Merge tag 'rdma-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband

Pull infiniband updates from Roland Dreier:
 "Main batch of InfiniBand/RDMA changes for 3.15:

   - The biggest change is core API extensions and mlx5 low-level driver
     support for handling DIF/DIX-style protection information, and the
     addition of PI support to the iSER initiator.  Target support will
     be arriving shortly through the SCSI target tree.

   - A nice simplification to the "umem" memory pinning library now that
     we have chained sg lists.  Kudos to Yishai Hadas for realizing our
     code didn't have to be so crazy.

   - Another nice simplification to the sg wrappers used by qib, ipath
     and ehca to handle their mapping of memory to adapter.

   - The usual batch of fixes to bugs found by static checkers etc.
     from intrepid people like Dan Carpenter and Yann Droneaud.

   - A large batch of cxgb4, ocrdma, qib driver updates"

* tag 'rdma-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband: (102 commits)
  RDMA/ocrdma: Unregister inet notifier when unloading ocrdma
  RDMA/ocrdma: Fix warnings about pointer <-> integer casts
  RDMA/ocrdma: Code clean-up
  RDMA/ocrdma: Display FW version
  RDMA/ocrdma: Query controller information
  RDMA/ocrdma: Support non-embedded mailbox commands
  RDMA/ocrdma: Handle CQ overrun error
  RDMA/ocrdma: Display proper value for max_mw
  RDMA/ocrdma: Use non-zero tag in SRQ posting
  RDMA/ocrdma: Memory leak fix in ocrdma_dereg_mr()
  RDMA/ocrdma: Increment abi version count
  RDMA/ocrdma: Update version string
  be2net: Add abi version between be2net and ocrdma
  RDMA/ocrdma: ABI versioning between ocrdma and be2net
  RDMA/ocrdma: Allow DPP QP creation
  RDMA/ocrdma: Read ASIC_ID register to select asic_gen
  RDMA/ocrdma: SQ and RQ doorbell offset clean up
  RDMA/ocrdma: EQ full catastrophe avoidance
  RDMA/cxgb4: Disable DSGL use by default
  RDMA/cxgb4: rx_data() needs to hold the ep mutex
  ...

drivers/infiniband/core/cm.c

@@ -349,23 +349,6 @@ static void cm_init_av_for_response(struct cm_port *port, struct ib_wc *wc,
 			   grh, &av->ah_attr);
 }
 
-int ib_update_cm_av(struct ib_cm_id *id, const u8 *smac, const u8 *alt_smac)
-{
-	struct cm_id_private *cm_id_priv;
-
-	cm_id_priv = container_of(id, struct cm_id_private, id);
-
-	if (smac != NULL)
-		memcpy(cm_id_priv->av.smac, smac, sizeof(cm_id_priv->av.smac));
-
-	if (alt_smac != NULL)
-		memcpy(cm_id_priv->alt_av.smac, alt_smac,
-		       sizeof(cm_id_priv->alt_av.smac));
-
-	return 0;
-}
-EXPORT_SYMBOL(ib_update_cm_av);
-
 static int cm_init_av_by_path(struct ib_sa_path_rec *path, struct cm_av *av)
 {
 	struct cm_device *cm_dev;

drivers/infiniband/core/cma.c

@@ -1284,15 +1284,6 @@ static int cma_req_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event)
 	struct rdma_id_private *listen_id, *conn_id;
 	struct rdma_cm_event event;
 	int offset, ret;
-	u8 smac[ETH_ALEN];
-	u8 alt_smac[ETH_ALEN];
-	u8 *psmac = smac;
-	u8 *palt_smac = alt_smac;
-	int is_iboe = ((rdma_node_get_transport(cm_id->device->node_type) ==
-			RDMA_TRANSPORT_IB) &&
-		       (rdma_port_get_link_layer(cm_id->device,
-			ib_event->param.req_rcvd.port) ==
-			IB_LINK_LAYER_ETHERNET));
 
 	listen_id = cm_id->context;
 	if (!cma_check_req_qp_type(&listen_id->id, ib_event))
@@ -1336,28 +1327,11 @@ static int cma_req_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event)
 	ret = conn_id->id.event_handler(&conn_id->id, &event);
 	if (ret)
 		goto err3;
-
-	if (is_iboe) {
-		if (ib_event->param.req_rcvd.primary_path != NULL)
-			rdma_addr_find_smac_by_sgid(
-				&ib_event->param.req_rcvd.primary_path->sgid,
-				psmac, NULL);
-		else
-			psmac = NULL;
-		if (ib_event->param.req_rcvd.alternate_path != NULL)
-			rdma_addr_find_smac_by_sgid(
-				&ib_event->param.req_rcvd.alternate_path->sgid,
-				palt_smac, NULL);
-		else
-			palt_smac = NULL;
-	}
 	/*
 	 * Acquire mutex to prevent user executing rdma_destroy_id()
 	 * while we're accessing the cm_id.
 	 */
 	mutex_lock(&lock);
-	if (is_iboe)
-		ib_update_cm_av(cm_id, psmac, palt_smac);
 	if (cma_comp(conn_id, RDMA_CM_CONNECT) &&
 	    (conn_id->id.qp_type != IB_QPT_UD))
 		ib_send_cm_mra(cm_id, CMA_CM_MRA_SETTING, NULL, 0);

drivers/infiniband/core/mad.c

@@ -1022,12 +1022,21 @@ int ib_send_mad(struct ib_mad_send_wr_private *mad_send_wr)
 					mad_send_wr->send_buf.mad,
 					sge[0].length,
 					DMA_TO_DEVICE);
+	if (unlikely(ib_dma_mapping_error(mad_agent->device, sge[0].addr)))
+		return -ENOMEM;
+
 	mad_send_wr->header_mapping = sge[0].addr;
 
 	sge[1].addr = ib_dma_map_single(mad_agent->device,
 					ib_get_payload(mad_send_wr),
 					sge[1].length,
 					DMA_TO_DEVICE);
+	if (unlikely(ib_dma_mapping_error(mad_agent->device, sge[1].addr))) {
+		ib_dma_unmap_single(mad_agent->device,
+				    mad_send_wr->header_mapping,
+				    sge[0].length, DMA_TO_DEVICE);
+		return -ENOMEM;
+	}
 	mad_send_wr->payload_mapping = sge[1].addr;
 
 	spin_lock_irqsave(&qp_info->send_queue.lock, flags);
@@ -2590,6 +2599,11 @@ static int ib_mad_post_receive_mads(struct ib_mad_qp_info *qp_info,
 						 sizeof *mad_priv -
 						   sizeof mad_priv->header,
 						 DMA_FROM_DEVICE);
+		if (unlikely(ib_dma_mapping_error(qp_info->port_priv->device,
+						  sg_list.addr))) {
+			ret = -ENOMEM;
+			break;
+		}
 		mad_priv->header.mapping = sg_list.addr;
 		recv_wr.wr_id = (unsigned long)&mad_priv->header.mad_list;
 		mad_priv->header.mad_list.mad_queue = recv_queue;

drivers/infiniband/core/umem.c

@@ -42,29 +42,29 @@
 
 #include "uverbs.h"
 
-#define IB_UMEM_MAX_PAGE_CHUNK						\
-	((PAGE_SIZE - offsetof(struct ib_umem_chunk, page_list)) /	\
-	 ((void *) &((struct ib_umem_chunk *) 0)->page_list[1] -	\
-	  (void *) &((struct ib_umem_chunk *) 0)->page_list[0]))
 
 static void __ib_umem_release(struct ib_device *dev, struct ib_umem *umem, int dirty)
 {
-	struct ib_umem_chunk *chunk, *tmp;
+	struct scatterlist *sg;
+	struct page *page;
 	int i;
 
-	list_for_each_entry_safe(chunk, tmp, &umem->chunk_list, list) {
-		ib_dma_unmap_sg(dev, chunk->page_list,
-				chunk->nents, DMA_BIDIRECTIONAL);
-		for (i = 0; i < chunk->nents; ++i) {
-			struct page *page = sg_page(&chunk->page_list[i]);
+	if (umem->nmap > 0)
+		ib_dma_unmap_sg(dev, umem->sg_head.sgl,
+				umem->nmap,
+				DMA_BIDIRECTIONAL);
 
-			if (umem->writable && dirty)
-				set_page_dirty_lock(page);
-			put_page(page);
-		}
+	for_each_sg(umem->sg_head.sgl, sg, umem->npages, i) {
 
-		kfree(chunk);
+		page = sg_page(sg);
+		if (umem->writable && dirty)
+			set_page_dirty_lock(page);
+		put_page(page);
 	}
+
+	sg_free_table(&umem->sg_head);
+	return;
+
 }
 
 /**
@@ -81,15 +81,15 @@ struct ib_umem *ib_umem_get(struct ib_ucontext *context, unsigned long addr,
 	struct ib_umem *umem;
 	struct page **page_list;
 	struct vm_area_struct **vma_list;
-	struct ib_umem_chunk *chunk;
 	unsigned long locked;
 	unsigned long lock_limit;
 	unsigned long cur_base;
 	unsigned long npages;
 	int ret;
-	int off;
 	int i;
 	DEFINE_DMA_ATTRS(attrs);
+	struct scatterlist *sg, *sg_list_start;
+	int need_release = 0;
 
 	if (dmasync)
 		dma_set_attr(DMA_ATTR_WRITE_BARRIER, &attrs);
@@ -97,7 +97,7 @@ struct ib_umem *ib_umem_get(struct ib_ucontext *context, unsigned long addr,
 	if (!can_do_mlock())
 		return ERR_PTR(-EPERM);
 
-	umem = kmalloc(sizeof *umem, GFP_KERNEL);
+	umem = kzalloc(sizeof *umem, GFP_KERNEL);
 	if (!umem)
 		return ERR_PTR(-ENOMEM);
 
@@ -117,8 +117,6 @@ struct ib_umem *ib_umem_get(struct ib_ucontext *context, unsigned long addr,
 	/* We assume the memory is from hugetlb until proved otherwise */
 	umem->hugetlb   = 1;
 
-	INIT_LIST_HEAD(&umem->chunk_list);
-
 	page_list = (struct page **) __get_free_page(GFP_KERNEL);
 	if (!page_list) {
 		kfree(umem);
@@ -147,7 +145,18 @@ struct ib_umem *ib_umem_get(struct ib_ucontext *context, unsigned long addr,
 
 	cur_base = addr & PAGE_MASK;
 
-	ret = 0;
+	if (npages == 0) {
+		ret = -EINVAL;
+		goto out;
+	}
+
+	ret = sg_alloc_table(&umem->sg_head, npages, GFP_KERNEL);
+	if (ret)
+		goto out;
+
+	need_release = 1;
+	sg_list_start = umem->sg_head.sgl;
+
 	while (npages) {
 		ret = get_user_pages(current, current->mm, cur_base,
 				     min_t(unsigned long, npages,
@@ -157,54 +166,38 @@ struct ib_umem *ib_umem_get(struct ib_ucontext *context, unsigned long addr,
 		if (ret < 0)
 			goto out;
 
+		umem->npages += ret;
 		cur_base += ret * PAGE_SIZE;
 		npages   -= ret;
 
-		off = 0;
-
-		while (ret) {
-			chunk = kmalloc(sizeof *chunk + sizeof (struct scatterlist) *
-					min_t(int, ret, IB_UMEM_MAX_PAGE_CHUNK),
-					GFP_KERNEL);
-			if (!chunk) {
-				ret = -ENOMEM;
-				goto out;
-			}
-
-			chunk->nents = min_t(int, ret, IB_UMEM_MAX_PAGE_CHUNK);
-			sg_init_table(chunk->page_list, chunk->nents);
-			for (i = 0; i < chunk->nents; ++i) {
-				if (vma_list &&
-				    !is_vm_hugetlb_page(vma_list[i + off]))
-					umem->hugetlb = 0;
-				sg_set_page(&chunk->page_list[i], page_list[i + off], PAGE_SIZE, 0);
-			}
-
-			chunk->nmap = ib_dma_map_sg_attrs(context->device,
-							  &chunk->page_list[0],
-							  chunk->nents,
-							  DMA_BIDIRECTIONAL,
-							  &attrs);
-			if (chunk->nmap <= 0) {
-				for (i = 0; i < chunk->nents; ++i)
-					put_page(sg_page(&chunk->page_list[i]));
-				kfree(chunk);
-
-				ret = -ENOMEM;
-				goto out;
-			}
-
-			ret -= chunk->nents;
-			off += chunk->nents;
-			list_add_tail(&chunk->list, &umem->chunk_list);
+		for_each_sg(sg_list_start, sg, ret, i) {
+			if (vma_list && !is_vm_hugetlb_page(vma_list[i]))
+				umem->hugetlb = 0;
+
+			sg_set_page(sg, page_list[i], PAGE_SIZE, 0);
 		}
 
-		ret = 0;
+		/* preparing for next loop */
+		sg_list_start = sg;
 	}
 
+	umem->nmap = ib_dma_map_sg_attrs(context->device,
+				  umem->sg_head.sgl,
+				  umem->npages,
+				  DMA_BIDIRECTIONAL,
+				  &attrs);
+
+	if (umem->nmap <= 0) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	ret = 0;
+
 out:
 	if (ret < 0) {
-		__ib_umem_release(context->device, umem, 0);
+		if (need_release)
+			__ib_umem_release(context->device, umem, 0);
 		kfree(umem);
 	} else
 		current->mm->pinned_vm = locked;
@@ -278,17 +271,16 @@ EXPORT_SYMBOL(ib_umem_release);
 
 int ib_umem_page_count(struct ib_umem *umem)
 {
-	struct ib_umem_chunk *chunk;
 	int shift;
 	int i;
 	int n;
+	struct scatterlist *sg;
 
 	shift = ilog2(umem->page_size);
 
 	n = 0;
-	list_for_each_entry(chunk, &umem->chunk_list, list)
-		for (i = 0; i < chunk->nmap; ++i)
-			n += sg_dma_len(&chunk->page_list[i]) >> shift;
+	for_each_sg(umem->sg_head.sgl, sg, umem->nmap, i)
+		n += sg_dma_len(sg) >> shift;
 
 	return n;
 }

drivers/infiniband/core/verbs.c

@@ -1169,6 +1169,45 @@ int ib_dereg_mr(struct ib_mr *mr)
 }
 EXPORT_SYMBOL(ib_dereg_mr);
 
+struct ib_mr *ib_create_mr(struct ib_pd *pd,
+			   struct ib_mr_init_attr *mr_init_attr)
+{
+	struct ib_mr *mr;
+
+	if (!pd->device->create_mr)
+		return ERR_PTR(-ENOSYS);
+
+	mr = pd->device->create_mr(pd, mr_init_attr);
+
+	if (!IS_ERR(mr)) {
+		mr->device  = pd->device;
+		mr->pd      = pd;
+		mr->uobject = NULL;
+		atomic_inc(&pd->usecnt);
+		atomic_set(&mr->usecnt, 0);
+	}
+
+	return mr;
+}
+EXPORT_SYMBOL(ib_create_mr);
+
+int ib_destroy_mr(struct ib_mr *mr)
+{
+	struct ib_pd *pd;
+	int ret;
+
+	if (atomic_read(&mr->usecnt))
+		return -EBUSY;
+
+	pd = mr->pd;
+	ret = mr->device->destroy_mr(mr);
+	if (!ret)
+		atomic_dec(&pd->usecnt);
+
+	return ret;
+}
+EXPORT_SYMBOL(ib_destroy_mr);
+
 struct ib_mr *ib_alloc_fast_reg_mr(struct ib_pd *pd, int max_page_list_len)
 {
 	struct ib_mr *mr;
@@ -1398,3 +1437,11 @@ int ib_destroy_flow(struct ib_flow *flow_id)
 	return err;
 }
 EXPORT_SYMBOL(ib_destroy_flow);
+
+int ib_check_mr_status(struct ib_mr *mr, u32 check_mask,
+		       struct ib_mr_status *mr_status)
+{
+	return mr->device->check_mr_status ?
+		mr->device->check_mr_status(mr, check_mask, mr_status) : -ENOSYS;
+}
+EXPORT_SYMBOL(ib_check_mr_status);

drivers/infiniband/hw/amso1100/c2_provider.c

@@ -431,9 +431,9 @@ static struct ib_mr *c2_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 	u64 *pages;
 	u64 kva = 0;
 	int shift, n, len;
-	int i, j, k;
+	int i, k, entry;
 	int err = 0;
-	struct ib_umem_chunk *chunk;
+	struct scatterlist *sg;
 	struct c2_pd *c2pd = to_c2pd(pd);
 	struct c2_mr *c2mr;
 
@@ -452,10 +452,7 @@ static struct ib_mr *c2_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 	}
 
 	shift = ffs(c2mr->umem->page_size) - 1;
-
-	n = 0;
-	list_for_each_entry(chunk, &c2mr->umem->chunk_list, list)
-		n += chunk->nents;
+	n = c2mr->umem->nmap;
 
 	pages = kmalloc(n * sizeof(u64), GFP_KERNEL);
 	if (!pages) {
@@ -464,14 +461,12 @@ static struct ib_mr *c2_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 	}
 
 	i = 0;
-	list_for_each_entry(chunk, &c2mr->umem->chunk_list, list) {
-		for (j = 0; j < chunk->nmap; ++j) {
-			len = sg_dma_len(&chunk->page_list[j]) >> shift;
-			for (k = 0; k < len; ++k) {
-				pages[i++] =
-					sg_dma_address(&chunk->page_list[j]) +
-					(c2mr->umem->page_size * k);
-			}
+	for_each_sg(c2mr->umem->sg_head.sgl, sg, c2mr->umem->nmap, entry) {
+		len = sg_dma_len(sg) >> shift;
+		for (k = 0; k < len; ++k) {
+			pages[i++] =
+				sg_dma_address(sg) +
+				(c2mr->umem->page_size * k);
 		}
 	}
 

drivers/infiniband/hw/cxgb3/iwch_provider.c

@@ -618,14 +618,13 @@ static struct ib_mr *iwch_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 {
 	__be64 *pages;
 	int shift, n, len;
-	int i, j, k;
+	int i, k, entry;
 	int err = 0;
-	struct ib_umem_chunk *chunk;
 	struct iwch_dev *rhp;
 	struct iwch_pd *php;
 	struct iwch_mr *mhp;
 	struct iwch_reg_user_mr_resp uresp;
-
+	struct scatterlist *sg;
 	PDBG("%s ib_pd %p\n", __func__, pd);
 
 	php = to_iwch_pd(pd);
@@ -645,9 +644,7 @@ static struct ib_mr *iwch_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 
 	shift = ffs(mhp->umem->page_size) - 1;
 
-	n = 0;
-	list_for_each_entry(chunk, &mhp->umem->chunk_list, list)
-		n += chunk->nents;
+	n = mhp->umem->nmap;
 
 	err = iwch_alloc_pbl(mhp, n);
 	if (err)
@@ -661,12 +658,10 @@ static struct ib_mr *iwch_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 
 	i = n = 0;
 
-	list_for_each_entry(chunk, &mhp->umem->chunk_list, list)
-		for (j = 0; j < chunk->nmap; ++j) {
-			len = sg_dma_len(&chunk->page_list[j]) >> shift;
+	for_each_sg(mhp->umem->sg_head.sgl, sg, mhp->umem->nmap, entry) {
+			len = sg_dma_len(sg) >> shift;
 			for (k = 0; k < len; ++k) {
-				pages[i++] = cpu_to_be64(sg_dma_address(
-					&chunk->page_list[j]) +
+				pages[i++] = cpu_to_be64(sg_dma_address(sg) +
 					mhp->umem->page_size * k);
 				if (i == PAGE_SIZE / sizeof *pages) {
 					err = iwch_write_pbl(mhp, pages, i, n);
@@ -676,7 +671,7 @@ static struct ib_mr *iwch_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 					i = 0;
 				}
 			}
-		}
+	}
 
 	if (i)
 		err = iwch_write_pbl(mhp, pages, i, n);

drivers/infiniband/hw/cxgb4/cm.c

@@ -98,9 +98,9 @@ int c4iw_debug;
 module_param(c4iw_debug, int, 0644);
 MODULE_PARM_DESC(c4iw_debug, "Enable debug logging (default=0)");
 
-static int peer2peer;
+static int peer2peer = 1;
 module_param(peer2peer, int, 0644);
-MODULE_PARM_DESC(peer2peer, "Support peer2peer ULPs (default=0)");
+MODULE_PARM_DESC(peer2peer, "Support peer2peer ULPs (default=1)");
 
 static int p2p_type = FW_RI_INIT_P2PTYPE_READ_REQ;
 module_param(p2p_type, int, 0644);
@@ -400,7 +400,8 @@ static struct dst_entry *find_route(struct c4iw_dev *dev, __be32 local_ip,
 	n = dst_neigh_lookup(&rt->dst, &peer_ip);
 	if (!n)
 		return NULL;
-	if (!our_interface(dev, n->dev)) {
+	if (!our_interface(dev, n->dev) &&
+	    !(n->dev->flags & IFF_LOOPBACK)) {
 		dst_release(&rt->dst);
 		return NULL;
 	}
@@ -759,8 +760,9 @@ static void send_mpa_req(struct c4iw_ep *ep, struct sk_buff *skb,
 	ep->mpa_skb = skb;
 	c4iw_l2t_send(&ep->com.dev->rdev, skb, ep->l2t);
 	start_ep_timer(ep);
-	state_set(&ep->com, MPA_REQ_SENT);
+	__state_set(&ep->com, MPA_REQ_SENT);
 	ep->mpa_attr.initiator = 1;
+	ep->snd_seq += mpalen;
 	return;
 }
 
@@ -840,6 +842,7 @@ static int send_mpa_reject(struct c4iw_ep *ep, const void *pdata, u8 plen)
 	t4_set_arp_err_handler(skb, NULL, arp_failure_discard);
 	BUG_ON(ep->mpa_skb);
 	ep->mpa_skb = skb;
+	ep->snd_seq += mpalen;
 	return c4iw_l2t_send(&ep->com.dev->rdev, skb, ep->l2t);
 }
 
@@ -923,7 +926,8 @@ static int send_mpa_reply(struct c4iw_ep *ep, const void *pdata, u8 plen)
 	skb_get(skb);
 	t4_set_arp_err_handler(skb, NULL, arp_failure_discard);
 	ep->mpa_skb = skb;
-	state_set(&ep->com, MPA_REP_SENT);
+	__state_set(&ep->com, MPA_REP_SENT);
+	ep->snd_seq += mpalen;
 	return c4iw_l2t_send(&ep->com.dev->rdev, skb, ep->l2t);
 }
 
@@ -940,6 +944,7 @@ static int act_establish(struct c4iw_dev *dev, struct sk_buff *skb)
 	PDBG("%s ep %p tid %u snd_isn %u rcv_isn %u\n", __func__, ep, tid,
 	     be32_to_cpu(req->snd_isn), be32_to_cpu(req->rcv_isn));
 
+	mutex_lock(&ep->com.mutex);
 	dst_confirm(ep->dst);
 
 	/* setup the hwtid for this connection */
@@ -963,17 +968,18 @@ static int act_establish(struct c4iw_dev *dev, struct sk_buff *skb)
 		send_mpa_req(ep, skb, 1);
 	else
 		send_mpa_req(ep, skb, mpa_rev);
-
+	mutex_unlock(&ep->com.mutex);
 	return 0;
 }
 
-static void close_complete_upcall(struct c4iw_ep *ep)
+static void close_complete_upcall(struct c4iw_ep *ep, int status)
 {
 	struct iw_cm_event event;
 
 	PDBG("%s ep %p tid %u\n", __func__, ep, ep->hwtid);
 	memset(&event, 0, sizeof(event));
 	event.event = IW_CM_EVENT_CLOSE;
+	event.status = status;
 	if (ep->com.cm_id) {
 		PDBG("close complete delivered ep %p cm_id %p tid %u\n",
 		     ep, ep->com.cm_id, ep->hwtid);
@@ -987,7 +993,6 @@ static void close_complete_upcall(struct c4iw_ep *ep)
 static int abort_connection(struct c4iw_ep *ep, struct sk_buff *skb, gfp_t gfp)
 {
 	PDBG("%s ep %p tid %u\n", __func__, ep, ep->hwtid);
-	close_complete_upcall(ep);
 	state_set(&ep->com, ABORTING);
 	set_bit(ABORT_CONN, &ep->com.history);
 	return send_abort(ep, skb, gfp);
@@ -1066,9 +1071,10 @@ static void connect_reply_upcall(struct c4iw_ep *ep, int status)
 	}
 }
 
-static void connect_request_upcall(struct c4iw_ep *ep)
+static int connect_request_upcall(struct c4iw_ep *ep)
 {
 	struct iw_cm_event event;
+	int ret;
 
 	PDBG("%s ep %p tid %u\n", __func__, ep, ep->hwtid);
 	memset(&event, 0, sizeof(event));
@@ -1093,15 +1099,14 @@ static void connect_request_upcall(struct c4iw_ep *ep)
 		event.private_data_len = ep->plen;
 		event.private_data = ep->mpa_pkt + sizeof(struct mpa_message);
 	}
-	if (state_read(&ep->parent_ep->com) != DEAD) {
-		c4iw_get_ep(&ep->com);
-		ep->parent_ep->com.cm_id->event_handler(
-						ep->parent_ep->com.cm_id,
-						&event);
-	}
+	c4iw_get_ep(&ep->com);
+	ret = ep->parent_ep->com.cm_id->event_handler(ep->parent_ep->com.cm_id,
+						      &event);
+	if (ret)
+		c4iw_put_ep(&ep->com);
 	set_bit(CONNREQ_UPCALL, &ep->com.history);
 	c4iw_put_ep(&ep->parent_ep->com);
-	ep->parent_ep = NULL;
+	return ret;
 }
 
 static void established_upcall(struct c4iw_ep *ep)
@@ -1165,7 +1170,7 @@ static void process_mpa_reply(struct c4iw_ep *ep, struct sk_buff *skb)
 	 * the connection.
 	 */
 	stop_ep_timer(ep);
-	if (state_read(&ep->com) != MPA_REQ_SENT)
+	if (ep->com.state != MPA_REQ_SENT)
 		return;
 
 	/*
@@ -1240,7 +1245,7 @@ static void process_mpa_reply(struct c4iw_ep *ep, struct sk_buff *skb)
 	 * start reply message including private data. And
 	 * the MPA header is valid.
 	 */
-	state_set(&ep->com, FPDU_MODE);
+	__state_set(&ep->com, FPDU_MODE);
 	ep->mpa_attr.crc_enabled = (mpa->flags & MPA_CRC) | crc_enabled ? 1 : 0;
 	ep->mpa_attr.recv_marker_enabled = markers_enabled;
 	ep->mpa_attr.xmit_marker_enabled = mpa->flags & MPA_MARKERS ? 1 : 0;
@@ -1355,7 +1360,7 @@ static void process_mpa_reply(struct c4iw_ep *ep, struct sk_buff *skb)
 	}
 	goto out;
 err:
-	state_set(&ep->com, ABORTING);
+	__state_set(&ep->com, ABORTING);
 	send_abort(ep, skb, GFP_KERNEL);
 out:
 	connect_reply_upcall(ep, err);
@@ -1370,7 +1375,7 @@ static void process_mpa_request(struct c4iw_ep *ep, struct sk_buff *skb)
 
 	PDBG("%s ep %p tid %u\n", __func__, ep, ep->hwtid);
 
-	if (state_read(&ep->com) != MPA_REQ_WAIT)
+	if (ep->com.state != MPA_REQ_WAIT)
 		return;
 
 	/*
@@ -1400,7 +1405,6 @@ static void process_mpa_request(struct c4iw_ep *ep, struct sk_buff *skb)
 		return;
 
 	PDBG("%s enter (%s line %u)\n", __func__, __FILE__, __LINE__);
-	stop_ep_timer(ep);
 	mpa = (struct mpa_message *) ep->mpa_pkt;
 
 	/*
@@ -1492,10 +1496,18 @@ static void process_mpa_request(struct c4iw_ep *ep, struct sk_buff *skb)
 	     ep->mpa_attr.xmit_marker_enabled, ep->mpa_attr.version,
 	     ep->mpa_attr.p2p_type);
 
-	state_set(&ep->com, MPA_REQ_RCVD);
+	__state_set(&ep->com, MPA_REQ_RCVD);
+	stop_ep_timer(ep);
 
 	/* drive upcall */
-	connect_request_upcall(ep);
+	mutex_lock(&ep->parent_ep->com.mutex);
+	if (ep->parent_ep->com.state != DEAD) {
+		if (connect_request_upcall(ep))
+			abort_connection(ep, skb, GFP_KERNEL);
+	} else {
+		abort_connection(ep, skb, GFP_KERNEL);
+	}
+	mutex_unlock(&ep->parent_ep->com.mutex);
 	return;
 }
 
@@ -1509,14 +1521,17 @@ static int rx_data(struct c4iw_dev *dev, struct sk_buff *skb)
 	__u8 status = hdr->status;
 
 	ep = lookup_tid(t, tid);
+	if (!ep)
+		return 0;
 	PDBG("%s ep %p tid %u dlen %u\n", __func__, ep, ep->hwtid, dlen);
 	skb_pull(skb, sizeof(*hdr));
 	skb_trim(skb, dlen);
+	mutex_lock(&ep->com.mutex);
 
 	/* update RX credits */
 	update_rx_credits(ep, dlen);
 
-	switch (state_read(&ep->com)) {
+	switch (ep->com.state) {
 	case MPA_REQ_SENT:
 		ep->rcv_seq += dlen;
 		process_mpa_reply(ep, skb);
@@ -1532,7 +1547,7 @@ static int rx_data(struct c4iw_dev *dev, struct sk_buff *skb)
 			pr_err("%s Unexpected streaming data." \
 			       " qpid %u ep %p state %d tid %u status %d\n",
 			       __func__, ep->com.qp->wq.sq.qid, ep,
-			       state_read(&ep->com), ep->hwtid, status);
+			       ep->com.state, ep->hwtid, status);
 		attrs.next_state = C4IW_QP_STATE_TERMINATE;
 		c4iw_modify_qp(ep->com.qp->rhp, ep->com.qp,
 			       C4IW_QP_ATTR_NEXT_STATE, &attrs, 0);
@@ -1541,6 +1556,7 @@ static int rx_data(struct c4iw_dev *dev, struct sk_buff *skb)
 	default:
 		break;
 	}
+	mutex_unlock(&ep->com.mutex);
 	return 0;
 }
 
@@ -2255,7 +2271,7 @@ static int peer_close(struct c4iw_dev *dev, struct sk_buff *skb)
 			c4iw_modify_qp(ep->com.qp->rhp, ep->com.qp,
 				       C4IW_QP_ATTR_NEXT_STATE, &attrs, 1);
 		}
-		close_complete_upcall(ep);
+		close_complete_upcall(ep, 0);
 		__state_set(&ep->com, DEAD);
 		release = 1;
 		disconnect = 0;
@@ -2425,7 +2441,7 @@ static int close_con_rpl(struct c4iw_dev *dev, struct sk_buff *skb)
 					     C4IW_QP_ATTR_NEXT_STATE,
 					     &attrs, 1);
 		}
-		close_complete_upcall(ep);
+		close_complete_upcall(ep, 0);
 		__state_set(&ep->com, DEAD);
 		release = 1;
 		break;
@@ -2500,22 +2516,28 @@ static int fw4_ack(struct c4iw_dev *dev, struct sk_buff *skb)
 
 int c4iw_reject_cr(struct iw_cm_id *cm_id, const void *pdata, u8 pdata_len)
 {
-	int err;
+	int err = 0;
+	int disconnect = 0;
 	struct c4iw_ep *ep = to_ep(cm_id);
 	PDBG("%s ep %p tid %u\n", __func__, ep, ep->hwtid);
 
-	if (state_read(&ep->com) == DEAD) {
+	mutex_lock(&ep->com.mutex);
+	if (ep->com.state == DEAD) {
+		mutex_unlock(&ep->com.mutex);
 		c4iw_put_ep(&ep->com);
 		return -ECONNRESET;
 	}
 	set_bit(ULP_REJECT, &ep->com.history);
-	BUG_ON(state_read(&ep->com) != MPA_REQ_RCVD);
+	BUG_ON(ep->com.state != MPA_REQ_RCVD);
 	if (mpa_rev == 0)
 		abort_connection(ep, NULL, GFP_KERNEL);
 	else {
 		err = send_mpa_reject(ep, pdata, pdata_len);
-		err = c4iw_ep_disconnect(ep, 0, GFP_KERNEL);
+		disconnect = 1;
 	}
+	mutex_unlock(&ep->com.mutex);
+	if (disconnect)
+		err = c4iw_ep_disconnect(ep, 0, GFP_KERNEL);
 	c4iw_put_ep(&ep->com);
 	return 0;
 }
@@ -2530,12 +2552,14 @@ int c4iw_accept_cr(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
 	struct c4iw_qp *qp = get_qhp(h, conn_param->qpn);
 
 	PDBG("%s ep %p tid %u\n", __func__, ep, ep->hwtid);
-	if (state_read(&ep->com) == DEAD) {
+
+	mutex_lock(&ep->com.mutex);
+	if (ep->com.state == DEAD) {
 		err = -ECONNRESET;
 		goto err;
 	}
 
-	BUG_ON(state_read(&ep->com) != MPA_REQ_RCVD);
+	BUG_ON(ep->com.state != MPA_REQ_RCVD);
 	BUG_ON(!qp);
 
 	set_bit(ULP_ACCEPT, &ep->com.history);
@@ -2604,14 +2628,16 @@ int c4iw_accept_cr(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
 	if (err)
 		goto err1;
 
-	state_set(&ep->com, FPDU_MODE);
+	__state_set(&ep->com, FPDU_MODE);
 	established_upcall(ep);
+	mutex_unlock(&ep->com.mutex);
 	c4iw_put_ep(&ep->com);
 	return 0;
 err1:
 	ep->com.cm_id = NULL;
 	cm_id->rem_ref(cm_id);
 err:
+	mutex_unlock(&ep->com.mutex);
 	c4iw_put_ep(&ep->com);
 	return err;
 }
@@ -2980,7 +3006,7 @@ int c4iw_ep_disconnect(struct c4iw_ep *ep, int abrupt, gfp_t gfp)
 	rdev = &ep->com.dev->rdev;
 	if (c4iw_fatal_error(rdev)) {
 		fatal = 1;
-		close_complete_upcall(ep);
+		close_complete_upcall(ep, -EIO);
 		ep->com.state = DEAD;
 	}
 	switch (ep->com.state) {
@@ -3022,7 +3048,7 @@ int c4iw_ep_disconnect(struct c4iw_ep *ep, int abrupt, gfp_t gfp)
 	if (close) {
 		if (abrupt) {
 			set_bit(EP_DISC_ABORT, &ep->com.history);
-			close_complete_upcall(ep);
+			close_complete_upcall(ep, -ECONNRESET);
 			ret = send_abort(ep, NULL, gfp);
 		} else {
 			set_bit(EP_DISC_CLOSE, &ep->com.history);
@@ -3203,6 +3229,7 @@ static void send_fw_pass_open_req(struct c4iw_dev *dev, struct sk_buff *skb,
 	struct sk_buff *req_skb;
 	struct fw_ofld_connection_wr *req;
 	struct cpl_pass_accept_req *cpl = cplhdr(skb);
+	int ret;
 
 	req_skb = alloc_skb(sizeof(struct fw_ofld_connection_wr), GFP_KERNEL);
 	req = (struct fw_ofld_connection_wr *)__skb_put(req_skb, sizeof(*req));
@@ -3239,7 +3266,13 @@ static void send_fw_pass_open_req(struct c4iw_dev *dev, struct sk_buff *skb,
 	req->cookie = (unsigned long)skb;
 
 	set_wr_txq(req_skb, CPL_PRIORITY_CONTROL, port_id);
-	cxgb4_ofld_send(dev->rdev.lldi.ports[0], req_skb);
+	ret = cxgb4_ofld_send(dev->rdev.lldi.ports[0], req_skb);
+	if (ret < 0) {
+		pr_err("%s - cxgb4_ofld_send error %d - dropping\n", __func__,
+		       ret);
+		kfree_skb(skb);
+		kfree_skb(req_skb);
+	}
 }
 
 /*
@@ -3346,13 +3379,13 @@ static int rx_pkt(struct c4iw_dev *dev, struct sk_buff *skb)
 		pi = (struct port_info *)netdev_priv(pdev);
 		tx_chan = cxgb4_port_chan(pdev);
 	}
+	neigh_release(neigh);
 	if (!e) {
 		pr_err("%s - failed to allocate l2t entry!\n",
 		       __func__);
 		goto free_dst;
 	}
 
-	neigh_release(neigh);
 	step = dev->rdev.lldi.nrxq / dev->rdev.lldi.nchan;
 	rss_qid = dev->rdev.lldi.rxq_ids[pi->port_id * step];
 	window = (__force u16) htons((__force u16)tcph->window);
@@ -3427,6 +3460,7 @@ static void process_timeout(struct c4iw_ep *ep)
 				     &attrs, 1);
 		}
 		__state_set(&ep->com, ABORTING);
+		close_complete_upcall(ep, -ETIMEDOUT);
 		break;
 	default:
 		WARN(1, "%s unexpected state ep %p tid %u state %u\n",

drivers/infiniband/hw/cxgb4/cq.c

@@ -365,8 +365,14 @@ void c4iw_flush_hw_cq(struct c4iw_cq *chp)
 
 		if (CQE_OPCODE(hw_cqe) == FW_RI_READ_RESP) {
 
-			/*
-			 * drop peer2peer RTR reads.
+			/* If we have reached here because of async
+			 * event or other error, and have egress error
+			 * then drop
+			 */
+			if (CQE_TYPE(hw_cqe) == 1)
+				goto next_cqe;
+
+			/* drop peer2peer RTR reads.
 			 */
 			if (CQE_WRID_STAG(hw_cqe) == 1)
 				goto next_cqe;
@@ -511,8 +517,18 @@ static int poll_cq(struct t4_wq *wq, struct t4_cq *cq, struct t4_cqe *cqe,
 	 */
 	if (RQ_TYPE(hw_cqe) && (CQE_OPCODE(hw_cqe) == FW_RI_READ_RESP)) {
 
-		/*
-		 * If this is an unsolicited read response, then the read
+		/* If we have reached here because of async
+		 * event or other error, and have egress error
+		 * then drop
+		 */
+		if (CQE_TYPE(hw_cqe) == 1) {
+			if (CQE_STATUS(hw_cqe))
+				t4_set_wq_in_error(wq);
+			ret = -EAGAIN;
+			goto skip_cqe;
+		}
+
+		/* If this is an unsolicited read response, then the read
 		 * was generated by the kernel driver as part of peer-2-peer
 		 * connection setup.  So ignore the completion.
 		 */
@@ -603,7 +619,7 @@ proc_cqe:
 	 */
 	if (SQ_TYPE(hw_cqe)) {
 		int idx = CQE_WRID_SQ_IDX(hw_cqe);
-		BUG_ON(idx > wq->sq.size);
+		BUG_ON(idx >= wq->sq.size);
 
 		/*
 		* Account for any unsignaled completions completed by
@@ -617,7 +633,7 @@ proc_cqe:
 			wq->sq.in_use -= wq->sq.size + idx - wq->sq.cidx;
 		else
 			wq->sq.in_use -= idx - wq->sq.cidx;
-		BUG_ON(wq->sq.in_use < 0 && wq->sq.in_use < wq->sq.size);
+		BUG_ON(wq->sq.in_use <= 0 && wq->sq.in_use >= wq->sq.size);
 
 		wq->sq.cidx = (uint16_t)idx;
 		PDBG("%s completing sq idx %u\n", __func__, wq->sq.cidx);
@@ -881,7 +897,7 @@ struct ib_cq *c4iw_create_cq(struct ib_device *ibdev, int entries,
 	/*
 	 * Make actual HW queue 2x to avoid cdix_inc overflows.
 	 */
-	hwentries = entries * 2;
+	hwentries = min(entries * 2, T4_MAX_IQ_SIZE);
 
 	/*
 	 * Make HW queue at least 64 entries so GTS updates aren't too
@@ -930,6 +946,7 @@ struct ib_cq *c4iw_create_cq(struct ib_device *ibdev, int entries,
 		if (!mm2)
 			goto err4;
 
+		memset(&uresp, 0, sizeof(uresp));
 		uresp.qid_mask = rhp->rdev.cqmask;
 		uresp.cqid = chp->cq.cqid;
 		uresp.size = chp->cq.size;

drivers/infiniband/hw/cxgb4/device.c

@@ -898,11 +898,13 @@ static int c4iw_uld_rx_handler(void *handle, const __be64 *rsp,
 	}
 
 	opcode = *(u8 *)rsp;
-	if (c4iw_handlers[opcode])
+	if (c4iw_handlers[opcode]) {
 		c4iw_handlers[opcode](dev, skb);
-	else
+	} else {
 		pr_info("%s no handler opcode 0x%x...\n", __func__,
 		       opcode);
+		kfree_skb(skb);
+	}
 
 	return 0;
 nomem:

drivers/infiniband/hw/cxgb4/iw_cxgb4.h

@@ -373,6 +373,7 @@ struct c4iw_fr_page_list {
 	DEFINE_DMA_UNMAP_ADDR(mapping);
 	dma_addr_t dma_addr;
 	struct c4iw_dev *dev;
+	int pll_len;
 };
 
 static inline struct c4iw_fr_page_list *to_c4iw_fr_page_list(
@@ -446,6 +447,7 @@ struct c4iw_qp {
 	atomic_t refcnt;
 	wait_queue_head_t wait;
 	struct timer_list timer;
+	int sq_sig_all;
 };
 
 static inline struct c4iw_qp *to_c4iw_qp(struct ib_qp *ibqp)

drivers/infiniband/hw/cxgb4/mem.c

@@ -37,9 +37,9 @@
 
 #include "iw_cxgb4.h"
 
-int use_dsgl = 1;
+int use_dsgl = 0;
 module_param(use_dsgl, int, 0644);
-MODULE_PARM_DESC(use_dsgl, "Use DSGL for PBL/FastReg (default=1)");
+MODULE_PARM_DESC(use_dsgl, "Use DSGL for PBL/FastReg (default=0)");
 
 #define T4_ULPTX_MIN_IO 32
 #define C4IW_MAX_INLINE_SIZE 96
@@ -678,9 +678,9 @@ struct ib_mr *c4iw_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 {
 	__be64 *pages;
 	int shift, n, len;
-	int i, j, k;
+	int i, k, entry;
 	int err = 0;
-	struct ib_umem_chunk *chunk;
+	struct scatterlist *sg;
 	struct c4iw_dev *rhp;
 	struct c4iw_pd *php;
 	struct c4iw_mr *mhp;
@@ -710,10 +710,7 @@ struct ib_mr *c4iw_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 
 	shift = ffs(mhp->umem->page_size) - 1;
 
-	n = 0;
-	list_for_each_entry(chunk, &mhp->umem->chunk_list, list)
-		n += chunk->nents;
-
+	n = mhp->umem->nmap;
 	err = alloc_pbl(mhp, n);
 	if (err)
 		goto err;
@@ -726,24 +723,22 @@ struct ib_mr *c4iw_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 
 	i = n = 0;
 
-	list_for_each_entry(chunk, &mhp->umem->chunk_list, list)
-		for (j = 0; j < chunk->nmap; ++j) {
-			len = sg_dma_len(&chunk->page_list[j]) >> shift;
-			for (k = 0; k < len; ++k) {
-				pages[i++] = cpu_to_be64(sg_dma_address(
-					&chunk->page_list[j]) +
-					mhp->umem->page_size * k);
-				if (i == PAGE_SIZE / sizeof *pages) {
-					err = write_pbl(&mhp->rhp->rdev,
-					      pages,
-					      mhp->attr.pbl_addr + (n << 3), i);
-					if (err)
-						goto pbl_done;
-					n += i;
-					i = 0;
-				}
+	for_each_sg(mhp->umem->sg_head.sgl, sg, mhp->umem->nmap, entry) {
+		len = sg_dma_len(sg) >> shift;
+		for (k = 0; k < len; ++k) {
+			pages[i++] = cpu_to_be64(sg_dma_address(sg) +
+				mhp->umem->page_size * k);
+			if (i == PAGE_SIZE / sizeof *pages) {
+				err = write_pbl(&mhp->rhp->rdev,
+				      pages,
+				      mhp->attr.pbl_addr + (n << 3), i);
+				if (err)
+					goto pbl_done;
+				n += i;
+				i = 0;
 			}
 		}
+	}
 
 	if (i)
 		err = write_pbl(&mhp->rhp->rdev, pages,
@@ -903,7 +898,11 @@ struct ib_fast_reg_page_list *c4iw_alloc_fastreg_pbl(struct ib_device *device,
 	dma_unmap_addr_set(c4pl, mapping, dma_addr);
 	c4pl->dma_addr = dma_addr;
 	c4pl->dev = dev;
-	c4pl->ibpl.max_page_list_len = pll_len;
+	c4pl->pll_len = pll_len;
+
+	PDBG("%s c4pl %p pll_len %u page_list %p dma_addr %pad\n",
+	     __func__, c4pl, c4pl->pll_len, c4pl->ibpl.page_list,
+	     &c4pl->dma_addr);
 
 	return &c4pl->ibpl;
 }
@@ -912,8 +911,12 @@ void c4iw_free_fastreg_pbl(struct ib_fast_reg_page_list *ibpl)
 {
 	struct c4iw_fr_page_list *c4pl = to_c4iw_fr_page_list(ibpl);
 
+	PDBG("%s c4pl %p pll_len %u page_list %p dma_addr %pad\n",
+	     __func__, c4pl, c4pl->pll_len, c4pl->ibpl.page_list,
+	     &c4pl->dma_addr);
+
 	dma_free_coherent(&c4pl->dev->rdev.lldi.pdev->dev,
-			  c4pl->ibpl.max_page_list_len,
+			  c4pl->pll_len,
 			  c4pl->ibpl.page_list, dma_unmap_addr(c4pl, mapping));
 	kfree(c4pl);
 }

drivers/infiniband/hw/cxgb4/qp.c

@@ -715,7 +715,7 @@ int c4iw_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 		fw_flags = 0;
 		if (wr->send_flags & IB_SEND_SOLICITED)
 			fw_flags |= FW_RI_SOLICITED_EVENT_FLAG;
-		if (wr->send_flags & IB_SEND_SIGNALED)
+		if (wr->send_flags & IB_SEND_SIGNALED || qhp->sq_sig_all)
 			fw_flags |= FW_RI_COMPLETION_FLAG;
 		swsqe = &qhp->wq.sq.sw_sq[qhp->wq.sq.pidx];
 		switch (wr->opcode) {
@@ -776,7 +776,8 @@ int c4iw_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 		}
 		swsqe->idx = qhp->wq.sq.pidx;
 		swsqe->complete = 0;
-		swsqe->signaled = (wr->send_flags & IB_SEND_SIGNALED);
+		swsqe->signaled = (wr->send_flags & IB_SEND_SIGNALED) ||
+				  qhp->sq_sig_all;
 		swsqe->flushed = 0;
 		swsqe->wr_id = wr->wr_id;
 
@@ -1529,7 +1530,7 @@ struct ib_qp *c4iw_create_qp(struct ib_pd *pd, struct ib_qp_init_attr *attrs,
 	struct c4iw_cq *schp;
 	struct c4iw_cq *rchp;
 	struct c4iw_create_qp_resp uresp;
-	int sqsize, rqsize;
+	unsigned int sqsize, rqsize;
 	struct c4iw_ucontext *ucontext;
 	int ret;
 	struct c4iw_mm_entry *mm1, *mm2, *mm3, *mm4, *mm5 = NULL;
@@ -1601,6 +1602,7 @@ struct ib_qp *c4iw_create_qp(struct ib_pd *pd, struct ib_qp_init_attr *attrs,
 	qhp->attr.enable_bind = 1;
 	qhp->attr.max_ord = 1;
 	qhp->attr.max_ird = 1;
+	qhp->sq_sig_all = attrs->sq_sig_type == IB_SIGNAL_ALL_WR;
 	spin_lock_init(&qhp->lock);
 	mutex_init(&qhp->mutex);
 	init_waitqueue_head(&qhp->wait);

drivers/infiniband/hw/ehca/ehca_classes.h

@@ -322,7 +322,7 @@ struct ehca_mr_pginfo {
 		} phy;
 		struct { /* type EHCA_MR_PGI_USER section */
 			struct ib_umem *region;
-			struct ib_umem_chunk *next_chunk;
+			struct scatterlist *next_sg;
 			u64 next_nmap;
 		} usr;
 		struct { /* type EHCA_MR_PGI_FMR section */

drivers/infiniband/hw/ehca/ehca_cq.c

@@ -283,6 +283,7 @@ struct ib_cq *ehca_create_cq(struct ib_device *device, int cqe, int comp_vector,
 			(my_cq->galpas.user.fw_handle & (PAGE_SIZE - 1));
 		if (ib_copy_to_udata(udata, &resp, sizeof(resp))) {
 			ehca_err(device, "Copy to udata failed.");
+			cq = ERR_PTR(-EFAULT);
 			goto create_cq_exit4;
 		}
 	}

drivers/infiniband/hw/ehca/ehca_mrmw.c

@@ -400,10 +400,7 @@ reg_user_mr_fallback:
 	pginfo.num_hwpages = num_hwpages;
 	pginfo.u.usr.region = e_mr->umem;
 	pginfo.next_hwpage = e_mr->umem->offset / hwpage_size;
-	pginfo.u.usr.next_chunk = list_prepare_entry(pginfo.u.usr.next_chunk,
-						     (&e_mr->umem->chunk_list),
-						     list);
-
+	pginfo.u.usr.next_sg = pginfo.u.usr.region->sg_head.sgl;
 	ret = ehca_reg_mr(shca, e_mr, (u64 *)virt, length, mr_access_flags,
 			  e_pd, &pginfo, &e_mr->ib.ib_mr.lkey,
 			  &e_mr->ib.ib_mr.rkey, EHCA_REG_MR);
@@ -1858,61 +1855,39 @@ static int ehca_set_pagebuf_user1(struct ehca_mr_pginfo *pginfo,
 				  u64 *kpage)
 {
 	int ret = 0;
-	struct ib_umem_chunk *prev_chunk;
-	struct ib_umem_chunk *chunk;
 	u64 pgaddr;
-	u32 i = 0;
 	u32 j = 0;
 	int hwpages_per_kpage = PAGE_SIZE / pginfo->hwpage_size;
-
-	/* loop over desired chunk entries */
-	chunk      = pginfo->u.usr.next_chunk;
-	prev_chunk = pginfo->u.usr.next_chunk;
-	list_for_each_entry_continue(
-		chunk, (&(pginfo->u.usr.region->chunk_list)), list) {
-		for (i = pginfo->u.usr.next_nmap; i < chunk->nmap; ) {
-			pgaddr = page_to_pfn(sg_page(&chunk->page_list[i]))
-				<< PAGE_SHIFT ;
-			*kpage = pgaddr + (pginfo->next_hwpage *
-					   pginfo->hwpage_size);
-			if ( !(*kpage) ) {
-				ehca_gen_err("pgaddr=%llx "
-					     "chunk->page_list[i]=%llx "
-					     "i=%x next_hwpage=%llx",
-					     pgaddr, (u64)sg_dma_address(
-						     &chunk->page_list[i]),
-					     i, pginfo->next_hwpage);
-				return -EFAULT;
-			}
-			(pginfo->hwpage_cnt)++;
-			(pginfo->next_hwpage)++;
-			kpage++;
-			if (pginfo->next_hwpage % hwpages_per_kpage == 0) {
-				(pginfo->kpage_cnt)++;
-				(pginfo->u.usr.next_nmap)++;
-				pginfo->next_hwpage = 0;
-				i++;
-			}
-			j++;
-			if (j >= number) break;
+	struct scatterlist **sg = &pginfo->u.usr.next_sg;
+
+	while (*sg != NULL) {
+		pgaddr = page_to_pfn(sg_page(*sg))
+			<< PAGE_SHIFT;
+		*kpage = pgaddr + (pginfo->next_hwpage *
+				   pginfo->hwpage_size);
+		if (!(*kpage)) {
+			ehca_gen_err("pgaddr=%llx "
+				     "sg_dma_address=%llx "
+				     "entry=%llx next_hwpage=%llx",
+				     pgaddr, (u64)sg_dma_address(*sg),
+				     pginfo->u.usr.next_nmap,
+				     pginfo->next_hwpage);
+			return -EFAULT;
 		}
-		if ((pginfo->u.usr.next_nmap >= chunk->nmap) &&
-		    (j >= number)) {
-			pginfo->u.usr.next_nmap = 0;
-			prev_chunk = chunk;
-			break;
-		} else if (pginfo->u.usr.next_nmap >= chunk->nmap) {
-			pginfo->u.usr.next_nmap = 0;
-			prev_chunk = chunk;
-		} else if (j >= number)
+		(pginfo->hwpage_cnt)++;
+		(pginfo->next_hwpage)++;
+		kpage++;
+		if (pginfo->next_hwpage % hwpages_per_kpage == 0) {
+			(pginfo->kpage_cnt)++;
+			(pginfo->u.usr.next_nmap)++;
+			pginfo->next_hwpage = 0;
+			*sg = sg_next(*sg);
+		}
+		j++;
+		if (j >= number)
 			break;
-		else
-			prev_chunk = chunk;
 	}
-	pginfo->u.usr.next_chunk =
-		list_prepare_entry(prev_chunk,
-				   (&(pginfo->u.usr.region->chunk_list)),
-				   list);
+
 	return ret;
 }
 
@@ -1920,20 +1895,19 @@ static int ehca_set_pagebuf_user1(struct ehca_mr_pginfo *pginfo,
  * check given pages for contiguous layout
  * last page addr is returned in prev_pgaddr for further check
  */
-static int ehca_check_kpages_per_ate(struct scatterlist *page_list,
-				     int start_idx, int end_idx,
+static int ehca_check_kpages_per_ate(struct scatterlist **sg,
+				     int num_pages,
 				     u64 *prev_pgaddr)
 {
-	int t;
-	for (t = start_idx; t <= end_idx; t++) {
-		u64 pgaddr = page_to_pfn(sg_page(&page_list[t])) << PAGE_SHIFT;
+	for (; *sg && num_pages > 0; *sg = sg_next(*sg), num_pages--) {
+		u64 pgaddr = page_to_pfn(sg_page(*sg)) << PAGE_SHIFT;
 		if (ehca_debug_level >= 3)
 			ehca_gen_dbg("chunk_page=%llx value=%016llx", pgaddr,
 				     *(u64 *)__va(pgaddr));
 		if (pgaddr - PAGE_SIZE != *prev_pgaddr) {
 			ehca_gen_err("uncontiguous page found pgaddr=%llx "
-				     "prev_pgaddr=%llx page_list_i=%x",
-				     pgaddr, *prev_pgaddr, t);
+				     "prev_pgaddr=%llx entries_left_in_hwpage=%x",
+				     pgaddr, *prev_pgaddr, num_pages);
 			return -EINVAL;
 		}
 		*prev_pgaddr = pgaddr;
@@ -1947,111 +1921,80 @@ static int ehca_set_pagebuf_user2(struct ehca_mr_pginfo *pginfo,
 				  u64 *kpage)
 {
 	int ret = 0;
-	struct ib_umem_chunk *prev_chunk;
-	struct ib_umem_chunk *chunk;
 	u64 pgaddr, prev_pgaddr;
-	u32 i = 0;
 	u32 j = 0;
 	int kpages_per_hwpage = pginfo->hwpage_size / PAGE_SIZE;
 	int nr_kpages = kpages_per_hwpage;
+	struct scatterlist **sg = &pginfo->u.usr.next_sg;
+
+	while (*sg != NULL) {
 
-	/* loop over desired chunk entries */
-	chunk      = pginfo->u.usr.next_chunk;
-	prev_chunk = pginfo->u.usr.next_chunk;
-	list_for_each_entry_continue(
-		chunk, (&(pginfo->u.usr.region->chunk_list)), list) {
-		for (i = pginfo->u.usr.next_nmap; i < chunk->nmap; ) {
-			if (nr_kpages == kpages_per_hwpage) {
-				pgaddr = ( page_to_pfn(sg_page(&chunk->page_list[i]))
-					   << PAGE_SHIFT );
-				*kpage = pgaddr;
-				if ( !(*kpage) ) {
-					ehca_gen_err("pgaddr=%llx i=%x",
-						     pgaddr, i);
+		if (nr_kpages == kpages_per_hwpage) {
+			pgaddr = (page_to_pfn(sg_page(*sg))
+				   << PAGE_SHIFT);
+			*kpage = pgaddr;
+			if (!(*kpage)) {
+				ehca_gen_err("pgaddr=%llx entry=%llx",
+					     pgaddr, pginfo->u.usr.next_nmap);
+				ret = -EFAULT;
+				return ret;
+			}
+			/*
+			 * The first page in a hwpage must be aligned;
+			 * the first MR page is exempt from this rule.
+			 */
+			if (pgaddr & (pginfo->hwpage_size - 1)) {
+				if (pginfo->hwpage_cnt) {
+					ehca_gen_err(
+						"invalid alignment "
+						"pgaddr=%llx entry=%llx "
+						"mr_pgsize=%llx",
+						pgaddr, pginfo->u.usr.next_nmap,
+						pginfo->hwpage_size);
 					ret = -EFAULT;
 					return ret;
 				}
-				/*
-				 * The first page in a hwpage must be aligned;
-				 * the first MR page is exempt from this rule.
-				 */
-				if (pgaddr & (pginfo->hwpage_size - 1)) {
-					if (pginfo->hwpage_cnt) {
-						ehca_gen_err(
-							"invalid alignment "
-							"pgaddr=%llx i=%x "
-							"mr_pgsize=%llx",
-							pgaddr, i,
-							pginfo->hwpage_size);
-						ret = -EFAULT;
-						return ret;
-					}
-					/* first MR page */
-					pginfo->kpage_cnt =
-						(pgaddr &
-						 (pginfo->hwpage_size - 1)) >>
-						PAGE_SHIFT;
-					nr_kpages -= pginfo->kpage_cnt;
-					*kpage = pgaddr &
-						 ~(pginfo->hwpage_size - 1);
-				}
-				if (ehca_debug_level >= 3) {
-					u64 val = *(u64 *)__va(pgaddr);
-					ehca_gen_dbg("kpage=%llx chunk_page=%llx "
-						     "value=%016llx",
-						     *kpage, pgaddr, val);
-				}
-				prev_pgaddr = pgaddr;
-				i++;
-				pginfo->kpage_cnt++;
-				pginfo->u.usr.next_nmap++;
-				nr_kpages--;
-				if (!nr_kpages)
-					goto next_kpage;
-				continue;
+				/* first MR page */
+				pginfo->kpage_cnt =
+					(pgaddr &
+					 (pginfo->hwpage_size - 1)) >>
+					PAGE_SHIFT;
+				nr_kpages -= pginfo->kpage_cnt;
+				*kpage = pgaddr &
+					 ~(pginfo->hwpage_size - 1);
 			}
-			if (i + nr_kpages > chunk->nmap) {
-				ret = ehca_check_kpages_per_ate(
-					chunk->page_list, i,
-					chunk->nmap - 1, &prev_pgaddr);
-				if (ret) return ret;
-				pginfo->kpage_cnt += chunk->nmap - i;
-				pginfo->u.usr.next_nmap += chunk->nmap - i;
-				nr_kpages -= chunk->nmap - i;
-				break;
+			if (ehca_debug_level >= 3) {
+				u64 val = *(u64 *)__va(pgaddr);
+				ehca_gen_dbg("kpage=%llx page=%llx "
+					     "value=%016llx",
+					     *kpage, pgaddr, val);
 			}
+			prev_pgaddr = pgaddr;
+			*sg = sg_next(*sg);
+			pginfo->kpage_cnt++;
+			pginfo->u.usr.next_nmap++;
+			nr_kpages--;
+			if (!nr_kpages)
+				goto next_kpage;
+			continue;
+		}
+
+		ret = ehca_check_kpages_per_ate(sg, nr_kpages,
+						&prev_pgaddr);
+		if (ret)
+			return ret;
+		pginfo->kpage_cnt += nr_kpages;
+		pginfo->u.usr.next_nmap += nr_kpages;
 
-			ret = ehca_check_kpages_per_ate(chunk->page_list, i,
-							i + nr_kpages - 1,
-							&prev_pgaddr);
-			if (ret) return ret;
-			i += nr_kpages;
-			pginfo->kpage_cnt += nr_kpages;
-			pginfo->u.usr.next_nmap += nr_kpages;
 next_kpage:
-			nr_kpages = kpages_per_hwpage;
-			(pginfo->hwpage_cnt)++;
-			kpage++;
-			j++;
-			if (j >= number) break;
-		}
-		if ((pginfo->u.usr.next_nmap >= chunk->nmap) &&
-		    (j >= number)) {
-			pginfo->u.usr.next_nmap = 0;
-			prev_chunk = chunk;
-			break;
-		} else if (pginfo->u.usr.next_nmap >= chunk->nmap) {
-			pginfo->u.usr.next_nmap = 0;
-			prev_chunk = chunk;
-		} else if (j >= number)
+		nr_kpages = kpages_per_hwpage;
+		(pginfo->hwpage_cnt)++;
+		kpage++;
+		j++;
+		if (j >= number)
 			break;
-		else
-			prev_chunk = chunk;
 	}
-	pginfo->u.usr.next_chunk =
-		list_prepare_entry(prev_chunk,
-				   (&(pginfo->u.usr.region->chunk_list)),
-				   list);
+
 	return ret;
 }
 
@@ -2591,16 +2534,6 @@ static void ehca_dma_unmap_sg(struct ib_device *dev, struct scatterlist *sg,
 	/* This is only a stub; nothing to be done here */
 }
 
-static u64 ehca_dma_address(struct ib_device *dev, struct scatterlist *sg)
-{
-	return sg->dma_address;
-}
-
-static unsigned int ehca_dma_len(struct ib_device *dev, struct scatterlist *sg)
-{
-	return sg->length;
-}
-
 static void ehca_dma_sync_single_for_cpu(struct ib_device *dev, u64 addr,
 					 size_t size,
 					 enum dma_data_direction dir)
@@ -2653,8 +2586,6 @@ struct ib_dma_mapping_ops ehca_dma_mapping_ops = {
 	.unmap_page             = ehca_dma_unmap_page,
 	.map_sg                 = ehca_dma_map_sg,
 	.unmap_sg               = ehca_dma_unmap_sg,
-	.dma_address            = ehca_dma_address,
-	.dma_len                = ehca_dma_len,
 	.sync_single_for_cpu    = ehca_dma_sync_single_for_cpu,
 	.sync_single_for_device = ehca_dma_sync_single_for_device,
 	.alloc_coherent         = ehca_dma_alloc_coherent,

drivers/infiniband/hw/ipath/ipath_diag.c

@@ -326,7 +326,7 @@ static ssize_t ipath_diagpkt_write(struct file *fp,
 				   size_t count, loff_t *off)
 {
 	u32 __iomem *piobuf;
-	u32 plen, clen, pbufn;
+	u32 plen, pbufn, maxlen_reserve;
 	struct ipath_diag_pkt odp;
 	struct ipath_diag_xpkt dp;
 	u32 *tmpbuf = NULL;
@@ -335,51 +335,29 @@ static ssize_t ipath_diagpkt_write(struct file *fp,
 	u64 val;
 	u32 l_state, lt_state; /* LinkState, LinkTrainingState */
 
-	if (count < sizeof(odp)) {
-		ret = -EINVAL;
-		goto bail;
-	}
 
 	if (count == sizeof(dp)) {
 		if (copy_from_user(&dp, data, sizeof(dp))) {
 			ret = -EFAULT;
 			goto bail;
 		}
-	} else if (copy_from_user(&odp, data, sizeof(odp))) {
-		ret = -EFAULT;
+	} else if (count == sizeof(odp)) {
+		if (copy_from_user(&odp, data, sizeof(odp))) {
+			ret = -EFAULT;
+			goto bail;
+		}
+	} else {
+		ret = -EINVAL;
 		goto bail;
 	}
 
-	/*
-	 * Due to padding/alignment issues (lessened with new struct)
-	 * the old and new structs are the same length. We need to
-	 * disambiguate them, which we can do because odp.len has never
-	 * been less than the total of LRH+BTH+DETH so far, while
-	 * dp.unit (same offset) unit is unlikely to get that high.
-	 * Similarly, dp.data, the pointer to user at the same offset
-	 * as odp.unit, is almost certainly at least one (512byte)page
-	 * "above" NULL. The if-block below can be omitted if compatibility
-	 * between a new driver and older diagnostic code is unimportant.
-	 * compatibility the other direction (new diags, old driver) is
-	 * handled in the diagnostic code, with a warning.
-	 */
-	if (dp.unit >= 20 && dp.data < 512) {
-		/* very probable version mismatch. Fix it up */
-		memcpy(&odp, &dp, sizeof(odp));
-		/* We got a legacy dp, copy elements to dp */
-		dp.unit = odp.unit;
-		dp.data = odp.data;
-		dp.len = odp.len;
-		dp.pbc_wd = 0; /* Indicate we need to compute PBC wd */
-	}
-
 	/* send count must be an exact number of dwords */
 	if (dp.len & 3) {
 		ret = -EINVAL;
 		goto bail;
 	}
 
-	clen = dp.len >> 2;
+	plen = dp.len >> 2;
 
 	dd = ipath_lookup(dp.unit);
 	if (!dd || !(dd->ipath_flags & IPATH_PRESENT) ||
@@ -422,16 +400,22 @@ static ssize_t ipath_diagpkt_write(struct file *fp,
 		goto bail;
 	}
 
-	/* need total length before first word written */
-	/* +1 word is for the qword padding */
-	plen = sizeof(u32) + dp.len;
-
-	if ((plen + 4) > dd->ipath_ibmaxlen) {
+	/*
+	 * need total length before first word written, plus 2 Dwords. One Dword
+	 * is for padding so we get the full user data when not aligned on
+	 * a word boundary. The other Dword is to make sure we have room for the
+	 * ICRC which gets tacked on later.
+	 */
+	maxlen_reserve = 2 * sizeof(u32);
+	if (dp.len > dd->ipath_ibmaxlen - maxlen_reserve) {
 		ipath_dbg("Pkt len 0x%x > ibmaxlen %x\n",
-			  plen - 4, dd->ipath_ibmaxlen);
+			  dp.len, dd->ipath_ibmaxlen);
 		ret = -EINVAL;
-		goto bail;	/* before writing pbc */
+		goto bail;
 	}
+
+	plen = sizeof(u32) + dp.len;
+
 	tmpbuf = vmalloc(plen);
 	if (!tmpbuf) {
 		dev_info(&dd->pcidev->dev, "Unable to allocate tmp buffer, "
@@ -473,11 +457,11 @@ static ssize_t ipath_diagpkt_write(struct file *fp,
 	 */
 	if (dd->ipath_flags & IPATH_PIO_FLUSH_WC) {
 		ipath_flush_wc();
-		__iowrite32_copy(piobuf + 2, tmpbuf, clen - 1);
+		__iowrite32_copy(piobuf + 2, tmpbuf, plen - 1);
 		ipath_flush_wc();
-		__raw_writel(tmpbuf[clen - 1], piobuf + clen + 1);
+		__raw_writel(tmpbuf[plen - 1], piobuf + plen + 1);
 	} else
-		__iowrite32_copy(piobuf + 2, tmpbuf, clen);
+		__iowrite32_copy(piobuf + 2, tmpbuf, plen);
 
 	ipath_flush_wc();
 

drivers/infiniband/hw/ipath/ipath_dma.c

@@ -115,6 +115,10 @@ static int ipath_map_sg(struct ib_device *dev, struct scatterlist *sgl,
 			ret = 0;
 			break;
 		}
+		sg->dma_address = addr + sg->offset;
+#ifdef CONFIG_NEED_SG_DMA_LENGTH
+		sg->dma_length = sg->length;
+#endif
 	}
 	return ret;
 }
@@ -126,21 +130,6 @@ static void ipath_unmap_sg(struct ib_device *dev,
 	BUG_ON(!valid_dma_direction(direction));
 }
 
-static u64 ipath_sg_dma_address(struct ib_device *dev, struct scatterlist *sg)
-{
-	u64 addr = (u64) page_address(sg_page(sg));
-
-	if (addr)
-		addr += sg->offset;
-	return addr;
-}
-
-static unsigned int ipath_sg_dma_len(struct ib_device *dev,
-				     struct scatterlist *sg)
-{
-	return sg->length;
-}
-
 static void ipath_sync_single_for_cpu(struct ib_device *dev,
 				      u64 addr,
 				      size_t size,
@@ -176,17 +165,15 @@ static void ipath_dma_free_coherent(struct ib_device *dev, size_t size,
 }
 
 struct ib_dma_mapping_ops ipath_dma_mapping_ops = {
-	ipath_mapping_error,
-	ipath_dma_map_single,
-	ipath_dma_unmap_single,
-	ipath_dma_map_page,
-	ipath_dma_unmap_page,
-	ipath_map_sg,
-	ipath_unmap_sg,
-	ipath_sg_dma_address,
-	ipath_sg_dma_len,
-	ipath_sync_single_for_cpu,
-	ipath_sync_single_for_device,
-	ipath_dma_alloc_coherent,
-	ipath_dma_free_coherent
+	.mapping_error = ipath_mapping_error,
+	.map_single = ipath_dma_map_single,
+	.unmap_single = ipath_dma_unmap_single,
+	.map_page = ipath_dma_map_page,
+	.unmap_page = ipath_dma_unmap_page,
+	.map_sg = ipath_map_sg,
+	.unmap_sg = ipath_unmap_sg,
+	.sync_single_for_cpu = ipath_sync_single_for_cpu,
+	.sync_single_for_device = ipath_sync_single_for_device,
+	.alloc_coherent = ipath_dma_alloc_coherent,
+	.free_coherent = ipath_dma_free_coherent
 };

drivers/infiniband/hw/ipath/ipath_mr.c

@@ -188,8 +188,8 @@ struct ib_mr *ipath_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 {
 	struct ipath_mr *mr;
 	struct ib_umem *umem;
-	struct ib_umem_chunk *chunk;
-	int n, m, i;
+	int n, m, entry;
+	struct scatterlist *sg;
 	struct ib_mr *ret;
 
 	if (length == 0) {
@@ -202,10 +202,7 @@ struct ib_mr *ipath_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 	if (IS_ERR(umem))
 		return (void *) umem;
 
-	n = 0;
-	list_for_each_entry(chunk, &umem->chunk_list, list)
-		n += chunk->nents;
-
+	n = umem->nmap;
 	mr = alloc_mr(n, &to_idev(pd->device)->lk_table);
 	if (!mr) {
 		ret = ERR_PTR(-ENOMEM);
@@ -224,22 +221,20 @@ struct ib_mr *ipath_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 
 	m = 0;
 	n = 0;
-	list_for_each_entry(chunk, &umem->chunk_list, list) {
-		for (i = 0; i < chunk->nents; i++) {
-			void *vaddr;
-
-			vaddr = page_address(sg_page(&chunk->page_list[i]));
-			if (!vaddr) {
-				ret = ERR_PTR(-EINVAL);
-				goto bail;
-			}
-			mr->mr.map[m]->segs[n].vaddr = vaddr;
-			mr->mr.map[m]->segs[n].length = umem->page_size;
-			n++;
-			if (n == IPATH_SEGSZ) {
-				m++;
-				n = 0;
-			}
+	for_each_sg(umem->sg_head.sgl, sg, umem->nmap, entry) {
+		void *vaddr;
+
+		vaddr = page_address(sg_page(sg));
+		if (!vaddr) {
+			ret = ERR_PTR(-EINVAL);
+			goto bail;
+		}
+		mr->mr.map[m]->segs[n].vaddr = vaddr;
+		mr->mr.map[m]->segs[n].length = umem->page_size;
+		n++;
+		if (n == IPATH_SEGSZ) {
+			m++;
+			n = 0;
 		}
 	}
 	ret = &mr->ibmr;

drivers/infiniband/hw/mlx4/doorbell.c

@@ -45,7 +45,6 @@ int mlx4_ib_db_map_user(struct mlx4_ib_ucontext *context, unsigned long virt,
 			struct mlx4_db *db)
 {
 	struct mlx4_ib_user_db_page *page;
-	struct ib_umem_chunk *chunk;
 	int err = 0;
 
 	mutex_lock(&context->db_page_mutex);
@@ -73,8 +72,7 @@ int mlx4_ib_db_map_user(struct mlx4_ib_ucontext *context, unsigned long virt,
 	list_add(&page->list, &context->db_page_list);
 
 found:
-	chunk = list_entry(page->umem->chunk_list.next, struct ib_umem_chunk, list);
-	db->dma		= sg_dma_address(chunk->page_list) + (virt & ~PAGE_MASK);
+	db->dma = sg_dma_address(page->umem->sg_head.sgl) + (virt & ~PAGE_MASK);
 	db->u.user_page = page;
 	++page->refcnt;
 

drivers/infiniband/hw/mlx4/main.c

@@ -1803,7 +1803,7 @@ static void init_pkeys(struct mlx4_ib_dev *ibdev)
 
 static void mlx4_ib_alloc_eqs(struct mlx4_dev *dev, struct mlx4_ib_dev *ibdev)
 {
-	char name[32];
+	char name[80];
 	int eq_per_port = 0;
 	int added_eqs = 0;
 	int total_eqs = 0;
@@ -1833,8 +1833,8 @@ static void mlx4_ib_alloc_eqs(struct mlx4_dev *dev, struct mlx4_ib_dev *ibdev)
 	eq = 0;
 	mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_IB) {
 		for (j = 0; j < eq_per_port; j++) {
-			sprintf(name, "mlx4-ib-%d-%d@%s",
-				i, j, dev->pdev->bus->name);
+			snprintf(name, sizeof(name), "mlx4-ib-%d-%d@%s",
+				 i, j, dev->pdev->bus->name);
 			/* Set IRQ for specific name (per ring) */
 			if (mlx4_assign_eq(dev, name, NULL,
 					   &ibdev->eq_table[eq])) {
@@ -2048,8 +2048,9 @@ static void *mlx4_ib_add(struct mlx4_dev *dev)
 			err = mlx4_counter_alloc(ibdev->dev, &ibdev->counters[i]);
 			if (err)
 				ibdev->counters[i] = -1;
-		} else
-				ibdev->counters[i] = -1;
+		} else {
+			ibdev->counters[i] = -1;
+		}
 	}
 
 	mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_IB)

drivers/infiniband/hw/mlx4/mr.c

@@ -90,11 +90,11 @@ int mlx4_ib_umem_write_mtt(struct mlx4_ib_dev *dev, struct mlx4_mtt *mtt,
 			   struct ib_umem *umem)
 {
 	u64 *pages;
-	struct ib_umem_chunk *chunk;
-	int i, j, k;
+	int i, k, entry;
 	int n;
 	int len;
 	int err = 0;
+	struct scatterlist *sg;
 
 	pages = (u64 *) __get_free_page(GFP_KERNEL);
 	if (!pages)
@@ -102,26 +102,25 @@ int mlx4_ib_umem_write_mtt(struct mlx4_ib_dev *dev, struct mlx4_mtt *mtt,
 
 	i = n = 0;
 
-	list_for_each_entry(chunk, &umem->chunk_list, list)
-		for (j = 0; j < chunk->nmap; ++j) {
-			len = sg_dma_len(&chunk->page_list[j]) >> mtt->page_shift;
-			for (k = 0; k < len; ++k) {
-				pages[i++] = sg_dma_address(&chunk->page_list[j]) +
-					umem->page_size * k;
-				/*
-				 * Be friendly to mlx4_write_mtt() and
-				 * pass it chunks of appropriate size.
-				 */
-				if (i == PAGE_SIZE / sizeof (u64)) {
-					err = mlx4_write_mtt(dev->dev, mtt, n,
-							     i, pages);
-					if (err)
-						goto out;
-					n += i;
-					i = 0;
-				}
+	for_each_sg(umem->sg_head.sgl, sg, umem->nmap, entry) {
+		len = sg_dma_len(sg) >> mtt->page_shift;
+		for (k = 0; k < len; ++k) {
+			pages[i++] = sg_dma_address(sg) +
+				umem->page_size * k;
+			/*
+			 * Be friendly to mlx4_write_mtt() and
+			 * pass it chunks of appropriate size.
+			 */
+			if (i == PAGE_SIZE / sizeof (u64)) {
+				err = mlx4_write_mtt(dev->dev, mtt, n,
+						     i, pages);
+				if (err)
+					goto out;
+				n += i;
+				i = 0;
 			}
 		}
+	}
 
 	if (i)
 		err = mlx4_write_mtt(dev->dev, mtt, n, i, pages);

drivers/infiniband/hw/mlx4/qp.c

@@ -2082,7 +2082,7 @@ static int build_mlx_header(struct mlx4_ib_sqp *sqp, struct ib_send_wr *wr,
 				return err;
 		}
 
-		if (ah->av.eth.vlan != 0xffff) {
+		if (ah->av.eth.vlan != cpu_to_be16(0xffff)) {
 			vlan = be16_to_cpu(ah->av.eth.vlan) & 0x0fff;
 			is_vlan = 1;
 		}

drivers/infiniband/hw/mlx5/cq.c

@@ -366,6 +366,38 @@ static void free_cq_buf(struct mlx5_ib_dev *dev, struct mlx5_ib_cq_buf *buf)
 	mlx5_buf_free(&dev->mdev, &buf->buf);
 }
 
+static void get_sig_err_item(struct mlx5_sig_err_cqe *cqe,
+			     struct ib_sig_err *item)
+{
+	u16 syndrome = be16_to_cpu(cqe->syndrome);
+
+#define GUARD_ERR   (1 << 13)
+#define APPTAG_ERR  (1 << 12)
+#define REFTAG_ERR  (1 << 11)
+
+	if (syndrome & GUARD_ERR) {
+		item->err_type = IB_SIG_BAD_GUARD;
+		item->expected = be32_to_cpu(cqe->expected_trans_sig) >> 16;
+		item->actual = be32_to_cpu(cqe->actual_trans_sig) >> 16;
+	} else
+	if (syndrome & REFTAG_ERR) {
+		item->err_type = IB_SIG_BAD_REFTAG;
+		item->expected = be32_to_cpu(cqe->expected_reftag);
+		item->actual = be32_to_cpu(cqe->actual_reftag);
+	} else
+	if (syndrome & APPTAG_ERR) {
+		item->err_type = IB_SIG_BAD_APPTAG;
+		item->expected = be32_to_cpu(cqe->expected_trans_sig) & 0xffff;
+		item->actual = be32_to_cpu(cqe->actual_trans_sig) & 0xffff;
+	} else {
+		pr_err("Got signature completion error with bad syndrome %04x\n",
+		       syndrome);
+	}
+
+	item->sig_err_offset = be64_to_cpu(cqe->err_offset);
+	item->key = be32_to_cpu(cqe->mkey);
+}
+
 static int mlx5_poll_one(struct mlx5_ib_cq *cq,
 			 struct mlx5_ib_qp **cur_qp,
 			 struct ib_wc *wc)
@@ -375,6 +407,9 @@ static int mlx5_poll_one(struct mlx5_ib_cq *cq,
 	struct mlx5_cqe64 *cqe64;
 	struct mlx5_core_qp *mqp;
 	struct mlx5_ib_wq *wq;
+	struct mlx5_sig_err_cqe *sig_err_cqe;
+	struct mlx5_core_mr *mmr;
+	struct mlx5_ib_mr *mr;
 	uint8_t opcode;
 	uint32_t qpn;
 	u16 wqe_ctr;
@@ -475,6 +510,33 @@ repoll:
 			}
 		}
 		break;
+	case MLX5_CQE_SIG_ERR:
+		sig_err_cqe = (struct mlx5_sig_err_cqe *)cqe64;
+
+		read_lock(&dev->mdev.priv.mr_table.lock);
+		mmr = __mlx5_mr_lookup(&dev->mdev,
+				       mlx5_base_mkey(be32_to_cpu(sig_err_cqe->mkey)));
+		if (unlikely(!mmr)) {
+			read_unlock(&dev->mdev.priv.mr_table.lock);
+			mlx5_ib_warn(dev, "CQE@CQ %06x for unknown MR %6x\n",
+				     cq->mcq.cqn, be32_to_cpu(sig_err_cqe->mkey));
+			return -EINVAL;
+		}
+
+		mr = to_mibmr(mmr);
+		get_sig_err_item(sig_err_cqe, &mr->sig->err_item);
+		mr->sig->sig_err_exists = true;
+		mr->sig->sigerr_count++;
+
+		mlx5_ib_warn(dev, "CQN: 0x%x Got SIGERR on key: 0x%x err_type %x err_offset %llx expected %x actual %x\n",
+			     cq->mcq.cqn, mr->sig->err_item.key,
+			     mr->sig->err_item.err_type,
+			     mr->sig->err_item.sig_err_offset,
+			     mr->sig->err_item.expected,
+			     mr->sig->err_item.actual);
+
+		read_unlock(&dev->mdev.priv.mr_table.lock);
+		goto repoll;
 	}
 
 	return 0;

drivers/infiniband/hw/mlx5/doorbell.c

@@ -47,7 +47,6 @@ int mlx5_ib_db_map_user(struct mlx5_ib_ucontext *context, unsigned long virt,
 			struct mlx5_db *db)
 {
 	struct mlx5_ib_user_db_page *page;
-	struct ib_umem_chunk *chunk;
 	int err = 0;
 
 	mutex_lock(&context->db_page_mutex);
@@ -75,8 +74,7 @@ int mlx5_ib_db_map_user(struct mlx5_ib_ucontext *context, unsigned long virt,
 	list_add(&page->list, &context->db_page_list);
 
 found:
-	chunk = list_entry(page->umem->chunk_list.next, struct ib_umem_chunk, list);
-	db->dma		= sg_dma_address(chunk->page_list) + (virt & ~PAGE_MASK);
+	db->dma = sg_dma_address(page->umem->sg_head.sgl) + (virt & ~PAGE_MASK);
 	db->u.user_page = page;
 	++page->refcnt;
 

drivers/infiniband/hw/mlx5/main.c

@@ -273,6 +273,15 @@ static int mlx5_ib_query_device(struct ib_device *ibdev,
 	if (flags & MLX5_DEV_CAP_FLAG_XRC)
 		props->device_cap_flags |= IB_DEVICE_XRC;
 	props->device_cap_flags |= IB_DEVICE_MEM_MGT_EXTENSIONS;
+	if (flags & MLX5_DEV_CAP_FLAG_SIG_HAND_OVER) {
+		props->device_cap_flags |= IB_DEVICE_SIGNATURE_HANDOVER;
+		/* At this stage no support for signature handover */
+		props->sig_prot_cap = IB_PROT_T10DIF_TYPE_1 |
+				      IB_PROT_T10DIF_TYPE_2 |
+				      IB_PROT_T10DIF_TYPE_3;
+		props->sig_guard_cap = IB_GUARD_T10DIF_CRC |
+				       IB_GUARD_T10DIF_CSUM;
+	}
 
 	props->vendor_id	   = be32_to_cpup((__be32 *)(out_mad->data + 36)) &
 		0xffffff;
@@ -1423,12 +1432,15 @@ static int init_one(struct pci_dev *pdev,
 	dev->ib_dev.get_dma_mr		= mlx5_ib_get_dma_mr;
 	dev->ib_dev.reg_user_mr		= mlx5_ib_reg_user_mr;
 	dev->ib_dev.dereg_mr		= mlx5_ib_dereg_mr;
+	dev->ib_dev.destroy_mr		= mlx5_ib_destroy_mr;
 	dev->ib_dev.attach_mcast	= mlx5_ib_mcg_attach;
 	dev->ib_dev.detach_mcast	= mlx5_ib_mcg_detach;
 	dev->ib_dev.process_mad		= mlx5_ib_process_mad;
+	dev->ib_dev.create_mr		= mlx5_ib_create_mr;
 	dev->ib_dev.alloc_fast_reg_mr	= mlx5_ib_alloc_fast_reg_mr;
 	dev->ib_dev.alloc_fast_reg_page_list = mlx5_ib_alloc_fast_reg_page_list;
 	dev->ib_dev.free_fast_reg_page_list  = mlx5_ib_free_fast_reg_page_list;
+	dev->ib_dev.check_mr_status	= mlx5_ib_check_mr_status;
 
 	if (mdev->caps.flags & MLX5_DEV_CAP_FLAG_XRC) {
 		dev->ib_dev.alloc_xrcd = mlx5_ib_alloc_xrcd;

drivers/infiniband/hw/mlx5/mem.c

@@ -44,16 +44,17 @@
 void mlx5_ib_cont_pages(struct ib_umem *umem, u64 addr, int *count, int *shift,
 			int *ncont, int *order)
 {
-	struct ib_umem_chunk *chunk;
 	unsigned long tmp;
 	unsigned long m;
-	int i, j, k;
+	int i, k;
 	u64 base = 0;
 	int p = 0;
 	int skip;
 	int mask;
 	u64 len;
 	u64 pfn;
+	struct scatterlist *sg;
+	int entry;
 
 	addr = addr >> PAGE_SHIFT;
 	tmp = (unsigned long)addr;
@@ -61,32 +62,31 @@ void mlx5_ib_cont_pages(struct ib_umem *umem, u64 addr, int *count, int *shift,
 	skip = 1 << m;
 	mask = skip - 1;
 	i = 0;
-	list_for_each_entry(chunk, &umem->chunk_list, list)
-		for (j = 0; j < chunk->nmap; j++) {
-			len = sg_dma_len(&chunk->page_list[j]) >> PAGE_SHIFT;
-			pfn = sg_dma_address(&chunk->page_list[j]) >> PAGE_SHIFT;
-			for (k = 0; k < len; k++) {
-				if (!(i & mask)) {
-					tmp = (unsigned long)pfn;
-					m = min(m, find_first_bit(&tmp, sizeof(tmp)));
+	for_each_sg(umem->sg_head.sgl, sg, umem->nmap, entry) {
+		len = sg_dma_len(sg) >> PAGE_SHIFT;
+		pfn = sg_dma_address(sg) >> PAGE_SHIFT;
+		for (k = 0; k < len; k++) {
+			if (!(i & mask)) {
+				tmp = (unsigned long)pfn;
+				m = min(m, find_first_bit(&tmp, sizeof(tmp)));
+				skip = 1 << m;
+				mask = skip - 1;
+				base = pfn;
+				p = 0;
+			} else {
+				if (base + p != pfn) {
+					tmp = (unsigned long)p;
+					m = find_first_bit(&tmp, sizeof(tmp));
 					skip = 1 << m;
 					mask = skip - 1;
 					base = pfn;
 					p = 0;
-				} else {
-					if (base + p != pfn) {
-						tmp = (unsigned long)p;
-						m = find_first_bit(&tmp, sizeof(tmp));
-						skip = 1 << m;
-						mask = skip - 1;
-						base = pfn;
-						p = 0;
-					}
 				}
-				p++;
-				i++;
 			}
+			p++;
+			i++;
 		}
+	}
 
 	if (i) {
 		m = min_t(unsigned long, ilog2(roundup_pow_of_two(i)), m);
@@ -112,32 +112,32 @@ void mlx5_ib_populate_pas(struct mlx5_ib_dev *dev, struct ib_umem *umem,
 {
 	int shift = page_shift - PAGE_SHIFT;
 	int mask = (1 << shift) - 1;
-	struct ib_umem_chunk *chunk;
-	int i, j, k;
+	int i, k;
 	u64 cur = 0;
 	u64 base;
 	int len;
+	struct scatterlist *sg;
+	int entry;
 
 	i = 0;
-	list_for_each_entry(chunk, &umem->chunk_list, list)
-		for (j = 0; j < chunk->nmap; j++) {
-			len = sg_dma_len(&chunk->page_list[j]) >> PAGE_SHIFT;
-			base = sg_dma_address(&chunk->page_list[j]);
-			for (k = 0; k < len; k++) {
-				if (!(i & mask)) {
-					cur = base + (k << PAGE_SHIFT);
-					if (umr)
-						cur |= 3;
+	for_each_sg(umem->sg_head.sgl, sg, umem->nmap, entry) {
+		len = sg_dma_len(sg) >> PAGE_SHIFT;
+		base = sg_dma_address(sg);
+		for (k = 0; k < len; k++) {
+			if (!(i & mask)) {
+				cur = base + (k << PAGE_SHIFT);
+				if (umr)
+					cur |= 3;
 
-					pas[i >> shift] = cpu_to_be64(cur);
-					mlx5_ib_dbg(dev, "pas[%d] 0x%llx\n",
-						    i >> shift, be64_to_cpu(pas[i >> shift]));
-				}  else
-					mlx5_ib_dbg(dev, "=====> 0x%llx\n",
-						    base + (k << PAGE_SHIFT));
-				i++;
-			}
+				pas[i >> shift] = cpu_to_be64(cur);
+				mlx5_ib_dbg(dev, "pas[%d] 0x%llx\n",
+					    i >> shift, be64_to_cpu(pas[i >> shift]));
+			}  else
+				mlx5_ib_dbg(dev, "=====> 0x%llx\n",
+					    base + (k << PAGE_SHIFT));
+			i++;
 		}
+	}
 }
 
 int mlx5_ib_get_buf_offset(u64 addr, int page_shift, u32 *offset)

drivers/infiniband/hw/mlx5/mlx5_ib.h

@@ -189,6 +189,9 @@ struct mlx5_ib_qp {
 
 	int			create_type;
 	u32			pa_lkey;
+
+	/* Store signature errors */
+	bool			signature_en;
 };
 
 struct mlx5_ib_cq_buf {
@@ -265,6 +268,7 @@ struct mlx5_ib_mr {
 	enum ib_wc_status	status;
 	struct mlx5_ib_dev     *dev;
 	struct mlx5_create_mkey_mbox_out out;
+	struct mlx5_core_sig_ctx    *sig;
 };
 
 struct mlx5_ib_fast_reg_page_list {
@@ -396,6 +400,11 @@ static inline struct mlx5_ib_qp *to_mibqp(struct mlx5_core_qp *mqp)
 	return container_of(mqp, struct mlx5_ib_qp, mqp);
 }
 
+static inline struct mlx5_ib_mr *to_mibmr(struct mlx5_core_mr *mmr)
+{
+	return container_of(mmr, struct mlx5_ib_mr, mmr);
+}
+
 static inline struct mlx5_ib_pd *to_mpd(struct ib_pd *ibpd)
 {
 	return container_of(ibpd, struct mlx5_ib_pd, ibpd);
@@ -495,6 +504,9 @@ struct ib_mr *mlx5_ib_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 				  u64 virt_addr, int access_flags,
 				  struct ib_udata *udata);
 int mlx5_ib_dereg_mr(struct ib_mr *ibmr);
+int mlx5_ib_destroy_mr(struct ib_mr *ibmr);
+struct ib_mr *mlx5_ib_create_mr(struct ib_pd *pd,
+				struct ib_mr_init_attr *mr_init_attr);
 struct ib_mr *mlx5_ib_alloc_fast_reg_mr(struct ib_pd *pd,
 					int max_page_list_len);
 struct ib_fast_reg_page_list *mlx5_ib_alloc_fast_reg_page_list(struct ib_device *ibdev,
@@ -530,6 +542,8 @@ int mlx5_mr_cache_init(struct mlx5_ib_dev *dev);
 int mlx5_mr_cache_cleanup(struct mlx5_ib_dev *dev);
 int mlx5_mr_ib_cont_pages(struct ib_umem *umem, u64 addr, int *count, int *shift);
 void mlx5_umr_cq_handler(struct ib_cq *cq, void *cq_context);
+int mlx5_ib_check_mr_status(struct ib_mr *ibmr, u32 check_mask,
+			    struct ib_mr_status *mr_status);
 
 static inline void init_query_mad(struct ib_smp *mad)
 {

drivers/infiniband/hw/mlx5/mr.c

@@ -992,6 +992,122 @@ int mlx5_ib_dereg_mr(struct ib_mr *ibmr)
 	return 0;
 }
 
+struct ib_mr *mlx5_ib_create_mr(struct ib_pd *pd,
+				struct ib_mr_init_attr *mr_init_attr)
+{
+	struct mlx5_ib_dev *dev = to_mdev(pd->device);
+	struct mlx5_create_mkey_mbox_in *in;
+	struct mlx5_ib_mr *mr;
+	int access_mode, err;
+	int ndescs = roundup(mr_init_attr->max_reg_descriptors, 4);
+
+	mr = kzalloc(sizeof(*mr), GFP_KERNEL);
+	if (!mr)
+		return ERR_PTR(-ENOMEM);
+
+	in = kzalloc(sizeof(*in), GFP_KERNEL);
+	if (!in) {
+		err = -ENOMEM;
+		goto err_free;
+	}
+
+	in->seg.status = 1 << 6; /* free */
+	in->seg.xlt_oct_size = cpu_to_be32(ndescs);
+	in->seg.qpn_mkey7_0 = cpu_to_be32(0xffffff << 8);
+	in->seg.flags_pd = cpu_to_be32(to_mpd(pd)->pdn);
+	access_mode = MLX5_ACCESS_MODE_MTT;
+
+	if (mr_init_attr->flags & IB_MR_SIGNATURE_EN) {
+		u32 psv_index[2];
+
+		in->seg.flags_pd = cpu_to_be32(be32_to_cpu(in->seg.flags_pd) |
+							   MLX5_MKEY_BSF_EN);
+		in->seg.bsfs_octo_size = cpu_to_be32(MLX5_MKEY_BSF_OCTO_SIZE);
+		mr->sig = kzalloc(sizeof(*mr->sig), GFP_KERNEL);
+		if (!mr->sig) {
+			err = -ENOMEM;
+			goto err_free_in;
+		}
+
+		/* create mem & wire PSVs */
+		err = mlx5_core_create_psv(&dev->mdev, to_mpd(pd)->pdn,
+					   2, psv_index);
+		if (err)
+			goto err_free_sig;
+
+		access_mode = MLX5_ACCESS_MODE_KLM;
+		mr->sig->psv_memory.psv_idx = psv_index[0];
+		mr->sig->psv_wire.psv_idx = psv_index[1];
+
+		mr->sig->sig_status_checked = true;
+		mr->sig->sig_err_exists = false;
+		/* Next UMR, Arm SIGERR */
+		++mr->sig->sigerr_count;
+	}
+
+	in->seg.flags = MLX5_PERM_UMR_EN | access_mode;
+	err = mlx5_core_create_mkey(&dev->mdev, &mr->mmr, in, sizeof(*in),
+				    NULL, NULL, NULL);
+	if (err)
+		goto err_destroy_psv;
+
+	mr->ibmr.lkey = mr->mmr.key;
+	mr->ibmr.rkey = mr->mmr.key;
+	mr->umem = NULL;
+	kfree(in);
+
+	return &mr->ibmr;
+
+err_destroy_psv:
+	if (mr->sig) {
+		if (mlx5_core_destroy_psv(&dev->mdev,
+					  mr->sig->psv_memory.psv_idx))
+			mlx5_ib_warn(dev, "failed to destroy mem psv %d\n",
+				     mr->sig->psv_memory.psv_idx);
+		if (mlx5_core_destroy_psv(&dev->mdev,
+					  mr->sig->psv_wire.psv_idx))
+			mlx5_ib_warn(dev, "failed to destroy wire psv %d\n",
+				     mr->sig->psv_wire.psv_idx);
+	}
+err_free_sig:
+	kfree(mr->sig);
+err_free_in:
+	kfree(in);
+err_free:
+	kfree(mr);
+	return ERR_PTR(err);
+}
+
+int mlx5_ib_destroy_mr(struct ib_mr *ibmr)
+{
+	struct mlx5_ib_dev *dev = to_mdev(ibmr->device);
+	struct mlx5_ib_mr *mr = to_mmr(ibmr);
+	int err;
+
+	if (mr->sig) {
+		if (mlx5_core_destroy_psv(&dev->mdev,
+					  mr->sig->psv_memory.psv_idx))
+			mlx5_ib_warn(dev, "failed to destroy mem psv %d\n",
+				     mr->sig->psv_memory.psv_idx);
+		if (mlx5_core_destroy_psv(&dev->mdev,
+					  mr->sig->psv_wire.psv_idx))
+			mlx5_ib_warn(dev, "failed to destroy wire psv %d\n",
+				     mr->sig->psv_wire.psv_idx);
+		kfree(mr->sig);
+	}
+
+	err = mlx5_core_destroy_mkey(&dev->mdev, &mr->mmr);
+	if (err) {
+		mlx5_ib_warn(dev, "failed to destroy mkey 0x%x (%d)\n",
+			     mr->mmr.key, err);
+		return err;
+	}
+
+	kfree(mr);
+
+	return err;
+}
+
 struct ib_mr *mlx5_ib_alloc_fast_reg_mr(struct ib_pd *pd,
 					int max_page_list_len)
 {
@@ -1077,3 +1193,44 @@ void mlx5_ib_free_fast_reg_page_list(struct ib_fast_reg_page_list *page_list)
 	kfree(mfrpl->ibfrpl.page_list);
 	kfree(mfrpl);
 }
+
+int mlx5_ib_check_mr_status(struct ib_mr *ibmr, u32 check_mask,
+			    struct ib_mr_status *mr_status)
+{
+	struct mlx5_ib_mr *mmr = to_mmr(ibmr);
+	int ret = 0;
+
+	if (check_mask & ~IB_MR_CHECK_SIG_STATUS) {
+		pr_err("Invalid status check mask\n");
+		ret = -EINVAL;
+		goto done;
+	}
+
+	mr_status->fail_status = 0;
+	if (check_mask & IB_MR_CHECK_SIG_STATUS) {
+		if (!mmr->sig) {
+			ret = -EINVAL;
+			pr_err("signature status check requested on a non-signature enabled MR\n");
+			goto done;
+		}
+
+		mmr->sig->sig_status_checked = true;
+		if (!mmr->sig->sig_err_exists)
+			goto done;
+
+		if (ibmr->lkey == mmr->sig->err_item.key)
+			memcpy(&mr_status->sig_err, &mmr->sig->err_item,
+			       sizeof(mr_status->sig_err));
+		else {
+			mr_status->sig_err.err_type = IB_SIG_BAD_GUARD;
+			mr_status->sig_err.sig_err_offset = 0;
+			mr_status->sig_err.key = mmr->sig->err_item.key;
+		}
+
+		mmr->sig->sig_err_exists = false;
+		mr_status->fail_status |= IB_MR_CHECK_SIG_STATUS;
+	}
+
+done:
+	return ret;
+}

drivers/infiniband/hw/mlx5/qp.c

@@ -256,8 +256,11 @@ static int calc_send_wqe(struct ib_qp_init_attr *attr)
 	}
 
 	size += attr->cap.max_send_sge * sizeof(struct mlx5_wqe_data_seg);
-
-	return ALIGN(max_t(int, inl_size, size), MLX5_SEND_WQE_BB);
+	if (attr->create_flags & IB_QP_CREATE_SIGNATURE_EN &&
+	    ALIGN(max_t(int, inl_size, size), MLX5_SEND_WQE_BB) < MLX5_SIG_WQE_SIZE)
+			return MLX5_SIG_WQE_SIZE;
+	else
+		return ALIGN(max_t(int, inl_size, size), MLX5_SEND_WQE_BB);
 }
 
 static int calc_sq_size(struct mlx5_ib_dev *dev, struct ib_qp_init_attr *attr,
@@ -284,6 +287,9 @@ static int calc_sq_size(struct mlx5_ib_dev *dev, struct ib_qp_init_attr *attr,
 		sizeof(struct mlx5_wqe_inline_seg);
 	attr->cap.max_inline_data = qp->max_inline_data;
 
+	if (attr->create_flags & IB_QP_CREATE_SIGNATURE_EN)
+		qp->signature_en = true;
+
 	wq_size = roundup_pow_of_two(attr->cap.max_send_wr * wqe_size);
 	qp->sq.wqe_cnt = wq_size / MLX5_SEND_WQE_BB;
 	if (qp->sq.wqe_cnt > dev->mdev.caps.max_wqes) {
@@ -665,7 +671,7 @@ static int create_kernel_qp(struct mlx5_ib_dev *dev,
 	int err;
 
 	uuari = &dev->mdev.priv.uuari;
-	if (init_attr->create_flags)
+	if (init_attr->create_flags & ~IB_QP_CREATE_SIGNATURE_EN)
 		return -EINVAL;
 
 	if (init_attr->qp_type == MLX5_IB_QPT_REG_UMR)
@@ -1771,6 +1777,27 @@ static __be64 frwr_mkey_mask(void)
 	return cpu_to_be64(result);
 }
 
+static __be64 sig_mkey_mask(void)
+{
+	u64 result;
+
+	result = MLX5_MKEY_MASK_LEN		|
+		MLX5_MKEY_MASK_PAGE_SIZE	|
+		MLX5_MKEY_MASK_START_ADDR	|
+		MLX5_MKEY_MASK_EN_SIGERR	|
+		MLX5_MKEY_MASK_EN_RINVAL	|
+		MLX5_MKEY_MASK_KEY		|
+		MLX5_MKEY_MASK_LR		|
+		MLX5_MKEY_MASK_LW		|
+		MLX5_MKEY_MASK_RR		|
+		MLX5_MKEY_MASK_RW		|
+		MLX5_MKEY_MASK_SMALL_FENCE	|
+		MLX5_MKEY_MASK_FREE		|
+		MLX5_MKEY_MASK_BSF_EN;
+
+	return cpu_to_be64(result);
+}
+
 static void set_frwr_umr_segment(struct mlx5_wqe_umr_ctrl_seg *umr,
 				 struct ib_send_wr *wr, int li)
 {
@@ -1826,7 +1853,7 @@ static u8 get_umr_flags(int acc)
 	       (acc & IB_ACCESS_REMOTE_WRITE  ? MLX5_PERM_REMOTE_WRITE : 0) |
 	       (acc & IB_ACCESS_REMOTE_READ   ? MLX5_PERM_REMOTE_READ  : 0) |
 	       (acc & IB_ACCESS_LOCAL_WRITE   ? MLX5_PERM_LOCAL_WRITE  : 0) |
-		MLX5_PERM_LOCAL_READ | MLX5_PERM_UMR_EN | MLX5_ACCESS_MODE_MTT;
+		MLX5_PERM_LOCAL_READ | MLX5_PERM_UMR_EN;
 }
 
 static void set_mkey_segment(struct mlx5_mkey_seg *seg, struct ib_send_wr *wr,
@@ -1838,7 +1865,8 @@ static void set_mkey_segment(struct mlx5_mkey_seg *seg, struct ib_send_wr *wr,
 		return;
 	}
 
-	seg->flags = get_umr_flags(wr->wr.fast_reg.access_flags);
+	seg->flags = get_umr_flags(wr->wr.fast_reg.access_flags) |
+		     MLX5_ACCESS_MODE_MTT;
 	*writ = seg->flags & (MLX5_PERM_LOCAL_WRITE | IB_ACCESS_REMOTE_WRITE);
 	seg->qpn_mkey7_0 = cpu_to_be32((wr->wr.fast_reg.rkey & 0xff) | 0xffffff00);
 	seg->flags_pd = cpu_to_be32(MLX5_MKEY_REMOTE_INVAL);
@@ -1954,6 +1982,342 @@ static int set_data_inl_seg(struct mlx5_ib_qp *qp, struct ib_send_wr *wr,
 	return 0;
 }
 
+static u16 prot_field_size(enum ib_signature_type type)
+{
+	switch (type) {
+	case IB_SIG_TYPE_T10_DIF:
+		return MLX5_DIF_SIZE;
+	default:
+		return 0;
+	}
+}
+
+static u8 bs_selector(int block_size)
+{
+	switch (block_size) {
+	case 512:	    return 0x1;
+	case 520:	    return 0x2;
+	case 4096:	    return 0x3;
+	case 4160:	    return 0x4;
+	case 1073741824:    return 0x5;
+	default:	    return 0;
+	}
+}
+
+static int format_selector(struct ib_sig_attrs *attr,
+			   struct ib_sig_domain *domain,
+			   int *selector)
+{
+
+#define FORMAT_DIF_NONE		0
+#define FORMAT_DIF_CRC_INC	8
+#define FORMAT_DIF_CRC_NO_INC	12
+#define FORMAT_DIF_CSUM_INC	13
+#define FORMAT_DIF_CSUM_NO_INC	14
+
+	switch (domain->sig.dif.type) {
+	case IB_T10DIF_NONE:
+		/* No DIF */
+		*selector = FORMAT_DIF_NONE;
+		break;
+	case IB_T10DIF_TYPE1: /* Fall through */
+	case IB_T10DIF_TYPE2:
+		switch (domain->sig.dif.bg_type) {
+		case IB_T10DIF_CRC:
+			*selector = FORMAT_DIF_CRC_INC;
+			break;
+		case IB_T10DIF_CSUM:
+			*selector = FORMAT_DIF_CSUM_INC;
+			break;
+		default:
+			return 1;
+		}
+		break;
+	case IB_T10DIF_TYPE3:
+		switch (domain->sig.dif.bg_type) {
+		case IB_T10DIF_CRC:
+			*selector = domain->sig.dif.type3_inc_reftag ?
+					   FORMAT_DIF_CRC_INC :
+					   FORMAT_DIF_CRC_NO_INC;
+			break;
+		case IB_T10DIF_CSUM:
+			*selector = domain->sig.dif.type3_inc_reftag ?
+					   FORMAT_DIF_CSUM_INC :
+					   FORMAT_DIF_CSUM_NO_INC;
+			break;
+		default:
+			return 1;
+		}
+		break;
+	default:
+		return 1;
+	}
+
+	return 0;
+}
+
+static int mlx5_set_bsf(struct ib_mr *sig_mr,
+			struct ib_sig_attrs *sig_attrs,
+			struct mlx5_bsf *bsf, u32 data_size)
+{
+	struct mlx5_core_sig_ctx *msig = to_mmr(sig_mr)->sig;
+	struct mlx5_bsf_basic *basic = &bsf->basic;
+	struct ib_sig_domain *mem = &sig_attrs->mem;
+	struct ib_sig_domain *wire = &sig_attrs->wire;
+	int ret, selector;
+
+	switch (sig_attrs->mem.sig_type) {
+	case IB_SIG_TYPE_T10_DIF:
+		if (sig_attrs->wire.sig_type != IB_SIG_TYPE_T10_DIF)
+			return -EINVAL;
+
+		/* Input domain check byte mask */
+		basic->check_byte_mask = sig_attrs->check_mask;
+		if (mem->sig.dif.pi_interval == wire->sig.dif.pi_interval &&
+		    mem->sig.dif.type == wire->sig.dif.type) {
+			/* Same block structure */
+			basic->bsf_size_sbs = 1 << 4;
+			if (mem->sig.dif.bg_type == wire->sig.dif.bg_type)
+				basic->wire.copy_byte_mask = 0xff;
+			else
+				basic->wire.copy_byte_mask = 0x3f;
+		} else
+			basic->wire.bs_selector = bs_selector(wire->sig.dif.pi_interval);
+
+		basic->mem.bs_selector = bs_selector(mem->sig.dif.pi_interval);
+		basic->raw_data_size = cpu_to_be32(data_size);
+
+		ret = format_selector(sig_attrs, mem, &selector);
+		if (ret)
+			return -EINVAL;
+		basic->m_bfs_psv = cpu_to_be32(selector << 24 |
+					       msig->psv_memory.psv_idx);
+
+		ret = format_selector(sig_attrs, wire, &selector);
+		if (ret)
+			return -EINVAL;
+		basic->w_bfs_psv = cpu_to_be32(selector << 24 |
+					       msig->psv_wire.psv_idx);
+		break;
+
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int set_sig_data_segment(struct ib_send_wr *wr, struct mlx5_ib_qp *qp,
+				void **seg, int *size)
+{
+	struct ib_sig_attrs *sig_attrs = wr->wr.sig_handover.sig_attrs;
+	struct ib_mr *sig_mr = wr->wr.sig_handover.sig_mr;
+	struct mlx5_bsf *bsf;
+	u32 data_len = wr->sg_list->length;
+	u32 data_key = wr->sg_list->lkey;
+	u64 data_va = wr->sg_list->addr;
+	int ret;
+	int wqe_size;
+
+	if (!wr->wr.sig_handover.prot) {
+		/**
+		 * Source domain doesn't contain signature information
+		 * So need construct:
+		 *                  ------------------
+		 *                 |     data_klm     |
+		 *                  ------------------
+		 *                 |       BSF        |
+		 *                  ------------------
+		 **/
+		struct mlx5_klm *data_klm = *seg;
+
+		data_klm->bcount = cpu_to_be32(data_len);
+		data_klm->key = cpu_to_be32(data_key);
+		data_klm->va = cpu_to_be64(data_va);
+		wqe_size = ALIGN(sizeof(*data_klm), 64);
+	} else {
+		/**
+		 * Source domain contains signature information
+		 * So need construct a strided block format:
+		 *               ---------------------------
+		 *              |     stride_block_ctrl     |
+		 *               ---------------------------
+		 *              |          data_klm         |
+		 *               ---------------------------
+		 *              |          prot_klm         |
+		 *               ---------------------------
+		 *              |             BSF           |
+		 *               ---------------------------
+		 **/
+		struct mlx5_stride_block_ctrl_seg *sblock_ctrl;
+		struct mlx5_stride_block_entry *data_sentry;
+		struct mlx5_stride_block_entry *prot_sentry;
+		u32 prot_key = wr->wr.sig_handover.prot->lkey;
+		u64 prot_va = wr->wr.sig_handover.prot->addr;
+		u16 block_size = sig_attrs->mem.sig.dif.pi_interval;
+		int prot_size;
+
+		sblock_ctrl = *seg;
+		data_sentry = (void *)sblock_ctrl + sizeof(*sblock_ctrl);
+		prot_sentry = (void *)data_sentry + sizeof(*data_sentry);
+
+		prot_size = prot_field_size(sig_attrs->mem.sig_type);
+		if (!prot_size) {
+			pr_err("Bad block size given: %u\n", block_size);
+			return -EINVAL;
+		}
+		sblock_ctrl->bcount_per_cycle = cpu_to_be32(block_size +
+							    prot_size);
+		sblock_ctrl->op = cpu_to_be32(MLX5_STRIDE_BLOCK_OP);
+		sblock_ctrl->repeat_count = cpu_to_be32(data_len / block_size);
+		sblock_ctrl->num_entries = cpu_to_be16(2);
+
+		data_sentry->bcount = cpu_to_be16(block_size);
+		data_sentry->key = cpu_to_be32(data_key);
+		data_sentry->va = cpu_to_be64(data_va);
+		prot_sentry->bcount = cpu_to_be16(prot_size);
+		prot_sentry->key = cpu_to_be32(prot_key);
+
+		if (prot_key == data_key && prot_va == data_va) {
+			/**
+			 * The data and protection are interleaved
+			 * in a single memory region
+			 **/
+			prot_sentry->va = cpu_to_be64(data_va + block_size);
+			prot_sentry->stride = cpu_to_be16(block_size + prot_size);
+			data_sentry->stride = prot_sentry->stride;
+		} else {
+			/* The data and protection are two different buffers */
+			prot_sentry->va = cpu_to_be64(prot_va);
+			data_sentry->stride = cpu_to_be16(block_size);
+			prot_sentry->stride = cpu_to_be16(prot_size);
+		}
+		wqe_size = ALIGN(sizeof(*sblock_ctrl) + sizeof(*data_sentry) +
+				 sizeof(*prot_sentry), 64);
+	}
+
+	*seg += wqe_size;
+	*size += wqe_size / 16;
+	if (unlikely((*seg == qp->sq.qend)))
+		*seg = mlx5_get_send_wqe(qp, 0);
+
+	bsf = *seg;
+	ret = mlx5_set_bsf(sig_mr, sig_attrs, bsf, data_len);
+	if (ret)
+		return -EINVAL;
+
+	*seg += sizeof(*bsf);
+	*size += sizeof(*bsf) / 16;
+	if (unlikely((*seg == qp->sq.qend)))
+		*seg = mlx5_get_send_wqe(qp, 0);
+
+	return 0;
+}
+
+static void set_sig_mkey_segment(struct mlx5_mkey_seg *seg,
+				 struct ib_send_wr *wr, u32 nelements,
+				 u32 length, u32 pdn)
+{
+	struct ib_mr *sig_mr = wr->wr.sig_handover.sig_mr;
+	u32 sig_key = sig_mr->rkey;
+	u8 sigerr = to_mmr(sig_mr)->sig->sigerr_count & 1;
+
+	memset(seg, 0, sizeof(*seg));
+
+	seg->flags = get_umr_flags(wr->wr.sig_handover.access_flags) |
+				   MLX5_ACCESS_MODE_KLM;
+	seg->qpn_mkey7_0 = cpu_to_be32((sig_key & 0xff) | 0xffffff00);
+	seg->flags_pd = cpu_to_be32(MLX5_MKEY_REMOTE_INVAL | sigerr << 26 |
+				    MLX5_MKEY_BSF_EN | pdn);
+	seg->len = cpu_to_be64(length);
+	seg->xlt_oct_size = cpu_to_be32(be16_to_cpu(get_klm_octo(nelements)));
+	seg->bsfs_octo_size = cpu_to_be32(MLX5_MKEY_BSF_OCTO_SIZE);
+}
+
+static void set_sig_umr_segment(struct mlx5_wqe_umr_ctrl_seg *umr,
+				struct ib_send_wr *wr, u32 nelements)
+{
+	memset(umr, 0, sizeof(*umr));
+
+	umr->flags = MLX5_FLAGS_INLINE | MLX5_FLAGS_CHECK_FREE;
+	umr->klm_octowords = get_klm_octo(nelements);
+	umr->bsf_octowords = cpu_to_be16(MLX5_MKEY_BSF_OCTO_SIZE);
+	umr->mkey_mask = sig_mkey_mask();
+}
+
+
+static int set_sig_umr_wr(struct ib_send_wr *wr, struct mlx5_ib_qp *qp,
+			  void **seg, int *size)
+{
+	struct mlx5_ib_mr *sig_mr = to_mmr(wr->wr.sig_handover.sig_mr);
+	u32 pdn = get_pd(qp)->pdn;
+	u32 klm_oct_size;
+	int region_len, ret;
+
+	if (unlikely(wr->num_sge != 1) ||
+	    unlikely(wr->wr.sig_handover.access_flags &
+		     IB_ACCESS_REMOTE_ATOMIC) ||
+	    unlikely(!sig_mr->sig) || unlikely(!qp->signature_en) ||
+	    unlikely(!sig_mr->sig->sig_status_checked))
+		return -EINVAL;
+
+	/* length of the protected region, data + protection */
+	region_len = wr->sg_list->length;
+	if (wr->wr.sig_handover.prot)
+		region_len += wr->wr.sig_handover.prot->length;
+
+	/**
+	 * KLM octoword size - if protection was provided
+	 * then we use strided block format (3 octowords),
+	 * else we use single KLM (1 octoword)
+	 **/
+	klm_oct_size = wr->wr.sig_handover.prot ? 3 : 1;
+
+	set_sig_umr_segment(*seg, wr, klm_oct_size);
+	*seg += sizeof(struct mlx5_wqe_umr_ctrl_seg);
+	*size += sizeof(struct mlx5_wqe_umr_ctrl_seg) / 16;
+	if (unlikely((*seg == qp->sq.qend)))
+		*seg = mlx5_get_send_wqe(qp, 0);
+
+	set_sig_mkey_segment(*seg, wr, klm_oct_size, region_len, pdn);
+	*seg += sizeof(struct mlx5_mkey_seg);
+	*size += sizeof(struct mlx5_mkey_seg) / 16;
+	if (unlikely((*seg == qp->sq.qend)))
+		*seg = mlx5_get_send_wqe(qp, 0);
+
+	ret = set_sig_data_segment(wr, qp, seg, size);
+	if (ret)
+		return ret;
+
+	sig_mr->sig->sig_status_checked = false;
+	return 0;
+}
+
+static int set_psv_wr(struct ib_sig_domain *domain,
+		      u32 psv_idx, void **seg, int *size)
+{
+	struct mlx5_seg_set_psv *psv_seg = *seg;
+
+	memset(psv_seg, 0, sizeof(*psv_seg));
+	psv_seg->psv_num = cpu_to_be32(psv_idx);
+	switch (domain->sig_type) {
+	case IB_SIG_TYPE_T10_DIF:
+		psv_seg->transient_sig = cpu_to_be32(domain->sig.dif.bg << 16 |
+						     domain->sig.dif.app_tag);
+		psv_seg->ref_tag = cpu_to_be32(domain->sig.dif.ref_tag);
+
+		*seg += sizeof(*psv_seg);
+		*size += sizeof(*psv_seg) / 16;
+		break;
+
+	default:
+		pr_err("Bad signature type given.\n");
+		return 1;
+	}
+
+	return 0;
+}
+
 static int set_frwr_li_wr(void **seg, struct ib_send_wr *wr, int *size,
 			  struct mlx5_core_dev *mdev, struct mlx5_ib_pd *pd, struct mlx5_ib_qp *qp)
 {
@@ -2041,6 +2405,59 @@ static u8 get_fence(u8 fence, struct ib_send_wr *wr)
 	}
 }
 
+static int begin_wqe(struct mlx5_ib_qp *qp, void **seg,
+		     struct mlx5_wqe_ctrl_seg **ctrl,
+		     struct ib_send_wr *wr, int *idx,
+		     int *size, int nreq)
+{
+	int err = 0;
+
+	if (unlikely(mlx5_wq_overflow(&qp->sq, nreq, qp->ibqp.send_cq))) {
+		err = -ENOMEM;
+		return err;
+	}
+
+	*idx = qp->sq.cur_post & (qp->sq.wqe_cnt - 1);
+	*seg = mlx5_get_send_wqe(qp, *idx);
+	*ctrl = *seg;
+	*(uint32_t *)(*seg + 8) = 0;
+	(*ctrl)->imm = send_ieth(wr);
+	(*ctrl)->fm_ce_se = qp->sq_signal_bits |
+		(wr->send_flags & IB_SEND_SIGNALED ?
+		 MLX5_WQE_CTRL_CQ_UPDATE : 0) |
+		(wr->send_flags & IB_SEND_SOLICITED ?
+		 MLX5_WQE_CTRL_SOLICITED : 0);
+
+	*seg += sizeof(**ctrl);
+	*size = sizeof(**ctrl) / 16;
+
+	return err;
+}
+
+static void finish_wqe(struct mlx5_ib_qp *qp,
+		       struct mlx5_wqe_ctrl_seg *ctrl,
+		       u8 size, unsigned idx, u64 wr_id,
+		       int nreq, u8 fence, u8 next_fence,
+		       u32 mlx5_opcode)
+{
+	u8 opmod = 0;
+
+	ctrl->opmod_idx_opcode = cpu_to_be32(((u32)(qp->sq.cur_post) << 8) |
+					     mlx5_opcode | ((u32)opmod << 24));
+	ctrl->qpn_ds = cpu_to_be32(size | (qp->mqp.qpn << 8));
+	ctrl->fm_ce_se |= fence;
+	qp->fm_cache = next_fence;
+	if (unlikely(qp->wq_sig))
+		ctrl->signature = wq_sig(ctrl);
+
+	qp->sq.wrid[idx] = wr_id;
+	qp->sq.w_list[idx].opcode = mlx5_opcode;
+	qp->sq.wqe_head[idx] = qp->sq.head + nreq;
+	qp->sq.cur_post += DIV_ROUND_UP(size * 16, MLX5_SEND_WQE_BB);
+	qp->sq.w_list[idx].next = qp->sq.cur_post;
+}
+
+
 int mlx5_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 		      struct ib_send_wr **bad_wr)
 {
@@ -2048,13 +2465,13 @@ int mlx5_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 	struct mlx5_ib_dev *dev = to_mdev(ibqp->device);
 	struct mlx5_core_dev *mdev = &dev->mdev;
 	struct mlx5_ib_qp *qp = to_mqp(ibqp);
+	struct mlx5_ib_mr *mr;
 	struct mlx5_wqe_data_seg *dpseg;
 	struct mlx5_wqe_xrc_seg *xrc;
 	struct mlx5_bf *bf = qp->bf;
 	int uninitialized_var(size);
 	void *qend = qp->sq.qend;
 	unsigned long flags;
-	u32 mlx5_opcode;
 	unsigned idx;
 	int err = 0;
 	int inl = 0;
@@ -2063,7 +2480,6 @@ int mlx5_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 	int nreq;
 	int i;
 	u8 next_fence = 0;
-	u8 opmod = 0;
 	u8 fence;
 
 	spin_lock_irqsave(&qp->sq.lock, flags);
@@ -2076,36 +2492,23 @@ int mlx5_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 			goto out;
 		}
 
-		if (unlikely(mlx5_wq_overflow(&qp->sq, nreq, qp->ibqp.send_cq))) {
+		fence = qp->fm_cache;
+		num_sge = wr->num_sge;
+		if (unlikely(num_sge > qp->sq.max_gs)) {
 			mlx5_ib_warn(dev, "\n");
 			err = -ENOMEM;
 			*bad_wr = wr;
 			goto out;
 		}
 
-		fence = qp->fm_cache;
-		num_sge = wr->num_sge;
-		if (unlikely(num_sge > qp->sq.max_gs)) {
+		err = begin_wqe(qp, &seg, &ctrl, wr, &idx, &size, nreq);
+		if (err) {
 			mlx5_ib_warn(dev, "\n");
 			err = -ENOMEM;
 			*bad_wr = wr;
 			goto out;
 		}
 
-		idx = qp->sq.cur_post & (qp->sq.wqe_cnt - 1);
-		seg = mlx5_get_send_wqe(qp, idx);
-		ctrl = seg;
-		*(uint32_t *)(seg + 8) = 0;
-		ctrl->imm = send_ieth(wr);
-		ctrl->fm_ce_se = qp->sq_signal_bits |
-			(wr->send_flags & IB_SEND_SIGNALED ?
-			 MLX5_WQE_CTRL_CQ_UPDATE : 0) |
-			(wr->send_flags & IB_SEND_SOLICITED ?
-			 MLX5_WQE_CTRL_SOLICITED : 0);
-
-		seg += sizeof(*ctrl);
-		size = sizeof(*ctrl) / 16;
-
 		switch (ibqp->qp_type) {
 		case IB_QPT_XRC_INI:
 			xrc = seg;
@@ -2158,6 +2561,73 @@ int mlx5_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 				num_sge = 0;
 				break;
 
+			case IB_WR_REG_SIG_MR:
+				qp->sq.wr_data[idx] = IB_WR_REG_SIG_MR;
+				mr = to_mmr(wr->wr.sig_handover.sig_mr);
+
+				ctrl->imm = cpu_to_be32(mr->ibmr.rkey);
+				err = set_sig_umr_wr(wr, qp, &seg, &size);
+				if (err) {
+					mlx5_ib_warn(dev, "\n");
+					*bad_wr = wr;
+					goto out;
+				}
+
+				finish_wqe(qp, ctrl, size, idx, wr->wr_id,
+					   nreq, get_fence(fence, wr),
+					   next_fence, MLX5_OPCODE_UMR);
+				/*
+				 * SET_PSV WQEs are not signaled and solicited
+				 * on error
+				 */
+				wr->send_flags &= ~IB_SEND_SIGNALED;
+				wr->send_flags |= IB_SEND_SOLICITED;
+				err = begin_wqe(qp, &seg, &ctrl, wr,
+						&idx, &size, nreq);
+				if (err) {
+					mlx5_ib_warn(dev, "\n");
+					err = -ENOMEM;
+					*bad_wr = wr;
+					goto out;
+				}
+
+				err = set_psv_wr(&wr->wr.sig_handover.sig_attrs->mem,
+						 mr->sig->psv_memory.psv_idx, &seg,
+						 &size);
+				if (err) {
+					mlx5_ib_warn(dev, "\n");
+					*bad_wr = wr;
+					goto out;
+				}
+
+				finish_wqe(qp, ctrl, size, idx, wr->wr_id,
+					   nreq, get_fence(fence, wr),
+					   next_fence, MLX5_OPCODE_SET_PSV);
+				err = begin_wqe(qp, &seg, &ctrl, wr,
+						&idx, &size, nreq);
+				if (err) {
+					mlx5_ib_warn(dev, "\n");
+					err = -ENOMEM;
+					*bad_wr = wr;
+					goto out;
+				}
+
+				next_fence = MLX5_FENCE_MODE_INITIATOR_SMALL;
+				err = set_psv_wr(&wr->wr.sig_handover.sig_attrs->wire,
+						 mr->sig->psv_wire.psv_idx, &seg,
+						 &size);
+				if (err) {
+					mlx5_ib_warn(dev, "\n");
+					*bad_wr = wr;
+					goto out;
+				}
+
+				finish_wqe(qp, ctrl, size, idx, wr->wr_id,
+					   nreq, get_fence(fence, wr),
+					   next_fence, MLX5_OPCODE_SET_PSV);
+				num_sge = 0;
+				goto skip_psv;
+
 			default:
 				break;
 			}
@@ -2238,22 +2708,10 @@ int mlx5_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 			}
 		}
 
-		mlx5_opcode = mlx5_ib_opcode[wr->opcode];
-		ctrl->opmod_idx_opcode = cpu_to_be32(((u32)(qp->sq.cur_post) << 8)	|
-						     mlx5_opcode			|
-						     ((u32)opmod << 24));
-		ctrl->qpn_ds = cpu_to_be32(size | (qp->mqp.qpn << 8));
-		ctrl->fm_ce_se |= get_fence(fence, wr);
-		qp->fm_cache = next_fence;
-		if (unlikely(qp->wq_sig))
-			ctrl->signature = wq_sig(ctrl);
-
-		qp->sq.wrid[idx] = wr->wr_id;
-		qp->sq.w_list[idx].opcode = mlx5_opcode;
-		qp->sq.wqe_head[idx] = qp->sq.head + nreq;
-		qp->sq.cur_post += DIV_ROUND_UP(size * 16, MLX5_SEND_WQE_BB);
-		qp->sq.w_list[idx].next = qp->sq.cur_post;
-
+		finish_wqe(qp, ctrl, size, idx, wr->wr_id, nreq,
+			   get_fence(fence, wr), next_fence,
+			   mlx5_ib_opcode[wr->opcode]);
+skip_psv:
 		if (0)
 			dump_wqe(qp, idx, size);
 	}

drivers/infiniband/hw/mthca/mthca_provider.c

@@ -695,6 +695,7 @@ static struct ib_cq *mthca_create_cq(struct ib_device *ibdev, int entries,
 
 	if (context && ib_copy_to_udata(udata, &cq->cqn, sizeof (__u32))) {
 		mthca_free_cq(to_mdev(ibdev), cq);
+		err = -EFAULT;
 		goto err_free;
 	}
 
@@ -976,12 +977,12 @@ static struct ib_mr *mthca_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 				       u64 virt, int acc, struct ib_udata *udata)
 {
 	struct mthca_dev *dev = to_mdev(pd->device);
-	struct ib_umem_chunk *chunk;
+	struct scatterlist *sg;
 	struct mthca_mr *mr;
 	struct mthca_reg_mr ucmd;
 	u64 *pages;
 	int shift, n, len;
-	int i, j, k;
+	int i, k, entry;
 	int err = 0;
 	int write_mtt_size;
 
@@ -1009,10 +1010,7 @@ static struct ib_mr *mthca_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 	}
 
 	shift = ffs(mr->umem->page_size) - 1;
-
-	n = 0;
-	list_for_each_entry(chunk, &mr->umem->chunk_list, list)
-		n += chunk->nents;
+	n = mr->umem->nmap;
 
 	mr->mtt = mthca_alloc_mtt(dev, n);
 	if (IS_ERR(mr->mtt)) {
@@ -1030,25 +1028,24 @@ static struct ib_mr *mthca_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 
 	write_mtt_size = min(mthca_write_mtt_size(dev), (int) (PAGE_SIZE / sizeof *pages));
 
-	list_for_each_entry(chunk, &mr->umem->chunk_list, list)
-		for (j = 0; j < chunk->nmap; ++j) {
-			len = sg_dma_len(&chunk->page_list[j]) >> shift;
-			for (k = 0; k < len; ++k) {
-				pages[i++] = sg_dma_address(&chunk->page_list[j]) +
-					mr->umem->page_size * k;
-				/*
-				 * Be friendly to write_mtt and pass it chunks
-				 * of appropriate size.
-				 */
-				if (i == write_mtt_size) {
-					err = mthca_write_mtt(dev, mr->mtt, n, pages, i);
-					if (err)
-						goto mtt_done;
-					n += i;
-					i = 0;
-				}
+	for_each_sg(mr->umem->sg_head.sgl, sg, mr->umem->nmap, entry) {
+		len = sg_dma_len(sg) >> shift;
+		for (k = 0; k < len; ++k) {
+			pages[i++] = sg_dma_address(sg) +
+				mr->umem->page_size * k;
+			/*
+			 * Be friendly to write_mtt and pass it chunks
+			 * of appropriate size.
+			 */
+			if (i == write_mtt_size) {
+				err = mthca_write_mtt(dev, mr->mtt, n, pages, i);
+				if (err)
+					goto mtt_done;
+				n += i;
+				i = 0;
 			}
 		}
+	}
 
 	if (i)
 		err = mthca_write_mtt(dev, mr->mtt, n, pages, i);

drivers/infiniband/hw/nes/nes_cm.c

@@ -128,6 +128,7 @@ static void build_mpa_v1(struct nes_cm_node *, void *, u8);
 static void build_rdma0_msg(struct nes_cm_node *, struct nes_qp **);
 
 static void print_core(struct nes_cm_core *core);
+static void record_ird_ord(struct nes_cm_node *, u16, u16);
 
 /* External CM API Interface */
 /* instance of function pointers for client API */
@@ -317,7 +318,6 @@ static int parse_mpa(struct nes_cm_node *cm_node, u8 *buffer, u32 *type,
 		}
 	}
 
-
 	if (priv_data_len + mpa_hdr_len != len) {
 		nes_debug(NES_DBG_CM, "The received ietf buffer was not right"
 			" complete (%x + %x != %x)\n",
@@ -356,25 +356,57 @@ static int parse_mpa(struct nes_cm_node *cm_node, u8 *buffer, u32 *type,
 			/* send reset */
 			return -EINVAL;
 		}
+		if (ird_size == IETF_NO_IRD_ORD || ord_size == IETF_NO_IRD_ORD)
+			cm_node->mpav2_ird_ord = IETF_NO_IRD_ORD;
 
-		if (cm_node->state != NES_CM_STATE_MPAREQ_SENT) {
+		if (cm_node->mpav2_ird_ord != IETF_NO_IRD_ORD) {
 			/* responder */
-			if (cm_node->ord_size > ird_size)
-				cm_node->ord_size = ird_size;
-		} else {
-			/* initiator */
-			if (cm_node->ord_size > ird_size)
-				cm_node->ord_size = ird_size;
-
-			if (cm_node->ird_size < ord_size) {
-				/* no resources available */
-				/* send terminate message */
-				return -EINVAL;
+			if (cm_node->state != NES_CM_STATE_MPAREQ_SENT) {
+				/* we are still negotiating */
+				if (ord_size > NES_MAX_IRD) {
+					cm_node->ird_size = NES_MAX_IRD;
+				} else {
+					cm_node->ird_size = ord_size;
+					if (ord_size == 0 &&
+					(rtr_ctrl_ord & IETF_RDMA0_READ)) {
+						cm_node->ird_size = 1;
+						nes_debug(NES_DBG_CM,
+						"%s: Remote peer doesn't support RDMA0_READ (ord=%u)\n",
+							__func__, ord_size);
+					}
+				}
+				if (ird_size > NES_MAX_ORD)
+					cm_node->ord_size = NES_MAX_ORD;
+				else
+					cm_node->ord_size = ird_size;
+			} else { /* initiator */
+				if (ord_size > NES_MAX_IRD) {
+					nes_debug(NES_DBG_CM,
+					"%s: Unable to support the requested (ord =%u)\n",
+							__func__, ord_size);
+					return -EINVAL;
+				}
+				cm_node->ird_size = ord_size;
+
+				if (ird_size > NES_MAX_ORD) {
+					cm_node->ord_size = NES_MAX_ORD;
+				} else {
+					if (ird_size == 0 &&
+					(rtr_ctrl_ord & IETF_RDMA0_READ)) {
+						nes_debug(NES_DBG_CM,
+						"%s: Remote peer doesn't support RDMA0_READ (ird=%u)\n",
+							__func__, ird_size);
+						return -EINVAL;
+					} else {
+						cm_node->ord_size = ird_size;
+					}
+				}
 			}
 		}
 
 		if (rtr_ctrl_ord & IETF_RDMA0_READ) {
 			cm_node->send_rdma0_op = SEND_RDMA_READ_ZERO;
+
 		} else if (rtr_ctrl_ord & IETF_RDMA0_WRITE) {
 			cm_node->send_rdma0_op = SEND_RDMA_WRITE_ZERO;
 		} else {        /* Not supported RDMA0 operation */
@@ -514,6 +546,19 @@ static void print_core(struct nes_cm_core *core)
 	nes_debug(NES_DBG_CM, "-------------- end core ---------------\n");
 }
 
+static void record_ird_ord(struct nes_cm_node *cm_node,
+					u16 conn_ird, u16 conn_ord)
+{
+	if (conn_ird > NES_MAX_IRD)
+		conn_ird = NES_MAX_IRD;
+
+	if (conn_ord > NES_MAX_ORD)
+		conn_ord = NES_MAX_ORD;
+
+	cm_node->ird_size = conn_ird;
+	cm_node->ord_size = conn_ord;
+}
+
 /**
  * cm_build_mpa_frame - build a MPA V1 frame or MPA V2 frame
  */
@@ -557,11 +602,13 @@ static void build_mpa_v2(struct nes_cm_node *cm_node,
 	mpa_frame->priv_data_len += htons(IETF_RTR_MSG_SIZE);
 
 	/* initialize RTR msg */
-	ctrl_ird = (cm_node->ird_size > IETF_NO_IRD_ORD) ?
-			    IETF_NO_IRD_ORD : cm_node->ird_size;
-	ctrl_ord = (cm_node->ord_size > IETF_NO_IRD_ORD) ?
-			    IETF_NO_IRD_ORD : cm_node->ord_size;
-
+	if (cm_node->mpav2_ird_ord == IETF_NO_IRD_ORD) {
+		ctrl_ird = IETF_NO_IRD_ORD;
+		ctrl_ord = IETF_NO_IRD_ORD;
+	} else {
+		ctrl_ird = cm_node->ird_size & IETF_NO_IRD_ORD;
+		ctrl_ord = cm_node->ord_size & IETF_NO_IRD_ORD;
+	}
 	ctrl_ird |= IETF_PEER_TO_PEER;
 	ctrl_ird |= IETF_FLPDU_ZERO_LEN;
 
@@ -610,7 +657,7 @@ static void build_rdma0_msg(struct nes_cm_node *cm_node, struct nes_qp **nesqp_a
 	struct nes_qp *nesqp = *nesqp_addr;
 	struct nes_hw_qp_wqe *wqe = &nesqp->hwqp.sq_vbase[0];
 
-	u64temp = (unsigned long)nesqp;
+	u64temp = (unsigned long)nesqp->nesuqp_addr;
 	u64temp |= NES_SW_CONTEXT_ALIGN >> 1;
 	set_wqe_64bit_value(wqe->wqe_words, NES_IWARP_SQ_WQE_COMP_CTX_LOW_IDX, u64temp);
 
@@ -1409,8 +1456,9 @@ static struct nes_cm_node *make_cm_node(struct nes_cm_core *cm_core,
 
 	cm_node->mpa_frame_rev = mpa_version;
 	cm_node->send_rdma0_op = SEND_RDMA_READ_ZERO;
-	cm_node->ird_size = IETF_NO_IRD_ORD;
-	cm_node->ord_size = IETF_NO_IRD_ORD;
+	cm_node->mpav2_ird_ord = 0;
+	cm_node->ird_size = 0;
+	cm_node->ord_size = 0;
 
 	nes_debug(NES_DBG_CM, "Make node addresses : loc = %pI4:%x, rem = %pI4:%x\n",
 		  &cm_node->loc_addr, cm_node->loc_port,
@@ -3027,11 +3075,11 @@ int nes_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
 		rem_ref_cm_node(cm_node->cm_core, cm_node);
 		return -ECONNRESET;
 	}
-
 	/* associate the node with the QP */
 	nesqp->cm_node = (void *)cm_node;
 	cm_node->nesqp = nesqp;
 
+
 	nes_debug(NES_DBG_CM, "QP%u, cm_node=%p, jiffies = %lu listener = %p\n",
 		nesqp->hwqp.qp_id, cm_node, jiffies, cm_node->listener);
 	atomic_inc(&cm_accepts);
@@ -3054,6 +3102,11 @@ int nes_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
 	if (cm_node->mpa_frame_rev == IETF_MPA_V1)
 		mpa_frame_offset = 4;
 
+	if (cm_node->mpa_frame_rev == IETF_MPA_V1 ||
+			cm_node->mpav2_ird_ord == IETF_NO_IRD_ORD) {
+		record_ird_ord(cm_node, (u16)conn_param->ird, (u16)conn_param->ord);
+	}
+
 	memcpy(mpa_v2_frame->priv_data, conn_param->private_data,
 	       conn_param->private_data_len);
 
@@ -3117,7 +3170,6 @@ int nes_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
 	}
 	nesqp->skip_lsmm = 1;
 
-
 	/* Cache the cm_id in the qp */
 	nesqp->cm_id = cm_id;
 	cm_node->cm_id = cm_id;
@@ -3154,7 +3206,7 @@ int nes_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
 	nesqp->nesqp_context->ird_ord_sizes |= cpu_to_le32(
 		((u32)1 << NES_QPCONTEXT_ORDIRD_IWARP_MODE_SHIFT));
 	nesqp->nesqp_context->ird_ord_sizes |=
-		cpu_to_le32((u32)conn_param->ord);
+		cpu_to_le32((u32)cm_node->ord_size);
 
 	memset(&nes_quad, 0, sizeof(nes_quad));
 	nes_quad.DstIpAdrIndex =
@@ -3194,6 +3246,9 @@ int nes_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
 	cm_event.remote_addr = cm_id->remote_addr;
 	cm_event.private_data = NULL;
 	cm_event.private_data_len = 0;
+	cm_event.ird = cm_node->ird_size;
+	cm_event.ord = cm_node->ord_size;
+
 	ret = cm_id->event_handler(cm_id, &cm_event);
 	attr.qp_state = IB_QPS_RTS;
 	nes_modify_qp(&nesqp->ibqp, &attr, IB_QP_STATE, NULL);
@@ -3290,14 +3345,8 @@ int nes_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
 
 	/* cache the cm_id in the qp */
 	nesqp->cm_id = cm_id;
-
 	cm_id->provider_data = nesqp;
-
 	nesqp->private_data_len = conn_param->private_data_len;
-	nesqp->nesqp_context->ird_ord_sizes |= cpu_to_le32((u32)conn_param->ord);
-	/* space for rdma0 read msg */
-	if (conn_param->ord == 0)
-		nesqp->nesqp_context->ird_ord_sizes |= cpu_to_le32(1);
 
 	nes_debug(NES_DBG_CM, "requested ord = 0x%08X.\n", (u32)conn_param->ord);
 	nes_debug(NES_DBG_CM, "mpa private data len =%u\n",
@@ -3334,6 +3383,11 @@ int nes_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
 		return -ENOMEM;
 	}
 
+	record_ird_ord(cm_node, (u16)conn_param->ird, (u16)conn_param->ord);
+	if (cm_node->send_rdma0_op == SEND_RDMA_READ_ZERO &&
+				cm_node->ord_size == 0)
+		cm_node->ord_size = 1;
+
 	cm_node->apbvt_set = apbvt_set;
 	nesqp->cm_node = cm_node;
 	cm_node->nesqp = nesqp;
@@ -3530,6 +3584,8 @@ static void cm_event_connected(struct nes_cm_event *event)
 	nesqp->nesqp_context->ird_ord_sizes |=
 			cpu_to_le32((u32)1 <<
 			NES_QPCONTEXT_ORDIRD_IWARP_MODE_SHIFT);
+	nesqp->nesqp_context->ird_ord_sizes |=
+			cpu_to_le32((u32)cm_node->ord_size);
 
 	/* Adjust tail for not having a LSMM */
 	/*nesqp->hwqp.sq_tail = 1;*/
@@ -3742,8 +3798,13 @@ static void cm_event_mpa_req(struct nes_cm_event *event)
 	cm_event_raddr->sin_addr.s_addr = htonl(event->cm_info.rem_addr);
 	cm_event.private_data = cm_node->mpa_frame_buf;
 	cm_event.private_data_len = (u8)cm_node->mpa_frame_size;
+	if (cm_node->mpa_frame_rev == IETF_MPA_V1) {
+		cm_event.ird = NES_MAX_IRD;
+		cm_event.ord = NES_MAX_ORD;
+	} else {
 	cm_event.ird = cm_node->ird_size;
 	cm_event.ord = cm_node->ord_size;
+	}
 
 	ret = cm_id->event_handler(cm_id, &cm_event);
 	if (ret)

drivers/infiniband/hw/nes/nes_cm.h

@@ -58,6 +58,8 @@
 #define IETF_RDMA0_WRITE        0x8000
 #define IETF_RDMA0_READ         0x4000
 #define IETF_NO_IRD_ORD         0x3FFF
+#define NES_MAX_IRD		 0x40
+#define NES_MAX_ORD		 0x7F
 
 enum ietf_mpa_flags {
 	IETF_MPA_FLAGS_MARKERS = 0x80,	/* receive Markers */
@@ -333,6 +335,7 @@ struct nes_cm_node {
 	enum mpa_frame_version    mpa_frame_rev;
 	u16			  ird_size;
 	u16                       ord_size;
+	u16			  mpav2_ird_ord;
 
 	u16                       mpa_frame_size;
 	struct iw_cm_id           *cm_id;

drivers/infiniband/hw/nes/nes_user.h

@@ -39,8 +39,8 @@
 
 #include <linux/types.h>
 
-#define NES_ABI_USERSPACE_VER 1
-#define NES_ABI_KERNEL_VER    1
+#define NES_ABI_USERSPACE_VER 2
+#define NES_ABI_KERNEL_VER    2
 
 /*
  * Make sure that all structs defined in this file remain laid out so
@@ -78,6 +78,7 @@ struct nes_create_cq_req {
 
 struct nes_create_qp_req {
 	__u64 user_wqe_buffers;
+	__u64 user_qp_buffer;
 };
 
 enum iwnes_memreg_type {

drivers/infiniband/hw/nes/nes_verbs.c

@@ -1186,11 +1186,13 @@ static struct ib_qp *nes_create_qp(struct ib_pd *ibpd,
 					nes_free_resource(nesadapter, nesadapter->allocated_qps, qp_num);
 					kfree(nesqp->allocated_buffer);
 					nes_debug(NES_DBG_QP, "ib_copy_from_udata() Failed \n");
-					return NULL;
+					return ERR_PTR(-EFAULT);
 				}
 				if (req.user_wqe_buffers) {
 					virt_wqs = 1;
 				}
+				if (req.user_qp_buffer)
+					nesqp->nesuqp_addr = req.user_qp_buffer;
 				if ((ibpd->uobject) && (ibpd->uobject->context)) {
 					nesqp->user_mode = 1;
 					nes_ucontext = to_nesucontext(ibpd->uobject->context);
@@ -2307,7 +2309,7 @@ static struct ib_mr *nes_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 	struct nes_device *nesdev = nesvnic->nesdev;
 	struct nes_adapter *nesadapter = nesdev->nesadapter;
 	struct ib_mr *ibmr = ERR_PTR(-EINVAL);
-	struct ib_umem_chunk *chunk;
+	struct scatterlist *sg;
 	struct nes_ucontext *nes_ucontext;
 	struct nes_pbl *nespbl;
 	struct nes_mr *nesmr;
@@ -2315,7 +2317,7 @@ static struct ib_mr *nes_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 	struct nes_mem_reg_req req;
 	struct nes_vpbl vpbl;
 	struct nes_root_vpbl root_vpbl;
-	int nmap_index, page_index;
+	int entry, page_index;
 	int page_count = 0;
 	int err, pbl_depth = 0;
 	int chunk_pages;
@@ -2330,6 +2332,7 @@ static struct ib_mr *nes_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 	u16 pbl_count;
 	u8 single_page = 1;
 	u8 stag_key;
+	int first_page = 1;
 
 	region = ib_umem_get(pd->uobject->context, start, length, acc, 0);
 	if (IS_ERR(region)) {
@@ -2380,128 +2383,125 @@ static struct ib_mr *nes_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 			}
 			nesmr->region = region;
 
-			list_for_each_entry(chunk, &region->chunk_list, list) {
-				nes_debug(NES_DBG_MR, "Chunk: nents = %u, nmap = %u .\n",
-						chunk->nents, chunk->nmap);
-				for (nmap_index = 0; nmap_index < chunk->nmap; ++nmap_index) {
-					if (sg_dma_address(&chunk->page_list[nmap_index]) & ~PAGE_MASK) {
-						ib_umem_release(region);
-						nes_free_resource(nesadapter, nesadapter->allocated_mrs, stag_index);
-						nes_debug(NES_DBG_MR, "Unaligned Memory Buffer: 0x%x\n",
-								(unsigned int) sg_dma_address(&chunk->page_list[nmap_index]));
-						ibmr = ERR_PTR(-EINVAL);
-						kfree(nesmr);
-						goto reg_user_mr_err;
-					}
+			for_each_sg(region->sg_head.sgl, sg, region->nmap, entry) {
+				if (sg_dma_address(sg) & ~PAGE_MASK) {
+					ib_umem_release(region);
+					nes_free_resource(nesadapter, nesadapter->allocated_mrs, stag_index);
+					nes_debug(NES_DBG_MR, "Unaligned Memory Buffer: 0x%x\n",
+						  (unsigned int) sg_dma_address(sg));
+					ibmr = ERR_PTR(-EINVAL);
+					kfree(nesmr);
+					goto reg_user_mr_err;
+				}
 
-					if (!sg_dma_len(&chunk->page_list[nmap_index])) {
-						ib_umem_release(region);
-						nes_free_resource(nesadapter, nesadapter->allocated_mrs,
-								stag_index);
-						nes_debug(NES_DBG_MR, "Invalid Buffer Size\n");
-						ibmr = ERR_PTR(-EINVAL);
-						kfree(nesmr);
-						goto reg_user_mr_err;
-					}
+				if (!sg_dma_len(sg)) {
+					ib_umem_release(region);
+					nes_free_resource(nesadapter, nesadapter->allocated_mrs,
+							  stag_index);
+					nes_debug(NES_DBG_MR, "Invalid Buffer Size\n");
+					ibmr = ERR_PTR(-EINVAL);
+					kfree(nesmr);
+					goto reg_user_mr_err;
+				}
 
-					region_length += sg_dma_len(&chunk->page_list[nmap_index]);
-					chunk_pages = sg_dma_len(&chunk->page_list[nmap_index]) >> 12;
-					region_length -= skip_pages << 12;
-					for (page_index=skip_pages; page_index < chunk_pages; page_index++) {
-						skip_pages = 0;
-						if ((page_count!=0)&&(page_count<<12)-(region->offset&(4096-1))>=region->length)
-							goto enough_pages;
-						if ((page_count&0x01FF) == 0) {
-							if (page_count >= 1024 * 512) {
+				region_length += sg_dma_len(sg);
+				chunk_pages = sg_dma_len(sg) >> 12;
+				region_length -= skip_pages << 12;
+				for (page_index = skip_pages; page_index < chunk_pages; page_index++) {
+					skip_pages = 0;
+					if ((page_count != 0) && (page_count<<12)-(region->offset&(4096-1)) >= region->length)
+						goto enough_pages;
+					if ((page_count&0x01FF) == 0) {
+						if (page_count >= 1024 * 512) {
+							ib_umem_release(region);
+							nes_free_resource(nesadapter,
+									  nesadapter->allocated_mrs, stag_index);
+							kfree(nesmr);
+							ibmr = ERR_PTR(-E2BIG);
+							goto reg_user_mr_err;
+						}
+						if (root_pbl_index == 1) {
+							root_vpbl.pbl_vbase = pci_alloc_consistent(nesdev->pcidev,
+									8192, &root_vpbl.pbl_pbase);
+							nes_debug(NES_DBG_MR, "Allocating root PBL, va = %p, pa = 0x%08X\n",
+								  root_vpbl.pbl_vbase, (unsigned int)root_vpbl.pbl_pbase);
+							if (!root_vpbl.pbl_vbase) {
 								ib_umem_release(region);
-								nes_free_resource(nesadapter,
-										nesadapter->allocated_mrs, stag_index);
+								pci_free_consistent(nesdev->pcidev, 4096, vpbl.pbl_vbase,
+										    vpbl.pbl_pbase);
+								nes_free_resource(nesadapter, nesadapter->allocated_mrs,
+										  stag_index);
 								kfree(nesmr);
-								ibmr = ERR_PTR(-E2BIG);
+								ibmr = ERR_PTR(-ENOMEM);
 								goto reg_user_mr_err;
 							}
-							if (root_pbl_index == 1) {
-								root_vpbl.pbl_vbase = pci_alloc_consistent(nesdev->pcidev,
-										8192, &root_vpbl.pbl_pbase);
-								nes_debug(NES_DBG_MR, "Allocating root PBL, va = %p, pa = 0x%08X\n",
-										root_vpbl.pbl_vbase, (unsigned int)root_vpbl.pbl_pbase);
-								if (!root_vpbl.pbl_vbase) {
-									ib_umem_release(region);
-									pci_free_consistent(nesdev->pcidev, 4096, vpbl.pbl_vbase,
-											vpbl.pbl_pbase);
-									nes_free_resource(nesadapter, nesadapter->allocated_mrs,
-											stag_index);
-									kfree(nesmr);
-									ibmr = ERR_PTR(-ENOMEM);
-									goto reg_user_mr_err;
-								}
-								root_vpbl.leaf_vpbl = kzalloc(sizeof(*root_vpbl.leaf_vpbl)*1024,
-										GFP_KERNEL);
-								if (!root_vpbl.leaf_vpbl) {
-									ib_umem_release(region);
-									pci_free_consistent(nesdev->pcidev, 8192, root_vpbl.pbl_vbase,
-											root_vpbl.pbl_pbase);
-									pci_free_consistent(nesdev->pcidev, 4096, vpbl.pbl_vbase,
-											vpbl.pbl_pbase);
-									nes_free_resource(nesadapter, nesadapter->allocated_mrs,
-											stag_index);
-									kfree(nesmr);
-									ibmr = ERR_PTR(-ENOMEM);
-									goto reg_user_mr_err;
-								}
-								root_vpbl.pbl_vbase[0].pa_low =
-										cpu_to_le32((u32)vpbl.pbl_pbase);
-								root_vpbl.pbl_vbase[0].pa_high =
-										cpu_to_le32((u32)((((u64)vpbl.pbl_pbase) >> 32)));
-								root_vpbl.leaf_vpbl[0] = vpbl;
-							}
-							vpbl.pbl_vbase = pci_alloc_consistent(nesdev->pcidev, 4096,
-									&vpbl.pbl_pbase);
-							nes_debug(NES_DBG_MR, "Allocating leaf PBL, va = %p, pa = 0x%08X\n",
-									vpbl.pbl_vbase, (unsigned int)vpbl.pbl_pbase);
-							if (!vpbl.pbl_vbase) {
+							root_vpbl.leaf_vpbl = kzalloc(sizeof(*root_vpbl.leaf_vpbl)*1024,
+									GFP_KERNEL);
+							if (!root_vpbl.leaf_vpbl) {
 								ib_umem_release(region);
-								nes_free_resource(nesadapter, nesadapter->allocated_mrs, stag_index);
-								ibmr = ERR_PTR(-ENOMEM);
+								pci_free_consistent(nesdev->pcidev, 8192, root_vpbl.pbl_vbase,
+										    root_vpbl.pbl_pbase);
+								pci_free_consistent(nesdev->pcidev, 4096, vpbl.pbl_vbase,
+										    vpbl.pbl_pbase);
+								nes_free_resource(nesadapter, nesadapter->allocated_mrs,
+										  stag_index);
 								kfree(nesmr);
+								ibmr = ERR_PTR(-ENOMEM);
 								goto reg_user_mr_err;
 							}
-							if (1 <= root_pbl_index) {
-								root_vpbl.pbl_vbase[root_pbl_index].pa_low =
-										cpu_to_le32((u32)vpbl.pbl_pbase);
-								root_vpbl.pbl_vbase[root_pbl_index].pa_high =
-										cpu_to_le32((u32)((((u64)vpbl.pbl_pbase)>>32)));
-								root_vpbl.leaf_vpbl[root_pbl_index] = vpbl;
-							}
-							root_pbl_index++;
-							cur_pbl_index = 0;
+							root_vpbl.pbl_vbase[0].pa_low =
+									cpu_to_le32((u32)vpbl.pbl_pbase);
+							root_vpbl.pbl_vbase[0].pa_high =
+									cpu_to_le32((u32)((((u64)vpbl.pbl_pbase) >> 32)));
+							root_vpbl.leaf_vpbl[0] = vpbl;
 						}
-						if (single_page) {
-							if (page_count != 0) {
-								if ((last_dma_addr+4096) !=
-										(sg_dma_address(&chunk->page_list[nmap_index])+
-										(page_index*4096)))
-									single_page = 0;
-								last_dma_addr = sg_dma_address(&chunk->page_list[nmap_index])+
-										(page_index*4096);
-							} else {
-								first_dma_addr = sg_dma_address(&chunk->page_list[nmap_index])+
-										(page_index*4096);
-								last_dma_addr = first_dma_addr;
-							}
+						vpbl.pbl_vbase = pci_alloc_consistent(nesdev->pcidev, 4096,
+								&vpbl.pbl_pbase);
+						nes_debug(NES_DBG_MR, "Allocating leaf PBL, va = %p, pa = 0x%08X\n",
+							  vpbl.pbl_vbase, (unsigned int)vpbl.pbl_pbase);
+						if (!vpbl.pbl_vbase) {
+							ib_umem_release(region);
+							nes_free_resource(nesadapter, nesadapter->allocated_mrs, stag_index);
+							ibmr = ERR_PTR(-ENOMEM);
+							kfree(nesmr);
+							goto reg_user_mr_err;
+						}
+						if (1 <= root_pbl_index) {
+							root_vpbl.pbl_vbase[root_pbl_index].pa_low =
+									cpu_to_le32((u32)vpbl.pbl_pbase);
+							root_vpbl.pbl_vbase[root_pbl_index].pa_high =
+									cpu_to_le32((u32)((((u64)vpbl.pbl_pbase)>>32)));
+							root_vpbl.leaf_vpbl[root_pbl_index] = vpbl;
+						}
+						root_pbl_index++;
+						cur_pbl_index = 0;
+					}
+					if (single_page) {
+						if (page_count != 0) {
+							if ((last_dma_addr+4096) !=
+									(sg_dma_address(sg)+
+									(page_index*4096)))
+								single_page = 0;
+							last_dma_addr = sg_dma_address(sg)+
+									(page_index*4096);
+						} else {
+							first_dma_addr = sg_dma_address(sg)+
+									(page_index*4096);
+							last_dma_addr = first_dma_addr;
 						}
-
-						vpbl.pbl_vbase[cur_pbl_index].pa_low =
-								cpu_to_le32((u32)(sg_dma_address(&chunk->page_list[nmap_index])+
-								(page_index*4096)));
-						vpbl.pbl_vbase[cur_pbl_index].pa_high =
-								cpu_to_le32((u32)((((u64)(sg_dma_address(&chunk->page_list[nmap_index])+
-								(page_index*4096))) >> 32)));
-						cur_pbl_index++;
-						page_count++;
 					}
+
+					vpbl.pbl_vbase[cur_pbl_index].pa_low =
+							cpu_to_le32((u32)(sg_dma_address(sg)+
+							(page_index*4096)));
+					vpbl.pbl_vbase[cur_pbl_index].pa_high =
+							cpu_to_le32((u32)((((u64)(sg_dma_address(sg)+
+							(page_index*4096))) >> 32)));
+					cur_pbl_index++;
+					page_count++;
 				}
 			}
+
 			enough_pages:
 			nes_debug(NES_DBG_MR, "calculating stag, stag_index=0x%08x, driver_key=0x%08x,"
 					" stag_key=0x%08x\n",
@@ -2613,25 +2613,28 @@ static struct ib_mr *nes_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 				  nespbl->pbl_size, (unsigned long) nespbl->pbl_pbase,
 				  (void *) nespbl->pbl_vbase, nespbl->user_base);
 
-			list_for_each_entry(chunk, &region->chunk_list, list) {
-				for (nmap_index = 0; nmap_index < chunk->nmap; ++nmap_index) {
-					chunk_pages = sg_dma_len(&chunk->page_list[nmap_index]) >> 12;
-					chunk_pages += (sg_dma_len(&chunk->page_list[nmap_index]) & (4096-1)) ? 1 : 0;
-					nespbl->page = sg_page(&chunk->page_list[0]);
-					for (page_index=0; page_index<chunk_pages; page_index++) {
-						((__le32 *)pbl)[0] = cpu_to_le32((u32)
-								(sg_dma_address(&chunk->page_list[nmap_index])+
-								(page_index*4096)));
-						((__le32 *)pbl)[1] = cpu_to_le32(((u64)
-								(sg_dma_address(&chunk->page_list[nmap_index])+
-								(page_index*4096)))>>32);
-						nes_debug(NES_DBG_MR, "pbl=%p, *pbl=0x%016llx, 0x%08x%08x\n", pbl,
-								(unsigned long long)*pbl,
-								le32_to_cpu(((__le32 *)pbl)[1]), le32_to_cpu(((__le32 *)pbl)[0]));
-						pbl++;
-					}
+			for_each_sg(region->sg_head.sgl, sg, region->nmap, entry) {
+				chunk_pages = sg_dma_len(sg) >> 12;
+				chunk_pages += (sg_dma_len(sg) & (4096-1)) ? 1 : 0;
+				if (first_page) {
+					nespbl->page = sg_page(sg);
+					first_page = 0;
+				}
+
+				for (page_index = 0; page_index < chunk_pages; page_index++) {
+					((__le32 *)pbl)[0] = cpu_to_le32((u32)
+							(sg_dma_address(sg)+
+							(page_index*4096)));
+					((__le32 *)pbl)[1] = cpu_to_le32(((u64)
+							(sg_dma_address(sg)+
+							(page_index*4096)))>>32);
+					nes_debug(NES_DBG_MR, "pbl=%p, *pbl=0x%016llx, 0x%08x%08x\n", pbl,
+						  (unsigned long long)*pbl,
+						  le32_to_cpu(((__le32 *)pbl)[1]), le32_to_cpu(((__le32 *)pbl)[0]));
+					pbl++;
 				}
 			}
+
 			if (req.reg_type == IWNES_MEMREG_TYPE_QP) {
 				list_add_tail(&nespbl->list, &nes_ucontext->qp_reg_mem_list);
 			} else {
@@ -3134,9 +3137,7 @@ int nes_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
 				" original_last_aeq = 0x%04X. last_aeq = 0x%04X.\n",
 				nesqp->hwqp.qp_id, atomic_read(&nesqp->refcount),
 				original_last_aeq, nesqp->last_aeq);
-		if ((!ret) ||
-				((original_last_aeq != NES_AEQE_AEID_RDMAP_ROE_BAD_LLP_CLOSE) &&
-				(ret))) {
+		if (!ret || original_last_aeq != NES_AEQE_AEID_RDMAP_ROE_BAD_LLP_CLOSE) {
 			if (dont_wait) {
 				if (nesqp->cm_id && nesqp->hw_tcp_state != 0) {
 					nes_debug(NES_DBG_MOD_QP, "QP%u Queuing fake disconnect for QP refcount (%d),"

drivers/infiniband/hw/nes/nes_verbs.h

@@ -184,5 +184,6 @@ struct nes_qp {
 	u8                    pau_busy;
 	u8                    pau_pending;
 	u8                    pau_state;
+	__u64                 nesuqp_addr;
 };
 #endif			/* NES_VERBS_H */

drivers/infiniband/hw/ocrdma/Makefile

@@ -2,4 +2,4 @@ ccflags-y := -Idrivers/net/ethernet/emulex/benet
 
 obj-$(CONFIG_INFINIBAND_OCRDMA)	+= ocrdma.o
 
-ocrdma-y :=	ocrdma_main.o ocrdma_verbs.o ocrdma_hw.o ocrdma_ah.o
+ocrdma-y :=	ocrdma_main.o ocrdma_verbs.o ocrdma_hw.o ocrdma_ah.o ocrdma_stats.o

drivers/infiniband/hw/ocrdma/ocrdma.h

@@ -35,17 +35,27 @@
 
 #include <rdma/ib_verbs.h>
 #include <rdma/ib_user_verbs.h>
+#include <rdma/ib_addr.h>
 
 #include <be_roce.h>
 #include "ocrdma_sli.h"
 
-#define OCRDMA_ROCE_DEV_VERSION "1.0.0"
+#define OCRDMA_ROCE_DRV_VERSION "10.2.145.0u"
+
+#define OCRDMA_ROCE_DRV_DESC "Emulex OneConnect RoCE Driver"
 #define OCRDMA_NODE_DESC "Emulex OneConnect RoCE HCA"
 
+#define OC_NAME_SH	OCRDMA_NODE_DESC "(Skyhawk)"
+#define OC_NAME_UNKNOWN OCRDMA_NODE_DESC "(Unknown)"
+
+#define OC_SKH_DEVICE_PF 0x720
+#define OC_SKH_DEVICE_VF 0x728
 #define OCRDMA_MAX_AH 512
 
 #define OCRDMA_UVERBS(CMD_NAME) (1ull << IB_USER_VERBS_CMD_##CMD_NAME)
 
+#define convert_to_64bit(lo, hi) ((u64)hi << 32 | (u64)lo)
+
 struct ocrdma_dev_attr {
 	u8 fw_ver[32];
 	u32 vendor_id;
@@ -65,6 +75,7 @@ struct ocrdma_dev_attr {
 	int max_mr;
 	u64 max_mr_size;
 	u32 max_num_mr_pbl;
+	int max_mw;
 	int max_fmr;
 	int max_map_per_fmr;
 	int max_pages_per_frmr;
@@ -83,6 +94,12 @@ struct ocrdma_dev_attr {
 	u8 num_ird_pages;
 };
 
+struct ocrdma_dma_mem {
+	void *va;
+	dma_addr_t pa;
+	u32 size;
+};
+
 struct ocrdma_pbl {
 	void *va;
 	dma_addr_t pa;
@@ -148,6 +165,26 @@ struct ocrdma_mr {
 	struct ocrdma_hw_mr hwmr;
 };
 
+struct ocrdma_stats {
+	u8 type;
+	struct ocrdma_dev *dev;
+};
+
+struct stats_mem {
+	struct ocrdma_mqe mqe;
+	void *va;
+	dma_addr_t pa;
+	u32 size;
+	char *debugfs_mem;
+};
+
+struct phy_info {
+	u16 auto_speeds_supported;
+	u16 fixed_speeds_supported;
+	u16 phy_type;
+	u16 interface_type;
+};
+
 struct ocrdma_dev {
 	struct ib_device ibdev;
 	struct ocrdma_dev_attr attr;
@@ -191,12 +228,30 @@ struct ocrdma_dev {
 	struct mqe_ctx mqe_ctx;
 
 	struct be_dev_info nic_info;
+	struct phy_info phy;
+	char model_number[32];
+	u32 hba_port_num;
 
 	struct list_head entry;
 	struct rcu_head rcu;
 	int id;
-	struct ocrdma_mr *stag_arr[OCRDMA_MAX_STAG];
+	u64 stag_arr[OCRDMA_MAX_STAG];
 	u16 pvid;
+	u32 asic_id;
+
+	ulong last_stats_time;
+	struct mutex stats_lock; /* provide synch for debugfs operations */
+	struct stats_mem stats_mem;
+	struct ocrdma_stats rsrc_stats;
+	struct ocrdma_stats rx_stats;
+	struct ocrdma_stats wqe_stats;
+	struct ocrdma_stats tx_stats;
+	struct ocrdma_stats db_err_stats;
+	struct ocrdma_stats tx_qp_err_stats;
+	struct ocrdma_stats rx_qp_err_stats;
+	struct ocrdma_stats tx_dbg_stats;
+	struct ocrdma_stats rx_dbg_stats;
+	struct dentry *dir;
 };
 
 struct ocrdma_cq {
@@ -209,8 +264,8 @@ struct ocrdma_cq {
 			 */
 	u32 max_hw_cqe;
 	bool phase_change;
-	bool armed, solicited;
-	bool arm_needed;
+	bool deferred_arm, deferred_sol;
+	bool first_arm;
 
 	spinlock_t cq_lock ____cacheline_aligned; /* provide synchronization
 						   * to cq polling
@@ -223,6 +278,7 @@ struct ocrdma_cq {
 	struct ocrdma_ucontext *ucontext;
 	dma_addr_t pa;
 	u32 len;
+	u32 cqe_cnt;
 
 	/* head of all qp's sq and rq for which cqes need to be flushed
 	 * by the software.
@@ -232,7 +288,6 @@ struct ocrdma_cq {
 
 struct ocrdma_pd {
 	struct ib_pd ibpd;
-	struct ocrdma_dev *dev;
 	struct ocrdma_ucontext *uctx;
 	u32 id;
 	int num_dpp_qp;
@@ -317,10 +372,8 @@ struct ocrdma_qp {
 	bool dpp_enabled;
 	u8 *ird_q_va;
 	bool signaled;
-	u16 db_cache;
 };
 
-
 struct ocrdma_ucontext {
 	struct ib_ucontext ibucontext;
 
@@ -385,13 +438,6 @@ static inline struct ocrdma_srq *get_ocrdma_srq(struct ib_srq *ibsrq)
 	return container_of(ibsrq, struct ocrdma_srq, ibsrq);
 }
 
-
-static inline int ocrdma_get_num_posted_shift(struct ocrdma_qp *qp)
-{
-	return ((qp->dev->nic_info.dev_family == OCRDMA_GEN2_FAMILY &&
-		 qp->id < 128) ? 24 : 16);
-}
-
 static inline int is_cqe_valid(struct ocrdma_cq *cq, struct ocrdma_cqe *cqe)
 {
 	int cqe_valid;
@@ -436,4 +482,40 @@ static inline int ocrdma_resolve_dmac(struct ocrdma_dev *dev,
 	return 0;
 }
 
+static inline char *hca_name(struct ocrdma_dev *dev)
+{
+	switch (dev->nic_info.pdev->device) {
+	case OC_SKH_DEVICE_PF:
+	case OC_SKH_DEVICE_VF:
+		return OC_NAME_SH;
+	default:
+		return OC_NAME_UNKNOWN;
+	}
+}
+
+static inline int ocrdma_get_eq_table_index(struct ocrdma_dev *dev,
+		int eqid)
+{
+	int indx;
+
+	for (indx = 0; indx < dev->eq_cnt; indx++) {
+		if (dev->eq_tbl[indx].q.id == eqid)
+			return indx;
+	}
+
+	return -EINVAL;
+}
+
+static inline u8 ocrdma_get_asic_type(struct ocrdma_dev *dev)
+{
+	if (dev->nic_info.dev_family == 0xF && !dev->asic_id) {
+		pci_read_config_dword(
+			dev->nic_info.pdev,
+			OCRDMA_SLI_ASIC_ID_OFFSET, &dev->asic_id);
+	}
+
+	return (dev->asic_id & OCRDMA_SLI_ASIC_GEN_NUM_MASK) >>
+				OCRDMA_SLI_ASIC_GEN_NUM_SHIFT;
+}
+
 #endif

drivers/infiniband/hw/ocrdma/ocrdma_abi.h

@@ -28,7 +28,8 @@
 #ifndef __OCRDMA_ABI_H__
 #define __OCRDMA_ABI_H__
 
-#define OCRDMA_ABI_VERSION 1
+#define OCRDMA_ABI_VERSION 2
+#define OCRDMA_BE_ROCE_ABI_VERSION 1
 /* user kernel communication data structures. */
 
 struct ocrdma_alloc_ucontext_resp {
@@ -107,9 +108,7 @@ struct ocrdma_create_qp_uresp {
 	u32 db_sq_offset;
 	u32 db_rq_offset;
 	u32 db_shift;
-	u64 rsvd1;
-	u64 rsvd2;
-	u64 rsvd3;
+	u64 rsvd[11];
 } __packed;
 
 struct ocrdma_create_srq_uresp {

drivers/infiniband/hw/ocrdma/ocrdma_ah.c

@@ -100,7 +100,7 @@ struct ib_ah *ocrdma_create_ah(struct ib_pd *ibpd, struct ib_ah_attr *attr)
 	if (!(attr->ah_flags & IB_AH_GRH))
 		return ERR_PTR(-EINVAL);
 
-	ah = kzalloc(sizeof *ah, GFP_ATOMIC);
+	ah = kzalloc(sizeof(*ah), GFP_ATOMIC);
 	if (!ah)
 		return ERR_PTR(-ENOMEM);
 

drivers/infiniband/hw/ocrdma/ocrdma_hw.c

@@ -32,7 +32,6 @@
 
 #include <rdma/ib_verbs.h>
 #include <rdma/ib_user_verbs.h>
-#include <rdma/ib_addr.h>
 
 #include "ocrdma.h"
 #include "ocrdma_hw.h"
@@ -243,6 +242,23 @@ static int ocrdma_get_mbx_errno(u32 status)
 	return err_num;
 }
 
+char *port_speed_string(struct ocrdma_dev *dev)
+{
+	char *str = "";
+	u16 speeds_supported;
+
+	speeds_supported = dev->phy.fixed_speeds_supported |
+				dev->phy.auto_speeds_supported;
+	if (speeds_supported & OCRDMA_PHY_SPEED_40GBPS)
+		str = "40Gbps ";
+	else if (speeds_supported & OCRDMA_PHY_SPEED_10GBPS)
+		str = "10Gbps ";
+	else if (speeds_supported & OCRDMA_PHY_SPEED_1GBPS)
+		str = "1Gbps ";
+
+	return str;
+}
+
 static int ocrdma_get_mbx_cqe_errno(u16 cqe_status)
 {
 	int err_num = -EINVAL;
@@ -332,6 +348,11 @@ static void *ocrdma_init_emb_mqe(u8 opcode, u32 cmd_len)
 	return mqe;
 }
 
+static void *ocrdma_alloc_mqe(void)
+{
+	return kzalloc(sizeof(struct ocrdma_mqe), GFP_KERNEL);
+}
+
 static void ocrdma_free_q(struct ocrdma_dev *dev, struct ocrdma_queue_info *q)
 {
 	dma_free_coherent(&dev->nic_info.pdev->dev, q->size, q->va, q->dma);
@@ -364,8 +385,8 @@ static void ocrdma_build_q_pages(struct ocrdma_pa *q_pa, int cnt,
 	}
 }
 
-static int ocrdma_mbx_delete_q(struct ocrdma_dev *dev, struct ocrdma_queue_info *q,
-			       int queue_type)
+static int ocrdma_mbx_delete_q(struct ocrdma_dev *dev,
+			       struct ocrdma_queue_info *q, int queue_type)
 {
 	u8 opcode = 0;
 	int status;
@@ -444,7 +465,7 @@ mbx_err:
 	return status;
 }
 
-static int ocrdma_get_irq(struct ocrdma_dev *dev, struct ocrdma_eq *eq)
+int ocrdma_get_irq(struct ocrdma_dev *dev, struct ocrdma_eq *eq)
 {
 	int irq;
 
@@ -574,6 +595,7 @@ static int ocrdma_create_mq(struct ocrdma_dev *dev)
 	if (status)
 		goto alloc_err;
 
+	dev->eq_tbl[0].cq_cnt++;
 	status = ocrdma_mbx_mq_cq_create(dev, &dev->mq.cq, &dev->eq_tbl[0].q);
 	if (status)
 		goto mbx_cq_free;
@@ -639,7 +661,7 @@ static void ocrdma_dispatch_ibevent(struct ocrdma_dev *dev,
 {
 	struct ocrdma_qp *qp = NULL;
 	struct ocrdma_cq *cq = NULL;
-	struct ib_event ib_evt;
+	struct ib_event ib_evt = { 0 };
 	int cq_event = 0;
 	int qp_event = 1;
 	int srq_event = 0;
@@ -664,6 +686,8 @@ static void ocrdma_dispatch_ibevent(struct ocrdma_dev *dev,
 	case OCRDMA_CQ_OVERRUN_ERROR:
 		ib_evt.element.cq = &cq->ibcq;
 		ib_evt.event = IB_EVENT_CQ_ERR;
+		cq_event = 1;
+		qp_event = 0;
 		break;
 	case OCRDMA_CQ_QPCAT_ERROR:
 		ib_evt.element.qp = &qp->ibqp;
@@ -725,6 +749,7 @@ static void ocrdma_dispatch_ibevent(struct ocrdma_dev *dev,
 						     qp->srq->ibsrq.
 						     srq_context);
 	} else if (dev_event) {
+		pr_err("%s: Fatal event received\n", dev->ibdev.name);
 		ib_dispatch_event(&ib_evt);
 	}
 
@@ -752,7 +777,6 @@ static void ocrdma_process_grp5_aync(struct ocrdma_dev *dev,
 	}
 }
 
-
 static void ocrdma_process_acqe(struct ocrdma_dev *dev, void *ae_cqe)
 {
 	/* async CQE processing */
@@ -799,8 +823,6 @@ static int ocrdma_mq_cq_handler(struct ocrdma_dev *dev, u16 cq_id)
 			ocrdma_process_acqe(dev, cqe);
 		else if (cqe->valid_ae_cmpl_cons & OCRDMA_MCQE_CMPL_MASK)
 			ocrdma_process_mcqe(dev, cqe);
-		else
-			pr_err("%s() cqe->compl is not set.\n", __func__);
 		memset(cqe, 0, sizeof(struct ocrdma_mcqe));
 		ocrdma_mcq_inc_tail(dev);
 	}
@@ -858,16 +880,8 @@ static void ocrdma_qp_cq_handler(struct ocrdma_dev *dev, u16 cq_idx)
 		BUG();
 
 	cq = dev->cq_tbl[cq_idx];
-	if (cq == NULL) {
-		pr_err("%s%d invalid id=0x%x\n", __func__, dev->id, cq_idx);
+	if (cq == NULL)
 		return;
-	}
-	spin_lock_irqsave(&cq->cq_lock, flags);
-	cq->armed = false;
-	cq->solicited = false;
-	spin_unlock_irqrestore(&cq->cq_lock, flags);
-
-	ocrdma_ring_cq_db(dev, cq->id, false, false, 0);
 
 	if (cq->ibcq.comp_handler) {
 		spin_lock_irqsave(&cq->comp_handler_lock, flags);
@@ -892,27 +906,35 @@ static irqreturn_t ocrdma_irq_handler(int irq, void *handle)
 	struct ocrdma_dev *dev = eq->dev;
 	struct ocrdma_eqe eqe;
 	struct ocrdma_eqe *ptr;
-	u16 eqe_popped = 0;
 	u16 cq_id;
-	while (1) {
+	int budget = eq->cq_cnt;
+
+	do {
 		ptr = ocrdma_get_eqe(eq);
 		eqe = *ptr;
 		ocrdma_le32_to_cpu(&eqe, sizeof(eqe));
 		if ((eqe.id_valid & OCRDMA_EQE_VALID_MASK) == 0)
 			break;
-		eqe_popped += 1;
+
 		ptr->id_valid = 0;
+		/* ring eq doorbell as soon as its consumed. */
+		ocrdma_ring_eq_db(dev, eq->q.id, false, true, 1);
 		/* check whether its CQE or not. */
 		if ((eqe.id_valid & OCRDMA_EQE_FOR_CQE_MASK) == 0) {
 			cq_id = eqe.id_valid >> OCRDMA_EQE_RESOURCE_ID_SHIFT;
 			ocrdma_cq_handler(dev, cq_id);
 		}
 		ocrdma_eq_inc_tail(eq);
-	}
-	ocrdma_ring_eq_db(dev, eq->q.id, true, true, eqe_popped);
-	/* Ring EQ doorbell with num_popped to 0 to enable interrupts again. */
-	if (dev->nic_info.intr_mode == BE_INTERRUPT_MODE_INTX)
-		ocrdma_ring_eq_db(dev, eq->q.id, true, true, 0);
+
+		/* There can be a stale EQE after the last bound CQ is
+		 * destroyed. EQE valid and budget == 0 implies this.
+		 */
+		if (budget)
+			budget--;
+
+	} while (budget);
+
+	ocrdma_ring_eq_db(dev, eq->q.id, true, true, 0);
 	return IRQ_HANDLED;
 }
 
@@ -949,7 +971,8 @@ static int ocrdma_mbx_cmd(struct ocrdma_dev *dev, struct ocrdma_mqe *mqe)
 {
 	int status = 0;
 	u16 cqe_status, ext_status;
-	struct ocrdma_mqe *rsp;
+	struct ocrdma_mqe *rsp_mqe;
+	struct ocrdma_mbx_rsp *rsp = NULL;
 
 	mutex_lock(&dev->mqe_ctx.lock);
 	ocrdma_post_mqe(dev, mqe);
@@ -958,23 +981,61 @@ static int ocrdma_mbx_cmd(struct ocrdma_dev *dev, struct ocrdma_mqe *mqe)
 		goto mbx_err;
 	cqe_status = dev->mqe_ctx.cqe_status;
 	ext_status = dev->mqe_ctx.ext_status;
-	rsp = ocrdma_get_mqe_rsp(dev);
-	ocrdma_copy_le32_to_cpu(mqe, rsp, (sizeof(*mqe)));
+	rsp_mqe = ocrdma_get_mqe_rsp(dev);
+	ocrdma_copy_le32_to_cpu(mqe, rsp_mqe, (sizeof(*mqe)));
+	if ((mqe->hdr.spcl_sge_cnt_emb & OCRDMA_MQE_HDR_EMB_MASK) >>
+				OCRDMA_MQE_HDR_EMB_SHIFT)
+		rsp = &mqe->u.rsp;
+
 	if (cqe_status || ext_status) {
-		pr_err("%s() opcode=0x%x, cqe_status=0x%x, ext_status=0x%x\n",
-		       __func__,
-		     (rsp->u.rsp.subsys_op & OCRDMA_MBX_RSP_OPCODE_MASK) >>
-		     OCRDMA_MBX_RSP_OPCODE_SHIFT, cqe_status, ext_status);
+		pr_err("%s() cqe_status=0x%x, ext_status=0x%x,",
+		       __func__, cqe_status, ext_status);
+		if (rsp) {
+			/* This is for embedded cmds. */
+			pr_err("opcode=0x%x, subsystem=0x%x\n",
+			       (rsp->subsys_op & OCRDMA_MBX_RSP_OPCODE_MASK) >>
+				OCRDMA_MBX_RSP_OPCODE_SHIFT,
+				(rsp->subsys_op & OCRDMA_MBX_RSP_SUBSYS_MASK) >>
+				OCRDMA_MBX_RSP_SUBSYS_SHIFT);
+		}
 		status = ocrdma_get_mbx_cqe_errno(cqe_status);
 		goto mbx_err;
 	}
-	if (mqe->u.rsp.status & OCRDMA_MBX_RSP_STATUS_MASK)
+	/* For non embedded, rsp errors are handled in ocrdma_nonemb_mbx_cmd */
+	if (rsp && (mqe->u.rsp.status & OCRDMA_MBX_RSP_STATUS_MASK))
 		status = ocrdma_get_mbx_errno(mqe->u.rsp.status);
 mbx_err:
 	mutex_unlock(&dev->mqe_ctx.lock);
 	return status;
 }
 
+static int ocrdma_nonemb_mbx_cmd(struct ocrdma_dev *dev, struct ocrdma_mqe *mqe,
+				 void *payload_va)
+{
+	int status = 0;
+	struct ocrdma_mbx_rsp *rsp = payload_va;
+
+	if ((mqe->hdr.spcl_sge_cnt_emb & OCRDMA_MQE_HDR_EMB_MASK) >>
+				OCRDMA_MQE_HDR_EMB_SHIFT)
+		BUG();
+
+	status = ocrdma_mbx_cmd(dev, mqe);
+	if (!status)
+		/* For non embedded, only CQE failures are handled in
+		 * ocrdma_mbx_cmd. We need to check for RSP errors.
+		 */
+		if (rsp->status & OCRDMA_MBX_RSP_STATUS_MASK)
+			status = ocrdma_get_mbx_errno(rsp->status);
+
+	if (status)
+		pr_err("opcode=0x%x, subsystem=0x%x\n",
+		       (rsp->subsys_op & OCRDMA_MBX_RSP_OPCODE_MASK) >>
+			OCRDMA_MBX_RSP_OPCODE_SHIFT,
+			(rsp->subsys_op & OCRDMA_MBX_RSP_SUBSYS_MASK) >>
+			OCRDMA_MBX_RSP_SUBSYS_SHIFT);
+	return status;
+}
+
 static void ocrdma_get_attr(struct ocrdma_dev *dev,
 			      struct ocrdma_dev_attr *attr,
 			      struct ocrdma_mbx_query_config *rsp)
@@ -985,6 +1046,9 @@ static void ocrdma_get_attr(struct ocrdma_dev *dev,
 	attr->max_qp =
 	    (rsp->qp_srq_cq_ird_ord & OCRDMA_MBX_QUERY_CFG_MAX_QP_MASK) >>
 	    OCRDMA_MBX_QUERY_CFG_MAX_QP_SHIFT;
+	attr->max_srq =
+		(rsp->max_srq_rpir_qps & OCRDMA_MBX_QUERY_CFG_MAX_SRQ_MASK) >>
+		OCRDMA_MBX_QUERY_CFG_MAX_SRQ_OFFSET;
 	attr->max_send_sge = ((rsp->max_write_send_sge &
 			       OCRDMA_MBX_QUERY_CFG_MAX_SEND_SGE_MASK) >>
 			      OCRDMA_MBX_QUERY_CFG_MAX_SEND_SGE_SHIFT);
@@ -1000,9 +1064,6 @@ static void ocrdma_get_attr(struct ocrdma_dev *dev,
 	attr->max_ord_per_qp = (rsp->max_ird_ord_per_qp &
 				OCRDMA_MBX_QUERY_CFG_MAX_ORD_PER_QP_MASK) >>
 	    OCRDMA_MBX_QUERY_CFG_MAX_ORD_PER_QP_SHIFT;
-	attr->max_srq =
-		(rsp->max_srq_rpir_qps & OCRDMA_MBX_QUERY_CFG_MAX_SRQ_MASK) >>
-		OCRDMA_MBX_QUERY_CFG_MAX_SRQ_OFFSET;
 	attr->max_ird_per_qp = (rsp->max_ird_ord_per_qp &
 				OCRDMA_MBX_QUERY_CFG_MAX_IRD_PER_QP_MASK) >>
 	    OCRDMA_MBX_QUERY_CFG_MAX_IRD_PER_QP_SHIFT;
@@ -1015,6 +1076,7 @@ static void ocrdma_get_attr(struct ocrdma_dev *dev,
 	attr->local_ca_ack_delay = (rsp->max_pd_ca_ack_delay &
 				    OCRDMA_MBX_QUERY_CFG_CA_ACK_DELAY_MASK) >>
 	    OCRDMA_MBX_QUERY_CFG_CA_ACK_DELAY_SHIFT;
+	attr->max_mw = rsp->max_mw;
 	attr->max_mr = rsp->max_mr;
 	attr->max_mr_size = ~0ull;
 	attr->max_fmr = 0;
@@ -1036,7 +1098,7 @@ static void ocrdma_get_attr(struct ocrdma_dev *dev,
 	attr->max_inline_data =
 	    attr->wqe_size - (sizeof(struct ocrdma_hdr_wqe) +
 			      sizeof(struct ocrdma_sge));
-	if (dev->nic_info.dev_family == OCRDMA_GEN2_FAMILY) {
+	if (ocrdma_get_asic_type(dev) == OCRDMA_ASIC_GEN_SKH_R) {
 		attr->ird = 1;
 		attr->ird_page_size = OCRDMA_MIN_Q_PAGE_SIZE;
 		attr->num_ird_pages = MAX_OCRDMA_IRD_PAGES;
@@ -1110,6 +1172,96 @@ mbx_err:
 	return status;
 }
 
+int ocrdma_mbx_rdma_stats(struct ocrdma_dev *dev, bool reset)
+{
+	struct ocrdma_rdma_stats_req *req = dev->stats_mem.va;
+	struct ocrdma_mqe *mqe = &dev->stats_mem.mqe;
+	struct ocrdma_rdma_stats_resp *old_stats = NULL;
+	int status;
+
+	old_stats = kzalloc(sizeof(*old_stats), GFP_KERNEL);
+	if (old_stats == NULL)
+		return -ENOMEM;
+
+	memset(mqe, 0, sizeof(*mqe));
+	mqe->hdr.pyld_len = dev->stats_mem.size;
+	mqe->hdr.spcl_sge_cnt_emb |=
+			(1 << OCRDMA_MQE_HDR_SGE_CNT_SHIFT) &
+				OCRDMA_MQE_HDR_SGE_CNT_MASK;
+	mqe->u.nonemb_req.sge[0].pa_lo = (u32) (dev->stats_mem.pa & 0xffffffff);
+	mqe->u.nonemb_req.sge[0].pa_hi = (u32) upper_32_bits(dev->stats_mem.pa);
+	mqe->u.nonemb_req.sge[0].len = dev->stats_mem.size;
+
+	/* Cache the old stats */
+	memcpy(old_stats, req, sizeof(struct ocrdma_rdma_stats_resp));
+	memset(req, 0, dev->stats_mem.size);
+
+	ocrdma_init_mch((struct ocrdma_mbx_hdr *)req,
+			OCRDMA_CMD_GET_RDMA_STATS,
+			OCRDMA_SUBSYS_ROCE,
+			dev->stats_mem.size);
+	if (reset)
+		req->reset_stats = reset;
+
+	status = ocrdma_nonemb_mbx_cmd(dev, mqe, dev->stats_mem.va);
+	if (status)
+		/* Copy from cache, if mbox fails */
+		memcpy(req, old_stats, sizeof(struct ocrdma_rdma_stats_resp));
+	else
+		ocrdma_le32_to_cpu(req, dev->stats_mem.size);
+
+	kfree(old_stats);
+	return status;
+}
+
+static int ocrdma_mbx_get_ctrl_attribs(struct ocrdma_dev *dev)
+{
+	int status = -ENOMEM;
+	struct ocrdma_dma_mem dma;
+	struct ocrdma_mqe *mqe;
+	struct ocrdma_get_ctrl_attribs_rsp *ctrl_attr_rsp;
+	struct mgmt_hba_attribs *hba_attribs;
+
+	mqe = ocrdma_alloc_mqe();
+	if (!mqe)
+		return status;
+	memset(mqe, 0, sizeof(*mqe));
+
+	dma.size = sizeof(struct ocrdma_get_ctrl_attribs_rsp);
+	dma.va	 = dma_alloc_coherent(&dev->nic_info.pdev->dev,
+					dma.size, &dma.pa, GFP_KERNEL);
+	if (!dma.va)
+		goto free_mqe;
+
+	mqe->hdr.pyld_len = dma.size;
+	mqe->hdr.spcl_sge_cnt_emb |=
+			(1 << OCRDMA_MQE_HDR_SGE_CNT_SHIFT) &
+			OCRDMA_MQE_HDR_SGE_CNT_MASK;
+	mqe->u.nonemb_req.sge[0].pa_lo = (u32) (dma.pa & 0xffffffff);
+	mqe->u.nonemb_req.sge[0].pa_hi = (u32) upper_32_bits(dma.pa);
+	mqe->u.nonemb_req.sge[0].len = dma.size;
+
+	memset(dma.va, 0, dma.size);
+	ocrdma_init_mch((struct ocrdma_mbx_hdr *)dma.va,
+			OCRDMA_CMD_GET_CTRL_ATTRIBUTES,
+			OCRDMA_SUBSYS_COMMON,
+			dma.size);
+
+	status = ocrdma_nonemb_mbx_cmd(dev, mqe, dma.va);
+	if (!status) {
+		ctrl_attr_rsp = (struct ocrdma_get_ctrl_attribs_rsp *)dma.va;
+		hba_attribs = &ctrl_attr_rsp->ctrl_attribs.hba_attribs;
+
+		dev->hba_port_num = hba_attribs->phy_port;
+		strncpy(dev->model_number,
+			hba_attribs->controller_model_number, 31);
+	}
+	dma_free_coherent(&dev->nic_info.pdev->dev, dma.size, dma.va, dma.pa);
+free_mqe:
+	kfree(mqe);
+	return status;
+}
+
 static int ocrdma_mbx_query_dev(struct ocrdma_dev *dev)
 {
 	int status = -ENOMEM;
@@ -1157,6 +1309,35 @@ mbx_err:
 	return status;
 }
 
+static int ocrdma_mbx_get_phy_info(struct ocrdma_dev *dev)
+{
+	int status = -ENOMEM;
+	struct ocrdma_mqe *cmd;
+	struct ocrdma_get_phy_info_rsp *rsp;
+
+	cmd = ocrdma_init_emb_mqe(OCRDMA_CMD_PHY_DETAILS, sizeof(*cmd));
+	if (!cmd)
+		return status;
+
+	ocrdma_init_mch((struct ocrdma_mbx_hdr *)&cmd->u.cmd[0],
+			OCRDMA_CMD_PHY_DETAILS, OCRDMA_SUBSYS_COMMON,
+			sizeof(*cmd));
+
+	status = ocrdma_mbx_cmd(dev, (struct ocrdma_mqe *)cmd);
+	if (status)
+		goto mbx_err;
+
+	rsp = (struct ocrdma_get_phy_info_rsp *)cmd;
+	dev->phy.phy_type = le16_to_cpu(rsp->phy_type);
+	dev->phy.auto_speeds_supported  =
+			le16_to_cpu(rsp->auto_speeds_supported);
+	dev->phy.fixed_speeds_supported =
+			le16_to_cpu(rsp->fixed_speeds_supported);
+mbx_err:
+	kfree(cmd);
+	return status;
+}
+
 int ocrdma_mbx_alloc_pd(struct ocrdma_dev *dev, struct ocrdma_pd *pd)
 {
 	int status = -ENOMEM;
@@ -1226,7 +1407,7 @@ static int ocrdma_build_q_conf(u32 *num_entries, int entry_size,
 
 static int ocrdma_mbx_create_ah_tbl(struct ocrdma_dev *dev)
 {
-	int i ;
+	int i;
 	int status = 0;
 	int max_ah;
 	struct ocrdma_create_ah_tbl *cmd;
@@ -1357,12 +1538,10 @@ static void ocrdma_unbind_eq(struct ocrdma_dev *dev, u16 eq_id)
 	int i;
 
 	mutex_lock(&dev->dev_lock);
-	for (i = 0; i < dev->eq_cnt; i++) {
-		if (dev->eq_tbl[i].q.id != eq_id)
-			continue;
-		dev->eq_tbl[i].cq_cnt -= 1;
-		break;
-	}
+	i = ocrdma_get_eq_table_index(dev, eq_id);
+	if (i == -EINVAL)
+		BUG();
+	dev->eq_tbl[i].cq_cnt -= 1;
 	mutex_unlock(&dev->dev_lock);
 }
 
@@ -1380,7 +1559,7 @@ int ocrdma_mbx_create_cq(struct ocrdma_dev *dev, struct ocrdma_cq *cq,
 		       __func__, dev->id, dev->attr.max_cqe, entries);
 		return -EINVAL;
 	}
-	if (dpp_cq && (dev->nic_info.dev_family != OCRDMA_GEN2_FAMILY))
+	if (dpp_cq && (ocrdma_get_asic_type(dev) != OCRDMA_ASIC_GEN_SKH_R))
 		return -EINVAL;
 
 	if (dpp_cq) {
@@ -1417,6 +1596,7 @@ int ocrdma_mbx_create_cq(struct ocrdma_dev *dev, struct ocrdma_cq *cq,
 	cq->eqn = ocrdma_bind_eq(dev);
 	cmd->cmd.req.rsvd_version = OCRDMA_CREATE_CQ_VER3;
 	cqe_count = cq->len / cqe_size;
+	cq->cqe_cnt = cqe_count;
 	if (cqe_count > 1024) {
 		/* Set cnt to 3 to indicate more than 1024 cq entries */
 		cmd->cmd.ev_cnt_flags |= (0x3 << OCRDMA_CREATE_CQ_CNT_SHIFT);
@@ -1439,7 +1619,7 @@ int ocrdma_mbx_create_cq(struct ocrdma_dev *dev, struct ocrdma_cq *cq,
 	}
 	/* shared eq between all the consumer cqs. */
 	cmd->cmd.eqn = cq->eqn;
-	if (dev->nic_info.dev_family == OCRDMA_GEN2_FAMILY) {
+	if (ocrdma_get_asic_type(dev) == OCRDMA_ASIC_GEN_SKH_R) {
 		if (dpp_cq)
 			cmd->cmd.pgsz_pgcnt |= OCRDMA_CREATE_CQ_DPP <<
 				OCRDMA_CREATE_CQ_TYPE_SHIFT;
@@ -1484,12 +1664,9 @@ int ocrdma_mbx_destroy_cq(struct ocrdma_dev *dev, struct ocrdma_cq *cq)
 	    (cq->id << OCRDMA_DESTROY_CQ_QID_SHIFT) &
 	    OCRDMA_DESTROY_CQ_QID_MASK;
 
-	ocrdma_unbind_eq(dev, cq->eqn);
 	status = ocrdma_mbx_cmd(dev, (struct ocrdma_mqe *)cmd);
-	if (status)
-		goto mbx_err;
+	ocrdma_unbind_eq(dev, cq->eqn);
 	dma_free_coherent(&dev->nic_info.pdev->dev, cq->len, cq->va, cq->pa);
-mbx_err:
 	kfree(cmd);
 	return status;
 }
@@ -2029,8 +2206,7 @@ int ocrdma_mbx_create_qp(struct ocrdma_qp *qp, struct ib_qp_init_attr *attrs,
 				OCRDMA_CREATE_QP_REQ_RQ_CQID_MASK;
 	qp->rq_cq = cq;
 
-	if (pd->dpp_enabled && attrs->cap.max_inline_data && pd->num_dpp_qp &&
-	    (attrs->cap.max_inline_data <= dev->attr.max_inline_data)) {
+	if (pd->dpp_enabled && pd->num_dpp_qp) {
 		ocrdma_set_create_qp_dpp_cmd(cmd, pd, qp, enable_dpp_cq,
 					     dpp_cq_id);
 	}
@@ -2099,7 +2275,7 @@ static int ocrdma_set_av_params(struct ocrdma_qp *qp,
 	memcpy(&cmd->params.dgid[0], &ah_attr->grh.dgid.raw[0],
 	       sizeof(cmd->params.dgid));
 	status = ocrdma_query_gid(&qp->dev->ibdev, 1,
-			 ah_attr->grh.sgid_index, &sgid);
+			ah_attr->grh.sgid_index, &sgid);
 	if (status)
 		return status;
 
@@ -2127,8 +2303,7 @@ static int ocrdma_set_av_params(struct ocrdma_qp *qp,
 
 static int ocrdma_set_qp_params(struct ocrdma_qp *qp,
 				struct ocrdma_modify_qp *cmd,
-				struct ib_qp_attr *attrs, int attr_mask,
-				enum ib_qp_state old_qps)
+				struct ib_qp_attr *attrs, int attr_mask)
 {
 	int status = 0;
 
@@ -2233,8 +2408,7 @@ pmtu_err:
 }
 
 int ocrdma_mbx_modify_qp(struct ocrdma_dev *dev, struct ocrdma_qp *qp,
-			 struct ib_qp_attr *attrs, int attr_mask,
-			 enum ib_qp_state old_qps)
+			 struct ib_qp_attr *attrs, int attr_mask)
 {
 	int status = -ENOMEM;
 	struct ocrdma_modify_qp *cmd;
@@ -2257,7 +2431,7 @@ int ocrdma_mbx_modify_qp(struct ocrdma_dev *dev, struct ocrdma_qp *qp,
 		    OCRDMA_QP_PARAMS_STATE_MASK;
 	}
 
-	status = ocrdma_set_qp_params(qp, cmd, attrs, attr_mask, old_qps);
+	status = ocrdma_set_qp_params(qp, cmd, attrs, attr_mask);
 	if (status)
 		goto mbx_err;
 	status = ocrdma_mbx_cmd(dev, (struct ocrdma_mqe *)cmd);
@@ -2488,7 +2662,7 @@ static int ocrdma_create_eqs(struct ocrdma_dev *dev)
 
 	for (i = 0; i < num_eq; i++) {
 		status = ocrdma_create_eq(dev, &dev->eq_tbl[i],
-					  OCRDMA_EQ_LEN);
+					OCRDMA_EQ_LEN);
 		if (status) {
 			status = -EINVAL;
 			break;
@@ -2533,6 +2707,13 @@ int ocrdma_init_hw(struct ocrdma_dev *dev)
 	status = ocrdma_mbx_create_ah_tbl(dev);
 	if (status)
 		goto conf_err;
+	status = ocrdma_mbx_get_phy_info(dev);
+	if (status)
+		goto conf_err;
+	status = ocrdma_mbx_get_ctrl_attribs(dev);
+	if (status)
+		goto conf_err;
+
 	return 0;
 
 conf_err:

drivers/infiniband/hw/ocrdma/ocrdma_hw.h

@@ -112,8 +112,7 @@ int ocrdma_mbx_create_qp(struct ocrdma_qp *, struct ib_qp_init_attr *attrs,
 			 u8 enable_dpp_cq, u16 dpp_cq_id, u16 *dpp_offset,
 			 u16 *dpp_credit_lmt);
 int ocrdma_mbx_modify_qp(struct ocrdma_dev *, struct ocrdma_qp *,
-			 struct ib_qp_attr *attrs, int attr_mask,
-			 enum ib_qp_state old_qps);
+			 struct ib_qp_attr *attrs, int attr_mask);
 int ocrdma_mbx_query_qp(struct ocrdma_dev *, struct ocrdma_qp *,
 			struct ocrdma_qp_params *param);
 int ocrdma_mbx_destroy_qp(struct ocrdma_dev *, struct ocrdma_qp *);
@@ -132,5 +131,8 @@ int ocrdma_qp_state_change(struct ocrdma_qp *, enum ib_qp_state new_state,
 bool ocrdma_is_qp_in_sq_flushlist(struct ocrdma_cq *, struct ocrdma_qp *);
 bool ocrdma_is_qp_in_rq_flushlist(struct ocrdma_cq *, struct ocrdma_qp *);
 void ocrdma_flush_qp(struct ocrdma_qp *);
+int ocrdma_get_irq(struct ocrdma_dev *dev, struct ocrdma_eq *eq);
 
+int ocrdma_mbx_rdma_stats(struct ocrdma_dev *, bool reset);
+char *port_speed_string(struct ocrdma_dev *dev);
 #endif				/* __OCRDMA_HW_H__ */

drivers/infiniband/hw/ocrdma/ocrdma_main.c

@@ -39,10 +39,11 @@
 #include "ocrdma_ah.h"
 #include "be_roce.h"
 #include "ocrdma_hw.h"
+#include "ocrdma_stats.h"
 #include "ocrdma_abi.h"
 
-MODULE_VERSION(OCRDMA_ROCE_DEV_VERSION);
-MODULE_DESCRIPTION("Emulex RoCE HCA Driver");
+MODULE_VERSION(OCRDMA_ROCE_DRV_VERSION);
+MODULE_DESCRIPTION(OCRDMA_ROCE_DRV_DESC " " OCRDMA_ROCE_DRV_VERSION);
 MODULE_AUTHOR("Emulex Corporation");
 MODULE_LICENSE("GPL");
 
@@ -286,7 +287,7 @@ static int ocrdma_register_device(struct ocrdma_dev *dev)
 
 	dev->ibdev.process_mad = ocrdma_process_mad;
 
-	if (dev->nic_info.dev_family == OCRDMA_GEN2_FAMILY) {
+	if (ocrdma_get_asic_type(dev) == OCRDMA_ASIC_GEN_SKH_R) {
 		dev->ibdev.uverbs_cmd_mask |=
 		     OCRDMA_UVERBS(CREATE_SRQ) |
 		     OCRDMA_UVERBS(MODIFY_SRQ) |
@@ -338,9 +339,42 @@ static void ocrdma_free_resources(struct ocrdma_dev *dev)
 	kfree(dev->sgid_tbl);
 }
 
+/* OCRDMA sysfs interface */
+static ssize_t show_rev(struct device *device, struct device_attribute *attr,
+			char *buf)
+{
+	struct ocrdma_dev *dev = dev_get_drvdata(device);
+
+	return scnprintf(buf, PAGE_SIZE, "0x%x\n", dev->nic_info.pdev->vendor);
+}
+
+static ssize_t show_fw_ver(struct device *device, struct device_attribute *attr,
+			char *buf)
+{
+	struct ocrdma_dev *dev = dev_get_drvdata(device);
+
+	return scnprintf(buf, PAGE_SIZE, "%s", &dev->attr.fw_ver[0]);
+}
+
+static DEVICE_ATTR(hw_rev, S_IRUGO, show_rev, NULL);
+static DEVICE_ATTR(fw_ver, S_IRUGO, show_fw_ver, NULL);
+
+static struct device_attribute *ocrdma_attributes[] = {
+	&dev_attr_hw_rev,
+	&dev_attr_fw_ver
+};
+
+static void ocrdma_remove_sysfiles(struct ocrdma_dev *dev)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(ocrdma_attributes); i++)
+		device_remove_file(&dev->ibdev.dev, ocrdma_attributes[i]);
+}
+
 static struct ocrdma_dev *ocrdma_add(struct be_dev_info *dev_info)
 {
-	int status = 0;
+	int status = 0, i;
 	struct ocrdma_dev *dev;
 
 	dev = (struct ocrdma_dev *)ib_alloc_device(sizeof(struct ocrdma_dev));
@@ -369,11 +403,25 @@ static struct ocrdma_dev *ocrdma_add(struct be_dev_info *dev_info)
 	if (status)
 		goto alloc_err;
 
+	for (i = 0; i < ARRAY_SIZE(ocrdma_attributes); i++)
+		if (device_create_file(&dev->ibdev.dev, ocrdma_attributes[i]))
+			goto sysfs_err;
 	spin_lock(&ocrdma_devlist_lock);
 	list_add_tail_rcu(&dev->entry, &ocrdma_dev_list);
 	spin_unlock(&ocrdma_devlist_lock);
+	/* Init stats */
+	ocrdma_add_port_stats(dev);
+
+	pr_info("%s %s: %s \"%s\" port %d\n",
+		dev_name(&dev->nic_info.pdev->dev), hca_name(dev),
+		port_speed_string(dev), dev->model_number,
+		dev->hba_port_num);
+	pr_info("%s ocrdma%d driver loaded successfully\n",
+		dev_name(&dev->nic_info.pdev->dev), dev->id);
 	return dev;
 
+sysfs_err:
+	ocrdma_remove_sysfiles(dev);
 alloc_err:
 	ocrdma_free_resources(dev);
 	ocrdma_cleanup_hw(dev);
@@ -400,6 +448,9 @@ static void ocrdma_remove(struct ocrdma_dev *dev)
 	/* first unregister with stack to stop all the active traffic
 	 * of the registered clients.
 	 */
+	ocrdma_rem_port_stats(dev);
+	ocrdma_remove_sysfiles(dev);
+
 	ib_unregister_device(&dev->ibdev);
 
 	spin_lock(&ocrdma_devlist_lock);
@@ -437,7 +488,7 @@ static int ocrdma_close(struct ocrdma_dev *dev)
 		cur_qp = dev->qp_tbl;
 		for (i = 0; i < OCRDMA_MAX_QP; i++) {
 			qp = cur_qp[i];
-			if (qp) {
+			if (qp && qp->ibqp.qp_type != IB_QPT_GSI) {
 				/* change the QP state to ERROR */
 				_ocrdma_modify_qp(&qp->ibqp, &attrs, attr_mask);
 
@@ -478,6 +529,7 @@ static struct ocrdma_driver ocrdma_drv = {
 	.add			= ocrdma_add,
 	.remove			= ocrdma_remove,
 	.state_change_handler	= ocrdma_event_handler,
+	.be_abi_version		= OCRDMA_BE_ROCE_ABI_VERSION,
 };
 
 static void ocrdma_unregister_inet6addr_notifier(void)
@@ -487,10 +539,17 @@ static void ocrdma_unregister_inet6addr_notifier(void)
 #endif
 }
 
+static void ocrdma_unregister_inetaddr_notifier(void)
+{
+	unregister_inetaddr_notifier(&ocrdma_inetaddr_notifier);
+}
+
 static int __init ocrdma_init_module(void)
 {
 	int status;
 
+	ocrdma_init_debugfs();
+
 	status = register_inetaddr_notifier(&ocrdma_inetaddr_notifier);
 	if (status)
 		return status;
@@ -498,13 +557,19 @@ static int __init ocrdma_init_module(void)
 #if IS_ENABLED(CONFIG_IPV6)
 	status = register_inet6addr_notifier(&ocrdma_inet6addr_notifier);
 	if (status)
-		return status;
+		goto err_notifier6;
 #endif
 
 	status = be_roce_register_driver(&ocrdma_drv);
 	if (status)
-		ocrdma_unregister_inet6addr_notifier();
+		goto err_be_reg;
 
+	return 0;
+
+err_be_reg:
+	ocrdma_unregister_inet6addr_notifier();
+err_notifier6:
+	ocrdma_unregister_inetaddr_notifier();
 	return status;
 }
 
@@ -512,6 +577,8 @@ static void __exit ocrdma_exit_module(void)
 {
 	be_roce_unregister_driver(&ocrdma_drv);
 	ocrdma_unregister_inet6addr_notifier();
+	ocrdma_unregister_inetaddr_notifier();
+	ocrdma_rem_debugfs();
 }
 
 module_init(ocrdma_init_module);

drivers/infiniband/hw/ocrdma/ocrdma_sli.h

@@ -30,8 +30,16 @@
 
 #define Bit(_b) (1 << (_b))
 
-#define OCRDMA_GEN1_FAMILY	0xB
-#define OCRDMA_GEN2_FAMILY	0x0F
+enum {
+	OCRDMA_ASIC_GEN_SKH_R = 0x04,
+	OCRDMA_ASIC_GEN_LANCER = 0x0B
+};
+
+enum {
+	OCRDMA_ASIC_REV_A0 = 0x00,
+	OCRDMA_ASIC_REV_B0 = 0x10,
+	OCRDMA_ASIC_REV_C0 = 0x20
+};
 
 #define OCRDMA_SUBSYS_ROCE 10
 enum {
@@ -64,6 +72,7 @@ enum {
 
 	OCRDMA_CMD_ATTACH_MCAST,
 	OCRDMA_CMD_DETACH_MCAST,
+	OCRDMA_CMD_GET_RDMA_STATS,
 
 	OCRDMA_CMD_MAX
 };
@@ -74,12 +83,14 @@ enum {
 	OCRDMA_CMD_CREATE_CQ		= 12,
 	OCRDMA_CMD_CREATE_EQ		= 13,
 	OCRDMA_CMD_CREATE_MQ		= 21,
+	OCRDMA_CMD_GET_CTRL_ATTRIBUTES  = 32,
 	OCRDMA_CMD_GET_FW_VER		= 35,
 	OCRDMA_CMD_DELETE_MQ		= 53,
 	OCRDMA_CMD_DELETE_CQ		= 54,
 	OCRDMA_CMD_DELETE_EQ		= 55,
 	OCRDMA_CMD_GET_FW_CONFIG	= 58,
-	OCRDMA_CMD_CREATE_MQ_EXT	= 90
+	OCRDMA_CMD_CREATE_MQ_EXT	= 90,
+	OCRDMA_CMD_PHY_DETAILS		= 102
 };
 
 enum {
@@ -103,7 +114,10 @@ enum {
 	OCRDMA_DB_GEN2_SRQ_OFFSET	= OCRDMA_DB_GEN2_RQ_OFFSET,
 	OCRDMA_DB_CQ_OFFSET		= 0x120,
 	OCRDMA_DB_EQ_OFFSET		= OCRDMA_DB_CQ_OFFSET,
-	OCRDMA_DB_MQ_OFFSET		= 0x140
+	OCRDMA_DB_MQ_OFFSET		= 0x140,
+
+	OCRDMA_DB_SQ_SHIFT		= 16,
+	OCRDMA_DB_RQ_SHIFT		= 24
 };
 
 #define OCRDMA_DB_CQ_RING_ID_MASK       0x3FF	/* bits 0 - 9 */
@@ -138,6 +152,10 @@ enum {
 #define OCRDMA_MIN_Q_PAGE_SIZE (4096)
 #define OCRDMA_MAX_Q_PAGES     (8)
 
+#define OCRDMA_SLI_ASIC_ID_OFFSET	0x9C
+#define OCRDMA_SLI_ASIC_REV_MASK	0x000000FF
+#define OCRDMA_SLI_ASIC_GEN_NUM_MASK	0x0000FF00
+#define OCRDMA_SLI_ASIC_GEN_NUM_SHIFT	0x08
 /*
 # 0: 4K Bytes
 # 1: 8K Bytes
@@ -562,6 +580,30 @@ enum {
 	OCRDMA_FN_MODE_RDMA	= 0x4
 };
 
+struct ocrdma_get_phy_info_rsp {
+	struct ocrdma_mqe_hdr hdr;
+	struct ocrdma_mbx_rsp rsp;
+
+	u16 phy_type;
+	u16 interface_type;
+	u32 misc_params;
+	u16 ext_phy_details;
+	u16 rsvd;
+	u16 auto_speeds_supported;
+	u16 fixed_speeds_supported;
+	u32 future_use[2];
+};
+
+enum {
+	OCRDMA_PHY_SPEED_ZERO = 0x0,
+	OCRDMA_PHY_SPEED_10MBPS = 0x1,
+	OCRDMA_PHY_SPEED_100MBPS = 0x2,
+	OCRDMA_PHY_SPEED_1GBPS = 0x4,
+	OCRDMA_PHY_SPEED_10GBPS = 0x8,
+	OCRDMA_PHY_SPEED_40GBPS = 0x20
+};
+
+
 struct ocrdma_get_link_speed_rsp {
 	struct ocrdma_mqe_hdr hdr;
 	struct ocrdma_mbx_rsp rsp;
@@ -590,7 +632,7 @@ enum {
 
 enum {
 	OCRDMA_CREATE_CQ_VER2			= 2,
-	OCRDMA_CREATE_CQ_VER3                   = 3,
+	OCRDMA_CREATE_CQ_VER3			= 3,
 
 	OCRDMA_CREATE_CQ_PAGE_CNT_MASK		= 0xFFFF,
 	OCRDMA_CREATE_CQ_PAGE_SIZE_SHIFT	= 16,
@@ -1050,6 +1092,7 @@ enum {
 	OCRDMA_MODIFY_QP_RSP_MAX_ORD_MASK	= 0xFFFF <<
 					OCRDMA_MODIFY_QP_RSP_MAX_ORD_SHIFT
 };
+
 struct ocrdma_modify_qp_rsp {
 	struct ocrdma_mqe_hdr hdr;
 	struct ocrdma_mbx_rsp rsp;
@@ -1062,8 +1105,8 @@ struct ocrdma_query_qp {
 	struct ocrdma_mqe_hdr hdr;
 	struct ocrdma_mbx_hdr req;
 
-#define OCRDMA_QUERY_UP_QP_ID_SHIFT 0
-#define OCRDMA_QUERY_UP_QP_ID_MASK   0xFFFFFF
+#define OCRDMA_QUERY_UP_QP_ID_SHIFT	0
+#define OCRDMA_QUERY_UP_QP_ID_MASK	0xFFFFFF
 	u32 qp_id;
 };
 
@@ -1703,4 +1746,208 @@ struct ocrdma_av {
 	u32 valid;
 } __packed;
 
+struct ocrdma_rsrc_stats {
+	u32 dpp_pds;
+	u32 non_dpp_pds;
+	u32 rc_dpp_qps;
+	u32 uc_dpp_qps;
+	u32 ud_dpp_qps;
+	u32 rc_non_dpp_qps;
+	u32 rsvd;
+	u32 uc_non_dpp_qps;
+	u32 ud_non_dpp_qps;
+	u32 rsvd1;
+	u32 srqs;
+	u32 rbqs;
+	u32 r64K_nsmr;
+	u32 r64K_to_2M_nsmr;
+	u32 r2M_to_44M_nsmr;
+	u32 r44M_to_1G_nsmr;
+	u32 r1G_to_4G_nsmr;
+	u32 nsmr_count_4G_to_32G;
+	u32 r32G_to_64G_nsmr;
+	u32 r64G_to_128G_nsmr;
+	u32 r128G_to_higher_nsmr;
+	u32 embedded_nsmr;
+	u32 frmr;
+	u32 prefetch_qps;
+	u32 ondemand_qps;
+	u32 phy_mr;
+	u32 mw;
+	u32 rsvd2[7];
+};
+
+struct ocrdma_db_err_stats {
+	u32 sq_doorbell_errors;
+	u32 cq_doorbell_errors;
+	u32 rq_srq_doorbell_errors;
+	u32 cq_overflow_errors;
+	u32 rsvd[4];
+};
+
+struct ocrdma_wqe_stats {
+	u32 large_send_rc_wqes_lo;
+	u32 large_send_rc_wqes_hi;
+	u32 large_write_rc_wqes_lo;
+	u32 large_write_rc_wqes_hi;
+	u32 rsvd[4];
+	u32 read_wqes_lo;
+	u32 read_wqes_hi;
+	u32 frmr_wqes_lo;
+	u32 frmr_wqes_hi;
+	u32 mw_bind_wqes_lo;
+	u32 mw_bind_wqes_hi;
+	u32 invalidate_wqes_lo;
+	u32 invalidate_wqes_hi;
+	u32 rsvd1[2];
+	u32 dpp_wqe_drops;
+	u32 rsvd2[5];
+};
+
+struct ocrdma_tx_stats {
+	u32 send_pkts_lo;
+	u32 send_pkts_hi;
+	u32 write_pkts_lo;
+	u32 write_pkts_hi;
+	u32 read_pkts_lo;
+	u32 read_pkts_hi;
+	u32 read_rsp_pkts_lo;
+	u32 read_rsp_pkts_hi;
+	u32 ack_pkts_lo;
+	u32 ack_pkts_hi;
+	u32 send_bytes_lo;
+	u32 send_bytes_hi;
+	u32 write_bytes_lo;
+	u32 write_bytes_hi;
+	u32 read_req_bytes_lo;
+	u32 read_req_bytes_hi;
+	u32 read_rsp_bytes_lo;
+	u32 read_rsp_bytes_hi;
+	u32 ack_timeouts;
+	u32 rsvd[5];
+};
+
+
+struct ocrdma_tx_qp_err_stats {
+	u32 local_length_errors;
+	u32 local_protection_errors;
+	u32 local_qp_operation_errors;
+	u32 retry_count_exceeded_errors;
+	u32 rnr_retry_count_exceeded_errors;
+	u32 rsvd[3];
+};
+
+struct ocrdma_rx_stats {
+	u32 roce_frame_bytes_lo;
+	u32 roce_frame_bytes_hi;
+	u32 roce_frame_icrc_drops;
+	u32 roce_frame_payload_len_drops;
+	u32 ud_drops;
+	u32 qp1_drops;
+	u32 psn_error_request_packets;
+	u32 psn_error_resp_packets;
+	u32 rnr_nak_timeouts;
+	u32 rnr_nak_receives;
+	u32 roce_frame_rxmt_drops;
+	u32 nak_count_psn_sequence_errors;
+	u32 rc_drop_count_lookup_errors;
+	u32 rq_rnr_naks;
+	u32 srq_rnr_naks;
+	u32 roce_frames_lo;
+	u32 roce_frames_hi;
+	u32 rsvd;
+};
+
+struct ocrdma_rx_qp_err_stats {
+	u32 nak_invalid_requst_errors;
+	u32 nak_remote_operation_errors;
+	u32 nak_count_remote_access_errors;
+	u32 local_length_errors;
+	u32 local_protection_errors;
+	u32 local_qp_operation_errors;
+	u32 rsvd[2];
+};
+
+struct ocrdma_tx_dbg_stats {
+	u32 data[100];
+};
+
+struct ocrdma_rx_dbg_stats {
+	u32 data[200];
+};
+
+struct ocrdma_rdma_stats_req {
+	struct ocrdma_mbx_hdr hdr;
+	u8 reset_stats;
+	u8 rsvd[3];
+} __packed;
+
+struct ocrdma_rdma_stats_resp {
+	struct ocrdma_mbx_hdr hdr;
+	struct ocrdma_rsrc_stats act_rsrc_stats;
+	struct ocrdma_rsrc_stats th_rsrc_stats;
+	struct ocrdma_db_err_stats	db_err_stats;
+	struct ocrdma_wqe_stats		wqe_stats;
+	struct ocrdma_tx_stats		tx_stats;
+	struct ocrdma_tx_qp_err_stats	tx_qp_err_stats;
+	struct ocrdma_rx_stats		rx_stats;
+	struct ocrdma_rx_qp_err_stats	rx_qp_err_stats;
+	struct ocrdma_tx_dbg_stats	tx_dbg_stats;
+	struct ocrdma_rx_dbg_stats	rx_dbg_stats;
+} __packed;
+
+
+struct mgmt_hba_attribs {
+	u8 flashrom_version_string[32];
+	u8 manufacturer_name[32];
+	u32 supported_modes;
+	u32 rsvd0[3];
+	u8 ncsi_ver_string[12];
+	u32 default_extended_timeout;
+	u8 controller_model_number[32];
+	u8 controller_description[64];
+	u8 controller_serial_number[32];
+	u8 ip_version_string[32];
+	u8 firmware_version_string[32];
+	u8 bios_version_string[32];
+	u8 redboot_version_string[32];
+	u8 driver_version_string[32];
+	u8 fw_on_flash_version_string[32];
+	u32 functionalities_supported;
+	u16 max_cdblength;
+	u8 asic_revision;
+	u8 generational_guid[16];
+	u8 hba_port_count;
+	u16 default_link_down_timeout;
+	u8 iscsi_ver_min_max;
+	u8 multifunction_device;
+	u8 cache_valid;
+	u8 hba_status;
+	u8 max_domains_supported;
+	u8 phy_port;
+	u32 firmware_post_status;
+	u32 hba_mtu[8];
+	u32 rsvd1[4];
+};
+
+struct mgmt_controller_attrib {
+	struct mgmt_hba_attribs hba_attribs;
+	u16 pci_vendor_id;
+	u16 pci_device_id;
+	u16 pci_sub_vendor_id;
+	u16 pci_sub_system_id;
+	u8 pci_bus_number;
+	u8 pci_device_number;
+	u8 pci_function_number;
+	u8 interface_type;
+	u64 unique_identifier;
+	u32 rsvd0[5];
+};
+
+struct ocrdma_get_ctrl_attribs_rsp {
+	struct ocrdma_mbx_hdr hdr;
+	struct mgmt_controller_attrib ctrl_attribs;
+};
+
+
 #endif				/* __OCRDMA_SLI_H__ */

drivers/infiniband/hw/ocrdma/ocrdma_stats.c

@@ -0,0 +1,623 @@
+/*******************************************************************
+ * This file is part of the Emulex RoCE Device Driver for          *
+ * RoCE (RDMA over Converged Ethernet) adapters.                   *
+ * Copyright (C) 2008-2014 Emulex. All rights reserved.            *
+ * EMULEX and SLI are trademarks of Emulex.                        *
+ * www.emulex.com                                                  *
+ *                                                                 *
+ * This program is free software; you can redistribute it and/or   *
+ * modify it under the terms of version 2 of the GNU General       *
+ * Public License as published by the Free Software Foundation.    *
+ * This program is distributed in the hope that it will be useful. *
+ * ALL EXPRESS OR IMPLIED CONDITIONS, REPRESENTATIONS AND          *
+ * WARRANTIES, INCLUDING ANY IMPLIED WARRANTY OF MERCHANTABILITY,  *
+ * FITNESS FOR A PARTICULAR PURPOSE, OR NON-INFRINGEMENT, ARE      *
+ * DISCLAIMED, EXCEPT TO THE EXTENT THAT SUCH DISCLAIMERS ARE HELD *
+ * TO BE LEGALLY INVALID.  See the GNU General Public License for  *
+ * more details, a copy of which can be found in the file COPYING  *
+ * included with this package.                                     *
+ *
+ * Contact Information:
+ * linux-drivers@emulex.com
+ *
+ * Emulex
+ * 3333 Susan Street
+ * Costa Mesa, CA 92626
+ *******************************************************************/
+
+#include <rdma/ib_addr.h>
+#include "ocrdma_stats.h"
+
+static struct dentry *ocrdma_dbgfs_dir;
+
+static int ocrdma_add_stat(char *start, char *pcur,
+				char *name, u64 count)
+{
+	char buff[128] = {0};
+	int cpy_len = 0;
+
+	snprintf(buff, 128, "%s: %llu\n", name, count);
+	cpy_len = strlen(buff);
+
+	if (pcur + cpy_len > start + OCRDMA_MAX_DBGFS_MEM) {
+		pr_err("%s: No space in stats buff\n", __func__);
+		return 0;
+	}
+
+	memcpy(pcur, buff, cpy_len);
+	return cpy_len;
+}
+
+static bool ocrdma_alloc_stats_mem(struct ocrdma_dev *dev)
+{
+	struct stats_mem *mem = &dev->stats_mem;
+
+	/* Alloc mbox command mem*/
+	mem->size = max_t(u32, sizeof(struct ocrdma_rdma_stats_req),
+			sizeof(struct ocrdma_rdma_stats_resp));
+
+	mem->va   = dma_alloc_coherent(&dev->nic_info.pdev->dev, mem->size,
+					 &mem->pa, GFP_KERNEL);
+	if (!mem->va) {
+		pr_err("%s: stats mbox allocation failed\n", __func__);
+		return false;
+	}
+
+	memset(mem->va, 0, mem->size);
+
+	/* Alloc debugfs mem */
+	mem->debugfs_mem = kzalloc(OCRDMA_MAX_DBGFS_MEM, GFP_KERNEL);
+	if (!mem->debugfs_mem) {
+		pr_err("%s: stats debugfs mem allocation failed\n", __func__);
+		return false;
+	}
+
+	return true;
+}
+
+static void ocrdma_release_stats_mem(struct ocrdma_dev *dev)
+{
+	struct stats_mem *mem = &dev->stats_mem;
+
+	if (mem->va)
+		dma_free_coherent(&dev->nic_info.pdev->dev, mem->size,
+				  mem->va, mem->pa);
+	kfree(mem->debugfs_mem);
+}
+
+static char *ocrdma_resource_stats(struct ocrdma_dev *dev)
+{
+	char *stats = dev->stats_mem.debugfs_mem, *pcur;
+	struct ocrdma_rdma_stats_resp *rdma_stats =
+			(struct ocrdma_rdma_stats_resp *)dev->stats_mem.va;
+	struct ocrdma_rsrc_stats *rsrc_stats = &rdma_stats->act_rsrc_stats;
+
+	memset(stats, 0, (OCRDMA_MAX_DBGFS_MEM));
+
+	pcur = stats;
+	pcur += ocrdma_add_stat(stats, pcur, "active_dpp_pds",
+				(u64)rsrc_stats->dpp_pds);
+	pcur += ocrdma_add_stat(stats, pcur, "active_non_dpp_pds",
+				(u64)rsrc_stats->non_dpp_pds);
+	pcur += ocrdma_add_stat(stats, pcur, "active_rc_dpp_qps",
+				(u64)rsrc_stats->rc_dpp_qps);
+	pcur += ocrdma_add_stat(stats, pcur, "active_uc_dpp_qps",
+				(u64)rsrc_stats->uc_dpp_qps);
+	pcur += ocrdma_add_stat(stats, pcur, "active_ud_dpp_qps",
+				(u64)rsrc_stats->ud_dpp_qps);
+	pcur += ocrdma_add_stat(stats, pcur, "active_rc_non_dpp_qps",
+				(u64)rsrc_stats->rc_non_dpp_qps);
+	pcur += ocrdma_add_stat(stats, pcur, "active_uc_non_dpp_qps",
+				(u64)rsrc_stats->uc_non_dpp_qps);
+	pcur += ocrdma_add_stat(stats, pcur, "active_ud_non_dpp_qps",
+				(u64)rsrc_stats->ud_non_dpp_qps);
+	pcur += ocrdma_add_stat(stats, pcur, "active_srqs",
+				(u64)rsrc_stats->srqs);
+	pcur += ocrdma_add_stat(stats, pcur, "active_rbqs",
+				(u64)rsrc_stats->rbqs);
+	pcur += ocrdma_add_stat(stats, pcur, "active_64K_nsmr",
+				(u64)rsrc_stats->r64K_nsmr);
+	pcur += ocrdma_add_stat(stats, pcur, "active_64K_to_2M_nsmr",
+				(u64)rsrc_stats->r64K_to_2M_nsmr);
+	pcur += ocrdma_add_stat(stats, pcur, "active_2M_to_44M_nsmr",
+				(u64)rsrc_stats->r2M_to_44M_nsmr);
+	pcur += ocrdma_add_stat(stats, pcur, "active_44M_to_1G_nsmr",
+				(u64)rsrc_stats->r44M_to_1G_nsmr);
+	pcur += ocrdma_add_stat(stats, pcur, "active_1G_to_4G_nsmr",
+				(u64)rsrc_stats->r1G_to_4G_nsmr);
+	pcur += ocrdma_add_stat(stats, pcur, "active_nsmr_count_4G_to_32G",
+				(u64)rsrc_stats->nsmr_count_4G_to_32G);
+	pcur += ocrdma_add_stat(stats, pcur, "active_32G_to_64G_nsmr",
+				(u64)rsrc_stats->r32G_to_64G_nsmr);
+	pcur += ocrdma_add_stat(stats, pcur, "active_64G_to_128G_nsmr",
+				(u64)rsrc_stats->r64G_to_128G_nsmr);
+	pcur += ocrdma_add_stat(stats, pcur, "active_128G_to_higher_nsmr",
+				(u64)rsrc_stats->r128G_to_higher_nsmr);
+	pcur += ocrdma_add_stat(stats, pcur, "active_embedded_nsmr",
+				(u64)rsrc_stats->embedded_nsmr);
+	pcur += ocrdma_add_stat(stats, pcur, "active_frmr",
+				(u64)rsrc_stats->frmr);
+	pcur += ocrdma_add_stat(stats, pcur, "active_prefetch_qps",
+				(u64)rsrc_stats->prefetch_qps);
+	pcur += ocrdma_add_stat(stats, pcur, "active_ondemand_qps",
+				(u64)rsrc_stats->ondemand_qps);
+	pcur += ocrdma_add_stat(stats, pcur, "active_phy_mr",
+				(u64)rsrc_stats->phy_mr);
+	pcur += ocrdma_add_stat(stats, pcur, "active_mw",
+				(u64)rsrc_stats->mw);
+
+	/* Print the threshold stats */
+	rsrc_stats = &rdma_stats->th_rsrc_stats;
+
+	pcur += ocrdma_add_stat(stats, pcur, "threshold_dpp_pds",
+				(u64)rsrc_stats->dpp_pds);
+	pcur += ocrdma_add_stat(stats, pcur, "threshold_non_dpp_pds",
+				(u64)rsrc_stats->non_dpp_pds);
+	pcur += ocrdma_add_stat(stats, pcur, "threshold_rc_dpp_qps",
+				(u64)rsrc_stats->rc_dpp_qps);
+	pcur += ocrdma_add_stat(stats, pcur, "threshold_uc_dpp_qps",
+				(u64)rsrc_stats->uc_dpp_qps);
+	pcur += ocrdma_add_stat(stats, pcur, "threshold_ud_dpp_qps",
+				(u64)rsrc_stats->ud_dpp_qps);
+	pcur += ocrdma_add_stat(stats, pcur, "threshold_rc_non_dpp_qps",
+				(u64)rsrc_stats->rc_non_dpp_qps);
+	pcur += ocrdma_add_stat(stats, pcur, "threshold_uc_non_dpp_qps",
+				(u64)rsrc_stats->uc_non_dpp_qps);
+	pcur += ocrdma_add_stat(stats, pcur, "threshold_ud_non_dpp_qps",
+				(u64)rsrc_stats->ud_non_dpp_qps);
+	pcur += ocrdma_add_stat(stats, pcur, "threshold_srqs",
+				(u64)rsrc_stats->srqs);
+	pcur += ocrdma_add_stat(stats, pcur, "threshold_rbqs",
+				(u64)rsrc_stats->rbqs);
+	pcur += ocrdma_add_stat(stats, pcur, "threshold_64K_nsmr",
+				(u64)rsrc_stats->r64K_nsmr);
+	pcur += ocrdma_add_stat(stats, pcur, "threshold_64K_to_2M_nsmr",
+				(u64)rsrc_stats->r64K_to_2M_nsmr);
+	pcur += ocrdma_add_stat(stats, pcur, "threshold_2M_to_44M_nsmr",
+				(u64)rsrc_stats->r2M_to_44M_nsmr);
+	pcur += ocrdma_add_stat(stats, pcur, "threshold_44M_to_1G_nsmr",
+				(u64)rsrc_stats->r44M_to_1G_nsmr);
+	pcur += ocrdma_add_stat(stats, pcur, "threshold_1G_to_4G_nsmr",
+				(u64)rsrc_stats->r1G_to_4G_nsmr);
+	pcur += ocrdma_add_stat(stats, pcur, "threshold_nsmr_count_4G_to_32G",
+				(u64)rsrc_stats->nsmr_count_4G_to_32G);
+	pcur += ocrdma_add_stat(stats, pcur, "threshold_32G_to_64G_nsmr",
+				(u64)rsrc_stats->r32G_to_64G_nsmr);
+	pcur += ocrdma_add_stat(stats, pcur, "threshold_64G_to_128G_nsmr",
+				(u64)rsrc_stats->r64G_to_128G_nsmr);
+	pcur += ocrdma_add_stat(stats, pcur, "threshold_128G_to_higher_nsmr",
+				(u64)rsrc_stats->r128G_to_higher_nsmr);
+	pcur += ocrdma_add_stat(stats, pcur, "threshold_embedded_nsmr",
+				(u64)rsrc_stats->embedded_nsmr);
+	pcur += ocrdma_add_stat(stats, pcur, "threshold_frmr",
+				(u64)rsrc_stats->frmr);
+	pcur += ocrdma_add_stat(stats, pcur, "threshold_prefetch_qps",
+				(u64)rsrc_stats->prefetch_qps);
+	pcur += ocrdma_add_stat(stats, pcur, "threshold_ondemand_qps",
+				(u64)rsrc_stats->ondemand_qps);
+	pcur += ocrdma_add_stat(stats, pcur, "threshold_phy_mr",
+				(u64)rsrc_stats->phy_mr);
+	pcur += ocrdma_add_stat(stats, pcur, "threshold_mw",
+				(u64)rsrc_stats->mw);
+	return stats;
+}
+
+static char *ocrdma_rx_stats(struct ocrdma_dev *dev)
+{
+	char *stats = dev->stats_mem.debugfs_mem, *pcur;
+	struct ocrdma_rdma_stats_resp *rdma_stats =
+		(struct ocrdma_rdma_stats_resp *)dev->stats_mem.va;
+	struct ocrdma_rx_stats *rx_stats = &rdma_stats->rx_stats;
+
+	memset(stats, 0, (OCRDMA_MAX_DBGFS_MEM));
+
+	pcur = stats;
+	pcur += ocrdma_add_stat
+		(stats, pcur, "roce_frame_bytes",
+		 convert_to_64bit(rx_stats->roce_frame_bytes_lo,
+		 rx_stats->roce_frame_bytes_hi));
+	pcur += ocrdma_add_stat(stats, pcur, "roce_frame_icrc_drops",
+				(u64)rx_stats->roce_frame_icrc_drops);
+	pcur += ocrdma_add_stat(stats, pcur, "roce_frame_payload_len_drops",
+				(u64)rx_stats->roce_frame_payload_len_drops);
+	pcur += ocrdma_add_stat(stats, pcur, "ud_drops",
+				(u64)rx_stats->ud_drops);
+	pcur += ocrdma_add_stat(stats, pcur, "qp1_drops",
+				(u64)rx_stats->qp1_drops);
+	pcur += ocrdma_add_stat(stats, pcur, "psn_error_request_packets",
+				(u64)rx_stats->psn_error_request_packets);
+	pcur += ocrdma_add_stat(stats, pcur, "psn_error_resp_packets",
+				(u64)rx_stats->psn_error_resp_packets);
+	pcur += ocrdma_add_stat(stats, pcur, "rnr_nak_timeouts",
+				(u64)rx_stats->rnr_nak_timeouts);
+	pcur += ocrdma_add_stat(stats, pcur, "rnr_nak_receives",
+				(u64)rx_stats->rnr_nak_receives);
+	pcur += ocrdma_add_stat(stats, pcur, "roce_frame_rxmt_drops",
+				(u64)rx_stats->roce_frame_rxmt_drops);
+	pcur += ocrdma_add_stat(stats, pcur, "nak_count_psn_sequence_errors",
+				(u64)rx_stats->nak_count_psn_sequence_errors);
+	pcur += ocrdma_add_stat(stats, pcur, "rc_drop_count_lookup_errors",
+				(u64)rx_stats->rc_drop_count_lookup_errors);
+	pcur += ocrdma_add_stat(stats, pcur, "rq_rnr_naks",
+				(u64)rx_stats->rq_rnr_naks);
+	pcur += ocrdma_add_stat(stats, pcur, "srq_rnr_naks",
+				(u64)rx_stats->srq_rnr_naks);
+	pcur += ocrdma_add_stat(stats, pcur, "roce_frames",
+				convert_to_64bit(rx_stats->roce_frames_lo,
+						 rx_stats->roce_frames_hi));
+
+	return stats;
+}
+
+static char *ocrdma_tx_stats(struct ocrdma_dev *dev)
+{
+	char *stats = dev->stats_mem.debugfs_mem, *pcur;
+	struct ocrdma_rdma_stats_resp *rdma_stats =
+		(struct ocrdma_rdma_stats_resp *)dev->stats_mem.va;
+	struct ocrdma_tx_stats *tx_stats = &rdma_stats->tx_stats;
+
+	memset(stats, 0, (OCRDMA_MAX_DBGFS_MEM));
+
+	pcur = stats;
+	pcur += ocrdma_add_stat(stats, pcur, "send_pkts",
+				convert_to_64bit(tx_stats->send_pkts_lo,
+						 tx_stats->send_pkts_hi));
+	pcur += ocrdma_add_stat(stats, pcur, "write_pkts",
+				convert_to_64bit(tx_stats->write_pkts_lo,
+						 tx_stats->write_pkts_hi));
+	pcur += ocrdma_add_stat(stats, pcur, "read_pkts",
+				convert_to_64bit(tx_stats->read_pkts_lo,
+						 tx_stats->read_pkts_hi));
+	pcur += ocrdma_add_stat(stats, pcur, "read_rsp_pkts",
+				convert_to_64bit(tx_stats->read_rsp_pkts_lo,
+						 tx_stats->read_rsp_pkts_hi));
+	pcur += ocrdma_add_stat(stats, pcur, "ack_pkts",
+				convert_to_64bit(tx_stats->ack_pkts_lo,
+						 tx_stats->ack_pkts_hi));
+	pcur += ocrdma_add_stat(stats, pcur, "send_bytes",
+				convert_to_64bit(tx_stats->send_bytes_lo,
+						 tx_stats->send_bytes_hi));
+	pcur += ocrdma_add_stat(stats, pcur, "write_bytes",
+				convert_to_64bit(tx_stats->write_bytes_lo,
+						 tx_stats->write_bytes_hi));
+	pcur += ocrdma_add_stat(stats, pcur, "read_req_bytes",
+				convert_to_64bit(tx_stats->read_req_bytes_lo,
+						 tx_stats->read_req_bytes_hi));
+	pcur += ocrdma_add_stat(stats, pcur, "read_rsp_bytes",
+				convert_to_64bit(tx_stats->read_rsp_bytes_lo,
+						 tx_stats->read_rsp_bytes_hi));
+	pcur += ocrdma_add_stat(stats, pcur, "ack_timeouts",
+				(u64)tx_stats->ack_timeouts);
+
+	return stats;
+}
+
+static char *ocrdma_wqe_stats(struct ocrdma_dev *dev)
+{
+	char *stats = dev->stats_mem.debugfs_mem, *pcur;
+	struct ocrdma_rdma_stats_resp *rdma_stats =
+		(struct ocrdma_rdma_stats_resp *)dev->stats_mem.va;
+	struct ocrdma_wqe_stats	*wqe_stats = &rdma_stats->wqe_stats;
+
+	memset(stats, 0, (OCRDMA_MAX_DBGFS_MEM));
+
+	pcur = stats;
+	pcur += ocrdma_add_stat(stats, pcur, "large_send_rc_wqes",
+		convert_to_64bit(wqe_stats->large_send_rc_wqes_lo,
+				 wqe_stats->large_send_rc_wqes_hi));
+	pcur += ocrdma_add_stat(stats, pcur, "large_write_rc_wqes",
+		convert_to_64bit(wqe_stats->large_write_rc_wqes_lo,
+				 wqe_stats->large_write_rc_wqes_hi));
+	pcur += ocrdma_add_stat(stats, pcur, "read_wqes",
+				convert_to_64bit(wqe_stats->read_wqes_lo,
+						 wqe_stats->read_wqes_hi));
+	pcur += ocrdma_add_stat(stats, pcur, "frmr_wqes",
+				convert_to_64bit(wqe_stats->frmr_wqes_lo,
+						 wqe_stats->frmr_wqes_hi));
+	pcur += ocrdma_add_stat(stats, pcur, "mw_bind_wqes",
+				convert_to_64bit(wqe_stats->mw_bind_wqes_lo,
+						 wqe_stats->mw_bind_wqes_hi));
+	pcur += ocrdma_add_stat(stats, pcur, "invalidate_wqes",
+		convert_to_64bit(wqe_stats->invalidate_wqes_lo,
+				 wqe_stats->invalidate_wqes_hi));
+	pcur += ocrdma_add_stat(stats, pcur, "dpp_wqe_drops",
+				(u64)wqe_stats->dpp_wqe_drops);
+	return stats;
+}
+
+static char *ocrdma_db_errstats(struct ocrdma_dev *dev)
+{
+	char *stats = dev->stats_mem.debugfs_mem, *pcur;
+	struct ocrdma_rdma_stats_resp *rdma_stats =
+		(struct ocrdma_rdma_stats_resp *)dev->stats_mem.va;
+	struct ocrdma_db_err_stats *db_err_stats = &rdma_stats->db_err_stats;
+
+	memset(stats, 0, (OCRDMA_MAX_DBGFS_MEM));
+
+	pcur = stats;
+	pcur += ocrdma_add_stat(stats, pcur, "sq_doorbell_errors",
+				(u64)db_err_stats->sq_doorbell_errors);
+	pcur += ocrdma_add_stat(stats, pcur, "cq_doorbell_errors",
+				(u64)db_err_stats->cq_doorbell_errors);
+	pcur += ocrdma_add_stat(stats, pcur, "rq_srq_doorbell_errors",
+				(u64)db_err_stats->rq_srq_doorbell_errors);
+	pcur += ocrdma_add_stat(stats, pcur, "cq_overflow_errors",
+				(u64)db_err_stats->cq_overflow_errors);
+	return stats;
+}
+
+static char *ocrdma_rxqp_errstats(struct ocrdma_dev *dev)
+{
+	char *stats = dev->stats_mem.debugfs_mem, *pcur;
+	struct ocrdma_rdma_stats_resp *rdma_stats =
+		(struct ocrdma_rdma_stats_resp *)dev->stats_mem.va;
+	struct ocrdma_rx_qp_err_stats *rx_qp_err_stats =
+		 &rdma_stats->rx_qp_err_stats;
+
+	memset(stats, 0, (OCRDMA_MAX_DBGFS_MEM));
+
+	pcur = stats;
+	pcur += ocrdma_add_stat(stats, pcur, "nak_invalid_requst_errors",
+			(u64)rx_qp_err_stats->nak_invalid_requst_errors);
+	pcur += ocrdma_add_stat(stats, pcur, "nak_remote_operation_errors",
+			(u64)rx_qp_err_stats->nak_remote_operation_errors);
+	pcur += ocrdma_add_stat(stats, pcur, "nak_count_remote_access_errors",
+			(u64)rx_qp_err_stats->nak_count_remote_access_errors);
+	pcur += ocrdma_add_stat(stats, pcur, "local_length_errors",
+			(u64)rx_qp_err_stats->local_length_errors);
+	pcur += ocrdma_add_stat(stats, pcur, "local_protection_errors",
+			(u64)rx_qp_err_stats->local_protection_errors);
+	pcur += ocrdma_add_stat(stats, pcur, "local_qp_operation_errors",
+			(u64)rx_qp_err_stats->local_qp_operation_errors);
+	return stats;
+}
+
+static char *ocrdma_txqp_errstats(struct ocrdma_dev *dev)
+{
+	char *stats = dev->stats_mem.debugfs_mem, *pcur;
+	struct ocrdma_rdma_stats_resp *rdma_stats =
+		(struct ocrdma_rdma_stats_resp *)dev->stats_mem.va;
+	struct ocrdma_tx_qp_err_stats *tx_qp_err_stats =
+		&rdma_stats->tx_qp_err_stats;
+
+	memset(stats, 0, (OCRDMA_MAX_DBGFS_MEM));
+
+	pcur = stats;
+	pcur += ocrdma_add_stat(stats, pcur, "local_length_errors",
+			(u64)tx_qp_err_stats->local_length_errors);
+	pcur += ocrdma_add_stat(stats, pcur, "local_protection_errors",
+			(u64)tx_qp_err_stats->local_protection_errors);
+	pcur += ocrdma_add_stat(stats, pcur, "local_qp_operation_errors",
+			(u64)tx_qp_err_stats->local_qp_operation_errors);
+	pcur += ocrdma_add_stat(stats, pcur, "retry_count_exceeded_errors",
+			(u64)tx_qp_err_stats->retry_count_exceeded_errors);
+	pcur += ocrdma_add_stat(stats, pcur, "rnr_retry_count_exceeded_errors",
+			(u64)tx_qp_err_stats->rnr_retry_count_exceeded_errors);
+	return stats;
+}
+
+static char *ocrdma_tx_dbg_stats(struct ocrdma_dev *dev)
+{
+	int i;
+	char *pstats = dev->stats_mem.debugfs_mem;
+	struct ocrdma_rdma_stats_resp *rdma_stats =
+		(struct ocrdma_rdma_stats_resp *)dev->stats_mem.va;
+	struct ocrdma_tx_dbg_stats *tx_dbg_stats =
+		&rdma_stats->tx_dbg_stats;
+
+	memset(pstats, 0, (OCRDMA_MAX_DBGFS_MEM));
+
+	for (i = 0; i < 100; i++)
+		pstats += snprintf(pstats, 80, "DW[%d] = 0x%x\n", i,
+				 tx_dbg_stats->data[i]);
+
+	return dev->stats_mem.debugfs_mem;
+}
+
+static char *ocrdma_rx_dbg_stats(struct ocrdma_dev *dev)
+{
+	int i;
+	char *pstats = dev->stats_mem.debugfs_mem;
+	struct ocrdma_rdma_stats_resp *rdma_stats =
+		(struct ocrdma_rdma_stats_resp *)dev->stats_mem.va;
+	struct ocrdma_rx_dbg_stats *rx_dbg_stats =
+		&rdma_stats->rx_dbg_stats;
+
+	memset(pstats, 0, (OCRDMA_MAX_DBGFS_MEM));
+
+	for (i = 0; i < 200; i++)
+		pstats += snprintf(pstats, 80, "DW[%d] = 0x%x\n", i,
+				 rx_dbg_stats->data[i]);
+
+	return dev->stats_mem.debugfs_mem;
+}
+
+static void ocrdma_update_stats(struct ocrdma_dev *dev)
+{
+	ulong now = jiffies, secs;
+	int status = 0;
+
+	secs = jiffies_to_msecs(now - dev->last_stats_time) / 1000U;
+	if (secs) {
+		/* update */
+		status = ocrdma_mbx_rdma_stats(dev, false);
+		if (status)
+			pr_err("%s: stats mbox failed with status = %d\n",
+			       __func__, status);
+		dev->last_stats_time = jiffies;
+	}
+}
+
+static ssize_t ocrdma_dbgfs_ops_read(struct file *filp, char __user *buffer,
+					size_t usr_buf_len, loff_t *ppos)
+{
+	struct ocrdma_stats *pstats = filp->private_data;
+	struct ocrdma_dev *dev = pstats->dev;
+	ssize_t status = 0;
+	char *data = NULL;
+
+	/* No partial reads */
+	if (*ppos != 0)
+		return 0;
+
+	mutex_lock(&dev->stats_lock);
+
+	ocrdma_update_stats(dev);
+
+	switch (pstats->type) {
+	case OCRDMA_RSRC_STATS:
+		data = ocrdma_resource_stats(dev);
+		break;
+	case OCRDMA_RXSTATS:
+		data = ocrdma_rx_stats(dev);
+		break;
+	case OCRDMA_WQESTATS:
+		data = ocrdma_wqe_stats(dev);
+		break;
+	case OCRDMA_TXSTATS:
+		data = ocrdma_tx_stats(dev);
+		break;
+	case OCRDMA_DB_ERRSTATS:
+		data = ocrdma_db_errstats(dev);
+		break;
+	case OCRDMA_RXQP_ERRSTATS:
+		data = ocrdma_rxqp_errstats(dev);
+		break;
+	case OCRDMA_TXQP_ERRSTATS:
+		data = ocrdma_txqp_errstats(dev);
+		break;
+	case OCRDMA_TX_DBG_STATS:
+		data = ocrdma_tx_dbg_stats(dev);
+		break;
+	case OCRDMA_RX_DBG_STATS:
+		data = ocrdma_rx_dbg_stats(dev);
+		break;
+
+	default:
+		status = -EFAULT;
+		goto exit;
+	}
+
+	if (usr_buf_len < strlen(data)) {
+		status = -ENOSPC;
+		goto exit;
+	}
+
+	status = simple_read_from_buffer(buffer, usr_buf_len, ppos, data,
+					 strlen(data));
+exit:
+	mutex_unlock(&dev->stats_lock);
+	return status;
+}
+
+static int ocrdma_debugfs_open(struct inode *inode, struct file *file)
+{
+	if (inode->i_private)
+		file->private_data = inode->i_private;
+	return 0;
+}
+
+static const struct file_operations ocrdma_dbg_ops = {
+	.owner = THIS_MODULE,
+	.open = ocrdma_debugfs_open,
+	.read = ocrdma_dbgfs_ops_read,
+};
+
+void ocrdma_add_port_stats(struct ocrdma_dev *dev)
+{
+	if (!ocrdma_dbgfs_dir)
+		return;
+
+	/* Create post stats base dir */
+	dev->dir = debugfs_create_dir(dev->ibdev.name, ocrdma_dbgfs_dir);
+	if (!dev->dir)
+		goto err;
+
+	dev->rsrc_stats.type = OCRDMA_RSRC_STATS;
+	dev->rsrc_stats.dev = dev;
+	if (!debugfs_create_file("resource_stats", S_IRUSR, dev->dir,
+				 &dev->rsrc_stats, &ocrdma_dbg_ops))
+		goto err;
+
+	dev->rx_stats.type = OCRDMA_RXSTATS;
+	dev->rx_stats.dev = dev;
+	if (!debugfs_create_file("rx_stats", S_IRUSR, dev->dir,
+				 &dev->rx_stats, &ocrdma_dbg_ops))
+		goto err;
+
+	dev->wqe_stats.type = OCRDMA_WQESTATS;
+	dev->wqe_stats.dev = dev;
+	if (!debugfs_create_file("wqe_stats", S_IRUSR, dev->dir,
+				 &dev->wqe_stats, &ocrdma_dbg_ops))
+		goto err;
+
+	dev->tx_stats.type = OCRDMA_TXSTATS;
+	dev->tx_stats.dev = dev;
+	if (!debugfs_create_file("tx_stats", S_IRUSR, dev->dir,
+				 &dev->tx_stats, &ocrdma_dbg_ops))
+		goto err;
+
+	dev->db_err_stats.type = OCRDMA_DB_ERRSTATS;
+	dev->db_err_stats.dev = dev;
+	if (!debugfs_create_file("db_err_stats", S_IRUSR, dev->dir,
+				 &dev->db_err_stats, &ocrdma_dbg_ops))
+		goto err;
+
+
+	dev->tx_qp_err_stats.type = OCRDMA_TXQP_ERRSTATS;
+	dev->tx_qp_err_stats.dev = dev;
+	if (!debugfs_create_file("tx_qp_err_stats", S_IRUSR, dev->dir,
+				 &dev->tx_qp_err_stats, &ocrdma_dbg_ops))
+		goto err;
+
+	dev->rx_qp_err_stats.type = OCRDMA_RXQP_ERRSTATS;
+	dev->rx_qp_err_stats.dev = dev;
+	if (!debugfs_create_file("rx_qp_err_stats", S_IRUSR, dev->dir,
+				 &dev->rx_qp_err_stats, &ocrdma_dbg_ops))
+		goto err;
+
+
+	dev->tx_dbg_stats.type = OCRDMA_TX_DBG_STATS;
+	dev->tx_dbg_stats.dev = dev;
+	if (!debugfs_create_file("tx_dbg_stats", S_IRUSR, dev->dir,
+				 &dev->tx_dbg_stats, &ocrdma_dbg_ops))
+		goto err;
+
+	dev->rx_dbg_stats.type = OCRDMA_RX_DBG_STATS;
+	dev->rx_dbg_stats.dev = dev;
+	if (!debugfs_create_file("rx_dbg_stats", S_IRUSR, dev->dir,
+				 &dev->rx_dbg_stats, &ocrdma_dbg_ops))
+		goto err;
+
+	/* Now create dma_mem for stats mbx command */
+	if (!ocrdma_alloc_stats_mem(dev))
+		goto err;
+
+	mutex_init(&dev->stats_lock);
+
+	return;
+err:
+	ocrdma_release_stats_mem(dev);
+	debugfs_remove_recursive(dev->dir);
+	dev->dir = NULL;
+}
+
+void ocrdma_rem_port_stats(struct ocrdma_dev *dev)
+{
+	if (!dev->dir)
+		return;
+	mutex_destroy(&dev->stats_lock);
+	ocrdma_release_stats_mem(dev);
+	debugfs_remove(dev->dir);
+}
+
+void ocrdma_init_debugfs(void)
+{
+	/* Create base dir in debugfs root dir */
+	ocrdma_dbgfs_dir = debugfs_create_dir("ocrdma", NULL);
+}
+
+void ocrdma_rem_debugfs(void)
+{
+	debugfs_remove_recursive(ocrdma_dbgfs_dir);
+}

drivers/infiniband/hw/ocrdma/ocrdma_stats.h

@@ -0,0 +1,54 @@
+/*******************************************************************
+ * This file is part of the Emulex RoCE Device Driver for          *
+ * RoCE (RDMA over Converged Ethernet) adapters.                   *
+ * Copyright (C) 2008-2014 Emulex. All rights reserved.            *
+ * EMULEX and SLI are trademarks of Emulex.                        *
+ * www.emulex.com                                                  *
+ *                                                                 *
+ * This program is free software; you can redistribute it and/or   *
+ * modify it under the terms of version 2 of the GNU General       *
+ * Public License as published by the Free Software Foundation.    *
+ * This program is distributed in the hope that it will be useful. *
+ * ALL EXPRESS OR IMPLIED CONDITIONS, REPRESENTATIONS AND          *
+ * WARRANTIES, INCLUDING ANY IMPLIED WARRANTY OF MERCHANTABILITY,  *
+ * FITNESS FOR A PARTICULAR PURPOSE, OR NON-INFRINGEMENT, ARE      *
+ * DISCLAIMED, EXCEPT TO THE EXTENT THAT SUCH DISCLAIMERS ARE HELD *
+ * TO BE LEGALLY INVALID.  See the GNU General Public License for  *
+ * more details, a copy of which can be found in the file COPYING  *
+ * included with this package.                                     *
+ *
+ * Contact Information:
+ * linux-drivers@emulex.com
+ *
+ * Emulex
+ * 3333 Susan Street
+ * Costa Mesa, CA 92626
+ *******************************************************************/
+
+#ifndef __OCRDMA_STATS_H__
+#define __OCRDMA_STATS_H__
+
+#include <linux/debugfs.h>
+#include "ocrdma.h"
+#include "ocrdma_hw.h"
+
+#define OCRDMA_MAX_DBGFS_MEM 4096
+
+enum OCRDMA_STATS_TYPE {
+	OCRDMA_RSRC_STATS,
+	OCRDMA_RXSTATS,
+	OCRDMA_WQESTATS,
+	OCRDMA_TXSTATS,
+	OCRDMA_DB_ERRSTATS,
+	OCRDMA_RXQP_ERRSTATS,
+	OCRDMA_TXQP_ERRSTATS,
+	OCRDMA_TX_DBG_STATS,
+	OCRDMA_RX_DBG_STATS
+};
+
+void ocrdma_rem_debugfs(void);
+void ocrdma_init_debugfs(void);
+void ocrdma_rem_port_stats(struct ocrdma_dev *dev);
+void ocrdma_add_port_stats(struct ocrdma_dev *dev);
+
+#endif	/* __OCRDMA_STATS_H__ */

drivers/infiniband/hw/ocrdma/ocrdma_verbs.c

@@ -53,7 +53,7 @@ int ocrdma_query_gid(struct ib_device *ibdev, u8 port,
 
 	dev = get_ocrdma_dev(ibdev);
 	memset(sgid, 0, sizeof(*sgid));
-	if (index >= OCRDMA_MAX_SGID)
+	if (index > OCRDMA_MAX_SGID)
 		return -EINVAL;
 
 	memcpy(sgid, &dev->sgid_tbl[index], sizeof(*sgid));
@@ -89,7 +89,7 @@ int ocrdma_query_device(struct ib_device *ibdev, struct ib_device_attr *attr)
 	attr->max_cq = dev->attr.max_cq;
 	attr->max_cqe = dev->attr.max_cqe;
 	attr->max_mr = dev->attr.max_mr;
-	attr->max_mw = 0;
+	attr->max_mw = dev->attr.max_mw;
 	attr->max_pd = dev->attr.max_pd;
 	attr->atomic_cap = 0;
 	attr->max_fmr = 0;
@@ -144,7 +144,6 @@ static inline void get_link_speed_and_width(struct ocrdma_dev *dev,
 	}
 }
 
-
 int ocrdma_query_port(struct ib_device *ibdev,
 		      u8 port, struct ib_port_attr *props)
 {
@@ -267,7 +266,7 @@ static struct ocrdma_pd *_ocrdma_alloc_pd(struct ocrdma_dev *dev,
 
 	if (udata && uctx) {
 		pd->dpp_enabled =
-			dev->nic_info.dev_family == OCRDMA_GEN2_FAMILY;
+			ocrdma_get_asic_type(dev) == OCRDMA_ASIC_GEN_SKH_R;
 		pd->num_dpp_qp =
 			pd->dpp_enabled ? OCRDMA_PD_MAX_DPP_ENABLED_QP : 0;
 	}
@@ -726,10 +725,10 @@ static void build_user_pbes(struct ocrdma_dev *dev, struct ocrdma_mr *mr,
 			    u32 num_pbes)
 {
 	struct ocrdma_pbe *pbe;
-	struct ib_umem_chunk *chunk;
+	struct scatterlist *sg;
 	struct ocrdma_pbl *pbl_tbl = mr->hwmr.pbl_table;
 	struct ib_umem *umem = mr->umem;
-	int i, shift, pg_cnt, pages, pbe_cnt, total_num_pbes = 0;
+	int shift, pg_cnt, pages, pbe_cnt, entry, total_num_pbes = 0;
 
 	if (!mr->hwmr.num_pbes)
 		return;
@@ -739,39 +738,37 @@ static void build_user_pbes(struct ocrdma_dev *dev, struct ocrdma_mr *mr,
 
 	shift = ilog2(umem->page_size);
 
-	list_for_each_entry(chunk, &umem->chunk_list, list) {
-		/* get all the dma regions from the chunk. */
-		for (i = 0; i < chunk->nmap; i++) {
-			pages = sg_dma_len(&chunk->page_list[i]) >> shift;
-			for (pg_cnt = 0; pg_cnt < pages; pg_cnt++) {
-				/* store the page address in pbe */
-				pbe->pa_lo =
-				    cpu_to_le32(sg_dma_address
-						(&chunk->page_list[i]) +
-						(umem->page_size * pg_cnt));
-				pbe->pa_hi =
-				    cpu_to_le32(upper_32_bits
-						((sg_dma_address
-						  (&chunk->page_list[i]) +
-						  umem->page_size * pg_cnt)));
-				pbe_cnt += 1;
-				total_num_pbes += 1;
-				pbe++;
-
-				/* if done building pbes, issue the mbx cmd. */
-				if (total_num_pbes == num_pbes)
-					return;
-
-				/* if the given pbl is full storing the pbes,
-				 * move to next pbl.
-				 */
-				if (pbe_cnt ==
-					(mr->hwmr.pbl_size / sizeof(u64))) {
-					pbl_tbl++;
-					pbe = (struct ocrdma_pbe *)pbl_tbl->va;
-					pbe_cnt = 0;
-				}
+	for_each_sg(umem->sg_head.sgl, sg, umem->nmap, entry) {
+		pages = sg_dma_len(sg) >> shift;
+		for (pg_cnt = 0; pg_cnt < pages; pg_cnt++) {
+			/* store the page address in pbe */
+			pbe->pa_lo =
+			    cpu_to_le32(sg_dma_address
+					(sg) +
+					(umem->page_size * pg_cnt));
+			pbe->pa_hi =
+			    cpu_to_le32(upper_32_bits
+					((sg_dma_address
+					  (sg) +
+					  umem->page_size * pg_cnt)));
+			pbe_cnt += 1;
+			total_num_pbes += 1;
+			pbe++;
+
+			/* if done building pbes, issue the mbx cmd. */
+			if (total_num_pbes == num_pbes)
+				return;
+
+			/* if the given pbl is full storing the pbes,
+			 * move to next pbl.
+			 */
+			if (pbe_cnt ==
+				(mr->hwmr.pbl_size / sizeof(u64))) {
+				pbl_tbl++;
+				pbe = (struct ocrdma_pbe *)pbl_tbl->va;
+				pbe_cnt = 0;
 			}
+
 		}
 	}
 }
@@ -840,8 +837,7 @@ int ocrdma_dereg_mr(struct ib_mr *ib_mr)
 
 	status = ocrdma_mbx_dealloc_lkey(dev, mr->hwmr.fr_mr, mr->hwmr.lkey);
 
-	if (mr->hwmr.fr_mr == 0)
-		ocrdma_free_mr_pbl_tbl(dev, &mr->hwmr);
+	ocrdma_free_mr_pbl_tbl(dev, &mr->hwmr);
 
 	/* it could be user registered memory. */
 	if (mr->umem)
@@ -910,6 +906,7 @@ struct ib_cq *ocrdma_create_cq(struct ib_device *ibdev, int entries, int vector,
 	spin_lock_init(&cq->comp_handler_lock);
 	INIT_LIST_HEAD(&cq->sq_head);
 	INIT_LIST_HEAD(&cq->rq_head);
+	cq->first_arm = true;
 
 	if (ib_ctx) {
 		uctx = get_ocrdma_ucontext(ib_ctx);
@@ -927,9 +924,7 @@ struct ib_cq *ocrdma_create_cq(struct ib_device *ibdev, int entries, int vector,
 			goto ctx_err;
 	}
 	cq->phase = OCRDMA_CQE_VALID;
-	cq->arm_needed = true;
 	dev->cq_tbl[cq->id] = cq;
-
 	return &cq->ibcq;
 
 ctx_err:
@@ -952,15 +947,52 @@ int ocrdma_resize_cq(struct ib_cq *ibcq, int new_cnt,
 	return status;
 }
 
+static void ocrdma_flush_cq(struct ocrdma_cq *cq)
+{
+	int cqe_cnt;
+	int valid_count = 0;
+	unsigned long flags;
+
+	struct ocrdma_dev *dev = get_ocrdma_dev(cq->ibcq.device);
+	struct ocrdma_cqe *cqe = NULL;
+
+	cqe = cq->va;
+	cqe_cnt = cq->cqe_cnt;
+
+	/* Last irq might have scheduled a polling thread
+	 * sync-up with it before hard flushing.
+	 */
+	spin_lock_irqsave(&cq->cq_lock, flags);
+	while (cqe_cnt) {
+		if (is_cqe_valid(cq, cqe))
+			valid_count++;
+		cqe++;
+		cqe_cnt--;
+	}
+	ocrdma_ring_cq_db(dev, cq->id, false, false, valid_count);
+	spin_unlock_irqrestore(&cq->cq_lock, flags);
+}
+
 int ocrdma_destroy_cq(struct ib_cq *ibcq)
 {
 	int status;
 	struct ocrdma_cq *cq = get_ocrdma_cq(ibcq);
+	struct ocrdma_eq *eq = NULL;
 	struct ocrdma_dev *dev = get_ocrdma_dev(ibcq->device);
 	int pdid = 0;
+	u32 irq, indx;
 
-	status = ocrdma_mbx_destroy_cq(dev, cq);
+	dev->cq_tbl[cq->id] = NULL;
+	indx = ocrdma_get_eq_table_index(dev, cq->eqn);
+	if (indx == -EINVAL)
+		BUG();
 
+	eq = &dev->eq_tbl[indx];
+	irq = ocrdma_get_irq(dev, eq);
+	synchronize_irq(irq);
+	ocrdma_flush_cq(cq);
+
+	status = ocrdma_mbx_destroy_cq(dev, cq);
 	if (cq->ucontext) {
 		pdid = cq->ucontext->cntxt_pd->id;
 		ocrdma_del_mmap(cq->ucontext, (u64) cq->pa,
@@ -969,7 +1001,6 @@ int ocrdma_destroy_cq(struct ib_cq *ibcq)
 				ocrdma_get_db_addr(dev, pdid),
 				dev->nic_info.db_page_size);
 	}
-	dev->cq_tbl[cq->id] = NULL;
 
 	kfree(cq);
 	return status;
@@ -1092,15 +1123,9 @@ static int ocrdma_copy_qp_uresp(struct ocrdma_qp *qp,
 	}
 	uresp.db_page_addr = usr_db;
 	uresp.db_page_size = dev->nic_info.db_page_size;
-	if (dev->nic_info.dev_family == OCRDMA_GEN2_FAMILY) {
-		uresp.db_sq_offset = OCRDMA_DB_GEN2_SQ_OFFSET;
-		uresp.db_rq_offset = OCRDMA_DB_GEN2_RQ_OFFSET;
-		uresp.db_shift = 24;
-	} else {
-		uresp.db_sq_offset = OCRDMA_DB_SQ_OFFSET;
-		uresp.db_rq_offset = OCRDMA_DB_RQ_OFFSET;
-		uresp.db_shift = 16;
-	}
+	uresp.db_sq_offset = OCRDMA_DB_GEN2_SQ_OFFSET;
+	uresp.db_rq_offset = OCRDMA_DB_GEN2_RQ_OFFSET;
+	uresp.db_shift = OCRDMA_DB_RQ_SHIFT;
 
 	if (qp->dpp_enabled) {
 		uresp.dpp_credit = dpp_credit_lmt;
@@ -1132,7 +1157,7 @@ err:
 static void ocrdma_set_qp_db(struct ocrdma_dev *dev, struct ocrdma_qp *qp,
 			     struct ocrdma_pd *pd)
 {
-	if (dev->nic_info.dev_family == OCRDMA_GEN2_FAMILY) {
+	if (ocrdma_get_asic_type(dev) == OCRDMA_ASIC_GEN_SKH_R) {
 		qp->sq_db = dev->nic_info.db +
 			(pd->id * dev->nic_info.db_page_size) +
 			OCRDMA_DB_GEN2_SQ_OFFSET;
@@ -1182,7 +1207,6 @@ static void ocrdma_set_qp_init_params(struct ocrdma_qp *qp,
 	qp->signaled = (attrs->sq_sig_type == IB_SIGNAL_ALL_WR) ? true : false;
 }
 
-
 static void ocrdma_store_gsi_qp_cq(struct ocrdma_dev *dev,
 				   struct ib_qp_init_attr *attrs)
 {
@@ -1268,17 +1292,6 @@ gen_err:
 	return ERR_PTR(status);
 }
 
-
-static void ocrdma_flush_rq_db(struct ocrdma_qp *qp)
-{
-	if (qp->db_cache) {
-		u32 val = qp->rq.dbid | (qp->db_cache <<
-				ocrdma_get_num_posted_shift(qp));
-		iowrite32(val, qp->rq_db);
-		qp->db_cache = 0;
-	}
-}
-
 int _ocrdma_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
 		      int attr_mask)
 {
@@ -1296,9 +1309,7 @@ int _ocrdma_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
 	 */
 	if (status < 0)
 		return status;
-	status = ocrdma_mbx_modify_qp(dev, qp, attr, attr_mask, old_qps);
-	if (!status && attr_mask & IB_QP_STATE && attr->qp_state == IB_QPS_RTR)
-		ocrdma_flush_rq_db(qp);
+	status = ocrdma_mbx_modify_qp(dev, qp, attr, attr_mask);
 
 	return status;
 }
@@ -1510,7 +1521,7 @@ static void ocrdma_discard_cqes(struct ocrdma_qp *qp, struct ocrdma_cq *cq)
 	int discard_cnt = 0;
 	u32 cur_getp, stop_getp;
 	struct ocrdma_cqe *cqe;
-	u32 qpn = 0;
+	u32 qpn = 0, wqe_idx = 0;
 
 	spin_lock_irqsave(&cq->cq_lock, cq_flags);
 
@@ -1539,24 +1550,29 @@ static void ocrdma_discard_cqes(struct ocrdma_qp *qp, struct ocrdma_cq *cq)
 		if (qpn == 0 || qpn != qp->id)
 			goto skip_cqe;
 
-		/* mark cqe discarded so that it is not picked up later
-		 * in the poll_cq().
-		 */
-		discard_cnt += 1;
-		cqe->cmn.qpn = 0;
 		if (is_cqe_for_sq(cqe)) {
 			ocrdma_hwq_inc_tail(&qp->sq);
 		} else {
 			if (qp->srq) {
+				wqe_idx = (le32_to_cpu(cqe->rq.buftag_qpn) >>
+					OCRDMA_CQE_BUFTAG_SHIFT) &
+					qp->srq->rq.max_wqe_idx;
+				if (wqe_idx < 1)
+					BUG();
 				spin_lock_irqsave(&qp->srq->q_lock, flags);
 				ocrdma_hwq_inc_tail(&qp->srq->rq);
-				ocrdma_srq_toggle_bit(qp->srq, cur_getp);
+				ocrdma_srq_toggle_bit(qp->srq, wqe_idx - 1);
 				spin_unlock_irqrestore(&qp->srq->q_lock, flags);
 
 			} else {
 				ocrdma_hwq_inc_tail(&qp->rq);
 			}
 		}
+		/* mark cqe discarded so that it is not picked up later
+		 * in the poll_cq().
+		 */
+		discard_cnt += 1;
+		cqe->cmn.qpn = 0;
 skip_cqe:
 		cur_getp = (cur_getp + 1) % cq->max_hw_cqe;
 	} while (cur_getp != stop_getp);
@@ -1659,7 +1675,7 @@ static int ocrdma_copy_srq_uresp(struct ocrdma_dev *dev, struct ocrdma_srq *srq,
 	    (srq->pd->id * dev->nic_info.db_page_size);
 	uresp.db_page_size = dev->nic_info.db_page_size;
 	uresp.num_rqe_allocated = srq->rq.max_cnt;
-	if (dev->nic_info.dev_family == OCRDMA_GEN2_FAMILY) {
+	if (ocrdma_get_asic_type(dev) == OCRDMA_ASIC_GEN_SKH_R) {
 		uresp.db_rq_offset = OCRDMA_DB_GEN2_RQ_OFFSET;
 		uresp.db_shift = 24;
 	} else {
@@ -2009,15 +2025,15 @@ static int ocrdma_build_fr(struct ocrdma_qp *qp, struct ocrdma_hdr_wqe *hdr,
 	fast_reg->num_sges = wr->wr.fast_reg.page_list_len;
 	fast_reg->size_sge =
 		get_encoded_page_size(1 << wr->wr.fast_reg.page_shift);
-	mr = (struct ocrdma_mr *) (unsigned long) qp->dev->stag_arr[(hdr->lkey >> 8) &
-		(OCRDMA_MAX_STAG - 1)];
+	mr = (struct ocrdma_mr *) (unsigned long)
+		qp->dev->stag_arr[(hdr->lkey >> 8) & (OCRDMA_MAX_STAG - 1)];
 	build_frmr_pbes(wr, mr->hwmr.pbl_table, &mr->hwmr);
 	return 0;
 }
 
 static void ocrdma_ring_sq_db(struct ocrdma_qp *qp)
 {
-	u32 val = qp->sq.dbid | (1 << 16);
+	u32 val = qp->sq.dbid | (1 << OCRDMA_DB_SQ_SHIFT);
 
 	iowrite32(val, qp->sq_db);
 }
@@ -2122,12 +2138,9 @@ int ocrdma_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 
 static void ocrdma_ring_rq_db(struct ocrdma_qp *qp)
 {
-	u32 val = qp->rq.dbid | (1 << ocrdma_get_num_posted_shift(qp));
+	u32 val = qp->rq.dbid | (1 << OCRDMA_DB_RQ_SHIFT);
 
-	if (qp->state != OCRDMA_QPS_INIT)
-		iowrite32(val, qp->rq_db);
-	else
-		qp->db_cache++;
+	iowrite32(val, qp->rq_db);
 }
 
 static void ocrdma_build_rqe(struct ocrdma_hdr_wqe *rqe, struct ib_recv_wr *wr,
@@ -2213,7 +2226,7 @@ static int ocrdma_srq_get_idx(struct ocrdma_srq *srq)
 
 	if (row == srq->bit_fields_len)
 		BUG();
-	return indx;
+	return indx + 1; /* Use from index 1 */
 }
 
 static void ocrdma_ring_srq_db(struct ocrdma_srq *srq)
@@ -2550,10 +2563,13 @@ static void ocrdma_update_free_srq_cqe(struct ib_wc *ibwc,
 
 	srq = get_ocrdma_srq(qp->ibqp.srq);
 	wqe_idx = (le32_to_cpu(cqe->rq.buftag_qpn) >>
-			OCRDMA_CQE_BUFTAG_SHIFT) & srq->rq.max_wqe_idx;
+		OCRDMA_CQE_BUFTAG_SHIFT) & srq->rq.max_wqe_idx;
+	if (wqe_idx < 1)
+		BUG();
+
 	ibwc->wr_id = srq->rqe_wr_id_tbl[wqe_idx];
 	spin_lock_irqsave(&srq->q_lock, flags);
-	ocrdma_srq_toggle_bit(srq, wqe_idx);
+	ocrdma_srq_toggle_bit(srq, wqe_idx - 1);
 	spin_unlock_irqrestore(&srq->q_lock, flags);
 	ocrdma_hwq_inc_tail(&srq->rq);
 }
@@ -2705,10 +2721,18 @@ expand_cqe:
 	}
 stop_cqe:
 	cq->getp = cur_getp;
-	if (polled_hw_cqes || expand || stop) {
-		ocrdma_ring_cq_db(dev, cq->id, cq->armed, cq->solicited,
+	if (cq->deferred_arm) {
+		ocrdma_ring_cq_db(dev, cq->id, true, cq->deferred_sol,
+				  polled_hw_cqes);
+		cq->deferred_arm = false;
+		cq->deferred_sol = false;
+	} else {
+		/* We need to pop the CQE. No need to arm */
+		ocrdma_ring_cq_db(dev, cq->id, false, cq->deferred_sol,
 				  polled_hw_cqes);
+		cq->deferred_sol = false;
 	}
+
 	return i;
 }
 
@@ -2780,30 +2804,28 @@ int ocrdma_arm_cq(struct ib_cq *ibcq, enum ib_cq_notify_flags cq_flags)
 	struct ocrdma_cq *cq = get_ocrdma_cq(ibcq);
 	struct ocrdma_dev *dev = get_ocrdma_dev(ibcq->device);
 	u16 cq_id;
-	u16 cur_getp;
-	struct ocrdma_cqe *cqe;
 	unsigned long flags;
+	bool arm_needed = false, sol_needed = false;
 
 	cq_id = cq->id;
 
 	spin_lock_irqsave(&cq->cq_lock, flags);
 	if (cq_flags & IB_CQ_NEXT_COMP || cq_flags & IB_CQ_SOLICITED)
-		cq->armed = true;
+		arm_needed = true;
 	if (cq_flags & IB_CQ_SOLICITED)
-		cq->solicited = true;
-
-	cur_getp = cq->getp;
-	cqe = cq->va + cur_getp;
+		sol_needed = true;
 
-	/* check whether any valid cqe exist or not, if not then safe to
-	 * arm. If cqe is not yet consumed, then let it get consumed and then
-	 * we arm it to avoid false interrupts.
-	 */
-	if (!is_cqe_valid(cq, cqe) || cq->arm_needed) {
-		cq->arm_needed = false;
-		ocrdma_ring_cq_db(dev, cq_id, cq->armed, cq->solicited, 0);
+	if (cq->first_arm) {
+		ocrdma_ring_cq_db(dev, cq_id, arm_needed, sol_needed, 0);
+		cq->first_arm = false;
+		goto skip_defer;
 	}
+	cq->deferred_arm = true;
+
+skip_defer:
+	cq->deferred_sol = sol_needed;
 	spin_unlock_irqrestore(&cq->cq_lock, flags);
+
 	return 0;
 }
 
@@ -2838,7 +2860,8 @@ struct ib_mr *ocrdma_alloc_frmr(struct ib_pd *ibpd, int max_page_list_len)
 		goto mbx_err;
 	mr->ibmr.rkey = mr->hwmr.lkey;
 	mr->ibmr.lkey = mr->hwmr.lkey;
-	dev->stag_arr[(mr->hwmr.lkey >> 8) & (OCRDMA_MAX_STAG - 1)] = mr;
+	dev->stag_arr[(mr->hwmr.lkey >> 8) & (OCRDMA_MAX_STAG - 1)] =
+		(unsigned long) mr;
 	return &mr->ibmr;
 mbx_err:
 	ocrdma_free_mr_pbl_tbl(dev, &mr->hwmr);

drivers/infiniband/hw/qib/qib.h

@@ -868,8 +868,10 @@ struct qib_devdata {
 	/* last buffer for user use */
 	u32 lastctxt_piobuf;
 
-	/* saturating counter of (non-port-specific) device interrupts */
-	u32 int_counter;
+	/* reset value */
+	u64 z_int_counter;
+	/* percpu intcounter */
+	u64 __percpu *int_counter;
 
 	/* pio bufs allocated per ctxt */
 	u32 pbufsctxt;
@@ -1184,7 +1186,7 @@ int qib_setup_eagerbufs(struct qib_ctxtdata *);
 void qib_set_ctxtcnt(struct qib_devdata *);
 int qib_create_ctxts(struct qib_devdata *dd);
 struct qib_ctxtdata *qib_create_ctxtdata(struct qib_pportdata *, u32, int);
-void qib_init_pportdata(struct qib_pportdata *, struct qib_devdata *, u8, u8);
+int qib_init_pportdata(struct qib_pportdata *, struct qib_devdata *, u8, u8);
 void qib_free_ctxtdata(struct qib_devdata *, struct qib_ctxtdata *);
 
 u32 qib_kreceive(struct qib_ctxtdata *, u32 *, u32 *);
@@ -1449,6 +1451,10 @@ void qib_nomsi(struct qib_devdata *);
 void qib_nomsix(struct qib_devdata *);
 void qib_pcie_getcmd(struct qib_devdata *, u16 *, u8 *, u8 *);
 void qib_pcie_reenable(struct qib_devdata *, u16, u8, u8);
+/* interrupts for device */
+u64 qib_int_counter(struct qib_devdata *);
+/* interrupt for all devices */
+u64 qib_sps_ints(void);
 
 /*
  * dma_addr wrappers - all 0's invalid for hw

drivers/infiniband/hw/qib/qib_diag.c

@@ -546,7 +546,7 @@ static ssize_t qib_diagpkt_write(struct file *fp,
 				 size_t count, loff_t *off)
 {
 	u32 __iomem *piobuf;
-	u32 plen, clen, pbufn;
+	u32 plen, pbufn, maxlen_reserve;
 	struct qib_diag_xpkt dp;
 	u32 *tmpbuf = NULL;
 	struct qib_devdata *dd;
@@ -590,15 +590,20 @@ static ssize_t qib_diagpkt_write(struct file *fp,
 	}
 	ppd = &dd->pport[dp.port - 1];
 
-	/* need total length before first word written */
-	/* +1 word is for the qword padding */
-	plen = sizeof(u32) + dp.len;
-	clen = dp.len >> 2;
-
-	if ((plen + 4) > ppd->ibmaxlen) {
+	/*
+	 * need total length before first word written, plus 2 Dwords. One Dword
+	 * is for padding so we get the full user data when not aligned on
+	 * a word boundary. The other Dword is to make sure we have room for the
+	 * ICRC which gets tacked on later.
+	 */
+	maxlen_reserve = 2 * sizeof(u32);
+	if (dp.len > ppd->ibmaxlen - maxlen_reserve) {
 		ret = -EINVAL;
-		goto bail;      /* before writing pbc */
+		goto bail;
 	}
+
+	plen = sizeof(u32) + dp.len;
+
 	tmpbuf = vmalloc(plen);
 	if (!tmpbuf) {
 		qib_devinfo(dd->pcidev,
@@ -638,11 +643,11 @@ static ssize_t qib_diagpkt_write(struct file *fp,
 	 */
 	if (dd->flags & QIB_PIO_FLUSH_WC) {
 		qib_flush_wc();
-		qib_pio_copy(piobuf + 2, tmpbuf, clen - 1);
+		qib_pio_copy(piobuf + 2, tmpbuf, plen - 1);
 		qib_flush_wc();
-		__raw_writel(tmpbuf[clen - 1], piobuf + clen + 1);
+		__raw_writel(tmpbuf[plen - 1], piobuf + plen + 1);
 	} else
-		qib_pio_copy(piobuf + 2, tmpbuf, clen);
+		qib_pio_copy(piobuf + 2, tmpbuf, plen);
 
 	if (dd->flags & QIB_USE_SPCL_TRIG) {
 		u32 spcl_off = (pbufn >= dd->piobcnt2k) ? 2047 : 1023;
@@ -689,28 +694,23 @@ int qib_register_observer(struct qib_devdata *dd,
 			  const struct diag_observer *op)
 {
 	struct diag_observer_list_elt *olp;
-	int ret = -EINVAL;
+	unsigned long flags;
 
 	if (!dd || !op)
-		goto bail;
-	ret = -ENOMEM;
+		return -EINVAL;
 	olp = vmalloc(sizeof *olp);
 	if (!olp) {
 		pr_err("vmalloc for observer failed\n");
-		goto bail;
+		return -ENOMEM;
 	}
-	if (olp) {
-		unsigned long flags;
 
-		spin_lock_irqsave(&dd->qib_diag_trans_lock, flags);
-		olp->op = op;
-		olp->next = dd->diag_observer_list;
-		dd->diag_observer_list = olp;
-		spin_unlock_irqrestore(&dd->qib_diag_trans_lock, flags);
-		ret = 0;
-	}
-bail:
-	return ret;
+	spin_lock_irqsave(&dd->qib_diag_trans_lock, flags);
+	olp->op = op;
+	olp->next = dd->diag_observer_list;
+	dd->diag_observer_list = olp;
+	spin_unlock_irqrestore(&dd->qib_diag_trans_lock, flags);
+
+	return 0;
 }
 
 /* Remove all registered observers when device is closed */

drivers/infiniband/hw/qib/qib_dma.c

@@ -108,6 +108,10 @@ static int qib_map_sg(struct ib_device *dev, struct scatterlist *sgl,
 			ret = 0;
 			break;
 		}
+		sg->dma_address = addr + sg->offset;
+#ifdef CONFIG_NEED_SG_DMA_LENGTH
+		sg->dma_length = sg->length;
+#endif
 	}
 	return ret;
 }
@@ -119,21 +123,6 @@ static void qib_unmap_sg(struct ib_device *dev,
 	BUG_ON(!valid_dma_direction(direction));
 }
 
-static u64 qib_sg_dma_address(struct ib_device *dev, struct scatterlist *sg)
-{
-	u64 addr = (u64) page_address(sg_page(sg));
-
-	if (addr)
-		addr += sg->offset;
-	return addr;
-}
-
-static unsigned int qib_sg_dma_len(struct ib_device *dev,
-				   struct scatterlist *sg)
-{
-	return sg->length;
-}
-
 static void qib_sync_single_for_cpu(struct ib_device *dev, u64 addr,
 				    size_t size, enum dma_data_direction dir)
 {
@@ -173,8 +162,6 @@ struct ib_dma_mapping_ops qib_dma_mapping_ops = {
 	.unmap_page = qib_dma_unmap_page,
 	.map_sg = qib_map_sg,
 	.unmap_sg = qib_unmap_sg,
-	.dma_address = qib_sg_dma_address,
-	.dma_len = qib_sg_dma_len,
 	.sync_single_for_cpu = qib_sync_single_for_cpu,
 	.sync_single_for_device = qib_sync_single_for_device,
 	.alloc_coherent = qib_dma_alloc_coherent,

drivers/infiniband/hw/qib/qib_file_ops.c

@@ -1459,7 +1459,7 @@ static int get_a_ctxt(struct file *fp, const struct qib_user_info *uinfo,
 					cused++;
 				else
 					cfree++;
-			if (pusable && cfree && cused < inuse) {
+			if (cfree && cused < inuse) {
 				udd = dd;
 				inuse = cused;
 			}
@@ -1578,7 +1578,7 @@ static int do_qib_user_sdma_queue_create(struct file *fp)
 	struct qib_ctxtdata *rcd = fd->rcd;
 	struct qib_devdata *dd = rcd->dd;
 
-	if (dd->flags & QIB_HAS_SEND_DMA)
+	if (dd->flags & QIB_HAS_SEND_DMA) {
 
 		fd->pq = qib_user_sdma_queue_create(&dd->pcidev->dev,
 						    dd->unit,
@@ -1586,6 +1586,7 @@ static int do_qib_user_sdma_queue_create(struct file *fp)
 						    fd->subctxt);
 		if (!fd->pq)
 			return -ENOMEM;
+	}
 
 	return 0;
 }

drivers/infiniband/hw/qib/qib_fs.c

@@ -105,6 +105,7 @@ static int create_file(const char *name, umode_t mode,
 static ssize_t driver_stats_read(struct file *file, char __user *buf,
 				 size_t count, loff_t *ppos)
 {
+	qib_stats.sps_ints = qib_sps_ints();
 	return simple_read_from_buffer(buf, count, ppos, &qib_stats,
 				       sizeof qib_stats);
 }

drivers/infiniband/hw/qib/qib_iba6120.c

@@ -1634,9 +1634,7 @@ static irqreturn_t qib_6120intr(int irq, void *data)
 		goto bail;
 	}
 
-	qib_stats.sps_ints++;
-	if (dd->int_counter != (u32) -1)
-		dd->int_counter++;
+	this_cpu_inc(*dd->int_counter);
 
 	if (unlikely(istat & (~QLOGIC_IB_I_BITSEXTANT |
 			      QLOGIC_IB_I_GPIO | QLOGIC_IB_I_ERROR)))
@@ -1808,7 +1806,8 @@ static int qib_6120_setup_reset(struct qib_devdata *dd)
 	 * isn't set.
 	 */
 	dd->flags &= ~(QIB_INITTED | QIB_PRESENT);
-	dd->int_counter = 0; /* so we check interrupts work again */
+	/* so we check interrupts work again */
+	dd->z_int_counter = qib_int_counter(dd);
 	val = dd->control | QLOGIC_IB_C_RESET;
 	writeq(val, &dd->kregbase[kr_control]);
 	mb(); /* prevent compiler re-ordering around actual reset */
@@ -3266,7 +3265,9 @@ static int init_6120_variables(struct qib_devdata *dd)
 
 	dd->eep_st_masks[2].errs_to_log = ERR_MASK(ResetNegated);
 
-	qib_init_pportdata(ppd, dd, 0, 1);
+	ret = qib_init_pportdata(ppd, dd, 0, 1);
+	if (ret)
+		goto bail;
 	ppd->link_width_supported = IB_WIDTH_1X | IB_WIDTH_4X;
 	ppd->link_speed_supported = QIB_IB_SDR;
 	ppd->link_width_enabled = IB_WIDTH_4X;

drivers/infiniband/hw/qib/qib_iba7220.c

@@ -1962,10 +1962,7 @@ static irqreturn_t qib_7220intr(int irq, void *data)
 		goto bail;
 	}
 
-	qib_stats.sps_ints++;
-	if (dd->int_counter != (u32) -1)
-		dd->int_counter++;
-
+	this_cpu_inc(*dd->int_counter);
 	if (unlikely(istat & (~QLOGIC_IB_I_BITSEXTANT |
 			      QLOGIC_IB_I_GPIO | QLOGIC_IB_I_ERROR)))
 		unlikely_7220_intr(dd, istat);
@@ -2120,7 +2117,8 @@ static int qib_setup_7220_reset(struct qib_devdata *dd)
 	 * isn't set.
 	 */
 	dd->flags &= ~(QIB_INITTED | QIB_PRESENT);
-	dd->int_counter = 0; /* so we check interrupts work again */
+	/* so we check interrupts work again */
+	dd->z_int_counter = qib_int_counter(dd);
 	val = dd->control | QLOGIC_IB_C_RESET;
 	writeq(val, &dd->kregbase[kr_control]);
 	mb(); /* prevent compiler reordering around actual reset */
@@ -4061,7 +4059,9 @@ static int qib_init_7220_variables(struct qib_devdata *dd)
 	init_waitqueue_head(&cpspec->autoneg_wait);
 	INIT_DELAYED_WORK(&cpspec->autoneg_work, autoneg_7220_work);
 
-	qib_init_pportdata(ppd, dd, 0, 1);
+	ret = qib_init_pportdata(ppd, dd, 0, 1);
+	if (ret)
+		goto bail;
 	ppd->link_width_supported = IB_WIDTH_1X | IB_WIDTH_4X;
 	ppd->link_speed_supported = QIB_IB_SDR | QIB_IB_DDR;
 

drivers/infiniband/hw/qib/qib_iba7322.c

@@ -3115,9 +3115,7 @@ static irqreturn_t qib_7322intr(int irq, void *data)
 		goto bail;
 	}
 
-	qib_stats.sps_ints++;
-	if (dd->int_counter != (u32) -1)
-		dd->int_counter++;
+	this_cpu_inc(*dd->int_counter);
 
 	/* handle "errors" of various kinds first, device ahead of port */
 	if (unlikely(istat & (~QIB_I_BITSEXTANT | QIB_I_GPIO |
@@ -3186,9 +3184,7 @@ static irqreturn_t qib_7322pintr(int irq, void *data)
 		 */
 		return IRQ_HANDLED;
 
-	qib_stats.sps_ints++;
-	if (dd->int_counter != (u32) -1)
-		dd->int_counter++;
+	this_cpu_inc(*dd->int_counter);
 
 	/* Clear the interrupt bit we expect to be set. */
 	qib_write_kreg(dd, kr_intclear, ((1ULL << QIB_I_RCVAVAIL_LSB) |
@@ -3215,9 +3211,7 @@ static irqreturn_t qib_7322bufavail(int irq, void *data)
 		 */
 		return IRQ_HANDLED;
 
-	qib_stats.sps_ints++;
-	if (dd->int_counter != (u32) -1)
-		dd->int_counter++;
+	this_cpu_inc(*dd->int_counter);
 
 	/* Clear the interrupt bit we expect to be set. */
 	qib_write_kreg(dd, kr_intclear, QIB_I_SPIOBUFAVAIL);
@@ -3248,9 +3242,7 @@ static irqreturn_t sdma_intr(int irq, void *data)
 		 */
 		return IRQ_HANDLED;
 
-	qib_stats.sps_ints++;
-	if (dd->int_counter != (u32) -1)
-		dd->int_counter++;
+	this_cpu_inc(*dd->int_counter);
 
 	/* Clear the interrupt bit we expect to be set. */
 	qib_write_kreg(dd, kr_intclear, ppd->hw_pidx ?
@@ -3277,9 +3269,7 @@ static irqreturn_t sdma_idle_intr(int irq, void *data)
 		 */
 		return IRQ_HANDLED;
 
-	qib_stats.sps_ints++;
-	if (dd->int_counter != (u32) -1)
-		dd->int_counter++;
+	this_cpu_inc(*dd->int_counter);
 
 	/* Clear the interrupt bit we expect to be set. */
 	qib_write_kreg(dd, kr_intclear, ppd->hw_pidx ?
@@ -3306,9 +3296,7 @@ static irqreturn_t sdma_progress_intr(int irq, void *data)
 		 */
 		return IRQ_HANDLED;
 
-	qib_stats.sps_ints++;
-	if (dd->int_counter != (u32) -1)
-		dd->int_counter++;
+	this_cpu_inc(*dd->int_counter);
 
 	/* Clear the interrupt bit we expect to be set. */
 	qib_write_kreg(dd, kr_intclear, ppd->hw_pidx ?
@@ -3336,9 +3324,7 @@ static irqreturn_t sdma_cleanup_intr(int irq, void *data)
 		 */
 		return IRQ_HANDLED;
 
-	qib_stats.sps_ints++;
-	if (dd->int_counter != (u32) -1)
-		dd->int_counter++;
+	this_cpu_inc(*dd->int_counter);
 
 	/* Clear the interrupt bit we expect to be set. */
 	qib_write_kreg(dd, kr_intclear, ppd->hw_pidx ?
@@ -3723,7 +3709,8 @@ static int qib_do_7322_reset(struct qib_devdata *dd)
 	dd->pport->cpspec->ibsymdelta = 0;
 	dd->pport->cpspec->iblnkerrdelta = 0;
 	dd->pport->cpspec->ibmalfdelta = 0;
-	dd->int_counter = 0; /* so we check interrupts work again */
+	/* so we check interrupts work again */
+	dd->z_int_counter = qib_int_counter(dd);
 
 	/*
 	 * Keep chip from being accessed until we are ready.  Use
@@ -6557,7 +6544,11 @@ static int qib_init_7322_variables(struct qib_devdata *dd)
 		}
 
 		dd->num_pports++;
-		qib_init_pportdata(ppd, dd, pidx, dd->num_pports);
+		ret = qib_init_pportdata(ppd, dd, pidx, dd->num_pports);
+		if (ret) {
+			dd->num_pports--;
+			goto bail;
+		}
 
 		ppd->link_width_supported = IB_WIDTH_1X | IB_WIDTH_4X;
 		ppd->link_width_enabled = IB_WIDTH_4X;

drivers/infiniband/hw/qib/qib_init.c

@@ -130,7 +130,6 @@ void qib_set_ctxtcnt(struct qib_devdata *dd)
 int qib_create_ctxts(struct qib_devdata *dd)
 {
 	unsigned i;
-	int ret;
 	int local_node_id = pcibus_to_node(dd->pcidev->bus);
 
 	if (local_node_id < 0)
@@ -145,8 +144,7 @@ int qib_create_ctxts(struct qib_devdata *dd)
 	if (!dd->rcd) {
 		qib_dev_err(dd,
 			"Unable to allocate ctxtdata array, failing\n");
-		ret = -ENOMEM;
-		goto done;
+		return -ENOMEM;
 	}
 
 	/* create (one or more) kctxt */
@@ -163,15 +161,14 @@ int qib_create_ctxts(struct qib_devdata *dd)
 		if (!rcd) {
 			qib_dev_err(dd,
 				"Unable to allocate ctxtdata for Kernel ctxt, failing\n");
-			ret = -ENOMEM;
-			goto done;
+			kfree(dd->rcd);
+			dd->rcd = NULL;
+			return -ENOMEM;
 		}
 		rcd->pkeys[0] = QIB_DEFAULT_P_KEY;
 		rcd->seq_cnt = 1;
 	}
-	ret = 0;
-done:
-	return ret;
+	return 0;
 }
 
 /*
@@ -233,7 +230,7 @@ struct qib_ctxtdata *qib_create_ctxtdata(struct qib_pportdata *ppd, u32 ctxt,
 /*
  * Common code for initializing the physical port structure.
  */
-void qib_init_pportdata(struct qib_pportdata *ppd, struct qib_devdata *dd,
+int qib_init_pportdata(struct qib_pportdata *ppd, struct qib_devdata *dd,
 			u8 hw_pidx, u8 port)
 {
 	int size;
@@ -243,6 +240,7 @@ void qib_init_pportdata(struct qib_pportdata *ppd, struct qib_devdata *dd,
 
 	spin_lock_init(&ppd->sdma_lock);
 	spin_lock_init(&ppd->lflags_lock);
+	spin_lock_init(&ppd->cc_shadow_lock);
 	init_waitqueue_head(&ppd->state_wait);
 
 	init_timer(&ppd->symerr_clear_timer);
@@ -250,8 +248,10 @@ void qib_init_pportdata(struct qib_pportdata *ppd, struct qib_devdata *dd,
 	ppd->symerr_clear_timer.data = (unsigned long)ppd;
 
 	ppd->qib_wq = NULL;
-
-	spin_lock_init(&ppd->cc_shadow_lock);
+	ppd->ibport_data.pmastats =
+		alloc_percpu(struct qib_pma_counters);
+	if (!ppd->ibport_data.pmastats)
+		return -ENOMEM;
 
 	if (qib_cc_table_size < IB_CCT_MIN_ENTRIES)
 		goto bail;
@@ -299,7 +299,7 @@ void qib_init_pportdata(struct qib_pportdata *ppd, struct qib_devdata *dd,
 		goto bail_3;
 	}
 
-	return;
+	return 0;
 
 bail_3:
 	kfree(ppd->ccti_entries_shadow);
@@ -313,7 +313,7 @@ bail_1:
 bail:
 	/* User is intentionally disabling the congestion control agent */
 	if (!qib_cc_table_size)
-		return;
+		return 0;
 
 	if (qib_cc_table_size < IB_CCT_MIN_ENTRIES) {
 		qib_cc_table_size = 0;
@@ -324,7 +324,7 @@ bail:
 
 	qib_dev_err(dd, "Congestion Control Agent disabled for port %d\n",
 		port);
-	return;
+	return 0;
 }
 
 static int init_pioavailregs(struct qib_devdata *dd)
@@ -525,6 +525,7 @@ static void enable_chip(struct qib_devdata *dd)
 static void verify_interrupt(unsigned long opaque)
 {
 	struct qib_devdata *dd = (struct qib_devdata *) opaque;
+	u64 int_counter;
 
 	if (!dd)
 		return; /* being torn down */
@@ -533,7 +534,8 @@ static void verify_interrupt(unsigned long opaque)
 	 * If we don't have a lid or any interrupts, let the user know and
 	 * don't bother checking again.
 	 */
-	if (dd->int_counter == 0) {
+	int_counter = qib_int_counter(dd) - dd->z_int_counter;
+	if (int_counter == 0) {
 		if (!dd->f_intr_fallback(dd))
 			dev_err(&dd->pcidev->dev,
 				"No interrupts detected, not usable.\n");
@@ -633,6 +635,12 @@ wq_error:
 	return -ENOMEM;
 }
 
+static void qib_free_pportdata(struct qib_pportdata *ppd)
+{
+	free_percpu(ppd->ibport_data.pmastats);
+	ppd->ibport_data.pmastats = NULL;
+}
+
 /**
  * qib_init - do the actual initialization sequence on the chip
  * @dd: the qlogic_ib device
@@ -920,6 +928,7 @@ static void qib_shutdown_device(struct qib_devdata *dd)
 			destroy_workqueue(ppd->qib_wq);
 			ppd->qib_wq = NULL;
 		}
+		qib_free_pportdata(ppd);
 	}
 
 	qib_update_eeprom_log(dd);
@@ -1079,9 +1088,34 @@ void qib_free_devdata(struct qib_devdata *dd)
 #ifdef CONFIG_DEBUG_FS
 	qib_dbg_ibdev_exit(&dd->verbs_dev);
 #endif
+	free_percpu(dd->int_counter);
 	ib_dealloc_device(&dd->verbs_dev.ibdev);
 }
 
+u64 qib_int_counter(struct qib_devdata *dd)
+{
+	int cpu;
+	u64 int_counter = 0;
+
+	for_each_possible_cpu(cpu)
+		int_counter += *per_cpu_ptr(dd->int_counter, cpu);
+	return int_counter;
+}
+
+u64 qib_sps_ints(void)
+{
+	unsigned long flags;
+	struct qib_devdata *dd;
+	u64 sps_ints = 0;
+
+	spin_lock_irqsave(&qib_devs_lock, flags);
+	list_for_each_entry(dd, &qib_dev_list, list) {
+		sps_ints += qib_int_counter(dd);
+	}
+	spin_unlock_irqrestore(&qib_devs_lock, flags);
+	return sps_ints;
+}
+
 /*
  * Allocate our primary per-unit data structure.  Must be done via verbs
  * allocator, because the verbs cleanup process both does cleanup and
@@ -1097,14 +1131,10 @@ struct qib_devdata *qib_alloc_devdata(struct pci_dev *pdev, size_t extra)
 	int ret;
 
 	dd = (struct qib_devdata *) ib_alloc_device(sizeof(*dd) + extra);
-	if (!dd) {
-		dd = ERR_PTR(-ENOMEM);
-		goto bail;
-	}
+	if (!dd)
+		return ERR_PTR(-ENOMEM);
 
-#ifdef CONFIG_DEBUG_FS
-	qib_dbg_ibdev_init(&dd->verbs_dev);
-#endif
+	INIT_LIST_HEAD(&dd->list);
 
 	idr_preload(GFP_KERNEL);
 	spin_lock_irqsave(&qib_devs_lock, flags);
@@ -1121,11 +1151,13 @@ struct qib_devdata *qib_alloc_devdata(struct pci_dev *pdev, size_t extra)
 	if (ret < 0) {
 		qib_early_err(&pdev->dev,
 			      "Could not allocate unit ID: error %d\n", -ret);
-#ifdef CONFIG_DEBUG_FS
-		qib_dbg_ibdev_exit(&dd->verbs_dev);
-#endif
-		ib_dealloc_device(&dd->verbs_dev.ibdev);
-		dd = ERR_PTR(ret);
+		goto bail;
+	}
+	dd->int_counter = alloc_percpu(u64);
+	if (!dd->int_counter) {
+		ret = -ENOMEM;
+		qib_early_err(&pdev->dev,
+			      "Could not allocate per-cpu int_counter\n");
 		goto bail;
 	}
 
@@ -1139,9 +1171,15 @@ struct qib_devdata *qib_alloc_devdata(struct pci_dev *pdev, size_t extra)
 			qib_early_err(&pdev->dev,
 				"Could not alloc cpulist info, cpu affinity might be wrong\n");
 	}
-
-bail:
+#ifdef CONFIG_DEBUG_FS
+	qib_dbg_ibdev_init(&dd->verbs_dev);
+#endif
 	return dd;
+bail:
+	if (!list_empty(&dd->list))
+		list_del_init(&dd->list);
+	ib_dealloc_device(&dd->verbs_dev.ibdev);
+	return ERR_PTR(ret);;
 }
 
 /*

drivers/infiniband/hw/qib/qib_mad.c

@@ -1634,6 +1634,23 @@ static int pma_get_portcounters_cong(struct ib_pma_mad *pmp,
 	return reply((struct ib_smp *)pmp);
 }
 
+static void qib_snapshot_pmacounters(
+	struct qib_ibport *ibp,
+	struct qib_pma_counters *pmacounters)
+{
+	struct qib_pma_counters *p;
+	int cpu;
+
+	memset(pmacounters, 0, sizeof(*pmacounters));
+	for_each_possible_cpu(cpu) {
+		p = per_cpu_ptr(ibp->pmastats, cpu);
+		pmacounters->n_unicast_xmit += p->n_unicast_xmit;
+		pmacounters->n_unicast_rcv += p->n_unicast_rcv;
+		pmacounters->n_multicast_xmit += p->n_multicast_xmit;
+		pmacounters->n_multicast_rcv += p->n_multicast_rcv;
+	}
+}
+
 static int pma_get_portcounters_ext(struct ib_pma_mad *pmp,
 				    struct ib_device *ibdev, u8 port)
 {
@@ -1642,6 +1659,7 @@ static int pma_get_portcounters_ext(struct ib_pma_mad *pmp,
 	struct qib_ibport *ibp = to_iport(ibdev, port);
 	struct qib_pportdata *ppd = ppd_from_ibp(ibp);
 	u64 swords, rwords, spkts, rpkts, xwait;
+	struct qib_pma_counters pma;
 	u8 port_select = p->port_select;
 
 	memset(pmp->data, 0, sizeof(pmp->data));
@@ -1664,10 +1682,17 @@ static int pma_get_portcounters_ext(struct ib_pma_mad *pmp,
 	p->port_rcv_data = cpu_to_be64(rwords);
 	p->port_xmit_packets = cpu_to_be64(spkts);
 	p->port_rcv_packets = cpu_to_be64(rpkts);
-	p->port_unicast_xmit_packets = cpu_to_be64(ibp->n_unicast_xmit);
-	p->port_unicast_rcv_packets = cpu_to_be64(ibp->n_unicast_rcv);
-	p->port_multicast_xmit_packets = cpu_to_be64(ibp->n_multicast_xmit);
-	p->port_multicast_rcv_packets = cpu_to_be64(ibp->n_multicast_rcv);
+
+	qib_snapshot_pmacounters(ibp, &pma);
+
+	p->port_unicast_xmit_packets = cpu_to_be64(pma.n_unicast_xmit
+		- ibp->z_unicast_xmit);
+	p->port_unicast_rcv_packets = cpu_to_be64(pma.n_unicast_rcv
+		- ibp->z_unicast_rcv);
+	p->port_multicast_xmit_packets = cpu_to_be64(pma.n_multicast_xmit
+		- ibp->z_multicast_xmit);
+	p->port_multicast_rcv_packets = cpu_to_be64(pma.n_multicast_rcv
+		- ibp->z_multicast_rcv);
 
 bail:
 	return reply((struct ib_smp *) pmp);
@@ -1795,6 +1820,7 @@ static int pma_set_portcounters_ext(struct ib_pma_mad *pmp,
 	struct qib_ibport *ibp = to_iport(ibdev, port);
 	struct qib_pportdata *ppd = ppd_from_ibp(ibp);
 	u64 swords, rwords, spkts, rpkts, xwait;
+	struct qib_pma_counters pma;
 
 	qib_snapshot_counters(ppd, &swords, &rwords, &spkts, &rpkts, &xwait);
 
@@ -1810,17 +1836,19 @@ static int pma_set_portcounters_ext(struct ib_pma_mad *pmp,
 	if (p->counter_select & IB_PMA_SELX_PORT_RCV_PACKETS)
 		ibp->z_port_rcv_packets = rpkts;
 
+	qib_snapshot_pmacounters(ibp, &pma);
+
 	if (p->counter_select & IB_PMA_SELX_PORT_UNI_XMIT_PACKETS)
-		ibp->n_unicast_xmit = 0;
+		ibp->z_unicast_xmit = pma.n_unicast_xmit;
 
 	if (p->counter_select & IB_PMA_SELX_PORT_UNI_RCV_PACKETS)
-		ibp->n_unicast_rcv = 0;
+		ibp->z_unicast_rcv = pma.n_unicast_rcv;
 
 	if (p->counter_select & IB_PMA_SELX_PORT_MULTI_XMIT_PACKETS)
-		ibp->n_multicast_xmit = 0;
+		ibp->z_multicast_xmit = pma.n_multicast_xmit;
 
 	if (p->counter_select & IB_PMA_SELX_PORT_MULTI_RCV_PACKETS)
-		ibp->n_multicast_rcv = 0;
+		ibp->z_multicast_rcv = pma.n_multicast_rcv;
 
 	return pma_get_portcounters_ext(pmp, ibdev, port);
 }

drivers/infiniband/hw/qib/qib_mr.c

@@ -232,8 +232,8 @@ struct ib_mr *qib_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 {
 	struct qib_mr *mr;
 	struct ib_umem *umem;
-	struct ib_umem_chunk *chunk;
-	int n, m, i;
+	struct scatterlist *sg;
+	int n, m, entry;
 	struct ib_mr *ret;
 
 	if (length == 0) {
@@ -246,9 +246,7 @@ struct ib_mr *qib_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 	if (IS_ERR(umem))
 		return (void *) umem;
 
-	n = 0;
-	list_for_each_entry(chunk, &umem->chunk_list, list)
-		n += chunk->nents;
+	n = umem->nmap;
 
 	mr = alloc_mr(n, pd);
 	if (IS_ERR(mr)) {
@@ -268,11 +266,10 @@ struct ib_mr *qib_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 		mr->mr.page_shift = ilog2(umem->page_size);
 	m = 0;
 	n = 0;
-	list_for_each_entry(chunk, &umem->chunk_list, list) {
-		for (i = 0; i < chunk->nents; i++) {
+	for_each_sg(umem->sg_head.sgl, sg, umem->nmap, entry) {
 			void *vaddr;
 
-			vaddr = page_address(sg_page(&chunk->page_list[i]));
+			vaddr = page_address(sg_page(sg));
 			if (!vaddr) {
 				ret = ERR_PTR(-EINVAL);
 				goto bail;
@@ -284,7 +281,6 @@ struct ib_mr *qib_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
 				m++;
 				n = 0;
 			}
-		}
 	}
 	ret = &mr->ibmr;
 

drivers/infiniband/hw/qib/qib_rc.c

@@ -752,7 +752,7 @@ void qib_send_rc_ack(struct qib_qp *qp)
 	qib_flush_wc();
 	qib_sendbuf_done(dd, pbufn);
 
-	ibp->n_unicast_xmit++;
+	this_cpu_inc(ibp->pmastats->n_unicast_xmit);
 	goto done;
 
 queue_ack:

drivers/infiniband/hw/qib/qib_ruc.c

@@ -703,6 +703,7 @@ void qib_make_ruc_header(struct qib_qp *qp, struct qib_other_headers *ohdr,
 	ohdr->bth[0] = cpu_to_be32(bth0);
 	ohdr->bth[1] = cpu_to_be32(qp->remote_qpn);
 	ohdr->bth[2] = cpu_to_be32(bth2);
+	this_cpu_inc(ibp->pmastats->n_unicast_xmit);
 }
 
 /**

drivers/infiniband/hw/qib/qib_ud.c

@@ -280,11 +280,11 @@ int qib_make_ud_req(struct qib_qp *qp)
 	ah_attr = &to_iah(wqe->wr.wr.ud.ah)->attr;
 	if (ah_attr->dlid >= QIB_MULTICAST_LID_BASE) {
 		if (ah_attr->dlid != QIB_PERMISSIVE_LID)
-			ibp->n_multicast_xmit++;
+			this_cpu_inc(ibp->pmastats->n_multicast_xmit);
 		else
-			ibp->n_unicast_xmit++;
+			this_cpu_inc(ibp->pmastats->n_unicast_xmit);
 	} else {
-		ibp->n_unicast_xmit++;
+		this_cpu_inc(ibp->pmastats->n_unicast_xmit);
 		lid = ah_attr->dlid & ~((1 << ppd->lmc) - 1);
 		if (unlikely(lid == ppd->lid)) {
 			/*

drivers/infiniband/hw/qib/qib_user_sdma.c

@@ -52,6 +52,17 @@
 /* attempt to drain the queue for 5secs */
 #define QIB_USER_SDMA_DRAIN_TIMEOUT 500
 
+/*
+ * track how many times a process open this driver.
+ */
+static struct rb_root qib_user_sdma_rb_root = RB_ROOT;
+
+struct qib_user_sdma_rb_node {
+	struct rb_node node;
+	int refcount;
+	pid_t pid;
+};
+
 struct qib_user_sdma_pkt {
 	struct list_head list;  /* list element */
 
@@ -120,15 +131,60 @@ struct qib_user_sdma_queue {
 	/* dma page table */
 	struct rb_root dma_pages_root;
 
+	struct qib_user_sdma_rb_node *sdma_rb_node;
+
 	/* protect everything above... */
 	struct mutex lock;
 };
 
+static struct qib_user_sdma_rb_node *
+qib_user_sdma_rb_search(struct rb_root *root, pid_t pid)
+{
+	struct qib_user_sdma_rb_node *sdma_rb_node;
+	struct rb_node *node = root->rb_node;
+
+	while (node) {
+		sdma_rb_node = container_of(node,
+			struct qib_user_sdma_rb_node, node);
+		if (pid < sdma_rb_node->pid)
+			node = node->rb_left;
+		else if (pid > sdma_rb_node->pid)
+			node = node->rb_right;
+		else
+			return sdma_rb_node;
+	}
+	return NULL;
+}
+
+static int
+qib_user_sdma_rb_insert(struct rb_root *root, struct qib_user_sdma_rb_node *new)
+{
+	struct rb_node **node = &(root->rb_node);
+	struct rb_node *parent = NULL;
+	struct qib_user_sdma_rb_node *got;
+
+	while (*node) {
+		got = container_of(*node, struct qib_user_sdma_rb_node, node);
+		parent = *node;
+		if (new->pid < got->pid)
+			node = &((*node)->rb_left);
+		else if (new->pid > got->pid)
+			node = &((*node)->rb_right);
+		else
+			return 0;
+	}
+
+	rb_link_node(&new->node, parent, node);
+	rb_insert_color(&new->node, root);
+	return 1;
+}
+
 struct qib_user_sdma_queue *
 qib_user_sdma_queue_create(struct device *dev, int unit, int ctxt, int sctxt)
 {
 	struct qib_user_sdma_queue *pq =
 		kmalloc(sizeof(struct qib_user_sdma_queue), GFP_KERNEL);
+	struct qib_user_sdma_rb_node *sdma_rb_node;
 
 	if (!pq)
 		goto done;
@@ -138,6 +194,7 @@ qib_user_sdma_queue_create(struct device *dev, int unit, int ctxt, int sctxt)
 	pq->num_pending = 0;
 	pq->num_sending = 0;
 	pq->added = 0;
+	pq->sdma_rb_node = NULL;
 
 	INIT_LIST_HEAD(&pq->sent);
 	spin_lock_init(&pq->sent_lock);
@@ -163,8 +220,30 @@ qib_user_sdma_queue_create(struct device *dev, int unit, int ctxt, int sctxt)
 
 	pq->dma_pages_root = RB_ROOT;
 
+	sdma_rb_node = qib_user_sdma_rb_search(&qib_user_sdma_rb_root,
+					current->pid);
+	if (sdma_rb_node) {
+		sdma_rb_node->refcount++;
+	} else {
+		int ret;
+		sdma_rb_node = kmalloc(sizeof(
+			struct qib_user_sdma_rb_node), GFP_KERNEL);
+		if (!sdma_rb_node)
+			goto err_rb;
+
+		sdma_rb_node->refcount = 1;
+		sdma_rb_node->pid = current->pid;
+
+		ret = qib_user_sdma_rb_insert(&qib_user_sdma_rb_root,
+					sdma_rb_node);
+		BUG_ON(ret == 0);
+	}
+	pq->sdma_rb_node = sdma_rb_node;
+
 	goto done;
 
+err_rb:
+	dma_pool_destroy(pq->header_cache);
 err_slab:
 	kmem_cache_destroy(pq->pkt_slab);
 err_kfree:
@@ -1020,8 +1099,13 @@ void qib_user_sdma_queue_destroy(struct qib_user_sdma_queue *pq)
 	if (!pq)
 		return;
 
-	kmem_cache_destroy(pq->pkt_slab);
+	pq->sdma_rb_node->refcount--;
+	if (pq->sdma_rb_node->refcount == 0) {
+		rb_erase(&pq->sdma_rb_node->node, &qib_user_sdma_rb_root);
+		kfree(pq->sdma_rb_node);
+	}
 	dma_pool_destroy(pq->header_cache);
+	kmem_cache_destroy(pq->pkt_slab);
 	kfree(pq);
 }
 
@@ -1241,26 +1325,52 @@ static int qib_user_sdma_push_pkts(struct qib_pportdata *ppd,
 				 struct qib_user_sdma_queue *pq,
 				 struct list_head *pktlist, int count)
 {
-	int ret = 0;
 	unsigned long flags;
 
 	if (unlikely(!(ppd->lflags & QIBL_LINKACTIVE)))
 		return -ECOMM;
 
-	spin_lock_irqsave(&ppd->sdma_lock, flags);
-
-	if (unlikely(!__qib_sdma_running(ppd))) {
-		ret = -ECOMM;
-		goto unlock;
+	/* non-blocking mode */
+	if (pq->sdma_rb_node->refcount > 1) {
+		spin_lock_irqsave(&ppd->sdma_lock, flags);
+		if (unlikely(!__qib_sdma_running(ppd))) {
+			spin_unlock_irqrestore(&ppd->sdma_lock, flags);
+			return -ECOMM;
+		}
+		pq->num_pending += count;
+		list_splice_tail_init(pktlist, &ppd->sdma_userpending);
+		qib_user_sdma_send_desc(ppd, &ppd->sdma_userpending);
+		spin_unlock_irqrestore(&ppd->sdma_lock, flags);
+		return 0;
 	}
 
+	/* In this case, descriptors from this process are not
+	 * linked to ppd pending queue, interrupt handler
+	 * won't update this process, it is OK to directly
+	 * modify without sdma lock.
+	 */
+
+
 	pq->num_pending += count;
-	list_splice_tail_init(pktlist, &ppd->sdma_userpending);
-	qib_user_sdma_send_desc(ppd, &ppd->sdma_userpending);
+	/*
+	 * Blocking mode for single rail process, we must
+	 * release/regain sdma_lock to give other process
+	 * chance to make progress. This is important for
+	 * performance.
+	 */
+	do {
+		spin_lock_irqsave(&ppd->sdma_lock, flags);
+		if (unlikely(!__qib_sdma_running(ppd))) {
+			spin_unlock_irqrestore(&ppd->sdma_lock, flags);
+			return -ECOMM;
+		}
+		qib_user_sdma_send_desc(ppd, pktlist);
+		if (!list_empty(pktlist))
+			qib_sdma_make_progress(ppd);
+		spin_unlock_irqrestore(&ppd->sdma_lock, flags);
+	} while (!list_empty(pktlist));
 
-unlock:
-	spin_unlock_irqrestore(&ppd->sdma_lock, flags);
-	return ret;
+	return 0;
 }
 
 int qib_user_sdma_writev(struct qib_ctxtdata *rcd,
@@ -1290,7 +1400,7 @@ int qib_user_sdma_writev(struct qib_ctxtdata *rcd,
 		qib_user_sdma_queue_clean(ppd, pq);
 
 	while (dim) {
-		int mxp = 8;
+		int mxp = 1;
 		int ndesc = 0;
 
 		ret = qib_user_sdma_queue_pkts(dd, ppd, pq,

drivers/infiniband/hw/qib/qib_verbs.c

@@ -662,7 +662,7 @@ void qib_ib_rcv(struct qib_ctxtdata *rcd, void *rhdr, void *data, u32 tlen)
 		mcast = qib_mcast_find(ibp, &hdr->u.l.grh.dgid);
 		if (mcast == NULL)
 			goto drop;
-		ibp->n_multicast_rcv++;
+		this_cpu_inc(ibp->pmastats->n_multicast_rcv);
 		list_for_each_entry_rcu(p, &mcast->qp_list, list)
 			qib_qp_rcv(rcd, hdr, 1, data, tlen, p->qp);
 		/*
@@ -678,8 +678,8 @@ void qib_ib_rcv(struct qib_ctxtdata *rcd, void *rhdr, void *data, u32 tlen)
 					&rcd->lookaside_qp->refcount))
 					wake_up(
 					 &rcd->lookaside_qp->wait);
-					rcd->lookaside_qp = NULL;
-				}
+				rcd->lookaside_qp = NULL;
+			}
 		}
 		if (!rcd->lookaside_qp) {
 			qp = qib_lookup_qpn(ibp, qp_num);
@@ -689,7 +689,7 @@ void qib_ib_rcv(struct qib_ctxtdata *rcd, void *rhdr, void *data, u32 tlen)
 			rcd->lookaside_qpn = qp_num;
 		} else
 			qp = rcd->lookaside_qp;
-		ibp->n_unicast_rcv++;
+		this_cpu_inc(ibp->pmastats->n_unicast_rcv);
 		qib_qp_rcv(rcd, hdr, lnh == QIB_LRH_GRH, data, tlen, qp);
 	}
 	return;

drivers/infiniband/hw/qib/qib_verbs.h

@@ -664,6 +664,13 @@ struct qib_opcode_stats_perctx {
 	struct qib_opcode_stats stats[128];
 };
 
+struct qib_pma_counters {
+	u64 n_unicast_xmit;     /* total unicast packets sent */
+	u64 n_unicast_rcv;      /* total unicast packets received */
+	u64 n_multicast_xmit;   /* total multicast packets sent */
+	u64 n_multicast_rcv;    /* total multicast packets received */
+};
+
 struct qib_ibport {
 	struct qib_qp __rcu *qp0;
 	struct qib_qp __rcu *qp1;
@@ -680,10 +687,11 @@ struct qib_ibport {
 	__be64 mkey;
 	__be64 guids[QIB_GUIDS_PER_PORT	- 1];	/* writable GUIDs */
 	u64 tid;		/* TID for traps */
-	u64 n_unicast_xmit;     /* total unicast packets sent */
-	u64 n_unicast_rcv;      /* total unicast packets received */
-	u64 n_multicast_xmit;   /* total multicast packets sent */
-	u64 n_multicast_rcv;    /* total multicast packets received */
+	struct qib_pma_counters __percpu *pmastats;
+	u64 z_unicast_xmit;     /* starting count for PMA */
+	u64 z_unicast_rcv;      /* starting count for PMA */
+	u64 z_multicast_xmit;   /* starting count for PMA */
+	u64 z_multicast_rcv;    /* starting count for PMA */
 	u64 z_symbol_error_counter;             /* starting count for PMA */
 	u64 z_link_error_recovery_counter;      /* starting count for PMA */
 	u64 z_link_downed_counter;              /* starting count for PMA */

drivers/infiniband/hw/usnic/usnic_uiom.c

@@ -286,7 +286,7 @@ iter_chunk:
 				err = iommu_map(pd->domain, va_start, pa_start,
 							size, flags);
 				if (err) {
-					usnic_err("Failed to map va 0x%lx pa 0x%pa size 0x%zx with err %d\n",
+					usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n",
 						va_start, &pa_start, size, err);
 					goto err_out;
 				}

drivers/infiniband/ulp/iser/iscsi_iser.c

@@ -5,7 +5,7 @@
  * Copyright (C) 2004 Alex Aizman
  * Copyright (C) 2005 Mike Christie
  * Copyright (c) 2005, 2006 Voltaire, Inc. All rights reserved.
- * Copyright (c) 2013 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2013-2014 Mellanox Technologies. All rights reserved.
  * maintained by openib-general@openib.org
  *
  * This software is available to you under a choice of one of two
@@ -82,6 +82,8 @@ static unsigned int iscsi_max_lun = 512;
 module_param_named(max_lun, iscsi_max_lun, uint, S_IRUGO);
 
 int iser_debug_level = 0;
+bool iser_pi_enable = false;
+int iser_pi_guard = 0;
 
 MODULE_DESCRIPTION("iSER (iSCSI Extensions for RDMA) Datamover");
 MODULE_LICENSE("Dual BSD/GPL");
@@ -91,6 +93,12 @@ MODULE_VERSION(DRV_VER);
 module_param_named(debug_level, iser_debug_level, int, 0644);
 MODULE_PARM_DESC(debug_level, "Enable debug tracing if > 0 (default:disabled)");
 
+module_param_named(pi_enable, iser_pi_enable, bool, 0644);
+MODULE_PARM_DESC(pi_enable, "Enable T10-PI offload support (default:disabled)");
+
+module_param_named(pi_guard, iser_pi_guard, int, 0644);
+MODULE_PARM_DESC(pi_guard, "T10-PI guard_type, 0:CRC|1:IP_CSUM (default:CRC)");
+
 struct iser_global ig;
 
 void
@@ -138,8 +146,8 @@ static int iscsi_iser_pdu_alloc(struct iscsi_task *task, uint8_t opcode)
 int iser_initialize_task_headers(struct iscsi_task *task,
 						struct iser_tx_desc *tx_desc)
 {
-	struct iscsi_iser_conn *iser_conn = task->conn->dd_data;
-	struct iser_device     *device    = iser_conn->ib_conn->device;
+	struct iser_conn       *ib_conn   = task->conn->dd_data;
+	struct iser_device     *device    = ib_conn->device;
 	struct iscsi_iser_task *iser_task = task->dd_data;
 	u64 dma_addr;
 
@@ -153,7 +161,7 @@ int iser_initialize_task_headers(struct iscsi_task *task,
 	tx_desc->tx_sg[0].length = ISER_HEADERS_LEN;
 	tx_desc->tx_sg[0].lkey   = device->mr->lkey;
 
-	iser_task->iser_conn		= iser_conn;
+	iser_task->ib_conn = ib_conn;
 	return 0;
 }
 /**
@@ -176,6 +184,8 @@ iscsi_iser_task_init(struct iscsi_task *task)
 
 	iser_task->command_sent = 0;
 	iser_task_rdma_init(iser_task);
+	iser_task->sc = task->sc;
+
 	return 0;
 }
 
@@ -278,10 +288,9 @@ iscsi_iser_task_xmit(struct iscsi_task *task)
 static void iscsi_iser_cleanup_task(struct iscsi_task *task)
 {
 	struct iscsi_iser_task *iser_task = task->dd_data;
-	struct iser_tx_desc	*tx_desc = &iser_task->desc;
-
-	struct iscsi_iser_conn *iser_conn = task->conn->dd_data;
-	struct iser_device     *device    = iser_conn->ib_conn->device;
+	struct iser_tx_desc    *tx_desc   = &iser_task->desc;
+	struct iser_conn       *ib_conn	  = task->conn->dd_data;
+	struct iser_device     *device	  = ib_conn->device;
 
 	ib_dma_unmap_single(device->ib_device,
 		tx_desc->dma_addr, ISER_HEADERS_LEN, DMA_TO_DEVICE);
@@ -296,14 +305,25 @@ static void iscsi_iser_cleanup_task(struct iscsi_task *task)
 	}
 }
 
+static u8 iscsi_iser_check_protection(struct iscsi_task *task, sector_t *sector)
+{
+	struct iscsi_iser_task *iser_task = task->dd_data;
+
+	if (iser_task->dir[ISER_DIR_IN])
+		return iser_check_task_pi_status(iser_task, ISER_DIR_IN,
+						 sector);
+	else
+		return iser_check_task_pi_status(iser_task, ISER_DIR_OUT,
+						 sector);
+}
+
 static struct iscsi_cls_conn *
 iscsi_iser_conn_create(struct iscsi_cls_session *cls_session, uint32_t conn_idx)
 {
 	struct iscsi_conn *conn;
 	struct iscsi_cls_conn *cls_conn;
-	struct iscsi_iser_conn *iser_conn;
 
-	cls_conn = iscsi_conn_setup(cls_session, sizeof(*iser_conn), conn_idx);
+	cls_conn = iscsi_conn_setup(cls_session, 0, conn_idx);
 	if (!cls_conn)
 		return NULL;
 	conn = cls_conn->dd_data;
@@ -314,10 +334,6 @@ iscsi_iser_conn_create(struct iscsi_cls_session *cls_session, uint32_t conn_idx)
 	 */
 	conn->max_recv_dlength = ISER_RECV_DATA_SEG_LEN;
 
-	iser_conn = conn->dd_data;
-	conn->dd_data = iser_conn;
-	iser_conn->iscsi_conn = conn;
-
 	return cls_conn;
 }
 
@@ -325,8 +341,7 @@ static void
 iscsi_iser_conn_destroy(struct iscsi_cls_conn *cls_conn)
 {
 	struct iscsi_conn *conn = cls_conn->dd_data;
-	struct iscsi_iser_conn *iser_conn = conn->dd_data;
-	struct iser_conn *ib_conn = iser_conn->ib_conn;
+	struct iser_conn *ib_conn = conn->dd_data;
 
 	iscsi_conn_teardown(cls_conn);
 	/*
@@ -335,7 +350,7 @@ iscsi_iser_conn_destroy(struct iscsi_cls_conn *cls_conn)
 	 * we free it here.
 	 */
 	if (ib_conn) {
-		ib_conn->iser_conn = NULL;
+		ib_conn->iscsi_conn = NULL;
 		iser_conn_put(ib_conn, 1); /* deref iscsi/ib conn unbinding */
 	}
 }
@@ -346,7 +361,6 @@ iscsi_iser_conn_bind(struct iscsi_cls_session *cls_session,
 		     int is_leading)
 {
 	struct iscsi_conn *conn = cls_conn->dd_data;
-	struct iscsi_iser_conn *iser_conn;
 	struct iscsi_session *session;
 	struct iser_conn *ib_conn;
 	struct iscsi_endpoint *ep;
@@ -373,11 +387,11 @@ iscsi_iser_conn_bind(struct iscsi_cls_session *cls_session,
 	/* binds the iSER connection retrieved from the previously
 	 * connected ep_handle to the iSCSI layer connection. exchanges
 	 * connection pointers */
-	iser_info("binding iscsi/iser conn %p %p to ib_conn %p\n",
-		  conn, conn->dd_data, ib_conn);
-	iser_conn = conn->dd_data;
-	ib_conn->iser_conn = iser_conn;
-	iser_conn->ib_conn  = ib_conn;
+	iser_info("binding iscsi conn %p to ib_conn %p\n", conn, ib_conn);
+
+	conn->dd_data = ib_conn;
+	ib_conn->iscsi_conn = conn;
+
 	iser_conn_get(ib_conn); /* ref iscsi/ib conn binding */
 	return 0;
 }
@@ -386,8 +400,7 @@ static void
 iscsi_iser_conn_stop(struct iscsi_cls_conn *cls_conn, int flag)
 {
 	struct iscsi_conn *conn = cls_conn->dd_data;
-	struct iscsi_iser_conn *iser_conn = conn->dd_data;
-	struct iser_conn *ib_conn = iser_conn->ib_conn;
+	struct iser_conn *ib_conn = conn->dd_data;
 
 	/*
 	 * Userspace may have goofed up and not bound the connection or
@@ -401,7 +414,7 @@ iscsi_iser_conn_stop(struct iscsi_cls_conn *cls_conn, int flag)
 		 */
 		iser_conn_put(ib_conn, 1); /* deref iscsi/ib conn unbinding */
 	}
-	iser_conn->ib_conn = NULL;
+	conn->dd_data = NULL;
 }
 
 static void iscsi_iser_session_destroy(struct iscsi_cls_session *cls_session)
@@ -413,6 +426,17 @@ static void iscsi_iser_session_destroy(struct iscsi_cls_session *cls_session)
 	iscsi_host_free(shost);
 }
 
+static inline unsigned int
+iser_dif_prot_caps(int prot_caps)
+{
+	return ((prot_caps & IB_PROT_T10DIF_TYPE_1) ? SHOST_DIF_TYPE1_PROTECTION |
+						      SHOST_DIX_TYPE1_PROTECTION : 0) |
+	       ((prot_caps & IB_PROT_T10DIF_TYPE_2) ? SHOST_DIF_TYPE2_PROTECTION |
+						      SHOST_DIX_TYPE2_PROTECTION : 0) |
+	       ((prot_caps & IB_PROT_T10DIF_TYPE_3) ? SHOST_DIF_TYPE3_PROTECTION |
+						      SHOST_DIX_TYPE3_PROTECTION : 0);
+}
+
 static struct iscsi_cls_session *
 iscsi_iser_session_create(struct iscsi_endpoint *ep,
 			  uint16_t cmds_max, uint16_t qdepth,
@@ -437,8 +461,18 @@ iscsi_iser_session_create(struct iscsi_endpoint *ep,
 	 * older userspace tools (before 2.0-870) did not pass us
 	 * the leading conn's ep so this will be NULL;
 	 */
-	if (ep)
+	if (ep) {
 		ib_conn = ep->dd_data;
+		if (ib_conn->pi_support) {
+			u32 sig_caps = ib_conn->device->dev_attr.sig_prot_cap;
+
+			scsi_host_set_prot(shost, iser_dif_prot_caps(sig_caps));
+			if (iser_pi_guard)
+				scsi_host_set_guard(shost, SHOST_DIX_GUARD_IP);
+			else
+				scsi_host_set_guard(shost, SHOST_DIX_GUARD_CRC);
+		}
+	}
 
 	if (iscsi_host_add(shost,
 			   ep ? ib_conn->device->ib_device->dma_device : NULL))
@@ -618,7 +652,7 @@ iscsi_iser_ep_disconnect(struct iscsi_endpoint *ep)
 	struct iser_conn *ib_conn;
 
 	ib_conn = ep->dd_data;
-	if (ib_conn->iser_conn)
+	if (ib_conn->iscsi_conn)
 		/*
 		 * Must suspend xmit path if the ep is bound to the
 		 * iscsi_conn, so we know we are not accessing the ib_conn
@@ -626,7 +660,7 @@ iscsi_iser_ep_disconnect(struct iscsi_endpoint *ep)
 		 *
 		 * This may not be bound if the ep poll failed.
 		 */
-		iscsi_suspend_tx(ib_conn->iser_conn->iscsi_conn);
+		iscsi_suspend_tx(ib_conn->iscsi_conn);
 
 
 	iser_info("ib conn %p state %d\n", ib_conn, ib_conn->state);
@@ -732,6 +766,7 @@ static struct iscsi_transport iscsi_iser_transport = {
 	.xmit_task		= iscsi_iser_task_xmit,
 	.cleanup_task		= iscsi_iser_cleanup_task,
 	.alloc_pdu		= iscsi_iser_pdu_alloc,
+	.check_protection	= iscsi_iser_check_protection,
 	/* recovery */
 	.session_recovery_timedout = iscsi_session_recovery_timedout,
 

drivers/infiniband/ulp/iser/iscsi_iser.h

@@ -8,7 +8,7 @@
  *
  * Copyright (c) 2004, 2005, 2006 Voltaire, Inc. All rights reserved.
  * Copyright (c) 2005, 2006 Cisco Systems.  All rights reserved.
- * Copyright (c) 2013 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2013-2014 Mellanox Technologies. All rights reserved.
  *
  * This software is available to you under a choice of one of two
  * licenses.  You may choose to be licensed under the terms of the GNU
@@ -46,6 +46,8 @@
 #include <linux/printk.h>
 #include <scsi/libiscsi.h>
 #include <scsi/scsi_transport_iscsi.h>
+#include <scsi/scsi_cmnd.h>
+#include <scsi/scsi_device.h>
 
 #include <linux/interrupt.h>
 #include <linux/wait.h>
@@ -67,7 +69,7 @@
 
 #define DRV_NAME	"iser"
 #define PFX		DRV_NAME ": "
-#define DRV_VER		"1.1"
+#define DRV_VER		"1.3"
 
 #define iser_dbg(fmt, arg...)				\
 	do {						\
@@ -134,10 +136,21 @@
 					ISER_MAX_TX_MISC_PDUS        + \
 					ISER_MAX_RX_MISC_PDUS)
 
+/* Max registration work requests per command */
+#define ISER_MAX_REG_WR_PER_CMD		5
+
+/* For Signature we don't support DATAOUTs so no need to make room for them */
+#define ISER_QP_SIG_MAX_REQ_DTOS	(ISER_DEF_XMIT_CMDS_MAX	*       \
+					(1 + ISER_MAX_REG_WR_PER_CMD) + \
+					ISER_MAX_TX_MISC_PDUS         + \
+					ISER_MAX_RX_MISC_PDUS)
+
 #define ISER_VER			0x10
 #define ISER_WSV			0x08
 #define ISER_RSV			0x04
 
+#define ISER_FASTREG_LI_WRID		0xffffffffffffffffULL
+
 struct iser_hdr {
 	u8      flags;
 	u8      rsvd[3];
@@ -201,7 +214,6 @@ struct iser_data_buf {
 /* fwd declarations */
 struct iser_device;
 struct iser_cq_desc;
-struct iscsi_iser_conn;
 struct iscsi_iser_task;
 struct iscsi_endpoint;
 
@@ -258,6 +270,7 @@ struct iscsi_iser_task;
 struct iser_device {
 	struct ib_device             *ib_device;
 	struct ib_pd	             *pd;
+	struct ib_device_attr	     dev_attr;
 	struct ib_cq	             *rx_cq[ISER_MAX_CQ];
 	struct ib_cq	             *tx_cq[ISER_MAX_CQ];
 	struct ib_mr	             *mr;
@@ -277,17 +290,35 @@ struct iser_device {
 							    enum iser_data_dir cmd_dir);
 };
 
+#define ISER_CHECK_GUARD	0xc0
+#define ISER_CHECK_REFTAG	0x0f
+#define ISER_CHECK_APPTAG	0x30
+
+enum iser_reg_indicator {
+	ISER_DATA_KEY_VALID	= 1 << 0,
+	ISER_PROT_KEY_VALID	= 1 << 1,
+	ISER_SIG_KEY_VALID	= 1 << 2,
+	ISER_FASTREG_PROTECTED	= 1 << 3,
+};
+
+struct iser_pi_context {
+	struct ib_mr                   *prot_mr;
+	struct ib_fast_reg_page_list   *prot_frpl;
+	struct ib_mr                   *sig_mr;
+};
+
 struct fast_reg_descriptor {
 	struct list_head		  list;
 	/* For fast registration - FRWR */
 	struct ib_mr			 *data_mr;
 	struct ib_fast_reg_page_list     *data_frpl;
-	/* Valid for fast registration flag */
-	bool				  valid;
+	struct iser_pi_context		 *pi_ctx;
+	/* registration indicators container */
+	u8				  reg_indicators;
 };
 
 struct iser_conn {
-	struct iscsi_iser_conn       *iser_conn; /* iser conn for upcalls  */
+	struct iscsi_conn	     *iscsi_conn;
 	struct iscsi_endpoint	     *ep;
 	enum iser_ib_conn_state	     state;	    /* rdma connection state   */
 	atomic_t		     refcount;
@@ -310,6 +341,9 @@ struct iser_conn {
 	unsigned int 		     rx_desc_head;
 	struct iser_rx_desc	     *rx_descs;
 	struct ib_recv_wr	     rx_wr[ISER_MIN_POSTED_RX];
+	bool			     pi_support;
+
+	/* Connection memory registration pool */
 	union {
 		struct {
 			struct ib_fmr_pool      *pool;	   /* pool of IB FMRs         */
@@ -319,24 +353,22 @@ struct iser_conn {
 		struct {
 			struct list_head	pool;
 			int			pool_size;
-		} frwr;
-	} fastreg;
-};
-
-struct iscsi_iser_conn {
-	struct iscsi_conn            *iscsi_conn;/* ptr to iscsi conn */
-	struct iser_conn             *ib_conn;   /* iSER IB conn      */
+		} fastreg;
+	};
 };
 
 struct iscsi_iser_task {
 	struct iser_tx_desc          desc;
-	struct iscsi_iser_conn	     *iser_conn;
+	struct iser_conn	     *ib_conn;
 	enum iser_task_status 	     status;
+	struct scsi_cmnd	     *sc;
 	int                          command_sent;  /* set if command  sent  */
 	int                          dir[ISER_DIRS_NUM];      /* set if dir use*/
 	struct iser_regd_buf         rdma_regd[ISER_DIRS_NUM];/* regd rdma buf */
 	struct iser_data_buf         data[ISER_DIRS_NUM];     /* orig. data des*/
 	struct iser_data_buf         data_copy[ISER_DIRS_NUM];/* contig. copy  */
+	struct iser_data_buf         prot[ISER_DIRS_NUM];     /* prot desc     */
+	struct iser_data_buf         prot_copy[ISER_DIRS_NUM];/* prot copy     */
 };
 
 struct iser_page_vec {
@@ -362,6 +394,8 @@ struct iser_global {
 
 extern struct iser_global ig;
 extern int iser_debug_level;
+extern bool iser_pi_enable;
+extern int iser_pi_guard;
 
 /* allocate connection resources needed for rdma functionality */
 int iser_conn_set_full_featured_mode(struct iscsi_conn *conn);
@@ -401,13 +435,15 @@ void iser_task_rdma_finalize(struct iscsi_iser_task *task);
 
 void iser_free_rx_descriptors(struct iser_conn *ib_conn);
 
-void iser_finalize_rdma_unaligned_sg(struct iscsi_iser_task *task,
-				     enum iser_data_dir         cmd_dir);
+void iser_finalize_rdma_unaligned_sg(struct iscsi_iser_task *iser_task,
+				     struct iser_data_buf *mem,
+				     struct iser_data_buf *mem_copy,
+				     enum iser_data_dir cmd_dir);
 
 int  iser_reg_rdma_mem_fmr(struct iscsi_iser_task *task,
 			   enum iser_data_dir cmd_dir);
-int  iser_reg_rdma_mem_frwr(struct iscsi_iser_task *task,
-			    enum iser_data_dir cmd_dir);
+int  iser_reg_rdma_mem_fastreg(struct iscsi_iser_task *task,
+			       enum iser_data_dir cmd_dir);
 
 int  iser_connect(struct iser_conn   *ib_conn,
 		  struct sockaddr_in *src_addr,
@@ -420,8 +456,8 @@ int  iser_reg_page_vec(struct iser_conn     *ib_conn,
 
 void iser_unreg_mem_fmr(struct iscsi_iser_task *iser_task,
 			enum iser_data_dir cmd_dir);
-void iser_unreg_mem_frwr(struct iscsi_iser_task *iser_task,
-			 enum iser_data_dir cmd_dir);
+void iser_unreg_mem_fastreg(struct iscsi_iser_task *iser_task,
+			    enum iser_data_dir cmd_dir);
 
 int  iser_post_recvl(struct iser_conn *ib_conn);
 int  iser_post_recvm(struct iser_conn *ib_conn, int count);
@@ -432,12 +468,15 @@ int iser_dma_map_task_data(struct iscsi_iser_task *iser_task,
 			    enum   iser_data_dir       iser_dir,
 			    enum   dma_data_direction  dma_dir);
 
-void iser_dma_unmap_task_data(struct iscsi_iser_task *iser_task);
+void iser_dma_unmap_task_data(struct iscsi_iser_task *iser_task,
+			      struct iser_data_buf *data);
 int  iser_initialize_task_headers(struct iscsi_task *task,
 			struct iser_tx_desc *tx_desc);
 int iser_alloc_rx_descriptors(struct iser_conn *ib_conn, struct iscsi_session *session);
 int iser_create_fmr_pool(struct iser_conn *ib_conn, unsigned cmds_max);
 void iser_free_fmr_pool(struct iser_conn *ib_conn);
-int iser_create_frwr_pool(struct iser_conn *ib_conn, unsigned cmds_max);
-void iser_free_frwr_pool(struct iser_conn *ib_conn);
+int iser_create_fastreg_pool(struct iser_conn *ib_conn, unsigned cmds_max);
+void iser_free_fastreg_pool(struct iser_conn *ib_conn);
+u8 iser_check_task_pi_status(struct iscsi_iser_task *iser_task,
+			     enum iser_data_dir cmd_dir, sector_t *sector);
 #endif

drivers/infiniband/ulp/iser/iser_initiator.c

@@ -1,6 +1,6 @@
 /*
  * Copyright (c) 2004, 2005, 2006 Voltaire, Inc. All rights reserved.
- * Copyright (c) 2013 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2013-2014 Mellanox Technologies. All rights reserved.
  *
  * This software is available to you under a choice of one of two
  * licenses.  You may choose to be licensed under the terms of the GNU
@@ -49,7 +49,7 @@ static int iser_prepare_read_cmd(struct iscsi_task *task,
 
 {
 	struct iscsi_iser_task *iser_task = task->dd_data;
-	struct iser_device  *device = iser_task->iser_conn->ib_conn->device;
+	struct iser_device  *device = iser_task->ib_conn->device;
 	struct iser_regd_buf *regd_buf;
 	int err;
 	struct iser_hdr *hdr = &iser_task->desc.iser_header;
@@ -62,11 +62,22 @@ static int iser_prepare_read_cmd(struct iscsi_task *task,
 	if (err)
 		return err;
 
+	if (scsi_prot_sg_count(iser_task->sc)) {
+		struct iser_data_buf *pbuf_in = &iser_task->prot[ISER_DIR_IN];
+
+		err = iser_dma_map_task_data(iser_task,
+					     pbuf_in,
+					     ISER_DIR_IN,
+					     DMA_FROM_DEVICE);
+		if (err)
+			return err;
+	}
+
 	if (edtl > iser_task->data[ISER_DIR_IN].data_len) {
 		iser_err("Total data length: %ld, less than EDTL: "
 			 "%d, in READ cmd BHS itt: %d, conn: 0x%p\n",
 			 iser_task->data[ISER_DIR_IN].data_len, edtl,
-			 task->itt, iser_task->iser_conn);
+			 task->itt, iser_task->ib_conn);
 		return -EINVAL;
 	}
 
@@ -99,7 +110,7 @@ iser_prepare_write_cmd(struct iscsi_task *task,
 		       unsigned int edtl)
 {
 	struct iscsi_iser_task *iser_task = task->dd_data;
-	struct iser_device  *device = iser_task->iser_conn->ib_conn->device;
+	struct iser_device  *device = iser_task->ib_conn->device;
 	struct iser_regd_buf *regd_buf;
 	int err;
 	struct iser_hdr *hdr = &iser_task->desc.iser_header;
@@ -113,6 +124,17 @@ iser_prepare_write_cmd(struct iscsi_task *task,
 	if (err)
 		return err;
 
+	if (scsi_prot_sg_count(iser_task->sc)) {
+		struct iser_data_buf *pbuf_out = &iser_task->prot[ISER_DIR_OUT];
+
+		err = iser_dma_map_task_data(iser_task,
+					     pbuf_out,
+					     ISER_DIR_OUT,
+					     DMA_TO_DEVICE);
+		if (err)
+			return err;
+	}
+
 	if (edtl > iser_task->data[ISER_DIR_OUT].data_len) {
 		iser_err("Total data length: %ld, less than EDTL: %d, "
 			 "in WRITE cmd BHS itt: %d, conn: 0x%p\n",
@@ -327,7 +349,7 @@ free_login_buf:
 
 static int iser_post_rx_bufs(struct iscsi_conn *conn, struct iscsi_hdr *req)
 {
-	struct iscsi_iser_conn *iser_conn = conn->dd_data;
+	struct iser_conn *ib_conn = conn->dd_data;
 	struct iscsi_session *session = conn->session;
 
 	iser_dbg("req op %x flags %x\n", req->opcode, req->flags);
@@ -340,19 +362,18 @@ static int iser_post_rx_bufs(struct iscsi_conn *conn, struct iscsi_hdr *req)
 	 * response) and no posted send buffers left - they must have been
 	 * consumed during previous login phases.
 	 */
-	WARN_ON(iser_conn->ib_conn->post_recv_buf_count != 1);
-	WARN_ON(atomic_read(&iser_conn->ib_conn->post_send_buf_count) != 0);
+	WARN_ON(ib_conn->post_recv_buf_count != 1);
+	WARN_ON(atomic_read(&ib_conn->post_send_buf_count) != 0);
 
 	if (session->discovery_sess) {
 		iser_info("Discovery session, re-using login RX buffer\n");
 		return 0;
 	} else
 		iser_info("Normal session, posting batch of RX %d buffers\n",
-			  iser_conn->ib_conn->min_posted_rx);
+			  ib_conn->min_posted_rx);
 
 	/* Initial post receive buffers */
-	if (iser_post_recvm(iser_conn->ib_conn,
-			    iser_conn->ib_conn->min_posted_rx))
+	if (iser_post_recvm(ib_conn, ib_conn->min_posted_rx))
 		return -ENOMEM;
 
 	return 0;
@@ -364,11 +385,11 @@ static int iser_post_rx_bufs(struct iscsi_conn *conn, struct iscsi_hdr *req)
 int iser_send_command(struct iscsi_conn *conn,
 		      struct iscsi_task *task)
 {
-	struct iscsi_iser_conn *iser_conn = conn->dd_data;
+	struct iser_conn *ib_conn = conn->dd_data;
 	struct iscsi_iser_task *iser_task = task->dd_data;
 	unsigned long edtl;
 	int err;
-	struct iser_data_buf *data_buf;
+	struct iser_data_buf *data_buf, *prot_buf;
 	struct iscsi_scsi_req *hdr = (struct iscsi_scsi_req *)task->hdr;
 	struct scsi_cmnd *sc  =  task->sc;
 	struct iser_tx_desc *tx_desc = &iser_task->desc;
@@ -377,20 +398,28 @@ int iser_send_command(struct iscsi_conn *conn,
 
 	/* build the tx desc regd header and add it to the tx desc dto */
 	tx_desc->type = ISCSI_TX_SCSI_COMMAND;
-	iser_create_send_desc(iser_conn->ib_conn, tx_desc);
+	iser_create_send_desc(ib_conn, tx_desc);
 
-	if (hdr->flags & ISCSI_FLAG_CMD_READ)
+	if (hdr->flags & ISCSI_FLAG_CMD_READ) {
 		data_buf = &iser_task->data[ISER_DIR_IN];
-	else
+		prot_buf = &iser_task->prot[ISER_DIR_IN];
+	} else {
 		data_buf = &iser_task->data[ISER_DIR_OUT];
+		prot_buf = &iser_task->prot[ISER_DIR_OUT];
+	}
 
 	if (scsi_sg_count(sc)) { /* using a scatter list */
 		data_buf->buf  = scsi_sglist(sc);
 		data_buf->size = scsi_sg_count(sc);
 	}
-
 	data_buf->data_len = scsi_bufflen(sc);
 
+	if (scsi_prot_sg_count(sc)) {
+		prot_buf->buf  = scsi_prot_sglist(sc);
+		prot_buf->size = scsi_prot_sg_count(sc);
+		prot_buf->data_len = sc->prot_sdb->length;
+	}
+
 	if (hdr->flags & ISCSI_FLAG_CMD_READ) {
 		err = iser_prepare_read_cmd(task, edtl);
 		if (err)
@@ -408,7 +437,7 @@ int iser_send_command(struct iscsi_conn *conn,
 
 	iser_task->status = ISER_TASK_STATUS_STARTED;
 
-	err = iser_post_send(iser_conn->ib_conn, tx_desc);
+	err = iser_post_send(ib_conn, tx_desc);
 	if (!err)
 		return 0;
 
@@ -424,7 +453,7 @@ int iser_send_data_out(struct iscsi_conn *conn,
 		       struct iscsi_task *task,
 		       struct iscsi_data *hdr)
 {
-	struct iscsi_iser_conn *iser_conn = conn->dd_data;
+	struct iser_conn *ib_conn = conn->dd_data;
 	struct iscsi_iser_task *iser_task = task->dd_data;
 	struct iser_tx_desc *tx_desc = NULL;
 	struct iser_regd_buf *regd_buf;
@@ -473,7 +502,7 @@ int iser_send_data_out(struct iscsi_conn *conn,
 		 itt, buf_offset, data_seg_len);
 
 
-	err = iser_post_send(iser_conn->ib_conn, tx_desc);
+	err = iser_post_send(ib_conn, tx_desc);
 	if (!err)
 		return 0;
 
@@ -486,19 +515,18 @@ send_data_out_error:
 int iser_send_control(struct iscsi_conn *conn,
 		      struct iscsi_task *task)
 {
-	struct iscsi_iser_conn *iser_conn = conn->dd_data;
+	struct iser_conn *ib_conn = conn->dd_data;
 	struct iscsi_iser_task *iser_task = task->dd_data;
 	struct iser_tx_desc *mdesc = &iser_task->desc;
 	unsigned long data_seg_len;
 	int err = 0;
 	struct iser_device *device;
-	struct iser_conn *ib_conn = iser_conn->ib_conn;
 
 	/* build the tx desc regd header and add it to the tx desc dto */
 	mdesc->type = ISCSI_TX_CONTROL;
-	iser_create_send_desc(iser_conn->ib_conn, mdesc);
+	iser_create_send_desc(ib_conn, mdesc);
 
-	device = iser_conn->ib_conn->device;
+	device = ib_conn->device;
 
 	data_seg_len = ntoh24(task->hdr->dlength);
 
@@ -513,14 +541,13 @@ int iser_send_control(struct iscsi_conn *conn,
 			ib_conn->login_req_dma, task->data_count,
 			DMA_TO_DEVICE);
 
-		memcpy(iser_conn->ib_conn->login_req_buf, task->data,
-							task->data_count);
+		memcpy(ib_conn->login_req_buf, task->data, task->data_count);
 
 		ib_dma_sync_single_for_device(device->ib_device,
 			ib_conn->login_req_dma, task->data_count,
 			DMA_TO_DEVICE);
 
-		tx_dsg->addr    = iser_conn->ib_conn->login_req_dma;
+		tx_dsg->addr    = ib_conn->login_req_dma;
 		tx_dsg->length  = task->data_count;
 		tx_dsg->lkey    = device->mr->lkey;
 		mdesc->num_sge = 2;
@@ -529,7 +556,7 @@ int iser_send_control(struct iscsi_conn *conn,
 	if (task == conn->login_task) {
 		iser_dbg("op %x dsl %lx, posting login rx buffer\n",
 			 task->hdr->opcode, data_seg_len);
-		err = iser_post_recvl(iser_conn->ib_conn);
+		err = iser_post_recvl(ib_conn);
 		if (err)
 			goto send_control_error;
 		err = iser_post_rx_bufs(conn, task->hdr);
@@ -537,7 +564,7 @@ int iser_send_control(struct iscsi_conn *conn,
 			goto send_control_error;
 	}
 
-	err = iser_post_send(iser_conn->ib_conn, mdesc);
+	err = iser_post_send(ib_conn, mdesc);
 	if (!err)
 		return 0;
 
@@ -553,7 +580,6 @@ void iser_rcv_completion(struct iser_rx_desc *rx_desc,
 			 unsigned long rx_xfer_len,
 			 struct iser_conn *ib_conn)
 {
-	struct iscsi_iser_conn *conn = ib_conn->iser_conn;
 	struct iscsi_hdr *hdr;
 	u64 rx_dma;
 	int rx_buflen, outstanding, count, err;
@@ -575,17 +601,17 @@ void iser_rcv_completion(struct iser_rx_desc *rx_desc,
 	iser_dbg("op 0x%x itt 0x%x dlen %d\n", hdr->opcode,
 			hdr->itt, (int)(rx_xfer_len - ISER_HEADERS_LEN));
 
-	iscsi_iser_recv(conn->iscsi_conn, hdr,
-		rx_desc->data, rx_xfer_len - ISER_HEADERS_LEN);
+	iscsi_iser_recv(ib_conn->iscsi_conn, hdr, rx_desc->data,
+			rx_xfer_len - ISER_HEADERS_LEN);
 
 	ib_dma_sync_single_for_device(ib_conn->device->ib_device, rx_dma,
-			rx_buflen, DMA_FROM_DEVICE);
+				      rx_buflen, DMA_FROM_DEVICE);
 
 	/* decrementing conn->post_recv_buf_count only --after-- freeing the   *
 	 * task eliminates the need to worry on tasks which are completed in   *
 	 * parallel to the execution of iser_conn_term. So the code that waits *
 	 * for the posted rx bufs refcount to become zero handles everything   */
-	conn->ib_conn->post_recv_buf_count--;
+	ib_conn->post_recv_buf_count--;
 
 	if (rx_dma == ib_conn->login_resp_dma)
 		return;
@@ -635,6 +661,9 @@ void iser_task_rdma_init(struct iscsi_iser_task *iser_task)
 	iser_task->data[ISER_DIR_IN].data_len  = 0;
 	iser_task->data[ISER_DIR_OUT].data_len = 0;
 
+	iser_task->prot[ISER_DIR_IN].data_len  = 0;
+	iser_task->prot[ISER_DIR_OUT].data_len = 0;
+
 	memset(&iser_task->rdma_regd[ISER_DIR_IN], 0,
 	       sizeof(struct iser_regd_buf));
 	memset(&iser_task->rdma_regd[ISER_DIR_OUT], 0,
@@ -643,28 +672,63 @@ void iser_task_rdma_init(struct iscsi_iser_task *iser_task)
 
 void iser_task_rdma_finalize(struct iscsi_iser_task *iser_task)
 {
-	struct iser_device *device = iser_task->iser_conn->ib_conn->device;
-	int is_rdma_aligned = 1;
+	struct iser_device *device = iser_task->ib_conn->device;
+	int is_rdma_data_aligned = 1;
+	int is_rdma_prot_aligned = 1;
+	int prot_count = scsi_prot_sg_count(iser_task->sc);
 
 	/* if we were reading, copy back to unaligned sglist,
 	 * anyway dma_unmap and free the copy
 	 */
 	if (iser_task->data_copy[ISER_DIR_IN].copy_buf != NULL) {
-		is_rdma_aligned = 0;
-		iser_finalize_rdma_unaligned_sg(iser_task, ISER_DIR_IN);
+		is_rdma_data_aligned = 0;
+		iser_finalize_rdma_unaligned_sg(iser_task,
+						&iser_task->data[ISER_DIR_IN],
+						&iser_task->data_copy[ISER_DIR_IN],
+						ISER_DIR_IN);
 	}
+
 	if (iser_task->data_copy[ISER_DIR_OUT].copy_buf != NULL) {
-		is_rdma_aligned = 0;
-		iser_finalize_rdma_unaligned_sg(iser_task, ISER_DIR_OUT);
+		is_rdma_data_aligned = 0;
+		iser_finalize_rdma_unaligned_sg(iser_task,
+						&iser_task->data[ISER_DIR_OUT],
+						&iser_task->data_copy[ISER_DIR_OUT],
+						ISER_DIR_OUT);
+	}
+
+	if (iser_task->prot_copy[ISER_DIR_IN].copy_buf != NULL) {
+		is_rdma_prot_aligned = 0;
+		iser_finalize_rdma_unaligned_sg(iser_task,
+						&iser_task->prot[ISER_DIR_IN],
+						&iser_task->prot_copy[ISER_DIR_IN],
+						ISER_DIR_IN);
+	}
+
+	if (iser_task->prot_copy[ISER_DIR_OUT].copy_buf != NULL) {
+		is_rdma_prot_aligned = 0;
+		iser_finalize_rdma_unaligned_sg(iser_task,
+						&iser_task->prot[ISER_DIR_OUT],
+						&iser_task->prot_copy[ISER_DIR_OUT],
+						ISER_DIR_OUT);
 	}
 
-	if (iser_task->dir[ISER_DIR_IN])
+	if (iser_task->dir[ISER_DIR_IN]) {
 		device->iser_unreg_rdma_mem(iser_task, ISER_DIR_IN);
+		if (is_rdma_data_aligned)
+			iser_dma_unmap_task_data(iser_task,
+						 &iser_task->data[ISER_DIR_IN]);
+		if (prot_count && is_rdma_prot_aligned)
+			iser_dma_unmap_task_data(iser_task,
+						 &iser_task->prot[ISER_DIR_IN]);
+	}
 
-	if (iser_task->dir[ISER_DIR_OUT])
+	if (iser_task->dir[ISER_DIR_OUT]) {
 		device->iser_unreg_rdma_mem(iser_task, ISER_DIR_OUT);
-
-       /* if the data was unaligned, it was already unmapped and then copied */
-       if (is_rdma_aligned)
-		iser_dma_unmap_task_data(iser_task);
+		if (is_rdma_data_aligned)
+			iser_dma_unmap_task_data(iser_task,
+						 &iser_task->data[ISER_DIR_OUT]);
+		if (prot_count && is_rdma_prot_aligned)
+			iser_dma_unmap_task_data(iser_task,
+						 &iser_task->prot[ISER_DIR_OUT]);
+	}
 }

drivers/infiniband/ulp/iser/iser_memory.c

@@ -1,6 +1,6 @@
 /*
  * Copyright (c) 2004, 2005, 2006 Voltaire, Inc. All rights reserved.
- * Copyright (c) 2013 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2013-2014 Mellanox Technologies. All rights reserved.
  *
  * This software is available to you under a choice of one of two
  * licenses.  You may choose to be licensed under the terms of the GNU
@@ -45,13 +45,19 @@
  * iser_start_rdma_unaligned_sg
  */
 static int iser_start_rdma_unaligned_sg(struct iscsi_iser_task *iser_task,
+					struct iser_data_buf *data,
+					struct iser_data_buf *data_copy,
 					enum iser_data_dir cmd_dir)
 {
-	int dma_nents;
-	struct ib_device *dev;
+	struct ib_device *dev = iser_task->ib_conn->device->ib_device;
+	struct scatterlist *sgl = (struct scatterlist *)data->buf;
+	struct scatterlist *sg;
 	char *mem = NULL;
-	struct iser_data_buf *data = &iser_task->data[cmd_dir];
-	unsigned long  cmd_data_len = data->data_len;
+	unsigned long  cmd_data_len = 0;
+	int dma_nents, i;
+
+	for_each_sg(sgl, sg, data->size, i)
+		cmd_data_len += ib_sg_dma_len(dev, sg);
 
 	if (cmd_data_len > ISER_KMALLOC_THRESHOLD)
 		mem = (void *)__get_free_pages(GFP_ATOMIC,
@@ -61,17 +67,16 @@ static int iser_start_rdma_unaligned_sg(struct iscsi_iser_task *iser_task,
 
 	if (mem == NULL) {
 		iser_err("Failed to allocate mem size %d %d for copying sglist\n",
-			 data->size,(int)cmd_data_len);
+			 data->size, (int)cmd_data_len);
 		return -ENOMEM;
 	}
 
 	if (cmd_dir == ISER_DIR_OUT) {
 		/* copy the unaligned sg the buffer which is used for RDMA */
-		struct scatterlist *sgl = (struct scatterlist *)data->buf;
-		struct scatterlist *sg;
 		int i;
 		char *p, *from;
 
+		sgl = (struct scatterlist *)data->buf;
 		p = mem;
 		for_each_sg(sgl, sg, data->size, i) {
 			from = kmap_atomic(sg_page(sg));
@@ -83,39 +88,37 @@ static int iser_start_rdma_unaligned_sg(struct iscsi_iser_task *iser_task,
 		}
 	}
 
-	sg_init_one(&iser_task->data_copy[cmd_dir].sg_single, mem, cmd_data_len);
-	iser_task->data_copy[cmd_dir].buf  =
-		&iser_task->data_copy[cmd_dir].sg_single;
-	iser_task->data_copy[cmd_dir].size = 1;
+	sg_init_one(&data_copy->sg_single, mem, cmd_data_len);
+	data_copy->buf = &data_copy->sg_single;
+	data_copy->size = 1;
+	data_copy->copy_buf = mem;
 
-	iser_task->data_copy[cmd_dir].copy_buf  = mem;
-
-	dev = iser_task->iser_conn->ib_conn->device->ib_device;
-	dma_nents = ib_dma_map_sg(dev,
-				  &iser_task->data_copy[cmd_dir].sg_single,
-				  1,
+	dma_nents = ib_dma_map_sg(dev, &data_copy->sg_single, 1,
 				  (cmd_dir == ISER_DIR_OUT) ?
 				  DMA_TO_DEVICE : DMA_FROM_DEVICE);
 	BUG_ON(dma_nents == 0);
 
-	iser_task->data_copy[cmd_dir].dma_nents = dma_nents;
+	data_copy->dma_nents = dma_nents;
+	data_copy->data_len = cmd_data_len;
+
 	return 0;
 }
 
 /**
  * iser_finalize_rdma_unaligned_sg
  */
+
 void iser_finalize_rdma_unaligned_sg(struct iscsi_iser_task *iser_task,
-				     enum iser_data_dir         cmd_dir)
+				     struct iser_data_buf *data,
+				     struct iser_data_buf *data_copy,
+				     enum iser_data_dir cmd_dir)
 {
 	struct ib_device *dev;
-	struct iser_data_buf *mem_copy;
 	unsigned long  cmd_data_len;
 
-	dev = iser_task->iser_conn->ib_conn->device->ib_device;
-	mem_copy = &iser_task->data_copy[cmd_dir];
+	dev = iser_task->ib_conn->device->ib_device;
 
-	ib_dma_unmap_sg(dev, &mem_copy->sg_single, 1,
+	ib_dma_unmap_sg(dev, &data_copy->sg_single, 1,
 			(cmd_dir == ISER_DIR_OUT) ?
 			DMA_TO_DEVICE : DMA_FROM_DEVICE);
 
@@ -127,10 +130,10 @@ void iser_finalize_rdma_unaligned_sg(struct iscsi_iser_task *iser_task,
 		int i;
 
 		/* copy back read RDMA to unaligned sg */
-		mem	= mem_copy->copy_buf;
+		mem = data_copy->copy_buf;
 
-		sgl	= (struct scatterlist *)iser_task->data[ISER_DIR_IN].buf;
-		sg_size = iser_task->data[ISER_DIR_IN].size;
+		sgl = (struct scatterlist *)data->buf;
+		sg_size = data->size;
 
 		p = mem;
 		for_each_sg(sgl, sg, sg_size, i) {
@@ -143,15 +146,15 @@ void iser_finalize_rdma_unaligned_sg(struct iscsi_iser_task *iser_task,
 		}
 	}
 
-	cmd_data_len = iser_task->data[cmd_dir].data_len;
+	cmd_data_len = data->data_len;
 
 	if (cmd_data_len > ISER_KMALLOC_THRESHOLD)
-		free_pages((unsigned long)mem_copy->copy_buf,
+		free_pages((unsigned long)data_copy->copy_buf,
 			   ilog2(roundup_pow_of_two(cmd_data_len)) - PAGE_SHIFT);
 	else
-		kfree(mem_copy->copy_buf);
+		kfree(data_copy->copy_buf);
 
-	mem_copy->copy_buf = NULL;
+	data_copy->copy_buf = NULL;
 }
 
 #define IS_4K_ALIGNED(addr)	((((unsigned long)addr) & ~MASK_4K) == 0)
@@ -319,7 +322,7 @@ int iser_dma_map_task_data(struct iscsi_iser_task *iser_task,
 	struct ib_device *dev;
 
 	iser_task->dir[iser_dir] = 1;
-	dev = iser_task->iser_conn->ib_conn->device->ib_device;
+	dev = iser_task->ib_conn->device->ib_device;
 
 	data->dma_nents = ib_dma_map_sg(dev, data->buf, data->size, dma_dir);
 	if (data->dma_nents == 0) {
@@ -329,31 +332,23 @@ int iser_dma_map_task_data(struct iscsi_iser_task *iser_task,
 	return 0;
 }
 
-void iser_dma_unmap_task_data(struct iscsi_iser_task *iser_task)
+void iser_dma_unmap_task_data(struct iscsi_iser_task *iser_task,
+			      struct iser_data_buf *data)
 {
 	struct ib_device *dev;
-	struct iser_data_buf *data;
 
-	dev = iser_task->iser_conn->ib_conn->device->ib_device;
-
-	if (iser_task->dir[ISER_DIR_IN]) {
-		data = &iser_task->data[ISER_DIR_IN];
-		ib_dma_unmap_sg(dev, data->buf, data->size, DMA_FROM_DEVICE);
-	}
-
-	if (iser_task->dir[ISER_DIR_OUT]) {
-		data = &iser_task->data[ISER_DIR_OUT];
-		ib_dma_unmap_sg(dev, data->buf, data->size, DMA_TO_DEVICE);
-	}
+	dev = iser_task->ib_conn->device->ib_device;
+	ib_dma_unmap_sg(dev, data->buf, data->size, DMA_FROM_DEVICE);
 }
 
 static int fall_to_bounce_buf(struct iscsi_iser_task *iser_task,
 			      struct ib_device *ibdev,
+			      struct iser_data_buf *mem,
+			      struct iser_data_buf *mem_copy,
 			      enum iser_data_dir cmd_dir,
 			      int aligned_len)
 {
-	struct iscsi_conn    *iscsi_conn = iser_task->iser_conn->iscsi_conn;
-	struct iser_data_buf *mem = &iser_task->data[cmd_dir];
+	struct iscsi_conn    *iscsi_conn = iser_task->ib_conn->iscsi_conn;
 
 	iscsi_conn->fmr_unalign_cnt++;
 	iser_warn("rdma alignment violation (%d/%d aligned) or FMR not supported\n",
@@ -363,12 +358,12 @@ static int fall_to_bounce_buf(struct iscsi_iser_task *iser_task,
 		iser_data_buf_dump(mem, ibdev);
 
 	/* unmap the command data before accessing it */
-	iser_dma_unmap_task_data(iser_task);
+	iser_dma_unmap_task_data(iser_task, mem);
 
 	/* allocate copy buf, if we are writing, copy the */
 	/* unaligned scatterlist, dma map the copy        */
-	if (iser_start_rdma_unaligned_sg(iser_task, cmd_dir) != 0)
-			return -ENOMEM;
+	if (iser_start_rdma_unaligned_sg(iser_task, mem, mem_copy, cmd_dir) != 0)
+		return -ENOMEM;
 
 	return 0;
 }
@@ -382,7 +377,7 @@ static int fall_to_bounce_buf(struct iscsi_iser_task *iser_task,
 int iser_reg_rdma_mem_fmr(struct iscsi_iser_task *iser_task,
 			  enum iser_data_dir cmd_dir)
 {
-	struct iser_conn     *ib_conn = iser_task->iser_conn->ib_conn;
+	struct iser_conn     *ib_conn = iser_task->ib_conn;
 	struct iser_device   *device = ib_conn->device;
 	struct ib_device     *ibdev = device->ib_device;
 	struct iser_data_buf *mem = &iser_task->data[cmd_dir];
@@ -396,7 +391,8 @@ int iser_reg_rdma_mem_fmr(struct iscsi_iser_task *iser_task,
 
 	aligned_len = iser_data_buf_aligned_len(mem, ibdev);
 	if (aligned_len != mem->dma_nents) {
-		err = fall_to_bounce_buf(iser_task, ibdev,
+		err = fall_to_bounce_buf(iser_task, ibdev, mem,
+					 &iser_task->data_copy[cmd_dir],
 					 cmd_dir, aligned_len);
 		if (err) {
 			iser_err("failed to allocate bounce buffer\n");
@@ -422,8 +418,8 @@ int iser_reg_rdma_mem_fmr(struct iscsi_iser_task *iser_task,
 			 (unsigned long)regd_buf->reg.va,
 			 (unsigned long)regd_buf->reg.len);
 	} else { /* use FMR for multiple dma entries */
-		iser_page_vec_build(mem, ib_conn->fastreg.fmr.page_vec, ibdev);
-		err = iser_reg_page_vec(ib_conn, ib_conn->fastreg.fmr.page_vec,
+		iser_page_vec_build(mem, ib_conn->fmr.page_vec, ibdev);
+		err = iser_reg_page_vec(ib_conn, ib_conn->fmr.page_vec,
 					&regd_buf->reg);
 		if (err && err != -EAGAIN) {
 			iser_data_buf_dump(mem, ibdev);
@@ -431,12 +427,12 @@ int iser_reg_rdma_mem_fmr(struct iscsi_iser_task *iser_task,
 				 mem->dma_nents,
 				 ntoh24(iser_task->desc.iscsi_header.dlength));
 			iser_err("page_vec: data_size = 0x%x, length = %d, offset = 0x%x\n",
-				 ib_conn->fastreg.fmr.page_vec->data_size,
-				 ib_conn->fastreg.fmr.page_vec->length,
-				 ib_conn->fastreg.fmr.page_vec->offset);
-			for (i = 0; i < ib_conn->fastreg.fmr.page_vec->length; i++)
+				 ib_conn->fmr.page_vec->data_size,
+				 ib_conn->fmr.page_vec->length,
+				 ib_conn->fmr.page_vec->offset);
+			for (i = 0; i < ib_conn->fmr.page_vec->length; i++)
 				iser_err("page_vec[%d] = 0x%llx\n", i,
-					 (unsigned long long) ib_conn->fastreg.fmr.page_vec->pages[i]);
+					 (unsigned long long) ib_conn->fmr.page_vec->pages[i]);
 		}
 		if (err)
 			return err;
@@ -444,94 +440,280 @@ int iser_reg_rdma_mem_fmr(struct iscsi_iser_task *iser_task,
 	return 0;
 }
 
-static int iser_fast_reg_mr(struct fast_reg_descriptor *desc,
-			    struct iser_conn *ib_conn,
+static inline enum ib_t10_dif_type
+scsi2ib_prot_type(unsigned char prot_type)
+{
+	switch (prot_type) {
+	case SCSI_PROT_DIF_TYPE0:
+		return IB_T10DIF_NONE;
+	case SCSI_PROT_DIF_TYPE1:
+		return IB_T10DIF_TYPE1;
+	case SCSI_PROT_DIF_TYPE2:
+		return IB_T10DIF_TYPE2;
+	case SCSI_PROT_DIF_TYPE3:
+		return IB_T10DIF_TYPE3;
+	default:
+		return IB_T10DIF_NONE;
+	}
+}
+
+
+static int
+iser_set_sig_attrs(struct scsi_cmnd *sc, struct ib_sig_attrs *sig_attrs)
+{
+	unsigned char scsi_ptype = scsi_get_prot_type(sc);
+
+	sig_attrs->mem.sig_type = IB_SIG_TYPE_T10_DIF;
+	sig_attrs->wire.sig_type = IB_SIG_TYPE_T10_DIF;
+	sig_attrs->mem.sig.dif.pi_interval = sc->device->sector_size;
+	sig_attrs->wire.sig.dif.pi_interval = sc->device->sector_size;
+
+	switch (scsi_get_prot_op(sc)) {
+	case SCSI_PROT_WRITE_INSERT:
+	case SCSI_PROT_READ_STRIP:
+		sig_attrs->mem.sig.dif.type = IB_T10DIF_NONE;
+		sig_attrs->wire.sig.dif.type = scsi2ib_prot_type(scsi_ptype);
+		sig_attrs->wire.sig.dif.bg_type = IB_T10DIF_CRC;
+		sig_attrs->wire.sig.dif.ref_tag = scsi_get_lba(sc) &
+						  0xffffffff;
+		break;
+	case SCSI_PROT_READ_INSERT:
+	case SCSI_PROT_WRITE_STRIP:
+		sig_attrs->mem.sig.dif.type = scsi2ib_prot_type(scsi_ptype);
+		sig_attrs->mem.sig.dif.bg_type = IB_T10DIF_CRC;
+		sig_attrs->mem.sig.dif.ref_tag = scsi_get_lba(sc) &
+						 0xffffffff;
+		sig_attrs->wire.sig.dif.type = IB_T10DIF_NONE;
+		break;
+	case SCSI_PROT_READ_PASS:
+	case SCSI_PROT_WRITE_PASS:
+		sig_attrs->mem.sig.dif.type = scsi2ib_prot_type(scsi_ptype);
+		sig_attrs->mem.sig.dif.bg_type = IB_T10DIF_CRC;
+		sig_attrs->mem.sig.dif.ref_tag = scsi_get_lba(sc) &
+						 0xffffffff;
+		sig_attrs->wire.sig.dif.type = scsi2ib_prot_type(scsi_ptype);
+		sig_attrs->wire.sig.dif.bg_type = IB_T10DIF_CRC;
+		sig_attrs->wire.sig.dif.ref_tag = scsi_get_lba(sc) &
+						  0xffffffff;
+		break;
+	default:
+		iser_err("Unsupported PI operation %d\n",
+			 scsi_get_prot_op(sc));
+		return -EINVAL;
+	}
+	return 0;
+}
+
+
+static int
+iser_set_prot_checks(struct scsi_cmnd *sc, u8 *mask)
+{
+	switch (scsi_get_prot_type(sc)) {
+	case SCSI_PROT_DIF_TYPE0:
+		*mask = 0x0;
+		break;
+	case SCSI_PROT_DIF_TYPE1:
+	case SCSI_PROT_DIF_TYPE2:
+		*mask = ISER_CHECK_GUARD | ISER_CHECK_REFTAG;
+		break;
+	case SCSI_PROT_DIF_TYPE3:
+		*mask = ISER_CHECK_GUARD;
+		break;
+	default:
+		iser_err("Unsupported protection type %d\n",
+			 scsi_get_prot_type(sc));
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int
+iser_reg_sig_mr(struct iscsi_iser_task *iser_task,
+		struct fast_reg_descriptor *desc, struct ib_sge *data_sge,
+		struct ib_sge *prot_sge, struct ib_sge *sig_sge)
+{
+	struct iser_conn *ib_conn = iser_task->ib_conn;
+	struct iser_pi_context *pi_ctx = desc->pi_ctx;
+	struct ib_send_wr sig_wr, inv_wr;
+	struct ib_send_wr *bad_wr, *wr = NULL;
+	struct ib_sig_attrs sig_attrs;
+	int ret;
+	u32 key;
+
+	memset(&sig_attrs, 0, sizeof(sig_attrs));
+	ret = iser_set_sig_attrs(iser_task->sc, &sig_attrs);
+	if (ret)
+		goto err;
+
+	ret = iser_set_prot_checks(iser_task->sc, &sig_attrs.check_mask);
+	if (ret)
+		goto err;
+
+	if (!(desc->reg_indicators & ISER_SIG_KEY_VALID)) {
+		memset(&inv_wr, 0, sizeof(inv_wr));
+		inv_wr.opcode = IB_WR_LOCAL_INV;
+		inv_wr.wr_id = ISER_FASTREG_LI_WRID;
+		inv_wr.ex.invalidate_rkey = pi_ctx->sig_mr->rkey;
+		wr = &inv_wr;
+		/* Bump the key */
+		key = (u8)(pi_ctx->sig_mr->rkey & 0x000000FF);
+		ib_update_fast_reg_key(pi_ctx->sig_mr, ++key);
+	}
+
+	memset(&sig_wr, 0, sizeof(sig_wr));
+	sig_wr.opcode = IB_WR_REG_SIG_MR;
+	sig_wr.wr_id = ISER_FASTREG_LI_WRID;
+	sig_wr.sg_list = data_sge;
+	sig_wr.num_sge = 1;
+	sig_wr.wr.sig_handover.sig_attrs = &sig_attrs;
+	sig_wr.wr.sig_handover.sig_mr = pi_ctx->sig_mr;
+	if (scsi_prot_sg_count(iser_task->sc))
+		sig_wr.wr.sig_handover.prot = prot_sge;
+	sig_wr.wr.sig_handover.access_flags = IB_ACCESS_LOCAL_WRITE |
+					      IB_ACCESS_REMOTE_READ |
+					      IB_ACCESS_REMOTE_WRITE;
+
+	if (!wr)
+		wr = &sig_wr;
+	else
+		wr->next = &sig_wr;
+
+	ret = ib_post_send(ib_conn->qp, wr, &bad_wr);
+	if (ret) {
+		iser_err("reg_sig_mr failed, ret:%d\n", ret);
+		goto err;
+	}
+	desc->reg_indicators &= ~ISER_SIG_KEY_VALID;
+
+	sig_sge->lkey = pi_ctx->sig_mr->lkey;
+	sig_sge->addr = 0;
+	sig_sge->length = data_sge->length + prot_sge->length;
+	if (scsi_get_prot_op(iser_task->sc) == SCSI_PROT_WRITE_INSERT ||
+	    scsi_get_prot_op(iser_task->sc) == SCSI_PROT_READ_STRIP) {
+		sig_sge->length += (data_sge->length /
+				   iser_task->sc->device->sector_size) * 8;
+	}
+
+	iser_dbg("sig_sge: addr: 0x%llx  length: %u lkey: 0x%x\n",
+		 sig_sge->addr, sig_sge->length,
+		 sig_sge->lkey);
+err:
+	return ret;
+}
+
+static int iser_fast_reg_mr(struct iscsi_iser_task *iser_task,
 			    struct iser_regd_buf *regd_buf,
-			    u32 offset, unsigned int data_size,
-			    unsigned int page_list_len)
+			    struct iser_data_buf *mem,
+			    enum iser_reg_indicator ind,
+			    struct ib_sge *sge)
 {
+	struct fast_reg_descriptor *desc = regd_buf->reg.mem_h;
+	struct iser_conn *ib_conn = iser_task->ib_conn;
+	struct iser_device *device = ib_conn->device;
+	struct ib_device *ibdev = device->ib_device;
+	struct ib_mr *mr;
+	struct ib_fast_reg_page_list *frpl;
 	struct ib_send_wr fastreg_wr, inv_wr;
 	struct ib_send_wr *bad_wr, *wr = NULL;
 	u8 key;
-	int ret;
+	int ret, offset, size, plen;
+
+	/* if there a single dma entry, dma mr suffices */
+	if (mem->dma_nents == 1) {
+		struct scatterlist *sg = (struct scatterlist *)mem->buf;
 
-	if (!desc->valid) {
+		sge->lkey = device->mr->lkey;
+		sge->addr   = ib_sg_dma_address(ibdev, &sg[0]);
+		sge->length  = ib_sg_dma_len(ibdev, &sg[0]);
+
+		iser_dbg("Single DMA entry: lkey=0x%x, addr=0x%llx, length=0x%x\n",
+			 sge->lkey, sge->addr, sge->length);
+		return 0;
+	}
+
+	if (ind == ISER_DATA_KEY_VALID) {
+		mr = desc->data_mr;
+		frpl = desc->data_frpl;
+	} else {
+		mr = desc->pi_ctx->prot_mr;
+		frpl = desc->pi_ctx->prot_frpl;
+	}
+
+	plen = iser_sg_to_page_vec(mem, device->ib_device, frpl->page_list,
+				   &offset, &size);
+	if (plen * SIZE_4K < size) {
+		iser_err("fast reg page_list too short to hold this SG\n");
+		return -EINVAL;
+	}
+
+	if (!(desc->reg_indicators & ind)) {
 		memset(&inv_wr, 0, sizeof(inv_wr));
+		inv_wr.wr_id = ISER_FASTREG_LI_WRID;
 		inv_wr.opcode = IB_WR_LOCAL_INV;
-		inv_wr.send_flags = IB_SEND_SIGNALED;
-		inv_wr.ex.invalidate_rkey = desc->data_mr->rkey;
+		inv_wr.ex.invalidate_rkey = mr->rkey;
 		wr = &inv_wr;
 		/* Bump the key */
-		key = (u8)(desc->data_mr->rkey & 0x000000FF);
-		ib_update_fast_reg_key(desc->data_mr, ++key);
+		key = (u8)(mr->rkey & 0x000000FF);
+		ib_update_fast_reg_key(mr, ++key);
 	}
 
 	/* Prepare FASTREG WR */
 	memset(&fastreg_wr, 0, sizeof(fastreg_wr));
+	fastreg_wr.wr_id = ISER_FASTREG_LI_WRID;
 	fastreg_wr.opcode = IB_WR_FAST_REG_MR;
-	fastreg_wr.send_flags = IB_SEND_SIGNALED;
-	fastreg_wr.wr.fast_reg.iova_start = desc->data_frpl->page_list[0] + offset;
-	fastreg_wr.wr.fast_reg.page_list = desc->data_frpl;
-	fastreg_wr.wr.fast_reg.page_list_len = page_list_len;
+	fastreg_wr.wr.fast_reg.iova_start = frpl->page_list[0] + offset;
+	fastreg_wr.wr.fast_reg.page_list = frpl;
+	fastreg_wr.wr.fast_reg.page_list_len = plen;
 	fastreg_wr.wr.fast_reg.page_shift = SHIFT_4K;
-	fastreg_wr.wr.fast_reg.length = data_size;
-	fastreg_wr.wr.fast_reg.rkey = desc->data_mr->rkey;
+	fastreg_wr.wr.fast_reg.length = size;
+	fastreg_wr.wr.fast_reg.rkey = mr->rkey;
 	fastreg_wr.wr.fast_reg.access_flags = (IB_ACCESS_LOCAL_WRITE  |
 					       IB_ACCESS_REMOTE_WRITE |
 					       IB_ACCESS_REMOTE_READ);
 
-	if (!wr) {
+	if (!wr)
 		wr = &fastreg_wr;
-		atomic_inc(&ib_conn->post_send_buf_count);
-	} else {
+	else
 		wr->next = &fastreg_wr;
-		atomic_add(2, &ib_conn->post_send_buf_count);
-	}
 
 	ret = ib_post_send(ib_conn->qp, wr, &bad_wr);
 	if (ret) {
-		if (bad_wr->next)
-			atomic_sub(2, &ib_conn->post_send_buf_count);
-		else
-			atomic_dec(&ib_conn->post_send_buf_count);
 		iser_err("fast registration failed, ret:%d\n", ret);
 		return ret;
 	}
-	desc->valid = false;
+	desc->reg_indicators &= ~ind;
 
-	regd_buf->reg.mem_h = desc;
-	regd_buf->reg.lkey = desc->data_mr->lkey;
-	regd_buf->reg.rkey = desc->data_mr->rkey;
-	regd_buf->reg.va = desc->data_frpl->page_list[0] + offset;
-	regd_buf->reg.len = data_size;
-	regd_buf->reg.is_mr = 1;
+	sge->lkey = mr->lkey;
+	sge->addr = frpl->page_list[0] + offset;
+	sge->length = size;
 
 	return ret;
 }
 
 /**
- * iser_reg_rdma_mem_frwr - Registers memory intended for RDMA,
+ * iser_reg_rdma_mem_fastreg - Registers memory intended for RDMA,
  * using Fast Registration WR (if possible) obtaining rkey and va
  *
  * returns 0 on success, errno code on failure
  */
-int iser_reg_rdma_mem_frwr(struct iscsi_iser_task *iser_task,
-			   enum iser_data_dir cmd_dir)
+int iser_reg_rdma_mem_fastreg(struct iscsi_iser_task *iser_task,
+			      enum iser_data_dir cmd_dir)
 {
-	struct iser_conn *ib_conn = iser_task->iser_conn->ib_conn;
+	struct iser_conn *ib_conn = iser_task->ib_conn;
 	struct iser_device *device = ib_conn->device;
 	struct ib_device *ibdev = device->ib_device;
 	struct iser_data_buf *mem = &iser_task->data[cmd_dir];
 	struct iser_regd_buf *regd_buf = &iser_task->rdma_regd[cmd_dir];
-	struct fast_reg_descriptor *desc;
-	unsigned int data_size, page_list_len;
+	struct fast_reg_descriptor *desc = NULL;
+	struct ib_sge data_sge;
 	int err, aligned_len;
 	unsigned long flags;
-	u32 offset;
 
 	aligned_len = iser_data_buf_aligned_len(mem, ibdev);
 	if (aligned_len != mem->dma_nents) {
-		err = fall_to_bounce_buf(iser_task, ibdev,
+		err = fall_to_bounce_buf(iser_task, ibdev, mem,
+					 &iser_task->data_copy[cmd_dir],
 					 cmd_dir, aligned_len);
 		if (err) {
 			iser_err("failed to allocate bounce buffer\n");
@@ -540,41 +722,79 @@ int iser_reg_rdma_mem_frwr(struct iscsi_iser_task *iser_task,
 		mem = &iser_task->data_copy[cmd_dir];
 	}
 
-	/* if there a single dma entry, dma mr suffices */
-	if (mem->dma_nents == 1) {
-		struct scatterlist *sg = (struct scatterlist *)mem->buf;
-
-		regd_buf->reg.lkey = device->mr->lkey;
-		regd_buf->reg.rkey = device->mr->rkey;
-		regd_buf->reg.len  = ib_sg_dma_len(ibdev, &sg[0]);
-		regd_buf->reg.va   = ib_sg_dma_address(ibdev, &sg[0]);
-		regd_buf->reg.is_mr = 0;
-	} else {
+	if (mem->dma_nents != 1 ||
+	    scsi_get_prot_op(iser_task->sc) != SCSI_PROT_NORMAL) {
 		spin_lock_irqsave(&ib_conn->lock, flags);
-		desc = list_first_entry(&ib_conn->fastreg.frwr.pool,
+		desc = list_first_entry(&ib_conn->fastreg.pool,
 					struct fast_reg_descriptor, list);
 		list_del(&desc->list);
 		spin_unlock_irqrestore(&ib_conn->lock, flags);
-		page_list_len = iser_sg_to_page_vec(mem, device->ib_device,
-						    desc->data_frpl->page_list,
-						    &offset, &data_size);
-
-		if (page_list_len * SIZE_4K < data_size) {
-			iser_err("fast reg page_list too short to hold this SG\n");
-			err = -EINVAL;
-			goto err_reg;
+		regd_buf->reg.mem_h = desc;
+	}
+
+	err = iser_fast_reg_mr(iser_task, regd_buf, mem,
+			       ISER_DATA_KEY_VALID, &data_sge);
+	if (err)
+		goto err_reg;
+
+	if (scsi_get_prot_op(iser_task->sc) != SCSI_PROT_NORMAL) {
+		struct ib_sge prot_sge, sig_sge;
+
+		memset(&prot_sge, 0, sizeof(prot_sge));
+		if (scsi_prot_sg_count(iser_task->sc)) {
+			mem = &iser_task->prot[cmd_dir];
+			aligned_len = iser_data_buf_aligned_len(mem, ibdev);
+			if (aligned_len != mem->dma_nents) {
+				err = fall_to_bounce_buf(iser_task, ibdev, mem,
+							 &iser_task->prot_copy[cmd_dir],
+							 cmd_dir, aligned_len);
+				if (err) {
+					iser_err("failed to allocate bounce buffer\n");
+					return err;
+				}
+				mem = &iser_task->prot_copy[cmd_dir];
+			}
+
+			err = iser_fast_reg_mr(iser_task, regd_buf, mem,
+					       ISER_PROT_KEY_VALID, &prot_sge);
+			if (err)
+				goto err_reg;
 		}
 
-		err = iser_fast_reg_mr(desc, ib_conn, regd_buf,
-				       offset, data_size, page_list_len);
-		if (err)
-			goto err_reg;
+		err = iser_reg_sig_mr(iser_task, desc, &data_sge,
+				      &prot_sge, &sig_sge);
+		if (err) {
+			iser_err("Failed to register signature mr\n");
+			return err;
+		}
+		desc->reg_indicators |= ISER_FASTREG_PROTECTED;
+
+		regd_buf->reg.lkey = sig_sge.lkey;
+		regd_buf->reg.rkey = desc->pi_ctx->sig_mr->rkey;
+		regd_buf->reg.va = sig_sge.addr;
+		regd_buf->reg.len = sig_sge.length;
+		regd_buf->reg.is_mr = 1;
+	} else {
+		if (desc) {
+			regd_buf->reg.rkey = desc->data_mr->rkey;
+			regd_buf->reg.is_mr = 1;
+		} else {
+			regd_buf->reg.rkey = device->mr->rkey;
+			regd_buf->reg.is_mr = 0;
+		}
+
+		regd_buf->reg.lkey = data_sge.lkey;
+		regd_buf->reg.va = data_sge.addr;
+		regd_buf->reg.len = data_sge.length;
 	}
 
 	return 0;
 err_reg:
-	spin_lock_irqsave(&ib_conn->lock, flags);
-	list_add_tail(&desc->list, &ib_conn->fastreg.frwr.pool);
-	spin_unlock_irqrestore(&ib_conn->lock, flags);
+	if (desc) {
+		spin_lock_irqsave(&ib_conn->lock, flags);
+		list_add_tail(&desc->list, &ib_conn->fastreg.pool);
+		spin_unlock_irqrestore(&ib_conn->lock, flags);
+	}
+
 	return err;
 }

drivers/infiniband/ulp/iser/iser_verbs.c

@@ -1,7 +1,7 @@
 /*
  * Copyright (c) 2004, 2005, 2006 Voltaire, Inc. All rights reserved.
  * Copyright (c) 2005, 2006 Cisco Systems.  All rights reserved.
- * Copyright (c) 2013 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2013-2014 Mellanox Technologies. All rights reserved.
  *
  * This software is available to you under a choice of one of two
  * licenses.  You may choose to be licensed under the terms of the GNU
@@ -71,17 +71,14 @@ static void iser_event_handler(struct ib_event_handler *handler,
  */
 static int iser_create_device_ib_res(struct iser_device *device)
 {
-	int i, j;
 	struct iser_cq_desc *cq_desc;
-	struct ib_device_attr *dev_attr;
+	struct ib_device_attr *dev_attr = &device->dev_attr;
+	int ret, i, j;
 
-	dev_attr = kmalloc(sizeof(*dev_attr), GFP_KERNEL);
-	if (!dev_attr)
-		return -ENOMEM;
-
-	if (ib_query_device(device->ib_device, dev_attr)) {
+	ret = ib_query_device(device->ib_device, dev_attr);
+	if (ret) {
 		pr_warn("Query device failed for %s\n", device->ib_device->name);
-		goto dev_attr_err;
+		return ret;
 	}
 
 	/* Assign function handles  - based on FMR support */
@@ -94,14 +91,14 @@ static int iser_create_device_ib_res(struct iser_device *device)
 		device->iser_unreg_rdma_mem = iser_unreg_mem_fmr;
 	} else
 	if (dev_attr->device_cap_flags & IB_DEVICE_MEM_MGT_EXTENSIONS) {
-		iser_info("FRWR supported, using FRWR for registration\n");
-		device->iser_alloc_rdma_reg_res = iser_create_frwr_pool;
-		device->iser_free_rdma_reg_res = iser_free_frwr_pool;
-		device->iser_reg_rdma_mem = iser_reg_rdma_mem_frwr;
-		device->iser_unreg_rdma_mem = iser_unreg_mem_frwr;
+		iser_info("FastReg supported, using FastReg for registration\n");
+		device->iser_alloc_rdma_reg_res = iser_create_fastreg_pool;
+		device->iser_free_rdma_reg_res = iser_free_fastreg_pool;
+		device->iser_reg_rdma_mem = iser_reg_rdma_mem_fastreg;
+		device->iser_unreg_rdma_mem = iser_unreg_mem_fastreg;
 	} else {
-		iser_err("IB device does not support FMRs nor FRWRs, can't register memory\n");
-		goto dev_attr_err;
+		iser_err("IB device does not support FMRs nor FastRegs, can't register memory\n");
+		return -1;
 	}
 
 	device->cqs_used = min(ISER_MAX_CQ, device->ib_device->num_comp_vectors);
@@ -158,7 +155,6 @@ static int iser_create_device_ib_res(struct iser_device *device)
 	if (ib_register_event_handler(&device->event_handler))
 		goto handler_err;
 
-	kfree(dev_attr);
 	return 0;
 
 handler_err:
@@ -178,8 +174,6 @@ pd_err:
 	kfree(device->cq_desc);
 cq_desc_err:
 	iser_err("failed to allocate an IB resource\n");
-dev_attr_err:
-	kfree(dev_attr);
 	return -1;
 }
 
@@ -221,13 +215,13 @@ int iser_create_fmr_pool(struct iser_conn *ib_conn, unsigned cmds_max)
 	struct ib_fmr_pool_param params;
 	int ret = -ENOMEM;
 
-	ib_conn->fastreg.fmr.page_vec = kmalloc(sizeof(struct iser_page_vec) +
-						(sizeof(u64)*(ISCSI_ISER_SG_TABLESIZE + 1)),
-						GFP_KERNEL);
-	if (!ib_conn->fastreg.fmr.page_vec)
+	ib_conn->fmr.page_vec = kmalloc(sizeof(*ib_conn->fmr.page_vec) +
+					(sizeof(u64)*(ISCSI_ISER_SG_TABLESIZE + 1)),
+					GFP_KERNEL);
+	if (!ib_conn->fmr.page_vec)
 		return ret;
 
-	ib_conn->fastreg.fmr.page_vec->pages = (u64 *)(ib_conn->fastreg.fmr.page_vec + 1);
+	ib_conn->fmr.page_vec->pages = (u64 *)(ib_conn->fmr.page_vec + 1);
 
 	params.page_shift        = SHIFT_4K;
 	/* when the first/last SG element are not start/end *
@@ -243,16 +237,16 @@ int iser_create_fmr_pool(struct iser_conn *ib_conn, unsigned cmds_max)
 				    IB_ACCESS_REMOTE_WRITE |
 				    IB_ACCESS_REMOTE_READ);
 
-	ib_conn->fastreg.fmr.pool = ib_create_fmr_pool(device->pd, &params);
-	if (!IS_ERR(ib_conn->fastreg.fmr.pool))
+	ib_conn->fmr.pool = ib_create_fmr_pool(device->pd, &params);
+	if (!IS_ERR(ib_conn->fmr.pool))
 		return 0;
 
 	/* no FMR => no need for page_vec */
-	kfree(ib_conn->fastreg.fmr.page_vec);
-	ib_conn->fastreg.fmr.page_vec = NULL;
+	kfree(ib_conn->fmr.page_vec);
+	ib_conn->fmr.page_vec = NULL;
 
-	ret = PTR_ERR(ib_conn->fastreg.fmr.pool);
-	ib_conn->fastreg.fmr.pool = NULL;
+	ret = PTR_ERR(ib_conn->fmr.pool);
+	ib_conn->fmr.pool = NULL;
 	if (ret != -ENOSYS) {
 		iser_err("FMR allocation failed, err %d\n", ret);
 		return ret;
@@ -268,93 +262,173 @@ int iser_create_fmr_pool(struct iser_conn *ib_conn, unsigned cmds_max)
 void iser_free_fmr_pool(struct iser_conn *ib_conn)
 {
 	iser_info("freeing conn %p fmr pool %p\n",
-		  ib_conn, ib_conn->fastreg.fmr.pool);
+		  ib_conn, ib_conn->fmr.pool);
+
+	if (ib_conn->fmr.pool != NULL)
+		ib_destroy_fmr_pool(ib_conn->fmr.pool);
+
+	ib_conn->fmr.pool = NULL;
+
+	kfree(ib_conn->fmr.page_vec);
+	ib_conn->fmr.page_vec = NULL;
+}
+
+static int
+iser_create_fastreg_desc(struct ib_device *ib_device, struct ib_pd *pd,
+			 bool pi_enable, struct fast_reg_descriptor *desc)
+{
+	int ret;
+
+	desc->data_frpl = ib_alloc_fast_reg_page_list(ib_device,
+						      ISCSI_ISER_SG_TABLESIZE + 1);
+	if (IS_ERR(desc->data_frpl)) {
+		ret = PTR_ERR(desc->data_frpl);
+		iser_err("Failed to allocate ib_fast_reg_page_list err=%d\n",
+			 ret);
+		return PTR_ERR(desc->data_frpl);
+	}
+
+	desc->data_mr = ib_alloc_fast_reg_mr(pd, ISCSI_ISER_SG_TABLESIZE + 1);
+	if (IS_ERR(desc->data_mr)) {
+		ret = PTR_ERR(desc->data_mr);
+		iser_err("Failed to allocate ib_fast_reg_mr err=%d\n", ret);
+		goto fast_reg_mr_failure;
+	}
+	desc->reg_indicators |= ISER_DATA_KEY_VALID;
+
+	if (pi_enable) {
+		struct ib_mr_init_attr mr_init_attr = {0};
+		struct iser_pi_context *pi_ctx = NULL;
 
-	if (ib_conn->fastreg.fmr.pool != NULL)
-		ib_destroy_fmr_pool(ib_conn->fastreg.fmr.pool);
+		desc->pi_ctx = kzalloc(sizeof(*desc->pi_ctx), GFP_KERNEL);
+		if (!desc->pi_ctx) {
+			iser_err("Failed to allocate pi context\n");
+			ret = -ENOMEM;
+			goto pi_ctx_alloc_failure;
+		}
+		pi_ctx = desc->pi_ctx;
+
+		pi_ctx->prot_frpl = ib_alloc_fast_reg_page_list(ib_device,
+						    ISCSI_ISER_SG_TABLESIZE);
+		if (IS_ERR(pi_ctx->prot_frpl)) {
+			ret = PTR_ERR(pi_ctx->prot_frpl);
+			iser_err("Failed to allocate prot frpl ret=%d\n",
+				 ret);
+			goto prot_frpl_failure;
+		}
+
+		pi_ctx->prot_mr = ib_alloc_fast_reg_mr(pd,
+						ISCSI_ISER_SG_TABLESIZE + 1);
+		if (IS_ERR(pi_ctx->prot_mr)) {
+			ret = PTR_ERR(pi_ctx->prot_mr);
+			iser_err("Failed to allocate prot frmr ret=%d\n",
+				 ret);
+			goto prot_mr_failure;
+		}
+		desc->reg_indicators |= ISER_PROT_KEY_VALID;
+
+		mr_init_attr.max_reg_descriptors = 2;
+		mr_init_attr.flags |= IB_MR_SIGNATURE_EN;
+		pi_ctx->sig_mr = ib_create_mr(pd, &mr_init_attr);
+		if (IS_ERR(pi_ctx->sig_mr)) {
+			ret = PTR_ERR(pi_ctx->sig_mr);
+			iser_err("Failed to allocate signature enabled mr err=%d\n",
+				 ret);
+			goto sig_mr_failure;
+		}
+		desc->reg_indicators |= ISER_SIG_KEY_VALID;
+	}
+	desc->reg_indicators &= ~ISER_FASTREG_PROTECTED;
 
-	ib_conn->fastreg.fmr.pool = NULL;
+	iser_dbg("Create fr_desc %p page_list %p\n",
+		 desc, desc->data_frpl->page_list);
 
-	kfree(ib_conn->fastreg.fmr.page_vec);
-	ib_conn->fastreg.fmr.page_vec = NULL;
+	return 0;
+sig_mr_failure:
+	ib_dereg_mr(desc->pi_ctx->prot_mr);
+prot_mr_failure:
+	ib_free_fast_reg_page_list(desc->pi_ctx->prot_frpl);
+prot_frpl_failure:
+	kfree(desc->pi_ctx);
+pi_ctx_alloc_failure:
+	ib_dereg_mr(desc->data_mr);
+fast_reg_mr_failure:
+	ib_free_fast_reg_page_list(desc->data_frpl);
+
+	return ret;
 }
 
 /**
- * iser_create_frwr_pool - Creates pool of fast_reg descriptors
+ * iser_create_fastreg_pool - Creates pool of fast_reg descriptors
  * for fast registration work requests.
  * returns 0 on success, or errno code on failure
  */
-int iser_create_frwr_pool(struct iser_conn *ib_conn, unsigned cmds_max)
+int iser_create_fastreg_pool(struct iser_conn *ib_conn, unsigned cmds_max)
 {
 	struct iser_device	*device = ib_conn->device;
 	struct fast_reg_descriptor	*desc;
 	int i, ret;
 
-	INIT_LIST_HEAD(&ib_conn->fastreg.frwr.pool);
-	ib_conn->fastreg.frwr.pool_size = 0;
+	INIT_LIST_HEAD(&ib_conn->fastreg.pool);
+	ib_conn->fastreg.pool_size = 0;
 	for (i = 0; i < cmds_max; i++) {
-		desc = kmalloc(sizeof(*desc), GFP_KERNEL);
+		desc = kzalloc(sizeof(*desc), GFP_KERNEL);
 		if (!desc) {
 			iser_err("Failed to allocate a new fast_reg descriptor\n");
 			ret = -ENOMEM;
 			goto err;
 		}
 
-		desc->data_frpl = ib_alloc_fast_reg_page_list(device->ib_device,
-							 ISCSI_ISER_SG_TABLESIZE + 1);
-		if (IS_ERR(desc->data_frpl)) {
-			ret = PTR_ERR(desc->data_frpl);
-			iser_err("Failed to allocate ib_fast_reg_page_list err=%d\n", ret);
-			goto fast_reg_page_failure;
+		ret = iser_create_fastreg_desc(device->ib_device, device->pd,
+					       ib_conn->pi_support, desc);
+		if (ret) {
+			iser_err("Failed to create fastreg descriptor err=%d\n",
+				 ret);
+			kfree(desc);
+			goto err;
 		}
 
-		desc->data_mr = ib_alloc_fast_reg_mr(device->pd,
-						     ISCSI_ISER_SG_TABLESIZE + 1);
-		if (IS_ERR(desc->data_mr)) {
-			ret = PTR_ERR(desc->data_mr);
-			iser_err("Failed to allocate ib_fast_reg_mr err=%d\n", ret);
-			goto fast_reg_mr_failure;
-		}
-		desc->valid = true;
-		list_add_tail(&desc->list, &ib_conn->fastreg.frwr.pool);
-		ib_conn->fastreg.frwr.pool_size++;
+		list_add_tail(&desc->list, &ib_conn->fastreg.pool);
+		ib_conn->fastreg.pool_size++;
 	}
 
 	return 0;
 
-fast_reg_mr_failure:
-	ib_free_fast_reg_page_list(desc->data_frpl);
-fast_reg_page_failure:
-	kfree(desc);
 err:
-	iser_free_frwr_pool(ib_conn);
+	iser_free_fastreg_pool(ib_conn);
 	return ret;
 }
 
 /**
- * iser_free_frwr_pool - releases the pool of fast_reg descriptors
+ * iser_free_fastreg_pool - releases the pool of fast_reg descriptors
  */
-void iser_free_frwr_pool(struct iser_conn *ib_conn)
+void iser_free_fastreg_pool(struct iser_conn *ib_conn)
 {
 	struct fast_reg_descriptor *desc, *tmp;
 	int i = 0;
 
-	if (list_empty(&ib_conn->fastreg.frwr.pool))
+	if (list_empty(&ib_conn->fastreg.pool))
 		return;
 
-	iser_info("freeing conn %p frwr pool\n", ib_conn);
+	iser_info("freeing conn %p fr pool\n", ib_conn);
 
-	list_for_each_entry_safe(desc, tmp, &ib_conn->fastreg.frwr.pool, list) {
+	list_for_each_entry_safe(desc, tmp, &ib_conn->fastreg.pool, list) {
 		list_del(&desc->list);
 		ib_free_fast_reg_page_list(desc->data_frpl);
 		ib_dereg_mr(desc->data_mr);
+		if (desc->pi_ctx) {
+			ib_free_fast_reg_page_list(desc->pi_ctx->prot_frpl);
+			ib_dereg_mr(desc->pi_ctx->prot_mr);
+			ib_destroy_mr(desc->pi_ctx->sig_mr);
+			kfree(desc->pi_ctx);
+		}
 		kfree(desc);
 		++i;
 	}
 
-	if (i < ib_conn->fastreg.frwr.pool_size)
+	if (i < ib_conn->fastreg.pool_size)
 		iser_warn("pool still has %d regions registered\n",
-			  ib_conn->fastreg.frwr.pool_size - i);
+			  ib_conn->fastreg.pool_size - i);
 }
 
 /**
@@ -389,12 +463,17 @@ static int iser_create_ib_conn_res(struct iser_conn *ib_conn)
 	init_attr.qp_context	= (void *)ib_conn;
 	init_attr.send_cq	= device->tx_cq[min_index];
 	init_attr.recv_cq	= device->rx_cq[min_index];
-	init_attr.cap.max_send_wr  = ISER_QP_MAX_REQ_DTOS;
 	init_attr.cap.max_recv_wr  = ISER_QP_MAX_RECV_DTOS;
 	init_attr.cap.max_send_sge = 2;
 	init_attr.cap.max_recv_sge = 1;
 	init_attr.sq_sig_type	= IB_SIGNAL_REQ_WR;
 	init_attr.qp_type	= IB_QPT_RC;
+	if (ib_conn->pi_support) {
+		init_attr.cap.max_send_wr = ISER_QP_SIG_MAX_REQ_DTOS;
+		init_attr.create_flags |= IB_QP_CREATE_SIGNATURE_EN;
+	} else {
+		init_attr.cap.max_send_wr  = ISER_QP_MAX_REQ_DTOS;
+	}
 
 	ret = rdma_create_qp(ib_conn->cma_id, device->pd, &init_attr);
 	if (ret)
@@ -591,6 +670,19 @@ static int iser_addr_handler(struct rdma_cm_id *cma_id)
 	ib_conn = (struct iser_conn *)cma_id->context;
 	ib_conn->device = device;
 
+	/* connection T10-PI support */
+	if (iser_pi_enable) {
+		if (!(device->dev_attr.device_cap_flags &
+		      IB_DEVICE_SIGNATURE_HANDOVER)) {
+			iser_warn("T10-PI requested but not supported on %s, "
+				  "continue without T10-PI\n",
+				  ib_conn->device->ib_device->name);
+			ib_conn->pi_support = false;
+		} else {
+			ib_conn->pi_support = true;
+		}
+	}
+
 	ret = rdma_resolve_route(cma_id, 1000);
 	if (ret) {
 		iser_err("resolve route failed: %d\n", ret);
@@ -636,6 +728,11 @@ failure:
 static void iser_connected_handler(struct rdma_cm_id *cma_id)
 {
 	struct iser_conn *ib_conn;
+	struct ib_qp_attr attr;
+	struct ib_qp_init_attr init_attr;
+
+	(void)ib_query_qp(cma_id->qp, &attr, ~0, &init_attr);
+	iser_info("remote qpn:%x my qpn:%x\n", attr.dest_qp_num, cma_id->qp->qp_num);
 
 	ib_conn = (struct iser_conn *)cma_id->context;
 	ib_conn->state = ISER_CONN_UP;
@@ -653,9 +750,8 @@ static int iser_disconnected_handler(struct rdma_cm_id *cma_id)
 	 * terminated asynchronously from the iSCSI layer's perspective.  */
 	if (iser_conn_state_comp_exch(ib_conn, ISER_CONN_UP,
 					ISER_CONN_TERMINATING)){
-		if (ib_conn->iser_conn)
-			iscsi_conn_failure(ib_conn->iser_conn->iscsi_conn,
-					   ISCSI_ERR_CONN_FAILED);
+		if (ib_conn->iscsi_conn)
+			iscsi_conn_failure(ib_conn->iscsi_conn, ISCSI_ERR_CONN_FAILED);
 		else
 			iser_err("iscsi_iser connection isn't bound\n");
 	}
@@ -801,7 +897,7 @@ int iser_reg_page_vec(struct iser_conn     *ib_conn,
 	page_list = page_vec->pages;
 	io_addr	  = page_list[0];
 
-	mem  = ib_fmr_pool_map_phys(ib_conn->fastreg.fmr.pool,
+	mem  = ib_fmr_pool_map_phys(ib_conn->fmr.pool,
 				    page_list,
 				    page_vec->length,
 				    io_addr);
@@ -855,11 +951,11 @@ void iser_unreg_mem_fmr(struct iscsi_iser_task *iser_task,
 	reg->mem_h = NULL;
 }
 
-void iser_unreg_mem_frwr(struct iscsi_iser_task *iser_task,
-			 enum iser_data_dir cmd_dir)
+void iser_unreg_mem_fastreg(struct iscsi_iser_task *iser_task,
+			    enum iser_data_dir cmd_dir)
 {
 	struct iser_mem_reg *reg = &iser_task->rdma_regd[cmd_dir].reg;
-	struct iser_conn *ib_conn = iser_task->iser_conn->ib_conn;
+	struct iser_conn *ib_conn = iser_task->ib_conn;
 	struct fast_reg_descriptor *desc = reg->mem_h;
 
 	if (!reg->is_mr)
@@ -868,7 +964,7 @@ void iser_unreg_mem_frwr(struct iscsi_iser_task *iser_task,
 	reg->mem_h = NULL;
 	reg->is_mr = 0;
 	spin_lock_bh(&ib_conn->lock);
-	list_add_tail(&desc->list, &ib_conn->fastreg.frwr.pool);
+	list_add_tail(&desc->list, &ib_conn->fastreg.pool);
 	spin_unlock_bh(&ib_conn->lock);
 }
 
@@ -969,7 +1065,7 @@ static void iser_handle_comp_error(struct iser_tx_desc *desc,
 		 * perspective.                                             */
 		if (iser_conn_state_comp_exch(ib_conn, ISER_CONN_UP,
 		    ISER_CONN_TERMINATING))
-			iscsi_conn_failure(ib_conn->iser_conn->iscsi_conn,
+			iscsi_conn_failure(ib_conn->iscsi_conn,
 					   ISCSI_ERR_CONN_FAILED);
 
 		/* no more non completed posts to the QP, complete the
@@ -993,18 +1089,16 @@ static int iser_drain_tx_cq(struct iser_device  *device, int cq_index)
 		if (wc.status == IB_WC_SUCCESS) {
 			if (wc.opcode == IB_WC_SEND)
 				iser_snd_completion(tx_desc, ib_conn);
-			else if (wc.opcode == IB_WC_LOCAL_INV ||
-				 wc.opcode == IB_WC_FAST_REG_MR) {
-				atomic_dec(&ib_conn->post_send_buf_count);
-				continue;
-			} else
+			else
 				iser_err("expected opcode %d got %d\n",
 					IB_WC_SEND, wc.opcode);
 		} else {
 			iser_err("tx id %llx status %d vend_err %x\n",
-				wc.wr_id, wc.status, wc.vendor_err);
-			atomic_dec(&ib_conn->post_send_buf_count);
-			iser_handle_comp_error(tx_desc, ib_conn);
+				 wc.wr_id, wc.status, wc.vendor_err);
+			if (wc.wr_id != ISER_FASTREG_LI_WRID) {
+				atomic_dec(&ib_conn->post_send_buf_count);
+				iser_handle_comp_error(tx_desc, ib_conn);
+			}
 		}
 		completed_tx++;
 	}
@@ -1022,8 +1116,12 @@ static void iser_cq_tasklet_fn(unsigned long data)
 	 struct iser_rx_desc *desc;
 	 unsigned long	     xfer_len;
 	struct iser_conn *ib_conn;
-	int completed_tx, completed_rx;
-	completed_tx = completed_rx = 0;
+	int completed_tx, completed_rx = 0;
+
+	/* First do tx drain, so in a case where we have rx flushes and a successful
+	 * tx completion we will still go through completion error handling.
+	 */
+	completed_tx = iser_drain_tx_cq(device, cq_index);
 
 	while (ib_poll_cq(cq, 1, &wc) == 1) {
 		desc	 = (struct iser_rx_desc *) (unsigned long) wc.wr_id;
@@ -1051,7 +1149,6 @@ static void iser_cq_tasklet_fn(unsigned long data)
 	 * " would not cause interrupts to be missed"                       */
 	ib_req_notify_cq(cq, IB_CQ_NEXT_COMP);
 
-	completed_tx += iser_drain_tx_cq(device, cq_index);
 	iser_dbg("got %d rx %d tx completions\n", completed_rx, completed_tx);
 }
 
@@ -1063,3 +1160,51 @@ static void iser_cq_callback(struct ib_cq *cq, void *cq_context)
 
 	tasklet_schedule(&device->cq_tasklet[cq_index]);
 }
+
+u8 iser_check_task_pi_status(struct iscsi_iser_task *iser_task,
+			     enum iser_data_dir cmd_dir, sector_t *sector)
+{
+	struct iser_mem_reg *reg = &iser_task->rdma_regd[cmd_dir].reg;
+	struct fast_reg_descriptor *desc = reg->mem_h;
+	unsigned long sector_size = iser_task->sc->device->sector_size;
+	struct ib_mr_status mr_status;
+	int ret;
+
+	if (desc && desc->reg_indicators & ISER_FASTREG_PROTECTED) {
+		desc->reg_indicators &= ~ISER_FASTREG_PROTECTED;
+		ret = ib_check_mr_status(desc->pi_ctx->sig_mr,
+					 IB_MR_CHECK_SIG_STATUS, &mr_status);
+		if (ret) {
+			pr_err("ib_check_mr_status failed, ret %d\n", ret);
+			goto err;
+		}
+
+		if (mr_status.fail_status & IB_MR_CHECK_SIG_STATUS) {
+			sector_t sector_off = mr_status.sig_err.sig_err_offset;
+
+			do_div(sector_off, sector_size + 8);
+			*sector = scsi_get_lba(iser_task->sc) + sector_off;
+
+			pr_err("PI error found type %d at sector %llx "
+			       "expected %x vs actual %x\n",
+			       mr_status.sig_err.err_type,
+			       (unsigned long long)*sector,
+			       mr_status.sig_err.expected,
+			       mr_status.sig_err.actual);
+
+			switch (mr_status.sig_err.err_type) {
+			case IB_SIG_BAD_GUARD:
+				return 0x1;
+			case IB_SIG_BAD_REFTAG:
+				return 0x3;
+			case IB_SIG_BAD_APPTAG:
+				return 0x2;
+			}
+		}
+	}
+
+	return 0;
+err:
+	/* Not alot we can do here, return ambiguous guard error */
+	return 0x1;
+}

drivers/infiniband/ulp/srp/ib_srp.c

@@ -411,6 +411,8 @@ static void srp_path_rec_completion(int status,
 
 static int srp_lookup_path(struct srp_target_port *target)
 {
+	int ret;
+
 	target->path.numb_path = 1;
 
 	init_completion(&target->done);
@@ -431,7 +433,9 @@ static int srp_lookup_path(struct srp_target_port *target)
 	if (target->path_query_id < 0)
 		return target->path_query_id;
 
-	wait_for_completion(&target->done);
+	ret = wait_for_completion_interruptible(&target->done);
+	if (ret < 0)
+		return ret;
 
 	if (target->status < 0)
 		shost_printk(KERN_WARNING, target->scsi_host,
@@ -710,7 +714,9 @@ static int srp_connect_target(struct srp_target_port *target)
 		ret = srp_send_req(target);
 		if (ret)
 			return ret;
-		wait_for_completion(&target->done);
+		ret = wait_for_completion_interruptible(&target->done);
+		if (ret < 0)
+			return ret;
 
 		/*
 		 * The CM event handling code will set status to
@@ -777,6 +783,7 @@ static void srp_unmap_data(struct scsi_cmnd *scmnd,
  * srp_claim_req - Take ownership of the scmnd associated with a request.
  * @target: SRP target port.
  * @req: SRP request.
+ * @sdev: If not NULL, only take ownership for this SCSI device.
  * @scmnd: If NULL, take ownership of @req->scmnd. If not NULL, only take
  *         ownership of @req->scmnd if it equals @scmnd.
  *
@@ -785,16 +792,17 @@ static void srp_unmap_data(struct scsi_cmnd *scmnd,
  */
 static struct scsi_cmnd *srp_claim_req(struct srp_target_port *target,
 				       struct srp_request *req,
+				       struct scsi_device *sdev,
 				       struct scsi_cmnd *scmnd)
 {
 	unsigned long flags;
 
 	spin_lock_irqsave(&target->lock, flags);
-	if (!scmnd) {
+	if (req->scmnd &&
+	    (!sdev || req->scmnd->device == sdev) &&
+	    (!scmnd || req->scmnd == scmnd)) {
 		scmnd = req->scmnd;
 		req->scmnd = NULL;
-	} else if (req->scmnd == scmnd) {
-		req->scmnd = NULL;
 	} else {
 		scmnd = NULL;
 	}
@@ -821,9 +829,10 @@ static void srp_free_req(struct srp_target_port *target,
 }
 
 static void srp_finish_req(struct srp_target_port *target,
-			   struct srp_request *req, int result)
+			   struct srp_request *req, struct scsi_device *sdev,
+			   int result)
 {
-	struct scsi_cmnd *scmnd = srp_claim_req(target, req, NULL);
+	struct scsi_cmnd *scmnd = srp_claim_req(target, req, sdev, NULL);
 
 	if (scmnd) {
 		srp_free_req(target, req, scmnd, 0);
@@ -835,11 +844,20 @@ static void srp_finish_req(struct srp_target_port *target,
 static void srp_terminate_io(struct srp_rport *rport)
 {
 	struct srp_target_port *target = rport->lld_data;
+	struct Scsi_Host *shost = target->scsi_host;
+	struct scsi_device *sdev;
 	int i;
 
+	/*
+	 * Invoking srp_terminate_io() while srp_queuecommand() is running
+	 * is not safe. Hence the warning statement below.
+	 */
+	shost_for_each_device(sdev, shost)
+		WARN_ON_ONCE(sdev->request_queue->request_fn_active);
+
 	for (i = 0; i < target->req_ring_size; ++i) {
 		struct srp_request *req = &target->req_ring[i];
-		srp_finish_req(target, req, DID_TRANSPORT_FAILFAST << 16);
+		srp_finish_req(target, req, NULL, DID_TRANSPORT_FAILFAST << 16);
 	}
 }
 
@@ -876,7 +894,7 @@ static int srp_rport_reconnect(struct srp_rport *rport)
 
 	for (i = 0; i < target->req_ring_size; ++i) {
 		struct srp_request *req = &target->req_ring[i];
-		srp_finish_req(target, req, DID_RESET << 16);
+		srp_finish_req(target, req, NULL, DID_RESET << 16);
 	}
 
 	INIT_LIST_HEAD(&target->free_tx);
@@ -1284,7 +1302,7 @@ static void srp_process_rsp(struct srp_target_port *target, struct srp_rsp *rsp)
 		complete(&target->tsk_mgmt_done);
 	} else {
 		req = &target->req_ring[rsp->tag];
-		scmnd = srp_claim_req(target, req, NULL);
+		scmnd = srp_claim_req(target, req, NULL, NULL);
 		if (!scmnd) {
 			shost_printk(KERN_ERR, target->scsi_host,
 				     "Null scmnd for RSP w/tag %016llx\n",
@@ -1804,8 +1822,10 @@ static void srp_cm_rej_handler(struct ib_cm_id *cm_id,
 				shost_printk(KERN_WARNING, shost,
 					     PFX "SRP_LOGIN_REJ: requested max_it_iu_len too large\n");
 			else
-				shost_printk(KERN_WARNING, shost,
-					    PFX "SRP LOGIN REJECTED, reason 0x%08x\n", reason);
+				shost_printk(KERN_WARNING, shost, PFX
+					     "SRP LOGIN from %pI6 to %pI6 REJECTED, reason 0x%08x\n",
+					     target->path.sgid.raw,
+					     target->orig_dgid, reason);
 		} else
 			shost_printk(KERN_WARNING, shost,
 				     "  REJ reason: IB_CM_REJ_CONSUMER_DEFINED,"
@@ -1863,6 +1883,7 @@ static int srp_cm_handler(struct ib_cm_id *cm_id, struct ib_cm_event *event)
 	case IB_CM_TIMEWAIT_EXIT:
 		shost_printk(KERN_ERR, target->scsi_host,
 			     PFX "connection closed\n");
+		comp = 1;
 
 		target->status = 0;
 		break;
@@ -1999,7 +2020,7 @@ static int srp_abort(struct scsi_cmnd *scmnd)
 
 	shost_printk(KERN_ERR, target->scsi_host, "SRP abort called\n");
 
-	if (!req || !srp_claim_req(target, req, scmnd))
+	if (!req || !srp_claim_req(target, req, NULL, scmnd))
 		return SUCCESS;
 	if (srp_send_tsk_mgmt(target, req->index, scmnd->device->lun,
 			      SRP_TSK_ABORT_TASK) == 0)
@@ -2030,8 +2051,7 @@ static int srp_reset_device(struct scsi_cmnd *scmnd)
 
 	for (i = 0; i < target->req_ring_size; ++i) {
 		struct srp_request *req = &target->req_ring[i];
-		if (req->scmnd && req->scmnd->device == scmnd->device)
-			srp_finish_req(target, req, DID_RESET << 16);
+		srp_finish_req(target, req, scmnd->device, DID_RESET << 16);
 	}
 
 	return SUCCESS;
@@ -2612,6 +2632,8 @@ static ssize_t srp_create_target(struct device *dev,
 	target->tl_retry_count	= 7;
 	target->queue_size	= SRP_DEFAULT_QUEUE_SIZE;
 
+	mutex_lock(&host->add_target_mutex);
+
 	ret = srp_parse_options(buf, target);
 	if (ret)
 		goto err;
@@ -2649,16 +2671,9 @@ static ssize_t srp_create_target(struct device *dev,
 	if (ret)
 		goto err_free_mem;
 
-	ib_query_gid(ibdev, host->port, 0, &target->path.sgid);
-
-	shost_printk(KERN_DEBUG, target->scsi_host, PFX
-		     "new target: id_ext %016llx ioc_guid %016llx pkey %04x "
-		     "service_id %016llx dgid %pI6\n",
-	       (unsigned long long) be64_to_cpu(target->id_ext),
-	       (unsigned long long) be64_to_cpu(target->ioc_guid),
-	       be16_to_cpu(target->path.pkey),
-	       (unsigned long long) be64_to_cpu(target->service_id),
-	       target->path.dgid.raw);
+	ret = ib_query_gid(ibdev, host->port, 0, &target->path.sgid);
+	if (ret)
+		goto err_free_mem;
 
 	ret = srp_create_target_ib(target);
 	if (ret)
@@ -2679,7 +2694,19 @@ static ssize_t srp_create_target(struct device *dev,
 	if (ret)
 		goto err_disconnect;
 
-	return count;
+	shost_printk(KERN_DEBUG, target->scsi_host, PFX
+		     "new target: id_ext %016llx ioc_guid %016llx pkey %04x service_id %016llx sgid %pI6 dgid %pI6\n",
+		     be64_to_cpu(target->id_ext),
+		     be64_to_cpu(target->ioc_guid),
+		     be16_to_cpu(target->path.pkey),
+		     be64_to_cpu(target->service_id),
+		     target->path.sgid.raw, target->path.dgid.raw);
+
+	ret = count;
+
+out:
+	mutex_unlock(&host->add_target_mutex);
+	return ret;
 
 err_disconnect:
 	srp_disconnect_target(target);
@@ -2695,8 +2722,7 @@ err_free_mem:
 
 err:
 	scsi_host_put(target_host);
-
-	return ret;
+	goto out;
 }
 
 static DEVICE_ATTR(add_target, S_IWUSR, NULL, srp_create_target);
@@ -2732,6 +2758,7 @@ static struct srp_host *srp_add_port(struct srp_device *device, u8 port)
 	INIT_LIST_HEAD(&host->target_list);
 	spin_lock_init(&host->target_lock);
 	init_completion(&host->released);
+	mutex_init(&host->add_target_mutex);
 	host->srp_dev = device;
 	host->port = port;
 

drivers/infiniband/ulp/srp/ib_srp.h

@@ -105,6 +105,7 @@ struct srp_host {
 	spinlock_t		target_lock;
 	struct completion	released;
 	struct list_head	list;
+	struct mutex		add_target_mutex;
 };
 
 struct srp_request {

drivers/net/ethernet/emulex/benet/be_roce.c

@@ -35,6 +35,12 @@ static void _be_roce_dev_add(struct be_adapter *adapter)
 
 	if (!ocrdma_drv)
 		return;
+
+	if (ocrdma_drv->be_abi_version != BE_ROCE_ABI_VERSION) {
+		dev_warn(&pdev->dev, "Cannot initialize RoCE due to ocrdma ABI mismatch\n");
+		return;
+	}
+
 	if (pdev->device == OC_DEVICE_ID5) {
 		/* only msix is supported on these devices */
 		if (!msix_enabled(adapter))

drivers/net/ethernet/emulex/benet/be_roce.h

@@ -21,6 +21,8 @@
 #include <linux/pci.h>
 #include <linux/netdevice.h>
 
+#define BE_ROCE_ABI_VERSION	1
+
 struct ocrdma_dev;
 
 enum be_interrupt_mode {
@@ -52,6 +54,7 @@ struct be_dev_info {
 /* ocrdma driver register's the callback functions with nic driver. */
 struct ocrdma_driver {
 	unsigned char name[32];
+	u32 be_abi_version;
 	struct ocrdma_dev *(*add) (struct be_dev_info *dev_info);
 	void (*remove) (struct ocrdma_dev *);
 	void (*state_change_handler) (struct ocrdma_dev *, u32 new_state);

drivers/net/ethernet/mellanox/mlx5/core/main.c

@@ -440,6 +440,7 @@ int mlx5_dev_init(struct mlx5_core_dev *dev, struct pci_dev *pdev)
 	mlx5_init_cq_table(dev);
 	mlx5_init_qp_table(dev);
 	mlx5_init_srq_table(dev);
+	mlx5_init_mr_table(dev);
 
 	return 0;
 

drivers/net/ethernet/mellanox/mlx5/core/mr.c

@@ -36,11 +36,24 @@
 #include <linux/mlx5/cmd.h>
 #include "mlx5_core.h"
 
+void mlx5_init_mr_table(struct mlx5_core_dev *dev)
+{
+	struct mlx5_mr_table *table = &dev->priv.mr_table;
+
+	rwlock_init(&table->lock);
+	INIT_RADIX_TREE(&table->tree, GFP_ATOMIC);
+}
+
+void mlx5_cleanup_mr_table(struct mlx5_core_dev *dev)
+{
+}
+
 int mlx5_core_create_mkey(struct mlx5_core_dev *dev, struct mlx5_core_mr *mr,
 			  struct mlx5_create_mkey_mbox_in *in, int inlen,
 			  mlx5_cmd_cbk_t callback, void *context,
 			  struct mlx5_create_mkey_mbox_out *out)
 {
+	struct mlx5_mr_table *table = &dev->priv.mr_table;
 	struct mlx5_create_mkey_mbox_out lout;
 	int err;
 	u8 key;
@@ -73,14 +86,21 @@ int mlx5_core_create_mkey(struct mlx5_core_dev *dev, struct mlx5_core_mr *mr,
 	mlx5_core_dbg(dev, "out 0x%x, key 0x%x, mkey 0x%x\n",
 		      be32_to_cpu(lout.mkey), key, mr->key);
 
+	/* connect to MR tree */
+	write_lock_irq(&table->lock);
+	err = radix_tree_insert(&table->tree, mlx5_base_mkey(mr->key), mr);
+	write_unlock_irq(&table->lock);
+
 	return err;
 }
 EXPORT_SYMBOL(mlx5_core_create_mkey);
 
 int mlx5_core_destroy_mkey(struct mlx5_core_dev *dev, struct mlx5_core_mr *mr)
 {
+	struct mlx5_mr_table *table = &dev->priv.mr_table;
 	struct mlx5_destroy_mkey_mbox_in in;
 	struct mlx5_destroy_mkey_mbox_out out;
+	unsigned long flags;
 	int err;
 
 	memset(&in, 0, sizeof(in));
@@ -95,6 +115,10 @@ int mlx5_core_destroy_mkey(struct mlx5_core_dev *dev, struct mlx5_core_mr *mr)
 	if (out.hdr.status)
 		return mlx5_cmd_status_to_err(&out.hdr);
 
+	write_lock_irqsave(&table->lock, flags);
+	radix_tree_delete(&table->tree, mlx5_base_mkey(mr->key));
+	write_unlock_irqrestore(&table->lock, flags);
+
 	return err;
 }
 EXPORT_SYMBOL(mlx5_core_destroy_mkey);
@@ -144,3 +168,64 @@ int mlx5_core_dump_fill_mkey(struct mlx5_core_dev *dev, struct mlx5_core_mr *mr,
 	return err;
 }
 EXPORT_SYMBOL(mlx5_core_dump_fill_mkey);
+
+int mlx5_core_create_psv(struct mlx5_core_dev *dev, u32 pdn,
+			 int npsvs, u32 *sig_index)
+{
+	struct mlx5_allocate_psv_in in;
+	struct mlx5_allocate_psv_out out;
+	int i, err;
+
+	if (npsvs > MLX5_MAX_PSVS)
+		return -EINVAL;
+
+	memset(&in, 0, sizeof(in));
+	memset(&out, 0, sizeof(out));
+
+	in.hdr.opcode = cpu_to_be16(MLX5_CMD_OP_CREATE_PSV);
+	in.npsv_pd = cpu_to_be32((npsvs << 28) | pdn);
+	err = mlx5_cmd_exec(dev, &in, sizeof(in), &out, sizeof(out));
+	if (err) {
+		mlx5_core_err(dev, "cmd exec failed %d\n", err);
+		return err;
+	}
+
+	if (out.hdr.status) {
+		mlx5_core_err(dev, "create_psv bad status %d\n", out.hdr.status);
+		return mlx5_cmd_status_to_err(&out.hdr);
+	}
+
+	for (i = 0; i < npsvs; i++)
+		sig_index[i] = be32_to_cpu(out.psv_idx[i]) & 0xffffff;
+
+	return err;
+}
+EXPORT_SYMBOL(mlx5_core_create_psv);
+
+int mlx5_core_destroy_psv(struct mlx5_core_dev *dev, int psv_num)
+{
+	struct mlx5_destroy_psv_in in;
+	struct mlx5_destroy_psv_out out;
+	int err;
+
+	memset(&in, 0, sizeof(in));
+	memset(&out, 0, sizeof(out));
+
+	in.psv_number = cpu_to_be32(psv_num);
+	in.hdr.opcode = cpu_to_be16(MLX5_CMD_OP_DESTROY_PSV);
+	err = mlx5_cmd_exec(dev, &in, sizeof(in), &out, sizeof(out));
+	if (err) {
+		mlx5_core_err(dev, "destroy_psv cmd exec failed %d\n", err);
+		goto out;
+	}
+
+	if (out.hdr.status) {
+		mlx5_core_err(dev, "destroy_psv bad status %d\n", out.hdr.status);
+		err = mlx5_cmd_status_to_err(&out.hdr);
+		goto out;
+	}
+
+out:
+	return err;
+}
+EXPORT_SYMBOL(mlx5_core_destroy_psv);

drivers/scsi/libiscsi.c

@@ -387,6 +387,10 @@ static int iscsi_prep_scsi_cmd_pdu(struct iscsi_task *task)
 		if (rc)
 			return rc;
 	}
+
+	if (scsi_get_prot_op(sc) != SCSI_PROT_NORMAL)
+		task->protected = true;
+
 	if (sc->sc_data_direction == DMA_TO_DEVICE) {
 		unsigned out_len = scsi_out(sc)->length;
 		struct iscsi_r2t_info *r2t = &task->unsol_r2t;
@@ -822,6 +826,33 @@ static void iscsi_scsi_cmd_rsp(struct iscsi_conn *conn, struct iscsi_hdr *hdr,
 
 	sc->result = (DID_OK << 16) | rhdr->cmd_status;
 
+	if (task->protected) {
+		sector_t sector;
+		u8 ascq;
+
+		/**
+		 * Transports that didn't implement check_protection
+		 * callback but still published T10-PI support to scsi-mid
+		 * deserve this BUG_ON.
+		 **/
+		BUG_ON(!session->tt->check_protection);
+
+		ascq = session->tt->check_protection(task, &sector);
+		if (ascq) {
+			sc->result = DRIVER_SENSE << 24 |
+				     SAM_STAT_CHECK_CONDITION;
+			scsi_build_sense_buffer(1, sc->sense_buffer,
+						ILLEGAL_REQUEST, 0x10, ascq);
+			sc->sense_buffer[7] = 0xc; /* Additional sense length */
+			sc->sense_buffer[8] = 0;   /* Information desc type */
+			sc->sense_buffer[9] = 0xa; /* Additional desc length */
+			sc->sense_buffer[10] = 0x80; /* Validity bit */
+
+			put_unaligned_be64(sector, &sc->sense_buffer[12]);
+			goto out;
+		}
+	}
+
 	if (rhdr->response != ISCSI_STATUS_CMD_COMPLETED) {
 		sc->result = DID_ERROR << 16;
 		goto out;
@@ -1582,6 +1613,7 @@ static inline struct iscsi_task *iscsi_alloc_task(struct iscsi_conn *conn,
 	task->have_checked_conn = false;
 	task->last_timeout = jiffies;
 	task->last_xfer = jiffies;
+	task->protected = false;
 	INIT_LIST_HEAD(&task->running);
 	return task;
 }

drivers/scsi/scsi_transport_srp.c

@@ -810,6 +810,7 @@ EXPORT_SYMBOL_GPL(srp_remove_host);
 
 /**
  * srp_stop_rport_timers - stop the transport layer recovery timers
+ * @rport: SRP remote port for which to stop the timers.
  *
  * Must be called after srp_remove_host() and scsi_remove_host(). The caller
  * must hold a reference on the rport (rport->dev) and on the SCSI host

include/linux/mlx5/cq.h

@@ -80,6 +80,7 @@ enum {
 	MLX5_CQE_RESP_SEND_IMM	= 3,
 	MLX5_CQE_RESP_SEND_INV	= 4,
 	MLX5_CQE_RESIZE_CQ	= 5,
+	MLX5_CQE_SIG_ERR	= 12,
 	MLX5_CQE_REQ_ERR	= 13,
 	MLX5_CQE_RESP_ERR	= 14,
 	MLX5_CQE_INVALID	= 15,

include/linux/mlx5/device.h

@@ -48,6 +48,8 @@ enum {
 	MLX5_MAX_COMMANDS		= 32,
 	MLX5_CMD_DATA_BLOCK_SIZE	= 512,
 	MLX5_PCI_CMD_XPORT		= 7,
+	MLX5_MKEY_BSF_OCTO_SIZE		= 4,
+	MLX5_MAX_PSVS			= 4,
 };
 
 enum {
@@ -116,6 +118,7 @@ enum {
 	MLX5_MKEY_MASK_START_ADDR	= 1ull << 6,
 	MLX5_MKEY_MASK_PD		= 1ull << 7,
 	MLX5_MKEY_MASK_EN_RINVAL	= 1ull << 8,
+	MLX5_MKEY_MASK_EN_SIGERR	= 1ull << 9,
 	MLX5_MKEY_MASK_BSF_EN		= 1ull << 12,
 	MLX5_MKEY_MASK_KEY		= 1ull << 13,
 	MLX5_MKEY_MASK_QPN		= 1ull << 14,
@@ -555,6 +558,23 @@ struct mlx5_cqe64 {
 	u8		op_own;
 };
 
+struct mlx5_sig_err_cqe {
+	u8		rsvd0[16];
+	__be32		expected_trans_sig;
+	__be32		actual_trans_sig;
+	__be32		expected_reftag;
+	__be32		actual_reftag;
+	__be16		syndrome;
+	u8		rsvd22[2];
+	__be32		mkey;
+	__be64		err_offset;
+	u8		rsvd30[8];
+	__be32		qpn;
+	u8		rsvd38[2];
+	u8		signature;
+	u8		op_own;
+};
+
 struct mlx5_wqe_srq_next_seg {
 	u8			rsvd0[2];
 	__be16			next_wqe_index;
@@ -936,4 +956,27 @@ enum {
 	MLX_EXT_PORT_CAP_FLAG_EXTENDED_PORT_INFO	= 1 <<  0
 };
 
+struct mlx5_allocate_psv_in {
+	struct mlx5_inbox_hdr   hdr;
+	__be32			npsv_pd;
+	__be32			rsvd_psv0;
+};
+
+struct mlx5_allocate_psv_out {
+	struct mlx5_outbox_hdr  hdr;
+	u8			rsvd[8];
+	__be32			psv_idx[4];
+};
+
+struct mlx5_destroy_psv_in {
+	struct mlx5_inbox_hdr	hdr;
+	__be32                  psv_number;
+	u8                      rsvd[4];
+};
+
+struct mlx5_destroy_psv_out {
+	struct mlx5_outbox_hdr  hdr;
+	u8                      rsvd[8];
+};
+
 #endif /* MLX5_DEVICE_H */

include/linux/mlx5/driver.h

@@ -401,6 +401,26 @@ struct mlx5_eq {
 	struct mlx5_rsc_debug	*dbg;
 };
 
+struct mlx5_core_psv {
+	u32	psv_idx;
+	struct psv_layout {
+		u32	pd;
+		u16	syndrome;
+		u16	reserved;
+		u16	bg;
+		u16	app_tag;
+		u32	ref_tag;
+	} psv;
+};
+
+struct mlx5_core_sig_ctx {
+	struct mlx5_core_psv	psv_memory;
+	struct mlx5_core_psv	psv_wire;
+	struct ib_sig_err       err_item;
+	bool			sig_status_checked;
+	bool			sig_err_exists;
+	u32			sigerr_count;
+};
 
 struct mlx5_core_mr {
 	u64			iova;
@@ -475,6 +495,13 @@ struct mlx5_srq_table {
 	struct radix_tree_root	tree;
 };
 
+struct mlx5_mr_table {
+	/* protect radix tree
+	 */
+	rwlock_t		lock;
+	struct radix_tree_root	tree;
+};
+
 struct mlx5_priv {
 	char			name[MLX5_MAX_NAME_LEN];
 	struct mlx5_eq_table	eq_table;
@@ -504,6 +531,10 @@ struct mlx5_priv {
 	struct mlx5_cq_table	cq_table;
 	/* end: cq staff */
 
+	/* start: mr staff */
+	struct mlx5_mr_table	mr_table;
+	/* end: mr staff */
+
 	/* start: alloc staff */
 	struct mutex            pgdir_mutex;
 	struct list_head        pgdir_list;
@@ -651,6 +682,11 @@ static inline void mlx5_vfree(const void *addr)
 		kfree(addr);
 }
 
+static inline u32 mlx5_base_mkey(const u32 key)
+{
+	return key & 0xffffff00u;
+}
+
 int mlx5_dev_init(struct mlx5_core_dev *dev, struct pci_dev *pdev);
 void mlx5_dev_cleanup(struct mlx5_core_dev *dev);
 int mlx5_cmd_init(struct mlx5_core_dev *dev);
@@ -685,6 +721,8 @@ int mlx5_core_query_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq,
 			struct mlx5_query_srq_mbox_out *out);
 int mlx5_core_arm_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq,
 		      u16 lwm, int is_srq);
+void mlx5_init_mr_table(struct mlx5_core_dev *dev);
+void mlx5_cleanup_mr_table(struct mlx5_core_dev *dev);
 int mlx5_core_create_mkey(struct mlx5_core_dev *dev, struct mlx5_core_mr *mr,
 			  struct mlx5_create_mkey_mbox_in *in, int inlen,
 			  mlx5_cmd_cbk_t callback, void *context,
@@ -746,6 +784,9 @@ void mlx5_db_free(struct mlx5_core_dev *dev, struct mlx5_db *db);
 const char *mlx5_command_str(int command);
 int mlx5_cmdif_debugfs_init(struct mlx5_core_dev *dev);
 void mlx5_cmdif_debugfs_cleanup(struct mlx5_core_dev *dev);
+int mlx5_core_create_psv(struct mlx5_core_dev *dev, u32 pdn,
+			 int npsvs, u32 *sig_index);
+int mlx5_core_destroy_psv(struct mlx5_core_dev *dev, int psv_num);
 
 static inline u32 mlx5_mkey_to_idx(u32 mkey)
 {

include/linux/mlx5/qp.h

@@ -37,6 +37,9 @@
 #include <linux/mlx5/driver.h>
 
 #define MLX5_INVALID_LKEY	0x100
+#define MLX5_SIG_WQE_SIZE	(MLX5_SEND_WQE_BB * 5)
+#define MLX5_DIF_SIZE		8
+#define MLX5_STRIDE_BLOCK_OP	0x400
 
 enum mlx5_qp_optpar {
 	MLX5_QP_OPTPAR_ALT_ADDR_PATH		= 1 << 0,
@@ -151,6 +154,11 @@ enum {
 	MLX5_SND_DBR	= 1,
 };
 
+enum {
+	MLX5_FLAGS_INLINE	= 1<<7,
+	MLX5_FLAGS_CHECK_FREE   = 1<<5,
+};
+
 struct mlx5_wqe_fmr_seg {
 	__be32			flags;
 	__be32			mem_key;
@@ -278,6 +286,60 @@ struct mlx5_wqe_inline_seg {
 	__be32	byte_count;
 };
 
+struct mlx5_bsf {
+	struct mlx5_bsf_basic {
+		u8		bsf_size_sbs;
+		u8		check_byte_mask;
+		union {
+			u8	copy_byte_mask;
+			u8	bs_selector;
+			u8	rsvd_wflags;
+		} wire;
+		union {
+			u8	bs_selector;
+			u8	rsvd_mflags;
+		} mem;
+		__be32		raw_data_size;
+		__be32		w_bfs_psv;
+		__be32		m_bfs_psv;
+	} basic;
+	struct mlx5_bsf_ext {
+		__be32		t_init_gen_pro_size;
+		__be32		rsvd_epi_size;
+		__be32		w_tfs_psv;
+		__be32		m_tfs_psv;
+	} ext;
+	struct mlx5_bsf_inl {
+		__be32		w_inl_vld;
+		__be32		w_rsvd;
+		__be64		w_block_format;
+		__be32		m_inl_vld;
+		__be32		m_rsvd;
+		__be64		m_block_format;
+	} inl;
+};
+
+struct mlx5_klm {
+	__be32		bcount;
+	__be32		key;
+	__be64		va;
+};
+
+struct mlx5_stride_block_entry {
+	__be16		stride;
+	__be16		bcount;
+	__be32		key;
+	__be64		va;
+};
+
+struct mlx5_stride_block_ctrl_seg {
+	__be32		bcount_per_cycle;
+	__be32		op;
+	__be32		repeat_count;
+	u16		rsvd;
+	__be16		num_entries;
+};
+
 struct mlx5_core_qp {
 	void (*event)		(struct mlx5_core_qp *, int);
 	int			qpn;
@@ -444,6 +506,11 @@ static inline struct mlx5_core_qp *__mlx5_qp_lookup(struct mlx5_core_dev *dev, u
 	return radix_tree_lookup(&dev->priv.qp_table.tree, qpn);
 }
 
+static inline struct mlx5_core_mr *__mlx5_mr_lookup(struct mlx5_core_dev *dev, u32 key)
+{
+	return radix_tree_lookup(&dev->priv.mr_table.tree, key);
+}
+
 int mlx5_core_create_qp(struct mlx5_core_dev *dev,
 			struct mlx5_core_qp *qp,
 			struct mlx5_create_qp_mbox_in *in,

include/rdma/ib_cm.h

@@ -601,5 +601,4 @@ struct ib_cm_sidr_rep_param {
 int ib_send_cm_sidr_rep(struct ib_cm_id *cm_id,
 			struct ib_cm_sidr_rep_param *param);
 
-int ib_update_cm_av(struct ib_cm_id *id, const u8 *smac, const u8 *alt_smac);
 #endif /* IB_CM_H */

include/rdma/ib_umem.h

@@ -46,17 +46,12 @@ struct ib_umem {
 	int			page_size;
 	int                     writable;
 	int                     hugetlb;
-	struct list_head	chunk_list;
 	struct work_struct	work;
 	struct mm_struct       *mm;
 	unsigned long		diff;
-};
-
-struct ib_umem_chunk {
-	struct list_head	list;
-	int                     nents;
-	int                     nmap;
-	struct scatterlist      page_list[0];
+	struct sg_table sg_head;
+	int             nmap;
+	int             npages;
 };
 
 #ifdef CONFIG_INFINIBAND_USER_MEM

include/rdma/ib_verbs.h

@@ -122,7 +122,19 @@ enum ib_device_cap_flags {
 	IB_DEVICE_BLOCK_MULTICAST_LOOPBACK = (1<<22),
 	IB_DEVICE_MEM_WINDOW_TYPE_2A	= (1<<23),
 	IB_DEVICE_MEM_WINDOW_TYPE_2B	= (1<<24),
-	IB_DEVICE_MANAGED_FLOW_STEERING = (1<<29)
+	IB_DEVICE_MANAGED_FLOW_STEERING = (1<<29),
+	IB_DEVICE_SIGNATURE_HANDOVER	= (1<<30)
+};
+
+enum ib_signature_prot_cap {
+	IB_PROT_T10DIF_TYPE_1 = 1,
+	IB_PROT_T10DIF_TYPE_2 = 1 << 1,
+	IB_PROT_T10DIF_TYPE_3 = 1 << 2,
+};
+
+enum ib_signature_guard_cap {
+	IB_GUARD_T10DIF_CRC	= 1,
+	IB_GUARD_T10DIF_CSUM	= 1 << 1,
 };
 
 enum ib_atomic_cap {
@@ -172,6 +184,8 @@ struct ib_device_attr {
 	unsigned int		max_fast_reg_page_list_len;
 	u16			max_pkeys;
 	u8			local_ca_ack_delay;
+	int			sig_prot_cap;
+	int			sig_guard_cap;
 };
 
 enum ib_mtu {
@@ -461,6 +475,130 @@ int ib_rate_to_mult(enum ib_rate rate) __attribute_const__;
  */
 int ib_rate_to_mbps(enum ib_rate rate) __attribute_const__;
 
+enum ib_mr_create_flags {
+	IB_MR_SIGNATURE_EN = 1,
+};
+
+/**
+ * ib_mr_init_attr - Memory region init attributes passed to routine
+ *     ib_create_mr.
+ * @max_reg_descriptors: max number of registration descriptors that
+ *     may be used with registration work requests.
+ * @flags: MR creation flags bit mask.
+ */
+struct ib_mr_init_attr {
+	int	    max_reg_descriptors;
+	u32	    flags;
+};
+
+enum ib_signature_type {
+	IB_SIG_TYPE_T10_DIF,
+};
+
+/**
+ * T10-DIF Signature types
+ * T10-DIF types are defined by SCSI
+ * specifications.
+ */
+enum ib_t10_dif_type {
+	IB_T10DIF_NONE,
+	IB_T10DIF_TYPE1,
+	IB_T10DIF_TYPE2,
+	IB_T10DIF_TYPE3
+};
+
+/**
+ * Signature T10-DIF block-guard types
+ * IB_T10DIF_CRC: Corresponds to T10-PI mandated CRC checksum rules.
+ * IB_T10DIF_CSUM: Corresponds to IP checksum rules.
+ */
+enum ib_t10_dif_bg_type {
+	IB_T10DIF_CRC,
+	IB_T10DIF_CSUM
+};
+
+/**
+ * struct ib_t10_dif_domain - Parameters specific for T10-DIF
+ *     domain.
+ * @type: T10-DIF type (0|1|2|3)
+ * @bg_type: T10-DIF block guard type (CRC|CSUM)
+ * @pi_interval: protection information interval.
+ * @bg: seed of guard computation.
+ * @app_tag: application tag of guard block
+ * @ref_tag: initial guard block reference tag.
+ * @type3_inc_reftag: T10-DIF type 3 does not state
+ *     about the reference tag, it is the user
+ *     choice to increment it or not.
+ */
+struct ib_t10_dif_domain {
+	enum ib_t10_dif_type	type;
+	enum ib_t10_dif_bg_type bg_type;
+	u16			pi_interval;
+	u16			bg;
+	u16			app_tag;
+	u32			ref_tag;
+	bool			type3_inc_reftag;
+};
+
+/**
+ * struct ib_sig_domain - Parameters for signature domain
+ * @sig_type: specific signauture type
+ * @sig: union of all signature domain attributes that may
+ *     be used to set domain layout.
+ */
+struct ib_sig_domain {
+	enum ib_signature_type sig_type;
+	union {
+		struct ib_t10_dif_domain dif;
+	} sig;
+};
+
+/**
+ * struct ib_sig_attrs - Parameters for signature handover operation
+ * @check_mask: bitmask for signature byte check (8 bytes)
+ * @mem: memory domain layout desciptor.
+ * @wire: wire domain layout desciptor.
+ */
+struct ib_sig_attrs {
+	u8			check_mask;
+	struct ib_sig_domain	mem;
+	struct ib_sig_domain	wire;
+};
+
+enum ib_sig_err_type {
+	IB_SIG_BAD_GUARD,
+	IB_SIG_BAD_REFTAG,
+	IB_SIG_BAD_APPTAG,
+};
+
+/**
+ * struct ib_sig_err - signature error descriptor
+ */
+struct ib_sig_err {
+	enum ib_sig_err_type	err_type;
+	u32			expected;
+	u32			actual;
+	u64			sig_err_offset;
+	u32			key;
+};
+
+enum ib_mr_status_check {
+	IB_MR_CHECK_SIG_STATUS = 1,
+};
+
+/**
+ * struct ib_mr_status - Memory region status container
+ *
+ * @fail_status: Bitmask of MR checks status. For each
+ *     failed check a corresponding status bit is set.
+ * @sig_err: Additional info for IB_MR_CEHCK_SIG_STATUS
+ *     failure.
+ */
+struct ib_mr_status {
+	u32		    fail_status;
+	struct ib_sig_err   sig_err;
+};
+
 /**
  * mult_to_ib_rate - Convert a multiple of 2.5 Gbit/sec to an IB rate
  * enum.
@@ -644,6 +782,7 @@ enum ib_qp_create_flags {
 	IB_QP_CREATE_IPOIB_UD_LSO		= 1 << 0,
 	IB_QP_CREATE_BLOCK_MULTICAST_LOOPBACK	= 1 << 1,
 	IB_QP_CREATE_NETIF_QP			= 1 << 5,
+	IB_QP_CREATE_SIGNATURE_EN		= 1 << 6,
 	/* reserve bits 26-31 for low level drivers' internal use */
 	IB_QP_CREATE_RESERVED_START		= 1 << 26,
 	IB_QP_CREATE_RESERVED_END		= 1 << 31,
@@ -808,6 +947,7 @@ enum ib_wr_opcode {
 	IB_WR_MASKED_ATOMIC_CMP_AND_SWP,
 	IB_WR_MASKED_ATOMIC_FETCH_AND_ADD,
 	IB_WR_BIND_MW,
+	IB_WR_REG_SIG_MR,
 	/* reserve values for low level drivers' internal use.
 	 * These values will not be used at all in the ib core layer.
 	 */
@@ -913,6 +1053,12 @@ struct ib_send_wr {
 			u32                      rkey;
 			struct ib_mw_bind_info   bind_info;
 		} bind_mw;
+		struct {
+			struct ib_sig_attrs    *sig_attrs;
+			struct ib_mr	       *sig_mr;
+			int			access_flags;
+			struct ib_sge	       *prot;
+		} sig_handover;
 	} wr;
 	u32			xrc_remote_srq_num;	/* XRC TGT QPs only */
 };
@@ -1266,10 +1412,6 @@ struct ib_dma_mapping_ops {
 	void		(*unmap_sg)(struct ib_device *dev,
 				    struct scatterlist *sg, int nents,
 				    enum dma_data_direction direction);
-	u64		(*dma_address)(struct ib_device *dev,
-				       struct scatterlist *sg);
-	unsigned int	(*dma_len)(struct ib_device *dev,
-				   struct scatterlist *sg);
 	void		(*sync_single_for_cpu)(struct ib_device *dev,
 					       u64 dma_handle,
 					       size_t size,
@@ -1407,6 +1549,9 @@ struct ib_device {
 	int                        (*query_mr)(struct ib_mr *mr,
 					       struct ib_mr_attr *mr_attr);
 	int                        (*dereg_mr)(struct ib_mr *mr);
+	int                        (*destroy_mr)(struct ib_mr *mr);
+	struct ib_mr *		   (*create_mr)(struct ib_pd *pd,
+						struct ib_mr_init_attr *mr_init_attr);
 	struct ib_mr *		   (*alloc_fast_reg_mr)(struct ib_pd *pd,
 					       int max_page_list_len);
 	struct ib_fast_reg_page_list * (*alloc_fast_reg_page_list)(struct ib_device *device,
@@ -1455,6 +1600,8 @@ struct ib_device {
 						  *flow_attr,
 						  int domain);
 	int			   (*destroy_flow)(struct ib_flow *flow_id);
+	int			   (*check_mr_status)(struct ib_mr *mr, u32 check_mask,
+						      struct ib_mr_status *mr_status);
 
 	struct ib_dma_mapping_ops   *dma_ops;
 
@@ -2089,12 +2236,13 @@ static inline void ib_dma_unmap_sg_attrs(struct ib_device *dev,
  * ib_sg_dma_address - Return the DMA address from a scatter/gather entry
  * @dev: The device for which the DMA addresses were created
  * @sg: The scatter/gather entry
+ *
+ * Note: this function is obsolete. To do: change all occurrences of
+ * ib_sg_dma_address() into sg_dma_address().
  */
 static inline u64 ib_sg_dma_address(struct ib_device *dev,
 				    struct scatterlist *sg)
 {
-	if (dev->dma_ops)
-		return dev->dma_ops->dma_address(dev, sg);
 	return sg_dma_address(sg);
 }
 
@@ -2102,12 +2250,13 @@ static inline u64 ib_sg_dma_address(struct ib_device *dev,
  * ib_sg_dma_len - Return the DMA length from a scatter/gather entry
  * @dev: The device for which the DMA addresses were created
  * @sg: The scatter/gather entry
+ *
+ * Note: this function is obsolete. To do: change all occurrences of
+ * ib_sg_dma_len() into sg_dma_len().
  */
 static inline unsigned int ib_sg_dma_len(struct ib_device *dev,
 					 struct scatterlist *sg)
 {
-	if (dev->dma_ops)
-		return dev->dma_ops->dma_len(dev, sg);
 	return sg_dma_len(sg);
 }
 
@@ -2250,6 +2399,25 @@ int ib_query_mr(struct ib_mr *mr, struct ib_mr_attr *mr_attr);
  */
 int ib_dereg_mr(struct ib_mr *mr);
 
+
+/**
+ * ib_create_mr - Allocates a memory region that may be used for
+ *     signature handover operations.
+ * @pd: The protection domain associated with the region.
+ * @mr_init_attr: memory region init attributes.
+ */
+struct ib_mr *ib_create_mr(struct ib_pd *pd,
+			   struct ib_mr_init_attr *mr_init_attr);
+
+/**
+ * ib_destroy_mr - Destroys a memory region that was created using
+ *     ib_create_mr and removes it from HW translation tables.
+ * @mr: The memory region to destroy.
+ *
+ * This function can fail, if the memory region has memory windows bound to it.
+ */
+int ib_destroy_mr(struct ib_mr *mr);
+
 /**
  * ib_alloc_fast_reg_mr - Allocates memory region usable with the
  *   IB_WR_FAST_REG_MR send work request.
@@ -2435,4 +2603,19 @@ static inline int ib_check_mr_access(int flags)
 	return 0;
 }
 
+/**
+ * ib_check_mr_status: lightweight check of MR status.
+ *     This routine may provide status checks on a selected
+ *     ib_mr. first use is for signature status check.
+ *
+ * @mr: A memory region.
+ * @check_mask: Bitmask of which checks to perform from
+ *     ib_mr_status_check enumeration.
+ * @mr_status: The container of relevant status checks.
+ *     failed checks will be indicated in the status bitmask
+ *     and the relevant info shall be in the error item.
+ */
+int ib_check_mr_status(struct ib_mr *mr, u32 check_mask,
+		       struct ib_mr_status *mr_status);
+
 #endif /* IB_VERBS_H */

include/scsi/libiscsi.h

@@ -133,6 +133,10 @@ struct iscsi_task {
 	unsigned long		last_xfer;
 	unsigned long		last_timeout;
 	bool			have_checked_conn;
+
+	/* T10 protection information */
+	bool			protected;
+
 	/* state set/tested under session->lock */
 	int			state;
 	atomic_t		refcount;

include/scsi/scsi_transport_iscsi.h

@@ -167,6 +167,7 @@ struct iscsi_transport {
 				 struct iscsi_bus_flash_conn *fnode_conn);
 	int (*logout_flashnode_sid) (struct iscsi_cls_session *cls_sess);
 	int (*get_host_stats) (struct Scsi_Host *shost, char *buf, int len);
+	u8 (*check_protection)(struct iscsi_task *task, sector_t *sector);
 };
 
 /*

include/scsi/scsi_transport_srp.h

@@ -41,7 +41,6 @@ enum srp_rport_state {
  * @mutex:             Protects against concurrent rport reconnect /
  *                     fast_io_fail / dev_loss_tmo activity.
  * @state:             rport state.
- * @deleted:           Whether or not srp_rport_del() has already been invoked.
  * @reconnect_delay:   Reconnect delay in seconds.
  * @failed_reconnects: Number of failed reconnect attempts.
  * @reconnect_work:    Work structure used for scheduling reconnect attempts.