Andreas Beckmann anbe at debian.org
Thu Dec 17 11:34:45 UTC 2015

On 2015-12-17 12:15, alberto fuentes wrote:
> Thanks for being so detailed!
> On Thu, Dec 17, 2015 at 1:27 AM, Andreas Beckmann <anbe at debian.org> wrote:
>> Take a look at
>> https://wiki.debian.org/KernelModesetting
>> and try booting with "nomodeset"
>> and maybe try disabling any graphics stuff in /etc/default/grub
>> (needs running 'update-grub' after editing)
> I included nomodeset on /etc/default/grub and updated grub afterwards. I
> dont think I have any special configuration there (it doesnt show up while
> running debsumbs -ce nor any other file that seems that might affect). I
> attach my grub file nonetheless
> This didnt do anything

since it has a typo ... so perhaps try again after removing the diversion

You should also try the

>> You can try to forcibly move the nouveau module away:
>> dpkg-divert --rename
>> /lib/modules/4.3.0-1-amd64/kernel/drivers/gpu/drm/nouveau/nouveau.ko
>> (that's the "friendly" way, it tells apt about the rename, and apt will
>> respect
>> it on kernel upgrades (as long as it stays in the same path, i.e. until
>> you upgrade to
>> 4.4 or an ABI bump for 3.3 happens (the ABI version is the '1' in
>> '4.3.0-1-amd64'))
>> To undo this later on use
>> dpkg-divert --remove --rename
>> /lib/modules/4.3.0-1-amd64/kernel/drivers/gpu/drm/nouveau/nouveau.ko
>> Then rebuild the initramfs (update-initramfs -u) just in case ...
>> any maybe while rebooting you find something that complains that it cannot
>> load the nouveau module
>> Thereafter (hopefully) the nvidia module should be loadable.
> This finally did it!
> sddm always starts with a weird resolution and i dont see the user or
> password boxes, but i enter my password blindly, press enters and it logins
> with my usual resolutions (this might have to do with the fact that xrandr
> showed a ghost vga-1 along with my other two dvi-0 and vga-0)
> # journalctl -b and searching for nouveau showed nothing, but in tty1
> outside the logs and before the login prompt I can read the text attached
> in nolog.output
> How that text scape the logs is a good question. I dont have or never had
> raids... not sure why is giving the other error

The kernel has all these things enabled and tries to find them, it falls
safely back to "no raid". So there is no need for special kernels for
raid setups.


More information about the pkg-nvidia-devel mailing list