On Wed, 20 Aug 2014 18:07:19 +0800
Kinglong Mee <kinglongmee@xxxxxxxxx> wrote:
> Fix left code by Joe Perches's patch,
> "locks: Remove unused conf argument from lm_grant"
>
> Signed-off-by: Kinglong Mee <kinglongmee@xxxxxxxxx>
> ---
> fs/lockd/svclock.c | 26 +++++---------------------
> include/linux/lockd/lockd.h | 1 -
> 2 files changed, 5 insertions(+), 22 deletions(-)
>
> diff --git a/fs/lockd/svclock.c b/fs/lockd/svclock.c
> index 2a61701..796e63b 100644
> --- a/fs/lockd/svclock.c
> +++ b/fs/lockd/svclock.c
> @@ -245,7 +245,6 @@ nlmsvc_create_block(struct svc_rqst *rqstp, struct nlm_host *host,
> block->b_daemon = rqstp->rq_server;
> block->b_host = host;
> block->b_file = file;
> - block->b_fl = NULL;
> file->f_count++;
>
> /* Add to file's list of blocks */
> @@ -295,7 +294,6 @@ static void nlmsvc_free_block(struct kref *kref)
> nlmsvc_freegrantargs(block->b_call);
> nlmsvc_release_call(block->b_call);
> nlm_release_file(block->b_file);
> - kfree(block->b_fl);
> kfree(block);
> }
>
> @@ -523,20 +521,13 @@ nlmsvc_testlock(struct svc_rqst *rqstp, struct nlm_file *file,
> block = nlmsvc_lookup_block(file, lock);
>
> if (block == NULL) {
> - struct file_lock *conf = kzalloc(sizeof(*conf), GFP_KERNEL);
> -
I think the problem here is that we're allocating a file_lock with
kzalloc instead of using locks_alloc_lock. I think we should change
this code to use that, and the kfree in nlmsvc_free_block to use
locks_free_lock. Maybe something like this? (untested, but it compiles).
I'll plan to resend the patch below for 3.18 once I've tested it out and
Trond, Bruce and I can work out who should merge it.
-------------------------[snip]-------------------------
[PATCH] lockd: switch allocation of conflock to standard lock allocation routines
lockd currently allocates a struct file_lock with kzalloc to use as a
conflock. Change it to use locks_alloc_lock and locks_free_lock instead.
In the event that someone were to add lm_get_owner/lm_put_owner ops for
lockd, then this would help ensure that things get cleaned up properly.
It's also less wasteful with memory since locks_alloc_lock allocates
from a dedicated slabcache.
Cc: Kinglong Mee <kinglongmee@xxxxxxxxx>
Signed-off-by: Jeff Layton <jlayton@xxxxxxxxxxxxxxx>
---
fs/lockd/svclock.c | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/fs/lockd/svclock.c b/fs/lockd/svclock.c
index 2a6170133c1d..1eb2ae47e6b1 100644
--- a/fs/lockd/svclock.c
+++ b/fs/lockd/svclock.c
@@ -295,7 +295,8 @@ static void nlmsvc_free_block(struct kref *kref)
nlmsvc_freegrantargs(block->b_call);
nlmsvc_release_call(block->b_call);
nlm_release_file(block->b_file);
- kfree(block->b_fl);
+ if (block->b_fl)
+ locks_free_lock(block->b_fl);
kfree(block);
}
@@ -523,13 +524,13 @@ nlmsvc_testlock(struct svc_rqst *rqstp, struct nlm_file *file,
block = nlmsvc_lookup_block(file, lock);
if (block == NULL) {
- struct file_lock *conf = kzalloc(sizeof(*conf), GFP_KERNEL);
+ struct file_lock *conf = locks_alloc_lock();
if (conf == NULL)
return nlm_granted;
block = nlmsvc_create_block(rqstp, host, file, lock, cookie);
if (block == NULL) {
- kfree(conf);
+ locks_free_lock(conf);
return nlm_granted;
}
block->b_fl = conf;
--
1.9.3
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
