linux-kernel - Re: [PATCH] clk: vt8500: don't return possibly uninitialized data

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20160203113626.4a3f4f8a@v1ron-s7>
Date:	Wed, 3 Feb 2016 11:36:26 +0300
From:	Roman Volkov <v1ron@...l.ru>
To:	Arnd Bergmann <arnd@...db.de>
Cc:	linux-arm-kernel@...ts.infradead.org,
	Stephen Boyd <sboyd@...eaurora.org>,
	Michael Turquette <mturquette@...libre.com>,
	linux-kernel@...r.kernel.org, Tony Prisk <linux@...sktech.co.nz>,
	Andrzej Hajda <a.hajda@...sung.com>, linux-clk@...r.kernel.org
Subject: Re: [PATCH] clk: vt8500: don't return possibly uninitialized data

В Tue, 02 Feb 2016 21:00:29 +0100
Arnd Bergmann <arnd@...db.de> пишет:

> On Tuesday 02 February 2016 11:47:13 Stephen Boyd wrote:
> > On 02/02, Arnd Bergmann wrote:  
> > > On Monday 01 February 2016 17:15:45 Stephen Boyd wrote:
> > > 
> > > I see what you mean now. I checked different gcc versions, and
> > > with my patch I get the warnings for 4.6 through 4.9, but not for
> > > 5.x.
> > > 
> > > In general, I tried to only address warnings I still see with
> > > newer gcc version, as they are better about false positives. Do
> > > you think it's ok to take the patch as is then? Otherwise we
> > > probably have to add fake initializations which would shut up the
> > > warnings but not help with the code quality. 
> > 
> > Sure. I was hoping something could be done to restructure the
> > code to make it easier for the compiler to figure out the
> > variables would be initialized. But you're the one who's sending
> > the patch to silence them so if you're satisfied I'm not going to
> > spend too much time on this.
> >   
> 
> Ok, thanks!
> 
> Note that this one patch was for a real bug involving undefined
> behavior that is now fixed.
> 
> 	Arnd

Hi Arnd,

Thanks for fixing this code! Did someone reproduce this bug, or this
is something theoretical, based on the code analysis? I just never
heard about the issue. I can look into the code on the weekends too, I
have WM8505\WM8650 machines to test.

Is it enough to run the regular kernel build for WM8650 to see the
warnings, or there are special options in the kernel to run the compiler
test?

Regards
Roman