linux-kernel - Re: [PATCH] x86/fpu: Revert earlier patch of Disable AVX when eagerfpu is off

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:	Wed, 9 Mar 2016 10:51:45 -0800
From:	"H. Peter Anvin" <hpa@...or.com>
To:	Ingo Molnar <mingo@...nel.org>, Yu-cheng Yu <yu-cheng.yu@...el.com>
Cc:	linux-kernel@...r.kernel.org, x86@...nel.org,
	Thomas Gleixner <tglx@...utronix.de>,
	Dave Hansen <dave.hansen@...ux.intel.com>,
	Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...e.de>,
	Sai Praneeth Prakhya <sai.praneeth.prakhya@...el.com>,
	"Ravi V. Shankar" <ravi.v.shankar@...el.com>,
	Leonid Shatz <leonid.shatz@...ellosystems.com>,
	Fenghua Yu <fenghua.yu@...el.com>,
	Andy Lutomirski <luto@...nel.org>,
	Linus Torvalds <torvalds@...ux-foundation.org>,
	Andrew Morton <akpm@...ux-foundation.org>,
	Peter Zijlstra <a.p.zijlstra@...llo.nl>
Subject: Re: [PATCH] x86/fpu: Revert earlier patch of Disable AVX when
 eagerfpu is off

On 03/09/2016 04:46 AM, Ingo Molnar wrote:
> 
> * Yu-cheng Yu <yu-cheng.yu@...el.com> wrote:
> 
>> AVX was mistakenly believed to be dependent on eagerfpu switch.
>> This turns out to be false. The earlier patch should be reverted.
>>
>> Original patch:
>>     http://git.kernel.org/tip/394db20ca240741a08d472173db13d6f6a6e5a28
> 
> So the original patch had a whole host of explanations of why that change is 
> correct. This revert should explain where that argumentation was wrong.
> 
> Also note that eagerfpu=off is on the way out, on v4.6 we'll switch all CPUs to 
> eagerfpu:
> 
>   58122bf1d856 x86/fpu: Default eagerfpu=on on all CPUs
> 
> and in the not so distant future, if everything goes fine with the eager mode, I'd 
> like to drop the lazy FPU context switching code altogether - which will simplify 
> a lot of code. At that point the 'eagerfpu' boot option will go away as well.
> 

Last I heard the only use case which was seeing trouble with eagerfpu
was KVM during service of I/O events.  Might be worth checking into...

	-hpa