linux-kernel - Re: syscon regmap for disabled node?

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] [day] [month] [year] [list]

Date:   Mon, 19 Feb 2018 18:06:59 -0600
From:   Suman Anna <s-anna@...com>
To:     Arnd Bergmann <arnd@...db.de>
CC:     Lee Jones <lee.jones@...aro.org>,
        Pankaj Dubey <pankaj.dubey@...sung.com>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Linux ARM <linux-arm-kernel@...ts.infradead.org>
Subject: Re: syscon regmap for disabled node?

On 02/17/2018 01:51 PM, Arnd Bergmann wrote:
> On Sat, Feb 17, 2018 at 2:40 AM, Suman Anna <s-anna@...com> wrote:
>> Hi Pankaj, Arnd, Lee,
>>
>> I am testing some code to use a syscon/regmap interface and I find that
>> the syscon/regmap is initialized even on a disabled device node using a
>> "syscon" compatible property when I have expected it to fail. Prior to
>> commit bdb0066df96e ("mfd: syscon: Decouple syscon interface from
>> platform devices"), the driver would have never probed, and the
>> of_syscon_register() only checks for the compatible, but not if the
>> device node is available. Is this intentional or a bug?
> 
> My guess would be that nothing relies on the current behavior and that
> it could be changed if it causes problems, but it's hard to tell for the general
> case.
> 
> Can you describe why you have a disabled syscon node? Most uses of the
> status property are for on-chip devices that we disable because nothing is
> connected to them, but that seems unlikely for a syscon.

I am just testing the failure paths in my code by specifically disabling
the syscon node, and to my surprise, the code went through fine when I
had expected the syscon_node_to_regmap() call to fail.

regards
Suman