Re: [PATCH v4] mm/gup: honour FOLL_PIN in NOMMU __get_user_pages_locked()
From: Lance Yang
Date: Mon Sep 21 2026 - 08:47:17 EST
On 2026/9/21 17:57, David Hildenbrand (Arm) wrote:
From: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
The !CONFIG_MMU implementation of __get_user_pages_locked() takes a bare
get_page() reference for each page regardless of foll_flags:
if (pages[i])
get_page(pages[i]);
This is reached from pin_user_pages*() with FOLL_PIN set.
unpin_user_page() is shared between MMU and NOMMU configurations and
unconditionally calls gup_put_folio(..., FOLL_PIN), which subtracts
GUP_PIN_COUNTING_BIAS (1024) from the folio refcount.
This means that pin adds 1, and then unpin will subtract 1024.
If a user maps a page (refcount 1), registers it 1023 times as an
io_uring fixed buffer (1023 pin_user_pages calls -> refcount 1024), then
unregisters: the first unpin_user_page subtracts 1024, refcount hits 0,
the page is freed and returned to the buddy allocator. The remaining
1022 unpins write into whatever was reallocated, and the user's VMA
still maps the freed page (NOMMU has no MMU to invalidate it).
Reallocating the page for an io_uring pbuf_ring then lets userspace
corrupt the new owner's data through the stale mapping.
Use try_grab_folio() which adds GUP_PIN_COUNTING_BIAS for FOLL_PIN and 1
for FOLL_GET, mirroring the CONFIG_MMU path so pin and unpin are
symmetric. Keep supporting the traditional behavior where users specify
a pages array but don't set FOLL_GET.
While at it, don't return NULL pointers in the page array,
as this is really not expected for GUP users; instead, just fail and return
-EFAULT.
[ david: support traditional behavior with no FOLL_GET, extend
description ]
Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
Cc: David Hildenbrand <david@xxxxxxxxxx>
Cc: Jason Gunthorpe <jgg@xxxxxxxx>
Cc: John Hubbard <jhubbard@xxxxxxxxxx>
Cc: Peter Xu <peterx@xxxxxxxxxx>
Cc: Lance Yang <lance.yang@xxxxxxxxx>
Reported-by: Anthropic
Fixes: 3faa52c03f44 ("mm/gup: track FOLL_PIN pages")
Assisted-by: gkh_clanker_t1000
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
Signed-off-by: David Hildenbrand (Arm) <david@xxxxxxxxxx>
---
Tested-by: Lance Yang <lance.yang@xxxxxxxxx>