From: Yan, Zheng <zheng.z.yan@intel.com>
Date: Tue, 15 Apr 2014 08:06:07 +0000 (+0800)
Subject: mds: share max size to client who is allowed for WR cap
X-Git-Tag: v0.80-rc1~22^2~1
X-Git-Url: http://git-server-git.apps.pok.os.sepia.ceph.com/?a=commitdiff_plain;h=f6db1bc2b684a7fb1884aa300ba95d3928342391;p=ceph.git

mds: share max size to client who is allowed for WR cap

WR cap is allowed for the loner client when filelock is in excl->mix
state. MDS should share max size with the loner client in this case.
Otherwise the client may wait for the max size forever.

Fixes: #8092
Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
---

diff --git a/src/mds/Locker.cc b/src/mds/Locker.cc
index b24aec9bcfc6..e5fe00c2a679 100644
--- a/src/mds/Locker.cc
+++ b/src/mds/Locker.cc
@@ -1674,7 +1674,8 @@ void Locker::file_update_finish(CInode *in, MutationRef& mut, bool share, client
       issue_caps(in, cap);
     }
   
-    if (share && in->is_auth() && in->filelock.is_stable())
+    if (share && in->is_auth() &&
+	(in->filelock.gcaps_allowed(CAP_LONER) & (CEPH_CAP_GWR|CEPH_CAP_GBUFFER)))
       share_inode_max_size(in);
   }
   issue_caps_set(need_issue);