[PATCH 0/3] apple rtkit: add support for atomic operations

[Sven Peter]

On Sat, Sep 17, 2022, at 13:21, Russell King (Oracle) wrote:
> On Sat, Sep 17, 2022 at 12:42:30PM +0200, Sven Peter wrote:
>> Hi,
>> 
>> On Wed, Sep 14, 2022, at 10:34, Russell King (Oracle) wrote:
>> > Hi,
>> >
>> > This series adds support to the Apple rtkit for atomic messaging,
>> > which will be necessary for SMC support on these platforms.
>> >
>> > SMC support requires a bit of rework after recent review comments,
>> > so I'm sending this separately in the hope that we can move forward
>> > with upstream kernels.
>> 
>> Maybe I misremember or maybe you forgot to pick up my reviewed-by tags
>> for the mailbox commits.
>
> It's the first time I've sent the mailbox commits after Hector
> mentioned them.

There are just too many things in flight right now which is part of the
reason I'm happy you're trying to get all this upstream!

https://lore.kernel.org/asahi/YxYoW2IZFK%2FJdKgH@shell.armlinux.org.uk/


>
>> Either way, for the entire series:
>> 
>> Reviewed-by: Sven Peter <sven at svenpeter.dev>
>> 
>> 
>> We'd usually take the last soc/rtkit commit through the Asahi platform and then the
>> soc tree. Taking it through the mailbox tree will make things more complicated
>> if SMC (which is the only driver that uses the poll API and hard depends on it) is
>> also ready for this merge window. I don't think that's very likely so taking it
>> through mailbox is also fine.
>
> I'm sending these seperately because I don't think the platform/mfd
> issues will be sorted before the merge window. We've fallen into the
> trap where I have a load of updates to the patches in my git tree
> that aren't in the Asahi tree, but Lee wants some substantial changes
> to the way the MFD part is handled which I think Hector needs to do.
> So somehow I need to get my changes back to Hector (I don't do github)
> which basically means digging out the original patches from the Asahi
> kernel tree and supplying diffs to Hector for those - along with all
> the acks/reviewed-bys etc. This *really* isn't nice.
>
> I didn't follow that discussion very well - with about 130 messages
> in total in that thread, it just became too much for me to keep track
> of, as I explained as the discussion drew to a close. That was the
> whole reason for me sending v2, which essentially mostly got ignored
> and discussion continued on v1.
>
> So, basically, trying to get these patches into mainline has created
> a hell of a lot of extra work and difficulty - and at this point I'm
> wondering whether it really is a good idea to continue with this.
>
> It's fine if the comments are all minor issues that can be quickly
> resolved, but a major restructuring is a completely different matter.
>
> So, to at least get some progress, I decided to send what would be
> possible to hopefully merge during the remainder of this window.

yeah, I'm just glad this is finally getting some momentum. there's just
too many things in-flight in our downstream trees!
I didn't follow that discussion entirely either but I also don't think
SMC will land in this merge window so taking this through mailbox is
completely fine with me.


Sven