By making maybe_unlock_mmap_for_io() handle the VMA lock correctly,
we make fault_dirty_shared_page() safe to be called without the mmap
lock held.
Signed-off-by: Matthew Wilcox (Oracle) <willy@xxxxxxxxxxxxx>
Reported-by: David Hildenbrand <david@xxxxxxxxxx>
Tested-by: Suren Baghdasaryan <surenb@xxxxxxxxxx>
---
Andrew, can you insert this before "mm: handle faults that merely update
the accessed bit under the VMA lock" please? It could be handled as a fix
patch, but it actually stands on its own as a separate patch. No big deal
if it has to go in after that patch.
mm/internal.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/mm/internal.h b/mm/internal.h
index 8611f7c5bd16..c7720e83cb3c 100644
--- a/mm/internal.h
+++ b/mm/internal.h
@@ -706,7 +706,7 @@ static inline struct file *maybe_unlock_mmap_for_io(struct vm_fault *vmf,
if (fault_flag_allow_retry_first(flags) &&
!(flags & FAULT_FLAG_RETRY_NOWAIT)) {
fpin = get_file(vmf->vma->vm_file);
- mmap_read_unlock(vmf->vma->vm_mm);
+ release_fault_lock(vmf);
}
return fpin;
}
--
2.40.1
