linux-kernel - Re: [PATCH 0/2] Introduce STM32MP1 Reset driver

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:   Wed, 14 Mar 2018 10:12:02 +0100
From:   Philipp Zabel <p.zabel@...gutronix.de>
To:     gabriel.fernandez@...com, Rob Herring <robh+dt@...nel.org>,
        Mark Rutland <mark.rutland@....com>,
        Maxime Coquelin <mcoquelin.stm32@...il.com>,
        Alexandre Torgue <alexandre.torgue@...com>
Cc:     linux-kernel@...r.kernel.org, devicetree@...r.kernel.org,
        linux-arm-kernel@...ts.infradead.org,
        gabriel.fernandez.st@...il.com,
        Loic PALLARDY <loic.pallardy@...com>,
        benjamin GAIGNARD <benjamin.gaignard@...com>,
        Michael Turquette <mturquette@...libre.com>, sboyd@...nel.org
Subject: Re: [PATCH 0/2] Introduce STM32MP1 Reset driver

Hi Gabriel,

On Tue, 2018-03-13 at 17:34 +0100, gabriel.fernandez@...com wrote:
> From: Gabriel Fernandez <gabriel.fernandez@...com>
> 
> This patch-set enables the reset of STM32MP1.
> It uses the reset simple driver by introducing the clear register offset
> parameter.
> STM32MP1 reset IP has a register to assert by writing '1' and another
> register to de-assert by writing '1'.
> The offset between this two registers is '0x4'.

I worry a bit about feature creep in the simple-reset driver.
Your patch on its own is simple enough, and I'm not opposed to add a
SET/CLR feature on principle, but there are a few issues:

The RESET_SIMPLE Kconfig description currently says:
  "This enables a simple reset controller driver for reset lines that
   that can be asserted and deasserted by toggling bits in a contiguous,
   exclusive register space."
That would have to be extended to mention SET/CLR register pairs as an
alternative.

What about status (reset_simple_status)? Can current reset line status
be read back from the SET register, as is currently tried? If not, is
there a way to read current reset line status back at all?

The data->lock spinlock is only needed to protect the read-modify-write
cycle on a toggle register, for separate SET/CLR register access the
locking is not necessary.

At this point, it may or may not be easier to add a custom reset driver.
Either way you go, this is missing binding documentation for the
st,stm32mp1-rcc compatible in Documentation/devicetree/bindings/reset.

> The patch 'dt-bindings: reset: add STM32MP1 resets' could be squashed
> with the patch:
> 'dt-bindings: Document STM32MP1 Reset Clock Controller (RCC) bindings'
> commit 3830681d354f
>
> Gabriel Fernandez (2):
>   dt-bindings: reset: add STM32MP1 resets
>   reset: simple: Enable stm32mp1 reset driver
> 
>  drivers/reset/reset-simple.c                |  27 +++++--
>  drivers/reset/reset-simple.h                |   1 +
>  include/dt-bindings/reset/stm32mp1-resets.h | 108 ++++++++++++++++++++++++++++
>  3 files changed, 130 insertions(+), 6 deletions(-)
>  create mode 100644 include/dt-bindings/reset/stm32mp1-resets.h

regards
Philipp