[Xenomai] Spurious interrupt are raised with 2.6.1 xenomai patch

[Xenomai] Spurious interrupt are raised with 2.6.1 xenomai patch

[Srikanth C]

Hi,

Spurious interrupts are raised on 2.6.1 Xenomai patch
with Linux Kernel version 3.2.34

for every pci interrupt raise  there is one spurious interrupt also getting
raised.

Without Xenomai patch pci interrupts are working fine (Spurious interrupts
are not raised).


Thanks and Regards,
Srikanth

Re: [Xenomai] Spurious interrupt are raised with 2.6.1 xenomai patch

[Gilles Chanteperdrix]

On 04/24/2013 08:29 AM, Srikanth C wrote:

> Hi,
> 
> Spurious interrupts are raised on 2.6.1 Xenomai patch
> with Linux Kernel version 3.2.34
> 
> for every pci interrupt raise  there is one spurious interrupt also getting
> raised.
> 
> Without Xenomai patch pci interrupts are working fine (Spurious interrupts
> are not raised).


What architecture? What kind of interrupt? Do you know the latest
version of Xenomai is 2.6.2.1, not 2.6.1? Where did you get an I-pipe
patch for 3.2.34 as the one we provide is for 3.2.21?

-- 
                                                                Gilles.

Re: [Xenomai] Spurious interrupt are raised with 2.6.1 xenomai patch

[Srikanth C]

Arch is PPC and interrupt is PCI, level sensitive interrupt.

ipipe 3.2.21 patch is applied to kernel 3.2.24 and even I tried with
ipipe-core-3.5.7-powerpc-3.patch with 3.5.7 kernel
same result is observed.



On Wed, Apr 24, 2013 at 12:19 PM, Gilles Chanteperdrix <
gilles.chanteperdrix@xenomai.org> wrote:

> On 04/24/2013 08:29 AM, Srikanth C wrote:
>
> > Hi,
> >
> > Spurious interrupts are raised on 2.6.1 Xenomai patch
> > with Linux Kernel version 3.2.34
> >
> > for every pci interrupt raise  there is one spurious interrupt also
> getting
> > raised.
> >
> > Without Xenomai patch pci interrupts are working fine (Spurious
> interrupts
> > are not raised).
>
>
> What architecture? What kind of interrupt? Do you know the latest
> version of Xenomai is 2.6.2.1, not 2.6.1? Where did you get an I-pipe
> patch for 3.2.34 as the one we provide is for 3.2.21?
>
> --
>                                                                 Gilles.
>

Re: [Xenomai] Spurious interrupt are raised with 2.6.1 xenomai patch

[Philippe Gerum]

On 04/24/2013 08:57 AM, Srikanth C wrote:
> Arch is PPC and interrupt is PCI, level sensitive interrupt.
>
> ipipe 3.2.21 patch is applied to kernel 3.2.24 and even I tried with
> ipipe-core-3.5.7-powerpc-3.patch with 3.5.7 kernel
> same result is observed.

Which powerpc platform? What is the test case? How do you notice the 
spurious interrupt? Is there any kernel log info you may share that 
reports it?

Please, help us to help you.

>
>
>
> On Wed, Apr 24, 2013 at 12:19 PM, Gilles Chanteperdrix <
> gilles.chanteperdrix@xenomai.org> wrote:
>
>> On 04/24/2013 08:29 AM, Srikanth C wrote:
>>
>>> Hi,
>>>
>>> Spurious interrupts are raised on 2.6.1 Xenomai patch
>>> with Linux Kernel version 3.2.34
>>>
>>> for every pci interrupt raise  there is one spurious interrupt also
>> getting
>>> raised.
>>>
>>> Without Xenomai patch pci interrupts are working fine (Spurious
>> interrupts
>>> are not raised).
>>
>>
>> What architecture? What kind of interrupt? Do you know the latest
>> version of Xenomai is 2.6.2.1, not 2.6.1? Where did you get an I-pipe
>> patch for 3.2.34 as the one we provide is for 3.2.21?
>>
>> --
>>                                                                  Gilles.
>>
> _______________________________________________
> Xenomai mailing list
> Xenomai@xenomai.org
> http://www.xenomai.org/mailman/listinfo/xenomai
>


-- 
Philippe.

Re: [Xenomai] Spurious interrupt are raised with 2.6.1 xenomai patch

[Srikanth C]

Sorry there is typo mistake in kernel version 3.2.24 instead of 3.2.34

Please find the attached boot log and interrupts(/proc/interuupts) log

Added a debug message in _ipipe_grab_irq function to verify the spurious
interrupt occurrence after IACK read
test code pasted below :
count variable is global variable and initialised to zero
*****************************************************************
count++;
irq = ppc_md.get_irq();
        if (unlikely(irq == NO_IRQ)) {
                __get_cpu_var(irq_stat).spurious_irqs++;
                printk("S. %d",count);
                return __ipipe_exit_irq(regs);
        }
******************************************************************
Continuous odd numbers is printed for spurious interrupt
S. 1
S. 3
S. 5



On Wed, Apr 24, 2013 at 2:29 PM, Philippe Gerum <rpm@xenomai.org> wrote:

> On 04/24/2013 08:57 AM, Srikanth C wrote:
>
>> Arch is PPC and interrupt is PCI, level sensitive interrupt.
>>
>> ipipe 3.2.21 patch is applied to kernel 3.2.24 and even I tried with
>> ipipe-core-3.5.7-powerpc-3.**patch with 3.5.7 kernel
>> same result is observed.
>>
>
> Which powerpc platform? What is the test case? How do you notice the
> spurious interrupt? Is there any kernel log info you may share that reports
> it?
>
> Please, help us to help you.
>
>
>>
>>
>> On Wed, Apr 24, 2013 at 12:19 PM, Gilles Chanteperdrix <
>> gilles.chanteperdrix@xenomai.**org <gilles.chanteperdrix@xenomai.org>>
>> wrote:
>>
>>  On 04/24/2013 08:29 AM, Srikanth C wrote:
>>>
>>>  Hi,
>>>>
>>>> Spurious interrupts are raised on 2.6.1 Xenomai patch
>>>> with Linux Kernel version 3.2.34
>>>>
>>>> for every pci interrupt raise  there is one spurious interrupt also
>>>>
>>> getting
>>>
>>>> raised.
>>>>
>>>> Without Xenomai patch pci interrupts are working fine (Spurious
>>>>
>>> interrupts
>>>
>>>> are not raised).
>>>>
>>>
>>>
>>> What architecture? What kind of interrupt? Do you know the latest
>>> version of Xenomai is 2.6.2.1, not 2.6.1? Where did you get an I-pipe
>>> patch for 3.2.34 as the one we provide is for 3.2.21?
>>>
>>> --
>>>                                                                  Gilles.
>>>
>>>  ______________________________**_________________
>> Xenomai mailing list
>> Xenomai@xenomai.org
>> http://www.xenomai.org/**mailman/listinfo/xenomai<http://www.xenomai.org/mailman/listinfo/xenomai>
>>
>>
>
> --
> Philippe.
>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: dmesg_log
Type: application/octet-stream
Size: 38565 bytes
Desc: not available
URL: <http://www.xenomai.org/pipermail/xenomai/attachments/20130424/29f2a97b/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: proc_log
Type: application/octet-stream
Size: 414 bytes
Desc: not available
URL: <http://www.xenomai.org/pipermail/xenomai/attachments/20130424/29f2a97b/attachment-0001.obj>

Re: [Xenomai] Spurious interrupt are raised with 2.6.1 xenomai patch

[Srikanth C]

Test Results with adeos-ipipe patch:

H/W Setup is same, S/W Changed to adeos-ipipe with 3.0.8 kernel.

adeos-ipipe-3.0.8-powerpc-2.13-04.patch applied to 3.0.8 kernel, spurious
interrupt is not raised with this patch.

All ipipe series patches raising spurious interrupt in powerpc.



On Wed, Apr 24, 2013 at 3:03 PM, Srikanth C <srikanthc9@gmail.com> wrote:

> Sorry there is typo mistake in kernel version 3.2.24 instead of 3.2.34
>
> Please find the attached boot log and interrupts(/proc/interuupts) log
>
> Added a debug message in _ipipe_grab_irq function to verify the spurious
> interrupt occurrence after IACK read
> test code pasted below :
> count variable is global variable and initialised to zero
> *****************************************************************
> count++;
> irq = ppc_md.get_irq();
>         if (unlikely(irq == NO_IRQ)) {
>                 __get_cpu_var(irq_stat).spurious_irqs++;
>                 printk("S. %d",count);
>                 return __ipipe_exit_irq(regs);
>         }
> ******************************************************************
> Continuous odd numbers is printed for spurious interrupt
> S. 1
> S. 3
> S. 5
>
>
>
> On Wed, Apr 24, 2013 at 2:29 PM, Philippe Gerum <rpm@xenomai.org> wrote:
>
>> On 04/24/2013 08:57 AM, Srikanth C wrote:
>>
>>> Arch is PPC and interrupt is PCI, level sensitive interrupt.
>>>
>>> ipipe 3.2.21 patch is applied to kernel 3.2.24 and even I tried with
>>> ipipe-core-3.5.7-powerpc-3.**patch with 3.5.7 kernel
>>> same result is observed.
>>>
>>
>> Which powerpc platform? What is the test case? How do you notice the
>> spurious interrupt? Is there any kernel log info you may share that reports
>> it?
>>
>> Please, help us to help you.
>>
>>
>>>
>>>
>>> On Wed, Apr 24, 2013 at 12:19 PM, Gilles Chanteperdrix <
>>> gilles.chanteperdrix@xenomai.**org <gilles.chanteperdrix@xenomai.org>>
>>> wrote:
>>>
>>>  On 04/24/2013 08:29 AM, Srikanth C wrote:
>>>>
>>>>  Hi,
>>>>>
>>>>> Spurious interrupts are raised on 2.6.1 Xenomai patch
>>>>> with Linux Kernel version 3.2.34
>>>>>
>>>>> for every pci interrupt raise  there is one spurious interrupt also
>>>>>
>>>> getting
>>>>
>>>>> raised.
>>>>>
>>>>> Without Xenomai patch pci interrupts are working fine (Spurious
>>>>>
>>>> interrupts
>>>>
>>>>> are not raised).
>>>>>
>>>>
>>>>
>>>> What architecture? What kind of interrupt? Do you know the latest
>>>> version of Xenomai is 2.6.2.1, not 2.6.1? Where did you get an I-pipe
>>>> patch for 3.2.34 as the one we provide is for 3.2.21?
>>>>
>>>> --
>>>>                                                                  Gilles.
>>>>
>>>>  ______________________________**_________________
>>> Xenomai mailing list
>>> Xenomai@xenomai.org
>>> http://www.xenomai.org/**mailman/listinfo/xenomai<http://www.xenomai.org/mailman/listinfo/xenomai>
>>>
>>>
>>
>> --
>> Philippe.
>>
>
>

Re: [Xenomai] Spurious interrupt are raised with 2.6.1 xenomai patch

[Philippe Gerum]

On 04/25/2013 09:57 AM, Srikanth C wrote:
> Test Results with adeos-ipipe patch:
>
> H/W Setup is same, S/W Changed to adeos-ipipe with 3.0.8 kernel.
>
> adeos-ipipe-3.0.8-powerpc-2.13-04.patch applied to 3.0.8 kernel,
> spurious interrupt is not raised with this patch.
>

I still don't know what ppc platform you are running on, which does not 
allow me to figure out which interrupt controller might be involved, and 
how to try reproducing this issue.

Again and I won't ask for this once more:

- what is your ppc platform: 40x, 44x, 52xx, 82xx, 83xx, 85xx, 86xx, 
anything else?

> All ipipe series patches raising spurious interrupt in powerpc.
>

The few patches > 3.0.8 you tried on your own PCI setup and kernel 
configuration raise spurious interrupts on your own board, would be a 
more informative way to describe the situation.

-- 
Philippe.

Re: [Xenomai] Spurious interrupt are raised with 2.6.1 xenomai patch

[Philippe Gerum]

On 04/25/2013 11:03 AM, Philippe Gerum wrote:
> On 04/25/2013 09:57 AM, Srikanth C wrote:
>> Test Results with adeos-ipipe patch:
>>
>> H/W Setup is same, S/W Changed to adeos-ipipe with 3.0.8 kernel.
>>
>> adeos-ipipe-3.0.8-powerpc-2.13-04.patch applied to 3.0.8 kernel,
>> spurious interrupt is not raised with this patch.
>>
>
> I still don't know what ppc platform you are running on, which does not
> allow me to figure out which interrupt controller might be involved, and
> how to try reproducing this issue.
>
> Again and I won't ask for this once more:
>
> - what is your ppc platform: 40x, 44x, 52xx, 82xx, 83xx, 85xx, 86xx,
> anything else?

I need the cpu type _and_ the board type. Attaching your kernel .config 
would provide the required information.

>
>> All ipipe series patches raising spurious interrupt in powerpc.
>>
>
> The few patches > 3.0.8 you tried on your own PCI setup and kernel
> configuration raise spurious interrupts on your own board, would be a
> more informative way to describe the situation.
>


-- 
Philippe.

Re: [Xenomai] Spurious interrupt are raised with 2.6.1 xenomai patch

[Philippe Gerum]

On 04/25/2013 11:06 AM, Philippe Gerum wrote:
> On 04/25/2013 11:03 AM, Philippe Gerum wrote:
>> On 04/25/2013 09:57 AM, Srikanth C wrote:
>>> Test Results with adeos-ipipe patch:
>>>
>>> H/W Setup is same, S/W Changed to adeos-ipipe with 3.0.8 kernel.
>>>
>>> adeos-ipipe-3.0.8-powerpc-2.13-04.patch applied to 3.0.8 kernel,
>>> spurious interrupt is not raised with this patch.
>>>
>>
>> I still don't know what ppc platform you are running on, which does not
>> allow me to figure out which interrupt controller might be involved, and
>> how to try reproducing this issue.
>>
>> Again and I won't ask for this once more:
>>
>> - what is your ppc platform: 40x, 44x, 52xx, 82xx, 83xx, 85xx, 86xx,
>> anything else?
>
> I need the cpu type _and_ the board type. Attaching your kernel .config
> would provide the required information.

CONFIG_PCI_MSI as found in your configuration won't work, with any past 
or current pipeline release for powerpc (3.0.8 was certainly a fluke, or 
MSI was disabled there). The IRQ chip handlers are still attempting to 
mask the MSI source upon interrupt, which won't fly in the I-pipe case.

-- 
Philippe.

Re: [Xenomai] Spurious interrupt are raised with 2.6.1 xenomai patch

[Srikanth C]

CONFIG_PCI_MSI is disabled in kernel, still spurious interrupts are raised.

I tried with 3.1.10 kernel with adeos 3.1.10 patch with the same config
file but  no spurious interrupts are raised.

adeos patches are working fine, problem only with ipipe series.

why ppc_md.get_irq is reading FF when interrupt is raised (this is
happening after patching the xenomai, without xenomai patch same kernel
with same configuration is working fine).


On Fri, Apr 26, 2013 at 8:16 PM, Philippe Gerum <rpm@xenomai.org> wrote:

> On 04/25/2013 11:06 AM, Philippe Gerum wrote:
>
>> On 04/25/2013 11:03 AM, Philippe Gerum wrote:
>>
>>> On 04/25/2013 09:57 AM, Srikanth C wrote:
>>>
>>>> Test Results with adeos-ipipe patch:
>>>>
>>>> H/W Setup is same, S/W Changed to adeos-ipipe with 3.0.8 kernel.
>>>>
>>>> adeos-ipipe-3.0.8-powerpc-2.**13-04.patch applied to 3.0.8 kernel,
>>>> spurious interrupt is not raised with this patch.
>>>>
>>>>
>>> I still don't know what ppc platform you are running on, which does not
>>> allow me to figure out which interrupt controller might be involved, and
>>> how to try reproducing this issue.
>>>
>>> Again and I won't ask for this once more:
>>>
>>> - what is your ppc platform: 40x, 44x, 52xx, 82xx, 83xx, 85xx, 86xx,
>>> anything else?
>>>
>>
>> I need the cpu type _and_ the board type. Attaching your kernel .config
>> would provide the required information.
>>
>
> CONFIG_PCI_MSI as found in your configuration won't work, with any past or
> current pipeline release for powerpc (3.0.8 was certainly a fluke, or MSI
> was disabled there). The IRQ chip handlers are still attempting to mask the
> MSI source upon interrupt, which won't fly in the I-pipe case.
>
> --
> Philippe.
>

Re: [Xenomai] Spurious interrupt are raised with 2.6.1 xenomai patch

[Srikanth C]

Problem is resolved, issue with the hardware.

Thanks for your valuable time and support.

Regards,
Srikanth
On Apr 27, 2013 12:19 AM, "Srikanth C" <srikanthc9@gmail.com> wrote:

>
> CONFIG_PCI_MSI is disabled in kernel, still spurious interrupts are raised.
>
> I tried with 3.1.10 kernel with adeos 3.1.10 patch with the same config
> file but  no spurious interrupts are raised.
>
> adeos patches are working fine, problem only with ipipe series.
>
> why ppc_md.get_irq is reading FF when interrupt is raised (this is
> happening after patching the xenomai, without xenomai patch same kernel
> with same configuration is working fine).
>
>
> On Fri, Apr 26, 2013 at 8:16 PM, Philippe Gerum <rpm@xenomai.org> wrote:
>
>> On 04/25/2013 11:06 AM, Philippe Gerum wrote:
>>
>>> On 04/25/2013 11:03 AM, Philippe Gerum wrote:
>>>
>>>> On 04/25/2013 09:57 AM, Srikanth C wrote:
>>>>
>>>>> Test Results with adeos-ipipe patch:
>>>>>
>>>>> H/W Setup is same, S/W Changed to adeos-ipipe with 3.0.8 kernel.
>>>>>
>>>>> adeos-ipipe-3.0.8-powerpc-2.**13-04.patch applied to 3.0.8 kernel,
>>>>> spurious interrupt is not raised with this patch.
>>>>>
>>>>>
>>>> I still don't know what ppc platform you are running on, which does not
>>>> allow me to figure out which interrupt controller might be involved, and
>>>> how to try reproducing this issue.
>>>>
>>>> Again and I won't ask for this once more:
>>>>
>>>> - what is your ppc platform: 40x, 44x, 52xx, 82xx, 83xx, 85xx, 86xx,
>>>> anything else?
>>>>
>>>
>>> I need the cpu type _and_ the board type. Attaching your kernel .config
>>> would provide the required information.
>>>
>>
>> CONFIG_PCI_MSI as found in your configuration won't work, with any past
>> or current pipeline release for powerpc (3.0.8 was certainly a fluke, or
>> MSI was disabled there). The IRQ chip handlers are still attempting to mask
>> the MSI source upon interrupt, which won't fly in the I-pipe case.
>>
>> --
>> Philippe.
>>
>
>