linux-kernel - Re: [RFC][PATCH 3/3] x86: make MP a required-feature on 64-bit

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: <53A473AA.7050301@zytor.com>
Date:	Fri, 20 Jun 2014 10:47:22 -0700
From:	"H. Peter Anvin" <hpa@...or.com>
To:	Borislav Petkov <bp@...en8.de>
CC:	Dave Hansen <dave@...1.net>, linux-kernel@...r.kernel.org,
	x86@...nel.org, Andi Kleen <ak@...ux.intel.com>
Subject: Re: [RFC][PATCH 3/3] x86: make MP  a required-feature on 64-bit

On 06/20/2014 10:43 AM, Borislav Petkov wrote:
> On Fri, Jun 20, 2014 at 09:37:57AM -0700, H. Peter Anvin wrote:
>> We probably should just the cpu_has_mp macro entirely.  All it is used
>> for is printing a warning in amd_k7_smp_check().
>>
>> Andi, Borislav -- as far as I can tell, we have *never* enforced this on
>> the 64-bit kernel, although we have enforced it on 64-bit processors
>> running the 32-bit kernel.  We should either enforce it on both or just
>> drop it.  What is your opinion?
> 
> Well, my AMD CPUID guide says CPUID Fn8000_0001_EDX[19] is reserved,
> i.e. that X86_FEATURE_MP bit will be probably cleared.
> 
> And so it is, on my F15h AMD it is not set:
> 
> eax in: 0x80000001, eax = 00600f20 ebx = 10000000 ecx = 01ebbfff edx = 2fd3fbff
> 									  ^
> 								----------|
> 
> So I don't think we should enforce it on 64-bit.
> 
> I guess we can leave the check in amd_k7_smp_check() though but remove that
> nasty cpu_has_mp macro and do static_cpu_has() instead.
> 
> My 2¢ only.
> 

This is run before static_cpu_has().

The point, though, was that we "enforce" (taint) on 32 bits but not on
64 bits, which is clearly wrong.

My inclination is to completely kill amd_k7_smp_check() entirely, since
noone seems to know when it actually matters and it is clearly historic.

	-hpa

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/