[PATCH 2/2] sem restore: don't free sma if it's an error

Serge E. Hallyn serue at us.ibm.com
Wed Jan 13 12:17:30 PST 2010


Quoting Serge E. Hallyn (serue at us.ibm.com):
> Quoting serue at us.ibm.com (serue at us.ibm.com):
> > From: Serge Hallyn <serue at us.ibm.com>
> > 
> > Signed-off-by: Serge Hallyn <serue at us.ibm.com>
> > ---
> >  ipc/checkpoint_sem.c |    1 +
> >  1 files changed, 1 insertions(+), 0 deletions(-)
> > 
> > diff --git a/ipc/checkpoint_sem.c b/ipc/checkpoint_sem.c
> > index 53a19ed..395c84d 100644
> > --- a/ipc/checkpoint_sem.c
> > +++ b/ipc/checkpoint_sem.c
> > @@ -186,6 +186,7 @@ int restore_ipc_sem(struct ckpt_ctx *ctx, struct ipc_namespace *ns)
> >  	sma = restore_sem_array(ctx, h->sem_nsems);
> >  	if (IS_ERR(sma)) {
> >  		ret = PTR_ERR(sma);
> > +		sma = NULL;
> >  		goto out;
> >  	}
> > 
> > -- 
> > 1.6.0.6
> 
> So FWIW on x86-64 many of the tests current pass - the 3 failures I'm seeing
> are due to:
> 
> 	1. message queues for some reason aren't seen by the restarted
> 	   task (restart does think it is restoring them according to debug
> 	   info)
> 	2. self-checkpoint not yet supported
> 	3. restart block not set up right (or at least cr_tests/sleep fails
> 	   for some reason, haven't yet looked at why)
> 
> I'm trying to track down #1 right now

Oops, haha - bad selinux policy.  Down to two errors then.

-serge


More information about the Containers mailing list