Post RECV WRs in batches to reduce the hardware doorbell rate per
transport. This helps the RPC-over-RDMA client scale better in
number of transports.
Signed-off-by: Chuck Lever <chuck.lever@xxxxxxxxxx>
---
net/sunrpc/xprtrdma/verbs.c | 2 ++
net/sunrpc/xprtrdma/xprt_rdma.h | 10 ++++++++++
2 files changed, 12 insertions(+)
diff --git a/net/sunrpc/xprtrdma/verbs.c b/net/sunrpc/xprtrdma/verbs.c
index 4994e75..b4e997d 100644
--- a/net/sunrpc/xprtrdma/verbs.c
+++ b/net/sunrpc/xprtrdma/verbs.c
@@ -1480,6 +1480,8 @@ struct rpcrdma_regbuf *
if (ep->rep_receive_count > needed)
goto out;
needed -= ep->rep_receive_count;
+ if (!temp)
+ needed += RPCRDMA_MAX_RECV_BATCH;
count = 0;
wr = NULL;
diff --git a/net/sunrpc/xprtrdma/xprt_rdma.h b/net/sunrpc/xprtrdma/xprt_rdma.h
index 33db208..10f6593 100644
--- a/net/sunrpc/xprtrdma/xprt_rdma.h
+++ b/net/sunrpc/xprtrdma/xprt_rdma.h
@@ -205,6 +205,16 @@ struct rpcrdma_rep {
struct ib_recv_wr rr_recv_wr;
};
+/* To reduce the rate at which a transport invokes ib_post_recv
+ * (and thus the hardware doorbell rate), xprtrdma posts Receive
+ * WRs in batches.
+ *
+ * Setting this to zero disables Receive post batching.
+ */
+enum {
+ RPCRDMA_MAX_RECV_BATCH = 7,
+};
+
/* struct rpcrdma_sendctx - DMA mapped SGEs to unmap after Send completes
*/
struct rpcrdma_req;
