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;
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
From:   Andy Lutomirski <luto@amacapital.net>
Mime-Version: 1.0 (1.0)
Subject: Re: [PATCH v2 1/4] x86/fpu: Add kernel_fpu_begin_mask() to selectively initialize state
Date:   Wed, 20 Jan 2021 04:19:24 -0800
Message-Id: <F60CEFE2-22A4-405A-9BBC-7D84AFDA07C5@amacapital.net>
References: <20210120115324.GE825@zn.tnic>
Cc:     Andy Lutomirski <luto@kernel.org>, x86@kernel.org,
        LKML <linux-kernel@vger.kernel.org>,
        Krzysztof Mazur <krzysiek@podlesie.net>,
        =?utf-8?Q?Krzysztof_Ol=C4=99dzki?= <ole@ans.pl>,
        Arnd Bergmann <arnd@arndb.de>
In-Reply-To: <20210120115324.GE825@zn.tnic>
To:     Borislav Petkov <bp@alien8.de>
Precedence: bulk


> On Jan 20, 2021, at 3:53 AM, Borislav Petkov <bp@alien8.de> wrote:
>=20
> =EF=BB=BFOn Tue, Jan 19, 2021 at 09:38:59AM -0800, Andy Lutomirski wrote:
>> Currently, requesting kernel FPU access doesn't distinguish which parts o=
f
>> the extended ("FPU") state are needed.  This is nice for simplicity, but
>> there are a few cases in which it's suboptimal:
>>=20
>> - The vast majority of in-kernel FPU users want XMM/YMM/ZMM state but do
>>   not use legacy 387 state.  These users want MXCSR initialized but don't=

>>   care about the FPU control word.  Skipping FNINIT would save time.
>>   (Empirically, FNINIT is several times slower than LDMXCSR.)
>>=20
>> - Code that wants MMX doesn't want or need MXCSR initialized.
>>   _mmx_memcpy(), for example, can run before CR4.OSFXSR gets set, and
>>   initializing MXCSR will fail.
>>=20
>> - Any future in-kernel users of XFD (eXtended Feature Disable)-capable
>>   dynamic states will need special handling.
>>=20
>> This patch adds a more specific API that allows callers specify exactly
>> what they want.
>=20
> Same nitpicks:
>=20
> https://lkml.kernel.org/r/20210119110834.GH27433@zn.tnic

I would have sworn I fixed those.  Sorry!

>=20
> --=20
> Regards/Gruss,
>    Boris.
>=20
> https://people.kernel.org/tglx/notes-about-netiquette