FS#9627 - Enable new bluetooth driver to reduce wakeups

Attached to Project: Arch Linux
Opened by Dennis Brendel (buddabrod) - Wednesday, 20 February 2008, 10:33 GMT
Last edited by Tobias Powalowski (tpowa) - Monday, 03 March 2008, 22:22 GMT
Task Type Feature Request
Category Kernel
Status Closed
Assigned To Tobias Powalowski (tpowa)
Architecture All
Severity Low
Priority Normal
Reported Version 2007.08-2
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 1
Private No

Details

On the intel power mailing list this was talked about:

Peter F. Patel-Schneider wrote:
> Hi:
>
> One of the suggestions that Powertop makes is to turn off bluetooth.
> That gets rid of the interrupts from the bluetooth device but at the
> cost of removing bluetooth functionality. Is there any effort to fix
> the cause and get a bluetooth system that doesn't produce 100 interrupts
> per second even when no devices are connected?
>

yes just enable the CONFIG_BT_HCIBTUSB config option; that's a rewritten, much
improved driver that doesn't have this problem.
(you may need to disable CONFIG_BT_HCIUSB to see this option)

It would be nice to enable the new driver in order to massively reduce wakeups and power consumption.
This task depends upon

Closed by  Tobias Powalowski (tpowa)
Monday, 03 March 2008, 22:22 GMT
Reason for closing:  Won't implement
Comment by Tom Killian (tomk) - Wednesday, 20 February 2008, 11:32 GMT
In time, this will probably be a better driver, but it is still experimental, and it has no SCO support yet. It cannot be provided alongside the current HCI USB driver, so enabling it would mean disabling SCO.
Comment by Dennis Brendel (buddabrod) - Wednesday, 20 February 2008, 11:49 GMT
Isn't it possible to build both drivers as modules, independent from each other and leave the user the choice which driver he wants?
Or is that too complex?
Comment by Tobias Powalowski (tpowa) - Wednesday, 20 February 2008, 17:54 GMT
if you don't see both options, it's not possible
Comment by Dennis Brendel (buddabrod) - Wednesday, 20 February 2008, 20:06 GMT
I meant building them totally independant from each other (maybe by making a separate kernel package which might be overkill for this small enhancement) in order to just create a package which ships the modules only. Might be overkill as well since the new driver is maturing very fast and will then replace the old thing.
Comment by Tom Killian (tomk) - Thursday, 21 February 2008, 01:21 GMT
I'd suggest that you build a replacement kernel package with BT_HCIUSB=n and BT_HCIBTUSB=m. I'm doing this myself at the moment, as I want to see how this driver is doing. SCO isn't critical for me, so that's OK, but we can't assume that's the case for all Arch users.
Comment by Tobias Powalowski (tpowa) - Monday, 03 March 2008, 22:22 GMT
until the new subsystem get's default we stay with old subsystem

Loading...