[Bug 828377] New: Asus Zenbook Ux31a Ivy Bridge boots to blank screen with kernel greater than 3.7 series
https://bugzilla.novell.com/show_bug.cgi?id=828377 https://bugzilla.novell.com/show_bug.cgi?id=828377#c0 Summary: Asus Zenbook Ux31a Ivy Bridge boots to blank screen with kernel greater than 3.7 series Classification: openSUSE Product: openSUSE 12.3 Version: Final Platform: x86-64 OS/Version: openSUSE 12.3 Status: NEW Severity: Critical Priority: P5 - None Component: Kernel AssignedTo: kernel-maintainers@forge.provo.novell.com ReportedBy: matt.mccann.8@gmail.com QAContact: qa-bugs@suse.de Found By: --- Blocker: --- User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:22.0) Gecko/20100101 Firefox/22.0 After installing 12.3 and updating to the latest Kernel my Asus UX31a laptop will not even show the UEFI/BIOS. It powers up but displays nothing and no combination of keys does anything. The only solution was to pop open the computer, disconnect the SSD and boot the computer. The UEFI/BIOS then appeared. I then reconnected the drive and was able to reinstall openSUSE 12.3. Leaving the machine at the default 3.7 kernel this time I have had no issue with the machine booting. Reproducible: Always Steps to Reproduce: 1.Install openSUSE 12.3 2.Update kernel to something greater than the 3.7 series 3. Laptop is no longer able to boot Actual Results: After installing the new kernel the laptop will only boot to a blank screen. No UEFI/BIOS, no GRUB screen. Nothing. Expected Results: The computer should start and offer me the Grub screen to boot into openSUSE 12.3. My guess, and I am in no way an expert, is that when the new kernel gets installed it corrupts the UEFI boot partition in some way that renders the computer completely unbootable. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=828377
https://bugzilla.novell.com/show_bug.cgi?id=828377#c1
--- Comment #1 from Matt McCann
https://bugzilla.novell.com/show_bug.cgi?id=828377
https://bugzilla.novell.com/show_bug.cgi?id=828377#c2
--- Comment #2 from Matt McCann
https://bugzilla.novell.com/show_bug.cgi?id=828377
https://bugzilla.novell.com/show_bug.cgi?id=828377#c
Jeff Mahoney
https://bugzilla.novell.com/show_bug.cgi?id=828377
https://bugzilla.novell.com/show_bug.cgi?id=828377#c3
Joey Lee
https://bugzilla.novell.com/show_bug.cgi?id=828377
https://bugzilla.novell.com/show_bug.cgi?id=828377#c4
--- Comment #4 from Matt McCann
Hi Matt,
Does this issue also happen on v3.10 vanilla kernel?
I asked because there have wall EFI clock change in v3.8 kernel but it should make things on Asus machine better but not brick it. On the other hand, that also strange for remove SSD could recover the problem.
Thanks
Joey, Unfortunately, I didn't try the vanilla kernel prior to my latest fresh install. Once I got the machine to a useable state I left it on the stock 3.7 desktop kernel. I can't do it just now as I need the machine to be useable, but hopefully in early August I will have time where I can test new kernels and see what happens. Best, Matt -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=828377
https://bugzilla.novell.com/show_bug.cgi?id=828377#c5
Joey Lee
(In reply to comment #3)
Hi Matt,
Does this issue also happen on v3.10 vanilla kernel?
I asked because there have wall EFI clock change in v3.8 kernel but it should make things on Asus machine better but not brick it. On the other hand, that also strange for remove SSD could recover the problem.
Thanks
Joey,
Unfortunately, I didn't try the vanilla kernel prior to my latest fresh install. Once I got the machine to a useable state I left it on the stock 3.7 desktop kernel. I can't do it just now as I need the machine to be useable, but hopefully in early August I will have time where I can test new kernels and see what happens.
Best, Matt
Thanks for your quick response. As far as i know the only possible issue causes brick machine is the flash space of EFI variables full but BIOS didn't trigger garbage collection. But it's not make sense remove SSD could rescued machine back. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=828377
https://bugzilla.novell.com/show_bug.cgi?id=828377#c6
--- Comment #6 from Michael Chang
https://bugzilla.novell.com/show_bug.cgi?id=828377
https://bugzilla.novell.com/show_bug.cgi?id=828377#c7
--- Comment #7 from Matt McCann
(( OOPs, mid air collision ... resend. ))
In openSUSE, updating kernel will trigger mkinitrd and in turn performs bootloader update that will also manipulate efi boot variables as a side effect.
Actually, for kernel update, any re-installation of bootloader is unneeded and merely a waste of cpu cycles , all we need is only update the bootloader config to reflect the kernel changes (and that's solely the purpose of running update-bootloader for mkinitrd).
Anyway above could be off-topic, but it reveals some potential causes to this problem, the (although unintended) uefi variable update accompanying with kernel update process, as well as re-installation of grub and it's config to the EFI System Partition.
What I'm not clear from your report is whether you could boot into the problematic kernel (>3.7 series) by reconnecting the drive, for doing some tests and grabbing some logs to narrow down the cause ?
Thanks.
Michael, No I was not able to boot into the system after reconnecting the drive. The system recognized the drive, but dumped me straight into the UEFI/Bios and wouldn't allow me to boot to the SSD. I was only able to boot to the SSD after a full re-installation of openSUSE from a USB stick. Though I am no expert, from reading through various posts etc. it seemed like my issue had something to do with the updating of the bootloader/altering the EFI partition and not necessarily the kernel itself, but as you mentioned, installing a new kernel in openSUSE will trigger such an update which is why I reported it as a kernel bug. At first I was worried I had an issue similar to the Samsung laptop issue that was actually bricking laptops. Best, Matt -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=828377
https://bugzilla.novell.com/show_bug.cgi?id=828377#c8
--- Comment #8 from Michael Chang
(In reply to comment #6)
No I was not able to boot into the system after reconnecting the drive. The system recognized the drive, but dumped me straight into the UEFI/Bios and wouldn't allow me to boot to the SSD. I was only able to boot to the SSD after a full re-installation of openSUSE from a USB stick.
Looks really like ESP somehow contaminated. Or the EFI Boot Manager has problem in booting the boot loaders (could the boot variable somehow destroyed or some other changes in hardware configuration .. ??)
Though I am no expert, from reading through various posts etc. it seemed like my issue had something to do with the updating of the bootloader/altering the EFI partition and not necessarily the kernel itself, but as you mentioned,
Yes, I'm also suspecting it .. Would you mind to run stress mkinitrd/reboot cycle on your up-and-running 3.7 kernel to see what's happened ??
installing a new kernel in openSUSE will trigger such an update which is why I reported it as a kernel bug. At first I was worried I had an issue similar to the Samsung laptop issue that was actually bricking laptops.
Yes, you're right. This is also pretty much kernel bug per the report. What I'm feeling not right is that we can avoid the mess of identifying kernel or bootloader update process if we do things right. Now we are taking extra effort in segregating them. :( Thanks, Michael
Best, Matt
-- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=828377
https://bugzilla.novell.com/show_bug.cgi?id=828377#c9
--- Comment #9 from Matt McCann
(In reply to comment #7)
(In reply to comment #6)
No I was not able to boot into the system after reconnecting the drive. The system recognized the drive, but dumped me straight into the UEFI/Bios and wouldn't allow me to boot to the SSD. I was only able to boot to the SSD after a full re-installation of openSUSE from a USB stick.
Looks really like ESP somehow contaminated. Or the EFI Boot Manager has problem in booting the boot loaders (could the boot variable somehow destroyed or some other changes in hardware configuration .. ??)
Though I am no expert, from reading through various posts etc. it seemed like my issue had something to do with the updating of the bootloader/altering the EFI partition and not necessarily the kernel itself, but as you mentioned,
Yes, I'm also suspecting it ..
Would you mind to run stress mkinitrd/reboot cycle on your up-and-running 3.7 kernel to see what's happened ??
installing a new kernel in openSUSE will trigger such an update which is why I reported it as a kernel bug. At first I was worried I had an issue similar to the Samsung laptop issue that was actually bricking laptops.
Yes, you're right. This is also pretty much kernel bug per the report. What I'm feeling not right is that we can avoid the mess of identifying kernel or bootloader update process if we do things right. Now we are taking extra effort in segregating them. :(
Thanks, Michael
Best, Matt
How do I go about doing a mkinitrd/reboot cycle stress test? Thanks. Best, Matt -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=828377
https://bugzilla.novell.com/show_bug.cgi?id=828377#c10
--- Comment #10 from Michael Chang
How do I go about doing a mkinitrd/reboot cycle stress test? Thanks.
Run the two commands several times (can't exactly tell how much times, 3 - 5 should suffice ..) $ mkinitrd $ reboot This is to simulate the bootloader update happened during kernel update. If we passed the test that means this is a kernel regression bug, otherwise not. Thanks. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=828377
https://bugzilla.novell.com/show_bug.cgi?id=828377#c11
--- Comment #11 from Matt McCann
(In reply to comment #9)
How do I go about doing a mkinitrd/reboot cycle stress test? Thanks.
Run the two commands several times (can't exactly tell how much times, 3 - 5 should suffice ..)
$ mkinitrd $ reboot
This is to simulate the bootloader update happened during kernel update. If we passed the test that means this is a kernel regression bug, otherwise not.
Thanks.
I'll give it a go and report back. Thanks. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=828377
https://bugzilla.novell.com/show_bug.cgi?id=828377#c12
--- Comment #12 from Matt McCann
https://bugzilla.novell.com/show_bug.cgi?id=828377
https://bugzilla.novell.com/show_bug.cgi?id=828377#c13
--- Comment #13 from Joey Lee
So I decided to try the latest and greatest kernel on my Zenbook to see what happens. I compiled the 3.11-RC2 kernel from kernel.org. It compiled and installed without issue. The system also boots after repeated restarts. It seems there is no corruption occurring. I then compiled the 3.10.2 kernel and also did not have issues.
Thanks for your testing result, I will check anything about EFI and EFI variable filesystem from v3.7 to v3.10. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=828377
https://bugzilla.novell.com/show_bug.cgi?id=828377#c14
Joey Lee
participants (1)
-
bugzilla_noreply@novell.com