* Re: [PATCH] sched: Document that RT task priorities are 1…99
2019-04-03 21:08 [PATCH] sched: Document that RT task priorities are 1…99 Sebastian Andrzej Siewior
@ 2019-04-03 21:17 ` John Ogness
2019-04-04 7:45 ` John Ogness
2019-04-16 8:13 ` [tip:sched/core] sched/core: Document that RT task priorities are 1...99 tip-bot for Sebastian Andrzej Siewior
2019-06-17 12:24 ` [PATCH] sched: Document that RT task priorities are 1…99 Peter Zijlstra
2 siblings, 1 reply; 10+ messages in thread
From: John Ogness @ 2019-04-03 21:17 UTC (permalink / raw)
To: Sebastian Andrzej Siewior; +Cc: linux-kernel, Ingo Molnar, Peter Zijlstra, tglx
On 2019-04-03, Sebastian Andrzej Siewior <bigeasy@linutronix.de> wrote:
> John identified three files which claim that RT task priorities start at
> zero. As far as I understand, 0 is used for DL and has nothing to do
> wihich RT priorities as identified by the RT policy.
>
> Correct the comment, valid RT priorities are in the range from 1 to 99.
>
> Reported-by: John Ogness <john.ogness@linutronix.de>
> Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
> ---
> Documentation/scheduler/sched-rt-group.txt | 2 +-
> include/linux/sched/prio.h | 2 +-
> kernel/sched/cpupri.h | 2 +-
> 3 files changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/Documentation/scheduler/sched-rt-group.txt b/Documentation/scheduler/sched-rt-group.txt
> index d8fce3e784574..23f8f8465a775 100644
> --- a/Documentation/scheduler/sched-rt-group.txt
> +++ b/Documentation/scheduler/sched-rt-group.txt
> @@ -175,7 +175,7 @@ get their allocated time.
>
> Implementing SCHED_EDF might take a while to complete. Priority Inheritance is
> the biggest challenge as the current linux PI infrastructure is geared towards
> -the limited static priority levels 0-99. With deadline scheduling you need to
> +the limited static priority levels 1-99. With deadline scheduling you need to
> do deadline inheritance (since priority is inversely proportional to the
> deadline delta (deadline - now)).
>
> diff --git a/include/linux/sched/prio.h b/include/linux/sched/prio.h
> index 7d64feafc408e..6986c32356842 100644
> --- a/include/linux/sched/prio.h
> +++ b/include/linux/sched/prio.h
> @@ -8,7 +8,7 @@
>
> /*
> * Priority of a process goes from 0..MAX_PRIO-1, valid RT
> - * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
> + * priority is 1..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
Actually, valid RT priority is 0..MAX_RT_PRIO-2 (0-98). This comment is
talking about the kernel representation, not the userspace one.
> * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
> * values are inverted: lower p->prio value means higher priority.
> *
> diff --git a/kernel/sched/cpupri.h b/kernel/sched/cpupri.h
> index 7dc20a3232e72..40257a97fb8f2 100644
> --- a/kernel/sched/cpupri.h
> +++ b/kernel/sched/cpupri.h
> @@ -5,7 +5,7 @@
> #define CPUPRI_INVALID -1
> #define CPUPRI_IDLE 0
> #define CPUPRI_NORMAL 1
> -/* values 2-101 are RT priorities 0-99 */
> +/* values 2-101 are RT priorities 1-99 */
I suppose this also should be 0-98.
>
> struct cpupri_vec {
> atomic_t count;
IMHO it is a bit crazy that userspace RT prio 99 maps to kernel prio
0. This leaves a hole at kernel prio 99. Wouldn't it be better just to
map userspace RT prio 1-99 to kernel prio 99-1?
John Ogness
^ permalink raw reply [flat|nested] 10+ messages in thread* Re: [PATCH] sched: Document that RT task priorities are 1…99
2019-04-03 21:17 ` John Ogness
@ 2019-04-04 7:45 ` John Ogness
0 siblings, 0 replies; 10+ messages in thread
From: John Ogness @ 2019-04-04 7:45 UTC (permalink / raw)
To: Sebastian Andrzej Siewior; +Cc: linux-kernel, Ingo Molnar, Peter Zijlstra, tglx
On 2019-04-03, John Ogness <john.ogness@linutronix.de> wrote:
> IMHO it is a bit crazy that userspace RT prio 99 maps to kernel prio
> 0. This leaves a hole at kernel prio 99. Wouldn't it be better just to
> map userspace RT prio 1-99 to kernel prio 99-1?
FWIW, the current mapping:
userspace kernel
--------- ------
rt 1...99 98...0
nice -20...+19 100...139
has been in place since the introduction of Ingo's scalable scheduler:
Author: linus1 <torvalds@athlon.transmeta.com>
Date: Sun Dec 16 12:00:00 2001 -0800
v2.5.1.9 -> v2.5.1.10
- Kai Germaschewski: ISDN updates
- Al Viro: start moving buffer cache indexing to "struct
- block_device *"
- Greg KH: USB update
- Russell King: fix up some ARM merge issues
- Ingo Molnar: scalable scheduler
With that commit, the then separate nice and rt_priority task_struct
fields were combined to the new prio field, and thus the mapping was
born.
John Ogness
^ permalink raw reply [flat|nested] 10+ messages in thread
* [tip:sched/core] sched/core: Document that RT task priorities are 1...99
2019-04-03 21:08 [PATCH] sched: Document that RT task priorities are 1…99 Sebastian Andrzej Siewior
2019-04-03 21:17 ` John Ogness
@ 2019-04-16 8:13 ` tip-bot for Sebastian Andrzej Siewior
2019-04-16 9:36 ` Peter Zijlstra
2019-06-17 12:24 ` [PATCH] sched: Document that RT task priorities are 1…99 Peter Zijlstra
2 siblings, 1 reply; 10+ messages in thread
From: tip-bot for Sebastian Andrzej Siewior @ 2019-04-16 8:13 UTC (permalink / raw)
To: linux-tip-commits
Cc: bigeasy, peterz, hpa, mingo, linux-kernel, torvalds, john.ogness, tglx
Commit-ID: 37a8f8590f307db8f8cfa62dfd82b3fc0978607c
Gitweb: https://git.kernel.org/tip/37a8f8590f307db8f8cfa62dfd82b3fc0978607c
Author: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
AuthorDate: Wed, 3 Apr 2019 23:08:21 +0200
Committer: Ingo Molnar <mingo@kernel.org>
CommitDate: Tue, 16 Apr 2019 10:06:08 +0200
sched/core: Document that RT task priorities are 1...99
John identified three files which claim that RT task priorities start at
zero. As far as I understand, 0 is used for DL and has nothing to do
wihich RT priorities as identified by the RT policy.
Correct the comment, valid RT priorities are in the range from 1 to 99.
Reported-by: John Ogness <john.ogness@linutronix.de>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/20190403210821.10916-1-bigeasy@linutronix.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
Documentation/scheduler/sched-rt-group.txt | 2 +-
include/linux/sched/prio.h | 2 +-
kernel/sched/cpupri.h | 2 +-
3 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/Documentation/scheduler/sched-rt-group.txt b/Documentation/scheduler/sched-rt-group.txt
index d8fce3e78457..23f8f8465a77 100644
--- a/Documentation/scheduler/sched-rt-group.txt
+++ b/Documentation/scheduler/sched-rt-group.txt
@@ -175,7 +175,7 @@ get their allocated time.
Implementing SCHED_EDF might take a while to complete. Priority Inheritance is
the biggest challenge as the current linux PI infrastructure is geared towards
-the limited static priority levels 0-99. With deadline scheduling you need to
+the limited static priority levels 1-99. With deadline scheduling you need to
do deadline inheritance (since priority is inversely proportional to the
deadline delta (deadline - now)).
diff --git a/include/linux/sched/prio.h b/include/linux/sched/prio.h
index 7d64feafc408..6986c3235684 100644
--- a/include/linux/sched/prio.h
+++ b/include/linux/sched/prio.h
@@ -8,7 +8,7 @@
/*
* Priority of a process goes from 0..MAX_PRIO-1, valid RT
- * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
+ * priority is 1..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
* tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
* values are inverted: lower p->prio value means higher priority.
*
diff --git a/kernel/sched/cpupri.h b/kernel/sched/cpupri.h
index 7dc20a3232e7..40257a97fb8f 100644
--- a/kernel/sched/cpupri.h
+++ b/kernel/sched/cpupri.h
@@ -5,7 +5,7 @@
#define CPUPRI_INVALID -1
#define CPUPRI_IDLE 0
#define CPUPRI_NORMAL 1
-/* values 2-101 are RT priorities 0-99 */
+/* values 2-101 are RT priorities 1-99 */
struct cpupri_vec {
atomic_t count;
^ permalink raw reply related [flat|nested] 10+ messages in thread* Re: [tip:sched/core] sched/core: Document that RT task priorities are 1...99
2019-04-16 8:13 ` [tip:sched/core] sched/core: Document that RT task priorities are 1...99 tip-bot for Sebastian Andrzej Siewior
@ 2019-04-16 9:36 ` Peter Zijlstra
2019-04-16 9:47 ` Sebastian Andrzej Siewior
0 siblings, 1 reply; 10+ messages in thread
From: Peter Zijlstra @ 2019-04-16 9:36 UTC (permalink / raw)
To: bigeasy, hpa, mingo, linux-kernel, torvalds, tglx, john.ogness
Cc: linux-tip-commits
On Tue, Apr 16, 2019 at 01:13:11AM -0700, tip-bot for Sebastian Andrzej Siewior wrote:
> Commit-ID: 37a8f8590f307db8f8cfa62dfd82b3fc0978607c
> Gitweb: https://git.kernel.org/tip/37a8f8590f307db8f8cfa62dfd82b3fc0978607c
> Author: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
> AuthorDate: Wed, 3 Apr 2019 23:08:21 +0200
> Committer: Ingo Molnar <mingo@kernel.org>
> CommitDate: Tue, 16 Apr 2019 10:06:08 +0200
>
> sched/core: Document that RT task priorities are 1...99
>
> John identified three files which claim that RT task priorities start at
> zero. As far as I understand, 0 is used for DL and has nothing to do
> wihich RT priorities as identified by the RT policy.
DL is -1, 0 is still very much FIFO/FF
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [tip:sched/core] sched/core: Document that RT task priorities are 1...99
2019-04-16 9:36 ` Peter Zijlstra
@ 2019-04-16 9:47 ` Sebastian Andrzej Siewior
2019-04-16 11:22 ` Peter Zijlstra
2019-04-16 12:43 ` John Ogness
0 siblings, 2 replies; 10+ messages in thread
From: Sebastian Andrzej Siewior @ 2019-04-16 9:47 UTC (permalink / raw)
To: Peter Zijlstra
Cc: hpa, mingo, linux-kernel, torvalds, tglx, john.ogness, linux-tip-commits
On 2019-04-16 11:36:34 [+0200], Peter Zijlstra wrote:
> On Tue, Apr 16, 2019 at 01:13:11AM -0700, tip-bot for Sebastian Andrzej Siewior wrote:
> > Commit-ID: 37a8f8590f307db8f8cfa62dfd82b3fc0978607c
> > Gitweb: https://git.kernel.org/tip/37a8f8590f307db8f8cfa62dfd82b3fc0978607c
> > Author: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
> > AuthorDate: Wed, 3 Apr 2019 23:08:21 +0200
> > Committer: Ingo Molnar <mingo@kernel.org>
> > CommitDate: Tue, 16 Apr 2019 10:06:08 +0200
> >
> > sched/core: Document that RT task priorities are 1...99
> >
> > John identified three files which claim that RT task priorities start at
> > zero. As far as I understand, 0 is used for DL and has nothing to do
> > wihich RT priorities as identified by the RT policy.
>
> DL is -1, 0 is still very much FIFO/FF
but it can't be set. The lowest we can set is 1. Is this a bug then?
Sebastian
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [tip:sched/core] sched/core: Document that RT task priorities are 1...99
2019-04-16 9:47 ` Sebastian Andrzej Siewior
@ 2019-04-16 11:22 ` Peter Zijlstra
2019-04-16 12:08 ` Sebastian Andrzej Siewior
2019-04-16 12:43 ` John Ogness
1 sibling, 1 reply; 10+ messages in thread
From: Peter Zijlstra @ 2019-04-16 11:22 UTC (permalink / raw)
To: Sebastian Andrzej Siewior
Cc: hpa, mingo, linux-kernel, torvalds, tglx, john.ogness, linux-tip-commits
On Tue, Apr 16, 2019 at 11:47:27AM +0200, Sebastian Andrzej Siewior wrote:
> On 2019-04-16 11:36:34 [+0200], Peter Zijlstra wrote:
> > On Tue, Apr 16, 2019 at 01:13:11AM -0700, tip-bot for Sebastian Andrzej Siewior wrote:
> > > Commit-ID: 37a8f8590f307db8f8cfa62dfd82b3fc0978607c
> > > Gitweb: https://git.kernel.org/tip/37a8f8590f307db8f8cfa62dfd82b3fc0978607c
> > > Author: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
> > > AuthorDate: Wed, 3 Apr 2019 23:08:21 +0200
> > > Committer: Ingo Molnar <mingo@kernel.org>
> > > CommitDate: Tue, 16 Apr 2019 10:06:08 +0200
> > >
> > > sched/core: Document that RT task priorities are 1...99
> > >
> > > John identified three files which claim that RT task priorities start at
> > > zero. As far as I understand, 0 is used for DL and has nothing to do
> > > wihich RT priorities as identified by the RT policy.
> >
> > DL is -1, 0 is still very much FIFO/FF
>
> but it can't be set. The lowest we can set is 1. Is this a bug then?
What cannot be set?
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [tip:sched/core] sched/core: Document that RT task priorities are 1...99
2019-04-16 11:22 ` Peter Zijlstra
@ 2019-04-16 12:08 ` Sebastian Andrzej Siewior
0 siblings, 0 replies; 10+ messages in thread
From: Sebastian Andrzej Siewior @ 2019-04-16 12:08 UTC (permalink / raw)
To: Peter Zijlstra
Cc: hpa, mingo, linux-kernel, torvalds, tglx, john.ogness, linux-tip-commits
On 2019-04-16 13:22:35 [+0200], Peter Zijlstra wrote:
> On Tue, Apr 16, 2019 at 11:47:27AM +0200, Sebastian Andrzej Siewior wrote:
> > On 2019-04-16 11:36:34 [+0200], Peter Zijlstra wrote:
> > > On Tue, Apr 16, 2019 at 01:13:11AM -0700, tip-bot for Sebastian Andrzej Siewior wrote:
> > > > Commit-ID: 37a8f8590f307db8f8cfa62dfd82b3fc0978607c
> > > > Gitweb: https://git.kernel.org/tip/37a8f8590f307db8f8cfa62dfd82b3fc0978607c
> > > > Author: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
> > > > AuthorDate: Wed, 3 Apr 2019 23:08:21 +0200
> > > > Committer: Ingo Molnar <mingo@kernel.org>
> > > > CommitDate: Tue, 16 Apr 2019 10:06:08 +0200
> > > >
> > > > sched/core: Document that RT task priorities are 1...99
> > > >
> > > > John identified three files which claim that RT task priorities start at
> > > > zero. As far as I understand, 0 is used for DL and has nothing to do
> > > > wihich RT priorities as identified by the RT policy.
> > >
> > > DL is -1, 0 is still very much FIFO/FF
> >
> > but it can't be set. The lowest we can set is 1. Is this a bug then?
>
> What cannot be set?
| {
| struct sched_param sp;
| int r;
|
| sp.sched_priority = 0;
| r = sched_setscheduler_nocheck(current, SCHED_FIFO, &sp);
| pr_err("%s(%d): %d\n", __func__, __LINE__, r);
|
| }
-22 due to
4160 static int __sched_setscheduler(struct task_struct *p,
…
4191 /*
4192 * Valid priorities for SCHED_FIFO and SCHED_RR are
4193 * 1..MAX_USER_RT_PRIO-1, valid priority for SCHED_NORMAL,
4194 * SCHED_BATCH and SCHED_IDLE is 0.
4195 */
4196 if ((p->mm && attr->sched_priority > MAX_USER_RT_PRIO-1) ||
4197 (!p->mm && attr->sched_priority > MAX_RT_PRIO-1))
4198 return -EINVAL;
4199 if ((dl_policy(policy) && !__checkparam_dl(attr)) ||
4200 (rt_policy(policy) != (attr->sched_priority != 0)))
4201 return -EINVAL;
↑
Sebastian
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [tip:sched/core] sched/core: Document that RT task priorities are 1...99
2019-04-16 9:47 ` Sebastian Andrzej Siewior
2019-04-16 11:22 ` Peter Zijlstra
@ 2019-04-16 12:43 ` John Ogness
1 sibling, 0 replies; 10+ messages in thread
From: John Ogness @ 2019-04-16 12:43 UTC (permalink / raw)
To: Sebastian Andrzej Siewior
Cc: Peter Zijlstra, hpa, mingo, linux-kernel, torvalds, tglx,
linux-tip-commits
On 2019-04-16, Sebastian Andrzej Siewior <bigeasy@linutronix.de> wrote:
> On 2019-04-16 11:36:34 [+0200], Peter Zijlstra wrote:
>> On Tue, Apr 16, 2019 at 01:13:11AM -0700, tip-bot for Sebastian Andrzej Siewior wrote:
>> > Commit-ID: 37a8f8590f307db8f8cfa62dfd82b3fc0978607c
>> > Gitweb: https://git.kernel.org/tip/37a8f8590f307db8f8cfa62dfd82b3fc0978607c
>> > Author: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
>> > AuthorDate: Wed, 3 Apr 2019 23:08:21 +0200
>> > Committer: Ingo Molnar <mingo@kernel.org>
>> > CommitDate: Tue, 16 Apr 2019 10:06:08 +0200
>> >
>> > sched/core: Document that RT task priorities are 1...99
>> >
>> > John identified three files which claim that RT task priorities start at
>> > zero. As far as I understand, 0 is used for DL and has nothing to do
>> > wihich RT priorities as identified by the RT policy.
>>
>> DL is -1, 0 is still very much FIFO/FF
>
> but it can't be set. The lowest we can set is 1. Is this a bug then?
I think it is important to clarify if you are talking about userspace or
kernel. They are different. For userspace 1(low)-99(high) is
allowed. For kernel 0(high)-98(low) is used.
John Ogness
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH] sched: Document that RT task priorities are 1…99
2019-04-03 21:08 [PATCH] sched: Document that RT task priorities are 1…99 Sebastian Andrzej Siewior
2019-04-03 21:17 ` John Ogness
2019-04-16 8:13 ` [tip:sched/core] sched/core: Document that RT task priorities are 1...99 tip-bot for Sebastian Andrzej Siewior
@ 2019-06-17 12:24 ` Peter Zijlstra
2 siblings, 0 replies; 10+ messages in thread
From: Peter Zijlstra @ 2019-06-17 12:24 UTC (permalink / raw)
To: Sebastian Andrzej Siewior
Cc: linux-kernel, Ingo Molnar, tglx, John Ogness, richard
On Wed, Apr 03, 2019 at 11:08:21PM +0200, Sebastian Andrzej Siewior wrote:
> John identified three files which claim that RT task priorities start at
> zero. As far as I understand, 0 is used for DL and has nothing to do
> wihich RT priorities as identified by the RT policy.
>
> Correct the comment, valid RT priorities are in the range from 1 to 99.
It all depends on what view I'm afraid. User priorities go from 1-99, as
per sched_get_priority_{min,max}(), but Kernel priority is:
kernel_prio := MAX_RT_PRIO-1 - user_prio
and that then gives [0-98], where 0 is max and 98 is min (see
__sched_setscheduler() and normal_prio()).
And DL uses kernel prio -1 (there is no user prio equivalent).
Nice maps to:
kernel_prio := MAX_RT_PRIO + (MAX_NICE - MIN_NICE + 1) / 2 + user_nice
which is: [100-139].
Which then, as derRichard says, leaves (kernel) 99 unaccounted for.
> Reported-by: John Ogness <john.ogness@linutronix.de>
> Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
> ---
> Documentation/scheduler/sched-rt-group.txt | 2 +-
> include/linux/sched/prio.h | 2 +-
> kernel/sched/cpupri.h | 2 +-
> 3 files changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/Documentation/scheduler/sched-rt-group.txt b/Documentation/scheduler/sched-rt-group.txt
> index d8fce3e784574..23f8f8465a775 100644
> --- a/Documentation/scheduler/sched-rt-group.txt
> +++ b/Documentation/scheduler/sched-rt-group.txt
> @@ -175,7 +175,7 @@ get their allocated time.
>
> Implementing SCHED_EDF might take a while to complete. Priority Inheritance is
> the biggest challenge as the current linux PI infrastructure is geared towards
> -the limited static priority levels 0-99. With deadline scheduling you need to
> +the limited static priority levels 1-99. With deadline scheduling you need to
> do deadline inheritance (since priority is inversely proportional to the
> deadline delta (deadline - now)).
>
This might be correct,.. but I feel we should strive to delete
everything rt groups.
> diff --git a/include/linux/sched/prio.h b/include/linux/sched/prio.h
> index 7d64feafc408e..6986c32356842 100644
> --- a/include/linux/sched/prio.h
> +++ b/include/linux/sched/prio.h
> @@ -8,7 +8,7 @@
>
> /*
> * Priority of a process goes from 0..MAX_PRIO-1, valid RT
> - * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
> + * priority is 1..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
> * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
> * values are inverted: lower p->prio value means higher priority.
> *
So that comment talks about kernel prio, and there, as we've shown, 0 is
an actual valid RR/FIFO priority (in fact, the highest).
> diff --git a/kernel/sched/cpupri.h b/kernel/sched/cpupri.h
> index 7dc20a3232e72..40257a97fb8f2 100644
> --- a/kernel/sched/cpupri.h
> +++ b/kernel/sched/cpupri.h
> @@ -5,7 +5,7 @@
> #define CPUPRI_INVALID -1
> #define CPUPRI_IDLE 0
> #define CPUPRI_NORMAL 1
> -/* values 2-101 are RT priorities 0-99 */
> +/* values 2-101 are RT priorities 1-99 */
Again, this is kernel prios, not user prios.
> struct cpupri_vec {
> atomic_t count;
> --
> 2.20.1
>
^ permalink raw reply [flat|nested] 10+ messages in thread