| Message ID | 1467383054-1959-1-git-send-email-shreyas@linux.vnet.ibm.com (mailing list archive) |
|---|---|
| State | Accepted, archived |
| Delegated to: | Rafael Wysocki |
| Headers | show |
On 07/01/2016 04:24 PM, Shreyas B. Prabhu wrote: > Snooze is a poll idle state in powernv and pseries platforms. Snooze > has a timeout so that if a cpu stays in snooze for more than target > residency of the next available idle state, then it would exit thereby > giving chance to the cpuidle governor to re-evaluate and > promote the cpu to a deeper idle state. Therefore whenever snooze exits > due to this timeout, its last_residency will be target_residency of next > deeper state. > > commit e93e59ce5b85 ("cpuidle: Replace ktime_get() with local_clock()") > changed the math around last_residency calculation. Specifically, while > converting last_residency value from nanoseconds to microseconds it does > right shift by 10. Due to this, in snooze timeout exit scenarios > last_residency calculated is roughly 2.3% less than target_residency of > next available state. This pattern is picked up get_typical_interval() > in the menu governor and therefore expected_interval in menu_select() is > frequently less than the target_residency of any state but snooze. > > Due to this we are entering snooze at a higher rate, thereby affecting > the single thread performance. > > Fix this by using precise division via ktime_us_delta. > > Reported-by: Anton Blanchard <anton@samba.org> > Bisected-by: Shilpasri G Bhat <shilpa.bhat@linux.vnet.ibm.com> > Signed-off-by: Shreyas B. Prabhu <shreyas@linux.vnet.ibm.com> > --- Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
On 02/07/16 00:24, Shreyas B. Prabhu wrote: > Snooze is a poll idle state in powernv and pseries platforms. Snooze > has a timeout so that if a cpu stays in snooze for more than target > residency of the next available idle state, then it would exit thereby > giving chance to the cpuidle governor to re-evaluate and > promote the cpu to a deeper idle state. Therefore whenever snooze exits > due to this timeout, its last_residency will be target_residency of next > deeper state. > > commit e93e59ce5b85 ("cpuidle: Replace ktime_get() with local_clock()") > changed the math around last_residency calculation. Specifically, while > converting last_residency value from nanoseconds to microseconds it does > right shift by 10. Due to this, in snooze timeout exit scenarios > last_residency calculated is roughly 2.3% less than target_residency of > next available state. This pattern is picked up get_typical_interval() > in the menu governor and therefore expected_interval in menu_select() is > frequently less than the target_residency of any state but snooze. > > Due to this we are entering snooze at a higher rate, thereby affecting > the single thread performance. > > Fix this by using precise division via ktime_us_delta. > > Reported-by: Anton Blanchard <anton@samba.org> > Bisected-by: Shilpasri G Bhat <shilpa.bhat@linux.vnet.ibm.com> > Signed-off-by: Shreyas B. Prabhu <shreyas@linux.vnet.ibm.com> > --- This looks so much cleaner :) Acked-by: Balbir Singh <bsingharora@gmail.com> Balbir -- To unsubscribe from this list: send the line "unsubscribe linux-pm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Mon, Jul 4, 2016 at 8:10 AM, Balbir Singh <bsingharora@gmail.com> wrote: > > > On 02/07/16 00:24, Shreyas B. Prabhu wrote: >> Snooze is a poll idle state in powernv and pseries platforms. Snooze >> has a timeout so that if a cpu stays in snooze for more than target >> residency of the next available idle state, then it would exit thereby >> giving chance to the cpuidle governor to re-evaluate and >> promote the cpu to a deeper idle state. Therefore whenever snooze exits >> due to this timeout, its last_residency will be target_residency of next >> deeper state. >> >> commit e93e59ce5b85 ("cpuidle: Replace ktime_get() with local_clock()") >> changed the math around last_residency calculation. Specifically, while >> converting last_residency value from nanoseconds to microseconds it does >> right shift by 10. Due to this, in snooze timeout exit scenarios >> last_residency calculated is roughly 2.3% less than target_residency of >> next available state. This pattern is picked up get_typical_interval() >> in the menu governor and therefore expected_interval in menu_select() is >> frequently less than the target_residency of any state but snooze. >> >> Due to this we are entering snooze at a higher rate, thereby affecting >> the single thread performance. >> >> Fix this by using precise division via ktime_us_delta. >> >> Reported-by: Anton Blanchard <anton@samba.org> >> Bisected-by: Shilpasri G Bhat <shilpa.bhat@linux.vnet.ibm.com> >> Signed-off-by: Shreyas B. Prabhu <shreyas@linux.vnet.ibm.com> >> --- > > This looks so much cleaner :) > > Acked-by: Balbir Singh <bsingharora@gmail.com> Patch applied, thanks! -- To unsubscribe from this list: send the line "unsubscribe linux-pm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
============= - Replacing approximation division with ktime_us_delta. Changes in v4 ============= - Increasing the threshold upto which approximation can be used. - Removed explicit cast. Instead added a comment saying why cast is safe. Changes in v3 ============= - Using approximation suggested by David Changes in v2 ============= - Fixing it in the cpuidle core code instead of driver code. drivers/cpuidle/cpuidle.c | 12 ++++-------- 1 file changed, 4 insertions(+), 8 deletions(-) diff --git a/drivers/cpuidle/cpuidle.c b/drivers/cpuidle/cpuidle.c index a4d0059..c73207a 100644 --- a/drivers/cpuidle/cpuidle.c +++ b/drivers/cpuidle/cpuidle.c @@ -173,7 +173,7 @@ int cpuidle_enter_state(struct cpuidle_device *dev, struct cpuidle_driver *drv, struct cpuidle_state *target_state = &drv->states[index]; bool broadcast = !!(target_state->flags & CPUIDLE_FLAG_TIMER_STOP); - u64 time_start, time_end; + ktime_t time_start, time_end; s64 diff; /* @@ -195,13 +195,13 @@ int cpuidle_enter_state(struct cpuidle_device *dev, struct cpuidle_driver *drv, sched_idle_set_state(target_state); trace_cpu_idle_rcuidle(index, dev->cpu); - time_start = local_clock(); + time_start = ns_to_ktime(local_clock()); stop_critical_timings(); entered_state = target_state->enter(dev, drv, index); start_critical_timings(); - time_end = local_clock(); + time_end = ns_to_ktime(local_clock()); trace_cpu_idle_rcuidle(PWR_EVENT_EXIT, dev->cpu); /* The cpu is no longer idle or about to enter idle. */ @@ -217,11 +217,7 @@ int cpuidle_enter_state(struct cpuidle_device *dev, struct cpuidle_driver *drv, if (!cpuidle_state_is_coupled(drv, index)) local_irq_enable(); - /* - * local_clock() returns the time in nanosecond, let's shift - * by 10 (divide by 1024) to have microsecond based time. - */ - diff = (time_end - time_start) >> 10; + diff = ktime_us_delta(time_end, time_start); if (diff > INT_MAX) diff = INT_MAX;
Snooze is a poll idle state in powernv and pseries platforms. Snooze has a timeout so that if a cpu stays in snooze for more than target residency of the next available idle state, then it would exit thereby giving chance to the cpuidle governor to re-evaluate and promote the cpu to a deeper idle state. Therefore whenever snooze exits due to this timeout, its last_residency will be target_residency of next deeper state. commit e93e59ce5b85 ("cpuidle: Replace ktime_get() with local_clock()") changed the math around last_residency calculation. Specifically, while converting last_residency value from nanoseconds to microseconds it does right shift by 10. Due to this, in snooze timeout exit scenarios last_residency calculated is roughly 2.3% less than target_residency of next available state. This pattern is picked up get_typical_interval() in the menu governor and therefore expected_interval in menu_select() is frequently less than the target_residency of any state but snooze. Due to this we are entering snooze at a higher rate, thereby affecting the single thread performance. Fix this by using precise division via ktime_us_delta. Reported-by: Anton Blanchard <anton@samba.org> Bisected-by: Shilpasri G Bhat <shilpa.bhat@linux.vnet.ibm.com> Signed-off-by: Shreyas B. Prabhu <shreyas@linux.vnet.ibm.com> --- Changes in v5