ompi icon indicating copy to clipboard operation
ompi copied to clipboard

Revoke coll/han subcomms when a comm is revoked (and general framework)

Open Matthew-Whitlock opened this issue 2 months ago • 1 comments

Without a way to revoke the subcomms created by the coll modules, processes can get stuck waiting on sub-operations of a collective on a revoked communicator. I chose to put the top-level coll_revoke_local into the comm's c_coll object as a function pointer, but that's only to maintain the same programming pattern as the other functions. The top-level function is static, and calls coll_revoke_local on the comm for each of it's active modules.

Matthew-Whitlock avatar Sep 30 '25 20:09 Matthew-Whitlock