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

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband:
  IB/core: Use kmemdup() instead of kmalloc()+memcpy()
  IB/iser: Fix error flow in iser_create_ib_conn_res()
  IB/iser: Enhance disconnection logic for multi-pathing
  IB/iser: Remove buggy back-pointer setting
  IB/iser: Add asynchronous event handler
  MAINTAINERS: Add cxgb4 and iw_cxgb4 entries
  RDMA/cxgb3: Shrink .text with compile-time init of handlers arrays
  IPoIB: Allow disabling/enabling TSO on the fly through ethtool
  IB/mlx4: Add support for masked atomic operations
  IB/core: Add support for masked atomic operations
  RDMA/cma: Randomize local port allocation
  RDMA/nes: Make unnecessarily global functions static
  RDMA/nes: Make nesadapter->phy_lock usage consistent
  RDMA/cxgb4: Add driver for Chelsio T4 RNIC
  IB/mthca: Use the dma state API instead of pci equivalents
  RDMA/amso1100: Use the dma state API instead of pci equivalents
  RDMA/cxgb3: Don't free skbs on NET_XMIT_* indications from LLD
  RDMA/cxgb3: Use the dma state API instead of pci equivalents
  IB: Explicitly rule out llseek to avoid BKL in default_llseek()

MAINTAINERS

@@ -1749,6 +1749,20 @@ W:	http://www.openfabrics.org
 S:	Supported
 F:	drivers/infiniband/hw/cxgb3/
 
+CXGB4 ETHERNET DRIVER (CXGB4)
+M:	Dimitris Michailidis <dm@chelsio.com>
+L:	netdev@vger.kernel.org
+W:	http://www.chelsio.com
+S:	Supported
+F:	drivers/net/cxgb4/
+
+CXGB4 IWARP RNIC DRIVER (IW_CXGB4)
+M:	Steve Wise <swise@chelsio.com>
+L:	linux-rdma@vger.kernel.org
+W:	http://www.openfabrics.org
+S:	Supported
+F:	drivers/infiniband/hw/cxgb4/
+
 CYBERPRO FB DRIVER
 M:	Russell King <linux@arm.linux.org.uk>
 L:	linux-arm-kernel@lists.infradead.org (moderated for non-subscribers)

drivers/infiniband/Kconfig

@@ -46,6 +46,7 @@ source "drivers/infiniband/hw/ipath/Kconfig"
 source "drivers/infiniband/hw/ehca/Kconfig"
 source "drivers/infiniband/hw/amso1100/Kconfig"
 source "drivers/infiniband/hw/cxgb3/Kconfig"
+source "drivers/infiniband/hw/cxgb4/Kconfig"
 source "drivers/infiniband/hw/mlx4/Kconfig"
 source "drivers/infiniband/hw/nes/Kconfig"
 

drivers/infiniband/Makefile

@@ -4,6 +4,7 @@ obj-$(CONFIG_INFINIBAND_IPATH)		+= hw/ipath/
 obj-$(CONFIG_INFINIBAND_EHCA)		+= hw/ehca/
 obj-$(CONFIG_INFINIBAND_AMSO1100)	+= hw/amso1100/
 obj-$(CONFIG_INFINIBAND_CXGB3)		+= hw/cxgb3/
+obj-$(CONFIG_INFINIBAND_CXGB4)		+= hw/cxgb4/
 obj-$(CONFIG_MLX4_INFINIBAND)		+= hw/mlx4/
 obj-$(CONFIG_INFINIBAND_NES)		+= hw/nes/
 obj-$(CONFIG_INFINIBAND_IPOIB)		+= ulp/ipoib/

drivers/infiniband/core/cma.c

@@ -79,7 +79,6 @@ static DEFINE_IDR(sdp_ps);
 static DEFINE_IDR(tcp_ps);
 static DEFINE_IDR(udp_ps);
 static DEFINE_IDR(ipoib_ps);
-static int next_port;
 
 struct cma_device {
 	struct list_head	list;
@@ -1677,13 +1676,13 @@ int rdma_set_ib_paths(struct rdma_cm_id *id,
 	if (!cma_comp_exch(id_priv, CMA_ADDR_RESOLVED, CMA_ROUTE_RESOLVED))
 		return -EINVAL;
 
-	id->route.path_rec = kmalloc(sizeof *path_rec * num_paths, GFP_KERNEL);
+	id->route.path_rec = kmemdup(path_rec, sizeof *path_rec * num_paths,
+				     GFP_KERNEL);
 	if (!id->route.path_rec) {
 		ret = -ENOMEM;
 		goto err;
 	}
 
-	memcpy(id->route.path_rec, path_rec, sizeof *path_rec * num_paths);
 	id->route.num_paths = num_paths;
 	return 0;
 err:
@@ -1970,47 +1969,33 @@ static int cma_alloc_port(struct idr *ps, struct rdma_id_private *id_priv,
 
 static int cma_alloc_any_port(struct idr *ps, struct rdma_id_private *id_priv)
 {
-	struct rdma_bind_list *bind_list;
-	int port, ret, low, high;
-
-	bind_list = kzalloc(sizeof *bind_list, GFP_KERNEL);
-	if (!bind_list)
-		return -ENOMEM;
-
-retry:
-	/* FIXME: add proper port randomization per like inet_csk_get_port */
-	do {
-		ret = idr_get_new_above(ps, bind_list, next_port, &port);
-	} while ((ret == -EAGAIN) && idr_pre_get(ps, GFP_KERNEL));
-
-	if (ret)
-		goto err1;
+	static unsigned int last_used_port;
+	int low, high, remaining;
+	unsigned int rover;
 
 	inet_get_local_port_range(&low, &high);
-	if (port > high) {
-		if (next_port != low) {
-			idr_remove(ps, port);
-			next_port = low;
-			goto retry;
-		}
-		ret = -EADDRNOTAVAIL;
-		goto err2;
+	remaining = (high - low) + 1;
+	rover = net_random() % remaining + low;
+retry:
+	if (last_used_port != rover &&
+	    !idr_find(ps, (unsigned short) rover)) {
+		int ret = cma_alloc_port(ps, id_priv, rover);
+		/*
+		 * Remember previously used port number in order to avoid
+		 * re-using same port immediately after it is closed.
+		 */
+		if (!ret)
+			last_used_port = rover;
+		if (ret != -EADDRNOTAVAIL)
+			return ret;
 	}
-
-	if (port == high)
-		next_port = low;
-	else
-		next_port = port + 1;
-
-	bind_list->ps = ps;
-	bind_list->port = (unsigned short) port;
-	cma_bind_port(bind_list, id_priv);
-	return 0;
-err2:
-	idr_remove(ps, port);
-err1:
-	kfree(bind_list);
-	return ret;
+	if (--remaining) {
+		rover++;
+		if ((rover < low) || (rover > high))
+			rover = low;
+		goto retry;
+	}
+	return -EADDRNOTAVAIL;
 }
 
 static int cma_use_port(struct idr *ps, struct rdma_id_private *id_priv)
@@ -2995,12 +2980,7 @@ static void cma_remove_one(struct ib_device *device)
 
 static int __init cma_init(void)
 {
-	int ret, low, high, remaining;
-
-	get_random_bytes(&next_port, sizeof next_port);
-	inet_get_local_port_range(&low, &high);
-	remaining = (high - low) + 1;
-	next_port = ((unsigned int) next_port % remaining) + low;
+	int ret;
 
 	cma_wq = create_singlethread_workqueue("rdma_cm");
 	if (!cma_wq)

drivers/infiniband/core/mad.c

@@ -291,13 +291,11 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device,
 	}
 
 	if (mad_reg_req) {
-		reg_req = kmalloc(sizeof *reg_req, GFP_KERNEL);
+		reg_req = kmemdup(mad_reg_req, sizeof *reg_req, GFP_KERNEL);
 		if (!reg_req) {
 			ret = ERR_PTR(-ENOMEM);
 			goto error3;
 		}
-		/* Make a copy of the MAD registration request */
-		memcpy(reg_req, mad_reg_req, sizeof *reg_req);
 	}
 
 	/* Now, fill in the various structures */

drivers/infiniband/core/ucm.c

@@ -1181,7 +1181,7 @@ static int ib_ucm_open(struct inode *inode, struct file *filp)
 	file->filp = filp;
 	file->device = container_of(inode->i_cdev, struct ib_ucm_device, cdev);
 
-	return 0;
+	return nonseekable_open(inode, filp);
 }
 
 static int ib_ucm_close(struct inode *inode, struct file *filp)
@@ -1229,6 +1229,7 @@ static const struct file_operations ucm_fops = {
 	.release = ib_ucm_close,
 	.write	 = ib_ucm_write,
 	.poll    = ib_ucm_poll,
+	.llseek	 = no_llseek,
 };
 
 static ssize_t show_ibdev(struct device *dev, struct device_attribute *attr,

drivers/infiniband/core/ucma.c

@@ -1220,7 +1220,8 @@ static int ucma_open(struct inode *inode, struct file *filp)
 
 	filp->private_data = file;
 	file->filp = filp;
-	return 0;
+
+	return nonseekable_open(inode, filp);
 }
 
 static int ucma_close(struct inode *inode, struct file *filp)
@@ -1250,6 +1251,7 @@ static const struct file_operations ucma_fops = {
 	.release = ucma_close,
 	.write	 = ucma_write,
 	.poll    = ucma_poll,
+	.llseek	 = no_llseek,
 };
 
 static struct miscdevice ucma_misc = {

drivers/infiniband/core/user_mad.c

@@ -781,7 +781,7 @@ static int ib_umad_open(struct inode *inode, struct file *filp)
 {
 	struct ib_umad_port *port;
 	struct ib_umad_file *file;
-	int ret = 0;
+	int ret;
 
 	port = container_of(inode->i_cdev, struct ib_umad_port, cdev);
 	if (port)
@@ -814,6 +814,8 @@ static int ib_umad_open(struct inode *inode, struct file *filp)
 
 	list_add_tail(&file->port_list, &port->file_list);
 
+	ret = nonseekable_open(inode, filp);
+
 out:
 	mutex_unlock(&port->file_mutex);
 	return ret;
@@ -866,7 +868,8 @@ static const struct file_operations umad_fops = {
 	.compat_ioctl	= ib_umad_compat_ioctl,
 #endif
 	.open		= ib_umad_open,
-	.release	= ib_umad_close
+	.release	= ib_umad_close,
+	.llseek		= no_llseek,
 };
 
 static int ib_umad_sm_open(struct inode *inode, struct file *filp)
@@ -903,7 +906,7 @@ static int ib_umad_sm_open(struct inode *inode, struct file *filp)
 
 	filp->private_data = port;
 
-	return 0;
+	return nonseekable_open(inode, filp);
 
 fail:
 	kref_put(&port->umad_dev->ref, ib_umad_release_dev);
@@ -933,7 +936,8 @@ static int ib_umad_sm_close(struct inode *inode, struct file *filp)
 static const struct file_operations umad_sm_fops = {
 	.owner	 = THIS_MODULE,
 	.open	 = ib_umad_sm_open,
-	.release = ib_umad_sm_close
+	.release = ib_umad_sm_close,
+	.llseek	 = no_llseek,
 };
 
 static struct ib_client umad_client = {

drivers/infiniband/core/uverbs_main.c

@@ -369,7 +369,8 @@ static const struct file_operations uverbs_event_fops = {
 	.read	 = ib_uverbs_event_read,
 	.poll    = ib_uverbs_event_poll,
 	.release = ib_uverbs_event_close,
-	.fasync  = ib_uverbs_event_fasync
+	.fasync  = ib_uverbs_event_fasync,
+	.llseek	 = no_llseek,
 };
 
 void ib_uverbs_comp_handler(struct ib_cq *cq, void *cq_context)
@@ -623,7 +624,7 @@ static int ib_uverbs_open(struct inode *inode, struct file *filp)
 
 	filp->private_data = file;
 
-	return 0;
+	return nonseekable_open(inode, filp);
 
 err_module:
 	module_put(dev->ib_dev->owner);
@@ -651,7 +652,8 @@ static const struct file_operations uverbs_fops = {
 	.owner	 = THIS_MODULE,
 	.write	 = ib_uverbs_write,
 	.open	 = ib_uverbs_open,
-	.release = ib_uverbs_close
+	.release = ib_uverbs_close,
+	.llseek	 = no_llseek,
 };
 
 static const struct file_operations uverbs_mmap_fops = {
@@ -659,7 +661,8 @@ static const struct file_operations uverbs_mmap_fops = {
 	.write	 = ib_uverbs_write,
 	.mmap    = ib_uverbs_mmap,
 	.open	 = ib_uverbs_open,
-	.release = ib_uverbs_close
+	.release = ib_uverbs_close,
+	.llseek	 = no_llseek,
 };
 
 static struct ib_client uverbs_client = {

drivers/infiniband/hw/amso1100/c2.h

@@ -250,7 +250,7 @@ struct c2_array {
 struct sp_chunk {
 	struct sp_chunk *next;
 	dma_addr_t dma_addr;
-	DECLARE_PCI_UNMAP_ADDR(mapping);
+	DEFINE_DMA_UNMAP_ADDR(mapping);
 	u16 head;
 	u16 shared_ptr[0];
 };

drivers/infiniband/hw/amso1100/c2_alloc.c

@@ -49,7 +49,7 @@ static int c2_alloc_mqsp_chunk(struct c2_dev *c2dev, gfp_t gfp_mask,
 		return -ENOMEM;
 
 	new_head->dma_addr = dma_addr;
-	pci_unmap_addr_set(new_head, mapping, new_head->dma_addr);
+	dma_unmap_addr_set(new_head, mapping, new_head->dma_addr);
 
 	new_head->next = NULL;
 	new_head->head = 0;
@@ -81,7 +81,7 @@ void c2_free_mqsp_pool(struct c2_dev *c2dev, struct sp_chunk *root)
 	while (root) {
 		next = root->next;
 		dma_free_coherent(&c2dev->pcidev->dev, PAGE_SIZE, root,
-				  pci_unmap_addr(root, mapping));
+				  dma_unmap_addr(root, mapping));
 		root = next;
 	}
 }

drivers/infiniband/hw/amso1100/c2_cq.c

@@ -257,7 +257,7 @@ int c2_arm_cq(struct ib_cq *ibcq, enum ib_cq_notify_flags notify_flags)
 static void c2_free_cq_buf(struct c2_dev *c2dev, struct c2_mq *mq)
 {
 	dma_free_coherent(&c2dev->pcidev->dev, mq->q_size * mq->msg_size,
-			  mq->msg_pool.host, pci_unmap_addr(mq, mapping));
+			  mq->msg_pool.host, dma_unmap_addr(mq, mapping));
 }
 
 static int c2_alloc_cq_buf(struct c2_dev *c2dev, struct c2_mq *mq, int q_size,
@@ -278,7 +278,7 @@ static int c2_alloc_cq_buf(struct c2_dev *c2dev, struct c2_mq *mq, int q_size,
 		       NULL,	/* peer (currently unknown) */
 		       C2_MQ_HOST_TARGET);
 
-	pci_unmap_addr_set(mq, mapping, mq->host_dma);
+	dma_unmap_addr_set(mq, mapping, mq->host_dma);
 
 	return 0;
 }

drivers/infiniband/hw/amso1100/c2_mq.h

@@ -71,7 +71,7 @@ struct c2_mq {
 		u8 __iomem *adapter;
 	} msg_pool;
 	dma_addr_t host_dma;
-	DECLARE_PCI_UNMAP_ADDR(mapping);
+	DEFINE_DMA_UNMAP_ADDR(mapping);
 	u16 hint_count;
 	u16 priv;
 	struct c2_mq_shared __iomem *peer;

drivers/infiniband/hw/amso1100/c2_provider.h

@@ -50,7 +50,7 @@
 
 struct c2_buf_list {
 	void *buf;
-	 DECLARE_PCI_UNMAP_ADDR(mapping)
+	DEFINE_DMA_UNMAP_ADDR(mapping);
 };
 
 

drivers/infiniband/hw/amso1100/c2_rnic.c

@@ -524,7 +524,7 @@ int __devinit c2_rnic_init(struct c2_dev *c2dev)
 		err = -ENOMEM;
 		goto bail1;
 	}
-	pci_unmap_addr_set(&c2dev->rep_vq, mapping, c2dev->rep_vq.host_dma);
+	dma_unmap_addr_set(&c2dev->rep_vq, mapping, c2dev->rep_vq.host_dma);
 	pr_debug("%s rep_vq va %p dma %llx\n", __func__, q1_pages,
 		 (unsigned long long) c2dev->rep_vq.host_dma);
 	c2_mq_rep_init(&c2dev->rep_vq,
@@ -545,7 +545,7 @@ int __devinit c2_rnic_init(struct c2_dev *c2dev)
 		err = -ENOMEM;
 		goto bail2;
 	}
-	pci_unmap_addr_set(&c2dev->aeq, mapping, c2dev->aeq.host_dma);
+	dma_unmap_addr_set(&c2dev->aeq, mapping, c2dev->aeq.host_dma);
 	pr_debug("%s aeq va %p dma %llx\n", __func__, q2_pages,
 		 (unsigned long long) c2dev->aeq.host_dma);
 	c2_mq_rep_init(&c2dev->aeq,
@@ -596,11 +596,11 @@ int __devinit c2_rnic_init(struct c2_dev *c2dev)
       bail3:
 	dma_free_coherent(&c2dev->pcidev->dev,
 			  c2dev->aeq.q_size * c2dev->aeq.msg_size,
-			  q2_pages, pci_unmap_addr(&c2dev->aeq, mapping));
+			  q2_pages, dma_unmap_addr(&c2dev->aeq, mapping));
       bail2:
 	dma_free_coherent(&c2dev->pcidev->dev,
 			  c2dev->rep_vq.q_size * c2dev->rep_vq.msg_size,
-			  q1_pages, pci_unmap_addr(&c2dev->rep_vq, mapping));
+			  q1_pages, dma_unmap_addr(&c2dev->rep_vq, mapping));
       bail1:
 	c2_free_mqsp_pool(c2dev, c2dev->kern_mqsp_pool);
       bail0:
@@ -637,13 +637,13 @@ void __devexit c2_rnic_term(struct c2_dev *c2dev)
 	dma_free_coherent(&c2dev->pcidev->dev,
 			  c2dev->aeq.q_size * c2dev->aeq.msg_size,
 			  c2dev->aeq.msg_pool.host,
-			  pci_unmap_addr(&c2dev->aeq, mapping));
+			  dma_unmap_addr(&c2dev->aeq, mapping));
 
 	/* Free the verbs reply queue */
 	dma_free_coherent(&c2dev->pcidev->dev,
 			  c2dev->rep_vq.q_size * c2dev->rep_vq.msg_size,
 			  c2dev->rep_vq.msg_pool.host,
-			  pci_unmap_addr(&c2dev->rep_vq, mapping));
+			  dma_unmap_addr(&c2dev->rep_vq, mapping));
 
 	/* Free the MQ shared pointer pool */
 	c2_free_mqsp_pool(c2dev, c2dev->kern_mqsp_pool);

drivers/infiniband/hw/cxgb3/cxio_hal.c

@@ -174,7 +174,7 @@ int cxio_create_cq(struct cxio_rdev *rdev_p, struct t3_cq *cq, int kernel)
 		kfree(cq->sw_queue);
 		return -ENOMEM;
 	}
-	pci_unmap_addr_set(cq, mapping, cq->dma_addr);
+	dma_unmap_addr_set(cq, mapping, cq->dma_addr);
 	memset(cq->queue, 0, size);
 	setup.id = cq->cqid;
 	setup.base_addr = (u64) (cq->dma_addr);
@@ -297,7 +297,7 @@ int cxio_create_qp(struct cxio_rdev *rdev_p, u32 kernel_domain,
 		goto err4;
 
 	memset(wq->queue, 0, depth * sizeof(union t3_wr));
-	pci_unmap_addr_set(wq, mapping, wq->dma_addr);
+	dma_unmap_addr_set(wq, mapping, wq->dma_addr);
 	wq->doorbell = (void __iomem *)rdev_p->rnic_info.kdb_addr;
 	if (!kernel_domain)
 		wq->udb = (u64)rdev_p->rnic_info.udbell_physbase +
@@ -325,7 +325,7 @@ int cxio_destroy_cq(struct cxio_rdev *rdev_p, struct t3_cq *cq)
 	dma_free_coherent(&(rdev_p->rnic_info.pdev->dev),
 			  (1UL << (cq->size_log2))
 			  * sizeof(struct t3_cqe), cq->queue,
-			  pci_unmap_addr(cq, mapping));
+			  dma_unmap_addr(cq, mapping));
 	cxio_hal_put_cqid(rdev_p->rscp, cq->cqid);
 	return err;
 }
@@ -336,7 +336,7 @@ int cxio_destroy_qp(struct cxio_rdev *rdev_p, struct t3_wq *wq,
 	dma_free_coherent(&(rdev_p->rnic_info.pdev->dev),
 			  (1UL << (wq->size_log2))
 			  * sizeof(union t3_wr), wq->queue,
-			  pci_unmap_addr(wq, mapping));
+			  dma_unmap_addr(wq, mapping));
 	kfree(wq->sq);
 	cxio_hal_rqtpool_free(rdev_p, wq->rq_addr, (1UL << wq->rq_size_log2));
 	kfree(wq->rq);
@@ -537,7 +537,7 @@ static int cxio_hal_init_ctrl_qp(struct cxio_rdev *rdev_p)
 		err = -ENOMEM;
 		goto err;
 	}
-	pci_unmap_addr_set(&rdev_p->ctrl_qp, mapping,
+	dma_unmap_addr_set(&rdev_p->ctrl_qp, mapping,
 			   rdev_p->ctrl_qp.dma_addr);
 	rdev_p->ctrl_qp.doorbell = (void __iomem *)rdev_p->rnic_info.kdb_addr;
 	memset(rdev_p->ctrl_qp.workq, 0,
@@ -583,7 +583,7 @@ static int cxio_hal_destroy_ctrl_qp(struct cxio_rdev *rdev_p)
 	dma_free_coherent(&(rdev_p->rnic_info.pdev->dev),
 			  (1UL << T3_CTRL_QP_SIZE_LOG2)
 			  * sizeof(union t3_wr), rdev_p->ctrl_qp.workq,
-			  pci_unmap_addr(&rdev_p->ctrl_qp, mapping));
+			  dma_unmap_addr(&rdev_p->ctrl_qp, mapping));
 	return cxio_hal_clear_qp_ctx(rdev_p, T3_CTRL_QP_ID);
 }
 

drivers/infiniband/hw/cxgb3/cxio_hal.h

@@ -71,7 +71,7 @@ struct cxio_hal_ctrl_qp {
 	wait_queue_head_t waitq;/* wait for RspQ/CQE msg */
 	union t3_wr *workq;	/* the work request queue */
 	dma_addr_t dma_addr;	/* pci bus address of the workq */
-	DECLARE_PCI_UNMAP_ADDR(mapping)
+	DEFINE_DMA_UNMAP_ADDR(mapping);
 	void __iomem *doorbell;
 };
 

drivers/infiniband/hw/cxgb3/cxio_wr.h

@@ -691,7 +691,7 @@ struct t3_swrq {
 struct t3_wq {
 	union t3_wr *queue;		/* DMA accessable memory */
 	dma_addr_t dma_addr;		/* DMA address for HW */
-	DECLARE_PCI_UNMAP_ADDR(mapping)	/* unmap kruft */
+	DEFINE_DMA_UNMAP_ADDR(mapping); /* unmap kruft */
 	u32 error;			/* 1 once we go to ERROR */
 	u32 qpid;
 	u32 wptr;			/* idx to next available WR slot */
@@ -718,7 +718,7 @@ struct t3_cq {
 	u32 wptr;
 	u32 size_log2;
 	dma_addr_t dma_addr;
-	DECLARE_PCI_UNMAP_ADDR(mapping)
+	DEFINE_DMA_UNMAP_ADDR(mapping);
 	struct t3_cqe *queue;
 	struct t3_cqe *sw_queue;
 	u32 sw_rptr;

drivers/infiniband/hw/cxgb3/iwch.c

@@ -47,8 +47,6 @@ MODULE_DESCRIPTION("Chelsio T3 RDMA Driver");
 MODULE_LICENSE("Dual BSD/GPL");
 MODULE_VERSION(DRV_VERSION);
 
-cxgb3_cpl_handler_func t3c_handlers[NUM_CPL_CMDS];
-
 static void open_rnic_dev(struct t3cdev *);
 static void close_rnic_dev(struct t3cdev *);
 static void iwch_event_handler(struct t3cdev *, u32, u32);

drivers/infiniband/hw/cxgb3/iwch_cm.c

@@ -102,12 +102,9 @@ static unsigned int cong_flavor = 1;
 module_param(cong_flavor, uint, 0644);
 MODULE_PARM_DESC(cong_flavor, "TCP Congestion control flavor (default=1)");
 
-static void process_work(struct work_struct *work);
 static struct workqueue_struct *workq;
-static DECLARE_WORK(skb_work, process_work);
 
 static struct sk_buff_head rxq;
-static cxgb3_cpl_handler_func work_handlers[NUM_CPL_CMDS];
 
 static struct sk_buff *get_skb(struct sk_buff *skb, int len, gfp_t gfp);
 static void ep_timeout(unsigned long arg);
@@ -151,7 +148,7 @@ int iwch_l2t_send(struct t3cdev *tdev, struct sk_buff *skb, struct l2t_entry *l2
 		return -EIO;
 	}
 	error = l2t_send(tdev, skb, l2e);
-	if (error)
+	if (error < 0)
 		kfree_skb(skb);
 	return error;
 }
@@ -167,7 +164,7 @@ int iwch_cxgb3_ofld_send(struct t3cdev *tdev, struct sk_buff *skb)
 		return -EIO;
 	}
 	error = cxgb3_ofld_send(tdev, skb);
-	if (error)
+	if (error < 0)
 		kfree_skb(skb);
 	return error;
 }
@@ -302,27 +299,6 @@ static void release_ep_resources(struct iwch_ep *ep)
 	put_ep(&ep->com);
 }
 
-static void process_work(struct work_struct *work)
-{
-	struct sk_buff *skb = NULL;
-	void *ep;
-	struct t3cdev *tdev;
-	int ret;
-
-	while ((skb = skb_dequeue(&rxq))) {
-		ep = *((void **) (skb->cb));
-		tdev = *((struct t3cdev **) (skb->cb + sizeof(void *)));
-		ret = work_handlers[G_OPCODE(ntohl((__force __be32)skb->csum))](tdev, skb, ep);
-		if (ret & CPL_RET_BUF_DONE)
-			kfree_skb(skb);
-
-		/*
-		 * ep was referenced in sched(), and is freed here.
-		 */
-		put_ep((struct iwch_ep_common *)ep);
-	}
-}
-
 static int status2errno(int status)
 {
 	switch (status) {
@@ -2157,7 +2133,49 @@ int iwch_ep_redirect(void *ctx, struct dst_entry *old, struct dst_entry *new,
 
 /*
  * All the CM events are handled on a work queue to have a safe context.
+ * These are the real handlers that are called from the work queue.
  */
+static const cxgb3_cpl_handler_func work_handlers[NUM_CPL_CMDS] = {
+	[CPL_ACT_ESTABLISH]	= act_establish,
+	[CPL_ACT_OPEN_RPL]	= act_open_rpl,
+	[CPL_RX_DATA]		= rx_data,
+	[CPL_TX_DMA_ACK]	= tx_ack,
+	[CPL_ABORT_RPL_RSS]	= abort_rpl,
+	[CPL_ABORT_RPL]		= abort_rpl,
+	[CPL_PASS_OPEN_RPL]	= pass_open_rpl,
+	[CPL_CLOSE_LISTSRV_RPL]	= close_listsrv_rpl,
+	[CPL_PASS_ACCEPT_REQ]	= pass_accept_req,
+	[CPL_PASS_ESTABLISH]	= pass_establish,
+	[CPL_PEER_CLOSE]	= peer_close,
+	[CPL_ABORT_REQ_RSS]	= peer_abort,
+	[CPL_CLOSE_CON_RPL]	= close_con_rpl,
+	[CPL_RDMA_TERMINATE]	= terminate,
+	[CPL_RDMA_EC_STATUS]	= ec_status,
+};
+
+static void process_work(struct work_struct *work)
+{
+	struct sk_buff *skb = NULL;
+	void *ep;
+	struct t3cdev *tdev;
+	int ret;
+
+	while ((skb = skb_dequeue(&rxq))) {
+		ep = *((void **) (skb->cb));
+		tdev = *((struct t3cdev **) (skb->cb + sizeof(void *)));
+		ret = work_handlers[G_OPCODE(ntohl((__force __be32)skb->csum))](tdev, skb, ep);
+		if (ret & CPL_RET_BUF_DONE)
+			kfree_skb(skb);
+
+		/*
+		 * ep was referenced in sched(), and is freed here.
+		 */
+		put_ep((struct iwch_ep_common *)ep);
+	}
+}
+
+static DECLARE_WORK(skb_work, process_work);
+
 static int sched(struct t3cdev *tdev, struct sk_buff *skb, void *ctx)
 {
 	struct iwch_ep_common *epc = ctx;
@@ -2189,6 +2207,29 @@ static int set_tcb_rpl(struct t3cdev *tdev, struct sk_buff *skb, void *ctx)
 	return CPL_RET_BUF_DONE;
 }
 
+/*
+ * All upcalls from the T3 Core go to sched() to schedule the
+ * processing on a work queue.
+ */
+cxgb3_cpl_handler_func t3c_handlers[NUM_CPL_CMDS] = {
+	[CPL_ACT_ESTABLISH]	= sched,
+	[CPL_ACT_OPEN_RPL]	= sched,
+	[CPL_RX_DATA]		= sched,
+	[CPL_TX_DMA_ACK]	= sched,
+	[CPL_ABORT_RPL_RSS]	= sched,
+	[CPL_ABORT_RPL]		= sched,
+	[CPL_PASS_OPEN_RPL]	= sched,
+	[CPL_CLOSE_LISTSRV_RPL]	= sched,
+	[CPL_PASS_ACCEPT_REQ]	= sched,
+	[CPL_PASS_ESTABLISH]	= sched,
+	[CPL_PEER_CLOSE]	= sched,
+	[CPL_CLOSE_CON_RPL]	= sched,
+	[CPL_ABORT_REQ_RSS]	= sched,
+	[CPL_RDMA_TERMINATE]	= sched,
+	[CPL_RDMA_EC_STATUS]	= sched,
+	[CPL_SET_TCB_RPL]	= set_tcb_rpl,
+};
+
 int __init iwch_cm_init(void)
 {
 	skb_queue_head_init(&rxq);
@@ -2197,46 +2238,6 @@ int __init iwch_cm_init(void)
 	if (!workq)
 		return -ENOMEM;
 
-	/*
-	 * All upcalls from the T3 Core go to sched() to
-	 * schedule the processing on a work queue.
-	 */
-	t3c_handlers[CPL_ACT_ESTABLISH] = sched;
-	t3c_handlers[CPL_ACT_OPEN_RPL] = sched;
-	t3c_handlers[CPL_RX_DATA] = sched;
-	t3c_handlers[CPL_TX_DMA_ACK] = sched;
-	t3c_handlers[CPL_ABORT_RPL_RSS] = sched;
-	t3c_handlers[CPL_ABORT_RPL] = sched;
-	t3c_handlers[CPL_PASS_OPEN_RPL] = sched;
-	t3c_handlers[CPL_CLOSE_LISTSRV_RPL] = sched;
-	t3c_handlers[CPL_PASS_ACCEPT_REQ] = sched;
-	t3c_handlers[CPL_PASS_ESTABLISH] = sched;
-	t3c_handlers[CPL_PEER_CLOSE] = sched;
-	t3c_handlers[CPL_CLOSE_CON_RPL] = sched;
-	t3c_handlers[CPL_ABORT_REQ_RSS] = sched;
-	t3c_handlers[CPL_RDMA_TERMINATE] = sched;
-	t3c_handlers[CPL_RDMA_EC_STATUS] = sched;
-	t3c_handlers[CPL_SET_TCB_RPL] = set_tcb_rpl;
-
-	/*
-	 * These are the real handlers that are called from a
-	 * work queue.
-	 */
-	work_handlers[CPL_ACT_ESTABLISH] = act_establish;
-	work_handlers[CPL_ACT_OPEN_RPL] = act_open_rpl;
-	work_handlers[CPL_RX_DATA] = rx_data;
-	work_handlers[CPL_TX_DMA_ACK] = tx_ack;
-	work_handlers[CPL_ABORT_RPL_RSS] = abort_rpl;
-	work_handlers[CPL_ABORT_RPL] = abort_rpl;
-	work_handlers[CPL_PASS_OPEN_RPL] = pass_open_rpl;
-	work_handlers[CPL_CLOSE_LISTSRV_RPL] = close_listsrv_rpl;
-	work_handlers[CPL_PASS_ACCEPT_REQ] = pass_accept_req;
-	work_handlers[CPL_PASS_ESTABLISH] = pass_establish;
-	work_handlers[CPL_PEER_CLOSE] = peer_close;
-	work_handlers[CPL_ABORT_REQ_RSS] = peer_abort;
-	work_handlers[CPL_CLOSE_CON_RPL] = close_con_rpl;
-	work_handlers[CPL_RDMA_TERMINATE] = terminate;
-	work_handlers[CPL_RDMA_EC_STATUS] = ec_status;
 	return 0;
 }
 

drivers/infiniband/hw/cxgb4/Kconfig

+config INFINIBAND_CXGB4
+	tristate "Chelsio T4 RDMA Driver"
+	depends on CHELSIO_T4 && INET
+	select GENERIC_ALLOCATOR
+	---help---
+	  This is an iWARP/RDMA driver for the Chelsio T4 1GbE and
+	  10GbE adapters.
+
+	  For general information about Chelsio and our products, visit
+	  our website at <http://www.chelsio.com>.
+
+	  For customer support, please visit our customer support page at
+	  <http://www.chelsio.com/support.htm>.
+
+	  Please send feedback to <linux-bugs@chelsio.com>.
+
+	  To compile this driver as a module, choose M here: the module
+	  will be called iw_cxgb4.

drivers/infiniband/hw/cxgb4/Makefile

+EXTRA_CFLAGS += -Idrivers/net/cxgb4
+
+obj-$(CONFIG_INFINIBAND_CXGB4) += iw_cxgb4.o
+
+iw_cxgb4-y :=  device.o cm.o provider.o mem.o cq.o qp.o resource.o ev.o

drivers/infiniband/hw/cxgb4/ev.c

+/*
+ * Copyright (c) 2009-2010 Chelsio, Inc. 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
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+#include <linux/slab.h>
+#include <linux/mman.h>
+#include <net/sock.h>
+
+#include "iw_cxgb4.h"
+
+static void post_qp_event(struct c4iw_dev *dev, struct c4iw_cq *chp,
+			  struct c4iw_qp *qhp,
+			  struct t4_cqe *err_cqe,
+			  enum ib_event_type ib_event)
+{
+	struct ib_event event;
+	struct c4iw_qp_attributes attrs;
+
+	if ((qhp->attr.state == C4IW_QP_STATE_ERROR) ||
+	    (qhp->attr.state == C4IW_QP_STATE_TERMINATE)) {
+		PDBG("%s AE received after RTS - "
+		     "qp state %d qpid 0x%x status 0x%x\n", __func__,
+		     qhp->attr.state, qhp->wq.sq.qid, CQE_STATUS(err_cqe));
+		return;
+	}
+
+	printk(KERN_ERR MOD "AE qpid 0x%x opcode %d status 0x%x "
+	       "type %d wrid.hi 0x%x wrid.lo 0x%x\n",
+	       CQE_QPID(err_cqe), CQE_OPCODE(err_cqe),
+	       CQE_STATUS(err_cqe), CQE_TYPE(err_cqe),
+	       CQE_WRID_HI(err_cqe), CQE_WRID_LOW(err_cqe));
+
+	if (qhp->attr.state == C4IW_QP_STATE_RTS) {
+		attrs.next_state = C4IW_QP_STATE_TERMINATE;
+		c4iw_modify_qp(qhp->rhp, qhp, C4IW_QP_ATTR_NEXT_STATE,
+			       &attrs, 1);
+	}
+
+	event.event = ib_event;
+	event.device = chp->ibcq.device;
+	if (ib_event == IB_EVENT_CQ_ERR)
+		event.element.cq = &chp->ibcq;
+	else
+		event.element.qp = &qhp->ibqp;
+	if (qhp->ibqp.event_handler)
+		(*qhp->ibqp.event_handler)(&event, qhp->ibqp.qp_context);
+
+	(*chp->ibcq.comp_handler)(&chp->ibcq, chp->ibcq.cq_context);
+}
+
+void c4iw_ev_dispatch(struct c4iw_dev *dev, struct t4_cqe *err_cqe)
+{
+	struct c4iw_cq *chp;
+	struct c4iw_qp *qhp;
+	u32 cqid;
+
+	spin_lock(&dev->lock);
+	qhp = get_qhp(dev, CQE_QPID(err_cqe));
+	if (!qhp) {
+		printk(KERN_ERR MOD "BAD AE qpid 0x%x opcode %d "
+		       "status 0x%x type %d wrid.hi 0x%x wrid.lo 0x%x\n",
+		       CQE_QPID(err_cqe),
+		       CQE_OPCODE(err_cqe), CQE_STATUS(err_cqe),
+		       CQE_TYPE(err_cqe), CQE_WRID_HI(err_cqe),
+		       CQE_WRID_LOW(err_cqe));
+		spin_unlock(&dev->lock);
+		goto out;
+	}
+
+	if (SQ_TYPE(err_cqe))
+		cqid = qhp->attr.scq;
+	else
+		cqid = qhp->attr.rcq;
+	chp = get_chp(dev, cqid);
+	if (!chp) {
+		printk(KERN_ERR MOD "BAD AE cqid 0x%x qpid 0x%x opcode %d "
+		       "status 0x%x type %d wrid.hi 0x%x wrid.lo 0x%x\n",
+		       cqid, CQE_QPID(err_cqe),
+		       CQE_OPCODE(err_cqe), CQE_STATUS(err_cqe),
+		       CQE_TYPE(err_cqe), CQE_WRID_HI(err_cqe),
+		       CQE_WRID_LOW(err_cqe));
+		spin_unlock(&dev->lock);
+		goto out;
+	}
+
+	c4iw_qp_add_ref(&qhp->ibqp);
+	atomic_inc(&chp->refcnt);
+	spin_unlock(&dev->lock);
+
+	/* Bad incoming write */
+	if (RQ_TYPE(err_cqe) &&
+	    (CQE_OPCODE(err_cqe) == FW_RI_RDMA_WRITE)) {
+		post_qp_event(dev, chp, qhp, err_cqe, IB_EVENT_QP_REQ_ERR);
+		goto done;
+	}
+
+	switch (CQE_STATUS(err_cqe)) {
+
+	/* Completion Events */
+	case T4_ERR_SUCCESS:
+		printk(KERN_ERR MOD "AE with status 0!\n");
+		break;
+
+	case T4_ERR_STAG:
+	case T4_ERR_PDID:
+	case T4_ERR_QPID:
+	case T4_ERR_ACCESS:
+	case T4_ERR_WRAP:
+	case T4_ERR_BOUND:
+	case T4_ERR_INVALIDATE_SHARED_MR:
+	case T4_ERR_INVALIDATE_MR_WITH_MW_BOUND:
+		post_qp_event(dev, chp, qhp, err_cqe, IB_EVENT_QP_ACCESS_ERR);
+		break;
+
+	/* Device Fatal Errors */
+	case T4_ERR_ECC:
+	case T4_ERR_ECC_PSTAG:
+	case T4_ERR_INTERNAL_ERR:
+		post_qp_event(dev, chp, qhp, err_cqe, IB_EVENT_DEVICE_FATAL);
+		break;
+
+	/* QP Fatal Errors */
+	case T4_ERR_OUT_OF_RQE:
+	case T4_ERR_PBL_ADDR_BOUND:
+	case T4_ERR_CRC:
+	case T4_ERR_MARKER:
+	case T4_ERR_PDU_LEN_ERR:
+	case T4_ERR_DDP_VERSION:
+	case T4_ERR_RDMA_VERSION:
+	case T4_ERR_OPCODE:
+	case T4_ERR_DDP_QUEUE_NUM:
+	case T4_ERR_MSN:
+	case T4_ERR_TBIT:
+	case T4_ERR_MO:
+	case T4_ERR_MSN_GAP:
+	case T4_ERR_MSN_RANGE:
+	case T4_ERR_RQE_ADDR_BOUND:
+	case T4_ERR_IRD_OVERFLOW:
+		post_qp_event(dev, chp, qhp, err_cqe, IB_EVENT_QP_FATAL);
+		break;
+
+	default:
+		printk(KERN_ERR MOD "Unknown T4 status 0x%x QPID 0x%x\n",
+		       CQE_STATUS(err_cqe), qhp->wq.sq.qid);
+		post_qp_event(dev, chp, qhp, err_cqe, IB_EVENT_QP_FATAL);
+		break;
+	}
+done:
+	if (atomic_dec_and_test(&chp->refcnt))
+		wake_up(&chp->wait);
+	c4iw_qp_rem_ref(&qhp->ibqp);
+out:
+	return;
+}
+
+int c4iw_ev_handler(struct c4iw_dev *dev, u32 qid)
+{
+	struct c4iw_cq *chp;
+
+	chp = get_chp(dev, qid);
+	if (chp)
+		(*chp->ibcq.comp_handler)(&chp->ibcq, chp->ibcq.cq_context);
+	else
+		PDBG("%s unknown cqid 0x%x\n", __func__, qid);
+	return 0;
+}

drivers/infiniband/hw/cxgb4/user.h

+/*
+ * Copyright (c) 2009-2010 Chelsio, Inc. 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
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+#ifndef __C4IW_USER_H__
+#define __C4IW_USER_H__
+
+#define C4IW_UVERBS_ABI_VERSION	1
+
+/*
+ * Make sure that all structs defined in this file remain laid out so
+ * that they pack the same way on 32-bit and 64-bit architectures (to
+ * avoid incompatibility between 32-bit userspace and 64-bit kernels).
+ * In particular do not use pointer types -- pass pointers in __u64
+ * instead.
+ */
+struct c4iw_create_cq_resp {
+	__u64 key;
+	__u64 gts_key;
+	__u64 memsize;
+	__u32 cqid;
+	__u32 size;
+	__u32 qid_mask;
+};
+
+struct c4iw_create_qp_resp {
+	__u64 sq_key;
+	__u64 rq_key;
+	__u64 sq_db_gts_key;
+	__u64 rq_db_gts_key;
+	__u64 sq_memsize;
+	__u64 rq_memsize;
+	__u32 sqid;
+	__u32 rqid;
+	__u32 sq_size;
+	__u32 rq_size;
+	__u32 qid_mask;
+};
+#endif

drivers/infiniband/hw/mlx4/cq.c

@@ -661,6 +661,14 @@ static int mlx4_ib_poll_one(struct mlx4_ib_cq *cq,
 			wc->opcode    = IB_WC_FETCH_ADD;
 			wc->byte_len  = 8;
 			break;
+		case MLX4_OPCODE_MASKED_ATOMIC_CS:
+			wc->opcode    = IB_WC_MASKED_COMP_SWAP;
+			wc->byte_len  = 8;
+			break;
+		case MLX4_OPCODE_MASKED_ATOMIC_FA:
+			wc->opcode    = IB_WC_MASKED_FETCH_ADD;
+			wc->byte_len  = 8;
+			break;
 		case MLX4_OPCODE_BIND_MW:
 			wc->opcode    = IB_WC_BIND_MW;
 			break;

drivers/infiniband/hw/mlx4/main.c

@@ -139,6 +139,7 @@ static int mlx4_ib_query_device(struct ib_device *ibdev,
 	props->local_ca_ack_delay  = dev->dev->caps.local_ca_ack_delay;
 	props->atomic_cap	   = dev->dev->caps.flags & MLX4_DEV_CAP_FLAG_ATOMIC ?
 		IB_ATOMIC_HCA : IB_ATOMIC_NONE;
+	props->masked_atomic_cap   = IB_ATOMIC_HCA;
 	props->max_pkeys	   = dev->dev->caps.pkey_table_len[1];
 	props->max_mcast_grp	   = dev->dev->caps.num_mgms + dev->dev->caps.num_amgms;
 	props->max_mcast_qp_attach = dev->dev->caps.num_qp_per_mgm;

drivers/infiniband/hw/mlx4/qp.c

@@ -74,17 +74,19 @@ enum {
 };
 
 static const __be32 mlx4_ib_opcode[] = {
-	[IB_WR_SEND]			= cpu_to_be32(MLX4_OPCODE_SEND),
-	[IB_WR_LSO]			= cpu_to_be32(MLX4_OPCODE_LSO),
-	[IB_WR_SEND_WITH_IMM]		= cpu_to_be32(MLX4_OPCODE_SEND_IMM),
-	[IB_WR_RDMA_WRITE]		= cpu_to_be32(MLX4_OPCODE_RDMA_WRITE),
-	[IB_WR_RDMA_WRITE_WITH_IMM]	= cpu_to_be32(MLX4_OPCODE_RDMA_WRITE_IMM),
-	[IB_WR_RDMA_READ]		= cpu_to_be32(MLX4_OPCODE_RDMA_READ),
-	[IB_WR_ATOMIC_CMP_AND_SWP]	= cpu_to_be32(MLX4_OPCODE_ATOMIC_CS),
-	[IB_WR_ATOMIC_FETCH_AND_ADD]	= cpu_to_be32(MLX4_OPCODE_ATOMIC_FA),
-	[IB_WR_SEND_WITH_INV]		= cpu_to_be32(MLX4_OPCODE_SEND_INVAL),
-	[IB_WR_LOCAL_INV]		= cpu_to_be32(MLX4_OPCODE_LOCAL_INVAL),
-	[IB_WR_FAST_REG_MR]		= cpu_to_be32(MLX4_OPCODE_FMR),
+	[IB_WR_SEND]				= cpu_to_be32(MLX4_OPCODE_SEND),
+	[IB_WR_LSO]				= cpu_to_be32(MLX4_OPCODE_LSO),
+	[IB_WR_SEND_WITH_IMM]			= cpu_to_be32(MLX4_OPCODE_SEND_IMM),
+	[IB_WR_RDMA_WRITE]			= cpu_to_be32(MLX4_OPCODE_RDMA_WRITE),
+	[IB_WR_RDMA_WRITE_WITH_IMM]		= cpu_to_be32(MLX4_OPCODE_RDMA_WRITE_IMM),
+	[IB_WR_RDMA_READ]			= cpu_to_be32(MLX4_OPCODE_RDMA_READ),
+	[IB_WR_ATOMIC_CMP_AND_SWP]		= cpu_to_be32(MLX4_OPCODE_ATOMIC_CS),
+	[IB_WR_ATOMIC_FETCH_AND_ADD]		= cpu_to_be32(MLX4_OPCODE_ATOMIC_FA),
+	[IB_WR_SEND_WITH_INV]			= cpu_to_be32(MLX4_OPCODE_SEND_INVAL),
+	[IB_WR_LOCAL_INV]			= cpu_to_be32(MLX4_OPCODE_LOCAL_INVAL),
+	[IB_WR_FAST_REG_MR]			= cpu_to_be32(MLX4_OPCODE_FMR),
+	[IB_WR_MASKED_ATOMIC_CMP_AND_SWP]	= cpu_to_be32(MLX4_OPCODE_MASKED_ATOMIC_CS),
+	[IB_WR_MASKED_ATOMIC_FETCH_AND_ADD]	= cpu_to_be32(MLX4_OPCODE_MASKED_ATOMIC_FA),
 };
 
 static struct mlx4_ib_sqp *to_msqp(struct mlx4_ib_qp *mqp)
@@ -1407,6 +1409,9 @@ static void set_atomic_seg(struct mlx4_wqe_atomic_seg *aseg, struct ib_send_wr *
 	if (wr->opcode == IB_WR_ATOMIC_CMP_AND_SWP) {
 		aseg->swap_add = cpu_to_be64(wr->wr.atomic.swap);
 		aseg->compare  = cpu_to_be64(wr->wr.atomic.compare_add);
+	} else if (wr->opcode == IB_WR_MASKED_ATOMIC_FETCH_AND_ADD) {
+		aseg->swap_add = cpu_to_be64(wr->wr.atomic.compare_add);
+		aseg->compare  = cpu_to_be64(wr->wr.atomic.compare_add_mask);
 	} else {
 		aseg->swap_add = cpu_to_be64(wr->wr.atomic.compare_add);
 		aseg->compare  = 0;
@@ -1414,6 +1419,15 @@ static void set_atomic_seg(struct mlx4_wqe_atomic_seg *aseg, struct ib_send_wr *
 
 }
 
+static void set_masked_atomic_seg(struct mlx4_wqe_masked_atomic_seg *aseg,
+				  struct ib_send_wr *wr)
+{
+	aseg->swap_add		= cpu_to_be64(wr->wr.atomic.swap);
+	aseg->swap_add_mask	= cpu_to_be64(wr->wr.atomic.swap_mask);
+	aseg->compare		= cpu_to_be64(wr->wr.atomic.compare_add);
+	aseg->compare_mask	= cpu_to_be64(wr->wr.atomic.compare_add_mask);
+}
+
 static void set_datagram_seg(struct mlx4_wqe_datagram_seg *dseg,
 			     struct ib_send_wr *wr)
 {
@@ -1567,6 +1581,7 @@ int mlx4_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 			switch (wr->opcode) {
 			case IB_WR_ATOMIC_CMP_AND_SWP:
 			case IB_WR_ATOMIC_FETCH_AND_ADD:
+			case IB_WR_MASKED_ATOMIC_FETCH_AND_ADD:
 				set_raddr_seg(wqe, wr->wr.atomic.remote_addr,
 					      wr->wr.atomic.rkey);
 				wqe  += sizeof (struct mlx4_wqe_raddr_seg);
@@ -1579,6 +1594,19 @@ int mlx4_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 
 				break;
 
+			case IB_WR_MASKED_ATOMIC_CMP_AND_SWP:
+				set_raddr_seg(wqe, wr->wr.atomic.remote_addr,
+					      wr->wr.atomic.rkey);
+				wqe  += sizeof (struct mlx4_wqe_raddr_seg);
+
+				set_masked_atomic_seg(wqe, wr);
+				wqe  += sizeof (struct mlx4_wqe_masked_atomic_seg);
+
+				size += (sizeof (struct mlx4_wqe_raddr_seg) +
+					 sizeof (struct mlx4_wqe_masked_atomic_seg)) / 16;
+
+				break;
+
 			case IB_WR_RDMA_READ:
 			case IB_WR_RDMA_WRITE:
 			case IB_WR_RDMA_WRITE_WITH_IMM:

drivers/infiniband/hw/mthca/mthca_allocator.c

@@ -211,7 +211,7 @@ int mthca_buf_alloc(struct mthca_dev *dev, int size, int max_direct,
 		if (!buf->direct.buf)
 			return -ENOMEM;
 
-		pci_unmap_addr_set(&buf->direct, mapping, t);
+		dma_unmap_addr_set(&buf->direct, mapping, t);
 
 		memset(buf->direct.buf, 0, size);
 
@@ -251,7 +251,7 @@ int mthca_buf_alloc(struct mthca_dev *dev, int size, int max_direct,
 				goto err_free;
 
 			dma_list[i] = t;
-			pci_unmap_addr_set(&buf->page_list[i], mapping, t);
+			dma_unmap_addr_set(&buf->page_list[i], mapping, t);
 
 			clear_page(buf->page_list[i].buf);
 		}
@@ -289,12 +289,12 @@ void mthca_buf_free(struct mthca_dev *dev, int size, union mthca_buf *buf,
 
 	if (is_direct)
 		dma_free_coherent(&dev->pdev->dev, size, buf->direct.buf,
-				  pci_unmap_addr(&buf->direct, mapping));
+				  dma_unmap_addr(&buf->direct, mapping));
 	else {
 		for (i = 0; i < (size + PAGE_SIZE - 1) / PAGE_SIZE; ++i)
 			dma_free_coherent(&dev->pdev->dev, PAGE_SIZE,
 					  buf->page_list[i].buf,
-					  pci_unmap_addr(&buf->page_list[i],
+					  dma_unmap_addr(&buf->page_list[i],
 							 mapping));
 		kfree(buf->page_list);
 	}

drivers/infiniband/hw/mthca/mthca_eq.c

@@ -504,7 +504,7 @@ static int mthca_create_eq(struct mthca_dev *dev,
 			goto err_out_free_pages;
 
 		dma_list[i] = t;
-		pci_unmap_addr_set(&eq->page_list[i], mapping, t);
+		dma_unmap_addr_set(&eq->page_list[i], mapping, t);
 
 		clear_page(eq->page_list[i].buf);
 	}
@@ -579,7 +579,7 @@ static int mthca_create_eq(struct mthca_dev *dev,
 		if (eq->page_list[i].buf)
 			dma_free_coherent(&dev->pdev->dev, PAGE_SIZE,
 					  eq->page_list[i].buf,
-					  pci_unmap_addr(&eq->page_list[i],
+					  dma_unmap_addr(&eq->page_list[i],
 							 mapping));
 
 	mthca_free_mailbox(dev, mailbox);
@@ -629,7 +629,7 @@ static void mthca_free_eq(struct mthca_dev *dev,
 	for (i = 0; i < npages; ++i)
 		pci_free_consistent(dev->pdev, PAGE_SIZE,
 				    eq->page_list[i].buf,
-				    pci_unmap_addr(&eq->page_list[i], mapping));
+				    dma_unmap_addr(&eq->page_list[i], mapping));
 
 	kfree(eq->page_list);
 	mthca_free_mailbox(dev, mailbox);

drivers/infiniband/hw/mthca/mthca_provider.h

@@ -46,7 +46,7 @@
 
 struct mthca_buf_list {
 	void *buf;
-	DECLARE_PCI_UNMAP_ADDR(mapping)
+	DEFINE_DMA_UNMAP_ADDR(mapping);
 };
 
 union mthca_buf {

drivers/infiniband/hw/nes/nes_hw.c

@@ -1297,7 +1297,7 @@ int nes_destroy_cqp(struct nes_device *nesdev)
 /**
  * nes_init_1g_phy
  */
-int nes_init_1g_phy(struct nes_device *nesdev, u8 phy_type, u8 phy_index)
+static int nes_init_1g_phy(struct nes_device *nesdev, u8 phy_type, u8 phy_index)
 {
 	u32 counter = 0;
 	u16 phy_data;
@@ -1351,7 +1351,7 @@ int nes_init_1g_phy(struct nes_device *nesdev, u8 phy_type, u8 phy_index)
 /**
  * nes_init_2025_phy
  */
-int nes_init_2025_phy(struct nes_device *nesdev, u8 phy_type, u8 phy_index)
+static int nes_init_2025_phy(struct nes_device *nesdev, u8 phy_type, u8 phy_index)
 {
 	u32 temp_phy_data = 0;
 	u32 temp_phy_data2 = 0;
@@ -2458,7 +2458,6 @@ static void nes_process_mac_intr(struct nes_device *nesdev, u32 mac_number)
 		return;
 	}
 	nesadapter->mac_sw_state[mac_number] = NES_MAC_SW_INTERRUPT;
-	spin_unlock_irqrestore(&nesadapter->phy_lock, flags);
 
 	/* ack the MAC interrupt */
 	mac_status = nes_read_indexed(nesdev, NES_IDX_MAC_INT_STATUS + (mac_index * 0x200));
@@ -2469,11 +2468,9 @@ static void nes_process_mac_intr(struct nes_device *nesdev, u32 mac_number)
 
 	if (mac_status & (NES_MAC_INT_LINK_STAT_CHG | NES_MAC_INT_XGMII_EXT)) {
 		nesdev->link_status_interrupts++;
-		if (0 == (++nesadapter->link_interrupt_count[mac_index] % ((u16)NES_MAX_LINK_INTERRUPTS))) {
-			spin_lock_irqsave(&nesadapter->phy_lock, flags);
+		if (0 == (++nesadapter->link_interrupt_count[mac_index] % ((u16)NES_MAX_LINK_INTERRUPTS)))
 			nes_reset_link(nesdev, mac_index);
-			spin_unlock_irqrestore(&nesadapter->phy_lock, flags);
-		}
+
 		/* read the PHY interrupt status register */
 		if ((nesadapter->OneG_Mode) &&
 		(nesadapter->phy_type[mac_index] != NES_PHY_TYPE_PUMA_1G)) {
@@ -2587,6 +2584,7 @@ static void nes_process_mac_intr(struct nes_device *nesdev, u32 mac_number)
 				break;
 			}
 		}
+		spin_unlock_irqrestore(&nesadapter->phy_lock, flags);
 
 		if (phy_data & 0x0004) {
 			if (wide_ppm_offset &&

drivers/infiniband/hw/nes/nes_nic.c

@@ -1461,11 +1461,14 @@ static int nes_netdev_get_settings(struct net_device *netdev, struct ethtool_cmd
 			et_cmd->transceiver = XCVR_INTERNAL;
 			et_cmd->phy_address = mac_index;
 		} else {
+			unsigned long flags;
 			et_cmd->supported   = SUPPORTED_1000baseT_Full
 					    | SUPPORTED_Autoneg;
 			et_cmd->advertising = ADVERTISED_1000baseT_Full
 					    | ADVERTISED_Autoneg;
+			spin_lock_irqsave(&nesadapter->phy_lock, flags);
 			nes_read_1G_phy_reg(nesdev, 0, phy_index, &phy_data);
+			spin_unlock_irqrestore(&nesadapter->phy_lock, flags);
 			if (phy_data & 0x1000)
 				et_cmd->autoneg = AUTONEG_ENABLE;
 			else
@@ -1503,12 +1506,15 @@ static int nes_netdev_set_settings(struct net_device *netdev, struct ethtool_cmd
 	struct nes_vnic *nesvnic = netdev_priv(netdev);
 	struct nes_device *nesdev = nesvnic->nesdev;
 	struct nes_adapter *nesadapter = nesdev->nesadapter;
-	u16 phy_data;
 
 	if ((nesadapter->OneG_Mode) &&
 	    (nesadapter->phy_type[nesdev->mac_index] != NES_PHY_TYPE_PUMA_1G)) {
-		nes_read_1G_phy_reg(nesdev, 0, nesadapter->phy_index[nesdev->mac_index],
-				&phy_data);
+		unsigned long flags;
+		u16 phy_data;
+		u8 phy_index = nesadapter->phy_index[nesdev->mac_index];
+
+		spin_lock_irqsave(&nesadapter->phy_lock, flags);
+		nes_read_1G_phy_reg(nesdev, 0, phy_index, &phy_data);
 		if (et_cmd->autoneg) {
 			/* Turn on Full duplex, Autoneg, and restart autonegotiation */
 			phy_data |= 0x1300;
@@ -1516,8 +1522,8 @@ static int nes_netdev_set_settings(struct net_device *netdev, struct ethtool_cmd
 			/* Turn off autoneg */
 			phy_data &= ~0x1000;
 		}
-		nes_write_1G_phy_reg(nesdev, 0, nesadapter->phy_index[nesdev->mac_index],
-				phy_data);
+		nes_write_1G_phy_reg(nesdev, 0, phy_index, phy_data);
+		spin_unlock_irqrestore(&nesadapter->phy_lock, flags);
 	}
 
 	return 0;

drivers/infiniband/hw/nes/nes_utils.c

@@ -381,12 +381,8 @@ static u16 nes_read16_eeprom(void __iomem *addr, u16 offset)
  */
 void nes_write_1G_phy_reg(struct nes_device *nesdev, u8 phy_reg, u8 phy_addr, u16 data)
 {
-	struct nes_adapter *nesadapter = nesdev->nesadapter;
 	u32 u32temp;
 	u32 counter;
-	unsigned long flags;
-
-	spin_lock_irqsave(&nesadapter->phy_lock, flags);
 
 	nes_write_indexed(nesdev, NES_IDX_MAC_MDIO_CONTROL,
 			0x50020000 | data | ((u32)phy_reg << 18) | ((u32)phy_addr << 23));
@@ -402,8 +398,6 @@ void nes_write_1G_phy_reg(struct nes_device *nesdev, u8 phy_reg, u8 phy_addr, u1
 	if (!(u32temp & 1))
 		nes_debug(NES_DBG_PHY, "Phy is not responding. interrupt status = 0x%X.\n",
 				u32temp);
-
-	spin_unlock_irqrestore(&nesadapter->phy_lock, flags);
 }
 
 
@@ -414,14 +408,11 @@ void nes_write_1G_phy_reg(struct nes_device *nesdev, u8 phy_reg, u8 phy_addr, u1
  */
 void nes_read_1G_phy_reg(struct nes_device *nesdev, u8 phy_reg, u8 phy_addr, u16 *data)
 {
-	struct nes_adapter *nesadapter = nesdev->nesadapter;
 	u32 u32temp;
 	u32 counter;
-	unsigned long flags;
 
 	/* nes_debug(NES_DBG_PHY, "phy addr = %d, mac_index = %d\n",
 			phy_addr, nesdev->mac_index); */
-	spin_lock_irqsave(&nesadapter->phy_lock, flags);
 
 	nes_write_indexed(nesdev, NES_IDX_MAC_MDIO_CONTROL,
 			0x60020000 | ((u32)phy_reg << 18) | ((u32)phy_addr << 23));
@@ -441,7 +432,6 @@ void nes_read_1G_phy_reg(struct nes_device *nesdev, u8 phy_reg, u8 phy_addr, u16
 	} else {
 		*data = (u16)nes_read_indexed(nesdev, NES_IDX_MAC_MDIO_CONTROL);
 	}
-	spin_unlock_irqrestore(&nesadapter->phy_lock, flags);
 }
 
 

drivers/infiniband/hw/nes/nes_verbs.c

@@ -374,7 +374,7 @@ static int alloc_fast_reg_mr(struct nes_device *nesdev, struct nes_pd *nespd,
 /*
  * nes_alloc_fast_reg_mr
  */
-struct ib_mr *nes_alloc_fast_reg_mr(struct ib_pd *ibpd, int max_page_list_len)
+static struct ib_mr *nes_alloc_fast_reg_mr(struct ib_pd *ibpd, int max_page_list_len)
 {
 	struct nes_pd *nespd = to_nespd(ibpd);
 	struct nes_vnic *nesvnic = to_nesvnic(ibpd->device);

drivers/infiniband/ulp/iser/iscsi_iser.h

@@ -232,6 +232,7 @@ struct iser_device {
 	struct ib_cq	             *tx_cq;
 	struct ib_mr	             *mr;
 	struct tasklet_struct	     cq_tasklet;
+	struct ib_event_handler      event_handler;
 	struct list_head             ig_list; /* entry in ig devices list */
 	int                          refcount;
 };
@@ -246,7 +247,6 @@ struct iser_conn {
 	struct rdma_cm_id            *cma_id;       /* CMA ID		       */
 	struct ib_qp	             *qp;           /* QP 		       */
 	struct ib_fmr_pool           *fmr_pool;     /* pool of IB FMRs         */
-	int                          disc_evt_flag; /* disconn event delivered */
 	wait_queue_head_t	     wait;          /* waitq for conn/disconn  */
 	int                          post_recv_buf_count; /* posted rx count  */
 	atomic_t                     post_send_buf_count; /* posted tx count   */
@@ -320,7 +320,7 @@ void iser_conn_init(struct iser_conn *ib_conn);
 
 void iser_conn_get(struct iser_conn *ib_conn);
 
-void iser_conn_put(struct iser_conn *ib_conn);
+int iser_conn_put(struct iser_conn *ib_conn, int destroy_cma_id_allowed);
 
 void iser_conn_terminate(struct iser_conn *ib_conn);