From: kosaki.motohiro@gmail.com
To: linux-kernel@vger.kernel.org
Cc: Olivier Langlois <olivier@trillion01.com>,
Thomas Gleixner <tglx@linutronix.de>,
Frederic Weisbecker <fweisbec@gmail.com>,
Ingo Molnar <mingo@kernel.org>,
Peter Zijlstra <peterz@infradead.org>,
KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Subject: [PATCH 2/8] posix-cpu-timers: fix acounting delta_exec twice
Date: Sun, 26 May 2013 17:35:42 -0400 [thread overview]
Message-ID: <1369604149-13016-4-git-send-email-kosaki.motohiro@gmail.com> (raw)
In-Reply-To: <1369604149-13016-1-git-send-email-kosaki.motohiro@gmail.com>
From: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Currently glibc rt/tst-cpuclock2 test(*) sporadically fails because
scheduler delta can be accounted twice from thread_group_cputimer()
and account_group_exec_runtime().
Finally, clock_nanosleep() wakes up earlier than an argument. This is posix
violation. This issue was introduced by commit d670ec1317 (posix-cpu-timers:
Cure SMP wobbles).
(*) http://sourceware.org/git/?p=glibc.git;a=blob;f=rt/tst-cpuclock2.c;h=6752721717f959e89c0d692b3f1ee082d507eec2;hb=HEAD
Cc: Olivier Langlois <olivier@trillion01.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
---
fs/binfmt_elf.c | 2 +-
fs/binfmt_elf_fdpic.c | 2 +-
include/linux/sched.h | 4 ++--
kernel/posix-cpu-timers.c | 15 ++++++++++-----
kernel/sched/core.c | 6 ++++--
kernel/sched/cputime.c | 6 +++---
6 files changed, 21 insertions(+), 14 deletions(-)
diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c
index f8a0b0e..9f7e13a 100644
--- a/fs/binfmt_elf.c
+++ b/fs/binfmt_elf.c
@@ -1326,7 +1326,7 @@ static void fill_prstatus(struct elf_prstatus *prstatus,
* This is the record for the group leader. It shows the
* group-wide total, not its individual thread total.
*/
- thread_group_cputime(p, &cputime);
+ thread_group_cputime(p, true, &cputime);
cputime_to_timeval(cputime.utime, &prstatus->pr_utime);
cputime_to_timeval(cputime.stime, &prstatus->pr_stime);
} else {
diff --git a/fs/binfmt_elf_fdpic.c b/fs/binfmt_elf_fdpic.c
index c166f32..fbcaeeb 100644
--- a/fs/binfmt_elf_fdpic.c
+++ b/fs/binfmt_elf_fdpic.c
@@ -1368,7 +1368,7 @@ static void fill_prstatus(struct elf_prstatus *prstatus,
* This is the record for the group leader. It shows the
* group-wide total, not its individual thread total.
*/
- thread_group_cputime(p, &cputime);
+ thread_group_cputime(p, true, &cputime);
cputime_to_timeval(cputime.utime, &prstatus->pr_utime);
cputime_to_timeval(cputime.stime, &prstatus->pr_stime);
} else {
diff --git a/include/linux/sched.h b/include/linux/sched.h
index 178a8d9..4878579 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -1841,7 +1841,7 @@ static inline void disable_sched_clock_irqtime(void) {}
#endif
extern unsigned long long
-task_sched_runtime(struct task_struct *task);
+task_sched_runtime(struct task_struct *task, bool add_delta);
/* sched_exec is called by processes performing an exec */
#ifdef CONFIG_SMP
@@ -2502,7 +2502,7 @@ static inline void current_clr_polling(void) { }
/*
* Thread group CPU time accounting.
*/
-void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times);
+void thread_group_cputime(struct task_struct *tsk, bool add_delta, struct task_cputime *times);
void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times);
static inline void thread_group_cputime_init(struct signal_struct *sig)
diff --git a/kernel/posix-cpu-timers.c b/kernel/posix-cpu-timers.c
index 42670e9..25447c5 100644
--- a/kernel/posix-cpu-timers.c
+++ b/kernel/posix-cpu-timers.c
@@ -237,7 +237,7 @@ static int cpu_clock_sample(const clockid_t which_clock, struct task_struct *p,
cpu->cpu = virt_ticks(p);
break;
case CPUCLOCK_SCHED:
- cpu->sched = task_sched_runtime(p);
+ cpu->sched = task_sched_runtime(p, true);
break;
}
return 0;
@@ -267,8 +267,13 @@ void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times)
* values through the TIMER_ABSTIME flag, therefore we have
* to synchronize the timer to the clock every time we start
* it.
+ *
+ * Do not add the current delta, because
+ * account_group_exec_runtime() will also this delta and we
+ * wouldn't want to double account time and get ahead of
+ * ourselves.
*/
- thread_group_cputime(tsk, &sum);
+ thread_group_cputime(tsk, false, &sum);
raw_spin_lock_irqsave(&cputimer->lock, flags);
cputimer->running = 1;
update_gt_cputime(&cputimer->cputime, &sum);
@@ -292,15 +297,15 @@ static int cpu_clock_sample_group(const clockid_t which_clock,
default:
return -EINVAL;
case CPUCLOCK_PROF:
- thread_group_cputime(p, &cputime);
+ thread_group_cputime(p, true, &cputime);
cpu->cpu = cputime.utime + cputime.stime;
break;
case CPUCLOCK_VIRT:
- thread_group_cputime(p, &cputime);
+ thread_group_cputime(p, true, &cputime);
cpu->cpu = cputime.utime;
break;
case CPUCLOCK_SCHED:
- thread_group_cputime(p, &cputime);
+ thread_group_cputime(p, true, &cputime);
cpu->sched = cputime.sum_exec_runtime;
break;
}
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 58453b8..a1e823b 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -2699,14 +2699,16 @@ unsigned long long task_delta_exec(struct task_struct *p)
* In case the task is currently running, return the runtime plus current's
* pending runtime that have not been accounted yet.
*/
-unsigned long long task_sched_runtime(struct task_struct *p)
+unsigned long long task_sched_runtime(struct task_struct *p, bool add_delta)
{
unsigned long flags;
struct rq *rq;
u64 ns = 0;
rq = task_rq_lock(p, &flags);
- ns = p->se.sum_exec_runtime + do_task_delta_exec(p, rq);
+ ns = p->se.sum_exec_runtime;
+ if (add_delta)
+ ns += do_task_delta_exec(p, rq);
task_rq_unlock(rq, p, &flags);
return ns;
diff --git a/kernel/sched/cputime.c b/kernel/sched/cputime.c
index cc2dc3e..3ca432c 100644
--- a/kernel/sched/cputime.c
+++ b/kernel/sched/cputime.c
@@ -277,7 +277,7 @@ static __always_inline bool steal_account_process_tick(void)
* Accumulate raw cputime values of dead tasks (sig->[us]time) and live
* tasks (sum on group iteration) belonging to @tsk's group.
*/
-void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times)
+void thread_group_cputime(struct task_struct *tsk, bool add_delta, struct task_cputime *times)
{
struct signal_struct *sig = tsk->signal;
cputime_t utime, stime;
@@ -297,7 +297,7 @@ void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times)
task_cputime(t, &utime, &stime);
times->utime += utime;
times->stime += stime;
- times->sum_exec_runtime += task_sched_runtime(t);
+ times->sum_exec_runtime += task_sched_runtime(t, add_delta);
} while_each_thread(tsk, t);
out:
rcu_read_unlock();
@@ -628,7 +628,7 @@ void thread_group_cputime_adjusted(struct task_struct *p, cputime_t *ut, cputime
{
struct task_cputime cputime;
- thread_group_cputime(p, &cputime);
+ thread_group_cputime(p, true, &cputime);
cputime_adjust(&cputime, &p->signal->prev_cputime, ut, st);
}
#endif /* !CONFIG_VIRT_CPU_ACCOUNTING_NATIVE */
--
1.7.1
next prev parent reply other threads:[~2013-05-26 21:36 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-05-26 21:35 [PATCH v5 0/8] posix timers fixlet kosaki.motohiro
2013-05-26 21:35 ` [PATCH 1/8] posix-cpu-timers: don't account cpu timer after stopped thread runtime accounting kosaki.motohiro
2013-05-26 21:35 ` kosaki.motohiro [this message]
2013-06-03 19:07 ` [PATCH 2/8] posix-cpu-timers: fix acounting delta_exec twice Frederic Weisbecker
2013-05-26 21:35 ` kosaki.motohiro
2013-05-26 21:35 ` [PATCH 3/8] posix-cpu-timers: fix wrong timer initialization kosaki.motohiro
2013-06-18 14:20 ` Frederic Weisbecker
2013-06-18 15:12 ` KOSAKI Motohiro
2013-06-19 20:56 ` Frederic Weisbecker
2013-05-26 21:35 ` [PATCH 4/8] posix-cpu-timers: timer functions should use timer time instead of clock time kosaki.motohiro
2013-05-26 21:35 ` [PATCH 5/8] posix-cpu-timers: check_thread_timers() uses task_sched_runtime() kosaki.motohiro
2013-05-26 21:35 ` [PATCH 6/8] sched: task_sched_runtime introduce micro optimization kosaki.motohiro
2013-06-18 14:27 ` Frederic Weisbecker
2013-06-18 15:17 ` KOSAKI Motohiro
2013-06-18 15:28 ` KOSAKI Motohiro
2013-06-18 17:18 ` Frederic Weisbecker
2013-06-20 8:41 ` KOSAKI Motohiro
2013-05-26 21:35 ` [PATCH 7/8] posix-cpu-timers: cleanup cpu_{clock,timer}_sample{,_group} kosaki.motohiro
2013-05-26 21:35 ` [PATCH 8/8] posix-cpu-timers: fix cputimer initialization mistake for {u,s}time kosaki.motohiro
2013-05-31 16:26 ` [PATCH v5 0/8] posix timers fixlet KOSAKI Motohiro
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=1369604149-13016-4-git-send-email-kosaki.motohiro@gmail.com \
--to=kosaki.motohiro@gmail.com \
--cc=fweisbec@gmail.com \
--cc=kosaki.motohiro@jp.fujitsu.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@kernel.org \
--cc=olivier@trillion01.com \
--cc=peterz@infradead.org \
--cc=tglx@linutronix.de \
/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
all inboxes | Powered by JetHome®