MIssed interrupts

Hi,

If a particual IRQ line has been disabled (disable_irq()), what
happens if the device on that IRQ line generates interrupt? Does it
result in missed interrupt?

And in case of Uniprocessor machine, what happens if interrupts are
generated while spinlock_irq_save() etc?

TIA,

Rick
-
To unsubscribe from this list: send the line "unsubscribe linux-newbie" in
the body of a message to majordomo [at] vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.linux-learn.org/faqs
Rick Brown [ Do, 07 Februar 2008 06:12 ] [ ID #1922975 ]

Re: MIssed interrupts

Hi....

On Feb 7, 2008 12:12 PM, Rick Brown <rick.brown.3 [at] gmail.com> wrote:
> Hi,
>
> If a particual IRQ line has been disabled (disable_irq()), what
> happens if the device on that IRQ line generates interrupt? Does it
> result in missed interrupt?

AFAIK at the time it's in disabled state...yes it will be missed. But
once you enabled it again.... missing interrupts will be received. It
is different when interrupt is masked...it will be simply ignored.

regards.

Mulyadi.
-
To unsubscribe from this list: send the line "unsubscribe linux-newbie" in
the body of a message to majordomo [at] vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.linux-learn.org/faqs
Mulyadi Santosa [ Do, 07 Februar 2008 16:50 ] [ ID #1922985 ]

Re: MIssed interrupts

On 2/7/08, Rick Brown <rick.brown.3 [at] gmail.com> wrote:
> Hi,
>
> If a particual IRQ line has been disabled (disable_irq()), what
> happens if the device on that IRQ line generates interrupt? Does it
> result in missed interrupt?

Yes, and there are a few other problems with interrupts too. For
missed interrupt, we solved by "bottom half". In essence the
interrupt handler has to be very short (which is immuned to another
interrupt trigger), and complete as soon as possible to get ready to
receive another interrupt, while the "deferred context" can execute
slowly, but is all the time interrupt-ready.

>
> And in case of Uniprocessor machine, what happens if interrupts are
> generated while spinlock_irq_save() etc?
>

Same, missed interrupts.

Check this out:

http://www.cs.utah.edu/classes/cs5785/slides/08-6up.pdf
--
To unsubscribe from this list: send the line "unsubscribe linux-newbie" in
the body of a message to majordomo [at] vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.linux-learn.org/faqs
Peter Teoh [ Fr, 29 Februar 2008 04:35 ] [ ID #1926755 ]

Re: MIssed interrupts

On Thu, Feb 7, 2008 at 11:50 PM, Mulyadi Santosa
<mulyadi.santosa [at] gmail.com> wrote:
> Hi....
>
> AFAIK at the time it's in disabled state...yes it will be missed. But
> once you enabled it again.... missing interrupts will be received.

Do you mean that the PIC(Local APIC or something else) will retain the
disabled interrupt, and deliver it once again to the processor core as
soon as be re-enabled?

Regards,
Jike
--
To unsubscribe from this list: send the line "unsubscribe linux-newbie" in
the body of a message to majordomo [at] vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.linux-learn.org/faqs
Jike Song [ Fr, 29 Februar 2008 08:21 ] [ ID #1926756 ]

Re: MIssed interrupts

Hi...

On Fri, Feb 29, 2008 at 2:21 PM, Jike Song <albcamus [at] gmail.com> wrote:
> Do you mean that the PIC(Local APIC or something else) will retain the
> disabled interrupt, and deliver it once again to the processor core as
> soon as be re-enabled?

As far as I know, yes

regards,

Mulyadi.
--
To unsubscribe from this list: send the line "unsubscribe linux-newbie" in
the body of a message to majordomo [at] vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.linux-learn.org/faqs
Mulyadi Santosa [ Fr, 29 Februar 2008 12:43 ] [ ID #1926757 ]

Re: MIssed interrupts

Hi Mulyadi,

On Fri, Feb 29, 2008 at 3:21 PM, Jike Song <albcamus [at] gmail.com> wrote:
> On Thu, Feb 7, 2008 at 11:50 PM, Mulyadi Santosa
> <mulyadi.santosa [at] gmail.com> wrote:
> > Hi....
> >
> > AFAIK at the time it's in disabled state...yes it will be missed. But
> > once you enabled it again.... missing interrupts will be received.
>
> Do you mean that the PIC(Local APIC or something else) will retain the
> disabled interrupt, and deliver it once again to the processor core as
> soon as be re-enabled?
>

This is something new to me. After some some search, Mulyadi is
right. It is called "Interrupt Request Buffering". And according to
Unabridged Pentium 4 (pg 1537) while processing one interrupt, it can
receive another interrupt, but the 3rd one that comes in will be
discarded, and it is all on the same interrupt vector. Thanks for
the sharing everyone....

> Regards,
> Jike
--
To unsubscribe from this list: send the line "unsubscribe linux-newbie" in
the body of a message to majordomo [at] vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.linux-learn.org/faqs
Peter Teoh [ Fr, 29 Februar 2008 15:34 ] [ ID #1926758 ]

Re: MIssed interrupts

Hi...

On Fri, Feb 29, 2008 at 9:34 PM, Peter Teoh <htmldeveloper [at] gmail.com> wrote:
> This is something new to me. After some some search, Mulyadi is
> right. It is called "Interrupt Request Buffering". And according to
> Unabridged Pentium 4 (pg 1537) while processing one interrupt, it can
> receive another interrupt, but the 3rd one that comes in will be
> discarded, and it is all on the same interrupt vector. Thanks for
> the sharing everyone....

Well, I just recall the best I could from Understanding the Linux
kernel 3rd edition. If somebody can give further confirmation, that
would be great.

regards,

Mulyadi.
--
To unsubscribe from this list: send the line "unsubscribe linux-newbie" in
the body of a message to majordomo [at] vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.linux-learn.org/faqs
Mulyadi Santosa [ Sa, 01 März 2008 09:36 ] [ ID #1927004 ]

Re: MIssed interrupts

On Fri, Feb 29, 2008 at 7:43 PM, Mulyadi Santosa
<mulyadi.santosa [at] gmail.com> wrote:
> As far as I know, yes
>
> regards,
>
> Mulyadi.
>

Thanks for the confirmation.

Regards,
Jike
--
To unsubscribe from this list: send the line "unsubscribe linux-newbie" in
the body of a message to majordomo [at] vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.linux-learn.org/faqs
Jike Song [ Mo, 03 März 2008 02:10 ] [ ID #1927182 ]

Re: MIssed interrupts

More information on missing interrupts, esp on ARM platform:

http://www.embeddedrelated.com/usenet/embedded/show/72620-1. php

On Mon, Mar 3, 2008 at 9:10 AM, Jike Song <albcamus [at] gmail.com> wrote:
> On Fri, Feb 29, 2008 at 7:43 PM, Mulyadi Santosa
>
> <mulyadi.santosa [at] gmail.com> wrote:
>
> > As far as I know, yes
> >
> > regards,
> >
> > Mulyadi.
> >
>
> Thanks for the confirmation.
>
> Regards,
> Jike
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-newbie" in
> the body of a message to majordomo [at] vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.linux-learn.org/faqs
>



--
Regards,
Peter Teoh
--
To unsubscribe from this list: send the line "unsubscribe linux-newbie" in
the body of a message to majordomo [at] vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.linux-learn.org/faqs
Peter Teoh [ Mi, 02 April 2008 10:09 ] [ ID #1933570 ]
Linux » gmane.linux.newbie » MIssed interrupts

Vorheriges Thema: Correlation between linux-2.6-stable kernel tree and linux-2.6.24.4.tar.bz2
Nächstes Thema: What's the deal with 1/3G separation between kernel address space and user address space?