Re: [PATCH v6 07/11] x86/smpboot: Disable parallel boot for AMD CPUs

[David Woodhouse]

On Tue, 2023-02-07 at 01:23 +0100, Thomas Gleixner wrote:
> On Sat, Feb 04 2023 at 15:40, David Woodhouse wrote:
> > On Fri, 2023-02-03 at 13:48 -0600, Kim Phillips wrote:
> > > Then:
> > > 
> > >   - a Ryzen 3000 (Picasso A1/Zen+) notebook I have access to fails to boot.
> > >   - Zen 2,3,4-based servers boot fine
> > >   - a Zen1-based server doesn't boot.
> > 
> > I've changed it to use CPUID 0xb only if we're actually in x2apic mode,
> > which Boris tells me won't be the case on Zen1 because that doesn't
> > support X2APIC.
> 
> Correct.
> 
> > When we're not in x2apic mode, we can use CPUID 0x1 because the 8 bits
> > of APIC ID we find there are perfectly sufficient.
> 
> Is that worth the trouble?

Well, that's what was being debated. I think the conclusion that was
bring reached was that it *is* worth the trouble, because there will be
a number of physical and especially virtual machines which have a high
CPU count but which don't actually use X2APIC mode. And which might not
even *support* CPUID 0xb.

So using CPUID 0x1 when there is no APIC ID greater than 254 does seem
to make sense.


> > Even though we *can* support non-X2APIC processors, we *might* want to
> > play it safe and not go back that far; only enabling parallel bringup
> > on machines with X2APIC which roughly correlates with "lots of CPUs"
> > since that's where the benefit is.
> 
> The parallel bringup code is complex enough already, so please don't
> optimize for the non-interesting case in the first place. When this has
> stabilized then the CPUID 0x1 mechanism can be added if anyone thinks
> it's interesting. KISS is still the best engineering principle.

Actually it ends up being trivial. It probably makes sense to keep it
in there even if it can only be exercised by a deliberate opt-in on
older CPUs. I reworked the register usage from your original anyway,
which helps a little.

	testl	$STARTUP_APICID_CPUID_0B, %edx
	jnz	.Luse_cpuid_0b
	testl	$STARTUP_APICID_CPUID_01, %edx
	jnz	.Luse_cpuid_01
	andl	$0x0FFFFFFF, %edx
	jmp	.Lsetup_AP

.Luse_cpuid_01:
	mov	$0x01, %eax
	cpuid
	mov	%ebx, %edx
	shr	$24, %edx
	jmp	.Lsetup_AP

.Luse_cpuid_0b:
	mov	$0x0B, %eax
	xorl	%ecx, %ecx
	cpuid

.Lsetup_AP:
	/* EDX contains the APICID of the current CPU */

Attachment: smime.p7s
Description: S/MIME cryptographic signature