[PATCH 3.2 062/110] ext4: avoid deadlocks in the writeback path by using sb_getblk_gfp

From: Ben Hutchings
Date: Mon Aug 10 2015 - 06:25:21 EST


3.2.71-rc1 review patch. If anyone has any objections, please let me know.

------------------

From: Nikolay Borisov <kernel@xxxxxxxx>

commit c45653c341f5c8a0ce19c8f0ad4678640849cb86 upstream.

Switch ext4 to using sb_getblk_gfp with GFP_NOFS added to fix possible
deadlocks in the page writeback path.

Signed-off-by: Nikolay Borisov <kernel@xxxxxxxx>
Signed-off-by: Theodore Ts'o <tytso@xxxxxxx>
[bwh: Backported to 3.2: adjust context]
Signed-off-by: Ben Hutchings <ben@xxxxxxxxxxxxxxx>
---
--- a/fs/ext4/extents.c
+++ b/fs/ext4/extents.c
@@ -700,7 +700,8 @@ ext4_ext_find_extent(struct inode *inode
path[ppos].p_depth = i;
path[ppos].p_ext = NULL;

- bh = sb_getblk(inode->i_sb, path[ppos].p_block);
+ bh = sb_getblk_gfp(inode->i_sb, path[ppos].p_block,
+ __GFP_MOVABLE | GFP_NOFS);
if (unlikely(!bh)) {
ret = -ENOMEM;
goto err;
@@ -905,7 +906,7 @@ static int ext4_ext_split(handle_t *hand
err = -EIO;
goto cleanup;
}
- bh = sb_getblk(inode->i_sb, newblock);
+ bh = sb_getblk_gfp(inode->i_sb, newblock, __GFP_MOVABLE | GFP_NOFS);
if (!bh) {
err = -ENOMEM;
goto cleanup;
@@ -1089,7 +1090,7 @@ static int ext4_ext_grow_indepth(handle_
if (newblock == 0)
return err;

- bh = sb_getblk(inode->i_sb, newblock);
+ bh = sb_getblk_gfp(inode->i_sb, newblock, __GFP_MOVABLE | GFP_NOFS);
if (!bh)
return -ENOMEM;
lock_buffer(bh);

--
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/