PPCGeeks - NAND Testing - 05-25 Update: New LK, Recovery.img, Kernel Updates through Recovery

Page 169 of 274

« First

169

Show 40 post(s) from this thread on one page

PPCGeeks (http://forum.ppcgeeks.com/index.php)

- Android On TP2 Development (http://forum.ppcgeeks.com/forumdisplay.php?f=319)

- - NAND Testing - 05-25 Update: New LK, Recovery.img, Kernel Updates through Recovery (http://forum.ppcgeeks.com/showthread.php?t=134598)

lionedas

03-11-2011 08:20 PM

Re: NAND Boot Testing - 03-08: Full NAND FRX05 Testing (Data not working for some)

OK

This is kinda weird, but wifi just .............. co ... con....... connected,
Thanks guys, I would like to thank you all by name, but I would spill your names wrong, and then look stupid, so thanks to all who help with this project.

Now for the stupid question, would you guys not have the same connection problem as me if you were doing a fresh install, or are you guys just doing updates...
cuz if you are just doing updates wouldnt it be harder to know if something else broke.... remember this is like talking to me about event horizons, and sub atomic particles, or worst women!

natemcnutty

03-11-2011 08:24 PM

Re: NAND Boot Testing - 03-08: Full NAND FRX05 Testing (Data not working for some)

Quote:

Originally Posted by Lmiller1708 (Post 2067298)

Nate. Look at the install-seq.sh now.

The reason why Wifi didn't work was because it was not getting past the #echo "Installing...Recovery" section. We will just need to leave that out until we can get the recovery part working. :)

You must have had an old install-seq.sh file if it was breaking there. I did a commit a couple days ago to fix that issue. I put it up on git so we would all use the same one for easier troubleshooting.

That being said, I think I may have the modules thing ironed out for multiple kernel versions. Here's the code that I'm testing right now in install-seq.sh:

Code:

if [ ! -d /data/modules/`uname -r` ] ; then

    mkdir /data/modules/`uname -r`

    cp -a /data/modules /data/modules/`uname -r`

    cp -R /system/etc/wifi/* /data/modules/`uname -r`

    rm /system/lib/modules/`uname -r`

    ln -s /data/modules/`uname -r` /system/lib/modules/`uname -r`

elif [ ! -e /system/lib/modules/`uname -r` ] ; then

    cp -R /system/etc/wifi/* /data/modules/`uname -r`

    ln -s /data/modules/`uname -r` /system/lib/modules/`uname -r`

fi

Also, if we make this change, we need to fix up init, so I did this:

Code:

if [ ! -d /data/modules/`uname -r` ] ; then

    echo "Modules are not installed. Download the matching"

    fail "androidupdate.tgz and put it in your andboot folder"

elif [ ! -e /system/lib/modules/`uname -r` ] ; then

    cp -R /system/etc/wifi/* /data/modules/`uname -r`

    ln -s /data/modules/`uname -r` /system/lib/modules/`uname -r`

fi

If this works the way I'm expecting, I'll attach the install-seq.sh, NBH, and androidupdate.tgz to a post for people to test with. Assuming it works properly for them too, then I'll commit it. I guess on ln, the last folder listed in the command is finnicky. If that folder exists, it will create an additional folder under it named the last folder in the target directory name. If it doesn't exist, it will simply create it there. Finally got that all figured out since I was a little confused :)

natemcnutty

03-11-2011 08:27 PM

Re: NAND Boot Testing - 03-08: Full NAND FRX05 Testing (Data not working for some)

Quote:

Originally Posted by lionedas (Post 2067303)

At this point, we each have our own clone locally going where we are making changes. We should be testing each change we make, and then commit them if they are good. Unfortunately, sometimes we miss work that we have done to fix something, it isn't committed with the other commits, and then it works for us but not everyone else ;)

As for the androidupdate.tgz, I'm hoping we move away from having anything but modules in those. It's useful for now to make system changes and all that, but eventually those should be cooked into the system rather than applied from the update.

natemcnutty

03-11-2011 08:41 PM

Re: NAND Boot Testing - 03-08: Full NAND FRX05 Testing (Data not working for some)

Well, data is looking awesome, but I back to the drawing board on that code for modules :(

Edit: Holy crap is this thing smooth... :D

siraloz

03-11-2011 11:54 PM

Re: NAND Boot Testing - 03-08: Full NAND FRX05 Testing (Data not working for some)

Nice. Smooth. Sexy.

Job well done!

Observations:
FR05, camera no worky...
Gmail Sync'ed... Calendar Sync'ed... GPS location in the ocean...
Youtube -- "There was a problem while playing" Message
USB Drive works!

[ACL]

03-12-2011 01:43 AM

Re: NAND Boot Testing - 03-08: Full NAND FRX05 Testing (Data not working for some)

1 Attachment(s)

Quote:

Originally Posted by siraloz (Post 2067380)

you look like someone who wants to keep a running list of whats up.... we need it. So if you can keep track if it, awesome..

nate: Thanks alot for tackling that module thing. i got my pot of coffee ready so im hopefully finishing up this recovery soon. THe downside right now is that everyone will need to reflash eventually:-p .. yup.. until we get around the multi kernel thing we need to wipe everything, but lmiller had success with the backup and restore so its not so bad ..

This is the latest screenshot off the recovery .. It's no magldr, but it will suffice for us. I'm keeping the backend script based so we all can maintain it. Only part that will require recompiles is the UI itself.

natemcnutty

03-12-2011 02:59 AM

Re: NAND Boot Testing - 03-08: Full NAND FRX05 Testing (Data not working for some)

Quote:

Originally Posted by siraloz (Post 2067380)

Thanks, I added those to the first post. If you'd like to keep a running tab of working/not working, I can link that in the first post as well. If not, I can just keep updating the first post with everyone's experiences ;)

natemcnutty

03-12-2011 03:05 AM

Re: NAND Boot Testing - 03-11: FRX05 on NAND (data working again!)

So ACL/LMiller, I was wondering if we should consider creating a new git for the android install and android update files. For android install, we would only need to maintain the files that deviate from FRX05. This way we can keep track of sysinit.rc and maybe try using /system/init instead of the init.android in initrd :)

What do you guys think?

[ACL]

03-12-2011 03:45 AM

Re: NAND Boot Testing - 03-11: FRX05 on NAND (data working again!)

Quote:

Originally Posted by natemcnutty (Post 2067452)

I think we stopped using the init.android in initrd. If we havent then we need to start including init inside /system. The init.android inside the initrd will be replaced by the recovery version (which is very similar if not almost identical) this way recovery is always possible even if android poops. The initrd automatically checks for init inside system (as it should) so no changes are needed there when moving the correct init inside /system.

Eventually im hoping we can include our work in the xdandroid git. So i can ask stine to create a small project for us in there so we can be officially part of xdandroid. If not, well then we can always just make our own. I'll email him and let yall know. This way to create frx5, all we would need to do is pull the files over the base img and we should be golden.

how does that sound?

natemcnutty

03-12-2011 04:59 AM

Re: NAND Boot Testing - 03-11: FRX05 on NAND (data working again!)

Quote:

Originally Posted by [ACL] (Post 2067457)

Sounds great to me. You're the man ;)

All times are GMT -4. The time now is 06:39 PM.

Page 169 of 274