Received-SPF: softfail (google.com: domain of transitioning linux-kernel-owner@vger.kernel.org does not designate 23.128.96.19 as permitted sender) client-ip=23.128.96.19;
From:   Guangguan Wang <guangguan.wang@linux.alibaba.com>
To:     kgraul@linux.ibm.com, davem@davemloft.net, edumazet@google.com,
        kuba@kernel.org, pabeni@redhat.com
Cc:     linux-s390@vger.kernel.org, netdev@vger.kernel.org,
        linux-kernel@vger.kernel.org
Subject: [PATCH net-next 2/2] net/smc: rdma write inline if qp has sufficient inline space
Date:   Fri, 13 May 2022 15:15:51 +0800
Message-Id: <20220513071551.22065-3-guangguan.wang@linux.alibaba.com>
In-Reply-To: <20220513071551.22065-1-guangguan.wang@linux.alibaba.com>
References: <20220513071551.22065-1-guangguan.wang@linux.alibaba.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

Rdma write with inline flag when sending small packages,
whose length is shorter than the qp's max_inline_data, can
help reducing latency.

In my test environment, which are 2 VMs running on the same
physical host and whose NICs(ConnectX-4Lx) are working on
SR-IOV mode, qperf shows 0.5us-0.7us improvement in latency.

Test command:
server: smc_run taskset -c 1 qperf
client: smc_run taskset -c 1 qperf <server ip> -oo \
		msg_size:1:2K:*2 -t 30 -vu tcp_lat

The results shown below:
msgsize     before       after
1B          11.2 us      10.6 us (-0.6 us)
2B          11.2 us      10.7 us (-0.5 us)
4B          11.3 us      10.7 us (-0.6 us)
8B          11.2 us      10.6 us (-0.6 us)
16B         11.3 us      10.7 us (-0.6 us)
32B         11.3 us      10.6 us (-0.7 us)
64B         11.2 us      11.2 us (0 us)
128B        11.2 us      11.2 us (0 us)
256B        11.2 us      11.2 us (0 us)
512B        11.4 us      11.3 us (-0.1 us)
1KB         11.4 us      11.5 us (0.1 us)
2KB         11.5 us      11.5 us (0 us)

Signed-off-by: Guangguan Wang <guangguan.wang@linux.alibaba.com>
---
 net/smc/smc_tx.c | 17 ++++++++++++-----
 1 file changed, 12 insertions(+), 5 deletions(-)

diff --git a/net/smc/smc_tx.c b/net/smc/smc_tx.c
index 98ca9229fe87..4294259b3588 100644
--- a/net/smc/smc_tx.c
+++ b/net/smc/smc_tx.c
@@ -391,12 +391,20 @@ static int smcr_tx_rdma_writes(struct smc_connection *conn, size_t len,
 	int rc;
 
 	for (dstchunk = 0; dstchunk < 2; dstchunk++) {
-		struct ib_sge *sge =
-			wr_rdma_buf->wr_tx_rdma[dstchunk].wr.sg_list;
+		struct ib_rdma_wr *wr = &wr_rdma_buf->wr_tx_rdma[dstchunk];
+		struct ib_sge *sge = wr->wr.sg_list;
+		u64 base_addr = dma_addr;
+
+		if (dst_len <= link->qp_attr.cap.max_inline_data) {
+			base_addr = (u64)conn->sndbuf_desc->cpu_addr;
+			wr->wr.send_flags |= IB_SEND_INLINE;
+		} else {
+			wr->wr.send_flags &= ~IB_SEND_INLINE;
+		}
 
 		num_sges = 0;
 		for (srcchunk = 0; srcchunk < 2; srcchunk++) {
-			sge[srcchunk].addr = dma_addr + src_off;
+			sge[srcchunk].addr = base_addr + src_off;
 			sge[srcchunk].length = src_len;
 			num_sges++;
 
@@ -410,8 +418,7 @@ static int smcr_tx_rdma_writes(struct smc_connection *conn, size_t len,
 			src_len = dst_len - src_len; /* remainder */
 			src_len_sum += src_len;
 		}
-		rc = smc_tx_rdma_write(conn, dst_off, num_sges,
-				       &wr_rdma_buf->wr_tx_rdma[dstchunk]);
+		rc = smc_tx_rdma_write(conn, dst_off, num_sges, wr);
 		if (rc)
 			return rc;
 		if (dst_len_sum == len)
-- 
2.24.3 (Apple Git-128)