Kernel 3.7 doesn't boot using EFI Stub

fauxCoder · 2013-04-15 22:19:17

UEFI gets a mention in the 3.8.7 change logs: https://www.kernel.org/pub/linux/kernel … eLog-3.8.7

Is this perhaps what solved the problem?

cfr · 2013-04-16 00:54:33

It hasn't fixed the issue for me. Nothing in 3.8.* works.

cdrc · 2013-04-16 19:56:10

Nothing in 3.8.* worked for me either until today. I realized I never updated the uefi (or whatyoucallit on the motherboard, which is an Asrock H61iCafe).

I had boot failures after the update even with the 3.7.9-2 that I downgraded to, but I did not copy the refind file since the last time I tried to solve this a month ago. Now it’s fine.

andyc · 2013-04-16 22:19:00

3.8.7 update failed to boot here too using refind. Motherboard is Gigabyte X79-UP4.
Failed to load kernel modules. Failed to mount /boot/efi.

Downgrade to 3.8.6 works again.
Happy to help with info to fix, just need to know what info required and how to find it. New to arch.

PaulSamways · 2013-04-16 22:25:12

Just for reference for those of you running Lenovo systems, I've had good success with the last few updates (3.8.6+) on my T430.

cfr · 2013-04-16 23:06:30

It seems to be more specific than that. I'm on Lenovo, too, and my firmware is up to date.

I'm not sure what is meant by the refind file. I double-checked to be sure but the refind .efi is identical to that stored in /usr/lib which makes sense as it hasn't been updated since I last did it. I don't keep the kernel etc. on the ESP but use the ext4 driver to load from /boot. I'm not sure what else might be meant...

cdrc · 2013-04-17 00:15:30

I meant refind_x64.efi, which has been updated since the last time I did it.

The only observation I wanted to make was that without an up to date firmware, I experienced the problem exactly the way you did ; with it, well I don't know, I’ll take an hour this weekend to see what works and what doesn’t exactly but it should be the same as some other people earlier in this thread.

cfr · 2013-04-17 01:05:22

Thanks, I checked and mine is up to date as is the firmware, the kernel... Still doesn't work, though.

catharsis · 2013-04-18 12:18:32

3.8.7 works for me.

teateawhy · 2013-04-19 15:35:25

Works for me:
linux 3.8.7-1
refind_x64.efi from this thread. ( http://www.rodsbooks.com/refind_x64_gnuefi_3.6.8.efi )
Lenovo x220

cdrc · 2013-04-19 21:34:01

3.8.7-1 and refind 0.6.8-1 from Extra works if:

an USB flash drive or game controller is plugged in
USB 3 is disabled
USB 3 legacy support is enabled

cfr · 2013-04-23 22:01:59

3.8.8-1 also fails here.

MFserver · 2013-04-24 08:59:47

3.8.7-1 works for me with gummiboot from the repositories, without any extra quirks. Lenovo Thinkpad W530.

delaerep · 2013-04-28 17:08:18

3.8.10-1 fails with refind on ASUS F2A85-M LE motherboard as did every kernel above 3.7.10

version 3.9.1-1 from testing works

Last edited by delaerep (2013-05-09 11:31:33)

65kid · 2013-04-28 18:24:29

funny, I never had this problem, now it finally got me too.
Kernel 3.8.8-2 works fine, kernel 3.8.10 just hangs there with a black screen (Samsung Series 9 900x4d with gummiboot). And there wasn't even a single EFI related change between these kernel releases... seriously, wtf?

WonderWoofy · 2013-04-28 18:31:07

65kid wrote:

funny, I never had this problem, now it finally got me too.
Kernel 3.8.8-2 works fine, kernel 3.8.10 just hangs there with a black screen (Samsung Series 9 900x4d with gummiboot). And there wasn't even a single EFI related change between these kernel releases... seriously, wtf?

Have you tried anything besides gummiboot? I got bit by this very briefly, and so I now have rEFInd as well as gummiboot. When it was doing it, sometimes gummiboot would work with a kernel and sometimes rEFInd would work... but there was never any indication as to why these changes were happening. So I also keep elilo on my system just in case.

65kid · 2013-04-28 18:53:53

WonderWoofy wrote:

65kid wrote:
funny, I never had this problem, now it finally got me too.
Kernel 3.8.8-2 works fine, kernel 3.8.10 just hangs there with a black screen (Samsung Series 9 900x4d with gummiboot). And there wasn't even a single EFI related change between these kernel releases... seriously, wtf?
Have you tried anything besides gummiboot? I got bit by this very briefly, and so I now have rEFInd as well as gummiboot. When it was doing it, sometimes gummiboot would work with a kernel and sometimes rEFInd would work... but there was never any indication as to why these changes were happening. So I also keep elilo on my system just in case.

I'm too lazy to set up another boot loader right now. I just tried linux-mainline (3.9-rc8), which boots just fine, so I'm gonna hope and assume that it will magically be fixed with the next stable kernel release. Seeing as how this problems seems to happen totally randomly for people, I think this is a fair assumption.

btw, I haven't read the whole thread, has there been some kind of comparison between firmwares? bootctl shows me the following:

Firmware: UEFI 2.31 (Phoenix Technologies Ltd. 4660.22136)

srs5694 · 2013-04-28 22:26:41

FWIW, there's been a very recent kernel change (posted four days ago) that may fix this problem on at least some systems:

http://git.kernel.org/cgit/linux/kernel … 56c0ddac45

My suspicion, though, is that there are actually two problems: One seems to affect mostly Mac users and is very reliable, in the sense that it's easy to reproduce and it always prevents booting. That's the problem that the new patch will fix. The other problem is more arbitrary: It can manifest when launching from boot manager X but not from boot manager Y, or from build A but not build B of the same boot manager. Some sort of alignment issue (as in, the alignment of the kernel in memory to word or long word boundaries) might explain this problem, but I haven't studied the relevant code to know if that's really plausible.

cfr · 2013-04-29 01:37:49

The latest fails for me as has everything 3.8.* up.

I couldn't find any relevant-looking bugs on kernel.org but maybe I just searched for the wrong thing? I was just curious what, if anything, was being looked at in relation to this. Not really understanding how this stuff works, my capacity to do anything is pretty much limited to observer status... (And saying "still doesn't work" periodically.)

I'm curious as to why mine is so consistent. It seems like most people it works - doesn't - works - fails - works etc. apparently randomly with kernel changes etc. But for me it seems to be just everything after a certain point fails and everything before worked fine. Yet I'm not on a Mac. (Unless the mere presence of a non-EFI Mac in the same room affects booting - maybe ESP is meant to carry a double-meaning...)

cfr · 2013-04-29 01:41:44

65kid wrote:

bootctl shows me the following:
Firmware: UEFI 2.31 (Phoenix Technologies Ltd. 4660.22136)

Not sure if I'm doing something wrong, but I get:

No suitable data is provided by the boot manager. See:
  http://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
  http://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
for details.

I am, however, using Phoenix SecureCore Tiano.

EDIT: I guess only gummiboot supports bootctl?

Last edited by cfr (2013-04-29 01:53:08)

srs5694 · 2013-04-29 17:34:02

You can learn about your firmware with dmesg:

dmesg | grep -i efi
[    0.000000] efi: EFI v2.00 by American Megatrends
[    0.000000] efi:  ACPI 2.0=0x7bb5df98  SMBIOS=0x7bb88e18 
[    0.000000] efi: mem00: type=3, attr=0xf, range=[0x0000000000000000-0x0000000000048000) (0MB)
[    0.000000] efi: mem01: type=2, attr=0xf, range=[0x0000000000048000-0x0000000000050000) (0MB)
...
[    3.385592] efifb: probing for efifb
[    3.386083] efifb: framebuffer at 0xc0000000, mapped to 0xffffc90011a00000, using 6144k, total 32704k
[    3.386088] efifb: mode is 1024x768x32, linelength=4096, pages=1
[    3.386090] efifb: scrolling: redraw
[    3.386093] efifb: Truecolor: size=8:8:8:8, shift=24:16:8:0
[    3.389666] fb0: EFI VGA frame buffer device
[    3.579290] EFI Variables Facility v0.08 2004-May-17
[   16.494986] fb: conflicting fb hw usage inteldrmfb vs EFI VGA - removing generic driver

I've trimmed over 100 lines of "efi: mem???" lines from this output. Note that "dmesg" displays the kernel ring buffer, which is limited in size and cyclical -- after the system has been up for long enough, the older messages will be removed. Thus, you might get this type of EFI information only shortly after booting. What "shortly" is depends on how many kernel messages get generated -- it could be weeks, or just seconds if a driver is having problems and generating lots of messages.

Although the bug fix to which I linked specifically mentioned Apple firmware, it's entirely plausible that the same bug exists in non-Apple implementations.

cfr · 2013-04-29 23:29:36

Thanks. Here's some info:

[    0.000000] efi: EFI v2.00 by Phoenix Technologies Ltd.
[    0.000000] efi:  ACPI=0xbaffe000  ACPI 2.0=0xbaffe014  SMBIOS=0xbae65000 
...
[    0.000000] ACPI: UEFI 00000000bafe3000 0003E (v01 LENOVO TP-8Q    00001180 PTL  00000002)
[    0.000000] ACPI: UEFI 00000000bafe2000 00042 (v01 PTL      COMBUF 00000001 PTL  00000001)
[    0.000000] ACPI: UEFI 00000000bafe1000 00292 (v01 LENOVO TP-8Q    00001180 PTL  00000002)
...
[    1.849485] efifb: probing for efifb
[    1.850049] efifb: framebuffer at 0xc0000000, mapped to 0xffffc90004f00000, using 3072k, total 3072k
[    1.850052] efifb: mode is 1024x768x32, linelength=4096, pages=1
[    1.850053] efifb: scrolling: redraw
[    1.850055] efifb: Truecolor: size=8:8:8:8, shift=24:16:8:0
[    1.857201] fb0: EFI VGA frame buffer device
[    1.874278] EFI Variables Facility v0.08 2004-May-17
[    2.135646] fb: conflicting fb hw usage inteldrmfb vs EFI VGA - removing generic driver
[    2.826366] tsc: Refined TSC clocksource calibration: 1396.829 MHz

Why are the framebuffer sizes so different in my case?

Last edited by cfr (2013-04-29 23:31:02)

WonderWoofy · 2013-04-30 00:46:13

Mine are pretty different too...

[    1.093382] efifb: framebuffer at 0xe0000000, mapped to 0xffffc90004f00000, using 4128k, total 4128k

and although mine too is a Pheonix SecureCore Tiano as well, it will only tell me it is a "Lenovo" bios.

[    0.000000] efi: EFI v2.31 by Lenovo

srs5694 · 2013-04-30 03:00:42

cfr, your EFI version is fairly old, although it looks like the recent kernel patch is returning EFI_UNSUPPORTED only for pre-2.0 EFIs, so if the cause for you is that the call isn't supported on your EFI, that patch won't help -- but another one that's a bit "smarter" might.

I don't know why the framebuffer sizes on our three systems are different -- although WonderWoofy hasn't posted resolution data, so his might be running at a different resolution than the 1024x768x32 used by both my and cfr's computers.

WonderWoofy · 2013-04-30 03:34:27

Oh sorry, maybe I should have been more complete if I was trying to help here...

[    0.000000] efi: EFI v2.31 by Lenovo
[    0.000000] efi:  ACPI 2.0=0xdaffe014  ACPI=0xdaffe000  SMBIOS=0xdae9e000 
[    0.000000] efi: mem00: type=7, attr=0xf, range=[0x0000000000000000-0x0000000000087000) (0MB)
...
[    0.000000] ACPI: UEFI 00000000dafe5000 0003E (v01 LENOVO TP-H0    00002510 PTL  00000002)
[    0.000000] ACPI: UEFI 00000000dafe4000 00042 (v01 PTL      COMBUF 00000001 PTL  00000001)
[    0.000000] ACPI: UEFI 00000000dafe3000 0026A (v01 LENOVO TP-H0    00002510 PTL  00000002)

...
[    1.043522] Switching to clocksource refined-jiffies
[    1.085358] efifb: probing for efifb
[    1.086041] efifb: framebuffer at 0xe0000000, mapped to 0xffffc90004f00000, using 4128k, total 4128k
[    1.086042] efifb: mode is 1366x768x32, linelength=5504, pages=1
[    1.086043] efifb: scrolling: redraw
[    1.086045] efifb: Truecolor: size=8:8:8:8, shift=24:16:8:0
[    1.090774] fb0: EFI VGA frame buffer device
[    1.103205] EFI Variables Facility v0.08 2004-May-17
[    1.143338] fb: conflicting fb hw usage inteldrmfb vs EFI VGA - removing generic driver
[    2.075018] tsc: Refined TSC clocksource calibration: 2494.333 MHz

Arch Linux

#201 2013-04-15 22:19:17

Re: Kernel 3.7 doesn't boot using EFI Stub

#202 2013-04-16 00:54:33

Re: Kernel 3.7 doesn't boot using EFI Stub

#203 2013-04-16 19:56:10

Re: Kernel 3.7 doesn't boot using EFI Stub

#204 2013-04-16 22:19:00

Re: Kernel 3.7 doesn't boot using EFI Stub

#205 2013-04-16 22:25:12

Re: Kernel 3.7 doesn't boot using EFI Stub

#206 2013-04-16 23:06:30

Re: Kernel 3.7 doesn't boot using EFI Stub

#207 2013-04-17 00:15:30

Re: Kernel 3.7 doesn't boot using EFI Stub

#208 2013-04-17 01:05:22

Re: Kernel 3.7 doesn't boot using EFI Stub

#209 2013-04-18 12:18:32

Re: Kernel 3.7 doesn't boot using EFI Stub

#210 2013-04-19 15:35:25

Re: Kernel 3.7 doesn't boot using EFI Stub

#211 2013-04-19 21:34:01

Re: Kernel 3.7 doesn't boot using EFI Stub

#212 2013-04-23 22:01:59

Re: Kernel 3.7 doesn't boot using EFI Stub

#213 2013-04-24 08:59:47

Re: Kernel 3.7 doesn't boot using EFI Stub

#214 2013-04-28 17:08:18

Re: Kernel 3.7 doesn't boot using EFI Stub

#215 2013-04-28 18:24:29

Re: Kernel 3.7 doesn't boot using EFI Stub

#216 2013-04-28 18:31:07

Re: Kernel 3.7 doesn't boot using EFI Stub

#217 2013-04-28 18:53:53

Re: Kernel 3.7 doesn't boot using EFI Stub

#218 2013-04-28 22:26:41

Re: Kernel 3.7 doesn't boot using EFI Stub

#219 2013-04-29 01:37:49

Re: Kernel 3.7 doesn't boot using EFI Stub

#220 2013-04-29 01:41:44

Re: Kernel 3.7 doesn't boot using EFI Stub

#221 2013-04-29 17:34:02

Re: Kernel 3.7 doesn't boot using EFI Stub

#222 2013-04-29 23:29:36

Re: Kernel 3.7 doesn't boot using EFI Stub

#223 2013-04-30 00:46:13

Re: Kernel 3.7 doesn't boot using EFI Stub

#224 2013-04-30 03:00:42

Re: Kernel 3.7 doesn't boot using EFI Stub

#225 2013-04-30 03:34:27

Re: Kernel 3.7 doesn't boot using EFI Stub

Board footer