You are not logged in.
Hello
I just had the problem that after the Update to Kernel 3.12 my TrackPoint/TouchPad and Keyboard of my Thinkpad X220 didn't work and so I was not able to login with gdm.
I managed to downgrade the kernel from cache back to 3.11.
But since sooner or later I would have to update again I thought I should find out what caused the problem.
Unfortunatly I have no idea which logs to check and what to do with the things I will find. So I was hoping that somebody could tell me which files are needed and maybe check the, after I paste them somewhere.
thanks in advance.
Offline
What do you see in your dmesg output relating to errors or warnings when booting into version 3.12? You will need to ssh into the box if your keyboard doesn't work to check obviously. Related question: are you building a module manually for the trackpoint?
CPU-optimized Linux-ck packages @ Repo-ck • AUR packages • Zsh and other configs
Offline
I never used ssh and would have to set it up first.
Would it work the same way if I install 3.12 again and boot with kernel option "systemd.unit=multi-user.target". Got that from a user in the arch irc channel. Booted without a gui.
Related question: are you building a module manually for the trackpoint?
I did nothing special for the trackpoint, touchpad or keyboard. It always worked out of the box. So no.
Offline
Would it work the same way if I install 3.12 again and boot with kernel option "systemd.unit=multi-user.target".
It might work but if it does not, then you can rule out something under X as the cause. I do not think X is to blame but you are welcome to try. Using ssh from another machine is your next option.
CPU-optimized Linux-ck packages @ Repo-ck • AUR packages • Zsh and other configs
Offline
Can you try to wait for like 1 or 2 minutes to see if they come up? I had a similar problem with my k750 from logitech. On some kernels (don't know why), it takes time for the usb to fully initialize and detect the keyboard.
Offline
Can you try to wait for like 1 or 2 minutes to see if they come up? I had a similar problem with my k750 from logitech. On some kernels (don't know why), it takes time for the usb to fully initialize and detect the keyboard.
Will do that.
Also, after downgrading the kernel back it worked on the first boot. But this morning, after booting up i can se the cursor but it keeps going back into the upper right corner -- as if it is magnetic.
After logging in the mouse cursors behavior is normal again. Maybe it actually is a gnome problem?
Offline
Just installed the 3.12 kernel again. Saw this time, that it shows me a warning when installing, that I seem to have a boot partition which is not mounted. That is correct... it's a uefi partition.
[alex@X220 ~]$ sudo fdisk -l
Disk /dev/sda: 119,2 GiB, 128035676160 bytes, 250069680 sectors
Units: Sektoren of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: gpt
Disk identifier: 0CECEF99-D34B-4793-97FA-D02082BD7A50
Gerät Start Ende Size Typ
/dev/sda1 2048 1050623 512M EFI System
/dev/sda2 1050624 233291775 110,8G Linux filesystem
/dev/sda3 233291776 250068991 8G Linux swap
That it is not mounted is also correct, I have set the noauto flag in the fstab for this partition:
#
# /etc/fstab: static file system information
#
# <file system> <dir> <type> <options> <dump> <pass>
# /dev/sda2
UUID=7e2eb40a-0592-4404-b7c7-ba3afc900e0d / ext4 defaults,noatime,discard,rw,relatime,data=ordered 0 1
# /dev/sda1
UUID=4933-7757 /boot vfat defaults,noauto,rw,relatime,fmask=0022,dmask=0022,codepage=437,iocharset=iso8859-1,shortname=mixed,errors=remount-ro 0 2
# /dev/sda3
#UUID=267b22e1-6f59-4855-a6e9-514c248f119b none swap defaults 0 0
If I remove the noauto flag my notebook won't boot at all.
I also uploaded dmesg output and Xorg log bevore downgrading again.
dmesg: http://pastie.org/8482076
Xorg: http://pastie.org/8482079
Waiting 5 minutes didn't show the cursor.
Hope somebody can help.
Offline
You have to update your kernel after you mount the boot partition. So, first mount your device:
sudo mount /dev/sda1 /boot
Then (re)install, your kernel:
sudo pacman -Syu
sudo pacman -S linux
Remove noauto in your fstab file and reboot.
I hope that helps!
Offline
So I discovered that if you mount the ESP at /boot, systemd wants to set it up as an x-systemd.automount,noauto situation (via a generator). This would seem a great idea, as there is no real reason to have your /boot mounted unless you are going to access it for some reason. But what ends up happening in Arch is no bueno.
Because we don't keep old cruft around, kernels and their supporting modules replace the old rather than being installed side-by-side. So because the ESP /boot is not mounted, the vfat module is also not loaded yet. So when you update the kernel and it has not had a reason beforehand to mount the ESP, the modules get replaced. Then it can't actually mount your ESP because there is no longer a matching vfat module to load.
So now I mount it at /boot/efi, partially for the issue above, and partially because I want my kernel to be snapshotted with the rest of my rootfs on my btrfs filesystem.
Recently though I did run into an issue where the kernel and initramfs was not copying itself over from /boot to /boot/efi/EFI/arch. I was using a systemd.path to trigger a oneshot service. I had it set to trigger on "PathChanged=/boot/initramfs-linux-fallback.img". For some reason this no longer works, but changing it to "PathModified" seems to work as it did before. I am not sure if this constitutes a bug in systemd or Arch or if it just might be my improper understanding of which Path* I should have been using. Nor do I know how to debug such a thing to figure it out... oh well.
Edit: BTW, there are a couple other potential solutions for the ESP at /boot as well. On is to explicitly load the vfat module via /etc/modules-load.d. Otherwise, one could also simply mask 'boot.automount' so that the generated unit will not take effect, and the systemd-fstab-generator's boot.mount (created from your fstab) will be applied right away.
Last edited by WonderWoofy (2013-11-27 16:02:01)
Offline
I am having the same problem on a UEFI. The only thing that is displayed is
Loading linux-ck kernel....
Loading initial ramdisk....
and that is it!
Most of the time I use that box as topless so I didn't notice exactly since when it is in that state. I would say that it coincide with 3.12 upgrade but I am not sure.
The funny thing is that beside not having boot msgs and the login prompt, the system is normally functionning. I do have a ssh server and if I login with ssh, I can launch X and start using it with the my local keyboard and mouse.
Only the login prompt is gone. I think that it could be related to systemd but I never have problem with the console before so I have no clue what to look to fix it.
Offline
Ok, I think that I have solved my problem. I'm not sure anymore that it is the same problem than the one from the original poster.
I have ommited to say that I have a custom kernel that I have compiled myself. It turns out that 3.12 has introduced several FB related config options. By comparing these config options of my kernel with the ones used by Arch stock kernel, one diff between my kernel and Arch kernel was
CONFIG_X86_SYSFB that was set on my kernel and not in Arch. I did turn off that option in my kernel and on the next reboot my console came back.
I need to dig to understand what are the 3.12 FB related new features.
What is doing X86_SYSFB ? What is the SIMPLEFB? How do they fit with existing FB drivers?
Offline