Bogus interrupt

Bogus interrupt

[Stefano Coluccini]

Hi everyone,
  I'm trying to run kernel 2.5 on Motorola MVME5100 and, after the sti
istruction in start_kernel, the kernel hangs due to a spurious interrupt.
(Bogus interrupt -1)
Is this a well-known problem also in other ports ? or have anyone of you a
suggestion to solve this ?

Thaks in advance.
Stefano.


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

Re: Bogus interrupt

[Matt Porter]

On Mon, Feb 26, 2001 at 12:54:19PM +0100, Stefano Coluccini wrote:
>
> Hi everyone,
>   I'm trying to run kernel 2.5 on Motorola MVME5100 and, after the sti
> istruction in start_kernel, the kernel hangs due to a spurious interrupt.
> (Bogus interrupt -1)
> Is this a well-known problem also in other ports ? or have anyone of you a
> suggestion to solve this ?

If you are using the default openpic initsenses in prep_pci.c then you
might see this problem since MPIC input 1 now has the 550 UART.  You'll
need to change the initsenses since the edge triggered ECC interrupt
used to be on input 1 on the older boards.

--
Matt Porter
MontaVista Software, Inc.
mporter@home.com

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

RE: Bogus interrupt

[Stefano Coluccini]

> > Hi everyone,
> >   I'm trying to run kernel 2.5 on Motorola MVME5100 and, after the sti
> > istruction in start_kernel, the kernel hangs due to a spurious
> interrupt.
> > (Bogus interrupt -1)
> > Is this a well-known problem also in other ports ? or have
> anyone of you a
> > suggestion to solve this ?
>
> If you are using the default openpic initsenses in prep_pci.c then you
> might see this problem since MPIC input 1 now has the 550 UART.  You'll
> need to change the initsenses since the edge triggered ECC interrupt
> used to be on input 1 on the older boards.
>
I've started working on MCPN765 port, since I think MVME5100 is an MCPN765
with a VME Bridge and without an ISA bridge, so I think the senses of
interrupt are correct because the 5100 and the 765 manual page related to
the interrupt senses is exactly the same (or it is a cut&paste error ??? :))
However this problem, like many other I've encountered, are all due to a bad
cross-compiler :(. I start working today with a good environment after a
month of nightmares !

Thank you very much for the help.
Stefano.


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

Re: Bogus interrupt

[Matt Porter]

On Wed, Feb 28, 2001 at 09:29:18AM +0100, Stefano Coluccini wrote:
> > > Hi everyone,
> > >   I'm trying to run kernel 2.5 on Motorola MVME5100 and, after the sti
> > > istruction in start_kernel, the kernel hangs due to a spurious
> > interrupt.
> > > (Bogus interrupt -1)
> > > Is this a well-known problem also in other ports ? or have
> > anyone of you a
> > > suggestion to solve this ?
> >
> > If you are using the default openpic initsenses in prep_pci.c then you
> > might see this problem since MPIC input 1 now has the 550 UART.  You'll
> > need to change the initsenses since the edge triggered ECC interrupt
> > used to be on input 1 on the older boards.
> >
> I've started working on MCPN765 port, since I think MVME5100 is an MCPN765
> with a VME Bridge and without an ISA bridge, so I think the senses of
> interrupt are correct because the 5100 and the 765 manual page related to
> the interrupt senses is exactly the same (or it is a cut&paste error ??? :))
> However this problem, like many other I've encountered, are all due to a bad
> cross-compiler :(. I start working today with a good environment after a
> month of nightmares !

Ok, I finally got time today to powerup the 5100 that's been laying
around here.  I just pushed the port to the linuxppc_2_5 tree.
Hopefully this saves some effort for others.

--
Matt Porter
MontaVista Software, Inc.
mporter@mvista.com

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

RE: Bogus interrupt + little Makefile bug

[Stefano Coluccini]

> Ok, I finally got time today to powerup the 5100 that's been laying
> around here.  I just pushed the port to the linuxppc_2_5 tree.
> Hopefully this saves some effort for others.

Very great work !!!

I have found a little bug in arch/ppc/boot/pp3/Makefile:
the rule to make zImage.initrd miss the -pbp option in the mkprep
invocation.

I've seen that in the make menuconfig, selecting the MVME5100, does not
display the possibility to specify the default kernel command line. Don't
you think this is an useful option ?

Bye.
Stefano.

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

MVME5100 was RE:Bogus interrupt + little Makefile bug

[Matt Porter]

[Followups trimmed to only the embedded list]

On Thu, Mar 01, 2001 at 10:55:33AM +0100, Stefano Coluccini wrote:
> > Ok, I finally got time today to powerup the 5100 that's been laying
> > around here.  I just pushed the port to the linuxppc_2_5 tree.
> > Hopefully this saves some effort for others.
>
> Very great work !!!
>
> I have found a little bug in arch/ppc/boot/pp3/Makefile:
> the rule to make zImage.initrd miss the -pbp option in the mkprep
> invocation.

Ah, thanks.  A typo from the great bootloader reorganization, it seems.

> I've seen that in the make menuconfig, selecting the MVME5100, does not
> display the possibility to specify the default kernel command line. Don't
> you think this is an useful option ?

Oversight.  I just turned it on for a bunch of boards.

--
Matt Porter
MontaVista Software, Inc.
mporter@mvista.com

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

Re: MVME5100 was RE:Bogus interrupt + little Makefile bug

[Tom Rini]

On Thu, Mar 01, 2001 at 08:17:48AM -0700, Matt Porter wrote:
>
> [Followups trimmed to only the embedded list]
>
> On Thu, Mar 01, 2001 at 10:55:33AM +0100, Stefano Coluccini wrote:
> > > Ok, I finally got time today to powerup the 5100 that's been laying
> > > around here.  I just pushed the port to the linuxppc_2_5 tree.
> > > Hopefully this saves some effort for others.
> >
> > Very great work !!!
> >
> > I have found a little bug in arch/ppc/boot/pp3/Makefile:
> > the rule to make zImage.initrd miss the -pbp option in the mkprep
> > invocation.
>
> Ah, thanks.  A typo from the great bootloader reorganization, it seems.

Quite probably.  If people in the know (Matt.. :) could check the other
prep-ish board-dirs to make sure mkprep is invoked properly..

> > I've seen that in the make menuconfig, selecting the MVME5100, does not
> > display the possibility to specify the default kernel command line. Don't
> > you think this is an useful option ?
>
> Oversight.  I just turned it on for a bunch of boards.

...and will "soon" be always an option.  I think it's only useless right now
on CHRP/PMAC stand alone kernels.  I of course need to check this out before I
turn it on, and can probably make them use it too..

--
Tom Rini (TR1265)
http://gate.crashing.org/~trini/

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