The immediate cause of the segfault is that in
test_trace_library_read(), we have:
...
handle = tracecmd_open(TRACECMD_FILE, 0);
CU_TEST(handle != NULL);
ret = tracecmd_iterate_events(handle, NULL, 0, read_events, &data);
CU_TEST(ret == 0);
...
and tracecmd_open() returns NULL. But, CU_TEST() failing is not fatal,
so the tracecmd_iterate_events() call still happens.
However, *why* tracecmd_open() fails (and returns NULL) on s390x only is
not clear. I found *where* it's failing, which is in handle_section():
...
section->data_offset = lseek(handle->fd, 0, SEEK_CUR);
if ((section->flags & TRACECMD_SEC_FL_COMPRESS) &&
in_uncompress_block(handle))
return -1;
...
Adding --compression none to the relevant trace-cmd invocations does
mask the issue.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2114514
Title:
trace-cmd: autopkgtest fails on s390x with seg fault
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/trace-cmd/+bug/2114514/+subscriptions
--
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
