linux-kernel - Re: [PATCH] hwmon: coretemp: fix oops on cpu unplug

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: <1335805164.2351.116.camel@groeck-laptop>
Date:	Mon, 30 Apr 2012 09:59:24 -0700
From:	Guenter Roeck <guenter.roeck@...csson.com>
To:	"Kirill A. Shutemov" <kirill.shutemov@...ux.intel.com>
CC:	Fenghua Yu <fenghua.yu@...el.com>, Andi Kleen <ak@...ux.intel.com>,
	Jean Delvare <khali@...ux-fr.org>,
	"lm-sensors@...sensors.org" <lm-sensors@...sensors.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] hwmon: coretemp: fix oops on cpu unplug

On Mon, 2012-04-30 at 12:19 -0400, Kirill A. Shutemov wrote:
> On Mon, Apr 30, 2012 at 08:28:10AM -0700, Guenter Roeck wrote:
> > Hi,
> > 
> > good catch. Couple of problems, though.
> > 
> > First, what number of cores are we talking about ? We should probably
> > increase NUM_REAL_CORES as well. Long term, we should get rid of the
> > dependency to prevent that problem from happening again, but that is a
> > different issue.
> 
> It triggered by another bug. Cores on my 10-core processor have ids 0, 1,
> 2, 8, 9, 16, 17, 18, 24, 25. :-/
> No need to change NUM_REAL_CORES at the moment.
> 
The problem, though, is that core_data[] is indexed by the core ID. So,
NUM_REAL_CORES is a bit misleading. It should probably be named
NUM_CORE_ID or something similar.

> > Second, we'll need the same code in get_core_online(). Otherwise the
> > platform device can be created for the new core (if the core is
> > re-enabled) but will never be deleted.
> 
> It has already handled in create_core_data().
> 
... which is called from coretemp_add_core(), which does not return an
error. Besides, it is called after the call to coretemp_device_add().

As a result, the platform device for the unsupported core(s) will be
created, even if the core is not supported. I just don't think that is a
good idea.

Besides, if you'd have the check in get_core_online(), you would not
even hit the problem in put_core_offline(), since the platform device
would not exist and the function would bail out because of that.

Thanks,
Guenter


--
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/