boot hangs without error message - what now?

commander-keen · n00b Joined: 09 Oct 2013 Posts: 47 Location: Europe

Hello,

I have a Gentoo installation that crashes on boot. [Kernel: 3.14.19]
Only problem is, I do not have an error message. However the last log comes from Nouveau:

NeddySeagoon · Posted: Fri Oct 31, 2014 2:07 pm Post subject:

commander-keen,

Two things may have happened.
The boot may have contined normally but with no console output. In this case you can ssh in from another system and post dmesg.
dmesg will be helpful.

From your image, you are using the nouveau driver but not its free framebuffer console. There is no tux logo as a framebuffer provides.

dmesg (from the failed boot) and your kernel .config both on a pastebin would be useful but you may not be able to get dmesg
The output of lspci may be useful too.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.

commander-keen · n00b Joined: 09 Oct 2013 Posts: 47 Location: Europe

Hi Neddy

thanks a lot for your help.
I booted the machine connected to a router (with a dhcp server) by ethernet. But the computer did not show up in the network so I guess it really gets stuck on boot.

I also checked the framebuffer console in the kernel config. Looks fine to me. :-(

Anyways here is my kernel config:
http://students.fim.uni-passau.de/~schieder/.config

And the output of lspci:

NeddySeagoon · Posted: Fri Oct 31, 2014 4:41 pm Post subject:

commander-keen,

You may not have both

commander-keen · n00b Joined: 09 Oct 2013 Posts: 47 Location: Europe

Hi,

thank you for pointing out the conflicting Framebuffer drivers. I disabled the NVIDIA support.
But no luck so far, the problem remains. The boot still crashes at the exact same part, same output. :-(

Max

NeddySeagoon · Posted: Fri Oct 31, 2014 5:57 pm Post subject:

commander-keen,

OK, I was lazy - that looked so obvious. I'll look some more. Meanwhile, get into your chroot and check some timestamps. Just to make sure you really are trying your new kernel. It won't be the first time a user has solved their problem but not known it because they did not boot the new kernel :)

/usr/src/linux/.config should be the oldest file. It drives the build
/usr/src/linux/arch/x86/boot/bzImage should be newer, if not, it did not come from the .config file
/boot/<kernel_name> should have the same timestamp as bzImage.

Lastly, your bootloader should be loading /boot/<kernel_name>, have a look at its configuration file.
lilo users need to remember to run /sbin/lilo
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.

DONAHUE · Posted: Fri Oct 31, 2014 6:04 pm Post subject:

you might try enabling <*> /dev/agpgart (AGP Support) --->
I don't have mcp79 chipset but have not had luck with any chipset without enabling agpgart including a much older nvidia chipset.
_________________
Defund the FCC.

Ant P. · Watchman Joined: 18 Apr 2009 Posts: 6920

Try also enabling CONFIG_X86_SYSFB and CONFIG_FB_SIMPLE, they may increase your chances of getting *some* output.

DONAHUE · Posted: Fri Oct 31, 2014 9:08 pm Post subject:

rather than adding I'd suggest reducing:

commander-keen · n00b Joined: 09 Oct 2013 Posts: 47 Location: Europe

Hi everyone,

first of all thank you all very much for your ideas. Today I once more experimented with lots of different kernel configurations following your hints. But no matter what I tried I always ended with the kernel crashing at the very same step. By comparing to a proper nouveau bootup on a different machine I figured the crash occurs shortly before turning on KMS. So I eventually figured I might sidestep the problem rather than solving it. Long story short, staring the kernel with the argument