Re: N900 as a Wifi Hotspot, part 2
 

That's why I'm asking the question. If I knew the answer I wouldn't be asking. Joikuspot was working fine, the next day I installed Mobile Hotspot and auto-disconnect and it stopped working. Both were immediately removed and Joikuspot re-installed and now I have not been able to get it working since. I don't know what could have screwed things up.

As for the kernel comment I have no idea what if anything the kernel has to do with it. Someone gave a suggestion to check it and I did. I'm open to all suggestions.

Re: N900 as a Wifi Hotspot, part 2
 

I didn't know I had installed a titan kernel. I tried it and it didn't make any difference. I'll probably just reflash my phone and start from scratch.

Re: N900 as a Wifi Hotspot, part 2
 

Ok, this explains a lot. Titans kernel got installed because mobilehotspot depends on kernel-feature-netfilter (which titans kernel provides). Reinstall stock kernel and you should be good to go (uninstalling just mobilehotspot will not uninstall titans kernel and I'm not 100% titans kernel automatically reflashes stock kernel after uninstall [which would be kinda smart to do]).

I kinda suppose (yes, supposion is the mother of all f*ckups) that people testing stuff from extras-devel seeing the package description (where it's stated that a custom kernel is required and might be installed automatically) know what it means.

edit: Do you mean reflashing stock kernel made no difference to Joikuspot ? Interesting but can't really help with that.

Re: N900 as a Wifi Hotspot, part 2
 

Just wanted to tell how impressed I am. I was a more-or-less happy user from Joikuspot. There were issues with not connection linux boxes, extensive CPU usage (far more than seemed needed, even in the lo-fi client screen) but all got worse during my OC kernel trials.

I couldn't correctly get it to run anymore. So I decided to give mobilehotspot a chance and I must say : amazing. very little config (just enough for me), using the latest Titan kernel (power25) and it just works.

Very low overhead, no compatibility issues so far (eg. DHCP server not working), straight connection from my linux box ... Only thing I might like is a sort of 'connected clients' screen, to monitor a bit what's happening. low-fi is ok, CPU cycles sometimes quite scarce :)

Anyway, one word : super !! I don't think I'll go back to try to get Joikuspot working again unless I REALLY have to ... I just hope whenever PR1.2 comes out, Titan will update his kernel ASAP. I'm getting addicted to these enhanced community kernels :)

Re: N900 as a Wifi Hotspot, part 2
 

AFAIK Titan's Kernel is already based on PR1.2 source tree so it's PR1.2 kernel with enhancements.

EDIT: My only concern is that WEP seems to always be off even if I activate it via the frontend. If someone have a solution for that, I'll be glad to read it.

Re: N900 as a Wifi Hotspot, part 2
 

Post screenshots of the config (remember to change the key to something to style of "example" first), then if you know how to read gconf data from the python shell read values

/apps/mobilehotspot/encryption_algo
/apps/mobilehotspot/encryption_key

and post them.

Then from terminal run hotspot_backend manually (via sudo) and post output.

Re: N900 as a Wifi Hotspot, part 2
 

Here is what I get:
The strange part is that 'iwconfig wlan0 key restricted' is executed (I verified it wasn't here for an open connection)

Re: N900 as a Wifi Hotspot, part 2
 

@rambo: this is a simple bug in helpers.hex_encode… there is
"for char in output_str:" when is should say "for char in input_str:"

Re: N900 as a Wifi Hotspot, part 2
 

Figured as much as soon as I saw that 'iwconfig wlan0 key restricted' , but I'm sure I fixed that typo "somewhere", it seems I never committed that change...

I'll be rolling a new package later today, meanwhile get r168 of helpers.py and overwrite the old one (or just use vi to fix that line...)

Edit: package built by autobuilder has landed to the repo.

Re: N900 as a Wifi Hotspot, part 2
 

The "hudge cpu consumption" complaints of Joiku makes me think that they do screen updates even when background which would be a monumental mistake, of course doing NAT in userspace is an overhead but I have no idea of the magnitude (yet... we'll be using weks approach for that with mobilehotspot soonish)

I have been thinking about that too and maybe some sort of simple data counter. But since both of those you can check from the terminal (ifconfig and iwconfig are your friends) if you're really interested I think I won't be adding them anytime soon (feel free to submit patches though).

Another idea I have floated around in the back of my head would be to add a led-pattern that the hotspot backend would use when active (this is one reason why I wrote mceledpattern)