Thread Sanitizer CI Crashes

[jblueh]

Observed behaviour Thread sanitizer tests in the CI pipeline fail across PRs, e.g. #2158, with errors along the lines

FATAL: ThreadSanitizer: unexpected memory mapping 0x7d81770ae000-0x7d8177500000

Background and possible causes

There is some incompatibility between sanitizers and high degrees of address space layout randomization (https://github.com/google/sanitizers/issues/1716), it seems to be triggered by a combination of too recent kernel and/or too old LLVM and affects also github runners (https://github.com/actions/runner-images/issues/9491).

Possible fixes

From what I understand in https://github.com/actions/runner-images/issues/9491 the corresponding parameters will be adapted in the runner, but it seems that there is also a workaround at the level of the workflow (https://github.com/openvswitch/ovs/commit/fa0dfa18d5344c82bc8387b4f79f4257a59d6f58).