eboot.c regression allows boot, from v3.7 on- no boot with UEFI, only legacy, bisect results

eboot.c regression allows boot, from v3.7 on- no boot with UEFI, only legacy, bisect results

[Daniel McLellan]

System: Hewlett-Packard product: HP EliteBook 2570p
CPU~Dual core Intel Core i7-3520M CPU (-HT-MCP-)
Kernel: 3.12.4-2-hplove x86_64 (64 bit) Desktop: Enlightenment 0.18.99.18202
Distro: Arch Linux


I believe this issue is common to only a few iterations of ivy bridge

https://bbs.archlinux.org/viewtopic.php?id=156670
https://bugs.archlinux.org/task/33745?project=1

I had been patching eboot.c to v3.7 with the attached
efi_hp_boot.patch, but due to the recent changes in eboot.c (in v3.13)
I decided to bisect, and identified the following:

# first bad commit: [dd5fc854de5fd37adfcef8a366cd21a55aa01d3d] EFI:
Stash ROMs if they're not in the PCI BAR

My next step is diffing the eboot.c from v3.13.

or would a better path be to check for ivy bridge?

Re: eboot.c regression allows boot, from v3.7 on- no boot with UEFI, only legacy, bisect results

[Matt Fleming]

(Adding linux-efi and correct Matthew's address)

On Fri, 28 Mar, at 04:08:36PM, Daniel McLellan wrote:
> System: Hewlett-Packard product: HP EliteBook 2570p
> CPU~Dual core Intel Core i7-3520M CPU (-HT-MCP-)
> Kernel: 3.12.4-2-hplove x86_64 (64 bit) Desktop: Enlightenment 0.18.99.18202
> Distro: Arch Linux
> 
> 
> I believe this issue is common to only a few iterations of ivy bridge
> 
> https://bbs.archlinux.org/viewtopic.php?id=156670
> https://bugs.archlinux.org/task/33745?project=1
> 
> I had been patching eboot.c to v3.7 with the attached
> efi_hp_boot.patch, but due to the recent changes in eboot.c (in v3.13)
> I decided to bisect, and identified the following:
> 
> # first bad commit: [dd5fc854de5fd37adfcef8a366cd21a55aa01d3d] EFI:
> Stash ROMs if they're not in the PCI BAR
> 
> My next step is diffing the eboot.c from v3.13.
> 
> or would a better path be to check for ivy bridge?

Daniel, could you try booting a recent kernel, without your patch
applied and with the earlyprintk=efi,keep kernel parameter.

Meanwhile I'm going to try and reproduce this error.

-- 
Matt Fleming, Intel Open Source Technology Center

Re: eboot.c regression allows boot, from v3.7 on- no boot with UEFI, only legacy, bisect results

[Matt Fleming]

(Adding addresses for real this time... )

On Thu, 03 Apr, at 10:03:42PM, Matt Fleming wrote:
> (Adding linux-efi and correct Matthew's address)
> 
> On Fri, 28 Mar, at 04:08:36PM, Daniel McLellan wrote:
> > System: Hewlett-Packard product: HP EliteBook 2570p
> > CPU~Dual core Intel Core i7-3520M CPU (-HT-MCP-)
> > Kernel: 3.12.4-2-hplove x86_64 (64 bit) Desktop: Enlightenment 0.18.99.18202
> > Distro: Arch Linux
> > 
> > 
> > I believe this issue is common to only a few iterations of ivy bridge
> > 
> > https://bbs.archlinux.org/viewtopic.php?id=156670
> > https://bugs.archlinux.org/task/33745?project=1
> > 
> > I had been patching eboot.c to v3.7 with the attached
> > efi_hp_boot.patch, but due to the recent changes in eboot.c (in v3.13)
> > I decided to bisect, and identified the following:
> > 
> > # first bad commit: [dd5fc854de5fd37adfcef8a366cd21a55aa01d3d] EFI:
> > Stash ROMs if they're not in the PCI BAR
> > 
> > My next step is diffing the eboot.c from v3.13.
> > 
> > or would a better path be to check for ivy bridge?
> 
> Daniel, could you try booting a recent kernel, without your patch
> applied and with the earlyprintk=efi,keep kernel parameter.
> 
> Meanwhile I'm going to try and reproduce this error.
> 
> -- 
> Matt Fleming, Intel Open Source Technology Center

-- 
Matt Fleming, Intel Open Source Technology Center

Re: eboot.c regression allows boot, from v3.7 on- no boot with UEFI, only legacy, bisect results

[Matt Fleming]

(Adding addresses for real this time... )

On Thu, 03 Apr, at 10:03:42PM, Matt Fleming wrote:
> (Adding linux-efi and correct Matthew's address)
> 
> On Fri, 28 Mar, at 04:08:36PM, Daniel McLellan wrote:
> > System: Hewlett-Packard product: HP EliteBook 2570p
> > CPU~Dual core Intel Core i7-3520M CPU (-HT-MCP-)
> > Kernel: 3.12.4-2-hplove x86_64 (64 bit) Desktop: Enlightenment 0.18.99.18202
> > Distro: Arch Linux
> > 
> > 
> > I believe this issue is common to only a few iterations of ivy bridge
> > 
> > https://bbs.archlinux.org/viewtopic.php?id=156670
> > https://bugs.archlinux.org/task/33745?project=1
> > 
> > I had been patching eboot.c to v3.7 with the attached
> > efi_hp_boot.patch, but due to the recent changes in eboot.c (in v3.13)
> > I decided to bisect, and identified the following:
> > 
> > # first bad commit: [dd5fc854de5fd37adfcef8a366cd21a55aa01d3d] EFI:
> > Stash ROMs if they're not in the PCI BAR
> > 
> > My next step is diffing the eboot.c from v3.13.
> > 
> > or would a better path be to check for ivy bridge?
> 
> Daniel, could you try booting a recent kernel, without your patch
> applied and with the earlyprintk=efi,keep kernel parameter.
> 
> Meanwhile I'm going to try and reproduce this error.
> 
> -- 
> Matt Fleming, Intel Open Source Technology Center

-- 
Matt Fleming, Intel Open Source Technology Center