linux-kernel - Re: [PATCH v3 4/8] iio: adc: ad4030: Reduce register access transfer 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] [day] [month] [year] [list]

Message-ID: <aNlDfJvyXjnfINy3@debian-BULLSEYE-live-builder-AMD64>
Date: Sun, 28 Sep 2025 11:17:32 -0300
From: Marcelo Schmitt <marcelo.schmitt1@...il.com>
To: Jonathan Cameron <jic23@...nel.org>
Cc: Marcelo Schmitt <marcelo.schmitt@...log.com>, linux-iio@...r.kernel.org,
	devicetree@...r.kernel.org, linux-doc@...r.kernel.org,
	linux-spi@...r.kernel.org, linux-kernel@...r.kernel.org,
	michael.hennerich@...log.com, nuno.sa@...log.com,
	eblanc@...libre.com, dlechner@...libre.com, andy@...nel.org,
	robh@...nel.org, krzk+dt@...nel.org, conor+dt@...nel.org,
	corbet@....net
Subject: Re: [PATCH v3 4/8] iio: adc: ad4030: Reduce register access transfer
 speed

On 09/28, Jonathan Cameron wrote:
> On Fri, 26 Sep 2025 17:39:42 -0300
> Marcelo Schmitt <marcelo.schmitt@...log.com> wrote:
> 
> > Configuration register accesses are not considered a critical path in terms
> > of time to complete. Even though register access transfers can run at high
> > speeds, nanosecond completion times are not required as device
> > configuration is usually done step by step from user space. Also, high
> > frequency transfers hinder debug with external tools since they require
> > faster clocked equipment. Reduce register access transfer speed.
> 
> So making debug with external tools easier isn't usually a justification we'd
> make to slow things down by default.
> 
> Is there another reason for this being useful as opposed to not a problem
> to do?   If it had been done this way in the first place I wouldn't have
> minded, but to make a change I'd like either some others to jump in and
> say, yes please do this, or a reason beyond you are using tooling that can't
> cope with 80 MHz and don't want to hack the driver when you need
> to slow it down (my tools can't cope with that rate either!)

Main motivation for this was a suggestion from David.
https://lore.kernel.org/linux-iio/30659b16-290d-4ae5-a644-214c106bbe87@baylibre.com/
By the way, if he agrees with, I'll add a suggested-by tag (if we decide to keep
this patch).

Reasoning a bit more about this, lowering reg access speed may help debug with
external tools, but it won't help debug transfers ran by SPI offload hw because
those transfers will be fast anyway. Maybe a more relevant potential benefit of
lowering transfer speeds would be to make it more "friendly" to slower
controllers. E.g. raspberry pi controller reaches 32 MHz maximum so, unless SPI
core can adapt transfers in those cases, it wouldn't work on a rpi (if anyone
ever connects this to a rpi).

Me, I only have remote access to a setup with adaq4216 connected to a zedboard
so I won't be connecting any external tool for debugging. 

Another thing that came to mind now is we could just not set speed_hz of
spi_transfers. AFAIC, those are not required.