linux-kernel - Re: [PATCH 1/2] dt-bindings: spi: fsl-lpspi: Option to allow stalling

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:   Thu, 27 Sep 2018 23:47:23 +0100
From:   Mark Brown <broonie@...nel.org>
To:     Đặng Trần Hiếu 
        <dangtranhieu2012@...il.com>
Cc:     robh@...nel.org, linux-spi@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/2] dt-bindings: spi: fsl-lpspi: Option to allow stalling

On Wed, Sep 26, 2018 at 09:37:39PM +0700, Đặng Trần Hiếu wrote:

> Default value of the register is to allow stalling (NOSTALL bit not
> set) but the spi-fsl-lpspi driver defaults to setting the NOSTALL bit
> in CFGR1. To me, it's more logical to leave the NOSTALL bit off with
> fsl,nostall binding to set the bit but as I am not sure if there are
> other drivers depending on the NOSTALL bit being set and not wanting
> to break other drivers hence introduction of this binding to keep the
> current default behavior.

I can't see a situation where you'd actively want to report an error
rather than stall, stalling allows us to handle things gracefully by
restarting things.  Even if it's a timeout situation it sounds like we
can unblock by reading/writing the stalled FIFO, and normally you'd be
resetting the entire IP anyway.  I'd imagine the driver is this way
either through an oversight or because there's some bug on some silicon
which means that stalling breaks.

Probably best to just always enable stalling unless I'm misreading
things, if it is silicon bugs on some versions then either a whitelist
or blacklist of SoCs to enable on (depending on how common the bug is)
would be the way forwards - that way SoCs where it works get the benefit.

Download attachment "signature.asc" of type "application/pgp-signature" (489 bytes)