On Wed, 19 Dec 2018 at 17:36, Ulf Hansson wrote:
>
> On Wed, 19 Dec 2018 at 14:26, Vincent Guittot
> wrote:
> >
> > On Wed, 19 Dec 2018 at 11:43, Ulf Hansson wrote:
> > >
> > > On Wed, 19 Dec 2018 at 11:34, Vincent Guittot
> > > wrote:
> >
On Wed, 19 Dec 2018 at 16:30, Valentin Schneider
wrote:
>
> On 19/12/2018 15:20, Vincent Guittot wrote:
> [...]
> >> Oh yes, I never said it didn't work - I was doing some investigation on
> >> the reason as to why we'd need this fix, because it's wa
On Wed, 19 Dec 2018 at 10:58, Ulf Hansson wrote:
>
> On Tue, 18 Dec 2018 at 15:55, Vincent Guittot
> wrote:
> >
> > Some drivers (like i915/drm) need to get the accounted suspended time.
> > pm_runtime_accounted_time_get() will return the suspended or active
> >
On Wed, 19 Dec 2018 at 16:11, Valentin Schneider
wrote:
>
> On 19/12/2018 15:05, Vincent Guittot wrote:
> [...]
> >> True, I had a look at the trace and there doesn't seem to be any running
> >> task on that CPU. That's a separate matter however - the rounding
On Wed, 19 Dec 2018 at 15:59, Valentin Schneider
wrote:
>
>
>
> On 19/12/2018 13:39, Vincent Guittot wrote:
> [...]
> >> I used that setup out of convenience for myself, but AFAICT that use-case
> >> just stresses that issue.
> >
> > After looking at y
On Wed, 19 Dec 2018 at 12:58, Valentin Schneider
wrote:
>
> On 19/12/2018 08:32, Vincent Guittot wrote:
> [...]
> > This is another UC, asym packing is used at SMT level for now and we
> > don't face this kind of problem, it has been also tested and DynamiQ
> > c
On Wed, 19 Dec 2018 at 12:16, Valentin Schneider
wrote:
>
> On 19/12/2018 08:27, Vincent Guittot wrote:
> [...]
> >> Wouldn't LBF_ALL_PINNED cover all relevant cases? It's set at the very top
> >> of the 'if (busiest->nr_running > 1)' block a
On Wed, 19 Dec 2018 at 11:43, Ulf Hansson wrote:
>
> On Wed, 19 Dec 2018 at 11:34, Vincent Guittot
> wrote:
> >
> > On Wed, 19 Dec 2018 at 11:21, Ulf Hansson wrote:
> > >
> > > > > >
> > > > > > diff --git a/driver
On Wed, 19 Dec 2018 at 11:21, Ulf Hansson wrote:
>
> On Wed, 19 Dec 2018 at 11:11, Vincent Guittot
> wrote:
> >
> > On Wed, 19 Dec 2018 at 10:58, Ulf Hansson wrote:
> > >
> > > On Tue, 18 Dec 2018 at 15:55, Vincent Guittot
> > > wrote:
> >
On Wed, 19 Dec 2018 at 10:58, Ulf Hansson wrote:
>
> On Tue, 18 Dec 2018 at 15:55, Vincent Guittot
> wrote:
> >
> > Some drivers (like i915/drm) need to get the accounted suspended time.
> > pm_runtime_accounted_time_get() will return the suspended or active
> >
On Tue, 18 Dec 2018 at 18:32, Valentin Schneider
wrote:
>
> On 14/12/2018 16:01, Vincent Guittot wrote:
> > When check_asym_packing() is triggered, the imbalance is set to :
> > busiest_stat.avg_load * busiest_stat.group_capacity / SCHED_CAPACITY_SCALE
> > busiest_stat.av
On Tue, 18 Dec 2018 at 15:09, Valentin Schneider
wrote:
>
> On 18/12/2018 13:23, Vincent Guittot wrote:
> [...]
> >> Ah, I think I get it: you're saying that this balance_interval increase
> >> is done because it is always assumed we do an active balance w
Some drivers (like i915/drm) need to get the accounted suspended time.
pm_runtime_accounted_time_get() will return the suspended or active
accounted time until now.
Signed-off-by: Vincent Guittot
---
drivers/base/power/runtime.c | 26 ++
include/linux/pm_runtime.h | 2
Use the new pm runtime interface to get the accounted suspended time:
pm_runtime_accounted_time_get()
Signed-off-by: Vincent Guittot
---
drivers/gpu/drm/i915/i915_pmu.c | 18 --
drivers/gpu/drm/i915/i915_pmu.h | 4 ++--
2 files changed, 10 insertions(+), 12 deletions(-)
diff
compilation error
Changes since v1:
- updated commit message of patch 1
- Added patches 2 & 3 to move runtime_pm accounting on raw ns
Thara Gopinath (1):
PM/runtime:Replace jiffies based accounting with ktime based
accounting
Vincent Guittot (2):
PM/runtime: Add a new interface to
from ktime to raw nsec]
Signed-off-by: Vincent Guittot
---
drivers/base/power/runtime.c | 10 +-
drivers/base/power/sysfs.c | 11 ---
include/linux/pm.h | 6 +++---
3 files changed, 16 insertions(+), 11 deletions(-)
diff --git a/drivers/base/power/runtime.c b/drivers
On Tue, 18 Dec 2018 at 12:46, Valentin Schneider
wrote:
>
> On 18/12/2018 09:32, Vincent Guittot wrote:
> [...]
> > In this asym packing case, It has nothing to do with pinned tasks and
> > that's the root cause of the problem:
> > the active balance triggered by
On Tue, 18 Dec 2018 at 11:03, Rafael J. Wysocki wrote:
>
> On Tue, Dec 18, 2018 at 10:58 AM Vincent Guittot
> wrote:
> >
> > On Tue, 18 Dec 2018 at 10:57, Rafael J. Wysocki wrote:
> > >
> > > On Mon, Dec 17, 2018 at 3:22 PM Vincent Guittot
> > >
On Tue, 18 Dec 2018 at 10:57, Rafael J. Wysocki wrote:
>
> On Mon, Dec 17, 2018 at 3:22 PM Vincent Guittot
> wrote:
> >
> > On Fri, 14 Dec 2018 at 15:36, Ulf Hansson wrote:
> > >
> > > On Fri, 14 Dec 2018 at 15:22, Vincent Guittot
> > > wrote:
&
Hi Valentin,
On Mon, 17 Dec 2018 at 17:56, Valentin Schneider
wrote:
>
> Hi Vincent,
>
> About time I had a look at this one...
>
> On 14/12/2018 16:01, Vincent Guittot wrote:
> > In case of active balance, we increase the balance interval to cover
> > pin
On Mon, 17 Dec 2018 at 17:59, Valentin Schneider
wrote:
>
> Hi Vincent,
>
> On 14/12/2018 16:01, Vincent Guittot wrote:
> > newly idle load balance is not always triggered when a cpu becomes idle.
> > This prevent the scheduler to get a chance to migrate task for asym pa
On Fri, 14 Dec 2018 at 15:36, Ulf Hansson wrote:
>
> On Fri, 14 Dec 2018 at 15:22, Vincent Guittot
> wrote:
> >
> > With jiffies been replaced by raw ns in PM core accounting, 915 driver is
> > updated to use this new time infrastructure.
> >
&
enough to skip the rq in find_busiest_queue and prevents asym
migration to happen.
Add 1 to the avg_load to make sure that the targeted cpu will not be
skipped unexpectidly.
Signed-off-by: Vincent Guittot
---
kernel/sched/fair.c | 6 ++
1 file changed, 6 insertions(+)
diff --git a/kernel
newly idle load balance is not always triggered when a cpu becomes idle.
This prevent the scheduler to get a chance to migrate task for asym packing.
Enable active migration because of asym packing during idle load balance too.
Signed-off-by: Vincent Guittot
---
kernel/sched/fair.c | 2 +-
1
for asym_packing can be easily delayed for hundreds of ms
because of this all_pinned detection mecanism.
Signed-off-by: Vincent Guittot
---
kernel/sched/fair.c | 27 +++
1 file changed, 15 insertions(+), 12 deletions(-)
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
Vincent Guittot (3):
sched/fair: fix rounding issue for asym packing
sched/fair: trigger asym_packing during idle load balance
sched/fair: fix unnecessary increase of balance interval
kernel/sched/fair.c | 33 +
1 file changed, 21 insertions(+), 12 deletions
filename
133042996805276 402768 20512343138fe57 vmlinux
Signed-off-by: Vincent Guittot
---
drivers/base/power/runtime.c | 63
include/linux/pm.h | 5 ++--
include/linux/pm_runtime.h | 6 ++---
3 files changed, 40
from ktime to raw nsec]
Signed-off-by: Vincent Guittot
---
drivers/base/power/runtime.c | 10 +-
drivers/base/power/sysfs.c | 11 ---
include/linux/pm.h | 6 +++---
3 files changed, 16 insertions(+), 11 deletions(-)
diff --git a/drivers/base/power/runtime.c b/drivers
With jiffies been replaced by raw ns in PM core accounting, 915 driver is
updated to use this new time infrastructure.
Signed-off-by: Vincent Guittot
---
drivers/gpu/drm/i915/i915_pmu.c | 12 ++--
drivers/gpu/drm/i915/i915_pmu.h | 4 ++--
2 files changed, 8 insertions(+), 8 deletions
drm/i915 driver that uses PM core fields
Changes since v1:
- updated commit message of patch 1
- Added patches 2 & 3 to move runtime_pm accounting on raw ns
Thara Gopinath (1):
PM/runtime:Replace jiffies based accouting with ktime based accounting
Vincent Guittot (2):
PM/pm_runtime:
On Thu, 13 Dec 2018 at 14:52, Peter Zijlstra wrote:
>
> On Tue, Aug 07, 2018 at 05:56:27PM +0200, Vincent Guittot wrote:
> > +static inline bool
> > +asym_active_balance(enum cpu_idle_type idle, unsigned int flags, int dst,
> > int src)
> > {
>
On Thu, 13 Dec 2018 at 14:12, Peter Zijlstra wrote:
>
> On Thu, Dec 13, 2018 at 12:04:20PM +0100, Vincent Guittot wrote:
> > On Thu, 13 Dec 2018 at 11:44, Peter Zijlstra wrote:
> > >
> > > On Mon, Sep 10, 2018 at 04:43:09PM +0200, Vincent Guittot wrote:
> > &
On Thu, 13 Dec 2018 at 11:44, Peter Zijlstra wrote:
>
> On Mon, Sep 10, 2018 at 04:43:09PM +0200, Vincent Guittot wrote:
> > When CPUs have different capacity because of RT/DL tasks or
> > micro-architecture or max frequency differences, there are situation where
> &
Hi Peter, Ingo,
On Fri, 14 Sep 2018 at 18:26, Vincent Guittot
wrote:
>
> On Fri, 14 Sep 2018 at 05:22, Valentin Schneider
> wrote:
> >
> > Hi,
> >
> > On 10/09/18 07:43, Vincent Guittot wrote:
> > > When CPUs have different capacity because of RT/D
Commit-ID: 765d0af19f5f388a34bf4533378f8398b72ded46
Gitweb: https://git.kernel.org/tip/765d0af19f5f388a34bf4533378f8398b72ded46
Author: Vincent Guittot
AuthorDate: Wed, 29 Aug 2018 15:19:11 +0200
Committer: Ingo Molnar
CommitDate: Tue, 11 Dec 2018 15:16:57 +0100
sched/topology: Remove
Hi Arnd,
On Mon, 10 Dec 2018 at 22:01, Arnd Bergmann wrote:
>
> Marking the variable static showed that it's only used for
> SMP builds, as seen from this warning:
>
> kernel/sched/fair.c:119:21: error: 'capacity_margin' defined but not used
> [-Werror=unused-variable]
> static unsigned int cap
On Mon, 10 Dec 2018 at 17:33, Vincent Guittot
wrote:
>
> On Mon, 10 Dec 2018 at 17:29, Steven Sistare
> wrote:
> >
> > On 12/10/2018 11:10 AM, Vincent Guittot wrote:
> > > Hi Steven,
> > >
> > > On Thu, 6 Dec 2018 at 22:38, Steve Sistare
>
On Mon, 10 Dec 2018 at 17:29, Steven Sistare wrote:
>
> On 12/10/2018 11:10 AM, Vincent Guittot wrote:
> > Hi Steven,
> >
> > On Thu, 6 Dec 2018 at 22:38, Steve Sistare
> > wrote:
> >>
> >> When a CPU has no more CFS tasks to run, and idle_bala
Hi Steven,
On Thu, 6 Dec 2018 at 22:38, Steve Sistare wrote:
>
> When a CPU has no more CFS tasks to run, and idle_balance() fails to
> find a task, then attempt to steal a task from an overloaded CPU in the
> same LLC. Maintain and use a bitmap of overloaded CPUs to efficiently
> identify candid
Hi Thara,
On Tue, 9 Oct 2018 at 18:25, Thara Gopinath wrote:
>
> Introduce support in CFS periodic tick to trigger the process of
> computing average thermal pressure for a cpu.
>
> Signed-off-by: Thara Gopinath
> ---
> kernel/sched/fair.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --g
On Thu, 29 Nov 2018 at 16:00, Patrick Bellasi wrote:
>
> On 29-Nov 11:43, Vincent Guittot wrote:
> > On Wed, 28 Nov 2018 at 17:35, Patrick Bellasi
> > wrote:
> > > On 28-Nov 16:42, Vincent Guittot wrote:
> > > > On Wed, 28 Nov 2018 at 16:21, Patrick Bella
On Wed, 28 Nov 2018 at 17:35, Patrick Bellasi wrote:
>
> On 28-Nov 16:42, Vincent Guittot wrote:
> > On Wed, 28 Nov 2018 at 16:21, Patrick Bellasi
> > wrote:
> > >
> > > On 28-Nov 15:55, Vincent Guittot wrote:
> > > > On Wed, 28 No
On Wed, 28 Nov 2018 at 16:21, Patrick Bellasi wrote:
>
> On 28-Nov 15:55, Vincent Guittot wrote:
> > On Wed, 28 Nov 2018 at 15:40, Patrick Bellasi
> > wrote:
> > >
> > > On 28-Nov 14:33, Vincent Guittot wrote:
> > > > On Wed, 28 No
On Wed, 28 Nov 2018 at 15:40, Patrick Bellasi wrote:
>
> On 28-Nov 14:33, Vincent Guittot wrote:
> > On Wed, 28 Nov 2018 at 12:53, Patrick Bellasi
> > wrote:
> > >
> > > On 28-Nov 11:02, Peter Zijlstra wrote:
> > > > On Wed, Nov 28,
On Wed, 28 Nov 2018 at 14:33, Vincent Guittot
wrote:
>
> On Wed, 28 Nov 2018 at 12:53, Patrick Bellasi wrote:
> >
> > On 28-Nov 11:02, Peter Zijlstra wrote:
> > > On Wed, Nov 28, 2018 at 10:54:13AM +0100, Vincent Guittot wrote:
> > >
> > > > I
On Wed, 28 Nov 2018 at 12:53, Patrick Bellasi wrote:
>
> On 28-Nov 11:02, Peter Zijlstra wrote:
> > On Wed, Nov 28, 2018 at 10:54:13AM +0100, Vincent Guittot wrote:
> >
> > > Is there anything else that I should do for these patches ?
> >
> > IIRC, Morten m
Hi,
On Tue, 20 Nov 2018 at 11:55, Vincent Guittot
wrote:
>
> The current implementation of load tracking invariance scales the
> contribution with current frequency and uarch performance (only for
> utilization) of the CPU. One main result of this formula is that the
> figure
ader with misleading/out-of-date comments.
>
> [1]
> http://lkml.kernel.org/r/cakftptbja-ocbrko6__npqwl3+hljzk7riccpu1r7ydo-ep...@mail.gmail.com
>
> Suggested-by: Vincent Guittot
> Signed-off-by: Andrea Parri
> Cc: Ingo Molnar
> Cc: Peter Zijlstra
> Cc: Vincent Guittot
On Mon, 26 Nov 2018 at 21:44, Andrea Parri
wrote:
>
> On Mon, Nov 26, 2018 at 12:37:00PM +0100, Vincent Guittot wrote:
> > On Mon, 26 Nov 2018 at 10:30, Peter Zijlstra wrote:
> > >
> > > On Wed, Nov 21, 2018 at 11:34:53PM +0100, Andrea Parri wrote:
> > > &g
On Mon, 26 Nov 2018 at 10:30, Peter Zijlstra wrote:
>
> On Wed, Nov 21, 2018 at 11:34:53PM +0100, Andrea Parri wrote:
> > Hi,
> >
> > The comment for the atomic_fetch_andnot() in nohz_idle_balance() says:
> >
> > "barrier, pairs with nohz_balance_enter_idle(), ensures ..."
> >
> > which, well, d
13ms 32ms 26ms
On my hikey (octo Arm64 platform) with schedutil governor, the time to
reach max OPP when starting from a null utilization, decreases from 223ms
with current scale invariance down to 121ms with the new algorithm.
Signed-off-by: Vincent Guittot
---
incl
Move rq_of() helper function so it can be used in pelt.c
Signed-off-by: Vincent Guittot
---
kernel/sched/fair.c | 13 -
kernel/sched/sched.h | 16
2 files changed, 16 insertions(+), 13 deletions(-)
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index
ts the new scaling algorithm
Changes since v5:
- Fix niptick raised by Dietmar
- Upodated some comments
- Remove some unused variables
- No functional change
Vincent Guittot (2):
sched/fair: move rq_of helper function
sched/fair: update scale invariance of PELT
include/linux/sched
On Mon, 12 Nov 2018 at 18:53, Dietmar Eggemann wrote:
>
> On 11/9/18 8:20 AM, Vincent Guittot wrote:
>
> [...]
>
> > In order to achieve this time scaling, a new clock_pelt is created per rq.
> > The increase of this clock scales with current capacity when somethi
32ms 26ms
On my hikey (octo Arm64 platform) with schedutil governor, the time to reach
max OPP when starting from a null utilization, decreases from 223ms with
current scale invariance down to 121ms with the new algorithm.
Signed-off-by: Vincent Guittot
---
kernel/sched/core
Move rq_of() helper function so it can be used in pelt.c
Signed-off-by: Vincent Guittot
---
kernel/sched/fair.c | 13 -
kernel/sched/sched.h | 16
2 files changed, 16 insertions(+), 13 deletions(-)
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index
ts the new scaling algorithm
Changes since v5:
- Fix running_sum scaling in update_tg_cfs_runnable() raised by Dietmar
- Remove unused cpu parameters raised by Dietmar
Vincent Guittot (2):
sched/fair: move rq_of helper function
sched/fair: update scale invariance of PELT
kernel/sched/core.c
On Thu, 8 Nov 2018 at 12:35, Quentin Perret wrote:
>
> On Wednesday 07 Nov 2018 at 11:47:09 (+0100), Dietmar Eggemann wrote:
> > The important bit for EAS is that it only uses utilization in the
> > non-overutilized case. Here, utilization signals should look the same
> > between the two approache
On Wed, 7 Nov 2018 at 18:02, Quentin Perret wrote:
>
> Hi Vincent,
>
> On Wednesday 07 Nov 2018 at 17:32:32 (+0100), Vincent Guittot wrote:
> > Hi Quentin,
> >
> > On Tue, 16 Oct 2018 at 12:15, Quentin Perret wrote:
> > >
> >
> > > +
>
Hi Quentin,
On Tue, 16 Oct 2018 at 12:15, Quentin Perret wrote:
>
> +
> +/**
> + * em_pd_energy() - Estimates the energy consumed by the CPUs of a perf.
> domain
> + * @pd : performance domain for which energy has to be estimated
> + * @max_util : highest utilization among CPUs of the
On Wed, 7 Nov 2018 at 11:47, Dietmar Eggemann wrote:
>
> On 11/5/18 10:10 AM, Vincent Guittot wrote:
> > On Fri, 2 Nov 2018 at 16:36, Dietmar Eggemann
> > wrote:
> >>
> >> On 10/26/18 6:11 PM, Vincent Guittot wrote:
>
> [...]
>
> >> Thi
On Mon, 5 Nov 2018 at 15:59, Morten Rasmussen wrote:
>
> On Mon, Nov 05, 2018 at 10:10:34AM +0100, Vincent Guittot wrote:
> > On Fri, 2 Nov 2018 at 16:36, Dietmar Eggemann
> > wrote:
> > >
...
> > > >
> > > > In order to achieve this time s
On Fri, 2 Nov 2018 at 16:36, Dietmar Eggemann wrote:
>
> On 10/26/18 6:11 PM, Vincent Guittot wrote:
> > The current implementation of load tracking invariance scales the
> > contribution with current frequency and uarch performance (only for
> > utilization) of the CPU.
On Thu, 1 Nov 2018 at 10:38, Dietmar Eggemann wrote:
>
> On 10/31/18 10:18 AM, Vincent Guittot wrote:
> > Hi Dietmar,
> >
> > On Wed, 31 Oct 2018 at 08:20, Dietmar Eggemann
> > wrote:
> >>
> >> On 10/26/18 6:11 PM, Vincent Guittot
Hi Dietmar,
On Wed, 31 Oct 2018 at 08:20, Dietmar Eggemann wrote:
>
> On 10/26/18 6:11 PM, Vincent Guittot wrote:
>
> [...]
>
> > static int select_idle_sibling(struct task_struct *p, int prev_cpu, int
> > cpu);
> > static unsigned long task_h_load(stru
Hi Pavan,
On Tue, 30 Oct 2018 at 10:19, Pavan Kondeti wrote:
>
> Hi Vincent,
>
> On Fri, Oct 26, 2018 at 06:11:43PM +0200, Vincent Guittot wrote:
> > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> > index 6806c27..7a69673 100644
> > --- a/kernel/sched
Move rq_of() helper function so it can be used in pelt.c
Signed-off-by: Vincent Guittot
---
kernel/sched/fair.c | 13 -
kernel/sched/sched.h | 16
2 files changed, 16 insertions(+), 13 deletions(-)
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index
32ms 26ms
On my hikey (octo ARM platform) with schedutil governor, the time to reach
max OPP when starting from a null utilization, decreases from 223ms with
current scale invariance down to 121ms with the new algorithm.
Signed-off-by: Vincent Guittot
---
kernel/sched/core
ew scaling algorithm
Changes since v4:
- Fix compile issue for !SMP raised by kbuild test robot
- Fix lost_idle_time update raised by Pavan
Vincent Guittot (2):
sched/fair: move rq_of helper function
sched/fair: update scale invariance of PELT
kernel/sched/core.c | 1 +
kernel/sched/
On Thu, 25 Oct 2018 at 13:29, Steven Sistare wrote:
>
> On 10/25/2018 3:50 AM, Vincent Guittot wrote:
> > Hi Steve,
> >
> > On Mon, 22 Oct 2018 at 17:10, Steve Sistare
> > wrote:
> >>
> >> When a CPU has no more CFS tasks to run, and idle_bala
On Thu, 25 Oct 2018 at 12:36, Dietmar Eggemann wrote:
>
> Hi Vincent,
>
> On 10/19/18 6:17 PM, Vincent Guittot wrote:
> > The current implementation of load tracking invariance scales the
> > contribution with current frequency and uarch performance (only for
> > uti
Hi Steve,
On Mon, 22 Oct 2018 at 17:10, Steve Sistare wrote:
>
> When a CPU has no more CFS tasks to run, and idle_balance() fails to
> find a task, then attempt to steal a task from an overloaded CPU in the
> same LLC. Maintain and use a bitmap of overloaded CPUs to efficiently
> identify candid
Hi Pavan,
On Wed, 24 Oct 2018 at 06:53, Pavan Kondeti wrote:
>
> Hi Vincent,
>
> Thanks for the detailed explanation.
>
> On Tue, Oct 23, 2018 at 02:15:08PM +0200, Vincent Guittot wrote:
> > Hi Pavan,
> >
> > On Tue, 23 Oct 2018 at 07:59, Pavan Ko
Hi Pavan,
On Tue, 23 Oct 2018 at 07:59, Pavan Kondeti wrote:
>
> Hi Vincent,
>
> On Fri, Oct 19, 2018 at 06:17:51PM +0200, Vincent Guittot wrote:
> >
> > /*
> > + * The clock_pelt scales the time to reflect the effective amount of
> > + * computation done du
On Tue, 23 Oct 2018 at 12:01, Peter Zijlstra wrote:
>
> On Fri, Oct 19, 2018 at 06:17:51PM +0200, Vincent Guittot wrote:
> > In order to achieve this time scaling, a new clock_pelt is created per rq.
>
>
> > diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
&g
rnor, the time to reach
max OPP when starting from a null utilization, decreases from 223ms with
current scale invariance down to 121ms with the new algorithm. For this
test, I have enable arch_scale_freq for arm64.
Signed-off-by: Vincent Guittot
---
kernel/sched/core.c | 2 +-
kernel
ts the new scaling algorithm
Vincent Guittot (2):
sched/fair: move rq_of helper function
sched/fair: update scale invariance of PELT
kernel/sched/core.c | 2 +-
kernel/sched/deadline.c | 6 ++--
kernel/sched/fair.c | 29 +---
kernel/sched/pelt.
Move rq_of() helper function so it can be used in pelt.c
Signed-off-by: Vincent Guittot
---
kernel/sched/fair.c | 13 -
kernel/sched/sched.h | 16
2 files changed, 16 insertions(+), 13 deletions(-)
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index
Hi Lukasz,
On Thu, 11 Oct 2018 at 13:10, Lukasz Luba wrote:
>
>
>
> On 10/10/2018 07:30 PM, Thara Gopinath wrote:
> > Hello Lukasz,
> >
> > On 10/10/2018 11:35 AM, Lukasz Luba wrote:
> >> Hi Thara,
> >>
> >> I have run it on Exynos5433 mainline.
> >> When it is enabled with step_wise thermal gove
On Wed, 10 Oct 2018 at 15:48, Quentin Perret wrote:
>
> On Wednesday 10 Oct 2018 at 15:27:57 (+0200), Vincent Guittot wrote:
> > On Wed, 10 Oct 2018 at 15:05, Quentin Perret wrote:
> > >
> > > On Wednesday 10 Oct 2018 at 14:04:40 (+0200), Vincent Guittot wrote:
On Wed, 10 Oct 2018 at 15:35, Juri Lelli wrote:
>
> On 10/10/18 15:08, Vincent Guittot wrote:
> > On Wed, 10 Oct 2018 at 14:50, Juri Lelli wrote:
> > >
> > > On 10/10/18 14:34, Vincent Guittot wrote:
> > > > Hi Juri,
> > > >
&
On Wed, 10 Oct 2018 at 15:05, Quentin Perret wrote:
>
> On Wednesday 10 Oct 2018 at 14:04:40 (+0200), Vincent Guittot wrote:
> > This patchset doesn't touch cpu_capacity_orig and doesn't need to as
> > it assume that the max capacity is unchanged but some capacit
On Wed, 10 Oct 2018 at 14:50, Juri Lelli wrote:
>
> On 10/10/18 14:34, Vincent Guittot wrote:
> > Hi Juri,
> >
> > On Wed, 10 Oct 2018 at 14:23, Juri Lelli wrote:
> > >
> > > On 10/10/18 14:04, Vincent Guittot wrote:
> > >
> > > [
Hi Juri,
On Wed, 10 Oct 2018 at 14:23, Juri Lelli wrote:
>
> On 10/10/18 14:04, Vincent Guittot wrote:
>
> [...]
>
> > The problem was the same with RT, the cfs utilization was lower than
> > reality because RT steals soem cycle to CFS
> > So schedutil was sel
On Wed, 10 Oct 2018 at 12:36, Quentin Perret wrote:
>
> On Wednesday 10 Oct 2018 at 12:14:32 (+0200), Vincent Guittot wrote:
> > On Wed, 10 Oct 2018 at 11:55, Quentin Perret wrote:
> > >
> > > Hi Vincent,
> > >
> > > On Wednesday 10 Oct
On Wed, 10 Oct 2018 at 11:55, Quentin Perret wrote:
>
> Hi Vincent,
>
> On Wednesday 10 Oct 2018 at 10:50:05 (+0200), Vincent Guittot wrote:
> > The problem with reflecting directly the capping is that it happens
> > far more often than the pace at which cpu_capacity_or
On Wed, 10 Oct 2018 at 10:29, Quentin Perret wrote:
>
> Hi Thara,
>
> On Wednesday 10 Oct 2018 at 08:17:51 (+0200), Ingo Molnar wrote:
> >
> > * Thara Gopinath wrote:
> >
> > > Thermal governors can respond to an overheat event for a cpu by
> > > capping the cpu's maximum possible frequency. This
Commit-ID: 11d4afd4ff667f9b6178ee8c142c36cb78bd84db
Gitweb: https://git.kernel.org/tip/11d4afd4ff667f9b6178ee8c142c36cb78bd84db
Author: Vincent Guittot
AuthorDate: Tue, 25 Sep 2018 11:17:42 +0200
Committer: Ingo Molnar
CommitDate: Tue, 2 Oct 2018 09:45:00 +0200
sched/pelt: Fix warning
Scaling the utilization of CPUs with irq util_avg in schedutil doesn't give
any benefit and just waste CPU cycles when irq time is not accounted but
only steal time.
Skip the irq scaling when irq time is not accounted
Suggested-by: Wanpeng Li
Signed-off-by: Vincent Guittot
---
kernel/
for asym_packing can be easily delayed for hundreds of ms
because of this all_pinned detection mecanism.
Signed-off-by: Vincent Guittot
---
kernel/sched/fair.c | 27 +++
1 file changed, 19 insertions(+), 8 deletions(-)
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
patches fixes the problems raised for lmbench and the rt-app UC that
creates 2 tasks that start as small tasks and then become suddenly always
running tasks. (I can provide the rt-app json is needed)
- Rebase on latest tip/sched/core
Vincent Guittot (3):
sched/fair: fix rounding issue for asym
enough to skip the rq in find_busiest_queue and prevents asym
migration to happen.
Add 1 to the avg_load to make sure that the targeted cpu will not be
skipped unexpectidly.
Signed-off-by: Vincent Guittot
---
kernel/sched/fair.c | 6 ++
1 file changed, 6 insertions(+)
diff --git a/kernel
newly idle load balance is not always triggered when a cpu becomes idle.
This prevent the scheduler to get a chance to migrate task for asym packing.
Enable active migration because of asym packing during idle load balance too.
Signed-off-by: Vincent Guittot
---
kernel/sched/fair.c | 2 +-
1
On Tue, 25 Sep 2018 at 11:17, Vincent Guittot
wrote:
>
> Create a config for enabling irq load tracking in the scheduler.
> irq load tracking is useful only when irq or paravirtual time is
> accounted but it's only possible with SMP for now.
>
> Also use __maybe_unused to
On Wed, 26 Sep 2018 at 15:17, Peter Zijlstra wrote:
>
> On Wed, Sep 26, 2018 at 12:33:25PM +0200, Vincent Guittot wrote:
> > On Wed, 26 Sep 2018 at 11:35, Valentin Schneider
>
> > > library I use) has some phase where it spawns at lot of tasks at once to
> >
On Wed, 26 Sep 2018 at 15:16, Valentin Schneider
wrote:
>
> On 26/09/18 11:33, Vincent Guittot wrote:
> > On Wed, 26 Sep 2018 at 11:35, Valentin Schneider
> > [...]
> >>> Can you give us details about the use case that you care about ?
> >>>
> &
On Wed, 26 Sep 2018 at 11:35, Valentin Schneider
wrote:
>
> Hi,
>
> On 26/09/18 09:13, Vincent Guittot wrote:
> > On Tue, 25 Sep 2018 at 19:38, Valentin Schneider
> > wrote:
> >>
> >> When load_balance() fails to move some load because of t
On Tue, 25 Sep 2018 at 19:38, Valentin Schneider
wrote:
>
> When load_balance() fails to move some load because of task affinity,
> we end up increasing sd->balance_interval to delay the next periodic
> balance in the hopes that next time we look, that annoying pinned
> task(s) will be gone.
It's
ed by:
commit 2e62c4743adc ("sched/fair: Remove #ifdefs from scale_rt_capacity()")
Reported-by: Dou Liyang
Reported-by: Miguel Ojeda
Suggested-by: Ingo Molnar
Fixes: 2e62c4743adc ("sched/fair: Remove #ifdefs from scale_rt_capacity()")
Signed-off-by: Vincent Guittot
---
On Sat, 15 Sep 2018 at 14:30, Ingo Molnar wrote:
>
>
> * Vincent Guittot wrote:
>
> > Hi Ingo,
> >
> > On Sat, 15 Sep 2018 at 13:47, Ingo Molnar wrote:
> > >
> > >
> > > * tip-bot for Vincent Guittot wrote:
> > >
> >
801 - 900 of 1975 matches
Mail list logo
