You are not logged in.
Hi
Every kernel after 3.14.5-1-ARCH (even the newest 3.15.x release) causes boot problems on my Intel NUC 2820
It is not able to boot into the system, not even to the stage where systemd is taking over.
Sometimes the kernel hangs at the AUX port, sometimes at the "usbcore" initialization, you cannot toggle a certain pattern.
I use the newest BIOS (at the moment v. 0034).
Can someone confirm the same behavior?
Any ideas, thoughts?
Thank you!
Cheers,
mijenix
Offline
I'm having a problem booting 3.15 too, although I get to systemd and a prompt. I just downgraded to 3.14 by chrooting with the install usb by following this archwiki: https://wiki.archlinux.org/index.php/Kernel_Panics.
Offline
I'm having a problem booting 3.15 too, although I get to systemd and a prompt. I just downgraded to 3.14 by chrooting with the install usb by following this archwiki: https://wiki.archlinux.org/index.php/Kernel_Panics.
do you also use UEFI only boot on x86_64?
Offline
I can't boot correctly on my L540 too.
The strange thing is that after I got it through systemd (by disabling mount of a vfat partition) the output of uname -r was 3.14. Oh and there were no detected devices (soundcard, network interfaces, video card).
Reverting to 3.14 by booting from a life-disk solved the problem.
Does anyone know what the problem is? I'm kinda inexperienced with kernel upgrade problems.
/e I'm using gummiboot with EFI on x86_64
Last edited by Krypt0n (2014-06-19 14:57:20)
Offline
@mijenix: Yes, UEFI boot only on x86_64.
Offline
Yes, UEFI boot only on x86_64.
I was able to boot till I get a prompt too (at least with some newer >= 3.14.6) but after a few minutes the computer freezed and I had to restart it
@sfabius: Did you test your box with a newer kernel version a little longer or did you switch immediately back to an older kernel?
Last edited by mijenix (2014-06-20 09:38:03)
Offline
I only got to the fallback prompt, pressing ctl-D and then dropping into a root shell. But nothing was mounted and I could mount nothing manually. So at that point I was stuck and went back to previous kernel.
Offline
Quickest way to downgrade the kernel if you have no live system at hand:
Boot the bad kernel with maxcpus=1, then downgrade the kernel: https://wiki.archlinux.org/index.php/Kernel_Panics#Roll_back_to_previous_kernel_version
Offline
And for the future - install linux-lts.
Offline
Solution for me in similar case - removing i915 from MODULES line in /etc/mkinitcpio.conf
(my configuration - lenovo ideapad u330p (Intel(R) Core(TM) i3-4010U CPU, HD 4400 ), gummiboot, x86_64)
Offline
Solution for me in similar case - removing i915 from MODULES line in /etc/mkinitcpio.conf
(my configuration - lenovo ideapad u330p (Intel(R) Core(TM) i3-4010U CPU, HD 4400 ), gummiboot, x86_64)
Does not work for me, still the same problem.
Offline
Should we file a bug report? Or should we do that upstream?
Does someone has any news?
Offline
I have the same problem. Intel claims that the only officially supported OSes are Windows 8.1 and 7, so I don't know how far we'll get upstream.
Offline
I have the same problem. Intel claims that the only officially supported OSes are Windows 8.1 and 7, so I don't know how far we'll get upstream.
Yes, I created a thread in the Intel Forum as well!
Offline
I can confirm that I can boot with either "maxcpus=1" or "nosmp", but without it the boot fails completely, with no error reported even with all kernel logging enabled. Why would SMP cause a boot failure without an error message?
Offline
Okay, if you disable ASPM and the other intel smart power management thing in the BIOS, it boots fine without any kernel parameters needed.
Offline
Okay, if you disable ASPM and the other intel smart power management thing in the BIOS, it boots fine without any kernel parameters needed.
Yes, I can confirm that helped for me as well, I hope the kernel devs or the intel devs can solve the problem
Offline
I have the same issue.
Problem with disabling that stuff is that CPU freq is fixed at 1.6Ghz
Offline
Offline
Where should we report this bug?
Offline
Solution for me in similar case - removing i915 from MODULES line in /etc/mkinitcpio.conf
(my configuration - lenovo ideapad u330p (Intel(R) Core(TM) i3-4010U CPU, HD 4400 ), gummiboot, x86_64)
Yes I've got a problem with i915 and this kernel. I need the module as I'm running an Intel GM965/GL960 graphics chip, although it usually works without including it in etc/mkincpio.conf. With the 3.15 kernel my computer crashes to coloured vertical stripes at tty login screen. Disabling modesetting allows tty login but that's not useful as it stops X working.
(my config - Toshiba Satellite Pro - Intel Core2Duo T5450)
Yes, I created a thread in the Intel Forum as well!
I think the problem is with kernel rather than Intel driver as latest driver (xf86-video-intel-2.99.912-2-x86_64) works fine with lts and 3.14-5 kernels. For now my solution is kernel downgrade.
Offline
Curiouser and curiouser - i915, linux-3.15.3-1, xf86-video-intel and xorg-server all work fine from the live CD. But not when installed to HD. Anybody got any ideas?
Offline
I also have an Intel NUC 2820 which would not boot properly until I turned off the Intel Speedstep option in the BIOS.
However, if I disable the intel pstate cpu driver (with intel_pstate=disable on the kernel commandline) then I can boot with the Intel Speedstep options turned on in the BIOS. I guess this means there is a bug or a missing quirk in the intel_pstate driver for this chip.
I think the people with issues with the i915 driver are running into an unrelated bug.
Offline
Thanks ikandros, moved the i915 stuff to a different thread
Offline
Has anyone been brave enough to try the intel_pstate driver with kernel 3.16-2? I don't have a usb keyboard handy to recover if it won't boot.
Offline