linux-kernel - Re: [PATCH] pinctrl: use non-devm kmalloc versions for free functions

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, 5 May 2017 17:49:32 -0400
From:   Tejun Heo <tj@...nel.org>
To:     Maxime Ripard <maxime.ripard@...e-electrons.com>
Cc:     Andre Przywara <andre.przywara@....com>,
        Linus Walleij <linus.walleij@...aro.org>,
        Icenowy Zheng <icenowy@...c.xyz>,
        Adam Borowski <kilobyte@...band.pl>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Chen-Yu Tsai <wens@...e.org>, linux-sunxi@...glegroups.com,
        linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] pinctrl: use non-devm kmalloc versions for free functions

Hello, Maxime.

On Fri, May 05, 2017 at 09:55:18PM +0200, Maxime Ripard wrote:
> > It doesn't make any sense to use the managed functions from the
> > release functions and if you're always matching devm_kmalloc() with
> > devm_kfree(), the only thing it'd do is confusing its readers.
> 
> I wouldn't say that being able to recover and free whatever memory
> leak we might have not making sense, but ok. That was one of the
> options, let's discard it.
>
> The other one is: refactor the rest of the allocations so that you
> don't have a mix of devm_kmalloc / devm_kfree and kmalloc / kfree for
> the same purpose in the same function.

So, I think the issues here are

1. Use of devm functions in a devres release function.  This just
   doesn't make sense.

2. If a resource is always allocated and freed in the same function,
   there's no reason to use devres for it.  I understand that there
   can be exceptions for this, e.g. consistency, but it doesn't seem
   to apply here.

3. Last but not least, allocating memory to destroy a radix tree.  It
   should be able to do that without allocating any memory.

Thanks.

-- 
tejun