Quill OS - Sony PRS-T1

https://prodgpl.blob.core.windows.net/download/Audio/common/2UrziX9Zanih0kfZRY4U3Q/linux-2.6.35.2-20140522-2.tgz

Source code download

5:06 PM

5:07 PM

https://oss.sony.net/Products/Linux/Audio/PRS-T1JP_20140702.html

Source code download

Nice

tux-linux

U-Boot version, U-Boot log, kernel position in MMC, load address, etc

besides watching the uart can you explain more verbosely where these things are so I can find them for you maybe through the recovery console?

You need to get into U-Boot, execute help, look for version command, then execute it, execute boot and gather all the output even from the begginging, thats it

8:15 PM

Propably

Do you think there is a way to "get into U-Boot" from the ADB Shell?

No

10:40 PM

Usually you interrupt the boot process by pressing a key on the keyboard right after you press the power button

8.11 KB

MX50_ARM2 U-Boot > help ? - alias for 'help' autoscr - DEPRECATED - use "source" command instead base - print or set address offset bdinfo - print Board Info structure boot - boot default, i.e., run 'bootcmd' bootd - boot default, i.e., run 'bootcmd' bootm - boot application image from memory bootp - boot image via network using BOOTP/TFTP protocol cls - clear screen cmp - memory compare coninfo - print console devices and information cp - memory copy crc32 - checksum calculation dhcp - boot image via network using DHCP/TFTP protocol echo - echo args to console erase - erase FLASH memory ext2load- load binary file from a Ext2 filesystem ext2ls - list files in a directory (default /) fatinfo - print information about filesystem fatload - load binary file from a dos filesystem fatls - list files in a directory (default /) flinfo - print FLASH memory information go - start application at address 'addr' help - print online help i2c - I2C sub-system iminfo - print header information for application image imxtract- extract a part of a multi-image itest - return true/false on integer compare loadb - load binary file over serial line (kermit mode) loads - load S-Record file over serial line loady - load binary file over serial line (ymodem mode) loop - infinite loop on address range md - memory display mii - MII utility commands mm - memory modify (auto-incrementing address) mmc - MMC sub system mmcinfo - mmcinfo <dev num>-- display MMC info (edited)

https://discord.com/channels/809205711778480158/943711473341960253/963910834751426634

2.27 KB

1:09 AM

mtest - simple RAM read/write test mw - memory write (fill) nfs - boot image via network using NFS protocol nm - memory modify (constant address) ping - send ICMP ECHO_REQUEST to network host printenv- print environment variables protect - enable or disable FLASH write protection rarpboot- boot image via network using RARP/TFTP protocol reset - Perform RESET of the CPU run - run commands in an environment variable saveenv - save environment variables to persistent storage setenv - set environment variables sf - SPI flash sub-system sleep - delay execution for some time source - run script from memory sspi - SPI utility commands tftpboot- boot image via network using TFTP protocol version - print monitor version MX50_ARM2 U-Boot >

MX50_ARM2 U-Boot > base Base Address: 0x00000000

1:23 AM

MX50_ARM2 U-Boot > coninfo List of available devices: lcd 00000002 ..O serial 80000003 SIO stdin stdout stderr nulldev 80000003 SIO

1:24 AM

MX50_ARM2 U-Boot > version U-Boot 2009.08 (Dec 21 2011 - 14:48:38)

MX50_ARM2 U-Boot > iminfo

Checking Image at 70800000 ...

Unknown image format!

1:32 AM

@tux-linux anything else?

There’s the faulty Sandisk eMMC under the e-ink flex cable.

That PRSTUX set this for their build...

stdin=serial
stdout=serial
stderr=serial
ethact=FEC0
loadaddr=0x70800000
loadaddr_ramdisk=0x70C00000
bootdev=2
rawtable=0xF40000
bootcmd=mmc read 2 ${loadaddr} 0x800 0x1400;bootm ${loadaddr}
bootargs=root=/dev/mmcblk0p1 rootfstype=ext4 rw rootwait init=/sbin/init bootdev=2 rawtable=0xF40000

Thanks for the logs It seems pretty similar to the Touch B on the bare hardware part

What do you think would happen if I just tried to boot the touch B image?

12:40 PM

Is that load addr the right position for that?

1. Nothing, it panicks 2. magic smoke comes out of some chip 3. the battery starts burning

Are you speaking from exp?

not experience, from my knowledge

@tux-linux what version of Debian do you use for compiling?

it isin't really that much diffrence, just not an old one. use 11 or 10

Szybet

1. Nothing, it panicks 2. magic smoke comes out of some chip 3. the battery starts burning

The eink screen will probably go bye-bye

1:15 PM

(forever)

hm, why exactly? details?

NonaSuomy

Is that load addr the right position for that?

It should load the kernel fine But don't try it

Szybet

hm, why exactly? details?

Well, I'm pretty certain it's a completely different eink screen, so the waveform will mess it up completely

sad that it doesn't have any protection

NonaSuomy

@tux-linux what version of Debian do you use for compiling?

11, 10 will do

1:20 PM

Also the Touch B image is too big for the MMC

Are you talking about the 2GB eMMC that is corrupted? Using an external 4GB SDCard

waveform is on the internal emmc right?...

3:47 PM

you won't get far if you don't replace this flash :/

I already have it booting from the SDCard with the OEM os.

its steal reading from the flash

3:49 PM

it will kill it one day completly

We’ll deal with that when that time comes.

3:51 PM

What does the waveform data do? Controls how the screen updates?

its the information for the kernel / cpu how and what to write to the e-ink

Where is it normally stored?

on a specified location on a flash

What does it look like

3:54 PM

here it is from the nia

3:54 PM

its model specific, but still maybe you will find it

NonaSuomy

Are you talking about the 2GB eMMC that is corrupted? Using an external 4GB SDCard

Does everything load from the external MMC?

it would be nice if it did

Well holding onto home and menu triggers the device to update from SD so maybe if you know what to do from there.

4:43 PM

Does the device know what to do with those buttons before it touches loading commands from the eMMC? If there was no internal SD card in the KOBO would it still do something?

4:45 PM

Is there some kind of hardware bios before the internal memory module.

Someone should make a bga to sdcard slot

7:31 PM

Lots of devices to go from eMMC to SD but not SD to eMMC

NonaSuomy

There’s the faulty Sandisk eMMC under the e-ink flex cable.

https://www.digchip.com/datasheets/download_datasheet.php?id=7693769&part-number=SDIN5D1-2G-LT

7:38 PM

I got the folders docker-build 14.04 / rootfs / splash to make but not boot or 12.04 (if this is needed?)

7:48 PM

https://gitlab.com/phreakuencies/prstux/-/blob/master/boot/Makefile

boot/Makefile · master · Matias N. / prstux

PRS-TUX: Linux system for Sony PRS-T* ebook readers

7:48 PM

Anything there standout that may be a tool that I don't have?

7:49 PM

think this mkubootenv maybe something.

7:49 PM

maybe it's this https://github.com/tklauser/mkubootenv

GitHub - tklauser/mkubootenv: U-Boot environment image creator

U-Boot environment image creator. Contribute to tklauser/mkubootenv development by creating an account on GitHub.

7:51 PM

I think that checks out... usage: mkubootenv [-s <size>] <source file> <target file> mkubootenv/mkubootenv -s 131072 boot-env.txt bootenv

Probably, but why not use setenv in U-Boot console?

not sure I'm just trying to get this to build the prstux image not trying to do anything a different way just yet.

1

7:55 PM

Are the steps some place to build the InkBox Touch B build?

Hmm... Not really You can try imgtool, but there's not a written walkthrough

What dependencies would be required?

A lot

Can you list them for me?

7:56 PM

I'll start installing everything in this VM

Too much to remember I'll try to write them down later Look in the emu repo, in the readme there are some you should install

What do you think would have to be mainly modified in the Touch B firmware to get this working on the PRS-T1? (edited)

Eink related stuff

I just tried to run it. It strangely blanked the screen then showed the regular loading screen but the bar never moved heard some squealing so I just reset it and put the other card back in seems ok.

Whoa, I'm impressed it got there

8:54 PM

Well, first thing would be to compile a kernel from Sony's sources

How do you go about that?

Have no idea what I built with prstux repo but it completed and after a small tweak of pointing it to a file instead of the card it spit out a 900K image not sure what it’s for.

10:43 PM

I see this here but also not sure what it’s for… http://build.koreader.rocks/download/nightly/v2022.03.1-40-g4e517f48_2022-04-22/koreader-sony-prstux-v2022.03.1-40-g4e517f48_2022-04-22.zip

10:46 PM

Pokes @NiLuJe

NonaSuomy

Have no idea what I built with prstux repo but it completed and after a small tweak of pointing it to a file instead of the card it spit out a 900K image not sure what it’s for.

I think you need to set up a block device with a file Some things (partition, dd) won't work otherwise Read about qemu-nbd

11:18 PM

qemu-img create -f qcow2 prstux.img 3.6G
modprobe nbd
qemu-nbd --connect /dev/nbd0 prstux.img

Then point your script to nbd0 (edited)

11:19 PM

To write it to the SD, dd from nbd0 and not the image (edited)

Building PRSTUX repo from fresh debian 11 install (not sure what I'm doing yet...)

sudo apt install git  linux-headers-$(uname -r) build-essential docker curl debootstrap qemu qemu-utils qemu-user-static binfmt-support imagemagick vim
sudo modprobe binfmt_misc
sudo curl -sSL https://get.docker.com/ | sh
mkdir code
cd code
git clone https://github.com/tklauser/mkubootenv.git 
cd mkubootenv
sudo make
cd ..
git clone https://gitlab.com/phreakuencies/prstux.git
cd prstux
cp ~/code/mkubootenv/mkubootenv ~/code/prstux/mkubootenv/

Fix broken Makefile replace spaces with tabs

vim /home/nonasuomy/code/prstux/splash/Makefile
:%s/^[ ]\+/\t/g

remove the first ../ in front of the script path as well. sudo make flash DEVICE=./prstux.img (edited)

10:22 AM

Output

~/code/prstux$ make flash DEVICE=./prstux.img
make -C boot
make[1]: Entering directory '/home/nonasuomy/code/prstux/boot'
mkubootenv/mkubootenv -s 131072 boot-env.txt bootenv
make[1]: Leaving directory '/home/nonasuomy/code/prstux/boot'
make -C boot flash
make[1]: Entering directory '/home/nonasuomy/code/prstux/boot'
Flashing...
../scripts/flash_normal_bootenv.sh ./prstux.img
Will flash into device: ./prstux.img
Continue? Type 'yes': yes
+ echo 'Zeroing normal env area'
Zeroing normal env area
+ sudo dd if=/dev/zero of=./prstux.img bs=512 seek=30720 count=256
256+0 records in
256+0 records out
131072 bytes (131 kB, 128 KiB) copied, 0.00130886 s, 100 MB/s
+ echo 'Flashing env'
Flashing env
+ sudo dd if=bootenv of=./prstux.img bs=512 seek=30720 count=256
256+0 records in
256+0 records out
131072 bytes (131 kB, 128 KiB) copied, 0.00254849 s, 51.4 MB/s
+ sudo dd if=bootenv of=./prstux.img bs=512 seek=1536 count=256
256+0 records in
256+0 records out
131072 bytes (131 kB, 128 KiB) copied, 0.00657853 s, 19.9 MB/s
+ echo Synching
Synching
+ sync
+ sync
+ sync
Done.
make[1]: Leaving directory '/home/nonasuomy/code/prstux/boot'
make -C splash flash
make[1]: Entering directory '/home/nonasuomy/code/prstux/splash'
convert splash.png -size 800x600 -depth 8 gray:bootsplash.bin
Flashing...
../scripts/flash_bootsplash.sh ./prstux.img
Flashing bootsplash
480000+0 records in
480000+0 records out
480000 bytes (480 kB, 469 KiB) copied, 1.40263 s, 342 kB/s
Synching
Done.
make[1]: Leaving directory '/home/nonasuomy/code/prstux/splash'

tux-linux

qemu-img create -f qcow2 prstux.img 3.6G
modprobe nbd
qemu-nbd --connect /dev/nbd0 prstux.img

Then point your script to nbd0 (edited)

ok so you don't need to do this?

10:22 AM

and provide nbd0 as DEVICE env var?

not sure where are you finding this?

Because dd-ing to files doesn't work as you expect

Do you see that process in any of their build files?

No But they're using the MMC as DEVICE

10:24 AM

Which is a block device, just like nbd0, and not a file

NonaSuomy

Output

~/code/prstux$ make flash DEVICE=./prstux.img
make -C boot
make[1]: Entering directory '/home/nonasuomy/code/prstux/boot'
mkubootenv/mkubootenv -s 131072 boot-env.txt bootenv
make[1]: Leaving directory '/home/nonasuomy/code/prstux/boot'
make -C boot flash
make[1]: Entering directory '/home/nonasuomy/code/prstux/boot'
Flashing...
../scripts/flash_normal_bootenv.sh ./prstux.img
Will flash into device: ./prstux.img
Continue? Type 'yes': yes
+ echo 'Zeroing normal env area'
Zeroing normal env area
+ sudo dd if=/dev/zero of=./prstux.img bs=512 seek=30720 count=256
256+0 records in
256+0 records out
131072 bytes (131 kB, 128 KiB) copied, 0.00130886 s, 100 MB/s
+ echo 'Flashing env'
Flashing env
+ sudo dd if=bootenv of=./prstux.img bs=512 seek=30720 count=256
256+0 records in
256+0 records out
131072 bytes (131 kB, 128 KiB) copied, 0.00254849 s, 51.4 MB/s
+ sudo dd if=bootenv of=./prstux.img bs=512 seek=1536 count=256
256+0 records in
256+0 records out
131072 bytes (131 kB, 128 KiB) copied, 0.00657853 s, 19.9 MB/s
+ echo Synching
Synching
+ sync
+ sync
+ sync
Done.
make[1]: Leaving directory '/home/nonasuomy/code/prstux/boot'
make -C splash flash
make[1]: Entering directory '/home/nonasuomy/code/prstux/splash'
convert splash.png -size 800x600 -depth 8 gray:bootsplash.bin
Flashing...
../scripts/flash_bootsplash.sh ./prstux.img
Flashing bootsplash
480000+0 records in
480000+0 records out
480000 bytes (480 kB, 469 KiB) copied, 1.40263 s, 342 kB/s
Synching
Done.
make[1]: Leaving directory '/home/nonasuomy/code/prstux/splash'

Is it still a 900K image?

little bit larger now after I corrected the Makefile error as it seems to have added the splash haha.

10:25 AM

-rw-r--r-- 1 nonasuomy nonasuomy 17257216 Apr 24 10:17 prstux.img

hmm what if you flash it to an MMC?

Does 17MB seem like enough?

...no? But I don't really know PRSTux, maybe it is enough

I feel like nothing is in there no kernel or system files etc

then try with the qemu-img thing above

Wouldn't that just make an empty image that is 3.6G then it would splice that 17MB into it?

10:36 AM

I feel like something is missing

qcow2 is dynamically allocated (edited)

Where are the linux os files coming from

so if you try to see the size of the img file, it will show how much has been written to it (approximately)

10:37 AM

that's why you need to dd from nbd0 and not the img itself

He seems to only have modules from sony in the directory no kernel

it says it, there are two parts

10:38 AM

first that 17mb nonsense, then apparently there is a prstux.img somewhere

10:39 AM

unless there isn't (edited)

there isn't

but you should make all then

10:40 AM

Hmm maybe this prstux-base image...

10:43 AM

sudo make all
[sudo] password for nonasuomy: 
make -C boot
make[1]: Entering directory '/home/nonasuomy/code/prstux/boot'
make[1]: Nothing to be done for 'all'.
make[1]: Leaving directory '/home/nonasuomy/code/prstux/boot'
make -C splash
make[1]: Entering directory '/home/nonasuomy/code/prstux/splash'
make[1]: 'bootsplash.bin' is up to date.
make[1]: Leaving directory '/home/nonasuomy/code/prstux/splash'
make -C docker-build 14.04
make[1]: Entering directory '/home/nonasuomy/code/prstux/docker-build'
docker build -t prstux-dev:14.04 14.04
Sending build context to Docker daemon   2.56kB
Step 1/8 : FROM ubuntu:14.04
 ---> 13b66b487594
Step 2/8 : RUN apt-get -y update
 ---> Using cache
 ---> e2966287f94f
Step 3/8 : RUN apt-get install -y emdebian-archive-keyring
 ---> Using cache
 ---> 68f078683adf
Step 4/8 : RUN apt-get install -y build-essential git bison autoconf gperf flex texinfo help2man gawk libtool libncurses5-dev unzip bc
 ---> Using cache
 ---> 510d48b1e3ab
Step 5/8 : RUN apt-get install -y pkg-config-arm-linux-gnueabihf
 ---> Using cache
 ---> df7b42790e3a
Step 6/8 : RUN apt-get install -y u-boot-tools
 ---> Using cache
 ---> 5f70785bb8bb
Step 7/8 : RUN apt-get install -y wget vim
 ---> Using cache
 ---> 46afb16fe8b3
Step 8/8 : RUN apt-get install -y debootstrap qemu-user-static
 ---> Using cache
 ---> 322a2df2c107
Successfully built 322a2df2c107
Successfully tagged prstux-dev:14.04
make[1]: Leaving directory '/home/nonasuomy/code/prstux/docker-build'
make -C rootfs
make[1]: Entering directory '/home/nonasuomy/code/prstux/rootfs'
./make-base.sh
prstux-base docker image already exists. use make clean_base to remove if you want to rebuild it
make[1]: Leaving directory '/home/nonasuomy/code/prstux/rootfs'

10:45 AM

He says this:

Build
TODO: not finished
In case you prefer to recreate PRSTUX image yourself, you can do so by cloning the repository and simply do:

$ make


This will produce a .tar file containing the contents of the root filesystem and a ready-made image appropriate for burning into an SD card. The .tar file is useful if you want to update PRSTUX after an update but want to leave your data (such as eBooks) intact.

10:46 AM

I'm guessing by the terminal output above that it created a docker container but never output a build file from it as a tar as said above.

10:49 AM

https://gitlab.com/phreakuencies/prstux/-/blob/master/rootfs/make-base.sh

rootfs/make-base.sh · master · Matias N. / prstux

PRS-TUX: Linux system for Sony PRS-T* ebook readers

10:51 AM

How would we output that docker to a writable image? (edited)

Hmm I guess it does…

11:25 AM

‘’’sudo tar -C rootfs -c . | docker import - prstux-base:latest sudo rm -rf rootfs/‘’’

Hmm maybe because I ran everything separately to figure out the dependancies and maybe it broke the flow with those files already existing in each of the folders.

I guess thats actually dumping the contents of rootfs to the docker running.

that's what it seems it does, yeah

mount: permission denied
update-binfmts: warning: Couldn't mount the binfmt_misc filesystem on /proc/sys/fs/binfmt_misc.

(edited)

you need to enable that feature in your kernel

1:13 PM

or maybe it is availble as a module

I believe I modprobed it above

that's strange

1:16 PM

it's a warning, I don't know if it is fatal or not

Does it need sudo modprobe binfmt_misc inside the docker build or binfmt_misc is for the host?

afaik, docker doesn't use its own kernel/VM

1:29 PM

so, on the host

It says mount: permission denied above the warning (edited)

probably sudo then

1:31 PM

and docker probably isn't running as root

1:31 PM

modprobe on your host

binfmt is trying to register qemu as a binary loader so you don't need to invoke qemu everytime you need to run a binary with it. To do that it requires privileges that the container doesn't have.

Ignore those warnings because it doesn't make sense to register binfmt on the host while the image is being built.

where's that written?

https://stackoverflow.com/questions/54951262/binfmt-misc-problems-in-ubuntu18-04-under-docker

binfmt_misc problems in ubuntu18.04 under docker

Trying to get a cross chroot running under ubuntu:18.04 in a docker container. While trying to install binfmt-support (dependency of qemu-user-static) this happens: ~ docker run -it ubuntu:18.04 ...

well, from an educated guess, prstux img builder is probably trying to use qemu-arm or qemu-user-arm, which requires binfmt most of the time

1:36 PM

it would make sense, given it's building an image for an ARM device

no expliations though conflicts with what this person is saying as well https://community.nxp.com/t5/Layerscape/LSDK1909-weston-compilation-failure-on-custom-docker-container/m-p/1026748

LSDK1909 weston compilation failure on custom docker container

I am having trouble building a ls1088ardb_pb image (with CONFIG_APP_SECURE_OBJ and CONFIG_APP_OPTEE enabled) via LSDK 19.09 on a docker amd64 Ubuntu 18.04.3 container. I have seen previous questions about this (LSDK weston build error , compilation failure flexbuild_lsdk1909_update_221019 "app" com...

1:38 PM

"I figured out the issue: I had been installing “sudo apt install -y binfmt-support qemu-system-common qemu-user-static" within the container, rather than the host.After installing it on the host, the issue was resolved."

(edited)

Tried both ways and still has the error so we're just going to ignore it like you said haha... (edited)

A semi-appropriate thing to do would be to actually install Ubuntu 14.04 on a VM and use the tools that were available at the time

1:53 PM

Though I don't know if you could still access the packages server

1:53 PM

Probably archive.ubuntu.com

Maybe

1:59 PM

I: Base system installed successfully.
sha256:f60291bf47337abb8193d56f1dd1b8dfae0bafd245cb93f6ece11260a4f64e60
make[1]: Leaving directory '/home/nonasuomy/code/prstux/rootfs'

2:00 PM

ok so what now build thing...

that's suspicious

2:00 PM

but anyway, we'll see

Checked all the folders nothing came out

*yay*

2:03 PM

we really should just get the kernel sources

2:03 PM

after that we can setup inkbox

I think the scripts are missing the output part from docker after it does this massaging

2:05 PM

Lead the way senpai https://prodgpl.blob.core.windows.net/download/Audio/common/2UrziX9Zanih0kfZRY4U3Q/linux-2.6.35.2-20140522-2.tgz

but this is not for android right?

This device stock runs android 2.2.1 is that what you are asking?

yes, there are differences between android and linux kernel

2:08 PM

even though the Android one is Linux-based

2:08 PM

but maybe they don't matter

:~/code/prstux$ sudo docker images
REPOSITORY    TAG       IMAGE ID       CREATED             SIZE
prstux-base   latest    f60291bf4733   13 minutes ago      342MB
prstux-dev    14.04     9fb33fb8ac9b   22 minutes ago      527MB
<none>        <none>    d97cb24c140d   About an hour ago   524MB
ubuntu        14.04     13b66b487594   13 months ago       197MB

2:17 PM

I'm guessing the image would have to be close to one of those in size.

2:22 PM

I think scripting like this is missing https://iximiuz.com/en/posts/from-docker-container-to-bootable-linux-disk-image/

From Docker Container to Bootable Linux Disk Image

Dump Docker container's filesystem to tar archive and covert it to bootable Linux disk image.

2:25 PM

"We couldn't find any code results matching docker export in project Matias N. / prstux"

Maybe it didn't exist at the time

Maybe or they got side tracked after getting that far...

2:44 PM

CID=$(docker run -d prstux-base /bin/true)

docker export -o prstux.tar ${CID}

sudo tar -tf 
prstux.tar | grep -E '^[^/]*/?$'
.dockerenv
bin/
boot/
dev/
etc/
home/
lib/
media/
mnt/
opt/
proc/
root/
run/
sbin/
srv/
sys/
tmp/
usr/
var/

(edited)

tux-linux

but maybe they don't matter

So what needs to happen?

first, try to compile a kernel from those sources and get it to boot on the Sony

5:03 PM

after that we can include it in the official kernel repo and start porting

tux-linux

first, try to compile a kernel from those sources and get it to boot on the Sony

How do you accomplish this process?

git clone https://github.com/Kobo-InkBox/kernel
export PATH="${PATH}:/kernel/toolchain/gcc-4.8/bin"
cd "/sony/kernel/source"
make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- uImage

(edited)

5:09 PM

Then dd if=uImage of=/dev/mmcblk0 bs=512 seek=2048 the seek value may not be the same if sony put the load sector elsewhere

NonaSuomy

That PRSTUX set this for their build...

stdin=serial
stdout=serial
stderr=serial
ethact=FEC0
loadaddr=0x70800000
loadaddr_ramdisk=0x70C00000
bootdev=2
rawtable=0xF40000
bootcmd=mmc read 2 ${loadaddr} 0x800 0x1400;bootm ${loadaddr}
bootargs=root=/dev/mmcblk0p1 rootfstype=ext4 rw rootwait init=/sbin/init bootdev=2 rawtable=0xF40000

it will be 2048 (0x800) (edited)

https://discord.com/channels/809205711778480158/966554812567130152/967330075253043230

5:11 PM

Does that explain where it is or this is something different?

the bootcmd part tells it

5:12 PM

mmc read 2 ${loadaddr} 0x800 0x1400 means read the MMC 2 from sector 2048 to sector 5120

Ok will see how this goes!

1

Too many cooks: what was the actual ping about?

NiLuJe

Too many cooks: what was the actual ping about?

https://youtu.be/QrGrOK8oZG8

Adult Swim

Too Many Cooks | Adult Swim

6:55 PM

Did you have a PRS-T2/T1 and have that PRSTUX running? (edited)

tux-linux

git clone https://github.com/Kobo-InkBox/kernel
export PATH="${PATH}:/kernel/toolchain/gcc-4.8/bin"
cd "/sony/kernel/source"
make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- uImage

(edited)

~/code/linux-2.6.35.2$ sudo make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- uImage
make: arm-linux-gnueabihf-gcc: No such file or directory
  HOSTCC  scripts/basic/fixdep
  HOSTCC  scripts/basic/docproc
  HOSTCC  scripts/basic/hash
  HOSTCC  scripts/kconfig/conf.o
scripts/kconfig/conf.c: In function ‘conf_sym’:
scripts/kconfig/conf.c:159:6: warning: variable ‘type’ set but not used [-Wunused-but-set-variable]
  159 |  int type;
      |      ^~~~
scripts/kconfig/conf.c: In function ‘conf_choice’:
scripts/kconfig/conf.c:231:6: warning: variable ‘type’ set but not used [-Wunused-but-set-variable]
  231 |  int type;
      |      ^~~~
  HOSTCC  scripts/kconfig/kxgettext.o
  SHIPPED scripts/kconfig/zconf.tab.c
  SHIPPED scripts/kconfig/lex.zconf.c
  SHIPPED scripts/kconfig/zconf.hash.c
  HOSTCC  scripts/kconfig/zconf.tab.o
  HOSTLD  scripts/kconfig/conf
scripts/kconfig/conf -s arch/arm/Kconfig
***
*** You have not yet configured your kernel!
*** (missing kernel config file ".config")
***
*** Please run some configurator (e.g. "make oldconfig" or
*** "make menuconfig" or "make xconfig").
***
make[2]: *** [/home/nonasuomy/code/linux-2.6.35.2/scripts/kconfig/Makefile:34: silentoldconfig] Error 1
make[1]: *** [/home/nonasuomy/code/linux-2.6.35.2/Makefile:464: silentoldconfig] Error 2
make: *** No rule to make target 'include/config/auto.conf', needed by 'include/config/kernel.release'.  Stop.

env | grep PATH?

env | grep PATH
PATH=/home/nonasuomy/bin:/usr/local/bin:/usr/bin:/bin:/usr/local/games:/usr/games:/snap/bin:/home/nonasuomy/code/kernel/toolchain/gcc-4.8/bin

If you run arm-linux-gnueabihf-gcc, what does it output?

8:03 PM

don't use sudo to make

8:04 PM

i think it might actually be your problem

make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- uImage
scripts/kconfig/conf -s arch/arm/Kconfig
***
*** You have not yet configured your kernel!
*** (missing kernel config file ".config")
***
*** Please run some configurator (e.g. "make oldconfig" or
*** "make menuconfig" or "make xconfig").
***
make[2]: *** [/home/nonasuomy/code/linux-2.6.35.2/scripts/kconfig/Makefile:34: silentoldconfig] Error 1
make[1]: *** [/home/nonasuomy/code/linux-2.6.35.2/Makefile:464: silentoldconfig] Error 2
make: *** No rule to make target 'include/config/auto.conf', needed by 'include/config/kernel.release'.  Stop.

then, the toolchain problem is fine now fiddle in arch/arm/configs and try to see if you see something relevant for the Sony

~/code/linux-2.6.35.2/arch/arm/configs$ ls
acs5k_defconfig             mx21_defconfig
acs5k_tiny_defconfig        mx27_defconfig
afeb9260_defconfig          mx31pdk_defconfig
am200epdkit_defconfig       mx3_defconfig
am3517_evm_defconfig        mx3_updater_defconfig
ams_delta_defconfig         mx51_defconfig
ap4evb_defconfig            n770_defconfig
assabet_defconfig           n8x0_defconfig
at572d940hfek_defconfig     neocore926_defconfig
at91cap9adk_defconfig       neponset_defconfig
at91rm9200dk_defconfig      netwinder_defconfig
at91rm9200ek_defconfig      netx_defconfig
at91sam9260ek_defconfig     nhk8815_defconfig
at91sam9261ek_defconfig     ns9xxx_defconfig
at91sam9263ek_defconfig     nuc910_defconfig
at91sam9g20ek_defconfig     nuc950_defconfig
at91sam9rlek_defconfig      nuc960_defconfig
ateb9200_defconfig          omap_2430sdp_defconfig
badge4_defconfig            omap_3430sdp_defconfig
bcmring_defconfig           omap_3630sdp_defconfig
cam60_defconfig             omap3_beagle_defconfig
carmeva_defconfig           omap3_defconfig
cerfcube_defconfig          omap3_evm_defconfig
cm_t35_defconfig            omap3_pandora_defconfig
cm_x2xx_defconfig           omap3_stalker_lks_defconfig
cm_x300_defconfig           omap3_touchbook_defconfig
cns3420vb_defconfig         omap_4430sdp_defconfig
colibri_pxa270_defconfig    omap_apollon_2420_defconfig
colibri_pxa300_defconfig    omap_generic_1510_defconfig
collie_defconfig            omap_generic_1610_defconfig
corgi_defconfig             omap_generic_1710_defconfig
cpu9260_defconfig           omap_generic_2420_defconfig
cpu9g20_defconfig           omap_h2_1610_defconfig
cpuat91_defconfig           omap_h4_2420_defconfig
csb337_defconfig            omap_innovator_1510_defconfig
csb637_defconfig            omap_innovator_1610_defconfig
da8xx_omapl_defconfig       omap_ldp_defconfig
davinci_all_defconfig       omap_osk_5912_defconfig
devkit8000_defconfig        omap_perseus2_730_defconfig

8:10 PM

dove_defconfig              omap_zoom2_defconfig
ebsa110_defconfig           omap_zoom3_defconfig
ecbat91_defconfig           onearm_defconfig
edb7211_defconfig           orion5x_defconfig
em_x270_defconfig           overo_defconfig
ep93xx_defconfig            palmte_defconfig
eseries_pxa_defconfig       palmtt_defconfig
ezx_defconfig               palmz71_defconfig
footbridge_defconfig        palmz72_defconfig
fortunet_defconfig          pcm027_defconfig
g3evm_defconfig             picotux200_defconfig
g4evm_defconfig             pleb_defconfig
h3600_defconfig             pnx4008_defconfig
h5000_defconfig             pxa168_defconfig
h7201_defconfig             pxa255-idp_defconfig
h7202_defconfig             pxa3xx_defconfig
hackkit_defconfig           pxa910_defconfig
htcherald_defconfig         qil-a9260_defconfig
igep0020_defconfig          raumfeld_defconfig
imote2_defconfig            realview_defconfig
imx23evk_defconfig          realview-smp_defconfig
imx23evk_updater_defconfig  rpc_defconfig
imx25_3stack_defconfig      rx51_defconfig
imx25_updater_defconfig     s3c2410_defconfig
imx28evk_defconfig          s3c6400_defconfig
imx28evk_updater_defconfig  s5p6440_defconfig
imx5_android_defconfig      s5p6442_defconfig
imx5_defconfig              s5pc100_defconfig
imx5_updater_defconfig      s5pc110_defconfig

8:11 PM

integrator_defconfig        s5pv210_defconfig
iop13xx_defconfig           sam9_l9260_defconfig
iop32x_defconfig            shannon_defconfig
iop33x_defconfig            shark_defconfig
ixp2000_defconfig           simpad_defconfig
ixp23xx_defconfig           spear300_defconfig
ixp4xx_defconfig            spear310_defconfig
jornada720_defconfig        spear320_defconfig
kafa_defconfig              spear600_defconfig
kb9202_defconfig            spitz_defconfig
kirkwood_defconfig          stamp9g20_defconfig
ks8695_defconfig            stmp378x_defconfig
lart_defconfig              stmp37xx_defconfig
loki_defconfig              sx1_defconfig
lpd270_defconfig            tct_hammer_defconfig
lpd7a400_defconfig          trizeps4_defconfig
lpd7a404_defconfig          u300_defconfig
lubbock_defconfig           u8500_defconfig
lusl7200_defconfig          usb-a9260_defconfig
magician_defconfig          usb-a9263_defconfig
mainstone_defconfig         versatile_defconfig
mini2440_defconfig          viper_defconfig
mmp2_defconfig              xcep_defconfig
msm_defconfig               yl9200_defconfig
mv78xx0_defconfig           zeus_defconfig
mx1_defconfig

8:11 PM

This one is highlighted imx5_android_defconfig

8:16 PM

imx5_android_defconfig

53.7 KB

8:16 PM

See anything there that might need changing?

8:22 PM

Definitely the right file "CONFIG_TOUCHSCREEN_SONYIR=y" What do you do with this file now?

make ARCH=arm imx5_android_defconfig CROSS_COMPILE=arm-linux-gnueabihf- uImage
arch/arm/configs/imx5_android_defconfig:514:warning: override: reassigning to symbol NETFILTER_XT_CONNMARK
#
# configuration written to .config
#
scripts/kconfig/conf -s arch/arm/Kconfig

*** Error during update of the kernel configuration.

make[3]: *** [/home/nonasuomy/code/linux-2.6.35.2/scripts/kconfig/Makefile:34: silentoldconfig] Error 1
make[2]: *** [/home/nonasuomy/code/linux-2.6.35.2/Makefile:464: silentoldconfig] Error 2
make[1]: *** No rule to make target 'include/config/auto.conf', needed by 'include/config/kernel.release'.  Stop.
make: *** [Makefile:444: uImage] Error 2

try the two commands separately?

8:40 PM

make ARCH=arm CORSS_COMPILE=arm-linux-gnueabihf- imx5_android_defconfig make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- uImage

nonasuomy@DebianCompiler:~/code/linux-2.6.35.2$ make ARCH=arm CORSS_COMPILE=arm-linux-gnueabihf- imx5_android_defconfig
arch/arm/configs/imx5_android_defconfig:514:warning: override: reassigning to symbol NETFILTER_XT_CONNMARK
#
# configuration written to .config
#
nonasuomy@DebianCompiler:~/code/linux-2.6.35.2$ make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- uImage
scripts/kconfig/conf -s arch/arm/Kconfig

*** Error during update of the kernel configuration.

make[2]: *** [/home/nonasuomy/code/linux-2.6.35.2/scripts/kconfig/Makefile:34: silentoldconfig] Error 1
make[1]: *** [/home/nonasuomy/code/linux-2.6.35.2/Makefile:464: silentoldconfig] Error 2
make: *** No rule to make target 'include/config/auto.conf', needed by 'include/config/kernel.release'.  Stop.

strange.

8:43 PM

make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- distclean make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- mrproper

8:43 PM

then run the two commands above again

yay it's doing stuff!

1

8:45 PM

Can't use 'defined(@array)' (Maybe you should just omit the defined()?) at kernel/timeconst.pl line 373.
make[1]: *** [/home/nonasuomy/code/linux-2.6.35.2/kernel/Makefile:138: kernel/timeconst.h] Error 255
make: *** [Makefile:884: kernel] Error 2

vi kernel/timeconst.pl go to line 373 replace the existing statement with if (!@val) {

8:47 PM

so it looks like this

Compiling again...

1

8:58 PM

Did you have to do the same?

8:58 PM

for the kobo

on those kernels, yes

8:58 PM

not the distclean/mrproper bit though

any more of those gotchas?

i don't think so

1

I always shoot myself in the foot with vm size I told myself to always use 40GB and some dumb reason I put it to 20GB and have like 99mb so this compile will probably die

9:01 PM

prob need to expand the vdi.

yea

  CC      net/core/gen_stats.o
In file included from net/core/gen_stats.c:18:0:
include/linux/module.h:212:22: fatal error: closing dependency file net/core/.gen_stats.o.d: No space left on device
  static const struct kernel_symbol __ksymtab_##sym \
                      ^
include/linux/module.h:218:2: note: in expansion of macro ‘__EXPORT_SYMBOL’
  __EXPORT_SYMBOL(sym, "")
  ^
net/core/gen_stats.c:252:1: note: in expansion of macro ‘EXPORT_SYMBOL’
 EXPORT_SYMBOL(gnet_stats_finish_copy);
 ^
compilation terminated.
make[2]: *** [scripts/Makefile.build:231: net/core/gen_stats.o] Error 1
make[1]: *** [scripts/Makefile.build:366: net/core] Error 2
make: *** [Makefile:884: net] Error 2

9:03 PM

haha

9:03 PM

no space left on device

well...

9:03 PM

expand your disk

speak of the devil

or compile in ram

9:04 PM

(if you have enough)

how much space in total does it require?

not that much

9:04 PM

512M more I would say

Purged all the Sony files going again for now 450MB free will see if there's anything else hiding.

9:09 PM

596MB free hope that's enough (edited)

9:10 PM

How long did the compile take on your vm?

~50s

1

9:11 PM

but my VM is fast

9:11 PM

you can try with make jobs -j4, -j8, etc.

  CC      net/socket.o
net/core/.scm.o.cmd:3: *** unterminated variable reference.  Stop.
make[1]: *** [scripts/Makefile.build:366: net/core] Error 2
make: *** [Makefile:884: net] Error 2

no idea

(edited)

But maybe the fix is easy

scm.o.cmd

20 KB

I may have got it someone said quote marks shouldn't be in make files even though it wasn't directly related I removed that and it started compiling again

1

9:43 PM

-D"KBUILD_STR(s)=#s" 
-D"KBUILD_BASENAME=KBUILD_STR(scm)"  
-D"KBUILD_MODNAME=KBUILD_STR(scm)"

to

-DKBUILD_STR(s)=#s 
-DKBUILD_BASENAME=KBUILD_STR(scm)  
-DKBUILD_MODNAME=KBUILD_STR(scm)

(edited)

9:46 PM

nope

9:47 PM

  CC      net/socket.o
net/core/.scm.o.cmd:3: *** unterminated variable reference.  Stop.
make[1]: *** [scripts/Makefile.build:366: net/core] Error 2
make: *** [Makefile:884: net] Error 2

Sooo not saying this is sane to do but I only found 3 instance of that file mentioned in google and did a search in this repo https://github.com/ananjaser1211/Refinedkernel_2/search?q=.scm.o.cmd

Search · .scm.o.cmd · ananjaser1211/Refinedkernel_2

Refined Kernel For Galaxy J5/J7 2017 - V2.8. Contribute to ananjaser1211/Refinedkernel_2 development by creating an account on GitHub.

10:18 PM

/.gitignore
net/core/.request_sock.o.cmd
net/core/.rtnetlink.o.cmd
net/core/.scm.o.cmd
net/core/.secure_seq.o.cmd
net/core/.skbuff.o.cmd

10:19 PM

I saw here that it was on the ignore list... so I deleted the file...

hmm and?

Image Name:   Linux-2.6.35.3
Created:      Sun Apr 24 22:16:10 2022
Image Type:   ARM Linux Kernel Image (uncompressed)
Data Size:    2246720 Bytes = 2194.06 KiB = 2.14 MiB
Load Address: 70008000
Entry Point:  70008000
  Image arch/arm/boot/uImage is ready

yay!

10:21 PM

but it's possible it won't load because it seems it doesn't load enough from the MMC

10:21 PM

you'll have to change the env try it before

What do you mean?

It might not load all of the uImage

10:22 PM

only read the MMC up to a certain point

tux-linux

Then dd if=uImage of=/dev/mmcblk0 bs=512 seek=2048 the seek value may not be the same if sony put the load sector elsewhere

So now this... I'm just writing this to a blank SD card?

No, the SD card needs to contain at least U-Boot

./tux-linux --verbose

1

tl;dr, just dd the kernel to an existing working sony firmware card

What about if I do it to this Rupor's recovery will that work?

i don't really know

I'm just not sure what android will do with it if I do it to the official os

10:30 PM

I guess we can just try them all...

10:31 PM

do we care about the full os loading or just enough to get a console?

10:31 PM

rupors would just be a console

enough to get a kernel panic (edited)

10:32 PM

that will let us know that it's working (roughly)

NonaSuomy

Did you have a PRS-T2/T1 and have that PRSTUX running? (edited)

Nope, but supposedly the author did ;p

1

Nothing so far

11:56 PM

not even uboot

12:01 AM

just an orange flashing light

Ok got something

12:27 AM

U-Boot 2009.08 (Dec 21 2011 - 14:48:38)

CPU:   Freescale i.MX50 family 1.1V at 800 MHz
mx50 pll1: 800MHz
mx50 pll2: 400MHz
mx50 pll3: 216MHz
ipg clock     : 66666666Hz
ipg per clock : 66666666Hz
uart clock    : 24000000Hz
ahb clock     : 133333333Hz
axi_a clock   : 400000000Hz
axi_b clock   : 200000000Hz
weim_clock    : 100000000Hz
ddr clock     : 160000000Hz
esdhc1 clock  : 80000000Hz
esdhc2 clock  : 80000000Hz
esdhc3 clock  : 80000000Hz
esdhc4 clock  : 80000000Hz
Board: MX50 ARM2 board
Boot Reason: [POR]
Boot Device: MMC
I2C:   ready
DRAM:  256 MB
MMC:   FSL_ESDHC: 0, FSL_ESDHC: 1, FSL_ESDHC: 2
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
temperature 31
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
In:    serial
Out:   serial
Err:   serial
Net:   got MAC address from IIM: 00:00:00:00:00:00
FEC0
Hit any key to stop autoboot:  0

MMC read: dev # 0, block # 2048, count 5120 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
SD set clock 50MHz
5120 blocks read: OK

12:27 AM

MMC read: dev # 0, block # 10240, count 600 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
SD set clock 50MHz
600 blocks read: OK
## Booting kernel from Legacy Image at 70800000 ...
   Image Name:   Linux-2.6.35.3
   Image Type:   ARM Linux Kernel Image (uncompressed)
   Data Size:    2246720 Bytes =  2.1 MB
   Load Address: 70008000
   Entry Point:  70008000
   Verifying Checksum ... OK
## Loading init Ramdisk from Legacy Image at 70c00000 ...
   Image Name:   Normal Rootfs
   Image Type:   ARM Linux RAMDisk Image (uncompressed)
   Data Size:    286848 Bytes = 280.1 kB
   Load Address: 70308000
   Entry Point:  70308000
   Verifying Checksum ... OK
   Loading Kernel Image ... OK
OK

12:27 AM

Starting kernel ...

Linux version 2.6.35.3 (nonasuomy@DebianCompiler) (gcc version 4.8.1 20130401 (prerelease) (crosstool-NG linaro-1.13.1-4.8-2013.04-20130417 - Linaro GCC 2013.04) ) #1 PREEMPT Sun Apr 24 22:16:04 EDT 2022
CPU: ARMv7 Processor [412fc085] revision 5 (ARMv7), cr=10c53c7f
CPU: VIPT nonaliasing data cache, VIPT nonaliasing instruction cache
Machine: Freescale MX50 ARM2 Board
Ignoring unrecognised tag 0x54410008
Memory policy: ECC disabled, Data cache writeback
Built 1 zonelists in Zone order, mobility grouping on.  Total pages: 63754
Kernel command line: console=ttymxc2,115200 init=/init bootdev=0 rawtable=0xF40000
PID hash table entries: 1024 (order: 0, 4096 bytes)
Dentry cache hash table entries: 32768 (order: 5, 131072 bytes)
Inode-cache hash table entries: 16384 (order: 4, 65536 bytes)
Memory: 251MB = 251MB total
Memory: 249236k/249236k available, 7788k reserved, 0K highmem
Virtual kernel memory layout:
    vector  : 0xffff0000 - 0xffff1000   (   4 kB)
    fixmap  : 0xfff00000 - 0xfffe0000   ( 896 kB)
    DMA     : 0xf9e00000 - 0xffe00000   (  96 MB)
    vmalloc : 0x90000000 - 0xf4000000   (1600 MB)
    lowmem  : 0x80000000 - 0x8fb00000   ( 251 MB)
    pkmap   : 0x7fe00000 - 0x80000000   (   2 MB)
    modules : 0x7f000000 - 0x7fe00000   (  14 MB)
      .init : 0x80008000 - 0x80025000   ( 116 kB)
      .text : 0x80025000 - 0x80412000   (4020 kB)
      .data : 0x8042e000 - 0x804a86c0   ( 490 kB)
SLUB: Genslabs=9, HWalign=64, Order=0-3, MinObjects=0, CPUs=1, Nodes=1
Hierarchical RCU implementation.
        RCU-based detection of stalled CPUs is disabled.
        Verbose stalled-CPUs detection is disabled.
NR_IRQS:396
MXC GPIO hardware
MXC IRQ initialized
You should not call the gpmi_set_parent
MXC_Early serial console at MMIO 0x63f90000 (options '115200')
bootconsole [ttymxc2] enabled
Console: colour dummy device 80x30
Mix log static memory : 7fe00000 - 7fffffff
              address : 90400000
  There is not the last log.

12:28 AM

Initialize a static memory for mix_logger.

12:31 AM

So nothing panic'd?

12:32 AM

What would be the next step?

12:36 AM

uImage

2.14 MB

12:39 AM

I used the Stock OS here https://jpa.kapsi.fi/stuff/other/sd_card_PRS-T1_1.0.07_adb_enabled.bin.xz

ok, well, even better if nothing panicked

8:37 AM

congrats, we have a working kernel

8:37 AM

the screen flashes alright?

Flashes but then restarts and flashes again.

9:25 AM

I believe there is supposed to be more to the log but maybe that is because no os booted?

9:26 AM

I searched "There is not the last log." and there is only one result of someone capturing the boot process in gist... https://gist.github.com/nrdmn/c75f8c8011ca82def7f28e4e23d05ed8

Sony PRS-T1

Sony PRS-T1. GitHub Gist: instantly share code, notes, and snippets.

9:29 AM

Maybe it died Initializing a static memory for mix_logger?

9:29 AM

What do you need next?

NonaSuomy

Flashes but then restarts and flashes again.

Is it with the stock Sony system?

yes

Whoops

10:06 AM

Do you know how to embed an initrd in a kernel?

No sir.

10:08 AM

What need to be done to do that?

Ok

10:08 AM

Get busybox source

10:08 AM

Tell me when you have

Any specific version?

No

10:09 AM

Latest

10:09 AM

Latest stable

3 May 2021 -- BusyBox 1.33.1 (stable) Does this version make sense?

Ok yeah

30 November 2021 -- BusyBox 1.33.2 (stable)

Latest then

26 December 2021 -- BusyBox 1.35.0 (unstable)

Latest stable

10:12 AM

So November

10:12 AM

Do you have gcc 4.8 in your path?

Isn't that what was in your kernel package?

Yes

10:13 AM

So get it in your path

10:14 AM

Then make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf defconfig

10:14 AM

make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf -j8 install

10:14 AM

Tell me when it's done

Rebuilding the VM will be back in action soon.

1

ok back!

3:52 PM

export PATH="${PATH}:/home/nonasuomy/code/busybox"
make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- imx5_android_defconfig
make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf -j8 install

(edited)

3:57 PM

hmmm it's trying to install that kernel on my machine... and getting a permission denied

what exactly error?

sbin/installkernel 42 cannot create /boot/vmlinuz-2.6.35.3: Permission denied

why are you installing kernel for PRS-T1 on your machine...

lurk up.

tux-linux

make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf -j8 install

you need here to change a env variable to install it somewhere else

3:58 PM

or just install it there and then copy

how do you give it a different directory?

4:06 PM

Apparently export INSTALL_PATH=/home/nonasuomy/code/kernelinstall (edited)

yea

Kernel: arch/arm/boot/Image is ready
/bin/bash /home/nonasuomy/code/linux-2.6.35.2/arch/arm/boot/install.sh 2.6.35.3 \
arch/arm/boot/Image System.map "/home/nonasuomy/code/kernelinstall/"

NonaSuomy

Rebuilding the VM will be back in action soon.

Are you compiling this in a VM?

Yes sir

It is not efficient... Try a container

4:52 PM

Distrobox, docker

4:52 PM

There was another one too, orange logo

...

Well for now stick with a VM, but for the future use containers

Or do what actually works for you ;). (VMs are fine, I hate docker).

Why do you hate it?

Because I can't be arsed to learn to use it, so it's just an impediment to actually getting shit done.

6:21 PM

And getting shit done > learning to use crap to maybe get shit done slower

6:21 PM

6:22 PM

(I can see the appeal for some things, these are... not it).

6:22 PM

I don't even use a VM at all to cross-compile, FWIW.

6:22 PM

And, surprise: it works

6:22 PM

(Requires not running a shitty distro, though)

I always assume when people say they hate something is because they don't want to learn something new.

I also hate indirection layers with a fiery passion

6:24 PM

And it's the very definition of a system-level indirection layer ;o)

6:24 PM

If you actually need to automate builds or some such, sure, it's "fun" (not), but practical and useful (edited)

6:24 PM

If you're actually just building stuff for the first time by hand: ahahaha, nope. (edited)

6:25 PM

Cross compilation is already a balancing act, that would feel akin to doing that with your eyes closed and your hands bound behind your back. (edited)

Yeah I didn't understand why they were suggesting that when we have yet to get anything working properly yet.

I agree from an automation stand point it is pretty important though when you are trying to do something large scale. One off nothingness testing... not so much.

6:50 PM

@tux-linux what needs to be done next with the install kernel?

sorry, i really don't feel like it now the goal would be to get a working busybox initrd with a console+shell

are you ok?

yea kinda tired among others

Ok well I hope you feel better.

1

Well, what's working as of right now ?

we attempted that uImage boot it got so far and then you said whoops. Then told me to grab busybox. I did so and seemed like the compile completed.

8:26 PM

I'm just not sure what to do with the output of that now.

do you have an _install folder in the busybox source?

Checking

1

8:29 PM

~/code/kernelinstall$ ls
config-2.6.35.3  System.map-2.6.35.3  vmlinuz-2.6.35.3

8:30 PM

~/code/busybox$ cd _install
bash: cd: _install: No such file or directory

ls -shal ?

Nothing

nothing, as in no _install?

https://stackoverflow.com/questions/52035914/earlyprintk-not-work-as-expected-kernel-hung-during-booting

3.5 KB

ok, make distclean

export INSTALL_PATH=/home/nonasuomy/code/kernelinstall
export PATH="${PATH}:/home/nonasuomy/code/busybox"
make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- imx5_android_defconfig
make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf -j8 install

8:36 PM

this is all I have done.

i'm trying it on my box right now

You may have set that path _install under busybox I set the path to ~/code/kernelinstall

oh

NonaSuomy

~/code/kernelinstall$ ls
config-2.6.35.3  System.map-2.6.35.3  vmlinuz-2.6.35.3

were these things not what you were looking for?

so there is something in ~/code/kernelinstall

8:39 PM

no

8:39 PM

like this

nicolas@gentoo-ryzen ~/D/busybox-1.35.0> ls -shal _install
total 20K
4.0K drwxr-xr-x  5 nicolas nicolas 4.0K Apr 25 20:37 ./
4.0K drwxr-xr-x 36 nicolas nicolas 4.0K Apr 25 20:37 ../
4.0K drwxr-xr-x  2 nicolas nicolas 4.0K Apr 25 20:37 bin/
   0 lrwxrwxrwx  1 nicolas nicolas   11 Apr 25 20:37 linuxrc -> bin/busybox*
4.0K drwxr-xr-x  2 nicolas nicolas 4.0K Apr 25 20:37 sbin/
4.0K drwxr-xr-x  4 nicolas nicolas 4.0K Apr 25 20:37 usr/
nicolas@gentoo-ryzen ~/D/busybox-1.35.0>

8:40 PM

also, I forgot something, you need to enable static build in menuconfig

8:40 PM

Settings -> Static binary

8:40 PM

(make menuconfig)

installing ncurses stuff...

8:43 PM

fml "your display is too small to run menuconfig...

you could also edit .config

the virtualbox linux additions are a PoS

8:44 PM

don't let the desktop load

I don't remember what's its name

NonaSuomy

the virtualbox linux additions are a PoS

ssh instead?

ok

8:46 PM

fixing networking...

1

8:46 PM

Need bridge not Nat

you could just unzoom the terminal otherwise

ok got it but no settings...

CTRL+- and CTRL+SHIFT++

Looks like I need to be in the busybox directory not the linux kernel directory?

yes?

probably why _install doesn't exist as well..

9:00 PM

minor details are missing...

why did you think I wanted you to get busybox source?

I thought adding it to path made it magically use it somehow from the kernel directory

welp, ok

9:02 PM

well no we know why it didn't work... ;p

1

"Build static binary (no shared libs) (new) ?

yes that one

9:02 PM

space to select it

That is all just exit and save?

yes

9:03 PM

then compile and install with make install (edited)

9:04 PM

(with ARCH, CROSS_COMPILE env vars, etc.)

cd ~/code/busybox
export PATH="/usr/local/bin:/usr/bin:/bin:/usr/local/games:/usr/games:/home/nonasuomy/code/kernel/toolchain/gcc-4.8/bin:/home/nonasuomy/code/busybox"
make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- distclean
make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- mrproper
make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- imx5_android_defconfig
make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- -j8 install

9:14 PM

Look ok?

no!

ok

you mixed up the kernel and busybox source

9:16 PM

cd ~/code/busybox
export PATH="/usr/local/bin:/usr/bin:/bin:/usr/local/games:/usr/games:/home/nonasuomy/code/kernel/toolchain/gcc-4.8/bin"
make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- distclean
make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- menuconfig
# Enable static build #
make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- -j8 install

(edited)

tux-linux

cd ~/code/busybox
export PATH="/usr/local/bin:/usr/bin:/bin:/usr/local/games:/usr/games:/home/nonasuomy/code/kernel/toolchain/gcc-4.8/bin"
make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- distclean
make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- menuconfig
# Enable static build #
make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- -j8 install

(edited)

Clear as a bell thank you.

1

9:19 PM

compiling...

9:20 PM

--------------------------------------------------
You will probably need to make your busybox binary
setuid root to ensure all configured applets will
work properly.
--------------------------------------------------

9:20 PM

hopefully that means all good.

not to worry about

9:21 PM

now, go to kernel source

cd ~/code/linux-2.6.35.2

9:21 PM

here.

2.6.35.2? (edited)

9:21 PM

that's interesting

yes

9:22 PM

what is kobo?

2.6.35.3 (edited)

niceeee

*not nice

sadface.

I made a typo

9:22 PM

well, I don't know

9:22 PM

make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- menuconfig

done

hang on

hanging!

9:24 PM

Go in General setup and make that part look like this

9:24 PM

Change intramfs source file(s) path to /home/nonasuomy/code/busybox/_install

9:27 PM

For some reason it says .3 at the top...

9:27 PM

so maybe just a mislabel on sony's part of the folder.

probably (edited)

more to do in here or exit save?

yes

9:29 PM

go into boot options

9:29 PM

are you able to access your sony's recovery console now?

9:29 PM

fine actually

9:29 PM

i've got it with the gist from earlier

1

9:30 PM

console=ttymxc2,115200 rawtable=0xF40000 root=/dev/ram0 rdinit=/sbin/init rootfstype=ramfs set this as the kernel command line (edited)

9:30 PM

i modified the message, take what is there now

9:31 PM

then enable "always use the default kernel command string"

anything there that may understand that we are going to sdcard and not the eMMC? or nothing related to that currently?

oh wait

9:32 PM

I forgot some things in the cmdline

9:32 PM

well, it will just boot the initrd from ram, so nothing to do with the emmc

9:32 PM

nor the sd

9:32 PM

ok the command line should be fine now

9:32 PM

then, exit save

9:34 PM

vi ~/code/busybox/_install/etc/init.d/rcS create etc/init.d if not existing populate rcS with this content:

#!/bin/sh

mkdir -p /sys /proc /dev
mount -t sysfs sysfs /sys
mount -t proc proc /proc
mount -t devtmpfs devtmpfs /dev

/bin/sh

then recompile the kernel (edited)

noinitrd console=ttymxc0,115200 root=/dev/mtdblock2 rw rootfstype=jffs2 ip=off

9:34 PM

this is what used to be there just for ref...

outdated it's not even using the right serial console

I'm removing that noinitrd as well at the start?

absolutely all of it

9:35 PM

and paste the cmdline above

kk

9:36 PM

check always use...

yes

9:38 PM

all good exit save then do the instructions above

on my way...

1

tux-linux

vi ~/code/busybox/_install/etc/init.d/rcS create etc/init.d if not existing populate rcS with this content:

#!/bin/sh

mkdir -p /sys /proc /dev
mount -t sysfs sysfs /sys
mount -t proc proc /proc
mount -t devtmpfs devtmpfs /dev

/bin/sh

then recompile the kernel (edited)

create etc/init.d if not existing ... they didn't

it's normal

Ok so just for clarity and future generations reading this

cd ~/code/linux-2.6.35.2
export PATH="/usr/local/bin:/usr/bin:/bin:/usr/local/games:/usr/games:/home/nonasuomy/code/kernel/toolchain/gcc-4.8/bin"
make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- distclean
make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- mrproper
make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- imx5_android_defconfig
make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- menuconfig

Go in General setup and make that part look like this Change intramfs source file(s) path to /home/nonasuomy/code/busybox/_install Turn off Support initial ramdisks compressed using gzip Turn on Support initial ramdisks compressed using LZMA Exit General setup Go into Boot options Delete everything in the kernel command string noinitrd console=ttymxc0,115200 root=/dev/mtdblock2 rw rootfstype=jffs2 ip=off change it to this console=ttymxc2,115200 rawtable=0xF40000 root=/dev/ram0 rdinit=/sbin/init rootfstype=ramfs Then enable "Always use the default kernel command string" Exit Boot Options Exit Main Menu then save yes. make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- -j8 uImage Flash to working Sony SD Image dd if=uImage of=/dev/mmcblk0 bs=512 seek=2048 (edited)

9:45 PM

?

NOOOOOO

haha

you need to redo the kernel menuconfig

9:48 PM

the imx5_android_defconfig reset .config

ok I never did anything yet.

9:48 PM

so just do make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- -j8 install

9:49 PM

nothing else

not install

9:49 PM

uImage

9:49 PM

cd ~/code/linux-2.6.35.2
export PATH="/usr/local/bin:/usr/bin:/bin:/usr/local/games:/usr/games:/home/nonasuomy/code/kernel/toolchain/gcc-4.8/bin"
make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- -j8 uImage

9:49 PM

distclean and mrproper will clean .config too

compiling...

  LD      arch/arm/boot/compressed/vmlinux
  OBJCOPY arch/arm/boot/zImage
  Kernel: arch/arm/boot/zImage is ready
  UIMAGE  arch/arm/boot/uImage
Image Name:   Linux-2.6.35.3
Created:      Mon Apr 25 21:54:31 2022
Image Type:   ARM Linux Kernel Image (uncompressed)
Data Size:    3303116 Bytes = 3225.70 KiB = 3.15 MiB
Load Address: 70008000
Entry Point:  70008000
  Image arch/arm/boot/uImage is ready

ok

10:03 PM

now try to flash it to the mmc

3.2M Apr 25 21:54 uImage (edited)

uImage

does that size seem ok

yes

should it be bigger now with the inclusion of busybox?

10:05 PM

than the prior one we made

yes

NonaSuomy

Click to see attachment 🖼️

looks like it 2.14MB (edited)

10:08 PM

uImage

3.15 MB

dd it to the sd

tux-linux

Then dd if=uImage of=/dev/mmcblk0 bs=512 seek=2048 the seek value may not be the same if sony put the load sector elsewhere

This

yes

kk

10:14 PM

uart right nothing will be going to the screen yet?

10:18 PM

Nothing just red flashing so far

10:18 PM

flashes red a few times then dead

10:21 PM

U-Boot 2009.08 (Dec 21 2011 - 14:48:38)

CPU:   Freescale i.MX50 family 1.1V at 800 MHz
mx50 pll1: 800MHz
mx50 pll2: 400MHz
mx50 pll3: 216MHz
ipg clock     : 66666666Hz
ipg per clock : 66666666Hz
uart clock    : 24000000Hz
ahb clock     : 133333333Hz
axi_a clock   : 400000000Hz
axi_b clock   : 200000000Hz
weim_clock    : 100000000Hz
ddr clock     : 160000000Hz
esdhc1 clock  : 80000000Hz
esdhc2 clock  : 80000000Hz
esdhc3 clock  : 80000000Hz
esdhc4 clock  : 80000000Hz
Board: MX50 ARM2 board
Boot Reason: [POR]
Boot Device: MMC
I2C:   ready
DRAM:  256 MB
MMC:   FSL_ESDHC: 0, FSL_ESDHC: 1, FSL_ESDHC: 2
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
temperature 35
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
In:    serial
Out:   serial
Err:   serial
Net:   got MAC address from IIM: 00:00:00:00:00:00
FEC0
Hit any key to stop autoboot:  0

MMC read: dev # 0, block # 2048, count 5120 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
SD set clock 50MHz
5120 blocks read: OK

MMC read: dev # 0, block # 10240, count 600 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
SD set clock 50MHz
600 blocks read: OK

(edited)

## Booting kernel from Legacy Image at 70800000 ...
   Image Name:   Linux-2.6.35.3
   Image Type:   ARM Linux Kernel Image (uncompressed)
   Data Size:    3303116 Bytes =  3.2 MB
   Load Address: 70008000
   Entry Point:  70008000
   Verifying Checksum ... Bad Data CRC
ERROR: can't get kernel image!
MX50_ARM2 U-Boot >

Ok it's the load size

10:35 PM

printenv ?

MX50_ARM2 U-Boot > printenv
stdin=serial
stdout=serial
stderr=serial
ethact=FEC0
loadaddr=0x70800000
loadaddr_ramdisk=0x70C00000
rawtable=0xF40000
bootargs=console=ttymxc2,115200 init=/init bootdev=0 rawtable=0xF40000
bootcmd=mmc read 0 ${loadaddr} 0x800 0x1400;mmc read 0 ${loadaddr_ramdisk} 0x2800 0x258;bootm ${loadaddr} ${loadaddr_ramdisk}
bootdev=0

Environment size: 326/131068 bytes

Change bootcmd for mmc read 0 ${loadaddr} 0x800 0x2710; bootm ${loadaddr}

10:44 PM

Escape ";" when setenv

10:44 PM

Basically, setenv bootcmd mmc read...

10:45 PM

saveenv when done, then reset

MX50_ARM2 U-Boot > mmc read 0 ${loadaddr} 0x800 0x2710; bootm ${loadaddr}

MMC read: dev # 0, block # 2048, count 10000 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
SD set clock 50MHz
10000 blocks read: OK
## Booting kernel from Legacy Image at 70800000 ...
   Image Name:   Linux-2.6.35.3
   Image Type:   ARM Linux Kernel Image (uncompressed)
   Data Size:    3303116 Bytes =  3.2 MB
   Load Address: 70008000
   Entry Point:  70008000
   Verifying Checksum ... OK
   Loading Kernel Image ... OK
OK

Starting kernel ...



U-Boot 2009.08 (Dec 21 2011 - 14:48:38)

10:47 PM

then it said bad again

10:47 PM

didnt get a chance to do anything else

Maybe earlyprintk can help Time to sleep here

ok I'll wait for you have a good rest.

1

tux-linux

Maybe earlyprintk can help Time to sleep here

Try to enable it

Enable what?

earlyprintk

earlyprintk=serial,ttymxc2,115200 init=/init bootdev=0 rawtable=0xF40000 root=/dev/ram0 rdinit=/sbin/init rootfstype=ramfs

(edited)

10:53 PM

can you change that from u-boot or have to rebuild?

Rebuild

"earlyprintk=serial,ttyS0,115200 enable earlyprintk on serial port. Pretty much useless on sunxi since some u-boot specific mechanism of printing to u-boot enabled serial port is used anyway and duplicates the earlyprintk functionality. Might come handy if you want to use another port, however."

We're not on sunxi hardware

Szybet

Well for now stick with a VM, but for the future use containers

VMs are efficient with QEMU/KVM or Xen hypervisors VirtualBox a bit less though (edited)

Should we change some size stuff before compile so we don't have that crc error?

No

11:05 PM

Use the same command to start it from U-Boot

It basically just crashes and reboots endlessly

11:08 PM

Someone said this is in the kernel config as well does it need to be enabled as well?

11:08 PM

CONF_PRINTK=y CONF_EARLY_PRINTK=y CONFIG_DEBUG_LL=y (low-level debug) # CONFIG_DEBUG_ICEDCC is not set

earlyprintk not work as expected, kernel hung during booting

I am booting the kernel (arm, cortexa15), and the kernel hung at:

Booting kernel from Legacy Image at 40007fc0 ...

Image Name: Linux-4.4.138-cip25-rt19-yocto-s Image Type: ARM Linux K...

Finally found it

Kernel hacking  ---> 
[*] Kernel debugging
[*] Kernel low-level debugging functions
[*]   Early printk

won't compile now with that enabled

12:14 AM

just drops out after:

kernel/power/fbearlysuspend.c: In function ‘stop_drawing_early_suspend’:
kernel/power/fbearlysuspend.c:40:6: warning: variable ‘ret’ set but not used [-Wunused-but-set-variable]
  int ret;
      ^
  CC      kernel/time/tick-oneshot.o
  CC      kernel/time/tick-sched.o
  LD      kernel/irq/built-in.o
  LD      kernel/power/built-in.o
  LD      kernel/time/built-in.o
  LD      kernel/built-in.o
~/code/linux-2.6.35.2$

(edited)

12:20 AM

If I remove these

[*] Kernel low-level debugging functions
[*]   Early printk

it compiles again...

tux-linux

VMs are efficient with QEMU/KVM or Xen hypervisors VirtualBox a bit less though (edited)

Qemu is pain ;p

...no

It was for me with virt manager

virt-manager is a pain

8:16 AM

qemu-img create -f qcow2 hd.img 40G qemu-system-x86_64 -cdrom linux-livecd.iso -hda hd.img -m 4G -enable-kvm -vga virtio As simple as that

I dont have time in school for figuring out how to add a folder to the VM using cmd

?

A shared folder

8:17 AM

An example

networking is more complicated but it's feasible

Well, time is a problem

NonaSuomy

just drops out after:

kernel/power/fbearlysuspend.c: In function ‘stop_drawing_early_suspend’:
kernel/power/fbearlysuspend.c:40:6: warning: variable ‘ret’ set but not used [-Wunused-but-set-variable]
  int ret;
      ^
  CC      kernel/time/tick-oneshot.o
  CC      kernel/time/tick-sched.o
  LD      kernel/irq/built-in.o
  LD      kernel/power/built-in.o
  LD      kernel/time/built-in.o
  LD      kernel/built-in.o
~/code/linux-2.6.35.2$

(edited)

That's just a warning The error is above somewhere

I know was just showing that it dies there. Seems silly to not dump your actual die error at the end.

(edited)

NonaSuomy

I know was just showing that it dies there. Seems silly to not dump your actual die error at the end.

(edited)

"Ignore errors, it is fine, propably. Propably actually not"

My whole network revolves around qemu kvm virt-manager

8:38 AM

http://nonasuomy.github.io/Arch-Linux-Router-VM-Systemd-networkd/

Arch Linux Infrastructure - Brouter Inception - Part 4a - Arch Linu...

NonaSuomy

I know was just showing that it dies there. Seems silly to not dump your actual die error at the end.

(edited)

I don't recall if this changed in recent kernels, but keep in mind that this is 11 years old code... (edited)

https://www.spinics.net/lists/arm-kernel/msg69936.html

10.89 KB

9:20 AM

Here's the full log with earlyprintk enabled again...

9:24 AM

Ignore the first part didn't have the export set after reconnecting to the session export PATH="/usr/local/bin:/usr/bin:/bin:/usr/local/games:/usr/games:/home/nonasuomy/code/kernel/toolchain/gcc-4.8/bin" (edited)

  AS      arch/arm/kernel/debug.o
In file included from arch/arm/plat-mxc/include/mach/debug-macro.S:52:0,
                 from arch/arm/kernel/debug.S:130:
arch/arm/plat-mxc/include/mach/mx5x.h:23:2: error: #error "Do not include directly."
 #error "Do not include directly."
  ^
make[1]: *** [scripts/Makefile.build:266: arch/arm/kernel/debug.o] Error 1
make[1]: *** Waiting for unfinished jobs....

10:00 AM

[PATCH 07/24] mxc: remove do not include directly — ARM, OMAP, Xsca...

Linux ARM, OMAP, Xscale Kernel: [PATCH 07/24] mxc: remove do not include directly

10:02 AM

//#ifndef __ASM_ARCH_MXC_HARDWARE_H__
//#error "Do not include directly."
//#endif

10:05 AM

  CC      arch/arm/mach-mx5/cpu.o
arch/arm/kernel/debug.S: Assembler messages:
arch/arm/kernel/debug.S:167: Error: garbage following instruction -- `ldrne r3,=MX51_AIPS1_IO_ADDRESS(MX51_UART1_BASE_ADDR)'
arch/arm/kernel/debug.S:183: Error: garbage following instruction -- `ldrne r3,=MX51_AIPS1_IO_ADDRESS(MX51_UART1_BASE_ADDR)'
make[1]: *** [scripts/Makefile.build:266: arch/arm/kernel/debug.o] Error 1
make[1]: *** Waiting for unfinished jobs....
  CC      arch/arm/mach-mx5/mm.o
  CC      arch/arm/mm/init.o
  CC      arch/arm/mm/iomap.o
  CC      arch/arm/mach-mx5/devices.o
  CC      arch/arm/mach-mx5/serial.o
  CC      arch/arm/mach-mx5/dma.o
  CC      arch/arm/mach-mx5/lpmodes.o
make: *** [Makefile:884: arch/arm/kernel] Error 2
make: *** Waiting for unfinished jobs....
  CC      arch/arm/mach-mx5/pm.o

10:09 AM

ASM file

10:10 AM

"Error: garbage following instruction Check your instruction syntax, comments, etc. Check that you have the .syntax unified declaration at the top of your module."

NonaSuomy

https://www.spinics.net/lists/arm-kernel/msg69936.html

yea, there is some playing arround with patches

10:10 AM

i hope you get it done

That worked for that file but alas not sure about this one.

10:15 AM

http://r40eubuntu.blogspot.com/2011/09/ turn on translate and search "garbage"

R40eUbuntu-Sempron26Debian

ubuntu 在 R40e 上還有 Debian 在 Sempron 2600 上

10:15 AM

this is a little above my head...

10:19 AM

arch/arm/kernel/debug.S: Assembler messages:
arch/arm/kernel/debug.S:167: Error: garbage following instruction -- `ldrne r3,=MX51_AIPS1_IO_ADDRESS(MX51_UART1_BASE_ADDR)'
arch/arm/kernel/debug.S:183: Error: garbage following instruction -- `ldrne r3,=MX51_AIPS1_IO_ADDRESS(MX51_UART1_BASE_ADDR)'
make[1]: *** [arch/arm/kernel/debug.o] Error 1
make: *** [arch/arm/kernel] Error 2
make: *** Waiting for unfinished jobs....
 
trace debug.S: 167
        addruart r3, r1
addruart is defined in debug-macro.S:
        .macro    addruart, rx, tmp
        mrc    p15, 0, \rx, c1, c0
        tst    \rx, #1            @ MMU enabled?
        ldreq    \rx, =UART_PADDR    @ physical
        ldrne    \rx, =UART_VADDR    @ virtual
so is the definition of UART_VADDR: (debug-macro.S): MX51_AIPS1_IO_ADDRESS( ) is not expanded (defined). This is defined in mx51.h. Use git blame to find debug-macro.S, and found that it was include mx51.h. Later, after adding mx5x.h, it was changed to include mx51.h.. Change it back and build OK ... But. .After adding earlyprintk to boot argument, after , there is no message...
#include 
#define UART_PADDR    MX51_UART1_BASE_ADDR
#define UART_VADDR    MX51_AIPS1_IO_ADDRESS(MX51_UART1_BASE_ADDR)

10:20 AM

Something may be lost in translation but seems that even after the fix nothing showed up on the console still.

//#ifndef __ASM_ARCH_MXC_HARDWARE_H__
//#error "Do not include directly."
//#endif

// Issue: arch/arm/kernel/debug.S:167/183: Error: garbage following instruction -- `ldrne  issue...
#define UART_PADDR    MX51_UART1_BASE_ADDR
#define UART_VADDR    MX51_AIPS1_IO_ADDRESS(MX51_UART1_BASE_ADDR)

11:10 AM

Still failed

11:11 AM

seems like this is already there in debug-macro.S

#include <mach/mx5x.h>
#define UART_PADDR    MX51_UART1_BASE_ADDR
#define UART_VADDR    MX51_AIPS1_IO_ADDRESS(MX51_UART1_BASE_ADDR)
#endif

11:16 AM

Shakes fist at Ben...

/* arch/arm/mach-imx/include/mach/debug-macro.S
 *
 * Debugging macro include header
 *
 *  Copyright (C) 1994-1999 Russell King
 *  Moved from linux/arch/arm/kernel/debug.S by Ben Dooks

11:21 AM

Does EarlyPrintK work with kobo kernel?

11:27 AM

Assuming it will have the same issue with Ben... haha

Looks like Ben made that change in 2.6.27-rc3

11:54 AM

Guess he just moved it used to be #include <asm/arch/debug-macro.S>

moved this under the #endif which got past it.. debug.S

#else
#include <mach/debug-macro.S>
#endif    /* CONFIG_DEBUG_ICEDCC */
#include <mach/debug-macro.S>
/*
 * Useful debugging routines
 */

(edited)

12:08 PM

Next up...

  CC      arch/arm/kernel/early_printk.o
In file included from arch/arm/kernel/debug.S:132:0:
arch/arm/plat-mxc/include/mach/debug-macro.S:51:2: error: #error "CONFIG_DEBUG_LL is incompatible with multiple archs"
 #error "CONFIG_DEBUG_LL is incompatible with multiple archs"
  ^

12:09 PM

I'm guessing that someplace it's not defining which processor it is for these files to work properly

https://patchwork.ozlabs.org/project/linux-imx/patch/1313729819-30301-1-git-send-email-shawn.guo@linaro.org/#257981

NonaSuomy

Does EarlyPrintK work with kobo kernel?

Never tried it...

do you know where you select the processor type in menuconfig?

Somewhere in general setup?

12:32 PM

The kernel booted before we added the initrd, right?

Well it got to that one point

12:33 PM

about the mem log

Is there fatload in your U-Boot?

is there in kobo?

I'm not sure

12:34 PM

In recent U-Boots, yes

which version is kobo u-boot?

12:34 PM

on the touch a/b (edited)

2009.08

12:35 PM

Just try fatload cmd

ok so same as the sony prst1

Might not be the same config

NonaSuomy

MX50_ARM2 U-Boot > help ? - alias for 'help' autoscr - DEPRECATED - use "source" command instead base - print or set address offset bdinfo - print Board Info structure boot - boot default, i.e., run 'bootcmd' bootd - boot default, i.e., run 'bootcmd' bootm - boot application image from memory bootp - boot image via network using BOOTP/TFTP protocol cls - clear screen cmp - memory compare coninfo - print console devices and information cp - memory copy crc32 - checksum calculation dhcp - boot image via network using DHCP/TFTP protocol echo - echo args to console erase - erase FLASH memory ext2load- load binary file from a Ext2 filesystem ext2ls - list files in a directory (default /) fatinfo - print information about filesystem fatload - load binary file from a dos filesystem fatls - list files in a directory (default /) flinfo - print FLASH memory information go - start application at address 'addr' help - print online help i2c - I2C sub-system iminfo - print header information for application image imxtract- extract a part of a multi-image itest - return true/false on integer compare loadb - load binary file over serial line (kermit mode) loads - load S-Record file over serial line loady - load binary file over serial line (ymodem mode) loop - infinite loop on address range md - memory display mii - MII utility commands mm - memory modify (auto-incrementing address) mmc - MMC sub system mmcinfo - mmcinfo <dev num>-- display MMC info (edited)

So glad I just keep dumping everything here... (edited)

12:36 PM

fatload - load binary file from a dos filesystem

Nice The goal would be to put the kernel in a fat partition and load it from there

Is that how you got it working on the kobo?

No but success rate is quite higher that way

I found this python file under bootfs called undocker.py and these options in bootfs make stage2 make stage3 PRSTUX which produced final_root.tar not sure what to do with it though... thinking there would maybe a step after that, that produced the prstux.img

1:06 AM

https://gitlab.com/phreakuencies/prstux/-/blob/master/rootfs/undocker.py

rootfs/undocker.py · master · Matias N. / prstux · GitLab

PRS-TUX: Linux system for Sony PRS-T* ebook readers

1:06 AM

https://gitlab.com/phreakuencies/prstux/-/blob/master/rootfs/Makefile

rootfs/Makefile · master · Matias N. / prstux · GitLab

PRS-TUX: Linux system for Sony PRS-T* ebook readers

1:07 AM

stage1:
    ./make-base.sh
    
stage2: stage1
    docker build -t prstux .
    
stage3: final_root.tar

final_root.tar: stage2 koreader
    ./make-rootfs.sh

1:07 AM

https://gitlab.com/phreakuencies/prstux/-/blob/master/rootfs/make-rootfs.sh

rootfs/make-rootfs.sh · master · Matias N. / prstux · GitLab

PRS-TUX: Linux system for Sony PRS-T* ebook readers

1:08 AM

the tar it produces has a dot folder in it with all the os directories.

1:08 AM

not sure why they are in a . folder though seems strange...

1:11 AM

Not sure if the kernel is in there some place...

1:12 AM

I think Sony classified the device as "mx50_arm2" which is why the debug code is not engaging properly

3:50 PM

In file included from arch/arm/kernel/debug.S:132:0: arch/arm/plat-mxc/include/mach/debug-macro.S:51:2: error: #error "CONFIG_DEBUG_LL is incompatible with multiple archs" #error "CONFIG_DEBUG_LL is incompatible with multiple archs"

I contacted the Matias Nitsche from PRSTUX but he said he doesn't have any of the files and doesn't remember much about it and that he can't help

1

10:24 AM

Final_root.tar that his build spits out.

treedirectory.txt

968.86 KB

What is this processor considered to the kernel MX51?

NonaSuomy

I contacted the Matias Nitsche from PRSTUX but he said he doesn't have any of the files and doesn't remember much about it and that he can't help

I'm surprised you got a reply

1

NonaSuomy

Final_root.tar that his build spits out.

Seems legit

tux-linux

Seems legit

what else would be required for it to boot do you think?

A working kernel

1

2:44 PM

What do you want, it's like you want to eat a hamburger without the meatball

tux-linux

A working kernel

Any chance when you have a moment can enable earlyprintk and see what would be needed to get it to process?

2:44 PM

for the kobo etc

I suspect it's not loading properly Look into fatload or extload in U-Boot and load the uImage from a partition (edited)

there's like a million and one patches but not sure what any of them are talking about

2:45 PM

https://www.google.com/search?q=plat-mxc+config_debug_ll+is+incompatible+with+multiple+archs&rlz=1C1CHBF_enCA906CA906&oq=plat-mxc+config_debug_ll+is+incompatible+with+multiple+archs&aqs=chrome..69i57.116704j0j7&sourceid=chrome&ie=UTF-8

what is the main problem once more?

Usually I don't have much hope with 12+ years old code (edited)

2:50 PM

This problem happened before on my Kobos, and more often than not it was related to the load size

2:50 PM

(kernel not booting)

I feel like this one would be not too bad to get past I just lack any low level understanding.

2:52 PM

I think that is just telling it where to dump the extra data to. Just that stuff was moved around so it doesn't know which arch it's actually trying to apply it to.

2:54 PM

You will probably be able to solve it right away soon as you enable it and see what it's doing haha.

tux-linux

I suspect it's not loading properly Look into fatload or extload in U-Boot and load the uImage from a partition (edited)

That would be my best bet

I'm not sure why you would take a different route than what you already did with the kobo's feels like then I'm just tossing myself into uncharted waters where I won't get anything functioning.

Szybet

what is the main problem once more?

?

We already had the kernel compiling what went wrong after that?

2:59 PM

Why did we need to compile busybox into the kernel did you do this for the Kobo?

NonaSuomy

Why did we need to compile busybox into the kernel did you do this for the Kobo?

yes, it is like that

Then why does the kobo not need fatload or extload in U-Boot

3:01 PM

They are basically the same hardware just with some sony flair.

NonaSuomy

Then why does the kobo not need fatload or extload in U-Boot

Because kernel loading is handled differently, directly in the C code, not with a command in U-Boot env (edited)

Is that something kobo changed in the kernel source?

What needs to be done to do that?

Maybe... It could be possible to use the Kobo U-Boot But I would really try to load the kernel from a partition Or via serial transfer (loady cmd in U-Boot)

i dont think this is a good idea

Which one?

this is a terrible idea? u-boot has preconfigured usb ports, serial etc. its the same cpu, but there are many usb ports for those cpus. it could not work at all after this

tux-linux

Which one?

u-boot

It's the same soc U-Boot shouldn't touch the epdc

4:44 PM

I would try serial loading before

im talking about other things. maybe low-level things at boot are configured by u-boot, like voltage regulators

tux-linux

I would try serial loading before

to boot u-boot?

4:45 PM

is it possible?

No, to boot the kernel

why not? i dont even know whats the problem for now, i got lost in this thread

How does serial booting work?

https://www.mobileread.com/forums/showpost.php?p=2472765&postcount=15

MobileRead Forums - View Single Post - USB OTG on Sony PRS-T1? scam?

MobileRead - the resource for mobile geeks seeking information and advice for keeping their gadgets happy.

Talks about kernel building and drawing to the screen.

8:42 PM

Guy never makes his own post though and it just dies there

@tux-linux what would happen once the busybox kernel booted? Would it then be a login console prompt of some kind? (edited)

If we get that done, it's probably half the way of the InkBox port, as @Szybet once said

Oh sorry I mean like what would I see on the console if it worked...

A shell prompt

9:59 PM

And kernel messages

9:59 PM

Try loady + minicom

9:59 PM

Have you got it nearby?

yeah I'm just compiling now

1

After loady just type bootm and it should boot

how do you select what its loady'n?

Ctrl A Ctrl S

10:04 PM

More like ctrl a+s

10:04 PM

But a first, then s

10:05 PM

Idk Something like that should open a menu

Am I still writing the too large kernel to the sdcard first?

10:08 PM

then booting from that

No

then trying loading from u-boot?

10:08 PM

loady

You boot from mem

10:08 PM

Yes

So just stock kernel and try to trigger uboot then run loady.

yes

10:15 PM

u-boot is the bootloader so it won't try to run the kernel (edited)

Can't seem to get minicom to take input

Really?

Ah had to set hardware flow control to off

10:30 PM

zmodem ?

Ymodem

hmm not sure what zmodem did when I selected that first but the ttyUSB0 no longer functions

10:38 PM

Tried restarting still shows up under /dev

strange

nm me being dumb

10:39 PM

forgot to put /dev/ in the mincom line

Transfer incomplete

10:47 PM

retry count exceeded

reset the board and try again

10:48 PM

don't touch the wires at all

I think this boot process auto times out and restarts

10:48 PM

some watchdog or something

run i2c probe

10:48 PM

maybe it can help

10:48 PM

for some reason

valid chip addresses: 0048

yeah, i don't need the output

10:52 PM

on kobos it prevents serial console watchdog

10:52 PM

for some reason

it keeps trying to boot and I have to type keys again

there is extload in there right?

10:52 PM

or ext2load

yes

is there an ext2 partition where you could put the uImage?

10:54 PM

or a fat partition, with fatload

10:55 PM

Do it like this

kobo => ext4load mmc 0:1 0x70800000 uImage-root
Loading file "uImage-root" from mmc device 0:1 (xxa1)
9254016 bytes read
kobo =>

(edited)

cat /sys/module/rawdatatable/parameters/rawdata_table
MBR                             :0x00000000:0x00000400
uBoot                           :0x00000400:0x000bfc00
Boot Env                        :0x000c0000:0x00020000
Reserved1                       :0x000e0000:0x00020000
Normal Kernel                   :0x00100000:0x00400000
Normal Rootfs                   :0x00500000:0x00100000
Recovery Kernel                 :0x00600000:0x00400000
Reserved2                       :0x00a00000:0x00500000
Normal Boot Env                 :0x00f00000:0x00020000
Recovery Boot Env               :0x00f20000:0x00020000
Raw Data Table                  :0x00f40000:0x00020000
Info                            :0x00f60000:0x00020000
Id                              :0x00f80000:0x00020000
Reserved3                       :0x00fa0000:0x00060000
Boot Image                      :0x01000000:0x00100000
Waveform                        :0x01100000:0x00200000
LOG                             :0x01300000:0x00500000

10:57 PM

Partitions
Disk /dev/mmcblk2: 1958 MB, 1958739968 bytes
4 heads, 16 sectors/track, 59776 cylinders
Units = cylinders of 64 * 512 = 32768 bytes

        Device Boot      Start         End      Blocks  Id System
/dev/mmcblk2p1             801        1120       10240  83 Linux
/dev/mmcblk2p2            1121        1440       10240  83 Linux
/dev/mmcblk2p3            1441       27008      818176   5 Extended
Partition 3 does not end on cylinder boundary
/dev/mmcblk2p4           27009       59776     1048576   b Win95 FAT32
Partition 4 does not end on cylinder boundary
/dev/mmcblk2p5            1441        2721       40984  83 Linux
/dev/mmcblk2p6            2722        7074      139288  83 Linux
/dev/mmcblk2p7            7075        7395       10264  83 Linux
/dev/mmcblk2p8            7396        9188       57368  83 Linux
/dev/mmcblk2p9            9189       22399      422743  82 Linux swap
/dev/mmcblk2p10          22400       27008      147480  82 Linux swap

10:57 PM

correspond to:

P1: recovery initrd. seems to test for update image on SD card and chroots into actual recovery root otherwise
P2: recovery root filesystem or "diagnostic" system
P3: extended partition
P4: user-data partition ("READER")
P5: font partition
P6: dictionary partition
P7: "setup" partition (contains Sony applications for Windows)
P8: "preload" partition (free ebooks and example content)
P9: Android data partition
P10: Android system partition

NonaSuomy

cat /sys/module/rawdatatable/parameters/rawdata_table
MBR                             :0x00000000:0x00000400
uBoot                           :0x00000400:0x000bfc00
Boot Env                        :0x000c0000:0x00020000
Reserved1                       :0x000e0000:0x00020000
Normal Kernel                   :0x00100000:0x00400000
Normal Rootfs                   :0x00500000:0x00100000
Recovery Kernel                 :0x00600000:0x00400000
Reserved2                       :0x00a00000:0x00500000
Normal Boot Env                 :0x00f00000:0x00020000
Recovery Boot Env               :0x00f20000:0x00020000
Raw Data Table                  :0x00f40000:0x00020000
Info                            :0x00f60000:0x00020000
Id                              :0x00f80000:0x00020000
Reserved3                       :0x00fa0000:0x00060000
Boot Image                      :0x01000000:0x00100000
Waveform                        :0x01100000:0x00200000
LOG                             :0x01300000:0x00500000

that's useful

10:57 PM

awesome

Unmapped offsets

Bootsplash image: 0x01000000, 800x600 = 480000 bytes.

so put it in p4

ok

fatload mmc 2:4 0x70800000 uImage; bootm (edited)

Said wrong image format for bootm

ok, well, good that it loaded at least

tux-linux pinned a message to this channel. 2022-04-28 11:06 PM

bootm 0x708000000?

nop

and if you try with the smaller uImage of a few days ago?

the full message was

** Unable to read "uImage" from mmc 2:4 **
Wrong Image Format for bootm command
ERROR: can't get kernel image!

is there an uImage file in p4?

11:17 PM

what's the command?

yes

11:18 PM

the first one you said

what's mmcinfo?

11:28 PM

Sorry for the delay it’s hard to keep that watchdog happy then it starts booting and it doesn’t seem to like to reset until I unplug the usb cable. Takes a bit to get it going again.

11:30 PM

4GB SDCard

11:31 PM

so it checks out with the capacity there

11:31 PM

Internal is only 2GB

tux-linux

and if you try with the smaller uImage of a few days ago?

did you try that?

11:39 PM

what's fatls mmc 0:2?

Not yet it doesn't seem like it found the file to begin with. Will try

11:40 PM

...fatls

11:46 PM

how do you look in a folder

11:46 PM

I think 2:4 was correct

fatls mmc 2:4 folder/?

MX50_ARM2 U-Boot > fatls mmc 2:4                  
   521090   e2fsck 
            sony_reader/
            books/
            dcim/
   214586   badblocks 

2 file(s), 3 dir(s)

11:48 PM

missed the watchdog think its in that sony_reader/ folder

11:48 PM

will be a min

11:50 PM

didnt like that

11:50 PM

just times out then tries to boot

well at least we can see that the uImage isn't there

11:51 PM

where could it be then?...

It's in that sony_reader folder for sure.

11:52 PM

I would like to verify that through uboot though

then fatload mmc 2:4 0x70800000 sony_reader/uImage (edited)

MX50_ARM2 U-Boot > fatload mmc 2:4 0x70800000 sony_reader/uImage reading sony_reader/uImage

paste that in batch fatload mmc 2:4 0x70800000 sony_reader/uImage; bootm (edited)

watchdog kills it every time

11:59 PM

GitHub - twhitehead/notes-kobohacking

7.92 KB

12:00 AM

(Linux version 2.6.35.3 (hudson@devuntu5)

12:00 AM

should say nonasuomy

yes

12:02 AM

i hate this u-boot

1

NonaSuomy

hmm not sure what zmodem did when I selected that first but the ttyUSB0 no longer functions

It sends the file? You need to wait, there should be a progress bar

NonaSuomy

retry count exceeded

Sound like a problem with the connection, maybe gry to solder :D

Nah it doesn't uboot watchdog kills it after about 7 secs

NonaSuomy

Nah it doesn't uboot watchdog kills it after about 7 secs

Bad connection it is

no it is not.

1

12:37 AM

it's a feature of u-boot

Also i couldn't figure out how to boot from zmodem too, i have no idea

"Method 2:a. Boot into U-boot and enter this command:mw fffffd44 00008000[Enter]You may not be quick enough to enter this command by hand. To be quick enough, you should copy the above command and then when U-boot gives its prompt you just do a paste and press the [Enter] key.This command will prevent the box from rebooting.b. setenv wdt_dis mw.l fffffd44 00008000[Enter]Note: letter l (el) not number 1. c. save[Enter]d. save[Enter]Now, you can do whatever within U-boot and it will not reset. Once you're done with U-boot, you can issue the command: reset[Enter] and the unit will reboot and re-enable the WDT.For future flash updates after having done the above Method 2, you can just go into U-Boot and issue the command and then do your update:run wdt_dis[Enter]Once done, reset the unit and the WDT will automatically re-enable."

12:38 AM

this is for a different processor

Szybet

Also i couldn't figure out how to boot from zmodem too, i have no idea

And i have tryied for some time on the nia

maybe we can do this to disable the WDT

12:38 AM

Just not sure if that memory location is correct for the MXC

Maybe try fastboot boot? (edited)

12:39 AM

Its not like it will work, but maybe

12:41 AM

But i think booting from a partition would be the best for testing, less headache to make it work

We tried that above but still not fast enough

NonaSuomy

it's a feature of u-boot

that's not a joke, U-Boot's really programmed like that

wdt kills everything with in 7 seconds

NonaSuomy

wdt kills everything with in 7 seconds

WTF, why

Szybet

Maybe try fastboot boot? (edited)

2009.08 u-boots I encountered often have broken fastboot/not working at all fastboot

Szybet

WTF, why

because

any chance you would know where the memory location would be on these?

tux-linux

2009.08 u-boots I encountered often have broken fastboot/not working at all fastboot

Maybe Sony wanted fastboot

memory location of what?

I tested setenv wdt_dis mw.l fffffd44 00008000

Szybet

Maybe Sony wanted fastboot

since it's using Android, maybe there are more chances it will work

mw fffffd44 00008000

12:44 AM

nothing happened with setenv so prob wrong memory location

Szybet

Also i couldn't figure out how to boot from zmodem too, i have no idea

on the nia it's more difficult on old devices, I sometimes load the kernel to RAM via loady and ymodem and then boot it

NonaSuomy

nothing happened with setenv so prob wrong memory location

you're trying to disable the watchdog?

yes

maybe you can just recompile u-boot

12:44 AM

or maybe the watchdog is an soc feature

could be

which is why it needs that memory write thing

12:45 AM

sony must do it at some point

12:45 AM

maybe booted in linux

12:45 AM

OH!

12:45 AM

I know why the kernel did not boot

12:47 AM

I just remembered something from a year ago Linux kernel, somehow, on imx507/imx508s boards, disables the watchdog. But it isn't disabled until the kernel is started. So, usually bigger kernels load slower, that's why the kernel we compiled didn't boot because it is always interrupted in the middle of the load and never started. So yeah, we need to find a way to disable this watchdog. Or load a really small kernel which would load a bigger one afterwards (if kexec works there, that is). On Kobos, I added i2c probe at the beginning of the bootcmd to disable the watchdog and didn't have problems after that. (edited)

12:54 AM

Is there an msp430 controller on your board?

1:02 AM

Maybe that’s a 450

1:03 AM

Nope 430

1:03 AM

M430

do that https://github.com/twhitehead/notes-kobohacking#watchdog

Contribute to twhitehead/notes-kobohacking development by creating an account on GitHub.

So i2c probe shows me 00 48 should I use that instead of 43?

1:13 AM

MX50_ARM2 U-Boot > i2c md 0x43 0x0.1 0x2 i2c_addr:chip address cycle fail(a1) i2c_addr failed Error reading the chip. MX50_ARM2 U-Boot > i2c md 0x48 0x0.1 0x2 0000: 1c 1f ..

NonaSuomy

any chance you would know where the memory location would be on these?

CPU datasheet

MX50_ARM2 U-Boot > i2c mm 0x48 0x16.1 00000016: 00 ?

i don't know

NonaSuomy

So i2c probe shows me 00 48 should I use that instead of 43?

try to run i2cdetect on the sony in linux

You want to see if linux sees more devices than u-boot?

11:33 AM

Why don't we take a step back and analyze what prstux did? I don't have a total understanding what I'm looking at but with your help we can probably map it out.

11:33 AM

It's not totally complete but maybe we can fill in the gaps with the kernel we compiled

11:35 AM

PRSTUX
Installation

Flashing
WARNING install PRSTUX at your own risk. If you make a mistake or there's a problem during flashing, you risk bricking your device.
In order to install PRSTUX, you first need to clone/download the repository. Once you have booted into recovery mode (on PRS-T2 you need to hold Home and Menu keys while booting until you see the progress bar moving), plug the device into your computer using the USB cable. You should take note of the mass storage device name which is detected in your computer. You can retrieve this information running dmesg or having udiskctl monitor running before plugging in and noting the output. This guide will assume your device is named /dev/sdb.
Once plugged in, enter prstux directory and run

$ make flash DEVICE=/dev/sdb


replacing your device name as appropriate.
This will build the boot configuration or environment and also a custom boot splash image with the PRSTUX logo and then flash it to the reader. Before flashing, it will take a backup of the sections to be modified, just in case something goes wrong.

11:36 AM

Install PRSTUX into SD card
The next step is to write the PRSTUX image into an SD card. You should have a SD card with at least 4GB. To burn the image you should insert your SD card into a reader and do:

$ ddrescue --force --synchronous prstux.img <sd card device>


Note that you should use the entire device (such as /dev/sdc) and not of any existing partition in it (such as /dev/sdc1). Note that this creates an image with the root partition and a separate data partition. If your SD card is bigger than 4GB you can use gparted to increase the size of the data partition if you like.
After this step, poweroff your reader, insert the SD card and power on again. You should see the PRSTUX logo during boot.

11:36 AM

Usage
There are no specific instructions on how to use PRSTUX. Upon booting Koreader will open. Wifi works as expected. Pressing the power button will make the reader go to suspend mode (disabling devices including WiFi).
You can access the reader via SSH. If connected via USB, it will create a USB-based network. The device is set with IP 192.168.101.1 (so you can set yourself .2, for example). Otherwise, you can access via WiFi.
To add books, you should use calibre wireless connection (see Koreader wiki). At the moment, loading via USB is not possible.
KOreader settings are stored in /mnt/data/koreader. You should load data directly into /mnt/data.

NonaSuomy

PRSTUX
Installation

Flashing
WARNING install PRSTUX at your own risk. If you make a mistake or there's a problem during flashing, you risk bricking your device.
In order to install PRSTUX, you first need to clone/download the repository. Once you have booted into recovery mode (on PRS-T2 you need to hold Home and Menu keys while booting until you see the progress bar moving), plug the device into your computer using the USB cable. You should take note of the mass storage device name which is detected in your computer. You can retrieve this information running dmesg or having udiskctl monitor running before plugging in and noting the output. This guide will assume your device is named /dev/sdb.
Once plugged in, enter prstux directory and run

$ make flash DEVICE=/dev/sdb


replacing your device name as appropriate.
This will build the boot configuration or environment and also a custom boot splash image with the PRSTUX logo and then flash it to the reader. Before flashing, it will take a backup of the sections to be modified, just in case something goes wrong.

For this part I believe is modifying the eMMC to tell it to boot from SD which maybe you can confirm. PRSTUX Makefile ~/code/prstux/Makefile

all: boot splash rootfs

.PHONY: boot splash rootfs docker-build

boot:
    $(MAKE) -C boot

splash:
    $(MAKE) -C splash

rootfs: docker-build
    $(MAKE) -C rootfs

docker-build:
    $(MAKE) -C docker-build 14.04

flash: boot
    $(MAKE) -C boot flash
    $(MAKE) -C splash flash

clean:
    $(MAKE) -C boot clean
    $(MAKE) -C splash clean
    $(MAKE) -C rootfs clean

flash: boot $(MAKE) -C boot flash $(MAKE) -C splash flash ~/code/prstux/boot/Makefile

.PHONY: clean flash

DEVICE ?=

all: bootenv

mkubootenv/mkubootenv:
    $(MAKE) -C mkubootenv

# boot environment creation
bootenv: boot-env.txt mkubootenv/mkubootenv
    mkubootenv/mkubootenv -s 131072 boot-env.txt bootenv

# backup of boot environment
bootenv.bak: 
ifeq ($(DEVICE),)
    @echo Please specify target device with the DEVICE variable
else
    @echo Backing up current bootenv
    ../scripts/extract_normal_bootenv.sh $(DEVICE) > bootenv.bak
endif

# boot environment and splash image flashing
flash: bootenv bootenv.bak
ifeq ($(DEVICE),)
    @echo Please specify target device with the DEVICE variable
else
    @echo Flashing...
    ../scripts/flash_normal_bootenv.sh $(DEVICE)
    @echo Done.
endif

clean:
    rm -f bootenv

11:45 AM

~/code/prstux/boot/boot-env.txt

stdin=serial
stdout=serial
stderr=serial
ethact=FEC0
loadaddr=0x70800000
loadaddr_ramdisk=0x70C00000
bootdev=2
rawtable=0xF40000
bootcmd=mmc read 2 ${loadaddr} 0x800 0x1400;bootm ${loadaddr}
bootargs=root=/dev/mmcblk0p1 rootfstype=ext4 rw rootwait init=/sbin/init bootdev=2 rawtable=0xF40000

11:49 AM

~/code/prstux/scripts/extract_normal_bootenv.sh

#!/bin/bash
set -x
if [ "$1" != "" ]; then
  sudo dd if=$1 bs=512 skip=$((0x00f00000/512)) count=$((0x00020000/512))
else
  echo "usage: $(basename $0) <flash device>"
  exit 1
fi

11:49 AM

~/code/prstux/scripts/extract_bootsplash.sh

#!/bin/bash
if [ "$1" != "" ]; then
  dd if=$1 bs=1 skip=$((0x01000000)) count=$((800*600))
else
  echo "usage: $(basename $0): <device>"
  exit 1
fi

11:51 AM

~/code/prstux/scripts/flash_normal_bootenv.sh

#!/bin/bash
if [ "$1" != "" ]; then
  echo "Will flash into device: $1"
  echo -n "Continue? Type 'yes': "
  read answer
  if [ "$answer" = "yes" ]; then
    set -x
    echo "Zeroing normal env area"
    sudo dd if=/dev/zero of=$1 bs=512 seek=$((0x00f00000/512)) count=$((0x00020000/512))
    echo "Flashing env"
    sudo dd if=bootenv of=$1 bs=512 seek=$((0x00f00000/512)) count=$((0x00020000/512))
    sudo dd if=bootenv of=$1 bs=512 seek=$((0x000c0000/512)) count=$((0x00020000/512))
    echo "Synching"
    sync; sync; sync
  else
    echo "Aborted"
  fi
else
  echo "usage: $(basename $0): <flash device>"
  exit 1
fi

11:53 AM

~/code/prstux/scripts/flash_bootsplash.sh

#!/bin/bash
if [ "$1" != "" ]; then
  echo "Flashing bootsplash"
  dd if=bootsplash.bin of=$1 bs=1 seek=$((0x01000000)) count=$((800*600))
  echo "Synching"
  sync; sync; sync
else
  echo "usage $(basename $0): <device>"
  exit 1
fi

11:54 AM

~/code/prstux/scripts/flash_normal_kernel.sh

#!/bin/bash
echo "Zeroing normal kernel area"
dd if=/dev/zero of=/dev/mmcblk2 bs=512 seek=$((0x00100000/512)) count=$((0x00400000/512))
echo "Flashing kernel"
dd if=uImage of=/dev/mmcblk2 bs=512 seek=$((0x00100000/512))
echo "Synching"
sync; sync; sync

11:54 AM

~/code/prstux/scripts/flash_normal_ramdisk.sh

#!/bin/bash
echo "Zeroing normal rootfs area"
dd if=/dev/zero of=/dev/mmcblk2 bs=512 seek=$((0x00500000/512)) count=$((0x00100000/512))
echo "Flashing kernel"
dd if=root.img of=/dev/mmcblk2 bs=512 seek=$((0x00500000/512))
echo "Synching"
sync; sync; sync

11:58 AM

~/code/prstux/splash/Makefile

.PHONY: flash clean

DEVICE ?=

bootsplash.bin: splash.png
    convert splash.png -size 800x600 -depth 8 gray:bootsplash.bin

bootsplash.bin.bak: 
ifeq ($(DEVICE),)
    @echo Please specify target device with the DEVICE variable
else
        @echo Backing up current bootplash image
        ../../scripts/extract_bootsplash.sh $(DEVICE) > bootsplash.bin.bak
endif

flash: bootsplash.bin bootsplash.bin.bak
ifeq ($(DEVICE),)
    @echo Please specify target device with the DEVICE variable
else
    @echo Flashing...
    ../../scripts/flash_bootsplash.sh $(DEVICE)
    @echo Done.
endif

clean:
    rm -f bootsplash.bin

all: bootsplash.bin

~/code/prstux/rootfs/configs/etc/fstab

# /etc/fstab: static file system information.
#
# file system   mount       type    options           dump    pass
/dev/mmcblk0p2  /mnt/data   ext4    defaults          0       0
/dev/mmcblk2p10 /system     ext4    defaults,ro       0       0 
proc            /proc       proc    defaults          0       0
devpts          /dev/pts    devpts  gid=5,mode=620    0       0
sysfs           /sys        sysfs   defaults          0       0

NonaSuomy

Why don't we take a step back and analyze what prstux did? I don't have a total understanding what I'm looking at but with your help we can probably map it out.

I think prstux didn't even care about fixing the watchdog because the kernel already did and prstux probably has a small kernel, as small as Sony's

propably going with kexec is the easiest and fastest solution?

if it works

whats the kernel version?

2.6.35.3

well, maybe

Do you need busybox built into the kernel for inkbox to load?

its designed this way

1:19 PM

but it could be loaded after booting

Like sony normally just has the busybox tools in a folder to use on the drive I used them to fix some badblocks on the eMMC one time.

1:20 PM

With the stock os.

inkbox kernel/initrd needs busybox

couldn't it be loaded a second after?

what do you mean?

do all the tools get stored in the kernel with it? is there a way to slimline busybox to only use the parts you need to load?

load busybox from a partition

no, all the tools need to be in the initrd/kernel

why?

Szybet

load busybox from a partition

maybe, but it would not be practical and it would be subject to partition corruption

it would be practical, this would solve kernel size issues

NonaSuomy

why?

handle boot switching between diagnostics/main, usb debug mode, print error splashes, verify digital signatures, etc.

Szybet

it would be practical, this would solve kernel size issues

it would also make the system less secure

1:26 PM

if someone can hack into busybox then they can compile the kernel from kobo etc

not necessarily

just... dont

don't do what?

"inkbox cannot support this device becouse of security" sound ridicoulus

Think what you want, I stand on my position.

1:29 PM

The best solution would be to have kexec working Or maybe disable the watchdog

1:30 PM

somehow

How does the watchdog work for Kobo what is the process?

1:59 PM

How large is kobo kernel + busybox?

~8M

for nia its 10.5

So the source is there for uboot as well is there a way to see if it’s enabling a watchdog or where?

uboot-110901.tgz

2:30 PM

https://oss.sony.net/Products/Linux/Audio/PRS-T1JP_20140702.html

https://prodgpl.blob.core.windows.net/download/Audio/common/tUY0WsGdtWcZ3qa4GZ8R4A/i2c-tools-3.0.2.tar.bz2

Source code download

#define WDOG1_BASE_ADDR (AIPS1_BASE_ADDR + 0x00098000)

3:47 PM

MX50_PIN_WDOG = _MXC_BUILD_GPIO_PIN(5, 28, 1, 0x68, 0x314),

WDOG memory map
Absolute
address
(hex)
Register name Width
(in bits) Access Reset value Section/
page
53F9_8000 Watchdog Control Register (WDOG_WCR) 16 R/W 0030h 57.6.1/3418
53F9_8002 Watchdog Service Register (WDOG_WSR) 16 R/W 0000h 57.6.2/3420
53F9_8004 Watchdog Reset Status Register (WDOG_WRSR) 16 R 0000h 57.6.3/3421
53F9_8006 Watchdog Interrupt Control Register (WDOG_WICR) 16 R/W 0004h 57.6.4/3422
53F9_8008 Watchdog Miscellaneous Control Register (WDOG_WMCR) 16 R/W 0001h 57.6.5/3423

Hardware

PCB PRS-T1/G1 M5047 1-884-610-12 BW1118073202953 Cast 2012/05/05A
MCIMX508CVK8B NXP i.MX50 32-bit MPU, ARM Cortex-A8 core, 800MHz, MAPBGA 416.
K4X2G323PC-8GD8 Samsung DRAM Chip Mobile LPDDR SDRAM 256Mbit 64Mx32 90-Pin FBGA
SDIN5D1-2G SanDisk e.MMC 2GB 4.41 BGA
F0513A Renesas MCU 8-bit 78K0 SISC 32KB Flash 5V 44-Pin LQFP (SDCard Controller?)
WM8321G Wolfson Processor Power Management Subsystem (PMIC)
CS42L52-CNZ Cirrus Logic Low-power, Stereo CODEC w/ Headphone & Speaker Amps
3611 WC120 Atheros 6k (6002?)
TPS65180 Texas Instruments PMIC for E Ink® Vizplex™-Enabled Electronic Paper Display
6 inch E-ink LCD (PRS-T1/NOOK Simple/ KOBO N905A/B/C) ED060SCE (LF)C1 RET60C4026(C118) E4K0C6B82M4VS5374AY -2.46
VM15AB LMV65 National Semiconductor  LMV65x 12-MHz, Low Voltage (5V?), Low Power Amplifiers
2 x HC4067B 16-channel analog multiplexer/demultiplexer
LIS1476MHPPC(SY6) 1-853-104-11 Sony 01 W US304362 3.7V 3.6Wh 1500mA Max Charge 4.30V

(edited)

do you have a version of i2cdetect that works on these units?

10:05 PM

Guess it's already included

10:05 PM

Not sure how I would get access though

10:06 PM

This latest firmware has no rooting information about it.

10:06 PM

ADB is running though

NonaSuomy

do you have a version of i2cdetect that works on these units?

busybox i2cdetect

10:24 PM

Transplanting the binary in _install/bin/busybox should work

Tried it in the adb shell

# busybox i2cdetect
i2cdetect: applet not found

10:36 PM

# busybox --help
BusyBox v1.17.4 (2010-12-13 13:27:50 JST) multi-call binary.
Copyright (C) 1998-2009 Erik Andersen, Rob Landley, Denys Vlasenko
and others. Licensed under GPLv2.
See source distribution for full notice.

Usage: busybox [function] [arguments]...
   or: function [arguments]...

        BusyBox is a multi-call binary that combines many common Unix
        utilities into a single executable.  Most people will create a
        link to busybox for each function they wish to use and BusyBox
        will act like whatever it was invoked as.

Currently defined functions:
        [, [[, acpid, addgroup, adduser, adjtimex, arp, arping, ash, awk,
        basename, beep, blkid, bootchartd, brctl, bunzip2, bzcat, bzip2, cal,
        cat, catv, chat, chattr, chgrp, chmod, chown, chpasswd, chpst, chroot,
        chrt, chvt, cksum, clear, cmp, comm, cp, cpio, crond, crontab, cryptpw,
        cttyhack, cut, date, dc, dd, deallocvt, delgroup, deluser, depmod,
        devmem, df, dhcprelay, diff, dirname, dmesg, dnsd, dnsdomainname,
        dos2unix, du, dumpkmap, dumpleases, echo, ed, egrep, eject, env,
        envdir, envuidgid, ether-wake, expand, expr, fakeidentd, false, fbset,
        fbsplash, fdflush, fdformat, fdisk, fgconsole, fgrep, find, findfs,
        flock, fold, free, freeramdisk, fsck, fsck.minix, fsync, ftpd, ftpget,
        ftpput, fuser, getopt, getty, grep, gunzip, gzip, halt, hd, hdparm,
        head, hexdump, hostid, hostname, httpd, hush, hwclock, id, ifconfig,
        ifdown, ifenslave, ifplugd, ifup, inetd, init, insmod, install, ionice,
        ip, ipaddr, ipcalc, ipcrm, ipcs, iplink, iproute, iprule, iptunnel,
        kbd_mode, kill, killall, killall5, klogd, last, length, less, linux32,
        linux64, linuxrc, ln, loadfont, loadkmap, logger, login, logname,
        logread, losetup, lpd, lpq, lpr, ls, lsattr, lsmod, lspci, lsusb,
        lzcat, lzma, lzop, lzopcat, makedevs, makemime, man, md5sum, mdev,

10:36 PM

        mesg, microcom, mkdir, mkdosfs, mke2fs, mkfifo, mkfs.ext2, mkfs.minix,
        mkfs.vfat, mknod, mkpasswd, mkswap, mktemp, modinfo, modprobe, more,
        mount, mountpoint, mt, mv, nameif, nc, netstat, nice, nmeter, nohup,
        nslookup, ntpd, od, openvt, passwd, patch, pgrep, pidof, ping, ping6,
        pipe_progress, pivot_root, pkill, popmaildir, poweroff, printenv,
        printf, ps, pscan, pwd, raidautorun, rdate, rdev, readahead, readlink,
        readprofile, realpath, reboot, reformime, renice, reset, resize, rev,
        rm, rmdir, rmmod, route, rpm, rpm2cpio, rtcwake, run-parts, runlevel,
        runsv, runsvdir, rx, script, scriptreplay, sed, sendmail, seq, setarch,
        setconsole, setfont, setkeycodes, setlogcons, setsid, setuidgid, sh,
        sha1sum, sha256sum, sha512sum, showkey, slattach, sleep, smemcap,
        softlimit, sort, split, start-stop-daemon, stat, strings, stty, su,
        sulogin, sum, sv, svlogd, swapoff, swapon, switch_root, sync, sysctl,
        syslogd, tac, tail, tar, tcpsvd, tee, telnet, telnetd, test, tftp,
        tftpd, time, timeout, top, touch, tr, traceroute, traceroute6, true,
        tty, ttysize, tunctl, udhcpc, udhcpd, udpsvd, umount, uname, unexpand,
        uniq, unix2dos, unlzma, unlzop, unxz, unzip, uptime, usleep, uudecode,
        uuencode, vconfig, vi, vlock, volname, wall, watch, watchdog, wc, wget,
        which, who, whoami, xargs, xz, xzcat, yes, zcat, zcip

10:36 PM

Checks out... busybox-1.17.4.tgz

10:37 PM

seems they have the i2c-tools-3.0.2.tar.bz2 package maybe it's hiding.

10:43 PM

how do you cross compile these apps sources from sony?

10:43 PM

10:50 PM

I guess just like this ~/code/i2c/i2c-tools-3.0.2/tools$ make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- -j8

10:50 PM

Now can I upload this through adb

maybe but just use the busybox you compiled for the initrd

10:56 PM

_install/bin/busybox

10:56 PM

it has i2c tools iirc

ok I'll give that a go as it said permission denied for these things I uploaded over adb

10:58 PM

tried to chmod +x them and it said bad mode

hmm strange

https://stackoverflow.com/questions/9285903/adb-shell-giving-bad-mode-when-executing-chmod-under-su

ADB Shell giving bad mode when executing chmod (under su)

I'm having an issue related to the Android ADB Shell. When I try to execute the command chmod he gives me a "Bad Mode". I don't get it why. I'm sure I'm executing under "su", as represented by the ...

10:59 PM

Apparently it doesn't understand +x need to use numbers

11:04 PM

maybe because this is not an android app (edited)

tux-linux

maybe but just use the busybox you compiled for the initrd

do you upload just the busybox binary or all the folders in there as well?

11:11 PM

guess it's just all symbolic links

11:13 PM

# ./busybox
./busybox: permission denied
# ls
database
i2cdetect
i2cdump
i2cget
i2cset
busybox

Note: These work fine if ran without ./ from rupor’s rescue image. (edited)

just the binary

platform-tools>adb root shell
adbd is already running as root

11:17 PM

I think you have to root the device but no idea how one accomplishes that.

11:19 PM

besides all these sketch tools do you happen to know how to root android 2.2.1 manually?

11:20 PM

What do all these sketchy tools do to root devices

nope, sorry

[root (ttyGS0)]# ./busybox i2cdetect -l
i2c-0   i2c             imx-i2c                                 I2C adapter
i2c-1   i2c             imx-i2c                                 I2C adapter
i2c-2   i2c             imx-i2c                                 I2C adapter

I was expecting to see hex values there...

1:08 AM

[root (ttyGS0)]# i2cdetect -y 0
     0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
00:          -- -- -- -- -- -- -- -- -- -- -- -- --
10: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
20: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
30: -- -- -- -- UU -- -- -- -- -- -- -- -- -- -- --
40: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
50: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
60: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
70: -- -- -- -- -- -- -- --

I think this means this device is on i2c 34 wm831x 0-0034: WM8321 revision C 48 tps65180 1-0048: PMIC TPS6518x for eInk display (edited)

1:19 AM

Touch About compiling the kernel (actually `g_ether' for now) - Mob...

11.1 KB

probably, this controller is there on kobos too

MXC WatchDog Driver 2.0
MXC Watchdog # 0 Timer: initial timeout 60 sec

1:23 AM

was there something else you wanted me to try with the busybox ?

no, it was to help you find the memory addresses for the watchdog

I'm wondering if they actually have different pmic chips and that's why the i2c values don't work

maybe

do you see that WM8321 revision C in the kobo log?

1:26 AM

can I kexec our kernel from here?

1:31 AM

Someone said the touch doesn't have kexec

1:31 AM

https://mobileread.com/forums/showthread.php?t=326168 (edited)

Touch About compiling the kernel (actually `g_ether' for now) Kobo Developer's Corner

the stock kernel doesn't, but you can enable CONFIG_KEXEC in kernel config

Yes but if we can't run a custom kernel yet how would we accomplish that?

i dont know

1:34 AM

i'm facing a similar problem on my kindle pw3

What's it doing?

can't boot the initramfs

1:39 AM

https://www.mobileread.com/forums/showthread.php?t=343310 i'm reading through this...

Tutorial How-to: Building a custom initramfs for fun and profit - M...

Tutorial How-to: Building a custom initramfs for fun and profit Kindle Developer's Corner

1:39 AM

(I'm trying to port InkBox to it)

Please post your antics some place while doing so!

1:45 AM

So I can read for something to do while spinning my tires.

NonaSuomy

Please post your antics some place while doing so!

what do you mean?

Basically like how I spam everything I'm doing here do the same for that device

1

ok, apparently I succeeded

2

Great work!

that was the last thing I expected

1

2:11 AM

join the kpw3 thread if you want

I find it weird that everyone does so much with all these different devices but nothing with sony

2:16 AM

all the posts are just changing official firmware versions and that's basically it.

2:16 AM

https://www.mobileread.com/forums/forumdisplay.php?f=141

Sony Reader Dev Corner - MobileRead Forums

Linux, hacking and development of software and hardware

so... nothing new here?

Do you have any new ideas>

well, if we really can't get this watchdog to shut up...

haha

i was thinking of kexec with a spl kernel

how does that work?

but I have to know if kexec works

7:54 PM

please compile a kernel that works with the stock firmware enable kexec Boot options -> Kexec system call

Well I guess we can just look at the source and see if it's enabled or not

probably not

tux-linux

please compile a kernel that works with the stock firmware enable kexec Boot options -> Kexec system call

When we compiled the kernel to begin with do we have to do something differently to make it "stock" ?

oh- you never had a kernel that worked with the stock system

7:57 PM

right?

Like an android kernel or something

7:57 PM

correct

7:57 PM

It froze up at the log starting part

NonaSuomy

Starting kernel ...

Linux version 2.6.35.3 (nonasuomy@DebianCompiler) (gcc version 4.8.1 20130401 (prerelease) (crosstool-NG linaro-1.13.1-4.8-2013.04-20130417 - Linaro GCC 2013.04) ) #1 PREEMPT Sun Apr 24 22:16:04 EDT 2022
CPU: ARMv7 Processor [412fc085] revision 5 (ARMv7), cr=10c53c7f
CPU: VIPT nonaliasing data cache, VIPT nonaliasing instruction cache
Machine: Freescale MX50 ARM2 Board
Ignoring unrecognised tag 0x54410008
Memory policy: ECC disabled, Data cache writeback
Built 1 zonelists in Zone order, mobility grouping on.  Total pages: 63754
Kernel command line: console=ttymxc2,115200 init=/init bootdev=0 rawtable=0xF40000
PID hash table entries: 1024 (order: 0, 4096 bytes)
Dentry cache hash table entries: 32768 (order: 5, 131072 bytes)
Inode-cache hash table entries: 16384 (order: 4, 65536 bytes)
Memory: 251MB = 251MB total
Memory: 249236k/249236k available, 7788k reserved, 0K highmem
Virtual kernel memory layout:
    vector  : 0xffff0000 - 0xffff1000   (   4 kB)
    fixmap  : 0xfff00000 - 0xfffe0000   ( 896 kB)
    DMA     : 0xf9e00000 - 0xffe00000   (  96 MB)
    vmalloc : 0x90000000 - 0xf4000000   (1600 MB)
    lowmem  : 0x80000000 - 0x8fb00000   ( 251 MB)
    pkmap   : 0x7fe00000 - 0x80000000   (   2 MB)
    modules : 0x7f000000 - 0x7fe00000   (  14 MB)
      .init : 0x80008000 - 0x80025000   ( 116 kB)
      .text : 0x80025000 - 0x80412000   (4020 kB)
      .data : 0x8042e000 - 0x804a86c0   ( 490 kB)
SLUB: Genslabs=9, HWalign=64, Order=0-3, MinObjects=0, CPUs=1, Nodes=1
Hierarchical RCU implementation.
        RCU-based detection of stalled CPUs is disabled.
        Verbose stalled-CPUs detection is disabled.
NR_IRQS:396
MXC GPIO hardware
MXC IRQ initialized
You should not call the gpmi_set_parent
MXC_Early serial console at MMIO 0x63f90000 (options '115200')
bootconsole [ttymxc2] enabled
Console: colour dummy device 80x30
Mix log static memory : 7fe00000 - 7fffffff
              address : 90400000
  There is not the last log.

This is what it did...

Ok, you have your busybox source?

yup

8:03 PM

We tried to compile that in and then made the uImage too large

let me make a very stripped down version

8:04 PM

install upx

ok

8:05 PM

installed

8:07 PM

Make it super bare metal just to see if we can even get a console at all maybe.

yes

I was reading some place that busybox adds padding as well but it may be off by default now

8:08 PM

might have been in earlier editions

8:14 PM

They were saying no matter what they changed in busybox the uImage always stayed the same size and that was the reason why.

config

33.58 KB

8:18 PM

please put this file as .config in your busybox source

8:18 PM

then make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- install

8:20 PM

then upx --ultra-brute _install/bin/busybox

8:20 PM

I got a 524K binary

1

8:20 PM

busybox

547.19 KB

8:23 PM

Very limited

kobo:/tmp# ./busybox 
BusyBox v1.34.1 (2022-04-30 20:19:26 EDT) multi-call binary.
BusyBox is copyrighted by many authors between 1998-2015.
Licensed under GPLv2. See source distribution for detailed
copyright notices.

Usage: busybox [function [arguments]...]
   or: busybox --list[-full]
   or: busybox --install [-s] [DIR]
   or: function [arguments]...

    BusyBox is a multi-call binary that combines many common Unix
    utilities into a single executable.  Most people will create a
    link to busybox for each function they wish to use and BusyBox
    will act like whatever it was invoked as.

Currently defined functions:
    [, [[, ash, awk, base64, cat, chattr, chgrp, chmod, chown, chroot, cp, cttyhack, cut, date, dd, devmem, dmesg, echo, env, false, fbset, fdformat, fold, free, fsck, fuser, getty, grep,
    halt, i2cdetect, i2cdump, i2cget, i2cset, i2ctransfer, ifconfig, ifdown, ifenslave, ifplugd, ifup, init, insmod, kill, killall, linuxrc, ln, login, losetup, ls, lsattr, lsmod, lsof, mkdir,
    mkfifo, mknod, mount, mv, nc, nohup, nproc, pidof, ping, poweroff, printf, ps, pwd, reboot, reset, rm, rmmod, route, run-init, rx, sed, sh, sha256sum, sleep, sort, split, stat, sync, test,
    timeout, touch, tr, true, umount, uname, unlink, vi, watchdog, xxd, yes
kobo:/tmp#

-rwxr-xr-x 1 nonasuomy nonasuomy 894K Apr 30 20:22 busybox

even with upx?

does the config just go in the root dir of busybox?

.config

That was the ticket

or just use my binary otherwise cp ~/Downloads/busybox ~/code/busybox/_install/bin/busybox

it said using .config now when it started.

ok

-rwxr-xr-x 1 nonasuomy nonasuomy 548K Apr 30 20:25 busybox

ok

8:26 PM

now recompile your kernel

8:26 PM

post its size here and try to boot it

Image Name:   Linux-2.6.35.3
Created:      Sat Apr 30 20:33:57 2022
Image Type:   ARM Linux Kernel Image (uncompressed)
Data Size:    2817996 Bytes = 2751.95 KiB = 2.69 MiB
Load Address: 70008000
Entry Point:  70008000
  Image arch/arm/boot/uImage is ready

Let me just verify that the _install directory was in place though.

8:35 PM

Looks good (/home/nonasuomy/code/busybox/_install) Initramfs source file(s) (edited)

Hmm can't seem to get it to see our newly compiled kernel... keeps just showing the old one may need to reflash this sdcard

9:21 PM

https://www.mobileread.com/forums/showpost.php?p=3981129&postcount=1

6.71 KB

9:26 PM

hmm we may have corrupted the sdcard

9:26 PM

died at 38%

Just noticed it said permission denied think it ran out of space...

9:37 PM

looks good now still going

I think it may have lost its direction and is no longer pointing to the SDCard

9:50 PM

Yup that was the culprit.

9:52 PM

Well at the least I found out that SDCard change gets factory reset easy.

Can no longer get it running from SDCard not sure what happened.

Does eMMC work?

Yes

10:13 PM

still boots the broke 07 firmware with no usb/wifi support

10:13 PM

Just to note look at this comment..

Can you try another SD card?

MobileRead Forums - View Single Post - PRS-T1 PRS-T1 run from SD ca...

MobileRead - the resource for mobile geeks seeking information and advice for keeping their gadgets happy.

10:15 PM

Writing to eMMC is easiest, but if your eMMC was totally broken (not just bad sectors), you'd have to modify the uboot image in NAND so that it would load its configuration from mmc0 instead of mmc2. To get recovery environment to boot, put rupor-rescue on the SD card partition 4 and then use this Python script to temporarily point uboot to load the rescue partition from SD card: Code:

import serial, time, sys
s = serial.Serial("/dev/ttyUSB0", 115200, timeout = 0.1)

def slowsend(x):
    for b in x:
        s.write(b)
        time.sleep(0.001)

while True:
    d = s.read(32)
    sys.stdout.write(d)
    s.write(' ' * 32)  # "Any key" for uboot
    if 'Hit' in d:
        slowsend("\r\nprintenv\r\n")
        time.sleep(1.5)

# To boot recovery from SD card
        slowsend("setenv bootargs 'root=/dev/mmcblk0p1 rootfstype=ext4 rw "
                + "rootwait init=/linuxrc console=ttymxc4,115200 bootdev=2 rawtable=0xF40000'\r\n")

# To boot normal system from SD card
#        slowsend("setenv bootargs 'console=ttymxc4,115200 init=/init bootdev=0 rawtable=0xF40000'\r\n")

        time.sleep(1.0)
        slowsend("boot\r\n")
        break

while True:
    d = s.read(64)
    sys.stdout.write(d)

You can also just copy-paste the commands to serial terminal, but that is a bit annoying as the watchdog will reboot the device if it is stuck in uboot for more than 10 seconds or so.

10:17 PM

Can we do the same but point it to our uImage instead

tux-linux

Can you try another SD card?

I don't believe it's even looking at the card slot

10:19 PM

just boots direct to the eMMC every time it starts now. Before it would just die if there was no SDCard in the slot.

10:23 PM

Maybe something got messed up on the eMMC and now it's using the nand instead for booting.

10:27 PM

I don't remember it saying "legacy" in the kernel boot

10:28 PM

## Booting kernel from Legacy Image at 70800000 ...
   Image Name:   Normal Kernel
   Image Type:   ARM Linux Kernel Image (uncompressed)
   Data Size:    2288064 Bytes =  2.2 MB
   Load Address: 70008000
   Entry Point:  70008000
   Verifying Checksum ... OK
## Loading init Ramdisk from Legacy Image at 70c00000 ...
   Image Name:   Normal Rootfs
   Image Type:   ARM Linux RAMDisk Image (uncompressed)
   Data Size:    214270 Bytes = 209.2 kB
   Load Address: 70308000
   Entry Point:  70308000
   Verifying Checksum ... OK
   Loading Kernel Image ... OK

10:29 PM

N/m it did say it...

10:30 PM

https://gist.github.com/nrdmn/c75f8c8011ca82def7f28e4e23d05ed8

Sony PRS-T1

Sony PRS-T1. GitHub Gist: instantly share code, notes, and snippets.

So I flashed over the backup uboot_old then uboot_new again and it's working again.

10:47 PM

so something went screwy there...

10:48 PM

but alas

MMC read: dev # 0, block # 10240, count 600 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
SD set clock 50MHz
600 blocks read: OK
## Booting kernel from Legacy Image at 70800000 ...
   Image Name:   Linux-2.6.35.3
   Image Type:   ARM Linux Kernel Image (uncompressed)
   Data Size:    2817996 Bytes =  2.7 MB
   Load Address: 70008000
   Entry Point:  70008000
   Verifying Checksum ... Bad Data CRC
ERROR: can't get kernel image!

10:48 PM

Size checks out 2.7 MB

10:49 PM

Stock kernel 2.2 MB (edited)

Can you set kernel compression mode to LZMA?

10:49 PM

In general setup Should really improve things

Think we did that

Yes, for the initrd

10:50 PM

But there's also an option for the kernel

Ok how do we do that?

Go in kernel menuconfig

10:50 PM

General setup

10:50 PM

Kernel compression mode or something like that

Ok found it

10:51 PM

Was on GZip

10:51 PM

Now on... Kernel compression mode (LZMA) --->

Yes

Compiling...

1

10:55 PM

Ideally were not going make it to 2.2MB do we need it to be exact?

10:57 PM

Nice it did make a difference

10:57 PM

2.12MB

2.12mb is ok

10:58 PM

try it

Hit any key to stop autoboot:  0

MMC read: dev # 0, block # 2048, count 5120 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
SD set clock 50MHz
5120 blocks read: OK

MMC read: dev # 0, block # 10240, count 600 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
SD set clock 50MHz
600 blocks read: OK
## Booting kernel from Legacy Image at 70800000 ...
   Image Name:   Linux-2.6.35.3
   Image Type:   ARM Linux Kernel Image (uncompressed)
   Data Size:    2226868 Bytes =  2.1 MB
   Load Address: 70008000
   Entry Point:  70008000
   Verifying Checksum ... OK
## Loading init Ramdisk from Legacy Image at 70c00000 ...
   Image Name:   Normal Rootfs
   Image Type:   ARM Linux RAMDisk Image (uncompressed)
   Data Size:    286848 Bytes = 280.1 kB
   Load Address: 70308000
   Entry Point:  70308000
   Verifying Checksum ... OK
   Loading Kernel Image ... OK
OK

Starting kernel ...

Linux version 2.6.35.3 (nonasuomy@Marceline) (gcc version 4.8.1 20130401 (prerelease) (crosstool-NG linaro-1.13.1-4.8-2013.04-20130417 - Linaro GCC 2013.04) ) #4 PREEMPT Sat Apr 30 22:57:06 EDT 2022
CPU: ARMv7 Processor [412fc085] revision 5 (ARMv7), cr=10c53c7f
CPU: VIPT nonaliasing data cache, VIPT nonaliasing instruction cache
Machine: Freescale MX50 ARM2 Board
Ignoring unrecognised tag 0x54410009
Ignoring unrecognised tag 0x54410008
Memory policy: ECC disabled, Data cache writeback
Built 1 zonelists in Zone order, mobility grouping on.  Total pages: 63754
Kernel command line: console=ttymxc2,115200 rawtable=0xF40000 root=/dev/ram0 rdinit=/sbin/init rootfstype=ramfs
PID hash table entries: 1024 (order: 0, 4096 bytes)
Dentry cache hash table entries: 32768 (order: 5, 131072 bytes)
Inode-cache hash table entries: 16384 (order: 4, 65536 bytes)
Memory: 251MB = 251MB total
Memory: 248600k/248600k available, 8424k reserved, 0K highmem

11:08 PM

Virtual kernel memory layout:
    vector  : 0xffff0000 - 0xffff1000   (   4 kB)
    fixmap  : 0xfff00000 - 0xfffe0000   ( 896 kB)
    DMA     : 0xf9e00000 - 0xffe00000   (  96 MB)
    vmalloc : 0x90000000 - 0xf4000000   (1600 MB)
    lowmem  : 0x80000000 - 0x8fb00000   ( 251 MB)
    pkmap   : 0x7fe00000 - 0x80000000   (   2 MB)
    modules : 0x7f000000 - 0x7fe00000   (  14 MB)
      .init : 0x80008000 - 0x800be000   ( 728 kB)
      .text : 0x800be000 - 0x804ad000   (4028 kB)
      .data : 0x804ca000 - 0x80544760   ( 490 kB)
SLUB: Genslabs=9, HWalign=64, Order=0-3, MinObjects=0, CPUs=1, Nodes=1
Hierarchical RCU implementation.
        RCU-based detection of stalled CPUs is disabled.
        Verbose stalled-CPUs detection is disabled.
NR_IRQS:396
MXC GPIO hardware
MXC IRQ initialized
You should not call the gpmi_set_parent
MXC_Early serial console at MMIO 0x63f90000 (options '115200')
bootconsole [ttymxc2] enabled
Console: colour dummy device 80x30
Mix log static memory : 7fe00000 - 7fffffff
              address : 90400000
  There is not the last log.
Initialize a static memory for mix_logger.

11:08 PM

stopped at mix_logger. again.

What is this mess

1

11:12 PM

Go into kernel source top level grep -r "There is not the last log" .

That's only a sony thing as that is how I found that gist above.

11:13 PM

engrish

11:14 PM

./drivers/staging/android/logger.c: printk(KERN_INFO " There is not the last log.\n");

11:18 PM

after that in the stock boot gist above comes

##### LOG START #####
Calibrating delay loop... 799.53 BogoMIPS (lpj=3997696)

11:18 PM

logger.c

39.34 KB

11:19 PM

which you can also find in there.

11:19 PM

not sure what happens between there.

Please count the time spent between when you press the power button and when you get the last message on serial

11:22 PM

How many seconds?

Oh you think its the watchdog killing it

11:23 PM

ok

I hope it's not

11:23 PM

But I can hardly see how a logger could hang the whole system (edited)

I would say about 15 seconds

11:25 PM

used a stopwatch but give or take for not knowing when it's actually starting at.

ok so no watchdog in the field

11:26 PM

good

11:26 PM

but also not good

It very well may be as it just keeps rebooting nonstop at that same point every time.

11:29 PM

I think I read the default watchdog was 16 seconds

oh

11:30 PM

then, in ultimate despair, a ping for @andi because I absolutely don't know what needs to be done to disable this @!@#$# watchdog

1

11:32 PM

andi1 is like a magician for low-level hardware things

11:32 PM

he has ported mainline Linux 5.16 kernel to Clara HD and has ongoing ports for the Libra and Aura 2/SE

Do you have a boot log of the Touch A/B can you see if it has that same WM8 chip for it's pmic (edited)

11:38 PM

(Power Management)

11:44 PM

Wonder what the stock Sony kernel sends to the mysterious watchdog to keep it happy

11:45 PM

Maybe this is missing in our custom version.

11:46 PM

Do you think it matters that the stock kernel was built with GCC 4.4 and ours is 4.8?

No

NonaSuomy

Do you have a boot log of the Touch A/B can you see if it has that same WM8 chip for it's pmic (edited)

I'll try to get one

Touch B kernel ring buffer (edited)

https://www.mobileread.com/forums/showpost.php?p=2472765&postcount=15

13.07 KB

Defo different mc13892 & MSP430

NonaSuomy

Virtual kernel memory layout:
    vector  : 0xffff0000 - 0xffff1000   (   4 kB)
    fixmap  : 0xfff00000 - 0xfffe0000   ( 896 kB)
    DMA     : 0xf9e00000 - 0xffe00000   (  96 MB)
    vmalloc : 0x90000000 - 0xf4000000   (1600 MB)
    lowmem  : 0x80000000 - 0x8fb00000   ( 251 MB)
    pkmap   : 0x7fe00000 - 0x80000000   (   2 MB)
    modules : 0x7f000000 - 0x7fe00000   (  14 MB)
      .init : 0x80008000 - 0x800be000   ( 728 kB)
      .text : 0x800be000 - 0x804ad000   (4028 kB)
      .data : 0x804ca000 - 0x80544760   ( 490 kB)
SLUB: Genslabs=9, HWalign=64, Order=0-3, MinObjects=0, CPUs=1, Nodes=1
Hierarchical RCU implementation.
        RCU-based detection of stalled CPUs is disabled.
        Verbose stalled-CPUs detection is disabled.
NR_IRQS:396
MXC GPIO hardware
MXC IRQ initialized
You should not call the gpmi_set_parent
MXC_Early serial console at MMIO 0x63f90000 (options '115200')
bootconsole [ttymxc2] enabled
Console: colour dummy device 80x30
Mix log static memory : 7fe00000 - 7fffffff
              address : 90400000
  There is not the last log.
Initialize a static memory for mix_logger.

What if you try Sony kernel with Sony system but without the initrd and with LZMA compression? This should make it really small (edited)

I'm game for anything just tell me what to do.

Disable initrd support

8:57 PM

simpler: save .config somewhere, then make with the default sony config, don't remember its name then go into menuconfig and change kernel compression mode to LZMA

I don't think we had any issues with size after the default compile but it still rebooted in the same spot after 15~16 seconds

the default kernel is compressed with gzip

9:00 PM

so i'd try lzma as a last effort

what do you expect to happen changing the compression?

the kernel size will be smaller

9:01 PM

like, 1.x mb

won't it take the same time to boot and then reset?

maybe but maybe not (edited)

9:01 PM

last thing I would try

I'll try it but I think there's something missing like you have to send a flag to the PMIC to stop it from power resetting.

sure, but i don't know what

I was looking all through the code today but found nothing

did you grep -r "watchdog" . or something?

(not really sure what I'm looking for or where it would do that)

9:04 PM

found all these things related to watchdog 3 letter abbrivations in the mx.i documentation and the wm8321

9:05 PM

Just not sure where something would send the signal also noted that you can run the wm8321 over spi so maybe that's why there's no i2c connection for it.

9:05 PM

also in busybox is a watchdog systemlink

9:07 PM

some of the acros are wdt wdi https://www.nxp.com/docs/en/data-sheet/MC13892.pdf

9:08 PM

Interesting block diagram here

so it is some gpio pin

28 I believe in the source.

NonaSuomy

MX50_PIN_WDOG = _MXC_BUILD_GPIO_PIN(5, 28, 1, 0x68, 0x314),

this one

9:11 PM

Though that is on the CPU it could just be power resetting on the WM8321 and not sending anything to the CPU

MobileRead Forums - View Single Post - USB OTG on Sony PRS-T1? scam?

MobileRead - the resource for mobile geeks seeking information and advice for keeping their gadgets happy.

9:12 PM

maybe... some help

maybe the CPU sends the PMIC something to say don't reset the power.

9:13 PM

Someone said to look in one of the linux folders under sys will show if it's enabled in the kernel.

NonaSuomy

https://www.mobileread.com/forums/showpost.php?p=2472765&postcount=15

Do we need these? make modules_install

no

9:18 PM

oh; look in kernel menuconfig, search with / for WATCHDOG

9:18 PM

maybe there is an option related to it

--- Watchdog Timer Support                  
[*]   Disable watchdog shutdown on close                                                                                     
      *** Watchdog Device Drivers ***
< >   Software watchdog                                                                                                      
< >   WM831x watchdog                                                                                                        
<*>   MXC watchdog                                                                                                           
< >   Max63xx watchdog                                                                                                       
< >   IMX2+ Watchdog                                                                                                         
      *** USB-based Watchdog Cards ***
< >   Berkshire Products USB-PC Watchdog

what's "help" of the first option?

interesting

9:29 PM

try disabling that

disable that too

9:30 PM

why in the world were those options enabled

9:30 PM

it must be something I missed (edited)

Probably if the ereader hangs on anything it reboots

9:31 PM

something probably keeps those heartbeats alive.

9:31 PM

something writing to /dev/watchdog

NonaSuomy

Probably if the ereader hangs on anything it reboots

well, nothing of that is used in inkbox, so we don't care

Interesting under drivers they don't enable CONFIG_WM831X_POWER

  x Say Y here to enable support for the power management unit                                                                                                                                                                           x
  x provided by Wolfson Microelectronics WM831x PMICs.

9:36 PM

If you look in the kernel config for kobo are any watchdogs enabled?

no, there were not iirc

Maybe have a quick looksee might surprise yourself.

9:38 PM

"< > WM831x watchdog" so this is not enabled as well.

CONFIG_WATCHDOG=y
CONFIG_WATCHDOG_NOWAYOUT=y
CONFIG_SOFT_WATCHDOG=m
CONFIG_MXC_WATCHDOG=m
# CONFIG_MAX63XX_WATCHDOG is not set
# CONFIG_USBPCWATCHDOG is not set

9:38 PM

CONFIG_WATCHDOG_NOWAYOUT=y quite funny

I'm going to try to disable these first and build the kernel from before.

1

9:41 PM

It maybe just that the kernel is hanging so the watchdog fires off as well..

9:41 PM

So we might just be met with no reset but nothing else.

idk

I always try to put my foot in my mouth seems to be good luck for me haha...

9:42 PM

"It will never do that..." - It does it...

9:42 PM

compiling...

9:45 PM

Could there be anything in the stock kernel config that is pointing to some place that we don't know about and it's not finding it because were telling it to goto busybox console

Same issue

oh well no clue there (edited)

Linux version 2.6.35.3 (nonasuomy@Marceline) (gcc version 4.8.1 20130401 (prerelease) (crosstool-NG linaro-1.13.1-4.8-2013.04-20130417 - Linaro GCC 2013.04) ) #4 PREEMPT Sat Apr 30 22:57:06 EDT 2022 <= Should this change on every compile?

9:55 PM

or does the device itself just think it is that date...

YES?!

9:55 PM

it is the kernel compile time so it didn't flash or something

hmm...

9:57 PM

9:59 PM

hmm... Flash drive has the right date...

9:59 PM

dd said everything was good...

10:00 PM

I also noticed with the image boot I can't trigger UBoot console for some reason

whoops

10:01 PM

that's not good

10:03 PM

you wrote Image ?!

Not sure what would cause that as I can still do it during the rupor's recovery ``` UIMAGE arch/arm/boot/uImage Image Name: Linux-2.6.35.3 Created: Sun May 1 21:44:35 2022 Image Type: ARM Linux Kernel Image (uncompressed) Data Size: 2225840 Bytes = 2173.67 KiB = 2.12 MiB Load Address: 70008000 Entry Point: 70008000 Image arch/arm/boot/uImage is ready

10:03 PM

no I wrote uImage

ok

10:04 PM

rupor's recovery u-boot triggers the watchdog too?

I think so but maybe I never tried

try!

sudo dd if=/media/nonasuomy/UNTITLED/uImage of=/dev/mmcblk0 bs=512 seek=2048
[sudo] password for nonasuomy: 
4347+1 records in
4347+1 records out
2225904 bytes (2.2 MB, 2.1 MiB) copied, 0.273714 s, 8.1 MB/s

sync too

Never did that before but I'll try

it's important

10:14 PM

or eject at least

could be the culprit

10:15 PM

just double checked here too just in case

mkimage -l uImage
Image Name:   Linux-2.6.35.3
Created:      Sun May  1 21:44:35 2022
Image Type:   ARM Linux Kernel Image (uncompressed)
Data Size:    2225840 Bytes = 2173.67 KiB = 2.12 MiB
Load Address: 70008000
Entry Point:  70008000

10:16 PM

says may now

10:17 PM

still dogs out.

10:17 PM

Starting kernel ...

Linux version 2.6.35.3 (nonasuomy@Marceline) (gcc version 4.8.1 20130401 (prerelease) (crosstool-NG linaro-1.13.1-4.8-2013.04-20130417 - Linaro GCC 2013.04) ) #5 PREEMPT Sun May 1 21:44:29 EDT 2022
CPU: ARMv7 Processor [412fc085] revision 5 (ARMv7), cr=10c53c7f
CPU: VIPT nonaliasing data cache, VIPT nonaliasing instruction cache
Machine: Freescale MX50 ARM2 Board
Ignoring unrecognised tag 0x54410009
Ignoring unrecognised tag 0x54410008
Memory policy: ECC disabled, Data cache writeback
Built 1 zonelists in Zone order, mobility grouping on.  Total pages: 63754
Kernel command line: console=ttymxc2,115200 rawtable=0xF40000 root=/dev/ram0 rdinit=/sbin/init rootfstype=ramfs
PID hash table entries: 1024 (order: 0, 4096 bytes)
Dentry cache hash table entries: 32768 (order: 5, 131072 bytes)
Inode-cache hash table entries: 16384 (order: 4, 65536 bytes)
Memory: 251MB = 251MB total
Memory: 248612k/248612k available, 8412k reserved, 0K highmem
Virtual kernel memory layout:
    vector  : 0xffff0000 - 0xffff1000   (   4 kB)
    fixmap  : 0xfff00000 - 0xfffe0000   ( 896 kB)
    DMA     : 0xf9e00000 - 0xffe00000   (  96 MB)
    vmalloc : 0x90000000 - 0xf4000000   (1600 MB)
    lowmem  : 0x80000000 - 0x8fb00000   ( 251 MB)
    pkmap   : 0x7fe00000 - 0x80000000   (   2 MB)
    modules : 0x7f000000 - 0x7fe00000   (  14 MB)
      .init : 0x80008000 - 0x800be000   ( 728 kB)
      .text : 0x800be000 - 0x804ac000   (4024 kB)
      .data : 0x804c8000 - 0x80542660   ( 490 kB)
SLUB: Genslabs=9, HWalign=64, Order=0-3, MinObjects=0, CPUs=1, Nodes=1
Hierarchical RCU implementation.
        RCU-based detection of stalled CPUs is disabled.
        Verbose stalled-CPUs detection is disabled.

10:17 PM

NR_IRQS:396
MXC GPIO hardware
MXC IRQ initialized
You should not call the gpmi_set_parent
MXC_Early serial console at MMIO 0x63f90000 (options '115200')
bootconsole [ttymxc2] enabled
Console: colour dummy device 80x30
Mix log static memory : 7fe00000 - 7fffffff
              address : 90400000
  There is not the last log.
Initialize a static memory for mix_logger.

Well, I'll ping again @andi in absolute despair or this port will likely come to an end unless a miracle happens To summarize, we have two problems: first, there is a watchdog that kicks us out no matter what we try to stop it, second, a custom compiled kernel hangs there without further explanation (see above boot log). I'm suspecting the watchdog but not sure. Hardware platform is imx508, could be great a triple port for that device, Kindle Touch and Kobo Touch B... Just suggestions (edited)

Did you see this comment the guy made: "this tells us that they use uboot for loading two different kernels. i replaced the second image, that starts at block 0x3000 (byte offset 0x600000) with a newly compiled one like this:"

Well, that's about the eMMC Not your microSD card

The sd is running a copy of eMMC

And U-Boot does the right thing, it loads your custom kernel

NonaSuomy

The sd is running a copy of eMMC

Hmm

10:25 PM

How do you switch between eMMC/SD?

Ripped the uboot out of the guys sd image and transplanted overtop of the eMMC uboot.

10:28 PM

dd if=/dev/mmcblk2 skip=786432 of=uboot_old.bin bs=1 count=131072
dd if=sd_card_PRS-T1_1.0.07_adb_enabled.bin skip=786432 of=uboot_new.bin bs=1 count=131072
dd if=uboot_new.bin of=/dev/mmcblk2 seek=786432 bs=1 count=131072

10:29 PM

"After that it should boot (in normal mode) from SD card that has my image on it. Rescue mode should still work as before, as this only changes the "active uboot env" instead of "rescue uboot env"."

10:32 PM

I would assume that means he already did everything needed to uboot.

10:33 PM

"make sure the kernel2backup file really starts with the uboot magic sequence: 27 05 19 56 " what is uboot magic sequence?

10:34 PM

He's copying the uImage to 0x3000

10:34 PM

"dd of=/dev/block/mmcblk2 if=uImage bs=512 seek=$((0x3000))"

10:35 PM

is that the same spot?

10:35 PM

(he's working with the eMMC)

10:36 PM

Wonder why he's telling you to compile UBoot

10:37 PM

hmm zcat /proc/config.gz > .config

10:38 PM

would this have the same stuff in it as that def android we used...

go compare!

where is /proc though..

10:39 PM

is that from the booted os

huh? oh no don't tell me proc doesn't exist in android

10:40 PM

iirc it does it does on my phone that is

I'm just saying he says get the sony kernel source and goto /proc

boot the stock os or recovery mode whatever kernel that works and extract the config then diff

wtf theres no proc in android???

10:41 PM

that would be WEIRD

I don't think that is in the kernel source?

not at all

You guys are jumping to conclusions...

10:41 PM

https://www.mobileread.com/forums/showpost.php?p=2472765&postcount=15

MobileRead Forums - View Single Post - USB OTG on Sony PRS-T1? scam?

MobileRead - the resource for mobile geeks seeking information and advice for keeping their gadgets happy.

10:42 PM

wget http://www.sony.net/Products/Linux/Audio/Download/common/L1KaMCC9mHxsZfoFPJnsBg/linux-2.6.35.3-110901.tgz
wget 

zcat /proc/config.gz > .config
make menuconfig (change what you want)
make uImage
make modules_install

this makes absolutely no sense at all

not sure where that second wget was going.

the command list literally says "get the kernel source, extract the kernel config from your x86_64 computer (totally incompatible) and feed it to the sony kernel compilation" (edited)

10:43 PM

like, ...?

I'm going to go boot the E-ink OS and see if something is laying in /proc/config.gz

1

NonaSuomy

wget http://www.sony.net/Products/Linux/Audio/Download/common/L1KaMCC9mHxsZfoFPJnsBg/linux-2.6.35.3-110901.tgz
wget 

zcat /proc/config.gz > .config
make menuconfig (change what you want)
make uImage
make modules_install

it seems the command list is also getting a different kernel source linux-2.6.35.3 compared to the wrong .2 you had on your source tarball

I think he was drunk while typing.

10:49 PM

110901 => https://prodgpl.blob.core.windows.net/download/Audio/common/wgUBcpExj80eqi5pkrlZfw/uboot-110901.tgz

10:49 PM

https://prodgpl.blob.core.windows.net/download/Audio/common/2UrziX9Zanih0kfZRY4U3Q/linux-2.6.35.2-20140522-2.tgz

10:49 PM

or maybe older source

10:51 PM

Maybe it used to compile but no longer.

Some how lost the ADB usb connection vanished from the USB menu...

11:43 PM

[root (ttyGS0)]# ls
1             193           332           driver        pwr3
1008          197           382           execdomains   sdma
1072          2             393           fb            self
1074          201           397           filesystems   slabinfo
1078          210           4             fs            softirqs
1100          213           5             interrupts    stat
1108          218           76            iomem         swaps
1110          2264          77            ioports       sys
1129          2275          8             irq           sysvipc
1131          2282          9             kallsyms      timer_list
1136          2286          apm           kmsg          tty
1144          2287          asound        kpagecount    uid_stat
1145          2288          buddyinfo     kpageflags    uptime
1150          2303          bus           loadavg       version
12            266           cmdline       locks         vmallocinfo
1229          3             config.gz     meminfo       vmstat
1230          306           cpu           misc          wakelocks
1247          308           cpuinfo       modules       zoneinfo
169           310           crypto        mounts
171           312           devices       net
173           315           diskstats     pagetypeinfo
191           322           dma-engine    partitions

11:43 PM

config.gz

Ok extract that

11:44 PM

Extract cmdline too

52.95 KB

11:52 PM

cat /proc/cmdline
root=/dev/mmcblk2p1 rootfstype=ext4 rw rootwait init=/linuxrc console=ttymxc4,115200 bootdev=2 rawtable=0xF40000

11:53 PM

ttymxc4?

11:53 PM

Maybe we're just dumping to the wrong console?

11:53 PM

n/m it restarts

11:54 PM

so it's not like it had time to do anything.

11:54 PM

Just to note this was from the older generated Rupors Recover build

11:58 PM

I'll see if I can reflash that SD image again and get ADB back to look at the newer build

Weird folder called dropbox/SYSTEM_BOOT@1651370276809.txt

Build: sony/sydney/EBX-5047/imx50_evk:2.2.1/FRG83/105:user/test-keys
Hardware: imx50_evk
Bootloader: unknown
Radio: unknown
Kernel: Linux version 2.6.35.3 (hudson@devuntu5) (gcc version 4.4.0 (GCC) ) #1 PREEMPT Wed May 7 15:03:54 JST 2014

(edited)

I was working too late and didn't realize I was on the eMMC for some reason and wiped it out haha... so may be a bit until I figure out how to recover as nothing works but a loading screen now and recover buttons no longer work. (Can't boot Rupor's Recovery anymore) (edited)

9:23 AM

"bricked until further notice"

9:24 AM

May have to connect the eMMC via that image above to write the partitions back to it.

9:29 AM

NonaSuomy

May have to connect the eMMC via that image above to write the partitions back to it.

i dont think that this will work, but hope dies the last

its really propably that there are somewhere pads to program this IC

Just need to get one of those level shifters as it works at 1.8v

Szybet

its really propably that there are somewhere pads to program this IC

It’s in the picture above

NonaSuomy

Just need to get one of those level shifters as it works at 1.8v

thats the tx50104?

NonaSuomy

It’s in the picture above

was this tested?

Yeah they call it the “iron” method in Russian forums.

i hope you get a logic shifter for 1.8, its not a popular voltage

Well it’s popular for embedded systems…

9:43 AM

Wonder why they are showing two different IC’s 104/102 in the image above. (edited)

9:46 AM

Wonder if txB is ok seems voltage compatible and some make premade breakouts for it. I may even have one…

9:48 AM

The TXB0104 is a 4-bit bidirectional voltage-level translator with automatic direction sensing. This 4-bit noninverting translator uses two separate configurable power-supply rails. The A port is designed to track VCCA. VCCA accepts any supply voltage from 1.2V to 3.6V. The B port is designed to track VCCB. VCCB accepts any supply voltage from 1.65V to 5.5V. This allows for universal low-voltage bidirectional translation between any of the 1.2-V, 1.5-V, 1.8-V, 2.5-V, 3.3-V, and 5-V voltage nodes.

9:52 AM

Looks like it “TXS0104E 4-Bit Bidirectional Voltage-Level Translator for Open-Drain and Push-Pull Applications”

9:56 AM

Think I have this one

9:57 AM

I could interface at full speed with all 4 data lines

10:18 AM

Those pins are so small

its time to learn soldering

Szybet

its time to learn soldering

There’s a “big” difference from through-hole to surface mount SMD 0201 soldering. Learning to solder is only half the battle there. The caps and resistors just stick to the end of your soldering iron at that size and burn up. You really need hot air tweezers and a microscope. (edited)

of course tweezers are needed, and this can be done without air. and without sticking to the solder iron

Make a video of you soldering smd 0201 by hand iron for me

I would need a microscope to record that, I already did ( i dont know if this exact size ) to fpv googles to set the brightness (edited)

11:16 AM

it took hours

11:16 AM

but its not impossible

I’ve done 0805 components by drag soldering but that’s really large in comparison to 0201

You can maybe trigger some sort of download mode by shorting eMMC pins

11:20 AM

Just an idea Might not work

its not a good idea

11:20 AM

it could kill the iC, its best too check in the datasheet

tux-linux

You can maybe trigger some sort of download mode by shorting eMMC pins

Any documentation you have seen preform this with an ereader you can paste a link to?

So this is weird it seems to still be booting the stock kernel...

https://fccid.io/AK8PRST1

7.96 KB

you said you wiped it out

Exactly.

Literally purged each partition off of it before realizing that I selected the wrong device.

1:13 PM

The guy that mentioned how to boot from SD said there was 3 spots to boot from I think... Something about the NAND. ( I thought the eMMC / Nand were both the same thing as the datasheet from the eMMC said iNAND on it)

1:15 PM

or maybe it only allowed me to purge up to the kernel partition...

1:16 PM

They said: " Then it is just a matter of pointing uboot to boot kernel from mmc0 instead of mmc2. There are three ways to do it: Write eMMC from Linux recovery environment Write eMMC from uboot command prompt Write NAND chip from uboot command prompt Writing to eMMC is easiest, but if your eMMC was totally broken (not just bad sectors), you'd have to modify the uboot image in NAND so that it would load its configuration from mmc0 instead of mmc2. "

NonaSuomy

Any documentation you have seen preform this with an ereader you can paste a link to?

The pw3 uses a similar mechanism

NonaSuomy

They said: " Then it is just a matter of pointing uboot to boot kernel from mmc0 instead of mmc2. There are three ways to do it: Write eMMC from Linux recovery environment Write eMMC from uboot command prompt Write NAND chip from uboot command prompt Writing to eMMC is easiest, but if your eMMC was totally broken (not just bad sectors), you'd have to modify the uboot image in NAND so that it would load its configuration from mmc0 instead of mmc2. "

But then to write to nand you need U-Boot Oops

We have access to U-Boot though? as said above it's still booting the uboot -> kernel It just has no files/folders.

1:46 PM

As I think it's getting it from the nand and not eMMC

Oh ok nice

MX50_ARM2 U-Boot > fatls mmc 2:4
** Partition 4 not valid on device 2 **

** Unable to use mmc 2:4 for fatls **

imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
SD set clock 50MHz
Device: FSL_ESDHC
Manufacturer ID: 2
OEM: 544d
Name: SD02G
Tran Speed: 25000000
Rd Block Len: 512
SD version 2.0
High Capacity: No
Capacity: 2013265920
Bus Width: 4-bit
Boot Partition Size: No boot partition available

Niceeee directing uboot to the sdcard boot worked...

2:38 PM

MX50_ARM2 U-Boot > setenv bootargs 'console=ttymxc4,115200 init=/init bootdev=0 rawtable=0xF40000
MX50_ARM2 U-Boot > boot

MMC read: dev # 2, block # 2048, count 5120 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
5120 blocks read: OK

MMC read: dev # 2, block # 10240, count 500 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
500 blocks read: OK
## Booting kernel from Legacy Image at 70800000 ...
   Image Name:   Normal Kernel
   Image Type:   ARM Linux Kernel Image (uncompressed)
   Data Size:    2288064 Bytes =  2.2 MB
   Load Address: 70008000
   Entry Point:  70008000
   Verifying Checksum ... OK
## Loading init Ramdisk from Legacy Image at 70c00000 ...
   Image Name:   Normal Rootfs
   Image Type:   ARM Linux RAMDisk Image (uncompressed)
   Data Size:    214270 Bytes = 209.2 kB
   Load Address: 70308000
   Entry Point:  70308000
   Verifying Checksum ... OK
   Loading Kernel Image ... OK
OK

Starting kernel ...

2:38 PM

Linux version 2.6.35.3 (hudson@devuntu5) (gcc version 4.4.0 (GCC) ) #1 PREEMPT Wed May 7 15:03:54 JST 2014
CPU: ARMv7 Processor [412fc085] revision 5 (ARMv7), cr=10c53c7f
CPU: VIPT nonaliasing data cache, VIPT nonaliasing instruction cache
Machine: Freescale MX50 ARM2 Board
Ignoring unrecognised tag 0x54410008
Memory policy: ECC disabled, Data cache writeback
Built 1 zonelists in Zone order, mobility grouping on.  Total pages: 63754
Kernel command line: console=ttymxc4,115200 init=/init bootdev=0 rawtable=0xF40000
PID hash table entries: 1024 (order: 0, 4096 bytes)
Dentry cache hash table entries: 32768 (order: 5, 131072 bytes)
Inode-cache hash table entries: 16384 (order: 4, 65536 bytes)
Memory: 251MB = 251MB total
Memory: 249220k/249220k available, 7804k reserved, 0K highmem
Virtual kernel memory layout:
    vector  : 0xffff0000 - 0xffff1000   (   4 kB)
    fixmap  : 0xfff00000 - 0xfffe0000   ( 896 kB)
    DMA     : 0xf9e00000 - 0xffe00000   (  96 MB)
    vmalloc : 0x90000000 - 0xf4000000   (1600 MB)
    lowmem  : 0x80000000 - 0x8fb00000   ( 251 MB)
    pkmap   : 0x7fe00000 - 0x80000000   (   2 MB)
    modules : 0x7f000000 - 0x7fe00000   (  14 MB)
      .init : 0x80008000 - 0x80025000   ( 116 kB)
      .text : 0x80025000 - 0x80424000   (4092 kB)
      .data : 0x80444000 - 0x804be700   ( 490 kB)
SLUB: Genslabs=9, HWalign=64, Order=0-3, MinObjects=0, CPUs=1, Nodes=1
Hierarchical RCU implementation.
        RCU-based detection of stalled CPUs is disabled.
        Verbose stalled-CPUs detection is disabled.
NR_IRQS:396
MXC GPIO hardware
MXC IRQ initialized
You should not call the gpmi_set_parent
MXC_Early serial console at MMIO 0x63f90000 (options '115200')
bootconsole [ttymxc4] enabled
Console: colour dummy device 80x30
Mix log static memory : 7fe00000 - 7fffffff
              address : 90400000
  There is not the last log.
Initialize a static memory for mix_logger.

2:39 PM

##### LOG START #####
Calibrating delay loop... 799.53 BogoMIPS (lpj=3997696)
pid_max: default: 32768 minimum: 301
Mount-cache hash table entries: 512
CPU: Testing write buffer coherency: ok
regulator: core version 0.5
NET: Registered protocol family 16
i.MX IRAM pool: 128 KB@0x90040000
CPU is i.MX50 Revision 1.1
IR Power is LDO4
 This is TPS65181.
PWR3DET:power 3 detect
FEC disable
Using SDMA I.API
MXC DMA API initialized
IMX usb wakeup probe
IMX usb wakeup probe
bio: create slab <bio-0> at 0
CSPI: mxc_spi-0 probed
CSPI: mxc_spi-2 probed
usbcore: registered new interface driver usbfs
usbcore: registered new interface driver hub
usbcore: registered new device driver usb
wm831x 0-0034: WM8321 revision C
wm831x 0-0034: Security key had non-zero value 1
regulator: VCORE: 850 <--> 1100 mV at 1000 mV at 0 uA
regulator: VPERI: 950 <--> 1250 mV at 1250 mV at 0 uA
regulator: DCDC3: at 1800 mV
regulator: DCDC4: at 2900 mV
regulator: LDO1: 900 <--> 3300 mV at 1200 mV
regulator: LDO2: 900 <--> 3300 mV at 1200 mV
regulator: LDO3: 900 <--> 3300 mV at 2500 mV
regulator: LDO4: 900 <--> 3300 mV at 3200 mV
regulator: LDO5: 900 <--> 3300 mV at 2900 mV
regulator: LDO6: 900 <--> 3300 mV at 2500 mV
regulator: LDO7: 1000 <--> 3500 mV at 1800 mV
regulator: LDO8: 1000 <--> 3500 mV at 2500 mV
regulator: LDO9: 1000 <--> 3500 mV at 3100 mV
regulator: LDO10: 1000 <--> 3500 mV at 3200 mV
regulator: LDO11: 800 <--> 1550 mV at 1200 mV
Advanced Linux Sound Architecture Driver Version 1.0.23.
Switching to clocksource mxc_timer1
mxsdhci: MXC Secure Digital Host Controller Interface driver
mxsdhci: MXC SDHCI Controller Driver.
SDHC:sdhci_set_power:slot pwr off
mmc_detect_change: id=0
mmc0: SDHCI detect irq 179 irq 1 INTERNAL DMA
mxsdhci: MXC SDHCI Controller Driver.
SDHC:sdhci_set_power:slot pwr on
mmc_detect_change: id=1
mmc1: SDHCI detect irq 0 irq 2 INTERNAL DMA
mxsdhci: MXC SDHCI Controller Driver.
mmc_detect_change: id=2

2:40 PM

mmc2: SDHCI detect irq 228 irq 3 INTERNAL DMA
mxsdhci: MXC SDHCI Controller Driver.
mmc_detect_change: id=3
mmc3: SDHCI detect irq 0 irq 4 INTERNAL DMA
NET: Registered protocol family 2
IP route cache hash table entries: 2048 (order: 1, 8192 bytes)
TCP established hash table entries: 8192 (order: 4, 65536 bytes)
TCP bind hash table entries: 8192 (order: 3, 32768 bytes)
TCP: Hash tables configured (established 8192 bind 8192)
TCP reno registered
UDP hash table entries: 256 (order: 0, 4096 bytes)
UDP-Lite hash table entries: 256 (order: 0, 4096 bytes)
NET: Registered protocol family 1
RPC: Registered udp transport module.
RPC: Registered tcp transport module.
RPC: Registered tcp NFSv4.1 backchannel transport module.
Unpacking initramfs...
Freeing initrd memory: 208K
LPMode driver module loaded
Static Power Management for Freescale i.MX5
PM driver module loaded
sdram autogating driver module loaded
Bus freq driver module loaded
mxc_dvfs_core_probe
DVFS driver module loaded
i.MXC CPU frequency driver
DVFS PER driver module loaded
ashmem: initialized
msgmni has been set to 487
alg: No test for stdrng (krng)
cryptodev: driver loaded.
io scheduler noop registered
io scheduler deadline registered
io scheduler cfq registered (default)
regulator: DISPLAY:
regulator: VCOM: at 1250 mV
regulator: V3P3_CTRL:
regulator: PMIC_TEMP:
regulator: PWR0_CTRL:
regulator: PWR2_CTRL:
regulator: VSYS_EPD:
tps65180 1-0048: PMIC TPS6518x for eInk display
SDHC:sdhci_set_power:slot pwr on
mmc0: new high speed SDHC card at address e624
mmc_rescan:wake_lock_timeout (0)
mmc_rescan:detect done(id=0)
Serial: MXC Internal UART driver
mxcuart register
mxcintuart.2: ttymxc2 at MMIO 0x5000c000 (irq = 33) is a Freescale i.MX
mxcintuart.3: ttymxc3 at MMIO 0x53ff0000 (irq = 13) is a Freescale i.MX
mxcintuart.4: ttymxc4 at MMIO 0x63f90000 (irq = 86) is a Freescale i.MX

2:40 PM

console [ttymxc4] enabled, bootconsole disabled
console [ttymxc4] enabled, bootconsole disabled
mxcswuart register
mxcintswuart.0: ttymxc0 at MMIO 0x53fbc000 (irq = 31) is a Freescale i.MX
SSW:ssw port register 0
mxcintswuart.1: ttymxc1 at MMIO 0x53fc0000 (irq = 32) is a Freescale i.MX
SSW:ssw port register 1
loop: module loaded
FEC Ethernet Driver
PPP generic driver version 2.4.2
usbcore: registered new interface driver asix
usbcore: registered new interface driver cdc_ether
usbcore: registered new interface driver net1080
usbcore: registered new interface driver cdc_subset
mmc2: new high speed MMC card at address 0001
input: gpio-keys as /devices/platform/gpio-keys/input/input0
mmc_rescan:wake_lock_timeout (2)
irtp_init
irtp_connect
mmc_rescan:wifi:wake unlock (3)
input: SONY IR Touchpanel as /devices/virtual/input/input1
IRTP: register early_suspend
input: wm831x_on as /devices/platform/imx-i2c.0/i2c-0/0-0034/wm831x-on/input/input2
input: phxlit_vbus as /devices/platform/phxlit_vbus.0/input/input3
phxlit_vbus phxlit_vbus.0: event(type 5 code 12)
wm831x-rtc wm831x-rtc: rtc core: registered wm831x as rtc0
mxc_rtc mxc_rtc.0: rtc core: registered mxc_rtc as rtc1
using rtc device, wm831x, for alarms
i2c /dev entries driver
APM Battery Driver
MXC WatchDog Driver 2.0
MXC Watchdog # 0 Timer: initial timeout 60 sec
mmcblk0: mmc0:e624 SU04G 3.69 GiB
 mmcblk0: p1 p2 p3 < p5 p6 p7 p8 p9 p10 > p4
Raw data table load
mmcblk2: mmc2:0001 SEM02G 1.82 GiB
BUG: recent printk recursion!
<6> mmcblk2:
logger: created 64K log 'log_main'
logger: created 64K log 'log_events'
logger: created 64K log 'log_radio'
logger: created 64K log 'log_system'
logger: created 2047K log 'log_mix'
logger: created 64K log 'log_kernel'
Past log manager.
CS42L52: CODEC revision C1
DMA Sound Buffers Allocated:UseIram=1 buf->addr=f8006000 buf->area=90046000 size=24576
DMA Sound Buffers Allocated:UseIram=1 buf->addr=7e360000 buf->area=f9e01000 size=24576

2:41 PM

asoc: CS42L52 dai <-> imx-ssi-2-0 mapping ok
ALSA device list:
  #0: imx-3stack (CS42L52 codec)
nf_conntrack version 0.5.0 (3897 buckets, 15588 max)
ip_tables: (C) 2000-2006 Netfilter Core Team
TCP cubic registered
NET: Registered protocol family 17
VFP support v0.3: implementor 41 architecture 3 part 30 variant c rev 2
tps65180_vsys_enable : wait VSYS_EPD.
regulator_init_complete: incomplete constraints, leaving PWR2_CTRL on
regulator_init_complete: incomplete constraints, leaving PMIC_TEMP on
epdc_progress_work_func start progress.
regulator_init_complete: incomplete constraints, leaving LDO11 on
regulator_init_complete: incomplete constraints, leaving LDO7 on
regulator_init_complete: incomplete constraints, leaving LDO6 on
regulator_init_complete: incomplete constraints, leaving LDO5 on
regulator_init_complete: incomplete constraints, leaving LDO3 on
regulator_init_complete: incomplete constraints, leaving LDO2 on
regulator_init_complete: incomplete constraints, leaving LDO1 on
input: sub_cpu_pwrbutton as /devices/platform/mxc_spi.2/spi3.2/sub_cpu_pwrbutton/input/input4
wm831x-rtc wm831x-rtc: setting system clock to 2010-01-01 00:01:59 UTC (1262304119)
Freeing init memory: 116K
init: cannot open '/initlogo.rle'
init: cannot find '/system/bin/switcher', disabling 'switcher'
init: cannot find '/system/bin/servicemanager', disabling 'servicemanager'
init: cannot find '/system/bin/vold', disabling 'vold'
init: cannot find '/system/bin/netd', disabling 'netd'
init: cannot find '/system/bin/dispd', disabling 'dispd'
init: cannot find '/system/bin/debuggerd', disabling 'debuggerd'
init: cannot find '/system/bin/rild', disabling 'ril-daemon'
init: cannot find '/system/bin/app_process', disabling 'zygote'
init: cannot find '/system/bin/mediaserver', disabling 'media'
init: cannot find '/system/bin/dbus-daemon', disabling 'dbus'
init: cannot find '/system/bin/installd', disabling 'installd'

2:41 PM

init: cannot find '/system/etc/install-recovery.sh', disabling 'flash_recovery'
init: cannot find '/system/etc/firewall.sh', disabling 'firewall'
init: cannot find '/system/bin/keystore', disabling 'keystore'
init: cannot find '/system/bin/logwrapper', disabling 'wpa_supplicant'
epdc_progress_work_func end progress.

2:41 PM

Seems to spit out a bit more log doing it that way...

2:45 PM

MXC WatchDog Driver 2.0
MXC Watchdog # 0 Timer: initial timeout 60 sec

2:48 PM

input: wm831x_on as /devices/platform/imx-i2c.0/i2c-0/0-0034/wm831x-on/input/input2

2:52 PM

wm831x-rtc wm831x-rtc: rtc core: registered wm831x as rtc0
mxc_rtc mxc_rtc.0: rtc core: registered mxc_rtc as rtc1
using rtc device, wm831x, for alarms

2:52 PM

uses the wm831 as the Real Time Clock, guess that makes sense keep track of the wdt.

To boot recovery from SD card

setenv bootargs 'root=/dev/mmcblk0p1 rootfstype=ext4 rw rootwait init=/linuxrc console=ttymxc4,115200 bootdev=2 rawtable=0xF40000'
boot

To boot normal system from SD card

setenv bootargs 'console=ttymxc4,115200 init=/init bootdev=0 rawtable=0xF40000'
boot

Are we writing our kernel to one of these spots? (edited)

3:12 PM

Tried both setenv but still got the stock kernel both times.

Setenv is only for kernel cmdline

3:14 PM

Don't have time right now, will get back to you tonight

ok no worries (I dd/sync'd our uImage over the stock SDCard Image so I assumed the stock kernel should no longer exist)

dd if=uImage of=/dev/mmcblk0 bs=512 seek=2048
sync

(edited)

Can you still boot our kernel?

NonaSuomy

Click to see attachment 🖼️

ttymxc2 is the problem I think

7:25 PM

It should be ttymxc4

7:25 PM

Try this Last hope

tux-linux

Can you still boot our kernel?

I'm not sure how to yet

9:41 PM

You said that line is only for after the kernel loads

9:42 PM

so to me that means it's always going to be loading the stock kernel

Yes, but it controls the boot console

9:42 PM

Yes

9:42 PM

Unless you find how to load something else

But I really think ttymxc4 is the key

But the "lock" is the watchdog woof.

I think the logger does something related to the console, it auto-switches it somehow, so that's why we can't see anything else on the log idk (edited)

Hopefully it unmasks something. First struggle will be to get the kernel booting again no idea yet how to accomplish that.

there is a way in u-boot to read from the external mmc if it is supported

Was hopping that I could boot rupor's recovery with that above but it never engaged the usb serial terminal like it normally did or I'm just not sure how it works in general.

tux-linux pinned a message to this channel. 2022-05-02 10:12 PM

the SD card of rupors just has a folder called OS Firmware\update.img on it

please do mmcinfo 0 in U-Boot

so I'm sure it doesn't know how to launch that.

10:14 PM

As the SD image that the other guy did has all the linux and partitions expanded

10:14 PM

so maybe we can use loady to launch that update.img

how much does it weigh?

Then get usb serial console back then flash all the partitions back (edited)

NonaSuomy

so maybe we can use loady to launch that update.img

unless update.img is a kernel, nothing will 'launch'

NonaSuomy

Click to see attachment 🖼️

Have a look here

loading 25mb with loady will take half an hour

10:17 PM

and it is not a kernel maybe there's some hope inside the file

The normal process is toss that folder on the SDCard then hold HOME + MENU physical buttons during boot then it launches that.

10:18 PM

(the Home + Menu process I'm guessing requires some of those partitions I purged

10:19 PM

as it no longer launches anything now.

tux-linux

please do mmcinfo 0 in U-Boot

can you do that, or U-Boot console is not accessible at all?

I still have U-Boot console from UART loading from what I believe is NAND

tux-linux

can you do that, or U-Boot console is not accessible at all?

Yes trying in the background

1

10:24 PM

MX50_ARM2 U-Boot > mmcinfo 0
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
SD set clock 50MHz
Device: FSL_ESDHC
Manufacturer ID: 2
OEM: 544d
Name: SD02G
Tran Speed: 25000000
Rd Block Len: 512
SD version 2.0
High Capacity: No
Capacity: 2013265920
Bus Width: 4-bit
Boot Partition Size: No boot partition available

ok, now, mmcinfo 1 if it doesn't work, mmcinfo 2

NonaSuomy

MX50_ARM2 U-Boot > mmcinfo 0
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
SD set clock 50MHz
Device: FSL_ESDHC
Manufacturer ID: 2
OEM: 544d
Name: SD02G
Tran Speed: 25000000
Rd Block Len: 512
SD version 2.0
High Capacity: No
Capacity: 2013265920
Bus Width: 4-bit
Boot Partition Size: No boot partition available

this seems to be the eMMC

MX50_ARM2 U-Boot > mmcinfo 1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
Card did not respond to voltage select!
MMC card init failed!
MX50_ARM2 U-Boot > mmcinfo 2
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
Device: FSL_ESDHC
Manufacturer ID: 45
OEM: 100
Name: SEM02
Tran Speed: 25000000
Rd Block Len: 512
MMC version 4.0
High Capacity: No
Capacity: 1958739968
Bus Width: 4-bit
Boot Partition Size:  1024KB
Current Partition for boot: Not bootable
MX50_ARM2 U-Boot > mmcinfo 3
MMC Device 3 not found

(edited)

ok nice is your external sd 2GB in capacity?

I have 3 that I was testing but yes the one in there now is 2GB

10:28 PM

2GB had rupors on it

good, so you can read in external mmc from internal u-boot (edited)

10:29 PM

that means you can load a kernel

4GB has that stock adb image with our flash over it and the 16gb just has the stock adb image on it.

let me try something on one of my kobo's u-boot

I read someplace to satisfy a /dev/watchdog you have to keep writing "v" to it. (this may not be something running yet though before it resets)

busybox watchdog does that iirc

If only we could get to busybox haha

10:37 PM

console=ttymxc4,115200 rawtable=0xF40000 root=/dev/ram0 rdinit=/sbin/init rootfstype=ramfs You want the default kernel command string to be changed to 2 -> 4 correct?

yes

I wish I was able to save /proc/config.gz from the latest build before I did that stupid mistake as that was what I was trying to clean the SDCard for

10:40 PM

As the next attempt was going to just be get the default compiled kernel booting the stock os. (edited)

10:42 PM

I wish people just brain dumped everything they were doing into git and then cleaned up the sanity later if they wanted it to help someone so at the least you could maybe put the puzzle pieces together.

10:43 PM

Maybe Neural link will do this.

10:44 PM

hmm lets compare our .config to rupors config.gz I dupped.

Does mmc read 1 0x800 0x1000 work or does it hang?

One moment

1

Things in our config that are different

Rupor
N/A
Ours
CONFIG_ARM_L1_CACHE_SHIFT_6=y

Rupor
CONFIG_KERNEL_GZIP=y
# CONFIG_KERNEL_LZMA is not set
Ours
# CONFIG_KERNEL_GZIP is not set
CONFIG_KERNEL_LZMA=y

Rupor
CONFIG_INITRAMFS_SOURCE=""                                         
CONFIG_RD_GZIP=y
# CONFIG_RD_BZIP2 is not set
# CONFIG_RD_LZMA is not set
# CONFIG_RD_LZO is not set
Ours
CONFIG_INITRAMFS_SOURCE="/home/nonasuomy/code/busybox/_install"
CONFIG_INITRAMFS_ROOT_UID=0
CONFIG_INITRAMFS_ROOT_GID=0
# CONFIG_RD_GZIP is not set
# CONFIG_RD_BZIP2 is not set
CONFIG_RD_LZMA=y
# CONFIG_RD_LZO is not set
CONFIG_INITRAMFS_COMPRESSION_NONE=y
# CONFIG_INITRAMFS_COMPRESSION_GZIP is not set
# CONFIG_INITRAMFS_COMPRESSION_BZIP2 is not set
# CONFIG_INITRAMFS_COMPRESSION_LZMA is not set
# CONFIG_INITRAMFS_COMPRESSION_LZO is not set

Rupor
CONFIG_ARM_L1_CACHE_SHIFT=5
Ours
CONFIG_ARM_L1_CACHE_SHIFT=6

Rupor
CONFIG_CMDLINE="noinitrd console=ttymxc0,115200 root=/dev/mtdblock2 rw rootfstype=jffs2 ip=off"
# CONFIG_CMDLINE_FORCE is not set
# CONFIG_XIP_KERNEL is not set
# CONFIG_KEXEC is not set
Ours
CONFIG_CMDLINE="console=ttymxc4,115200 rawtable=0xF40000 root=/dev/ram0 rdinit=/sbin/init rootfstype=ramfs"
CONFIG_CMDLINE_FORCE=y
# CONFIG_XIP_KERNEL is not set
CONFIG_KEXEC=y
CONFIG_ATAGS_PROC=y

(edited)

god, you did use diff, didn't you?

1

11:11 PM

thanks for the summary

Rupor
# CONFIG_NETWORK_SECMARK is not setONFIG_NETFILTER_XT_TARGET_CLASSIFY is not set
Ours
# CONFIG_NETWORK_SECMARK is not set
CONFIG_NETFILTER=y
# CONFIG_NETFILTER_DEBUG is not set
CONFIG_NETFILTER_ADVANCED=y

#
# Core Netfilter Configuration
#
# CONFIG_NETFILTER_NETLINK_QUEUE is not set
# CONFIG_NETFILTER_NETLINK_LOG is not set
CONFIG_NF_CONNTRACK=y
# CONFIG_NF_CT_ACCT is not set
# CONFIG_NF_CONNTRACK_MARK is not set
# CONFIG_NF_CONNTRACK_EVENTS is not set
# CONFIG_NF_CT_PROTO_DCCP is not set
# CONFIG_NF_CT_PROTO_SCTP is not set
# CONFIG_NF_CT_PROTO_UDPLITE is not set
# CONFIG_NF_CONNTRACK_AMANDA is not set
CONFIG_NF_CONNTRACK_FTP=y
# CONFIG_NF_CONNTRACK_H323 is not set
# CONFIG_NF_CONNTRACK_IRC is not set
# CONFIG_NF_CONNTRACK_NETBIOS_NS is not set
# CONFIG_NF_CONNTRACK_PPTP is not set
# CONFIG_NF_CONNTRACK_SANE is not set
# CONFIG_NF_CONNTRACK_SIP is not set
# CONFIG_NF_CONNTRACK_TFTP is not set
# CONFIG_NF_CT_NETLINK is not set
CONFIG_NETFILTER_XTABLES=y

#
# Xtables combined modules
#
# CONFIG_NETFILTER_XT_MARK is not set
# CONFIG_NETFILTER_XT_CONNMARK is not set

#
# Xtables targets
#
# CONFIG_NETFILTER_XT_TARGET_CLASSIFY is not set

Rupor
CONFIG_TOUCHSCREEN_SONYIR_NOFILM=y
Ours
# CONFIG_TOUCHSCREEN_SONYIR_NOFILM is not set

Rupor
CONFIG_WATCHDOG=y
CONFIG_WATCHDOG_NOWAYOUT=y

#
# Watchdog Device Drivers
#
# CONFIG_SOFT_WATCHDOG is not set
# CONFIG_WM831X_WATCHDOG is not set
CONFIG_MXC_WATCHDOG=y
# CONFIG_MAX63XX_WATCHDOG is not set
# CONFIG_IMX2_WDT is not set

#
# USB-based Watchdog Cards
#
# CONFIG_USBPCWATCHDOG is not set
ours
# CONFIG_WATCHDOG is not set

11:11 PM

Rupor
# USB Device Class dri
                      # CONFIG_USB_SERIAL_ARK3116 is not set
Ours
# USB Device Class drivers
#
# CONFIG_USB_ACM is not set
# CONFIG_USB_PRINTER is not set
# CONFIG_USB_WDM is not set
# CONFIG_USB_TMC is not set

#
# NOTE: USB_STORAGE depends on SCSI but BLK_DEV_SD may
#

#
# also be needed; see USB_STORAGE Help for more info
#
# CONFIG_USB_LIBUSUAL is not set

#
# USB Imaging devices
#
# CONFIG_USB_MDC800 is not set

#
# USB port drivers
#
CONFIG_USB_SERIAL=m
# CONFIG_USB_EZUSB is not set
# CONFIG_USB_SERIAL_GENERIC is not set
# CONFIG_USB_SERIAL_AIRCABLE is not set
# CONFIG_USB_SERIAL_ARK3116 is not set

Rupor
CONFIG_USB_G_SERIAL=m
Ours
# CONFIG_USB_G_SERIAL is not set

Rupor
# CONFIG_USB_ANDROID is not set
Ours
CONFIG_USB_ANDROID=m
CONFIG_USB_ANDROID_ADB=y
CONFIG_USB_ANDROID_MASS_STORAGE=y
CONFIG_USB_ANDROID_MASS_STORAGE_EXTSCSI=y
# CONFIG_USB_ANDROID_RNDIS is not set

Rupor
#
# MXC support drivers
#
CONF_SECURITY_RNG is not set
Ours
CONFIG_MXC_IPU=y
CONFIG_MXC_IPU_V3=y

#
# MXC SSI support
#
# CONFIG_MXC_SSI is not set

#
# MXC Digital Audio Multiplexer support
#
# CONFIG_MXC_DAM is not set

#
# MXC PMIC support
#
# CONFIG_MXC_PMIC_MC13783 is not set
# CONFIG_MXC_PMIC_MC13892 is not set
# CONFIG_MXC_PMIC_MC34704 is not set
# CONFIG_MXC_PMIC_MC9SDZ60 is not set
# CONFIG_MXC_PMIC_MC9S08DZ60 is not set

#
# MXC Security Drivers
#
# CONFIG_MXC_SECURITY_SCC is not set
# CONFIG_MXC_SECURITY_SCC2 is not set
# CONFIG_MXC_SECURITY_RNG is not set

#

Rupor
CONFIG_DECOMPRESS_GZIP=y
Ours
CONFIG_DECOMPRESS_LZMA=y

tux-linux

Does mmc read 1 0x800 0x1000 work or does it hang?

Do we want on the SDCard for the MMC command?

11:15 PM

Also do you mean 2?

11:16 PM

1 reported errors in our test above.

yes, it doesn't matter right now, i just wanted to test if mmc read wasn't unusable

11:16 PM

yeah, 2 then

MX50_ARM2 U-Boot > mmc read 2 0x800 0x1000

MMC read: dev # 2, block # 4096, count 0 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz

Do you get the prompt back after that?

Nope watchdog

Actually it should be this mmc read 2 0x70008000 0x800 0x1000 (edited)

12:19 AM

Wait

12:19 AM

Ok edited

MX50_ARM2 U-Boot > mmc read 2 0x70008000 0x800 0x1000

MMC read: dev # 2, block # 2048, count 4096 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
4096 blocks read: OK
MX50_ARM2 U-Boot >

Yay

I have a uImage sitting on that sdcard root just to note.

Try

mmc read 2 0x70008000 0x800 0x2000
bootm 0x70008000

NonaSuomy

I have a uImage sitting on that sdcard root just to note.

Is it written to unpartitioned space?

no just raw sitting there.

12:24 AM

MX50_ARM2 U-Boot > mmc read 2 0x70008000 0x800 0x2000

MMC read: dev # 2, block # 2048, count 8192 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz

That is unpartitioned space

NonaSuomy

MX50_ARM2 U-Boot > mmc read 2 0x70008000 0x800 0x2000

MMC read: dev # 2, block # 2048, count 8192 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz

Watchdog?

tux-linux

Watchdog?

yes

Crap

12:25 AM

Ok replace 0x2000 with 0x1400 (edited)

12:25 AM

Copy paste as soon as possible

MX50_ARM2 U-Boot > mmc read 2 0x70008000 0x800 0x1400

MMC read: dev # 2, block # 2048, count 5120 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
5120 blocks read: OK
MX50_ARM2 U-Boot >

Or setenv bootcmd that command

12:26 AM

Good but you forgot bootm

tux-linux

Try

mmc read 2 0x70008000 0x800 0x2000
bootm 0x70008000

.

MX50_ARM2 U-Boot > mmc read 2 0x70008000 0x800 0x1400 bootm 0x70008000
mmc - MMC sub system

Usage:
mmc mmc read <device num> addr blk# cnt [partition]
mmc write <device num> addr blk# cnt [partition]
mmc rescan <device num>
mmc list - lists available devices
MX50_ARM2 U-Boot >

no

12:29 AM

bootm is the second command

MX50_ARM2 U-Boot > mmc read 2 0x70008000 0x800 0x1400 MMC read: dev # 2, block # 2048, count 5120 partition # 0 ... imx_esdhc.c bus_width 0x1 imx_esdhc.c bus_width 0x1 imx_esdhc.c bus_width 0x4 imx_esdhc.c bus_width 0x4 MMC set clock 50MHz 5120 blocks read: OK MX50_ARM2 U-Boot > bootm 0x70008000

Booting kernel from Legacy Image at 70008000 ...

Image Name: Normal Kernel Image Type: ARM Linux Kernel Image (uncompressed) Data Size: 2288064 Bytes = 2.2 MB Load Address: 70008000 Entry Point: 70008000 Verifying Checksum ... OK XIP Kernel Image ... OK OK Starting kernel ... Watchdog...

Well we got to some point

12:32 AM

now

setenv bootcmd mmc read 2 0x70008000 0x800 0x1400\; bootm 0x70008000
saveenv
reset

(edited)

reset or boot?

setenv sets environment variables

12:32 AM

Bootm isn't executed

I'm just saying when I was launching before I just put boot after the setenv line

12:34 AM

(not bootm)

You need bootm

12:34 AM

then nothing worked

I thought reset would be like the watchdog triggering

tux-linux

now

setenv bootcmd mmc read 2 0x70008000 0x800 0x1400\; bootm 0x70008000
saveenv
reset

(edited)

Try this

12:35 AM

Caution in newlines

MX50_ARM2 U-Boot > setenv bootcmd mmc read 2 0x70008000 0x800 0x1400\; bootm
MX50_ARM2 U-Boot > saveenv
Saving Environment to MMC...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
Writing to MMC(2)... done
MX50_ARM2 U-Boot > reset
resetting ...
▒

U-Boot 2009.08 (Dec 21 2011 - 14:48:38)

CPU:   Freescale i.MX50 family 1.1V at 800 MHz
mx50 pll1: 800MHz
mx50 pll2: 400MHz
mx50 pll3: 216MHz
ipg clock     : 66666666Hz
ipg per clock : 66666666Hz
uart clock    : 24000000Hz
ahb clock     : 133333333Hz
axi_a clock   : 400000000Hz
axi_b clock   : 200000000Hz
weim_clock    : 100000000Hz
ddr clock     : 160000000Hz
esdhc1 clock  : 80000000Hz
esdhc2 clock  : 80000000Hz
esdhc3 clock  : 80000000Hz
esdhc4 clock  : 80000000Hz
Board: MX50 ARM2 board
Boot Reason: [WDOG]
Boot Device: MMC
I2C:   ready
DRAM:  256 MB
MMC:   FSL_ESDHC: 0, FSL_ESDHC: 1, FSL_ESDHC: 2
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
temperature 28
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
In:    serial
Out:   serial
Err:   serial
Net:   got MAC address from IIM: 00:00:00:00:00:00
FEC0
Hit any key to stop autoboot:  0

12:37 AM

I didn't let it boot

Why?

I wanted to know what that watchdog trigger was haha that I saw fly by

12:38 AM

will try again

1

12:40 AM

MX50_ARM2 U-Boot > setenv bootcmd mmc read 2 0x70008000 0x800 0x1400\; bootm 0x70008000
MX50_ARM2 U-Boot > saveenv
Saving Environment to MMC...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
Writing to MMC(2)... done
MX50_ARM2 U-Boot > reset
resetting ...
▒

U-Boot 2009.08 (Dec 21 2011 - 14:48:38)

CPU:   Freescale i.MX50 family 1.1V at 800 MHz
mx50 pll1: 800MHz
mx50 pll2: 400MHz
mx50 pll3: 216MHz
ipg clock     : 66666666Hz
ipg per clock : 66666666Hz
uart clock    : 24000000Hz
ahb clock     : 133333333Hz
axi_a clock   : 400000000Hz
axi_b clock   : 200000000Hz
weim_clock    : 100000000Hz
ddr clock     : 160000000Hz
esdhc1 clock  : 80000000Hz
esdhc2 clock  : 80000000Hz
esdhc3 clock  : 80000000Hz
esdhc4 clock  : 80000000Hz
Board: MX50 ARM2 board
Boot Reason: [WDOG]
Boot Device: MMC
I2C:   ready
DRAM:  256 MB
MMC:   FSL_ESDHC: 0, FSL_ESDHC: 1, FSL_ESDHC: 2
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
temperature 28
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
In:    serial
Out:   serial
Err:   serial
Net:   got MAC address from IIM: 00:00:00:00:00:00
FEC0
Hit any key to stop autoboot:  0

12:40 AM

MMC read: dev # 2, block # 2048, count 5120 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
5120 blocks read: OK
## Booting kernel from Legacy Image at 70008000 ...
   Image Name:   Normal Kernel
   Image Type:   ARM Linux Kernel Image (uncompressed)
   Data Size:    2288064 Bytes =  2.2 MB
   Load Address: 70008000
   Entry Point:  70008000
   Verifying Checksum ... OK
   XIP Kernel Image ... OK
OK

Starting kernel ...



U-Boot 2009.08 (Dec 21 2011 - 14:48:38)

12:41 AM

seems like it saved it's config though

12:41 AM

as now its watchdog every time now after starting kernel

12:42 AM

Saves me having to spam ever 15 seconds to prevent a lock up haha

12:43 AM

Just to note the SDCard in there now is just a blank fat card with rupor's raw file img and our raw file uImage sitting on it.

12:43 AM

(No os partitions etc.)

12:44 AM

I was trying to get the fatls working but can't seem to, to see if it can actually see the files.

XIP should not be there

12:45 AM

Can't understand why it is

12:45 AM

But it probably doesn't help

12:46 AM

This device is a load of crap (edited)

how do we form this command to see what's on the sd again

MX50_ARM2 U-Boot > fatls mmc 0
failed in set read bl len = 0x0
** Can't read from device 0 **

** Unable to use mmc 0:1 for fatls **

12:47 AM

MX50_ARM2 U-Boot > fatls mmc 2
** Partition 1 not valid on device 2 **

** Unable to use mmc 2:1 for fatls **
MX50_ARM2 U-Boot >

You should use MMC 2

12:47 AM

It's your external SD right?

12:47 AM

Or maybe not

12:49 AM

Oh

2 must be the nand

12:49 AM

as we are booting a kernel from it

Which part of the SD is it you've formatted in fat?

12:50 AM

Ok Then change setenv command

there is nothing on the SD cept raw files so that wouldn't work if that was 2 (edited)

setenv bootcmd mmc read 0 0x70008000 0x800 0x1400\; bootm 0x70008000
saveenv
reset

12:51 AM

It won't work but we will have at least the right MMC

Should we really save that

12:51 AM

instead of just booting it right away?

12:52 AM

then if it fails we will still have something?

12:52 AM

setenv bootcmd mmc read 0 0x70008000 0x800 0x1400\; bootm 0x70008000
boot

Actually writing the commands takes more time

12:52 AM

So less watchdog time

12:53 AM

If they are in the env they load and execute quicker (edited)

I'm just saying that "boot" is an actual U-boot command.

Yes but that is not what we want

ok

Boot will execute bootcmd

12:54 AM

So it makes no sense

tux-linux

setenv bootcmd mmc read 0 0x70008000 0x800 0x1400\; bootm 0x70008000
saveenv
reset

Ok going to do this

1

12:56 AM

Ok I see now

12:56 AM

right to uboot no issue

12:56 AM

Hit any key to stop autoboot:  0

MMC read: dev # 0, block # 2048, count 5120 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
SD set clock 50MHz
5120 blocks read: OK
Wrong Image Format for bootm command
ERROR: can't get kernel image!

12:57 AM

I was worried that we were not going to get uboot.

12:57 AM

then have nothing to work with

U-Boot doesn't have anything to do with this command

Understood now.

DD custom uimage to external MMC Sector 2048 (edited)

So basically just put the 4GB card in?

Yeah

or you want blank card with just the uImage written to it?

Whichever is fine

Hit any key to stop autoboot:  0

MMC read: dev # 0, block # 2048, count 5120 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
SD set clock 50MHz
5120 blocks read: OK
## Booting kernel from Legacy Image at 70008000 ...
   Image Name:   Linux-2.6.35.3
   Image Type:   ARM Linux Kernel Image (uncompressed)
   Data Size:    2225840 Bytes =  2.1 MB
   Load Address: 70008000
   Entry Point:  70008000
   Verifying Checksum ... OK
   XIP Kernel Image ... OK
OK

Starting kernel ...

1:02 AM

watchdog

1:03 AM

2.1 MB (our kernel)

Nonsense I'm giving up for tonight Good night!

haha ok thanks for playing hope you have no nightmares

1

No luck with the new kernel just to note. same thing

1

tux-linux

Just an idea Might not work

may use some resistor between the pins, that might help

1

Do you have any idea about how to disable the watchdog?

tux-linux

Do you have any idea about how to disable the watchdog?

@andi

4:55 PM

@tux-linux can you show me the same process with kobo compile kernel -> busybox prompt so I can have a general understanding what it looks like?

hmm, so what do we have involved here: watchdog of IMX, watchdog of PMIC, watchdog of some msp433 thingy?

4:57 PM

does just waiting at u-boot prompt trigger the watchdog?

Basically not sure about the msp430 the IC label is smudged off it says m430 on it but I don't believe they label their chips like that.

FCC ID is what?

I think its more just watchdog between IMX and PMIC

4:59 PM

the PMIC is a wm8123

andi

FCC ID is what?

Where does one find that?

5:00 PM

FCC ID AK8PRST1 Digital Book Reader by Sony Corporation

FCC ID application submitted by Sony Corporation for Digital Book Reader for FCC ID AK8PRST1 ( AK8 PRST1 ) User Manual, Frequency, Reports, Images and more.

5:03 PM

hmm those links would have been nice... Schematic Diagram Schematics Adobe Acrobat PDF (621 kB) 2011-07-28 Parts List Parts List/Tune Up Info Adobe Acrobat PDF (442 kB) 2011-07-28

5:03 PM

(broken links or maybe you have to pay or something to get them)

andi

hmm, so what do we have involved here: watchdog of IMX, watchdog of PMIC, watchdog of some msp433 thingy?

Yes it does (Regarding U-Boot, I mean) (edited)

NonaSuomy

hmm those links would have been nice... Schematic Diagram Schematics Adobe Acrobat PDF (621 kB) 2011-07-28 Parts List Parts List/Tune Up Info Adobe Acrobat PDF (442 kB) 2011-07-28

Internet archive be like

hmm, the worst internal photos I have seen, often the ones from FCC application are the best

Yes, they are horrible 2011 probably doesn't help

if you break into uboot, do you also have watchdog problems?

5:34 PM

what kind of commands does this u-boot have?

See pinned messages

NonaSuomy

MX50_ARM2 U-Boot > help ? - alias for 'help' autoscr - DEPRECATED - use "source" command instead base - print or set address offset bdinfo - print Board Info structure boot - boot default, i.e., run 'bootcmd' bootd - boot default, i.e., run 'bootcmd' bootm - boot application image from memory bootp - boot image via network using BOOTP/TFTP protocol cls - clear screen cmp - memory compare coninfo - print console devices and information cp - memory copy crc32 - checksum calculation dhcp - boot image via network using DHCP/TFTP protocol echo - echo args to console erase - erase FLASH memory ext2load- load binary file from a Ext2 filesystem ext2ls - list files in a directory (default /) fatinfo - print information about filesystem fatload - load binary file from a dos filesystem fatls - list files in a directory (default /) flinfo - print FLASH memory information go - start application at address 'addr' help - print online help i2c - I2C sub-system iminfo - print header information for application image imxtract- extract a part of a multi-image itest - return true/false on integer compare loadb - load binary file over serial line (kermit mode) loads - load S-Record file over serial line loady - load binary file over serial line (ymodem mode) loop - infinite loop on address range md - memory display mii - MII utility commands mm - memory modify (auto-incrementing address) mmc - MMC sub system mmcinfo - mmcinfo <dev num>-- display MMC info (edited)

.

NonaSuomy

mtest - simple RAM read/write test mw - memory write (fill) nfs - boot image via network using NFS protocol nm - memory modify (constant address) ping - send ICMP ECHO_REQUEST to network host printenv- print environment variables protect - enable or disable FLASH write protection rarpboot- boot image via network using RARP/TFTP protocol reset - Perform RESET of the CPU run - run commands in an environment variable saveenv - save environment variables to persistent storage setenv - set environment variables sf - SPI flash sub-system sleep - delay execution for some time source - run script from memory sspi - SPI utility commands tftpboot- boot image via network using TFTP protocol version - print monitor version MX50_ARM2 U-Boot >

.

andi

if you break into uboot, do you also have watchdog problems?

Yes

hmm, does the watchdog get turned off, when you do any command in u-boot which takes time (like loady and stuff)

No

5:48 PM

It's still triggered

Nope ever 15sec it gets triggered no matter what

unless something pets it

2

Correct

5:49 PM

We just don’t know which hand to pet it with

It's always biting so we can never know in time

well, the up side is that so people know there is a watchdog. Well, and dogs need to get out for a walk regularly. So for the next more dangerous covid mutant when you should stay at home again, you have an excuse to go out

1

The Sony kernel probably does some magic but we've yet to figure it out what (edited)

so we have one working setup

5:57 PM

to look at

Yes, @NonaSuomy bricked partially the device but U-Boot is accessible

and I just found out that I was too strict with noscript plugin, so that I could not filter pinned messages...

Oh

ok, one way to go: and or enable some printk at i2c writes in the kernel

5:59 PM

in that working sony kernel

But we can't boot a custom kernel It hangs

6:00 PM

Even the Sony one when recompiled

hmm, the timing is the same 15s

6:01 PM

so it really hangs because of the watchdog

Seemingly yes

6:01 PM

Can't understand why it doesn't with the default one in the device

6:01 PM

Maybe Sony did not publish some code

so by inserting some sleep command before bootm, the point where the kernel hangs can be moved?

It will just hang faster I guess?

6:04 PM

Try this @NonaSuomy

hmm, if we somehow geht the original kernel to boot, the whole timing of the boot process would be interesting to see what happens until or after the point when the watchdog would kick in

Yes, good idea

on debian there is a package moreutils with a tool called ts

6:08 PM

I often use ts </dev/ttyUSBx to get some timings, I have also seen terminal programs providing timestamps

Hmm Nice tool

maybe also inserting some sleep command before starting the original kernel

6:09 PM

and increase sleep until it hangs

6:10 PM

if it hangs before i2c initialisation, we have learned something

NonaSuomy did some things with i2cdetect while the system was working

NonaSuomy

[root (ttyGS0)]# ./busybox i2cdetect -l
i2c-0   i2c             imx-i2c                                 I2C adapter
i2c-1   i2c             imx-i2c                                 I2C adapter
i2c-2   i2c             imx-i2c                                 I2C adapter

.

6:14 PM

Thanks for your help

# remount SD card for writing
mount -o remount,rw /initrd/mnt/sd

# your screen driver
dd if=/dev/mmcblk2 of=/initrd/mnt/sd/WF.img skip=17408 count=2048 bs=1024
sync

# other important information, such as the operating voltage of the screen in line 5
dd if=/dev/mmcblk2 of=/initrd/mnt/sd/Vcom.img skip=15876 count=1 bs=1024
sync

# serial, MAC
dd if=/dev/mmcblk2 of=/initrd/mnt/sd/SN.img skip=15872 count=1 bs=1024
sync

Partitions

Disk /dev/sdb: 1.82 GiB, 1958739968 bytes, 3825664 sectors 
Disk model: PRS-XXX          
Units: 1 * 512 sectors = 512 bytes 
Sector size (logical/physical): 512 bytes / 512 bytes 
I/O size ( minimum/optimal): 512 bytes / 512 bytes 
Disk Label Type: dos 
Disk ID: 0x00000000 

Device Boot Start End Sectors Size Id Type 
/dev/sdb1 51200 71679 20480 10M 83 Linux 
/dev/sdb2 71680 92159 20480 10M 83 Linux 
/dev/sdb3 92160 866687 774528 378.2M 5 Extended 
/dev/sdb4 866688 3825663 2958976 1.4G 83 Linux
/dev/sdb5 92176 124991 32816 16M 83 Linux 
/dev/sdb6 125008 403583 278576 136M 83 Linux 
/dev/sdb7 403600 424127 20528 10M 83 Linux 
/dev/sdb8 424144 506111 81968 40M 83 Linux 
/dev/sdb9 506128 604479 98352 48M 83 Linux 
/dev/sdb10 604496 866687 262192 128M 83 Linux 

Partition table entries are not ordered as they are on disk.

(edited)

TranslatorComparison.pdf

1.28 MB

What has this to do with the watchdog?

Still in recovery mode

tux-linux

Try this @NonaSuomy

Have you tried this?

What was said I have no idea what it's about.

Can you still access U-Boot prompt?

Yep

Is there a sleep command there?

Yes it does nothing

It doesn't sleep?

I can tell it to sleep it sits at prompt does nothing for a bit then gives me the console back. It doesn't extend the watchdog still resets if you select too long of sleep.

Does i2c probe take some time?

(I first did this a while back when I thought U-Boot was controlling the watchdog (which it wasn't))

Oh so it works

11:44 PM

Ok

11:44 PM

Then can you sleep a little bit, then run boot?

?? Unsure this would just give you less time to boot if you sleep first as the watchdog will reset in 15 seconds no matter what. You set the sleep to 14 seconds after it comes back it will reset the device within 1 more second.

Yes, it is the goal

I'm not sure I understand how that helps can you explain further?

11:48 PM

So you want it to start booting the kernel after sleep and it only get a few lines in and then it dies and resets. (edited)

andi

if it hangs before i2c initialisation, we have learned something

To try to hang it before it can disable the watchdog, but just enough to get it to boot

I wish the chip wasn't bga so I could tap that pin and hold it high or low.

NonaSuomy

So you want it to start booting the kernel after sleep and it only get a few lines in and then it dies and resets. (edited)

Yes, basically

I don't understand how that helps but I will try anything.

Thanks It's to determine if the log error we have with custom kernel is caused by the watchdog or something else

All I see happening is Sleep -> partial kernel boot -> reset clean slate still watchdog enabled and same process happening.

Does the partial kernel boot hangs at the same place?

I found this ramdisk in a package I wonder if it has to run that first to disable the watchdog then it boots the mmc kernel or something like that.

tux-linux

Does the partial kernel boot hangs at the same place?

Will try in a moment

1

NonaSuomy

I found this ramdisk in a package I wonder if it has to run that first to disable the watchdog then it boots the mmc kernel or something like that.

Possible, might be writing to i2c from userspace

If we knew what to write.

1

11:55 PM

sniff the bus on boot.

11:56 PM

Would really love to know what prstux did

11:56 PM

I begged him he just repeated his before message about no access to the old files.

NonaSuomy

Would really love to know what prstux did

Maybe email the maintainer? If we have spent so much time on this watchdog, so he probably did; maybe he has some memories (edited)

I don't understand how you can spend so much work on something and then just want to see it die. (edited)

Next thing would be JTAG But I've never done that

11:59 PM

It's like bare hardware debugging

12:01 AM

There you can probably sniff the bus as you said

I have this some place http://dangerousprototypes.com/docs/Open_Bench_Logic_Sniffer

12:05 AM

Not sure if it ever worked properly though.

Interesting

Russian + English eMMC hookup (Google Translate) (edited)

SonyPRS-T1.pdf

1.68 MB

ironwayrecoverySonyPRS-T1.pdf

1.17 MB

12:20 AM

Was hunting this down for too long.

12:20 AM

So many broken links I went through.

eMMC hookup to computer (USB export)?

yes

12:22 AM

R/W

Awesome

12:23 AM

Did you try it?

Current issue is the level shifter I have is TXB not TXS

12:24 AM

https://www.aliexpress.com/wholesale?trafficChannel=main&d=y&CatId=0&SearchText=txs0108&ltype=wholesale&SortType=price_asc&groupsort=1&page=1

12:25 AM

TXS is meant for MMC level shifting. YF08E is the TXS0108. TXS0108 is for Open Drain circuits (like I2C, requires pull up or pull down resistors) but can be used in push-pull circuits, but you need to lower impedance by adding in some parallel resistors. (edited)

12:25 AM

TXB is meant for SPI level shifting. YE08 is the TXB0108. (edited)

12:26 AM

That was what the PDF above was explaining the comparison between the two https://discord.com/channels/809205711778480158/966554812567130152/971978847748751371 (edited)

12:26 AM

I may still try B as I still don't see why it would matter

12:27 AM

B supports even lower voltages than 1.8

12:27 AM

May have something to do with "OpenDrain" though as that seemed to be listed on the S and not the B

Try the level shifter in the Kindle touch thread ;)

Does it use the same chip?

https://www.mobileread.com/forums/showpost.php?p=1897643&postcount=1

MobileRead Forums - View Single Post - Kindle serial adapter level...

MobileRead - the resource for mobile geeks seeking information and advice for keeping their gadgets happy.

12:29 AM

Oh, it needs a specific chip?

12:29 AM

Mine is just between my UART host and Kindle

Ah just passive components

12:30 AM

That's just for the serial communication

Yes

Above is for directly interfacing to the eMMC https://discord.com/channels/809205711778480158/966554812567130152/971989544478064640 (edited)

1

12:31 AM

(We already have 3.3v UART communication on the PRS-T1) (edited)

NonaSuomy

Above is for directly interfacing to the eMMC https://discord.com/channels/809205711778480158/966554812567130152/971989544478064640 (edited)

It would basically be like plugging the eMMC into you computer like a USB Flash Drive.

1

12:34 AM

Would probably work for all readers if they had pins broken out to the eMMC

12:35 AM

I wish I knew how to map out what all these tests points are on the back of the PRS one I bet one is the watchdog pin haha.

12:36 AM

I could probably figure out a few with a mulitmeter as long as they don't go into BGA land.

NonaSuomy

I wish I knew how to map out what all these tests points are on the back of the PRS one I bet one is the watchdog pin haha.

That's probably true; try it A multimeter should be safe enough

Ok so set to go you want me to do like a 14 sec sleep then what?

maybe more like start at 5 secs and increase by 2 each time until it breaks or you see a difference

NonaSuomy

Would probably work for all readers if they had pins broken out to the eMMC

In recent U-Boots, there's the ums command which does that too (USB export) (edited)

example

Hit any key to stop autoboot:  0
MX50_ARM2 U-Boot > sleep 16
MX50_ARM2 U-Boot > sl <-- All I was able to type before lockup and watchdog reset.

1

12:51 AM

This may not be making sense or my timing was off... sleep 20 still getting a console back

hmm hmm

12:51 AM

interesting

sleep 25 still ok

12:53 AM

sleep 30 too much.

12:55 AM

sleep 28 console back

thanks for testing

I think it's a 30 sec watchdog

12:58 AM

Yup I was able to get a console back after 30 seconds now by typing 1 letter before the prompt then pasting it in the window sleep 30 then it dies right away after the prompt.

Oh! interesting

12:59 AM

if you try that with the custom kernel...? (edited)

It's already set to boot the kernel

## Booting kernel from Legacy Image at 70008000 ...
   Image Name:   Linux-2.6.35.3
   Image Type:   ARM Linux Kernel Image (uncompressed)
   Data Size:    2226004 Bytes =  2.1 MB
   Load Address: 70008000
   Entry Point:  70008000
   Verifying Checksum ... OK
   XIP Kernel Image ... OK
OK

Starting kernel ...

Do you want me to type

sleep 25
boot

(edited)

no, rather try to hit a key when it's seemingly hanging in the kernel boot process

1:01 AM

but sleep 25; boot could be interesting too (edited)

Hit any key to stop autoboot:  0
MX50_ARM2 U-Boot > sleep 25; boot

MMC read: dev # 0, block # 2048, count 5120 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
SD set clock 50MHz
5120 blocks read: OK
## Booting kernel from Legacy Image at 70008000 ...
   Image Name:   Linux-2.6.35.3
   Image Type:   ARM Linux Kernel Image (uncompressed)
   Data Size:    2226004 Bytes =  2.1 MB
   Load Address: 70008000
   Entry Point:  70008000
   Verifying Checksum ... OK
   XIP Kernel Image ... OK
OK

Starting kernel ...



U-Boot 2009.08 (Dec 21 2011 - 14:48:38)

CPU:   Freescale i.MX50 family 1.1V at 800 MHz
mx50 pll1: 800MHz
mx50 pll2: 400MHz
mx50 pll3: 216MHz
ipg clock     : 66666666Hz
ipg per clock : 66666666Hz
uart clock    : 24000000Hz
ahb clock     : 133333333Hz
axi_a clock   : 400000000Hz
axi_b clock   : 200000000Hz
weim_clock    : 100000000Hz
ddr clock     : 160000000Hz
esdhc1 clock  : 80000000Hz
esdhc2 clock  : 80000000Hz
esdhc3 clock  : 80000000Hz
esdhc4 clock  : 80000000Hz
Board: MX50 ARM2 board
Boot Reason: [POR]
Boot Device: MMC
I2C:   ready
DRAM:  256 MB
MMC:   FSL_ESDHC: 0, FSL_ESDHC: 1, FSL_ESDHC: 2
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
temperature 28
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
In:    serial
Out:   serial
Err:   serial
Net:   got MAC address from IIM: 00:00:00:00:00:00
FEC0
Hit any key to stop autoboot:  0

1:11 AM

MMC read: dev # 0, block # 2048, count 5120 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
SD set clock 50MHz
5120 blocks read: OK
## Booting kernel from Legacy Image at 70008000 ...
   Image Name:   Linux-2.6.35.3
   Image Type:   ARM Linux Kernel Image (uncompressed)
   Data Size:    2226004 Bytes =  2.1 MB
   Load Address: 70008000
   Entry Point:  70008000
   Verifying Checksum ... OK
   XIP Kernel Image ... OK
OK

Starting kernel ...

1:14 AM

sleep 28; boot

Hit any key to stop autoboot:  0
MX50_ARM2 U-Boot > sleep 28; boot

MMC read: dev # 0, block # 2048, count 5120 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
SD set clock 50MHz
5120 blocks read: OK
## Booting kernel from Legacy Image at 70008000 ...
   Image Name:   Linux-2.6.35.3
   Image Type:   ARM Linux Kernel Image (uncompressed)
   Data Size:    2226004 Bytes =  2.1 MB
   Load Address: 70008000
   Entry Point:  70008000
   Verifying Checksum ... <-- Watchdog'd here

(edited)

1:17 AM

sleep 29; boot

Hit any key to stop autoboot:  0
MX50_ARM2 U-Boot > sleep 29; boot

MMC read: dev # 0, block # 2048, count 5120 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
SD set clock 50MHz <-- Watchdog'd here

1:18 AM

sleep 30; boot

Hit any key to stop autoboot:  0
MX50_ARM2 U-Boot > sleep 30; boot

MMC read: dev # 0, block # 2048, count 5120 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1 <-- Watchdog'd here

(edited)

1:20 AM

This was my hypnosis above.

welp i never thought we would spend so much time on a @#$% watchdog

I wonder if the kernel actually booted if it would pet the dog properly

@andi here's some interesting logs for you as per your previous request

NonaSuomy

I could probably figure out a few with a mulitmeter as long as they don't go into BGA land.

I guess this is actually not wise as these circuits run at 1.8 -> 3.3v and continuity test run at 9v from your meter. (edited)

1

BUG: scheduling while atomic: swapper/1/0x00000002
Modules linked in:
[<8002a594>] (unwind_backtrace+0x0/0xf0) from [<8031ef5c>] (schedule+0x6c/0x324)
[<8031ef5c>] (schedule+0x6c/0x324) from [<8031f80c>] (schedule_timeout+0x18/0x1fc)
[<8031f80c>] (schedule_timeout+0x18/0x1fc) from [<8031f688>] (wait_for_common+0xf8/0x1b8)
[<8031f688>] (wait_for_common+0xf8/0x1b8) from [<80245930>] (mmc_wait_for_req+0x114/0x120)
[<80245930>] (mmc_wait_for_req+0x114/0x120) from [<8024d530>] (transfer+0x1b8/0x318)
[<8024d530>] (transfer+0x1b8/0x318) from [<8024d7f0>] (rawdata_write+0x160/0x238)
[<8024d7f0>] (rawdata_write+0x160/0x238) from [<8025ec34>] (past_mmc_header_write+0x48/0xb4)
[<8025ec34>] (past_mmc_header_write+0x48/0xb4) from [<8025f770>] (past_mmc_write+0x4e0/0x4fc)
[<8025f770>] (past_mmc_write+0x4e0/0x4fc) from [<8025e114>] (pastlog_mix2past_write+0x44/0x50)
[<8025e114>] (pastlog_mix2past_write+0x44/0x50) from [<8025e138>] (pastlog_mix2past_flush+0x18/0x28)
[<8025e138>] (pastlog_mix2past_flush+0x18/0x28) from [<8025e188>] (past_flush+0x40/0x54)
[<8025e188>] (past_flush+0x40/0x54) from [<8031ed40>] (panic+0x70/0xe4)
[<8031ed40>] (panic+0x70/0xe4) from [<80008e14>] (mount_block_root+0x1cc/0x20c)
[<80008e14>] (mount_block_root+0x1cc/0x20c) from [<8000906c>] (prepare_namespace+0x128/0x180)
[<8000906c>] (prepare_namespace+0x128/0x180) from [<80008484>] (kernel_init+0x124/0x16c)
[<80008484>] (kernel_init+0x124/0x16c) from [<800269e0>] (kernel_thread_exit+0x0/0x8)

2:05 AM

(After holding home + menu on boot)

2:05 AM

probably looking for the recover partition

Wooo

2:05 AM

We got to some point

Just endlessly repeats that with no watchdog reset

1

This seems to be an important milestone in the boot process

Maybe if you boot from recover mode it disabled the watchdog?

Try it?

I'm not sure how

2:09 AM

Because there is nothing on the eMMC I can't get rupor's recover working to do anything.

2:09 AM

I wish I could see what happened before that but too hard to hold the buttons and watch the uart

2:10 AM

It's also super fast repeating fills the buffer really fast

Try fun Tak, blue putty, whatever it's called to try to hold the wires a bit

maybe I can just slide the button part off the table

1

I wonder if it can detect if the screen is connected?

2:19 AM

Does the Kobo boot with no screen connected?

NonaSuomy

I wonder if it can detect if the screen is connected?

what

Got it this time....

2:41 AM

https://discord.com/channels/809205711778480158/966554812567130152/970169381759877170

13.68 KB

HOME + OPTION key input
...
Read recovery boot environment

At least I know it still detects pushing those buttons now. It just has no recovery data on the drive to point it to rupor recover.

NonaSuomy

Does the Kobo boot with no screen connected?

Yes

7:54 AM

Something to try might be to compile the kernel with the toolchain they seem to use (GCC 4.4.0, look at boot log)

7:55 AM

Did you try our custom kernel with ttymxc4?

Pretty old toolchain it is

tux-linux

Did you try our custom kernel with ttymxc4?

That 2.1MB kernel above has ttymxc4 set the stock kernel is 2.2MB https://discord.com/channels/809205711778480158/966554812567130152/972003313082044476 (edited)

In config_cmdline?

tux-linux

Something to try might be to compile the kernel with the toolchain they seem to use (GCC 4.4.0, look at boot log)

1:44 PM

How does one use older versions?

it just does

1:45 PM

try it

Well, now maybe it matters

1

2:55 PM

Like the Kindle

Can you modify this line to show me the command for GCC 4.4.0 compile make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- -j8 uImage

do you have the toolchain in you path?

2:57 PM

this 4.4.0

Szybet shhh

?

haha

2:57 PM

❤️

i dont really understand

Szybet isn't supposed to mean "tree"?

in what language does it mean tree?

4:32 PM

Szybet -> "Szyba" in polish -> Glass

Idk, look at Google translate

4:36 PM

It gets weirder with every translate

I know

4:36 PM

I translated from Polish to English and got "shaft"

4:36 PM

Arbre means tree in French

oh in french

4:37 PM

okay

NonaSuomy

❤️

I guess that Google translate showed "love" as the translation

NonaSuomy

Can you modify this line to show me the command for GCC 4.4.0 compile make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- -j8 uImage

You need to download the toolchain

4:40 PM

I'll get you a download link when I get home

4:40 PM

In an hour or so

Ok thank you

i looked up the repo

5:16 PM

https://gitlab.com/phreakuencies/prstux

Matias N. / prstux · GitLab

PRS-TUX: Linux system for Sony PRS-T* ebook readers

5:16 PM

its just so weird

5:16 PM

those makefiles

5:16 PM

there is nothing

1

tux-linux

I'll get you a download link when I get home

There https://osdn.net/frs/g_redir.php?m=jaist&f=iadfilehost%2Fdevtools%2Farm-2009q3-67-arm-none-linux-gnueabi-i686-pc-linux-gnu.tar.bz2

Download IAD Filehost from SourceForge.net

project page for IAD members

If it doesn't work you need to install 32 bit libs

tux-linux

In config_cmdline?

tux-linux

If it doesn't work you need to install 32 bit libs

I think I added them originally to compile the stock kernel.

Szybet

there is nothing

.gitignore files usually tell a story, sometimes give you hints of things missing. The one you see there https://gitlab.com/phreakuencies/prstux/-/blob/master/.gitignore it excludes the kernel folder. (edited)

.gitignore · master · Matias N. / prstux · GitLab

PRS-TUX: Linux system for Sony PRS-T* ebook readers

6:40 PM

He probably had some custom make files in there to do that job from the prior make files.

6:41 PM

Not just the kernel source and didn't realize it.

6:42 PM

Always when you make projects as large as these. Take your current dev box and exclude it from final build test environment. Rebuild a new machine so you know what other users will go through trying to compile your code on a fresh machine. (edited)

There they flash the kernel https://gitlab.com/phreakuencies/prstux/-/blob/master/scripts/flash_normal_kernel.sh

scripts/flash_normal_kernel.sh · master · Matias N. / prstux · GitL...

PRS-TUX: Linux system for Sony PRS-T* ebook readers

@tux-linux what is the ramdisk for? https://gitlab.com/phreakuencies/prstux/-/blob/master/scripts/flash_normal_ramdisk.sh

scripts/flash_normal_ramdisk.sh · master · Matias N. / prstux · Git...

PRS-TUX: Linux system for Sony PRS-T* ebook readers

7:59 PM

new.ramdisk

513 KB

7:59 PM

It's extractable with 7zip

7:59 PM

I saw it in other backups

7:59 PM

new.ramdisk.uimg

https://source.android.com/devices/bootloader/partitions/ramdisk-partitions

8:26 PM

Ramdisk (Android 8.x and lower) Contains a kernel and a ramdisk.img.

8:29 PM

Something about some devices use the ramdisk as the recovery.

tux-linux

There https://osdn.net/frs/g_redir.php?m=jaist&f=iadfilehost%2Fdevtools%2Farm-2009q3-67-arm-none-linux-gnueabi-i686-pc-linux-gnu.tar.bz2

How do you use it?

8:49 PM

Can I test this kernel we compiled first in arm emulator?

It will be harder to try there than on the actual device

Nice

qemu-system-arm -M versatilepb -cpu cortex-a8 -kernel /home/nonasuomy/code/linux-2.6.35.3/arch/arm/boot/uImage -m 128
pulseaudio: set_sink_input_volume() failed
pulseaudio: Reason: Invalid argument
pulseaudio: set_sink_input_mute() failed
pulseaudio: Reason: Invalid argume

It really boots?

yes

9:13 PM

to that point

9:13 PM

Then it says no screen

It doesn't boot It's QEMU error messages

ok

9:13 PM

I tried

9:13 PM

I didn't expect to see anything yet it said pulseaudio

It won't boot until you feed it with a zimage with versatilepb device tree

9:14 PM

And probably kernel >4.x, so basically building a completely different one, not really useful for our testing

Apparently you can get further with make ARCH=arm CROSS_COMPILE=arm-none-linux-gnueabi- versatile_defconfig

9:23 PM

People said you you have to use at the least 512mb as well but when I tried that it said max is 256mb

9:24 PM

also said there is a -M realview-pb-a8 -cpu cortex-a8 (edited)

9:24 PM

Someone said they got as far as (qemu) qemu: fatal: Trying to execute code outside RAM or ROM at 0x70008000

9:28 PM

They were using 2.6.31-207

9:34 PM

https://balau82.wordpress.com/2010/03/22/compiling-linux-kernel-for-qemu-arm-emulator/

Balau

Compiling Linux kernel for QEMU ARM emulator

How to compile a vanilla Linux kernel for the Versatile ARM platform, and run it using QEMU.

9:35 PM

(version 2.6.33)

NonaSuomy

How do you use it?

@tux-linux

9:58 PM

Minus the emulator talk.

extract tarball

10:11 PM

cd tarball/bin

10:12 PM

export PATH="${PATH}:${PWD}" (edited)

10:12 PM

then you can set CROSS_COMPILE to arm-none-linux-gnueabi-

Thank you

1

10:18 PM

compiling...

10:22 PM

Image Name: Linux-2.6.35.3 Created: Fri May 6 22:21:27 2022 Image Type: ARM Linux Kernel Image (uncompressed) Data Size: 2244756 Bytes = 2192.14 KiB = 2.14 MiB Load Address: 70008000 Entry Point: 70008000 Image arch/arm/boot/uImage is ready

1

Hit any key to stop autoboot:  0

MMC read: dev # 0, block # 2048, count 5120 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
SD set clock 50MHz
5120 blocks read: OK
## Booting kernel from Legacy Image at 70008000 ...
   Image Name:   Linux-2.6.35.3
   Image Type:   ARM Linux Kernel Image (uncompressed)
   Data Size:    2244756 Bytes =  2.1 MB
   Load Address: 70008000
   Entry Point:  70008000
   Verifying Checksum ... OK
   XIP Kernel Image ... OK
OK

Starting kernel ...

10:58 PM

not really sure what it's doing in it's current state

10:58 PM

it probably needs some more of those other partitions or something.

10:59 PM

I wonder if we loaded it from the recover partition spot if that would have different results

http://dangerousprototypes.com/docs/Bus_Pirate interesting

tux-linux

http://dangerousprototypes.com/docs/Bus_Pirate interesting

a very interesting device, but not usefull for hacking kobo's

it was for trying to sniff the watchdog activity via i2c

you can do this with an arduino propably ( well by skipping that you need to know which lines are i2c, and that device will not crash or something, maybe even damage itself ) (edited)

7:06 PM

and i dont think watchdog goes over i2c

7:06 PM

no way, watchdogs are cpu internal things

this one seems to be i2c though

whaaat

wm8321 it's a power management IC

1

and what is u-boot doing to him

Szybet

and what is u-boot doing to him

In the wm8321 manual some interesting things says you can lockout watchdog modifications with a code. Another is that you can set after so many watchdog triggers it won’t power the device back on (which it does) I was wondering why the device was off after leaving it on for a while…

well okay

Szybet

and what is u-boot doing to him

There are 3 ways you can do watchdogs U-Boot/Kernel/PMIC power management

7:25 PM

Uboot in this case does nothing

but i dont see how hardware-hacking would help to port inkbox to it

NonaSuomy

Uboot in this case does nothing

??? so what triggers the watchdog

Kernel enables watchdog after kernel loads to catch failures

7:26 PM

We don’t make it to this part so it’s the third the PMIC causing it to trigger when we don’t pet it.

Szybet

??? so what triggers the watchdog

The PMIC

not good

7:27 PM

kexec doesn't work on this device?

Szybet

but i dont see how hardware-hacking would help to port inkbox to it

We can’t get to that point because of the external watchdog

7:28 PM

Right when it is about to load busybox it takes too long and then resets

the only way i see this is to make the kernel smaller

It doesn’t matter about that

hardware - hacking is the wrong direction to go to if the port is intendeed for normal people too

The pmic would still trigger the watchdog as we didn’t pet it yet

Szybet

hardware - hacking is the wrong direction to go to if the port is intendeed for normal people too

Incorrect (edited)

7:30 PM

Once we find out how to pet the dog we do that in software before the boot

NonaSuomy

Once we find out how to pet the dog we do that in software before the boot

but if you load a smaller kernel it will not trigger it?

It will always trigger unless you pet it

NonaSuomy

Incorrect (edited)

well then there is a osciloscope needed with i2c support, not a universal hacking board

NonaSuomy

It will always trigger unless you pet it

a smaller kernel will pet it before

7:32 PM

thats how sony does it?

Doesn’t matter what device you use

7:33 PM

Oscilloscope with bus sniffing is way more expensive than a protocol sniffing board

7:34 PM

Your talking about $500-1000 vs ~$50 or less

Szybet

a smaller kernel will pet it before

A smaller kernel won’t do anything without knowing how to pet the external watchdog.

7:36 PM

We disabled all the kernel watchdog features and it still triggered as it is external to mx508

NonaSuomy

Your talking about $500-1000 vs ~$50 or less

overstatement

NonaSuomy

We disabled all the kernel watchdog features and it still triggered as it is external to mx508

but how sony does that? you have the kernel sources

Last oscilloscope I bought was just under $500 a Rigol and the one up from it does bus sniffing so whatever you say.

7:38 PM

Like one of the cheapest decent scope companies

osciloscope would be best for that becouse you could detect what is happening on this line, but yea to expensive for the most part

Szybet

the only way i see this is to make the kernel smaller

it doesn't work

Szybet

but how sony does that? you have the kernel sources

?

we didn't find yet

cant you use sony's kernel sources and adjust it to inkbox?

7:41 PM

what about this debian distro for this device? there wasn't kernel sources right?

that's what we are trying to do since two weeks (edited)

you use a smaller sony's kernel that is disabling the watchdog and then loading busybox, whats the problem with this?

it doesn't load

what

@NonaSuomy , did you get around the ttymxc2/mxc4 thing I suggested a few days ago?

7:43 PM

^unless that works we're basically screwed

7:44 PM

even the stock kernel sources don't work with the stock system

tux-linux

it doesn't load

any error? anything?

7:45 PM

first, can you load sony's kernel without any modyfications, self compiled?

I have to fix the eMMC first to test again.

I wonder what would happen if we enabled the wm8321 functions in the kernel

8:19 PM

PRS-T1_989329901_Service_Manual.pdf

1.46 MB

8:20 PM

Few interesting things in there about waveform and changing screens.

8:21 PM

Explains what the numbers mean on the ink screen flex cable

8:22 PM

Wonder if anyone had this tool "WaveForm file translation tool (wwv_parser.exe)" (edited)

8:26 PM

Szybet

first, can you load sony's kernel without any modyfications, self compiled?

no

well here is the problem, every other try won't work becouse of that (edited)

5:36 AM

why? any error?

@NonaSuomy got some interesting logs from the Kindle!

[  274.985211] wm8962_process_req
[  274.988969] wm8962 1-001a: Writing reg : 0x19 value : 0x0
[  274.990196] wm8962 1-001a: Writing reg : 0x1c value : 0x0
[  274.992090] wm8962 1-001a: Writing reg : 0x30 value : 0x8221
[  274.993315] wm8962 1-001a: Writing reg : 0x17 value : 0x0
[  274.994540] wm8962 1-001a: Writing reg : 0x30 value : 0x8021
[  274.995098] wm8962 1-001a: Writing reg : 0x81 value : 0x0

12:26 PM

it's a different watchdog, but things might be similar

you can write there using u-boot or busybox devmem

Can write what?

values to registers

1:28 PM

i think you can use mw in U-Boot

Sorry it didn’t update the chat I see them now. I think that edition of the chip is also the audio processor

1:41 PM

I was reading something of someone that added the 832x to the kernel spec and said most of the functions are the same with all their chips so maybe it might gives us a lot of hints

1:43 PM

When using memory write, how do you know where the memory locations are and what kind of data needs to be written there (edited)

i've not used it before

I saw an example that I pasted before that was for disabling the watchdog for a different chip (won’t be the same) https://discord.com/channels/809205711778480158/966554812567130152/969458955241783377

1:47 PM

Not sure how those values align to that chip or what they are setting exactly

1:48 PM

Above that though gave another little hint of what it was setting I think https://discord.com/channels/809205711778480158/966554812567130152/969458873905844224 (edited)

1:50 PM

So looks like they were making a env variable that triggers the watchdog to disable

1:51 PM

There is an option in the kernel under watchdog for wm831x is that enabled for the Amazon reader? (edited)

1:54 PM

We can probably correlate the data sheet from that IC to this one. If the contents about watchdog are the same we can probably just spam that.

1:55 PM

Maybe in the env somehow like above

hmm wm8962 is just a dac no power management.

3:36 PM

Also different company

3:36 PM

I wonder if they cooperate or bought one another.

3:37 PM

https://www.cirrus.com/products/wm8962-62b/

tux-linux

@NonaSuomy got some interesting logs from the Kindle!

[  274.985211] wm8962_process_req
[  274.988969] wm8962 1-001a: Writing reg : 0x19 value : 0x0
[  274.990196] wm8962 1-001a: Writing reg : 0x1c value : 0x0
[  274.992090] wm8962 1-001a: Writing reg : 0x30 value : 0x8221
[  274.993315] wm8962 1-001a: Writing reg : 0x17 value : 0x0
[  274.994540] wm8962 1-001a: Writing reg : 0x30 value : 0x8021
[  274.995098] wm8962 1-001a: Writing reg : 0x81 value : 0x0

0x19 looks like this deals with enabling dac functions 0x1c disables the start-up bias current generator 0x30 mic relations https://statics.cirrus.com/pubs/proDatasheet/WM8962_Rev4.4.pdf

https://www.mobileread.com/forums/showthread.php?t=189735

PRS-T1 Altering Root FS - MobileRead Forums

PRS-T1 Altering Root FS Sony Reader Dev Corner

7:57 PM

Post 5

find . -print | cpio -o -H newc | gzip -v9 > ../ramdisk-adb.gz
mkimage -A ARM -O Linux -T ramdisk -C gzip -a 0x70308000 -e 0x70308000 -n "Normal Rootfs" -d ramdisk-adb.gz ramdisk-adb.uimg

8:27 PM

https://balau82.wordpress.com/2010/04/12/booting-linux-with-u-boot-on-qemu-arm/

Balau

Booting Linux with U-Boot on QEMU ARM

Using QEMU to emulate the complete boot procedure of an ARM system using U-Boot, the Linux kernel and Busybox.

8:28 PM

http://www.maddes.net/files//u-boot/extract_uImage.sh

what are you trying to do with this script? (Well, why do you need the uImage?) (edited)

tux-linux

what are you trying to do with this script? (Well, why do you need the uImage?) (edited)

I dont understand him too

tux-linux

what are you trying to do with this script? (Well, why do you need the uImage?) (edited)

It pulls out the config.gz of a uImage.

Oh wait that was to extract that full uImage

10:26 AM

@tux-linux you use Gentoo correct?

10:27 AM

Apparently this exists in gentoo /usr/src/linux/scripts/extract-ikconfig and that extracts the config.gz from a uImage

10:29 AM

https://github.com/torvalds/linux/blob/master/scripts/extract-ikconfig

linux/extract-ikconfig at master · torvalds/linux

Linux kernel source tree. Contribute to torvalds/linux development by creating an account on GitHub.

NonaSuomy

https://www.mobileread.com/forums/showthread.php?t=189735

@tux-linux Any thoughts on this post 5 https://www.mobileread.com/forums/showpost.php?p=2207783&postcount=5

MobileRead Forums - View Single Post - PRS-T1 Altering Root FS

MobileRead - the resource for mobile geeks seeking information and advice for keeping their gadgets happy.

If you can't get the kernel to boot, modifying the rootfs is useless

I'm not sure it will allow it to boot without packaging it like that.

12:55 PM

I think it needs that rootfs to disable the watchdog then we point it from there to our kernel.

NonaSuomy

I think it needs that rootfs to disable the watchdog then we point it from there to our kernel.

wrong

12:56 PM

propably

Szybet with the most respect please try not to answer without knowing.

sure, well we will see (edited)

Sony seems to deal with things a far different way with everything they do. If you look up any of their readers nobody has really done anything with them. (edited)

1:14 PM

Ordered one of these to interface with the eMMC https://www.aliexpress.com/item/1005001619337949.html

0.4US $ | 8 Channel Logic Level Bi directional Converter Module TXS...

Smarter Shopping, Better Living! Aliexpress.com

https://web.archive.org/web/20150114012331/http://oss.sony.net:80/Products/Linux/Audio/category03.html

https://web.archive.org/web/20170509133232/http://oss.sony.net/Products/Linux/Audio/Download/common/L1KaMCC9mHxsZfoFPJnsBg/linux-2.6.35.3-110901.tgz (edited)

Source code download

12:01 AM

Seems they used to have every reader there before and you can still download the older kernels there too!

12:01 AM

12:06 AM

Spoke too soon most of the links look dead but hey at the least it had that one that I was wondering about

1

12:11 AM

haha wut...

VERSION = 2
PATCHLEVEL = 6
SUBLEVEL = 35
EXTRAVERSION = .3
NAME = Sheep on Meth

Looks like some "Butty" put this package together.

12:28 AM

At the least they used the right numbers for version in the older package.

12:34 AM

Linux code names are some weird stuff https://en.wikipedia.org/wiki/Linux_kernel_version_history

Linux kernel version history

This article denotes the version history of the Linux kernel. The Linux kernel is a free and open-source, monolithic, Unix-like operating system kernel. It was conceived and created in 1991 by Linus Torvalds.Linux kernels have different support level depending on version. Version 4.4, released in January 2016, was declared to have Long-Term Supp...

I don't understand I tried to make a rootfs it says to extract it out the main image. Did so. Seemed a little small. Said there was garbage data on the end that it ignored. I then compressed it back into the main image and it destroyed the whole thing. The 1.82GB image is now 5.3MB. I opened it and now there is only 4 img inside instead of 9. (edited)

cd ~/code
mkdir rootfs

dd if=t1_eng_\(1.82Gb\)_1.0.05.11130.img of=rootfs.img bs=1 skip=5242880 count=1048512
1048512+0 records in
1048512+0 records out
1048512 bytes (1.0 MB, 1.0 MiB) copied, 1.74376 s, 601 kB/s

dd if=t1_eng_\(1.82Gb\)_1.0.05.11130.img of=rootfs.img.gz bs=1 skip=5242944 count=1048512
1048512+0 records in
1048512+0 records out
1048512 bytes (1.0 MB, 1.0 MiB) copied, 1.76024 s, 596 kB/s

mv rootfs.img.gz ~/code/rootfs
cd ~/code/rootfs
gunzip -c rootfs.img.gz | cpio -i
842 blocks
gzip: rootfs.img.gz: decompression OK, trailing garbage ignored

mv rootfs.img.gz ../

10:06 AM

find . -print
.
./default.prop
./sys
./data
./ebook
./ebook/fonts
./ebook/preload
./ebook/dic
./sbin
./sbin/ts_calibrator
./init
./system
./init.rc
./lib
./lib/modules
./lib/modules/2.6.35.3
./lib/modules/2.6.35.3/build
./lib/modules/2.6.35.3/modules.usbmap
./lib/modules/2.6.35.3/modules.dep.bin
./lib/modules/2.6.35.3/modules.builtin
./lib/modules/2.6.35.3/modules.seriomap
./lib/modules/2.6.35.3/source
./lib/modules/2.6.35.3/modules.ofmap
./lib/modules/2.6.35.3/modules.isapnpmap
./lib/modules/2.6.35.3/modules.inputmap
./lib/modules/2.6.35.3/modules.alias.bin
./lib/modules/2.6.35.3/modules.dep
./lib/modules/2.6.35.3/modules.ccwmap
./lib/modules/2.6.35.3/kernel
./lib/modules/2.6.35.3/kernel/drivers
./lib/modules/2.6.35.3/kernel/drivers/usb
./lib/modules/2.6.35.3/kernel/drivers/usb/gadget
./lib/modules/2.6.35.3/kernel/drivers/usb/gadget/g_android.ko
./lib/modules/2.6.35.3/kernel/drivers/usb/gadget/arcotg_udc.ko
./lib/modules/2.6.35.3/kernel/drivers/usb/serial
./lib/modules/2.6.35.3/kernel/drivers/usb/serial/usbserial.ko
./lib/modules/2.6.35.3/kernel/drivers/usb/host
./lib/modules/2.6.35.3/kernel/drivers/usb/host/ehci-hcd.ko
./lib/modules/2.6.35.3/modules.builtin.bin
./lib/modules/2.6.35.3/modules.alias
./lib/modules/2.6.35.3/modules.pcimap
./lib/modules/2.6.35.3/modules.ieee1394map
./lib/modules/2.6.35.3/modules.order
./lib/modules/2.6.35.3/modules.symbols.bin
./lib/modules/2.6.35.3/modules.symbols
./init.goldfish.rc
./proc
./dev

10:06 AM

find . | cpio -o -H newc | gzip > ../newrootfs.img.gz
842 blocks

mkimage -A arm -O linux -C gzip -T ramdisk -n "Normal Rootfs" -a 0x500040 -e 0x500040 -d ../newrootfs.img.gz ../newrootfs.img
Image Name:   Normal Rootfs
Created:      Tue May 10 04:29:32 2022
Image Type:   ARM Linux RAMDisk Image (gzip compressed)
Data Size:    214702 Bytes = 209.67 KiB = 0.20 MiB
Load Address: 00500040
Entry Point:  00500040

dd if=newrootfs.img of=t1_eng_\(1.82Gb\)_1.0.05.11130.img bs=1 seek=5242880
214766+0 records in
214766+0 records out
214766 bytes (215 kB, 210 KiB) copied, 0.347159 s, 619 kB/s

4:43 PM

Another persons backup of the eMMC

6:06 PM

All those images show the resistors removed from the circled area of D3 DATO D1 D2 that's going to be rough to remove then put back, those solder pads are the size of the needle of a pin

https://www.ebay.ca/itm/325184352800?hash=item4bb67f7220:g:ZYcAAOSw0EJieuDZ

sony prs-t1 eReader | eBay

Find many great new & used options and get the best deals for sony prs-t1 eReader at the best online prices at eBay! Free shipping for many products!

this is an interesting price

9:39 PM

(well, current bid that is)

9:40 PM

maybe I could help if I owned one

Not sure why anyone would want it

9:41 PM

Probability is high you would win.

9:41 PM

with one bid.

well, total price would still be $30

9:41 PM

which, for a student, is kind of a lot

9:44 PM

hmm with the amount of work you have done though, I'm not sure actually trying things out on my side would really be useful

I have only done fumbling around in the dark collecting others information and processing your input.

9:45 PM

I'm sure you would eventually get it.

9:46 PM

prstux is based on the T2 which is basically the T1 without the sound processor and a higher clock 1GHz instead of 800MHz but same chip

hmm

I would send you mine but shipping in Canada is dumb and would cost probably the same as that full ebay unit

9:49 PM

Also it kinda starts you off on a bad foot with no eMMC files.

If you sent yours, it would actually be better, no difficult opening, serial already easier to setup, etc. and exact same hardware platform/bugs there might be But I understand the problem of shipping

9:51 PM

What country are you in?

Same as you.

9:53 PM

They are pretty easy to open

Well then it might not be that bad

like 15 plastic clips around the outside

Shipping for a package like this is around ~15$

9:54 PM

(from experience, I often hunt down FB marketplace, Kijiji for ads)

9:56 PM

But before we do this, if ever we do this, can you send me a link of the kernel source code archive you're currently using that produces the logger bug? I want to look around a bit

I will pay the price for you it will be more fun if we both have one unit.

9:59 PM

*if nobody else bids on it haha

well, thanks

10:00 PM

i guess we should wait until tuesday and see what's happened

10:00 PM

then bid

10:00 PM

(or not)

I don't care in general about the unit I've just been having fun hacking around with it with you.

1

10:02 PM

You will probably get it working in two seconds with that SD Loader forum post and that other vague one.

not so sure

10:03 PM

that watchdog is really a concern

10:03 PM

i can't believe we spent almost three weeks trying to pet a watchdog

I almost found another way through uboot to pet it but uboot wasn't compiled with the multi i2c flag so I can't switch i2c ports

10:05 PM

it's stuck on i2c 1 for some reason

10:05 PM

I think it's on 0

if you compile it with the flag, it fails?

I'm not sure how to compile uboot and what I would do with it after it was compiled

10:06 PM

I'm not sure I want to risk the current loading state of uboot as it's all I have left to play with the device

10:06 PM

unless you can load uboot from uboot.

you can

10:07 PM

look it up i've never done it on kobo but it is possible

10:07 PM

example: U-Boot SPL

does the kobo have multiple i2c ports?

10:08 PM

with different devices on each?

10:09 PM

if you are in uboot does this work: i2c dev 0 or 1 etc

10:11 PM

https://pastebin.com/dM149Hz4 this is the uboot i2c command code

Pastebin.com is the number one paste tool since 2002. Pastebin is a website where you can store text online for a set period of time.

10:11 PM

If you look in there for CONFIG_I2C_MULTI_BUS

10:11 PM

you will see how it works

10:11 PM

but you need that in your uboot config

10:13 PM

from what I can see the default for this sony config is

/*
 * I2C Configs
 */

#ifdef CONFIG_CMD_I2C
    #define CONFIG_HARD_I2C         1
    #define CONFIG_I2C_MXC          1
    #define CONFIG_SYS_I2C_PORT             I2C2_BASE_ADDR
    #define CONFIG_SYS_I2C_SPEED            100000
    #define CONFIG_SYS_I2C_SLAVE            0xfe
#endif

Hmm I'll try on a Kobo soon

uboot/include/configs/mx50_arm2.h

10:16 PM

Have you compiled uboot before?

The only down fall to this is if the watchdog is not this chip but I think it is highly likely. Doc: WM8321G Address R16388(4004h) Watchdog, Bit 14 WDOG_DEBUG, Default 0, Watchdog Pause 0=Disable 1=Enabled (halts the watchdog timer for system debugging) Protected by user key This would show if you can't modify the watchdog bit without the code

i2c dev 0
i2c probe
i2c md.w 0x34 0x4008   -> default value is 0x0000

We have to write a magic value 0x9716 to this register, to allow writes to WDOG_DEBUG etc. MX50_ARM2 U-Boot > i2c mm.w 0x34 0x4008 0x9716 Print out the bit set for 4004 MX50_ARM2 U-Boot > i2c md.w 0x34 0x4004 Then you should be abled to activate the watchdog debug mode MX50_ARM2 U-Boot > i2c mm.w 0x34 0x4004 (change bit 14 of whatever you get from above to 1) Then hopefully no more watch dog unless its that 8bit ic or that weird m430 doing it. (edited)

Hopefully they didn't upgrade it past 1.04 or 1.05 as then rooting is still possible

10:47 PM

Not that you may care about that but eh

don't worry, if we have access to U-Boot everything is possible

10:47 PM

even access for 15 secs only

You can still downgrade it from 1.07 but it's sketchy.

10:49 PM

Just to note the eMMC gets badblocks on these devices as well which cause it not to boot.

honestly the first thing I'll want to do once we have this watchdog fixed up will be to wipe the SD

So probably SD would be the best bet.

yes

Just realize as well though I don't think prstux had to do anything with the watchdog so there is also that avenue if it push comes to shove.

10:52 PM

I think it just involves packaging the stock rootfs properly.

10:52 PM

with mkimage

NonaSuomy

I think it just involves packaging the stock rootfs properly.

That is also how jpa got his stock image loading from SD.

11:05 PM

Here is the back cover with the clip locations the only spot you have to look out for is the part that goes around the headphones jack as you can see I broke it when I peeled it off.

11:07 PM

14 screws in the motherboard.

11:10 PM

Maybe even just make that guy a low offer and see if it accepts it as that button is enabled.

11:10 PM

It will usually give you 3 attempts

11:10 PM

To hit what number they set it to.

11:12 PM

I had one guy not accept the offer then he suggested $10 more for the device so I just put the same old offer I did before in to show him I was not bugging and then he finally accepted it.

1

11:12 PM

So maybe try something like that if it goes down.

11:14 PM

They usually have an auto amount so you will know right away if it is accepted or not.

11:14 PM

try like $5 then maybe $10

minimum is $11.20

haha we tried.

do you prefer to bid or I do?

No bid just make an offer

11:17 PM

You do it then it can ship right to you instead of double shipping

yeah ok

up to you some people are just thriftstore shoppers and just want to get their money back at the least.

11:20 PM

I wouldn't go higher than his starting bid though.

let's say I offer $12

no

no? hm

unless it rejected you at 10?

no

11:21 PM

it says 5 offers left

11:21 PM

i didn't offer anything yet

oh ok. low ball first it will auto reply right away if it was under his set make an offer price so it won't hurt you to do that.

ok $8 maybe?

The make an offer price can be set lower than your starting bid price just so you know.

11:23 PM

That's how it was set with the other item I got.

11:23 PM

go for it.

ok

11:23 PM

let's see

11:24 PM

if it doesn't come back right away then that means it was set even lower than that (edited)

or it wasn't set at all...

that too

11:27 PM

I do everything like this as I loathe waiting around for someone to snipe the bid on you at the last second and a lot of stuff on ebay never sees the light of day for any bids they just go stale.

11:28 PM

most are willing just to accept the make an offer.

we'll see

Realize he will always have doubt that nobody will give him any bids on it so you have that in your corner he won’t want to lose you on it.

1

11:53 PM

Don’t feel shy to slap him with the same price again or lower than his counter offer if its above the min bid price. That is why you want to low ball it first below the min bid as they will try counter offer it higher but usually they counter offer the min bid as it feels weird going higher than that depending on the person. (edited)

11:55 PM

If all else fails we just wait for another.

1

Yay figured out something...

https://github.com/onyx-intl/toolchain/blob/master/imx508/gcc-4.4.4-glibc-2.11.1-multilib-1.0.tar.gz

11.92 KB

2:03 PM

Got the SDCard booting again.

2:03 PM

nice extended log as well.

Ok awesome

NonaSuomy

Got the SDCard booting again.

How?

by loading the rootfs (edited)

Ok

Lucky I had a backup of uboot_new/old which I grabbed the loading command from then launched with that.

1

Is this where we were copying our custom kernel to or was just at some random partition space?

## Booting kernel from Legacy Image at 70800000 ...
   Image Name:   Normal Kernel
   Image Type:   ARM Linux Kernel Image (uncompressed)
   Data Size:    2288064 Bytes =  2.2 MB
   Load Address: 70008000
   Entry Point:  70008000
   Verifying Checksum ... OK
## Loading init Ramdisk from Legacy Image at 70c00000 ...
   Image Name:   Normal Rootfs
   Image Type:   ARM Linux RAMDisk Image (uncompressed)
   Data Size:    286848 Bytes = 280.1 kB
   Load Address: 70308000
   Entry Point:  70308000
   Verifying Checksum ... OK
   Loading Kernel Image ... OK
OK

Kernel is at sector 2048

2:57 PM

The load address is a memory address

# cat /sys/module/rawdatatable/parameters/rawdata_param
MBR                             :0x00000000:0x00000400
uBoot                           :0x00000400:0x000bfc00
Boot Env                        :0x000c0000:0x00020000
Reserved1                       :0x000e0000:0x00020000
Normal Kernel                   :0x00100000:0x00400000
Normal Rootfs                   :0x00500000:0x00100000
Recovery Kernel                 :0x00600000:0x00400000
Reserved2                       :0x00a00000:0x00500000
Normal Boot Env                 :0x00f00000:0x00020000
Recovery Boot Env               :0x00f20000:0x00020000
Raw Data Table                  :0x00f40000:0x00020000
Info                            :0x00f60000:0x00020000
Id                              :0x00f80000:0x00020000
Reserved3                       :0x00fa0000:0x00060000
Boot Image                      :0x01000000:0x00100000
Waveform                        :0x01100000:0x00200000
LOG                             :0x01300000:0x00500000

2:59 PM

so sector 2048 == Normal Kernel:0x00100000:0x00400000 ? (edited)

3:03 PM

config.gz

12.13 KB

3:03 PM

here's the original kernel config

3:07 PM

If you compare this /proc/config.gz to the Kobo Touch A/B /proc/config.gz are they pretty close? (edited)

Check yourself

I don't own an A/B

Go in kernel repo -> kernel/config/config-n905b

did you modify that one?

Yes It's InkBox config

you can show me the diffrences of these files

3:13 PM

and I can explain

3:14 PM

there was a nice GUI for diff

3:14 PM

i dont remember the name

https://www.nxp.com/docs/en/reference-manual/IMX50RM.pdf PG# 2910 chart 6.5 Boot Devices (Internal Boot) 6.5.3 Expansion Device Support

md 0x53fd0004 1
53fd0004: d8200064    d. .
d8200064 = 32'b 1101 1000 0010 0000 0000 0000 0110 0100
echo "obase=2; ibase=16; D8200064" | bc
11011000001000000000000001100100
31=1 30=1 29=0 28=1 27=1 26=0 25=0 24=0 23=0 22=0 21=1 20=0 19=0 18=0 17=0 16=0 15=0 14=0 13=0 12=0 11=0 10=0 9=0 8=0 7=0 6=1 5=1 4=0 3=0 2=1 1=0 0=0
SRC_SBMR[25:24] = 0
BOOT_MODE[1:0]=0 => internal boot
eMMC normal boot
imx5 loads U-Boot from raw eMMC offset 2 sectors from the start (offset 1kiB from start of storage)

(edited)

3:39 PM

raw u-boot.imx dd if=/dev/mmcblkX bs=512 skip=2 count=1 | hexdump -vC

3:39 PM

imx header should be there.

3:43 PM

MX50_ARM2 U-Boot > md $loadaddr 0x80
00000000: e59ff01c e59ff01c e59ff01c e59ff01c    ................
00000010: e59ff01c e59ff01c e59ff01c e59ff01c    ................
00000020: e59ff01c 0000e0fc f801ffbc f801ffc0    ................
00000030: f801ffc4 f801ffc8 f801ffcc f801ffd0    ................
00000040: f801ffd4 f801ffd8 00000011 20296328    ............(c)
00000050: 79706f43 68676972 30322074 202c3031    Copyright 2010,
00000060: 65657246 6c616373 65532065 6f63696d    Freescale Semico
00000070: 6375646e 2c726f74 636e4920 6c41202e    nductor, Inc. Al
00000080: 6972206c 73746867 73657220 65767265    l rights reserve
00000090: 00002e64 402c00dd 00007bf9 00007d8b    d.....,@.{...}..
000000a0: 00007dc7 00008107 000080ab 00007fc9    .}..............
000000b0: 00007e71 0000822b 00008271 000082dd    q~..+...q.......
000000c0: 000028a5 00002aad 0000307f 000043c5    .(...*...0...C..
000000d0: 0000453d 000067e5 00007693 f8524a22    =E...g...v.."JR.
000000e0: 68000020 f00040c8 47700001 2301b510     ..h.@....pG...#
000000f0: 3c184c1d f854408b 68010020 4319b10a    .L.<.@T. ..h...C
00000100: 4399e000 bd106001 0594b5f0 4d16d516    ...C.`.........M
00000110: 3407f3c2 f3c23518 f8552681 b2d75020    ...4.5...&U. P..
00000120: c024f855 f04f00f6 fa0e0eff 40b7fe06    U.$...O........@
00000130: 0c0eea2c 070cea47 7024f845 d50b05d2    ,...G...E.$p....
00000140: 4c092201 408a3c30 0020f854 b10b6801    .".L0<.@T. ..h..
00000150: e0004311 60014391 4b03bdf0 f8533318    .C...C.`...K.3S.
00000160: f8400020 47702021 0000e4f4 68534aff     .@.! pG.....JSh
00000170: 0302f023 68136053 0320f023 68936013    #...S`.h#. ..`.h
00000180: 03fff023 4bfa6093 0380eb03 3021f853    #....`.K....S.!0
00000190: 4bf76093 eb033310 f8530380 61133021    .`.K.3....S.!0.a
000001a0: 33204bf3 0080eb03 0021f850 681060d0    .K 3....P.!..`.h
000001b0: 0030f040 68506010 0002f040 68106050    @.0..`Ph@...P`.h

3:43 PM

000001c0: d0fc07c0 4aeb4770 f0436813 60134340    ....pG.J.hC.@C.`
000001d0: f0436853 60534340 f0236813 6013033f    ShC.@CS`.h#.?..`
000001e0: f0236853 6053033f 28083aac e8dfd21e    Sh#.?.S`.:.(....
000001f0: 2304f000 8a70523d b171bca4 d0162901    ...#=Rp...q..)..
MX50_ARM2 U-Boot >

3:49 PM

Compile U-Boot

cd uboot
make ARCH=arm CROSS_COMPILE=arm-none-linux-gnueabi- -j8 mx50_arm2_config
make ARCH=arm CROSS_COMPILE=arm-none-linux-gnueabi- -j8
...
arm-none-linux-gnueabi/4.4.1/armv4t -lgcc -Map u-boot.map -o u-boot
arm-none-linux-gnueabi-objcopy -O srec u-boot u-boot.srec
arm-none-linux-gnueabi-objcopy --gap-fill=0xff -O binary u-boot u-boot.bin

(edited)

1

3:49 PM

The worst thing about discord is it doesn't seem to search code tags very well

3:50 PM

You have to remember what you said around that point to get back to it.

Reads raw eMMC at offset 0x800 (in 512B block units) and it reads 0x1400 such blocks that is the Kernel. Then the ramdisk 0x2800 0x258 then jumps to it with bootm. They read uImage which wraps the kernel zImage and then read uImage which wraps the ramdisk two blobs and then they point the bootm command to both blobs to start kernel and provide pointer to ramdisk to the kernel. (edited)

MX50_ARM2 U-Boot > printenv
stdin=serial
stdout=serial
stderr=serial
ethact=FEC0
loadaddr=0x70800000
loadaddr_ramdisk=0x70C00000
bootdev=2
rawtable=0xF40000
bootargs=root=/dev/mmcblk0p1 rootfstype=ext4 rw rootwait init=/linuxrc console=ttymxc4,115200 bootdev=0 rawtable=0xF40000
bootcmd=mmc read 0 ${loadaddr} 0x800 0x1400;mmc read 0 ${loadaddr_ramdisk} 0x2800 0x258;bootm ${loadaddr} ${loadaddr_ramdisk} bootdev=0

Environment size: 387/131068 bytes

yes?

Just putting the working Stock boot from SDCard bootargs/bootcmd here case I break it again

1

7:02 PM

Going to change init=/system/bin/sh to see what happens

1

please try our kernel with the ttymxc4 thing (edited)

ok

7:08 PM

one sec.

1

7:08 PM

so all that I got from doing the prior thing was "init: cannot open '/initlogo.rle'"

7:15 PM

custkern.txt

14.24 KB

7:15 PM

So it boots with those two lines now.

7:15 PM

but dies with what looks like power issues

YESSSSSSS

7:16 PM

Progress

7:17 PM

what's the cmdline in default system?

Sorry my brain is rusty what are you asking for?

cat /proc/cmdline

tux-linux

YESSSSSSS

BTW this means we're done with the watchdog

1

do you want me to switch cards back to stock firmware to run that?

yes

7:20 PM

sorry

root=/dev/mmcblk0p1 rootfstype=ext4 rw rootwait init=/linuxrc console=ttymxc4,115200 bootdev=0 rawtable=0xF40000

hmm okay, nothing interesting there it seems

What were you expecting to see there?

7:23 PM

bootargs=root=/dev/mmcblk0p1 rootfstype=ext4 rw rootwait init=/linuxrc console=ttymxc4,115200 bootdev=0 rawtable=0xF40000

sometimes bootargs are in u-boot and other ones are in the kernel for some reason

7:23 PM

it fails to power the epdc up

Looks like it fails to find the data it needs from the eMMC

7:59 PM

tps65180_display_enable fail to read vcom from eMMC. -1

Any ideas?

Nope Are you using the last kernel source tarball Sony provided?

where we write the kernel to is that within the first 26mb of the sdcard? (edited)

Yes

I don't know if I remember this right but I think someone said that is where it gets the vcom settings from and other things.

10:01 PM

so maybe we are overwriting?

Maybe But if it works with the stock firmware I have no idea why it is not there

10:04 PM

Hmm

10:05 PM

Restore first 26M to external SD and try to flash the kernel there again and boot it

Here is a comment in the forum not that the person knows directly what they are dealing with: "One thing you should never mess with is the first 26MB of the internal flash device. There is a bunch of raw data hidden in it that cannot easily be restored and there are some folks on the forum that already bricked their devices (irreversibly!!)."

10:12 PM

Is this the same with every ereader?

No

10:13 PM

But often yes

tux-linux

Restore first 26M to external SD and try to flash the kernel there again and boot it

Will try

10:24 PM

the only other thing I can think of is ramdisk and our kernel have to match

10:25 PM

The last message is actually this

10:25 PM

Warning: unable to open an initial console.
Freeing init memory: 728K
Kernel panic - not syncing: Attempted to kill init!

(edited)

10:27 PM

Thoughts?

10:29 PM

Maybe we should try to compile the kernel first without injecting busybox into it and see if it boots.

10:30 PM

Would the warning there be because it was trying to open busybox

NonaSuomy

The last message is actually this

The first warning is easy to fix But it's not what's causing the kernel panic IMO

10:37 PM

You need to add ttymxc4 to your initrd root/dev

Which first one?

10:53 PM

I just noticed this as well

Unpacking initramfs...
Initramfs unpacking failed: compression method gzip not configured

10:53 PM

maybe that is where the vcom data is

whooops

10:56 PM

but no, has nothing to do with it

10:57 PM

what's the general setup page of menuconfig looking like?

config

55.29 KB

11:02 PM

CONFIG_INITRAMFS_SOURCE="/home/nonasuomy/code/busybox/_install"

NonaSuomy

I just noticed this as well

Unpacking initramfs...
Initramfs unpacking failed: compression method gzip not configured

I guess that is why busybox won't load though

CONFIG_INITRAMFS_SOURCE="/home/nonasuomy/code/busybox/_install"

11:06 PM

Just a guess but because we set LZMA compression and turned GZIP off that we then need to set this? # CONFIG_INITRAMFS_COMPRESSION_LZMA is not set

11:07 PM

Though above that it does say CONFIG_INITRAMFS_COMPRESSION_NONE=y

11:08 PM

so it should not technically be uncompressed but we never copied an intramfs to the card so maybe it's just looking at the existing one or is intramfs inside uImage?

tux-linux

You need to add ttymxc4 to your initrd root/dev

How do you add ttymxc4 to your initrd root/dev ?

11:12 PM

Interesting their stock kernel with no modifications is

Image Name:   Linux-2.6.35.3
Created:      Fri May 13 22:56:50 2022
Image Type:   ARM Linux Kernel Image (uncompressed)
Data Size:    1669524 Bytes = 1630.39 KiB = 1.59 MiB
Load Address: 70008000
Entry Point:  70008000
  Image arch/arm/boot/uImage is ready

1.59MB instead of what is on the device (2.2MB)

11:17 PM

Is this because they also include busybox to the uImage?

maybe

NonaSuomy

How do you add ttymxc4 to your initrd root/dev ?

what's ls -shal /dev/ttymxc4 in the stock system?

I can't do any -shal after ls in adb shell

can you launch busybox telnetd?

11:30 PM

or ls -l maybe?

nothing - works

11:31 PM

only ls

try busybox telnetd then

crw------- root root 207, 20 2010-01-01 02:00 ttymxc4

ok, so in _install, do sudo mknod dev/ttymxc4 c 207 20

from the busybox _install folder I run sudo mknod /dev/ttymxc4 c 207 20

11:54 PM

?

no

11:54 PM

dev/ttymxc4

11:54 PM

without the first slash

11:55 PM

if dev folder is not existing, create it

11:55 PM

in short:

mknod dev/console c 5 1
mknod dev/null c 1 3
mknod dev/ttymxc4 c 207 20

~/code/busybox/_install/dev$ ls -l
total 0
crw-r--r-- 1 root root   5,  1 May 13 23:53 console
crw-r--r-- 1 root root   1,  3 May 13 23:54 null
crw-r--r-- 1 root root 207, 20 May 13 23:54 ttymxc4

11:59 PM

look ok?

12:01 AM

Note nothing is in _install any more for some reason should something have been there beside the dev we put there?

why is there nothing in _install anymore?

what should be in there?

your busybox initrd

I compiled busybox earlier would it have purged that directory for some reason?

yes maybe

12:04 AM

did you compile with make install?

I did this now and it's all back

make ARCH=arm CROSS_COMPILE=arm-none-linux-gnueabi- -j8 install
upx --ultra-brute _install/bin/busybox

(edited)

ok

dev was still there after doing that

you can recompile the kernel

ok

12:16 AM

hmm that doesn't look right

  UIMAGE  arch/arm/boot/uImage
Image Name:   Linux-2.6.35.3
Created:      Sat May 14 00:10:45 2022
Image Type:   ARM Linux Kernel Image (uncompressed)
Data Size:    1673520 Bytes = 1634.30 KiB = 1.60 MiB
Load Address: 70008000
Entry Point:  70008000
  Image arch/arm/boot/uImage is ready

I did enable that lzma compression on initrams though.

what's CONFIG_INITRAMFS_SOURCE?

Maybe we could have just done that instead of stripping everything out

12:17 AM

Config above

ok it's fine

12:18 AM

try it

My only change was CONFIG_INITRAMFS_COMPRESSION_LZMA=y

it's good

12:19 AM

the reason why I didn't want to enable it before was the watchdog

gotcha

1

12:25 AM

So essentially we should have left gzip support on? (edited)

12:28 AM

this is a slow sd card will be 17mins

why 17mins?

12:34 AM

no, lzma is fine with the watchdog problem solved

no I meant for the error in the console about it not finding gzip

oh, I don't know

12:35 AM

we will see

NonaSuomy

I just noticed this as well

Unpacking initramfs...
Initramfs unpacking failed: compression method gzip not configured

.

tux-linux

why 17mins?

Writing the full eMMC backup to the SD again just in case.

why the full backup?

12:40 AM

dd if=backup.img of=/dev/mmcblk0 bs=1M count=26 or something like that should work (don't remember the exact sector numbers)? (edited)

In case we wrote over the vcom data with our extra large uImage that one time. (edited)

12:43 AM

dd if=uImage of=/dev/mmcblk0 bs=512 seek=2048
sync

that is ok

log.txt

16.3 KB

it looks like it hasn't extracted the initramfs

12:57 AM

but you can't load the existing initramfs in u-boot on top of that

12:57 AM

remove it from bootcmd

bootargs=root=/dev/mmcblk0p1 rootfstype=ext4 rw rootwait init=/linuxrc console=ttymxc4,115200 bootdev=0 rawtable=0xF40000
bootcmd=mmc read 0 ${loadaddr} 0x800 0x1400;mmc read 0 ${loadaddr_ramdisk} 0x2800 0x258;bootm ${loadaddr} ${loadaddr_ramdisk} bootdev=0

bootm ${loadaddr} ${loadaddr_ramdisk} bootdev=0 remove that part

1:00 AM

then saveenv

so

setenv bootcmd 'mmc read 0 ${loadaddr} 0x800 0x1400;mmc read 0 ${loadaddr_ramdisk} 0x2800 0x258'
saveenv

(edited)

no

1:02 AM

setenv bootcmd mmc read 0 ${loadaddr} 0x800 0x1400\;mmc read 0 ${loadaddr_ramdisk} 0x2800 0x258 (edited)

good catch bad paste.

1:05 AM

hmm something weird can't trigger uboot command prompt

1:11 AM

got it

1:11 AM

Hit any key to stop autoboot:  0

MMC read: dev # 0, block # 2048, count 5120 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
SD set clock 50MHz
5120 blocks read: OK

MMC read: dev # 0, block # 10240, count 600 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
SD set clock 50MHz
600 blocks read: OK
MX50_ARM2 U-Boot >

(edited)

1:11 AM

no longer loads the kernel now

1:18 AM

any ideas?

1:20 AM

note we removed bootm but I tried to just run it and got the same error.

what's the env?

1:21 AM

printenv

MX50_ARM2 U-Boot > printenv
stdin=serial
stdout=serial
stderr=serial
ethact=FEC0
loadaddr=0x70800000
loadaddr_ramdisk=0x70C00000
bootdev=2
rawtable=0xF40000
bootargs=root=/dev/mmcblk0p1 rootfstype=ext4 rw rootwait init=/linuxrc console=ttymxc4,115200 bootdev=0 rawtable=0xF40000
bootcmd=mmc read 0 0x70800000 0x800 0x1400;mmc read 0 0x70C00000 0x2800 0x258

Environment size: 329/131068 bytes

1:23 AM

should init be something else related to busybox now?

Does that look ok?

How would we extract their kernel and decompress it to dissect some things?

NonaSuomy

Does that look ok?

No

9:27 AM

You need to add \;bootm (edited)

setenv bootcmd mmc read 0 ${loadaddr} 0x800 0x1400\;mmc read 0 ${loadaddr_ramdisk} 0x2800 0x258\;bootm

9:33 AM

Note I looked through the kernel source and all these log messages were added by Sony dev as to why we wouldn’t find them googling for them.

9:34 AM

If I wanted to pull their uImage out of my eMMC backup what dd skip etc command would I use?

maybe dd if=emmc.img of=kernel.bin bs=512 skip=2048 count=3072 but don't expect a lot (edited)

bootlog.txt

17.74 KB

go into kernel config

10:09 AM

Device drivers -> Block devices

ugh machine rebooted last night and I think it corrupted the vm will be a bit

10:15 AM

getting kernel panic on vm boot.

Ok wiped out a physical machine and installed debian 11 resetting up build env. to have less problems hopefully

toolchain/gcc-4.4.4-glibc-2.11.1-multilib-1.0.tar.gz at master · on...

Toolchain for Onyx devices. Contribute to onyx-intl/toolchain development by creating an account on GitHub.

1:46 PM

https://github.com/onyx-intl/toolchain/blob/master/imx508/sdk_imx508.tar.gz

toolchain/sdk_imx508.tar.gz at master · onyx-intl/toolchain

Toolchain for Onyx devices. Contribute to onyx-intl/toolchain development by creating an account on GitHub.

Did you succeed with menuconfig?

I didn't realize how much you have to do to get a compile going haha still working

1

3:29 PM

trying to get busybox to compile

3:29 PM

networking/libiproute/ipaddress.c:345: error: 'IFA_F_DADFAILED' undeclared (first use in this function)

are you using the same compiler? (arm-none-linux-gnueabi)?

yes

3:30 PM

make ARCH=arm CROSS_COMPILE=arm-none-linux-gnueabi- -j8 install

and same busybox tarball?

nope 4.4.4

busybox tarball

sorry

3:30 PM

yes same 1.4.1

NonaSuomy

nope 4.4.4

then it's not the same

NonaSuomy

yes same 1.4.1

then it's not the same either

3:31 PM

busybox 1.4.1 was released in 2007

sorry one moment

1

3:32 PM

my heads in the clouds.

3:32 PM

https://busybox.net/downloads/busybox-1.34.1.tar.bz2

ok that one is ok

did you get this error before Failed: -Wl,--start-group -lcrypt -lm -lresolv -lrt -Wl,--end-group

no

which version of busybox do you use?

1.32.1 in InkBox I want to update it soon (edited)

what version will you goto?

latest stable

so 1.34.1

tux-linux

1.32.1 in InkBox I want to update it soon (edited)

you can update qt libs too...

probably

5:41 PM

yeah although it will be more complicated

becouse?

compiling qt is way harder than compiling busybox

becouse?

also I need to recompile the platform plugins

and?

Szybet

becouse?

*because you never compiled qt from source?

I feel like I'm watching dude where's my car right now

tux-linux

*because you never compiled qt from source?

i did, and i did that no one else needs to

5:43 PM

https://github.com/Szybet/kobo-nia-audio/blob/main/apps-on-kobo/qt-setup.md

kobo-nia-audio/qt-setup.md at main · Szybet/kobo-nia-audio

Kobo nia ereader with audio added. Contribute to Szybet/kobo-nia-audio development by creating an account on GitHub.

5:43 PM

everything described

5:43 PM

maybe the glo platform plugin thing not

nice you never shared it

i did this 4 times

1

?...

then i forgot

5:44 PM

yestarday too

but you compiled 5.15.2

whats the problem with that

well now why do we need to update the libs?

becouse my compiled libs aren't working with yours

did you use the older toolchain?

i used koxtoolchain - kobo (edited)

5:46 PM

self generated

ok then you need to switch to an older commit

why

to make your libs compatible

we were talking about updating the libs

5:47 PM

not about switching to older commits

if you generated this toolchain before may 15 2021 it would have worked (it was building gcc 7.5.0 at the time) https://github.com/koreader/koxtoolchain/commit/688c830b91b3944fca3031ff447044ce7477c47a (edited)

Resync ref scripts · koreader/koxtoolchain@688c830

5:48 PM

but they could benefit from an update

tux-linux

if you generated this toolchain before may 15 2021 it would have worked (it was building gcc 7.5.0 at the time) https://github.com/koreader/koxtoolchain/commit/688c830b91b3944fca3031ff447044ce7477c47a (edited)

what didn't worked

I mean, your binary would have been compatible with the existing libs

Szybet

we were talking about updating the libs

here

5:49 PM

i know but it would be cool if i didn't need to do that

well, if you want to update them now, clone https://github.com/Kobo-InkBox/gui-bundle and update the libs in content/qt

GitHub - Kobo-InkBox/gui-bundle: InkBox GUI bundle contents and gen...

InkBox GUI bundle contents and generation script. Contribute to Kobo-InkBox/gui-bundle development by creating an account on GitHub.

it would break thing for you propably

why?

your qt dev setup

no, if it's the same directory structure, it should work

5:52 PM

qt usually installs itself in ~/qt-bin

5:52 PM

then you can just move out the required libs to update them

without the updated toolchain, can you use this qt lib?

the toolchain has nothing to do with it

5:54 PM

did what you compile with newer koxtoolchain work on the kobo?

yes of course

then, why wouldn't it work with the updated libs

becouse it sounds to easy

5:56 PM

it gets suspicious

NonaSuomy

networking/libiproute/ipaddress.c:345: error: 'IFA_F_DADFAILED' undeclared (first use in this function)

In case you deal with this updating to 1.34.1

No idea how to deal with this broken things Failed lcrypt lm lresolv lrt

7:59 PM

lcrypt apparently is now libxcrypt or something.

Wonder if unstable fixed all these issues

it would be unlikely

8:24 PM

the problem is probably your debian system

8:25 PM

for some reason

This is a fresh build and this issue is listed all over the place.

oh ok

I just think its far too complex for me to patch it.

then something went really wrong on their side

8:28 PM

you can just use 1.32.1 meanwhile

I think your compressed config probably excluded all the broken things

8:30 PM

people say that busybox is almost uncomplilable with out musl

8:31 PM

in its current state.

maybe you can try with a newer compiler

8:31 PM

devices are more tolerant to them when it's not kernel things

8:32 PM

arm-nickel-linux-gnueabihf- for example

8:32 PM

or armv7l-linux-musleabihf-

8:32 PM

all there in [kernel repo]/toolchain

just to note nothing fixed in 1.35.0

yay...

was I supposed to apply some imx5 config to busybox?

9:02 PM

this is all that I'm doing using the same 4.4.1 as you sent last time

9:03 PM

Add config https://discordapp.com/channels/809205711778480158/966554812567130152/970116956122415154

make ARCH=arm CROSS_COMPILE=arm-none-linux-gnueabi- menuconfig
Make sure this is on [*] Build static binary (no shared libs)

make ARCH=arm CROSS_COMPILE=arm-none-linux-gnueabi- -j$(nproc)

make ARCH=arm CROSS_COMPILE=arm-none-linux-gnueabi- -j$(nproc) install

upx --ultra-brute _install/bin/busybox

(edited)

I'm going with fresh system is missing something no idea what

9:11 PM

I tried your shrunken config no go

9:11 PM

tried the build before that one no go

9:12 PM

Going to try 1.32.1 and it is going to do the same thing and I'm going to go crazy

I can build it for you at this point

echo $PATH /usr/local/bin:/usr/bin:/bin:/usr/local/games:/usr/games:/snap/bin:/home/nonasuomy/code/arm-2009q3-67-arm-none-linux-gnueabi-i686-pc-linux-gnu/arm-2009q3/bin

9:20 PM

how would that work?

9:22 PM

like you build busybox then I dump it in _install and then compile the kernel or more to it then that?

9:28 PM

yup this system is defo missing something went back to 1.32.1 and still same error

9:32 PM

I just noticed we were using 4.8 gcc to compile this first

9:32 PM

maybe it has the issues fixed

9:32 PM

then we were just always using the same build files after that and never cleaning the directory

9:33 PM

maybe that is the key

9:34 PM

If you make a clean directory of 1.34.1 busybox and compile it with 4.4.1 does it work?

Not a GCC issue

9:35 PM

Kernel (headers) are too old

?

Which is annoying, true, but easily remedied for simple constants like those

9:37 PM

c.f., https://github.com/koreader/koxtoolchain/blob/97e4dcd6e9e20c90062038dc6e2fb18adc0da40d/refs/x-compile.sh#L1944

koxtoolchain/x-compile.sh at 97e4dcd6e9e20c90062038dc6e2fb18adc0da4...

Build script for various toolchains used by KOReader - koxtoolchain/x-compile.sh at 97e4dcd6e9e20c90062038dc6e2fb18adc0da40d · koreader/koxtoolchain

9:37 PM

(Not the same exact issue, but close enough)

How do you fix it?

We may not actually be talking about the same thing. I'm talking about the ipaddress applet build issues on old kernel headers

9:40 PM

I have no idea what else you may be talking about since you haven't posted any build logs

NonaSuomy

did you get this error before Failed: -Wl,--start-group -lcrypt -lm -lresolv -lrt -Wl,--end-group

.

That's not a build log

9:42 PM

(verbose, ideally)

9:42 PM

(pass V=1 to make)

https://pastebin.com/raw/HtSeAFDn

9:44 PM

I'll do that next but for now here it is

Actual error being sync.c:(.text.sync_main+0x5c): undefined reference to syncfs'`

9:45 PM

Because syncfs() first appeared in Linux 2.6.39; library support was added to glibc in version 2.14.

9:45 PM

So sayeth man syncfs

Someone said to turn sync off to fix that

Stuff.. kinda relies on the sync applet being there.

but then you no longer have sync

9:47 PM

So I didn't really know what to do after that

//config:config FEATURE_SYNC_FANCY
//config:    bool "Enable -d and -f flags (requires syncfs(2) in libc)"

9:48 PM

Disable that in your bb config

[ ] Enable -d and -f flags (requires syncfs(2) in libc)

Something else is toggling FSYNC support, then

9:52 PM

Don't have a bb source tree checked out rn

9:52 PM

ripgrep is your friend

Just strolls in and fixes all my problems

9:53 PM

I struggled with that the whole day.

1

Misread the source actually, ENABLE_FSYNC does what it says on the tin, unrelated

9:53 PM

So nevermind that last bit ;p

9:53 PM

ripgrep is still your friend, though ;p (edited)

I usually just use / in menuconfig and the help button to make sure I'm at the right one haha.

9:55 PM

slower but works

I meant for build issues in general

9:57 PM

The search key is indeed your other friend when looking for a setting by its flag name in kconfig

oh so how do you use ripgrep?

9:57 PM

to solve build issues?

It's grep, but faster ;p

Doesn't rgrep do the same thing (or grep -r)? (edited)

So grep the offending function/constant/whatever to locate it in the sources, then look at the context around that piece of code (edited)

9:58 PM

It's a from-scratch grep implementation in rust

Ah

9:59 PM

Nice

9:59 PM

@NonaSuomy I will compile this damn busybox for you

no its ok

9:59 PM

they fixed it for me

9:59 PM

@NiLuJe

Ok, I misunderstood then

9:59 PM

Good

10:00 PM

Thanks NiLuJe

1

Years off my life right there.

That's cross-compilation for you.

1

10:02 PM

The script I linked to above is full of my ranting about that.

2

I'm sure it does amazing things as I was confused just looking at it all

I've given up on cross compilation on my side; I just use an arm vm

10:03 PM

No matter the time

Not really, it just builds the stuff I ship in my various packages

It will always be less than the time required to set up the cross compile

Stuff is slightly less painful the newer the target kernel & libc are

10:04 PM

And the less cmake is involved ;p

@NiLuJe at any time did anyone on the koreader team get a build of prstux by chance or was that dude that built it the only one that ever had it working?

10:05 PM

(talking about the os image)

He was probably the only user, yeah

1

10:06 PM

Might have been some other guy, but I would count the target audience on the fingers of one hand at best

I would like to find anyone at all. The guy that made it was zero help (edited)

10:06 PM

said he forgot everything and said he wouldn't help.

Check the PRs & issues that touch the frontend/device/sony-prstux path (edited)

the code he has online is not complete.

10:07 PM

missing kernel and other vital things.

10:08 PM

would love to get that running just to poke him with it haha.

10:09 PM

I only found a few comments where he was talking with you or maybe one of the build maintainers but nobody that actually was using it.

10:10 PM

and all he said to them was that he has no time to fix things. haha

10:10 PM

I sense a theme...

NonaSuomy

In case you deal with this updating to 1.34.1

@NiLuJe the ip problem you were talking about or this is the same fix? (edited)

10:17 PM

I don't remember doing this before but it did work then to get the next error

10:19 PM

(wondering if there is a better just turn xyz off than modifying the source)

10:21 PM

I got that fix from here if you want to see the full error https://blog.csdn.net/qq_16126823/article/details/121891466 Just to note google translate brakes the page and sends you to their homepage or something weird. (edited)

制作可以从SD卡加载文件系统的流程_qq_16126823的博客-CSDN博客_sd卡分区

环境操作平台：UBuntu18.04 64位目标平台：Linux 2.6.35.7 ARM Cortex-A(arm v7)编译工具：arm-2009q3准备工作下载busybox-1.34.1.tar.bz2源码https://busybox.net/downloads/一张SD卡一、制作根文件系统1、解压busybox-1.34.1.tar.bz2源，并进入到源码中tar -jxvf busybox-1.34.1.tar.bz2cd busybox-1.34.12、修改Makef

NonaSuomy

the code he has online is not complete.

Maybe there is something to look at in internet archive (edited)

tux-linux

Maybe there is something to look at in internet archive (edited)

I believe he got it working for himself then he attempted to put an automated build process together for everyone else then gave up halfway through.

10:24 PM

its basically just missing the kernel and the ramdisk ramfs stuff I think.

hmm

Hopefully we get InkBox running then we will understand more about the process then hopefully it will be a breeze right...

yep

NonaSuomy

its basically just missing the kernel and the ramdisk ramfs stuff I think.

There may even be some tips there to to getting stuff supported in InkBox as well that he already figured out. I noticed a few scripts that had some things in it about hardware management.

10:32 PM

Sleep functions and such

In any case, this device earns the no. 1 price for the most complicated, convoluted, nonsensical, and painful port ever (edited)

I pretty sure you just spelled S O N Y

Even porting to Kindle is easier... like... something's wrong

(edited)

complicationS cOnvoluted Nonsensical painfullY ported.

n o i c e

10:35 PM

n o y s

haha

10:37 PM

Now will the kernel compile

I totally forget what we were going to do next but

  UIMAGE  arch/arm/boot/uImage
Image Name:   Linux-2.6.35.3
Created:      Sat May 14 22:56:43 2022
Image Type:   ARM Linux Kernel Image (uncompressed)
Data Size:    2271056 Bytes = 2217.83 KiB = 2.17 MiB
Load Address: 70008000
Entry Point:  70008000
  Image arch/arm/boot/uImage is ready

2

just try to run it

I'm going to lurk up a bit as I think you told me to try something first.

tux-linux

Device drivers -> Block devices

^

oh yes that

tux-linux

vi ~/code/busybox/_install/etc/init.d/rcS create etc/init.d if not existing populate rcS with this content:

#!/bin/sh

mkdir -p /sys /proc /dev
mount -t sysfs sysfs /sys
mount -t proc proc /proc
mount -t devtmpfs devtmpfs /dev

/bin/sh

then recompile the kernel (edited)

Is this still valid?

yes, if it has been erased

(fresh os on fresh system)

yes

My VM got corrupted last night

yeah, so do it

kk

11:23 PM

When I add that stuff now that I already compiled "make install" can I just do that again and it will included those additions or have to clean (edited)

?

oh does the kernel compile process grab that _install folder and it has nothing to do with the busybox make install

it just grabs the folder

11:25 PM

it doesn't care about what's outside it

11:25 PM

don't forget to create device nodes

just did them

1

11:26 PM

so just make ARCH=arm CROSS_COMPILE=arm-none-linux-gnueabi- -j$(nproc) uImage

11:26 PM

and all that stuff will be added in

yes

Just saw these fly by on the compile console

  CC      drivers/regulator/tps65180-regulator.o
  CC      net/socket.o
drivers/regulator/tps65180-regulator.c: In function 'tps65180_vcom_get_voltage':
drivers/regulator/tps65180-regulator.c:838: warning: 'vcomValue' may be used uninitialized in this function

hmm hmm

11:32 PM

GREAT job sony

Maybe that will help find our issue location later

11:34 PM

Never ran UPX but were pretty close to the stock uImage size now. Data Size: 2291748 Bytes = 2238.04 KiB = 2.19 MiB

tux-linux

maybe dd if=emmc.img of=kernel.bin bs=512 skip=2048 count=3072 but don't expect a lot (edited)

When I ran this against the eMMC backup I got this with mkimage:

~/code$ mkimage -l kernel.bin 
GP Header: Size 27051956 LoadAddr c62947b6

when I run it against a uImage that we compiled I get:

~/code$ mkimage -l uImage
Image Name:   Linux-2.6.35.3
Created:      Sat May 14 00:10:45 2022
Image Type:   ARM Linux Kernel Image (uncompressed)
Data Size:    1673520 Bytes = 1634.30 KiB = 1.60 MiB
Load Address: 70008000
Entry Point:  70008000

(edited)

11:39 PM

Why do they not output the same information?

because kernel.bin probably contains other things

11:40 PM

and c62947b6 as a load address makes absolutely no sense

so probably not the right dd values?

yeah

11:42 PM

i don't know

Is there a safer way to insert the uImage into eMMC backup where the size of the uImage wouldn't matter?

not really

11:46 PM

not to my knowledge

basically like using 7zip manager to extract the current one then toss the other one in like a compressed file

11:52 PM

When I open this bin I see 10 more img files of the partitions the space we are working with is not one of these 10 correct but the raw space that we can't view as a file but is part of the start of the bin data correct? https://jpa.kapsi.fi/stuff/other/sd_card_PRS-T1_1.0.07_adb_enabled.bin.xz

it should

we should be able to see the 25-26MB raw data inside the bin? (edited)

11:55 PM

so it's one of those 10?

yes

hmm

use binwalk on the image

When I looked at them all with 7zip they all seemed very trivial like fonts or demo books or windows drivers, I'll have to take another look

11:57 PM

ok

BinWalkeMMC.txt

14.83 KB

12:05 AM

Think it crashed at the end but that's what it spit out

the uImage headers are important

12:05 AM

strange that there are three

So does that data line up with these locations?

# cat /sys/module/rawdatatable/parameters/rawdata_param
MBR                             :0x00000000:0x00000400
uBoot                           :0x00000400:0x000bfc00
Boot Env                        :0x000c0000:0x00020000
Reserved1                       :0x000e0000:0x00020000
Normal Kernel                   :0x00100000:0x00400000
Normal Rootfs                   :0x00500000:0x00100000
Recovery Kernel                 :0x00600000:0x00400000
Reserved2                       :0x00a00000:0x00500000
Normal Boot Env                 :0x00f00000:0x00020000
Recovery Boot Env               :0x00f20000:0x00020000
Raw Data Table                  :0x00f40000:0x00020000
Info                            :0x00f60000:0x00020000
Id                              :0x00f80000:0x00020000
Reserved3                       :0x00fa0000:0x00060000
Boot Image                      :0x01000000:0x00100000
Waveform                        :0x01100000:0x00200000
LOG                             :0x01300000:0x00500000

NonaSuomy

When I ran this against the eMMC backup I got this with mkimage:

~/code$ mkimage -l kernel.bin 
GP Header: Size 27051956 LoadAddr c62947b6

when I run it against a uImage that we compiled I get:

~/code$ mkimage -l uImage
Image Name:   Linux-2.6.35.3
Created:      Sat May 14 00:10:45 2022
Image Type:   ARM Linux Kernel Image (uncompressed)
Data Size:    1673520 Bytes = 1634.30 KiB = 1.60 MiB
Load Address: 70008000
Entry Point:  70008000

(edited)

This numbers familiar header CRC: 0xC62947B6

it looks strange

12:12 AM

can you just boot the kernel you compiled?

boot1000.txt

17.23 KB

chmod +x etc/init.d/rcS there has been some progress

12:27 AM

the initrd starts

Will that fix the tty repeating thing as well?

maybe

oh my there is actually a prompt under all those tty spams

12:29 AM

can't open /dev/tty2: No such file or directory ls bin dev etc linuxrc sbin usr

cool

one moment while I recompile

12:31 AM

or could I have technically did that from that prompt

No change

12:40 AM

maybe it didn't sync properly -rw-r--r-- 1 1000 1000 128 May 15 2022 rcS

it's impossible

I did it from the console though maybe it will keep?

12:41 AM

-rwxr-xr-x 1 1000 1000 128 May 15 2022 rcS

12:42 AM

or is it a read only file system?

12:43 AM

didn't keep

12:44 AM

Does it cache those files in the kernel folder?

12:46 AM

should I chown it to root root as well?

12:46 AM

instead of my username

12:47 AM

~/code/busybox/_install/etc/init.d$ ls -l
total 4
-rwxr-xr-x 1 nonasuomy nonasuomy 128 May 14 23:19 rcS

NonaSuomy

I did it from the console though maybe it will keep?

It will never keep anything

NonaSuomy

~/code/busybox/_install/etc/init.d$ ls -l
total 4
-rwxr-xr-x 1 nonasuomy nonasuomy 128 May 14 23:19 rcS

Doesn't matter

ok I will try to flash again then

1

12:53 AM

I defo did sudo sync and sync just for good measure though.

1

12:59 AM

I guess I just maybe pulled the card a little quick last time

1:00 AM

tps65180_display_enable fail to read vcom from eMMC. -1
------------[ cut here ]------------
WARNING: at drivers/regulator/core.c:1423 _regulator_disable+0x30/0x118()
unbalanced disables for DISPLAY
Modules linked in:
[<800c066c>] (unwind_backtrace+0x0/0xec) from [<800e0ca8>] (warn_slowpath_common+0x4c/0x64)
[<800e0ca8>] (warn_slowpath_common+0x4c/0x64) from [<800e0d40>] (warn_slowpath_fmt+0x2c/0x3c)
[<800e0d40>] (warn_slowpath_fmt+0x2c/0x3c) from [<8025017c>] (_regulator_disable+0x30/0x118)
[<8025017c>] (_regulator_disable+0x30/0x118) from [<8025033c>] (regulator_disable+0x1c/0x30)
[<8025033c>] (regulator_disable+0x1c/0x30) from [<802bfdb4>] (get_temperature+0x40/0x50)
[<802bfdb4>] (get_temperature+0x40/0x50) from [<802bfdd0>] (calculate_battery_capacity+0xc/0x8c)
[<802bfdd0>] (calculate_battery_capacity+0xc/0x8c) from [<803aa2f8>] (sub_cpu_bci_probe+0x2b4/0x36c)
[<803aa2f8>] (sub_cpu_bci_probe+0x2b4/0x36c) from [<80279908>] (platform_drv_probe+0x18/0x1c)
[<80279908>] (platform_drv_probe+0x18/0x1c) from [<80278aa8>] (driver_probe_device+0xc8/0x184)
[<80278aa8>] (driver_probe_device+0xc8/0x184) from [<80278bc4>] (__driver_attach+0x60/0x84)
[<80278bc4>] (__driver_attach+0x60/0x84) from [<802782c4>] (bus_for_each_dev+0x44/0x74)
(kernel_init+0xa8/0x168)
[<80008408>] (kernel_init+0xa8/0x168) from [<800bc9c0>] (kernel_thread_exit+0x0/0x8)
---[ end trace 37afc84e26cca609 ]---
Freeing init memory: 716K
mount: mounting devtmpfs on /dev failed: No such device
/bin/sh: can't access tty; job control turned off
/ #

good at. last. (edited)

and yes it cured the tty repeating

1:00 AM

what is devtmpfs and why did it fail?

1:02 AM

and /bin/sh

what is in device drivers->generic driver options?

NonaSuomy

and /bin/sh

not to worry about at all

config

53.3 KB

you need to enable devtmpfs then

1:08 AM

but not the automount devtmpfs at dev thing

1:08 AM

yes

Anything else?

1:12 AM

or compile

compile

was I supposed to remove something from rcS?

no

ok

1:17 AM

---[ end trace f3fb3a274bc2f334 ]---
Freeing init memory: 716K
/bin/sh: can't access tty; job control turned off
/ #

Good That means we can start InkBox port

1

1:22 AM

Please make clean your kernel source and create a tarball of it. Then share it with me

tux-linux

Good That means we can start InkBox port

Well, after vcom issue if fixed

so just run make clean in the kernel source directory and send you a tar ball of it even without the vcom issue fixed?

Maybe fix the vcom issue before yeah

1:27 AM

But make a tarball for yourself if you mess things up

haha ok

With the .config

what is involved in the inkbox port process?

A lot of things

I would like to understand all the steps

Look in the Libra H2O porting thread

ok is there a specific point to start at?

The beginning

ok

Just hit home until you reach the top

make ARCH=arm CROSS_COMPILE=arm-none-linux-gnueabi- -j$(nproc) modules
  CHK     include/linux/version.h
  CHK     include/generated/utsrelease.h
  CC      scripts/mod/empty.o
  MKELF   scripts/mod/elfconfig.h
  HOSTCC  scripts/mod/file2alias.o
make[1]: 'include/generated/mach-types.h' is up to date.
  HOSTCC  scripts/mod/modpost.o
  HOSTCC  scripts/mod/sumversion.o
  CC      kernel/bounds.s
  GEN     include/generated/bounds.h
  CC      arch/arm/kernel/asm-offsets.s
  GEN     include/generated/asm-offsets.h
  CALL    scripts/checksyscalls.sh
  HOSTLD  scripts/mod/modpost
  CC [M]  drivers/usb/gadget/arcotg_udc.o
  CC [M]  drivers/usb/host/ehci-hcd.o
  CC [M]  drivers/usb/gadget/serial.o
drivers/usb/gadget/arcotg_udc.c:686: warning: 'udc_wait_b_session_low' defined but not used
drivers/usb/gadget/arcotg_udc.c:2423: warning: 'wake_up_irq' defined but not used
drivers/usb/host/ehci-hub.c:109: warning: 'ehci_adjust_port_wakeup_flags' defined but not used
  CC [M]  drivers/usb/serial/usb-serial.o
  CC [M]  drivers/usb/serial/generic.o
  CC [M]  drivers/usb/serial/bus.o
  LD [M]  drivers/usb/gadget/g_serial.o
  LD [M]  drivers/usb/serial/usbserial.o
  Building modules, stage 2.
  MODPOST 4 modules
WARNING: modpost: Found 3 section mismatch(es).
To see full details build your kernel with:
'make CONFIG_DEBUG_SECTION_MISMATCH=y'
  CC      drivers/usb/gadget/arcotg_udc.mod.o
  CC      drivers/usb/gadget/g_serial.mod.o
  CC      drivers/usb/host/ehci-hcd.mod.o
  CC      drivers/usb/serial/usbserial.mod.o
  LD [M]  drivers/usb/host/ehci-hcd.ko
  LD [M]  drivers/usb/gadget/g_serial.ko
  LD [M]  drivers/usb/serial/usbserial.ko
  LD [M]  drivers/usb/gadget/arcotg_udc.ko

But does it compile or stops?

Compiles as far as I know what I am doing which is little.

3:17 AM

I'm not sure if doing that includes them in the uImage or you have to manually place them on the sdcard some place

warnings are fine

3:41 AM

1

I found GCC 4.4.0 (original stock compile version) for fun as well https://microsoft.github.io/mu/dyn/mu_tiano_plus/CryptoPkg/Library/OpensslLib/openssl/boringssl/third_party/android-cmake/ndk_links/ In package http://dl.google.com/android/ndk/android-ndk-r5c-darwin-x86.tar.bz2

ndk links - Project Mu

Project Mu Documentation

3:45 AM

basically every version is there (edited)

NonaSuomy

Compiles as far as I know what I am doing which is little.

You have to make modules_install with INSTALL_MOD_PATH

8:07 AM

Then make a tarball or a squashfs and put it in your initrd root

8:08 AM

That might tank the watchdog though

tux-linux

You have to make modules_install with INSTALL_MOD_PATH

I think that's exactly what I was looking for as I did that but it tried to install it to my local debian

make ARCH=arm CROSS_COMPILE=arm-none-linux-gnueabi- -j$(nproc) EXTRA_CFLAGS=-fno-pic INSTALL_MOD_PATH=sonymodules modules

make ARCH=arm CROSS_COMPILE=arm-none-linux-gnueabi- -j$(nproc) uImage EXTRA_CFLAGS=-fno-pic INSTALL_MOD_PATH=sonymodules modules_install

(edited)

tux-linux

Then make a tarball or a squashfs and put it in your initrd root

How do I do this process? Have you done it with the rest of the different devices? I would like to do it as close as plausible to how you did it for the others.

11:43 AM

I have this sonymodules directory now

11:45 AM

sonymodules/
└── lib
    └── modules
        └── 2.6.35.3
            ├── build -> /home/nonasuomy/code/linux-2.6.35.2
            ├── kernel
            │   └── drivers
            │       └── usb
            │           ├── gadget
            │           │   ├── arcotg_udc.ko
            │           │   └── g_serial.ko
            │           ├── host
            │           │   └── ehci-hcd.ko
            │           └── serial
            │               └── usbserial.ko
            ├── modules.alias
            ├── modules.alias.bin
            ├── modules.builtin
            ├── modules.builtin.alias.bin
            ├── modules.builtin.bin
            ├── modules.dep
            ├── modules.dep.bin
            ├── modules.devname
            ├── modules.order
            ├── modules.softdep
            ├── modules.symbols
            ├── modules.symbols.bin
            └── source -> /home/nonasuomy/code/linux-2.6.35.2

Put it somewhere in your initrd

./tux-linux --verbose

Mv modules folder to _install

modules folder as in the one I named sonymodules or the sub folder in the tree above named modules?

12:00 PM

~/code/busybox/_install/sonymodules ?

12:03 PM

~/code/busybox/_install$ ls
bin  dev  etc  linuxrc  sbin  sonymodules  usr

That is fine

You are correct watchdog triggers now (Not sure if the image is too large but not getting that too large issue just resets)

1:45 PM

2.23MB

1:46 PM

Also I found an interesting directory on the eMMC directory two sh scripts called setvcom and getvcom

1:51 PM

vcom_get.sh

#!/bin/sh

vcom_uV=`epd_fb_test getvcom | awk '{ print $4 }'`
vcom_mV=`expr $vcom_uV / 1000`
echo ${vcom_mV}mV

1:52 PM

vcom_set.sh

#!/bin/sh


if [ $# -ne 1 ] ; then
    echo "Usage : vcom_set.sh vcom_value|off"
    exit 1
fi



case $1 in
[Oo][Ff][Ff])
    voltage.sh off epd
    echo "Power off OK"
    exit
    ;;
[0-9]*)
    vcom_mV=$1
    vcom_mV=`expr $vcom_mV / 10`
    vcom_mV=`expr $vcom_mV \* 10`
    vcom_uV=`expr $vcom_mV \* 1000`

    echo "setVCOM ${vcom_mV}mV"
    echo "VCOM write to eMMC"

    epd_fb_test setvcom $vcom_uV > /dev/null
    if [ $? -ne 0 ]; then
        echo "NG : fail set vcom."
        exit
    fi

    voltage.sh on epd
    echo "Power on OK"
    ;;
*)
    echo "NG : parameter error $1"
    exit 1
    ;;
esac
exit 0

1:55 PM

voltage.sh

#!/bin/sh




# 機能ブロックの全リスト
blocks_all="sd wan wifi epd"
power_impl=/proc/pwr3
power_block=phoenix
if [ -f $power_impl ]; then
  if [ $(cat $power_impl) -ne 0 ]; then
    power_block=wolfson
  fi
fi
blocks_all="$blocks_all $power_block"




# 引数なしは NGとする
if [ -z "$1" ]; then
  echo NG
  exit 1
fi



#
# 小文字変換関数
#  $1    変換対象
#
tolower()
{
  echo $1 | tr '[A-Z]' '[a-z]'
}



#
# Regulator制御関数
#  $1    on/off
#  $2... Regulatorの名称
#
regulator_control()
{
  # 引数なしはエラーとする
  if [ -z "$*" ]; then
    return 1
  fi

  # on/offで指定文字列を設定する
  case $1 in
  on)
    control=enabled
    ;;
  off)
    control=disabled
    ;;
  *)
    return 1
    ;;
  esac

  shift
  remain=$*

  # 処理対象の Regulatorを検索する
  for regulator in $*; do

    # Regulator名称を参照する
    for name in $(ls /sys/class/regulator/regulator.*/name); do
      if [ $regulator != $(cat $name) ]; then
        # 対象外だったため次を参照する
        continue
      fi

      # 対象の Regulatorへ値を書き込む
      node=$(echo $name | sed -e 's/name$/state/')
      echo $control > $node

      # 処理が済んだ Regulatorを外す
      remain=$(echo $remain | sed -e "s/$regulator//")
      break
    done

    # すべての Regulatorへ値を書き込んだか
    if [ -z "$remain" ]; then
      # すべて書き込んだ
      break
    fi

  done

  return 0
}




diag_sd()
{
  case $1 in
  on)
    gpio 5 12 0
    ;;
  off)
    gpio 5 12 1
    ;;
  *)
    return 1
    ;;
  esac

  return 0
}



diag_wan()
{
  case $1 in
  on)
    gpio 4 3 1
    ;;
  off)
    gpio 4 3 0
    ;;
  *)
    return 1
    ;;
  esac

  return 0
}



diag_wifi()
{
  case $1 in
  on)
    gpio 5 25 1
    ;;
  off)
    gpio 5 25 0
    ;;
  *)
    return 1
    ;;
  esac

  return 0
}



diag_epd()
{
  case $1 in
  on)
    regulators="$regulators DISPLAY V3P3_CTRL PWR0_CTRL VCOM"
    ;;
  off)
    regulators="$regulators VCOM PWR0_CTRL V3P3_CTRL DISPLAY"
    ;;
  *)
    return 1
    ;;
  esac

  return 0
}

1:55 PM

diag_wolfson()
{
  case $1 in
  on)
    ldo=1
    while [ $ldo -le 11 ]; do
      regulators="$regulators LDO$ldo"
      let ldo=$ldo+1
    done
    ;;
  off)
    ;;
  *)
    return 1
    ;;
  esac

  return 0
}



diag_phoenix()
{
  case $1 in
  on)
    regulators="$regulators LDO1 LDO2 LDO3 LDO4 LDO5 LDO6 LDO7 VANA LDOLN LDOUSB SMPS3 SMPS4 VIO"
    ;;
  off)
    ;;
  *)
    return 1
    ;;
  esac

  return 0
}




if [ -z "$2" -o "$(tolower $2)" == "all" ]; then
  blocks=$blocks_all
elif echo $blocks_all | grep "\<$2\>" 2>&1 > /dev/null; then
  blocks=$2
else
  echo NG
  exit 1
fi



for block in $blocks; do
  if ! eval diag_$block $(tolower $1); then
    break
  fi
done

regulator_control $1 $regulators



if [ $? -eq 0 ]; then
  echo OK
else
  echo NG
fi

StockLogWithoutRamDisk.txt

13.29 KB

@tux-linux do these numbers mean anything to you across any of the other build for vcom?

BW1120617204466                 BU1125868900232                 148427501482734                 PVI6inchE120C                   
65                  
1990
AudioOK                         2011/11/14-7:53:48

no?

7:51 PM

where did you get ithis?

NonaSuomy

# remount SD card for writing
mount -o remount,rw /initrd/mnt/sd

# your screen driver
dd if=/dev/mmcblk2 of=/initrd/mnt/sd/WF.img skip=17408 count=2048 bs=1024
sync

# other important information, such as the operating voltage of the screen in line 5
dd if=/dev/mmcblk2 of=/initrd/mnt/sd/Vcom.img skip=15876 count=1 bs=1024
sync

# serial, MAC
dd if=/dev/mmcblk2 of=/initrd/mnt/sd/SN.img skip=15872 count=1 bs=1024
sync

Partitions

Disk /dev/sdb: 1.82 GiB, 1958739968 bytes, 3825664 sectors 
Disk model: PRS-XXX          
Units: 1 * 512 sectors = 512 bytes 
Sector size (logical/physical): 512 bytes / 512 bytes 
I/O size ( minimum/optimal): 512 bytes / 512 bytes 
Disk Label Type: dos 
Disk ID: 0x00000000 

Device Boot Start End Sectors Size Id Type 
/dev/sdb1 51200 71679 20480 10M 83 Linux 
/dev/sdb2 71680 92159 20480 10M 83 Linux 
/dev/sdb3 92160 866687 774528 378.2M 5 Extended 
/dev/sdb4 866688 3825663 2958976 1.4G 83 Linux
/dev/sdb5 92176 124991 32816 16M 83 Linux 
/dev/sdb6 125008 403583 278576 136M 83 Linux 
/dev/sdb7 403600 424127 20528 10M 83 Linux 
/dev/sdb8 424144 506111 81968 40M 83 Linux 
/dev/sdb9 506128 604479 98352 48M 83 Linux 
/dev/sdb10 604496 866687 262192 128M 83 Linux 

Partition table entries are not ordered as they are on disk.

(edited)

Here

7:54 PM

Except I tried to hit the eMMC backup for it so may have not been the right spot

7:56 PM

What do the other builds look like for how they store VCOM?

8:03 PM

https://4pda.to/forum/index.php?s=&showtopic=301887&view=findpost&p=88936369

Sony PRS-T1 - прошивка (US-версия) - 4PDA

Sony PRS-T1 - прошивка (US-версия), Русификация и установка дополнительных приложений (amutin)

8:05 PM

#!/bin/sh 

PATH=/diag/bin:/bin:/sbin:/usr/bin:/usr/sbin:/usr/local/bin 

mount -t vfat -o rw,shortname=winnt /dev/mmcblk0p1 /initrd/mnt/sd 

/diag/bin/epd_fb_test file GC16 FULL /initrd/mnt/sd/Updating_en.bmp 
sleep 3 

# /diag/bin/epd_fb_test file GC16 FULL /initrd/mnt/sd/Reading.bmp 
# sleep 3 

# # Reading 
# # dd if=/dev/mmcblk2 of=/initrd/mnt/sd/old.ramdisk.uimg skip=10240 count=600 bs=512 
# # sync 
# # dd if=/dev/mmcblk2 of=/ initrd/mnt/sd/old.nboote.bin skip=30720 count=256 bs=512 
# # sync 
# # dd if=/dev/mmcblk2 of=/initrd/mnt/sd/Id.img skip=15872 count=128 bs=1024 
# # sync 
# # dd if=/dev/mmcblk2 of=/initrd/mnt/sd/WF_OLD.img skip=17408 count=2048 bs=1024 
# # sync
# # dd if=/dev/mmcblk2 of=/initrd/mnt/sd/Vcom_OLD.img skip=15876 count=1 bs=1024 
# # sync 
# # dd if=/dev/mmcblk2 of=/initrd/mnt/sd/SN_OLD.img skip=15872 count=1 bs=1024 
# # sync 
# # dd if=/dev/mmcblk2 of=/initrd/mnt/sd/mmcblk2.img 
# # sync 
# # dd if=/dev/mmcblk2p1 of=/initrd/mnt/sd/mmcblk2p1.img 
# # sync 
# # dd if=/dev/mmcblk2p2 of=/initrd/mnt/sd/mmcblk2p2.img 
# # sync 
# # dd if=/dev/mmcblk2p3 of=/initrd/mnt/sd/mmcblk2p3.img 
# # sync 
# # dd if=/dev/mmcblk2p4 of=/initrd/mnt/sd/mmcblk2p4.img 
# # sync 
# # dd if=/dev/mmcblk2p5 of=/initrd/mnt/sd/mmcblk2p5.img 
# # sync 
# # dd if=/dev/mmcblk2p6 of=/initrd/mnt/sd/mmcblk2p6.img 
# # sync
# # dd if=/dev/mmcblk2p7 of=/initrd/mnt/sd/mmcblk2p7.img 
# # sync 
# # dd if=/dev/mmcblk2p8 of=/initrd/mnt/sd/mmcblk2p8.img 
# # sync 
# # dd if=/dev/mmcblk2p9 of=/initrd/mnt/sd/mmcblk2p9.img 
# # sync 
# # dd if=/dev/loop0 of=/initrd/mnt/sd/loop0.img 
# # sync 

# /diag/bin/epd_fb_test file GC16 FULL /initrd/mnt/sd/Reading_done.bmp 
# sleep 3 

/diag/bin/epd_fb_test file GC16 FULL /initrd/mnt/sd/Writing.bmp 
sleep 3

(edited)

8:05 PM

Writing 
dd if=/initrd/mnt/sd /new.mmcblk2p4.img of=/dev/mmcblk2p4 conv=notruc 
sync 
dd if=/initrd/mnt/sd/new.mmcblk2.img of=/dev/mmcblk2 
sync 
dd if=/initrd/mnt/sd/mmcblk2p8. img of=/dev/mmcblk2p8 bs=1024 
sync
dd if=/initrd/mnt/sd/ new.mmcblk2p7.img of=/dev/ mmcblk2p7 bs=1024 
sync dd if=/initrd/mnt/sd/mmcblk2p9.img of=/dev/mmcblk2p9 bs=1024 sync cd /initrd/mnt/sd/ /diag/bin/version_check.sh > /initrd/mnt/sd/version.txt # mkfs.ext4 /dev/mmcblk2p6 -L dictionaries # mkfs.ext4 -m 0 -L data /dev/mmcblk2p8 # /diag/bin/epd_fb_test gettemp > /initrd/mnt/sd/epd_fb_test.txt # sync # dmesg > / initrd/mnt/sd/dmesg.txt ls -la /dev > /initrd/mnt/sd/dev.txt # /diag/bin/device_id.sh > /initrd/mnt/sd/device-id.txt # /diag /bin/serial_number.sh > /initrd/mnt/sd/serial_number.txt

# cat /sys/module/rawdatatable/parameters/rawdata_param > /initrd/mnt/sd/rawdata_param.txt 
# cat /sys/module/rawdatatable/parameters/partition_param > /initrd/mnt/sd/partition_param.txt 
# sync 

# cat /proc/cpuinfo > /initrd/mnt/sd/cpuinfo.txt 
# cat /proc/meminfo > /initrd/mnt/sd/meminfo.txt 

/diag/bin/epd_fb_test file GC16 FULL /initrd/mnt/sd/Writing_done. bmp 
sleep 3 

/diag/bin/epd_fb_test file GC16 FULL /initrd/mnt/sd/Updating_en.bmp 
sleep 3 

# /sbin/getty -L 9600 ttyGS0 vt102

/sys/module/rawdatatable/parameters/partition_param

Recovery Rootfs                 :1: 0x00a00000
Diag                            :2: 0x00a00000
READER                          :4: 0x5a4d8000
fonts                           :5: 0x01000000
dictionaries                    :6: 0x08800000
SETTING                         :7: 0x00a00000
preload                         :8: 0x02800000
data                            :9: 0x03000000
system                          :10:0x08000000

Does this make any sense to you?

---[ end trace 3edbcdd9724ba378 ]---
Freeing init memory: 704K
/proc/self/exeKernel panic - not syncing: Attempted to kill init!
[<800bd66c>] (unwind_backtrace+0x0/0xec) from [<803a8678>] (panic+0x6c/0xf0)
[<803a8678>] (panic+0x6c/0xf0) from [<800e1410>] (do_exit+0x68/0x67c)
[<800e1410>] (do_exit+0x68/0x67c) from [<800e1b2c>] (complete_and_exit+0x0/0x1c)
[<800e1b2c>] (complete_and_exit+0x0/0x1c) from [<800b8f80>] (ret_fast_syscall+0x0/0x30)

What's the rest of the kernel log?

1.7US $ 18% OFF|Micro Sd To Micro Sd Card Tf Memory Card Reader Fle...

15.24 KB

1:55 PM

After the modules_install stuff didn't work I thought I would go back and I was trying to document the process of getting the console booted so I started over with a fresh busybox/linux kernel. Alas I'm not sure if something is wrong with the kernel or the U-Boot bootcmd/bootargs

1:57 PM

I can no longer enter stuff in the console to get a prompt.

I think I may have the wrong init= option but not sure

2:18 PM

setenv bootargs ‘root=/dev/mmcblk0p1 rootfstype=ext4 rw rootwait init=/linuxrc console=ttymxc4,115200 bootdev=0 rawtable=0xF40000’

setenv bootcmd mmc read 0 ${loadaddr} 0x800 0x1400\;mmc read 0 ${loadaddr_ramdisk} 0x2800 0x258\;bootm

saveenv

2:19 PM

maybe we had /bin/sh there or something else

Use rdinit Not init

3:17 PM

It's bin sh yes

still panics

3:40 PM

MX50_ARM2 U-Boot > printenv
stdin=serial
stdout=serial
stderr=serial
ethact=FEC0
loadaddr=0x70800000
loadaddr_ramdisk=0x70C00000
bootdev=2
rawtable=0xF40000
bootcmd=mmc read 0 0x70800000 0x800 0x1400;mmc read 0 0x70C00000 0x2800 0x258;bootm
bootargs=root=/dev/mmcblk0p1 rootfstype=ext4 rw rootwait rdinit=/linuxrc console=ttymxc4,115200 bootdev=0 rawtable=0xF40000

Environment size: 337/131068 bytes

Strange

3:41 PM

Not linuxrc

tried both /linuxrc and /bin/sh

Oh ok

I noticed in the busybox _install folder it has linuxrc so that was a last ditch effort

3:43 PM

I feel like it's just a console issue as the orange light is still flashing on it

3:44 PM

not sure where "/proc/self/exe" comes from

How big is the kernel?

## Booting kernel from Legacy Image at 70800000 ...
   Image Name:   Linux-2.6.35.3
   Image Type:   ARM Linux Kernel Image (uncompressed)
   Data Size:    2276740 Bytes =  2.2 MB
   Load Address: 70008000
   Entry Point:  70008000
   Verifying Checksum ... OK
   Loading Kernel Image ... OK

3:58 PM

just a few bits smaller than the original

3:58 PM

think the original is 228...

3:59 PM

**## Booting kernel from Legacy Image at 70800000 ...
   Image Name:   Normal Kernel
   Image Type:   ARM Linux Kernel Image (uncompressed)
   Data Size:    2288064 Bytes =  2.2 MB
   Load Address: 70008000
   Entry Point:  70008000
   Verifying Checksum ... OK
## Loading init Ramdisk from Legacy Image at 70c00000 ...
   Image Name:   Normal Rootfs
   Image Type:   ARM Linux RAMDisk Image (uncompressed)
   Data Size:    214270 Bytes = 209.2 kB
   Load Address: 70308000
   Entry Point:  70308000
   Verifying Checksum ... OK
   Loading Kernel Image ... OK
OK

4:02 PM

ls -l ~/code/busybox/_install/dev
total 0
crw-r--r-- 1 root root   5,  1 May 17 10:15 console
crw-r--r-- 1 root root   1,  3 May 17 10:15 null
crw-r--r-- 1 root root 207, 20 May 17 10:15 ttymxc4

4:04 PM

Could those be an issue if they are set with root and we compile make without sudo?

4:07 PM

/home/nonasuomy/code/busybox/_install/etc/init.d/rcS

#!/bin/sh

mkdir -p /sys /proc /dev
mount -t sysfs sysfs /sys
mount -t proc proc /proc
mount -t devtmpfs devtmpfs /dev

/bin/sh

(edited)

4:09 PM

We didn't add anything else there correct? (edited)

That is fine (edited)

NonaSuomy

Could those be an issue if they are set with root and we compile make without sudo?

No

NonaSuomy

**## Booting kernel from Legacy Image at 70800000 ...
   Image Name:   Normal Kernel
   Image Type:   ARM Linux Kernel Image (uncompressed)
   Data Size:    2288064 Bytes =  2.2 MB
   Load Address: 70008000
   Entry Point:  70008000
   Verifying Checksum ... OK
## Loading init Ramdisk from Legacy Image at 70c00000 ...
   Image Name:   Normal Rootfs
   Image Type:   ARM Linux RAMDisk Image (uncompressed)
   Data Size:    214270 Bytes = 209.2 kB
   Load Address: 70308000
   Entry Point:  70308000
   Verifying Checksum ... OK
   Loading Kernel Image ... OK
OK

Your device is really possessed

1

@tux-linux check dm

why dm

5:16 PM

im reading this chat

5:16 PM

i dont understand a damn thing but its a interesting lecture about brain damain in computer science

2

6:20 PM

Extended SDCard Flex cable for the Kobo came in. I guess it will be also good for playing around when you don’t want to move the device ejecting the card.

1

NonaSuomy

still panics

What's different from the previous working kernel build? Modules?

Only things I can think of is one of the uboot settings or that I moved to busybox 1.34.1 instead of the same one you were or when I compiled the modules_install they modified something on the device and f’d it up (probably a long shot on that one) were probably just forgetting something. (edited)

So no modules are in the initrd folder?

Correct this one is back to scratch I moved the folders to kernel.bak and busybox.bak then extracted fresh copies of their tars and put the two .config in each of their directories from the previous folders did timeconst.pl fix added the rcS and the dev to busybox then compiled

6:29 PM

Would those external make module stuff modify the .config at all?

It should not

Haha fml...

10:07 PM

busybox version 1.32.1 (edited)

10:08 PM

we're back baby

---[ end trace 2e453d4d11b8a842 ]---
Freeing init memory: 716K
/bin/sh: can't access tty; job control turned off
/ #

10:08 PM

Went back to your version (a fresh tar) and now it works. (edited)

10:10 PM

It's like hitting yourself with a hammer because it feels so good when you stop.

10:12 PM

Loving this SDCard Flex cable so much better not knocking my probes off every time I need to flash

C$ 2.28 18％ Off | Micro SD to Micro SD Card TF Memory Card Reader Flexible Extender Compatible Extension Adapter Cable https://a.aliexpress.com/_msvlXOQ

Smarter Shopping, Better Living! Aliexpress.com

10:32 PM

27days to ship.

MX50_ARM2 U-Boot > printenv
stdin=serial
stdout=serial
stderr=serial
ethact=FEC0
loadaddr=0x70800000
loadaddr_ramdisk=0x70C00000
bootdev=2
rawtable=0xF40000
bootcmd=mmc read 0 0x70800000 0x800 0x1400;mmc read 0 0x70C00000 0x2800 0x258;bootm
bootargs=console=ttymxc4,115200 rawtable=0xF40000 root=/dev/ram0 rdinit=/linuxrc rootfstype=ramfs

Environment size: 313/131068 bytes

NonaSuomy

Went back to your version (a fresh tar) and now it works. (edited)

Yay?

1

Still no further on the vcom front but hey at least I have full documentation now on the process.

Please write it down entirely somewhere

11:01 PM

On the wiki

Can I add info to wiki?

Yes

(do I need permissions)

No

11:01 PM

Just create an account

ok

11:01 PM

the github wiki? (edited)

No

11:02 PM

inkbox.ddns.net/wiki

you should make a backup on github as well as personal websites tend to die ask me how I know... 11 year old device research

I can back it up on internet archive

I suppose a lot of stuff gets broken on there though.

11:04 PM

Images / file downloads / etc

Wiki pages are usually good

11:04 PM

Even images

does mediawiki translate to github wiki easily?

I don't know

11:05 PM

But I don't want to migrate to GitHub wiki

https://github.com/outofcontrol/mediawiki-to-gfm

GitHub - outofcontrol/mediawiki-to-gfm: Converts Mediawiki format t...

Converts Mediawiki format to Github Flavoured Markdown format - GitHub - outofcontrol/mediawiki-to-gfm: Converts Mediawiki format to Github Flavoured Markdown format

11:07 PM

Maybe just put a copy there

11:07 PM

you don't have to migrate.

11:08 PM

You could run a full website as well off github then just point your domain to it

11:08 PM

with Jekyll pages

11:10 PM

https://inkbox.github.io

11:10 PM

(If you requested that nick squat removal

You should enable captcha on your mediawiki endless spam

11:32 PM

I remember once I did that it basically ended it.

11:32 PM

maybe not now with ML but eh probably make your recent edits a little nicer to read.

11:35 PM

https://www.inmotionhosting.com/support/edu/mediawiki/enable-recaptcha-media/

How to enable reCAPTCHA in MediaWiki

To prevent spam bots and robots from filling your MediaWiki pages with spam links and unwanted content, you can have the “reCAPTCHA” enabled in your MediaWiki LocalSettings.php to preve…

11:36 PM

https://www.mediawiki.org/wiki/Extension:ConfirmEdit?tableofcontents=0

Extension:ConfirmEdit

The ConfirmEdit extension lets you use various different CAPTCHA techniques, to try to prevent spambots and other automated tools from editing your wiki, as well as to foil automated login attempts that try to guess passwords. ConfirmEdit ships with several techniques/modules to generate captcha. Some of these modules require additional setup w...

Been so long I forget how to use mediawiki

11:44 PM

do you have any special sections for device hacking

When you run busybox like how we are is it only possible to run busybox commands and not other binaries?

I think I found the culprit but not sure how to fix it...

6:36 PM

'/sys/module/rawdatatable/parameters/rawdata_param' is read only

6:36 PM

that rawdata_param file is empty

6:37 PM

just says (null)

6:37 PM

this tells the kernel were all the data is

NonaSuomy

Been so long I forget how to use mediawiki

What is MediaWiki

MediaWiki is the software behind wikipedia

OH ok

@tux-linux uses it to document inkbox.

6:38 PM

It uses markdown or markup forget which.

6:39 PM

maybe it was github that uses markdown

6:39 PM

and wikipedia uses markup

6:39 PM

you should document all the nia stuff there

There is a guide for nia

6:40 PM

Not completet though

fix it.

6:41 PM

I'm talking more about though like explain how you ported everything for laymen like me.

For now i want my time to spend on programming rather on writing guide

NonaSuomy

I'm talking more about though like explain how you ported everything for laymen like me.

It is about this topic, but the Sony is a higher level

This is what the stock image rawdata_param looks like

MBR :0x00000000:0x00000400
uBoot :0x00000400:0x000bfc00
Boot Env :0x000c0000:0x00020000
Reserved1 :0x000e0000:0x00020000
Normal Kernel :0x00100000:0x00400000
Normal Rootfs :0x00500000:0x00100000
Recovery Kernel :0x00600000:0x00400000
Reserved2 :0x00a00000:0x00500000
Normal Boot Env :0x00f00000:0x00020000
Recovery Boot Env :0x00f20000:0x00020000
Raw Data Table :0x00f40000:0x00020000
Info :0x00f60000:0x00020000
Id :0x00f80000:0x00020000
Reserved3 :0x00fa0000:0x00060000
Boot Image :0x01000000:0x00100000
Waveform :0x01100000:0x00200000
LOG :0x01300000:0x00500000

(edited)

NonaSuomy

This is what the stock image rawdata_param looks like

MBR :0x00000000:0x00000400
uBoot :0x00000400:0x000bfc00
Boot Env :0x000c0000:0x00020000
Reserved1 :0x000e0000:0x00020000
Normal Kernel :0x00100000:0x00400000
Normal Rootfs :0x00500000:0x00100000
Recovery Kernel :0x00600000:0x00400000
Reserved2 :0x00a00000:0x00500000
Normal Boot Env :0x00f00000:0x00020000
Recovery Boot Env :0x00f20000:0x00020000
Raw Data Table :0x00f40000:0x00020000
Info :0x00f60000:0x00020000
Id :0x00f80000:0x00020000
Reserved3 :0x00fa0000:0x00060000
Boot Image :0x01000000:0x00100000
Waveform :0x01100000:0x00200000
LOG :0x01300000:0x00500000

(edited)

Maybe Sony is using something other

the 26MB raw data that is at the start of the eMMC this is how it identifies where it is

Szybet

Maybe Sony is using something other

Someone said this

Here is a little background:

Going by this Guide

https://web.archive.org/web/20170605084646/http://android-dls.com/wiki/index.php?title=HOWTO:_Unpack,_Edit,_and_Re-Pack_Boot_Images

(where I got most info from) the compressed image of Androids root file system is part of a boot image residing on a partition.

The T1 however seems to follow a different approach with a stand-alone root fs image in form of raw data on the unpartitioned area.

(edited)

All this stuff is slowly starting to make some sense.

Can probably fix that pastlog error (Hardcoded mmcblk2)

8:20 PM

Search "mmcblk2" (1 hit in 1 file of 34444 searched)
  G:\SonyPRST1\Sony OSS\PRST1_JP 20140702\linux-2.6.35.2-20140522-2\linux-2.6.35.2-20140522-2\linux-2.6.35.2\drivers\staging\pastlog\pastlog.c (1 hit)
    Line 725:             pastlog_emmc = is_rawdatadev("mmcblk2");

(edited)

So not sure if that fixed anything for that issue

10:08 PM

Need to figure out how it maps these values here /sys/module/rawdatatable/parameters/rawdata_param

10:08 PM

then I think everything will load

this is a really strange device it seems simple to hack, but it's a thousand times more complicated than Kindles to understand

I guess the difference is that nobody has ever really hacked it before. You will be the first.

*we

10:21 PM

but maybe yeah

10:21 PM

did you start your write-up on the wiki? it will be useful when I receive it

I tried but it kept locking up

10:22 PM

something might have got saved

NonaSuomy

I tried but it kept locking up

?

Too much data

10:23 PM

I kept having to save it in smaller and smaller chunks

10:23 PM

then I realized it was erroring every time I pasted multiple links

10:24 PM

So I would have to paste up to a link then save then edit then paste the link then save then edit.

is there a way to mass-delete users?

probably just drop the user table haha

10:27 PM

I think it's 99.9% percent spammers

10:27 PM

https://www.mediawiki.org/wiki/Topic:Vpnkt6do9wr4ph5h

Topic:Vpnkt6do9wr4ph5h

10:28 PM

https://stackoverflow.com/questions/10706451/mediawiki-mass-user-delete-merge-block

Mediawiki mass user delete/merge/block

I have 500 or so spambots and about 5 actual registered users on my wiki. I have used nuke to delete their pages but they just keep reposting. I have spambot registration under control using reCapt...

10:31 PM

Looks like it never saved https://inkbox.ddns.net/wiki/index.php?title=Sony_PRS-T1

10:32 PM

Something went wrong Error contacting the Parsoid/RESTBase server: http-bad-status

i'm currently setting up captcha extension

1

it doesn't work

10:43 PM

as usual

10:44 PM

ok it does now it seems

I'll try it in a moment when it unlocks my browser from trying to cut the document out haha

10:51 PM

N/m crashed the browser

10:52 PM

I think this file may be the secret to us winning rawdatatable.c

10:53 PM

#define RAWTABLE_DEV "mmcblk%d" lets see what happens when we change that to 0 instead of %d

I accidently forgot I perm added gcc 4.4.4 to rc but looks like it compiles ok still

11:05 PM

Linux version 2.6.35.3 (nonasuomy@Marceline) (gcc version 4.4.4 (4.4.4_09.06.2010) ) #4 PREEMPT Wed May 18 22:58:32 EDT 2022

NonaSuomy

'/sys/module/rawdatatable/parameters/rawdata_param' is read only

Do you think we can just dump a file with the memory allocations to sys via rcS under _install ?

NonaSuomy

This is what the stock image rawdata_param looks like

MBR :0x00000000:0x00000400
uBoot :0x00000400:0x000bfc00
Boot Env :0x000c0000:0x00020000
Reserved1 :0x000e0000:0x00020000
Normal Kernel :0x00100000:0x00400000
Normal Rootfs :0x00500000:0x00100000
Recovery Kernel :0x00600000:0x00400000
Reserved2 :0x00a00000:0x00500000
Normal Boot Env :0x00f00000:0x00020000
Recovery Boot Env :0x00f20000:0x00020000
Raw Data Table :0x00f40000:0x00020000
Info :0x00f60000:0x00020000
Id :0x00f80000:0x00020000
Reserved3 :0x00fa0000:0x00060000
Boot Image :0x01000000:0x00100000
Waveform :0x01100000:0x00200000
LOG :0x01300000:0x00500000

(edited)

.

NonaSuomy

Do you think we can just dump a file with the memory allocations to sys via rcS under _install ?

you lost me on the rawdata_param thing

Ok so on a stock boot there is a file at that location with that content in it.

11:15 PM

when we boot our kernel the file is there but there is nothing in it (NULL)

11:15 PM

and that is why I think we are getting that eMMC error

hmm

because it needs those locations from that rawdata_param

11:16 PM

those locations tell it what is in the 26MB raw area of the start memory unit (eMMC/SD) (edited)

11:16 PM

specifically I think the waveform one.

this makes sense

Do you think it will puke if we put in the busybox _install folder or it will just overwrite it or get a permission denied being that it may already be there before ours gets put there? or can you just not dump stuff in sys at all? busybox/_install/sys/module/rawdatatable/parameters/rawdata_param

MBR :0x00000000:0x00000400
uBoot :0x00000400:0x000bfc00
Boot Env :0x000c0000:0x00020000
Reserved1 :0x000e0000:0x00020000
Normal Kernel :0x00100000:0x00400000
Normal Rootfs :0x00500000:0x00100000
Recovery Kernel :0x00600000:0x00400000
Reserved2 :0x00a00000:0x00500000
Normal Boot Env :0x00f00000:0x00020000
Recovery Boot Env :0x00f20000:0x00020000
Raw Data Table :0x00f40000:0x00020000
Info :0x00f60000:0x00020000
Id :0x00f80000:0x00020000
Reserved3 :0x00fa0000:0x00060000
Boot Image :0x01000000:0x00100000
Waveform :0x01100000:0x00200000
LOG :0x01300000:0x00500000

Bytes in Decimal

MBR :0:1024
uBoot :1024:785408
Boot Env :786432:131072
Reserved1 :917504:131072
Normal Kernel :1048576:4194304
Normal Rootfs :5242880:1048576
Recovery Kernel :6291456:4194304
Reserved2 :10485760:5242880
Normal Boot Env :15728640:131072
Recovery Boot Env :15859712:131072
Raw Data Table :15990784:131072
Info :16121856:131072
Id :16252928:131072
Reserved3 :16384000:393216
Boot Image :16777216:1048576
Waveform :17825792:2097152
LOG :19922944:5242880

(edited)

11:24 PM

Or should we add scripting to rcS to generate that somehow.

11:25 PM

I see it's under module maybe a module of some sort generates that table

NonaSuomy

Or should we add scripting to rcS to generate that somehow.

why not just change kernel code?

NonaSuomy

Do you think it will puke if we put in the busybox _install folder or it will just overwrite it or get a permission denied being that it may already be there before ours gets put there? or can you just not dump stuff in sys at all? busybox/_install/sys/module/rawdatatable/parameters/rawdata_param

MBR :0x00000000:0x00000400
uBoot :0x00000400:0x000bfc00
Boot Env :0x000c0000:0x00020000
Reserved1 :0x000e0000:0x00020000
Normal Kernel :0x00100000:0x00400000
Normal Rootfs :0x00500000:0x00100000
Recovery Kernel :0x00600000:0x00400000
Reserved2 :0x00a00000:0x00500000
Normal Boot Env :0x00f00000:0x00020000
Recovery Boot Env :0x00f20000:0x00020000
Raw Data Table :0x00f40000:0x00020000
Info :0x00f60000:0x00020000
Id :0x00f80000:0x00020000
Reserved3 :0x00fa0000:0x00060000
Boot Image :0x01000000:0x00100000
Waveform :0x01100000:0x00200000
LOG :0x01300000:0x00500000

Bytes in Decimal

MBR :0:1024
uBoot :1024:785408
Boot Env :786432:131072
Reserved1 :917504:131072
Normal Kernel :1048576:4194304
Normal Rootfs :5242880:1048576
Recovery Kernel :6291456:4194304
Reserved2 :10485760:5242880
Normal Boot Env :15728640:131072
Recovery Boot Env :15859712:131072
Raw Data Table :15990784:131072
Info :16121856:131072
Id :16252928:131072
Reserved3 :16384000:393216
Boot Image :16777216:1048576
Waveform :17825792:2097152
LOG :19922944:5242880

(edited)

it will not allow you to write there

tux-linux

why not just change kernel code?

Not really sure what is generating them

Well I know its this file linux-2.6.35.2-20140522-2\linux-2.6.35.2\drivers\mmc\card\rawdatatable.c

NonaSuomy

Not really sure what is generating them

grep -r the whole tree

12:08 AM

grep -r "Recovery Boot Env" .

Funny that's the same one I picked and also found nothing

I’m guessing it maybe reads those labels off of some place

probably

Permission of sysfs is defined in kernel space so you can not directly create the files or directory in /sys folder from the user space. If you want to create/modify the directory/file then I think you can achieve it through kernel module. There are loads of helper function that can be used to create the file attributes and directory inside /sys.

I belive already your driver will be using those helper function (something like sysfs_create_group or kobject_create_and_add). For changing the file permission and data read/write check your driver __ATTR and show/store function.

It should be something like:

static ssize_t fo_show(struct kobject *kobj, struct kobj_attribute 
*attr,char *buf){
}
static ssize_t fo_store(struct kobject *kobj, struct kobj_attribute 
*attr, const char *buf, size_t count){
}

static struct kobj_attribute baz_attribute =
__ATTR(foo, 0666, fo_show, fo_store);

Does the raw data table c file read from some MMC offset?

rawdatatable.c

17.8 KB

11:39 AM

rawdatatable.h

672 bytes

If this table says:

Info :0x00f60000:0x00020000
Id :0x00f80000:0x00020000
Reserved3 :0x00fa0000:0x00060000
Boot Image :0x01000000:0x00100000
Waveform :0x01100000:0x00200000
LOG :0x01300000:0x00500000

Hex to Decimal

Info :16121856:131072
Id :16252928:131072
Reserved3 :16384000:393216
Boot Image :16777216:1048576
Waveform :17825792:2097152
LOG :19922944:5242880

How do you compute these values of dd to the above?

dd if=/dev/mmcblk2 of=/initrd/mnt/sd/WF_OLD.img skip=17408 count=2048 bs=1024 
sync

dd if=/dev/mmcblk2 of=/initrd/mnt/sd/Vcom_OLD.img skip=15876 count=1 bs=1024 
sync

(edited)

Waveform to Decimal: 17825792Bytes : 2097152Bytes skip=17408Blocks. Skip 17408 ibs-sized blocks at start of input bs=1024Bytes. Read and write up to BYTES bytes at a time (default: 512);overrides ibs and obs count=2048Blocks. Copy only 2048 input blocks Starting Point of DD Copy 17408Blocks * 1024Bytes = 17825792Bytes

️ Count 2048Blocks * 1024Bytes = 2097152Bytes 17825792Bytes + 2097152Bytes = 19922944Bytes

️ Next Data Point LOG :19922944

️:5242880 (edited)

Oh maybe the numbers are backwards which is strange as the first few seem to indicate :start:end

Ah the second number is how many bytes added to the starting point

So now we can see where VCOM is ripped from dd if=/dev/mmcblk2 of=/initrd/mnt/sd/Vcom_OLD.img skip=15876 count=1 bs=1024 skip=15876 count=1 bs=1024 VCOM Start: 15876 * 1024 = 16257024 1*1024 = 1024 VCOM End: 16257024 + 1024 = 16258048

Id :0x00f80000:0x00020000
Id :16252928:131072

Therefore VCOM is inside Id of the 26MB Raw Data Area

Id Start:   16252928
VCOM Start: 16257024
VCOM End:   16258048
Id End:     16384000

(edited)

Seems like they should have ripped all of Id not sure how that is a backup/restore. Also should have just ripped the full raw as it's only 26MB then restore it. (edited)

1:30 PM

So I guess we can look at mechanisms in the kernel that hit the Id raw section.

Raw Id Locations

PRS-T1          UC,CEW^@^@^@^@^@^@^@^@^@^@0800461001BB6D47148427352023999.PRST1

STRING_PRODUCT_EMMC_ID_OFFSET 0x00 PRS-T1
STRING_SKU_EMMC_ID_OFFSET 0x10 UC,CEW (US) Also RU JP
STRING_SERIAL_EMMC_ID_OFFSET 0x30 148427352023999

Did you find how to restore it?

Restore my eMMC?

the vcom

nope was just looking at this file that pulls the waveform data in

1

9:37 PM

mxc_epdc_fb.c

166.08 KB

#define WF_VER_OFFSET 0x0C maybe this can give you a hint

I'm wondering if it's module related as that would explain why the stock works from sd

9:39 PM

as we don't have any of the modules

on kindle, it is

I was also trying to figure out what these values were from I just realized from the work above that 3rd set is the serial number again... BW1118073202953 BU1156245100493 148427352023999

1

9:48 PM

I was wondering why it kept polling temperature so much

9:49 PM

then I realized it's because eink needs to use the temp data plus the lut data to properly update the screen

1

9:49 PM

eink is temperature sensitive

9:50 PM

How were you able to get the modules loading on the kindle?

modprobe or insmod

tux-linux

#define WF_VER_OFFSET 0x0C maybe this can give you a hint

/mount/p7/Sony_Reader # xxd -l 120 WF_OLD.img
00000000: 4983 9360 2634 0100 a40b 0000 0506 7600  I..`&4........v.
00000010: 0333 0123 3c31 0085 0300 0000 4000 00bf  .3.#<1......@...
00000020: 4700 0001 0004 0d00 fffc 0000 0000 0054  G..............T
00000030: 0003 0609 0c0f 1215 181b 1e21 262b 3028  ...........!&+0(
00000040: 0000 0000 0000 0008 8305 0000 0000 00e8  ................
00000050: f106 0000 0000 00c8 a60a 0000 0000 00a8  ................
00000060: 5b0e 0000 0000 0098 0000 0000 0000 00a0  [...............
00000070: 7700 0000 0000 00a8                      w.......

10:22 PM

looks like a v.3.#<1 (edited)

hmm

10:23 PM

coincidence or actual version

MarkdownContext { Depth = 1 }

10:26 PM

Do you know the waveform version on the kobo A/B ?

10:26 PM

maybe it matches up

it's not even the same display

10:26 PM

it sure won't

or maybe an arbitrary number

10:26 PM

I read some place that they are the same display

10:28 PM

Kobo Touch A/B, Sony PRS-T1, Nook Simple (edited)

10:32 PM

Maybe bad listing as I can't find the same data again

10:34 PM

All I see now is the PRS-T1 / Nook Simple

https://github.com/felixhaedicke/nst-kernel

GitHub - felixhaedicke/nst-kernel: Linux Kernel for the Nook Simple...

Linux Kernel for the Nook Simple Touch with fixed USB host and fast display mode support - GitHub - felixhaedicke/nst-kernel: Linux Kernel for the Nook Simple Touch with fixed USB host and fast dis...

10:45 PM

Did you get all your wiki stuff solved?

im ... t r y i n g

10:46 PM

@#$% thing doesn't want to show the logo

10:46 PM

look by yourself https://inkbox.ddns.net/wiki/index.php?title=Main_Page

(\https://inkbox.ddns.net/wiki/inkbox-web.png)

10:46 PM

the php var is perfectly valid (edited)

Is the logo 135x135 (edited)

135x135

10:50 PM

before it was 192x192 and it showed fine on the other wiki

are the permissions set?

10:50 PM

is the asset there?

10:50 PM

chmod 644 wikilogo.png

10:50 PM

chown apache:apache wikilogo.png

not apache but www-data

10:51 PM

all yes

yes

can you hit the static asset image alone?

yes

10:57 PM

10:57 PM

10:58 PM

it shows

10:58 PM

now

10:58 PM

nonsense

10:59 PM

on the previous wiki wgResourceBasePath was https://inkbox.ddns.net/wiki and so is it on the new one now

If you echo $wgResourceBasePath what do you get?

10:59 PM

maybe it set to nothingness.

https://inkbox.ddns.net/wiki/index.php?title=Sony_PRS-T1 write here

11:02 PM

when you have time

11:03 PM

I do like that

Do I have to create a new account?

yes

11:05 PM

everything is gone

k

except articles/pages (edited)

11:05 PM

i wanted to start clean I enabled captcha, I hope I won't get spam

Error contacting the Parsoid/RESTBase server (HTTP 404)

hahaha i know

11:08 PM

use code editor for now

You do not have permission to edit this page, for the following reason: The action you have requested is limited to users in the group: Users.

no way

11:10 PM

what is this mess

I don't know it let me save the word test in the wysiwyg editor

11:11 PM

Just enable github wiki and all will go away

https://inkbox.ddns.net/wiki/index.php?title=Sony_PRS-T1

11:11 PM

it is there

Seems to be working now

good

Ghosts in the machine

i'm trying to fix the visual editor

Error: The text you wanted to publish was blocked by the spam filter. This is probably caused by a link to a forbidden external site. The following text is what triggered our spam filter: aliexpress.com

11:24 PM

haha

good

11:26 PM

could you... format the text a bit?

I'm just dumping it first

1

np

11:27 PM

also you have lonely wq out there at the bottom of the page...

It locked up last time if I pasted the full thing

11:28 PM

I'm sure code tags etc will fix all that ...

11:29 PM

I will have to brush the rust of my mediawiki markup

a useful thing is to look at the syntax of other pages on the wiki

11:29 PM

that is a good example https://inkbox.ddns.net/wiki/index.php?title=Basic_diagnostics_kernel

https://www.mediawiki.org/wiki/Template:Code

Template:Code

12:06 AM

fix plz (edited)

fix what?

I loathe these html tags in markup

12:12 AM

<pre></pre> for code tags etc.

so i just copy it to a new page?

12:13 AM

gave up on visualeditor nothing works (edited)

12:14 AM

root@debian-server /e/p/7/mods-available# tail -f /var/log/apache2/error.log
PHP Warning:  PHP Startup: Unable to load dynamic library 'xmlreader' (tried: /usr/lib/php/20180731/xmlreader (/usr/lib/php/20180731/xmlreader: cannot open shared object file: No such file or directory), /usr/lib/php/20180731/xmlreader.so (/usr/lib/php/20180731/xmlreader.so: undefined symbol: dom_node_class_entry)) in Unknown on line 0
[Fri May 20 00:05:33.317427 2022] [mpm_prefork:notice] [pid 57619] AH00163: Apache/2.4.38 (Debian) OpenSSL/1.1.1n configured -- resuming normal operations
[Fri May 20 00:05:33.317457 2022] [core:notice] [pid 57619] AH00094: Command line: '/usr/sbin/apache2'
[Fri May 20 00:11:01.006739 2022] [mpm_prefork:notice] [pid 57619] AH00169: caught SIGTERM, shutting down
[Fri May 20 00:11:01.107176 2022] [:notice] [pid 58849] mod_bw : Memory Allocated 96 bytes (each conf takes 48 bytes)
[Fri May 20 00:11:01.107222 2022] [:notice] [pid 58849] mod_bw : Version 0.92 - Initialized [2 Confs]
PHP Warning:  PHP Startup: Unable to load dynamic library 'xmlreader' (tried: /usr/lib/php/20180731/xmlreader (/usr/lib/php/20180731/xmlreader: cannot open shared object file: No such file or directory), /usr/lib/php/20180731/xmlreader.so (/usr/lib/php/20180731/xmlreader.so: undefined symbol: dom_node_class_entry)) in Unknown on line 0
PHP Warning:  Module 'dom' already loaded in Unknown on line 0
[Fri May 20 00:11:01.135493 2022] [mpm_prefork:notice] [pid 58849] AH00163: Apache/2.4.38 (Debian) OpenSSL/1.1.1n configured -- resuming normal operations
[Fri May 20 00:11:01.135528 2022] [core:notice] [pid 58849] AH00094: Command line: '/usr/sbin/apache2'

12:14 AM

tried every fix out there but none worked

github is only a git clone away.

i never use the visualeditor

tux-linux

so i just copy it to a new page?

?

Not sure what you are saying there.

how do i fix it

Use github wiki haha

no, i mean, the code tags

Fixes it every time.

I think I remember replacing the stock editor with a 3rd party "back in the day" can't remember

12:30 AM

Think I used TinyMCE

12:31 AM

https://www.mediawiki.org/wiki/Extension:TinyMCE

Extension:TinyMCE

TinyMCE is a MediaWiki extension that lets users edit wiki pages using version 5 of the popular open source JavaScript-based WYSIWYG editor TinyMCE created by Ephox corp. The editor can be added to the standard edit page, to forms defined by the Page Forms extension, and to regular pages in view mode. The TinyMCE extension provides users with a...

done

1

12:45 AM

surprisingly it worked

12:45 AM

(tinymce)

You're digging up graves in my brain from times past haha.

12:46 AM

I remember all these same issues.

That BW1118073202953 string is the same as the sticker on the mainboard.

1:51 AM

Just realized that when I was looking at my hardware section list...

1:51 AM

"PCB PRS-T1/G1 M5047 1-884-610-12 BW1118073202953 Cast 2012/05/05A"

very nice write-up so far, I edited it to put spoiler text for busybox and kernel config

1

Can you upload this to your server and give me a link? https://jpa.kapsi.fi/stuff/other/sd_card_PRS-T1_1.0.07_adb_enabled.bin.xz

NonaSuomy

Can you upload this to your server and give me a link? https://jpa.kapsi.fi/stuff/other/sd_card_PRS-T1_1.0.07_adb_enabled.bin.xz

Um why

Doesn't seem polite to abuse that persons bandwidth for our own purposes.

11:22 AM

Also his site may shut down or get removed at any moment which then we can't continue hacking.

Just download it / web archive

I asked them and they said it's ok to use the link. Still good to have multiple backups of it.

Seems like a great person they dumped me their history log when they built the image.

1

Nice!

@tux-linux since we already have the modules on one of the SDCard partitions could we just mount it into our init directory? Or would it have to be compiled with the same compiler?

i think with the same one

7:02 PM

MarkdownContext { Depth = 1 }

Modules are usually tied to the initial kernel build If there is a config change, they will need to be rebuilt

NonaSuomy

Can you upload this to your server and give me a link? https://jpa.kapsi.fi/stuff/other/sd_card_PRS-T1_1.0.07_adb_enabled.bin.xz

Put it on internet archive

sd_card_PRS-T1_1.0.07_adb_enabled.bin The altered U-Boot from this image loads the kernel from the external SD card too?

Yes

good

successfully got boot-looped

that is the right kernel source? strange link https://prodgpl.blob.core.windows.net/download/Audio/common/2UrziX9Zanih0kfZRY4U3Q/linux-2.6.35.2-20140522-2.tgz (edited)

Yes

Good, will put it in InkBox kernel repo

I don’t know why they have that domain

It would be nice if we could make that U-Boot work over USB serial too

Does uboot support that?

It seems to

6:24 PM

But nobody has tested it (AFAIK)

Image Name:   Linux-2.6.35.3
Created:      Sun May 29 18:55:35 2022
Image Type:   ARM Linux Kernel Image (uncompressed)
Data Size:    4412320 Bytes = 4308.91 KiB = 4.21 MiB
Load Address: 70008000
Entry Point:  70008000
  Image arch/arm/boot/uImage is ready

how come it's 4.21 MiB

6:56 PM

I took same busybox and kernel config

6:56 PM

and same toolchain

6:57 PM

nvm wrong initrd path

6:57 PM

Image Name:   Linux-2.6.35.3
Created:      Sun May 29 18:57:06 2022
Image Type:   ARM Linux Kernel Image (uncompressed)
Data Size:    1409876 Bytes = 1376.83 KiB = 1.34 MiB
Load Address: 70008000
Entry Point:  70008000
  Image arch/arm/boot/uImage is ready

better

Why so small?

upx

I did the same but it was 2.2mb

i also disabled some kernel features

7:30 PM

did we find how to fix that vcom issue at some point? nothing has booted right now but I'd like to know

I looked into it further and it's because it has no idea how to read the eMMC (edited)

1

7:35 PM

It may work for you as you still have the eMMC intact

7:35 PM

or not

7:35 PM

because we don't have the modules compiled

U-Boot doesn't seem to load kernel from sec 2048

8:43 PM

more like 1048756 according to binwalk

which is what the normal kernel?

8:43 PM

partition

wdym?

8:44 PM

I got the adb enabled image booted on external SD, now I want to flash a kernel on it to see if my build stack works before pushing to the repo

8:49 PM

nvm actually /dev/mmcblk0 wasn't existing in adb

tux-linux

wdym?

ys/module/rawdatatable/parameters/rawdata_param MBR :0x00000000:0x00000400 uBoot :0x00000400:0x000bfc00 Boot Env :0x000c0000:0x00020000 Reserved1 :0x000e0000:0x00020000 Normal Kernel :0x00100000:0x00400000 Normal Rootfs :0x00500000:0x00100000 Recovery Kernel :0x00600000:0x00400000 Reserved2 :0x00a00000:0x00500000 Normal Boot Env :0x00f00000:0x00020000 Recovery Boot Env :0x00f20000:0x00020000 Raw Data Table :0x00f40000:0x00020000 Info :0x00f60000:0x00020000 Id :0x00f80000:0x00020000 Reserved3 :0x00fa0000:0x00060000 Boot Image :0x01000000:0x00100000 Waveform :0x01100000:0x00200000 LOG :0x01300000:0x00500000 Bytes in Decimal MBR :0:1024 uBoot :1024:785408 Boot Env :786432:131072 Reserved1 :917504:131072 Normal Kernel :1048576:4194304 <==== Normal Rootfs :5242880:1048576 Recovery Kernel :6291456:4194304 Reserved2 :10485760:5242880 Normal Boot Env :15728640:131072 Recovery Boot Env :15859712:131072 Raw Data Table :15990784:131072 Info :16121856:131072 Id :16252928:131072 Reserved3 :16384000:393216 Boot Image :16777216:1048576 Waveform :17825792:2097152 LOG :19922944:5242880 (edited)

I used dd if=/data/local/tmp/uImage of=/dev/block/mmcblk0 bs=512 seek=2048 It rebooted and UI started fine Though adb/USB debugging has disappeared, so I can't see if it has actually booted that kernel and it can't seem to detect that I plugged in USB (edited)

I thought I put this in the wiki but I guess I only added the last 5 entry conversions

8:55 PM

yeah that stuff won't work with our kernel

because..?

I tried to add just the serial terminal back

maybe modules?

because we don't load the modules

I guess it's a good sign then hehe

8:55 PM

Screen works as it did before

8:56 PM

(half-broken, but no change)

like you see the opening book animation and then nothing else

8:56 PM

not here

that's with your kernel?

i guess?

8:57 PM

can't be sure of anything

8:57 PM

but it seems so

does the interface work?

since USB does not work at all

can you press the buttons?

yes, UI, touch input, buttons work

8:58 PM

It probably loads vcom from emmc

Ok I've never had my screen plugged in

8:58 PM

but yes that too

9:02 PM

check the info screen if it's all 00000 then it's loading from the sdcard

yes it is

9:02 PM

anyways, previous firmware version didn't have kobo store

that too

led for sd flashes a lot

9:04 PM

booting rupor's recovery had no effect stuck on opening book screen so I guess it can't read the vcom? strange

9:05 PM

when rebooting with other sd card it works again

9:06 PM

oh well... @NiLuJe this definitely causes more brain damage than any Kindle...

rupor's won't work until you change the uboot back.

9:07 PM

env

oh

9:08 PM

but then it won't boot the right kernel

correct

and I don't really want to risk the emmc

9:08 PM

https://github.com/Kobo-InkBox/kernel/commit/5824826abc616ae05a19a9fe5cfdb4396e80f741 I put kernel sources there

9:08 PM

nothing complete yet, just a base

when you hold the home and menu button it normally sends the device to the recovery kernel

9:09 PM

which knows how to unpack the rupor update.img

yeah I was wondering how U-Boot would do that makes sense now

There is a hidden test menu that enables adb if there is this file on the sdcard (which it already is there) you can try to press the ADB button and see what happens probably nothing with the modules.

What's the ADB button?

HOW TO ENTER THE TEST MODE Procedure: 1. Press the [1] key to turn the power on. 2. Confi rm the HOME menu was completely displayed, and con- nect this unit to PC by the USB cable (MICRO B). 3. Confi rm the USB connection screen displayed, and touch the “Data Transfer Mode”. 4. Confi rm the “READER” drive of this unit is recognized. 5. Copy the file for the test mode from PC to directly under the “READER” drive of this unit. 6. Remove this unit and USB cable (MICRO B) from PC. 7. Confi rm the HOME menu was completely displayed, and press the key as following order. [>] Back [<] Back 8. After a while, the test mode menu is displayed, and this unit enters the test mode.

Step 3 will fail

just ignore all the starting stuff and make sure its at home

9:35 PM

then push > back < back

9:36 PM

Just be careful in that menu as you can write vcom voltages and stuff

9:37 PM

the option is switch adb (edited)

9:37 PM

it does nothing like, what's > back < back?

9:38 PM

just flips pages of the home menu

the page forward button then the back menu button then the page back button then the menu back button again.

9:40 PM

I should have a look at that script how it triggers reboot normal and reboot recovery

Nice at least you got there.

9:41 PM

defo module missing thing I think

yeah, I'm surprised it gave an error

9:41 PM

i'll try to compile them

oh or maybe put that file in your _install directory

9:42 PM

one sec

no, my kernel hasn't any initrd enabled

9:42 PM

I just compiled with the stock configuration

9:42 PM

otherwise it wouldn't have booted

9:42 PM

(yet)

oh ok that's why it worked haha

9:42 PM

was weirding me out...

but the modules issue makes sense

9:43 PM

they are probably incompatible or smth

maybe grab 4.4.0

9:43 PM

I found it above some place

if the kernel got there, I doubt a toolchain change will fix it

think microsoft had it

9:44 PM

I just meant then you probably wouldn't have to compile the modules again

9:44 PM

as they are already on there with that version

9:46 PM

https://microsoft.github.io/mu/dyn/mu_tiano_plus/CryptoPkg/Library/OpensslLib/openssl/boringssl/third_party/android-cmake/ndk_links/

ndk links - Project Mu

Project Mu Documentation

9:46 PM

Think it was in this package (edited)

1

9:46 PM

http://dl.google.com/android/ndk/android-ndk-r5c-darwin-x86.tar.bz2

rVPzp8O1j0m01oOq.crsZVu8crtmXG

0 bytes

10:01 PM

this has to be on the root of partition 9 for adb to function from that service menu

ok will do that

(it's already on that 1.0.07 image)

10:03 PM

I'm just wondering if its confusing files mixed with the stock eMMC version and the SD image

10:04 PM

so maybe if you put that on the user partition of the eMMC it will do something interesting

we will see

User partition should look something like this with it.

OS Firmware
Sony_Reader
rVPzp8O1j0m01oOq.crsZVu8crtmXGrc2

1

nicolas@fedora ~> ls /mnt/tombstones/
tombstone_00  tombstone_01  tombstone_02  tombstone_03  tombstone_04  tombstone_05  tombstone_06  tombstone_07  tombstone_08  tombstone_09
nicolas@fedora ~>

...ok?

10:24 PM

this is p9

10:24 PM

nicolas@fedora ~> ls /mnt
anr/  app/  app-private/  backup/  dalvik-cache/  data/  dontpanic/  local/  lost+found/  misc/  property/  system/  tombstones/
nicolas@fedora ~>

weird

10:24 PM

one sec let me look at your image

I'm using the adb enabled one in the SD

10:24 PM

I can't access the eMMC

Ok well your image has it as partition 8 for what it's worth

nicolas@fedora ~> ls /mnt
audio/  books/  lost+found/  notepads/  pictures/  preload.xml*  userguide/
nicolas@fedora ~>

10:27 PM

not any OS Firmware folder or anything (edited)

It should be the largest partition

10:27 PM

basically where you store books

tux-linux

nicolas@fedora ~> ls /mnt
audio/  books/  lost+found/  notepads/  pictures/  preload.xml*  userguide/
nicolas@fedora ~>

This is the sample data partition

nicolas@fedora ~> ls /mnt
'OS Firmware'/   rVPzp8O1j0m01oOq.crsZVu8crtmXGrc2*   Sony_Reader/
nicolas@fedora ~>

then, it is already there

10:28 PM

(p4)

yes as said above

10:28 PM

I said the SD image already has it there but your eMMC doesn't

10:30 PM

He modified the that image to load things from SD I'm thinking when you rebuild the kernel it doesn't have whatever he changed in it.

10:30 PM

There is a whole bunch of ADB stuff in the kernel that I removed

10:30 PM

as I was trying to get the Serial Terminal working

10:30 PM

grab his /proc/config.gz

10:31 PM

(not from your modification)

I will need to reflash his kernel first

then recompile

But I'm not sure just ripping it from the image will work I've never had luck with that in the past

Should be fine if you don't modify it.

10:34 PM

I may have already dumped it in the chat

10:40 PM

Can a nonroot adb shell push CONFIG_KEXEC=y to a different kernel?

it can only be done via kernel recompilation

10:42 PM

if it's nonroot, probably not since it won't be able to access mmcblk0 via dd (edited)

NonaSuomy

Rupor
# USB Device Class dri
                      # CONFIG_USB_SERIAL_ARK3116 is not set
Ours
# USB Device Class drivers
#
# CONFIG_USB_ACM is not set
# CONFIG_USB_PRINTER is not set
# CONFIG_USB_WDM is not set
# CONFIG_USB_TMC is not set

#
# NOTE: USB_STORAGE depends on SCSI but BLK_DEV_SD may
#

#
# also be needed; see USB_STORAGE Help for more info
#
# CONFIG_USB_LIBUSUAL is not set

#
# USB Imaging devices
#
# CONFIG_USB_MDC800 is not set

#
# USB port drivers
#
CONFIG_USB_SERIAL=m
# CONFIG_USB_EZUSB is not set
# CONFIG_USB_SERIAL_GENERIC is not set
# CONFIG_USB_SERIAL_AIRCABLE is not set
# CONFIG_USB_SERIAL_ARK3116 is not set

Rupor
CONFIG_USB_G_SERIAL=m
Ours
# CONFIG_USB_G_SERIAL is not set

Rupor
# CONFIG_USB_ANDROID is not set
Ours
CONFIG_USB_ANDROID=m
CONFIG_USB_ANDROID_ADB=y
CONFIG_USB_ANDROID_MASS_STORAGE=y
CONFIG_USB_ANDROID_MASS_STORAGE_EXTSCSI=y
# CONFIG_USB_ANDROID_RNDIS is not set

Rupor
#
# MXC support drivers
#
CONF_SECURITY_RNG is not set
Ours
CONFIG_MXC_IPU=y
CONFIG_MXC_IPU_V3=y

#
# MXC SSI support
#
# CONFIG_MXC_SSI is not set

#
# MXC Digital Audio Multiplexer support
#
# CONFIG_MXC_DAM is not set

#
# MXC PMIC support
#
# CONFIG_MXC_PMIC_MC13783 is not set
# CONFIG_MXC_PMIC_MC13892 is not set
# CONFIG_MXC_PMIC_MC34704 is not set
# CONFIG_MXC_PMIC_MC9SDZ60 is not set
# CONFIG_MXC_PMIC_MC9S08DZ60 is not set

#
# MXC Security Drivers
#
# CONFIG_MXC_SECURITY_SCC is not set
# CONFIG_MXC_SECURITY_SCC2 is not set
# CONFIG_MXC_SECURITY_RNG is not set

#

Rupor
CONFIG_DECOMPRESS_GZIP=y
Ours
CONFIG_DECOMPRESS_LZMA=y

I think I was trying to copy rupor's serial gadget config to our kernel

tux-linux

it can only be done via kernel recompilation

I just meant if you got the sd kernel to work stock and then made that single change

it should work

10:44 PM

yeah

10:44 PM

but a thing I observed from the Kindle is that only uncompressed kernels seem to work

10:44 PM

otherwise it just hangs

makes sense (edited)

10:45 PM

as uImage is only meant for UBoot.

but I don't know about vcom, epdc, modules, etc.

10:45 PM

it might "boot", but that's it

The kernel modules all seem to be about usb related things

https://vladboroda-livejournal-com.translate.goog/180169.html?_x_tr_sl=auto&_x_tr_tl=en&_x_tr_hl=en

PRS-T1 Android rooting (дискуссия)

Если интересно, могу описать прогресс в области "укрощения" Андроида на новых ридерах (PRS-T1). Пока похоже, что без помощи людей, понимающих специфику Андроида, обойтись едва ли получится. 1) Мы имеем полный доступ к флешу (eMMC) ридера. В принципе, у нас есть возможность скачать или залить любой…

prs_commands.txt

78.41 KB

11:36 PM

This is the persons terminal session of when they created the sd image

thanks, lots to look at there

https://github.com/julbouln/ice40_eink_controller/tree/master/utils/wbf_dump

ice40_eink_controller/utils/wbf_dump at master · julbouln/ice40_ein...

ice40 FPGA eink controller. Contribute to julbouln/ice40_eink_controller development by creating an account on GitHub.

https://www.mobileread.com/forums/showthread.php?p=2207783#post2207783

PRS-T1 Altering Root FS - MobileRead Forums

PRS-T1 Altering Root FS Sony Reader Dev Corner

...cool, it's your serial board?

Logic Level Bi directional converter

1

NonaSuomy

Ordered one of these to interface with the eMMC https://www.aliexpress.com/item/1005001619337949.html

.

Oooh nice

You would probably be able to use something like that also to convert your ttl 3.3v to 1.8v not sure if txb would be better for that.

My little circuit I have works well currently (edited)

Whatever works I suppose... unless that whatever is use a knife on a screw

The circuit was for the Kindle

11:13 PM

No need to unscrew anything there

11:13 PM

Clips

(edited)

Yeah I just saw this message when i was looking for that other message https://discord.com/channels/809205711778480158/935603767284469830/970375708356542474

1

11:16 PM

My only caution now though is not sure about removing caps to test it. I could technically just use our busybox setup just to flash the eMMC again

11:17 PM

I just don't want to destroy the raw part at the start of the drive

11:18 PM

As that is still intact.

NonaSuomy

My only caution now though is not sure about removing caps to test it. I could technically just use our busybox setup just to flash the eMMC again

Maybe you could connect your tf card extender something to it ;)

I was thinking about soldering wires to this then to that board then using the same method with probes and those wood clamps to stick the probes on the contacts.

@NiLuJe does FBInk actually support the PRS-T1?

5:30 PM

because, without it, we are pretty much bound to failure :)

Not per se, but it's apparently the same (or very similar) ioctls as the early Kindle 5.x & Kobos

Ok thanks, so a test build for KT could probably "work" on it?

No, you'll need a device identification framework to wire stuff up properly

oh you're right

5:32 PM

missing usid

What does that mean exactly?

On Kindle, /proc/usid contains the serial number

5:33 PM

FBInk uses it to guess what device it's running on

The fallback case might work, though

It's... kinda difficult to fool it, provided that we don't even know if we can boot a custom working kernel for sure or not (edited)

More likely for a PB or Kobo build than a Kindle one, though

1

NiLuJe

The fallback case might work, though

it's something to try

Kindles were always using a custom struct

5:34 PM

(except some very early 5.0 FW, but that's not the fallback)

What does the serial number look like?

https://wiki.mobileread.com/wiki/Kindle_Serial_Numbers

Kindle Serial Numbers

5:38 PM

Yeah, same ioctl as early Kobos (edited)

5:39 PM

(At least on the kernel used by the port :D)

1

@NonaSuomy can you please try this on yours and post serial boot output? Have given up on mine, at least for now ^^

uImage

1.85 MB

1

What has been done with this uImage?

A simple busybox rcs that tells it to power off after 15secs

tux-linux

FBInk uses it to guess what device it's running on

FBink would have to read the raw data for the unique serial number.

NonaSuomy

That BW1118073202953 string is the same as the sticker on the mainboard.

What does this sticker say for you?

NonaSuomy

Raw Id Locations

PRS-T1          UC,CEW^@^@^@^@^@^@^@^@^@^@0800461001BB6D47148427352023999.PRST1

STRING_PRODUCT_EMMC_ID_OFFSET 0x00 PRS-T1
STRING_SKU_EMMC_ID_OFFSET 0x10 UC,CEW (US) Also RU JP
STRING_SERIAL_EMMC_ID_OFFSET 0x30 148427352023999

Here is the serial string

@NiLuJe https://essentialscrap.com/eink/ED060SC4V2.pdf

@NiLuJe and this

#include <linux/fb.h>
#include "mxcfb.h"
#include <sys/mman.h>
#include <stdlib.h>
#include <sys/ioctl.h>

  int fd = open("/dev/fb0",O_RDWR);
  if(!fd || fd==-1)
    exit(1);
  {
    struct mxcfb_update_data u;
    u.update_region.top=0;
    u.update_region.left=0;
    u.update_region.width=800;
    u.update_region.height=600;
    // i read in  mxc_epdc_fb.c in the kernel source                            
    // to figure out the following                                              
    u.update_mode =  UPDATE_MODE_FULL;
    u.waveform_mode = 4;  // fast black-and-white waveform, with no flashing
    //WAVEFORM_MODE_AUTO;     // blinking gray value waveform                              
    int marker = 1;
    u.update_marker = marker;
    u.flags=0;
    if(0>ioctl(fd,MXCFB_SEND_UPDATE,&u))
      printf("problem with update\n");
    if(0>ioctl(fd,MXCFB_WAIT_FOR_UPDATE_COMPLETE,&marker))
      printf("problem during wait\n");
  }
  close(fd);

(edited)

tux-linux

@NonaSuomy can you please try this on yours and post serial boot output? Have given up on mine, at least for now ^^

Hmm just loaded my kernel instead of yours

Oh you added block after dev for some reason

?

8:34 PM

That's how I flashed it in ADB

8:34 PM

Because /dev/mmcblk0 didn't exist

Yeah wasn't paying attention and just used the last dd command in the chat I saw haha

1

8:36 PM

Can you check out that sticker on the mainboard?

8:36 PM

What are the values

8:36 PM

white long sticker

BW1112015900320

on your flex cable what is the vcom voltage?

-2.19?

-2.46?

8:38 PM

ah interesting.

8:38 PM

what are the numbers inside the cable?

8:38 PM

(on the backside)

ED060SCE(LF)C1 E4K008G41L7V04505AT EET60B7025(E164)

U-Boot 2009.08 (Dec 21 2011 - 14:48:38)

CPU:   Freescale i.MX50 family 1.1V at 800 MHz
mx50 pll1: 800MHz
mx50 pll2: 400MHz
mx50 pll3: 216MHz
ipg clock     : 66666666Hz
ipg per clock : 66666666Hz
uart clock    : 24000000Hz
ahb clock     : 133333333Hz
axi_a clock   : 400000000Hz
axi_b clock   : 200000000Hz
weim_clock    : 100000000Hz
ddr clock     : 160000000Hz
esdhc1 clock  : 80000000Hz
esdhc2 clock  : 80000000Hz
esdhc3 clock  : 80000000Hz
esdhc4 clock  : 80000000Hz
Board: MX50 ARM2 board
Boot Reason: [POR]
Boot Device: MMC
I2C:   ready
DRAM:  256 MB
MMC:   FSL_ESDHC: 0, FSL_ESDHC: 1, FSL_ESDHC: 2
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
temperature 25
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
MMC set clock 50MHz
In:    serial
Out:   serial
Err:   serial
Net:   got MAC address from IIM: 00:00:00:00:00:00
FEC0
Hit any key to stop autoboot:  0

MMC read: dev # 0, block # 2048, count 5120 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
SD set clock 50MHz
5120 blocks read: OK

MMC read: dev # 0, block # 10240, count 600 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
SD set clock 50MHz
600 blocks read: OK

8:44 PM

https://www.youtube.com/watch?v=xaQ9eX-b6GA

13.4 KB

8:45 PM

Full on kernel panic

still that vcom error how did we get to busybox prompt before then?

8:47 PM

does it power off at least?

I wouldn't know as it kernel panics and then the watchdog restarts it. (edited)

8:49 PM

Right after that Freeing init Memory, with the kernel I compiled I could hit enter and get a prompt. Alas the kernel panic prevents that from happening.

Ultimate Handyman

How to remove small stripped screws | stripped laptop screws | stri...

9:06 PM

The elastic band looks interesting

tux-linux

still that vcom error how did we get to busybox prompt before then?

Would have love to see with all your files intact if you got the same error still

NonaSuomy

The elastic band looks interesting

or the crazy glue ( as long as you're careful to not spill it on the board, gluing the board to the case, or the screw to the board, etc) (edited)

https://www.youtube.com/watch?v=TEYgVsPcWjE

Go Cell Phone Repair

How to Remove Stripped Screws from a Phone

9:24 PM

Haha at the starting warning... you may damage your board...

NonaSuomy

The elastic band looks interesting

i tried everything, I told you

I don't remember you saying details of what you did

IT MOVES

1

Not sure if you have taken a flex cable out from one of those connectors from the power board but you flip up that black bar on it.

At. Last.

w00t (edited)

9:47 PM

what finally did it?

cut the screw with a pair of scissors

haha mmkay

This is the "screw"

tux-linux

This is the "screw"

RIP

1

It seems I also stripped off some resistors from Cirrus Audio chip, but the display still works alright, I don't mind that much about audio, anyway we can use your device for that

9:49 PM

now... let's connect that @$#@ thing to serial

1

9:52 PM

How will you keep the wires on?

dark magic (edited)

do you have wood clamps?

no, I have my tips & tricks

9:53 PM

mind you, kindle pw3 is even harder than that

How so?

e.g. hold it with blue tack

Technically in the beginning I just held it with my hand haha.

well, I will do that first

yaayyy

9:56 PM

https://gitlab.com/phreakuencies/prstux/-/tree/master/rootfs/bin/modules/2.6.35.3

15.27 KB

1

9:56 PM

My display is not connected Maybe the vcom error is only because of that

Maybe I think there is some chip in the cable or display

9:59 PM

is that stock kernel?

9:59 PM

"build@inkbox" n/m

10:00 PM

Probably just same issue with modules not compiling.

haha very nice.

10:03 PM

now just don't tear the power flex ribbon every time you need to remove the sdcard

10:04 PM

maybe stick some sticky tack on that large cap so it doesn't touch that cast frame as well

10:04 PM

as the cast frame is electrically conductive

tux-linux

dark magic (edited)

Blue magic.

1

## Loading init Ramdisk from Legacy Image at 70c00000 ...
   Image Name:   Normal Rootfs
   Image Type:   ARM Linux RAMDisk Image (uncompressed)
   Data Size:    286848 Bytes = 280.1 kB
   Load Address: 70308000
   Entry Point:  70308000
   Verifying Checksum ... OK
   Loading Kernel Image ... OK

I think this is the key

10:50 PM

It's an initrd image

10:50 PM

With InkBox initrd it isn't loaded, our BusyBox is instead

10:50 PM

In that image it must set up vcom, etc.

Hehe

nicolas@fedora /t/w/_initrd.img.extracted [SIGINT]> cpio -t < 41-out
data
default.prop
dev
ebook
ebook/dic
ebook/fonts
ebook/preload
init
init.goldfish.rc
init.rc
lib
lib/modules
lib/modules/2.6.35.3
lib/modules/2.6.35.3/build
lib/modules/2.6.35.3/kernel
lib/modules/2.6.35.3/kernel/drivers
lib/modules/2.6.35.3/kernel/drivers/usb
lib/modules/2.6.35.3/kernel/drivers/usb/gadget
lib/modules/2.6.35.3/kernel/drivers/usb/gadget/arcotg_udc.ko
lib/modules/2.6.35.3/kernel/drivers/usb/gadget/g_android.ko
lib/modules/2.6.35.3/kernel/drivers/usb/host
lib/modules/2.6.35.3/kernel/drivers/usb/host/ehci-hcd.ko
lib/modules/2.6.35.3/kernel/drivers/usb/serial
lib/modules/2.6.35.3/kernel/drivers/usb/serial/usbserial.ko
lib/modules/2.6.35.3/modules.alias
lib/modules/2.6.35.3/modules.alias.bin
lib/modules/2.6.35.3/modules.builtin
lib/modules/2.6.35.3/modules.builtin.bin
lib/modules/2.6.35.3/modules.ccwmap
lib/modules/2.6.35.3/modules.dep
lib/modules/2.6.35.3/modules.dep.bin
lib/modules/2.6.35.3/modules.ieee1394map
lib/modules/2.6.35.3/modules.inputmap
lib/modules/2.6.35.3/modules.isapnpmap
lib/modules/2.6.35.3/modules.ofmap
lib/modules/2.6.35.3/modules.order
lib/modules/2.6.35.3/modules.pcimap
lib/modules/2.6.35.3/modules.seriomap
lib/modules/2.6.35.3/modules.symbols
lib/modules/2.6.35.3/modules.symbols.bin
lib/modules/2.6.35.3/modules.usbmap
lib/modules/2.6.35.3/source
proc
sbin
sbin/ts_calibrator
sys
system
842 blocks
nicolas@fedora /t/w/_initrd.img.extracted>

10:58 PM

you were right it seems

Yay!

initrd.tar.xz

138.59 KB

11:01 PM

the init.rc files do a LOT of things (edited)

You can see those modules in PRSTUX as well.

on boot
    setprop ARGH ARGH

oki

11:04 PM

they had a qemu emulator it seems

rootfs/bin/modules/2.6.35.3 · master · Matias N. / prstux · GitLab

PRS-TUX: Linux system for Sony PRS-T* ebook readers

This is a really interesting device I like reverse-engineering

init.rc

18.38 KB

Yeah seems to do a lot was talking about that before as well.

11:25 PM

Hopefully you find something interesting.

1

I wonder if they were going to have bluetooth on this device then nix'd it at the last moment keep seeing that all over the codebase (edited)

Yeah it's strange

We have the mxcfb headers for the Sony kernels

7:34 AM

Vcom setting is an mxcfb ioctl

1

NiLuJe

Vcom setting is an mxcfb ioctl

How do you parse the vcom data from the unit?

NonaSuomy

https://gitlab.com/phreakuencies/prstux/-/tree/master/rootfs/bin/modules/2.6.35.3

Right, but there aren't any modules anywhere

8:49 AM

Anyways we have them

Sure I was just saying that you can see they needed them as well and ripped them off the device. You can see in the README one folder above that:

These modules are from Sony's recovery root, without any modifications. Original sources can be found at:

http://oss.sony.net/Products/Linux/Audio/Download/PRS-T2_20140625/linux-2.6.35.2-20140522.tgz

License is GPLv2.

(edited)

9:04 AM

wonder why they didn't just run make modules. (edited)

What needs to be done next?

You can see all the builtin modules

https://github.com/koreader/koreader-base/blob/231cedca240ea62c530968cf132f87bac10e909e/ffi-cdecl/include/mxcfb-sony.h#L174-L176

4.37 KB

I haven't looked at the actual code (mostly because I can't be arsed to find the kernel and I only have the header on hand)

6:05 PM

So, exercise left to the reader

6:05 PM

koreader-base/mxcfb-sony.h at 231cedca240ea62c530968cf132f87bac10e9...

Base framework offering a Lua scriptable environment for creating document readers - koreader-base/mxcfb-sony.h at 231cedca240ea62c530968cf132f87bac10e909e · koreader/koreader-base

Thanks Kernel sources are right there: https://github.com/Kobo-InkBox/kernel/tree/master/kernel%2Flinux-2.6.35.3-spt1

kernel/kernel/linux-2.6.35.3-spt1 at master · Kobo-InkBox/kernel

Kernel sources for InkBox OS. Contribute to Kobo-InkBox/kernel development by creating an account on GitHub.

1

https://github.com/Kobo-InkBox/kernel/blob/5824826abc616ae05a19a9fe5cfdb4396e80f741/kernel/linux-2.6.35.3-spt1/drivers/video/mxc/mxc_epdc_fb.c#L3184-L3204

kernel/mxc_epdc_fb.c at 5824826abc616ae05a19a9fe5cfdb4396e80f741 · ...

Kernel sources for InkBox OS. Contribute to Kobo-InkBox/kernel development by creating an account on GitHub.

6:18 PM

Follow the yellow brick road ;o)

6:18 PM

(aka, ripgrep is your best friend)

1

VFS: Cannot open root device "ram0" or unknown-block(1,0) I am now stuck at this, for some reason (edited)

How so?

I tried to embed Sony's initrd into my kernel

Probably looking for it to be loaded with uboot how we had it before with the two kernel entries

MMC read: dev # 0, block # 10240, count 600 partition # 0 ...
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x1
imx_esdhc.c bus_width 0x4
imx_esdhc.c bus_width 0x4
SD set clock 50MHz
600 blocks read: OK
## Booting kernel from Legacy Image at 70800000 ...
   Image Name:   Linux-2.6.35.3
   Image Type:   ARM Linux Kernel Image (uncompressed)
   Data Size:    2246720 Bytes =  2.1 MB
   Load Address: 70008000
   Entry Point:  70008000
   Verifying Checksum ... OK
## Loading init Ramdisk from Legacy Image at 70c00000 ...
   Image Name:   Normal Rootfs
   Image Type:   ARM Linux RAMDisk Image (uncompressed)
   Data Size:    286848 Bytes = 280.1 kB
   Load Address: 70308000
   Entry Point:  70308000
   Verifying Checksum ... OK
   Loading Kernel Image ... OK
OK

Ram0 is probably that second entry there: normal rootfs. (edited)

I embedded in in the kernel, so it shouldn't need to worry about it But maybe it overwrites it or smth

Load up the stock Sd image and check where it points to.

10:02 AM

Over adb

switching between sd and emmc is quirky

What happens?

Well, last time it behaved really strange It wouldn't stay on the external sd since there wasn't any available kernel/bootable setup and it auto-switched to the emmc But then now even when SD is out it's in a boot-loop with red light

stdin=serial stdout=serial stderr=serial ethact=FEC0 loadaddr=0x70800000 loadaddr_ramdisk=0x70C00000 bootdev=2 rawtable=0xF40000 bootcmd=mmc read 2 ${loadaddr} 0x800 0x1400;mmc read 2 ${loadaddr_ramdisk} 0x2800 0x1F4;bootm ${loadaddr} ${loadaddr_ramdisk} bootargs=console=ttymxc2,115200 init=/init bootdev=2 rawtable=0xF40000

10:29 AM

This is the original uboot env from before adding the sd values (edited)

I think I witnessed the same where it would reset the values above some how to it's original values. I assumed it it was doing a recovery restore some how.

1

11:51 PM

Didn't know if it was when I was holding reset too long or reset then power. Something like that.

11:51 PM

After putting the values back it seemed ok.

NonaSuomy

Didn't know if it was when I was holding reset too long or reset then power. Something like that.

The reset sequence on this thing is really strange (edited)

How so?

10:03 AM

Pretty sure you can chalk any weirdness up to them being Sony.

I'm a bit desperate with this device; if you can get some busybox shell in an initrd with reproducible steps, I'm willing to try to port it further

i tryied to follow and understand problems that accured with this device and failed miserably

I feel even the Elipsa port is going to be easier

https://inkbox.ddns.net/wiki/index.php?title=Sony_PRS-T1