From: Peter Zijlstra <peterz@infradead.org>
To: Mel Gorman <mgorman@techsingularity.net>
Cc: linux-kernel@vger.kernel.org, Ingo Molnar <mingo@redhat.com>,
Juri Lelli <juri.lelli@redhat.com>,
Dietmar Eggemann <dietmar.eggemann@arm.com>,
Valentin Schneider <vschneid@redhat.com>,
Chris Mason <clm@meta.com>
Subject: Re: [PATCH 2/2] sched/fair: Reimplement NEXT_BUDDY to align with EEVDF goals
Date: Tue, 28 Oct 2025 16:47:34 +0100 [thread overview]
Message-ID: <20251028154734.GS4068168@noisy.programming.kicks-ass.net> (raw)
In-Reply-To: <20251028153351.GN3245006@noisy.programming.kicks-ass.net>
On Tue, Oct 28, 2025 at 04:33:51PM +0100, Peter Zijlstra wrote:
> On Mon, Oct 27, 2025 at 01:39:15PM +0000, Mel Gorman wrote:
> > @@ -8783,7 +8862,7 @@ static void check_preempt_wakeup_fair(struct rq *rq, struct task_struct *p, int
> > * When non-idle entity preempt an idle entity,
> > * don't give idle entity slice protection.
> > */
> > - do_preempt_short = true;
> > + do_preempt_short = PREEMPT_WAKEUP_NEXT;
> > goto preempt;
> > }
>
> I'm confused, should this not be WAKEUP_RESCHED?
It doesn't matter, you cannot end up with !do_preempt_short at preempt:,
so that condition is always true and can thus be deleted, at which point
the value of do_preempt_short is irrelevant and doesn't need to be set.
A little something like so perhaps...
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -8727,7 +8727,7 @@ static void set_next_buddy(struct sched_
}
enum preempt_wakeup_action {
- PREEMPT_WAKEUP_NONE, /* No action on the buddy */
+ PREEMPT_WAKEUP_NONE = 0, /* No action on the buddy */
PREEMPT_WAKEUP_NEXT, /* Check next is most eligible
* before rescheduling.
*/
@@ -8814,7 +8814,7 @@ __do_preempt_buddy(struct rq *rq, struct
*/
static void check_preempt_wakeup_fair(struct rq *rq, struct task_struct *p, int wake_flags)
{
- enum preempt_wakeup_action do_preempt_short = PREEMPT_WAKEUP_NONE;
+ enum preempt_wakeup_action preempt_action = PREEMPT_WAKEUP_NONE;
struct task_struct *donor = rq->donor;
struct sched_entity *se = &donor->se, *pse = &p->se;
struct cfs_rq *cfs_rq = task_cfs_rq(donor);
@@ -8863,7 +8863,6 @@ static void check_preempt_wakeup_fair(st
* When non-idle entity preempt an idle entity,
* don't give idle entity slice protection.
*/
- do_preempt_short = PREEMPT_WAKEUP_NEXT;
goto preempt;
}
@@ -8883,17 +8882,17 @@ static void check_preempt_wakeup_fair(st
* current's slice protection in order to allow preemption.
*/
if (sched_feat(PREEMPT_SHORT) && (pse->slice < se->slice)) {
- do_preempt_short = PREEMPT_WAKEUP_NEXT;
+ preempt_action = PREEMPT_WAKEUP_NEXT;
} else {
/*
* If @p potentially is completing work required by current then
* consider preemption.
*/
- do_preempt_short = __do_preempt_buddy(rq, cfs_rq, wake_flags,
+ preempt_action = __do_preempt_buddy(rq, cfs_rq, wake_flags,
pse, se);
}
- switch (do_preempt_short) {
+ switch (preempt_action) {
case PREEMPT_WAKEUP_NONE:
return;
case PREEMPT_WAKEUP_RESCHED:
@@ -8905,18 +8904,16 @@ static void check_preempt_wakeup_fair(st
/*
* If @p has become the most eligible task, force preemption.
*/
- if (__pick_eevdf(cfs_rq, !do_preempt_short) == pse)
+ if (__pick_eevdf(cfs_rq, false) == pse)
goto preempt;
- if (sched_feat(RUN_TO_PARITY) && do_preempt_short != PREEMPT_WAKEUP_NONE)
+ if (sched_feat(RUN_TO_PARITY))
update_protect_slice(cfs_rq, se);
return;
preempt:
- if (do_preempt_short)
- cancel_protect_slice(se);
-
+ cancel_protect_slice(se);
resched_curr_lazy(rq);
}
next prev parent reply other threads:[~2025-10-28 15:47 UTC|newest]
Thread overview: 23+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <20251027133915.4103633-1-mgorman@techsingularity.net>
2025-10-27 13:39 ` [PATCH 1/2] sched/fair: Enable scheduler feature NEXT_BUDDY Mel Gorman
2025-10-28 14:37 ` Peter Zijlstra
2025-10-27 13:39 ` [PATCH 2/2] sched/fair: Reimplement NEXT_BUDDY to align with EEVDF goals Mel Gorman
2025-10-28 15:05 ` Peter Zijlstra
2025-10-31 9:46 ` Mel Gorman
2025-10-28 15:09 ` Peter Zijlstra
2025-10-31 9:48 ` Mel Gorman
2025-10-28 15:33 ` Peter Zijlstra
2025-10-28 15:47 ` Peter Zijlstra [this message]
2025-10-30 9:10 ` Peter Zijlstra
2025-10-31 10:27 ` Mel Gorman
2025-11-12 12:25 [PATCH 0/2 v5] Reintroduce NEXT_BUDDY for EEVDF Mel Gorman
[not found] ` <20251112122521.1331238-3-mgorman@techsingularity.net>
2025-11-12 14:48 ` [PATCH 2/2] sched/fair: Reimplement NEXT_BUDDY to align with EEVDF goals Peter Zijlstra
2025-11-13 8:26 ` Madadi Vineeth Reddy
2025-11-13 9:04 ` Mel Gorman
2025-11-14 12:13 ` Peter Zijlstra
-- strict thread matches above, loose matches on Subject: below --
2025-11-03 11:04 [PATCH 0/2 v4] Reintroduce NEXT_BUDDY for EEVDF Mel Gorman
2025-11-03 11:04 ` [PATCH 2/2] sched/fair: Reimplement NEXT_BUDDY to align with EEVDF goals Mel Gorman
2025-11-03 14:07 ` Peter Zijlstra
2025-11-03 14:14 ` Peter Zijlstra
2025-11-05 21:48 ` Madadi Vineeth Reddy
2025-11-07 8:53 ` Mel Gorman
2025-10-21 14:28 [RFC PATCH 0/2] Reintroduce NEXT_BUDDY for EEVDF v2 Mel Gorman
2025-10-21 14:28 ` [PATCH 2/2] sched/fair: Reimplement NEXT_BUDDY to align with EEVDF goals Mel Gorman
2025-10-23 6:29 ` K Prateek Nayak
[not found] <20250714134429.19624-1-mgorman@techsingularity.net>
2025-07-14 13:44 ` Mel Gorman
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20251028154734.GS4068168@noisy.programming.kicks-ass.net \
--to=peterz@infradead.org \
--cc=clm@meta.com \
--cc=dietmar.eggemann@arm.com \
--cc=juri.lelli@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mgorman@techsingularity.net \
--cc=mingo@redhat.com \
--cc=vschneid@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox