On Thu, 2013-07-11 at 17:26 -0400, Adam Jackson wrote: > On Thu, 2013-07-11 at 14:12 -0700, Adam Williamson wrote: > > > We probably need to test it more, but the impression I was getting from > > the bug is that we don't necessarily want to simply use 'nomodeset' as > > the 'standard fallback path' for UEFI, that there may be UEFI systems on > > which we might want to do KMS but use a generic driver, for > > instance...is that incorrect? > > I guess it depends which part you're trying to fall back from: > accelerated drawing, or any kernel code entirely. The conservative > thing is saying nomodeset: you get the dumb framebuffer the firmware > hands you and that's it. > > If we thought the modesetting part of KMS worked but the accelerated > drawing didn't, then we might want to instead make the X server pick > xorg-x11-drv-modesetting instead of a native driver. Currently there's > no convenient mechanism for that, but I'm hard pressed to think of a > time when I'd have wanted it. > > There are KMS-but-not-accelerated drivers - cirrus, mgag200, ast - and > those do use the -modesetting X driver already; 'fallback' for them > could only possibly mean nomodeset. So in the interests of symmetry... So if this just boils down to 'fallback mode is whatever you get when you pass nomodeset', all we need to do is ensure the UEFI bootloaders have a 'Basic Graphics Mode' option which appends nomodeset, and drop the xdriver=vesa from the BIOS bootloaders (though it's not terribly important), yes? -- Adam Williamson Fedora QA Community Monkey IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora http://www.happyassassin.net -- test mailing list test@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe: https://admin.fedoraproject.org/mailman/listinfo/test
