Re: [coreboot] about Ivy Bridge CPU with QM67 PCH

[Iru Cai]

Hi,

On Sun, Oct 18, 2015 at 06:44:09PM +0200, Peter Stuge wrote:
> Hi,
> 
> Iru Cai wrote:
> > I've been testing the Lenovo T420 port recently, and now I can
> > install an Ivy Bridge processor and run fine on Linux(except some
> > thermal issues).
> 
> That's pretty nice I think.
> 
> 
> > However, the native graphics initialization doesn't work properly
> ..
> > I don't know if someone has tried using mixed generation of CPU and
> > chipset combination with coreboot before,
> 
> Probably not.
I finally found it's lvds_num_lanes in devicetree.cb that breaks
native graphics initilization, because SNB and IVB use different values.
And it has been solved in this upstream commit:

commit c48f5ef3cc623a4b1bccdbc9cb3e1d15505b7ad4
Author: Vladimir Serbinenko 
Date:   Sun Oct 11 02:05:55 2015 +0200

Kill lvds_num_lanes

Only one value would work with corresponding gma code currently (which one
depends on board). Going forward, it's possible to compute which number can
be used, so there is no need to keep this info around.

Thank you Vladimir Serbinenko.

> 
> 
> > I have pushed the patch that makes coreboot support both SNB and
> > IVB processor for review:
> > http://review.coreboot.org/#/c/12087/
> 
> Thanks! There is a source code style issue - you used two spaces
> instead of one tab to indent - please fix that, but other than that
> the patch looks fine to me.
> 
> But the two files gma_{sandy,ivy}bridge_lvds.c are a bit of a mess -
> they need to be cleaned up and can quite easily be unified into a
> single file, without having it become the mess that the i915 kernel
> code is. The key is to have a good data model, and the data model
> seems clear to me from the existing code.
> 
> For fun (or pain), take a look at: git diff 758a41 7137c5
> (sandy..ivy)
> 
> 
> //Peter
> 
> -- 
> coreboot mailing list: coreboot@coreboot.org
> http://www.coreboot.org/mailman/listinfo/coreboot

Iru Cai


signature.asc
Description: PGP signature
-- 
coreboot mailing list: coreboot@coreboot.org
http://www.coreboot.org/mailman/listinfo/coreboot

[coreboot] about Ivy Bridge CPU with QM67 PCH

[Iru Cai]

Hi,

I've been testing the Lenovo T420 port recently, and now I can install an Ivy 
Bridge processor and run fine on Linux(except some thermal issues). However, 
the native graphics initialization doesn't work properly before Linux kernel 
boots. I can see the GRUB interface but it displays bad as if it's on a panel 
larger than the laptop panel. Then, Linux kernel boots and the graphics become 
good again. I read some of the i915 driver code of Linux, and saw some code 
about PCH, so I think this issue has something to do about the southbridge. I 
don't know if someone has tried using mixed generation of CPU and chipset 
combination with coreboot before, and I hope this issue can be resolved.

I have pushed the patch that makes coreboot support both SNB and IVB processor 
for review:
http://review.coreboot.org/#/c/12087/

Iru Cai


signature.asc
Description: PGP signature
-- 
coreboot mailing list: coreboot@coreboot.org
http://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] about Ivy Bridge CPU with QM67 PCH

[Peter Stuge]

Hi,

Iru Cai wrote:
> I've been testing the Lenovo T420 port recently, and now I can
> install an Ivy Bridge processor and run fine on Linux(except some
> thermal issues).

That's pretty nice I think.


> However, the native graphics initialization doesn't work properly
..
> I don't know if someone has tried using mixed generation of CPU and
> chipset combination with coreboot before,

Probably not.


> I have pushed the patch that makes coreboot support both SNB and
> IVB processor for review:
> http://review.coreboot.org/#/c/12087/

Thanks! There is a source code style issue - you used two spaces
instead of one tab to indent - please fix that, but other than that
the patch looks fine to me.

But the two files gma_{sandy,ivy}bridge_lvds.c are a bit of a mess -
they need to be cleaned up and can quite easily be unified into a
single file, without having it become the mess that the i915 kernel
code is. The key is to have a good data model, and the data model
seems clear to me from the existing code.

For fun (or pain), take a look at: git diff 758a41 7137c5
(sandy..ivy)


//Peter

-- 
coreboot mailing list: coreboot@coreboot.org
http://www.coreboot.org/mailman/listinfo/coreboot