[PATCH v2] vfs: __read_cache_page should use gfp argument ratherthan GFP_KERNEL

[Dave Kleikamp]

[ updated to remove now-obsolete comment in read_cache_page_gfp()]

lockdep reports a deadlock in jfs because a special inode's rw semaphore
is taken recursively. The mapping's gfp mask is GFP_NOFS, but is not used
when __read_cache_page() calls add_to_page_cache_lru().

Signed-off-by: Dave Kleikamp <dave.kleikamp@xxxxxxxxxx>
Acked-by: Hugh Dickins <hughd@xxxxxxxxxx>
Acked-by: Al Viro <viro@xxxxxxxxxxxxxxxxxx>

diff --git a/mm/filemap.c b/mm/filemap.c
index c106d3b..5f0a3c9 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -1828,7 +1828,7 @@ repeat:
 		page = __page_cache_alloc(gfp | __GFP_COLD);
 		if (!page)
 			return ERR_PTR(-ENOMEM);
-		err = add_to_page_cache_lru(page, mapping, index, GFP_KERNEL);
+		err = add_to_page_cache_lru(page, mapping, index, gfp);
 		if (unlikely(err)) {
 			page_cache_release(page);
 			if (err == -EEXIST)
@@ -1925,10 +1925,7 @@ static struct page *wait_on_page_read(struct page *page)
  * @gfp:	the page allocator flags to use if allocating
  *
  * This is the same as "read_mapping_page(mapping, index, NULL)", but with
- * any new page allocations done using the specified allocation flags. Note
- * that the Radix tree operations will still use GFP_KERNEL, so you can't
- * expect to do this atomically or anything like that - but you can pass in
- * other page requirements.
+ * any new page allocations done using the specified allocation flags.
  *
  * If the page does not get brought uptodate, return -EIO.
  */
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/