Subject: Re: [PATCH] x86/cpuid: Deal with broken firmware once more
To: Peter Zijlstra <peterz@infradead.org>,
        Thomas Gleixner <tglx@linutronix.de>
References: <20161102122557.qs4rl6mb7n7l7j7p@linutronix.de>
 <24e69019-60d0-29e7-e31f-c6f00f9ed98a@brocade.com>
 <alpine.DEB.2.20.1611071717000.3709@nanos>
 <58e229e2-91f4-a97f-1b9f-089f48ef994a@brocade.com>
 <alpine.DEB.2.20.1611072117040.3501@nanos>
 <e3a6f9cb-ee46-6e5c-7b2e-9a6ed2fc949e@brocade.com>
 <alpine.DEB.2.20.1611081525160.3501@nanos>
 <86609338-2b45-ed7e-fb07-99421e43a2f1@brocade.com>
 <alpine.DEB.2.20.1611081721330.3501@nanos>
 <alpine.DEB.2.20.1611091613540.3501@nanos>
 <20161109160324.GR3142@twins.programming.kicks-ass.net>
CC: Sebastian Andrzej Siewior <bigeasy@linutronix.de>, <x86@kernel.org>,
        LKML <linux-kernel@vger.kernel.org>,
        "M. Vefa Bicakci" <m.v.b@runbox.com>, Borislav Petkov <bp@alien8.de>
From: "Charles (Chas) Williams" <ciwillia@brocade.com>
Message-ID: <9badae12-f07e-b2e3-c1ce-c2017cd5e510@brocade.com>
Date: Wed, 9 Nov 2016 11:34:52 -0500
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101
 Icedove/45.3.0
MIME-Version: 1.0
In-Reply-To: <20161109160324.GR3142@twins.programming.kicks-ass.net>
Content-Type: text/plain; charset="windows-1252"; format=flowed
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1527
Lines: 42


On 11/09/2016 11:03 AM, Peter Zijlstra wrote:
> On Wed, Nov 09, 2016 at 04:35:51PM +0100, Thomas Gleixner wrote:
>> Both ACPI and MP specifications require that the APIC id in the respective
>> tables must be the same as the APIC id in CPUID.
>>
>> The kernel retrieves the physical package id from the APIC id during the
>> ACPI/MP table scan and builds the physical to logical package map.
>>
>> There exist Virtualbox and Xen implementations which violate the spec. As a
>
> ISTR it was VMware, not VirtualBox, but whatever.. they're both crazy
> virt stuff.

Yes, this was VMware in particular.  It would be good to get this comment
right so as not to mislead anyone.

>>  /*
>> + * The physical to logical package id mapping is initialized from the
>> + * acpi/mptables information. Make sure that CPUID actually agrees with
>> + * that.
>> + */
>> +static void sanitize_package_id(struct cpuinfo_x86 *c)
>> +{
>> +#ifdef CONFIG_SMP
>> +	unsigned int pkg, apicid, cpu = smp_processor_id();
>> +
>> +	apicid = apic->cpu_present_to_apicid(cpu);
>> +	pkg = apicid >> boot_cpu_data.x86_coreid_bits;
>> +
>> +	if (apicid != c->initial_apicid) {
>> +		pr_err(FW_BUG "CPU%u: APIC id mismatch. Firmware: %x CPUID: %x\n",
>> +		       cpu, apicid, c->initial_apicid);
>
> Should we not also 'fix' c->initial_apicid ?

Since we have c->apicid and c->initial_apicid it seems reasonable to keep one
set to the "correct" value.  I don't think c->initial_apicid is used past
this.

I should have some tests on this patch later today.