linux-kernel - Re: [PATCH 2/3] x86/apic: Add TSC

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: <alpine.DEB.2.20.1706011148330.2470@nanos>
Date:   Thu, 1 Jun 2017 11:58:08 +0200 (CEST)
From:   Thomas Gleixner <tglx@...utronix.de>
To:     Peter Zijlstra <peterz@...radead.org>
cc:     Henrique de Moraes Holschuh <hmh@....eng.br>, x86@...nel.org,
        linux-kernel@...r.kernel.org, kevin.b.stanton@...el.com,
        bp@...en8.de
Subject: Re: [PATCH 2/3] x86/apic: Add TSC_DEADLINE quirk due to errata

On Thu, 1 Jun 2017, Peter Zijlstra wrote:
> That commit also states that is avoids a superfluous microcode load. And
> we've verified on affected systems (both Thomas and I have a SKL system
> with the PRMRR bit set in MTRRCAP) that when we manually load the
> microcode image the reported revision number matches the one from the
> image.
> 
>  [    0.000000] microcode: microcode updated early to revision 0xba, date = 2017-04-09
>  [    2.297894] microcode: sig=0x506e3, pf=0x2, revision=0xba
> 
> And:
> 
> # hexdump /lib/firmware/intel-ucode/06-5e-03 | head -1
> 0000000 0001 0000 00ba 0000 2017 0409 06e3 0005
>                   ^^^^^^^^^
> 
> So aside from a possible OS re-load of the microcode, the issue doesn't
> appear to have any negative effect.
> 
> 
> The microcode people (Cc'ed) might want to further look into this is
> they care about avoiding the superfluous reload, but for the purpose of
> this patch all is well.

Avoiding the superflous reload is dangerous. If the BIOS/FIT nonsense gets
fixed (however unlikely that is) then we run into an even worse problem
because then the kernel will not load version 5 if the CPU says version 4.
That's way worse than reloading the same version for nothing.

Thanks,

	tglx