Re: NITdroid 0.4.x
 

Hmm. I tested it out and I've got some comments.

*The keyboard is severely broken. Capitalization seems not to work and it crashes constantly.

*Wireless has bugs as well. If you try to connect to a secured network that it finds, it only accepts the key in hexadecimal, not ASCII.

*The "Boot to Maemo" option doesn't work for me, even with serial-console enabled-- I was under the impression that it was operational; Was this assumption incorrect?

*Thanks for keeping the symlinks-- my script still works.

*The app installer is a nice touch, but there are very few apps available.

*Trying to scroll is harder, it seems. When you press lightly, it clicks an option. When you press harder, it still clicks. Go figure.

Overall though, it's still an excellent port.

Thanks Solca!

Re: NITdroid 0.4.x
 

Wait, this really exists? Does it really work in 0.4.2 and how would I set this flag?

Re: NITdroid 0.4.x
 

You can set that flag with flasher (linux or mac one). There's a bug in kexec so that flag needs to be enabled. Good luck again, :)

Re: NITdroid 0.4.x
 

Thanks. I guess flasher --help will give me the flag I need.
I wonder if there is a Mac Intel version of flasher anywhere... (Nokia only offers Mac PPC)

UPDATE: I flashed the nitdroid kernel with --set-rd-flags=serial-console (which flasher reported was set, numerically anyway), but when I use the Maemo command in the menu of NITdroid, the tablet just reboots back into NITdroid. :(

Re: NITdroid 0.4.x
 

There is a contributed Mac Intel ADB binary in the project's site which you can use to connect to your NIT.

Then, when you are in, please type 'boot_maemo' in the shell and if it works there is a bug in the GUI part, as meizirkki said, you need to set the R&D flag 'serial-console'.

Then report your results.

HTH.

Re: NITdroid 0.4.x
 

Just watched Ricky's new video. I think he did a great job. Hit on all the key points. My only concern was the way he described flashing to Maemo to charge the battery.. kinda came across like it takes an hour to switch back and forth instead of the few seconds it really takes.

Still no word on the approval from nongnu.org for formalized hosting. Its amazing they get anything done over there.

I've had a few crazy days and haven't been in the IRC channel. Has any progress been made on initializing the DSP for sound? Do you guys log the chats with the channel bot? Do you object to having the chats archived for posting?

Re: NITdroid 0.4.x
 

Nope, dspgateway have been removed from latest kernels and priors doesn't work, still investigating. There is a possibility of ARM sound (without DSP) but still doesn't work.

By now everything should be working but is not that simple for non "open" devices.

You are welcome! really, anything you can do to help is warmly welcomed by all of us. Thanks again for your dedication.

Re: NITdroid 0.4.x
 

I've been trying without success to get this working on the WiMAX Edition. I am almost sure that the rootfs is not being chrooted, and I am stuck trying to debug or verify this. @solca: Can you please explain how your kernel image works? Does it (or rather, does nolo) boot the kernel into the standard initfs, and then chroot to the real partition? I've read somewhere that the standard android install for the G1 works in a different way because the device does not have dedicated space for both partitions, is this the case for nitdroid?

The reason why I ask is that the initfs from the fiasco image for WE is different from the stock N810, and has a bunch of extra routines for WiMAX.

Re: NITdroid 0.4.x
 

You can try debug-kernel to see what goes wrong.

Re: NITdroid 0.4.x
 

Yeah, try with a debugging kernel first. NITdroid doesn't use the /linuxrc script in the initfs partition.

Now, this is how NITdroid boots:

1. NOLO does its init thing and finally loads and uncompress the flashed kernel to RAM.
2. Control is passed by NOLO to the kernel loaded in RAM.
3. If that flashed kernel is NITdroid it contains an embedded RAM filesystem (initramfs) which is / (root dir) all the time mimicking other Android devices.
4. Kernel pass control to /init in the ramfs.
5. /init loads instructions from /init.rc and /init.nokia.rc.
6. /init.rc tells init to mounts /dev/block/mmcblk0p3 (ext3) under /nitdroid.
7. /init.nokia.rc tells init to run nitboot which loads and run the Nokia proprietary stuff that lives in NITdroid under /nokia.
8. /init launches all remaining Android stuff.

Visually:

1. Initial white Nokia screen is loaded by NOLO.
2. First NITdroid screen (static image) is loaded by /init when run.
3. Screen dims when DSME proprietary daemon is loaded by nitboot.
4. Second NITdroid screen (animated) is loaded by the "remaining Android stuff" when loading.
5. Finally the Android GUI is loaded.

Differences for the debugging kernel:

1. Kernel and console output are directed to screen and serial.
2. ADB daemon is always loaded early in the init sequence.
3. First NITdroid screen (static) is not loaded to maximize console output.
4. Second NITdroid screen (animated) is not loaded but screen turns black until the GUI loads or an error is shown.

You should tell your visual sequence and better if you can connect with ADB and get a 'logcat' so we can determine where is the problem...