View previous topic :: View next topic |
Author |
Message |
ulenrich Veteran
Joined: 10 Oct 2010 Posts: 1480
|
Posted: Tue Apr 23, 2013 12:27 pm Post subject: [solved] EFI stub regression at linux-3.9-rc8 |
|
|
Is someone able to run the new Linux-3.9 with EFI-stub directly?
I am only able to boot Linux-3.9 in compatibility mode
Last edited by ulenrich on Sun Apr 28, 2013 11:22 pm; edited 2 times in total |
|
Back to top |
|
|
khayyam Watchman
Joined: 07 Jun 2012 Posts: 6227 Location: Room 101
|
Posted: Tue Apr 23, 2013 1:33 pm Post subject: Re: Linux-3.8 runs as EFI stub but not linux-3.9-rc8 on Mac- |
|
|
ulenrich wrote: | Is someone able to run the new Linux-3.9 with EFI-stub directly? I am only able to boot Linux-3.9 in compatibility mode |
ulenrich ... I imagine this the same issue I had with eboot.c and 3.8.x ... try copying arch/x86/boot/compressed/eboot.c from 3.7 to the same location in 3.9 (probably a good idea to make a backup of eboot.c before hand).
best ... khay |
|
Back to top |
|
|
ulenrich Veteran
Joined: 10 Oct 2010 Posts: 1480
|
Posted: Tue Apr 23, 2013 3:41 pm Post subject: |
|
|
diff linux-3.8/.../eboot.c linux-3.9/.../eboot.c
reveals a brand new function:
setup_efi_vars(boot_params);
... and one call of it, should I just comment out this one call?
But diffs on other files reveal a lot more changed regarding efi in Linux-3.9 ... |
|
Back to top |
|
|
khayyam Watchman
Joined: 07 Jun 2012 Posts: 6227 Location: Room 101
|
Posted: Tue Apr 23, 2013 4:19 pm Post subject: |
|
|
ulenrich wrote: | diff linux-3.8/.../eboot.c linux-3.9/.../eboot.c
reveals a brand new function: setup_efi_vars(boot_params); ... and one call of it, should I just comment out this one call? |
ulenrich ... as I said, I don't know if your issue is the same, different kernel, different issues ... perhaps. I simply suggested you try the solution that worked for 3.8.8.
ulenrich wrote: | But diffs on other files reveal a lot more changed regarding efi in Linux-3.9 ... |
Thats quite possible, but it may function all the same with the eboot.c from 3.7 as this is used for memory addressing on boot and so not actually efi itself. Anyhow, I'd try and see, it takes only a minute to replace the file, recompile bzImage, and boot.
best ... khay |
|
Back to top |
|
|
chithanh Developer
Joined: 05 Aug 2006 Posts: 2158 Location: Berlin, Germany
|
Posted: Tue Apr 23, 2013 10:33 pm Post subject: |
|
|
I am booting 3.9_rc7 EFI stub kernel on ASUS M5A97 R2.0 fine. Sometimes I noticed that kernels would fail to boot after upgrade, but it worked after powering off the computer. |
|
Back to top |
|
|
ulenrich Veteran
Joined: 10 Oct 2010 Posts: 1480
|
Posted: Wed Apr 24, 2013 8:46 am Post subject: simply a regression |
|
|
chithanh wrote: | I am booting 3.9_rc7 EFI stub kernel on ASUS M5A97 R2.0 fine. Sometimes I noticed that kernels would fail to boot after upgrade, but it worked after powering off the computer. | @chithanh,
very thanks for your info. This motivated me to retry with older Linux- 3.9_rc7
I can confirm a regression for working with EFI with a 3 years old Mac-Mini:
Linux- 3.9_rc7 efi works
Linux- 3.9_rc8 efi does NOT work
Code: | boot # diff config-3.9.0-9997 config-3.9.0-9998
53c53
< CONFIG_LOCALVERSION="-9997"
---
> CONFIG_LOCALVERSION="-9998"
3653a3654
> CONFIG_UCS2_STRING=y |
9998 == rc8 I do patch away Makefile EXTRAVERSION - I don't like.
I didn't manually enable UCS2_STRING for Linux- 3.9_rc8! |
|
Back to top |
|
|
chithanh Developer
Joined: 05 Aug 2006 Posts: 2158 Location: Berlin, Germany
|
|
Back to top |
|
|
khayyam Watchman
Joined: 07 Jun 2012 Posts: 6227 Location: Room 101
|
Posted: Sun Apr 28, 2013 6:11 pm Post subject: |
|
|
chithanh wrote: | There was one EFI boot fix after 3.9_rc8 which might apply to your situation. |
chithanh ... that seems to be the fix for the eboot.c issue I mention above.
best ... khay |
|
Back to top |
|
|
ulenrich Veteran
Joined: 10 Oct 2010 Posts: 1480
|
Posted: Sun Apr 28, 2013 10:38 pm Post subject: |
|
|
@chithanh, thanks!
Yes, this seems to be the bug with my older MacMini:
J.Boyer wrote: | Older implementations of EFI don't have QueryVariableInfo but the runtime is a smaller structure, so the pointer to it may be pointing off into garbage. This is apparently the case with several Apple firmwares that support EFI 1.10, and the current check causes them to no longer boot. |
[edit]Yes, boots again!
This will be no issue anymore with coming release (tomorrow?). |
|
Back to top |
|
|
|