tests: msgq: Fix uncheck return value of coverity scan #31788
Conversation
enjiamai
requested review from
andrewboie,
andyross and
nashif
as code owners
github-actions
bot
added
area: Kernel
area: Tests
Add a return value check of message queue test cases, to fix issue zephyrproject-rtos#31680, which is for CID :216795. Signed-off-by: Enjia Mai <enjiax.mai@intel.com>
enjiamai
force-pushed
the
fix-coverity-uncheck-return-value
branch
from
30efb8d to
d228115
Compare
| zassert_equal(ret, 0, NULL); | ||
|
|
||
| ret = k_msgq_put(&msgq1, &data[0], K_NO_WAIT); | ||
| zassert_equal(ret, 0, NULL); |
There was a problem hiding this comment.
Can you please help me understand? This test is put message to a full queue, and should get a -ENOMSG if a timeout is set to K_NO_WAIT (from description).
We init'd the msgq with size 1, so these lines here are to fill that buffer. And the assert checks that the message was sent. Then the next part (k_thread_create) is what executes our test scenario of a thread putting more data on a full queue. Is my understanding correct, or close?
If so, how do we confirm the test scenario was handled properly? Would we not need a zassert_equal(ret, -ENOMSG, NULL)?
There was a problem hiding this comment.
Considering this PR is to fix the unchecked return value, my clarification above should not block this PR.
There was a problem hiding this comment.
@jenmwms thanks for your review and approval, this is indeed to fix the coverity issue.
When it first call k_msgq_put (see line 435), it won't get a -ENOMSG because it is still not empty, (msgq->used == 0, msgq->max_msgs==1 , as long as used < max, it will return 0. see "kernel/msg_q.c" line 126, 148 please.
And your understanding is correct. If we keep calling k_msgq_put again with K_NO_WAIT, we will get -ENOMSG. like you can see in the created thread (line 231).
Hope this answer your question, thanks!
4 participants
Add a return value check of message queue test cases, to fix
issue #31680, which is for CID :216795.
Fixes #31680
Signed-off-by: Enjia Mai enjiax.mai@intel.com