2019-05-23 17:14:39 +08:00
|
|
|
// SPDX-License-Identifier: GPL-2.0-or-later
|
2007-10-22 09:03:40 +08:00
|
|
|
/* Virtio ring implementation.
|
|
|
|
*
|
|
|
|
* Copyright 2007 Rusty Russell IBM Corporation
|
|
|
|
*/
|
|
|
|
#include <linux/virtio.h>
|
|
|
|
#include <linux/virtio_ring.h>
|
2008-07-26 01:06:13 +08:00
|
|
|
#include <linux/virtio_config.h>
|
2007-10-22 09:03:40 +08:00
|
|
|
#include <linux/device.h>
|
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 16:04:11 +08:00
|
|
|
#include <linux/slab.h>
|
2011-07-04 04:20:30 +08:00
|
|
|
#include <linux/module.h>
|
2012-01-12 13:14:43 +08:00
|
|
|
#include <linux/hrtimer.h>
|
2016-02-03 13:46:36 +08:00
|
|
|
#include <linux/dma-mapping.h>
|
2022-09-15 23:03:56 +08:00
|
|
|
#include <linux/kmsan.h>
|
2021-08-11 00:26:05 +08:00
|
|
|
#include <linux/spinlock.h>
|
2016-02-03 13:46:40 +08:00
|
|
|
#include <xen/xen.h>
|
2007-10-22 09:03:40 +08:00
|
|
|
|
|
|
|
#ifdef DEBUG
|
|
|
|
/* For development, we want to crash whenever the ring is screwed. */
|
2009-06-13 12:16:35 +08:00
|
|
|
#define BAD_RING(_vq, fmt, args...) \
|
|
|
|
do { \
|
|
|
|
dev_err(&(_vq)->vq.vdev->dev, \
|
|
|
|
"%s:"fmt, (_vq)->vq.name, ##args); \
|
|
|
|
BUG(); \
|
|
|
|
} while (0)
|
2009-03-31 11:55:22 +08:00
|
|
|
/* Caller is supposed to guarantee no reentry. */
|
|
|
|
#define START_USE(_vq) \
|
|
|
|
do { \
|
|
|
|
if ((_vq)->in_use) \
|
2009-06-13 12:16:35 +08:00
|
|
|
panic("%s:in_use = %i\n", \
|
|
|
|
(_vq)->vq.name, (_vq)->in_use); \
|
2009-03-31 11:55:22 +08:00
|
|
|
(_vq)->in_use = __LINE__; \
|
2009-06-13 12:16:35 +08:00
|
|
|
} while (0)
|
2009-01-22 23:42:57 +08:00
|
|
|
#define END_USE(_vq) \
|
2010-02-25 04:22:22 +08:00
|
|
|
do { BUG_ON(!(_vq)->in_use); (_vq)->in_use = 0; } while(0)
|
2018-11-21 18:03:22 +08:00
|
|
|
#define LAST_ADD_TIME_UPDATE(_vq) \
|
|
|
|
do { \
|
|
|
|
ktime_t now = ktime_get(); \
|
|
|
|
\
|
|
|
|
/* No kick or get, with .1 second between? Warn. */ \
|
|
|
|
if ((_vq)->last_add_time_valid) \
|
|
|
|
WARN_ON(ktime_to_ms(ktime_sub(now, \
|
|
|
|
(_vq)->last_add_time)) > 100); \
|
|
|
|
(_vq)->last_add_time = now; \
|
|
|
|
(_vq)->last_add_time_valid = true; \
|
|
|
|
} while (0)
|
|
|
|
#define LAST_ADD_TIME_CHECK(_vq) \
|
|
|
|
do { \
|
|
|
|
if ((_vq)->last_add_time_valid) { \
|
|
|
|
WARN_ON(ktime_to_ms(ktime_sub(ktime_get(), \
|
|
|
|
(_vq)->last_add_time)) > 100); \
|
|
|
|
} \
|
|
|
|
} while (0)
|
|
|
|
#define LAST_ADD_TIME_INVALID(_vq) \
|
|
|
|
((_vq)->last_add_time_valid = false)
|
2007-10-22 09:03:40 +08:00
|
|
|
#else
|
2009-06-13 12:16:35 +08:00
|
|
|
#define BAD_RING(_vq, fmt, args...) \
|
|
|
|
do { \
|
|
|
|
dev_err(&_vq->vq.vdev->dev, \
|
|
|
|
"%s:"fmt, (_vq)->vq.name, ##args); \
|
|
|
|
(_vq)->broken = true; \
|
|
|
|
} while (0)
|
2007-10-22 09:03:40 +08:00
|
|
|
#define START_USE(vq)
|
|
|
|
#define END_USE(vq)
|
2018-11-21 18:03:22 +08:00
|
|
|
#define LAST_ADD_TIME_UPDATE(vq)
|
|
|
|
#define LAST_ADD_TIME_CHECK(vq)
|
|
|
|
#define LAST_ADD_TIME_INVALID(vq)
|
2007-10-22 09:03:40 +08:00
|
|
|
#endif
|
|
|
|
|
2018-11-21 18:03:24 +08:00
|
|
|
struct vring_desc_state_split {
|
2016-02-03 13:46:36 +08:00
|
|
|
void *data; /* Data for callback. */
|
|
|
|
struct vring_desc *indir_desc; /* Indirect descriptor, if any. */
|
|
|
|
};
|
|
|
|
|
2018-11-21 18:03:27 +08:00
|
|
|
struct vring_desc_state_packed {
|
|
|
|
void *data; /* Data for callback. */
|
|
|
|
struct vring_packed_desc *indir_desc; /* Indirect descriptor, if any. */
|
|
|
|
u16 num; /* Descriptor list length. */
|
|
|
|
u16 last; /* The last desc state in a list. */
|
|
|
|
};
|
|
|
|
|
2021-06-04 13:53:45 +08:00
|
|
|
struct vring_desc_extra {
|
2021-10-19 15:01:48 +08:00
|
|
|
dma_addr_t addr; /* Descriptor DMA addr. */
|
|
|
|
u32 len; /* Descriptor length. */
|
2018-11-21 18:03:27 +08:00
|
|
|
u16 flags; /* Descriptor flags. */
|
2021-06-04 13:53:44 +08:00
|
|
|
u16 next; /* The next desc state in a list. */
|
2018-11-21 18:03:27 +08:00
|
|
|
};
|
|
|
|
|
2022-08-01 14:38:25 +08:00
|
|
|
struct vring_virtqueue_split {
|
|
|
|
/* Actual memory layout for this queue. */
|
|
|
|
struct vring vring;
|
|
|
|
|
|
|
|
/* Last written value to avail->flags */
|
|
|
|
u16 avail_flags_shadow;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Last written value to avail->idx in
|
|
|
|
* guest byte order.
|
|
|
|
*/
|
|
|
|
u16 avail_idx_shadow;
|
|
|
|
|
|
|
|
/* Per-descriptor state. */
|
|
|
|
struct vring_desc_state_split *desc_state;
|
|
|
|
struct vring_desc_extra *desc_extra;
|
|
|
|
|
|
|
|
/* DMA address and size information */
|
|
|
|
dma_addr_t queue_dma_addr;
|
|
|
|
size_t queue_size_in_bytes;
|
2022-08-01 14:38:35 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* The parameters for creating vrings are reserved for creating new
|
|
|
|
* vring.
|
|
|
|
*/
|
|
|
|
u32 vring_align;
|
|
|
|
bool may_reduce_num;
|
2022-08-01 14:38:25 +08:00
|
|
|
};
|
|
|
|
|
|
|
|
struct vring_virtqueue_packed {
|
|
|
|
/* Actual memory layout for this queue. */
|
|
|
|
struct {
|
|
|
|
unsigned int num;
|
|
|
|
struct vring_packed_desc *desc;
|
|
|
|
struct vring_packed_desc_event *driver;
|
|
|
|
struct vring_packed_desc_event *device;
|
|
|
|
} vring;
|
|
|
|
|
|
|
|
/* Driver ring wrap counter. */
|
|
|
|
bool avail_wrap_counter;
|
|
|
|
|
|
|
|
/* Avail used flags. */
|
|
|
|
u16 avail_used_flags;
|
|
|
|
|
|
|
|
/* Index of the next avail descriptor. */
|
|
|
|
u16 next_avail_idx;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Last written value to driver->flags in
|
|
|
|
* guest byte order.
|
|
|
|
*/
|
|
|
|
u16 event_flags_shadow;
|
|
|
|
|
|
|
|
/* Per-descriptor state. */
|
|
|
|
struct vring_desc_state_packed *desc_state;
|
|
|
|
struct vring_desc_extra *desc_extra;
|
|
|
|
|
|
|
|
/* DMA address and size information */
|
|
|
|
dma_addr_t ring_dma_addr;
|
|
|
|
dma_addr_t driver_event_dma_addr;
|
|
|
|
dma_addr_t device_event_dma_addr;
|
|
|
|
size_t ring_size_in_bytes;
|
|
|
|
size_t event_size_in_bytes;
|
|
|
|
};
|
|
|
|
|
2015-01-15 19:33:31 +08:00
|
|
|
struct vring_virtqueue {
|
2007-10-22 09:03:40 +08:00
|
|
|
struct virtqueue vq;
|
|
|
|
|
2018-11-21 18:03:27 +08:00
|
|
|
/* Is this a packed ring? */
|
|
|
|
bool packed_ring;
|
|
|
|
|
2018-11-21 18:03:26 +08:00
|
|
|
/* Is DMA API used? */
|
|
|
|
bool use_dma_api;
|
|
|
|
|
2012-01-12 13:14:42 +08:00
|
|
|
/* Can we use weak barriers? */
|
|
|
|
bool weak_barriers;
|
|
|
|
|
2007-10-22 09:03:40 +08:00
|
|
|
/* Other side has made a mess, don't try any more. */
|
|
|
|
bool broken;
|
|
|
|
|
2009-05-12 01:11:45 +08:00
|
|
|
/* Host supports indirect buffers */
|
|
|
|
bool indirect;
|
|
|
|
|
2011-05-20 07:10:44 +08:00
|
|
|
/* Host publishes avail event idx */
|
|
|
|
bool event;
|
|
|
|
|
2023-08-10 20:30:48 +08:00
|
|
|
/* Do DMA mapping by driver */
|
|
|
|
bool premapped;
|
|
|
|
|
2023-08-10 20:30:51 +08:00
|
|
|
/* Do unmap or not for desc. Just when premapped is False and
|
|
|
|
* use_dma_api is true, this is true.
|
|
|
|
*/
|
|
|
|
bool do_unmap;
|
|
|
|
|
2007-10-22 09:03:40 +08:00
|
|
|
/* Head of free buffer list. */
|
|
|
|
unsigned int free_head;
|
|
|
|
/* Number we've added since last sync. */
|
|
|
|
unsigned int num_added;
|
|
|
|
|
2022-06-17 10:04:11 +08:00
|
|
|
/* Last used index we've seen.
|
|
|
|
* for split ring, it just contains last used index
|
|
|
|
* for packed ring:
|
|
|
|
* bits up to VRING_PACKED_EVENT_F_WRAP_CTR include the last used index.
|
|
|
|
* bits from VRING_PACKED_EVENT_F_WRAP_CTR include the used wrap counter.
|
|
|
|
*/
|
2007-11-08 05:49:24 +08:00
|
|
|
u16 last_used_idx;
|
2007-10-22 09:03:40 +08:00
|
|
|
|
2021-04-13 13:19:16 +08:00
|
|
|
/* Hint for event idx: already triggered no need to disable. */
|
|
|
|
bool event_triggered;
|
|
|
|
|
2018-11-21 18:03:27 +08:00
|
|
|
union {
|
|
|
|
/* Available for split ring */
|
2022-08-01 14:38:25 +08:00
|
|
|
struct vring_virtqueue_split split;
|
2018-11-21 18:03:21 +08:00
|
|
|
|
2018-11-21 18:03:27 +08:00
|
|
|
/* Available for packed ring */
|
2022-08-01 14:38:25 +08:00
|
|
|
struct vring_virtqueue_packed packed;
|
2018-11-21 18:03:27 +08:00
|
|
|
};
|
virtio_ring: shadow available ring flags & index
Improves cacheline transfer flow of available ring header.
Virtqueues are implemented as a pair of rings, one producer->consumer
avail ring and one consumer->producer used ring; preceding the
avail ring in memory are two contiguous u16 fields -- avail->flags
and avail->idx. A producer posts work by writing to avail->idx and
a consumer reads avail->idx.
The flags and idx fields only need to be written by a producer CPU
and only read by a consumer CPU; when the producer and consumer are
running on different CPUs and the virtio_ring code is structured to
only have source writes/sink reads, we can continuously transfer the
avail header cacheline between 'M' states between cores. This flow
optimizes core -> core bandwidth on certain CPUs.
(see: "Software Optimization Guide for AMD Family 15h Processors",
Section 11.6; similar language appears in the 10h guide and should
apply to CPUs w/ exclusive caches, using LLC as a transfer cache)
Unfortunately the existing virtio_ring code issued reads to the
avail->idx and read-modify-writes to avail->flags on the producer.
This change shadows the flags and index fields in producer memory;
the vring code now reads from the shadows and only ever writes to
avail->flags and avail->idx, allowing the cacheline to transfer
core -> core optimally.
In a concurrent version of vring_bench, the time required for
10,000,000 buffer checkout/returns was reduced by ~2% (average
across many runs) on an AMD Piledriver (15h) CPU:
(w/o shadowing):
Performance counter stats for './vring_bench':
5,451,082,016 L1-dcache-loads
...
2.221477739 seconds time elapsed
(w/ shadowing):
Performance counter stats for './vring_bench':
5,405,701,361 L1-dcache-loads
...
2.168405376 seconds time elapsed
The further away (in a NUMA sense) virtio producers and consumers are
from each other, the more we expect to benefit. Physical implementations
of virtio devices and implementations of virtio where the consumer polls
vring avail indexes (vhost) should also benefit.
Signed-off-by: Venkatesh Srinivas <venkateshs@google.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2015-11-11 08:21:07 +08:00
|
|
|
|
2007-10-22 09:03:40 +08:00
|
|
|
/* How to notify other side. FIXME: commonalize hcalls! */
|
2013-10-29 07:08:50 +08:00
|
|
|
bool (*notify)(struct virtqueue *vq);
|
2007-10-22 09:03:40 +08:00
|
|
|
|
2016-02-03 13:46:37 +08:00
|
|
|
/* DMA, allocation, and size information */
|
|
|
|
bool we_own_ring;
|
|
|
|
|
2023-01-19 14:15:21 +08:00
|
|
|
/* Device used for doing DMA */
|
|
|
|
struct device *dma_dev;
|
|
|
|
|
2007-10-22 09:03:40 +08:00
|
|
|
#ifdef DEBUG
|
|
|
|
/* They're supposed to lock for us. */
|
|
|
|
unsigned int in_use;
|
2012-01-12 13:14:43 +08:00
|
|
|
|
|
|
|
/* Figure out if their kicks are too delayed. */
|
|
|
|
bool last_add_time_valid;
|
|
|
|
ktime_t last_add_time;
|
2007-10-22 09:03:40 +08:00
|
|
|
#endif
|
|
|
|
};
|
|
|
|
|
2022-08-01 14:38:27 +08:00
|
|
|
static struct virtqueue *__vring_new_virtqueue(unsigned int index,
|
2022-08-01 14:38:28 +08:00
|
|
|
struct vring_virtqueue_split *vring_split,
|
2022-08-01 14:38:27 +08:00
|
|
|
struct virtio_device *vdev,
|
|
|
|
bool weak_barriers,
|
|
|
|
bool context,
|
|
|
|
bool (*notify)(struct virtqueue *),
|
|
|
|
void (*callback)(struct virtqueue *),
|
2023-01-19 14:15:21 +08:00
|
|
|
const char *name,
|
|
|
|
struct device *dma_dev);
|
2022-08-01 14:38:31 +08:00
|
|
|
static struct vring_desc_extra *vring_alloc_desc_extra(unsigned int num);
|
2022-08-01 14:38:36 +08:00
|
|
|
static void vring_free(struct virtqueue *_vq);
|
2018-11-21 18:03:20 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Helpers.
|
|
|
|
*/
|
|
|
|
|
2023-03-10 13:34:28 +08:00
|
|
|
#define to_vvq(_vq) container_of_const(_vq, struct vring_virtqueue, vq)
|
2007-10-22 09:03:40 +08:00
|
|
|
|
2023-03-10 13:34:28 +08:00
|
|
|
static bool virtqueue_use_indirect(const struct vring_virtqueue *vq,
|
2023-03-10 13:34:27 +08:00
|
|
|
unsigned int total_sg)
|
2018-11-21 18:03:23 +08:00
|
|
|
{
|
|
|
|
/*
|
|
|
|
* If the host supports indirect descriptor tables, and we have multiple
|
|
|
|
* buffers, then go indirect. FIXME: tune this threshold
|
|
|
|
*/
|
|
|
|
return (vq->indirect && total_sg > 1 && vq->vq.num_free);
|
|
|
|
}
|
|
|
|
|
2016-02-03 13:46:35 +08:00
|
|
|
/*
|
virtio: new feature to detect IOMMU device quirk
The interaction between virtio and IOMMUs is messy.
On most systems with virtio, physical addresses match bus addresses,
and it doesn't particularly matter which one we use to program
the device.
On some systems, including Xen and any system with a physical device
that speaks virtio behind a physical IOMMU, we must program the IOMMU
for virtio DMA to work at all.
On other systems, including SPARC and PPC64, virtio-pci devices are
enumerated as though they are behind an IOMMU, but the virtio host
ignores the IOMMU, so we must either pretend that the IOMMU isn't
there or somehow map everything as the identity.
Add a feature bit to detect that quirk: VIRTIO_F_IOMMU_PLATFORM.
Any device with this feature bit set to 0 needs a quirk and has to be
passed physical addresses (as opposed to bus addresses) even though
the device is behind an IOMMU.
Note: it has to be a per-device quirk because for example, there could
be a mix of passed-through and virtual virtio devices. As another
example, some devices could be implemented by an out of process
hypervisor backend (in case of qemu vhost, or vhost-user) and so support
for an IOMMU needs to be coded up separately.
It would be cleanest to handle this in IOMMU core code, but that needs
per-device DMA ops. While we are waiting for that to be implemented, use
a work-around in virtio core.
Note: a "noiommu" feature is a quirk - add a wrapper to make
that clear.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2016-04-18 17:58:14 +08:00
|
|
|
* Modern virtio devices have feature bits to specify whether they need a
|
|
|
|
* quirk and bypass the IOMMU. If not there, just use the DMA API.
|
|
|
|
*
|
|
|
|
* If there, the interaction between virtio and DMA API is messy.
|
2016-02-03 13:46:35 +08:00
|
|
|
*
|
|
|
|
* On most systems with virtio, physical addresses match bus addresses,
|
|
|
|
* and it doesn't particularly matter whether we use the DMA API.
|
|
|
|
*
|
|
|
|
* On some systems, including Xen and any system with a physical device
|
|
|
|
* that speaks virtio behind a physical IOMMU, we must use the DMA API
|
|
|
|
* for virtio DMA to work at all.
|
|
|
|
*
|
|
|
|
* On other systems, including SPARC and PPC64, virtio-pci devices are
|
|
|
|
* enumerated as though they are behind an IOMMU, but the virtio host
|
|
|
|
* ignores the IOMMU, so we must either pretend that the IOMMU isn't
|
|
|
|
* there or somehow map everything as the identity.
|
|
|
|
*
|
|
|
|
* For the time being, we preserve historic behavior and bypass the DMA
|
|
|
|
* API.
|
virtio: new feature to detect IOMMU device quirk
The interaction between virtio and IOMMUs is messy.
On most systems with virtio, physical addresses match bus addresses,
and it doesn't particularly matter which one we use to program
the device.
On some systems, including Xen and any system with a physical device
that speaks virtio behind a physical IOMMU, we must program the IOMMU
for virtio DMA to work at all.
On other systems, including SPARC and PPC64, virtio-pci devices are
enumerated as though they are behind an IOMMU, but the virtio host
ignores the IOMMU, so we must either pretend that the IOMMU isn't
there or somehow map everything as the identity.
Add a feature bit to detect that quirk: VIRTIO_F_IOMMU_PLATFORM.
Any device with this feature bit set to 0 needs a quirk and has to be
passed physical addresses (as opposed to bus addresses) even though
the device is behind an IOMMU.
Note: it has to be a per-device quirk because for example, there could
be a mix of passed-through and virtual virtio devices. As another
example, some devices could be implemented by an out of process
hypervisor backend (in case of qemu vhost, or vhost-user) and so support
for an IOMMU needs to be coded up separately.
It would be cleanest to handle this in IOMMU core code, but that needs
per-device DMA ops. While we are waiting for that to be implemented, use
a work-around in virtio core.
Note: a "noiommu" feature is a quirk - add a wrapper to make
that clear.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2016-04-18 17:58:14 +08:00
|
|
|
*
|
|
|
|
* TODO: install a per-device DMA ops structure that does the right thing
|
|
|
|
* taking into account all the above quirks, and use the DMA API
|
|
|
|
* unconditionally on data path.
|
2016-02-03 13:46:35 +08:00
|
|
|
*/
|
|
|
|
|
2023-03-10 13:34:28 +08:00
|
|
|
static bool vring_use_dma_api(const struct virtio_device *vdev)
|
2016-02-03 13:46:35 +08:00
|
|
|
{
|
2020-06-25 07:17:04 +08:00
|
|
|
if (!virtio_has_dma_quirk(vdev))
|
virtio: new feature to detect IOMMU device quirk
The interaction between virtio and IOMMUs is messy.
On most systems with virtio, physical addresses match bus addresses,
and it doesn't particularly matter which one we use to program
the device.
On some systems, including Xen and any system with a physical device
that speaks virtio behind a physical IOMMU, we must program the IOMMU
for virtio DMA to work at all.
On other systems, including SPARC and PPC64, virtio-pci devices are
enumerated as though they are behind an IOMMU, but the virtio host
ignores the IOMMU, so we must either pretend that the IOMMU isn't
there or somehow map everything as the identity.
Add a feature bit to detect that quirk: VIRTIO_F_IOMMU_PLATFORM.
Any device with this feature bit set to 0 needs a quirk and has to be
passed physical addresses (as opposed to bus addresses) even though
the device is behind an IOMMU.
Note: it has to be a per-device quirk because for example, there could
be a mix of passed-through and virtual virtio devices. As another
example, some devices could be implemented by an out of process
hypervisor backend (in case of qemu vhost, or vhost-user) and so support
for an IOMMU needs to be coded up separately.
It would be cleanest to handle this in IOMMU core code, but that needs
per-device DMA ops. While we are waiting for that to be implemented, use
a work-around in virtio core.
Note: a "noiommu" feature is a quirk - add a wrapper to make
that clear.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2016-04-18 17:58:14 +08:00
|
|
|
return true;
|
|
|
|
|
|
|
|
/* Otherwise, we are left to guess. */
|
2016-02-03 13:46:40 +08:00
|
|
|
/*
|
|
|
|
* In theory, it's possible to have a buggy QEMU-supposed
|
|
|
|
* emulated Q35 IOMMU and Xen enabled at the same time. On
|
|
|
|
* such a configuration, virtio has never worked and will
|
|
|
|
* not work without an even larger kludge. Instead, enable
|
|
|
|
* the DMA API if we're a Xen guest, which at least allows
|
|
|
|
* all of the sensible Xen configurations to work correctly.
|
|
|
|
*/
|
|
|
|
if (xen_domain())
|
|
|
|
return true;
|
|
|
|
|
2016-02-03 13:46:35 +08:00
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2023-03-10 13:34:28 +08:00
|
|
|
size_t virtio_max_dma_size(const struct virtio_device *vdev)
|
2019-02-07 19:59:16 +08:00
|
|
|
{
|
|
|
|
size_t max_segment_size = SIZE_MAX;
|
|
|
|
|
|
|
|
if (vring_use_dma_api(vdev))
|
2021-12-01 19:20:18 +08:00
|
|
|
max_segment_size = dma_max_mapping_size(vdev->dev.parent);
|
2019-02-07 19:59:16 +08:00
|
|
|
|
|
|
|
return max_segment_size;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtio_max_dma_size);
|
|
|
|
|
2018-11-21 18:03:25 +08:00
|
|
|
static void *vring_alloc_queue(struct virtio_device *vdev, size_t size,
|
2023-01-19 14:15:21 +08:00
|
|
|
dma_addr_t *dma_handle, gfp_t flag,
|
|
|
|
struct device *dma_dev)
|
2018-11-21 18:03:25 +08:00
|
|
|
{
|
|
|
|
if (vring_use_dma_api(vdev)) {
|
2023-01-19 14:15:21 +08:00
|
|
|
return dma_alloc_coherent(dma_dev, size,
|
2018-11-21 18:03:25 +08:00
|
|
|
dma_handle, flag);
|
|
|
|
} else {
|
|
|
|
void *queue = alloc_pages_exact(PAGE_ALIGN(size), flag);
|
|
|
|
|
|
|
|
if (queue) {
|
|
|
|
phys_addr_t phys_addr = virt_to_phys(queue);
|
|
|
|
*dma_handle = (dma_addr_t)phys_addr;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Sanity check: make sure we dind't truncate
|
|
|
|
* the address. The only arches I can find that
|
|
|
|
* have 64-bit phys_addr_t but 32-bit dma_addr_t
|
|
|
|
* are certain non-highmem MIPS and x86
|
|
|
|
* configurations, but these configurations
|
|
|
|
* should never allocate physical pages above 32
|
|
|
|
* bits, so this is fine. Just in case, throw a
|
|
|
|
* warning and abort if we end up with an
|
|
|
|
* unrepresentable address.
|
|
|
|
*/
|
|
|
|
if (WARN_ON_ONCE(*dma_handle != phys_addr)) {
|
|
|
|
free_pages_exact(queue, PAGE_ALIGN(size));
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return queue;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void vring_free_queue(struct virtio_device *vdev, size_t size,
|
2023-01-19 14:15:21 +08:00
|
|
|
void *queue, dma_addr_t dma_handle,
|
|
|
|
struct device *dma_dev)
|
2018-11-21 18:03:25 +08:00
|
|
|
{
|
|
|
|
if (vring_use_dma_api(vdev))
|
2023-01-19 14:15:21 +08:00
|
|
|
dma_free_coherent(dma_dev, size, queue, dma_handle);
|
2018-11-21 18:03:25 +08:00
|
|
|
else
|
|
|
|
free_pages_exact(queue, PAGE_ALIGN(size));
|
|
|
|
}
|
|
|
|
|
2016-02-03 13:46:36 +08:00
|
|
|
/*
|
|
|
|
* The DMA ops on various arches are rather gnarly right now, and
|
|
|
|
* making all of the arch DMA ops work on the vring device itself
|
2023-01-19 14:15:21 +08:00
|
|
|
* is a mess.
|
2016-02-03 13:46:36 +08:00
|
|
|
*/
|
2023-03-10 13:34:27 +08:00
|
|
|
static struct device *vring_dma_dev(const struct vring_virtqueue *vq)
|
2016-02-03 13:46:36 +08:00
|
|
|
{
|
2023-01-19 14:15:21 +08:00
|
|
|
return vq->dma_dev;
|
2016-02-03 13:46:36 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/* Map one sg entry. */
|
2023-08-10 20:30:47 +08:00
|
|
|
static int vring_map_one_sg(const struct vring_virtqueue *vq, struct scatterlist *sg,
|
|
|
|
enum dma_data_direction direction, dma_addr_t *addr)
|
2016-02-03 13:46:36 +08:00
|
|
|
{
|
2023-08-10 20:30:49 +08:00
|
|
|
if (vq->premapped) {
|
|
|
|
*addr = sg_dma_address(sg);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2022-09-15 23:03:56 +08:00
|
|
|
if (!vq->use_dma_api) {
|
|
|
|
/*
|
|
|
|
* If DMA is not used, KMSAN doesn't know that the scatterlist
|
|
|
|
* is initialized by the hardware. Explicitly check/unpoison it
|
|
|
|
* depending on the direction.
|
|
|
|
*/
|
|
|
|
kmsan_handle_dma(sg_page(sg), sg->offset, sg->length, direction);
|
2023-08-10 20:30:47 +08:00
|
|
|
*addr = (dma_addr_t)sg_phys(sg);
|
|
|
|
return 0;
|
2022-09-15 23:03:56 +08:00
|
|
|
}
|
2016-02-03 13:46:36 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* We can't use dma_map_sg, because we don't use scatterlists in
|
|
|
|
* the way it expects (we don't guarantee that the scatterlist
|
|
|
|
* will exist for the lifetime of the mapping).
|
|
|
|
*/
|
2023-08-10 20:30:47 +08:00
|
|
|
*addr = dma_map_page(vring_dma_dev(vq),
|
2016-02-03 13:46:36 +08:00
|
|
|
sg_page(sg), sg->offset, sg->length,
|
|
|
|
direction);
|
2023-08-10 20:30:47 +08:00
|
|
|
|
|
|
|
if (dma_mapping_error(vring_dma_dev(vq), *addr))
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
return 0;
|
2016-02-03 13:46:36 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static dma_addr_t vring_map_single(const struct vring_virtqueue *vq,
|
|
|
|
void *cpu_addr, size_t size,
|
|
|
|
enum dma_data_direction direction)
|
|
|
|
{
|
2018-11-21 18:03:26 +08:00
|
|
|
if (!vq->use_dma_api)
|
2016-02-03 13:46:36 +08:00
|
|
|
return (dma_addr_t)virt_to_phys(cpu_addr);
|
|
|
|
|
|
|
|
return dma_map_single(vring_dma_dev(vq),
|
|
|
|
cpu_addr, size, direction);
|
|
|
|
}
|
|
|
|
|
2018-11-21 18:03:20 +08:00
|
|
|
static int vring_mapping_error(const struct vring_virtqueue *vq,
|
|
|
|
dma_addr_t addr)
|
|
|
|
{
|
2018-11-21 18:03:26 +08:00
|
|
|
if (!vq->use_dma_api)
|
2018-11-21 18:03:20 +08:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
return dma_mapping_error(vring_dma_dev(vq), addr);
|
|
|
|
}
|
|
|
|
|
2022-08-01 14:38:26 +08:00
|
|
|
static void virtqueue_init(struct vring_virtqueue *vq, u32 num)
|
|
|
|
{
|
|
|
|
vq->vq.num_free = num;
|
|
|
|
|
|
|
|
if (vq->packed_ring)
|
|
|
|
vq->last_used_idx = 0 | (1 << VRING_PACKED_EVENT_F_WRAP_CTR);
|
|
|
|
else
|
|
|
|
vq->last_used_idx = 0;
|
|
|
|
|
|
|
|
vq->event_triggered = false;
|
|
|
|
vq->num_added = 0;
|
|
|
|
|
|
|
|
#ifdef DEBUG
|
|
|
|
vq->in_use = false;
|
|
|
|
vq->last_add_time_valid = false;
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
|
2018-11-21 18:03:20 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Split ring specific functions - *_split().
|
|
|
|
*/
|
|
|
|
|
2021-06-04 13:53:50 +08:00
|
|
|
static void vring_unmap_one_split_indirect(const struct vring_virtqueue *vq,
|
2023-03-10 13:34:28 +08:00
|
|
|
const struct vring_desc *desc)
|
2016-02-03 13:46:36 +08:00
|
|
|
{
|
|
|
|
u16 flags;
|
|
|
|
|
2023-08-10 20:30:51 +08:00
|
|
|
if (!vq->do_unmap)
|
2016-02-03 13:46:36 +08:00
|
|
|
return;
|
|
|
|
|
|
|
|
flags = virtio16_to_cpu(vq->vq.vdev, desc->flags);
|
|
|
|
|
2022-02-24 19:03:55 +08:00
|
|
|
dma_unmap_page(vring_dma_dev(vq),
|
|
|
|
virtio64_to_cpu(vq->vq.vdev, desc->addr),
|
|
|
|
virtio32_to_cpu(vq->vq.vdev, desc->len),
|
|
|
|
(flags & VRING_DESC_F_WRITE) ?
|
|
|
|
DMA_FROM_DEVICE : DMA_TO_DEVICE);
|
2016-02-03 13:46:36 +08:00
|
|
|
}
|
|
|
|
|
2021-06-04 13:53:50 +08:00
|
|
|
static unsigned int vring_unmap_one_split(const struct vring_virtqueue *vq,
|
|
|
|
unsigned int i)
|
|
|
|
{
|
|
|
|
struct vring_desc_extra *extra = vq->split.desc_extra;
|
|
|
|
u16 flags;
|
|
|
|
|
|
|
|
flags = extra[i].flags;
|
|
|
|
|
|
|
|
if (flags & VRING_DESC_F_INDIRECT) {
|
2023-08-10 20:30:51 +08:00
|
|
|
if (!vq->use_dma_api)
|
|
|
|
goto out;
|
|
|
|
|
2021-06-04 13:53:50 +08:00
|
|
|
dma_unmap_single(vring_dma_dev(vq),
|
|
|
|
extra[i].addr,
|
|
|
|
extra[i].len,
|
|
|
|
(flags & VRING_DESC_F_WRITE) ?
|
|
|
|
DMA_FROM_DEVICE : DMA_TO_DEVICE);
|
|
|
|
} else {
|
2023-08-10 20:30:51 +08:00
|
|
|
if (!vq->do_unmap)
|
|
|
|
goto out;
|
|
|
|
|
2021-06-04 13:53:50 +08:00
|
|
|
dma_unmap_page(vring_dma_dev(vq),
|
|
|
|
extra[i].addr,
|
|
|
|
extra[i].len,
|
|
|
|
(flags & VRING_DESC_F_WRITE) ?
|
|
|
|
DMA_FROM_DEVICE : DMA_TO_DEVICE);
|
|
|
|
}
|
|
|
|
|
|
|
|
out:
|
|
|
|
return extra[i].next;
|
|
|
|
}
|
|
|
|
|
2018-11-21 18:03:19 +08:00
|
|
|
static struct vring_desc *alloc_indirect_split(struct virtqueue *_vq,
|
|
|
|
unsigned int total_sg,
|
|
|
|
gfp_t gfp)
|
2009-05-12 01:11:45 +08:00
|
|
|
{
|
|
|
|
struct vring_desc *desc;
|
2014-09-11 08:47:38 +08:00
|
|
|
unsigned int i;
|
2009-05-12 01:11:45 +08:00
|
|
|
|
2012-10-19 21:03:33 +08:00
|
|
|
/*
|
|
|
|
* We require lowmem mappings for the descriptors because
|
|
|
|
* otherwise virt_to_phys will give us bogus addresses in the
|
|
|
|
* virtqueue.
|
|
|
|
*/
|
2015-12-01 22:32:49 +08:00
|
|
|
gfp &= ~__GFP_HIGHMEM;
|
2012-10-19 21:03:33 +08:00
|
|
|
|
treewide: kmalloc() -> kmalloc_array()
The kmalloc() function has a 2-factor argument form, kmalloc_array(). This
patch replaces cases of:
kmalloc(a * b, gfp)
with:
kmalloc_array(a * b, gfp)
as well as handling cases of:
kmalloc(a * b * c, gfp)
with:
kmalloc(array3_size(a, b, c), gfp)
as it's slightly less ugly than:
kmalloc_array(array_size(a, b), c, gfp)
This does, however, attempt to ignore constant size factors like:
kmalloc(4 * 1024, gfp)
though any constants defined via macros get caught up in the conversion.
Any factors with a sizeof() of "unsigned char", "char", and "u8" were
dropped, since they're redundant.
The tools/ directory was manually excluded, since it has its own
implementation of kmalloc().
The Coccinelle script used for this was:
// Fix redundant parens around sizeof().
@@
type TYPE;
expression THING, E;
@@
(
kmalloc(
- (sizeof(TYPE)) * E
+ sizeof(TYPE) * E
, ...)
|
kmalloc(
- (sizeof(THING)) * E
+ sizeof(THING) * E
, ...)
)
// Drop single-byte sizes and redundant parens.
@@
expression COUNT;
typedef u8;
typedef __u8;
@@
(
kmalloc(
- sizeof(u8) * (COUNT)
+ COUNT
, ...)
|
kmalloc(
- sizeof(__u8) * (COUNT)
+ COUNT
, ...)
|
kmalloc(
- sizeof(char) * (COUNT)
+ COUNT
, ...)
|
kmalloc(
- sizeof(unsigned char) * (COUNT)
+ COUNT
, ...)
|
kmalloc(
- sizeof(u8) * COUNT
+ COUNT
, ...)
|
kmalloc(
- sizeof(__u8) * COUNT
+ COUNT
, ...)
|
kmalloc(
- sizeof(char) * COUNT
+ COUNT
, ...)
|
kmalloc(
- sizeof(unsigned char) * COUNT
+ COUNT
, ...)
)
// 2-factor product with sizeof(type/expression) and identifier or constant.
@@
type TYPE;
expression THING;
identifier COUNT_ID;
constant COUNT_CONST;
@@
(
- kmalloc
+ kmalloc_array
(
- sizeof(TYPE) * (COUNT_ID)
+ COUNT_ID, sizeof(TYPE)
, ...)
|
- kmalloc
+ kmalloc_array
(
- sizeof(TYPE) * COUNT_ID
+ COUNT_ID, sizeof(TYPE)
, ...)
|
- kmalloc
+ kmalloc_array
(
- sizeof(TYPE) * (COUNT_CONST)
+ COUNT_CONST, sizeof(TYPE)
, ...)
|
- kmalloc
+ kmalloc_array
(
- sizeof(TYPE) * COUNT_CONST
+ COUNT_CONST, sizeof(TYPE)
, ...)
|
- kmalloc
+ kmalloc_array
(
- sizeof(THING) * (COUNT_ID)
+ COUNT_ID, sizeof(THING)
, ...)
|
- kmalloc
+ kmalloc_array
(
- sizeof(THING) * COUNT_ID
+ COUNT_ID, sizeof(THING)
, ...)
|
- kmalloc
+ kmalloc_array
(
- sizeof(THING) * (COUNT_CONST)
+ COUNT_CONST, sizeof(THING)
, ...)
|
- kmalloc
+ kmalloc_array
(
- sizeof(THING) * COUNT_CONST
+ COUNT_CONST, sizeof(THING)
, ...)
)
// 2-factor product, only identifiers.
@@
identifier SIZE, COUNT;
@@
- kmalloc
+ kmalloc_array
(
- SIZE * COUNT
+ COUNT, SIZE
, ...)
// 3-factor product with 1 sizeof(type) or sizeof(expression), with
// redundant parens removed.
@@
expression THING;
identifier STRIDE, COUNT;
type TYPE;
@@
(
kmalloc(
- sizeof(TYPE) * (COUNT) * (STRIDE)
+ array3_size(COUNT, STRIDE, sizeof(TYPE))
, ...)
|
kmalloc(
- sizeof(TYPE) * (COUNT) * STRIDE
+ array3_size(COUNT, STRIDE, sizeof(TYPE))
, ...)
|
kmalloc(
- sizeof(TYPE) * COUNT * (STRIDE)
+ array3_size(COUNT, STRIDE, sizeof(TYPE))
, ...)
|
kmalloc(
- sizeof(TYPE) * COUNT * STRIDE
+ array3_size(COUNT, STRIDE, sizeof(TYPE))
, ...)
|
kmalloc(
- sizeof(THING) * (COUNT) * (STRIDE)
+ array3_size(COUNT, STRIDE, sizeof(THING))
, ...)
|
kmalloc(
- sizeof(THING) * (COUNT) * STRIDE
+ array3_size(COUNT, STRIDE, sizeof(THING))
, ...)
|
kmalloc(
- sizeof(THING) * COUNT * (STRIDE)
+ array3_size(COUNT, STRIDE, sizeof(THING))
, ...)
|
kmalloc(
- sizeof(THING) * COUNT * STRIDE
+ array3_size(COUNT, STRIDE, sizeof(THING))
, ...)
)
// 3-factor product with 2 sizeof(variable), with redundant parens removed.
@@
expression THING1, THING2;
identifier COUNT;
type TYPE1, TYPE2;
@@
(
kmalloc(
- sizeof(TYPE1) * sizeof(TYPE2) * COUNT
+ array3_size(COUNT, sizeof(TYPE1), sizeof(TYPE2))
, ...)
|
kmalloc(
- sizeof(TYPE1) * sizeof(THING2) * (COUNT)
+ array3_size(COUNT, sizeof(TYPE1), sizeof(TYPE2))
, ...)
|
kmalloc(
- sizeof(THING1) * sizeof(THING2) * COUNT
+ array3_size(COUNT, sizeof(THING1), sizeof(THING2))
, ...)
|
kmalloc(
- sizeof(THING1) * sizeof(THING2) * (COUNT)
+ array3_size(COUNT, sizeof(THING1), sizeof(THING2))
, ...)
|
kmalloc(
- sizeof(TYPE1) * sizeof(THING2) * COUNT
+ array3_size(COUNT, sizeof(TYPE1), sizeof(THING2))
, ...)
|
kmalloc(
- sizeof(TYPE1) * sizeof(THING2) * (COUNT)
+ array3_size(COUNT, sizeof(TYPE1), sizeof(THING2))
, ...)
)
// 3-factor product, only identifiers, with redundant parens removed.
@@
identifier STRIDE, SIZE, COUNT;
@@
(
kmalloc(
- (COUNT) * STRIDE * SIZE
+ array3_size(COUNT, STRIDE, SIZE)
, ...)
|
kmalloc(
- COUNT * (STRIDE) * SIZE
+ array3_size(COUNT, STRIDE, SIZE)
, ...)
|
kmalloc(
- COUNT * STRIDE * (SIZE)
+ array3_size(COUNT, STRIDE, SIZE)
, ...)
|
kmalloc(
- (COUNT) * (STRIDE) * SIZE
+ array3_size(COUNT, STRIDE, SIZE)
, ...)
|
kmalloc(
- COUNT * (STRIDE) * (SIZE)
+ array3_size(COUNT, STRIDE, SIZE)
, ...)
|
kmalloc(
- (COUNT) * STRIDE * (SIZE)
+ array3_size(COUNT, STRIDE, SIZE)
, ...)
|
kmalloc(
- (COUNT) * (STRIDE) * (SIZE)
+ array3_size(COUNT, STRIDE, SIZE)
, ...)
|
kmalloc(
- COUNT * STRIDE * SIZE
+ array3_size(COUNT, STRIDE, SIZE)
, ...)
)
// Any remaining multi-factor products, first at least 3-factor products,
// when they're not all constants...
@@
expression E1, E2, E3;
constant C1, C2, C3;
@@
(
kmalloc(C1 * C2 * C3, ...)
|
kmalloc(
- (E1) * E2 * E3
+ array3_size(E1, E2, E3)
, ...)
|
kmalloc(
- (E1) * (E2) * E3
+ array3_size(E1, E2, E3)
, ...)
|
kmalloc(
- (E1) * (E2) * (E3)
+ array3_size(E1, E2, E3)
, ...)
|
kmalloc(
- E1 * E2 * E3
+ array3_size(E1, E2, E3)
, ...)
)
// And then all remaining 2 factors products when they're not all constants,
// keeping sizeof() as the second factor argument.
@@
expression THING, E1, E2;
type TYPE;
constant C1, C2, C3;
@@
(
kmalloc(sizeof(THING) * C2, ...)
|
kmalloc(sizeof(TYPE) * C2, ...)
|
kmalloc(C1 * C2 * C3, ...)
|
kmalloc(C1 * C2, ...)
|
- kmalloc
+ kmalloc_array
(
- sizeof(TYPE) * (E2)
+ E2, sizeof(TYPE)
, ...)
|
- kmalloc
+ kmalloc_array
(
- sizeof(TYPE) * E2
+ E2, sizeof(TYPE)
, ...)
|
- kmalloc
+ kmalloc_array
(
- sizeof(THING) * (E2)
+ E2, sizeof(THING)
, ...)
|
- kmalloc
+ kmalloc_array
(
- sizeof(THING) * E2
+ E2, sizeof(THING)
, ...)
|
- kmalloc
+ kmalloc_array
(
- (E1) * E2
+ E1, E2
, ...)
|
- kmalloc
+ kmalloc_array
(
- (E1) * (E2)
+ E1, E2
, ...)
|
- kmalloc
+ kmalloc_array
(
- E1 * E2
+ E1, E2
, ...)
)
Signed-off-by: Kees Cook <keescook@chromium.org>
2018-06-13 04:55:00 +08:00
|
|
|
desc = kmalloc_array(total_sg, sizeof(struct vring_desc), gfp);
|
2009-05-12 01:11:45 +08:00
|
|
|
if (!desc)
|
2014-09-11 08:47:38 +08:00
|
|
|
return NULL;
|
2009-05-12 01:11:45 +08:00
|
|
|
|
2014-09-11 08:47:38 +08:00
|
|
|
for (i = 0; i < total_sg; i++)
|
2014-10-22 20:42:09 +08:00
|
|
|
desc[i].next = cpu_to_virtio16(_vq->vdev, i + 1);
|
2014-09-11 08:47:38 +08:00
|
|
|
return desc;
|
2009-05-12 01:11:45 +08:00
|
|
|
}
|
|
|
|
|
2021-06-04 13:53:48 +08:00
|
|
|
static inline unsigned int virtqueue_add_desc_split(struct virtqueue *vq,
|
|
|
|
struct vring_desc *desc,
|
|
|
|
unsigned int i,
|
|
|
|
dma_addr_t addr,
|
|
|
|
unsigned int len,
|
2021-06-04 13:53:50 +08:00
|
|
|
u16 flags,
|
|
|
|
bool indirect)
|
2021-06-04 13:53:48 +08:00
|
|
|
{
|
2021-06-04 13:53:50 +08:00
|
|
|
struct vring_virtqueue *vring = to_vvq(vq);
|
|
|
|
struct vring_desc_extra *extra = vring->split.desc_extra;
|
|
|
|
u16 next;
|
|
|
|
|
2021-06-04 13:53:48 +08:00
|
|
|
desc[i].flags = cpu_to_virtio16(vq->vdev, flags);
|
|
|
|
desc[i].addr = cpu_to_virtio64(vq->vdev, addr);
|
|
|
|
desc[i].len = cpu_to_virtio32(vq->vdev, len);
|
|
|
|
|
2021-06-04 13:53:50 +08:00
|
|
|
if (!indirect) {
|
|
|
|
next = extra[i].next;
|
|
|
|
desc[i].next = cpu_to_virtio16(vq->vdev, next);
|
|
|
|
|
|
|
|
extra[i].addr = addr;
|
|
|
|
extra[i].len = len;
|
|
|
|
extra[i].flags = flags;
|
|
|
|
} else
|
|
|
|
next = virtio16_to_cpu(vq->vdev, desc[i].next);
|
|
|
|
|
|
|
|
return next;
|
2021-06-04 13:53:48 +08:00
|
|
|
}
|
|
|
|
|
2018-11-21 18:03:19 +08:00
|
|
|
static inline int virtqueue_add_split(struct virtqueue *_vq,
|
|
|
|
struct scatterlist *sgs[],
|
|
|
|
unsigned int total_sg,
|
|
|
|
unsigned int out_sgs,
|
|
|
|
unsigned int in_sgs,
|
|
|
|
void *data,
|
|
|
|
void *ctx,
|
|
|
|
gfp_t gfp)
|
2007-10-22 09:03:40 +08:00
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
2013-03-20 13:07:09 +08:00
|
|
|
struct scatterlist *sg;
|
2014-09-11 08:47:38 +08:00
|
|
|
struct vring_desc *desc;
|
treewide: Remove uninitialized_var() usage
Using uninitialized_var() is dangerous as it papers over real bugs[1]
(or can in the future), and suppresses unrelated compiler warnings
(e.g. "unused variable"). If the compiler thinks it is uninitialized,
either simply initialize the variable or make compiler changes.
In preparation for removing[2] the[3] macro[4], remove all remaining
needless uses with the following script:
git grep '\buninitialized_var\b' | cut -d: -f1 | sort -u | \
xargs perl -pi -e \
's/\buninitialized_var\(([^\)]+)\)/\1/g;
s:\s*/\* (GCC be quiet|to make compiler happy) \*/$::g;'
drivers/video/fbdev/riva/riva_hw.c was manually tweaked to avoid
pathological white-space.
No outstanding warnings were found building allmodconfig with GCC 9.3.0
for x86_64, i386, arm64, arm, powerpc, powerpc64le, s390x, mips, sparc64,
alpha, and m68k.
[1] https://lore.kernel.org/lkml/20200603174714.192027-1-glider@google.com/
[2] https://lore.kernel.org/lkml/CA+55aFw+Vbj0i=1TGqCR5vQkCzWJ0QxK6CernOU6eedsudAixw@mail.gmail.com/
[3] https://lore.kernel.org/lkml/CA+55aFwgbgqhbp1fkxvRKEpzyR5J8n1vKT1VZdz9knmPuXhOeg@mail.gmail.com/
[4] https://lore.kernel.org/lkml/CA+55aFz2500WfbKXAx8s67wrm9=yVJu65TpLgN_ybYNv0VEOKA@mail.gmail.com/
Reviewed-by: Leon Romanovsky <leonro@mellanox.com> # drivers/infiniband and mlx4/mlx5
Acked-by: Jason Gunthorpe <jgg@mellanox.com> # IB
Acked-by: Kalle Valo <kvalo@codeaurora.org> # wireless drivers
Reviewed-by: Chao Yu <yuchao0@huawei.com> # erofs
Signed-off-by: Kees Cook <keescook@chromium.org>
2020-06-04 04:09:38 +08:00
|
|
|
unsigned int i, n, avail, descs_used, prev, err_idx;
|
2010-07-26 15:25:30 +08:00
|
|
|
int head;
|
2014-09-11 08:47:38 +08:00
|
|
|
bool indirect;
|
2007-10-22 09:03:40 +08:00
|
|
|
|
2009-05-12 01:11:45 +08:00
|
|
|
START_USE(vq);
|
|
|
|
|
2007-10-22 09:03:40 +08:00
|
|
|
BUG_ON(data == NULL);
|
2017-02-07 12:15:13 +08:00
|
|
|
BUG_ON(ctx && vq->indirect);
|
2009-05-12 01:11:45 +08:00
|
|
|
|
2014-03-13 08:53:40 +08:00
|
|
|
if (unlikely(vq->broken)) {
|
|
|
|
END_USE(vq);
|
|
|
|
return -EIO;
|
|
|
|
}
|
|
|
|
|
2018-11-21 18:03:22 +08:00
|
|
|
LAST_ADD_TIME_UPDATE(vq);
|
2012-01-12 13:14:43 +08:00
|
|
|
|
2014-09-11 08:47:38 +08:00
|
|
|
BUG_ON(total_sg == 0);
|
|
|
|
|
|
|
|
head = vq->free_head;
|
|
|
|
|
2022-03-28 18:58:16 +08:00
|
|
|
if (virtqueue_use_indirect(vq, total_sg))
|
2018-11-21 18:03:19 +08:00
|
|
|
desc = alloc_indirect_split(_vq, total_sg, gfp);
|
2017-08-11 00:56:51 +08:00
|
|
|
else {
|
2014-09-11 08:47:38 +08:00
|
|
|
desc = NULL;
|
2018-11-21 18:03:21 +08:00
|
|
|
WARN_ON_ONCE(total_sg > vq->split.vring.num && !vq->indirect);
|
2017-08-11 00:56:51 +08:00
|
|
|
}
|
2014-09-11 08:47:38 +08:00
|
|
|
|
|
|
|
if (desc) {
|
|
|
|
/* Use a single buffer which doesn't continue */
|
2016-02-03 13:46:36 +08:00
|
|
|
indirect = true;
|
2014-09-11 08:47:38 +08:00
|
|
|
/* Set up rest to use this indirect table. */
|
|
|
|
i = 0;
|
|
|
|
descs_used = 1;
|
|
|
|
} else {
|
2016-02-03 13:46:36 +08:00
|
|
|
indirect = false;
|
2018-11-21 18:03:21 +08:00
|
|
|
desc = vq->split.vring.desc;
|
2014-09-11 08:47:38 +08:00
|
|
|
i = head;
|
|
|
|
descs_used = total_sg;
|
2009-05-12 01:11:45 +08:00
|
|
|
}
|
|
|
|
|
2022-03-28 18:58:17 +08:00
|
|
|
if (unlikely(vq->vq.num_free < descs_used)) {
|
2007-10-22 09:03:40 +08:00
|
|
|
pr_debug("Can't add buf len %i - avail = %i\n",
|
2014-09-11 08:47:38 +08:00
|
|
|
descs_used, vq->vq.num_free);
|
2008-07-26 01:06:04 +08:00
|
|
|
/* FIXME: for historical reasons, we force a notify here if
|
|
|
|
* there are outgoing parts to the buffer. Presumably the
|
|
|
|
* host should service the ring ASAP. */
|
2013-03-20 13:07:09 +08:00
|
|
|
if (out_sgs)
|
2008-07-26 01:06:04 +08:00
|
|
|
vq->notify(&vq->vq);
|
2016-08-02 22:16:31 +08:00
|
|
|
if (indirect)
|
|
|
|
kfree(desc);
|
2007-10-22 09:03:40 +08:00
|
|
|
END_USE(vq);
|
|
|
|
return -ENOSPC;
|
|
|
|
}
|
|
|
|
|
2013-03-20 13:07:09 +08:00
|
|
|
for (n = 0; n < out_sgs; n++) {
|
2014-09-11 08:47:37 +08:00
|
|
|
for (sg = sgs[n]; sg; sg = sg_next(sg)) {
|
2023-08-10 20:30:47 +08:00
|
|
|
dma_addr_t addr;
|
|
|
|
|
|
|
|
if (vring_map_one_sg(vq, sg, DMA_TO_DEVICE, &addr))
|
2016-02-03 13:46:36 +08:00
|
|
|
goto unmap_release;
|
|
|
|
|
2013-03-20 13:07:09 +08:00
|
|
|
prev = i;
|
2021-06-04 13:53:50 +08:00
|
|
|
/* Note that we trust indirect descriptor
|
|
|
|
* table since it use stream DMA mapping.
|
|
|
|
*/
|
2021-06-04 13:53:48 +08:00
|
|
|
i = virtqueue_add_desc_split(_vq, desc, i, addr, sg->length,
|
2021-06-04 13:53:50 +08:00
|
|
|
VRING_DESC_F_NEXT,
|
|
|
|
indirect);
|
2013-03-20 13:07:09 +08:00
|
|
|
}
|
2007-10-22 09:03:40 +08:00
|
|
|
}
|
2013-03-20 13:07:09 +08:00
|
|
|
for (; n < (out_sgs + in_sgs); n++) {
|
2014-09-11 08:47:37 +08:00
|
|
|
for (sg = sgs[n]; sg; sg = sg_next(sg)) {
|
2023-08-10 20:30:47 +08:00
|
|
|
dma_addr_t addr;
|
|
|
|
|
|
|
|
if (vring_map_one_sg(vq, sg, DMA_FROM_DEVICE, &addr))
|
2016-02-03 13:46:36 +08:00
|
|
|
goto unmap_release;
|
|
|
|
|
2013-03-20 13:07:09 +08:00
|
|
|
prev = i;
|
2021-06-04 13:53:50 +08:00
|
|
|
/* Note that we trust indirect descriptor
|
|
|
|
* table since it use stream DMA mapping.
|
|
|
|
*/
|
2021-06-04 13:53:48 +08:00
|
|
|
i = virtqueue_add_desc_split(_vq, desc, i, addr,
|
|
|
|
sg->length,
|
|
|
|
VRING_DESC_F_NEXT |
|
2021-06-04 13:53:50 +08:00
|
|
|
VRING_DESC_F_WRITE,
|
|
|
|
indirect);
|
2013-03-20 13:07:09 +08:00
|
|
|
}
|
2007-10-22 09:03:40 +08:00
|
|
|
}
|
|
|
|
/* Last one doesn't continue. */
|
2014-10-22 20:42:09 +08:00
|
|
|
desc[prev].flags &= cpu_to_virtio16(_vq->vdev, ~VRING_DESC_F_NEXT);
|
2023-08-10 20:30:51 +08:00
|
|
|
if (!indirect && vq->do_unmap)
|
2021-10-26 21:31:00 +08:00
|
|
|
vq->split.desc_extra[prev & (vq->split.vring.num - 1)].flags &=
|
2021-06-04 13:53:50 +08:00
|
|
|
~VRING_DESC_F_NEXT;
|
2007-10-22 09:03:40 +08:00
|
|
|
|
2016-02-03 13:46:36 +08:00
|
|
|
if (indirect) {
|
|
|
|
/* Now that the indirect table is filled in, map it. */
|
|
|
|
dma_addr_t addr = vring_map_single(
|
|
|
|
vq, desc, total_sg * sizeof(struct vring_desc),
|
|
|
|
DMA_TO_DEVICE);
|
2023-08-10 20:30:49 +08:00
|
|
|
if (vring_mapping_error(vq, addr)) {
|
|
|
|
if (vq->premapped)
|
|
|
|
goto free_indirect;
|
|
|
|
|
2016-02-03 13:46:36 +08:00
|
|
|
goto unmap_release;
|
2023-08-10 20:30:49 +08:00
|
|
|
}
|
2016-02-03 13:46:36 +08:00
|
|
|
|
2021-06-04 13:53:48 +08:00
|
|
|
virtqueue_add_desc_split(_vq, vq->split.vring.desc,
|
|
|
|
head, addr,
|
|
|
|
total_sg * sizeof(struct vring_desc),
|
2021-06-04 13:53:50 +08:00
|
|
|
VRING_DESC_F_INDIRECT,
|
|
|
|
false);
|
2016-02-03 13:46:36 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/* We're using some buffers from the free list. */
|
|
|
|
vq->vq.num_free -= descs_used;
|
|
|
|
|
2007-10-22 09:03:40 +08:00
|
|
|
/* Update free pointer */
|
2014-09-11 08:47:38 +08:00
|
|
|
if (indirect)
|
2021-06-04 13:53:50 +08:00
|
|
|
vq->free_head = vq->split.desc_extra[head].next;
|
2014-09-11 08:47:38 +08:00
|
|
|
else
|
|
|
|
vq->free_head = i;
|
2007-10-22 09:03:40 +08:00
|
|
|
|
2016-02-03 13:46:36 +08:00
|
|
|
/* Store token and indirect buffer state. */
|
2018-11-21 18:03:24 +08:00
|
|
|
vq->split.desc_state[head].data = data;
|
2016-02-03 13:46:36 +08:00
|
|
|
if (indirect)
|
2018-11-21 18:03:24 +08:00
|
|
|
vq->split.desc_state[head].indir_desc = desc;
|
2017-07-19 16:54:45 +08:00
|
|
|
else
|
2018-11-21 18:03:24 +08:00
|
|
|
vq->split.desc_state[head].indir_desc = ctx;
|
2007-10-22 09:03:40 +08:00
|
|
|
|
|
|
|
/* Put entry in available array (but don't update avail->idx until they
|
2012-01-12 13:14:43 +08:00
|
|
|
* do sync). */
|
2018-11-21 18:03:21 +08:00
|
|
|
avail = vq->split.avail_idx_shadow & (vq->split.vring.num - 1);
|
|
|
|
vq->split.vring.avail->ring[avail] = cpu_to_virtio16(_vq->vdev, head);
|
2007-10-22 09:03:40 +08:00
|
|
|
|
2012-01-12 13:14:43 +08:00
|
|
|
/* Descriptors and available array need to be set before we expose the
|
|
|
|
* new available array entries. */
|
2013-03-18 10:52:19 +08:00
|
|
|
virtio_wmb(vq->weak_barriers);
|
2018-11-21 18:03:21 +08:00
|
|
|
vq->split.avail_idx_shadow++;
|
|
|
|
vq->split.vring.avail->idx = cpu_to_virtio16(_vq->vdev,
|
|
|
|
vq->split.avail_idx_shadow);
|
2012-01-12 13:14:43 +08:00
|
|
|
vq->num_added++;
|
|
|
|
|
2015-02-11 12:31:13 +08:00
|
|
|
pr_debug("Added buffer head %i to %p\n", head, vq);
|
|
|
|
END_USE(vq);
|
|
|
|
|
2012-01-12 13:14:43 +08:00
|
|
|
/* This is very unlikely, but theoretically possible. Kick
|
|
|
|
* just in case. */
|
|
|
|
if (unlikely(vq->num_added == (1 << 16) - 1))
|
|
|
|
virtqueue_kick(_vq);
|
|
|
|
|
2012-10-16 21:26:15 +08:00
|
|
|
return 0;
|
2016-02-03 13:46:36 +08:00
|
|
|
|
|
|
|
unmap_release:
|
|
|
|
err_idx = i;
|
2019-09-06 22:59:01 +08:00
|
|
|
|
|
|
|
if (indirect)
|
|
|
|
i = 0;
|
|
|
|
else
|
|
|
|
i = head;
|
2016-02-03 13:46:36 +08:00
|
|
|
|
|
|
|
for (n = 0; n < total_sg; n++) {
|
|
|
|
if (i == err_idx)
|
|
|
|
break;
|
2021-06-04 13:53:50 +08:00
|
|
|
if (indirect) {
|
|
|
|
vring_unmap_one_split_indirect(vq, &desc[i]);
|
|
|
|
i = virtio16_to_cpu(_vq->vdev, desc[i].next);
|
|
|
|
} else
|
|
|
|
i = vring_unmap_one_split(vq, i);
|
2016-02-03 13:46:36 +08:00
|
|
|
}
|
|
|
|
|
2023-08-10 20:30:49 +08:00
|
|
|
free_indirect:
|
2016-02-03 13:46:36 +08:00
|
|
|
if (indirect)
|
|
|
|
kfree(desc);
|
|
|
|
|
2016-08-03 12:18:51 +08:00
|
|
|
END_USE(vq);
|
2019-11-14 20:46:46 +08:00
|
|
|
return -ENOMEM;
|
2007-10-22 09:03:40 +08:00
|
|
|
}
|
2013-03-20 13:07:09 +08:00
|
|
|
|
2018-11-21 18:03:19 +08:00
|
|
|
static bool virtqueue_kick_prepare_split(struct virtqueue *_vq)
|
2007-10-22 09:03:40 +08:00
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
2011-05-20 07:10:44 +08:00
|
|
|
u16 new, old;
|
2012-01-12 13:14:43 +08:00
|
|
|
bool needs_kick;
|
|
|
|
|
2007-10-22 09:03:40 +08:00
|
|
|
START_USE(vq);
|
2012-01-20 16:17:08 +08:00
|
|
|
/* We need to expose available array entries before checking avail
|
|
|
|
* event. */
|
2013-03-18 10:52:19 +08:00
|
|
|
virtio_mb(vq->weak_barriers);
|
2007-10-22 09:03:40 +08:00
|
|
|
|
2018-11-21 18:03:21 +08:00
|
|
|
old = vq->split.avail_idx_shadow - vq->num_added;
|
|
|
|
new = vq->split.avail_idx_shadow;
|
2007-10-22 09:03:40 +08:00
|
|
|
vq->num_added = 0;
|
|
|
|
|
2018-11-21 18:03:22 +08:00
|
|
|
LAST_ADD_TIME_CHECK(vq);
|
|
|
|
LAST_ADD_TIME_INVALID(vq);
|
2012-01-12 13:14:43 +08:00
|
|
|
|
2012-01-12 13:14:43 +08:00
|
|
|
if (vq->event) {
|
2018-11-21 18:03:21 +08:00
|
|
|
needs_kick = vring_need_event(virtio16_to_cpu(_vq->vdev,
|
|
|
|
vring_avail_event(&vq->split.vring)),
|
2012-01-12 13:14:43 +08:00
|
|
|
new, old);
|
|
|
|
} else {
|
2018-11-21 18:03:21 +08:00
|
|
|
needs_kick = !(vq->split.vring.used->flags &
|
|
|
|
cpu_to_virtio16(_vq->vdev,
|
|
|
|
VRING_USED_F_NO_NOTIFY));
|
2012-01-12 13:14:43 +08:00
|
|
|
}
|
2007-10-22 09:03:40 +08:00
|
|
|
END_USE(vq);
|
2012-01-12 13:14:43 +08:00
|
|
|
return needs_kick;
|
|
|
|
}
|
2018-11-21 18:03:19 +08:00
|
|
|
|
|
|
|
static void detach_buf_split(struct vring_virtqueue *vq, unsigned int head,
|
|
|
|
void **ctx)
|
2007-10-22 09:03:40 +08:00
|
|
|
{
|
2016-02-03 13:46:36 +08:00
|
|
|
unsigned int i, j;
|
2016-11-22 13:51:50 +08:00
|
|
|
__virtio16 nextflag = cpu_to_virtio16(vq->vq.vdev, VRING_DESC_F_NEXT);
|
2007-10-22 09:03:40 +08:00
|
|
|
|
|
|
|
/* Clear data ptr. */
|
2018-11-21 18:03:24 +08:00
|
|
|
vq->split.desc_state[head].data = NULL;
|
2007-10-22 09:03:40 +08:00
|
|
|
|
2016-02-03 13:46:36 +08:00
|
|
|
/* Put back on free list: unmap first-level descriptors and find end */
|
2007-10-22 09:03:40 +08:00
|
|
|
i = head;
|
2009-05-12 01:11:45 +08:00
|
|
|
|
2018-11-21 18:03:21 +08:00
|
|
|
while (vq->split.vring.desc[i].flags & nextflag) {
|
2021-06-04 13:53:50 +08:00
|
|
|
vring_unmap_one_split(vq, i);
|
|
|
|
i = vq->split.desc_extra[i].next;
|
2012-10-16 21:26:14 +08:00
|
|
|
vq->vq.num_free++;
|
2007-10-22 09:03:40 +08:00
|
|
|
}
|
|
|
|
|
2021-06-04 13:53:50 +08:00
|
|
|
vring_unmap_one_split(vq, i);
|
|
|
|
vq->split.desc_extra[i].next = vq->free_head;
|
2007-10-22 09:03:40 +08:00
|
|
|
vq->free_head = head;
|
2016-02-03 13:46:36 +08:00
|
|
|
|
2007-10-22 09:03:40 +08:00
|
|
|
/* Plus final descriptor */
|
2012-10-16 21:26:14 +08:00
|
|
|
vq->vq.num_free++;
|
2016-02-03 13:46:36 +08:00
|
|
|
|
2017-02-07 12:15:13 +08:00
|
|
|
if (vq->indirect) {
|
2018-11-21 18:03:24 +08:00
|
|
|
struct vring_desc *indir_desc =
|
|
|
|
vq->split.desc_state[head].indir_desc;
|
2017-02-07 12:15:13 +08:00
|
|
|
u32 len;
|
|
|
|
|
|
|
|
/* Free the indirect table, if any, now that it's unmapped. */
|
|
|
|
if (!indir_desc)
|
|
|
|
return;
|
|
|
|
|
2021-06-04 13:53:50 +08:00
|
|
|
len = vq->split.desc_extra[head].len;
|
2016-02-03 13:46:36 +08:00
|
|
|
|
2021-06-04 13:53:50 +08:00
|
|
|
BUG_ON(!(vq->split.desc_extra[head].flags &
|
|
|
|
VRING_DESC_F_INDIRECT));
|
2016-02-03 13:46:36 +08:00
|
|
|
BUG_ON(len == 0 || len % sizeof(struct vring_desc));
|
|
|
|
|
2023-08-10 20:30:51 +08:00
|
|
|
if (vq->do_unmap) {
|
2023-08-10 20:30:46 +08:00
|
|
|
for (j = 0; j < len / sizeof(struct vring_desc); j++)
|
|
|
|
vring_unmap_one_split_indirect(vq, &indir_desc[j]);
|
|
|
|
}
|
2016-02-03 13:46:36 +08:00
|
|
|
|
2017-02-07 12:15:13 +08:00
|
|
|
kfree(indir_desc);
|
2018-11-21 18:03:24 +08:00
|
|
|
vq->split.desc_state[head].indir_desc = NULL;
|
2017-02-07 12:15:13 +08:00
|
|
|
} else if (ctx) {
|
2018-11-21 18:03:24 +08:00
|
|
|
*ctx = vq->split.desc_state[head].indir_desc;
|
2016-02-03 13:46:36 +08:00
|
|
|
}
|
2007-10-22 09:03:40 +08:00
|
|
|
}
|
|
|
|
|
2023-03-10 13:34:27 +08:00
|
|
|
static bool more_used_split(const struct vring_virtqueue *vq)
|
2007-10-22 09:03:40 +08:00
|
|
|
{
|
2018-11-21 18:03:21 +08:00
|
|
|
return vq->last_used_idx != virtio16_to_cpu(vq->vq.vdev,
|
|
|
|
vq->split.vring.used->idx);
|
2007-10-22 09:03:40 +08:00
|
|
|
}
|
|
|
|
|
2018-11-21 18:03:19 +08:00
|
|
|
static void *virtqueue_get_buf_ctx_split(struct virtqueue *_vq,
|
|
|
|
unsigned int *len,
|
|
|
|
void **ctx)
|
2007-10-22 09:03:40 +08:00
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
void *ret;
|
|
|
|
unsigned int i;
|
2012-01-12 13:14:43 +08:00
|
|
|
u16 last_used;
|
2007-10-22 09:03:40 +08:00
|
|
|
|
|
|
|
START_USE(vq);
|
|
|
|
|
2008-05-03 10:50:43 +08:00
|
|
|
if (unlikely(vq->broken)) {
|
|
|
|
END_USE(vq);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2018-11-21 18:03:19 +08:00
|
|
|
if (!more_used_split(vq)) {
|
2007-10-22 09:03:40 +08:00
|
|
|
pr_debug("No more buffers in queue\n");
|
|
|
|
END_USE(vq);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2009-10-25 21:28:53 +08:00
|
|
|
/* Only get used array entries after they have been exposed by host. */
|
2013-03-18 10:52:19 +08:00
|
|
|
virtio_rmb(vq->weak_barriers);
|
2009-10-25 21:28:53 +08:00
|
|
|
|
2018-11-21 18:03:21 +08:00
|
|
|
last_used = (vq->last_used_idx & (vq->split.vring.num - 1));
|
|
|
|
i = virtio32_to_cpu(_vq->vdev,
|
|
|
|
vq->split.vring.used->ring[last_used].id);
|
|
|
|
*len = virtio32_to_cpu(_vq->vdev,
|
|
|
|
vq->split.vring.used->ring[last_used].len);
|
2007-10-22 09:03:40 +08:00
|
|
|
|
2018-11-21 18:03:21 +08:00
|
|
|
if (unlikely(i >= vq->split.vring.num)) {
|
2007-10-22 09:03:40 +08:00
|
|
|
BAD_RING(vq, "id %u out of range\n", i);
|
|
|
|
return NULL;
|
|
|
|
}
|
2018-11-21 18:03:24 +08:00
|
|
|
if (unlikely(!vq->split.desc_state[i].data)) {
|
2007-10-22 09:03:40 +08:00
|
|
|
BAD_RING(vq, "id %u is not a head!\n", i);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2018-11-21 18:03:19 +08:00
|
|
|
/* detach_buf_split clears data, so grab it now. */
|
2018-11-21 18:03:24 +08:00
|
|
|
ret = vq->split.desc_state[i].data;
|
2018-11-21 18:03:19 +08:00
|
|
|
detach_buf_split(vq, i, ctx);
|
2007-10-22 09:03:40 +08:00
|
|
|
vq->last_used_idx++;
|
2011-05-20 07:10:44 +08:00
|
|
|
/* If we expect an interrupt for the next entry, tell host
|
|
|
|
* by writing event index and flush out the write before
|
|
|
|
* the read in the next get_buf call. */
|
2018-11-21 18:03:21 +08:00
|
|
|
if (!(vq->split.avail_flags_shadow & VRING_AVAIL_F_NO_INTERRUPT))
|
2015-12-17 18:20:39 +08:00
|
|
|
virtio_store_mb(vq->weak_barriers,
|
2018-11-21 18:03:21 +08:00
|
|
|
&vring_used_event(&vq->split.vring),
|
2015-12-17 18:20:39 +08:00
|
|
|
cpu_to_virtio16(_vq->vdev, vq->last_used_idx));
|
2011-05-20 07:10:44 +08:00
|
|
|
|
2018-11-21 18:03:22 +08:00
|
|
|
LAST_ADD_TIME_INVALID(vq);
|
2012-01-12 13:14:43 +08:00
|
|
|
|
2007-10-22 09:03:40 +08:00
|
|
|
END_USE(vq);
|
|
|
|
return ret;
|
|
|
|
}
|
2018-11-21 18:03:19 +08:00
|
|
|
|
|
|
|
static void virtqueue_disable_cb_split(struct virtqueue *_vq)
|
2008-02-05 12:49:57 +08:00
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
|
2018-11-21 18:03:21 +08:00
|
|
|
if (!(vq->split.avail_flags_shadow & VRING_AVAIL_F_NO_INTERRUPT)) {
|
|
|
|
vq->split.avail_flags_shadow |= VRING_AVAIL_F_NO_INTERRUPT;
|
2023-03-29 18:23:00 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* If device triggered an event already it won't trigger one again:
|
|
|
|
* no need to disable.
|
|
|
|
*/
|
|
|
|
if (vq->event_triggered)
|
|
|
|
return;
|
|
|
|
|
2021-04-13 13:19:16 +08:00
|
|
|
if (vq->event)
|
|
|
|
/* TODO: this is a hack. Figure out a cleaner value to write. */
|
|
|
|
vring_used_event(&vq->split.vring) = 0x0;
|
|
|
|
else
|
2018-11-21 18:03:21 +08:00
|
|
|
vq->split.vring.avail->flags =
|
|
|
|
cpu_to_virtio16(_vq->vdev,
|
|
|
|
vq->split.avail_flags_shadow);
|
virtio_ring: shadow available ring flags & index
Improves cacheline transfer flow of available ring header.
Virtqueues are implemented as a pair of rings, one producer->consumer
avail ring and one consumer->producer used ring; preceding the
avail ring in memory are two contiguous u16 fields -- avail->flags
and avail->idx. A producer posts work by writing to avail->idx and
a consumer reads avail->idx.
The flags and idx fields only need to be written by a producer CPU
and only read by a consumer CPU; when the producer and consumer are
running on different CPUs and the virtio_ring code is structured to
only have source writes/sink reads, we can continuously transfer the
avail header cacheline between 'M' states between cores. This flow
optimizes core -> core bandwidth on certain CPUs.
(see: "Software Optimization Guide for AMD Family 15h Processors",
Section 11.6; similar language appears in the 10h guide and should
apply to CPUs w/ exclusive caches, using LLC as a transfer cache)
Unfortunately the existing virtio_ring code issued reads to the
avail->idx and read-modify-writes to avail->flags on the producer.
This change shadows the flags and index fields in producer memory;
the vring code now reads from the shadows and only ever writes to
avail->flags and avail->idx, allowing the cacheline to transfer
core -> core optimally.
In a concurrent version of vring_bench, the time required for
10,000,000 buffer checkout/returns was reduced by ~2% (average
across many runs) on an AMD Piledriver (15h) CPU:
(w/o shadowing):
Performance counter stats for './vring_bench':
5,451,082,016 L1-dcache-loads
...
2.221477739 seconds time elapsed
(w/ shadowing):
Performance counter stats for './vring_bench':
5,405,701,361 L1-dcache-loads
...
2.168405376 seconds time elapsed
The further away (in a NUMA sense) virtio producers and consumers are
from each other, the more we expect to benefit. Physical implementations
of virtio devices and implementations of virtio where the consumer polls
vring avail indexes (vhost) should also benefit.
Signed-off-by: Venkatesh Srinivas <venkateshs@google.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2015-11-11 08:21:07 +08:00
|
|
|
}
|
2008-02-05 12:49:57 +08:00
|
|
|
}
|
|
|
|
|
2022-04-18 10:54:35 +08:00
|
|
|
static unsigned int virtqueue_enable_cb_prepare_split(struct virtqueue *_vq)
|
2007-10-22 09:03:40 +08:00
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
2013-07-09 18:19:18 +08:00
|
|
|
u16 last_used_idx;
|
2007-10-22 09:03:40 +08:00
|
|
|
|
|
|
|
START_USE(vq);
|
|
|
|
|
|
|
|
/* We optimistically turn back on interrupts, then check if there was
|
|
|
|
* more to do. */
|
2011-05-20 07:10:44 +08:00
|
|
|
/* Depending on the VIRTIO_RING_F_EVENT_IDX feature, we need to
|
|
|
|
* either clear the flags bit or point the event index at the next
|
|
|
|
* entry. Always do both to keep code simple. */
|
2018-11-21 18:03:21 +08:00
|
|
|
if (vq->split.avail_flags_shadow & VRING_AVAIL_F_NO_INTERRUPT) {
|
|
|
|
vq->split.avail_flags_shadow &= ~VRING_AVAIL_F_NO_INTERRUPT;
|
2016-08-31 20:00:04 +08:00
|
|
|
if (!vq->event)
|
2018-11-21 18:03:21 +08:00
|
|
|
vq->split.vring.avail->flags =
|
|
|
|
cpu_to_virtio16(_vq->vdev,
|
|
|
|
vq->split.avail_flags_shadow);
|
virtio_ring: shadow available ring flags & index
Improves cacheline transfer flow of available ring header.
Virtqueues are implemented as a pair of rings, one producer->consumer
avail ring and one consumer->producer used ring; preceding the
avail ring in memory are two contiguous u16 fields -- avail->flags
and avail->idx. A producer posts work by writing to avail->idx and
a consumer reads avail->idx.
The flags and idx fields only need to be written by a producer CPU
and only read by a consumer CPU; when the producer and consumer are
running on different CPUs and the virtio_ring code is structured to
only have source writes/sink reads, we can continuously transfer the
avail header cacheline between 'M' states between cores. This flow
optimizes core -> core bandwidth on certain CPUs.
(see: "Software Optimization Guide for AMD Family 15h Processors",
Section 11.6; similar language appears in the 10h guide and should
apply to CPUs w/ exclusive caches, using LLC as a transfer cache)
Unfortunately the existing virtio_ring code issued reads to the
avail->idx and read-modify-writes to avail->flags on the producer.
This change shadows the flags and index fields in producer memory;
the vring code now reads from the shadows and only ever writes to
avail->flags and avail->idx, allowing the cacheline to transfer
core -> core optimally.
In a concurrent version of vring_bench, the time required for
10,000,000 buffer checkout/returns was reduced by ~2% (average
across many runs) on an AMD Piledriver (15h) CPU:
(w/o shadowing):
Performance counter stats for './vring_bench':
5,451,082,016 L1-dcache-loads
...
2.221477739 seconds time elapsed
(w/ shadowing):
Performance counter stats for './vring_bench':
5,405,701,361 L1-dcache-loads
...
2.168405376 seconds time elapsed
The further away (in a NUMA sense) virtio producers and consumers are
from each other, the more we expect to benefit. Physical implementations
of virtio devices and implementations of virtio where the consumer polls
vring avail indexes (vhost) should also benefit.
Signed-off-by: Venkatesh Srinivas <venkateshs@google.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2015-11-11 08:21:07 +08:00
|
|
|
}
|
2018-11-21 18:03:21 +08:00
|
|
|
vring_used_event(&vq->split.vring) = cpu_to_virtio16(_vq->vdev,
|
|
|
|
last_used_idx = vq->last_used_idx);
|
2013-07-09 18:19:18 +08:00
|
|
|
END_USE(vq);
|
|
|
|
return last_used_idx;
|
|
|
|
}
|
2018-11-21 18:03:19 +08:00
|
|
|
|
2022-04-18 10:54:35 +08:00
|
|
|
static bool virtqueue_poll_split(struct virtqueue *_vq, unsigned int last_used_idx)
|
2018-11-21 18:03:19 +08:00
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
|
|
|
|
return (u16)last_used_idx != virtio16_to_cpu(_vq->vdev,
|
2018-11-21 18:03:21 +08:00
|
|
|
vq->split.vring.used->idx);
|
2018-11-21 18:03:19 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static bool virtqueue_enable_cb_delayed_split(struct virtqueue *_vq)
|
2011-05-20 07:11:14 +08:00
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
u16 bufs;
|
|
|
|
|
|
|
|
START_USE(vq);
|
|
|
|
|
|
|
|
/* We optimistically turn back on interrupts, then check if there was
|
|
|
|
* more to do. */
|
|
|
|
/* Depending on the VIRTIO_RING_F_USED_EVENT_IDX feature, we need to
|
|
|
|
* either clear the flags bit or point the event index at the next
|
2016-08-31 20:00:04 +08:00
|
|
|
* entry. Always update the event index to keep code simple. */
|
2018-11-21 18:03:21 +08:00
|
|
|
if (vq->split.avail_flags_shadow & VRING_AVAIL_F_NO_INTERRUPT) {
|
|
|
|
vq->split.avail_flags_shadow &= ~VRING_AVAIL_F_NO_INTERRUPT;
|
2016-08-31 20:00:04 +08:00
|
|
|
if (!vq->event)
|
2018-11-21 18:03:21 +08:00
|
|
|
vq->split.vring.avail->flags =
|
|
|
|
cpu_to_virtio16(_vq->vdev,
|
|
|
|
vq->split.avail_flags_shadow);
|
virtio_ring: shadow available ring flags & index
Improves cacheline transfer flow of available ring header.
Virtqueues are implemented as a pair of rings, one producer->consumer
avail ring and one consumer->producer used ring; preceding the
avail ring in memory are two contiguous u16 fields -- avail->flags
and avail->idx. A producer posts work by writing to avail->idx and
a consumer reads avail->idx.
The flags and idx fields only need to be written by a producer CPU
and only read by a consumer CPU; when the producer and consumer are
running on different CPUs and the virtio_ring code is structured to
only have source writes/sink reads, we can continuously transfer the
avail header cacheline between 'M' states between cores. This flow
optimizes core -> core bandwidth on certain CPUs.
(see: "Software Optimization Guide for AMD Family 15h Processors",
Section 11.6; similar language appears in the 10h guide and should
apply to CPUs w/ exclusive caches, using LLC as a transfer cache)
Unfortunately the existing virtio_ring code issued reads to the
avail->idx and read-modify-writes to avail->flags on the producer.
This change shadows the flags and index fields in producer memory;
the vring code now reads from the shadows and only ever writes to
avail->flags and avail->idx, allowing the cacheline to transfer
core -> core optimally.
In a concurrent version of vring_bench, the time required for
10,000,000 buffer checkout/returns was reduced by ~2% (average
across many runs) on an AMD Piledriver (15h) CPU:
(w/o shadowing):
Performance counter stats for './vring_bench':
5,451,082,016 L1-dcache-loads
...
2.221477739 seconds time elapsed
(w/ shadowing):
Performance counter stats for './vring_bench':
5,405,701,361 L1-dcache-loads
...
2.168405376 seconds time elapsed
The further away (in a NUMA sense) virtio producers and consumers are
from each other, the more we expect to benefit. Physical implementations
of virtio devices and implementations of virtio where the consumer polls
vring avail indexes (vhost) should also benefit.
Signed-off-by: Venkatesh Srinivas <venkateshs@google.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2015-11-11 08:21:07 +08:00
|
|
|
}
|
2011-05-20 07:11:14 +08:00
|
|
|
/* TODO: tune this threshold */
|
2018-11-21 18:03:21 +08:00
|
|
|
bufs = (u16)(vq->split.avail_idx_shadow - vq->last_used_idx) * 3 / 4;
|
2015-12-17 18:20:39 +08:00
|
|
|
|
|
|
|
virtio_store_mb(vq->weak_barriers,
|
2018-11-21 18:03:21 +08:00
|
|
|
&vring_used_event(&vq->split.vring),
|
2015-12-17 18:20:39 +08:00
|
|
|
cpu_to_virtio16(_vq->vdev, vq->last_used_idx + bufs));
|
|
|
|
|
2018-11-21 18:03:21 +08:00
|
|
|
if (unlikely((u16)(virtio16_to_cpu(_vq->vdev, vq->split.vring.used->idx)
|
|
|
|
- vq->last_used_idx) > bufs)) {
|
2011-05-20 07:11:14 +08:00
|
|
|
END_USE(vq);
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
|
|
|
END_USE(vq);
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
2018-11-21 18:03:19 +08:00
|
|
|
static void *virtqueue_detach_unused_buf_split(struct virtqueue *_vq)
|
2010-01-18 21:45:23 +08:00
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
unsigned int i;
|
|
|
|
void *buf;
|
|
|
|
|
|
|
|
START_USE(vq);
|
|
|
|
|
2018-11-21 18:03:21 +08:00
|
|
|
for (i = 0; i < vq->split.vring.num; i++) {
|
2018-11-21 18:03:24 +08:00
|
|
|
if (!vq->split.desc_state[i].data)
|
2010-01-18 21:45:23 +08:00
|
|
|
continue;
|
2018-11-21 18:03:19 +08:00
|
|
|
/* detach_buf_split clears data, so grab it now. */
|
2018-11-21 18:03:24 +08:00
|
|
|
buf = vq->split.desc_state[i].data;
|
2018-11-21 18:03:19 +08:00
|
|
|
detach_buf_split(vq, i, NULL);
|
2018-11-21 18:03:21 +08:00
|
|
|
vq->split.avail_idx_shadow--;
|
|
|
|
vq->split.vring.avail->idx = cpu_to_virtio16(_vq->vdev,
|
|
|
|
vq->split.avail_idx_shadow);
|
2010-01-18 21:45:23 +08:00
|
|
|
END_USE(vq);
|
|
|
|
return buf;
|
|
|
|
}
|
|
|
|
/* That should have freed everything. */
|
2018-11-21 18:03:21 +08:00
|
|
|
BUG_ON(vq->vq.num_free != vq->split.vring.num);
|
2010-01-18 21:45:23 +08:00
|
|
|
|
|
|
|
END_USE(vq);
|
|
|
|
return NULL;
|
|
|
|
}
|
2018-11-21 18:03:19 +08:00
|
|
|
|
2022-08-01 14:38:32 +08:00
|
|
|
static void virtqueue_vring_init_split(struct vring_virtqueue_split *vring_split,
|
|
|
|
struct vring_virtqueue *vq)
|
|
|
|
{
|
|
|
|
struct virtio_device *vdev;
|
|
|
|
|
|
|
|
vdev = vq->vq.vdev;
|
|
|
|
|
|
|
|
vring_split->avail_flags_shadow = 0;
|
|
|
|
vring_split->avail_idx_shadow = 0;
|
|
|
|
|
|
|
|
/* No callback? Tell other side not to bother us. */
|
|
|
|
if (!vq->vq.callback) {
|
|
|
|
vring_split->avail_flags_shadow |= VRING_AVAIL_F_NO_INTERRUPT;
|
|
|
|
if (!vq->event)
|
|
|
|
vring_split->vring.avail->flags = cpu_to_virtio16(vdev,
|
|
|
|
vring_split->avail_flags_shadow);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2022-08-01 14:38:34 +08:00
|
|
|
static void virtqueue_reinit_split(struct vring_virtqueue *vq)
|
|
|
|
{
|
|
|
|
int num;
|
|
|
|
|
|
|
|
num = vq->split.vring.num;
|
|
|
|
|
|
|
|
vq->split.vring.avail->flags = 0;
|
|
|
|
vq->split.vring.avail->idx = 0;
|
|
|
|
|
|
|
|
/* reset avail event */
|
|
|
|
vq->split.vring.avail->ring[num] = 0;
|
|
|
|
|
|
|
|
vq->split.vring.used->flags = 0;
|
|
|
|
vq->split.vring.used->idx = 0;
|
|
|
|
|
|
|
|
/* reset used event */
|
|
|
|
*(__virtio16 *)&(vq->split.vring.used->ring[num]) = 0;
|
|
|
|
|
|
|
|
virtqueue_init(vq, num);
|
|
|
|
|
|
|
|
virtqueue_vring_init_split(&vq->split, vq);
|
|
|
|
}
|
|
|
|
|
2022-08-01 14:38:33 +08:00
|
|
|
static void virtqueue_vring_attach_split(struct vring_virtqueue *vq,
|
|
|
|
struct vring_virtqueue_split *vring_split)
|
|
|
|
{
|
|
|
|
vq->split = *vring_split;
|
|
|
|
|
|
|
|
/* Put everything in free lists. */
|
|
|
|
vq->free_head = 0;
|
|
|
|
}
|
|
|
|
|
2022-08-01 14:38:31 +08:00
|
|
|
static int vring_alloc_state_extra_split(struct vring_virtqueue_split *vring_split)
|
|
|
|
{
|
|
|
|
struct vring_desc_state_split *state;
|
|
|
|
struct vring_desc_extra *extra;
|
|
|
|
u32 num = vring_split->vring.num;
|
|
|
|
|
|
|
|
state = kmalloc_array(num, sizeof(struct vring_desc_state_split), GFP_KERNEL);
|
|
|
|
if (!state)
|
|
|
|
goto err_state;
|
|
|
|
|
|
|
|
extra = vring_alloc_desc_extra(num);
|
|
|
|
if (!extra)
|
|
|
|
goto err_extra;
|
|
|
|
|
|
|
|
memset(state, 0, num * sizeof(struct vring_desc_state_split));
|
|
|
|
|
|
|
|
vring_split->desc_state = state;
|
|
|
|
vring_split->desc_extra = extra;
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
err_extra:
|
|
|
|
kfree(state);
|
|
|
|
err_state:
|
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
|
2022-08-01 14:38:29 +08:00
|
|
|
static void vring_free_split(struct vring_virtqueue_split *vring_split,
|
2023-01-19 14:15:21 +08:00
|
|
|
struct virtio_device *vdev, struct device *dma_dev)
|
2022-08-01 14:38:29 +08:00
|
|
|
{
|
|
|
|
vring_free_queue(vdev, vring_split->queue_size_in_bytes,
|
|
|
|
vring_split->vring.desc,
|
2023-01-19 14:15:21 +08:00
|
|
|
vring_split->queue_dma_addr,
|
|
|
|
dma_dev);
|
2022-08-01 14:38:29 +08:00
|
|
|
|
|
|
|
kfree(vring_split->desc_state);
|
|
|
|
kfree(vring_split->desc_extra);
|
|
|
|
}
|
|
|
|
|
2022-08-01 14:38:30 +08:00
|
|
|
static int vring_alloc_queue_split(struct vring_virtqueue_split *vring_split,
|
|
|
|
struct virtio_device *vdev,
|
|
|
|
u32 num,
|
|
|
|
unsigned int vring_align,
|
2023-01-19 14:15:21 +08:00
|
|
|
bool may_reduce_num,
|
|
|
|
struct device *dma_dev)
|
2018-11-21 18:03:25 +08:00
|
|
|
{
|
|
|
|
void *queue = NULL;
|
|
|
|
dma_addr_t dma_addr;
|
|
|
|
|
|
|
|
/* We assume num is a power of 2. */
|
2022-10-21 14:27:34 +08:00
|
|
|
if (!is_power_of_2(num)) {
|
2018-11-21 18:03:25 +08:00
|
|
|
dev_warn(&vdev->dev, "Bad virtqueue length %u\n", num);
|
2022-08-01 14:38:30 +08:00
|
|
|
return -EINVAL;
|
2018-11-21 18:03:25 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/* TODO: allocate each queue chunk individually */
|
|
|
|
for (; num && vring_size(num, vring_align) > PAGE_SIZE; num /= 2) {
|
|
|
|
queue = vring_alloc_queue(vdev, vring_size(num, vring_align),
|
|
|
|
&dma_addr,
|
2023-01-19 14:15:21 +08:00
|
|
|
GFP_KERNEL | __GFP_NOWARN | __GFP_ZERO,
|
|
|
|
dma_dev);
|
2018-11-21 18:03:25 +08:00
|
|
|
if (queue)
|
|
|
|
break;
|
2019-04-08 20:33:22 +08:00
|
|
|
if (!may_reduce_num)
|
2022-08-01 14:38:30 +08:00
|
|
|
return -ENOMEM;
|
2018-11-21 18:03:25 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
if (!num)
|
2022-08-01 14:38:30 +08:00
|
|
|
return -ENOMEM;
|
2018-11-21 18:03:25 +08:00
|
|
|
|
|
|
|
if (!queue) {
|
|
|
|
/* Try to get a single page. You are my only hope! */
|
|
|
|
queue = vring_alloc_queue(vdev, vring_size(num, vring_align),
|
2023-01-19 14:15:21 +08:00
|
|
|
&dma_addr, GFP_KERNEL | __GFP_ZERO,
|
|
|
|
dma_dev);
|
2018-11-21 18:03:25 +08:00
|
|
|
}
|
|
|
|
if (!queue)
|
2022-08-01 14:38:30 +08:00
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
vring_init(&vring_split->vring, num, queue, vring_align);
|
|
|
|
|
|
|
|
vring_split->queue_dma_addr = dma_addr;
|
|
|
|
vring_split->queue_size_in_bytes = vring_size(num, vring_align);
|
2018-11-21 18:03:25 +08:00
|
|
|
|
2022-08-01 14:38:35 +08:00
|
|
|
vring_split->vring_align = vring_align;
|
|
|
|
vring_split->may_reduce_num = may_reduce_num;
|
|
|
|
|
2022-08-01 14:38:30 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct virtqueue *vring_create_virtqueue_split(
|
|
|
|
unsigned int index,
|
|
|
|
unsigned int num,
|
|
|
|
unsigned int vring_align,
|
|
|
|
struct virtio_device *vdev,
|
|
|
|
bool weak_barriers,
|
|
|
|
bool may_reduce_num,
|
|
|
|
bool context,
|
|
|
|
bool (*notify)(struct virtqueue *),
|
|
|
|
void (*callback)(struct virtqueue *),
|
2023-01-19 14:15:21 +08:00
|
|
|
const char *name,
|
|
|
|
struct device *dma_dev)
|
2022-08-01 14:38:30 +08:00
|
|
|
{
|
|
|
|
struct vring_virtqueue_split vring_split = {};
|
|
|
|
struct virtqueue *vq;
|
|
|
|
int err;
|
|
|
|
|
|
|
|
err = vring_alloc_queue_split(&vring_split, vdev, num, vring_align,
|
2023-01-19 14:15:21 +08:00
|
|
|
may_reduce_num, dma_dev);
|
2022-08-01 14:38:30 +08:00
|
|
|
if (err)
|
|
|
|
return NULL;
|
2018-11-21 18:03:25 +08:00
|
|
|
|
2022-08-01 14:38:28 +08:00
|
|
|
vq = __vring_new_virtqueue(index, &vring_split, vdev, weak_barriers,
|
2023-01-19 14:15:21 +08:00
|
|
|
context, notify, callback, name, dma_dev);
|
2018-11-21 18:03:25 +08:00
|
|
|
if (!vq) {
|
2023-01-19 14:15:21 +08:00
|
|
|
vring_free_split(&vring_split, vdev, dma_dev);
|
2018-11-21 18:03:25 +08:00
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
to_vvq(vq)->we_own_ring = true;
|
|
|
|
|
|
|
|
return vq;
|
|
|
|
}
|
|
|
|
|
2022-08-01 14:38:36 +08:00
|
|
|
static int virtqueue_resize_split(struct virtqueue *_vq, u32 num)
|
|
|
|
{
|
|
|
|
struct vring_virtqueue_split vring_split = {};
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
struct virtio_device *vdev = _vq->vdev;
|
|
|
|
int err;
|
|
|
|
|
|
|
|
err = vring_alloc_queue_split(&vring_split, vdev, num,
|
|
|
|
vq->split.vring_align,
|
2023-01-19 14:15:21 +08:00
|
|
|
vq->split.may_reduce_num,
|
|
|
|
vring_dma_dev(vq));
|
2022-08-01 14:38:36 +08:00
|
|
|
if (err)
|
|
|
|
goto err;
|
|
|
|
|
|
|
|
err = vring_alloc_state_extra_split(&vring_split);
|
|
|
|
if (err)
|
|
|
|
goto err_state_extra;
|
|
|
|
|
|
|
|
vring_free(&vq->vq);
|
|
|
|
|
|
|
|
virtqueue_vring_init_split(&vring_split, vq);
|
|
|
|
|
|
|
|
virtqueue_init(vq, vring_split.vring.num);
|
|
|
|
virtqueue_vring_attach_split(vq, &vring_split);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
err_state_extra:
|
2023-01-19 14:15:21 +08:00
|
|
|
vring_free_split(&vring_split, vdev, vring_dma_dev(vq));
|
2022-08-01 14:38:36 +08:00
|
|
|
err:
|
|
|
|
virtqueue_reinit_split(vq);
|
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
|
2018-11-21 18:03:20 +08:00
|
|
|
|
2018-11-21 18:03:27 +08:00
|
|
|
/*
|
|
|
|
* Packed ring specific functions - *_packed().
|
|
|
|
*/
|
2023-03-10 13:34:27 +08:00
|
|
|
static bool packed_used_wrap_counter(u16 last_used_idx)
|
2022-06-17 10:04:11 +08:00
|
|
|
{
|
|
|
|
return !!(last_used_idx & (1 << VRING_PACKED_EVENT_F_WRAP_CTR));
|
|
|
|
}
|
|
|
|
|
2023-03-10 13:34:27 +08:00
|
|
|
static u16 packed_last_used(u16 last_used_idx)
|
2022-06-17 10:04:11 +08:00
|
|
|
{
|
|
|
|
return last_used_idx & ~(-(1 << VRING_PACKED_EVENT_F_WRAP_CTR));
|
|
|
|
}
|
2018-11-21 18:03:27 +08:00
|
|
|
|
2022-02-24 19:03:54 +08:00
|
|
|
static void vring_unmap_extra_packed(const struct vring_virtqueue *vq,
|
2023-03-10 13:34:28 +08:00
|
|
|
const struct vring_desc_extra *extra)
|
2018-11-21 18:03:27 +08:00
|
|
|
{
|
|
|
|
u16 flags;
|
|
|
|
|
2022-02-24 19:03:54 +08:00
|
|
|
flags = extra->flags;
|
2018-11-21 18:03:27 +08:00
|
|
|
|
|
|
|
if (flags & VRING_DESC_F_INDIRECT) {
|
2023-08-10 20:30:51 +08:00
|
|
|
if (!vq->use_dma_api)
|
|
|
|
return;
|
|
|
|
|
2018-11-21 18:03:27 +08:00
|
|
|
dma_unmap_single(vring_dma_dev(vq),
|
2022-02-24 19:03:54 +08:00
|
|
|
extra->addr, extra->len,
|
2018-11-21 18:03:27 +08:00
|
|
|
(flags & VRING_DESC_F_WRITE) ?
|
|
|
|
DMA_FROM_DEVICE : DMA_TO_DEVICE);
|
|
|
|
} else {
|
2023-08-10 20:30:51 +08:00
|
|
|
if (!vq->do_unmap)
|
|
|
|
return;
|
|
|
|
|
2018-11-21 18:03:27 +08:00
|
|
|
dma_unmap_page(vring_dma_dev(vq),
|
2022-02-24 19:03:54 +08:00
|
|
|
extra->addr, extra->len,
|
2018-11-21 18:03:27 +08:00
|
|
|
(flags & VRING_DESC_F_WRITE) ?
|
|
|
|
DMA_FROM_DEVICE : DMA_TO_DEVICE);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void vring_unmap_desc_packed(const struct vring_virtqueue *vq,
|
2023-03-10 13:34:28 +08:00
|
|
|
const struct vring_packed_desc *desc)
|
2018-11-21 18:03:27 +08:00
|
|
|
{
|
|
|
|
u16 flags;
|
|
|
|
|
2023-08-10 20:30:51 +08:00
|
|
|
if (!vq->do_unmap)
|
2018-11-21 18:03:27 +08:00
|
|
|
return;
|
|
|
|
|
|
|
|
flags = le16_to_cpu(desc->flags);
|
|
|
|
|
2022-02-24 19:03:56 +08:00
|
|
|
dma_unmap_page(vring_dma_dev(vq),
|
|
|
|
le64_to_cpu(desc->addr),
|
|
|
|
le32_to_cpu(desc->len),
|
|
|
|
(flags & VRING_DESC_F_WRITE) ?
|
|
|
|
DMA_FROM_DEVICE : DMA_TO_DEVICE);
|
2018-11-21 18:03:27 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static struct vring_packed_desc *alloc_indirect_packed(unsigned int total_sg,
|
|
|
|
gfp_t gfp)
|
|
|
|
{
|
|
|
|
struct vring_packed_desc *desc;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* We require lowmem mappings for the descriptors because
|
|
|
|
* otherwise virt_to_phys will give us bogus addresses in the
|
|
|
|
* virtqueue.
|
|
|
|
*/
|
|
|
|
gfp &= ~__GFP_HIGHMEM;
|
|
|
|
|
|
|
|
desc = kmalloc_array(total_sg, sizeof(struct vring_packed_desc), gfp);
|
|
|
|
|
|
|
|
return desc;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int virtqueue_add_indirect_packed(struct vring_virtqueue *vq,
|
2021-10-20 19:23:22 +08:00
|
|
|
struct scatterlist *sgs[],
|
|
|
|
unsigned int total_sg,
|
|
|
|
unsigned int out_sgs,
|
|
|
|
unsigned int in_sgs,
|
|
|
|
void *data,
|
|
|
|
gfp_t gfp)
|
2018-11-21 18:03:27 +08:00
|
|
|
{
|
|
|
|
struct vring_packed_desc *desc;
|
|
|
|
struct scatterlist *sg;
|
|
|
|
unsigned int i, n, err_idx;
|
|
|
|
u16 head, id;
|
|
|
|
dma_addr_t addr;
|
|
|
|
|
|
|
|
head = vq->packed.next_avail_idx;
|
|
|
|
desc = alloc_indirect_packed(total_sg, gfp);
|
2021-10-20 19:23:23 +08:00
|
|
|
if (!desc)
|
|
|
|
return -ENOMEM;
|
2018-11-21 18:03:27 +08:00
|
|
|
|
|
|
|
if (unlikely(vq->vq.num_free < 1)) {
|
|
|
|
pr_debug("Can't add buf len 1 - avail = 0\n");
|
2019-03-12 15:06:53 +08:00
|
|
|
kfree(desc);
|
2018-11-21 18:03:27 +08:00
|
|
|
END_USE(vq);
|
|
|
|
return -ENOSPC;
|
|
|
|
}
|
|
|
|
|
|
|
|
i = 0;
|
|
|
|
id = vq->free_head;
|
|
|
|
BUG_ON(id == vq->packed.vring.num);
|
|
|
|
|
|
|
|
for (n = 0; n < out_sgs + in_sgs; n++) {
|
|
|
|
for (sg = sgs[n]; sg; sg = sg_next(sg)) {
|
2023-08-10 20:30:47 +08:00
|
|
|
if (vring_map_one_sg(vq, sg, n < out_sgs ?
|
|
|
|
DMA_TO_DEVICE : DMA_FROM_DEVICE, &addr))
|
2018-11-21 18:03:27 +08:00
|
|
|
goto unmap_release;
|
|
|
|
|
|
|
|
desc[i].flags = cpu_to_le16(n < out_sgs ?
|
|
|
|
0 : VRING_DESC_F_WRITE);
|
|
|
|
desc[i].addr = cpu_to_le64(addr);
|
|
|
|
desc[i].len = cpu_to_le32(sg->length);
|
|
|
|
i++;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Now that the indirect table is filled in, map it. */
|
|
|
|
addr = vring_map_single(vq, desc,
|
|
|
|
total_sg * sizeof(struct vring_packed_desc),
|
|
|
|
DMA_TO_DEVICE);
|
2023-08-10 20:30:49 +08:00
|
|
|
if (vring_mapping_error(vq, addr)) {
|
|
|
|
if (vq->premapped)
|
|
|
|
goto free_desc;
|
|
|
|
|
2018-11-21 18:03:27 +08:00
|
|
|
goto unmap_release;
|
2023-08-10 20:30:49 +08:00
|
|
|
}
|
2018-11-21 18:03:27 +08:00
|
|
|
|
|
|
|
vq->packed.vring.desc[head].addr = cpu_to_le64(addr);
|
|
|
|
vq->packed.vring.desc[head].len = cpu_to_le32(total_sg *
|
|
|
|
sizeof(struct vring_packed_desc));
|
|
|
|
vq->packed.vring.desc[head].id = cpu_to_le16(id);
|
|
|
|
|
2024-02-23 15:18:33 +08:00
|
|
|
if (vq->use_dma_api) {
|
2018-11-21 18:03:27 +08:00
|
|
|
vq->packed.desc_extra[id].addr = addr;
|
|
|
|
vq->packed.desc_extra[id].len = total_sg *
|
|
|
|
sizeof(struct vring_packed_desc);
|
|
|
|
vq->packed.desc_extra[id].flags = VRING_DESC_F_INDIRECT |
|
|
|
|
vq->packed.avail_used_flags;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* A driver MUST NOT make the first descriptor in the list
|
|
|
|
* available before all subsequent descriptors comprising
|
|
|
|
* the list are made available.
|
|
|
|
*/
|
|
|
|
virtio_wmb(vq->weak_barriers);
|
|
|
|
vq->packed.vring.desc[head].flags = cpu_to_le16(VRING_DESC_F_INDIRECT |
|
|
|
|
vq->packed.avail_used_flags);
|
|
|
|
|
|
|
|
/* We're using some buffers from the free list. */
|
|
|
|
vq->vq.num_free -= 1;
|
|
|
|
|
|
|
|
/* Update free pointer */
|
|
|
|
n = head + 1;
|
|
|
|
if (n >= vq->packed.vring.num) {
|
|
|
|
n = 0;
|
|
|
|
vq->packed.avail_wrap_counter ^= 1;
|
|
|
|
vq->packed.avail_used_flags ^=
|
|
|
|
1 << VRING_PACKED_DESC_F_AVAIL |
|
|
|
|
1 << VRING_PACKED_DESC_F_USED;
|
|
|
|
}
|
|
|
|
vq->packed.next_avail_idx = n;
|
2021-06-04 13:53:44 +08:00
|
|
|
vq->free_head = vq->packed.desc_extra[id].next;
|
2018-11-21 18:03:27 +08:00
|
|
|
|
|
|
|
/* Store token and indirect buffer state. */
|
|
|
|
vq->packed.desc_state[id].num = 1;
|
|
|
|
vq->packed.desc_state[id].data = data;
|
|
|
|
vq->packed.desc_state[id].indir_desc = desc;
|
|
|
|
vq->packed.desc_state[id].last = id;
|
|
|
|
|
|
|
|
vq->num_added += 1;
|
|
|
|
|
|
|
|
pr_debug("Added buffer head %i to %p\n", head, vq);
|
|
|
|
END_USE(vq);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
unmap_release:
|
|
|
|
err_idx = i;
|
|
|
|
|
|
|
|
for (i = 0; i < err_idx; i++)
|
|
|
|
vring_unmap_desc_packed(vq, &desc[i]);
|
|
|
|
|
2023-08-10 20:30:49 +08:00
|
|
|
free_desc:
|
2018-11-21 18:03:27 +08:00
|
|
|
kfree(desc);
|
|
|
|
|
|
|
|
END_USE(vq);
|
2019-11-14 20:46:46 +08:00
|
|
|
return -ENOMEM;
|
2018-11-21 18:03:27 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static inline int virtqueue_add_packed(struct virtqueue *_vq,
|
|
|
|
struct scatterlist *sgs[],
|
|
|
|
unsigned int total_sg,
|
|
|
|
unsigned int out_sgs,
|
|
|
|
unsigned int in_sgs,
|
|
|
|
void *data,
|
|
|
|
void *ctx,
|
|
|
|
gfp_t gfp)
|
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
struct vring_packed_desc *desc;
|
|
|
|
struct scatterlist *sg;
|
|
|
|
unsigned int i, n, c, descs_used, err_idx;
|
treewide: Remove uninitialized_var() usage
Using uninitialized_var() is dangerous as it papers over real bugs[1]
(or can in the future), and suppresses unrelated compiler warnings
(e.g. "unused variable"). If the compiler thinks it is uninitialized,
either simply initialize the variable or make compiler changes.
In preparation for removing[2] the[3] macro[4], remove all remaining
needless uses with the following script:
git grep '\buninitialized_var\b' | cut -d: -f1 | sort -u | \
xargs perl -pi -e \
's/\buninitialized_var\(([^\)]+)\)/\1/g;
s:\s*/\* (GCC be quiet|to make compiler happy) \*/$::g;'
drivers/video/fbdev/riva/riva_hw.c was manually tweaked to avoid
pathological white-space.
No outstanding warnings were found building allmodconfig with GCC 9.3.0
for x86_64, i386, arm64, arm, powerpc, powerpc64le, s390x, mips, sparc64,
alpha, and m68k.
[1] https://lore.kernel.org/lkml/20200603174714.192027-1-glider@google.com/
[2] https://lore.kernel.org/lkml/CA+55aFw+Vbj0i=1TGqCR5vQkCzWJ0QxK6CernOU6eedsudAixw@mail.gmail.com/
[3] https://lore.kernel.org/lkml/CA+55aFwgbgqhbp1fkxvRKEpzyR5J8n1vKT1VZdz9knmPuXhOeg@mail.gmail.com/
[4] https://lore.kernel.org/lkml/CA+55aFz2500WfbKXAx8s67wrm9=yVJu65TpLgN_ybYNv0VEOKA@mail.gmail.com/
Reviewed-by: Leon Romanovsky <leonro@mellanox.com> # drivers/infiniband and mlx4/mlx5
Acked-by: Jason Gunthorpe <jgg@mellanox.com> # IB
Acked-by: Kalle Valo <kvalo@codeaurora.org> # wireless drivers
Reviewed-by: Chao Yu <yuchao0@huawei.com> # erofs
Signed-off-by: Kees Cook <keescook@chromium.org>
2020-06-04 04:09:38 +08:00
|
|
|
__le16 head_flags, flags;
|
|
|
|
u16 head, id, prev, curr, avail_used_flags;
|
2021-10-20 19:23:23 +08:00
|
|
|
int err;
|
2018-11-21 18:03:27 +08:00
|
|
|
|
|
|
|
START_USE(vq);
|
|
|
|
|
|
|
|
BUG_ON(data == NULL);
|
|
|
|
BUG_ON(ctx && vq->indirect);
|
|
|
|
|
|
|
|
if (unlikely(vq->broken)) {
|
|
|
|
END_USE(vq);
|
|
|
|
return -EIO;
|
|
|
|
}
|
|
|
|
|
|
|
|
LAST_ADD_TIME_UPDATE(vq);
|
|
|
|
|
|
|
|
BUG_ON(total_sg == 0);
|
|
|
|
|
2022-03-28 18:58:16 +08:00
|
|
|
if (virtqueue_use_indirect(vq, total_sg)) {
|
2021-10-20 19:23:23 +08:00
|
|
|
err = virtqueue_add_indirect_packed(vq, sgs, total_sg, out_sgs,
|
|
|
|
in_sgs, data, gfp);
|
2022-01-06 20:57:46 +08:00
|
|
|
if (err != -ENOMEM) {
|
|
|
|
END_USE(vq);
|
2021-10-20 19:23:23 +08:00
|
|
|
return err;
|
2022-01-06 20:57:46 +08:00
|
|
|
}
|
2021-10-20 19:23:23 +08:00
|
|
|
|
|
|
|
/* fall back on direct */
|
|
|
|
}
|
2018-11-21 18:03:27 +08:00
|
|
|
|
|
|
|
head = vq->packed.next_avail_idx;
|
|
|
|
avail_used_flags = vq->packed.avail_used_flags;
|
|
|
|
|
|
|
|
WARN_ON_ONCE(total_sg > vq->packed.vring.num && !vq->indirect);
|
|
|
|
|
|
|
|
desc = vq->packed.vring.desc;
|
|
|
|
i = head;
|
|
|
|
descs_used = total_sg;
|
|
|
|
|
|
|
|
if (unlikely(vq->vq.num_free < descs_used)) {
|
|
|
|
pr_debug("Can't add buf len %i - avail = %i\n",
|
|
|
|
descs_used, vq->vq.num_free);
|
|
|
|
END_USE(vq);
|
|
|
|
return -ENOSPC;
|
|
|
|
}
|
|
|
|
|
|
|
|
id = vq->free_head;
|
|
|
|
BUG_ON(id == vq->packed.vring.num);
|
|
|
|
|
|
|
|
curr = id;
|
|
|
|
c = 0;
|
|
|
|
for (n = 0; n < out_sgs + in_sgs; n++) {
|
|
|
|
for (sg = sgs[n]; sg; sg = sg_next(sg)) {
|
2023-08-10 20:30:47 +08:00
|
|
|
dma_addr_t addr;
|
|
|
|
|
|
|
|
if (vring_map_one_sg(vq, sg, n < out_sgs ?
|
|
|
|
DMA_TO_DEVICE : DMA_FROM_DEVICE, &addr))
|
2018-11-21 18:03:27 +08:00
|
|
|
goto unmap_release;
|
|
|
|
|
|
|
|
flags = cpu_to_le16(vq->packed.avail_used_flags |
|
|
|
|
(++c == total_sg ? 0 : VRING_DESC_F_NEXT) |
|
|
|
|
(n < out_sgs ? 0 : VRING_DESC_F_WRITE));
|
|
|
|
if (i == head)
|
|
|
|
head_flags = flags;
|
|
|
|
else
|
|
|
|
desc[i].flags = flags;
|
|
|
|
|
|
|
|
desc[i].addr = cpu_to_le64(addr);
|
|
|
|
desc[i].len = cpu_to_le32(sg->length);
|
|
|
|
desc[i].id = cpu_to_le16(id);
|
|
|
|
|
2024-02-23 15:18:33 +08:00
|
|
|
if (unlikely(vq->use_dma_api)) {
|
2018-11-21 18:03:27 +08:00
|
|
|
vq->packed.desc_extra[curr].addr = addr;
|
|
|
|
vq->packed.desc_extra[curr].len = sg->length;
|
|
|
|
vq->packed.desc_extra[curr].flags =
|
|
|
|
le16_to_cpu(flags);
|
|
|
|
}
|
|
|
|
prev = curr;
|
2021-06-04 13:53:44 +08:00
|
|
|
curr = vq->packed.desc_extra[curr].next;
|
2018-11-21 18:03:27 +08:00
|
|
|
|
|
|
|
if ((unlikely(++i >= vq->packed.vring.num))) {
|
|
|
|
i = 0;
|
|
|
|
vq->packed.avail_used_flags ^=
|
|
|
|
1 << VRING_PACKED_DESC_F_AVAIL |
|
|
|
|
1 << VRING_PACKED_DESC_F_USED;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
virtio_ring: fix avail_wrap_counter in virtqueue_add_packed
In current packed virtqueue implementation, the avail_wrap_counter won't
flip, in the case when the driver supplies a descriptor chain with a
length equals to the queue size; total_sg == vq->packed.vring.num.
Let’s assume the following situation:
vq->packed.vring.num=4
vq->packed.next_avail_idx: 1
vq->packed.avail_wrap_counter: 0
Then the driver adds a descriptor chain containing 4 descriptors.
We expect the following result with avail_wrap_counter flipped:
vq->packed.next_avail_idx: 1
vq->packed.avail_wrap_counter: 1
But, the current implementation gives the following result:
vq->packed.next_avail_idx: 1
vq->packed.avail_wrap_counter: 0
To reproduce the bug, you can set a packed queue size as small as
possible, so that the driver is more likely to provide a descriptor
chain with a length equal to the packed queue size. For example, in
qemu run following commands:
sudo qemu-system-x86_64 \
-enable-kvm \
-nographic \
-kernel "path/to/kernel_image" \
-m 1G \
-drive file="path/to/rootfs",if=none,id=disk \
-device virtio-blk,drive=disk \
-drive file="path/to/disk_image",if=none,id=rwdisk \
-device virtio-blk,drive=rwdisk,packed=on,queue-size=4,\
indirect_desc=off \
-append "console=ttyS0 root=/dev/vda rw init=/bin/bash"
Inside the VM, create a directory and mount the rwdisk device on it. The
rwdisk will hang and mount operation will not complete.
This commit fixes the wrap counter error by flipping the
packed.avail_wrap_counter, when start of descriptor chain equals to the
end of descriptor chain (head == i).
Fixes: 1ce9e6055fa0 ("virtio_ring: introduce packed ring support")
Signed-off-by: Yuan Yao <yuanyaogoog@chromium.org>
Message-Id: <20230808051110.3492693-1-yuanyaogoog@chromium.org>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2023-08-08 13:10:59 +08:00
|
|
|
if (i <= head)
|
2018-11-21 18:03:27 +08:00
|
|
|
vq->packed.avail_wrap_counter ^= 1;
|
|
|
|
|
|
|
|
/* We're using some buffers from the free list. */
|
|
|
|
vq->vq.num_free -= descs_used;
|
|
|
|
|
|
|
|
/* Update free pointer */
|
|
|
|
vq->packed.next_avail_idx = i;
|
|
|
|
vq->free_head = curr;
|
|
|
|
|
|
|
|
/* Store token. */
|
|
|
|
vq->packed.desc_state[id].num = descs_used;
|
|
|
|
vq->packed.desc_state[id].data = data;
|
|
|
|
vq->packed.desc_state[id].indir_desc = ctx;
|
|
|
|
vq->packed.desc_state[id].last = prev;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* A driver MUST NOT make the first descriptor in the list
|
|
|
|
* available before all subsequent descriptors comprising
|
|
|
|
* the list are made available.
|
|
|
|
*/
|
|
|
|
virtio_wmb(vq->weak_barriers);
|
|
|
|
vq->packed.vring.desc[head].flags = head_flags;
|
|
|
|
vq->num_added += descs_used;
|
|
|
|
|
|
|
|
pr_debug("Added buffer head %i to %p\n", head, vq);
|
|
|
|
END_USE(vq);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
unmap_release:
|
|
|
|
err_idx = i;
|
|
|
|
i = head;
|
2021-06-04 13:53:47 +08:00
|
|
|
curr = vq->free_head;
|
2018-11-21 18:03:27 +08:00
|
|
|
|
|
|
|
vq->packed.avail_used_flags = avail_used_flags;
|
|
|
|
|
|
|
|
for (n = 0; n < total_sg; n++) {
|
|
|
|
if (i == err_idx)
|
|
|
|
break;
|
2022-02-24 19:03:54 +08:00
|
|
|
vring_unmap_extra_packed(vq, &vq->packed.desc_extra[curr]);
|
2021-06-04 13:53:47 +08:00
|
|
|
curr = vq->packed.desc_extra[curr].next;
|
2018-11-21 18:03:27 +08:00
|
|
|
i++;
|
|
|
|
if (i >= vq->packed.vring.num)
|
|
|
|
i = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
END_USE(vq);
|
|
|
|
return -EIO;
|
|
|
|
}
|
|
|
|
|
|
|
|
static bool virtqueue_kick_prepare_packed(struct virtqueue *_vq)
|
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
2018-11-21 18:03:28 +08:00
|
|
|
u16 new, old, off_wrap, flags, wrap_counter, event_idx;
|
2018-11-21 18:03:27 +08:00
|
|
|
bool needs_kick;
|
|
|
|
union {
|
|
|
|
struct {
|
|
|
|
__le16 off_wrap;
|
|
|
|
__le16 flags;
|
|
|
|
};
|
|
|
|
u32 u32;
|
|
|
|
} snapshot;
|
|
|
|
|
|
|
|
START_USE(vq);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* We need to expose the new flags value before checking notification
|
|
|
|
* suppressions.
|
|
|
|
*/
|
|
|
|
virtio_mb(vq->weak_barriers);
|
|
|
|
|
2018-11-21 18:03:28 +08:00
|
|
|
old = vq->packed.next_avail_idx - vq->num_added;
|
|
|
|
new = vq->packed.next_avail_idx;
|
2018-11-21 18:03:27 +08:00
|
|
|
vq->num_added = 0;
|
|
|
|
|
|
|
|
snapshot.u32 = *(u32 *)vq->packed.vring.device;
|
|
|
|
flags = le16_to_cpu(snapshot.flags);
|
|
|
|
|
|
|
|
LAST_ADD_TIME_CHECK(vq);
|
|
|
|
LAST_ADD_TIME_INVALID(vq);
|
|
|
|
|
2018-11-21 18:03:28 +08:00
|
|
|
if (flags != VRING_PACKED_EVENT_FLAG_DESC) {
|
|
|
|
needs_kick = (flags != VRING_PACKED_EVENT_FLAG_DISABLE);
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
off_wrap = le16_to_cpu(snapshot.off_wrap);
|
|
|
|
|
|
|
|
wrap_counter = off_wrap >> VRING_PACKED_EVENT_F_WRAP_CTR;
|
|
|
|
event_idx = off_wrap & ~(1 << VRING_PACKED_EVENT_F_WRAP_CTR);
|
|
|
|
if (wrap_counter != vq->packed.avail_wrap_counter)
|
|
|
|
event_idx -= vq->packed.vring.num;
|
|
|
|
|
|
|
|
needs_kick = vring_need_event(event_idx, new, old);
|
|
|
|
out:
|
2018-11-21 18:03:27 +08:00
|
|
|
END_USE(vq);
|
|
|
|
return needs_kick;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void detach_buf_packed(struct vring_virtqueue *vq,
|
|
|
|
unsigned int id, void **ctx)
|
|
|
|
{
|
|
|
|
struct vring_desc_state_packed *state = NULL;
|
|
|
|
struct vring_packed_desc *desc;
|
|
|
|
unsigned int i, curr;
|
|
|
|
|
|
|
|
state = &vq->packed.desc_state[id];
|
|
|
|
|
|
|
|
/* Clear data ptr. */
|
|
|
|
state->data = NULL;
|
|
|
|
|
2021-06-04 13:53:44 +08:00
|
|
|
vq->packed.desc_extra[state->last].next = vq->free_head;
|
2018-11-21 18:03:27 +08:00
|
|
|
vq->free_head = id;
|
|
|
|
vq->vq.num_free += state->num;
|
|
|
|
|
2024-02-23 15:18:33 +08:00
|
|
|
if (unlikely(vq->use_dma_api)) {
|
2018-11-21 18:03:27 +08:00
|
|
|
curr = id;
|
|
|
|
for (i = 0; i < state->num; i++) {
|
2022-02-24 19:03:54 +08:00
|
|
|
vring_unmap_extra_packed(vq,
|
|
|
|
&vq->packed.desc_extra[curr]);
|
2021-06-04 13:53:44 +08:00
|
|
|
curr = vq->packed.desc_extra[curr].next;
|
2018-11-21 18:03:27 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (vq->indirect) {
|
|
|
|
u32 len;
|
|
|
|
|
|
|
|
/* Free the indirect table, if any, now that it's unmapped. */
|
|
|
|
desc = state->indir_desc;
|
|
|
|
if (!desc)
|
|
|
|
return;
|
|
|
|
|
2023-08-10 20:30:51 +08:00
|
|
|
if (vq->do_unmap) {
|
2018-11-21 18:03:27 +08:00
|
|
|
len = vq->packed.desc_extra[id].len;
|
|
|
|
for (i = 0; i < len / sizeof(struct vring_packed_desc);
|
|
|
|
i++)
|
|
|
|
vring_unmap_desc_packed(vq, &desc[i]);
|
|
|
|
}
|
|
|
|
kfree(desc);
|
|
|
|
state->indir_desc = NULL;
|
|
|
|
} else if (ctx) {
|
|
|
|
*ctx = state->indir_desc;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline bool is_used_desc_packed(const struct vring_virtqueue *vq,
|
|
|
|
u16 idx, bool used_wrap_counter)
|
|
|
|
{
|
|
|
|
bool avail, used;
|
|
|
|
u16 flags;
|
|
|
|
|
|
|
|
flags = le16_to_cpu(vq->packed.vring.desc[idx].flags);
|
|
|
|
avail = !!(flags & (1 << VRING_PACKED_DESC_F_AVAIL));
|
|
|
|
used = !!(flags & (1 << VRING_PACKED_DESC_F_USED));
|
|
|
|
|
|
|
|
return avail == used && used == used_wrap_counter;
|
|
|
|
}
|
|
|
|
|
2023-03-10 13:34:27 +08:00
|
|
|
static bool more_used_packed(const struct vring_virtqueue *vq)
|
2018-11-21 18:03:27 +08:00
|
|
|
{
|
2022-06-17 10:04:11 +08:00
|
|
|
u16 last_used;
|
|
|
|
u16 last_used_idx;
|
|
|
|
bool used_wrap_counter;
|
|
|
|
|
|
|
|
last_used_idx = READ_ONCE(vq->last_used_idx);
|
|
|
|
last_used = packed_last_used(last_used_idx);
|
|
|
|
used_wrap_counter = packed_used_wrap_counter(last_used_idx);
|
|
|
|
return is_used_desc_packed(vq, last_used, used_wrap_counter);
|
2018-11-21 18:03:27 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static void *virtqueue_get_buf_ctx_packed(struct virtqueue *_vq,
|
|
|
|
unsigned int *len,
|
|
|
|
void **ctx)
|
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
2022-06-17 10:04:11 +08:00
|
|
|
u16 last_used, id, last_used_idx;
|
|
|
|
bool used_wrap_counter;
|
2018-11-21 18:03:27 +08:00
|
|
|
void *ret;
|
|
|
|
|
|
|
|
START_USE(vq);
|
|
|
|
|
|
|
|
if (unlikely(vq->broken)) {
|
|
|
|
END_USE(vq);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!more_used_packed(vq)) {
|
|
|
|
pr_debug("No more buffers in queue\n");
|
|
|
|
END_USE(vq);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Only get used elements after they have been exposed by host. */
|
|
|
|
virtio_rmb(vq->weak_barriers);
|
|
|
|
|
2022-06-17 10:04:11 +08:00
|
|
|
last_used_idx = READ_ONCE(vq->last_used_idx);
|
|
|
|
used_wrap_counter = packed_used_wrap_counter(last_used_idx);
|
|
|
|
last_used = packed_last_used(last_used_idx);
|
2018-11-21 18:03:27 +08:00
|
|
|
id = le16_to_cpu(vq->packed.vring.desc[last_used].id);
|
|
|
|
*len = le32_to_cpu(vq->packed.vring.desc[last_used].len);
|
|
|
|
|
|
|
|
if (unlikely(id >= vq->packed.vring.num)) {
|
|
|
|
BAD_RING(vq, "id %u out of range\n", id);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
if (unlikely(!vq->packed.desc_state[id].data)) {
|
|
|
|
BAD_RING(vq, "id %u is not a head!\n", id);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* detach_buf_packed clears data, so grab it now. */
|
|
|
|
ret = vq->packed.desc_state[id].data;
|
|
|
|
detach_buf_packed(vq, id, ctx);
|
|
|
|
|
2022-06-17 10:04:11 +08:00
|
|
|
last_used += vq->packed.desc_state[id].num;
|
|
|
|
if (unlikely(last_used >= vq->packed.vring.num)) {
|
|
|
|
last_used -= vq->packed.vring.num;
|
|
|
|
used_wrap_counter ^= 1;
|
2018-11-21 18:03:27 +08:00
|
|
|
}
|
|
|
|
|
2022-06-17 10:04:11 +08:00
|
|
|
last_used = (last_used | (used_wrap_counter << VRING_PACKED_EVENT_F_WRAP_CTR));
|
|
|
|
WRITE_ONCE(vq->last_used_idx, last_used);
|
|
|
|
|
2018-11-21 18:03:28 +08:00
|
|
|
/*
|
|
|
|
* If we expect an interrupt for the next entry, tell host
|
|
|
|
* by writing event index and flush out the write before
|
|
|
|
* the read in the next get_buf call.
|
|
|
|
*/
|
|
|
|
if (vq->packed.event_flags_shadow == VRING_PACKED_EVENT_FLAG_DESC)
|
|
|
|
virtio_store_mb(vq->weak_barriers,
|
|
|
|
&vq->packed.vring.driver->off_wrap,
|
2022-06-17 10:04:11 +08:00
|
|
|
cpu_to_le16(vq->last_used_idx));
|
2018-11-21 18:03:28 +08:00
|
|
|
|
2018-11-21 18:03:27 +08:00
|
|
|
LAST_ADD_TIME_INVALID(vq);
|
|
|
|
|
|
|
|
END_USE(vq);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void virtqueue_disable_cb_packed(struct virtqueue *_vq)
|
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
|
|
|
|
if (vq->packed.event_flags_shadow != VRING_PACKED_EVENT_FLAG_DISABLE) {
|
|
|
|
vq->packed.event_flags_shadow = VRING_PACKED_EVENT_FLAG_DISABLE;
|
2023-03-29 18:23:00 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* If device triggered an event already it won't trigger one again:
|
|
|
|
* no need to disable.
|
|
|
|
*/
|
|
|
|
if (vq->event_triggered)
|
|
|
|
return;
|
|
|
|
|
2018-11-21 18:03:27 +08:00
|
|
|
vq->packed.vring.driver->flags =
|
|
|
|
cpu_to_le16(vq->packed.event_flags_shadow);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2022-04-18 10:54:35 +08:00
|
|
|
static unsigned int virtqueue_enable_cb_prepare_packed(struct virtqueue *_vq)
|
2018-11-21 18:03:27 +08:00
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
|
|
|
|
START_USE(vq);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* We optimistically turn back on interrupts, then check if there was
|
|
|
|
* more to do.
|
|
|
|
*/
|
|
|
|
|
2018-11-21 18:03:28 +08:00
|
|
|
if (vq->event) {
|
|
|
|
vq->packed.vring.driver->off_wrap =
|
2022-06-17 10:04:11 +08:00
|
|
|
cpu_to_le16(vq->last_used_idx);
|
2018-11-21 18:03:28 +08:00
|
|
|
/*
|
|
|
|
* We need to update event offset and event wrap
|
|
|
|
* counter first before updating event flags.
|
|
|
|
*/
|
|
|
|
virtio_wmb(vq->weak_barriers);
|
|
|
|
}
|
|
|
|
|
2018-11-21 18:03:27 +08:00
|
|
|
if (vq->packed.event_flags_shadow == VRING_PACKED_EVENT_FLAG_DISABLE) {
|
2018-11-21 18:03:28 +08:00
|
|
|
vq->packed.event_flags_shadow = vq->event ?
|
|
|
|
VRING_PACKED_EVENT_FLAG_DESC :
|
|
|
|
VRING_PACKED_EVENT_FLAG_ENABLE;
|
2018-11-21 18:03:27 +08:00
|
|
|
vq->packed.vring.driver->flags =
|
|
|
|
cpu_to_le16(vq->packed.event_flags_shadow);
|
|
|
|
}
|
|
|
|
|
|
|
|
END_USE(vq);
|
2022-06-17 10:04:11 +08:00
|
|
|
return vq->last_used_idx;
|
2018-11-21 18:03:27 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static bool virtqueue_poll_packed(struct virtqueue *_vq, u16 off_wrap)
|
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
bool wrap_counter;
|
|
|
|
u16 used_idx;
|
|
|
|
|
|
|
|
wrap_counter = off_wrap >> VRING_PACKED_EVENT_F_WRAP_CTR;
|
|
|
|
used_idx = off_wrap & ~(1 << VRING_PACKED_EVENT_F_WRAP_CTR);
|
|
|
|
|
|
|
|
return is_used_desc_packed(vq, used_idx, wrap_counter);
|
|
|
|
}
|
|
|
|
|
|
|
|
static bool virtqueue_enable_cb_delayed_packed(struct virtqueue *_vq)
|
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
2022-06-17 10:04:11 +08:00
|
|
|
u16 used_idx, wrap_counter, last_used_idx;
|
2018-11-21 18:03:28 +08:00
|
|
|
u16 bufs;
|
2018-11-21 18:03:27 +08:00
|
|
|
|
|
|
|
START_USE(vq);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* We optimistically turn back on interrupts, then check if there was
|
|
|
|
* more to do.
|
|
|
|
*/
|
|
|
|
|
2018-11-21 18:03:28 +08:00
|
|
|
if (vq->event) {
|
|
|
|
/* TODO: tune this threshold */
|
|
|
|
bufs = (vq->packed.vring.num - vq->vq.num_free) * 3 / 4;
|
2022-06-17 10:04:11 +08:00
|
|
|
last_used_idx = READ_ONCE(vq->last_used_idx);
|
|
|
|
wrap_counter = packed_used_wrap_counter(last_used_idx);
|
2018-11-21 18:03:28 +08:00
|
|
|
|
2022-06-17 10:04:11 +08:00
|
|
|
used_idx = packed_last_used(last_used_idx) + bufs;
|
2018-11-21 18:03:28 +08:00
|
|
|
if (used_idx >= vq->packed.vring.num) {
|
|
|
|
used_idx -= vq->packed.vring.num;
|
|
|
|
wrap_counter ^= 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
vq->packed.vring.driver->off_wrap = cpu_to_le16(used_idx |
|
|
|
|
(wrap_counter << VRING_PACKED_EVENT_F_WRAP_CTR));
|
|
|
|
|
|
|
|
/*
|
|
|
|
* We need to update event offset and event wrap
|
|
|
|
* counter first before updating event flags.
|
|
|
|
*/
|
|
|
|
virtio_wmb(vq->weak_barriers);
|
|
|
|
}
|
2018-11-21 18:03:27 +08:00
|
|
|
|
|
|
|
if (vq->packed.event_flags_shadow == VRING_PACKED_EVENT_FLAG_DISABLE) {
|
2018-11-21 18:03:28 +08:00
|
|
|
vq->packed.event_flags_shadow = vq->event ?
|
|
|
|
VRING_PACKED_EVENT_FLAG_DESC :
|
|
|
|
VRING_PACKED_EVENT_FLAG_ENABLE;
|
2018-11-21 18:03:27 +08:00
|
|
|
vq->packed.vring.driver->flags =
|
|
|
|
cpu_to_le16(vq->packed.event_flags_shadow);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* We need to update event suppression structure first
|
|
|
|
* before re-checking for more used buffers.
|
|
|
|
*/
|
|
|
|
virtio_mb(vq->weak_barriers);
|
|
|
|
|
2022-06-17 10:04:11 +08:00
|
|
|
last_used_idx = READ_ONCE(vq->last_used_idx);
|
|
|
|
wrap_counter = packed_used_wrap_counter(last_used_idx);
|
|
|
|
used_idx = packed_last_used(last_used_idx);
|
|
|
|
if (is_used_desc_packed(vq, used_idx, wrap_counter)) {
|
2018-11-21 18:03:27 +08:00
|
|
|
END_USE(vq);
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
|
|
|
END_USE(vq);
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void *virtqueue_detach_unused_buf_packed(struct virtqueue *_vq)
|
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
unsigned int i;
|
|
|
|
void *buf;
|
|
|
|
|
|
|
|
START_USE(vq);
|
|
|
|
|
|
|
|
for (i = 0; i < vq->packed.vring.num; i++) {
|
|
|
|
if (!vq->packed.desc_state[i].data)
|
|
|
|
continue;
|
|
|
|
/* detach_buf clears data, so grab it now. */
|
|
|
|
buf = vq->packed.desc_state[i].data;
|
|
|
|
detach_buf_packed(vq, i, NULL);
|
|
|
|
END_USE(vq);
|
|
|
|
return buf;
|
|
|
|
}
|
|
|
|
/* That should have freed everything. */
|
|
|
|
BUG_ON(vq->vq.num_free != vq->packed.vring.num);
|
|
|
|
|
|
|
|
END_USE(vq);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2022-06-24 10:55:45 +08:00
|
|
|
static struct vring_desc_extra *vring_alloc_desc_extra(unsigned int num)
|
2021-06-04 13:53:46 +08:00
|
|
|
{
|
|
|
|
struct vring_desc_extra *desc_extra;
|
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
desc_extra = kmalloc_array(num, sizeof(struct vring_desc_extra),
|
|
|
|
GFP_KERNEL);
|
|
|
|
if (!desc_extra)
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
memset(desc_extra, 0, num * sizeof(struct vring_desc_extra));
|
|
|
|
|
|
|
|
for (i = 0; i < num - 1; i++)
|
|
|
|
desc_extra[i].next = i + 1;
|
|
|
|
|
|
|
|
return desc_extra;
|
|
|
|
}
|
|
|
|
|
2022-08-01 14:38:37 +08:00
|
|
|
static void vring_free_packed(struct vring_virtqueue_packed *vring_packed,
|
2023-01-19 14:15:21 +08:00
|
|
|
struct virtio_device *vdev,
|
|
|
|
struct device *dma_dev)
|
2022-08-01 14:38:37 +08:00
|
|
|
{
|
|
|
|
if (vring_packed->vring.desc)
|
|
|
|
vring_free_queue(vdev, vring_packed->ring_size_in_bytes,
|
|
|
|
vring_packed->vring.desc,
|
2023-01-19 14:15:21 +08:00
|
|
|
vring_packed->ring_dma_addr,
|
|
|
|
dma_dev);
|
2022-08-01 14:38:37 +08:00
|
|
|
|
|
|
|
if (vring_packed->vring.driver)
|
|
|
|
vring_free_queue(vdev, vring_packed->event_size_in_bytes,
|
|
|
|
vring_packed->vring.driver,
|
2023-01-19 14:15:21 +08:00
|
|
|
vring_packed->driver_event_dma_addr,
|
|
|
|
dma_dev);
|
2022-08-01 14:38:37 +08:00
|
|
|
|
|
|
|
if (vring_packed->vring.device)
|
|
|
|
vring_free_queue(vdev, vring_packed->event_size_in_bytes,
|
|
|
|
vring_packed->vring.device,
|
2023-01-19 14:15:21 +08:00
|
|
|
vring_packed->device_event_dma_addr,
|
|
|
|
dma_dev);
|
2022-08-01 14:38:37 +08:00
|
|
|
|
|
|
|
kfree(vring_packed->desc_state);
|
|
|
|
kfree(vring_packed->desc_extra);
|
|
|
|
}
|
|
|
|
|
2022-08-01 14:38:38 +08:00
|
|
|
static int vring_alloc_queue_packed(struct vring_virtqueue_packed *vring_packed,
|
|
|
|
struct virtio_device *vdev,
|
2023-01-19 14:15:21 +08:00
|
|
|
u32 num, struct device *dma_dev)
|
2018-11-21 18:03:27 +08:00
|
|
|
{
|
|
|
|
struct vring_packed_desc *ring;
|
|
|
|
struct vring_packed_desc_event *driver, *device;
|
|
|
|
dma_addr_t ring_dma_addr, driver_event_dma_addr, device_event_dma_addr;
|
|
|
|
size_t ring_size_in_bytes, event_size_in_bytes;
|
|
|
|
|
|
|
|
ring_size_in_bytes = num * sizeof(struct vring_packed_desc);
|
|
|
|
|
|
|
|
ring = vring_alloc_queue(vdev, ring_size_in_bytes,
|
|
|
|
&ring_dma_addr,
|
2023-01-19 14:15:21 +08:00
|
|
|
GFP_KERNEL | __GFP_NOWARN | __GFP_ZERO,
|
|
|
|
dma_dev);
|
2018-11-21 18:03:27 +08:00
|
|
|
if (!ring)
|
2022-08-01 14:38:38 +08:00
|
|
|
goto err;
|
|
|
|
|
|
|
|
vring_packed->vring.desc = ring;
|
|
|
|
vring_packed->ring_dma_addr = ring_dma_addr;
|
|
|
|
vring_packed->ring_size_in_bytes = ring_size_in_bytes;
|
2018-11-21 18:03:27 +08:00
|
|
|
|
|
|
|
event_size_in_bytes = sizeof(struct vring_packed_desc_event);
|
|
|
|
|
|
|
|
driver = vring_alloc_queue(vdev, event_size_in_bytes,
|
|
|
|
&driver_event_dma_addr,
|
2023-01-19 14:15:21 +08:00
|
|
|
GFP_KERNEL | __GFP_NOWARN | __GFP_ZERO,
|
|
|
|
dma_dev);
|
2018-11-21 18:03:27 +08:00
|
|
|
if (!driver)
|
2022-08-01 14:38:38 +08:00
|
|
|
goto err;
|
|
|
|
|
|
|
|
vring_packed->vring.driver = driver;
|
|
|
|
vring_packed->event_size_in_bytes = event_size_in_bytes;
|
|
|
|
vring_packed->driver_event_dma_addr = driver_event_dma_addr;
|
2018-11-21 18:03:27 +08:00
|
|
|
|
|
|
|
device = vring_alloc_queue(vdev, event_size_in_bytes,
|
|
|
|
&device_event_dma_addr,
|
2023-01-19 14:15:21 +08:00
|
|
|
GFP_KERNEL | __GFP_NOWARN | __GFP_ZERO,
|
|
|
|
dma_dev);
|
2018-11-21 18:03:27 +08:00
|
|
|
if (!device)
|
2022-08-01 14:38:38 +08:00
|
|
|
goto err;
|
|
|
|
|
|
|
|
vring_packed->vring.device = device;
|
|
|
|
vring_packed->device_event_dma_addr = device_event_dma_addr;
|
|
|
|
|
|
|
|
vring_packed->vring.num = num;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
err:
|
2023-01-19 14:15:21 +08:00
|
|
|
vring_free_packed(vring_packed, vdev, dma_dev);
|
2022-08-01 14:38:38 +08:00
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
|
2022-08-01 14:38:39 +08:00
|
|
|
static int vring_alloc_state_extra_packed(struct vring_virtqueue_packed *vring_packed)
|
|
|
|
{
|
|
|
|
struct vring_desc_state_packed *state;
|
|
|
|
struct vring_desc_extra *extra;
|
|
|
|
u32 num = vring_packed->vring.num;
|
|
|
|
|
|
|
|
state = kmalloc_array(num, sizeof(struct vring_desc_state_packed), GFP_KERNEL);
|
|
|
|
if (!state)
|
|
|
|
goto err_desc_state;
|
|
|
|
|
|
|
|
memset(state, 0, num * sizeof(struct vring_desc_state_packed));
|
|
|
|
|
|
|
|
extra = vring_alloc_desc_extra(num);
|
|
|
|
if (!extra)
|
|
|
|
goto err_desc_extra;
|
|
|
|
|
|
|
|
vring_packed->desc_state = state;
|
|
|
|
vring_packed->desc_extra = extra;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
err_desc_extra:
|
|
|
|
kfree(state);
|
|
|
|
err_desc_state:
|
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
|
2022-08-01 14:38:40 +08:00
|
|
|
static void virtqueue_vring_init_packed(struct vring_virtqueue_packed *vring_packed,
|
|
|
|
bool callback)
|
|
|
|
{
|
|
|
|
vring_packed->next_avail_idx = 0;
|
|
|
|
vring_packed->avail_wrap_counter = 1;
|
|
|
|
vring_packed->event_flags_shadow = 0;
|
|
|
|
vring_packed->avail_used_flags = 1 << VRING_PACKED_DESC_F_AVAIL;
|
|
|
|
|
|
|
|
/* No callback? Tell other side not to bother us. */
|
|
|
|
if (!callback) {
|
|
|
|
vring_packed->event_flags_shadow = VRING_PACKED_EVENT_FLAG_DISABLE;
|
|
|
|
vring_packed->vring.driver->flags =
|
|
|
|
cpu_to_le16(vring_packed->event_flags_shadow);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2022-08-01 14:38:41 +08:00
|
|
|
static void virtqueue_vring_attach_packed(struct vring_virtqueue *vq,
|
|
|
|
struct vring_virtqueue_packed *vring_packed)
|
|
|
|
{
|
|
|
|
vq->packed = *vring_packed;
|
|
|
|
|
|
|
|
/* Put everything in free lists. */
|
|
|
|
vq->free_head = 0;
|
|
|
|
}
|
|
|
|
|
2022-08-01 14:38:42 +08:00
|
|
|
static void virtqueue_reinit_packed(struct vring_virtqueue *vq)
|
|
|
|
{
|
|
|
|
memset(vq->packed.vring.device, 0, vq->packed.event_size_in_bytes);
|
|
|
|
memset(vq->packed.vring.driver, 0, vq->packed.event_size_in_bytes);
|
|
|
|
|
|
|
|
/* we need to reset the desc.flags. For more, see is_used_desc_packed() */
|
|
|
|
memset(vq->packed.vring.desc, 0, vq->packed.ring_size_in_bytes);
|
|
|
|
|
|
|
|
virtqueue_init(vq, vq->packed.vring.num);
|
|
|
|
virtqueue_vring_init_packed(&vq->packed, !!vq->vq.callback);
|
|
|
|
}
|
|
|
|
|
2022-08-01 14:38:38 +08:00
|
|
|
static struct virtqueue *vring_create_virtqueue_packed(
|
|
|
|
unsigned int index,
|
|
|
|
unsigned int num,
|
|
|
|
unsigned int vring_align,
|
|
|
|
struct virtio_device *vdev,
|
|
|
|
bool weak_barriers,
|
|
|
|
bool may_reduce_num,
|
|
|
|
bool context,
|
|
|
|
bool (*notify)(struct virtqueue *),
|
|
|
|
void (*callback)(struct virtqueue *),
|
2023-01-19 14:15:21 +08:00
|
|
|
const char *name,
|
|
|
|
struct device *dma_dev)
|
2022-08-01 14:38:38 +08:00
|
|
|
{
|
|
|
|
struct vring_virtqueue_packed vring_packed = {};
|
|
|
|
struct vring_virtqueue *vq;
|
2022-08-01 14:38:39 +08:00
|
|
|
int err;
|
2022-08-01 14:38:38 +08:00
|
|
|
|
2023-01-19 14:15:21 +08:00
|
|
|
if (vring_alloc_queue_packed(&vring_packed, vdev, num, dma_dev))
|
2022-08-01 14:38:38 +08:00
|
|
|
goto err_ring;
|
2018-11-21 18:03:27 +08:00
|
|
|
|
|
|
|
vq = kmalloc(sizeof(*vq), GFP_KERNEL);
|
|
|
|
if (!vq)
|
|
|
|
goto err_vq;
|
|
|
|
|
|
|
|
vq->vq.callback = callback;
|
|
|
|
vq->vq.vdev = vdev;
|
|
|
|
vq->vq.name = name;
|
|
|
|
vq->vq.index = index;
|
2022-08-01 14:38:48 +08:00
|
|
|
vq->vq.reset = false;
|
2018-11-21 18:03:27 +08:00
|
|
|
vq->we_own_ring = true;
|
|
|
|
vq->notify = notify;
|
|
|
|
vq->weak_barriers = weak_barriers;
|
2022-06-22 09:29:40 +08:00
|
|
|
#ifdef CONFIG_VIRTIO_HARDEN_NOTIFICATION
|
2022-05-27 14:01:19 +08:00
|
|
|
vq->broken = true;
|
2022-06-22 09:29:40 +08:00
|
|
|
#else
|
|
|
|
vq->broken = false;
|
|
|
|
#endif
|
2018-11-21 18:03:27 +08:00
|
|
|
vq->packed_ring = true;
|
2023-01-19 14:15:21 +08:00
|
|
|
vq->dma_dev = dma_dev;
|
2018-11-21 18:03:27 +08:00
|
|
|
vq->use_dma_api = vring_use_dma_api(vdev);
|
2023-08-10 20:30:48 +08:00
|
|
|
vq->premapped = false;
|
2023-08-10 20:30:51 +08:00
|
|
|
vq->do_unmap = vq->use_dma_api;
|
2018-11-21 18:03:27 +08:00
|
|
|
|
|
|
|
vq->indirect = virtio_has_feature(vdev, VIRTIO_RING_F_INDIRECT_DESC) &&
|
|
|
|
!context;
|
|
|
|
vq->event = virtio_has_feature(vdev, VIRTIO_RING_F_EVENT_IDX);
|
|
|
|
|
2019-01-23 17:50:26 +08:00
|
|
|
if (virtio_has_feature(vdev, VIRTIO_F_ORDER_PLATFORM))
|
|
|
|
vq->weak_barriers = false;
|
|
|
|
|
2022-08-01 14:38:39 +08:00
|
|
|
err = vring_alloc_state_extra_packed(&vring_packed);
|
|
|
|
if (err)
|
|
|
|
goto err_state_extra;
|
2018-11-21 18:03:27 +08:00
|
|
|
|
2022-08-01 14:38:40 +08:00
|
|
|
virtqueue_vring_init_packed(&vring_packed, !!callback);
|
2018-11-21 18:03:27 +08:00
|
|
|
|
2022-08-01 14:38:26 +08:00
|
|
|
virtqueue_init(vq, num);
|
2022-08-01 14:38:41 +08:00
|
|
|
virtqueue_vring_attach_packed(vq, &vring_packed);
|
2022-08-01 14:38:26 +08:00
|
|
|
|
2021-07-21 22:26:47 +08:00
|
|
|
spin_lock(&vdev->vqs_list_lock);
|
2020-12-04 22:23:36 +08:00
|
|
|
list_add_tail(&vq->vq.list, &vdev->vqs);
|
2021-07-21 22:26:47 +08:00
|
|
|
spin_unlock(&vdev->vqs_list_lock);
|
2018-11-21 18:03:27 +08:00
|
|
|
return &vq->vq;
|
|
|
|
|
2022-08-01 14:38:39 +08:00
|
|
|
err_state_extra:
|
2018-11-21 18:03:27 +08:00
|
|
|
kfree(vq);
|
|
|
|
err_vq:
|
2023-01-19 14:15:21 +08:00
|
|
|
vring_free_packed(&vring_packed, vdev, dma_dev);
|
2018-11-21 18:03:27 +08:00
|
|
|
err_ring:
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2022-08-01 14:38:43 +08:00
|
|
|
static int virtqueue_resize_packed(struct virtqueue *_vq, u32 num)
|
|
|
|
{
|
|
|
|
struct vring_virtqueue_packed vring_packed = {};
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
struct virtio_device *vdev = _vq->vdev;
|
|
|
|
int err;
|
|
|
|
|
2023-01-19 14:15:21 +08:00
|
|
|
if (vring_alloc_queue_packed(&vring_packed, vdev, num, vring_dma_dev(vq)))
|
2022-08-01 14:38:43 +08:00
|
|
|
goto err_ring;
|
|
|
|
|
|
|
|
err = vring_alloc_state_extra_packed(&vring_packed);
|
|
|
|
if (err)
|
|
|
|
goto err_state_extra;
|
|
|
|
|
|
|
|
vring_free(&vq->vq);
|
|
|
|
|
|
|
|
virtqueue_vring_init_packed(&vring_packed, !!vq->vq.callback);
|
|
|
|
|
|
|
|
virtqueue_init(vq, vring_packed.vring.num);
|
|
|
|
virtqueue_vring_attach_packed(vq, &vring_packed);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
err_state_extra:
|
2023-01-19 14:15:21 +08:00
|
|
|
vring_free_packed(&vring_packed, vdev, vring_dma_dev(vq));
|
2022-08-01 14:38:43 +08:00
|
|
|
err_ring:
|
|
|
|
virtqueue_reinit_packed(vq);
|
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
|
2023-08-10 20:30:53 +08:00
|
|
|
static int virtqueue_disable_and_recycle(struct virtqueue *_vq,
|
|
|
|
void (*recycle)(struct virtqueue *vq, void *buf))
|
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
struct virtio_device *vdev = vq->vq.vdev;
|
|
|
|
void *buf;
|
|
|
|
int err;
|
|
|
|
|
|
|
|
if (!vq->we_own_ring)
|
|
|
|
return -EPERM;
|
|
|
|
|
|
|
|
if (!vdev->config->disable_vq_and_reset)
|
|
|
|
return -ENOENT;
|
|
|
|
|
|
|
|
if (!vdev->config->enable_vq_after_reset)
|
|
|
|
return -ENOENT;
|
|
|
|
|
|
|
|
err = vdev->config->disable_vq_and_reset(_vq);
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
|
|
|
|
while ((buf = virtqueue_detach_unused_buf(_vq)) != NULL)
|
|
|
|
recycle(_vq, buf);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int virtqueue_enable_after_reset(struct virtqueue *_vq)
|
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
struct virtio_device *vdev = vq->vq.vdev;
|
|
|
|
|
|
|
|
if (vdev->config->enable_vq_after_reset(_vq))
|
|
|
|
return -EBUSY;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2018-11-21 18:03:27 +08:00
|
|
|
|
2018-11-21 18:03:20 +08:00
|
|
|
/*
|
|
|
|
* Generic functions and exported symbols.
|
|
|
|
*/
|
|
|
|
|
|
|
|
static inline int virtqueue_add(struct virtqueue *_vq,
|
|
|
|
struct scatterlist *sgs[],
|
|
|
|
unsigned int total_sg,
|
|
|
|
unsigned int out_sgs,
|
|
|
|
unsigned int in_sgs,
|
|
|
|
void *data,
|
|
|
|
void *ctx,
|
|
|
|
gfp_t gfp)
|
|
|
|
{
|
2018-11-21 18:03:27 +08:00
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
|
|
|
|
return vq->packed_ring ? virtqueue_add_packed(_vq, sgs, total_sg,
|
|
|
|
out_sgs, in_sgs, data, ctx, gfp) :
|
|
|
|
virtqueue_add_split(_vq, sgs, total_sg,
|
|
|
|
out_sgs, in_sgs, data, ctx, gfp);
|
2018-11-21 18:03:20 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* virtqueue_add_sgs - expose buffers to other end
|
2019-04-23 18:25:12 +08:00
|
|
|
* @_vq: the struct virtqueue we're talking about.
|
2018-11-21 18:03:20 +08:00
|
|
|
* @sgs: array of terminated scatterlists.
|
2019-04-23 18:25:12 +08:00
|
|
|
* @out_sgs: the number of scatterlists readable by other side
|
|
|
|
* @in_sgs: the number of scatterlists which are writable (after readable ones)
|
2018-11-21 18:03:20 +08:00
|
|
|
* @data: the token identifying the buffer.
|
|
|
|
* @gfp: how to do memory allocations (if necessary).
|
|
|
|
*
|
|
|
|
* Caller must ensure we don't call this with other virtqueue operations
|
|
|
|
* at the same time (except where noted).
|
|
|
|
*
|
|
|
|
* Returns zero or a negative error (ie. ENOSPC, ENOMEM, EIO).
|
|
|
|
*/
|
|
|
|
int virtqueue_add_sgs(struct virtqueue *_vq,
|
|
|
|
struct scatterlist *sgs[],
|
|
|
|
unsigned int out_sgs,
|
|
|
|
unsigned int in_sgs,
|
|
|
|
void *data,
|
|
|
|
gfp_t gfp)
|
|
|
|
{
|
|
|
|
unsigned int i, total_sg = 0;
|
|
|
|
|
|
|
|
/* Count them first. */
|
|
|
|
for (i = 0; i < out_sgs + in_sgs; i++) {
|
|
|
|
struct scatterlist *sg;
|
|
|
|
|
|
|
|
for (sg = sgs[i]; sg; sg = sg_next(sg))
|
|
|
|
total_sg++;
|
|
|
|
}
|
|
|
|
return virtqueue_add(_vq, sgs, total_sg, out_sgs, in_sgs,
|
|
|
|
data, NULL, gfp);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_add_sgs);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* virtqueue_add_outbuf - expose output buffers to other end
|
|
|
|
* @vq: the struct virtqueue we're talking about.
|
|
|
|
* @sg: scatterlist (must be well-formed and terminated!)
|
|
|
|
* @num: the number of entries in @sg readable by other side
|
|
|
|
* @data: the token identifying the buffer.
|
|
|
|
* @gfp: how to do memory allocations (if necessary).
|
|
|
|
*
|
|
|
|
* Caller must ensure we don't call this with other virtqueue operations
|
|
|
|
* at the same time (except where noted).
|
|
|
|
*
|
|
|
|
* Returns zero or a negative error (ie. ENOSPC, ENOMEM, EIO).
|
|
|
|
*/
|
|
|
|
int virtqueue_add_outbuf(struct virtqueue *vq,
|
|
|
|
struct scatterlist *sg, unsigned int num,
|
|
|
|
void *data,
|
|
|
|
gfp_t gfp)
|
|
|
|
{
|
|
|
|
return virtqueue_add(vq, &sg, num, 1, 0, data, NULL, gfp);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_add_outbuf);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* virtqueue_add_inbuf - expose input buffers to other end
|
|
|
|
* @vq: the struct virtqueue we're talking about.
|
|
|
|
* @sg: scatterlist (must be well-formed and terminated!)
|
|
|
|
* @num: the number of entries in @sg writable by other side
|
|
|
|
* @data: the token identifying the buffer.
|
|
|
|
* @gfp: how to do memory allocations (if necessary).
|
|
|
|
*
|
|
|
|
* Caller must ensure we don't call this with other virtqueue operations
|
|
|
|
* at the same time (except where noted).
|
|
|
|
*
|
|
|
|
* Returns zero or a negative error (ie. ENOSPC, ENOMEM, EIO).
|
|
|
|
*/
|
|
|
|
int virtqueue_add_inbuf(struct virtqueue *vq,
|
|
|
|
struct scatterlist *sg, unsigned int num,
|
|
|
|
void *data,
|
|
|
|
gfp_t gfp)
|
|
|
|
{
|
|
|
|
return virtqueue_add(vq, &sg, num, 0, 1, data, NULL, gfp);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_add_inbuf);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* virtqueue_add_inbuf_ctx - expose input buffers to other end
|
|
|
|
* @vq: the struct virtqueue we're talking about.
|
|
|
|
* @sg: scatterlist (must be well-formed and terminated!)
|
|
|
|
* @num: the number of entries in @sg writable by other side
|
|
|
|
* @data: the token identifying the buffer.
|
|
|
|
* @ctx: extra context for the token
|
|
|
|
* @gfp: how to do memory allocations (if necessary).
|
|
|
|
*
|
|
|
|
* Caller must ensure we don't call this with other virtqueue operations
|
|
|
|
* at the same time (except where noted).
|
|
|
|
*
|
|
|
|
* Returns zero or a negative error (ie. ENOSPC, ENOMEM, EIO).
|
|
|
|
*/
|
|
|
|
int virtqueue_add_inbuf_ctx(struct virtqueue *vq,
|
|
|
|
struct scatterlist *sg, unsigned int num,
|
|
|
|
void *data,
|
|
|
|
void *ctx,
|
|
|
|
gfp_t gfp)
|
|
|
|
{
|
|
|
|
return virtqueue_add(vq, &sg, num, 0, 1, data, ctx, gfp);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_add_inbuf_ctx);
|
|
|
|
|
2023-08-10 20:30:50 +08:00
|
|
|
/**
|
|
|
|
* virtqueue_dma_dev - get the dma dev
|
|
|
|
* @_vq: the struct virtqueue we're talking about.
|
|
|
|
*
|
|
|
|
* Returns the dma dev. That can been used for dma api.
|
|
|
|
*/
|
|
|
|
struct device *virtqueue_dma_dev(struct virtqueue *_vq)
|
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
|
|
|
|
if (vq->use_dma_api)
|
|
|
|
return vring_dma_dev(vq);
|
|
|
|
else
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_dma_dev);
|
|
|
|
|
2018-11-21 18:03:20 +08:00
|
|
|
/**
|
|
|
|
* virtqueue_kick_prepare - first half of split virtqueue_kick call.
|
2019-04-23 18:25:12 +08:00
|
|
|
* @_vq: the struct virtqueue
|
2018-11-21 18:03:20 +08:00
|
|
|
*
|
|
|
|
* Instead of virtqueue_kick(), you can do:
|
|
|
|
* if (virtqueue_kick_prepare(vq))
|
|
|
|
* virtqueue_notify(vq);
|
|
|
|
*
|
|
|
|
* This is sometimes useful because the virtqueue_kick_prepare() needs
|
|
|
|
* to be serialized, but the actual virtqueue_notify() call does not.
|
|
|
|
*/
|
|
|
|
bool virtqueue_kick_prepare(struct virtqueue *_vq)
|
|
|
|
{
|
2018-11-21 18:03:27 +08:00
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
|
|
|
|
return vq->packed_ring ? virtqueue_kick_prepare_packed(_vq) :
|
|
|
|
virtqueue_kick_prepare_split(_vq);
|
2018-11-21 18:03:20 +08:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_kick_prepare);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* virtqueue_notify - second half of split virtqueue_kick call.
|
2019-04-23 18:25:12 +08:00
|
|
|
* @_vq: the struct virtqueue
|
2018-11-21 18:03:20 +08:00
|
|
|
*
|
|
|
|
* This does not need to be serialized.
|
|
|
|
*
|
|
|
|
* Returns false if host notify failed or queue is broken, otherwise true.
|
|
|
|
*/
|
|
|
|
bool virtqueue_notify(struct virtqueue *_vq)
|
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
|
|
|
|
if (unlikely(vq->broken))
|
|
|
|
return false;
|
|
|
|
|
|
|
|
/* Prod other side to tell it about changes. */
|
|
|
|
if (!vq->notify(_vq)) {
|
|
|
|
vq->broken = true;
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_notify);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* virtqueue_kick - update after add_buf
|
|
|
|
* @vq: the struct virtqueue
|
|
|
|
*
|
|
|
|
* After one or more virtqueue_add_* calls, invoke this to kick
|
|
|
|
* the other side.
|
|
|
|
*
|
|
|
|
* Caller must ensure we don't call this with other virtqueue
|
|
|
|
* operations at the same time (except where noted).
|
|
|
|
*
|
|
|
|
* Returns false if kick failed, otherwise true.
|
|
|
|
*/
|
|
|
|
bool virtqueue_kick(struct virtqueue *vq)
|
|
|
|
{
|
|
|
|
if (virtqueue_kick_prepare(vq))
|
|
|
|
return virtqueue_notify(vq);
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_kick);
|
|
|
|
|
|
|
|
/**
|
2021-05-26 11:12:11 +08:00
|
|
|
* virtqueue_get_buf_ctx - get the next used buffer
|
2019-04-23 18:25:12 +08:00
|
|
|
* @_vq: the struct virtqueue we're talking about.
|
2018-11-21 18:03:20 +08:00
|
|
|
* @len: the length written into the buffer
|
2019-04-23 18:25:12 +08:00
|
|
|
* @ctx: extra context for the token
|
2018-11-21 18:03:20 +08:00
|
|
|
*
|
|
|
|
* If the device wrote data into the buffer, @len will be set to the
|
|
|
|
* amount written. This means you don't need to clear the buffer
|
|
|
|
* beforehand to ensure there's no data leakage in the case of short
|
|
|
|
* writes.
|
|
|
|
*
|
|
|
|
* Caller must ensure we don't call this with other virtqueue
|
|
|
|
* operations at the same time (except where noted).
|
|
|
|
*
|
|
|
|
* Returns NULL if there are no used buffers, or the "data" token
|
|
|
|
* handed to virtqueue_add_*().
|
|
|
|
*/
|
|
|
|
void *virtqueue_get_buf_ctx(struct virtqueue *_vq, unsigned int *len,
|
|
|
|
void **ctx)
|
|
|
|
{
|
2018-11-21 18:03:27 +08:00
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
|
|
|
|
return vq->packed_ring ? virtqueue_get_buf_ctx_packed(_vq, len, ctx) :
|
|
|
|
virtqueue_get_buf_ctx_split(_vq, len, ctx);
|
2018-11-21 18:03:20 +08:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_get_buf_ctx);
|
|
|
|
|
|
|
|
void *virtqueue_get_buf(struct virtqueue *_vq, unsigned int *len)
|
|
|
|
{
|
|
|
|
return virtqueue_get_buf_ctx(_vq, len, NULL);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_get_buf);
|
|
|
|
/**
|
|
|
|
* virtqueue_disable_cb - disable callbacks
|
2019-04-23 18:25:12 +08:00
|
|
|
* @_vq: the struct virtqueue we're talking about.
|
2018-11-21 18:03:20 +08:00
|
|
|
*
|
|
|
|
* Note that this is not necessarily synchronous, hence unreliable and only
|
|
|
|
* useful as an optimization.
|
|
|
|
*
|
|
|
|
* Unlike other operations, this need not be serialized.
|
|
|
|
*/
|
|
|
|
void virtqueue_disable_cb(struct virtqueue *_vq)
|
|
|
|
{
|
2018-11-21 18:03:27 +08:00
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
|
|
|
|
if (vq->packed_ring)
|
|
|
|
virtqueue_disable_cb_packed(_vq);
|
|
|
|
else
|
|
|
|
virtqueue_disable_cb_split(_vq);
|
2018-11-21 18:03:20 +08:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_disable_cb);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* virtqueue_enable_cb_prepare - restart callbacks after disable_cb
|
2019-04-23 18:25:12 +08:00
|
|
|
* @_vq: the struct virtqueue we're talking about.
|
2018-11-21 18:03:20 +08:00
|
|
|
*
|
|
|
|
* This re-enables callbacks; it returns current queue state
|
|
|
|
* in an opaque unsigned value. This value should be later tested by
|
|
|
|
* virtqueue_poll, to detect a possible race between the driver checking for
|
|
|
|
* more work, and enabling callbacks.
|
|
|
|
*
|
|
|
|
* Caller must ensure we don't call this with other virtqueue
|
|
|
|
* operations at the same time (except where noted).
|
|
|
|
*/
|
2022-04-18 10:54:35 +08:00
|
|
|
unsigned int virtqueue_enable_cb_prepare(struct virtqueue *_vq)
|
2018-11-21 18:03:20 +08:00
|
|
|
{
|
2018-11-21 18:03:27 +08:00
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
|
2021-04-13 13:19:16 +08:00
|
|
|
if (vq->event_triggered)
|
|
|
|
vq->event_triggered = false;
|
|
|
|
|
2018-11-21 18:03:27 +08:00
|
|
|
return vq->packed_ring ? virtqueue_enable_cb_prepare_packed(_vq) :
|
|
|
|
virtqueue_enable_cb_prepare_split(_vq);
|
2018-11-21 18:03:20 +08:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_enable_cb_prepare);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* virtqueue_poll - query pending used buffers
|
2019-04-23 18:25:12 +08:00
|
|
|
* @_vq: the struct virtqueue we're talking about.
|
2018-11-21 18:03:20 +08:00
|
|
|
* @last_used_idx: virtqueue state (from call to virtqueue_enable_cb_prepare).
|
|
|
|
*
|
|
|
|
* Returns "true" if there are pending used buffers in the queue.
|
|
|
|
*
|
|
|
|
* This does not need to be serialized.
|
|
|
|
*/
|
2022-04-18 10:54:35 +08:00
|
|
|
bool virtqueue_poll(struct virtqueue *_vq, unsigned int last_used_idx)
|
2018-11-21 18:03:20 +08:00
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
|
2020-08-02 15:44:09 +08:00
|
|
|
if (unlikely(vq->broken))
|
|
|
|
return false;
|
|
|
|
|
2018-11-21 18:03:20 +08:00
|
|
|
virtio_mb(vq->weak_barriers);
|
2018-11-21 18:03:27 +08:00
|
|
|
return vq->packed_ring ? virtqueue_poll_packed(_vq, last_used_idx) :
|
|
|
|
virtqueue_poll_split(_vq, last_used_idx);
|
2018-11-21 18:03:20 +08:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_poll);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* virtqueue_enable_cb - restart callbacks after disable_cb.
|
2019-04-23 18:25:12 +08:00
|
|
|
* @_vq: the struct virtqueue we're talking about.
|
2018-11-21 18:03:20 +08:00
|
|
|
*
|
|
|
|
* This re-enables callbacks; it returns "false" if there are pending
|
|
|
|
* buffers in the queue, to detect a possible race between the driver
|
|
|
|
* checking for more work, and enabling callbacks.
|
|
|
|
*
|
|
|
|
* Caller must ensure we don't call this with other virtqueue
|
|
|
|
* operations at the same time (except where noted).
|
|
|
|
*/
|
|
|
|
bool virtqueue_enable_cb(struct virtqueue *_vq)
|
|
|
|
{
|
2022-04-18 10:54:35 +08:00
|
|
|
unsigned int last_used_idx = virtqueue_enable_cb_prepare(_vq);
|
2018-11-21 18:03:20 +08:00
|
|
|
|
|
|
|
return !virtqueue_poll(_vq, last_used_idx);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_enable_cb);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* virtqueue_enable_cb_delayed - restart callbacks after disable_cb.
|
2019-04-23 18:25:12 +08:00
|
|
|
* @_vq: the struct virtqueue we're talking about.
|
2018-11-21 18:03:20 +08:00
|
|
|
*
|
|
|
|
* This re-enables callbacks but hints to the other side to delay
|
|
|
|
* interrupts until most of the available buffers have been processed;
|
|
|
|
* it returns "false" if there are many pending buffers in the queue,
|
|
|
|
* to detect a possible race between the driver checking for more work,
|
|
|
|
* and enabling callbacks.
|
|
|
|
*
|
|
|
|
* Caller must ensure we don't call this with other virtqueue
|
|
|
|
* operations at the same time (except where noted).
|
|
|
|
*/
|
|
|
|
bool virtqueue_enable_cb_delayed(struct virtqueue *_vq)
|
|
|
|
{
|
2018-11-21 18:03:27 +08:00
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
|
2021-04-13 13:19:16 +08:00
|
|
|
if (vq->event_triggered)
|
|
|
|
vq->event_triggered = false;
|
|
|
|
|
2018-11-21 18:03:27 +08:00
|
|
|
return vq->packed_ring ? virtqueue_enable_cb_delayed_packed(_vq) :
|
|
|
|
virtqueue_enable_cb_delayed_split(_vq);
|
2018-11-21 18:03:20 +08:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_enable_cb_delayed);
|
|
|
|
|
2018-11-21 18:03:19 +08:00
|
|
|
/**
|
|
|
|
* virtqueue_detach_unused_buf - detach first unused buffer
|
2019-04-23 18:25:12 +08:00
|
|
|
* @_vq: the struct virtqueue we're talking about.
|
2018-11-21 18:03:19 +08:00
|
|
|
*
|
|
|
|
* Returns NULL or the "data" token handed to virtqueue_add_*().
|
2022-08-01 14:38:23 +08:00
|
|
|
* This is not valid on an active queue; it is useful for device
|
|
|
|
* shutdown or the reset queue.
|
2018-11-21 18:03:19 +08:00
|
|
|
*/
|
|
|
|
void *virtqueue_detach_unused_buf(struct virtqueue *_vq)
|
|
|
|
{
|
2018-11-21 18:03:27 +08:00
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
|
|
|
|
return vq->packed_ring ? virtqueue_detach_unused_buf_packed(_vq) :
|
|
|
|
virtqueue_detach_unused_buf_split(_vq);
|
2018-11-21 18:03:19 +08:00
|
|
|
}
|
2010-04-12 21:19:07 +08:00
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_detach_unused_buf);
|
2010-01-18 21:45:23 +08:00
|
|
|
|
2018-11-21 18:03:19 +08:00
|
|
|
static inline bool more_used(const struct vring_virtqueue *vq)
|
|
|
|
{
|
2018-11-21 18:03:27 +08:00
|
|
|
return vq->packed_ring ? more_used_packed(vq) : more_used_split(vq);
|
2018-11-21 18:03:19 +08:00
|
|
|
}
|
|
|
|
|
2022-08-10 17:40:03 +08:00
|
|
|
/**
|
|
|
|
* vring_interrupt - notify a virtqueue on an interrupt
|
|
|
|
* @irq: the IRQ number (ignored)
|
|
|
|
* @_vq: the struct virtqueue to notify
|
|
|
|
*
|
|
|
|
* Calls the callback function of @_vq to process the virtqueue
|
|
|
|
* notification.
|
|
|
|
*/
|
2007-10-22 09:03:40 +08:00
|
|
|
irqreturn_t vring_interrupt(int irq, void *_vq)
|
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
|
|
|
|
if (!more_used(vq)) {
|
|
|
|
pr_debug("virtqueue interrupt with no work for %p\n", vq);
|
|
|
|
return IRQ_NONE;
|
|
|
|
}
|
|
|
|
|
2022-05-27 14:01:19 +08:00
|
|
|
if (unlikely(vq->broken)) {
|
2022-06-22 09:29:40 +08:00
|
|
|
#ifdef CONFIG_VIRTIO_HARDEN_NOTIFICATION
|
2022-05-27 14:01:19 +08:00
|
|
|
dev_warn_once(&vq->vq.vdev->dev,
|
|
|
|
"virtio vring IRQ raised before DRIVER_OK");
|
|
|
|
return IRQ_NONE;
|
2022-06-22 09:29:40 +08:00
|
|
|
#else
|
|
|
|
return IRQ_HANDLED;
|
|
|
|
#endif
|
2022-05-27 14:01:19 +08:00
|
|
|
}
|
2007-10-22 09:03:40 +08:00
|
|
|
|
2021-04-13 13:19:16 +08:00
|
|
|
/* Just a hint for performance: so it's ok that this can be racy! */
|
|
|
|
if (vq->event)
|
|
|
|
vq->event_triggered = true;
|
|
|
|
|
2007-10-22 09:03:40 +08:00
|
|
|
pr_debug("virtqueue callback for %p (%p)\n", vq, vq->vq.callback);
|
2008-02-05 12:49:57 +08:00
|
|
|
if (vq->vq.callback)
|
|
|
|
vq->vq.callback(&vq->vq);
|
2007-10-22 09:03:40 +08:00
|
|
|
|
|
|
|
return IRQ_HANDLED;
|
|
|
|
}
|
2008-02-05 12:50:05 +08:00
|
|
|
EXPORT_SYMBOL_GPL(vring_interrupt);
|
2007-10-22 09:03:40 +08:00
|
|
|
|
2018-11-21 18:03:27 +08:00
|
|
|
/* Only available for split ring */
|
2022-08-01 14:38:27 +08:00
|
|
|
static struct virtqueue *__vring_new_virtqueue(unsigned int index,
|
2022-08-01 14:38:28 +08:00
|
|
|
struct vring_virtqueue_split *vring_split,
|
2022-08-01 14:38:27 +08:00
|
|
|
struct virtio_device *vdev,
|
|
|
|
bool weak_barriers,
|
|
|
|
bool context,
|
|
|
|
bool (*notify)(struct virtqueue *),
|
|
|
|
void (*callback)(struct virtqueue *),
|
2023-01-19 14:15:21 +08:00
|
|
|
const char *name,
|
|
|
|
struct device *dma_dev)
|
2007-10-22 09:03:40 +08:00
|
|
|
{
|
2016-02-03 13:46:37 +08:00
|
|
|
struct vring_virtqueue *vq;
|
2022-08-01 14:38:31 +08:00
|
|
|
int err;
|
2007-10-22 09:03:40 +08:00
|
|
|
|
2018-11-21 18:03:27 +08:00
|
|
|
if (virtio_has_feature(vdev, VIRTIO_F_RING_PACKED))
|
|
|
|
return NULL;
|
|
|
|
|
2018-11-21 18:03:24 +08:00
|
|
|
vq = kmalloc(sizeof(*vq), GFP_KERNEL);
|
2007-10-22 09:03:40 +08:00
|
|
|
if (!vq)
|
|
|
|
return NULL;
|
|
|
|
|
2018-11-21 18:03:27 +08:00
|
|
|
vq->packed_ring = false;
|
2007-10-22 09:03:40 +08:00
|
|
|
vq->vq.callback = callback;
|
|
|
|
vq->vq.vdev = vdev;
|
2009-06-13 12:16:35 +08:00
|
|
|
vq->vq.name = name;
|
2012-10-16 21:26:14 +08:00
|
|
|
vq->vq.index = index;
|
2022-08-01 14:38:48 +08:00
|
|
|
vq->vq.reset = false;
|
2016-02-03 13:46:37 +08:00
|
|
|
vq->we_own_ring = false;
|
2007-10-22 09:03:40 +08:00
|
|
|
vq->notify = notify;
|
2012-01-12 13:14:42 +08:00
|
|
|
vq->weak_barriers = weak_barriers;
|
2022-06-22 09:29:40 +08:00
|
|
|
#ifdef CONFIG_VIRTIO_HARDEN_NOTIFICATION
|
2022-05-27 14:01:19 +08:00
|
|
|
vq->broken = true;
|
2022-06-22 09:29:40 +08:00
|
|
|
#else
|
|
|
|
vq->broken = false;
|
|
|
|
#endif
|
2023-01-19 14:15:21 +08:00
|
|
|
vq->dma_dev = dma_dev;
|
2018-11-21 18:03:26 +08:00
|
|
|
vq->use_dma_api = vring_use_dma_api(vdev);
|
2023-08-10 20:30:48 +08:00
|
|
|
vq->premapped = false;
|
2023-08-10 20:30:51 +08:00
|
|
|
vq->do_unmap = vq->use_dma_api;
|
2007-10-22 09:03:40 +08:00
|
|
|
|
2017-02-07 12:15:13 +08:00
|
|
|
vq->indirect = virtio_has_feature(vdev, VIRTIO_RING_F_INDIRECT_DESC) &&
|
|
|
|
!context;
|
2011-05-20 07:10:44 +08:00
|
|
|
vq->event = virtio_has_feature(vdev, VIRTIO_RING_F_EVENT_IDX);
|
2009-05-12 01:11:45 +08:00
|
|
|
|
2019-01-23 17:50:26 +08:00
|
|
|
if (virtio_has_feature(vdev, VIRTIO_F_ORDER_PLATFORM))
|
|
|
|
vq->weak_barriers = false;
|
|
|
|
|
2022-08-01 14:38:31 +08:00
|
|
|
err = vring_alloc_state_extra_split(vring_split);
|
|
|
|
if (err) {
|
|
|
|
kfree(vq);
|
|
|
|
return NULL;
|
|
|
|
}
|
2021-06-04 13:53:50 +08:00
|
|
|
|
2022-08-01 14:38:32 +08:00
|
|
|
virtqueue_vring_init_split(vring_split, vq);
|
|
|
|
|
2022-08-01 14:38:28 +08:00
|
|
|
virtqueue_init(vq, vring_split->vring.num);
|
2022-08-01 14:38:33 +08:00
|
|
|
virtqueue_vring_attach_split(vq, vring_split);
|
2022-08-01 14:38:26 +08:00
|
|
|
|
2021-07-21 22:26:47 +08:00
|
|
|
spin_lock(&vdev->vqs_list_lock);
|
2020-12-04 22:23:36 +08:00
|
|
|
list_add_tail(&vq->vq.list, &vdev->vqs);
|
2021-07-21 22:26:47 +08:00
|
|
|
spin_unlock(&vdev->vqs_list_lock);
|
2007-10-22 09:03:40 +08:00
|
|
|
return &vq->vq;
|
|
|
|
}
|
2016-02-03 13:46:37 +08:00
|
|
|
|
|
|
|
struct virtqueue *vring_create_virtqueue(
|
|
|
|
unsigned int index,
|
|
|
|
unsigned int num,
|
|
|
|
unsigned int vring_align,
|
|
|
|
struct virtio_device *vdev,
|
|
|
|
bool weak_barriers,
|
|
|
|
bool may_reduce_num,
|
2017-03-07 00:32:29 +08:00
|
|
|
bool context,
|
2016-02-03 13:46:37 +08:00
|
|
|
bool (*notify)(struct virtqueue *),
|
|
|
|
void (*callback)(struct virtqueue *),
|
|
|
|
const char *name)
|
|
|
|
{
|
2018-11-21 18:03:27 +08:00
|
|
|
|
|
|
|
if (virtio_has_feature(vdev, VIRTIO_F_RING_PACKED))
|
|
|
|
return vring_create_virtqueue_packed(index, num, vring_align,
|
|
|
|
vdev, weak_barriers, may_reduce_num,
|
2023-01-19 14:15:21 +08:00
|
|
|
context, notify, callback, name, vdev->dev.parent);
|
2018-11-21 18:03:27 +08:00
|
|
|
|
2018-11-21 18:03:25 +08:00
|
|
|
return vring_create_virtqueue_split(index, num, vring_align,
|
|
|
|
vdev, weak_barriers, may_reduce_num,
|
2023-01-19 14:15:21 +08:00
|
|
|
context, notify, callback, name, vdev->dev.parent);
|
2016-02-03 13:46:37 +08:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(vring_create_virtqueue);
|
|
|
|
|
2023-01-19 14:15:21 +08:00
|
|
|
struct virtqueue *vring_create_virtqueue_dma(
|
|
|
|
unsigned int index,
|
|
|
|
unsigned int num,
|
|
|
|
unsigned int vring_align,
|
|
|
|
struct virtio_device *vdev,
|
|
|
|
bool weak_barriers,
|
|
|
|
bool may_reduce_num,
|
|
|
|
bool context,
|
|
|
|
bool (*notify)(struct virtqueue *),
|
|
|
|
void (*callback)(struct virtqueue *),
|
|
|
|
const char *name,
|
|
|
|
struct device *dma_dev)
|
|
|
|
{
|
|
|
|
|
|
|
|
if (virtio_has_feature(vdev, VIRTIO_F_RING_PACKED))
|
|
|
|
return vring_create_virtqueue_packed(index, num, vring_align,
|
|
|
|
vdev, weak_barriers, may_reduce_num,
|
|
|
|
context, notify, callback, name, dma_dev);
|
|
|
|
|
|
|
|
return vring_create_virtqueue_split(index, num, vring_align,
|
|
|
|
vdev, weak_barriers, may_reduce_num,
|
|
|
|
context, notify, callback, name, dma_dev);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(vring_create_virtqueue_dma);
|
|
|
|
|
2022-08-01 14:38:44 +08:00
|
|
|
/**
|
|
|
|
* virtqueue_resize - resize the vring of vq
|
|
|
|
* @_vq: the struct virtqueue we're talking about.
|
|
|
|
* @num: new ring num
|
2023-08-10 20:30:52 +08:00
|
|
|
* @recycle: callback to recycle unused buffers
|
2022-08-01 14:38:44 +08:00
|
|
|
*
|
|
|
|
* When it is really necessary to create a new vring, it will set the current vq
|
|
|
|
* into the reset state. Then call the passed callback to recycle the buffer
|
|
|
|
* that is no longer used. Only after the new vring is successfully created, the
|
|
|
|
* old vring will be released.
|
|
|
|
*
|
|
|
|
* Caller must ensure we don't call this with other virtqueue operations
|
|
|
|
* at the same time (except where noted).
|
|
|
|
*
|
|
|
|
* Returns zero or a negative error.
|
|
|
|
* 0: success.
|
|
|
|
* -ENOMEM: Failed to allocate a new ring, fall back to the original ring size.
|
|
|
|
* vq can still work normally
|
|
|
|
* -EBUSY: Failed to sync with device, vq may not work properly
|
|
|
|
* -ENOENT: Transport or device not supported
|
|
|
|
* -E2BIG/-EINVAL: num error
|
|
|
|
* -EPERM: Operation not permitted
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
int virtqueue_resize(struct virtqueue *_vq, u32 num,
|
|
|
|
void (*recycle)(struct virtqueue *vq, void *buf))
|
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
int err;
|
|
|
|
|
|
|
|
if (num > vq->vq.num_max)
|
|
|
|
return -E2BIG;
|
|
|
|
|
|
|
|
if (!num)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
if ((vq->packed_ring ? vq->packed.vring.num : vq->split.vring.num) == num)
|
|
|
|
return 0;
|
|
|
|
|
2023-08-10 20:30:53 +08:00
|
|
|
err = virtqueue_disable_and_recycle(_vq, recycle);
|
2022-08-01 14:38:44 +08:00
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
|
|
|
|
if (vq->packed_ring)
|
|
|
|
err = virtqueue_resize_packed(_vq, num);
|
|
|
|
else
|
|
|
|
err = virtqueue_resize_split(_vq, num);
|
|
|
|
|
2023-08-10 20:30:53 +08:00
|
|
|
return virtqueue_enable_after_reset(_vq);
|
2022-08-01 14:38:44 +08:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_resize);
|
|
|
|
|
2023-08-10 20:30:48 +08:00
|
|
|
/**
|
|
|
|
* virtqueue_set_dma_premapped - set the vring premapped mode
|
|
|
|
* @_vq: the struct virtqueue we're talking about.
|
|
|
|
*
|
|
|
|
* Enable the premapped mode of the vq.
|
|
|
|
*
|
|
|
|
* The vring in premapped mode does not do dma internally, so the driver must
|
|
|
|
* do dma mapping in advance. The driver must pass the dma_address through
|
|
|
|
* dma_address of scatterlist. When the driver got a used buffer from
|
|
|
|
* the vring, it has to unmap the dma address.
|
|
|
|
*
|
|
|
|
* This function must be called immediately after creating the vq, or after vq
|
|
|
|
* reset, and before adding any buffers to it.
|
|
|
|
*
|
|
|
|
* Caller must ensure we don't call this with other virtqueue operations
|
|
|
|
* at the same time (except where noted).
|
|
|
|
*
|
|
|
|
* Returns zero or a negative error.
|
|
|
|
* 0: success.
|
|
|
|
* -EINVAL: vring does not use the dma api, so we can not enable premapped mode.
|
|
|
|
*/
|
|
|
|
int virtqueue_set_dma_premapped(struct virtqueue *_vq)
|
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
u32 num;
|
|
|
|
|
|
|
|
START_USE(vq);
|
|
|
|
|
|
|
|
num = vq->packed_ring ? vq->packed.vring.num : vq->split.vring.num;
|
|
|
|
|
|
|
|
if (num != vq->vq.num_free) {
|
|
|
|
END_USE(vq);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!vq->use_dma_api) {
|
|
|
|
END_USE(vq);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
vq->premapped = true;
|
2023-08-10 20:30:51 +08:00
|
|
|
vq->do_unmap = false;
|
2023-08-10 20:30:48 +08:00
|
|
|
|
|
|
|
END_USE(vq);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_set_dma_premapped);
|
|
|
|
|
2023-08-10 20:30:54 +08:00
|
|
|
/**
|
|
|
|
* virtqueue_reset - detach and recycle all unused buffers
|
|
|
|
* @_vq: the struct virtqueue we're talking about.
|
|
|
|
* @recycle: callback to recycle unused buffers
|
|
|
|
*
|
|
|
|
* Caller must ensure we don't call this with other virtqueue operations
|
|
|
|
* at the same time (except where noted).
|
|
|
|
*
|
|
|
|
* Returns zero or a negative error.
|
|
|
|
* 0: success.
|
|
|
|
* -EBUSY: Failed to sync with device, vq may not work properly
|
|
|
|
* -ENOENT: Transport or device not supported
|
|
|
|
* -EPERM: Operation not permitted
|
|
|
|
*/
|
|
|
|
int virtqueue_reset(struct virtqueue *_vq,
|
|
|
|
void (*recycle)(struct virtqueue *vq, void *buf))
|
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
int err;
|
|
|
|
|
|
|
|
err = virtqueue_disable_and_recycle(_vq, recycle);
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
|
|
|
|
if (vq->packed_ring)
|
|
|
|
virtqueue_reinit_packed(vq);
|
|
|
|
else
|
|
|
|
virtqueue_reinit_split(vq);
|
|
|
|
|
|
|
|
return virtqueue_enable_after_reset(_vq);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_reset);
|
|
|
|
|
2018-11-21 18:03:27 +08:00
|
|
|
/* Only available for split ring */
|
2016-02-03 13:46:37 +08:00
|
|
|
struct virtqueue *vring_new_virtqueue(unsigned int index,
|
|
|
|
unsigned int num,
|
|
|
|
unsigned int vring_align,
|
|
|
|
struct virtio_device *vdev,
|
|
|
|
bool weak_barriers,
|
2017-03-07 00:32:29 +08:00
|
|
|
bool context,
|
2016-02-03 13:46:37 +08:00
|
|
|
void *pages,
|
|
|
|
bool (*notify)(struct virtqueue *vq),
|
|
|
|
void (*callback)(struct virtqueue *vq),
|
|
|
|
const char *name)
|
|
|
|
{
|
2022-08-01 14:38:28 +08:00
|
|
|
struct vring_virtqueue_split vring_split = {};
|
2018-11-21 18:03:27 +08:00
|
|
|
|
|
|
|
if (virtio_has_feature(vdev, VIRTIO_F_RING_PACKED))
|
|
|
|
return NULL;
|
|
|
|
|
2022-08-01 14:38:28 +08:00
|
|
|
vring_init(&vring_split.vring, num, pages, vring_align);
|
|
|
|
return __vring_new_virtqueue(index, &vring_split, vdev, weak_barriers,
|
2023-01-19 14:15:21 +08:00
|
|
|
context, notify, callback, name,
|
|
|
|
vdev->dev.parent);
|
2016-02-03 13:46:37 +08:00
|
|
|
}
|
2008-02-05 12:50:05 +08:00
|
|
|
EXPORT_SYMBOL_GPL(vring_new_virtqueue);
|
2007-10-22 09:03:40 +08:00
|
|
|
|
2022-08-01 14:38:24 +08:00
|
|
|
static void vring_free(struct virtqueue *_vq)
|
2007-10-22 09:03:40 +08:00
|
|
|
{
|
2016-02-03 13:46:37 +08:00
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
|
|
|
|
if (vq->we_own_ring) {
|
2018-11-21 18:03:27 +08:00
|
|
|
if (vq->packed_ring) {
|
|
|
|
vring_free_queue(vq->vq.vdev,
|
|
|
|
vq->packed.ring_size_in_bytes,
|
|
|
|
vq->packed.vring.desc,
|
2023-01-19 14:15:21 +08:00
|
|
|
vq->packed.ring_dma_addr,
|
|
|
|
vring_dma_dev(vq));
|
2018-11-21 18:03:27 +08:00
|
|
|
|
|
|
|
vring_free_queue(vq->vq.vdev,
|
|
|
|
vq->packed.event_size_in_bytes,
|
|
|
|
vq->packed.vring.driver,
|
2023-01-19 14:15:21 +08:00
|
|
|
vq->packed.driver_event_dma_addr,
|
|
|
|
vring_dma_dev(vq));
|
2018-11-21 18:03:27 +08:00
|
|
|
|
|
|
|
vring_free_queue(vq->vq.vdev,
|
|
|
|
vq->packed.event_size_in_bytes,
|
|
|
|
vq->packed.vring.device,
|
2023-01-19 14:15:21 +08:00
|
|
|
vq->packed.device_event_dma_addr,
|
|
|
|
vring_dma_dev(vq));
|
2018-11-21 18:03:27 +08:00
|
|
|
|
|
|
|
kfree(vq->packed.desc_state);
|
|
|
|
kfree(vq->packed.desc_extra);
|
|
|
|
} else {
|
|
|
|
vring_free_queue(vq->vq.vdev,
|
|
|
|
vq->split.queue_size_in_bytes,
|
|
|
|
vq->split.vring.desc,
|
2023-01-19 14:15:21 +08:00
|
|
|
vq->split.queue_dma_addr,
|
|
|
|
vring_dma_dev(vq));
|
2018-11-21 18:03:27 +08:00
|
|
|
}
|
2016-02-03 13:46:37 +08:00
|
|
|
}
|
2021-06-04 13:53:50 +08:00
|
|
|
if (!vq->packed_ring) {
|
2020-02-25 05:26:43 +08:00
|
|
|
kfree(vq->split.desc_state);
|
2021-06-04 13:53:50 +08:00
|
|
|
kfree(vq->split.desc_extra);
|
|
|
|
}
|
2022-08-01 14:38:24 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
void vring_del_virtqueue(struct virtqueue *_vq)
|
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
|
|
|
|
spin_lock(&vq->vq.vdev->vqs_list_lock);
|
|
|
|
list_del(&_vq->list);
|
|
|
|
spin_unlock(&vq->vq.vdev->vqs_list_lock);
|
|
|
|
|
|
|
|
vring_free(_vq);
|
|
|
|
|
2016-02-03 13:46:37 +08:00
|
|
|
kfree(vq);
|
2007-10-22 09:03:40 +08:00
|
|
|
}
|
2008-02-05 12:50:05 +08:00
|
|
|
EXPORT_SYMBOL_GPL(vring_del_virtqueue);
|
2007-10-22 09:03:40 +08:00
|
|
|
|
2023-04-13 16:18:54 +08:00
|
|
|
u32 vring_notification_data(struct virtqueue *_vq)
|
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
u16 next;
|
|
|
|
|
|
|
|
if (vq->packed_ring)
|
|
|
|
next = (vq->packed.next_avail_idx &
|
|
|
|
~(-(1 << VRING_PACKED_EVENT_F_WRAP_CTR))) |
|
|
|
|
vq->packed.avail_wrap_counter <<
|
|
|
|
VRING_PACKED_EVENT_F_WRAP_CTR;
|
|
|
|
else
|
|
|
|
next = vq->split.avail_idx_shadow;
|
|
|
|
|
|
|
|
return next << 16 | _vq->index;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(vring_notification_data);
|
|
|
|
|
2008-07-26 01:06:13 +08:00
|
|
|
/* Manipulates transport-specific feature bits. */
|
|
|
|
void vring_transport_features(struct virtio_device *vdev)
|
|
|
|
{
|
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
for (i = VIRTIO_TRANSPORT_F_START; i < VIRTIO_TRANSPORT_F_END; i++) {
|
|
|
|
switch (i) {
|
2009-05-12 01:11:45 +08:00
|
|
|
case VIRTIO_RING_F_INDIRECT_DESC:
|
|
|
|
break;
|
2011-05-20 07:10:44 +08:00
|
|
|
case VIRTIO_RING_F_EVENT_IDX:
|
|
|
|
break;
|
2014-12-01 21:52:40 +08:00
|
|
|
case VIRTIO_F_VERSION_1:
|
|
|
|
break;
|
2020-06-25 06:24:33 +08:00
|
|
|
case VIRTIO_F_ACCESS_PLATFORM:
|
virtio: new feature to detect IOMMU device quirk
The interaction between virtio and IOMMUs is messy.
On most systems with virtio, physical addresses match bus addresses,
and it doesn't particularly matter which one we use to program
the device.
On some systems, including Xen and any system with a physical device
that speaks virtio behind a physical IOMMU, we must program the IOMMU
for virtio DMA to work at all.
On other systems, including SPARC and PPC64, virtio-pci devices are
enumerated as though they are behind an IOMMU, but the virtio host
ignores the IOMMU, so we must either pretend that the IOMMU isn't
there or somehow map everything as the identity.
Add a feature bit to detect that quirk: VIRTIO_F_IOMMU_PLATFORM.
Any device with this feature bit set to 0 needs a quirk and has to be
passed physical addresses (as opposed to bus addresses) even though
the device is behind an IOMMU.
Note: it has to be a per-device quirk because for example, there could
be a mix of passed-through and virtual virtio devices. As another
example, some devices could be implemented by an out of process
hypervisor backend (in case of qemu vhost, or vhost-user) and so support
for an IOMMU needs to be coded up separately.
It would be cleanest to handle this in IOMMU core code, but that needs
per-device DMA ops. While we are waiting for that to be implemented, use
a work-around in virtio core.
Note: a "noiommu" feature is a quirk - add a wrapper to make
that clear.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2016-04-18 17:58:14 +08:00
|
|
|
break;
|
2018-11-21 18:03:30 +08:00
|
|
|
case VIRTIO_F_RING_PACKED:
|
|
|
|
break;
|
2019-01-23 17:50:26 +08:00
|
|
|
case VIRTIO_F_ORDER_PLATFORM:
|
|
|
|
break;
|
2023-04-13 16:18:54 +08:00
|
|
|
case VIRTIO_F_NOTIFICATION_DATA:
|
|
|
|
break;
|
2008-07-26 01:06:13 +08:00
|
|
|
default:
|
|
|
|
/* We don't understand this bit. */
|
2014-10-07 22:39:42 +08:00
|
|
|
__virtio_clear_bit(vdev, i);
|
2008-07-26 01:06:13 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(vring_transport_features);
|
|
|
|
|
2012-01-12 13:14:42 +08:00
|
|
|
/**
|
|
|
|
* virtqueue_get_vring_size - return the size of the virtqueue's vring
|
2019-04-23 18:25:12 +08:00
|
|
|
* @_vq: the struct virtqueue containing the vring of interest.
|
2012-01-12 13:14:42 +08:00
|
|
|
*
|
|
|
|
* Returns the size of the vring. This is mainly used for boasting to
|
|
|
|
* userspace. Unlike other operations, this need not be serialized.
|
|
|
|
*/
|
2023-03-10 13:34:28 +08:00
|
|
|
unsigned int virtqueue_get_vring_size(const struct virtqueue *_vq)
|
2011-10-19 16:10:59 +08:00
|
|
|
{
|
|
|
|
|
2023-03-10 13:34:28 +08:00
|
|
|
const struct vring_virtqueue *vq = to_vvq(_vq);
|
2011-10-19 16:10:59 +08:00
|
|
|
|
2018-11-21 18:03:27 +08:00
|
|
|
return vq->packed_ring ? vq->packed.vring.num : vq->split.vring.num;
|
2011-10-19 16:10:59 +08:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_get_vring_size);
|
|
|
|
|
2022-08-01 14:38:46 +08:00
|
|
|
/*
|
|
|
|
* This function should only be called by the core, not directly by the driver.
|
|
|
|
*/
|
|
|
|
void __virtqueue_break(struct virtqueue *_vq)
|
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
|
|
|
|
/* Pairs with READ_ONCE() in virtqueue_is_broken(). */
|
|
|
|
WRITE_ONCE(vq->broken, true);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(__virtqueue_break);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* This function should only be called by the core, not directly by the driver.
|
|
|
|
*/
|
|
|
|
void __virtqueue_unbreak(struct virtqueue *_vq)
|
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
|
|
|
|
/* Pairs with READ_ONCE() in virtqueue_is_broken(). */
|
|
|
|
WRITE_ONCE(vq->broken, false);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(__virtqueue_unbreak);
|
|
|
|
|
2023-03-10 13:34:28 +08:00
|
|
|
bool virtqueue_is_broken(const struct virtqueue *_vq)
|
2013-10-29 07:10:19 +08:00
|
|
|
{
|
2023-03-10 13:34:28 +08:00
|
|
|
const struct vring_virtqueue *vq = to_vvq(_vq);
|
2013-10-29 07:10:19 +08:00
|
|
|
|
2021-07-21 22:26:45 +08:00
|
|
|
return READ_ONCE(vq->broken);
|
2013-10-29 07:10:19 +08:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_is_broken);
|
|
|
|
|
2014-04-28 09:45:08 +08:00
|
|
|
/*
|
|
|
|
* This should prevent the device from being used, allowing drivers to
|
|
|
|
* recover. You may need to grab appropriate locks to flush.
|
|
|
|
*/
|
|
|
|
void virtio_break_device(struct virtio_device *dev)
|
|
|
|
{
|
|
|
|
struct virtqueue *_vq;
|
|
|
|
|
2021-07-21 22:26:47 +08:00
|
|
|
spin_lock(&dev->vqs_list_lock);
|
2014-04-28 09:45:08 +08:00
|
|
|
list_for_each_entry(_vq, &dev->vqs, list) {
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
2021-07-21 22:26:45 +08:00
|
|
|
|
|
|
|
/* Pairs with READ_ONCE() in virtqueue_is_broken(). */
|
|
|
|
WRITE_ONCE(vq->broken, true);
|
2014-04-28 09:45:08 +08:00
|
|
|
}
|
2021-07-21 22:26:47 +08:00
|
|
|
spin_unlock(&dev->vqs_list_lock);
|
2014-04-28 09:45:08 +08:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtio_break_device);
|
|
|
|
|
2022-05-27 14:01:18 +08:00
|
|
|
/*
|
|
|
|
* This should allow the device to be used by the driver. You may
|
|
|
|
* need to grab appropriate locks to flush the write to
|
|
|
|
* vq->broken. This should only be used in some specific case e.g
|
|
|
|
* (probing and restoring). This function should only be called by the
|
|
|
|
* core, not directly by the driver.
|
|
|
|
*/
|
|
|
|
void __virtio_unbreak_device(struct virtio_device *dev)
|
|
|
|
{
|
|
|
|
struct virtqueue *_vq;
|
|
|
|
|
|
|
|
spin_lock(&dev->vqs_list_lock);
|
|
|
|
list_for_each_entry(_vq, &dev->vqs, list) {
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
|
|
|
|
/* Pairs with READ_ONCE() in virtqueue_is_broken(). */
|
|
|
|
WRITE_ONCE(vq->broken, false);
|
|
|
|
}
|
|
|
|
spin_unlock(&dev->vqs_list_lock);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(__virtio_unbreak_device);
|
|
|
|
|
2023-03-10 13:34:28 +08:00
|
|
|
dma_addr_t virtqueue_get_desc_addr(const struct virtqueue *_vq)
|
2014-10-07 22:39:47 +08:00
|
|
|
{
|
2023-03-10 13:34:28 +08:00
|
|
|
const struct vring_virtqueue *vq = to_vvq(_vq);
|
2014-10-07 22:39:47 +08:00
|
|
|
|
2016-02-03 13:46:37 +08:00
|
|
|
BUG_ON(!vq->we_own_ring);
|
|
|
|
|
2018-11-21 18:03:27 +08:00
|
|
|
if (vq->packed_ring)
|
|
|
|
return vq->packed.ring_dma_addr;
|
|
|
|
|
2018-11-21 18:03:25 +08:00
|
|
|
return vq->split.queue_dma_addr;
|
2014-10-07 22:39:47 +08:00
|
|
|
}
|
2016-02-03 13:46:37 +08:00
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_get_desc_addr);
|
2014-10-07 22:39:47 +08:00
|
|
|
|
2023-03-10 13:34:28 +08:00
|
|
|
dma_addr_t virtqueue_get_avail_addr(const struct virtqueue *_vq)
|
2014-10-07 22:39:47 +08:00
|
|
|
{
|
2023-03-10 13:34:28 +08:00
|
|
|
const struct vring_virtqueue *vq = to_vvq(_vq);
|
2014-10-07 22:39:47 +08:00
|
|
|
|
2016-02-03 13:46:37 +08:00
|
|
|
BUG_ON(!vq->we_own_ring);
|
|
|
|
|
2018-11-21 18:03:27 +08:00
|
|
|
if (vq->packed_ring)
|
|
|
|
return vq->packed.driver_event_dma_addr;
|
|
|
|
|
2018-11-21 18:03:25 +08:00
|
|
|
return vq->split.queue_dma_addr +
|
2018-11-21 18:03:21 +08:00
|
|
|
((char *)vq->split.vring.avail - (char *)vq->split.vring.desc);
|
2016-02-03 13:46:37 +08:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_get_avail_addr);
|
|
|
|
|
2023-03-10 13:34:28 +08:00
|
|
|
dma_addr_t virtqueue_get_used_addr(const struct virtqueue *_vq)
|
2016-02-03 13:46:37 +08:00
|
|
|
{
|
2023-03-10 13:34:28 +08:00
|
|
|
const struct vring_virtqueue *vq = to_vvq(_vq);
|
2016-02-03 13:46:37 +08:00
|
|
|
|
|
|
|
BUG_ON(!vq->we_own_ring);
|
|
|
|
|
2018-11-21 18:03:27 +08:00
|
|
|
if (vq->packed_ring)
|
|
|
|
return vq->packed.device_event_dma_addr;
|
|
|
|
|
2018-11-21 18:03:25 +08:00
|
|
|
return vq->split.queue_dma_addr +
|
2018-11-21 18:03:21 +08:00
|
|
|
((char *)vq->split.vring.used - (char *)vq->split.vring.desc);
|
2016-02-03 13:46:37 +08:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_get_used_addr);
|
|
|
|
|
2018-11-21 18:03:27 +08:00
|
|
|
/* Only available for split ring */
|
2023-03-10 13:34:28 +08:00
|
|
|
const struct vring *virtqueue_get_vring(const struct virtqueue *vq)
|
2016-02-03 13:46:37 +08:00
|
|
|
{
|
2018-11-21 18:03:21 +08:00
|
|
|
return &to_vvq(vq)->split.vring;
|
2014-10-07 22:39:47 +08:00
|
|
|
}
|
2016-02-03 13:46:37 +08:00
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_get_vring);
|
2014-10-07 22:39:47 +08:00
|
|
|
|
2023-08-10 20:30:55 +08:00
|
|
|
/**
|
|
|
|
* virtqueue_dma_map_single_attrs - map DMA for _vq
|
|
|
|
* @_vq: the struct virtqueue we're talking about.
|
|
|
|
* @ptr: the pointer of the buffer to do dma
|
|
|
|
* @size: the size of the buffer to do dma
|
|
|
|
* @dir: DMA direction
|
|
|
|
* @attrs: DMA Attrs
|
|
|
|
*
|
|
|
|
* The caller calls this to do dma mapping in advance. The DMA address can be
|
|
|
|
* passed to this _vq when it is in pre-mapped mode.
|
|
|
|
*
|
|
|
|
* return DMA address. Caller should check that by virtqueue_dma_mapping_error().
|
|
|
|
*/
|
|
|
|
dma_addr_t virtqueue_dma_map_single_attrs(struct virtqueue *_vq, void *ptr,
|
|
|
|
size_t size,
|
|
|
|
enum dma_data_direction dir,
|
|
|
|
unsigned long attrs)
|
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
|
|
|
|
if (!vq->use_dma_api)
|
|
|
|
return (dma_addr_t)virt_to_phys(ptr);
|
|
|
|
|
|
|
|
return dma_map_single_attrs(vring_dma_dev(vq), ptr, size, dir, attrs);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_dma_map_single_attrs);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* virtqueue_dma_unmap_single_attrs - unmap DMA for _vq
|
|
|
|
* @_vq: the struct virtqueue we're talking about.
|
|
|
|
* @addr: the dma address to unmap
|
|
|
|
* @size: the size of the buffer
|
|
|
|
* @dir: DMA direction
|
|
|
|
* @attrs: DMA Attrs
|
|
|
|
*
|
|
|
|
* Unmap the address that is mapped by the virtqueue_dma_map_* APIs.
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
void virtqueue_dma_unmap_single_attrs(struct virtqueue *_vq, dma_addr_t addr,
|
|
|
|
size_t size, enum dma_data_direction dir,
|
|
|
|
unsigned long attrs)
|
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
|
|
|
|
if (!vq->use_dma_api)
|
|
|
|
return;
|
|
|
|
|
|
|
|
dma_unmap_single_attrs(vring_dma_dev(vq), addr, size, dir, attrs);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_dma_unmap_single_attrs);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* virtqueue_dma_mapping_error - check dma address
|
|
|
|
* @_vq: the struct virtqueue we're talking about.
|
|
|
|
* @addr: DMA address
|
|
|
|
*
|
|
|
|
* Returns 0 means dma valid. Other means invalid dma address.
|
|
|
|
*/
|
|
|
|
int virtqueue_dma_mapping_error(struct virtqueue *_vq, dma_addr_t addr)
|
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
|
|
|
|
if (!vq->use_dma_api)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
return dma_mapping_error(vring_dma_dev(vq), addr);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_dma_mapping_error);
|
|
|
|
|
2023-08-10 20:30:56 +08:00
|
|
|
/**
|
|
|
|
* virtqueue_dma_need_sync - check a dma address needs sync
|
|
|
|
* @_vq: the struct virtqueue we're talking about.
|
|
|
|
* @addr: DMA address
|
|
|
|
*
|
|
|
|
* Check if the dma address mapped by the virtqueue_dma_map_* APIs needs to be
|
|
|
|
* synchronized
|
|
|
|
*
|
|
|
|
* return bool
|
|
|
|
*/
|
|
|
|
bool virtqueue_dma_need_sync(struct virtqueue *_vq, dma_addr_t addr)
|
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
|
|
|
|
if (!vq->use_dma_api)
|
|
|
|
return false;
|
|
|
|
|
|
|
|
return dma_need_sync(vring_dma_dev(vq), addr);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_dma_need_sync);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* virtqueue_dma_sync_single_range_for_cpu - dma sync for cpu
|
|
|
|
* @_vq: the struct virtqueue we're talking about.
|
|
|
|
* @addr: DMA address
|
|
|
|
* @offset: DMA address offset
|
|
|
|
* @size: buf size for sync
|
|
|
|
* @dir: DMA direction
|
|
|
|
*
|
|
|
|
* Before calling this function, use virtqueue_dma_need_sync() to confirm that
|
|
|
|
* the DMA address really needs to be synchronized
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
void virtqueue_dma_sync_single_range_for_cpu(struct virtqueue *_vq,
|
|
|
|
dma_addr_t addr,
|
|
|
|
unsigned long offset, size_t size,
|
|
|
|
enum dma_data_direction dir)
|
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
struct device *dev = vring_dma_dev(vq);
|
|
|
|
|
|
|
|
if (!vq->use_dma_api)
|
|
|
|
return;
|
|
|
|
|
2023-12-01 11:33:03 +08:00
|
|
|
dma_sync_single_range_for_cpu(dev, addr, offset, size, dir);
|
2023-08-10 20:30:56 +08:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_dma_sync_single_range_for_cpu);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* virtqueue_dma_sync_single_range_for_device - dma sync for device
|
|
|
|
* @_vq: the struct virtqueue we're talking about.
|
|
|
|
* @addr: DMA address
|
|
|
|
* @offset: DMA address offset
|
|
|
|
* @size: buf size for sync
|
|
|
|
* @dir: DMA direction
|
|
|
|
*
|
|
|
|
* Before calling this function, use virtqueue_dma_need_sync() to confirm that
|
|
|
|
* the DMA address really needs to be synchronized
|
|
|
|
*/
|
|
|
|
void virtqueue_dma_sync_single_range_for_device(struct virtqueue *_vq,
|
|
|
|
dma_addr_t addr,
|
|
|
|
unsigned long offset, size_t size,
|
|
|
|
enum dma_data_direction dir)
|
|
|
|
{
|
|
|
|
struct vring_virtqueue *vq = to_vvq(_vq);
|
|
|
|
struct device *dev = vring_dma_dev(vq);
|
|
|
|
|
|
|
|
if (!vq->use_dma_api)
|
|
|
|
return;
|
|
|
|
|
2023-12-01 11:33:03 +08:00
|
|
|
dma_sync_single_range_for_device(dev, addr, offset, size, dir);
|
2023-08-10 20:30:56 +08:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(virtqueue_dma_sync_single_range_for_device);
|
|
|
|
|
2008-02-05 12:50:05 +08:00
|
|
|
MODULE_LICENSE("GPL");
|