linux-kernel - Re: [PATCH V4] mfd: wm8994-core: Don't use managed regulator bulk get API

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]

Date:   Fri, 25 Nov 2016 15:46:21 +0530
From:   Viresh Kumar <viresh.kumar@...aro.org>
To:     Lee Jones <lee.jones@...aro.org>
Cc:     linaro-kernel@...ts.linaro.org, broonie@...nel.org,
        ckeepax@...nsource.wolfsonmicro.com,
        patches@...nsource.wolfsonmicro.com, linux-kernel@...r.kernel.org
Subject: Re: [PATCH V4] mfd: wm8994-core: Don't use managed regulator bulk
 get API

On 25-11-16, 10:14, Lee Jones wrote:
> On Thu, 27 Oct 2016, Viresh Kumar wrote:
> 
> > The kernel WARNs and then crashes today if wm8994_device_init() fails
> > after calling devm_regulator_bulk_get().
> > 
> > That happens because there are multiple devices involved here and the
> > order in which managed resources are freed isn't correct.
> > 
> > The regulators are added as children of wm8994->dev.  Whereas,
> > devm_regulator_bulk_get() receives wm8994->dev as the device, though it
> > gets the same regulators which were added as children of wm8994->dev
> > earlier.
> > 
> > During failures, the children are removed first and the core eventually
> > calls regulator_unregister() for them. As regulator_put() was never done
> > for them (opposite of devm_regulator_bulk_get()), the kernel WARNs at
> > 
> > 	WARN_ON(rdev->open_count);
> > 
> > And eventually it crashes from debugfs_remove_recursive().
> 
> Is ...
> 
>    mfd: wm8994-core: disable regulators before removing them
> 
> ... required as well, or is that separate?

It would be better if we get that too. Anyway, the $subject patch has a
dependency on it..

-- 
viresh