linux-kernel - Re: [PATCH AUTOSEL 5.10 09/10] x86/barrier: Do not serialize MSR accesses on AMD

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: Tue, 30 Jan 2024 15:58:21 -0500
From: Sasha Levin <sashal@...nel.org>
To: Pavel Machek <pavel@...x.de>
Cc: linux-kernel@...r.kernel.org, stable@...r.kernel.org,
	"Borislav Petkov (AMD)" <bp@...en8.de>, tglx@...utronix.de,
	mingo@...hat.com, dave.hansen@...ux.intel.com, x86@...nel.org,
	puwen@...on.cn, seanjc@...gle.com, kim.phillips@....com,
	reinette.chatre@...el.com, babu.moger@....com, jmattson@...gle.com,
	peterz@...radead.org, ashok.raj@...el.com,
	rick.p.edgecombe@...el.com, brgerst@...il.com, mjguzik@...il.com,
	jpoimboe@...nel.org, nik.borisov@...e.com, aik@....com,
	vegard.nossum@...cle.com, daniel.sneddon@...ux.intel.com,
	acdunlap@...gle.com
Subject: Re: [PATCH AUTOSEL 5.10 09/10] x86/barrier: Do not serialize MSR
 accesses on AMD

On Tue, Jan 16, 2024 at 09:44:01PM +0100, Pavel Machek wrote:
>Hi!
>
>> From: "Borislav Petkov (AMD)" <bp@...en8.de>
>>
>> [ Upstream commit 04c3024560d3a14acd18d0a51a1d0a89d29b7eb5 ]
>>
>> AMD does not have the requirement for a synchronization barrier when
>> acccessing a certain group of MSRs. Do not incur that unnecessary
>> penalty there.
>...
>> Performance captured using an unmodified ipi-bench using the 'mesh-ipi' option
>> with and without weak_wrmsr_fence() on a Zen4 system also showed significant
>> performance improvement without weak_wrmsr_fence(). The 'mesh-ipi' option ignores
>> CCX or CCD and just picks random vCPU.
>>
>>   Average throughput (10 iterations) with weak_wrmsr_fence(),
>>         Cumulative throughput: 4933374 IPI/s
>>
>>   Average throughput (10 iterations) without weak_wrmsr_fence(),
>>         Cumulative throughput: 6355156 IPI/s
>>
>> [1] https://github.com/bytedance/kvm-utils/tree/master/microbenchmark/ipi-bench
>
>Speed improvement, not a bugfix. Please drop.

Dropped, thanks!

-- 
Thanks,
Sasha