linux-kernel - simple-reset to de-duplicate reset drivers

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-next>] [day] [month] [year] [list]

Date:   Thu, 10 Aug 2017 17:16:12 -0700
From:   Alexandru Gagniuc <alex.g@...ptrum.com>
To:     p.zabel@...gutronix.de
Cc:     linux-kernel@...r.kernel.org,
        Michal Simek <michal.simek@...inx.com>,
        soren.brinkmann@...inx.com, linux-arm-kernel@...ts.infradead.org,
        jun.nie@...aro.org, baoyou.xie@...aro.org,
        maxime.ripard@...e-electrons.com, wens@...e.org,
        mcoquelin.stm32@...il.com,
        Alexandre TORGUE <alexandre.torgue@...com>,
        narmstrong@...libre.com, carlo@...one.org, khilman@...libre.com
Subject: simple-reset to de-duplicate reset drivers

Hi,

I was looking at implementing a reset driver for our in-house SoC. It's 
essentially a long bitfield, each bit controlling one sort of reset. 
That seems to be surprisingly common. I've identified the following 
drivers which control a very similar reset:

* reset-zynq
* reset-zx2967
* reset-sunxi
* reset-stm32
* reset-socfpga
* reset-oxynas
* reset-meson
* reset-berlin

All of these have in common some form of another of:

	int bank = id / BITS_PER_LONG;
	int offset = id % BITS_PER_LONG;

It doesn't make sense to me why all these would be duplicated for every 
reset controller. I think it would make sense to have a common driver to 
handle these simple resets -- call it "simple-reset". I'm thinking of 
something similar to drivers/tty/serial/8250, where the following 
parameters can be specified:

- reg-offset
- reg-shift
- reg-io-width

I think a generic "simple-reset" driver which is configurable by a 
similar set of parameters would be a suitable replacement for all the 
aforementioned drivers. As far as our SoC goes, I've just added

	compatible = "st,stm32-rcc";

to our devicetree, and I already have a fully-working reset driver. Do 
you think we should proceed in the direction of "simple-reset", and what 
other features do you estimate we'll need?

Alex