[beagleboard] Beagle Board kernel - omapfb: no displays

I have tried to compile my own 2.6.34 kernel from

http://git.kernel.org/?p=linux/kernel/git/tmlind/linux-omap-2.6.git;a=summary

but I constantly get a message during boot

omapfb omapfb: no displays

And finally FB initialization fails. I have compared my .config with
several others and tried many different configurations, but with no
success. No DVI, no LCD. In /sys/bus/omapdss/drivers I can see
generic_panel and a LCD driver I have configured.

Is it a .config problem? Does it need special bootargs? Do I have to
register something in my board-file?

Any hints where to look at are appreciated.

Nikolaus

Why don't you build a know working kernel like http://gitorious.org/beagleboard-validation/linux first before trying to build a known broken one? I say known broken, since upstream doesn't have DSS2 support for beagle yet.

Hi Koen,
I obviously assumed that if DSS2 and beagleboard configs are available
in http://git.kernel.org/?p=linux/kernel/git/tmlind/linux-omap-2.6.git;a
that they work together. How should I know that it is *known* to be
broken? Maybe, the BB project should put more focus on fixing upstream
first before developing new things? This fragmentation is driving me
crazy…

So I will try to get the beagleboard.c file from your link.

BR,
Nikolaus

And of course you did your homework and checked if 'the BB project' hadn't already fixed that and checked when the kernel merge windows are? Lo and behold: http://gitorious.org/linux-omap-dss2/linux/commits/for-next , beagle DSS2 support waiting for the .35 merge window to open to go into mainline linux.

So now that you are pointing fingers at us about fixing things upstream, how is your track record? I gather the Acer N10 is fully supported by 2.6.34?

Dear Prof Koen,
I just wonder why both

a) http://git.kernel.org/?p=linux/kernel/git/tmlind/linux-omap-2.6.git;a=summary
b) http://gitorious.org/beagleboard-validation/linux

(and probably more repositories) exist and have BB support. I thought
the tmlind branch is the one where OMAP development goes on. This is
simply confusing me (I don't know if it is confusing others as well).

I don't expect you to take any actions on my observations and
opinions.

BR,
Nikolaus

Ah,
now i understand (at least I think).

You just committed http://gitorious.org/linux-omap-dss2/linux/commit/1f7dac135e9a98248fa4f89a79397898ae5fa14a
on Wednesday which says "board-omap3-beagle: add DSS2 support".

Please, please don't expect that I am able to read and follow your
commit logs 24 hours a day (expecially when I don't understand the
relationship of the gitorius repos to the tmlind branch). I spent my
time since Wednesday with configuring, compiling and testing...

Sorry for creating confusion,
Nikolaus

Maybe, the BB project should put more focus on fixing upstream
first before developing new things?

And of course you did your homework and checked if 'the BB project' hadn't already fixed that and checked when the kernel merge windows are? Lo and behold:http://gitorious.org/linux-omap-dss2/linux/commits/for-next, beagle DSS2 support waiting for the .35 merge window to open to go into mainline linux.

Ah,
now i understand (at least I think).

You just committed http://gitorious.org/linux-omap-dss2/linux/commit/1f7dac135e9a98248fa4f89a79397898ae5fa14a
on Wednesday which says "board-omap3-beagle: add DSS2 support".

No.

Read that page:

• Tomi Valkeinen (Committer) 2 days ago
• Koen Kooi (Author) 23 days ago

So Tomi (the DSS2 maintainer) commited that 2 days ago to his for-next branch that will go into 2.6.35 during the merge window. The patch was done more than 3 weeks ago. It was based on this one: http://gitorious.org/beagleboard-validation/linux/commit/8a3d9bff79f71a2733607e409f820e79142592d5.patch which was based on an older patch by Steve Sakoman, which was based on an even older patch by Tomi. The DSS2 patch for beagle probably goes back more than a year.

Please, please don't expect that I am able to read and follow your
commit logs 24 hours a day (expecially when I don't understand the
relationship of the gitorius repos to the tmlind branch).

First, they are not my commit logs, but I do expect people to pay attention when Jason says that the code we're going to use from now on is located at http://gitorious.org/beagleboard-validation . I especially expect people to do so when that start pointing fingers.

> Please, please don't expect that I am able to read and follow your
> commit logs 24 hours a day (expecially when I don't understand the
> relationship of the gitorius repos to the tmlind branch).

First, they are not my commit logs, but I do expect people to pay attention when Jason says that the code we're going to use from now on is located at http://gitorious.org/beagleboard-validation. I especially expect

I mean "you" as those actively doing a lot of work on the kernel (not
you personally).

Hm. I must have missed the statement from Jason. And could not locate
it in this list. By the way, I do not really expect a project called
"validation" to be the basis for a production system...

people to do so when that start pointing fingers.

People usually start when the are lost in information overload and
don't find the single tree in the forest... Take my criticism as
engagement from someone who really wants to get things working.

Anyway, I have been sucessful in compiling the kernel from
http://gitorious.org/beagleboard-validation/linux (beagleboadXM
branch) except an issue with net/netfilter/xt_rateest.ko. And, the
kernel initializes DVI well. Applying the DSS2 for-next to my 2.6.34
did not make it work.

But the validation sources appear to build a 2.6.32 kernel. So I am
still in doubt if I should take this as the basis of my patches and
individual .config. But it appears to be the easier way than starting
at 2.6.34 without working DSS2 and fixing that...

Thanks for the pointer to http://gitorious.org/beagleboard-validation

Nikolaus

I don't mean for it to 'production' for anybody's product, just used to validate the board and demonstrate the capabilities. The real place for collaboration is the linux-omap list, but we are trying to throw something quick together at the gitorious site. We'll try to align more and more with linux-omap moving forward.