Re: RE : MPC5200Lite and PCI card IRQ

[Benjamin Herrenschmidt <benh@kernel.crashing.org>]

On Wed, 2004-06-02 at 09:21, Bertrand Baudet wrote:
> Well,
>
> I have the answer to my question:
> Yes, the IRQ is automatically unmasked with the IRQ# from the PCI conf.
>
> In fact the INT A of the PCI is linked to the IRQ0 of the MPC5200.
>
> I'm trying to add a SATA PCI card (with Sii3112 chipset).
>
> But it looks like in ide-probe.c if hwif->irq is set to 0, the IRQ is not
> set. Unfortunately, in the case of the MPC5200, the IRQ0 has a value of 0.
> Which disable the device.

This is a none bug of some old drivers. 0 should be replaced by the
arch-defined NO_IRQ constant, which on ppc can conveniently by set
to -1. In the meantime, just hack around :)

> After commenting out:
> /*
> if (!hwif->irq) {
> 	if (!(hwif->irq = ide_default_irq(hwif->io_ports[IDE_DATA_OFFSET])))
>       {
>               printk("%s: DISABLED, NO IRQ\n", hwif->name);
>               return (hwif->present = 0);
>       }
> }
> */
> the SATA card works properly.
>
>
> Can someone correct me if I'm wrong: setting 0 to hwif->irq means that
> there is no IRQ# defined. Right?

Yes, and it's an incorrect assumption of the IDE code.
>
>
> Regards,
> Bertrand
>
> -----Message d'origine-----
> De : owner-linuxppc-dev@lists.linuxppc.org [mailto:owner-linuxppc-dev@lists.linuxppc.org] De la part de Bertrand Baudet
> Envoyé : mercredi 2 juin 2004 11:20
> À : linuxppc-dev@lists.linuxppc.org
> Objet : MPC5200Lite and PCI card IRQ
>
>
> Hi,
>
> I'm trying to plug a PCI card on my MPC5200Lite board but I have some
> trouble enabling the IRQ.
>
> >From U-Boot, I'm setting the interruption line to IRQ 49. By default,
> the interruption line is set to 0.
>
> But when the driver is loaded, it doesn't look like the IRQ is unmasked.
> The IRQ# is reported by the driver but when I look at /proc/interrupt,
> I have:
>            CPU0
> 49:          0  MPC5xxx Edge
>
>
> What needs to be done in order to have the IRQ automatically enabled for
> the
> PCI card installed? (using linux kernel from DENX devel cvs)
>
> Actually, how does it work? When the PCI driver probes the devices, will
> it
> enable the IRQ set for the PCI device? Or is it the driver for this PCI
> card
> that needs to enable the IRQ?
>
> I tried to enable the IRQ once the drive reports the IRQ# by calling
> enable_irq directly, but I'm getting a "enable_irq(49) unbalanced"
>
>
>
>
> Regards,
> Bertrand
>
>
--
Benjamin Herrenschmidt <benh@kernel.crashing.org>


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