| Message ID | 1588227169-83237-1-git-send-email-wubo40@huawei.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [V3] fs/ceph:fix double unlock in handle_cap_export() | expand |
On Thu, Apr 30, 2020 at 2:13 PM Wu Bo <wubo40@huawei.com> wrote: > > If the ceph_mdsc_open_export_target_session() return fails, > we should add mutex_lock(&session->s_mutex) on IS_ERR(tsession) block > to avoid twice unlocking. because the session->s_mutex will be unlock > at the out_unlock lable. > > -- > v2 -> v3: > - Rewrite solution, adding a mutex_lock(&session->s_mutex) > to the IS_ERR(tsession) block. > - Modify the comment more clearly. > v1 -> v2: > - add spin_lock(&ci->i_ceph_lock) before goto out_unlock lable > > > Signed-off-by: Wu Bo <wubo40@huawei.com> > --- > fs/ceph/caps.c | 1 + > 1 file changed, 1 insertion(+) > > diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c > index 185db76..d27d778 100644 > --- a/fs/ceph/caps.c > +++ b/fs/ceph/caps.c > @@ -3746,6 +3746,7 @@ static void handle_cap_export(struct inode *inode, struct ceph_mds_caps *ex, > WARN_ON(1); > tsession = NULL; > target = -1; > + mutex_lock(&session->s_mutex); > } > goto retry; > > -- > 1.8.3.1 > Reviewed-by: "Yan, Zheng" <zyan@redhat.com>
On Thu, 2020-04-30 at 14:12 +0800, Wu Bo wrote: > If the ceph_mdsc_open_export_target_session() return fails, > we should add mutex_lock(&session->s_mutex) on IS_ERR(tsession) block > to avoid twice unlocking. because the session->s_mutex will be unlock > at the out_unlock lable. > > -- > v2 -> v3: > - Rewrite solution, adding a mutex_lock(&session->s_mutex) > to the IS_ERR(tsession) block. > - Modify the comment more clearly. > v1 -> v2: > - add spin_lock(&ci->i_ceph_lock) before goto out_unlock lable > For future reference, notes like this should go a little lower... > > > Signed-off-by: Wu Bo <wubo40@huawei.com> > --- > fs/ceph/caps.c | 1 + > 1 file changed, 1 insertion(+) > If you put the patch version notes here, they'll be stripped out when we go to merge the patch with git-am. > diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c > index 185db76..d27d778 100644 > --- a/fs/ceph/caps.c > +++ b/fs/ceph/caps.c > @@ -3746,6 +3746,7 @@ static void handle_cap_export(struct inode *inode, struct ceph_mds_caps *ex, > WARN_ON(1); > tsession = NULL; > target = -1; > + mutex_lock(&session->s_mutex); > } > goto retry; > Looks good. Merged, but I did clean up the verbiage in the changelog a bit. Thanks!
diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index 185db76..d27d778 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -3746,6 +3746,7 @@ static void handle_cap_export(struct inode *inode, struct ceph_mds_caps *ex, WARN_ON(1); tsession = NULL; target = -1; + mutex_lock(&session->s_mutex); } goto retry;
If the ceph_mdsc_open_export_target_session() return fails, we should add mutex_lock(&session->s_mutex) on IS_ERR(tsession) block to avoid twice unlocking. because the session->s_mutex will be unlock at the out_unlock lable. -- v2 -> v3: - Rewrite solution, adding a mutex_lock(&session->s_mutex) to the IS_ERR(tsession) block. - Modify the comment more clearly. v1 -> v2: - add spin_lock(&ci->i_ceph_lock) before goto out_unlock lable Signed-off-by: Wu Bo <wubo40@huawei.com> --- fs/ceph/caps.c | 1 + 1 file changed, 1 insertion(+)