On Sat, 29 Sep 2012, Lassi Väätämöinen wrote: > Ok, I got a new set of traces in Launchpad, with the added usb device > list, which I forgot to post last time: > https://bugs.launchpad.net/ubuntustudio/+bug/1035602/comments/5 > https://bugs.launchpad.net/ubuntustudio/+bug/1035602/comments/6 > https://bugs.launchpad.net/ubuntustudio/+bug/1035602/comments/7 > > I took a quick peek myself to get the hang of it. There seems to be a > ECONNRESET (-104) coming in from somewhere, and then it stops. Otherwise > I just see -115 (EINPROGRESS), which is normal behavior, right?. Yes. But what isn't so apparent is that things are missing! Here's the first place where that happens: ffff8801c9816700 2022630636 C Zi:4:003:5 0:1:43924:0 1 0:0:264 294 = b1feff47 feffc9fe ffc4ffff bbffff81 0300beff ff1dfdff 930000b6 fdfff5fe ffff8801c9816700 2022630641 S Zi:4:003:5 -115:1:43924 1 -18:0:294 294 < ffff8801c9816d00 2022631634 C Zi:4:003:5 0:1:43925:0 1 0:0:264 294 = 9b000041 02000d00 007a0100 ffffff66 ffffed00 00b9fcff 74ffff86 ffffb0ff ffff8801c9816d00 2022631642 S Zi:4:003:5 -115:1:43925 1 -18:0:294 294 < ffff8801c9816e00 2022632635 C Zi:4:003:5 0:1:43926:0 1 0:0:264 294 = 92000074 faff0401 00adfeff fcfffff8 feff6400 00e4feff b3ffffec 0000f0fe ffff8801c9816e00 2022632645 S Zi:4:003:5 -115:1:43926 1 -18:0:294 294 < ffff8801c9816600 2022634635 C Zi:4:003:5 0:1:43928:0 1 0:0:264 294 = a8feff58 fdffd5ff ffecffff 1f0000ad 03004100 00ceffff ffffff8e fcff8300 ffff8801c9816600 2022634646 S Zi:4:003:5 -115:1:43928 1 -18:0:294 294 < Look at the frame number values in the 'C' rows (fifth column, third number). They go: 43924, 43925, 43926, 43928, ... What happened to 43927? Look also at the timestamps (second column). They are in microseconds, so ignore the first 4 and the last 3 digits to get just the milliseconds. They go: 630, 631, 632, 634, ... What happened to 633? The entries weren't simply omitted from the trace. You can tell by looking at the URB addresses (first column). There are eight Zi URBs, and up to this point they repeat over and over the same cycle (ignoring the first 12 hex digits): 6400, 6300, 6200, 6700, 6d00, 6e00, 6900, 6600. But here we see that 6900 is missing, and it never reappears later on. The same is true with the Zo URBs. There are only three of them, and they repeat over and over: 9800, 9600, 9e00. But the 9600 URB should appear at the 634 ms mark, and it doesn't -- it doesn't appear at all from that point on. The same thing eventually happens with the 9800 and then the 9e00 URBs, at which point you're not getting any audio output at all. > But this time in the testing I did not even get to the part where I'd > start the playback. The crash probably happened right after I started > Jack with QJackCtl app. That's good! In fact, for any future tests, _don't_ start the playback. The less things happening, the better. It's seems pretty clear that these problems are the result of a hardware bug in your OHCI controller. There's no simple way to work around it in the driver. But you can work around it by not attaching the sound card to that controller. Buy a USB-2.0 hub and plug the sound card into it (but don't attach the keyboard or mouse to the hub; leave them plugged directly into the computer). 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
