k3flasher, a USB download mode utility - Page 6

knc1 · 08-25-2012, 10:39 AM

Devices (well, device registers) are memory mapped . . .
When the iROM code starts, there is a single, 32bit, identity mapping installed in the TLB . . .
The iROM code can read (and write) memory locations . . .

If the battery management chip's "battery charge level" address can be found, it might be possible to read the battery charge level before continuing with the download of the "RAM kernel".

If it is mapped into the higher 60Gbytes of the address space, then another TLB entry would have to be made to read it.

This would give the application a chance to advise the user if the K3's battery should be re-charged before continuing.

- - - -

I seem to recall seeing battery level check messages in the kernel's dmesg listing during boot.
If so, it might be possible to find that "battery charge level" address in the (Amazon patched) kernel source code.

I'll PM you the details if I stumble across them, but don't wait on me, I may never stumble on them.

hawhill · 08-25-2012, 11:30 AM

Oh, this is a great idea! Will need to dig a bit into the kernel/reference manual to see where SoC registers are mapped... However, it might be more complicated, the battery management might be external and connected via SPI or I2C or the likes...

geekmaster · 08-25-2012, 11:58 AM

The u-boot source code should have an example of how to read battery status, because it decides how to boot depending on battery charge. I remember seeing that code when I made the patched u-boot images for the select boot thread. Of course, that was for K4/K5, but it should apply to the K3 as well.

I will not be available for many hours starting now, or I would check into it myself. If nobody else finds it, I will look in u-boot for the battery status check code (when I get back)...

knc1 · 08-25-2012, 12:27 PM

Quote:

Originally Posted by hawhill

Oh, this is a great idea! Will need to dig a bit into the kernel/reference manual to see where SoC registers are mapped... However, it might be more complicated, the battery management might be external and connected via SPI or I2C or the likes...

I don't recall if iROM can start the SPI, I2C and/or the one-wire controllers.
It has been over a week since I read the documents, will have to re-read them.

But checking the battery charge level **before** starting up the eDDR2 ram would be nice.
(That is what some of the memory addresses being poked at do, start the eDDR2 ram in self-refresh mode.)

In case you missed it: an3996.pdf is the document to have on-hand.

- - - -
O.T:
I don't recall if I took your username/password off the webdav://kloud - I can put them back if you want. That protected kloud is gathering a lot of "limited-distribution" documents.

Xaber · 08-26-2012, 04:01 AM

I try to flash my kindle.
I use command

Code:

sudo ./k3flasher mx35to2_mmc.bin program rootfs mmcblk0p1.dmp

In result i have

Code:

I: flashed 0x00010000 (=65536) bytes
I: flashed 0x00020000 (=131072) bytes
I: flashed 0x00030000 (=196608) bytes
I: flashed 0x00040000 (=262144) bytes
I: flashed 0x00050000 (=327680) bytes
I: flashed 0x00060000 (=393216) bytes
I: flashed 0x00070000 (=458752) bytes
I: flashed 0x00080000 (=524288) bytes
I: flashed 0x00090000 (=589824) bytes
I: flashed 0x000a0000 (=655360) bytes
I: flashed 0x000b0000 (=720896) bytes
I: flashed 0x000c0000 (=786432) bytes
I: flashed 0x000d0000 (=851968) bytes
I: flashed 0x000e0000 (=917504) bytes
I: flashed 0x000f0000 (=983040) bytes
I: flashed 0x00100000 (=1048576) bytes
I: flashed 0x00110000 (=1114112) bytes
I: flashed 0x00120000 (=1179648) bytes
I: flashed 0x00130000 (=1245184) bytes
I: flashed 0x00140000 (=1310720) bytes
I: flashed 0x00150000 (=1376256) bytes
I: flashed 0x00160000 (=1441792) bytes
I: flashed 0x00170000 (=1507328) bytes
I: flashed 0x00180000 (=1572864) bytes
I: flashed 0x00190000 (=1638400) bytes
I: flashed 0x001a0000 (=1703936) bytes
I: flashed 0x001b0000 (=1769472) bytes
I: flashed 0x001c0000 (=1835008) bytes
I: flashed 0x001d0000 (=1900544) bytes
I: flashed 0x001e0000 (=1966080) bytes
I: flashed 0x001f0000 (=2031616) bytes
I: flashed 0x00200000 (=2097152) bytes
I: flashing of 0x00200000 (=2097152) bytes complete
I: writing 0x00200000 (=2097152) bytes to address 0x059c1000, waiting for completion...
I: wrote 0x00200000 (=2097152) bytes, waiting for completion...
I: flashed 0x00010000 (=65536) bytes
I: flashed 0x00020000 (=131072) bytes
I: flashed 0x00030000 (=196608) bytes
I: flashed 0x00040000 (=262144) bytes
I: flashed 0x00050000 (=327680) bytes
I: flashed 0x00060000 (=393216) bytes
I: flashed 0x00070000 (=458752) bytes
I: flashed 0x00080000 (=524288) bytes
I: flashed 0x00090000 (=589824) bytes
I: flashed 0x000a0000 (=655360) bytes
I: flashed 0x000b0000 (=720896) bytes
I: flashed 0x000c0000 (=786432) bytes
I: flashed 0x000d0000 (=851968) bytes
I: flashed 0x000e0000 (=917504) bytes
I: flashed 0x000f0000 (=983040) bytes
I: flashed 0x00100000 (=1048576) bytes
I: flashed 0x00110000 (=1114112) bytes
I: flashed 0x00120000 (=1179648) bytes
I: flashed 0x00130000 (=1245184) bytes
I: flashed 0x00140000 (=1310720) bytes
I: flashed 0x00150000 (=1376256) bytes
I: flashed 0x00160000 (=1441792) bytes
I: flashed 0x00170000 (=1507328) bytes
I: flashed 0x00180000 (=1572864) bytes
I: flashed 0x00190000 (=1638400) bytes
I: flashed 0x001a0000 (=1703936) bytes
I: flashed 0x001b0000 (=1769472) bytes
I: flashed 0x001c0000 (=1835008) bytes
I: flashed 0x001d0000 (=1900544) bytes
I: flashed 0x001e0000 (=1966080) bytes
I: flashed 0x001f0000 (=2031616) bytes
I: flashed 0x00200000 (=2097152) bytes
I: flashing of 0x00200000 (=2097152) bytes complete
I: writing 0x00200000 (=2097152) bytes to address 0x05bc1000, waiting for completion...
I: wrote 0x00200000 (=2097152) bytes, waiting for completion...

Again and again!

Xaber · 08-26-2012, 06:08 AM

I try with kernel and understood! maybe in v8 you can make progressbar, because its long time process, and on display we see same information long time.

hawhill · 08-26-2012, 07:27 AM

Well, I'm a hacker/programmer guy, I hate progress bars :-) But then I guess you're right, it could display a constantly incrementing value instead of displaying only the steps up to 2MByte. But if you look carefully, you'll see it's not identical over and over again: The address (displayed for each 2MByte block) is incrementing. But I see the point - and then, not everyone is fluent with hexadecimal notation either...

Glad to hear you worked it out in the end :-)

knc1 · 08-26-2012, 07:54 AM

Quote:

Originally Posted by Xaber

I try to flash my kindle.
I use command

Code:

sudo ./k3flasher mx35to2_mmc.bin program rootfs mmcblk0p1.dmp

In result i have

Code:

I: flashed 0x00010000 (=65536) bytes
I: flashed 0x00020000 (=131072) bytes
I: flashed 0x00030000 (=196608) bytes
I: flashed 0x00040000 (=262144) bytes
I: flashed 0x00050000 (=327680) bytes
I: flashed 0x00060000 (=393216) bytes
I: flashed 0x00070000 (=458752) bytes
I: flashed 0x00080000 (=524288) bytes
I: flashed 0x00090000 (=589824) bytes
I: flashed 0x000a0000 (=655360) bytes
I: flashed 0x000b0000 (=720896) bytes
I: flashed 0x000c0000 (=786432) bytes
I: flashed 0x000d0000 (=851968) bytes
I: flashed 0x000e0000 (=917504) bytes
I: flashed 0x000f0000 (=983040) bytes
I: flashed 0x00100000 (=1048576) bytes
I: flashed 0x00110000 (=1114112) bytes
I: flashed 0x00120000 (=1179648) bytes
I: flashed 0x00130000 (=1245184) bytes
I: flashed 0x00140000 (=1310720) bytes
I: flashed 0x00150000 (=1376256) bytes
I: flashed 0x00160000 (=1441792) bytes
I: flashed 0x00170000 (=1507328) bytes
I: flashed 0x00180000 (=1572864) bytes
I: flashed 0x00190000 (=1638400) bytes
I: flashed 0x001a0000 (=1703936) bytes
I: flashed 0x001b0000 (=1769472) bytes
I: flashed 0x001c0000 (=1835008) bytes
I: flashed 0x001d0000 (=1900544) bytes
I: flashed 0x001e0000 (=1966080) bytes
I: flashed 0x001f0000 (=2031616) bytes
I: flashed 0x00200000 (=2097152) bytes
I: flashing of 0x00200000 (=2097152) bytes complete
I: writing 0x00200000 (=2097152) bytes to address 0x059c1000, waiting for completion...
I: wrote 0x00200000 (=2097152) bytes, waiting for completion...
I: flashed 0x00010000 (=65536) bytes
I: flashed 0x00020000 (=131072) bytes
I: flashed 0x00030000 (=196608) bytes
I: flashed 0x00040000 (=262144) bytes
I: flashed 0x00050000 (=327680) bytes
I: flashed 0x00060000 (=393216) bytes
I: flashed 0x00070000 (=458752) bytes
I: flashed 0x00080000 (=524288) bytes
I: flashed 0x00090000 (=589824) bytes
I: flashed 0x000a0000 (=655360) bytes
I: flashed 0x000b0000 (=720896) bytes
I: flashed 0x000c0000 (=786432) bytes
I: flashed 0x000d0000 (=851968) bytes
I: flashed 0x000e0000 (=917504) bytes
I: flashed 0x000f0000 (=983040) bytes
I: flashed 0x00100000 (=1048576) bytes
I: flashed 0x00110000 (=1114112) bytes
I: flashed 0x00120000 (=1179648) bytes
I: flashed 0x00130000 (=1245184) bytes
I: flashed 0x00140000 (=1310720) bytes
I: flashed 0x00150000 (=1376256) bytes
I: flashed 0x00160000 (=1441792) bytes
I: flashed 0x00170000 (=1507328) bytes
I: flashed 0x00180000 (=1572864) bytes
I: flashed 0x00190000 (=1638400) bytes
I: flashed 0x001a0000 (=1703936) bytes
I: flashed 0x001b0000 (=1769472) bytes
I: flashed 0x001c0000 (=1835008) bytes
I: flashed 0x001d0000 (=1900544) bytes
I: flashed 0x001e0000 (=1966080) bytes
I: flashed 0x001f0000 (=2031616) bytes
I: flashed 0x00200000 (=2097152) bytes
I: flashing of 0x00200000 (=2097152) bytes complete
I: writing 0x00200000 (=2097152) bytes to address 0x05bc1000, waiting for completion...
I: wrote 0x00200000 (=2097152) bytes, waiting for completion...

Again and again!

Those messages are reporting the progress of each 2Mbyte chunk.
How big is the image? 600Mbytes or something like that?
So expect that message to be repeated 300 times.

The remote (Kindle end) of the software is designed to work with a large number of devices.
To keep it simple, it is written to the lowest common denominator of all those device's features.
Translation: Its slow, its meant to be that way by the author (Freescale), just wait it out.

qlob · 08-26-2012, 02:47 PM

@Hawhill:
Did you ever end up adding the RAM kernel to K3flasher? I only ask becuase I wiped my Windows partition and now have no easy way of obtaining it. (I really, really, really, don't want to install Windows.......again........)

knc1 · 08-26-2012, 03:35 PM

Quote:

Originally Posted by qlob

@Hawhill:
Did you ever end up adding the RAM kernel to K3flasher? I only ask becuase I wiped my Windows partition and now have no easy way of obtaining it. (I really, really, really, don't want to install Windows.......again........)

If you have a Linux install of recent vintage, just double click on the *.exe, it will open in Wine and install. No M$-Windows required.

By default, the path will be: ~/.wine/drive_c/<path seen in Windows>

OR

Depending on your distribution, the *.exe will open in the archive manager. Which also gives you access to the RAM kernels.

geekmaster · 08-26-2012, 03:50 PM

AFAIK Wine has rather limited USB support.

knc1 · 08-26-2012, 03:55 PM

Quote:

Originally Posted by geekmaster

AFAIK Wine has rather limited USB support.

You don't need USB support to **install** the files in the *.exe cabinet.
And that was the question, gaining access to the ram kernel file.

qlob · 08-26-2012, 04:25 PM

Quote:

Originally Posted by knc1

If you have a Linux install of recent vintage, just double click on the *.exe, it will open in Wine and install. No M$-Windows required.

By default, the path will be: ~/.wine/drive_c/<path seen in Windows>

OR

Depending on your distribution, the *.exe will open in the archive manager. Which also gives you access to the RAM kernels.

My package manager is giving me errors about wine and broken dependencies.......

The archive manager gives me an error about the exe....

I already tried both of these.... Probably should have mentioned that.

knc1 · 08-26-2012, 04:46 PM

Please mention the name/version of your distribution.
Perhaps also the link to the atk binary your trying to open.

I have to go deal with other things at the moment, but someone can help you.

qlob · 08-26-2012, 06:16 PM

My distribution is Fedora 16/17 (see spoiler for details)

Spoiler:

I got the ATK binary from freescale's website.
Direct Link: http://styles.freescale.com/files/so...L_STD_1_67.zip

08-25-2012, 10:39 AM	#76
knc1 Going Viral Posts: 17,212 Karma: 18210809 Join Date: Feb 2012 Location: Central Texas Device: No K1, PW2, KV, KOA	@ Hawhill Devices (well, device registers) are memory mapped . . . When the iROM code starts, there is a single, 32bit, identity mapping installed in the TLB . . . The iROM code can read (and write) memory locations . . . If the battery management chip's "battery charge level" address can be found, it might be possible to read the battery charge level before continuing with the download of the "RAM kernel". If it is mapped into the higher 60Gbytes of the address space, then another TLB entry would have to be made to read it. This would give the application a chance to advise the user if the K3's battery should be re-charged before continuing. - - - - I seem to recall seeing battery level check messages in the kernel's dmesg listing during boot. If so, it might be possible to find that "battery charge level" address in the (Amazon patched) kernel source code. I'll PM you the details if I stumble across them, but don't wait on me, I may never stumble on them.

08-25-2012, 11:58 AM	#78
geekmaster Carpe diem, c'est la vie. Posts: 6,433 Karma: 10773668 Join Date: Nov 2011 Location: Multiverse 6627A Device: K1 to PW3	The u-boot source code should have an example of how to read battery status, because it decides how to boot depending on battery charge. I remember seeing that code when I made the patched u-boot images for the select boot thread. Of course, that was for K4/K5, but it should apply to the K3 as well. I will not be available for many hours starting now, or I would check into it myself. If nobody else finds it, I will look in u-boot for the battery status check code (when I get back)... Last edited by geekmaster; 08-25-2012 at 12:01 PM.

08-26-2012, 06:16 PM	#90
qlob Official Lurker Posts: 1,050 Karma: 7096675 Join Date: Apr 2012 Device: Kindle 3.4	My distribution is Fedora 16/17 (see spoiler for details) Spoiler: I used Preupgrade to upgrade from 16 to 17, has caused some errors in the dependencies department... for info: https://fedoraproject.org/wiki/How_to_use_PreUpgrade I got the ATK binary from freescale's website. Direct Link: http://styles.freescale.com/files/so...L_STD_1_67.zip

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[Kindle Touch] Boot over USB HID serial / "USB downloader" mode	eureka	Kindle Developer's Corner	16	02-25-2012 10:21 PM
USB Host Mode (Master Mode) on K3	ericepe	Kindle Developer's Corner	1	01-24-2012 04:59 AM
USB Drive Mode	copyrite	Amazon Kindle	7	02-08-2011 10:47 AM
USB Drive Mode on K3	Paulinafrica	Amazon Kindle	5	12-17-2010 12:43 AM
Where did the USB transfer mode go?	codo coderson	HanLin eBook	2	11-28-2010 07:04 AM

08-25-2012, 11:30 AM	#77
hawhill Wizard Posts: 1,379 Karma: 2155307 Join Date: Nov 2010 Location: Goettingen, Germany Device: Kindle Paperwhite, Kobo Mini	Oh, this is a great idea! Will need to dig a bit into the kernel/reference manual to see where SoC registers are mapped... However, it might be more complicated, the battery management might be external and connected via SPI or I2C or the likes...

08-26-2012, 06:08 AM	#81
Xaber Junior Member Posts: 2 Karma: 10 Join Date: Aug 2012 Device: Kindle 3	I try with kernel and understood! maybe in v8 you can make progressbar, because its long time process, and on display we see same information long time.

08-26-2012, 07:27 AM	#82
hawhill Wizard Posts: 1,379 Karma: 2155307 Join Date: Nov 2010 Location: Goettingen, Germany Device: Kindle Paperwhite, Kobo Mini	Well, I'm a hacker/programmer guy, I hate progress bars :-) But then I guess you're right, it could display a constantly incrementing value instead of displaying only the steps up to 2MByte. But if you look carefully, you'll see it's not identical over and over again: The address (displayed for each 2MByte block) is incrementing. But I see the point - and then, not everyone is fluent with hexadecimal notation either... Glad to hear you worked it out in the end :-)

08-26-2012, 02:47 PM	#84
qlob Official Lurker Posts: 1,050 Karma: 7096675 Join Date: Apr 2012 Device: Kindle 3.4	@Hawhill: Did you ever end up adding the RAM kernel to K3flasher? I only ask becuase I wiped my Windows partition and now have no easy way of obtaining it. (I really, really, really, don't want to install Windows.......again........)

08-26-2012, 03:50 PM	#86
geekmaster Carpe diem, c'est la vie. Posts: 6,433 Karma: 10773668 Join Date: Nov 2011 Location: Multiverse 6627A Device: K1 to PW3	AFAIK Wine has rather limited USB support.

08-26-2012, 04:46 PM	#89
knc1 Going Viral Posts: 17,212 Karma: 18210809 Join Date: Feb 2012 Location: Central Texas Device: No K1, PW2, KV, KOA	Please mention the name/version of your distribution. Perhaps also the link to the atk binary your trying to open. I have to go deal with other things at the moment, but someone can help you.

Advert

Advert