Message ID | 63ADC13FD55D6546B7DECE290D39E373CED7AA61@H3CMLB14-EX.srv.huawei-3com.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Hi Changwei, Now, this seem a good fix. But it would be better if you can add a *Fixs* tag. BTW, I think this problem also exists in the "-stable" tree and you may cc <stable@vger.kernel.org>. Thanks, Alex On 2017/11/14 17:15, Changwei Ge wrote: > Sorry for previous patch's format. > So I resend it. > > When dlm_add_migration_mle returns -EEXIST, previously input mle will > not be initialized. So we can't use its associated dlm object. > And we truly don't need this mle for already launched migration > progress, since oldmle has taken this role. > > Thanks, > Changwei > > Signed-off-by: Changwei Ge <ge.changwei@h3c.com> > --- > fs/ocfs2/dlm/dlmmaster.c | 4 +++- > 1 file changed, 3 insertions(+), 1 deletion(-) > > diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c > index 3e04279446e8..9c3e0f13ca87 100644 > --- a/fs/ocfs2/dlm/dlmmaster.c > +++ b/fs/ocfs2/dlm/dlmmaster.c > @@ -2616,7 +2616,9 @@ static int dlm_migrate_lockres(struct dlm_ctxt *dlm, > * otherwise the assert_master from the new > * master will destroy this. > */ > - dlm_get_mle_inuse(mle); > + if (ret != -EEXIST) > + dlm_get_mle_inuse(mle); > + > spin_unlock(&dlm->master_lock); > spin_unlock(&dlm->spinlock); >
On 17/11/14 17:15, Changwei Ge wrote: > Sorry for previous patch's format. > So I resend it. > > When dlm_add_migration_mle returns -EEXIST, previously input mle will > not be initialized. So we can't use its associated dlm object. > And we truly don't need this mle for already launched migration > progress, since oldmle has taken this role. > > Thanks, > Changwei > > Signed-off-by: Changwei Ge <ge.changwei@h3c.com> Reviewed-by: Joseph Qi <jiangqi903@gmail.com> > --- > fs/ocfs2/dlm/dlmmaster.c | 4 +++- > 1 file changed, 3 insertions(+), 1 deletion(-) > > diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c > index 3e04279446e8..9c3e0f13ca87 100644 > --- a/fs/ocfs2/dlm/dlmmaster.c > +++ b/fs/ocfs2/dlm/dlmmaster.c > @@ -2616,7 +2616,9 @@ static int dlm_migrate_lockres(struct dlm_ctxt *dlm, > * otherwise the assert_master from the new > * master will destroy this. > */ > - dlm_get_mle_inuse(mle); > + if (ret != -EEXIST) > + dlm_get_mle_inuse(mle); > + > spin_unlock(&dlm->master_lock); > spin_unlock(&dlm->spinlock); >
diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c index 3e04279446e8..9c3e0f13ca87 100644 --- a/fs/ocfs2/dlm/dlmmaster.c +++ b/fs/ocfs2/dlm/dlmmaster.c @@ -2616,7 +2616,9 @@ static int dlm_migrate_lockres(struct dlm_ctxt *dlm, * otherwise the assert_master from the new * master will destroy this. */ - dlm_get_mle_inuse(mle); + if (ret != -EEXIST) + dlm_get_mle_inuse(mle); + spin_unlock(&dlm->master_lock); spin_unlock(&dlm->spinlock);
Sorry for previous patch's format. So I resend it. When dlm_add_migration_mle returns -EEXIST, previously input mle will not be initialized. So we can't use its associated dlm object. And we truly don't need this mle for already launched migration progress, since oldmle has taken this role. Thanks, Changwei Signed-off-by: Changwei Ge <ge.changwei@h3c.com> --- fs/ocfs2/dlm/dlmmaster.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)