Re: [PATCH] uio: pruss: Deprecate use of this driver

From: Andrew Davis
Date: Wed Apr 10 2024 - 10:53:49 EST


On 4/5/24 10:39 AM, Robert Nelson wrote:
On Tue, Mar 26, 2024 at 12:36 PM Andrew Davis <afd@xxxxxx> wrote:

On 3/26/24 12:24 PM, Matthijs van Duin wrote:
I'll write a more in-depth reply when I have a moment, but right now I'd
like to point out that the uio-pruss driver in mainline linux is for the
pru subsystem on the freon/primus family of ARM9-based SoCs (OMAP-L1xx /
AM17xx / AM18xx / TMS320C674x / DA8xx), which is not currently supported
by remoteproc-pru.


I'll wait for your full reply, but a quick note, for those devices listed
this driver isn't usable either after they all moved to DT. As this driver
never got a DT port and relies on platform data (which is gone for those
couple devices).

Andrew

Andrew, I think we are okay with nuking the whole uio driver in
mainline, I'll ack it..


Okay, will send that patch.

I figured we'd have more community response.. We gave them a chance..

The only issue I personally have with supporting remoteproc-pruss

We have a few users on 4.19.x-ti, 5.10.x-ti (remoteproc_pruss) can we
please make sure this project:

https://git.ti.com/gitweb?p=pru-software-support-package/pru-software-support-package.git;a=summary

actually works on mainline remoteproc_pruss ?

Watching the shortlog, it must break on every single TI LTS release:


Yeah, before we got the remoteproc driver upstream we managed
to break it just about every LTS, it *should* be a bit more
stable now that it is upstream..

https://git.ti.com/gitweb?p=pru-software-support-package/pru-software-support-package.git;a=shortlog

Whereas uio... well same firmware from 3.8.x ;)


That is one benefit of the microkernel style "just expose raw memory
to userspace as a driver", it sure looks simple :)

Andrew

(i will keep our uio fork alive, but we just use an overlay to switch
to between remoteproc_pruss and uio)

Regards,