* [Xenomai-help] Compile error
@ 2012-01-18 15:12 Terry Fryar
2012-01-18 19:22 ` Gilles Chanteperdrix
0 siblings, 1 reply; 4+ messages in thread
From: Terry Fryar @ 2012-01-18 15:12 UTC (permalink / raw)
To: xenomai
[-- Attachment #1: Type: text/plain, Size: 642 bytes --]
Using xenomai 2.6.0 with 2.6.38 kernel. Turned off
CONFIG_XENO_OPT_PERVASIVE, and now I can't compile kernel:
CC kernel/xenomai/nucleus/synch.o
kernel/xenomai/nucleus/synch.c: In function 'xnsynch_release_thread':
kernel/xenomai/nucleus/synch.c:688:3: error: implicit declaration of
function 'xnthread_get_rescnt'
kernel/xenomai/nucleus/synch.c:689:4: error: implicit declaration of
function 'xnshadow_send_sig'
There are some ifdef's for this flag....how to compile xenomai kernel
without pervasive user support? Only using kernel modules with xenomai at
this point, and I'm trying to optimize the kernel performance....
Thanks!
[-- Attachment #2: Type: text/html, Size: 1406 bytes --]
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [Xenomai-help] Compile error
2012-01-18 15:12 [Xenomai-help] Compile error Terry Fryar
@ 2012-01-18 19:22 ` Gilles Chanteperdrix
2012-01-18 21:44 ` Terry Fryar
0 siblings, 1 reply; 4+ messages in thread
From: Gilles Chanteperdrix @ 2012-01-18 19:22 UTC (permalink / raw)
To: Terry Fryar; +Cc: xenomai
On 01/18/2012 04:12 PM, Terry Fryar wrote:
> Using xenomai 2.6.0 with 2.6.38 kernel. Turned off
> CONFIG_XENO_OPT_PERVASIVE, and now I can't compile kernel:
>
> CC kernel/xenomai/nucleus/synch.o
> kernel/xenomai/nucleus/synch.c: In function 'xnsynch_release_thread':
> kernel/xenomai/nucleus/synch.c:688:3: error: implicit declaration of
> function 'xnthread_get_rescnt'
> kernel/xenomai/nucleus/synch.c:689:4: error: implicit declaration of
> function 'xnshadow_send_sig'
>
> There are some ifdef's for this flag....how to compile xenomai kernel
> without pervasive user support? Only using kernel modules with xenomai at
> this point, and I'm trying to optimize the kernel performance....
I am afraid nobody has tried compiling without pervasive support for a
long time. Anyway, you will not gain enough that way: you will not
reduce the interrupt latency from 20us to 500ns. As I said, you can not
have a 500ns latency on any processor where xenomai is available.
--
Gilles.
^ permalink raw reply [flat|nested] 4+ messages in thread* Re: [Xenomai-help] Compile error
2012-01-18 19:22 ` Gilles Chanteperdrix
@ 2012-01-18 21:44 ` Terry Fryar
2012-01-18 22:32 ` Gilles Chanteperdrix
0 siblings, 1 reply; 4+ messages in thread
From: Terry Fryar @ 2012-01-18 21:44 UTC (permalink / raw)
To: 'Gilles Chanteperdrix'; +Cc: xenomai
I knew I wouldn't get enough...but I have other projects in between those
timing requirements and I wanted to see how little latency I could get.
Using an RTDM driver we wrote, I am measuring about 2us (not 20us) by
triggering an irq off a gpio and pulsing another gpio line and measuring the
difference between the irq edge and the pulsed edge using a logic analyzer.
I have a board that runs a suhdol linux sbc that is somehow handling gpio
interactions within nanosecond ranges, and I have no idea how that is being
done. Even flipping a gpio pin with direct register access inside a KLM is
around 62ns. I just don't know how they are managing to detect a gpio pin
state change and respond to several things all within a couple hundred ns??
I've looked at their app files and it's just a gpio driver and what appears
to be a single userspace app. A mystery I guess.....
-----Original Message-----
From: Gilles Chanteperdrix [mailto:gilles.chanteperdrix@xenomai.org]
Sent: Wednesday, January 18, 2012 1:23 PM
To: Terry Fryar
Cc: xenomai@xenomai.org
Subject: Re: [Xenomai-help] Compile error
On 01/18/2012 04:12 PM, Terry Fryar wrote:
> Using xenomai 2.6.0 with 2.6.38 kernel. Turned off
> CONFIG_XENO_OPT_PERVASIVE, and now I can't compile kernel:
>
> CC kernel/xenomai/nucleus/synch.o
> kernel/xenomai/nucleus/synch.c: In function 'xnsynch_release_thread':
> kernel/xenomai/nucleus/synch.c:688:3: error: implicit declaration of
> function 'xnthread_get_rescnt'
> kernel/xenomai/nucleus/synch.c:689:4: error: implicit declaration of
> function 'xnshadow_send_sig'
>
> There are some ifdef's for this flag....how to compile xenomai kernel
> without pervasive user support? Only using kernel modules with
> xenomai at this point, and I'm trying to optimize the kernel
performance....
I am afraid nobody has tried compiling without pervasive support for a long
time. Anyway, you will not gain enough that way: you will not reduce the
interrupt latency from 20us to 500ns. As I said, you can not have a 500ns
latency on any processor where xenomai is available.
--
Gilles.
^ permalink raw reply [flat|nested] 4+ messages in thread* Re: [Xenomai-help] Compile error
2012-01-18 21:44 ` Terry Fryar
@ 2012-01-18 22:32 ` Gilles Chanteperdrix
0 siblings, 0 replies; 4+ messages in thread
From: Gilles Chanteperdrix @ 2012-01-18 22:32 UTC (permalink / raw)
To: Terry Fryar; +Cc: xenomai
On 01/18/2012 10:44 PM, Terry Fryar wrote:
> I knew I wouldn't get enough...but I have other projects in between those
> timing requirements and I wanted to see how little latency I could get.
> Using an RTDM driver we wrote, I am measuring about 2us (not 20us) by
> triggering an irq off a gpio and pulsing another gpio line and measuring the
> difference between the irq edge and the pulsed edge using a logic analyzer.
That is probably because you are not measuring latency under load. So,
you do not see the worst case.
> I have a board that runs a suhdol linux sbc that is somehow handling gpio
> interactions within nanosecond ranges, and I have no idea how that is being
> done. Even flipping a gpio pin with direct register access inside a KLM is
> around 62ns. I just don't know how they are managing to detect a gpio pin
> state change and respond to several things all within a couple hundred ns??
>
> I've looked at their app files and it's just a gpio driver and what appears
> to be a single userspace app. A mystery I guess.....
I am talking about worst case here, not about average case.
--
Gilles.
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2012-01-18 22:32 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2012-01-18 15:12 [Xenomai-help] Compile error Terry Fryar
2012-01-18 19:22 ` Gilles Chanteperdrix
2012-01-18 21:44 ` Terry Fryar
2012-01-18 22:32 ` Gilles Chanteperdrix
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.