On 1/2/17, 5:42 AM, "Bart Van Assche" <Bart.VanAssche@xxxxxxxxxxx> wrote:
>Change do_complete() such that an error completion is not only
>generated if a QP is in the error state but also if a work request
>failed.
>
>Signed-off-by: Bart Van Assche <bart.vanassche@xxxxxxxxxxx>
>Cc: Moni Shoua <monis@xxxxxxxxxxxx>
>Cc: Andrew Boyer <andrew.boyer@xxxxxxxx>
>---
> drivers/infiniband/sw/rxe/rxe_comp.c | 3 ++-
> drivers/infiniband/sw/rxe/rxe_loc.h | 1 +
> drivers/infiniband/sw/rxe/rxe_req.c | 18 +++++++-----------
> drivers/infiniband/sw/rxe/rxe_resp.c | 2 +-
> 4 files changed, 11 insertions(+), 13 deletions(-)
>
>diff --git a/drivers/infiniband/sw/rxe/rxe_comp.c
>b/drivers/infiniband/sw/rxe/rxe_comp.c
>index 6769a075501e..d1c3411c5478 100644
>--- a/drivers/infiniband/sw/rxe/rxe_comp.c
>+++ b/drivers/infiniband/sw/rxe/rxe_comp.c
>@@ -418,7 +418,7 @@ static void do_complete(struct rxe_qp *qp, struct
>rxe_send_wqe *wqe)
>
> if ((qp->sq_sig_type == IB_SIGNAL_ALL_WR) ||
> (wqe->wr.send_flags & IB_SEND_SIGNALED) ||
>- (qp->req.state == QP_STATE_ERROR)) {
>+ wqe->status != IB_WC_SUCCESS) {
> make_send_cqe(qp, wqe, &cqe);
> advance_consumer(qp->sq.queue);
> rxe_cq_post(qp->scq, &cqe, 0);
>@@ -709,6 +709,7 @@ int rxe_completer(void *arg)
> break;
>
> case COMPST_ERROR:
>+ WARN_ON_ONCE(wqe->status == IB_WC_SUCCESS);
> do_complete(qp, wqe);
> rxe_qp_error(qp);
>
>diff --git a/drivers/infiniband/sw/rxe/rxe_loc.h
>b/drivers/infiniband/sw/rxe/rxe_loc.h
>index da191d7acb6f..bdec460f1fce 100644
>--- a/drivers/infiniband/sw/rxe/rxe_loc.h
>+++ b/drivers/infiniband/sw/rxe/rxe_loc.h
>@@ -225,6 +225,7 @@ extern struct ib_dma_mapping_ops rxe_dma_mapping_ops;
>
> void rxe_release(struct kref *kref);
>
>+void rxe_drain_req_pkts(struct rxe_qp *qp, bool notify);
> int rxe_completer(void *arg);
> int rxe_requester(void *arg);
> int rxe_responder(void *arg);
>diff --git a/drivers/infiniband/sw/rxe/rxe_req.c
>b/drivers/infiniband/sw/rxe/rxe_req.c
>index b95b6035e988..64999f529981 100644
>--- a/drivers/infiniband/sw/rxe/rxe_req.c
>+++ b/drivers/infiniband/sw/rxe/rxe_req.c
>@@ -594,9 +594,14 @@ int rxe_requester(void *arg)
> rxe_add_ref(qp);
>
> next_wqe:
>- if (unlikely(!qp->valid || qp->req.state == QP_STATE_ERROR))
>+ if (unlikely(!qp->valid))
> goto exit;
>
>+ if (unlikely(qp->req.state == QP_STATE_ERROR)) {
>+ rxe_drain_req_pkts(qp, true);
>+ goto exit;
>+ }
>+
> if (unlikely(qp->req.state == QP_STATE_RESET)) {
> qp->req.wqe_index = consumer_index(qp->sq.queue);
> qp->req.opcode = -1;
>@@ -743,17 +748,8 @@ int rxe_requester(void *arg)
> kfree_skb(skb);
> wqe->status = IB_WC_LOC_PROT_ERR;
> wqe->state = wqe_state_error;
>-
>- /*
>- * IBA Spec. Section 10.7.3.1 SIGNALED COMPLETIONS
>- * ---------8<---------8<-------------
>- * ...Note that if a completion error occurs, a Work Completion
>- * will always be generated, even if the signaling
>- * indicator requests an Unsignaled Completion.
>- * ---------8<---------8<-------------
>- */
It wouldn¹t hurt to move this comment over to the top of do_complete() to
explain the wqe->status != IB_WC_SUCCESS test.
Reviewed-by: Andrew Boyer <andrew.boyer@xxxxxxxx>
>- wqe->wr.send_flags |= IB_SEND_SIGNALED;
> __rxe_do_task(&qp->comp.task);
>+
> exit:
> rxe_drop_ref(qp);
> return -EAGAIN;
>diff --git a/drivers/infiniband/sw/rxe/rxe_resp.c
>b/drivers/infiniband/sw/rxe/rxe_resp.c
>index 25951e9413b8..33defaddc000 100644
>--- a/drivers/infiniband/sw/rxe/rxe_resp.c
>+++ b/drivers/infiniband/sw/rxe/rxe_resp.c
>@@ -1207,7 +1207,7 @@ static enum resp_states do_class_d1e_error(struct
>rxe_qp *qp)
> }
> }
>
>-static void rxe_drain_req_pkts(struct rxe_qp *qp, bool notify)
>+void rxe_drain_req_pkts(struct rxe_qp *qp, bool notify)
> {
> struct sk_buff *skb;
>
>--
>2.11.0
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
