Since atomic operations have been adopted in DPDK now[1], change rte_atomicNN_xxx APIs to compiler's atomic built-ins in test cases.
[1] https://www.dpdk.org/blog/2021/03/26/dpdk-adopts-the-c11-memory-model/ v3: Change 'GCC atomic builtins' to 'compiler atomic builtins' as clang adopted the same syntax as GCC.<Thomas Monjalon> v2: Use rte_wait_until_equal() instead of original sync loops. <David Marchand> v1: The initial version. Joyce Kong (8): test/ticketlock: use compiler atomics for lcores sync test/spinlock: use compile atomics for lcores sync test/rwlock: use compiler atomics for lcores sync test/mcslock: use compiler atomics for lcores sync test/mempool: remove unused variable for lcores sync test/mempool_perf: use compiler atomics for lcores sync test/service_cores: use compiler atomics for lock sync test/rcu: use compiler atomics for data sync app/test/test_mcslock.c | 14 +++-- app/test/test_mempool.c | 5 -- app/test/test_mempool_perf.c | 11 ++-- app/test/test_rcu_qsbr_perf.c | 98 +++++++++++++++++------------------ app/test/test_rwlock.c | 10 ++-- app/test/test_service_cores.c | 36 +++++++------ app/test/test_spinlock.c | 9 ++-- app/test/test_ticketlock.c | 10 ++-- 8 files changed, 91 insertions(+), 102 deletions(-) -- 2.17.1
