Re: [PATCH RFC 0/5] mm: GUP dirty + ->release_folio() fixes
From: John Hubbard
Date: Thu Nov 19 2026 - 19:55:55 EST
On 11/19/26 12:10 PM, Pedro Falcato wrote:
> Sooo.... It turns out GUP dirty and ->release_folio() interact really
> weirdly. This was initially reported by some AI agent, and David let me
> know.
>
> The basic issue is:
The basic issue is covered in a series of LWN articles that span 5 or
6 years:
https://lwn.net/Kernel/Index/#Memory_management-get_user_pages
Probably good to read and absorb that. There's a lot.
The problem remains not fully solved. pin_user_pages() is a building
block, but we need to fully coordinate with filesystems in order to
finish. It's good that you are looking at this.
> - ->dirty_folio() relies on buffers/IFS/etc being attached to the
> folios. It _cannot_ attach them itself, as dirty_folio() cannot
> block.
> - certain MM paths (reclaim, khugepaged, folio split, etc) are very
> trigger-happy when it comes to asking filesystems to release them. In
> most cases, barring any dirty folios, buffers, it will just drop
> them.
> - there is no problem in usual IO submission, but issues start to pop
> up with GUP, where the code flow (roughly) becomes:
> [get_user_pages(), etc]
> ->page_mkwrite() called here, or before
> ->private correctly populated with whatever the filesystem
> expects.
> folio is marked as dirty, initially.
> *write finishes*
> folio_lock();
> folio_mark_dirty();
> ->dirty_folio()
> dirty_folio expects ->private to be populated by
> ->page_mkwrite(), ->write_begin().
> folio_unlock();
>
> What happens if any filemap_release_folio() runs (and succeeds) between
> the initial dirtying (plus eventually writeback) and the write to the
> folio completes? Welp, it's unpredictable. ext4 for example will
> WARN_ONCE, and likely drop data (see ext4_dirty_folio()).
>
> So, what can we do about it? It's complicated, hence why this is an RFC.
> For the possible solutions:
> - ->dirty_folio() attaching this FS information itself is not really
> possible in many cases. ->dirty_folio() can't block, nor is it even
> guaranteed to hold the folio lock while being called.
> - Can pincount help? Kind of. Things like folio_maybe_dma_pinned() are
> rough, and don't deal with the usage of get_user_pages(FOLL_WRITE).
> pincount itself isn't available in the common order-0 case, and not
> available on order-1 as well on 32-bit, but who cares about that^W^W^W^W^W
> - Could get_user_pages() ban FOLL_WRITE? Mayhaps. I would not be
> opposed to that, but that's probably a larger change than ideal for stable,
> and I'm not familiar enough with GUP requirements to decide to go that way.
> This would require migrating all FOLL_WRITE users to
> pin_user_pages().
> - Can the whole system be reworked? Probably, but not for stable.
>
> This patch series is really rough but kind-of works. Instead of using
> folio_maybe_dma_pinned(), which only interacts with pin_user_pages(), I
pin_user_pages() was specifically introduced for this exact class of
problem. get_user_pages() is now meant for other, much more limited cases.
I've attempted to document that, here:
https://docs.kernel.org/core-api/pin_user_pages.html
> introduce a new helper that helps places deal with the folio refcount
> directly. In that case, we can also deal with GUP. On the other end,
> moar problems:
> - NFS private data itself takes refs on folios. Now private data (that
> we could perhaps release if only we tried) is keeping us from
> releasing that same data. What?
> - ext4 data=journal has some funky usage of buffer_heads. Data journalled
> filesystems will not, in any way, dirty BHs. Instead it will,
> *only if it was DMA pinned*, set the checked flag. ext4 skips
> dropping those buffers if the checked flag is set.
> - ext4 move_extents seems to have suspicious usage of
> filemap_release_folio() as well.
>
> Many of these have suspicious Fixes: (which are definitely wrong). I
> targeted these patches to complement Jan's original fix. Anything before
> that (or which is missing that patch) is hopelessly broken anyway.
> Proper Fixes: can be added, but many of them go back quite far.
>
> Signed-off-by: Pedro Falcato <pfalcato@xxxxxxx>
> ---
> Pedro Falcato (5):
> mm/khugepaged: fix filemap_release_folio() interaction with GUP dirty
> mm/huge_memory: fix filemap_release_folio() interaction with GUP dirty
> mm/vmscan: avoid releasing FS metadata for !FOLL_PIN gup folios
> mm/vmscan: be more careful when releasing FS metadata if buffer_heads_over_limit
> mm/filemap: refuse releasing maybe_dma_pinned folios
>
> mm/filemap.c | 3 +++
> mm/huge_memory.c | 22 +++++++++++-----------
> mm/internal.h | 48 ++++++++++++++++++++++++++++++++++++++++++++++++
> mm/khugepaged.c | 19 +++++++++++++------
> mm/vmscan.c | 18 ++++++------------
> 5 files changed, 81 insertions(+), 29 deletions(-)
> ---
> base-commit: 9663aa0da7070d958e333477ce0720512e9fa016
> change-id: 20261116-gup-dirty-what-b73da748f5d1
A git alias to wrap "git format-patch" with a sed(1) invocation is a great way
to avoid accidentally "leaking" gerrit change-id's into a posting, fwiw.
thanks,
--
John Hubbard