Re: [GIT pull] scheduler updates for 5.1

From: Thomas Gleixner
Date: Sun Mar 24 2019 - 14:39:48 EST


On Sun, 24 Mar 2019, Linus Torvalds wrote:

> On Sun, Mar 24, 2019 at 7:14 AM Thomas Gleixner <tglx@xxxxxxxxxxxxx> wrote:
> >
> > Second more careful attempt for this set of fixes:
>
> What? No. It's the exact same garbage as last time, with no more care
> shown anywhere:
>
> kernel/sched/cpufreq_schedutil.c:346:26: note: in expansion of macro âminâ
> sg_cpu->iowait_boost = min(sg_cpu->iowait_boost << 1,
> SCHED_CAPACITY_SCALE);
>
> I told you *exactly* what the problem was the last time, and it's
> *exactly* the same now too.
>
> Hmm. Looking closer, I see that the patch you *claim* to send me has
> "min_t()". But the tree I actually got has the same old garbage.
>
> I suspect you didn't update the git repo properly or something.

Duh. Just did

git checkout sched-urgent-for-linus
git diff sched/urgent -> empty
git diff origin/sched/urgent -> empty
git diff origin/sched-urgent-for-linus -> empty

Now doing a fresh clone from git.k.org (not ssh) shows me the stale commit
from last week.

I'm sure that I force pushed and my pull request script checks whether all
branches are up to date before creating the pull request mail. No idea at
all what went wrong.

I just force pushed the thing once more. The head commit is:

b9a7b8831600: sched/fair: Skip LLC NOHZ logic for asymmetric systems

I'll add that to the pull request mail scripts in future.

It has propagated to gitweb now as well.

Thanks,

tglx