pthread_attr can go in psxtests, cpu_set_t with 1 cpu can go in sptests. SMP-aware versions should go in smptests, which should continue to be conditional on --enable-smp switch.
There may need to be some psxsmptests in the future. They could perhaps be integrated in the psxtests directory. On Fri, Nov 8, 2013 at 10:46 AM, Joel Sherrill <joel.sherr...@oarcorp.com> wrote: > Following up on Jennifer. The "real functional" tests > require SMP to be enabled. But cpu_set_t manipulation > and the pthread_attr methods are always available. > Newlib doesn't know SMP vs. non-SMP. > > So the basic tests of cpu_set_t and pthread_attr_XXX_np > are what we are talking about. > > Tests which check that the affinity requested is honored > is a different matter. [And may require knowing which > CPU you are running on.. hmm.. have to think more about > testing affinity.] > > --joel > > On 11/8/2013 9:38 AM, Jennifer Averett wrote: >> Since the header information is in newlib it will always be there. >> >> Where should the affinity tests be added? >> >> >> >> A) In the smp tests and change the smptests so those are always >> compiled >> >> B) In psxtests >> >> >> >> Any other suggestions? >> >> >> >> Jennifer Averett >> >> On-Line Applications Research >> >> 256-319-2752 >> >> >> > > > > -- > Joel Sherrill, Ph.D. Director of Research & Development > joel.sherr...@oarcorp.com On-Line Applications Research > Ask me about RTEMS: a free RTOS Huntsville AL 35805 > Support Available (256) 722-9985 > _______________________________________________ > rtems-devel mailing list > rtems-devel@rtems.org > http://www.rtems.org/mailman/listinfo/rtems-devel _______________________________________________ rtems-devel mailing list rtems-devel@rtems.org http://www.rtems.org/mailman/listinfo/rtems-devel