On 03/19/2013 01:23 PM, Daniel Baumann wrote: > ok, so the bug isn't in the runtime configuration (live-config) but in > fglrx doesn't work with kms which we'll need to workaround in the > buildtime configuration (live-images) by conditionally adding nomodeset > to bootappend in case non-free was enabled.
Ugh. That is going to be crappy for systems which don't actually require a non-free driver and may behave badly if it is disabled (e.g. intel). Use case: User wants a single image to use on two home systems, one of which requires non-free and one doesn't. probably not all that rare ... Wouldn't it be better to disable modeset for specific chipsets, and wouldn't that be better to do in the package providing the driver that behaves badly with modeset enabled? e.g. intel driver turns it on by providing /etc/modprobe.d/i915-kms.conf which contains: options i915 modeset=1 So, the opposite of that, and with the option only being applied to radeon (and/or nvidia if it turns out it needs it too). Do *all* radeons behave better with modeset disabled, or just some? (and are there any that behave well only with modeset enabled and not with it disabled)? I'm very uneasy about this change without checking it out on a wider range of hardware. Ben -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
