https://bugzilla.novell.com/show_bug.cgi?id=464998 User m.munnix@redcross.be added comment https://bugzilla.novell.com/show_bug.cgi?id=464998#c1 --- Comment #1 from michel munnix 2009-01-09 13:11:12 MST --- Created an attachment (id=264232) --> (https://bugzilla.novell.com/attachment.cgi?id=264232) example y2log with debug enabled here an example debug log when moving a slider from 78 to 93, it takes about 5 seconds : seems as if it's writing config for every single change in value -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=464998 User lslezak@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=464998#c2 Ladislav Slezak changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |INVALID --- Comment #2 from Ladislav Slezak 2009-01-12 08:18:24 MST --- I have reproduced it - the problem is caused by the enabled debug mode. Every single volume change results in about 4kB text written to the yast log. If you move a slider very fast the result might be even about 1MB written to the log. And that's why it's so slow. If you disable the debug mode then the sliders should move immediately. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=464998 User m.munnix@redcross.be added comment https://bugzilla.novell.com/show_bug.cgi?id=464998#c4 --- Comment #4 from michel munnix 2009-01-12 13:56:22 MST --- Created an attachment (id=264620) --> (https://bugzilla.novell.com/attachment.cgi?id=264620) /var/log/boot.msg wondering if this could be a kernel / timing issue know any benchmark I could run to sort this out ? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=464998 User m.munnix@redcross.be added comment https://bugzilla.novell.com/show_bug.cgi?id=464998#c6 michel munnix changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |REOPENED Info Provider|m.munnix@redcross.be | --- Comment #6 from michel munnix 2009-01-13 14:15:58 MST --- in kmix, response time is satisfactory because it adapts pace while moving. but when I drag slowly, I can see it hop about 3 times per second. If I drag fast, it will make bigger hops. I rebooted also on my old installation on that machine (opensuse 10.2) : here 'yast2 sound' behaves even faster than kmix, the cursor is *very* responsive. Installed latest bios (P1.90 -> P2.80) Installed X driver from Nvidia repository, I'm impressed with the desktop effects :) but this didn't change anything for our problem. I then tried an "strace -Tttv -s 128 -p 6213 -o sound.trace" command on the y2base process, this seems to be very informative : One iteration takes about 350 ms, (81 syscalls) cumulated syscall time accounts for most of it. Especially ioctl's on the /dev/snd/controlC0 account for about 320 ms /dev/snd/controlC0 is opened RO / closed /opened RW /closed at each iteration, why not leave it open ? are really that much ioctl (70) necessary to change the value of one control ? so the added latency of the usb communication amplifies the effect on my system -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=464998 User lslezak@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=464998#c8 Ladislav Slezak changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |NEW Component|YaST2 |Sound AssignedTo|lslezak@novell.com |tiwai@novell.com QAContact|jsrain@novell.com |qa@suse.de --- Comment #8 from Ladislav Slezak 2009-01-14 07:18:59 MST --- Thank you for the details. The mixer part in yast2-sound has not been changed since 10.2, the problem is in kernel (USB) or ALSA (snd-usb-audio driver). -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=464998 User m.munnix@redcross.be added comment https://bugzilla.novell.com/show_bug.cgi?id=464998#c10 michel munnix changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW Info Provider|m.munnix@redcross.be | --- Comment #10 from michel munnix 2009-01-14 15:13:22 MST --- Sorry, amixer and alsamixer insist on using pulseaudio, even if I tell yast not to enable it, it starts. When I do not start kde, it doesn't work either. Any hint or an alternative mixer to try? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=464998 User m.munnix@redcross.be added comment https://bugzilla.novell.com/show_bug.cgi?id=464998#c12 michel munnix changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW --- Comment #12 from michel munnix 2009-01-15 03:15:29 MST --- Created an attachment (id=265240) --> (https://bugzilla.novell.com/attachment.cgi?id=265240) strace output for amixer -Dhw:0 --stdin I pasted repetitively (40x) the command "set PCM 1%-" to amixer there are 8 ioctl per iteration accounting for 52ms to 150ms, mean 66ms so yast2: 70 ioctl -> 350 ms amixer: 8 ioctl -> 66 ms I'd say same problem for amixer even if it uses ioctl's more effectively -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=464998 User m.munnix@redcross.be added comment https://bugzilla.novell.com/show_bug.cgi?id=464998#c14 michel munnix changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW Info Provider|m.munnix@redcross.be | --- Comment #14 from michel munnix 2009-01-16 11:14:20 MST --- tried alsa-driver-unstable-kmp-default-1.0.18.20090116_2.6.27.7_9.1-2.1 with kernel-default-2.6.27.7-9.1 It does indeed improve, it shows normal behavior in "yast sound volume" module I guess there are other changes also in these modules, because in playback I have these messages in /var/log/messages and playback is suspended after about 25 seconds: Jan 16 18:47:25 linux-bcpx pulseaudio[4859]: cpulimit.c: Recevied request to terminate due to CPU overload. Jan 16 18:48:37 linux-bcpx pulseaudio[5464]: main.c: High-priority scheduling enabled in configuration but not allowed by policy. Jan 16 18:48:37 linux-bcpx pulseaudio[5464]: core-util.c: setpriority(): Permission denied Jan 16 18:48:37 linux-bcpx pulseaudio[5464]: ltdl-bind-now.c: Failed to find original dlopen loader. Jan 16 18:48:37 linux-bcpx pulseaudio[5466]: module-alsa-source.c: Your kernel driver is broken: it reports a volume range from -16.00 dB to -16.00 dB which makes no sense. Jan 16 18:48:55 linux-bcpx pulseaudio[5466]: module-alsa-sink.c: Increasing wakeup watermark to 40.00 ms Jan 16 18:49:21 linux-bcpx pulseaudio[5466]: cpulimit.c: Recevied request to terminate due to CPU overload. to 2 "top" most processes 5936 mm 20 0 321m 4992 3752 S 88 0.2 0:07.86 pulseaudio 6 root 15 -5 0 0 0 S 12 0.0 0:17.40 ksoftirqd/1 -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=464998 User tiwai@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=464998#c16 --- Comment #16 from Takashi Iwai 2009-01-17 03:58:48 MST --- Thanks, good to hear that something gets improved. About the error messages -- the "-16.00 dB to -16.0dB..." should have appears in the original driver, too, as there is no change regarding this. Basically it's no real driver bug. It's a hardware that gives the broken information. I'll back-merge the patch to the main tree so that you can try alsa-driver-kmp (no stable version) later. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=464998 User m.munnix@redcross.be added comment https://bugzilla.novell.com/show_bug.cgi?id=464998#c18 michel munnix changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW Info Provider|m.munnix@redcross.be | --- Comment #18 from michel munnix 2009-01-17 08:06:25 MST --- I tested with the new alsa-driver-kmp-default-1.0.18.20090117_2.6.27.7_9.1-2.1, this works also fine and no more problems with pulseaudio. About the "-16.00 dB to -16.0dB..." error messages, they didn't appear with the original driver and do no longer appear now that I installed this driver. I am happy with this fix, thanks a lot -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.