09 Dec, 2013
10 commits
-
The posix cpu timers code makes a heavy use of BUG_ON()
but none of these concern fatal issues that require
to stop the machine. So let's just warn the user when
some internal state slips out of our hands.Signed-off-by: Frederic Weisbecker
Cc: Thomas Gleixner
Cc: Ingo Molnar
Cc: Peter Zijlstra
Cc: Oleg Nesterov
Cc: Kosaki Motohiro
Cc: Andrew Morton -
The remaining uses of tasklist_lock were mostly about synchronizing
against sighand modifications, getting coherent and safe group samples
and also thread/process wide timers list handling.All of this is already safely synchronizable with the target's
sighand lock. Let's use it on these places instead.Also update the comments about locking.
Signed-off-by: Frederic Weisbecker
Cc: Thomas Gleixner
Cc: Ingo Molnar
Cc: Peter Zijlstra
Cc: Oleg Nesterov
Cc: Kosaki Motohiro
Cc: Andrew Morton -
Timer deletion doesn't need the tasklist lock.
We need to protect against:* concurrent access to the lists p->cputime_expires and
p->sighand->cputime_expires* task reaping that may also delete the timer list entry
* timer firing
We already hold the timer lock which protects us against concurrent
timer firing.The rest only need the targets sighand to be locked.
So hold it and drop the use of tasklist_lock there.Signed-off-by: Frederic Weisbecker
Cc: Thomas Gleixner
Cc: Ingo Molnar
Cc: Peter Zijlstra
Cc: Oleg Nesterov
Cc: Kosaki Motohiro
Cc: Andrew Morton -
There is no need for the tasklist_lock just to take a process
wide clock sample.All we need is to get a coherent sample that doesn't race with
exit() and exec():* exit() may be concurrently reaping a task and flushing its time
* sighand is unstable under exit() and exec(), and the latter also
result in group leader that can changeTo protect against these, locking the target's sighand is enough.
Signed-off-by: Frederic Weisbecker
Cc: Thomas Gleixner
Cc: Ingo Molnar
Cc: Peter Zijlstra
Cc: Oleg Nesterov
Cc: Kosaki Motohiro
Cc: Andrew Morton -
Consolidate the clock sampling common code used for both local
and remote targets.Note that this introduces a tiny user ABI change: if a
PID is passed to clock_gettime() along the clockid,
we used to forbid a process wide clock sample when that
PID doesn't belong to a group leader. Now after this patch
we allow process wide clock samples if that PID belongs to
the current task, even if the current task is not the
group leader.But local process wide clock samples are allowed if PID == 0
(current task) even if the current task is not the group leader.
So in the end this should be no big deal as this actually harmonize
the behaviour when the remote sample is actually a local one.Signed-off-by: Frederic Weisbecker
Cc: Thomas Gleixner
Cc: Ingo Molnar
Cc: Peter Zijlstra
Cc: Oleg Nesterov
Cc: Kosaki Motohiro
Cc: Andrew Morton -
a0b2062b0904ef07944c4a6e4d0f88ee44f1e9f2
("posix_timers: fix racy timer delta caching on task exit") forgot
to remove the arguments used for timer caching.Fix this leftover.
Signed-off-by: Frederic Weisbecker
Cc: Thomas Gleixner
Cc: Ingo Molnar
Cc: Peter Zijlstra
Cc: Oleg Nesterov
Cc: Kosaki Motohiro
Cc: Andrew Morton -
Now that we've removed all the optimizations that could
result in NULL timer's targets, we can remove all the
associated special case handling.Also add some warnings on NULL targets to spot any possible
leftover.Signed-off-by: Frederic Weisbecker
Cc: Thomas Gleixner
Cc: Ingo Molnar
Cc: Peter Zijlstra
Cc: Oleg Nesterov
Cc: Kosaki Motohiro
Cc: Andrew Morton -
When a timer's target is seen to be buried, for example on calls
to timer_gettime(), the posix cpu timers code behaves a bit
like a garbage collector and releases early the reference to the
task.Then again, this optimization complicates the code for no much
value: it's up to the user to release the timer and its associated
ressources by calling timer_delete() after it buries the target
tasks.Remove this to simplify the code.
Signed-off-by: Frederic Weisbecker
Cc: Thomas Gleixner
Cc: Ingo Molnar
Cc: Peter Zijlstra
Cc: Oleg Nesterov
Cc: Kosaki Motohiro
Cc: Andrew Morton -
Now that we removed dead thread posix cpu timers caching,
lets remove the dead process wide version. This caching
is similar to the per thread version but it should be even
more rare:* If the process id dead, we are not reading its timers
status from a thread belonging to its group since they
are all dead. So this caching only concern remote process
timers reads. Now posix cpu timers using itimers or timer_settime()
can't do remote process timers anyway so it's not even clear if there
is actually a user for this caching.* Unlike per thread timers caching, this only applies to
zombies targets. Buried targets' process wide timers return
0 values. But then again, timer_gettime() can't read remote
process timers, so if the process is dead, there can't be
any reader left anyway.Then again this caching seem to complicate the code for
corner cases that are probably not worth it. So lets get
rid of it.Also remove the sample snapshot on dying process timer
that is now useless, as suggested by Kosaki.Signed-off-by: Frederic Weisbecker
Cc: Thomas Gleixner
Cc: Ingo Molnar
Cc: Peter Zijlstra
Cc: Oleg Nesterov
Cc: Kosaki Motohiro
Cc: Andrew Morton -
When a task is exiting or has exited, its posix cpu timers
don't tick anymore and won't elapse further. It's too late
for them to expire.So any further call to timer_gettime() on these timers will
return the same remaining expiry time.The current code optimize this by caching the remaining delta
and storing it where we use to save the absolute expiration time.
This way, the future calls to timer_gettime() won't need to
compute the difference between the absolute expiration time and
the current time anymore.Now this optimization doesn't seem to bring much value. Computing
the timer remaining delta is not very costly. Fetching the timer
value OTOH can be costly in two ways:* CPUCLOCK_SCHED read requires to lock the target's rq. But some
optimizations are on the way to make task_sched_runtime() not holding
the rq lock of a non-running target.* CPUCLOCK_VIRT/CPUCLOCK_PROF read simply consist in fetching
current->utime/current->stime except when the system uses full
dynticks cputime accounting. The latter requires a per task lock
in order to correctly compute user and system time. But once the
target is dead, this lock shouldn't be contended anyway.All in one this caching doesn't seem to be justified.
Given that it complicates the code significantly for
few wins, let's remove it on single thread timers.Signed-off-by: Frederic Weisbecker
Cc: Thomas Gleixner
Cc: Ingo Molnar
Cc: Peter Zijlstra
Cc: Oleg Nesterov
Cc: Kosaki Motohiro
Cc: Andrew Morton
03 Dec, 2013
2 commits
-
A posix CPU timer can be rearmed while it is firing or after it is
notified with a signal. This can happen for example with timers that
were set with a non zero interval in timer_settime().This rearming can happen in two places:
1) On timer firing time, which happens on the target's tick. If the timer
can't trigger a signal because it is ignored, it reschedules itself
to honour the timer interval.2) On signal handling from the timer's notification target. This one
can be a different task than the timer's target itself. Once the
signal is notified, the notification target rearms the timer, again
to honour the timer interval.When a timer is rearmed, we need to notify the full dynticks CPUs
such that they restart their tick in case they are running tasks that
may have a share in elapsing this timer.Now the 1st case above handles full dynticks CPUs with a call to
posix_cpu_timer_kick_nohz() from the posix cpu timer firing code. But
the second case ignores the fact that some CPUs may run non-idle tasks
with their tick off. As a result, when a timer is resheduled after its signal
notification, the full dynticks CPUs may completely ignore it and not
tick on the timer as expectedThis patch fixes this bug by handling both cases in one. All we need
is to move the kick to the rearming common code in posix_cpu_timer_schedule().Signed-off-by: Frederic Weisbecker
Cc: Thomas Gleixner
Cc: Ingo Molnar
Cc: Peter Zijlstra
Cc: Oleg Nesterov
Cc: Steven Rostedt
Cc: Olivier Langlois -
After a posix cpu timer is set, a workqueue is scheduled in order to
kick the full dynticks CPUs and let them restart their tick if
necessary in case the task they are running is concerned by the
new timer.This kick is implemented by way of IPIs, which require interrupts
to be enabled, hence the need for a workqueue to raise them because
the posix cpu timer set path has interrupts disabled.Now if there is no full dynticks CPU on the system, the workqueue is
still scheduled but it simply won't send any IPI and return immediately.So lets spare that worqueue when it is not needed.
Signed-off-by: Frederic Weisbecker
Cc: Thomas Gleixner
Cc: Ingo Molnar
Cc: Peter Zijlstra
Cc: Oleg Nesterov
Cc: Steven Rostedt
03 Jul, 2013
5 commits
-
When a task exits, we perform a caching of the remaining cputime delta
before expiring of its timers.This is done from the following places:
* When the task is reaped. We iterate through its list of
posix cpu timers and store the remaining timer delta to
the timer struct instead of the absolute value.
(See posix_cpu_timers_exit() / posix_cpu_timers_exit_group() )* When we call posix_cpu_timer_get() or posix_cpu_timer_schedule().
If the timer's task is considered dying when watched from these
places, the same conversion from absolute to relative expiry time
is performed. Then the given task's reference is released.
(See clear_dead_task() ).The relevance of this caching is questionable but this is another
and deeper debate.The big issue here is that these two sources of caching don't mix
up very well together.More specifically, the caching can easily be done twice, resulting
in a wrong delta as it gets spuriously substracted a second time by
the elapsed clock. This can happen in the following scenario:1) The task exits and gets reaped: we call posix_cpu_timers_exit()
and the absolute timer expiry values are converted to a relative
delta.2) timer_gettime() -> posix_cpu_timer_get() is called and relies on
clear_dead_task() because tsk->exit_state == EXIT_DEAD.
The delta gets substracted again by the elapsed clock and we return
a wrong result.To fix this, just remove the caching done on task reaping time. It
doesn't bring much value on its own. The caching done from
posix_cpu_timer_get/schedule is enough.And it would also be hard to get it really right: we could make it put and
clear the target task in the timer struct so that readers know if they are
dealing with a relative cached of absolute value. But it would be racy.
The only safe way to do it would be to lock the itimer->it_lock so that we
know nobody reads the cputime expiry value while we modify it and its
target task reference. Doing so would involve some funny workarounds to
avoid circular lock against the sighand lock. There is just no reason to
maintain this.The user visible effect of this patch can be observed by running the
following code: it creates a subthread that launches a posix cputimer
which expires after 10 seconds. But then the subthread only busy loops for 2
seconds and exits. The parent reaps the subthread and read the timer value.
Its expected value should the be the initial timer's expiration value
minus the cputime elapsed in the subthread. Roughly 10 - 2 = 8 seconds:#include
#include
#include
#include
#includestatic timer_t id;
static struct itimerspec val = { .it_value.tv_sec = 10, }, new;static void *thread(void *unused)
{
int err;
struct timeval start, end, diff;timer_create(CLOCK_THREAD_CPUTIME_ID, NULL, &id);
if (err < 0) {
perror("Can't create timer\n");
return NULL;
}/* Arm 10 sec timer */
err = timer_settime(id, 0, &val, NULL);
if (err < 0) {
perror("Can't set timer\n");
return NULL;
}/* Exit after 2 seconds of execution */
gettimeofday(&start, NULL);
do {
gettimeofday(&end, NULL);
timersub(&end, &start, &diff);
} while (diff.tv_sec < 2);return NULL;
}int main(int argc, char **argv)
{
pthread_t pthread;
int err;err = pthread_create(&pthread, NULL, thread, NULL);
if (err) {
perror("Can't create thread\n");
return -1;
}
pthread_join(pthread, NULL);
/* Just wait a little bit to make sure the child got reaped */
sleep(1);
err = timer_gettime(id, &new);
if (err)
perror("Can't get timer value\n");
printf("%d %ld\n", new.it_value.tv_sec, new.it_value.tv_nsec);return 0;
}Before the patch:
$ ./posix_cpu_timers
6 2278074After the patch:
$ ./posix_cpu_timers
8 1158766Before the patch, the elapsed time got two more seconds spuriously accounted.
Signed-off-by: Frederic Weisbecker
Cc: Stanislaw Gruszka
Cc: Thomas Gleixner
Cc: Peter Zijlstra
Cc: Ingo Molnar
Cc: Oleg Nesterov
Cc: KOSAKI Motohiro
Cc: Olivier Langlois
Signed-off-by: Andrew Morton -
In order to re-arm a timer after it fired, we take a sample of the current
process or thread cputime.If the task is dying though, we don't arm anything but we cache the
remaining timer expiration delta for further reads.Something similar is performed in posix_cpu_timer_get() but here we forget
to take the process wide cputime sample before caching it.As a result we are storing random stack content, leading every further
reads of that timer to return junk values.Fix this by taking the appropriate sample in the case of process wide
timers.This probably doesn't matter much in practice because, at this stage, the
thread is the last one in the group and we reached exit_notify(). This
implies that we called exit_itimers() and there should be no more timers
to handle for that task.So this is likely dead code anyway but let's fix the current logic
and the warning that came along:kernel/posix-cpu-timers.c: In function 'posix_cpu_timer_schedule':
kernel/posix-cpu-timers.c:1127: warning: 'now' may be used uninitialized in this functionThen we can start to think further about cleaning up that code.
Reported-by: Andrew Morton
Reported-by: Chen Gang
Signed-off-by: Frederic Weisbecker
Cc: Stanislaw Gruszka
Cc: Thomas Gleixner
Cc: Peter Zijlstra
Cc: Ingo Molnar
Cc: Oleg Nesterov
Cc: Chen Gang
Cc: KOSAKI Motohiro
Cc: Olivier Langlois
Signed-off-by: Andrew Morton -
Consolidate the common code amongst per thread and per process timers list
on tick time.List traversal, expiry check and subsequent updates can be shared in a
common helper.Signed-off-by: Frederic Weisbecker
Cc: Stanislaw Gruszka
Cc: Thomas Gleixner
Cc: Peter Zijlstra
Cc: Ingo Molnar
Cc: Oleg Nesterov
Cc: KOSAKI Motohiro
Cc: Olivier Langlois
Signed-off-by: Andrew Morton -
Cleaning up the posix cpu timers on task exit shares some common code
among timer list types, most notably the list traversal and expiry time
update.Unify this in a common helper.
Signed-off-by: Frederic Weisbecker
Cc: Stanislaw Gruszka
Cc: Thomas Gleixner
Cc: Peter Zijlstra
Cc: Ingo Molnar
Cc: Oleg Nesterov
Cc: KOSAKI Motohiro
Cc: Olivier Langlois
Signed-off-by: Andrew Morton -
The posix cpu timer expiry time is stored in a union of two types: a 64
bits field if we rely on scheduler precise accounting, or a cputime_t if
we rely on jiffies.This results in quite some duplicate code and special cases to handle the
two types.Just unify this into a single 64 bits field. cputime_t can always fit
into it.Signed-off-by: Frederic Weisbecker
Cc: Stanislaw Gruszka
Cc: Thomas Gleixner
Cc: Peter Zijlstra
Cc: Ingo Molnar
Cc: Oleg Nesterov
Cc: KOSAKI Motohiro
Cc: Olivier Langlois
Signed-off-by: Andrew Morton
23 Apr, 2013
1 commit
-
The test that checks if a CPU can stop its tick from posix CPU
timers angle was mistakenly inverted.What we want is to prevent the tick from being stopped as long
as the current CPU's task runs a posix CPU timer.Fix this.
Signed-off-by: Frederic Weisbecker
Cc: Chris Metcalf
Cc: Christoph Lameter
Cc: Geoff Levand
Cc: Gilad Ben Yossef
Cc: Hakan Akkan
Cc: Ingo Molnar
Cc: Kevin Hilman
Cc: Li Zhong
Cc: Oleg Nesterov
Cc: Paul E. McKenney
Cc: Paul Gortmaker
Cc: Peter Zijlstra
Cc: Steven Rostedt
Cc: Thomas Gleixner
19 Apr, 2013
2 commits
-
Bring a new helper that the full dynticks infrastructure can
call in order to know if it can safely stop the tick from
the posix cpu timers subsystem point of view.Signed-off-by: Frederic Weisbecker
Cc: Chris Metcalf
Cc: Christoph Lameter
Cc: Geoff Levand
Cc: Gilad Ben Yossef
Cc: Hakan Akkan
Cc: Ingo Molnar
Cc: Kevin Hilman
Cc: Li Zhong
Cc: Oleg Nesterov
Cc: Paul E. McKenney
Cc: Paul Gortmaker
Cc: Peter Zijlstra
Cc: Steven Rostedt
Cc: Thomas Gleixner -
Kick the full dynticks CPUs when a posix cpu timer is enqueued by
way of a standard call to posix_cpu_timer_set() or set_process_cpu_timer().
This also include rescheduled firing timers.This way they can re-evaluate the state of (and possibly restart) their
tick against the new expiry modification.Signed-off-by: Frederic Weisbecker
Cc: Chris Metcalf
Cc: Christoph Lameter
Cc: Geoff Levand
Cc: Gilad Ben Yossef
Cc: Hakan Akkan
Cc: Ingo Molnar
Cc: Kevin Hilman
Cc: Li Zhong
Cc: Oleg Nesterov
Cc: Paul E. McKenney
Cc: Paul Gortmaker
Cc: Peter Zijlstra
Cc: Steven Rostedt
Cc: Thomas Gleixner
20 Feb, 2013
1 commit
-
Pull timer changes from Ingo Molnar:
"Main changes:- ntp: Add CONFIG_RTC_SYSTOHC: a generic RTC driver facility
complementing the existing CONFIG_RTC_HCTOSYS, which uses NTP to
keep the hardware clock updated.- posix-timers: Fix clock_adjtime to always return timex data on
success. This is changing the ABI, but no breakage was expected
and found - caution is warranted nevertheless.- platform persistent clock improvements/cleanups.
- clockevents: refactor timer broadcast handling to be more generic
and less duplicated with matching architecture code (mostly ARM
motivated.)- various fixes and cleanups"
* 'timers-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
timers/x86/hpet: Use HPET_COUNTER to specify the hpet counter in vread_hpet()
posix-cpu-timers: Fix nanosleep task_struct leak
clockevents: Fix generic broadcast for FEAT_C3STOP
time, Fix setting of hardware clock in NTP code
hrtimer: Prevent hrtimer_enqueue_reprogram race
clockevents: Add generic timer broadcast function
clockevents: Add generic timer broadcast receiver
timekeeping: Switch HAS_PERSISTENT_CLOCK to ALWAYS_USE_PERSISTENT_CLOCK
x86/time/rtc: Don't print extended CMOS year when reading RTC
x86: Select HAS_PERSISTENT_CLOCK on x86
timekeeping: Add CONFIG_HAS_PERSISTENT_CLOCK option
rtc: Skip the suspend/resume handling if persistent clock exist
timekeeping: Add persistent_clock_exist flag
posix-timers: Fix clock_adjtime to always return timex data on success
Round the calculated scale factor in set_cyc2ns_scale()
NTP: Add a CONFIG_RTC_SYSTOHC configuration
MAINTAINERS: Update John Stultz's email
time: create __getnstimeofday for WARNless calls
15 Feb, 2013
1 commit
-
The trinity fuzzer triggered a task_struct reference leak via
clock_nanosleep with CPU_TIMERs. do_cpu_nanosleep() calls
posic_cpu_timer_create(), but misses a corresponding
posix_cpu_timer_del() which leads to the task_struct reference leak.Reported-and-tested-by: Tommi Rantala
Signed-off-by: Stanislaw Gruszka
Cc: Dave Jones
Cc: John Stultz
Cc: Oleg Nesterov
Cc: stable@vger.kernel.org
Link: http://lkml.kernel.org/r/20130215100810.GF4392@redhat.com
Signed-off-by: Thomas Gleixner
28 Jan, 2013
1 commit
-
This is in preparation for the full dynticks feature. While
remotely reading the cputime of a task running in a full
dynticks CPU, we'll need to do some extra-computation. This
way we can account the time it spent tickless in userspace
since its last cputime snapshot.Signed-off-by: Frederic Weisbecker
Cc: Andrew Morton
Cc: Ingo Molnar
Cc: Li Zhong
Cc: Namhyung Kim
Cc: Paul E. McKenney
Cc: Paul Gortmaker
Cc: Peter Zijlstra
Cc: Steven Rostedt
Cc: Thomas Gleixner
20 Dec, 2012
1 commit
-
Pull random updates from Ted Ts'o:
"A few /dev/random improvements for the v3.8 merge window."* tag 'random_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/random:
random: Mix cputime from each thread that exits to the pool
random: prime last_data value per fips requirements
random: fix debug format strings
random: make it possible to enable debugging without rebuild
17 Dec, 2012
1 commit
-
When a thread exits mix it's cputime (userspace + kernelspace) to the entropy pool.
We don't know how "random" this is, so we use add_device_randomness that doesn't mess
with entropy count.Signed-off-by: Nick Kossifidis
Signed-off-by: Theodore Ts'o
29 Nov, 2012
1 commit
-
thread_group_cputime() is a general cputime API that is not only
used by posix cpu timer. Let's move this helper to sched code.Signed-off-by: Frederic Weisbecker
Cc: Ingo Molnar
Cc: Peter Zijlstra
Cc: Thomas Gleixner
Cc: Steven Rostedt
Cc: Paul Gortmaker
15 Dec, 2011
1 commit
-
Make cputime_t and cputime64_t nocast to enable sparse checking to
detect incorrect use of cputime. Drop the cputime macros for simple
scalar operations. The conversion macros are still needed.Signed-off-by: Martin Schwidefsky
26 Oct, 2011
1 commit
-
* 'core-locking-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (27 commits)
rtmutex: Add missing rcu_read_unlock() in debug_rt_mutex_print_deadlock()
lockdep: Comment all warnings
lib: atomic64: Change the type of local lock to raw_spinlock_t
locking, lib/atomic64: Annotate atomic64_lock::lock as raw
locking, x86, iommu: Annotate qi->q_lock as raw
locking, x86, iommu: Annotate irq_2_ir_lock as raw
locking, x86, iommu: Annotate iommu->register_lock as raw
locking, dma, ipu: Annotate bank_lock as raw
locking, ARM: Annotate low level hw locks as raw
locking, drivers/dca: Annotate dca_lock as raw
locking, powerpc: Annotate uic->lock as raw
locking, x86: mce: Annotate cmci_discover_lock as raw
locking, ACPI: Annotate c3_lock as raw
locking, oprofile: Annotate oprofilefs lock as raw
locking, video: Annotate vga console lock as raw
locking, latencytop: Annotate latency_lock as raw
locking, timer_stats: Annotate table_lock as raw
locking, rwsem: Annotate inner lock as raw
locking, semaphores: Annotate inner lock as raw
locking, sched: Annotate thread_group_cputimer as raw
...Fix up conflicts in kernel/posix-cpu-timers.c manually: making
cputimer->cputime a raw lock conflicted with the ABBA fix in commit
bcd5cff7216f ("cputimer: Cure lock inversion").
18 Oct, 2011
1 commit
-
There's a lock inversion between the cputimer->lock and rq->lock;
notably the two callchains involved are:update_rlimit_cpu()
sighand->siglock
set_process_cpu_timer()
cpu_timer_sample_group()
thread_group_cputimer()
cputimer->lock
thread_group_cputime()
task_sched_runtime()
->pi_lock
rq->lockscheduler_tick()
rq->lock
task_tick_fair()
update_curr()
account_group_exec()
cputimer->lockWhere the first one is enabling a CLOCK_PROCESS_CPUTIME_ID timer, and
the second one is keeping up-to-date.This problem was introduced by e8abccb7193 ("posix-cpu-timers: Cure
SMP accounting oddities").Cure the problem by removing the cputimer->lock and rq->lock nesting,
this leaves concurrent enablers doing duplicate work, but the time
wasted should be on the same order otherwise wasted spinning on the
lock and the greater-than assignment filter should ensure we preserve
monotonicity.Reported-by: Dave Jones
Reported-by: Simon Kirby
Signed-off-by: Peter Zijlstra
Cc: stable@kernel.org
Cc: Linus Torvalds
Cc: Martin Schwidefsky
Link: http://lkml.kernel.org/r/1318928713.21167.4.camel@twins
Signed-off-by: Thomas Gleixner
30 Sep, 2011
1 commit
-
David reported:
Attached below is a watered-down version of rt/tst-cpuclock2.c from
GLIBC. Just build it with "gcc -o test test.c -lpthread -lrt" or
similar.Run it several times, and you will see cases where the main thread
will measure a process clock difference before and after the nanosleep
which is smaller than the cpu-burner thread's individual thread clock
difference. This doesn't make any sense since the cpu-burner thread
is part of the top-level process's thread group.I've reproduced this on both x86-64 and sparc64 (using both 32-bit and
64-bit binaries).For example:
[davem@boricha build-x86_64-linux]$ ./test
process: before(0.001221967) after(0.498624371) diff(497402404)
thread: before(0.000081692) after(0.498316431) diff(498234739)
self: before(0.001223521) after(0.001240219) diff(16698)
[davem@boricha build-x86_64-linux]$The diff of 'process' should always be >= the diff of 'thread'.
I make sure to wrap the 'thread' clock measurements the most tightly
around the nanosleep() call, and that the 'process' clock measurements
are the outer-most ones.---
#include
#include
#include
#include
#include
#include
#include
#includestatic pthread_barrier_t barrier;
static void *chew_cpu(void *arg)
{
pthread_barrier_wait(&barrier);
while (1)
__asm__ __volatile__("" : : : "memory");
return NULL;
}int main(void)
{
clockid_t process_clock, my_thread_clock, th_clock;
struct timespec process_before, process_after;
struct timespec me_before, me_after;
struct timespec th_before, th_after;
struct timespec sleeptime;
unsigned long diff;
pthread_t th;
int err;err = clock_getcpuclockid(0, &process_clock);
if (err)
return 1;err = pthread_getcpuclockid(pthread_self(), &my_thread_clock);
if (err)
return 1;pthread_barrier_init(&barrier, NULL, 2);
err = pthread_create(&th, NULL, chew_cpu, NULL);
if (err)
return 1;err = pthread_getcpuclockid(th, &th_clock);
if (err)
return 1;pthread_barrier_wait(&barrier);
err = clock_gettime(process_clock, &process_before);
if (err)
return 1;err = clock_gettime(my_thread_clock, &me_before);
if (err)
return 1;err = clock_gettime(th_clock, &th_before);
if (err)
return 1;sleeptime.tv_sec = 0;
sleeptime.tv_nsec = 500000000;
nanosleep(&sleeptime, NULL);err = clock_gettime(th_clock, &th_after);
if (err)
return 1;err = clock_gettime(my_thread_clock, &me_after);
if (err)
return 1;err = clock_gettime(process_clock, &process_after);
if (err)
return 1;diff = process_after.tv_nsec - process_before.tv_nsec;
printf("process: before(%lu.%.9lu) after(%lu.%.9lu) diff(%lu)\n",
process_before.tv_sec, process_before.tv_nsec,
process_after.tv_sec, process_after.tv_nsec, diff);
diff = th_after.tv_nsec - th_before.tv_nsec;
printf("thread: before(%lu.%.9lu) after(%lu.%.9lu) diff(%lu)\n",
th_before.tv_sec, th_before.tv_nsec,
th_after.tv_sec, th_after.tv_nsec, diff);
diff = me_after.tv_nsec - me_before.tv_nsec;
printf("self: before(%lu.%.9lu) after(%lu.%.9lu) diff(%lu)\n",
me_before.tv_sec, me_before.tv_nsec,
me_after.tv_sec, me_after.tv_nsec, diff);return 0;
}This is due to us using p->se.sum_exec_runtime in
thread_group_cputime() where we iterate the thread group and sum all
data. This does not take time since the last schedule operation (tick
or otherwise) into account. We can cure this by using
task_sched_runtime() at the cost of having to take locks.This also means we can (and must) do away with
thread_group_sched_runtime() since the modified thread_group_cputime()
is now more accurate and would deadlock when called from
thread_group_sched_runtime().Aside of that it makes the function safe on 32 bit systems. The old
code added t->se.sum_exec_runtime unprotected. sum_exec_runtime is a
64bit value and could be changed on another cpu at the same time.Reported-by: David Miller
Signed-off-by: Peter Zijlstra
Cc: stable@kernel.org
Link: http://lkml.kernel.org/r/1314874459.7945.22.camel@twins
Tested-by: David Miller
Signed-off-by: Thomas Gleixner
13 Sep, 2011
1 commit
-
The thread_group_cputimer lock can be taken in atomic context and therefore
cannot be preempted on -rt - annotate it.In mainline this change documents the low level nature of
the lock - otherwise there's no functional difference. Lockdep
and Sparse checking will work as usual.Signed-off-by: Thomas Gleixner
Signed-off-by: Ingo Molnar
23 May, 2011
1 commit
-
Instead of iterating over all possible timer bases avoid it by marking
the active bases in the cpu base.Signed-off-by: Thomas Gleixner
Reviewed-by: Peter Zijlstra
31 Mar, 2011
1 commit
-
Fixes generated by 'codespell' and manually reviewed.
Signed-off-by: Lucas De Marchi
02 Feb, 2011
7 commits
-
Rename register_posix_clock() to posix_timers_register_clock(). That's
what the function really does. As a side effect this cleans up the
posix_clock namespace for the upcoming dynamic posix_clock
infrastructure.Signed-off-by: Thomas Gleixner
Tested-by: Richard Cochran
Cc: John Stultz
LKML-Reference: -
All functions are accessed via clock_posix_cpu now. So make them static.
Signed-off-by: Thomas Gleixner
Acked-by: John Stultz
Tested-by: Richard Cochran
LKML-Reference: -
Use the new kclock decoding function in clock_settime and cleanup all
kclocks which use the default functions. Rename the misnomed
common_clock_set() to posix_clock_realtime_set().Signed-off-by: Thomas Gleixner
Acked-by: John Stultz
Tested-by: Richard Cochran
LKML-Reference: -
CLOCK_THREAD_CPUTIME_ID implements stub functions for nanosleep and
nanosleep_restart, which return -EINVAL. That return value is
wrong. The correct return value is -ENOTSUP.Remove the stubs and let the new dispatch code return the correct
error code.Signed-off-by: Thomas Gleixner
Acked-by: John Stultz
Tested-by: Richard Cochran
LKML-Reference: -
posix timers still use the legacy arg0-arg3 members of
restart_block. Use restart_block.nanosleep insteadSigned-off-by: Thomas Gleixner
Acked-by: John Stultz
Tested-by: Richard Cochran
LKML-Reference: -
The CLOCK_DISPATCH() macro is a horrible magic. We call common
functions if a function pointer is not set. That's just backwards.To support dynamic file decriptor based clocks we need to cleanup that
dispatch logic.Create a k_clock struct clock_posix_cpu which has all the
posix-cpu-timer functions filled in. After the cleanup the functions
can be made static.Signed-off-by: Thomas Gleixner
Acked-by: John Stultz
Tested-by: Richard Cochran
LKML-Reference: -
Cosmetic. No functional change
Signed-off-by: Thomas Gleixner
Acked-by: John Stultz
Tested-by: Richard Cochran
LKML-Reference: