return; @@ -519,27 +525,20 @@ static void test_lru_sanity5(int map_type, int map_flags) key = 0; assert(!bpf_map_update(map_fd, &key, value, BPF_NOEXIST)); - for (i = 0; i < nr_cpus; i++) { + while (sched_next_online(0, &next_cpu) != -1) { pid_t pid; pid = fork(); if (pid == 0) { - next_sched_cpu = sched_next_online(0, next_sched_cpu); - if (next_sched_cpu != -1) - do_test_lru_sanity5(key, map_fd); + do_test_lru_sanity5(key, map_fd); exit(0); } else if (pid == -1) { - printf("couldn't spawn #%d process\n", i); + printf("couldn't spawn process to test key:%llu\n", + key); exit(1); } else { int status; - /* It is mostly redundant and just allow the parent - * process to update next_shced_cpu for the next child - * process - */ - next_sched_cpu = sched_next_online(pid, next_sched_cpu); - assert(waitpid(pid, &status, 0) == pid); assert(status == 0); key++; @@ -547,6 +546,8 @@ static void test_lru_sanity5(int map_type, int map_flags) } close(map_fd); + /* At least one key should be tested */ + assert(key > 0); printf("Pass\n"); } From 93b43fd137cd8865adf9978ab9870a344365d3af Mon Sep 17 00:00:00 2001 From: "Alvaro G. M" Date: Tue, 17 Jan 2017 09:08:16 +0100 Subject: [PATCH 095/343] net: phy: dp83848: add DP83620 PHY support This PHY with fiber support is register compatible with DP83848, so add support for it. Signed-off-by: Alvaro Gamez Machado Signed-off-by: David S. Miller --- drivers/net/phy/dp83848.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/net/phy/dp83848.c b/drivers/net/phy/dp83848.c index 800b39f06279..a10d0e7fc5f7 100644 --- a/drivers/net/phy/dp83848.c +++ b/drivers/net/phy/dp83848.c @@ -17,6 +17,7 @@ #include #define TI_DP83848C_PHY_ID 0x20005ca0 +#define TI_DP83620_PHY_ID 0x20005ce0 #define NS_DP83848C_PHY_ID 0x20005c90 #define TLK10X_PHY_ID 0x2000a210 #define TI_DP83822_PHY_ID 0x2000a240 @@ -77,6 +78,7 @@ static int dp83848_config_intr(struct phy_device *phydev) static struct mdio_device_id __maybe_unused dp83848_tbl[] = { { TI_DP83848C_PHY_ID, 0xfffffff0 }, { NS_DP83848C_PHY_ID, 0xfffffff0 }, + { TI_DP83620_PHY_ID, 0xfffffff0 }, { TLK10X_PHY_ID, 0xfffffff0 }, { TI_DP83822_PHY_ID, 0xfffffff0 }, { } @@ -106,6 +108,7 @@ MODULE_DEVICE_TABLE(mdio, dp83848_tbl); static struct phy_driver dp83848_driver[] = { DP83848_PHY_DRIVER(TI_DP83848C_PHY_ID, "TI DP83848C 10/100 Mbps PHY"), DP83848_PHY_DRIVER(NS_DP83848C_PHY_ID, "NS DP83848C 10/100 Mbps PHY"), + DP83848_PHY_DRIVER(TI_DP83620_PHY_ID, "TI DP83620 10/100 Mbps PHY"), DP83848_PHY_DRIVER(TLK10X_PHY_ID, "TI TLK10X 10/100 Mbps PHY"), DP83848_PHY_DRIVER(TI_DP83822_PHY_ID, "TI DP83822 10/100 Mbps PHY"), }; From c97c52be78b8463ac5407f1cf1f22f8f6cf93a37 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Einar=20J=C3=B3n?= Date: Fri, 12 Aug 2016 13:50:41 +0200 Subject: [PATCH 096/343] can: c_can_pci: fix null-pointer-deref in c_can_start() - set device pointer MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The priv->device pointer for c_can_pci is never set, but it is used without a NULL check in c_can_start(). Setting it in c_can_pci_probe() like c_can_plat_probe() prevents c_can_pci.ko from crashing, with and without CONFIG_PM. This might also cause the pm_runtime_*() functions in c_can.c to actually be executed for c_can_pci devices - they are the only other place where priv->device is used, but they all contain a null check. Signed-off-by: Einar Jón Cc: linux-stable Signed-off-by: Marc Kleine-Budde --- drivers/net/can/c_can/c_can_pci.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/net/can/c_can/c_can_pci.c b/drivers/net/can/c_can/c_can_pci.c index 7be393c96b1a..cf7c18947189 100644 --- a/drivers/net/can/c_can/c_can_pci.c +++ b/drivers/net/can/c_can/c_can_pci.c @@ -161,6 +161,7 @@ static int c_can_pci_probe(struct pci_dev *pdev, dev->irq = pdev->irq; priv->base = addr; + priv->device = &pdev->dev; if (!c_can_pci_data->freq) { dev_err(&pdev->dev, "no clock frequency defined\n"); From befa60113ce7ea270cb51eada28443ca2756f480 Mon Sep 17 00:00:00 2001 From: Yegor Yefremov Date: Wed, 18 Jan 2017 11:35:57 +0100 Subject: [PATCH 097/343] can: ti_hecc: add missing prepare and unprepare of the clock In order to make the driver work with the common clock framework, this patch converts the clk_enable()/clk_disable() to clk_prepare_enable()/clk_disable_unprepare(). Also add error checking for clk_prepare_enable(). Signed-off-by: Yegor Yefremov Cc: linux-stable Signed-off-by: Marc Kleine-Budde --- drivers/net/can/ti_hecc.c | 16 ++++++++++++---- 1 file changed, 12 insertions(+), 4 deletions(-) diff --git a/drivers/net/can/ti_hecc.c b/drivers/net/can/ti_hecc.c index 680d1ff07a55..6749b1829469 100644 --- a/drivers/net/can/ti_hecc.c +++ b/drivers/net/can/ti_hecc.c @@ -948,7 +948,12 @@ static int ti_hecc_probe(struct platform_device *pdev) netif_napi_add(ndev, &priv->napi, ti_hecc_rx_poll, HECC_DEF_NAPI_WEIGHT); - clk_enable(priv->clk); + err = clk_prepare_enable(priv->clk); + if (err) { + dev_err(&pdev->dev, "clk_prepare_enable() failed\n"); + goto probe_exit_clk; + } + err = register_candev(ndev); if (err) { dev_err(&pdev->dev, "register_candev() failed\n"); @@ -981,7 +986,7 @@ static int ti_hecc_remove(struct platform_device *pdev) struct ti_hecc_priv *priv = netdev_priv(ndev); unregister_candev(ndev); - clk_disable(priv->clk); + clk_disable_unprepare(priv->clk); clk_put(priv->clk); res = platform_get_resource(pdev, IORESOURCE_MEM, 0); iounmap(priv->base); @@ -1006,7 +1011,7 @@ static int ti_hecc_suspend(struct platform_device *pdev, pm_message_t state) hecc_set_bit(priv, HECC_CANMC, HECC_CANMC_PDR); priv->can.state = CAN_STATE_SLEEPING; - clk_disable(priv->clk); + clk_disable_unprepare(priv->clk); return 0; } @@ -1015,8 +1020,11 @@ static int ti_hecc_resume(struct platform_device *pdev) { struct net_device *dev = platform_get_drvdata(pdev); struct ti_hecc_priv *priv = netdev_priv(dev); + int err; - clk_enable(priv->clk); + err = clk_prepare_enable(priv->clk); + if (err) + return err; hecc_clear_bit(priv, HECC_CANMC, HECC_CANMC_PDR); priv->can.state = CAN_STATE_ERROR_ACTIVE; From 4fc020d864647ea3ae8cb8f17d63e48e87ebd0bf Mon Sep 17 00:00:00 2001 From: Francisco Jerez Date: Thu, 12 Jan 2017 12:44:54 +0200 Subject: [PATCH 098/343] drm/i915: Remove WaDisableLSQCROPERFforOCL KBL workaround. The WaDisableLSQCROPERFforOCL workaround has the side effect of disabling an L3SQ optimization that has huge performance implications and is unlikely to be necessary for the correct functioning of usual graphic workloads. Userspace is free to re-enable the workaround on demand, and is generally in a better position to determine whether the workaround is necessary than the DRM is (e.g. only during the execution of compute kernels that rely on both L3 fences and HDC R/W requests). The same workaround seems to apply to BDW (at least to production stepping G1) and SKL as well (the internal workaround database claims that it does for all steppings, while the BSpec workaround table only mentions pre-production steppings), but the DRM doesn't do anything beyond whitelisting the L3SQCREG4 register so userspace can enable it when it sees fit. Do the same on KBL platforms. Improves performance of the GFXBench4 gl_manhattan31 benchmark by 60%, and gl_4 (AKA car chase) by 14% on a KBL GT2 running Mesa master -- This is followed by a regression of 35% and 10% respectively for the same benchmarks and platform caused by my recent patch series switching userspace to use the dataport constant cache instead of the sampler to implement uniform pull constant loads, which caused us to hit more heavily the L3 cache (and on platforms other than KBL had the opposite effect of improving performance of the same two benchmarks). The overall effect on KBL of this change combined with the recent userspace change is respectively 4.6% and 2.6%. SynMark2 OglShMapPcf was affected by the constant cache changes (though it improved as it did on other platforms rather than regressing), but is not significantly affected by this patch (with statistical significance of 5% and sample size 20). v2: Drop some more code to avoid unused variable warning. Fixes: 738fa1b3123f ("drm/i915/kbl: Add WaDisableLSQCROPERFforOCL") Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=99256 Signed-off-by: Francisco Jerez Cc: Matthew Auld Cc: Eero Tamminen Cc: Jani Nikula Cc: Mika Kuoppala Cc: beignet@lists.freedesktop.org Cc: # v4.7+ Reviewed-by: Mika Kuoppala [Removed double Fixes tag] Signed-off-by: Mika Kuoppala Link: http://patchwork.freedesktop.org/patch/msgid/1484217894-20505-1-git-send-email-mika.kuoppala@intel.com (cherry picked from commit 8726f2faa371514fba2f594d799db95203dfeee0) Signed-off-by: Jani Nikula --- drivers/gpu/drm/i915/intel_lrc.c | 10 ---------- drivers/gpu/drm/i915/intel_ringbuffer.c | 8 -------- 2 files changed, 18 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index d4961fa20c73..beabc17e7c8a 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -979,18 +979,8 @@ static inline int gen8_emit_flush_coherentl3_wa(struct intel_engine_cs *engine, uint32_t *batch, uint32_t index) { - struct drm_i915_private *dev_priv = engine->i915; uint32_t l3sqc4_flush = (0x40400000 | GEN8_LQSC_FLUSH_COHERENT_LINES); - /* - * WaDisableLSQCROPERFforOCL:kbl - * This WA is implemented in skl_init_clock_gating() but since - * this batch updates GEN8_L3SQCREG4 with default value we need to - * set this bit here to retain the WA during flush. - */ - if (IS_KBL_REVID(dev_priv, 0, KBL_REVID_E0)) - l3sqc4_flush |= GEN8_LQSC_RO_PERF_DIS; - wa_ctx_emit(batch, index, (MI_STORE_REGISTER_MEM_GEN8 | MI_SRM_LRM_GLOBAL_GTT)); wa_ctx_emit_reg(batch, index, GEN8_L3SQCREG4); diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index aeb637dc1fdf..91cb4c422ad5 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -1095,14 +1095,6 @@ static int kbl_init_workarounds(struct intel_engine_cs *engine) WA_SET_BIT_MASKED(HDC_CHICKEN0, HDC_FENCE_DEST_SLM_DISABLE); - /* GEN8_L3SQCREG4 has a dependency with WA batch so any new changes - * involving this register should also be added to WA batch as required. - */ - if (IS_KBL_REVID(dev_priv, 0, KBL_REVID_E0)) - /* WaDisableLSQCROPERFforOCL:kbl */ - I915_WRITE(GEN8_L3SQCREG4, I915_READ(GEN8_L3SQCREG4) | - GEN8_LQSC_RO_PERF_DIS); - /* WaToEnableHwFixForPushConstHWBug:kbl */ if (IS_KBL_REVID(dev_priv, KBL_REVID_C0, REVID_FOREVER)) WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2, From 3bfdfdcbce2796ce75bf2d85fd8471858d702e5d Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Mon, 7 Nov 2016 22:20:54 +0200 Subject: [PATCH 099/343] drm/i915: Ignore bogus plane coordinates on SKL when the plane is not visible MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When the plane is invisible we may have all sorts of bogus stuff in the coordinates, which we must ignore or else we might fail the plane update. This started to happen on SKL when I moved the plane offset computation to happen in the check phase. Previously we happily ignored it all since we never called the update_plane hook with an invisible plane. Cc: Sivakumar Thulasimani Cc: drm-intel-fixes@lists.freedesktop.org Fixes: b63a16f6cd89 ("drm/i915: Compute display surface offset in the plane check hook for SKL+") Signed-off-by: Ville Syrjälä Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=98258 Testcase: igt/pm_rpm/legacy-planes Testcase: igt/pm_rpm/universal-planes Reviewed-by: Matt Roper Signed-off-by: Matt Roper Link: http://patchwork.freedesktop.org/patch/msgid/1478550057-24864-3-git-send-email-ville.syrjala@linux.intel.com (cherry picked from commit a5e4c7d0aa6784d8abe95c3ceef0da9656d17468) Signed-off-by: Jani Nikula --- drivers/gpu/drm/i915/intel_display.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 3dc8724df400..8d702cf1a616 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -2967,6 +2967,9 @@ int skl_check_plane_surface(struct intel_plane_state *plane_state) unsigned int rotation = plane_state->base.rotation; int ret; + if (!plane_state->base.visible) + return 0; + /* Rotate src coordinates to match rotated GTT view */ if (drm_rotation_90_or_270(rotation)) drm_rect_rotate(&plane_state->base.src, From 31a86d137219373c3222ca5f4f912e9a4d8065bb Mon Sep 17 00:00:00 2001 From: Eran Ben Elisha Date: Tue, 17 Jan 2017 19:19:17 +0200 Subject: [PATCH 100/343] net: ethtool: Initialize buffer when querying device channel settings Ethtool channels respond struct was uninitialized when querying device channel boundaries settings. As a result, unreported fields by the driver hold garbage. This may cause sending unsupported params to driver. Fixes: 8bf368620486 ('ethtool: ensure channel counts are within bounds ...') Signed-off-by: Eran Ben Elisha Signed-off-by: Tariq Toukan CC: John W. Linville Signed-off-by: David S. Miller --- net/core/ethtool.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/core/ethtool.c b/net/core/ethtool.c index e23766c7e3ba..236a21e3c878 100644 --- a/net/core/ethtool.c +++ b/net/core/ethtool.c @@ -1712,7 +1712,7 @@ static noinline_for_stack int ethtool_get_channels(struct net_device *dev, static noinline_for_stack int ethtool_set_channels(struct net_device *dev, void __user *useraddr) { - struct ethtool_channels channels, max; + struct ethtool_channels channels, max = { .cmd = ETHTOOL_GCHANNELS }; u32 max_rx_in_use = 0; if (!dev->ethtool_ops->set_channels || !dev->ethtool_ops->get_channels) From e91ef71dfe834e11b57411f1715cd2e2bb4401f1 Mon Sep 17 00:00:00 2001 From: Eran Ben Elisha Date: Tue, 17 Jan 2017 19:19:18 +0200 Subject: [PATCH 101/343] net/mlx4_en: Remove unnecessary checks when setting num channels Boundaries checks for the number of RX, TX, other and combined channels should be checked by the caller and not in the driver. In addition, remove wrong memset on get channels as it overrides the cmd field in the requester struct. Signed-off-by: Eran Ben Elisha Signed-off-by: Tariq Toukan Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx4/en_ethtool.c | 7 +------ 1 file changed, 1 insertion(+), 6 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx4/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx4/en_ethtool.c index d9c9f86a30df..d5a9372ed84d 100644 --- a/drivers/net/ethernet/mellanox/mlx4/en_ethtool.c +++ b/drivers/net/ethernet/mellanox/mlx4/en_ethtool.c @@ -1732,8 +1732,6 @@ static void mlx4_en_get_channels(struct net_device *dev, { struct mlx4_en_priv *priv = netdev_priv(dev); - memset(channel, 0, sizeof(*channel)); - channel->max_rx = MAX_RX_RINGS; channel->max_tx = MLX4_EN_MAX_TX_RING_P_UP; @@ -1752,10 +1750,7 @@ static int mlx4_en_set_channels(struct net_device *dev, int xdp_count; int err = 0; - if (channel->other_count || channel->combined_count || - channel->tx_count > MLX4_EN_MAX_TX_RING_P_UP || - channel->rx_count > MAX_RX_RINGS || - !channel->tx_count || !channel->rx_count) + if (!channel->tx_count || !channel->rx_count) return -EINVAL; tmp = kzalloc(sizeof(*tmp), GFP_KERNEL); From 639e9e94160e59469305fc2c5e6f9c2733744958 Mon Sep 17 00:00:00 2001 From: Eran Ben Elisha Date: Tue, 17 Jan 2017 19:19:19 +0200 Subject: [PATCH 102/343] net/mlx5e: Remove unnecessary checks when setting num channels Boundaries checks for the number of RX and TX should be checked by the caller and not in the driver. Signed-off-by: Eran Ben Elisha Signed-off-by: Tariq Toukan Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c | 10 ---------- 1 file changed, 10 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c index 33a399a8b5d5..b1b9eb6ee135 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c @@ -554,16 +554,6 @@ static int mlx5e_set_channels(struct net_device *dev, __func__); return -EINVAL; } - if (ch->rx_count || ch->tx_count) { - netdev_info(dev, "%s: separate rx/tx count not supported\n", - __func__); - return -EINVAL; - } - if (count > ncv) { - netdev_info(dev, "%s: count (%d) > max (%d)\n", - __func__, count, ncv); - return -EINVAL; - } if (priv->params.num_channels == count) return 0; From 3fd0b634de7d6b9a85f34a4cf9d8afc1df465cc9 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Fri, 13 Jan 2017 16:41:03 +0100 Subject: [PATCH 103/343] netfilter: ipt_CLUSTERIP: fix build error without procfs We can't access c->pde if CONFIG_PROC_FS is disabled: net/ipv4/netfilter/ipt_CLUSTERIP.c: In function 'clusterip_config_find_get': net/ipv4/netfilter/ipt_CLUSTERIP.c:147:9: error: 'struct clusterip_config' has no member named 'pde' This moves the check inside of another #ifdef. Fixes: 6c5d5cfbe3c5 ("netfilter: ipt_CLUSTERIP: check duplicate config when initializing") Signed-off-by: Arnd Bergmann Signed-off-by: Pablo Neira Ayuso --- net/ipv4/netfilter/ipt_CLUSTERIP.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/net/ipv4/netfilter/ipt_CLUSTERIP.c b/net/ipv4/netfilter/ipt_CLUSTERIP.c index a6b8c1a4102b..0a783cd73faf 100644 --- a/net/ipv4/netfilter/ipt_CLUSTERIP.c +++ b/net/ipv4/netfilter/ipt_CLUSTERIP.c @@ -144,7 +144,12 @@ clusterip_config_find_get(struct net *net, __be32 clusterip, int entry) rcu_read_lock_bh(); c = __clusterip_config_find(net, clusterip); if (c) { - if (!c->pde || unlikely(!atomic_inc_not_zero(&c->refcount))) +#ifdef CONFIG_PROC_FS + if (!c->pde) + c = NULL; + else +#endif + if (unlikely(!atomic_inc_not_zero(&c->refcount))) c = NULL; else if (entry) atomic_inc(&c->entries); From 9a6cdf52b85ea5fb21d2bb31e4a7bc61b79923a7 Mon Sep 17 00:00:00 2001 From: Igor Druzhinin Date: Tue, 17 Jan 2017 20:49:37 +0000 Subject: [PATCH 104/343] xen-netback: fix memory leaks on XenBus disconnect Eliminate memory leaks introduced several years ago by cleaning the queue resources which are allocated on XenBus connection event. Namely, queue structure array and pages used for IO rings. Signed-off-by: Igor Druzhinin Reviewed-by: Paul Durrant Acked-by: Wei Liu Signed-off-by: David S. Miller --- drivers/net/xen-netback/xenbus.c | 11 +++++++++++ 1 file changed, 11 insertions(+) diff --git a/drivers/net/xen-netback/xenbus.c b/drivers/net/xen-netback/xenbus.c index 3124eaec9427..34e4af9272ea 100644 --- a/drivers/net/xen-netback/xenbus.c +++ b/drivers/net/xen-netback/xenbus.c @@ -493,11 +493,20 @@ static int backend_create_xenvif(struct backend_info *be) static void backend_disconnect(struct backend_info *be) { if (be->vif) { + unsigned int queue_index; + xen_unregister_watchers(be->vif); #ifdef CONFIG_DEBUG_FS xenvif_debugfs_delif(be->vif); #endif /* CONFIG_DEBUG_FS */ xenvif_disconnect_data(be->vif); + for (queue_index = 0; queue_index < be->vif->num_queues; ++queue_index) + xenvif_deinit_queue(&be->vif->queues[queue_index]); + + vfree(be->vif->queues); + be->vif->num_queues = 0; + be->vif->queues = NULL; + xenvif_disconnect_ctrl(be->vif); } } @@ -1034,6 +1043,8 @@ static void connect(struct backend_info *be) err: if (be->vif->num_queues > 0) xenvif_disconnect_data(be->vif); /* Clean up existing queues */ + for (queue_index = 0; queue_index < be->vif->num_queues; ++queue_index) + xenvif_deinit_queue(&be->vif->queues[queue_index]); vfree(be->vif->queues); be->vif->queues = NULL; be->vif->num_queues = 0; From f16f1df65f1cf139ff9e9f84661e6573d6bb27fc Mon Sep 17 00:00:00 2001 From: Igor Druzhinin Date: Tue, 17 Jan 2017 20:49:38 +0000 Subject: [PATCH 105/343] xen-netback: protect resource cleaning on XenBus disconnect vif->lock is used to protect statistics gathering agents from using the queue structure during cleaning. Signed-off-by: Igor Druzhinin Acked-by: Wei Liu Reviewed-by: Paul Durrant Signed-off-by: David S. Miller --- drivers/net/xen-netback/interface.c | 6 ++++-- drivers/net/xen-netback/xenbus.c | 2 ++ 2 files changed, 6 insertions(+), 2 deletions(-) diff --git a/drivers/net/xen-netback/interface.c b/drivers/net/xen-netback/interface.c index e30ffd29b7e9..579521327b03 100644 --- a/drivers/net/xen-netback/interface.c +++ b/drivers/net/xen-netback/interface.c @@ -221,18 +221,18 @@ static struct net_device_stats *xenvif_get_stats(struct net_device *dev) { struct xenvif *vif = netdev_priv(dev); struct xenvif_queue *queue = NULL; - unsigned int num_queues = vif->num_queues; unsigned long rx_bytes = 0; unsigned long rx_packets = 0; unsigned long tx_bytes = 0; unsigned long tx_packets = 0; unsigned int index; + spin_lock(&vif->lock); if (vif->queues == NULL) goto out; /* Aggregate tx and rx stats from each queue */ - for (index = 0; index < num_queues; ++index) { + for (index = 0; index < vif->num_queues; ++index) { queue = &vif->queues[index]; rx_bytes += queue->stats.rx_bytes; rx_packets += queue->stats.rx_packets; @@ -241,6 +241,8 @@ static struct net_device_stats *xenvif_get_stats(struct net_device *dev) } out: + spin_unlock(&vif->lock); + vif->dev->stats.rx_bytes = rx_bytes; vif->dev->stats.rx_packets = rx_packets; vif->dev->stats.tx_bytes = tx_bytes; diff --git a/drivers/net/xen-netback/xenbus.c b/drivers/net/xen-netback/xenbus.c index 34e4af9272ea..85b742e1c42f 100644 --- a/drivers/net/xen-netback/xenbus.c +++ b/drivers/net/xen-netback/xenbus.c @@ -503,9 +503,11 @@ static void backend_disconnect(struct backend_info *be) for (queue_index = 0; queue_index < be->vif->num_queues; ++queue_index) xenvif_deinit_queue(&be->vif->queues[queue_index]); + spin_lock(&be->vif->lock); vfree(be->vif->queues); be->vif->num_queues = 0; be->vif->queues = NULL; + spin_unlock(&be->vif->lock); xenvif_disconnect_ctrl(be->vif); } From 7be2c82cfd5d28d7adb66821a992604eb6dd112e Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Wed, 18 Jan 2017 12:12:17 -0800 Subject: [PATCH 106/343] net: fix harmonize_features() vs NETIF_F_HIGHDMA Ashizuka reported a highmem oddity and sent a patch for freescale fec driver. But the problem root cause is that core networking stack must ensure no skb with highmem fragment is ever sent through a device that does not assert NETIF_F_HIGHDMA in its features. We need to call illegal_highdma() from harmonize_features() regardless of CSUM checks. Fixes: ec5f06156423 ("net: Kill link between CSUM and SG features.") Signed-off-by: Eric Dumazet Cc: Pravin Shelar Reported-by: "Ashizuka, Yuusuke" Signed-off-by: David S. Miller --- net/core/dev.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/net/core/dev.c b/net/core/dev.c index 07b307b0b414..7f218e095361 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2795,9 +2795,9 @@ static netdev_features_t harmonize_features(struct sk_buff *skb, if (skb->ip_summed != CHECKSUM_NONE && !can_checksum_protocol(features, type)) { features &= ~(NETIF_F_CSUM_MASK | NETIF_F_GSO_MASK); - } else if (illegal_highdma(skb->dev, skb)) { - features &= ~NETIF_F_SG; } + if (illegal_highdma(skb->dev, skb)) + features &= ~NETIF_F_SG; return features; } From cd33b3e0da43522ff8e8f2b2b71d3d08298512b0 Mon Sep 17 00:00:00 2001 From: Daniel Gonzalez Cabanelas Date: Tue, 17 Jan 2017 16:26:55 -0800 Subject: [PATCH 107/343] net: phy: bcm63xx: Utilize correct config_intr function Commit a1cba5613edf ("net: phy: Add Broadcom phy library for common interfaces") make the BCM63xx PHY driver utilize bcm_phy_config_intr() which would appear to do the right thing, except that it does not write to the MII_BCM63XX_IR register but to MII_BCM54XX_ECR which is different. This would be causing invalid link parameters and events from being generated by the PHY interrupt. Fixes: a1cba5613edf ("net: phy: Add Broadcom phy library for common interfaces") Signed-off-by: Daniel Gonzalez Cabanelas Signed-off-by: Florian Fainelli Signed-off-by: David S. Miller --- drivers/net/phy/bcm63xx.c | 21 +++++++++++++++++++-- 1 file changed, 19 insertions(+), 2 deletions(-) diff --git a/drivers/net/phy/bcm63xx.c b/drivers/net/phy/bcm63xx.c index e741bf614c4e..b0492ef2cdaa 100644 --- a/drivers/net/phy/bcm63xx.c +++ b/drivers/net/phy/bcm63xx.c @@ -21,6 +21,23 @@ MODULE_DESCRIPTION("Broadcom 63xx internal PHY driver"); MODULE_AUTHOR("Maxime Bizon "); MODULE_LICENSE("GPL"); +static int bcm63xx_config_intr(struct phy_device *phydev) +{ + int reg, err; + + reg = phy_read(phydev, MII_BCM63XX_IR); + if (reg < 0) + return reg; + + if (phydev->interrupts == PHY_INTERRUPT_ENABLED) + reg &= ~MII_BCM63XX_IR_GMASK; + else + reg |= MII_BCM63XX_IR_GMASK; + + err = phy_write(phydev, MII_BCM63XX_IR, reg); + return err; +} + static int bcm63xx_config_init(struct phy_device *phydev) { int reg, err; @@ -55,7 +72,7 @@ static struct phy_driver bcm63xx_driver[] = { .config_aneg = genphy_config_aneg, .read_status = genphy_read_status, .ack_interrupt = bcm_phy_ack_intr, - .config_intr = bcm_phy_config_intr, + .config_intr = bcm63xx_config_intr, }, { /* same phy as above, with just a different OUI */ .phy_id = 0x002bdc00, @@ -67,7 +84,7 @@ static struct phy_driver bcm63xx_driver[] = { .config_aneg = genphy_config_aneg, .read_status = genphy_read_status, .ack_interrupt = bcm_phy_ack_intr, - .config_intr = bcm_phy_config_intr, + .config_intr = bcm63xx_config_intr, } }; module_phy_driver(bcm63xx_driver); From 719ca8111402aa6157bd83a3c966d184db0d8956 Mon Sep 17 00:00:00 2001 From: Michael Chan Date: Tue, 17 Jan 2017 22:07:19 -0500 Subject: [PATCH 108/343] bnxt_en: Fix "uninitialized variable" bug in TPA code path. In the TPA GRO code path, initialize the tcp_opt_len variable to 0 so that it will be correct for packets without TCP timestamps. The bug caused the SKB fields to be incorrectly set up for packets without TCP timestamps, leading to these packets being rejected by the stack. Reported-by: Andy Gospodarek Acked-by: Andy Gospodarek Signed-off-by: Michael Chan Signed-off-by: David S. Miller --- drivers/net/ethernet/broadcom/bnxt/bnxt.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c index 9608cb49a11c..53e686fdf2ce 100644 --- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c @@ -1099,7 +1099,7 @@ static struct sk_buff *bnxt_gro_func_5730x(struct bnxt_tpa_info *tpa_info, { #ifdef CONFIG_INET struct tcphdr *th; - int len, nw_off, tcp_opt_len; + int len, nw_off, tcp_opt_len = 0; if (tcp_ts) tcp_opt_len = 12; From 9ed59592e3e379b2e9557dc1d9e9ec8fcbb33f16 Mon Sep 17 00:00:00 2001 From: David Ahern Date: Tue, 17 Jan 2017 14:57:36 -0800 Subject: [PATCH 109/343] lwtunnel: fix autoload of lwt modules Trying to add an mpls encap route when the MPLS modules are not loaded hangs. For example: CONFIG_MPLS=y CONFIG_NET_MPLS_GSO=m CONFIG_MPLS_ROUTING=m CONFIG_MPLS_IPTUNNEL=m $ ip route add 10.10.10.10/32 encap mpls 100 via inet 10.100.1.2 The ip command hangs: root 880 826 0 21:25 pts/0 00:00:00 ip route add 10.10.10.10/32 encap mpls 100 via inet 10.100.1.2 $ cat /proc/880/stack [] call_usermodehelper_exec+0xd6/0x134 [] __request_module+0x27b/0x30a [] lwtunnel_build_state+0xe4/0x178 [] fib_create_info+0x47f/0xdd4 [] fib_table_insert+0x90/0x41f [] inet_rtm_newroute+0x4b/0x52 ... modprobe is trying to load rtnl-lwt-MPLS: root 881 5 0 21:25 ? 00:00:00 /sbin/modprobe -q -- rtnl-lwt-MPLS and it hangs after loading mpls_router: $ cat /proc/881/stack [] rtnl_lock+0x12/0x14 [] register_netdevice_notifier+0x16/0x179 [] mpls_init+0x25/0x1000 [mpls_router] [] do_one_initcall+0x8e/0x13f [] do_init_module+0x5a/0x1e5 [] load_module+0x13bd/0x17d6 ... The problem is that lwtunnel_build_state is called with rtnl lock held preventing mpls_init from registering. Given the potential references held by the time lwtunnel_build_state it can not drop the rtnl lock to the load module. So, extract the module loading code from lwtunnel_build_state into a new function to validate the encap type. The new function is called while converting the user request into a fib_config which is well before any table, device or fib entries are examined. Fixes: 745041e2aaf1 ("lwtunnel: autoload of lwt modules") Signed-off-by: David Ahern Signed-off-by: David S. Miller --- include/net/lwtunnel.h | 11 ++++++ net/core/lwtunnel.c | 74 ++++++++++++++++++++++++++++++++++------- net/ipv4/fib_frontend.c | 8 +++++ net/ipv6/route.c | 12 ++++++- 4 files changed, 92 insertions(+), 13 deletions(-) diff --git a/include/net/lwtunnel.h b/include/net/lwtunnel.h index d4c1c75b8862..0b585f1fd340 100644 --- a/include/net/lwtunnel.h +++ b/include/net/lwtunnel.h @@ -105,6 +105,8 @@ int lwtunnel_encap_add_ops(const struct lwtunnel_encap_ops *op, unsigned int num); int lwtunnel_encap_del_ops(const struct lwtunnel_encap_ops *op, unsigned int num); +int lwtunnel_valid_encap_type(u16 encap_type); +int lwtunnel_valid_encap_type_attr(struct nlattr *attr, int len); int lwtunnel_build_state(struct net_device *dev, u16 encap_type, struct nlattr *encap, unsigned int family, const void *cfg, @@ -168,6 +170,15 @@ static inline int lwtunnel_encap_del_ops(const struct lwtunnel_encap_ops *op, return -EOPNOTSUPP; } +static inline int lwtunnel_valid_encap_type(u16 encap_type) +{ + return -EOPNOTSUPP; +} +static inline int lwtunnel_valid_encap_type_attr(struct nlattr *attr, int len) +{ + return -EOPNOTSUPP; +} + static inline int lwtunnel_build_state(struct net_device *dev, u16 encap_type, struct nlattr *encap, unsigned int family, const void *cfg, diff --git a/net/core/lwtunnel.c b/net/core/lwtunnel.c index a5d4e866ce88..47b1dd65947b 100644 --- a/net/core/lwtunnel.c +++ b/net/core/lwtunnel.c @@ -26,6 +26,7 @@ #include #include #include +#include #ifdef CONFIG_MODULES @@ -114,18 +115,6 @@ int lwtunnel_build_state(struct net_device *dev, u16 encap_type, ret = -EOPNOTSUPP; rcu_read_lock(); ops = rcu_dereference(lwtun_encaps[encap_type]); -#ifdef CONFIG_MODULES - if (!ops) { - const char *encap_type_str = lwtunnel_encap_str(encap_type); - - if (encap_type_str) { - rcu_read_unlock(); - request_module("rtnl-lwt-%s", encap_type_str); - rcu_read_lock(); - ops = rcu_dereference(lwtun_encaps[encap_type]); - } - } -#endif if (likely(ops && ops->build_state)) ret = ops->build_state(dev, encap, family, cfg, lws); rcu_read_unlock(); @@ -134,6 +123,67 @@ int lwtunnel_build_state(struct net_device *dev, u16 encap_type, } EXPORT_SYMBOL(lwtunnel_build_state); +int lwtunnel_valid_encap_type(u16 encap_type) +{ + const struct lwtunnel_encap_ops *ops; + int ret = -EINVAL; + + if (encap_type == LWTUNNEL_ENCAP_NONE || + encap_type > LWTUNNEL_ENCAP_MAX) + return ret; + + rcu_read_lock(); + ops = rcu_dereference(lwtun_encaps[encap_type]); + rcu_read_unlock(); +#ifdef CONFIG_MODULES + if (!ops) { + const char *encap_type_str = lwtunnel_encap_str(encap_type); + + if (encap_type_str) { + __rtnl_unlock(); + request_module("rtnl-lwt-%s", encap_type_str); + rtnl_lock(); + + rcu_read_lock(); + ops = rcu_dereference(lwtun_encaps[encap_type]); + rcu_read_unlock(); + } + } +#endif + return ops ? 0 : -EOPNOTSUPP; +} +EXPORT_SYMBOL(lwtunnel_valid_encap_type); + +int lwtunnel_valid_encap_type_attr(struct nlattr *attr, int remaining) +{ + struct rtnexthop *rtnh = (struct rtnexthop *)attr; + struct nlattr *nla_entype; + struct nlattr *attrs; + struct nlattr *nla; + u16 encap_type; + int attrlen; + + while (rtnh_ok(rtnh, remaining)) { + attrlen = rtnh_attrlen(rtnh); + if (attrlen > 0) { + attrs = rtnh_attrs(rtnh); + nla = nla_find(attrs, attrlen, RTA_ENCAP); + nla_entype = nla_find(attrs, attrlen, RTA_ENCAP_TYPE); + + if (nla_entype) { + encap_type = nla_get_u16(nla_entype); + + if (lwtunnel_valid_encap_type(encap_type) != 0) + return -EOPNOTSUPP; + } + } + rtnh = rtnh_next(rtnh, &remaining); + } + + return 0; +} +EXPORT_SYMBOL(lwtunnel_valid_encap_type_attr); + void lwtstate_free(struct lwtunnel_state *lws) { const struct lwtunnel_encap_ops *ops = lwtun_encaps[lws->type]; diff --git a/net/ipv4/fib_frontend.c b/net/ipv4/fib_frontend.c index eae0332b0e8c..7db2ad2e82d3 100644 --- a/net/ipv4/fib_frontend.c +++ b/net/ipv4/fib_frontend.c @@ -46,6 +46,7 @@ #include #include #include +#include #include #ifndef CONFIG_IP_MULTIPLE_TABLES @@ -677,6 +678,10 @@ static int rtm_to_fib_config(struct net *net, struct sk_buff *skb, cfg->fc_mx_len = nla_len(attr); break; case RTA_MULTIPATH: + err = lwtunnel_valid_encap_type_attr(nla_data(attr), + nla_len(attr)); + if (err < 0) + goto errout; cfg->fc_mp = nla_data(attr); cfg->fc_mp_len = nla_len(attr); break; @@ -691,6 +696,9 @@ static int rtm_to_fib_config(struct net *net, struct sk_buff *skb, break; case RTA_ENCAP_TYPE: cfg->fc_encap_type = nla_get_u16(attr); + err = lwtunnel_valid_encap_type(cfg->fc_encap_type); + if (err < 0) + goto errout; break; } } diff --git a/net/ipv6/route.c b/net/ipv6/route.c index 4f6b067c8753..7ea85370c11c 100644 --- a/net/ipv6/route.c +++ b/net/ipv6/route.c @@ -2896,6 +2896,11 @@ static int rtm_to_fib6_config(struct sk_buff *skb, struct nlmsghdr *nlh, if (tb[RTA_MULTIPATH]) { cfg->fc_mp = nla_data(tb[RTA_MULTIPATH]); cfg->fc_mp_len = nla_len(tb[RTA_MULTIPATH]); + + err = lwtunnel_valid_encap_type_attr(cfg->fc_mp, + cfg->fc_mp_len); + if (err < 0) + goto errout; } if (tb[RTA_PREF]) { @@ -2909,9 +2914,14 @@ static int rtm_to_fib6_config(struct sk_buff *skb, struct nlmsghdr *nlh, if (tb[RTA_ENCAP]) cfg->fc_encap = tb[RTA_ENCAP]; - if (tb[RTA_ENCAP_TYPE]) + if (tb[RTA_ENCAP_TYPE]) { cfg->fc_encap_type = nla_get_u16(tb[RTA_ENCAP_TYPE]); + err = lwtunnel_valid_encap_type(cfg->fc_encap_type); + if (err < 0) + goto errout; + } + if (tb[RTA_EXPIRES]) { unsigned long timeout = addrconf_timeout_fixup(nla_get_u32(tb[RTA_EXPIRES]), HZ); From d407bd25a204bd66b7346dde24bd3d37ef0e0b05 Mon Sep 17 00:00:00 2001 From: Daniel Borkmann Date: Wed, 18 Jan 2017 15:14:17 +0100 Subject: [PATCH 110/343] bpf: don't trigger OOM killer under pressure with map alloc This patch adds two helpers, bpf_map_area_alloc() and bpf_map_area_free(), that are to be used for map allocations. Using kmalloc() for very large allocations can cause excessive work within the page allocator, so i) fall back earlier to vmalloc() when the attempt is considered costly anyway, and even more importantly ii) don't trigger OOM killer with any of the allocators. Since this is based on a user space request, for example, when creating maps with element pre-allocation, we really want such requests to fail instead of killing other user space processes. Also, don't spam the kernel log with warnings should any of the allocations fail under pressure. Given that, we can make backend selection in bpf_map_area_alloc() generic, and convert all maps over to use this API for spots with potentially large allocation requests. Note, replacing the one kmalloc_array() is fine as overflow checks happen earlier in htab_map_alloc(), since it must also protect the multiplication for vmalloc() should kmalloc_array() fail. Signed-off-by: Daniel Borkmann Acked-by: Alexei Starovoitov Signed-off-by: David S. Miller --- include/linux/bpf.h | 2 ++ kernel/bpf/arraymap.c | 18 +++++++----------- kernel/bpf/hashtab.c | 22 +++++++++------------- kernel/bpf/stackmap.c | 20 ++++++++------------ kernel/bpf/syscall.c | 26 ++++++++++++++++++++++++++ 5 files changed, 52 insertions(+), 36 deletions(-) diff --git a/include/linux/bpf.h b/include/linux/bpf.h index 05cf951df3fe..3ed1f3b1d594 100644 --- a/include/linux/bpf.h +++ b/include/linux/bpf.h @@ -247,6 +247,8 @@ struct bpf_map * __must_check bpf_map_inc(struct bpf_map *map, bool uref); void bpf_map_put_with_uref(struct bpf_map *map); void bpf_map_put(struct bpf_map *map); int bpf_map_precharge_memlock(u32 pages); +void *bpf_map_area_alloc(size_t size); +void bpf_map_area_free(void *base); extern int sysctl_unprivileged_bpf_disabled; diff --git a/kernel/bpf/arraymap.c b/kernel/bpf/arraymap.c index 229a5d5df977..3d55d95dcf49 100644 --- a/kernel/bpf/arraymap.c +++ b/kernel/bpf/arraymap.c @@ -11,7 +11,6 @@ */ #include #include -#include #include #include #include @@ -74,14 +73,10 @@ static struct bpf_map *array_map_alloc(union bpf_attr *attr) if (array_size >= U32_MAX - PAGE_SIZE) return ERR_PTR(-ENOMEM); - /* allocate all map elements and zero-initialize them */ - array = kzalloc(array_size, GFP_USER | __GFP_NOWARN); - if (!array) { - array = vzalloc(array_size); - if (!array) - return ERR_PTR(-ENOMEM); - } + array = bpf_map_area_alloc(array_size); + if (!array) + return ERR_PTR(-ENOMEM); /* copy mandatory map attributes */ array->map.map_type = attr->map_type; @@ -97,7 +92,7 @@ static struct bpf_map *array_map_alloc(union bpf_attr *attr) if (array_size >= U32_MAX - PAGE_SIZE || elem_size > PCPU_MIN_UNIT_SIZE || bpf_array_alloc_percpu(array)) { - kvfree(array); + bpf_map_area_free(array); return ERR_PTR(-ENOMEM); } out: @@ -262,7 +257,7 @@ static void array_map_free(struct bpf_map *map) if (array->map.map_type == BPF_MAP_TYPE_PERCPU_ARRAY) bpf_array_free_percpu(array); - kvfree(array); + bpf_map_area_free(array); } static const struct bpf_map_ops array_ops = { @@ -319,7 +314,8 @@ static void fd_array_map_free(struct bpf_map *map) /* make sure it's empty */ for (i = 0; i < array->map.max_entries; i++) BUG_ON(array->ptrs[i] != NULL); - kvfree(array); + + bpf_map_area_free(array); } static void *fd_array_map_lookup_elem(struct bpf_map *map, void *key) diff --git a/kernel/bpf/hashtab.c b/kernel/bpf/hashtab.c index 3f2bb58952d8..a753bbe7df0a 100644 --- a/kernel/bpf/hashtab.c +++ b/kernel/bpf/hashtab.c @@ -13,7 +13,6 @@ #include #include #include -#include #include "percpu_freelist.h" #include "bpf_lru_list.h" @@ -103,7 +102,7 @@ static void htab_free_elems(struct bpf_htab *htab) free_percpu(pptr); } free_elems: - vfree(htab->elems); + bpf_map_area_free(htab->elems); } static struct htab_elem *prealloc_lru_pop(struct bpf_htab *htab, void *key, @@ -125,7 +124,8 @@ static int prealloc_init(struct bpf_htab *htab) { int err = -ENOMEM, i; - htab->elems = vzalloc(htab->elem_size * htab->map.max_entries); + htab->elems = bpf_map_area_alloc(htab->elem_size * + htab->map.max_entries); if (!htab->elems) return -ENOMEM; @@ -320,14 +320,10 @@ static struct bpf_map *htab_map_alloc(union bpf_attr *attr) goto free_htab; err = -ENOMEM; - htab->buckets = kmalloc_array(htab->n_buckets, sizeof(struct bucket), - GFP_USER | __GFP_NOWARN); - - if (!htab->buckets) { - htab->buckets = vmalloc(htab->n_buckets * sizeof(struct bucket)); - if (!htab->buckets) - goto free_htab; - } + htab->buckets = bpf_map_area_alloc(htab->n_buckets * + sizeof(struct bucket)); + if (!htab->buckets) + goto free_htab; for (i = 0; i < htab->n_buckets; i++) { INIT_HLIST_HEAD(&htab->buckets[i].head); @@ -354,7 +350,7 @@ static struct bpf_map *htab_map_alloc(union bpf_attr *attr) free_extra_elems: free_percpu(htab->extra_elems); free_buckets: - kvfree(htab->buckets); + bpf_map_area_free(htab->buckets); free_htab: kfree(htab); return ERR_PTR(err); @@ -1014,7 +1010,7 @@ static void htab_map_free(struct bpf_map *map) prealloc_destroy(htab); free_percpu(htab->extra_elems); - kvfree(htab->buckets); + bpf_map_area_free(htab->buckets); kfree(htab); } diff --git a/kernel/bpf/stackmap.c b/kernel/bpf/stackmap.c index 732ae16d12b7..be8519148c25 100644 --- a/kernel/bpf/stackmap.c +++ b/kernel/bpf/stackmap.c @@ -7,7 +7,6 @@ #include #include #include -#include #include #include #include "percpu_freelist.h" @@ -32,7 +31,7 @@ static int prealloc_elems_and_freelist(struct bpf_stack_map *smap) u32 elem_size = sizeof(struct stack_map_bucket) + smap->map.value_size; int err; - smap->elems = vzalloc(elem_size * smap->map.max_entries); + smap->elems = bpf_map_area_alloc(elem_size * smap->map.max_entries); if (!smap->elems) return -ENOMEM; @@ -45,7 +44,7 @@ static int prealloc_elems_and_freelist(struct bpf_stack_map *smap) return 0; free_elems: - vfree(smap->elems); + bpf_map_area_free(smap->elems); return err; } @@ -76,12 +75,9 @@ static struct bpf_map *stack_map_alloc(union bpf_attr *attr) if (cost >= U32_MAX - PAGE_SIZE) return ERR_PTR(-E2BIG); - smap = kzalloc(cost, GFP_USER | __GFP_NOWARN); - if (!smap) { - smap = vzalloc(cost); - if (!smap) - return ERR_PTR(-ENOMEM); - } + smap = bpf_map_area_alloc(cost); + if (!smap) + return ERR_PTR(-ENOMEM); err = -E2BIG; cost += n_buckets * (value_size + sizeof(struct stack_map_bucket)); @@ -112,7 +108,7 @@ static struct bpf_map *stack_map_alloc(union bpf_attr *attr) put_buffers: put_callchain_buffers(); free_smap: - kvfree(smap); + bpf_map_area_free(smap); return ERR_PTR(err); } @@ -262,9 +258,9 @@ static void stack_map_free(struct bpf_map *map) /* wait for bpf programs to complete before freeing stack map */ synchronize_rcu(); - vfree(smap->elems); + bpf_map_area_free(smap->elems); pcpu_freelist_destroy(&smap->freelist); - kvfree(smap); + bpf_map_area_free(smap); put_callchain_buffers(); } diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c index 1d6b29e4e2c3..19b6129eab23 100644 --- a/kernel/bpf/syscall.c +++ b/kernel/bpf/syscall.c @@ -12,6 +12,8 @@ #include #include #include +#include +#include #include #include #include @@ -49,6 +51,30 @@ void bpf_register_map_type(struct bpf_map_type_list *tl) list_add(&tl->list_node, &bpf_map_types); } +void *bpf_map_area_alloc(size_t size) +{ + /* We definitely need __GFP_NORETRY, so OOM killer doesn't + * trigger under memory pressure as we really just want to + * fail instead. + */ + const gfp_t flags = __GFP_NOWARN | __GFP_NORETRY | __GFP_ZERO; + void *area; + + if (size <= (PAGE_SIZE << PAGE_ALLOC_COSTLY_ORDER)) { + area = kmalloc(size, GFP_USER | flags); + if (area != NULL) + return area; + } + + return __vmalloc(size, GFP_KERNEL | __GFP_HIGHMEM | flags, + PAGE_KERNEL); +} + +void bpf_map_area_free(void *area) +{ + kvfree(area); +} + int bpf_map_precharge_memlock(u32 pages) { struct user_struct *user = get_current_user(); From 739e6f5945d88dcee01590913f6886132a10c215 Mon Sep 17 00:00:00 2001 From: Linus Walleij Date: Wed, 11 Jan 2017 13:37:07 +0100 Subject: [PATCH 111/343] gpio: provide lockdep keys for nested/unnested irqchips The helper function for adding a GPIO chip compiles in a lockdep key for debugging, the same key is needed for nested chips as well. The macro construction is unreadable, replace this with two static inlines instead. The _gpiochip_irqchip_add prefixed function is not helpful, rename it with gpiochip_irqchip_add_key() that tell us what the function is actually doing. Fixes: d245b3f9bd36 ("gpio: simplify adding threaded interrupts") Cc: Roger Quadros Reported-by: Clemens Gruber Reported-by: Roger Quadros Reported-by: Grygorii Strashko Tested-by: Clemens Gruber Tested-by: Grygorii Strashko Signed-off-by: Linus Walleij --- drivers/gpio/gpiolib.c | 18 ++++----- include/linux/gpio/driver.h | 74 ++++++++++++++++++++++++++----------- 2 files changed, 61 insertions(+), 31 deletions(-) diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c index 86bf3b84ada5..a07ae9e37930 100644 --- a/drivers/gpio/gpiolib.c +++ b/drivers/gpio/gpiolib.c @@ -1723,7 +1723,7 @@ static void gpiochip_irqchip_remove(struct gpio_chip *gpiochip) } /** - * _gpiochip_irqchip_add() - adds an irqchip to a gpiochip + * gpiochip_irqchip_add_key() - adds an irqchip to a gpiochip * @gpiochip: the gpiochip to add the irqchip to * @irqchip: the irqchip to add to the gpiochip * @first_irq: if not dynamically assigned, the base (first) IRQ to @@ -1749,13 +1749,13 @@ static void gpiochip_irqchip_remove(struct gpio_chip *gpiochip) * the pins on the gpiochip can generate a unique IRQ. Everything else * need to be open coded. */ -int _gpiochip_irqchip_add(struct gpio_chip *gpiochip, - struct irq_chip *irqchip, - unsigned int first_irq, - irq_flow_handler_t handler, - unsigned int type, - bool nested, - struct lock_class_key *lock_key) +int gpiochip_irqchip_add_key(struct gpio_chip *gpiochip, + struct irq_chip *irqchip, + unsigned int first_irq, + irq_flow_handler_t handler, + unsigned int type, + bool nested, + struct lock_class_key *lock_key) { struct device_node *of_node; bool irq_base_set = false; @@ -1840,7 +1840,7 @@ int _gpiochip_irqchip_add(struct gpio_chip *gpiochip, return 0; } -EXPORT_SYMBOL_GPL(_gpiochip_irqchip_add); +EXPORT_SYMBOL_GPL(gpiochip_irqchip_add_key); #else /* CONFIG_GPIOLIB_IRQCHIP */ diff --git a/include/linux/gpio/driver.h b/include/linux/gpio/driver.h index c2748accea71..e973faba69dc 100644 --- a/include/linux/gpio/driver.h +++ b/include/linux/gpio/driver.h @@ -274,37 +274,67 @@ void gpiochip_set_nested_irqchip(struct gpio_chip *gpiochip, struct irq_chip *irqchip, int parent_irq); -int _gpiochip_irqchip_add(struct gpio_chip *gpiochip, - struct irq_chip *irqchip, - unsigned int first_irq, - irq_flow_handler_t handler, - unsigned int type, - bool nested, - struct lock_class_key *lock_key); +int gpiochip_irqchip_add_key(struct gpio_chip *gpiochip, + struct irq_chip *irqchip, + unsigned int first_irq, + irq_flow_handler_t handler, + unsigned int type, + bool nested, + struct lock_class_key *lock_key); + +#ifdef CONFIG_LOCKDEP + +/* + * Lockdep requires that each irqchip instance be created with a + * unique key so as to avoid unnecessary warnings. This upfront + * boilerplate static inlines provides such a key for each + * unique instance. + */ +static inline int gpiochip_irqchip_add(struct gpio_chip *gpiochip, + struct irq_chip *irqchip, + unsigned int first_irq, + irq_flow_handler_t handler, + unsigned int type) +{ + static struct lock_class_key key; + + return gpiochip_irqchip_add_key(gpiochip, irqchip, first_irq, + handler, type, false, &key); +} -/* FIXME: I assume threaded IRQchips do not have the lockdep problem */ static inline int gpiochip_irqchip_add_nested(struct gpio_chip *gpiochip, struct irq_chip *irqchip, unsigned int first_irq, irq_flow_handler_t handler, unsigned int type) { - return _gpiochip_irqchip_add(gpiochip, irqchip, first_irq, - handler, type, true, NULL); + + static struct lock_class_key key; + + return gpiochip_irqchip_add_key(gpiochip, irqchip, first_irq, + handler, type, true, &key); +} +#else +static inline int gpiochip_irqchip_add(struct gpio_chip *gpiochip, + struct irq_chip *irqchip, + unsigned int first_irq, + irq_flow_handler_t handler, + unsigned int type) +{ + return gpiochip_irqchip_add_key(gpiochip, irqchip, first_irq, + handler, type, false, NULL); } -#ifdef CONFIG_LOCKDEP -#define gpiochip_irqchip_add(...) \ -( \ - ({ \ - static struct lock_class_key _key; \ - _gpiochip_irqchip_add(__VA_ARGS__, false, &_key); \ - }) \ -) -#else -#define gpiochip_irqchip_add(...) \ - _gpiochip_irqchip_add(__VA_ARGS__, false, NULL) -#endif +static inline int gpiochip_irqchip_add_nested(struct gpio_chip *gpiochip, + struct irq_chip *irqchip, + unsigned int first_irq, + irq_flow_handler_t handler, + unsigned int type) +{ + return gpiochip_irqchip_add_key(gpiochip, irqchip, first_irq, + handler, type, true, NULL); +} +#endif /* CONFIG_LOCKDEP */ #endif /* CONFIG_GPIOLIB_IRQCHIP */ From b27e36482c02a94194fec71fb29696f4c8e9241c Mon Sep 17 00:00:00 2001 From: Martin Blumenstingl Date: Sun, 15 Jan 2017 23:20:28 +0100 Subject: [PATCH 112/343] pinctrl: meson: fix uart_ao_b for GXBB and GXL/GXM The GXBB and GXL/GXM pinctrl drivers had a configuration which conflicts with uart_ao_a. According to the GXBB ("S905") datasheet the AO UART functions are: - GPIOAO_0: Func1 = UART_TX_AO_A (bit 12), Func2 = UART_TX_AO_B (bit 26) - GPIOAO_1: Func1 = UART_RX_AO_A (bit 11), Func2 = UART_RX_AO_B (bit 25) - GPIOAO_4: Func2 = UART_TX_AO_B (bit 24) - GPIOAO_5: Func2 = UART_RX_AO_B (bit 25) The existing definition for uart_AO_A already uses GPIOAO_0 and GPIOAO_1. The old definition of uart_AO_B however was broken, as it used GPIOAO_0 for TX (which would be fine) and two pins (GPIOAO_1 and GPIOAO_5) for RX (which does not make any sense). This fixes the uart_AO_B configuration by moving it to GPIOAO_4 and GPIOAO_5 (it would be possible to use GPIOAO_0 and GPIOAO_1 in theory, but all existing hardware uses uart_AO_A there). The fix for GXBB and GXL/GXM is identical since it seems that these specific pins are identical on both SoC variants. Signed-off-by: Martin Blumenstingl Reviewed-by: Kevin Hilman Signed-off-by: Linus Walleij --- drivers/pinctrl/meson/pinctrl-meson-gxbb.c | 7 +++---- drivers/pinctrl/meson/pinctrl-meson-gxl.c | 7 +++---- 2 files changed, 6 insertions(+), 8 deletions(-) diff --git a/drivers/pinctrl/meson/pinctrl-meson-gxbb.c b/drivers/pinctrl/meson/pinctrl-meson-gxbb.c index c3928aa3fefa..e0bca4df2a2f 100644 --- a/drivers/pinctrl/meson/pinctrl-meson-gxbb.c +++ b/drivers/pinctrl/meson/pinctrl-meson-gxbb.c @@ -253,9 +253,8 @@ static const unsigned int uart_tx_ao_a_pins[] = { PIN(GPIOAO_0, 0) }; static const unsigned int uart_rx_ao_a_pins[] = { PIN(GPIOAO_1, 0) }; static const unsigned int uart_cts_ao_a_pins[] = { PIN(GPIOAO_2, 0) }; static const unsigned int uart_rts_ao_a_pins[] = { PIN(GPIOAO_3, 0) }; -static const unsigned int uart_tx_ao_b_pins[] = { PIN(GPIOAO_0, 0) }; -static const unsigned int uart_rx_ao_b_pins[] = { PIN(GPIOAO_1, 0), - PIN(GPIOAO_5, 0) }; +static const unsigned int uart_tx_ao_b_pins[] = { PIN(GPIOAO_4, 0) }; +static const unsigned int uart_rx_ao_b_pins[] = { PIN(GPIOAO_5, 0) }; static const unsigned int uart_cts_ao_b_pins[] = { PIN(GPIOAO_2, 0) }; static const unsigned int uart_rts_ao_b_pins[] = { PIN(GPIOAO_3, 0) }; @@ -498,7 +497,7 @@ static struct meson_pmx_group meson_gxbb_aobus_groups[] = { GPIO_GROUP(GPIOAO_13, 0), /* bank AO */ - GROUP(uart_tx_ao_b, 0, 26), + GROUP(uart_tx_ao_b, 0, 24), GROUP(uart_rx_ao_b, 0, 25), GROUP(uart_tx_ao_a, 0, 12), GROUP(uart_rx_ao_a, 0, 11), diff --git a/drivers/pinctrl/meson/pinctrl-meson-gxl.c b/drivers/pinctrl/meson/pinctrl-meson-gxl.c index 25694f7094c7..b69743b07a1d 100644 --- a/drivers/pinctrl/meson/pinctrl-meson-gxl.c +++ b/drivers/pinctrl/meson/pinctrl-meson-gxl.c @@ -214,9 +214,8 @@ static const unsigned int uart_tx_ao_a_pins[] = { PIN(GPIOAO_0, 0) }; static const unsigned int uart_rx_ao_a_pins[] = { PIN(GPIOAO_1, 0) }; static const unsigned int uart_cts_ao_a_pins[] = { PIN(GPIOAO_2, 0) }; static const unsigned int uart_rts_ao_a_pins[] = { PIN(GPIOAO_3, 0) }; -static const unsigned int uart_tx_ao_b_pins[] = { PIN(GPIOAO_0, 0) }; -static const unsigned int uart_rx_ao_b_pins[] = { PIN(GPIOAO_1, 0), - PIN(GPIOAO_5, 0) }; +static const unsigned int uart_tx_ao_b_pins[] = { PIN(GPIOAO_4, 0) }; +static const unsigned int uart_rx_ao_b_pins[] = { PIN(GPIOAO_5, 0) }; static const unsigned int uart_cts_ao_b_pins[] = { PIN(GPIOAO_2, 0) }; static const unsigned int uart_rts_ao_b_pins[] = { PIN(GPIOAO_3, 0) }; @@ -409,7 +408,7 @@ static struct meson_pmx_group meson_gxl_aobus_groups[] = { GPIO_GROUP(GPIOAO_9, 0), /* bank AO */ - GROUP(uart_tx_ao_b, 0, 26), + GROUP(uart_tx_ao_b, 0, 24), GROUP(uart_rx_ao_b, 0, 25), GROUP(uart_tx_ao_a, 0, 12), GROUP(uart_rx_ao_a, 0, 11), From df1539c25cce98e2ac69881958850c6535240707 Mon Sep 17 00:00:00 2001 From: Masahiro Yamada Date: Tue, 17 Jan 2017 19:52:54 +0900 Subject: [PATCH 113/343] pinctrl: uniphier: fix Ethernet (RMII) pin-mux setting for LD20 Fix the pin-mux values for the MDC, MDIO, MDIO_INTL, PHYRSTL pins. Fixes: 1e359ab1285e ("pinctrl: uniphier: add Ethernet pin-mux settings") Signed-off-by: Masahiro Yamada Signed-off-by: Linus Walleij --- drivers/pinctrl/uniphier/pinctrl-uniphier-ld20.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/pinctrl/uniphier/pinctrl-uniphier-ld20.c b/drivers/pinctrl/uniphier/pinctrl-uniphier-ld20.c index aa8bd9794683..96686336e3a3 100644 --- a/drivers/pinctrl/uniphier/pinctrl-uniphier-ld20.c +++ b/drivers/pinctrl/uniphier/pinctrl-uniphier-ld20.c @@ -561,7 +561,7 @@ static const int ether_rgmii_muxvals[] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}; static const unsigned ether_rmii_pins[] = {30, 31, 32, 33, 34, 35, 36, 37, 39, 41, 42, 45}; -static const int ether_rmii_muxvals[] = {1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1}; +static const int ether_rmii_muxvals[] = {0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1}; static const unsigned i2c0_pins[] = {63, 64}; static const int i2c0_muxvals[] = {0, 0}; static const unsigned i2c1_pins[] = {65, 66}; From 524b698db06b9b6da7192e749f637904e2f62d7b Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Mon, 16 Jan 2017 18:24:56 +0100 Subject: [PATCH 114/343] netfilter: conntrack: remove GC_MAX_EVICTS break Instead of breaking loop and instant resched, don't bother checking this in first place (the loop calls cond_resched for every bucket anyway). Suggested-by: Nicolas Dichtel Signed-off-by: Florian Westphal Acked-by: Nicolas Dichtel Signed-off-by: Pablo Neira Ayuso --- net/netfilter/nf_conntrack_core.c | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-) diff --git a/net/netfilter/nf_conntrack_core.c b/net/netfilter/nf_conntrack_core.c index 3a073cd9fcf4..6feb5d370319 100644 --- a/net/netfilter/nf_conntrack_core.c +++ b/net/netfilter/nf_conntrack_core.c @@ -88,8 +88,6 @@ static __read_mostly bool nf_conntrack_locks_all; #define GC_MAX_BUCKETS_DIV 64u /* upper bound of scan intervals */ #define GC_INTERVAL_MAX (2 * HZ) -/* maximum conntracks to evict per gc run */ -#define GC_MAX_EVICTS 256u static struct conntrack_gc_work conntrack_gc_work; @@ -979,8 +977,7 @@ static void gc_worker(struct work_struct *work) */ rcu_read_unlock(); cond_resched_rcu_qs(); - } while (++buckets < goal && - expired_count < GC_MAX_EVICTS); + } while (++buckets < goal); if (gc_work->exiting) return; @@ -1005,7 +1002,7 @@ static void gc_worker(struct work_struct *work) * In case we have lots of evictions next scan is done immediately. */ ratio = scanned ? expired_count * 100 / scanned : 0; - if (ratio >= 90 || expired_count == GC_MAX_EVICTS) { + if (ratio >= 90) { gc_work->next_gc_run = 0; next_run = 0; } else if (expired_count) { From e5072053b09642b8ff417d47da05b84720aea3ee Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Wed, 18 Jan 2017 02:01:22 +0100 Subject: [PATCH 115/343] netfilter: conntrack: refine gc worker heuristics, redux This further refines the changes made to conntrack gc_worker in commit e0df8cae6c16 ("netfilter: conntrack: refine gc worker heuristics"). The main idea of that change was to reduce the scan interval when evictions take place. However, on the reporters' setup, there are 1-2 million conntrack entries in total and roughly 8k new (and closing) connections per second. In this case we'll always evict at least one entry per gc cycle and scan interval is always at 1 jiffy because of this test: } else if (expired_count) { gc_work->next_gc_run /= 2U; next_run = msecs_to_jiffies(1); being true almost all the time. Given we scan ~10k entries per run its clearly wrong to reduce interval based on nonzero eviction count, it will only waste cpu cycles since a vast majorities of conntracks are not timed out. Thus only look at the ratio (scanned entries vs. evicted entries) to make a decision on whether to reduce or not. Because evictor is supposed to only kick in when system turns idle after a busy period, pick a high ratio -- this makes it 50%. We thus keep the idea of increasing scan rate when its likely that table contains many expired entries. In order to not let timed-out entries hang around for too long (important when using event logging, in which case we want to timely destroy events), we now scan the full table within at most GC_MAX_SCAN_JIFFIES (16 seconds) even in worst-case scenario where all timed-out entries sit in same slot. I tested this with a vm under synflood (with sysctl net.netfilter.nf_conntrack_tcp_timeout_syn_recv=3). While flood is ongoing, interval now stays at its max rate (GC_MAX_SCAN_JIFFIES / GC_MAX_BUCKETS_DIV -> 125ms). With feedback from Nicolas Dichtel. Reported-by: Denys Fedoryshchenko Cc: Nicolas Dichtel Fixes: b87a2f9199ea82eaadc ("netfilter: conntrack: add gc worker to remove timed-out entries") Signed-off-by: Florian Westphal Tested-by: Nicolas Dichtel Acked-by: Nicolas Dichtel Tested-by: Denys Fedoryshchenko Signed-off-by: Pablo Neira Ayuso --- net/netfilter/nf_conntrack_core.c | 39 ++++++++++++++++--------------- 1 file changed, 20 insertions(+), 19 deletions(-) diff --git a/net/netfilter/nf_conntrack_core.c b/net/netfilter/nf_conntrack_core.c index 6feb5d370319..4e8083c5e01d 100644 --- a/net/netfilter/nf_conntrack_core.c +++ b/net/netfilter/nf_conntrack_core.c @@ -85,9 +85,11 @@ static __read_mostly DEFINE_SPINLOCK(nf_conntrack_locks_all_lock); static __read_mostly bool nf_conntrack_locks_all; /* every gc cycle scans at most 1/GC_MAX_BUCKETS_DIV part of table */ -#define GC_MAX_BUCKETS_DIV 64u -/* upper bound of scan intervals */ -#define GC_INTERVAL_MAX (2 * HZ) +#define GC_MAX_BUCKETS_DIV 128u +/* upper bound of full table scan */ +#define GC_MAX_SCAN_JIFFIES (16u * HZ) +/* desired ratio of entries found to be expired */ +#define GC_EVICT_RATIO 50u static struct conntrack_gc_work conntrack_gc_work; @@ -936,6 +938,7 @@ static noinline int early_drop(struct net *net, unsigned int _hash) static void gc_worker(struct work_struct *work) { + unsigned int min_interval = max(HZ / GC_MAX_BUCKETS_DIV, 1u); unsigned int i, goal, buckets = 0, expired_count = 0; struct conntrack_gc_work *gc_work; unsigned int ratio, scanned = 0; @@ -994,27 +997,25 @@ static void gc_worker(struct work_struct *work) * 1. Minimize time until we notice a stale entry * 2. Maximize scan intervals to not waste cycles * - * Normally, expired_count will be 0, this increases the next_run time - * to priorize 2) above. + * Normally, expire ratio will be close to 0. * - * As soon as a timed-out entry is found, move towards 1) and increase - * the scan frequency. - * In case we have lots of evictions next scan is done immediately. + * As soon as a sizeable fraction of the entries have expired + * increase scan frequency. */ ratio = scanned ? expired_count * 100 / scanned : 0; - if (ratio >= 90) { - gc_work->next_gc_run = 0; - next_run = 0; - } else if (expired_count) { - gc_work->next_gc_run /= 2U; - next_run = msecs_to_jiffies(1); + if (ratio > GC_EVICT_RATIO) { + gc_work->next_gc_run = min_interval; } else { - if (gc_work->next_gc_run < GC_INTERVAL_MAX) - gc_work->next_gc_run += msecs_to_jiffies(1); + unsigned int max = GC_MAX_SCAN_JIFFIES / GC_MAX_BUCKETS_DIV; - next_run = gc_work->next_gc_run; + BUILD_BUG_ON((GC_MAX_SCAN_JIFFIES / GC_MAX_BUCKETS_DIV) == 0); + + gc_work->next_gc_run += min_interval; + if (gc_work->next_gc_run > max) + gc_work->next_gc_run = max; } + next_run = gc_work->next_gc_run; gc_work->last_bucket = i; queue_delayed_work(system_long_wq, &gc_work->dwork, next_run); } @@ -1022,7 +1023,7 @@ static void gc_worker(struct work_struct *work) static void conntrack_gc_work_init(struct conntrack_gc_work *gc_work) { INIT_DELAYED_WORK(&gc_work->dwork, gc_worker); - gc_work->next_gc_run = GC_INTERVAL_MAX; + gc_work->next_gc_run = HZ; gc_work->exiting = false; } @@ -1914,7 +1915,7 @@ int nf_conntrack_init_start(void) nf_ct_untracked_status_or(IPS_CONFIRMED | IPS_UNTRACKED); conntrack_gc_work_init(&conntrack_gc_work); - queue_delayed_work(system_long_wq, &conntrack_gc_work.dwork, GC_INTERVAL_MAX); + queue_delayed_work(system_long_wq, &conntrack_gc_work.dwork, HZ); return 0; From 59cfa789d04c35b6c647aacf4cc89b3d4d430cfe Mon Sep 17 00:00:00 2001 From: Hariprasad Shenai Date: Thu, 19 Jan 2017 11:50:32 +0530 Subject: [PATCH 116/343] MAINTAINERS: update cxgb4 maintainer Ganesg will be taking over as maintainer from now Signed-off-by: Hariprasad Shenai Signed-off-by: David S. Miller --- MAINTAINERS | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/MAINTAINERS b/MAINTAINERS index c36976d3bd1a..af3456d33a18 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -3564,7 +3564,7 @@ F: drivers/infiniband/hw/cxgb3/ F: include/uapi/rdma/cxgb3-abi.h CXGB4 ETHERNET DRIVER (CXGB4) -M: Hariprasad S +M: Ganesh Goudar L: netdev@vger.kernel.org W: http://www.chelsio.com S: Supported From 03e4deff4987f79c34112c5ba4eb195d4f9382b0 Mon Sep 17 00:00:00 2001 From: Kefeng Wang Date: Thu, 19 Jan 2017 16:26:21 +0800 Subject: [PATCH 117/343] ipv6: addrconf: Avoid addrconf_disable_change() using RCU read-side lock Just like commit 4acd4945cd1e ("ipv6: addrconf: Avoid calling netdevice notifiers with RCU read-side lock"), it is unnecessary to make addrconf_disable_change() use RCU iteration over the netdev list, since it already holds the RTNL lock, or we may meet Illegal context switch in RCU read-side critical section. Signed-off-by: Kefeng Wang Signed-off-by: David S. Miller --- net/ipv6/addrconf.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c index c1e124bc8e1e..f60e88e56255 100644 --- a/net/ipv6/addrconf.c +++ b/net/ipv6/addrconf.c @@ -5540,8 +5540,7 @@ static void addrconf_disable_change(struct net *net, __s32 newf) struct net_device *dev; struct inet6_dev *idev; - rcu_read_lock(); - for_each_netdev_rcu(net, dev) { + for_each_netdev(net, dev) { idev = __in6_dev_get(dev); if (idev) { int changed = (!idev->cnf.disable_ipv6) ^ (!newf); @@ -5550,7 +5549,6 @@ static void addrconf_disable_change(struct net *net, __s32 newf) dev_disable_change(idev); } } - rcu_read_unlock(); } static int addrconf_disable_ipv6(struct ctl_table *table, int *p, int newf) From ad05df399f3343b10664827a3860669a8a80782d Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Thu, 19 Jan 2017 10:33:29 +0100 Subject: [PATCH 118/343] net/mlx5e: Remove unused variable A cleanup removed the only user of this variable mlx5/core/en_ethtool.c: In function 'mlx5e_set_channels': mlx5/core/en_ethtool.c:546:6: error: unused variable 'ncv' [-Werror=unused-variable] Let's remove the declaration as well. Fixes: 639e9e94160e ("net/mlx5e: Remove unnecessary checks when setting num channels") Signed-off-by: Arnd Bergmann Reported-by: Stephen Rothwell Reviewed-by: Tariq Toukan Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c | 1 - 1 file changed, 1 deletion(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c index b1b9eb6ee135..5197817e4b2f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c @@ -543,7 +543,6 @@ static int mlx5e_set_channels(struct net_device *dev, struct ethtool_channels *ch) { struct mlx5e_priv *priv = netdev_priv(dev); - int ncv = mlx5e_get_max_num_channels(priv->mdev); unsigned int count = ch->combined_count; bool arfs_enabled; bool was_opened; From 0dbd7ff3ac5017a46033a9d0a87a8267d69119d9 Mon Sep 17 00:00:00 2001 From: Alexey Kodanev Date: Thu, 19 Jan 2017 16:36:39 +0300 Subject: [PATCH 119/343] tcp: initialize max window for a new fastopen socket Found that if we run LTP netstress test with large MSS (65K), the first attempt from server to send data comparable to this MSS on fastopen connection will be delayed by the probe timer. Here is an example: < S seq 0:0 win 43690 options [mss 65495 wscale 7 tfo cookie] length 32 > S. seq 0:0 ack 1 win 43690 options [mss 65495 wscale 7] length 0 < . ack 1 win 342 length 0 Inside tcp_sendmsg(), tcp_send_mss() returns max MSS in 'mss_now', as well as in 'size_goal'. This results the segment not queued for transmition until all the data copied from user buffer. Then, inside __tcp_push_pending_frames(), it breaks on send window test and continues with the check probe timer. Fragmentation occurs in tcp_write_wakeup()... +0.2 > P. seq 1:43777 ack 1 win 342 length 43776 < . ack 43777, win 1365 length 0 > P. seq 43777:65001 ack 1 win 342 options [...] length 21224 ... This also contradicts with the fact that we should bound to the half of the window if it is large. Fix this flaw by correctly initializing max_window. Before that, it could have large values that affect further calculations of 'size_goal'. Fixes: 168a8f58059a ("tcp: TCP Fast Open Server - main code path") Signed-off-by: Alexey Kodanev Acked-by: Eric Dumazet Signed-off-by: David S. Miller --- net/ipv4/tcp_fastopen.c | 1 + 1 file changed, 1 insertion(+) diff --git a/net/ipv4/tcp_fastopen.c b/net/ipv4/tcp_fastopen.c index f51919535ca7..dd2560c83a85 100644 --- a/net/ipv4/tcp_fastopen.c +++ b/net/ipv4/tcp_fastopen.c @@ -205,6 +205,7 @@ static struct sock *tcp_fastopen_create_child(struct sock *sk, * scaled. So correct it appropriately. */ tp->snd_wnd = ntohs(tcp_hdr(skb)->window); + tp->max_window = tp->snd_wnd; /* Activate the retrans timer so that SYNACK can be retransmitted. * The request socket is not added to the ehash From 69fed99baac186013840ced3524562841296034f Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Wed, 18 Jan 2017 19:44:42 -0800 Subject: [PATCH 120/343] gianfar: Do not reuse pages from emergency reserve A driver using dev_alloc_page() must not reuse a page that had to use emergency memory reserve. Otherwise all packets using this page will be immediately dropped, unless for very specific sockets having SOCK_MEMALLOC bit set. This issue might be hard to debug, because only a fraction of the RX ring buffer would suffer from drops. Fixes: 75354148ce69 ("gianfar: Add paged allocation and Rx S/G") Signed-off-by: Eric Dumazet Cc: Claudiu Manoil Acked-by: Claudiu Manoil Signed-off-by: David S. Miller --- drivers/net/ethernet/freescale/gianfar.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/freescale/gianfar.c b/drivers/net/ethernet/freescale/gianfar.c index a6e7afa878be..c1b671667920 100644 --- a/drivers/net/ethernet/freescale/gianfar.c +++ b/drivers/net/ethernet/freescale/gianfar.c @@ -2948,7 +2948,7 @@ static bool gfar_add_rx_frag(struct gfar_rx_buff *rxb, u32 lstatus, } /* try reuse page */ - if (unlikely(page_count(page) != 1)) + if (unlikely(page_count(page) != 1 || page_is_pfmemalloc(page))) return false; /* change offset to the other half */ From 47b5d64691350e116093c9b47b55ca6b9433bc50 Mon Sep 17 00:00:00 2001 From: Wang Xiaoguang Date: Wed, 7 Sep 2016 20:17:38 +0800 Subject: [PATCH 121/343] btrfs: fix false enospc error when truncating heavily reflinked file Below test script can reveal this bug: dd if=/dev/zero of=fs.img bs=$((1024*1024)) count=100 dev=$(losetup --show -f fs.img) mkdir -p /mnt/mntpoint mkfs.btrfs -f $dev mount $dev /mnt/mntpoint cd /mnt/mntpoint echo "workdir is: /mnt/mntpoint" blocksize=$((128 * 1024)) dd if=/dev/zero of=testfile bs=$blocksize count=1 sync count=$((17*1024*1024*1024/blocksize)) echo "file size is:" $((count*blocksize)) for ((i = 1; i <= $count; i++)); do dst_offset=$((blocksize * i)) xfs_io -f -c "reflink testfile 0 $dst_offset $blocksize"\ testfile > /dev/null done sync truncate --size 0 testfile The last truncate operation will fail for ENOSPC reason, but indeed it should not fail. In btrfs_truncate(), we use a temporary block_rsv to do truncate operation. With every btrfs_truncate_inode_items() call, we migrate space to this block_rsv, but forget to cleanup previous reservation, which will make this block_rsv's reserved bytes keep growing, and this reserved space will only be released in the end of btrfs_truncate(), this metadata leak will impact other's metadata reservation. In this case, it's "btrfs_start_transaction(root, 2);" fails for enospc error, which make this truncate operation fail. Call btrfs_block_rsv_release() to fix this bug. Signed-off-by: Wang Xiaoguang Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/inode.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 128e52489f81..415a6ad5ba22 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -9215,6 +9215,7 @@ static int btrfs_truncate(struct inode *inode) break; } + btrfs_block_rsv_release(fs_info, rsv, -1); ret = btrfs_block_rsv_migrate(&fs_info->trans_block_rsv, rsv, min_size, 0); BUG_ON(ret); /* shouldn't happen */ From 97dcdea076ecef41ea4aaa23d4397c2f622e4265 Mon Sep 17 00:00:00 2001 From: Chandan Rajendra Date: Fri, 23 Dec 2016 15:00:18 +0530 Subject: [PATCH 122/343] Btrfs: Fix deadlock between direct IO and fast fsync The following deadlock is seen when executing generic/113 test, ---------------------------------------------------------+---------------------------------------------------- Direct I/O task Fast fsync task ---------------------------------------------------------+---------------------------------------------------- btrfs_direct_IO __blockdev_direct_IO do_blockdev_direct_IO do_direct_IO btrfs_get_blocks_direct while (blocks needs to written) get_more_blocks (first iteration) btrfs_get_blocks_direct btrfs_create_dio_extent down_read(&BTRFS_I(inode) >dio_sem) Create and add extent map and ordered extent up_read(&BTRFS_I(inode) >dio_sem) btrfs_sync_file btrfs_log_dentry_safe btrfs_log_inode_parent btrfs_log_inode btrfs_log_changed_extents down_write(&BTRFS_I(inode) >dio_sem) Collect new extent maps and ordered extents wait for ordered extent completion get_more_blocks (second iteration) btrfs_get_blocks_direct btrfs_create_dio_extent down_read(&BTRFS_I(inode) >dio_sem) -------------------------------------------------------------------------------------------------------------- In the above description, Btrfs direct I/O code path has not yet started submitting bios for file range covered by the initial ordered extent. Meanwhile, The fast fsync task obtains the write semaphore and waits for I/O on the ordered extent to get completed. However, the Direct I/O task is now blocked on obtaining the read semaphore. To resolve the deadlock, this commit modifies the Direct I/O code path to obtain the read semaphore before invoking __blockdev_direct_IO(). The semaphore is then given up after __blockdev_direct_IO() returns. This allows the Direct I/O code to complete I/O on all the ordered extents it creates. Signed-off-by: Chandan Rajendra Reviewed-by: Filipe Manana Signed-off-by: David Sterba --- fs/btrfs/inode.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 415a6ad5ba22..bed6cf5f67a4 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -7215,7 +7215,6 @@ static struct extent_map *btrfs_create_dio_extent(struct inode *inode, struct extent_map *em = NULL; int ret; - down_read(&BTRFS_I(inode)->dio_sem); if (type != BTRFS_ORDERED_NOCOW) { em = create_pinned_em(inode, start, len, orig_start, block_start, block_len, orig_block_len, @@ -7234,7 +7233,6 @@ static struct extent_map *btrfs_create_dio_extent(struct inode *inode, em = ERR_PTR(ret); } out: - up_read(&BTRFS_I(inode)->dio_sem); return em; } @@ -8695,6 +8693,7 @@ static ssize_t btrfs_direct_IO(struct kiocb *iocb, struct iov_iter *iter) dio_data.unsubmitted_oe_range_start = (u64)offset; dio_data.unsubmitted_oe_range_end = (u64)offset; current->journal_info = &dio_data; + down_read(&BTRFS_I(inode)->dio_sem); } else if (test_bit(BTRFS_INODE_READDIO_NEED_LOCK, &BTRFS_I(inode)->runtime_flags)) { inode_dio_end(inode); @@ -8707,6 +8706,7 @@ static ssize_t btrfs_direct_IO(struct kiocb *iocb, struct iov_iter *iter) iter, btrfs_get_blocks_direct, NULL, btrfs_submit_direct, flags); if (iov_iter_rw(iter) == WRITE) { + up_read(&BTRFS_I(inode)->dio_sem); current->journal_info = NULL; if (ret < 0 && ret != -EIOCBQUEUED) { if (dio_data.reserve) From 91298eec05cd8d4e828cf7ee5d4a6334f70cf69a Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Thu, 1 Dec 2016 13:43:31 -0800 Subject: [PATCH 123/343] Btrfs: fix truncate down when no_holes feature is enabled For such a file mapping, [0-4k][hole][8k-12k] In NO_HOLES mode, we don't have the [hole] extent any more. Commit c1aa45759e90 ("Btrfs: fix shrinking truncate when the no_holes feature is enabled") fixed disk isize not being updated in NO_HOLES mode when data is not flushed. However, even if data has been flushed, we can still have trouble in updating disk isize since we updated disk isize to 'start' of the last evicted extent. Reviewed-by: Chris Mason Signed-off-by: Liu Bo Signed-off-by: David Sterba --- fs/btrfs/inode.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-) diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index bed6cf5f67a4..177e7284909c 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -4505,8 +4505,19 @@ search_again: if (found_type > min_type) { del_item = 1; } else { - if (item_end < new_size) + if (item_end < new_size) { + /* + * With NO_HOLES mode, for the following mapping + * + * [0-4k][hole][8k-12k] + * + * if truncating isize down to 6k, it ends up + * isize being 8k. + */ + if (btrfs_fs_incompat(root->fs_info, NO_HOLES)) + last_size = new_size; break; + } if (found_key.offset >= new_size) del_item = 1; else From e326ce013a8e851193eb337aafb1aa396c533a61 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Fri, 20 Jan 2017 03:25:34 +0100 Subject: [PATCH 124/343] Revert "PM / sleep / ACPI: Use the ACPI_FADT_LOW_POWER_S0 flag" Revert commit 08b98d329165 (PM / sleep / ACPI: Use the ACPI_FADT_LOW_POWER_S0 flag) as it caused system suspend (in the default configuration) to fail on Dell XPS13 (9360) with the Kaby Lake processor. Fixes: 08b98d329165 (PM / sleep / ACPI: Use the ACPI_FADT_LOW_POWER_S0 flag) Reported-by: Paul Menzel Signed-off-by: Rafael J. Wysocki --- Documentation/power/states.txt | 4 +--- drivers/acpi/sleep.c | 8 -------- include/linux/suspend.h | 2 -- kernel/power/suspend.c | 4 ++-- 4 files changed, 3 insertions(+), 15 deletions(-) diff --git a/Documentation/power/states.txt b/Documentation/power/states.txt index 8a39ce45d8a0..008ecb588317 100644 --- a/Documentation/power/states.txt +++ b/Documentation/power/states.txt @@ -35,9 +35,7 @@ only one way to cause the system to go into the Suspend-To-RAM state (write The default suspend mode (ie. the one to be used without writing anything into /sys/power/mem_sleep) is either "deep" (if Suspend-To-RAM is supported) or "s2idle", but it can be overridden by the value of the "mem_sleep_default" -parameter in the kernel command line. On some ACPI-based systems, depending on -the information in the FADT, the default may be "s2idle" even if Suspend-To-RAM -is supported. +parameter in the kernel command line. The properties of all of the sleep states are described below. diff --git a/drivers/acpi/sleep.c b/drivers/acpi/sleep.c index ce1855fd584b..deb0ff78eba8 100644 --- a/drivers/acpi/sleep.c +++ b/drivers/acpi/sleep.c @@ -691,14 +691,6 @@ static void acpi_sleep_suspend_setup(void) if (acpi_sleep_state_supported(i)) sleep_states[i] = 1; - /* - * Use suspend-to-idle by default if ACPI_FADT_LOW_POWER_S0 is set and - * the default suspend mode was not selected from the command line. - */ - if (acpi_gbl_FADT.flags & ACPI_FADT_LOW_POWER_S0 && - mem_sleep_default > PM_SUSPEND_MEM) - mem_sleep_default = PM_SUSPEND_FREEZE; - suspend_set_ops(old_suspend_ordering ? &acpi_suspend_ops_old : &acpi_suspend_ops); freeze_set_ops(&acpi_freeze_ops); diff --git a/include/linux/suspend.h b/include/linux/suspend.h index 0c729c3c8549..d9718378a8be 100644 --- a/include/linux/suspend.h +++ b/include/linux/suspend.h @@ -194,8 +194,6 @@ struct platform_freeze_ops { }; #ifdef CONFIG_SUSPEND -extern suspend_state_t mem_sleep_default; - /** * suspend_set_ops - set platform dependent suspend operations * @ops: The new suspend operations to set. diff --git a/kernel/power/suspend.c b/kernel/power/suspend.c index f67ceb7768b8..15e6baef5c73 100644 --- a/kernel/power/suspend.c +++ b/kernel/power/suspend.c @@ -46,7 +46,7 @@ static const char * const mem_sleep_labels[] = { const char *mem_sleep_states[PM_SUSPEND_MAX]; suspend_state_t mem_sleep_current = PM_SUSPEND_FREEZE; -suspend_state_t mem_sleep_default = PM_SUSPEND_MAX; +static suspend_state_t mem_sleep_default = PM_SUSPEND_MEM; unsigned int pm_suspend_global_flags; EXPORT_SYMBOL_GPL(pm_suspend_global_flags); @@ -168,7 +168,7 @@ void suspend_set_ops(const struct platform_suspend_ops *ops) } if (valid_state(PM_SUSPEND_MEM)) { mem_sleep_states[PM_SUSPEND_MEM] = mem_sleep_labels[PM_SUSPEND_MEM]; - if (mem_sleep_default >= PM_SUSPEND_MEM) + if (mem_sleep_default == PM_SUSPEND_MEM) mem_sleep_current = PM_SUSPEND_MEM; } From 1443ebbacfd7f8d84fbbbf629ef66a12dc8a4b4e Mon Sep 17 00:00:00 2001 From: Srinivas Pandruvada Date: Wed, 18 Jan 2017 10:48:22 -0800 Subject: [PATCH 125/343] cpufreq: intel_pstate: Fix sysfs limits enforcement for performance policy A side effect of keeping intel_pstate sysfs limits in sync with cpufreq is that the now sysfs limits can't enforced under performance policy. For example, if the max_perf_pct is changed from 100 to 80, this will call intel_pstate_set_policy(), which will change the max_perf to 100 again for performance policy. Same issue happens, when no_turbo is set. This change calculates max and min frequency using sysfs performance limits in intel_pstate_verify_policy() and adjusts policy limits by calling cpufreq_verify_within_limits(). Also, it causes the setting of performance limits to be skipped if no_turbo is set. Fixes: 111b8b3fe4fa (cpufreq: intel_pstate: Always keep all limits settings in sync) Signed-off-by: Srinivas Pandruvada Signed-off-by: Rafael J. Wysocki --- drivers/cpufreq/intel_pstate.c | 14 +++++++++++++- 1 file changed, 13 insertions(+), 1 deletion(-) diff --git a/drivers/cpufreq/intel_pstate.c b/drivers/cpufreq/intel_pstate.c index e261438cd690..ed215c5592f6 100644 --- a/drivers/cpufreq/intel_pstate.c +++ b/drivers/cpufreq/intel_pstate.c @@ -1951,7 +1951,8 @@ static int intel_pstate_set_policy(struct cpufreq_policy *policy) limits = &performance_limits; perf_limits = limits; } - if (policy->max >= policy->cpuinfo.max_freq) { + if (policy->max >= policy->cpuinfo.max_freq && + !limits->no_turbo) { pr_debug("set performance\n"); intel_pstate_set_performance_limits(perf_limits); goto out; @@ -1993,6 +1994,17 @@ static int intel_pstate_verify_policy(struct cpufreq_policy *policy) policy->policy != CPUFREQ_POLICY_PERFORMANCE) return -EINVAL; + /* When per-CPU limits are used, sysfs limits are not used */ + if (!per_cpu_limits) { + unsigned int max_freq, min_freq; + + max_freq = policy->cpuinfo.max_freq * + limits->max_sysfs_pct / 100; + min_freq = policy->cpuinfo.max_freq * + limits->min_sysfs_pct / 100; + cpufreq_verify_within_limits(policy, min_freq, max_freq); + } + return 0; } From 7a37052adb5e843bcfff6c98aee9b60bb087b910 Mon Sep 17 00:00:00 2001 From: Lv Zheng Date: Thu, 19 Jan 2017 15:21:34 +0800 Subject: [PATCH 126/343] ACPICA: Tables: Fix hidden logic related to acpi_tb_install_standard_table() There is a hidden logic for acpi_tb_install_standard_table() as it can be invoked from the boot stage and during runtime. 1. When it is invoked from the OS boot stage, the ACPICA mutex may not have been initialized yet and so acpi_ut_acquire_mutex()/acpi_ut_release_mutex() are not invoked in these code paths: acpi_initialize_tables acpi_tb_parse_root_table acpi_tb_install_standard_table (4 invocations) acpi_install_table acpi_tb_install_standard_table 2. When it is invoked during the runtime, ACPICA mutex is used as appropriate: acpi_ex_load_op acpi_tb_install_and_load_table acpi_tb_install_standard_table acpi_load_table acpi_tb_install_and_load_table acpi_tb_install_standard_table The mutex is now used in acpi_tb_install_and_load_table(), while it actually should be in acpi_tb_install_standard_table(). This introduces another problem in acpi_tb_install_standard_table() where acpi_gbl_table_handler is invoked from and the lock contexts are thus not consistent for the table handlers. This triggers a regression when acpi_get_table()/acpi_put_table() start to hold table mutex during runtime. The regression is noticed by LKP as new errors reported by ACPICA mutex debugging facility. [ 2.043693] ACPI Error: Mutex [ACPI_MTX_Tables] already acquired by this thread [497483776] (20160930/utmutex-254) [ 2.054084] ACPI Error: Mutex [0x2] is not acquired, cannot release (20160930/utmutex-326) And it triggers a deadlock: [ 247.066214] INFO: task swapper/0:1 blocked for more than 120 seconds. ... [ 247.091271] Call Trace: ... [ 247.121523] down_timeout+0x47/0x50 [ 247.125065] acpi_os_wait_semaphore+0x47/0x62 [ 247.129475] acpi_ut_acquire_mutex+0x43/0x81 [ 247.133798] acpi_get_table+0x2d/0x84 [ 247.137513] acpi_table_attr_init+0xcd/0x100 [ 247.146590] acpi_sysfs_table_handler+0x5d/0xb8 [ 247.151174] acpi_bus_table_handler+0x23/0x2a [ 247.155583] acpi_tb_install_standard_table+0xe0/0x213 [ 247.164489] acpi_tb_install_and_load_table+0x3a/0x82 [ 247.169592] acpi_ex_load_op+0x194/0x201 ... [ 247.200108] acpi_ns_evaluate+0x1bb/0x247 [ 247.204170] acpi_evaluate_object+0x178/0x274 [ 247.213249] acpi_processor_set_pdc+0x154/0x17b ... The table mutex is held in acpi_tb_install_and_load_table() and is re-visited by acpi_get_table(). Noticing that the early mutex requirement actually belongs to the OSL layer and has already been handled in acpi_os_wait_semaphore()/acpi_os_signal_semaphore(), the regression canbe fixed by removing this hidden logic from the ACPICA core to the OS-specific code. Fixes: 174cc7187e6f ("ACPICA: Tables: Back port acpi_get_table_with_size() and early_acpi_os_unmap_memory() from Linux kernel") Reported-and-tested-by: Tomi Sarvela Reported-by: Ye Xiaolong Signed-off-by: Lv Zheng Signed-off-by: Rafael J. Wysocki --- drivers/acpi/acpica/tbdata.c | 9 ++------- drivers/acpi/acpica/tbinstal.c | 17 +++++++++++++++-- 2 files changed, 17 insertions(+), 9 deletions(-) diff --git a/drivers/acpi/acpica/tbdata.c b/drivers/acpi/acpica/tbdata.c index 82b0b5710979..b0399e8f6d27 100644 --- a/drivers/acpi/acpica/tbdata.c +++ b/drivers/acpi/acpica/tbdata.c @@ -852,23 +852,18 @@ acpi_tb_install_and_load_table(acpi_physical_address address, ACPI_FUNCTION_TRACE(tb_install_and_load_table); - (void)acpi_ut_acquire_mutex(ACPI_MTX_TABLES); - /* Install the table and load it into the namespace */ status = acpi_tb_install_standard_table(address, flags, TRUE, override, &i); if (ACPI_FAILURE(status)) { - goto unlock_and_exit; + goto exit; } - (void)acpi_ut_release_mutex(ACPI_MTX_TABLES); status = acpi_tb_load_table(i, acpi_gbl_root_node); - (void)acpi_ut_acquire_mutex(ACPI_MTX_TABLES); -unlock_and_exit: +exit: *table_index = i; - (void)acpi_ut_release_mutex(ACPI_MTX_TABLES); return_ACPI_STATUS(status); } diff --git a/drivers/acpi/acpica/tbinstal.c b/drivers/acpi/acpica/tbinstal.c index 5fdf251a9f97..01e1b3d63fc0 100644 --- a/drivers/acpi/acpica/tbinstal.c +++ b/drivers/acpi/acpica/tbinstal.c @@ -217,6 +217,10 @@ acpi_tb_install_standard_table(acpi_physical_address address, goto release_and_exit; } + /* Acquire the table lock */ + + (void)acpi_ut_acquire_mutex(ACPI_MTX_TABLES); + if (reload) { /* * Validate the incoming table signature. @@ -244,7 +248,7 @@ acpi_tb_install_standard_table(acpi_physical_address address, new_table_desc.signature.integer)); status = AE_BAD_SIGNATURE; - goto release_and_exit; + goto unlock_and_exit; } /* Check if table is already registered */ @@ -279,7 +283,7 @@ acpi_tb_install_standard_table(acpi_physical_address address, /* Table is still loaded, this is an error */ status = AE_ALREADY_EXISTS; - goto release_and_exit; + goto unlock_and_exit; } else { /* * Table was unloaded, allow it to be reloaded. @@ -290,6 +294,7 @@ acpi_tb_install_standard_table(acpi_physical_address address, * indicate the re-installation. */ acpi_tb_uninstall_table(&new_table_desc); + (void)acpi_ut_release_mutex(ACPI_MTX_TABLES); *table_index = i; return_ACPI_STATUS(AE_OK); } @@ -303,11 +308,19 @@ acpi_tb_install_standard_table(acpi_physical_address address, /* Invoke table handler if present */ + (void)acpi_ut_release_mutex(ACPI_MTX_TABLES); if (acpi_gbl_table_handler) { (void)acpi_gbl_table_handler(ACPI_TABLE_EVENT_INSTALL, new_table_desc.pointer, acpi_gbl_table_handler_context); } + (void)acpi_ut_acquire_mutex(ACPI_MTX_TABLES); + +unlock_and_exit: + + /* Release the table lock */ + + (void)acpi_ut_release_mutex(ACPI_MTX_TABLES); release_and_exit: From 4151e9a61c26bc86a356edfea713c0f913582760 Mon Sep 17 00:00:00 2001 From: Andrzej Hajda Date: Tue, 17 Jan 2017 15:15:18 +0100 Subject: [PATCH 127/343] drm/exynos/decon5433: do not disable video after reset decon_commit is called just after reset so video is disabled anyway. Signed-off-by: Andrzej Hajda Signed-off-by: Inki Dae --- drivers/gpu/drm/exynos/exynos5433_drm_decon.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/drivers/gpu/drm/exynos/exynos5433_drm_decon.c b/drivers/gpu/drm/exynos/exynos5433_drm_decon.c index 09e8cc36948e..d4920091e35c 100644 --- a/drivers/gpu/drm/exynos/exynos5433_drm_decon.c +++ b/drivers/gpu/drm/exynos/exynos5433_drm_decon.c @@ -142,8 +142,6 @@ static void decon_commit(struct exynos_drm_crtc *crtc) m->crtc_vsync_end = m->crtc_vsync_start + 1; } - decon_set_bits(ctx, DECON_VIDCON0, VIDCON0_ENVID, 0); - /* enable clock gate */ val = CMU_CLKGAGE_MODE_SFR_F | CMU_CLKGAGE_MODE_MEM_F; writel(val, ctx->addr + DECON_CMU); From 1202a096328ed3de59e2a722038c4d80ec59a958 Mon Sep 17 00:00:00 2001 From: Andrzej Hajda Date: Tue, 17 Jan 2017 15:15:19 +0100 Subject: [PATCH 128/343] drm/exynos/decon5433: fix CMU programming DECON_CMU register has reserved bits which should not be zeroed, otherwise IP can behave strangely and cause IOMMU faults. Signed-off-by: Andrzej Hajda Signed-off-by: Inki Dae --- drivers/gpu/drm/exynos/exynos5433_drm_decon.c | 4 ---- 1 file changed, 4 deletions(-) diff --git a/drivers/gpu/drm/exynos/exynos5433_drm_decon.c b/drivers/gpu/drm/exynos/exynos5433_drm_decon.c index d4920091e35c..3a9aca1c62f9 100644 --- a/drivers/gpu/drm/exynos/exynos5433_drm_decon.c +++ b/drivers/gpu/drm/exynos/exynos5433_drm_decon.c @@ -142,10 +142,6 @@ static void decon_commit(struct exynos_drm_crtc *crtc) m->crtc_vsync_end = m->crtc_vsync_start + 1; } - /* enable clock gate */ - val = CMU_CLKGAGE_MODE_SFR_F | CMU_CLKGAGE_MODE_MEM_F; - writel(val, ctx->addr + DECON_CMU); - if (ctx->out_type & (IFTYPE_I80 | I80_HW_TRG)) decon_setup_trigger(ctx); From 11d8bcef7a0399e1d2519f207fd575fc404306b4 Mon Sep 17 00:00:00 2001 From: Andrzej Hajda Date: Tue, 17 Jan 2017 15:15:20 +0100 Subject: [PATCH 129/343] drm/exynos/decon5433: set STANDALONE_UPDATE_F on output enablement DECON_TV requires STANDALONE_UPDATE after output enabling, otherwise it does not start. This change is neutral for DECON. Signed-off-by: Andrzej Hajda Signed-off-by: Inki Dae --- drivers/gpu/drm/exynos/exynos5433_drm_decon.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/gpu/drm/exynos/exynos5433_drm_decon.c b/drivers/gpu/drm/exynos/exynos5433_drm_decon.c index 3a9aca1c62f9..75eeb831ed6a 100644 --- a/drivers/gpu/drm/exynos/exynos5433_drm_decon.c +++ b/drivers/gpu/drm/exynos/exynos5433_drm_decon.c @@ -183,6 +183,8 @@ static void decon_commit(struct exynos_drm_crtc *crtc) /* enable output and display signal */ decon_set_bits(ctx, DECON_VIDCON0, VIDCON0_ENVID | VIDCON0_ENVID_F, ~0); + + decon_set_bits(ctx, DECON_UPDATE, STANDALONE_UPDATE_F, ~0); } static void decon_win_set_pixfmt(struct decon_context *ctx, unsigned int win, From f1225ee4c8fcf09afaa199b8b1f0450f38b8cd11 Mon Sep 17 00:00:00 2001 From: Stefano Stabellini Date: Thu, 19 Jan 2017 10:39:09 -0800 Subject: [PATCH 130/343] swiotlb-xen: update dev_addr after swapping pages In xen_swiotlb_map_page and xen_swiotlb_map_sg_attrs, if the original page is not suitable, we swap it for another page from the swiotlb pool. In these cases, we don't update the previously calculated dma address for the page before calling xen_dma_map_page. Thus, we end up calling xen_dma_map_page passing the wrong dev_addr, resulting in xen_dma_map_page mistakenly assuming that the page is foreign when it is local. Fix the bug by updating dev_addr appropriately. This change has no effect on x86, because xen_dma_map_page is a stub there. Signed-off-by: Stefano Stabellini Signed-off-by: Pooya Keshavarzi Tested-by: Pooya Keshavarzi Reviewed-by: Boris Ostrovsky Signed-off-by: Konrad Rzeszutek Wilk --- drivers/xen/swiotlb-xen.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/drivers/xen/swiotlb-xen.c b/drivers/xen/swiotlb-xen.c index f905d6eeb048..f8afc6dcc29f 100644 --- a/drivers/xen/swiotlb-xen.c +++ b/drivers/xen/swiotlb-xen.c @@ -414,9 +414,9 @@ dma_addr_t xen_swiotlb_map_page(struct device *dev, struct page *page, if (map == SWIOTLB_MAP_ERROR) return DMA_ERROR_CODE; + dev_addr = xen_phys_to_bus(map); xen_dma_map_page(dev, pfn_to_page(map >> PAGE_SHIFT), dev_addr, map & ~PAGE_MASK, size, dir, attrs); - dev_addr = xen_phys_to_bus(map); /* * Ensure that the address returned is DMA'ble @@ -575,13 +575,14 @@ xen_swiotlb_map_sg_attrs(struct device *hwdev, struct scatterlist *sgl, sg_dma_len(sgl) = 0; return 0; } + dev_addr = xen_phys_to_bus(map); xen_dma_map_page(hwdev, pfn_to_page(map >> PAGE_SHIFT), dev_addr, map & ~PAGE_MASK, sg->length, dir, attrs); - sg->dma_address = xen_phys_to_bus(map); + sg->dma_address = dev_addr; } else { /* we are not interested in the dma_addr returned by * xen_dma_map_page, only in the potential cache flushes executed From 6391a4481ba0796805d6581e42f9f0418c099e34 Mon Sep 17 00:00:00 2001 From: Jason Wang Date: Fri, 20 Jan 2017 14:32:42 +0800 Subject: [PATCH 131/343] virtio-net: restore VIRTIO_HDR_F_DATA_VALID on receiving Commit 501db511397f ("virtio: don't set VIRTIO_NET_HDR_F_DATA_VALID on xmit") in fact disables VIRTIO_HDR_F_DATA_VALID on receiving path too, fixing this by adding a hint (has_data_valid) and set it only on the receiving path. Cc: Rolf Neugebauer Signed-off-by: Jason Wang Acked-by: Rolf Neugebauer Signed-off-by: David S. Miller --- drivers/net/macvtap.c | 2 +- drivers/net/tun.c | 2 +- drivers/net/virtio_net.c | 2 +- include/linux/virtio_net.h | 6 +++++- net/packet/af_packet.c | 4 ++-- 5 files changed, 10 insertions(+), 6 deletions(-) diff --git a/drivers/net/macvtap.c b/drivers/net/macvtap.c index 5c26653eceb5..402618565838 100644 --- a/drivers/net/macvtap.c +++ b/drivers/net/macvtap.c @@ -825,7 +825,7 @@ static ssize_t macvtap_put_user(struct macvtap_queue *q, return -EINVAL; if (virtio_net_hdr_from_skb(skb, &vnet_hdr, - macvtap_is_little_endian(q))) + macvtap_is_little_endian(q), true)) BUG(); if (copy_to_iter(&vnet_hdr, sizeof(vnet_hdr), iter) != diff --git a/drivers/net/tun.c b/drivers/net/tun.c index cd8e02c94be0..2cd10b26b650 100644 --- a/drivers/net/tun.c +++ b/drivers/net/tun.c @@ -1360,7 +1360,7 @@ static ssize_t tun_put_user(struct tun_struct *tun, return -EINVAL; if (virtio_net_hdr_from_skb(skb, &gso, - tun_is_little_endian(tun))) { + tun_is_little_endian(tun), true)) { struct skb_shared_info *sinfo = skb_shinfo(skb); pr_err("unexpected GSO type: " "0x%x, gso_size %d, hdr_len %d\n", diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c index 4a105006ca63..347424351ade 100644 --- a/drivers/net/virtio_net.c +++ b/drivers/net/virtio_net.c @@ -1104,7 +1104,7 @@ static int xmit_skb(struct send_queue *sq, struct sk_buff *skb) hdr = skb_vnet_hdr(skb); if (virtio_net_hdr_from_skb(skb, &hdr->hdr, - virtio_is_little_endian(vi->vdev))) + virtio_is_little_endian(vi->vdev), false)) BUG(); if (vi->mergeable_rx_bufs) diff --git a/include/linux/virtio_net.h b/include/linux/virtio_net.h index 56436472ccc7..5209b5ed2a64 100644 --- a/include/linux/virtio_net.h +++ b/include/linux/virtio_net.h @@ -56,7 +56,8 @@ static inline int virtio_net_hdr_to_skb(struct sk_buff *skb, static inline int virtio_net_hdr_from_skb(const struct sk_buff *skb, struct virtio_net_hdr *hdr, - bool little_endian) + bool little_endian, + bool has_data_valid) { memset(hdr, 0, sizeof(*hdr)); /* no info leak */ @@ -91,6 +92,9 @@ static inline int virtio_net_hdr_from_skb(const struct sk_buff *skb, skb_checksum_start_offset(skb)); hdr->csum_offset = __cpu_to_virtio16(little_endian, skb->csum_offset); + } else if (has_data_valid && + skb->ip_summed == CHECKSUM_UNNECESSARY) { + hdr->flags = VIRTIO_NET_HDR_F_DATA_VALID; } /* else everything is zero */ return 0; diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c index b9e1a13b4ba3..3d555c79a7b5 100644 --- a/net/packet/af_packet.c +++ b/net/packet/af_packet.c @@ -1976,7 +1976,7 @@ static int packet_rcv_vnet(struct msghdr *msg, const struct sk_buff *skb, return -EINVAL; *len -= sizeof(vnet_hdr); - if (virtio_net_hdr_from_skb(skb, &vnet_hdr, vio_le())) + if (virtio_net_hdr_from_skb(skb, &vnet_hdr, vio_le(), true)) return -EINVAL; return memcpy_to_msg(msg, (void *)&vnet_hdr, sizeof(vnet_hdr)); @@ -2237,7 +2237,7 @@ static int tpacket_rcv(struct sk_buff *skb, struct net_device *dev, if (po->has_vnet_hdr) { if (virtio_net_hdr_from_skb(skb, h.raw + macoff - sizeof(struct virtio_net_hdr), - vio_le())) { + vio_le(), true)) { spin_lock(&sk->sk_receive_queue.lock); goto drop_n_account; } From 2c561b2b728ca4013e76d6439bde2c137503745e Mon Sep 17 00:00:00 2001 From: hayeswang Date: Fri, 20 Jan 2017 14:33:55 +0800 Subject: [PATCH 132/343] r8152: fix rtl8152_post_reset function The rtl8152_post_reset() should sumbit rx urb and interrupt transfer, otherwise the rx wouldn't work and the linking change couldn't be detected. Signed-off-by: Hayes Wang Signed-off-by: David S. Miller --- drivers/net/usb/r8152.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/net/usb/r8152.c b/drivers/net/usb/r8152.c index f3b48ad90865..0e99af090734 100644 --- a/drivers/net/usb/r8152.c +++ b/drivers/net/usb/r8152.c @@ -3545,12 +3545,14 @@ static int rtl8152_post_reset(struct usb_interface *intf) if (netif_carrier_ok(netdev)) { mutex_lock(&tp->control); tp->rtl_ops.enable(tp); + rtl_start_rx(tp); rtl8152_set_rx_mode(netdev); mutex_unlock(&tp->control); netif_wake_queue(netdev); } napi_enable(&tp->napi); + usb_submit_urb(tp->intr_urb, GFP_KERNEL); return 0; } From e363116b90906f326c9cde5473b4b9a99ba476df Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Fri, 20 Jan 2017 07:57:42 -0800 Subject: [PATCH 133/343] ipv6: seg6_genl_set_tunsrc() must check kmemdup() return value seg6_genl_get_tunsrc() and set_tun_src() do not handle tun_src being possibly NULL, so we must check kmemdup() return value and abort if it is NULL Fixes: 915d7e5e5930 ("ipv6: sr: add code base for control plane support of SR-IPv6") Signed-off-by: Eric Dumazet Cc: David Lebrun Acked-by: David Lebrun Signed-off-by: David S. Miller --- net/ipv6/seg6.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/net/ipv6/seg6.c b/net/ipv6/seg6.c index b172d85c650a..a855eb325b03 100644 --- a/net/ipv6/seg6.c +++ b/net/ipv6/seg6.c @@ -176,6 +176,8 @@ static int seg6_genl_set_tunsrc(struct sk_buff *skb, struct genl_info *info) val = nla_data(info->attrs[SEG6_ATTR_DST]); t_new = kmemdup(val, sizeof(*val), GFP_KERNEL); + if (!t_new) + return -ENOMEM; mutex_lock(&sdata->lock); From 0e73fc9a56f22f2eec4d2b2910c649f7af67b74d Mon Sep 17 00:00:00 2001 From: Colin Ian King Date: Fri, 20 Jan 2017 13:01:57 +0000 Subject: [PATCH 134/343] net: sctp: fix array overrun read on sctp_timer_tbl The comparison on the timeout can lead to an array overrun read on sctp_timer_tbl because of an off-by-one error. Fix this by using < instead of <= and also compare to the array size rather than SCTP_EVENT_TIMEOUT_MAX. Fixes CoverityScan CID#1397639 ("Out-of-bounds read") Signed-off-by: Colin Ian King Signed-off-by: David S. Miller --- net/sctp/debug.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/sctp/debug.c b/net/sctp/debug.c index 95d7b15dad21..e371a0d90068 100644 --- a/net/sctp/debug.c +++ b/net/sctp/debug.c @@ -166,7 +166,7 @@ static const char *const sctp_timer_tbl[] = { /* Lookup timer debug name. */ const char *sctp_tname(const sctp_subtype_t id) { - if (id.timeout <= SCTP_EVENT_TIMEOUT_MAX) + if (id.timeout < ARRAY_SIZE(sctp_timer_tbl)) return sctp_timer_tbl[id.timeout]; return "unknown_timer"; } From 91e744653cb80554f3fdfd1d31c5ddf7b6169f37 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Fri, 20 Jan 2017 11:29:43 -0500 Subject: [PATCH 135/343] Revert "net: sctp: fix array overrun read on sctp_timer_tbl" This reverts commit 0e73fc9a56f22f2eec4d2b2910c649f7af67b74d. This fix wasn't correct, a better one is coming right up. Signed-off-by: David S. Miller --- net/sctp/debug.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/sctp/debug.c b/net/sctp/debug.c index e371a0d90068..95d7b15dad21 100644 --- a/net/sctp/debug.c +++ b/net/sctp/debug.c @@ -166,7 +166,7 @@ static const char *const sctp_timer_tbl[] = { /* Lookup timer debug name. */ const char *sctp_tname(const sctp_subtype_t id) { - if (id.timeout < ARRAY_SIZE(sctp_timer_tbl)) + if (id.timeout <= SCTP_EVENT_TIMEOUT_MAX) return sctp_timer_tbl[id.timeout]; return "unknown_timer"; } From 0629a330cf55454962168dd3ee46fad53a39323e Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Wed, 18 Jan 2017 15:52:52 +0100 Subject: [PATCH 136/343] qed: avoid possible stack overflow in qed_ll2_acquire_connection struct qed_ll2_info is rather large, so putting it on the stack can cause an overflow, as this warning tries to tell us: drivers/net/ethernet/qlogic/qed/qed_ll2.c: In function 'qed_ll2_start': drivers/net/ethernet/qlogic/qed/qed_ll2.c:2159:1: error: the frame size of 1056 bytes is larger than 1024 bytes [-Werror=frame-larger-than=] qed_ll2_start_ooo() already uses a dynamic allocation for the structure to work around that problem, and we could do the same in qed_ll2_start() as well as qed_roce_ll2_start(), but since the structure is only used to pass a couple of initialization values here, it seems nicer to replace it with a different structure. Lacking any idea for better naming, I'm adding 'struct qed_ll2_conn', which now contains all the initialization data, and this now simply gets copied into struct qed_ll2_info rather than assigning all members one by one. Signed-off-by: Arnd Bergmann Acked-by: Yuval Mintz Signed-off-by: David S. Miller --- drivers/net/ethernet/qlogic/qed/qed_ll2.c | 86 ++++++++++------------ drivers/net/ethernet/qlogic/qed/qed_ll2.h | 24 +++--- drivers/net/ethernet/qlogic/qed/qed_roce.c | 2 +- 3 files changed, 52 insertions(+), 60 deletions(-) diff --git a/drivers/net/ethernet/qlogic/qed/qed_ll2.c b/drivers/net/ethernet/qlogic/qed/qed_ll2.c index 8e5cb7605b0f..873ce2cd76ba 100644 --- a/drivers/net/ethernet/qlogic/qed/qed_ll2.c +++ b/drivers/net/ethernet/qlogic/qed/qed_ll2.c @@ -297,7 +297,7 @@ static void qed_ll2_txq_flush(struct qed_hwfn *p_hwfn, u8 connection_handle) list_del(&p_pkt->list_entry); b_last_packet = list_empty(&p_tx->active_descq); list_add_tail(&p_pkt->list_entry, &p_tx->free_descq); - if (p_ll2_conn->conn_type == QED_LL2_TYPE_ISCSI_OOO) { + if (p_ll2_conn->conn.conn_type == QED_LL2_TYPE_ISCSI_OOO) { struct qed_ooo_buffer *p_buffer; p_buffer = (struct qed_ooo_buffer *)p_pkt->cookie; @@ -309,7 +309,7 @@ static void qed_ll2_txq_flush(struct qed_hwfn *p_hwfn, u8 connection_handle) b_last_frag = p_tx->cur_completing_bd_idx == p_pkt->bd_used; tx_frag = p_pkt->bds_set[0].tx_frag; - if (p_ll2_conn->gsi_enable) + if (p_ll2_conn->conn.gsi_enable) qed_ll2b_release_tx_gsi_packet(p_hwfn, p_ll2_conn-> my_id, @@ -378,7 +378,7 @@ static int qed_ll2_txq_completion(struct qed_hwfn *p_hwfn, void *p_cookie) spin_unlock_irqrestore(&p_tx->lock, flags); tx_frag = p_pkt->bds_set[0].tx_frag; - if (p_ll2_conn->gsi_enable) + if (p_ll2_conn->conn.gsi_enable) qed_ll2b_complete_tx_gsi_packet(p_hwfn, p_ll2_conn->my_id, p_pkt->cookie, @@ -550,7 +550,7 @@ static void qed_ll2_rxq_flush(struct qed_hwfn *p_hwfn, u8 connection_handle) list_move_tail(&p_pkt->list_entry, &p_rx->free_descq); - if (p_ll2_conn->conn_type == QED_LL2_TYPE_ISCSI_OOO) { + if (p_ll2_conn->conn.conn_type == QED_LL2_TYPE_ISCSI_OOO) { struct qed_ooo_buffer *p_buffer; p_buffer = (struct qed_ooo_buffer *)p_pkt->cookie; @@ -738,7 +738,7 @@ qed_ooo_submit_tx_buffers(struct qed_hwfn *p_hwfn, rc = qed_ll2_prepare_tx_packet(p_hwfn, p_ll2_conn->my_id, 1, p_buffer->vlan, bd_flags, l4_hdr_offset_w, - p_ll2_conn->tx_dest, 0, + p_ll2_conn->conn.tx_dest, 0, first_frag, p_buffer->packet_length, p_buffer, true); @@ -858,7 +858,7 @@ qed_ll2_acquire_connection_ooo(struct qed_hwfn *p_hwfn, u16 buf_idx; int rc = 0; - if (p_ll2_info->conn_type != QED_LL2_TYPE_ISCSI_OOO) + if (p_ll2_info->conn.conn_type != QED_LL2_TYPE_ISCSI_OOO) return rc; if (!rx_num_ooo_buffers) @@ -901,7 +901,7 @@ static void qed_ll2_establish_connection_ooo(struct qed_hwfn *p_hwfn, struct qed_ll2_info *p_ll2_conn) { - if (p_ll2_conn->conn_type != QED_LL2_TYPE_ISCSI_OOO) + if (p_ll2_conn->conn.conn_type != QED_LL2_TYPE_ISCSI_OOO) return; qed_ooo_release_all_isles(p_hwfn, p_hwfn->p_ooo_info); @@ -913,7 +913,7 @@ static void qed_ll2_release_connection_ooo(struct qed_hwfn *p_hwfn, { struct qed_ooo_buffer *p_buffer; - if (p_ll2_conn->conn_type != QED_LL2_TYPE_ISCSI_OOO) + if (p_ll2_conn->conn.conn_type != QED_LL2_TYPE_ISCSI_OOO) return; qed_ooo_release_all_isles(p_hwfn, p_hwfn->p_ooo_info); @@ -945,23 +945,19 @@ static int qed_ll2_start_ooo(struct qed_dev *cdev, { struct qed_hwfn *hwfn = QED_LEADING_HWFN(cdev); u8 *handle = &hwfn->pf_params.iscsi_pf_params.ll2_ooo_queue_id; - struct qed_ll2_info *ll2_info; + struct qed_ll2_conn ll2_info; int rc; - ll2_info = kzalloc(sizeof(*ll2_info), GFP_KERNEL); - if (!ll2_info) - return -ENOMEM; - ll2_info->conn_type = QED_LL2_TYPE_ISCSI_OOO; - ll2_info->mtu = params->mtu; - ll2_info->rx_drop_ttl0_flg = params->drop_ttl0_packets; - ll2_info->rx_vlan_removal_en = params->rx_vlan_stripping; - ll2_info->tx_tc = OOO_LB_TC; - ll2_info->tx_dest = CORE_TX_DEST_LB; + ll2_info.conn_type = QED_LL2_TYPE_ISCSI_OOO; + ll2_info.mtu = params->mtu; + ll2_info.rx_drop_ttl0_flg = params->drop_ttl0_packets; + ll2_info.rx_vlan_removal_en = params->rx_vlan_stripping; + ll2_info.tx_tc = OOO_LB_TC; + ll2_info.tx_dest = CORE_TX_DEST_LB; - rc = qed_ll2_acquire_connection(hwfn, ll2_info, + rc = qed_ll2_acquire_connection(hwfn, &ll2_info, QED_LL2_RX_SIZE, QED_LL2_TX_SIZE, handle); - kfree(ll2_info); if (rc) { DP_INFO(cdev, "Failed to acquire LL2 OOO connection\n"); goto out; @@ -1006,7 +1002,7 @@ static int qed_sp_ll2_rx_queue_start(struct qed_hwfn *p_hwfn, struct qed_ll2_info *p_ll2_conn, u8 action_on_error) { - enum qed_ll2_conn_type conn_type = p_ll2_conn->conn_type; + enum qed_ll2_conn_type conn_type = p_ll2_conn->conn.conn_type; struct qed_ll2_rx_queue *p_rx = &p_ll2_conn->rx_queue; struct core_rx_start_ramrod_data *p_ramrod = NULL; struct qed_spq_entry *p_ent = NULL; @@ -1032,7 +1028,7 @@ static int qed_sp_ll2_rx_queue_start(struct qed_hwfn *p_hwfn, p_ramrod->sb_index = p_rx->rx_sb_index; p_ramrod->complete_event_flg = 1; - p_ramrod->mtu = cpu_to_le16(p_ll2_conn->mtu); + p_ramrod->mtu = cpu_to_le16(p_ll2_conn->conn.mtu); DMA_REGPAIR_LE(p_ramrod->bd_base, p_rx->rxq_chain.p_phys_addr); cqe_pbl_size = (u16)qed_chain_get_page_cnt(&p_rx->rcq_chain); @@ -1040,8 +1036,8 @@ static int qed_sp_ll2_rx_queue_start(struct qed_hwfn *p_hwfn, DMA_REGPAIR_LE(p_ramrod->cqe_pbl_addr, qed_chain_get_pbl_phys(&p_rx->rcq_chain)); - p_ramrod->drop_ttl0_flg = p_ll2_conn->rx_drop_ttl0_flg; - p_ramrod->inner_vlan_removal_en = p_ll2_conn->rx_vlan_removal_en; + p_ramrod->drop_ttl0_flg = p_ll2_conn->conn.rx_drop_ttl0_flg; + p_ramrod->inner_vlan_removal_en = p_ll2_conn->conn.rx_vlan_removal_en; p_ramrod->queue_id = p_ll2_conn->queue_id; p_ramrod->main_func_queue = (conn_type == QED_LL2_TYPE_ISCSI_OOO) ? 0 : 1; @@ -1056,14 +1052,14 @@ static int qed_sp_ll2_rx_queue_start(struct qed_hwfn *p_hwfn, } p_ramrod->action_on_error.error_type = action_on_error; - p_ramrod->gsi_offload_flag = p_ll2_conn->gsi_enable; + p_ramrod->gsi_offload_flag = p_ll2_conn->conn.gsi_enable; return qed_spq_post(p_hwfn, p_ent, NULL); } static int qed_sp_ll2_tx_queue_start(struct qed_hwfn *p_hwfn, struct qed_ll2_info *p_ll2_conn) { - enum qed_ll2_conn_type conn_type = p_ll2_conn->conn_type; + enum qed_ll2_conn_type conn_type = p_ll2_conn->conn.conn_type; struct qed_ll2_tx_queue *p_tx = &p_ll2_conn->tx_queue; struct core_tx_start_ramrod_data *p_ramrod = NULL; struct qed_spq_entry *p_ent = NULL; @@ -1075,7 +1071,7 @@ static int qed_sp_ll2_tx_queue_start(struct qed_hwfn *p_hwfn, if (!QED_LL2_TX_REGISTERED(p_ll2_conn)) return 0; - if (p_ll2_conn->conn_type == QED_LL2_TYPE_ISCSI_OOO) + if (p_ll2_conn->conn.conn_type == QED_LL2_TYPE_ISCSI_OOO) p_ll2_conn->tx_stats_en = 0; else p_ll2_conn->tx_stats_en = 1; @@ -1096,7 +1092,7 @@ static int qed_sp_ll2_tx_queue_start(struct qed_hwfn *p_hwfn, p_ramrod->sb_id = cpu_to_le16(qed_int_get_sp_sb_id(p_hwfn)); p_ramrod->sb_index = p_tx->tx_sb_index; - p_ramrod->mtu = cpu_to_le16(p_ll2_conn->mtu); + p_ramrod->mtu = cpu_to_le16(p_ll2_conn->conn.mtu); p_ramrod->stats_en = p_ll2_conn->tx_stats_en; p_ramrod->stats_id = p_ll2_conn->tx_stats_id; @@ -1106,7 +1102,7 @@ static int qed_sp_ll2_tx_queue_start(struct qed_hwfn *p_hwfn, p_ramrod->pbl_size = cpu_to_le16(pbl_size); memset(&pq_params, 0, sizeof(pq_params)); - pq_params.core.tc = p_ll2_conn->tx_tc; + pq_params.core.tc = p_ll2_conn->conn.tx_tc; pq_id = qed_get_qm_pq(p_hwfn, PROTOCOLID_CORE, &pq_params); p_ramrod->qm_pq_id = cpu_to_le16(pq_id); @@ -1123,7 +1119,7 @@ static int qed_sp_ll2_tx_queue_start(struct qed_hwfn *p_hwfn, DP_NOTICE(p_hwfn, "Unknown connection type: %d\n", conn_type); } - p_ramrod->gsi_offload_flag = p_ll2_conn->gsi_enable; + p_ramrod->gsi_offload_flag = p_ll2_conn->conn.gsi_enable; return qed_spq_post(p_hwfn, p_ent, NULL); } @@ -1224,7 +1220,7 @@ qed_ll2_acquire_connection_rx(struct qed_hwfn *p_hwfn, DP_VERBOSE(p_hwfn, QED_MSG_LL2, "Allocated LL2 Rxq [Type %08x] with 0x%08x buffers\n", - p_ll2_info->conn_type, rx_num_desc); + p_ll2_info->conn.conn_type, rx_num_desc); out: return rc; @@ -1262,7 +1258,7 @@ static int qed_ll2_acquire_connection_tx(struct qed_hwfn *p_hwfn, DP_VERBOSE(p_hwfn, QED_MSG_LL2, "Allocated LL2 Txq [Type %08x] with 0x%08x buffers\n", - p_ll2_info->conn_type, tx_num_desc); + p_ll2_info->conn.conn_type, tx_num_desc); out: if (rc) @@ -1273,7 +1269,7 @@ out: } int qed_ll2_acquire_connection(struct qed_hwfn *p_hwfn, - struct qed_ll2_info *p_params, + struct qed_ll2_conn *p_params, u16 rx_num_desc, u16 tx_num_desc, u8 *p_connection_handle) @@ -1302,15 +1298,7 @@ int qed_ll2_acquire_connection(struct qed_hwfn *p_hwfn, if (!p_ll2_info) return -EBUSY; - p_ll2_info->conn_type = p_params->conn_type; - p_ll2_info->mtu = p_params->mtu; - p_ll2_info->rx_drop_ttl0_flg = p_params->rx_drop_ttl0_flg; - p_ll2_info->rx_vlan_removal_en = p_params->rx_vlan_removal_en; - p_ll2_info->tx_tc = p_params->tx_tc; - p_ll2_info->tx_dest = p_params->tx_dest; - p_ll2_info->ai_err_packet_too_big = p_params->ai_err_packet_too_big; - p_ll2_info->ai_err_no_buf = p_params->ai_err_no_buf; - p_ll2_info->gsi_enable = p_params->gsi_enable; + p_ll2_info->conn = *p_params; rc = qed_ll2_acquire_connection_rx(p_hwfn, p_ll2_info, rx_num_desc); if (rc) @@ -1371,9 +1359,9 @@ static int qed_ll2_establish_connection_rx(struct qed_hwfn *p_hwfn, SET_FIELD(action_on_error, CORE_RX_ACTION_ON_ERROR_PACKET_TOO_BIG, - p_ll2_conn->ai_err_packet_too_big); + p_ll2_conn->conn.ai_err_packet_too_big); SET_FIELD(action_on_error, - CORE_RX_ACTION_ON_ERROR_NO_BUFF, p_ll2_conn->ai_err_no_buf); + CORE_RX_ACTION_ON_ERROR_NO_BUFF, p_ll2_conn->conn.ai_err_no_buf); return qed_sp_ll2_rx_queue_start(p_hwfn, p_ll2_conn, action_on_error); } @@ -1600,7 +1588,7 @@ static void qed_ll2_prepare_tx_packet_set_bd(struct qed_hwfn *p_hwfn, "LL2 [q 0x%02x cid 0x%08x type 0x%08x] Tx Producer at [0x%04x] - set with a %04x bytes %02x BDs buffer at %08x:%08x\n", p_ll2->queue_id, p_ll2->cid, - p_ll2->conn_type, + p_ll2->conn.conn_type, prod_idx, first_frag_len, num_of_bds, @@ -1676,7 +1664,7 @@ static void qed_ll2_tx_packet_notify(struct qed_hwfn *p_hwfn, (NETIF_MSG_TX_QUEUED | QED_MSG_LL2), "LL2 [q 0x%02x cid 0x%08x type 0x%08x] Doorbelled [producer 0x%04x]\n", p_ll2_conn->queue_id, - p_ll2_conn->cid, p_ll2_conn->conn_type, db_msg.spq_prod); + p_ll2_conn->cid, p_ll2_conn->conn.conn_type, db_msg.spq_prod); } int qed_ll2_prepare_tx_packet(struct qed_hwfn *p_hwfn, @@ -1817,7 +1805,7 @@ int qed_ll2_terminate_connection(struct qed_hwfn *p_hwfn, u8 connection_handle) qed_ll2_rxq_flush(p_hwfn, connection_handle); } - if (p_ll2_conn->conn_type == QED_LL2_TYPE_ISCSI_OOO) + if (p_ll2_conn->conn.conn_type == QED_LL2_TYPE_ISCSI_OOO) qed_ooo_release_all_isles(p_hwfn, p_hwfn->p_ooo_info); return rc; @@ -1993,7 +1981,7 @@ static void qed_ll2_register_cb_ops(struct qed_dev *cdev, static int qed_ll2_start(struct qed_dev *cdev, struct qed_ll2_params *params) { - struct qed_ll2_info ll2_info; + struct qed_ll2_conn ll2_info; struct qed_ll2_buffer *buffer, *tmp_buffer; enum qed_ll2_conn_type conn_type; struct qed_ptt *p_ptt; @@ -2041,6 +2029,7 @@ static int qed_ll2_start(struct qed_dev *cdev, struct qed_ll2_params *params) /* Prepare the temporary ll2 information */ memset(&ll2_info, 0, sizeof(ll2_info)); + ll2_info.conn_type = conn_type; ll2_info.mtu = params->mtu; ll2_info.rx_drop_ttl0_flg = params->drop_ttl0_packets; @@ -2120,7 +2109,6 @@ static int qed_ll2_start(struct qed_dev *cdev, struct qed_ll2_params *params) } ether_addr_copy(cdev->ll2_mac_address, params->ll2_mac_address); - return 0; release_terminate_all: diff --git a/drivers/net/ethernet/qlogic/qed/qed_ll2.h b/drivers/net/ethernet/qlogic/qed/qed_ll2.h index 6625a3ae5a33..31417928b635 100644 --- a/drivers/net/ethernet/qlogic/qed/qed_ll2.h +++ b/drivers/net/ethernet/qlogic/qed/qed_ll2.h @@ -112,15 +112,8 @@ struct qed_ll2_tx_queue { bool b_completing_packet; }; -struct qed_ll2_info { - /* Lock protecting the state of LL2 */ - struct mutex mutex; +struct qed_ll2_conn { enum qed_ll2_conn_type conn_type; - u32 cid; - u8 my_id; - u8 queue_id; - u8 tx_stats_id; - bool b_active; u16 mtu; u8 rx_drop_ttl0_flg; u8 rx_vlan_removal_en; @@ -128,10 +121,21 @@ struct qed_ll2_info { enum core_tx_dest tx_dest; enum core_error_handle ai_err_packet_too_big; enum core_error_handle ai_err_no_buf; + u8 gsi_enable; +}; + +struct qed_ll2_info { + /* Lock protecting the state of LL2 */ + struct mutex mutex; + struct qed_ll2_conn conn; + u32 cid; + u8 my_id; + u8 queue_id; + u8 tx_stats_id; + bool b_active; u8 tx_stats_en; struct qed_ll2_rx_queue rx_queue; struct qed_ll2_tx_queue tx_queue; - u8 gsi_enable; }; /** @@ -149,7 +153,7 @@ struct qed_ll2_info { * @return 0 on success, failure otherwise */ int qed_ll2_acquire_connection(struct qed_hwfn *p_hwfn, - struct qed_ll2_info *p_params, + struct qed_ll2_conn *p_params, u16 rx_num_desc, u16 tx_num_desc, u8 *p_connection_handle); diff --git a/drivers/net/ethernet/qlogic/qed/qed_roce.c b/drivers/net/ethernet/qlogic/qed/qed_roce.c index 2a16547c8966..2dbdb3298991 100644 --- a/drivers/net/ethernet/qlogic/qed/qed_roce.c +++ b/drivers/net/ethernet/qlogic/qed/qed_roce.c @@ -2632,7 +2632,7 @@ static int qed_roce_ll2_start(struct qed_dev *cdev, { struct qed_hwfn *hwfn = QED_LEADING_HWFN(cdev); struct qed_roce_ll2_info *roce_ll2; - struct qed_ll2_info ll2_params; + struct qed_ll2_conn ll2_params; int rc; if (!params) { From df384d435a5c034c735df3d9ea87a03172c59b56 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Wed, 18 Jan 2017 15:52:53 +0100 Subject: [PATCH 137/343] bcm63xx_enet: avoid uninitialized variable warning gcc-7 and probably earlier versions get confused by this function and print a harmless warning: drivers/net/ethernet/broadcom/bcm63xx_enet.c: In function 'bcm_enet_open': drivers/net/ethernet/broadcom/bcm63xx_enet.c:1130:3: error: 'phydev' may be used uninitialized in this function [-Werror=maybe-uninitialized] This adds an initialization for the 'phydev' variable when it is unused and changes the check to test for that NULL pointer to make it clear that we always pass a valid pointer here. Signed-off-by: Arnd Bergmann Signed-off-by: David S. Miller --- drivers/net/ethernet/broadcom/bcm63xx_enet.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/broadcom/bcm63xx_enet.c b/drivers/net/ethernet/broadcom/bcm63xx_enet.c index 3b14d5144228..c483618b57bd 100644 --- a/drivers/net/ethernet/broadcom/bcm63xx_enet.c +++ b/drivers/net/ethernet/broadcom/bcm63xx_enet.c @@ -913,6 +913,8 @@ static int bcm_enet_open(struct net_device *dev) priv->old_link = 0; priv->old_duplex = -1; priv->old_pause = -1; + } else { + phydev = NULL; } /* mask all interrupts and request them */ @@ -1083,7 +1085,7 @@ static int bcm_enet_open(struct net_device *dev) enet_dmac_writel(priv, priv->dma_chan_int_mask, ENETDMAC_IRMASK, priv->tx_chan); - if (priv->has_phy) + if (phydev) phy_start(phydev); else bcm_enet_adjust_link(dev); @@ -1126,7 +1128,7 @@ out_freeirq: free_irq(dev->irq, dev); out_phy_disconnect: - if (priv->has_phy) + if (phydev) phy_disconnect(phydev); return ret; From cdb749cef16bceb74950fc8668f2632ff7cac9aa Mon Sep 17 00:00:00 2001 From: Jesper Dangaard Brouer Date: Wed, 18 Jan 2017 17:19:00 +0100 Subject: [PATCH 138/343] bpf: fix samples xdp_tx_iptunnel and tc_l2_redirect with fake KBUILD_MODNAME Fix build errors for samples/bpf xdp_tx_iptunnel and tc_l2_redirect, when dynamic debugging is enabled (CONFIG_DYNAMIC_DEBUG) by defining a fake KBUILD_MODNAME. Just like Daniel Borkmann fixed other samples/bpf in commit 96a8eb1eeed2 ("bpf: fix samples to add fake KBUILD_MODNAME"). Fixes: 12d8bb64e3f6 ("bpf: xdp: Add XDP example for head adjustment") Fixes: 90e02896f1a4 ("bpf: Add test for bpf_redirect to ipip/ip6tnl") Signed-off-by: Jesper Dangaard Brouer Signed-off-by: David S. Miller --- samples/bpf/tc_l2_redirect_kern.c | 1 + samples/bpf/xdp_tx_iptunnel_kern.c | 1 + 2 files changed, 2 insertions(+) diff --git a/samples/bpf/tc_l2_redirect_kern.c b/samples/bpf/tc_l2_redirect_kern.c index 92a44729dbe4..7ef2a12b25b2 100644 --- a/samples/bpf/tc_l2_redirect_kern.c +++ b/samples/bpf/tc_l2_redirect_kern.c @@ -4,6 +4,7 @@ * modify it under the terms of version 2 of the GNU General Public * License as published by the Free Software Foundation. */ +#define KBUILD_MODNAME "foo" #include #include #include diff --git a/samples/bpf/xdp_tx_iptunnel_kern.c b/samples/bpf/xdp_tx_iptunnel_kern.c index 85c38ecd3a2d..0f4f6e8c8611 100644 --- a/samples/bpf/xdp_tx_iptunnel_kern.c +++ b/samples/bpf/xdp_tx_iptunnel_kern.c @@ -8,6 +8,7 @@ * encapsulating the incoming packet in an IPv4/v6 header * and then XDP_TX it out. */ +#define KBUILD_MODNAME "foo" #include #include #include From e048fc50d7bde23136e098e04a324d7e3404408d Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Wed, 18 Jan 2017 23:03:08 -0800 Subject: [PATCH 139/343] net/mlx5e: Do not recycle pages from emergency reserve A driver using dev_alloc_page() must not reuse a page allocated from emergency memory reserve. Otherwise all packets using this page will be immediately dropped, unless for very specific sockets having SOCK_MEMALLOC bit set. This issue might be hard to debug, because only a fraction of received packets would be dropped. Fixes: 4415a0319f92 ("net/mlx5e: Implement RX mapped page cache for page recycle") Signed-off-by: Eric Dumazet Cc: Tariq Toukan Cc: Saeed Mahameed Acked-by: Saeed Mahameed Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx5/core/en_rx.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c index 0e2fb3ed1790..06d5e6fecb0a 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c @@ -193,6 +193,9 @@ static inline bool mlx5e_rx_cache_put(struct mlx5e_rq *rq, return false; } + if (unlikely(page_is_pfmemalloc(dma_info->page))) + return false; + cache->page_cache[cache->tail] = *dma_info; cache->tail = tail_next; return true; From 90c311b0eeead647b708a723dbdde1eda3dcad05 Mon Sep 17 00:00:00 2001 From: Vineeth Remanan Pillai Date: Thu, 19 Jan 2017 08:35:39 -0800 Subject: [PATCH 140/343] xen-netfront: Fix Rx stall during network stress and OOM During an OOM scenario, request slots could not be created as skb allocation fails. So the netback cannot pass in packets and netfront wrongly assumes that there is no more work to be done and it disables polling. This causes Rx to stall. The issue is with the retry logic which schedules the timer if the created slots are less than NET_RX_SLOTS_MIN. The count of new request slots to be pushed are calculated as a difference between new req_prod and rsp_cons which could be more than the actual slots, if there are unconsumed responses. The fix is to calculate the count of newly created slots as the difference between new req_prod and old req_prod. Signed-off-by: Vineeth Remanan Pillai Reviewed-by: Juergen Gross Signed-off-by: David S. Miller --- drivers/net/xen-netfront.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/xen-netfront.c b/drivers/net/xen-netfront.c index a479cd99911d..8315fe73ecd0 100644 --- a/drivers/net/xen-netfront.c +++ b/drivers/net/xen-netfront.c @@ -321,7 +321,7 @@ static void xennet_alloc_rx_buffers(struct netfront_queue *queue) queue->rx.req_prod_pvt = req_prod; /* Not enough requests? Try again later. */ - if (req_prod - queue->rx.rsp_cons < NET_RX_SLOTS_MIN) { + if (req_prod - queue->rx.sring->req_prod < NET_RX_SLOTS_MIN) { mod_timer(&queue->rx_refill_timer, jiffies + (HZ/10)); return; } From b6677449dff674cf5b81429b11d5c7f358852ef9 Mon Sep 17 00:00:00 2001 From: Ivan Vecera Date: Fri, 20 Jan 2017 18:12:17 +0100 Subject: [PATCH 141/343] bridge: netlink: call br_changelink() during br_dev_newlink() Any bridge options specified during link creation (e.g. ip link add) are ignored as br_dev_newlink() does not process them. Use br_changelink() to do it. Fixes: 133235161721 ("bridge: implement rtnl_link_ops->changelink") Signed-off-by: Ivan Vecera Reviewed-by: Jiri Pirko Signed-off-by: David S. Miller --- net/bridge/br_netlink.c | 33 +++++++++++++++++++-------------- 1 file changed, 19 insertions(+), 14 deletions(-) diff --git a/net/bridge/br_netlink.c b/net/bridge/br_netlink.c index 71c7453268c1..7109b389ea58 100644 --- a/net/bridge/br_netlink.c +++ b/net/bridge/br_netlink.c @@ -781,20 +781,6 @@ static int br_validate(struct nlattr *tb[], struct nlattr *data[]) return 0; } -static int br_dev_newlink(struct net *src_net, struct net_device *dev, - struct nlattr *tb[], struct nlattr *data[]) -{ - struct net_bridge *br = netdev_priv(dev); - - if (tb[IFLA_ADDRESS]) { - spin_lock_bh(&br->lock); - br_stp_change_bridge_id(br, nla_data(tb[IFLA_ADDRESS])); - spin_unlock_bh(&br->lock); - } - - return register_netdevice(dev); -} - static int br_port_slave_changelink(struct net_device *brdev, struct net_device *dev, struct nlattr *tb[], @@ -1115,6 +1101,25 @@ static int br_changelink(struct net_device *brdev, struct nlattr *tb[], return 0; } +static int br_dev_newlink(struct net *src_net, struct net_device *dev, + struct nlattr *tb[], struct nlattr *data[]) +{ + struct net_bridge *br = netdev_priv(dev); + int err; + + if (tb[IFLA_ADDRESS]) { + spin_lock_bh(&br->lock); + br_stp_change_bridge_id(br, nla_data(tb[IFLA_ADDRESS])); + spin_unlock_bh(&br->lock); + } + + err = br_changelink(dev, tb, data); + if (err) + return err; + + return register_netdevice(dev); +} + static size_t br_get_size(const struct net_device *brdev) { return nla_total_size(sizeof(u32)) + /* IFLA_BR_FORWARD_DELAY */ From 63d762b88cb5510f2bfdb5112ced18cde867ae61 Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Sat, 7 Jan 2017 09:33:34 +0300 Subject: [PATCH 142/343] platform/x86: mlx-platform: free first dev on error There is an off-by-one error so we don't unregister priv->pdev_mux[0]. Also it's slightly simpler as a while loop instead of a for loop. Fixes: 58cbbee2391c ("x86/platform/mellanox: Introduce support for Mellanox systems platform") Signed-off-by: Dan Carpenter Acked-by: Vadim Pasternak Signed-off-by: Andy Shevchenko --- drivers/platform/x86/mlx-platform.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/platform/x86/mlx-platform.c b/drivers/platform/x86/mlx-platform.c index 97b4c3a219c0..25f15df5c2d7 100644 --- a/drivers/platform/x86/mlx-platform.c +++ b/drivers/platform/x86/mlx-platform.c @@ -326,7 +326,7 @@ static int __init mlxplat_init(void) return 0; fail_platform_mux_register: - for (i--; i > 0 ; i--) + while (--i >= 0) platform_device_unregister(priv->pdev_mux[i]); platform_device_unregister(priv->pdev_i2c); fail_alloc: From 44e6861646748a21b55725adc0780342f4440066 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Tue, 10 Jan 2017 16:28:47 +0100 Subject: [PATCH 143/343] platform/x86: surface3-wmi: Shut up unused-function warning The newly added driver guards its "resume" callback with an warning in some configurations: drivers/platform/x86/surface3-wmi.c:248:12: error: 's3_wmi_resume' defined but not used [-Werror=unused-function] Using a __maybe_unused annotation without an #ifdef avoids the mistake more reliably. Fixes: 3dda3b3798f9 ("platform/x86: Add custom surface3 platform device for controlling LID") Signed-off-by: Arnd Bergmann Reviewed-by: Benjamin Tissoires Signed-off-by: Darren Hart --- drivers/platform/x86/surface3-wmi.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/drivers/platform/x86/surface3-wmi.c b/drivers/platform/x86/surface3-wmi.c index cbf4d83a7271..6ac1d1349d7e 100644 --- a/drivers/platform/x86/surface3-wmi.c +++ b/drivers/platform/x86/surface3-wmi.c @@ -244,13 +244,11 @@ static int s3_wmi_remove(struct platform_device *device) return 0; } -#ifdef CONFIG_PM -static int s3_wmi_resume(struct device *dev) +static int __maybe_unused s3_wmi_resume(struct device *dev) { s3_wmi_send_lid_state(); return 0; } -#endif static SIMPLE_DEV_PM_OPS(s3_wmi_pm, NULL, s3_wmi_resume); static struct platform_driver s3_wmi_driver = { From e95ac4574b23a5fd8f5c5f2c19ef69ac15b7252c Mon Sep 17 00:00:00 2001 From: Benjamin Tissoires Date: Wed, 18 Jan 2017 09:13:46 +0100 Subject: [PATCH 144/343] platform/x86: surface3-wmi: fix uninitialized symbol The patch 3dda3b3798f9: "platform/x86: Add custom surface3 platform device for controlling LID" from Nov 25, 2016, leads to the following static checker warning: drivers/platform/x86/surface3-wmi.c:168 s3_wmi_check_platform_device() error: uninitialized symbol 'ts_adev'. Reported-by: Dan Carpenter Signed-off-by: Benjamin Tissoires Signed-off-by: Andy Shevchenko --- drivers/platform/x86/surface3-wmi.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/platform/x86/surface3-wmi.c b/drivers/platform/x86/surface3-wmi.c index 6ac1d1349d7e..25b176996cb7 100644 --- a/drivers/platform/x86/surface3-wmi.c +++ b/drivers/platform/x86/surface3-wmi.c @@ -139,7 +139,7 @@ static acpi_status s3_wmi_attach_spi_device(acpi_handle handle, static int s3_wmi_check_platform_device(struct device *dev, void *data) { - struct acpi_device *adev, *ts_adev; + struct acpi_device *adev, *ts_adev = NULL; acpi_handle handle; acpi_status status; From 5a00b6c2438460b870a451f14593fc40d3c7edf6 Mon Sep 17 00:00:00 2001 From: Andy Shevchenko Date: Thu, 19 Jan 2017 18:39:40 +0200 Subject: [PATCH 145/343] platform/x86: intel_mid_powerbtn: Set IRQ_ONESHOT The commit 1c6c69525b40 ("genirq: Reject bogus threaded irq requests") starts refusing misconfigured interrupt handlers. This makes intel_mid_powerbtn not working anymore. Add a mandatory flag to a threaded IRQ request in the driver. Fixes: 1c6c69525b40 ("genirq: Reject bogus threaded irq requests") Signed-off-by: Andy Shevchenko --- drivers/platform/x86/intel_mid_powerbtn.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/platform/x86/intel_mid_powerbtn.c b/drivers/platform/x86/intel_mid_powerbtn.c index 1fc0de870ff8..361770568ad0 100644 --- a/drivers/platform/x86/intel_mid_powerbtn.c +++ b/drivers/platform/x86/intel_mid_powerbtn.c @@ -77,7 +77,7 @@ static int mfld_pb_probe(struct platform_device *pdev) input_set_capability(input, EV_KEY, KEY_POWER); - error = request_threaded_irq(irq, NULL, mfld_pb_isr, 0, + error = request_threaded_irq(irq, NULL, mfld_pb_isr, IRQF_ONESHOT, DRIVER_NAME, input); if (error) { dev_err(&pdev->dev, "Unable to request irq %d for mfld power" From cfee5d63767b2e7997c1f36420d008abbe61565c Mon Sep 17 00:00:00 2001 From: Zach Ploskey Date: Sun, 22 Jan 2017 00:47:19 -0800 Subject: [PATCH 146/343] platform/x86: ideapad-laptop: handle ACPI event 1 On Ideapad laptops, ACPI event 1 is currently not handled. Many models log "ideapad_laptop: Unknown event: 1" every 20 seconds or so while running on battery power. Some convertible laptops receive this event when switching in and out of tablet mode. This adds and additional case for event 1 in ideapad_acpi_notify to call ideapad_input_report(priv, vpc_bit), so that the event is reported to userspace and we avoid unnecessary logging. Fixes bug #107481 (https://bugzilla.kernel.org/show_bug.cgi?id=107481) Fixes bug #65751 (https://bugzilla.kernel.org/show_bug.cgi?id=65751) Signed-off-by: Zach Ploskey Signed-off-by: Andy Shevchenko --- drivers/platform/x86/ideapad-laptop.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/platform/x86/ideapad-laptop.c b/drivers/platform/x86/ideapad-laptop.c index 410741acb3c9..f46ece2ce3c4 100644 --- a/drivers/platform/x86/ideapad-laptop.c +++ b/drivers/platform/x86/ideapad-laptop.c @@ -813,6 +813,7 @@ static void ideapad_acpi_notify(acpi_handle handle, u32 event, void *data) case 8: case 7: case 6: + case 1: ideapad_input_report(priv, vpc_bit); break; case 5: From 4eccbfc36186926b570310bfbd44f4216cd05c63 Mon Sep 17 00:00:00 2001 From: "Lendacky, Thomas" Date: Fri, 20 Jan 2017 12:14:03 -0600 Subject: [PATCH 147/343] amd-xgbe: Add a hardware quirk for register definitions A newer version of the hardware is using the same PCI ids for the network device but has altered register definitions for determining the window settings for the indirect PCS access. Add support to check for this hardware and if found use the new register values. Signed-off-by: Tom Lendacky Signed-off-by: David S. Miller --- drivers/net/ethernet/amd/xgbe/xgbe-common.h | 2 ++ drivers/net/ethernet/amd/xgbe/xgbe-dev.c | 4 ++-- drivers/net/ethernet/amd/xgbe/xgbe-pci.c | 15 ++++++++++++++- drivers/net/ethernet/amd/xgbe/xgbe.h | 2 ++ 4 files changed, 20 insertions(+), 3 deletions(-) diff --git a/drivers/net/ethernet/amd/xgbe/xgbe-common.h b/drivers/net/ethernet/amd/xgbe/xgbe-common.h index 5b7ba25e0065..8a280e7d66bd 100644 --- a/drivers/net/ethernet/amd/xgbe/xgbe-common.h +++ b/drivers/net/ethernet/amd/xgbe/xgbe-common.h @@ -891,6 +891,8 @@ #define PCS_V1_WINDOW_SELECT 0x03fc #define PCS_V2_WINDOW_DEF 0x9060 #define PCS_V2_WINDOW_SELECT 0x9064 +#define PCS_V2_RV_WINDOW_DEF 0x1060 +#define PCS_V2_RV_WINDOW_SELECT 0x1064 /* PCS register entry bit positions and sizes */ #define PCS_V2_WINDOW_DEF_OFFSET_INDEX 6 diff --git a/drivers/net/ethernet/amd/xgbe/xgbe-dev.c b/drivers/net/ethernet/amd/xgbe/xgbe-dev.c index aaf0350076a9..c8e8a4a271ec 100644 --- a/drivers/net/ethernet/amd/xgbe/xgbe-dev.c +++ b/drivers/net/ethernet/amd/xgbe/xgbe-dev.c @@ -1151,7 +1151,7 @@ static int xgbe_read_mmd_regs_v2(struct xgbe_prv_data *pdata, int prtad, offset = pdata->xpcs_window + (mmd_address & pdata->xpcs_window_mask); spin_lock_irqsave(&pdata->xpcs_lock, flags); - XPCS32_IOWRITE(pdata, PCS_V2_WINDOW_SELECT, index); + XPCS32_IOWRITE(pdata, pdata->xpcs_window_sel_reg, index); mmd_data = XPCS16_IOREAD(pdata, offset); spin_unlock_irqrestore(&pdata->xpcs_lock, flags); @@ -1183,7 +1183,7 @@ static void xgbe_write_mmd_regs_v2(struct xgbe_prv_data *pdata, int prtad, offset = pdata->xpcs_window + (mmd_address & pdata->xpcs_window_mask); spin_lock_irqsave(&pdata->xpcs_lock, flags); - XPCS32_IOWRITE(pdata, PCS_V2_WINDOW_SELECT, index); + XPCS32_IOWRITE(pdata, pdata->xpcs_window_sel_reg, index); XPCS16_IOWRITE(pdata, offset, mmd_data); spin_unlock_irqrestore(&pdata->xpcs_lock, flags); } diff --git a/drivers/net/ethernet/amd/xgbe/xgbe-pci.c b/drivers/net/ethernet/amd/xgbe/xgbe-pci.c index e76b7f65b805..c2730f15bd8b 100644 --- a/drivers/net/ethernet/amd/xgbe/xgbe-pci.c +++ b/drivers/net/ethernet/amd/xgbe/xgbe-pci.c @@ -265,6 +265,7 @@ static int xgbe_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id) struct xgbe_prv_data *pdata; struct device *dev = &pdev->dev; void __iomem * const *iomap_table; + struct pci_dev *rdev; unsigned int ma_lo, ma_hi; unsigned int reg; int bar_mask; @@ -326,8 +327,20 @@ static int xgbe_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id) if (netif_msg_probe(pdata)) dev_dbg(dev, "xpcs_regs = %p\n", pdata->xpcs_regs); + /* Set the PCS indirect addressing definition registers */ + rdev = pci_get_domain_bus_and_slot(0, 0, PCI_DEVFN(0, 0)); + if (rdev && + (rdev->vendor == PCI_VENDOR_ID_AMD) && (rdev->device == 0x15d0)) { + pdata->xpcs_window_def_reg = PCS_V2_RV_WINDOW_DEF; + pdata->xpcs_window_sel_reg = PCS_V2_RV_WINDOW_SELECT; + } else { + pdata->xpcs_window_def_reg = PCS_V2_WINDOW_DEF; + pdata->xpcs_window_sel_reg = PCS_V2_WINDOW_SELECT; + } + pci_dev_put(rdev); + /* Configure the PCS indirect addressing support */ - reg = XPCS32_IOREAD(pdata, PCS_V2_WINDOW_DEF); + reg = XPCS32_IOREAD(pdata, pdata->xpcs_window_def_reg); pdata->xpcs_window = XPCS_GET_BITS(reg, PCS_V2_WINDOW_DEF, OFFSET); pdata->xpcs_window <<= 6; pdata->xpcs_window_size = XPCS_GET_BITS(reg, PCS_V2_WINDOW_DEF, SIZE); diff --git a/drivers/net/ethernet/amd/xgbe/xgbe.h b/drivers/net/ethernet/amd/xgbe/xgbe.h index f52a9bd05bac..00108815b55e 100644 --- a/drivers/net/ethernet/amd/xgbe/xgbe.h +++ b/drivers/net/ethernet/amd/xgbe/xgbe.h @@ -955,6 +955,8 @@ struct xgbe_prv_data { /* XPCS indirect addressing lock */ spinlock_t xpcs_lock; + unsigned int xpcs_window_def_reg; + unsigned int xpcs_window_sel_reg; unsigned int xpcs_window; unsigned int xpcs_window_size; unsigned int xpcs_window_mask; From 738f7f647371ff4cfc9646c99dba5b58ad142db3 Mon Sep 17 00:00:00 2001 From: "Lendacky, Thomas" Date: Fri, 20 Jan 2017 12:14:13 -0600 Subject: [PATCH 148/343] amd-xgbe: Check xgbe_init() return code The xgbe_init() routine returns a return code indicating success or failure, but the return code is not checked. Add code to xgbe_init() to issue a message when failures are seen and add code to check the xgbe_init() return code. Signed-off-by: Tom Lendacky Signed-off-by: David S. Miller --- drivers/net/ethernet/amd/xgbe/xgbe-dev.c | 4 +++- drivers/net/ethernet/amd/xgbe/xgbe-drv.c | 4 +++- 2 files changed, 6 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/amd/xgbe/xgbe-dev.c b/drivers/net/ethernet/amd/xgbe/xgbe-dev.c index c8e8a4a271ec..a7d16db5c4b2 100644 --- a/drivers/net/ethernet/amd/xgbe/xgbe-dev.c +++ b/drivers/net/ethernet/amd/xgbe/xgbe-dev.c @@ -3407,8 +3407,10 @@ static int xgbe_init(struct xgbe_prv_data *pdata) /* Flush Tx queues */ ret = xgbe_flush_tx_queues(pdata); - if (ret) + if (ret) { + netdev_err(pdata->netdev, "error flushing TX queues\n"); return ret; + } /* * Initialize DMA related features diff --git a/drivers/net/ethernet/amd/xgbe/xgbe-drv.c b/drivers/net/ethernet/amd/xgbe/xgbe-drv.c index 9943629fcbf9..1c87cc204075 100644 --- a/drivers/net/ethernet/amd/xgbe/xgbe-drv.c +++ b/drivers/net/ethernet/amd/xgbe/xgbe-drv.c @@ -1070,7 +1070,9 @@ static int xgbe_start(struct xgbe_prv_data *pdata) DBGPR("-->xgbe_start\n"); - hw_if->init(pdata); + ret = hw_if->init(pdata); + if (ret) + return ret; xgbe_napi_enable(pdata, 1); From e9748e0364fe82dc037d22900ff13a62d04518bf Mon Sep 17 00:00:00 2001 From: Ziyuan Xu Date: Tue, 17 Jan 2017 09:22:56 +0800 Subject: [PATCH 149/343] mmc: dw_mmc: force setup bus if active slots exist It's necessary to setup bus if any slots are present. - update clock after ctrl reset - if the host has genpd node, we can guarantee the clock is available before starting request. Otherwies, the clock register is reset once power off the pd, and host can't output the active clock during communication. Fixes: e9ed8835e990 ("mmc: dw_mmc: add runtime PM callback") Fixes: df9bcc2bc0a1 ("mmc: dw_mmc: add missing codes for runtime resume") cc: Reported-by: Randy Li Reported-by: S. Gilles Signed-off-by: Ziyuan Xu Signed-off-by: Shawn Lin Signed-off-by: Ulf Hansson --- drivers/mmc/host/dw_mmc.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/drivers/mmc/host/dw_mmc.c b/drivers/mmc/host/dw_mmc.c index b44306b886cb..73db08558e4d 100644 --- a/drivers/mmc/host/dw_mmc.c +++ b/drivers/mmc/host/dw_mmc.c @@ -3354,10 +3354,11 @@ int dw_mci_runtime_resume(struct device *dev) if (!slot) continue; - if (slot->mmc->pm_flags & MMC_PM_KEEP_POWER) { + if (slot->mmc->pm_flags & MMC_PM_KEEP_POWER) dw_mci_set_ios(slot->mmc, &slot->mmc->ios); - dw_mci_setup_bus(slot, true); - } + + /* Force setup bus to guarantee available clock output */ + dw_mci_setup_bus(slot, true); } /* Now that slots are all setup, we can enable card detect */ From 880a38547ff08715ce4f1daf9a4bb30c87676e68 Mon Sep 17 00:00:00 2001 From: Nikolay Borisov Date: Fri, 20 Jan 2017 15:21:35 +0200 Subject: [PATCH 150/343] userns: Make ucounts lock irq-safe The ucounts_lock is being used to protect various ucounts lifecycle management functionalities. However, those services can also be invoked when a pidns is being freed in an RCU callback (e.g. softirq context). This can lead to deadlocks. There were already efforts trying to prevent similar deadlocks in add7c65ca426 ("pid: fix lockdep deadlock warning due to ucount_lock"), however they just moved the context from hardirq to softrq. Fix this issue once and for all by explictly making the lock disable irqs altogether. Dmitry Vyukov reported: > I've got the following deadlock report while running syzkaller fuzzer > on eec0d3d065bfcdf9cd5f56dd2a36b94d12d32297 of linux-next (on odroid > device if it matters): > > ================================= > [ INFO: inconsistent lock state ] > 4.10.0-rc3-next-20170112-xc2-dirty #6 Not tainted > --------------------------------- > inconsistent {SOFTIRQ-ON-W} -> {IN-SOFTIRQ-W} usage. > swapper/2/0 [HC0[0]:SC1[1]:HE1:SE0] takes: > (ucounts_lock){+.?...}, at: [< inline >] spin_lock > ./include/linux/spinlock.h:302 > (ucounts_lock){+.?...}, at: [] > put_ucounts+0x60/0x138 kernel/ucount.c:162 > {SOFTIRQ-ON-W} state was registered at: > [] mark_lock+0x220/0xb60 kernel/locking/lockdep.c:3054 > [< inline >] mark_irqflags kernel/locking/lockdep.c:2941 > [] __lock_acquire+0x388/0x3260 kernel/locking/lockdep.c:3295 > [] lock_acquire+0xa4/0x138 kernel/locking/lockdep.c:3753 > [< inline >] __raw_spin_lock ./include/linux/spinlock_api_smp.h:144 > [] _raw_spin_lock+0x90/0xd0 kernel/locking/spinlock.c:151 > [< inline >] spin_lock ./include/linux/spinlock.h:302 > [< inline >] get_ucounts kernel/ucount.c:131 > [] inc_ucount+0x80/0x6c8 kernel/ucount.c:189 > [< inline >] inc_mnt_namespaces fs/namespace.c:2818 > [] alloc_mnt_ns+0x78/0x3a8 fs/namespace.c:2849 > [] create_mnt_ns+0x28/0x200 fs/namespace.c:2959 > [< inline >] init_mount_tree fs/namespace.c:3199 > [] mnt_init+0x258/0x384 fs/namespace.c:3251 > [] vfs_caches_init+0x6c/0x80 fs/dcache.c:3626 > [] start_kernel+0x414/0x460 init/main.c:648 > [] __primary_switched+0x6c/0x70 arch/arm64/kernel/head.S:456 > irq event stamp: 2316924 > hardirqs last enabled at (2316924): [< inline >] rcu_do_batch > kernel/rcu/tree.c:2911 > hardirqs last enabled at (2316924): [< inline >] > invoke_rcu_callbacks kernel/rcu/tree.c:3182 > hardirqs last enabled at (2316924): [< inline >] > __rcu_process_callbacks kernel/rcu/tree.c:3149 > hardirqs last enabled at (2316924): [] > rcu_process_callbacks+0x7a4/0xc28 kernel/rcu/tree.c:3166 > hardirqs last disabled at (2316923): [< inline >] rcu_do_batch > kernel/rcu/tree.c:2900 > hardirqs last disabled at (2316923): [< inline >] > invoke_rcu_callbacks kernel/rcu/tree.c:3182 > hardirqs last disabled at (2316923): [< inline >] > __rcu_process_callbacks kernel/rcu/tree.c:3149 > hardirqs last disabled at (2316923): [] > rcu_process_callbacks+0x210/0xc28 kernel/rcu/tree.c:3166 > softirqs last enabled at (2316912): [] > _local_bh_enable+0x4c/0x80 kernel/softirq.c:155 > softirqs last disabled at (2316913): [< inline >] > do_softirq_own_stack ./include/linux/interrupt.h:488 > softirqs last disabled at (2316913): [< inline >] > invoke_softirq kernel/softirq.c:371 > softirqs last disabled at (2316913): [] > irq_exit+0x264/0x308 kernel/softirq.c:405 > > other info that might help us debug this: > Possible unsafe locking scenario: > > CPU0 > ---- > lock(ucounts_lock); > > lock(ucounts_lock); > > *** DEADLOCK *** > > 1 lock held by swapper/2/0: > #0: (rcu_callback){......}, at: [< inline >] __rcu_reclaim > kernel/rcu/rcu.h:108 > #0: (rcu_callback){......}, at: [< inline >] rcu_do_batch > kernel/rcu/tree.c:2919 > #0: (rcu_callback){......}, at: [< inline >] > invoke_rcu_callbacks kernel/rcu/tree.c:3182 > #0: (rcu_callback){......}, at: [< inline >] > __rcu_process_callbacks kernel/rcu/tree.c:3149 > #0: (rcu_callback){......}, at: [] > rcu_process_callbacks+0x720/0xc28 kernel/rcu/tree.c:3166 > > stack backtrace: > CPU: 2 PID: 0 Comm: swapper/2 Not tainted 4.10.0-rc3-next-20170112-xc2-dirty #6 > Hardware name: Hardkernel ODROID-C2 (DT) > Call trace: > [] dump_backtrace+0x0/0x440 arch/arm64/kernel/traps.c:500 > [] show_stack+0x20/0x30 arch/arm64/kernel/traps.c:225 > [] dump_stack+0x110/0x168 > [] print_usage_bug.part.27+0x49c/0x4bc > kernel/locking/lockdep.c:2387 > [< inline >] print_usage_bug kernel/locking/lockdep.c:2357 > [< inline >] valid_state kernel/locking/lockdep.c:2400 > [< inline >] mark_lock_irq kernel/locking/lockdep.c:2617 > [] mark_lock+0x934/0xb60 kernel/locking/lockdep.c:3065 > [< inline >] mark_irqflags kernel/locking/lockdep.c:2923 > [] __lock_acquire+0x640/0x3260 kernel/locking/lockdep.c:3295 > [] lock_acquire+0xa4/0x138 kernel/locking/lockdep.c:3753 > [< inline >] __raw_spin_lock ./include/linux/spinlock_api_smp.h:144 > [] _raw_spin_lock+0x90/0xd0 kernel/locking/spinlock.c:151 > [< inline >] spin_lock ./include/linux/spinlock.h:302 > [] put_ucounts+0x60/0x138 kernel/ucount.c:162 > [] dec_ucount+0xf4/0x158 kernel/ucount.c:214 > [< inline >] dec_pid_namespaces kernel/pid_namespace.c:89 > [] delayed_free_pidns+0x40/0xe0 kernel/pid_namespace.c:156 > [< inline >] __rcu_reclaim kernel/rcu/rcu.h:118 > [< inline >] rcu_do_batch kernel/rcu/tree.c:2919 > [< inline >] invoke_rcu_callbacks kernel/rcu/tree.c:3182 > [< inline >] __rcu_process_callbacks kernel/rcu/tree.c:3149 > [] rcu_process_callbacks+0x768/0xc28 kernel/rcu/tree.c:3166 > [] __do_softirq+0x324/0x6e0 kernel/softirq.c:284 > [< inline >] do_softirq_own_stack ./include/linux/interrupt.h:488 > [< inline >] invoke_softirq kernel/softirq.c:371 > [] irq_exit+0x264/0x308 kernel/softirq.c:405 > [] __handle_domain_irq+0xc0/0x150 kernel/irq/irqdesc.c:636 > [] gic_handle_irq+0x68/0xd8 > Exception stack(0xffff8000648e7dd0 to 0xffff8000648e7f00) > 7dc0: ffff8000648d4b3c 0000000000000007 > 7de0: 0000000000000000 1ffff0000c91a967 1ffff0000c91a967 1ffff0000c91a967 > 7e00: ffff20000a4b6b68 0000000000000001 0000000000000007 0000000000000001 > 7e20: 1fffe4000149ae90 ffff200009d35000 0000000000000000 0000000000000002 > 7e40: 0000000000000000 0000000000000000 0000000002624a1a 0000000000000000 > 7e60: 0000000000000000 ffff200009cbcd88 000060006d2ed000 0000000000000140 > 7e80: ffff200009cff000 ffff200009cb6000 ffff200009cc2020 ffff200009d2159d > 7ea0: 0000000000000000 ffff8000648d4380 0000000000000000 ffff8000648e7f00 > 7ec0: ffff20000820a478 ffff8000648e7f00 ffff20000820a47c 0000000010000145 > 7ee0: 0000000000000140 dfff200000000000 ffffffffffffffff ffff20000820a478 > [] el1_irq+0xb8/0x130 arch/arm64/kernel/entry.S:486 > [< inline >] arch_local_irq_restore > ./arch/arm64/include/asm/irqflags.h:81 > [] rcu_idle_exit+0x64/0xa8 kernel/rcu/tree.c:1030 > [< inline >] cpuidle_idle_call kernel/sched/idle.c:200 > [] do_idle+0x1dc/0x2d0 kernel/sched/idle.c:243 > [] cpu_startup_entry+0x24/0x28 kernel/sched/idle.c:345 > [] secondary_start_kernel+0x2cc/0x358 > arch/arm64/kernel/smp.c:276 > [<000000000279f1a4>] 0x279f1a4 Reported-by: Dmitry Vyukov Tested-by: Dmitry Vyukov Fixes: add7c65ca426 ("pid: fix lockdep deadlock warning due to ucount_lock") Fixes: f333c700c610 ("pidns: Add a limit on the number of pid namespaces") Cc: stable@vger.kernel.org Link: https://www.spinics.net/lists/kernel/msg2426637.html Signed-off-by: Nikolay Borisov Signed-off-by: Eric W. Biederman --- kernel/ucount.c | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-) diff --git a/kernel/ucount.c b/kernel/ucount.c index 9d20d5dd298a..4bbd38ec3788 100644 --- a/kernel/ucount.c +++ b/kernel/ucount.c @@ -128,10 +128,10 @@ static struct ucounts *get_ucounts(struct user_namespace *ns, kuid_t uid) struct hlist_head *hashent = ucounts_hashentry(ns, uid); struct ucounts *ucounts, *new; - spin_lock(&ucounts_lock); + spin_lock_irq(&ucounts_lock); ucounts = find_ucounts(ns, uid, hashent); if (!ucounts) { - spin_unlock(&ucounts_lock); + spin_unlock_irq(&ucounts_lock); new = kzalloc(sizeof(*new), GFP_KERNEL); if (!new) @@ -141,7 +141,7 @@ static struct ucounts *get_ucounts(struct user_namespace *ns, kuid_t uid) new->uid = uid; atomic_set(&new->count, 0); - spin_lock(&ucounts_lock); + spin_lock_irq(&ucounts_lock); ucounts = find_ucounts(ns, uid, hashent); if (ucounts) { kfree(new); @@ -152,16 +152,18 @@ static struct ucounts *get_ucounts(struct user_namespace *ns, kuid_t uid) } if (!atomic_add_unless(&ucounts->count, 1, INT_MAX)) ucounts = NULL; - spin_unlock(&ucounts_lock); + spin_unlock_irq(&ucounts_lock); return ucounts; } static void put_ucounts(struct ucounts *ucounts) { + unsigned long flags; + if (atomic_dec_and_test(&ucounts->count)) { - spin_lock(&ucounts_lock); + spin_lock_irqsave(&ucounts_lock, flags); hlist_del_init(&ucounts->node); - spin_unlock(&ucounts_lock); + spin_unlock_irqrestore(&ucounts_lock, flags); kfree(ucounts); } From 9f427a0e474a67b454420c131709600d44850486 Mon Sep 17 00:00:00 2001 From: David Ahern Date: Fri, 20 Jan 2017 12:58:34 -0800 Subject: [PATCH 151/343] net: mpls: Fix multipath selection for LSR use case MPLS multipath for LSR is broken -- always selecting the first nexthop in the one label case. For example: $ ip -f mpls ro ls 100 nexthop as to 200 via inet 172.16.2.2 dev virt12 nexthop as to 300 via inet 172.16.3.2 dev virt13 101 nexthop as to 201 via inet6 2000:2::2 dev virt12 nexthop as to 301 via inet6 2000:3::2 dev virt13 In this example incoming packets have a single MPLS labels which means BOS bit is set. The BOS bit is passed from mpls_forward down to mpls_multipath_hash which never processes the hash loop because BOS is 1. Update mpls_multipath_hash to process the entire label stack. mpls_hdr_len tracks the total mpls header length on each pass (on pass N mpls_hdr_len is N * sizeof(mpls_shim_hdr)). When the label is found with the BOS set it verifies the skb has sufficient header for ipv4 or ipv6, and find the IPv4 and IPv6 header by using the last mpls_hdr pointer and adding 1 to advance past it. With these changes I have verified the code correctly sees the label, BOS, IPv4 and IPv6 addresses in the network header and icmp/tcp/udp traffic for ipv4 and ipv6 are distributed across the nexthops. Fixes: 1c78efa8319ca ("mpls: flow-based multipath selection") Acked-by: Robert Shearman Signed-off-by: David Ahern Signed-off-by: David S. Miller --- net/mpls/af_mpls.c | 48 ++++++++++++++++++++++++---------------------- 1 file changed, 25 insertions(+), 23 deletions(-) diff --git a/net/mpls/af_mpls.c b/net/mpls/af_mpls.c index 15fe97644ffe..5b77377e5a15 100644 --- a/net/mpls/af_mpls.c +++ b/net/mpls/af_mpls.c @@ -98,18 +98,19 @@ bool mpls_pkt_too_big(const struct sk_buff *skb, unsigned int mtu) } EXPORT_SYMBOL_GPL(mpls_pkt_too_big); -static u32 mpls_multipath_hash(struct mpls_route *rt, - struct sk_buff *skb, bool bos) +static u32 mpls_multipath_hash(struct mpls_route *rt, struct sk_buff *skb) { struct mpls_entry_decoded dec; + unsigned int mpls_hdr_len = 0; struct mpls_shim_hdr *hdr; bool eli_seen = false; int label_index; u32 hash = 0; - for (label_index = 0; label_index < MAX_MP_SELECT_LABELS && !bos; + for (label_index = 0; label_index < MAX_MP_SELECT_LABELS; label_index++) { - if (!pskb_may_pull(skb, sizeof(*hdr) * label_index)) + mpls_hdr_len += sizeof(*hdr); + if (!pskb_may_pull(skb, mpls_hdr_len)) break; /* Read and decode the current label */ @@ -134,37 +135,38 @@ static u32 mpls_multipath_hash(struct mpls_route *rt, eli_seen = true; } - bos = dec.bos; - if (bos && pskb_may_pull(skb, sizeof(*hdr) * label_index + - sizeof(struct iphdr))) { + if (!dec.bos) + continue; + + /* found bottom label; does skb have room for a header? */ + if (pskb_may_pull(skb, mpls_hdr_len + sizeof(struct iphdr))) { const struct iphdr *v4hdr; - v4hdr = (const struct iphdr *)(mpls_hdr(skb) + - label_index); + v4hdr = (const struct iphdr *)(hdr + 1); if (v4hdr->version == 4) { hash = jhash_3words(ntohl(v4hdr->saddr), ntohl(v4hdr->daddr), v4hdr->protocol, hash); } else if (v4hdr->version == 6 && - pskb_may_pull(skb, sizeof(*hdr) * label_index + - sizeof(struct ipv6hdr))) { + pskb_may_pull(skb, mpls_hdr_len + + sizeof(struct ipv6hdr))) { const struct ipv6hdr *v6hdr; - v6hdr = (const struct ipv6hdr *)(mpls_hdr(skb) + - label_index); - + v6hdr = (const struct ipv6hdr *)(hdr + 1); hash = __ipv6_addr_jhash(&v6hdr->saddr, hash); hash = __ipv6_addr_jhash(&v6hdr->daddr, hash); hash = jhash_1word(v6hdr->nexthdr, hash); } } + + break; } return hash; } static struct mpls_nh *mpls_select_multipath(struct mpls_route *rt, - struct sk_buff *skb, bool bos) + struct sk_buff *skb) { int alive = ACCESS_ONCE(rt->rt_nhn_alive); u32 hash = 0; @@ -180,7 +182,7 @@ static struct mpls_nh *mpls_select_multipath(struct mpls_route *rt, if (alive <= 0) return NULL; - hash = mpls_multipath_hash(rt, skb, bos); + hash = mpls_multipath_hash(rt, skb); nh_index = hash % alive; if (alive == rt->rt_nhn) goto out; @@ -278,17 +280,11 @@ static int mpls_forward(struct sk_buff *skb, struct net_device *dev, hdr = mpls_hdr(skb); dec = mpls_entry_decode(hdr); - /* Pop the label */ - skb_pull(skb, sizeof(*hdr)); - skb_reset_network_header(skb); - - skb_orphan(skb); - rt = mpls_route_input_rcu(net, dec.label); if (!rt) goto drop; - nh = mpls_select_multipath(rt, skb, dec.bos); + nh = mpls_select_multipath(rt, skb); if (!nh) goto drop; @@ -297,6 +293,12 @@ static int mpls_forward(struct sk_buff *skb, struct net_device *dev, if (!mpls_output_possible(out_dev)) goto drop; + /* Pop the label */ + skb_pull(skb, sizeof(*hdr)); + skb_reset_network_header(skb); + + skb_orphan(skb); + if (skb_warn_if_lro(skb)) goto drop; From b32728ffef7f233dbdabb3f11814bdf692aaf501 Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Mon, 23 Jan 2017 08:12:19 -0700 Subject: [PATCH 152/343] xen-blkfront: feature flags handling adjustments Don't truncate the "feature-persistent" value read from xenstore: Any non-zero value is supposed to enable the feature, just like is already being done for feature_secdiscard. Just like the other feature_* fields, feature_flush and feature_fua are boolean flags, and hence fit well into a single bit. Keep all bit fields together to limit gaps. Signed-off-by: Jan Beulich Signed-off-by: Konrad Rzeszutek Wilk --- drivers/block/xen-blkfront.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/drivers/block/xen-blkfront.c b/drivers/block/xen-blkfront.c index b2bdfa81f929..386b17208605 100644 --- a/drivers/block/xen-blkfront.c +++ b/drivers/block/xen-blkfront.c @@ -197,13 +197,13 @@ struct blkfront_info /* Number of pages per ring buffer. */ unsigned int nr_ring_pages; struct request_queue *rq; - unsigned int feature_flush; - unsigned int feature_fua; + unsigned int feature_flush:1; + unsigned int feature_fua:1; unsigned int feature_discard:1; unsigned int feature_secdiscard:1; + unsigned int feature_persistent:1; unsigned int discard_granularity; unsigned int discard_alignment; - unsigned int feature_persistent:1; /* Number of 4KB segments handled */ unsigned int max_indirect_segments; int is_ready; @@ -2323,8 +2323,8 @@ static void blkfront_gather_backend_features(struct blkfront_info *info) blkfront_setup_discard(info); info->feature_persistent = - xenbus_read_unsigned(info->xbdev->otherend, - "feature-persistent", 0); + !!xenbus_read_unsigned(info->xbdev->otherend, + "feature-persistent", 0); indirect_segments = xenbus_read_unsigned(info->xbdev->otherend, "feature-max-indirect-segments", 0); From 3b4f18843e511193e7eb616710e838f5852e661d Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Mon, 23 Jan 2017 08:11:37 -0700 Subject: [PATCH 153/343] xen-blkfront: correct maximum segment accounting Making use of "max_indirect_segments=" has issues: - blkfront_setup_indirect() may end up with zero psegs when PAGE_SIZE is sufficiently much larger than XEN_PAGE_SIZE - the variable driven by the command line option (xen_blkif_max_segments) has a somewhat different purpose, and hence should namely never end up being zero - as long as the specified value is lower than the legacy default, we better don't use indirect segments at all (or we'd in fact lower throughput) Signed-off-by: Jan Beulich Signed-off-by: Konrad Rzeszutek Wilk --- drivers/block/xen-blkfront.c | 12 +++++++++--- 1 file changed, 9 insertions(+), 3 deletions(-) diff --git a/drivers/block/xen-blkfront.c b/drivers/block/xen-blkfront.c index 386b17208605..265f1a7072e9 100644 --- a/drivers/block/xen-blkfront.c +++ b/drivers/block/xen-blkfront.c @@ -2223,7 +2223,7 @@ static int blkfront_setup_indirect(struct blkfront_ring_info *rinfo) } else grants = info->max_indirect_segments; - psegs = grants / GRANTS_PER_PSEG; + psegs = DIV_ROUND_UP(grants, GRANTS_PER_PSEG); err = fill_grant_buffer(rinfo, (grants + INDIRECT_GREFS(grants)) * BLK_RING_SIZE(info)); @@ -2328,8 +2328,11 @@ static void blkfront_gather_backend_features(struct blkfront_info *info) indirect_segments = xenbus_read_unsigned(info->xbdev->otherend, "feature-max-indirect-segments", 0); - info->max_indirect_segments = min(indirect_segments, - xen_blkif_max_segments); + if (indirect_segments > xen_blkif_max_segments) + indirect_segments = xen_blkif_max_segments; + if (indirect_segments <= BLKIF_MAX_SEGMENTS_PER_REQUEST) + indirect_segments = 0; + info->max_indirect_segments = indirect_segments; } /* @@ -2652,6 +2655,9 @@ static int __init xlblk_init(void) if (!xen_domain()) return -ENODEV; + if (xen_blkif_max_segments < BLKIF_MAX_SEGMENTS_PER_REQUEST) + xen_blkif_max_segments = BLKIF_MAX_SEGMENTS_PER_REQUEST; + if (xen_blkif_max_ring_order > XENBUS_MAX_RING_GRANT_ORDER) { pr_info("Invalid max_ring_order (%d), will use default max: %d.\n", xen_blkif_max_ring_order, XENBUS_MAX_RING_GRANT_ORDER); From d2b3964a0780d2d2994eba57f950d6c9fe489ed8 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 20 Jan 2017 09:31:54 -0800 Subject: [PATCH 154/343] xfs: fix COW writeback race Due to the way how xfs_iomap_write_allocate tries to convert the whole found extents from delalloc to real space we can run into a race condition with multiple threads doing writes to this same extent. For the non-COW case that is harmless as the only thing that can happen is that we call xfs_bmapi_write on an extent that has already been converted to a real allocation. For COW writes where we move the extent from the COW to the data fork after I/O completion the race is, however, not quite as harmless. In the worst case we are now calling xfs_bmapi_write on a region that contains hole in the COW work, which will trip up an assert in debug builds or lead to file system corruption in non-debug builds. This seems to be reproducible with workloads of small O_DSYNC write, although so far I've not managed to come up with a with an isolated reproducer. The fix for the issue is relatively simple: tell xfs_bmapi_write that we are only asked to convert delayed allocations and skip holes in that case. Signed-off-by: Christoph Hellwig Reviewed-by: Brian Foster Reviewed-by: Darrick J. Wong Signed-off-by: Darrick J. Wong --- fs/xfs/libxfs/xfs_bmap.c | 44 +++++++++++++++++++++++++++++----------- fs/xfs/libxfs/xfs_bmap.h | 6 +++++- fs/xfs/xfs_iomap.c | 2 +- 3 files changed, 38 insertions(+), 14 deletions(-) diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c index 44773c9eb957..ab82dd4a4980 100644 --- a/fs/xfs/libxfs/xfs_bmap.c +++ b/fs/xfs/libxfs/xfs_bmap.c @@ -4514,8 +4514,6 @@ xfs_bmapi_write( int n; /* current extent index */ xfs_fileoff_t obno; /* old block number (offset) */ int whichfork; /* data or attr fork */ - char inhole; /* current location is hole in file */ - char wasdelay; /* old extent was delayed */ #ifdef DEBUG xfs_fileoff_t orig_bno; /* original block number value */ @@ -4603,22 +4601,44 @@ xfs_bmapi_write( bma.firstblock = firstblock; while (bno < end && n < *nmap) { - inhole = eof || bma.got.br_startoff > bno; - wasdelay = !inhole && isnullstartblock(bma.got.br_startblock); + bool need_alloc = false, wasdelay = false; - /* - * Make sure we only reflink into a hole. - */ - if (flags & XFS_BMAPI_REMAP) - ASSERT(inhole); - if (flags & XFS_BMAPI_COWFORK) - ASSERT(!inhole); + /* in hole or beyoned EOF? */ + if (eof || bma.got.br_startoff > bno) { + if (flags & XFS_BMAPI_DELALLOC) { + /* + * For the COW fork we can reasonably get a + * request for converting an extent that races + * with other threads already having converted + * part of it, as there converting COW to + * regular blocks is not protected using the + * IOLOCK. + */ + ASSERT(flags & XFS_BMAPI_COWFORK); + if (!(flags & XFS_BMAPI_COWFORK)) { + error = -EIO; + goto error0; + } + + if (eof || bno >= end) + break; + } else { + need_alloc = true; + } + } else { + /* + * Make sure we only reflink into a hole. + */ + ASSERT(!(flags & XFS_BMAPI_REMAP)); + if (isnullstartblock(bma.got.br_startblock)) + wasdelay = true; + } /* * First, deal with the hole before the allocated space * that we found, if any. */ - if (inhole || wasdelay) { + if (need_alloc || wasdelay) { bma.eof = eof; bma.conv = !!(flags & XFS_BMAPI_CONVERT); bma.wasdel = wasdelay; diff --git a/fs/xfs/libxfs/xfs_bmap.h b/fs/xfs/libxfs/xfs_bmap.h index cecd094404cc..cdef87db5262 100644 --- a/fs/xfs/libxfs/xfs_bmap.h +++ b/fs/xfs/libxfs/xfs_bmap.h @@ -110,6 +110,9 @@ struct xfs_extent_free_item /* Map something in the CoW fork. */ #define XFS_BMAPI_COWFORK 0x200 +/* Only convert delalloc space, don't allocate entirely new extents */ +#define XFS_BMAPI_DELALLOC 0x400 + #define XFS_BMAPI_FLAGS \ { XFS_BMAPI_ENTIRE, "ENTIRE" }, \ { XFS_BMAPI_METADATA, "METADATA" }, \ @@ -120,7 +123,8 @@ struct xfs_extent_free_item { XFS_BMAPI_CONVERT, "CONVERT" }, \ { XFS_BMAPI_ZERO, "ZERO" }, \ { XFS_BMAPI_REMAP, "REMAP" }, \ - { XFS_BMAPI_COWFORK, "COWFORK" } + { XFS_BMAPI_COWFORK, "COWFORK" }, \ + { XFS_BMAPI_DELALLOC, "DELALLOC" } static inline int xfs_bmapi_aflag(int w) diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c index 0d147428971e..1aa3abd67b36 100644 --- a/fs/xfs/xfs_iomap.c +++ b/fs/xfs/xfs_iomap.c @@ -681,7 +681,7 @@ xfs_iomap_write_allocate( xfs_trans_t *tp; int nimaps; int error = 0; - int flags = 0; + int flags = XFS_BMAPI_DELALLOC; int nres; if (whichfork == XFS_COW_FORK) From eab127717a6af54401ba534790c793ec143cd1fc Mon Sep 17 00:00:00 2001 From: Florian Fainelli Date: Fri, 20 Jan 2017 15:31:52 -0800 Subject: [PATCH 155/343] net: phy: Avoid deadlock during phy_error() phy_error() is called in the PHY state machine workqueue context, and calls phy_trigger_machine() which does a cancel_delayed_work_sync() of the workqueue we execute from, causing a deadlock situation. Augment phy_trigger_machine() machine with a sync boolean indicating whether we should use cancel_*_sync() or just cancel_*_work(). Fixes: 3c293f4e08b5 ("net: phy: Trigger state machine on state change and not polling.") Reported-by: Russell King Signed-off-by: Florian Fainelli Signed-off-by: David S. Miller --- drivers/net/phy/phy.c | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-) diff --git a/drivers/net/phy/phy.c b/drivers/net/phy/phy.c index 48da6e93c3f7..e687a9cb4a37 100644 --- a/drivers/net/phy/phy.c +++ b/drivers/net/phy/phy.c @@ -649,14 +649,18 @@ void phy_start_machine(struct phy_device *phydev) * phy_trigger_machine - trigger the state machine to run * * @phydev: the phy_device struct + * @sync: indicate whether we should wait for the workqueue cancelation * * Description: There has been a change in state which requires that the * state machine runs. */ -static void phy_trigger_machine(struct phy_device *phydev) +static void phy_trigger_machine(struct phy_device *phydev, bool sync) { - cancel_delayed_work_sync(&phydev->state_queue); + if (sync) + cancel_delayed_work_sync(&phydev->state_queue); + else + cancel_delayed_work(&phydev->state_queue); queue_delayed_work(system_power_efficient_wq, &phydev->state_queue, 0); } @@ -693,7 +697,7 @@ static void phy_error(struct phy_device *phydev) phydev->state = PHY_HALTED; mutex_unlock(&phydev->lock); - phy_trigger_machine(phydev); + phy_trigger_machine(phydev, false); } /** @@ -840,7 +844,7 @@ void phy_change(struct phy_device *phydev) } /* reschedule state queue work to run as soon as possible */ - phy_trigger_machine(phydev); + phy_trigger_machine(phydev, true); return; ignore: @@ -942,7 +946,7 @@ void phy_start(struct phy_device *phydev) if (do_resume) phy_resume(phydev); - phy_trigger_machine(phydev); + phy_trigger_machine(phydev, true); } EXPORT_SYMBOL(phy_start); From 4078b76cac68e50ccf1f76a74e7d3d5788aec3fe Mon Sep 17 00:00:00 2001 From: Florian Fainelli Date: Fri, 20 Jan 2017 16:05:05 -0800 Subject: [PATCH 156/343] net: dsa: Check return value of phy_connect_direct() We need to check the return value of phy_connect_direct() in dsa_slave_phy_connect() otherwise we may be continuing the initialization of a slave network device with a PHY that already attached somewhere else and which will soon be in error because the PHY device is in error. The conditions for such an error to occur are that we have a port of our switch that is not disabled, and has the same port number as a PHY address (say both 5) that can be probed using the DSA slave MII bus. We end-up having this slave network device find a PHY at the same address as our port number, and we try to attach to it. A slave network (e.g: port 0) has already attached to our PHY device, and we try to re-attach it with a different network device, but since we ignore the error we would end-up initializating incorrect device references by the time the slave network interface is opened. The code has been (re)organized several times, making it hard to provide an exact Fixes tag, this is a bugfix nonetheless. Signed-off-by: Florian Fainelli Signed-off-by: David S. Miller --- net/dsa/slave.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/net/dsa/slave.c b/net/dsa/slave.c index 68c9eea00518..ba1b6b9630d2 100644 --- a/net/dsa/slave.c +++ b/net/dsa/slave.c @@ -1105,10 +1105,8 @@ static int dsa_slave_phy_connect(struct dsa_slave_priv *p, /* Use already configured phy mode */ if (p->phy_interface == PHY_INTERFACE_MODE_NA) p->phy_interface = p->phy->interface; - phy_connect_direct(slave_dev, p->phy, dsa_slave_adjust_link, - p->phy_interface); - - return 0; + return phy_connect_direct(slave_dev, p->phy, dsa_slave_adjust_link, + p->phy_interface); } static int dsa_slave_phy_setup(struct dsa_slave_priv *p, From c5f21c9f878b8dcd54d0b9739c025ca73cb4c091 Mon Sep 17 00:00:00 2001 From: Ding Pixel Date: Wed, 18 Jan 2017 17:26:38 +0800 Subject: [PATCH 157/343] drm/amdgpu: check ring being ready before using MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Return success when the ring is properly initialized, otherwise return failure. Tonga SRIOV VF doesn't have UVD and VCE engines, the initialization of these IPs is bypassed. The system crashes if application submit IB to their rings which are not ready to use. It could be a common issue if IP having ring buffer is disabled for some reason on specific ASIC, so it should check the ring being ready to use. Bug: amdgpu_test crashes system on Tonga VF. Signed-off-by: Ding Pixel Reviewed-by: Christian König Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index 29d6d84d1c28..41e41f90265d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -83,6 +83,13 @@ int amdgpu_cs_get_ring(struct amdgpu_device *adev, u32 ip_type, } break; } + + if (!(*out_ring && (*out_ring)->adev)) { + DRM_ERROR("Ring %d is not initialized on IP %d\n", + ring, ip_type); + return -EINVAL; + } + return 0; } From 3a1d19a29670aa7eb58576a31883d0aa9fb77549 Mon Sep 17 00:00:00 2001 From: Xiangliang Yu Date: Thu, 19 Jan 2017 09:57:41 +0800 Subject: [PATCH 158/343] drm/amdgpu: fix unload driver issue for virtual display MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Virtual display doesn't allocate amdgpu_encoder when initializing, so will get invaild pointer if try to free amdgpu_encoder when unloading driver. Signed-off-by: Xiangliang Yu Reviewed-by: Alex Deucher Acked-by: Christian König Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/dce_virtual.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/dce_virtual.c b/drivers/gpu/drm/amd/amdgpu/dce_virtual.c index 762f8e82ceb7..e9a176891e13 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_virtual.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_virtual.c @@ -627,11 +627,8 @@ static const struct drm_encoder_helper_funcs dce_virtual_encoder_helper_funcs = static void dce_virtual_encoder_destroy(struct drm_encoder *encoder) { - struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder); - - kfree(amdgpu_encoder->enc_priv); drm_encoder_cleanup(encoder); - kfree(amdgpu_encoder); + kfree(encoder); } static const struct drm_encoder_funcs dce_virtual_encoder_funcs = { From b2fbd04498789def80ceba3d5bbc5af7f2f70a5f Mon Sep 17 00:00:00 2001 From: Liping Zhang Date: Fri, 20 Jan 2017 21:03:03 +0800 Subject: [PATCH 159/343] netfilter: nf_tables: validate the name size when possible Currently, if the user add a stateful object with the name size exceed NFT_OBJ_MAXNAMELEN - 1 (i.e. 31), we truncate it down to 31 silently. This is not friendly, furthermore, this will cause duplicated stateful objects when the first 31 characters of the name is same. So limit the stateful object's name size to NFT_OBJ_MAXNAMELEN - 1. After apply this patch, error message will be printed out like this: # name_32=$(printf "%0.sQ" {1..32}) # nft add counter filter $name_32 :1:1-52: Error: Could not process rule: Numerical result out of range add counter filter QQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQQ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Also this patch cleans up the codes which missing the name size limit validation in nftables. Fixes: e50092404c1b ("netfilter: nf_tables: add stateful objects") Signed-off-by: Liping Zhang Signed-off-by: Pablo Neira Ayuso --- net/netfilter/nf_tables_api.c | 21 ++++++++++++++------- net/netfilter/nft_dynset.c | 3 ++- net/netfilter/nft_lookup.c | 3 ++- net/netfilter/nft_objref.c | 6 ++++-- 4 files changed, 22 insertions(+), 11 deletions(-) diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c index 091d2dcc63b2..b84c7b25219b 100644 --- a/net/netfilter/nf_tables_api.c +++ b/net/netfilter/nf_tables_api.c @@ -928,7 +928,8 @@ static struct nft_chain *nf_tables_chain_lookup(const struct nft_table *table, } static const struct nla_policy nft_chain_policy[NFTA_CHAIN_MAX + 1] = { - [NFTA_CHAIN_TABLE] = { .type = NLA_STRING }, + [NFTA_CHAIN_TABLE] = { .type = NLA_STRING, + .len = NFT_TABLE_MAXNAMELEN - 1 }, [NFTA_CHAIN_HANDLE] = { .type = NLA_U64 }, [NFTA_CHAIN_NAME] = { .type = NLA_STRING, .len = NFT_CHAIN_MAXNAMELEN - 1 }, @@ -1854,7 +1855,8 @@ static struct nft_rule *nf_tables_rule_lookup(const struct nft_chain *chain, } static const struct nla_policy nft_rule_policy[NFTA_RULE_MAX + 1] = { - [NFTA_RULE_TABLE] = { .type = NLA_STRING }, + [NFTA_RULE_TABLE] = { .type = NLA_STRING, + .len = NFT_TABLE_MAXNAMELEN - 1 }, [NFTA_RULE_CHAIN] = { .type = NLA_STRING, .len = NFT_CHAIN_MAXNAMELEN - 1 }, [NFTA_RULE_HANDLE] = { .type = NLA_U64 }, @@ -2443,7 +2445,8 @@ nft_select_set_ops(const struct nlattr * const nla[], } static const struct nla_policy nft_set_policy[NFTA_SET_MAX + 1] = { - [NFTA_SET_TABLE] = { .type = NLA_STRING }, + [NFTA_SET_TABLE] = { .type = NLA_STRING, + .len = NFT_TABLE_MAXNAMELEN - 1 }, [NFTA_SET_NAME] = { .type = NLA_STRING, .len = NFT_SET_MAXNAMELEN - 1 }, [NFTA_SET_FLAGS] = { .type = NLA_U32 }, @@ -3192,8 +3195,10 @@ static const struct nla_policy nft_set_elem_policy[NFTA_SET_ELEM_MAX + 1] = { }; static const struct nla_policy nft_set_elem_list_policy[NFTA_SET_ELEM_LIST_MAX + 1] = { - [NFTA_SET_ELEM_LIST_TABLE] = { .type = NLA_STRING }, - [NFTA_SET_ELEM_LIST_SET] = { .type = NLA_STRING }, + [NFTA_SET_ELEM_LIST_TABLE] = { .type = NLA_STRING, + .len = NFT_TABLE_MAXNAMELEN - 1 }, + [NFTA_SET_ELEM_LIST_SET] = { .type = NLA_STRING, + .len = NFT_SET_MAXNAMELEN - 1 }, [NFTA_SET_ELEM_LIST_ELEMENTS] = { .type = NLA_NESTED }, [NFTA_SET_ELEM_LIST_SET_ID] = { .type = NLA_U32 }, }; @@ -4032,8 +4037,10 @@ struct nft_object *nf_tables_obj_lookup(const struct nft_table *table, EXPORT_SYMBOL_GPL(nf_tables_obj_lookup); static const struct nla_policy nft_obj_policy[NFTA_OBJ_MAX + 1] = { - [NFTA_OBJ_TABLE] = { .type = NLA_STRING }, - [NFTA_OBJ_NAME] = { .type = NLA_STRING }, + [NFTA_OBJ_TABLE] = { .type = NLA_STRING, + .len = NFT_TABLE_MAXNAMELEN - 1 }, + [NFTA_OBJ_NAME] = { .type = NLA_STRING, + .len = NFT_OBJ_MAXNAMELEN - 1 }, [NFTA_OBJ_TYPE] = { .type = NLA_U32 }, [NFTA_OBJ_DATA] = { .type = NLA_NESTED }, }; diff --git a/net/netfilter/nft_dynset.c b/net/netfilter/nft_dynset.c index 7de2f46734a4..049ad2d9ee66 100644 --- a/net/netfilter/nft_dynset.c +++ b/net/netfilter/nft_dynset.c @@ -98,7 +98,8 @@ out: } static const struct nla_policy nft_dynset_policy[NFTA_DYNSET_MAX + 1] = { - [NFTA_DYNSET_SET_NAME] = { .type = NLA_STRING }, + [NFTA_DYNSET_SET_NAME] = { .type = NLA_STRING, + .len = NFT_SET_MAXNAMELEN - 1 }, [NFTA_DYNSET_SET_ID] = { .type = NLA_U32 }, [NFTA_DYNSET_OP] = { .type = NLA_U32 }, [NFTA_DYNSET_SREG_KEY] = { .type = NLA_U32 }, diff --git a/net/netfilter/nft_lookup.c b/net/netfilter/nft_lookup.c index d4f97fa7e21d..e21aea7e5ec8 100644 --- a/net/netfilter/nft_lookup.c +++ b/net/netfilter/nft_lookup.c @@ -49,7 +49,8 @@ static void nft_lookup_eval(const struct nft_expr *expr, } static const struct nla_policy nft_lookup_policy[NFTA_LOOKUP_MAX + 1] = { - [NFTA_LOOKUP_SET] = { .type = NLA_STRING }, + [NFTA_LOOKUP_SET] = { .type = NLA_STRING, + .len = NFT_SET_MAXNAMELEN - 1 }, [NFTA_LOOKUP_SET_ID] = { .type = NLA_U32 }, [NFTA_LOOKUP_SREG] = { .type = NLA_U32 }, [NFTA_LOOKUP_DREG] = { .type = NLA_U32 }, diff --git a/net/netfilter/nft_objref.c b/net/netfilter/nft_objref.c index 415a65ba2b85..1ae8c49ca4a1 100644 --- a/net/netfilter/nft_objref.c +++ b/net/netfilter/nft_objref.c @@ -193,10 +193,12 @@ nft_objref_select_ops(const struct nft_ctx *ctx, } static const struct nla_policy nft_objref_policy[NFTA_OBJREF_MAX + 1] = { - [NFTA_OBJREF_IMM_NAME] = { .type = NLA_STRING }, + [NFTA_OBJREF_IMM_NAME] = { .type = NLA_STRING, + .len = NFT_OBJ_MAXNAMELEN - 1 }, [NFTA_OBJREF_IMM_TYPE] = { .type = NLA_U32 }, [NFTA_OBJREF_SET_SREG] = { .type = NLA_U32 }, - [NFTA_OBJREF_SET_NAME] = { .type = NLA_STRING }, + [NFTA_OBJREF_SET_NAME] = { .type = NLA_STRING, + .len = NFT_SET_MAXNAMELEN - 1 }, [NFTA_OBJREF_SET_ID] = { .type = NLA_U32 }, }; From 8ac092519ad91931c96d306c4bfae2c6587c325f Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 23 Jan 2017 22:44:12 -0500 Subject: [PATCH 160/343] NFSv4.1: Fix a deadlock in layoutget We cannot call nfs4_handle_exception() without first ensuring that the slot has been freed. If not, we end up deadlocking with the process waiting for recovery to complete, and recovery waiting for the slot table to drain. Fixes: 2e80dbe7ac51 ("NFSv4.1: Close callback races for OPEN, LAYOUTGET...") Cc: stable@vger.kernel.org # v4.8+ Signed-off-by: Trond Myklebust --- fs/nfs/nfs4proc.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index ecc151697fd4..59bb574d7d7c 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -8490,6 +8490,7 @@ nfs4_layoutget_handle_exception(struct rpc_task *task, goto out; } + nfs4_sequence_free_slot(&lgp->res.seq_res); err = nfs4_handle_exception(server, nfs4err, exception); if (!status) { if (exception->retry) From 9dce990d2cf57b5ed4e71a9cdbd7eae4335111ff Mon Sep 17 00:00:00 2001 From: Martin Schwidefsky Date: Tue, 24 Jan 2017 08:05:52 +0100 Subject: [PATCH 161/343] s390/ptrace: Preserve previous registers for short regset write Ensure that if userspace supplies insufficient data to PTRACE_SETREGSET to fill all the registers, the thread's old registers are preserved. convert_vx_to_fp() is adapted to handle only a specified number of registers rather than unconditionally handling all of them: other callers of this function are adapted appropriately. Based on an initial patch by Dave Martin. Cc: stable@vger.kernel.org Reported-by: Dave Martin Signed-off-by: Martin Schwidefsky --- arch/s390/kernel/ptrace.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/arch/s390/kernel/ptrace.c b/arch/s390/kernel/ptrace.c index 7447ba509c30..12020b55887b 100644 --- a/arch/s390/kernel/ptrace.c +++ b/arch/s390/kernel/ptrace.c @@ -963,6 +963,11 @@ static int s390_fpregs_set(struct task_struct *target, if (target == current) save_fpu_regs(); + if (MACHINE_HAS_VX) + convert_vx_to_fp(fprs, target->thread.fpu.vxrs); + else + memcpy(&fprs, target->thread.fpu.fprs, sizeof(fprs)); + /* If setting FPC, must validate it first. */ if (count > 0 && pos < offsetof(s390_fp_regs, fprs)) { u32 ufpc[2] = { target->thread.fpu.fpc, 0 }; @@ -1067,6 +1072,9 @@ static int s390_vxrs_low_set(struct task_struct *target, if (target == current) save_fpu_regs(); + for (i = 0; i < __NUM_VXRS_LOW; i++) + vxrs[i] = *((__u64 *)(target->thread.fpu.vxrs + i) + 1); + rc = user_regset_copyin(&pos, &count, &kbuf, &ubuf, vxrs, 0, -1); if (rc == 0) for (i = 0; i < __NUM_VXRS_LOW; i++) From 690e5325b8c7d5db05fc569c0f7b888bb4248272 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 24 Jan 2017 14:50:19 +0100 Subject: [PATCH 162/343] block: fix use after free in __blkdev_direct_IO MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We can't dereference the dio structure after submitting the last bio for this request, as I/O completion might have happened before the code is run. Introduce a local is_sync variable instead. Fixes: 542ff7bf ("block: new direct I/O implementation") Signed-off-by: Christoph Hellwig Reported-by: Matias Bjørling Tested-by: Matias Bjørling Signed-off-by: Jens Axboe --- fs/block_dev.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/fs/block_dev.c b/fs/block_dev.c index 5db5d1340d69..3c47614a4b32 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -331,7 +331,7 @@ __blkdev_direct_IO(struct kiocb *iocb, struct iov_iter *iter, int nr_pages) struct blk_plug plug; struct blkdev_dio *dio; struct bio *bio; - bool is_read = (iov_iter_rw(iter) == READ); + bool is_read = (iov_iter_rw(iter) == READ), is_sync; loff_t pos = iocb->ki_pos; blk_qc_t qc = BLK_QC_T_NONE; int ret; @@ -344,7 +344,7 @@ __blkdev_direct_IO(struct kiocb *iocb, struct iov_iter *iter, int nr_pages) bio_get(bio); /* extra ref for the completion handler */ dio = container_of(bio, struct blkdev_dio, bio); - dio->is_sync = is_sync_kiocb(iocb); + dio->is_sync = is_sync = is_sync_kiocb(iocb); if (dio->is_sync) dio->waiter = current; else @@ -398,7 +398,7 @@ __blkdev_direct_IO(struct kiocb *iocb, struct iov_iter *iter, int nr_pages) } blk_finish_plug(&plug); - if (!dio->is_sync) + if (!is_sync) return -EIOCBQUEUED; for (;;) { From 0d6da872d3e4a60f43c295386d7ff9a4cdcd57e9 Mon Sep 17 00:00:00 2001 From: Christian Borntraeger Date: Mon, 23 Jan 2017 22:59:44 +0100 Subject: [PATCH 163/343] s390/mm: Fix cmma unused transfer from pgste into pte The last pgtable rework silently disabled the CMMA unused state by setting a local pte variable (a parameter) instead of propagating it back into the caller. Fix it. Fixes: ebde765c0e85 ("s390/mm: uninline ptep_xxx functions from pgtable.h") Cc: stable@vger.kernel.org # v4.6+ Cc: Martin Schwidefsky Cc: Claudio Imbrenda Signed-off-by: Christian Borntraeger Signed-off-by: Martin Schwidefsky --- arch/s390/mm/pgtable.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/arch/s390/mm/pgtable.c b/arch/s390/mm/pgtable.c index 7a1897c51c54..d56ef26d4681 100644 --- a/arch/s390/mm/pgtable.c +++ b/arch/s390/mm/pgtable.c @@ -202,7 +202,7 @@ static inline pgste_t ptep_xchg_start(struct mm_struct *mm, return pgste; } -static inline void ptep_xchg_commit(struct mm_struct *mm, +static inline pte_t ptep_xchg_commit(struct mm_struct *mm, unsigned long addr, pte_t *ptep, pgste_t pgste, pte_t old, pte_t new) { @@ -220,6 +220,7 @@ static inline void ptep_xchg_commit(struct mm_struct *mm, } else { *ptep = new; } + return old; } pte_t ptep_xchg_direct(struct mm_struct *mm, unsigned long addr, @@ -231,7 +232,7 @@ pte_t ptep_xchg_direct(struct mm_struct *mm, unsigned long addr, preempt_disable(); pgste = ptep_xchg_start(mm, addr, ptep); old = ptep_flush_direct(mm, addr, ptep); - ptep_xchg_commit(mm, addr, ptep, pgste, old, new); + old = ptep_xchg_commit(mm, addr, ptep, pgste, old, new); preempt_enable(); return old; } @@ -246,7 +247,7 @@ pte_t ptep_xchg_lazy(struct mm_struct *mm, unsigned long addr, preempt_disable(); pgste = ptep_xchg_start(mm, addr, ptep); old = ptep_flush_lazy(mm, addr, ptep); - ptep_xchg_commit(mm, addr, ptep, pgste, old, new); + old = ptep_xchg_commit(mm, addr, ptep, pgste, old, new); preempt_enable(); return old; } From 115865fa0826ed18ca04717cf72d0fe874c0fe7f Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Mon, 23 Jan 2017 09:29:09 +0100 Subject: [PATCH 164/343] mac80211: don't try to sleep in rate_control_rate_init() In my previous patch, I missed that rate_control_rate_init() is called from some places that cannot sleep, so it cannot call ieee80211_recalc_min_chandef(). Remove that call for now to fix the context bug, we'll have to find a different way to fix the minimum channel width issue. Fixes: 96aa2e7cf126 ("mac80211: calculate min channel width correctly") Reported-by: Xiaolong Ye (via lkp-robot) Signed-off-by: Johannes Berg --- net/mac80211/rate.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/net/mac80211/rate.c b/net/mac80211/rate.c index 9e2641d45587..206698bc93f4 100644 --- a/net/mac80211/rate.c +++ b/net/mac80211/rate.c @@ -40,8 +40,6 @@ void rate_control_rate_init(struct sta_info *sta) ieee80211_sta_set_rx_nss(sta); - ieee80211_recalc_min_chandef(sta->sdata); - if (!ref) return; From ad8e66b4a80182174f73487ed25fd2140cf43361 Mon Sep 17 00:00:00 2001 From: Israel Rukshin Date: Wed, 28 Dec 2016 12:48:28 +0200 Subject: [PATCH 165/343] IB/srp: fix mr allocation when the device supports sg gaps If the device support arbitrary sg list mapping (device cap IB_DEVICE_SG_GAPS_REG set) we allocate the memory regions with IB_MR_TYPE_SG_GAPS. Fixes: 509c5f33f4f6 ("IB/srp: Prevent mapping failures") Cc: # 4.7+ Signed-off-by: Israel Rukshin Signed-off-by: Max Gurtovoy Reviewed-by: Leon Romanovsky Reviewed-by: Mark Bloch Reviewed-by: Yuval Shaia Reviewed-by: Bart Van Assche Signed-off-by: Doug Ledford --- drivers/infiniband/ulp/srp/ib_srp.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) diff --git a/drivers/infiniband/ulp/srp/ib_srp.c b/drivers/infiniband/ulp/srp/ib_srp.c index 8ddc07123193..0f67cf909462 100644 --- a/drivers/infiniband/ulp/srp/ib_srp.c +++ b/drivers/infiniband/ulp/srp/ib_srp.c @@ -371,6 +371,7 @@ static struct srp_fr_pool *srp_create_fr_pool(struct ib_device *device, struct srp_fr_desc *d; struct ib_mr *mr; int i, ret = -EINVAL; + enum ib_mr_type mr_type; if (pool_size <= 0) goto err; @@ -384,9 +385,13 @@ static struct srp_fr_pool *srp_create_fr_pool(struct ib_device *device, spin_lock_init(&pool->lock); INIT_LIST_HEAD(&pool->free_list); + if (device->attrs.device_cap_flags & IB_DEVICE_SG_GAPS_REG) + mr_type = IB_MR_TYPE_SG_GAPS; + else + mr_type = IB_MR_TYPE_MEM_REG; + for (i = 0, d = &pool->desc[0]; i < pool->size; i++, d++) { - mr = ib_alloc_mr(pd, IB_MR_TYPE_MEM_REG, - max_page_list_len); + mr = ib_alloc_mr(pd, mr_type, max_page_list_len); if (IS_ERR(mr)) { ret = PTR_ERR(mr); if (ret == -ENOMEM) From 0a475ef4226e305bdcffe12b401ca1eab06c4913 Mon Sep 17 00:00:00 2001 From: Israel Rukshin Date: Wed, 4 Jan 2017 15:59:37 +0200 Subject: [PATCH 166/343] IB/srp: fix invalid indirect_sg_entries parameter value After setting indirect_sg_entries module_param to huge value (e.g 500,000), srp_alloc_req_data() fails to allocate indirect descriptors for the request ring (kmalloc fails). This commit enforces the maximum value of indirect_sg_entries to be SG_MAX_SEGMENTS as signified in module param description. Fixes: 65e8617fba17 (scsi: rename SCSI_MAX_{SG, SG_CHAIN}_SEGMENTS) Fixes: c07d424d6118 (IB/srp: add support for indirect tables that don't fit in SRP_CMD) Cc: stable@vger.kernel.org # 4.7+ Signed-off-by: Israel Rukshin Signed-off-by: Max Gurtovoy Reviewed-by: Laurence Oberman Reviewed-by: Bart Van Assche -- Signed-off-by: Doug Ledford --- drivers/infiniband/ulp/srp/ib_srp.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/drivers/infiniband/ulp/srp/ib_srp.c b/drivers/infiniband/ulp/srp/ib_srp.c index 0f67cf909462..79bf48477ddb 100644 --- a/drivers/infiniband/ulp/srp/ib_srp.c +++ b/drivers/infiniband/ulp/srp/ib_srp.c @@ -3699,6 +3699,12 @@ static int __init srp_init_module(void) indirect_sg_entries = cmd_sg_entries; } + if (indirect_sg_entries > SG_MAX_SEGMENTS) { + pr_warn("Clamping indirect_sg_entries to %u\n", + SG_MAX_SEGMENTS); + indirect_sg_entries = SG_MAX_SEGMENTS; + } + srp_remove_wq = create_workqueue("srp_remove"); if (!srp_remove_wq) { ret = -ENOMEM; From 7b9e1d89e1b6a3b99a8fdd949aa0f98dd5bf2f6b Mon Sep 17 00:00:00 2001 From: Andy Shevchenko Date: Tue, 24 Jan 2017 17:22:01 +0200 Subject: [PATCH 167/343] MAINTAINERS: Add myself to X86 PLATFORM DRIVERS as a co-maintainer For last few months Darren and I are co-maintaining PDx86 subsystem. Make this fact official by updating MAINTAINERS database. Acked-by: Darren Hart Signed-off-by: Andy Shevchenko --- MAINTAINERS | 1 + 1 file changed, 1 insertion(+) diff --git a/MAINTAINERS b/MAINTAINERS index c36976d3bd1a..235d66237938 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -13436,6 +13436,7 @@ F: arch/x86/ X86 PLATFORM DRIVERS M: Darren Hart +M: Andy Shevchenko L: platform-driver-x86@vger.kernel.org T: git git://git.infradead.org/users/dvhart/linux-platform-drivers-x86.git S: Maintained From 1e5db6c31ade4150c2e2b1a21e39f776c38fea39 Mon Sep 17 00:00:00 2001 From: Max Gurtovoy Date: Wed, 18 Jan 2017 00:40:39 +0200 Subject: [PATCH 168/343] IB/iser: Fix sg_tablesize calculation For devices that can register page list that is bigger than USHRT_MAX, we actually take the wrong value for sg_tablesize. E.g: for CX4 max_fast_reg_page_list_len is 65536 (bigger than USHRT_MAX) so we set sg_tablesize to 0 by mistake. Therefore, each IO that is bigger than 4k splitted to "< 4k" chunks that cause performance degredation. Remove wrong sg_tablesize assignment, and use the value that was set during address resolution handler with the needed casting. Cc: # v4.5+ Signed-off-by: Max Gurtovoy Reviewed-by: Sagi Grimberg Signed-off-by: Doug Ledford --- drivers/infiniband/ulp/iser/iscsi_iser.c | 7 ------- 1 file changed, 7 deletions(-) diff --git a/drivers/infiniband/ulp/iser/iscsi_iser.c b/drivers/infiniband/ulp/iser/iscsi_iser.c index 9104e6b8cac9..1c911876556e 100644 --- a/drivers/infiniband/ulp/iser/iscsi_iser.c +++ b/drivers/infiniband/ulp/iser/iscsi_iser.c @@ -651,13 +651,6 @@ iscsi_iser_session_create(struct iscsi_endpoint *ep, SHOST_DIX_GUARD_CRC); } - /* - * Limit the sg_tablesize and max_sectors based on the device - * max fastreg page list length. - */ - shost->sg_tablesize = min_t(unsigned short, shost->sg_tablesize, - ib_conn->device->ib_device->attrs.max_fast_reg_page_list_len); - if (iscsi_host_add(shost, ib_conn->device->ib_device->dma_device)) { mutex_unlock(&iser_conn->state_mutex); From 83236f0157feec0f01bf688a1474b889bdcc5ad0 Mon Sep 17 00:00:00 2001 From: Max Gurtovoy Date: Wed, 18 Jan 2017 00:40:40 +0200 Subject: [PATCH 169/343] IB/iser: remove unused variable from iser_conn struct max_sectors calculation was fixed in commit: 9c674815d346 ("IB/iser: Fix max_sectors calculation"). Thus, iser_conn variable scsi_max_sectors is not needed anymore. Signed-off-by: Max Gurtovoy Reviewed-by: Sagi Grimberg Tested-by: Raju Rangoju Signed-off-by: Doug Ledford --- drivers/infiniband/ulp/iser/iscsi_iser.c | 4 ++++ drivers/infiniband/ulp/iser/iscsi_iser.h | 2 -- drivers/infiniband/ulp/iser/iser_verbs.c | 13 +------------ 3 files changed, 5 insertions(+), 14 deletions(-) diff --git a/drivers/infiniband/ulp/iser/iscsi_iser.c b/drivers/infiniband/ulp/iser/iscsi_iser.c index 1c911876556e..e71af717e71b 100644 --- a/drivers/infiniband/ulp/iser/iscsi_iser.c +++ b/drivers/infiniband/ulp/iser/iscsi_iser.c @@ -672,6 +672,10 @@ iscsi_iser_session_create(struct iscsi_endpoint *ep, max_fr_sectors = ((shost->sg_tablesize - 1) * PAGE_SIZE) >> 9; shost->max_sectors = min(iser_max_sectors, max_fr_sectors); + iser_dbg("iser_conn %p, sg_tablesize %u, max_sectors %u\n", + iser_conn, shost->sg_tablesize, + shost->max_sectors); + if (cmds_max > max_cmds) { iser_info("cmds_max changed from %u to %u\n", cmds_max, max_cmds); diff --git a/drivers/infiniband/ulp/iser/iscsi_iser.h b/drivers/infiniband/ulp/iser/iscsi_iser.h index 0be6a7c5ddb5..9d0b22ad58c1 100644 --- a/drivers/infiniband/ulp/iser/iscsi_iser.h +++ b/drivers/infiniband/ulp/iser/iscsi_iser.h @@ -496,7 +496,6 @@ struct ib_conn { * @rx_descs: rx buffers array (cyclic buffer) * @num_rx_descs: number of rx descriptors * @scsi_sg_tablesize: scsi host sg_tablesize - * @scsi_max_sectors: scsi host max sectors */ struct iser_conn { struct ib_conn ib_conn; @@ -519,7 +518,6 @@ struct iser_conn { struct iser_rx_desc *rx_descs; u32 num_rx_descs; unsigned short scsi_sg_tablesize; - unsigned int scsi_max_sectors; bool snd_w_inv; }; diff --git a/drivers/infiniband/ulp/iser/iser_verbs.c b/drivers/infiniband/ulp/iser/iser_verbs.c index 8ae7a3beddb7..6a9d1cb548ee 100644 --- a/drivers/infiniband/ulp/iser/iser_verbs.c +++ b/drivers/infiniband/ulp/iser/iser_verbs.c @@ -707,18 +707,7 @@ iser_calc_scsi_params(struct iser_conn *iser_conn, sup_sg_tablesize = min_t(unsigned, ISCSI_ISER_MAX_SG_TABLESIZE, device->ib_device->attrs.max_fast_reg_page_list_len); - if (sg_tablesize > sup_sg_tablesize) { - sg_tablesize = sup_sg_tablesize; - iser_conn->scsi_max_sectors = sg_tablesize * SIZE_4K / 512; - } else { - iser_conn->scsi_max_sectors = max_sectors; - } - - iser_conn->scsi_sg_tablesize = sg_tablesize; - - iser_dbg("iser_conn %p, sg_tablesize %u, max_sectors %u\n", - iser_conn, iser_conn->scsi_sg_tablesize, - iser_conn->scsi_max_sectors); + iser_conn->scsi_sg_tablesize = min(sg_tablesize, sup_sg_tablesize); } /** From bd00fdf198e2da475a2f4265a83686ab42d998a8 Mon Sep 17 00:00:00 2001 From: Greg Kurz Date: Tue, 24 Jan 2017 17:50:26 +0100 Subject: [PATCH 170/343] vfio/spapr: fail tce_iommu_attach_group() when iommu_data is null The recently added mediated VFIO driver doesn't know about powerpc iommu. It thus doesn't register a struct iommu_table_group in the iommu group upon device creation. The iommu_data pointer hence remains null. This causes a kernel oops when userspace tries to set the iommu type of a container associated with a mediated device to VFIO_SPAPR_TCE_v2_IOMMU. [ 82.585440] mtty mtty: MDEV: Registered [ 87.655522] iommu: Adding device 83b8f4f2-509f-382f-3c1e-e6bfe0fa1001 to group 10 [ 87.655527] vfio_mdev 83b8f4f2-509f-382f-3c1e-e6bfe0fa1001: MDEV: group_id = 10 [ 116.297184] Unable to handle kernel paging request for data at address 0x00000030 [ 116.297389] Faulting instruction address: 0xd000000007870524 [ 116.297465] Oops: Kernel access of bad area, sig: 11 [#1] [ 116.297611] SMP NR_CPUS=2048 [ 116.297611] NUMA [ 116.297627] PowerNV ... [ 116.297954] CPU: 33 PID: 7067 Comm: qemu-system-ppc Not tainted 4.10.0-rc5-mdev-test #8 [ 116.297993] task: c000000e7718b680 task.stack: c000000e77214000 [ 116.298025] NIP: d000000007870524 LR: d000000007870518 CTR: 0000000000000000 [ 116.298064] REGS: c000000e77217990 TRAP: 0300 Not tainted (4.10.0-rc5-mdev-test) [ 116.298103] MSR: 9000000000009033 [ 116.298107] CR: 84004444 XER: 00000000 [ 116.298154] CFAR: c00000000000888c DAR: 0000000000000030 DSISR: 40000000 SOFTE: 1 GPR00: d000000007870518 c000000e77217c10 d00000000787b0ed c000000eed2103c0 GPR04: 0000000000000000 0000000000000000 c000000eed2103e0 0000000f24320000 GPR08: 0000000000000104 0000000000000001 0000000000000000 d0000000078729b0 GPR12: c00000000025b7e0 c00000000fe08400 0000000000000001 000001002d31d100 GPR16: 000001002c22c850 00003ffff315c750 0000000043145680 0000000043141bc0 GPR20: ffffffffffffffed fffffffffffff000 0000000020003b65 d000000007706018 GPR24: c000000f16cf0d98 d000000007706000 c000000003f42980 c000000003f42980 GPR28: c000000f1575ac00 c000000003f429c8 0000000000000000 c000000eed2103c0 [ 116.298504] NIP [d000000007870524] tce_iommu_attach_group+0x10c/0x360 [vfio_iommu_spapr_tce] [ 116.298555] LR [d000000007870518] tce_iommu_attach_group+0x100/0x360 [vfio_iommu_spapr_tce] [ 116.298601] Call Trace: [ 116.298610] [c000000e77217c10] [d000000007870518] tce_iommu_attach_group+0x100/0x360 [vfio_iommu_spapr_tce] (unreliable) [ 116.298671] [c000000e77217cb0] [d0000000077033a0] vfio_fops_unl_ioctl+0x278/0x3e0 [vfio] [ 116.298713] [c000000e77217d40] [c0000000002a3ebc] do_vfs_ioctl+0xcc/0x8b0 [ 116.298745] [c000000e77217de0] [c0000000002a4700] SyS_ioctl+0x60/0xc0 [ 116.298782] [c000000e77217e30] [c00000000000b220] system_call+0x38/0xfc [ 116.298812] Instruction dump: [ 116.298828] 7d3f4b78 409effc8 3d220000 e9298020 3c800140 38a00018 608480c0 e8690028 [ 116.298869] 4800249d e8410018 7c7f1b79 41820230 2fa90000 419e0114 e9090020 [ 116.298914] ---[ end trace 1e10b0ced08b9120 ]--- This patch fixes the oops. Reported-by: Vaibhav Jain Signed-off-by: Greg Kurz Signed-off-by: Alex Williamson --- drivers/vfio/vfio_iommu_spapr_tce.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/drivers/vfio/vfio_iommu_spapr_tce.c b/drivers/vfio/vfio_iommu_spapr_tce.c index c8823578a1b2..128d10282d16 100644 --- a/drivers/vfio/vfio_iommu_spapr_tce.c +++ b/drivers/vfio/vfio_iommu_spapr_tce.c @@ -1270,6 +1270,10 @@ static int tce_iommu_attach_group(void *iommu_data, /* pr_debug("tce_vfio: Attaching group #%u to iommu %p\n", iommu_group_id(iommu_group), iommu_group); */ table_group = iommu_group_get_iommudata(iommu_group); + if (!table_group) { + ret = -ENODEV; + goto unlock_exit; + } if (tce_groups_attached(container) && (!table_group->ops || !table_group->ops->take_ownership || From b1a27eac7fefff33ccf6acc919fc0725bf9815fb Mon Sep 17 00:00:00 2001 From: Nicolas Iooss Date: Sun, 22 Jan 2017 14:41:22 +0100 Subject: [PATCH 171/343] IB/cxgb3: fix misspelling in header guard Use CXGB3_... instead of CXBG3_... Fixes: a85fb3383340 ("IB/cxgb3: Move user vendor structures") Cc: stable@vger.kernel.org # 4.9 Signed-off-by: Nicolas Iooss Reviewed-by: Leon Romanovsky Acked-by: Steve Wise Signed-off-by: Doug Ledford --- include/uapi/rdma/cxgb3-abi.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/include/uapi/rdma/cxgb3-abi.h b/include/uapi/rdma/cxgb3-abi.h index 48a19bda071b..d24eee12128f 100644 --- a/include/uapi/rdma/cxgb3-abi.h +++ b/include/uapi/rdma/cxgb3-abi.h @@ -30,7 +30,7 @@ * SOFTWARE. */ #ifndef CXGB3_ABI_USER_H -#define CXBG3_ABI_USER_H +#define CXGB3_ABI_USER_H #include From f39aac7e839368e3895dff952f3bfa0a22e20060 Mon Sep 17 00:00:00 2001 From: Jingju Hou Date: Sun, 22 Jan 2017 18:20:56 +0800 Subject: [PATCH 172/343] net: phy: marvell: Add Wake from LAN support for 88E1510 PHY Signed-off-by: Jingju Hou Signed-off-by: David S. Miller --- drivers/net/phy/marvell.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/net/phy/marvell.c b/drivers/net/phy/marvell.c index 0b78210c0fa7..ed0d235cf850 100644 --- a/drivers/net/phy/marvell.c +++ b/drivers/net/phy/marvell.c @@ -1679,6 +1679,8 @@ static struct phy_driver marvell_drivers[] = { .ack_interrupt = &marvell_ack_interrupt, .config_intr = &marvell_config_intr, .did_interrupt = &m88e1121_did_interrupt, + .get_wol = &m88e1318_get_wol, + .set_wol = &m88e1318_set_wol, .resume = &marvell_resume, .suspend = &marvell_suspend, .get_sset_count = marvell_get_sset_count, From 059aa734824165507c65fd30a55ff000afd14983 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 22 Jan 2017 14:04:29 -0500 Subject: [PATCH 173/343] nfs: Don't increment lock sequence ID after NFS4ERR_MOVED Xuan Qi reports that the Linux NFSv4 client failed to lock a file that was migrated. The steps he observed on the wire: 1. The client sent a LOCK request to the source server 2. The source server replied NFS4ERR_MOVED 3. The client switched to the destination server 4. The client sent the same LOCK request to the destination server with a bumped lock sequence ID 5. The destination server rejected the LOCK request with NFS4ERR_BAD_SEQID RFC 3530 section 8.1.5 provides a list of NFS errors which do not bump a lock sequence ID. However, RFC 3530 is now obsoleted by RFC 7530. In RFC 7530 section 9.1.7, this list has been updated by the addition of NFS4ERR_MOVED. Reported-by: Xuan Qi Signed-off-by: Chuck Lever Cc: stable@vger.kernel.org # v3.7+ Signed-off-by: Trond Myklebust --- include/linux/nfs4.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/include/linux/nfs4.h b/include/linux/nfs4.h index bca536341d1a..1b1ca04820a3 100644 --- a/include/linux/nfs4.h +++ b/include/linux/nfs4.h @@ -282,7 +282,7 @@ enum nfsstat4 { static inline bool seqid_mutating_err(u32 err) { - /* rfc 3530 section 8.1.5: */ + /* See RFC 7530, section 9.1.7 */ switch (err) { case NFS4ERR_STALE_CLIENTID: case NFS4ERR_STALE_STATEID: @@ -291,6 +291,7 @@ static inline bool seqid_mutating_err(u32 err) case NFS4ERR_BADXDR: case NFS4ERR_RESOURCE: case NFS4ERR_NOFILEHANDLE: + case NFS4ERR_MOVED: return false; }; return true; From a430607b2ef7c3be090f88c71cfcb1b3988aa7c0 Mon Sep 17 00:00:00 2001 From: Benjamin Coddington Date: Tue, 24 Jan 2017 11:34:20 -0500 Subject: [PATCH 174/343] NFSv4.0: always send mode in SETATTR after EXCLUSIVE4 Some nfsv4.0 servers may return a mode for the verifier following an open with EXCLUSIVE4 createmode, but this does not mean the client should skip setting the mode in the following SETATTR. It should only do that for EXCLUSIVE4_1 or UNGAURDED createmode. Fixes: 5334c5bdac92 ("NFS: Send attributes in OPEN request for NFS4_CREATE_EXCLUSIVE4_1") Signed-off-by: Benjamin Coddington Cc: stable@vger.kernel.org # v4.3+ Signed-off-by: Trond Myklebust --- fs/nfs/nfs4proc.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 59bb574d7d7c..0a0eaecf9676 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -2700,7 +2700,8 @@ static inline void nfs4_exclusive_attrset(struct nfs4_opendata *opendata, sattr->ia_valid |= ATTR_MTIME; /* Except MODE, it seems harmless of setting twice. */ - if ((attrset[1] & FATTR4_WORD1_MODE)) + if (opendata->o_arg.createmode != NFS4_CREATE_EXCLUSIVE && + attrset[1] & FATTR4_WORD1_MODE) sattr->ia_valid &= ~ATTR_MODE; if (attrset[2] & FATTR4_WORD2_SECURITY_LABEL) From 92fdb527eecff7e5eb945a3fbf4743110f5c1171 Mon Sep 17 00:00:00 2001 From: Yuriy Kolerov Date: Wed, 28 Dec 2016 11:46:26 +0300 Subject: [PATCH 175/343] ARCv2: MCIP: Deprecate setting of affinity in Device Tree Ignore value of interrupt distribution mode for common interrupts in IDU since setting of affinity using value from Device Tree is deprecated in ARC. Originally it is done in idu_irq_xlate() function and it is semantically wrong and does not guaranty that an affinity value will be set properly. idu_irq_enable() function is better place for initialization of common interrupts. By default send all common interrupts to all available online CPUs. The affinity of common interrupts in IDU must be set manually since in some cases the kernel will not call irq_set_affinity() by itself: 1. When the kernel is not configured with support of SMP. 2. When the kernel is configured with support of SMP but upper interrupt controllers does not support setting of the affinity and cannot propagate it to IDU. Signed-off-by: Yuriy Kolerov Signed-off-by: Vineet Gupta --- .../snps,archs-idu-intc.txt | 3 ++ arch/arc/kernel/mcip.c | 52 ++++++++----------- 2 files changed, 25 insertions(+), 30 deletions(-) diff --git a/Documentation/devicetree/bindings/interrupt-controller/snps,archs-idu-intc.txt b/Documentation/devicetree/bindings/interrupt-controller/snps,archs-idu-intc.txt index 0dcb7c7d3e40..944657684d73 100644 --- a/Documentation/devicetree/bindings/interrupt-controller/snps,archs-idu-intc.txt +++ b/Documentation/devicetree/bindings/interrupt-controller/snps,archs-idu-intc.txt @@ -15,6 +15,9 @@ Properties: Second cell specifies the irq distribution mode to cores 0=Round Robin; 1=cpu0, 2=cpu1, 4=cpu2, 8=cpu3 + The second cell in interrupts property is deprecated and may be ignored by + the kernel. + intc accessed via the special ARC AUX register interface, hence "reg" property is not specified. diff --git a/arch/arc/kernel/mcip.c b/arch/arc/kernel/mcip.c index 9274f8ade8c7..9988b427a1e0 100644 --- a/arch/arc/kernel/mcip.c +++ b/arch/arc/kernel/mcip.c @@ -175,7 +175,6 @@ static void idu_irq_unmask(struct irq_data *data) raw_spin_unlock_irqrestore(&mcip_lock, flags); } -#ifdef CONFIG_SMP static int idu_irq_set_affinity(struct irq_data *data, const struct cpumask *cpumask, bool force) @@ -205,12 +204,27 @@ idu_irq_set_affinity(struct irq_data *data, const struct cpumask *cpumask, return IRQ_SET_MASK_OK; } -#endif + +static void idu_irq_enable(struct irq_data *data) +{ + /* + * By default send all common interrupts to all available online CPUs. + * The affinity of common interrupts in IDU must be set manually since + * in some cases the kernel will not call irq_set_affinity() by itself: + * 1. When the kernel is not configured with support of SMP. + * 2. When the kernel is configured with support of SMP but upper + * interrupt controllers does not support setting of the affinity + * and cannot propagate it to IDU. + */ + idu_irq_set_affinity(data, cpu_online_mask, false); + idu_irq_unmask(data); +} static struct irq_chip idu_irq_chip = { .name = "MCIP IDU Intc", .irq_mask = idu_irq_mask, .irq_unmask = idu_irq_unmask, + .irq_enable = idu_irq_enable, #ifdef CONFIG_SMP .irq_set_affinity = idu_irq_set_affinity, #endif @@ -243,36 +257,14 @@ static int idu_irq_xlate(struct irq_domain *d, struct device_node *n, const u32 *intspec, unsigned int intsize, irq_hw_number_t *out_hwirq, unsigned int *out_type) { - irq_hw_number_t hwirq = *out_hwirq = intspec[0]; - int distri = intspec[1]; - unsigned long flags; - + /* + * Ignore value of interrupt distribution mode for common interrupts in + * IDU which resides in intspec[1] since setting an affinity using value + * from Device Tree is deprecated in ARC. + */ + *out_hwirq = intspec[0]; *out_type = IRQ_TYPE_NONE; - /* XXX: validate distribution scheme again online cpu mask */ - if (distri == 0) { - /* 0 - Round Robin to all cpus, otherwise 1 bit per core */ - raw_spin_lock_irqsave(&mcip_lock, flags); - idu_set_dest(hwirq, BIT(num_online_cpus()) - 1); - idu_set_mode(hwirq, IDU_M_TRIG_LEVEL, IDU_M_DISTRI_RR); - raw_spin_unlock_irqrestore(&mcip_lock, flags); - } else { - /* - * DEST based distribution for Level Triggered intr can only - * have 1 CPU, so generalize it to always contain 1 cpu - */ - int cpu = ffs(distri); - - if (cpu != fls(distri)) - pr_warn("IDU irq %lx distri mode set to cpu %x\n", - hwirq, cpu); - - raw_spin_lock_irqsave(&mcip_lock, flags); - idu_set_dest(hwirq, cpu); - idu_set_mode(hwirq, IDU_M_TRIG_LEVEL, IDU_M_DISTRI_DEST); - raw_spin_unlock_irqrestore(&mcip_lock, flags); - } - return 0; } From 7630ea4bda18df2ee1c64dfdca1724a9cc32f920 Mon Sep 17 00:00:00 2001 From: Martin Blumenstingl Date: Sun, 22 Jan 2017 17:41:32 +0100 Subject: [PATCH 176/343] Documentation: net: phy: improve explanation when to specify the PHY ID The old description basically read like "ethernet-phy-idAAAA.BBBB" can be specified when you know the actual PHY ID. However, specifying this has a side-effect: it forces Linux to bind to a certain PHY driver (the one that matches the ID given in the compatible string), ignoring the ID which is reported by the actual PHY. Whenever a device is shipped with (multiple) different PHYs during it's production lifetime then explicitly specifying "ethernet-phy-idAAAA.BBBB" could break certain revisions of that device. Signed-off-by: Martin Blumenstingl Reviewed-by: Andrew Lunn Acked-by: Rob Herring Reviewed-by: Florian Fainelli Signed-off-by: David S. Miller --- Documentation/devicetree/bindings/net/phy.txt | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/Documentation/devicetree/bindings/net/phy.txt b/Documentation/devicetree/bindings/net/phy.txt index ff1bc4b1bb3b..fb5056b22685 100644 --- a/Documentation/devicetree/bindings/net/phy.txt +++ b/Documentation/devicetree/bindings/net/phy.txt @@ -19,8 +19,9 @@ Optional Properties: specifications. If neither of these are specified, the default is to assume clause 22. - If the phy's identifier is known then the list may contain an entry - of the form: "ethernet-phy-idAAAA.BBBB" where + If the PHY reports an incorrect ID (or none at all) then the + "compatible" list may contain an entry with the correct PHY ID in the + form: "ethernet-phy-idAAAA.BBBB" where AAAA - The value of the 16 bit Phy Identifier 1 register as 4 hex digits. This is the chip vendor OUI bits 3:18 BBBB - The value of the 16 bit Phy Identifier 2 register as From 6a0b76c04ec157c88ca943debf78a8ee58469f2d Mon Sep 17 00:00:00 2001 From: hayeswang Date: Mon, 23 Jan 2017 14:18:43 +0800 Subject: [PATCH 177/343] r8152: don't execute runtime suspend if the tx is not empty Runtime suspend shouldn't be executed if the tx queue is not empty, because the device is not idle. Signed-off-by: Hayes Wang Signed-off-by: David S. Miller --- drivers/net/usb/r8152.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/net/usb/r8152.c b/drivers/net/usb/r8152.c index 0e99af090734..e1466b4d2b6c 100644 --- a/drivers/net/usb/r8152.c +++ b/drivers/net/usb/r8152.c @@ -32,7 +32,7 @@ #define NETNEXT_VERSION "08" /* Information for net */ -#define NET_VERSION "6" +#define NET_VERSION "7" #define DRIVER_VERSION "v1." NETNEXT_VERSION "." NET_VERSION #define DRIVER_AUTHOR "Realtek linux nic maintainers " @@ -3574,6 +3574,8 @@ static bool delay_autosuspend(struct r8152 *tp) */ if (!sw_linking && tp->rtl_ops.in_nway(tp)) return true; + else if (!skb_queue_empty(&tp->tx_queue)) + return true; else return false; } From a59b7e0246774e28193126fe7fdbbd0ae9c67dcc Mon Sep 17 00:00:00 2001 From: Ido Schimmel Date: Mon, 23 Jan 2017 11:11:42 +0100 Subject: [PATCH 178/343] mlxsw: spectrum_router: Correctly reallocate adjacency entries mlxsw_sp_nexthop_group_mac_update() is called in one of two cases: 1) When the MAC of a nexthop needs to be updated 2) When the size of a nexthop group has changed In the second case the adjacency entries for the nexthop group need to be reallocated from the adjacency table. In this case we must write to the entries the MAC addresses of all the nexthops that should be offloaded and not only those whose MAC changed. Otherwise, these entries would be filled with garbage data, resulting in packet loss. Fixes: a7ff87acd995 ("mlxsw: spectrum_router: Implement next-hop routing") Signed-off-by: Ido Schimmel Signed-off-by: Jiri Pirko Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c index 01d0efa9c5c7..9e494a446b7e 100644 --- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c +++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c @@ -1172,7 +1172,8 @@ static int mlxsw_sp_nexthop_mac_update(struct mlxsw_sp *mlxsw_sp, u32 adj_index, static int mlxsw_sp_nexthop_group_mac_update(struct mlxsw_sp *mlxsw_sp, - struct mlxsw_sp_nexthop_group *nh_grp) + struct mlxsw_sp_nexthop_group *nh_grp, + bool reallocate) { u32 adj_index = nh_grp->adj_index; /* base */ struct mlxsw_sp_nexthop *nh; @@ -1187,7 +1188,7 @@ mlxsw_sp_nexthop_group_mac_update(struct mlxsw_sp *mlxsw_sp, continue; } - if (nh->update) { + if (nh->update || reallocate) { err = mlxsw_sp_nexthop_mac_update(mlxsw_sp, adj_index, nh); if (err) @@ -1248,7 +1249,8 @@ mlxsw_sp_nexthop_group_refresh(struct mlxsw_sp *mlxsw_sp, /* Nothing was added or removed, so no need to reallocate. Just * update MAC on existing adjacency indexes. */ - err = mlxsw_sp_nexthop_group_mac_update(mlxsw_sp, nh_grp); + err = mlxsw_sp_nexthop_group_mac_update(mlxsw_sp, nh_grp, + false); if (err) { dev_warn(mlxsw_sp->bus_info->dev, "Failed to update neigh MAC in adjacency table.\n"); goto set_trap; @@ -1276,7 +1278,7 @@ mlxsw_sp_nexthop_group_refresh(struct mlxsw_sp *mlxsw_sp, nh_grp->adj_index_valid = 1; nh_grp->adj_index = adj_index; nh_grp->ecmp_size = ecmp_size; - err = mlxsw_sp_nexthop_group_mac_update(mlxsw_sp, nh_grp); + err = mlxsw_sp_nexthop_group_mac_update(mlxsw_sp, nh_grp, true); if (err) { dev_warn(mlxsw_sp->bus_info->dev, "Failed to update neigh MAC in adjacency table.\n"); goto set_trap; From 36425cd67052e3becf325fd4d3ba5691791ef7e4 Mon Sep 17 00:00:00 2001 From: Vineet Gupta Date: Tue, 24 Jan 2017 10:23:42 -0800 Subject: [PATCH 179/343] ARC: udelay: fix inline assembler by adding LP_COUNT to clobber list commit 3c7c7a2fc8811bc ("ARC: Don't use "+l" inline asm constraint") modified the inline assembly to setup LP_COUNT register manually and NOT rely on gcc to do it (with the +l inline assembler contraint hint, now being retired in the compiler) However the fix was flawed as we didn't add LP_COUNT to asm clobber list, meaning gcc doesn't know that LP_COUNT or zero-delay-loops are in action in the inline asm. This resulted in some fun - as nested ZOL loops were being generared | mov lp_count,250000 ;16 # tmp235, | lp .L__GCC__LP14 # <======= OUTER LOOP (gcc generated) | .L14: | ld r2, [r5] # MEM[(volatile u32 *)prephitmp_43], w | dmb 1 | breq r2, -1, @.L21 #, w,, | bbit0 r2,1,@.L13 # w,, | ld r4,[r7] ;25 # loops_per_jiffy, loops_per_jiffy | mpymu r3,r4,r6 #, loops_per_jiffy, tmp234 | | mov lp_count, r3 # <====== INNER LOOP (from inline asm) | lp 1f | nop | 1: | nop_s | .L__GCC__LP14: ; loop end, start is @.L14 #, This caused issues with drivers relying on sane behaviour of udelay friends. With LP_COUNT added to clobber list, gcc doesn't generate the outer loop in say above case. Addresses STAR 9001146134 Reported-by: Joao Pinto Fixes: 3c7c7a2fc8811bc ("ARC: Don't use "+l" inline asm constraint") Cc: stable@vger.kernel.org Signed-off-by: Vineet Gupta --- arch/arc/include/asm/delay.h | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/arch/arc/include/asm/delay.h b/arch/arc/include/asm/delay.h index a36e8601114d..d5da2115d78a 100644 --- a/arch/arc/include/asm/delay.h +++ b/arch/arc/include/asm/delay.h @@ -26,7 +26,9 @@ static inline void __delay(unsigned long loops) " lp 1f \n" " nop \n" "1: \n" - : : "r"(loops)); + : + : "r"(loops) + : "lp_count"); } extern void __bad_udelay(void); From 517e7610d2ce04d1b8d8b6c6d1a36dcce5cac6ab Mon Sep 17 00:00:00 2001 From: Vineet Gupta Date: Thu, 19 Jan 2017 17:05:00 -0800 Subject: [PATCH 180/343] ARCv2: MCIP: update the BCR per current changes Signed-off-by: Vineet Gupta --- arch/arc/kernel/mcip.c | 3 +-- include/soc/arc/mcip.h | 16 ++++++++-------- 2 files changed, 9 insertions(+), 10 deletions(-) diff --git a/arch/arc/kernel/mcip.c b/arch/arc/kernel/mcip.c index 9988b427a1e0..9f6b68fd4f3b 100644 --- a/arch/arc/kernel/mcip.c +++ b/arch/arc/kernel/mcip.c @@ -93,11 +93,10 @@ static void mcip_probe_n_setup(void) READ_BCR(ARC_REG_MCIP_BCR, mp); sprintf(smp_cpuinfo_buf, - "Extn [SMP]\t: ARConnect (v%d): %d cores with %s%s%s%s%s\n", + "Extn [SMP]\t: ARConnect (v%d): %d cores with %s%s%s%s\n", mp.ver, mp.num_cores, IS_AVAIL1(mp.ipi, "IPI "), IS_AVAIL1(mp.idu, "IDU "), - IS_AVAIL1(mp.llm, "LLM "), IS_AVAIL1(mp.dbg, "DEBUG "), IS_AVAIL1(mp.gfrc, "GFRC")); diff --git a/include/soc/arc/mcip.h b/include/soc/arc/mcip.h index 6902c2a8bd23..4b6b489a8d7c 100644 --- a/include/soc/arc/mcip.h +++ b/include/soc/arc/mcip.h @@ -55,17 +55,17 @@ struct mcip_cmd { struct mcip_bcr { #ifdef CONFIG_CPU_BIG_ENDIAN - unsigned int pad3:8, - idu:1, llm:1, num_cores:6, - iocoh:1, gfrc:1, dbg:1, pad2:1, - msg:1, sem:1, ipi:1, pad:1, + unsigned int pad4:6, pw_dom:1, pad3:1, + idu:1, pad2:1, num_cores:6, + pad:1, gfrc:1, dbg:1, pw:1, + msg:1, sem:1, ipi:1, slv:1, ver:8; #else unsigned int ver:8, - pad:1, ipi:1, sem:1, msg:1, - pad2:1, dbg:1, gfrc:1, iocoh:1, - num_cores:6, llm:1, idu:1, - pad3:8; + slv:1, ipi:1, sem:1, msg:1, + pw:1, dbg:1, gfrc:1, pad:1, + num_cores:6, pad2:1, idu:1, + pad3:1, pw_dom:1, pad4:6; #endif }; From bf02454a741b58682a82c314a9a46bed930ed2f7 Mon Sep 17 00:00:00 2001 From: Vineet Gupta Date: Thu, 12 Jan 2017 14:30:29 -0800 Subject: [PATCH 181/343] ARC: smp-boot: Decouple Non masters waiting API from jump to entry point For run-on-reset SMP configs, non master cores call a routine which waits until Master gives it a "go" signal (currently using a shared mem flag). The same routine then jumps off the well known entry point of all non Master cores i.e. @first_lines_of_secondary This patch moves out the last part into one single place in early boot code. This is better in terms of absraction (the wait API only waits) and returns, leaving out the "jump off to" part. In actual implementation this requires some restructuring of the early boot code as well as Master now jumps to BSS setup explicitly, vs. falling thru into it before. Technically this patch doesn't cause any functional change, it just moves the ugly #ifdef'ry from assembly code to "C" Signed-off-by: Vineet Gupta --- arch/arc/kernel/head.S | 14 +++++++------- arch/arc/kernel/smp.c | 6 ++++-- 2 files changed, 11 insertions(+), 9 deletions(-) diff --git a/arch/arc/kernel/head.S b/arch/arc/kernel/head.S index 689dd867fdff..8b90d25a15cc 100644 --- a/arch/arc/kernel/head.S +++ b/arch/arc/kernel/head.S @@ -71,14 +71,14 @@ ENTRY(stext) GET_CPU_ID r5 cmp r5, 0 mov.nz r0, r5 -#ifdef CONFIG_ARC_SMP_HALT_ON_RESET - ; Non-Master can proceed as system would be booted sufficiently - jnz first_lines_of_secondary -#else + bz .Lmaster_proceed + ; Non-Masters wait for Master to boot enough and bring them up - jnz arc_platform_smp_wait_to_boot -#endif - ; Master falls thru + ; when they resume, tail-call to entry point + mov blink, @first_lines_of_secondary + j arc_platform_smp_wait_to_boot + +.Lmaster_proceed: #endif ; Clear BSS before updating any globals diff --git a/arch/arc/kernel/smp.c b/arch/arc/kernel/smp.c index 88674d972c9d..44a0d21ed342 100644 --- a/arch/arc/kernel/smp.c +++ b/arch/arc/kernel/smp.c @@ -98,14 +98,16 @@ static void arc_default_smp_cpu_kick(int cpu, unsigned long pc) void arc_platform_smp_wait_to_boot(int cpu) { + /* for halt-on-reset, we've waited already */ + if (IS_ENABLED(CONFIG_ARC_SMP_HALT_ON_RESET)) + return; + while (wake_flag != cpu) ; wake_flag = 0; - __asm__ __volatile__("j @first_lines_of_secondary \n"); } - const char *arc_platform_smp_cpuinfo(void) { return plat_smp_ops.info ? : ""; From 7d211c81e97ef8505610ef82e14e302ab415bad1 Mon Sep 17 00:00:00 2001 From: Adit Ranadive Date: Thu, 19 Jan 2017 13:20:39 -0800 Subject: [PATCH 182/343] IB/vmw_pvrdma: Don't leak info from alloc_ucontext Clear out the user response struct correctly. Fixes: 29c8d9eba550 ("IB: Add vmw_pvrdma driver") Reported-by: Dan Carpenter Signed-off-by: Adit Ranadive Signed-off-by: Doug Ledford --- drivers/infiniband/hw/vmw_pvrdma/pvrdma_verbs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/infiniband/hw/vmw_pvrdma/pvrdma_verbs.c b/drivers/infiniband/hw/vmw_pvrdma/pvrdma_verbs.c index 54891370d18a..c2aa52638dcb 100644 --- a/drivers/infiniband/hw/vmw_pvrdma/pvrdma_verbs.c +++ b/drivers/infiniband/hw/vmw_pvrdma/pvrdma_verbs.c @@ -306,7 +306,7 @@ struct ib_ucontext *pvrdma_alloc_ucontext(struct ib_device *ibdev, union pvrdma_cmd_resp rsp; struct pvrdma_cmd_create_uc *cmd = &req.create_uc; struct pvrdma_cmd_create_uc_resp *resp = &rsp.create_uc_resp; - struct pvrdma_alloc_ucontext_resp uresp; + struct pvrdma_alloc_ucontext_resp uresp = {0}; int ret; void *ptr; From ff89b070b7c98eb6782361310ca7a15186f15b2c Mon Sep 17 00:00:00 2001 From: Adit Ranadive Date: Thu, 19 Jan 2017 13:20:40 -0800 Subject: [PATCH 183/343] IB/vmw_pvrdma: Fix incorrect cleanup on pvrdma_pci_probe error path If the interrupt allocation failed we should start freeing the CQ rings rather than unregistering the netdev notifier. Fixes: 29c8d9eba550 ("IB: Add vmw_pvrdma driver") Signed-off-by: Adit Ranadive Reviewed-by: Yuval Shaia Signed-off-by: Doug Ledford --- drivers/infiniband/hw/vmw_pvrdma/pvrdma_main.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/drivers/infiniband/hw/vmw_pvrdma/pvrdma_main.c b/drivers/infiniband/hw/vmw_pvrdma/pvrdma_main.c index 231a1ce1f4be..bd8fbd3d2032 100644 --- a/drivers/infiniband/hw/vmw_pvrdma/pvrdma_main.c +++ b/drivers/infiniband/hw/vmw_pvrdma/pvrdma_main.c @@ -1029,7 +1029,7 @@ static int pvrdma_pci_probe(struct pci_dev *pdev, if (ret) { dev_err(&pdev->dev, "failed to allocate interrupts\n"); ret = -ENOMEM; - goto err_netdevice; + goto err_free_cq_ring; } /* Allocate UAR table. */ @@ -1092,8 +1092,6 @@ err_free_uar_table: err_free_intrs: pvrdma_free_irq(dev); pvrdma_disable_msi_all(dev); -err_netdevice: - unregister_netdevice_notifier(&dev->nb_netdev); err_free_cq_ring: pvrdma_page_dir_cleanup(dev, &dev->cq_pdir); err_free_async_ring: From d46d29f072accb069cb42b5fbebcc77d9094a785 Mon Sep 17 00:00:00 2001 From: Shaohua Li Date: Wed, 11 Jan 2017 13:38:52 -0800 Subject: [PATCH 184/343] md/raid5-cache: delete meaningless code sector_t is unsigned long, it's never < 0 Reported-by: Julia Lawall Signed-off-by: Shaohua Li --- drivers/md/raid5-cache.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/drivers/md/raid5-cache.c b/drivers/md/raid5-cache.c index 0e8ed2c327b0..e0d6dd1835d0 100644 --- a/drivers/md/raid5-cache.c +++ b/drivers/md/raid5-cache.c @@ -1393,8 +1393,6 @@ static void r5l_do_reclaim(struct r5l_log *log) next_checkpoint = r5c_calculate_new_cp(conf); spin_unlock_irq(&log->io_list_lock); - BUG_ON(reclaimable < 0); - if (reclaimable == 0 || !write_super) return; From 86aa1397ddfde563b3692adadb8b8e32e97b4e5e Mon Sep 17 00:00:00 2001 From: Song Liu Date: Thu, 12 Jan 2017 17:22:41 -0800 Subject: [PATCH 185/343] md/r5cache: read data into orig_page for prexor of cached data With write back cache, we use orig_page to do prexor. This patch makes sure we read data into orig_page for it. Flag R5_OrigPageUPTDODATE is added to show whether orig_page has the latest data from raid disk. We introduce a helper function uptodate_for_rmw() to simplify the a couple conditions in handle_stripe_dirtying(). Signed-off-by: Song Liu Signed-off-by: Shaohua Li --- drivers/md/raid5-cache.c | 2 ++ drivers/md/raid5.c | 44 ++++++++++++++++++++++++++++++++-------- drivers/md/raid5.h | 5 +++++ 3 files changed, 42 insertions(+), 9 deletions(-) diff --git a/drivers/md/raid5-cache.c b/drivers/md/raid5-cache.c index e0d6dd1835d0..95dcaa022e1f 100644 --- a/drivers/md/raid5-cache.c +++ b/drivers/md/raid5-cache.c @@ -2349,6 +2349,8 @@ void r5c_release_extra_page(struct stripe_head *sh) struct page *p = sh->dev[i].orig_page; sh->dev[i].orig_page = sh->dev[i].page; + clear_bit(R5_OrigPageUPTDODATE, &sh->dev[i].flags); + if (!using_disk_info_extra_page) put_page(p); } diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c index 36c13e4be9c9..7780ae44f355 100644 --- a/drivers/md/raid5.c +++ b/drivers/md/raid5.c @@ -1015,7 +1015,17 @@ again: if (test_bit(R5_SkipCopy, &sh->dev[i].flags)) WARN_ON(test_bit(R5_UPTODATE, &sh->dev[i].flags)); - sh->dev[i].vec.bv_page = sh->dev[i].page; + + if (!op_is_write(op) && + test_bit(R5_InJournal, &sh->dev[i].flags)) + /* + * issuing read for a page in journal, this + * must be preparing for prexor in rmw; read + * the data into orig_page + */ + sh->dev[i].vec.bv_page = sh->dev[i].orig_page; + else + sh->dev[i].vec.bv_page = sh->dev[i].page; bi->bi_vcnt = 1; bi->bi_io_vec[0].bv_len = STRIPE_SIZE; bi->bi_io_vec[0].bv_offset = 0; @@ -2380,6 +2390,13 @@ static void raid5_end_read_request(struct bio * bi) } else if (test_bit(R5_ReadNoMerge, &sh->dev[i].flags)) clear_bit(R5_ReadNoMerge, &sh->dev[i].flags); + if (test_bit(R5_InJournal, &sh->dev[i].flags)) + /* + * end read for a page in journal, this + * must be preparing for prexor in rmw + */ + set_bit(R5_OrigPageUPTDODATE, &sh->dev[i].flags); + if (atomic_read(&rdev->read_errors)) atomic_set(&rdev->read_errors, 0); } else { @@ -3594,6 +3611,21 @@ unhash: break_stripe_batch_list(head_sh, STRIPE_EXPAND_SYNC_FLAGS); } +/* + * For RMW in write back cache, we need extra page in prexor to store the + * old data. This page is stored in dev->orig_page. + * + * This function checks whether we have data for prexor. The exact logic + * is: + * R5_UPTODATE && (!R5_InJournal || R5_OrigPageUPTDODATE) + */ +static inline bool uptodate_for_rmw(struct r5dev *dev) +{ + return (test_bit(R5_UPTODATE, &dev->flags)) && + (!test_bit(R5_InJournal, &dev->flags) || + test_bit(R5_OrigPageUPTDODATE, &dev->flags)); +} + static int handle_stripe_dirtying(struct r5conf *conf, struct stripe_head *sh, struct stripe_head_state *s, @@ -3625,9 +3657,7 @@ static int handle_stripe_dirtying(struct r5conf *conf, if ((dev->towrite || i == sh->pd_idx || i == sh->qd_idx || test_bit(R5_InJournal, &dev->flags)) && !test_bit(R5_LOCKED, &dev->flags) && - !((test_bit(R5_UPTODATE, &dev->flags) && - (!test_bit(R5_InJournal, &dev->flags) || - dev->page != dev->orig_page)) || + !(uptodate_for_rmw(dev) || test_bit(R5_Wantcompute, &dev->flags))) { if (test_bit(R5_Insync, &dev->flags)) rmw++; @@ -3639,7 +3669,6 @@ static int handle_stripe_dirtying(struct r5conf *conf, i != sh->pd_idx && i != sh->qd_idx && !test_bit(R5_LOCKED, &dev->flags) && !(test_bit(R5_UPTODATE, &dev->flags) || - test_bit(R5_InJournal, &dev->flags) || test_bit(R5_Wantcompute, &dev->flags))) { if (test_bit(R5_Insync, &dev->flags)) rcw++; @@ -3693,9 +3722,7 @@ static int handle_stripe_dirtying(struct r5conf *conf, i == sh->pd_idx || i == sh->qd_idx || test_bit(R5_InJournal, &dev->flags)) && !test_bit(R5_LOCKED, &dev->flags) && - !((test_bit(R5_UPTODATE, &dev->flags) && - (!test_bit(R5_InJournal, &dev->flags) || - dev->page != dev->orig_page)) || + !(uptodate_for_rmw(dev) || test_bit(R5_Wantcompute, &dev->flags)) && test_bit(R5_Insync, &dev->flags)) { if (test_bit(STRIPE_PREREAD_ACTIVE, @@ -3722,7 +3749,6 @@ static int handle_stripe_dirtying(struct r5conf *conf, i != sh->pd_idx && i != sh->qd_idx && !test_bit(R5_LOCKED, &dev->flags) && !(test_bit(R5_UPTODATE, &dev->flags) || - test_bit(R5_InJournal, &dev->flags) || test_bit(R5_Wantcompute, &dev->flags))) { rcw++; if (test_bit(R5_Insync, &dev->flags) && diff --git a/drivers/md/raid5.h b/drivers/md/raid5.h index ed8e1362ab36..461df197d157 100644 --- a/drivers/md/raid5.h +++ b/drivers/md/raid5.h @@ -322,6 +322,11 @@ enum r5dev_flags { * data and parity being written are in the journal * device */ + R5_OrigPageUPTDODATE, /* with write back cache, we read old data into + * dev->orig_page for prexor. When this flag is + * set, orig_page contains latest data in the + * raid disk. + */ }; /* From ba02684daf7fb4a827580f909b7c7db61c05ae7d Mon Sep 17 00:00:00 2001 From: Song Liu Date: Thu, 12 Jan 2017 17:22:42 -0800 Subject: [PATCH 186/343] md/raid5: move comment of fetch_block to right location Signed-off-by: Song Liu Signed-off-by: Shaohua Li --- drivers/md/raid5.c | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-) diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c index 7780ae44f355..13d76767c2cf 100644 --- a/drivers/md/raid5.c +++ b/drivers/md/raid5.c @@ -3312,13 +3312,6 @@ static int want_replace(struct stripe_head *sh, int disk_idx) return rv; } -/* fetch_block - checks the given member device to see if its data needs - * to be read or computed to satisfy a request. - * - * Returns 1 when no more member devices need to be checked, otherwise returns - * 0 to tell the loop in handle_stripe_fill to continue - */ - static int need_this_block(struct stripe_head *sh, struct stripe_head_state *s, int disk_idx, int disks) { @@ -3409,6 +3402,12 @@ static int need_this_block(struct stripe_head *sh, struct stripe_head_state *s, return 0; } +/* fetch_block - checks the given member device to see if its data needs + * to be read or computed to satisfy a request. + * + * Returns 1 when no more member devices need to be checked, otherwise returns + * 0 to tell the loop in handle_stripe_fill to continue + */ static int fetch_block(struct stripe_head *sh, struct stripe_head_state *s, int disk_idx, int disks) { From a85dd7b8df52e35d8ee3794c65cac5c39128fd80 Mon Sep 17 00:00:00 2001 From: Song Liu Date: Mon, 23 Jan 2017 17:12:57 -0800 Subject: [PATCH 187/343] md/r5cache: flush data only stripes in r5l_recovery_log() For safer operation, all arrays start in write-through mode, which has been better tested and is more mature. And actually the write-through/write-mode isn't persistent after array restarted, so we always start array in write-through mode. However, if recovery found data-only stripes before the shutdown (from previous write-back mode), it is not safe to start the array in write-through mode, as write-through mode can not handle stripes with data in write-back cache. To solve this problem, we flush all data-only stripes in r5l_recovery_log(). When r5l_recovery_log() returns, the array starts with empty cache in write-through mode. This logic is implemented in r5c_recovery_flush_data_only_stripes(): 1. enable write back cache 2. flush all stripes 3. wake up conf->mddev->thread 4. wait for all stripes get flushed (reuse wait_for_quiescent) 5. disable write back cache The wait in 4 will be waked up in release_inactive_stripe_list() when conf->active_stripes reaches 0. It is safe to wake up mddev->thread here because all the resource required for the thread has been initialized. Signed-off-by: Song Liu Signed-off-by: Shaohua Li --- drivers/md/md.c | 5 ++++ drivers/md/raid5-cache.c | 56 ++++++++++++++++++++++++++++------------ 2 files changed, 45 insertions(+), 16 deletions(-) diff --git a/drivers/md/md.c b/drivers/md/md.c index 82821ee0d57f..01175dac0db6 100644 --- a/drivers/md/md.c +++ b/drivers/md/md.c @@ -5291,6 +5291,11 @@ int md_run(struct mddev *mddev) if (start_readonly && mddev->ro == 0) mddev->ro = 2; /* read-only, but switch on first write */ + /* + * NOTE: some pers->run(), for example r5l_recovery_log(), wakes + * up mddev->thread. It is important to initialize critical + * resources for mddev->thread BEFORE calling pers->run(). + */ err = pers->run(mddev); if (err) pr_warn("md: pers->run() failed ...\n"); diff --git a/drivers/md/raid5-cache.c b/drivers/md/raid5-cache.c index 95dcaa022e1f..3d7dda85494c 100644 --- a/drivers/md/raid5-cache.c +++ b/drivers/md/raid5-cache.c @@ -2060,7 +2060,7 @@ static int r5c_recovery_rewrite_data_only_stripes(struct r5l_log *log, struct r5l_recovery_ctx *ctx) { - struct stripe_head *sh, *next; + struct stripe_head *sh; struct mddev *mddev = log->rdev->mddev; struct page *page; sector_t next_checkpoint = MaxSector; @@ -2074,7 +2074,7 @@ r5c_recovery_rewrite_data_only_stripes(struct r5l_log *log, WARN_ON(list_empty(&ctx->cached_list)); - list_for_each_entry_safe(sh, next, &ctx->cached_list, lru) { + list_for_each_entry(sh, &ctx->cached_list, lru) { struct r5l_meta_block *mb; int i; int offset; @@ -2124,14 +2124,39 @@ r5c_recovery_rewrite_data_only_stripes(struct r5l_log *log, ctx->pos = write_pos; ctx->seq += 1; next_checkpoint = sh->log_start; - list_del_init(&sh->lru); - raid5_release_stripe(sh); } log->next_checkpoint = next_checkpoint; __free_page(page); return 0; } +static void r5c_recovery_flush_data_only_stripes(struct r5l_log *log, + struct r5l_recovery_ctx *ctx) +{ + struct mddev *mddev = log->rdev->mddev; + struct r5conf *conf = mddev->private; + struct stripe_head *sh, *next; + + if (ctx->data_only_stripes == 0) + return; + + log->r5c_journal_mode = R5C_JOURNAL_MODE_WRITE_BACK; + + list_for_each_entry_safe(sh, next, &ctx->cached_list, lru) { + r5c_make_stripe_write_out(sh); + set_bit(STRIPE_HANDLE, &sh->state); + list_del_init(&sh->lru); + raid5_release_stripe(sh); + } + + md_wakeup_thread(conf->mddev->thread); + /* reuse conf->wait_for_quiescent in recovery */ + wait_event(conf->wait_for_quiescent, + atomic_read(&conf->active_stripes) == 0); + + log->r5c_journal_mode = R5C_JOURNAL_MODE_WRITE_THROUGH; +} + static int r5l_recovery_log(struct r5l_log *log) { struct mddev *mddev = log->rdev->mddev; @@ -2158,32 +2183,31 @@ static int r5l_recovery_log(struct r5l_log *log) pos = ctx.pos; ctx.seq += 10000; - if (ctx.data_only_stripes == 0) { - log->next_checkpoint = ctx.pos; - r5l_log_write_empty_meta_block(log, ctx.pos, ctx.seq++); - ctx.pos = r5l_ring_add(log, ctx.pos, BLOCK_SECTORS); - } if ((ctx.data_only_stripes == 0) && (ctx.data_parity_stripes == 0)) pr_debug("md/raid:%s: starting from clean shutdown\n", mdname(mddev)); - else { + else pr_debug("md/raid:%s: recovering %d data-only stripes and %d data-parity stripes\n", mdname(mddev), ctx.data_only_stripes, ctx.data_parity_stripes); - if (ctx.data_only_stripes > 0) - if (r5c_recovery_rewrite_data_only_stripes(log, &ctx)) { - pr_err("md/raid:%s: failed to rewrite stripes to journal\n", - mdname(mddev)); - return -EIO; - } + if (ctx.data_only_stripes == 0) { + log->next_checkpoint = ctx.pos; + r5l_log_write_empty_meta_block(log, ctx.pos, ctx.seq++); + ctx.pos = r5l_ring_add(log, ctx.pos, BLOCK_SECTORS); + } else if (r5c_recovery_rewrite_data_only_stripes(log, &ctx)) { + pr_err("md/raid:%s: failed to rewrite stripes to journal\n", + mdname(mddev)); + return -EIO; } log->log_start = ctx.pos; log->seq = ctx.seq; log->last_checkpoint = pos; r5l_write_super(log, pos); + + r5c_recovery_flush_data_only_stripes(log, &ctx); return 0; } From 07e83364845e1e1c7e189a01206a9d7d33831568 Mon Sep 17 00:00:00 2001 From: Song Liu Date: Mon, 23 Jan 2017 17:12:58 -0800 Subject: [PATCH 188/343] md/r5cache: shift complex rmw from read path to write path Write back cache requires a complex RMW mechanism, where old data is read into dev->orig_page for prexor, and then xor is done with dev->page. This logic is already implemented in the write path. However, current read path is not awared of this requirement. When the array is optimal, the RMW is not required, as the data are read from raid disks. However, when the target stripe is degraded, complex RMW is required to generate right data. To keep read path as clean as possible, we handle read path by flushing degraded, in-journal stripes before processing reads to missing dev. Specifically, when there is read requests to a degraded stripe with data in journal, handle_stripe_fill() calls r5c_make_stripe_write_out() and exits. Then handle_stripe_dirtying() will do the complex RMW and flush the stripe to RAID disks. After that, read requests are handled. There is one more corner case when there is non-overwrite bio for the missing (or out of sync) dev. handle_stripe_dirtying() will not be able to process the non-overwrite bios without constructing the data in handle_stripe_fill(). This is fixed by delaying non-overwrite bios in handle_stripe_dirtying(). So handle_stripe_fill() works on these bios after the stripe is flushed to raid disks. Signed-off-by: Song Liu Signed-off-by: Shaohua Li --- drivers/md/raid5.c | 49 ++++++++++++++++++++++++++++++++++++++++++---- 1 file changed, 45 insertions(+), 4 deletions(-) diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c index 13d76767c2cf..dc83da69ca7c 100644 --- a/drivers/md/raid5.c +++ b/drivers/md/raid5.c @@ -2897,6 +2897,30 @@ sector_t raid5_compute_blocknr(struct stripe_head *sh, int i, int previous) return r_sector; } +/* + * There are cases where we want handle_stripe_dirtying() and + * schedule_reconstruction() to delay towrite to some dev of a stripe. + * + * This function checks whether we want to delay the towrite. Specifically, + * we delay the towrite when: + * + * 1. degraded stripe has a non-overwrite to the missing dev, AND this + * stripe has data in journal (for other devices). + * + * In this case, when reading data for the non-overwrite dev, it is + * necessary to handle complex rmw of write back cache (prexor with + * orig_page, and xor with page). To keep read path simple, we would + * like to flush data in journal to RAID disks first, so complex rmw + * is handled in the write patch (handle_stripe_dirtying). + * + */ +static inline bool delay_towrite(struct r5dev *dev, + struct stripe_head_state *s) +{ + return !test_bit(R5_OVERWRITE, &dev->flags) && + !test_bit(R5_Insync, &dev->flags) && s->injournal; +} + static void schedule_reconstruction(struct stripe_head *sh, struct stripe_head_state *s, int rcw, int expand) @@ -2917,7 +2941,7 @@ schedule_reconstruction(struct stripe_head *sh, struct stripe_head_state *s, for (i = disks; i--; ) { struct r5dev *dev = &sh->dev[i]; - if (dev->towrite) { + if (dev->towrite && !delay_towrite(dev, s)) { set_bit(R5_LOCKED, &dev->flags); set_bit(R5_Wantdrain, &dev->flags); if (!expand) @@ -3494,10 +3518,26 @@ static void handle_stripe_fill(struct stripe_head *sh, * midst of changing due to a write */ if (!test_bit(STRIPE_COMPUTE_RUN, &sh->state) && !sh->check_state && - !sh->reconstruct_state) + !sh->reconstruct_state) { + + /* + * For degraded stripe with data in journal, do not handle + * read requests yet, instead, flush the stripe to raid + * disks first, this avoids handling complex rmw of write + * back cache (prexor with orig_page, and then xor with + * page) in the read path + */ + if (s->injournal && s->failed) { + if (test_bit(STRIPE_R5C_CACHING, &sh->state)) + r5c_make_stripe_write_out(sh); + goto out; + } + for (i = disks; i--; ) if (fetch_block(sh, s, i, disks)) break; + } +out: set_bit(STRIPE_HANDLE, &sh->state); } @@ -3653,7 +3693,8 @@ static int handle_stripe_dirtying(struct r5conf *conf, } else for (i = disks; i--; ) { /* would I have to read this buffer for read_modify_write */ struct r5dev *dev = &sh->dev[i]; - if ((dev->towrite || i == sh->pd_idx || i == sh->qd_idx || + if (((dev->towrite && !delay_towrite(dev, s)) || + i == sh->pd_idx || i == sh->qd_idx || test_bit(R5_InJournal, &dev->flags)) && !test_bit(R5_LOCKED, &dev->flags) && !(uptodate_for_rmw(dev) || @@ -3717,7 +3758,7 @@ static int handle_stripe_dirtying(struct r5conf *conf, for (i = disks; i--; ) { struct r5dev *dev = &sh->dev[i]; - if ((dev->towrite || + if (((dev->towrite && !delay_towrite(dev, s)) || i == sh->pd_idx || i == sh->qd_idx || test_bit(R5_InJournal, &dev->flags)) && !test_bit(R5_LOCKED, &dev->flags) && From 2e38a37f23c98d7fad87ff022670060b8a0e2bf5 Mon Sep 17 00:00:00 2001 From: Song Liu Date: Tue, 24 Jan 2017 10:45:30 -0800 Subject: [PATCH 189/343] md/r5cache: disable write back for degraded array write-back cache in degraded mode introduces corner cases to the array. Although we try to cover all these corner cases, it is safer to just disable write-back cache when the array is in degraded mode. In this patch, we disable writeback cache for degraded mode: 1. On device failure, if the array enters degraded mode, raid5_error() will submit async job r5c_disable_writeback_async to disable writeback; 2. In r5c_journal_mode_store(), it is invalid to enable writeback in degraded mode; 3. In r5c_try_caching_write(), stripes with s->failed>0 will be handled in write-through mode. Signed-off-by: Song Liu Signed-off-by: Shaohua Li --- drivers/md/raid5-cache.c | 46 ++++++++++++++++++++++++++++++++++++++++ drivers/md/raid5.c | 15 +++++++------ drivers/md/raid5.h | 2 ++ 3 files changed, 56 insertions(+), 7 deletions(-) diff --git a/drivers/md/raid5-cache.c b/drivers/md/raid5-cache.c index 3d7dda85494c..302dea3296ba 100644 --- a/drivers/md/raid5-cache.c +++ b/drivers/md/raid5-cache.c @@ -162,6 +162,8 @@ struct r5l_log { /* to submit async io_units, to fulfill ordering of flush */ struct work_struct deferred_io_work; + /* to disable write back during in degraded mode */ + struct work_struct disable_writeback_work; }; /* @@ -611,6 +613,21 @@ static void r5l_submit_io_async(struct work_struct *work) r5l_do_submit_io(log, io); } +static void r5c_disable_writeback_async(struct work_struct *work) +{ + struct r5l_log *log = container_of(work, struct r5l_log, + disable_writeback_work); + struct mddev *mddev = log->rdev->mddev; + + if (log->r5c_journal_mode == R5C_JOURNAL_MODE_WRITE_THROUGH) + return; + pr_info("md/raid:%s: Disabling writeback cache for degraded array.\n", + mdname(mddev)); + mddev_suspend(mddev); + log->r5c_journal_mode = R5C_JOURNAL_MODE_WRITE_THROUGH; + mddev_resume(mddev); +} + static void r5l_submit_current_io(struct r5l_log *log) { struct r5l_io_unit *io = log->current_io; @@ -2269,6 +2286,10 @@ static ssize_t r5c_journal_mode_store(struct mddev *mddev, val > R5C_JOURNAL_MODE_WRITE_BACK) return -EINVAL; + if (raid5_calc_degraded(conf) > 0 && + val == R5C_JOURNAL_MODE_WRITE_BACK) + return -EINVAL; + mddev_suspend(mddev); conf->log->r5c_journal_mode = val; mddev_resume(mddev); @@ -2323,6 +2344,16 @@ int r5c_try_caching_write(struct r5conf *conf, set_bit(STRIPE_R5C_CACHING, &sh->state); } + /* + * When run in degraded mode, array is set to write-through mode. + * This check helps drain pending write safely in the transition to + * write-through mode. + */ + if (s->failed) { + r5c_make_stripe_write_out(sh); + return -EAGAIN; + } + for (i = disks; i--; ) { dev = &sh->dev[i]; /* if non-overwrite, use writing-out phase */ @@ -2579,6 +2610,19 @@ ioerr: return ret; } +void r5c_update_on_rdev_error(struct mddev *mddev) +{ + struct r5conf *conf = mddev->private; + struct r5l_log *log = conf->log; + + if (!log) + return; + + if (raid5_calc_degraded(conf) > 0 && + conf->log->r5c_journal_mode == R5C_JOURNAL_MODE_WRITE_BACK) + schedule_work(&log->disable_writeback_work); +} + int r5l_init_log(struct r5conf *conf, struct md_rdev *rdev) { struct request_queue *q = bdev_get_queue(rdev->bdev); @@ -2651,6 +2695,7 @@ int r5l_init_log(struct r5conf *conf, struct md_rdev *rdev) spin_lock_init(&log->no_space_stripes_lock); INIT_WORK(&log->deferred_io_work, r5l_submit_io_async); + INIT_WORK(&log->disable_writeback_work, r5c_disable_writeback_async); log->r5c_journal_mode = R5C_JOURNAL_MODE_WRITE_THROUGH; INIT_LIST_HEAD(&log->stripe_in_journal_list); @@ -2683,6 +2728,7 @@ io_kc: void r5l_exit_log(struct r5l_log *log) { + flush_work(&log->disable_writeback_work); md_unregister_thread(&log->reclaim_thread); mempool_destroy(log->meta_pool); bioset_free(log->bs); diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c index dc83da69ca7c..3c7e106c12a2 100644 --- a/drivers/md/raid5.c +++ b/drivers/md/raid5.c @@ -556,7 +556,7 @@ static struct stripe_head *__find_stripe(struct r5conf *conf, sector_t sector, * of the two sections, and some non-in_sync devices may * be insync in the section most affected by failed devices. */ -static int calc_degraded(struct r5conf *conf) +int raid5_calc_degraded(struct r5conf *conf) { int degraded, degraded2; int i; @@ -619,7 +619,7 @@ static int has_failed(struct r5conf *conf) if (conf->mddev->reshape_position == MaxSector) return conf->mddev->degraded > conf->max_degraded; - degraded = calc_degraded(conf); + degraded = raid5_calc_degraded(conf); if (degraded > conf->max_degraded) return 1; return 0; @@ -2555,7 +2555,7 @@ static void raid5_error(struct mddev *mddev, struct md_rdev *rdev) spin_lock_irqsave(&conf->device_lock, flags); clear_bit(In_sync, &rdev->flags); - mddev->degraded = calc_degraded(conf); + mddev->degraded = raid5_calc_degraded(conf); spin_unlock_irqrestore(&conf->device_lock, flags); set_bit(MD_RECOVERY_INTR, &mddev->recovery); @@ -2569,6 +2569,7 @@ static void raid5_error(struct mddev *mddev, struct md_rdev *rdev) bdevname(rdev->bdev, b), mdname(mddev), conf->raid_disks - mddev->degraded); + r5c_update_on_rdev_error(mddev); } /* @@ -7091,7 +7092,7 @@ static int raid5_run(struct mddev *mddev) /* * 0 for a fully functional array, 1 or 2 for a degraded array. */ - mddev->degraded = calc_degraded(conf); + mddev->degraded = raid5_calc_degraded(conf); if (has_failed(conf)) { pr_crit("md/raid:%s: not enough operational devices (%d/%d failed)\n", @@ -7338,7 +7339,7 @@ static int raid5_spare_active(struct mddev *mddev) } } spin_lock_irqsave(&conf->device_lock, flags); - mddev->degraded = calc_degraded(conf); + mddev->degraded = raid5_calc_degraded(conf); spin_unlock_irqrestore(&conf->device_lock, flags); print_raid5_conf(conf); return count; @@ -7698,7 +7699,7 @@ static int raid5_start_reshape(struct mddev *mddev) * pre and post number of devices. */ spin_lock_irqsave(&conf->device_lock, flags); - mddev->degraded = calc_degraded(conf); + mddev->degraded = raid5_calc_degraded(conf); spin_unlock_irqrestore(&conf->device_lock, flags); } mddev->raid_disks = conf->raid_disks; @@ -7786,7 +7787,7 @@ static void raid5_finish_reshape(struct mddev *mddev) } else { int d; spin_lock_irq(&conf->device_lock); - mddev->degraded = calc_degraded(conf); + mddev->degraded = raid5_calc_degraded(conf); spin_unlock_irq(&conf->device_lock); for (d = conf->raid_disks ; d < conf->raid_disks - mddev->delta_disks; diff --git a/drivers/md/raid5.h b/drivers/md/raid5.h index 461df197d157..1440fa26e296 100644 --- a/drivers/md/raid5.h +++ b/drivers/md/raid5.h @@ -758,6 +758,7 @@ extern sector_t raid5_compute_sector(struct r5conf *conf, sector_t r_sector, extern struct stripe_head * raid5_get_active_stripe(struct r5conf *conf, sector_t sector, int previous, int noblock, int noquiesce); +extern int raid5_calc_degraded(struct r5conf *conf); extern int r5l_init_log(struct r5conf *conf, struct md_rdev *rdev); extern void r5l_exit_log(struct r5l_log *log); extern int r5l_write_stripe(struct r5l_log *log, struct stripe_head *head_sh); @@ -786,4 +787,5 @@ extern void r5c_flush_cache(struct r5conf *conf, int num); extern void r5c_check_stripe_cache_usage(struct r5conf *conf); extern void r5c_check_cached_full_stripe(struct r5conf *conf); extern struct md_sysfs_entry r5c_journal_mode; +extern void r5c_update_on_rdev_error(struct mddev *mddev); #endif From 0fb44559ffd67de8517098b81f675fa0210f13f0 Mon Sep 17 00:00:00 2001 From: WANG Cong Date: Mon, 23 Jan 2017 11:17:35 -0800 Subject: [PATCH 190/343] af_unix: move unix_mknod() out of bindlock Dmitry reported a deadlock scenario: unix_bind() path: u->bindlock ==> sb_writer do_splice() path: sb_writer ==> pipe->mutex ==> u->bindlock In the unix_bind() code path, unix_mknod() does not have to be done with u->bindlock held, since it is a pure fs operation, so we can just move unix_mknod() out. Reported-by: Dmitry Vyukov Tested-by: Dmitry Vyukov Cc: Rainer Weikusat Cc: Al Viro Signed-off-by: Cong Wang Signed-off-by: David S. Miller --- net/unix/af_unix.c | 27 ++++++++++++++++----------- 1 file changed, 16 insertions(+), 11 deletions(-) diff --git a/net/unix/af_unix.c b/net/unix/af_unix.c index 127656ebe7be..cef79873b09d 100644 --- a/net/unix/af_unix.c +++ b/net/unix/af_unix.c @@ -995,6 +995,7 @@ static int unix_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len) unsigned int hash; struct unix_address *addr; struct hlist_head *list; + struct path path = { NULL, NULL }; err = -EINVAL; if (sunaddr->sun_family != AF_UNIX) @@ -1010,9 +1011,20 @@ static int unix_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len) goto out; addr_len = err; + if (sun_path[0]) { + umode_t mode = S_IFSOCK | + (SOCK_INODE(sock)->i_mode & ~current_umask()); + err = unix_mknod(sun_path, mode, &path); + if (err) { + if (err == -EEXIST) + err = -EADDRINUSE; + goto out; + } + } + err = mutex_lock_interruptible(&u->bindlock); if (err) - goto out; + goto out_put; err = -EINVAL; if (u->addr) @@ -1029,16 +1041,6 @@ static int unix_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len) atomic_set(&addr->refcnt, 1); if (sun_path[0]) { - struct path path; - umode_t mode = S_IFSOCK | - (SOCK_INODE(sock)->i_mode & ~current_umask()); - err = unix_mknod(sun_path, mode, &path); - if (err) { - if (err == -EEXIST) - err = -EADDRINUSE; - unix_release_addr(addr); - goto out_up; - } addr->hash = UNIX_HASH_SIZE; hash = d_backing_inode(path.dentry)->i_ino & (UNIX_HASH_SIZE - 1); spin_lock(&unix_table_lock); @@ -1065,6 +1067,9 @@ out_unlock: spin_unlock(&unix_table_lock); out_up: mutex_unlock(&u->bindlock); +out_put: + if (err) + path_put(&path); out: return err; } From d0fa28f00052391b5df328f502fbbdd4444938b7 Mon Sep 17 00:00:00 2001 From: "Michael S. Tsirkin" Date: Mon, 23 Jan 2017 21:37:52 +0200 Subject: [PATCH 191/343] virtio_net: fix PAGE_SIZE > 64k I don't have any guests with PAGE_SIZE > 64k but the code seems to be clearly broken in that case as PAGE_SIZE / MERGEABLE_BUFFER_ALIGN will need more than 8 bit and so the code in mergeable_ctx_to_buf_address does not give us the actual true size. Cc: John Fastabend Signed-off-by: Michael S. Tsirkin Signed-off-by: David S. Miller --- drivers/net/virtio_net.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c index 347424351ade..3d1519ea0669 100644 --- a/drivers/net/virtio_net.c +++ b/drivers/net/virtio_net.c @@ -48,8 +48,16 @@ module_param(gso, bool, 0444); */ DECLARE_EWMA(pkt_len, 1, 64) +/* With mergeable buffers we align buffer address and use the low bits to + * encode its true size. Buffer size is up to 1 page so we need to align to + * square root of page size to ensure we reserve enough bits to encode the true + * size. + */ +#define MERGEABLE_BUFFER_MIN_ALIGN_SHIFT ((PAGE_SHIFT + 1) / 2) + /* Minimum alignment for mergeable packet buffers. */ -#define MERGEABLE_BUFFER_ALIGN max(L1_CACHE_BYTES, 256) +#define MERGEABLE_BUFFER_ALIGN max(L1_CACHE_BYTES, \ + 1 << MERGEABLE_BUFFER_MIN_ALIGN_SHIFT) #define VIRTNET_DRIVER_VERSION "1.0.0" From 21b995a9cb093fff33ec91d7cb3822b882a90a1e Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Mon, 23 Jan 2017 16:43:05 -0800 Subject: [PATCH 192/343] ip6_tunnel: must reload ipv6h in ip6ip6_tnl_xmit() Since ip6_tnl_parse_tlv_enc_lim() can call pskb_may_pull(), we must reload any pointer that was related to skb->head (or skb->data), or risk use after free. Fixes: c12b395a4664 ("gre: Support GRE over IPv6") Signed-off-by: Eric Dumazet Cc: Dmitry Kozlov Signed-off-by: David S. Miller --- net/ipv6/ip6_gre.c | 3 +++ net/ipv6/ip6_tunnel.c | 2 ++ 2 files changed, 5 insertions(+) diff --git a/net/ipv6/ip6_gre.c b/net/ipv6/ip6_gre.c index 75b6108234dd..558631860d91 100644 --- a/net/ipv6/ip6_gre.c +++ b/net/ipv6/ip6_gre.c @@ -582,6 +582,9 @@ static inline int ip6gre_xmit_ipv6(struct sk_buff *skb, struct net_device *dev) return -1; offset = ip6_tnl_parse_tlv_enc_lim(skb, skb_network_header(skb)); + /* ip6_tnl_parse_tlv_enc_lim() might have reallocated skb->head */ + ipv6h = ipv6_hdr(skb); + if (offset > 0) { struct ipv6_tlv_tnl_enc_lim *tel; tel = (struct ipv6_tlv_tnl_enc_lim *)&skb_network_header(skb)[offset]; diff --git a/net/ipv6/ip6_tunnel.c b/net/ipv6/ip6_tunnel.c index 753d6d0860fb..02923f956ac8 100644 --- a/net/ipv6/ip6_tunnel.c +++ b/net/ipv6/ip6_tunnel.c @@ -1303,6 +1303,8 @@ ip6ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev) fl6.flowlabel = key->label; } else { offset = ip6_tnl_parse_tlv_enc_lim(skb, skb_network_header(skb)); + /* ip6_tnl_parse_tlv_enc_lim() might have reallocated skb->head */ + ipv6h = ipv6_hdr(skb); if (offset > 0) { struct ipv6_tlv_tnl_enc_lim *tel; From fbfa743a9d2a0ffa24251764f10afc13eb21e739 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Mon, 23 Jan 2017 16:43:06 -0800 Subject: [PATCH 193/343] ipv6: fix ip6_tnl_parse_tlv_enc_lim() This function suffers from multiple issues. First one is that pskb_may_pull() may reallocate skb->head, so the 'raw' pointer needs either to be reloaded or not used at all. Second issue is that NEXTHDR_DEST handling does not validate that the options are present in skb->data, so we might read garbage or access non existent memory. With help from Willem de Bruijn. Signed-off-by: Eric Dumazet Reported-by: Dmitry Vyukov Cc: Willem de Bruijn Signed-off-by: David S. Miller --- net/ipv6/ip6_tunnel.c | 34 ++++++++++++++++++++++------------ 1 file changed, 22 insertions(+), 12 deletions(-) diff --git a/net/ipv6/ip6_tunnel.c b/net/ipv6/ip6_tunnel.c index 02923f956ac8..ff8ee06491c3 100644 --- a/net/ipv6/ip6_tunnel.c +++ b/net/ipv6/ip6_tunnel.c @@ -400,18 +400,19 @@ ip6_tnl_dev_uninit(struct net_device *dev) __u16 ip6_tnl_parse_tlv_enc_lim(struct sk_buff *skb, __u8 *raw) { - const struct ipv6hdr *ipv6h = (const struct ipv6hdr *) raw; - __u8 nexthdr = ipv6h->nexthdr; - __u16 off = sizeof(*ipv6h); + const struct ipv6hdr *ipv6h = (const struct ipv6hdr *)raw; + unsigned int nhoff = raw - skb->data; + unsigned int off = nhoff + sizeof(*ipv6h); + u8 next, nexthdr = ipv6h->nexthdr; while (ipv6_ext_hdr(nexthdr) && nexthdr != NEXTHDR_NONE) { - __u16 optlen = 0; struct ipv6_opt_hdr *hdr; - if (raw + off + sizeof(*hdr) > skb->data && - !pskb_may_pull(skb, raw - skb->data + off + sizeof (*hdr))) + u16 optlen; + + if (!pskb_may_pull(skb, off + sizeof(*hdr))) break; - hdr = (struct ipv6_opt_hdr *) (raw + off); + hdr = (struct ipv6_opt_hdr *)(skb->data + off); if (nexthdr == NEXTHDR_FRAGMENT) { struct frag_hdr *frag_hdr = (struct frag_hdr *) hdr; if (frag_hdr->frag_off) @@ -422,20 +423,29 @@ __u16 ip6_tnl_parse_tlv_enc_lim(struct sk_buff *skb, __u8 *raw) } else { optlen = ipv6_optlen(hdr); } + /* cache hdr->nexthdr, since pskb_may_pull() might + * invalidate hdr + */ + next = hdr->nexthdr; if (nexthdr == NEXTHDR_DEST) { - __u16 i = off + 2; + u16 i = 2; + + /* Remember : hdr is no longer valid at this point. */ + if (!pskb_may_pull(skb, off + optlen)) + break; + while (1) { struct ipv6_tlv_tnl_enc_lim *tel; /* No more room for encapsulation limit */ - if (i + sizeof (*tel) > off + optlen) + if (i + sizeof(*tel) > optlen) break; - tel = (struct ipv6_tlv_tnl_enc_lim *) &raw[i]; + tel = (struct ipv6_tlv_tnl_enc_lim *) skb->data + off + i; /* return index of option if found and valid */ if (tel->type == IPV6_TLV_TNL_ENCAP_LIMIT && tel->length == 1) - return i; + return i + off - nhoff; /* else jump to next option */ if (tel->type) i += tel->length + 2; @@ -443,7 +453,7 @@ __u16 ip6_tnl_parse_tlv_enc_lim(struct sk_buff *skb, __u8 *raw) i++; } } - nexthdr = hdr->nexthdr; + nexthdr = next; off += optlen; } return 0; From 8b3f9337e17aaf710c79e65fd0a3c572a075f498 Mon Sep 17 00:00:00 2001 From: Roopa Prabhu Date: Mon, 23 Jan 2017 20:44:32 -0800 Subject: [PATCH 194/343] vxlan: don't flush static fdb entries on admin down This patch skips flushing static fdb entries in ndo_stop, but flushes all fdb entries during vxlan device delete. This is consistent with the bridge driver fdb Signed-off-by: Roopa Prabhu Signed-off-by: David S. Miller --- drivers/net/vxlan.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c index 8a79cfcac889..746a819f0a41 100644 --- a/drivers/net/vxlan.c +++ b/drivers/net/vxlan.c @@ -2354,7 +2354,7 @@ static int vxlan_open(struct net_device *dev) } /* Purge the forwarding table */ -static void vxlan_flush(struct vxlan_dev *vxlan) +static void vxlan_flush(struct vxlan_dev *vxlan, bool do_all) { unsigned int h; @@ -2364,6 +2364,8 @@ static void vxlan_flush(struct vxlan_dev *vxlan) hlist_for_each_safe(p, n, &vxlan->fdb_head[h]) { struct vxlan_fdb *f = container_of(p, struct vxlan_fdb, hlist); + if (!do_all && (f->state & (NUD_PERMANENT | NUD_NOARP))) + continue; /* the all_zeros_mac entry is deleted at vxlan_uninit */ if (!is_zero_ether_addr(f->eth_addr)) vxlan_fdb_destroy(vxlan, f); @@ -2385,7 +2387,7 @@ static int vxlan_stop(struct net_device *dev) del_timer_sync(&vxlan->age_timer); - vxlan_flush(vxlan); + vxlan_flush(vxlan, false); vxlan_sock_release(vxlan); return ret; @@ -3058,6 +3060,8 @@ static void vxlan_dellink(struct net_device *dev, struct list_head *head) struct vxlan_dev *vxlan = netdev_priv(dev); struct vxlan_net *vn = net_generic(vxlan->net, vxlan_net_id); + vxlan_flush(vxlan, true); + spin_lock(&vn->sock_lock); if (!hlist_unhashed(&vxlan->hlist)) hlist_del_rcu(&vxlan->hlist); From efb5f68f32995c146944a9d4257c3cf8eae2c4a1 Mon Sep 17 00:00:00 2001 From: Balakrishnan Raman Date: Mon, 23 Jan 2017 20:44:33 -0800 Subject: [PATCH 195/343] vxlan: do not age static remote mac entries Mac aging is applicable only for dynamically learnt remote mac entries. Check for user configured static remote mac entries and skip aging. Signed-off-by: Balakrishnan Raman Signed-off-by: Roopa Prabhu Signed-off-by: David S. Miller --- drivers/net/vxlan.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c index 746a819f0a41..50b62db213b0 100644 --- a/drivers/net/vxlan.c +++ b/drivers/net/vxlan.c @@ -2268,7 +2268,7 @@ static void vxlan_cleanup(unsigned long arg) = container_of(p, struct vxlan_fdb, hlist); unsigned long timeout; - if (f->state & NUD_PERMANENT) + if (f->state & (NUD_PERMANENT | NUD_NOARP)) continue; timeout = f->used + vxlan->cfg.age_interval * HZ; From 23d28a859fb847fd7fcfbd31acb3b160abb5d6ae Mon Sep 17 00:00:00 2001 From: Thomas Huth Date: Tue, 24 Jan 2017 07:28:41 +0100 Subject: [PATCH 196/343] ibmveth: Add a proper check for the availability of the checksum features When using the ibmveth driver in a KVM/QEMU based VM, it currently always prints out a scary error message like this when it is started: ibmveth 71000003 (unregistered net_device): unable to change checksum offload settings. 1 rc=-2 ret_attr=71000003 This happens because the driver always tries to enable the checksum offloading without checking for the availability of this feature first. QEMU does not support checksum offloading for the spapr-vlan device, thus we always get the error message here. According to the LoPAPR specification, the "ibm,illan-options" property of the corresponding device tree node should be checked first to see whether the H_ILLAN_ATTRIUBTES hypercall and thus the checksum offloading feature is available. Thus let's do this in the ibmveth driver, too, so that the error message is really only limited to cases where something goes wrong, and does not occur if the feature is just missing. Signed-off-by: Thomas Huth Signed-off-by: David S. Miller --- drivers/net/ethernet/ibm/ibmveth.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/ibm/ibmveth.c b/drivers/net/ethernet/ibm/ibmveth.c index a831f947ca8c..309f5c66083c 100644 --- a/drivers/net/ethernet/ibm/ibmveth.c +++ b/drivers/net/ethernet/ibm/ibmveth.c @@ -1601,8 +1601,11 @@ static int ibmveth_probe(struct vio_dev *dev, const struct vio_device_id *id) netdev->netdev_ops = &ibmveth_netdev_ops; netdev->ethtool_ops = &netdev_ethtool_ops; SET_NETDEV_DEV(netdev, &dev->dev); - netdev->hw_features = NETIF_F_SG | NETIF_F_RXCSUM | - NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM; + netdev->hw_features = NETIF_F_SG; + if (vio_get_attribute(dev, "ibm,illan-options", NULL) != NULL) { + netdev->hw_features |= NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM | + NETIF_F_RXCSUM; + } netdev->features |= netdev->hw_features; From 83d230eb5c638949350f4761acdfc0af5cb1bc00 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Mon, 23 Jan 2017 19:43:00 -0800 Subject: [PATCH 197/343] xfs: verify dirblocklog correctly sb_dirblklog is added to sb_blocklog to compute the directory block size in bytes. Therefore, we must compare the sum of both those values against XFS_MAX_BLOCKSIZE_LOG, not just dirblklog. Signed-off-by: Darrick J. Wong Reviewed-by: Eric Sandeen Reviewed-by: Christoph Hellwig --- fs/xfs/libxfs/xfs_sb.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/xfs/libxfs/xfs_sb.c b/fs/xfs/libxfs/xfs_sb.c index 2580262e4ea0..584ec896a533 100644 --- a/fs/xfs/libxfs/xfs_sb.c +++ b/fs/xfs/libxfs/xfs_sb.c @@ -242,7 +242,7 @@ xfs_mount_validate_sb( sbp->sb_blocklog < XFS_MIN_BLOCKSIZE_LOG || sbp->sb_blocklog > XFS_MAX_BLOCKSIZE_LOG || sbp->sb_blocksize != (1 << sbp->sb_blocklog) || - sbp->sb_dirblklog > XFS_MAX_BLOCKSIZE_LOG || + sbp->sb_dirblklog + sbp->sb_blocklog > XFS_MAX_BLOCKSIZE_LOG || sbp->sb_inodesize < XFS_DINODE_MIN_SIZE || sbp->sb_inodesize > XFS_DINODE_MAX_SIZE || sbp->sb_inodelog < XFS_DINODE_MIN_LOG || From 5b9f57516337b523f7466a53939aaaea7b78141b Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Bj=C3=B8rn=20Mork?= Date: Tue, 24 Jan 2017 10:45:38 +0100 Subject: [PATCH 198/343] qmi_wwan/cdc_ether: add device ID for HP lt2523 (Novatel E371) WWAN card MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Another rebranded Novatel E371. qmi_wwan should drive this device, while cdc_ether should ignore it. Even though the USB descriptors are plain CDC-ETHER that USB interface is a QMI interface. Ref commit 7fdb7846c9ca ("qmi_wwan/cdc_ether: add device IDs for Dell 5804 (Novatel E371) WWAN card") Cc: Dan Williams Signed-off-by: Bjørn Mork Signed-off-by: David S. Miller --- drivers/net/usb/cdc_ether.c | 8 ++++++++ drivers/net/usb/qmi_wwan.c | 7 +++++++ 2 files changed, 15 insertions(+) diff --git a/drivers/net/usb/cdc_ether.c b/drivers/net/usb/cdc_ether.c index fe7b2886cb6b..86144f9a80ee 100644 --- a/drivers/net/usb/cdc_ether.c +++ b/drivers/net/usb/cdc_ether.c @@ -531,6 +531,7 @@ static const struct driver_info wwan_info = { #define SAMSUNG_VENDOR_ID 0x04e8 #define LENOVO_VENDOR_ID 0x17ef #define NVIDIA_VENDOR_ID 0x0955 +#define HP_VENDOR_ID 0x03f0 static const struct usb_device_id products[] = { /* BLACKLIST !! @@ -677,6 +678,13 @@ static const struct usb_device_id products[] = { .driver_info = 0, }, +/* HP lt2523 (Novatel E371) - handled by qmi_wwan */ +{ + USB_DEVICE_AND_INTERFACE_INFO(HP_VENDOR_ID, 0x421d, USB_CLASS_COMM, + USB_CDC_SUBCLASS_ETHERNET, USB_CDC_PROTO_NONE), + .driver_info = 0, +}, + /* AnyDATA ADU960S - handled by qmi_wwan */ { USB_DEVICE_AND_INTERFACE_INFO(0x16d5, 0x650a, USB_CLASS_COMM, diff --git a/drivers/net/usb/qmi_wwan.c b/drivers/net/usb/qmi_wwan.c index 6fe1cdb0174f..24d5272cdce5 100644 --- a/drivers/net/usb/qmi_wwan.c +++ b/drivers/net/usb/qmi_wwan.c @@ -654,6 +654,13 @@ static const struct usb_device_id products[] = { USB_CDC_PROTO_NONE), .driver_info = (unsigned long)&qmi_wwan_info, }, + { /* HP lt2523 (Novatel E371) */ + USB_DEVICE_AND_INTERFACE_INFO(0x03f0, 0x421d, + USB_CLASS_COMM, + USB_CDC_SUBCLASS_ETHERNET, + USB_CDC_PROTO_NONE), + .driver_info = (unsigned long)&qmi_wwan_info, + }, { /* HP lt4112 LTE/HSPA+ Gobi 4G Module (Huawei me906e) */ USB_DEVICE_AND_INTERFACE_INFO(0x03f0, 0x581d, USB_CLASS_VENDOR_SPEC, 1, 7), .driver_info = (unsigned long)&qmi_wwan_info, From f1db5c101cd48b5555ed9e061dcc49ed329812ea Mon Sep 17 00:00:00 2001 From: Tobias Regnery Date: Tue, 24 Jan 2017 14:34:22 +0100 Subject: [PATCH 199/343] alx: fix wrong condition to free descriptor memory The condition to free the descriptor memory is wrong, we want to free the memory if it is set and not if it is unset. Invert the test to fix this issue. Fixes: b0999223f224b ("alx: add ability to allocate and free alx_napi structures") Signed-off-by: Tobias Regnery Signed-off-by: David S. Miller --- drivers/net/ethernet/atheros/alx/main.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/atheros/alx/main.c b/drivers/net/ethernet/atheros/alx/main.c index c8f525574d68..765306bd78c2 100644 --- a/drivers/net/ethernet/atheros/alx/main.c +++ b/drivers/net/ethernet/atheros/alx/main.c @@ -703,7 +703,7 @@ static void alx_free_rings(struct alx_priv *alx) if (alx->qnapi[0] && alx->qnapi[0]->rxq) kfree(alx->qnapi[0]->rxq->bufs); - if (!alx->descmem.virt) + if (alx->descmem.virt) dma_free_coherent(&alx->hw.pdev->dev, alx->descmem.size, alx->descmem.virt, From 37187a016c37d7e550544544dba25399ce4589c9 Mon Sep 17 00:00:00 2001 From: Tobias Regnery Date: Tue, 24 Jan 2017 14:34:23 +0100 Subject: [PATCH 200/343] alx: fix fallback to msi or legacy interrupts If requesting msi-x interrupts fails we should fall back to msi or legacy interrupts. However alx_realloc_ressources don't call alx_init_intr, so we fail to set the right number of tx queues. This results in watchdog timeouts and a nonfunctional adapter. Fixes: d768319cd427 ("alx: enable multiple tx queues") Signed-off-by: Tobias Regnery Signed-off-by: David S. Miller --- drivers/net/ethernet/atheros/alx/main.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/net/ethernet/atheros/alx/main.c b/drivers/net/ethernet/atheros/alx/main.c index 765306bd78c2..75cbd46e429d 100644 --- a/drivers/net/ethernet/atheros/alx/main.c +++ b/drivers/net/ethernet/atheros/alx/main.c @@ -984,6 +984,7 @@ static int alx_realloc_resources(struct alx_priv *alx) alx_free_rings(alx); alx_free_napis(alx); alx_disable_advanced_intr(alx); + alx_init_intr(alx, false); err = alx_alloc_napis(alx); if (err) From 185aceefd80f98dc5b9d73eb6cbb70739a5ce4ea Mon Sep 17 00:00:00 2001 From: Tobias Regnery Date: Tue, 24 Jan 2017 14:34:24 +0100 Subject: [PATCH 201/343] alx: work around hardware bug in interrupt fallback path If requesting msi-x interrupts fails in alx_request_irq we fall back to a single tx queue and msi or legacy interrupts. Currently the adapter stops working in this case and we get tx watchdog timeouts. For reasons unknown the adapter gets confused when we load the dma adresses to the chip in alx_init_ring_ptrs twice: the first time with multiple queues and the second time in the fallback case with a single queue. To fix this move the the call to alx_reinit_rings (which calls alx_init_ring_ptrs) after alx_request_irq. At this time it is clear how much tx queues we have and which dma addresses we use. Fixes: d768319cd427 ("alx: enable multiple tx queues") Signed-off-by: Tobias Regnery Signed-off-by: David S. Miller --- drivers/net/ethernet/atheros/alx/main.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/atheros/alx/main.c b/drivers/net/ethernet/atheros/alx/main.c index 75cbd46e429d..7dcc907a449d 100644 --- a/drivers/net/ethernet/atheros/alx/main.c +++ b/drivers/net/ethernet/atheros/alx/main.c @@ -685,8 +685,6 @@ static int alx_alloc_rings(struct alx_priv *alx) return -ENOMEM; } - alx_reinit_rings(alx); - return 0; } @@ -1242,6 +1240,12 @@ static int __alx_open(struct alx_priv *alx, bool resume) if (err) goto out_free_rings; + /* must be called after alx_request_irq because the chip stops working + * if we copy the dma addresses in alx_init_ring_ptrs twice when + * requesting msi-x interrupts failed + */ + alx_reinit_rings(alx); + netif_set_real_num_tx_queues(alx->dev, alx->num_txq); netif_set_real_num_rx_queues(alx->dev, alx->num_rxq); From c929ea0b910355e1876c64431f3d5802f95b3d75 Mon Sep 17 00:00:00 2001 From: Kinglong Mee Date: Fri, 20 Jan 2017 16:48:39 +0800 Subject: [PATCH 202/343] SUNRPC: cleanup ida information when removing sunrpc module After removing sunrpc module, I get many kmemleak information as, unreferenced object 0xffff88003316b1e0 (size 544): comm "gssproxy", pid 2148, jiffies 4294794465 (age 4200.081s) hex dump (first 32 bytes): 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ backtrace: [] kmemleak_alloc+0x4a/0xa0 [] kmem_cache_alloc+0x15e/0x1f0 [] ida_pre_get+0xaa/0x150 [] ida_simple_get+0xad/0x180 [] nlmsvc_lookup_host+0x4ab/0x7f0 [lockd] [] lockd+0x4d/0x270 [lockd] [] param_set_timeout+0x55/0x100 [lockd] [] svc_defer+0x114/0x3f0 [sunrpc] [] svc_defer+0x2d7/0x3f0 [sunrpc] [] rpc_show_info+0x8a/0x110 [sunrpc] [] proc_reg_write+0x7f/0xc0 [] __vfs_write+0xdf/0x3c0 [] vfs_write+0xef/0x240 [] SyS_write+0xad/0x130 [] entry_SYSCALL_64_fastpath+0x1a/0xa9 [] 0xffffffffffffffff I found, the ida information (dynamic memory) isn't cleanup. Signed-off-by: Kinglong Mee Fixes: 2f048db4680a ("SUNRPC: Add an identifier for struct rpc_clnt") Cc: stable@vger.kernel.org # v3.12+ Signed-off-by: Trond Myklebust --- include/linux/sunrpc/clnt.h | 1 + net/sunrpc/clnt.c | 5 +++++ net/sunrpc/sunrpc_syms.c | 1 + 3 files changed, 7 insertions(+) diff --git a/include/linux/sunrpc/clnt.h b/include/linux/sunrpc/clnt.h index 85cc819676e8..333ad11b3dd9 100644 --- a/include/linux/sunrpc/clnt.h +++ b/include/linux/sunrpc/clnt.h @@ -216,5 +216,6 @@ void rpc_clnt_xprt_switch_put(struct rpc_clnt *); void rpc_clnt_xprt_switch_add_xprt(struct rpc_clnt *, struct rpc_xprt *); bool rpc_clnt_xprt_switch_has_addr(struct rpc_clnt *clnt, const struct sockaddr *sap); +void rpc_cleanup_clids(void); #endif /* __KERNEL__ */ #endif /* _LINUX_SUNRPC_CLNT_H */ diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index 1efbe48e794f..1dc9f3bac099 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -336,6 +336,11 @@ out: static DEFINE_IDA(rpc_clids); +void rpc_cleanup_clids(void) +{ + ida_destroy(&rpc_clids); +} + static int rpc_alloc_clid(struct rpc_clnt *clnt) { int clid; diff --git a/net/sunrpc/sunrpc_syms.c b/net/sunrpc/sunrpc_syms.c index d1c330a7953a..c73de181467a 100644 --- a/net/sunrpc/sunrpc_syms.c +++ b/net/sunrpc/sunrpc_syms.c @@ -119,6 +119,7 @@ out: static void __exit cleanup_sunrpc(void) { + rpc_cleanup_clids(); rpcauth_remove_module(); cleanup_socket_xprt(); svc_cleanup_xprt_sock(); From d3f4aadd614c4627244452ad64eaf351179f2c31 Mon Sep 17 00:00:00 2001 From: "Amrani, Ram" Date: Mon, 26 Dec 2016 08:40:57 +0200 Subject: [PATCH 203/343] RDMA/core: Add the function ib_mtu_int_to_enum As the functionality to convert the MTU from a number to enum_ib_mtu is ubiquitous, define a dedicated function and remove the duplicated code. Signed-off-by: Ram Amrani Reviewed-by: Leon Romanovsky Signed-off-by: Doug Ledford --- drivers/infiniband/hw/cxgb3/iwch_provider.c | 11 +---------- drivers/infiniband/hw/cxgb4/provider.c | 11 +---------- drivers/infiniband/hw/i40iw/i40iw_verbs.c | 11 +---------- drivers/infiniband/hw/nes/nes_verbs.c | 12 +----------- include/rdma/ib_verbs.h | 14 ++++++++++++++ 5 files changed, 18 insertions(+), 41 deletions(-) diff --git a/drivers/infiniband/hw/cxgb3/iwch_provider.c b/drivers/infiniband/hw/cxgb3/iwch_provider.c index 9d5fe1853da4..6262dc035f3c 100644 --- a/drivers/infiniband/hw/cxgb3/iwch_provider.c +++ b/drivers/infiniband/hw/cxgb3/iwch_provider.c @@ -1135,16 +1135,7 @@ static int iwch_query_port(struct ib_device *ibdev, memset(props, 0, sizeof(struct ib_port_attr)); props->max_mtu = IB_MTU_4096; - if (netdev->mtu >= 4096) - props->active_mtu = IB_MTU_4096; - else if (netdev->mtu >= 2048) - props->active_mtu = IB_MTU_2048; - else if (netdev->mtu >= 1024) - props->active_mtu = IB_MTU_1024; - else if (netdev->mtu >= 512) - props->active_mtu = IB_MTU_512; - else - props->active_mtu = IB_MTU_256; + props->active_mtu = ib_mtu_int_to_enum(netdev->mtu); if (!netif_carrier_ok(netdev)) props->state = IB_PORT_DOWN; diff --git a/drivers/infiniband/hw/cxgb4/provider.c b/drivers/infiniband/hw/cxgb4/provider.c index fa64f5d93b11..3345e1c312f7 100644 --- a/drivers/infiniband/hw/cxgb4/provider.c +++ b/drivers/infiniband/hw/cxgb4/provider.c @@ -373,16 +373,7 @@ static int c4iw_query_port(struct ib_device *ibdev, u8 port, memset(props, 0, sizeof(struct ib_port_attr)); props->max_mtu = IB_MTU_4096; - if (netdev->mtu >= 4096) - props->active_mtu = IB_MTU_4096; - else if (netdev->mtu >= 2048) - props->active_mtu = IB_MTU_2048; - else if (netdev->mtu >= 1024) - props->active_mtu = IB_MTU_1024; - else if (netdev->mtu >= 512) - props->active_mtu = IB_MTU_512; - else - props->active_mtu = IB_MTU_256; + props->active_mtu = ib_mtu_int_to_enum(netdev->mtu); if (!netif_carrier_ok(netdev)) props->state = IB_PORT_DOWN; diff --git a/drivers/infiniband/hw/i40iw/i40iw_verbs.c b/drivers/infiniband/hw/i40iw/i40iw_verbs.c index 29e97df9e1a7..4c000d60d5c6 100644 --- a/drivers/infiniband/hw/i40iw/i40iw_verbs.c +++ b/drivers/infiniband/hw/i40iw/i40iw_verbs.c @@ -100,16 +100,7 @@ static int i40iw_query_port(struct ib_device *ibdev, memset(props, 0, sizeof(*props)); props->max_mtu = IB_MTU_4096; - if (netdev->mtu >= 4096) - props->active_mtu = IB_MTU_4096; - else if (netdev->mtu >= 2048) - props->active_mtu = IB_MTU_2048; - else if (netdev->mtu >= 1024) - props->active_mtu = IB_MTU_1024; - else if (netdev->mtu >= 512) - props->active_mtu = IB_MTU_512; - else - props->active_mtu = IB_MTU_256; + props->active_mtu = ib_mtu_int_to_enum(netdev->mtu); props->lid = 1; if (netif_carrier_ok(iwdev->netdev)) diff --git a/drivers/infiniband/hw/nes/nes_verbs.c b/drivers/infiniband/hw/nes/nes_verbs.c index aff9fb14768b..5a31f3c6a421 100644 --- a/drivers/infiniband/hw/nes/nes_verbs.c +++ b/drivers/infiniband/hw/nes/nes_verbs.c @@ -478,17 +478,7 @@ static int nes_query_port(struct ib_device *ibdev, u8 port, struct ib_port_attr memset(props, 0, sizeof(*props)); props->max_mtu = IB_MTU_4096; - - if (netdev->mtu >= 4096) - props->active_mtu = IB_MTU_4096; - else if (netdev->mtu >= 2048) - props->active_mtu = IB_MTU_2048; - else if (netdev->mtu >= 1024) - props->active_mtu = IB_MTU_1024; - else if (netdev->mtu >= 512) - props->active_mtu = IB_MTU_512; - else - props->active_mtu = IB_MTU_256; + props->active_mtu = ib_mtu_int_to_enum(netdev->mtu); props->lid = 1; props->lmc = 0; diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h index 958a24d8fae7..b567e4452a47 100644 --- a/include/rdma/ib_verbs.h +++ b/include/rdma/ib_verbs.h @@ -352,6 +352,20 @@ static inline int ib_mtu_enum_to_int(enum ib_mtu mtu) } } +static inline enum ib_mtu ib_mtu_int_to_enum(int mtu) +{ + if (mtu >= 4096) + return IB_MTU_4096; + else if (mtu >= 2048) + return IB_MTU_2048; + else if (mtu >= 1024) + return IB_MTU_1024; + else if (mtu >= 512) + return IB_MTU_512; + else + return IB_MTU_256; +} + enum ib_port_state { IB_PORT_NOP = 0, IB_PORT_DOWN = 1, From 097b615965fb1af714fbc2311f68839b1086ebcb Mon Sep 17 00:00:00 2001 From: "Amrani, Ram" Date: Mon, 26 Dec 2016 08:40:58 +0200 Subject: [PATCH 204/343] RDMA/qedr: Fix MTU returned from QP query MTU value returned from QP query should include overhead. Signed-off-by: Ram Amrani Signed-off-by: Doug Ledford --- drivers/infiniband/hw/qedr/verbs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/infiniband/hw/qedr/verbs.c b/drivers/infiniband/hw/qedr/verbs.c index 57c8de208077..84dcd73bd57b 100644 --- a/drivers/infiniband/hw/qedr/verbs.c +++ b/drivers/infiniband/hw/qedr/verbs.c @@ -2016,7 +2016,7 @@ int qedr_query_qp(struct ib_qp *ibqp, qp_attr->qp_state = qedr_get_ibqp_state(params.state); qp_attr->cur_qp_state = qedr_get_ibqp_state(params.state); - qp_attr->path_mtu = iboe_get_mtu(params.mtu); + qp_attr->path_mtu = ib_mtu_int_to_enum(params.mtu); qp_attr->path_mig_state = IB_MIG_MIGRATED; qp_attr->rq_psn = params.rq_psn; qp_attr->sq_psn = params.sq_psn; From 20f5e10ef8bcf29a915642245b66e5a132e38fc4 Mon Sep 17 00:00:00 2001 From: "Amrani, Ram" Date: Tue, 24 Jan 2017 12:01:31 +0200 Subject: [PATCH 205/343] RDMA/qedr: Add uapi header qedr-abi.h Signed-off-by: Ram Amrani Signed-off-by: Doug Ledford --- include/uapi/rdma/Kbuild | 1 + 1 file changed, 1 insertion(+) diff --git a/include/uapi/rdma/Kbuild b/include/uapi/rdma/Kbuild index 82bdf5626859..bb68cb1b04ed 100644 --- a/include/uapi/rdma/Kbuild +++ b/include/uapi/rdma/Kbuild @@ -16,3 +16,4 @@ header-y += nes-abi.h header-y += ocrdma-abi.h header-y += hns-abi.h header-y += vmw_pvrdma-abi.h +header-y += qedr-abi.h From 865cea40b69741c3da2574176876463233b2b67c Mon Sep 17 00:00:00 2001 From: Ram Amrani Date: Tue, 24 Jan 2017 13:50:34 +0200 Subject: [PATCH 206/343] RDMA/qedr: Return success when not changing QP state If the user is requesting us to change the QP state to the same state that it is already in, return success instead of failure. Signed-off-by: Ram Amrani Signed-off-by: Michal Kalderon Signed-off-by: Doug Ledford --- drivers/infiniband/hw/qedr/verbs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/infiniband/hw/qedr/verbs.c b/drivers/infiniband/hw/qedr/verbs.c index 84dcd73bd57b..27d90a82d731 100644 --- a/drivers/infiniband/hw/qedr/verbs.c +++ b/drivers/infiniband/hw/qedr/verbs.c @@ -1657,7 +1657,7 @@ static int qedr_update_qp_state(struct qedr_dev *dev, int status = 0; if (new_state == qp->state) - return 1; + return 0; switch (qp->state) { case QED_ROCE_QP_STATE_RESET: From 59e8970b3798e4cbe575ed9cf4d53098760a2a86 Mon Sep 17 00:00:00 2001 From: Ram Amrani Date: Tue, 24 Jan 2017 13:50:35 +0200 Subject: [PATCH 207/343] RDMA/qedr: Return max inline data in QP query result Return the maximum supported amount of inline data, not the qp's current configured inline data size, when filling out the results of a query qp call. Signed-off-by: Ram Amrani Signed-off-by: Michal Kalderon Signed-off-by: Doug Ledford --- drivers/infiniband/hw/qedr/verbs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/infiniband/hw/qedr/verbs.c b/drivers/infiniband/hw/qedr/verbs.c index 27d90a82d731..bb907b14c988 100644 --- a/drivers/infiniband/hw/qedr/verbs.c +++ b/drivers/infiniband/hw/qedr/verbs.c @@ -2028,7 +2028,7 @@ int qedr_query_qp(struct ib_qp *ibqp, qp_attr->cap.max_recv_wr = qp->rq.max_wr; qp_attr->cap.max_send_sge = qp->sq.max_sges; qp_attr->cap.max_recv_sge = qp->rq.max_sges; - qp_attr->cap.max_inline_data = qp->max_inline_data; + qp_attr->cap.max_inline_data = ROCE_REQ_MAX_INLINE_DATA_SIZE; qp_init_attr->cap = qp_attr->cap; memcpy(&qp_attr->ah_attr.grh.dgid.raw[0], ¶ms.dgid.bytes[0], From 91bff997db2ec04f9ba761a55c21642f9803b06c Mon Sep 17 00:00:00 2001 From: Ram Amrani Date: Tue, 24 Jan 2017 13:50:36 +0200 Subject: [PATCH 208/343] RDMA/qedr: Remove CQ spinlock from CM completion handlers There is only a single event queue that triggers the completion events for the RDMA CM and it is being processed serially. This means that inherently there can no parallelism of CQ completion handler callbacks, hence the lock is redundant. Signed-off-by: Ram Amrani Signed-off-by: Michal Kalderon Reviewed-by: Leon Romanovsky Signed-off-by: Doug Ledford --- drivers/infiniband/hw/qedr/qedr.h | 3 --- drivers/infiniband/hw/qedr/qedr_cm.c | 10 ++-------- 2 files changed, 2 insertions(+), 11 deletions(-) diff --git a/drivers/infiniband/hw/qedr/qedr.h b/drivers/infiniband/hw/qedr/qedr.h index 620badd7d4fb..94319abb0df2 100644 --- a/drivers/infiniband/hw/qedr/qedr.h +++ b/drivers/infiniband/hw/qedr/qedr.h @@ -251,9 +251,6 @@ struct qedr_cq { u16 icid; - /* Lock to protect completion handler */ - spinlock_t comp_handler_lock; - /* Lock to protect multiplem CQ's */ spinlock_t cq_lock; u8 arm_flags; diff --git a/drivers/infiniband/hw/qedr/qedr_cm.c b/drivers/infiniband/hw/qedr/qedr_cm.c index 63890ebb72bd..00361f310d15 100644 --- a/drivers/infiniband/hw/qedr/qedr_cm.c +++ b/drivers/infiniband/hw/qedr/qedr_cm.c @@ -87,11 +87,8 @@ void qedr_ll2_tx_cb(void *_qdev, struct qed_roce_ll2_packet *pkt) qedr_inc_sw_gsi_cons(&qp->sq); spin_unlock_irqrestore(&qp->q_lock, flags); - if (cq->ibcq.comp_handler) { - spin_lock_irqsave(&cq->comp_handler_lock, flags); + if (cq->ibcq.comp_handler) (*cq->ibcq.comp_handler) (&cq->ibcq, cq->ibcq.cq_context); - spin_unlock_irqrestore(&cq->comp_handler_lock, flags); - } } void qedr_ll2_rx_cb(void *_dev, struct qed_roce_ll2_packet *pkt, @@ -113,11 +110,8 @@ void qedr_ll2_rx_cb(void *_dev, struct qed_roce_ll2_packet *pkt, spin_unlock_irqrestore(&qp->q_lock, flags); - if (cq->ibcq.comp_handler) { - spin_lock_irqsave(&cq->comp_handler_lock, flags); + if (cq->ibcq.comp_handler) (*cq->ibcq.comp_handler) (&cq->ibcq, cq->ibcq.cq_context); - spin_unlock_irqrestore(&cq->comp_handler_lock, flags); - } } static void qedr_destroy_gsi_cq(struct qedr_dev *dev, From c78c31496111f497b4a03f955c100091185da8b6 Mon Sep 17 00:00:00 2001 From: Ram Amrani Date: Tue, 24 Jan 2017 13:50:37 +0200 Subject: [PATCH 209/343] RDMA/qedr: Don't spam dmesg if QP is in error state It is normal to flush CQEs if the QP is in error state. Hence there's no use in printing a message per CQE to dmesg. Signed-off-by: Ram Amrani Signed-off-by: Michal Kalderon Signed-off-by: Doug Ledford --- drivers/infiniband/hw/qedr/verbs.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/drivers/infiniband/hw/qedr/verbs.c b/drivers/infiniband/hw/qedr/verbs.c index bb907b14c988..0688dce54f92 100644 --- a/drivers/infiniband/hw/qedr/verbs.c +++ b/drivers/infiniband/hw/qedr/verbs.c @@ -3234,9 +3234,10 @@ static int qedr_poll_cq_req(struct qedr_dev *dev, IB_WC_SUCCESS, 0); break; case RDMA_CQE_REQ_STS_WORK_REQUEST_FLUSHED_ERR: - DP_ERR(dev, - "Error: POLL CQ with RDMA_CQE_REQ_STS_WORK_REQUEST_FLUSHED_ERR. CQ icid=0x%x, QP icid=0x%x\n", - cq->icid, qp->icid); + if (qp->state != QED_ROCE_QP_STATE_ERR) + DP_ERR(dev, + "Error: POLL CQ with RDMA_CQE_REQ_STS_WORK_REQUEST_FLUSHED_ERR. CQ icid=0x%x, QP icid=0x%x\n", + cq->icid, qp->icid); cnt = process_req(dev, qp, cq, num_entries, wc, req->sq_cons, IB_WC_WR_FLUSH_ERR, 1); break; From 933e6dcaa0f65eb2f624ad760274020874a1f35e Mon Sep 17 00:00:00 2001 From: Ram Amrani Date: Tue, 24 Jan 2017 13:50:38 +0200 Subject: [PATCH 210/343] RDMA/qedr: Don't reset QP when queues aren't flushed Fail QP state transition from error to reset if SQ/RQ are not empty and still in the process of flushing out the queued work entries. Signed-off-by: Ram Amrani Signed-off-by: Michal Kalderon Signed-off-by: Doug Ledford --- drivers/infiniband/hw/qedr/verbs.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/drivers/infiniband/hw/qedr/verbs.c b/drivers/infiniband/hw/qedr/verbs.c index 0688dce54f92..3dae9641f821 100644 --- a/drivers/infiniband/hw/qedr/verbs.c +++ b/drivers/infiniband/hw/qedr/verbs.c @@ -1733,6 +1733,14 @@ static int qedr_update_qp_state(struct qedr_dev *dev, /* ERR->XXX */ switch (new_state) { case QED_ROCE_QP_STATE_RESET: + if ((qp->rq.prod != qp->rq.cons) || + (qp->sq.prod != qp->sq.cons)) { + DP_NOTICE(dev, + "Error->Reset with rq/sq not empty rq.prod=%x rq.cons=%x sq.prod=%x sq.cons=%x\n", + qp->rq.prod, qp->rq.cons, qp->sq.prod, + qp->sq.cons); + status = -EINVAL; + } break; default: status = -EINVAL; From 27a4b1a6d6fcf09314359bacefa1e106927ae21b Mon Sep 17 00:00:00 2001 From: Ram Amrani Date: Tue, 24 Jan 2017 13:51:39 +0200 Subject: [PATCH 211/343] RDMA/qedr: Mark three functions as static mark qedr_get_state_from_ibqp(), __qedr_alloc_mr() and __qedr_post_send() as static since they are only used in the same file. Signed-off-by: Ram Amrani Signed-off-by: Ariel Elior Signed-off-by: Doug Ledford --- drivers/infiniband/hw/qedr/verbs.c | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-) diff --git a/drivers/infiniband/hw/qedr/verbs.c b/drivers/infiniband/hw/qedr/verbs.c index 3dae9641f821..1b9d55965479 100644 --- a/drivers/infiniband/hw/qedr/verbs.c +++ b/drivers/infiniband/hw/qedr/verbs.c @@ -1600,7 +1600,7 @@ err0: return ERR_PTR(-EFAULT); } -enum ib_qp_state qedr_get_ibqp_state(enum qed_roce_qp_state qp_state) +static enum ib_qp_state qedr_get_ibqp_state(enum qed_roce_qp_state qp_state) { switch (qp_state) { case QED_ROCE_QP_STATE_RESET: @@ -1621,7 +1621,8 @@ enum ib_qp_state qedr_get_ibqp_state(enum qed_roce_qp_state qp_state) return IB_QPS_ERR; } -enum qed_roce_qp_state qedr_get_state_from_ibqp(enum ib_qp_state qp_state) +static enum qed_roce_qp_state qedr_get_state_from_ibqp( + enum ib_qp_state qp_state) { switch (qp_state) { case IB_QPS_RESET: @@ -2310,7 +2311,8 @@ int qedr_dereg_mr(struct ib_mr *ib_mr) return rc; } -struct qedr_mr *__qedr_alloc_mr(struct ib_pd *ibpd, int max_page_list_len) +static struct qedr_mr *__qedr_alloc_mr(struct ib_pd *ibpd, + int max_page_list_len) { struct qedr_pd *pd = get_qedr_pd(ibpd); struct qedr_dev *dev = get_qedr_dev(ibpd->device); @@ -2712,7 +2714,7 @@ static int qedr_prepare_reg(struct qedr_qp *qp, return 0; } -enum ib_wc_opcode qedr_ib_to_wc_opcode(enum ib_wr_opcode opcode) +static enum ib_wc_opcode qedr_ib_to_wc_opcode(enum ib_wr_opcode opcode) { switch (opcode) { case IB_WR_RDMA_WRITE: @@ -2737,7 +2739,7 @@ enum ib_wc_opcode qedr_ib_to_wc_opcode(enum ib_wr_opcode opcode) } } -inline bool qedr_can_post_send(struct qedr_qp *qp, struct ib_send_wr *wr) +static inline bool qedr_can_post_send(struct qedr_qp *qp, struct ib_send_wr *wr) { int wq_is_full, err_wr, pbl_is_full; struct qedr_dev *dev = qp->dev; @@ -2774,7 +2776,7 @@ inline bool qedr_can_post_send(struct qedr_qp *qp, struct ib_send_wr *wr) return true; } -int __qedr_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr, +static int __qedr_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr, struct ib_send_wr **bad_wr) { struct qedr_dev *dev = get_qedr_dev(ibqp->device); From 1a59075197976611bacaa383a6673f9e57e9e98b Mon Sep 17 00:00:00 2001 From: Ram Amrani Date: Tue, 24 Jan 2017 13:51:40 +0200 Subject: [PATCH 212/343] RDMA/qedr: Fix formatting Remove standalone ';'. List function's parameters in a single line. Signed-off-by: Ram Amrani Signed-off-by: Ariel Elior Reviewed-by: Leon Romanovsky Signed-off-by: Doug Ledford --- drivers/infiniband/hw/qedr/main.c | 3 +-- drivers/infiniband/hw/qedr/verbs.c | 1 - 2 files changed, 1 insertion(+), 3 deletions(-) diff --git a/drivers/infiniband/hw/qedr/main.c b/drivers/infiniband/hw/qedr/main.c index 7b74d09a8217..eac0bfc1fc99 100644 --- a/drivers/infiniband/hw/qedr/main.c +++ b/drivers/infiniband/hw/qedr/main.c @@ -576,8 +576,7 @@ static int qedr_set_device_attr(struct qedr_dev *dev) return 0; } -void qedr_unaffiliated_event(void *context, - u8 event_code) +void qedr_unaffiliated_event(void *context, u8 event_code) { pr_err("unaffiliated event not implemented yet\n"); } diff --git a/drivers/infiniband/hw/qedr/verbs.c b/drivers/infiniband/hw/qedr/verbs.c index 1b9d55965479..c4f4c2131a45 100644 --- a/drivers/infiniband/hw/qedr/verbs.c +++ b/drivers/infiniband/hw/qedr/verbs.c @@ -1874,7 +1874,6 @@ int qedr_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr, qp_params.sgid.dwords[2], qp_params.sgid.dwords[3]); DP_DEBUG(dev, QEDR_MSG_QP, "remote_mac=[%pM]\n", qp_params.remote_mac_addr); -; qp_params.mtu = qp->mtu; qp_params.lb_indication = false; From af2b14b8b8ae21b0047a52c767ac8b44f435a280 Mon Sep 17 00:00:00 2001 From: Ram Amrani Date: Tue, 24 Jan 2017 13:51:41 +0200 Subject: [PATCH 213/343] RDMA/qedr: Fix RDMA CM loopback The loopback logic in RDMA CM packets compares Ethernet addresses and was accidently inverse. Signed-off-by: Ram Amrani Signed-off-by: Ariel Elior Signed-off-by: Doug Ledford --- drivers/infiniband/hw/qedr/qedr_cm.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/infiniband/hw/qedr/qedr_cm.c b/drivers/infiniband/hw/qedr/qedr_cm.c index 00361f310d15..a9a8d8745d2e 100644 --- a/drivers/infiniband/hw/qedr/qedr_cm.c +++ b/drivers/infiniband/hw/qedr/qedr_cm.c @@ -398,9 +398,9 @@ static inline int qedr_gsi_build_packet(struct qedr_dev *dev, } if (ether_addr_equal(udh.eth.smac_h, udh.eth.dmac_h)) - packet->tx_dest = QED_ROCE_LL2_TX_DEST_NW; - else packet->tx_dest = QED_ROCE_LL2_TX_DEST_LB; + else + packet->tx_dest = QED_ROCE_LL2_TX_DEST_NW; packet->roce_mode = roce_mode; memcpy(packet->header.vaddr, ud_header_buffer, header_size); From 9c1e0228ab35e52d30abf4b5629c28350833fbcb Mon Sep 17 00:00:00 2001 From: Ram Amrani Date: Tue, 24 Jan 2017 13:51:42 +0200 Subject: [PATCH 214/343] RDMA/qedr: Fix and simplify memory leak in PD alloc Free the PD if no internal resources were available. Move userspace code under the relevant 'if'. Signed-off-by: Ram Amrani Signed-off-by: Ariel Elior Signed-off-by: Doug Ledford --- drivers/infiniband/hw/qedr/verbs.c | 26 ++++++++++++++++++-------- 1 file changed, 18 insertions(+), 8 deletions(-) diff --git a/drivers/infiniband/hw/qedr/verbs.c b/drivers/infiniband/hw/qedr/verbs.c index c4f4c2131a45..c7d6c9a783bd 100644 --- a/drivers/infiniband/hw/qedr/verbs.c +++ b/drivers/infiniband/hw/qedr/verbs.c @@ -471,8 +471,6 @@ struct ib_pd *qedr_alloc_pd(struct ib_device *ibdev, struct ib_ucontext *context, struct ib_udata *udata) { struct qedr_dev *dev = get_qedr_dev(ibdev); - struct qedr_ucontext *uctx = NULL; - struct qedr_alloc_pd_uresp uresp; struct qedr_pd *pd; u16 pd_id; int rc; @@ -489,21 +487,33 @@ struct ib_pd *qedr_alloc_pd(struct ib_device *ibdev, if (!pd) return ERR_PTR(-ENOMEM); - dev->ops->rdma_alloc_pd(dev->rdma_ctx, &pd_id); + rc = dev->ops->rdma_alloc_pd(dev->rdma_ctx, &pd_id); + if (rc) + goto err; - uresp.pd_id = pd_id; pd->pd_id = pd_id; if (udata && context) { + struct qedr_alloc_pd_uresp uresp; + + uresp.pd_id = pd_id; + rc = ib_copy_to_udata(udata, &uresp, sizeof(uresp)); - if (rc) + if (rc) { DP_ERR(dev, "copy error pd_id=0x%x.\n", pd_id); - uctx = get_qedr_ucontext(context); - uctx->pd = pd; - pd->uctx = uctx; + dev->ops->rdma_dealloc_pd(dev->rdma_ctx, pd_id); + goto err; + } + + pd->uctx = get_qedr_ucontext(context); + pd->uctx->pd = pd; } return &pd->ibpd; + +err: + kfree(pd); + return ERR_PTR(rc); } int qedr_dealloc_pd(struct ib_pd *ibpd) From f449c7a2d822c2d81b5bcb2c50eec80796766726 Mon Sep 17 00:00:00 2001 From: Ram Amrani Date: Tue, 24 Jan 2017 13:51:43 +0200 Subject: [PATCH 215/343] RDMA/qedr: Dispatch port active event from qedr_add Relying on qede to trigger qedr on startup is problematic. When probing both if qedr loads slowly then qede can assume qedr is missing and not trigger it. This patch adds a triggering from qedr and protects against a race via an atomic bit. Signed-off-by: Ram Amrani Signed-off-by: Ariel Elior Signed-off-by: Doug Ledford --- drivers/infiniband/hw/qedr/main.c | 20 ++++++++++++++------ drivers/infiniband/hw/qedr/qedr.h | 5 +++++ 2 files changed, 19 insertions(+), 6 deletions(-) diff --git a/drivers/infiniband/hw/qedr/main.c b/drivers/infiniband/hw/qedr/main.c index eac0bfc1fc99..3ac8aa5ef37d 100644 --- a/drivers/infiniband/hw/qedr/main.c +++ b/drivers/infiniband/hw/qedr/main.c @@ -791,6 +791,9 @@ static struct qedr_dev *qedr_add(struct qed_dev *cdev, struct pci_dev *pdev, if (device_create_file(&dev->ibdev.dev, qedr_attributes[i])) goto sysfs_err; + if (!test_and_set_bit(QEDR_ENET_STATE_BIT, &dev->enet_state)) + qedr_ib_dispatch_event(dev, QEDR_PORT, IB_EVENT_PORT_ACTIVE); + DP_DEBUG(dev, QEDR_MSG_INIT, "qedr driver loaded successfully\n"); return dev; @@ -823,11 +826,10 @@ static void qedr_remove(struct qedr_dev *dev) ib_dealloc_device(&dev->ibdev); } -static int qedr_close(struct qedr_dev *dev) +static void qedr_close(struct qedr_dev *dev) { - qedr_ib_dispatch_event(dev, 1, IB_EVENT_PORT_ERR); - - return 0; + if (test_and_clear_bit(QEDR_ENET_STATE_BIT, &dev->enet_state)) + qedr_ib_dispatch_event(dev, QEDR_PORT, IB_EVENT_PORT_ERR); } static void qedr_shutdown(struct qedr_dev *dev) @@ -836,6 +838,12 @@ static void qedr_shutdown(struct qedr_dev *dev) qedr_remove(dev); } +static void qedr_open(struct qedr_dev *dev) +{ + if (!test_and_set_bit(QEDR_ENET_STATE_BIT, &dev->enet_state)) + qedr_ib_dispatch_event(dev, QEDR_PORT, IB_EVENT_PORT_ACTIVE); +} + static void qedr_mac_address_change(struct qedr_dev *dev) { union ib_gid *sgid = &dev->sgid_tbl[0]; @@ -862,7 +870,7 @@ static void qedr_mac_address_change(struct qedr_dev *dev) ether_addr_copy(dev->gsi_ll2_mac_address, dev->ndev->dev_addr); - qedr_ib_dispatch_event(dev, 1, IB_EVENT_GID_CHANGE); + qedr_ib_dispatch_event(dev, QEDR_PORT, IB_EVENT_GID_CHANGE); if (rc) DP_ERR(dev, "Error updating mac filter\n"); @@ -876,7 +884,7 @@ static void qedr_notify(struct qedr_dev *dev, enum qede_roce_event event) { switch (event) { case QEDE_UP: - qedr_ib_dispatch_event(dev, 1, IB_EVENT_PORT_ACTIVE); + qedr_open(dev); break; case QEDE_DOWN: qedr_close(dev); diff --git a/drivers/infiniband/hw/qedr/qedr.h b/drivers/infiniband/hw/qedr/qedr.h index 94319abb0df2..bb32e4792ec9 100644 --- a/drivers/infiniband/hw/qedr/qedr.h +++ b/drivers/infiniband/hw/qedr/qedr.h @@ -113,6 +113,8 @@ struct qedr_device_attr { struct qed_rdma_events events; }; +#define QEDR_ENET_STATE_BIT (0) + struct qedr_dev { struct ib_device ibdev; struct qed_dev *cdev; @@ -153,6 +155,8 @@ struct qedr_dev { struct qedr_cq *gsi_sqcq; struct qedr_cq *gsi_rqcq; struct qedr_qp *gsi_qp; + + unsigned long enet_state; }; #define QEDR_MAX_SQ_PBL (0x8000) @@ -188,6 +192,7 @@ struct qedr_dev { #define QEDR_ROCE_MAX_CNQ_SIZE (0x4000) #define QEDR_MAX_PORT (1) +#define QEDR_PORT (1) #define QEDR_UVERBS(CMD_NAME) (1ull << IB_USER_VERBS_CMD_##CMD_NAME) From 828f6fa65ce7e80f77f5ab12942e44eb3d9d174e Mon Sep 17 00:00:00 2001 From: Kenneth Lee Date: Thu, 5 Jan 2017 15:00:05 +0800 Subject: [PATCH 216/343] IB/umem: Release pid in error and ODP flow 1. Release pid before enter odp flow 2. Release pid when fail to allocate memory Fixes: 87773dd56d54 ("IB: ib_umem_release() should decrement mm->pinned_vm from ib_umem_get") Fixes: 8ada2c1c0c1d ("IB/core: Add support for on demand paging regions") Signed-off-by: Kenneth Lee Reviewed-by: Haggai Eran Reviewed-by: Leon Romanovsky Signed-off-by: Doug Ledford --- drivers/infiniband/core/umem.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/infiniband/core/umem.c b/drivers/infiniband/core/umem.c index 1e62a5f0cb28..4609b921f899 100644 --- a/drivers/infiniband/core/umem.c +++ b/drivers/infiniband/core/umem.c @@ -134,6 +134,7 @@ struct ib_umem *ib_umem_get(struct ib_ucontext *context, unsigned long addr, IB_ACCESS_REMOTE_ATOMIC | IB_ACCESS_MW_BIND)); if (access & IB_ACCESS_ON_DEMAND) { + put_pid(umem->pid); ret = ib_umem_odp_get(context, umem); if (ret) { kfree(umem); @@ -149,6 +150,7 @@ struct ib_umem *ib_umem_get(struct ib_ucontext *context, unsigned long addr, page_list = (struct page **) __get_free_page(GFP_KERNEL); if (!page_list) { + put_pid(umem->pid); kfree(umem); return ERR_PTR(-ENOMEM); } From 5ce6b04ce96896e8a79e6f60740ced911eaac7a4 Mon Sep 17 00:00:00 2001 From: Liping Zhang Date: Sun, 22 Jan 2017 22:10:32 +0800 Subject: [PATCH 217/343] netfilter: nft_log: restrict the log prefix length to 127 First, log prefix will be truncated to NF_LOG_PREFIXLEN-1, i.e. 127, at nf_log_packet(), so the extra part is useless. Second, after adding a log rule with a very very long prefix, we will fail to dump the nft rules after this _special_ one, but acctually, they do exist. For example: # name_65000=$(printf "%0.sQ" {1..65000}) # nft add rule filter output log prefix "$name_65000" # nft add rule filter output counter # nft add rule filter output counter # nft list chain filter output table ip filter { chain output { type filter hook output priority 0; policy accept; } } So now, restrict the log prefix length to NF_LOG_PREFIXLEN-1. Fixes: 96518518cc41 ("netfilter: add nftables") Signed-off-by: Liping Zhang Signed-off-by: Pablo Neira Ayuso --- include/uapi/linux/netfilter/nf_log.h | 2 ++ net/netfilter/nf_log.c | 1 - net/netfilter/nft_log.c | 3 ++- 3 files changed, 4 insertions(+), 2 deletions(-) diff --git a/include/uapi/linux/netfilter/nf_log.h b/include/uapi/linux/netfilter/nf_log.h index 8be21e02387d..d0b5fa91ff54 100644 --- a/include/uapi/linux/netfilter/nf_log.h +++ b/include/uapi/linux/netfilter/nf_log.h @@ -9,4 +9,6 @@ #define NF_LOG_MACDECODE 0x20 /* Decode MAC header */ #define NF_LOG_MASK 0x2f +#define NF_LOG_PREFIXLEN 128 + #endif /* _NETFILTER_NF_LOG_H */ diff --git a/net/netfilter/nf_log.c b/net/netfilter/nf_log.c index 3dca90dc24ad..ffb9e8ada899 100644 --- a/net/netfilter/nf_log.c +++ b/net/netfilter/nf_log.c @@ -13,7 +13,6 @@ /* Internal logging interface, which relies on the real LOG target modules */ -#define NF_LOG_PREFIXLEN 128 #define NFLOGGER_NAME_LEN 64 static struct nf_logger __rcu *loggers[NFPROTO_NUMPROTO][NF_LOG_TYPE_MAX] __read_mostly; diff --git a/net/netfilter/nft_log.c b/net/netfilter/nft_log.c index 6271e40a3dd6..6f6e64423643 100644 --- a/net/netfilter/nft_log.c +++ b/net/netfilter/nft_log.c @@ -39,7 +39,8 @@ static void nft_log_eval(const struct nft_expr *expr, static const struct nla_policy nft_log_policy[NFTA_LOG_MAX + 1] = { [NFTA_LOG_GROUP] = { .type = NLA_U16 }, - [NFTA_LOG_PREFIX] = { .type = NLA_STRING }, + [NFTA_LOG_PREFIX] = { .type = NLA_STRING, + .len = NF_LOG_PREFIXLEN - 1 }, [NFTA_LOG_SNAPLEN] = { .type = NLA_U32 }, [NFTA_LOG_QTHRESHOLD] = { .type = NLA_U16 }, [NFTA_LOG_LEVEL] = { .type = NLA_U32 }, From 35d0ac9070ef619e3bf44324375878a1c540387b Mon Sep 17 00:00:00 2001 From: Pablo Neira Ayuso Date: Tue, 24 Jan 2017 00:51:32 +0100 Subject: [PATCH 218/343] netfilter: nf_tables: fix set->nelems counting with no NLM_F_EXCL If the element exists and no NLM_F_EXCL is specified, do not bump set->nelems, otherwise we leak one set element slot. This problem amplifies if the set is full since the abort path always decrements the counter for the -ENFILE case too, giving one spare extra slot. Fix this by moving set->nelems update to nft_add_set_elem() after successful element insertion. Moreover, remove the element if the set is full so there is no need to rely on the abort path to undo things anymore. Fixes: c016c7e45ddf ("netfilter: nf_tables: honor NLM_F_EXCL flag in set element insertion") Signed-off-by: Pablo Neira Ayuso --- net/netfilter/nf_tables_api.c | 16 +++++++++------- 1 file changed, 9 insertions(+), 7 deletions(-) diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c index b84c7b25219b..831a9a16f563 100644 --- a/net/netfilter/nf_tables_api.c +++ b/net/netfilter/nf_tables_api.c @@ -3745,10 +3745,18 @@ static int nft_add_set_elem(struct nft_ctx *ctx, struct nft_set *set, goto err5; } + if (set->size && + !atomic_add_unless(&set->nelems, 1, set->size + set->ndeact)) { + err = -ENFILE; + goto err6; + } + nft_trans_elem(trans) = elem; list_add_tail(&trans->list, &ctx->net->nft.commit_list); return 0; +err6: + set->ops->remove(set, &elem); err5: kfree(trans); err4: @@ -3795,15 +3803,9 @@ static int nf_tables_newsetelem(struct net *net, struct sock *nlsk, return -EBUSY; nla_for_each_nested(attr, nla[NFTA_SET_ELEM_LIST_ELEMENTS], rem) { - if (set->size && - !atomic_add_unless(&set->nelems, 1, set->size + set->ndeact)) - return -ENFILE; - err = nft_add_set_elem(&ctx, set, attr, nlh->nlmsg_flags); - if (err < 0) { - atomic_dec(&set->nelems); + if (err < 0) break; - } } return err; } From de70185de0333783154863278ac87bfbbc54e384 Mon Sep 17 00:00:00 2001 From: Pablo Neira Ayuso Date: Tue, 24 Jan 2017 00:51:41 +0100 Subject: [PATCH 219/343] netfilter: nf_tables: deconstify walk callback function The flush operation needs to modify set and element objects, so let's deconstify this. Signed-off-by: Pablo Neira Ayuso --- include/net/netfilter/nf_tables.h | 6 +++--- net/netfilter/nf_tables_api.c | 24 ++++++++++++------------ net/netfilter/nft_set_hash.c | 2 +- net/netfilter/nft_set_rbtree.c | 2 +- 4 files changed, 17 insertions(+), 17 deletions(-) diff --git a/include/net/netfilter/nf_tables.h b/include/net/netfilter/nf_tables.h index 924325c46aab..7dfdb517f0be 100644 --- a/include/net/netfilter/nf_tables.h +++ b/include/net/netfilter/nf_tables.h @@ -207,9 +207,9 @@ struct nft_set_iter { unsigned int skip; int err; int (*fn)(const struct nft_ctx *ctx, - const struct nft_set *set, + struct nft_set *set, const struct nft_set_iter *iter, - const struct nft_set_elem *elem); + struct nft_set_elem *elem); }; /** @@ -301,7 +301,7 @@ struct nft_set_ops { void (*remove)(const struct nft_set *set, const struct nft_set_elem *elem); void (*walk)(const struct nft_ctx *ctx, - const struct nft_set *set, + struct nft_set *set, struct nft_set_iter *iter); unsigned int (*privsize)(const struct nlattr * const nla[]); diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c index 831a9a16f563..5bd0068320fb 100644 --- a/net/netfilter/nf_tables_api.c +++ b/net/netfilter/nf_tables_api.c @@ -3087,9 +3087,9 @@ static int nf_tables_delset(struct net *net, struct sock *nlsk, } static int nf_tables_bind_check_setelem(const struct nft_ctx *ctx, - const struct nft_set *set, + struct nft_set *set, const struct nft_set_iter *iter, - const struct nft_set_elem *elem) + struct nft_set_elem *elem) { const struct nft_set_ext *ext = nft_set_elem_ext(set, elem->priv); enum nft_registers dreg; @@ -3308,9 +3308,9 @@ struct nft_set_dump_args { }; static int nf_tables_dump_setelem(const struct nft_ctx *ctx, - const struct nft_set *set, + struct nft_set *set, const struct nft_set_iter *iter, - const struct nft_set_elem *elem) + struct nft_set_elem *elem) { struct nft_set_dump_args *args; @@ -3322,7 +3322,7 @@ static int nf_tables_dump_set(struct sk_buff *skb, struct netlink_callback *cb) { struct net *net = sock_net(skb->sk); u8 genmask = nft_genmask_cur(net); - const struct nft_set *set; + struct nft_set *set; struct nft_set_dump_args args; struct nft_ctx ctx; struct nlattr *nla[NFTA_SET_ELEM_LIST_MAX + 1]; @@ -3890,9 +3890,9 @@ err1: } static int nft_flush_set(const struct nft_ctx *ctx, - const struct nft_set *set, + struct nft_set *set, const struct nft_set_iter *iter, - const struct nft_set_elem *elem) + struct nft_set_elem *elem) { struct nft_trans *trans; int err; @@ -3907,8 +3907,8 @@ static int nft_flush_set(const struct nft_ctx *ctx, goto err1; } - nft_trans_elem_set(trans) = (struct nft_set *)set; - nft_trans_elem(trans) = *((struct nft_set_elem *)elem); + nft_trans_elem_set(trans) = set; + nft_trans_elem(trans) = *elem; list_add_tail(&trans->list, &ctx->net->nft.commit_list); return 0; @@ -5019,9 +5019,9 @@ static int nf_tables_check_loops(const struct nft_ctx *ctx, const struct nft_chain *chain); static int nf_tables_loop_check_setelem(const struct nft_ctx *ctx, - const struct nft_set *set, + struct nft_set *set, const struct nft_set_iter *iter, - const struct nft_set_elem *elem) + struct nft_set_elem *elem) { const struct nft_set_ext *ext = nft_set_elem_ext(set, elem->priv); const struct nft_data *data; @@ -5045,7 +5045,7 @@ static int nf_tables_check_loops(const struct nft_ctx *ctx, { const struct nft_rule *rule; const struct nft_expr *expr, *last; - const struct nft_set *set; + struct nft_set *set; struct nft_set_binding *binding; struct nft_set_iter iter; diff --git a/net/netfilter/nft_set_hash.c b/net/netfilter/nft_set_hash.c index 1e20e2bbb6d9..e36069fb76ae 100644 --- a/net/netfilter/nft_set_hash.c +++ b/net/netfilter/nft_set_hash.c @@ -212,7 +212,7 @@ static void nft_hash_remove(const struct nft_set *set, rhashtable_remove_fast(&priv->ht, &he->node, nft_hash_params); } -static void nft_hash_walk(const struct nft_ctx *ctx, const struct nft_set *set, +static void nft_hash_walk(const struct nft_ctx *ctx, struct nft_set *set, struct nft_set_iter *iter) { struct nft_hash *priv = nft_set_priv(set); diff --git a/net/netfilter/nft_set_rbtree.c b/net/netfilter/nft_set_rbtree.c index 08376e50f6cd..f06f55ee516d 100644 --- a/net/netfilter/nft_set_rbtree.c +++ b/net/netfilter/nft_set_rbtree.c @@ -221,7 +221,7 @@ static void *nft_rbtree_deactivate(const struct net *net, } static void nft_rbtree_walk(const struct nft_ctx *ctx, - const struct nft_set *set, + struct nft_set *set, struct nft_set_iter *iter) { const struct nft_rbtree *priv = nft_set_priv(set); From b2c11e4b9536ebab6b39929e1fe15f57039ab445 Mon Sep 17 00:00:00 2001 From: Pablo Neira Ayuso Date: Tue, 24 Jan 2017 00:51:48 +0100 Subject: [PATCH 220/343] netfilter: nf_tables: bump set->ndeact on set flush Add missing set->ndeact update on each deactivated element from the set flush path. Otherwise, sets with fixed size break after flush since accounting breaks. # nft add set x y { type ipv4_addr\; size 2\; } # nft add element x y { 1.1.1.1 } # nft add element x y { 1.1.1.2 } # nft flush set x y # nft add element x y { 1.1.1.1 } :1:1-28: Error: Could not process rule: Too many open files in system Fixes: 8411b6442e59 ("netfilter: nf_tables: support for set flushing") Reported-by: Elise Lennion Signed-off-by: Pablo Neira Ayuso --- net/netfilter/nf_tables_api.c | 1 + 1 file changed, 1 insertion(+) diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c index 5bd0068320fb..1b913760f205 100644 --- a/net/netfilter/nf_tables_api.c +++ b/net/netfilter/nf_tables_api.c @@ -3906,6 +3906,7 @@ static int nft_flush_set(const struct nft_ctx *ctx, err = -ENOENT; goto err1; } + set->ndeact++; nft_trans_elem_set(trans) = set; nft_trans_elem(trans) = *elem; From 93f955aad4bacee5acebad141d1a03cd51f27b4e Mon Sep 17 00:00:00 2001 From: Parthasarathy Bhuvaragan Date: Tue, 24 Jan 2017 13:00:43 +0100 Subject: [PATCH 221/343] tipc: fix nametbl_lock soft lockup at node/link events We trigger a soft lockup as we grab nametbl_lock twice if the node has a pending node up/down or link up/down event while: - we process an incoming named message in tipc_named_rcv() and perform an tipc_update_nametbl(). - we have pending backlog items in the name distributor queue during a nametable update using tipc_nametbl_publish() or tipc_nametbl_withdraw(). The following are the call chain associated: tipc_named_rcv() Grabs nametbl_lock tipc_update_nametbl() (publish/withdraw) tipc_node_subscribe()/unsubscribe() tipc_node_write_unlock() << lockup occurs if an outstanding node/link event exits, as we grabs nametbl_lock again >> tipc_nametbl_withdraw() Grab nametbl_lock tipc_named_process_backlog() tipc_update_nametbl() << rest as above >> The function tipc_node_write_unlock(), in addition to releasing the lock processes the outstanding node/link up/down events. To do this, we need to grab the nametbl_lock again leading to the lockup. In this commit we fix the soft lockup by introducing a fast variant of node_unlock(), where we just release the lock. We adapt the node_subscribe()/node_unsubscribe() to use the fast variants. Reported-and-Tested-by: John Thompson Acked-by: Ying Xue Acked-by: Jon Maloy Signed-off-by: Parthasarathy Bhuvaragan Signed-off-by: David S. Miller --- net/tipc/node.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) diff --git a/net/tipc/node.c b/net/tipc/node.c index 9d2f4c2b08ab..27753325e06e 100644 --- a/net/tipc/node.c +++ b/net/tipc/node.c @@ -263,6 +263,11 @@ static void tipc_node_write_lock(struct tipc_node *n) write_lock_bh(&n->lock); } +static void tipc_node_write_unlock_fast(struct tipc_node *n) +{ + write_unlock_bh(&n->lock); +} + static void tipc_node_write_unlock(struct tipc_node *n) { struct net *net = n->net; @@ -417,7 +422,7 @@ void tipc_node_subscribe(struct net *net, struct list_head *subscr, u32 addr) } tipc_node_write_lock(n); list_add_tail(subscr, &n->publ_list); - tipc_node_write_unlock(n); + tipc_node_write_unlock_fast(n); tipc_node_put(n); } @@ -435,7 +440,7 @@ void tipc_node_unsubscribe(struct net *net, struct list_head *subscr, u32 addr) } tipc_node_write_lock(n); list_del_init(subscr); - tipc_node_write_unlock(n); + tipc_node_write_unlock_fast(n); tipc_node_put(n); } From d094c4d5f5c7e1b225e94227ca3f007be3adc4e8 Mon Sep 17 00:00:00 2001 From: Parthasarathy Bhuvaragan Date: Tue, 24 Jan 2017 13:00:44 +0100 Subject: [PATCH 222/343] tipc: add subscription refcount to avoid invalid delete Until now, the subscribers keep track of the subscriptions using reference count at subscriber level. At subscription cancel or subscriber delete, we delete the subscription only if the timer was pending for the subscription. This approach is incorrect as: 1. del_timer() is not SMP safe, if on CPU0 the check for pending timer returns true but CPU1 might schedule the timer callback thereby deleting the subscription. Thus when CPU0 is scheduled, it deletes an invalid subscription. 2. We export tipc_subscrp_report_overlap(), which accesses the subscription pointer multiple times. Meanwhile the subscription timer can expire thereby freeing the subscription and we might continue to access the subscription pointer leading to memory violations. In this commit, we introduce subscription refcount to avoid deleting an invalid subscription. Reported-and-Tested-by: John Thompson Acked-by: Ying Xue Acked-by: Jon Maloy Signed-off-by: Parthasarathy Bhuvaragan Signed-off-by: David S. Miller --- net/tipc/subscr.c | 126 ++++++++++++++++++++++++++-------------------- net/tipc/subscr.h | 1 + 2 files changed, 72 insertions(+), 55 deletions(-) diff --git a/net/tipc/subscr.c b/net/tipc/subscr.c index 0dd02244e21d..9d94e65d0894 100644 --- a/net/tipc/subscr.c +++ b/net/tipc/subscr.c @@ -54,6 +54,8 @@ struct tipc_subscriber { static void tipc_subscrp_delete(struct tipc_subscription *sub); static void tipc_subscrb_put(struct tipc_subscriber *subscriber); +static void tipc_subscrp_put(struct tipc_subscription *subscription); +static void tipc_subscrp_get(struct tipc_subscription *subscription); /** * htohl - convert value to endianness used by destination @@ -123,6 +125,7 @@ void tipc_subscrp_report_overlap(struct tipc_subscription *sub, u32 found_lower, { struct tipc_name_seq seq; + tipc_subscrp_get(sub); tipc_subscrp_convert_seq(&sub->evt.s.seq, sub->swap, &seq); if (!tipc_subscrp_check_overlap(&seq, found_lower, found_upper)) return; @@ -132,30 +135,23 @@ void tipc_subscrp_report_overlap(struct tipc_subscription *sub, u32 found_lower, tipc_subscrp_send_event(sub, found_lower, found_upper, event, port_ref, node); + tipc_subscrp_put(sub); } static void tipc_subscrp_timeout(unsigned long data) { struct tipc_subscription *sub = (struct tipc_subscription *)data; - struct tipc_subscriber *subscriber = sub->subscriber; /* Notify subscriber of timeout */ tipc_subscrp_send_event(sub, sub->evt.s.seq.lower, sub->evt.s.seq.upper, TIPC_SUBSCR_TIMEOUT, 0, 0); - spin_lock_bh(&subscriber->lock); - tipc_subscrp_delete(sub); - spin_unlock_bh(&subscriber->lock); - - tipc_subscrb_put(subscriber); + tipc_subscrp_put(sub); } static void tipc_subscrb_kref_release(struct kref *kref) { - struct tipc_subscriber *subcriber = container_of(kref, - struct tipc_subscriber, kref); - - kfree(subcriber); + kfree(container_of(kref,struct tipc_subscriber, kref)); } static void tipc_subscrb_put(struct tipc_subscriber *subscriber) @@ -168,6 +164,59 @@ static void tipc_subscrb_get(struct tipc_subscriber *subscriber) kref_get(&subscriber->kref); } +static void tipc_subscrp_kref_release(struct kref *kref) +{ + struct tipc_subscription *sub = container_of(kref, + struct tipc_subscription, + kref); + struct tipc_net *tn = net_generic(sub->net, tipc_net_id); + struct tipc_subscriber *subscriber = sub->subscriber; + + spin_lock_bh(&subscriber->lock); + tipc_nametbl_unsubscribe(sub); + list_del(&sub->subscrp_list); + atomic_dec(&tn->subscription_count); + spin_unlock_bh(&subscriber->lock); + kfree(sub); + tipc_subscrb_put(subscriber); +} + +static void tipc_subscrp_put(struct tipc_subscription *subscription) +{ + kref_put(&subscription->kref, tipc_subscrp_kref_release); +} + +static void tipc_subscrp_get(struct tipc_subscription *subscription) +{ + kref_get(&subscription->kref); +} + +/* tipc_subscrb_subscrp_delete - delete a specific subscription or all + * subscriptions for a given subscriber. + */ +static void tipc_subscrb_subscrp_delete(struct tipc_subscriber *subscriber, + struct tipc_subscr *s) +{ + struct list_head *subscription_list = &subscriber->subscrp_list; + struct tipc_subscription *sub, *temp; + + spin_lock_bh(&subscriber->lock); + list_for_each_entry_safe(sub, temp, subscription_list, subscrp_list) { + if (s && memcmp(s, &sub->evt.s, sizeof(struct tipc_subscr))) + continue; + + tipc_subscrp_get(sub); + spin_unlock_bh(&subscriber->lock); + tipc_subscrp_delete(sub); + tipc_subscrp_put(sub); + spin_lock_bh(&subscriber->lock); + + if (s) + break; + } + spin_unlock_bh(&subscriber->lock); +} + static struct tipc_subscriber *tipc_subscrb_create(int conid) { struct tipc_subscriber *subscriber; @@ -177,8 +226,8 @@ static struct tipc_subscriber *tipc_subscrb_create(int conid) pr_warn("Subscriber rejected, no memory\n"); return NULL; } - kref_init(&subscriber->kref); INIT_LIST_HEAD(&subscriber->subscrp_list); + kref_init(&subscriber->kref); subscriber->conid = conid; spin_lock_init(&subscriber->lock); @@ -187,55 +236,22 @@ static struct tipc_subscriber *tipc_subscrb_create(int conid) static void tipc_subscrb_delete(struct tipc_subscriber *subscriber) { - struct tipc_subscription *sub, *temp; - u32 timeout; - - spin_lock_bh(&subscriber->lock); - /* Destroy any existing subscriptions for subscriber */ - list_for_each_entry_safe(sub, temp, &subscriber->subscrp_list, - subscrp_list) { - timeout = htohl(sub->evt.s.timeout, sub->swap); - if ((timeout == TIPC_WAIT_FOREVER) || del_timer(&sub->timer)) { - tipc_subscrp_delete(sub); - tipc_subscrb_put(subscriber); - } - } - spin_unlock_bh(&subscriber->lock); - + tipc_subscrb_subscrp_delete(subscriber, NULL); tipc_subscrb_put(subscriber); } static void tipc_subscrp_delete(struct tipc_subscription *sub) { - struct tipc_net *tn = net_generic(sub->net, tipc_net_id); + u32 timeout = htohl(sub->evt.s.timeout, sub->swap); - tipc_nametbl_unsubscribe(sub); - list_del(&sub->subscrp_list); - kfree(sub); - atomic_dec(&tn->subscription_count); + if (timeout == TIPC_WAIT_FOREVER || del_timer(&sub->timer)) + tipc_subscrp_put(sub); } static void tipc_subscrp_cancel(struct tipc_subscr *s, struct tipc_subscriber *subscriber) { - struct tipc_subscription *sub, *temp; - u32 timeout; - - spin_lock_bh(&subscriber->lock); - /* Find first matching subscription, exit if not found */ - list_for_each_entry_safe(sub, temp, &subscriber->subscrp_list, - subscrp_list) { - if (!memcmp(s, &sub->evt.s, sizeof(struct tipc_subscr))) { - timeout = htohl(sub->evt.s.timeout, sub->swap); - if ((timeout == TIPC_WAIT_FOREVER) || - del_timer(&sub->timer)) { - tipc_subscrp_delete(sub); - tipc_subscrb_put(subscriber); - } - break; - } - } - spin_unlock_bh(&subscriber->lock); + tipc_subscrb_subscrp_delete(subscriber, s); } static struct tipc_subscription *tipc_subscrp_create(struct net *net, @@ -272,6 +288,7 @@ static struct tipc_subscription *tipc_subscrp_create(struct net *net, sub->swap = swap; memcpy(&sub->evt.s, s, sizeof(*s)); atomic_inc(&tn->subscription_count); + kref_init(&sub->kref); return sub; } @@ -288,17 +305,16 @@ static void tipc_subscrp_subscribe(struct net *net, struct tipc_subscr *s, spin_lock_bh(&subscriber->lock); list_add(&sub->subscrp_list, &subscriber->subscrp_list); - tipc_subscrb_get(subscriber); sub->subscriber = subscriber; tipc_nametbl_subscribe(sub); + tipc_subscrb_get(subscriber); spin_unlock_bh(&subscriber->lock); - timeout = htohl(sub->evt.s.timeout, swap); - if (timeout == TIPC_WAIT_FOREVER) - return; - setup_timer(&sub->timer, tipc_subscrp_timeout, (unsigned long)sub); - mod_timer(&sub->timer, jiffies + msecs_to_jiffies(timeout)); + timeout = htohl(sub->evt.s.timeout, swap); + + if (timeout != TIPC_WAIT_FOREVER) + mod_timer(&sub->timer, jiffies + msecs_to_jiffies(timeout)); } /* Handle one termination request for the subscriber */ diff --git a/net/tipc/subscr.h b/net/tipc/subscr.h index be60103082c9..ffdc214c117a 100644 --- a/net/tipc/subscr.h +++ b/net/tipc/subscr.h @@ -57,6 +57,7 @@ struct tipc_subscriber; * @evt: template for events generated by subscription */ struct tipc_subscription { + struct kref kref; struct tipc_subscriber *subscriber; struct net *net; struct timer_list timer; From fc0adfc8fd18b61b6f7a3f28b429e134d6f3a008 Mon Sep 17 00:00:00 2001 From: Parthasarathy Bhuvaragan Date: Tue, 24 Jan 2017 13:00:45 +0100 Subject: [PATCH 223/343] tipc: fix connection refcount error Until now, the generic server framework maintains the connection id's per subscriber in server's conn_idr. At tipc_close_conn, we remove the connection id from the server list, but the connection is valid until we call the refcount cleanup. Hence we have a window where the server allocates the same connection to an new subscriber leading to inconsistent reference count. We have another refcount warning we grab the refcount in tipc_conn_lookup() for connections with flag with CF_CONNECTED not set. This usually occurs at shutdown when the we stop the topology server and withdraw TIPC_CFG_SRV publication thereby triggering a withdraw message to subscribers. In this commit, we: 1. remove the connection from the server list at recount cleanup. 2. grab the refcount for a connection only if CF_CONNECTED is set. Tested-by: John Thompson Acked-by: Ying Xue Acked-by: Jon Maloy Signed-off-by: Parthasarathy Bhuvaragan Signed-off-by: David S. Miller --- net/tipc/server.c | 19 ++++++++++--------- 1 file changed, 10 insertions(+), 9 deletions(-) diff --git a/net/tipc/server.c b/net/tipc/server.c index 215849ce453d..2e803601aa99 100644 --- a/net/tipc/server.c +++ b/net/tipc/server.c @@ -91,7 +91,8 @@ static void tipc_sock_release(struct tipc_conn *con); static void tipc_conn_kref_release(struct kref *kref) { struct tipc_conn *con = container_of(kref, struct tipc_conn, kref); - struct sockaddr_tipc *saddr = con->server->saddr; + struct tipc_server *s = con->server; + struct sockaddr_tipc *saddr = s->saddr; struct socket *sock = con->sock; struct sock *sk; @@ -106,6 +107,11 @@ static void tipc_conn_kref_release(struct kref *kref) tipc_sock_release(con); sock_release(sock); con->sock = NULL; + + spin_lock_bh(&s->idr_lock); + idr_remove(&s->conn_idr, con->conid); + s->idr_in_use--; + spin_unlock_bh(&s->idr_lock); } tipc_clean_outqueues(con); @@ -128,8 +134,10 @@ static struct tipc_conn *tipc_conn_lookup(struct tipc_server *s, int conid) spin_lock_bh(&s->idr_lock); con = idr_find(&s->conn_idr, conid); - if (con) + if (con && test_bit(CF_CONNECTED, &con->flags)) conn_get(con); + else + con = NULL; spin_unlock_bh(&s->idr_lock); return con; } @@ -198,15 +206,8 @@ static void tipc_sock_release(struct tipc_conn *con) static void tipc_close_conn(struct tipc_conn *con) { - struct tipc_server *s = con->server; - if (test_and_clear_bit(CF_CONNECTED, &con->flags)) { - spin_lock_bh(&s->idr_lock); - idr_remove(&s->conn_idr, con->conid); - s->idr_in_use--; - spin_unlock_bh(&s->idr_lock); - /* We shouldn't flush pending works as we may be in the * thread. In fact the races with pending rx/tx work structs * are harmless for us here as we have already deleted this From 9dc3abdd1f7ea524e8552e0a3ef01219892ed1f4 Mon Sep 17 00:00:00 2001 From: Parthasarathy Bhuvaragan Date: Tue, 24 Jan 2017 13:00:46 +0100 Subject: [PATCH 224/343] tipc: fix nametbl_lock soft lockup at module exit Commit 333f796235a527 ("tipc: fix a race condition leading to subscriber refcnt bug") reveals a soft lockup while acquiring nametbl_lock. Before commit 333f796235a527, we call tipc_conn_shutdown() from tipc_close_conn() in the context of tipc_topsrv_stop(). In that context, we are allowed to grab the nametbl_lock. Commit 333f796235a527, moved tipc_conn_release (renamed from tipc_conn_shutdown) to the connection refcount cleanup. This allows either tipc_nametbl_withdraw() or tipc_topsrv_stop() to the cleanup. Since tipc_exit_net() first calls tipc_topsrv_stop() and then tipc_nametble_withdraw() increases the chances for the later to perform the connection cleanup. The soft lockup occurs in the call chain of tipc_nametbl_withdraw(), when it performs the tipc_conn_kref_release() as it tries to grab nametbl_lock again while holding it already. tipc_nametbl_withdraw() grabs nametbl_lock tipc_nametbl_remove_publ() tipc_subscrp_report_overlap() tipc_subscrp_send_event() tipc_conn_sendmsg() << if (con->flags != CF_CONNECTED) we do conn_put(), triggering the cleanup as refcount=0. >> tipc_conn_kref_release tipc_sock_release tipc_conn_release tipc_subscrb_delete tipc_subscrp_delete tipc_nametbl_unsubscribe << Soft Lockup >> The previous changes in this series fixes the race conditions fixed by commit 333f796235a527. Hence we can now revert the commit. Fixes: 333f796235a52727 ("tipc: fix a race condition leading to subscriber refcnt bug") Reported-and-Tested-by: John Thompson Acked-by: Ying Xue Acked-by: Jon Maloy Signed-off-by: Parthasarathy Bhuvaragan Signed-off-by: David S. Miller --- net/tipc/server.c | 16 +++++----------- 1 file changed, 5 insertions(+), 11 deletions(-) diff --git a/net/tipc/server.c b/net/tipc/server.c index 2e803601aa99..826cde2c401e 100644 --- a/net/tipc/server.c +++ b/net/tipc/server.c @@ -86,7 +86,6 @@ struct outqueue_entry { static void tipc_recv_work(struct work_struct *work); static void tipc_send_work(struct work_struct *work); static void tipc_clean_outqueues(struct tipc_conn *con); -static void tipc_sock_release(struct tipc_conn *con); static void tipc_conn_kref_release(struct kref *kref) { @@ -104,7 +103,6 @@ static void tipc_conn_kref_release(struct kref *kref) } saddr->scope = -TIPC_NODE_SCOPE; kernel_bind(sock, (struct sockaddr *)saddr, sizeof(*saddr)); - tipc_sock_release(con); sock_release(sock); con->sock = NULL; @@ -194,19 +192,15 @@ static void tipc_unregister_callbacks(struct tipc_conn *con) write_unlock_bh(&sk->sk_callback_lock); } -static void tipc_sock_release(struct tipc_conn *con) +static void tipc_close_conn(struct tipc_conn *con) { struct tipc_server *s = con->server; - if (con->conid) - s->tipc_conn_release(con->conid, con->usr_data); - - tipc_unregister_callbacks(con); -} - -static void tipc_close_conn(struct tipc_conn *con) -{ if (test_and_clear_bit(CF_CONNECTED, &con->flags)) { + tipc_unregister_callbacks(con); + + if (con->conid) + s->tipc_conn_release(con->conid, con->usr_data); /* We shouldn't flush pending works as we may be in the * thread. In fact the races with pending rx/tx work structs From 4c887aa65d38633885010277f3482400681be719 Mon Sep 17 00:00:00 2001 From: Parthasarathy Bhuvaragan Date: Tue, 24 Jan 2017 13:00:47 +0100 Subject: [PATCH 225/343] tipc: ignore requests when the connection state is not CONNECTED In tipc_conn_sendmsg(), we first queue the request to the outqueue followed by the connection state check. If the connection is not connected, we should not queue this message. In this commit, we reject the messages if the connection state is not CF_CONNECTED. Acked-by: Ying Xue Acked-by: Jon Maloy Tested-by: John Thompson Signed-off-by: Parthasarathy Bhuvaragan Signed-off-by: David S. Miller --- net/tipc/server.c | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) diff --git a/net/tipc/server.c b/net/tipc/server.c index 826cde2c401e..04ff441b8065 100644 --- a/net/tipc/server.c +++ b/net/tipc/server.c @@ -453,6 +453,11 @@ int tipc_conn_sendmsg(struct tipc_server *s, int conid, if (!con) return -EINVAL; + if (!test_bit(CF_CONNECTED, &con->flags)) { + conn_put(con); + return 0; + } + e = tipc_alloc_entry(data, len); if (!e) { conn_put(con); @@ -466,12 +471,8 @@ int tipc_conn_sendmsg(struct tipc_server *s, int conid, list_add_tail(&e->list, &con->outqueue); spin_unlock_bh(&con->outqueue_lock); - if (test_bit(CF_CONNECTED, &con->flags)) { - if (!queue_work(s->send_wq, &con->swork)) - conn_put(con); - } else { + if (!queue_work(s->send_wq, &con->swork)) conn_put(con); - } return 0; } @@ -495,7 +496,7 @@ static void tipc_send_to_sock(struct tipc_conn *con) int ret; spin_lock_bh(&con->outqueue_lock); - while (1) { + while (test_bit(CF_CONNECTED, &con->flags)) { e = list_entry(con->outqueue.next, struct outqueue_entry, list); if ((struct list_head *) e == &con->outqueue) From 35e22e49a5d6a741ebe7f2dd280b2052c3003ef7 Mon Sep 17 00:00:00 2001 From: Parthasarathy Bhuvaragan Date: Tue, 24 Jan 2017 13:00:48 +0100 Subject: [PATCH 226/343] tipc: fix cleanup at module unload In tipc_server_stop(), we iterate over the connections with limiting factor as server's idr_in_use. We ignore the fact that this variable is decremented in tipc_close_conn(), leading to premature exit. In this commit, we iterate until the we have no connections left. Acked-by: Ying Xue Acked-by: Jon Maloy Tested-by: John Thompson Signed-off-by: Parthasarathy Bhuvaragan Signed-off-by: David S. Miller --- net/tipc/server.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/net/tipc/server.c b/net/tipc/server.c index 04ff441b8065..3cd6402e812c 100644 --- a/net/tipc/server.c +++ b/net/tipc/server.c @@ -619,14 +619,12 @@ int tipc_server_start(struct tipc_server *s) void tipc_server_stop(struct tipc_server *s) { struct tipc_conn *con; - int total = 0; int id; spin_lock_bh(&s->idr_lock); - for (id = 0; total < s->idr_in_use; id++) { + for (id = 0; s->idr_in_use; id++) { con = idr_find(&s->conn_idr, id); if (con) { - total++; spin_unlock_bh(&s->idr_lock); tipc_close_conn(con); spin_lock_bh(&s->idr_lock); From f39f775218a7520e3700de2003c84a042c3b5972 Mon Sep 17 00:00:00 2001 From: Maor Gottlieb Date: Thu, 19 Jan 2017 15:25:58 +0200 Subject: [PATCH 227/343] IB/rxe: Fix rxe dev insertion to rxe_dev_list The first argument of list_add_tail is the new item and the second is the head of the list. Fix the code to pass arguments in the right order, otherwise not all the rxe devices will be removed during teardown. Fixes: 8700e3e7c4857 ('Soft RoCE driver') Signed-off-by: Maor Gottlieb Reviewed-by: Moni Shoua Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- drivers/infiniband/sw/rxe/rxe_net.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/infiniband/sw/rxe/rxe_net.c b/drivers/infiniband/sw/rxe/rxe_net.c index 342e78163613..4abdeb359fb4 100644 --- a/drivers/infiniband/sw/rxe/rxe_net.c +++ b/drivers/infiniband/sw/rxe/rxe_net.c @@ -555,7 +555,7 @@ struct rxe_dev *rxe_net_add(struct net_device *ndev) } spin_lock_bh(&dev_list_lock); - list_add_tail(&rxe_dev_list, &rxe->list); + list_add_tail(&rxe->list, &rxe_dev_list); spin_unlock_bh(&dev_list_lock); return rxe; } From 2d4b21e0a2913612274a69a3ba1bfee4cffc6e77 Mon Sep 17 00:00:00 2001 From: Yonatan Cohen Date: Thu, 19 Jan 2017 15:25:59 +0200 Subject: [PATCH 228/343] IB/rxe: Prevent from completer to operate on non valid QP On UD QP completer tasklet is scheduled for each packet sent. If it is followed by a destroy_qp(), the kernel panic will happen as the completer tries to operate on a destroyed QP. Fixes: 8700e3e7c485 ("Soft RoCE driver") Signed-off-by: Yonatan Cohen Reviewed-by: Moni Shoua Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- drivers/infiniband/sw/rxe/rxe_qp.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/infiniband/sw/rxe/rxe_qp.c b/drivers/infiniband/sw/rxe/rxe_qp.c index 486d576e55bc..44b2108253bd 100644 --- a/drivers/infiniband/sw/rxe/rxe_qp.c +++ b/drivers/infiniband/sw/rxe/rxe_qp.c @@ -813,8 +813,7 @@ void rxe_qp_destroy(struct rxe_qp *qp) del_timer_sync(&qp->rnr_nak_timer); rxe_cleanup_task(&qp->req.task); - if (qp_type(qp) == IB_QPT_RC) - rxe_cleanup_task(&qp->comp.task); + rxe_cleanup_task(&qp->comp.task); /* flush out any receive wr's or pending requests */ __rxe_do_task(&qp->req.task); From 88ff7334f25909802140e690c0e16433e485b0a0 Mon Sep 17 00:00:00 2001 From: Robert Shearman Date: Tue, 24 Jan 2017 16:26:47 +0000 Subject: [PATCH 229/343] net: Specify the owning module for lwtunnel ops Modules implementing lwtunnel ops should not be allowed to unload while there is state alive using those ops, so specify the owning module for all lwtunnel ops. Signed-off-by: Robert Shearman Signed-off-by: David S. Miller --- include/net/lwtunnel.h | 2 ++ net/core/lwt_bpf.c | 1 + net/ipv4/ip_tunnel_core.c | 2 ++ net/ipv6/ila/ila_lwt.c | 1 + net/ipv6/seg6_iptunnel.c | 1 + net/mpls/mpls_iptunnel.c | 1 + 6 files changed, 8 insertions(+) diff --git a/include/net/lwtunnel.h b/include/net/lwtunnel.h index 0b585f1fd340..73dd87647460 100644 --- a/include/net/lwtunnel.h +++ b/include/net/lwtunnel.h @@ -44,6 +44,8 @@ struct lwtunnel_encap_ops { int (*get_encap_size)(struct lwtunnel_state *lwtstate); int (*cmp_encap)(struct lwtunnel_state *a, struct lwtunnel_state *b); int (*xmit)(struct sk_buff *skb); + + struct module *owner; }; #ifdef CONFIG_LWTUNNEL diff --git a/net/core/lwt_bpf.c b/net/core/lwt_bpf.c index 71bb3e2eca08..b3eef90b2df9 100644 --- a/net/core/lwt_bpf.c +++ b/net/core/lwt_bpf.c @@ -386,6 +386,7 @@ static const struct lwtunnel_encap_ops bpf_encap_ops = { .fill_encap = bpf_fill_encap_info, .get_encap_size = bpf_encap_nlsize, .cmp_encap = bpf_encap_cmp, + .owner = THIS_MODULE, }; static int __init bpf_lwt_init(void) diff --git a/net/ipv4/ip_tunnel_core.c b/net/ipv4/ip_tunnel_core.c index fed3d29f9eb3..0fd1976ab63b 100644 --- a/net/ipv4/ip_tunnel_core.c +++ b/net/ipv4/ip_tunnel_core.c @@ -313,6 +313,7 @@ static const struct lwtunnel_encap_ops ip_tun_lwt_ops = { .fill_encap = ip_tun_fill_encap_info, .get_encap_size = ip_tun_encap_nlsize, .cmp_encap = ip_tun_cmp_encap, + .owner = THIS_MODULE, }; static const struct nla_policy ip6_tun_policy[LWTUNNEL_IP6_MAX + 1] = { @@ -403,6 +404,7 @@ static const struct lwtunnel_encap_ops ip6_tun_lwt_ops = { .fill_encap = ip6_tun_fill_encap_info, .get_encap_size = ip6_tun_encap_nlsize, .cmp_encap = ip_tun_cmp_encap, + .owner = THIS_MODULE, }; void __init ip_tunnel_core_init(void) diff --git a/net/ipv6/ila/ila_lwt.c b/net/ipv6/ila/ila_lwt.c index a7bc54ab46e2..13b5e85fe0d5 100644 --- a/net/ipv6/ila/ila_lwt.c +++ b/net/ipv6/ila/ila_lwt.c @@ -238,6 +238,7 @@ static const struct lwtunnel_encap_ops ila_encap_ops = { .fill_encap = ila_fill_encap_info, .get_encap_size = ila_encap_nlsize, .cmp_encap = ila_encap_cmp, + .owner = THIS_MODULE, }; int ila_lwt_init(void) diff --git a/net/ipv6/seg6_iptunnel.c b/net/ipv6/seg6_iptunnel.c index 1d60cb132835..c46f8cbf5ab5 100644 --- a/net/ipv6/seg6_iptunnel.c +++ b/net/ipv6/seg6_iptunnel.c @@ -422,6 +422,7 @@ static const struct lwtunnel_encap_ops seg6_iptun_ops = { .fill_encap = seg6_fill_encap_info, .get_encap_size = seg6_encap_nlsize, .cmp_encap = seg6_encap_cmp, + .owner = THIS_MODULE, }; int __init seg6_iptunnel_init(void) diff --git a/net/mpls/mpls_iptunnel.c b/net/mpls/mpls_iptunnel.c index 2f7ccd934416..1d281c1ff7c1 100644 --- a/net/mpls/mpls_iptunnel.c +++ b/net/mpls/mpls_iptunnel.c @@ -215,6 +215,7 @@ static const struct lwtunnel_encap_ops mpls_iptun_ops = { .fill_encap = mpls_fill_encap_info, .get_encap_size = mpls_encap_nlsize, .cmp_encap = mpls_encap_cmp, + .owner = THIS_MODULE, }; static int __init mpls_iptunnel_init(void) From 85c814016ce3b371016c2c054a905fa2492f5a65 Mon Sep 17 00:00:00 2001 From: Robert Shearman Date: Tue, 24 Jan 2017 16:26:48 +0000 Subject: [PATCH 230/343] lwtunnel: Fix oops on state free after encap module unload When attempting to free lwtunnel state after the module for the encap has been unloaded an oops occurs: BUG: unable to handle kernel NULL pointer dereference at 0000000000000008 IP: lwtstate_free+0x18/0x40 [..] task: ffff88003e372380 task.stack: ffffc900001fc000 RIP: 0010:lwtstate_free+0x18/0x40 RSP: 0018:ffff88003fd83e88 EFLAGS: 00010246 RAX: 0000000000000000 RBX: ffff88002bbb3380 RCX: ffff88000c91a300 [..] Call Trace: free_fib_info_rcu+0x195/0x1a0 ? rt_fibinfo_free+0x50/0x50 rcu_process_callbacks+0x2d3/0x850 ? rcu_process_callbacks+0x296/0x850 __do_softirq+0xe4/0x4cb irq_exit+0xb0/0xc0 smp_apic_timer_interrupt+0x3d/0x50 apic_timer_interrupt+0x93/0xa0 [..] Code: e8 6e c6 fc ff 89 d8 5b 5d c3 bb de ff ff ff eb f4 66 90 66 66 66 66 90 55 48 89 e5 53 0f b7 07 48 89 fb 48 8b 04 c5 00 81 d5 81 <48> 8b 40 08 48 85 c0 74 13 ff d0 48 8d 7b 20 be 20 00 00 00 e8 The problem is after the module for the encap can be unloaded the corresponding ops is removed and is thus NULL here. Modules implementing lwtunnel ops should not be allowed to unload while there is state alive using those ops, so grab the module reference for the ops on creating lwtunnel state and of course release the reference when freeing the state. Fixes: 1104d9ba443a ("lwtunnel: Add destroy state operation") Signed-off-by: Robert Shearman Signed-off-by: David S. Miller --- net/core/lwtunnel.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/net/core/lwtunnel.c b/net/core/lwtunnel.c index 47b1dd65947b..c23465005f2f 100644 --- a/net/core/lwtunnel.c +++ b/net/core/lwtunnel.c @@ -115,8 +115,11 @@ int lwtunnel_build_state(struct net_device *dev, u16 encap_type, ret = -EOPNOTSUPP; rcu_read_lock(); ops = rcu_dereference(lwtun_encaps[encap_type]); - if (likely(ops && ops->build_state)) + if (likely(ops && ops->build_state && try_module_get(ops->owner))) { ret = ops->build_state(dev, encap, family, cfg, lws); + if (ret) + module_put(ops->owner); + } rcu_read_unlock(); return ret; @@ -194,6 +197,7 @@ void lwtstate_free(struct lwtunnel_state *lws) } else { kfree(lws); } + module_put(ops->owner); } EXPORT_SYMBOL(lwtstate_free); From 78f824d4312a8944f5340c6b161bba3bf2c81096 Mon Sep 17 00:00:00 2001 From: Vineet Gupta Date: Tue, 21 Jun 2016 14:24:33 +0530 Subject: [PATCH 231/343] ARCv2: smp-boot: wake_flag polling by non-Masters needs to be uncached This is needed on HS38 cores, for setting up IO-Coherency aperture properly The polling could perturb the caches and coherecy fabric which could be wrong in the small window when Master is setting up IOC aperture etc in arc_cache_init() We do it only for ARCv2 based builds to not affect EZChip ARCompact based platform. Signed-off-by: Vineet Gupta --- arch/arc/kernel/smp.c | 19 ++++++++++++++++--- 1 file changed, 16 insertions(+), 3 deletions(-) diff --git a/arch/arc/kernel/smp.c b/arch/arc/kernel/smp.c index 44a0d21ed342..2afbafadb6ab 100644 --- a/arch/arc/kernel/smp.c +++ b/arch/arc/kernel/smp.c @@ -90,10 +90,23 @@ void __init smp_cpus_done(unsigned int max_cpus) */ static volatile int wake_flag; +#ifdef CONFIG_ISA_ARCOMPACT + +#define __boot_read(f) f +#define __boot_write(f, v) f = v + +#else + +#define __boot_read(f) arc_read_uncached_32(&f) +#define __boot_write(f, v) arc_write_uncached_32(&f, v) + +#endif + static void arc_default_smp_cpu_kick(int cpu, unsigned long pc) { BUG_ON(cpu == 0); - wake_flag = cpu; + + __boot_write(wake_flag, cpu); } void arc_platform_smp_wait_to_boot(int cpu) @@ -102,10 +115,10 @@ void arc_platform_smp_wait_to_boot(int cpu) if (IS_ENABLED(CONFIG_ARC_SMP_HALT_ON_RESET)) return; - while (wake_flag != cpu) + while (__boot_read(wake_flag) != cpu) ; - wake_flag = 0; + __boot_write(wake_flag, 0); } const char *arc_platform_smp_cpuinfo(void) From 0516ffd88fa0d006ee80389ce14a9ca5ae45e845 Mon Sep 17 00:00:00 2001 From: Stefan Hajnoczi Date: Thu, 19 Jan 2017 10:43:53 +0000 Subject: [PATCH 232/343] vhost/vsock: handle vhost_vq_init_access() error Propagate the error when vhost_vq_init_access() fails and set vq->private_data to NULL. Signed-off-by: Stefan Hajnoczi Signed-off-by: Michael S. Tsirkin --- drivers/vhost/vsock.c | 13 +++++++++---- 1 file changed, 9 insertions(+), 4 deletions(-) diff --git a/drivers/vhost/vsock.c b/drivers/vhost/vsock.c index bbbf588540ed..ce5e63d2c66a 100644 --- a/drivers/vhost/vsock.c +++ b/drivers/vhost/vsock.c @@ -373,6 +373,7 @@ static void vhost_vsock_handle_rx_kick(struct vhost_work *work) static int vhost_vsock_start(struct vhost_vsock *vsock) { + struct vhost_virtqueue *vq; size_t i; int ret; @@ -383,19 +384,20 @@ static int vhost_vsock_start(struct vhost_vsock *vsock) goto err; for (i = 0; i < ARRAY_SIZE(vsock->vqs); i++) { - struct vhost_virtqueue *vq = &vsock->vqs[i]; + vq = &vsock->vqs[i]; mutex_lock(&vq->mutex); if (!vhost_vq_access_ok(vq)) { ret = -EFAULT; - mutex_unlock(&vq->mutex); goto err_vq; } if (!vq->private_data) { vq->private_data = vsock; - vhost_vq_init_access(vq); + ret = vhost_vq_init_access(vq); + if (ret) + goto err_vq; } mutex_unlock(&vq->mutex); @@ -405,8 +407,11 @@ static int vhost_vsock_start(struct vhost_vsock *vsock) return 0; err_vq: + vq->private_data = NULL; + mutex_unlock(&vq->mutex); + for (i = 0; i < ARRAY_SIZE(vsock->vqs); i++) { - struct vhost_virtqueue *vq = &vsock->vqs[i]; + vq = &vsock->vqs[i]; mutex_lock(&vq->mutex); vq->private_data = NULL; From f7f6634d23830ff74335734fbdb28ea109c1f349 Mon Sep 17 00:00:00 2001 From: Robin Murphy Date: Tue, 10 Jan 2017 17:51:17 +0000 Subject: [PATCH 233/343] virtio_mmio: Set DMA masks appropriately Once DMA API usage is enabled, it becomes apparent that virtio-mmio is inadvertently relying on the default 32-bit DMA mask, which leads to problems like rapidly exhausting SWIOTLB bounce buffers. Ensure that we set the appropriate 64-bit DMA mask whenever possible, with the coherent mask suitably limited for the legacy vring as per a0be1db4304f ("virtio_pci: Limit DMA mask to 44 bits for legacy virtio devices"). Cc: Andy Lutomirski Cc: Michael S. Tsirkin Reported-by: Jean-Philippe Brucker Fixes: b42111382f0e ("virtio_mmio: Use the DMA API if enabled") Signed-off-by: Robin Murphy Signed-off-by: Michael S. Tsirkin --- drivers/virtio/virtio_mmio.c | 20 +++++++++++++++++++- 1 file changed, 19 insertions(+), 1 deletion(-) diff --git a/drivers/virtio/virtio_mmio.c b/drivers/virtio/virtio_mmio.c index d47a2fcef818..c71fde5fe835 100644 --- a/drivers/virtio/virtio_mmio.c +++ b/drivers/virtio/virtio_mmio.c @@ -59,6 +59,7 @@ #define pr_fmt(fmt) "virtio-mmio: " fmt #include +#include #include #include #include @@ -498,6 +499,7 @@ static int virtio_mmio_probe(struct platform_device *pdev) struct virtio_mmio_device *vm_dev; struct resource *mem; unsigned long magic; + int rc; mem = platform_get_resource(pdev, IORESOURCE_MEM, 0); if (!mem) @@ -547,9 +549,25 @@ static int virtio_mmio_probe(struct platform_device *pdev) } vm_dev->vdev.id.vendor = readl(vm_dev->base + VIRTIO_MMIO_VENDOR_ID); - if (vm_dev->version == 1) + if (vm_dev->version == 1) { writel(PAGE_SIZE, vm_dev->base + VIRTIO_MMIO_GUEST_PAGE_SIZE); + rc = dma_set_mask(&pdev->dev, DMA_BIT_MASK(64)); + /* + * In the legacy case, ensure our coherently-allocated virtio + * ring will be at an address expressable as a 32-bit PFN. + */ + if (!rc) + dma_set_coherent_mask(&pdev->dev, + DMA_BIT_MASK(32 + PAGE_SHIFT)); + } else { + rc = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(64)); + } + if (rc) + rc = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(32)); + if (rc) + dev_warn(&pdev->dev, "Failed to enable 64-bit or 32-bit DMA. Trying to continue, but this might not work.\n"); + platform_set_drvdata(pdev, vm_dev); return register_virtio_device(&vm_dev->vdev); From c7070619f3408d9a0dffbed9149e6f00479cf43b Mon Sep 17 00:00:00 2001 From: Will Deacon Date: Fri, 20 Jan 2017 10:33:32 +0000 Subject: [PATCH 234/343] vring: Force use of DMA API for ARM-based systems with legacy devices Booting Linux on an ARM fastmodel containing an SMMU emulation results in an unexpected I/O page fault from the legacy virtio-blk PCI device: [ 1.211721] arm-smmu-v3 2b400000.smmu: event 0x10 received: [ 1.211800] arm-smmu-v3 2b400000.smmu: 0x00000000fffff010 [ 1.211880] arm-smmu-v3 2b400000.smmu: 0x0000020800000000 [ 1.211959] arm-smmu-v3 2b400000.smmu: 0x00000008fa081002 [ 1.212075] arm-smmu-v3 2b400000.smmu: 0x0000000000000000 [ 1.212155] arm-smmu-v3 2b400000.smmu: event 0x10 received: [ 1.212234] arm-smmu-v3 2b400000.smmu: 0x00000000fffff010 [ 1.212314] arm-smmu-v3 2b400000.smmu: 0x0000020800000000 [ 1.212394] arm-smmu-v3 2b400000.smmu: 0x00000008fa081000 [ 1.212471] arm-smmu-v3 2b400000.smmu: 0x0000000000000000 This is because the legacy virtio-blk device is behind an SMMU, so we have consequently swizzled its DMA ops and configured the SMMU to translate accesses. This then requires the vring code to use the DMA API to establish translations, otherwise all transactions will result in fatal faults and termination. Given that ARM-based systems only see an SMMU if one is really present (the topology is all described by firmware tables such as device-tree or IORT), then we can safely use the DMA API for all legacy virtio devices. Modern devices can advertise the prescense of an IOMMU using the VIRTIO_F_IOMMU_PLATFORM feature flag. Cc: Andy Lutomirski Cc: Michael S. Tsirkin Cc: Fixes: 876945dbf649 ("arm64: Hook up IOMMU dma_ops") Signed-off-by: Will Deacon Signed-off-by: Michael S. Tsirkin Acked-by: Marc Zyngier --- drivers/virtio/virtio_ring.c | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c index 409aeaa49246..7e38ed79c3fc 100644 --- a/drivers/virtio/virtio_ring.c +++ b/drivers/virtio/virtio_ring.c @@ -159,6 +159,13 @@ static bool vring_use_dma_api(struct virtio_device *vdev) if (xen_domain()) return true; + /* + * On ARM-based machines, the DMA ops will do the right thing, + * so always use them with legacy devices. + */ + if (IS_ENABLED(CONFIG_ARM) || IS_ENABLED(CONFIG_ARM64)) + return !virtio_has_feature(vdev, VIRTIO_F_VERSION_1); + return false; } From 8a1f780e7f28c7c1d640118242cf68d528c456cd Mon Sep 17 00:00:00 2001 From: Yasuaki Ishimatsu Date: Tue, 24 Jan 2017 15:17:45 -0800 Subject: [PATCH 235/343] memory_hotplug: make zone_can_shift() return a boolean value online_{kernel|movable} is used to change the memory zone to ZONE_{NORMAL|MOVABLE} and online the memory. To check that memory zone can be changed, zone_can_shift() is used. Currently the function returns minus integer value, plus integer value and 0. When the function returns minus or plus integer value, it means that the memory zone can be changed to ZONE_{NORNAL|MOVABLE}. But when the function returns 0, there are two meanings. One of the meanings is that the memory zone does not need to be changed. For example, when memory is in ZONE_NORMAL and onlined by online_kernel the memory zone does not need to be changed. Another meaning is that the memory zone cannot be changed. When memory is in ZONE_NORMAL and onlined by online_movable, the memory zone may not be changed to ZONE_MOVALBE due to memory online limitation(see Documentation/memory-hotplug.txt). In this case, memory must not be onlined. The patch changes the return type of zone_can_shift() so that memory online operation fails when memory zone cannot be changed as follows: Before applying patch: # grep -A 35 "Node 2" /proc/zoneinfo Node 2, zone Normal node_scanned 0 spanned 8388608 present 7864320 managed 7864320 # echo online_movable > memory4097/state # grep -A 35 "Node 2" /proc/zoneinfo Node 2, zone Normal node_scanned 0 spanned 8388608 present 8388608 managed 8388608 online_movable operation succeeded. But memory is onlined as ZONE_NORMAL, not ZONE_MOVABLE. After applying patch: # grep -A 35 "Node 2" /proc/zoneinfo Node 2, zone Normal node_scanned 0 spanned 8388608 present 7864320 managed 7864320 # echo online_movable > memory4097/state bash: echo: write error: Invalid argument # grep -A 35 "Node 2" /proc/zoneinfo Node 2, zone Normal node_scanned 0 spanned 8388608 present 7864320 managed 7864320 online_movable operation failed because of failure of changing the memory zone from ZONE_NORMAL to ZONE_MOVABLE Fixes: df429ac03936 ("memory-hotplug: more general validation of zone during online") Link: http://lkml.kernel.org/r/2f9c3837-33d7-b6e5-59c0-6ca4372b2d84@gmail.com Signed-off-by: Yasuaki Ishimatsu Reviewed-by: Reza Arbab Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- drivers/base/memory.c | 4 ++-- include/linux/memory_hotplug.h | 4 ++-- mm/memory_hotplug.c | 28 +++++++++++++++++----------- 3 files changed, 21 insertions(+), 15 deletions(-) diff --git a/drivers/base/memory.c b/drivers/base/memory.c index 8ab8ea1253e6..dacb6a8418aa 100644 --- a/drivers/base/memory.c +++ b/drivers/base/memory.c @@ -408,14 +408,14 @@ static ssize_t show_valid_zones(struct device *dev, sprintf(buf, "%s", zone->name); /* MMOP_ONLINE_KERNEL */ - zone_shift = zone_can_shift(start_pfn, nr_pages, ZONE_NORMAL); + zone_can_shift(start_pfn, nr_pages, ZONE_NORMAL, &zone_shift); if (zone_shift) { strcat(buf, " "); strcat(buf, (zone + zone_shift)->name); } /* MMOP_ONLINE_MOVABLE */ - zone_shift = zone_can_shift(start_pfn, nr_pages, ZONE_MOVABLE); + zone_can_shift(start_pfn, nr_pages, ZONE_MOVABLE, &zone_shift); if (zone_shift) { strcat(buf, " "); strcat(buf, (zone + zone_shift)->name); diff --git a/include/linux/memory_hotplug.h b/include/linux/memory_hotplug.h index 01033fadea47..c1784c0b4f35 100644 --- a/include/linux/memory_hotplug.h +++ b/include/linux/memory_hotplug.h @@ -284,7 +284,7 @@ extern void sparse_remove_one_section(struct zone *zone, struct mem_section *ms, unsigned long map_offset); extern struct page *sparse_decode_mem_map(unsigned long coded_mem_map, unsigned long pnum); -extern int zone_can_shift(unsigned long pfn, unsigned long nr_pages, - enum zone_type target); +extern bool zone_can_shift(unsigned long pfn, unsigned long nr_pages, + enum zone_type target, int *zone_shift); #endif /* __LINUX_MEMORY_HOTPLUG_H */ diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c index e43142c15631..ca2723d47338 100644 --- a/mm/memory_hotplug.c +++ b/mm/memory_hotplug.c @@ -1033,36 +1033,39 @@ static void node_states_set_node(int node, struct memory_notify *arg) node_set_state(node, N_MEMORY); } -int zone_can_shift(unsigned long pfn, unsigned long nr_pages, - enum zone_type target) +bool zone_can_shift(unsigned long pfn, unsigned long nr_pages, + enum zone_type target, int *zone_shift) { struct zone *zone = page_zone(pfn_to_page(pfn)); enum zone_type idx = zone_idx(zone); int i; + *zone_shift = 0; + if (idx < target) { /* pages must be at end of current zone */ if (pfn + nr_pages != zone_end_pfn(zone)) - return 0; + return false; /* no zones in use between current zone and target */ for (i = idx + 1; i < target; i++) if (zone_is_initialized(zone - idx + i)) - return 0; + return false; } if (target < idx) { /* pages must be at beginning of current zone */ if (pfn != zone->zone_start_pfn) - return 0; + return false; /* no zones in use between current zone and target */ for (i = target + 1; i < idx; i++) if (zone_is_initialized(zone - idx + i)) - return 0; + return false; } - return target - idx; + *zone_shift = target - idx; + return true; } /* Must be protected by mem_hotplug_begin() */ @@ -1089,10 +1092,13 @@ int __ref online_pages(unsigned long pfn, unsigned long nr_pages, int online_typ !can_online_high_movable(zone)) return -EINVAL; - if (online_type == MMOP_ONLINE_KERNEL) - zone_shift = zone_can_shift(pfn, nr_pages, ZONE_NORMAL); - else if (online_type == MMOP_ONLINE_MOVABLE) - zone_shift = zone_can_shift(pfn, nr_pages, ZONE_MOVABLE); + if (online_type == MMOP_ONLINE_KERNEL) { + if (!zone_can_shift(pfn, nr_pages, ZONE_NORMAL, &zone_shift)) + return -EINVAL; + } else if (online_type == MMOP_ONLINE_MOVABLE) { + if (!zone_can_shift(pfn, nr_pages, ZONE_MOVABLE, &zone_shift)) + return -EINVAL; + } zone = move_pfn_range(zone_shift, pfn, pfn + nr_pages); if (!zone) From 8310d48b125d19fcd9521d83b8293e63eb1646aa Mon Sep 17 00:00:00 2001 From: Keno Fischer Date: Tue, 24 Jan 2017 15:17:48 -0800 Subject: [PATCH 236/343] mm/huge_memory.c: respect FOLL_FORCE/FOLL_COW for thp In commit 19be0eaffa3a ("mm: remove gup_flags FOLL_WRITE games from __get_user_pages()"), the mm code was changed from unsetting FOLL_WRITE after a COW was resolved to setting the (newly introduced) FOLL_COW instead. Simultaneously, the check in gup.c was updated to still allow writes with FOLL_FORCE set if FOLL_COW had also been set. However, a similar check in huge_memory.c was forgotten. As a result, remote memory writes to ro regions of memory backed by transparent huge pages cause an infinite loop in the kernel (handle_mm_fault sets FOLL_COW and returns 0 causing a retry, but follow_trans_huge_pmd bails out immidiately because `(flags & FOLL_WRITE) && !pmd_write(*pmd)` is true. While in this state the process is stil SIGKILLable, but little else works (e.g. no ptrace attach, no other signals). This is easily reproduced with the following code (assuming thp are set to always): #include #include #include #include #include #include #include #include #include #include #define TEST_SIZE 5 * 1024 * 1024 int main(void) { int status; pid_t child; int fd = open("/proc/self/mem", O_RDWR); void *addr = mmap(NULL, TEST_SIZE, PROT_READ, MAP_ANONYMOUS | MAP_PRIVATE, 0, 0); assert(addr != MAP_FAILED); pid_t parent_pid = getpid(); if ((child = fork()) == 0) { void *addr2 = mmap(NULL, TEST_SIZE, PROT_READ | PROT_WRITE, MAP_ANONYMOUS | MAP_PRIVATE, 0, 0); assert(addr2 != MAP_FAILED); memset(addr2, 'a', TEST_SIZE); pwrite(fd, addr2, TEST_SIZE, (uintptr_t)addr); return 0; } assert(child == waitpid(child, &status, 0)); assert(WIFEXITED(status) && WEXITSTATUS(status) == 0); return 0; } Fix this by updating follow_trans_huge_pmd in huge_memory.c analogously to the update in gup.c in the original commit. The same pattern exists in follow_devmap_pmd. However, we should not be able to reach that check with FOLL_COW set, so add WARN_ONCE to make sure we notice if we ever do. [akpm@linux-foundation.org: coding-style fixes] Link: http://lkml.kernel.org/r/20170106015025.GA38411@juliacomputing.com Signed-off-by: Keno Fischer Acked-by: Kirill A. Shutemov Cc: Greg Thelen Cc: Nicholas Piggin Cc: Willy Tarreau Cc: Oleg Nesterov Cc: Kees Cook Cc: Andy Lutomirski Cc: Michal Hocko Cc: Hugh Dickins Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/huge_memory.c | 18 +++++++++++++++++- 1 file changed, 17 insertions(+), 1 deletion(-) diff --git a/mm/huge_memory.c b/mm/huge_memory.c index 9a6bd6c8d55a..5f3ad65c85de 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -783,6 +783,12 @@ struct page *follow_devmap_pmd(struct vm_area_struct *vma, unsigned long addr, assert_spin_locked(pmd_lockptr(mm, pmd)); + /* + * When we COW a devmap PMD entry, we split it into PTEs, so we should + * not be in this function with `flags & FOLL_COW` set. + */ + WARN_ONCE(flags & FOLL_COW, "mm: In follow_devmap_pmd with FOLL_COW set"); + if (flags & FOLL_WRITE && !pmd_write(*pmd)) return NULL; @@ -1128,6 +1134,16 @@ out_unlock: return ret; } +/* + * FOLL_FORCE can write to even unwritable pmd's, but only + * after we've gone through a COW cycle and they are dirty. + */ +static inline bool can_follow_write_pmd(pmd_t pmd, unsigned int flags) +{ + return pmd_write(pmd) || + ((flags & FOLL_FORCE) && (flags & FOLL_COW) && pmd_dirty(pmd)); +} + struct page *follow_trans_huge_pmd(struct vm_area_struct *vma, unsigned long addr, pmd_t *pmd, @@ -1138,7 +1154,7 @@ struct page *follow_trans_huge_pmd(struct vm_area_struct *vma, assert_spin_locked(pmd_lockptr(mm, pmd)); - if (flags & FOLL_WRITE && !pmd_write(*pmd)) + if (flags & FOLL_WRITE && !can_follow_write_pmd(*pmd, flags)) goto out; /* Avoid dumping huge zero page */ From 6affb9d7b137fc93d86c926a5587e77b8bc64255 Mon Sep 17 00:00:00 2001 From: Ross Zwisler Date: Tue, 24 Jan 2017 15:17:51 -0800 Subject: [PATCH 237/343] dax: fix build warnings with FS_DAX and !FS_IOMAP As reported by Arnd: https://lkml.org/lkml/2017/1/10/756 Compiling with the following configuration: # CONFIG_EXT2_FS is not set # CONFIG_EXT4_FS is not set # CONFIG_XFS_FS is not set # CONFIG_FS_IOMAP depends on the above filesystems, as is not set CONFIG_FS_DAX=y generates build warnings about unused functions in fs/dax.c: fs/dax.c:878:12: warning: `dax_insert_mapping' defined but not used [-Wunused-function] static int dax_insert_mapping(struct address_space *mapping, ^~~~~~~~~~~~~~~~~~ fs/dax.c:572:12: warning: `copy_user_dax' defined but not used [-Wunused-function] static int copy_user_dax(struct block_device *bdev, sector_t sector, size_t size, ^~~~~~~~~~~~~ fs/dax.c:542:12: warning: `dax_load_hole' defined but not used [-Wunused-function] static int dax_load_hole(struct address_space *mapping, void **entry, ^~~~~~~~~~~~~ fs/dax.c:312:14: warning: `grab_mapping_entry' defined but not used [-Wunused-function] static void *grab_mapping_entry(struct address_space *mapping, pgoff_t index, ^~~~~~~~~~~~~~~~~~ Now that the struct buffer_head based DAX fault paths and I/O path have been removed we really depend on iomap support being present for DAX. Make this explicit by selecting FS_IOMAP if we compile in DAX support. This allows us to remove conditional selections of FS_IOMAP when FS_DAX was present for ext2 and ext4, and to remove an #ifdef in fs/dax.c. Link: http://lkml.kernel.org/r/1484087383-29478-1-git-send-email-ross.zwisler@linux.intel.com Signed-off-by: Ross Zwisler Reported-by: Arnd Bergmann Reviewed-by: Christoph Hellwig Reviewed-by: Jan Kara Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/Kconfig | 1 + fs/dax.c | 2 -- fs/ext2/Kconfig | 1 - fs/ext4/Kconfig | 1 - 4 files changed, 1 insertion(+), 4 deletions(-) diff --git a/fs/Kconfig b/fs/Kconfig index c2a377cdda2b..83eab52fb3f6 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -38,6 +38,7 @@ config FS_DAX bool "Direct Access (DAX) support" depends on MMU depends on !(ARM || MIPS || SPARC) + select FS_IOMAP help Direct Access (DAX) can be used on memory-backed block devices. If the block device supports DAX and the filesystem supports DAX, diff --git a/fs/dax.c b/fs/dax.c index ddcddfeaa03b..3af2da5e64ce 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -990,7 +990,6 @@ int __dax_zero_page_range(struct block_device *bdev, sector_t sector, } EXPORT_SYMBOL_GPL(__dax_zero_page_range); -#ifdef CONFIG_FS_IOMAP static sector_t dax_iomap_sector(struct iomap *iomap, loff_t pos) { return iomap->blkno + (((pos & PAGE_MASK) - iomap->offset) >> 9); @@ -1428,4 +1427,3 @@ int dax_iomap_pmd_fault(struct vm_area_struct *vma, unsigned long address, } EXPORT_SYMBOL_GPL(dax_iomap_pmd_fault); #endif /* CONFIG_FS_DAX_PMD */ -#endif /* CONFIG_FS_IOMAP */ diff --git a/fs/ext2/Kconfig b/fs/ext2/Kconfig index 36bea5adcaba..c634874e12d9 100644 --- a/fs/ext2/Kconfig +++ b/fs/ext2/Kconfig @@ -1,6 +1,5 @@ config EXT2_FS tristate "Second extended fs support" - select FS_IOMAP if FS_DAX help Ext2 is a standard Linux file system for hard disks. diff --git a/fs/ext4/Kconfig b/fs/ext4/Kconfig index 7b90691e98c4..e38039fd96ff 100644 --- a/fs/ext4/Kconfig +++ b/fs/ext4/Kconfig @@ -37,7 +37,6 @@ config EXT4_FS select CRC16 select CRYPTO select CRYPTO_CRC32C - select FS_IOMAP if FS_DAX help This is the next generation of the ext3 filesystem. From b94f51183b0617e7b9b4fb4137d4cf1cab7547c2 Mon Sep 17 00:00:00 2001 From: Don Zickus Date: Tue, 24 Jan 2017 15:17:53 -0800 Subject: [PATCH 238/343] kernel/watchdog: prevent false hardlockup on overloaded system On an overloaded system, it is possible that a change in the watchdog threshold can be delayed long enough to trigger a false positive. This can easily be achieved by having a cpu spinning indefinitely on a task, while another cpu updates watchdog threshold. What happens is while trying to park the watchdog threads, the hrtimers on the other cpus trigger and reprogram themselves with the new slower watchdog threshold. Meanwhile, the nmi watchdog is still programmed with the old faster threshold. Because the one cpu is blocked, it prevents the thread parking on the other cpus from completing, which is needed to shutdown the nmi watchdog and reprogram it correctly. As a result, a false positive from the nmi watchdog is reported. Fix this by setting a park_in_progress flag to block all lockups until the parking is complete. Fix provided by Ulrich Obergfell. [akpm@linux-foundation.org: s/park_in_progress/watchdog_park_in_progress/] Link: http://lkml.kernel.org/r/1481041033-192236-1-git-send-email-dzickus@redhat.com Signed-off-by: Don Zickus Reviewed-by: Aaron Tomlin Cc: Ulrich Obergfell Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/nmi.h | 1 + kernel/watchdog.c | 9 +++++++++ kernel/watchdog_hld.c | 3 +++ 3 files changed, 13 insertions(+) diff --git a/include/linux/nmi.h b/include/linux/nmi.h index aacca824a6ae..0a3fadc32693 100644 --- a/include/linux/nmi.h +++ b/include/linux/nmi.h @@ -110,6 +110,7 @@ extern int watchdog_user_enabled; extern int watchdog_thresh; extern unsigned long watchdog_enabled; extern unsigned long *watchdog_cpumask_bits; +extern atomic_t watchdog_park_in_progress; #ifdef CONFIG_SMP extern int sysctl_softlockup_all_cpu_backtrace; extern int sysctl_hardlockup_all_cpu_backtrace; diff --git a/kernel/watchdog.c b/kernel/watchdog.c index d4b0fa01cae3..63177be0159e 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -49,6 +49,8 @@ unsigned long *watchdog_cpumask_bits = cpumask_bits(&watchdog_cpumask); #define for_each_watchdog_cpu(cpu) \ for_each_cpu_and((cpu), cpu_online_mask, &watchdog_cpumask) +atomic_t watchdog_park_in_progress = ATOMIC_INIT(0); + /* * The 'watchdog_running' variable is set to 1 when the watchdog threads * are registered/started and is set to 0 when the watchdog threads are @@ -260,6 +262,9 @@ static enum hrtimer_restart watchdog_timer_fn(struct hrtimer *hrtimer) int duration; int softlockup_all_cpu_backtrace = sysctl_softlockup_all_cpu_backtrace; + if (atomic_read(&watchdog_park_in_progress) != 0) + return HRTIMER_NORESTART; + /* kick the hardlockup detector */ watchdog_interrupt_count(); @@ -467,12 +472,16 @@ static int watchdog_park_threads(void) { int cpu, ret = 0; + atomic_set(&watchdog_park_in_progress, 1); + for_each_watchdog_cpu(cpu) { ret = kthread_park(per_cpu(softlockup_watchdog, cpu)); if (ret) break; } + atomic_set(&watchdog_park_in_progress, 0); + return ret; } diff --git a/kernel/watchdog_hld.c b/kernel/watchdog_hld.c index 84016c8aee6b..12b8dd640786 100644 --- a/kernel/watchdog_hld.c +++ b/kernel/watchdog_hld.c @@ -84,6 +84,9 @@ static void watchdog_overflow_callback(struct perf_event *event, /* Ensure the watchdog never gets throttled */ event->hw.interrupts = 0; + if (atomic_read(&watchdog_park_in_progress) != 0) + return; + if (__this_cpu_read(watchdog_nmi_touch) == true) { __this_cpu_write(watchdog_nmi_touch, false); return; From de182cc8e882f74af2a112e09f148ce646937232 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Tue, 24 Jan 2017 15:17:56 -0800 Subject: [PATCH 239/343] drivers/memstick/core/memstick.c: avoid -Wnonnull warning gcc-7 produces a harmless false-postive warning about a possible NULL pointer access: drivers/memstick/core/memstick.c: In function 'h_memstick_read_dev_id': drivers/memstick/core/memstick.c:309:3: error: argument 2 null where non-null expected [-Werror=nonnull] memcpy(mrq->data, buf, mrq->data_len); This can't happen because the caller sets the command to 'MS_TPC_READ_REG', which causes the data direction to be 'READ' and the NULL pointer not accessed. As a simple workaround for the warning, we can pass a pointer to the data that we actually want to read into. This is not needed here, but also harmless, and lets the compiler know that the access is ok. Link: http://lkml.kernel.org/r/20170111144143.548867-1-arnd@arndb.de Signed-off-by: Arnd Bergmann Cc: Alex Dubov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- drivers/memstick/core/memstick.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/memstick/core/memstick.c b/drivers/memstick/core/memstick.c index a0547dbf9806..76382c858c35 100644 --- a/drivers/memstick/core/memstick.c +++ b/drivers/memstick/core/memstick.c @@ -330,7 +330,7 @@ static int h_memstick_read_dev_id(struct memstick_dev *card, struct ms_id_register id_reg; if (!(*mrq)) { - memstick_init_req(&card->current_mrq, MS_TPC_READ_REG, NULL, + memstick_init_req(&card->current_mrq, MS_TPC_READ_REG, &id_reg, sizeof(struct ms_id_register)); *mrq = &card->current_mrq; return 0; From 15a77c6fe494f4b1757d30cd137fe66ab06a38c3 Mon Sep 17 00:00:00 2001 From: Andrea Arcangeli Date: Tue, 24 Jan 2017 15:17:59 -0800 Subject: [PATCH 240/343] userfaultfd: fix SIGBUS resulting from false rwsem wakeups With >=32 CPUs the userfaultfd selftest triggered a graceful but unexpected SIGBUS because VM_FAULT_RETRY was returned by handle_userfault() despite the UFFDIO_COPY wasn't completed. This seems caused by rwsem waking the thread blocked in handle_userfault() and we can't run up_read() before the wait_event sequence is complete. Keeping the wait_even sequence identical to the first one, would require running userfaultfd_must_wait() again to know if the loop should be repeated, and it would also require retaking the rwsem and revalidating the whole vma status. It seems simpler to wait the targeted wakeup so that if false wakeups materialize we still wait for our specific wakeup event, unless of course there are signals or the uffd was released. Debug code collecting the stack trace of the wakeup showed this: $ ./userfaultfd 100 99999 nr_pages: 25600, nr_pages_per_cpu: 800 bounces: 99998, mode: racing ver poll, userfaults: 32 35 90 232 30 138 69 82 34 30 139 40 40 31 20 19 43 13 15 28 27 38 21 43 56 22 1 17 31 8 4 2 bounces: 99997, mode: rnd ver poll, Bus error (core dumped) save_stack_trace+0x2b/0x50 try_to_wake_up+0x2a6/0x580 wake_up_q+0x32/0x70 rwsem_wake+0xe0/0x120 call_rwsem_wake+0x1b/0x30 up_write+0x3b/0x40 vm_mmap_pgoff+0x9c/0xc0 SyS_mmap_pgoff+0x1a9/0x240 SyS_mmap+0x22/0x30 entry_SYSCALL_64_fastpath+0x1f/0xbd 0xffffffffffffffff FAULT_FLAG_ALLOW_RETRY missing 70 CPU: 24 PID: 1054 Comm: userfaultfd Tainted: G W 4.8.0+ #30 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.9.3-0-ge2fc41e-prebuilt.qemu-project.org 04/01/2014 Call Trace: dump_stack+0xb8/0x112 handle_userfault+0x572/0x650 handle_mm_fault+0x12cb/0x1520 __do_page_fault+0x175/0x500 trace_do_page_fault+0x61/0x270 do_async_page_fault+0x19/0x90 async_page_fault+0x25/0x30 This always happens when the main userfault selftest thread is running clone() while glibc runs either mprotect or mmap (both taking mmap_sem down_write()) to allocate the thread stack of the background threads, while locking/userfault threads already run at full throttle and are susceptible to false wakeups that may cause handle_userfault() to return before than expected (which results in graceful SIGBUS at the next attempt). This was reproduced only with >=32 CPUs because the loop to start the thread where clone() is too quick with fewer CPUs, while with 32 CPUs there's already significant activity on ~32 locking and userfault threads when the last background threads are started with clone(). This >=32 CPUs SMP race condition is likely reproducible only with the selftest because of the much heavier userfault load it generates if compared to real apps. We'll have to allow "one more" VM_FAULT_RETRY for the WP support and a patch floating around that provides it also hidden this problem but in reality only is successfully at hiding the problem. False wakeups could still happen again the second time handle_userfault() is invoked, even if it's a so rare race condition that getting false wakeups twice in a row is impossible to reproduce. This full fix is needed for correctness, the only alternative would be to allow VM_FAULT_RETRY to be returned infinitely. With this fix the WP support can stick to a strict "one more" VM_FAULT_RETRY logic (no need of returning it infinite times to avoid the SIGBUS). Link: http://lkml.kernel.org/r/20170111005535.13832-2-aarcange@redhat.com Signed-off-by: Andrea Arcangeli Reported-by: Shubham Kumar Sharma Tested-by: Mike Kravetz Acked-by: Hillf Danton Cc: Michael Rapoport Cc: "Dr. David Alan Gilbert" Cc: Pavel Emelyanov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/userfaultfd.c | 37 +++++++++++++++++++++++++++++++++++-- 1 file changed, 35 insertions(+), 2 deletions(-) diff --git a/fs/userfaultfd.c b/fs/userfaultfd.c index d96e2f30084b..43953e03c356 100644 --- a/fs/userfaultfd.c +++ b/fs/userfaultfd.c @@ -63,6 +63,7 @@ struct userfaultfd_wait_queue { struct uffd_msg msg; wait_queue_t wq; struct userfaultfd_ctx *ctx; + bool waken; }; struct userfaultfd_wake_range { @@ -86,6 +87,12 @@ static int userfaultfd_wake_function(wait_queue_t *wq, unsigned mode, if (len && (start > uwq->msg.arg.pagefault.address || start + len <= uwq->msg.arg.pagefault.address)) goto out; + WRITE_ONCE(uwq->waken, true); + /* + * The implicit smp_mb__before_spinlock in try_to_wake_up() + * renders uwq->waken visible to other CPUs before the task is + * waken. + */ ret = wake_up_state(wq->private, mode); if (ret) /* @@ -264,6 +271,7 @@ int handle_userfault(struct vm_fault *vmf, unsigned long reason) struct userfaultfd_wait_queue uwq; int ret; bool must_wait, return_to_userland; + long blocking_state; BUG_ON(!rwsem_is_locked(&mm->mmap_sem)); @@ -334,10 +342,13 @@ int handle_userfault(struct vm_fault *vmf, unsigned long reason) uwq.wq.private = current; uwq.msg = userfault_msg(vmf->address, vmf->flags, reason); uwq.ctx = ctx; + uwq.waken = false; return_to_userland = (vmf->flags & (FAULT_FLAG_USER|FAULT_FLAG_KILLABLE)) == (FAULT_FLAG_USER|FAULT_FLAG_KILLABLE); + blocking_state = return_to_userland ? TASK_INTERRUPTIBLE : + TASK_KILLABLE; spin_lock(&ctx->fault_pending_wqh.lock); /* @@ -350,8 +361,7 @@ int handle_userfault(struct vm_fault *vmf, unsigned long reason) * following the spin_unlock to happen before the list_add in * __add_wait_queue. */ - set_current_state(return_to_userland ? TASK_INTERRUPTIBLE : - TASK_KILLABLE); + set_current_state(blocking_state); spin_unlock(&ctx->fault_pending_wqh.lock); must_wait = userfaultfd_must_wait(ctx, vmf->address, vmf->flags, @@ -364,6 +374,29 @@ int handle_userfault(struct vm_fault *vmf, unsigned long reason) wake_up_poll(&ctx->fd_wqh, POLLIN); schedule(); ret |= VM_FAULT_MAJOR; + + /* + * False wakeups can orginate even from rwsem before + * up_read() however userfaults will wait either for a + * targeted wakeup on the specific uwq waitqueue from + * wake_userfault() or for signals or for uffd + * release. + */ + while (!READ_ONCE(uwq.waken)) { + /* + * This needs the full smp_store_mb() + * guarantee as the state write must be + * visible to other CPUs before reading + * uwq.waken from other CPUs. + */ + set_current_state(blocking_state); + if (READ_ONCE(uwq.waken) || + READ_ONCE(ctx->released) || + (return_to_userland ? signal_pending(current) : + fatal_signal_pending(current))) + break; + schedule(); + } } __set_current_state(TASK_RUNNING); From aa2efd5ea4041754da4046c3d2e7edaac9526258 Mon Sep 17 00:00:00 2001 From: Daniel Thompson Date: Tue, 24 Jan 2017 15:18:02 -0800 Subject: [PATCH 241/343] mm/slub.c: trace free objects at KERN_INFO Currently when trace is enabled (e.g. slub_debug=T,kmalloc-128 ) the trace messages are mostly output at KERN_INFO. However the trace code also calls print_section() to hexdump the head of a free object. This is hard coded to use KERN_ERR, meaning the console is deluged with trace messages even if we've asked for quiet. Fix this the obvious way but adding a level parameter to print_section(), allowing calls from the trace code to use the same trace level as other trace messages. Link: http://lkml.kernel.org/r/20170113154850.518-1-daniel.thompson@linaro.org Signed-off-by: Daniel Thompson Acked-by: Christoph Lameter Acked-by: David Rientjes Cc: Pekka Enberg Cc: Joonsoo Kim Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/slub.c | 23 +++++++++++++---------- 1 file changed, 13 insertions(+), 10 deletions(-) diff --git a/mm/slub.c b/mm/slub.c index 067598a00849..7aa6f433f4de 100644 --- a/mm/slub.c +++ b/mm/slub.c @@ -496,10 +496,11 @@ static inline int check_valid_pointer(struct kmem_cache *s, return 1; } -static void print_section(char *text, u8 *addr, unsigned int length) +static void print_section(char *level, char *text, u8 *addr, + unsigned int length) { metadata_access_enable(); - print_hex_dump(KERN_ERR, text, DUMP_PREFIX_ADDRESS, 16, 1, addr, + print_hex_dump(level, text, DUMP_PREFIX_ADDRESS, 16, 1, addr, length, 1); metadata_access_disable(); } @@ -636,14 +637,15 @@ static void print_trailer(struct kmem_cache *s, struct page *page, u8 *p) p, p - addr, get_freepointer(s, p)); if (s->flags & SLAB_RED_ZONE) - print_section("Redzone ", p - s->red_left_pad, s->red_left_pad); + print_section(KERN_ERR, "Redzone ", p - s->red_left_pad, + s->red_left_pad); else if (p > addr + 16) - print_section("Bytes b4 ", p - 16, 16); + print_section(KERN_ERR, "Bytes b4 ", p - 16, 16); - print_section("Object ", p, min_t(unsigned long, s->object_size, - PAGE_SIZE)); + print_section(KERN_ERR, "Object ", p, + min_t(unsigned long, s->object_size, PAGE_SIZE)); if (s->flags & SLAB_RED_ZONE) - print_section("Redzone ", p + s->object_size, + print_section(KERN_ERR, "Redzone ", p + s->object_size, s->inuse - s->object_size); if (s->offset) @@ -658,7 +660,8 @@ static void print_trailer(struct kmem_cache *s, struct page *page, u8 *p) if (off != size_from_object(s)) /* Beginning of the filler is the free pointer */ - print_section("Padding ", p + off, size_from_object(s) - off); + print_section(KERN_ERR, "Padding ", p + off, + size_from_object(s) - off); dump_stack(); } @@ -820,7 +823,7 @@ static int slab_pad_check(struct kmem_cache *s, struct page *page) end--; slab_err(s, page, "Padding overwritten. 0x%p-0x%p", fault, end - 1); - print_section("Padding ", end - remainder, remainder); + print_section(KERN_ERR, "Padding ", end - remainder, remainder); restore_bytes(s, "slab padding", POISON_INUSE, end - remainder, end); return 0; @@ -973,7 +976,7 @@ static void trace(struct kmem_cache *s, struct page *page, void *object, page->freelist); if (!alloc) - print_section("Object ", (void *)object, + print_section(KERN_INFO, "Object ", (void *)object, s->object_size); dump_stack(); From 424f6c4818bbf1b8ccf58aa012ecc19c0bb9b446 Mon Sep 17 00:00:00 2001 From: Lucas Stach Date: Tue, 24 Jan 2017 15:18:05 -0800 Subject: [PATCH 242/343] mm: alloc_contig: re-allow CMA to compact FS pages Commit 73e64c51afc5 ("mm, compaction: allow compaction for GFP_NOFS requests") changed compation to skip FS pages if not explicitly allowed to touch them, but missed to update the CMA compact_control. This leads to a very high isolation failure rate, crippling performance of CMA even on a lightly loaded system. Re-allow CMA to compact FS pages by setting the correct GFP flags, restoring CMA behavior and performance to the kernel 4.9 level. Fixes: 73e64c51afc5 (mm, compaction: allow compaction for GFP_NOFS requests) Link: http://lkml.kernel.org/r/20170113115155.24335-1-l.stach@pengutronix.de Signed-off-by: Lucas Stach Acked-by: Michal Hocko Acked-by: Vlastimil Babka Cc: Joonsoo Kim Cc: Mel Gorman Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 1 + 1 file changed, 1 insertion(+) diff --git a/mm/page_alloc.c b/mm/page_alloc.c index d604d2596b7b..41d5e2e2d8ce 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -7248,6 +7248,7 @@ int alloc_contig_range(unsigned long start, unsigned long end, .zone = page_zone(pfn_to_page(start)), .mode = MIGRATE_SYNC, .ignore_skip_hint = true, + .gfp_mask = GFP_KERNEL, }; INIT_LIST_HEAD(&cc.migratepages); From 3ba4bceef23206349d4130ddf140819b365de7c8 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Tue, 24 Jan 2017 15:18:07 -0800 Subject: [PATCH 243/343] proc: add a schedule point in proc_pid_readdir() We have seen proc_pid_readdir() invocations holding cpu for more than 50 ms. Add a cond_resched() to be gentle with other tasks. [akpm@linux-foundation.org: coding style fix] Link: http://lkml.kernel.org/r/1484238380.15816.42.camel@edumazet-glaptop3.roam.corp.google.com Signed-off-by: Eric Dumazet Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/base.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/fs/proc/base.c b/fs/proc/base.c index 8e7e61b28f31..87c9a9aacda3 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -3179,6 +3179,8 @@ int proc_pid_readdir(struct file *file, struct dir_context *ctx) iter.tgid += 1, iter = next_tgid(ns, iter)) { char name[PROC_NUMBUF]; int len; + + cond_resched(); if (!has_pid_permissions(ns, iter.task, 2)) continue; From 3674534b775354516e5c148ea48f51d4d1909a78 Mon Sep 17 00:00:00 2001 From: David Rientjes Date: Tue, 24 Jan 2017 15:18:10 -0800 Subject: [PATCH 244/343] mm, memcg: do not retry precharge charges When memory.move_charge_at_immigrate is enabled and precharges are depleted during move, mem_cgroup_move_charge_pte_range() will attempt to increase the size of the precharge. Prevent precharges from ever looping by setting __GFP_NORETRY. This was probably the intention of the GFP_KERNEL & ~__GFP_NORETRY, which is pointless as written. Fixes: 0029e19ebf84 ("mm: memcontrol: remove explicit OOM parameter in charge path") Link: http://lkml.kernel.org/r/alpine.DEB.2.10.1701130208510.69402@chino.kir.corp.google.com Signed-off-by: David Rientjes Acked-by: Michal Hocko Cc: Johannes Weiner Cc: Vladimir Davydov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/memcontrol.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/mm/memcontrol.c b/mm/memcontrol.c index a63a8f832664..b822e158b319 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -4353,9 +4353,9 @@ static int mem_cgroup_do_precharge(unsigned long count) return ret; } - /* Try charges one by one with reclaim */ + /* Try charges one by one with reclaim, but do not retry */ while (count--) { - ret = try_charge(mc.to, GFP_KERNEL & ~__GFP_NORETRY, 1); + ret = try_charge(mc.to, GFP_KERNEL | __GFP_NORETRY, 1); if (ret) return ret; mc.precharge++; From bbd88e1d53a84df9f57a2e37acc15518c3d304db Mon Sep 17 00:00:00 2001 From: Fabian Frederick Date: Tue, 24 Jan 2017 15:18:13 -0800 Subject: [PATCH 245/343] Documentation/filesystems/proc.txt: add VmPin Commit bc3e53f682d9 ("mm: distinguish between mlocked and pinned pages") added VmPin in /proc//status. Report that in Documentation/filesystems/proc.txt Also move Umask after Name to keep correct order. Link: http://lkml.kernel.org/r/20170114201219.30387-1-fabf@skynet.be Signed-off-by: Fabian Frederick Cc: Christoph Lameter Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- Documentation/filesystems/proc.txt | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/Documentation/filesystems/proc.txt b/Documentation/filesystems/proc.txt index 72624a16b792..c94b4675d021 100644 --- a/Documentation/filesystems/proc.txt +++ b/Documentation/filesystems/proc.txt @@ -212,10 +212,11 @@ asynchronous manner and the value may not be very precise. To see a precise snapshot of a moment, you can see /proc//smaps file and scan page table. It's slow but very precise. -Table 1-2: Contents of the status files (as of 4.1) +Table 1-2: Contents of the status files (as of 4.8) .............................................................................. Field Content Name filename of the executable + Umask file mode creation mask State state (R is running, S is sleeping, D is sleeping in an uninterruptible wait, Z is zombie, T is traced or stopped) @@ -226,7 +227,6 @@ Table 1-2: Contents of the status files (as of 4.1) TracerPid PID of process tracing this process (0 if not) Uid Real, effective, saved set, and file system UIDs Gid Real, effective, saved set, and file system GIDs - Umask file mode creation mask FDSize number of file descriptor slots currently allocated Groups supplementary group list NStgid descendant namespace thread group ID hierarchy @@ -236,6 +236,7 @@ Table 1-2: Contents of the status files (as of 4.1) VmPeak peak virtual memory size VmSize total program size VmLck locked memory size + VmPin pinned memory size VmHWM peak resident set size ("high water mark") VmRSS size of memory portions. It contains the three following parts (VmRSS = RssAnon + RssFile + RssShmem) From dd040b6f6d5630202e185399a2ff7ab356ed469c Mon Sep 17 00:00:00 2001 From: Matthew Wilcox Date: Tue, 24 Jan 2017 15:18:16 -0800 Subject: [PATCH 246/343] radix-tree: fix private list warnings The newly introduced warning in radix_tree_free_nodes() was testing the wrong variable; it should have been 'old' instead of 'node'. Fixes: ea07b862ac8e ("mm: workingset: fix use-after-free in shadow node shrinker") Link: http://lkml.kernel.org/r/20170118163746.GA32495@cmpxchg.org Signed-off-by: Matthew Wilcox Signed-off-by: Johannes Weiner Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- lib/radix-tree.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/lib/radix-tree.c b/lib/radix-tree.c index 0b92d605fb69..84812a9fb16f 100644 --- a/lib/radix-tree.c +++ b/lib/radix-tree.c @@ -769,7 +769,7 @@ static void radix_tree_free_nodes(struct radix_tree_node *node) struct radix_tree_node *old = child; offset = child->offset + 1; child = child->parent; - WARN_ON_ONCE(!list_empty(&node->private_list)); + WARN_ON_ONCE(!list_empty(&old->private_list)); radix_tree_node_free(old); if (old == entry_to_node(node)) return; From d51e9894d27492783fc6d1b489070b4ba66ce969 Mon Sep 17 00:00:00 2001 From: Vlastimil Babka Date: Tue, 24 Jan 2017 15:18:18 -0800 Subject: [PATCH 247/343] mm/mempolicy.c: do not put mempolicy before using its nodemask Since commit be97a41b291e ("mm/mempolicy.c: merge alloc_hugepage_vma to alloc_pages_vma") alloc_pages_vma() can potentially free a mempolicy by mpol_cond_put() before accessing the embedded nodemask by __alloc_pages_nodemask(). The commit log says it's so "we can use a single exit path within the function" but that's clearly wrong. We can still do that when doing mpol_cond_put() after the allocation attempt. Make sure the mempolicy is not freed prematurely, otherwise __alloc_pages_nodemask() can end up using a bogus nodemask, which could lead e.g. to premature OOM. Fixes: be97a41b291e ("mm/mempolicy.c: merge alloc_hugepage_vma to alloc_pages_vma") Link: http://lkml.kernel.org/r/20170118141124.8345-1-vbabka@suse.cz Signed-off-by: Vlastimil Babka Acked-by: Kirill A. Shutemov Acked-by: Michal Hocko Acked-by: David Rientjes Cc: Aneesh Kumar K.V Cc: Andrea Arcangeli Cc: [4.0+] Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/mempolicy.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/mm/mempolicy.c b/mm/mempolicy.c index 2e346645eb80..1e7873e40c9a 100644 --- a/mm/mempolicy.c +++ b/mm/mempolicy.c @@ -2017,8 +2017,8 @@ retry_cpuset: nmask = policy_nodemask(gfp, pol); zl = policy_zonelist(gfp, pol, node); - mpol_cond_put(pol); page = __alloc_pages_nodemask(gfp, order, zl, nmask); + mpol_cond_put(pol); out: if (unlikely(!page && read_mems_allowed_retry(cpuset_mems_cookie))) goto retry_cpuset; From 545d58f677b21401f6de1ac12c25cc109f903ace Mon Sep 17 00:00:00 2001 From: Sudip Mukherjee Date: Tue, 24 Jan 2017 15:18:21 -0800 Subject: [PATCH 248/343] frv: add atomic64_add_unless() The build of frv allmodconfig was failing with the error: lib/atomic64_test.c:209:9: error: implicit declaration of function 'atomic64_add_unless' All the atomic64 operations were defined in frv, but atomic64_add_unless() was not done. Implement atomic64_add_unless() as done in other arches. Link: http://lkml.kernel.org/r/1484781236-6698-1-git-send-email-sudipm.mukherjee@gmail.com Signed-off-by: Sudip Mukherjee Cc: David Howells Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/frv/include/asm/atomic.h | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) diff --git a/arch/frv/include/asm/atomic.h b/arch/frv/include/asm/atomic.h index 1c2a5e264fc7..994ed3d5ca08 100644 --- a/arch/frv/include/asm/atomic.h +++ b/arch/frv/include/asm/atomic.h @@ -161,6 +161,22 @@ static __inline__ int __atomic_add_unless(atomic_t *v, int a, int u) return c; } +static inline int atomic64_add_unless(atomic64_t *v, long long i, long long u) +{ + long long c, old; + + c = atomic64_read(v); + for (;;) { + if (unlikely(c == u)) + break; + old = atomic64_cmpxchg(v, c, c + i); + if (likely(old == c)) + break; + c = old; + } + return c != u; +} + #define ATOMIC_OP(op) \ static inline int atomic_fetch_##op(int i, atomic_t *v) \ { \ From 2dc705a9930b4806250fbf5a76e55266e59389f2 Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Tue, 24 Jan 2017 15:18:24 -0800 Subject: [PATCH 249/343] fbdev: color map copying bounds checking Copying color maps to userspace doesn't check the value of to->start, which will cause kernel heap buffer OOB read due to signedness wraps. CVE-2016-8405 Link: http://lkml.kernel.org/r/20170105224249.GA50925@beast Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Signed-off-by: Kees Cook Reported-by: Peter Pi (@heisecode) of Trend Micro Cc: Min Chong Cc: Dan Carpenter Cc: Tomi Valkeinen Cc: Bartlomiej Zolnierkiewicz Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- drivers/video/fbdev/core/fbcmap.c | 26 ++++++++++++++------------ 1 file changed, 14 insertions(+), 12 deletions(-) diff --git a/drivers/video/fbdev/core/fbcmap.c b/drivers/video/fbdev/core/fbcmap.c index f89245b8ba8e..68a113594808 100644 --- a/drivers/video/fbdev/core/fbcmap.c +++ b/drivers/video/fbdev/core/fbcmap.c @@ -163,17 +163,18 @@ void fb_dealloc_cmap(struct fb_cmap *cmap) int fb_copy_cmap(const struct fb_cmap *from, struct fb_cmap *to) { - int tooff = 0, fromoff = 0; - int size; + unsigned int tooff = 0, fromoff = 0; + size_t size; if (to->start > from->start) fromoff = to->start - from->start; else tooff = from->start - to->start; - size = to->len - tooff; - if (size > (int) (from->len - fromoff)) - size = from->len - fromoff; - if (size <= 0) + if (fromoff >= from->len || tooff >= to->len) + return -EINVAL; + + size = min_t(size_t, to->len - tooff, from->len - fromoff); + if (size == 0) return -EINVAL; size *= sizeof(u16); @@ -187,17 +188,18 @@ int fb_copy_cmap(const struct fb_cmap *from, struct fb_cmap *to) int fb_cmap_to_user(const struct fb_cmap *from, struct fb_cmap_user *to) { - int tooff = 0, fromoff = 0; - int size; + unsigned int tooff = 0, fromoff = 0; + size_t size; if (to->start > from->start) fromoff = to->start - from->start; else tooff = from->start - to->start; - size = to->len - tooff; - if (size > (int) (from->len - fromoff)) - size = from->len - fromoff; - if (size <= 0) + if (fromoff >= from->len || tooff >= to->len) + return -EINVAL; + + size = min_t(size_t, to->len - tooff, from->len - fromoff); + if (size == 0) return -EINVAL; size *= sizeof(u16); From ff7a28a074ccbea999dadbb58c46212cf90984c6 Mon Sep 17 00:00:00 2001 From: Jiri Slaby Date: Tue, 24 Jan 2017 15:18:29 -0800 Subject: [PATCH 250/343] kernel/panic.c: add missing \n When a system panics, the "Rebooting in X seconds.." message is never printed because it lacks a new line. Fix it. Link: http://lkml.kernel.org/r/20170119114751.2724-1-jslaby@suse.cz Signed-off-by: Jiri Slaby Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/panic.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/panic.c b/kernel/panic.c index 901c4fb46002..08aa88dde7de 100644 --- a/kernel/panic.c +++ b/kernel/panic.c @@ -249,7 +249,7 @@ void panic(const char *fmt, ...) * Delay timeout seconds before rebooting the machine. * We can't use the "normal" timers since we just panicked. */ - pr_emerg("Rebooting in %d seconds..", panic_timeout); + pr_emerg("Rebooting in %d seconds..\n", panic_timeout); for (i = 0; i < panic_timeout * 1000; i += PANIC_TIMER_STEP) { touch_nmi_watchdog(); From ea57485af8f4221312a5a95d63c382b45e7840dc Mon Sep 17 00:00:00 2001 From: Vlastimil Babka Date: Tue, 24 Jan 2017 15:18:32 -0800 Subject: [PATCH 251/343] mm, page_alloc: fix check for NULL preferred_zone Patch series "fix premature OOM regression in 4.7+ due to cpuset races". This is v2 of my attempt to fix the recent report based on LTP cpuset stress test [1]. The intention is to go to stable 4.9 LTSS with this, as triggering repeated OOMs is not nice. That's why the patches try to be not too intrusive. Unfortunately why investigating I found that modifying the testcase to use per-VMA policies instead of per-task policies will bring the OOM's back, but that seems to be much older and harder to fix problem. I have posted a RFC [2] but I believe that fixing the recent regressions has a higher priority. Longer-term we might try to think how to fix the cpuset mess in a better and less error prone way. I was for example very surprised to learn, that cpuset updates change not only task->mems_allowed, but also nodemask of mempolicies. Until now I expected the parameter to alloc_pages_nodemask() to be stable. I wonder why do we then treat cpusets specially in get_page_from_freelist() and distinguish HARDWALL etc, when there's unconditional intersection between mempolicy and cpuset. I would expect the nodemask adjustment for saving overhead in g_p_f(), but that clearly doesn't happen in the current form. So we have both crazy complexity and overhead, AFAICS. [1] https://lkml.kernel.org/r/CAFpQJXUq-JuEP=QPidy4p_=FN0rkH5Z-kfB4qBvsf6jMS87Edg@mail.gmail.com [2] https://lkml.kernel.org/r/7c459f26-13a6-a817-e508-b65b903a8378@suse.cz This patch (of 4): Since commit c33d6c06f60f ("mm, page_alloc: avoid looking up the first zone in a zonelist twice") we have a wrong check for NULL preferred_zone, which can theoretically happen due to concurrent cpuset modification. We check the zoneref pointer which is never NULL and we should check the zone pointer. Also document this in first_zones_zonelist() comment per Michal Hocko. Fixes: c33d6c06f60f ("mm, page_alloc: avoid looking up the first zone in a zonelist twice") Link: http://lkml.kernel.org/r/20170120103843.24587-2-vbabka@suse.cz Signed-off-by: Vlastimil Babka Acked-by: Mel Gorman Acked-by: Hillf Danton Cc: Ganapatrao Kulkarni Cc: Michal Hocko Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/mmzone.h | 6 +++++- mm/page_alloc.c | 2 +- 2 files changed, 6 insertions(+), 2 deletions(-) diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index 36d9896fbc1e..f4aac87adcc3 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h @@ -972,12 +972,16 @@ static __always_inline struct zoneref *next_zones_zonelist(struct zoneref *z, * @zonelist - The zonelist to search for a suitable zone * @highest_zoneidx - The zone index of the highest zone to return * @nodes - An optional nodemask to filter the zonelist with - * @zone - The first suitable zone found is returned via this parameter + * @return - Zoneref pointer for the first suitable zone found (see below) * * This function returns the first zone at or below a given zone index that is * within the allowed nodemask. The zoneref returned is a cursor that can be * used to iterate the zonelist with next_zones_zonelist by advancing it by * one before calling. + * + * When no eligible zone is found, zoneref->zone is NULL (zoneref itself is + * never NULL). This may happen either genuinely, or due to concurrent nodemask + * update due to cpuset modification. */ static inline struct zoneref *first_zones_zonelist(struct zonelist *zonelist, enum zone_type highest_zoneidx, diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 41d5e2e2d8ce..85cf0f715eb0 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -3784,7 +3784,7 @@ retry_cpuset: */ ac.preferred_zoneref = first_zones_zonelist(ac.zonelist, ac.high_zoneidx, ac.nodemask); - if (!ac.preferred_zoneref) { + if (!ac.preferred_zoneref->zone) { page = NULL; goto no_zone; } From 16096c25bf0ca5d87e4fa6ec6108ba53feead212 Mon Sep 17 00:00:00 2001 From: Vlastimil Babka Date: Tue, 24 Jan 2017 15:18:35 -0800 Subject: [PATCH 252/343] mm, page_alloc: fix fast-path race with cpuset update or removal Ganapatrao Kulkarni reported that the LTP test cpuset01 in stress mode triggers OOM killer in few seconds, despite lots of free memory. The test attempts to repeatedly fault in memory in one process in a cpuset, while changing allowed nodes of the cpuset between 0 and 1 in another process. One possible cause is that in the fast path we find the preferred zoneref according to current mems_allowed, so that it points to the middle of the zonelist, skipping e.g. zones of node 1 completely. If the mems_allowed is updated to contain only node 1, we never reach it in the zonelist, and trigger OOM before checking the cpuset_mems_cookie. This patch fixes the particular case by redoing the preferred zoneref search if we switch back to the original nodemask. The condition is also slightly changed so that when the last non-root cpuset is removed, we don't miss it. Note that this is not a full fix, and more patches will follow. Link: http://lkml.kernel.org/r/20170120103843.24587-3-vbabka@suse.cz Fixes: 682a3385e773 ("mm, page_alloc: inline the fast path of the zonelist iterator") Signed-off-by: Vlastimil Babka Reported-by: Ganapatrao Kulkarni Acked-by: Michal Hocko Acked-by: Mel Gorman Acked-by: Hillf Danton Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 85cf0f715eb0..6f28b7e926d1 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -3804,9 +3804,17 @@ retry_cpuset: /* * Restore the original nodemask if it was potentially replaced with * &cpuset_current_mems_allowed to optimize the fast-path attempt. + * Also recalculate the starting point for the zonelist iterator or + * we could end up iterating over non-eligible zones endlessly. */ - if (cpusets_enabled()) + if (unlikely(ac.nodemask != nodemask)) { ac.nodemask = nodemask; + ac.preferred_zoneref = first_zones_zonelist(ac.zonelist, + ac.high_zoneidx, ac.nodemask); + if (!ac.preferred_zoneref->zone) + goto no_zone; + } + page = __alloc_pages_slowpath(alloc_mask, order, &ac); no_zone: From 5ce9bfef1d27944c119a397a9d827bef795487ce Mon Sep 17 00:00:00 2001 From: Vlastimil Babka Date: Tue, 24 Jan 2017 15:18:38 -0800 Subject: [PATCH 253/343] mm, page_alloc: move cpuset seqcount checking to slowpath This is a preparation for the following patch to make review simpler. While the primary motivation is a bug fix, this also simplifies the fast path, although the moved code is only enabled when cpusets are in use. Link: http://lkml.kernel.org/r/20170120103843.24587-4-vbabka@suse.cz Signed-off-by: Vlastimil Babka Acked-by: Mel Gorman Acked-by: Hillf Danton Cc: Ganapatrao Kulkarni Cc: Michal Hocko Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 47 ++++++++++++++++++++++++++--------------------- 1 file changed, 26 insertions(+), 21 deletions(-) diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 6f28b7e926d1..0df3c089d3af 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -3523,12 +3523,13 @@ __alloc_pages_slowpath(gfp_t gfp_mask, unsigned int order, struct page *page = NULL; unsigned int alloc_flags; unsigned long did_some_progress; - enum compact_priority compact_priority = DEF_COMPACT_PRIORITY; + enum compact_priority compact_priority; enum compact_result compact_result; - int compaction_retries = 0; - int no_progress_loops = 0; + int compaction_retries; + int no_progress_loops; unsigned long alloc_start = jiffies; unsigned int stall_timeout = 10 * HZ; + unsigned int cpuset_mems_cookie; /* * In the slowpath, we sanity check order to avoid ever trying to @@ -3549,6 +3550,12 @@ __alloc_pages_slowpath(gfp_t gfp_mask, unsigned int order, (__GFP_ATOMIC|__GFP_DIRECT_RECLAIM))) gfp_mask &= ~__GFP_ATOMIC; +retry_cpuset: + compaction_retries = 0; + no_progress_loops = 0; + compact_priority = DEF_COMPACT_PRIORITY; + cpuset_mems_cookie = read_mems_allowed_begin(); + /* * The fast path uses conservative alloc_flags to succeed only until * kswapd needs to be woken up, and to avoid the cost of setting up @@ -3720,6 +3727,15 @@ retry: } nopage: + /* + * When updating a task's mems_allowed, it is possible to race with + * parallel threads in such a way that an allocation can fail while + * the mask is being updated. If a page allocation is about to fail, + * check if the cpuset changed during allocation and if so, retry. + */ + if (read_mems_allowed_retry(cpuset_mems_cookie)) + goto retry_cpuset; + warn_alloc(gfp_mask, "page allocation failure: order:%u", order); got_pg: @@ -3734,7 +3750,6 @@ __alloc_pages_nodemask(gfp_t gfp_mask, unsigned int order, struct zonelist *zonelist, nodemask_t *nodemask) { struct page *page; - unsigned int cpuset_mems_cookie; unsigned int alloc_flags = ALLOC_WMARK_LOW; gfp_t alloc_mask = gfp_mask; /* The gfp_t that was actually used for allocation */ struct alloc_context ac = { @@ -3771,9 +3786,6 @@ __alloc_pages_nodemask(gfp_t gfp_mask, unsigned int order, if (IS_ENABLED(CONFIG_CMA) && ac.migratetype == MIGRATE_MOVABLE) alloc_flags |= ALLOC_CMA; -retry_cpuset: - cpuset_mems_cookie = read_mems_allowed_begin(); - /* Dirty zone balancing only done in the fast path */ ac.spread_dirty_pages = (gfp_mask & __GFP_WRITE); @@ -3786,6 +3798,11 @@ retry_cpuset: ac.high_zoneidx, ac.nodemask); if (!ac.preferred_zoneref->zone) { page = NULL; + /* + * This might be due to race with cpuset_current_mems_allowed + * update, so make sure we retry with original nodemask in the + * slow path. + */ goto no_zone; } @@ -3794,6 +3811,7 @@ retry_cpuset: if (likely(page)) goto out; +no_zone: /* * Runtime PM, block IO and its error handling path can deadlock * because I/O on the device might not complete. @@ -3811,24 +3829,11 @@ retry_cpuset: ac.nodemask = nodemask; ac.preferred_zoneref = first_zones_zonelist(ac.zonelist, ac.high_zoneidx, ac.nodemask); - if (!ac.preferred_zoneref->zone) - goto no_zone; + /* If we have NULL preferred zone, slowpath wll handle that */ } page = __alloc_pages_slowpath(alloc_mask, order, &ac); -no_zone: - /* - * When updating a task's mems_allowed, it is possible to race with - * parallel threads in such a way that an allocation can fail while - * the mask is being updated. If a page allocation is about to fail, - * check if the cpuset changed during allocation and if so, retry. - */ - if (unlikely(!page && read_mems_allowed_retry(cpuset_mems_cookie))) { - alloc_mask = gfp_mask; - goto retry_cpuset; - } - out: if (memcg_kmem_enabled() && (gfp_mask & __GFP_ACCOUNT) && page && unlikely(memcg_kmem_charge(page, gfp_mask, order) != 0)) { From e47483bca2cc59a4593b37a270b16ee42b1d9f08 Mon Sep 17 00:00:00 2001 From: Vlastimil Babka Date: Tue, 24 Jan 2017 15:18:41 -0800 Subject: [PATCH 254/343] mm, page_alloc: fix premature OOM when racing with cpuset mems update Ganapatrao Kulkarni reported that the LTP test cpuset01 in stress mode triggers OOM killer in few seconds, despite lots of free memory. The test attempts to repeatedly fault in memory in one process in a cpuset, while changing allowed nodes of the cpuset between 0 and 1 in another process. The problem comes from insufficient protection against cpuset changes, which can cause get_page_from_freelist() to consider all zones as non-eligible due to nodemask and/or current->mems_allowed. This was masked in the past by sufficient retries, but since commit 682a3385e773 ("mm, page_alloc: inline the fast path of the zonelist iterator") we fix the preferred_zoneref once, and don't iterate over the whole zonelist in further attempts, thus the only eligible zones might be placed in the zonelist before our starting point and we always miss them. A previous patch fixed this problem for current->mems_allowed. However, cpuset changes also update the task's mempolicy nodemask. The fix has two parts. We have to repeat the preferred_zoneref search when we detect cpuset update by way of seqcount, and we have to check the seqcount before considering OOM. [akpm@linux-foundation.org: fix typo in comment] Link: http://lkml.kernel.org/r/20170120103843.24587-5-vbabka@suse.cz Fixes: c33d6c06f60f ("mm, page_alloc: avoid looking up the first zone in a zonelist twice") Signed-off-by: Vlastimil Babka Reported-by: Ganapatrao Kulkarni Acked-by: Mel Gorman Acked-by: Hillf Danton Cc: Michal Hocko Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 35 ++++++++++++++++++++++++----------- 1 file changed, 24 insertions(+), 11 deletions(-) diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 0df3c089d3af..f3e0c69a97b7 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -3555,6 +3555,17 @@ retry_cpuset: no_progress_loops = 0; compact_priority = DEF_COMPACT_PRIORITY; cpuset_mems_cookie = read_mems_allowed_begin(); + /* + * We need to recalculate the starting point for the zonelist iterator + * because we might have used different nodemask in the fast path, or + * there was a cpuset modification and we are retrying - otherwise we + * could end up iterating over non-eligible zones endlessly. + */ + ac->preferred_zoneref = first_zones_zonelist(ac->zonelist, + ac->high_zoneidx, ac->nodemask); + if (!ac->preferred_zoneref->zone) + goto nopage; + /* * The fast path uses conservative alloc_flags to succeed only until @@ -3715,6 +3726,13 @@ retry: &compaction_retries)) goto retry; + /* + * It's possible we raced with cpuset update so the OOM would be + * premature (see below the nopage: label for full explanation). + */ + if (read_mems_allowed_retry(cpuset_mems_cookie)) + goto retry_cpuset; + /* Reclaim has failed us, start killing things */ page = __alloc_pages_may_oom(gfp_mask, order, ac, &did_some_progress); if (page) @@ -3728,10 +3746,11 @@ retry: nopage: /* - * When updating a task's mems_allowed, it is possible to race with - * parallel threads in such a way that an allocation can fail while - * the mask is being updated. If a page allocation is about to fail, - * check if the cpuset changed during allocation and if so, retry. + * When updating a task's mems_allowed or mempolicy nodemask, it is + * possible to race with parallel threads in such a way that our + * allocation can fail while the mask is being updated. If we are about + * to fail, check if the cpuset changed during allocation and if so, + * retry. */ if (read_mems_allowed_retry(cpuset_mems_cookie)) goto retry_cpuset; @@ -3822,15 +3841,9 @@ no_zone: /* * Restore the original nodemask if it was potentially replaced with * &cpuset_current_mems_allowed to optimize the fast-path attempt. - * Also recalculate the starting point for the zonelist iterator or - * we could end up iterating over non-eligible zones endlessly. */ - if (unlikely(ac.nodemask != nodemask)) { + if (unlikely(ac.nodemask != nodemask)) ac.nodemask = nodemask; - ac.preferred_zoneref = first_zones_zonelist(ac.zonelist, - ac.high_zoneidx, ac.nodemask); - /* If we have NULL preferred zone, slowpath wll handle that */ - } page = __alloc_pages_slowpath(alloc_mask, order, &ac); From 4180c4c170a5a33b9987b314d248a9d572d89ab0 Mon Sep 17 00:00:00 2001 From: Sudip Mukherjee Date: Tue, 24 Jan 2017 15:18:43 -0800 Subject: [PATCH 255/343] frv: add missing atomic64 operations Some more atomic64 operations were missing and as a result frv allmodconfig was failing. Add the missing operations. Link: http://lkml.kernel.org/r/1485193844-12850-1-git-send-email-sudip.mukherjee@codethink.co.uk Signed-off-by: Sudip Mukherjee Cc: David Howells Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/frv/include/asm/atomic.h | 19 ++++++++++++++++++- 1 file changed, 18 insertions(+), 1 deletion(-) diff --git a/arch/frv/include/asm/atomic.h b/arch/frv/include/asm/atomic.h index 994ed3d5ca08..e93c9494503a 100644 --- a/arch/frv/include/asm/atomic.h +++ b/arch/frv/include/asm/atomic.h @@ -139,7 +139,7 @@ static inline void atomic64_dec(atomic64_t *v) #define atomic64_sub_and_test(i,v) (atomic64_sub_return((i), (v)) == 0) #define atomic64_dec_and_test(v) (atomic64_dec_return((v)) == 0) #define atomic64_inc_and_test(v) (atomic64_inc_return((v)) == 0) - +#define atomic64_inc_not_zero(v) atomic64_add_unless((v), 1, 0) #define atomic_cmpxchg(v, old, new) (cmpxchg(&(v)->counter, old, new)) #define atomic_xchg(v, new) (xchg(&(v)->counter, new)) @@ -177,6 +177,23 @@ static inline int atomic64_add_unless(atomic64_t *v, long long i, long long u) return c != u; } +static inline long long atomic64_dec_if_positive(atomic64_t *v) +{ + long long c, old, dec; + + c = atomic64_read(v); + for (;;) { + dec = c - 1; + if (unlikely(dec < 0)) + break; + old = atomic64_cmpxchg((v), c, dec); + if (likely(old == c)) + break; + c = old; + } + return dec; +} + #define ATOMIC_OP(op) \ static inline int atomic_fetch_##op(int i, atomic_t *v) \ { \ From f598f82e204ec0b17797caaf1b0311c52d43fb9a Mon Sep 17 00:00:00 2001 From: Coly Li Date: Tue, 24 Jan 2017 15:18:46 -0800 Subject: [PATCH 256/343] romfs: use different way to generate fsid for BLOCK or MTD Commit 8a59f5d25265 ("fs/romfs: return f_fsid for statfs(2)") generates a 64bit id from sb->s_bdev->bd_dev. This is only correct when romfs is defined with CONFIG_ROMFS_ON_BLOCK. If romfs is only defined with CONFIG_ROMFS_ON_MTD, sb->s_bdev is NULL, referencing sb->s_bdev->bd_dev will triger an oops. Richard Weinberger points out that when CONFIG_ROMFS_BACKED_BY_BOTH=y, both CONFIG_ROMFS_ON_BLOCK and CONFIG_ROMFS_ON_MTD are defined. Therefore when calling huge_encode_dev() to generate a 64bit id, I use the follow order to choose parameter, - CONFIG_ROMFS_ON_BLOCK defined use sb->s_bdev->bd_dev - CONFIG_ROMFS_ON_BLOCK undefined and CONFIG_ROMFS_ON_MTD defined use sb->s_dev when, - both CONFIG_ROMFS_ON_BLOCK and CONFIG_ROMFS_ON_MTD undefined leave id as 0 When CONFIG_ROMFS_ON_MTD is defined and sb->s_mtd is not NULL, sb->s_dev is set to a device ID generated by MTD_BLOCK_MAJOR and mtd index, otherwise sb->s_dev is 0. This is a try-best effort to generate a uniq file system ID, if all the above conditions are not meet, f_fsid of this romfs instance will be 0. Generally only one romfs can be built on single MTD block device, this method is enough to identify multiple romfs instances in a computer. Link: http://lkml.kernel.org/r/1482928596-115155-1-git-send-email-colyli@suse.de Signed-off-by: Coly Li Reported-by: Nong Li Tested-by: Nong Li Cc: Richard Weinberger Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/romfs/super.c | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/fs/romfs/super.c b/fs/romfs/super.c index d0f8a38dfafa..0186fe6d39f3 100644 --- a/fs/romfs/super.c +++ b/fs/romfs/super.c @@ -74,6 +74,7 @@ #include #include #include +#include #include "internal.h" static struct kmem_cache *romfs_inode_cachep; @@ -416,7 +417,22 @@ static void romfs_destroy_inode(struct inode *inode) static int romfs_statfs(struct dentry *dentry, struct kstatfs *buf) { struct super_block *sb = dentry->d_sb; - u64 id = huge_encode_dev(sb->s_bdev->bd_dev); + u64 id = 0; + + /* When calling huge_encode_dev(), + * use sb->s_bdev->bd_dev when, + * - CONFIG_ROMFS_ON_BLOCK defined + * use sb->s_dev when, + * - CONFIG_ROMFS_ON_BLOCK undefined and + * - CONFIG_ROMFS_ON_MTD defined + * leave id as 0 when, + * - CONFIG_ROMFS_ON_BLOCK undefined and + * - CONFIG_ROMFS_ON_MTD undefined + */ + if (sb->s_bdev) + id = huge_encode_dev(sb->s_bdev->bd_dev); + else if (sb->s_dev) + id = huge_encode_dev(sb->s_dev); buf->f_type = ROMFS_MAGIC; buf->f_namelen = ROMFS_MAXFN; @@ -489,6 +505,11 @@ static int romfs_fill_super(struct super_block *sb, void *data, int silent) sb->s_flags |= MS_RDONLY | MS_NOATIME; sb->s_op = &romfs_super_ops; +#ifdef CONFIG_ROMFS_ON_MTD + /* Use same dev ID from the underlying mtdblock device */ + if (sb->s_mtd) + sb->s_dev = MKDEV(MTD_BLOCK_MAJOR, sb->s_mtd->index); +#endif /* read the image superblock and check it */ rsb = kmalloc(512, GFP_KERNEL); if (!rsb) From 3705ccfdd1e8b539225ce20e3925a945cc788d67 Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Tue, 24 Jan 2017 15:18:49 -0800 Subject: [PATCH 257/343] mn10300: fix build error of missing fpu_save() When CONFIG_FPU is not enabled on arch/mn10300, causes a build error with a call to fpu_save(): kernel/built-in.o: In function `.L410': core.c:(.sched.text+0x28a): undefined reference to `fpu_save' Fix this by including in so that an empty static inline fpu_save() is defined. Link: http://lkml.kernel.org/r/dc421c4f-4842-4429-1b99-92865c2f24b6@infradead.org Signed-off-by: Randy Dunlap Reported-by: kbuild test robot Reviewed-by: David Howells Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/mn10300/include/asm/switch_to.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/mn10300/include/asm/switch_to.h b/arch/mn10300/include/asm/switch_to.h index 393d311735c8..67e333aa7629 100644 --- a/arch/mn10300/include/asm/switch_to.h +++ b/arch/mn10300/include/asm/switch_to.h @@ -16,7 +16,7 @@ struct task_struct; struct thread_struct; -#if !defined(CONFIG_LAZY_SAVE_FPU) +#if defined(CONFIG_FPU) && !defined(CONFIG_LAZY_SAVE_FPU) struct fpu_state_struct; extern asmlinkage void fpu_save(struct fpu_state_struct *); #define switch_fpu(prev, next) \ From 3277953de2f31dd03c6375e9a9f680ac37fc9d27 Mon Sep 17 00:00:00 2001 From: zhong jiang Date: Tue, 24 Jan 2017 15:18:52 -0800 Subject: [PATCH 258/343] mm: do not export ioremap_page_range symbol for external module Recently, I've found cases in which ioremap_page_range was used incorrectly, in external modules, leading to crashes. This can be partly attributed to the fact that ioremap_page_range is lower-level, with fewer protections, as compared to the other functions that an external module would typically call. Those include: ioremap_cache ioremap_nocache ioremap_prot ioremap_uc ioremap_wc ioremap_wt ...each of which wraps __ioremap_caller, which in turn provides a safer way to achieve the mapping. Therefore, stop EXPORT-ing ioremap_page_range. Link: http://lkml.kernel.org/r/1485173220-29010-1-git-send-email-zhongjiang@huawei.com Signed-off-by: zhong jiang Reviewed-by: John Hubbard Suggested-by: John Hubbard Acked-by: Michal Hocko Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- lib/ioremap.c | 1 - 1 file changed, 1 deletion(-) diff --git a/lib/ioremap.c b/lib/ioremap.c index 86c8911b0e3a..a3e14ce92a56 100644 --- a/lib/ioremap.c +++ b/lib/ioremap.c @@ -144,4 +144,3 @@ int ioremap_page_range(unsigned long addr, return err; } -EXPORT_SYMBOL_GPL(ioremap_page_range); From 534c9dc982aca01b630297ad5637f6e95e94c1e2 Mon Sep 17 00:00:00 2001 From: Dan Streetman Date: Tue, 24 Jan 2017 15:18:55 -0800 Subject: [PATCH 259/343] MAINTAINERS: add Dan Streetman to zswap maintainers Add myself as zswap maintainer. Link: http://lkml.kernel.org/r/20170124212200.19052-1-ddstreet@ieee.org Signed-off-by: Dan Streetman Acked-by: Seth Jennings Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- MAINTAINERS | 1 + 1 file changed, 1 insertion(+) diff --git a/MAINTAINERS b/MAINTAINERS index 2fd11b439ab3..c79b519fb422 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -13679,6 +13679,7 @@ F: Documentation/vm/zsmalloc.txt ZSWAP COMPRESSED SWAP CACHING M: Seth Jennings +M: Dan Streetman L: linux-mm@kvack.org S: Maintained F: mm/zswap.c From aab45453ff5c77200c6da4ac909f7a4392aed17e Mon Sep 17 00:00:00 2001 From: Dan Streetman Date: Tue, 24 Jan 2017 15:18:57 -0800 Subject: [PATCH 260/343] MAINTAINERS: add Dan Streetman to zbud maintainers Add myself as zbud maintainer. Link: http://lkml.kernel.org/r/20170124221705.26523-1-ddstreet@ieee.org Signed-off-by: Dan Streetman Cc: Seth Jennings Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- MAINTAINERS | 1 + 1 file changed, 1 insertion(+) diff --git a/MAINTAINERS b/MAINTAINERS index c79b519fb422..bbf74a9d6b47 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -13624,6 +13624,7 @@ F: drivers/net/hamradio/z8530.h ZBUD COMPRESSED PAGE ALLOCATOR M: Seth Jennings +M: Dan Streetman L: linux-mm@kvack.org S: Maintained F: mm/zbud.c From 3feb479cea37fc623cf4e705631b2e679cbfbd7a Mon Sep 17 00:00:00 2001 From: Fabio Estevam Date: Mon, 23 Jan 2017 13:13:58 -0200 Subject: [PATCH 261/343] Revert "thermal: thermal_hwmon: Convert to hwmon_device_register_with_info()" This reverts commit 7611fb68062f ("thermal: thermal_hwmon: Convert to hwmon_device_register_with_info()"). Pavel Machek reported breakage in the Nokia N900 due to this commit. We can revisit a proper fix for the warning later. Reported-by: Pavel Machek Signed-off-by: Fabio Estevam Acked-by: Guenter Roeck Acked-by: Pavel Machek Signed-off-by: Zhang Rui --- drivers/thermal/thermal_hwmon.c | 20 +++++++++++++++++--- 1 file changed, 17 insertions(+), 3 deletions(-) diff --git a/drivers/thermal/thermal_hwmon.c b/drivers/thermal/thermal_hwmon.c index c4a508a124dc..541af5946203 100644 --- a/drivers/thermal/thermal_hwmon.c +++ b/drivers/thermal/thermal_hwmon.c @@ -58,6 +58,14 @@ static LIST_HEAD(thermal_hwmon_list); static DEFINE_MUTEX(thermal_hwmon_list_lock); +static ssize_t +name_show(struct device *dev, struct device_attribute *attr, char *buf) +{ + struct thermal_hwmon_device *hwmon = dev_get_drvdata(dev); + return sprintf(buf, "%s\n", hwmon->type); +} +static DEVICE_ATTR_RO(name); + static ssize_t temp_input_show(struct device *dev, struct device_attribute *attr, char *buf) { @@ -157,12 +165,15 @@ int thermal_add_hwmon_sysfs(struct thermal_zone_device *tz) INIT_LIST_HEAD(&hwmon->tz_list); strlcpy(hwmon->type, tz->type, THERMAL_NAME_LENGTH); - hwmon->device = hwmon_device_register_with_info(NULL, hwmon->type, - hwmon, NULL, NULL); + hwmon->device = hwmon_device_register(NULL); if (IS_ERR(hwmon->device)) { result = PTR_ERR(hwmon->device); goto free_mem; } + dev_set_drvdata(hwmon->device, hwmon); + result = device_create_file(hwmon->device, &dev_attr_name); + if (result) + goto free_mem; register_sys_interface: temp = kzalloc(sizeof(*temp), GFP_KERNEL); @@ -211,8 +222,10 @@ int thermal_add_hwmon_sysfs(struct thermal_zone_device *tz) free_temp_mem: kfree(temp); unregister_name: - if (new_hwmon_device) + if (new_hwmon_device) { + device_remove_file(hwmon->device, &dev_attr_name); hwmon_device_unregister(hwmon->device); + } free_mem: if (new_hwmon_device) kfree(hwmon); @@ -254,6 +267,7 @@ void thermal_remove_hwmon_sysfs(struct thermal_zone_device *tz) list_del(&hwmon->node); mutex_unlock(&thermal_hwmon_list_lock); + device_remove_file(hwmon->device, &dev_attr_name); hwmon_device_unregister(hwmon->device); kfree(hwmon); } From bdbfd5196d24a6d0845b549eba6ce8e6fa8bb3d0 Mon Sep 17 00:00:00 2001 From: Alex Williamson Date: Tue, 24 Jan 2017 12:53:45 -0700 Subject: [PATCH 262/343] drm/i915/gvt/kvmgt: mdev ABI is available_instances, not available_instance Per the ABI specification[1], each mdev_supported_types entry should have an available_instances, with an "s", not available_instance. [1] Documentation/ABI/testing/sysfs-bus-vfio-mdev Signed-off-by: Alex Williamson Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/kvmgt.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/i915/gvt/kvmgt.c b/drivers/gpu/drm/i915/gvt/kvmgt.c index 0c9234a87a20..3f656e3a6e5a 100644 --- a/drivers/gpu/drm/i915/gvt/kvmgt.c +++ b/drivers/gpu/drm/i915/gvt/kvmgt.c @@ -230,8 +230,8 @@ static struct intel_vgpu_type *intel_gvt_find_vgpu_type(struct intel_gvt *gvt, return NULL; } -static ssize_t available_instance_show(struct kobject *kobj, struct device *dev, - char *buf) +static ssize_t available_instances_show(struct kobject *kobj, + struct device *dev, char *buf) { struct intel_vgpu_type *type; unsigned int num = 0; @@ -269,12 +269,12 @@ static ssize_t description_show(struct kobject *kobj, struct device *dev, type->fence); } -static MDEV_TYPE_ATTR_RO(available_instance); +static MDEV_TYPE_ATTR_RO(available_instances); static MDEV_TYPE_ATTR_RO(device_api); static MDEV_TYPE_ATTR_RO(description); static struct attribute *type_attrs[] = { - &mdev_type_attr_available_instance.attr, + &mdev_type_attr_available_instances.attr, &mdev_type_attr_device_api.attr, &mdev_type_attr_description.attr, NULL, From 7283accfaef66e6a64f7d3ec0672596dd8e5b144 Mon Sep 17 00:00:00 2001 From: Alex Williamson Date: Tue, 24 Jan 2017 13:15:43 -0700 Subject: [PATCH 263/343] drm/i915/gvt: Fix kmem_cache_create() name According to kmem_cache_sanity_check(), spaces are not allowed in the name of a cache and results in a kernel oops with CONFIG_DEBUG_VM. Convert to underscores. Signed-off-by: Alex Williamson Signed-off-by: Zhenyu Wang --- drivers/gpu/drm/i915/gvt/execlist.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/gvt/execlist.c b/drivers/gpu/drm/i915/gvt/execlist.c index fb852c51d00e..34083731669d 100644 --- a/drivers/gpu/drm/i915/gvt/execlist.c +++ b/drivers/gpu/drm/i915/gvt/execlist.c @@ -798,7 +798,7 @@ int intel_vgpu_init_execlist(struct intel_vgpu *vgpu) INIT_LIST_HEAD(&vgpu->workload_q_head[i]); } - vgpu->workloads = kmem_cache_create("gvt-g vgpu workload", + vgpu->workloads = kmem_cache_create("gvt-g_vgpu_workload", sizeof(struct intel_vgpu_workload), 0, SLAB_HWCACHE_ALIGN, NULL); From ba7addcd805e5c83e201b118a2693b921a980b44 Mon Sep 17 00:00:00 2001 From: Zhenyu Wang Date: Wed, 25 Jan 2017 10:30:02 +0800 Subject: [PATCH 264/343] MAINTAINERS: update new mail list for intel gvt driver We've moved to lists.freedesktop.org from lists.01.org. Update info in MAINTAINERS. Signed-off-by: Zhenyu Wang --- MAINTAINERS | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/MAINTAINERS b/MAINTAINERS index 5f0420a0da5b..5bd03d5cb28b 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -4153,7 +4153,7 @@ F: Documentation/gpu/i915.rst INTEL GVT-g DRIVERS (Intel GPU Virtualization) M: Zhenyu Wang M: Zhi Wang -L: igvt-g-dev@lists.01.org +L: intel-gvt-dev@lists.freedesktop.org L: intel-gfx@lists.freedesktop.org W: https://01.org/igvt-g T: git https://github.com/01org/gvt-linux.git From b78671591a10218ab18bbea120fd05df7a002e88 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Sun, 15 Jan 2017 12:58:25 +0000 Subject: [PATCH 265/343] drm/i915: Avoid drm_atomic_state_put(NULL) in intel_display_resume intel_display_resume() may be called without an atomic state to restore, i.e. dev_priv->modeset_reset_restore state is NULL. One such case is following a lid open/close event and the forced modeset in intel_lid_notify(). Reported-by: Stefan Seyfried Tested-by: Stefan Seyfried Fixes: 0853695c3ba4 ("drm: Add reference counting to drm_atomic_state") Signed-off-by: Chris Wilson Cc: Daniel Vetter Cc: Jani Nikula Cc: # v4.10-rc1+ Link: http://patchwork.freedesktop.org/patch/msgid/20170115125825.18597-1-chris@chris-wilson.co.uk Reviewed-by: Ander Conselvan de Oliveira (cherry picked from commit 3c5e37f169cb67cbd03c6116fbc93e0805815d29) Signed-off-by: Jani Nikula --- drivers/gpu/drm/i915/intel_display.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 8d702cf1a616..1a01d4189c3f 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -17027,7 +17027,8 @@ void intel_display_resume(struct drm_device *dev) if (ret) DRM_ERROR("Restoring old state failed with %i\n", ret); - drm_atomic_state_put(state); + if (state) + drm_atomic_state_put(state); } void intel_modeset_gem_init(struct drm_device *dev) From 27892bbdc9233f33bf0f44e08aab8f12e0dec142 Mon Sep 17 00:00:00 2001 From: Clint Taylor Date: Wed, 18 Jan 2017 13:38:43 -0800 Subject: [PATCH 266/343] drm/i915: prevent crash with .disable_display parameter The .disable_display parameter was causing a fatal crash when fbdev was dereferenced during driver init. V1: protection in i915_drv.c V2: Moved protection to intel_fbdev.c Fixes: 43cee314345a ("drm/i915/fbdev: Limit the global async-domain synchronization") Testcase: igt/drv_module_reload/basic-no-display Cc: Chris Wilson Signed-off-by: Clint Taylor Link: http://patchwork.freedesktop.org/patch/msgid/1484775523-29428-1-git-send-email-clinton.a.taylor@intel.com Reviewed-by: Chris Wilson Cc: Lukas Wunner Cc: Daniel Vetter Cc: Jani Nikula Cc: # v4.8+ Signed-off-by: Chris Wilson (cherry picked from commit 5b8cd0755f8a06a851c436a013e7be0823fb155a) Signed-off-by: Jani Nikula --- drivers/gpu/drm/i915/intel_fbdev.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/gpu/drm/i915/intel_fbdev.c b/drivers/gpu/drm/i915/intel_fbdev.c index beb08982dc0b..8cf2d80f2254 100644 --- a/drivers/gpu/drm/i915/intel_fbdev.c +++ b/drivers/gpu/drm/i915/intel_fbdev.c @@ -742,6 +742,9 @@ void intel_fbdev_initial_config_async(struct drm_device *dev) { struct intel_fbdev *ifbdev = to_i915(dev)->fbdev; + if (!ifbdev) + return; + ifbdev->cookie = async_schedule(intel_fbdev_initial_config, ifbdev); } From a38a7bd1766b42ea0ed14b99be23a653922ed5c8 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Thu, 19 Jan 2017 11:37:49 +0000 Subject: [PATCH 267/343] drm/i915: Release temporary load-detect state upon switching MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit After we call drm_atomic_commit() on the load-detect state, we can free our local reference. Upon restore, we only apply and free the previous state. Fixes: 0853695c3ba4 ("drm: Add reference counting to drm_atomic_state") Signed-off-by: Chris Wilson Cc: Chris Wilson Cc: Daniel Vetter Cc: # v4.10-rc1+ Link: http://patchwork.freedesktop.org/patch/msgid/20170119113749.2517-1-chris@chris-wilson.co.uk Reviewed-by: Ville Syrjälä (cherry picked from commit 7abbd11f344aa7abe29befb218774a1ea26018ac) Signed-off-by: Jani Nikula --- drivers/gpu/drm/i915/intel_display.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 1a01d4189c3f..f313d62a7e74 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -11246,6 +11246,7 @@ found: } old->restore_state = restore_state; + drm_atomic_state_put(state); /* let the connector get through one full cycle before testing */ intel_wait_for_vblank(dev_priv, intel_crtc->pipe); From c34f078675f505c4437919bb1897b1351f16a050 Mon Sep 17 00:00:00 2001 From: Ander Conselvan de Oliveira Date: Fri, 20 Jan 2017 16:28:42 +0200 Subject: [PATCH 268/343] drm/i915: Don't leak edid in intel_crt_detect_ddc() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit In the path where intel_crt_detect_ddc() detects a CRT, if would return true without freeing the edid. Fixes: a2bd1f541f19 ("drm/i915: check whether we actually received an edid in detect_ddc") Cc: Chris Wilson Cc: Daniel Vetter Cc: Daniel Vetter Cc: Jani Nikula Cc: intel-gfx@lists.freedesktop.org Cc: # v3.6+ Signed-off-by: Ander Conselvan de Oliveira Reviewed-by: Ville Syrjälä Reviewed-by: Jani Nikula Link: http://patchwork.freedesktop.org/patch/msgid/1484922525-6131-1-git-send-email-ander.conselvan.de.oliveira@intel.com (cherry picked from commit c96b63a6a7ac4bd670ec2e663793a9a31418b790) Signed-off-by: Jani Nikula --- drivers/gpu/drm/i915/intel_crt.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_crt.c b/drivers/gpu/drm/i915/intel_crt.c index 86ecec5601d4..588470eb8d39 100644 --- a/drivers/gpu/drm/i915/intel_crt.c +++ b/drivers/gpu/drm/i915/intel_crt.c @@ -499,6 +499,7 @@ static bool intel_crt_detect_ddc(struct drm_connector *connector) struct drm_i915_private *dev_priv = to_i915(crt->base.base.dev); struct edid *edid; struct i2c_adapter *i2c; + bool ret = false; BUG_ON(crt->base.type != INTEL_OUTPUT_ANALOG); @@ -515,17 +516,17 @@ static bool intel_crt_detect_ddc(struct drm_connector *connector) */ if (!is_digital) { DRM_DEBUG_KMS("CRT detected via DDC:0x50 [EDID]\n"); - return true; + ret = true; + } else { + DRM_DEBUG_KMS("CRT not detected via DDC:0x50 [EDID reports a digital panel]\n"); } - - DRM_DEBUG_KMS("CRT not detected via DDC:0x50 [EDID reports a digital panel]\n"); } else { DRM_DEBUG_KMS("CRT not detected via DDC:0x50 [no valid EDID found]\n"); } kfree(edid); - return false; + return ret; } static enum drm_connector_status From 21d6e0bde50713922a6520ef84e5fd245b05d468 Mon Sep 17 00:00:00 2001 From: Ander Conselvan de Oliveira Date: Fri, 20 Jan 2017 16:28:43 +0200 Subject: [PATCH 269/343] drm/i915: Don't init hpd polling for vlv and chv from runtime_suspend() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit An error in the condition for avoiding the call to intel_hpd_poll_init() for valleyview and cherryview from intel_runtime_suspend() caused it to be called unconditionally. Fix it. Fixes: 19625e85c6ec ("drm/i915: Enable polling when we don't have hpd") Cc: stable@vger.kernel.org Cc: Ville Syrjälä Cc: Daniel Vetter Cc: Lyude Cc: Daniel Vetter Cc: Jani Nikula Cc: intel-gfx@lists.freedesktop.org Cc: # v4.9+ Signed-off-by: Ander Conselvan de Oliveira Reviewed-by: Ville Syrjälä Link: http://patchwork.freedesktop.org/patch/msgid/1484922525-6131-2-git-send-email-ander.conselvan.de.oliveira@intel.com (cherry picked from commit 04313b00b79405f86d815100f85c47a2ee5b8ca0) Signed-off-by: Jani Nikula --- drivers/gpu/drm/i915/i915_drv.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index 445fec9c2841..b2c4a0b8a627 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -2378,7 +2378,7 @@ static int intel_runtime_suspend(struct device *kdev) assert_forcewakes_inactive(dev_priv); - if (!IS_VALLEYVIEW(dev_priv) || !IS_CHERRYVIEW(dev_priv)) + if (!IS_VALLEYVIEW(dev_priv) && !IS_CHERRYVIEW(dev_priv)) intel_hpd_poll_init(dev_priv); DRM_DEBUG_KMS("Device suspended\n"); From 3781bd6e7d64d5f5bea9fdee11ab9460a700c0e4 Mon Sep 17 00:00:00 2001 From: Ander Conselvan de Oliveira Date: Fri, 20 Jan 2017 16:28:44 +0200 Subject: [PATCH 270/343] drm/i915: Fix calculation of rotated x and y offsets for planar formats MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Parameters tile_size, tile_width and tile_height were passed in the wrong order to _intel_adjust_tile_offset() when calculating the rotated offsets. This doesn't fix any user visible bug, since for packed formats new and old offset are the same and the rotated offsets are within a tile before they are fed to _intel_adjust_tile_offset(). In that case, the offsets are unchanged. That is not true for planar formats, but those are currently not supported. Fixes: 66a2d927cb0e ("drm/i915: Make intel_adjust_tile_offset() work for linear buffers") Cc: Ville Syrjälä Cc: Sivakumar Thulasimani Cc: Daniel Vetter Cc: Jani Nikula Cc: intel-gfx@lists.freedesktop.org Cc: # v4.9+ Signed-off-by: Ander Conselvan de Oliveira Reviewed-by: Ville Syrjälä Link: http://patchwork.freedesktop.org/patch/msgid/1484922525-6131-3-git-send-email-ander.conselvan.de.oliveira@intel.com (cherry picked from commit 46a1bd289507dfcc428fb9daf65421ed6be6af8b) Signed-off-by: Jani Nikula --- drivers/gpu/drm/i915/intel_display.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index f313d62a7e74..7a9040fc6e5a 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -2585,8 +2585,9 @@ intel_fill_fb_info(struct drm_i915_private *dev_priv, * We only keep the x/y offsets, so push all of the * gtt offset into the x/y offsets. */ - _intel_adjust_tile_offset(&x, &y, tile_size, - tile_width, tile_height, pitch_tiles, + _intel_adjust_tile_offset(&x, &y, + tile_width, tile_height, + tile_size, pitch_tiles, gtt_offset_rotated * tile_size, 0); gtt_offset_rotated += rot_info->plane[i].width * rot_info->plane[i].height; From 6d1d427a4e24c403b4adf928d61994bdaa0ca03a Mon Sep 17 00:00:00 2001 From: Ander Conselvan de Oliveira Date: Fri, 20 Jan 2017 16:28:45 +0200 Subject: [PATCH 271/343] drm/i915: Check for NULL atomic state in intel_crtc_disable_noatomic() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit In intel_crtc_disable_noatomic(), bail on a failure to allocate an atomic state to avoid a NULL pointer dereference. Fixes: 4a80655827af ("drm/i915: Pass atomic state to crtc enable/disable functions") Cc: Maarten Lankhorst Cc: Daniel Vetter Cc: Daniel Vetter Cc: Jani Nikula Cc: intel-gfx@lists.freedesktop.org Cc: # v4.9+ Signed-off-by: Ander Conselvan de Oliveira Reviewed-by: Ville Syrjälä Link: http://patchwork.freedesktop.org/patch/msgid/1484922525-6131-4-git-send-email-ander.conselvan.de.oliveira@intel.com (cherry picked from commit 31bb2ef97ea9db343348f9b5ccaa9bb6f48fc655) Signed-off-by: Jani Nikula --- drivers/gpu/drm/i915/intel_display.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 7a9040fc6e5a..f6dd073afdc9 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -6850,6 +6850,12 @@ static void intel_crtc_disable_noatomic(struct drm_crtc *crtc) } state = drm_atomic_state_alloc(crtc->dev); + if (!state) { + DRM_DEBUG_KMS("failed to disable [CRTC:%d:%s], out of memory", + crtc->base.id, crtc->name); + return; + } + state->acquire_ctx = crtc->dev->mode_config.acquire_ctx; /* Everything's already locked, -EDEADLK can't happen. */ From 6f0f02dc56f18760b46dc1bf5b3f7386869d4162 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Mon, 23 Jan 2017 21:29:39 +0000 Subject: [PATCH 272/343] drm/i915: Move atomic state free from out of fence release Fences are required to support being released from under an atomic context. The drm_atomic_state struct may take a mutex when being released and so we cannot drop a reference to the drm_atomic_state from the fence release path directly, and so we need to defer that unreference to a worker. [ 326.576697] WARNING: CPU: 2 PID: 366 at kernel/sched/core.c:7737 __might_sleep+0x5d/0x80 [ 326.576816] do not call blocking ops when !TASK_RUNNING; state=1 set at [] intel_breadcrumbs_signaler+0x59/0x270 [i915] [ 326.576818] Modules linked in: rfcomm fuse snd_hda_codec_hdmi bnep snd_hda_codec_realtek snd_hda_codec_generic snd_hda_intel snd_hda_codec snd_hwdep snd_hda_core snd_pcm snd_seq_midi snd_seq_midi_event snd_rawmidi snd_seq snd_seq_device snd_timer input_leds led_class snd punit_atom_debug btusb btrtl btbcm btintel intel_rapl bluetooth i915 drm_kms_helper syscopyarea sysfillrect iwlwifi sysimgblt soundcore fb_sys_fops mei_txe cfg80211 drm pwm_lpss_platform pwm_lpss pinctrl_cherryview fjes acpi_pad parport_pc ppdev parport autofs4 [ 326.576899] CPU: 2 PID: 366 Comm: i915/signal:0 Tainted: G U 4.10.0-rc3-patser+ #5030 [ 326.576902] Hardware name: /NUC5PPYB, BIOS PYBSWCEL.86A.0031.2015.0601.1712 06/01/2015 [ 326.576905] Call Trace: [ 326.576920] dump_stack+0x4d/0x6d [ 326.576926] __warn+0xc0/0xe0 [ 326.576931] warn_slowpath_fmt+0x5a/0x80 [ 326.577004] ? intel_breadcrumbs_signaler+0x59/0x270 [i915] [ 326.577075] ? intel_breadcrumbs_signaler+0x59/0x270 [i915] [ 326.577079] __might_sleep+0x5d/0x80 [ 326.577087] mutex_lock+0x1b/0x40 [ 326.577133] drm_property_free_blob+0x1e/0x80 [drm] [ 326.577167] ? drm_property_destroy+0xe0/0xe0 [drm] [ 326.577200] drm_mode_object_unreference+0x5c/0x70 [drm] [ 326.577233] drm_property_unreference_blob+0xe/0x10 [drm] [ 326.577260] __drm_atomic_helper_crtc_destroy_state+0x14/0x40 [drm_kms_helper] [ 326.577278] drm_atomic_helper_crtc_destroy_state+0x10/0x20 [drm_kms_helper] [ 326.577352] intel_crtc_destroy_state+0x9/0x10 [i915] [ 326.577388] drm_atomic_state_default_clear+0xea/0x1d0 [drm] [ 326.577462] intel_atomic_state_clear+0xd/0x20 [i915] [ 326.577497] drm_atomic_state_clear+0x1a/0x30 [drm] [ 326.577532] __drm_atomic_state_free+0x13/0x60 [drm] [ 326.577607] intel_atomic_commit_ready+0x6f/0x78 [i915] [ 326.577670] i915_sw_fence_release+0x3a/0x50 [i915] [ 326.577733] dma_i915_sw_fence_wake+0x39/0x80 [i915] [ 326.577741] dma_fence_signal+0xda/0x120 [ 326.577812] ? intel_breadcrumbs_signaler+0x59/0x270 [i915] [ 326.577884] intel_breadcrumbs_signaler+0xb1/0x270 [i915] [ 326.577889] kthread+0x127/0x130 [ 326.577961] ? intel_engine_remove_wait+0x1a0/0x1a0 [i915] [ 326.577964] ? kthread_stop+0x120/0x120 [ 326.577970] ret_from_fork+0x22/0x30 Fixes: c004a90b7263 ("drm/i915: Restore nonblocking awaits for modesetting") Reported-by: Maarten Lankhorst Signed-off-by: Chris Wilson Cc: Chris Wilson Cc: Joonas Lahtinen Cc: Maarten Lankhorst Cc: Daniel Vetter Link: http://patchwork.freedesktop.org/patch/msgid/20170123212939.30345-1-chris@chris-wilson.co.uk Cc: # v4.10-rc1+ Reviewed-by: Joonas Lahtinen (cherry picked from commit eb955eee27d9dc176871540c43c9070ee4701642) Signed-off-by: Jani Nikula --- drivers/gpu/drm/i915/i915_drv.h | 5 +++++ drivers/gpu/drm/i915/intel_display.c | 28 ++++++++++++++++++++++++++-- drivers/gpu/drm/i915/intel_drv.h | 2 ++ 3 files changed, 33 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 243224aeabf8..69bc3b0c4390 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -1977,6 +1977,11 @@ struct drm_i915_private { struct i915_frontbuffer_tracking fb_tracking; + struct intel_atomic_helper { + struct llist_head free_list; + struct work_struct free_work; + } atomic_helper; + u16 orig_clock; bool mchbar_need_disable; diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index f6dd073afdc9..77f7b1d849a4 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -14523,8 +14523,14 @@ intel_atomic_commit_ready(struct i915_sw_fence *fence, break; case FENCE_FREE: - drm_atomic_state_put(&state->base); - break; + { + struct intel_atomic_helper *helper = + &to_i915(state->base.dev)->atomic_helper; + + if (llist_add(&state->freed, &helper->free_list)) + schedule_work(&helper->free_work); + break; + } } return NOTIFY_DONE; @@ -16403,6 +16409,18 @@ fail: drm_modeset_acquire_fini(&ctx); } +static void intel_atomic_helper_free_state(struct work_struct *work) +{ + struct drm_i915_private *dev_priv = + container_of(work, typeof(*dev_priv), atomic_helper.free_work); + struct intel_atomic_state *state, *next; + struct llist_node *freed; + + freed = llist_del_all(&dev_priv->atomic_helper.free_list); + llist_for_each_entry_safe(state, next, freed, freed) + drm_atomic_state_put(&state->base); +} + int intel_modeset_init(struct drm_device *dev) { struct drm_i915_private *dev_priv = to_i915(dev); @@ -16422,6 +16440,9 @@ int intel_modeset_init(struct drm_device *dev) dev->mode_config.funcs = &intel_mode_funcs; + INIT_WORK(&dev_priv->atomic_helper.free_work, + intel_atomic_helper_free_state); + intel_init_quirks(dev); intel_init_pm(dev_priv); @@ -17106,6 +17127,9 @@ void intel_modeset_cleanup(struct drm_device *dev) { struct drm_i915_private *dev_priv = to_i915(dev); + flush_work(&dev_priv->atomic_helper.free_work); + WARN_ON(!llist_empty(&dev_priv->atomic_helper.free_list)); + intel_disable_gt_powersave(dev_priv); /* diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index cd132c216a67..cd72ae171eeb 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -370,6 +370,8 @@ struct intel_atomic_state { struct skl_wm_values wm_results; struct i915_sw_fence commit_ready; + + struct llist_node freed; }; struct intel_plane_state { From 2f5db26c2ecb248bdc319feb2990453cb02fc950 Mon Sep 17 00:00:00 2001 From: Daniele Ceraolo Spurio Date: Fri, 20 Jan 2017 13:51:23 -0800 Subject: [PATCH 273/343] drm/i915: reinstate call to trace_i915_vma_bind The call went away in: commit 3b16525cc4c1a43e9053cfdc414356eea24bdfad Author: Chris Wilson Date: Thu Aug 4 16:32:25 2016 +0100 drm/i915: Split insertion/binding of an object into the VM It is useful to have this trace as it pairs nicely with the vma_unbind one to track vma activity. Added inside the i915_vma_bind function (was outside before) to keep a similar placement as trace_i915_vma_unbind. v2: print bind_flags instead of flags (Chris) Fixes: 3b16525cc4c1 ("drm/i915: Split insertion/binding of an object into the VM") Cc: Chris Wilson Signed-off-by: Daniele Ceraolo Spurio Link: http://patchwork.freedesktop.org/patch/msgid/1484949083-11430-1-git-send-email-daniele.ceraolospurio@intel.com Reviewed-by: Chris Wilson Signed-off-by: Chris Wilson (cherry picked from commit 6146e6da5c961735dacf9b6c0c8b5f1382193ee2) Signed-off-by: Jani Nikula --- drivers/gpu/drm/i915/i915_vma.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/gpu/drm/i915/i915_vma.c b/drivers/gpu/drm/i915/i915_vma.c index a792dcb902b5..e924a9516079 100644 --- a/drivers/gpu/drm/i915/i915_vma.c +++ b/drivers/gpu/drm/i915/i915_vma.c @@ -185,6 +185,7 @@ int i915_vma_bind(struct i915_vma *vma, enum i915_cache_level cache_level, return ret; } + trace_i915_vma_bind(vma, bind_flags); ret = vma->vm->bind_vma(vma, cache_level, bind_flags); if (ret) return ret; From fd25ea29093e275195d0ae8b2573021a1c98959f Mon Sep 17 00:00:00 2001 From: Hans de Goede Date: Sun, 22 Jan 2017 13:24:05 +0100 Subject: [PATCH 274/343] Revert "ACPI / video: Add force_native quirk for HP Pavilion dv6" Revert commit 6276e53fa8c0 (ACPI / video: Add force_native quirk for HP Pavilion dv6). In the commit message for the quirk this revert removes I wrote: "Note that there are quite a few HP Pavilion dv6 variants, some woth ATI and some with NVIDIA hybrid gfx, both seem to need this quirk to have working backlight control. There are also some versions with only Intel integrated gfx, these may not need this quirk, but it should not hurt there." Unfortunately that seems wrong, I've already received 2 reports of this commit causing regressions on some dv6 variants (at least one of which actually has a nvidia GPU). So it seems that HP has made a mess here by using the same model-name both in marketing and in the DMI data for many different variants. Some of which need acpi_backlight=native for functional backlight control (as the quirk this commit reverts was doing), where as others are broken by it. So lets get back to the old sitation so as to avoid regressing on models which used to work without any kernel cmdline arguments before. Fixes: 6276e53fa8c0 (ACPI / video: Add force_native quirk for HP Pavilion dv6) Signed-off-by: Hans de Goede Signed-off-by: Rafael J. Wysocki --- drivers/acpi/video_detect.c | 11 ----------- 1 file changed, 11 deletions(-) diff --git a/drivers/acpi/video_detect.c b/drivers/acpi/video_detect.c index 02ded25c82e4..7f48156cbc0c 100644 --- a/drivers/acpi/video_detect.c +++ b/drivers/acpi/video_detect.c @@ -305,17 +305,6 @@ static const struct dmi_system_id video_detect_dmi_table[] = { DMI_MATCH(DMI_PRODUCT_NAME, "Dell System XPS L702X"), }, }, - { - /* https://bugzilla.redhat.com/show_bug.cgi?id=1204476 */ - /* https://bugs.launchpad.net/ubuntu/+source/linux-lts-trusty/+bug/1416940 */ - .callback = video_detect_force_native, - .ident = "HP Pavilion dv6", - .matches = { - DMI_MATCH(DMI_SYS_VENDOR, "Hewlett-Packard"), - DMI_MATCH(DMI_PRODUCT_NAME, "HP Pavilion dv6 Notebook PC"), - }, - }, - { }, }; From 4dfa2b84118fd6c95202ae87e62adf5000ccd4d0 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 25 Jan 2017 07:49:34 -0800 Subject: [PATCH 275/343] xfs: only update mount/resv fields on success in __xfs_ag_resv_init Try to reserve the blocks first and only then update the fields in or hanging off the mount structure. This way we can call __xfs_ag_resv_init again after a previous failure. Signed-off-by: Christoph Hellwig Reviewed-by: Darrick J. Wong Signed-off-by: Darrick J. Wong --- fs/xfs/libxfs/xfs_ag_resv.c | 23 ++++++++++++++--------- 1 file changed, 14 insertions(+), 9 deletions(-) diff --git a/fs/xfs/libxfs/xfs_ag_resv.c b/fs/xfs/libxfs/xfs_ag_resv.c index d346d42c54d1..94234bff40dc 100644 --- a/fs/xfs/libxfs/xfs_ag_resv.c +++ b/fs/xfs/libxfs/xfs_ag_resv.c @@ -200,22 +200,27 @@ __xfs_ag_resv_init( struct xfs_mount *mp = pag->pag_mount; struct xfs_ag_resv *resv; int error; + xfs_extlen_t reserved; - resv = xfs_perag_resv(pag, type); if (used > ask) ask = used; - resv->ar_asked = ask; - resv->ar_reserved = resv->ar_orig_reserved = ask - used; - mp->m_ag_max_usable -= ask; + reserved = ask - used; - trace_xfs_ag_resv_init(pag, type, ask); - - error = xfs_mod_fdblocks(mp, -(int64_t)resv->ar_reserved, true); - if (error) + error = xfs_mod_fdblocks(mp, -(int64_t)reserved, true); + if (error) { trace_xfs_ag_resv_init_error(pag->pag_mount, pag->pag_agno, error, _RET_IP_); + return error; + } - return error; + mp->m_ag_max_usable -= ask; + + resv = xfs_perag_resv(pag, type); + resv->ar_asked = ask; + resv->ar_reserved = resv->ar_orig_reserved = reserved; + + trace_xfs_ag_resv_init(pag, type, ask); + return 0; } /* Create a per-AG block reservation. */ From 76d771b4cbe33c581bd6ca2710c120be51172440 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 25 Jan 2017 07:49:35 -0800 Subject: [PATCH 276/343] xfs: use per-AG reservations for the finobt Currently we try to rely on the global reserved block pool for block allocations for the free inode btree, but I have customer reports (fairly complex workload, need to find an easier reproducer) where that is not enough as the AG where we free an inode that requires a new finobt block is entirely full. This causes us to cancel a dirty transaction and thus a file system shutdown. I think the right way to guard against this is to treat the finot the same way as the refcount btree and have a per-AG reservations for the possible worst case size of it, and the patch below implements that. Note that this could increase mount times with large finobt trees. In an ideal world we would have added a field for the number of finobt fields to the AGI, similar to what we did for the refcount blocks. We should do add it next time we rev the AGI or AGF format by adding new fields. Signed-off-by: Christoph Hellwig Reviewed-by: Darrick J. Wong Signed-off-by: Darrick J. Wong --- fs/xfs/libxfs/xfs_ag_resv.c | 47 ++++++++++++++--- fs/xfs/libxfs/xfs_ialloc_btree.c | 90 ++++++++++++++++++++++++++++++-- fs/xfs/libxfs/xfs_ialloc_btree.h | 3 ++ fs/xfs/xfs_inode.c | 23 ++++---- fs/xfs/xfs_mount.h | 1 + 5 files changed, 144 insertions(+), 20 deletions(-) diff --git a/fs/xfs/libxfs/xfs_ag_resv.c b/fs/xfs/libxfs/xfs_ag_resv.c index 94234bff40dc..33db69be4832 100644 --- a/fs/xfs/libxfs/xfs_ag_resv.c +++ b/fs/xfs/libxfs/xfs_ag_resv.c @@ -39,6 +39,7 @@ #include "xfs_rmap_btree.h" #include "xfs_btree.h" #include "xfs_refcount_btree.h" +#include "xfs_ialloc_btree.h" /* * Per-AG Block Reservations @@ -210,6 +211,9 @@ __xfs_ag_resv_init( if (error) { trace_xfs_ag_resv_init_error(pag->pag_mount, pag->pag_agno, error, _RET_IP_); + xfs_warn(mp, +"Per-AG reservation for AG %u failed. Filesystem may run out of space.", + pag->pag_agno); return error; } @@ -228,6 +232,8 @@ int xfs_ag_resv_init( struct xfs_perag *pag) { + struct xfs_mount *mp = pag->pag_mount; + xfs_agnumber_t agno = pag->pag_agno; xfs_extlen_t ask; xfs_extlen_t used; int error = 0; @@ -236,23 +242,45 @@ xfs_ag_resv_init( if (pag->pag_meta_resv.ar_asked == 0) { ask = used = 0; - error = xfs_refcountbt_calc_reserves(pag->pag_mount, - pag->pag_agno, &ask, &used); + error = xfs_refcountbt_calc_reserves(mp, agno, &ask, &used); + if (error) + goto out; + + error = xfs_finobt_calc_reserves(mp, agno, &ask, &used); if (error) goto out; error = __xfs_ag_resv_init(pag, XFS_AG_RESV_METADATA, ask, used); - if (error) - goto out; + if (error) { + /* + * Because we didn't have per-AG reservations when the + * finobt feature was added we might not be able to + * reserve all needed blocks. Warn and fall back to the + * old and potentially buggy code in that case, but + * ensure we do have the reservation for the refcountbt. + */ + ask = used = 0; + + mp->m_inotbt_nores = true; + + error = xfs_refcountbt_calc_reserves(mp, agno, &ask, + &used); + if (error) + goto out; + + error = __xfs_ag_resv_init(pag, XFS_AG_RESV_METADATA, + ask, used); + if (error) + goto out; + } } /* Create the AGFL metadata reservation */ if (pag->pag_agfl_resv.ar_asked == 0) { ask = used = 0; - error = xfs_rmapbt_calc_reserves(pag->pag_mount, pag->pag_agno, - &ask, &used); + error = xfs_rmapbt_calc_reserves(mp, agno, &ask, &used); if (error) goto out; @@ -261,9 +289,16 @@ xfs_ag_resv_init( goto out; } +#ifdef DEBUG + /* need to read in the AGF for the ASSERT below to work */ + error = xfs_alloc_pagf_init(pag->pag_mount, NULL, pag->pag_agno, 0); + if (error) + return error; + ASSERT(xfs_perag_resv(pag, XFS_AG_RESV_METADATA)->ar_reserved + xfs_perag_resv(pag, XFS_AG_RESV_AGFL)->ar_reserved <= pag->pagf_freeblks + pag->pagf_flcount); +#endif out: return error; } diff --git a/fs/xfs/libxfs/xfs_ialloc_btree.c b/fs/xfs/libxfs/xfs_ialloc_btree.c index 0fd086d03d41..7c471881c9a6 100644 --- a/fs/xfs/libxfs/xfs_ialloc_btree.c +++ b/fs/xfs/libxfs/xfs_ialloc_btree.c @@ -82,11 +82,12 @@ xfs_finobt_set_root( } STATIC int -xfs_inobt_alloc_block( +__xfs_inobt_alloc_block( struct xfs_btree_cur *cur, union xfs_btree_ptr *start, union xfs_btree_ptr *new, - int *stat) + int *stat, + enum xfs_ag_resv_type resv) { xfs_alloc_arg_t args; /* block allocation args */ int error; /* error return value */ @@ -103,6 +104,7 @@ xfs_inobt_alloc_block( args.maxlen = 1; args.prod = 1; args.type = XFS_ALLOCTYPE_NEAR_BNO; + args.resv = resv; error = xfs_alloc_vextent(&args); if (error) { @@ -122,6 +124,27 @@ xfs_inobt_alloc_block( return 0; } +STATIC int +xfs_inobt_alloc_block( + struct xfs_btree_cur *cur, + union xfs_btree_ptr *start, + union xfs_btree_ptr *new, + int *stat) +{ + return __xfs_inobt_alloc_block(cur, start, new, stat, XFS_AG_RESV_NONE); +} + +STATIC int +xfs_finobt_alloc_block( + struct xfs_btree_cur *cur, + union xfs_btree_ptr *start, + union xfs_btree_ptr *new, + int *stat) +{ + return __xfs_inobt_alloc_block(cur, start, new, stat, + XFS_AG_RESV_METADATA); +} + STATIC int xfs_inobt_free_block( struct xfs_btree_cur *cur, @@ -328,7 +351,7 @@ static const struct xfs_btree_ops xfs_finobt_ops = { .dup_cursor = xfs_inobt_dup_cursor, .set_root = xfs_finobt_set_root, - .alloc_block = xfs_inobt_alloc_block, + .alloc_block = xfs_finobt_alloc_block, .free_block = xfs_inobt_free_block, .get_minrecs = xfs_inobt_get_minrecs, .get_maxrecs = xfs_inobt_get_maxrecs, @@ -480,3 +503,64 @@ xfs_inobt_rec_check_count( return 0; } #endif /* DEBUG */ + +static xfs_extlen_t +xfs_inobt_max_size( + struct xfs_mount *mp) +{ + /* Bail out if we're uninitialized, which can happen in mkfs. */ + if (mp->m_inobt_mxr[0] == 0) + return 0; + + return xfs_btree_calc_size(mp, mp->m_inobt_mnr, + (uint64_t)mp->m_sb.sb_agblocks * mp->m_sb.sb_inopblock / + XFS_INODES_PER_CHUNK); +} + +static int +xfs_inobt_count_blocks( + struct xfs_mount *mp, + xfs_agnumber_t agno, + xfs_btnum_t btnum, + xfs_extlen_t *tree_blocks) +{ + struct xfs_buf *agbp; + struct xfs_btree_cur *cur; + int error; + + error = xfs_ialloc_read_agi(mp, NULL, agno, &agbp); + if (error) + return error; + + cur = xfs_inobt_init_cursor(mp, NULL, agbp, agno, btnum); + error = xfs_btree_count_blocks(cur, tree_blocks); + xfs_btree_del_cursor(cur, error ? XFS_BTREE_ERROR : XFS_BTREE_NOERROR); + xfs_buf_relse(agbp); + + return error; +} + +/* + * Figure out how many blocks to reserve and how many are used by this btree. + */ +int +xfs_finobt_calc_reserves( + struct xfs_mount *mp, + xfs_agnumber_t agno, + xfs_extlen_t *ask, + xfs_extlen_t *used) +{ + xfs_extlen_t tree_len = 0; + int error; + + if (!xfs_sb_version_hasfinobt(&mp->m_sb)) + return 0; + + error = xfs_inobt_count_blocks(mp, agno, XFS_BTNUM_FINO, &tree_len); + if (error) + return error; + + *ask += xfs_inobt_max_size(mp); + *used += tree_len; + return 0; +} diff --git a/fs/xfs/libxfs/xfs_ialloc_btree.h b/fs/xfs/libxfs/xfs_ialloc_btree.h index bd88453217ce..aa81e2e63f3f 100644 --- a/fs/xfs/libxfs/xfs_ialloc_btree.h +++ b/fs/xfs/libxfs/xfs_ialloc_btree.h @@ -72,4 +72,7 @@ int xfs_inobt_rec_check_count(struct xfs_mount *, #define xfs_inobt_rec_check_count(mp, rec) 0 #endif /* DEBUG */ +int xfs_finobt_calc_reserves(struct xfs_mount *mp, xfs_agnumber_t agno, + xfs_extlen_t *ask, xfs_extlen_t *used); + #endif /* __XFS_IALLOC_BTREE_H__ */ diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index b9557795eb74..de32f0fe47c8 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -1792,22 +1792,23 @@ xfs_inactive_ifree( int error; /* - * The ifree transaction might need to allocate blocks for record - * insertion to the finobt. We don't want to fail here at ENOSPC, so - * allow ifree to dip into the reserved block pool if necessary. - * - * Freeing large sets of inodes generally means freeing inode chunks, - * directory and file data blocks, so this should be relatively safe. - * Only under severe circumstances should it be possible to free enough - * inodes to exhaust the reserve block pool via finobt expansion while - * at the same time not creating free space in the filesystem. + * We try to use a per-AG reservation for any block needed by the finobt + * tree, but as the finobt feature predates the per-AG reservation + * support a degraded file system might not have enough space for the + * reservation at mount time. In that case try to dip into the reserved + * pool and pray. * * Send a warning if the reservation does happen to fail, as the inode * now remains allocated and sits on the unlinked list until the fs is * repaired. */ - error = xfs_trans_alloc(mp, &M_RES(mp)->tr_ifree, - XFS_IFREE_SPACE_RES(mp), 0, XFS_TRANS_RESERVE, &tp); + if (unlikely(mp->m_inotbt_nores)) { + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_ifree, + XFS_IFREE_SPACE_RES(mp), 0, XFS_TRANS_RESERVE, + &tp); + } else { + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_ifree, 0, 0, 0, &tp); + } if (error) { if (error == -ENOSPC) { xfs_warn_ratelimited(mp, diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index 84f785218907..7f351f706b7a 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -140,6 +140,7 @@ typedef struct xfs_mount { int m_fixedfsid[2]; /* unchanged for life of FS */ uint m_dmevmask; /* DMI events for this FS */ __uint64_t m_flags; /* global mount flags */ + bool m_inotbt_nores; /* no per-AG finobt resv. */ int m_ialloc_inos; /* inodes in inode allocation */ int m_ialloc_blks; /* blocks in inode allocation */ int m_ialloc_min_blks;/* min blocks in sparse inode From 5a93790d4e2df73e30c965ec6e49be82fc3ccfce Mon Sep 17 00:00:00 2001 From: Brian Foster Date: Wed, 25 Jan 2017 07:53:43 -0800 Subject: [PATCH 277/343] xfs: remove racy hasattr check from attr ops xfs_attr_[get|remove]() have unlocked attribute fork checks to optimize away a lock cycle in cases where the fork does not exist or is otherwise empty. This check is not safe, however, because an attribute fork short form to extent format conversion includes a transient state that causes the xfs_inode_hasattr() check to fail. Specifically, xfs_attr_shortform_to_leaf() creates an empty extent format attribute fork and then adds the existing shortform attributes to it. This means that lookup of an existing xattr can spuriously return -ENOATTR when racing against a setxattr that causes the associated format conversion. This was originally reproduced by an untar on a particularly configured glusterfs volume, but can also be reproduced on demand with properly crafted xattr requests. The format conversion occurs under the exclusive ilock. xfs_attr_get() and xfs_attr_remove() already have the proper locking and checks further down in the functions to handle this situation correctly. Drop the unlocked checks to avoid the spurious failure and rely on the existing logic. Signed-off-by: Brian Foster Reviewed-by: Christoph Hellwig Reviewed-by: Darrick J. Wong Signed-off-by: Darrick J. Wong --- fs/xfs/libxfs/xfs_attr.c | 6 ------ 1 file changed, 6 deletions(-) diff --git a/fs/xfs/libxfs/xfs_attr.c b/fs/xfs/libxfs/xfs_attr.c index af1ecb19121e..6622d46ddec3 100644 --- a/fs/xfs/libxfs/xfs_attr.c +++ b/fs/xfs/libxfs/xfs_attr.c @@ -131,9 +131,6 @@ xfs_attr_get( if (XFS_FORCED_SHUTDOWN(ip->i_mount)) return -EIO; - if (!xfs_inode_hasattr(ip)) - return -ENOATTR; - error = xfs_attr_args_init(&args, ip, name, flags); if (error) return error; @@ -392,9 +389,6 @@ xfs_attr_remove( if (XFS_FORCED_SHUTDOWN(dp->i_mount)) return -EIO; - if (!xfs_inode_hasattr(dp)) - return -ENOATTR; - error = xfs_attr_args_init(&args, dp, name, flags); if (error) return error; From 493611ebd62673f39e2f52c2561182c558a21cb6 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 25 Jan 2017 08:59:43 -0800 Subject: [PATCH 278/343] xfs: extsize hints are not unlikely in xfs_bmap_btalloc With COW files they are the hotpath, just like for files with the extent size hint attribute. We really shouldn't micro-manage anything but failure cases with unlikely. Additionally Arnd Bergmann recently reported that one of these two unlikely annotations causes link failures together with an upcoming kernel instrumentation patch, so let's get rid of it ASAP. Signed-off-by: Christoph Hellwig Reported-by: Arnd Bergmann Reviewed-by: Darrick J. Wong Signed-off-by: Darrick J. Wong --- fs/xfs/libxfs/xfs_bmap.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c index ab82dd4a4980..bfc00de5c6f1 100644 --- a/fs/xfs/libxfs/xfs_bmap.c +++ b/fs/xfs/libxfs/xfs_bmap.c @@ -3629,7 +3629,7 @@ xfs_bmap_btalloc( align = xfs_get_cowextsz_hint(ap->ip); else if (xfs_alloc_is_userdata(ap->datatype)) align = xfs_get_extsz_hint(ap->ip); - if (unlikely(align)) { + if (align) { error = xfs_bmap_extsize_align(mp, &ap->got, &ap->prev, align, 0, ap->eof, 0, ap->conv, &ap->offset, &ap->length); @@ -3701,7 +3701,7 @@ xfs_bmap_btalloc( args.minlen = ap->minlen; } /* apply extent size hints if obtained earlier */ - if (unlikely(align)) { + if (align) { args.prod = align; if ((args.mod = (xfs_extlen_t)do_mod(ap->offset, args.prod))) args.mod = (xfs_extlen_t)(args.prod - args.mod); From 6f29a130613191d3c6335169febe002cba00edf5 Mon Sep 17 00:00:00 2001 From: Xin Long Date: Tue, 24 Jan 2017 14:01:53 +0800 Subject: [PATCH 279/343] sctp: sctp_addr_id2transport should verify the addr before looking up assoc sctp_addr_id2transport is a function for sockopt to look up assoc by address. As the address is from userspace, it can be a v4-mapped v6 address. But in sctp protocol stack, it always handles a v4-mapped v6 address as a v4 address. So it's necessary to convert it to a v4 address before looking up assoc by address. This patch is to fix it by calling sctp_verify_addr in which it can do this conversion before calling sctp_endpoint_lookup_assoc, just like what sctp_sendmsg and __sctp_connect do for the address from users. Signed-off-by: Xin Long Acked-by: Neil Horman Signed-off-by: David S. Miller --- net/sctp/socket.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/net/sctp/socket.c b/net/sctp/socket.c index 318c6786d653..37eeab7899fc 100644 --- a/net/sctp/socket.c +++ b/net/sctp/socket.c @@ -235,8 +235,12 @@ static struct sctp_transport *sctp_addr_id2transport(struct sock *sk, sctp_assoc_t id) { struct sctp_association *addr_asoc = NULL, *id_asoc = NULL; - struct sctp_transport *transport; + struct sctp_af *af = sctp_get_af_specific(addr->ss_family); union sctp_addr *laddr = (union sctp_addr *)addr; + struct sctp_transport *transport; + + if (sctp_verify_addr(sk, laddr, af->sockaddr_len)) + return NULL; addr_asoc = sctp_endpoint_lookup_assoc(sctp_sk(sk)->ep, laddr, From 5207f3996338e1db71363fe381c81aaf1e54e4e3 Mon Sep 17 00:00:00 2001 From: Xin Long Date: Tue, 24 Jan 2017 14:05:16 +0800 Subject: [PATCH 280/343] sctp: sctp gso should set feature with NETIF_F_SG when calling skb_segment Now sctp gso puts segments into skb's frag_list, then processes these segments in skb_segment. But skb_segment handles them only when gs is enabled, as it's in the same branch with skb's frags. Although almost all the NICs support sg other than some old ones, but since commit 1e16aa3ddf86 ("net: gso: use feature flag argument in all protocol gso handlers"), features &= skb->dev->hw_enc_features, and xfrm_output_gso call skb_segment with features = 0, which means sctp gso would call skb_segment with sg = 0, and skb_segment would not work as expected. This patch is to fix it by setting features param with NETIF_F_SG when calling skb_segment so that it can go the right branch to process the skb's frag_list. Signed-off-by: Xin Long Signed-off-by: David S. Miller --- net/sctp/offload.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/sctp/offload.c b/net/sctp/offload.c index 7e869d0cca69..4f5a2b580aa5 100644 --- a/net/sctp/offload.c +++ b/net/sctp/offload.c @@ -68,7 +68,7 @@ static struct sk_buff *sctp_gso_segment(struct sk_buff *skb, goto out; } - segs = skb_segment(skb, features | NETIF_F_HW_CSUM); + segs = skb_segment(skb, features | NETIF_F_HW_CSUM | NETIF_F_SG); if (IS_ERR(segs)) goto out; From 56d806222ace4c3aeae516cd7a855340fb2839d8 Mon Sep 17 00:00:00 2001 From: Jason Baron Date: Tue, 24 Jan 2017 21:49:41 -0500 Subject: [PATCH 281/343] tcp: correct memory barrier usage in tcp_check_space() sock_reset_flag() maps to __clear_bit() not the atomic version clear_bit(). Thus, we need smp_mb(), smp_mb__after_atomic() is not sufficient. Fixes: 3c7151275c0c ("tcp: add memory barriers to write space paths") Cc: Eric Dumazet Cc: Oleg Nesterov Signed-off-by: Jason Baron Acked-by: Eric Dumazet Reported-by: Oleg Nesterov Signed-off-by: David S. Miller --- net/ipv4/tcp_input.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 6c790754ae3e..41dcbd568cbe 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -5078,7 +5078,7 @@ static void tcp_check_space(struct sock *sk) if (sock_flag(sk, SOCK_QUEUE_SHRUNK)) { sock_reset_flag(sk, SOCK_QUEUE_SHRUNK); /* pairs with tcp_poll() */ - smp_mb__after_atomic(); + smp_mb(); if (sk->sk_socket && test_bit(SOCK_NOSPACE, &sk->sk_socket->flags)) { tcp_new_space(sk); From a551ee94ea723b4af9b827c7460f108bc13425ee Mon Sep 17 00:00:00 2001 From: Michael Chan Date: Wed, 25 Jan 2017 02:55:07 -0500 Subject: [PATCH 282/343] bnxt_en: Fix bnxt_reset() in the slow path task. In bnxt_sp_task(), we set a bit BNXT_STATE_IN_SP_TASK so that bnxt_close() will synchronize and wait for bnxt_sp_task() to finish. Some functions in bnxt_sp_task() require us to clear BNXT_STATE_IN_SP_TASK and then acquire rtnl_lock() to prevent race conditions. There are some bugs related to this logic. This patch refactors the code to have common bnxt_rtnl_lock_sp() and bnxt_rtnl_unlock_sp() to handle the RTNL and the clearing/setting of the bit. Multiple functions will need the same logic. We also need to move bnxt_reset() to the end of bnxt_sp_task(). Functions that clear BNXT_STATE_IN_SP_TASK must be the last functions to be called in bnxt_sp_task(). The common scheme will handle the condition properly. Signed-off-by: Michael Chan Signed-off-by: David S. Miller --- drivers/net/ethernet/broadcom/bnxt/bnxt.c | 38 +++++++++++++++-------- 1 file changed, 25 insertions(+), 13 deletions(-) diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c index 53e686fdf2ce..30d7d643bdbe 100644 --- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c @@ -6200,23 +6200,32 @@ bnxt_restart_timer: mod_timer(&bp->timer, jiffies + bp->current_interval); } -/* Only called from bnxt_sp_task() */ -static void bnxt_reset(struct bnxt *bp, bool silent) +static void bnxt_rtnl_lock_sp(struct bnxt *bp) { - /* bnxt_reset_task() calls bnxt_close_nic() which waits - * for BNXT_STATE_IN_SP_TASK to clear. - * If there is a parallel dev_close(), bnxt_close() may be holding + /* We are called from bnxt_sp_task which has BNXT_STATE_IN_SP_TASK + * set. If the device is being closed, bnxt_close() may be holding * rtnl() and waiting for BNXT_STATE_IN_SP_TASK to clear. So we * must clear BNXT_STATE_IN_SP_TASK before holding rtnl(). */ clear_bit(BNXT_STATE_IN_SP_TASK, &bp->state); rtnl_lock(); - if (test_bit(BNXT_STATE_OPEN, &bp->state)) - bnxt_reset_task(bp, silent); +} + +static void bnxt_rtnl_unlock_sp(struct bnxt *bp) +{ set_bit(BNXT_STATE_IN_SP_TASK, &bp->state); rtnl_unlock(); } +/* Only called from bnxt_sp_task() */ +static void bnxt_reset(struct bnxt *bp, bool silent) +{ + bnxt_rtnl_lock_sp(bp); + if (test_bit(BNXT_STATE_OPEN, &bp->state)) + bnxt_reset_task(bp, silent); + bnxt_rtnl_unlock_sp(bp); +} + static void bnxt_cfg_ntp_filters(struct bnxt *); static void bnxt_sp_task(struct work_struct *work) @@ -6266,18 +6275,21 @@ static void bnxt_sp_task(struct work_struct *work) bnxt_hwrm_tunnel_dst_port_free( bp, TUNNEL_DST_PORT_FREE_REQ_TUNNEL_TYPE_GENEVE); } - if (test_and_clear_bit(BNXT_RESET_TASK_SP_EVENT, &bp->sp_event)) - bnxt_reset(bp, false); - - if (test_and_clear_bit(BNXT_RESET_TASK_SILENT_SP_EVENT, &bp->sp_event)) - bnxt_reset(bp, true); - if (test_and_clear_bit(BNXT_HWRM_PORT_MODULE_SP_EVENT, &bp->sp_event)) bnxt_get_port_module_status(bp); if (test_and_clear_bit(BNXT_PERIODIC_STATS_SP_EVENT, &bp->sp_event)) bnxt_hwrm_port_qstats(bp); + /* These functions below will clear BNXT_STATE_IN_SP_TASK. They + * must be the last functions to be called before exiting. + */ + if (test_and_clear_bit(BNXT_RESET_TASK_SP_EVENT, &bp->sp_event)) + bnxt_reset(bp, false); + + if (test_and_clear_bit(BNXT_RESET_TASK_SILENT_SP_EVENT, &bp->sp_event)) + bnxt_reset(bp, true); + smp_mb__before_atomic(); clear_bit(BNXT_STATE_IN_SP_TASK, &bp->state); } From 0eaa24b971ae251ae9d3be23f77662a655532063 Mon Sep 17 00:00:00 2001 From: Michael Chan Date: Wed, 25 Jan 2017 02:55:08 -0500 Subject: [PATCH 283/343] bnxt_en: Fix RTNL lock usage on bnxt_update_link(). bnxt_update_link() is called from multiple code paths. Most callers, such as open, ethtool, already hold RTNL. Only the caller bnxt_sp_task() does not. So it is a bug to take RTNL inside bnxt_update_link(). Fix it by removing the RTNL inside bnxt_update_link(). The function now expects the caller to always hold RTNL. In bnxt_sp_task(), call bnxt_rtnl_lock_sp() before calling bnxt_update_link(). We also need to move the call to the end of bnxt_sp_task() since it will be clearing the BNXT_STATE_IN_SP_TASK bit. Signed-off-by: Michael Chan Signed-off-by: David S. Miller --- drivers/net/ethernet/broadcom/bnxt/bnxt.c | 37 +++++++++++------------ 1 file changed, 18 insertions(+), 19 deletions(-) diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c index 30d7d643bdbe..69b0fce4f5b3 100644 --- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c @@ -5314,17 +5314,12 @@ static int bnxt_update_link(struct bnxt *bp, bool chng_link_state) if ((link_info->support_auto_speeds | diff) != link_info->support_auto_speeds) { /* An advertised speed is no longer supported, so we need to - * update the advertisement settings. See bnxt_reset() for - * comments about the rtnl_lock() sequence below. + * update the advertisement settings. Caller holds RTNL + * so we can modify link settings. */ - clear_bit(BNXT_STATE_IN_SP_TASK, &bp->state); - rtnl_lock(); link_info->advertising = link_info->support_auto_speeds; - if (test_bit(BNXT_STATE_OPEN, &bp->state) && - (link_info->autoneg & BNXT_AUTONEG_SPEED)) + if (link_info->autoneg & BNXT_AUTONEG_SPEED) bnxt_hwrm_set_link_setting(bp, true, false); - set_bit(BNXT_STATE_IN_SP_TASK, &bp->state); - rtnl_unlock(); } return 0; } @@ -6231,7 +6226,6 @@ static void bnxt_cfg_ntp_filters(struct bnxt *); static void bnxt_sp_task(struct work_struct *work) { struct bnxt *bp = container_of(work, struct bnxt, sp_task); - int rc; set_bit(BNXT_STATE_IN_SP_TASK, &bp->state); smp_mb__after_atomic(); @@ -6245,16 +6239,6 @@ static void bnxt_sp_task(struct work_struct *work) if (test_and_clear_bit(BNXT_RX_NTP_FLTR_SP_EVENT, &bp->sp_event)) bnxt_cfg_ntp_filters(bp); - if (test_and_clear_bit(BNXT_LINK_CHNG_SP_EVENT, &bp->sp_event)) { - if (test_and_clear_bit(BNXT_LINK_SPEED_CHNG_SP_EVENT, - &bp->sp_event)) - bnxt_hwrm_phy_qcaps(bp); - - rc = bnxt_update_link(bp, true); - if (rc) - netdev_err(bp->dev, "SP task can't update link (rc: %x)\n", - rc); - } if (test_and_clear_bit(BNXT_HWRM_EXEC_FWD_REQ_SP_EVENT, &bp->sp_event)) bnxt_hwrm_exec_fwd_req(bp); if (test_and_clear_bit(BNXT_VXLAN_ADD_PORT_SP_EVENT, &bp->sp_event)) { @@ -6284,6 +6268,21 @@ static void bnxt_sp_task(struct work_struct *work) /* These functions below will clear BNXT_STATE_IN_SP_TASK. They * must be the last functions to be called before exiting. */ + if (test_and_clear_bit(BNXT_LINK_CHNG_SP_EVENT, &bp->sp_event)) { + int rc = 0; + + if (test_and_clear_bit(BNXT_LINK_SPEED_CHNG_SP_EVENT, + &bp->sp_event)) + bnxt_hwrm_phy_qcaps(bp); + + bnxt_rtnl_lock_sp(bp); + if (test_bit(BNXT_STATE_OPEN, &bp->state)) + rc = bnxt_update_link(bp, true); + bnxt_rtnl_unlock_sp(bp); + if (rc) + netdev_err(bp->dev, "SP task can't update link (rc: %x)\n", + rc); + } if (test_and_clear_bit(BNXT_RESET_TASK_SP_EVENT, &bp->sp_event)) bnxt_reset(bp, false); From 90c694bb71819fb5bd3501ac397307d7e41ddeca Mon Sep 17 00:00:00 2001 From: Michael Chan Date: Wed, 25 Jan 2017 02:55:09 -0500 Subject: [PATCH 284/343] bnxt_en: Fix RTNL lock usage on bnxt_get_port_module_status(). bnxt_get_port_module_status() calls bnxt_update_link() which expects RTNL to be held. In bnxt_sp_task() that does not hold RTNL, we need to call it with a prior call to bnxt_rtnl_lock_sp() and the call needs to be moved to the end of bnxt_sp_task(). Signed-off-by: Michael Chan Signed-off-by: David S. Miller --- drivers/net/ethernet/broadcom/bnxt/bnxt.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c index 69b0fce4f5b3..4fcc6a84a087 100644 --- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c @@ -6259,9 +6259,6 @@ static void bnxt_sp_task(struct work_struct *work) bnxt_hwrm_tunnel_dst_port_free( bp, TUNNEL_DST_PORT_FREE_REQ_TUNNEL_TYPE_GENEVE); } - if (test_and_clear_bit(BNXT_HWRM_PORT_MODULE_SP_EVENT, &bp->sp_event)) - bnxt_get_port_module_status(bp); - if (test_and_clear_bit(BNXT_PERIODIC_STATS_SP_EVENT, &bp->sp_event)) bnxt_hwrm_port_qstats(bp); @@ -6283,6 +6280,12 @@ static void bnxt_sp_task(struct work_struct *work) netdev_err(bp->dev, "SP task can't update link (rc: %x)\n", rc); } + if (test_and_clear_bit(BNXT_HWRM_PORT_MODULE_SP_EVENT, &bp->sp_event)) { + bnxt_rtnl_lock_sp(bp); + if (test_bit(BNXT_STATE_OPEN, &bp->state)) + bnxt_get_port_module_status(bp); + bnxt_rtnl_unlock_sp(bp); + } if (test_and_clear_bit(BNXT_RESET_TASK_SP_EVENT, &bp->sp_event)) bnxt_reset(bp, false); From 61976fff20f92aceecc3670f6168bfc57a79e047 Mon Sep 17 00:00:00 2001 From: John Crispin Date: Wed, 25 Jan 2017 09:20:54 +0100 Subject: [PATCH 285/343] Documentation: devicetree: change the mediatek ethernet compatible string When the binding was defined, I was not aware that mt2701 was an earlier version of the SoC. For sake of consistency, the ethernet driver should use mt2701 inside the compat string as this is the earliest SoC with the ethernet core. The ethernet driver is currently of no real use until we finish and upstream the DSA driver. There are no users of this binding yet. It should be safe to fix this now before it is too late and we need to provide backward compatibility for the mt7623-eth compat string. Reported-by: Sean Wang Signed-off-by: John Crispin Reviewed-by: Matthias Brugger Signed-off-by: David S. Miller --- Documentation/devicetree/bindings/net/mediatek-net.txt | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Documentation/devicetree/bindings/net/mediatek-net.txt b/Documentation/devicetree/bindings/net/mediatek-net.txt index c010fafc66a8..c7194e87d5f4 100644 --- a/Documentation/devicetree/bindings/net/mediatek-net.txt +++ b/Documentation/devicetree/bindings/net/mediatek-net.txt @@ -7,7 +7,7 @@ have dual GMAC each represented by a child node.. * Ethernet controller node Required properties: -- compatible: Should be "mediatek,mt7623-eth" +- compatible: Should be "mediatek,mt2701-eth" - reg: Address and length of the register set for the device - interrupts: Should contain the three frame engines interrupts in numeric order. These are fe_int0, fe_int1 and fe_int2. From 8b901f6bbcf12a20e43105d161bedde093431e61 Mon Sep 17 00:00:00 2001 From: John Crispin Date: Wed, 25 Jan 2017 09:20:55 +0100 Subject: [PATCH 286/343] net-next: ethernet: mediatek: change the compatible string When the binding was defined, I was not aware that mt2701 was an earlier version of the SoC. For sake of consistency, the ethernet driver should use mt2701 inside the compat string as this is the earliest SoC with the ethernet core. The ethernet driver is currently of no real use until we finish and upstream the DSA driver. There are no users of this binding yet. It should be safe to fix this now before it is too late and we need to provide backward compatibility for the mt7623-eth compat string. Reported-by: Sean Wang Signed-off-by: John Crispin Signed-off-by: David S. Miller --- drivers/net/ethernet/mediatek/mtk_eth_soc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/mediatek/mtk_eth_soc.c b/drivers/net/ethernet/mediatek/mtk_eth_soc.c index 3dd87889e67e..1c29c86f8709 100644 --- a/drivers/net/ethernet/mediatek/mtk_eth_soc.c +++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.c @@ -2517,7 +2517,7 @@ static int mtk_remove(struct platform_device *pdev) } const struct of_device_id of_mtk_match[] = { - { .compatible = "mediatek,mt7623-eth" }, + { .compatible = "mediatek,mt2701-eth" }, {}, }; MODULE_DEVICE_TABLE(of, of_mtk_match); From 8a87fca8dd5879eb05a0903cb7ea4fd2a3876ae0 Mon Sep 17 00:00:00 2001 From: Geert Uytterhoeven Date: Wed, 25 Jan 2017 11:39:48 +0100 Subject: [PATCH 287/343] net: phy: leds: Clear phy_num_led_triggers on failure to avoid crash phy_attach_direct() ignores errors returned by phy_led_triggers_register(). I think that's OK, as LED triggers can be considered a non-critical feature. However, this causes problems later: - phy_led_trigger_change_speed() will access the array phy_device.phy_led_triggers, which has been freed in the error path of phy_led_triggers_register(), which may lead to a crash. - phy_led_triggers_unregister() will access the same array, leading to crashes during s2ram or poweroff, like: Unable to handle kernel NULL pointer dereference at virtual address 00000000 ... [] (__list_del_entry_valid) from [] (led_trigger_unregister+0x34/0xcc) [] (led_trigger_unregister) from [] (phy_led_triggers_unregister+0x28/0x34) [] (phy_led_triggers_unregister) from [] (phy_detach+0x30/0x74) [] (phy_detach) from [] (sh_eth_close+0x64/0x9c) [] (sh_eth_close) from [] (dpm_run_callback+0x48/0xc8) or: list_del corruption. prev->next should be dede6540, but was 2e323931 ------------[ cut here ]------------ kernel BUG at lib/list_debug.c:52! ... [] (__list_del_entry_valid) from [] (led_trigger_unregister+0x34/0xcc) [] (led_trigger_unregister) from [] (phy_led_triggers_unregister+0x28/0x34) [] (phy_led_triggers_unregister) from [] (phy_detach+0x30/0x74) [] (phy_detach) from [] (sh_eth_close+0x6c/0xa4) [] (sh_eth_close) from [] (__dev_close_many+0xac/0xd0) To fix this, clear phy_device.phy_num_led_triggers in the error path of phy_led_triggers_register() fails. Note that the "No phy led trigger registered for speed" message will still be printed on link speed changes, which is a good cue that something went wrong with the LED triggers. Fixes: 2e0bc452f4721520 ("net: phy: leds: add support for led triggers on phy link state change") Signed-off-by: Geert Uytterhoeven Reviewed-by: Florian Fainelli Signed-off-by: David S. Miller --- drivers/net/phy/phy_led_triggers.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/drivers/net/phy/phy_led_triggers.c b/drivers/net/phy/phy_led_triggers.c index fa62bdf2f526..3f619e7371e9 100644 --- a/drivers/net/phy/phy_led_triggers.c +++ b/drivers/net/phy/phy_led_triggers.c @@ -102,8 +102,10 @@ int phy_led_triggers_register(struct phy_device *phy) sizeof(struct phy_led_trigger) * phy->phy_num_led_triggers, GFP_KERNEL); - if (!phy->phy_led_triggers) - return -ENOMEM; + if (!phy->phy_led_triggers) { + err = -ENOMEM; + goto out_clear; + } for (i = 0; i < phy->phy_num_led_triggers; i++) { err = phy_led_trigger_register(phy, &phy->phy_led_triggers[i], @@ -120,6 +122,8 @@ out_unreg: while (i--) phy_led_trigger_unregister(&phy->phy_led_triggers[i]); devm_kfree(&phy->mdio.dev, phy->phy_led_triggers); +out_clear: + phy->phy_num_led_triggers = 0; return err; } EXPORT_SYMBOL_GPL(phy_led_triggers_register); From d6f8cfa3dea294eabf8f302e90176dd6381fb66e Mon Sep 17 00:00:00 2001 From: Geert Uytterhoeven Date: Wed, 25 Jan 2017 11:39:49 +0100 Subject: [PATCH 288/343] net: phy: leds: Break dependency of phy.h on phy_led_triggers.h includes , which is not really needed. Drop the include from , and add it to all users that didn't include it explicitly. Suggested-by: Andrew Lunn Signed-off-by: Geert Uytterhoeven Reviewed-by: Andrew Lunn Signed-off-by: David S. Miller --- drivers/net/phy/phy.c | 1 + drivers/net/phy/phy_led_triggers.c | 1 + include/linux/phy.h | 1 - 3 files changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/net/phy/phy.c b/drivers/net/phy/phy.c index e687a9cb4a37..7cc1b7dcfe05 100644 --- a/drivers/net/phy/phy.c +++ b/drivers/net/phy/phy.c @@ -29,6 +29,7 @@ #include #include #include +#include #include #include #include diff --git a/drivers/net/phy/phy_led_triggers.c b/drivers/net/phy/phy_led_triggers.c index 3f619e7371e9..94ca42e630bb 100644 --- a/drivers/net/phy/phy_led_triggers.c +++ b/drivers/net/phy/phy_led_triggers.c @@ -12,6 +12,7 @@ */ #include #include +#include #include static struct phy_led_trigger *phy_speed_to_led_trigger(struct phy_device *phy, diff --git a/include/linux/phy.h b/include/linux/phy.h index f7d95f644eed..7fc1105605bf 100644 --- a/include/linux/phy.h +++ b/include/linux/phy.h @@ -25,7 +25,6 @@ #include #include #include -#include #include From 3c880eb0205222bb062970085ebedc73ec8dfd14 Mon Sep 17 00:00:00 2001 From: Geert Uytterhoeven Date: Wed, 25 Jan 2017 11:39:50 +0100 Subject: [PATCH 289/343] net: phy: leds: Fix truncated LED trigger names Commit 4567d686f5c6d955 ("phy: increase size of MII_BUS_ID_SIZE and bus_id") increased the size of MII bus IDs, but forgot to update the private definition in . This may cause: 1. Truncation of LED trigger names, 2. Duplicate LED trigger names, 3. Failures registering LED triggers, 4. Crashes due to bad error handling in the LED trigger failure path. To fix this, and prevent the definitions going out of sync again in the future, let the PHY LED trigger code use the existing MII_BUS_ID_SIZE definition. Example: - Before I had triggers "ee700000.etherne:01:100Mbps" and "ee700000.etherne:01:10Mbps", - After the increase of MII_BUS_ID_SIZE, both became "ee700000.ethernet-ffffffff:01:" => FAIL, - Now, the triggers are "ee700000.ethernet-ffffffff:01:100Mbps" and "ee700000.ethernet-ffffffff:01:10Mbps", which are unique again. Fixes: 4567d686f5c6d955 ("phy: increase size of MII_BUS_ID_SIZE and bus_id") Fixes: 2e0bc452f4721520 ("net: phy: leds: add support for led triggers on phy link state change") Signed-off-by: Geert Uytterhoeven Reviewed-by: Andrew Lunn Signed-off-by: David S. Miller --- include/linux/phy_led_triggers.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/include/linux/phy_led_triggers.h b/include/linux/phy_led_triggers.h index a2daea0a37d2..b37b05bfd1a6 100644 --- a/include/linux/phy_led_triggers.h +++ b/include/linux/phy_led_triggers.h @@ -18,11 +18,11 @@ struct phy_device; #ifdef CONFIG_LED_TRIGGER_PHY #include +#include #define PHY_LED_TRIGGER_SPEED_SUFFIX_SIZE 10 -#define PHY_MII_BUS_ID_SIZE (20 - 3) -#define PHY_LINK_LED_TRIGGER_NAME_SIZE (PHY_MII_BUS_ID_SIZE + \ +#define PHY_LINK_LED_TRIGGER_NAME_SIZE (MII_BUS_ID_SIZE + \ FIELD_SIZEOF(struct mdio_device, addr)+\ PHY_LED_TRIGGER_SPEED_SUFFIX_SIZE) From f154be241d22298d2b63c9b613f619fa1086ea75 Mon Sep 17 00:00:00 2001 From: Florian Fainelli Date: Wed, 25 Jan 2017 09:10:41 -0800 Subject: [PATCH 290/343] net: dsa: Bring back device detaching in dsa_slave_suspend() Commit 448b4482c671 ("net: dsa: Add lockdep class to tx queues to avoid lockdep splat") removed the netif_device_detach() call done in dsa_slave_suspend() which is necessary, and paired with a corresponding netif_device_attach(), bring it back. Fixes: 448b4482c671 ("net: dsa: Add lockdep class to tx queues to avoid lockdep splat") Signed-off-by: Florian Fainelli Reviewed-by: Andrew Lunn Signed-off-by: David S. Miller --- net/dsa/slave.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/net/dsa/slave.c b/net/dsa/slave.c index ba1b6b9630d2..7d4596110851 100644 --- a/net/dsa/slave.c +++ b/net/dsa/slave.c @@ -1201,6 +1201,8 @@ int dsa_slave_suspend(struct net_device *slave_dev) { struct dsa_slave_priv *p = netdev_priv(slave_dev); + netif_device_detach(slave_dev); + if (p->phy) { phy_stop(p->phy); p->old_pause = -1; From 54a07c7bb0da0343734c78212bbe9f3735394962 Mon Sep 17 00:00:00 2001 From: Dave Airlie Date: Thu, 26 Jan 2017 06:44:03 +1000 Subject: [PATCH 291/343] Revert "drm/probe-helpers: Drop locking from poll_enable" This reverts commit 3846fd9b86001bea171943cc3bb9222cb6da6b42. There were some precursor commits missing for this around connector locking, we should probably merge Lyude's nouveau avoid the problem patch. --- drivers/gpu/drm/drm_probe_helper.c | 51 +++++++++++++++++----------- drivers/gpu/drm/i915/intel_hotplug.c | 4 +-- include/drm/drm_crtc_helper.h | 1 + 3 files changed, 34 insertions(+), 22 deletions(-) diff --git a/drivers/gpu/drm/drm_probe_helper.c b/drivers/gpu/drm/drm_probe_helper.c index b452a7ccd84b..cf8f0128c161 100644 --- a/drivers/gpu/drm/drm_probe_helper.c +++ b/drivers/gpu/drm/drm_probe_helper.c @@ -115,27 +115,24 @@ static int drm_helper_probe_add_cmdline_mode(struct drm_connector *connector) #define DRM_OUTPUT_POLL_PERIOD (10*HZ) /** - * drm_kms_helper_poll_enable - re-enable output polling. + * drm_kms_helper_poll_enable_locked - re-enable output polling. * @dev: drm_device * - * This function re-enables the output polling work, after it has been - * temporarily disabled using drm_kms_helper_poll_disable(), for example over - * suspend/resume. + * This function re-enables the output polling work without + * locking the mode_config mutex. * - * Drivers can call this helper from their device resume implementation. It is - * an error to call this when the output polling support has not yet been set - * up. - * - * Note that calls to enable and disable polling must be strictly ordered, which - * is automatically the case when they're only call from suspend/resume - * callbacks. + * This is like drm_kms_helper_poll_enable() however it is to be + * called from a context where the mode_config mutex is locked + * already. */ -void drm_kms_helper_poll_enable(struct drm_device *dev) +void drm_kms_helper_poll_enable_locked(struct drm_device *dev) { bool poll = false; struct drm_connector *connector; unsigned long delay = DRM_OUTPUT_POLL_PERIOD; + WARN_ON(!mutex_is_locked(&dev->mode_config.mutex)); + if (!dev->mode_config.poll_enabled || !drm_kms_helper_poll) return; @@ -163,7 +160,7 @@ void drm_kms_helper_poll_enable(struct drm_device *dev) if (poll) schedule_delayed_work(&dev->mode_config.output_poll_work, delay); } -EXPORT_SYMBOL(drm_kms_helper_poll_enable); +EXPORT_SYMBOL(drm_kms_helper_poll_enable_locked); static enum drm_connector_status drm_connector_detect(struct drm_connector *connector, bool force) @@ -290,7 +287,7 @@ int drm_helper_probe_single_connector_modes(struct drm_connector *connector, /* Re-enable polling in case the global poll config changed. */ if (drm_kms_helper_poll != dev->mode_config.poll_running) - drm_kms_helper_poll_enable(dev); + drm_kms_helper_poll_enable_locked(dev); dev->mode_config.poll_running = drm_kms_helper_poll; @@ -482,12 +479,8 @@ out: * This function disables the output polling work. * * Drivers can call this helper from their device suspend implementation. It is - * not an error to call this even when output polling isn't enabled or already - * disabled. Polling is re-enabled by calling drm_kms_helper_poll_enable(). - * - * Note that calls to enable and disable polling must be strictly ordered, which - * is automatically the case when they're only call from suspend/resume - * callbacks. + * not an error to call this even when output polling isn't enabled or arlready + * disabled. */ void drm_kms_helper_poll_disable(struct drm_device *dev) { @@ -497,6 +490,24 @@ void drm_kms_helper_poll_disable(struct drm_device *dev) } EXPORT_SYMBOL(drm_kms_helper_poll_disable); +/** + * drm_kms_helper_poll_enable - re-enable output polling. + * @dev: drm_device + * + * This function re-enables the output polling work. + * + * Drivers can call this helper from their device resume implementation. It is + * an error to call this when the output polling support has not yet been set + * up. + */ +void drm_kms_helper_poll_enable(struct drm_device *dev) +{ + mutex_lock(&dev->mode_config.mutex); + drm_kms_helper_poll_enable_locked(dev); + mutex_unlock(&dev->mode_config.mutex); +} +EXPORT_SYMBOL(drm_kms_helper_poll_enable); + /** * drm_kms_helper_poll_init - initialize and enable output polling * @dev: drm_device diff --git a/drivers/gpu/drm/i915/intel_hotplug.c b/drivers/gpu/drm/i915/intel_hotplug.c index b62e3f8ad415..3d546c019de0 100644 --- a/drivers/gpu/drm/i915/intel_hotplug.c +++ b/drivers/gpu/drm/i915/intel_hotplug.c @@ -180,7 +180,7 @@ static void intel_hpd_irq_storm_disable(struct drm_i915_private *dev_priv) /* Enable polling and queue hotplug re-enabling. */ if (hpd_disabled) { - drm_kms_helper_poll_enable(dev); + drm_kms_helper_poll_enable_locked(dev); mod_delayed_work(system_wq, &dev_priv->hotplug.reenable_work, msecs_to_jiffies(HPD_STORM_REENABLE_DELAY)); } @@ -511,7 +511,7 @@ static void i915_hpd_poll_init_work(struct work_struct *work) } if (enabled) - drm_kms_helper_poll_enable(dev); + drm_kms_helper_poll_enable_locked(dev); mutex_unlock(&dev->mode_config.mutex); diff --git a/include/drm/drm_crtc_helper.h b/include/drm/drm_crtc_helper.h index d026f5017c33..982c299e435a 100644 --- a/include/drm/drm_crtc_helper.h +++ b/include/drm/drm_crtc_helper.h @@ -73,5 +73,6 @@ extern void drm_kms_helper_hotplug_event(struct drm_device *dev); extern void drm_kms_helper_poll_disable(struct drm_device *dev); extern void drm_kms_helper_poll_enable(struct drm_device *dev); +extern void drm_kms_helper_poll_enable_locked(struct drm_device *dev); #endif From 0e1929dedea36781e25902118c93edd8d8f09af1 Mon Sep 17 00:00:00 2001 From: Mike Looijmans Date: Mon, 16 Jan 2017 15:49:38 +0100 Subject: [PATCH 292/343] i2c: i2c-cadence: Initialize configuration before probing devices The cadence I2C driver calls cdns_i2c_writereg(..) to setup a workaround in the controller, but did so after calling i2c_add_adapter() which starts probing devices on the bus. Change the order so that the configuration is completely finished before using the adapter. Signed-off-by: Mike Looijmans Signed-off-by: Wolfram Sang --- drivers/i2c/busses/i2c-cadence.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/i2c/busses/i2c-cadence.c b/drivers/i2c/busses/i2c-cadence.c index 686971263bef..45d6771fac8c 100644 --- a/drivers/i2c/busses/i2c-cadence.c +++ b/drivers/i2c/busses/i2c-cadence.c @@ -962,10 +962,6 @@ static int cdns_i2c_probe(struct platform_device *pdev) goto err_clk_dis; } - ret = i2c_add_adapter(&id->adap); - if (ret < 0) - goto err_clk_dis; - /* * Cadence I2C controller has a bug wherein it generates * invalid read transaction after HW timeout in master receiver mode. @@ -975,6 +971,10 @@ static int cdns_i2c_probe(struct platform_device *pdev) */ cdns_i2c_writereg(CDNS_I2C_TIMEOUT_MAX, CDNS_I2C_TIME_OUT_OFFSET); + ret = i2c_add_adapter(&id->adap); + if (ret < 0) + goto err_clk_dis; + dev_info(&pdev->dev, "%u kHz mmio %08lx irq %d\n", id->i2c_clk / 1000, (unsigned long)r_mem->start, id->irq); From b9b487e494712c8e5905b724e12f5ef17e9ae6f9 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Wed, 25 Jan 2017 12:00:29 -0500 Subject: [PATCH 293/343] Revert "drm/radeon: always apply pci shutdown callbacks" This seems to break reboot on some evergreen systems. bugs: https://bugs.freedesktop.org/show_bug.cgi?id=99524 https://bugzilla.kernel.org/show_bug.cgi?id=192271 This reverts commit a481daa88fd4d6b54f25348972bba10b5f6a84d0. Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org --- drivers/gpu/drm/radeon/radeon_drv.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c index 00ea0002b539..e0c143b865f3 100644 --- a/drivers/gpu/drm/radeon/radeon_drv.c +++ b/drivers/gpu/drm/radeon/radeon_drv.c @@ -366,11 +366,10 @@ static void radeon_pci_shutdown(struct pci_dev *pdev) { /* if we are running in a VM, make sure the device - * torn down properly on reboot/shutdown. - * unfortunately we can't detect certain - * hypervisors so just do this all the time. + * torn down properly on reboot/shutdown */ - radeon_pci_remove(pdev); + if (radeon_device_is_virtual()) + radeon_pci_remove(pdev); } static int radeon_pmops_suspend(struct device *dev) From e13fe92bb58cf9b8f709ec18267ffc9e6ffeb016 Mon Sep 17 00:00:00 2001 From: Gao Pan Date: Tue, 17 Jan 2017 18:20:55 +0800 Subject: [PATCH 294/343] i2c: imx-lpi2c: add VLLS mode support When system enters VLLS mode, module power is turned off. As a result, all registers are reset to HW default value. After exiting VLLS mode, registers are still in default mode. As a result, the pinctrl settings are incorrect, which will affect the module function. The patch recovers the pinctrl setting when exit VLLS mode. Signed-off-by: Gao Pan Reviewed-by: Vladimir Zapolskiy [wsa: added missing include] Signed-off-by: Wolfram Sang --- drivers/i2c/busses/i2c-imx-lpi2c.c | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+) diff --git a/drivers/i2c/busses/i2c-imx-lpi2c.c b/drivers/i2c/busses/i2c-imx-lpi2c.c index c62b7cd475f8..3310f2e0dbd3 100644 --- a/drivers/i2c/busses/i2c-imx-lpi2c.c +++ b/drivers/i2c/busses/i2c-imx-lpi2c.c @@ -28,6 +28,7 @@ #include #include #include +#include #include #include #include @@ -636,12 +637,31 @@ static int lpi2c_imx_remove(struct platform_device *pdev) return 0; } +#ifdef CONFIG_PM_SLEEP +static int lpi2c_imx_suspend(struct device *dev) +{ + pinctrl_pm_select_sleep_state(dev); + + return 0; +} + +static int lpi2c_imx_resume(struct device *dev) +{ + pinctrl_pm_select_default_state(dev); + + return 0; +} +#endif + +static SIMPLE_DEV_PM_OPS(imx_lpi2c_pm, lpi2c_imx_suspend, lpi2c_imx_resume); + static struct platform_driver lpi2c_imx_driver = { .probe = lpi2c_imx_probe, .remove = lpi2c_imx_remove, .driver = { .name = DRIVER_NAME, .of_match_table = lpi2c_imx_of_match, + .pm = &imx_lpi2c_pm, }, }; From 26afec39306926654e9cd320f19bbf3685bb0997 Mon Sep 17 00:00:00 2001 From: hayeswang Date: Thu, 26 Jan 2017 09:38:31 +0800 Subject: [PATCH 295/343] r8152: avoid start_xmit to call napi_schedule during autosuspend Adjust the setting of the flag of SELECTIVE_SUSPEND to prevent start_xmit() from calling napi_schedule() directly during runtime suspend. After calling napi_disable() or clearing the flag of WORK_ENABLE, scheduling the napi is useless. Signed-off-by: Hayes Wang Signed-off-by: David S. Miller --- drivers/net/usb/r8152.c | 12 +++++++++--- 1 file changed, 9 insertions(+), 3 deletions(-) diff --git a/drivers/net/usb/r8152.c b/drivers/net/usb/r8152.c index e1466b4d2b6c..23bef8e6e0bd 100644 --- a/drivers/net/usb/r8152.c +++ b/drivers/net/usb/r8152.c @@ -3585,10 +3585,15 @@ static int rtl8152_rumtime_suspend(struct r8152 *tp) struct net_device *netdev = tp->netdev; int ret = 0; + set_bit(SELECTIVE_SUSPEND, &tp->flags); + smp_mb__after_atomic(); + if (netif_running(netdev) && test_bit(WORK_ENABLE, &tp->flags)) { u32 rcr = 0; if (delay_autosuspend(tp)) { + clear_bit(SELECTIVE_SUSPEND, &tp->flags); + smp_mb__after_atomic(); ret = -EBUSY; goto out1; } @@ -3605,6 +3610,8 @@ static int rtl8152_rumtime_suspend(struct r8152 *tp) if (!(ocp_data & RXFIFO_EMPTY)) { rxdy_gated_en(tp, false); ocp_write_dword(tp, MCU_TYPE_PLA, PLA_RCR, rcr); + clear_bit(SELECTIVE_SUSPEND, &tp->flags); + smp_mb__after_atomic(); ret = -EBUSY; goto out1; } @@ -3624,8 +3631,6 @@ static int rtl8152_rumtime_suspend(struct r8152 *tp) } } - set_bit(SELECTIVE_SUSPEND, &tp->flags); - out1: return ret; } @@ -3681,12 +3686,13 @@ static int rtl8152_resume(struct usb_interface *intf) if (netif_running(tp->netdev) && tp->netdev->flags & IFF_UP) { if (test_bit(SELECTIVE_SUSPEND, &tp->flags)) { tp->rtl_ops.autosuspend_en(tp, false); - clear_bit(SELECTIVE_SUSPEND, &tp->flags); napi_disable(&tp->napi); set_bit(WORK_ENABLE, &tp->flags); if (netif_carrier_ok(tp->netdev)) rtl_start_rx(tp); napi_enable(&tp->napi); + clear_bit(SELECTIVE_SUSPEND, &tp->flags); + smp_mb__after_atomic(); } else { tp->rtl_ops.up(tp); netif_carrier_off(tp->netdev); From de9bf29dd6e4a8a874cb92f8901aed50a9d0b1d3 Mon Sep 17 00:00:00 2001 From: hayeswang Date: Thu, 26 Jan 2017 09:38:32 +0800 Subject: [PATCH 296/343] r8152: avoid start_xmit to schedule napi when napi is disabled Stop the tx when the napi is disabled to prevent napi_schedule() is called. Signed-off-by: Hayes Wang Signed-off-by: David S. Miller --- drivers/net/usb/r8152.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/drivers/net/usb/r8152.c b/drivers/net/usb/r8152.c index 23bef8e6e0bd..ec882be4cfaf 100644 --- a/drivers/net/usb/r8152.c +++ b/drivers/net/usb/r8152.c @@ -3155,10 +3155,13 @@ static void set_carrier(struct r8152 *tp) if (!netif_carrier_ok(netdev)) { tp->rtl_ops.enable(tp); set_bit(RTL8152_SET_RX_MODE, &tp->flags); + netif_stop_queue(netdev); napi_disable(&tp->napi); netif_carrier_on(netdev); rtl_start_rx(tp); napi_enable(&tp->napi); + netif_wake_queue(netdev); + netif_info(tp, link, netdev, "carrier on\n"); } } else { if (netif_carrier_ok(netdev)) { @@ -3166,6 +3169,7 @@ static void set_carrier(struct r8152 *tp) napi_disable(&tp->napi); tp->rtl_ops.disable(tp); napi_enable(&tp->napi); + netif_info(tp, link, netdev, "carrier off\n"); } } } @@ -3515,12 +3519,12 @@ static int rtl8152_pre_reset(struct usb_interface *intf) if (!netif_running(netdev)) return 0; + netif_stop_queue(netdev); napi_disable(&tp->napi); clear_bit(WORK_ENABLE, &tp->flags); usb_kill_urb(tp->intr_urb); cancel_delayed_work_sync(&tp->schedule); if (netif_carrier_ok(netdev)) { - netif_stop_queue(netdev); mutex_lock(&tp->control); tp->rtl_ops.disable(tp); mutex_unlock(&tp->control); @@ -3548,10 +3552,10 @@ static int rtl8152_post_reset(struct usb_interface *intf) rtl_start_rx(tp); rtl8152_set_rx_mode(netdev); mutex_unlock(&tp->control); - netif_wake_queue(netdev); } napi_enable(&tp->napi); + netif_wake_queue(netdev); usb_submit_urb(tp->intr_urb, GFP_KERNEL); return 0; From 248b213ad908b88db15941202ef7cb7eb137c1a0 Mon Sep 17 00:00:00 2001 From: hayeswang Date: Thu, 26 Jan 2017 09:38:33 +0800 Subject: [PATCH 297/343] r8152: re-schedule napi for tx Re-schedule napi after napi_complete() for tx, if it is necessay. In r8152_poll(), if the tx is completed after tx_bottom() and before napi_complete(), the scheduling of napi would be lost. Then, no one handles the next tx until the next napi_schedule() is called. Signed-off-by: Hayes Wang Signed-off-by: David S. Miller --- drivers/net/usb/r8152.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/net/usb/r8152.c b/drivers/net/usb/r8152.c index ec882be4cfaf..4785d2b9d80f 100644 --- a/drivers/net/usb/r8152.c +++ b/drivers/net/usb/r8152.c @@ -1936,6 +1936,9 @@ static int r8152_poll(struct napi_struct *napi, int budget) napi_complete(napi); if (!list_empty(&tp->rx_done)) napi_schedule(napi); + else if (!skb_queue_empty(&tp->tx_queue) && + !list_empty(&tp->tx_free)) + napi_schedule(napi); } return work_done; From 7489bdadb7d17d3c81e39b85688500f700beb790 Mon Sep 17 00:00:00 2001 From: hayeswang Date: Thu, 26 Jan 2017 09:38:34 +0800 Subject: [PATCH 298/343] r8152: check rx after napi is enabled Schedule the napi after napi_enable() for rx, if it is necessary. If the rx is completed when napi is disabled, the sheduling of napi would be lost. Then, no one handles the rx packet until next napi is scheduled. Signed-off-by: Hayes Wang Signed-off-by: David S. Miller --- drivers/net/usb/r8152.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/drivers/net/usb/r8152.c b/drivers/net/usb/r8152.c index 4785d2b9d80f..ad42295356dd 100644 --- a/drivers/net/usb/r8152.c +++ b/drivers/net/usb/r8152.c @@ -32,7 +32,7 @@ #define NETNEXT_VERSION "08" /* Information for net */ -#define NET_VERSION "7" +#define NET_VERSION "8" #define DRIVER_VERSION "v1." NETNEXT_VERSION "." NET_VERSION #define DRIVER_AUTHOR "Realtek linux nic maintainers " @@ -3561,6 +3561,9 @@ static int rtl8152_post_reset(struct usb_interface *intf) netif_wake_queue(netdev); usb_submit_urb(tp->intr_urb, GFP_KERNEL); + if (!list_empty(&tp->rx_done)) + napi_schedule(&tp->napi); + return 0; } @@ -3700,6 +3703,8 @@ static int rtl8152_resume(struct usb_interface *intf) napi_enable(&tp->napi); clear_bit(SELECTIVE_SUSPEND, &tp->flags); smp_mb__after_atomic(); + if (!list_empty(&tp->rx_done)) + napi_schedule(&tp->napi); } else { tp->rtl_ops.up(tp); netif_carrier_off(tp->netdev); From b68df015609eac67f045c155cb3195e5a1061d66 Mon Sep 17 00:00:00 2001 From: John Fastabend Date: Wed, 25 Jan 2017 18:22:48 -0800 Subject: [PATCH 299/343] virtio_net: use dev_kfree_skb for small buffer XDP receive In the small buffer case during driver unload we currently use put_page instead of dev_kfree_skb. Resolve this by adding a check for virtnet mode when checking XDP queue type. Also name the function so that the code reads correctly to match the additional check. Fixes: bb91accf2733 ("virtio-net: XDP support for small buffers") Signed-off-by: John Fastabend Acked-by: Jason Wang Acked-by: Michael S. Tsirkin Signed-off-by: David S. Miller --- drivers/net/virtio_net.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c index 3d1519ea0669..4c6e9b43f1d3 100644 --- a/drivers/net/virtio_net.c +++ b/drivers/net/virtio_net.c @@ -1898,8 +1898,12 @@ static void free_receive_page_frags(struct virtnet_info *vi) put_page(vi->rq[i].alloc_frag.page); } -static bool is_xdp_queue(struct virtnet_info *vi, int q) +static bool is_xdp_raw_buffer_queue(struct virtnet_info *vi, int q) { + /* For small receive mode always use kfree_skb variants */ + if (!vi->mergeable_rx_bufs) + return false; + if (q < (vi->curr_queue_pairs - vi->xdp_queue_pairs)) return false; else if (q < vi->curr_queue_pairs) @@ -1916,7 +1920,7 @@ static void free_unused_bufs(struct virtnet_info *vi) for (i = 0; i < vi->max_queue_pairs; i++) { struct virtqueue *vq = vi->sq[i].vq; while ((buf = virtqueue_detach_unused_buf(vq)) != NULL) { - if (!is_xdp_queue(vi, i)) + if (!is_xdp_raw_buffer_queue(vi, i)) dev_kfree_skb(buf); else put_page(virt_to_head_page(buf)); From 529ec6ac26656378435eb0396a780f017d51e105 Mon Sep 17 00:00:00 2001 From: Jakub Kicinski Date: Wed, 25 Jan 2017 14:56:36 -0800 Subject: [PATCH 300/343] virtio_net: reject XDP programs using header adjustment commit 17bedab27231 ("bpf: xdp: Allow head adjustment in XDP prog") added a new XDP helper to prepend and remove data from a frame. Make virtio_net reject programs making use of this helper until proper support is added. Signed-off-by: Jakub Kicinski Acked-by: John Fastabend Acked-by: Jason Wang Acked-by: Michael S. Tsirkin Signed-off-by: David S. Miller --- drivers/net/virtio_net.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c index 4c6e9b43f1d3..765c2d6358da 100644 --- a/drivers/net/virtio_net.c +++ b/drivers/net/virtio_net.c @@ -1715,6 +1715,11 @@ static int virtnet_xdp_set(struct net_device *dev, struct bpf_prog *prog) u16 xdp_qp = 0, curr_qp; int i, err; + if (prog && prog->xdp_adjust_head) { + netdev_warn(dev, "Does not support bpf_xdp_adjust_head()\n"); + return -EOPNOTSUPP; + } + if (virtio_has_feature(vi->vdev, VIRTIO_NET_F_GUEST_TSO4) || virtio_has_feature(vi->vdev, VIRTIO_NET_F_GUEST_TSO6) || virtio_has_feature(vi->vdev, VIRTIO_NET_F_GUEST_ECN) || From 2aa6ba7b5ad3189cc27f14540aa2f57f0ed8df4b Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 25 Jan 2017 20:24:57 -0800 Subject: [PATCH 301/343] xfs: clear _XBF_PAGES from buffers when readahead page If we try to allocate memory pages to back an xfs_buf that we're trying to read, it's possible that we'll be so short on memory that the page allocation fails. For a blocking read we'll just wait, but for readahead we simply dump all the pages we've collected so far. Unfortunately, after dumping the pages we neglect to clear the _XBF_PAGES state, which means that the subsequent call to xfs_buf_free thinks that b_pages still points to pages we own. It then double-frees the b_pages pages. This results in screaming about negative page refcounts from the memory manager, which xfs oughtn't be triggering. To reproduce this case, mount a filesystem where the size of the inodes far outweighs the availalble memory (a ~500M inode filesystem on a VM with 300MB memory did the trick here) and run bulkstat in parallel with other memory eating processes to put a huge load on the system. The "check summary" phase of xfs_scrub also works for this purpose. Signed-off-by: Darrick J. Wong Reviewed-by: Eric Sandeen --- fs/xfs/xfs_buf.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index 7f0a01f7b592..ac3b4db519df 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -422,6 +422,7 @@ retry: out_free_pages: for (i = 0; i < bp->b_page_count; i++) __free_page(bp->b_pages[i]); + bp->b_flags &= ~_XBF_PAGES; return error; } From c81e55e057b6458aac6d96a6429ef021b7f6f62c Mon Sep 17 00:00:00 2001 From: James Smart Date: Sat, 24 Dec 2016 09:46:43 -0800 Subject: [PATCH 302/343] nvmet_fc: correct logic in disconnect queue LS handling Correct logic in disconnect queue LS handling. Rework so that queue searching and error reporting is above the section to send back a ls rjt Signed-off-by: James Smart Signed-off-by: Sagi Grimberg --- drivers/nvme/target/fc.c | 36 ++++++++++++++++++++++-------------- 1 file changed, 22 insertions(+), 14 deletions(-) diff --git a/drivers/nvme/target/fc.c b/drivers/nvme/target/fc.c index 173e842f19c9..ba57f9852bde 100644 --- a/drivers/nvme/target/fc.c +++ b/drivers/nvme/target/fc.c @@ -1314,7 +1314,7 @@ nvmet_fc_ls_disconnect(struct nvmet_fc_tgtport *tgtport, (struct fcnvme_ls_disconnect_rqst *)iod->rqstbuf; struct fcnvme_ls_disconnect_acc *acc = (struct fcnvme_ls_disconnect_acc *)iod->rspbuf; - struct nvmet_fc_tgt_queue *queue; + struct nvmet_fc_tgt_queue *queue = NULL; struct nvmet_fc_tgt_assoc *assoc; int ret = 0; bool del_assoc = false; @@ -1348,7 +1348,18 @@ nvmet_fc_ls_disconnect(struct nvmet_fc_tgtport *tgtport, assoc = nvmet_fc_find_target_assoc(tgtport, be64_to_cpu(rqst->associd.association_id)); iod->assoc = assoc; - if (!assoc) + if (assoc) { + if (rqst->discon_cmd.scope == + FCNVME_DISCONN_CONNECTION) { + queue = nvmet_fc_find_target_queue(tgtport, + be64_to_cpu( + rqst->discon_cmd.id)); + if (!queue) { + nvmet_fc_tgt_a_put(assoc); + ret = VERR_NO_CONN; + } + } + } else ret = VERR_NO_ASSOC; } @@ -1373,21 +1384,18 @@ nvmet_fc_ls_disconnect(struct nvmet_fc_tgtport *tgtport, FCNVME_LS_DISCONNECT); - if (rqst->discon_cmd.scope == FCNVME_DISCONN_CONNECTION) { - queue = nvmet_fc_find_target_queue(tgtport, - be64_to_cpu(rqst->discon_cmd.id)); - if (queue) { - int qid = queue->qid; + /* are we to delete a Connection ID (queue) */ + if (queue) { + int qid = queue->qid; - nvmet_fc_delete_target_queue(queue); + nvmet_fc_delete_target_queue(queue); - /* release the get taken by find_target_queue */ - nvmet_fc_tgt_q_put(queue); + /* release the get taken by find_target_queue */ + nvmet_fc_tgt_q_put(queue); - /* tear association down if io queue terminated */ - if (!qid) - del_assoc = true; - } + /* tear association down if io queue terminated */ + if (!qid) + del_assoc = true; } /* release get taken in nvmet_fc_find_target_assoc */ From 344770b07b7ae70639ebf110010eb6156a6e55e9 Mon Sep 17 00:00:00 2001 From: Sagi Grimberg Date: Sun, 27 Nov 2016 22:29:17 +0200 Subject: [PATCH 303/343] nvmet: delete controllers deletion upon subsystem release No reason for them to be kept around if we are deleting the subsystem, so instead of passively wait for the host to disconnect, actively delete the controllers. Signed-off-by: Sagi Grimberg Reviewed-by: Johannes Thumshirn Reviewed-by: Max Gurtovoy Reviewed-by: Christoph Hellwig --- drivers/nvme/target/configfs.c | 1 + drivers/nvme/target/core.c | 10 ++++++++++ drivers/nvme/target/nvmet.h | 1 + 3 files changed, 12 insertions(+) diff --git a/drivers/nvme/target/configfs.c b/drivers/nvme/target/configfs.c index 6f5074153dcd..be8c800078e2 100644 --- a/drivers/nvme/target/configfs.c +++ b/drivers/nvme/target/configfs.c @@ -631,6 +631,7 @@ static void nvmet_subsys_release(struct config_item *item) { struct nvmet_subsys *subsys = to_subsys(item); + nvmet_subsys_del_ctrls(subsys); nvmet_subsys_put(subsys); } diff --git a/drivers/nvme/target/core.c b/drivers/nvme/target/core.c index b1d66ed655c9..4a367549eb93 100644 --- a/drivers/nvme/target/core.c +++ b/drivers/nvme/target/core.c @@ -935,6 +935,16 @@ static void nvmet_subsys_free(struct kref *ref) kfree(subsys); } +void nvmet_subsys_del_ctrls(struct nvmet_subsys *subsys) +{ + struct nvmet_ctrl *ctrl; + + mutex_lock(&subsys->lock); + list_for_each_entry(ctrl, &subsys->ctrls, subsys_entry) + ctrl->ops->delete_ctrl(ctrl); + mutex_unlock(&subsys->lock); +} + void nvmet_subsys_put(struct nvmet_subsys *subsys) { kref_put(&subsys->ref, nvmet_subsys_free); diff --git a/drivers/nvme/target/nvmet.h b/drivers/nvme/target/nvmet.h index 23d5eb1c944f..cc7ad06b43a7 100644 --- a/drivers/nvme/target/nvmet.h +++ b/drivers/nvme/target/nvmet.h @@ -282,6 +282,7 @@ void nvmet_ctrl_put(struct nvmet_ctrl *ctrl); struct nvmet_subsys *nvmet_subsys_alloc(const char *subsysnqn, enum nvme_subsys_type type); void nvmet_subsys_put(struct nvmet_subsys *subsys); +void nvmet_subsys_del_ctrls(struct nvmet_subsys *subsys); struct nvmet_ns *nvmet_find_namespace(struct nvmet_ctrl *ctrl, __le32 nsid); void nvmet_put_namespace(struct nvmet_ns *ns); From 06406d81a2d7cfb8abcc4fa6cdfeb8e5897007c5 Mon Sep 17 00:00:00 2001 From: Sagi Grimberg Date: Sun, 1 Jan 2017 13:41:56 +0200 Subject: [PATCH 304/343] nvmet: cancel fatal error and flush async work before free controller Make sure they are not running and we can free the controller safely. Signed-off-by: Roy Shterman Signed-off-by: Sagi Grimberg Reviewed-by: Christoph Hellwig --- drivers/nvme/target/core.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/nvme/target/core.c b/drivers/nvme/target/core.c index 4a367549eb93..a327a43f8166 100644 --- a/drivers/nvme/target/core.c +++ b/drivers/nvme/target/core.c @@ -816,6 +816,9 @@ static void nvmet_ctrl_free(struct kref *ref) list_del(&ctrl->subsys_entry); mutex_unlock(&subsys->lock); + flush_work(&ctrl->async_event_work); + cancel_work_sync(&ctrl->fatal_err_work); + ida_simple_remove(&subsys->cntlid_ida, ctrl->cntlid); nvmet_subsys_put(subsys); From 23a8ed4a624324dc696c328f09bd502c4a3816f0 Mon Sep 17 00:00:00 2001 From: Sagi Grimberg Date: Sun, 1 Jan 2017 13:18:26 +0200 Subject: [PATCH 305/343] nvmet: Call fatal_error from keep-alive timout expiration We only need to call delete_ctrl once, so given that both keep-alive timeout and any other fatal error can trigger it, just make sure we only call delete_ctrl once. Signed-off-by: Sagi Grimberg Reviewed-by: Christoph Hellwig --- drivers/nvme/target/core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/nvme/target/core.c b/drivers/nvme/target/core.c index a327a43f8166..fc5ba2f9e15f 100644 --- a/drivers/nvme/target/core.c +++ b/drivers/nvme/target/core.c @@ -200,7 +200,7 @@ static void nvmet_keep_alive_timer(struct work_struct *work) pr_err("ctrl %d keep-alive timer (%d seconds) expired!\n", ctrl->cntlid, ctrl->kato); - ctrl->ops->delete_ctrl(ctrl); + nvmet_ctrl_fatal_error(ctrl); } static void nvmet_start_keep_alive_timer(struct nvmet_ctrl *ctrl) From 748ff8408f8e208f279ba221e5c12612fbb4dddb Mon Sep 17 00:00:00 2001 From: Parav Pandit Date: Thu, 19 Jan 2017 09:55:08 -0600 Subject: [PATCH 306/343] nvmet-rdma: Fix missing dma sync to nvme data structures This patch performs dma sync operations on nvme_command and nvme_completion. nvme_command is synced (a) on receiving of the recv queue completion for cpu access. (b) before posting recv wqe back to rdma adapter for device access. nvme_completion is synced (a) on receiving of the recv queue completion of associated nvme_command for cpu access. (b) before posting send wqe to rdma adapter for device access. This patch is generated for git://git.infradead.org/nvme-fabrics.git Branch: nvmf-4.10 Signed-off-by: Parav Pandit Reviewed-by: Max Gurtovoy --- drivers/nvme/target/rdma.c | 17 +++++++++++++++++ 1 file changed, 17 insertions(+) diff --git a/drivers/nvme/target/rdma.c b/drivers/nvme/target/rdma.c index 8c3760a78ac0..60990220bd83 100644 --- a/drivers/nvme/target/rdma.c +++ b/drivers/nvme/target/rdma.c @@ -438,6 +438,10 @@ static int nvmet_rdma_post_recv(struct nvmet_rdma_device *ndev, { struct ib_recv_wr *bad_wr; + ib_dma_sync_single_for_device(ndev->device, + cmd->sge[0].addr, cmd->sge[0].length, + DMA_FROM_DEVICE); + if (ndev->srq) return ib_post_srq_recv(ndev->srq, &cmd->wr, &bad_wr); return ib_post_recv(cmd->queue->cm_id->qp, &cmd->wr, &bad_wr); @@ -538,6 +542,11 @@ static void nvmet_rdma_queue_response(struct nvmet_req *req) first_wr = &rsp->send_wr; nvmet_rdma_post_recv(rsp->queue->dev, rsp->cmd); + + ib_dma_sync_single_for_device(rsp->queue->dev->device, + rsp->send_sge.addr, rsp->send_sge.length, + DMA_TO_DEVICE); + if (ib_post_send(cm_id->qp, first_wr, &bad_wr)) { pr_err("sending cmd response failed\n"); nvmet_rdma_release_rsp(rsp); @@ -698,6 +707,14 @@ static void nvmet_rdma_handle_command(struct nvmet_rdma_queue *queue, cmd->n_rdma = 0; cmd->req.port = queue->port; + + ib_dma_sync_single_for_cpu(queue->dev->device, + cmd->cmd->sge[0].addr, cmd->cmd->sge[0].length, + DMA_FROM_DEVICE); + ib_dma_sync_single_for_cpu(queue->dev->device, + cmd->send_sge.addr, cmd->send_sge.length, + DMA_TO_DEVICE); + if (!nvmet_req_init(&cmd->req, &queue->nvme_cq, &queue->nvme_sq, &nvmet_rdma_ops)) return; From 19e420bb4076ace670addc55300e3b8c4a02dfc6 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 19 Jan 2017 16:55:57 +0100 Subject: [PATCH 307/343] nvme-fc: use blk_rq_nr_phys_segments Without this deallocate won't work properly due to the mismatch of the bio/request size and the actual payload size. Signed-off-by: Christoph Hellwig Reviewed-by: James Smart Reviewed-by: Johannes Thumshirn Signed-off-by: Sagi Grimberg --- drivers/nvme/host/fc.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/nvme/host/fc.c b/drivers/nvme/host/fc.c index fcc9dcfdf675..e65041c640cb 100644 --- a/drivers/nvme/host/fc.c +++ b/drivers/nvme/host/fc.c @@ -1663,13 +1663,13 @@ nvme_fc_map_data(struct nvme_fc_ctrl *ctrl, struct request *rq, return 0; freq->sg_table.sgl = freq->first_sgl; - ret = sg_alloc_table_chained(&freq->sg_table, rq->nr_phys_segments, - freq->sg_table.sgl); + ret = sg_alloc_table_chained(&freq->sg_table, + blk_rq_nr_phys_segments(rq), freq->sg_table.sgl); if (ret) return -ENOMEM; op->nents = blk_rq_map_sg(rq->q, rq, freq->sg_table.sgl); - WARN_ON(op->nents > rq->nr_phys_segments); + WARN_ON(op->nents > blk_rq_nr_phys_segments(rq)); dir = (rq_data_dir(rq) == WRITE) ? DMA_TO_DEVICE : DMA_FROM_DEVICE; freq->sg_cnt = fc_dma_map_sg(ctrl->lport->dev, freq->sg_table.sgl, op->nents, dir); From 08965c2eba135bdfb6e86cf25308e01421c7e0ce Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Wed, 25 Jan 2017 13:43:56 -0800 Subject: [PATCH 308/343] Revert "sd: remove __data_len hack for WRITE SAME" This patch reverts commit f80de881d8df and avoids that sending a WRITE SAME command to the iSCSI initiator triggers the following: BUG: unable to handle kernel NULL pointer dereference at 0000000000000014 TARGET_CORE[iSCSI]: Expected Transfer Length: 260096 does not match SCSI CDB Length: 512 for SAM Opcode: 0x41 IP: iscsi_tcp_segment_done+0x20b/0x310 [libiscsi_tcp] Oops: 0000 [#1] SMP Modules linked in: target_core_user uio target_core_iblock target_core_file iscsi_target_mod target_core_mod netconsole configfs crct10dif_pclmul crc32_pclmul ghash_clmulni_intel aesni_intel aes_x86_64 crypto_simd cryptd glue_helper virtio_console virtio_rng virtio_balloon serio_raw i2c_piix4 acpi_cpufreq button iscsi_tcp libiscsi_tcp libiscsi scsi_transport_iscsi ext4 jbd2 mbcache virtio_blk virtio_net psmouse floppy drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops ttm drm virtio_pci CPU: 2 PID: 5 Comm: kworker/u8:0 Not tainted 4.10.0-rc5-debug+ #3 Workqueue: iscsi_q_0 iscsi_xmitworker [libiscsi] RIP: 0010:iscsi_tcp_segment_done+0x20b/0x310 [libiscsi_tcp] Call Trace: iscsi_sw_tcp_xmit_segment+0x84/0x120 [iscsi_tcp] iscsi_sw_tcp_pdu_xmit+0x51/0x180 [iscsi_tcp] iscsi_tcp_task_xmit+0xb3/0x290 [libiscsi_tcp] iscsi_xmit_task+0x4e/0xc0 [libiscsi] iscsi_xmitworker+0x243/0x330 [libiscsi] process_one_work+0x1d8/0x4b0 worker_thread+0x49/0x4a0 kthread+0x102/0x140 Fixes: f80de881d8df ("sd: remove __data_len hack for WRITE SAME") Signed-off-by: Bart Van Assche Cc: Hannes Reinecke Cc: Sagi Grimberg Cc: Jens Axboe Cc: Lee Duncan Cc: Chris Leech Acked-by: Christoph Hellwig Acked-by: Martin K. Petersen Signed-off-by: Jens Axboe --- drivers/scsi/sd.c | 17 ++++++++++++++++- 1 file changed, 16 insertions(+), 1 deletion(-) diff --git a/drivers/scsi/sd.c b/drivers/scsi/sd.c index 0b09638fa39b..1f5d92a25a49 100644 --- a/drivers/scsi/sd.c +++ b/drivers/scsi/sd.c @@ -836,6 +836,7 @@ static int sd_setup_write_same_cmnd(struct scsi_cmnd *cmd) struct bio *bio = rq->bio; sector_t sector = blk_rq_pos(rq); unsigned int nr_sectors = blk_rq_sectors(rq); + unsigned int nr_bytes = blk_rq_bytes(rq); int ret; if (sdkp->device->no_write_same) @@ -868,7 +869,21 @@ static int sd_setup_write_same_cmnd(struct scsi_cmnd *cmd) cmd->transfersize = sdp->sector_size; cmd->allowed = SD_MAX_RETRIES; - return scsi_init_io(cmd); + + /* + * For WRITE SAME the data transferred via the DATA OUT buffer is + * different from the amount of data actually written to the target. + * + * We set up __data_len to the amount of data transferred via the + * DATA OUT buffer so that blk_rq_map_sg sets up the proper S/G list + * to transfer a single sector of data first, but then reset it to + * the amount of data to be written right after so that the I/O path + * knows how much to actually write. + */ + rq->__data_len = sdp->sector_size; + ret = scsi_init_io(cmd); + rq->__data_len = nr_bytes; + return ret; } static int sd_setup_flush_cmnd(struct scsi_cmnd *cmd) From ff9f8a7cf935468a94d9927c68b00daae701667e Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Wed, 25 Jan 2017 18:20:55 -0800 Subject: [PATCH 309/343] sysctl: fix proc_doulongvec_ms_jiffies_minmax() We perform the conversion between kernel jiffies and ms only when exporting kernel value to user space. We need to do the opposite operation when value is written by user. Only matters when HZ != 1000 Signed-off-by: Eric Dumazet Cc: stable@vger.kernel.org Signed-off-by: Linus Torvalds --- kernel/sysctl.c | 1 + 1 file changed, 1 insertion(+) diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 8dbaec0e4f7f..1aea594a54db 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -2475,6 +2475,7 @@ static int __do_proc_doulongvec_minmax(void *data, struct ctl_table *table, int break; if (neg) continue; + val = convmul * val / convdiv; if ((min && val < *min) || (max && val > *max)) continue; *i = val; From c364b6d0b6cda1cd5d9ab689489adda3e82529aa Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Thu, 26 Jan 2017 09:50:30 -0800 Subject: [PATCH 310/343] xfs: fix bmv_count confusion w/ shared extents In a bmapx call, bmv_count is the total size of the array, including the zeroth element that userspace uses to supply the search key. The output array starts at offset 1 so that we can set up the user for the next invocation. Since we now can split an extent into multiple bmap records due to shared/unshared status, we have to be careful that we don't overflow the output array. In the original patch f86f403794b ("xfs: teach get_bmapx about shared extents and the CoW fork") I used cur_ext (the output index) to check for overflows, albeit with an off-by-one error. Since nexleft no longer describes the number of unfilled slots in the output, we can rip all that out and use cur_ext for the overflow check directly. Failure to do this causes heap corruption in bmapx callers such as xfs_io and xfs_scrub. xfs/328 can reproduce this problem. Reviewed-by: Eric Sandeen Signed-off-by: Darrick J. Wong --- fs/xfs/xfs_bmap_util.c | 28 ++++++++++++++++++---------- 1 file changed, 18 insertions(+), 10 deletions(-) diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c index b9abce524c33..c1417919ab0a 100644 --- a/fs/xfs/xfs_bmap_util.c +++ b/fs/xfs/xfs_bmap_util.c @@ -528,7 +528,6 @@ xfs_getbmap( xfs_bmbt_irec_t *map; /* buffer for user's data */ xfs_mount_t *mp; /* file system mount point */ int nex; /* # of user extents can do */ - int nexleft; /* # of user extents left */ int subnex; /* # of bmapi's can do */ int nmap; /* number of map entries */ struct getbmapx *out; /* output structure */ @@ -686,10 +685,8 @@ xfs_getbmap( goto out_free_map; } - nexleft = nex; - do { - nmap = (nexleft > subnex) ? subnex : nexleft; + nmap = (nex> subnex) ? subnex : nex; error = xfs_bmapi_read(ip, XFS_BB_TO_FSBT(mp, bmv->bmv_offset), XFS_BB_TO_FSB(mp, bmv->bmv_length), map, &nmap, bmapi_flags); @@ -697,8 +694,8 @@ xfs_getbmap( goto out_free_map; ASSERT(nmap <= subnex); - for (i = 0; i < nmap && nexleft && bmv->bmv_length && - cur_ext < bmv->bmv_count; i++) { + for (i = 0; i < nmap && bmv->bmv_length && + cur_ext < bmv->bmv_count - 1; i++) { out[cur_ext].bmv_oflags = 0; if (map[i].br_state == XFS_EXT_UNWRITTEN) out[cur_ext].bmv_oflags |= BMV_OF_PREALLOC; @@ -760,16 +757,27 @@ xfs_getbmap( continue; } + /* + * In order to report shared extents accurately, + * we report each distinct shared/unshared part + * of a single bmbt record using multiple bmap + * extents. To make that happen, we iterate the + * same map array item multiple times, each + * time trimming out the subextent that we just + * reported. + * + * Because of this, we must check the out array + * index (cur_ext) directly against bmv_count-1 + * to avoid overflows. + */ if (inject_map.br_startblock != NULLFSBLOCK) { map[i] = inject_map; i--; - } else - nexleft--; + } bmv->bmv_entries++; cur_ext++; } - } while (nmap && nexleft && bmv->bmv_length && - cur_ext < bmv->bmv_count); + } while (nmap && bmv->bmv_length && cur_ext < bmv->bmv_count - 1); out_free_map: kmem_free(map); From 406dab8450ec76eca88a1af2fc15d18a2b36ca49 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Thu, 26 Jan 2017 15:14:52 -0500 Subject: [PATCH 311/343] nfs: Fix "Don't increment lock sequence ID after NFS4ERR_MOVED" Lock sequence IDs are bumped in decode_lock by calling nfs_increment_seqid(). nfs_increment_sequid() does not use the seqid_mutating_err() function fixed in commit 059aa7348241 ("Don't increment lock sequence ID after NFS4ERR_MOVED"). Fixes: 059aa7348241 ("Don't increment lock sequence ID after ...") Signed-off-by: Chuck Lever Tested-by: Xuan Qi Cc: stable@vger.kernel.org # v3.7+ Signed-off-by: Trond Myklebust --- fs/nfs/nfs4state.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index 90e6193ce6be..daeb94e3acd4 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -1091,6 +1091,7 @@ static void nfs_increment_seqid(int status, struct nfs_seqid *seqid) case -NFS4ERR_BADXDR: case -NFS4ERR_RESOURCE: case -NFS4ERR_NOFILEHANDLE: + case -NFS4ERR_MOVED: /* Non-seqid mutating errors */ return; }; From ee6625a948d2e47267ec8fd97307fdd67d0f8a5b Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 26 Jan 2017 15:50:41 -0500 Subject: [PATCH 312/343] pNFS: Fix a reference leak in _pnfs_return_layout IF NFS_LAYOUT_RETURN_REQUESTED is not set, then we currently exit without freeing the list of invalidated layout segments, leading to a reference leak. Reported-by: Olga Kornievskaia Fixes: 24408f5282 ("pNFS: Fix bugs in _pnfs_return_layout") Signed-off-by: Trond Myklebust --- fs/nfs/pnfs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index 59554f3adf29..dd042498ce7c 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -1200,10 +1200,10 @@ _pnfs_return_layout(struct inode *ino) send = pnfs_prepare_layoutreturn(lo, &stateid, NULL); spin_unlock(&ino->i_lock); - pnfs_free_lseg_list(&tmp_list); if (send) status = pnfs_send_layoutreturn(lo, &stateid, IOMODE_ANY, true); out_put_layout_hdr: + pnfs_free_lseg_list(&tmp_list); pnfs_put_layout_hdr(lo); out: dprintk("<-- %s status: %d\n", __func__, status); From a47b70ea86bdeb3091341f5ae3ef580f1a1ad822 Mon Sep 17 00:00:00 2001 From: Kazuya Mizuguchi Date: Thu, 26 Jan 2017 14:29:27 +0100 Subject: [PATCH 313/343] ravb: unmap descriptors when freeing rings "swiotlb buffer is full" errors occur after repeated initialisation of a device - f.e. suspend/resume or ip link set up/down. This is because memory mapped using dma_map_single() in ravb_ring_format() and ravb_start_xmit() is not released. Resolve this problem by unmapping descriptors when freeing rings. Fixes: c156633f1353 ("Renesas Ethernet AVB driver proper") Signed-off-by: Kazuya Mizuguchi [simon: reworked] Signed-off-by: Simon Horman Acked-by: Sergei Shtylyov Signed-off-by: David S. Miller --- drivers/net/ethernet/renesas/ravb_main.c | 112 +++++++++++++---------- 1 file changed, 64 insertions(+), 48 deletions(-) diff --git a/drivers/net/ethernet/renesas/ravb_main.c b/drivers/net/ethernet/renesas/ravb_main.c index 89ac1e3f6175..301f48755093 100644 --- a/drivers/net/ethernet/renesas/ravb_main.c +++ b/drivers/net/ethernet/renesas/ravb_main.c @@ -179,6 +179,49 @@ static struct mdiobb_ops bb_ops = { .get_mdio_data = ravb_get_mdio_data, }; +/* Free TX skb function for AVB-IP */ +static int ravb_tx_free(struct net_device *ndev, int q, bool free_txed_only) +{ + struct ravb_private *priv = netdev_priv(ndev); + struct net_device_stats *stats = &priv->stats[q]; + struct ravb_tx_desc *desc; + int free_num = 0; + int entry; + u32 size; + + for (; priv->cur_tx[q] - priv->dirty_tx[q] > 0; priv->dirty_tx[q]++) { + bool txed; + + entry = priv->dirty_tx[q] % (priv->num_tx_ring[q] * + NUM_TX_DESC); + desc = &priv->tx_ring[q][entry]; + txed = desc->die_dt == DT_FEMPTY; + if (free_txed_only && !txed) + break; + /* Descriptor type must be checked before all other reads */ + dma_rmb(); + size = le16_to_cpu(desc->ds_tagl) & TX_DS; + /* Free the original skb. */ + if (priv->tx_skb[q][entry / NUM_TX_DESC]) { + dma_unmap_single(ndev->dev.parent, le32_to_cpu(desc->dptr), + size, DMA_TO_DEVICE); + /* Last packet descriptor? */ + if (entry % NUM_TX_DESC == NUM_TX_DESC - 1) { + entry /= NUM_TX_DESC; + dev_kfree_skb_any(priv->tx_skb[q][entry]); + priv->tx_skb[q][entry] = NULL; + if (txed) + stats->tx_packets++; + } + free_num++; + } + if (txed) + stats->tx_bytes += size; + desc->die_dt = DT_EEMPTY; + } + return free_num; +} + /* Free skb's and DMA buffers for Ethernet AVB */ static void ravb_ring_free(struct net_device *ndev, int q) { @@ -194,19 +237,21 @@ static void ravb_ring_free(struct net_device *ndev, int q) kfree(priv->rx_skb[q]); priv->rx_skb[q] = NULL; - /* Free TX skb ringbuffer */ - if (priv->tx_skb[q]) { - for (i = 0; i < priv->num_tx_ring[q]; i++) - dev_kfree_skb(priv->tx_skb[q][i]); - } - kfree(priv->tx_skb[q]); - priv->tx_skb[q] = NULL; - /* Free aligned TX buffers */ kfree(priv->tx_align[q]); priv->tx_align[q] = NULL; if (priv->rx_ring[q]) { + for (i = 0; i < priv->num_rx_ring[q]; i++) { + struct ravb_ex_rx_desc *desc = &priv->rx_ring[q][i]; + + if (!dma_mapping_error(ndev->dev.parent, + le32_to_cpu(desc->dptr))) + dma_unmap_single(ndev->dev.parent, + le32_to_cpu(desc->dptr), + PKT_BUF_SZ, + DMA_FROM_DEVICE); + } ring_size = sizeof(struct ravb_ex_rx_desc) * (priv->num_rx_ring[q] + 1); dma_free_coherent(ndev->dev.parent, ring_size, priv->rx_ring[q], @@ -215,12 +260,20 @@ static void ravb_ring_free(struct net_device *ndev, int q) } if (priv->tx_ring[q]) { + ravb_tx_free(ndev, q, false); + ring_size = sizeof(struct ravb_tx_desc) * (priv->num_tx_ring[q] * NUM_TX_DESC + 1); dma_free_coherent(ndev->dev.parent, ring_size, priv->tx_ring[q], priv->tx_desc_dma[q]); priv->tx_ring[q] = NULL; } + + /* Free TX skb ringbuffer. + * SKBs are freed by ravb_tx_free() call above. + */ + kfree(priv->tx_skb[q]); + priv->tx_skb[q] = NULL; } /* Format skb and descriptor buffer for Ethernet AVB */ @@ -431,44 +484,6 @@ static int ravb_dmac_init(struct net_device *ndev) return 0; } -/* Free TX skb function for AVB-IP */ -static int ravb_tx_free(struct net_device *ndev, int q) -{ - struct ravb_private *priv = netdev_priv(ndev); - struct net_device_stats *stats = &priv->stats[q]; - struct ravb_tx_desc *desc; - int free_num = 0; - int entry; - u32 size; - - for (; priv->cur_tx[q] - priv->dirty_tx[q] > 0; priv->dirty_tx[q]++) { - entry = priv->dirty_tx[q] % (priv->num_tx_ring[q] * - NUM_TX_DESC); - desc = &priv->tx_ring[q][entry]; - if (desc->die_dt != DT_FEMPTY) - break; - /* Descriptor type must be checked before all other reads */ - dma_rmb(); - size = le16_to_cpu(desc->ds_tagl) & TX_DS; - /* Free the original skb. */ - if (priv->tx_skb[q][entry / NUM_TX_DESC]) { - dma_unmap_single(ndev->dev.parent, le32_to_cpu(desc->dptr), - size, DMA_TO_DEVICE); - /* Last packet descriptor? */ - if (entry % NUM_TX_DESC == NUM_TX_DESC - 1) { - entry /= NUM_TX_DESC; - dev_kfree_skb_any(priv->tx_skb[q][entry]); - priv->tx_skb[q][entry] = NULL; - stats->tx_packets++; - } - free_num++; - } - stats->tx_bytes += size; - desc->die_dt = DT_EEMPTY; - } - return free_num; -} - static void ravb_get_tx_tstamp(struct net_device *ndev) { struct ravb_private *priv = netdev_priv(ndev); @@ -902,7 +917,7 @@ static int ravb_poll(struct napi_struct *napi, int budget) spin_lock_irqsave(&priv->lock, flags); /* Clear TX interrupt */ ravb_write(ndev, ~mask, TIS); - ravb_tx_free(ndev, q); + ravb_tx_free(ndev, q, true); netif_wake_subqueue(ndev, q); mmiowb(); spin_unlock_irqrestore(&priv->lock, flags); @@ -1567,7 +1582,8 @@ static netdev_tx_t ravb_start_xmit(struct sk_buff *skb, struct net_device *ndev) priv->cur_tx[q] += NUM_TX_DESC; if (priv->cur_tx[q] - priv->dirty_tx[q] > - (priv->num_tx_ring[q] - 1) * NUM_TX_DESC && !ravb_tx_free(ndev, q)) + (priv->num_tx_ring[q] - 1) * NUM_TX_DESC && + !ravb_tx_free(ndev, q, true)) netif_stop_subqueue(ndev, q); exit: From 67ade058ef2c65a3e56878af9c293ec76722a2e5 Mon Sep 17 00:00:00 2001 From: Omar Sandoval Date: Wed, 25 Jan 2017 17:06:38 -0800 Subject: [PATCH 314/343] Btrfs: remove old tree_root case in btrfs_read_locked_inode() As Jeff explained in c2951f32d36c ("btrfs: remove old tree_root dirent processing in btrfs_real_readdir()"), supporting this old format is no longer necessary since the Btrfs magic number has been updated since we changed to the current format. There are other places where we still handle this old format, but since this is part of a fix that is going to stable, I'm only removing this one for now. Cc: # 4.9.x Signed-off-by: Omar Sandoval Reviewed-by: David Sterba Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 177e7284909c..e2b961e20ec1 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -3835,10 +3835,7 @@ cache_acl: break; case S_IFDIR: inode->i_fop = &btrfs_dir_file_operations; - if (root == fs_info->tree_root) - inode->i_op = &btrfs_dir_ro_inode_operations; - else - inode->i_op = &btrfs_dir_inode_operations; + inode->i_op = &btrfs_dir_inode_operations; break; case S_IFLNK: inode->i_op = &btrfs_symlink_inode_operations; From 1fdf41941b8010691679638f8d0c8d08cfee7726 Mon Sep 17 00:00:00 2001 From: Omar Sandoval Date: Wed, 25 Jan 2017 17:06:39 -0800 Subject: [PATCH 315/343] Btrfs: disable xattr operations on subvolume directories When you snapshot a subvolume containing a subvolume, you get a placeholder directory where the subvolume would be. These directory inodes have ->i_ops set to btrfs_dir_ro_inode_operations. Previously, these i_ops didn't include the xattr operation callbacks. The conversion to xattr_handlers missed this case, leading to bogus attempts to set xattrs on these inodes. This manifested itself as failures when running delayed inodes. To fix this, clear IOP_XATTR in ->i_opflags on these inodes. Fixes: 6c6ef9f26e59 ("xattr: Stop calling {get,set,remove}xattr inode operations") Cc: Andreas Gruenbacher Reported-by: Chris Murphy Tested-by: Chris Murphy Cc: # 4.9.x Signed-off-by: Omar Sandoval Reviewed-by: David Sterba Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index e2b961e20ec1..16694325a449 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -5718,6 +5718,7 @@ static struct inode *new_simple_dir(struct super_block *s, inode->i_ino = BTRFS_EMPTY_SUBVOL_DIR_OBJECTID; inode->i_op = &btrfs_dir_ro_inode_operations; + inode->i_opflags &= ~IOP_XATTR; inode->i_fop = &simple_dir_operations; inode->i_mode = S_IFDIR | S_IRUGO | S_IWUSR | S_IXUGO; inode->i_mtime = current_time(inode); From 57b59ed2e5b91e958843609c7884794e29e6c4cb Mon Sep 17 00:00:00 2001 From: Omar Sandoval Date: Wed, 25 Jan 2017 17:06:40 -0800 Subject: [PATCH 316/343] Btrfs: remove ->{get, set}_acl() from btrfs_dir_ro_inode_operations Subvolume directory inodes can't have ACLs. Cc: # 4.9.x Signed-off-by: Omar Sandoval Reviewed-by: David Sterba Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 16694325a449..8a8e719778bd 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -10592,8 +10592,6 @@ static const struct inode_operations btrfs_dir_inode_operations = { static const struct inode_operations btrfs_dir_ro_inode_operations = { .lookup = btrfs_lookup, .permission = btrfs_permission, - .get_acl = btrfs_get_acl, - .set_acl = btrfs_set_acl, .update_time = btrfs_update_time, }; From 6c971c09f38704513c426ba6515f22fb3d6c87d5 Mon Sep 17 00:00:00 2001 From: "Y.C. Chen" Date: Thu, 26 Jan 2017 09:45:40 +0800 Subject: [PATCH 317/343] drm/ast: Fixed system hanged if disable P2A The original ast driver will access some BMC configuration through P2A bridge that can be disabled since AST2300 and after. It will cause system hanged if P2A bridge is disabled. Here is the update to fix it. Signed-off-by: Y.C. Chen Signed-off-by: Dave Airlie --- drivers/gpu/drm/ast/ast_drv.h | 1 + drivers/gpu/drm/ast/ast_main.c | 157 +++++++++++++++++---------------- drivers/gpu/drm/ast/ast_post.c | 18 ++-- 3 files changed, 97 insertions(+), 79 deletions(-) diff --git a/drivers/gpu/drm/ast/ast_drv.h b/drivers/gpu/drm/ast/ast_drv.h index 908011d2c8f5..7abda94fc2cf 100644 --- a/drivers/gpu/drm/ast/ast_drv.h +++ b/drivers/gpu/drm/ast/ast_drv.h @@ -113,6 +113,7 @@ struct ast_private { struct ttm_bo_kmap_obj cache_kmap; int next_cursor; bool support_wide_screen; + bool DisableP2A; enum ast_tx_chip tx_chip_type; u8 dp501_maxclk; diff --git a/drivers/gpu/drm/ast/ast_main.c b/drivers/gpu/drm/ast/ast_main.c index f75c6421db62..533e762d036d 100644 --- a/drivers/gpu/drm/ast/ast_main.c +++ b/drivers/gpu/drm/ast/ast_main.c @@ -124,6 +124,12 @@ static int ast_detect_chip(struct drm_device *dev, bool *need_post) } else *need_post = false; + /* Check P2A Access */ + ast->DisableP2A = true; + data = ast_read32(ast, 0xf004); + if (data != 0xFFFFFFFF) + ast->DisableP2A = false; + /* Check if we support wide screen */ switch (ast->chip) { case AST1180: @@ -140,15 +146,17 @@ static int ast_detect_chip(struct drm_device *dev, bool *need_post) ast->support_wide_screen = true; else { ast->support_wide_screen = false; - /* Read SCU7c (silicon revision register) */ - ast_write32(ast, 0xf004, 0x1e6e0000); - ast_write32(ast, 0xf000, 0x1); - data = ast_read32(ast, 0x1207c); - data &= 0x300; - if (ast->chip == AST2300 && data == 0x0) /* ast1300 */ - ast->support_wide_screen = true; - if (ast->chip == AST2400 && data == 0x100) /* ast1400 */ - ast->support_wide_screen = true; + if (ast->DisableP2A == false) { + /* Read SCU7c (silicon revision register) */ + ast_write32(ast, 0xf004, 0x1e6e0000); + ast_write32(ast, 0xf000, 0x1); + data = ast_read32(ast, 0x1207c); + data &= 0x300; + if (ast->chip == AST2300 && data == 0x0) /* ast1300 */ + ast->support_wide_screen = true; + if (ast->chip == AST2400 && data == 0x100) /* ast1400 */ + ast->support_wide_screen = true; + } } break; } @@ -216,80 +224,81 @@ static int ast_get_dram_info(struct drm_device *dev) uint32_t data, data2; uint32_t denum, num, div, ref_pll; - ast_write32(ast, 0xf004, 0x1e6e0000); - ast_write32(ast, 0xf000, 0x1); - - - ast_write32(ast, 0x10000, 0xfc600309); - - do { - if (pci_channel_offline(dev->pdev)) - return -EIO; - } while (ast_read32(ast, 0x10000) != 0x01); - data = ast_read32(ast, 0x10004); - - if (data & 0x40) + if (ast->DisableP2A) + { ast->dram_bus_width = 16; + ast->dram_type = AST_DRAM_1Gx16; + ast->mclk = 396; + } else - ast->dram_bus_width = 32; + { + ast_write32(ast, 0xf004, 0x1e6e0000); + ast_write32(ast, 0xf000, 0x1); + data = ast_read32(ast, 0x10004); - if (ast->chip == AST2300 || ast->chip == AST2400) { - switch (data & 0x03) { - case 0: - ast->dram_type = AST_DRAM_512Mx16; - break; - default: - case 1: - ast->dram_type = AST_DRAM_1Gx16; + if (data & 0x40) + ast->dram_bus_width = 16; + else + ast->dram_bus_width = 32; + + if (ast->chip == AST2300 || ast->chip == AST2400) { + switch (data & 0x03) { + case 0: + ast->dram_type = AST_DRAM_512Mx16; + break; + default: + case 1: + ast->dram_type = AST_DRAM_1Gx16; + break; + case 2: + ast->dram_type = AST_DRAM_2Gx16; + break; + case 3: + ast->dram_type = AST_DRAM_4Gx16; + break; + } + } else { + switch (data & 0x0c) { + case 0: + case 4: + ast->dram_type = AST_DRAM_512Mx16; + break; + case 8: + if (data & 0x40) + ast->dram_type = AST_DRAM_1Gx16; + else + ast->dram_type = AST_DRAM_512Mx32; + break; + case 0xc: + ast->dram_type = AST_DRAM_1Gx32; + break; + } + } + + data = ast_read32(ast, 0x10120); + data2 = ast_read32(ast, 0x10170); + if (data2 & 0x2000) + ref_pll = 14318; + else + ref_pll = 12000; + + denum = data & 0x1f; + num = (data & 0x3fe0) >> 5; + data = (data & 0xc000) >> 14; + switch (data) { + case 3: + div = 0x4; break; case 2: - ast->dram_type = AST_DRAM_2Gx16; + case 1: + div = 0x2; break; - case 3: - ast->dram_type = AST_DRAM_4Gx16; - break; - } - } else { - switch (data & 0x0c) { - case 0: - case 4: - ast->dram_type = AST_DRAM_512Mx16; - break; - case 8: - if (data & 0x40) - ast->dram_type = AST_DRAM_1Gx16; - else - ast->dram_type = AST_DRAM_512Mx32; - break; - case 0xc: - ast->dram_type = AST_DRAM_1Gx32; + default: + div = 0x1; break; } + ast->mclk = ref_pll * (num + 2) / (denum + 2) * (div * 1000); } - - data = ast_read32(ast, 0x10120); - data2 = ast_read32(ast, 0x10170); - if (data2 & 0x2000) - ref_pll = 14318; - else - ref_pll = 12000; - - denum = data & 0x1f; - num = (data & 0x3fe0) >> 5; - data = (data & 0xc000) >> 14; - switch (data) { - case 3: - div = 0x4; - break; - case 2: - case 1: - div = 0x2; - break; - default: - div = 0x1; - break; - } - ast->mclk = ref_pll * (num + 2) / (denum + 2) * (div * 1000); return 0; } diff --git a/drivers/gpu/drm/ast/ast_post.c b/drivers/gpu/drm/ast/ast_post.c index 810c51d92b99..5331ee1df086 100644 --- a/drivers/gpu/drm/ast/ast_post.c +++ b/drivers/gpu/drm/ast/ast_post.c @@ -379,12 +379,20 @@ void ast_post_gpu(struct drm_device *dev) ast_open_key(ast); ast_set_def_ext_reg(dev); - if (ast->chip == AST2300 || ast->chip == AST2400) - ast_init_dram_2300(dev); - else - ast_init_dram_reg(dev); + if (ast->DisableP2A == false) + { + if (ast->chip == AST2300 || ast->chip == AST2400) + ast_init_dram_2300(dev); + else + ast_init_dram_reg(dev); - ast_init_3rdtx(dev); + ast_init_3rdtx(dev); + } + else + { + if (ast->tx_chip_type != AST_TX_NONE) + ast_set_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xa3, 0xcf, 0x80); /* Enable DVO */ + } } /* AST 2300 DRAM settings */ From cae9ff036eea577856d5b12860b4c79c5e71db4a Mon Sep 17 00:00:00 2001 From: Lyude Paul Date: Wed, 11 Jan 2017 21:25:23 -0500 Subject: [PATCH 318/343] drm/nouveau: Don't enabling polling twice on runtime resume As it turns out, on cards that actually have CRTCs on them we're already calling drm_kms_helper_poll_enable(drm_dev) from nouveau_display_resume() before we call it in nouveau_pmops_runtime_resume(). This leads us to accidentally trying to enable polling twice, which results in a potential deadlock between the RPM locks and drm_dev->mode_config.mutex if we end up trying to enable polling the second time while output_poll_execute is running and holding the mode_config lock. As such, make sure we only enable polling in nouveau_pmops_runtime_resume() if we need to. This fixes hangs observed on the ThinkPad W541 Signed-off-by: Lyude Cc: Hans de Goede Cc: Kilian Singer Cc: Lukas Wunner Cc: David Airlie Signed-off-by: Dave Airlie --- drivers/gpu/drm/nouveau/nouveau_display.c | 3 ++- drivers/gpu/drm/nouveau/nouveau_drm.c | 5 ++++- 2 files changed, 6 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/nouveau/nouveau_display.c b/drivers/gpu/drm/nouveau/nouveau_display.c index cef08da1da4e..6a157763dfc3 100644 --- a/drivers/gpu/drm/nouveau/nouveau_display.c +++ b/drivers/gpu/drm/nouveau/nouveau_display.c @@ -411,7 +411,8 @@ nouveau_display_init(struct drm_device *dev) return ret; /* enable polling for external displays */ - drm_kms_helper_poll_enable(dev); + if (!dev->mode_config.poll_enabled) + drm_kms_helper_poll_enable(dev); /* enable hotplug interrupts */ list_for_each_entry(connector, &dev->mode_config.connector_list, head) { diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.c b/drivers/gpu/drm/nouveau/nouveau_drm.c index 59348fc41c77..bc85a45f91cd 100644 --- a/drivers/gpu/drm/nouveau/nouveau_drm.c +++ b/drivers/gpu/drm/nouveau/nouveau_drm.c @@ -773,7 +773,10 @@ nouveau_pmops_runtime_resume(struct device *dev) pci_set_master(pdev); ret = nouveau_do_resume(drm_dev, true); - drm_kms_helper_poll_enable(drm_dev); + + if (!drm_dev->mode_config.poll_enabled) + drm_kms_helper_poll_enable(drm_dev); + /* do magic */ nvif_mask(&device->object, 0x088488, (1 << 25), (1 << 25)); vga_switcheroo_set_dynamic_switch(pdev, VGA_SWITCHEROO_ON); From 15266ae38fe09dae07bd8812cb7a7717b1e1d992 Mon Sep 17 00:00:00 2001 From: Lyude Paul Date: Wed, 11 Jan 2017 21:25:24 -0500 Subject: [PATCH 319/343] drm/nouveau: Handle fbcon suspend/resume in seperate worker Resuming from RPM can happen while already holding dev->mode_config.mutex. This means we can't actually handle fbcon in any RPM resume workers, since restoring fbcon requires grabbing dev->mode_config.mutex again. So move the fbcon suspend/resume code into it's own worker, and rely on that instead to avoid deadlocking. This fixes more deadlocks for runtime suspending the GPU on the ThinkPad W541. Reproduction recipe: - Get a machine with both optimus and a nvidia card with connectors attached to it - Wait for the nvidia GPU to suspend - Attempt to manually reprobe any of the connectors on the nvidia GPU using sysfs - *deadlock* [airlied: use READ_ONCE to address Hans's comment] Signed-off-by: Lyude Cc: Hans de Goede Cc: Kilian Singer Cc: Lukas Wunner Cc: David Airlie Signed-off-by: Dave Airlie --- drivers/gpu/drm/nouveau/nouveau_drv.h | 2 ++ drivers/gpu/drm/nouveau/nouveau_fbcon.c | 43 +++++++++++++++++++------ 2 files changed, 36 insertions(+), 9 deletions(-) diff --git a/drivers/gpu/drm/nouveau/nouveau_drv.h b/drivers/gpu/drm/nouveau/nouveau_drv.h index 8d5ed5bfdacb..42c1fa53d431 100644 --- a/drivers/gpu/drm/nouveau/nouveau_drv.h +++ b/drivers/gpu/drm/nouveau/nouveau_drv.h @@ -165,6 +165,8 @@ struct nouveau_drm { struct backlight_device *backlight; struct list_head bl_connectors; struct work_struct hpd_work; + struct work_struct fbcon_work; + int fbcon_new_state; #ifdef CONFIG_ACPI struct notifier_block acpi_nb; #endif diff --git a/drivers/gpu/drm/nouveau/nouveau_fbcon.c b/drivers/gpu/drm/nouveau/nouveau_fbcon.c index 2f2a3dcd4ad7..fa2d0a978ccc 100644 --- a/drivers/gpu/drm/nouveau/nouveau_fbcon.c +++ b/drivers/gpu/drm/nouveau/nouveau_fbcon.c @@ -470,19 +470,43 @@ static const struct drm_fb_helper_funcs nouveau_fbcon_helper_funcs = { .fb_probe = nouveau_fbcon_create, }; +static void +nouveau_fbcon_set_suspend_work(struct work_struct *work) +{ + struct nouveau_drm *drm = container_of(work, typeof(*drm), fbcon_work); + int state = READ_ONCE(drm->fbcon_new_state); + + if (state == FBINFO_STATE_RUNNING) + pm_runtime_get_sync(drm->dev->dev); + + console_lock(); + if (state == FBINFO_STATE_RUNNING) + nouveau_fbcon_accel_restore(drm->dev); + drm_fb_helper_set_suspend(&drm->fbcon->helper, state); + if (state != FBINFO_STATE_RUNNING) + nouveau_fbcon_accel_save_disable(drm->dev); + console_unlock(); + + if (state == FBINFO_STATE_RUNNING) { + pm_runtime_mark_last_busy(drm->dev->dev); + pm_runtime_put_sync(drm->dev->dev); + } +} + void nouveau_fbcon_set_suspend(struct drm_device *dev, int state) { struct nouveau_drm *drm = nouveau_drm(dev); - if (drm->fbcon) { - console_lock(); - if (state == FBINFO_STATE_RUNNING) - nouveau_fbcon_accel_restore(dev); - drm_fb_helper_set_suspend(&drm->fbcon->helper, state); - if (state != FBINFO_STATE_RUNNING) - nouveau_fbcon_accel_save_disable(dev); - console_unlock(); - } + + if (!drm->fbcon) + return; + + drm->fbcon_new_state = state; + /* Since runtime resume can happen as a result of a sysfs operation, + * it's possible we already have the console locked. So handle fbcon + * init/deinit from a seperate work thread + */ + schedule_work(&drm->fbcon_work); } int @@ -502,6 +526,7 @@ nouveau_fbcon_init(struct drm_device *dev) return -ENOMEM; drm->fbcon = fbcon; + INIT_WORK(&drm->fbcon_work, nouveau_fbcon_set_suspend_work); drm_fb_helper_prepare(dev, &fbcon->helper, &nouveau_fbcon_helper_funcs); From 606f42265d384b9149bfb953c5dfc6d4710fef4c Mon Sep 17 00:00:00 2001 From: Prashanth Prakash Date: Thu, 26 Jan 2017 11:08:32 -0700 Subject: [PATCH 320/343] arm64: skip register_cpufreq_notifier on ACPI-based systems On ACPI based systems where the topology is setup using the API store_cpu_topology, at the moment we do not have necessary code to parse cpu capacity and handle cpufreq notifier, thus resulting in a kernel panic. Stack: init_cpu_capacity_callback+0xb4/0x1c8 notifier_call_chain+0x5c/0xa0 __blocking_notifier_call_chain+0x58/0xa0 blocking_notifier_call_chain+0x3c/0x50 cpufreq_set_policy+0xe4/0x328 cpufreq_init_policy+0x80/0x100 cpufreq_online+0x418/0x710 cpufreq_add_dev+0x118/0x180 subsys_interface_register+0xa4/0xf8 cpufreq_register_driver+0x1c0/0x298 cppc_cpufreq_init+0xdc/0x1000 [cppc_cpufreq] do_one_initcall+0x5c/0x168 do_init_module+0x64/0x1e4 load_module+0x130c/0x14d0 SyS_finit_module+0x108/0x120 el0_svc_naked+0x24/0x28 Fixes: 7202bde8b7ae ("arm64: parse cpu capacity-dmips-mhz from DT") Acked-by: Will Deacon Signed-off-by: Prashanth Prakash Signed-off-by: Catalin Marinas --- arch/arm64/kernel/topology.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/arch/arm64/kernel/topology.c b/arch/arm64/kernel/topology.c index 23e9e13bd2aa..655e65f38f31 100644 --- a/arch/arm64/kernel/topology.c +++ b/arch/arm64/kernel/topology.c @@ -11,6 +11,7 @@ * for more details. */ +#include #include #include #include @@ -209,7 +210,12 @@ static struct notifier_block init_cpu_capacity_notifier = { static int __init register_cpufreq_notifier(void) { - if (cap_parsing_failed) + /* + * on ACPI-based systems we need to use the default cpu capacity + * until we have the necessary code to parse the cpu capacity, so + * skip registering cpufreq notifier. + */ + if (!acpi_disabled || cap_parsing_failed) return -EINVAL; if (!alloc_cpumask_var(&cpus_to_visit, GFP_KERNEL)) { From 92e55f412cffd016cc245a74278cb4d7b89bb3bc Mon Sep 17 00:00:00 2001 From: Pablo Neira Date: Thu, 26 Jan 2017 22:56:21 +0100 Subject: [PATCH 321/343] tcp: don't annotate mark on control socket from tcp_v6_send_response() Unlike ipv4, this control socket is shared by all cpus so we cannot use it as scratchpad area to annotate the mark that we pass to ip6_xmit(). Add a new parameter to ip6_xmit() to indicate the mark. The SCTP socket family caches the flowi6 structure in the sctp_transport structure, so we cannot use to carry the mark unless we later on reset it back, which I discarded since it looks ugly to me. Fixes: bf99b4ded5f8 ("tcp: fix mark propagation with fwmark_reflect enabled") Suggested-by: Eric Dumazet Signed-off-by: Pablo Neira Ayuso Signed-off-by: David S. Miller --- include/net/ipv6.h | 2 +- net/dccp/ipv6.c | 4 ++-- net/ipv6/inet6_connection_sock.c | 2 +- net/ipv6/ip6_output.c | 4 ++-- net/ipv6/tcp_ipv6.c | 5 ++--- net/sctp/ipv6.c | 3 ++- 6 files changed, 10 insertions(+), 10 deletions(-) diff --git a/include/net/ipv6.h b/include/net/ipv6.h index 487e57391664..7afe991e900e 100644 --- a/include/net/ipv6.h +++ b/include/net/ipv6.h @@ -871,7 +871,7 @@ int ip6_rcv_finish(struct net *net, struct sock *sk, struct sk_buff *skb); * upper-layer output functions */ int ip6_xmit(const struct sock *sk, struct sk_buff *skb, struct flowi6 *fl6, - struct ipv6_txoptions *opt, int tclass); + __u32 mark, struct ipv6_txoptions *opt, int tclass); int ip6_find_1stfragopt(struct sk_buff *skb, u8 **nexthdr); diff --git a/net/dccp/ipv6.c b/net/dccp/ipv6.c index adfc790f7193..c4e879c02186 100644 --- a/net/dccp/ipv6.c +++ b/net/dccp/ipv6.c @@ -227,7 +227,7 @@ static int dccp_v6_send_response(const struct sock *sk, struct request_sock *req opt = ireq->ipv6_opt; if (!opt) opt = rcu_dereference(np->opt); - err = ip6_xmit(sk, skb, &fl6, opt, np->tclass); + err = ip6_xmit(sk, skb, &fl6, sk->sk_mark, opt, np->tclass); rcu_read_unlock(); err = net_xmit_eval(err); } @@ -281,7 +281,7 @@ static void dccp_v6_ctl_send_reset(const struct sock *sk, struct sk_buff *rxskb) dst = ip6_dst_lookup_flow(ctl_sk, &fl6, NULL); if (!IS_ERR(dst)) { skb_dst_set(skb, dst); - ip6_xmit(ctl_sk, skb, &fl6, NULL, 0); + ip6_xmit(ctl_sk, skb, &fl6, 0, NULL, 0); DCCP_INC_STATS(DCCP_MIB_OUTSEGS); DCCP_INC_STATS(DCCP_MIB_OUTRSTS); return; diff --git a/net/ipv6/inet6_connection_sock.c b/net/ipv6/inet6_connection_sock.c index 7396e75e161b..75c308239243 100644 --- a/net/ipv6/inet6_connection_sock.c +++ b/net/ipv6/inet6_connection_sock.c @@ -176,7 +176,7 @@ int inet6_csk_xmit(struct sock *sk, struct sk_buff *skb, struct flowi *fl_unused /* Restore final destination back after routing done */ fl6.daddr = sk->sk_v6_daddr; - res = ip6_xmit(sk, skb, &fl6, rcu_dereference(np->opt), + res = ip6_xmit(sk, skb, &fl6, sk->sk_mark, rcu_dereference(np->opt), np->tclass); rcu_read_unlock(); return res; diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c index 38122d04fadc..2c0df09e9036 100644 --- a/net/ipv6/ip6_output.c +++ b/net/ipv6/ip6_output.c @@ -172,7 +172,7 @@ int ip6_output(struct net *net, struct sock *sk, struct sk_buff *skb) * which are using proper atomic operations or spinlocks. */ int ip6_xmit(const struct sock *sk, struct sk_buff *skb, struct flowi6 *fl6, - struct ipv6_txoptions *opt, int tclass) + __u32 mark, struct ipv6_txoptions *opt, int tclass) { struct net *net = sock_net(sk); const struct ipv6_pinfo *np = inet6_sk(sk); @@ -240,7 +240,7 @@ int ip6_xmit(const struct sock *sk, struct sk_buff *skb, struct flowi6 *fl6, skb->protocol = htons(ETH_P_IPV6); skb->priority = sk->sk_priority; - skb->mark = sk->sk_mark; + skb->mark = mark; mtu = dst_mtu(dst); if ((skb->len <= mtu) || skb->ignore_df || skb_is_gso(skb)) { diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index 2b20622a5824..cb8929681dc7 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -469,7 +469,7 @@ static int tcp_v6_send_synack(const struct sock *sk, struct dst_entry *dst, opt = ireq->ipv6_opt; if (!opt) opt = rcu_dereference(np->opt); - err = ip6_xmit(sk, skb, fl6, opt, np->tclass); + err = ip6_xmit(sk, skb, fl6, sk->sk_mark, opt, np->tclass); rcu_read_unlock(); err = net_xmit_eval(err); } @@ -840,8 +840,7 @@ static void tcp_v6_send_response(const struct sock *sk, struct sk_buff *skb, u32 dst = ip6_dst_lookup_flow(ctl_sk, &fl6, NULL); if (!IS_ERR(dst)) { skb_dst_set(buff, dst); - ctl_sk->sk_mark = fl6.flowi6_mark; - ip6_xmit(ctl_sk, buff, &fl6, NULL, tclass); + ip6_xmit(ctl_sk, buff, &fl6, fl6.flowi6_mark, NULL, tclass); TCP_INC_STATS(net, TCP_MIB_OUTSEGS); if (rst) TCP_INC_STATS(net, TCP_MIB_OUTRSTS); diff --git a/net/sctp/ipv6.c b/net/sctp/ipv6.c index 5ed8e79bf102..64dfd35ccdcc 100644 --- a/net/sctp/ipv6.c +++ b/net/sctp/ipv6.c @@ -222,7 +222,8 @@ static int sctp_v6_xmit(struct sk_buff *skb, struct sctp_transport *transport) SCTP_INC_STATS(sock_net(sk), SCTP_MIB_OUTSCTPPACKS); rcu_read_lock(); - res = ip6_xmit(sk, skb, fl6, rcu_dereference(np->opt), np->tclass); + res = ip6_xmit(sk, skb, fl6, sk->sk_mark, rcu_dereference(np->opt), + np->tclass); rcu_read_unlock(); return res; } From ab729823ec16aef384f09fd2cffe0b3d3f6e6cba Mon Sep 17 00:00:00 2001 From: Andreas Schultz Date: Fri, 27 Jan 2017 10:40:56 +0100 Subject: [PATCH 322/343] gtp: add genl family modules alias Auto-load the module when userspace asks for the gtp netlink family. Signed-off-by: Andreas Schultz Acked-by: Harald Welte Acked-by: Pablo Neira Ayuso Signed-off-by: David S. Miller --- drivers/net/gtp.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/net/gtp.c b/drivers/net/gtp.c index 8b6810bad54b..7580ccc50e1d 100644 --- a/drivers/net/gtp.c +++ b/drivers/net/gtp.c @@ -1376,3 +1376,4 @@ MODULE_LICENSE("GPL"); MODULE_AUTHOR("Harald Welte "); MODULE_DESCRIPTION("Interface driver for GTP encapsulated traffic"); MODULE_ALIAS_RTNL_LINK("gtp"); +MODULE_ALIAS_GENL_FAMILY("gtp"); From c6ce1d08eede4c2968ed08aafa3165e8e183c5a1 Mon Sep 17 00:00:00 2001 From: Andreas Schultz Date: Fri, 27 Jan 2017 10:40:57 +0100 Subject: [PATCH 323/343] gtp: clear DF bit on GTP packet tx 3GPP TS 29.281 and 3GPP TS 29.060 imply that GTP-U packets should be sent with the DF bit cleared. For example 3GPP TS 29.060, Release 8, Section 13.2.2: > Backbone router: Any router in the backbone may fragment the GTP > packet if needed, according to IPv4. Signed-off-by: Andreas Schultz Acked-by: Harald Welte Acked-by: Pablo Neira Ayuso Signed-off-by: David S. Miller --- drivers/net/gtp.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/gtp.c b/drivers/net/gtp.c index 7580ccc50e1d..1df54d626f20 100644 --- a/drivers/net/gtp.c +++ b/drivers/net/gtp.c @@ -612,7 +612,7 @@ static netdev_tx_t gtp_dev_xmit(struct sk_buff *skb, struct net_device *dev) pktinfo.fl4.saddr, pktinfo.fl4.daddr, pktinfo.iph->tos, ip4_dst_hoplimit(&pktinfo.rt->dst), - htons(IP_DF), + 0, pktinfo.gtph_port, pktinfo.gtph_port, true, false); break; From 3ab1b469e847ba425af3c5ad5068cc94b55b38d0 Mon Sep 17 00:00:00 2001 From: Andreas Schultz Date: Fri, 27 Jan 2017 10:40:58 +0100 Subject: [PATCH 324/343] gtp: fix cross netns recv on gtp socket The use of the passed through netlink src_net to check for a cross netns operation was wrong. Using the GTP socket and the GTP netdevice is always correct (even if the netdev has been moved to new netns after link creation). Remove the now obsolete net field from gtp_dev. Signed-off-by: Andreas Schultz Acked-by: Pablo Neira Ayuso Signed-off-by: David S. Miller --- drivers/net/gtp.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-) diff --git a/drivers/net/gtp.c b/drivers/net/gtp.c index 1df54d626f20..99d3df788ce8 100644 --- a/drivers/net/gtp.c +++ b/drivers/net/gtp.c @@ -69,7 +69,6 @@ struct gtp_dev { struct socket *sock0; struct socket *sock1u; - struct net *net; struct net_device *dev; unsigned int hash_size; @@ -316,7 +315,7 @@ static int gtp_encap_recv(struct sock *sk, struct sk_buff *skb) netdev_dbg(gtp->dev, "encap_recv sk=%p\n", sk); - xnet = !net_eq(gtp->net, dev_net(gtp->dev)); + xnet = !net_eq(sock_net(sk), dev_net(gtp->dev)); switch (udp_sk(sk)->encap_type) { case UDP_ENCAP_GTP0: @@ -658,7 +657,7 @@ static void gtp_link_setup(struct net_device *dev) static int gtp_hashtable_new(struct gtp_dev *gtp, int hsize); static void gtp_hashtable_free(struct gtp_dev *gtp); static int gtp_encap_enable(struct net_device *dev, struct gtp_dev *gtp, - int fd_gtp0, int fd_gtp1, struct net *src_net); + int fd_gtp0, int fd_gtp1); static int gtp_newlink(struct net *src_net, struct net_device *dev, struct nlattr *tb[], struct nlattr *data[]) @@ -675,7 +674,7 @@ static int gtp_newlink(struct net *src_net, struct net_device *dev, fd0 = nla_get_u32(data[IFLA_GTP_FD0]); fd1 = nla_get_u32(data[IFLA_GTP_FD1]); - err = gtp_encap_enable(dev, gtp, fd0, fd1, src_net); + err = gtp_encap_enable(dev, gtp, fd0, fd1); if (err < 0) goto out_err; @@ -821,7 +820,7 @@ static void gtp_hashtable_free(struct gtp_dev *gtp) } static int gtp_encap_enable(struct net_device *dev, struct gtp_dev *gtp, - int fd_gtp0, int fd_gtp1, struct net *src_net) + int fd_gtp0, int fd_gtp1) { struct udp_tunnel_sock_cfg tuncfg = {NULL}; struct socket *sock0, *sock1u; @@ -858,7 +857,6 @@ static int gtp_encap_enable(struct net_device *dev, struct gtp_dev *gtp, gtp->sock0 = sock0; gtp->sock1u = sock1u; - gtp->net = src_net; tuncfg.sk_user_data = gtp; tuncfg.encap_rcv = gtp_encap_recv; From 9d162ed69f51cbd9ee5a0c7e82aba7acc96362ff Mon Sep 17 00:00:00 2001 From: Sean Nyekjaer Date: Fri, 27 Jan 2017 08:46:23 +0100 Subject: [PATCH 325/343] net: phy: micrel: add support for KSZ8795 This is adds support for the PHYs in the KSZ8795 5port managed switch. It will allow to detect the link between the switch and the soc and uses the same read_status functions as the KSZ8873MLL switch. Signed-off-by: Sean Nyekjaer Signed-off-by: David S. Miller --- drivers/net/phy/micrel.c | 14 ++++++++++++++ include/linux/micrel_phy.h | 2 ++ 2 files changed, 16 insertions(+) diff --git a/drivers/net/phy/micrel.c b/drivers/net/phy/micrel.c index 9a77289109b7..e55809c5beb7 100644 --- a/drivers/net/phy/micrel.c +++ b/drivers/net/phy/micrel.c @@ -1008,6 +1008,20 @@ static struct phy_driver ksphy_driver[] = { .get_stats = kszphy_get_stats, .suspend = genphy_suspend, .resume = genphy_resume, +}, { + .phy_id = PHY_ID_KSZ8795, + .phy_id_mask = MICREL_PHY_ID_MASK, + .name = "Micrel KSZ8795", + .features = (SUPPORTED_Pause | SUPPORTED_Asym_Pause), + .flags = PHY_HAS_MAGICANEG | PHY_HAS_INTERRUPT, + .config_init = kszphy_config_init, + .config_aneg = ksz8873mll_config_aneg, + .read_status = ksz8873mll_read_status, + .get_sset_count = kszphy_get_sset_count, + .get_strings = kszphy_get_strings, + .get_stats = kszphy_get_stats, + .suspend = genphy_suspend, + .resume = genphy_resume, } }; module_phy_driver(ksphy_driver); diff --git a/include/linux/micrel_phy.h b/include/linux/micrel_phy.h index 257173e0095e..f541da68d1e7 100644 --- a/include/linux/micrel_phy.h +++ b/include/linux/micrel_phy.h @@ -35,6 +35,8 @@ #define PHY_ID_KSZ886X 0x00221430 #define PHY_ID_KSZ8863 0x00221435 +#define PHY_ID_KSZ8795 0x00221550 + /* struct phy_device dev_flags definitions */ #define MICREL_PHY_50MHZ_CLK 0x00000001 #define MICREL_PHY_FXEN 0x00000002 From 950eabbd6ddedc1b08350b9169a6a51b130ebaaf Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Fri, 27 Jan 2017 13:32:14 +0100 Subject: [PATCH 326/343] ISDN: eicon: silence misleading array-bounds warning With some gcc versions, we get a warning about the eicon driver, and that currently shows up as the only remaining warning in one of the build bots: In file included from ../drivers/isdn/hardware/eicon/message.c:30:0: eicon/message.c: In function 'mixer_notify_update': eicon/platform.h:333:18: warning: array subscript is above array bounds [-Warray-bounds] The code is easily changed to open-code the unusual PUT_WORD() line causing this to avoid the warning. Cc: stable@vger.kernel.org Link: http://arm-soc.lixom.net/buildlogs/stable-rc/v4.4.45/ Signed-off-by: Arnd Bergmann Signed-off-by: David S. Miller --- drivers/isdn/hardware/eicon/message.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/isdn/hardware/eicon/message.c b/drivers/isdn/hardware/eicon/message.c index 1a1d99704fe6..296f1411fe84 100644 --- a/drivers/isdn/hardware/eicon/message.c +++ b/drivers/isdn/hardware/eicon/message.c @@ -11297,7 +11297,8 @@ static void mixer_notify_update(PLCI *plci, byte others) ((CAPI_MSG *) msg)->header.ncci = 0; ((CAPI_MSG *) msg)->info.facility_req.Selector = SELECTOR_LINE_INTERCONNECT; ((CAPI_MSG *) msg)->info.facility_req.structs[0] = 3; - PUT_WORD(&(((CAPI_MSG *) msg)->info.facility_req.structs[1]), LI_REQ_SILENT_UPDATE); + ((CAPI_MSG *) msg)->info.facility_req.structs[1] = LI_REQ_SILENT_UPDATE & 0xff; + ((CAPI_MSG *) msg)->info.facility_req.structs[2] = LI_REQ_SILENT_UPDATE >> 8; ((CAPI_MSG *) msg)->info.facility_req.structs[3] = 0; w = api_put(notify_plci->appl, (CAPI_MSG *) msg); if (w != _QUEUE_FULL) From e0d76fa4475ef2cf4b52d18588b8ce95153d021b Mon Sep 17 00:00:00 2001 From: Brian Foster Date: Thu, 26 Jan 2017 13:18:09 -0800 Subject: [PATCH 327/343] xfs: prevent quotacheck from overloading inode lru Quotacheck runs at mount time in situations where quota accounting must be recalculated. In doing so, it uses bulkstat to visit every inode in the filesystem. Historically, every inode processed during quotacheck was released and immediately tagged for reclaim because quotacheck runs before the superblock is marked active by the VFS. In other words, the final iput() lead to an immediate ->destroy_inode() call, which allowed the XFS background reclaim worker to start reclaiming inodes. Commit 17c12bcd3 ("xfs: when replaying bmap operations, don't let unlinked inodes get reaped") marks the XFS superblock active sooner as part of the mount process to support caching inodes processed during log recovery. This occurs before quotacheck and thus means all inodes processed by quotacheck are inserted to the LRU on release. The s_umount lock is held until the mount has completed and thus prevents the shrinkers from operating on the sb. This means that quotacheck can excessively populate the inode LRU and lead to OOM conditions on systems without sufficient RAM. Update the quotacheck bulkstat handler to set XFS_IGET_DONTCACHE on inodes processed by quotacheck. This causes ->drop_inode() to return 1 and in turn causes iput_final() to evict the inode. This preserves the original quotacheck behavior and prevents it from overloading the LRU and running out of memory. CC: stable@vger.kernel.org # v4.9 Reported-by: Martin Svec Signed-off-by: Brian Foster Reviewed-by: Eric Sandeen Reviewed-by: Darrick J. Wong Signed-off-by: Darrick J. Wong --- fs/xfs/xfs_qm.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c index 45e50ea90769..b669b123287b 100644 --- a/fs/xfs/xfs_qm.c +++ b/fs/xfs/xfs_qm.c @@ -1177,7 +1177,8 @@ xfs_qm_dqusage_adjust( * the case in all other instances. It's OK that we do this because * quotacheck is done only at mount time. */ - error = xfs_iget(mp, NULL, ino, 0, XFS_ILOCK_EXCL, &ip); + error = xfs_iget(mp, NULL, ino, XFS_IGET_DONTCACHE, XFS_ILOCK_EXCL, + &ip); if (error) { *res = BULKSTAT_RV_NOTHING; return error; From 9aed02feae57bf7a40cb04ea0e3017cb7a998db4 Mon Sep 17 00:00:00 2001 From: Vineet Gupta Date: Fri, 27 Jan 2017 10:45:27 -0800 Subject: [PATCH 328/343] ARC: [arcompact] handle unaligned access delay slot corner case After emulating an unaligned access in delay slot of a branch, we pretend as the delay slot never happened - so return back to actual branch target (or next PC if branch was not taken). Curently we did this by handling STATUS32.DE, we also need to clear the BTA.T bit, which is disregarded when returning from original misaligned exception, but could cause weirdness if it took the interrupt return path (in case interrupt was acive too) One ARC700 customer ran into this when enabling unaligned access fixup for kernel mode accesses as well Cc: stable@vger.kernel.org Signed-off-by: Vineet Gupta --- arch/arc/kernel/unaligned.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/arch/arc/kernel/unaligned.c b/arch/arc/kernel/unaligned.c index abd961f3e763..91ebe382147f 100644 --- a/arch/arc/kernel/unaligned.c +++ b/arch/arc/kernel/unaligned.c @@ -241,8 +241,9 @@ int misaligned_fixup(unsigned long address, struct pt_regs *regs, if (state.fault) goto fault; + /* clear any remanants of delay slot */ if (delay_mode(regs)) { - regs->ret = regs->bta; + regs->ret = regs->bta ~1U; regs->status32 &= ~STATUS_DE_MASK; } else { regs->ret += state.instr_len; From b4cfe3971f6eab542dd7ecc398bfa1aeec889934 Mon Sep 17 00:00:00 2001 From: Jack Morgenstein Date: Sun, 15 Jan 2017 20:15:00 +0200 Subject: [PATCH 329/343] RDMA/cma: Fix unknown symbol when CONFIG_IPV6 is not enabled If IPV6 has not been enabled in the underlying kernel, we must avoid calling IPV6 procedures in rdma_cm.ko. This requires using "IS_ENABLED(CONFIG_IPV6)" in "if" statements surrounding any code which calls external IPV6 procedures. In the instance fixed here, procedure cma_bind_addr() called ipv6_addr_type() -- which resulted in calling external procedure __ipv6_addr_type(). Fixes: 6c26a77124ff ("RDMA/cma: fix IPv6 address resolution") Cc: # v4.2+ Cc: Spencer Baugh Signed-off-by: Jack Morgenstein Reviewed-by: Moni Shoua Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- drivers/infiniband/core/cma.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/infiniband/core/cma.c b/drivers/infiniband/core/cma.c index e7dcfac877ca..3e70a9c5d79d 100644 --- a/drivers/infiniband/core/cma.c +++ b/drivers/infiniband/core/cma.c @@ -2811,7 +2811,8 @@ static int cma_bind_addr(struct rdma_cm_id *id, struct sockaddr *src_addr, if (!src_addr || !src_addr->sa_family) { src_addr = (struct sockaddr *) &id->route.addr.src_addr; src_addr->sa_family = dst_addr->sa_family; - if (dst_addr->sa_family == AF_INET6) { + if (IS_ENABLED(CONFIG_IPV6) && + dst_addr->sa_family == AF_INET6) { struct sockaddr_in6 *src_addr6 = (struct sockaddr_in6 *) src_addr; struct sockaddr_in6 *dst_addr6 = (struct sockaddr_in6 *) dst_addr; src_addr6->sin6_scope_id = dst_addr6->sin6_scope_id; From 83b5d1e3d3013dbf90645a5d07179d018c8243fa Mon Sep 17 00:00:00 2001 From: Helge Deller Date: Tue, 3 Jan 2017 22:55:50 +0100 Subject: [PATCH 330/343] parisc, parport_gsc: Fixes for printk continuation lines Signed-off-by: Helge Deller --- drivers/parport/parport_gsc.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/parport/parport_gsc.c b/drivers/parport/parport_gsc.c index dd6d4ccb41e4..3858b87fd0bb 100644 --- a/drivers/parport/parport_gsc.c +++ b/drivers/parport/parport_gsc.c @@ -293,7 +293,7 @@ struct parport *parport_gsc_probe_port(unsigned long base, p->irq = PARPORT_IRQ_NONE; } if (p->irq != PARPORT_IRQ_NONE) { - printk(", irq %d", p->irq); + pr_cont(", irq %d", p->irq); if (p->dma == PARPORT_DMA_AUTO) { p->dma = PARPORT_DMA_NONE; @@ -303,8 +303,8 @@ struct parport *parport_gsc_probe_port(unsigned long base, is mandatory (see above) */ p->dma = PARPORT_DMA_NONE; - printk(" ["); -#define printmode(x) {if(p->modes&PARPORT_MODE_##x){printk("%s%s",f?",":"",#x);f++;}} + pr_cont(" ["); +#define printmode(x) {if(p->modes&PARPORT_MODE_##x){pr_cont("%s%s",f?",":"",#x);f++;}} { int f = 0; printmode(PCSPP); @@ -315,7 +315,7 @@ struct parport *parport_gsc_probe_port(unsigned long base, // printmode(DMA); } #undef printmode - printk("]\n"); + pr_cont("]\n"); if (p->irq != PARPORT_IRQ_NONE) { if (request_irq (p->irq, parport_irq_handler, From 2ad5d52d42810bed95100a3d912679d8864421ec Mon Sep 17 00:00:00 2001 From: Helge Deller Date: Sat, 28 Jan 2017 11:52:02 +0100 Subject: [PATCH 331/343] parisc: Don't use BITS_PER_LONG in userspace-exported swab.h header In swab.h the "#if BITS_PER_LONG > 32" breaks compiling userspace programs if BITS_PER_LONG is #defined by userspace with the sizeof() compiler builtin. Solve this problem by using __BITS_PER_LONG instead. Since we now #include asm/bitsperlong.h avoid further potential userspace pollution by moving the #define of SHIFT_PER_LONG to bitops.h which is not exported to userspace. This patch unbreaks compiling qemu on hppa/parisc. Signed-off-by: Helge Deller Cc: --- arch/parisc/include/asm/bitops.h | 8 +++++++- arch/parisc/include/uapi/asm/bitsperlong.h | 2 -- arch/parisc/include/uapi/asm/swab.h | 5 +++-- 3 files changed, 10 insertions(+), 5 deletions(-) diff --git a/arch/parisc/include/asm/bitops.h b/arch/parisc/include/asm/bitops.h index 3f9406d9b9d6..da87943328a5 100644 --- a/arch/parisc/include/asm/bitops.h +++ b/arch/parisc/include/asm/bitops.h @@ -6,7 +6,7 @@ #endif #include -#include /* for BITS_PER_LONG/SHIFT_PER_LONG */ +#include #include #include #include @@ -17,6 +17,12 @@ * to include/asm-i386/bitops.h or kerneldoc */ +#if __BITS_PER_LONG == 64 +#define SHIFT_PER_LONG 6 +#else +#define SHIFT_PER_LONG 5 +#endif + #define CHOP_SHIFTCOUNT(x) (((unsigned long) (x)) & (BITS_PER_LONG - 1)) diff --git a/arch/parisc/include/uapi/asm/bitsperlong.h b/arch/parisc/include/uapi/asm/bitsperlong.h index e0a23c7bdd43..07fa7e50bdc0 100644 --- a/arch/parisc/include/uapi/asm/bitsperlong.h +++ b/arch/parisc/include/uapi/asm/bitsperlong.h @@ -3,10 +3,8 @@ #if defined(__LP64__) #define __BITS_PER_LONG 64 -#define SHIFT_PER_LONG 6 #else #define __BITS_PER_LONG 32 -#define SHIFT_PER_LONG 5 #endif #include diff --git a/arch/parisc/include/uapi/asm/swab.h b/arch/parisc/include/uapi/asm/swab.h index e78403b129ef..928e1bbac98f 100644 --- a/arch/parisc/include/uapi/asm/swab.h +++ b/arch/parisc/include/uapi/asm/swab.h @@ -1,6 +1,7 @@ #ifndef _PARISC_SWAB_H #define _PARISC_SWAB_H +#include #include #include @@ -38,7 +39,7 @@ static inline __attribute_const__ __u32 __arch_swab32(__u32 x) } #define __arch_swab32 __arch_swab32 -#if BITS_PER_LONG > 32 +#if __BITS_PER_LONG > 32 /* ** From "PA-RISC 2.0 Architecture", HP Professional Books. ** See Appendix I page 8 , "Endian Byte Swapping". @@ -61,6 +62,6 @@ static inline __attribute_const__ __u64 __arch_swab64(__u64 x) return x; } #define __arch_swab64 __arch_swab64 -#endif /* BITS_PER_LONG > 32 */ +#endif /* __BITS_PER_LONG > 32 */ #endif /* _PARISC_SWAB_H */ From 39cb2c9a316e77f6dfba96c543e55b6672d5a37e Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Sun, 29 Jan 2017 13:50:06 -0800 Subject: [PATCH 332/343] drm/i915: Check for NULL i915_vma in intel_unpin_fb_obj() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit I've seen this trigger twice now, where the i915_gem_object_to_ggtt() call in intel_unpin_fb_obj() returns NULL, resulting in an oops immediately afterwards as the (inlined) call to i915_vma_unpin_fence() tries to dereference it. It seems to be some race condition where the object is going away at shutdown time, since both times happened when shutting down the X server. The call chains were different: - VT ioctl(KDSETMODE, KD_TEXT): intel_cleanup_plane_fb+0x5b/0xa0 [i915] drm_atomic_helper_cleanup_planes+0x6f/0x90 [drm_kms_helper] intel_atomic_commit_tail+0x749/0xfe0 [i915] intel_atomic_commit+0x3cb/0x4f0 [i915] drm_atomic_commit+0x4b/0x50 [drm] restore_fbdev_mode+0x14c/0x2a0 [drm_kms_helper] drm_fb_helper_restore_fbdev_mode_unlocked+0x34/0x80 [drm_kms_helper] drm_fb_helper_set_par+0x2d/0x60 [drm_kms_helper] intel_fbdev_set_par+0x18/0x70 [i915] fb_set_var+0x236/0x460 fbcon_blank+0x30f/0x350 do_unblank_screen+0xd2/0x1a0 vt_ioctl+0x507/0x12a0 tty_ioctl+0x355/0xc30 do_vfs_ioctl+0xa3/0x5e0 SyS_ioctl+0x79/0x90 entry_SYSCALL_64_fastpath+0x13/0x94 - i915 unpin_work workqueue: intel_unpin_work_fn+0x58/0x140 [i915] process_one_work+0x1f1/0x480 worker_thread+0x48/0x4d0 kthread+0x101/0x140 and this patch purely papers over the issue by adding a NULL pointer check and a WARN_ON_ONCE() to avoid the oops that would then generally make the machine unresponsive. Other callers of i915_gem_object_to_ggtt() seem to also check for the returned pointer being NULL and warn about it, so this clearly has happened before in other places. [ Reported it originally to the i915 developers on Jan 8, applying the ugly workaround on my own now after triggering the problem for the second time with no feedback. This is likely to be the same bug reported as https://bugs.freedesktop.org/show_bug.cgi?id=98829 https://bugs.freedesktop.org/show_bug.cgi?id=99134 which has a patch for the underlying problem, but it hasn't gotten to me, so I'm applying the workaround. ] Cc: Daniel Vetter Cc: Jani Nikula Cc: Ville Syrjälä Cc: Chris Wilson Cc: Maarten Lankhorst Cc: Tvrtko Ursulin Cc: Imre Deak Signed-off-by: Linus Torvalds --- drivers/gpu/drm/i915/intel_display.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 77f7b1d849a4..f0b9aa7a0483 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -2251,6 +2251,9 @@ void intel_unpin_fb_obj(struct drm_framebuffer *fb, unsigned int rotation) intel_fill_fb_ggtt_view(&view, fb, rotation); vma = i915_gem_object_to_ggtt(obj, &view); + if (WARN_ON_ONCE(!vma)) + return; + i915_vma_unpin_fence(vma); i915_gem_object_unpin_from_display_plane(vma); } From 566cf877a1fcb6d6dc0126b076aad062054c2637 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Sun, 29 Jan 2017 14:25:17 -0800 Subject: [PATCH 333/343] Linux 4.10-rc6 --- Makefile | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/Makefile b/Makefile index 098840012b9b..96b27a888285 100644 --- a/Makefile +++ b/Makefile @@ -1,8 +1,8 @@ VERSION = 4 PATCHLEVEL = 10 SUBLEVEL = 0 -EXTRAVERSION = -rc5 -NAME = Anniversary Edition +EXTRAVERSION = -rc6 +NAME = Fearless Coyote # *DOCUMENTATION* # To see a list of typical targets execute "make help" From 61babe943938bb41fcd7e2b1dec7d1537ea3892f Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Mon, 21 Nov 2016 18:32:08 -0800 Subject: [PATCH 334/343] mtd: nand: fix nand.h kernel-doc warnings Fix kernel-doc warnings in : ..//include/linux/mtd/nand.h:658: warning: No description found for parameter 'tCEH_min' ..//include/linux/mtd/nand.h:877: warning: No description found for parameter 'data_interface' Fixes: eee64b700e26 ("mtd: nand: Introduce nand_data_interface") Signed-off-by: Randy Dunlap Signed-off-by: Boris Brezillon --- include/linux/mtd/nand.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/include/linux/mtd/nand.h b/include/linux/mtd/nand.h index c5f3a012ae62..f67915c7726f 100644 --- a/include/linux/mtd/nand.h +++ b/include/linux/mtd/nand.h @@ -615,7 +615,7 @@ struct nand_buffers { * @tALS_min: ALE setup time * @tAR_min: ALE to RE# delay * @tCEA_max: CE# access time - * @tCEH_min: + * @tCEH_min: CE# high hold time * @tCH_min: CE# hold time * @tCHZ_max: CE# high to output hi-Z * @tCLH_min: CLE hold time @@ -801,6 +801,7 @@ nand_get_sdr_timings(const struct nand_data_interface *conf) * supported, 0 otherwise. * @jedec_params: [INTERN] holds the JEDEC parameter page when JEDEC is * supported, 0 otherwise. + * @data_interface: [INTERN] NAND interface timing information * @read_retries: [INTERN] the number of read retry modes supported * @onfi_set_features: [REPLACEABLE] set the features for ONFI nand * @onfi_get_features: [REPLACEABLE] get the features for ONFI nand From 4339b7fdbb90ab579862b0fcde6ee0f342f12692 Mon Sep 17 00:00:00 2001 From: Fabio Estevam Date: Tue, 29 Nov 2016 13:28:52 -0200 Subject: [PATCH 335/343] mtd: nand: lpc32xx_slc: Remove unneeded NULL check on 'rc' devm_ioremap_resource() does a NULL check on the 'rc' argument, so remove the unneeded manual NULL check. While at it, place the 'rc' assignment just before devm_ioremap_resource() to improve readability. Signed-off-by: Fabio Estevam Acked-by: Marek Vasut Acked-by: Vladimir Zapolskiy Signed-off-by: Boris Brezillon --- drivers/mtd/nand/lpc32xx_slc.c | 9 ++------- 1 file changed, 2 insertions(+), 7 deletions(-) diff --git a/drivers/mtd/nand/lpc32xx_slc.c b/drivers/mtd/nand/lpc32xx_slc.c index 53bafe23ab39..a0669a33f8fe 100644 --- a/drivers/mtd/nand/lpc32xx_slc.c +++ b/drivers/mtd/nand/lpc32xx_slc.c @@ -797,22 +797,17 @@ static int lpc32xx_nand_probe(struct platform_device *pdev) struct resource *rc; int res; - rc = platform_get_resource(pdev, IORESOURCE_MEM, 0); - if (rc == NULL) { - dev_err(&pdev->dev, "No memory resource found for device\n"); - return -EBUSY; - } - /* Allocate memory for the device structure (and zero it) */ host = devm_kzalloc(&pdev->dev, sizeof(*host), GFP_KERNEL); if (!host) return -ENOMEM; - host->io_base_dma = rc->start; + rc = platform_get_resource(pdev, IORESOURCE_MEM, 0); host->io_base = devm_ioremap_resource(&pdev->dev, rc); if (IS_ERR(host->io_base)) return PTR_ERR(host->io_base); + host->io_base_dma = rc->start; if (pdev->dev.of_node) host->ncfg = lpc32xx_parse_dt(&pdev->dev); if (!host->ncfg) { From 4404d7d821c33ac8105f1d52deb60f736d7c6a06 Mon Sep 17 00:00:00 2001 From: Linus Walleij Date: Sun, 18 Dec 2016 12:34:55 +0100 Subject: [PATCH 336/343] mtd: nand: fsmc: remove stale non-DT probe path The FSMC driver has an execution path and a header file in that serves to support passing in platform data through board files, albeit no upstream users of this mechanism exist. The header file also contains function headers for functions that do not exist in the kernel. Delete this and move the platform data struct, parsing and handling into the driver, assume we are using OF and make the driver depend on OF, remove the ifdefs making that optional. Cc: Viresh Kumar Cc: Stefan Roese Cc: Vipin Kumar Signed-off-by: Linus Walleij Reviewed-by: Stefan Roese Acked-by: Viresh Kumar Signed-off-by: Boris Brezillon --- drivers/mtd/nand/Kconfig | 1 + drivers/mtd/nand/fsmc_nand.c | 153 +++++++++++++++++++++++++++++----- include/linux/mtd/fsmc.h | 156 ----------------------------------- 3 files changed, 133 insertions(+), 177 deletions(-) delete mode 100644 include/linux/mtd/fsmc.h diff --git a/drivers/mtd/nand/Kconfig b/drivers/mtd/nand/Kconfig index 9ce5dcb4abd0..c6585454f102 100644 --- a/drivers/mtd/nand/Kconfig +++ b/drivers/mtd/nand/Kconfig @@ -535,6 +535,7 @@ config MTD_NAND_JZ4780 config MTD_NAND_FSMC tristate "Support for NAND on ST Micros FSMC" + depends on OF depends on PLAT_SPEAR || ARCH_NOMADIK || ARCH_U8500 || MACH_U300 help Enables support for NAND Flash chips on the ST Microelectronics diff --git a/drivers/mtd/nand/fsmc_nand.c b/drivers/mtd/nand/fsmc_nand.c index 4924b43977ef..bda1e4667138 100644 --- a/drivers/mtd/nand/fsmc_nand.c +++ b/drivers/mtd/nand/fsmc_nand.c @@ -35,10 +35,133 @@ #include #include #include -#include #include #include +#define FSMC_NAND_BW8 1 +#define FSMC_NAND_BW16 2 + +#define FSMC_MAX_NOR_BANKS 4 +#define FSMC_MAX_NAND_BANKS 4 + +#define FSMC_FLASH_WIDTH8 1 +#define FSMC_FLASH_WIDTH16 2 + +/* fsmc controller registers for NOR flash */ +#define CTRL 0x0 + /* ctrl register definitions */ + #define BANK_ENABLE (1 << 0) + #define MUXED (1 << 1) + #define NOR_DEV (2 << 2) + #define WIDTH_8 (0 << 4) + #define WIDTH_16 (1 << 4) + #define RSTPWRDWN (1 << 6) + #define WPROT (1 << 7) + #define WRT_ENABLE (1 << 12) + #define WAIT_ENB (1 << 13) + +#define CTRL_TIM 0x4 + /* ctrl_tim register definitions */ + +#define FSMC_NOR_BANK_SZ 0x8 +#define FSMC_NOR_REG_SIZE 0x40 + +#define FSMC_NOR_REG(base, bank, reg) (base + \ + FSMC_NOR_BANK_SZ * (bank) + \ + reg) + +/* fsmc controller registers for NAND flash */ +#define PC 0x00 + /* pc register definitions */ + #define FSMC_RESET (1 << 0) + #define FSMC_WAITON (1 << 1) + #define FSMC_ENABLE (1 << 2) + #define FSMC_DEVTYPE_NAND (1 << 3) + #define FSMC_DEVWID_8 (0 << 4) + #define FSMC_DEVWID_16 (1 << 4) + #define FSMC_ECCEN (1 << 6) + #define FSMC_ECCPLEN_512 (0 << 7) + #define FSMC_ECCPLEN_256 (1 << 7) + #define FSMC_TCLR_1 (1) + #define FSMC_TCLR_SHIFT (9) + #define FSMC_TCLR_MASK (0xF) + #define FSMC_TAR_1 (1) + #define FSMC_TAR_SHIFT (13) + #define FSMC_TAR_MASK (0xF) +#define STS 0x04 + /* sts register definitions */ + #define FSMC_CODE_RDY (1 << 15) +#define COMM 0x08 + /* comm register definitions */ + #define FSMC_TSET_0 0 + #define FSMC_TSET_SHIFT 0 + #define FSMC_TSET_MASK 0xFF + #define FSMC_TWAIT_6 6 + #define FSMC_TWAIT_SHIFT 8 + #define FSMC_TWAIT_MASK 0xFF + #define FSMC_THOLD_4 4 + #define FSMC_THOLD_SHIFT 16 + #define FSMC_THOLD_MASK 0xFF + #define FSMC_THIZ_1 1 + #define FSMC_THIZ_SHIFT 24 + #define FSMC_THIZ_MASK 0xFF +#define ATTRIB 0x0C +#define IOATA 0x10 +#define ECC1 0x14 +#define ECC2 0x18 +#define ECC3 0x1C +#define FSMC_NAND_BANK_SZ 0x20 + +#define FSMC_NAND_REG(base, bank, reg) (base + FSMC_NOR_REG_SIZE + \ + (FSMC_NAND_BANK_SZ * (bank)) + \ + reg) + +#define FSMC_BUSY_WAIT_TIMEOUT (1 * HZ) + +struct fsmc_nand_timings { + uint8_t tclr; + uint8_t tar; + uint8_t thiz; + uint8_t thold; + uint8_t twait; + uint8_t tset; +}; + +enum access_mode { + USE_DMA_ACCESS = 1, + USE_WORD_ACCESS, +}; + +/** + * fsmc_nand_platform_data - platform specific NAND controller config + * @nand_timings: timing setup for the physical NAND interface + * @partitions: partition table for the platform, use a default fallback + * if this is NULL + * @nr_partitions: the number of partitions in the previous entry + * @options: different options for the driver + * @width: bus width + * @bank: default bank + * @select_bank: callback to select a certain bank, this is + * platform-specific. If the controller only supports one bank + * this may be set to NULL + */ +struct fsmc_nand_platform_data { + struct fsmc_nand_timings *nand_timings; + struct mtd_partition *partitions; + unsigned int nr_partitions; + unsigned int options; + unsigned int width; + unsigned int bank; + + enum access_mode mode; + + void (*select_bank)(uint32_t bank, uint32_t busw); + + /* priv structures for dma accesses */ + void *read_dma_priv; + void *write_dma_priv; +}; + static int fsmc_ecc1_ooblayout_ecc(struct mtd_info *mtd, int section, struct mtd_oob_region *oobregion) { @@ -714,7 +837,6 @@ static bool filter(struct dma_chan *chan, void *slave) return true; } -#ifdef CONFIG_OF static int fsmc_nand_probe_config_dt(struct platform_device *pdev, struct device_node *np) { @@ -757,13 +879,6 @@ static int fsmc_nand_probe_config_dt(struct platform_device *pdev, } return 0; } -#else -static int fsmc_nand_probe_config_dt(struct platform_device *pdev, - struct device_node *np) -{ - return -ENOSYS; -} -#endif /* * fsmc_nand_probe - Probe function @@ -782,19 +897,15 @@ static int __init fsmc_nand_probe(struct platform_device *pdev) u32 pid; int i; - if (np) { - pdata = devm_kzalloc(&pdev->dev, sizeof(*pdata), GFP_KERNEL); - pdev->dev.platform_data = pdata; - ret = fsmc_nand_probe_config_dt(pdev, np); - if (ret) { - dev_err(&pdev->dev, "no platform data\n"); - return -ENODEV; - } - } + pdata = devm_kzalloc(&pdev->dev, sizeof(*pdata), GFP_KERNEL); + if (!pdata) + return -ENOMEM; - if (!pdata) { - dev_err(&pdev->dev, "platform data is NULL\n"); - return -EINVAL; + pdev->dev.platform_data = pdata; + ret = fsmc_nand_probe_config_dt(pdev, np); + if (ret) { + dev_err(&pdev->dev, "no platform data\n"); + return -ENODEV; } /* Allocate memory for the device structure (and zero it) */ diff --git a/include/linux/mtd/fsmc.h b/include/linux/mtd/fsmc.h deleted file mode 100644 index ad3c3488073c..000000000000 --- a/include/linux/mtd/fsmc.h +++ /dev/null @@ -1,156 +0,0 @@ -/* - * incude/mtd/fsmc.h - * - * ST Microelectronics - * Flexible Static Memory Controller (FSMC) - * platform data interface and header file - * - * Copyright © 2010 ST Microelectronics - * Vipin Kumar - * - * This file is licensed under the terms of the GNU General Public - * License version 2. This program is licensed "as is" without any - * warranty of any kind, whether express or implied. - */ - -#ifndef __MTD_FSMC_H -#define __MTD_FSMC_H - -#include -#include -#include -#include -#include -#include - -#define FSMC_NAND_BW8 1 -#define FSMC_NAND_BW16 2 - -#define FSMC_MAX_NOR_BANKS 4 -#define FSMC_MAX_NAND_BANKS 4 - -#define FSMC_FLASH_WIDTH8 1 -#define FSMC_FLASH_WIDTH16 2 - -/* fsmc controller registers for NOR flash */ -#define CTRL 0x0 - /* ctrl register definitions */ - #define BANK_ENABLE (1 << 0) - #define MUXED (1 << 1) - #define NOR_DEV (2 << 2) - #define WIDTH_8 (0 << 4) - #define WIDTH_16 (1 << 4) - #define RSTPWRDWN (1 << 6) - #define WPROT (1 << 7) - #define WRT_ENABLE (1 << 12) - #define WAIT_ENB (1 << 13) - -#define CTRL_TIM 0x4 - /* ctrl_tim register definitions */ - -#define FSMC_NOR_BANK_SZ 0x8 -#define FSMC_NOR_REG_SIZE 0x40 - -#define FSMC_NOR_REG(base, bank, reg) (base + \ - FSMC_NOR_BANK_SZ * (bank) + \ - reg) - -/* fsmc controller registers for NAND flash */ -#define PC 0x00 - /* pc register definitions */ - #define FSMC_RESET (1 << 0) - #define FSMC_WAITON (1 << 1) - #define FSMC_ENABLE (1 << 2) - #define FSMC_DEVTYPE_NAND (1 << 3) - #define FSMC_DEVWID_8 (0 << 4) - #define FSMC_DEVWID_16 (1 << 4) - #define FSMC_ECCEN (1 << 6) - #define FSMC_ECCPLEN_512 (0 << 7) - #define FSMC_ECCPLEN_256 (1 << 7) - #define FSMC_TCLR_1 (1) - #define FSMC_TCLR_SHIFT (9) - #define FSMC_TCLR_MASK (0xF) - #define FSMC_TAR_1 (1) - #define FSMC_TAR_SHIFT (13) - #define FSMC_TAR_MASK (0xF) -#define STS 0x04 - /* sts register definitions */ - #define FSMC_CODE_RDY (1 << 15) -#define COMM 0x08 - /* comm register definitions */ - #define FSMC_TSET_0 0 - #define FSMC_TSET_SHIFT 0 - #define FSMC_TSET_MASK 0xFF - #define FSMC_TWAIT_6 6 - #define FSMC_TWAIT_SHIFT 8 - #define FSMC_TWAIT_MASK 0xFF - #define FSMC_THOLD_4 4 - #define FSMC_THOLD_SHIFT 16 - #define FSMC_THOLD_MASK 0xFF - #define FSMC_THIZ_1 1 - #define FSMC_THIZ_SHIFT 24 - #define FSMC_THIZ_MASK 0xFF -#define ATTRIB 0x0C -#define IOATA 0x10 -#define ECC1 0x14 -#define ECC2 0x18 -#define ECC3 0x1C -#define FSMC_NAND_BANK_SZ 0x20 - -#define FSMC_NAND_REG(base, bank, reg) (base + FSMC_NOR_REG_SIZE + \ - (FSMC_NAND_BANK_SZ * (bank)) + \ - reg) - -#define FSMC_BUSY_WAIT_TIMEOUT (1 * HZ) - -struct fsmc_nand_timings { - uint8_t tclr; - uint8_t tar; - uint8_t thiz; - uint8_t thold; - uint8_t twait; - uint8_t tset; -}; - -enum access_mode { - USE_DMA_ACCESS = 1, - USE_WORD_ACCESS, -}; - -/** - * fsmc_nand_platform_data - platform specific NAND controller config - * @nand_timings: timing setup for the physical NAND interface - * @partitions: partition table for the platform, use a default fallback - * if this is NULL - * @nr_partitions: the number of partitions in the previous entry - * @options: different options for the driver - * @width: bus width - * @bank: default bank - * @select_bank: callback to select a certain bank, this is - * platform-specific. If the controller only supports one bank - * this may be set to NULL - */ -struct fsmc_nand_platform_data { - struct fsmc_nand_timings *nand_timings; - struct mtd_partition *partitions; - unsigned int nr_partitions; - unsigned int options; - unsigned int width; - unsigned int bank; - - enum access_mode mode; - - void (*select_bank)(uint32_t bank, uint32_t busw); - - /* priv structures for dma accesses */ - void *read_dma_priv; - void *write_dma_priv; -}; - -extern int __init fsmc_nor_init(struct platform_device *pdev, - unsigned long base, uint32_t bank, uint32_t width); -extern void __init fsmc_init_board_info(struct platform_device *pdev, - struct mtd_partition *partitions, unsigned int nr_partitions, - unsigned int width); - -#endif /* __MTD_FSMC_H */ From de0b34455afd1204a3644c9eb25b0b3b0aeb2fe9 Mon Sep 17 00:00:00 2001 From: Jean Delvare Date: Mon, 9 Jan 2017 11:58:14 +0100 Subject: [PATCH 337/343] mtd: nand: Add OX820 NAND hardware dependency The oxnas NAND driver is only needed for a specific platform, do not propose it on other platforms unless build-testing. Signed-off-by: Jean Delvare Fixes: 668592492409 ("mtd: nand: Add OX820 NAND Support") Cc: Rob Herring Cc: Neil Armstrong Cc: Boris Brezillon Cc: Richard Weinberger Cc: David Woodhouse Cc: Brian Norris Cc: Marek Vasut Cc: Cyrille Pitchen Acked-by: Marek Vasut Acked-by: Neil Armstrong Signed-off-by: Boris Brezillon --- drivers/mtd/nand/Kconfig | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/mtd/nand/Kconfig b/drivers/mtd/nand/Kconfig index c6585454f102..6d4d5672d1d8 100644 --- a/drivers/mtd/nand/Kconfig +++ b/drivers/mtd/nand/Kconfig @@ -426,6 +426,7 @@ config MTD_NAND_ORION config MTD_NAND_OXNAS tristate "NAND Flash support for Oxford Semiconductor SoC" + depends on ARCH_OXNAS || COMPILE_TEST depends on HAS_IOMEM help This enables the NAND flash controller on Oxford Semiconductor SoCs. From 07f76443ef81e2b76bb6f7858ba183cd60357ada Mon Sep 17 00:00:00 2001 From: Wei Yongjun Date: Tue, 17 Jan 2017 14:22:46 +0000 Subject: [PATCH 338/343] mtd: nand: mediatek: remove redundant dev_err call in mtk_nfc_probe() There is a error message within devm_ioremap_resource already, so remove the dev_err call to avoid redundant error message. Signed-off-by: Wei Yongjun Acked-by: Marek Vasut Signed-off-by: Boris Brezillon --- drivers/mtd/nand/mtk_nand.c | 1 - 1 file changed, 1 deletion(-) diff --git a/drivers/mtd/nand/mtk_nand.c b/drivers/mtd/nand/mtk_nand.c index 6c3eed3c2094..6c517c682939 100644 --- a/drivers/mtd/nand/mtk_nand.c +++ b/drivers/mtd/nand/mtk_nand.c @@ -1383,7 +1383,6 @@ static int mtk_nfc_probe(struct platform_device *pdev) nfc->regs = devm_ioremap_resource(dev, res); if (IS_ERR(nfc->regs)) { ret = PTR_ERR(nfc->regs); - dev_err(dev, "no nfi base\n"); goto release_ecc; } From 656441478ed55d960df5f3ccdf5a0f8c61dfd0b3 Mon Sep 17 00:00:00 2001 From: Mark Marshall Date: Thu, 26 Jan 2017 16:18:27 +0100 Subject: [PATCH 339/343] mtd: nand: ifc: Fix location of eccstat registers for IFC V1.0 The commit 7a654172161c ("mtd/ifc: Add support for IFC controller version 2.0") added support for version 2.0 of the IFC controller. The version 2.0 controller has the ECC status registers at a different location to the previous versions. Correct the fsl_ifc_nand structure so that the ECC status can be read from the correct location for both version 1.0 and 2.0 of the controller. Cc: stable@vger.kernel.org Fixes: 7a654172161c ("mtd/ifc: Add support for IFC controller version 2.0") Signed-off-by: Mark Marshall Signed-off-by: Boris Brezillon --- drivers/mtd/nand/fsl_ifc_nand.c | 8 +++++++- include/linux/fsl_ifc.h | 8 ++++++-- 2 files changed, 13 insertions(+), 3 deletions(-) diff --git a/drivers/mtd/nand/fsl_ifc_nand.c b/drivers/mtd/nand/fsl_ifc_nand.c index 0a177b1bfe3e..d1570f512f0b 100644 --- a/drivers/mtd/nand/fsl_ifc_nand.c +++ b/drivers/mtd/nand/fsl_ifc_nand.c @@ -258,9 +258,15 @@ static void fsl_ifc_run_command(struct mtd_info *mtd) int bufnum = nctrl->page & priv->bufnum_mask; int sector = bufnum * chip->ecc.steps; int sector_end = sector + chip->ecc.steps - 1; + __be32 *eccstat_regs; + + if (ctrl->version >= FSL_IFC_VERSION_2_0_0) + eccstat_regs = ifc->ifc_nand.v2_nand_eccstat; + else + eccstat_regs = ifc->ifc_nand.v1_nand_eccstat; for (i = sector / 4; i <= sector_end / 4; i++) - eccstat[i] = ifc_in32(&ifc->ifc_nand.nand_eccstat[i]); + eccstat[i] = ifc_in32(&eccstat_regs[i]); for (i = sector; i <= sector_end; i++) { errors = check_read_ecc(mtd, ctrl, eccstat, i); diff --git a/include/linux/fsl_ifc.h b/include/linux/fsl_ifc.h index 3f9778cbc79d..c332f0a45607 100644 --- a/include/linux/fsl_ifc.h +++ b/include/linux/fsl_ifc.h @@ -733,8 +733,12 @@ struct fsl_ifc_nand { __be32 nand_erattr1; u32 res19[0x10]; __be32 nand_fsr; - u32 res20[0x3]; - __be32 nand_eccstat[6]; + u32 res20; + /* The V1 nand_eccstat is actually 4 words that overlaps the + * V2 nand_eccstat. + */ + __be32 v1_nand_eccstat[2]; + __be32 v2_nand_eccstat[6]; u32 res21[0x1c]; __be32 nanndcr; u32 res22[0x2]; From 19649e2c16fbc94b664f7074ec4fa9f15292fdce Mon Sep 17 00:00:00 2001 From: Boris Brezillon Date: Fri, 6 Jan 2017 10:42:05 +0100 Subject: [PATCH 340/343] mtd: nand: sunxi: Fix the non-polling case in sunxi_nfc_wait_events() wait_for_completion_timeout() returns 0 if a timeout occurred, 1 otherwise. Fix the sunxi_nfc_wait_events() accordingly. Signed-off-by: Boris Brezillon --- drivers/mtd/nand/sunxi_nand.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/drivers/mtd/nand/sunxi_nand.c b/drivers/mtd/nand/sunxi_nand.c index e40482a65de6..ba78e13a3570 100644 --- a/drivers/mtd/nand/sunxi_nand.c +++ b/drivers/mtd/nand/sunxi_nand.c @@ -321,6 +321,10 @@ static int sunxi_nfc_wait_events(struct sunxi_nfc *nfc, u32 events, ret = wait_for_completion_timeout(&nfc->complete, msecs_to_jiffies(timeout_ms)); + if (!ret) + ret = -ETIMEDOUT; + else + ret = 0; writel(0, nfc->regs + NFC_REG_INT); } else { From 8de15e1feb232086d1d14e01a6a7648d10002c19 Mon Sep 17 00:00:00 2001 From: Boris Brezillon Date: Fri, 6 Jan 2017 10:42:06 +0100 Subject: [PATCH 341/343] mtd: nand: sunxi: Stop using polling mode when waiting for long operations Some operations, like read/write an entire page of data with the ECC engine enabled, are known to take a lot of time. Use the interrupt-based waiting mode in these situation. Signed-off-by: Boris Brezillon --- drivers/mtd/nand/sunxi_nand.c | 24 ++++++++++++++++++------ 1 file changed, 18 insertions(+), 6 deletions(-) diff --git a/drivers/mtd/nand/sunxi_nand.c b/drivers/mtd/nand/sunxi_nand.c index ba78e13a3570..76449f79d4c5 100644 --- a/drivers/mtd/nand/sunxi_nand.c +++ b/drivers/mtd/nand/sunxi_nand.c @@ -522,6 +522,8 @@ static void sunxi_nfc_read_buf(struct mtd_info *mtd, uint8_t *buf, int len) u32 tmp; while (len > offs) { + bool poll = false; + cnt = min(len - offs, NFC_SRAM_SIZE); ret = sunxi_nfc_wait_cmd_fifo_empty(nfc); @@ -532,7 +534,11 @@ static void sunxi_nfc_read_buf(struct mtd_info *mtd, uint8_t *buf, int len) tmp = NFC_DATA_TRANS | NFC_DATA_SWAP_METHOD; writel(tmp, nfc->regs + NFC_REG_CMD); - ret = sunxi_nfc_wait_events(nfc, NFC_CMD_INT_FLAG, true, 0); + /* Arbitrary limit for polling mode */ + if (cnt < 64) + poll = true; + + ret = sunxi_nfc_wait_events(nfc, NFC_CMD_INT_FLAG, poll, 0); if (ret) break; @@ -555,6 +561,8 @@ static void sunxi_nfc_write_buf(struct mtd_info *mtd, const uint8_t *buf, u32 tmp; while (len > offs) { + bool poll = false; + cnt = min(len - offs, NFC_SRAM_SIZE); ret = sunxi_nfc_wait_cmd_fifo_empty(nfc); @@ -567,7 +575,11 @@ static void sunxi_nfc_write_buf(struct mtd_info *mtd, const uint8_t *buf, NFC_ACCESS_DIR; writel(tmp, nfc->regs + NFC_REG_CMD); - ret = sunxi_nfc_wait_events(nfc, NFC_CMD_INT_FLAG, true, 0); + /* Arbitrary limit for polling mode */ + if (cnt < 64) + poll = true; + + ret = sunxi_nfc_wait_events(nfc, NFC_CMD_INT_FLAG, poll, 0); if (ret) break; @@ -961,7 +973,7 @@ static int sunxi_nfc_hw_ecc_read_chunk(struct mtd_info *mtd, writel(NFC_DATA_TRANS | NFC_DATA_SWAP_METHOD | NFC_ECC_OP, nfc->regs + NFC_REG_CMD); - ret = sunxi_nfc_wait_events(nfc, NFC_CMD_INT_FLAG, true, 0); + ret = sunxi_nfc_wait_events(nfc, NFC_CMD_INT_FLAG, false, 0); sunxi_nfc_randomizer_disable(mtd); if (ret) return ret; @@ -1073,7 +1085,7 @@ static int sunxi_nfc_hw_ecc_read_chunks_dma(struct mtd_info *mtd, uint8_t *buf, writel(NFC_PAGE_OP | NFC_DATA_SWAP_METHOD | NFC_DATA_TRANS, nfc->regs + NFC_REG_CMD); - ret = sunxi_nfc_wait_events(nfc, NFC_CMD_INT_FLAG, true, 0); + ret = sunxi_nfc_wait_events(nfc, NFC_CMD_INT_FLAG, false, 0); if (ret) dmaengine_terminate_all(nfc->dmac); @@ -1193,7 +1205,7 @@ static int sunxi_nfc_hw_ecc_write_chunk(struct mtd_info *mtd, NFC_ACCESS_DIR | NFC_ECC_OP, nfc->regs + NFC_REG_CMD); - ret = sunxi_nfc_wait_events(nfc, NFC_CMD_INT_FLAG, true, 0); + ret = sunxi_nfc_wait_events(nfc, NFC_CMD_INT_FLAG, false, 0); sunxi_nfc_randomizer_disable(mtd); if (ret) return ret; @@ -1432,7 +1444,7 @@ static int sunxi_nfc_hw_ecc_write_page_dma(struct mtd_info *mtd, NFC_DATA_TRANS | NFC_ACCESS_DIR, nfc->regs + NFC_REG_CMD); - ret = sunxi_nfc_wait_events(nfc, NFC_CMD_INT_FLAG, true, 0); + ret = sunxi_nfc_wait_events(nfc, NFC_CMD_INT_FLAG, false, 0); if (ret) dmaengine_terminate_all(nfc->dmac); From cad32741218299eee3f75de21ac9ccab8485bd62 Mon Sep 17 00:00:00 2001 From: Boris Brezillon Date: Fri, 6 Jan 2017 10:42:07 +0100 Subject: [PATCH 342/343] mtd: nand: sunxi: Improve sunxi_nfc_cmd_ctrl() We only need to call sunxi_nfc_wait_cmd_fifo_empty() if we want to send a new command. Move the sunxi_nfc_wait_cmd_fifo_empty() call to right place to avoid extra register reads. Signed-off-by: Boris Brezillon --- drivers/mtd/nand/sunxi_nand.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/mtd/nand/sunxi_nand.c b/drivers/mtd/nand/sunxi_nand.c index 76449f79d4c5..0eeeb8b889ea 100644 --- a/drivers/mtd/nand/sunxi_nand.c +++ b/drivers/mtd/nand/sunxi_nand.c @@ -604,10 +604,6 @@ static void sunxi_nfc_cmd_ctrl(struct mtd_info *mtd, int dat, struct sunxi_nfc *nfc = to_sunxi_nfc(sunxi_nand->nand.controller); int ret; - ret = sunxi_nfc_wait_cmd_fifo_empty(nfc); - if (ret) - return; - if (dat == NAND_CMD_NONE && (ctrl & NAND_NCE) && !(ctrl & (NAND_CLE | NAND_ALE))) { u32 cmd = 0; @@ -637,6 +633,10 @@ static void sunxi_nfc_cmd_ctrl(struct mtd_info *mtd, int dat, writel(sunxi_nand->addr[1], nfc->regs + NFC_REG_ADDR_HIGH); + ret = sunxi_nfc_wait_cmd_fifo_empty(nfc); + if (ret) + return; + writel(cmd, nfc->regs + NFC_REG_CMD); sunxi_nand->addr[0] = 0; sunxi_nand->addr[1] = 0; From a4077ce5871304f8a78f80b74b18b6052a410f1a Mon Sep 17 00:00:00 2001 From: "Andrey Jr. Melnikov" Date: Thu, 8 Dec 2016 19:57:08 +0300 Subject: [PATCH 343/343] mtd: nand: Add Winbond manufacturer id Add WINBOND manufacturer id. Signed-off-by: Andrey Jr. Melnikov Signed-off-by: Boris Brezillon --- drivers/mtd/nand/nand_ids.c | 1 + include/linux/mtd/nand.h | 1 + 2 files changed, 2 insertions(+) diff --git a/drivers/mtd/nand/nand_ids.c b/drivers/mtd/nand/nand_ids.c index b3a332f37e14..4a2f75b0c200 100644 --- a/drivers/mtd/nand/nand_ids.c +++ b/drivers/mtd/nand/nand_ids.c @@ -185,6 +185,7 @@ struct nand_manufacturers nand_manuf_ids[] = { {NAND_MFR_SANDISK, "SanDisk"}, {NAND_MFR_INTEL, "Intel"}, {NAND_MFR_ATO, "ATO"}, + {NAND_MFR_WINBOND, "Winbond"}, {0x0, "Unknown"} }; diff --git a/include/linux/mtd/nand.h b/include/linux/mtd/nand.h index f67915c7726f..f6017a1a35d6 100644 --- a/include/linux/mtd/nand.h +++ b/include/linux/mtd/nand.h @@ -959,6 +959,7 @@ static inline void nand_set_controller_data(struct nand_chip *chip, void *priv) #define NAND_MFR_SANDISK 0x45 #define NAND_MFR_INTEL 0x89 #define NAND_MFR_ATO 0x9b +#define NAND_MFR_WINBOND 0xef /* The maximum expected count of bytes in the NAND ID sequence */ #define NAND_MAX_ID_LEN 8