ASUS WL-700gE custom firmware released - (1.0.4.2-1)

[MoD]

If you could do a release with ipkg and uclibc built in, that'd be great

The only package I am personally after is a working bittorrent client - not sure how much better enhanced-ctorrent is than normal ctorrent, but if it handles the download stability issues from the original asus firmware, then it's a definite improvement!

I would appreciate some help and advice in doing this everybody is waiting for custom firmware. lets share what we have done so far. i see in this forum that not only me is working on this.

For example I do not know why does telnet has bad responsiveness, I do not know why shell scripts are not executed when you create them, but executes if you copy and edit from existing one.


personally I see a lot of things how to use ipkg - dp500serv for kiss DVD player, home automation, asterix pbx etc...

[Mpu]

The reason why I choose GPL 1.0.4.2 was that GPL_WL700gE_1042.zip has 602M of source code when GPL_WL700gE_1046.zip has only 391M. I'm not sure how much of source code is there. Definitelly there is no router startup code in directory rc.asus witch I needed to create changes that are in my release of firmware.

I guess that it is possible to use latest GPL_WL700gE_1046.zip code and take the missing code from GPL_WL700gE_1042.zip, but I'm not sure how stable would be such a firmware. I'm expecting to get some feedback from my release to decide what to do next. If ASUS will not prowide new code, I would use this aproach in the future.

There is good change that rc.asus original source files are same in GPL_WL700gE_1042 and GPL_WL700gE_1046/GPL_WL700gE_1041, because rc.asus/rc binaries in 1.0.4.1 and 1.0.4.6 are same size(and diff cant find any differeces).

So maybe its safe to use rc.asus-1.0.4.2 files in latest GPL_WL700gE_1046 code?

[d3viant]

So maybe its safe to use rc.asus-1.0.4.2 files in latest GPL_WL700gE_1046 code?

I am willing to test that if somebody releases something with combined 1042 and 1046 files - just don't have the know-how to compile it myself

MoD - I haven't had any trouble with response times through telnet to the router (although SSH support would be probably faster and more secure) The shell script thing is probably permissions related - it may have to have the user and group ID of the routers root user to be executed correctly - it does work, but it cannot set path variables by the looks of it (I am guessing because the router executes it in a seperate shell session, so the variables don't get carried across...)

[MoD]

There is good change that rc.asus original source files are same in GPL_WL700gE_1042 and GPL_WL700gE_1046/GPL_WL700gE_1041, because rc.asus/rc binaries in 1.0.4.1 and 1.0.4.6 are same size(and diff cant find any differeces).

So maybe its safe to use rc.asus-1.0.4.2 files in latest GPL_WL700gE_1046 code?

first I have to figure out how to implement all custom features in one GPL source. The next step could me to mix GPL sources. I already wrote about it.

[d3viant]

first I have to figure out how to implement all custom features in one GPL source. The next step could me to mix GPL sources. I already wrote about it.

Theoretically, can't you just do the following:

1. unzip the 1042 source
2. unzip the 1046 source, then copy it to the 1042 tree, overwriting anything it asks about
3. Put in the changes you have made to the firmware to add telnet support and the startup script nvram variable - obviously into the newer version of the file - so that you don't mess up and new features in 1046 (if there is a newer version of those files that is...)

[MoD]

Theoretically, can't you just do the following:

1. unzip the 1042 source
2. unzip the 1046 source, then copy it to the 1042 tree, overwriting anything it asks about
3. Put in the changes you have made to the firmware to add telnet support and the startup script nvram variable - obviously into the newer version of the file - so that you don't mess up and new features in 1046 (if there is a newer version of those files that is...)

..and theoretically there can be problems. When you are programming, there are always errors. Before start to deal with problems that can arise when you mix different GPL sources, I want to create really usefull customized firmware with following features:

1. password protected shell
2. implemented ipkg system
3. startup script that could launch ipkg installed applications

that is my first goal.

Later I can use my expierence to mix GPL sources to create customized firmware that fixes ASUS bugs. Plz don't ask to mix GPL code yet. Right now my customized firmware maybe is not so usefull but it is created to test if customized features works. I released it to rise discussion and to get some feedback if anyone is testing it. If it is usefull for anyone, it is welcome.

We are on a way to create usefull customized firmware. I'm posting information I have on this forum even before I have implemented it in my code. I'm doing this because I hope to get support/help/futher directions. I'm getting it. Thanks to all who are responding!

[d3viant]

Sorry MoD - didn't mean to seem ungrateful - your modified firmware is already an improvement over the original ASUS one - I have been able to logon and customise iptables properly

If you need anything testing, I am sure plenty of people here will be willing to test it (I will be at least )

[MoD]

Sorry MoD - didn't mean to seem ungrateful - your modified firmware is already an improvement over the original ASUS one - I have been able to logon and customise iptables properly

If you need anything testing, I am sure plenty of people here will be willing to test it (I will be at least )

never mind, d3viant. i count on you!

[fireflash]

why not just tag a startup script onto the rcex file in 1.0.4.6? Seems like a good way to go and you can run anything from there.

[MoD]

why not just tag a startup script onto the rcex file in 1.0.4.6? Seems like a good way to go and you can run anything from there.

Yes, maybe this is a better aproach. I see that there is a big progress in doing this (by kfurge): http://wl500g.info/showpost.php?p=40489&postcount=27

[kfurge]

For example I do not know why does telnet has bad responsiveness, I do not know why shell scripts are not executed when you create them, but executes if you copy and edit from existing one.

I'm not seeing any problems with script execution.

A few things to check... Did you set the 'x' bit (i.e. chmod 755 scriptname)? Also, does your script start with the "#!/bin/sh" or "#!/bin/ash" magic? Did you try running your script by executing '/bin/ash scriptname'?

I've enabled utelnetd without password support and I see no delay whatsoever. If I remember correctly, you enabled password logins. Can you describe how you did that? I'll try to replicate the problem at my place.

Also, generically speaking, when the initial socket connection to the telnet port is fast, but the login prompt takes a long time to show, it's been my experience that the problem is usually an issue with the reverse DNS lookup of the client's IP address. Expecially if the IP address comes from an internal network.

- K.C.

[fireflash]

Yes, maybe this is a better aproach. I see that there is a big progress in doing this (by kfurge): http://wl500g.info/showpost.php?p=40489&postcount=27

Not really a big process. I had the files and such modified and nvram variable loaded up in no time. I've compiled my firmware with the mods, but I haven't had a chance to flash it and see how it works yet. I think before I do that I'm gonna try and sneak in a working fstab file like kfurge did.

I did, however, manually run the commands kfurge mentioned in an earlier post to get the HD to spin down, but it never did. The only thing I couldn't run was the mount command because it was looking for a non-existant fstab file. So I think that's my next move before I take another stab at bricking my router.

[d3viant]

fireflash - I managed to get my hard drive to spin down by modifying kfurges instructions slightly - due to the lack of the fstab, you have to define where to do the remount to:

killall -q -9 watchdog
mount -o remount,noatime,nodiratime /dev/se/2 /shares/MYVOLUME1
hdparm -S 240 /dev/ide/host2/bus0/target0/lun0/disc
smartctl -dT /dev/ide/host2/bus0/target0/lun0/disc


Change the time after -S to whatever you want to set the time before it spins down to (240 is 20minutes I believe)

I have these commands running on the rcinit startup script in MoD's firmware, so if the router gets rebooted for some reason, then the hard drive parameters still apply.

Now I have to figure out how to turn on/off the LEDS on the front - so I can sleep with it on

[fireflash]

fireflash - I managed to get my hard drive to spin down by modifying kfurges instructions slightly - due to the lack of the fstab, you have to define where to do the remount to:

killall -q -9 watchdog
mount -o remount,noatime,nodiratime /dev/se/2 /shares/MYVOLUME1
hdparm -S 240 /dev/ide/host2/bus0/target0/lun0/disc
smartctl -dT /dev/ide/host2/bus0/target0/lun0/disc


Change the time after -S to whatever you want to set the time before it spins down to (240 is 20minutes I believe)

I have these commands running on the rcinit startup script in MoD's firmware, so if the router gets rebooted for some reason, then the hard drive parameters still apply.

Now I have to figure out how to turn on/off the LEDS on the front - so I can sleep with it on

I think i see where I goofed. On the mount command I'm sure I forgot a space between /dev/se/2 and /shares/MYVOLUME1

I'll try it when I get home later, and if it works I'll finish adding the chances to the rcex file and re-compile, and see if there's smoke when it flashes.

Now to turn the lights off... I think most of the lights are GPIO triggered, so it may be possible to trigger the GPIO's to turn them off.. But that requires digging inside a whole other part of the firmware that controls the lights and such. If it could be tricked to emulate a shutdown state without actually turning it off, then you're good to go. Otherwise do what I do, strategically place your chair between your bed and router so you don't get the lights shining directly at you. A lot of times I turn it off at night also, to save the HD and conserve a little power.. But the only device I have that'll access the internet at a time where I may not be home is my TiVo, but it can live without a day or two of net access.

EDIT: I found another reason why my HD wouldn't shut down the last time I tried it. It would seem I left 2 downloads queued up, so ctorrent was running causing disk access. It got real quiet when I killed those, and it even shut down. Kinda scared me when it went down lol

[d3viant]

I was hoping that the lights would be controlled from somewhere within the linux /proc structure - but thats a pure guess

If theyre in there somewhere, its just a case of finding them and changing a one to a zero in the file or something. Otherwise it becomes harder as a dig through the firmware source code is required. It is ok for the time being as the torrent client doesn't work amazingly well - so it isnt downloading anything (so I turn it off at night) but once downloads work properly, I will be wanting to leave it running :P