summary refs log tree commit diff
diff options
context:
space:
mode:
authorJeff Layton <jlayton@redhat.com>2014-03-26 07:24:23 -0700
committerSteve French <smfrench@gmail.com>2014-04-24 22:37:03 -0500
commita87c9ad956676d84d459739fc14ec5a3c3565717 (patch)
tree5e5da8ae521dc675775ee65f0964121a5d752d26
parent76429f1dedbc1815d875c9b68eefd30c95cb8fa9 (diff)
downloadlinux-a87c9ad956676d84d459739fc14ec5a3c3565717.tar.gz
cifs: fix actimeo=0 corner case when cifs_i->time == jiffies
actimeo=0 is supposed to be a special case that ensures that inode
attributes are always refetched from the server instead of trusting the
cache. The cifs code however uses time_in_range() to determine whether
the attributes have timed out. In the case where cifs_i->time equals
jiffies, this leads to the cifs code not refetching the inode attributes
when it should.

Fix this by explicitly testing for actimeo=0, and handling it as a
special case.

Reported-and-tested-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
-rw-r--r--fs/cifs/inode.c3
1 files changed, 3 insertions, 0 deletions
diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c
index aadc2b68678b..a22d667f1069 100644
--- a/fs/cifs/inode.c
+++ b/fs/cifs/inode.c
@@ -1737,6 +1737,9 @@ cifs_inode_needs_reval(struct inode *inode)
 	if (cifs_i->time == 0)
 		return true;
 
+	if (!cifs_sb->actimeo)
+		return true;
+
 	if (!time_in_range(jiffies, cifs_i->time,
 				cifs_i->time + cifs_sb->actimeo))
 		return true;