[unstable] attempting to bond crashes hcid

[Mercury]

When I attempt to bond (or even start up passkey-agent) according to the instructions on the bluez site ​http://wiki.bluez.org/wiki/HOWTO/Bonding, hcid crashes and must be killed with -9

My log buffer is flooded with the following endlessly:

---

Aug 27 17:16:03 lisa daemon.err hcid[2127]: register_passkey_agent called without any adapter info!
Aug 27 17:16:03 lisa daemon.err hcid[2127]: register_passkey_agent called without any adapter info!
Aug 27 17:16:03 lisa daemon.err hcid[2127]: register_passkey_agent called without any adapter info!
Aug 27 17:16:03 lisa daemon.err hcid[2127]: register_passkey_agent called without any adapter info!
Aug 27 17:16:03 lisa daemon.err hcid[2127]: register_passkey_agent called without any adapter info!
Aug 27 17:16:03 lisa daemon.err hcid[2127]: register_passkey_agent called without any adapter info!
Aug 27 17:16:03 lisa daemon.err hcid[2127]: register_passkey_agent called without any adapter info!
Aug 27 17:16:03 lisa daemon.err hcid[2127]: register_passkey_agent called without any adapter info!
Aug 27 17:16:03 lisa daemon.err hcid[2127]: register_passkey_agent called without any adapter info!

---

I'm using the unstable opkg feed from downloads.openmoko.org

[zecke]

Aeh? My logic scanner failed on crashing AND you need to kill it. If it crashed you don't need to kill it...

Please provide the output of opkg list_installed as of the ​http://wiki.openmoko.org/wiki/Bug_Filing_Policy and then provide instructions on how to reproduce and what you actually do...

Provide enough information to be able to help you...

[Mercury]

Your logic scanner has a more rigid definition of crashing than mine. It's locks up competely.

I also posted a link to a tutorial which has instructions on how to bond with a device. I could enumerate the steps in greater detail if that link isn't sufficient? This should work (Just change over to the unstable feed, run opkg upgrade, and run this.) :

1) make sure hcid is running
2) Run passkey-agent 0000 aa:bb:cc:dd:ee:ff (or any other mac)
3) logread -f <-- flood o' text.

Also I apologize for not sending 'opkg list_installed' At the time I filed this ticket I don't believe that policy was in place. However, perhaps instead of requiring a list of package versions and such it may be better to just instruct people to upgrade to the most recent versions on the feed instead. Since that may actually eliminate bugs before they're reported, and it's a lot easier to type (As I did in the ticket description) 'I'm using the unstable opkg feed from downloads.openmoko.org'

Unfortunately, I have since downgrated my bluez-utils and the problem is gone now, and I don't plan on upgrading again, so if I generated 'opkg list_installed' right now it won't be as useful for you.

[john_lee]

hmm, closed for now. if it's still there please kindly reopen.

[mwester]

Confirming this problem still exists; seems to be an upstream issue that has also been noted by the Debian community:
​http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=501222

The workaround suggested (use --default instead of a specific bt address) works, although admittedly it is a less-than-perfect solution.

[roh]

maintainer change

[Nytowl]

This is a bluez 3.x problem

We are moving forward with Bluez 4