Skip to content

bugfix: coll/han double free'd subcomms #13398

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Sep 26, 2025
Merged

bugfix: coll/han double free'd subcomms #13398

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
22 changes: 22 additions & 0 deletions ompi/mca/coll/han/coll_han_subcomms.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -46,6 +46,15 @@
(COMM)->c_coll->coll_##COLL##_module = (FALLBACKS).COLL.module; \
} while (0)

#define HAN_SUBCOM_EXTRA_RETAIN(COMM, PARENT_COMM) \
do \
{ \
if (OMPI_COMM_CID_IS_LOWER(COMM, PARENT_COMM)) { \
OMPI_COMM_SET_EXTRA_RETAIN(COMM); \
OBJ_RETAIN(COMM); \
} \
} while (0)

/*
* Routine that creates the local hierarchical sub-communicators
* Called each time a collective is called.
Expand Down Expand Up @@ -206,6 +215,11 @@ int mca_coll_han_comm_create_new(struct ompi_communicator_t *comm,
HAN_SUBCOM_RESTORE_COLLECTIVE(fallbacks, comm, han_module, scatterv);

OBJ_DESTRUCT(&comm_info);

/* Ensure these communicators aren't released before the parent comm */
HAN_SUBCOM_EXTRA_RETAIN(*low_comm, comm);
HAN_SUBCOM_EXTRA_RETAIN(*up_comm, comm);

return OMPI_SUCCESS;

return_with_error:
Expand Down Expand Up @@ -376,6 +390,14 @@ int mca_coll_han_comm_create(struct ompi_communicator_t *comm,
han_module->cached_up_comms = up_comms;
han_module->cached_vranks = vranks;

/* Ensure these communicators aren't released before the parent comm */
for(int i = 0; i < COLL_HAN_LOW_MODULES; i++) {
HAN_SUBCOM_EXTRA_RETAIN(low_comms[i], comm);
}
for(int i = 0; i < COLL_HAN_UP_MODULES; i++) {
HAN_SUBCOM_EXTRA_RETAIN(up_comms[i], comm);
}

/* Reset the saved collectives to point back to HAN */
HAN_SUBCOM_RESTORE_COLLECTIVE(fallbacks, comm, han_module, alltoall);
HAN_SUBCOM_RESTORE_COLLECTIVE(fallbacks, comm, han_module, alltoallv);
Expand Down