[Pkg-utopia-maintainers] Bug#1093142: more information

[Russell Coker]

I currently can't reproduce the situation of taking more than 1 CPU core.  I 
didn't do anything differently, just rebooted it a day later and got different 
results.

Now I'm seeing the following, not a huge amount of CPU time, but more than 
expected from an idle system with the screen locked and nothing happening.

    PID USER      PR  NI    VIRT    RES    SHR S  %CPU  %MEM     TIME+ COMMAND                                                                                                                                                                 
   1913 mobian    20   0 5447112 410056 196544 S   2.0   5.3   0:24.27 
plasmashell                                                                                                                                                             
   1537 mobian    20   0  106568  10732   6916 S   1.3   0.1   0:07.15 
pipewire                                                                                                                                                                
   1544 mobian    20   0  186116  17092   7492 S   1.3   0.2   0:07.00 
pipewire-pulse  

root at op6a:~# strace -p 1537
strace: Process 1537 attached
epoll_pwait(5, 

The pipewire process is waiting, how can it be using CPU time?

root at op6a:~# strace -p 1544 2>&1|head 
strace: Process 1544 attached
epoll_pwait(4, [{events=EPOLLIN, data={u32=201999296, u64=187651618128832}}], 
32, -1, NULL, 8) = 1
read(5, "\1\0\0\0\0\0\0\0", 8)          = 8
epoll_ctl(4, EPOLL_CTL_MOD, 44, {events=EPOLLIN|EPOLLOUT|EPOLLERR|EPOLLHUP, 
data={u32=205036208, u64=187651621165744}}) = 0
epoll_pwait(4, [{events=EPOLLOUT, data={u32=205036208, u64=187651621165744}}], 
32, -1, NULL, 8) = 1
sendto(44, "\0\0\0\4\0\0\0\1\0\0\0\0\0\0\0\0\0\0\0\0", 20, MSG_DONTWAIT|
MSG_NOSIGNAL, NULL, 0) = 20
sendto(44, "\0\0\0\0", 4, MSG_DONTWAIT|MSG_NOSIGNAL, NULL, 0) = 4
epoll_ctl(4, EPOLL_CTL_MOD, 44, {events=EPOLLIN|EPOLLERR|EPOLLHUP, 
data={u32=205036208, u64=187651621165744}}) = 0
epoll_pwait(4, [{events=EPOLLIN, data={u32=201999296, u64=187651618128832}}], 
32, -1, NULL, 8) = 1
read(5, "\1\0\0\0\0\0\0\0", 8)          = 8
root at op6a:~# strace -p 1544 2>&1|head -20
strace: Process 1544 attached
epoll_pwait(4, [{events=EPOLLIN, data={u32=201999296, u64=187651618128832}}], 
32, -1, NULL, 8) = 1
read(5, "\1\0\0\0\0\0\0\0", 8)          = 8
epoll_ctl(4, EPOLL_CTL_MOD, 44, {events=EPOLLIN|EPOLLOUT|EPOLLERR|EPOLLHUP, 
data={u32=205036208, u64=187651621165744}}) = 0
epoll_pwait(4, [{events=EPOLLOUT, data={u32=205036208, u64=187651621165744}}], 
32, -1, NULL, 8) = 1
sendto(44, "\0\0\0\4\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0", 20, MSG_DONTWAIT|
MSG_NOSIGNAL, NULL, 0) = 20
sendto(44, "\0\0\0\0", 4, MSG_DONTWAIT|MSG_NOSIGNAL, NULL, 0) = 4
epoll_ctl(4, EPOLL_CTL_MOD, 44, {events=EPOLLIN|EPOLLERR|EPOLLHUP, 
data={u32=205036208, u64=187651621165744}}) = 0
epoll_pwait(4, [{events=EPOLLIN, data={u32=201999296, u64=187651618128832}}], 
32, -1, NULL, 8) = 1
read(5, "\1\0\0\0\0\0\0\0", 8)          = 8
epoll_ctl(4, EPOLL_CTL_MOD, 44, {events=EPOLLIN|EPOLLOUT|EPOLLERR|EPOLLHUP, 
data={u32=205036208, u64=187651621165744}}) = 0
epoll_pwait(4, [{events=EPOLLOUT, data={u32=205036208, u64=187651621165744}}], 
32, -1, NULL, 8) = 1
sendto(44, "\0\0\0\4\0\0\0\1\0\0\0\0\0\0\0\0\0\0\0\0", 20, MSG_DONTWAIT|
MSG_NOSIGNAL, NULL, 0) = 20
sendto(44, "\0\0\0\0", 4, MSG_DONTWAIT|MSG_NOSIGNAL, NULL, 0) = 4
epoll_ctl(4, EPOLL_CTL_MOD, 44, {events=EPOLLIN|EPOLLERR|EPOLLHUP, 
data={u32=205036208, u64=187651621165744}}) = 0
epoll_pwait(4, [{events=EPOLLIN, data={u32=201999296, u64=187651618128832}}], 
32, -1, NULL, 8) = 1
read(5, "\1\0\0\0\0\0\0\0", 8)          = 8
epoll_ctl(4, EPOLL_CTL_MOD, 44, {events=EPOLLIN|EPOLLOUT|EPOLLERR|EPOLLHUP, 
data={u32=205036208, u64=187651621165744}}) = 0
epoll_pwait(4, [{events=EPOLLOUT, data={u32=205036208, u64=187651621165744}}], 
32, -1, NULL, 8) = 1
sendto(44, "\0\0\0\4\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0", 20, MSG_DONTWAIT|
MSG_NOSIGNAL, NULL, 0) = 20
root at op6a:~# ls -l /proc/1544/fd/5
lrwx------ 1 mobian mobian 64 Jan 17 00:08 /proc/1544/fd/5 -> 'anon_inode:
[eventfd]'
root at op6a:~# ls -l /proc/1544/fd/44
lrwx------ 1 mobian mobian 64 Jan 17 00:02 /proc/1544/fd/44 -> 'socket:
[18425]'

The pipewire-pulse program is looking on an eventfd and writing to a socket in 
a fairly tight loop.

As for plasmashell, what it does is more complex and I'm not sure if it's 
connected to the pipewire CPU use.

For desktop use probably most people would be happy to let these processes use 
1% CPU time each for as long as the system is used.  For even a laptop this 
probably isn't a problem.  But for a phone it's a real problem.

-- 
My Main Blog         http://etbe.coker.com.au/
My Documents Blog    http://doc.coker.com.au/