You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Like in this comment, currently uftrace gets SIG failure (meaning it's terminated by a signal) for every unit test function on RISC-V when it's built by gcc. Clang built versions are fine, so it should be a compiler issue.
The assembly output shows that the uftrace.unit_test section didn't have any data and filled with 0s. This info is needed for the unittest binary to find the test name and function address to run. We can skip the test with 0 value to prevent the segfault, but it'd be meaningless when it cannot run any tests. So I'd like to keep this issue open and to track any progress on it.
On RISC-V, at least some GCC doesn't set the uftrace test data in
the uftrace.unit_test section. I'm not sure what's the problem
exactly but we can handle that by filling the symbol and string from
the binary.
Fixed: #1833
Signed-off-by: Namhyung Kim <[email protected]>
Like in this comment, currently uftrace gets SIG failure (meaning it's terminated by a signal) for every unit test function on RISC-V when it's built by gcc. Clang built versions are fine, so it should be a compiler issue.
The assembly output shows that the
uftrace.unit_test
section didn't have any data and filled with 0s. This info is needed for the unittest binary to find the test name and function address to run. We can skip the test with 0 value to prevent the segfault, but it'd be meaningless when it cannot run any tests. So I'd like to keep this issue open and to track any progress on it.The text was updated successfully, but these errors were encountered: