commit e775ac4770bb1e5dfbfe22c5b56752ed4a2317c5
author Hirokazu Honda <hiroh@chromium.org> Tue Jan 17 17:54:02 2023
committer Chromium LUCI CQ <chromium-scoped@luci-project-accounts.iam.gserviceaccount.com> Tue Jan 17 17:54:02 2023
tree 4c266d595ccc839b7f82d95352d1b92efe3ed101
parent 62047f16f71ee6253709671a7cc4a2a83605fecc

Revert "media/gpu/v4l2VEA,IP: Workaround to USERPTR API against read-only buf"

This reverts commit 203fdd03a4317f2eac57161ee101515a82df704f.

Reason for revert: This workaround is no longer required because of
the kernel patch was reverted.

Original change's description:
> media/gpu/v4l2VEA,IP: Workaround to USERPTR API against read-only buf
>
> VideoFrame fed in VEA::Encode() is read-only since R107 if the
> VideoFrame has a shared memory. V4L2VideoEncodeAccelerator fails
> because VIDOC_QBUF with USERPTR buffer fails if the pointers
> references a read-only buffer.
> The USERPTR API issue is apparently to be fixed by a kernel side.
> In the meantime, this CL adds the workaround to the issue; the read
> only buffer is copied to a writable temporary buffer every Encode
> before VIDIOC_QBUF.
> Not that VideoFrame has a shared memory in the case of screen share
> because a camera stack produces GpuMemoryBuffer.
>
> Bug: b:243883312
> Test: video_encode_accelerator_tests on elm, kevin and trogdor
> Test: screen share in Google Meet on elm, kevin and trogdor
> Change-Id: I922d98eaf52f80d9cd6c3784a8544bffb1232856
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/3891421
> Reviewed-by: Nathan Hebert <nhebert@chromium.org>
> Commit-Queue: Hirokazu Honda <hiroh@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#1046659}

Bug: b:243883312, b:261660224
Change-Id: I6fe37276f2ebe8c8730cfd3dd766b04277b32a67
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/4091205
Reviewed-by: Nathan Hebert <nhebert@chromium.org>
Commit-Queue: Hirokazu Honda <hiroh@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1093388}