ok, I found the bug in the timer implementation, a fix is coming soon. On Mon, Feb 16, 2009 at 10:27 AM, David Turner <di...@android.com> wrote:
> this one's for me :-) > > Can you send me a small compilable test program to reproduce this (it'd be > much better for me since > I don't know what your timer_handler function is doing there). > > It might be a bug in the SIGEV_THREAD timer implementation, I checked and > tested it for correctness but didn't look at CPU usage, I admit. > > By the way, sigev_signo isn't used with SIGEV_THREAD timers. > > > On Mon, Feb 16, 2009 at 4:13 AM, henry.lon...@gmail.com < > henry.lon...@gmail.com> wrote: > >> >> hi, >> >> i played with timer_create routine with the following code. it seems >> the timer is working, but the cpu utilization is really high(over 90%) >> if i run "top". Any idea what's going on on G1? >> >> Thanks! >> >> >> ----------------------------------------------------------------------------------------------------- >> se.sigev_notify = SIGEV_THREAD; >> se.sigev_notify_function = timer_handler; >> se.sigev_signo = SIGUSR1; >> timer_create(CLOCK_REALTIME, &se, &tid); >> >> memset(&ts, 0, sizeof(struct itimerspec)); >> ts.it_value.tv_sec = 0; >> ts.it_value.tv_nsec = 30000000; >> ts.it_interval.tv_sec = 0; >> ts.it_interval.tv_nsec = 30000000; >> result = timer_settime(tid, 0, &ts, 0); >> >> >> >> > --~--~---------~--~----~------------~-------~--~----~ unsubscribe: android-porting+unsubscr...@googlegroups.com website: http://groups.google.com/group/android-porting -~----------~----~----~----~------~----~------~--~---
