Oliver Neukum <[EMAIL PROTECTED]> wrote: > + (*) entering and returning from interrupt handlers implies a full barrier
That shouldn't really be under the "MISCELLANEOUS FUNCTIONS" subheading as it's not precisely describing a callable function that has barrier effects. Hmmm... also it isn't quite true: the individual handlers in an action chain are not barriered with respect to each other. How about the following: INTERRUPT HANDLING ------------------ Execution of the interrupt handler chain for an interrupt not bound to a CPU is bounded by a lock/unlock at either side. Furthermore, recurrence of that interrupt is suppressed whilst its interrupt handler chain is executing. This means that with respect to itself, an interrupt handler chain can be assumed to have a full memory barrier between executions, no matter if it migrates to another CPU between invocations. David ------------------------------------------------------------------------- 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
