From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.12) Gecko/20050922 Fedora/1.0.7-1.1.fc4 Firefox/1.0.7 Description of problem: After upgrading to the latest kernel BT uplink, ie. sound input, gets distorted Version-Release number of selected component (if applicable): kernel-2.6.13-1.1526_FC4 How reproducible: Always Steps to Reproduce: 1. Fetch bluetooth-alsa from sf.net CVS 2. Compile & install kernel module and btsco user space daemon 3. Plugin USB BT dongle and start BT service 4. Load kernel module and start btsco daemon to create connection to BT headset 5. Make a call to an echo server, e.g. with kphone Actual Results: Sound on downlink is OK Sound recording gets distorted, i.e. the echo comes back disrupted by pauses and long delays. Sounds like it gets recorded in pieces and between each piece a fixed pause gets inserted. Expected Results: Sound OK on both up & downlink Additional info: I know bluetooth-alsa is still alpha but it worked OK on every FC4 kernel before the 2.6.13 rebase. Here is a list of all involved programs/subsystems: [kphone] [alsa-lib] ALSA kernel stack [snd-bt-sco module (provides the "sound card")] [btsco (bridges between snd-bt-sco.ko and BT stack)] [bluez-libs] BT kernel stack hci_usb module USB kernel stack uhci_hcd module [USB 1.1 BT dongle] [BT Headset] Everything in [] is unchanged from before the kernel upgrade.
Please report this to the upstream maintainers of that module.
Just for the fun of it I retested the new kernel with a normal USB Headset: Same problem. So it's not related to bluetooth-alsa, it's the kernel -> REOPENED. Based on the retest we can now narrow it down to the following subcomponents: [kphone] [alsa-lib] ALSA kernel stack [snd_usb_audio module (provides the "sound card")] USB kernel stack uhci_hcd module [USB 1.1 Headset]
Retested with kernel-2.6.13-1.1532_FC4: Still the same problem.
Mass update to all FC4 bugs: An update has been released (2.6.14-1.1637_FC4) which rebases to a new upstream kernel (2.6.13.2). As there were ~3500 changes upstream between this and the previous kernel, it's possible your bug has been fixed already. Please retest with this update, and update this bug if necessary. Thanks.
Retested with kernel-2.6.14-1.1637_FC4: Still the same problem.
It maybe that only kphone is affected by this. A simple recording loop arecord -Dplughw:2 | aplay -Dplughw:2 doesn't have this problem. That could mean that the actual problem might be network stack related, ie. delayed or dropped packets.
Retested with kernel-2.6.14-1.1644_FC4: Still the same problem. Recompilation of kphone didn't help either, as I expected...
Found another VoIP/SIP application to test with: Twinkle (http://www.twinklephone.com). It exhibits the same problem as kphone under kernel >2.6.12.
In light of my test results I've updated the summary again. I've now compiled a plain 2.6.14.3 kernel using the config from the kernel-2.6.14-1.1644_FC4: this kernel does *not* exhibit the problem. So it must be caused by a Fedora patch.
[LATER] Argggg... Now twinkle works OK on kernel-2.6.14-1.1644_FC4. Don't ask me why. It didn't when I first tested it and exhibited the same problem as kphone. I shutdown the computer completely and removed the power cord, but somehow I can't reproduce the error with twinkle anymore :-[ kphone on the other hand has the problem also on the plain 2.6.14.3 kernel with FC4 config. So it's something the kernel does which makes kphone stumble on sound input. Maybe kphone does something stupid which worked OK until 2.6.12... I'll resolve the bug as WORKSFORME for now as I have a working alternative.