linux-kernel - Re: [PATCH v7 05/26] x86/fpu/xstate: Introduce fpu_guest

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

Message-ID: <c7c7295a-ae5e-420a-837e-5bc6e18ee90c@intel.com>
Date:   Thu, 30 Nov 2023 21:28:02 +0800
From:   "Yang, Weijiang" <weijiang.yang@...el.com>
To:     "Edgecombe, Rick P" <rick.p.edgecombe@...el.com>
CC:     "Gao, Chao" <chao.gao@...el.com>,
        "Hansen, Dave" <dave.hansen@...el.com>,
        "seanjc@...gle.com" <seanjc@...gle.com>,
        "peterz@...radead.org" <peterz@...radead.org>,
        "john.allen@....com" <john.allen@....com>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "mlevitsk@...hat.com" <mlevitsk@...hat.com>,
        "kvm@...r.kernel.org" <kvm@...r.kernel.org>,
        "pbonzini@...hat.com" <pbonzini@...hat.com>
Subject: Re: [PATCH v7 05/26] x86/fpu/xstate: Introduce fpu_guest_cfg for
 guest FPU configuration

On 11/30/2023 1:08 AM, Edgecombe, Rick P wrote:
> On Wed, 2023-11-29 at 22:12 +0800, Yang, Weijiang wrote:
>> On 11/28/2023 10:58 PM, Edgecombe, Rick P wrote:
>>> On Fri, 2023-11-24 at 00:53 -0500, Yang Weijiang wrote:
>>>> +       /*
>>>> +        * Set guest's __user_state_size to
>>>> fpu_user_cfg.default_size
>>>> so that
>>>> +        * existing uAPIs can still work.
>>>> +        */
>>>> +       fpu->guest_perm.__user_state_size =
>>>> fpu_user_cfg.default_size;
>>> It seems like an appropriate value, but where does this come into
>>> play
>>> exactly for guest FPUs?
>> I don't see there's special usage of this field for vCPU in VMM
>> userspace(QEMU).
>> Maybe it's mainly for AMX resulted usespace fault handling? For vCPU
>> thread,
>> it's only  referenced when AMX is enabled via __xfd_enable_feature()
>> .
>>
> In that case the "so that existing uAPIs can still work" comment seems
> misleading. Maybe "this doesn't come into play for guest FPUs, but set
> it to a reasonable value"?

Ah, I mistook it for uabi_size and added comments. Will reword it properly.
Thank you for bringing it up!