Gilles Chanteperdrix wrote: > Now that the big context switches bugs have been solved, here is a patch > that adds a unit test for context switches and FPU switches > with various type of threads (kernel, user, user in secondary mode, > not using FPU, using FPU, etc...).
Very good idea! Maybe it's a start to create more of such useful tests. > As is the case of the latency test > there is a small RTDM driver in kernel-space, put in the benchmark > class, even though this test is for unit testing, not for benchmarking. "Benchmark" is likely not the right term for such devices anymore. I guess no one will miss rtbenchmark.h, so what about this: rttesting.h: RTDM_CLASS_TESTING | +-- RTDM_SUBCLASS_TIMER | +-- RTDM_SUBCLASS_SWITCH In fact, you may even consider to add some performance tests to your driver one day. I would find it very interesting to have some numbers on average domain migration delays or on those various other switching variants - both from user and kernel contexts. This could then serve users to decide how to design critical parts of their system (and us to look for performance regressions). > > The FPU switches need a small piece of code architecture dependent, > put in <asm/xenomai/fptest.h>, currently only implemented for x86. > > The kernel-space driver is called xeno_switchtest.ko, the user-space > testing tool is called switchtest, because there is already a context > switch benchmarking tool called "switch". > By combining both feature sets, the need for different names would disappear. :) Jan
Description: OpenPGP digital signature
_______________________________________________ Xenomai-core mailing list Xenomaiemail@example.com https://mail.gna.org/listinfo/xenomai-core