Re: [patch 2/2] x86: mce: Implement cmci poll mode for intel machines

[Chen Gong <gong.chen@linux.intel.com>]

Hi, Thomas

I have some confusion in your patch please help to give me some updates.

于 2012/5/25 1:54, Thomas Gleixner 写道:
> Intentionally left blank to be filled out by someone who wants that and
> can explain the reason for this better than me.
> 
> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
> ---
>  arch/x86/kernel/cpu/mcheck/mce-internal.h |    8 ++
>  arch/x86/kernel/cpu/mcheck/mce.c          |   41 +++++++++++++--
>  arch/x86/kernel/cpu/mcheck/mce_intel.c    |   81 +++++++++++++++++++++++++++++-
>  3 files changed, 125 insertions(+), 5 deletions(-)
> 
> Index: linux-2.6/arch/x86/kernel/cpu/mcheck/mce-internal.h
> ===================================================================
> --- linux-2.6.orig/arch/x86/kernel/cpu/mcheck/mce-internal.h
> +++ linux-2.6/arch/x86/kernel/cpu/mcheck/mce-internal.h
> @@ -28,6 +28,14 @@ extern int mce_ser;
>  
>  extern struct mce_bank *mce_banks;
>  
> +#ifdef CONFIG_X86_MCE_INTEL
> +unsigned long mce_intel_adjust_timer(unsigned long interval);
> +#else
> +# define mce_intel_adjust_timer mce_adjust_timer_default
> +#endif
> +
> +void mce_timer_kick(unsigned long interval);
> +
>  #ifdef CONFIG_ACPI_APEI
>  int apei_write_mce(struct mce *m);
>  ssize_t apei_read_mce(struct mce *m, u64 *record_id);
> Index: linux-2.6/arch/x86/kernel/cpu/mcheck/mce.c
> ===================================================================
> --- linux-2.6.orig/arch/x86/kernel/cpu/mcheck/mce.c
> +++ linux-2.6/arch/x86/kernel/cpu/mcheck/mce.c
> @@ -1242,6 +1242,14 @@ static unsigned long check_interval = 5 
>  static DEFINE_PER_CPU(unsigned long, mce_next_interval); /* in jiffies */
>  static DEFINE_PER_CPU(struct timer_list, mce_timer);
>  
> +static unsigned long mce_adjust_timer_default(unsigned long interval)
> +{
> +	return interval;
> +}
> +
> +static unsigned long (*mce_adjust_timer)(unsigned long interval) =
> +	mce_adjust_timer_default;
> +
>  static void mce_timer_fn(unsigned long data)
>  {
>  	struct timer_list *t = &__get_cpu_var(mce_timer);
> @@ -1259,14 +1267,37 @@ static void mce_timer_fn(unsigned long d
>  	 * polling interval, otherwise increase the polling interval.
>  	 */
>  	iv = __this_cpu_read(mce_next_interval);
> -	if (mce_notify_irq())
> +	if (mce_notify_irq()) {
>  		iv = max(iv, (unsigned long) HZ/100);
> -	else
> +	} else {
>  		iv = min(iv * 2, round_jiffies_relative(check_interval * HZ));
> +		iv = mce_adjust_timer(iv);
> +	}
>  	__this_cpu_write(mce_next_interval, iv);
> +	/* Might have become 0 after CMCI storm subsided */
> +	if (iv) {
> +		t->expires = jiffies + iv;
> +		add_timer_on(t, smp_processor_id());
> +	}
> +}
>  
> -	t->expires = jiffies + iv;
> -	add_timer_on(t, smp_processor_id());
> +/*
> + * Ensure that the timer is firing in @interval from now.
> + */
> +void mce_timer_kick(unsigned long interval)
> +{
> +	struct timer_list *t = &__get_cpu_var(mce_timer);
> +	unsigned long when = jiffies + interval;
> +	unsigned long iv = __this_cpu_read(mce_next_interval);
> +
> +	if (time_before(when, t->expires) && timer_pending(t)) {
> +		mod_timer(t, when);
> +	} else if (!mce_next_interval) {

Why using mce_next_interval, it is a per_cpu var, should be non-NULL
definitely, right? How about using *iv* here?

> +		t->expires = round_jiffies(jiffies + iv);
> +		add_timer_on(t, smp_processor_id());
> +	}
> +	if (interval < iv)
> +		__this_cpu_write(mce_next_interval, iv);
>  }

This code should be __this_cpu_write(mce_next_interval, interval);?

>  
>  /* Must not be called in IRQ context where del_timer_sync() can deadlock */
> @@ -1531,6 +1562,7 @@ static void __mcheck_cpu_init_vendor(str
>  	switch (c->x86_vendor) {
>  	case X86_VENDOR_INTEL:
>  		mce_intel_feature_init(c);
> +		mce_adjust_timer = mce_intel_adjust_timer;
>  		break;
>  	case X86_VENDOR_AMD:
>  		mce_amd_feature_init(c);
> @@ -1550,6 +1582,7 @@ static void __mcheck_cpu_init_timer(void
>  	if (mce_ignore_ce)
>  		return;
>  
> +	iv = mce_adjust_timer(check_interval * HZ);
>  	__this_cpu_write(mce_next_interval, iv);
>  	if (!iv)
>  		return;