Re: PCI MSI breaks when booting with nosmp

[Jean Delvare]

Hi Jesse,

Le lundi 21 avril 2008, Jesse Barnes a écrit :
> On Monday, April 21, 2008 2:09 pm Andi Kleen wrote:
> > > Both call disable_ioapic_setup(), so how can they not be the same?
> >
> > See Jesse's explanation. Basically there is special code in the smp
> > boot up to handle 0 CPUs and it disables all APICs.
> >
> > This is very old and crufty and somewhat obsolete logic,
> > but it is like that.
> 
> This might be closer to a real fix.  Can you confirm that this works for you 
> Jean?
> 
> Thanks,
> Jesse
> 
> diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c
> index e6abe8a..7cdf930 100644
> --- a/arch/x86/kernel/smpboot.c
> +++ b/arch/x86/kernel/smpboot.c
> @@ -1138,14 +1138,10 @@ static int __init smp_sanity_check(unsigned max_cpus)
>  				 "forcing use of dummy APIC emulation.\n");
>  		smpboot_clear_io_apic();
>  #ifdef CONFIG_X86_32
> -		if (nmi_watchdog == NMI_LOCAL_APIC) {
> -			printk(KERN_INFO "activating minimal APIC for"
> -					 "NMI watchdog use.\n");
> -			connect_bsp_APIC();
> -			setup_local_APIC();
> -			end_local_APIC_setup();
> -		}
> +		connect_bsp_APIC();
>  #endif
> +		setup_local_APIC();
> +		end_local_APIC_setup();
>  		return -1;
>  	}
>  
> 

Sorry for the late reply. I tested the patch above on top of 2.6.25-git4
and yes, it fixes my problem: booting with "nosmp" works now. Thanks!

Is this patch good enough to go upstream, and if so, through whose
tree? Adding Thomas Gleixner to Cc.

-- 
Jean Delvare
Suse L3
--
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/