Skip to contentSkip to conversation

sched/pelt: simplify load_sum assignment code in attach_entity_load_avg()

PATCH In review Zhaoyu Liu 2 participants 5 messages Jun 21, 2022 – Jul 18, 2022 Scheduler

lore
Share & tools
lore

Zhaoyu Liu

author
#0 lore raw

PATCH 1/1 sched/pelt: simplify load_sum assignment code in attach_entity_load_avg()

In commit 40f5aa4c5eae ("sched/pelt: Fix attach_entity_load_avg() corner case"), these code was committed:

	if (se_weight(se) < se->avg.load_sum)
		se->avg.load_sum = div_u64(se->avg.load_sum, se_weight(se));
	else
		se->avg.load_sum = 1;

they could be replace with:

	se->avg.load_sum = div_u64(se->avg.load_sum, se_weight(se)) ?: 1;

to make the code cleaner.

---
 kernel/sched/fair.c | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)
kernel/sched/fair.c +1 -4
index 78795a997d9c..ed32f66bbd3d 100644
@@ -3881,10 +3881,7 @@ static void attach_entity_load_avg(struct cfs_rq *cfs_rq, struct sched_entity *s
 	se->avg.runnable_sum = se->avg.runnable_avg * divider;
 
 	se->avg.load_sum = se->avg.load_avg * divider;
-	if (se_weight(se) < se->avg.load_sum)
-		se->avg.load_sum = div_u64(se->avg.load_sum, se_weight(se));
-	else
-		se->avg.load_sum = 1;
+	se->avg.load_sum = div_u64(se->avg.load_sum, se_weight(se)) ?: 1;
 
 	enqueue_load_avg(cfs_rq, se);
 	cfs_rq->avg.util_avg += se->avg.util_avg;
Signature · 1 line
2.17.1
maintainer
#1 lore raw
Quoted 8 lines · Zhaoyu Liu

In commit 40f5aa4c5eae ("sched/pelt: Fix attach_entity_load_avg() corner case"), these code was committed:

        if (se_weight(se) < se->avg.load_sum)
                se->avg.load_sum = div_u64(se->avg.load_sum, se_weight(se));
        else
                se->avg.load_sum = 1;
Quoted 4 lines

they could be replace with:

        se->avg.load_sum = div_u64(se->avg.load_sum, se_weight(se)) ?: 1;

to make the code cleaner.

This quite subjective as I consider current version cleaner than your proposal

Quoted 25 lines
---
 kernel/sched/fair.c | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)
kernel/sched/fair.c +1 -5
index 78795a997d9c..ed32f66bbd3d 100644
@@ -3881,10 +3881,7 @@ static void attach_entity_load_avg(struct cfs_rq *cfs_rq, struct sched_entity *s
        se->avg.runnable_sum = se->avg.runnable_avg * divider;
 
        se->avg.load_sum = se->avg.load_avg * divider;
-       if (se_weight(se) < se->avg.load_sum)
-               se->avg.load_sum = div_u64(se->avg.load_sum, se_weight(se));
-       else
-               se->avg.load_sum = 1;
+       se->avg.load_sum = div_u64(se->avg.load_sum, se_weight(se)) ?: 1;
 
        enqueue_load_avg(cfs_rq, se);
        cfs_rq->avg.util_avg += se->avg.util_avg;
--

2.17.1

#2 lore raw
Quoted 10 lines · Vincent Guittot
Quoted 9 lines · Zhaoyu Liu

In commit 40f5aa4c5eae ("sched/pelt: Fix attach_entity_load_avg() corner case"),
these code was committed:

        if (se_weight(se) < se->avg.load_sum)
                se->avg.load_sum = div_u64(se->avg.load_sum, se_weight(se));
        else
                se->avg.load_sum = 1;
Quoted 7 lines
Quoted 5 lines

they could be replace with:

        se->avg.load_sum = div_u64(se->avg.load_sum, se_weight(se))

?: 1;

to make the code cleaner.

This quite subjective as I consider current version cleaner than your proposal

Thanks for your reply, vincent

Perhaps, this code is more concise, and this form can exist in many places in the kernel, and can be searched with 'grep "?: 1;" -nR kernel'

--
zackary

#3 lore raw
Quoted 21 lines · Zackary Liu
Quoted 10 lines · Vincent Guittot
Quoted 9 lines · Zhaoyu Liu

In commit 40f5aa4c5eae ("sched/pelt: Fix attach_entity_load_avg() corner case"),
these code was committed:

        if (se_weight(se) < se->avg.load_sum)
                se->avg.load_sum = div_u64(se->avg.load_sum, se_weight(se));
        else
                se->avg.load_sum = 1;
Quoted 7 lines
Quoted 5 lines

they could be replace with:

        se->avg.load_sum = div_u64(se->avg.load_sum, se_weight(se))

?: 1;

to make the code cleaner.

This quite subjective as I consider current version cleaner than your proposal

Quoted 7 lines

Thanks for your reply, vincent

Perhaps, this code is more concise, and this form can exist in many places in the kernel, and can be searched with 'grep "?: 1;" -nR kernel'

--
zackary

I have sent a patch couple days ago but still i don't get the reply, I am looking forward to your reply,
thank you

--
zackary

maintainer
#4 lore raw
Quoted 22 lines · Zackary Liu
Quoted 17 lines · Zackary Liu
Quoted 14 lines · Vincent Guittot
Quoted 13 lines · Zhaoyu Liu

In commit 40f5aa4c5eae ("sched/pelt: Fix attach_entity_load_avg() corner case"),
these code was committed:

        if (se_weight(se) < se->avg.load_sum)
                se->avg.load_sum = div_u64(se->avg.load_sum, se_weight(se));
        else
                se->avg.load_sum = 1;

they could be replace with:

        se->avg.load_sum = div_u64(se->avg.load_sum, se_weight(se))

?: 1;

Quoted 8 lines
Quoted 8 lines
Quoted 3 lines
Quoted 1 line

to make the code cleaner.

This quite subjective as I consider current version cleaner than your proposal

Thanks for your reply, vincent

Perhaps, this code is more concise, and this form can exist in many places in the kernel, and can be searched with 'grep "?: 1;" -nR kernel'

As said, the current code is cleaner and easier to read and I don't see any benefit from this patch.
So it's a NAK

Quoted 10 lines
Quoted 3 lines

--
zackary

I have sent a patch couple days ago but still i don't get the reply, I am looking forward to your reply,
thank you

--
zackary