Re: [patch] x86: 2.6.31-rc7 crash due to buggy flat_phys_pkg_id

[Ingo Molnar <mingo@elte.hu>]

* Yinghai Lu <yinghai@kernel.org> wrote:

> Ingo Molnar wrote:
> > * Linus Torvalds <torvalds@linux-foundation.org> wrote:
> > 
> >> On Tue, 25 Aug 2009, Yinghai Lu wrote:
> >>> initial apic id and apic id could be different.
> >>>
> >>> and we should use initial apic id to get correct phys pkg id in 
> >>> case BIOS set crazy apic id.
> >> Yinghai - I think you missed Cyrills' point.  Let me repeat it:
> >>
> >> 	"cpu_has_apic bit turned off"
> >>
> >> there's no apic. No "initial apic id". No "phys pkg id". No 
> >> nothing.
> >>
> >> Discussions about "correct phys pkg id" are pointless.
> > 
> > that's not the case here though:
> > 
> > [    8.713916] Total of 32 processors activated (162314.96 BogoMIPS).
> > 
> > so APICs are active. The real difference is i think this aspect of 
> > commit 2759c3287:
> > 
> >  static int flat_phys_pkg_id(int initial_apic_id, int index_msb)
> >  {
> > -       return hard_smp_processor_id() >> index_msb;
> > +       return initial_apic_id >> index_msb;
> >  }
> > 
> > We need to revert back to .30 behavior here. (In case of which 
> > environment to trust we almost always trust whatever booted millions 
> > of Linux boxes in the past already.)
> > 
> > Furthermore, commit 2759c3287 did not declare any side-effects and 
> > clearly causes a side-effect on vSMP which apparently has an 
> > overlapping set of initial APIC ids.
> > 
> > Ravikiran, your patch does not do a clear revert of this bit though. 
> > If you do a plain revert of the line above alone, does that fix the 
> > problem too?
> 
> how about patch phys_pkg_id for vsmp?
> 
> diff --git a/arch/x86/kernel/apic/probe_64.c b/arch/x86/kernel/apic/probe_64.c
> index f3b1037..65edc18 100644
> --- a/arch/x86/kernel/apic/probe_64.c
> +++ b/arch/x86/kernel/apic/probe_64.c
> @@ -44,6 +44,11 @@ static struct apic *apic_probe[] __initdata = {
>  	NULL,
>  };
>  
> +static int apicid_phys_pkg_id(int initial_apic_id, int index_msb)
> +{
> +	return hard_smp_processor_id() >> index_msb;
> +}
> +
>  /*
>   * Check the APIC IDs in bios_cpu_apicid and choose the APIC mode.
>   */
> @@ -69,6 +74,11 @@ void __init default_setup_apic_routing(void)
>  		printk(KERN_INFO "Setting APIC routing to %s\n", apic->name);
>  	}
>  
> +	if (is_vsmp_box()) {
> +		/* need to update phys_pkg_id */
> +		apic->phys_pkg_id = apicid_phys_pkg_id;
> +	}

Hm, this is rather tempting simply because it only affects vSMP 
systems and we are in late -rc's. Ravikiran, does Yinghai's patch 
solve the crash for you too?

	Ingo