linux-kernel - Re: [PATCH] soc: qcom: aoss: Fix the out of bound usage of cooling

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: <YNsrcHOf90rZl44z@google.com>
Date:   Tue, 29 Jun 2021 07:17:20 -0700
From:   Matthias Kaehlcke <mka@...omium.org>
To:     Manivannan Sadhasivam <manivannan.sadhasivam@...aro.org>
Cc:     bjorn.andersson@...aro.org, linux-arm-msm@...r.kernel.org,
        linux-kernel@...r.kernel.org, thara.gopinath@...aro.org,
        stable@...r.kernel.org
Subject: Re: [PATCH] soc: qcom: aoss: Fix the out of bound usage of
 cooling_devs

On Tue, Jun 29, 2021 at 09:55:58AM +0530, Manivannan Sadhasivam wrote:
> On Mon, Jun 28, 2021 at 04:03:14PM -0700, Matthias Kaehlcke wrote:
> 
> [...]
> 
> > 
> > 
> > A few more previous lines of code for context:
> > 
> >   int count = QMP_NUM_COOLING_RESOURCES;
> > 
> >   qmp->cooling_devs = devm_kcalloc(qmp->dev, count,
> >                                    sizeof(*qmp->cooling_devs),
> >                                    GFP_KERNEL);
> > 
> > I would suggest to initialize 'count' to 0 from the start and pass
> > QMP_NUM_COOLING_RESOURCES to devm_kcalloc() rather than 'count',
> > instead of resetting 'count' afterwards.
> 
> Yeah, I thought about it but the actual bug in the code is not resetting
> the count value to 0. So fixing this way seems a better option.

I don't agree that it's the better option. IMO it's clearer to pass
the constant QMP_NUM_COOLING_RESOURCES directly to devm_kcalloc(),
rather than giving the impression that the number of allocated items
is variable. Repurposing variables can be confusing and led to this
bug. Also the resulting code doesn't need to re-initialize 'count'.