net/smc: preallocated memory for rdma work requests

The work requests for rdma writes are built in local variables within
function smc_tx_rdma_write(). This violates the rule that the work
request storage has to stay till the work request is confirmed by
a completion queue response.
This patch introduces preallocated memory for these work requests.
The storage is allocated, once a link (and thus a queue pair) is
established.
Signed-off-by: Ursula Braun <ubraun@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/smc/smc_cdc.c

@@ -21,13 +21,6 @@
 
 /********************************** send *************************************/
 
-struct smc_cdc_tx_pend {
-	struct smc_connection	*conn;		/* socket connection */
-	union smc_host_cursor	cursor;	/* tx sndbuf cursor sent */
-	union smc_host_cursor	p_cursor;	/* rx RMBE cursor produced */
-	u16			ctrl_seq;	/* conn. tx sequence # */
-};
-
 /* handler for send/transmission completion of a CDC msg */
 static void smc_cdc_tx_handler(struct smc_wr_tx_pend_priv *pnd_snd,
 			       struct smc_link *link,
@@ -61,12 +54,14 @@ static void smc_cdc_tx_handler(struct smc_wr_tx_pend_priv *pnd_snd,
 
 int smc_cdc_get_free_slot(struct smc_connection *conn,
 			  struct smc_wr_buf **wr_buf,
+			  struct smc_rdma_wr **wr_rdma_buf,
 			  struct smc_cdc_tx_pend **pend)
 {
 	struct smc_link *link = &conn->lgr->lnk[SMC_SINGLE_LINK];
 	int rc;
 
 	rc = smc_wr_tx_get_free_slot(link, smc_cdc_tx_handler, wr_buf,
+				     wr_rdma_buf,
 				     (struct smc_wr_tx_pend_priv **)pend);
 	if (!conn->alert_token_local)
 		/* abnormal termination */
@@ -121,7 +116,7 @@ static int smcr_cdc_get_slot_and_msg_send(struct smc_connection *conn)
 	struct smc_wr_buf *wr_buf;
 	int rc;
 
-	rc = smc_cdc_get_free_slot(conn, &wr_buf, &pend);
+	rc = smc_cdc_get_free_slot(conn, &wr_buf, NULL, &pend);
 	if (rc)
 		return rc;
 

net/smc/smc_cdc.h

@@ -270,10 +270,16 @@ static inline void smc_cdc_msg_to_host(struct smc_host_cdc_msg *local,
 		smcr_cdc_msg_to_host(local, peer, conn);
 }
 
-struct smc_cdc_tx_pend;
+struct smc_cdc_tx_pend {
+	struct smc_connection	*conn;		/* socket connection */
+	union smc_host_cursor	cursor;		/* tx sndbuf cursor sent */
+	union smc_host_cursor	p_cursor;	/* rx RMBE cursor produced */
+	u16			ctrl_seq;	/* conn. tx sequence # */
+};
 
 int smc_cdc_get_free_slot(struct smc_connection *conn,
 			  struct smc_wr_buf **wr_buf,
+			  struct smc_rdma_wr **wr_rdma_buf,
 			  struct smc_cdc_tx_pend **pend);
 void smc_cdc_tx_dismiss_slots(struct smc_connection *conn);
 int smc_cdc_msg_send(struct smc_connection *conn, struct smc_wr_buf *wr_buf,

net/smc/smc_core.h

@@ -52,6 +52,24 @@ enum smc_wr_reg_state {
 	FAILED		/* ib_wr_reg_mr response: failure */
 };
 
+struct smc_rdma_sge {				/* sges for RDMA writes */
+	struct ib_sge		wr_tx_rdma_sge[SMC_IB_MAX_SEND_SGE];
+};
+
+#define SMC_MAX_RDMA_WRITES	2		/* max. # of RDMA writes per
+						 * message send
+						 */
+
+struct smc_rdma_sges {				/* sges per message send */
+	struct smc_rdma_sge	tx_rdma_sge[SMC_MAX_RDMA_WRITES];
+};
+
+struct smc_rdma_wr {				/* work requests per message
+						 * send
+						 */
+	struct ib_rdma_wr	wr_tx_rdma[SMC_MAX_RDMA_WRITES];
+};
+
 struct smc_link {
 	struct smc_ib_device	*smcibdev;	/* ib-device */
 	u8			ibport;		/* port - values 1 | 2 */
@@ -64,6 +82,8 @@ struct smc_link {
 	struct smc_wr_buf	*wr_tx_bufs;	/* WR send payload buffers */
 	struct ib_send_wr	*wr_tx_ibs;	/* WR send meta data */
 	struct ib_sge		*wr_tx_sges;	/* WR send gather meta data */
+	struct smc_rdma_sges	*wr_tx_rdma_sges;/*RDMA WRITE gather meta data*/
+	struct smc_rdma_wr	*wr_tx_rdmas;	/* WR RDMA WRITE */
 	struct smc_wr_tx_pend	*wr_tx_pends;	/* WR send waiting for CQE */
 	/* above four vectors have wr_tx_cnt elements and use the same index */
 	dma_addr_t		wr_tx_dma_addr;	/* DMA address of wr_tx_bufs */

net/smc/smc_llc.c

@@ -166,7 +166,8 @@ static int smc_llc_add_pending_send(struct smc_link *link,
 {
 	int rc;
 
-	rc = smc_wr_tx_get_free_slot(link, smc_llc_tx_handler, wr_buf, pend);
+	rc = smc_wr_tx_get_free_slot(link, smc_llc_tx_handler, wr_buf, NULL,
+				     pend);
 	if (rc < 0)
 		return rc;
 	BUILD_BUG_ON_MSG(

net/smc/smc_tx.c

@@ -266,27 +266,23 @@ int smcd_tx_ism_write(struct smc_connection *conn, void *data, size_t len,
 
 /* sndbuf consumer: actual data transfer of one target chunk with RDMA write */
 static int smc_tx_rdma_write(struct smc_connection *conn, int peer_rmbe_offset,
-			     int num_sges, struct ib_sge sges[])
+			     int num_sges, struct ib_rdma_wr *rdma_wr)
 {
 	struct smc_link_group *lgr = conn->lgr;
-	struct ib_rdma_wr rdma_wr;
 	struct smc_link *link;
 	int rc;
 
-	memset(&rdma_wr, 0, sizeof(rdma_wr));
 	link = &lgr->lnk[SMC_SINGLE_LINK];
-	rdma_wr.wr.wr_id = smc_wr_tx_get_next_wr_id(link);
-	rdma_wr.wr.sg_list = sges;
-	rdma_wr.wr.num_sge = num_sges;
-	rdma_wr.wr.opcode = IB_WR_RDMA_WRITE;
-	rdma_wr.remote_addr =
+	rdma_wr->wr.wr_id = smc_wr_tx_get_next_wr_id(link);
+	rdma_wr->wr.num_sge = num_sges;
+	rdma_wr->remote_addr =
 		lgr->rtokens[conn->rtoken_idx][SMC_SINGLE_LINK].dma_addr +
 		/* RMBE within RMB */
 		conn->tx_off +
 		/* offset within RMBE */
 		peer_rmbe_offset;
-	rdma_wr.rkey = lgr->rtokens[conn->rtoken_idx][SMC_SINGLE_LINK].rkey;
-	rc = ib_post_send(link->roce_qp, &rdma_wr.wr, NULL);
+	rdma_wr->rkey = lgr->rtokens[conn->rtoken_idx][SMC_SINGLE_LINK].rkey;
+	rc = ib_post_send(link->roce_qp, &rdma_wr->wr, NULL);
 	if (rc) {
 		conn->local_tx_ctrl.conn_state_flags.peer_conn_abort = 1;
 		smc_lgr_terminate(lgr);
@@ -313,24 +309,25 @@ static inline void smc_tx_advance_cursors(struct smc_connection *conn,
 /* SMC-R helper for smc_tx_rdma_writes() */
 static int smcr_tx_rdma_writes(struct smc_connection *conn, size_t len,
 			       size_t src_off, size_t src_len,
-			       size_t dst_off, size_t dst_len)
+			       size_t dst_off, size_t dst_len,
+			       struct smc_rdma_wr *wr_rdma_buf)
 {
 	dma_addr_t dma_addr =
 		sg_dma_address(conn->sndbuf_desc->sgt[SMC_SINGLE_LINK].sgl);
-	struct smc_link *link = &conn->lgr->lnk[SMC_SINGLE_LINK];
 	int src_len_sum = src_len, dst_len_sum = dst_len;
-	struct ib_sge sges[SMC_IB_MAX_SEND_SGE];
 	int sent_count = src_off;
 	int srcchunk, dstchunk;
 	int num_sges;
 	int rc;
 
 	for (dstchunk = 0; dstchunk < 2; dstchunk++) {
+		struct ib_sge *sge =
+			wr_rdma_buf->wr_tx_rdma[dstchunk].wr.sg_list;
+
 		num_sges = 0;
 		for (srcchunk = 0; srcchunk < 2; srcchunk++) {
-			sges[srcchunk].addr = dma_addr + src_off;
-			sges[srcchunk].length = src_len;
-			sges[srcchunk].lkey = link->roce_pd->local_dma_lkey;
+			sge[srcchunk].addr = dma_addr + src_off;
+			sge[srcchunk].length = src_len;
 			num_sges++;
 
 			src_off += src_len;
@@ -343,7 +340,8 @@ 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, sges);
+		rc = smc_tx_rdma_write(conn, dst_off, num_sges,
+				       &wr_rdma_buf->wr_tx_rdma[dstchunk]);
 		if (rc)
 			return rc;
 		if (dst_len_sum == len)
@@ -402,7 +400,8 @@ static int smcd_tx_rdma_writes(struct smc_connection *conn, size_t len,
 /* sndbuf consumer: prepare all necessary (src&dst) chunks of data transmit;
  * usable snd_wnd as max transmit
  */
-static int smc_tx_rdma_writes(struct smc_connection *conn)
+static int smc_tx_rdma_writes(struct smc_connection *conn,
+			      struct smc_rdma_wr *wr_rdma_buf)
 {
 	size_t len, src_len, dst_off, dst_len; /* current chunk values */
 	union smc_host_cursor sent, prep, prod, cons;
@@ -463,7 +462,7 @@ static int smc_tx_rdma_writes(struct smc_connection *conn)
 					 dst_off, dst_len);
 	else
 		rc = smcr_tx_rdma_writes(conn, len, sent.count, src_len,
-					 dst_off, dst_len);
+					 dst_off, dst_len, wr_rdma_buf);
 	if (rc)
 		return rc;
 
@@ -484,11 +483,12 @@ static int smc_tx_rdma_writes(struct smc_connection *conn)
 static int smcr_tx_sndbuf_nonempty(struct smc_connection *conn)
 {
 	struct smc_cdc_producer_flags *pflags;
+	struct smc_rdma_wr *wr_rdma_buf;
 	struct smc_cdc_tx_pend *pend;
 	struct smc_wr_buf *wr_buf;
 	int rc;
 
-	rc = smc_cdc_get_free_slot(conn, &wr_buf, &pend);
+	rc = smc_cdc_get_free_slot(conn, &wr_buf, &wr_rdma_buf, &pend);
 	if (rc < 0) {
 		if (rc == -EBUSY) {
 			struct smc_sock *smc =
@@ -506,7 +506,7 @@ static int smcr_tx_sndbuf_nonempty(struct smc_connection *conn)
 
 	spin_lock_bh(&conn->send_lock);
 	if (!conn->local_tx_ctrl.prod_flags.urg_data_present) {
-		rc = smc_tx_rdma_writes(conn);
+		rc = smc_tx_rdma_writes(conn, wr_rdma_buf);
 		if (rc) {
 			smc_wr_tx_put_slot(&conn->lgr->lnk[SMC_SINGLE_LINK],
 					   (struct smc_wr_tx_pend_priv *)pend);
@@ -533,7 +533,7 @@ static int smcd_tx_sndbuf_nonempty(struct smc_connection *conn)
 
 	spin_lock_bh(&conn->send_lock);
 	if (!pflags->urg_data_present)
-		rc = smc_tx_rdma_writes(conn);
+		rc = smc_tx_rdma_writes(conn, NULL);
 	if (!rc)
 		rc = smcd_cdc_msg_send(conn);
 

net/smc/smc_wr.c

@@ -160,6 +160,7 @@ static inline int smc_wr_tx_get_free_slot_index(struct smc_link *link, u32 *idx)
  * @link:		Pointer to smc_link used to later send the message.
  * @handler:		Send completion handler function pointer.
  * @wr_buf:		Out value returns pointer to message buffer.
+ * @wr_rdma_buf:	Out value returns pointer to rdma work request.
  * @wr_pend_priv:	Out value returns pointer serving as handler context.
  *
  * Return: 0 on success, or -errno on error.
@@ -167,6 +168,7 @@ static inline int smc_wr_tx_get_free_slot_index(struct smc_link *link, u32 *idx)
 int smc_wr_tx_get_free_slot(struct smc_link *link,
 			    smc_wr_tx_handler handler,
 			    struct smc_wr_buf **wr_buf,
+			    struct smc_rdma_wr **wr_rdma_buf,
 			    struct smc_wr_tx_pend_priv **wr_pend_priv)
 {
 	struct smc_wr_tx_pend *wr_pend;
@@ -204,6 +206,8 @@ int smc_wr_tx_get_free_slot(struct smc_link *link,
 	wr_ib = &link->wr_tx_ibs[idx];
 	wr_ib->wr_id = wr_id;
 	*wr_buf = &link->wr_tx_bufs[idx];
+	if (wr_rdma_buf)
+		*wr_rdma_buf = &link->wr_tx_rdmas[idx];
 	*wr_pend_priv = &wr_pend->priv;
 	return 0;
 }
@@ -465,12 +469,26 @@ static void smc_wr_init_sge(struct smc_link *lnk)
 			lnk->wr_tx_dma_addr + i * SMC_WR_BUF_SIZE;
 		lnk->wr_tx_sges[i].length = SMC_WR_TX_SIZE;
 		lnk->wr_tx_sges[i].lkey = lnk->roce_pd->local_dma_lkey;
+		lnk->wr_tx_rdma_sges[i].tx_rdma_sge[0].wr_tx_rdma_sge[0].lkey =
+			lnk->roce_pd->local_dma_lkey;
+		lnk->wr_tx_rdma_sges[i].tx_rdma_sge[0].wr_tx_rdma_sge[1].lkey =
+			lnk->roce_pd->local_dma_lkey;
+		lnk->wr_tx_rdma_sges[i].tx_rdma_sge[1].wr_tx_rdma_sge[0].lkey =
+			lnk->roce_pd->local_dma_lkey;
+		lnk->wr_tx_rdma_sges[i].tx_rdma_sge[1].wr_tx_rdma_sge[1].lkey =
+			lnk->roce_pd->local_dma_lkey;
 		lnk->wr_tx_ibs[i].next = NULL;
 		lnk->wr_tx_ibs[i].sg_list = &lnk->wr_tx_sges[i];
 		lnk->wr_tx_ibs[i].num_sge = 1;
 		lnk->wr_tx_ibs[i].opcode = IB_WR_SEND;
 		lnk->wr_tx_ibs[i].send_flags =
 			IB_SEND_SIGNALED | IB_SEND_SOLICITED;
+		lnk->wr_tx_rdmas[i].wr_tx_rdma[0].wr.opcode = IB_WR_RDMA_WRITE;
+		lnk->wr_tx_rdmas[i].wr_tx_rdma[1].wr.opcode = IB_WR_RDMA_WRITE;
+		lnk->wr_tx_rdmas[i].wr_tx_rdma[0].wr.sg_list =
+			lnk->wr_tx_rdma_sges[i].tx_rdma_sge[0].wr_tx_rdma_sge;
+		lnk->wr_tx_rdmas[i].wr_tx_rdma[1].wr.sg_list =
+			lnk->wr_tx_rdma_sges[i].tx_rdma_sge[1].wr_tx_rdma_sge;
 	}
 	for (i = 0; i < lnk->wr_rx_cnt; i++) {
 		lnk->wr_rx_sges[i].addr =
@@ -521,8 +539,12 @@ void smc_wr_free_link_mem(struct smc_link *lnk)
 	lnk->wr_tx_mask = NULL;
 	kfree(lnk->wr_tx_sges);
 	lnk->wr_tx_sges = NULL;
+	kfree(lnk->wr_tx_rdma_sges);
+	lnk->wr_tx_rdma_sges = NULL;
 	kfree(lnk->wr_rx_sges);
 	lnk->wr_rx_sges = NULL;
+	kfree(lnk->wr_tx_rdmas);
+	lnk->wr_tx_rdmas = NULL;
 	kfree(lnk->wr_rx_ibs);
 	lnk->wr_rx_ibs = NULL;
 	kfree(lnk->wr_tx_ibs);
@@ -552,10 +574,20 @@ int smc_wr_alloc_link_mem(struct smc_link *link)
 				  GFP_KERNEL);
 	if (!link->wr_rx_ibs)
 		goto no_mem_wr_tx_ibs;
+	link->wr_tx_rdmas = kcalloc(SMC_WR_BUF_CNT,
+				    sizeof(link->wr_tx_rdmas[0]),
+				    GFP_KERNEL);
+	if (!link->wr_tx_rdmas)
+		goto no_mem_wr_rx_ibs;
+	link->wr_tx_rdma_sges = kcalloc(SMC_WR_BUF_CNT,
+					sizeof(link->wr_tx_rdma_sges[0]),
+					GFP_KERNEL);
+	if (!link->wr_tx_rdma_sges)
+		goto no_mem_wr_tx_rdmas;
 	link->wr_tx_sges = kcalloc(SMC_WR_BUF_CNT, sizeof(link->wr_tx_sges[0]),
 				   GFP_KERNEL);
 	if (!link->wr_tx_sges)
-		goto no_mem_wr_rx_ibs;
+		goto no_mem_wr_tx_rdma_sges;
 	link->wr_rx_sges = kcalloc(SMC_WR_BUF_CNT * 3,
 				   sizeof(link->wr_rx_sges[0]),
 				   GFP_KERNEL);
@@ -579,6 +611,10 @@ int smc_wr_alloc_link_mem(struct smc_link *link)
 	kfree(link->wr_rx_sges);
 no_mem_wr_tx_sges:
 	kfree(link->wr_tx_sges);
+no_mem_wr_tx_rdma_sges:
+	kfree(link->wr_tx_rdma_sges);
+no_mem_wr_tx_rdmas:
+	kfree(link->wr_tx_rdmas);
 no_mem_wr_rx_ibs:
 	kfree(link->wr_rx_ibs);
 no_mem_wr_tx_ibs:

net/smc/smc_wr.h

@@ -85,6 +85,7 @@ void smc_wr_add_dev(struct smc_ib_device *smcibdev);
 
 int smc_wr_tx_get_free_slot(struct smc_link *link, smc_wr_tx_handler handler,
 			    struct smc_wr_buf **wr_buf,
+			    struct smc_rdma_wr **wrs,
 			    struct smc_wr_tx_pend_priv **wr_pend_priv);
 int smc_wr_tx_put_slot(struct smc_link *link,
 		       struct smc_wr_tx_pend_priv *wr_pend_priv);