Test: deepsleep() API replacement #5147
Conversation
|
In these tests, we should not allow deepsleep when sleep() is called ? |
| /* Make sure deepsleep is allowed, to go to deepsleep */ | ||
| bool deep_sleep_allowed = sleep_manager_can_deep_sleep(); | ||
| TEST_ASSERT_TRUE_MESSAGE(deep_sleep_allowed, "Deep sleep should be allowed"); | ||
| sleep(); |
There was a problem hiding this comment.
this is assuming that using LowPower object does allow deepsleep, and using sleep() should go into deepsleep, but would it? :-) As this test is for drivers, this looks fine, what about the below test for hal? Same? no direct deepsleep hal call ?
There was a problem hiding this comment.
I was speaking about lp_timeout_1s_sleep function.
There was a problem hiding this comment.
I was speaking about lp_timeout_1s_sleep function.
Should have lock there.
Both changes are changes for deepsleep test cases, so deepsleep should be allowed and should be entered? |
|
Depends on #5148 |
|
As the dependency was merged, this is opened for reviews |
|
Provided updates for handling deep-sleep mode are valid. |
That's true, they should be documented. I can add a commit here, however might be better as a new patch |
|
Hi I assume current update is OK to force part 3. |
0xc0170
force-pushed
the
fix_deepsleep_tests
branch
2 times, most recently
from
2026024 to
e1d1c4d
Compare
| @@ -75,6 +78,8 @@ void lp_timeout_1s_sleep(void) | |||
| sleep_manager_lock_deep_sleep(); | |||
| timestamp_t start = us_ticker_read(); | |||
| lpt.attach(&cb_done, 1); | |||
| bool deep_sleep_allowed = sleep_manager_can_deep_sleep(); | |||
| TEST_ASSERT_FALSE_MESSAGE(deep_sleep_allowed, "Deep sleep should be allowed"); | |||
There was a problem hiding this comment.
Since deepsleep() is not allowed now (as it should) the message is invalid.
0xc0170
force-pushed
the
fix_deepsleep_tests
branch
from
e1d1c4d to
af141ac
Compare
Use sleep() as entry function + check to be certain we are entering deepsleep when required by test (should be allowed)
0xc0170
force-pushed
the
fix_deepsleep_tests
branch
from
af141ac to
b30c622
Compare
|
@fkjagodzinski Can you review again pls? |
There was a problem hiding this comment.
Missing test case description.
Please don't bother with comments for LowPowerTimeout test cases. I'll add them in #5106.
|
/morph test-nightly |
Result: SUCCESSYour command has finished executing! Here's what you wrote!
OutputAll builds and test passed! |
|
retest uvisor |
Build : SUCCESSBuild number : 102 Triggering tests/test mbed-os |
Test : SUCCESSBuild number : 52 |
Use sleep() as entry function + check to be certain we
are entering deepsleep when required by test (should be allowed as we use LowPower object).
This has dependency on fixing this issue #5076 (the test now fails, and will be fixed as today).
Tested using K64F and GCC ARM , these 2 tests affected, one failure at the moment
@jeromecoutant