dg_pthread_group_get_times(3T)



dg_pthread_group_get_times(3T)    SDK R4.11   dg_pthread_group_get_times(3T)


NAME
       dgpthreadgroupgettimes - get CPU times for a thread group

SYNOPSIS
       #include <pthread.h>
       #include <timers.h>

       int dgpthreadgroupgettimes(dgpthreadgroupt threadgroup,
                                      struct timespec *usertimeptr,
                                      struct timespec *systemtimeptr);

   where:
       threadgroup
              Target thread group ID
       usertimerptr
              Pointer to the location to receive the accumulated user CPU
              time
       systemtimerptr
              Pointer to the location to receive the accumulated system CPU
              time

DESCRIPTION
       This function returns the amount of user and system CPU time
       accumulated by the thread group with thread group ID given by
       threadgroup.   The CPU time is the summation of all CPU time
       accumulated by all of the thread group's past and present threads.

       This call succeeds even if the thread group's resources have not been
       freed.   In other words, this call succeeds even if the thread group
       has been marked as destroyed, as long as the thread group's resources
       have not yet been cleaned up due to the existence of active or
       undetached threads in the thread group.

DIAGNOSTICS
   Return Value
       If successful, dgpthreadgroupgettimes() returns 0, and saves the
       accumulated CPU times in the locations pointed to by usertimeptr
       and systemtimeptr.  Otherwise, it returns -1 and sets errno.

   Errors
       For each of the following conditions, dgpthreadgroupgettimes()
       returns -1 and sets errno to the corresponding value:

       [ESRCH]   The value specified by threadgroup does not refer to a
                 existing thread group.

SEE ALSO
       dgpthreadgroupcreate(3T), dgpthreadgroupdestroy(3T),
       dgpthreadgroupself(3T), dglwpinfo(2).

NOTES
       For efficieny reasons, CPU time is not accumulated on a per-thread
       basis, only on a thread-group basis.  This allows for the fastest
       possible thread-to-thread context switch within a thread group.

       For a thread group dedicated to a globally scheduled thread, this
       call returns the accumulated CPU time for the globally scheduled
       thread because no other thread can be created in the thread group;
       the same result is returned by a call to dg_lwp_info() for the
       globally scheduled thread.


Licensed material--property of copyright holder(s)
Museum
