In the FireWire OHCI interrupt handler, if a bus reset interrupt has
occurred, mask bus reset interrupts until bus_reset_work has serviced and
cleared the interrupt.
Normally, we always leave bus reset interrupts masked. We infer the bus
reset from the self-ID interrupt that happens shortly thereafter. A
scenario where we unmask bus reset interrupts was introduced in 2008 in
a007bb857e: If
OHCI_PARAM_DEBUG_BUSRESETS (8) is set in the debug parameter bitmask, we
will unmask bus reset interrupts so we can log them.
irq_handler logs the bus reset interrupt. However, we can't clear the bus
reset event flag in irq_handler, because we won't service the event until
later. irq_handler exits with the event flag still set. If the
corresponding interrupt is still unmasked, the first bus reset will
usually freeze the system due to irq_handler being called again each
time it exits. This freeze can be reproduced by loading firewire_ohci
with "modprobe firewire_ohci debug=-1" (to enable all debugging output).
Apparently there are also some cases where bus_reset_work will get called
soon enough to clear the event, and operation will continue normally.
This freeze was first reported a few months after a007bb85 was committed,
but until now it was never fixed. The debug level could safely be set
to -1 through sysfs after the module was loaded, but this would be
ineffectual in logging bus reset interrupts since they were only
unmasked during initialization.
irq_handler will now leave the event flag set but mask bus reset
interrupts, so irq_handler won't be called again and there will be no
freeze. If OHCI_PARAM_DEBUG_BUSRESETS is enabled, bus_reset_work will
unmask the interrupt after servicing the event, so future interrupts
will be caught as desired.
As a side effect to this change, OHCI_PARAM_DEBUG_BUSRESETS can now be
enabled through sysfs in addition to during initial module loading.
However, when enabled through sysfs, logging of bus reset interrupts will
be effective only starting with the second bus reset, after
bus_reset_work has executed.
Signed-off-by: Adam Goldman <adamg@pobox.com>
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
In the case of firewire core function, the caller of show functions for
device attributes is not only sysfs user, but also device initialization.
This commit adds memo about it against the typical assumption that the
functions are just dedicated to sysfs user.
Link: https://lore.kernel.org/lkml/20240318091759.678326-1-o-takashi@sakamocchi.jp/
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Per filesystems/sysfs.rst, show() should only use sysfs_emit()
or sysfs_emit_at() when formatting the value to be returned to user space.
coccinelle complains that there are still a couple of functions that use
snprintf(). Convert them to sysfs_emit().
> drivers/firewire/core-device.c:326:8-16: WARNING: please use sysfs_emit or sysfs_emit_at
No functional change intended
Signed-off-by: Li Zhijian <lizhijian@fujitsu.com>
Link: https://lore.kernel.org/r/20240122053942.80648-2-lizhijian@fujitsu.com
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Per Documentation/filesystems/sysfs.rst:
> sysfs allocates a buffer of size (PAGE_SIZE) and passes it to the
> method.
So we can kill the unnecessary buf check safely.
Signed-off-by: Li Zhijian <lizhijian@fujitsu.com>
Link: https://lore.kernel.org/r/20240122053942.80648-1-lizhijian@fujitsu.com
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Commit 5a95f1ded2 ("firewire: ohci: use devres for requested IRQ")
also removed the call to free_irq() in pci_remove(), leading to a
leftover irq of devm_request_irq() at pci_disable_msi() in pci_remove()
when unbinding the driver from the device
remove_proc_entry: removing non-empty directory 'irq/136', leaking at
least 'firewire_ohci'
Call Trace:
? remove_proc_entry+0x19c/0x1c0
? __warn+0x81/0x130
? remove_proc_entry+0x19c/0x1c0
? report_bug+0x171/0x1a0
? console_unlock+0x78/0x120
? handle_bug+0x3c/0x80
? exc_invalid_op+0x17/0x70
? asm_exc_invalid_op+0x1a/0x20
? remove_proc_entry+0x19c/0x1c0
unregister_irq_proc+0xf4/0x120
free_desc+0x3d/0xe0
? kfree+0x29f/0x2f0
irq_free_descs+0x47/0x70
msi_domain_free_locked.part.0+0x19d/0x1d0
msi_domain_free_irqs_all_locked+0x81/0xc0
pci_free_msi_irqs+0x12/0x40
pci_disable_msi+0x4c/0x60
pci_remove+0x9d/0xc0 [firewire_ohci
01b483699bebf9cb07a3d69df0aa2bee71db1b26]
pci_device_remove+0x37/0xa0
device_release_driver_internal+0x19f/0x200
unbind_store+0xa1/0xb0
remove irq with devm_free_irq() before pci_disable_msi()
also remove it in fail_msi: of pci_probe() as this would lead to
an identical leak
Cc: stable@vger.kernel.org
Fixes: 5a95f1ded2 ("firewire: ohci: use devres for requested IRQ")
Signed-off-by: Edmund Raile <edmund.raile@proton.me>
Link: https://lore.kernel.org/r/20240229144723.13047-2-edmund.raile@proton.me
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
When resetting the bus after a gap count error, use a long rather than
short bus reset.
IEEE 1394-1995 uses only long bus resets. IEEE 1394a adds the option of
short bus resets. When video or audio transmission is in progress and a
device is hot-plugged elsewhere on the bus, the resulting bus reset can
cause video frame drops or audio dropouts. Short bus resets reduce or
eliminate this problem. Accordingly, short bus resets are almost always
preferred.
However, on a mixed 1394/1394a bus, a short bus reset can trigger an
immediate additional bus reset. This double bus reset can be interpreted
differently by different nodes on the bus, resulting in an inconsistent gap
count after the bus reset. An inconsistent gap count will cause another bus
reset, leading to a neverending bus reset loop. This only happens for some
bus topologies, not for all mixed 1394/1394a buses.
By instead sending a long bus reset after a gap count inconsistency, we
avoid the doubled bus reset, restoring the bus to normal operation.
Signed-off-by: Adam Goldman <adamg@pobox.com>
Link: https://sourceforge.net/p/linux1394/mailman/message/58741624/
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
If we are bus manager and the bus has inconsistent gap counts, send a
bus reset immediately instead of trying to read the root node's config
ROM first. Otherwise, we could spend a lot of time trying to read the
config ROM but never succeeding.
This eliminates a 50+ second delay before the FireWire bus is usable after
a newly connected device is powered on in certain circumstances.
The delay occurs if a gap count inconsistency occurs, we are not the root
node, and we become bus manager. One scenario that causes this is with a TI
XIO2213B OHCI, the first time a Sony DSR-25 is powered on after being
connected to the FireWire cable. In this configuration, the Linux box will
not receive the initial PHY configuration packet sent by the DSR-25 as IRM,
resulting in the DSR-25 having a gap count of 44 while the Linux box has a
gap count of 63.
FireWire devices have a gap count parameter, which is set to 63 on power-up
and can be changed with a PHY configuration packet. This determines the
duration of the subaction and arbitration gaps. For reliable communication,
all nodes on a FireWire bus must have the same gap count.
A node may have zero or more of the following roles: root node, bus manager
(BM), isochronous resource manager (IRM), and cycle master. Unless a root
node was forced with a PHY configuration packet, any node might become root
node after a bus reset. Only the root node can become cycle master. If the
root node is not cycle master capable, the BM or IRM should force a change
of root node.
After a bus reset, each node sends a self-ID packet, which contains its
current gap count. A single bus reset does not change the gap count, but
two bus resets in a row will set the gap count to 63. Because a consistent
gap count is required for reliable communication, IEEE 1394a-2000 requires
that the bus manager generate a bus reset if it detects that the gap count
is inconsistent.
When the gap count is inconsistent, build_tree() will notice this after the
self identification process. It will set card->gap_count to the invalid
value 0. If we become bus master, this will force bm_work() to send a bus
reset when it performs gap count optimization.
After a bus reset, there is no bus manager. We will almost always try to
become bus manager. Once we become bus manager, we will first determine
whether the root node is cycle master capable. Then, we will determine if
the gap count should be changed. If either the root node or the gap count
should be changed, we will generate a bus reset.
To determine if the root node is cycle master capable, we read its
configuration ROM. bm_work() will wait until we have finished trying to
read the configuration ROM.
However, an inconsistent gap count can make this take a long time.
read_config_rom() will read the first few quadlets from the config ROM. Due
to the gap count inconsistency, eventually one of the reads will time out.
When read_config_rom() fails, fw_device_init() calls it again until
MAX_RETRIES is reached. This takes 50+ seconds.
Once we give up trying to read the configuration ROM, bm_work() will wake
up, assume that the root node is not cycle master capable, and do a bus
reset. Hopefully, this will resolve the gap count inconsistency.
This change makes bm_work() check for an inconsistent gap count before
waiting for the root node's configuration ROM. If the gap count is
inconsistent, bm_work() will immediately do a bus reset. This eliminates
the 50+ second delay and rapidly brings the bus to a working state.
I considered that if the gap count is inconsistent, a PHY configuration
packet might not be successful, so it could be desirable to skip the PHY
configuration packet before the bus reset in this case. However, IEEE
1394a-2000 and IEEE 1394-2008 say that the bus manager may transmit a PHY
configuration packet before a bus reset when correcting a gap count error.
Since the standard endorses this, I decided it's safe to retain the PHY
configuration packet transmission.
Normally, after a topology change, we will reset the bus a maximum of 5
times to change the root node and perform gap count optimization. However,
if there is a gap count inconsistency, we must always generate a bus reset.
Otherwise the gap count inconsistency will persist and communication will
be unreliable. For that reason, if there is a gap count inconstency, we
generate a bus reset even if we already reached the 5 reset limit.
Signed-off-by: Adam Goldman <adamg@pobox.com>
Reference: https://sourceforge.net/p/linux1394/mailman/message/58727806/
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
As the last part of support for legacy layout of configuration ROM, this
commit traverses vendor directory as well as root directory when
constructing modalias for unit device. The change brings loss of backward
compatibility since it can fill model field ('mo') which is 0 at current
implementation in the case. However, we can be optimistic against
regression for unit drivers in kernel, due to some points:
1. ALSA drivers for audio and music units use the model fields to match
device, however all of supported devices does not have such legacy
layout.
2. the other unit drivers (e.g. sbp2) does not use the model field to
match device.
The rest of concern is user space application. The most of applications
just take care of node device and does not use the modalias of unit
device, thus the change does not affect to them. But systemd project is
known to get affects from the change since it includes hwdb to take udev
to configure fw character device conveniently. I have a plan to work for
systemd so that the access permission of character device could be kept
across the change.
Suggested-by: Adam Goldman <adamg@pobox.com>
Link: https://lore.kernel.org/r/20231221134849.603857-9-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
As the part of support for legacy layout of configuration ROM, this
commit traverses vendor directory as well as root directory when showing
device attribute for node device. This change expects 'model_name'
attribute appears in node device, however it is probable to see the other
types of descriptor leaf if the vendor directory includes.
Suggested-by: Adam Goldman <adamg@pobox.com>
Link: https://lore.kernel.org/r/20231221134849.603857-8-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
As the part of support for legacy layout of configuration ROM, this
commit traverses vendor directory as well as root directory when showing
device attribute for node device. This change expects 'model' attribute
appears in node device, however it is probable to see the other types of
immediate values if the vendor directory includes.
Suggested-by: Adam Goldman <adamg@pobox.com>
Link: https://lore.kernel.org/r/20231221134849.603857-7-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Some legacy devices have configuration ROM against standard AV/C device.
They have vendor directory to store model identifier. It is described in
annex of the following document.
- Configuration ROM for AV/C Devices 1.0 (Dec. 12, 2000, 1394 Trading
Association)
In the case, current implementation of core function does not detect the
model identifier, thus device attributes and modalias of unit have lack of
it. Another KUnit test is required for the case, and this commit is for
the purpose.
The following output is the parse result for the hard-coded data, by
config-rom-pretty-printer in linux-firewire-utils
(https://git.kernel.org/pub/scm/utils/ieee1394/linux-firewire-utils.git/).
The data is written by my hand.
$ config-rom-pretty-printer < /tmp/rom.img
ROM header and bus information block
-----------------------------------------------------------------
1024 04199fe7 bus_info_length 4, crc_length 25, crc 40935
1028 31333934 bus_name "1394"
1032 e0644000 irmc 1, cmc 1, isc 1, bmc 0, cyc_clk_acc 100, max_rec 4 (32)
1036 00112233 company_id 001122 |
1040 44556677 device_id 220189779575 | EUI-64 4822678189205111
root directory
-----------------------------------------------------------------
1044 0005dace directory_length 5, crc 56014
1048 03012345 vendor
1052 0c0083c0 node capabilities: per IEEE 1394
1056 8d000009 --> eui-64 leaf at 1092
1060 d1000002 --> unit directory at 1068
1064 c3000004 --> vendor directory at 1080
unit directory at 1068
-----------------------------------------------------------------
1068 0002e107 directory_length 2, crc 57607
1072 12abcdef specifier id
1076 13543210 version
vendor directory at 1080
-----------------------------------------------------------------
1080 0002cb73 directory_length 2, crc 52083
1084 17fedcba model
1088 81000004 --> descriptor leaf at 1104
eui-64 leaf at 1092
-----------------------------------------------------------------
1092 00026dc1 leaf_length 2, crc 28097
1096 00112233 company_id 001122 |
1100 44556677 device_id 220189779575 | EUI-64 4822678189205111
descriptor leaf at 1104
-----------------------------------------------------------------
1104 00050e84 leaf_length 5, crc 3716
1108 00000000 textual descriptor
1112 00000000 minimal ASCII
1116 41424344 "ABCD"
1120 45464748 "EFGH"
1124 494a0000 "IJ"
Link: https://lore.kernel.org/r/20231221134849.603857-6-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
At present, core function can handle node which has configuration ROM
similar to standard AV/C device somehow. The standard layout of
configuration ROM is described in the following document.
- Configuration ROM for AV/C Devices 1.0 (Dec. 12, 2000, 1394 Trading
Association)
This commit adds a KUnit test for the above case.
The following output is the parse result for the hard-coded data, by
config-rom-pretty-printer in linux-firewire-utils
(https://git.kernel.org/pub/scm/utils/ieee1394/linux-firewire-utils.git/).
$ config-rom-pretty-printer < /tmp/rom.img
ROM header and bus information block
-----------------------------------------------------------------
1024 0404eabf bus_info_length 4, crc_length 4, crc 60095
1028 31333934 bus_name "1394"
1032 e0646102 irmc 1, cmc 1, isc 1, bmc 0, cyc_clk_acc 100, max_rec 6 (128)
1036 ffffffff company_id ffffff |
1040 ffffffff device_id 1099511627775 | EUI-64 18446744073709551615
root directory
-----------------------------------------------------------------
1044 00063287 directory_length 6, crc 12935
1048 03ffffff vendor
1052 8100000a --> descriptor leaf at 1092
1056 17ffffff model
1060 8100000e --> descriptor leaf at 1116
1064 0c0083c0 node capabilities: per IEEE 1394
1068 d1000001 --> unit directory at 1072
unit directory at 1072
-----------------------------------------------------------------
1072 0004442d directory_length 4, crc 17453
1076 1200a02d specifier id
1080 13010001 version
1084 17ffffff model
1088 81000007 --> descriptor leaf at 1116
descriptor leaf at 1092
-----------------------------------------------------------------
1092 0005c915 leaf_length 5, crc 51477
1096 00000000 textual descriptor
1100 00000000 minimal ASCII
1104 56656e64 "Vend"
1108 6f72204e "or N"
1112 616d6500 "ame"
descriptor leaf at 1116
-----------------------------------------------------------------
1116 00057f16 leaf_length 5, crc 32534
1120 00000000 textual descriptor
1124 00000000 minimal ASCII
1128 4d6f6465 "Mode"
1132 6c204e61 "l Na"
1136 6d650000 "me"
Link: https://lore.kernel.org/r/20231221134849.603857-5-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
The traverse over CSR space results in attributes of node and unit
devices. Any test of the traverse is useful.
This commit adds a skeleton of KUnit test for the purpose.
Link: https://lore.kernel.org/r/20231221134849.603857-4-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
In IEEE 1394 specification, the size of bus information block of
configuration ROM is fixed to 5, thus the offset of root directory is 5.
Current implementation to handle device structures has the hard-coded
offset.
This commit replaces the offset with macro.
Link: https://lore.kernel.org/r/20231221134849.603857-3-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Now that the driver core can properly handle constant struct bus_type,
move the fw_bus_type variable to be a constant structure as well,
placing it into read-only memory which can not be modified at runtime.
Cc: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Cc: linux1394-devel@lists.sourceforge.net
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Link: https://lore.kernel.org/r/2023121931-skydiver-dodgy-d1bd@gregkh
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
VIA VT6306/6307/6308 provides PCI interface compliant to 1394 OHCI. When
the hardware is combined with Asmedia ASM1083/1085 PCIe-to-PCI bus bridge,
it appears that accesses to its 'Isochronous Cycle Timer' register (offset
0xf0 on PCI memory space) often causes unexpected system reboot in any
type of AMD Ryzen machine (both 0x17 and 0x19 families). It does not
appears in the other type of machine (AMD pre-Ryzen machine, Intel
machine, at least), or in the other OHCI 1394 hardware (e.g. Texas
Instruments).
The issue explicitly appears at a commit dcadfd7f7c ("firewire: core:
use union for callback of transaction completion") added to v6.5 kernel.
It changed 1394 OHCI driver to access to the register every time to
dispatch local asynchronous transaction. However, the issue exists in
older version of kernel as long as it runs in AMD Ryzen machine, since
the access to the register is required to maintain bus time. It is not
hard to imagine that users experience the unexpected system reboot when
generating bus reset by plugging any devices in, or reading the register
by time-aware application programs; e.g. audio sample processing.
This commit suppresses the unexpected system reboot in the combination of
hardware. It avoids the access itself. As a result, the software stack can
not provide the hardware time anymore to unit drivers, userspace
applications, and nodes in the same IEEE 1394 bus. It brings apparent
disadvantage since time-aware application programs require it, while
time-unaware applications are available again; e.g. sbp2.
Cc: stable@vger.kernel.org
Reported-by: Jiri Slaby <jirislaby@kernel.org>
Closes: https://bugzilla.suse.com/show_bug.cgi?id=1215436
Reported-by: Mario Limonciello <mario.limonciello@amd.com>
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=217994
Reported-by: Tobias Gruetzmacher <tobias-lists@23.gs>
Closes: https://sourceforge.net/p/linux1394/mailman/message/58711901/
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=2240973
Closes: https://bugs.launchpad.net/linux/+bug/2043905
Link: https://lore.kernel.org/r/20240102110150.244475-1-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
This pull request includes a single patch to fix long-standing issue of
memory leak at failure of device registration for fw_unit. We rarely
encounter the issue, while it should be applied to stable releases, since
it fixes inappropriate API usage.
-----BEGIN PGP SIGNATURE-----
iHUEABYIAB0WIQQE66IEYNDXNBPeGKSsLtaWM8LwEwUCZWrgUQAKCRCsLtaWM8Lw
Ew8iAQDWPv5JItuZneeRs6NisGxtwNxyxP+O3qmSG4LxXOVfzAEA+K7ubreLee2a
lfZkWPeZnWtSIdN+eAm4JfKMhzDfrA4=
=oODB
-----END PGP SIGNATURE-----
Merge tag 'firewire-fixes-6.7-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394
Pull firewire fix from Takashi Sakamoto:
"A single patch to fix long-standing issue of memory leak at failure of
device registration for fw_unit. We rarely encounter the issue, but it
should be applied to stable releases, since it fixes inappropriate API
usage"
* tag 'firewire-fixes-6.7-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394:
firewire: core: fix possible memory leak in create_units()
If device_register() fails, the refcount of device is not 0, the name
allocated in dev_set_name() is leaked. To fix this by calling put_device(),
so that it will be freed in callback function kobject_cleanup().
unreferenced object 0xffff9d99035c7a90 (size 8):
comm "systemd-udevd", pid 168, jiffies 4294672386 (age 152.089s)
hex dump (first 8 bytes):
66 77 30 2e 30 00 ff ff fw0.0...
backtrace:
[<00000000e1d62bac>] __kmem_cache_alloc_node+0x1e9/0x360
[<00000000bbeaff31>] __kmalloc_node_track_caller+0x44/0x1a0
[<00000000491f2fb4>] kvasprintf+0x67/0xd0
[<000000005b960ddc>] kobject_set_name_vargs+0x1e/0x90
[<00000000427ac591>] dev_set_name+0x4e/0x70
[<000000003b4e447d>] create_units+0xc5/0x110
fw_unit_release() will be called in the error path, move fw_device_get()
before calling device_register() to keep balanced with fw_device_put() in
fw_unit_release().
Cc: stable@vger.kernel.org
Fixes: 1fa5ae857b ("driver core: get rid of struct device's bus_id string array")
Fixes: a1f64819fe ("firewire: struct device - replace bus_id with dev_name(), dev_set_name()")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Commit 3cc2ffe5c1 ("scsi: sd: Differentiate system and runtime start/stop
management") changed the single bit manage_start_stop flag into 2 boolean
fields of the SCSI device structure. Commit 24eca2dce0 ("scsi: sd:
Introduce manage_shutdown device flag") introduced the manage_shutdown
boolean field for the same structure. Together, these 2 commits increase
the size of struct scsi_device by 8 bytes by using booleans instead of
defining the manage_xxx fields as single bit flags, similarly to other
flags of this structure.
Avoid this unnecessary structure size increase and be consistent with the
definition of other flags by reverting the definitions of the manage_xxx
fields as single bit flags.
Fixes: 3cc2ffe5c1 ("scsi: sd: Differentiate system and runtime start/stop management")
Fixes: 24eca2dce0 ("scsi: sd: Introduce manage_shutdown device flag")
Cc: <stable@vger.kernel.org>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
Link: https://lore.kernel.org/r/20231120225631.37938-2-dlemoal@kernel.org
Reviewed-by: Niklas Cassel <niklas.cassel@wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Prepare for the coming implementation by GCC and Clang of the __counted_by
attribute. Flexible array members annotated with __counted_by can have
their accesses bounds-checked at run-time checking via CONFIG_UBSAN_BOUNDS
(for array indexing) and CONFIG_FORTIFY_SOURCE (for strcpy/memcpy-family
functions).
As found with Coccinelle[1], add __counted_by for struct fw_node.
[1] https://github.com/kees/kernel-tools/blob/trunk/coccinelle/examples/counted_by.cocci
Cc: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Cc: linux1394-devel@lists.sourceforge.net
Signed-off-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Link: https://lore.kernel.org/r/20230922175334.work.335-kees@kernel.org
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Commit aa3998dbeb ("ata: libata-scsi: Disable scsi device
manage_system_start_stop") change setting the manage_system_start_stop
flag to false for libata managed disks to enable libata internal
management of disk suspend/resume. However, a side effect of this change
is that on system shutdown, disks are no longer being stopped (set to
standby mode with the heads unloaded). While this is not a critical
issue, this unclean shutdown is not recommended and shows up with
increased smart counters (e.g. the unexpected power loss counter
"Unexpect_Power_Loss_Ct").
Instead of defining a shutdown driver method for all ATA adapter
drivers (not all of them define that operation), this patch resolves
this issue by further refining the sd driver start/stop control of disks
using the new flag manage_shutdown. If this new flag is set to true by
a low level driver, the function sd_shutdown() will issue a
START STOP UNIT command with the start argument set to 0 when a disk
needs to be powered off (suspended) on system power off, that is, when
system_state is equal to SYSTEM_POWER_OFF.
Similarly to the other manage_xxx flags, the new manage_shutdown flag is
exposed through sysfs as a read-write device attribute.
To avoid any confusion between manage_shutdown and
manage_system_start_stop, the comments describing these flags in
include/scsi/scsi.h are also improved.
Fixes: aa3998dbeb ("ata: libata-scsi: Disable scsi device manage_system_start_stop")
Cc: stable@vger.kernel.org
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=218038
Link: https://lore.kernel.org/all/cd397c88-bf53-4768-9ab8-9d107df9e613@gmail.com/
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
Reviewed-by: Niklas Cassel <niklas.cassel@wdc.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Acked-by: Martin K. Petersen <martin.petersen@oracle.com>
The underlying device and driver of a SCSI disk may have different
system and runtime power mode control requirements. This is because
runtime power management affects only the SCSI disk, while system level
power management affects all devices, including the controller for the
SCSI disk.
For instance, issuing a START STOP UNIT command when a SCSI disk is
runtime suspended and resumed is fine: the command is translated to a
STANDBY IMMEDIATE command to spin down the ATA disk and to a VERIFY
command to wake it up. The SCSI disk runtime operations have no effect
on the ata port device used to connect the ATA disk. However, for
system suspend/resume operations, the ATA port used to connect the
device will also be suspended and resumed, with the resume operation
requiring re-validating the device link and the device itself. In this
case, issuing a VERIFY command to spinup the disk must be done before
starting to revalidate the device, when the ata port is being resumed.
In such case, we must not allow the SCSI disk driver to issue START STOP
UNIT commands.
Allow a low level driver to refine the SCSI disk start/stop management
by differentiating system and runtime cases with two new SCSI device
flags: manage_system_start_stop and manage_runtime_start_stop. These new
flags replace the current manage_start_stop flag. Drivers setting the
manage_start_stop are modifed to set both new flags, thus preserving the
existing start/stop management behavior. For backward compatibility, the
old manage_start_stop sysfs device attribute is kept as a read-only
attribute showing a value of 1 for devices enabling both new flags and 0
otherwise.
Fixes: 0a85890559 ("ata,scsi: do not issue START STOP UNIT on resume")
Cc: stable@vger.kernel.org
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
The flag of GFP_ATOMIC is given to the call of kmalloc when building node
tree, but the call is not atomic context. The call of
fw_core_handle_bus_reset() and fw_core_remove_card() builds the tree,
while they are done in specific workqueue or pci remove callback.
This commit obsolete the usage of GFP_ATOMIC.
Link: https://lore.kernel.org/r/20230604070255.172700-1-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
The 1394 OHCI driver allocates several non-coherent DMA buffers for AR
request and response contexts. The buffers are mapped to kernel virtual
address (VMA) so that the first page locates after the last page. Even
when large payload of packet is handled crossing the boundary of buffers,
the driver operates continuously on VMA.
No kernel API is provided for this kind of mapping, while it is possible
to release the buffer when PCI device is going to be released.
This commit moves the call of release helper function to the callback
function of release resources.
Link: https://lore.kernel.org/r/20230604054451.161076-10-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
The 1394 OHCI driver allocates DMA coherent buffer to transfer content
of configuration ROM.
This commit utilizes managed device resource to maintain the lifetime of
buffer.
Link: https://lore.kernel.org/r/20230604054451.161076-9-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
The 1394 OHCI driver allocates DMA coherent buffer for descriptors of IT,
IR, AT receive, and AT request contexts by the same way.
This commit utilizes managed device resource to maintain the lifetime of
buffers.
Link: https://lore.kernel.org/r/20230604054451.161076-8-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
The 1394 OHCI driver allocates the list of isochronous contexts as much
as the hardware supports.
This commit utilizes managed device resource to maintain the lifetime of
list.
Link: https://lore.kernel.org/r/20230604054451.161076-7-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
The 1394 OHCI driver allocates a DMA coherent buffer for multi-purposes.
The buffer is split into three region for specific purposes; i.e. 1/4 for
context descriptors of AR request and response as well as 1/2 for self
ID handling.
This commit uses managed device resource to maintain the lifetime of
buffer.
Link: https://lore.kernel.org/r/20230604054451.161076-5-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
The PCI framework has the convenient helper function to check and map MMIO
region with managed device resource.
This commit elaborates 1394 OHCI driver to use the function.
Link: https://lore.kernel.org/r/20230604054451.161076-4-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
The managed device resource (devres) framework is convenient to maintain
lifetime of allocated memory object for device.
This commit utilizes the framework for the object of ohci structure. The
extra operation for power management is required in Apple PowerMac based
machines, thus release callback is assigned to the object to call the
operation.
Link: https://lore.kernel.org/r/20230604054451.161076-2-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
In 1394 OHCI, the OUTPUT_LAST descriptor of Asynchronous Transmit (AT)
context has timeStamp field, in which 1394 OHCI controller record the
isochronous cycle when the packet was sent for the request subaction.
Additionally, the trailing quadlet of Asynchronous Receive (AR) context
has timeStamp field as well in which 1394 OHCI controller record the
isochronous cycle when the packet arrived. The time stamps are also
available for the cases to send and receive phy packet.
This commit implements new events with time stamp field for user space.
Link: https://lore.kernel.org/r/20230529113406.986289-13-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
This commit adds new event to notify event of phy packet with time stamp
field.
Unlike the fw_cdev_event_request3 and fw_cdev_event_response2, the size
of new structure, fw_cdev_event_phy_packet2, is multiples of 8, thus
padding is not required to keep the same size between System V ABI for
different architectures.
It is noticeable that for the case of ping request 1394 OHCI controller
does not record the isochronous cycle at which the packet was sent for
the request subaction. Instead, it records round-trip count measured by
hardware at 42.195 MHz resolution.
Cc: kunit-dev@googlegroups.com
Link: https://lore.kernel.org/r/20230529113406.986289-12-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
In 1394 OHCI, both Asynchronous Transmit (AT) and Asynchronous Receive
(AR) contexts are used to deliver the phy packet of IEEE 1394. The time
stamp is available as well as the usual asynchronous transaction.
This commit is a preparation for future commit to handle the time stamp.
Link: https://lore.kernel.org/r/20230529113406.986289-11-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
The callback function now receives an argument for time stamps relevant
to asynchronous transaction. This commit implements a new event to
notify response subaction with the time stamps for user space.
Link: https://lore.kernel.org/r/20230529113406.986289-10-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
This commit adds new event to notify event of response subaction with
time stamp field.
Current compiler implementation of System V ABI selects one of structure
members which has the maximum alignment size in the structure to decide
the size of structure. In the case of fw_cdev_event_request3 structure,
it is closure member which has 8 byte storage. The size of alignment for
the type of 8 byte storage differs depending on architectures; 4 byte for
i386 architecture and 8 byte for the others including x32 architecture.
It is inconvenient to device driver developer to use structure layout
which varies between architectures since the developer takes care of ioctl
compat layer. This commit adds 32 bit member for padding to keep the
size of structure as multiples of 8.
Cc: kunit-dev@googlegroups.com
Link: https://lore.kernel.org/r/20230529113406.986289-9-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
In the previous commit, the core function of Linux FireWire subsystem
was changed for two cases to operate asynchronous transaction with or
without time stamp.
This commit changes kernel API for the two cases. Current kernel API,
fw_send_request(), is changed to be static inline function to call
__fw_send_request(), which receives two argument for union and flag of
callback function. The new kernel API, fw_send_request_with_tstamp() is
also added as static inline function, too. When calling, the two
arguments are copied to internal structure, then used in softIRQ
context.
Link: https://lore.kernel.org/r/20230529113406.986289-7-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
In 1394 OHCI, the OUTPUT_LAST descriptor of Asynchronous Transmit (AT)
request context has timeStamp field, in which 1394 OHCI controller
record the isochronous cycle when the packet was sent for the request
subaction. Additionally, for the case of split transaction in IEEE 1394,
Asynchronous Receive (AT) request context is used for response subaction
to finish the transaction. The trailer quadlet of descriptor in the
context has timeStamp field, in which 1394 OHCI controller records the
isochronous cycle when the packet arrived.
Current implementation of 1394 OHCI controller driver stores values of
both fields to internal structure as time stamp, while Linux FireWire
subsystem provides no way to access to it. When using asynchronous
transaction service provided by the subsystem, callback function is passed
to kernel API. The prototype of callback function has the lack of argument
for the values.
This commit adds a new callback function for the purpose. It has an
additional argument to point to the constant array with two elements. For
backward compatibility to kernel space, a new union is also adds to wrap
two different prototype of callback function. The fw_transaction structure
has the union as a member and a boolean flag to express which function
callback is available.
The core function is changed to handle the two cases; with or without
time stamp. For the error path to process transaction, the isochronous
cycle is computed by current value of CYCLE_TIMER register in 1394 OHCI
controller. Especially for the case of timeout of split transaction, the
expected isochronous cycle is computed.
Link: https://lore.kernel.org/r/20230529113406.986289-6-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
In 1394 OHCI, the trailer quadlet of descriptor in Asynchronous Receive
(AR) request context has timeStamp field, in which the 1394 OHCI
controller record the isochronous cycle when the packet arrived.
Current implementation of 1394 OHCI controller driver stores the value
of field to internal structure as time stamp, while the implementation
of FireWire character device doesn't have a field for the time stamp,
thus it is not available in user space. The time stamp is convenient to
some kind of application in which data from several sources are compared
in isochronous cycle unit.
This commit implement the new event, fw_cdev_event_request3, with an
additional field, tstamp.
Link: https://lore.kernel.org/r/20230529113406.986289-5-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
This commit adds new event to notify event of request subaction with
time stamp field.
Current compiler implementation of System V ABI selects one of structure
members which has the maximum alignment size in the structure to decide
the size of structure. In the case of fw_cdev_event_request3 structure,
it is closure member which has 8 byte storage. The size of alignment for
the type of 8 byte storage differs depending on architectures; 4 byte for
i386 architecture and 8 byte for the others including x32 architecture.
It is inconvenient to device driver developer to use structure layout
which varies between architectures since the developer takes care of ioctl
compat layer. This commit adds 32 bit member for padding to keep the
size of structure as multiples of 8.
Cc: kunit-dev@googlegroups.com
Link: https://lore.kernel.org/r/20230529113406.986289-4-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
