[Xenomai-help] Virtual IRQ Numbers in powerpc?

[Xenomai-help] Virtual IRQ Numbers in powerpc?

[A. Nolson]

Hi,

 i have a PowerPC 405 based system using a 2.6.30 Xenomai patched kernel
and I want to manage some interrupts in userspace for some non-critical
custom drivers. I have an OF device-tree where I indicate my HW
interrupts (the ones I need to use are not shared with any already
managed interrupt in the kernel). I have been trying to initialize an
interrupt handler in userspace with rt_intr_create by using the real
interrupt numbers indicated in the dts . The handler gets loaded but no
interrupt shows up. I have seen that rt_intr_create indicates that the
irq number is "architecture dependant", but I haven't been able to spot
further explanations on this. I know that arch=powerpc use virtual
interrupts and I am wondering if I first need to map the real interrupt
in a virtual interrupt number before using it( like one would do in a
normal linux kernel driver ).

I am doing this:

    if( (err = rt_intr_create(&(dev->intr_desc), "GPIO IRQ", irqno,
I_NOAUTOENA)) < 0 ){
        pdbg(DBG_WARN, "Cannot create interrupt for GPIO
rt_intr_create=%i\n", err);
        return err;
    }
       
    rt_intr_enable (&(dev->intr_desc));

    if( (err = rt_task_spawn(&(dev->interrupt_task), "Int", 0, irq_prio,
0,fisr, (void*)&(dev->intr_desc))) < 0){
        pdbg(DBG_WARN, "Cannot Spawn ISR for GPIO. err = %d\n", err);
        return err;
    }

 What am I missing?

Thanks!

 /Alfred

Re: [Xenomai-help] Virtual IRQ Numbers in powerpc?

[Philippe Gerum]

On Sat, 2009-10-24 at 19:05 +0200, A. Nolson wrote:
> Hi,
> 
>  i have a PowerPC 405 based system using a 2.6.30 Xenomai patched kernel
> and I want to manage some interrupts in userspace for some non-critical
> custom drivers. I have an OF device-tree where I indicate my HW
> interrupts (the ones I need to use are not shared with any already
> managed interrupt in the kernel). I have been trying to initialize an
> interrupt handler in userspace with rt_intr_create by using the real
> interrupt numbers indicated in the dts . The handler gets loaded but no
> interrupt shows up. I have seen that rt_intr_create indicates that the
> irq number is "architecture dependant", but I haven't been able to spot
> further explanations on this. I know that arch=powerpc use virtual
> interrupts and I am wondering if I first need to map the real interrupt
> in a virtual interrupt number before using it( like one would do in a
> normal linux kernel driver ).

Yes, Xenomai/ppc uses virtual interrupt mappings (which are NOT related
at all with virtual I-pipe interrupts though). You should always use the
same IRQ numbers as the vanilla kernel exposes to its drivers, this is a
rule of thumb with all archs Xenomai supports.

To sum up, you should replace the real (OF) hw IRQ number in the call to
rt_intr_create() by its remapped counterpart.

> 
> I am doing this:
> 
>     if( (err = rt_intr_create(&(dev->intr_desc), "GPIO IRQ", irqno,
> I_NOAUTOENA)) < 0 ){
>         pdbg(DBG_WARN, "Cannot create interrupt for GPIO
> rt_intr_create=%i\n", err);
>         return err;
>     }
>        
>     rt_intr_enable (&(dev->intr_desc));
> 
>     if( (err = rt_task_spawn(&(dev->interrupt_task), "Int", 0, irq_prio,
> 0,fisr, (void*)&(dev->intr_desc))) < 0){
>         pdbg(DBG_WARN, "Cannot Spawn ISR for GPIO. err = %d\n", err);
>         return err;
>     }
> 
>  What am I missing?
> 
> Thanks!
> 
>  /Alfred
> 
> _______________________________________________
> Xenomai-help mailing list
> Xenomai-help@domain.hid
> https://mail.gna.org/listinfo/xenomai-help
-- 
Philippe.

Re: [Xenomai-help] Virtual IRQ Numbers in powerpc?

[A. Nolson]

Philippe Gerum wrote:
> On Sat, 2009-10-24 at 19:05 +0200, A. Nolson wrote:
>   
>> Hi,
>>
>>  i have a PowerPC 405 based system using a 2.6.30 Xenomai patched kernel
>> and I want to manage some interrupts in userspace for some non-critical
>> custom drivers. I have an OF device-tree where I indicate my HW
>> interrupts (the ones I need to use are not shared with any already
>> managed interrupt in the kernel). I have been trying to initialize an
>> interrupt handler in userspace with rt_intr_create by using the real
>> interrupt numbers indicated in the dts . The handler gets loaded but no
>> interrupt shows up. I have seen that rt_intr_create indicates that the
>> irq number is "architecture dependant", but I haven't been able to spot
>> further explanations on this. I know that arch=powerpc use virtual
>> interrupts and I am wondering if I first need to map the real interrupt
>> in a virtual interrupt number before using it( like one would do in a
>> normal linux kernel driver ).
>>     
>
> Yes, Xenomai/ppc uses virtual interrupt mappings (which are NOT related
> at all with virtual I-pipe interrupts though). You should always use the
> same IRQ numbers as the vanilla kernel exposes to its drivers, this is a
> rule of thumb with all archs Xenomai supports.
>
> To sum up, you should replace the real (OF) hw IRQ number in the call to
> rt_intr_create() by its remapped counterpart.
>   
Thanks a lot for the info Phillippe. Now the question(maybe obvious)
would be, how do I remap the interrupts outside the kernel? I remember
doing this in a kernel driver, but never in userspace that is where I
need to develop now with Xenomai.
>   
>> I am doing this:
>>
>>     if( (err = rt_intr_create(&(dev->intr_desc), "GPIO IRQ", irqno,
>> I_NOAUTOENA)) < 0 ){
>>         pdbg(DBG_WARN, "Cannot create interrupt for GPIO
>> rt_intr_create=%i\n", err);
>>         return err;
>>     }
>>        
>>     rt_intr_enable (&(dev->intr_desc));
>>
>>     if( (err = rt_task_spawn(&(dev->interrupt_task), "Int", 0, irq_prio,
>> 0,fisr, (void*)&(dev->intr_desc))) < 0){
>>         pdbg(DBG_WARN, "Cannot Spawn ISR for GPIO. err = %d\n", err);
>>         return err;
>>     }
>>
>>  What am I missing?
>>
>> Thanks!
>>
>>  /Alfred
>>
>> _______________________________________________
>> Xenomai-help mailing list
>> Xenomai-help@domain.hid
>> https://mail.gna.org/listinfo/xenomai-help
>>     

Re: [Xenomai-help] Virtual IRQ Numbers in powerpc?

[Philippe Gerum]

On Sat, 2009-10-24 at 20:42 +0200, A. Nolson wrote:
> Philippe Gerum wrote:
> > On Sat, 2009-10-24 at 19:05 +0200, A. Nolson wrote:
> >   
> >> Hi,
> >>
> >>  i have a PowerPC 405 based system using a 2.6.30 Xenomai patched kernel
> >> and I want to manage some interrupts in userspace for some non-critical
> >> custom drivers. I have an OF device-tree where I indicate my HW
> >> interrupts (the ones I need to use are not shared with any already
> >> managed interrupt in the kernel). I have been trying to initialize an
> >> interrupt handler in userspace with rt_intr_create by using the real
> >> interrupt numbers indicated in the dts . The handler gets loaded but no
> >> interrupt shows up. I have seen that rt_intr_create indicates that the
> >> irq number is "architecture dependant", but I haven't been able to spot
> >> further explanations on this. I know that arch=powerpc use virtual
> >> interrupts and I am wondering if I first need to map the real interrupt
> >> in a virtual interrupt number before using it( like one would do in a
> >> normal linux kernel driver ).
> >>     
> >
> > Yes, Xenomai/ppc uses virtual interrupt mappings (which are NOT related
> > at all with virtual I-pipe interrupts though). You should always use the
> > same IRQ numbers as the vanilla kernel exposes to its drivers, this is a
> > rule of thumb with all archs Xenomai supports.
> >
> > To sum up, you should replace the real (OF) hw IRQ number in the call to
> > rt_intr_create() by its remapped counterpart.
> >   
> Thanks a lot for the info Phillippe. Now the question(maybe obvious)
> would be, how do I remap the interrupts outside the kernel? I remember
> doing this in a kernel driver, but never in userspace that is where I
> need to develop now with Xenomai.

You will still need a piece of kernel code to do that; Xenomai does not
provide any userland interface for this purpose.

Maybe you could just declare your custom devices in your dts file, with
the appropriate "interrupts" properties, so that they get virtualized
during the prom parsing stage, then mapped to the UIC they belong to?

> >   
> >> I am doing this:
> >>
> >>     if( (err = rt_intr_create(&(dev->intr_desc), "GPIO IRQ", irqno,
> >> I_NOAUTOENA)) < 0 ){
> >>         pdbg(DBG_WARN, "Cannot create interrupt for GPIO
> >> rt_intr_create=%i\n", err);
> >>         return err;
> >>     }
> >>        
> >>     rt_intr_enable (&(dev->intr_desc));
> >>
> >>     if( (err = rt_task_spawn(&(dev->interrupt_task), "Int", 0, irq_prio,
> >> 0,fisr, (void*)&(dev->intr_desc))) < 0){
> >>         pdbg(DBG_WARN, "Cannot Spawn ISR for GPIO. err = %d\n", err);
> >>         return err;
> >>     }
> >>
> >>  What am I missing?
> >>
> >> Thanks!
> >>
> >>  /Alfred
> >>
> >> _______________________________________________
> >> Xenomai-help mailing list
> >> Xenomai-help@domain.hid
> >> https://mail.gna.org/listinfo/xenomai-help
> >>     
> 
-- 
Philippe.

Re: [Xenomai-help] Virtual IRQ Numbers in powerpc?

[Philippe Gerum]

On Sat, 2009-10-24 at 19:05 +0200, A. Nolson wrote:
> Hi,
> 
>  i have a PowerPC 405 based system using a 2.6.30 Xenomai patched kernel

Btw,

adeos-ipipe-2.6.30-powerpc-DENX-2.7-00:   fine
adeos-ipipe-2.6.30.3-powerpc-DENX-2.7-01: BROKEN
adeos-ipipe-2.6.30.3-powerpc-DENX-2.7-02: fine

> and I want to manage some interrupts in userspace for some non-critical
> custom drivers. I have an OF device-tree where I indicate my HW
> interrupts (the ones I need to use are not shared with any already
> managed interrupt in the kernel). I have been trying to initialize an
> interrupt handler in userspace with rt_intr_create by using the real
> interrupt numbers indicated in the dts . The handler gets loaded but no
> interrupt shows up. I have seen that rt_intr_create indicates that the
> irq number is "architecture dependant", but I haven't been able to spot
> further explanations on this. I know that arch=powerpc use virtual
> interrupts and I am wondering if I first need to map the real interrupt
> in a virtual interrupt number before using it( like one would do in a
> normal linux kernel driver ).
> 
> I am doing this:
> 
>     if( (err = rt_intr_create(&(dev->intr_desc), "GPIO IRQ", irqno,
> I_NOAUTOENA)) < 0 ){
>         pdbg(DBG_WARN, "Cannot create interrupt for GPIO
> rt_intr_create=%i\n", err);
>         return err;
>     }
>        
>     rt_intr_enable (&(dev->intr_desc));
> 
>     if( (err = rt_task_spawn(&(dev->interrupt_task), "Int", 0, irq_prio,
> 0,fisr, (void*)&(dev->intr_desc))) < 0){
>         pdbg(DBG_WARN, "Cannot Spawn ISR for GPIO. err = %d\n", err);
>         return err;
>     }
> 
>  What am I missing?
> 
> Thanks!
> 
>  /Alfred
> 
> _______________________________________________
> Xenomai-help mailing list
> Xenomai-help@domain.hid
> https://mail.gna.org/listinfo/xenomai-help
-- 
Philippe.

[Xenomai-help] Supported Architecture?

[Matthias Frauendorf]

Hello,

does anybody know if xenomai is working with the following architecture:

Intel PC-Mainboard D945GSEJT
with integrated Intel Atom Processor N270
in Mini-ITX Configuration

http://www.intel.com/cd/products/services/emea/deu/motherboards/desktop/d945gsejt/overview/416876.htm

Thank you!

Kind Regards,
Matthias Frauendorf









 
-- 
GRATIS für alle GMX-Mitglieder: Die maxdome Movie-FLAT!
Jetzt freischalten unter http://portal.gmx.net/de/go/maxdome01

Re: [Xenomai-help] Supported Architecture?

[Pierre Ficheux]

Matthias Frauendorf a écrit :
> Hello,
> 
> does anybody know if xenomai is working with the following architecture:
> 
> Intel PC-Mainboard D945GSEJT
> with integrated Intel Atom Processor N270
> in Mini-ITX Configuration
> 
> http://www.intel.com/cd/products/services/emea/deu/motherboards/desktop/d945gsejt/overview/416876.htm
> 
> Thank you!
> 
> Kind Regards,
> Matthias Frauendorf
> 
>

Should work fine. I currently use Xenomai on my NetPC with Atom N270.


-- 
Pierre FICHEUX -/- CTO OW/OS4I, France -\- pierre.ficheux@domain.hid
                                         http://www.os4i.com
                                         http://www.ficheux.org
I would love to change the world, but they won't give me the source code

Re: [Xenomai-help] Virtual IRQ Numbers in powerpc?

[Richard Cochran]

On Sun, Oct 25, 2009 at 12:13:45AM +0200, Philippe Gerum wrote:
> 
> adeos-ipipe-2.6.30-powerpc-DENX-2.7-00:   fine
> adeos-ipipe-2.6.30.3-powerpc-DENX-2.7-01: BROKEN
> adeos-ipipe-2.6.30.3-powerpc-DENX-2.7-02: fine

I am using git to track ipipe, but these two commits

  da0e45d1e84389bdf5e750e9179767fd274f7031
  9010961a407699a3908d7c2a842a8ece6084cee3

don't show up when I pull from

  git://git.denx.de/ipipe-2.6.git

I think that is because the commit are not in any branch. I think they
should be in branch ipipe-2.6.30-powerpc.

See this:

  http://git.denx.de/?p=ipipe-2.6.git;a=shortlog;h=9010961a407699a3908d7c2a842a8ece6084cee3

Richard

Re: [Xenomai-help] Virtual IRQ Numbers in powerpc?

[Philippe Gerum]

On Mon, 2009-11-02 at 10:54 +0100, Richard Cochran wrote:
> On Sun, Oct 25, 2009 at 12:13:45AM +0200, Philippe Gerum wrote:
> > 
> > adeos-ipipe-2.6.30-powerpc-DENX-2.7-00:   fine
> > adeos-ipipe-2.6.30.3-powerpc-DENX-2.7-01: BROKEN
> > adeos-ipipe-2.6.30.3-powerpc-DENX-2.7-02: fine
> 
> I am using git to track ipipe, but these two commits
> 
>   da0e45d1e84389bdf5e750e9179767fd274f7031
>   9010961a407699a3908d7c2a842a8ece6084cee3
> 
> don't show up when I pull from
> 
>   git://git.denx.de/ipipe-2.6.git
> 
> I think that is because the commit are not in any branch. I think they
> should be in branch ipipe-2.6.30-powerpc.
> 
> See this:
> 
>   http://git.denx.de/?p=ipipe-2.6.git;a=shortlog;h=9010961a407699a3908d7c2a842a8ece6084cee3

Those commits seem to still be in flight from my remote staging tree to
the public one, waiting for a mirroring to happen. They are indeed part
of ipipe-2.6.30-powerpc; those are actually the latest commits on top of
2.7-01, leading to 2.7-02.

> 
> Richard
> 
> 
> 


-- 
Philippe.