Re: [PATCH v6 0/6] AMD Pstate Fixes And Enhancements

From: Borislav Petkov
Date: Fri Feb 09 2024 - 12:42:27 EST

On Fri, Feb 09, 2024 at 05:33:35PM +0000, Deucher, Alexander wrote:
> My reading of the rules is that you should wait before resending or
> pinging if you have not received feedback.

But the feedback you've received is not the whole feedback. Someone
might want to review them but not have the time right now. So at the
time we did agree that a week is kinda ok to let people have a look.

And from experience reviewers tend to ignore patchsets which get resent

> If you are actively receiving feedback, to me, it makes sense to
> rapidly iterate. If a patch is reviewed and comments are addressed,
> it can land rather than waiting an extra week or two.

There's the other problem - if you keep dealing with only a single
patchset, all the others who are waiting get starved. I, for example,
try to have at least some fairness when it comes to looking at people's
submissions and round-robin between them as much as I can.

> It also adds extra latency. With my maintainer hat on, I'd like to
> have new revisions rapidly. I guess it comes down to personal
> preference, but I don't do well with task switching.

I know *exactly* what you mean. I usually read my replies to the
previous submissions in order to swap in everything I was pointing at
the last time...

Can you review it all? Definitely not. :-\

> When a patch set is fresh in my mind, I'd rather see it finished off
> and committed sooner rather than lingering and then a week or two
> later, I'd need to page the whole discussion back into my head to make
> sure everything was addressed and all the tags were collected.

Yeah, I definitely see your point. But there are also those other
things I mentioned. I think one could find a good balance between the