On Sat, 31 Mar 2007, Oliver Neukum wrote: > Am Samstag, 31. März 2007 04:59 schrieb Alan Stern: > > isn't clean -- but several drivers do it this way. I'm not even certain > > that adding smp_mb() to the completion routine will really fix it; memory > > barriers have to be used in pairs, and so there would also need to be a > > barrier between the test of private_data->urb_done and the call to > > use(urb). > > > > In the example you have given smp_rmb() would be required. > However we have: > > CPU A CPU B > a = urb->status if (flag) > flag = 1 urb->status = x > > abstractly speaking. CPU A needs a memory barrier. > > Are you saying that CPUs may do speculative _writes_ ?
No, they don't. You're right that in this case CPU B doesn't need a memory barrier. Alan Stern ------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys-and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV _______________________________________________ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel