summaryrefslogtreecommitdiff
path: root/lib/fault-inject.c
diff options
context:
space:
mode:
authorJoel Becker <joel.becker@oracle.com>2009-08-19 02:13:50 -0700
committerJoel Becker <joel.becker@oracle.com>2010-02-26 15:41:13 -0800
commitd3981544d7a4ed276966cdd58fe2f5d6e8a585d9 (patch)
treea2a90d6a67c40fac4b56edd524b08a8c52179570 /lib/fault-inject.c
parentc5d95df5f78312c879f3058059c98a08821897a5 (diff)
ocfs2: Set xattr block entries with ocfs2_xa_set()
ocfs2_xattr_block_set() calls into ocfs2_xattr_set_entry() with just the HAS_XATTR flag. Most of the machinery of ocfs2_xattr_set_entry() is skipped. All that really happens other than the call to ocfs2_xa_set() is making sure the HAS_XATTR flag is set on the inode. But HAS_XATTR should be set when we also set di->i_xattr_loc. And that's done in ocfs2_create_xattr_block(). So let's move it there, and then ocfs2_xattr_block_set() can just call ocfs2_xa_set(). While we're there, ocfs2_create_xattr_block() can take the set_ctxt for a smaller argument list. It also learns to set HAS_XATTR_FL, because it knows for sure. ocfs2_create_empty_xatttr_block() in the reflink path fakes a set_ctxt to call ocfs2_create_xattr_block(). Signed-off-by: Joel Becker <joel.becker@oracle.com>
Diffstat (limited to 'lib/fault-inject.c')
0 files changed, 0 insertions, 0 deletions