Skip to contentSkip to conversation

cputime: Fixes

PATCH New Frederic Weisbecker 2 participants 5 messages Feb 23, 2013 – Feb 24, 2013

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

cover cputime: Fixes

Hi,

Just a couple of non-regression fixes on top of Linus tree.

Thanks.

Frederic Weisbecker (1):
cputime: Use local_clock() for full dynticks cputime accounting

Li Zhong (1):
cputime: fix compiler warning caused by timeval_to_cputime()

 include/asm-generic/cputime_nsecs.h |    2 +-
 kernel/sched/cputime.c              |    2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
Signature · 1 line
1.7.5.4
#1 lore raw

PATCH 1/2 cputime: fix compiler warning caused by timeval_to_cputime()

From: Li Zhong <[email protected]>

Saw the following compiler warning on next tree:

kernel/itimer.c: In function 'set_cpu_itimer':
kernel/itimer.c:152:2: warning: passing argument 1 of 'timeval_to_cputime' discards 'const' qualifier from pointer target type [enabled by default]
...

timeval_to_cputime() is always passed a constant timeval in argument, we need to teach the nsecs based cputime implementation about that.

Cc Ingo Molnar <[email protected]>
Cc Steven Rostedt <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Kevin Hilman <[email protected]>
---
 include/asm-generic/cputime_nsecs.h |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
include/asm-generic/cputime_nsecs.h +1 -1
index b6485ca..a8ece9a 100644
@@ -76,7 +76,7 @@ static inline void cputime_to_timespec(const cputime_t ct, struct timespec *val)
 /*
  * Convert cputime <-> timeval (msec)
  */
-static inline cputime_t timeval_to_cputime(struct timeval *val)
+static inline cputime_t timeval_to_cputime(const struct timeval *val)
 {
 	u64 ret = val->tv_sec * NSEC_PER_SEC + val->tv_usec * NSEC_PER_USEC;
 	return (__force cputime_t) ret;
Signature · 1 line
1.7.5.4
#2 lore raw

[tip:sched/urgent] cputime: Constify timeval_to_cputime(timeval) argument

cputime: Constify timeval_to_cputime(timeval) argument
Commit-ID:  c78a4bcd1a879b39fb7646c887b0c195f1018909
Gitweb:     http://git.kernel.org/tip/c78a4bcd1a879b39fb7646c887b0c195f1018909
Author:     Li Zhong <[email protected]>
AuthorDate: Sat, 23 Feb 2013 17:28:44 +0100
Committer:  Ingo Molnar <[email protected]>
CommitDate: Sun, 24 Feb 2013 12:57:15 +0100

cputime: Constify timeval_to_cputime(timeval) argument

Saw the following compiler warning on the linux-next tree:

kernel/itimer.c: In function 'set_cpu_itimer':
kernel/itimer.c:152:2: warning: passing argument 1 of 'timeval_to_cputime' discards 'const' qualifier from pointer target type [enabled by default]
...

timeval_to_cputime() is always passed a constant timeval in argument, we need to teach the nsecs based cputime implementation about that.

---
 include/asm-generic/cputime_nsecs.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
include/asm-generic/cputime_nsecs.h +1 -1
index b6485ca..a8ece9a 100644
@@ -76,7 +76,7 @@ static inline void cputime_to_timespec(const cputime_t ct, struct timespec *val)
 /*
  * Convert cputime <-> timeval (msec)
  */
-static inline cputime_t timeval_to_cputime(struct timeval *val)
+static inline cputime_t timeval_to_cputime(const struct timeval *val)
 {
 	u64 ret = val->tv_sec * NSEC_PER_SEC + val->tv_usec * NSEC_PER_USEC;
 	return (__force cputime_t) ret;
#3 lore raw

PATCH 2/2 cputime: Use local_clock() for full dynticks cputime accounting

In reply to Frederic Weisbecker

Running the full dynticks cputime accounting with preemptible kernel debugging trigger the following warning:

	[    4.488303] BUG: using smp_processor_id() in preemptible [00000000] code: init/1
	[    4.490971] caller is native_sched_clock+0x22/0x80
	[    4.493663] Pid: 1, comm: init Not tainted 3.8.0+ #13
	[    4.496376] Call Trace:
	[    4.498996]  [<ffffffff813410eb>] debug_smp_processor_id+0xdb/0xf0
	[    4.501716]  [<ffffffff8101e642>] native_sched_clock+0x22/0x80
	[    4.504434]  [<ffffffff8101db99>] sched_clock+0x9/0x10
	[    4.507185]  [<ffffffff81096ccd>] fetch_task_cputime+0xad/0x120
	[    4.509916]  [<ffffffff81096dd5>] task_cputime+0x35/0x60
	[    4.512622]  [<ffffffff810f146e>] acct_update_integrals+0x1e/0x40
	[    4.515372]  [<ffffffff8117d2cf>] do_execve_common+0x4ff/0x5c0
	[    4.518117]  [<ffffffff8117cf14>] ? do_execve_common+0x144/0x5c0
	[    4.520844]  [<ffffffff81867a10>] ? rest_init+0x160/0x160
	[    4.523554]  [<ffffffff8117d457>] do_execve+0x37/0x40
	[    4.526276]  [<ffffffff810021a3>] run_init_process+0x23/0x30
	[    4.528953]  [<ffffffff81867aac>] kernel_init+0x9c/0xf0
	[    4.531608]  [<ffffffff8188356c>] ret_from_fork+0x7c/0xb0

We use sched_clock() to perform and fixup the cputime accounting. However we are calling it with preemption enabled from the read side, which trigger the bug above.

To fix this up, use local_clock() instead. It takes care of preemption and also provide a more reliable clock source. This is welcome for this kind of statistic that is widely relied on in userspace.

---
 kernel/sched/cputime.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
kernel/sched/cputime.c +1 -1
index 9857329..ed12cbb 100644
@@ -604,7 +604,7 @@ static unsigned long long vtime_delta(struct task_struct *tsk)
 {
 	unsigned long long clock;
 
-	clock = sched_clock();
+	clock = local_clock();
 	if (clock < tsk->vtime_snap)
 		return 0;
 
Signature · 1 line
1.7.5.4
#4 lore raw

[tip:sched/urgent] cputime: Use local_clock() for full dynticks cputime accounting

cputime: Use local_clock() for full dynticks cputime accounting
Commit-ID:  7f6575f1fb963d5231afbceecd3feadb6ab58cd3
Gitweb:     http://git.kernel.org/tip/7f6575f1fb963d5231afbceecd3feadb6ab58cd3
Author:     Frederic Weisbecker <[email protected]>
AuthorDate: Sat, 23 Feb 2013 17:28:45 +0100
Committer:  Ingo Molnar <[email protected]>
CommitDate: Sun, 24 Feb 2013 12:57:16 +0100

cputime: Use local_clock() for full dynticks cputime accounting

Running the full dynticks cputime accounting with preemptible kernel debugging trigger the following warning:

	[    4.488303] BUG: using smp_processor_id() in preemptible [00000000] code: init/1
	[    4.490971] caller is native_sched_clock+0x22/0x80
	[    4.493663] Pid: 1, comm: init Not tainted 3.8.0+ #13
	[    4.496376] Call Trace:
	[    4.498996]  [<ffffffff813410eb>] debug_smp_processor_id+0xdb/0xf0
	[    4.501716]  [<ffffffff8101e642>] native_sched_clock+0x22/0x80
	[    4.504434]  [<ffffffff8101db99>] sched_clock+0x9/0x10
	[    4.507185]  [<ffffffff81096ccd>] fetch_task_cputime+0xad/0x120
	[    4.509916]  [<ffffffff81096dd5>] task_cputime+0x35/0x60
	[    4.512622]  [<ffffffff810f146e>] acct_update_integrals+0x1e/0x40
	[    4.515372]  [<ffffffff8117d2cf>] do_execve_common+0x4ff/0x5c0
	[    4.518117]  [<ffffffff8117cf14>] ? do_execve_common+0x144/0x5c0
	[    4.520844]  [<ffffffff81867a10>] ? rest_init+0x160/0x160
	[    4.523554]  [<ffffffff8117d457>] do_execve+0x37/0x40
	[    4.526276]  [<ffffffff810021a3>] run_init_process+0x23/0x30
	[    4.528953]  [<ffffffff81867aac>] kernel_init+0x9c/0xf0
	[    4.531608]  [<ffffffff8188356c>] ret_from_fork+0x7c/0xb0

We use sched_clock() to perform and fixup the cputime accounting. However we are calling it with preemption enabled from the read side, which trigger the bug above.

To fix this up, use local_clock() instead. It takes care of preemption and also provide a more reliable clock source. This is welcome for this kind of statistic that is widely relied on in userspace.

---
 kernel/sched/cputime.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
kernel/sched/cputime.c +1 -1
index 9857329..ed12cbb 100644
@@ -604,7 +604,7 @@ static unsigned long long vtime_delta(struct task_struct *tsk)
 {
 	unsigned long long clock;
 
-	clock = sched_clock();
+	clock = local_clock();
 	if (clock < tsk->vtime_snap)
 		return 0;