iLiad iLiad kernel development

[Antartica]

Originally Posted by alegeott

The operation should not be too invasive or dangerous, but obviously void your warranty. I think I will post a guide for the whole operation.

Thanks for the update alegeott. Please keep us posted on the results of your hardware hacking.

OTOH, I'm not good with the soldering iron, so probably I'll not dare to modify my iliad .

P.S.: Ulno, I've just seen your update to the old post stating that the the required USB controller support was merged into mainline and is in 2.6.32-rc5. Wow! Fantastic news indeed .

[ulno]

Kernel: arch/arm/boot/zImage is ready

Thanks Antarctica! - now I am eager to go home to try to execute this one ;-) (probably won't work as I made a bad typecast in the USB driver (from 8 to 16 bit to make it compile, might need to look closer into it.)
I will report, when back at home.

Best,
Ulno

[ulno]

Ok, I can kexec it, but of course, I don't se anything.

Just for toying a little bit more around with options, I tried to load my zImage in qemu-system-arm, but I just get a black screen. Is this supposed to happen?

Is there an option to pretest the kernel somehow?

I also found another very interesting thing in kernel 2.6.32-rc6: drivers/net/netconsole.c - seems to offer a console via network

Also a thing, which would be easily testable in qemu. Anybody has an idea what I have to do to run my kernel images in qemu?

Cheers,
Ulno

[Antartica]

Originally Posted by ulno

Ok, I can kexec it, but of course, I don't se anything.

As expected . Next step is trying with the USB controller support and the USB-to-serial adapter; alternatively there is the netconsole option.

Just for toying a little bit more around with options, I tried to load my zImage in qemu-system-arm, but I just get a black screen. Is this supposed to happen?

Is there an option to pretest the kernel somehow?

The problem is that ARM machines are different from x86 in the "standardization" aspect. Load addresses, I/O setup, etc are different in every board, and qemu has to be teached that. I don't know a way to pretest the kernel right now .

I also found another very interesting thing in kernel 2.6.32-rc6: drivers/net/netconsole.c - seems to offer a console via network

It was mentioned earlier in this thread, but nobody has tested it yet with the iliad. If you feel adventurous...

Also a thing, which would be easily testable in qemu. Anybody has an idea what I have to do to run my kernel images in qemu?

You would have to extend qemu to:

1. Support the specific iliad sub-architecture

2. Add the emulation bits for its specific hardware (delta display controller, buttons, etc).

Both of them require some intimate knowledge of the "emulated system" (the iliad), which I don't have. In all it would be quite a bit of work. iRex did some of it for its iON architecture (DR1000), which incorporates a modified qemu in the SDK; perhaps some of those modifications could be used as a starting point, but that is a project on its own.

IMHO the least-work-required way is to do one of:

1. Use USB-serial adapter

2. Use netconsole

3. Hack a driver so that the iliad use its own display for the boot messages.

I'm going to try the third option; if you can experiment the other two, it would be great.

Happy hacking!

[ulno]

Hi all,
just reporting my failures:
I tried to add all of nalims changes (as referred to by Antarctica) to the 2.6.32-rc6 kernel. It was mainly copying relevant stuff from nalim's kernel in the new one. However, I got already abad feeling when looking in all the changes which have happened in the arm architecture part in the new kernel, new interrupt handling totally new names for initialization routines, and more.
Even if I had a small configuration, letting netconsole run on qemu, I of course did not get any results (netcat stays dead) when kexecing the my own kernel.

It cost a weekend and was worth the try. However without a physical serial port, I think it's impossible to continue here.

I will think some days longer, but I am in the mood of trying to sell my iliad again and using something more pc like until I can afford an e-ink based programmable platform.

For now, thanks for your support,
Ulno

[Antartica]

Originally Posted by ulno

It cost a weekend and was worth the try. However without a physical serial port, I think it's impossible to continue here.

I will think some days longer, but I am in the mood of trying to sell my iliad again and using something more pc like until I can afford an e-ink based programmable platform.

First of all, thanks for your time and for activating this thread again.

About the prospects of having a working 2.6 kernel, I think it is doable, but it may need some months of light work to achieve it. With light work, I mean dedicating little spare time to this.

Anyway, you still have these options (hey, this sounds like a RPG!):

1.Use another approach to hacking the iliad. You still have the option of trying to achieve what you need of the iliad using the current 2.4 kernel. With this kernel you can use the arm port of Debian 4.0 Etch. It uses hardware FP emulation instead of softfloat and it will be slower than the current binaries of the iliad, but would be certainly more hackable.BTW, using an etch chroot in a CF card works (tested).

2. Hack another e-ink device. There are other e-ink devices that have been hacked (look at the devices supported by OpenInkpot), but AFAIK nither of them is as open/powerful as the iliad, nor has its nearly-pc-like ability (wacom touchscreen, ethernet+wifi, USB 2.0).

3. Schew the e-ink screen. Use a Nokia N810 or something similar, although the screen is too small for pdf files and comic reading. The Kohjinsha convertibles may fit the bill, but they are expensive.

4. Wait until the new crop of e-ink devices is here. Hopefully next year devices like the "enTourage Edge" will be on sale.

Anyway, I'll continue with the fake-console-driver attempt, so stay tuned.

[ulno]

Hi,
thanks for your encouragement Antarctica, sorry for the kind of strong reaction. I am just using the iliad at the moment to read some books - weird isn't it?

I hope I can afford an N900 at around Christmas (still the screen will be to small for serious reading), so this could add to some thinking about the N810.

Lots of good luck with your efforts and maybe I jump on board pretty soon again - probably need some two weeks of building up new courage - if I just would be able to get any feedback somewhere from the device, when I start the kernel. Maybe I have to start over with Nalim's kernel.

Did you gt this one running? Do you have a working configuration which let's Nalims kernel start and show at least something (not necessarily while booting - if I could login via ssh later and check some boot messages, this would be enough for a start)? Then I could try to backport netconsole to 2.6.21 and start the forward-port again.

Best,
Ulrich

[alegeott]

Hi guys...good news on the extraction of a UART (serial port), but not enough.
Me and my professor have opened my Iliad. We found the TP from which you can receive the signals output during boot.
Unfortunately, we have yet found 3 input TP, but none (for now) that accepts signals from which to interrupt the boot and let it continue step by step.
This is the output string shown at boot device.

iRex Technologies Sexy Program Loader for ER0100 series e-reader
(c) 2005, 2006 by iRex Technologies BV
TFFS Volume info unit size: 524288kByte
Reading SysSet info...Done
Hardware: 0007
VCOM: 2CB0
Source drivers: 0012
Initialising DELTA...

Input TP are 104, 113, 518; output TP are 114.
This TP are located on motherboard back, near the "extra connector".
https://wiki.mobileread.com/wiki/File:222115531.jpg
I post detail photos of the TP soon.

I hope the opening of my Iliad and it started without the display device does not have destroyed the chip DELTA screen management (could be a revival of power disproportionate to the chip). Ehm.

WARNING! YOU CAN HARM YOUR DEVICE, WITH ANY OF THIS ACTION!!!
CHILDREN, DO NOT DO IT AT HOME!

[Antartica]

Originally Posted by ulno

Did you gt this one running? Do you have a working configuration which let's Nalims kernel start and show at least something (not necessarily while booting - if I could login via ssh later and check some boot messages, this would be enough for a start)? Then I could try to backport netconsole to 2.6.21 and start the forward-port again.

Glad to be an encouragement source . About the N810: my wife has one; it's perfect for day-to-day web browsing (facebook, yahoo webmail, forums, news sites, youtube), and maybe for the occasional pdf, but that is it. It has not been used for ebooks having an iliad at home. I've read in it some comics (european ones, so large format and in color), and so much panning is a hassle. OTOH, as the OS is already debian-derived, it may fit your hacking interests. I don't know how much of this experience is translatable to the N900 with its smaller screen size.

About nalim's kernel: it has included an example .config for the iliad: /arch/arm/configs/irex_er0100_defconfig . I used that one (only modification: added NFS support). BUT the kernel had problems prior to executing the initrd scripts. To check that I put a program that shows a splash screen (gray16display, tested with 2.4) in the linuxrc and made it have the following lines at the beginning:

#!/bin/sh
mount -t proc proc /proc
/sbin/modprobe deltafb
/gray16display /splash.gray16

So the kernel has a panic or similar prior to executing the gray16display command. The deltafb driver that nalim wrote is specific for the "3" hardware version (i.e. alegeott has the "7" version, as can be seen in the output he posted), but I have the "3" version too:

ereader:~# dmesg | grep einkfb 
einkfb: Hardware version 3

This is why I'm right now doing the fake-console driver, to have information prior to the initrd.

Originally Posted by alegeott

Hi guys...good news on the extraction of a UART (serial port), but not enough.
Me and my professor of computer science have opened my Iliad. We found the TP from which you can receive the signals output during boot.
Unfortunately, we have yet found 3 input TP, but none (for now) that accepts signals from which to interrupt the boot and let it continue step by step.

Perfect . Of those TP you mention, which is the UART Tx, Rx and Ground? I assume that they are TTL level, nope? Just curiosity for now

[alegeott]

Many good news, but... not enough.
First of all, we closed the device and reconnected the screen.
The device has thus been able to finish the boot and we got to the login screen.
The TP functioning input was the No. 113; the presence of TP or their position it may change on a different hardware revision (mine is no.7): in the photos on wiki iliad and those attached at this post, there aren't TP near the debug connector.
At this point, we could then also login via serial port, but did not know the passwoord root (this isn't a standard word). Fortunately, I had already installed a version of Busybox. We have therefore launched a shell directly on the device and cleared the password field in file setup with the password (encrypted) to root.
Then you can login as root via serial port. The blank password is maintained even when you restart the device.
Soon I post a complete boot log file and a how-to for the entire procedure.

Unfortunately, you can't modify the boot process. It will be also difficult to recompile the kernel OpenEmbedded...
But I'll think about later.

Originally Posted by Antartica

Perfect . Of those TP you mention, which is the UART Tx, Rx and Ground? I assume that they are TTL level, nope? Just curiosity for now

I have not idea about it, sorry.

[alegeott]

I post here two log extracts (the attached txt file is a complete backup log).
This is the boot report:

iRex Technologies Sexy Program Loader for ER0100 series e-reader
(c) 2005, 2006 by iRex Technologies BV
TFFS Volume info unit size: 524288kByte
Reading SysSet info...Done
	Hardware:       0007
	VCOM:           2CB0
	Source drivers: 0012
Initialising DELTA...Done
Uploading splash screen...Done
Initialising keys...Done
Loading kernel image...Done
Loading initrd...Done
Jumping to kernel...

Uncompressing Linux............................................ done, booting the kernel.
Linux version 2.4.19-rmk7-pxa2-irex1 (buildmgr@buildserver) () #21100 Tue Oct 2 17:40:05 CEST 2007
CPU: XScale-PXA255 revision 6
Machine: iRex Technologies ER0100 eReader
Ignoring unrecognised tag 0x00000000
Memory clock: 99.53MHz (*27)
Run Mode clock: 398.13MHz (*4)
Turbo Mode clock: 398.13MHz (*1.0, active)
On node 0 totalpages: 16384
zone(0): 16384 pages.
zone(1): 0 pages.
zone(2): 0 pages.
Kernel command line: root=/dev/ram rw initrd=0xA1000000,4M rootfstype=ext2 init=/linuxrc rw console=ttyS2,115200 mem=64M
Console: colour dummy device 80x30
Calibrating delay loop... 397.31 BogoMIPS
Memory: 64MB = 64MB total
Memory: 59212KB available (1132K code, 275K data, 72K init)
Dentry cache hash table entries: 8192 (order: 4, 65536 bytes)
Inode cache hash table entries: 4096 (order: 3, 32768 bytes)
Mount-cache hash table entries: 1024 (order: 1, 8192 bytes)
Buffer-cache hash table entries: 4096 (order: 2, 16384 bytes)
Page-cache hash table entries: 16384 (order: 4, 65536 bytes)
POSIX conformance testing by UNIFIX
Linux NET4.0 for Linux 2.4
Based upon Swansea University Computer Society NET3.039
Initializing RT netlink socket
Using PXA255 frequency points.
Registering CPU frequency change support.
CPU clock: 398.131 MHz (33.000-400.000 MHz)
Starting kswapd
I2C: PXA algorithm module loaded.
enable_irq(11) unbalanced from c0085a84
I2C: Adding PXA-I2C-Adapter.
I2C: Scanning bus (51)
I2C: Successfully added bus
pty: 256 Unix98 ptys configured
Serial driver version 5.05c (2001-07-08) with no serial options enabled
ttyS00 at 0x0000 (irq = 15) is a PXA UART
ttyS01 at 0x0000 (irq = 14) is a PXA UART
ttyS02 at 0x0000 (irq = 13) is a PXA UART
SA1100 Real Time Clock driver v1.00
PCF8563: RTC Initialization Successfull
RAMDISK driver initialized: 16 RAM disks of 8192K size 1024 blocksize
ac97_codec: AC97 Audio codec, id: 0x5053:0x4304 (Philips UCB1400)
UCB1x00: probed IRQ55 correctly. Please remove machine dependencies from ucb1x00-core.c
Turning on amplifier
enable_irq(55) unbalanced from c00a5888
Linux Kernel Card Services 3.1.22
  options:  [pm]
Intel PXA250/210 PCMCIA (CS release 3.1.22)
usb.c: registered new driver usbdevfs
usb.c: registered new driver hub
NET4: Linux TCP/IP 1.0 for NET4.0
IP Protocols: ICMP, UDP, TCP
IP: routing cache hash table of 512 buckets, 4Kbytes
TCP: Hash tables configured (established 4096 bind 4096)
NET4: Unix domain sockets 1.0/SMP for Linux NET4.0.
NetWinder Floating Point Emulator V0.95 (c) 1998-1999 Rebel.com
irex_er0100_pcmcia_configure_socket: Resetting WLAN
RAMDISK: Compressed image found at block 0
Freeing initrd memory: 4096K
VFS: Mounted root (ext2 filesystem).
Freeing init memory: 72K
----------------------------------------
Starting initrd
----------------------------------------
Starting tffs filesystem
Using /lib/modules/tffs.o
tffs: TrueFFS driver 7100.76
tffs: will not use IRQ<7>tffs: TffsMapMem: request_mem_region(0,20000) OK
tffs: DOCH found<6>tffs: Socket 0 in addr 0x0
tffs: Device 0x0: size 0xe500000 HW sector 0x200 (recommended 0x1000)
tffs: use major device number 100
Partition check:
 tffsa: tffsa1 tffsa3 < tffsa5 tffsa6 tffsa7 >
tffs:     disk partition: dev_number=0x6401, 149084 sectors, start_sector=52
tffs:     disk partition: dev_number=0x6403, 2 sectors, start_sector=149136
tffs:     disk partition: dev_number=0x6405, 55484 sectors, start_sector=149188
tffs:     disk partition: dev_number=0x6406, 262652 sectors, start_sector=204724
tffs:     disk partition: dev_number=0x6407, 1196 sectors, start_sector=467428

Loading sysset and keystore modules
Using /lib/modules/sysset.o
iRex sysset module loaded.
Using /lib/modules/keystore.o
iRex keystore module loaded.

Starting delta ASIC with hwrev=7
Using /lib/modules/delta.o
einkfb: Hardware version 7
fbcon_setup: No support for fontwidth 8
fbcon_setup: type 0 (aux 0, depth 8) not supported
Console: switching to colour frame buffer device 96x64

Starting button device
Using /lib/modules/button.o
iRex ER0100 eReader Button Driver initialised
Kernel: Init WACOM On

Starting battery device
Using /lib/modules/battery.o
iRex ER0100 eReader Battery Gas Gauge Driver initialised

Check for downloaded firmware updates [os=0,fs=0]
EXT2-fs warning: mounting unchecked fs, running e2fsck is recommended
  os flag not set, no os.gz; nothing to do
  fs flag not set, no fs.gz; nothing to do
No updates, booting app fs...
EXT2-fs warning: mounting unchecked fs, running e2fsck is recommended
umount: Cannot open /proc/mounts
INIT: version 2.86 booting
EXT2-fs warning: mounting unchecked fs, running e2fsck is recommended
MSDOS FS: IO charset utf8
EXT2-fs warning: mounting unchecked fs, running e2fsck is recommended
Setting up IP spoofing protection: rp_filter.
Configuring network interfaces... done.
INIT: Entering runlevel: 5
Starting PCMCIA services: cardmgr[251]: watching 2 sockets
done.
cardmgr[252]: socket 1: Marvell
Starting MMC services
Using /lib/modules/2.4.19-rmk7-pxa2-irex1/kernel/drivers/ide/ide-mod.o
Uniform Multi-Platform E-IDE driver Revision: 6.31
ide: Assuming 50MHz system bus speed for PIO modes; override with idebus=xx
Using /lib/modules/2.4.19-rmk7-pxa2-irex1/kernel/drivers/ide/ide-disk.o
Using /lib/modules/2.4.19-rmk7-pxa2-irex1/kernel/drivers/ide/ide-probe-mod.o
cardmgr[252]: executing: 'modprobe cf8385 2>&1'
cardmgr[252]: + Using /lib/modules/2.4.19-rmk7-pxa2-irex1/kernel/drivers/pcmcia/cfio.o
(W)displayMgr.c:250,main() ePaper daemon (displayMgr 0.212)
(W)displayMgr.c:255,main() Display Manager is accessible to local processes only.
(W)erdmServer.c:62,initServer() Bound successfully to 127.0.0.1:50555.
Before calling register_pccard_driver
After calling register_pccard_driver
cardmgr[252]: + Using /lib/modules/2.4.19-rmk7-pxa2-irex1/kernel/drivers/pcmcia/cf8385.o
Entering cf_attach()
Before registering the client
irex_er0100_pcmcia_configure_socket: Resetting WLAN
BasePort1=0xf7000000, AssignedIRQ=68
(W)eripc.c:38,erIpcInitServer() Version: 256

(W)eripc.c:83,erIpcInitServer() Kernel: WACOM Off Delay start
Bound successfully to 127.0.0.1:50071.
After calling wlan_add_card function
Leaving cf_attach()
fbcon_setup: No support for fontwidth 8
fbcon_setup: type 0 (aux 0, depth 8) not supported
Waiting for mb-applet-icon-container ...
.cardmgr[252]: executing: './network start wlan0 2>&1'
cardmgr[252]: + Default Config
.cardmgr[252]: + wlan0     deepsleep:setting to on 
cardmgr[252]: + Sample private network setup
Could not init font path element /usr/lib/X11/fonts/misc, removing from list!
Could not init font path element /usr/lib/X11/fonts/100dpi, removing from list!
Could not init font path element /usr/lib/X11/fonts/75dpi, removing from list!
..(W)eripc.c:38,erIpcInitServer() Version: 256

(W)eripc.c:83,erIpcInitServer() Bound successfully to 127.0.0.1:50073.
..... wait done
Using /lib/modules/2.4.19-rmk7-pxa2-irex1/kernel/drivers/usb/hal.o
hal_pxa250.c: 1362-hal Initialization Success 
Using /lib/modules/2.4.19-rmk7-pxa2-irex1/kernel/drivers/usb/phci.o
td_cache = c022f9f0
ed_cache = c022fa60
usb.c: new USB bus registered, assigned bus number 1
(W)eripc.c:38,erIpcInitServer() Version: 256

(W)eripc.c:83,erIpcInitServer() Bound successfully to 127.0.0.1:50072.
hub.c: USB hub found
hub.c: 2 ports detected
USB HAL: HC suspend
usb_phci.c: usb-phci Initialization Success 
Using /lib/modules/2.4.19-rmk7-pxa2-irex1/kernel/drivers/scsi/scsi_mod.o
SCSI subsystem driver Revision: 1.00
Using /lib/modules/2.4.19-rmk7-pxa2-irex1/kernel/drivers/usb/storage/usb-storage.o
Initializing USB Mass Storage driver...
usb.c: registered new driver usb-storage
USB Mass Storage support registered.
(CL_W)main.c:315,main() gtk_rc_get_default_files [/etc/gtk-2.0/gtkrc]
(CL_W)main.c:315,main() gtk_rc_get_default_files [/home/root/.gtkrc-2.0]
$** can't synthesize pci hotplug events
AC97: Turning off sound
$** can't synthesize root hub events
Using /lib/modules/2.4.19-rmk7-pxa2-irex1/kernel/drivers/usb/pdc.o
USB HAL: DC suspend
usb_pdc.c: usb-pdc Initialization Success 
Using /lib/modules/2.4.19-rmk7-pxa2-irex1/kernel/drivers/usb/mscd.o
pdc_bus.c: Registered Driver devmscd
devmscd.c: devmscd Initialization Success 
(CL_E)erConnect.c:69,erConnectInit() Lock the connect mutex
(W)mb-applet-icon-container.c:1185,loadIcons() Can't find pRegIcon[56]
(W)mb-applet-icon-container.c:1185,loadIcons() Can't find pRegIcon[57]
(W)eripc.c:38,erIpcInitServer() Version: 256

(W)eripc.c:83,erIpcInitServer() Bound successfully to 127.0.0.1:50065.
(CL_E)metadataStoreManager.c:936,recreateArchives() Can not create directory /mnt/free/newspapers!
(CL_E)metadataStoreManager.c:947,recreateArchives() Can not create directory /mnt/free/books!
(CL_E)metadataStoreManager.c:958,recreateArchives() Can not create directory /mnt/free/documents!
(CL_E)metadataStoreManager.c:969,recreateArchives() Can not create directory /mnt/free/notes!
(CL_E)metadataStoreManager.c:921,recreateArchives() storage not mounted, storage type[5]
(W)eripc.c:38,erIpcInitServer() Version: 256

(W)eripc.c:83,erIpcInitServer() Bound successfully to 127.0.0.1:50070.
(CL_W)programManager.c:1076,pm_InitProgramTable() Application [APABIMOBI] executable [/usr/bin/xebreader] is not found: disabled it
(CL_W)programManager.c:1076,pm_InitProgramTable() Application [EBA] executable [/usr/bin/eba] is not found: disabled it
(CL_W)programManager.c:1076,pm_InitProgramTable() Application [ARINC] executable [/usr/bin/ereader_mapviewer] is not found: disabled it
(CL_W)programManager.c:1166,pm_AddExtensionsToTable() Program SHELL associated with [sh] without extensionInfo entry in registry. Consider fixing this!

OpenEmbedded Linux ereader ttyS2

ereader login: root
login[561]: root login  on `ttyS2'

This is the log message after 'halt' command.

root@ereader:~# halt

Broadcast message from root (ttyS2) (Thu Nov 19 16:58:25 2009):

The system is going down for system halt NOW!
INIT: Switching to runlevel: 0
INIT: Sending processes the TERM signal
root@ereader:~# INIT: Sending processes the KILL signal
Shutting down PCMCIA services: cardmgr[252]: executing: './network stop wlan0 2>&1'
cardmgr[252]: + /sbin/ifconfig wlan0 down
Before calling wlan_remove function
After calling wlan_remove function
cardmgr[252]: exiting
done.
Deconfiguring network interfaces... done.
usb.c: deregistering driver usb-storage
Sending all processes the TERM signal...
Can't deallocate console 2 errno 16
Sending all processes the KILL signal...
flushing ide devices: 
Power down.
Shutting off Power Supply

Walk on, soon I post the motherboard photos with TP connection...

[nrapallo]

Originally Posted by alegeott

I post here two log extracts (the attached txt file is a complete backup log).
This is the boot report:

I miei complimenti!

Keep up the good work!!

[Antartica]

Originally Posted by alegeott

I post here two log extracts (the attached txt file is a complete backup log).

Thanks. Reading through your logs I've realized one reason why my kexec'ed kernel didn't work: The initrd is limited to 4MB by linux kernel cmdline boot command! I had expanded mine to 8MB so that I coud fit more things there. Oh, my!

Keep up the good work .

[alegeott]

Now the photos...
The input TP is the No. 113; the output TP is the no. 114.
This is the link at photo of mb retro, with the DIY cable.
The link at photo of retro display; the link at photo of front mb; the link at photo of battery board.


The particular of mb retro with DIY cables. Note that the connected cables at TP 111 and TP 518 here are only a trial. The black cable is the cable grounding.


The detail of MB front at debug connector.

[granikus]

Hi all,

Did someone done some progress in 2.6 kernel porting since last report?