Re: [PATCH] x86: Enable NMI on all cpus on UV

[Ingo Molnar]

* Russ Anderson <rja@xxxxxxx> wrote:

> Enable NMI on all cpus in UV system and add an NMI handler
> to dump_stack on each cpu.
> 
> Signed-off-by: Russ Anderson <rja@xxxxxxx>
> 
> ---
> 
> By default on x86 all the cpus except the boot cpu have NMI
> masked off.  This patch enables NMI on all cpus in UV system
> and adds an NMI handler to dump_stack on each cpu.  This
> way if a system hangs we can NMI the machine and get a 
> backtrace from all the cpus.
> 
> 
>  arch/x86/include/asm/uv/uv.h       |    1 
>  arch/x86/kernel/apic/x2apic_uv_x.c |   49 +++++++++++++++++++++++++++++++++++++
>  arch/x86/kernel/smpboot.c          |    2 +
>  3 files changed, 52 insertions(+)
> 
> Index: linux/arch/x86/kernel/apic/x2apic_uv_x.c
> ===================================================================
> --- linux.orig/arch/x86/kernel/apic/x2apic_uv_x.c	2010-02-17 10:21:55.000000000 -0600
> +++ linux/arch/x86/kernel/apic/x2apic_uv_x.c	2010-02-17 10:32:20.000000000 -0600
> @@ -20,6 +20,7 @@
>  #include <linux/cpu.h>
>  #include <linux/init.h>
>  #include <linux/io.h>
> +#include <linux/kdebug.h>
>  
>  #include <asm/uv/uv_mmrs.h>
>  #include <asm/uv/uv_hub.h>
> @@ -39,6 +40,53 @@ static u64 gru_start_paddr, gru_end_padd
>  int uv_min_hub_revision_id;
>  EXPORT_SYMBOL_GPL(uv_min_hub_revision_id);
>  
> +int uv_handle_nmi(struct notifier_block *self,
> +		  unsigned long reason, void *data)
> +{
> +	unsigned long flags;
> +	static DEFINE_SPINLOCK(uv_nmi_lock);
> +
> +	if (reason != DIE_NMI_IPI)
> +		return NOTIFY_OK;
> +	/*
> +	 * Use a lock so only one cpu prints at a time
> +	 * to prevent intermixed output.
> +	 */
> +	spin_lock_irqsave(&uv_nmi_lock, flags);
> +	printk(KERN_INFO "NMI stack dump cpu %u:\n",
> +				smp_processor_id());
> +	dump_stack();
> +	spin_unlock_irqrestore(&uv_nmi_lock, flags);
> +
> +	return NOTIFY_STOP;
> +}
> +
> +static struct notifier_block uv_dump_stack_nmi_nb = {
> +  .notifier_call = uv_handle_nmi,
> +  .next = NULL,
> +  .priority = 0
> +};
> +
> +void uv_register_nmi_notifier(void)
> +{
> +	if (register_die_notifier(&uv_dump_stack_nmi_nb))
> +		printk(KERN_WARNING "UV NMI handler failed to register\n");
> +}
> +
> +/*
> + * Called on each cpu to unmask NMI.
> + */
> +void __cpuinit uv_nmi_init(void)
> +{
> +	unsigned int value;
> +
> +	/*
> +	 * Unmask NMI on all cpus
> +	 */
> +	value = apic_read(APIC_LVT1) | APIC_DM_NMI;
> +	value &= ~APIC_LVT_MASKED;
> +	apic_write(APIC_LVT1, value);
> +}
>  
>  static int is_GRU_range(u64 start, u64 end)
>  {
> @@ -718,5 +766,6 @@ void __init uv_system_init(void)
>  
>  	uv_cpu_init();
>  	uv_scir_register_cpu_notifier();
> +	uv_register_nmi_notifier();
>  	proc_mkdir("sgi_uv", NULL);
>  }
> Index: linux/arch/x86/include/asm/uv/uv.h
> ===================================================================
> --- linux.orig/arch/x86/include/asm/uv/uv.h	2010-02-17 10:21:55.000000000 -0600
> +++ linux/arch/x86/include/asm/uv/uv.h	2010-02-17 10:32:20.000000000 -0600
> @@ -11,6 +11,7 @@ struct mm_struct;
>  extern enum uv_system_type get_uv_system_type(void);
>  extern int is_uv_system(void);
>  extern void uv_cpu_init(void);
> +extern void uv_nmi_init(void);
>  extern void uv_system_init(void);
>  extern const struct cpumask *uv_flush_tlb_others(const struct cpumask *cpumask,
>  						 struct mm_struct *mm,
> Index: linux/arch/x86/kernel/smpboot.c
> ===================================================================
> --- linux.orig/arch/x86/kernel/smpboot.c	2010-02-17 10:21:55.000000000 -0600
> +++ linux/arch/x86/kernel/smpboot.c	2010-02-17 10:32:20.000000000 -0600
> @@ -320,6 +320,8 @@ notrace static void __cpuinit start_seco
>  	unlock_vector_lock();
>  	ipi_call_unlock();
>  	per_cpu(cpu_state, smp_processor_id()) = CPU_ONLINE;
> +	if (is_uv_system())
> +		uv_nmi_init();

Instead of cramming it into the init sequence open-coded, shouldnt this be 
done via the x86_platform driver mechanism?

Thanks,

	Ingo
--
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/