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

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)

Re: NAND Boot Testing - 01-07: Panel power on/off fixes

Quote:

Originally Posted by natemcnutty (Post 2060666)

yes, I am using the original method with the nbtools and all that. I will be playing with the cat method tomorrow just to see if it matters.

Works like a charm. I didnt commit those changes to tinboot but i made them on the autobuild so the rest of the peeps can test from a reference build.

Index of /files/rhodium_nand/

Thanks a lot fellas.. we are almost there.

Re: NAND Boot Testing - 01-07: Panel power on/off fixes

Quote:

Originally Posted by [ACL] (Post 2060675)

Thanks ACL. I've got a ton more testing to do, but I did find that the cat method leaves some issues in the partition table. If we had a kernel/initrd.gz that was too large, I think it would have issues. This is probably why the Vogue uses lzma now.

When we use the nbtools, it adjusts all of the partition information based on the size of the crap we are shoving into it. I think we might even be able to build any size initrd we want using this method since it auto-adjusts the table. As long as our mtdparts partition 1 does not overlap with the fatfs partition (which would require changing tinboot code), then we are good to go.

Re: NAND Boot Testing - 01-07: Panel power on/off fixes

Quote:

Originally Posted by [ACL] (Post 2060653)

also dont use the menu to install. The script is whacko and does other things its not supposed to do. So install via the script only for now.

Could this be why I'm having some trouble with this nbh? after A N D R O I D, I freeze up with "Unable to open persistent property directory /data/property errno: 2" using the files from the NAND server.

User error, or adjustment necessary?

Quote:

Originally Posted by [ACL] (Post 2060675)

Re: NAND Boot Testing - 01-07: Panel power on/off fixes

Quote:

Originally Posted by Boominsvx (Post 2060679)

Probably.. just dont use it.. use the install script.. in the andboot folder, if you have the install script it will automatically execute and do all the goodness. Once it runs it will disable itself so it does not run again.

Put this in your andboot folder with the android and updates .

http://htcdevs.wirelesstcp.net/files...install-seq.sh

Eventually this script will make it into the menu but its too xdandroid ish and wont support other builds.. so i need to come up with a happy medium where we should be able to install future builds instead of having to hack our initrd all the time.

Re: NAND Boot Testing - 01-07: Panel power on/off fixes

Quote:

Originally Posted by [ACL] (Post 2060680)

I think we provide the script as a template, and it should get modified by the chef if needed.

As for Boominsvx, the issue he is having could also be because he did the system install without doing the update part from replimenu.

Re: NAND Boot Testing - 01-07: Panel power on/off fixes

Quote:

Originally Posted by natemcnutty (Post 2060682)

yeah.. but then should we remove he ability to install from replimenu? Do we even need it ? I guess the wiping cache and permissions check are a good thing to have. But i wouldnt install through it. it doesnt get easier than the install script.. just sit back and let it do all the goodies.

ahh just wait till we do cyanogen.. its a different animal ..

Re: NAND Boot Testing - 01-07: Panel power on/off fixes

What are you guys still doing up!?!?! Dedication knowns no boundries I guess. Sorry, it was user error, but I've got the androidinstall.tgz in the RIGHT folder now, and also put the script in place. Since I'm on a RHOD400 anyway, it looks sufficient for my needs. Had to flash winmo because my card reader is dead, but she's booting now, and you're right, the install script is much easier.

Now, back to reading about 3G data on the Kaiser where i'm a quiet troll who doesn't distract anyone's focus. Thanks guys.

Edit: Here's the bootlog from the script install since it changed. Just want to make sure my info is accurate for you guys.

Re: NAND Boot Testing - 01-07: Panel power on/off fixes

So.... Taking the stuff from auto build. Just slap both files on the SD, flash and go?

Re: NAND Boot Testing - 01-07: Panel power on/off fixes

Ok. Here is what I found out...
First here is my partition table:

Code:

NBInfo 2.1rc2



'00_OS.nb.payload' has valid boot sector



Partition table:



Partition 0

-----------

 File System:    0x23 (XIP RAM)

 Start Sector:   0x00000002

 Total Sectors:  0x0000003e

 Boot indicator: 0x00

 First Head:     0x00

 First Sector:   0x00

 First Track:    0x00

 Last Head:      0x00

 Last Sector:    0x00

 Last Track:     0x00



Partition 1

-----------

 File System:    0x00 (unknown)

 Start Sector:   0x00000000

 Total Sectors:  0x00000000

 Boot indicator: 0x00

 First Head:     0x00

 First Sector:   0x00

 First Track:    0x00

 Last Head:      0x00

 Last Sector:    0x00

 Last Track:     0x00



Partition 2

-----------

 File System:    0x00 (unknown)

 Start Sector:   0x00000000

 Total Sectors:  0x00000000

 Boot indicator: 0x00

 First Head:     0x00

 First Sector:   0x00

 First Track:    0x00

 Last Head:      0x00

 Last Sector:    0x00

 Last Track:     0x00



Partition 3

-----------

 File System:    0x00 (unknown)

 Start Sector:   0x00000000

 Total Sectors:  0x00000000

 Boot indicator: 0x00

 First Head:     0x00

 First Sector:   0x00

 First Track:    0x00

 Last Head:      0x00

 Last Sector:    0x00

 Last Track:     0x00



Geometry: flash has 0 virtual heads





MSFLSH50 header found at offset 0x800

  (0 Reserved Entries, 0 Flash Region Entries)



Searching for IMGFS signature...

  Found IMGFS at byte 0x00380000 (sector 0x00380000).



  dwFSVersion:              00000001

  dwSectorsPerHeaderBlock:  00000001

  dwRunsPerFileHeader:      00000001

  dwBytesPerHeader:         00000034

  dwChunksPerSector:        00000020

  dwFirstHeaderBlockOffset: 00000800

  dwDataBlockSize:          00001000

  szCompressionType:        LZX

  dwFreeSectorCount:        00024682

  dwHiddenSectorCount:      00000040

  dwUpdateModeFlag:         00000000

---

As you can see partitions do not make that big of a difference. What seems to matter is having the imgfs found at 0x380000. I did the math in tinboot to do this. With a little playing around too.. ;)
This boots just fine now. (For me of course ;))

Nate I'm yours works because the tinboot is found somewhere on the on the partitions. Not sure why mine didn't work before. Your tinboot is getting inserted at 0x20000 to where mine is inserted at 0x2000. This is what Magldr roms use. :) I then had to change the .org in tinboot from 1000 to 2000. Now I'm not sure at all why this was not needed to be changed before but I think it has something to do with the not having the boot section in the partition table... Just a guess.

But I would like someone else to test my latest NBH.

Edit: Removed Attachment...

Re: NAND Boot Testing - 01-07: Panel power on/off fixes

Quote:

Originally Posted by coolsilver (Post 2060719)

So.... Taking the stuff from auto build. Just slap both files on the SD, flash and go?

No, you'll be missing a few files. Might want to go back and re-read the thread.

Androidupdate.tgz - Andboot folder
Androidinstall.tgz - Andboot Folder
install-seq.sh - Andboot Folder
RHODIMG.NBH - Sdcard Root

Still, re-read one more time before asking further support questions. I've done a good enough job at butting in to the development and I'm sure the developers wouldn't want to drag any more newbies in... if ya know what I mean. :mrgreen: