linux-kernel - Re: [PATCH v3 13/17] x86/xen: enable Hygon support to Xen

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Fri, 17 Aug 2018 00:37:53 -0400
From:   Boris Ostrovsky <boris.ostrovsky@...cle.com>
To:     Pu Wen <puwen@...on.cn>, Juergen Gross <jgross@...e.com>,
        tglx@...utronix.de, mingo@...hat.com, hpa@...or.com,
        x86@...nel.org, thomas.lendacky@....com, bp@...en8.de,
        pbonzini@...hat.com, JBeulich@...e.com
Cc:     linux-kernel@...r.kernel.org, linux-arch@...r.kernel.org,
        xen-devel@...ts.xenproject.org
Subject: Re: [PATCH v3 13/17] x86/xen: enable Hygon support to Xen

On 08/16/2018 09:29 AM, Pu Wen wrote:
> On 2018/8/12 21:26, Boris Ostrovsky wrote:
>> On 08/12/2018 04:55 AM, Juergen Gross wrote:
>>> On 11/08/18 16:34, Boris Ostrovsky wrote:
>>>> On 08/11/2018 09:29 AM, Pu Wen wrote:
>>>>>   bool pmu_msr_read(unsigned int msr, uint64_t *val, int *err)
>>>>>   {
>>>>> -    if (boot_cpu_data.x86_vendor == X86_VENDOR_AMD) {
>>>>> +    if (boot_cpu_data.x86_vendor == X86_VENDOR_AMD ||
>>>>> +        boot_cpu_data.x86_vendor == X86_VENDOR_HYGON) {
>>>>
>>>> 'if (boot_cpu_data.x86_vendor != X86_VENDOR_INTEL)' please.
>>> Really? Xen supports Centaur, too.
>>
>> VPMU doesn't --- hypervisor will not initialize it. Besides, the
>> existing code will steer non-AMD execution to Intel, which is not right
>> either.
>>
>> I'll add a check to bail if VPMU is not initialized properly, we seem to
>> ignore xen_pmu_init() failures. Which, BTW, makes this patch rather
>> pointless until there is support for Hygon Xen.
>
> So should it still need to test vendor Hygon here or wait for your check
> done?


I'd prefer checking for !Intel, as I suggested above. Centaur will fail
either way, but because we use safe versions of MSR access I now think
we don't need any extra checks for xen_pmu_init() result.

-boris