K4NT - Repair Needed, No drive found

[hostar]

Hello,

I messed up something with Kindle framework, but internally its fine,
only hasn't been closed properly.

Now it's stuck on "Kindle needs repair".
Problem is, it tries to connect like usb storage, but the disk seems like empty, so I can't access userstore.

When hard restart is performed, It will load progressbar to half and then jump back to Repair mode.

My questions are:
Have K4NT any "recovery mode" to connect over USB over fastboot or sb_loader?

Can someone upload compiled uboot.bin for K4NT or tell me, how to compile it with scratchbox?

Thanks

[geekmaster]

The empty drive is know in the GPL source code as "fstore" (fake storage) mode. It's purpose is to allow charging the kindle battery over a USB port, and it acts like a zero-byte USB Drive so that the host PC will keep supplying power to that USB port.

It is triggered automatically during boot if the battery voltage is too low to continue. It can also be triggered manually with a u-boot bist fstor command.

The "Kindle needs repair" can be repaired by resetting a boot counter in /var/local as documented in the "Kindle Touch Serial Port Recovery" thread, except now you can do that from a script file executed by ixtab's jailbreak method.

ixtab's method also works when booting from diagnostics mode, but the RUNME.sh script may need some adjustments (different root partition). You can mount the main root partition and make changes to it like this:
mount /dev/mmcblk0p1 /mnt/mmc
cd /mnt/mmc/etc/upstart
...
I have other ways to execute scripts that I am holding in reserve, so please use ixtab's method for now. Thanks.

I have custom patched k4nt u-boot.bin files named fastboot.bin, diag.bin, and main.bin. These can be loaded and run using MfgTool from freescale.com, but I have had very poor success being able to consistently run them using sb_loader. These patched and renamed u-boot.bin files boot the kindle 4 to the mode indicated by their name.

After you get MfgTool downloaded and installed in Windows, I can supply these patched files. In fact, I will post them to a new thread later today.

[hostar]

geekmaster: thanks for your reply.

According to your instructions, I was charging kindle for about 3 days,
but I'm still it that damn fakesore mode. Even on boot I can't see userstore for even one second.
So I couldn't to upload necessary files with ixtab's method.

I have managed to install MfgTool, sb_loader and fastboot.
I have also successfully boot to Recovery Mode.
In Recovery Mode I can see the kindle only with sb_loader.

I think, I can fix kindle if I would have access to userstore or usbnet.
Also kite runs successfully on boot.
But yifanlu wrote something about upcoming debricking tool,
so I don't know if I just shouldn't wait...

[geekmaster]

Here is another tool to help repairing kindle touch and k4nt:

https://www.mobileread.com/forums/sho...d.php?t=169645

[hostar]

With your KindleSelectBoot, I successfully (say 99%) debricked my Kindle.
Many thanks.

This 1% is this strange thing:
I can't run that "experimental" Internet browser.
I have checked for existing Browser.jar and browserd service.
When I stop browserd service and then run browserd binary, it writes this output:

Spoiler:

(process:4720): Gtk-WARNING **: Locale not supported by C library.
Using the fallback 'C' locale.

~~~~~~~~~~~~~~~~~~~~~~~~~~| DirectFB 1.2.0 |~~~~~~~~~~~~~~~~~~~~~~~~~~
(c) 2001-2008 The world wide DirectFB Open Source Community
(c) 2000-2004 Convergence (integrated media) GmbH
----------------------------------------------------------------

(*) DirectFB/Core: Single Application Core. (2011-09-03 04:53)
(!) Direct/Modules: Unable to dlopen `/usr/lib/directfb-1.2-0/systems/libdirectfb_devmem.so'!
--> /usr/lib/directfb-1.2-0/systems/libdirectfb_devmem.so: undefined symbol: system_set_mode
(*) Direct/Thread: Started 'Linux Input' (4723) [INPUT OTHER/OTHER 0/0] <8388608>...
(*) DirectFB/Input: tequila-keypad (1) 0.1 (directfb.org)
(*) Direct/Thread: Started 'Linux Input' (4724) [INPUT OTHER/OTHER 0/0] <8388608>...
(*) DirectFB/Input: fiveway (2) 0.1 (directfb.org)
(*) Direct/Thread: Started 'Linux Input' (4725) [INPUT OTHER/OTHER 0/0] <8388608>...
(*) DirectFB/Input: volume (3) 0.1 (directfb.org)
(*) DirectFB/Graphics: Generic Software Rasterizer 0.6 (directfb.org)
(*) DirectFB/Core/WM: Default 0.3 (directfb.org)
(*) FBDev/Mode: Setting 600x706 videomode x=600, y=800
(*) FBDev/Mode: Setting 600x706 videomode x=600, y=706
(*) FBDev/Mode: Setting 600x706 videomode x=800, y=506
(*) FBDev/Mode: Setting 600x706 videomode x=800, y=600
(*) FBDev/Mode: Setting 600x706 videomode x=600, y=800
(*) FBDev/Mode: Setting 600x706 videomode x=600, y=706
(*) FBDev/Mode: Setting 600x706 videomode x=800, y=506
(*) FBDev/Mode: Setting 600x706 videomode x=800, y=600
(*) FBDev/Mode: Setting 600x706 videomode x=600, y=800
(*) FBDev/Mode: Setting 600x706 videomode x=600, y=706
(*) FBDev/Mode: Setting 600x706 videomode x=800, y=506
(*) FBDev/Mode: Setting 600x706 videomode x=800, y=600
(*) FBDev/Mode: Setting 600x706 videomode x=600, y=800
(*) FBDev/Mode: Setting 600x706 videomode x=600, y=706
(*) FBDev/Mode: Setting 600x706 videomode x=800, y=506
(*) FBDev/Mode: Setting 600x706 videomode x=800, y=600
(*) FBDev/Mode: Setting 600x706 videomode x=600, y=800
(*) FBDev/Mode: Setting 600x706 videomode x=600, y=706
(*) FBDev/Mode: Setting 600x706 videomode x=800, y=506
(*) FBDev/Mode: Setting 600x706 videomode x=800, y=600
(*) FBDev/Surface: Allocated 600x706 8 bit LUT8 buffer (index 0) at offset 0 and pitch 600.
(*) FBDev/Mode: Setting 600x706 videomode x=600, y=800
(*) FBDev/Mode: Setting 600x706 videomode x=600, y=706
(*) FBDev/Mode: Setting 600x706 videomode x=800, y=506
(*) FBDev/Mode: Setting 600x706 videomode x=800, y=600
(!!!) *** WARNING [Non-flipping window surface and no 'autoflip-window' option used] *** [idirectfbsurface_window.c:310 in IDirectFBSurface_Window_Construct()]

(browserd:4720): Gdk-CRITICAL **: gdk_drawable_set_colormap: assertion `cmap == NULL || gdk_drawable_get_depth (drawable) == cmap->visual->depth' failed
(*) Direct/Thread: Started 'EventBufferFeed' (4726) [MESSAGING OTHER/OTHER 0/0] <8388608>...

I don't know, what should I check.
It's true, that I use that browser only rarely, but interested in fix it (just for fun).

[geekmaster]

Perhaps a missing or corrupted library or other file? You could loop mount a backup image and copy a replacement file from that.

[hostar]

First thing I've done when I found browser doesn't work was recovery
of Browser.jar from backup.

After I read your answer (thanks for that), I had an idea.
I checked (almost) all files on Kindle with md5sum,
and did same thing with files in backup.

Now the curious thing: even if I have already replaced the Browser.jar, it's md5 was different from backup!
Also some other files has different md5.
So I replaced the Browser.jar again.

But the browser actually run after correcting file called "internal.conf", where
was on last line:
com.lab126.booklet.a.Y instead of com.lab126.booklet.a.y

So everything is fine now, I'm writing this only because someone might have similar problem in future.

[geekmaster]

Originally Posted by hostar

First thing I've done when I found browser doesn't work was recovery
of Browser.jar from backup.

After I read your answer (thanks for that), I had an idea.
I checked (almost) all files on Kindle with md5sum,
and did same thing with files in backup.

Now the curious thing: even if I have already replaced the Browser.jar, it's md5 was different from backup!
Also some other files has different md5.
So I replaced the Browser.jar again.

But the browser actually run after correcting file called "internal.conf", where
was on last line:
com.lab126.booklet.a.Y instead of com.lab126.booklet.a.y

So everything is fine now, I'm writing this only because someone might have similar problem in future.

Thank you! That looks like something useful to add to my kindle "recovery system".

For people with bricked kindles who are anxiously waiting more progress on these tools, sorry about the delay -- a tiny mistake to my payload while testing new stuff bricked my touch in a way that I have not yet discovered how to repair, despite a lot of time diverted from development to repair attempts. I have also diverted a lot of time to helping others fix their kindles, in PMs and IRC private channel conversations (there has been a very high success rate). Now, if I could just succeed on my own Touch). Time to stop poking around in USB methods and go deeper into serial port recovery like I used the first time. I think the Touch does not work as well in fastboot mode as my K4NT did ("fastboot flash system mmcblk0p1.imf" reports "success" much to quickly, and it does not fix the problem). My soldering iron is warming up now...