Ticket #722 (new defect)
Opened 8 months ago
High CPU load during capture from the ALSA "default" device
| Reported by: | krnekit | Owned by: | ossman |
|---|---|---|---|
| Milestone: | Component: | alsa-plugins-pulse | |
| Keywords: | alsa capture default cpu load | Cc: |
Description
While doing capture from the ALSA "default" device the program have high cpu load (~20% of one core on my AMD Athlon X2). This can be reproduced by the following command:
arecord > /dev/null
Using the device, distinct from "default" (for example "plughw:0,0"), using capture through pulseaudio without emulation (with parecord) or doing playback works well without high cpu load.
This problem discovered in version 0.9.19 on Mandriva Linux 2010.0. Also it have been reproduced on Ubintu 9.10 and Fedora 12 with the same version of pulseaudio, and on Mandriva 2009.1. This problem is absent in pulseadio 0.9.12 (openSUSE 11.1). But in 0.9.10 on Mandriva 2009.0 it again gives high cpu load.
Calling strace on arecord gives information about enormous amount of poll() calls:
% time seconds usecs/call calls errors syscall ------ ----------- ----------- --------- --------- ---------------- 95.30 0.014735 0 62709 poll 2.43 0.000375 375 1 execve 2.27 0.000351 2 156 2 futex 0.00 0.000000 0 233 92 read 0.00 0.000000 0 167 write 0.00 0.000000 0 75 14 open 0.00 0.000000 0 62 close 0.00 0.000000 0 1 stat 0.00 0.000000 0 57 fstat 0.00 0.000000 0 96 mmap 0.00 0.000000 0 54 mprotect 0.00 0.000000 0 18 munmap 0.00 0.000000 0 4 brk 0.00 0.000000 0 5 rt_sigaction 0.00 0.000000 0 1 rt_sigprocmask 0.00 0.000000 0 17 writev 0.00 0.000000 0 6 3 access 0.00 0.000000 0 2 pipe 0.00 0.000000 0 2 socket 0.00 0.000000 0 2 connect 0.00 0.000000 0 1 1 sendto 0.00 0.000000 0 1 getsockname 0.00 0.000000 0 1 getpeername 0.00 0.000000 0 1 setsockopt 0.00 0.000000 0 1 clone 0.00 0.000000 0 8 1 kill 0.00 0.000000 0 1 uname 0.00 0.000000 0 23 fcntl 0.00 0.000000 0 1 ftruncate 0.00 0.000000 0 2 getdents 0.00 0.000000 0 1 unlink 0.00 0.000000 0 2 readlink 0.00 0.000000 0 1 getrlimit 0.00 0.000000 0 1 getuid 0.00 0.000000 0 1 statfs 0.00 0.000000 0 1 arch_prctl 0.00 0.000000 0 1 set_tid_address 0.00 0.000000 0 1 set_robust_list ------ ----------- ----------- --------- --------- ---------------- 100.00 0.015461 63717 113 total
This is for approximately ten seconds.
More detailed strace output shows, that this call:
poll([{fd=5, events=POLLIN|POLLERR|POLLNVAL}], 1, -1) = 1 ([{fd=5, revents=POLLIN}])
repeated ~1000 times.
Doing playback, poll() not called so frequently:
% time seconds usecs/call calls errors syscall ------ ----------- ----------- --------- --------- ---------------- 87.26 0.004999 54 92 poll 9.25 0.000530 17 32 2 futex 3.23 0.000185 2 77 write 0.26 0.000015 0 96 mmap 0.00 0.000000 0 267 75 read 0.00 0.000000 0 75 14 open 0.00 0.000000 0 62 close 0.00 0.000000 0 1 stat 0.00 0.000000 0 57 fstat 0.00 0.000000 0 54 mprotect 0.00 0.000000 0 17 munmap 0.00 0.000000 0 4 brk 0.00 0.000000 0 5 rt_sigaction 0.00 0.000000 0 1 rt_sigprocmask 0.00 0.000000 0 17 writev 0.00 0.000000 0 6 3 access 0.00 0.000000 0 2 pipe 0.00 0.000000 0 2 socket 0.00 0.000000 0 2 connect 0.00 0.000000 0 1 1 sendto 0.00 0.000000 0 1 getsockname 0.00 0.000000 0 1 getpeername 0.00 0.000000 0 1 setsockopt 0.00 0.000000 0 1 clone 0.00 0.000000 0 1 execve 0.00 0.000000 0 8 kill 0.00 0.000000 0 1 uname 0.00 0.000000 0 23 fcntl 0.00 0.000000 0 1 ftruncate 0.00 0.000000 0 2 getdents 0.00 0.000000 0 2 readlink 0.00 0.000000 0 1 getrlimit 0.00 0.000000 0 1 getuid 0.00 0.000000 0 1 statfs 0.00 0.000000 0 1 arch_prctl 0.00 0.000000 0 1 set_tid_address 0.00 0.000000 0 1 set_robust_list ------ ----------- ----------- --------- --------- ---------------- 100.00 0.005729 918 95 total
for the approximately same time. Playback is executed as:
aplay < /dev/zero
