linux-kernel - Re: [PATCH] spi: imx: add a check for speed

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, 8 Apr 2021 14:43:47 +0100
From:   Mark Brown <broonie@...nel.org>
To:     Clark Wang <xiaoning.wang@....com>
Cc:     shawnguo@...nel.org, s.hauer@...gutronix.de, festevam@...il.com,
        kernel@...gutronix.de, linux-imx@....com,
        linux-spi@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] spi: imx: add a check for speed_hz before calculating
 the clock

On Thu, Apr 08, 2021 at 06:33:47PM +0800, Clark Wang wrote:
> When some drivers use spi to send data, spi_transfer->speed_hz is
> not assigned. If spidev->max_speed_hz is not assigned as well, it
> will cause an error in configuring the clock.

> Add a check for these two values before configuring the clock. An
> error will be returned when they are not assigned.

For the case where the transfer speed is not set __spi_validate() will
take the controller's maximum speed so the controller should just be
able to unconditionally use the transfer's speed.  Your issue is
therefore that the controllers are sometimes not setting a maximum
speed which this doesn't seem to fix AFAICT?  I'd expect the driver to
be able to work one out based on the input clock.

>  struct spi_imx_devtype_data {
>  	void (*intctrl)(struct spi_imx_data *, int);
>  	int (*prepare_message)(struct spi_imx_data *, struct spi_message *);
> -	int (*prepare_transfer)(struct spi_imx_data *, struct spi_device *,
> -				struct spi_transfer *);
> +	int (*prepare_transfer)(struct spi_imx_data *, struct spi_device *);
>  	void (*trigger)(struct spi_imx_data *);
>  	int (*rx_available)(struct spi_imx_data *);
>  	void (*reset)(struct spi_imx_data *);

This seems to be a fairly big and surprising refactoring for the
described change.  It's quite hard to tie the change to the changelog.

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