summaryrefslogtreecommitdiffstats
path: root/spec/rtems/clock/req/get-uptime.yml
diff options
context:
space:
mode:
authorFrank Kühndel <frank.kuehndel@embedded-brains.de>2021-07-23 10:47:35 +0200
committerSebastian Huber <sebastian.huber@embedded-brains.de>2021-07-23 16:16:09 +0200
commitfc3586fe37824e51d47840bb97df827916ddd4d3 (patch)
tree74b8766b52f4ed20dfa15353813854cb291ff218 /spec/rtems/clock/req/get-uptime.yml
parentspec: Remove always enabled interrupt attribute (diff)
downloadrtems-central-fc3586fe37824e51d47840bb97df827916ddd4d3.tar.bz2
spec: Fix unstable test of clock manager
The tests of specification get-tod.yml and set.yml were expecting that no more than five ticks pass between consecutive calls to rtems_clock_set() and rtems_clock_get_tod(). Yet, in coverage tests this limit was exceeded. * All tests are placed in a test suite without clock driver so that all ticks must be triggered by the test suite. * The grace period of five ticks has been removed. * Test which require that the realtime clock has not been set manipulate RTEMS internal data with `_TOD.is_set = false`. The old function get_tod_before_set_tod() has been removed in favor of this new solution.
Diffstat (limited to 'spec/rtems/clock/req/get-uptime.yml')
-rw-r--r--spec/rtems/clock/req/get-uptime.yml4
1 files changed, 2 insertions, 2 deletions
diff --git a/spec/rtems/clock/req/get-uptime.yml b/spec/rtems/clock/req/get-uptime.yml
index 05ed4a7f..b63f7342 100644
--- a/spec/rtems/clock/req/get-uptime.yml
+++ b/spec/rtems/clock/req/get-uptime.yml
@@ -28,8 +28,8 @@ post-conditions:
- name: Set
test-code: |
T_eq_ptr( ctx->uptime, &ctx->uptime_value );
- T_gt_i64( ctx->uptime_value.tv_sec, 0LL );
- T_gt_long( ctx->uptime_value.tv_nsec, 0L );
+ T_ge_i64( ctx->uptime_value.tv_sec, 0LL );
+ T_ge_long( ctx->uptime_value.tv_nsec, 0L );
T_lt_long( ctx->uptime_value.tv_nsec, 1000000000L );
text: |
The value of the object referenced by the