Comment 12 for bug 288613
Revision history for this message
| Jamie Lokier (jamie-shareable) wrote | #12 |
© 2004
Canonical Ltd.
•
Terms of use
•
Data privacy
•
Contact Launchpad Support
•
Blog
•
Careers
•
System status
•
73416ba
(Get the code!)
I'm unable to connect to PAN on my mobile phone (Sony-Ericsson K850i) with the 2.6.27 kernels, but I was able to do so with Hardy's 2.6.24 kernel. I don't do file transfers, only PAN.
To be precise about the kernels: I ran Hardy the distro for a while with Intrepid beta kernels (to test the kernels), and connections made with "pand" were ultimately rejected with "Connection Refused" when running those Intrepid kernels (on a Hardy distro), but when booting a Hardy kernel "pand" worked fine.
It's a bit odd, in that the connection seems to work at first, with the phone's screen asking if it's ok to accept this connection. But on saying ok, then the phone asks a _second_ time the same question, and when it _asks_ the second time, "pand" reports "Connection Refused" immediately. Note that the laptop and phone are already known to each other, this is just the phone asking "Laptop wants to use the phone as a modem? yes or no". If I tell the phone to allow it without asking, there are no dialogs or delays, "pand" just gets to "Connection Refused" very quickly.)
Whereas with Hardy's own kernel, the phone asks just once, and the connection always succeeds in "pand"'s log.
Note the bluetooth userspace utilities were _exactly_ the same in these tests. So in my case, the ability to pair with the phone was affected by booting a different kernel *alone*, with 2.6.27 kernels failing and 2.6.24 kernels succeeding.
I'm now running Intrepid the whole distro (upgraded from Hardy), and generally up to date. The ability to pair (with "pand") is still broken, same as running Hardy with Intrepid's kernel. I haven't yet had a chance to try the other way around and boot the Hardy kernel on my Intrepid distro, to see if that fixes "pand".
So for people with pairing problems, it might be worth booting Hardy's latest kernel, to see if that fixes the problem. If it does, that will at least narrow it down to the kernel bluetooth stack. I suspect there's more than one different pairing problem, but this is one of them.