Problem requesting IRQ with 2.4.2 kernel

Problem requesting IRQ with 2.4.2 kernel

[xav]

Hi !

I'm using a clone of the bk tree to compile a 2.4.2 kernel.
I have also downloaded the code from Gabriel Paubert for the
pci2vme chip from TUNDRA.

Before with a 2.4.0 kernel I succesfully inmod the code.
But know I have the following error :
Unable to grab interrupt 16 for VME !
Trying to free free IRQ16

And if I have a look at /proc/pci I have the following :
PCI devices found:
  Bus  0, device   0, function  0:
    Host bridge: Motorola Raven (rev 5).
      IRQ 16.
      I/O at 0x0 [0xffffffff].
      Non-prefetchable 32 bit memory at 0xfc000000 [0xfc03ffff].
  Bus  0, device  11, function  0:
    ISA bridge: Symphony Labs W83C553 (rev 16).
      IRQ 16.
  Bus  0, device  11, function  1:
    IDE interface: Symphony Labs SL82c105 (rev 5).
      IRQ 16.
      Master Capable.  No bursts.  Min Gnt=2.Max Lat=40.
      I/O at 0xffffff0 [0xffffff7].
      I/O at 0xfffffec [0xfffffef].
      I/O at 0xfffffe0 [0xfffffe7].
      I/O at 0xfffffdc [0xfffffdf].
      I/O at 0xfffffc0 [0xfffffcf].
      I/O at 0xfffffb0 [0xfffffbf].
  Bus  0, device  13, function  0:
    Bridge: Tundra Semiconductor Corp. CA91C042 [Universe]
(rev1).
      IRQ 16.
      Master Capable.  No bursts.  Min Gnt=3.
      Non-prefetchable 32 bit memory at 0xeffef000 [0xeffeffff].
      I/O at 0xfffe000 [0xfffefff].


How can avoid all this pci board to have the same IRQ ?
Is this an openPic related problem ?

O+ xavier

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

Re: Problem requesting IRQ with 2.4.2 kernel

[Matt Porter]

On Tue, Feb 27, 2001 at 05:56:53PM +0100, xav wrote:
>
> Hi !
>
> I'm using a clone of the bk tree to compile a 2.4.2 kernel.
> I have also downloaded the code from Gabriel Paubert for the
> pci2vme chip from TUNDRA.

Which bk tree? linuxppc_2_4 or linuxppc_2_5?

> Before with a 2.4.0 kernel I succesfully inmod the code.
> But know I have the following error :
> Unable to grab interrupt 16 for VME !
> Trying to free free IRQ16

Look at Genesis2_pci_IRQ_map in prep_pci.c.  Unassigned interrupts
have zero which looks like IRQ 16 after the openpic offset is added.
This is actually the MPIC input for the 8259 cascade so anything
trying to use it will generate unhappy messages.

> And if I have a look at /proc/pci I have the following :
> PCI devices found:
>   Bus  0, device   0, function  0:
>     Host bridge: Motorola Raven (rev 5).
>       IRQ 16.
>       I/O at 0x0 [0xffffffff].
>       Non-prefetchable 32 bit memory at 0xfc000000 [0xfc03ffff].
>   Bus  0, device  11, function  0:
>     ISA bridge: Symphony Labs W83C553 (rev 16).
>       IRQ 16.
>   Bus  0, device  11, function  1:
>     IDE interface: Symphony Labs SL82c105 (rev 5).
>       IRQ 16.
>       Master Capable.  No bursts.  Min Gnt=2.Max Lat=40.
>       I/O at 0xffffff0 [0xffffff7].
>       I/O at 0xfffffec [0xfffffef].
>       I/O at 0xfffffe0 [0xfffffe7].
>       I/O at 0xfffffdc [0xfffffdf].
>       I/O at 0xfffffc0 [0xfffffcf].
>       I/O at 0xfffffb0 [0xfffffbf].
>   Bus  0, device  13, function  0:
>     Bridge: Tundra Semiconductor Corp. CA91C042 [Universe]
> (rev1).
>       IRQ 16.
>       Master Capable.  No bursts.  Min Gnt=3.
>       Non-prefetchable 32 bit memory at 0xeffef000 [0xeffeffff].
>       I/O at 0xfffe000 [0xfffefff].
>
>
> How can avoid all this pci board to have the same IRQ ?
> Is this an openPic related problem ?

Interrupt routing problem.  IIRC, put a 5 for the interrupt line in
the entry for Slot 13, Universe II LINT#0.  That should get you
going.  Check the 2300 programmer's reference if that interrupt
 doesn't work for you.  If the driver is using more LINT inputs than
that then you'll have to fix the routing as appropriate.  It can all
be rerouted via the Universe II but that's another game.

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

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

Re: Problem requesting IRQ with 2.4.2 kernel

[xav]

Matt Porter wrote:

> On Tue, Feb 27, 2001 at 05:56:53PM +0100, xav wrote:
> >
> > Hi !
> >
> > I'm using a clone of the bk tree to compile a 2.4.2 kernel.
> > I have also downloaded the code from Gabriel Paubert for the
> > pci2vme chip from TUNDRA.
>
> Which bk tree? linuxppc_2_4 or linuxppc_2_5?

linuxppc_2_4

> Interrupt routing problem.  IIRC, put a 5 for the interrupt line in
> the entry for Slot 13, Universe II LINT#0.  That should get you
> going.  Check the 2300 programmer's reference if that interrupt
>  doesn't work for you.  If the driver is using more LINT inputs than
> that then you'll have to fix the routing as appropriate.  It can all
> be rerouted via the Universe II but that's another game.

After I did it, I have the following :

00:0d.0 Bridge: Tundra Semiconductor Corp. CA91C042 [Universe] (rev 01)
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop-
ParErr- Stepping- SERR- FastB2B-
        Status: Cap- 66Mhz- UDF- FastB2B- ParErr- DEVSEL=medium >TAbort-
<TAbort- <MAbort- >SERR- <PERR-
        Latency: 0 (750ns min)
        Interrupt: pin A routed to IRQ 21
        Region 0: Memory at effef000 (32-bit, non-prefetchable)
[size=4K]
        Region 1: I/O ports at fffe000 [size=4K]

No more module problem ! Thanks a lot...

--
33 (0) 1 69 15 79 59
De chacun selon ses forces, a chacun selon ses besoins


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

Re: Problem requesting IRQ with 2.4.2 kernel

[Gabriel Paubert]

On Tue, 27 Feb 2001, Matt Porter wrote:

> >   Bus  0, device  13, function  0:
> >     Bridge: Tundra Semiconductor Corp. CA91C042 [Universe]
> > (rev1).
> >       IRQ 16.
> >       Master Capable.  No bursts.  Min Gnt=3.
> >       Non-prefetchable 32 bit memory at 0xeffef000 [0xeffeffff].
> >       I/O at 0xfffe000 [0xfffefff].
> >
> >
> > How can avoid all this pci board to have the same IRQ ?
> > Is this an openPic related problem ?
>
> Interrupt routing problem.  IIRC, put a 5 for the interrupt line in
> the entry for Slot 13, Universe II LINT#0.  That should get you
> going.  Check the 2300 programmer's reference if that interrupt
>  doesn't work for you.  If the driver is using more LINT inputs than
> that then you'll have to fix the routing as appropriate.  It can all
> be rerouted via the Universe II but that's another game.

No problem with this. I decided to use only one interrupt line in the
Universe driver. Everything is routed to LINT#0 (DMA and VIRQ#1 to VIRQ#7
are the only ones used right now, but I route the others to LINT#0 too).

	Gabriel.


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

Re: Problem requesting IRQ with 2.4.2 kernel

[xav]

Sorry I've forgot a few information:

I'm using a MVME2303 Board (ppc603)

--
33 (0) 1 69 15 79 59
De chacun selon ses forces, a chacun selon ses besoins


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

Re: Problem requesting IRQ with 2.4.2 kernel

[Konstantin Olchanski]

> Date: Tue, 27 Feb 2001 17:56:53 +0100
> From: xav <grave@ipno.in2p3.fr>
> Subject: Problem requesting IRQ with 2.4.2 kernel
>
> Hi !
>
> I'm using a clone of the bk tree to compile a 2.4.2 kernel.
> I have also downloaded the code from Gabriel Paubert for the
> pci2vme chip from TUNDRA.
>


I see the same problem- on my kernels (older bitkeeper 2.4 tree)
pdev->irq comes out as zero.

By experimenting, I determined (and hardwired into the driver)
a working Universe irq on the MVME230x: "11".

I do not know how to kludge the "prep" routing tables- none of the
numbers in those tables match the board and vxworks BSP documentation.

I guess this will be ultimately fixed by Matt Porter's new MVME support
in the 2.5 tree.


--
Konstantin Olchanski
Physics Department, Brookhaven National Laboratory, Long Island, New York
olchansk@bnl.gov

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