New Bluetooth Drivers

[Dave L.]

Tumnus,

You are the man  

I've been tearing my hair out trying to get PAN to work with my C860.  Finally came to the conclusion that the bnep module must be the problem (seemed to be no communication between UDP/TCP and L2CAP).  Thinking I'm out of luck, I did a search and here you are with an instant (and very new) solution !  Many Thanks.

By the way, I can't comment on the new suspend/resume - the old one worked fine for me.

By the way(2), any idea how to auto set the IP when bnep port becomes available on connection ?  All the searches I've done come up with a recommendation which doesn't work.

(C-860 Standard ROM, BlueMonkey (LSE039), PANning to XP with D-Link DBT-120)

[mimsmall]

Using BT2000E on 6000L.
Downloaded the driver and changed th (_) to (-) where required.
I got the driver installed and confirmed with hciconfig.

Then I began the connecting to bluetooth enabled phone.  Nokia 3660. It's bluetooth and Irda enabled.

The link to 'Bluetooth Dialup Applet' was broken so I went to the next suggestion ' Instead create a dialup connection with the Bluetooth entry in the Network application and set the rfcomm device to /dev/rfcomm0'

I can't find 'Bluetooth entry in the network application"

When I go to /dev I get a window that says 'can't show /dev directory'.

I have an Irda connection to my phone I use it for transferring files. I use the bluetooth connection to talk to another bluetooth cell phone. But I can't talk to my Zaurus.

[tumnus]

I've corrected the applet links in the Bluetooth howto.

But as for not using the applets, if you read the howto carefully, you create an IrDA network connection and then hack the PPP script to make it use the Bluetooth rfcomm device.

I don't know what you were using to get that error message about /dev, but in the terminal, if you type 'ls -l /dev/rfcomm*' it should list all of the rfcomm devices.

[pakman]

OK, here's a challenge to anyone with a serial based Bluetooth card (such as the Socket Rev G/H):

 - Get the latest ipk working with your card without hacking any default ROM files (i.e. /etc/pcmcia/serial) or at least tell me why using the same serial commands in /etc/pcmcia/bluetooth won't work

No one has reported whether the new susp-resume-bluez ipk works better either...
[div align=\"right\"][{POST_SNAPBACK}][/a][/div]

Hi, I'm back.

I'm going to try the new drivers soon (I'll probably have time to play over the weekend). I've just been looking over some of the old forum posts about this. For me, the show-stopper is described at [a href=\"https://www.oesf.org/forums/index.php?showtopic=2275&st=0&p=13402&#entry13402]https://www.oesf.org/forums/index.php?showt...402&#entry13402[/url] i.e. it is not that the commands don't work when invoked from /etc/pcmcia/bluetooth, but that /etc/pcmcia/bluetooth isn't run at all when the card description is in /etc/pcmcia/bluetooth.conf . The only way around this that I found was to remove the card description from bluetooth.conf, so my Z falls back on /etc/pcmcia/serial. I never managed to find out how to investigate this beyond the obvious/trivial (file permissions, etc.), so if anyone could give me any pointers I would be very grateful.

By the way, just unpacking the IPK and eyeballing your files, I found the following syntax error in /etc/pcmcia/bluetooth:

Code: [Select]

case "$1" in
start,resume)
...
stop,suspend)
should be

Code: [Select]

case "$1" in
start|resume)
...
stop|suspend)
But I don't think that is the problem from the past - the current version of this file on my Z (which would have been the last one I tried) is syntactially OK

Regards,
P.

[tumnus]

/etc/pcmcia/bluetooth isn't run at all when the card description is in /etc/pcmcia/bluetooth.conf .

In the latest bluetooth.conf, you should see a device entry called "serial_cs". Any card entry bound to "serial_cs" will run the script as defined by the class attribute of the device. So the latest script should in theory run /etc/pcmcia/bluetooth for serial cards after binding the serial driver to it.

By the way, just unpacking the IPK and eyeballing your files, I found the following syntax error in /etc/pcmcia/bluetooth:

Code: [Select]

case "$1" in
start,resume)
...
stop,suspend)should be

Code: [Select]

case "$1" in
start|resume)
...
stop|suspend)[div align=\"right\"][a href=\"index.php?act=findpost&pid=64704\"][{POST_SNAPBACK}][/a][/div]

Aha! Thanks for spotting that. Don't know how that got in there. Maybe that's why the script wasn't working  

I'll post a new ipk with the corrections shortly.

[tumnus]

I've corrected the syntax error spotted by pakman and included the hci_usb driver for SL-6000 users with USB dongles. The alpha3 package is now available on the first post in this topic.

[tumnus]

No one with a serial based Bluetooth card tried the alpha3 package yet?

[mimsmall]

I'm in the process. Making the Irda connection a bluetooth connection is my sticking point.

[lwong]

Tumnus,

Thanks for your effort to bring Bluetooth to Zaurus.

I had SL-6000 in stock Sharp ROM and Ambicom (old one) working with the old package. Just got a Socket N CF card lately and try to make it work with the new pacakge but didn;t quite work. The Z went as far as "connecting" (after "initialing modem") but never made the "connected" status. The phone (Siemens S56) also shows they were connecting so I think I am pretty close.

I also misplaced my Ambicom card so I have no way to test if that card will work with the new package. Where should I look to make the Socket card work with the new package?

Thanks

Frank

[lwong]

Ok, it's working now. All I did was eject the Bluetooth card and reinserted again. Work just like before and the card is flush with the Z.

However, the resume is not working. Even tried it one more time and same result. Have to re-insert the card everytime I power the Z off.

Frank

[sornette]

hey tumnus,
Here's another input.
I tried the alpha3 with a Socket revG and the new susp-resume-bluez (0.9.2) on SL-C860.

The alpha3 package seemed fine, but after a suspend I had to eject/insert the card to get it to work again.

Actually, I noticed 2 things when coming back from standby.
1. "hciconfig" would not display any info about the card.
2. cardctl status would show the card as being in "suspend" mode. Issuing a "cardctl resume" changes the status to ready, and both "hcitool dev" and "hciconfig" would now display the card information.
So I added "cardctl resume" inside the apm.d/bluetooth script (outside of "if" statements) and now my card seems to come back up without having to eject it.

I'm not sure what I'm doing, just trying things out. Hope this can make sense to you.

[tumnus]

Does the card automatically eject and reappear when you resume? Look at the CF icon on the taskbar when you resume. If nothing happens to it, did you remember to install the susp-resume and sudo packages too?

[pakman]

Hi Tumnus,

Finally got around to trying the new bluez-zaurus packages, and it works! I deinstalled the old bluez-zaurus, susp-resume-bluez and bluez-csr-serial packages (after carefully making copies of all the files involved so I wouldn't lose my hacks - I used 'ipkg files ...' to help with this).

I installed the new bluez-zaurus package, copied my previous /etc/bluetooth/rfcomm.conf and /etc/bluetooth/hcid.conf back, and everything worked!  (Pairings were still OK because this procedure doesn't touch /etc/bluetooth/link_key.) The name of the card now correctly appears when inserted or when I tap the eject icon. More importantly, when the Z is suspended, I only have to do a 'cardctl resume', and after a second or two, 'hciconfig' shows that the interface is up again - this is a real advance over the previous state of affairs where the interface stayed obstinately down unless I did '/etc/rc.d/init.d/bluetooth stop' followed by '/etc/rc.d/init.d/bluetooth start' . Unlike sornett, I didn't have to do an eject/insert to get this to work.

I haven't tried the new susp-resume scripts yet, because being paranoid I would rather leave the card suspended and explicitly enable it when I need it    My own scripts are pretty trivial now - a one-liner to start using the card, and a couple of lines to shut down any running PAN's before suspending the card when I have finished.

Many thanks to you and Maslovsky for sorting this out,
P.

[sornette]

Does the card automatically eject and reappear when you resume? Look at the CF icon on the taskbar when you resume. If nothing happens to it, did you remember to install the susp-resume and sudo packages too?
[div align=\"right\"][a href=\"index.php?act=findpost&pid=68085\"][{POST_SNAPBACK}][/a][/div]

hello tumnus.
Yes sudo and susp-resume are installed.
With a Socket Rev F card, I observed that the icon disappeared and reappeared.
But with an H/G card, it does not happen.

The script /etc/apm.d/bluetooth has the follwing section

Code: [Select]

if ["'/usr/sbin/hciconfig" != "" ]; then
   ...
fi
I tried commenting out all the resume lines in that script and issuing the commands manually. That's where I noticed that the hcidonfig command returned nothing, which means that the other commands in the if statement won;t be executed either.
The other thing was that cardctl status showed the card as suspended. resuming the card restored everything: hciconfig ok, hcitool ok etc...
that's why I simply added the cardctl resume statement to the apm.d/bluetooth script, and things are now working fine.
shouldn't the card manager take care of resuming the card ?

[tumnus]

sornette, thanks for pinpointing that. maslovsky just told me about that too.

I'll see if I can write some commands on resume that work with all cards. The reason for that if statement on resume was to only run those commands when there is a Bluetooth card inserted or was inserted when your suspended. For the non-serial based cards this seems to work.