linux-kernel - Re: [PATCH 3/3] spi: meson-spicc: Lower CS between bursts

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]

Message-ID: <Y3YUJjSTI9U9qz59@sirena.org.uk>
Date:   Thu, 17 Nov 2022 10:59:50 +0000
From:   Mark Brown <broonie@...nel.org>
To:     Carlo Caione <ccaione@...libre.com>
Cc:     Kamlesh Gurudasani <kamlesh.gurudasani@...il.com>,
        Neil Armstrong <neil.armstrong@...aro.org>,
        Jerome Brunet <jbrunet@...libre.com>,
        David Airlie <airlied@...il.com>,
        Martin Blumenstingl <martin.blumenstingl@...glemail.com>,
        Kevin Hilman <khilman@...libre.com>,
        Daniel Vetter <daniel@...ll.ch>,
        linux-arm-kernel@...ts.infradead.org,
        linux-amlogic@...ts.infradead.org, linux-kernel@...r.kernel.org,
        linux-spi@...r.kernel.org, dri-devel@...ts.freedesktop.org
Subject: Re: [PATCH 3/3] spi: meson-spicc: Lower CS between bursts

On Thu, Nov 17, 2022 at 09:47:41AM +0100, Carlo Caione wrote:
> On some hardware (reproduced on S905X) when a large payload is
> transmitted over SPI in bursts at the end of each burst, the clock line
> briefly fluctuates creating spurious clock transitions that are being
> recognised by the connected device as a genuine pulses, creating an
> offset in the data being transmitted.

> Lower the GPIO CS between bursts to avoid the clock being interpreted as
> valid.

This is just plain broken, *many* SPI devices attach meaning to
chip select edges - for example register writes will typically
have the register address followed by one or more register values
for sequential registers.  Bouncing chip select in the middle of
transfer will corrupt data.  If the device can't handle larger
transfers it needs to advertise this limit and refuse to handle
them.

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