Finally I tryed to control the update to 2.5 and I failed (ie I am fully updated) but it could be interesting to describe the process.

First there was a preupdate. It closes the network and tells you to shutdown the machine. If instead you enable networking again, you can see that the update installed a flashing process in rc0, ie in the shutdown process. This flash seems to install some improvements in the launch system and kernel modules, and it is numbered as 2.4.10 or similar. I copied it to the USB stick and left it update, as I was sure the main filesystem was not modifyed.

Then the real update was to begin. It proceeds by dowloading os.gz and fs.gz from iRex, storing it in a partition /mnt/protected/images/ and rebooting the machine so the update is done when starting again. Reading the starting script i imagined the trick: just rename os.gz and fs.gz and they will be not installed nor deleted. Then I should be able to copy/modify :D fs.gz and reenable the update flags (according the strip) to follow the update.

But where to install the renaming instruction? Well, I did it in the wrong place: :( , I tryed to modify the start script. It is unclear if such modifications survive reboot, and even if they did, it could be that "mv" is not defined in the /bin of the start system, I did not checked it. In any case, the process started, download from iRex was done (I could contemplate it via a sshd connection, the network being enabled) and it rebooted... and installed everything!!!

Now I think the good alternative was to follow exactly the same way that they did: to install a script in the rc0 section and rename the files at this opportunity, when it is doing the shutdown to reboot, just making sure the filesystem was still mounted.

While I was looking the download, I got a partial copy of fs.gz, and it is a compressed ext2, so really it was possible to mount and modify it if I had done it rightly. Damn!

If things are going this way, I vote for begin investigating the OpenEmbedded build process and create an OpenIliad distribution.

I buyed the iLiad for hacking and reading books, but if it's not hackable, there is no fun ;-)

The mayor problem right now is figuring a (safe) way to flash new versions of the firmware.

And that we don't know (yet) the proper method to use /dev/fb0 to draw small regions in the screen (it's used for the scribble function...), and the interface to the wacom tablet. For now we can use the irex kernel, but knowing the interfaces to the iliad specific functions would be very very handy...

And that we don't know (yet) the proper method to use /dev/fb0 to draw small regions in the screen (it's used for the scribble function...). For now we can use the irex kernel, but knowing the interfaces to the iliad specific functions would be very very handy...
Yes if should, and it fact it seems the main contribution of iReX to eink development.

i think this /dev/fb0 could provide for some quite interesting ways of book reading.

like autoscrolling where the new page is continously rendered from the top down so you can read from top down, as usual, and continue reading from top without page-switching break because the first half of the next page is alreay rendered.

i think this /dev/fb0 could provide for some quite interesting ways of book reading.

like autoscrolling where the new page is continously rendered from the top down so you can read from top down, as usual, and continue reading from top without page-switching break because the first half of the next page is alreay rendered.

Yup. That could be a funny way to read.

Perhaps is the time to read the documentation on standard way to access /dev/fb0, and try to do a small program to test it.

Then: there was a way to execute programs in 2.5 or not? (as I upgraded mine...)

well i had a try at catching the ids traffic with a mitm attack via lan/wlan with ettercap.
tried my router and setting my laptop at an open AP but i've given up, ettercap didn't seem to do anything on the router and the iliad wouldn't connect to the laptop :(

but i'm not really sure what i'm doing i've just tried a few examples i found on the net.

It is unclear if such modifications survive reboot, and even if they did, it could be that "mv" is not defined in the /bin of the start system, I did not checked it.
I checked contents, and it was the second: mv was not defined in the /bin/... Damn!

like autoscrolling where the new page is continously rendered from the top down so you can read from top down, as usual, and continue reading from top without page-switching break because the first half of the next page is alreay rendered.

That is actually a very good idea. But I would do it a little different. When you click the flipbar once the upper half of the screen will be updated, when you click it once more the lower half of the screen will be updated.
This way you could also read without interruption.

well i had a try at catching the ids traffic with a mitm attack via lan/wlan with ettercap.
tried my router and setting my laptop at an open AP but i've given up, ettercap didn't seem to do anything on the router and the iliad wouldn't connect to the laptop :(
Good idea, I brought it up earlier (http://www.mobileread.com/forums/showthread.php?t=7091&highlight=ettercap) but it didn't seem that too many people were interested. Let me help you with setting up the ARP spoofing:

Let's assume:

your router has the IP 192.168.0.1 and acts as the gateway
your iLiad has the IP 192.168.0.10
your laptop has the IP 192.168.0.11


Then use the following Ettercap commandline to log all traffic between your router and the iLiad:


ettercap -T -L /tmp/logfile.log -M arp:remote /192.168.0.1/ /192.168.0.10/


If iDS uses a secured connection (over SSL), you'll have to do some extra work to import the certificate into Ettercap.

i'm sure it tries to use https, i fell over at trying to figure out ssl cert spoofing and gave up.

deadite66, check again my thread on sniffing the iLiad (http://www.mobileread.com/forums/showthread.php?t=7091&highlight=ettercap), I just added a step by step guide how sniff the SSL connection.