On Thu, 4 Jul 2013, James Stone wrote: > > Another hint: The important lines are the ones containing "Zi:1:002:2" > > -- the actual string may be slightly different, but it will definitely > > contain "Zi" and it will match the " E " lines. > > > > > > OK. So I think this is something else then. There are no E lines, but > the device locks up and becomes unresponsive with the lower (128) > frames/period. Also, I think the performance is degraded vs. the 3.5 > kernel because I am getting xruns with the 256 frames/period setting, > with the device just sitting there not processing any audio. This > doesn't even happen with 64 frames/period with a generic 3.5 kernel. In that case, post what you're got and make sure it includes those "Zi:1:002:2" lines. Best would be if you can find a region of the file where those lines are present at the start but then disappear by the end. By the way, just to be clear, here is an example of some _uninteresting_ "Zi" lines from your recent trace: ffff8800a7c9a600 1731873904 C Zi:1:002:1 0:8:1199:0 1 0:0:4 4 = 08830500 ffff8800a7c9a600 1731873912 S Zi:1:002:1 -115:8:1199 1 -18:0:4 4 < These lines are the synch channel for the audio-out stream; they have no connection with the audio-in even though the "i" in "Zi" indicates an "input" direction of data flow. The distinguishing feature of the synch transfers is the "4" before the "=" and "<" symbols. The real audio-in lines will have much larger numbers there, like 1024 or 2048. Alan Stern -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
