T400 screen issue / EDID handling

Hello,

I allready contacted Francis from libreboot and he forwared me to this 
list, because it seems to be an issue regarding coreboot.

I have flashed a Lenovo T400 with libreboot.

The T400 has an Intel GPU and I am able to use it with an external 
display connected to the onboard vga-port.

The issue is, that I can't use the internal display, which is an 
1440x900 Samusung LTN141WD-L05, EDID dump:

https://paste.debian.net/plainh/b3699c60

It seems like the internal screen is powered on while booting but it 
stays black.

Francis assumes that this is due to bugs in how coreboot handles the 
EDID.

Here is an EDID dump of a working display:

http://paste.debian.net/plainh/776705d5

Anyone is able to help?

Thanks in advance!

[ron minnich]

Gee, what fun. There are lots of things that can go wrong, to get some
idea, see the
display handling for HP Falco, in which we managed 4 different types of
displays.

I'll try to take a look later, but just keep in mind that the displays are
frequently not all the same
on different production runs of a laptop so something that works fine on
one gives black
screen on another ...

ron

On Fri, Jul 3, 2015 at 7:35 AM  wrote:
> Hello,
>
> I allready contacted Francis from libreboot and he forwared me to this
> list, because it seems to be an issue regarding coreboot.
>
> I have flashed a Lenovo T400 with libreboot.
>
> The T400 has an Intel GPU and I am able to use it with an external
> display connected to the onboard vga-port.
>
> The issue is, that I can't use the internal display, which is an
> 1440x900 Samusung LTN141WD-L05, EDID dump:
>
> https://paste.debian.net/plainh/b3699c60
>
> It seems like the internal screen is powered on while booting but it
> stays black.
>
> Francis assumes that this is due to bugs in how coreboot handles the
> EDID.
>
> Here is an EDID dump of a working display:
>
> http://paste.debian.net/plainh/776705d5
>
> Anyone is able to help?
>
> Thanks in advance!

[Paul Menzel]

Dear Kroms,

Am Freitag, den 03.07.2015, 16:27 +0200 schrieb [e-mail]:
> I allready contacted Francis from libreboot and he forwared me to 
> this list, because it seems to be an issue regarding coreboot.

welcome and thank you for writing a message to the mailing list.
> I have flashed a Lenovo T400 with libreboot.
>
> The T400 has an Intel GPU and I am able to use it with an external 
> display connected to the onboard vga-port.
>
> The issue is, that I can't use the internal display, which is an 
> 1440x900 Samusung LTN141WD-L05, EDID dump:
>
> https://paste.debian.net/plainh/b3699c60

Please attach such information in the future instead of pasting it.
> It seems like the internal screen is powered on while booting but it 
> stays black.
>
> Francis assumes that this is due to bugs in how coreboot handles the 
> EDID.
>
> Here is an EDID dump of a working display:
>
> http://paste.debian.net/plainh/776705d5

Ditto.
> Anyone is able to help?

Please send us your configuration (.config of coreboot and the payload)
and the log files from coreboot (cbmem -c), the payload and Linux.
That’s one of the advantages of using coreboot, that you have logs.
Please also provide the output of `build/cbfstool build/coreboot.rom
print`.

Thanks,

Paul

Dear Paul, dear Ron,

thank you for your messages!

I used a pre-compiled libreboot ROM image for flashing my T400:

t400_8mb_ukqwerty_vesafb.rom

(source: http://mirrors.mit.edu/libreboot/20150518/rom/)

I attached most of the requested information, I hope I found the right 
information:

requested information -> attached file or remark

.config of coreboot -> config_t400_8mb_libreboot

and the payload -> the 4 files beginning with grub_

log files from coreboot (cbmem -c) -> cbmem_-c

the payload -> I dont know to get or find it.

and Linux -> Do you mean the output of dmesg?

output of `build/cbfstool build/coreboot.rom print -> 
cbfstool_t400_8mb_ukqwerty_vesafb.rom (I used cbfstool on the used 
pre-compiled rom)

I attached the EDID dumps from my non working and a working display as 
well.

I hope this helps, if you need further information please let me know 
and I will try to provide them.

If you still need the logs files from the payload and Linux please give 
me the commands I have to execute. I am kind of a noob using Linux.

Thanks again an best regards,

Kroms

Am 03.07.2015 23:19 schrieb Paul Menzel:
> Dear Kroms,
>
> Am Freitag, den 03.07.2015, 16:27 +0200 schrieb [e-mail]:
> > I allready contacted Francis from libreboot and he forwared me to
> > this list, because it seems to be an issue regarding coreboot.
>
> welcome and thank you for writing a message to the mailing list.
> > I have flashed a Lenovo T400 with libreboot.
> >
> > The T400 has an Intel GPU and I am able to use it with an external
> > display connected to the onboard vga-port.
> >
> > The issue is, that I can't use the internal display, which is an
> > 1440x900 Samusung LTN141WD-L05, EDID dump:
> >
> > https://paste.debian.net/plainh/b3699c60
>
> Please attach such information in the future instead of pasting it.
> > It seems like the internal screen is powered on while booting but it
> > stays black.
> >
> > Francis assumes that this is due to bugs in how coreboot handles the
> > EDID.
> >
> > Here is an EDID dump of a working display:
> >
> > http://paste.debian.net/plainh/776705d5
>
> Ditto.
> > Anyone is able to help?
>
> Please send us your configuration (.config of coreboot and the payload)
> and the log files from coreboot (cbmem -c), the payload and Linux.
> That’s one of the advantages of using coreboot, that you have logs.
> Please also provide the output of `build/cbfstool build/coreboot.rom
> print`.
>
> Thanks,
>
> Paul

[Paul Menzel]

Dear Kroms,

Am Samstag, den 04.07.2015, 11:11 +0200 schrieb [e-mail]:
> I used a pre-compiled libreboot ROM image for flashing my T400:
>
> t400_8mb_ukqwerty_vesafb.rom
>
> (source: http://mirrors.mit.edu/libreboot/20150518/rom/)
>
> I attached most of the requested information, I hope I found the 
> right information:
>
> requested information -> attached file or remark
>
> .config of coreboot -> config_t400_8mb_libreboot
>
> and the payload -> the 4 files beginning with grub_
>
> log files from coreboot (cbmem -c) -> cbmem_-c

Thank you. Unfortunately, I was unable to see anything suspicious in
there. Libreboot seems to strip of the coreboot revision it uses from
the banner, which is not good. It might be in the revision file in
CBFS.

build/cbfstool build/coreboot.rom extract -n revision -f revision.txt
> the payload -> I dont know to get or find it.

It is GRUB. Unfortunately, I do not see GRUB’s debug messages in CBMEM.
> and Linux -> Do you mean the output of dmesg?

Indeed, that’s one way of getting them. Please provide the output.

[…]

Thanks,

Paul

PS: It’s quite common to use interleaved posting style [1] on mailing
lists.

[1] https://en.wikipedia.org/wiki/Posting_style

[Francis Rowe]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
> Thank you. Unfortunately, I was unable to see anything suspicious
> in there. Libreboot seems to strip of the coreboot revision it uses
> from the banner, which is not good. It might be in the revision
> file in CBFS.

Libreboot builds coreboot without the .git directory, so it's not
possible to get the coreboot revision. You can find out what version
of coreboot it is by looking at
resources/scripts/helpers/download/coreboot

He is using libreboot 20150518, so the coreboot revision that he's
using is e19c8b0091022ae3f490601aed0c290cd5171b79

See
http://git.savannah.gnu.org/cgit/libreboot.git/tree/resources/scripts/helper...
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1

iQEcBAEBAgAGBQJVmUXaAAoJEP9Ft0z50c+UzDwIAIvUK5lpPHveEKdoGeHRIvZ/
wrfTuzqtM2snIT2baypZTpcirxCfcLonr4dntKG4VkL5edxzWTjWpPF855vLyCeQ
gLBdGjHlZ+NTyZZDvkPoHJNGaPhoYUaUk/2L6GafUgYBRgxWcOjJs5xyfHk1VU/L
BcXFztVKfPM2BDWW0nue8ycdxzo3fTfAwNdCCCFBaypi22eEQXqEt4WNjAxlKbwE
Y1VxUn7klMDqP0HYCtWDr2Gmgw58LirZ2ZCq+Q00oae8ZXYIJzpp/qM6si8n25HI
lzuU9AHIzDiMQ/hGAuU0J6VOK9uHw+DTgFjWyCgu69DgInA9mfihoQbLd+ro/Lo=
=1+ZG
-----END PGP SIGNATURE-----

[Francis Rowe]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 05/07/15 15:57, Francis Rowe wrote:
> > Thank you. Unfortunately, I was unable to see anything
> > suspicious in there. Libreboot seems to strip of the coreboot
> > revision it uses from the banner, which is not good. It might be
> > in the revision file in CBFS.
>
> Libreboot builds coreboot without the .git directory, so it's not 
> possible to get the coreboot revision. You can find out what
> version of coreboot it is by looking at 
> resources/scripts/helpers/download/coreboot
>
> He is using libreboot 20150518, so the coreboot revision that he's 
> using is e19c8b0091022ae3f490601aed0c290cd5171b79
>
> See 
> http://git.savannah.gnu.org/cgit/libreboot.git/tree/resources/scripts/helper...

Note: this bug also affects coreboot upstream, using no additional
patches, from the latest version in the "master" branch from
http://review.coreboot.org/coreboot
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1

iQEcBAEBAgAGBQJVmUZmAAoJEP9Ft0z50c+UGw8IAMCziSv5jfp078b9lokEapUb
kuWPtXAGPrAaZN92kaQRRlGPcp5OF0vAJxhpBPQlMBzl7taizhCb0+BSXKVncusY
GL/eXc2eKRt/sYRWCJxM8Eu3uukUH3ybhHLubR/kOmQ1gi2fuzYXbo+qb2ViJ5dZ
+wF+lyIHxj9RxCEY2TvuLO7c0MjdduC3GdUmlp9am7P+5qWbFr22lok0VznkZ54B
QKorLCxrLPrZnmpZJr1rL0UfkgGZ6uurtK6kbMYhdIsOXTlvgvuu22lnjqKr9fJ2
EwcLWAM9ywq8Aph4HyhEMcVCbj2izVdAURTsK7vGUAwaV7R9s2yfVCLcWmr7mXQ=
=71Qg
-----END PGP SIGNATURE-----

Dear Paul,

thank you for looking into the files!
> It might be in the revision file in CBFS.
> build/cbfstool build/coreboot.rom extract -n revision -f 
>
> revision.txt

Please find attached the revision.txt.
> > and Linux -> Do you mean the output of dmesg?
>
> Indeed, that’s one way of getting them. Please provide the output.

It is also attached. I hope the information will help you to get closer 
to the solution.
> PS: It’s quite common to use interleaved posting style [1] on mailing
> lists.
>
> [1] https://en.wikipedia.org/wiki/Posting_style

Thank you for the hint. I will do so, the next time I am posting a link 
to a mailing-list.

Best Regards,

Kroms

[Alexander Couzens]

On Sun, 05 Jul 2015 15:57:30 +0100
Francis Rowe  wrote:
> Libreboot builds coreboot without the .git directory, so it's not
> possible to get the coreboot revision. You can find out what version

I don't understand why you're removing the .git directory.
But anyhow, you should patch the coreboot version script. The script should
take your specific version of your libreboot build script git repo.
Please prefix the version with "libreboot", it makes all our live easier to know what version is running and from where the version comes from.
And btw. this makes your coreboot' build reproducible, which isn't it atm.

Best
lynxis

[Francis Rowe]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
> I don't understand why you're removing the .git directory. But
> anyhow, you should patch the coreboot version script. The script
> should take your specific version of your libreboot build script
> git repo. Please prefix the version with "libreboot", it makes all
> our live easier to know what version is running and from where the
> version comes from. And btw. this makes your coreboot' build
> reproducible, which isn't it atm.

Libreboot removes all of the blobs from coreboot. In doing so, a diff
showing those deletions (therefore, containing the blobs) is shown.
The git history in coreboot also shows blobs that have been added or
deleted, over time.

By not deleting the .git directory, libreboot would be distributing
blobs. That's why it's deleted.

However, your comment is valid; I should patch the build system in
coreboot so that it shows libreboot-<insert revision here> in the
coreboot-libre boot logs (as seen from serial output, EHCI debug
output or cbmem -c). I've added a note for myself, reminding me to do
this.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1

iQEcBAEBAgAGBQJVmZtuAAoJEP9Ft0z50c+UW3AH+QGjaH5w/rIEWhKCYvXUDFOj
U6yNgGFT8kkXEWaS+vZY1iVrDdIGbEx9SM27ZXroUWLlIS90XiHyzYGqztJgzfrB
3XaYFI17WYwpoQ5f3OGyF8uycc1JlQQQ8zs7xDZywNNVYR4pJUkXHleaYEOb9k3U
GJ4PdO22fqTcyZH5XZXk0ga0oZp8N3ChLJgA2gy7L0fGfX73VxQL0CvQ/bMdMsOG
hERR4wf7cy4Z6BV1p65lt4Hi60kbEiQ5JpTBawQ80xGUCq2yVyxjJDrY+Pn/1W2l
vHmyeM2iBpAroUEZ0f2XnWH8KDEAkoMiM1H6vu6ZHybGrK/JNqf+Le9AXA5oWfo=
=p9+b
-----END PGP SIGNATURE-----

[Francis Rowe]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 03/07/15 15:27, [e-mail] wrote:
> Hello,
>
> I allready contacted Francis from libreboot and he forwared me to
> this list, because it seems to be an issue regarding coreboot.
>
> I have flashed a Lenovo T400 with libreboot.
>
> The T400 has an Intel GPU and I am able to use it with an external 
> display connected to the onboard vga-port.
>
> The issue is, that I can't use the internal display, which is an 
> 1440x900 Samusung LTN141WD-L05, EDID dump:
>
> https://paste.debian.net/plainh/b3699c60
>
> It seems like the internal screen is powered on while booting but
> it stays black.
>
> Francis assumes that this is due to bugs in how coreboot handles
> the EDID.
>
> Here is an EDID dump of a working display:
>
> http://paste.debian.net/plainh/776705d5
>
> Anyone is able to help?
>
> Thanks in advance!

Dear coreboot community,

tty0_ in #coreboot/#libreboot found a partial fix.

See:
http://libreboot.org/docs/hcl/gm45_lcd.html
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1

iQEcBAEBAgAGBQJVobEOAAoJEP9Ft0z50c+UJSgH/3tj9DiBbitS1eCgJ98743Ij
sARuL4cFbUNMeT6nZmnhMEmv+O8Na9SO2yuvVb/iqcYLzbQc6gGmD8wzIln77c+c
jXZ+Tje7RSwwCJvt3gDtXG1UTvnDGthJy+qUh5JqQxyNljSNFUQNCpkZw/vhAgJa
aXcxO8w1W7eFRO5ueTR0HqF57JQXjyNBTicfCKqRpRJ/ac93ipbuKQBAIUywjjdD
+LD3a31RfEjmd8YpAzEXbvUhl3b4NWTuy0nqQj9amaQGeJKycNSAYisLKqriPZXe
r0WW9mfSUWWPTYpTtRREK8uBkKmRF0qMLKZES2B9SiDsZ3RcF0+CRAP+6UTwP+Y=
=gxaG
-----END PGP SIGNATURE-----

Please turn on CBMEM console for payloads in Libreboot (was: T400 screen issue / EDID handling)

[Paul Menzel]

Dear Libreboot folks,

Am Sonntag, den 05.07.2015, 09:42 +0200 schrieb Paul Menzel:
> Am Samstag, den 04.07.2015, 11:11 +0200 schrieb [e-mail]:

[…]
> > the payload -> I dont know to get or find it.
>
> It is GRUB. Unfortunately, I do not see GRUB’s debug messages in 
> CBMEM.

Could you please enable logging to CBMEM console in your payloads? For
GRUB you’d need to enable the CBMEM console terminal by adding the line
below below your terminal setting in your `etc/grub.cfg` in CBFS.

terminal_output --append cbmemc

Maybe Vladimir has some other suggestions.

Thanks,

Paul

[Paul Menzel]

Dear Kroms,

I am sorry for the long delay.

Am Sonntag, den 05.07.2015, 17:05 +0200 schrieb [e-mail]:
> > It might be in the revision file in CBFS.
> > build/cbfstool build/coreboot.rom extract -n revision -f revision.txt
>
> Please find attached the revision.txt.

Thank you! Unfortunately, it also does not contain the Git revision.

#define COREBOOT_ORIGIN_GIT_REVISION "Unknown"
> > > and Linux -> Do you mean the output of dmesg?
> >
> > Indeed, that’s one way of getting them. Please provide the output.
>
> It is also attached. I hope the information will help you to get 
> closer to the solution.

Thank you! Unfortunately, no useful i915 message are in there. Three
things to try and do.

1. Try to use a newer Linux kernel. Trying to fix the Linux kernel is
much easier and more likely, if the reports are against the latest
version. Do you have a spare disk with a newer Linux kernel
installed?Debian Sid/unstable currently ships Linux 4.0.8. Linux 4.1
could be installed from the Debian experimental archives.

2. Turn on debugging of the Linux i915 module [1].
> To obtain a dmesg with debug information, add drm.debug=0xe to your
> kernel command line or, if you have display issues with kernel 4.1
> or newer, use drm.debug=0x1e log_buf_len=1M instead. Then, reboot
> and reproduce the issue again. Make sure to attach the full dmesg
> all the way from boot.

3. Please also post a log with debugging enabled when using the vendor
firmware.

Thanks,

Paul

[1] https://01.org/linuxgraphics/documentation/how-report-bugs

Dear Paul,
> I am sorry for the long delay.

No problem, I am very happy that you are still trying to help!
> Thank you! Unfortunately, it also does not contain the Git revision.
> #define COREBOOT_ORIGIN_GIT_REVISION "Unknown"

I dont´t know if you noticed that Francis wrote in this thread on Sun 
Jul 5 16:57:30 CEST 2015:

"Libreboot builds coreboot without the .git directory, so it's not
possible to get the coreboot revision. You can find out what version
of coreboot it is by looking at
resources/scripts/helpers/download/coreboot

He is using libreboot 20150518, so the coreboot revision that he's
using is e19c8b0091022ae3f490601aed0c290cd5171b79"
> Try to use a newer Linux kernel.

I updated the Kernel to 4.1.2 from Debian experimental.
> Turn on debugging of the Linux i915 module [1].
> Please also post a log with debugging enabled when using the vendor
>
> firmware.

Please find attached the two outputs (libreboot and lenovo_bios) from 
dmesg with debugging enabled.

Thanks again,

Kroms

Please turn on CBMEM console for payloads in Libreboot

[Francis Rowe]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 12/07/15 10:21, Paul Menzel wrote:
> Dear Libreboot folks,
>
> Am Sonntag, den 05.07.2015, 09:42 +0200 schrieb Paul Menzel:
> > Am Samstag, den 04.07.2015, 11:11 +0200 schrieb [e-mail]:
>
> […]
> > > the payload -> I dont know to get or find it.
> >
> > It is GRUB. Unfortunately, I do not see GRUB’s debug messages in
> >  CBMEM.
>
> Could you please enable logging to CBMEM console in your payloads?
> For GRUB you’d need to enable the CBMEM console terminal by adding
> the line below below your terminal setting in your `etc/grub.cfg`
> in CBFS.
>
> terminal_output --append cbmemc

Done.

http://git.savannah.gnu.org/cgit/libreboot.git/commit/?id=08b4126023fa16f1d4...
> Maybe Vladimir has some other suggestions.

I will ask him.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1

iQEcBAEBAgAGBQJVotRBAAoJEP9Ft0z50c+U7VEIAIFwkPVg9J3qtkOlFj6AMR7x
gLXZjbZKIiFF67RI36eEJDWOzeMm8mYz4ZrijfZpOYcLQ0xkvsBGXJNtPQ18uMC5
OG+BEYf5Cw1yCR8IVAmIjGl2DjKo5Hr57HKTnU0q1mNNPKftnmgaWU0cCd1BGG5x
gjJSqY/0ZX8KSGPZMM1Oj163CtYsRJOv3JLPBugdsi8exyWeJja9MWSJrh7A81KW
V8/pFLvnEOMxyiSKVyd6M6i8zHQymmQIGOIizXPTUXVP+yigVbiDhQ57lviNvk8J
2QjJDEhzYQR1kam4OAVeZ2ONsLgY7KiTTsouxZeTGCrOwKw2KgIdPBtw7/TwRYA=
=kHvy
-----END PGP SIGNATURE-----

Syntax Highlighting for ACPI ASL, devicetree.cb and Kconfig files

[Martin Roth]

I've just posted syntax highlighting files for ASL, Kconfig and 
devicetree.cb for a few editors if anyone wants to try them out.

vim:
ASL - https://github.com/martinlroth/vim-acpi-asl
devicetree.cb - https://github.com/martinlroth/vim-devicetree
Kconfig is alredy supported in vim.

atom:
ASL - not done yet
devicetree.cb - https://github.com/martinlroth/language-devicetree
Kconfig - https://github.com/martinlroth/language-kconfig

Ultraedit:
ASL - https://github.com/martinlroth/wordfiles/blob/master/acpi.uew
devicetree.cb - 
https://github.com/martinlroth/wordfiles/blob/master/devicetree_cb.uew
Kconfig - https://github.com/martinlroth/wordfiles/blob/master/kconfig.uew

If there's interest in syntax highlighting for other editors,, let me 
know, and I'll see what I can do.

Martin

[Paul Menzel]

Dear Kroms,

Am Sonntag, den 12.07.2015, 22:17 +0200 schrieb [e-mail]:
> > Thank you! Unfortunately, it also does not contain the Git 
> > revision.
> > #define COREBOOT_ORIGIN_GIT_REVISION "Unknown"
>
> I dont´t know if you noticed that Francis wrote in this thread on Sun
>
> Jul 5 16:57:30 CEST 2015:
>
> "Libreboot builds coreboot without the .git directory, so it's not
> possible to get the coreboot revision. You can find out what version
> of coreboot it is by looking at
> resources/scripts/helpers/download/coreboot
>
> He is using libreboot 20150518, so the coreboot revision that he's
> using is e19c8b0091022ae3f490601aed0c290cd5171b79"

Thank you for looking this up!

commit e19c8b0091022ae3f490601aed0c290cd5171b79
    Author:     Alexander Couzens <    [e-mail]    >
    AuthorDate: Thu Apr 2 23:20:45 2015 +0200
    Commit:     Peter Stuge <    [e-mail]    >
    CommitDate: Sun Apr 5 04:07:25 2015 +0200

acpi: protect acpi generators from PRE_RAM & SMM
> > Try to use a newer Linux kernel.
>
> I updated the Kernel to 4.1.2 from Debian experimental.
> > Turn on debugging of the Linux i915 module [1].
> > Please also post a log with debugging enabled when using the
> >
> > vendor firmware.
>
> Please find attached the two outputs (libreboot and lenovo_bios) from
>
> dmesg with debugging enabled.

Awesome! Thank you very much.

$ grep 'drm:' dmesg_debug_libreboot.txt
    [   11.647383] [drm:i915_dump_device_info] i915 device info: gen=4, pciid=0x2a42 rev=0x07 flags=is_mobile,need_gfx_hws,is_g4x,has_fbc,has_pipe_cxsr,has_hotplug,supports_tv,
    [   11.647410] [drm:intel_detect_pch] No PCH found.
    [   11.647416] [drm:i915_gem_gtt_init] GMADR size = 256M
    [   11.647418] [drm:i915_gem_gtt_init] GTT stolen size = 32M
    [   11.647419] [drm:i915_gem_gtt_init] ppgtt mode: 0
    [   11.648615] [drm:intel_opregion_setup] graphic opregion physical addr: 0x0
    [   11.648618] [drm:intel_opregion_setup] ACPI OpRegion not supported!
    [   11.648665] [drm:intel_device_info_runtime_init] slice total: 0
    [   11.648666] [drm:intel_device_info_runtime_init] subslice total: 0
    [   11.648668] [drm:intel_device_info_runtime_init] subslice per slice: 0
    [   11.648670] [drm:intel_device_info_runtime_init] EU total: 0
    [   11.648671] [drm:intel_device_info_runtime_init] EU per subslice: 0
    [   11.648672] [drm:intel_device_info_runtime_init] has slice power gating: n
    [   11.648674] [drm:intel_device_info_runtime_init] has subslice power gating: n
    [   11.648676] [drm:intel_device_info_runtime_init] has EU power gating: n
    [   11.648680] [drm:init_vbt_defaults] Set default to SSC at 100000 kHz
    [   11.648684] [drm:validate_vbt] Using VBT from PCI ROM: $VBT IRONLAKE-MOBILEd
    [   11.648687] [drm:parse_general_features] BDB_GENERAL_FEATURES int_tv_support 0 int_crt_support 1 lvds_use_ssc 1 lvds_ssc_freq 96000 display_clock_mode 0 fdi_rx_polarity_inverted 0
    [   11.648689] [drm:parse_sdvo_device_mapping] No general definition block is found, unable to construct sdvo mapping.
    [   11.648691] [drm:parse_device_mapping] No general definition block is found, no devices defined.
    [   11.648693] [drm:parse_psr] No PSR BDB found.
    [   11.648700] [drm:intel_dsm_pci_probe] no _DSM method for intel device
    [   11.648707] [drm:i915_gem_init_stolen] found 33554432 bytes of stolen memory at be000000
    [   11.648710] [drm:intel_display_power_get] enabling always-on
    [   11.648744] [drm:intel_modeset_init] 2 display pipes available.
    [   11.648753] [drm:intel_modeset_init] pipe A sprite A init failed: -19
    [   11.648760] [drm:intel_modeset_init] pipe B sprite B init failed: -19
    [   11.718062] [drm:intel_lvds_init] using preferred mode from EDID: 
    [   11.718069] [drm:drm_mode_debug_printmodeline] Modeline 27:"1440x900" 60 101600 1440 1488 1520 1792 900 903 909 945 0x48 0xa
    [   11.718075] [drm:intel_lvds_init] detected single-link lvds configuration
    [   11.718147] [drm:intel_panel_setup_backlight] Connector LVDS-1 backlight initialized, enabled, brightness 98/98
    [   11.718205] [drm:intel_setup_outputs] probing SDVOB
    [   11.718209] [drm:intel_gmbus_force_bit] enabling bit-banging on i915 gmbus dpb. force bit now 1
    [   11.727609] [drm:intel_sdvo_read_byte] i2c transfer returned -6
    [   11.727611] [drm:intel_sdvo_init] No SDVO device found on SDVOB
    [   11.727642] [drm:intel_gmbus_force_bit] disabling bit-banging on i915 gmbus dpb. force bit now 0
    [   11.727644] [drm:intel_setup_outputs] probing HDMI on SDVOB
    [   11.727701] [drm:intel_dp_init_connector] Adding DP connector on port B
    [   11.727743] [drm:intel_dp_aux_init] registering DPDDC-B bus for card0-DP-1
    [   11.727781] [drm:intel_setup_outputs] probing SDVOC
    [   11.727783] [drm:intel_gmbus_force_bit] enabling bit-banging on i915 gmbus dpb. force bit now 1
    [   11.737209] [drm:intel_sdvo_read_byte] i2c transfer returned -6
    [   11.737212] [drm:intel_sdvo_init] No SDVO device found on SDVOC
    [   11.740675] [drm:intel_gmbus_force_bit] disabling bit-banging on i915 gmbus dpb. force bit now 0
    [   11.740681] [drm:intel_setup_outputs] probing HDMI on SDVOC
    [   11.740749] [drm:intel_dp_init_connector] Adding DP connector on port C
    [   11.740788] [drm:intel_dp_aux_init] registering DPDDC-C bus for card0-DP-2
    [   11.740838] [drm:intel_dp_init_connector] Adding DP connector on port D
    [   11.740879] [drm:intel_dp_aux_init] registering DPDDC-D bus for card0-DP-3
    [   11.740927] [drm:intel_modeset_readout_hw_state] [CRTC:20] hw state readout: enabled
    [   11.740930] [drm:intel_modeset_readout_hw_state] [CRTC:23] hw state readout: disabled
    [   11.740933] [drm:intel_modeset_readout_hw_state] [ENCODER:25:LVDS-25] hw state readout: enabled, pipe A
    [   11.740935] [drm:intel_modeset_readout_hw_state] [ENCODER:32:DAC-32] hw state readout: disabled, pipe A
    [   11.740938] [drm:intel_modeset_readout_hw_state] [ENCODER:33:TMDS-33] hw state readout: disabled, pipe A
    [   11.740941] [drm:intel_modeset_readout_hw_state] [ENCODER:38:TMDS-38] hw state readout: disabled, pipe A
    [   11.740943] [drm:intel_modeset_readout_hw_state] [ENCODER:40:TMDS-40] hw state readout: disabled, pipe A
    [   11.740945] [drm:intel_modeset_readout_hw_state] [ENCODER:42:TMDS-42] hw state readout: disabled, pipe A
    [   11.740948] [drm:intel_modeset_readout_hw_state] [ENCODER:44:TMDS-44] hw state readout: disabled, pipe A
    [   11.740950] [drm:intel_modeset_readout_hw_state] [CONNECTOR:24:LVDS-1] hw state readout: enabled
    [   11.740952] [drm:intel_modeset_readout_hw_state] [CONNECTOR:31:VGA-1] hw state readout: disabled
    [   11.740955] [drm:intel_modeset_readout_hw_state] [CONNECTOR:34:HDMI-A-1] hw state readout: disabled
    [   11.740957] [drm:intel_modeset_readout_hw_state] [CONNECTOR:39:DP-1] hw state readout: disabled
    [   11.740959] [drm:intel_modeset_readout_hw_state] [CONNECTOR:41:HDMI-A-2] hw state readout: disabled
    [   11.740961] [drm:intel_modeset_readout_hw_state] [CONNECTOR:43:DP-2] hw state readout: disabled
    [   11.740963] [drm:intel_modeset_readout_hw_state] [CONNECTOR:45:DP-3] hw state readout: disabled
    […]

With the vendor firmware the messages look like below.

$ grep 'drm:' dmesg_debug_lenovo_bios.txt
    [   13.567790] [drm:i915_dump_device_info] i915 device info: gen=4, pciid=0x2a42 rev=0x07 flags=is_mobile,need_gfx_hws,is_g4x,has_fbc,has_pipe_cxsr,has_hotplug,supports_tv,
    [   13.567817] [drm:intel_detect_pch] No PCH found.
    [   13.567822] [drm:i915_gem_gtt_init] GMADR size = 256M
    [   13.567824] [drm:i915_gem_gtt_init] GTT stolen size = 32M
    [   13.567825] [drm:i915_gem_gtt_init] ppgtt mode: 0
    [   13.569564] [drm:intel_opregion_setup] graphic opregion physical addr: 0xbdb6b0fa
    [   13.569573] [drm:intel_opregion_setup] Public ACPI methods supported
    [   13.569574] [drm:intel_opregion_setup] SWSCI supported
    [   13.575637] [drm:swsci_setup] SWSCI BIOS requested (00010673) SBCB callbacks that are not supported (00080673)
    [   13.575640] [drm:swsci_setup] SWSCI GBDA callbacks 00000cf3, SBCB callbacks 00010673
    [   13.575641] [drm:intel_opregion_setup] ASLE supported
    [   13.575697] [drm:intel_device_info_runtime_init] slice total: 0
    [   13.575698] [drm:intel_device_info_runtime_init] subslice total: 0
    [   13.575700] [drm:intel_device_info_runtime_init] subslice per slice: 0
    [   13.575701] [drm:intel_device_info_runtime_init] EU total: 0
    [   13.575702] [drm:intel_device_info_runtime_init] EU per subslice: 0
    [   13.575703] [drm:intel_device_info_runtime_init] has slice power gating: n
    [   13.575705] [drm:intel_device_info_runtime_init] has subslice power gating: n
    [   13.575706] [drm:intel_device_info_runtime_init] has EU power gating: n
    [   13.575711] [drm:init_vbt_defaults] Set default to SSC at 100000 kHz
    [   13.575714] [drm:validate_vbt] Using VBT from OpRegion: $VBT CANTIGA        d
    [   13.575716] [drm:parse_general_features] BDB_GENERAL_FEATURES int_tv_support 0 int_crt_support 1 lvds_use_ssc 1 lvds_ssc_freq 100000 display_clock_mode 0 fdi_rx_polarity_inverted 0
    [   13.575718] [drm:parse_general_definitions] crt_ddc_bus_pin: 2
    [   13.575720] [drm:parse_lfp_panel_data] DRRS supported mode is static
    [   13.575723] [drm:parse_lfp_panel_data] Found panel mode in BIOS VBT tables:
    [   13.575726] [drm:drm_mode_debug_printmodeline] Modeline 0:"1440x900" 0 101600 1440 1488 1520 1792 900 903 909 945 0x8 0xa
    [   13.575728] [drm:parse_lfp_panel_data] VBT initial LVDS value 4230033c
    [   13.575730] [drm:parse_lfp_backlight] VBT backlight PWM modulation frequency 220 Hz, active high, min brightness 0, level 0
    [   13.575732] [drm:parse_sdvo_panel_data] Found SDVO panel mode in BIOS VBT tables:
    [   13.575735] [drm:drm_mode_debug_printmodeline] Modeline 0:"1600x1200" 0 162000 1600 1664 1856 2160 1200 1201 1204 1250 0x8 0xa
    [   13.575737] [drm:parse_sdvo_device_mapping] No SDVO device info is found in VBT
    [   13.575739] [drm:parse_driver_features] DRRS State Enabled:0
    [   13.575740] [drm:parse_psr] No PSR BDB found.
    [   13.575750] [drm:intel_dsm_pci_probe] no _DSM method for intel device
    [   13.575760] [drm:i915_gem_init_stolen] found 33554432 bytes of stolen memory at be000000
    [   13.575763] [drm:intel_display_power_get] enabling always-on
    [   13.575801] [drm:intel_modeset_init] 2 display pipes available.
    [   13.575813] [drm:intel_modeset_init] pipe A sprite A init failed: -19
    [   13.575819] [drm:intel_modeset_init] pipe B sprite B init failed: -19
    [   13.646058] [drm:intel_lvds_init] using preferred mode from EDID: 
    [   13.646063] [drm:drm_mode_debug_printmodeline] Modeline 27:"1440x900" 60 101600 1440 1488 1520 1792 900 903 909 945 0x48 0xa
    [   13.646068] [drm:intel_lvds_init] detected dual-link lvds configuration
    [   13.646215] [drm:intel_panel_setup_backlight] Connector LVDS-1 backlight initialized, enabled, brightness 481695/2408475
    [   13.646327] [drm:intel_setup_outputs] probing SDVOB
    [   13.646331] [drm:intel_gmbus_force_bit] enabling bit-banging on i915 gmbus dpb. force bit now 1
    [   13.655959] [drm:intel_sdvo_read_byte] i2c transfer returned -6
    [   13.655961] [drm:intel_sdvo_init] No SDVO device found on SDVOB
    [   13.655990] [drm:intel_gmbus_force_bit] disabling bit-banging on i915 gmbus dpb. force bit now 0
    [   13.655992] [drm:intel_setup_outputs] probing HDMI on SDVOB
    [   13.656064] [drm:intel_dp_init_connector] Adding DP connector on port B
    [   13.656108] [drm:intel_dp_aux_init] registering DPDDC-B bus for card0-DP-1
    [   13.656144] [drm:intel_setup_outputs] probing SDVOC
    [   13.656148] [drm:intel_gmbus_force_bit] enabling bit-banging on i915 gmbus dpb. force bit now 1
    [   13.665729] [drm:intel_sdvo_read_byte] i2c transfer returned -6
    [   13.665731] [drm:intel_sdvo_init] No SDVO device found on SDVOC
    [   13.665758] [drm:intel_gmbus_force_bit] disabling bit-banging on i915 gmbus dpb. force bit now 0
    [   13.665760] [drm:intel_setup_outputs] probing HDMI on SDVOC
    [   13.665808] [drm:intel_dp_init_connector] Adding DP connector on port C
    [   13.665845] [drm:intel_dp_aux_init] registering DPDDC-C bus for card0-DP-2
    [   13.665888] [drm:intel_dp_init_connector] Adding DP connector on port D
    [   13.665927] [drm:intel_dp_aux_init] registering DPDDC-D bus for card0-DP-3
    [   13.665964] [drm:intel_tv_init] Integrated TV is not present.
    [   13.665970] [drm:intel_modeset_readout_hw_state] [CRTC:20] hw state readout: disabled
    [   13.665977] [drm:intel_modeset_readout_hw_state] [CRTC:23] hw state readout: enabled
    [   13.665980] [drm:intel_modeset_readout_hw_state] [ENCODER:25:LVDS-25] hw state readout: enabled, pipe B
    [   13.665982] [drm:intel_modeset_readout_hw_state] [ENCODER:32:DAC-32] hw state readout: disabled, pipe A
    [   13.665984] [drm:intel_modeset_readout_hw_state] [ENCODER:33:TMDS-33] hw state readout: disabled, pipe A
    [   13.665986] [drm:intel_modeset_readout_hw_state] [ENCODER:38:TMDS-38] hw state readout: disabled, pipe A
    [   13.665988] [drm:intel_modeset_readout_hw_state] [ENCODER:40:TMDS-40] hw state readout: disabled, pipe A
    [   13.665990] [drm:intel_modeset_readout_hw_state] [ENCODER:42:TMDS-42] hw state readout: disabled, pipe A
    [   13.665992] [drm:intel_modeset_readout_hw_state] [ENCODER:44:TMDS-44] hw state readout: disabled, pipe A
    [   13.665994] [drm:intel_modeset_readout_hw_state] [CONNECTOR:24:LVDS-1] hw state readout: enabled
    [   13.665995] [drm:intel_modeset_readout_hw_state] [CONNECTOR:31:VGA-1] hw state readout: disabled
    [   13.665997] [drm:intel_modeset_readout_hw_state] [CONNECTOR:34:HDMI-A-1] hw state readout: disabled
    [   13.665999] [drm:intel_modeset_readout_hw_state] [CONNECTOR:39:DP-1] hw state readout: disabled
    [   13.666001] [drm:intel_modeset_readout_hw_state] [CONNECTOR:41:HDMI-A-2] hw state readout: disabled
    [   13.666002] [drm:intel_modeset_readout_hw_state] [CONNECTOR:43:DP-2] hw state readout: disabled
    [   13.666004] [drm:intel_modeset_readout_hw_state] [CONNECTOR:45:DP-3] hw state readout: disabled
    […]

As you can see, the vendor firmware detects a dual-link LVDS
configuration.

[drm:intel_lvds_init] detected dual-link lvds configuration

With coreboot a single-link LVDS configuration is detected.

[drm:intel_lvds_init] detected single-link lvds configuration

Francis Rowe wrote to the list, that this is configured during compile
time in coreboot [1][2]. You’d need to rebuild your image. Then you
should have at least graphics when Linux started.

There are two more things two do.

1. Contact the Linux i915 folks and ask them, if the module is able to
detect the configuration itself and does not rely on the firmware or if
there is a way to override that setting from the Linux command line.

2. Improve coreboot so that the LVDS link configuration is detected
during runtime as the vendor firmware (VGA Option ROM/Video BIOS(?))
does.

Thanks,

Paul

[1] http://coreboot.org/pipermail/coreboot/2015-July/080109.html
[2] http://libreboot.org/docs/hcl/gm45_lcd.html

Dear Paul,
> As you can see, the vendor firmware detects a dual-link LVDS
> configuration.
>
> [drm:intel_lvds_init] detected dual-link lvds configuration
>
> With coreboot a single-link LVDS configuration is detected.
>
> [drm:intel_lvds_init] detected single-link lvds configuration

Thank you for checking the logs!
> Francis Rowe wrote to the list, that this is configured during compile
> time in coreboot [1][2]. You'd need to rebuild your image. Then you
> should have at least graphics when Linux started.

Yes, I am very happy that they found a workaround. This is the next
thing I am going to do.
Because I re-flashed the Lenovo Bios I have to disassemble the T400
again to reach the flash chip.
I wanted to provide you with the logs first.
> There are two more things two do.
> Contact the Linux i915 folks and ask them, if the module is able to
>
> detect the configuration itself and does not rely on the firmware or if
> there is a way to override that setting from the Linux command line.

Who is going to do this? Is this [1 [1]] the right community? I am not
sure, if I am able to explain it right.
If you want me to do this, what do you mean exactly with "the module",
is it "i915"?

On [2 [2]], is written:
"Linux (kernel) has its own init code (including EDID parsing). The
current suspicion is that coreboot's
own initialization code has buggy EDID parsing in some way, such that
the proper settings for the panel
are not used, leading to the failure."

Please tell me, if I understand it right:
The kernel module should not rely on coreboot concerning the settings
for the display and you want to know, 
if the kernel module is able to get the needed information direct from
the display (EDID)? If this is possible, you want to know,
if this is cofigurable with the Linux command line? Means "Linux command
line" the same as "kernel command line"?
> Improve coreboot so that the LVDS link configuration is detected
>
> during runtime as the vendor firmware (VGA Option ROM/Video BIOS(?))
> does.

While I am learning a lot of things working on this issue, this is far
above my skills.

Best regards,

Kroms

[1] https://01.org/linuxgraphics/community/kernel
 [1][2] http://libreboot.org/docs/hcl/gm45_lcd.html [2]
 [1]

Links:
------
[1] https://01.org/linuxgraphics/community/kernel
[2] http://libreboot.org/docs/hcl/gm45_lcd.html

[Paul Menzel]

Dear Kroms,

Am Montag, den 13.07.2015, 09:43 +0200 schrieb [e-mail]:

[…]
> > There are two more things two do.
> > Contact the Linux i915 folks and ask them, if the module is able
> >
> > to detect the configuration itself and does not rely on the 
> > firmware or if there is a way to override that setting from the 
> > Linux command line.
>
> Who is going to do this? Is this [1] the right community? I am not
> sure, if I am able to explain it right.
> If you want me to do this, what do you mean exactly with "the 
> module", is it "i915"?

Yes i915 and it’d be great if you could do it.

At least the second question should be answered. Looking at
`/sbin/modinfo i915` it shows that there is such an option already.

parm:           lvds_channel_mode:Specify LVDS channel mode (0=probe BIOS [default], 1=single-channel, 2=dual-channel) (int)

So you could add `i915.lvds_channel_mode=2`, I think, to the Linux
command line and see if Linux is able to initialize your system.
> On [2], is written:
> "Linux (kernel) has its own init code (including EDID parsing). The
> current suspicion is that coreboot's own initialization code has 
> buggy EDID parsing in some way, such that the proper settings for the 
> panel are not used, leading to the failure."
>
> Please tell me, if I understand it right:
> The kernel module should not rely on coreboot concerning the settings
> for the display and you want to know, 
> if the kernel module is able to get the needed information direct 
> from the display (EDID)? If this is possible, you want to know,
> if this is configurable with the Linux command line? Means "Linux 
> command line" the same as "kernel command line"?

Yes. But I am not sure if it’s related to EDID parsing at all. coreboot
just hard codes it.
> > Improve coreboot so that the LVDS link configuration is detected
> >
> > during runtime as the vendor firmware (VGA Option ROM/Video 
> > BIOS(?))
> > does.
>
> While I am learning a lot of things working on this issue, this is 
> far above my skills.

No problem. Maybe somebody else will do it.

Thanks,

Paul
> [1] https://01.org/linuxgraphics/community/kernel
> [2] http://libreboot.org/docs/hcl/gm45_lcd.html