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

Pull rdma updates from Doug Ledford:
 "Initial roundup of 4.5 merge window patches

   - Remove usage of ib_query_device and instead store attributes in
     ib_device struct

   - Move iopoll out of block and into lib, rename to irqpoll, and use
     in several places in the rdma stack as our new completion queue
     polling library mechanism.  Update the other block drivers that
     already used iopoll to use the new mechanism too.

   - Replace the per-entry GID table locks with a single GID table lock

   - IPoIB multicast cleanup

   - Cleanups to the IB MR facility

   - Add support for 64bit extended IB counters

   - Fix for netlink oops while parsing RDMA nl messages

   - RoCEv2 support for the core IB code

   - mlx4 RoCEv2 support

   - mlx5 RoCEv2 support

   - Cross Channel support for mlx5

   - Timestamp support for mlx5

   - Atomic support for mlx5

   - Raw QP support for mlx5

   - MAINTAINERS update for mlx4/mlx5

   - Misc ocrdma, qib, nes, usNIC, cxgb3, cxgb4, mlx4, mlx5 updates

   - Add support for remote invalidate to the iSER driver (pushed
     through the RDMA tree due to dependencies, acknowledged by nab)

   - Update to NFSoRDMA (pushed through the RDMA tree due to
     dependencies, acknowledged by Bruce)"

* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dledford/rdma: (169 commits)
  IB/mlx5: Unify CQ create flags check
  IB/mlx5: Expose Raw Packet QP to user space consumers
  {IB, net}/mlx5: Move the modify QP operation table to mlx5_ib
  IB/mlx5: Support setting Ethernet priority for Raw Packet QPs
  IB/mlx5: Add Raw Packet QP query functionality
  IB/mlx5: Add create and destroy functionality for Raw Packet QP
  IB/mlx5: Refactor mlx5_ib_qp to accommodate other QP types
  IB/mlx5: Allocate a Transport Domain for each ucontext
  net/mlx5_core: Warn on unsupported events of QP/RQ/SQ
  net/mlx5_core: Add RQ and SQ event handling
  net/mlx5_core: Export transport objects
  IB/mlx5: Expose CQE version to user-space
  IB/mlx5: Add CQE version 1 support to user QPs and SRQs
  IB/mlx5: Fix data validation in mlx5_ib_alloc_ucontext
  IB/sa: Fix netlink local service GFP crash
  IB/srpt: Remove redundant wc array
  IB/qib: Improve ipoib UD performance
  IB/mlx4: Advertise RoCE v2 support
  IB/mlx4: Create and use another QP1 for RoCEv2
  IB/mlx4: Enable send of RoCE QP1 packets with IP/UDP headers
  ...

1 files changed, 210 insertions, 113 deletions

diff --git a/drivers/infiniband/ulp/iser/iser_initiator.c b/drivers/infiniband/ulp/iser/iser_initiator.c
index ffd00c4..ed54b38 100644
--- a/drivers/infiniband/ulp/iser/iser_initiator.c
+++ b/drivers/infiniband/ulp/iser/iser_initiator.c
@@ -51,7 +51,7 @@ static int iser_prepare_read_cmd(struct iscsi_task *task)
 	struct iscsi_iser_task *iser_task = task->dd_data;
 	struct iser_mem_reg *mem_reg;
 	int err;
-	struct iser_hdr *hdr = &iser_task->desc.iser_header;
+	struct iser_ctrl *hdr = &iser_task->desc.iser_header;
 	struct iser_data_buf *buf_in = &iser_task->data[ISER_DIR_IN];
 
 	err = iser_dma_map_task_data(iser_task,
@@ -72,7 +72,7 @@ static int iser_prepare_read_cmd(struct iscsi_task *task)
 			return err;
 	}
 
-	err = iser_reg_rdma_mem(iser_task, ISER_DIR_IN);
+	err = iser_reg_rdma_mem(iser_task, ISER_DIR_IN, false);
 	if (err) {
 		iser_err("Failed to set up Data-IN RDMA\n");
 		return err;
@@ -104,7 +104,7 @@ iser_prepare_write_cmd(struct iscsi_task *task,
 	struct iscsi_iser_task *iser_task = task->dd_data;
 	struct iser_mem_reg *mem_reg;
 	int err;
-	struct iser_hdr *hdr = &iser_task->desc.iser_header;
+	struct iser_ctrl *hdr = &iser_task->desc.iser_header;
 	struct iser_data_buf *buf_out = &iser_task->data[ISER_DIR_OUT];
 	struct ib_sge *tx_dsg = &iser_task->desc.tx_sg[1];
 
@@ -126,7 +126,8 @@ iser_prepare_write_cmd(struct iscsi_task *task,
 			return err;
 	}
 
-	err = iser_reg_rdma_mem(iser_task, ISER_DIR_OUT);
+	err = iser_reg_rdma_mem(iser_task, ISER_DIR_OUT,
+				buf_out->data_len == imm_sz);
 	if (err != 0) {
 		iser_err("Failed to register write cmd RDMA mem\n");
 		return err;
@@ -166,7 +167,7 @@ static void iser_create_send_desc(struct iser_conn	*iser_conn,
 	ib_dma_sync_single_for_cpu(device->ib_device,
 		tx_desc->dma_addr, ISER_HEADERS_LEN, DMA_TO_DEVICE);
 
-	memset(&tx_desc->iser_header, 0, sizeof(struct iser_hdr));
+	memset(&tx_desc->iser_header, 0, sizeof(struct iser_ctrl));
 	tx_desc->iser_header.flags = ISER_VER;
 	tx_desc->num_sge = 1;
 }
@@ -174,73 +175,63 @@ static void iser_create_send_desc(struct iser_conn	*iser_conn,
 static void iser_free_login_buf(struct iser_conn *iser_conn)
 {
 	struct iser_device *device = iser_conn->ib_conn.device;
+	struct iser_login_desc *desc = &iser_conn->login_desc;
 
-	if (!iser_conn->login_buf)
+	if (!desc->req)
 		return;
 
-	if (iser_conn->login_req_dma)
-		ib_dma_unmap_single(device->ib_device,
-				    iser_conn->login_req_dma,
-				    ISCSI_DEF_MAX_RECV_SEG_LEN, DMA_TO_DEVICE);
+	ib_dma_unmap_single(device->ib_device, desc->req_dma,
+			    ISCSI_DEF_MAX_RECV_SEG_LEN, DMA_TO_DEVICE);
 
-	if (iser_conn->login_resp_dma)
-		ib_dma_unmap_single(device->ib_device,
-				    iser_conn->login_resp_dma,
-				    ISER_RX_LOGIN_SIZE, DMA_FROM_DEVICE);
+	ib_dma_unmap_single(device->ib_device, desc->rsp_dma,
+			    ISER_RX_LOGIN_SIZE, DMA_FROM_DEVICE);
 
-	kfree(iser_conn->login_buf);
+	kfree(desc->req);
+	kfree(desc->rsp);
 
 	/* make sure we never redo any unmapping */
-	iser_conn->login_req_dma = 0;
-	iser_conn->login_resp_dma = 0;
-	iser_conn->login_buf = NULL;
+	desc->req = NULL;
+	desc->rsp = NULL;
 }
 
 static int iser_alloc_login_buf(struct iser_conn *iser_conn)
 {
 	struct iser_device *device = iser_conn->ib_conn.device;
-	int			req_err, resp_err;
-
-	BUG_ON(device == NULL);
-
-	iser_conn->login_buf = kmalloc(ISCSI_DEF_MAX_RECV_SEG_LEN +
-				     ISER_RX_LOGIN_SIZE, GFP_KERNEL);
-	if (!iser_conn->login_buf)
-		goto out_err;
-
-	iser_conn->login_req_buf  = iser_conn->login_buf;
-	iser_conn->login_resp_buf = iser_conn->login_buf +
-						ISCSI_DEF_MAX_RECV_SEG_LEN;
-
-	iser_conn->login_req_dma = ib_dma_map_single(device->ib_device,
-						     iser_conn->login_req_buf,
-						     ISCSI_DEF_MAX_RECV_SEG_LEN,
-						     DMA_TO_DEVICE);
-
-	iser_conn->login_resp_dma = ib_dma_map_single(device->ib_device,
-						      iser_conn->login_resp_buf,
-						      ISER_RX_LOGIN_SIZE,
-						      DMA_FROM_DEVICE);
-
-	req_err  = ib_dma_mapping_error(device->ib_device,
-					iser_conn->login_req_dma);
-	resp_err = ib_dma_mapping_error(device->ib_device,
-					iser_conn->login_resp_dma);
-
-	if (req_err || resp_err) {
-		if (req_err)
-			iser_conn->login_req_dma = 0;
-		if (resp_err)
-			iser_conn->login_resp_dma = 0;
-		goto free_login_buf;
-	}
+	struct iser_login_desc *desc = &iser_conn->login_desc;
+
+	desc->req = kmalloc(ISCSI_DEF_MAX_RECV_SEG_LEN, GFP_KERNEL);
+	if (!desc->req)
+		return -ENOMEM;
+
+	desc->req_dma = ib_dma_map_single(device->ib_device, desc->req,
+					  ISCSI_DEF_MAX_RECV_SEG_LEN,
+					  DMA_TO_DEVICE);
+	if (ib_dma_mapping_error(device->ib_device,
+				desc->req_dma))
+		goto free_req;
+
+	desc->rsp = kmalloc(ISER_RX_LOGIN_SIZE, GFP_KERNEL);
+	if (!desc->rsp)
+		goto unmap_req;
+
+	desc->rsp_dma = ib_dma_map_single(device->ib_device, desc->rsp,
+					   ISER_RX_LOGIN_SIZE,
+					   DMA_FROM_DEVICE);
+	if (ib_dma_mapping_error(device->ib_device,
+				desc->rsp_dma))
+		goto free_rsp;
+
 	return 0;
 
-free_login_buf:
-	iser_free_login_buf(iser_conn);
+free_rsp:
+	kfree(desc->rsp);
+unmap_req:
+	ib_dma_unmap_single(device->ib_device, desc->req_dma,
+			    ISCSI_DEF_MAX_RECV_SEG_LEN,
+			    DMA_TO_DEVICE);
+free_req:
+	kfree(desc->req);
 
-out_err:
-	iser_err("unable to alloc or map login buf\n");
 	return -ENOMEM;
 }
 
@@ -280,11 +271,11 @@ int iser_alloc_rx_descriptors(struct iser_conn *iser_conn,
 			goto rx_desc_dma_map_failed;
 
 		rx_desc->dma_addr = dma_addr;
-
+		rx_desc->cqe.done = iser_task_rsp;
 		rx_sg = &rx_desc->rx_sg;
-		rx_sg->addr   = rx_desc->dma_addr;
+		rx_sg->addr = rx_desc->dma_addr;
 		rx_sg->length = ISER_RX_PAYLOAD_SIZE;
-		rx_sg->lkey   = device->pd->local_dma_lkey;
+		rx_sg->lkey = device->pd->local_dma_lkey;
 	}
 
 	iser_conn->rx_desc_head = 0;
@@ -383,6 +374,7 @@ 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;
+	tx_desc->cqe.done = iser_cmd_comp;
 	iser_create_send_desc(iser_conn, tx_desc);
 
 	if (hdr->flags & ISCSI_FLAG_CMD_READ) {
@@ -464,6 +456,7 @@ int iser_send_data_out(struct iscsi_conn *conn,
 	}
 
 	tx_desc->type = ISCSI_TX_DATAOUT;
+	tx_desc->cqe.done = iser_dataout_comp;
 	tx_desc->iser_header.flags = ISER_VER;
 	memcpy(&tx_desc->iscsi_header, hdr, sizeof(struct iscsi_hdr));
 
@@ -513,6 +506,7 @@ int iser_send_control(struct iscsi_conn *conn,
 
 	/* build the tx desc regd header and add it to the tx desc dto */
 	mdesc->type = ISCSI_TX_CONTROL;
+	mdesc->cqe.done = iser_ctrl_comp;
 	iser_create_send_desc(iser_conn, mdesc);
 
 	device = iser_conn->ib_conn.device;
@@ -520,25 +514,25 @@ int iser_send_control(struct iscsi_conn *conn,
 	data_seg_len = ntoh24(task->hdr->dlength);
 
 	if (data_seg_len > 0) {
+		struct iser_login_desc *desc = &iser_conn->login_desc;
 		struct ib_sge *tx_dsg = &mdesc->tx_sg[1];
+
 		if (task != conn->login_task) {
 			iser_err("data present on non login task!!!\n");
 			goto send_control_error;
 		}
 
-		ib_dma_sync_single_for_cpu(device->ib_device,
-			iser_conn->login_req_dma, task->data_count,
-			DMA_TO_DEVICE);
+		ib_dma_sync_single_for_cpu(device->ib_device, desc->req_dma,
+					   task->data_count, DMA_TO_DEVICE);
 
-		memcpy(iser_conn->login_req_buf, task->data, task->data_count);
+		memcpy(desc->req, task->data, task->data_count);
 
-		ib_dma_sync_single_for_device(device->ib_device,
-			iser_conn->login_req_dma, task->data_count,
-			DMA_TO_DEVICE);
+		ib_dma_sync_single_for_device(device->ib_device, desc->req_dma,
+					      task->data_count, DMA_TO_DEVICE);
 
-		tx_dsg->addr    = iser_conn->login_req_dma;
-		tx_dsg->length  = task->data_count;
-		tx_dsg->lkey    = device->pd->local_dma_lkey;
+		tx_dsg->addr = desc->req_dma;
+		tx_dsg->length = task->data_count;
+		tx_dsg->lkey = device->pd->local_dma_lkey;
 		mdesc->num_sge = 2;
 	}
 
@@ -562,41 +556,126 @@ send_control_error:
 	return err;
 }
 
-/**
- * iser_rcv_dto_completion - recv DTO completion
- */
-void iser_rcv_completion(struct iser_rx_desc *rx_desc,
-			 unsigned long rx_xfer_len,
-			 struct ib_conn *ib_conn)
+void iser_login_rsp(struct ib_cq *cq, struct ib_wc *wc)
 {
-	struct iser_conn *iser_conn = container_of(ib_conn, struct iser_conn,
-						   ib_conn);
+	struct ib_conn *ib_conn = wc->qp->qp_context;
+	struct iser_conn *iser_conn = to_iser_conn(ib_conn);
+	struct iser_login_desc *desc = iser_login(wc->wr_cqe);
 	struct iscsi_hdr *hdr;
-	u64 rx_dma;
-	int rx_buflen, outstanding, count, err;
+	char *data;
+	int length;
 
-	/* differentiate between login to all other PDUs */
-	if ((char *)rx_desc == iser_conn->login_resp_buf) {
-		rx_dma = iser_conn->login_resp_dma;
-		rx_buflen = ISER_RX_LOGIN_SIZE;
-	} else {
-		rx_dma = rx_desc->dma_addr;
-		rx_buflen = ISER_RX_PAYLOAD_SIZE;
+	if (unlikely(wc->status != IB_WC_SUCCESS)) {
+		iser_err_comp(wc, "login_rsp");
+		return;
+	}
+
+	ib_dma_sync_single_for_cpu(ib_conn->device->ib_device,
+				   desc->rsp_dma, ISER_RX_LOGIN_SIZE,
+				   DMA_FROM_DEVICE);
+
+	hdr = desc->rsp + sizeof(struct iser_ctrl);
+	data = desc->rsp + ISER_HEADERS_LEN;
+	length = wc->byte_len - ISER_HEADERS_LEN;
+
+	iser_dbg("op 0x%x itt 0x%x dlen %d\n", hdr->opcode,
+		 hdr->itt, length);
+
+	iscsi_iser_recv(iser_conn->iscsi_conn, hdr, data, length);
+
+	ib_dma_sync_single_for_device(ib_conn->device->ib_device,
+				      desc->rsp_dma, ISER_RX_LOGIN_SIZE,
+				      DMA_FROM_DEVICE);
+
+	ib_conn->post_recv_buf_count--;
+}
+
+static inline void
+iser_inv_desc(struct iser_fr_desc *desc, u32 rkey)
+{
+	if (likely(rkey == desc->rsc.mr->rkey))
+		desc->rsc.mr_valid = 0;
+	else if (likely(rkey == desc->pi_ctx->sig_mr->rkey))
+		desc->pi_ctx->sig_mr_valid = 0;
+}
+
+static int
+iser_check_remote_inv(struct iser_conn *iser_conn,
+		      struct ib_wc *wc,
+		      struct iscsi_hdr *hdr)
+{
+	if (wc->wc_flags & IB_WC_WITH_INVALIDATE) {
+		struct iscsi_task *task;
+		u32 rkey = wc->ex.invalidate_rkey;
+
+		iser_dbg("conn %p: remote invalidation for rkey %#x\n",
+			 iser_conn, rkey);
+
+		if (unlikely(!iser_conn->snd_w_inv)) {
+			iser_err("conn %p: unexepected remote invalidation, "
+				 "terminating connection\n", iser_conn);
+			return -EPROTO;
+		}
+
+		task = iscsi_itt_to_ctask(iser_conn->iscsi_conn, hdr->itt);
+		if (likely(task)) {
+			struct iscsi_iser_task *iser_task = task->dd_data;
+			struct iser_fr_desc *desc;
+
+			if (iser_task->dir[ISER_DIR_IN]) {
+				desc = iser_task->rdma_reg[ISER_DIR_IN].mem_h;
+				iser_inv_desc(desc, rkey);
+			}
+
+			if (iser_task->dir[ISER_DIR_OUT]) {
+				desc = iser_task->rdma_reg[ISER_DIR_OUT].mem_h;
+				iser_inv_desc(desc, rkey);
+			}
+		} else {
+			iser_err("failed to get task for itt=%d\n", hdr->itt);
+			return -EINVAL;
+		}
 	}
 
-	ib_dma_sync_single_for_cpu(ib_conn->device->ib_device, rx_dma,
-				   rx_buflen, DMA_FROM_DEVICE);
+	return 0;
+}
 
-	hdr = &rx_desc->iscsi_header;
+
+void iser_task_rsp(struct ib_cq *cq, struct ib_wc *wc)
+{
+	struct ib_conn *ib_conn = wc->qp->qp_context;
+	struct iser_conn *iser_conn = to_iser_conn(ib_conn);
+	struct iser_rx_desc *desc = iser_rx(wc->wr_cqe);
+	struct iscsi_hdr *hdr;
+	int length;
+	int outstanding, count, err;
+
+	if (unlikely(wc->status != IB_WC_SUCCESS)) {
+		iser_err_comp(wc, "task_rsp");
+		return;
+	}
+
+	ib_dma_sync_single_for_cpu(ib_conn->device->ib_device,
+				   desc->dma_addr, ISER_RX_PAYLOAD_SIZE,
+				   DMA_FROM_DEVICE);
+
+	hdr = &desc->iscsi_header;
+	length = wc->byte_len - ISER_HEADERS_LEN;
 
 	iser_dbg("op 0x%x itt 0x%x dlen %d\n", hdr->opcode,
-			hdr->itt, (int)(rx_xfer_len - ISER_HEADERS_LEN));
+		 hdr->itt, length);
+
+	if (iser_check_remote_inv(iser_conn, wc, hdr)) {
+		iscsi_conn_failure(iser_conn->iscsi_conn,
+				   ISCSI_ERR_CONN_FAILED);
+		return;
+	}
 
-	iscsi_iser_recv(iser_conn->iscsi_conn, hdr, rx_desc->data,
-			rx_xfer_len - ISER_HEADERS_LEN);
+	iscsi_iser_recv(iser_conn->iscsi_conn, hdr, desc->data, length);
 
-	ib_dma_sync_single_for_device(ib_conn->device->ib_device, rx_dma,
-				      rx_buflen, DMA_FROM_DEVICE);
+	ib_dma_sync_single_for_device(ib_conn->device->ib_device,
+				      desc->dma_addr, ISER_RX_PAYLOAD_SIZE,
+				      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   *
@@ -604,9 +683,6 @@ void iser_rcv_completion(struct iser_rx_desc *rx_desc,
 	 * for the posted rx bufs refcount to become zero handles everything   */
 	ib_conn->post_recv_buf_count--;
 
-	if (rx_dma == iser_conn->login_resp_dma)
-		return;
-
 	outstanding = ib_conn->post_recv_buf_count;
 	if (outstanding + iser_conn->min_posted_rx <= iser_conn->qp_max_recv_dtos) {
 		count = min(iser_conn->qp_max_recv_dtos - outstanding,
@@ -617,26 +693,47 @@ void iser_rcv_completion(struct iser_rx_desc *rx_desc,
 	}
 }
 
-void iser_snd_completion(struct iser_tx_desc *tx_desc,
-			struct ib_conn *ib_conn)
+void iser_cmd_comp(struct ib_cq *cq, struct ib_wc *wc)
 {
+	if (unlikely(wc->status != IB_WC_SUCCESS))
+		iser_err_comp(wc, "command");
+}
+
+void iser_ctrl_comp(struct ib_cq *cq, struct ib_wc *wc)
+{
+	struct iser_tx_desc *desc = iser_tx(wc->wr_cqe);
 	struct iscsi_task *task;
-	struct iser_device *device = ib_conn->device;
 
-	if (tx_desc->type == ISCSI_TX_DATAOUT) {
-		ib_dma_unmap_single(device->ib_device, tx_desc->dma_addr,
-					ISER_HEADERS_LEN, DMA_TO_DEVICE);
-		kmem_cache_free(ig.desc_cache, tx_desc);
-		tx_desc = NULL;
+	if (unlikely(wc->status != IB_WC_SUCCESS)) {
+		iser_err_comp(wc, "control");
+		return;
 	}
 
-	if (tx_desc && tx_desc->type == ISCSI_TX_CONTROL) {
-		/* this arithmetic is legal by libiscsi dd_data allocation */
-		task = (void *) ((long)(void *)tx_desc -
-				  sizeof(struct iscsi_task));
-		if (task->hdr->itt == RESERVED_ITT)
-			iscsi_put_task(task);
-	}
+	/* this arithmetic is legal by libiscsi dd_data allocation */
+	task = (void *)desc - sizeof(struct iscsi_task);
+	if (task->hdr->itt == RESERVED_ITT)
+		iscsi_put_task(task);
+}
+
+void iser_dataout_comp(struct ib_cq *cq, struct ib_wc *wc)
+{
+	struct iser_tx_desc *desc = iser_tx(wc->wr_cqe);
+	struct ib_conn *ib_conn = wc->qp->qp_context;
+	struct iser_device *device = ib_conn->device;
+
+	if (unlikely(wc->status != IB_WC_SUCCESS))
+		iser_err_comp(wc, "dataout");
+
+	ib_dma_unmap_single(device->ib_device, desc->dma_addr,
+			    ISER_HEADERS_LEN, DMA_TO_DEVICE);
+	kmem_cache_free(ig.desc_cache, desc);
+}
+
+void iser_last_comp(struct ib_cq *cq, struct ib_wc *wc)
+{
+	struct ib_conn *ib_conn = wc->qp->qp_context;
+
+	complete(&ib_conn->last_comp);
 }
 
 void iser_task_rdma_init(struct iscsi_iser_task *iser_task)