I missed that when extending the lockdep annotations to the nonblocking case. I missed this while testing since in the i915 mmu notifiers is hitting a nice lockdep splat already before the point of going into oom killer mode :-/ Reported-by: syzbot+aaedc50d99a03250fe1f@xxxxxxxxxxxxxxxxxxxxxxxxx Fixes: d2b219ed03d4 ("mm/mmu_notifiers: add a lockdep map for invalidate_range_start/end") Cc: Jason Gunthorpe <jgg@xxxxxxxxxxxx> Cc: Daniel Vetter <daniel.vetter@xxxxxxxxx> Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> Cc: "Jérôme Glisse" <jglisse@xxxxxxxxxx> Cc: Ralph Campbell <rcampbell@xxxxxxxxxx> Cc: Jason Gunthorpe <jgg@xxxxxxxx> Cc: Ira Weiny <ira.weiny@xxxxxxxxx> Cc: Michal Hocko <mhocko@xxxxxxxx> Cc: Daniel Vetter <daniel.vetter@xxxxxxxx> Cc: Sean Christopherson <sean.j.christopherson@xxxxxxxxx> Cc: Jean-Philippe Brucker <jean-philippe@xxxxxxxxxx> Cc: linux-mm@xxxxxxxxx Signed-off-by: Daniel Vetter <daniel.vetter@xxxxxxxxx> --- include/linux/mmu_notifier.h | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/include/linux/mmu_notifier.h b/include/linux/mmu_notifier.h index 5a03417e5bf7..4edd98b06834 100644 --- a/include/linux/mmu_notifier.h +++ b/include/linux/mmu_notifier.h @@ -356,13 +356,14 @@ mmu_notifier_invalidate_range_start(struct mmu_notifier_range *range) static inline int mmu_notifier_invalidate_range_start_nonblock(struct mmu_notifier_range *range) { + int ret = 0; lock_map_acquire(&__mmu_notifier_invalidate_range_start_map); if (mm_has_notifiers(range->mm)) { range->flags &= ~MMU_NOTIFIER_RANGE_BLOCKABLE; - return __mmu_notifier_invalidate_range_start(range); + ret = __mmu_notifier_invalidate_range_start(range); } lock_map_release(&__mmu_notifier_invalidate_range_start_map); - return 0; + return ret; } static inline void -- 2.23.0 _______________________________________________ dri-devel mailing list dri-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/dri-devel