linux-kernel - Re: [PATCH] regulator: core: Fix memory leak in regulator_resolve

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: <20150715112730.GN11162@sirena.org.uk>
Date:	Wed, 15 Jul 2015 12:27:30 +0100
From:	Mark Brown <broonie@...nel.org>
To:	Javier Martinez Canillas <javier@....samsung.com>
Cc:	Krzysztof Kozlowski <k.kozlowski@...sung.com>,
	Liam Girdwood <lgirdwood@...il.com>,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH] regulator: core: Fix memory leak in
 regulator_resolve_supply()

On Wed, Jul 15, 2015 at 10:38:38AM +0200, Javier Martinez Canillas wrote:
> On 07/15/2015 10:01 AM, Krzysztof Kozlowski wrote:

> > The _regulator_put() reverts more work than create_regulator() did,
> > e.g.: module_put and rdev->open_count--. Maybe you need a
> > destroy_regulator() function?

> Yes, it reverts more work than create_regulator() but the intention is to
> revert what set_supply() did. If you look at the set_supply() function,
> it does supply_rdev->open_count++.

> I did indeed missed the module_put() but now looking at the code again, I

Me too, I've dropped the patch.  At first glance everything looked safe
for multiple calls.

> wonder if the problem is not that set_supply() is missing a try_module_get()
> to be consistent with what the _regulator_get() function does.

The problem is more that it's a separate implementation and not just
using _regulator_get() I think.  A separate, rarely used, path is likely
to have this sort of issue.

Download attachment "signature.asc" of type "application/pgp-signature" (474 bytes)