Hi Christos, Some updates on the tests of priority protect. You need to apply my new patch here, https://github.com/ycui1984/posixtestsuite/blob/master/patches/PRIOPROTECT_AND_GETCLOCK/0005-extend-sched_param.patch
and run the test, https://github.com/ycui1984/posixtestsuite/blob/master/patches/PRIOPROTECT_AND_GETCLOCK/0006-unit-tests-for-priority-protect.patch You can see thread who is blocked on mutex has higher protect priority, which is considered by kernel when scheduling. Thanks, Charles 2016-06-20 18:58 GMT-07:00 Christos Zoulas <chris...@zoulas.com>: > On Jun 20, 1:07pm, charles.cui1...@gmail.com (Charles Cui) wrote: > -- Subject: Re: _SC_SIGQUEUE_MAX > > | I have write a draft of the benchmark, but have not verified successfully > | on the patched system. > | The idea is to > | 1. create a mutexattr data structure and set pthread priority protect > | protocol. > | 2. set the real time scheduling policy. > | 3. use the mutexattr to init a mutex. > | 4. create 2 threads to contend for the mutex > | 5. verify that the thread who is waiting for the mutex is executed on a > | higher priority(ceiling priority of mutex) > | I will notify you the new results once I have made progress. > | > | I will send you the summary and plan in a separate email later. > > Sounds like a good plan, thanks! > > christos >
