Re: [PATCH 1/2] IRQ: fix performance regression on large IA64 systems

[Jiri Slaby]

On 07/05/2009 09:26 PM, Thomas Gleixner wrote:
> On Fri, 3 Jul 2009, Jiri Slaby wrote:
>> Commit b60c1f6ffd88850079ae419aa933ab0eddbd5535
>> (drop note_interrupt() for per-CPU for proper scaling) removed call to
>> note_interrupt() in __do_IRQ(). Commit
>> d85a60d85ea5b7c597508c1510c88e657773d378
>> (Add IRQF_IRQPOLL flag (common code)) added it again, because it's needed
>> for irqpoll.
>>
>> This patch now introduces a new parameter 'only_fixup' for
>> note_interrupt(). This parameter determines two cases:
...
> Hmm. I'm not really happy about that patch. It's all about percpu
> interrupts which happen to have the same irq number. I think the
> correct thing to do is to use the handle_percpu_irq() handler and
> modify handle_percpu_irq to call note_interrupt() only when the return
> value of the action handler is IRQ_NONE. Otherwise we can leave
> everything untouched.

Sorry for the delay, I waited for test results. I have the patch
attached and the boot still lasts very long. Do you see any obvious
mistake which I don't see? (It's 2.6.27-based, therefore no irq_to_desc.)

Thanks.

---
 arch/ia64/kernel/irq_ia64.c |    9 ++++-----
 kernel/irq/chip.c           |    2 +-
 2 files changed, 5 insertions(+), 6 deletions(-)

--- a/arch/ia64/kernel/irq_ia64.c
+++ b/arch/ia64/kernel/irq_ia64.c
@@ -624,14 +624,13 @@ static struct irqaction tlb_irqaction =
 void
 ia64_native_register_percpu_irq (ia64_vector vec, struct irqaction *action)
 {
-       irq_desc_t *desc;
-       unsigned int irq;
+       unsigned int irq = vec;
+       irq_desc_t *desc = irq_desc + irq;

-       irq = vec;
        BUG_ON(bind_irq_vector(irq, vec, CPU_MASK_ALL));
-       desc = irq_desc + irq;
        desc->status |= IRQ_PER_CPU;
-       desc->chip = &irq_type_ia64_lsapic;
+       set_irq_chip_and_handler(irq, &irq_type_ia64_lsapic,
+                       &handle_percpu_irq);
        if (action)
                setup_irq(irq, action);
 }
--- a/kernel/irq/chip.c
+++ b/kernel/irq/chip.c
@@ -531,7 +531,7 @@ handle_percpu_irq(unsigned int irq, stru
                desc->chip->ack(irq);

        action_ret = handle_IRQ_event(irq, desc->action);
-       if (!noirqdebug)
+       if (!noirqdebug && action_ret == IRQ_NONE)
                note_interrupt(irq, desc, action_ret);

        if (desc->chip->eoi)
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/