Merge with /pub/scm/linux/kernel/git/torvalds/linux-2.6.git

10 files changed, 137 insertions, 29 deletions

diff --git a/drivers/infiniband/hw/mthca/mthca_catas.c b/drivers/infiniband/hw/mthca/mthca_catas.c
index 25ebab64bc4..c3bec7490f5 100644
--- a/drivers/infiniband/hw/mthca/mthca_catas.c
+++ b/drivers/infiniband/hw/mthca/mthca_catas.c
@@ -97,7 +97,7 @@ static void poll_catas(unsigned long dev_ptr)
 		}
 
 	spin_lock_irqsave(&catas_lock, flags);
-	if (dev->catas_err.stop)
+	if (!dev->catas_err.stop)
 		mod_timer(&dev->catas_err.timer,
 			  jiffies + MTHCA_CATAS_POLL_INTERVAL);
 	spin_unlock_irqrestore(&catas_lock, flags);
diff --git a/drivers/infiniband/hw/mthca/mthca_cmd.c b/drivers/infiniband/hw/mthca/mthca_cmd.c
index 49f211d55df..9ed34587fc5 100644
--- a/drivers/infiniband/hw/mthca/mthca_cmd.c
+++ b/drivers/infiniband/hw/mthca/mthca_cmd.c
@@ -1060,6 +1060,8 @@ int mthca_QUERY_DEV_LIM(struct mthca_dev *dev,
 		dev_lim->hca.arbel.resize_srq = field & 1;
 		MTHCA_GET(field, outbox, QUERY_DEV_LIM_MAX_SG_RQ_OFFSET);
 		dev_lim->max_sg = min_t(int, field, dev_lim->max_sg);
+		MTHCA_GET(size, outbox, QUERY_DEV_LIM_MAX_DESC_SZ_RQ_OFFSET);
+		dev_lim->max_desc_sz = min_t(int, size, dev_lim->max_desc_sz);
 		MTHCA_GET(size, outbox, QUERY_DEV_LIM_MPT_ENTRY_SZ_OFFSET);
 		dev_lim->mpt_entry_sz = size;
 		MTHCA_GET(field, outbox, QUERY_DEV_LIM_PBL_SZ_OFFSET);
diff --git a/drivers/infiniband/hw/mthca/mthca_cq.c b/drivers/infiniband/hw/mthca/mthca_cq.c
index f98e2355582..4a8adcef207 100644
--- a/drivers/infiniband/hw/mthca/mthca_cq.c
+++ b/drivers/infiniband/hw/mthca/mthca_cq.c
@@ -258,7 +258,7 @@ void mthca_cq_clean(struct mthca_dev *dev, u32 cqn, u32 qpn,
 {
 	struct mthca_cq *cq;
 	struct mthca_cqe *cqe;
-	int prod_index;
+	u32 prod_index;
 	int nfreed = 0;
 
 	spin_lock_irq(&dev->cq_table.lock);
@@ -293,19 +293,15 @@ void mthca_cq_clean(struct mthca_dev *dev, u32 cqn, u32 qpn,
 	 * Now sweep backwards through the CQ, removing CQ entries
 	 * that match our QP by copying older entries on top of them.
 	 */
-	while (prod_index > cq->cons_index) {
-		cqe = get_cqe(cq, (prod_index - 1) & cq->ibcq.cqe);
+	while ((int) --prod_index - (int) cq->cons_index >= 0) {
+		cqe = get_cqe(cq, prod_index & cq->ibcq.cqe);
 		if (cqe->my_qpn == cpu_to_be32(qpn)) {
 			if (srq)
 				mthca_free_srq_wqe(srq, be32_to_cpu(cqe->wqe));
 			++nfreed;
-		}
-		else if (nfreed)
-			memcpy(get_cqe(cq, (prod_index - 1 + nfreed) &
-				       cq->ibcq.cqe),
-			       cqe,
-			       MTHCA_CQ_ENTRY_SIZE);
-		--prod_index;
+		} else if (nfreed)
+			memcpy(get_cqe(cq, (prod_index + nfreed) & cq->ibcq.cqe),
+			       cqe, MTHCA_CQ_ENTRY_SIZE);
 	}
 
 	if (nfreed) {
diff --git a/drivers/infiniband/hw/mthca/mthca_dev.h b/drivers/infiniband/hw/mthca/mthca_dev.h
index e7e5d3b4f00..497ff794ef6 100644
--- a/drivers/infiniband/hw/mthca/mthca_dev.h
+++ b/drivers/infiniband/hw/mthca/mthca_dev.h
@@ -131,6 +131,7 @@ struct mthca_limits {
 	int      max_sg;
 	int      num_qps;
 	int      max_wqes;
+	int	 max_desc_sz;
 	int	 max_qp_init_rdma;
 	int      reserved_qps;
 	int      num_srqs;
@@ -154,6 +155,7 @@ struct mthca_limits {
 	int      reserved_mcgs;
 	int      num_pds;
 	int      reserved_pds;
+	u32      page_size_cap;
 	u32      flags;
 	u8       port_width_cap;
 };
diff --git a/drivers/infiniband/hw/mthca/mthca_main.c b/drivers/infiniband/hw/mthca/mthca_main.c
index 45c6328e780..147f248a807 100644
--- a/drivers/infiniband/hw/mthca/mthca_main.c
+++ b/drivers/infiniband/hw/mthca/mthca_main.c
@@ -168,6 +168,7 @@ static int __devinit mthca_dev_lim(struct mthca_dev *mdev, struct mthca_dev_lim
 	mdev->limits.max_srq_wqes       = dev_lim->max_srq_sz;
 	mdev->limits.reserved_srqs      = dev_lim->reserved_srqs;
 	mdev->limits.reserved_eecs      = dev_lim->reserved_eecs;
+	mdev->limits.max_desc_sz        = dev_lim->max_desc_sz;
 	/*
 	 * Subtract 1 from the limit because we need to allocate a
 	 * spare CQE so the HCA HW can tell the difference between an
@@ -181,6 +182,7 @@ static int __devinit mthca_dev_lim(struct mthca_dev *mdev, struct mthca_dev_lim
 	mdev->limits.reserved_uars      = dev_lim->reserved_uars;
 	mdev->limits.reserved_pds       = dev_lim->reserved_pds;
 	mdev->limits.port_width_cap     = dev_lim->max_port_width;
+	mdev->limits.page_size_cap      = ~(u32) (dev_lim->min_page_sz - 1);
 	mdev->limits.flags              = dev_lim->flags;
 
 	/* IB_DEVICE_RESIZE_MAX_WR not supported by driver.
diff --git a/drivers/infiniband/hw/mthca/mthca_provider.c b/drivers/infiniband/hw/mthca/mthca_provider.c
index 6b016666826..4cc7e2846df 100644
--- a/drivers/infiniband/hw/mthca/mthca_provider.c
+++ b/drivers/infiniband/hw/mthca/mthca_provider.c
@@ -90,6 +90,7 @@ static int mthca_query_device(struct ib_device *ibdev,
 	memcpy(&props->node_guid,      out_mad->data + 12, 8);
 
 	props->max_mr_size         = ~0ull;
+	props->page_size_cap       = mdev->limits.page_size_cap;
 	props->max_qp              = mdev->limits.num_qps - mdev->limits.reserved_qps;
 	props->max_qp_wr           = mdev->limits.max_wqes;
 	props->max_sge             = mdev->limits.max_sg;
@@ -615,11 +616,11 @@ static struct ib_qp *mthca_create_qp(struct ib_pd *pd,
 		return ERR_PTR(err);
 	}
 
-	init_attr->cap.max_inline_data = 0;
 	init_attr->cap.max_send_wr     = qp->sq.max;
 	init_attr->cap.max_recv_wr     = qp->rq.max;
 	init_attr->cap.max_send_sge    = qp->sq.max_gs;
 	init_attr->cap.max_recv_sge    = qp->rq.max_gs;
+	init_attr->cap.max_inline_data = qp->max_inline_data;
 
 	return &qp->ibqp;
 }
diff --git a/drivers/infiniband/hw/mthca/mthca_provider.h b/drivers/infiniband/hw/mthca/mthca_provider.h
index bcd4b01a339..1e73947b470 100644
--- a/drivers/infiniband/hw/mthca/mthca_provider.h
+++ b/drivers/infiniband/hw/mthca/mthca_provider.h
@@ -251,6 +251,7 @@ struct mthca_qp {
 	struct mthca_wq        sq;
 	enum ib_sig_type       sq_policy;
 	int                    send_wqe_offset;
+	int                    max_inline_data;
 
 	u64                   *wrid;
 	union mthca_buf	       queue;
diff --git a/drivers/infiniband/hw/mthca/mthca_qp.c b/drivers/infiniband/hw/mthca/mthca_qp.c
index 8852ea477c2..760c418d5bc 100644
--- a/drivers/infiniband/hw/mthca/mthca_qp.c
+++ b/drivers/infiniband/hw/mthca/mthca_qp.c
@@ -885,6 +885,48 @@ int mthca_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr, int attr_mask)
 	return err;
 }
 
+static void mthca_adjust_qp_caps(struct mthca_dev *dev,
+				 struct mthca_pd *pd,
+				 struct mthca_qp *qp)
+{
+	int max_data_size;
+
+	/*
+	 * Calculate the maximum size of WQE s/g segments, excluding
+	 * the next segment and other non-data segments.
+	 */
+	max_data_size = min(dev->limits.max_desc_sz, 1 << qp->sq.wqe_shift) -
+		sizeof (struct mthca_next_seg);
+
+	switch (qp->transport) {
+	case MLX:
+		max_data_size -= 2 * sizeof (struct mthca_data_seg);
+		break;
+
+	case UD:
+		if (mthca_is_memfree(dev))
+			max_data_size -= sizeof (struct mthca_arbel_ud_seg);
+		else
+			max_data_size -= sizeof (struct mthca_tavor_ud_seg);
+		break;
+
+	default:
+		max_data_size -= sizeof (struct mthca_raddr_seg);
+		break;
+	}
+
+	/* We don't support inline data for kernel QPs (yet). */
+	if (!pd->ibpd.uobject)
+		qp->max_inline_data = 0;
+        else
+		qp->max_inline_data = max_data_size - MTHCA_INLINE_HEADER_SIZE;
+
+	qp->sq.max_gs = max_data_size / sizeof (struct mthca_data_seg);
+	qp->rq.max_gs = (min(dev->limits.max_desc_sz, 1 << qp->rq.wqe_shift) -
+			sizeof (struct mthca_next_seg)) /
+			sizeof (struct mthca_data_seg);
+}
+
 /*
  * Allocate and register buffer for WQEs.  qp->rq.max, sq.max,
  * rq.max_gs and sq.max_gs must all be assigned.
@@ -902,27 +944,53 @@ static int mthca_alloc_wqe_buf(struct mthca_dev *dev,
 	size = sizeof (struct mthca_next_seg) +
 		qp->rq.max_gs * sizeof (struct mthca_data_seg);
 
+	if (size > dev->limits.max_desc_sz)
+		return -EINVAL;
+
 	for (qp->rq.wqe_shift = 6; 1 << qp->rq.wqe_shift < size;
 	     qp->rq.wqe_shift++)
 		; /* nothing */
 
-	size = sizeof (struct mthca_next_seg) +
-		qp->sq.max_gs * sizeof (struct mthca_data_seg);
+	size = qp->sq.max_gs * sizeof (struct mthca_data_seg);
 	switch (qp->transport) {
 	case MLX:
 		size += 2 * sizeof (struct mthca_data_seg);
 		break;
+
 	case UD:
-		if (mthca_is_memfree(dev))
-			size += sizeof (struct mthca_arbel_ud_seg);
-		else
-			size += sizeof (struct mthca_tavor_ud_seg);
+		size += mthca_is_memfree(dev) ?
+			sizeof (struct mthca_arbel_ud_seg) :
+			sizeof (struct mthca_tavor_ud_seg);
 		break;
+
+	case UC:
+		size += sizeof (struct mthca_raddr_seg);
+		break;
+
+	case RC:
+		size += sizeof (struct mthca_raddr_seg);
+		/*
+		 * An atomic op will require an atomic segment, a
+		 * remote address segment and one scatter entry.
+		 */
+		size = max_t(int, size,
+			     sizeof (struct mthca_atomic_seg) +
+			     sizeof (struct mthca_raddr_seg) +
+			     sizeof (struct mthca_data_seg));
+		break;
+
 	default:
-		/* bind seg is as big as atomic + raddr segs */
-		size += sizeof (struct mthca_bind_seg);
+		break;
 	}
 
+	/* Make sure that we have enough space for a bind request */
+	size = max_t(int, size, sizeof (struct mthca_bind_seg));
+
+	size += sizeof (struct mthca_next_seg);
+
+	if (size > dev->limits.max_desc_sz)
+		return -EINVAL;
+
 	for (qp->sq.wqe_shift = 6; 1 << qp->sq.wqe_shift < size;
 	     qp->sq.wqe_shift++)
 		; /* nothing */
@@ -1066,6 +1134,8 @@ static int mthca_alloc_qp_common(struct mthca_dev *dev,
 		return ret;
 	}
 
+	mthca_adjust_qp_caps(dev, pd, qp);
+
 	/*
 	 * If this is a userspace QP, we're done now.  The doorbells
 	 * will be allocated and buffers will be initialized in
@@ -1486,8 +1556,8 @@ int mthca_tavor_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 				}
 
 				wqe += sizeof (struct mthca_atomic_seg);
-				size += sizeof (struct mthca_raddr_seg) / 16 +
-					sizeof (struct mthca_atomic_seg);
+				size += (sizeof (struct mthca_raddr_seg) +
+					 sizeof (struct mthca_atomic_seg)) / 16;
 				break;
 
 			case IB_WR_RDMA_WRITE:
@@ -1637,6 +1707,7 @@ int mthca_tavor_post_receive(struct ib_qp *ibqp, struct ib_recv_wr *wr,
 {
 	struct mthca_dev *dev = to_mdev(ibqp->device);
 	struct mthca_qp *qp = to_mqp(ibqp);
+	__be32 doorbell[2];
 	unsigned long flags;
 	int err = 0;
 	int nreq;
@@ -1654,6 +1725,22 @@ int mthca_tavor_post_receive(struct ib_qp *ibqp, struct ib_recv_wr *wr,
 	ind = qp->rq.next_ind;
 
 	for (nreq = 0; wr; ++nreq, wr = wr->next) {
+		if (unlikely(nreq == MTHCA_TAVOR_MAX_WQES_PER_RECV_DB)) {
+			nreq = 0;
+
+			doorbell[0] = cpu_to_be32((qp->rq.next_ind << qp->rq.wqe_shift) | size0);
+			doorbell[1] = cpu_to_be32(qp->qpn << 8);
+
+			wmb();
+
+			mthca_write64(doorbell,
+				      dev->kar + MTHCA_RECEIVE_DOORBELL,
+				      MTHCA_GET_DOORBELL_LOCK(&dev->doorbell_lock));
+
+			qp->rq.head += MTHCA_TAVOR_MAX_WQES_PER_RECV_DB;
+			size0 = 0;
+		}
+
 		if (mthca_wq_overflow(&qp->rq, nreq, qp->ibqp.recv_cq)) {
 			mthca_err(dev, "RQ %06x full (%u head, %u tail,"
 					" %d max, %d nreq)\n", qp->qpn,
@@ -1711,8 +1798,6 @@ int mthca_tavor_post_receive(struct ib_qp *ibqp, struct ib_recv_wr *wr,
 
 out:
 	if (likely(nreq)) {
-		__be32 doorbell[2];
-
 		doorbell[0] = cpu_to_be32((qp->rq.next_ind << qp->rq.wqe_shift) | size0);
 		doorbell[1] = cpu_to_be32((qp->qpn << 8) | nreq);
 
@@ -1806,8 +1891,8 @@ int mthca_arbel_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 				}
 
 				wqe += sizeof (struct mthca_atomic_seg);
-				size += sizeof (struct mthca_raddr_seg) / 16 +
-					sizeof (struct mthca_atomic_seg);
+				size += (sizeof (struct mthca_raddr_seg) +
+					 sizeof (struct mthca_atomic_seg)) / 16;
 				break;
 
 			case IB_WR_RDMA_READ:
diff --git a/drivers/infiniband/hw/mthca/mthca_srq.c b/drivers/infiniband/hw/mthca/mthca_srq.c
index 26d5161fde0..f7d234295ef 100644
--- a/drivers/infiniband/hw/mthca/mthca_srq.c
+++ b/drivers/infiniband/hw/mthca/mthca_srq.c
@@ -417,6 +417,7 @@ int mthca_tavor_post_srq_recv(struct ib_srq *ibsrq, struct ib_recv_wr *wr,
 {
 	struct mthca_dev *dev = to_mdev(ibsrq->device);
 	struct mthca_srq *srq = to_msrq(ibsrq);
+	__be32 doorbell[2];
 	unsigned long flags;
 	int err = 0;
 	int first_ind;
@@ -432,6 +433,25 @@ int mthca_tavor_post_srq_recv(struct ib_srq *ibsrq, struct ib_recv_wr *wr,
 	first_ind = srq->first_free;
 
 	for (nreq = 0; wr; ++nreq, wr = wr->next) {
+		if (unlikely(nreq == MTHCA_TAVOR_MAX_WQES_PER_RECV_DB)) {
+			nreq = 0;
+
+			doorbell[0] = cpu_to_be32(first_ind << srq->wqe_shift);
+			doorbell[1] = cpu_to_be32(srq->srqn << 8);
+
+			/*
+			 * Make sure that descriptors are written
+			 * before doorbell is rung.
+			 */
+			wmb();
+
+			mthca_write64(doorbell,
+				      dev->kar + MTHCA_RECEIVE_DOORBELL,
+				      MTHCA_GET_DOORBELL_LOCK(&dev->doorbell_lock));
+
+			first_ind = srq->first_free;
+		}
+
 		ind = srq->first_free;
 
 		if (ind < 0) {
@@ -494,8 +514,6 @@ int mthca_tavor_post_srq_recv(struct ib_srq *ibsrq, struct ib_recv_wr *wr,
 	}
 
 	if (likely(nreq)) {
-		__be32 doorbell[2];
-
 		doorbell[0] = cpu_to_be32(first_ind << srq->wqe_shift);
 		doorbell[1] = cpu_to_be32((srq->srqn << 8) | nreq);
 
diff --git a/drivers/infiniband/hw/mthca/mthca_wqe.h b/drivers/infiniband/hw/mthca/mthca_wqe.h
index 1f4c0ff28f7..73f1c0b9021 100644
--- a/drivers/infiniband/hw/mthca/mthca_wqe.h
+++ b/drivers/infiniband/hw/mthca/mthca_wqe.h
@@ -49,7 +49,8 @@ enum {
 };
 
 enum {
-	MTHCA_INVAL_LKEY = 0x100
+	MTHCA_INVAL_LKEY			= 0x100,
+	MTHCA_TAVOR_MAX_WQES_PER_RECV_DB	= 256
 };
 
 struct mthca_next_seg {