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: Fri, 23 Feb 2024 15:01:30 +0200
From: Andy Shevchenko <andy.shevchenko@...il.com>
To: Quentin Schulz <foss+kernel@...il.net>
Cc: Jonathan Cameron <jic23@...nel.org>, Lars-Peter Clausen <lars@...afoo.de>, Heiko Stuebner <heiko@...ech.de>, 
	AngeloGioacchino Del Regno <angelogioacchino.delregno@...labora.com>, 
	Shreeya Patel <shreeya.patel@...labora.com>, Simon Xue <xxm@...k-chips.com>, 
	Philipp Zabel <p.zabel@...gutronix.de>, Jonathan Cameron <Jonathan.Cameron@...wei.com>, 
	linux-iio@...r.kernel.org, linux-arm-kernel@...ts.infradead.org, 
	linux-rockchip@...ts.infradead.org, linux-kernel@...r.kernel.org, 
	Quentin Schulz <quentin.schulz@...obroma-systems.com>
Subject: Re: [PATCH 0/3] iio: adc: rockchip_saradc: fix bitmasking and remove
 custom logic for getting reset

On Fri, Feb 23, 2024 at 2:46 PM Quentin Schulz <foss+kernel@...il.net> wrote:
>
> The mask for the channel selection is incorrect as it's specified to be
> 16b wide by is actually only 4.
>
> Also, the 16 lower bits in the SARADC_CONV_CON register are write
> protected. Whatever their value is can only be written to the hardware
> block if their associated bit in the higher 16 bits is set. Considering
> that the channel bitmask is 4b wide but that we can write e.g. 0 in
> there, we shouldn't use the value shifted by 16 as a mask but rather the
> bitmask for that value shifted by 16. This is currently NOT an issue
> because the only SoC with SARADCv2 IP is the RK3588 which has a reset
> defined in the SoC DTSI. When that is the case, the reset is asserted
> before every channel conversion is started. This means the registers are
> reset so effectively, we do not need to write zeros so the wrong mask
> still works because where we should be writing zeroes, there are already
> zeroes. However, let's fix this in case there comes a day there's an SoC
> which doesn't require to reset the controller before every channel
> conversion is started.
>
> Lastly, let's use the appropriate function from the reset subsystem
> for getting an optional exclusive reset instead of rolling out our own
> logic.
>
> Those three patches should not be changing any behavior.

Nice series, I have the comments in patch 3, but no need to resend
until Jonathan asks for. He might address that whilst applying.

Reviewed-by: Andy Shevchenko <andy.shevchenko@...il.com>

-- 
With Best Regards,
Andy Shevchenko

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ