RDS: IB: Support Fastreg MR (FRMR) memory registration mode

Fastreg MR(FRMR) is another method with which one can
register memory to HCA. Some of the newer HCAs supports only fastreg
mr mode, so we need to add support for it to have RDS functional
on them.
Signed-off-by: Santosh Shilimkar <ssantosh@kernel.org>
Signed-off-by: Avinash Repaka <avinash.repaka@oracle.com>
Signed-off-by: Santosh Shilimkar <santosh.shilimkar@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/rds/Makefile

@@ -6,7 +6,7 @@ rds-y :=	af_rds.o bind.o cong.o connection.o info.o message.o   \
 obj-$(CONFIG_RDS_RDMA) += rds_rdma.o
 rds_rdma-y :=	rdma_transport.o \
 			ib.o ib_cm.o ib_recv.o ib_ring.o ib_send.o ib_stats.o \
-			ib_sysctl.o ib_rdma.o ib_fmr.o
+			ib_sysctl.o ib_rdma.o ib_fmr.o ib_frmr.o
 
 
 obj-$(CONFIG_RDS_TCP) += rds_tcp.o

net/rds/ib.h

@@ -349,6 +349,7 @@ int rds_ib_update_ipaddr(struct rds_ib_device *rds_ibdev, __be32 ipaddr);
 void rds_ib_add_conn(struct rds_ib_device *rds_ibdev, struct rds_connection *conn);
 void rds_ib_remove_conn(struct rds_ib_device *rds_ibdev, struct rds_connection *conn);
 void rds_ib_destroy_nodev_conns(void);
+void rds_ib_mr_cqe_handler(struct rds_ib_connection *ic, struct ib_wc *wc);
 
 /* ib_recv.c */
 int rds_ib_recv_init(void);

net/rds/ib_cm.c

@@ -249,7 +249,12 @@ static void poll_scq(struct rds_ib_connection *ic, struct ib_cq *cq,
 				 (unsigned long long)wc->wr_id, wc->status,
 				 wc->byte_len, be32_to_cpu(wc->ex.imm_data));
 
-			rds_ib_send_cqe_handler(ic, wc);
+			if (wc->wr_id <= ic->i_send_ring.w_nr ||
+			    wc->wr_id == RDS_IB_ACK_WR_ID)
+				rds_ib_send_cqe_handler(ic, wc);
+			else
+				rds_ib_mr_cqe_handler(ic, wc);
+
 		}
 	}
 }

net/rds/ib_mr.h

@@ -48,6 +48,21 @@ struct rds_ib_fmr {
 	u64			*dma;
 };
 
+enum rds_ib_fr_state {
+	FRMR_IS_FREE,	/* mr invalidated & ready for use */
+	FRMR_IS_INUSE,	/* mr is in use or used & can be invalidated */
+	FRMR_IS_STALE,	/* Stale MR and needs to be dropped  */
+};
+
+struct rds_ib_frmr {
+	struct ib_mr		*mr;
+	enum rds_ib_fr_state	fr_state;
+	bool			fr_inv;
+	struct ib_send_wr	fr_wr;
+	unsigned int		dma_npages;
+	unsigned int		sg_byte_len;
+};
+
 /* This is stored as mr->r_trans_private. */
 struct rds_ib_mr {
 	struct rds_ib_device		*device;
@@ -66,6 +81,7 @@ struct rds_ib_mr {
 
 	union {
 		struct rds_ib_fmr	fmr;
+		struct rds_ib_frmr	frmr;
 	} u;
 };
 
@@ -88,6 +104,7 @@ struct rds_ib_mr_pool {
 	unsigned long		max_items_soft;
 	unsigned long		max_free_pinned;
 	struct ib_fmr_attr	fmr_attr;
+	bool			use_fastreg;
 };
 
 extern struct workqueue_struct *rds_ib_mr_wq;
@@ -121,4 +138,11 @@ struct rds_ib_mr *rds_ib_try_reuse_ibmr(struct rds_ib_mr_pool *);
 void rds_ib_unreg_fmr(struct list_head *, unsigned int *,
 		      unsigned long *, unsigned int);
 void rds_ib_free_fmr_list(struct rds_ib_mr *);
+struct rds_ib_mr *rds_ib_reg_frmr(struct rds_ib_device *rds_ibdev,
+				  struct rds_ib_connection *ic,
+				  struct scatterlist *sg,
+				  unsigned long nents, u32 *key);
+void rds_ib_unreg_frmr(struct list_head *list, unsigned int *nfreed,
+		       unsigned long *unpinned, unsigned int goal);
+void rds_ib_free_frmr_list(struct rds_ib_mr *);
 #endif

net/rds/ib_rdma.c

@@ -397,7 +397,10 @@ int rds_ib_flush_mr_pool(struct rds_ib_mr_pool *pool,
 	if (list_empty(&unmap_list))
 		goto out;
 
-	rds_ib_unreg_fmr(&unmap_list, &nfreed, &unpinned, free_goal);
+	if (pool->use_fastreg)
+		rds_ib_unreg_frmr(&unmap_list, &nfreed, &unpinned, free_goal);
+	else
+		rds_ib_unreg_fmr(&unmap_list, &nfreed, &unpinned, free_goal);
 
 	if (!list_empty(&unmap_list)) {
 		/* we have to make sure that none of the things we're about
@@ -489,7 +492,10 @@ void rds_ib_free_mr(void *trans_private, int invalidate)
 	rdsdebug("RDS/IB: free_mr nents %u\n", ibmr->sg_len);
 
 	/* Return it to the pool's free list */
-	rds_ib_free_fmr_list(ibmr);
+	if (rds_ibdev->use_fastreg)
+		rds_ib_free_frmr_list(ibmr);
+	else
+		rds_ib_free_fmr_list(ibmr);
 
 	atomic_add(ibmr->sg_len, &pool->free_pinned);
 	atomic_inc(&pool->dirty_count);
@@ -534,6 +540,7 @@ void *rds_ib_get_mr(struct scatterlist *sg, unsigned long nents,
 {
 	struct rds_ib_device *rds_ibdev;
 	struct rds_ib_mr *ibmr = NULL;
+	struct rds_ib_connection *ic = rs->rs_conn->c_transport_data;
 	int ret;
 
 	rds_ibdev = rds_ib_get_device(rs->rs_bound_addr);
@@ -547,7 +554,10 @@ void *rds_ib_get_mr(struct scatterlist *sg, unsigned long nents,
 		goto out;
 	}
 
-	ibmr = rds_ib_reg_fmr(rds_ibdev, sg, nents, key_ret);
+	if (rds_ibdev->use_fastreg)
+		ibmr = rds_ib_reg_frmr(rds_ibdev, ic, sg, nents, key_ret);
+	else
+		ibmr = rds_ib_reg_fmr(rds_ibdev, sg, nents, key_ret);
 	if (ibmr)
 		rds_ibdev = NULL;
 
@@ -601,6 +611,7 @@ struct rds_ib_mr_pool *rds_ib_create_mr_pool(struct rds_ib_device *rds_ibdev,
 	pool->fmr_attr.max_maps = rds_ibdev->fmr_max_remaps;
 	pool->fmr_attr.page_shift = PAGE_SHIFT;
 	pool->max_items_soft = rds_ibdev->max_mrs * 3 / 4;
+	pool->use_fastreg = rds_ibdev->use_fastreg;
 
 	return pool;
 }