The latest linux kernel doesn't boot on my computer
(h=21511abd0a248a3f225d3b611cfabb93124605a7).
elilo hangs while booting this kernel. 2.6.24 works.
[email protected] wrote:
> The latest linux kernel doesn't boot on my computer
> (h=21511abd0a248a3f225d3b611cfabb93124605a7).
>
> elilo hangs while booting this kernel. 2.6.24 works.
Wow, so we know it's affected with EFI, since you're using elilo.
You gave absolutely zero other information about your system or what
"doesn't boot" mean.
-hpa
Zitat von "H. Peter Anvin" <[email protected]>:
> [email protected] wrote:
> > The latest linux kernel doesn't boot on my computer
> > (h=21511abd0a248a3f225d3b611cfabb93124605a7).
> >
> > elilo hangs while booting this kernel. 2.6.24 works.
>
> Wow, so we know it's affected with EFI, since you're using elilo.
>
> You gave absolutely zero other information about your system or what
> "doesn't boot" mean.
It's a macbook pro with a core duo processor (the one with only 32bits).
Doesn't boot means: i enter the name of the kernel image i want to boot in elilo
and press enter and nothing happens, it just hangs.
i need to press the power button for a few seconds to power off.
H. Peter Anvin schrieb:
> [email protected] wrote:
>> The latest linux kernel doesn't boot on my computer
>> (h=21511abd0a248a3f225d3b611cfabb93124605a7).
>>
>> elilo hangs while booting this kernel. 2.6.24 works.
>
> Wow, so we know it's affected with EFI, since you're using elilo.
>
> You gave absolutely zero other information about your system or what
> "doesn't boot" mean.
Here is an older bisection try:
$ git bisect log
git-bisect start
# bad: [dd430ca20c40ecccd6954a7efd13d4398f507728] Merge
git://git.kernel.org/pub/scm/linux/kernel/git/x86/linux-2.6-x86
git-bisect bad dd430ca20c40ecccd6954a7efd13d4398f507728
# good: [49914084e797530d9baaf51df9eda77babc98fa8] Linux 2.6.24
git-bisect good 49914084e797530d9baaf51df9eda77babc98fa8
# good: [f4d900a2cae94256f56be7769734100c7054bf00] [NETLINK]: Mark
attribute construction exception unlikely
git-bisect good f4d900a2cae94256f56be7769734100c7054bf00
# good: [85040bcb4643cba578839e953f25e2d1965d83d0] [IPV6] ADDRLABEL: Fix
double free on label deletion.
git-bisect good 85040bcb4643cba578839e953f25e2d1965d83d0
# good: [70fd93c9d9e07dd2ea4465df76a0ddd414fd21ac] x86: export cpu_gdt_descr
git-bisect good 70fd93c9d9e07dd2ea4465df76a0ddd414fd21ac
# bad: [306c142f99caef5a30a0d61a39a956c069941b63] x86: pat: e820 cleanup
git-bisect bad 306c142f99caef5a30a0d61a39a956c069941b63
# bad: [7271339347653f6d652bb47aaee1f57d936002f6] x86:
arch/x86/kernel/cpu/mcheck/k7.c checkpatch fixes
git-bisect bad 7271339347653f6d652bb47aaee1f57d936002f6
# bad: [7818a1e0294debee02d5135e17b89f28b8871887] x86: provide 64-bit
with a load_sp0 function.
git-bisect bad 7818a1e0294debee02d5135e17b89f28b8871887
# bad: [2215e69d2cf5024647f9a034807990590d25dd4e] x86 boot: use E820
memory map on EFI 32 platform
git-bisect bad 2215e69d2cf5024647f9a034807990590d25dd4e
# good: [26048d75e8d6c840742468667f4a7ab8c2df74c9] x86: use the same
data type for tls_array.
git-bisect good 26048d75e8d6c840742468667f4a7ab8c2df74c9
# good: [341d8854d38fe6e7655f7748babae5c25a51aa70] x86: remove fastcall
from include/asm-x86
git-bisect good 341d8854d38fe6e7655f7748babae5c25a51aa70
# good: [8c8b8859b64baf6d7c33900e8720c7bafe775b2c] mcheck mce_64:
mce_read_sem to mutex
git-bisect good 8c8b8859b64baf6d7c33900e8720c7bafe775b2c
# good: [de18c850af701ac9512b7239e88fa45e4c168771] x86: EFI runtime
service support: EFI runtime services
git-bisect good de18c850af701ac9512b7239e88fa45e4c168771
# bad: [e429795c68d3001ecae74f6465420c9f043b0ece] x86: EFI runtime
service support: remove duplicated code from efi_32.c
git-bisect bad e429795c68d3001ecae74f6465420c9f043b0ece
# good: [9ad65e4748f55e3159283d7fa9d54fb30c086113] x86: EFI runtime
service support: document for EFI runtime services
git-bisect good 9ad65e4748f55e3159283d7fa9d54fb30c086113
I can not revert the commit e429795c68d3001ecae74f6465420c9f043b0ece. it
gave me errors.
but i'm also not sure what could be wrong with this commit!
my first idea was that the commit
"[2215e69d2cf5024647f9a034807990590d25dd4e] x86 boot: use E820 memory
map on EFI 32 platform" is the offender, because the macbook pro doesn't
provide a e820 memory map at all. but bisect showed that i'm wrong here?!
confused greeting
thomas
Thomas Meyer wrote:
>
> I can not revert the commit e429795c68d3001ecae74f6465420c9f043b0ece. it
> gave me errors.
> but i'm also not sure what could be wrong with this commit!
>
> my first idea was that the commit
> "[2215e69d2cf5024647f9a034807990590d25dd4e] x86 boot: use E820 memory
> map on EFI 32 platform" is the offender, because the macbook pro doesn't
> provide a e820 memory map at all. but bisect showed that i'm wrong here?!
>
ELILO is expected to convert the memory map from EFI to e820 format.
EFI32 is likely to be poorly supported, since Apple is likely to be the
only platform to *ever* ship EFI32 in mass production, and it has a BIOS
emulation layer available.
-hpa
H. Peter Anvin schrieb:
> Thomas Meyer wrote:
>>
>> I can not revert the commit e429795c68d3001ecae74f6465420c9f043b0ece.
>> it gave me errors.
>> but i'm also not sure what could be wrong with this commit!
>>
>> my first idea was that the commit
>> "[2215e69d2cf5024647f9a034807990590d25dd4e] x86 boot: use E820 memory
>> map on EFI 32 platform" is the offender, because the macbook pro
>> doesn't provide a e820 memory map at all. but bisect showed that i'm
>> wrong here?!
>>
>
> ELILO is expected to convert the memory map from EFI to e820 format.
I'm using elilo-3.6-ia32.efi provided at the elilo homepage. i wasn't
able to create a working efi toolchain :-(
Does elilo-3.6 already do the converting of the memory map, or not?
> EFI32 is likely to be poorly supported, since Apple is likely to be
> the only platform to *ever* ship EFI32 in mass production, and it has
> a BIOS emulation layer available.
What do you want so say? Does linux support EFI or not?
>
> -hpa
>
mfg
thomas
Thomas Meyer wrote:
> H. Peter Anvin schrieb:
>> Thomas Meyer wrote:
>>>
>>> I can not revert the commit e429795c68d3001ecae74f6465420c9f043b0ece.
>>> it gave me errors.
>>> but i'm also not sure what could be wrong with this commit!
>>>
>>> my first idea was that the commit
>>> "[2215e69d2cf5024647f9a034807990590d25dd4e] x86 boot: use E820 memory
>>> map on EFI 32 platform" is the offender, because the macbook pro
>>> doesn't provide a e820 memory map at all. but bisect showed that i'm
>>> wrong here?!
>>>
>>
>> ELILO is expected to convert the memory map from EFI to e820 format.
> I'm using elilo-3.6-ia32.efi provided at the elilo homepage. i wasn't
> able to create a working efi toolchain :-(
> Does elilo-3.6 already do the converting of the memory map, or not?
Don't know.
>> EFI32 is likely to be poorly supported, since Apple is likely to be
>> the only platform to *ever* ship EFI32 in mass production, and it has
>> a BIOS emulation layer available.
> What do you want so say? Does linux support EFI or not?
I'm saying that in my opinion support for EFI32 is likely to be
unreliable and never make it out of "experimental" status. EFI is
enough of a heap of junk on the 64-bit side.
Ying, do you have any insights?
-hpa
H. Peter Anvin schrieb:
> Thomas Meyer wrote:
>> H. Peter Anvin schrieb:
>>> Thomas Meyer wrote:
>>>>
>>>> I can not revert the commit
>>>> e429795c68d3001ecae74f6465420c9f043b0ece. it gave me errors.
>>>> but i'm also not sure what could be wrong with this commit!
>>>>
>>>> my first idea was that the commit
>>>> "[2215e69d2cf5024647f9a034807990590d25dd4e] x86 boot: use E820
>>>> memory map on EFI 32 platform" is the offender, because the macbook
>>>> pro doesn't provide a e820 memory map at all. but bisect showed
>>>> that i'm wrong here?!
>>>>
>>>
>>> ELILO is expected to convert the memory map from EFI to e820 format.
>> I'm using elilo-3.6-ia32.efi provided at the elilo homepage. i
>> wasn't able to create a working efi toolchain :-(
>> Does elilo-3.6 already do the converting of the memory map, or not?
>
> Don't know.
3.6 doesn't do the mapping in ia32/system.c create_boot_params().
3.8 does.
The problem is, that i wasn't able to produce a working elilo.efi files
with my gentoo toolchain. the files elilo-3.8.efi provided by the elilo
project also doesn't work (gives a load error from the efi runtime).
So i guess i'm stuck here.
It should be noted somewhere that with Huangs patches, elilo version 3.8
is needed.
On Sat, 2008-02-09 at 19:39 +0100, Thomas Meyer wrote:
> H. Peter Anvin schrieb:
> > Thomas Meyer wrote:
> >> H. Peter Anvin schrieb:
> >>> Thomas Meyer wrote:
> >>>>
> >>>> I can not revert the commit
> >>>> e429795c68d3001ecae74f6465420c9f043b0ece. it gave me errors.
> >>>> but i'm also not sure what could be wrong with this commit!
> >>>>
> >>>> my first idea was that the commit
> >>>> "[2215e69d2cf5024647f9a034807990590d25dd4e] x86 boot: use E820
> >>>> memory map on EFI 32 platform" is the offender, because the macbook
> >>>> pro doesn't provide a e820 memory map at all. but bisect showed
> >>>> that i'm wrong here?!
> >>>>
> >>>
> >>> ELILO is expected to convert the memory map from EFI to e820 format.
> >> I'm using elilo-3.6-ia32.efi provided at the elilo homepage. i
> >> wasn't able to create a working efi toolchain :-(
> >> Does elilo-3.6 already do the converting of the memory map, or not?
> >
> > Don't know.
> 3.6 doesn't do the mapping in ia32/system.c create_boot_params().
> 3.8 does.
>
> The problem is, that i wasn't able to produce a working elilo.efi files
> with my gentoo toolchain. the files elilo-3.8.efi provided by the elilo
> project also doesn't work (gives a load error from the efi runtime).
>
> So i guess i'm stuck here.
>
> It should be noted somewhere that with Huangs patches, elilo version 3.8
> is needed.
>
The release notes with elilo3.7 beta (first integration of Intel EFI 2
patches) does cover this, as well as the toolchain changes required to
build it. fyi
-Jason
On Sat, 2008-02-09 at 19:39 +0100, Thomas Meyer wrote:
> H. Peter Anvin schrieb:
> > Thomas Meyer wrote:
> >> H. Peter Anvin schrieb:
> >>> Thomas Meyer wrote:
> >>>>
> >>>> I can not revert the commit
> >>>> e429795c68d3001ecae74f6465420c9f043b0ece. it gave me errors.
> >>>> but i'm also not sure what could be wrong with this commit!
> >>>>
> >>>> my first idea was that the commit
> >>>> "[2215e69d2cf5024647f9a034807990590d25dd4e] x86 boot: use E820
> >>>> memory map on EFI 32 platform" is the offender, because the macbook
> >>>> pro doesn't provide a e820 memory map at all. but bisect showed
> >>>> that i'm wrong here?!
> >>>>
> >>>
> >>> ELILO is expected to convert the memory map from EFI to e820 format.
> >> I'm using elilo-3.6-ia32.efi provided at the elilo homepage. i
> >> wasn't able to create a working efi toolchain :-(
> >> Does elilo-3.6 already do the converting of the memory map, or not?
> >
> > Don't know.
> 3.6 doesn't do the mapping in ia32/system.c create_boot_params().
> 3.8 does.
>
> The problem is, that i wasn't able to produce a working elilo.efi files
> with my gentoo toolchain. the files elilo-3.8.efi provided by the elilo
> project also doesn't work (gives a load error from the efi runtime).
>
> So i guess i'm stuck here.
Can you attach the error message with elilo 3.8?
> It should be noted somewhere that with Huangs patches, elilo version 3.8
> is needed.
Yes. I will add this.
Best Regards,
Huang Ying