Kindle 3.1 software for Kindle 2, 2i, & DX (under construction)

[pmul]

Yes, I understood that the file was not actually named AT_WAKE_UP, and that that was somewhere there is something like AT_WAKE_UP=/etc/init.d/whatever, or similar. I was unable to locate the file.

[twobob]

ah. yeah, just checking. lol

Perhaps some other kind soul will wade in with their infinite knowledge of "all things 'startup'", that would be nice.

nice find on the WIFIOFF thingy

[pmul]

Thanks! While we're on the subject there is also:
/mnt/us/WANOFF
/mnt/us/DO_NOT_KILL_UART
/mnt/us/WIFION
I won't have time to try your suggestion until the weekend, so stay tuned! Even if I don't find a solution, this has still been fun.

[twobob]

DO_NOT_KILL_UART is for the serial IIRC.

plenty more fun to be had. you should come help us hack prboom together. that would be fun.

hope you get to the bottom of it.

[pmul]

Fun, indeed, I just wish I had more time. I used to do lots of this kind of stuff. My Linux knowledge is a bit rudimentary, but that's a good impetus to do a bit of learning-by-doing I suppose?

[twobob]

monkey see monkey do. That's how I learned. Pretty much everything that was significant was documented by me if I thought it missing or by others if it was already there.

There is a wealth of information on these threads and someone with even a bit of knowledge can definitely get a project done - start to end - in a variety of ways.

Native, QT, SDL, GM's FB work, WAF's, Kindlets, GTK, scripts and a whole scary pile of libraries probably included webkit and others
(In fact Niluje has just written this somewhere else in a similar context IIRC - now I hear myself type it - but meh.)

If you have a a nice set goal, then the tools, info and help are - pretty much - in place.
There are a few esoteric goals that remain elusive but in the main the beast is tamed.

: )

[pmul]

Got some more information. First of all WIFIOFF did not help. Secondly, I have been perusing the showlog log, and the process that stackdumps during a long wake-up is always this (different pid each time obviously):

121005:181152 cmd[20111]: I def:starting::Starting Connection Manager
121005:181152 cmd[20111]: I lipc:gsp:prop=state, source=com.lab126.powerd:Get string property
121005:181152 cmd[20111]: I lipc:gip:prop=driveModeState, source=com.lab126.volumd:Get int property
121005:181152 cmd[20111]: I lipc:gsp:prop=cmState, source=com.lab126.wan:Get string property
121005:181152 cmd[20111]: I def:prev_status_init:wan:Initializing prev_status for interface
121005:181152 cmd[20111]: I def:intfna::No connected network interface
121005:181152 cmd[20111]: I lipc:gip:prop=isCharging, source=com.lab126.powerd:Get int property
121005:181152 cmd[20111]: I def:started::Connection Manager started

A bit further down, I can see this:

121005:233739 powerd[2110]: I def:resume::System resumed
121005:233739 cmd[20111]: I def:sysactive::System Resuming
121005:233739 cmd[20111]: I lipc:gsp:prop=state, source=com.lab126.powerd:Get string property
121005:233739 cvm[3427]: I LipcService:EventArrived:source=com.lab126.powerd,name=resuming,arg0=<None>,arg1=<None>:
121005:233739 cvm[3427]: I PowerImpl:Information::handlePowerStateChangeEvent 7 0
121005:233739 cvm[3427]: I PrefixFreeVolume:ForceRescan:path=/mnt/us/documents:
121005:233739 cvm[3427]: I PrefixFreeVolume:ForceRescan:path=/mnt/us/audible:
121005:233739 cvm[3427]: I PrefixFreeVolume:ForceRescan:path=/mnt/us/pictures:
121005:233739 cvm[3427]: I PrefixFreeVolume:ForceRescan:path=/opt/amazon/dict:
121005:233739 cvm[3427]: I lipc:gsp:prop=activeInterface, source=com.lab126.cmd:Get string property
121005:233739 powerd[2110]: I lipc:evts:name=resuming, origin=com.lab126.powerd:Event sent
121005:233739 powerd[2110]: I def:statech:prev=SUSPENDED, next=SCREEN SAVER:State change: SUSPENDED -> SCREEN SAVER
121005:233739 powerd[2110]: I def:battinfo:cap=27%, mAh=422mAh, volt=3738mV, current=-81mA, temp=75F, bp=3740mV, lmd=2110mAh, cycl=-1091302392, cyct=1:
121005:233739 powerd[2110]: I lipc:evts:name=battLevelChanged, origin=com.lab126.powerd, fparam=27:Event sent
121005:233739 cvm[3427]: I HTTPUtil:clearDNSCache:cache_size=0:clearing DNS cacheMap
121005:233740 system: I dump-stack:dumpstack:cmd:pid=20111:begin native threads ==============================

So it's the "Connection Manager", ring any bells?

[twobob]

okay. have you done a shutdown -r now since you fiddled with the upstart scripts?
any chance they are cached? I have no idea.

Connection manager sounds a lot like a wifi thing. I would have to dig around in my 3 to check I suppose. perhaps killing some likely processes could give me some helpful log output.

like
(you would want a seperate shell for this ideally)
tail -f /var/log/messages &

and then go killing stuff in the other shell
See what process names spawns that log


or grep -r "Connection Manager" /var (and other places) to see if you can dig up some log configs...(This may hang your DX badly)

Those would be my tests. in the absence of better info.

[knc1]

Ah, a better place to start:
/etc/network/interfaces.
Remove the entries for the non-existent devices (such as; wlan0)

And/Or
View /etc/inittab
note that the default runlevel is 2

Next, /etc/rc2.d
you will find (if this is 3.2.1) there isn't any S##network entry

Next, /etc/rcS.d
there you will find a link (S10network) to ../init.d/network
After reading the script (via the /etc/rcS.d/S10network link) decide if you want to modify it.

Note: The above is firmware version specific (3.2.1) newer firmwares use a different initialization system.

Smaller note: The reader will notice that script already has conditionals for running the kernel with the rootfs provided by a NFS server (significant for another thread here).

zzzzz.....
as contributor goes back to sleep.

[pmul]

Tried all of the above (from knc1), the only thing I found was more references to the WIIFIOFF file. I rebooted to make sure that the WIFIOFF file was being seen at the the appropriate moments, alas no no avail. I will try to carry out twobob's suggestions at the weekend.

[alex2345]

Reading discussion, I am in great hope that your efforts will make possible the update 3.4 for Kindle DX! This update is promising to improve fonts making them more contrast and more crisp.

[knc1]

Originally Posted by alex2345

Reading discussion, I am in great hope that your efforts will make possible the update 3.4 for Kindle DX! This update is promising to improve fonts making them more contrast and more crisp.

You should only have to move (install) the font files, not the entire firmware installation.

[twobob]

Originally Posted by pmul

I will try to carry out twobob's suggestions at the weekend.

Gah. I hope it bears you fruit mate.

[pmul]

A little bit further:
There does not seem to be a "wake up script". A daemon called "powerd" reacts to the power switch and triggers other sleeping processes to wake up. One of them is the "Connection Manager", which, as it's last act, performs this:

121005:181134 cmd[19772]: I lipc:gsp:prop=state, source=com.lab126.powerd:Get string property

Which seems to be the cause of the exception, some property is requested, maybe one that does not exist in "Connection Manager" as running on my DXG, and this leads to the exception. I believe I need the help of someone a bit more expert in the Java side of the Kindle OS, com.lab126.powerd looks like a typical Java class name, complete with package name.

[pmul]

Fascinating!
Sorry if this is well known, but I just discovered this: in /usr/bin/ are native executables that allow one to get/set properties send events, and probably more.
/usr/bin/lipc-get-prop
/usr/bin/lipc-send-event
/usr/bin/lipc-daemon
/usr/bin/lipc-probe
/usr/bin/lipc-wait-event
/usr/bin/lipc-set-prop
/usr/bin/lipc-hash-prop
And I've been playing with them. Contrary to my previous thoughts, it's this:

121005:181135 lipc-send-event[19940]: I lipc:evts:name=orientationFaceUp, origin=com.lab126.hal:Event sent

that causes "Connection Manger" to crash. Possibly because 'orientationFaceUp' means nothing to the some part of the K3's firmware, it being a DXG specific name, given that the DXG has an accelerometer and the K3 does not? This looks like a promising path.....
Now if only I could find the files representing "Connection Manager"