Re: PCI, sort of working now...

[Geert U. <ge...@li...>]

On Sat, 10 Feb 2001, Glenn Hisdal wrote:
> PCI: Resource ef000000-ef01ffff (f=200, d=0, p=0)
> PCI: Cannot allocate resource region 0 of device 00:01.0
> PCI: Resource e0000000-e07fffff (f=200, d=0, p=0)
> PCI: Cannot allocate resource region 1 of device 00:01.0
> PCI: Resource e1000000-e17fffff (f=200, d=0, p=0)
> PCI: Cannot allocate resource region 2 of device 00:01.0

This usually means that those regions are not inside the bus resource regions
(or the bus doesn't have any bus resources filled in).

> PCI: Error while updating region 00:01.0/1 (10800000 != e1000000)
>   got res[10800000:10ffffff] for resource 1 of Texas Instruments TVP4020 [Permedia 2]
> apus_pci: apus_pcibios_write_config_dword: writelb: 0x0, d: 0x1, f: 0x0, o: 0x18, v: 0x11000000
> apus_pci: apus_pcibios_read_config_dword: readlb: 0x0, d: 0x1, f: 0x0, o: 0x18, v: 0x11000000
>   got res[11000000:117fffff] for resource 2 of Texas Instruments TVP4020 [Permedia 2]

And thus the PCI assignment code tries to assign new regions from the bus
resource, or from pci_mem if the bus resources are empty (it looks like they
are empty, 0x10000000 is the default start of pci mem).

I guess the errors come from the hardcoded CyberVision PCI bridge?

On Sun, 11 Feb 2001, Glenn Hisdal wrote:
> On 11-Feb-01, Glenn wrote:
> > On 11-Feb-01, Michel wrote:
> >> Glenn Hisdal wrote:
> 
> >>> PCI: Resource ef000000-ef01ffff (f=200, d=0, p=0)
> >>> PCI: Cannot allocate resource region 0 of device 00:01.0
> >>> PCI: Resource e0000000-e07fffff (f=200, d=0, p=0)
> >>> PCI: Cannot allocate resource region 1 of device 00:01.0
> >>> PCI: Resource e1000000-e17fffff (f=200, d=0, p=0)
> >>> PCI: Cannot allocate resource region 2 of device 00:01.0
> 
> > I have managed to remove these errors.
> > I'll try to make a diff...
> 
> ok. there should be a diff attached to this mail (i hope).
> I don't know if it's any good, but it stopped complaining about the resource
> allocation.
> I just did a little cut and paste from chrp_pci.c
> 
> - glenn
> 

> Index: 2.3/arch/ppc/kernel/apus_pci.c
> ===================================================================
> RCS file: /cvsroot/linux-apus/2.3/arch/ppc/kernel/apus_pci.c,v
> retrieving revision 1.6
> diff -r1.6 apus_pci.c
> 106a107
> > static struct resource pci_mem = { "B/CVisionPPC PCI mem", CVPPC_FB_APERTURE_ONE, CVPPC_PCI_CONFIG, IORESOURCE_MEM };
> 115a117,119
> > 
> >         request_resource(&iomem_resource, &pci_mem);
> >     	printk("%s: PCI mem resource requested\n", __FUNCTION__);
> 120,123c124
> < 	struct resource pci_mem = { "B/CVisionPPC PCI mem", CVPPC_FB_APERTURE_ONE, CVPPC_PCI_CONFIG, IORESOURCE_MEM };
> < 
> < 	request_resource(&iomem_resource, &pci_mem);
> <     	printk("%s: PCI mem resource requested\n", __FUNCTION__);
> ---
> >         bus->resource[1] = &pci_mem;

This looks fine to me. BTW, please use `diff -u' next time (or add the line

    diff	-u

to your ~/.cvsrc to make it the default).

Now please install pciutils and post the output of `lspci -vv'. Thx!

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- ge...@li...

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
							    -- Linus Torvalds