Btrfs: Add some debugging around the ENOSPC bugs
authorJosef Bacik <jbacik@redhat.com>
Tue, 18 Nov 2008 02:12:00 +0000 (21:12 -0500)
committerChris Mason <chris.mason@oracle.com>
Tue, 18 Nov 2008 02:12:00 +0000 (21:12 -0500)
Some people are still reporting problems with early enospc.  This
will help narrown down the cause.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
fs/btrfs/extent-tree.c

index 6c29669d81a29e7bd78db36c6d0fea099cd7026c..b0f2241274fdc02c59b47810e59b63c9f05fed97 100644 (file)
@@ -2984,6 +2984,9 @@ loop_check:
                        *last_ptr = ins->objectid + ins->offset;
                ret = 0;
        } else if (!ret) {
+               printk(KERN_ERR "we were searching for %Lu bytes, num_bytes %Lu,"
+                      " loop %d, allowed_alloc %d\n", total_needed, num_bytes,
+                      loop, allowed_chunk_alloc);
                ret = -ENOSPC;
        }