kernel-3.15 series fail to load gnome desktop

poma pomidorabelisima at gmail.com
Thu Apr 17 04:19:06 UTC 2014


On 17.04.2014 04:34, Clyde E. Kunkel wrote:
> On Thu, 17 Apr 2014 02:28:11 +0200
> poma <pomidorabelisima at gmail.com> wrote:
> 
>> On 17.04.2014 02:11, Adam Williamson wrote:
>>> On Wed, 2014-04-16 at 17:00 -0700, Adam Williamson wrote:
>>>> On Wed, 2014-04-16 at 17:41 -0600, Pete Travis wrote:
>>>>> On Apr 16, 2014 9:12 AM, "Clyde E. Kunkel"
>>>>> <clydekunkel7734 at verizon.net> wrote:
>>>>>>
>>>>>> With all of the kernels in the rawhide 3.15 series so far the
>>>>>> gnome desktop fails to load.  Kernels in the 3.14 series are
>>>>>> fine.
>>>>>>
>>>>>> All I can see is that the X server seems to start (ps -A | grep
>>>>>> Xorg returns a PID), but the Xorg.0.log is empty.  No obvious
>>>>>> failures in journalctl -xb.  I see gdm is started, but at this
>>>>>> point lots of disk activity but then nothing.  c-a-f2 gets me to
>>>>>> tty and I can login.  Top looks normal for a non-gnome session.
>>>>>> Killing X sometimes brings up the GDM login, but after entering
>>>>>> password, just a blank screen and the Xorg log file is still
>>>>>> empty.
>>>>>>
>>>>>> Have tried every 3.15 kernel so far without luck.
>>>>>>
>>>>>> I don't see any obvious bz, and would be happy to enter one, but
>>>>>> don't know what information should be included.
>>>>>>
>>>>>> Any help appreciated.
>>>>>>
>>>>>> TIA
>>>>>> --
>>>>>>
>>>>>
>>>>> I haven't had gdm display anything for the last several kernel
>>>>> updates , but I'm not convinced the kernel is at fault as older
>>>>> ones stopped gdm from working too.
>>>>>
>>>>> GDM *thinks* it is working - `journalctl -u gdm` is where you can
>>>>> find the Xorg log output these days - but the vterm is black.
>>>>> Lightdm and sddm display, though poorly, and neither produce a
>>>>> working gnome session.
>>>>>
>>>>> No one thing I've poked at so far appeared to be the cause, so I
>>>>> haven't complained. Good to know it is not just me ;)
>>>>
>>>> Have you folks tried booting with enforcing=0 , just as a shot in
>>>> the dark?
>>>
>>> Hum, actually - it looks like on my tablet, dropping 'rhgb quiet'
>>> from the cmdline helps (no 'enforcing=0' needed). Does that apply
>>> to others?
>>>
>>
>> # systemctl stop plymouth-halt plymouth-kexec plymouth-poweroff
>> plymouth-quit-wait plymouth-quit plymouth-read-write plymouth-reboot
>> plymouth-start plymouth-switch-root systemd-ask-password-plymouth
>>
>> # systemctl mask plymouth-halt plymouth-kexec plymouth-poweroff
>> plymouth-quit-wait plymouth-quit plymouth-read-write plymouth-reboot
>> plymouth-start plymouth-switch-root systemd-ask-password-plymouth
>>
>> # ll /etc/systemd/system/*plymouth*
>>
>> # vi /etc/dracut.conf.d/omit_dracut-module-plymouth.conf
>> omit_dracutmodules+=" plymouth "
>>
>> # dracut -f -v
>>
>> Will set you free.
>>
>>
>> poma
>>
>>
> 
> Why? What does plymouth have to do with the problem?  I don't have RHGB
> or quiet in the cmdline.
> 

You'll learn eventually, don't worry. :)


poma




More information about the test mailing list