IB/srpt: Limit the send and receive queue sizes to what the HCA supports

Additionally, correct the comment about ch->rq_size.

Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Doug Ledford <dledford@redhat.com>

diff --git a/drivers/infiniband/ulp/srpt/ib_srpt.c b/drivers/infiniband/ulp/srpt/ib_srpt.c
index 95178b4e3565d46e6349a21319a8aad4ca30f571..76370e39857d49ea6e7eaf3b2ed6dc76f665608b 100644 (file)
--- a/drivers/infiniband/ulp/srpt/ib_srpt.c
+++ b/drivers/infiniband/ulp/srpt/ib_srpt.c
@@ -1650,7 +1650,7 @@ retry:
         * both both, as RDMA contexts will also post completions for the
         * RDMA READ case.
         */
-       qp_init->cap.max_send_wr = srp_sq_size / 2;
+       qp_init->cap.max_send_wr = min(srp_sq_size / 2, attrs->max_qp_wr + 0U);
        qp_init->cap.max_rdma_ctxs = srp_sq_size / 2;
        qp_init->cap.max_send_sge = min(attrs->max_sge, SRPT_MAX_SG_PER_WQE);
        qp_init->port_num = ch->sport->port;
@@ -1953,10 +1953,11 @@ static int srpt_cm_req_recv(struct ib_cm_id *cm_id,
        ch->cm_id = cm_id;
        cm_id->context = ch;
        /*
-        * Avoid QUEUE_FULL conditions by limiting the number of buffers used
-        * for the SRP protocol to the command queue size.
+        * ch->rq_size should be at least as large as the initiator queue
+        * depth to avoid that the initiator driver has to report QUEUE_FULL
+        * to the SCSI mid-layer.
         */
-       ch->rq_size = SRPT_RQ_SIZE;
+       ch->rq_size = min(SRPT_RQ_SIZE, sdev->device->attrs.max_qp_wr);
        spin_lock_init(&ch->spinlock);
        ch->state = CH_CONNECTING;
        INIT_LIST_HEAD(&ch->cmd_wait_list);