make start_test() optional: part 1 #1170
Conversation
This comment was marked as resolved.
This comment was marked as resolved.
This comment was marked as outdated.
This comment was marked as outdated.
Extract new playback_subtest() function to de-duplicate. Pure clean-up, zero functional change. Signed-off-by: Marc Herbert <marc.herbert@intel.com>
Extract new print_test_result_exit() function from exit handler(). This make it possible for some "light-weight" tests like verify-kernel-boot-log.sh to not depend on the heavy and start_test() and exit handler machinery. Signed-off-by: Marc Herbert <marc.herbert@intel.com>
This unblocks thesofproject#1170 and allows "installing" the func_exit_handler() trap in start_test() which will allow making start_test() optional. Signed-off-by: Marc Herbert <marc.herbert@intel.com>
func_exit_handler() makes sense only when used with start_test() start_test() is a "mandatory" right now because it's invoked from lib.sh. But this is wrong because it should be possible to import lib.sh functions without actually running anything and it's causing various issues like sof_remove.sh depending on the firmware being loaded, see discussion in thesofproject#1169 This commit will make it possible NOT to use start_test(). Signed-off-by: Marc Herbert <marc.herbert@intel.com>
marc-hb
force-pushed
the
opt-start-test-1
branch
from
28e1868
to
dd906bf
Compare
|
GLK BOB hung in https://sof-ci.01.org/softestpr/PR1170/build311/devicetest/index.html?model=GLK_BOB_DA7219-ipc3&testcase=check-kmod-load-unload but that happened before this PR, already filed at:
CAVS https://sof-ci.01.org/softestpr/PR1170/build312/devicetest/index.html is all green. ACE https://sof-ci.01.org/softestpr/PR1170/build313/devicetest/index.html is all green. This is good to go! EDIT: and now tests have started running once again, no idea why... even better: GLK BOB (and everything else) passed this time: https://sof-ci.01.org/softestpr/PR1170/build314/devicetest/index.html ACE https://sof-ci.01.org/softestpr/PR1170/build315/devicetest/index.html and CAVS https://sof-ci.01.org/softestpr/PR1170/build316/devicetest/index.html all green again |
|
SOFCI TEST |
|
The device test is triggered to test lnl pr test |
|
There's no DSP runtime PM on I just reproduced even without this PR, so it's unrelated. Merging. |
This unblocks #1170 and allows "installing" the func_exit_handler() trap in start_test() which will allow making start_test() optional. Signed-off-by: Marc Herbert <marc.herbert@intel.com>
|
That's because |
This is the first step towards fixing: