fix(test/drivers): fix ENOSYS check and test skipping. #1737
Conversation
poiana
added
kind/bug
| * dealing with ENOSYS syscalls, ie: syscalls that are defined but unimplemented, | ||
| * skipping the test. | ||
| */ | ||
| #define assert_syscall_state(syscall_state, syscall_name, ...) \ |
There was a problem hiding this comment.
First 3 parameters are mandatory, thus we are sure that this macro always expands to the correct _assert_syscall_state call, even if it is called with just 3 parameters, since the ... will then expand to third(, fourth, fifth) parameter.
|
/hold |
FedeDP
force-pushed
the
fix/enosys_drivers_test
branch
from
e95f833 to
104cfe5
Compare
We cannot call GTEST_SKIP() in a sub-routine of a test, but only in the test body. Signed-off-by: Federico Di Pierro <[email protected]>
FedeDP
force-pushed
the
fix/enosys_drivers_test
branch
from
104cfe5 to
a78a15f
Compare
… tests. Signed-off-by: Federico Di Pierro <[email protected]>
| int64_t errno_value = -errno; | ||
| assert_syscall_state(SYSCALL_FAILURE, "dup3", res); |
There was a problem hiding this comment.
Since assert_syscall_state now resets errno to 0, we must take errno value before calling the macro.
Same below.
| { | ||
| bool match = false; | ||
|
|
||
| if (errno == ENOSYS) | ||
| { | ||
| GTEST_SKIP() << "Syscall " << syscall_name << " not implemented" << std::endl; | ||
| // it is managed upward by assert_syscall_state macro. |
There was a problem hiding this comment.
We need to short-circuit it here, because otherwise we would FAIL the test below.
|
/unhold |
| m_fd = syscall(__NR_openat, dirfd, filename, flags, 0); | ||
| assert_syscall_state(SYSCALL_SUCCESS, "openat", m_fd, NOT_EQUAL, -1); | ||
| // Using the macro that deals with ENOSYS produces a build issue since GTEST_SKIP returns value in a ctor. | ||
| _assert_syscall_state(SYSCALL_SUCCESS, "openat", m_fd, NOT_EQUAL, -1); |
There was a problem hiding this comment.
not clear who calls GTEST_SKIP() here if we use the internal method and not the macro
There was a problem hiding this comment.
nobody :/ this is just a ctor unfortunately
AFAIK openat should be implemented everywhere...hopefully.
There was a problem hiding this comment.
Anyway, we will notice immediately since it would break the tests.
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: Andreagit97, FedeDP, LucaGuerra The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
What type of PR is this?
/kind bug
Any specific area of the project related to this PR?
/area tests
Does this PR require a change in the driver versions?
What this PR does / why we need it:
We cannot call GTEST_SKIP() in a sub-routine of a test, but only in the test body.
I kept the changes at the bare minumum to avoid touching all of the test cases...i had to introduce a simple macro though.
Which issue(s) this PR fixes:
Fixes #
Special notes for your reviewer:
Does this PR introduce a user-facing change?: