Re: [PATCH] update MMConfig patches w/915 support

[Olivier Galibert <galibert@pobox.com>]

On Sun, Jan 07, 2007 at 11:42:09AM -0800, Jesse Barnes wrote:
> This patch updates Oliver's MMConfig bridge detection patches with support
> for 915G bridges.  It seems to work ok on my 915GM laptop.

Looks ok to me.


> I also tried adding 965 support, but it doesn't work (at least not on my
> G965 box).  When I enable MMConfig support when the register value is
> 0xf00000003 (should be a 256M enabled window at 0xf0000000) the box hangs
> at boot, so I'm not sure what I'm doing wrong...
> 
> The routines could probably be consolidated into a single probe_intel_9xx
> routine or something, but I really looked at that yet (though there are
> many similarities between the 91[05], 945 and 965 families, they may not
> be enough that the code would actually be simpler if shared.

The individual functions are so simple, it's probably way better for
maintainance simplicity to keep them separate, at least for now.


> +	pci_conf1_read(0, 0, PCI_DEVFN(0,0), 0x48, 4, &pciexbar);
> +
> +	/* No enable bit or size field, so assume 256M range is enabled. */
> +	len = 0x10000000U;
> +	pci_mmcfg_config_num = 1;
> +
> +	pci_mmcfg_config = kzalloc(sizeof(pci_mmcfg_config[0]), GFP_KERNEL);
> +	pci_mmcfg_config[0].base_address = pciexbar;

Hmmm, I'd mask out the reserved bits if I were you.  Paranoia :-)

  OG.