Bug#732692: loading radeonkms results in unusable screen
On 06/01/14 22:26, Robert Millan wrote:
> The proper fix (running the actual initialization before sysctl_init) is not
> so easy, as the codepaths are too separate (and this separation is even exposed
> to userland).
Really strangely, if the firmware is not found for my card, it isn't
registering the sysctl. Xorg detects this and falls back to a usable
vesa desktop, fixed at 1280x1024 (no xrandr) cloned on both monitors.
kldstat still shows radeonkms as being 'loaded' in this case.
I've broken networking on my test system which complicates testing and I
couldn't save the kernel messages. I remember messages prefixed by
[drm] about loading microcode, specifically by some ni_* functions which
must be specific to my chipset (Radeon HD 6800 - Northern Islands / Barts).
The behaviour I get sounds fine - it gives me something that works, or
otherwise I merely have to copy the non-free firmware modules to the
right place for KMS to work.
It is most interesting that your HD 2400 / RV610 isn't falling back in
the same way as described above.
Regards,
--
Steven Chamberlain
steven@pyro.eu.org
Reply to: