linux-kernel - Re: [PATCH] spi: spidev: Add compatible for external SPI ports on Kontron boards

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: <a5b88ad9-3884-1d9c-c4ad-057266f84261@kontron.de>
Date:   Mon, 13 Jul 2020 15:19:52 +0200
From:   Frieder Schrempf <frieder.schrempf@...tron.de>
To:     Mark Brown <broonie@...nel.org>
Cc:     linux-kernel@...r.kernel.org, linux-spi@...r.kernel.org
Subject: Re: [PATCH] spi: spidev: Add compatible for external SPI ports on
 Kontron boards

On 02.07.20 18:24, Frieder Schrempf wrote:
> On 02.07.20 17:07, Mark Brown wrote:
>> On Thu, Jul 02, 2020 at 04:46:09PM +0200, Frieder Schrempf wrote:
>>
>>> My intention is to use the spidev driver in the default board DT for an
>>> interface that is routed to an extension connector and has no dedicated
>>> slave device attached onboard. So users can attach sensors, etc. with
>>> userspace drivers without touching the kernel or DT.
>>
>> The expected way of doing this is to describe whatever was attached via
>> DT when it's attached - the device is what has the compatible, not some
>> connector in the middle of the connection.  The way you've got things
>> set up if the device has a driver then they won't be able to instantiate
>> the driver.
> 
> Ok thanks, got it. I will remove the spi device from the board DT and 
> use an overlay if required. Now I got a reason to learn writing DT 
> overlays ;)

I need to come back to this for another question. I would very much like 
to do it the way Mark described it above: Use a DT overlay to describe 
the device and load this overlay when the device is attached.

But if I get this correctly it requires me to write a specific driver 
that handles the loading of the overlay, which is way too much overhead 
for the simple issue I'm trying to solve.

I would have expected that there is some kind of existing userspace API 
to load an overlay manually, but it seems like there isn't!?

So what's the reasoning behind this? How can I solve this in a 
mainline-compliant way, meaning without either keeping downstream 
patches to bind spidev to my device or writing and maintaining code that 
does the overlay loading?