This patch adds support for the V4L2_BUF_FLAG_M2M_HOLD_CAPTURE_BUF
flag.
It also adds a new V4L2_BUF_CAP_SUPPORTS_M2M_HOLD_CAPTURE_BUF
capability.
Drivers should set vb2_queue->subsystem_flags to
VB2_V4L2_FL_SUPPORTS_M2M_HOLD_CAPTURE_BUF to indicate support
for this flag.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
This is rougly equivalent to ca0e68e21a (drm/prime: skip CPU sync
in map/unmap dma_buf). The contig memory allocated is already device
coherent memory, so there is no point in doing a CPU sync when
mapping it to another device. Also most importers currently cache
the mapping so the CPU sync would only happen on the first import,
so we are better off with not pretending to do a cache synchronization
at all.
This gets rid of a lot of CPU overhead in uses where those dma-bufs
are regularily imported and detached again, like Weston is currently
doing in the DRM compositor.
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Acked-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
[hverkuil-cisco@xs4all.nl: fix checkpatch warnings]
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
When reaching the end of stream, V4L2 clients may expect the
V4L2_EOS_EVENT before being able to dequeue the last buffer, which has
the V4L2_BUF_FLAG_LAST flag set.
If the vb2_poll() function first checks for events and afterwards if
buffers are available, a driver can queue the V4L2_EOS_EVENT event and
return the buffer after the check for events but before the check for
buffers. This causes vb2_poll() to signal that the buffer with
V4L2_BUF_FLAG_LAST can be read without the V4L2_EOS_EVENT being
available.
First, check for available buffers and afterwards for events to ensure
that if vb2_poll() signals POLLIN | POLLRDNORM for the
V4L2_BUF_FLAG_LAST buffer, it also signals POLLPRI for the
V4L2_EOS_EVENT.
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
-----BEGIN PGP SIGNATURE-----
iQFSBAABCAA8FiEEq68RxlopcLEwq+PEeb4+QwBBGIYFAlz8fAYeHHRvcnZhbGRz
QGxpbnV4LWZvdW5kYXRpb24ub3JnAAoJEHm+PkMAQRiG1asH/3ySguxqtqL1MCBa
4/SZ37PHeWKMerfX6ZyJdgEqK3B+PWlmuLiOMNK5h2bPLzeQQQAmHU/mfKmpXqgB
dHwUbG9yNnyUtTfsfRqAnCA6vpuw9Yb1oIzTCVQrgJLSWD0j7scBBvmzYqguOkto
ThwigLUq3AILr8EfR4rh+GM+5Dn9OTEFAxwil9fPHQo7QoczwZxpURhScT6Co9TB
DqLA3fvXbBvLs/CZy/S5vKM9hKzC+p39ApFTURvFPrelUVnythAM0dPDJg3pIn5u
g+/+gDxDFa+7ANxvxO2ng1sJPDqJMeY/xmjJYlYyLpA33B7zLNk2vDHhAP06VTtr
XCMhQ9s=
=cb80
-----END PGP SIGNATURE-----
Merge tag 'v5.2-rc4' into media/master
There are some conflicts due to SPDX changes. We also have more
patches being merged via media tree touching them.
So, let's merge back from upstream and address those.
Linux 5.2-rc4
* tag 'v5.2-rc4': (767 commits)
Linux 5.2-rc4
MAINTAINERS: Karthikeyan Ramasubramanian is MIA
i2c: xiic: Add max_read_len quirk
lockref: Limit number of cmpxchg loop retries
uaccess: add noop untagged_addr definition
x86/insn-eval: Fix use-after-free access to LDT entry
kbuild: use more portable 'command -v' for cc-cross-prefix
s390/unwind: correct stack switching during unwind
block, bfq: add weight symlink to the bfq.weight cgroup parameter
cgroup: let a symlink too be created with a cftype file
drm/nouveau/secboot/gp10[2467]: support newer FW to fix SEC2 failures on some boards
drm/nouveau/secboot: enable loading of versioned LS PMU/SEC2 ACR msgqueue FW
drm/nouveau/secboot: split out FW version-specific LS function pointers
drm/nouveau/secboot: pass max supported FW version to LS load funcs
drm/nouveau/core: support versioned firmware loading
drm/nouveau/core: pass subdev into nvkm_firmware_get, rather than device
block: free sched's request pool in blk_cleanup_queue
pktgen: do not sleep with the thread lock held.
net: mvpp2: Use strscpy to handle stat strings
net: rds: fix memory leak in rds_ib_flush_mr_pool
...
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Based on 1 normalized pattern(s):
this program is free software you can redistribute it and or modify
it under the terms of the gnu general public license as published by
the free software foundation either version 2 of the license or at
your option any later version
extracted by the scancode license scanner the SPDX license identifier
GPL-2.0-or-later
has been chosen to replace the boilerplate/reference in 3029 file(s).
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Allison Randal <allison@lohutok.net>
Cc: linux-spdx@vger.kernel.org
Link: https://lkml.kernel.org/r/20190527070032.746973796@linutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In vb2_vmalloc_get_userptr() the framevector is created with the
'write' argument set to false when vb2_create_framevec() is called
for OUTPUT buffers. So the pages are marked as read-only.
However, userspace will write to these buffers since it will fill
in the data to output. Since get_userptr is only called if the userptr
of the queued buffer has changed since the last time that same buffer
was queued, this will fail when the buffer contents is updated and the
buffer is queued again.
E.g., userspace fills buffer 1 with the output video and queues it.
The first time get_userptr is called and the pages are grabbed and
pinned in memory and marked read-only. The second time buffer 1 is
filled with different video data and queued again. Since the userptr
hasn't changed the get_userptr() callback isn't called again. Since
the pages were marked as read-only the new contents isn't updated.
Just always call vb2_create_framevec() with FOLL_WRITE to always
allow writing to the buffers.
Using USERPTR streaming with OUTPUT devices is almost never done. And
when it is done it is via v4l2-compliance and a driver like vim2m. But
since v4l2-compliance doesn't actually inspect the capture buffer and
compare it to the original output buffer, this issue was never noticed.
But the vicodec driver actually needs to parse the bitstream in the
OUTPUT buffers and any errors there will be immediately noticed. So
this time v4l2-compliance failed the USERPTR streaming test.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
last_buffer_dequeued was set to true in __fill_v4l2_buffer, but this
is called for qbuf as well. Move it to vb2_dqbuf.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
-----BEGIN PGP SIGNATURE-----
iQFSBAABCAA8FiEEq68RxlopcLEwq+PEeb4+QwBBGIYFAlzrJgUeHHRvcnZhbGRz
QGxpbnV4LWZvdW5kYXRpb24ub3JnAAoJEHm+PkMAQRiG1XAIAJajJpeKc9MVQipi
zWiI+WgHDg8gG/HnLKzP2Zp4MUjTCz30t0rnBbdo6AyK6LnPBL53yxg9q64XN7vI
p/h2ys+/DvqhIYSPWX6C++HYQT4Cb5ghxJABEIsztd3G4nfB9L2vgq9zKSNFusTD
UtfUxufciZPX515TuE4IqWMwS4Ut5daH7V3jjZunhPiayTFv64e1KbpKLLzHR0NR
DfFbrPcdp7VyCT/A5AKrqfqeB6O9dm/Fe4CsdsaKuzASCXJ6GUc/A7iZDndv/KA4
f5Xh7SGlQ2TfT1ud/aMfdw3AcMjqB8zgMJenovJ7UuHDfULBWNeAmMh/cTDwY9eY
Xgyvfm0=
=S8p9
-----END PGP SIGNATURE-----
Merge tag 'v5.2-rc2' into patchwork
Merge back from upstream into media tree, as there are some
patches merged upstream that has pontential of causing
conflicts (one actually rised a conflict already).
Linux 5.2-rc2
* tag 'v5.2-rc2': (377 commits)
Linux 5.2-rc2
random: fix soft lockup when trying to read from an uninitialized blocking pool
tracing: Silence GCC 9 array bounds warning
ext4: fix dcache lookup of !casefolded directories
locking/lock_events: Use this_cpu_add() when necessary
KVM: x86: fix return value for reserved EFER
tools/kvm_stat: fix fields filter for child events
KVM: selftests: Wrap vcpu_nested_state_get/set functions with x86 guard
kvm: selftests: aarch64: compile with warnings on
kvm: selftests: aarch64: fix default vm mode
kvm: selftests: aarch64: dirty_log_test: fix unaligned memslot size
KVM: s390: fix memory slot handling for KVM_SET_USER_MEMORY_REGION
KVM: x86/pmu: do not mask the value that is written to fixed PMUs
KVM: x86/pmu: mask the result of rdpmc according to the width of the counters
x86/kvm/pmu: Set AMD's virt PMU version to 1
KVM: x86: do not spam dmesg with VMCS/VMCB dumps
kvm: Check irqchip mode before assign irqfd
kvm: svm/avic: fix off-by-one in checking host APIC ID
KVM: selftests: do not blindly clobber registers in guest asm
KVM: selftests: Remove duplicated TEST_ASSERT in hyperv_cpuid.c
...
The size argument of the alloc memop, which allocates buffer memory, is
page aligned. Document it as such in the only caller as well as ops
documentation.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
buf->size is an unsigned long; casting that to int will lead to an
overflow if buf->size exceeds INT_MAX.
Fix this by changing the type to unsigned long instead. This is possible
as the buf->size is always aligned to PAGE_SIZE, and therefore the size
will never have values lesser than 0.
Note on backporting to stable: the file used to be under
drivers/media/v4l2-core, it was moved to the current location after 4.14.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Cc: stable@vger.kernel.org
Reviewed-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
PAGE_ALIGN() may wrap the buffer size around to 0. Prevent this by
checking that the aligned value is not smaller than the unaligned one.
Note on backporting to stable: the file used to be under
drivers/media/v4l2-core, it was moved to the current location after 4.14.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Cc: stable@vger.kernel.org
Reviewed-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Add SPDX license identifiers to all Make/Kconfig files which:
- Have no license information of any form
These files fall under the project license, GPL v2 only. The resulting SPDX
license identifier is:
GPL-2.0-only
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Convert to use vm_map_pages() to map range of kernel memory to user vma.
vm_pgoff is treated in V4L2 API as a 'cookie' to select a buffer, not as a
in-buffer offset by design and it always want to mmap a whole buffer from
its beginning.
Link: http://lkml.kernel.org/r/a953fe6b3056de1cc6eab654effdd4a22f125375.1552921225.git.jrdr.linux@gmail.com
Signed-off-by: Souptick Joarder <jrdr.linux@gmail.com>
Suggested-by: Marek Szyprowski <m.szyprowski@samsung.com>
Reviewed-by: Marek Szyprowski <m.szyprowski@samsung.com>
Cc: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Heiko Stuebner <heiko@sntech.de>
Cc: Joerg Roedel <joro@8bytes.org>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Juergen Gross <jgross@suse.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Kyungmin Park <kyungmin.park@samsung.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Mauro Carvalho Chehab <mchehab@infradead.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Mike Rapoport <rppt@linux.ibm.com>
Cc: Oleksandr Andrushchenko <oleksandr_andrushchenko@epam.com>
Cc: Pawel Osciak <pawel@osciak.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rik van Riel <riel@surriel.com>
Cc: Robin Murphy <robin.murphy@arm.com>
Cc: Russell King <linux@armlinux.org.uk>
Cc: Sandy Huang <hjc@rock-chips.com>
Cc: Stefan Richter <stefanr@s5r6.in-berlin.de>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Thierry Reding <treding@nvidia.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Calling VIDIOC_DQBUF can release the core serialization lock pointed to
by vb2_queue->lock if it has to wait for a new buffer to arrive.
However, if userspace dup()ped the video device filehandle, then it is
possible to read or call DQBUF from two filehandles at the same time.
It is also possible to call REQBUFS from one filehandle while the other
is waiting for a buffer. This will remove all the buffers and reallocate
new ones. Removing all the buffers isn't the problem here (that's already
handled correctly by DQBUF), but the reallocating part is: DQBUF isn't
aware that the buffers have changed.
This is fixed by setting a flag whenever the lock is released while waiting
for a buffer to arrive. And checking the flag where needed so we can return
-EBUSY.
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Reported-by: Syzbot <syzbot+4180ff9ca6810b06c1e9@syzkaller.appspotmail.com>
Reviewed-by: Tomasz Figa <tfiga@chromium.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
If a buffer is queued to a request, followed by an attempt to queue
the same buffer again, then the second qbuf returns an error since
the buffer is not in the DEQUEUED state anymore.
However, before it gets to that check it executes the code under the
'if (!vb->prepared)' condition. This clears previously set data needed
for request handling, and now querybuf will no longer report that this
buffer is part of a request.
Move the state check to before the 'if' and make sure to only do the
state check when called from QBUF and if V4L2_BUF_FLAG_REQUEST_FD is
set.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Minor fix for helper function and comment, s/v4l2_vb2_buffer/vb2_v4l2_buffer.
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
If requests are used when they shouldn't, or not used when they should,
then return EBADR (Invalid request descriptor) instead of EACCES.
The reason for this change is that EACCES has more to do with permissions
(not being the owner of the resource), but in this case the request file
descriptor is just wrong for the current mode of the device.
Update the documentation accordingly.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Stateless codecs require the use of the Request API as opposed of it
being optional.
So add a bit to indicate this and let vb2 check for this.
If an attempt is made to queue a buffer without an associated request,
then the EBADR error is returned to userspace.
Doing this check in the vb2 core simplifies drivers, since they
don't have to check for this, they can just set this flag.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Reviewed-by: Paul Kocialkowski <paul.kocialkowski@bootlin.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
There is no point to continuing assignment after memory allocation
failed, rather throw error immediately.
Signed-off-by: Souptick Joarder <jrdr.linux@gmail.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
[hverkuil-cisco@xs4all.nl: rebase and remove empty line before the if]
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
The last user of this state has been converted, so we can now drop
this. Requeueing causes the queue to become unordered, which causes
problems with requests and (in the future) fences.
Since it is no longer needed, just get rid of this.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
- add debugfs support for dumping dma-debug information (Corentin Labbe)
- Kconfig cleanups (Andy Shevchenko and me)
- debugfs cleanups (Greg Kroah-Hartman)
- improve dma_map_resource and use it in the media code
- arch_setup_dma_ops / arch_teardown_dma_ops cleanups
- various small cleanups and improvements for the per-device coherent
allocator
- make the DMA mask an upper bound and don't fail "too large" dma mask
in the remaning two architectures - this will allow big driver
cleanups in the following merge windows
-----BEGIN PGP SIGNATURE-----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=nAMI
-----END PGP SIGNATURE-----
Merge tag 'dma-mapping-5.1' of git://git.infradead.org/users/hch/dma-mapping
Pull DMA mapping updates from Christoph Hellwig:
- add debugfs support for dumping dma-debug information (Corentin
Labbe)
- Kconfig cleanups (Andy Shevchenko and me)
- debugfs cleanups (Greg Kroah-Hartman)
- improve dma_map_resource and use it in the media code
- arch_setup_dma_ops / arch_teardown_dma_ops cleanups
- various small cleanups and improvements for the per-device coherent
allocator
- make the DMA mask an upper bound and don't fail "too large" dma mask
in the remaning two architectures - this will allow big driver
cleanups in the following merge windows
* tag 'dma-mapping-5.1' of git://git.infradead.org/users/hch/dma-mapping: (21 commits)
Documentation/DMA-API-HOWTO: update dma_mask sections
sparc64/pci_sun4v: allow large DMA masks
sparc64/iommu: allow large DMA masks
sparc64: refactor the ali DMA quirk
ccio: allow large DMA masks
dma-mapping: remove the DMA_MEMORY_EXCLUSIVE flag
dma-mapping: remove dma_mark_declared_memory_occupied
dma-mapping: move CONFIG_DMA_CMA to kernel/dma/Kconfig
dma-mapping: improve selection of dma_declare_coherent availability
dma-mapping: remove an incorrect __iommem annotation
of: select OF_RESERVED_MEM automatically
device.h: dma_mem is only needed for HAVE_GENERIC_DMA_COHERENT
mfd/sm501: depend on HAS_DMA
dma-mapping: add a kconfig symbol for arch_teardown_dma_ops availability
dma-mapping: add a kconfig symbol for arch_setup_dma_ops availability
dma-mapping: move debug configuration options to kernel/dma
dma-debug: add dumping facility via debugfs
dma: debug: no need to check return value of debugfs_create functions
videobuf2: replace a layering violation with dma_map_resource
dma-mapping: don't BUG when calling dma_map_resource on RAM
...
Use codespell to fix lots of typos over frontends.
Manually verified to avoid false-positives.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
The epoll function expects that whenever the poll file op is
called, the poll_wait function is also called. That didn't
always happen in vb2_core_poll() and vb2_poll(). Fix this,
otherwise epoll() would timeout when it shouldn't.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Reported-by: Yi Qingliang <niqingliang2003@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
If a stream is stopped, or if a USERPTR/DMABUF buffer is queued
backed by a different user address or dmabuf fd, then the timestamp
should be skipped by vb2_find_timestamp since the memory it refers
to is no longer valid.
So keep track of a 'copied_timestamp' state: it is set when the
timestamp is copied from an output to a capture buffer, and is
cleared when it is no longer valid.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Reviewed-by: Paul Kocialkowski <paul.kocialkowski@bootlin.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
The bool type is not recommended for use in structs, so replace these
by bitfields.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Reviewed-by: Paul Kocialkowski <paul.kocialkowski@bootlin.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Commit 2cc1802f62 ("media: vb2: Keep dma-buf buffers mapped until
they are freed") removed code leaving a local variable unused.
Remove it to avoid a compiler warning.
Fixes: 2cc1802f62 ("media: vb2: Keep dma-buf buffers mapped until they are freed")
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
vb2_dc_get_userptr pokes into arm direct mapping details to get the
resemblance of a dma address for a a physical address that does is
not backed by a page struct. Not only is this not portable to other
architectures with dma direct mapping offsets, but also not to uses
of IOMMUs of any kind. Switch to the proper dma_map_resource /
dma_unmap_resource interface instead.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Tested-by: Marek Szyprowski <m.szyprowski@samsung.com>
The buf_out_validate is required for output queues in combination
with requests. Check this.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
When queueing a buffer to a request the 'field' value is not validated.
That field is only validated when the _buf_prepare() is called,
which happens when the request is queued.
However, this validation should happen at QBUF time, since you want
to know about this as soon as possible. Also, the spec requires that
the 'field' value is validated at QBUF time.
This patch adds a new buf_out_validate callback to validate the
output buffer at buf_prepare time or when QBUF queues an unprepared
buffer to a request. This callback is mandatory for output queues
that support requests.
This issue was found by v4l2-compliance since it failed to replace
V4L2_FIELD_ANY by a proper field value when testing the vivid video
output in combination with requests.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
When using vb2 for video decoding, dequeued capture buffers may still
be accessed by the hardware: this is the case when they are used as
reference frames for decoding subsequent frames.
When the buffer is imported with dma-buf, it needs to be mapped before
access. Until now, it was mapped when queuing and unmapped when
dequeuing, which doesn't work for access as a reference frames.
One way to solve this would be to map the buffer again when it is
needed as a reference, but the mapping/unmapping operations can
seriously impact performance. As a result, map the buffer once (when it
is first needed when queued) and keep it mapped until it is freed.
Reviewed-on: https://chromium-review.googlesource.com/334103
[Paul: Updated for mainline and changed commit message]
Signed-off-by: Pawel Osciak <posciak@chromium.org>
Signed-off-by: Paul Kocialkowski <paul.kocialkowski@bootlin.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
There really is no reason why vb2_find_timestamp can't just find
buffers in any state. Drop that part of the test.
This also means that vb->timestamp should only be set to 0 when
the driver doesn't copy timestamps.
This change allows for more efficient pipelining (i.e. you can use
a buffer for a reference frame even when it is queued).
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Reviewed-by: Tomasz Figa <tfiga@chromium.org>
Reviewed-by: Alexandre Courbot <acourbot@chromium.org>
Reviewed-by: Paul Kocialkowski <paul.kocialkowski@bootlin.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
My @samsung.com address is going to cease existing soon, so change it to
an address which can actually be used to contact me.
Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Use v4l2_timeval_to_ns instead of timeval_to_ns to ensure that
both kernelspace and userspace will use the same conversion
function.
Next add a new vb2_find_timestamp() function to find buffers
with a specific timestamp.
This function will only look at DEQUEUED and DONE buffers, i.e.
buffers that are already processed.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
-----BEGIN PGP SIGNATURE-----
iQFSBAABCAA8FiEEq68RxlopcLEwq+PEeb4+QwBBGIYFAlwW4/oeHHRvcnZhbGRz
QGxpbnV4LWZvdW5kYXRpb24ub3JnAAoJEHm+PkMAQRiG2QMH/Rl6iMpTUX23tMHe
eXQzAOSvQXaWlFoX25j1Jvt8nhS7Uy8vkdpYTCOI/7DF0Jg4O/6uxcZkErlwWxb8
MW1rMgpfO+OpDLSLXAO2GKxaKI3ArqF2BcOQA2mji1/jR2VUTqmIvBoudn5d+GYz
19aCyfdzmVTC38G9sBhhcqJ10EkxLiHe2K74bf4JxVuSf2EnTI4LYt5xJPDoT0/C
6fOeUNwVhvv5a4svvzJmortq7x7BwyxBQArc7PbO0MPhabLU4wyFUOTRszgsGd76
o5JuOFwgdIIHlSSacGla6rKq10nmkwR07fHfRFFwbvrfBOEHsXOP2hvzMZX+FLBK
IXOzdtc=
=XlMc
-----END PGP SIGNATURE-----
Merge tag 'v4.20-rc7' into patchwork
Linux 4.20-rc7
* tag 'v4.20-rc7': (403 commits)
Linux 4.20-rc7
scripts/spdxcheck.py: always open files in binary mode
checkstack.pl: fix for aarch64
userfaultfd: check VM_MAYWRITE was set after verifying the uffd is registered
fs/iomap.c: get/put the page in iomap_page_create/release()
hugetlbfs: call VM_BUG_ON_PAGE earlier in free_huge_page()
memblock: annotate memblock_is_reserved() with __init_memblock
psi: fix reference to kernel commandline enable
arch/sh/include/asm/io.h: provide prototypes for PCI I/O mapping in asm/io.h
mm/sparse: add common helper to mark all memblocks present
mm: introduce common STRUCT_PAGE_MAX_SHIFT define
alpha: fix hang caused by the bootmem removal
XArray: Fix xa_alloc when id exceeds max
drm/vmwgfx: Protect from excessive execbuf kernel memory allocations v3
MAINTAINERS: Daniel for drm co-maintainer
drm/amdgpu: drop fclk/gfxclk ratio setting
IB/core: Fix oops in netdev_next_upper_dev_rcu()
dm thin: bump target version
drm/vmwgfx: remove redundant return ret statement
drm/i915: Flush GPU relocs harder for gen3
...
The V4L2_BUF_TYPE_META_OUTPUT mirrors the V4L2_BUF_TYPE_META_CAPTURE with
the exception that it is an OUTPUT type. The use case for this is to pass
buffers to the device that are not image data but metadata. The formats,
just as the metadata capture formats, are typically device specific and
highly structured.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Acked-by: Hans Verkuil <hans.verkuil@cisco.com>
Reviewed-by: Tomasz Figa <tfiga@chromium.org>
Tested-by: Tian Shu Qiu <tian.shu.qiu@intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
The Request API is now merged to the kernel but the confidence on the
stability of that API is not great, especially regarding the interaction
with V4L2.
Add a Kconfig option for the API, with a scary-looking warning.
The patch itself disables request creation as well as does not advertise
them as buffer flags. The driver requiring requests (cedrus) now depends
on the Kconfig option as well.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Acked-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
When a buffer is returned to state QUEUED (that happens when
start_streaming fails), then do not unbind and put the object
from the request. Nothing has changed yet, so just keep it as
is.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
When vb2_buffer_done is called the buffer is unbound from the
request and put. The media_request_object_put also 'put's the
request reference. If the application has already closed the
request fd, then that means that the request reference at that
point goes to 0 and the whole request is released.
This means that the control handler associated with the request is
also freed and that causes this kernel oops:
[174705.995401] BUG: sleeping function called from invalid context at kernel/locking/mutex.c:908
[174705.995411] in_atomic(): 1, irqs_disabled(): 1, pid: 28071, name: vivid-000-vid-o
[174705.995416] 2 locks held by vivid-000-vid-o/28071:
[174705.995420] #0: 000000001ea3a232 (&dev->mutex#3){....}, at: vivid_thread_vid_out+0x3f5/0x550 [vivid]
[174705.995447] #1: 00000000e30a0d1e (&(&q->done_lock)->rlock){....}, at: vb2_buffer_done+0x92/0x1d0 [videobuf2_common]
[174705.995460] Preemption disabled at:
[174705.995461] [<0000000000000000>] (null)
[174705.995472] CPU: 11 PID: 28071 Comm: vivid-000-vid-o Tainted: G W 4.20.0-rc1-test-no #88
[174705.995476] Hardware name: VMware, Inc. VMware Virtual Platform/440BX Desktop Reference Platform, BIOS 6.00 05/19/2017
[174705.995481] Call Trace:
[174705.995500] dump_stack+0x46/0x60
[174705.995512] ___might_sleep.cold.79+0xe1/0xf1
[174705.995523] __mutex_lock+0x50/0x8f0
[174705.995531] ? find_held_lock+0x2d/0x90
[174705.995536] ? find_held_lock+0x2d/0x90
[174705.995542] ? find_held_lock+0x2d/0x90
[174705.995564] ? v4l2_ctrl_handler_free.part.13+0x44/0x1d0 [videodev]
[174705.995576] v4l2_ctrl_handler_free.part.13+0x44/0x1d0 [videodev]
[174705.995590] v4l2_ctrl_request_release+0x1c/0x30 [videodev]
[174705.995600] media_request_clean+0x64/0xe0 [media]
[174705.995609] media_request_release+0x19/0x40 [media]
[174705.995617] vb2_buffer_done+0xef/0x1d0 [videobuf2_common]
[174705.995630] vivid_thread_vid_out+0x2c1/0x550 [vivid]
[174705.995645] ? vivid_stop_generating_vid_cap+0x1c0/0x1c0 [vivid]
[174705.995653] kthread+0x113/0x130
[174705.995659] ? kthread_park+0x80/0x80
[174705.995667] ret_from_fork+0x35/0x40
The vb2_buffer_done function can be called from interrupt context, so
anything that sleeps is not allowed.
The solution is to increment the request refcount when the buffer is
queued and decrement it when the buffer is dequeued. Releasing the
request is fine if that happens from VIDIOC_DQBUF.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
VIDIOC_PREPARE_BUF should ignore V4L2_BUF_FLAG_REQUEST_FD since it isn't
doing anything with requests. So inform vb2_queue_or_prepare_buf whether
it is called from vb2_prepare_buf or vb2_qbuf and just return 0 in the
first case.
This was found when adding new v4l2-compliance checks.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
vb2_start_streaming() already rolls back the buffers, so there is no
need to call __vb2_queue_cancel(). Especially since __vb2_queue_cancel()
does too much, such as zeroing the q->queued_count value, causing vb2
to think that no buffers have been queued.
It appears that this call to __vb2_queue_cancel() is a left-over from
before commit b3379c6201.
Fixes: b3379c6201 ('vb2: only call start_streaming if sufficient buffers are queued')
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Cc: <stable@vger.kernel.org> # for v4.16 and up
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
As reported by smatch:
drivers/media/common/videobuf2/videobuf2-core.c: drivers/media/common/videobuf2/videobuf2-core.c:2159 vb2_mmap() warn: inconsistent returns 'mutex:&q->mmap_lock'.
Locked on: line 2148
Unlocked on: line 2100
line 2108
line 2113
line 2118
line 2156
line 2159
There is one error condition that doesn't unlock a mutex.
Fixes: cd26d1c4d1 ("media: vb2: vb2_mmap: move lock up")
Reviewed-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Userspace shouldn't set bytesused to 0 for output buffers.
vb2_warn_zero_bytesused() warns about this (only once!), but it also
calls WARN_ON(1), which is confusing since it is not immediately clear
that it warns about a 0 value for bytesused.
Just drop the WARN_ON as it serves no purpose.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Acked-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Videobuf2 presently does not allow VIDIOC_REQBUFS to destroy outstanding
buffers if the queue is of type V4L2_MEMORY_MMAP, and if the buffers are
considered "in use". This is different behavior than for other memory
types and prevents us from deallocating buffers in following two cases:
1) There are outstanding mmap()ed views on the buffer. However even if
we put the buffer in reqbufs(0), there will be remaining references,
due to vma .open/close() adjusting vb2 buffer refcount appropriately.
This means that the buffer will be in fact freed only when the last
mmap()ed view is unmapped.
2) Buffer has been exported as a DMABUF. Refcount of the vb2 buffer
is managed properly by VB2 DMABUF ops, i.e. incremented on DMABUF
get and decremented on DMABUF release. This means that the buffer
will be alive until all importers release it.
Considering both cases above, there does not seem to be any need to
prevent reqbufs(0) operation, because buffer lifetime is already
properly managed by both mmap() and DMABUF code paths. Let's remove it
and allow userspace freeing the queue (and potentially allocating a new
one) even though old buffers might be still in processing.
To let userspace know that the kernel now supports orphaning buffers
that are still in use, add a new V4L2_BUF_CAP_SUPPORTS_ORPHANED_BUFS
to be set by reqbufs and create_bufs.
[p.zabel@pengutronix.de: added V4L2_BUF_CAP_SUPPORTS_ORPHANED_BUFS,
updated documentation, and added back debug message]
Signed-off-by: John Sheu <sheu@chromium.org>
Reviewed-by: Pawel Osciak <posciak@chromium.org>
Signed-off-by: Tomasz Figa <tfiga@chromium.org>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
[hverkuil-cisco@xs4all.nl: added V4L2-BUF-CAP-SUPPORTS-ORPHANED-BUFS ref]
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
vb2_core_create_bufs did not check if the memory model for newly added
buffers is the same as for already existing buffers. It should return an
error if they aren't the same.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Reported-by: syzbot+e1fb118a2ebb88031d21@syzkaller.appspotmail.com
Cc: <stable@vger.kernel.org> # for v4.16 and up
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
If a filehandle is dup()ped, then it is possible to close it from one fd
and call mmap from the other. This creates a race condition in vb2_mmap
where it is using queue data that __vb2_queue_free (called from close())
is in the process of releasing.
By moving up the mutex_lock(mmap_lock) in vb2_mmap this race is avoided
since __vb2_queue_free is called with the same mutex locked. So vb2_mmap
now reads consistent buffer data.
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Reported-by: syzbot+be93025dd45dccd8923c@syzkaller.appspotmail.com
Signed-off-by: Hans Verkuil <hansverk@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
-----BEGIN PGP SIGNATURE-----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=V4X4
-----END PGP SIGNATURE-----
Merge tag 'media/v4.20-2' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media
Pull new experimental media request API from Mauro Carvalho Chehab:
"A new media request API
This API is needed to support device drivers that can dynamically
change their parameters for each new frame. The latest versions of
Google camera and codec HAL depends on such feature.
At this stage, it supports only stateless codecs.
It has been discussed for a long time (at least over the last 3-4
years), and we finally reached to something that seem to work.
This series contain both the API and core changes required to support
it and a new m2m decoder driver (cedrus).
As the current API is still experimental, the only real driver using
it (cedrus) was added at staging[1]. We intend to keep it there for a
while, in order to test the API. Only when we're sure that this API
works for other cases (like encoders), we'll move this driver out of
staging and set the API into a stone.
[1] We added support for the vivid virtual driver (used only for
testing) to it too, as it makes easier to test the API for the ones
that don't have the cedrus hardware"
* tag 'media/v4.20-2' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media: (53 commits)
media: dt-bindings: Document the Rockchip VPU bindings
media: platform: Add Cedrus VPU decoder driver
media: dt-bindings: media: Document bindings for the Cedrus VPU driver
media: v4l: Add definition for the Sunxi tiled NV12 format
media: v4l: Add definitions for MPEG-2 slice format and metadata
media: videobuf2-core: Rework and rename helper for request buffer count
media: v4l2-ctrls.c: initialize an error return code with zero
media: v4l2-compat-ioctl32.c: add missing documentation for a field
media: media-request: update documentation
media: media-request: EPERM -> EACCES/EBUSY
media: v4l2-ctrls: improve media_request_(un)lock_for_update
media: v4l2-ctrls: use media_request_(un)lock_for_access
media: media-request: add media_request_(un)lock_for_access
media: vb2: set reqbufs/create_bufs capabilities
media: videodev2.h: add new capabilities for buffer types
media: buffer.rst: only set V4L2_BUF_FLAG_REQUEST_FD for QBUF
media: v4l2-ctrls: return -EACCES if request wasn't completed
media: media-request: return -EINVAL for invalid request_fds
media: vivid: add request support
media: vivid: add mc
...
The helper indicating whether buffers are associated with the request is
reworked and renamed to return the number of associated buffer objects.
This is useful for drivers that need to check how many buffers are in
the request to validate it.
Existing users of the helper don't need particular adaptation since the
meaning of zero/non-zero remains consistent.
Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Warn and return error from the reqbufs ioctl when driver sets 0 number
of planes or 0 as plane sizes, as these values don't make any sense.
Checking this here stops obviously wrong values from propagating
further and causing various problems that are hard to trace back to
either of these values being 0.
Signed-off-by: Johan Fjeldtvedt <johfjeld@cisco.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
If requests are not supported by the driver, then return EACCES, not
EPERM.
If you attempt to mix queueing buffers directly and using requests,
then EBUSY is returned instead of EPERM: once a specific queueing mode
has been chosen the queue is 'busy' if you attempt the other mode
(i.e. direct queueing vs via a request).
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Reviewed-by: Tomasz Figa <tfiga@chromium.org>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
