after abort there are STILL 126 active TDs

[Wouter1]

After turning off the EMU device, you get (a.o.) this.

The message 'after abort there are STILL 126 active TDs' seems not ok.

Nov 27 19:18:10 vlieland kernel[0]: +EMUUSBAudioEngine[0xffffff80352f3800]::stopUSBStream ()
Nov 27 19:18:10 vlieland kernel[0]: USBF:    577.186    +AppleUSBXHCI[0xffffff81a8844000]::AbortIsochEP[0xffffff802ceba000] - done - pEP->outSlot (0x58) pEP->inSlot (0x81) activeTDs (126) onToDoList (0) todo (0) deferredTDs (0) deferred(0) scheduledTDs (126) onProducerQ (0) consumer (6233) producer (6233) onReversedList (0) onDoneQueue (0)  doneQueue (0)
Nov 27 19:18:10 --- last message repeated 1 time ---
Nov 27 19:18:10 vlieland kernel[0]: USBF:    577.187    +AppleUSBXHCI[0xffffff81a8844000]::AbortIsochEP[0xffffff802ceba000] - done - pEP->outSlot (0x58) pEP->inSlot (0x81) activeTDs (126) onToDoList (0) todo (0) deferredTDs (0) deferred(0) scheduledTDs (126) onProducerQ (0) consumer (6233) producer (6233) onReversedList (0) onDoneQueue (0)  doneQueue (0)
Nov 27 19:18:10 vlieland kernel[0]: USBF:    577.187    AppleUSBEHCI::DeleteIsochEP- after abort there are STILL 126 active TDs
Nov 27 19:18:10 vlieland kernel[0]: -EMUUSBAudioEngine[0xffffff80352f3800]::stopUSBStream ()

[Wouter1]

Might be related to #7.

[Wouter1]

When you stop the audio sampling you get this

Nov 29 12:38:00 -- kernel[0]: + readCompleted framelist 1  result e00002e7 frametime -548796683872 
Nov 29 12:38:00 -- kernel[0]: + read frameList 1 
Nov 29 12:38:00 -- kernel[0]: - readCompleted currentFrameList=2
Nov 29 12:38:00 -- kernel[0]: + readCompleted framelist 2  result e00002e7 frametime -548796683872 
Nov 29 12:38:00 -- kernel[0]: ****** OVERRUN: write head is overtaking read head!
Nov 29 12:38:00 -- kernel[0]: + read frameList 2 
Nov 29 12:38:00 -- kernel[0]: - readCompleted currentFrameList=3
Nov 29 12:38:00 -- kernel[0]: +EMUUSBAudioEngine[0xffffff80392b2800]::stopUSBStream ()
Nov 29 12:38:00 -- kernel[0]: + readCompleted framelist 3  result e00002eb frametime -548796683872 
Nov 29 12:38:00 -- kernel[0]: ++EMUUSBAudioEngine::readCompleted() - stopping: 1
Nov 29 12:38:00 -- kernel[0]: - readCompleted currentFrameList=3
Nov 29 12:38:00 -- kernel[0]: + readCompleted framelist 3  result e00002eb frametime -548796683872 
Nov 29 12:38:00 -- kernel[0]: ++EMUUSBAudioEngine::readCompleted() - stopping: 2
Nov 29 12:38:00 -- kernel[0]: - readCompleted currentFrameList=3
Nov 29 12:38:00 -- kernel[0]: USBF:    2801.912    +AppleUSBXHCI[0xffffff81b1b69000]::AbortIsochEP[0xffffff803f411000] - done - pEP->outSlot (0x58) pEP->inSlot (0x81) activeTDs (126) onToDoList (0) todo (0) deferredTDs (0) deferred(0) scheduledTDs (126) onProducerQ (0) consumer (4569) producer (4569) onReversedList (0) onDoneQueue (0)  doneQueue (0)
Nov 29 12:38:00 --- last message repeated 1 time ---
Nov 29 12:38:00 -- kernel[0]: USBF:    2801.913    +AppleUSBXHCI[0xffffff81b1b69000]::AbortIsochEP[0xffffff803f411000] - done - pEP->outSlot (0x58) pEP->inSlot (0x81) activeTDs (126) onToDoList (0) todo (0) deferredTDs (0) deferred(0) scheduledTDs (126) onProducerQ (0) consumer (4569) producer (4569) onReversedList (0) onDoneQueue (0)  doneQueue (0)
Nov 29 12:38:00 -- kernel[0]: USBF:    2801.913    AppleUSBEHCI::DeleteIsochEP- after abort there are STILL 126 active TDs
Nov 29 12:38:00 -- kernel[0]: -EMUUSBAudioEngine[0xffffff80392b2800]::stopUSBStream ()

[Wouter1]

Looks like we are stopping only 2 of the 4 USB input streams.

[Wouter1]

fixed by sleeping 1 second before forced-release of USB stream. Still a bit hacky, 1 second seems bit overkill. maybe we can do this in a nicer way

[Wouter1]

BTW we now also wait for all 4 input streams to close. The 1 second is kind of fallback now.