Bogus interrupt -1

Bogus interrupt -1

[Lars Brinkhoff]

I got these messages from benh's 2.4.4-pre1 running SMP:

  Bogus interrupt -1 from PC = c000781c
  Bogus interrupt -1 from PC = c0017ba4
  Bogus interrupt -1 from PC = c0024c90
  Bogus interrupt -1 from PC = c0017ba4
  Bogus interrupt -1 from PC = c0017c34
  Bogus interrupt -1 from PC = c0017ba4
  Bogus interrupt -1 from PC = c0006e38
  Bogus interrupt -1 from PC = c0006e38
  Bogus interrupt -1 from PC = c0006e38
  Bogus interrupt -1 from PC = c0006e38
  Bogus interrupt -1 from PC = c0006e38
  Bogus interrupt -1 from PC = c0006e38

According to my System.map:
c0006e38 = _nmask_and_or_msr + 0x10
c000781c = _switch_to + 0x184
c0017ba4 = schedule + 0x1084
c0017c34 = schedule + 0x1228
c0024c90 = del_timer_sync + 0x184

At first, the first six or so messages appeared, but everything seemed
fine.  Later, the last messages appeared, and it seemed like one of
the two processors in my 2001 Dual G4 had wedged.  According to "top",
there was always at least 50% idle time.

--
http://lars.nocrew.org/

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Bogus interrupt -1

[Benjamin Herrenschmidt]

>At first, the first six or so messages appeared, but everything seemed
>fine.  Later, the last messages appeared, and it seemed like one of
>the two processors in my 2001 Dual G4 had wedged.  According to "top",
>there was always at least 50% idle time.

Did you enable distribution of IRQs to all CPUs ?

I don't think it's completely safe yet.

Ben.


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Bogus interrupt -1

[Lars Brinkhoff]

Benjamin Herrenschmidt <benh@kernel.crashing.org> writes:
> >At first, the first six or so messages appeared, but everything seemed
> >fine.  Later, the last messages appeared, and it seemed like one of
> >the two processors in my 2001 Dual G4 had wedged.  According to "top",
> >there was always at least 50% idle time.
> Did you enable distribution of IRQs to all CPUs ?
> I don't think it's completely safe yet.

Yes, IRQ distribution was enabled.  I disabled it, and now the
messages don't appear any more.  However, it still seems like one CPU
hangs after some time.

--
http://lars.nocrew.org/

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Bogus interrupt -1

[Tony Mantler]

At 4:52 AM -0500 4/23/2001, Lars Brinkhoff wrote:
>Benjamin Herrenschmidt <benh@kernel.crashing.org> writes:
>> >At first, the first six or so messages appeared, but everything seemed
>> >fine.  Later, the last messages appeared, and it seemed like one of
>> >the two processors in my 2001 Dual G4 had wedged.  According to "top",
>> >there was always at least 50% idle time.
>> Did you enable distribution of IRQs to all CPUs ?
>> I don't think it's completely safe yet.
>
>Yes, IRQ distribution was enabled.  I disabled it, and now the
>messages don't appear any more.  However, it still seems like one CPU
>hangs after some time.

A friend of mine is having trouble with one CPU hanging in his dual G4
under MacOS X. You may want to check to make sure you don't have a hardware
problem.

The symptoms are pretty much (from what he's told me): The system runs for
a while, then one of the CPUs goes off into la-la land never to return
again. Not very complicated.


Cheers - Tony 'Nicoya' Mantler :)


--
Tony "Nicoya" Mantler - Renaissance Nerd Extraordinaire - nicoya@apia.dhs.org
Winnipeg, Manitoba, Canada           --           http://nicoya.feline.pp.se/


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Bogus interrupt -1

[Lars Brinkhoff]

Tony Mantler <nicoya@apia.dhs.org> writes:
> At 4:52 AM -0500 4/23/2001, Lars Brinkhoff wrote:
> > However, it still seems like one CPU hangs after some time.
>
> A friend of mine is having trouble with one CPU hanging in his dual
> G4 under MacOS X. The symptoms are pretty much (from what he's told
> me): The system runs for a while, then one of the CPUs goes off into
> la-la land never to return again.

The same thing happens to me, as far as I can tell.  Usually after a
few hours of operation.

> You may want to check to make sure you don't have a hardware
> problem.

Any idea how?  I ran the "Apple Hardware Test" CD without any failure
indications.

--
http://lars.nocrew.org/

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Bogus interrupt -1

[Tony Mantler]

At 7:24 AM -0500 4/23/2001, Lars Brinkhoff wrote:
>Tony Mantler <nicoya@apia.dhs.org> writes:
>> At 4:52 AM -0500 4/23/2001, Lars Brinkhoff wrote:
>> > However, it still seems like one CPU hangs after some time.
>>
>> A friend of mine is having trouble with one CPU hanging in his dual
>> G4 under MacOS X. The symptoms are pretty much (from what he's told
>> me): The system runs for a while, then one of the CPUs goes off into
>> la-la land never to return again.
>
>The same thing happens to me, as far as I can tell.  Usually after a
>few hours of operation.
>
>> You may want to check to make sure you don't have a hardware
>> problem.
>
>Any idea how?  I ran the "Apple Hardware Test" CD without any failure
>indications.

Other than "Run OS X (Or maybe Darwin) for a while and see if the same
thing happens", I don't know.


Cheers - Tony 'Nicoya' Mantler :)


--
Tony "Nicoya" Mantler - Renaissance Nerd Extraordinaire - nicoya@apia.dhs.org
Winnipeg, Manitoba, Canada           --           http://nicoya.feline.pp.se/


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Success? Both CPUs running (Was: Bogus interrupt -1)

[Lars Brinkhoff]

Tony Mantler <nicoya@apia.dhs.org> writes:
> At 4:52 AM -0500 4/23/2001, Lars Brinkhoff wrote:
> > However, it still seems like one CPU hangs after some time.
>
> A friend of mine is having trouble with one CPU hanging in his dual G4
> under MacOS X. You may want to check to make sure you don't have a hardware
> problem.
>
> The symptoms are pretty much (from what he's told me): The system runs for
> a while, then one of the CPUs goes off into la-la land never to return
> again. Not very complicated.

It seems to me there was a bug in benh 2.4.4-pre1.  I'm now running bk
2.4.5-pre1, and I have never seen a CPU hang.

--
http://lars.nocrew.org/

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Success? Both CPUs running (Was: Bogus interrupt -1)

[Benjamin Herrenschmidt]

>>
>> The symptoms are pretty much (from what he's told me): The system runs for
>> a while, then one of the CPUs goes off into la-la land never to return
>> again. Not very complicated.
>
>It seems to me there was a bug in benh 2.4.4-pre1.  I'm now running bk
>2.4.5-pre1, and I have never seen a CPU hang.

Can you tell me how my current 2.4.4 goes ?

Ben.


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Bogus interrupt -1

[Jiri Masik]

Michel Dänzer <michdaen@iiic.ethz.ch> writes:

> Benjamin Herrenschmidt wrote:
> >
> > >I have tons of messages like this in my logs:
> > >
> > >Jan 18 23:06:54 pismo kernel: Bogus interrupt -1 from PC = c00065d0
> > >
> > >Anyone knows where these (may) come from?
> >
> > Which kernel ?
>
> Paulus' 2.4.0 from the 11th.

I had Paulus' 2.4.0 from the 9th with the same behaviour. It also
occurs with 2.4.0-test13-pre3 and 2.4.1-pre7, 2.4.1-pre8 from bk.

        Jiri

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Bogus interrupt -1

[Michel Dänzer]

I have tons of messages like this in my logs:

Jan 18 23:06:54 pismo kernel: Bogus interrupt -1 from PC = c00065d0

Anyone knows where these (may) come from?


Michel


--
Earthling Michel Dänzer (MrCooper)    \   Debian GNU/Linux (powerpc) developer
CS student, Free Software enthusiast   \        XFree86 and DRI project member

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Bogus interrupt -1

[Benjamin Herrenschmidt]

>I have tons of messages like this in my logs:
>
>Jan 18 23:06:54 pismo kernel: Bogus interrupt -1 from PC = c00065d0
>
>Anyone knows where these (may) come from?

Which kernel ?

We do get tons of "short" interrupts on the Pismo and the iBook. I don't
know why exactly. I used to filter out the error message, it looks like
someone put it back.

Ben.


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Bogus interrupt -1

[Michel Dänzer]

Benjamin Herrenschmidt wrote:
>
> >I have tons of messages like this in my logs:
> >
> >Jan 18 23:06:54 pismo kernel: Bogus interrupt -1 from PC = c00065d0
> >
> >Anyone knows where these (may) come from?
>
> Which kernel ?

Paulus' 2.4.0 from the 11th.


Michel, rsyncing


--
Earthling Michel Dänzer (MrCooper)    \   Debian GNU/Linux (powerpc) developer
CS student, Free Software enthusiast   \        XFree86 and DRI project member

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/