[regression] Apple Magic Mouse stalls

[regression] Apple Magic Mouse stalls

[Ed Tomlinson]

Hi

I've been having fun with bluetooth in .38-rc.  My adapter was old (bt 1.1) and the bluetooth stack was having 'fun' with it.
Got tired of that a got asus doggle that supports 2.1+EDR.  Then I tried to use my magic mouse.  It works for a few
minutes and then stalls.  The stalls seem to be related to touches.  If I am careful only to touch with one finger it
seems to work.  It usually works with two fingers, with three it stalls.   Running hcidump sometimes lists a last 
packet with a dlen of 35 when it stalls.  Powering it on an off reset things 

> ACL data: handle 21 flags 0x02 dlen 27
    L2CAP(d): cid 0x0041 len 23 [psm 0]
> ACL data: handle 21 flags 0x02 dlen 27
    L2CAP(d): cid 0x0041 len 23 [psm 0]
> ACL data: handle 21 flags 0x02 dlen 27
    L2CAP(d): cid 0x0041 len 23 [psm 0]
> ACL data: handle 21 flags 0x02 dlen 35
    L2CAP(d): cid 0x0041 len 31 [psm 0]

This was not a problem pre .38.  It started happening in .38-rc along with other bluetooth problems.  I figured the bt
problem was the root cause - I was wrong.

This ring any bells?

TIA
Ed Tomlinson

Re: [regression] Apple Magic Mouse stalls

[Chase Douglas]

On 05/28/2011 10:06 AM, Ed Tomlinson wrote:
> Hi
> 
> I've been having fun with bluetooth in .38-rc.  My adapter was old (bt 1.1) and the bluetooth stack was having 'fun' with it.
> Got tired of that a got asus doggle that supports 2.1+EDR.  Then I tried to use my magic mouse.  It works for a few
> minutes and then stalls.  The stalls seem to be related to touches.  If I am careful only to touch with one finger it
> seems to work.  It usually works with two fingers, with three it stalls.   Running hcidump sometimes lists a last 
> packet with a dlen of 35 when it stalls.  Powering it on an off reset things 
> 
>> ACL data: handle 21 flags 0x02 dlen 27
>     L2CAP(d): cid 0x0041 len 23 [psm 0]
>> ACL data: handle 21 flags 0x02 dlen 27
>     L2CAP(d): cid 0x0041 len 23 [psm 0]
>> ACL data: handle 21 flags 0x02 dlen 27
>     L2CAP(d): cid 0x0041 len 23 [psm 0]
>> ACL data: handle 21 flags 0x02 dlen 35
>     L2CAP(d): cid 0x0041 len 31 [psm 0]
> 
> This was not a problem pre .38.  It started happening in .38-rc along with other bluetooth problems.  I figured the bt
> problem was the root cause - I was wrong.
> 
> This ring any bells?

Not for me... Nothing has changed in the driver that could cause such a
change, so if there is a regression it exists somewhere else in the stack.

Jiri, any thoughts?

-- Chase