Skip to contentSkip to conversation

ext4: use kobj_completion in ext4_sb_info

PATCH In review Jeff Mahoney 2 participants 3 messages Sep 10, 2013 – Sep 12, 2013 ext4 file system

lore
Share & tools
lore

Jeff Mahoney

author
#0 lore raw

PATCH 2/2 ext4: use kobj_completion in ext4_sb_info

Now that we have the kobj_completion structure available, let's use it in ext4.

Note that we aggregate the del/put/wait_for_completion in one kobj_completion_del_and_wait call. Since the kobject isn't used for anything while the fs is shutting down and we wait until it's released before continuing anyway, this is safe.

Signed-off-by Jeff Mahoney <[email protected]>
---
 fs/ext4/ext4.h  |    4 ++--
 fs/ext4/super.c |   41 ++++++++++++-----------------------------
 2 files changed, 14 insertions(+), 31 deletions(-)
fs/ext4/ext4.h +2 -2
@@ -30,6 +30,7 @@
 #include <linux/blockgroup_lock.h>
 #include <linux/percpu_counter.h>
 #include <crypto/hash.h>
+#include <linux/kobj_completion.h>
 #ifdef __KERNEL__
 #include <linux/compat.h>
 #endif

@@ -1198,8 +1199,7 @@ struct ext4_sb_info {
 	struct percpu_counter s_dirtyclusters_counter;
 	struct blockgroup_lock *s_blockgroup_lock;
 	struct proc_dir_entry *s_proc;
-	struct kobject s_kobj;
-	struct completion s_kobj_unregister;
+	struct kobj_completion s_kobjc;
 	struct super_block *s_sb;
 
 	/* Journaling */
fs/ext4/super.c +12 -29

96 lines not on this page · raw message

Signature · 2 lines
Jeff Mahoney
SUSE Labs
maintainer
#1 lore raw
Quoted 9 lines · Jeff Mahoney

Now that we have the kobj_completion structure available, let's use it in ext4.

Note that we aggregate the del/put/wait_for_completion in one kobj_completion_del_and_wait call. Since the kobject isn't used for anything while the fs is shutting down and we wait until it's released before continuing anyway, this is safe.

Signed-off-by Jeff Mahoney <[email protected]>

I don't have a huge objection to this, but given that it's not fixing a bug, and it's not saving _that_ many lines of code, if you want to run the first patch through Greg's tree (which is fine with me), I might wait to merge this patch until 3.13, so it's not obvious that it's worth dealing with the inter-tree dependency, since it complicates testing on my end.

Does that sound reasonable?

- Ted

author
#2 lore raw
Quoted 11 lines · Theodore Ts'o
Quoted 9 lines · Jeff Mahoney

Now that we have the kobj_completion structure available, let's use it in ext4.

Note that we aggregate the del/put/wait_for_completion in one kobj_completion_del_and_wait call. Since the kobject isn't used for anything while the fs is shutting down and we wait until it's released before continuing anyway, this is safe.

Signed-off-by Jeff Mahoney <[email protected]>
Quoted 8 lines

I don't have a huge objection to this, but given that it's not fixing a bug, and it's not saving _that_ many lines of code, if you want to run the first patch through Greg's tree (which is fine with me), I might wait to merge this patch until 3.13, so it's not obvious that it's worth dealing with the inter-tree dependency, since it complicates testing on my end.

Does that sound reasonable?

Works for me. The ext4 case isn't something I'm in a huge rush to get in -- it was just the most obvious (to me) existing users of the kobj/completion pair.

-Jeff

Signature · 2 lines
Jeff Mahoney
SUSE Labs