[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [Xen-devel] [PATCH 38 of 38] xen: make sure that softirqs get handled at the end of event processing
Make sure that irq_enter()/irq_exit() wrap the entire event processing loop, rather than each individual event invokation. This make sure that softirq processing is deferred until the of event processing, rather than in the middle with interrupts disabled. Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@xxxxxxxxxx> --- drivers/xen/events.c | 29 +++++++++-------------------- 1 file changed, 9 insertions(+), 20 deletions(-) diff --git a/drivers/xen/events.c b/drivers/xen/events.c --- a/drivers/xen/events.c +++ b/drivers/xen/events.c @@ -806,25 +806,6 @@ return IRQ_HANDLED; } - -static void xen_do_irq(unsigned irq, struct pt_regs *regs) -{ - struct pt_regs *old_regs = set_irq_regs(regs); - - if (WARN_ON(irq == -1)) - return; - - exit_idle(); - irq_enter(); - - //printk("cpu %d handling irq %d\n", smp_processor_id(), info->irq); - handle_irq(irq, regs); - - irq_exit(); - - set_irq_regs(old_regs); -} - /* * Search the CPUs pending events bitmasks. For each one found, map * the event number to an irq, and feed it into do_IRQ() for @@ -837,11 +818,15 @@ void xen_evtchn_do_upcall(struct pt_regs *regs) { int cpu = get_cpu(); + struct pt_regs *old_regs = set_irq_regs(regs); struct shared_info *s = HYPERVISOR_shared_info; struct vcpu_info *vcpu_info = __get_cpu_var(xen_vcpu); static DEFINE_PER_CPU(unsigned, nesting_count); unsigned count; + exit_idle(); + irq_enter(); + do { unsigned long pending_words; @@ -865,7 +850,8 @@ int port = (word_idx * BITS_PER_LONG) + bit_idx; int irq = evtchn_to_irq[port]; - xen_do_irq(irq, regs); + if (irq != -1) + handle_irq(irq, regs); } } @@ -876,6 +862,9 @@ } while(count != 1); out: + irq_exit(); + set_irq_regs(old_regs); + put_cpu(); } _______________________________________________ Xen-devel mailing list Xen-devel@xxxxxxxxxxxxxxxxxxx http://lists.xensource.com/xen-devel
|
Lists.xenproject.org is hosted with RackSpace, monitoring our |