| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch makes each test print start marker before sync. This will
allow us to distinguish the failure before even sync is attempted
(CTS_RC_DID_NOT_START, thus probably caused by the previous test)
and the failure caused by the hanging partner, in which case the one
good and alive will be stuck in sync (and should return _DID_NOT_END
or even better _BAD_SYNC once we implement timeout in sync).
This patch also:
* Adds did_not_start_test to and removes debug_test from meta suite
* Consolidates test runner loops into common cts_main_loop
* Removes dut_common.h and th_common.h
* Removes debug print macro and CTS_DEBUG
* Replaces all infinite loops after tests with task_wait_event(-1)
* Removes meaningless comments and debug printfs
* Removes CTS_TEST_ID_*
* Adds sync() to task suite
BUG=chromium:736104
BRANCH=none
TEST=Run run_ects.sh and verify all tests pass
Change-Id: I6ccdf26afac6b8e8cb16483c5d75e4e77e7962f4
Signed-off-by: Daisuke Nojiri <dnojiri@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/545176
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch makes the framework verify not only the result but also
the execution order of the tests. It also allows each test to
specify expected return code and strings printed by TH and DUT.
The final test results depends on the return code and the expectation.
Therefore, the output now includes 'RESULT' column showing PASS or FAIL:
test name TH_RETURN_CODE DUT_RETURN_CODE TH_STR DUT_STR RESULT
test_task_switch SUCCESS SUCCESS 1 1 PASS
test_task_priority SUCCESS FAILURE 1 1 FAIL
test_stack_overflow DID_NOT_END DID_NOT_END 1 1 PASS
Additionally, this patch:
* Adds CTS_RC_DID_NOT_START and CTS_RC_DID_NOT_END to indicate whether
the test did start or end, respectively.
* Makes stack overflow test check whether stack overflow was detected
and reboot occurred
* Removes post_corruption_test and conflict test since now
the test results are stricly compared against expected results.
* Fixes gpylint errors.
BUG=none
BRANCH=none
TEST=Run gpio, meta, timer, interrupt, and cts/cts.py -m task
Change-Id: I3b7005236e705dcac0c8f4711b44c85ff9a4f676
Reviewed-on: https://chromium-review.googlesource.com/538878
Commit-Ready: Daisuke Nojiri <dnojiri@chromium.org>
Tested-by: Daisuke Nojiri <dnojiri@chromium.org>
Reviewed-by: Daisuke Nojiri <dnojiri@chromium.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch adds a test for stack overflow detection. CTS task tries
to overflow by calling a function recursively. This function sleeps
to incur context switch, where stack overflow is checked. Reboot is
expected but not checked.
BUG=chromium:663873
BRANCH=none
TEST=cts.py -m task
Change-Id: I6ad98d209acf873d2d23b48bac6fc751fa8c2f74
Signed-off-by: Daisuke Nojiri <dnojiri@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/410283
Reviewed-by: Randall Spangler <rspangler@chromium.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
CTS task wakes up A and C then goes to sleep:
CTS -> A, C -> A -> B -> C
Since C has a higher priority, C should run first. This should result
in C running one more time than A (or B).
BUG=chromium:663873
BRANCH=none
TEST=cts.py -m task
Change-Id: I89c733ba3aab09b293edf8583d6ed73791531e59
Signed-off-by: Daisuke Nojiri <dnojiri@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/409535
Reviewed-by: Randall Spangler <rspangler@chromium.org>
|
|
Task A wakes up B and goes to sleep. Task B wakes up C then goes to
sleep. Task C wakes up A then goes to sleep. This is repeated
repeat_count times:
A -> B -> C -> A -> ...
It's expected all tasks to run exactly repeat_count times. Tick task
runs to inject some irregularity.
BUG=chromium:663873
BRANCH=none
TEST=cts.py -m task
Change-Id: Ib7227f05f09b7a49f8528aff6e6e8d3e6df93ba7
Signed-off-by: Daisuke Nojiri <dnojiri@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/409534
Reviewed-by: Randall Spangler <rspangler@chromium.org>
|