Re: Another SCHED_DEADLINE bug (with bisection and possible fix)

2015-02-02 Thread Juri Lelli
On 31/01/2015 09:56, Peter Zijlstra wrote: On Fri, Jan 30, 2015 at 10:35:02AM +, Juri Lelli wrote: So, we do the safe thing only in case of throttling. No, even for the !throttle aka running tasks. We only use dl_{runtime,deadline,period} for replenishment, until that time we observe

Re: [ANNOUNCE] 3.14.3-rt5

2015-02-16 Thread Juri Lelli
Hi Sebastian, On 16/02/15 11:29, Sebastian Andrzej Siewior wrote: * Juri Lelli | 2014-05-13 15:30:20 [+0200]: Hi, Hi Juri, Also SCHED_DEADLINE dies without the following. Thanks, - Juri ---From 3ca5943538c728399037823e5632431bc2da707c Mon Sep 17 00:00:00 2001 From: Juri Lelli

Re: [sched/deadline] kernel BUG at kernel/sched/deadline.c:805!

2015-02-16 Thread Juri Lelli
Hi, On 16/02/15 14:44, Peter Zijlstra wrote: On Mon, Feb 16, 2015 at 02:08:21PM +0100, Peter Zijlstra wrote: On Mon, Feb 16, 2015 at 03:38:34PM +0300, Kirill Tkhai wrote: We shouldn't enqueue migrating tasks. Please, try this one instead ;) Ha, we should amend that task-rq-lock loop for

Re: Another SCHED_DEADLINE bug (with bisection and possible fix)

2015-01-13 Thread Juri Lelli
Hi all, really sorry for the huge delay in replying to this! :( On 07/01/2015 12:29, Kirill Tkhai wrote: On Ср, 2015-01-07 at 08:01 +0100, Luca Abeni wrote: Hi Kirill, On Tue, 06 Jan 2015 02:07:21 +0300 Kirill Tkhai tk...@yandex.ru wrote: On Пн, 2015-01-05 at 16:21 +0100, Luca Abeni

Re: [PATCH RESEND] sched/deadline: don't need to check throttled status when switched to dl

2015-03-17 Thread Juri Lelli
Hi Ingo, On 17/03/15 08:08, Ingo Molnar wrote: * Wanpeng Li wanpeng...@linux.intel.com wrote: Ping Ingo, ;-) So the changelog is not very readable to me, it has various spelling mistakes that make it hard to read. Seems like Juri acked the patch, mind asking him to fix the changelog

Re: [PATCH v10] sched/deadline: support dl task migration during cpu hotplug

2015-03-06 Thread Juri Lelli
, and the push/pull mechanism should then move it around properly. Suggested-by: Juri Lelli juri.le...@arm.com Signed-off-by: Wanpeng Li wanpeng...@linux.intel.com I think we finally got it :). Thanks! Acked-by: Juri Lelli juri.le...@arm.com --- v9 - v10: * fix the WARNING: line over 80

Re: [PATCH RESEND v10] sched/deadline: support dl task migration during cpu hotplug

2015-03-24 Thread Juri Lelli
are destructive w.r.t. cpusets. Peter, how about we move the check you put in sched_cpu_inactive() to cpuset_cpu_inactive()? This way, if we fail, we don't need to destroy/rebuild the domains. Thanks, - Juri From 65e8033e05f8b70116747062d00d5a5c266699fb Mon Sep 17 00:00:00 2001 From: Juri Lelli

Re: [PATCH RESEND v10] sched/deadline: support dl task migration during cpu hotplug

2015-03-24 Thread Juri Lelli
On 24/03/15 09:13, Wanpeng Li wrote: Hi Juri, On Tue, Mar 24, 2015 at 09:27:09AM +, Juri Lelli wrote: Hi, On 23/03/2015 08:55, Peter Zijlstra wrote: On Mon, Mar 23, 2015 at 08:25:04AM +0100, Ingo Molnar wrote: + if (cpu = nr_cpu_ids

Re: [RFCv3 PATCH 33/48] sched: Energy-aware wake-up task placement

2015-03-25 Thread Juri Lelli
Hi Peter, On 24/03/15 16:35, Peter Zijlstra wrote: On Wed, Feb 04, 2015 at 06:31:10PM +, Morten Rasmussen wrote: +static int energy_aware_wake_cpu(struct task_struct *p) +{ +struct sched_domain *sd; +struct sched_group *sg, *sg_target; +int target_max_cap =

Re: [RFCv3 PATCH 33/48] sched: Energy-aware wake-up task placement

2015-03-26 Thread Juri Lelli
On 25/03/15 18:14, Peter Zijlstra wrote: On Wed, Mar 25, 2015 at 06:01:22PM +, Juri Lelli wrote: Yes and no, IMHO. It makes perfect sense to trigger cpufreq on the target_cpu's freq domain, as we know that we are going to add p's utilization there. Fair point; I mainly wanted to start

Re: [PATCH v13] sched/deadline: support dl task migration during cpu hotplug

2015-03-26 Thread Juri Lelli
, and the push/pull mechanism should then move it around properly. Suggested-and-acked-by: Juri Lelli juri.le...@arm.com Signed-off-by: Wanpeng Li wanpeng...@linux.intel.com --- v12 - 13: * move hotplug stuff to CONFIG_SMP in order to fix the error reported by kbuild test robot v11 - v12

[PATCH] sched/core: check for available -dl bandwidth in cpuset_cpu_inactive

2015-03-31 Thread Juri Lelli
to move the check we currently have in sched_cpu_inactive() to cpuset_cpu_inactive() to prevent useless cpusets reconfiguration in the CPU_DOWN_FAILED path. Signed-off-by: Juri Lelli juri.le...@arm.com Cc: Ingo Molnar mi...@redhat.com Cc: Peter Zijlstra pet...@infradead.org Cc: Juri Lelli juri.le

Re: [PATCH RESEND v10] sched/deadline: support dl task migration during cpu hotplug

2015-03-31 Thread Juri Lelli
Hi Peter, On 30/03/15 10:12, Peter Zijlstra wrote: On Tue, Mar 24, 2015 at 09:27:09AM +, Juri Lelli wrote: Right, it should not happen. It happens because hotplug operations are destructive w.r.t. cpusets. Peter, how about we move the check you put in sched_cpu_inactive

[PATCH] sched/deadline: always enqueue on previous rq when dl_task_timer fires

2015-03-31 Thread Juri Lelli
in dl_task_timer(), but we need to lock the task's previous one. Signed-off-by: Juri Lelli juri.le...@arm.com Tested-by: Wanpeng Li wanpeng...@linux.intel.com Acked-by: Kirill Tkhai ktk...@parallels.com Cc: Ingo Molnar mi...@redhat.com Cc: Peter Zijlstra pet...@infradead.org Cc: Kirill Tkhai ktk

Re: [PATCH v8] sched/deadline: support dl task migration during cpu hotplug

2015-03-02 Thread Juri Lelli
Hi, On 25/02/2015 11:50, Wanpeng Li wrote: I observe that dl task can't be migrated to other cpus during cpu hotplug, in addition, task may/may not be running again if cpu is added back. The root cause which I found is that dl task will be throtted and removed from dl rq after comsuming all

Re: [PATCH v9] sched/deadline: support dl task migration during cpu hotplug

2015-03-04 Thread Juri Lelli
Hi, I think we are still missing a corner case: no admission control, a task with an affinity mask of a single cpu, the cpu goes off. In this case we could try to let it run just somewhere else, as we don't guarantee anything from start. This applies on top of your patch, comments? Thanks, -

Re: [PATCH] sched/deadline: don't need to check throttled status when switched to dl

2015-03-05 Thread Juri Lelli
is cleared each time switch from dl, so throttled status always unset when switch back, there is no need to check throttled status, this patch drop the check. Signed-off-by: Wanpeng Li wanpeng...@linux.intel.com Acked-by: Juri Lelli juri.le...@arm.com --- kernel/sched/deadline.c | 8 1

Re: [PATCH RESEND] sched/deadline: fix pull if dl task who's prio changed is not on queue

2015-03-03 Thread Juri Lelli
is not on queue. So, this is something that was already raised by Kirill, but I always forgot to fix :/. Thanks for reminding me! I have this fix: From 8fcb04eee2d76042970e9561d253d1bc1fe4cc2b Mon Sep 17 00:00:00 2001 From: Juri Lelli juri.le...@arm.com Date: Tue, 3 Mar 2015 07:51:56 +

[PATCH 2/2] sched/deadline: always enqueue on previous rq when dl_task_timer fires

2015-02-24 Thread Juri Lelli
in dl_task_timer(), but we need to lock the task's previous one. Signed-off-by: Juri Lelli juri.le...@arm.com Cc: Ingo Molnar mi...@redhat.com Cc: Peter Zijlstra pet...@infradead.org Cc: Kirill Tkhai ktk...@parallels.com Cc: Juri Lelli juri.le...@gmail.com Cc: linux-kernel@vger.kernel.org Fixes: 3960c8c0c789

[PATCH 1/2] sched/deadline,core: fix bandwidth update when changing cpuset cpumask

2015-02-24 Thread Juri Lelli
methods to save and restore such bandwidth across cpuset reconfiguration. Signed-off-by: Juri Lelli juri.le...@arm.com Cc: Ingo Molnar mi...@redhat.com Cc: Peter Zijlstra pet...@infradead.org Cc: Juri Lelli juri.le...@gmail.com Cc: linux-kernel@vger.kernel.org --- kernel/sched/core.c | 3

Re: [PATCH RESEND v2] sched/deadline: fix rt runtime corrupt when dl refuse a smaller bandwidth

2015-02-26 Thread Juri Lelli
. Looks good, thanks! Acked-by: Juri Lelli juri.le...@arm.com Cc: Juri Lelli juri.le...@arm.com Signed-off-by: Wanpeng Li wanpeng...@linux.intel.com --- v1 - v2: * move sched_dl_global_constraints before sched_rt_global_constraints, and change the name of the former

Re: Another SCHED_DEADLINE bug (with bisection and possible fix)

2015-01-30 Thread Juri Lelli
Hi Peter, On 28/01/2015 14:08, Peter Zijlstra wrote: On Thu, Jan 15, 2015 at 02:35:46PM +0100, Luca Abeni wrote: From what I understand we should either modify the tasks run/sleep stats when we change its parameters or we should schedule a delayed release of the bandwidth delta (when it

Re: [PATCH 4/7] sched/deadline: reschedule if stop task slip in after pull operations

2015-04-21 Thread Juri Lelli
On 21/04/15 00:02, Wanpeng Li wrote: On Tue, Apr 21, 2015 at 06:59:02AM +0800, Wanpeng Li wrote: Hi Juri, On Mon, Apr 20, 2015 at 11:27:22AM +0100, Juri Lelli wrote: Hi, On 06/04/2015 09:53, Wanpeng Li wrote: pull_dl_task can drop (and re-acquire) rq-lock, this means a stop task can slip

Re: [PATCH 4/7] sched/deadline: reschedule if stop task slip in after pull operations

2015-04-20 Thread Juri Lelli
Hi, On 06/04/2015 09:53, Wanpeng Li wrote: pull_dl_task can drop (and re-acquire) rq-lock, this means a stop task can slip in, in which case we need to reschedule. This patch add the reschedule when the scenario occurs. Ok, I guess it can happen. Doesn't RT have the same problem? It seems

Re: [PATCH 6/7] sched/deadline: depend on clearing throttled status in replenish_dl_entity

2015-04-21 Thread Juri Lelli
Hi, On 06/04/2015 09:53, Wanpeng Li wrote: Since the natural place to clear -dl_throttled is in replenish_dl_entity(), and the task which is adjusted the priority is the current, it will be dequeued and then enqueued w/ replenish which can guarantee -dl_throttled can be cleared, this

Re: [PATCH 4/4] sched: cpufreq_cfs: pelt-based cpu frequency scaling

2015-05-05 Thread Juri Lelli
Hi Peter, thanks a lot for the fast reply! :) On 05/05/15 10:00, Peter Zijlstra wrote: On Mon, May 04, 2015 at 03:10:41PM -0700, Michael Turquette wrote: This policy is implemented using the cpufreq governor interface for two main reasons: 1) re-using the cpufreq machine drivers without

Re: [PATCH] sched/deadline: fix dl bandwidth of root domain overflow after dl task dead

2015-05-06 Thread Juri Lelli
Hi Wanpeng, I finally got to review this, sorry about the huge delay. On 07/04/2015 04:36, Wanpeng Li wrote: The total used dl bandwidth of each root domain will be reset to 0 after cpu hotplug when rebuild sched domains, since the call path is: _cpu_down cpuset_cpu_inactive()

Re: [PATCH 4/4] sched: cpufreq_cfs: pelt-based cpu frequency scaling

2015-05-07 Thread Juri Lelli
Hi Mike, On 07/05/15 05:17, Michael Turquette wrote: Quoting Peter Zijlstra (2015-05-06 05:22:40) On Tue, May 05, 2015 at 11:23:47AM -0700, Michael Turquette wrote: Quoting Peter Zijlstra (2015-05-05 02:00:42) On Mon, May 04, 2015 at 03:10:41PM -0700, Michael Turquette wrote: This policy is

Re: On resume from suspend only CPU 0 comes back on-line [REGRESSION][BISECTED]

2015-05-15 Thread Juri Lelli
On 14/05/15 23:12, Doug Smythies wrote: On 2015.05.14 10:48 Ingo Molnar wrote: * Juri Lelli juri.le...@arm.com wrote: On 14/05/15 15:41, Doug Smythies wrote: As of, or about, Kernel 4.1RC1 on resume from suspend only CPU 0 comes back on-line. The issue persists through Kernel 4.1RC3

Re: On resume from suspend only CPU 0 comes back on-line [REGRESSION][BISECTED]

2015-05-14 Thread Juri Lelli
3c18d447b3b36a8d3c90dc37dfbd363cdb685d0a Author: Juri Lelli juri.le...@arm.com Date: Tue Mar 31 09:53:37 2015 +0100 sched/core: Check for available DL bandwidth in cpuset_cpu_inactive() Hotplug operations are destructive w.r.t. cpusets. In case such an operation is performed

Re: [PATCH RFC v2 4/4] sched: cpufreq_cfs: pelt-based cpu frequency scaling

2015-05-18 Thread Juri Lelli
in the cpufreq core would also allow for peripheral devices to place constraint on cpu frequency without having to hack such behavior in at the governor level. Thanks to Juri Lelli juri.le...@arm.com for contributing design ideas, code and test results. [0] http://article.gmane.org

Re: [RFC][PATCH 3/4] sched: Change sched_class::set_cpus_allowed calling context

2015-05-18 Thread Juri Lelli
Hi Peter, On 05/18/2015 09:32 AM, Peter Zijlstra wrote: On Mon, May 18, 2015 at 03:37:43PM +0800, pang.xun...@zte.com.cn wrote: Hi Peter, With this modification, I think the pushing action in my previous patch Check to push the task away after its affinity was changed will not be able to

Re: [PATCH 2/7] sched/deadline: make init_sched_dl_class() __init

2015-04-07 Thread Juri Lelli
Hi, On 06/04/2015 09:53, Wanpeng Li wrote: It's a bootstrap function, make init_sched_dl_class() __init. Looks good, thanks! Best, - Juri Signed-off-by: Wanpeng Li wanpeng...@linux.intel.com --- kernel/sched/deadline.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff

Re: [PATCH] sched/core: Drop debugging leftover trace_printk call

2015-04-07 Thread Juri Lelli
00:00:00 2001 From: Juri Lelli juri.le...@arm.com Date: Sun, 5 Apr 2015 09:57:04 +0100 Subject: [PATCH] scripts/checkpatch: check for uses of trace_printk Production kernels will scream if trace_printk() is used (thanks to Rostedt's banner). Rather than waiting for that to happen, let's check

Re: [PATCH 3/7] sched/deadline: reduce rq lock contention by eliminating locking of non-feasible target

2015-04-07 Thread Juri Lelli
Hi, On 06/04/2015 09:53, Wanpeng Li wrote: This patch adds check that prevents futile attempts to move dl tasks to a CPU with active tasks of equal or earlier deadline. The same behavior as commit 80e3d87b2c55 (sched/rt: Reduce rq lock contention by eliminating locking of non-feasible target)

Re: [PATCH] sched/core: Drop debugging leftover trace_printk call

2015-04-07 Thread Juri Lelli
On 07/04/15 14:56, Steven Rostedt wrote: On Tue, 07 Apr 2015 14:47:50 +0100 Juri Lelli juri.le...@arm.com wrote: On 03/04/2015 09:42, Borislav Petkov wrote: From: Borislav Petkov b...@suse.de Commit 3c18d447b3b3 (sched/core: Check for available DL bandwidth in cpuset_cpu_inactive

Re: [PATCH 5/7] sched/deadline: drop duplicate init_sched_dl_class declaration

2015-04-08 Thread Juri Lelli
Hi, On 06/04/2015 09:53, Wanpeng Li wrote: There are two init_sched_dl_class declarations, this patch drop the duplicated one. I guess the changelog needs to be trimmed. Apart from this, the patch looks of course good, thanks! Best, - Juri Signed-off-by: Wanpeng Li

[PATCH v2] scripts/checkpatch: check for uses of trace_printk

2015-04-08 Thread Juri Lelli
Production kernels will scream if trace_printk() is used (thanks to Steve's banner). Rather than waiting for that to happen, let's check patches beforehand. Signed-off-by: Juri Lelli juri.le...@arm.com Cc: Andy Whitcroft a...@canonical.com Cc: Joe Perches j...@perches.com Cc: Steven Rostedt rost

Re: [PATCH] Documentation/scheduler/sched-deadline.txt: correct definition of density as C_i/min{D_i,P_i}

2015-04-08 Thread Juri Lelli
Hi Luca, On 03/04/15 11:52, Luca Abeni wrote: Hi, On Fri, 3 Apr 2015 16:18:33 +0800 Zhiqiang Zhang zhangzhiqiang.zh...@huawei.com wrote: From the contex,the definition of the destiny of a task C_i/min{D_i,T_i},where T_i is not referred before, should be substituted by C_i/min{D_i,P_i}.

Re: [RFC 4/4] Documentation/scheduler/sched-deadline.txt: add some references

2015-04-09 Thread Juri Lelli
On 08/04/15 12:59, Luca Abeni wrote: Add a description of the Dhall's effect, some discussion about schedulability tests for global EDF, and references to real-time literature, --- Documentation/scheduler/sched-deadline.txt | 81 1 file changed, 71

Re: [RFCv5 PATCH 41/46] sched/fair: add triggers for OPP change requests

2015-08-12 Thread Juri Lelli
On 11/08/15 17:37, Vincent Guittot wrote: On 11 August 2015 at 17:07, Juri Lelli juri.le...@arm.com wrote: Hi Vincent, On 11/08/15 12:41, Vincent Guittot wrote: On 11 August 2015 at 11:08, Juri Lelli juri.le...@arm.com wrote: On 10/08/15 16:07, Vincent Guittot wrote: On 10 August 2015 at 15

Re: [RFCv5 PATCH 45/46] sched/cpufreq_sched: modify pcpu_capacity handling

2015-08-17 Thread Juri Lelli
Hi Peter, On 16/08/15 21:35, Peter Zijlstra wrote: On Tue, Jul 07, 2015 at 07:24:28PM +0100, Morten Rasmussen wrote: From: Juri Lelli juri.le...@arm.com Use the cpu argument of cpufreq_sched_set_cap() to handle per_cpu writes, as the thing can be called remotely (e.g., from load balacing

Re: [RFCv5 PATCH 40/46] sched/cpufreq_sched: compute freq_new based on capacity_orig_of()

2015-08-17 Thread Juri Lelli
On 16/08/15 21:24, Peter Zijlstra wrote: On Sat, Aug 15, 2015 at 09:03:33PM -0700, Michael Turquette wrote: Quoting Peter Zijlstra (2015-08-15 05:46:38) On Tue, Jul 07, 2015 at 07:24:23PM +0100, Morten Rasmussen wrote: diff --git a/kernel/sched/cpufreq_sched.c b/kernel/sched/cpufreq_sched.c

Re: [RFCv5 PATCH 41/46] sched/fair: add triggers for OPP change requests

2015-08-11 Thread Juri Lelli
On 10/08/15 16:07, Vincent Guittot wrote: On 10 August 2015 at 15:43, Juri Lelli juri.le...@arm.com wrote: Hi Vincent, On 04/08/15 14:41, Vincent Guittot wrote: Hi Juri, On 7 July 2015 at 20:24, Morten Rasmussen morten.rasmus...@arm.com wrote: From: Juri Lelli juri.le...@arm.com Each

Re: [PATCH 1/2] sched/deadline: fix comment in push_dl_tasks

2015-08-11 Thread Juri Lelli
Hi Andrea, On 05/08/15 14:56, Andrea Parri wrote: The comment is misleading; fix it by adapting a comment from push_rt_tasks. Signed-off-by: Andrea Parri parri.and...@gmail.com --- kernel/sched/deadline.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git

Re: [RFCv5 PATCH 38/46] sched: scheduler-driven cpu frequency selection

2015-08-11 Thread Juri Lelli
is intentionally dumb. Note that this governor is event-driven. There is no polling loop to check cpu idle time nor any other method which is unsynchronized with the scheduler. Thanks to Juri Lelli juri.le...@arm.com for contributing design ideas, code and test results. [0] http://article.gmane.org

Re: Question about SCHED_DEADLINE and sched_yield() usage

2015-08-11 Thread Juri Lelli
Hi Michael, On 11/08/15 11:49, Michael Riesch wrote: Hi all, I connected two analog-to-digital converters to a BeagleBoneBlack (with kernel version 3.14.33-ti-r51.2) and tried to use the deadline scheduler to get samples at a constant rate. In my C++/Qt application the ADCs are represented

Re: [PATCH 2/2] sched/deadline: fix comment in enqueue_task_dl

2015-08-11 Thread Juri Lelli
Hi, On 05/08/15 14:56, Andrea Parri wrote: The flag dl_boosted is set by comparing *absolute* deadlines (c.f., rt_mutex_setprio). Signed-off-by: Andrea Parri parri.and...@gmail.com --- kernel/sched/deadline.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git

Re: [RFCv5 PATCH 41/46] sched/fair: add triggers for OPP change requests

2015-08-11 Thread Juri Lelli
Hi Vincent, On 11/08/15 12:41, Vincent Guittot wrote: On 11 August 2015 at 11:08, Juri Lelli juri.le...@arm.com wrote: On 10/08/15 16:07, Vincent Guittot wrote: On 10 August 2015 at 15:43, Juri Lelli juri.le...@arm.com wrote: Hi Vincent, On 04/08/15 14:41, Vincent Guittot wrote: Hi Juri

Re: [RFCv5 PATCH 41/46] sched/fair: add triggers for OPP change requests

2015-08-14 Thread Juri Lelli
Hi vincent, On 13/08/15 13:08, Vincent Guittot wrote: On 12 August 2015 at 17:15, Juri Lelli juri.le...@arm.com wrote: On 11/08/15 17:37, Vincent Guittot wrote: On 11 August 2015 at 17:07, Juri Lelli juri.le...@arm.com wrote: Hi Vincent, On 11/08/15 12:41, Vincent Guittot wrote: On 11

Re: [PATCH 07/10] sched: Migrate sched to use new tick dependency mask model

2015-08-10 Thread Juri Lelli
Hi, On 10/08/15 16:29, Frederic Weisbecker wrote: On Mon, Aug 10, 2015 at 05:11:51PM +0200, Peter Zijlstra wrote: On Mon, Aug 10, 2015 at 04:28:47PM +0200, Peter Zijlstra wrote: On Mon, Aug 10, 2015 at 04:16:58PM +0200, Frederic Weisbecker wrote: I considered many times relying on hrtick btw

Re: Question about SCHED_DEADLINE and sched_yield() usage

2015-08-12 Thread Juri Lelli
Hi, On 12/08/15 10:10, Michael Riesch wrote: Hi Juri, On 08/11/2015 01:55 PM, Juri Lelli wrote: As you are running a 3.14 kernel, you probably missed this fix 5bfd126e80dc sched/deadline: Fix sched_yield() behavior. Can you please check? I stumbled over this commit but somehow managed

Re: [RFCv5 PATCH 41/46] sched/fair: add triggers for OPP change requests

2015-08-10 Thread Juri Lelli
Hi Vincent, On 04/08/15 14:41, Vincent Guittot wrote: Hi Juri, On 7 July 2015 at 20:24, Morten Rasmussen morten.rasmus...@arm.com wrote: From: Juri Lelli juri.le...@arm.com Each time a task is {en,de}queued we might need to adapt the current frequency to the new usage. Add triggers on {en

Re: [PATCH 07/10] sched: Migrate sched to use new tick dependency mask model

2015-08-10 Thread Juri Lelli
Hi, On 04/08/15 08:41, Peter Zijlstra wrote: On Mon, Aug 03, 2015 at 07:30:32PM +0200, Frederic Weisbecker wrote: But you've forgotten about SCHED_DEADLINE, we count those in: rq-dl.dl_nr_running. Indeed. Hmm, there is no preemption between SCHED_DEALINE tasks, right? So I can treat it like

Re: [PATCH] sched/deadline: fix dl bandwidth of root domain overflow after dl task dead

2015-08-10 Thread Juri Lelli
On 06/08/15 09:39, Wanpeng Li wrote: Hi Juri, Hi, 2015-05-06 16:14 GMT+08:00 Juri Lelli juri.le...@arm.com mailto:juri.le...@arm.com: Hi Wanpeng, I finally got to review this, sorry about the huge delay. On 07/04/2015 04:36, Wanpeng Li wrote: The total used dl

Re: [RFC PATCH 0/3] CFS idle injection

2015-11-10 Thread Juri Lelli
Hi, On 9 November 2015 at 14:15, Peter Zijlstra wrote: > On Mon, Nov 09, 2015 at 11:56:51AM +, Punit Agrawal wrote: >> Jacob Pan writes: >> > My take is that RT and throttling will never go well together since they >> > are conflicting in

Re: [RFC PATCH 0/3] CFS idle injection

2015-11-10 Thread Juri Lelli
On 11/10/15, Peter Zijlstra wrote: > On Tue, Nov 10, 2015 at 10:07:35AM +0000, Juri Lelli wrote: > > Do you think that using SCHED_DEADLINE here would be completely > > foolish? I mean, we would have the duty_cycle/period thing for free, it > > would be know to the scheduler

Re: [PATCH] sched/deadline: fix earliest_dl.next is not the pushable candidate

2015-11-09 Thread Juri Lelli
Hi, On 10/22/15, Wanpeng Li wrote: > earliest_dl.next is used to cache the next earliest ready task > which also is pushable in order to be a candidate of pushable > tasks during pull algorithm. If the earliest_dl.next deadline > of the sr_rq is earlier than the earliest_dl.curr deadline of >

Re: [RFCv5 PATCH 38/46] sched: scheduler-driven cpu frequency selection

2015-10-08 Thread Juri Lelli
On 08/10/15 01:14, Steve Muckle wrote: > On 08/25/2015 03:45 AM, Juri Lelli wrote: >> But, it is true that if the above events happened the other way around >> (we trigger an update after load balancing and a new task arrives), we >> may miss the opportunity to jump to max wit

Re: [PATCH] Fix migration of SCHED_DEADLINE tasks

2015-10-16 Thread Juri Lelli
Hi, On 16/10/15 09:03, Luca Abeni wrote: > On 10/15/2015 06:40 PM, Juri Lelli wrote: >> On 15/10/15 12:09, Luca Abeni wrote: >>> Commit 9d5142624256 ("sched/deadline: Reduce rq lock contention by >>> eliminating locking of non-feasible

Re: [PATCH] Fix migration of SCHED_DEADLINE tasks

2015-10-15 Thread Juri Lelli
Hi Luca, On 15/10/15 12:09, Luca Abeni wrote: > Commit 9d5142624256 ("sched/deadline: Reduce rq lock contention by > eliminating locking of non-feasible target") broke select_task_rq_dl() > and find_lock_later_rq(), because it introduced a comparison between > the local task's deadline and

Re: [RFCv5 PATCH 43/46] sched/{fair,cpufreq_sched}: add reset_capacity interface

2015-10-09 Thread Juri Lelli
Hi Steve, On 08/10/15 21:40, Steve Muckle wrote: > Hi Juri, > > On 07/07/2015 11:24 AM, Morten Rasmussen wrote: >> From: Juri Lelli <juri.le...@arm.com> >> >> When a CPU is going idle it is pointless to ask for an OPP update as we >> would wake up another

Re: [RFCv5 PATCH 40/46] sched/cpufreq_sched: compute freq_new based on capacity_orig_of()

2015-07-09 Thread Juri Lelli
Hi Mike, On 08/07/15 16:22, Michael Turquette wrote: Quoting Morten Rasmussen (2015-07-07 11:24:23) From: Juri Lelli juri.le...@arm.com capacity is both cpu and freq scaled with EAS. We thus need to compute freq_new using capacity_orig_of(), so that we properly scale back the thing

Re: [RFCv5 PATCH 45/46] sched/cpufreq_sched: modify pcpu_capacity handling

2015-07-09 Thread Juri Lelli
Hi Mike, On 08/07/15 17:42, Michael Turquette wrote: Quoting Morten Rasmussen (2015-07-07 11:24:28) From: Juri Lelli juri.le...@arm.com Use the cpu argument of cpufreq_sched_set_cap() to handle per_cpu writes, as the thing can be called remotely (e.g., from load balacing code). cc: Ingo

Re: [RFCv5 PATCH 41/46] sched/fair: add triggers for OPP change requests

2015-07-09 Thread Juri Lelli
Hi Mike, On 08/07/15 16:42, Michael Turquette wrote: Hi Juri, Quoting Morten Rasmussen (2015-07-07 11:24:24) From: Juri Lelli juri.le...@arm.com Each time a task is {en,de}queued we might need to adapt the current frequency to the new usage. Add triggers on {en,de}queue_task_fair

Re: [RFCv5 PATCH 39/46] sched/cpufreq_sched: use static key for cpu frequency selection

2015-07-10 Thread Juri Lelli
Hi Mike, On 08/07/15 16:19, Michael Turquette wrote: Quoting Morten Rasmussen (2015-07-07 11:24:22) From: Juri Lelli juri.le...@arm.com Introduce a static key to only affect scheduler hot paths when sched governor is enabled. cc: Ingo Molnar mi...@redhat.com cc: Peter Zijlstra pet

Re: [RFCv5 PATCH 44/46] sched/fair: jump to max OPP when crossing UP threshold

2015-07-10 Thread Juri Lelli
Hi Mike, On 08/07/15 17:47, Michael Turquette wrote: Quoting Morten Rasmussen (2015-07-07 11:24:27) From: Juri Lelli juri.le...@arm.com Since the true utilization of a long running task is not detectable while it is running and might be bigger than the current cpu capacity, create

Re: [RFCv5 PATCH 38/46] sched: scheduler-driven cpu frequency selection

2015-08-25 Thread Juri Lelli
Hi Peter, On 15/08/15 14:05, Peter Zijlstra wrote: On Tue, Jul 07, 2015 at 07:24:21PM +0100, Morten Rasmussen wrote: +void cpufreq_sched_set_cap(int cpu, unsigned long capacity) +{ +unsigned int freq_new, cpu_tmp; +struct cpufreq_policy *policy; +struct gov_data *gd; +

Re: [PATCH] sched/deadline: fix dl bandwidth of root domain overflow after dl task dead

2015-09-01 Thread Juri Lelli
Hi, On 30/08/15 12:25, Wanpeng Li wrote: > On 8/10/15 10:10 PM, Juri Lelli wrote: >> On 06/08/15 09:39, Wanpeng Li wrote: >>> Hi Juri, >>> >> Hi, >> >>> 2015-05-06 16:14 GMT+08:00 Juri Lelli <juri.le...@arm.com >>> <mailto:juri.le...

Re: [RFCv5 PATCH 38/46] sched: scheduler-driven cpu frequency selection

2015-09-04 Thread Juri Lelli
On 15/08/15 13:35, Peter Zijlstra wrote: > On Tue, Jul 07, 2015 at 07:24:21PM +0100, Morten Rasmussen wrote: >> diff --git a/kernel/sched/cpufreq_sched.c b/kernel/sched/cpufreq_sched.c >> new file mode 100644 >> index 000..5020f24 >> --- /dev/null >> +++ b/kernel/sched/cpufreq_sched.c >> @@

[PATCH 1/4] sched/{cpuset,core}: restore complete root_domain status across hotplug

2015-09-02 Thread Juri Lelli
<mi...@redhat.com> Cc: Peter Zijlstra <pet...@infradead.org> Cc: Li Zefan <lize...@huawei.com> Cc: cgro...@vger.kernel.org Reported-by: Wanpeng Li <wanpeng...@linux.intel.com> Signed-off-by: Juri Lelli <juri.le...@arm.com> --- include/linux/sched.h | 2

[PATCH 3/4] locking/rtmutex: fix open coded check in rt_mutex_waiter_less()

2015-09-02 Thread Juri Lelli
rt_mutex_waiter_less() check of tasks deadlines is open coded. Since this is subject to wraparound issues, make it use the correct helper. Cc: Ingo Molnar <mi...@redhat.com> Cc: Peter Zijlstra <pet...@infradead.org> Reported-by: Luca Abeni <luca.ab...@unitn.it> Signed-off-by: J

[PATCH 4/4] sched/rt: make (do_)balance_runtime() return void

2015-09-02 Thread Juri Lelli
The return value of (do_)balance_runtime() is not consumed by anybody. Make them return void. Cc: Ingo Molnar <mi...@redhat.com> Cc: Peter Zijlstra <pet...@infradead.org> Signed-off-by: Juri Lelli <juri.le...@arm.com> --- kernel/sched/rt.c | 22 ++

[PATCH 2/4] sched/deadline: unify dl_time_before usage

2015-09-02 Thread Juri Lelli
Move dl_time_before static definition in include/linux/sched/deadline.h so that it can be used by different parties without being re-defined. Cc: Ingo Molnar <mi...@redhat.com> Cc: Peter Zijlstra <pet...@infradead.org> Reported-by: Luca Abeni <luca.ab...@unitn.it> Signed-

[PATCH 0/4] sched fixes and clean-ups

2015-09-02 Thread Juri Lelli
of admission control decisions (admitted bandwidth). This patch is invasive, but it seems to fix the problem (and I couldn't find cleaner fixes :-/). Please give special attention to it :-). Thanks! Best, - Juri Juri Lelli (4): sched/{cpuset,core}: restore complete root_domain status across

Re: [PATCH 1/4] sched/{cpuset,core}: restore complete root_domain status across hotplug

2015-09-10 Thread Juri Lelli
Hi Peter, On 09/09/15 16:11, Peter Zijlstra wrote: > On Wed, Sep 02, 2015 at 11:01:33AM +0100, Juri Lelli wrote: >> Hotplug operations are destructive w.r.t data associated with cpuset; >> in this case we care about root_domains. SCHED_DEADLINE puts bandwidth >> informati

Re: [RFCv5 PATCH 38/46] sched: scheduler-driven cpu frequency selection

2015-09-14 Thread Juri Lelli
On 04/09/15 14:27, Juri Lelli wrote: > On 15/08/15 13:35, Peter Zijlstra wrote: >> On Tue, Jul 07, 2015 at 07:24:21PM +0100, Morten Rasmussen wrote: >>> diff --git a/kernel/sched/cpufreq_sched.c b/kernel/sched/cpufreq_sched.c >>> new file mode 100644 >>> inde

Re: [RFC 08/14] sched/tune: add detailed documentation

2015-09-16 Thread Juri Lelli
Hi Steve, thanks a lot for this interesting discussion. On 16/09/15 00:55, Steve Muckle wrote: > On 09/15/2015 08:00 AM, Patrick Bellasi wrote: >>> Agreed, though I also think those tunable values might also change for a >>> given set of tasks in different circumstances. >> >> Could you provide

Re: [RFCv5 PATCH 38/46] sched: scheduler-driven cpu frequency selection

2015-09-15 Thread Juri Lelli
On 15/09/15 14:45, Peter Zijlstra wrote: > On Mon, Sep 14, 2015 at 04:57:35PM +0100, Juri Lelli wrote: >> On 04/09/15 14:27, Juri Lelli wrote: >>> So, just to recall what we discussed at LPC (I have Mike's slides >>> at hand :-)). It seems that key points are: >&g

Re: [PATCH 0/4] sched fixes and clean-ups

2015-09-18 Thread Juri Lelli
Hi Peter, On 02/09/15 11:01, Juri Lelli wrote: > Hi all, > > while trying to come up with 01/04, I collected the following set of fixes. > > Patches 02-04/04 are simple refactoring of code and clean-ups. > Patch 01/04 is instead my attempt to fix a problem highlighte

Re: [RFC PATCH 0/8] CPUs capacity information for heterogeneous systems

2015-12-07 Thread Juri Lelli
On 07/12/15 13:18, Russell King - ARM Linux wrote: > On Mon, Dec 07, 2015 at 12:36:54PM +0000, Juri Lelli wrote: > > On 07/12/15 12:11, Russell King - ARM Linux wrote: > > > It looks like the second patch doesn't yet have the backing from the DT > > > people, w

Re: [PATCH v5] sched/deadline: fix earliest_dl.next logic

2015-12-03 Thread Juri Lelli
Hi, On 03/12/15 16:59, Wanpeng Li wrote: > 2015-12-03 16:37 GMT+08:00 Luca Abeni : > > Hi, > > > > On 12/03/2015 03:25 AM, Wanpeng Li wrote: > > [...] > >>> > >>> @@ -202,16 +197,18 @@ static void dequeue_pushable_dl_task(struct rq *rq, > >>> struct task_struct *p) > >>> >

Re: [PATCH v6] sched/deadline: fix earliest_dl.next logic

2015-12-04 Thread Juri Lelli
hable_dl_tasks_leftmost = next_node; > + if (next_node) > + dl_rq->earliest_dl.next = rb_entry(next_node, > + struct task_struct, > pushable_dl_tasks)->dl.deadline; Small nitpick, we are breaking 80 columns here

Re: [RFC PATCH 0/8] CPUs capacity information for heterogeneous systems

2015-12-07 Thread Juri Lelli
On 07/12/15 12:11, Russell King - ARM Linux wrote: > On Mon, Dec 07, 2015 at 12:02:44PM +0000, Juri Lelli wrote: > > On 23/11/15 14:28, Juri Lelli wrote: > > > Hi all, > > > > > > > Hi again everybody, > > > > thanks a lot to Rob

Re: [RFC PATCH 0/8] CPUs capacity information for heterogeneous systems

2015-12-07 Thread Juri Lelli
On 23/11/15 14:28, Juri Lelli wrote: > Hi all, > Hi again everybody, thanks a lot to Rob and Vincent for feedback, but, IMHO, we'd need more discussion happening on this series to figure out how we move forward; so, this is a ping :-). It's a simple information that we need to figure o

Re: [PATCH v5] sched/deadline: fix earliest_dl.next logic

2015-12-02 Thread Juri Lelli
Hi, On 02/12/15 19:47, Wanpeng Li wrote: > earliest_dl.next should cache deadline of the earliest ready task that > is also enqueued in the pushable rbtree, as pull algorithm uses this > information to find candidates for migration: if the earliest_dl.next > deadline of source rq is earlier than

Re: [RFC PATCH 5/8] arm64: parse cpu capacity from DT

2015-12-11 Thread Juri Lelli
Hi, On 10/12/15 14:15, Dietmar Eggemann wrote: > On 23/11/15 14:28, Juri Lelli wrote: > > With the introduction of cpu capacity bindings, CPU capacities can now be > > extracted from DT. Add parsing of such information at boot time. Also, > > store such information using per

Re: [RFCv6 PATCH 07/10] sched/fair: jump to max OPP when crossing UP threshold

2015-12-11 Thread Juri Lelli
he cpu usage plus the capacity margin exceeds the > current capacity. This is also done to try to harm the performance of > a task the least. > > Original fair-class only version authored by Juri Lelli > <juri.le...@arm.com>. > > cc: Ingo Molnar <mi...@redhat.com&g

Re: [RFC PATCH 2/8] Documentation: arm: define DT cpu capacity bindings

2015-12-11 Thread Juri Lelli
Hi, On 10/12/15 14:14, Dietmar Eggemann wrote: > On 01/12/15 11:20, Juri Lelli wrote: > > Hi Vincent, > > > > On 30/11/15 10:59, Vincent Guittot wrote: > >> Hi Juri, > >> > >> On 24 November 2015

Re: [RFC PATCH 3/8] arm: parse cpu capacity from DT

2015-12-11 Thread Juri Lelli
Hi, On 10/12/15 14:14, Dietmar Eggemann wrote: > On 23/11/15 14:28, Juri Lelli wrote: > > With the introduction of cpu capacity bindings, CPU capacities can now be > > extracted from DT. Add parsing of such information at boot time. We keep > > code that can produce sa

Re: [RFCv6 PATCH 03/10] sched: scheduler-driven cpu frequency selection

2015-12-11 Thread Juri Lelli
here is no polling loop to > check cpu idle time nor any other method which is unsynchronized with > the scheduler, aside from a throttling mechanism to ensure frequency > changes are not attempted faster than the hardware can accommodate them. > > Thanks to Juri Lelli <juri.le...@ar

Re: [RFCv6 PATCH 10/10] sched: rt scheduler sets capacity requirement

2015-12-11 Thread Juri Lelli
On 08/12/15 22:19, Steve Muckle wrote: > From: Vincent Guittot > > RT tasks don't provide any running constraints like deadline ones > except their running priority. The only current usable input to > estimate the capacity needed by RT tasks is the rt_avg metric. We

Re: [RFC PATCH 8/8] arm64: add sysfs cpu_capacity attribute

2015-12-10 Thread Juri Lelli
Hi, On 10/12/15 15:59, Mark Brown wrote: > On Thu, Dec 10, 2015 at 02:15:04PM +, Dietmar Eggemann wrote: > > On 23/11/15 14:28, Juri Lelli wrote: > > > > The new attribute shows up as: > > > > /sys/devices/system/cpu/cpu*/cpu_capacity > > >

Re: [RFC PATCH 2/8] Documentation: arm: define DT cpu capacity bindings

2015-12-10 Thread Juri Lelli
Hi Mark, I certainly understand your (and Rob's) concerns, but let me try anyway to argument a bit more around this approach :-). On 10/12/15 15:30, Mark Brown wrote: > On Mon, Nov 23, 2015 at 08:06:31PM -0600, Rob Herring wrote: > > > I think you need something absolute and probably per MHz

Re: [RFC PATCH 2/8] Documentation: arm: define DT cpu capacity bindings

2015-12-17 Thread Juri Lelli
Hi, On 15/12/15 17:45, Mark Brown wrote: > On Tue, Dec 15, 2015 at 05:28:37PM +, Mark Rutland wrote: > > On Tue, Dec 15, 2015 at 05:17:13PM +, Mark Brown wrote: > > > > Obviously people are going to get upset if we introduce performance > > > regressions - but that's true always, we can

Re: [RFC PATCH 2/8] Documentation: arm: define DT cpu capacity bindings

2015-12-14 Thread Juri Lelli
Hi Mark, On 11/12/15 17:49, Mark Brown wrote: > On Thu, Dec 10, 2015 at 05:58:20PM +0000, Juri Lelli wrote: > > On 10/12/15 15:30, Mark Brown wrote: > > > On Mon, Nov 23, 2015 at 08:06:31PM -0600, Rob Herring wrote: > > > > > In other words, I want to see th

Re: [RFCv6 PATCH 09/10] sched: deadline: use deadline bandwidth in scale_rt_capacity

2015-12-14 Thread Juri Lelli
On 14/12/15 16:56, Vincent Guittot wrote: > On 14 December 2015 at 16:17, Peter Zijlstra wrote: > > On Tue, Dec 08, 2015 at 10:19:30PM -0800, Steve Muckle wrote: > >> From: Vincent Guittot > > > >> diff --git a/kernel/sched/deadline.c

Re: [RFCv6 PATCH 03/10] sched: scheduler-driven cpu frequency selection

2015-12-15 Thread Juri Lelli
On 14/12/15 18:02, Steve Muckle wrote: > Hi Juri, > > Thanks for the review. > > On 12/11/2015 03:04 AM, Juri Lelli wrote: > >> +config CPU_FREQ_GOV_SCHED > >> + bool "'sched' cpufreq governor" > >> + depends on CPU_FREQ > > > &

Re: [RFC PATCH 2/8] Documentation: arm: define DT cpu capacity bindings

2015-12-15 Thread Juri Lelli
On 14/12/15 16:59, Mark Brown wrote: > On Mon, Dec 14, 2015 at 12:36:16PM +0000, Juri Lelli wrote: > > On 11/12/15 17:49, Mark Brown wrote: > > > > The purpose of the capacity values is to influence the scheduler > > > behaviour and hence performance. Without

Re: [RFC PATCH 2/8] Documentation: arm: define DT cpu capacity bindings

2015-12-15 Thread Juri Lelli
Hi Mark, On 15/12/15 14:01, Mark Rutland wrote: > On Tue, Dec 15, 2015 at 01:39:51PM +, Mark Brown wrote: > > On Tue, Dec 15, 2015 at 12:22:38PM +0000, Juri Lelli wrote: > > > > So then why isn't it adequate to just have things like the core types in > > > > t

  1   2   3   4   5   6   7   8   9   10   >