Re: AGP problems.

[Adam K Kirchhoff <adamk@xxxxxxxxxxxx>]

http://memory.visualtech.com/dmesg.txt

Adam

On Wed, 24 Mar 2004, Matthew Dillon wrote:

>     Ok, do the following:
>
>     (1) cvsup and bring your source tree up to date
>
>     (2) Apply this debugging patch
>
>     (3) Build the agp module into the kernel 'device agp'.  agp is not loading
> 	properly via the module loader and I have not quite finished fixing
> 	that.
>
>     (4) When you boot the system, boot in verbose mode (boot -v).
>
> 						-Matt
>
> Index: agp_via.c
> ===================================================================
> RCS file: /cvs/src/sys/dev/agp/agp_via.c,v
> retrieving revision 1.4
> diff -u -r1.4 agp_via.c
> --- agp_via.c	9 Dec 2003 19:40:56 -0000	1.4
> +++ agp_via.c	24 Mar 2004 20:54:35 -0000
> @@ -55,12 +55,20 @@
>  static const char*
>  agp_via_match(device_t dev)
>  {
> +	printf("agp_via_match: %08x/%08x %08x/%08x vendor %04x\n",
> +		pci_get_class(dev), PCIC_BRIDGE,
> +		pci_get_subclass(dev), PCIS_BRIDGE_HOST,
> +		pci_get_vendor(dev)
> +	);
> +
>  	if (pci_get_class(dev) != PCIC_BRIDGE
>  	    || pci_get_subclass(dev) != PCIS_BRIDGE_HOST)
>  		return NULL;
>
> -	if (agp_find_caps(dev) == 0)
> +	if (agp_find_caps(dev) == 0) {
> +		printf("agp_via_match: agp_find_caps() failed\n");
>  		return NULL;
> +	}
>
>  	switch (pci_get_devid(dev)) {
>  	case 0x03051106:
> @@ -106,11 +114,13 @@
>  	int error;
>
>  	error = agp_generic_attach(dev);
> +	printf("agp_via_attach: generic attach code %d\n", error);
>  	if (error)
>  		return error;
>
>  	sc->initial_aperture = AGP_GET_APERTURE(dev);
>  	if (sc->initial_aperture == 0) {
> +		printf("agp_via_attach: aperture problem\n", error);
>  		device_printf(dev, "bad initial aperture size, disabling\n");
>  		return ENXIO;
>  	}
>