On Thu, 18 Aug 2022 18:59:00 +0200 Sebastian Andrzej Siewior wrote:

On 2022-08-18 09:02:00 [-0700], Jakub Kicinski wrote:

quoted

No ack, I'd much rather you waited for after the next merge window 
and queued this refactoring to net-next. Patch 9 is changing 70
files in networking. Unless I'm missing something and this is time
sensitive.  

It started with the clean up of the mess that has been made in the merge
and then it went on a little.

Any opinion on 8/9? It could wait for the next merge window if you want
to avoid a feature branch to pull from.

A question of priority, hard for me to judge how urgent any of it is.
But practically speaking the chances of a conflict on that one header
are pretty small, so ack if needed.

Regarding 9/9. This is a clean up, which is possible after 8/9. It can
definitely be applied later.
I assume you want only see the networking bits so I would split the
other subsystem out. I guess instead the big net patch I split them on
per driver vendor basis + net/ subsys?

Oh, don't care for per vendor split on a big refactoring like that.
You can split out the SPI patch and maybe BPF, the rest can be one 
big chunk AFAICT.

BTW, I have a hazy memory of people saying they switched to the _irq()
version because it reduced the number of retries significantly under
traffic. Dunno how practical that is but would you be willing to scan
the git history to double check that's not some of the motivation?