| Message ID | 1498026775-7927-1-git-send-email-Ram.Amrani@cavium.com (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
On Wed, Jun 21, 2017 at 09:32:55AM +0300, Ram Amrani wrote: > Enhanced Direct Packet Mode doesn't require flipping of bytes as it > is sent directly to the transmitting HW and isn't processed by the > FW that requires specific order. > > Fixes: c0965e4fe6fe ("libqedr (qelr) verbs") > > Signed-off-by: Ram Amrani <Ram.Amrani@cavium.com> > --- > providers/qedr/qelr_verbs.c | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) Please try to avoid usage of dot mark at the end of patch subject. Thanks > > diff --git a/providers/qedr/qelr_verbs.c b/providers/qedr/qelr_verbs.c > index 676b18c..68429df 100644 > --- a/providers/qedr/qelr_verbs.c > +++ b/providers/qedr/qelr_verbs.c > @@ -1259,7 +1259,7 @@ static int __qelr_post_send(struct qelr_devctx *cxt, struct qelr_qp *qp, > swqe->wqe_size = 2; > swqe2 = (struct rdma_sq_send_wqe_2st *)qelr_chain_produce(&qp->sq.chain); > swqe->inv_key_or_imm_data = htobe32(htole32(wr->imm_data)); > - qelr_edpm_set_inv_imm(qp, &edpm, swqe->inv_key_or_imm_data); > + qelr_edpm_set_inv_imm(qp, &edpm, wr->imm_data); > qelr_prepare_sq_send_data(qp, &edpm, data_size, swqe, swqe2, wr); > qelr_edpm_set_msg_data(qp, &edpm, > QELR_IB_OPCODE_SEND_ONLY_WITH_IMMEDIATE, > @@ -1291,7 +1291,7 @@ static int __qelr_post_send(struct qelr_devctx *cxt, struct qelr_qp *qp, > rwqe->imm_data = htobe32(htole32(wr->imm_data)); > qelr_edpm_set_rdma_ext(qp, &edpm, wr->wr.rdma.remote_addr, > wr->wr.rdma.rkey); > - qelr_edpm_set_inv_imm(qp, &edpm, rwqe->imm_data); > + qelr_edpm_set_inv_imm(qp, &edpm, wr->imm_data); > rwqe2 = (struct rdma_sq_rdma_wqe_2nd *)qelr_chain_produce(&qp->sq.chain); > qelr_prepare_sq_rdma_data(qp, &edpm, data_size, rwqe, rwqe2, wr); > qelr_edpm_set_msg_data(qp, &edpm, > -- > 1.8.3.1 >
diff --git a/providers/qedr/qelr_verbs.c b/providers/qedr/qelr_verbs.c index 676b18c..68429df 100644 --- a/providers/qedr/qelr_verbs.c +++ b/providers/qedr/qelr_verbs.c @@ -1259,7 +1259,7 @@ static int __qelr_post_send(struct qelr_devctx *cxt, struct qelr_qp *qp, swqe->wqe_size = 2; swqe2 = (struct rdma_sq_send_wqe_2st *)qelr_chain_produce(&qp->sq.chain); swqe->inv_key_or_imm_data = htobe32(htole32(wr->imm_data)); - qelr_edpm_set_inv_imm(qp, &edpm, swqe->inv_key_or_imm_data); + qelr_edpm_set_inv_imm(qp, &edpm, wr->imm_data); qelr_prepare_sq_send_data(qp, &edpm, data_size, swqe, swqe2, wr); qelr_edpm_set_msg_data(qp, &edpm, QELR_IB_OPCODE_SEND_ONLY_WITH_IMMEDIATE, @@ -1291,7 +1291,7 @@ static int __qelr_post_send(struct qelr_devctx *cxt, struct qelr_qp *qp, rwqe->imm_data = htobe32(htole32(wr->imm_data)); qelr_edpm_set_rdma_ext(qp, &edpm, wr->wr.rdma.remote_addr, wr->wr.rdma.rkey); - qelr_edpm_set_inv_imm(qp, &edpm, rwqe->imm_data); + qelr_edpm_set_inv_imm(qp, &edpm, wr->imm_data); rwqe2 = (struct rdma_sq_rdma_wqe_2nd *)qelr_chain_produce(&qp->sq.chain); qelr_prepare_sq_rdma_data(qp, &edpm, data_size, rwqe, rwqe2, wr); qelr_edpm_set_msg_data(qp, &edpm,
Enhanced Direct Packet Mode doesn't require flipping of bytes as it is sent directly to the transmitting HW and isn't processed by the FW that requires specific order. Fixes: c0965e4fe6fe ("libqedr (qelr) verbs") Signed-off-by: Ram Amrani <Ram.Amrani@cavium.com> --- providers/qedr/qelr_verbs.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)