diff options
author | Ricardo Labiaga <Ricardo.Labiaga@netapp.com> | 2010-10-13 01:30:05 +0200 |
---|---|---|
committer | Trond Myklebust <Trond.Myklebust@netapp.com> | 2010-10-24 23:59:56 +0200 |
commit | 6b96724e507fecc3e6440e86426fe4f44359ed66 (patch) | |
tree | 5b38131405301f4dd536a11f44ff0d5e6785d266 /fs/nfs/file.c | |
parent | SUNRPC: After calling xprt_release(), we must restart from call_reserve (diff) | |
download | linux-6b96724e507fecc3e6440e86426fe4f44359ed66.tar.xz linux-6b96724e507fecc3e6440e86426fe4f44359ed66.zip |
Revalidate caches on lock
Instead of blindly zapping the caches, attempt to revalidate them if
the server has indicated that it uses high resolution timestamps.
NFSv4 should be able to always revalidate the cache since the
protocol requires the update of the change attribute on modification of
the data. In reality, there are servers (the Linux NFS server
for example) that do not obey this requirement and use ctime as the
basis for change attribute. Long term, the server needs to be fixed.
At this time, and to be on the safe side, continue zapping caches if
the server indicates that it does not have a high resolution timestamp.
Signed-off-by: Ricardo Labiaga <Ricardo.Labiaga@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Diffstat (limited to 'fs/nfs/file.c')
-rw-r--r-- | fs/nfs/file.c | 19 |
1 files changed, 16 insertions, 3 deletions
diff --git a/fs/nfs/file.c b/fs/nfs/file.c index 39672b731736..c3f2477c16c1 100644 --- a/fs/nfs/file.c +++ b/fs/nfs/file.c @@ -758,6 +758,11 @@ do_unlk(struct file *filp, int cmd, struct file_lock *fl, int is_local) } static int +is_time_granular(struct timespec *ts) { + return ((ts->tv_sec == 0) && (ts->tv_nsec <= 1000)); +} + +static int do_setlk(struct file *filp, int cmd, struct file_lock *fl, int is_local) { struct inode *inode = filp->f_mapping->host; @@ -781,13 +786,21 @@ do_setlk(struct file *filp, int cmd, struct file_lock *fl, int is_local) status = do_vfs_lock(filp, fl); if (status < 0) goto out; + /* - * Make sure we clear the cache whenever we try to get the lock. + * Revalidate the cache if the server has time stamps granular + * enough to detect subsecond changes. Otherwise, clear the + * cache to prevent missing any changes. + * * This makes locking act as a cache coherency point. */ nfs_sync_mapping(filp->f_mapping); - if (!nfs_have_delegation(inode, FMODE_READ)) - nfs_zap_caches(inode); + if (!nfs_have_delegation(inode, FMODE_READ)) { + if (is_time_granular(&NFS_SERVER(inode)->time_delta)) + __nfs_revalidate_inode(NFS_SERVER(inode), inode); + else + nfs_zap_caches(inode); + } out: return status; } |