ext4: don't hold spinlock while calling ext4_issue_discard()
authorLukas Czerner <lczerner@redhat.com>
Thu, 28 Oct 2010 01:30:04 +0000 (21:30 -0400)
committerTheodore Ts'o <tytso@mit.edu>
Thu, 28 Oct 2010 01:30:04 +0000 (21:30 -0400)
We can't hold the block group spinlock because we ext4_issue_discard()
calls wait and hence can get rescheduled.

Google-Bug-Id: 3017678

Signed-off-by: Lukas Czerner <lczerner@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
fs/ext4/mballoc.c

index 53472e27b3277dc0c7bbcc287b7f765885cc5797..ccdfec6acb75c07209b63744a72f4f2cda3d1008 100644 (file)
@@ -4696,12 +4696,12 @@ do_more:
                 * with group lock held. generate_buddy look at
                 * them with group lock_held
                 */
+               if (test_opt(sb, DISCARD))
+                       ext4_issue_discard(sb, block_group, bit, count);
                ext4_lock_group(sb, block_group);
                mb_clear_bits(bitmap_bh->b_data, bit, count);
                mb_free_blocks(inode, &e4b, bit, count);
                ext4_mb_return_to_preallocation(inode, &e4b, block, count);
-               if (test_opt(sb, DISCARD))
-                       ext4_issue_discard(sb, block_group, bit, count);
        }
 
        ret = ext4_free_blks_count(sb, gdp) + count;