iw_cxgb4: advertise the correct device max attributes

Advertise the actual max limits for things like qp depths, number of
qps, cqs, etc.

Clean up the queue allocation for qps and cqs.

Signed-off-by: Steve Wise <swise@opengridcomputing.com>
Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/infiniband/hw/cxgb4/cq.c

@@ -913,14 +913,8 @@ struct ib_cq *c4iw_create_cq(struct ib_device *ibdev, int entries,
 	/*
 	 * memsize must be a multiple of the page size if its a user cq.
 	 */
-	if (ucontext) {
+	if (ucontext)
 		memsize = roundup(memsize, PAGE_SIZE);
-		hwentries = memsize / sizeof *chp->cq.queue;
-		while (hwentries > rhp->rdev.hw_queue.t4_max_iq_size) {
-			memsize -= PAGE_SIZE;
-			hwentries = memsize / sizeof *chp->cq.queue;
-		}
-	}
 	chp->cq.size = hwentries;
 	chp->cq.memsize = memsize;
 	chp->cq.vector = vector;

drivers/infiniband/hw/cxgb4/device.c

@@ -934,17 +934,17 @@ static struct c4iw_dev *c4iw_alloc(const struct cxgb4_lld_info *infop)
 
 	devp->rdev.hw_queue.t4_eq_status_entries =
 		devp->rdev.lldi.sge_ingpadboundary > 64 ? 2 : 1;
-	devp->rdev.hw_queue.t4_max_eq_size =
-		65520 - devp->rdev.hw_queue.t4_eq_status_entries;
-	devp->rdev.hw_queue.t4_max_iq_size = 65520 - 1;
-	devp->rdev.hw_queue.t4_max_rq_size =
-		8192 - devp->rdev.hw_queue.t4_eq_status_entries;
+	devp->rdev.hw_queue.t4_max_eq_size = 65520;
+	devp->rdev.hw_queue.t4_max_iq_size = 65520;
+	devp->rdev.hw_queue.t4_max_rq_size = 8192 -
+		devp->rdev.hw_queue.t4_eq_status_entries - 1;
 	devp->rdev.hw_queue.t4_max_sq_size =
-		devp->rdev.hw_queue.t4_max_eq_size - 1;
+		devp->rdev.hw_queue.t4_max_eq_size -
+		devp->rdev.hw_queue.t4_eq_status_entries - 1;
 	devp->rdev.hw_queue.t4_max_qp_depth =
-		devp->rdev.hw_queue.t4_max_rq_size - 1;
+		devp->rdev.hw_queue.t4_max_rq_size;
 	devp->rdev.hw_queue.t4_max_cq_depth =
-		devp->rdev.hw_queue.t4_max_iq_size - 1;
+		devp->rdev.hw_queue.t4_max_iq_size - 2;
 	devp->rdev.hw_queue.t4_stat_len =
 		devp->rdev.lldi.sge_egrstatuspagesize;
 

drivers/infiniband/hw/cxgb4/provider.c

@@ -318,7 +318,7 @@ static int c4iw_query_device(struct ib_device *ibdev,
 	props->vendor_id = (u32)dev->rdev.lldi.pdev->vendor;
 	props->vendor_part_id = (u32)dev->rdev.lldi.pdev->device;
 	props->max_mr_size = T4_MAX_MR_SIZE;
-	props->max_qp = T4_MAX_NUM_QP;
+	props->max_qp = dev->rdev.lldi.vr->qp.size / 2;
 	props->max_qp_wr = dev->rdev.hw_queue.t4_max_qp_depth;
 	props->max_sge = T4_MAX_RECV_SGE;
 	props->max_sge_rd = 1;
@@ -326,7 +326,7 @@ static int c4iw_query_device(struct ib_device *ibdev,
 	props->max_qp_rd_atom = min(dev->rdev.lldi.max_ordird_qp,
 				    c4iw_max_read_depth);
 	props->max_qp_init_rd_atom = props->max_qp_rd_atom;
-	props->max_cq = T4_MAX_NUM_CQ;
+	props->max_cq = dev->rdev.lldi.vr->qp.size;
 	props->max_cqe = dev->rdev.hw_queue.t4_max_cq_depth;
 	props->max_mr = c4iw_num_stags(&dev->rdev);
 	props->max_pd = T4_MAX_NUM_PD;

drivers/infiniband/hw/cxgb4/qp.c

@@ -205,9 +205,9 @@ static int create_qp(struct c4iw_rdev *rdev, struct t4_wq *wq,
 	}
 
 	/*
-	 * RQT must be a power of 2.
+	 * RQT must be a power of 2 and at least 16 deep.
 	 */
-	wq->rq.rqt_size = roundup_pow_of_two(wq->rq.size);
+	wq->rq.rqt_size = roundup_pow_of_two(max_t(u16, wq->rq.size, 16));
 	wq->rq.rqt_hwaddr = c4iw_rqtpool_alloc(rdev, wq->rq.rqt_size);
 	if (!wq->rq.rqt_hwaddr) {
 		ret = -ENOMEM;
@@ -1621,13 +1621,17 @@ struct ib_qp *c4iw_create_qp(struct ib_pd *pd, struct ib_qp_init_attr *attrs,
 	if (attrs->cap.max_inline_data > T4_MAX_SEND_INLINE)
 		return ERR_PTR(-EINVAL);
 
-	rqsize = roundup(attrs->cap.max_recv_wr + 1, 16);
-	if (rqsize > rhp->rdev.hw_queue.t4_max_rq_size)
+	if (attrs->cap.max_recv_wr > rhp->rdev.hw_queue.t4_max_rq_size)
 		return ERR_PTR(-E2BIG);
+	rqsize = attrs->cap.max_recv_wr + 1;
+	if (rqsize < 8)
+		rqsize = 8;
 
-	sqsize = roundup(attrs->cap.max_send_wr + 1, 16);
-	if (sqsize > rhp->rdev.hw_queue.t4_max_sq_size)
+	if (attrs->cap.max_send_wr > rhp->rdev.hw_queue.t4_max_sq_size)
 		return ERR_PTR(-E2BIG);
+	sqsize = attrs->cap.max_send_wr + 1;
+	if (sqsize < 8)
+		sqsize = 8;
 
 	ucontext = pd->uobject ? to_c4iw_ucontext(pd->uobject->context) : NULL;
 
@@ -1635,19 +1639,20 @@ struct ib_qp *c4iw_create_qp(struct ib_pd *pd, struct ib_qp_init_attr *attrs,
 	if (!qhp)
 		return ERR_PTR(-ENOMEM);
 	qhp->wq.sq.size = sqsize;
-	qhp->wq.sq.memsize = (sqsize + 1) * sizeof *qhp->wq.sq.queue;
+	qhp->wq.sq.memsize =
+		(sqsize + rhp->rdev.hw_queue.t4_eq_status_entries) *
+		sizeof(*qhp->wq.sq.queue) + 16 * sizeof(__be64);
 	qhp->wq.sq.flush_cidx = -1;
 	qhp->wq.rq.size = rqsize;
-	qhp->wq.rq.memsize = (rqsize + 1) * sizeof *qhp->wq.rq.queue;
+	qhp->wq.rq.memsize =
+		(rqsize + rhp->rdev.hw_queue.t4_eq_status_entries) *
+		sizeof(*qhp->wq.rq.queue);
 
 	if (ucontext) {
 		qhp->wq.sq.memsize = roundup(qhp->wq.sq.memsize, PAGE_SIZE);
 		qhp->wq.rq.memsize = roundup(qhp->wq.rq.memsize, PAGE_SIZE);
 	}
 
-	PDBG("%s sqsize %u sqmemsize %zu rqsize %u rqmemsize %zu\n",
-	     __func__, sqsize, qhp->wq.sq.memsize, rqsize, qhp->wq.rq.memsize);
-
 	ret = create_qp(&rhp->rdev, &qhp->wq, &schp->cq, &rchp->cq,
 			ucontext ? &ucontext->uctx : &rhp->rdev.uctx);
 	if (ret)
@@ -1766,9 +1771,11 @@ struct ib_qp *c4iw_create_qp(struct ib_pd *pd, struct ib_qp_init_attr *attrs,
 	qhp->ibqp.qp_num = qhp->wq.sq.qid;
 	init_timer(&(qhp->timer));
 	INIT_LIST_HEAD(&qhp->db_fc_entry);
-	PDBG("%s qhp %p sq_num_entries %d, rq_num_entries %d qpid 0x%0x\n",
-	     __func__, qhp, qhp->attr.sq_num_entries, qhp->attr.rq_num_entries,
-	     qhp->wq.sq.qid);
+	PDBG("%s sq id %u size %u memsize %zu num_entries %u "
+	     "rq id %u size %u memsize %zu num_entries %u\n", __func__,
+	     qhp->wq.sq.qid, qhp->wq.sq.size, qhp->wq.sq.memsize,
+	     attrs->cap.max_send_wr, qhp->wq.rq.qid, qhp->wq.rq.size,
+	     qhp->wq.rq.memsize, attrs->cap.max_recv_wr);
 	return &qhp->ibqp;
 err8:
 	kfree(mm5);

drivers/infiniband/hw/cxgb4/t4.h

@@ -36,8 +36,6 @@
 #include "t4_msg.h"
 #include "t4fw_ri_api.h"
 
-#define T4_MAX_NUM_QP 65536
-#define T4_MAX_NUM_CQ 65536
 #define T4_MAX_NUM_PD 65536
 #define T4_MAX_NUM_STAG (1<<15)
 #define T4_MAX_MR_SIZE (~0ULL)