Re: [PATCH v8 17/26] media/v4l2-core: set pages dirty upon releasing DMA buffers

[PATCH v8 00/26] mm/gup: track dma-pinned pages: FOLL_PIN · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
[PATCH v8 18/26] media/v4l2-core: pin_user_pages (FOLL_PIN) and put_user_page() conversion · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
[PATCH v8 20/26] powerpc: book3s64: convert to pin_user_pages() and put_user_page() · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
Re: [PATCH v8 20/26] powerpc: book3s64: convert to pin_user_pages() and put_user_page() · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
Re: [PATCH v8 20/26] powerpc: book3s64: convert to pin_user_pages() and put_user_page() · John Hubbard <jhubbard@nvidia.com> · 2019-12-10
Re: [PATCH v8 20/26] powerpc: book3s64: convert to pin_user_pages() and put_user_page() · Andrew Morton <akpm@linux-foundation.org> · 2019-12-11
[PATCH v8 25/26] mm/gup_benchmark: support pin_user_pages() and related calls · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
[PATCH v8 22/26] mm, tree-wide: rename put_user_page*() to unpin_user_page*() · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
[PATCH v8 24/26] mm/gup: track FOLL_PIN pages · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
Re: [PATCH v8 24/26] mm/gup: track FOLL_PIN pages · Jan Kara <jack@suse.cz> · 2019-12-10
Re: [PATCH v8 24/26] mm/gup: track FOLL_PIN pages · John Hubbard <jhubbard@nvidia.com> · 2019-12-11
[PATCH v8 26/26] selftests/vm: run_vmtests: invoke gup_benchmark with basic FOLL_PIN coverage · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
[PATCH v8 23/26] mm/gup: pass flags arg to __gup_device_* functions · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
Re: [PATCH v8 23/26] mm/gup: pass flags arg to __gup_device_* functions · Jan Kara <jack@suse.cz> · 2019-12-10
Re: [PATCH v8 23/26] mm/gup: pass flags arg to __gup_device_* functions · John Hubbard <jhubbard@nvidia.com> · 2019-12-11
[PATCH v8 12/26] IB/{core,hw,umem}: set FOLL_PIN via pin_user_pages*(), fix up ODP · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
[PATCH v8 21/26] mm/gup_benchmark: use proper FOLL_WRITE flags instead of hard-coding "1" · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
[PATCH v8 19/26] vfio, mm: pin_user_pages (FOLL_PIN) and put_user_page() conversion · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
[PATCH v8 15/26] fs/io_uring: set FOLL_PIN via pin_user_pages() · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
[PATCH v8 17/26] media/v4l2-core: set pages dirty upon releasing DMA buffers · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
Re: [PATCH v8 17/26] media/v4l2-core: set pages dirty upon releasing DMA buffers · Andrew Morton <akpm@linux-foundation.org> · 2019-12-10
Re: [PATCH v8 17/26] media/v4l2-core: set pages dirty upon releasing DMA buffers · John Hubbard <jhubbard@nvidia.com> · 2019-12-10
Re: [PATCH v8 17/26] media/v4l2-core: set pages dirty upon releasing DMA buffers · Jan Kara <jack@suse.cz> · 2019-12-10
[PATCH v8 14/26] drm/via: set FOLL_PIN via pin_user_pages_fast() · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
[PATCH v8 09/26] IB/umem: use get_user_pages_fast() to pin DMA pages · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
[PATCH v8 10/26] mm/gup: introduce pin_user_pages*() and FOLL_PIN · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
[PATCH v8 16/26] net/xdp: set FOLL_PIN via pin_user_pages() · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
[PATCH v8 13/26] mm/process_vm_access: set FOLL_PIN via pin_user_pages_remote() · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
[PATCH v8 06/26] mm: fix get_user_pages_remote()'s handling of FOLL_LONGTERM · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
[PATCH v8 08/26] mm/gup: allow FOLL_FORCE for get_user_pages_fast() · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
Re: [PATCH v8 08/26] mm/gup: allow FOLL_FORCE for get_user_pages_fast() · Jan Kara <jack@suse.cz> · 2019-12-10
[PATCH v8 11/26] goldish_pipe: convert to pin_user_pages() and put_user_page() · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
[PATCH v8 07/26] vfio: fix FOLL_LONGTERM use, simplify get_user_pages_remote() call · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
[PATCH v8 05/26] goldish_pipe: rename local pin_user_pages() routine · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
[PATCH v8 01/26] mm/gup: factor out duplicate code from four routines · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
[PATCH v8 04/26] mm: devmap: refactor 1-based refcounting for ZONE_DEVICE pages · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
[PATCH v8 03/26] mm: Cleanup __put_devmap_managed_page() vs ->page_free() · John Hubbard <jhubbard@nvidia.com> · 2019-12-09
[PATCH v8 02/26] mm/gup: move try_get_compound_head() to top, fix minor issues · John Hubbard <jhubbard@nvidia.com> · 2019-12-09

From: Andrew Morton <akpm@linux-foundation.org>
Date: 2019-12-10 00:56:38
Also in: bpf, dri-devel, kvm, linux-block, linux-doc, linux-fsdevel, linux-kselftest, linux-media, linux-mm, linux-rdma, lkml, netdev, stable

On Mon, 9 Dec 2019 14:53:35 -0800 John Hubbard [off-list ref] wrote:

After DMA is complete, and the device and CPU caches are synchronized,
it's still required to mark the CPU pages as dirty, if the data was
coming from the device. However, this driver was just issuing a
bare put_page() call, without any set_page_dirty*() call.

Fix the problem, by calling set_page_dirty_lock() if the CPU pages
were potentially receiving data from the device.

Reviewed-by: Christoph Hellwig <hch@lst.de>
Acked-by: Hans Verkuil <redacted>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: <redacted>

What are the user-visible effects of this change?

As it's cc:stable I'd normally send this to Linus within 1-2 weeks, or
sooner.  Please confirm that this is a standalone fix, independent of
the rest of this series.

`h`	back out one level
`j`	next message in thread
`k`	previous message in thread
`l`	drill in
`Esc`	close help / fold thread tree
`?`	toggle this help