Re: [coreboot] how to prevent legacy resource conflict with multiple VGA cards

[Scott Duplichan]

]> One problem is that the in-box ATI driver for the UMA graphics is
]> unstable.
]Even with the factory BIOS?

The reference BIOS works fine, so this is a coreboot problem. I have
added missing family 10h code to the coreboot RS780 GFX initialization,
and enabled HT3 for the link. I have added missing NP attributes to the
frame buffer mapping. But the driver still fails after a few seconds.
There is more debugging to do here. This is the most serious remaining
Win7 problem I know of for RS780/SB700 boards.

]> So to get Win7 installed, I must add a PCI video card. When Win7 is
]> installed with the PCI video card present, device manager reports that the
]> UMA graphics device cannot find enough free resources.
]> 
]> When a reference BIOS is used, the device manager warnings are not present
]> because the BIOS disables the UMA. One reference BIOS disables UMA only by
]> clearing the PCI command register bits for memory and I/O decode. A
]> different
]> reference BIOS skips UMA initialization altogether. I think for now, the
]> command register method might be easiest for coreboot.
]OK
]
]> The problem with the existing coreboot CONFIG_CONSOLE_VGA_MULTI feature is
]> that it can only disable graphics devices that are on an HT link. In the
]> case of UMA, non-AMD or non-HT systems, it doesn't work.
]
]I think the easiest thing to do would be to use the same mechanism to know
]whether or not to disable UMA.  If there is another VGA card added, then the
]VGA bits will be set on bridges that aren't associated with UMA.  The UMA
]device could check the tree for these bits and disable itself, or some more
]code could be added to generically do the same thing.

OK, thanks. I will eventually prototype something for review.

]Thanks,
]Myles