On Fri, Dec 10, 2010 at 3:17 AM, Joe Eykholt <jeykholt@xxxxxxxxx> wrote: > > > On 12/9/10 6:18 AM, Hillf Danton wrote: >> On Thu, Dec 9, 2010 at 6:06 AM, Joe Eykholt <jeykholt@xxxxxxxxx> wrote: >> >> Locking here looks unnecessary, since lport could not goto Mars. >> If could, lot of orphans have to stay on Earth with no more care. > > The lport could go away before the rport is freed by rcu. Yes, rcu was neglected by my fault. > Also, two rports could be freed in rcu callbacks at the same time, > so two decrements of nr_rdata could be concurrent and you would > get nondeterministic results since the decrement isn't atomic. > > >> >> When initing lport, clarity could be redundant. > > So, you agree or not? Agree with no doubt. > >> When destroying lport, it is the right place to report leakage. > > Yes, but there's a call from lport destroy into rport, and that's the > place to report rport leakage, not in the FCP module. ÂFor one thing, > not all LLDs use that fcp module. Great point. Redelivery soon out. good weekend Hillf -- To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
