All of lore.kernel.org
 help / color / mirror / Atom feed
From: Wolfgang Grandegger <wg@domain.hid>
To: roderik.wildenburg@domain.hid
Cc: xenomai@xenomai.org
Subject: Re: [Xenomai-help] Thread policy and priority in secondary mode
Date: Fri, 25 Jul 2008 10:09:07 +0200	[thread overview]
Message-ID: <48898A23.9030308@domain.hid> (raw)
In-Reply-To: <5D63919D95F87E4D9D34FF7748CE2C2A0137F8FD@ARVMAIL1.mra.roland-man.biz>

roderik.wildenburg@domain.hid wrote:
>>>> __real_pthread_setschedparam(). Note: this will propagate the 
>>>> priority to the underlying Xenomai thread as well.
>>>>
>>> Does this mean that the Xenomai thread will run with the new (Linux)
>>> priority when it returns to primary mode ?
>> Yes. However, the propagation only occurs when the skin and Linux have
>> compatible priority scales, which is the case with the POSIX skin.
>>
>>> And the other way round :
>>> when a posix thread switches to secondary mode (without using
>>> __real_pthread_setschedparam), does it run in Linux with 
>> the priority
>>> set in Xenomai (with pthread_setschedparam)?
>> Yes.
>>
>>> Or better : what is the Linux-prority when a Xenomai thread 
>> (posix or or
>>> any other Xenomai thread)switches to secondaary mode ?
>>>
>> It depends on the skin. When the priority scales are 
>> compatible (i.e. a given
>> priority level exists in both and scales are both upwarding, 
>> like POSIX skin <->
>> Linux, and native skin <-> Linux), you get the same priority 
>> level in both
>> modes, and changing the priority from one mode affects the 
>> other mode as well.
>>
>> When they are not compatible, the skin-based priority is 
>> "normalized" and
>> applied to the Linux side when needed (e.g. when VxWorks 
>> taskPrioritySet() is
>> invoked), but __real_pthread_setschedparam does not trigger 
>> the converse
>> propagation (since 2.4.4 actually, because experience showed 
>> that propagating in
>> that case made little sense, and would beget weird 
>> behaviours). The way priority
>> levels are normalized depends on the skin.
>>
> 
> Do you know how a Linux priority, set with pthread_setschedparam
> (real_pthread_setschedparam),
> correlates with the priorities shown by "ps l". In our case we realized
> that ps does not show the priority set with pthread_setschedparam. Is
> there a command which shows the "real" priority ?

With 2.6 you can list the priority with:

# ps -meo pid,tid,class,rtprio,ni,pri,psr,pcpu,stat,wchan:14,comm

But it's normally not available with the busybox ps command. Then you 
can look into /proc/<id>/stat directly. Here is the kernel's printk 
statement:

  http://lxr.linux.no/linux/fs/proc/array.c#L472

Unfortunately, this will not work for 2.4 either. The only simple way I 
see is to write a little program using sched_getscheduler() to read the 
RT priority and policy for the specified pid.

Wolfgang.


  reply	other threads:[~2008-07-25  8:09 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-07-25  7:40 [Xenomai-help] Thread policy and priority in secondary mode roderik.wildenburg
2008-07-25  8:09 ` Wolfgang Grandegger [this message]
2008-07-25  9:19 ` Philippe Gerum
     [not found] <5D63919D95F87E4D9D34FF7748CE2C2A0137F4FC@ARVMAIL1.mra.roland-man.biz>
2008-07-23 14:26 ` Philippe Gerum
  -- strict thread matches above, loose matches on Subject: below --
2008-07-23 10:03 Wolfgang Grandegger
2008-07-23 12:25 ` Philippe Gerum

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=48898A23.9030308@domain.hid \
    --to=wg@domain.hid \
    --cc=roderik.wildenburg@domain.hid \
    --cc=xenomai@xenomai.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.