From: Wengang Wang Date: Thu, 18 Nov 2010 00:44:02 +0800 Subject: [Ocfs2-devel] [PATCH] ocfs2/dlm: a missed DLM_LOCK_RES_IN_PROGRESS set In-Reply-To: <201011171556.oAH7aOht003969@rcsinet15.oracle.com> References: <201011171556.oAH7aOht003969@rcsinet15.oracle.com> Message-ID: <20101117164402.GA3401@laptop.jp.oracle.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: ocfs2-devel@oss.oracle.com Sorry, I want to take back the two patches. It seems DLM_LOCK_RES_IN_PROGRESS is needed only for remote work, and for master node, spin lock is enough. dlmunlock_common() does need some fix since it includes remote case. I will send the fix later. regards, wengang. On 10-11-17 23:53, Wengang Wang wrote: > We set DLM_LOCK_RES_IN_PROGRESS flag when we are processing a lock request. > There is a miss in dlmunlock_common. > > Signed-off-by: Wengang Wang > --- > fs/ocfs2/dlm/dlmunlock.c | 2 ++ > 1 files changed, 2 insertions(+), 0 deletions(-) > > diff --git a/fs/ocfs2/dlm/dlmunlock.c b/fs/ocfs2/dlm/dlmunlock.c > index 817287c..ef219ac 100644 > --- a/fs/ocfs2/dlm/dlmunlock.c > +++ b/fs/ocfs2/dlm/dlmunlock.c > @@ -137,6 +137,8 @@ static enum dlm_status dlmunlock_common(struct dlm_ctxt *dlm, > /* ok for this to sleep if not in a network handler */ > __dlm_wait_on_lockres(res); > res->state |= DLM_LOCK_RES_IN_PROGRESS; > + } else { > + res->state |= DLM_LOCK_RES_IN_PROGRESS; > } > spin_lock(&lock->spinlock); > > -- > 1.7.2.3 > > > _______________________________________________ > Ocfs2-devel mailing list > Ocfs2-devel at oss.oracle.com > http://oss.oracle.com/mailman/listinfo/ocfs2-devel