Bluetooth compatibility fix

Stefan Bader stefan.bader at canonical.com
Fri Feb 27 14:41:26 UTC 2009


Brian Rogers wrote:
> I'm trying to get the following Bluetooth compatibility bug fixed for 
> Intrepid and Jaunty:
> 
> https://bugs.launchpad.net/ubuntu/+source/linux/+bug/268502
> 
> What I've found is that the newer btusb driver never worked for certain 
> dongles, so when it replaced hci_usb in 2.6.27 those dongles no longer 
> worked. I also found that the problem was solved by the commit 
> 'Bluetooth: Submit bulk URBs along with interrupt URBs' in 
> git://git.kernel.org/pub/scm/linux/kernel/git/holtmann/bluetooth-testing.git
> 
> Unfortunately, there are other changes to the file 
> drivers/bluetooth/btusb.c in bluetooth-testing, before the fix, and 
> therefore the fix can't be cleanly cherry-picked by itself. But I did 
> verify that if you pull in the fix and its dependencies on top of either 
> the Intrepid or Jaunty kernel, it gets my problem dongle working.
> 
> Since just taking everything probably wouldn't be acceptable for the 
> Jaunty kernel or an Intrepid SRU, I've been able to pare down the 
> commits to cherry-pick to the following three (with some minor updates 
> and conflict resolution):
> 
> Bluetooth: Add fine grained mem_flags usage to btusb driver
> Bluetooth: Handle bulk URBs in btusb driver from notify callback
> Bluetooth: Submit bulk URBs along with interrupt URBs
> 
> Does this approach sound acceptable? I'll post my pared-down patches 
> after I do some testing. Alternatively, I could try backporting the fix 
> to stand by itself without the two prior changes.
> 
> Either way, the result will be applicable to both the Intrepid and 
> Jaunty kernels since btusb hasn't changed between them.
> 
> Brian
> 
> 

Hi Brian,

Usually the smaller the required change, the better. If it is hard to tune the 
patch to work without the required patches (and since the two patches look 
small and isolated enough to have less chance to regress), this is acceptable 
as well.
However I don't seem to be able to find the patch to solve the bug itself in 
Linus or tip-latest. Has that been accepted upstream already?

Stefan

-- 

When all other means of communication fail, try words!






More information about the kernel-team mailing list