Bluetooth mouse breaks on -current

[Eric Anderson]

Maksim Yevmenkin wrote:
> Eric,
> 
>>> Ok, I wish I could provide better information, but as of the last 
>>> couple of weeks (2 weeks maybe?), I've had this weird problem with my 
>>> bluetooth mouse (which had been working flawlessly until now).  Right 
>>> in the middle of use, my mouse will die.  It will lose association 
>>> (or whatever it's called).  Actually, the mouse seems to be ok, as 
>>> I've tried it on another computer and it works just fine with no 
>>> changes.  Once it dies, I have to muck with the bthidcontrol forget, 
>>> reboot, etc, so I'm not sure what I do to fix it, but eventually it 
>>> will be able to query the mouse again and then bthidd will connect 
>>> and everything will be fine.
> 
> 
> very strange... i do not think i committed any significant change to the 
> bluetooth code in a few months.
> 
>>> One thing I've noticed, is that now, when doing an l2ping, I get this:
>>> # l2ping -a mouse
>>> 44 bytes from mouse seq_no=0 time=59990.660 ms result=0x23 Resource 
>>> temporarily unavailable
>>> 44 bytes from mouse seq_no=1 time=59990.834 ms result=0x23 Resource 
>>> temporarily unavailable
>>> 44 bytes from mouse seq_no=2 time=59990.821 ms result=0x23 Resource 
>>> temporarily unavailable
>>> 44 bytes from mouse seq_no=3 time=59990.885 ms result=0x23 Resource 
>>> temporarily unavailable
>>> 44 bytes from mouse seq_no=4 time=59990.879 ms result=0x23 Resource 
>>> temporarily unavailable
>>>
>>> When my mouse is on or off.  Also, pinging other devices, returns 
>>> either the same result, or sometimes it will return ~5000ms returns, 
>>> as if it can ping it, but it is slow - even when the device isn't 
>>> powered on!
> 
> 
> i think its fine. what probably happens here is that bluetooth thinks 
> there is an active baseband connection to the device, but for what ever 
> reason no data is transferred (flow control issues?). l2ping has default 
> timeout of 60 seconds (time=59990.879) after that msleep() call will 
> timeout and return EAGAIN (35 decimal or 0x23)
> 
>>> What do I need to do to help diagnose?
> 
> 
> when this happens please do
> 
> # hccontrol -n ubt0hci read_connection_list
> 
> # l2control -a <bd_addr_of_LOCAL_device> read_connection_list
> 
> # l2control -a <bd_addr_of_LOCAL_device> read_channel_list
> 
> # btsockstat
> 
> and send me the output. note: <bd_addr_of_LOCAL_device> means bd_addr of 
> the device connected to the computer, not mouse.
> 
>> Seems that rebooting my FreeBSD laptop clears it up - so I'm thinking 
>> it has to do with the netgraph/bluetooth modules..
> 
> 
> you probably right. if you have time, you could try to revert your src/ 
> tree to, say, month ago and see if you still have the problem. using the 
> few iterations you could find the date after which your mouse no longer 
> works. this would help me a lot to figure out what change caused this.
> 
> also i would like to know if anybody else has the same issues.

Just letting you know that I haven't had a problem since cvsuping src 
and rebuilding world/kernel..

Eric



-- 
------------------------------------------------------------------------
Eric Anderson        Sr. Systems Administrator        Centaur Technology
Anything that works is better than anything that doesn't.
------------------------------------------------------------------------