top shows that tbp sometimes uses 0%CPU, and sometimes uses 1% CPU, so about ~.5% CPU would tie in with your estimate. Still a bit excessive though, I started tpb a few minutes ago and it's already used a second of CPU.
I also recompiled thinkpad-keys and changed the poll interval from 50 to 100 milliseconds. This reduces load to ~1%.
Attaching gdb to thinkpad-keys showed a backtrace waiting in nanosleep() and usleep():
(gdb) bt
#0 0xffffe410 in __kernel_vsyscall ()
#1 0xb7f1b110 in nanosleep () from /lib/tls/i686/cmov/libc.so.6
#2 0xb7f5198a in usleep () from /lib/tls/i686/cmov/libc.so.6
#3 0x08048c95 in main (argc=1, argv=0xbf8e8344) at thinkpad-keys.c:282
(gdb)
Would be interesting to know why this breaks on the X60. Mine is also the Core Duo.
Hi Paul,
OK, I ran tbp using this command (obtained from the tpb docs):
sudo tpb --osd=on --verbose --thinkpad= "/usr/bin/ X11/xterm -T ntpctl -e ntpctl"
top shows that tbp sometimes uses 0%CPU, and sometimes uses 1% CPU, so about ~.5% CPU would tie in with your estimate. Still a bit excessive though, I started tpb a few minutes ago and it's already used a second of CPU.
I also recompiled thinkpad-keys and changed the poll interval from 50 to 100 milliseconds. This reduces load to ~1%.
Attaching gdb to thinkpad-keys showed a backtrace waiting in nanosleep() and usleep():
(gdb) bt i686/cmov/ libc.so. 6 i686/cmov/ libc.so. 6
#0 0xffffe410 in __kernel_vsyscall ()
#1 0xb7f1b110 in nanosleep () from /lib/tls/
#2 0xb7f5198a in usleep () from /lib/tls/
#3 0x08048c95 in main (argc=1, argv=0xbf8e8344) at thinkpad-keys.c:282
(gdb)
Would be interesting to know why this breaks on the X60. Mine is also the Core Duo.