Re: [patch 1/1] Do not enforce unique IO_APIC_ID for Xeon processors in EM64T mode (x86_64)

[Andi Kleen <ak@suse.de>]

On Wed, May 11, 2005 at 11:21:15AM -0500, Protasevich, Natalie wrote:
> > > Looks like the need in the unique id can only be keyed of the local 
> > > APIC id, and probably it is a good idea to keep the NO_IOAPIC_CHECK 
> > > for subarchs that can override the heuristics?
> > 
> > I prefer not to do that. How about a simple
> > 
> > if (boot_cpu_data.x86_vendor == X86_VENDOR_INTEL && 
> > boot_cpu_data.x86 < 15)
> > 	/* do uniqueness check */
> > else
> > 	/* don't do it */
> > 
> > ?		
> > 
> > Rationale is that P4s and newer and systems not from Intel 
> > don't have serial APIC busses and don't need this uniqueness check.
> >
> 
> Yes, indeed this looks like the only undisputed (and sufficient)
> criteria. I tried the below with Xeon box and it worked fine:
> 
> --- mpparse.c.orig	2005-05-11 02:10:35.000000000 -0400
> +++ mpparse.c	2005-05-11 02:12:31.000000000 -0400
> @@ -912,7 +913,15 @@ void __init mp_register_ioapic (
>  	mp_ioapics[idx].mpc_apicaddr = address;
>  
>  	set_fixmap_nocache(FIX_IO_APIC_BASE_0 + idx, address);
> -	mp_ioapics[idx].mpc_apicid = io_apic_get_unique_id(idx, id);
> +	if ((boot_cpu_data.x86_vendor == X86_VENDOR_INTEL) &&
> (boot_cpu_data.x86 >= 15))
> +		mp_ioapics[idx].mpc_apicid = id;
> +	else

That's still wrong because it does not trigger on AMD systems. Please make it

	if (boot_cpu_data.x86_vendor == X86_VENDOR_INTEL && boot_cpu_data.x86 < 15)
		mp_ioapics[idx].mpc_apicid = io_apic_get_unique_id(idx,id);
	else
		mp_ioapics[idx].mpc_apicid = id;


-Andi