aboutsummaryrefslogtreecommitdiffstats
path: root/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c
diff options
context:
space:
mode:
authorVishal Moola (Oracle) <[email protected]>2024-04-15 21:17:47 +0000
committerAndrew Morton <[email protected]>2024-04-25 02:34:26 +0000
commit37641efaa3faa4b8292aba4bbd7d71c0b703a239 (patch)
treeba4426bd53e97e37951e37cfbe757ddd7998ecfe /drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c
parentmm: zswap: fix shrinker NULL crash with cgroup_disable=memory (diff)
downloadkernel-37641efaa3faa4b8292aba4bbd7d71c0b703a239.tar.gz
kernel-37641efaa3faa4b8292aba4bbd7d71c0b703a239.zip
hugetlb: check for anon_vma prior to folio allocation
Commit 9acad7ba3e25 ("hugetlb: use vmf_anon_prepare() instead of anon_vma_prepare()") may bailout after allocating a folio if we do not hold the mmap lock. When this occurs, vmf_anon_prepare() will release the vma lock. Hugetlb then attempts to call restore_reserve_on_error(), which depends on the vma lock being held. We can move vmf_anon_prepare() prior to the folio allocation in order to avoid calling restore_reserve_on_error() without the vma lock. Link: https://lkml.kernel.org/r/ZiFqSrSRLhIV91og@fedora Fixes: 9acad7ba3e25 ("hugetlb: use vmf_anon_prepare() instead of anon_vma_prepare()") Reported-by: [email protected] Signed-off-by: Vishal Moola (Oracle) <[email protected]> Cc: Muchun Song <[email protected]> Cc: <[email protected]> Signed-off-by: Andrew Morton <[email protected]>
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c')
0 files changed, 0 insertions, 0 deletions