Re: [PATCH 2/3] Retry page fault when blocking on disk transfer.

[Michel Lespinasse <walken@google.com>]

On Fri, Oct 08, 2010 at 06:22:04PM -0700, Michel Lespinasse wrote:
> Second try on adding the VM_FAULT_RETRY functionality to the swap in path.
> 
> This proposal would replace [patch 2/3] of this series (the initial
> version of it, which was approved by linus / rik / hpa).

Oh my. I've been evidently sloppy while merging the filemap and swap
versions of lock_page_or_retry(), and this issue slipped in...

Please apply attached change. Sorry for not catching this earlier :(

----------------------------- 8< --------------------------------

Release page reference during page fault retry

This slipped by when unifying the filemap and swap versions of
lock_page_or_retry()...

Signed-off-by: Michel Lespinasse <walken@google.com>
    
diff --git a/mm/filemap.c b/mm/filemap.c
index 33f8125..c479d9c 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -1563,8 +1563,10 @@ retry_find:
 			goto no_cached_page;
 	}
 
-	if (!lock_page_or_retry(page, vma->vm_mm, vmf->flags))
+	if (!lock_page_or_retry(page, vma->vm_mm, vmf->flags)) {
+		page_cache_release(page);
 		return ret | VM_FAULT_RETRY;
+	}
 
 	/* Did it get truncated? */
 	if (unlikely(page->mapping != mapping)) {

-- 
Michel "Walken" Lespinasse
A program is never fully debugged until the last user dies.