Unbriking Kindle DX

[twobob]

It's nice to see some real-world additions were made to the code.
Heart warming.

Awesome little tips as well. This is gold-dust info mate. thanks

[knc1]

here is another interesting bit:

/*
 * autoscript allows a remote host to download a command file and,
 * optionally, binary data for automatically updating the target. For
 * example, you create a new kernel image and want the user to be
 * able to simply download the image and the machine does the rest.
 * The kernel image is postprocessed with mkimage, which creates an
 * image with a script file prepended. If enabled, autoscript will
 * verify the script and contents of the download and execute the
 * script portion. This would be responsible for erasing flash,
 * copying the new image, and rebooting the machine.
 */

Now if we could make it work . . .
(And I can find what memory address to write the stinking file image to ...)

[twobob]

dang that would be SO peachy. This is So praise worthy. Qualitat, as Ixtab might say : )

[knc1]

[RANT]
Regular readers here are familar with GM's opinion on excessive whitespace in program code.

Here is mine:
Defines belong in the "include file" (*.h) ;
Code belongs in the "implementation file" (*.c) ;

That recommended practice is unknown to the folks at lab126.
[/RANT]

And to think, somebody actually paid those people to write this stuff.

[twobob]

They would be level 0 in your dojo sensei. The fight may be over but the war has just begun. Or other similar rousing Hollywood comments.

Looking at the Glasshouse of "functional" coding horrors my lineage have produced, well... Sufficed to say keeping ones head down is probably in order

Beautiful = expensive, every manager knows that.

[knc1]

But that organization which lab126 used means you find things where you least expect them.

Like the need to put the UART and charge controller setup into the *.c file where the defines are at - - even thought that *.c file has absolutely nothing at all to do with the UART or battery charger.
[/BAD_MOOD]

- - - -

I thought, thought I:
"Simple, review a few *.h files in the iMX31 branch of the u-boot file system tree, find everything I needed."

Yeah, Right! That was about a week ago. A week spent reading this stuff until my eyeballs feel like they are falling out.

Instead, what it takes (almost can write: what it took) is to build u-boot for the DX(G) and then read every file mentioned on the load map.

[twobob]

its a form of obfuscation far beyond simple name switching and function remapping.

It's a ninja level where at it's deepest roots something impossibly twisted lies.

And that's how "prime number" applications are born. Forged in the blazing heat of contracted managerial pressures. Divisible by nothing but itself, defying understandingly without grasping the complete whole at least once. and maybe not even then. Popping up when you least expect them. The larger the project the more impossible to envisage.

Ah. Analogies. Happy days.

[knc1]

It is a true, tri-color LED - -
There is White and Blue-Green also defined in here.

Blue-Green == Your device is sick and about to barf.

And if it starts blinking 4 sec on / 4 sec off - stand back.

(and you will not find that in the "LED code" file.)

[cavorite]

Aaaaaand we're back from our commercial break. I am now again talking and listening to the DX via the serial port. Situation is as it was before the converter died, in a loop, can get out at u-boot and it looks functional, can get out at recovery console but will not stop at the password prompt, back to loop.

When you have something to test, let me know.

[knc1]

Break out of the autoboot sequence, into the u-boot prompt -
Have USB cable plugged into something -
I learned today that the battery charger is enabled while sitting at the u-boot> prompt.

No need to open device and remove the battery - lab126 did that so they could take the weekend off without the Kindle they where working on, running down its battery.

[cavorite]

Done, but there's an inactivity timeout running that kicks out of u-boot if I'm not doing something with the keyboard. So now on the charger, monitoring serial port, stuck in loop.

[knc1]

Originally Posted by cavorite

Done, but there's an inactivity timeout running that kicks out of u-boot if I'm not doing something with the keyboard. So now on the charger, monitoring serial port, stuck in loop.

Maybe enter the first character of a command, let it charge while waiting for the next character.

Duh...
They did not give directions about how to make that "feature" work in the source code, just a note that it was added in by lab126.

[geekmaster]

Originally Posted by knc1

[RANT]
Regular readers here are familar with GM's opinion on excessive whitespace in program code.

Here is mine:
Defines belong in the "include file" (*.h) ;
Code belongs in the "implementation file" (*.c) ;

That recommended practice is unknown to the folks at lab126.
[/RANT]

And to think, somebody actually paid those people to write this stuff.

Sadly, we cannot use eink include files if we want to support 5.0.x and 5.1.x kindles in one source module. The files for both firmwares have the same file name. The structures passed to the eink update ioctl() call have the same structure name, but DIFFERENT layouts. Instead of adding the new fields to the end of the structure to maintain compatibility, the inserted them in the middle, pushing critical fields to a different position in the different firmware versions.

I like to use includes, because it uses less space than copying the stuctures into your main source module.

I really do not like excessive vertical scrolling, whether it is through large source files with too much white space, or scrolling through directories filled gazillions of tiny source modules. I have five monitors in front of me, and I have nowhere near enough pixels for my needs. I hate important information NOT being right there in front of me, but having to search in buried windows, or unopened files, or offscreen in some scroll buffer. Using the mouse too much hurts my hand and fingers... Yuck! I need a WALL of pixels.

[cavorite]

Nope, entered "h" and it hit timeout and exited to the loop.

It (the DX) has evidently has been continuously looping for the past few days, as it was immediately active on the serial port when I connected. Doesn't appear to have done it any harm. I hope.

[knc1]

Originally Posted by cavorite

Nope, entered "h" and it hit timeout and exited to the loop.

It (the DX) has evidently has been continuously looping for the past few days, as it was immediately active on the serial port when I connected. Doesn't appear to have done it any harm. I hope.

I gotta take a break, eyeballs have "red-lined" for today.

On a brighter (off-topic) note - -
My EeePC-701 (which runs 24/7 as my modem control, firewall and gateway device) reports an uptime of 214 days since I last re-booted it.

I think that is longer than my K3 lasted from arrival to death.