summaryrefslogtreecommitdiff
path: root/mm/memory.c
diff options
context:
space:
mode:
authorLinus Torvalds <torvalds@linux-foundation.org>2023-06-28 12:20:24 -0700
committerLinus Torvalds <torvalds@linux-foundation.org>2023-06-28 12:20:24 -0700
commit6581ccf03e717926be97dc3d27182ce351232f3c (patch)
treec4ca1b6867a5d72c755e5786130e1c97f5236cac /mm/memory.c
parent77b1a7f7a05c673c187894b4ae898a8c0cdc776c (diff)
mm: fix __access_remote_vm() GUP failure case
Commit ca5e863233e8 ("mm/gup: remove vmas parameter from get_user_pages_remote()") removed the vma argument from GUP handling, and instead added a helper function (get_user_page_vma_remote()) that looks it up separately using 'vma_lookup()'. And then converted existing users that needed a vma to use the helper instead. However, the helper function intentionally acts exactly like the old get_user_pages_remote() did, and only fills in 'vma' on successful page lookup. Fine so far. However, __access_remote_vm() wants the vma even for the unsuccessful case, and used to do a vma = vma_lookup(mm, addr); explicitly to look it up when the get_user_page() failed. However, that conversion commit incorrectly removed that vma lookup, thinking that get_user_page_vma_remote() would have done it. Not so. So add the vma_lookup() back in. Fixes: ca5e863233e8 ("mm/gup: remove vmas parameter from get_user_pages_remote()") Cc: Lorenzo Stoakes <lstoakes@gmail.com> Cc: David Hildenbrand <david@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Diffstat (limited to 'mm/memory.c')
-rw-r--r--mm/memory.c1
1 files changed, 1 insertions, 0 deletions
diff --git a/mm/memory.c b/mm/memory.c
index f758f59f3704..11882189864e 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -5592,6 +5592,7 @@ int __access_remote_vm(struct mm_struct *mm, unsigned long addr, void *buf,
* Check if this is a VM_IO | VM_PFNMAP VMA, which
* we can access using slightly different code.
*/
+ vma = vma_lookup(mm, addr);
if (!vma)
break;
if (vma->vm_ops && vma->vm_ops->access)