From: Valentin Schneider <valentin.schneider@arm.com>
To: Peter Zijlstra <peterz@infradead.org>
Cc: Vincent Guittot <vincent.guittot@linaro.org>,
linux-kernel <linux-kernel@vger.kernel.org>,
Ingo Molnar <mingo@kernel.org>,
Dietmar Eggemann <dietmar.eggemann@arm.com>,
Ingo Molnar <mingo@redhat.com>,
Juri Lelli <juri.lelli@redhat.com>,
Steven Rostedt <rostedt@goodmis.org>
Subject: Re: [PATCH v3 6/9] sched: Kill select_task_rq()'s sd_flag parameter
Date: Thu, 16 Apr 2020 12:43:54 +0100 [thread overview]
Message-ID: <jhjwo6f7j0l.mognet@arm.com> (raw)
In-Reply-To: <20200416104725.GM20730@hirez.programming.kicks-ass.net>
On 16/04/20 11:47, Peter Zijlstra wrote:
> On Thu, Apr 16, 2020 at 09:42:36AM +0200, Vincent Guittot wrote:
>> On Wed, 15 Apr 2020 at 23:05, Valentin Schneider
>> > @@ -6622,13 +6622,25 @@ static int find_energy_efficient_cpu(struct task_struct *p, int prev_cpu)
>> > * preempt must be disabled.
>> > */
>> > static int
>> > +select_task_rq_fair(struct task_struct *p, int prev_cpu, int wake_flags)
>> > {
>> > + int sync = (wake_flags & WF_SYNC) && !(current->flags & PF_EXITING);
>> > struct sched_domain *tmp, *sd = NULL;
>> > int cpu = smp_processor_id();
>> > int new_cpu = prev_cpu;
>> > int want_affine = 0;
>> > - int sync = (wake_flags & WF_SYNC) && !(current->flags & PF_EXITING);
>> > + int sd_flag;
>> > +
>> > + switch (wake_flags & (WF_TTWU | WF_FORK | WF_EXEC)) {
>>
>> You remove a function parameter, which was directly set with the right
>> flag, but then you add a switch case to recreate this sd_flag
>> internally. Not sure we can say that it's real benefit
>>
>> > + case WF_TTWU:
>> > + sd_flag = SD_BALANCE_WAKE;
>> > + break;
>> > + case WF_FORK:
>> > + sd_flag = SD_BALANCE_FORK;
>> > + break;
>> > + default:
>> > + sd_flag = SD_BALANCE_EXEC;
>> > + }
>
> Agreed, that's a bit yuck, how about something like so instead:
>
Aligning the SD_* and WF_* flags sure makes it simpler, I just wasn't
daring enough to do that. I suppose we'll want some BUILD_BUG_ON() checks
just for good measure.
Also, it doesn't directly solve the switch case of 9/9, but I may get out
of it with some hackery such as what I suggested in my reply to Vincent.
>
> --- a/include/linux/sched/topology.h
> +++ b/include/linux/sched/topology.h
> @@ -11,10 +11,12 @@
> */
> #ifdef CONFIG_SMP
>
> +/* First nibble of SD_flag is shared with WF_flag */
> #define SD_BALANCE_NEWIDLE 0x0001 /* Balance when about to become idle */
> #define SD_BALANCE_EXEC 0x0002 /* Balance on exec */
> #define SD_BALANCE_FORK 0x0004 /* Balance on fork, clone */
> #define SD_BALANCE_WAKE 0x0008 /* Balance on wakeup */
> +
> #define SD_WAKE_AFFINE 0x0010 /* Wake task to waking CPU */
> #define SD_ASYM_CPUCAPACITY 0x0020 /* Domain members have different CPU capacities */
> #define SD_SHARE_CPUCAPACITY 0x0040 /* Domain members share CPU capacity */
> --- a/kernel/sched/fair.c
> +++ b/kernel/sched/fair.c
> @@ -6635,16 +6635,8 @@ select_task_rq_fair(struct task_struct *
> int want_affine = 0;
> int sd_flag;
>
> - switch (wake_flags & (WF_TTWU | WF_FORK | WF_EXEC)) {
> - case WF_TTWU:
> - sd_flag = SD_BALANCE_WAKE;
> - break;
> - case WF_FORK:
> - sd_flag = SD_BALANCE_FORK;
> - break;
> - default:
> - sd_flag = SD_BALANCE_EXEC;
> - }
> + /* SD_flags and WF_flags share the first nibble */
> + sd_flag = wake_flags & 0xf;
>
> if (sd_flag & SD_BALANCE_WAKE) {
> record_wakee(p);
> --- a/kernel/sched/sched.h
> +++ b/kernel/sched/sched.h
> @@ -1685,11 +1685,12 @@ static inline int task_on_rq_migrating(s
> /*
> * Wake flags
> */
> -#define WF_SYNC 0x01 /* Waker goes to sleep after wakeup */
> -#define WF_TTWU 0x02 /* Regular task wakeup */
> -#define WF_FORK 0x04 /* Child wakeup after fork */
> -#define WF_EXEC 0x08 /* "Fake" wakeup at exec */
> -#define WF_MIGRATED 0x10 /* Internal use, task got migrated */
> +#define WF_EXEC 0x02 /* SD_BALANCE_EXEC */
> +#define WF_FORK 0x04 /* SD_BALANCE_FORK */
> +#define WF_TTWU 0x08 /* SD_BALANCE_WAKE */
> +
> +#define WF_SYNC 0x10 /* Waker goes to sleep after wakeup */
> +#define WF_MIGRATED 0x20 /* Internal use, task got migrated */
>
> /*
> * To aid in avoiding the subversion of "niceness" due to uneven distribution
next prev parent reply other threads:[~2020-04-16 11:45 UTC|newest]
Thread overview: 30+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-04-15 21:05 [PATCH v3 0/9] sched: Streamline select_task_rq() & select_task_rq_fair() Valentin Schneider
2020-04-15 21:05 ` [PATCH v3 1/9] sched/fair: find_idlest_group(): Remove unused sd_flag parameter Valentin Schneider
2020-04-16 7:40 ` Vincent Guittot
2020-05-01 18:22 ` [tip: sched/core] " tip-bot2 for Valentin Schneider
2020-04-15 21:05 ` [PATCH v3 2/9] sched/debug: Make sd->flags sysctl read-only Valentin Schneider
2020-05-01 18:22 ` [tip: sched/core] " tip-bot2 for Valentin Schneider
2020-04-15 21:05 ` [PATCH v3 3/9] sched: Remove checks against SD_LOAD_BALANCE Valentin Schneider
2020-05-01 18:22 ` [tip: sched/core] " tip-bot2 for Valentin Schneider
2020-04-15 21:05 ` [PATCH v3 4/9] sched/topology: Kill SD_LOAD_BALANCE Valentin Schneider
2020-05-01 18:22 ` [tip: sched/core] " tip-bot2 for Valentin Schneider
2020-04-15 21:05 ` [PATCH v3 5/9] sched: Add WF_TTWU, WF_EXEC wakeup flags Valentin Schneider
2020-04-15 21:05 ` [PATCH v3 6/9] sched: Kill select_task_rq()'s sd_flag parameter Valentin Schneider
2020-04-16 7:42 ` Vincent Guittot
2020-04-16 10:24 ` Valentin Schneider
2020-04-16 10:47 ` Peter Zijlstra
2020-04-16 11:43 ` Valentin Schneider [this message]
2020-04-15 21:05 ` [PATCH v3 7/9] sched/fair: Dissociate wakeup decisions from SD flag value Valentin Schneider
2020-04-15 21:05 ` [PATCH v3 8/9] sched/fair: Split select_task_rq_fair want_affine logic Valentin Schneider
2020-04-15 21:05 ` [PATCH v3 9/9] sched/topology: Define and use shortcut pointers for wakeup sd_flag scan Valentin Schneider
2020-04-16 7:46 ` Vincent Guittot
2020-04-16 10:24 ` Valentin Schneider
2020-04-16 13:04 ` Dietmar Eggemann
2020-04-16 13:36 ` Vincent Guittot
2020-04-16 15:27 ` Valentin Schneider
2020-04-16 15:58 ` Vincent Guittot
2020-04-16 20:54 ` Valentin Schneider
2020-04-16 10:58 ` [PATCH v3 0/9] sched: Streamline select_task_rq() & select_task_rq_fair() Peter Zijlstra
2020-04-16 11:00 ` Peter Zijlstra
2020-04-16 11:02 ` Valentin Schneider
2020-04-16 13:00 ` Vincent Guittot
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=jhjwo6f7j0l.mognet@arm.com \
--to=valentin.schneider@arm.com \
--cc=dietmar.eggemann@arm.com \
--cc=juri.lelli@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@kernel.org \
--cc=mingo@redhat.com \
--cc=peterz@infradead.org \
--cc=rostedt@goodmis.org \
--cc=vincent.guittot@linaro.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
Powered by JetHome