handle errors gracefuly to prevent SEGV #13194
Conversation
oob_allgather_test() do not check isend() call success, leading to the possibility to use oob_req->reqs[] un-initialized upon error and thus to SEGV. Signed-off-by: Bruno Faccini <[email protected]>
bfaccini
force-pushed
the
bfaccini/issue_13191
branch
from
3db0950 to
bc5e821
Compare
| MCA_PML_BASE_SEND_STANDARD, comm, &oob_req->reqs[0])); | ||
| MCA_PML_CALL(irecv(tmprecv, msglen, MPI_BYTE, recvfrom, | ||
| if (OMPI_SUCCESS != rc) { | ||
| return rc; |
There was a problem hiding this comment.
I'm not sure if conversion is safe here, return value of this function is ucc_status_t and might be not compatible with ompi error codes. If rc is positive number then it's actually not an error from ucc perspective since all errors in ucc are negative. Maybe just to make it safe the function returns UCC_ERR_NO_MESSAGE in case of error.
There was a problem hiding this comment.
I don't think this is correct. The macro ends up calling the .pml_isend field in the PML struct, which in the case of UCX points to mca_pml_ucx_isend, a function that returns OMPI errors.
jsquyres
dismissed
their stale review
Got the changes I asked but, but I'll still defer to those who are experts in this area of the code for providing a formal review
|
we will file a new PR for this change that adheres to our company policy |
|
we will file a new PR for this change that adheres to our company policy |
| oob_req->msglen = msglen; | ||
| oob_req->oob_coll_ctx = oob_coll_ctx; | ||
| oob_req->iter = 0; | ||
| oob_req->reqs[0] = NULL; |
There was a problem hiding this comment.
Technically I think this should be MPI_REQUEST_NULL.
4 participants
oob_allgather_test() do not check isend() call
success, leading to the possibility to use
oob_req->reqs[] un-initialized upon error and
thus to SEGV.