I’m really pissed.
It’s again the same shit as with my HTC dream (or should I say nightmare?)
Ages outdated kernel, branches with drivers which are that badly written that they can never be merged into Torvalds master again.
Dudes! Can’t you have a look over to Texas Instruments?!
The OMAP is well documented, there are free drivers available and there is always support in main.
WTF? What use is a kernel where no peripherial drivers are available within the official, recent kernel?!
Feel free to contact Notion Ink Support on support@notionink.com or Submit your query at http://support.notionink.com/index.php?/Tickets/Submit/
I’ll most likely do that as soon as I’ve got board-harmony running on top of linux-next.
Then I’ll use the ticketing system in order to get the pixel-qi-screen driver running properly.
A while ago I tried to boot a mainline kernel on one of the Tegra2 boards my co-workers are working on, and after Fastboot it just froze up. I guess if I ever need more power I will have a look at the TI Sitara series, which to me looks even better than the OMAP since it comes with SATA and gigabit Ethernet. I assume the Linux support is as good as it is for the OMAP.
It appears to be well supported.
But only a test can show us how well it goes with portability and reusability in practice.
http://processors.wiki.ti.com/index.php/Sitara_Linux_Software_Developer%E2%80%99s_Guide
Until there is proper Tegra2 support in mainline, there is not much point in getting device-specific stuff into mainline.
It’s missing something as important as a graphics driver – even fb.
That being said, most drivers for the adam 2.6.36+ kernels was initially written as a proof of concept. No one has bothered to rewrite them properly now.
Perhaps a task for you? 😉
In all seriousness: Writing kernels for tegra devices isn’t particularly easy and I’m sure you’ll realize this soon enough. The kernels are very peculiar and it doesn’t take a to to make them stop booting entirely.
I already found out that writing code for Tegra 2 can become kinda para informatics.
For example pinmuxing which has no effect or no output in any form on a specific GPIO pin at all…
Anyway, hopefully porting the adam board files to the recent nvidia kernel (3.1.10) will work.
It’s less of a work because it only needs some small modifications I guess…
But yeah, you’re right, it’s maybe a job for me to do the porting into upstream 🙂