> > This makes sense but I think you should also consider the fact that > > the tx_lock is taken once per per tx_completion so, with the patch, > > the driver spends less time under lock. > > I think we could get rid of the tx lock on completion entirely... > > I have the old patch below lying around. It no longer applies, and it > needs some careful thought to make sure there are no memory ordering > problems (which need to be addressed with appropriate barriers when > looking at tx_head/tx_tail)
netif_queue_stopped is an atomic, I think that would be sufficient, no? -- MST _______________________________________________ openib-general mailing list [email protected] http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
