Hi,

Dima Zavin wrote:
>  
>
>     So... I have tried the installer on a small desktop PC - a Asus
>     Nova Lite:
>
>     http://www.asus.com/news_show.aspx?id=11565
>
>
> Heh, cute. I should mention that we really should not be reusing the 
> same installer images as a "general purpose" installer for anything 
> x86. It really is meant to be configured individually for every target 
> device for reproducible builds/installs, per-device tweaks, etc.
>
>     and the installer seems as it has run its course well (although it
>     took
>     a lot of time - namely resizing the 250 GB partition), and the
>     nova lite
>     pc now boots from the hard drive.
>
>
> resize2fs is really really slow. I didn't get a chance to dig into it 
> to figure out why. It could probably be improved significantly with a 
> little TLC.
>
>
>     However, Android is not able to boot completely. After a few seconds,
>     the screen goes completely black and never proceeds. By pressing
>     alt+f1
>     I am able to access the system console (however the system keeps
>     trying
>     to initialize the graphic interface, and I have to press alt+f1 every
>     few seconds to access the console).
>
>
> Are the framebuffer drives loading correctly? The prebuilt drivers 
> were only verified to work on the i945g chipset that's in the eeepc 
> 701 (though it "probably" works on many other i915 derivatives?). As 
> mentioned I have only verified this on my eeepc, so ymmv.
>
> It's a bit of a pain right now since the drm/i915 modules need to be 
> built outside the kernel tree, and I really didn't want to mirror the 
> f.o drm trees since its under development, and I dont have the time to 
> keep up with it at the moment. I just threw together the precompiled 
> bins for convenience. The f.o guys are doing great work, and once 
> their stuff is pushed into mainline (.29ish timeframe i think?), the 
> fb drivers will be part of the kernel build and there will be much 
> rejoicing. It will be great.
>

Although the nova lite has a i945 intel board chipset, it uses an 620gle 
ati graphics card... that's probably a, or the, problem. Does the kernel 
of the installer include any ati drivers?

>     dmesg:
>
>     the surfaceflinger activity has exited due to a segfault - and I
>     can see
>     from the log that it keeps retrying forever.
>
>     Logcat :
>
>     /dev/pmem could not be found, as well as lighgl.so.
>     validate_display_surface has failed with error 300
>     (BAD_DISPLAY_SURFACE)
>     call to OpenGL API has been done with no current context.
>
>
> The pmem and hgl messages are non fatal in this case. Surfaceflinger 
> should fall-back onto ashmem heaps (iirc) and sw gl. It probably goes 
> haywire since your fb driver is not operational?
>  
>
>
>     Do you have any clue of what might be causing the surfaceflinger
>     to fail
>     (is it the lack of drivers?)? Also what do you recommend to
>     troubleshoot
>     these situations (how to get the complete logs out of the test
>     box, and
>     what other troubleshooting tools are available in the android
>     console).
>
>
> If this is not one of those silly umpc/netbook thingies and it has a 
> normal serial port, i would recommend hooking up to it. It will give 
> you a working shell, and let you debug a lot easier than on a 
> flickering console :)

:-)

>
> You should also build your own kernel and make sure you have the right 
> ethernet drivers to get access to it from the network. You can then 
> run adb over the network and essentially get a remote console.
>
> Hope this helps.

It does! Thanks! I tried this since I don't have a eee/netbook PC and 
was too anxious to see Android splashing on the big screen :-)... sorry 
for the trouble...

Cheers,
Filipe

>
> --Dima
>  
>
>
>     Cheers,
>     Filipe
>
>
>
>
>
>
>     >
>


--~--~---------~--~----~------------~-------~--~----~
unsubscribe: [email protected]
website: http://groups.google.com/group/android-porting
-~----------~----~----~----~------~----~------~--~---

Reply via email to