Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
To:     Peter Zijlstra <peterz@infradead.org>,
        Boris Ostrovsky <boris.ostrovsky@oracle.com>
Cc:     linux-kernel@vger.kernel.org, xen-devel@lists.xenproject.org,
        tglx@linutronix.de, bp@alien8.de, dave.hansen@linux.intel.com,
        x86@kernel.org, hpa@zytor.com, jgross@suse.com
References: <1635896196-18961-1-git-send-email-boris.ostrovsky@oracle.com>
 <YYk+D17oIOGKWNtN@hirez.programming.kicks-ass.net>
 <48fb48fa-c65d-8e38-dabb-cf9be21365ca@oracle.com>
 <YYqPd1c9HIQH9k/9@hirez.programming.kicks-ass.net>
From:   Andrew Cooper <amc96@srcf.net>
Subject: Re: [PATCH] x86/smp: Factor out parts of native_smp_prepare_cpus()
Message-ID: <aef8093d-f7da-c8b4-ebf6-6053b50b77ba@srcf.net>
Date:   Tue, 9 Nov 2021 15:28:32 +0000
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101
 Thunderbird/78.14.0
MIME-Version: 1.0
In-Reply-To: <YYqPd1c9HIQH9k/9@hirez.programming.kicks-ass.net>
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 8BIT
Content-Language: en-GB
Precedence: bulk

On 09/11/2021 15:10, Peter Zijlstra wrote:
> On Mon, Nov 08, 2021 at 12:20:26PM -0500, Boris Ostrovsky wrote:
>>> But looking at those functions; there seems to be more spurious
>>> differences. For example, the whole sched_topology thing.
>>
>> I did look at that and thought this should be benign given that Xen PV
>> is not really topology-aware. I didn't see anything that would be a
>> cause for concern but perhaps you can point me to things I missed.
> And me not being Xen aware... What does Xen-PV guests see of the CPUID
> topology fields? Does it fully sanitize the CPUID data, or is it a clean
> pass-through from whatever CPU the vCPU happens to run on at the time?

That depends on hardware support (CPUID Faulting or not), version of Xen
(anything before Xen 4.7 is totally insane.  Anything more recent is
only moderately insane), and whether the kernel asks via the enlightened
CPUID path or not.

On hardware lacking CPUID faulting, and for a kernel using
native_cpuid() where it ought to be using the PVOP, it sees the real
hardware value of the CPU it happens to be running on.

~Andrew