On Fri, 2012-04-27 at 01:50 +0200, Alexander Graf wrote: > Yup. We're running the entire guest with IRQs off. Do you have any better idea how to make sure that we're atomic wrt signal delivery? Something like this entry path: - local_irq_disable - check TIF_NEED_RESCHED, TIF_SIGPENDING,... (probably _TIF_USER_WORK_MASK in fact) - if any set, re-enable then -> back to qemu - ppc64: hard disable - ppc64: check irq_pending, something set (other than hard disable) -> local_irq_enable() and try again the whole sequence - call trace_irq_enable() (tell lockdep/irqtrace we are re-enabling) - enter the guest, last rfi will turn EE on And exit: - irqs are hard off from the interrupt handler - call local_irq_disable() to tell lockdep about it and mark soft-disabled on ppc64 - ppc64: maybe set PACA_HARD_IRQ_DIS (tbd) - do things like do_IRQ() etc... if needed - local_irq_enable() (will hard enable if PACA_HARD_IRQ_DIS was set) Cheers, Ben. -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
