Re: [PATCH] x86: Distinguish TLB shootdown interrupts from other functions call interrupts

[Alex Shi <alex.shi@intel.com>]

On 09/26/2012 10:11 AM, Tomoki Sekiyama wrote:

> Hi Alex,
> 
> On 2012/09/25 11:57, Alex Shi wrote:
>> On 09/24/2012 09:37 AM, Alex Shi wrote:
>>
>>> On 09/20/2012 04:50 PM, Tomoki Sekiyama wrote:
>>>
>>>> 	unsigned int irq_resched_count;
>>>> 	unsigned int irq_call_count;
>>>> +	/* irq_tlb_count is double-counted in irq_call_count, so it must be
>>>> +	   subtracted from irq_call_count when displaying irq_call_count */
>>>> 	unsigned int irq_tlb_count;
>>>
>>> Review again this patch, above comments is not kernel compatible format.
>>> Could you change it like standard comment format:
>>>
>>> /*
>>>  * xxxxxxx
>>>  * xxxx
>>>  */
>>>
>>
>> the 3.6 kernel will closed soon. it will be great to has this patch in.
>> So, could you like to refresh your patch with popular comments format? :)
> 
> Fixed patch is below.
> Thank you for the review again.
> 


Acked-by: Alex Shi <alex.shi@intel.com>

> --
> As TLB shootdown requests to other CPU cores are now using function call
> interrupts, TLB shootdowns entry in /proc/interrupts is always shown as 0.
> 
> This behavior change was introduced by commit 52aec3308db8 ("x86/tlb:
> replace INVALIDATE_TLB_VECTOR by CALL_FUNCTION_VECTOR").
> 
> This patch reverts TLB shootdowns entry in /proc/interrupts to count TLB
> shootdowns separately from the other function call interrupts.
> 
> Signed-off-by: Tomoki Sekiyama <tomoki.sekiyama.qu@hitachi.com>
> Cc: Thomas Gleixner <tglx@linutronix.de>
> Cc: Ingo Molnar <mingo@redhat.com>
> Cc: "H. Peter Anvin" <hpa@zytor.com>
> Cc: Alex Shi <alex.shi@intel.com>
> ---
>  arch/x86/include/asm/hardirq.h |    4 ++++
>  arch/x86/kernel/irq.c          |    4 ++--
>  arch/x86/mm/tlb.c              |    2 ++
>  3 files changed, 8 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/x86/include/asm/hardirq.h b/arch/x86/include/asm/hardirq.h
> index d3895db..81f04ce 100644
> --- a/arch/x86/include/asm/hardirq.h
> +++ b/arch/x86/include/asm/hardirq.h
> @@ -18,6 +18,10 @@ typedef struct {
>  #ifdef CONFIG_SMP
>  	unsigned int irq_resched_count;
>  	unsigned int irq_call_count;
> +	/*
> +	 * irq_tlb_count is double-counted in irq_call_count, so it must be
> +	 * subtracted from irq_call_count when displaying irq_call_count
> +	 */
>  	unsigned int irq_tlb_count;
>  #endif
>  #ifdef CONFIG_X86_THERMAL_VECTOR
> diff --git a/arch/x86/kernel/irq.c b/arch/x86/kernel/irq.c
> index d44f782..e4595f1 100644
> --- a/arch/x86/kernel/irq.c
> +++ b/arch/x86/kernel/irq.c
> @@ -92,7 +92,8 @@ int arch_show_interrupts(struct seq_file *p, int prec)
>  	seq_printf(p, "  Rescheduling interrupts\n");
>  	seq_printf(p, "%*s: ", prec, "CAL");
>  	for_each_online_cpu(j)
> -		seq_printf(p, "%10u ", irq_stats(j)->irq_call_count);
> +		seq_printf(p, "%10u ", irq_stats(j)->irq_call_count -
> +					irq_stats(j)->irq_tlb_count);
>  	seq_printf(p, "  Function call interrupts\n");
>  	seq_printf(p, "%*s: ", prec, "TLB");
>  	for_each_online_cpu(j)
> @@ -147,7 +148,6 @@ u64 arch_irq_stat_cpu(unsigned int cpu)
>  #ifdef CONFIG_SMP
>  	sum += irq_stats(cpu)->irq_resched_count;
>  	sum += irq_stats(cpu)->irq_call_count;
> -	sum += irq_stats(cpu)->irq_tlb_count;
>  #endif
>  #ifdef CONFIG_X86_THERMAL_VECTOR
>  	sum += irq_stats(cpu)->irq_thermal_count;
> diff --git a/arch/x86/mm/tlb.c b/arch/x86/mm/tlb.c
> index 613cd83..2d6d8ed 100644
> --- a/arch/x86/mm/tlb.c
> +++ b/arch/x86/mm/tlb.c
> @@ -98,6 +98,8 @@ static void flush_tlb_func(void *info)
>  {
>  	struct flush_tlb_info *f = info;
>  
> +	inc_irq_stat(irq_tlb_count);
> +
>  	if (f->flush_mm != this_cpu_read(cpu_tlbstate.active_mm))
>  		return;
>  
> 



-- 
Thanks
    Alex