linux-kernel - Re: [patch] x86: Rendezvous all the cpu's for MTRR/PAT init

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: <20090819065929.GA21824@sli10-desk.sh.intel.com>
Date:	Wed, 19 Aug 2009 14:59:29 +0800
From:	Shaohua Li <shaohua.li@...el.com>
To:	"Siddha, Suresh B" <suresh.b.siddha@...el.com>
Cc:	Andrew Morton <akpm@...ux-foundation.org>,
	Nick Piggin <npiggin@...e.de>, "mingo@...e.hu" <mingo@...e.hu>,
	"hpa@...or.com" <hpa@...or.com>,
	"tglx@...utronix.de" <tglx@...utronix.de>,
	"Pallipadi, Venkatesh" <venkatesh.pallipadi@...el.com>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [patch] x86: Rendezvous all the cpu's for MTRR/PAT init

On Wed, Aug 19, 2009 at 02:20:57PM +0800, Siddha, Suresh B wrote:
> On Tue, 2009-08-18 at 18:01 -0700, Andrew Morton wrote:
> > On Tue, 18 Aug 2009 17:30:35 -0700 Suresh Siddha <suresh.b.siddha@...el.com> wrote:
> > > +	/*
> > > +	 * Ideally we should hold mtrr_mutex here to avoid mtrr entries changed,
> > > +	 * but this routine will be called in cpu boot time, holding the lock
> > > +	 * breaks it. This routine is called in two cases: 1.very earily time
> > > +	 * of software resume, when there absolutely isn't mtrr entry changes;
> > > +	 * 2.cpu hotadd time. We let mtrr_add/del_page hold cpuhotplug lock to
> > > +	 * prevent mtrr entry changes
> > > +	 */
> > 
> > That's a tantalising little comment.  What does "breaks it" mean?  How
> > can reviewers and later code-readers possibly suggest alternative fixes
> > to this breakage if they aren't told what it is?
> 
> This is a cut and paste comment coming from the previous code. Shaohua
> added this comment originally and I think this is the case he is trying
> to avoid.
> 
> cpu - A modifying/adding a  MTRR register
> 
> cpu - B is coming online
> 
> if cpu - A doesn't take the cpu hotplug lock, then potentially what can
> happen is that cpu B will update its mtrr's with old state and now A can
> change the state and before B comes completely online, A can do send the
> MTRR update to all cpu's except B.
> 
> So Shaohua's code is taking cpu hotplug lock before A updates MTRR's so
> that B's MTRRs are always is in sync with rest of the cpu's in the
> system. Only the mtrr_mutex is not enough.
> 
> Nevertheless as far as this patch is concerned mtrr_aps_init() gets
> called during early boot/resume time and as such we never hit this
> condition. So I removed this comment in the new patch appended.
> 
> Shaohua if you agree with my explanation we can have a separate patch to
> make the original comment more meaningful.
Yes, your explanation is correct.

Thanks,
Shaohua
--
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/