Each time we perform full initialization procedure regardless to
whenever quota was already initialized for a given inode.
In fact dquot_initialize() called many times during inode life time,
which result in many useless quota lookup/dqput actions. It is
reasonable to optimize the case where inode already has quota initialized.
We can avoid locking here because:
* Serialization between dquot_initialize() and dquot_drop() is guaranteed
by caller.
* Other races (quota_off/quota_on) result in incorrect quota regardless to
locking in dquot_initialize().
Signed-off-by: Dmitry Monakhov <dmonakhov@xxxxxxxxxx>
---
fs/quota/dquot.c | 17 +++++++++++++++--
1 files changed, 15 insertions(+), 2 deletions(-)
diff --git a/fs/quota/dquot.c b/fs/quota/dquot.c
index d7ec471..c06f969 100644
--- a/fs/quota/dquot.c
+++ b/fs/quota/dquot.c
@@ -1462,10 +1462,23 @@ out_err:
/* Drop unused references */
dqput_all(got);
}
-
void dquot_initialize(struct inode *inode)
{
- __dquot_initialize(inode, -1);
+ int cnt;
+ /*
+ * We can check inodes quota pointers without lock, because
+ * serialization between dquot_init/dquot_drop is guaranteed by caller.
+ */
+ if (IS_NOQUOTA(inode))
+ return;
+ for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
+ if (!sb_has_quota_active(inode->i_sb, cnt))
+ continue;
+ if (!inode->i_dquot[cnt])
+ break;
+ }
+ if (cnt < MAXQUOTAS)
+ __dquot_initialize(inode, -1);
}
EXPORT_SYMBOL(dquot_initialize);
--
1.6.5.2
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
