linux-kernel - Re: [PATCH 0/2] x86/microcode/amd: Do not overwrite specific patch levels

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: <20160327123218.GA28778@khazad-dum.debian.net>
Date:	Sun, 27 Mar 2016 09:32:18 -0300
From:	Henrique de Moraes Holschuh <hmh@....eng.br>
To:	Borislav Petkov <bp@...en8.de>
Cc:	LKML <linux-kernel@...r.kernel.org>, X86 ML <x86@...nel.org>
Subject: Re: [PATCH 0/2] x86/microcode/amd: Do not overwrite specific patch
 levels

On Sun, 27 Mar 2016, Borislav Petkov wrote:
> On Sat, Mar 26, 2016 at 08:31:57PM -0300, Henrique de Moraes Holschuh wrote:
> > This patchset looks like it is pretty much a requirement for any distro that
> > ships AMD microcode updates...  Maybe the two commits should be sent to
> > -stable, now that they have seen lots of testing in mainline 4.4.x as well
> > as SuSE kernels?
> 
> I wouldn't say lots... :)
> 
> Do you have any specific bug report(s) or similar in mind? Or is it more
> of a "it would be wise to backport" sentiment?

Well, a Google search shows that microcodes 0x100098 and 0x100009f are not
that rare.  IMHO, it is a pretty safe bet that both Debian and Ubuntu have
some users of those microcodes.  Users who will have their systems rendered
unbootable (until we teach them about the dis_ucode_ldr parameter to the
kernel) if they ever install the amd64-microcode package in a kernel that
doesn't have this patchset.

So, it is really a bit of both: I had several "it doesn't work" type of
reports for both AMD and Intel over the years, and most often people won't
come back to the initial bug report, if they even go that far as to report a
bug in the first place: they just remove the microcode update packages and
disapear...  so, I wouldn't know if any were due to this specific issue
except by luck.

But I do assume there are at least 20 users having trouble that will never
report it for each single bug report I get, and it is likely to be a lot
more :-(

-- 
  "One disk to rule them all, One disk to find them. One disk to bring
  them all and in the darkness grind them. In the Land of Redmond
  where the shadows lie." -- The Silicon Valley Tarot
  Henrique Holschuh