lkml.org 
[lkml]   [2016]   [Sep]   [1]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
From
Subject[PATCH 2/3] sched/cputime: Make nr_migrations u32 on 32 bit cpus
Date
My previous patch:

("sched/cputime: Improve scalability of times()/clock_gettime() on 32 bit cpus")

add new sum_exec_runtime_seqclock fields to sched_entity struct.

This changed the alignment of other fields of the sched_entity struct.
To preserve previous alignment, which makes variables commonly used
together are kept in a cache line, make nr_migrations field u32 on
32 bit architectures.

This field is used only for debugging and basically nr_migrations
should not be enormous number.

Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
---
include/linux/sched.h | 5 +++--
kernel/sched/debug.c | 4 ++--
2 files changed, 5 insertions(+), 4 deletions(-)

diff --git a/include/linux/sched.h b/include/linux/sched.h
index 65714bc..b9d9333 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -1332,9 +1332,10 @@ struct sched_entity {
u64 vruntime;
u64 prev_sum_exec_runtime;

+#ifdef CONFIG_64BIT
u64 nr_migrations;
-
-#ifndef CONFIG_64BIT
+#else
+ u32 nr_migrations;
seqcount_t sum_exec_runtime_seqcount;
#endif

diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c
index 2a0a999..db85816 100644
--- a/kernel/sched/debug.c
+++ b/kernel/sched/debug.c
@@ -921,8 +921,8 @@ void proc_sched_show_task(struct task_struct *p, struct seq_file *m)

avg_per_cpu = p->se.sum_exec_runtime;
if (p->se.nr_migrations) {
- avg_per_cpu = div64_u64(avg_per_cpu,
- p->se.nr_migrations);
+ avg_per_cpu = div64_ul(avg_per_cpu,
+ p->se.nr_migrations);
} else {
avg_per_cpu = -1LL;
}
--
1.8.3.1
\
 
 \ /
  Last update: 2016-09-17 09:58    [W:2.209 / U:0.664 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site