Thread (5 messages) 5 messages, 2 authors, 2023-03-03

Re: [PATCH 1/1] mm/nommu: remove unnecessary VMA locking

From: Suren Baghdasaryan <surenb@google.com>
Date: 2023-03-03 01:34:49
Also in: linux-mm, lkml

On Thu, Mar 2, 2023 at 1:41 AM David Hildenbrand [off-list ref] wrote:
On 01.03.23 20:04, Suren Baghdasaryan wrote:
quoted
Since CONFIG_PER_VMA_LOCK depends on CONFIG_MMU, the changes in nommu
are not needed. Remove them.

Fixes: bad94decd6a4 ("mm: write-lock VMAs before removing them from VMA tree")
Reported-by: Hyeonggon Yoo <redacted>
Link: https://lore.kernel.org/all/Y%2F8CJQGNuMUTdLwP@localhost/ (local)
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
---
Fix cleanly applies over mm-unstable, SHA in "Fixes" is from that tree.

  mm/nommu.c | 5 -----
  1 file changed, 5 deletions(-)
diff --git a/mm/nommu.c b/mm/nommu.c
index 2ab162d773e2..57ba243c6a37 100644
--- a/mm/nommu.c
+++ b/mm/nommu.c
@@ -588,7 +588,6 @@ static int delete_vma_from_mm(struct vm_area_struct *vma)
                     current->pid);
              return -ENOMEM;
      }
-     vma_start_write(vma);
      cleanup_vma_from_mm(vma);

      /* remove from the MM's tree and list */
@@ -1520,10 +1519,6 @@ void exit_mmap(struct mm_struct *mm)
       */
      mmap_write_lock(mm);
      for_each_vma(vmi, vma) {
-             /*
-              * No need to lock VMA because this is the only mm user and no
-              * page fault handled can race with it.
-              */
              cleanup_vma_from_mm(vma);
              delete_vma(mm, vma);
              cond_resched();
So, i assume this should be squashed.
Yes, that would be best.
Reviewed-by: David Hildenbrand <redacted>
Thanks1

Just a general comment: usually, if review of the original series is
still going on, it makes a lot more sense to raise such things in the
original series so the author can fixup while things are still in
mm-unstable. Once the series is in mm-stable, it's a different story. In
that case, it is usually good to have the mail subjects be something
like  "[PATCH mm-stable 1/1] ...".
Ok... For my education, do you mean the title of this patch should
somehow reflect that it should be folded into the original patch? Just
trying to understand the actionable item here. How would you change
this patch when posting for mm-unstable and for mm-stable?
Thanks,
Suren.
--
Thanks,

David / dhildenb
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help