New update to 3.15.5 seems to break bootloader, getting black screen

dam5h · 2014-07-15 01:24:36

I did 'pacman -Syu' last night and today all I get is a black screen after post. It's the same behavior that I get with the live arch image when trying to use UEFI boot. I am posting this from a fedora live image right now. I originally had to install Arch via the Fedora live image via chroot, back in April. The machine is a ThinkPad T440s.

I have chrooted in to get the pacman log from last night. Here it is

[2014-07-13 21:03] [PACMAN] Running 'pacman -Syu'
[2014-07-13 21:03] [PACMAN] synchronizing package lists
[2014-07-13 21:03] [PACMAN] starting full system upgrade
[2014-07-13 21:03] [PACMAN] upgraded libdbus (1.8.4-1 -> 1.8.6-1)
[2014-07-13 21:03] [PACMAN] upgraded dbus (1.8.4-1 -> 1.8.6-1)
[2014-07-13 21:03] [PACMAN] upgraded elfutils (0.158-3 -> 0.159-1)
[2014-07-13 21:03] [ALPM] warning: /etc/group installed as /etc/group.pacnew
[2014-07-13 21:03] [ALPM] warning: /etc/gshadow installed as /etc/gshadow.pacnew
[2014-07-13 21:03] [PACMAN] upgraded filesystem (2014.06-2 -> 2014.07-1)
[2014-07-13 21:03] [PACMAN] upgraded flashplugin (11.2.202.378-1 -> 11.2.202.394-1)
[2014-07-13 21:03] [PACMAN] upgraded harfbuzz (0.9.29-1 -> 0.9.30-1)
[2014-07-13 21:03] [PACMAN] upgraded imlib2 (1.4.6-2 -> 1.4.6-3)
[2014-07-13 21:03] [PACMAN] upgraded libsystemd (214-2 -> 215-4)
[2014-07-13 21:03] [PACMAN] upgraded systemd (214-2 -> 215-4)
[2014-07-13 21:03] [PACMAN] upgraded mesa (10.2.2-1 -> 10.2.3-1)
[2014-07-13 21:03] [PACMAN] upgraded mesa-libgl (10.2.2-1 -> 10.2.3-1)
[2014-07-13 21:03] [PACMAN] upgraded intel-dri (10.2.2-1 -> 10.2.3-1)
[2014-07-13 21:03] [PACMAN] upgraded libxi (1.7.2-1 -> 1.7.3-1)
[2014-07-13 21:03] [ALPM-SCRIPTLET] >>> Updating module dependencies. Please wait ...
[2014-07-13 21:03] [ALPM-SCRIPTLET] >>> Generating initial ramdisk, using mkinitcpio.  Please wait...
[2014-07-13 21:03] [ALPM-SCRIPTLET] ==> Building image from preset: /etc/mkinitcpio.d/linux.preset: 'default'
[2014-07-13 21:03] [ALPM-SCRIPTLET]   -> -k /boot/vmlinuz-linux -c /etc/mkinitcpio.conf -g /boot/initramfs-linux.img
[2014-07-13 21:03] [ALPM-SCRIPTLET] ==> Starting build: 3.15.5-1-ARCH
[2014-07-13 21:03] [ALPM-SCRIPTLET]   -> Running build hook: [base]
[2014-07-13 21:03] [ALPM-SCRIPTLET]   -> Running build hook: [udev]
[2014-07-13 21:03] [ALPM-SCRIPTLET]   -> Running build hook: [autodetect]
[2014-07-13 21:03] [ALPM-SCRIPTLET]   -> Running build hook: [modconf]
[2014-07-13 21:03] [ALPM-SCRIPTLET]   -> Running build hook: [block]
[2014-07-13 21:03] [ALPM-SCRIPTLET]   -> Running build hook: [resume]
[2014-07-13 21:03] [ALPM-SCRIPTLET]   -> Running build hook: [filesystems]
[2014-07-13 21:03] [ALPM-SCRIPTLET]   -> Running build hook: [keyboard]
[2014-07-13 21:03] [ALPM-SCRIPTLET]   -> Running build hook: [fsck]
[2014-07-13 21:03] [ALPM-SCRIPTLET] ==> Generating module dependencies
[2014-07-13 21:03] [ALPM-SCRIPTLET] ==> Creating gzip initcpio image: /boot/initramfs-linux.img
[2014-07-13 21:03] [ALPM-SCRIPTLET] ==> Image generation successful
[2014-07-13 21:03] [ALPM-SCRIPTLET] ==> Building image from preset: /etc/mkinitcpio.d/linux.preset: 'fallback'
[2014-07-13 21:03] [ALPM-SCRIPTLET]   -> -k /boot/vmlinuz-linux -c /etc/mkinitcpio.conf -g /boot/initramfs-linux-fallback.img -S autodetect
[2014-07-13 21:03] [ALPM-SCRIPTLET] ==> Starting build: 3.15.5-1-ARCH
[2014-07-13 21:03] [ALPM-SCRIPTLET]   -> Running build hook: [base]
[2014-07-13 21:03] [ALPM-SCRIPTLET]   -> Running build hook: [udev]
[2014-07-13 21:03] [ALPM-SCRIPTLET]   -> Running build hook: [modconf]
[2014-07-13 21:03] [ALPM-SCRIPTLET]   -> Running build hook: [block]
[2014-07-13 21:03] [ALPM-SCRIPTLET] ==> WARNING: Possibly missing firmware for module: aic94xx
[2014-07-13 21:03] [ALPM-SCRIPTLET] ==> WARNING: Possibly missing firmware for module: smsmdtv
[2014-07-13 21:03] [ALPM-SCRIPTLET]   -> Running build hook: [resume]
[2014-07-13 21:03] [ALPM-SCRIPTLET]   -> Running build hook: [filesystems]
[2014-07-13 21:03] [ALPM-SCRIPTLET]   -> Running build hook: [keyboard]
[2014-07-13 21:03] [ALPM-SCRIPTLET]   -> Running build hook: [fsck]
[2014-07-13 21:03] [ALPM-SCRIPTLET] ==> Generating module dependencies
[2014-07-13 21:03] [ALPM-SCRIPTLET] ==> Creating gzip initcpio image: /boot/initramfs-linux-fallback.img
[2014-07-13 21:03] [ALPM-SCRIPTLET] ==> Image generation successful
[2014-07-13 21:03] [PACMAN] upgraded linux (3.15.3-1 -> 3.15.5-1)
[2014-07-13 21:03] [PACMAN] upgraded mpfr (3.1.2.p8-1 -> 3.1.2.p10-1)
[2014-07-13 21:03] [ALPM] warning: /etc/pacman.d/mirrorlist installed as /etc/pacman.d/mirrorlist.pacnew
[2014-07-13 21:03] [PACMAN] upgraded pacman-mirrorlist (20140531-1 -> 20140706-1)
[2014-07-13 21:03] [PACMAN] upgraded python2-setuptools (1:5.4-1 -> 1:5.4.1-1)
[2014-07-13 21:03] [PACMAN] upgraded systemd-sysvcompat (214-2 -> 215-4)
[2014-07-13 21:03] [PACMAN] upgraded xcb-util-image (0.3.9-1 -> 0.3.9-2)
[2014-07-13 21:03] [PACMAN] upgraded xcb-util-keysyms (0.3.9-1 -> 0.3.9-2)
[2014-07-13 21:03] [PACMAN] upgraded xorg-xkbcomp (1.2.4-1 -> 1.2.4-2)

Any ideas would be greatly appreciated. I can't really tell if it's getting to gummiboot or not, since I only ever had one entry it never went to a selection screen anyway. I added an entry in /boot/loader/entries that is basically a copy of the one I had in there already, except I removed the resume statement and resume_offset statements.

dam5h · 2014-07-15 02:17:40

Downgraded to 3.15.3 and all is back to normal. I will sit on this kernel for now.

Zzipo · 2014-07-16 16:34:09

Same here. 3.15.5-1 -> Black booting. Thought it was related with the card. Chrooted and downgraded virtualbox-host-modules and linux to 3.12.

One question related to this: Imagine that the 3.15.3 like yours works also in my PC (but I have only in my /var/cache the 3.12, how can I get that version if in the repos there are just the latest? Any idea? Any repos that keeps multiple versions?

Sanne · 2014-07-16 17:49:36

Zzipo wrote:

Any repos that keeps multiple versions?

Yup, we have the ARM.

das_coach · 2014-07-16 19:09:58

Happening on an Aspire E1-571G (EFI boot via rEFInd), too.
It starts with an "efi: System table signature incorrect!", followed by several "genirq: Flags mismatch irq 0. 00000080 (mmc0) vs. 00015a00 (timer)", (also for i801_smbus vs. timer and brcmsmac vs. timer) what results in "Failed to requestIRQ 0: -16" and "Found HC with no IRQ. Check BIOS/PCI [...] setup!".

I can't enter anything as the keyboard seems to be not initialized, but a downgrade to 3.14.6-1 (via livecd) also fixed it for me.

arboretumfess · 2014-07-17 00:01:23

Same problem. 3.15.4 works fine.

x33a · 2014-07-17 05:36:03

dam5h wrote:

I have chrooted in to get the pacman log from last night.

Slightly off-topic, but you don't need to to chroot just to access the pacman log.

Zzipo · 2014-07-17 17:08:44

Sanne wrote:

Zzipo wrote:
Any repos that keeps multiple versions?
Yup, we have the ARM.

Thanks a lot! I didn't know.

x33a wrote:

but you don't need to to chroot just to access the pacman log.

Althought it is not to me, do you mean from grub2? (in my case, because the boot loading process were stuck just after grub stage2.)

x33a · 2014-07-17 17:39:29

Zzipo wrote:

x33a wrote:
but you don't need to to chroot just to access the pacman log.
Althought it is not to me, do you mean from grub2? (in my case, because the boot loading process were stuck just after grub stage2.)

No, I mean that if you have access to a live media, you can simply mount any partition on your hard drive to access the files (including pacman log, for example), no need to chroot.

Zzipo · 2014-07-17 17:51:17

Ah Ok, perfect.

das_coach · 2014-07-17 18:09:45

@Zzipo:
Chroot'ing (changing the root that is used by your current command prompt) is used if you need to run commands as if you're using the chroot'ed environment (e.g. installing packages), mounting is used if you only need to access the files on the device (with programs that are available in the hosting os).

Last edited by das_coach (2014-07-17 18:10:31)

Zzipo · 2014-07-17 18:23:32

das_coach wrote:

@Zzipo:
Chroot'ing (changing the root that is used by your current command prompt) is used if you need to run commands as if you're using the chroot'ed environment (e.g. installing packages), mounting is used if you only need to access the files on the device (with programs that are available in the hosting os).

Yes yes, I knew that, but I was not sure if from grub (term) I could just do the same action or see what the logs had. Thanks anyway!

AleCon · 2014-07-17 21:15:57

same problem here with a fresh install and gummiboot:
"efi: System table signature incorrect!" message and no keyboard after boot.
I can access keyboard adding options on the boot comand line but hardware recognition seems to be completely crippled

ubunchu · 2014-07-18 00:07:41

Holy floating toilets!
I've been trying to install on an Acer Travelmate B113E/B113M, and a frikking version downgrade solved it! (3.15.4)
For days I thought I had GPT-gummiboot related issues. (issues here were: booting, but with errors, got to login tty, keyboard not working)
Thanks guys!

Last edited by ubunchu (2014-07-18 00:08:37)

ntony · 2014-07-18 08:25:45

Having this problem 2 days ago. Downgrading to linux kernel to 3.15.1-1 works. Anyone fired this bug? Whom to fire this bug on? Linus? or Gummiboot?

das_coach · 2014-07-18 13:44:01

@Zzipo: sorry, just wanted to make sure (and to get corrected if i am wrong^^)
@ntony: it's not gummiboot related, as it also occurs with rEFInd. i guess it's either the ramdisk or kernel, especially as it seem's that this bug breaks the whole hardware recognition (efi system partition, smbus, wlan, keyboard, mmc, etc.), but who to inform about that? good question:-/ i'm willing to create any log or debug information, but i'd need some details about which information is needed (and possibly sometimes how to acquire them, especially since changing to systemd).

arboretumfess · 2014-07-18 17:47:09

New 3.15.5-2 kernel works fine.

cybe-arch · 2014-07-18 22:23:34

Same issue here.

EFI (gummi) boot on X1C rev 2.
Downgrading kernel solves it.

ntony · 2014-07-24 09:37:10

cybe-arch wrote:

Same issue here.
EFI (gummi) boot on X1C rev 2.
Downgrading kernel solves it.

Which kernel version were you using?
3.15.5.-1 or 3.15.5-2?

ntony · 2014-07-24 09:51:23

arboretumfess wrote:

New 3.15.5-2 kernel works fine.

Same here. 3.15.5-2 fixed my problem too.

Arch Linux

#1 2014-07-15 01:24:36

New update to 3.15.5 seems to break bootloader, getting black screen

#2 2014-07-15 02:17:40

Re: New update to 3.15.5 seems to break bootloader, getting black screen

#3 2014-07-16 16:34:09

Re: New update to 3.15.5 seems to break bootloader, getting black screen

#4 2014-07-16 17:49:36

Re: New update to 3.15.5 seems to break bootloader, getting black screen

#5 2014-07-16 19:09:58

Re: New update to 3.15.5 seems to break bootloader, getting black screen

#6 2014-07-17 00:01:23

Re: New update to 3.15.5 seems to break bootloader, getting black screen

#7 2014-07-17 05:36:03

Re: New update to 3.15.5 seems to break bootloader, getting black screen

#8 2014-07-17 17:08:44

Re: New update to 3.15.5 seems to break bootloader, getting black screen

#9 2014-07-17 17:39:29

Re: New update to 3.15.5 seems to break bootloader, getting black screen

#10 2014-07-17 17:51:17

Re: New update to 3.15.5 seems to break bootloader, getting black screen

#11 2014-07-17 18:09:45

Re: New update to 3.15.5 seems to break bootloader, getting black screen

#12 2014-07-17 18:23:32

Re: New update to 3.15.5 seems to break bootloader, getting black screen

#13 2014-07-17 21:15:57

Re: New update to 3.15.5 seems to break bootloader, getting black screen

#14 2014-07-18 00:07:41

Re: New update to 3.15.5 seems to break bootloader, getting black screen

#15 2014-07-18 08:25:45

Re: New update to 3.15.5 seems to break bootloader, getting black screen

#16 2014-07-18 13:44:01

Re: New update to 3.15.5 seems to break bootloader, getting black screen

#17 2014-07-18 17:47:09

Re: New update to 3.15.5 seems to break bootloader, getting black screen

#18 2014-07-18 22:23:34

Re: New update to 3.15.5 seems to break bootloader, getting black screen

#19 2014-07-24 09:37:10

Re: New update to 3.15.5 seems to break bootloader, getting black screen

#20 2014-07-24 09:51:23

Re: New update to 3.15.5 seems to break bootloader, getting black screen

Board footer