Commit 25524288 authored by Chuck Lever's avatar Chuck Lever Committed by Anna Schumaker

xprtrdma: Fix corner cases when handling device removal

Michal Kalderon has found some corner cases around device unload
with active NFS mounts that I didn't have the imagination to test
when xprtrdma device removal was added last year.

- The ULP device removal handler is responsible for deallocating
  the PD. That wasn't clear to me initially, and my own testing
  suggested it was not necessary, but that is incorrect.

- The transport destruction path can no longer assume that there
  is a valid ID.

- When destroying a transport, ensure that ib_free_cq() is not
  invoked on a CQ that was already released.
Reported-by: default avatarMichal Kalderon <Michal.Kalderon@cavium.com>
Fixes: bebd0318 ("xprtrdma: Support unplugging an HCA from ...")
Signed-off-by: default avatarChuck Lever <chuck.lever@oracle.com>
Cc: stable@vger.kernel.org # v4.12+
Signed-off-by: default avatarAnna Schumaker <Anna.Schumaker@Netapp.com>
parent 57174593
...@@ -250,7 +250,6 @@ rpcrdma_conn_upcall(struct rdma_cm_id *id, struct rdma_cm_event *event) ...@@ -250,7 +250,6 @@ rpcrdma_conn_upcall(struct rdma_cm_id *id, struct rdma_cm_event *event)
wait_for_completion(&ia->ri_remove_done); wait_for_completion(&ia->ri_remove_done);
ia->ri_id = NULL; ia->ri_id = NULL;
ia->ri_pd = NULL;
ia->ri_device = NULL; ia->ri_device = NULL;
/* Return 1 to ensure the core destroys the id. */ /* Return 1 to ensure the core destroys the id. */
return 1; return 1;
...@@ -447,7 +446,9 @@ rpcrdma_ia_remove(struct rpcrdma_ia *ia) ...@@ -447,7 +446,9 @@ rpcrdma_ia_remove(struct rpcrdma_ia *ia)
ia->ri_id->qp = NULL; ia->ri_id->qp = NULL;
} }
ib_free_cq(ep->rep_attr.recv_cq); ib_free_cq(ep->rep_attr.recv_cq);
ep->rep_attr.recv_cq = NULL;
ib_free_cq(ep->rep_attr.send_cq); ib_free_cq(ep->rep_attr.send_cq);
ep->rep_attr.send_cq = NULL;
/* The ULP is responsible for ensuring all DMA /* The ULP is responsible for ensuring all DMA
* mappings and MRs are gone. * mappings and MRs are gone.
...@@ -460,6 +461,8 @@ rpcrdma_ia_remove(struct rpcrdma_ia *ia) ...@@ -460,6 +461,8 @@ rpcrdma_ia_remove(struct rpcrdma_ia *ia)
rpcrdma_dma_unmap_regbuf(req->rl_recvbuf); rpcrdma_dma_unmap_regbuf(req->rl_recvbuf);
} }
rpcrdma_mrs_destroy(buf); rpcrdma_mrs_destroy(buf);
ib_dealloc_pd(ia->ri_pd);
ia->ri_pd = NULL;
/* Allow waiters to continue */ /* Allow waiters to continue */
complete(&ia->ri_remove_done); complete(&ia->ri_remove_done);
...@@ -627,14 +630,16 @@ rpcrdma_ep_destroy(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia) ...@@ -627,14 +630,16 @@ rpcrdma_ep_destroy(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia)
{ {
cancel_delayed_work_sync(&ep->rep_connect_worker); cancel_delayed_work_sync(&ep->rep_connect_worker);
if (ia->ri_id->qp) { if (ia->ri_id && ia->ri_id->qp) {
rpcrdma_ep_disconnect(ep, ia); rpcrdma_ep_disconnect(ep, ia);
rdma_destroy_qp(ia->ri_id); rdma_destroy_qp(ia->ri_id);
ia->ri_id->qp = NULL; ia->ri_id->qp = NULL;
} }
ib_free_cq(ep->rep_attr.recv_cq); if (ep->rep_attr.recv_cq)
ib_free_cq(ep->rep_attr.send_cq); ib_free_cq(ep->rep_attr.recv_cq);
if (ep->rep_attr.send_cq)
ib_free_cq(ep->rep_attr.send_cq);
} }
/* Re-establish a connection after a device removal event. /* Re-establish a connection after a device removal event.
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment