Hi Ard, > But the fact remains that we are going about this the wrong way. > Whether a graphics card decodes legacy VGA ranges or not has *nothing* > to do with whether or not it is in fact the primary device on a > non-x86 system, and so I still think the VGA arbiter should be omitted > entirely for such platforms, and Xorg should be fixed instead.
OK, I see where you're coming from. I've been trying to keep my changes small as I don't want to end up on the hook for the almost limitless range of problems that changing this sort of code can have, but I do take your point that it's a bit of an ugly hack of a solution. Say we were to change Xorg instead. What would correct Xorg behaviour look like? Xorg would need to honour the boot_vga file if it existed so as not to break x86, etc. So your proposed Xorg - if it couldn't find a default card that way, and if there was no helpful config file info, would arbitrarily pick a card that has an Xorg driver? In other words, much like the proposed kernel approach but at a different level of the stack? Are there other graphical applications we care about (other than Xorg) that would need to be patched? I'm happy to do the Xorg patch, but I don't know if anything other than Xorg keys off the boot_vga file. I'm not fundamentally opposed to this approach if the Xorg community is happy with it, the kernel community is happy with it, and no-one expects me to provide patches to any other user-space applications that depend on boot_vga. Regards, Daniel
