For testing ensure syscall( __NR_clock_getres, CLOCK_MONOTONIC, &dummy ) would return zero (successfully). So that we can get a monotonic clock that does not depends on user date/time otherwise failback will be used. A Linux 2.6 kernel should have this support.
Denis Oliver Kropp wrote: > Keith Mok wrote: >> Hi all, >> >> Correction of 3 of 8) Add direct_util_monotonic_pthread_cond_init and >> direct_util_get_monotonic_pthread_timeout interface >> So pthread_cond_timedwait will not >> affected by system date/time changes. >> Previous one is wrong, Sequence of pthread_condattr_init in >> direct_util_monotonic_pthread_cond_init is wrong for obvious reason. >> Sorry for the noise. > > Thanks for all your patches regarding this important topic. > > Do I need special kernel support to test it? > _______________________________________________ directfb-dev mailing list directfb-dev@directfb.org http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-dev
