add idl4k kernel firmware version 1.13.0.105

kernel/drivers/net/mlx4/Makefile

@@ -0,0 +1,9 @@
+obj-$(CONFIG_MLX4_CORE)		+= mlx4_core.o
+
+mlx4_core-y :=	alloc.o catas.o cmd.o cq.o eq.o fw.o icm.o intf.o main.o mcg.o \
+		mr.o pd.o port.o profile.o qp.o reset.o sense.o srq.o
+
+obj-$(CONFIG_MLX4_EN)               += mlx4_en.o
+
+mlx4_en-y := 	en_main.o en_tx.o en_rx.o en_ethtool.o en_port.o en_cq.o \
+		en_resources.o en_netdev.o

kernel/drivers/net/mlx4/alloc.c

@@ -0,0 +1,436 @@
+/*
+ * Copyright (c) 2006, 2007 Cisco Systems, Inc.  All rights reserved.
+ * Copyright (c) 2007, 2008 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
+ * 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/errno.h>
+#include <linux/slab.h>
+#include <linux/mm.h>
+#include <linux/bitmap.h>
+#include <linux/dma-mapping.h>
+#include <linux/vmalloc.h>
+
+#include "mlx4.h"
+
+u32 mlx4_bitmap_alloc(struct mlx4_bitmap *bitmap)
+{
+	u32 obj;
+
+	spin_lock(&bitmap->lock);
+
+	obj = find_next_zero_bit(bitmap->table, bitmap->max, bitmap->last);
+	if (obj >= bitmap->max) {
+		bitmap->top = (bitmap->top + bitmap->max + bitmap->reserved_top)
+				& bitmap->mask;
+		obj = find_first_zero_bit(bitmap->table, bitmap->max);
+	}
+
+	if (obj < bitmap->max) {
+		set_bit(obj, bitmap->table);
+		bitmap->last = (obj + 1);
+		if (bitmap->last == bitmap->max)
+			bitmap->last = 0;
+		obj |= bitmap->top;
+	} else
+		obj = -1;
+
+	spin_unlock(&bitmap->lock);
+
+	return obj;
+}
+
+void mlx4_bitmap_free(struct mlx4_bitmap *bitmap, u32 obj)
+{
+	mlx4_bitmap_free_range(bitmap, obj, 1);
+}
+
+static unsigned long find_aligned_range(unsigned long *bitmap,
+					u32 start, u32 nbits,
+					int len, int align)
+{
+	unsigned long end, i;
+
+again:
+	start = ALIGN(start, align);
+
+	while ((start < nbits) && test_bit(start, bitmap))
+		start += align;
+
+	if (start >= nbits)
+		return -1;
+
+	end = start+len;
+	if (end > nbits)
+		return -1;
+
+	for (i = start + 1; i < end; i++) {
+		if (test_bit(i, bitmap)) {
+			start = i + 1;
+			goto again;
+		}
+	}
+
+	return start;
+}
+
+u32 mlx4_bitmap_alloc_range(struct mlx4_bitmap *bitmap, int cnt, int align)
+{
+	u32 obj, i;
+
+	if (likely(cnt == 1 && align == 1))
+		return mlx4_bitmap_alloc(bitmap);
+
+	spin_lock(&bitmap->lock);
+
+	obj = find_aligned_range(bitmap->table, bitmap->last,
+				 bitmap->max, cnt, align);
+	if (obj >= bitmap->max) {
+		bitmap->top = (bitmap->top + bitmap->max + bitmap->reserved_top)
+				& bitmap->mask;
+		obj = find_aligned_range(bitmap->table, 0, bitmap->max,
+					 cnt, align);
+	}
+
+	if (obj < bitmap->max) {
+		for (i = 0; i < cnt; i++)
+			set_bit(obj + i, bitmap->table);
+		if (obj == bitmap->last) {
+			bitmap->last = (obj + cnt);
+			if (bitmap->last >= bitmap->max)
+				bitmap->last = 0;
+		}
+		obj |= bitmap->top;
+	} else
+		obj = -1;
+
+	spin_unlock(&bitmap->lock);
+
+	return obj;
+}
+
+void mlx4_bitmap_free_range(struct mlx4_bitmap *bitmap, u32 obj, int cnt)
+{
+	u32 i;
+
+	obj &= bitmap->max + bitmap->reserved_top - 1;
+
+	spin_lock(&bitmap->lock);
+	for (i = 0; i < cnt; i++)
+		clear_bit(obj + i, bitmap->table);
+	bitmap->last = min(bitmap->last, obj);
+	bitmap->top = (bitmap->top + bitmap->max + bitmap->reserved_top)
+			& bitmap->mask;
+	spin_unlock(&bitmap->lock);
+}
+
+int mlx4_bitmap_init(struct mlx4_bitmap *bitmap, u32 num, u32 mask,
+		     u32 reserved_bot, u32 reserved_top)
+{
+	int i;
+
+	/* num must be a power of 2 */
+	if (num != roundup_pow_of_two(num))
+		return -EINVAL;
+
+	bitmap->last = 0;
+	bitmap->top  = 0;
+	bitmap->max  = num - reserved_top;
+	bitmap->mask = mask;
+	bitmap->reserved_top = reserved_top;
+	spin_lock_init(&bitmap->lock);
+	bitmap->table = kzalloc(BITS_TO_LONGS(bitmap->max) *
+				sizeof (long), GFP_KERNEL);
+	if (!bitmap->table)
+		return -ENOMEM;
+
+	for (i = 0; i < reserved_bot; ++i)
+		set_bit(i, bitmap->table);
+
+	return 0;
+}
+
+void mlx4_bitmap_cleanup(struct mlx4_bitmap *bitmap)
+{
+	kfree(bitmap->table);
+}
+
+/*
+ * Handling for queue buffers -- we allocate a bunch of memory and
+ * register it in a memory region at HCA virtual address 0.  If the
+ * requested size is > max_direct, we split the allocation into
+ * multiple pages, so we don't require too much contiguous memory.
+ */
+
+int mlx4_buf_alloc(struct mlx4_dev *dev, int size, int max_direct,
+		   struct mlx4_buf *buf)
+{
+	dma_addr_t t;
+
+	if (size <= max_direct) {
+		buf->nbufs        = 1;
+		buf->npages       = 1;
+		buf->page_shift   = get_order(size) + PAGE_SHIFT;
+		buf->direct.buf   = dma_alloc_coherent(&dev->pdev->dev,
+						       size, &t, GFP_KERNEL);
+		if (!buf->direct.buf)
+			return -ENOMEM;
+
+		buf->direct.map = t;
+
+		while (t & ((1 << buf->page_shift) - 1)) {
+			--buf->page_shift;
+			buf->npages *= 2;
+		}
+
+		memset(buf->direct.buf, 0, size);
+	} else {
+		int i;
+
+		buf->nbufs       = (size + PAGE_SIZE - 1) / PAGE_SIZE;
+		buf->npages      = buf->nbufs;
+		buf->page_shift  = PAGE_SHIFT;
+		buf->page_list   = kzalloc(buf->nbufs * sizeof *buf->page_list,
+					   GFP_KERNEL);
+		if (!buf->page_list)
+			return -ENOMEM;
+
+		for (i = 0; i < buf->nbufs; ++i) {
+			buf->page_list[i].buf =
+				dma_alloc_coherent(&dev->pdev->dev, PAGE_SIZE,
+						   &t, GFP_KERNEL);
+			if (!buf->page_list[i].buf)
+				goto err_free;
+
+			buf->page_list[i].map = t;
+
+			memset(buf->page_list[i].buf, 0, PAGE_SIZE);
+		}
+
+		if (BITS_PER_LONG == 64) {
+			struct page **pages;
+			pages = kmalloc(sizeof *pages * buf->nbufs, GFP_KERNEL);
+			if (!pages)
+				goto err_free;
+			for (i = 0; i < buf->nbufs; ++i)
+				pages[i] = virt_to_page(buf->page_list[i].buf);
+			buf->direct.buf = vmap(pages, buf->nbufs, VM_MAP, PAGE_KERNEL);
+			kfree(pages);
+			if (!buf->direct.buf)
+				goto err_free;
+		}
+	}
+
+	return 0;
+
+err_free:
+	mlx4_buf_free(dev, size, buf);
+
+	return -ENOMEM;
+}
+EXPORT_SYMBOL_GPL(mlx4_buf_alloc);
+
+void mlx4_buf_free(struct mlx4_dev *dev, int size, struct mlx4_buf *buf)
+{
+	int i;
+
+	if (buf->nbufs == 1)
+		dma_free_coherent(&dev->pdev->dev, size, buf->direct.buf,
+				  buf->direct.map);
+	else {
+		if (BITS_PER_LONG == 64)
+			vunmap(buf->direct.buf);
+
+		for (i = 0; i < buf->nbufs; ++i)
+			if (buf->page_list[i].buf)
+				dma_free_coherent(&dev->pdev->dev, PAGE_SIZE,
+						  buf->page_list[i].buf,
+						  buf->page_list[i].map);
+		kfree(buf->page_list);
+	}
+}
+EXPORT_SYMBOL_GPL(mlx4_buf_free);
+
+static struct mlx4_db_pgdir *mlx4_alloc_db_pgdir(struct device *dma_device)
+{
+	struct mlx4_db_pgdir *pgdir;
+
+	pgdir = kzalloc(sizeof *pgdir, GFP_KERNEL);
+	if (!pgdir)
+		return NULL;
+
+	bitmap_fill(pgdir->order1, MLX4_DB_PER_PAGE / 2);
+	pgdir->bits[0] = pgdir->order0;
+	pgdir->bits[1] = pgdir->order1;
+	pgdir->db_page = dma_alloc_coherent(dma_device, PAGE_SIZE,
+					    &pgdir->db_dma, GFP_KERNEL);
+	if (!pgdir->db_page) {
+		kfree(pgdir);
+		return NULL;
+	}
+
+	return pgdir;
+}
+
+static int mlx4_alloc_db_from_pgdir(struct mlx4_db_pgdir *pgdir,
+				    struct mlx4_db *db, int order)
+{
+	int o;
+	int i;
+
+	for (o = order; o <= 1; ++o) {
+		i = find_first_bit(pgdir->bits[o], MLX4_DB_PER_PAGE >> o);
+		if (i < MLX4_DB_PER_PAGE >> o)
+			goto found;
+	}
+
+	return -ENOMEM;
+
+found:
+	clear_bit(i, pgdir->bits[o]);
+
+	i <<= o;
+
+	if (o > order)
+		set_bit(i ^ 1, pgdir->bits[order]);
+
+	db->u.pgdir = pgdir;
+	db->index   = i;
+	db->db      = pgdir->db_page + db->index;
+	db->dma     = pgdir->db_dma  + db->index * 4;
+	db->order   = order;
+
+	return 0;
+}
+
+int mlx4_db_alloc(struct mlx4_dev *dev, struct mlx4_db *db, int order)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_db_pgdir *pgdir;
+	int ret = 0;
+
+	mutex_lock(&priv->pgdir_mutex);
+
+	list_for_each_entry(pgdir, &priv->pgdir_list, list)
+		if (!mlx4_alloc_db_from_pgdir(pgdir, db, order))
+			goto out;
+
+	pgdir = mlx4_alloc_db_pgdir(&(dev->pdev->dev));
+	if (!pgdir) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	list_add(&pgdir->list, &priv->pgdir_list);
+
+	/* This should never fail -- we just allocated an empty page: */
+	WARN_ON(mlx4_alloc_db_from_pgdir(pgdir, db, order));
+
+out:
+	mutex_unlock(&priv->pgdir_mutex);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(mlx4_db_alloc);
+
+void mlx4_db_free(struct mlx4_dev *dev, struct mlx4_db *db)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int o;
+	int i;
+
+	mutex_lock(&priv->pgdir_mutex);
+
+	o = db->order;
+	i = db->index;
+
+	if (db->order == 0 && test_bit(i ^ 1, db->u.pgdir->order0)) {
+		clear_bit(i ^ 1, db->u.pgdir->order0);
+		++o;
+	}
+	i >>= o;
+	set_bit(i, db->u.pgdir->bits[o]);
+
+	if (bitmap_full(db->u.pgdir->order1, MLX4_DB_PER_PAGE / 2)) {
+		dma_free_coherent(&(dev->pdev->dev), PAGE_SIZE,
+				  db->u.pgdir->db_page, db->u.pgdir->db_dma);
+		list_del(&db->u.pgdir->list);
+		kfree(db->u.pgdir);
+	}
+
+	mutex_unlock(&priv->pgdir_mutex);
+}
+EXPORT_SYMBOL_GPL(mlx4_db_free);
+
+int mlx4_alloc_hwq_res(struct mlx4_dev *dev, struct mlx4_hwq_resources *wqres,
+		       int size, int max_direct)
+{
+	int err;
+
+	err = mlx4_db_alloc(dev, &wqres->db, 1);
+	if (err)
+		return err;
+
+	*wqres->db.db = 0;
+
+	err = mlx4_buf_alloc(dev, size, max_direct, &wqres->buf);
+	if (err)
+		goto err_db;
+
+	err = mlx4_mtt_init(dev, wqres->buf.npages, wqres->buf.page_shift,
+			    &wqres->mtt);
+	if (err)
+		goto err_buf;
+
+	err = mlx4_buf_write_mtt(dev, &wqres->mtt, &wqres->buf);
+	if (err)
+		goto err_mtt;
+
+	return 0;
+
+err_mtt:
+	mlx4_mtt_cleanup(dev, &wqres->mtt);
+err_buf:
+	mlx4_buf_free(dev, size, &wqres->buf);
+err_db:
+	mlx4_db_free(dev, &wqres->db);
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_alloc_hwq_res);
+
+void mlx4_free_hwq_res(struct mlx4_dev *dev, struct mlx4_hwq_resources *wqres,
+		       int size)
+{
+	mlx4_mtt_cleanup(dev, &wqres->mtt);
+	mlx4_buf_free(dev, size, &wqres->buf);
+	mlx4_db_free(dev, &wqres->db);
+}
+EXPORT_SYMBOL_GPL(mlx4_free_hwq_res);

kernel/drivers/net/mlx4/catas.c

@@ -0,0 +1,156 @@
+/*
+ * Copyright (c) 2007 Cisco Systems, Inc. All rights reserved.
+ * Copyright (c) 2007, 2008 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
+ * 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/workqueue.h>
+
+#include "mlx4.h"
+
+enum {
+	MLX4_CATAS_POLL_INTERVAL	= 5 * HZ,
+};
+
+static DEFINE_SPINLOCK(catas_lock);
+
+static LIST_HEAD(catas_list);
+static struct work_struct catas_work;
+
+static int internal_err_reset = 1;
+module_param(internal_err_reset, int, 0644);
+MODULE_PARM_DESC(internal_err_reset,
+		 "Reset device on internal errors if non-zero (default 1)");
+
+static void dump_err_buf(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	int i;
+
+	mlx4_err(dev, "Internal error detected:\n");
+	for (i = 0; i < priv->fw.catas_size; ++i)
+		mlx4_err(dev, "  buf[%02x]: %08x\n",
+			 i, swab32(readl(priv->catas_err.map + i)));
+}
+
+static void poll_catas(unsigned long dev_ptr)
+{
+	struct mlx4_dev *dev = (struct mlx4_dev *) dev_ptr;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	if (readl(priv->catas_err.map)) {
+		dump_err_buf(dev);
+
+		mlx4_dispatch_event(dev, MLX4_DEV_EVENT_CATASTROPHIC_ERROR, 0);
+
+		if (internal_err_reset) {
+			spin_lock(&catas_lock);
+			list_add(&priv->catas_err.list, &catas_list);
+			spin_unlock(&catas_lock);
+
+			queue_work(mlx4_wq, &catas_work);
+		}
+	} else
+		mod_timer(&priv->catas_err.timer,
+			  round_jiffies(jiffies + MLX4_CATAS_POLL_INTERVAL));
+}
+
+static void catas_reset(struct work_struct *work)
+{
+	struct mlx4_priv *priv, *tmppriv;
+	struct mlx4_dev *dev;
+
+	LIST_HEAD(tlist);
+	int ret;
+
+	spin_lock_irq(&catas_lock);
+	list_splice_init(&catas_list, &tlist);
+	spin_unlock_irq(&catas_lock);
+
+	list_for_each_entry_safe(priv, tmppriv, &tlist, catas_err.list) {
+		struct pci_dev *pdev = priv->dev.pdev;
+
+		ret = mlx4_restart_one(priv->dev.pdev);
+		/* 'priv' now is not valid */
+		if (ret)
+			printk(KERN_ERR "mlx4 %s: Reset failed (%d)\n",
+				pci_name(pdev), ret);
+		else {
+			dev  = pci_get_drvdata(pdev);
+			mlx4_dbg(dev, "Reset succeeded\n");
+		}
+	}
+}
+
+void mlx4_start_catas_poll(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	unsigned long addr;
+
+	INIT_LIST_HEAD(&priv->catas_err.list);
+	init_timer(&priv->catas_err.timer);
+	priv->catas_err.map = NULL;
+
+	addr = pci_resource_start(dev->pdev, priv->fw.catas_bar) +
+		priv->fw.catas_offset;
+
+	priv->catas_err.map = ioremap(addr, priv->fw.catas_size * 4);
+	if (!priv->catas_err.map) {
+		mlx4_warn(dev, "Failed to map internal error buffer at 0x%lx\n",
+			  addr);
+		return;
+	}
+
+	priv->catas_err.timer.data     = (unsigned long) dev;
+	priv->catas_err.timer.function = poll_catas;
+	priv->catas_err.timer.expires  =
+		round_jiffies(jiffies + MLX4_CATAS_POLL_INTERVAL);
+	add_timer(&priv->catas_err.timer);
+}
+
+void mlx4_stop_catas_poll(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	del_timer_sync(&priv->catas_err.timer);
+
+	if (priv->catas_err.map)
+		iounmap(priv->catas_err.map);
+
+	spin_lock_irq(&catas_lock);
+	list_del(&priv->catas_err.list);
+	spin_unlock_irq(&catas_lock);
+}
+
+void  __init mlx4_catas_init(void)
+{
+	INIT_WORK(&catas_work, catas_reset);
+}

kernel/drivers/net/mlx4/cmd.c

@@ -0,0 +1,442 @@
+/*
+ * Copyright (c) 2004, 2005 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2005, 2006, 2007 Cisco Systems, 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/sched.h>
+#include <linux/pci.h>
+#include <linux/errno.h>
+
+#include <linux/mlx4/cmd.h>
+
+#include <asm/io.h>
+
+#include "mlx4.h"
+
+#define CMD_POLL_TOKEN 0xffff
+
+enum {
+	/* command completed successfully: */
+	CMD_STAT_OK		= 0x00,
+	/* Internal error (such as a bus error) occurred while processing command: */
+	CMD_STAT_INTERNAL_ERR	= 0x01,
+	/* Operation/command not supported or opcode modifier not supported: */
+	CMD_STAT_BAD_OP		= 0x02,
+	/* Parameter not supported or parameter out of range: */
+	CMD_STAT_BAD_PARAM	= 0x03,
+	/* System not enabled or bad system state: */
+	CMD_STAT_BAD_SYS_STATE	= 0x04,
+	/* Attempt to access reserved or unallocaterd resource: */
+	CMD_STAT_BAD_RESOURCE	= 0x05,
+	/* Requested resource is currently executing a command, or is otherwise busy: */
+	CMD_STAT_RESOURCE_BUSY	= 0x06,
+	/* Required capability exceeds device limits: */
+	CMD_STAT_EXCEED_LIM	= 0x08,
+	/* Resource is not in the appropriate state or ownership: */
+	CMD_STAT_BAD_RES_STATE	= 0x09,
+	/* Index out of range: */
+	CMD_STAT_BAD_INDEX	= 0x0a,
+	/* FW image corrupted: */
+	CMD_STAT_BAD_NVMEM	= 0x0b,
+	/* Error in ICM mapping (e.g. not enough auxiliary ICM pages to execute command): */
+	CMD_STAT_ICM_ERROR	= 0x0c,
+	/* Attempt to modify a QP/EE which is not in the presumed state: */
+	CMD_STAT_BAD_QP_STATE   = 0x10,
+	/* Bad segment parameters (Address/Size): */
+	CMD_STAT_BAD_SEG_PARAM	= 0x20,
+	/* Memory Region has Memory Windows bound to: */
+	CMD_STAT_REG_BOUND	= 0x21,
+	/* HCA local attached memory not present: */
+	CMD_STAT_LAM_NOT_PRE	= 0x22,
+	/* Bad management packet (silently discarded): */
+	CMD_STAT_BAD_PKT	= 0x30,
+	/* More outstanding CQEs in CQ than new CQ size: */
+	CMD_STAT_BAD_SIZE	= 0x40,
+	/* Multi Function device support required: */
+	CMD_STAT_MULTI_FUNC_REQ	= 0x50,
+};
+
+enum {
+	HCR_IN_PARAM_OFFSET	= 0x00,
+	HCR_IN_MODIFIER_OFFSET	= 0x08,
+	HCR_OUT_PARAM_OFFSET	= 0x0c,
+	HCR_TOKEN_OFFSET	= 0x14,
+	HCR_STATUS_OFFSET	= 0x18,
+
+	HCR_OPMOD_SHIFT		= 12,
+	HCR_T_BIT		= 21,
+	HCR_E_BIT		= 22,
+	HCR_GO_BIT		= 23
+};
+
+enum {
+	GO_BIT_TIMEOUT_MSECS	= 10000
+};
+
+struct mlx4_cmd_context {
+	struct completion	done;
+	int			result;
+	int			next;
+	u64			out_param;
+	u16			token;
+};
+
+static int mlx4_status_to_errno(u8 status)
+{
+	static const int trans_table[] = {
+		[CMD_STAT_INTERNAL_ERR]	  = -EIO,
+		[CMD_STAT_BAD_OP]	  = -EPERM,
+		[CMD_STAT_BAD_PARAM]	  = -EINVAL,
+		[CMD_STAT_BAD_SYS_STATE]  = -ENXIO,
+		[CMD_STAT_BAD_RESOURCE]	  = -EBADF,
+		[CMD_STAT_RESOURCE_BUSY]  = -EBUSY,
+		[CMD_STAT_EXCEED_LIM]	  = -ENOMEM,
+		[CMD_STAT_BAD_RES_STATE]  = -EBADF,
+		[CMD_STAT_BAD_INDEX]	  = -EBADF,
+		[CMD_STAT_BAD_NVMEM]	  = -EFAULT,
+		[CMD_STAT_ICM_ERROR]	  = -ENFILE,
+		[CMD_STAT_BAD_QP_STATE]   = -EINVAL,
+		[CMD_STAT_BAD_SEG_PARAM]  = -EFAULT,
+		[CMD_STAT_REG_BOUND]	  = -EBUSY,
+		[CMD_STAT_LAM_NOT_PRE]	  = -EAGAIN,
+		[CMD_STAT_BAD_PKT]	  = -EINVAL,
+		[CMD_STAT_BAD_SIZE]	  = -ENOMEM,
+		[CMD_STAT_MULTI_FUNC_REQ] = -EACCES,
+	};
+
+	if (status >= ARRAY_SIZE(trans_table) ||
+	    (status != CMD_STAT_OK && trans_table[status] == 0))
+		return -EIO;
+
+	return trans_table[status];
+}
+
+static int cmd_pending(struct mlx4_dev *dev)
+{
+	u32 status = readl(mlx4_priv(dev)->cmd.hcr + HCR_STATUS_OFFSET);
+
+	return (status & swab32(1 << HCR_GO_BIT)) ||
+		(mlx4_priv(dev)->cmd.toggle ==
+		 !!(status & swab32(1 << HCR_T_BIT)));
+}
+
+static int mlx4_cmd_post(struct mlx4_dev *dev, u64 in_param, u64 out_param,
+			 u32 in_modifier, u8 op_modifier, u16 op, u16 token,
+			 int event)
+{
+	struct mlx4_cmd *cmd = &mlx4_priv(dev)->cmd;
+	u32 __iomem *hcr = cmd->hcr;
+	int ret = -EAGAIN;
+	unsigned long end;
+
+	mutex_lock(&cmd->hcr_mutex);
+
+	end = jiffies;
+	if (event)
+		end += msecs_to_jiffies(GO_BIT_TIMEOUT_MSECS);
+
+	while (cmd_pending(dev)) {
+		if (time_after_eq(jiffies, end))
+			goto out;
+		cond_resched();
+	}
+
+	/*
+	 * We use writel (instead of something like memcpy_toio)
+	 * because writes of less than 32 bits to the HCR don't work
+	 * (and some architectures such as ia64 implement memcpy_toio
+	 * in terms of writeb).
+	 */
+	__raw_writel((__force u32) cpu_to_be32(in_param >> 32),		  hcr + 0);
+	__raw_writel((__force u32) cpu_to_be32(in_param & 0xfffffffful),  hcr + 1);
+	__raw_writel((__force u32) cpu_to_be32(in_modifier),		  hcr + 2);
+	__raw_writel((__force u32) cpu_to_be32(out_param >> 32),	  hcr + 3);
+	__raw_writel((__force u32) cpu_to_be32(out_param & 0xfffffffful), hcr + 4);
+	__raw_writel((__force u32) cpu_to_be32(token << 16),		  hcr + 5);
+
+	/* __raw_writel may not order writes. */
+	wmb();
+
+	__raw_writel((__force u32) cpu_to_be32((1 << HCR_GO_BIT)		|
+					       (cmd->toggle << HCR_T_BIT)	|
+					       (event ? (1 << HCR_E_BIT) : 0)	|
+					       (op_modifier << HCR_OPMOD_SHIFT) |
+					       op),			  hcr + 6);
+
+	/*
+	 * Make sure that our HCR writes don't get mixed in with
+	 * writes from another CPU starting a FW command.
+	 */
+	mmiowb();
+
+	cmd->toggle = cmd->toggle ^ 1;
+
+	ret = 0;
+
+out:
+	mutex_unlock(&cmd->hcr_mutex);
+	return ret;
+}
+
+static int mlx4_cmd_poll(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
+			 int out_is_imm, u32 in_modifier, u8 op_modifier,
+			 u16 op, unsigned long timeout)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	void __iomem *hcr = priv->cmd.hcr;
+	int err = 0;
+	unsigned long end;
+
+	down(&priv->cmd.poll_sem);
+
+	err = mlx4_cmd_post(dev, in_param, out_param ? *out_param : 0,
+			    in_modifier, op_modifier, op, CMD_POLL_TOKEN, 0);
+	if (err)
+		goto out;
+
+	end = msecs_to_jiffies(timeout) + jiffies;
+	while (cmd_pending(dev) && time_before(jiffies, end))
+		cond_resched();
+
+	if (cmd_pending(dev)) {
+		err = -ETIMEDOUT;
+		goto out;
+	}
+
+	if (out_is_imm)
+		*out_param =
+			(u64) be32_to_cpu((__force __be32)
+					  __raw_readl(hcr + HCR_OUT_PARAM_OFFSET)) << 32 |
+			(u64) be32_to_cpu((__force __be32)
+					  __raw_readl(hcr + HCR_OUT_PARAM_OFFSET + 4));
+
+	err = mlx4_status_to_errno(be32_to_cpu((__force __be32)
+					       __raw_readl(hcr + HCR_STATUS_OFFSET)) >> 24);
+
+out:
+	up(&priv->cmd.poll_sem);
+	return err;
+}
+
+void mlx4_cmd_event(struct mlx4_dev *dev, u16 token, u8 status, u64 out_param)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_cmd_context *context =
+		&priv->cmd.context[token & priv->cmd.token_mask];
+
+	/* previously timed out command completing at long last */
+	if (token != context->token)
+		return;
+
+	context->result    = mlx4_status_to_errno(status);
+	context->out_param = out_param;
+
+	complete(&context->done);
+}
+
+static int mlx4_cmd_wait(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
+			 int out_is_imm, u32 in_modifier, u8 op_modifier,
+			 u16 op, unsigned long timeout)
+{
+	struct mlx4_cmd *cmd = &mlx4_priv(dev)->cmd;
+	struct mlx4_cmd_context *context;
+	int err = 0;
+
+	down(&cmd->event_sem);
+
+	spin_lock(&cmd->context_lock);
+	BUG_ON(cmd->free_head < 0);
+	context = &cmd->context[cmd->free_head];
+	context->token += cmd->token_mask + 1;
+	cmd->free_head = context->next;
+	spin_unlock(&cmd->context_lock);
+
+	init_completion(&context->done);
+
+	mlx4_cmd_post(dev, in_param, out_param ? *out_param : 0,
+		      in_modifier, op_modifier, op, context->token, 1);
+
+	if (!wait_for_completion_timeout(&context->done, msecs_to_jiffies(timeout))) {
+		err = -EBUSY;
+		goto out;
+	}
+
+	err = context->result;
+	if (err)
+		goto out;
+
+	if (out_is_imm)
+		*out_param = context->out_param;
+
+out:
+	spin_lock(&cmd->context_lock);
+	context->next = cmd->free_head;
+	cmd->free_head = context - cmd->context;
+	spin_unlock(&cmd->context_lock);
+
+	up(&cmd->event_sem);
+	return err;
+}
+
+int __mlx4_cmd(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
+	       int out_is_imm, u32 in_modifier, u8 op_modifier,
+	       u16 op, unsigned long timeout)
+{
+	if (mlx4_priv(dev)->cmd.use_events)
+		return mlx4_cmd_wait(dev, in_param, out_param, out_is_imm,
+				     in_modifier, op_modifier, op, timeout);
+	else
+		return mlx4_cmd_poll(dev, in_param, out_param, out_is_imm,
+				     in_modifier, op_modifier, op, timeout);
+}
+EXPORT_SYMBOL_GPL(__mlx4_cmd);
+
+int mlx4_cmd_init(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	mutex_init(&priv->cmd.hcr_mutex);
+	sema_init(&priv->cmd.poll_sem, 1);
+	priv->cmd.use_events = 0;
+	priv->cmd.toggle     = 1;
+
+	priv->cmd.hcr = ioremap(pci_resource_start(dev->pdev, 0) + MLX4_HCR_BASE,
+				MLX4_HCR_SIZE);
+	if (!priv->cmd.hcr) {
+		mlx4_err(dev, "Couldn't map command register.");
+		return -ENOMEM;
+	}
+
+	priv->cmd.pool = pci_pool_create("mlx4_cmd", dev->pdev,
+					 MLX4_MAILBOX_SIZE,
+					 MLX4_MAILBOX_SIZE, 0);
+	if (!priv->cmd.pool) {
+		iounmap(priv->cmd.hcr);
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+void mlx4_cmd_cleanup(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	pci_pool_destroy(priv->cmd.pool);
+	iounmap(priv->cmd.hcr);
+}
+
+/*
+ * Switch to using events to issue FW commands (can only be called
+ * after event queue for command events has been initialized).
+ */
+int mlx4_cmd_use_events(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int i;
+
+	priv->cmd.context = kmalloc(priv->cmd.max_cmds *
+				   sizeof (struct mlx4_cmd_context),
+				   GFP_KERNEL);
+	if (!priv->cmd.context)
+		return -ENOMEM;
+
+	for (i = 0; i < priv->cmd.max_cmds; ++i) {
+		priv->cmd.context[i].token = i;
+		priv->cmd.context[i].next  = i + 1;
+	}
+
+	priv->cmd.context[priv->cmd.max_cmds - 1].next = -1;
+	priv->cmd.free_head = 0;
+
+	sema_init(&priv->cmd.event_sem, priv->cmd.max_cmds);
+	spin_lock_init(&priv->cmd.context_lock);
+
+	for (priv->cmd.token_mask = 1;
+	     priv->cmd.token_mask < priv->cmd.max_cmds;
+	     priv->cmd.token_mask <<= 1)
+		; /* nothing */
+	--priv->cmd.token_mask;
+
+	priv->cmd.use_events = 1;
+
+	down(&priv->cmd.poll_sem);
+
+	return 0;
+}
+
+/*
+ * Switch back to polling (used when shutting down the device)
+ */
+void mlx4_cmd_use_polling(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int i;
+
+	priv->cmd.use_events = 0;
+
+	for (i = 0; i < priv->cmd.max_cmds; ++i)
+		down(&priv->cmd.event_sem);
+
+	kfree(priv->cmd.context);
+
+	up(&priv->cmd.poll_sem);
+}
+
+struct mlx4_cmd_mailbox *mlx4_alloc_cmd_mailbox(struct mlx4_dev *dev)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+
+	mailbox = kmalloc(sizeof *mailbox, GFP_KERNEL);
+	if (!mailbox)
+		return ERR_PTR(-ENOMEM);
+
+	mailbox->buf = pci_pool_alloc(mlx4_priv(dev)->cmd.pool, GFP_KERNEL,
+				      &mailbox->dma);
+	if (!mailbox->buf) {
+		kfree(mailbox);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	return mailbox;
+}
+EXPORT_SYMBOL_GPL(mlx4_alloc_cmd_mailbox);
+
+void mlx4_free_cmd_mailbox(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *mailbox)
+{
+	if (!mailbox)
+		return;
+
+	pci_pool_free(mlx4_priv(dev)->cmd.pool, mailbox->buf, mailbox->dma);
+	kfree(mailbox);
+}
+EXPORT_SYMBOL_GPL(mlx4_free_cmd_mailbox);

kernel/drivers/net/mlx4/cq.c

@@ -0,0 +1,318 @@
+/*
+ * Copyright (c) 2004, 2005 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved.
+ * Copyright (c) 2005, 2006, 2007 Cisco Systems, Inc. All rights reserved.
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2004 Voltaire, 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/hardirq.h>
+
+#include <linux/mlx4/cmd.h>
+#include <linux/mlx4/cq.h>
+
+#include "mlx4.h"
+#include "icm.h"
+
+struct mlx4_cq_context {
+	__be32			flags;
+	u16			reserved1[3];
+	__be16			page_offset;
+	__be32			logsize_usrpage;
+	__be16			cq_period;
+	__be16			cq_max_count;
+	u8			reserved2[3];
+	u8			comp_eqn;
+	u8			log_page_size;
+	u8			reserved3[2];
+	u8			mtt_base_addr_h;
+	__be32			mtt_base_addr_l;
+	__be32			last_notified_index;
+	__be32			solicit_producer_index;
+	__be32			consumer_index;
+	__be32			producer_index;
+	u32			reserved4[2];
+	__be64			db_rec_addr;
+};
+
+#define MLX4_CQ_STATUS_OK		( 0 << 28)
+#define MLX4_CQ_STATUS_OVERFLOW		( 9 << 28)
+#define MLX4_CQ_STATUS_WRITE_FAIL	(10 << 28)
+#define MLX4_CQ_FLAG_CC			( 1 << 18)
+#define MLX4_CQ_FLAG_OI			( 1 << 17)
+#define MLX4_CQ_STATE_ARMED		( 9 <<  8)
+#define MLX4_CQ_STATE_ARMED_SOL		( 6 <<  8)
+#define MLX4_EQ_STATE_FIRED		(10 <<  8)
+
+void mlx4_cq_completion(struct mlx4_dev *dev, u32 cqn)
+{
+	struct mlx4_cq *cq;
+
+	cq = radix_tree_lookup(&mlx4_priv(dev)->cq_table.tree,
+			       cqn & (dev->caps.num_cqs - 1));
+	if (!cq) {
+		mlx4_warn(dev, "Completion event for bogus CQ %08x\n", cqn);
+		return;
+	}
+
+	++cq->arm_sn;
+
+	cq->comp(cq);
+}
+
+void mlx4_cq_event(struct mlx4_dev *dev, u32 cqn, int event_type)
+{
+	struct mlx4_cq_table *cq_table = &mlx4_priv(dev)->cq_table;
+	struct mlx4_cq *cq;
+
+	spin_lock(&cq_table->lock);
+
+	cq = radix_tree_lookup(&cq_table->tree, cqn & (dev->caps.num_cqs - 1));
+	if (cq)
+		atomic_inc(&cq->refcount);
+
+	spin_unlock(&cq_table->lock);
+
+	if (!cq) {
+		mlx4_warn(dev, "Async event for bogus CQ %08x\n", cqn);
+		return;
+	}
+
+	cq->event(cq, event_type);
+
+	if (atomic_dec_and_test(&cq->refcount))
+		complete(&cq->free);
+}
+
+static int mlx4_SW2HW_CQ(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *mailbox,
+			 int cq_num)
+{
+	return mlx4_cmd(dev, mailbox->dma, cq_num, 0, MLX4_CMD_SW2HW_CQ,
+			MLX4_CMD_TIME_CLASS_A);
+}
+
+static int mlx4_MODIFY_CQ(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *mailbox,
+			 int cq_num, u32 opmod)
+{
+	return mlx4_cmd(dev, mailbox->dma, cq_num, opmod, MLX4_CMD_MODIFY_CQ,
+			MLX4_CMD_TIME_CLASS_A);
+}
+
+static int mlx4_HW2SW_CQ(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *mailbox,
+			 int cq_num)
+{
+	return mlx4_cmd_box(dev, 0, mailbox ? mailbox->dma : 0, cq_num,
+			    mailbox ? 0 : 1, MLX4_CMD_HW2SW_CQ,
+			    MLX4_CMD_TIME_CLASS_A);
+}
+
+int mlx4_cq_modify(struct mlx4_dev *dev, struct mlx4_cq *cq,
+		   u16 count, u16 period)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_cq_context *cq_context;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	cq_context = mailbox->buf;
+	memset(cq_context, 0, sizeof *cq_context);
+
+	cq_context->cq_max_count = cpu_to_be16(count);
+	cq_context->cq_period    = cpu_to_be16(period);
+
+	err = mlx4_MODIFY_CQ(dev, mailbox, cq->cqn, 1);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_cq_modify);
+
+int mlx4_cq_resize(struct mlx4_dev *dev, struct mlx4_cq *cq,
+		   int entries, struct mlx4_mtt *mtt)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_cq_context *cq_context;
+	u64 mtt_addr;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	cq_context = mailbox->buf;
+	memset(cq_context, 0, sizeof *cq_context);
+
+	cq_context->logsize_usrpage = cpu_to_be32(ilog2(entries) << 24);
+	cq_context->log_page_size   = mtt->page_shift - 12;
+	mtt_addr = mlx4_mtt_addr(dev, mtt);
+	cq_context->mtt_base_addr_h = mtt_addr >> 32;
+	cq_context->mtt_base_addr_l = cpu_to_be32(mtt_addr & 0xffffffff);
+
+	err = mlx4_MODIFY_CQ(dev, mailbox, cq->cqn, 0);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_cq_resize);
+
+int mlx4_cq_alloc(struct mlx4_dev *dev, int nent, struct mlx4_mtt *mtt,
+		  struct mlx4_uar *uar, u64 db_rec, struct mlx4_cq *cq,
+		  unsigned vector, int collapsed)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_cq_table *cq_table = &priv->cq_table;
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_cq_context *cq_context;
+	u64 mtt_addr;
+	int err;
+
+	if (vector >= dev->caps.num_comp_vectors)
+		return -EINVAL;
+
+	cq->vector = vector;
+
+	cq->cqn = mlx4_bitmap_alloc(&cq_table->bitmap);
+	if (cq->cqn == -1)
+		return -ENOMEM;
+
+	err = mlx4_table_get(dev, &cq_table->table, cq->cqn);
+	if (err)
+		goto err_out;
+
+	err = mlx4_table_get(dev, &cq_table->cmpt_table, cq->cqn);
+	if (err)
+		goto err_put;
+
+	spin_lock_irq(&cq_table->lock);
+	err = radix_tree_insert(&cq_table->tree, cq->cqn, cq);
+	spin_unlock_irq(&cq_table->lock);
+	if (err)
+		goto err_cmpt_put;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox)) {
+		err = PTR_ERR(mailbox);
+		goto err_radix;
+	}
+
+	cq_context = mailbox->buf;
+	memset(cq_context, 0, sizeof *cq_context);
+
+	cq_context->flags	    = cpu_to_be32(!!collapsed << 18);
+	cq_context->logsize_usrpage = cpu_to_be32((ilog2(nent) << 24) | uar->index);
+	cq_context->comp_eqn	    = priv->eq_table.eq[vector].eqn;
+	cq_context->log_page_size   = mtt->page_shift - MLX4_ICM_PAGE_SHIFT;
+
+	mtt_addr = mlx4_mtt_addr(dev, mtt);
+	cq_context->mtt_base_addr_h = mtt_addr >> 32;
+	cq_context->mtt_base_addr_l = cpu_to_be32(mtt_addr & 0xffffffff);
+	cq_context->db_rec_addr     = cpu_to_be64(db_rec);
+
+	err = mlx4_SW2HW_CQ(dev, mailbox, cq->cqn);
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	if (err)
+		goto err_radix;
+
+	cq->cons_index = 0;
+	cq->arm_sn     = 1;
+	cq->uar        = uar;
+	atomic_set(&cq->refcount, 1);
+	init_completion(&cq->free);
+
+	return 0;
+
+err_radix:
+	spin_lock_irq(&cq_table->lock);
+	radix_tree_delete(&cq_table->tree, cq->cqn);
+	spin_unlock_irq(&cq_table->lock);
+
+err_cmpt_put:
+	mlx4_table_put(dev, &cq_table->cmpt_table, cq->cqn);
+
+err_put:
+	mlx4_table_put(dev, &cq_table->table, cq->cqn);
+
+err_out:
+	mlx4_bitmap_free(&cq_table->bitmap, cq->cqn);
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_cq_alloc);
+
+void mlx4_cq_free(struct mlx4_dev *dev, struct mlx4_cq *cq)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_cq_table *cq_table = &priv->cq_table;
+	int err;
+
+	err = mlx4_HW2SW_CQ(dev, NULL, cq->cqn);
+	if (err)
+		mlx4_warn(dev, "HW2SW_CQ failed (%d) for CQN %06x\n", err, cq->cqn);
+
+	synchronize_irq(priv->eq_table.eq[cq->vector].irq);
+
+	spin_lock_irq(&cq_table->lock);
+	radix_tree_delete(&cq_table->tree, cq->cqn);
+	spin_unlock_irq(&cq_table->lock);
+
+	if (atomic_dec_and_test(&cq->refcount))
+		complete(&cq->free);
+	wait_for_completion(&cq->free);
+
+	mlx4_table_put(dev, &cq_table->table, cq->cqn);
+	mlx4_bitmap_free(&cq_table->bitmap, cq->cqn);
+}
+EXPORT_SYMBOL_GPL(mlx4_cq_free);
+
+int mlx4_init_cq_table(struct mlx4_dev *dev)
+{
+	struct mlx4_cq_table *cq_table = &mlx4_priv(dev)->cq_table;
+	int err;
+
+	spin_lock_init(&cq_table->lock);
+	INIT_RADIX_TREE(&cq_table->tree, GFP_ATOMIC);
+
+	err = mlx4_bitmap_init(&cq_table->bitmap, dev->caps.num_cqs,
+			       dev->caps.num_cqs - 1, dev->caps.reserved_cqs, 0);
+	if (err)
+		return err;
+
+	return 0;
+}
+
+void mlx4_cleanup_cq_table(struct mlx4_dev *dev)
+{
+	/* Nothing to do to clean up radix_tree */
+	mlx4_bitmap_cleanup(&mlx4_priv(dev)->cq_table.bitmap);
+}

kernel/drivers/net/mlx4/en_cq.c

@@ -0,0 +1,154 @@
+/*
+ * Copyright (c) 2007 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
+ * 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/mlx4/cq.h>
+#include <linux/mlx4/qp.h>
+#include <linux/mlx4/cmd.h>
+
+#include "mlx4_en.h"
+
+static void mlx4_en_cq_event(struct mlx4_cq *cq, enum mlx4_event event)
+{
+	return;
+}
+
+
+int mlx4_en_create_cq(struct mlx4_en_priv *priv,
+		      struct mlx4_en_cq *cq,
+		      int entries, int ring, enum cq_type mode)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	int err;
+
+	cq->size = entries;
+	if (mode == RX) {
+		cq->buf_size = cq->size * sizeof(struct mlx4_cqe);
+		cq->vector   = ring % mdev->dev->caps.num_comp_vectors;
+	} else {
+		cq->buf_size = sizeof(struct mlx4_cqe);
+		cq->vector   = 0;
+	}
+
+	cq->ring = ring;
+	cq->is_tx = mode;
+	spin_lock_init(&cq->lock);
+
+	err = mlx4_alloc_hwq_res(mdev->dev, &cq->wqres,
+				cq->buf_size, 2 * PAGE_SIZE);
+	if (err)
+		return err;
+
+	err = mlx4_en_map_buffer(&cq->wqres.buf);
+	if (err)
+		mlx4_free_hwq_res(mdev->dev, &cq->wqres, cq->buf_size);
+	else
+		cq->buf = (struct mlx4_cqe *) cq->wqres.buf.direct.buf;
+
+	return err;
+}
+
+int mlx4_en_activate_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	int err;
+
+	cq->dev = mdev->pndev[priv->port];
+	cq->mcq.set_ci_db  = cq->wqres.db.db;
+	cq->mcq.arm_db     = cq->wqres.db.db + 1;
+	*cq->mcq.set_ci_db = 0;
+	*cq->mcq.arm_db    = 0;
+	memset(cq->buf, 0, cq->buf_size);
+
+	if (!cq->is_tx)
+		cq->size = priv->rx_ring[cq->ring].actual_size;
+
+	err = mlx4_cq_alloc(mdev->dev, cq->size, &cq->wqres.mtt, &mdev->priv_uar,
+			    cq->wqres.db.dma, &cq->mcq, cq->vector, cq->is_tx);
+	if (err)
+		return err;
+
+	cq->mcq.comp  = cq->is_tx ? mlx4_en_tx_irq : mlx4_en_rx_irq;
+	cq->mcq.event = mlx4_en_cq_event;
+
+	if (cq->is_tx) {
+		init_timer(&cq->timer);
+		cq->timer.function = mlx4_en_poll_tx_cq;
+		cq->timer.data = (unsigned long) cq;
+	} else {
+		netif_napi_add(cq->dev, &cq->napi, mlx4_en_poll_rx_cq, 64);
+		napi_enable(&cq->napi);
+	}
+
+	return 0;
+}
+
+void mlx4_en_destroy_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+
+	mlx4_en_unmap_buffer(&cq->wqres.buf);
+	mlx4_free_hwq_res(mdev->dev, &cq->wqres, cq->buf_size);
+	cq->buf_size = 0;
+	cq->buf = NULL;
+}
+
+void mlx4_en_deactivate_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+
+	if (cq->is_tx)
+		del_timer(&cq->timer);
+	else {
+		napi_disable(&cq->napi);
+		netif_napi_del(&cq->napi);
+	}
+
+	mlx4_cq_free(mdev->dev, &cq->mcq);
+}
+
+/* Set rx cq moderation parameters */
+int mlx4_en_set_cq_moder(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq)
+{
+	return mlx4_cq_modify(priv->mdev->dev, &cq->mcq,
+			      cq->moder_cnt, cq->moder_time);
+}
+
+int mlx4_en_arm_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq)
+{
+	mlx4_cq_arm(&cq->mcq, MLX4_CQ_DB_REQ_NOT, priv->mdev->uar_map,
+		    &priv->mdev->uar_lock);
+
+	return 0;
+}
+
+

kernel/drivers/net/mlx4/en_ethtool.c

@@ -0,0 +1,426 @@
+/*
+ * Copyright (c) 2007 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
+ * 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/kernel.h>
+#include <linux/ethtool.h>
+#include <linux/netdevice.h>
+
+#include "mlx4_en.h"
+#include "en_port.h"
+
+
+static void mlx4_en_update_lro_stats(struct mlx4_en_priv *priv)
+{
+	int i;
+
+	priv->port_stats.lro_aggregated = 0;
+	priv->port_stats.lro_flushed = 0;
+	priv->port_stats.lro_no_desc = 0;
+
+	for (i = 0; i < priv->rx_ring_num; i++) {
+		priv->port_stats.lro_aggregated += priv->rx_ring[i].lro.stats.aggregated;
+		priv->port_stats.lro_flushed += priv->rx_ring[i].lro.stats.flushed;
+		priv->port_stats.lro_no_desc += priv->rx_ring[i].lro.stats.no_desc;
+	}
+}
+
+static void
+mlx4_en_get_drvinfo(struct net_device *dev, struct ethtool_drvinfo *drvinfo)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+
+	sprintf(drvinfo->driver, DRV_NAME " (%s)", mdev->dev->board_id);
+	strncpy(drvinfo->version, DRV_VERSION " (" DRV_RELDATE ")", 32);
+	sprintf(drvinfo->fw_version, "%d.%d.%d",
+		(u16) (mdev->dev->caps.fw_ver >> 32),
+		(u16) ((mdev->dev->caps.fw_ver >> 16) & 0xffff),
+		(u16) (mdev->dev->caps.fw_ver & 0xffff));
+	strncpy(drvinfo->bus_info, pci_name(mdev->dev->pdev), 32);
+	drvinfo->n_stats = 0;
+	drvinfo->regdump_len = 0;
+	drvinfo->eedump_len = 0;
+}
+
+static u32 mlx4_en_get_tso(struct net_device *dev)
+{
+	return (dev->features & NETIF_F_TSO) != 0;
+}
+
+static int mlx4_en_set_tso(struct net_device *dev, u32 data)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+
+	if (data) {
+		if (!priv->mdev->LSO_support)
+			return -EPERM;
+		dev->features |= (NETIF_F_TSO | NETIF_F_TSO6);
+	} else
+		dev->features &= ~(NETIF_F_TSO | NETIF_F_TSO6);
+	return 0;
+}
+
+static u32 mlx4_en_get_rx_csum(struct net_device *dev)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	return priv->rx_csum;
+}
+
+static int mlx4_en_set_rx_csum(struct net_device *dev, u32 data)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	priv->rx_csum = (data != 0);
+	return 0;
+}
+
+static const char main_strings[][ETH_GSTRING_LEN] = {
+	"rx_packets", "tx_packets", "rx_bytes", "tx_bytes", "rx_errors",
+	"tx_errors", "rx_dropped", "tx_dropped", "multicast", "collisions",
+	"rx_length_errors", "rx_over_errors", "rx_crc_errors",
+	"rx_frame_errors", "rx_fifo_errors", "rx_missed_errors",
+	"tx_aborted_errors", "tx_carrier_errors", "tx_fifo_errors",
+	"tx_heartbeat_errors", "tx_window_errors",
+
+	/* port statistics */
+	"lro_aggregated", "lro_flushed", "lro_no_desc", "tso_packets",
+	"queue_stopped", "wake_queue", "tx_timeout", "rx_alloc_failed",
+	"rx_csum_good", "rx_csum_none", "tx_chksum_offload",
+
+	/* packet statistics */
+	"broadcast", "rx_prio_0", "rx_prio_1", "rx_prio_2", "rx_prio_3",
+	"rx_prio_4", "rx_prio_5", "rx_prio_6", "rx_prio_7", "tx_prio_0",
+	"tx_prio_1", "tx_prio_2", "tx_prio_3", "tx_prio_4", "tx_prio_5",
+	"tx_prio_6", "tx_prio_7",
+};
+#define NUM_MAIN_STATS	21
+#define NUM_ALL_STATS	(NUM_MAIN_STATS + NUM_PORT_STATS + NUM_PKT_STATS + NUM_PERF_STATS)
+
+static u32 mlx4_en_get_msglevel(struct net_device *dev)
+{
+	return ((struct mlx4_en_priv *) netdev_priv(dev))->msg_enable;
+}
+
+static void mlx4_en_set_msglevel(struct net_device *dev, u32 val)
+{
+	((struct mlx4_en_priv *) netdev_priv(dev))->msg_enable = val;
+}
+
+static void mlx4_en_get_wol(struct net_device *netdev,
+			    struct ethtool_wolinfo *wol)
+{
+	wol->supported = 0;
+	wol->wolopts = 0;
+
+	return;
+}
+
+static int mlx4_en_get_sset_count(struct net_device *dev, int sset)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+
+	if (sset != ETH_SS_STATS)
+		return -EOPNOTSUPP;
+
+	return NUM_ALL_STATS + (priv->tx_ring_num + priv->rx_ring_num) * 2;
+}
+
+static void mlx4_en_get_ethtool_stats(struct net_device *dev,
+		struct ethtool_stats *stats, uint64_t *data)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	int index = 0;
+	int i;
+
+	spin_lock_bh(&priv->stats_lock);
+
+	mlx4_en_update_lro_stats(priv);
+
+	for (i = 0; i < NUM_MAIN_STATS; i++)
+		data[index++] = ((unsigned long *) &priv->stats)[i];
+	for (i = 0; i < NUM_PORT_STATS; i++)
+		data[index++] = ((unsigned long *) &priv->port_stats)[i];
+	for (i = 0; i < priv->tx_ring_num; i++) {
+		data[index++] = priv->tx_ring[i].packets;
+		data[index++] = priv->tx_ring[i].bytes;
+	}
+	for (i = 0; i < priv->rx_ring_num; i++) {
+		data[index++] = priv->rx_ring[i].packets;
+		data[index++] = priv->rx_ring[i].bytes;
+	}
+	for (i = 0; i < NUM_PKT_STATS; i++)
+		data[index++] = ((unsigned long *) &priv->pkstats)[i];
+	spin_unlock_bh(&priv->stats_lock);
+
+}
+
+static void mlx4_en_get_strings(struct net_device *dev,
+				uint32_t stringset, uint8_t *data)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	int index = 0;
+	int i;
+
+	if (stringset != ETH_SS_STATS)
+		return;
+
+	/* Add main counters */
+	for (i = 0; i < NUM_MAIN_STATS; i++)
+		strcpy(data + (index++) * ETH_GSTRING_LEN, main_strings[i]);
+	for (i = 0; i < NUM_PORT_STATS; i++)
+		strcpy(data + (index++) * ETH_GSTRING_LEN,
+			main_strings[i + NUM_MAIN_STATS]);
+	for (i = 0; i < priv->tx_ring_num; i++) {
+		sprintf(data + (index++) * ETH_GSTRING_LEN,
+			"tx%d_packets", i);
+		sprintf(data + (index++) * ETH_GSTRING_LEN,
+			"tx%d_bytes", i);
+	}
+	for (i = 0; i < priv->rx_ring_num; i++) {
+		sprintf(data + (index++) * ETH_GSTRING_LEN,
+			"rx%d_packets", i);
+		sprintf(data + (index++) * ETH_GSTRING_LEN,
+			"rx%d_bytes", i);
+	}
+	for (i = 0; i < NUM_PKT_STATS; i++)
+		strcpy(data + (index++) * ETH_GSTRING_LEN,
+			main_strings[i + NUM_MAIN_STATS + NUM_PORT_STATS]);
+}
+
+static int mlx4_en_get_settings(struct net_device *dev, struct ethtool_cmd *cmd)
+{
+	cmd->autoneg = AUTONEG_DISABLE;
+	cmd->supported = SUPPORTED_10000baseT_Full;
+	cmd->advertising = ADVERTISED_1000baseT_Full;
+	if (netif_carrier_ok(dev)) {
+		cmd->speed = SPEED_10000;
+		cmd->duplex = DUPLEX_FULL;
+	} else {
+		cmd->speed = -1;
+		cmd->duplex = -1;
+	}
+	return 0;
+}
+
+static int mlx4_en_set_settings(struct net_device *dev, struct ethtool_cmd *cmd)
+{
+	if ((cmd->autoneg == AUTONEG_ENABLE) ||
+	    (cmd->speed != SPEED_10000) || (cmd->duplex != DUPLEX_FULL))
+		return -EINVAL;
+
+	/* Nothing to change */
+	return 0;
+}
+
+static int mlx4_en_get_coalesce(struct net_device *dev,
+			      struct ethtool_coalesce *coal)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+
+	coal->tx_coalesce_usecs = 0;
+	coal->tx_max_coalesced_frames = 0;
+	coal->rx_coalesce_usecs = priv->rx_usecs;
+	coal->rx_max_coalesced_frames = priv->rx_frames;
+
+	coal->pkt_rate_low = priv->pkt_rate_low;
+	coal->rx_coalesce_usecs_low = priv->rx_usecs_low;
+	coal->pkt_rate_high = priv->pkt_rate_high;
+	coal->rx_coalesce_usecs_high = priv->rx_usecs_high;
+	coal->rate_sample_interval = priv->sample_interval;
+	coal->use_adaptive_rx_coalesce = priv->adaptive_rx_coal;
+	return 0;
+}
+
+static int mlx4_en_set_coalesce(struct net_device *dev,
+			      struct ethtool_coalesce *coal)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	int err, i;
+
+	priv->rx_frames = (coal->rx_max_coalesced_frames ==
+			   MLX4_EN_AUTO_CONF) ?
+				MLX4_EN_RX_COAL_TARGET :
+				coal->rx_max_coalesced_frames;
+	priv->rx_usecs = (coal->rx_coalesce_usecs ==
+			  MLX4_EN_AUTO_CONF) ?
+				MLX4_EN_RX_COAL_TIME :
+				coal->rx_coalesce_usecs;
+
+	/* Set adaptive coalescing params */
+	priv->pkt_rate_low = coal->pkt_rate_low;
+	priv->rx_usecs_low = coal->rx_coalesce_usecs_low;
+	priv->pkt_rate_high = coal->pkt_rate_high;
+	priv->rx_usecs_high = coal->rx_coalesce_usecs_high;
+	priv->sample_interval = coal->rate_sample_interval;
+	priv->adaptive_rx_coal = coal->use_adaptive_rx_coalesce;
+	priv->last_moder_time = MLX4_EN_AUTO_CONF;
+	if (priv->adaptive_rx_coal)
+		return 0;
+
+	for (i = 0; i < priv->rx_ring_num; i++) {
+		priv->rx_cq[i].moder_cnt = priv->rx_frames;
+		priv->rx_cq[i].moder_time = priv->rx_usecs;
+		err = mlx4_en_set_cq_moder(priv, &priv->rx_cq[i]);
+		if (err)
+			return err;
+	}
+	return 0;
+}
+
+static int mlx4_en_set_pauseparam(struct net_device *dev,
+				struct ethtool_pauseparam *pause)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	int err;
+
+	priv->prof->tx_pause = pause->tx_pause != 0;
+	priv->prof->rx_pause = pause->rx_pause != 0;
+	err = mlx4_SET_PORT_general(mdev->dev, priv->port,
+				    priv->rx_skb_size + ETH_FCS_LEN,
+				    priv->prof->tx_pause,
+				    priv->prof->tx_ppp,
+				    priv->prof->rx_pause,
+				    priv->prof->rx_ppp);
+	if (err)
+		en_err(priv, "Failed setting pause params\n");
+
+	return err;
+}
+
+static void mlx4_en_get_pauseparam(struct net_device *dev,
+				 struct ethtool_pauseparam *pause)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+
+	pause->tx_pause = priv->prof->tx_pause;
+	pause->rx_pause = priv->prof->rx_pause;
+}
+
+static int mlx4_en_set_ringparam(struct net_device *dev,
+				 struct ethtool_ringparam *param)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	u32 rx_size, tx_size;
+	int port_up = 0;
+	int err = 0;
+
+	if (param->rx_jumbo_pending || param->rx_mini_pending)
+		return -EINVAL;
+
+	rx_size = roundup_pow_of_two(param->rx_pending);
+	rx_size = max_t(u32, rx_size, MLX4_EN_MIN_RX_SIZE);
+	rx_size = min_t(u32, rx_size, MLX4_EN_MAX_RX_SIZE);
+	tx_size = roundup_pow_of_two(param->tx_pending);
+	tx_size = max_t(u32, tx_size, MLX4_EN_MIN_TX_SIZE);
+	tx_size = min_t(u32, tx_size, MLX4_EN_MAX_TX_SIZE);
+
+	if (rx_size == priv->prof->rx_ring_size &&
+	    tx_size == priv->prof->tx_ring_size)
+		return 0;
+
+	mutex_lock(&mdev->state_lock);
+	if (priv->port_up) {
+		port_up = 1;
+		mlx4_en_stop_port(dev);
+	}
+
+	mlx4_en_free_resources(priv);
+
+	priv->prof->tx_ring_size = tx_size;
+	priv->prof->rx_ring_size = rx_size;
+
+	err = mlx4_en_alloc_resources(priv);
+	if (err) {
+		en_err(priv, "Failed reallocating port resources\n");
+		goto out;
+	}
+	if (port_up) {
+		err = mlx4_en_start_port(dev);
+		if (err)
+			en_err(priv, "Failed starting port\n");
+	}
+
+out:
+	mutex_unlock(&mdev->state_lock);
+	return err;
+}
+
+static void mlx4_en_get_ringparam(struct net_device *dev,
+				  struct ethtool_ringparam *param)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+
+	memset(param, 0, sizeof(*param));
+	param->rx_max_pending = MLX4_EN_MAX_RX_SIZE;
+	param->tx_max_pending = MLX4_EN_MAX_TX_SIZE;
+	param->rx_pending = mdev->profile.prof[priv->port].rx_ring_size;
+	param->tx_pending = mdev->profile.prof[priv->port].tx_ring_size;
+}
+
+const struct ethtool_ops mlx4_en_ethtool_ops = {
+	.get_drvinfo = mlx4_en_get_drvinfo,
+	.get_settings = mlx4_en_get_settings,
+	.set_settings = mlx4_en_set_settings,
+#ifdef NETIF_F_TSO
+	.get_tso = mlx4_en_get_tso,
+	.set_tso = mlx4_en_set_tso,
+#endif
+	.get_sg = ethtool_op_get_sg,
+	.set_sg = ethtool_op_set_sg,
+	.get_link = ethtool_op_get_link,
+	.get_rx_csum = mlx4_en_get_rx_csum,
+	.set_rx_csum = mlx4_en_set_rx_csum,
+	.get_tx_csum = ethtool_op_get_tx_csum,
+	.set_tx_csum = ethtool_op_set_tx_ipv6_csum,
+	.get_strings = mlx4_en_get_strings,
+	.get_sset_count = mlx4_en_get_sset_count,
+	.get_ethtool_stats = mlx4_en_get_ethtool_stats,
+	.get_wol = mlx4_en_get_wol,
+	.get_msglevel = mlx4_en_get_msglevel,
+	.set_msglevel = mlx4_en_set_msglevel,
+	.get_coalesce = mlx4_en_get_coalesce,
+	.set_coalesce = mlx4_en_set_coalesce,
+	.get_pauseparam = mlx4_en_get_pauseparam,
+	.set_pauseparam = mlx4_en_set_pauseparam,
+	.get_ringparam = mlx4_en_get_ringparam,
+	.set_ringparam = mlx4_en_set_ringparam,
+	.get_flags = ethtool_op_get_flags,
+	.set_flags = ethtool_op_set_flags,
+};
+
+
+
+
+

kernel/drivers/net/mlx4/en_main.c

@@ -0,0 +1,282 @@
+/*
+ * Copyright (c) 2007 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
+ * 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/cpumask.h>
+#include <linux/module.h>
+#include <linux/delay.h>
+#include <linux/netdevice.h>
+
+#include <linux/mlx4/driver.h>
+#include <linux/mlx4/device.h>
+#include <linux/mlx4/cmd.h>
+
+#include "mlx4_en.h"
+
+MODULE_AUTHOR("Liran Liss, Yevgeny Petrilin");
+MODULE_DESCRIPTION("Mellanox ConnectX HCA Ethernet driver");
+MODULE_LICENSE("Dual BSD/GPL");
+MODULE_VERSION(DRV_VERSION " ("DRV_RELDATE")");
+
+static const char mlx4_en_version[] =
+	DRV_NAME ": Mellanox ConnectX HCA Ethernet driver v"
+	DRV_VERSION " (" DRV_RELDATE ")\n";
+
+#define MLX4_EN_PARM_INT(X, def_val, desc) \
+	static unsigned int X = def_val;\
+	module_param(X , uint, 0444); \
+	MODULE_PARM_DESC(X, desc);
+
+
+/*
+ * Device scope module parameters
+ */
+
+
+/* Use a XOR rathern than Toeplitz hash function for RSS */
+MLX4_EN_PARM_INT(rss_xor, 0, "Use XOR hash function for RSS");
+
+/* RSS hash type mask - default to <saddr, daddr, sport, dport> */
+MLX4_EN_PARM_INT(rss_mask, 0xf, "RSS hash type bitmask");
+
+/* Number of LRO sessions per Rx ring (rounded up to a power of two) */
+MLX4_EN_PARM_INT(num_lro, MLX4_EN_MAX_LRO_DESCRIPTORS,
+		 "Number of LRO sessions per ring or disabled (0)");
+
+/* Priority pausing */
+MLX4_EN_PARM_INT(pfctx, 0, "Priority based Flow Control policy on TX[7:0]."
+			   " Per priority bit mask");
+MLX4_EN_PARM_INT(pfcrx, 0, "Priority based Flow Control policy on RX[7:0]."
+			   " Per priority bit mask");
+
+static int mlx4_en_get_profile(struct mlx4_en_dev *mdev)
+{
+	struct mlx4_en_profile *params = &mdev->profile;
+	int i;
+
+	params->rss_xor = (rss_xor != 0);
+	params->rss_mask = rss_mask & 0x1f;
+	params->num_lro = min_t(int, num_lro , MLX4_EN_MAX_LRO_DESCRIPTORS);
+	for (i = 1; i <= MLX4_MAX_PORTS; i++) {
+		params->prof[i].rx_pause = 1;
+		params->prof[i].rx_ppp = pfcrx;
+		params->prof[i].tx_pause = 1;
+		params->prof[i].tx_ppp = pfctx;
+		params->prof[i].tx_ring_size = MLX4_EN_DEF_TX_RING_SIZE;
+		params->prof[i].rx_ring_size = MLX4_EN_DEF_RX_RING_SIZE;
+		params->prof[i].tx_ring_num = MLX4_EN_NUM_TX_RINGS +
+			(!!pfcrx) * MLX4_EN_NUM_PPP_RINGS;
+	}
+
+	return 0;
+}
+
+static void mlx4_en_event(struct mlx4_dev *dev, void *endev_ptr,
+			  enum mlx4_dev_event event, int port)
+{
+	struct mlx4_en_dev *mdev = (struct mlx4_en_dev *) endev_ptr;
+	struct mlx4_en_priv *priv;
+
+	if (!mdev->pndev[port])
+		return;
+
+	priv = netdev_priv(mdev->pndev[port]);
+	switch (event) {
+	case MLX4_DEV_EVENT_PORT_UP:
+	case MLX4_DEV_EVENT_PORT_DOWN:
+		/* To prevent races, we poll the link state in a separate
+		  task rather than changing it here */
+		priv->link_state = event;
+		queue_work(mdev->workqueue, &priv->linkstate_task);
+		break;
+
+	case MLX4_DEV_EVENT_CATASTROPHIC_ERROR:
+		mlx4_err(mdev, "Internal error detected, restarting device\n");
+		break;
+
+	default:
+		mlx4_warn(mdev, "Unhandled event: %d\n", event);
+	}
+}
+
+static void mlx4_en_remove(struct mlx4_dev *dev, void *endev_ptr)
+{
+	struct mlx4_en_dev *mdev = endev_ptr;
+	int i;
+
+	mutex_lock(&mdev->state_lock);
+	mdev->device_up = false;
+	mutex_unlock(&mdev->state_lock);
+
+	mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_ETH)
+		if (mdev->pndev[i])
+			mlx4_en_destroy_netdev(mdev->pndev[i]);
+
+	flush_workqueue(mdev->workqueue);
+	destroy_workqueue(mdev->workqueue);
+	mlx4_mr_free(dev, &mdev->mr);
+	mlx4_uar_free(dev, &mdev->priv_uar);
+	mlx4_pd_free(dev, mdev->priv_pdn);
+	kfree(mdev);
+}
+
+static void *mlx4_en_add(struct mlx4_dev *dev)
+{
+	static int mlx4_en_version_printed;
+	struct mlx4_en_dev *mdev;
+	int i;
+	int err;
+
+	if (!mlx4_en_version_printed) {
+		printk(KERN_INFO "%s", mlx4_en_version);
+		mlx4_en_version_printed++;
+	}
+
+	mdev = kzalloc(sizeof *mdev, GFP_KERNEL);
+	if (!mdev) {
+		dev_err(&dev->pdev->dev, "Device struct alloc failed, "
+			"aborting.\n");
+		err = -ENOMEM;
+		goto err_free_res;
+	}
+
+	if (mlx4_pd_alloc(dev, &mdev->priv_pdn))
+		goto err_free_dev;
+
+	if (mlx4_uar_alloc(dev, &mdev->priv_uar))
+		goto err_pd;
+
+	mdev->uar_map = ioremap(mdev->priv_uar.pfn << PAGE_SHIFT, PAGE_SIZE);
+	if (!mdev->uar_map)
+		goto err_uar;
+	spin_lock_init(&mdev->uar_lock);
+
+	mdev->dev = dev;
+	mdev->dma_device = &(dev->pdev->dev);
+	mdev->pdev = dev->pdev;
+	mdev->device_up = false;
+
+	mdev->LSO_support = !!(dev->caps.flags & (1 << 15));
+	if (!mdev->LSO_support)
+		mlx4_warn(mdev, "LSO not supported, please upgrade to later "
+				"FW version to enable LSO\n");
+
+	if (mlx4_mr_alloc(mdev->dev, mdev->priv_pdn, 0, ~0ull,
+			 MLX4_PERM_LOCAL_WRITE |  MLX4_PERM_LOCAL_READ,
+			 0, 0, &mdev->mr)) {
+		mlx4_err(mdev, "Failed allocating memory region\n");
+		goto err_uar;
+	}
+	if (mlx4_mr_enable(mdev->dev, &mdev->mr)) {
+		mlx4_err(mdev, "Failed enabling memory region\n");
+		goto err_mr;
+	}
+
+	/* Build device profile according to supplied module parameters */
+	err = mlx4_en_get_profile(mdev);
+	if (err) {
+		mlx4_err(mdev, "Bad module parameters, aborting.\n");
+		goto err_mr;
+	}
+
+	/* Configure wich ports to start according to module parameters */
+	mdev->port_cnt = 0;
+	mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_ETH)
+		mdev->port_cnt++;
+
+	/* If we did not receive an explicit number of Rx rings, default to
+	 * the number of completion vectors populated by the mlx4_core */
+	mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_ETH) {
+		mlx4_info(mdev, "Using %d tx rings for port:%d\n",
+			  mdev->profile.prof[i].tx_ring_num, i);
+		mdev->profile.prof[i].rx_ring_num = min_t(int,
+			roundup_pow_of_two(dev->caps.num_comp_vectors),
+			MAX_RX_RINGS);
+		mlx4_info(mdev, "Defaulting to %d rx rings for port:%d\n",
+			  mdev->profile.prof[i].rx_ring_num, i);
+	}
+
+	/* Create our own workqueue for reset/multicast tasks
+	 * Note: we cannot use the shared workqueue because of deadlocks caused
+	 *       by the rtnl lock */
+	mdev->workqueue = create_singlethread_workqueue("mlx4_en");
+	if (!mdev->workqueue) {
+		err = -ENOMEM;
+		goto err_mr;
+	}
+
+	/* At this stage all non-port specific tasks are complete:
+	 * mark the card state as up */
+	mutex_init(&mdev->state_lock);
+	mdev->device_up = true;
+
+	/* Setup ports */
+
+	/* Create a netdev for each port */
+	mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_ETH) {
+		mlx4_info(mdev, "Activating port:%d\n", i);
+		if (mlx4_en_init_netdev(mdev, i, &mdev->profile.prof[i]))
+			mdev->pndev[i] = NULL;
+	}
+	return mdev;
+
+err_mr:
+	mlx4_mr_free(dev, &mdev->mr);
+err_uar:
+	mlx4_uar_free(dev, &mdev->priv_uar);
+err_pd:
+	mlx4_pd_free(dev, mdev->priv_pdn);
+err_free_dev:
+	kfree(mdev);
+err_free_res:
+	return NULL;
+}
+
+static struct mlx4_interface mlx4_en_interface = {
+	.add	= mlx4_en_add,
+	.remove	= mlx4_en_remove,
+	.event	= mlx4_en_event,
+};
+
+static int __init mlx4_en_init(void)
+{
+	return mlx4_register_interface(&mlx4_en_interface);
+}
+
+static void __exit mlx4_en_cleanup(void)
+{
+	mlx4_unregister_interface(&mlx4_en_interface);
+}
+
+module_init(mlx4_en_init);
+module_exit(mlx4_en_cleanup);
+

kernel/drivers/net/mlx4/en_port.c

@@ -0,0 +1,242 @@
+/*
+ * Copyright (c) 2007 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
+ * 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/if_vlan.h>
+
+#include <linux/mlx4/device.h>
+#include <linux/mlx4/cmd.h>
+
+#include "en_port.h"
+#include "mlx4_en.h"
+
+
+int mlx4_SET_MCAST_FLTR(struct mlx4_dev *dev, u8 port,
+			u64 mac, u64 clear, u8 mode)
+{
+	return mlx4_cmd(dev, (mac | (clear << 63)), port, mode,
+			MLX4_CMD_SET_MCAST_FLTR, MLX4_CMD_TIME_CLASS_B);
+}
+
+int mlx4_SET_VLAN_FLTR(struct mlx4_dev *dev, u8 port, struct vlan_group *grp)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_set_vlan_fltr_mbox *filter;
+	int i;
+	int j;
+	int index = 0;
+	u32 entry;
+	int err = 0;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	filter = mailbox->buf;
+	if (grp) {
+		memset(filter, 0, sizeof *filter);
+		for (i = VLAN_FLTR_SIZE - 1; i >= 0; i--) {
+			entry = 0;
+			for (j = 0; j < 32; j++)
+				if (vlan_group_get_device(grp, index++))
+					entry |= 1 << j;
+			filter->entry[i] = cpu_to_be32(entry);
+		}
+	} else {
+		/* When no vlans are configured we block all vlans */
+		memset(filter, 0, sizeof(*filter));
+	}
+	err = mlx4_cmd(dev, mailbox->dma, port, 0, MLX4_CMD_SET_VLAN_FLTR,
+		       MLX4_CMD_TIME_CLASS_B);
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+
+int mlx4_SET_PORT_general(struct mlx4_dev *dev, u8 port, int mtu,
+			  u8 pptx, u8 pfctx, u8 pprx, u8 pfcrx)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_set_port_general_context *context;
+	int err;
+	u32 in_mod;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	context = mailbox->buf;
+	memset(context, 0, sizeof *context);
+
+	context->flags = SET_PORT_GEN_ALL_VALID;
+	context->mtu = cpu_to_be16(mtu);
+	context->pptx = (pptx * (!pfctx)) << 7;
+	context->pfctx = pfctx;
+	context->pprx = (pprx * (!pfcrx)) << 7;
+	context->pfcrx = pfcrx;
+
+	in_mod = MLX4_SET_PORT_GENERAL << 8 | port;
+	err = mlx4_cmd(dev, mailbox->dma, in_mod, 1, MLX4_CMD_SET_PORT,
+		       MLX4_CMD_TIME_CLASS_B);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+int mlx4_SET_PORT_qpn_calc(struct mlx4_dev *dev, u8 port, u32 base_qpn,
+			   u8 promisc)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_set_port_rqp_calc_context *context;
+	int err;
+	u32 in_mod;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	context = mailbox->buf;
+	memset(context, 0, sizeof *context);
+
+	context->base_qpn = cpu_to_be32(base_qpn);
+	context->promisc = cpu_to_be32(promisc << SET_PORT_PROMISC_SHIFT | base_qpn);
+	context->mcast = cpu_to_be32(1 << SET_PORT_PROMISC_SHIFT | base_qpn);
+	context->intra_no_vlan = 0;
+	context->no_vlan = MLX4_NO_VLAN_IDX;
+	context->intra_vlan_miss = 0;
+	context->vlan_miss = MLX4_VLAN_MISS_IDX;
+
+	in_mod = MLX4_SET_PORT_RQP_CALC << 8 | port;
+	err = mlx4_cmd(dev, mailbox->dma, in_mod, 1, MLX4_CMD_SET_PORT,
+		       MLX4_CMD_TIME_CLASS_B);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+
+int mlx4_en_DUMP_ETH_STATS(struct mlx4_en_dev *mdev, u8 port, u8 reset)
+{
+	struct mlx4_en_stat_out_mbox *mlx4_en_stats;
+	struct mlx4_en_priv *priv = netdev_priv(mdev->pndev[port]);
+	struct net_device_stats *stats = &priv->stats;
+	struct mlx4_cmd_mailbox *mailbox;
+	u64 in_mod = reset << 8 | port;
+	int err;
+	int i;
+
+	mailbox = mlx4_alloc_cmd_mailbox(mdev->dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	memset(mailbox->buf, 0, sizeof(*mlx4_en_stats));
+	err = mlx4_cmd_box(mdev->dev, 0, mailbox->dma, in_mod, 0,
+			   MLX4_CMD_DUMP_ETH_STATS, MLX4_CMD_TIME_CLASS_B);
+	if (err)
+		goto out;
+
+	mlx4_en_stats = mailbox->buf;
+
+	spin_lock_bh(&priv->stats_lock);
+
+	stats->rx_packets = 0;
+	stats->rx_bytes = 0;
+	for (i = 0; i < priv->rx_ring_num; i++) {
+		stats->rx_packets += priv->rx_ring[i].packets;
+		stats->rx_bytes += priv->rx_ring[i].bytes;
+	}
+	stats->tx_packets = 0;
+	stats->tx_bytes = 0;
+	for (i = 0; i <= priv->tx_ring_num; i++) {
+		stats->tx_packets += priv->tx_ring[i].packets;
+		stats->tx_bytes += priv->tx_ring[i].bytes;
+	}
+
+	stats->rx_errors = be64_to_cpu(mlx4_en_stats->PCS) +
+			   be32_to_cpu(mlx4_en_stats->RdropLength) +
+			   be32_to_cpu(mlx4_en_stats->RJBBR) +
+			   be32_to_cpu(mlx4_en_stats->RCRC) +
+			   be32_to_cpu(mlx4_en_stats->RRUNT);
+	stats->tx_errors = be32_to_cpu(mlx4_en_stats->TDROP);
+	stats->multicast = be64_to_cpu(mlx4_en_stats->MCAST_prio_0) +
+			   be64_to_cpu(mlx4_en_stats->MCAST_prio_1) +
+			   be64_to_cpu(mlx4_en_stats->MCAST_prio_2) +
+			   be64_to_cpu(mlx4_en_stats->MCAST_prio_3) +
+			   be64_to_cpu(mlx4_en_stats->MCAST_prio_4) +
+			   be64_to_cpu(mlx4_en_stats->MCAST_prio_5) +
+			   be64_to_cpu(mlx4_en_stats->MCAST_prio_6) +
+			   be64_to_cpu(mlx4_en_stats->MCAST_prio_7) +
+			   be64_to_cpu(mlx4_en_stats->MCAST_novlan);
+	stats->collisions = 0;
+	stats->rx_length_errors = be32_to_cpu(mlx4_en_stats->RdropLength);
+	stats->rx_over_errors = be32_to_cpu(mlx4_en_stats->RdropOvflw);
+	stats->rx_crc_errors = be32_to_cpu(mlx4_en_stats->RCRC);
+	stats->rx_frame_errors = 0;
+	stats->rx_fifo_errors = be32_to_cpu(mlx4_en_stats->RdropOvflw);
+	stats->rx_missed_errors = be32_to_cpu(mlx4_en_stats->RdropOvflw);
+	stats->tx_aborted_errors = 0;
+	stats->tx_carrier_errors = 0;
+	stats->tx_fifo_errors = 0;
+	stats->tx_heartbeat_errors = 0;
+	stats->tx_window_errors = 0;
+
+	priv->pkstats.broadcast =
+				be64_to_cpu(mlx4_en_stats->RBCAST_prio_0) +
+				be64_to_cpu(mlx4_en_stats->RBCAST_prio_1) +
+				be64_to_cpu(mlx4_en_stats->RBCAST_prio_2) +
+				be64_to_cpu(mlx4_en_stats->RBCAST_prio_3) +
+				be64_to_cpu(mlx4_en_stats->RBCAST_prio_4) +
+				be64_to_cpu(mlx4_en_stats->RBCAST_prio_5) +
+				be64_to_cpu(mlx4_en_stats->RBCAST_prio_6) +
+				be64_to_cpu(mlx4_en_stats->RBCAST_prio_7) +
+				be64_to_cpu(mlx4_en_stats->RBCAST_novlan);
+	priv->pkstats.rx_prio[0] = be64_to_cpu(mlx4_en_stats->RTOT_prio_0);
+	priv->pkstats.rx_prio[1] = be64_to_cpu(mlx4_en_stats->RTOT_prio_1);
+	priv->pkstats.rx_prio[2] = be64_to_cpu(mlx4_en_stats->RTOT_prio_2);
+	priv->pkstats.rx_prio[3] = be64_to_cpu(mlx4_en_stats->RTOT_prio_3);
+	priv->pkstats.rx_prio[4] = be64_to_cpu(mlx4_en_stats->RTOT_prio_4);
+	priv->pkstats.rx_prio[5] = be64_to_cpu(mlx4_en_stats->RTOT_prio_5);
+	priv->pkstats.rx_prio[6] = be64_to_cpu(mlx4_en_stats->RTOT_prio_6);
+	priv->pkstats.rx_prio[7] = be64_to_cpu(mlx4_en_stats->RTOT_prio_7);
+	priv->pkstats.tx_prio[0] = be64_to_cpu(mlx4_en_stats->TTOT_prio_0);
+	priv->pkstats.tx_prio[1] = be64_to_cpu(mlx4_en_stats->TTOT_prio_1);
+	priv->pkstats.tx_prio[2] = be64_to_cpu(mlx4_en_stats->TTOT_prio_2);
+	priv->pkstats.tx_prio[3] = be64_to_cpu(mlx4_en_stats->TTOT_prio_3);
+	priv->pkstats.tx_prio[4] = be64_to_cpu(mlx4_en_stats->TTOT_prio_4);
+	priv->pkstats.tx_prio[5] = be64_to_cpu(mlx4_en_stats->TTOT_prio_5);
+	priv->pkstats.tx_prio[6] = be64_to_cpu(mlx4_en_stats->TTOT_prio_6);
+	priv->pkstats.tx_prio[7] = be64_to_cpu(mlx4_en_stats->TTOT_prio_7);
+	spin_unlock_bh(&priv->stats_lock);
+
+out:
+	mlx4_free_cmd_mailbox(mdev->dev, mailbox);
+	return err;
+}
+

kernel/drivers/net/mlx4/en_port.h

@@ -0,0 +1,570 @@
+/*
+ * Copyright (c) 2007 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
+ * 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 _MLX4_EN_PORT_H_
+#define _MLX4_EN_PORT_H_
+
+
+#define SET_PORT_GEN_ALL_VALID	0x7
+#define SET_PORT_PROMISC_SHIFT	31
+
+enum {
+	MLX4_CMD_SET_VLAN_FLTR  = 0x47,
+	MLX4_CMD_SET_MCAST_FLTR = 0x48,
+	MLX4_CMD_DUMP_ETH_STATS = 0x49,
+};
+
+struct mlx4_set_port_general_context {
+	u8 reserved[3];
+	u8 flags;
+	u16 reserved2;
+	__be16 mtu;
+	u8 pptx;
+	u8 pfctx;
+	u16 reserved3;
+	u8 pprx;
+	u8 pfcrx;
+	u16 reserved4;
+};
+
+struct mlx4_set_port_rqp_calc_context {
+	__be32 base_qpn;
+	__be32 flags;
+	u8 reserved[3];
+	u8 mac_miss;
+	u8 intra_no_vlan;
+	u8 no_vlan;
+	u8 intra_vlan_miss;
+	u8 vlan_miss;
+	u8 reserved2[3];
+	u8 no_vlan_prio;
+	__be32 promisc;
+	__be32 mcast;
+};
+
+#define VLAN_FLTR_SIZE	128
+struct mlx4_set_vlan_fltr_mbox {
+	__be32 entry[VLAN_FLTR_SIZE];
+};
+
+
+enum {
+	MLX4_MCAST_CONFIG       = 0,
+	MLX4_MCAST_DISABLE      = 1,
+	MLX4_MCAST_ENABLE       = 2,
+};
+
+
+struct mlx4_en_stat_out_mbox {
+	/* Received frames with a length of 64 octets */
+	__be64 R64_prio_0;
+	__be64 R64_prio_1;
+	__be64 R64_prio_2;
+	__be64 R64_prio_3;
+	__be64 R64_prio_4;
+	__be64 R64_prio_5;
+	__be64 R64_prio_6;
+	__be64 R64_prio_7;
+	__be64 R64_novlan;
+	/* Received frames with a length of 127 octets */
+	__be64 R127_prio_0;
+	__be64 R127_prio_1;
+	__be64 R127_prio_2;
+	__be64 R127_prio_3;
+	__be64 R127_prio_4;
+	__be64 R127_prio_5;
+	__be64 R127_prio_6;
+	__be64 R127_prio_7;
+	__be64 R127_novlan;
+	/* Received frames with a length of 255 octets */
+	__be64 R255_prio_0;
+	__be64 R255_prio_1;
+	__be64 R255_prio_2;
+	__be64 R255_prio_3;
+	__be64 R255_prio_4;
+	__be64 R255_prio_5;
+	__be64 R255_prio_6;
+	__be64 R255_prio_7;
+	__be64 R255_novlan;
+	/* Received frames with a length of 511 octets */
+	__be64 R511_prio_0;
+	__be64 R511_prio_1;
+	__be64 R511_prio_2;
+	__be64 R511_prio_3;
+	__be64 R511_prio_4;
+	__be64 R511_prio_5;
+	__be64 R511_prio_6;
+	__be64 R511_prio_7;
+	__be64 R511_novlan;
+	/* Received frames with a length of 1023 octets */
+	__be64 R1023_prio_0;
+	__be64 R1023_prio_1;
+	__be64 R1023_prio_2;
+	__be64 R1023_prio_3;
+	__be64 R1023_prio_4;
+	__be64 R1023_prio_5;
+	__be64 R1023_prio_6;
+	__be64 R1023_prio_7;
+	__be64 R1023_novlan;
+	/* Received frames with a length of 1518 octets */
+	__be64 R1518_prio_0;
+	__be64 R1518_prio_1;
+	__be64 R1518_prio_2;
+	__be64 R1518_prio_3;
+	__be64 R1518_prio_4;
+	__be64 R1518_prio_5;
+	__be64 R1518_prio_6;
+	__be64 R1518_prio_7;
+	__be64 R1518_novlan;
+	/* Received frames with a length of 1522 octets */
+	__be64 R1522_prio_0;
+	__be64 R1522_prio_1;
+	__be64 R1522_prio_2;
+	__be64 R1522_prio_3;
+	__be64 R1522_prio_4;
+	__be64 R1522_prio_5;
+	__be64 R1522_prio_6;
+	__be64 R1522_prio_7;
+	__be64 R1522_novlan;
+	/* Received frames with a length of 1548 octets */
+	__be64 R1548_prio_0;
+	__be64 R1548_prio_1;
+	__be64 R1548_prio_2;
+	__be64 R1548_prio_3;
+	__be64 R1548_prio_4;
+	__be64 R1548_prio_5;
+	__be64 R1548_prio_6;
+	__be64 R1548_prio_7;
+	__be64 R1548_novlan;
+	/* Received frames with a length of 1548 < octets < MTU */
+	__be64 R2MTU_prio_0;
+	__be64 R2MTU_prio_1;
+	__be64 R2MTU_prio_2;
+	__be64 R2MTU_prio_3;
+	__be64 R2MTU_prio_4;
+	__be64 R2MTU_prio_5;
+	__be64 R2MTU_prio_6;
+	__be64 R2MTU_prio_7;
+	__be64 R2MTU_novlan;
+	/* Received frames with a length of MTU< octets and good CRC */
+	__be64 RGIANT_prio_0;
+	__be64 RGIANT_prio_1;
+	__be64 RGIANT_prio_2;
+	__be64 RGIANT_prio_3;
+	__be64 RGIANT_prio_4;
+	__be64 RGIANT_prio_5;
+	__be64 RGIANT_prio_6;
+	__be64 RGIANT_prio_7;
+	__be64 RGIANT_novlan;
+	/* Received broadcast frames with good CRC */
+	__be64 RBCAST_prio_0;
+	__be64 RBCAST_prio_1;
+	__be64 RBCAST_prio_2;
+	__be64 RBCAST_prio_3;
+	__be64 RBCAST_prio_4;
+	__be64 RBCAST_prio_5;
+	__be64 RBCAST_prio_6;
+	__be64 RBCAST_prio_7;
+	__be64 RBCAST_novlan;
+	/* Received multicast frames with good CRC */
+	__be64 MCAST_prio_0;
+	__be64 MCAST_prio_1;
+	__be64 MCAST_prio_2;
+	__be64 MCAST_prio_3;
+	__be64 MCAST_prio_4;
+	__be64 MCAST_prio_5;
+	__be64 MCAST_prio_6;
+	__be64 MCAST_prio_7;
+	__be64 MCAST_novlan;
+	/* Received unicast not short or GIANT frames with good CRC */
+	__be64 RTOTG_prio_0;
+	__be64 RTOTG_prio_1;
+	__be64 RTOTG_prio_2;
+	__be64 RTOTG_prio_3;
+	__be64 RTOTG_prio_4;
+	__be64 RTOTG_prio_5;
+	__be64 RTOTG_prio_6;
+	__be64 RTOTG_prio_7;
+	__be64 RTOTG_novlan;
+
+	/* Count of total octets of received frames, includes framing characters */
+	__be64 RTTLOCT_prio_0;
+	/* Count of total octets of received frames, not including framing
+	   characters */
+	__be64 RTTLOCT_NOFRM_prio_0;
+	/* Count of Total number of octets received
+	   (only for frames without errors) */
+	__be64 ROCT_prio_0;
+
+	__be64 RTTLOCT_prio_1;
+	__be64 RTTLOCT_NOFRM_prio_1;
+	__be64 ROCT_prio_1;
+
+	__be64 RTTLOCT_prio_2;
+	__be64 RTTLOCT_NOFRM_prio_2;
+	__be64 ROCT_prio_2;
+
+	__be64 RTTLOCT_prio_3;
+	__be64 RTTLOCT_NOFRM_prio_3;
+	__be64 ROCT_prio_3;
+
+	__be64 RTTLOCT_prio_4;
+	__be64 RTTLOCT_NOFRM_prio_4;
+	__be64 ROCT_prio_4;
+
+	__be64 RTTLOCT_prio_5;
+	__be64 RTTLOCT_NOFRM_prio_5;
+	__be64 ROCT_prio_5;
+
+	__be64 RTTLOCT_prio_6;
+	__be64 RTTLOCT_NOFRM_prio_6;
+	__be64 ROCT_prio_6;
+
+	__be64 RTTLOCT_prio_7;
+	__be64 RTTLOCT_NOFRM_prio_7;
+	__be64 ROCT_prio_7;
+
+	__be64 RTTLOCT_novlan;
+	__be64 RTTLOCT_NOFRM_novlan;
+	__be64 ROCT_novlan;
+
+	/* Count of Total received frames including bad frames */
+	__be64 RTOT_prio_0;
+	/* Count of  Total number of received frames with 802.1Q encapsulation */
+	__be64 R1Q_prio_0;
+	__be64 reserved1;
+
+	__be64 RTOT_prio_1;
+	__be64 R1Q_prio_1;
+	__be64 reserved2;
+
+	__be64 RTOT_prio_2;
+	__be64 R1Q_prio_2;
+	__be64 reserved3;
+
+	__be64 RTOT_prio_3;
+	__be64 R1Q_prio_3;
+	__be64 reserved4;
+
+	__be64 RTOT_prio_4;
+	__be64 R1Q_prio_4;
+	__be64 reserved5;
+
+	__be64 RTOT_prio_5;
+	__be64 R1Q_prio_5;
+	__be64 reserved6;
+
+	__be64 RTOT_prio_6;
+	__be64 R1Q_prio_6;
+	__be64 reserved7;
+
+	__be64 RTOT_prio_7;
+	__be64 R1Q_prio_7;
+	__be64 reserved8;
+
+	__be64 RTOT_novlan;
+	__be64 R1Q_novlan;
+	__be64 reserved9;
+
+	/* Total number of Successfully Received Control Frames */
+	__be64 RCNTL;
+	__be64 reserved10;
+	__be64 reserved11;
+	__be64 reserved12;
+	/* Count of received frames with a length/type field  value between 46
+	   (42 for VLANtagged frames) and 1500 (also 1500 for VLAN-tagged frames),
+	   inclusive */
+	__be64 RInRangeLengthErr;
+	/* Count of received frames with length/type field between 1501 and 1535
+	   decimal, inclusive */
+	__be64 ROutRangeLengthErr;
+	/* Count of received frames that are longer than max allowed size for
+	   802.3 frames (1518/1522) */
+	__be64 RFrmTooLong;
+	/* Count frames received with PCS error */
+	__be64 PCS;
+
+	/* Transmit frames with a length of 64 octets */
+	__be64 T64_prio_0;
+	__be64 T64_prio_1;
+	__be64 T64_prio_2;
+	__be64 T64_prio_3;
+	__be64 T64_prio_4;
+	__be64 T64_prio_5;
+	__be64 T64_prio_6;
+	__be64 T64_prio_7;
+	__be64 T64_novlan;
+	__be64 T64_loopbk;
+	/* Transmit frames with a length of 65 to 127 octets. */
+	__be64 T127_prio_0;
+	__be64 T127_prio_1;
+	__be64 T127_prio_2;
+	__be64 T127_prio_3;
+	__be64 T127_prio_4;
+	__be64 T127_prio_5;
+	__be64 T127_prio_6;
+	__be64 T127_prio_7;
+	__be64 T127_novlan;
+	__be64 T127_loopbk;
+	/* Transmit frames with a length of 128 to 255 octets */
+	__be64 T255_prio_0;
+	__be64 T255_prio_1;
+	__be64 T255_prio_2;
+	__be64 T255_prio_3;
+	__be64 T255_prio_4;
+	__be64 T255_prio_5;
+	__be64 T255_prio_6;
+	__be64 T255_prio_7;
+	__be64 T255_novlan;
+	__be64 T255_loopbk;
+	/* Transmit frames with a length of 256 to 511 octets */
+	__be64 T511_prio_0;
+	__be64 T511_prio_1;
+	__be64 T511_prio_2;
+	__be64 T511_prio_3;
+	__be64 T511_prio_4;
+	__be64 T511_prio_5;
+	__be64 T511_prio_6;
+	__be64 T511_prio_7;
+	__be64 T511_novlan;
+	__be64 T511_loopbk;
+	/* Transmit frames with a length of 512 to 1023 octets */
+	__be64 T1023_prio_0;
+	__be64 T1023_prio_1;
+	__be64 T1023_prio_2;
+	__be64 T1023_prio_3;
+	__be64 T1023_prio_4;
+	__be64 T1023_prio_5;
+	__be64 T1023_prio_6;
+	__be64 T1023_prio_7;
+	__be64 T1023_novlan;
+	__be64 T1023_loopbk;
+	/* Transmit frames with a length of 1024 to 1518 octets */
+	__be64 T1518_prio_0;
+	__be64 T1518_prio_1;
+	__be64 T1518_prio_2;
+	__be64 T1518_prio_3;
+	__be64 T1518_prio_4;
+	__be64 T1518_prio_5;
+	__be64 T1518_prio_6;
+	__be64 T1518_prio_7;
+	__be64 T1518_novlan;
+	__be64 T1518_loopbk;
+	/* Counts transmit frames with a length of 1519 to 1522 bytes */
+	__be64 T1522_prio_0;
+	__be64 T1522_prio_1;
+	__be64 T1522_prio_2;
+	__be64 T1522_prio_3;
+	__be64 T1522_prio_4;
+	__be64 T1522_prio_5;
+	__be64 T1522_prio_6;
+	__be64 T1522_prio_7;
+	__be64 T1522_novlan;
+	__be64 T1522_loopbk;
+	/* Transmit frames with a length of 1523 to 1548 octets */
+	__be64 T1548_prio_0;
+	__be64 T1548_prio_1;
+	__be64 T1548_prio_2;
+	__be64 T1548_prio_3;
+	__be64 T1548_prio_4;
+	__be64 T1548_prio_5;
+	__be64 T1548_prio_6;
+	__be64 T1548_prio_7;
+	__be64 T1548_novlan;
+	__be64 T1548_loopbk;
+	/* Counts transmit frames with a length of 1549 to MTU bytes */
+	__be64 T2MTU_prio_0;
+	__be64 T2MTU_prio_1;
+	__be64 T2MTU_prio_2;
+	__be64 T2MTU_prio_3;
+	__be64 T2MTU_prio_4;
+	__be64 T2MTU_prio_5;
+	__be64 T2MTU_prio_6;
+	__be64 T2MTU_prio_7;
+	__be64 T2MTU_novlan;
+	__be64 T2MTU_loopbk;
+	/* Transmit frames with a length greater than MTU octets and a good CRC. */
+	__be64 TGIANT_prio_0;
+	__be64 TGIANT_prio_1;
+	__be64 TGIANT_prio_2;
+	__be64 TGIANT_prio_3;
+	__be64 TGIANT_prio_4;
+	__be64 TGIANT_prio_5;
+	__be64 TGIANT_prio_6;
+	__be64 TGIANT_prio_7;
+	__be64 TGIANT_novlan;
+	__be64 TGIANT_loopbk;
+	/* Transmit broadcast frames with a good CRC */
+	__be64 TBCAST_prio_0;
+	__be64 TBCAST_prio_1;
+	__be64 TBCAST_prio_2;
+	__be64 TBCAST_prio_3;
+	__be64 TBCAST_prio_4;
+	__be64 TBCAST_prio_5;
+	__be64 TBCAST_prio_6;
+	__be64 TBCAST_prio_7;
+	__be64 TBCAST_novlan;
+	__be64 TBCAST_loopbk;
+	/* Transmit multicast frames with a good CRC */
+	__be64 TMCAST_prio_0;
+	__be64 TMCAST_prio_1;
+	__be64 TMCAST_prio_2;
+	__be64 TMCAST_prio_3;
+	__be64 TMCAST_prio_4;
+	__be64 TMCAST_prio_5;
+	__be64 TMCAST_prio_6;
+	__be64 TMCAST_prio_7;
+	__be64 TMCAST_novlan;
+	__be64 TMCAST_loopbk;
+	/* Transmit good frames that are neither broadcast nor multicast */
+	__be64 TTOTG_prio_0;
+	__be64 TTOTG_prio_1;
+	__be64 TTOTG_prio_2;
+	__be64 TTOTG_prio_3;
+	__be64 TTOTG_prio_4;
+	__be64 TTOTG_prio_5;
+	__be64 TTOTG_prio_6;
+	__be64 TTOTG_prio_7;
+	__be64 TTOTG_novlan;
+	__be64 TTOTG_loopbk;
+
+	/* total octets of transmitted frames, including framing characters */
+	__be64 TTTLOCT_prio_0;
+	/* total octets of transmitted frames, not including framing characters */
+	__be64 TTTLOCT_NOFRM_prio_0;
+	/* ifOutOctets */
+	__be64 TOCT_prio_0;
+
+	__be64 TTTLOCT_prio_1;
+	__be64 TTTLOCT_NOFRM_prio_1;
+	__be64 TOCT_prio_1;
+
+	__be64 TTTLOCT_prio_2;
+	__be64 TTTLOCT_NOFRM_prio_2;
+	__be64 TOCT_prio_2;
+
+	__be64 TTTLOCT_prio_3;
+	__be64 TTTLOCT_NOFRM_prio_3;
+	__be64 TOCT_prio_3;
+
+	__be64 TTTLOCT_prio_4;
+	__be64 TTTLOCT_NOFRM_prio_4;
+	__be64 TOCT_prio_4;
+
+	__be64 TTTLOCT_prio_5;
+	__be64 TTTLOCT_NOFRM_prio_5;
+	__be64 TOCT_prio_5;
+
+	__be64 TTTLOCT_prio_6;
+	__be64 TTTLOCT_NOFRM_prio_6;
+	__be64 TOCT_prio_6;
+
+	__be64 TTTLOCT_prio_7;
+	__be64 TTTLOCT_NOFRM_prio_7;
+	__be64 TOCT_prio_7;
+
+	__be64 TTTLOCT_novlan;
+	__be64 TTTLOCT_NOFRM_novlan;
+	__be64 TOCT_novlan;
+
+	__be64 TTTLOCT_loopbk;
+	__be64 TTTLOCT_NOFRM_loopbk;
+	__be64 TOCT_loopbk;
+
+	/* Total frames transmitted with a good CRC that are not aborted  */
+	__be64 TTOT_prio_0;
+	/* Total number of frames transmitted with 802.1Q encapsulation */
+	__be64 T1Q_prio_0;
+	__be64 reserved13;
+
+	__be64 TTOT_prio_1;
+	__be64 T1Q_prio_1;
+	__be64 reserved14;
+
+	__be64 TTOT_prio_2;
+	__be64 T1Q_prio_2;
+	__be64 reserved15;
+
+	__be64 TTOT_prio_3;
+	__be64 T1Q_prio_3;
+	__be64 reserved16;
+
+	__be64 TTOT_prio_4;
+	__be64 T1Q_prio_4;
+	__be64 reserved17;
+
+	__be64 TTOT_prio_5;
+	__be64 T1Q_prio_5;
+	__be64 reserved18;
+
+	__be64 TTOT_prio_6;
+	__be64 T1Q_prio_6;
+	__be64 reserved19;
+
+	__be64 TTOT_prio_7;
+	__be64 T1Q_prio_7;
+	__be64 reserved20;
+
+	__be64 TTOT_novlan;
+	__be64 T1Q_novlan;
+	__be64 reserved21;
+
+	__be64 TTOT_loopbk;
+	__be64 T1Q_loopbk;
+	__be64 reserved22;
+
+	/* Received frames with a length greater than MTU octets and a bad CRC */
+	__be32 RJBBR;
+	/* Received frames with a bad CRC that are not runts, jabbers,
+	   or alignment errors */
+	__be32 RCRC;
+	/* Received frames with SFD with a length of less than 64 octets and a
+	   bad CRC */
+	__be32 RRUNT;
+	/* Received frames with a length less than 64 octets and a good CRC */
+	__be32 RSHORT;
+	/* Total Number of Received Packets Dropped */
+	__be32 RDROP;
+	/* Drop due to overflow  */
+	__be32 RdropOvflw;
+	/* Drop due to overflow */
+	__be32 RdropLength;
+	/* Total of good frames. Does not include frames received with
+	   frame-too-long, FCS, or length errors */
+	__be32 RTOTFRMS;
+	/* Total dropped Xmited packets */
+	__be32 TDROP;
+};
+
+
+#endif

kernel/drivers/net/mlx4/en_resources.c

@@ -0,0 +1,101 @@
+/*
+ * Copyright (c) 2007 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
+ * 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/vmalloc.h>
+#include <linux/mlx4/qp.h>
+
+#include "mlx4_en.h"
+
+void mlx4_en_fill_qp_context(struct mlx4_en_priv *priv, int size, int stride,
+			     int is_tx, int rss, int qpn, int cqn,
+			     struct mlx4_qp_context *context)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+
+	memset(context, 0, sizeof *context);
+	context->flags = cpu_to_be32(7 << 16 | rss << 13);
+	context->pd = cpu_to_be32(mdev->priv_pdn);
+	context->mtu_msgmax = 0xff;
+	if (!is_tx && !rss)
+		context->rq_size_stride = ilog2(size) << 3 | (ilog2(stride) - 4);
+	if (is_tx)
+		context->sq_size_stride = ilog2(size) << 3 | (ilog2(stride) - 4);
+	else
+		context->sq_size_stride = ilog2(TXBB_SIZE) - 4;
+	context->usr_page = cpu_to_be32(mdev->priv_uar.index);
+	context->local_qpn = cpu_to_be32(qpn);
+	context->pri_path.ackto = 1 & 0x07;
+	context->pri_path.sched_queue = 0x83 | (priv->port - 1) << 6;
+	context->pri_path.counter_index = 0xff;
+	context->cqn_send = cpu_to_be32(cqn);
+	context->cqn_recv = cpu_to_be32(cqn);
+	context->db_rec_addr = cpu_to_be64(priv->res.db.dma << 2);
+}
+
+
+int mlx4_en_map_buffer(struct mlx4_buf *buf)
+{
+	struct page **pages;
+	int i;
+
+	if (BITS_PER_LONG == 64 || buf->nbufs == 1)
+		return 0;
+
+	pages = kmalloc(sizeof *pages * buf->nbufs, GFP_KERNEL);
+	if (!pages)
+		return -ENOMEM;
+
+	for (i = 0; i < buf->nbufs; ++i)
+		pages[i] = virt_to_page(buf->page_list[i].buf);
+
+	buf->direct.buf = vmap(pages, buf->nbufs, VM_MAP, PAGE_KERNEL);
+	kfree(pages);
+	if (!buf->direct.buf)
+		return -ENOMEM;
+
+	return 0;
+}
+
+void mlx4_en_unmap_buffer(struct mlx4_buf *buf)
+{
+	if (BITS_PER_LONG == 64 || buf->nbufs == 1)
+		return;
+
+	vunmap(buf->direct.buf);
+}
+
+void mlx4_en_sqp_event(struct mlx4_qp *qp, enum mlx4_event event)
+{
+    return;
+}
+

kernel/drivers/net/mlx4/en_rx.c

@@ -0,0 +1,940 @@
+/*
+ * Copyright (c) 2007 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
+ * 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/mlx4/cq.h>
+#include <linux/mlx4/qp.h>
+#include <linux/skbuff.h>
+#include <linux/if_ether.h>
+#include <linux/if_vlan.h>
+#include <linux/vmalloc.h>
+
+#include "mlx4_en.h"
+
+
+static int mlx4_en_get_frag_header(struct skb_frag_struct *frags, void **mac_hdr,
+				   void **ip_hdr, void **tcpudp_hdr,
+				   u64 *hdr_flags, void *priv)
+{
+	*mac_hdr = page_address(frags->page) + frags->page_offset;
+	*ip_hdr = *mac_hdr + ETH_HLEN;
+	*tcpudp_hdr = (struct tcphdr *)(*ip_hdr + sizeof(struct iphdr));
+	*hdr_flags = LRO_IPV4 | LRO_TCP;
+
+	return 0;
+}
+
+static int mlx4_en_alloc_frag(struct mlx4_en_priv *priv,
+			      struct mlx4_en_rx_desc *rx_desc,
+			      struct skb_frag_struct *skb_frags,
+			      struct mlx4_en_rx_alloc *ring_alloc,
+			      int i)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_en_frag_info *frag_info = &priv->frag_info[i];
+	struct mlx4_en_rx_alloc *page_alloc = &ring_alloc[i];
+	struct page *page;
+	dma_addr_t dma;
+
+	if (page_alloc->offset == frag_info->last_offset) {
+		/* Allocate new page */
+		page = alloc_pages(GFP_ATOMIC | __GFP_COMP, MLX4_EN_ALLOC_ORDER);
+		if (!page)
+			return -ENOMEM;
+
+		skb_frags[i].page = page_alloc->page;
+		skb_frags[i].page_offset = page_alloc->offset;
+		page_alloc->page = page;
+		page_alloc->offset = frag_info->frag_align;
+	} else {
+		page = page_alloc->page;
+		get_page(page);
+
+		skb_frags[i].page = page;
+		skb_frags[i].page_offset = page_alloc->offset;
+		page_alloc->offset += frag_info->frag_stride;
+	}
+	dma = pci_map_single(mdev->pdev, page_address(skb_frags[i].page) +
+			     skb_frags[i].page_offset, frag_info->frag_size,
+			     PCI_DMA_FROMDEVICE);
+	rx_desc->data[i].addr = cpu_to_be64(dma);
+	return 0;
+}
+
+static int mlx4_en_init_allocator(struct mlx4_en_priv *priv,
+				  struct mlx4_en_rx_ring *ring)
+{
+	struct mlx4_en_rx_alloc *page_alloc;
+	int i;
+
+	for (i = 0; i < priv->num_frags; i++) {
+		page_alloc = &ring->page_alloc[i];
+		page_alloc->page = alloc_pages(GFP_ATOMIC | __GFP_COMP,
+					       MLX4_EN_ALLOC_ORDER);
+		if (!page_alloc->page)
+			goto out;
+
+		page_alloc->offset = priv->frag_info[i].frag_align;
+		en_dbg(DRV, priv, "Initialized allocator:%d with page:%p\n",
+		       i, page_alloc->page);
+	}
+	return 0;
+
+out:
+	while (i--) {
+		page_alloc = &ring->page_alloc[i];
+		put_page(page_alloc->page);
+		page_alloc->page = NULL;
+	}
+	return -ENOMEM;
+}
+
+static void mlx4_en_destroy_allocator(struct mlx4_en_priv *priv,
+				      struct mlx4_en_rx_ring *ring)
+{
+	struct mlx4_en_rx_alloc *page_alloc;
+	int i;
+
+	for (i = 0; i < priv->num_frags; i++) {
+		page_alloc = &ring->page_alloc[i];
+		en_dbg(DRV, priv, "Freeing allocator:%d count:%d\n",
+		       i, page_count(page_alloc->page));
+
+		put_page(page_alloc->page);
+		page_alloc->page = NULL;
+	}
+}
+
+
+static void mlx4_en_init_rx_desc(struct mlx4_en_priv *priv,
+				 struct mlx4_en_rx_ring *ring, int index)
+{
+	struct mlx4_en_rx_desc *rx_desc = ring->buf + ring->stride * index;
+	struct skb_frag_struct *skb_frags = ring->rx_info +
+					    (index << priv->log_rx_info);
+	int possible_frags;
+	int i;
+
+	/* Set size and memtype fields */
+	for (i = 0; i < priv->num_frags; i++) {
+		skb_frags[i].size = priv->frag_info[i].frag_size;
+		rx_desc->data[i].byte_count =
+			cpu_to_be32(priv->frag_info[i].frag_size);
+		rx_desc->data[i].lkey = cpu_to_be32(priv->mdev->mr.key);
+	}
+
+	/* If the number of used fragments does not fill up the ring stride,
+	 * remaining (unused) fragments must be padded with null address/size
+	 * and a special memory key */
+	possible_frags = (ring->stride - sizeof(struct mlx4_en_rx_desc)) / DS_SIZE;
+	for (i = priv->num_frags; i < possible_frags; i++) {
+		rx_desc->data[i].byte_count = 0;
+		rx_desc->data[i].lkey = cpu_to_be32(MLX4_EN_MEMTYPE_PAD);
+		rx_desc->data[i].addr = 0;
+	}
+}
+
+
+static int mlx4_en_prepare_rx_desc(struct mlx4_en_priv *priv,
+				   struct mlx4_en_rx_ring *ring, int index)
+{
+	struct mlx4_en_rx_desc *rx_desc = ring->buf + (index * ring->stride);
+	struct skb_frag_struct *skb_frags = ring->rx_info +
+					    (index << priv->log_rx_info);
+	int i;
+
+	for (i = 0; i < priv->num_frags; i++)
+		if (mlx4_en_alloc_frag(priv, rx_desc, skb_frags, ring->page_alloc, i))
+			goto err;
+
+	return 0;
+
+err:
+	while (i--)
+		put_page(skb_frags[i].page);
+	return -ENOMEM;
+}
+
+static inline void mlx4_en_update_rx_prod_db(struct mlx4_en_rx_ring *ring)
+{
+	*ring->wqres.db.db = cpu_to_be32(ring->prod & 0xffff);
+}
+
+static void mlx4_en_free_rx_desc(struct mlx4_en_priv *priv,
+				 struct mlx4_en_rx_ring *ring,
+				 int index)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct skb_frag_struct *skb_frags;
+	struct mlx4_en_rx_desc *rx_desc = ring->buf + (index << ring->log_stride);
+	dma_addr_t dma;
+	int nr;
+
+	skb_frags = ring->rx_info + (index << priv->log_rx_info);
+	for (nr = 0; nr < priv->num_frags; nr++) {
+		en_dbg(DRV, priv, "Freeing fragment:%d\n", nr);
+		dma = be64_to_cpu(rx_desc->data[nr].addr);
+
+		en_dbg(DRV, priv, "Unmaping buffer at dma:0x%llx\n", (u64) dma);
+		pci_unmap_single(mdev->pdev, dma, skb_frags[nr].size,
+				 PCI_DMA_FROMDEVICE);
+		put_page(skb_frags[nr].page);
+	}
+}
+
+static int mlx4_en_fill_rx_buffers(struct mlx4_en_priv *priv)
+{
+	struct mlx4_en_rx_ring *ring;
+	int ring_ind;
+	int buf_ind;
+	int new_size;
+
+	for (buf_ind = 0; buf_ind < priv->prof->rx_ring_size; buf_ind++) {
+		for (ring_ind = 0; ring_ind < priv->rx_ring_num; ring_ind++) {
+			ring = &priv->rx_ring[ring_ind];
+
+			if (mlx4_en_prepare_rx_desc(priv, ring,
+						    ring->actual_size)) {
+				if (ring->actual_size < MLX4_EN_MIN_RX_SIZE) {
+					en_err(priv, "Failed to allocate "
+						     "enough rx buffers\n");
+					return -ENOMEM;
+				} else {
+					new_size = rounddown_pow_of_two(ring->actual_size);
+					en_warn(priv, "Only %d buffers allocated "
+						      "reducing ring size to %d",
+						ring->actual_size, new_size);
+					goto reduce_rings;
+				}
+			}
+			ring->actual_size++;
+			ring->prod++;
+		}
+	}
+	return 0;
+
+reduce_rings:
+	for (ring_ind = 0; ring_ind < priv->rx_ring_num; ring_ind++) {
+		ring = &priv->rx_ring[ring_ind];
+		while (ring->actual_size > new_size) {
+			ring->actual_size--;
+			ring->prod--;
+			mlx4_en_free_rx_desc(priv, ring, ring->actual_size);
+		}
+		ring->size_mask = ring->actual_size - 1;
+	}
+
+	return 0;
+}
+
+static void mlx4_en_free_rx_buf(struct mlx4_en_priv *priv,
+				struct mlx4_en_rx_ring *ring)
+{
+	int index;
+
+	en_dbg(DRV, priv, "Freeing Rx buf - cons:%d prod:%d\n",
+	       ring->cons, ring->prod);
+
+	/* Unmap and free Rx buffers */
+	BUG_ON((u32) (ring->prod - ring->cons) > ring->actual_size);
+	while (ring->cons != ring->prod) {
+		index = ring->cons & ring->size_mask;
+		en_dbg(DRV, priv, "Processing descriptor:%d\n", index);
+		mlx4_en_free_rx_desc(priv, ring, index);
+		++ring->cons;
+	}
+}
+
+int mlx4_en_create_rx_ring(struct mlx4_en_priv *priv,
+			   struct mlx4_en_rx_ring *ring, u32 size, u16 stride)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	int err;
+	int tmp;
+
+
+	ring->prod = 0;
+	ring->cons = 0;
+	ring->size = size;
+	ring->size_mask = size - 1;
+	ring->stride = stride;
+	ring->log_stride = ffs(ring->stride) - 1;
+	ring->buf_size = ring->size * ring->stride + TXBB_SIZE;
+
+	tmp = size * roundup_pow_of_two(MLX4_EN_MAX_RX_FRAGS *
+					sizeof(struct skb_frag_struct));
+	ring->rx_info = vmalloc(tmp);
+	if (!ring->rx_info) {
+		en_err(priv, "Failed allocating rx_info ring\n");
+		return -ENOMEM;
+	}
+	en_dbg(DRV, priv, "Allocated rx_info ring at addr:%p size:%d\n",
+		 ring->rx_info, tmp);
+
+	err = mlx4_alloc_hwq_res(mdev->dev, &ring->wqres,
+				 ring->buf_size, 2 * PAGE_SIZE);
+	if (err)
+		goto err_ring;
+
+	err = mlx4_en_map_buffer(&ring->wqres.buf);
+	if (err) {
+		en_err(priv, "Failed to map RX buffer\n");
+		goto err_hwq;
+	}
+	ring->buf = ring->wqres.buf.direct.buf;
+
+	/* Configure lro mngr */
+	memset(&ring->lro, 0, sizeof(struct net_lro_mgr));
+	ring->lro.dev = priv->dev;
+	ring->lro.features = LRO_F_NAPI;
+	ring->lro.frag_align_pad = NET_IP_ALIGN;
+	ring->lro.ip_summed = CHECKSUM_UNNECESSARY;
+	ring->lro.ip_summed_aggr = CHECKSUM_UNNECESSARY;
+	ring->lro.max_desc = mdev->profile.num_lro;
+	ring->lro.max_aggr = MAX_SKB_FRAGS;
+	ring->lro.lro_arr = kzalloc(mdev->profile.num_lro *
+				    sizeof(struct net_lro_desc),
+				    GFP_KERNEL);
+	if (!ring->lro.lro_arr) {
+		en_err(priv, "Failed to allocate lro array\n");
+		goto err_map;
+	}
+	ring->lro.get_frag_header = mlx4_en_get_frag_header;
+
+	return 0;
+
+err_map:
+	mlx4_en_unmap_buffer(&ring->wqres.buf);
+err_hwq:
+	mlx4_free_hwq_res(mdev->dev, &ring->wqres, ring->buf_size);
+err_ring:
+	vfree(ring->rx_info);
+	ring->rx_info = NULL;
+	return err;
+}
+
+int mlx4_en_activate_rx_rings(struct mlx4_en_priv *priv)
+{
+	struct mlx4_en_rx_ring *ring;
+	int i;
+	int ring_ind;
+	int err;
+	int stride = roundup_pow_of_two(sizeof(struct mlx4_en_rx_desc) +
+					DS_SIZE * priv->num_frags);
+
+	for (ring_ind = 0; ring_ind < priv->rx_ring_num; ring_ind++) {
+		ring = &priv->rx_ring[ring_ind];
+
+		ring->prod = 0;
+		ring->cons = 0;
+		ring->actual_size = 0;
+		ring->cqn = priv->rx_cq[ring_ind].mcq.cqn;
+
+		ring->stride = stride;
+		if (ring->stride <= TXBB_SIZE)
+			ring->buf += TXBB_SIZE;
+
+		ring->log_stride = ffs(ring->stride) - 1;
+		ring->buf_size = ring->size * ring->stride;
+
+		memset(ring->buf, 0, ring->buf_size);
+		mlx4_en_update_rx_prod_db(ring);
+
+		/* Initailize all descriptors */
+		for (i = 0; i < ring->size; i++)
+			mlx4_en_init_rx_desc(priv, ring, i);
+
+		/* Initialize page allocators */
+		err = mlx4_en_init_allocator(priv, ring);
+		if (err) {
+			en_err(priv, "Failed initializing ring allocator\n");
+			ring_ind--;
+			goto err_allocator;
+		}
+	}
+	err = mlx4_en_fill_rx_buffers(priv);
+	if (err)
+		goto err_buffers;
+
+	for (ring_ind = 0; ring_ind < priv->rx_ring_num; ring_ind++) {
+		ring = &priv->rx_ring[ring_ind];
+
+		mlx4_en_update_rx_prod_db(ring);
+	}
+
+	return 0;
+
+err_buffers:
+	for (ring_ind = 0; ring_ind < priv->rx_ring_num; ring_ind++)
+		mlx4_en_free_rx_buf(priv, &priv->rx_ring[ring_ind]);
+
+	ring_ind = priv->rx_ring_num - 1;
+err_allocator:
+	while (ring_ind >= 0) {
+		mlx4_en_destroy_allocator(priv, &priv->rx_ring[ring_ind]);
+		ring_ind--;
+	}
+	return err;
+}
+
+void mlx4_en_destroy_rx_ring(struct mlx4_en_priv *priv,
+			     struct mlx4_en_rx_ring *ring)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+
+	kfree(ring->lro.lro_arr);
+	mlx4_en_unmap_buffer(&ring->wqres.buf);
+	mlx4_free_hwq_res(mdev->dev, &ring->wqres, ring->buf_size + TXBB_SIZE);
+	vfree(ring->rx_info);
+	ring->rx_info = NULL;
+}
+
+void mlx4_en_deactivate_rx_ring(struct mlx4_en_priv *priv,
+				struct mlx4_en_rx_ring *ring)
+{
+	mlx4_en_free_rx_buf(priv, ring);
+	if (ring->stride <= TXBB_SIZE)
+		ring->buf -= TXBB_SIZE;
+	mlx4_en_destroy_allocator(priv, ring);
+}
+
+
+/* Unmap a completed descriptor and free unused pages */
+static int mlx4_en_complete_rx_desc(struct mlx4_en_priv *priv,
+				    struct mlx4_en_rx_desc *rx_desc,
+				    struct skb_frag_struct *skb_frags,
+				    struct skb_frag_struct *skb_frags_rx,
+				    struct mlx4_en_rx_alloc *page_alloc,
+				    int length)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_en_frag_info *frag_info;
+	int nr;
+	dma_addr_t dma;
+
+	/* Collect used fragments while replacing them in the HW descirptors */
+	for (nr = 0; nr < priv->num_frags; nr++) {
+		frag_info = &priv->frag_info[nr];
+		if (length <= frag_info->frag_prefix_size)
+			break;
+
+		/* Save page reference in skb */
+		skb_frags_rx[nr].page = skb_frags[nr].page;
+		skb_frags_rx[nr].size = skb_frags[nr].size;
+		skb_frags_rx[nr].page_offset = skb_frags[nr].page_offset;
+		dma = be64_to_cpu(rx_desc->data[nr].addr);
+
+		/* Allocate a replacement page */
+		if (mlx4_en_alloc_frag(priv, rx_desc, skb_frags, page_alloc, nr))
+			goto fail;
+
+		/* Unmap buffer */
+		pci_unmap_single(mdev->pdev, dma, skb_frags[nr].size,
+				 PCI_DMA_FROMDEVICE);
+	}
+	/* Adjust size of last fragment to match actual length */
+	if (nr > 0)
+		skb_frags_rx[nr - 1].size = length -
+			priv->frag_info[nr - 1].frag_prefix_size;
+	return nr;
+
+fail:
+	/* Drop all accumulated fragments (which have already been replaced in
+	 * the descriptor) of this packet; remaining fragments are reused... */
+	while (nr > 0) {
+		nr--;
+		put_page(skb_frags_rx[nr].page);
+	}
+	return 0;
+}
+
+
+static struct sk_buff *mlx4_en_rx_skb(struct mlx4_en_priv *priv,
+				      struct mlx4_en_rx_desc *rx_desc,
+				      struct skb_frag_struct *skb_frags,
+				      struct mlx4_en_rx_alloc *page_alloc,
+				      unsigned int length)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct sk_buff *skb;
+	void *va;
+	int used_frags;
+	dma_addr_t dma;
+
+	skb = dev_alloc_skb(SMALL_PACKET_SIZE + NET_IP_ALIGN);
+	if (!skb) {
+		en_dbg(RX_ERR, priv, "Failed allocating skb\n");
+		return NULL;
+	}
+	skb->dev = priv->dev;
+	skb_reserve(skb, NET_IP_ALIGN);
+	skb->len = length;
+	skb->truesize = length + sizeof(struct sk_buff);
+
+	/* Get pointer to first fragment so we could copy the headers into the
+	 * (linear part of the) skb */
+	va = page_address(skb_frags[0].page) + skb_frags[0].page_offset;
+
+	if (length <= SMALL_PACKET_SIZE) {
+		/* We are copying all relevant data to the skb - temporarily
+		 * synch buffers for the copy */
+		dma = be64_to_cpu(rx_desc->data[0].addr);
+		dma_sync_single_range_for_cpu(&mdev->pdev->dev, dma, 0,
+					      length, DMA_FROM_DEVICE);
+		skb_copy_to_linear_data(skb, va, length);
+		dma_sync_single_range_for_device(&mdev->pdev->dev, dma, 0,
+						 length, DMA_FROM_DEVICE);
+		skb->tail += length;
+	} else {
+
+		/* Move relevant fragments to skb */
+		used_frags = mlx4_en_complete_rx_desc(priv, rx_desc, skb_frags,
+						      skb_shinfo(skb)->frags,
+						      page_alloc, length);
+		if (unlikely(!used_frags)) {
+			kfree_skb(skb);
+			return NULL;
+		}
+		skb_shinfo(skb)->nr_frags = used_frags;
+
+		/* Copy headers into the skb linear buffer */
+		memcpy(skb->data, va, HEADER_COPY_SIZE);
+		skb->tail += HEADER_COPY_SIZE;
+
+		/* Skip headers in first fragment */
+		skb_shinfo(skb)->frags[0].page_offset += HEADER_COPY_SIZE;
+
+		/* Adjust size of first fragment */
+		skb_shinfo(skb)->frags[0].size -= HEADER_COPY_SIZE;
+		skb->data_len = length - HEADER_COPY_SIZE;
+	}
+	return skb;
+}
+
+
+int mlx4_en_process_rx_cq(struct net_device *dev, struct mlx4_en_cq *cq, int budget)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_cqe *cqe;
+	struct mlx4_en_rx_ring *ring = &priv->rx_ring[cq->ring];
+	struct skb_frag_struct *skb_frags;
+	struct skb_frag_struct lro_frags[MLX4_EN_MAX_RX_FRAGS];
+	struct mlx4_en_rx_desc *rx_desc;
+	struct sk_buff *skb;
+	int index;
+	int nr;
+	unsigned int length;
+	int polled = 0;
+	int ip_summed;
+
+	if (!priv->port_up)
+		return 0;
+
+	/* We assume a 1:1 mapping between CQEs and Rx descriptors, so Rx
+	 * descriptor offset can be deduced from the CQE index instead of
+	 * reading 'cqe->index' */
+	index = cq->mcq.cons_index & ring->size_mask;
+	cqe = &cq->buf[index];
+
+	/* Process all completed CQEs */
+	while (XNOR(cqe->owner_sr_opcode & MLX4_CQE_OWNER_MASK,
+		    cq->mcq.cons_index & cq->size)) {
+
+		skb_frags = ring->rx_info + (index << priv->log_rx_info);
+		rx_desc = ring->buf + (index << ring->log_stride);
+
+		/*
+		 * make sure we read the CQE after we read the ownership bit
+		 */
+		rmb();
+
+		/* Drop packet on bad receive or bad checksum */
+		if (unlikely((cqe->owner_sr_opcode & MLX4_CQE_OPCODE_MASK) ==
+						MLX4_CQE_OPCODE_ERROR)) {
+			en_err(priv, "CQE completed in error - vendor "
+				  "syndrom:%d syndrom:%d\n",
+				  ((struct mlx4_err_cqe *) cqe)->vendor_err_syndrome,
+				  ((struct mlx4_err_cqe *) cqe)->syndrome);
+			goto next;
+		}
+		if (unlikely(cqe->badfcs_enc & MLX4_CQE_BAD_FCS)) {
+			en_dbg(RX_ERR, priv, "Accepted frame with bad FCS\n");
+			goto next;
+		}
+
+		/*
+		 * Packet is OK - process it.
+		 */
+		length = be32_to_cpu(cqe->byte_cnt);
+		ring->bytes += length;
+		ring->packets++;
+
+		if (likely(priv->rx_csum)) {
+			if ((cqe->status & cpu_to_be16(MLX4_CQE_STATUS_IPOK)) &&
+			    (cqe->checksum == cpu_to_be16(0xffff))) {
+				priv->port_stats.rx_chksum_good++;
+				/* This packet is eligible for LRO if it is:
+				 * - DIX Ethernet (type interpretation)
+				 * - TCP/IP (v4)
+				 * - without IP options
+				 * - not an IP fragment */
+				if (mlx4_en_can_lro(cqe->status) &&
+				    dev->features & NETIF_F_LRO) {
+
+					nr = mlx4_en_complete_rx_desc(
+						priv, rx_desc,
+						skb_frags, lro_frags,
+						ring->page_alloc, length);
+					if (!nr)
+						goto next;
+
+					if (priv->vlgrp && (cqe->vlan_my_qpn &
+							    cpu_to_be32(MLX4_CQE_VLAN_PRESENT_MASK))) {
+						lro_vlan_hwaccel_receive_frags(
+						       &ring->lro, lro_frags,
+						       length, length,
+						       priv->vlgrp,
+						       be16_to_cpu(cqe->sl_vid),
+						       NULL, 0);
+					} else
+						lro_receive_frags(&ring->lro,
+								  lro_frags,
+								  length,
+								  length,
+								  NULL, 0);
+
+					goto next;
+				}
+
+				/* LRO not possible, complete processing here */
+				ip_summed = CHECKSUM_UNNECESSARY;
+				INC_PERF_COUNTER(priv->pstats.lro_misses);
+			} else {
+				ip_summed = CHECKSUM_NONE;
+				priv->port_stats.rx_chksum_none++;
+			}
+		} else {
+			ip_summed = CHECKSUM_NONE;
+			priv->port_stats.rx_chksum_none++;
+		}
+
+		skb = mlx4_en_rx_skb(priv, rx_desc, skb_frags,
+				     ring->page_alloc, length);
+		if (!skb) {
+			priv->stats.rx_dropped++;
+			goto next;
+		}
+
+		skb->ip_summed = ip_summed;
+		skb->protocol = eth_type_trans(skb, dev);
+		skb_record_rx_queue(skb, cq->ring);
+
+		/* Push it up the stack */
+		if (priv->vlgrp && (be32_to_cpu(cqe->vlan_my_qpn) &
+				    MLX4_CQE_VLAN_PRESENT_MASK)) {
+			vlan_hwaccel_receive_skb(skb, priv->vlgrp,
+						be16_to_cpu(cqe->sl_vid));
+		} else
+			netif_receive_skb(skb);
+
+next:
+		++cq->mcq.cons_index;
+		index = (cq->mcq.cons_index) & ring->size_mask;
+		cqe = &cq->buf[index];
+		if (++polled == budget) {
+			/* We are here because we reached the NAPI budget -
+			 * flush only pending LRO sessions */
+			lro_flush_all(&ring->lro);
+			goto out;
+		}
+	}
+
+	/* If CQ is empty flush all LRO sessions unconditionally */
+	lro_flush_all(&ring->lro);
+
+out:
+	AVG_PERF_COUNTER(priv->pstats.rx_coal_avg, polled);
+	mlx4_cq_set_ci(&cq->mcq);
+	wmb(); /* ensure HW sees CQ consumer before we post new buffers */
+	ring->cons = cq->mcq.cons_index;
+	ring->prod += polled; /* Polled descriptors were realocated in place */
+	mlx4_en_update_rx_prod_db(ring);
+	return polled;
+}
+
+
+void mlx4_en_rx_irq(struct mlx4_cq *mcq)
+{
+	struct mlx4_en_cq *cq = container_of(mcq, struct mlx4_en_cq, mcq);
+	struct mlx4_en_priv *priv = netdev_priv(cq->dev);
+
+	if (priv->port_up)
+		napi_schedule(&cq->napi);
+	else
+		mlx4_en_arm_cq(priv, cq);
+}
+
+/* Rx CQ polling - called by NAPI */
+int mlx4_en_poll_rx_cq(struct napi_struct *napi, int budget)
+{
+	struct mlx4_en_cq *cq = container_of(napi, struct mlx4_en_cq, napi);
+	struct net_device *dev = cq->dev;
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	int done;
+
+	done = mlx4_en_process_rx_cq(dev, cq, budget);
+
+	/* If we used up all the quota - we're probably not done yet... */
+	if (done == budget)
+		INC_PERF_COUNTER(priv->pstats.napi_quota);
+	else {
+		/* Done for now */
+		napi_complete(napi);
+		mlx4_en_arm_cq(priv, cq);
+	}
+	return done;
+}
+
+
+/* Calculate the last offset position that accomodates a full fragment
+ * (assuming fagment size = stride-align) */
+static int mlx4_en_last_alloc_offset(struct mlx4_en_priv *priv, u16 stride, u16 align)
+{
+	u16 res = MLX4_EN_ALLOC_SIZE % stride;
+	u16 offset = MLX4_EN_ALLOC_SIZE - stride - res + align;
+
+	en_dbg(DRV, priv, "Calculated last offset for stride:%d align:%d "
+			    "res:%d offset:%d\n", stride, align, res, offset);
+	return offset;
+}
+
+
+static int frag_sizes[] = {
+	FRAG_SZ0,
+	FRAG_SZ1,
+	FRAG_SZ2,
+	FRAG_SZ3
+};
+
+void mlx4_en_calc_rx_buf(struct net_device *dev)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	int eff_mtu = dev->mtu + ETH_HLEN + VLAN_HLEN + ETH_LLC_SNAP_SIZE;
+	int buf_size = 0;
+	int i = 0;
+
+	while (buf_size < eff_mtu) {
+		priv->frag_info[i].frag_size =
+			(eff_mtu > buf_size + frag_sizes[i]) ?
+				frag_sizes[i] : eff_mtu - buf_size;
+		priv->frag_info[i].frag_prefix_size = buf_size;
+		if (!i)	{
+			priv->frag_info[i].frag_align = NET_IP_ALIGN;
+			priv->frag_info[i].frag_stride =
+				ALIGN(frag_sizes[i] + NET_IP_ALIGN, SMP_CACHE_BYTES);
+		} else {
+			priv->frag_info[i].frag_align = 0;
+			priv->frag_info[i].frag_stride =
+				ALIGN(frag_sizes[i], SMP_CACHE_BYTES);
+		}
+		priv->frag_info[i].last_offset = mlx4_en_last_alloc_offset(
+						priv, priv->frag_info[i].frag_stride,
+						priv->frag_info[i].frag_align);
+		buf_size += priv->frag_info[i].frag_size;
+		i++;
+	}
+
+	priv->num_frags = i;
+	priv->rx_skb_size = eff_mtu;
+	priv->log_rx_info = ROUNDUP_LOG2(i * sizeof(struct skb_frag_struct));
+
+	en_dbg(DRV, priv, "Rx buffer scatter-list (effective-mtu:%d "
+		  "num_frags:%d):\n", eff_mtu, priv->num_frags);
+	for (i = 0; i < priv->num_frags; i++) {
+		en_dbg(DRV, priv, "  frag:%d - size:%d prefix:%d align:%d "
+				"stride:%d last_offset:%d\n", i,
+				priv->frag_info[i].frag_size,
+				priv->frag_info[i].frag_prefix_size,
+				priv->frag_info[i].frag_align,
+				priv->frag_info[i].frag_stride,
+				priv->frag_info[i].last_offset);
+	}
+}
+
+/* RSS related functions */
+
+static int mlx4_en_config_rss_qp(struct mlx4_en_priv *priv, int qpn,
+				 struct mlx4_en_rx_ring *ring,
+				 enum mlx4_qp_state *state,
+				 struct mlx4_qp *qp)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_qp_context *context;
+	int err = 0;
+
+	context = kmalloc(sizeof *context , GFP_KERNEL);
+	if (!context) {
+		en_err(priv, "Failed to allocate qp context\n");
+		return -ENOMEM;
+	}
+
+	err = mlx4_qp_alloc(mdev->dev, qpn, qp);
+	if (err) {
+		en_err(priv, "Failed to allocate qp #%x\n", qpn);
+		goto out;
+	}
+	qp->event = mlx4_en_sqp_event;
+
+	memset(context, 0, sizeof *context);
+	mlx4_en_fill_qp_context(priv, ring->size, ring->stride, 0, 0,
+				qpn, ring->cqn, context);
+	context->db_rec_addr = cpu_to_be64(ring->wqres.db.dma);
+
+	err = mlx4_qp_to_ready(mdev->dev, &ring->wqres.mtt, context, qp, state);
+	if (err) {
+		mlx4_qp_remove(mdev->dev, qp);
+		mlx4_qp_free(mdev->dev, qp);
+	}
+	mlx4_en_update_rx_prod_db(ring);
+out:
+	kfree(context);
+	return err;
+}
+
+/* Allocate rx qp's and configure them according to rss map */
+int mlx4_en_config_rss_steer(struct mlx4_en_priv *priv)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_en_rss_map *rss_map = &priv->rss_map;
+	struct mlx4_qp_context context;
+	struct mlx4_en_rss_context *rss_context;
+	void *ptr;
+	int rss_xor = mdev->profile.rss_xor;
+	u8 rss_mask = mdev->profile.rss_mask;
+	int i, qpn;
+	int err = 0;
+	int good_qps = 0;
+
+	en_dbg(DRV, priv, "Configuring rss steering\n");
+	err = mlx4_qp_reserve_range(mdev->dev, priv->rx_ring_num,
+				    priv->rx_ring_num,
+				    &rss_map->base_qpn);
+	if (err) {
+		en_err(priv, "Failed reserving %d qps\n", priv->rx_ring_num);
+		return err;
+	}
+
+	for (i = 0; i < priv->rx_ring_num; i++) {
+		qpn = rss_map->base_qpn + i;
+		err = mlx4_en_config_rss_qp(priv, qpn, &priv->rx_ring[i],
+					    &rss_map->state[i],
+					    &rss_map->qps[i]);
+		if (err)
+			goto rss_err;
+
+		++good_qps;
+	}
+
+	/* Configure RSS indirection qp */
+	err = mlx4_qp_reserve_range(mdev->dev, 1, 1, &priv->base_qpn);
+	if (err) {
+		en_err(priv, "Failed to reserve range for RSS "
+			     "indirection qp\n");
+		goto rss_err;
+	}
+	err = mlx4_qp_alloc(mdev->dev, priv->base_qpn, &rss_map->indir_qp);
+	if (err) {
+		en_err(priv, "Failed to allocate RSS indirection QP\n");
+		goto reserve_err;
+	}
+	rss_map->indir_qp.event = mlx4_en_sqp_event;
+	mlx4_en_fill_qp_context(priv, 0, 0, 0, 1, priv->base_qpn,
+				priv->rx_ring[0].cqn, &context);
+
+	ptr = ((void *) &context) + 0x3c;
+	rss_context = (struct mlx4_en_rss_context *) ptr;
+	rss_context->base_qpn = cpu_to_be32(ilog2(priv->rx_ring_num) << 24 |
+					    (rss_map->base_qpn));
+	rss_context->default_qpn = cpu_to_be32(rss_map->base_qpn);
+	rss_context->hash_fn = rss_xor & 0x3;
+	rss_context->flags = rss_mask << 2;
+
+	err = mlx4_qp_to_ready(mdev->dev, &priv->res.mtt, &context,
+			       &rss_map->indir_qp, &rss_map->indir_state);
+	if (err)
+		goto indir_err;
+
+	return 0;
+
+indir_err:
+	mlx4_qp_modify(mdev->dev, NULL, rss_map->indir_state,
+		       MLX4_QP_STATE_RST, NULL, 0, 0, &rss_map->indir_qp);
+	mlx4_qp_remove(mdev->dev, &rss_map->indir_qp);
+	mlx4_qp_free(mdev->dev, &rss_map->indir_qp);
+reserve_err:
+	mlx4_qp_release_range(mdev->dev, priv->base_qpn, 1);
+rss_err:
+	for (i = 0; i < good_qps; i++) {
+		mlx4_qp_modify(mdev->dev, NULL, rss_map->state[i],
+			       MLX4_QP_STATE_RST, NULL, 0, 0, &rss_map->qps[i]);
+		mlx4_qp_remove(mdev->dev, &rss_map->qps[i]);
+		mlx4_qp_free(mdev->dev, &rss_map->qps[i]);
+	}
+	mlx4_qp_release_range(mdev->dev, rss_map->base_qpn, priv->rx_ring_num);
+	return err;
+}
+
+void mlx4_en_release_rss_steer(struct mlx4_en_priv *priv)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_en_rss_map *rss_map = &priv->rss_map;
+	int i;
+
+	mlx4_qp_modify(mdev->dev, NULL, rss_map->indir_state,
+		       MLX4_QP_STATE_RST, NULL, 0, 0, &rss_map->indir_qp);
+	mlx4_qp_remove(mdev->dev, &rss_map->indir_qp);
+	mlx4_qp_free(mdev->dev, &rss_map->indir_qp);
+	mlx4_qp_release_range(mdev->dev, priv->base_qpn, 1);
+
+	for (i = 0; i < priv->rx_ring_num; i++) {
+		mlx4_qp_modify(mdev->dev, NULL, rss_map->state[i],
+			       MLX4_QP_STATE_RST, NULL, 0, 0, &rss_map->qps[i]);
+		mlx4_qp_remove(mdev->dev, &rss_map->qps[i]);
+		mlx4_qp_free(mdev->dev, &rss_map->qps[i]);
+	}
+	mlx4_qp_release_range(mdev->dev, rss_map->base_qpn, priv->rx_ring_num);
+}
+
+
+
+
+

kernel/drivers/net/mlx4/en_tx.c

@@ -0,0 +1,776 @@
+/*
+ * Copyright (c) 2007 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
+ * 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 <asm/page.h>
+#include <linux/mlx4/cq.h>
+#include <linux/mlx4/qp.h>
+#include <linux/skbuff.h>
+#include <linux/if_vlan.h>
+#include <linux/vmalloc.h>
+
+#include "mlx4_en.h"
+
+enum {
+	MAX_INLINE = 104, /* 128 - 16 - 4 - 4 */
+};
+
+static int inline_thold __read_mostly = MAX_INLINE;
+
+module_param_named(inline_thold, inline_thold, int, 0444);
+MODULE_PARM_DESC(inline_thold, "treshold for using inline data");
+
+int mlx4_en_create_tx_ring(struct mlx4_en_priv *priv,
+			   struct mlx4_en_tx_ring *ring, u32 size,
+			   u16 stride)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	int tmp;
+	int err;
+
+	ring->size = size;
+	ring->size_mask = size - 1;
+	ring->stride = stride;
+
+	inline_thold = min(inline_thold, MAX_INLINE);
+
+	spin_lock_init(&ring->comp_lock);
+
+	tmp = size * sizeof(struct mlx4_en_tx_info);
+	ring->tx_info = vmalloc(tmp);
+	if (!ring->tx_info) {
+		en_err(priv, "Failed allocating tx_info ring\n");
+		return -ENOMEM;
+	}
+	en_dbg(DRV, priv, "Allocated tx_info ring at addr:%p size:%d\n",
+		 ring->tx_info, tmp);
+
+	ring->bounce_buf = kmalloc(MAX_DESC_SIZE, GFP_KERNEL);
+	if (!ring->bounce_buf) {
+		en_err(priv, "Failed allocating bounce buffer\n");
+		err = -ENOMEM;
+		goto err_tx;
+	}
+	ring->buf_size = ALIGN(size * ring->stride, MLX4_EN_PAGE_SIZE);
+
+	err = mlx4_alloc_hwq_res(mdev->dev, &ring->wqres, ring->buf_size,
+				 2 * PAGE_SIZE);
+	if (err) {
+		en_err(priv, "Failed allocating hwq resources\n");
+		goto err_bounce;
+	}
+
+	err = mlx4_en_map_buffer(&ring->wqres.buf);
+	if (err) {
+		en_err(priv, "Failed to map TX buffer\n");
+		goto err_hwq_res;
+	}
+
+	ring->buf = ring->wqres.buf.direct.buf;
+
+	en_dbg(DRV, priv, "Allocated TX ring (addr:%p) - buf:%p size:%d "
+	       "buf_size:%d dma:%llx\n", ring, ring->buf, ring->size,
+	       ring->buf_size, (unsigned long long) ring->wqres.buf.direct.map);
+
+	err = mlx4_qp_reserve_range(mdev->dev, 1, 1, &ring->qpn);
+	if (err) {
+		en_err(priv, "Failed reserving qp for tx ring.\n");
+		goto err_map;
+	}
+
+	err = mlx4_qp_alloc(mdev->dev, ring->qpn, &ring->qp);
+	if (err) {
+		en_err(priv, "Failed allocating qp %d\n", ring->qpn);
+		goto err_reserve;
+	}
+	ring->qp.event = mlx4_en_sqp_event;
+
+	return 0;
+
+err_reserve:
+	mlx4_qp_release_range(mdev->dev, ring->qpn, 1);
+err_map:
+	mlx4_en_unmap_buffer(&ring->wqres.buf);
+err_hwq_res:
+	mlx4_free_hwq_res(mdev->dev, &ring->wqres, ring->buf_size);
+err_bounce:
+	kfree(ring->bounce_buf);
+	ring->bounce_buf = NULL;
+err_tx:
+	vfree(ring->tx_info);
+	ring->tx_info = NULL;
+	return err;
+}
+
+void mlx4_en_destroy_tx_ring(struct mlx4_en_priv *priv,
+			     struct mlx4_en_tx_ring *ring)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	en_dbg(DRV, priv, "Destroying tx ring, qpn: %d\n", ring->qpn);
+
+	mlx4_qp_remove(mdev->dev, &ring->qp);
+	mlx4_qp_free(mdev->dev, &ring->qp);
+	mlx4_qp_release_range(mdev->dev, ring->qpn, 1);
+	mlx4_en_unmap_buffer(&ring->wqres.buf);
+	mlx4_free_hwq_res(mdev->dev, &ring->wqres, ring->buf_size);
+	kfree(ring->bounce_buf);
+	ring->bounce_buf = NULL;
+	vfree(ring->tx_info);
+	ring->tx_info = NULL;
+}
+
+int mlx4_en_activate_tx_ring(struct mlx4_en_priv *priv,
+			     struct mlx4_en_tx_ring *ring,
+			     int cq)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	int err;
+
+	ring->cqn = cq;
+	ring->prod = 0;
+	ring->cons = 0xffffffff;
+	ring->last_nr_txbb = 1;
+	ring->poll_cnt = 0;
+	ring->blocked = 0;
+	memset(ring->tx_info, 0, ring->size * sizeof(struct mlx4_en_tx_info));
+	memset(ring->buf, 0, ring->buf_size);
+
+	ring->qp_state = MLX4_QP_STATE_RST;
+	ring->doorbell_qpn = swab32(ring->qp.qpn << 8);
+
+	mlx4_en_fill_qp_context(priv, ring->size, ring->stride, 1, 0, ring->qpn,
+				ring->cqn, &ring->context);
+
+	err = mlx4_qp_to_ready(mdev->dev, &ring->wqres.mtt, &ring->context,
+			       &ring->qp, &ring->qp_state);
+
+	return err;
+}
+
+void mlx4_en_deactivate_tx_ring(struct mlx4_en_priv *priv,
+				struct mlx4_en_tx_ring *ring)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+
+	mlx4_qp_modify(mdev->dev, NULL, ring->qp_state,
+		       MLX4_QP_STATE_RST, NULL, 0, 0, &ring->qp);
+}
+
+
+static u32 mlx4_en_free_tx_desc(struct mlx4_en_priv *priv,
+				struct mlx4_en_tx_ring *ring,
+				int index, u8 owner)
+{
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_en_tx_info *tx_info = &ring->tx_info[index];
+	struct mlx4_en_tx_desc *tx_desc = ring->buf + index * TXBB_SIZE;
+	struct mlx4_wqe_data_seg *data = (void *) tx_desc + tx_info->data_offset;
+	struct sk_buff *skb = tx_info->skb;
+	struct skb_frag_struct *frag;
+	void *end = ring->buf + ring->buf_size;
+	int frags = skb_shinfo(skb)->nr_frags;
+	int i;
+	__be32 *ptr = (__be32 *)tx_desc;
+	__be32 stamp = cpu_to_be32(STAMP_VAL | (!!owner << STAMP_SHIFT));
+
+	/* Optimize the common case when there are no wraparounds */
+	if (likely((void *) tx_desc + tx_info->nr_txbb * TXBB_SIZE <= end)) {
+		if (!tx_info->inl) {
+			if (tx_info->linear) {
+				pci_unmap_single(mdev->pdev,
+					(dma_addr_t) be64_to_cpu(data->addr),
+					 be32_to_cpu(data->byte_count),
+					 PCI_DMA_TODEVICE);
+				++data;
+			}
+
+			for (i = 0; i < frags; i++) {
+				frag = &skb_shinfo(skb)->frags[i];
+				pci_unmap_page(mdev->pdev,
+					(dma_addr_t) be64_to_cpu(data[i].addr),
+					frag->size, PCI_DMA_TODEVICE);
+			}
+		}
+		/* Stamp the freed descriptor */
+		for (i = 0; i < tx_info->nr_txbb * TXBB_SIZE; i += STAMP_STRIDE) {
+			*ptr = stamp;
+			ptr += STAMP_DWORDS;
+		}
+
+	} else {
+		if (!tx_info->inl) {
+			if ((void *) data >= end) {
+				data = (struct mlx4_wqe_data_seg *)
+						(ring->buf + ((void *) data - end));
+			}
+
+			if (tx_info->linear) {
+				pci_unmap_single(mdev->pdev,
+					(dma_addr_t) be64_to_cpu(data->addr),
+					 be32_to_cpu(data->byte_count),
+					 PCI_DMA_TODEVICE);
+				++data;
+			}
+
+			for (i = 0; i < frags; i++) {
+				/* Check for wraparound before unmapping */
+				if ((void *) data >= end)
+					data = (struct mlx4_wqe_data_seg *) ring->buf;
+				frag = &skb_shinfo(skb)->frags[i];
+				pci_unmap_page(mdev->pdev,
+					(dma_addr_t) be64_to_cpu(data->addr),
+					 frag->size, PCI_DMA_TODEVICE);
+				++data;
+			}
+		}
+		/* Stamp the freed descriptor */
+		for (i = 0; i < tx_info->nr_txbb * TXBB_SIZE; i += STAMP_STRIDE) {
+			*ptr = stamp;
+			ptr += STAMP_DWORDS;
+			if ((void *) ptr >= end) {
+				ptr = ring->buf;
+				stamp ^= cpu_to_be32(0x80000000);
+			}
+		}
+
+	}
+	dev_kfree_skb_any(skb);
+	return tx_info->nr_txbb;
+}
+
+
+int mlx4_en_free_tx_buf(struct net_device *dev, struct mlx4_en_tx_ring *ring)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	int cnt = 0;
+
+	/* Skip last polled descriptor */
+	ring->cons += ring->last_nr_txbb;
+	en_dbg(DRV, priv, "Freeing Tx buf - cons:0x%x prod:0x%x\n",
+		 ring->cons, ring->prod);
+
+	if ((u32) (ring->prod - ring->cons) > ring->size) {
+		if (netif_msg_tx_err(priv))
+			en_warn(priv, "Tx consumer passed producer!\n");
+		return 0;
+	}
+
+	while (ring->cons != ring->prod) {
+		ring->last_nr_txbb = mlx4_en_free_tx_desc(priv, ring,
+						ring->cons & ring->size_mask,
+						!!(ring->cons & ring->size));
+		ring->cons += ring->last_nr_txbb;
+		cnt++;
+	}
+
+	if (cnt)
+		en_dbg(DRV, priv, "Freed %d uncompleted tx descriptors\n", cnt);
+
+	return cnt;
+}
+
+
+static void mlx4_en_process_tx_cq(struct net_device *dev, struct mlx4_en_cq *cq)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_cq *mcq = &cq->mcq;
+	struct mlx4_en_tx_ring *ring = &priv->tx_ring[cq->ring];
+	struct mlx4_cqe *cqe = cq->buf;
+	u16 index;
+	u16 new_index;
+	u32 txbbs_skipped = 0;
+	u32 cq_last_sav;
+
+	/* index always points to the first TXBB of the last polled descriptor */
+	index = ring->cons & ring->size_mask;
+	new_index = be16_to_cpu(cqe->wqe_index) & ring->size_mask;
+	if (index == new_index)
+		return;
+
+	if (!priv->port_up)
+		return;
+
+	/*
+	 * We use a two-stage loop:
+	 * - the first samples the HW-updated CQE
+	 * - the second frees TXBBs until the last sample
+	 * This lets us amortize CQE cache misses, while still polling the CQ
+	 * until is quiescent.
+	 */
+	cq_last_sav = mcq->cons_index;
+	do {
+		do {
+			/* Skip over last polled CQE */
+			index = (index + ring->last_nr_txbb) & ring->size_mask;
+			txbbs_skipped += ring->last_nr_txbb;
+
+			/* Poll next CQE */
+			ring->last_nr_txbb = mlx4_en_free_tx_desc(
+						priv, ring, index,
+						!!((ring->cons + txbbs_skipped) &
+						   ring->size));
+			++mcq->cons_index;
+
+		} while (index != new_index);
+
+		new_index = be16_to_cpu(cqe->wqe_index) & ring->size_mask;
+	} while (index != new_index);
+	AVG_PERF_COUNTER(priv->pstats.tx_coal_avg,
+			 (u32) (mcq->cons_index - cq_last_sav));
+
+	/*
+	 * To prevent CQ overflow we first update CQ consumer and only then
+	 * the ring consumer.
+	 */
+	mlx4_cq_set_ci(mcq);
+	wmb();
+	ring->cons += txbbs_skipped;
+
+	/* Wakeup Tx queue if this ring stopped it */
+	if (unlikely(ring->blocked)) {
+		if ((u32) (ring->prod - ring->cons) <=
+		     ring->size - HEADROOM - MAX_DESC_TXBBS) {
+			ring->blocked = 0;
+			netif_tx_wake_queue(netdev_get_tx_queue(dev, cq->ring));
+			priv->port_stats.wake_queue++;
+		}
+	}
+}
+
+void mlx4_en_tx_irq(struct mlx4_cq *mcq)
+{
+	struct mlx4_en_cq *cq = container_of(mcq, struct mlx4_en_cq, mcq);
+	struct mlx4_en_priv *priv = netdev_priv(cq->dev);
+	struct mlx4_en_tx_ring *ring = &priv->tx_ring[cq->ring];
+
+	if (!spin_trylock(&ring->comp_lock))
+		return;
+	mlx4_en_process_tx_cq(cq->dev, cq);
+	mod_timer(&cq->timer, jiffies + 1);
+	spin_unlock(&ring->comp_lock);
+}
+
+
+void mlx4_en_poll_tx_cq(unsigned long data)
+{
+	struct mlx4_en_cq *cq = (struct mlx4_en_cq *) data;
+	struct mlx4_en_priv *priv = netdev_priv(cq->dev);
+	struct mlx4_en_tx_ring *ring = &priv->tx_ring[cq->ring];
+	u32 inflight;
+
+	INC_PERF_COUNTER(priv->pstats.tx_poll);
+
+	if (!spin_trylock_irq(&ring->comp_lock)) {
+		mod_timer(&cq->timer, jiffies + MLX4_EN_TX_POLL_TIMEOUT);
+		return;
+	}
+	mlx4_en_process_tx_cq(cq->dev, cq);
+	inflight = (u32) (ring->prod - ring->cons - ring->last_nr_txbb);
+
+	/* If there are still packets in flight and the timer has not already
+	 * been scheduled by the Tx routine then schedule it here to guarantee
+	 * completion processing of these packets */
+	if (inflight && priv->port_up)
+		mod_timer(&cq->timer, jiffies + MLX4_EN_TX_POLL_TIMEOUT);
+
+	spin_unlock_irq(&ring->comp_lock);
+}
+
+static struct mlx4_en_tx_desc *mlx4_en_bounce_to_desc(struct mlx4_en_priv *priv,
+						      struct mlx4_en_tx_ring *ring,
+						      u32 index,
+						      unsigned int desc_size)
+{
+	u32 copy = (ring->size - index) * TXBB_SIZE;
+	int i;
+
+	for (i = desc_size - copy - 4; i >= 0; i -= 4) {
+		if ((i & (TXBB_SIZE - 1)) == 0)
+			wmb();
+
+		*((u32 *) (ring->buf + i)) =
+			*((u32 *) (ring->bounce_buf + copy + i));
+	}
+
+	for (i = copy - 4; i >= 4 ; i -= 4) {
+		if ((i & (TXBB_SIZE - 1)) == 0)
+			wmb();
+
+		*((u32 *) (ring->buf + index * TXBB_SIZE + i)) =
+			*((u32 *) (ring->bounce_buf + i));
+	}
+
+	/* Return real descriptor location */
+	return ring->buf + index * TXBB_SIZE;
+}
+
+static inline void mlx4_en_xmit_poll(struct mlx4_en_priv *priv, int tx_ind)
+{
+	struct mlx4_en_cq *cq = &priv->tx_cq[tx_ind];
+	struct mlx4_en_tx_ring *ring = &priv->tx_ring[tx_ind];
+	unsigned long flags;
+
+	/* If we don't have a pending timer, set one up to catch our recent
+	   post in case the interface becomes idle */
+	if (!timer_pending(&cq->timer))
+		mod_timer(&cq->timer, jiffies + MLX4_EN_TX_POLL_TIMEOUT);
+
+	/* Poll the CQ every mlx4_en_TX_MODER_POLL packets */
+	if ((++ring->poll_cnt & (MLX4_EN_TX_POLL_MODER - 1)) == 0)
+		if (spin_trylock_irqsave(&ring->comp_lock, flags)) {
+			mlx4_en_process_tx_cq(priv->dev, cq);
+			spin_unlock_irqrestore(&ring->comp_lock, flags);
+		}
+}
+
+static void *get_frag_ptr(struct sk_buff *skb)
+{
+	struct skb_frag_struct *frag =  &skb_shinfo(skb)->frags[0];
+	struct page *page = frag->page;
+	void *ptr;
+
+	ptr = page_address(page);
+	if (unlikely(!ptr))
+		return NULL;
+
+	return ptr + frag->page_offset;
+}
+
+static int is_inline(struct sk_buff *skb, void **pfrag)
+{
+	void *ptr;
+
+	if (inline_thold && !skb_is_gso(skb) && skb->len <= inline_thold) {
+		if (skb_shinfo(skb)->nr_frags == 1) {
+			ptr = get_frag_ptr(skb);
+			if (unlikely(!ptr))
+				return 0;
+
+			if (pfrag)
+				*pfrag = ptr;
+
+			return 1;
+		} else if (unlikely(skb_shinfo(skb)->nr_frags))
+			return 0;
+		else
+			return 1;
+	}
+
+	return 0;
+}
+
+static int inline_size(struct sk_buff *skb)
+{
+	if (skb->len + CTRL_SIZE + sizeof(struct mlx4_wqe_inline_seg)
+	    <= MLX4_INLINE_ALIGN)
+		return ALIGN(skb->len + CTRL_SIZE +
+			     sizeof(struct mlx4_wqe_inline_seg), 16);
+	else
+		return ALIGN(skb->len + CTRL_SIZE + 2 *
+			     sizeof(struct mlx4_wqe_inline_seg), 16);
+}
+
+static int get_real_size(struct sk_buff *skb, struct net_device *dev,
+			 int *lso_header_size)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	int real_size;
+
+	if (skb_is_gso(skb)) {
+		*lso_header_size = skb_transport_offset(skb) + tcp_hdrlen(skb);
+		real_size = CTRL_SIZE + skb_shinfo(skb)->nr_frags * DS_SIZE +
+			ALIGN(*lso_header_size + 4, DS_SIZE);
+		if (unlikely(*lso_header_size != skb_headlen(skb))) {
+			/* We add a segment for the skb linear buffer only if
+			 * it contains data */
+			if (*lso_header_size < skb_headlen(skb))
+				real_size += DS_SIZE;
+			else {
+				if (netif_msg_tx_err(priv))
+					en_warn(priv, "Non-linear headers\n");
+				return 0;
+			}
+		}
+	} else {
+		*lso_header_size = 0;
+		if (!is_inline(skb, NULL))
+			real_size = CTRL_SIZE + (skb_shinfo(skb)->nr_frags + 1) * DS_SIZE;
+		else
+			real_size = inline_size(skb);
+	}
+
+	return real_size;
+}
+
+static void build_inline_wqe(struct mlx4_en_tx_desc *tx_desc, struct sk_buff *skb,
+			     int real_size, u16 *vlan_tag, int tx_ind, void *fragptr)
+{
+	struct mlx4_wqe_inline_seg *inl = &tx_desc->inl;
+	int spc = MLX4_INLINE_ALIGN - CTRL_SIZE - sizeof *inl;
+
+	if (skb->len <= spc) {
+		inl->byte_count = cpu_to_be32(1 << 31 | skb->len);
+		skb_copy_from_linear_data(skb, inl + 1, skb_headlen(skb));
+		if (skb_shinfo(skb)->nr_frags)
+			memcpy(((void *)(inl + 1)) + skb_headlen(skb), fragptr,
+			       skb_shinfo(skb)->frags[0].size);
+
+	} else {
+		inl->byte_count = cpu_to_be32(1 << 31 | spc);
+		if (skb_headlen(skb) <= spc) {
+			skb_copy_from_linear_data(skb, inl + 1, skb_headlen(skb));
+			if (skb_headlen(skb) < spc) {
+				memcpy(((void *)(inl + 1)) + skb_headlen(skb),
+					fragptr, spc - skb_headlen(skb));
+				fragptr +=  spc - skb_headlen(skb);
+			}
+			inl = (void *) (inl + 1) + spc;
+			memcpy(((void *)(inl + 1)), fragptr, skb->len - spc);
+		} else {
+			skb_copy_from_linear_data(skb, inl + 1, spc);
+			inl = (void *) (inl + 1) + spc;
+			skb_copy_from_linear_data_offset(skb, spc, inl + 1,
+					skb_headlen(skb) - spc);
+			if (skb_shinfo(skb)->nr_frags)
+				memcpy(((void *)(inl + 1)) + skb_headlen(skb) - spc,
+					fragptr, skb_shinfo(skb)->frags[0].size);
+		}
+
+		wmb();
+		inl->byte_count = cpu_to_be32(1 << 31 | (skb->len - spc));
+	}
+	tx_desc->ctrl.vlan_tag = cpu_to_be16(*vlan_tag);
+	tx_desc->ctrl.ins_vlan = MLX4_WQE_CTRL_INS_VLAN * !!(*vlan_tag);
+	tx_desc->ctrl.fence_size = (real_size / 16) & 0x3f;
+}
+
+u16 mlx4_en_select_queue(struct net_device *dev, struct sk_buff *skb)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	u16 vlan_tag = 0;
+
+	/* If we support per priority flow control and the packet contains
+	 * a vlan tag, send the packet to the TX ring assigned to that priority
+	 */
+	if (priv->prof->rx_ppp && priv->vlgrp && vlan_tx_tag_present(skb)) {
+		vlan_tag = vlan_tx_tag_get(skb);
+		return MLX4_EN_NUM_TX_RINGS + (vlan_tag >> 13);
+	}
+
+	return skb_tx_hash(dev, skb);
+}
+
+netdev_tx_t mlx4_en_xmit(struct sk_buff *skb, struct net_device *dev)
+{
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+	struct mlx4_en_tx_ring *ring;
+	struct mlx4_en_cq *cq;
+	struct mlx4_en_tx_desc *tx_desc;
+	struct mlx4_wqe_data_seg *data;
+	struct skb_frag_struct *frag;
+	struct mlx4_en_tx_info *tx_info;
+	int tx_ind = 0;
+	int nr_txbb;
+	int desc_size;
+	int real_size;
+	dma_addr_t dma;
+	u32 index;
+	__be32 op_own;
+	u16 vlan_tag = 0;
+	int i;
+	int lso_header_size;
+	void *fragptr;
+
+	real_size = get_real_size(skb, dev, &lso_header_size);
+	if (unlikely(!real_size))
+		goto tx_drop;
+
+	/* Allign descriptor to TXBB size */
+	desc_size = ALIGN(real_size, TXBB_SIZE);
+	nr_txbb = desc_size / TXBB_SIZE;
+	if (unlikely(nr_txbb > MAX_DESC_TXBBS)) {
+		if (netif_msg_tx_err(priv))
+			en_warn(priv, "Oversized header or SG list\n");
+		goto tx_drop;
+	}
+
+	tx_ind = skb->queue_mapping;
+	ring = &priv->tx_ring[tx_ind];
+	if (priv->vlgrp && vlan_tx_tag_present(skb))
+		vlan_tag = vlan_tx_tag_get(skb);
+
+	/* Check available TXBBs And 2K spare for prefetch */
+	if (unlikely(((int)(ring->prod - ring->cons)) >
+		     ring->size - HEADROOM - MAX_DESC_TXBBS)) {
+		/* every full Tx ring stops queue */
+		netif_tx_stop_queue(netdev_get_tx_queue(dev, tx_ind));
+		ring->blocked = 1;
+		priv->port_stats.queue_stopped++;
+
+		/* Use interrupts to find out when queue opened */
+		cq = &priv->tx_cq[tx_ind];
+		mlx4_en_arm_cq(priv, cq);
+		return NETDEV_TX_BUSY;
+	}
+
+	/* Track current inflight packets for performance analysis */
+	AVG_PERF_COUNTER(priv->pstats.inflight_avg,
+			 (u32) (ring->prod - ring->cons - 1));
+
+	/* Packet is good - grab an index and transmit it */
+	index = ring->prod & ring->size_mask;
+
+	/* See if we have enough space for whole descriptor TXBB for setting
+	 * SW ownership on next descriptor; if not, use a bounce buffer. */
+	if (likely(index + nr_txbb <= ring->size))
+		tx_desc = ring->buf + index * TXBB_SIZE;
+	else
+		tx_desc = (struct mlx4_en_tx_desc *) ring->bounce_buf;
+
+	/* Save skb in tx_info ring */
+	tx_info = &ring->tx_info[index];
+	tx_info->skb = skb;
+	tx_info->nr_txbb = nr_txbb;
+
+	/* Prepare ctrl segement apart opcode+ownership, which depends on
+	 * whether LSO is used */
+	tx_desc->ctrl.vlan_tag = cpu_to_be16(vlan_tag);
+	tx_desc->ctrl.ins_vlan = MLX4_WQE_CTRL_INS_VLAN * !!vlan_tag;
+	tx_desc->ctrl.fence_size = (real_size / 16) & 0x3f;
+	tx_desc->ctrl.srcrb_flags = cpu_to_be32(MLX4_WQE_CTRL_CQ_UPDATE |
+						MLX4_WQE_CTRL_SOLICITED);
+	if (likely(skb->ip_summed == CHECKSUM_PARTIAL)) {
+		tx_desc->ctrl.srcrb_flags |= cpu_to_be32(MLX4_WQE_CTRL_IP_CSUM |
+							 MLX4_WQE_CTRL_TCP_UDP_CSUM);
+		priv->port_stats.tx_chksum_offload++;
+	}
+
+	/* Handle LSO (TSO) packets */
+	if (lso_header_size) {
+		/* Mark opcode as LSO */
+		op_own = cpu_to_be32(MLX4_OPCODE_LSO | (1 << 6)) |
+			((ring->prod & ring->size) ?
+				cpu_to_be32(MLX4_EN_BIT_DESC_OWN) : 0);
+
+		/* Fill in the LSO prefix */
+		tx_desc->lso.mss_hdr_size = cpu_to_be32(
+			skb_shinfo(skb)->gso_size << 16 | lso_header_size);
+
+		/* Copy headers;
+		 * note that we already verified that it is linear */
+		memcpy(tx_desc->lso.header, skb->data, lso_header_size);
+		data = ((void *) &tx_desc->lso +
+			ALIGN(lso_header_size + 4, DS_SIZE));
+
+		priv->port_stats.tso_packets++;
+		i = ((skb->len - lso_header_size) / skb_shinfo(skb)->gso_size) +
+			!!((skb->len - lso_header_size) % skb_shinfo(skb)->gso_size);
+		ring->bytes += skb->len + (i - 1) * lso_header_size;
+		ring->packets += i;
+	} else {
+		/* Normal (Non LSO) packet */
+		op_own = cpu_to_be32(MLX4_OPCODE_SEND) |
+			((ring->prod & ring->size) ?
+			 cpu_to_be32(MLX4_EN_BIT_DESC_OWN) : 0);
+		data = &tx_desc->data;
+		ring->bytes += max(skb->len, (unsigned int) ETH_ZLEN);
+		ring->packets++;
+
+	}
+	AVG_PERF_COUNTER(priv->pstats.tx_pktsz_avg, skb->len);
+
+
+	/* valid only for none inline segments */
+	tx_info->data_offset = (void *) data - (void *) tx_desc;
+
+	tx_info->linear = (lso_header_size < skb_headlen(skb) && !is_inline(skb, NULL)) ? 1 : 0;
+	data += skb_shinfo(skb)->nr_frags + tx_info->linear - 1;
+
+	if (!is_inline(skb, &fragptr)) {
+		/* Map fragments */
+		for (i = skb_shinfo(skb)->nr_frags - 1; i >= 0; i--) {
+			frag = &skb_shinfo(skb)->frags[i];
+			dma = pci_map_page(mdev->dev->pdev, frag->page, frag->page_offset,
+					   frag->size, PCI_DMA_TODEVICE);
+			data->addr = cpu_to_be64(dma);
+			data->lkey = cpu_to_be32(mdev->mr.key);
+			wmb();
+			data->byte_count = cpu_to_be32(frag->size);
+			--data;
+		}
+
+		/* Map linear part */
+		if (tx_info->linear) {
+			dma = pci_map_single(mdev->dev->pdev, skb->data + lso_header_size,
+					     skb_headlen(skb) - lso_header_size, PCI_DMA_TODEVICE);
+			data->addr = cpu_to_be64(dma);
+			data->lkey = cpu_to_be32(mdev->mr.key);
+			wmb();
+			data->byte_count = cpu_to_be32(skb_headlen(skb) - lso_header_size);
+		}
+		tx_info->inl = 0;
+	} else {
+		build_inline_wqe(tx_desc, skb, real_size, &vlan_tag, tx_ind, fragptr);
+		tx_info->inl = 1;
+	}
+
+	ring->prod += nr_txbb;
+
+	/* If we used a bounce buffer then copy descriptor back into place */
+	if (tx_desc == (struct mlx4_en_tx_desc *) ring->bounce_buf)
+		tx_desc = mlx4_en_bounce_to_desc(priv, ring, index, desc_size);
+
+	/* Run destructor before passing skb to HW */
+	if (likely(!skb_shared(skb)))
+		skb_orphan(skb);
+
+	/* Ensure new descirptor hits memory
+	 * before setting ownership of this descriptor to HW */
+	wmb();
+	tx_desc->ctrl.owner_opcode = op_own;
+
+	/* Ring doorbell! */
+	wmb();
+	writel(ring->doorbell_qpn, mdev->uar_map + MLX4_SEND_DOORBELL);
+
+	/* Poll CQ here */
+	mlx4_en_xmit_poll(priv, tx_ind);
+
+	return NETDEV_TX_OK;
+
+tx_drop:
+	dev_kfree_skb_any(skb);
+	priv->stats.tx_dropped++;
+	return NETDEV_TX_OK;
+}
+

kernel/drivers/net/mlx4/eq.c

@@ -0,0 +1,700 @@
+/*
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2005, 2006, 2007 Cisco Systems, 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/interrupt.h>
+#include <linux/mm.h>
+#include <linux/dma-mapping.h>
+
+#include <linux/mlx4/cmd.h>
+
+#include "mlx4.h"
+#include "fw.h"
+
+enum {
+	MLX4_IRQNAME_SIZE	= 64
+};
+
+enum {
+	MLX4_NUM_ASYNC_EQE	= 0x100,
+	MLX4_NUM_SPARE_EQE	= 0x80,
+	MLX4_EQ_ENTRY_SIZE	= 0x20
+};
+
+/*
+ * Must be packed because start is 64 bits but only aligned to 32 bits.
+ */
+struct mlx4_eq_context {
+	__be32			flags;
+	u16			reserved1[3];
+	__be16			page_offset;
+	u8			log_eq_size;
+	u8			reserved2[4];
+	u8			eq_period;
+	u8			reserved3;
+	u8			eq_max_count;
+	u8			reserved4[3];
+	u8			intr;
+	u8			log_page_size;
+	u8			reserved5[2];
+	u8			mtt_base_addr_h;
+	__be32			mtt_base_addr_l;
+	u32			reserved6[2];
+	__be32			consumer_index;
+	__be32			producer_index;
+	u32			reserved7[4];
+};
+
+#define MLX4_EQ_STATUS_OK	   ( 0 << 28)
+#define MLX4_EQ_STATUS_WRITE_FAIL  (10 << 28)
+#define MLX4_EQ_OWNER_SW	   ( 0 << 24)
+#define MLX4_EQ_OWNER_HW	   ( 1 << 24)
+#define MLX4_EQ_FLAG_EC		   ( 1 << 18)
+#define MLX4_EQ_FLAG_OI		   ( 1 << 17)
+#define MLX4_EQ_STATE_ARMED	   ( 9 <<  8)
+#define MLX4_EQ_STATE_FIRED	   (10 <<  8)
+#define MLX4_EQ_STATE_ALWAYS_ARMED (11 <<  8)
+
+#define MLX4_ASYNC_EVENT_MASK ((1ull << MLX4_EVENT_TYPE_PATH_MIG)	    | \
+			       (1ull << MLX4_EVENT_TYPE_COMM_EST)	    | \
+			       (1ull << MLX4_EVENT_TYPE_SQ_DRAINED)	    | \
+			       (1ull << MLX4_EVENT_TYPE_CQ_ERROR)	    | \
+			       (1ull << MLX4_EVENT_TYPE_WQ_CATAS_ERROR)	    | \
+			       (1ull << MLX4_EVENT_TYPE_EEC_CATAS_ERROR)    | \
+			       (1ull << MLX4_EVENT_TYPE_PATH_MIG_FAILED)    | \
+			       (1ull << MLX4_EVENT_TYPE_WQ_INVAL_REQ_ERROR) | \
+			       (1ull << MLX4_EVENT_TYPE_WQ_ACCESS_ERROR)    | \
+			       (1ull << MLX4_EVENT_TYPE_PORT_CHANGE)	    | \
+			       (1ull << MLX4_EVENT_TYPE_ECC_DETECT)	    | \
+			       (1ull << MLX4_EVENT_TYPE_SRQ_CATAS_ERROR)    | \
+			       (1ull << MLX4_EVENT_TYPE_SRQ_QP_LAST_WQE)    | \
+			       (1ull << MLX4_EVENT_TYPE_SRQ_LIMIT)	    | \
+			       (1ull << MLX4_EVENT_TYPE_CMD))
+
+struct mlx4_eqe {
+	u8			reserved1;
+	u8			type;
+	u8			reserved2;
+	u8			subtype;
+	union {
+		u32		raw[6];
+		struct {
+			__be32	cqn;
+		} __attribute__((packed)) comp;
+		struct {
+			u16	reserved1;
+			__be16	token;
+			u32	reserved2;
+			u8	reserved3[3];
+			u8	status;
+			__be64	out_param;
+		} __attribute__((packed)) cmd;
+		struct {
+			__be32	qpn;
+		} __attribute__((packed)) qp;
+		struct {
+			__be32	srqn;
+		} __attribute__((packed)) srq;
+		struct {
+			__be32	cqn;
+			u32	reserved1;
+			u8	reserved2[3];
+			u8	syndrome;
+		} __attribute__((packed)) cq_err;
+		struct {
+			u32	reserved1[2];
+			__be32	port;
+		} __attribute__((packed)) port_change;
+	}			event;
+	u8			reserved3[3];
+	u8			owner;
+} __attribute__((packed));
+
+static void eq_set_ci(struct mlx4_eq *eq, int req_not)
+{
+	__raw_writel((__force u32) cpu_to_be32((eq->cons_index & 0xffffff) |
+					       req_not << 31),
+		     eq->doorbell);
+	/* We still want ordering, just not swabbing, so add a barrier */
+	mb();
+}
+
+static struct mlx4_eqe *get_eqe(struct mlx4_eq *eq, u32 entry)
+{
+	unsigned long off = (entry & (eq->nent - 1)) * MLX4_EQ_ENTRY_SIZE;
+	return eq->page_list[off / PAGE_SIZE].buf + off % PAGE_SIZE;
+}
+
+static struct mlx4_eqe *next_eqe_sw(struct mlx4_eq *eq)
+{
+	struct mlx4_eqe *eqe = get_eqe(eq, eq->cons_index);
+	return !!(eqe->owner & 0x80) ^ !!(eq->cons_index & eq->nent) ? NULL : eqe;
+}
+
+static int mlx4_eq_int(struct mlx4_dev *dev, struct mlx4_eq *eq)
+{
+	struct mlx4_eqe *eqe;
+	int cqn;
+	int eqes_found = 0;
+	int set_ci = 0;
+	int port;
+
+	while ((eqe = next_eqe_sw(eq))) {
+		/*
+		 * Make sure we read EQ entry contents after we've
+		 * checked the ownership bit.
+		 */
+		rmb();
+
+		switch (eqe->type) {
+		case MLX4_EVENT_TYPE_COMP:
+			cqn = be32_to_cpu(eqe->event.comp.cqn) & 0xffffff;
+			mlx4_cq_completion(dev, cqn);
+			break;
+
+		case MLX4_EVENT_TYPE_PATH_MIG:
+		case MLX4_EVENT_TYPE_COMM_EST:
+		case MLX4_EVENT_TYPE_SQ_DRAINED:
+		case MLX4_EVENT_TYPE_SRQ_QP_LAST_WQE:
+		case MLX4_EVENT_TYPE_WQ_CATAS_ERROR:
+		case MLX4_EVENT_TYPE_PATH_MIG_FAILED:
+		case MLX4_EVENT_TYPE_WQ_INVAL_REQ_ERROR:
+		case MLX4_EVENT_TYPE_WQ_ACCESS_ERROR:
+			mlx4_qp_event(dev, be32_to_cpu(eqe->event.qp.qpn) & 0xffffff,
+				      eqe->type);
+			break;
+
+		case MLX4_EVENT_TYPE_SRQ_LIMIT:
+		case MLX4_EVENT_TYPE_SRQ_CATAS_ERROR:
+			mlx4_srq_event(dev, be32_to_cpu(eqe->event.srq.srqn) & 0xffffff,
+				      eqe->type);
+			break;
+
+		case MLX4_EVENT_TYPE_CMD:
+			mlx4_cmd_event(dev,
+				       be16_to_cpu(eqe->event.cmd.token),
+				       eqe->event.cmd.status,
+				       be64_to_cpu(eqe->event.cmd.out_param));
+			break;
+
+		case MLX4_EVENT_TYPE_PORT_CHANGE:
+			port = be32_to_cpu(eqe->event.port_change.port) >> 28;
+			if (eqe->subtype == MLX4_PORT_CHANGE_SUBTYPE_DOWN) {
+				mlx4_dispatch_event(dev, MLX4_DEV_EVENT_PORT_DOWN,
+						    port);
+				mlx4_priv(dev)->sense.do_sense_port[port] = 1;
+			} else {
+				mlx4_dispatch_event(dev, MLX4_DEV_EVENT_PORT_UP,
+						    port);
+				mlx4_priv(dev)->sense.do_sense_port[port] = 0;
+			}
+			break;
+
+		case MLX4_EVENT_TYPE_CQ_ERROR:
+			mlx4_warn(dev, "CQ %s on CQN %06x\n",
+				  eqe->event.cq_err.syndrome == 1 ?
+				  "overrun" : "access violation",
+				  be32_to_cpu(eqe->event.cq_err.cqn) & 0xffffff);
+			mlx4_cq_event(dev, be32_to_cpu(eqe->event.cq_err.cqn),
+				      eqe->type);
+			break;
+
+		case MLX4_EVENT_TYPE_EQ_OVERFLOW:
+			mlx4_warn(dev, "EQ overrun on EQN %d\n", eq->eqn);
+			break;
+
+		case MLX4_EVENT_TYPE_EEC_CATAS_ERROR:
+		case MLX4_EVENT_TYPE_ECC_DETECT:
+		default:
+			mlx4_warn(dev, "Unhandled event %02x(%02x) on EQ %d at index %u\n",
+				  eqe->type, eqe->subtype, eq->eqn, eq->cons_index);
+			break;
+		};
+
+		++eq->cons_index;
+		eqes_found = 1;
+		++set_ci;
+
+		/*
+		 * The HCA will think the queue has overflowed if we
+		 * don't tell it we've been processing events.  We
+		 * create our EQs with MLX4_NUM_SPARE_EQE extra
+		 * entries, so we must update our consumer index at
+		 * least that often.
+		 */
+		if (unlikely(set_ci >= MLX4_NUM_SPARE_EQE)) {
+			eq_set_ci(eq, 0);
+			set_ci = 0;
+		}
+	}
+
+	eq_set_ci(eq, 1);
+
+	return eqes_found;
+}
+
+static irqreturn_t mlx4_interrupt(int irq, void *dev_ptr)
+{
+	struct mlx4_dev *dev = dev_ptr;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int work = 0;
+	int i;
+
+	writel(priv->eq_table.clr_mask, priv->eq_table.clr_int);
+
+	for (i = 0; i < dev->caps.num_comp_vectors + 1; ++i)
+		work |= mlx4_eq_int(dev, &priv->eq_table.eq[i]);
+
+	return IRQ_RETVAL(work);
+}
+
+static irqreturn_t mlx4_msi_x_interrupt(int irq, void *eq_ptr)
+{
+	struct mlx4_eq  *eq  = eq_ptr;
+	struct mlx4_dev *dev = eq->dev;
+
+	mlx4_eq_int(dev, eq);
+
+	/* MSI-X vectors always belong to us */
+	return IRQ_HANDLED;
+}
+
+static int mlx4_MAP_EQ(struct mlx4_dev *dev, u64 event_mask, int unmap,
+			int eq_num)
+{
+	return mlx4_cmd(dev, event_mask, (unmap << 31) | eq_num,
+			0, MLX4_CMD_MAP_EQ, MLX4_CMD_TIME_CLASS_B);
+}
+
+static int mlx4_SW2HW_EQ(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *mailbox,
+			 int eq_num)
+{
+	return mlx4_cmd(dev, mailbox->dma, eq_num, 0, MLX4_CMD_SW2HW_EQ,
+			MLX4_CMD_TIME_CLASS_A);
+}
+
+static int mlx4_HW2SW_EQ(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *mailbox,
+			 int eq_num)
+{
+	return mlx4_cmd_box(dev, 0, mailbox->dma, eq_num, 0, MLX4_CMD_HW2SW_EQ,
+			    MLX4_CMD_TIME_CLASS_A);
+}
+
+static int mlx4_num_eq_uar(struct mlx4_dev *dev)
+{
+	/*
+	 * Each UAR holds 4 EQ doorbells.  To figure out how many UARs
+	 * we need to map, take the difference of highest index and
+	 * the lowest index we'll use and add 1.
+	 */
+	return (dev->caps.num_comp_vectors + 1 + dev->caps.reserved_eqs) / 4 -
+		dev->caps.reserved_eqs / 4 + 1;
+}
+
+static void __iomem *mlx4_get_eq_uar(struct mlx4_dev *dev, struct mlx4_eq *eq)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int index;
+
+	index = eq->eqn / 4 - dev->caps.reserved_eqs / 4;
+
+	if (!priv->eq_table.uar_map[index]) {
+		priv->eq_table.uar_map[index] =
+			ioremap(pci_resource_start(dev->pdev, 2) +
+				((eq->eqn / 4) << PAGE_SHIFT),
+				PAGE_SIZE);
+		if (!priv->eq_table.uar_map[index]) {
+			mlx4_err(dev, "Couldn't map EQ doorbell for EQN 0x%06x\n",
+				 eq->eqn);
+			return NULL;
+		}
+	}
+
+	return priv->eq_table.uar_map[index] + 0x800 + 8 * (eq->eqn % 4);
+}
+
+static int mlx4_create_eq(struct mlx4_dev *dev, int nent,
+			  u8 intr, struct mlx4_eq *eq)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_eq_context *eq_context;
+	int npages;
+	u64 *dma_list = NULL;
+	dma_addr_t t;
+	u64 mtt_addr;
+	int err = -ENOMEM;
+	int i;
+
+	eq->dev   = dev;
+	eq->nent  = roundup_pow_of_two(max(nent, 2));
+	npages = PAGE_ALIGN(eq->nent * MLX4_EQ_ENTRY_SIZE) / PAGE_SIZE;
+
+	eq->page_list = kmalloc(npages * sizeof *eq->page_list,
+				GFP_KERNEL);
+	if (!eq->page_list)
+		goto err_out;
+
+	for (i = 0; i < npages; ++i)
+		eq->page_list[i].buf = NULL;
+
+	dma_list = kmalloc(npages * sizeof *dma_list, GFP_KERNEL);
+	if (!dma_list)
+		goto err_out_free;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		goto err_out_free;
+	eq_context = mailbox->buf;
+
+	for (i = 0; i < npages; ++i) {
+		eq->page_list[i].buf = dma_alloc_coherent(&dev->pdev->dev,
+							  PAGE_SIZE, &t, GFP_KERNEL);
+		if (!eq->page_list[i].buf)
+			goto err_out_free_pages;
+
+		dma_list[i] = t;
+		eq->page_list[i].map = t;
+
+		memset(eq->page_list[i].buf, 0, PAGE_SIZE);
+	}
+
+	eq->eqn = mlx4_bitmap_alloc(&priv->eq_table.bitmap);
+	if (eq->eqn == -1)
+		goto err_out_free_pages;
+
+	eq->doorbell = mlx4_get_eq_uar(dev, eq);
+	if (!eq->doorbell) {
+		err = -ENOMEM;
+		goto err_out_free_eq;
+	}
+
+	err = mlx4_mtt_init(dev, npages, PAGE_SHIFT, &eq->mtt);
+	if (err)
+		goto err_out_free_eq;
+
+	err = mlx4_write_mtt(dev, &eq->mtt, 0, npages, dma_list);
+	if (err)
+		goto err_out_free_mtt;
+
+	memset(eq_context, 0, sizeof *eq_context);
+	eq_context->flags	  = cpu_to_be32(MLX4_EQ_STATUS_OK   |
+						MLX4_EQ_STATE_ARMED);
+	eq_context->log_eq_size	  = ilog2(eq->nent);
+	eq_context->intr	  = intr;
+	eq_context->log_page_size = PAGE_SHIFT - MLX4_ICM_PAGE_SHIFT;
+
+	mtt_addr = mlx4_mtt_addr(dev, &eq->mtt);
+	eq_context->mtt_base_addr_h = mtt_addr >> 32;
+	eq_context->mtt_base_addr_l = cpu_to_be32(mtt_addr & 0xffffffff);
+
+	err = mlx4_SW2HW_EQ(dev, mailbox, eq->eqn);
+	if (err) {
+		mlx4_warn(dev, "SW2HW_EQ failed (%d)\n", err);
+		goto err_out_free_mtt;
+	}
+
+	kfree(dma_list);
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+	eq->cons_index = 0;
+
+	return err;
+
+err_out_free_mtt:
+	mlx4_mtt_cleanup(dev, &eq->mtt);
+
+err_out_free_eq:
+	mlx4_bitmap_free(&priv->eq_table.bitmap, eq->eqn);
+
+err_out_free_pages:
+	for (i = 0; i < npages; ++i)
+		if (eq->page_list[i].buf)
+			dma_free_coherent(&dev->pdev->dev, PAGE_SIZE,
+					  eq->page_list[i].buf,
+					  eq->page_list[i].map);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+err_out_free:
+	kfree(eq->page_list);
+	kfree(dma_list);
+
+err_out:
+	return err;
+}
+
+static void mlx4_free_eq(struct mlx4_dev *dev,
+			 struct mlx4_eq *eq)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_cmd_mailbox *mailbox;
+	int err;
+	int npages = PAGE_ALIGN(MLX4_EQ_ENTRY_SIZE * eq->nent) / PAGE_SIZE;
+	int i;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return;
+
+	err = mlx4_HW2SW_EQ(dev, mailbox, eq->eqn);
+	if (err)
+		mlx4_warn(dev, "HW2SW_EQ failed (%d)\n", err);
+
+	if (0) {
+		mlx4_dbg(dev, "Dumping EQ context %02x:\n", eq->eqn);
+		for (i = 0; i < sizeof (struct mlx4_eq_context) / 4; ++i) {
+			if (i % 4 == 0)
+				printk("[%02x] ", i * 4);
+			printk(" %08x", be32_to_cpup(mailbox->buf + i * 4));
+			if ((i + 1) % 4 == 0)
+				printk("\n");
+		}
+	}
+
+	mlx4_mtt_cleanup(dev, &eq->mtt);
+	for (i = 0; i < npages; ++i)
+		pci_free_consistent(dev->pdev, PAGE_SIZE,
+				    eq->page_list[i].buf,
+				    eq->page_list[i].map);
+
+	kfree(eq->page_list);
+	mlx4_bitmap_free(&priv->eq_table.bitmap, eq->eqn);
+	mlx4_free_cmd_mailbox(dev, mailbox);
+}
+
+static void mlx4_free_irqs(struct mlx4_dev *dev)
+{
+	struct mlx4_eq_table *eq_table = &mlx4_priv(dev)->eq_table;
+	int i;
+
+	if (eq_table->have_irq)
+		free_irq(dev->pdev->irq, dev);
+	for (i = 0; i < dev->caps.num_comp_vectors + 1; ++i)
+		if (eq_table->eq[i].have_irq) {
+			free_irq(eq_table->eq[i].irq, eq_table->eq + i);
+			eq_table->eq[i].have_irq = 0;
+		}
+
+	kfree(eq_table->irq_names);
+}
+
+static int mlx4_map_clr_int(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	priv->clr_base = ioremap(pci_resource_start(dev->pdev, priv->fw.clr_int_bar) +
+				 priv->fw.clr_int_base, MLX4_CLR_INT_SIZE);
+	if (!priv->clr_base) {
+		mlx4_err(dev, "Couldn't map interrupt clear register, aborting.\n");
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+static void mlx4_unmap_clr_int(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	iounmap(priv->clr_base);
+}
+
+int mlx4_alloc_eq_table(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	priv->eq_table.eq = kcalloc(dev->caps.num_eqs - dev->caps.reserved_eqs,
+				    sizeof *priv->eq_table.eq, GFP_KERNEL);
+	if (!priv->eq_table.eq)
+		return -ENOMEM;
+
+	return 0;
+}
+
+void mlx4_free_eq_table(struct mlx4_dev *dev)
+{
+	kfree(mlx4_priv(dev)->eq_table.eq);
+}
+
+int mlx4_init_eq_table(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int err;
+	int i;
+
+	priv->eq_table.uar_map = kcalloc(sizeof *priv->eq_table.uar_map,
+					 mlx4_num_eq_uar(dev), GFP_KERNEL);
+	if (!priv->eq_table.uar_map) {
+		err = -ENOMEM;
+		goto err_out_free;
+	}
+
+	err = mlx4_bitmap_init(&priv->eq_table.bitmap, dev->caps.num_eqs,
+			       dev->caps.num_eqs - 1, dev->caps.reserved_eqs, 0);
+	if (err)
+		goto err_out_free;
+
+	for (i = 0; i < mlx4_num_eq_uar(dev); ++i)
+		priv->eq_table.uar_map[i] = NULL;
+
+	err = mlx4_map_clr_int(dev);
+	if (err)
+		goto err_out_bitmap;
+
+	priv->eq_table.clr_mask =
+		swab32(1 << (priv->eq_table.inta_pin & 31));
+	priv->eq_table.clr_int  = priv->clr_base +
+		(priv->eq_table.inta_pin < 32 ? 4 : 0);
+
+	priv->eq_table.irq_names =
+		kmalloc(MLX4_IRQNAME_SIZE * (dev->caps.num_comp_vectors + 1),
+			GFP_KERNEL);
+	if (!priv->eq_table.irq_names) {
+		err = -ENOMEM;
+		goto err_out_bitmap;
+	}
+
+	for (i = 0; i < dev->caps.num_comp_vectors; ++i) {
+		err = mlx4_create_eq(dev, dev->caps.num_cqs + MLX4_NUM_SPARE_EQE,
+				     (dev->flags & MLX4_FLAG_MSI_X) ? i : 0,
+				     &priv->eq_table.eq[i]);
+		if (err) {
+			--i;
+			goto err_out_unmap;
+		}
+	}
+
+	err = mlx4_create_eq(dev, MLX4_NUM_ASYNC_EQE + MLX4_NUM_SPARE_EQE,
+			     (dev->flags & MLX4_FLAG_MSI_X) ? dev->caps.num_comp_vectors : 0,
+			     &priv->eq_table.eq[dev->caps.num_comp_vectors]);
+	if (err)
+		goto err_out_comp;
+
+	if (dev->flags & MLX4_FLAG_MSI_X) {
+		const char *eq_name;
+
+		for (i = 0; i < dev->caps.num_comp_vectors + 1; ++i) {
+			if (i < dev->caps.num_comp_vectors) {
+				snprintf(priv->eq_table.irq_names +
+					 i * MLX4_IRQNAME_SIZE,
+					 MLX4_IRQNAME_SIZE,
+					 "mlx4-comp-%d@pci:%s", i,
+					 pci_name(dev->pdev));
+			} else {
+				snprintf(priv->eq_table.irq_names +
+					 i * MLX4_IRQNAME_SIZE,
+					 MLX4_IRQNAME_SIZE,
+					 "mlx4-async@pci:%s",
+					 pci_name(dev->pdev));
+			}
+
+			eq_name = priv->eq_table.irq_names +
+				  i * MLX4_IRQNAME_SIZE;
+			err = request_irq(priv->eq_table.eq[i].irq,
+					  mlx4_msi_x_interrupt, 0, eq_name,
+					  priv->eq_table.eq + i);
+			if (err)
+				goto err_out_async;
+
+			priv->eq_table.eq[i].have_irq = 1;
+		}
+	} else {
+		snprintf(priv->eq_table.irq_names,
+			 MLX4_IRQNAME_SIZE,
+			 DRV_NAME "@pci:%s",
+			 pci_name(dev->pdev));
+		err = request_irq(dev->pdev->irq, mlx4_interrupt,
+				  IRQF_SHARED, priv->eq_table.irq_names, dev);
+		if (err)
+			goto err_out_async;
+
+		priv->eq_table.have_irq = 1;
+	}
+
+	err = mlx4_MAP_EQ(dev, MLX4_ASYNC_EVENT_MASK, 0,
+			  priv->eq_table.eq[dev->caps.num_comp_vectors].eqn);
+	if (err)
+		mlx4_warn(dev, "MAP_EQ for async EQ %d failed (%d)\n",
+			   priv->eq_table.eq[dev->caps.num_comp_vectors].eqn, err);
+
+	for (i = 0; i < dev->caps.num_comp_vectors + 1; ++i)
+		eq_set_ci(&priv->eq_table.eq[i], 1);
+
+	return 0;
+
+err_out_async:
+	mlx4_free_eq(dev, &priv->eq_table.eq[dev->caps.num_comp_vectors]);
+
+err_out_comp:
+	i = dev->caps.num_comp_vectors - 1;
+
+err_out_unmap:
+	while (i >= 0) {
+		mlx4_free_eq(dev, &priv->eq_table.eq[i]);
+		--i;
+	}
+	mlx4_unmap_clr_int(dev);
+	mlx4_free_irqs(dev);
+
+err_out_bitmap:
+	mlx4_bitmap_cleanup(&priv->eq_table.bitmap);
+
+err_out_free:
+	kfree(priv->eq_table.uar_map);
+
+	return err;
+}
+
+void mlx4_cleanup_eq_table(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int i;
+
+	mlx4_MAP_EQ(dev, MLX4_ASYNC_EVENT_MASK, 1,
+		    priv->eq_table.eq[dev->caps.num_comp_vectors].eqn);
+
+	mlx4_free_irqs(dev);
+
+	for (i = 0; i < dev->caps.num_comp_vectors + 1; ++i)
+		mlx4_free_eq(dev, &priv->eq_table.eq[i]);
+
+	mlx4_unmap_clr_int(dev);
+
+	for (i = 0; i < mlx4_num_eq_uar(dev); ++i)
+		if (priv->eq_table.uar_map[i])
+			iounmap(priv->eq_table.uar_map[i]);
+
+	mlx4_bitmap_cleanup(&priv->eq_table.bitmap);
+
+	kfree(priv->eq_table.uar_map);
+}

kernel/drivers/net/mlx4/fw.c

@@ -0,0 +1,891 @@
+/*
+ * Copyright (c) 2004, 2005 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2005, 2006, 2007 Cisco Systems, 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/mlx4/cmd.h>
+#include <linux/cache.h>
+
+#include "fw.h"
+#include "icm.h"
+
+enum {
+	MLX4_COMMAND_INTERFACE_MIN_REV		= 2,
+	MLX4_COMMAND_INTERFACE_MAX_REV		= 3,
+	MLX4_COMMAND_INTERFACE_NEW_PORT_CMDS	= 3,
+};
+
+extern void __buggy_use_of_MLX4_GET(void);
+extern void __buggy_use_of_MLX4_PUT(void);
+
+static int enable_qos;
+module_param(enable_qos, bool, 0444);
+MODULE_PARM_DESC(enable_qos, "Enable Quality of Service support in the HCA (default: off)");
+
+#define MLX4_GET(dest, source, offset)				      \
+	do {							      \
+		void *__p = (char *) (source) + (offset);	      \
+		switch (sizeof (dest)) {			      \
+		case 1: (dest) = *(u8 *) __p;	    break;	      \
+		case 2: (dest) = be16_to_cpup(__p); break;	      \
+		case 4: (dest) = be32_to_cpup(__p); break;	      \
+		case 8: (dest) = be64_to_cpup(__p); break;	      \
+		default: __buggy_use_of_MLX4_GET();		      \
+		}						      \
+	} while (0)
+
+#define MLX4_PUT(dest, source, offset)				      \
+	do {							      \
+		void *__d = ((char *) (dest) + (offset));	      \
+		switch (sizeof(source)) {			      \
+		case 1: *(u8 *) __d = (source);		       break; \
+		case 2:	*(__be16 *) __d = cpu_to_be16(source); break; \
+		case 4:	*(__be32 *) __d = cpu_to_be32(source); break; \
+		case 8:	*(__be64 *) __d = cpu_to_be64(source); break; \
+		default: __buggy_use_of_MLX4_PUT();		      \
+		}						      \
+	} while (0)
+
+static void dump_dev_cap_flags(struct mlx4_dev *dev, u32 flags)
+{
+	static const char *fname[] = {
+		[ 0] = "RC transport",
+		[ 1] = "UC transport",
+		[ 2] = "UD transport",
+		[ 3] = "XRC transport",
+		[ 4] = "reliable multicast",
+		[ 5] = "FCoIB support",
+		[ 6] = "SRQ support",
+		[ 7] = "IPoIB checksum offload",
+		[ 8] = "P_Key violation counter",
+		[ 9] = "Q_Key violation counter",
+		[10] = "VMM",
+		[12] = "DPDP",
+		[16] = "MW support",
+		[17] = "APM support",
+		[18] = "Atomic ops support",
+		[19] = "Raw multicast support",
+		[20] = "Address vector port checking support",
+		[21] = "UD multicast support",
+		[24] = "Demand paging support",
+		[25] = "Router support"
+	};
+	int i;
+
+	mlx4_dbg(dev, "DEV_CAP flags:\n");
+	for (i = 0; i < ARRAY_SIZE(fname); ++i)
+		if (fname[i] && (flags & (1 << i)))
+			mlx4_dbg(dev, "    %s\n", fname[i]);
+}
+
+int mlx4_MOD_STAT_CFG(struct mlx4_dev *dev, struct mlx4_mod_stat_cfg *cfg)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	u32 *inbox;
+	int err = 0;
+
+#define MOD_STAT_CFG_IN_SIZE		0x100
+
+#define MOD_STAT_CFG_PG_SZ_M_OFFSET	0x002
+#define MOD_STAT_CFG_PG_SZ_OFFSET	0x003
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	inbox = mailbox->buf;
+
+	memset(inbox, 0, MOD_STAT_CFG_IN_SIZE);
+
+	MLX4_PUT(inbox, cfg->log_pg_sz, MOD_STAT_CFG_PG_SZ_OFFSET);
+	MLX4_PUT(inbox, cfg->log_pg_sz_m, MOD_STAT_CFG_PG_SZ_M_OFFSET);
+
+	err = mlx4_cmd(dev, mailbox->dma, 0, 0, MLX4_CMD_MOD_STAT_CFG,
+			MLX4_CMD_TIME_CLASS_A);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	u32 *outbox;
+	u8 field;
+	u16 size;
+	u16 stat_rate;
+	int err;
+	int i;
+
+#define QUERY_DEV_CAP_OUT_SIZE		       0x100
+#define QUERY_DEV_CAP_MAX_SRQ_SZ_OFFSET		0x10
+#define QUERY_DEV_CAP_MAX_QP_SZ_OFFSET		0x11
+#define QUERY_DEV_CAP_RSVD_QP_OFFSET		0x12
+#define QUERY_DEV_CAP_MAX_QP_OFFSET		0x13
+#define QUERY_DEV_CAP_RSVD_SRQ_OFFSET		0x14
+#define QUERY_DEV_CAP_MAX_SRQ_OFFSET		0x15
+#define QUERY_DEV_CAP_RSVD_EEC_OFFSET		0x16
+#define QUERY_DEV_CAP_MAX_EEC_OFFSET		0x17
+#define QUERY_DEV_CAP_MAX_CQ_SZ_OFFSET		0x19
+#define QUERY_DEV_CAP_RSVD_CQ_OFFSET		0x1a
+#define QUERY_DEV_CAP_MAX_CQ_OFFSET		0x1b
+#define QUERY_DEV_CAP_MAX_MPT_OFFSET		0x1d
+#define QUERY_DEV_CAP_RSVD_EQ_OFFSET		0x1e
+#define QUERY_DEV_CAP_MAX_EQ_OFFSET		0x1f
+#define QUERY_DEV_CAP_RSVD_MTT_OFFSET		0x20
+#define QUERY_DEV_CAP_MAX_MRW_SZ_OFFSET		0x21
+#define QUERY_DEV_CAP_RSVD_MRW_OFFSET		0x22
+#define QUERY_DEV_CAP_MAX_MTT_SEG_OFFSET	0x23
+#define QUERY_DEV_CAP_MAX_AV_OFFSET		0x27
+#define QUERY_DEV_CAP_MAX_REQ_QP_OFFSET		0x29
+#define QUERY_DEV_CAP_MAX_RES_QP_OFFSET		0x2b
+#define QUERY_DEV_CAP_MAX_GSO_OFFSET		0x2d
+#define QUERY_DEV_CAP_MAX_RDMA_OFFSET		0x2f
+#define QUERY_DEV_CAP_RSZ_SRQ_OFFSET		0x33
+#define QUERY_DEV_CAP_ACK_DELAY_OFFSET		0x35
+#define QUERY_DEV_CAP_MTU_WIDTH_OFFSET		0x36
+#define QUERY_DEV_CAP_VL_PORT_OFFSET		0x37
+#define QUERY_DEV_CAP_MAX_MSG_SZ_OFFSET		0x38
+#define QUERY_DEV_CAP_MAX_GID_OFFSET		0x3b
+#define QUERY_DEV_CAP_RATE_SUPPORT_OFFSET	0x3c
+#define QUERY_DEV_CAP_MAX_PKEY_OFFSET		0x3f
+#define QUERY_DEV_CAP_FLAGS_OFFSET		0x44
+#define QUERY_DEV_CAP_RSVD_UAR_OFFSET		0x48
+#define QUERY_DEV_CAP_UAR_SZ_OFFSET		0x49
+#define QUERY_DEV_CAP_PAGE_SZ_OFFSET		0x4b
+#define QUERY_DEV_CAP_BF_OFFSET			0x4c
+#define QUERY_DEV_CAP_LOG_BF_REG_SZ_OFFSET	0x4d
+#define QUERY_DEV_CAP_LOG_MAX_BF_REGS_PER_PAGE_OFFSET	0x4e
+#define QUERY_DEV_CAP_LOG_MAX_BF_PAGES_OFFSET	0x4f
+#define QUERY_DEV_CAP_MAX_SG_SQ_OFFSET		0x51
+#define QUERY_DEV_CAP_MAX_DESC_SZ_SQ_OFFSET	0x52
+#define QUERY_DEV_CAP_MAX_SG_RQ_OFFSET		0x55
+#define QUERY_DEV_CAP_MAX_DESC_SZ_RQ_OFFSET	0x56
+#define QUERY_DEV_CAP_MAX_QP_MCG_OFFSET		0x61
+#define QUERY_DEV_CAP_RSVD_MCG_OFFSET		0x62
+#define QUERY_DEV_CAP_MAX_MCG_OFFSET		0x63
+#define QUERY_DEV_CAP_RSVD_PD_OFFSET		0x64
+#define QUERY_DEV_CAP_MAX_PD_OFFSET		0x65
+#define QUERY_DEV_CAP_RDMARC_ENTRY_SZ_OFFSET	0x80
+#define QUERY_DEV_CAP_QPC_ENTRY_SZ_OFFSET	0x82
+#define QUERY_DEV_CAP_AUX_ENTRY_SZ_OFFSET	0x84
+#define QUERY_DEV_CAP_ALTC_ENTRY_SZ_OFFSET	0x86
+#define QUERY_DEV_CAP_EQC_ENTRY_SZ_OFFSET	0x88
+#define QUERY_DEV_CAP_CQC_ENTRY_SZ_OFFSET	0x8a
+#define QUERY_DEV_CAP_SRQ_ENTRY_SZ_OFFSET	0x8c
+#define QUERY_DEV_CAP_C_MPT_ENTRY_SZ_OFFSET	0x8e
+#define QUERY_DEV_CAP_MTT_ENTRY_SZ_OFFSET	0x90
+#define QUERY_DEV_CAP_D_MPT_ENTRY_SZ_OFFSET	0x92
+#define QUERY_DEV_CAP_BMME_FLAGS_OFFSET		0x94
+#define QUERY_DEV_CAP_RSVD_LKEY_OFFSET		0x98
+#define QUERY_DEV_CAP_MAX_ICM_SZ_OFFSET		0xa0
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	outbox = mailbox->buf;
+
+	err = mlx4_cmd_box(dev, 0, mailbox->dma, 0, 0, MLX4_CMD_QUERY_DEV_CAP,
+			   MLX4_CMD_TIME_CLASS_A);
+	if (err)
+		goto out;
+
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_RSVD_QP_OFFSET);
+	dev_cap->reserved_qps = 1 << (field & 0xf);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_QP_OFFSET);
+	dev_cap->max_qps = 1 << (field & 0x1f);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_RSVD_SRQ_OFFSET);
+	dev_cap->reserved_srqs = 1 << (field >> 4);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_SRQ_OFFSET);
+	dev_cap->max_srqs = 1 << (field & 0x1f);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_CQ_SZ_OFFSET);
+	dev_cap->max_cq_sz = 1 << field;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_RSVD_CQ_OFFSET);
+	dev_cap->reserved_cqs = 1 << (field & 0xf);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_CQ_OFFSET);
+	dev_cap->max_cqs = 1 << (field & 0x1f);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_MPT_OFFSET);
+	dev_cap->max_mpts = 1 << (field & 0x3f);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_RSVD_EQ_OFFSET);
+	dev_cap->reserved_eqs = 1 << (field & 0xf);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_EQ_OFFSET);
+	dev_cap->max_eqs = 1 << (field & 0xf);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_RSVD_MTT_OFFSET);
+	dev_cap->reserved_mtts = 1 << (field >> 4);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_MRW_SZ_OFFSET);
+	dev_cap->max_mrw_sz = 1 << field;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_RSVD_MRW_OFFSET);
+	dev_cap->reserved_mrws = 1 << (field & 0xf);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_MTT_SEG_OFFSET);
+	dev_cap->max_mtt_seg = 1 << (field & 0x3f);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_REQ_QP_OFFSET);
+	dev_cap->max_requester_per_qp = 1 << (field & 0x3f);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_RES_QP_OFFSET);
+	dev_cap->max_responder_per_qp = 1 << (field & 0x3f);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_GSO_OFFSET);
+	field &= 0x1f;
+	if (!field)
+		dev_cap->max_gso_sz = 0;
+	else
+		dev_cap->max_gso_sz = 1 << field;
+
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_RDMA_OFFSET);
+	dev_cap->max_rdma_global = 1 << (field & 0x3f);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_ACK_DELAY_OFFSET);
+	dev_cap->local_ca_ack_delay = field & 0x1f;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_VL_PORT_OFFSET);
+	dev_cap->num_ports = field & 0xf;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_MSG_SZ_OFFSET);
+	dev_cap->max_msg_sz = 1 << (field & 0x1f);
+	MLX4_GET(stat_rate, outbox, QUERY_DEV_CAP_RATE_SUPPORT_OFFSET);
+	dev_cap->stat_rate_support = stat_rate;
+	MLX4_GET(dev_cap->flags, outbox, QUERY_DEV_CAP_FLAGS_OFFSET);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_RSVD_UAR_OFFSET);
+	dev_cap->reserved_uars = field >> 4;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_UAR_SZ_OFFSET);
+	dev_cap->uar_size = 1 << ((field & 0x3f) + 20);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_PAGE_SZ_OFFSET);
+	dev_cap->min_page_sz = 1 << field;
+
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_BF_OFFSET);
+	if (field & 0x80) {
+		MLX4_GET(field, outbox, QUERY_DEV_CAP_LOG_BF_REG_SZ_OFFSET);
+		dev_cap->bf_reg_size = 1 << (field & 0x1f);
+		MLX4_GET(field, outbox, QUERY_DEV_CAP_LOG_MAX_BF_REGS_PER_PAGE_OFFSET);
+		dev_cap->bf_regs_per_page = 1 << (field & 0x3f);
+		mlx4_dbg(dev, "BlueFlame available (reg size %d, regs/page %d)\n",
+			 dev_cap->bf_reg_size, dev_cap->bf_regs_per_page);
+	} else {
+		dev_cap->bf_reg_size = 0;
+		mlx4_dbg(dev, "BlueFlame not available\n");
+	}
+
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_SG_SQ_OFFSET);
+	dev_cap->max_sq_sg = field;
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_MAX_DESC_SZ_SQ_OFFSET);
+	dev_cap->max_sq_desc_sz = size;
+
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_QP_MCG_OFFSET);
+	dev_cap->max_qp_per_mcg = 1 << field;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_RSVD_MCG_OFFSET);
+	dev_cap->reserved_mgms = field & 0xf;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_MCG_OFFSET);
+	dev_cap->max_mcgs = 1 << field;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_RSVD_PD_OFFSET);
+	dev_cap->reserved_pds = field >> 4;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_PD_OFFSET);
+	dev_cap->max_pds = 1 << (field & 0x3f);
+
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_RDMARC_ENTRY_SZ_OFFSET);
+	dev_cap->rdmarc_entry_sz = size;
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_QPC_ENTRY_SZ_OFFSET);
+	dev_cap->qpc_entry_sz = size;
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_AUX_ENTRY_SZ_OFFSET);
+	dev_cap->aux_entry_sz = size;
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_ALTC_ENTRY_SZ_OFFSET);
+	dev_cap->altc_entry_sz = size;
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_EQC_ENTRY_SZ_OFFSET);
+	dev_cap->eqc_entry_sz = size;
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_CQC_ENTRY_SZ_OFFSET);
+	dev_cap->cqc_entry_sz = size;
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_SRQ_ENTRY_SZ_OFFSET);
+	dev_cap->srq_entry_sz = size;
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_C_MPT_ENTRY_SZ_OFFSET);
+	dev_cap->cmpt_entry_sz = size;
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_MTT_ENTRY_SZ_OFFSET);
+	dev_cap->mtt_entry_sz = size;
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_D_MPT_ENTRY_SZ_OFFSET);
+	dev_cap->dmpt_entry_sz = size;
+
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_SRQ_SZ_OFFSET);
+	dev_cap->max_srq_sz = 1 << field;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_QP_SZ_OFFSET);
+	dev_cap->max_qp_sz = 1 << field;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_RSZ_SRQ_OFFSET);
+	dev_cap->resize_srq = field & 1;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_SG_RQ_OFFSET);
+	dev_cap->max_rq_sg = field;
+	MLX4_GET(size, outbox, QUERY_DEV_CAP_MAX_DESC_SZ_RQ_OFFSET);
+	dev_cap->max_rq_desc_sz = size;
+
+	MLX4_GET(dev_cap->bmme_flags, outbox,
+		 QUERY_DEV_CAP_BMME_FLAGS_OFFSET);
+	MLX4_GET(dev_cap->reserved_lkey, outbox,
+		 QUERY_DEV_CAP_RSVD_LKEY_OFFSET);
+	MLX4_GET(dev_cap->max_icm_sz, outbox,
+		 QUERY_DEV_CAP_MAX_ICM_SZ_OFFSET);
+
+	if (dev->flags & MLX4_FLAG_OLD_PORT_CMDS) {
+		for (i = 1; i <= dev_cap->num_ports; ++i) {
+			MLX4_GET(field, outbox, QUERY_DEV_CAP_VL_PORT_OFFSET);
+			dev_cap->max_vl[i]	   = field >> 4;
+			MLX4_GET(field, outbox, QUERY_DEV_CAP_MTU_WIDTH_OFFSET);
+			dev_cap->ib_mtu[i]	   = field >> 4;
+			dev_cap->max_port_width[i] = field & 0xf;
+			MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_GID_OFFSET);
+			dev_cap->max_gids[i]	   = 1 << (field & 0xf);
+			MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_PKEY_OFFSET);
+			dev_cap->max_pkeys[i]	   = 1 << (field & 0xf);
+		}
+	} else {
+#define QUERY_PORT_SUPPORTED_TYPE_OFFSET	0x00
+#define QUERY_PORT_MTU_OFFSET			0x01
+#define QUERY_PORT_ETH_MTU_OFFSET		0x02
+#define QUERY_PORT_WIDTH_OFFSET			0x06
+#define QUERY_PORT_MAX_GID_PKEY_OFFSET		0x07
+#define QUERY_PORT_MAX_MACVLAN_OFFSET		0x0a
+#define QUERY_PORT_MAX_VL_OFFSET		0x0b
+#define QUERY_PORT_MAC_OFFSET			0x10
+
+		for (i = 1; i <= dev_cap->num_ports; ++i) {
+			err = mlx4_cmd_box(dev, 0, mailbox->dma, i, 0, MLX4_CMD_QUERY_PORT,
+					   MLX4_CMD_TIME_CLASS_B);
+			if (err)
+				goto out;
+
+			MLX4_GET(field, outbox, QUERY_PORT_SUPPORTED_TYPE_OFFSET);
+			dev_cap->supported_port_types[i] = field & 3;
+			MLX4_GET(field, outbox, QUERY_PORT_MTU_OFFSET);
+			dev_cap->ib_mtu[i]	   = field & 0xf;
+			MLX4_GET(field, outbox, QUERY_PORT_WIDTH_OFFSET);
+			dev_cap->max_port_width[i] = field & 0xf;
+			MLX4_GET(field, outbox, QUERY_PORT_MAX_GID_PKEY_OFFSET);
+			dev_cap->max_gids[i]	   = 1 << (field >> 4);
+			dev_cap->max_pkeys[i]	   = 1 << (field & 0xf);
+			MLX4_GET(field, outbox, QUERY_PORT_MAX_VL_OFFSET);
+			dev_cap->max_vl[i]	   = field & 0xf;
+			MLX4_GET(field, outbox, QUERY_PORT_MAX_MACVLAN_OFFSET);
+			dev_cap->log_max_macs[i]  = field & 0xf;
+			dev_cap->log_max_vlans[i] = field >> 4;
+			MLX4_GET(dev_cap->eth_mtu[i], outbox, QUERY_PORT_ETH_MTU_OFFSET);
+			MLX4_GET(dev_cap->def_mac[i], outbox, QUERY_PORT_MAC_OFFSET);
+		}
+	}
+
+	mlx4_dbg(dev, "Base MM extensions: flags %08x, rsvd L_Key %08x\n",
+		 dev_cap->bmme_flags, dev_cap->reserved_lkey);
+
+	/*
+	 * Each UAR has 4 EQ doorbells; so if a UAR is reserved, then
+	 * we can't use any EQs whose doorbell falls on that page,
+	 * even if the EQ itself isn't reserved.
+	 */
+	dev_cap->reserved_eqs = max(dev_cap->reserved_uars * 4,
+				    dev_cap->reserved_eqs);
+
+	mlx4_dbg(dev, "Max ICM size %lld MB\n",
+		 (unsigned long long) dev_cap->max_icm_sz >> 20);
+	mlx4_dbg(dev, "Max QPs: %d, reserved QPs: %d, entry size: %d\n",
+		 dev_cap->max_qps, dev_cap->reserved_qps, dev_cap->qpc_entry_sz);
+	mlx4_dbg(dev, "Max SRQs: %d, reserved SRQs: %d, entry size: %d\n",
+		 dev_cap->max_srqs, dev_cap->reserved_srqs, dev_cap->srq_entry_sz);
+	mlx4_dbg(dev, "Max CQs: %d, reserved CQs: %d, entry size: %d\n",
+		 dev_cap->max_cqs, dev_cap->reserved_cqs, dev_cap->cqc_entry_sz);
+	mlx4_dbg(dev, "Max EQs: %d, reserved EQs: %d, entry size: %d\n",
+		 dev_cap->max_eqs, dev_cap->reserved_eqs, dev_cap->eqc_entry_sz);
+	mlx4_dbg(dev, "reserved MPTs: %d, reserved MTTs: %d\n",
+		 dev_cap->reserved_mrws, dev_cap->reserved_mtts);
+	mlx4_dbg(dev, "Max PDs: %d, reserved PDs: %d, reserved UARs: %d\n",
+		 dev_cap->max_pds, dev_cap->reserved_pds, dev_cap->reserved_uars);
+	mlx4_dbg(dev, "Max QP/MCG: %d, reserved MGMs: %d\n",
+		 dev_cap->max_pds, dev_cap->reserved_mgms);
+	mlx4_dbg(dev, "Max CQEs: %d, max WQEs: %d, max SRQ WQEs: %d\n",
+		 dev_cap->max_cq_sz, dev_cap->max_qp_sz, dev_cap->max_srq_sz);
+	mlx4_dbg(dev, "Local CA ACK delay: %d, max MTU: %d, port width cap: %d\n",
+		 dev_cap->local_ca_ack_delay, 128 << dev_cap->ib_mtu[1],
+		 dev_cap->max_port_width[1]);
+	mlx4_dbg(dev, "Max SQ desc size: %d, max SQ S/G: %d\n",
+		 dev_cap->max_sq_desc_sz, dev_cap->max_sq_sg);
+	mlx4_dbg(dev, "Max RQ desc size: %d, max RQ S/G: %d\n",
+		 dev_cap->max_rq_desc_sz, dev_cap->max_rq_sg);
+	mlx4_dbg(dev, "Max GSO size: %d\n", dev_cap->max_gso_sz);
+
+	dump_dev_cap_flags(dev, dev_cap->flags);
+
+out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+int mlx4_map_cmd(struct mlx4_dev *dev, u16 op, struct mlx4_icm *icm, u64 virt)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_icm_iter iter;
+	__be64 *pages;
+	int lg;
+	int nent = 0;
+	int i;
+	int err = 0;
+	int ts = 0, tc = 0;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	memset(mailbox->buf, 0, MLX4_MAILBOX_SIZE);
+	pages = mailbox->buf;
+
+	for (mlx4_icm_first(icm, &iter);
+	     !mlx4_icm_last(&iter);
+	     mlx4_icm_next(&iter)) {
+		/*
+		 * We have to pass pages that are aligned to their
+		 * size, so find the least significant 1 in the
+		 * address or size and use that as our log2 size.
+		 */
+		lg = ffs(mlx4_icm_addr(&iter) | mlx4_icm_size(&iter)) - 1;
+		if (lg < MLX4_ICM_PAGE_SHIFT) {
+			mlx4_warn(dev, "Got FW area not aligned to %d (%llx/%lx).\n",
+				   MLX4_ICM_PAGE_SIZE,
+				   (unsigned long long) mlx4_icm_addr(&iter),
+				   mlx4_icm_size(&iter));
+			err = -EINVAL;
+			goto out;
+		}
+
+		for (i = 0; i < mlx4_icm_size(&iter) >> lg; ++i) {
+			if (virt != -1) {
+				pages[nent * 2] = cpu_to_be64(virt);
+				virt += 1 << lg;
+			}
+
+			pages[nent * 2 + 1] =
+				cpu_to_be64((mlx4_icm_addr(&iter) + (i << lg)) |
+					    (lg - MLX4_ICM_PAGE_SHIFT));
+			ts += 1 << (lg - 10);
+			++tc;
+
+			if (++nent == MLX4_MAILBOX_SIZE / 16) {
+				err = mlx4_cmd(dev, mailbox->dma, nent, 0, op,
+						MLX4_CMD_TIME_CLASS_B);
+				if (err)
+					goto out;
+				nent = 0;
+			}
+		}
+	}
+
+	if (nent)
+		err = mlx4_cmd(dev, mailbox->dma, nent, 0, op, MLX4_CMD_TIME_CLASS_B);
+	if (err)
+		goto out;
+
+	switch (op) {
+	case MLX4_CMD_MAP_FA:
+		mlx4_dbg(dev, "Mapped %d chunks/%d KB for FW.\n", tc, ts);
+		break;
+	case MLX4_CMD_MAP_ICM_AUX:
+		mlx4_dbg(dev, "Mapped %d chunks/%d KB for ICM aux.\n", tc, ts);
+		break;
+	case MLX4_CMD_MAP_ICM:
+		mlx4_dbg(dev, "Mapped %d chunks/%d KB at %llx for ICM.\n",
+			  tc, ts, (unsigned long long) virt - (ts << 10));
+		break;
+	}
+
+out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+int mlx4_MAP_FA(struct mlx4_dev *dev, struct mlx4_icm *icm)
+{
+	return mlx4_map_cmd(dev, MLX4_CMD_MAP_FA, icm, -1);
+}
+
+int mlx4_UNMAP_FA(struct mlx4_dev *dev)
+{
+	return mlx4_cmd(dev, 0, 0, 0, MLX4_CMD_UNMAP_FA, MLX4_CMD_TIME_CLASS_B);
+}
+
+
+int mlx4_RUN_FW(struct mlx4_dev *dev)
+{
+	return mlx4_cmd(dev, 0, 0, 0, MLX4_CMD_RUN_FW, MLX4_CMD_TIME_CLASS_A);
+}
+
+int mlx4_QUERY_FW(struct mlx4_dev *dev)
+{
+	struct mlx4_fw  *fw  = &mlx4_priv(dev)->fw;
+	struct mlx4_cmd *cmd = &mlx4_priv(dev)->cmd;
+	struct mlx4_cmd_mailbox *mailbox;
+	u32 *outbox;
+	int err = 0;
+	u64 fw_ver;
+	u16 cmd_if_rev;
+	u8 lg;
+
+#define QUERY_FW_OUT_SIZE             0x100
+#define QUERY_FW_VER_OFFSET            0x00
+#define QUERY_FW_CMD_IF_REV_OFFSET     0x0a
+#define QUERY_FW_MAX_CMD_OFFSET        0x0f
+#define QUERY_FW_ERR_START_OFFSET      0x30
+#define QUERY_FW_ERR_SIZE_OFFSET       0x38
+#define QUERY_FW_ERR_BAR_OFFSET        0x3c
+
+#define QUERY_FW_SIZE_OFFSET           0x00
+#define QUERY_FW_CLR_INT_BASE_OFFSET   0x20
+#define QUERY_FW_CLR_INT_BAR_OFFSET    0x28
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	outbox = mailbox->buf;
+
+	err = mlx4_cmd_box(dev, 0, mailbox->dma, 0, 0, MLX4_CMD_QUERY_FW,
+			    MLX4_CMD_TIME_CLASS_A);
+	if (err)
+		goto out;
+
+	MLX4_GET(fw_ver, outbox, QUERY_FW_VER_OFFSET);
+	/*
+	 * FW subminor version is at more significant bits than minor
+	 * version, so swap here.
+	 */
+	dev->caps.fw_ver = (fw_ver & 0xffff00000000ull) |
+		((fw_ver & 0xffff0000ull) >> 16) |
+		((fw_ver & 0x0000ffffull) << 16);
+
+	MLX4_GET(cmd_if_rev, outbox, QUERY_FW_CMD_IF_REV_OFFSET);
+	if (cmd_if_rev < MLX4_COMMAND_INTERFACE_MIN_REV ||
+	    cmd_if_rev > MLX4_COMMAND_INTERFACE_MAX_REV) {
+		mlx4_err(dev, "Installed FW has unsupported "
+			 "command interface revision %d.\n",
+			 cmd_if_rev);
+		mlx4_err(dev, "(Installed FW version is %d.%d.%03d)\n",
+			 (int) (dev->caps.fw_ver >> 32),
+			 (int) (dev->caps.fw_ver >> 16) & 0xffff,
+			 (int) dev->caps.fw_ver & 0xffff);
+		mlx4_err(dev, "This driver version supports only revisions %d to %d.\n",
+			 MLX4_COMMAND_INTERFACE_MIN_REV, MLX4_COMMAND_INTERFACE_MAX_REV);
+		err = -ENODEV;
+		goto out;
+	}
+
+	if (cmd_if_rev < MLX4_COMMAND_INTERFACE_NEW_PORT_CMDS)
+		dev->flags |= MLX4_FLAG_OLD_PORT_CMDS;
+
+	MLX4_GET(lg, outbox, QUERY_FW_MAX_CMD_OFFSET);
+	cmd->max_cmds = 1 << lg;
+
+	mlx4_dbg(dev, "FW version %d.%d.%03d (cmd intf rev %d), max commands %d\n",
+		 (int) (dev->caps.fw_ver >> 32),
+		 (int) (dev->caps.fw_ver >> 16) & 0xffff,
+		 (int) dev->caps.fw_ver & 0xffff,
+		 cmd_if_rev, cmd->max_cmds);
+
+	MLX4_GET(fw->catas_offset, outbox, QUERY_FW_ERR_START_OFFSET);
+	MLX4_GET(fw->catas_size,   outbox, QUERY_FW_ERR_SIZE_OFFSET);
+	MLX4_GET(fw->catas_bar,    outbox, QUERY_FW_ERR_BAR_OFFSET);
+	fw->catas_bar = (fw->catas_bar >> 6) * 2;
+
+	mlx4_dbg(dev, "Catastrophic error buffer at 0x%llx, size 0x%x, BAR %d\n",
+		 (unsigned long long) fw->catas_offset, fw->catas_size, fw->catas_bar);
+
+	MLX4_GET(fw->fw_pages,     outbox, QUERY_FW_SIZE_OFFSET);
+	MLX4_GET(fw->clr_int_base, outbox, QUERY_FW_CLR_INT_BASE_OFFSET);
+	MLX4_GET(fw->clr_int_bar,  outbox, QUERY_FW_CLR_INT_BAR_OFFSET);
+	fw->clr_int_bar = (fw->clr_int_bar >> 6) * 2;
+
+	mlx4_dbg(dev, "FW size %d KB\n", fw->fw_pages >> 2);
+
+	/*
+	 * Round up number of system pages needed in case
+	 * MLX4_ICM_PAGE_SIZE < PAGE_SIZE.
+	 */
+	fw->fw_pages =
+		ALIGN(fw->fw_pages, PAGE_SIZE / MLX4_ICM_PAGE_SIZE) >>
+		(PAGE_SHIFT - MLX4_ICM_PAGE_SHIFT);
+
+	mlx4_dbg(dev, "Clear int @ %llx, BAR %d\n",
+		 (unsigned long long) fw->clr_int_base, fw->clr_int_bar);
+
+out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+static void get_board_id(void *vsd, char *board_id)
+{
+	int i;
+
+#define VSD_OFFSET_SIG1		0x00
+#define VSD_OFFSET_SIG2		0xde
+#define VSD_OFFSET_MLX_BOARD_ID	0xd0
+#define VSD_OFFSET_TS_BOARD_ID	0x20
+
+#define VSD_SIGNATURE_TOPSPIN	0x5ad
+
+	memset(board_id, 0, MLX4_BOARD_ID_LEN);
+
+	if (be16_to_cpup(vsd + VSD_OFFSET_SIG1) == VSD_SIGNATURE_TOPSPIN &&
+	    be16_to_cpup(vsd + VSD_OFFSET_SIG2) == VSD_SIGNATURE_TOPSPIN) {
+		strlcpy(board_id, vsd + VSD_OFFSET_TS_BOARD_ID, MLX4_BOARD_ID_LEN);
+	} else {
+		/*
+		 * The board ID is a string but the firmware byte
+		 * swaps each 4-byte word before passing it back to
+		 * us.  Therefore we need to swab it before printing.
+		 */
+		for (i = 0; i < 4; ++i)
+			((u32 *) board_id)[i] =
+				swab32(*(u32 *) (vsd + VSD_OFFSET_MLX_BOARD_ID + i * 4));
+	}
+}
+
+int mlx4_QUERY_ADAPTER(struct mlx4_dev *dev, struct mlx4_adapter *adapter)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	u32 *outbox;
+	int err;
+
+#define QUERY_ADAPTER_OUT_SIZE             0x100
+#define QUERY_ADAPTER_INTA_PIN_OFFSET      0x10
+#define QUERY_ADAPTER_VSD_OFFSET           0x20
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	outbox = mailbox->buf;
+
+	err = mlx4_cmd_box(dev, 0, mailbox->dma, 0, 0, MLX4_CMD_QUERY_ADAPTER,
+			   MLX4_CMD_TIME_CLASS_A);
+	if (err)
+		goto out;
+
+	MLX4_GET(adapter->inta_pin, outbox,    QUERY_ADAPTER_INTA_PIN_OFFSET);
+
+	get_board_id(outbox + QUERY_ADAPTER_VSD_OFFSET / 4,
+		     adapter->board_id);
+
+out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+int mlx4_INIT_HCA(struct mlx4_dev *dev, struct mlx4_init_hca_param *param)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	__be32 *inbox;
+	int err;
+
+#define INIT_HCA_IN_SIZE		 0x200
+#define INIT_HCA_VERSION_OFFSET		 0x000
+#define	 INIT_HCA_VERSION		 2
+#define INIT_HCA_CACHELINE_SZ_OFFSET	 0x0e
+#define INIT_HCA_FLAGS_OFFSET		 0x014
+#define INIT_HCA_QPC_OFFSET		 0x020
+#define	 INIT_HCA_QPC_BASE_OFFSET	 (INIT_HCA_QPC_OFFSET + 0x10)
+#define	 INIT_HCA_LOG_QP_OFFSET		 (INIT_HCA_QPC_OFFSET + 0x17)
+#define	 INIT_HCA_SRQC_BASE_OFFSET	 (INIT_HCA_QPC_OFFSET + 0x28)
+#define	 INIT_HCA_LOG_SRQ_OFFSET	 (INIT_HCA_QPC_OFFSET + 0x2f)
+#define	 INIT_HCA_CQC_BASE_OFFSET	 (INIT_HCA_QPC_OFFSET + 0x30)
+#define	 INIT_HCA_LOG_CQ_OFFSET		 (INIT_HCA_QPC_OFFSET + 0x37)
+#define	 INIT_HCA_ALTC_BASE_OFFSET	 (INIT_HCA_QPC_OFFSET + 0x40)
+#define	 INIT_HCA_AUXC_BASE_OFFSET	 (INIT_HCA_QPC_OFFSET + 0x50)
+#define	 INIT_HCA_EQC_BASE_OFFSET	 (INIT_HCA_QPC_OFFSET + 0x60)
+#define	 INIT_HCA_LOG_EQ_OFFSET		 (INIT_HCA_QPC_OFFSET + 0x67)
+#define	 INIT_HCA_RDMARC_BASE_OFFSET	 (INIT_HCA_QPC_OFFSET + 0x70)
+#define	 INIT_HCA_LOG_RD_OFFSET		 (INIT_HCA_QPC_OFFSET + 0x77)
+#define INIT_HCA_MCAST_OFFSET		 0x0c0
+#define	 INIT_HCA_MC_BASE_OFFSET	 (INIT_HCA_MCAST_OFFSET + 0x00)
+#define	 INIT_HCA_LOG_MC_ENTRY_SZ_OFFSET (INIT_HCA_MCAST_OFFSET + 0x12)
+#define	 INIT_HCA_LOG_MC_HASH_SZ_OFFSET	 (INIT_HCA_MCAST_OFFSET + 0x16)
+#define	 INIT_HCA_LOG_MC_TABLE_SZ_OFFSET (INIT_HCA_MCAST_OFFSET + 0x1b)
+#define INIT_HCA_TPT_OFFSET		 0x0f0
+#define	 INIT_HCA_DMPT_BASE_OFFSET	 (INIT_HCA_TPT_OFFSET + 0x00)
+#define	 INIT_HCA_LOG_MPT_SZ_OFFSET	 (INIT_HCA_TPT_OFFSET + 0x0b)
+#define	 INIT_HCA_MTT_BASE_OFFSET	 (INIT_HCA_TPT_OFFSET + 0x10)
+#define	 INIT_HCA_CMPT_BASE_OFFSET	 (INIT_HCA_TPT_OFFSET + 0x18)
+#define INIT_HCA_UAR_OFFSET		 0x120
+#define	 INIT_HCA_LOG_UAR_SZ_OFFSET	 (INIT_HCA_UAR_OFFSET + 0x0a)
+#define  INIT_HCA_UAR_PAGE_SZ_OFFSET     (INIT_HCA_UAR_OFFSET + 0x0b)
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	inbox = mailbox->buf;
+
+	memset(inbox, 0, INIT_HCA_IN_SIZE);
+
+	*((u8 *) mailbox->buf + INIT_HCA_VERSION_OFFSET) = INIT_HCA_VERSION;
+
+	*((u8 *) mailbox->buf + INIT_HCA_CACHELINE_SZ_OFFSET) =
+		(ilog2(cache_line_size()) - 4) << 5;
+
+#if defined(__LITTLE_ENDIAN)
+	*(inbox + INIT_HCA_FLAGS_OFFSET / 4) &= ~cpu_to_be32(1 << 1);
+#elif defined(__BIG_ENDIAN)
+	*(inbox + INIT_HCA_FLAGS_OFFSET / 4) |= cpu_to_be32(1 << 1);
+#else
+#error Host endianness not defined
+#endif
+	/* Check port for UD address vector: */
+	*(inbox + INIT_HCA_FLAGS_OFFSET / 4) |= cpu_to_be32(1);
+
+	/* Enable IPoIB checksumming if we can: */
+	if (dev->caps.flags & MLX4_DEV_CAP_FLAG_IPOIB_CSUM)
+		*(inbox + INIT_HCA_FLAGS_OFFSET / 4) |= cpu_to_be32(1 << 3);
+
+	/* Enable QoS support if module parameter set */
+	if (enable_qos)
+		*(inbox + INIT_HCA_FLAGS_OFFSET / 4) |= cpu_to_be32(1 << 2);
+
+	/* QPC/EEC/CQC/EQC/RDMARC attributes */
+
+	MLX4_PUT(inbox, param->qpc_base,      INIT_HCA_QPC_BASE_OFFSET);
+	MLX4_PUT(inbox, param->log_num_qps,   INIT_HCA_LOG_QP_OFFSET);
+	MLX4_PUT(inbox, param->srqc_base,     INIT_HCA_SRQC_BASE_OFFSET);
+	MLX4_PUT(inbox, param->log_num_srqs,  INIT_HCA_LOG_SRQ_OFFSET);
+	MLX4_PUT(inbox, param->cqc_base,      INIT_HCA_CQC_BASE_OFFSET);
+	MLX4_PUT(inbox, param->log_num_cqs,   INIT_HCA_LOG_CQ_OFFSET);
+	MLX4_PUT(inbox, param->altc_base,     INIT_HCA_ALTC_BASE_OFFSET);
+	MLX4_PUT(inbox, param->auxc_base,     INIT_HCA_AUXC_BASE_OFFSET);
+	MLX4_PUT(inbox, param->eqc_base,      INIT_HCA_EQC_BASE_OFFSET);
+	MLX4_PUT(inbox, param->log_num_eqs,   INIT_HCA_LOG_EQ_OFFSET);
+	MLX4_PUT(inbox, param->rdmarc_base,   INIT_HCA_RDMARC_BASE_OFFSET);
+	MLX4_PUT(inbox, param->log_rd_per_qp, INIT_HCA_LOG_RD_OFFSET);
+
+	/* multicast attributes */
+
+	MLX4_PUT(inbox, param->mc_base,		INIT_HCA_MC_BASE_OFFSET);
+	MLX4_PUT(inbox, param->log_mc_entry_sz, INIT_HCA_LOG_MC_ENTRY_SZ_OFFSET);
+	MLX4_PUT(inbox, param->log_mc_hash_sz,  INIT_HCA_LOG_MC_HASH_SZ_OFFSET);
+	MLX4_PUT(inbox, param->log_mc_table_sz, INIT_HCA_LOG_MC_TABLE_SZ_OFFSET);
+
+	/* TPT attributes */
+
+	MLX4_PUT(inbox, param->dmpt_base,  INIT_HCA_DMPT_BASE_OFFSET);
+	MLX4_PUT(inbox, param->log_mpt_sz, INIT_HCA_LOG_MPT_SZ_OFFSET);
+	MLX4_PUT(inbox, param->mtt_base,   INIT_HCA_MTT_BASE_OFFSET);
+	MLX4_PUT(inbox, param->cmpt_base,  INIT_HCA_CMPT_BASE_OFFSET);
+
+	/* UAR attributes */
+
+	MLX4_PUT(inbox, (u8) (PAGE_SHIFT - 12), INIT_HCA_UAR_PAGE_SZ_OFFSET);
+	MLX4_PUT(inbox, param->log_uar_sz,      INIT_HCA_LOG_UAR_SZ_OFFSET);
+
+	err = mlx4_cmd(dev, mailbox->dma, 0, 0, MLX4_CMD_INIT_HCA, 10000);
+
+	if (err)
+		mlx4_err(dev, "INIT_HCA returns %d\n", err);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+int mlx4_INIT_PORT(struct mlx4_dev *dev, int port)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	u32 *inbox;
+	int err;
+	u32 flags;
+	u16 field;
+
+	if (dev->flags & MLX4_FLAG_OLD_PORT_CMDS) {
+#define INIT_PORT_IN_SIZE          256
+#define INIT_PORT_FLAGS_OFFSET     0x00
+#define INIT_PORT_FLAG_SIG         (1 << 18)
+#define INIT_PORT_FLAG_NG          (1 << 17)
+#define INIT_PORT_FLAG_G0          (1 << 16)
+#define INIT_PORT_VL_SHIFT         4
+#define INIT_PORT_PORT_WIDTH_SHIFT 8
+#define INIT_PORT_MTU_OFFSET       0x04
+#define INIT_PORT_MAX_GID_OFFSET   0x06
+#define INIT_PORT_MAX_PKEY_OFFSET  0x0a
+#define INIT_PORT_GUID0_OFFSET     0x10
+#define INIT_PORT_NODE_GUID_OFFSET 0x18
+#define INIT_PORT_SI_GUID_OFFSET   0x20
+
+		mailbox = mlx4_alloc_cmd_mailbox(dev);
+		if (IS_ERR(mailbox))
+			return PTR_ERR(mailbox);
+		inbox = mailbox->buf;
+
+		memset(inbox, 0, INIT_PORT_IN_SIZE);
+
+		flags = 0;
+		flags |= (dev->caps.vl_cap[port] & 0xf) << INIT_PORT_VL_SHIFT;
+		flags |= (dev->caps.port_width_cap[port] & 0xf) << INIT_PORT_PORT_WIDTH_SHIFT;
+		MLX4_PUT(inbox, flags,		  INIT_PORT_FLAGS_OFFSET);
+
+		field = 128 << dev->caps.ib_mtu_cap[port];
+		MLX4_PUT(inbox, field, INIT_PORT_MTU_OFFSET);
+		field = dev->caps.gid_table_len[port];
+		MLX4_PUT(inbox, field, INIT_PORT_MAX_GID_OFFSET);
+		field = dev->caps.pkey_table_len[port];
+		MLX4_PUT(inbox, field, INIT_PORT_MAX_PKEY_OFFSET);
+
+		err = mlx4_cmd(dev, mailbox->dma, port, 0, MLX4_CMD_INIT_PORT,
+			       MLX4_CMD_TIME_CLASS_A);
+
+		mlx4_free_cmd_mailbox(dev, mailbox);
+	} else
+		err = mlx4_cmd(dev, 0, port, 0, MLX4_CMD_INIT_PORT,
+			       MLX4_CMD_TIME_CLASS_A);
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_INIT_PORT);
+
+int mlx4_CLOSE_PORT(struct mlx4_dev *dev, int port)
+{
+	return mlx4_cmd(dev, 0, port, 0, MLX4_CMD_CLOSE_PORT, 1000);
+}
+EXPORT_SYMBOL_GPL(mlx4_CLOSE_PORT);
+
+int mlx4_CLOSE_HCA(struct mlx4_dev *dev, int panic)
+{
+	return mlx4_cmd(dev, 0, 0, panic, MLX4_CMD_CLOSE_HCA, 1000);
+}
+
+int mlx4_SET_ICM_SIZE(struct mlx4_dev *dev, u64 icm_size, u64 *aux_pages)
+{
+	int ret = mlx4_cmd_imm(dev, icm_size, aux_pages, 0, 0,
+			       MLX4_CMD_SET_ICM_SIZE,
+			       MLX4_CMD_TIME_CLASS_A);
+	if (ret)
+		return ret;
+
+	/*
+	 * Round up number of system pages needed in case
+	 * MLX4_ICM_PAGE_SIZE < PAGE_SIZE.
+	 */
+	*aux_pages = ALIGN(*aux_pages, PAGE_SIZE / MLX4_ICM_PAGE_SIZE) >>
+		(PAGE_SHIFT - MLX4_ICM_PAGE_SHIFT);
+
+	return 0;
+}
+
+int mlx4_NOP(struct mlx4_dev *dev)
+{
+	/* Input modifier of 0x1f means "finish as soon as possible." */
+	return mlx4_cmd(dev, 0, 0x1f, 0, MLX4_CMD_NOP, 100);
+}

kernel/drivers/net/mlx4/fw.h

@@ -0,0 +1,177 @@
+/*
+ * Copyright (c) 2004, 2005 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2006, 2007 Cisco Systems.  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 MLX4_FW_H
+#define MLX4_FW_H
+
+#include "mlx4.h"
+#include "icm.h"
+
+struct mlx4_mod_stat_cfg {
+	u8 log_pg_sz;
+	u8 log_pg_sz_m;
+};
+
+struct mlx4_dev_cap {
+	int max_srq_sz;
+	int max_qp_sz;
+	int reserved_qps;
+	int max_qps;
+	int reserved_srqs;
+	int max_srqs;
+	int max_cq_sz;
+	int reserved_cqs;
+	int max_cqs;
+	int max_mpts;
+	int reserved_eqs;
+	int max_eqs;
+	int reserved_mtts;
+	int max_mrw_sz;
+	int reserved_mrws;
+	int max_mtt_seg;
+	int max_requester_per_qp;
+	int max_responder_per_qp;
+	int max_rdma_global;
+	int local_ca_ack_delay;
+	int num_ports;
+	u32 max_msg_sz;
+	int ib_mtu[MLX4_MAX_PORTS + 1];
+	int max_port_width[MLX4_MAX_PORTS + 1];
+	int max_vl[MLX4_MAX_PORTS + 1];
+	int max_gids[MLX4_MAX_PORTS + 1];
+	int max_pkeys[MLX4_MAX_PORTS + 1];
+	u64 def_mac[MLX4_MAX_PORTS + 1];
+	u16 eth_mtu[MLX4_MAX_PORTS + 1];
+	u16 stat_rate_support;
+	u32 flags;
+	int reserved_uars;
+	int uar_size;
+	int min_page_sz;
+	int bf_reg_size;
+	int bf_regs_per_page;
+	int max_sq_sg;
+	int max_sq_desc_sz;
+	int max_rq_sg;
+	int max_rq_desc_sz;
+	int max_qp_per_mcg;
+	int reserved_mgms;
+	int max_mcgs;
+	int reserved_pds;
+	int max_pds;
+	int qpc_entry_sz;
+	int rdmarc_entry_sz;
+	int altc_entry_sz;
+	int aux_entry_sz;
+	int srq_entry_sz;
+	int cqc_entry_sz;
+	int eqc_entry_sz;
+	int dmpt_entry_sz;
+	int cmpt_entry_sz;
+	int mtt_entry_sz;
+	int resize_srq;
+	u32 bmme_flags;
+	u32 reserved_lkey;
+	u64 max_icm_sz;
+	int max_gso_sz;
+	u8  supported_port_types[MLX4_MAX_PORTS + 1];
+	u8  log_max_macs[MLX4_MAX_PORTS + 1];
+	u8  log_max_vlans[MLX4_MAX_PORTS + 1];
+};
+
+struct mlx4_adapter {
+	char board_id[MLX4_BOARD_ID_LEN];
+	u8   inta_pin;
+};
+
+struct mlx4_init_hca_param {
+	u64 qpc_base;
+	u64 rdmarc_base;
+	u64 auxc_base;
+	u64 altc_base;
+	u64 srqc_base;
+	u64 cqc_base;
+	u64 eqc_base;
+	u64 mc_base;
+	u64 dmpt_base;
+	u64 cmpt_base;
+	u64 mtt_base;
+	u16 log_mc_entry_sz;
+	u16 log_mc_hash_sz;
+	u8  log_num_qps;
+	u8  log_num_srqs;
+	u8  log_num_cqs;
+	u8  log_num_eqs;
+	u8  log_rd_per_qp;
+	u8  log_mc_table_sz;
+	u8  log_mpt_sz;
+	u8  log_uar_sz;
+};
+
+struct mlx4_init_ib_param {
+	int port_width;
+	int vl_cap;
+	int mtu_cap;
+	u16 gid_cap;
+	u16 pkey_cap;
+	int set_guid0;
+	u64 guid0;
+	int set_node_guid;
+	u64 node_guid;
+	int set_si_guid;
+	u64 si_guid;
+};
+
+struct mlx4_set_ib_param {
+	int set_si_guid;
+	int reset_qkey_viol;
+	u64 si_guid;
+	u32 cap_mask;
+};
+
+int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap);
+int mlx4_MAP_FA(struct mlx4_dev *dev, struct mlx4_icm *icm);
+int mlx4_UNMAP_FA(struct mlx4_dev *dev);
+int mlx4_RUN_FW(struct mlx4_dev *dev);
+int mlx4_QUERY_FW(struct mlx4_dev *dev);
+int mlx4_QUERY_ADAPTER(struct mlx4_dev *dev, struct mlx4_adapter *adapter);
+int mlx4_INIT_HCA(struct mlx4_dev *dev, struct mlx4_init_hca_param *param);
+int mlx4_CLOSE_HCA(struct mlx4_dev *dev, int panic);
+int mlx4_map_cmd(struct mlx4_dev *dev, u16 op, struct mlx4_icm *icm, u64 virt);
+int mlx4_SET_ICM_SIZE(struct mlx4_dev *dev, u64 icm_size, u64 *aux_pages);
+int mlx4_MAP_ICM_AUX(struct mlx4_dev *dev, struct mlx4_icm *icm);
+int mlx4_UNMAP_ICM_AUX(struct mlx4_dev *dev);
+int mlx4_NOP(struct mlx4_dev *dev);
+int mlx4_MOD_STAT_CFG(struct mlx4_dev *dev, struct mlx4_mod_stat_cfg *cfg);
+
+#endif /* MLX4_FW_H */

kernel/drivers/net/mlx4/icm.c

@@ -0,0 +1,454 @@
+/*
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2006, 2007 Cisco Systems, 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/errno.h>
+#include <linux/mm.h>
+#include <linux/scatterlist.h>
+
+#include <linux/mlx4/cmd.h>
+
+#include "mlx4.h"
+#include "icm.h"
+#include "fw.h"
+
+/*
+ * We allocate in as big chunks as we can, up to a maximum of 256 KB
+ * per chunk.
+ */
+enum {
+	MLX4_ICM_ALLOC_SIZE	= 1 << 18,
+	MLX4_TABLE_CHUNK_SIZE	= 1 << 18
+};
+
+static void mlx4_free_icm_pages(struct mlx4_dev *dev, struct mlx4_icm_chunk *chunk)
+{
+	int i;
+
+	if (chunk->nsg > 0)
+		pci_unmap_sg(dev->pdev, chunk->mem, chunk->npages,
+			     PCI_DMA_BIDIRECTIONAL);
+
+	for (i = 0; i < chunk->npages; ++i)
+		__free_pages(sg_page(&chunk->mem[i]),
+			     get_order(chunk->mem[i].length));
+}
+
+static void mlx4_free_icm_coherent(struct mlx4_dev *dev, struct mlx4_icm_chunk *chunk)
+{
+	int i;
+
+	for (i = 0; i < chunk->npages; ++i)
+		dma_free_coherent(&dev->pdev->dev, chunk->mem[i].length,
+				  lowmem_page_address(sg_page(&chunk->mem[i])),
+				  sg_dma_address(&chunk->mem[i]));
+}
+
+void mlx4_free_icm(struct mlx4_dev *dev, struct mlx4_icm *icm, int coherent)
+{
+	struct mlx4_icm_chunk *chunk, *tmp;
+
+	if (!icm)
+		return;
+
+	list_for_each_entry_safe(chunk, tmp, &icm->chunk_list, list) {
+		if (coherent)
+			mlx4_free_icm_coherent(dev, chunk);
+		else
+			mlx4_free_icm_pages(dev, chunk);
+
+		kfree(chunk);
+	}
+
+	kfree(icm);
+}
+
+static int mlx4_alloc_icm_pages(struct scatterlist *mem, int order, gfp_t gfp_mask)
+{
+	struct page *page;
+
+	page = alloc_pages(gfp_mask, order);
+	if (!page)
+		return -ENOMEM;
+
+	sg_set_page(mem, page, PAGE_SIZE << order, 0);
+	return 0;
+}
+
+static int mlx4_alloc_icm_coherent(struct device *dev, struct scatterlist *mem,
+				    int order, gfp_t gfp_mask)
+{
+	void *buf = dma_alloc_coherent(dev, PAGE_SIZE << order,
+				       &sg_dma_address(mem), gfp_mask);
+	if (!buf)
+		return -ENOMEM;
+
+	sg_set_buf(mem, buf, PAGE_SIZE << order);
+	BUG_ON(mem->offset);
+	sg_dma_len(mem) = PAGE_SIZE << order;
+	return 0;
+}
+
+struct mlx4_icm *mlx4_alloc_icm(struct mlx4_dev *dev, int npages,
+				gfp_t gfp_mask, int coherent)
+{
+	struct mlx4_icm *icm;
+	struct mlx4_icm_chunk *chunk = NULL;
+	int cur_order;
+	int ret;
+
+	/* We use sg_set_buf for coherent allocs, which assumes low memory */
+	BUG_ON(coherent && (gfp_mask & __GFP_HIGHMEM));
+
+	icm = kmalloc(sizeof *icm, gfp_mask & ~(__GFP_HIGHMEM | __GFP_NOWARN));
+	if (!icm)
+		return NULL;
+
+	icm->refcount = 0;
+	INIT_LIST_HEAD(&icm->chunk_list);
+
+	cur_order = get_order(MLX4_ICM_ALLOC_SIZE);
+
+	while (npages > 0) {
+		if (!chunk) {
+			chunk = kmalloc(sizeof *chunk,
+					gfp_mask & ~(__GFP_HIGHMEM | __GFP_NOWARN));
+			if (!chunk)
+				goto fail;
+
+			sg_init_table(chunk->mem, MLX4_ICM_CHUNK_LEN);
+			chunk->npages = 0;
+			chunk->nsg    = 0;
+			list_add_tail(&chunk->list, &icm->chunk_list);
+		}
+
+		while (1 << cur_order > npages)
+			--cur_order;
+
+		if (coherent)
+			ret = mlx4_alloc_icm_coherent(&dev->pdev->dev,
+						      &chunk->mem[chunk->npages],
+						      cur_order, gfp_mask);
+		else
+			ret = mlx4_alloc_icm_pages(&chunk->mem[chunk->npages],
+						   cur_order, gfp_mask);
+
+		if (!ret) {
+			++chunk->npages;
+
+			if (coherent)
+				++chunk->nsg;
+			else if (chunk->npages == MLX4_ICM_CHUNK_LEN) {
+				chunk->nsg = pci_map_sg(dev->pdev, chunk->mem,
+							chunk->npages,
+							PCI_DMA_BIDIRECTIONAL);
+
+				if (chunk->nsg <= 0)
+					goto fail;
+			}
+
+			if (chunk->npages == MLX4_ICM_CHUNK_LEN)
+				chunk = NULL;
+
+			npages -= 1 << cur_order;
+		} else {
+			--cur_order;
+			if (cur_order < 0)
+				goto fail;
+		}
+	}
+
+	if (!coherent && chunk) {
+		chunk->nsg = pci_map_sg(dev->pdev, chunk->mem,
+					chunk->npages,
+					PCI_DMA_BIDIRECTIONAL);
+
+		if (chunk->nsg <= 0)
+			goto fail;
+	}
+
+	return icm;
+
+fail:
+	mlx4_free_icm(dev, icm, coherent);
+	return NULL;
+}
+
+static int mlx4_MAP_ICM(struct mlx4_dev *dev, struct mlx4_icm *icm, u64 virt)
+{
+	return mlx4_map_cmd(dev, MLX4_CMD_MAP_ICM, icm, virt);
+}
+
+int mlx4_UNMAP_ICM(struct mlx4_dev *dev, u64 virt, u32 page_count)
+{
+	return mlx4_cmd(dev, virt, page_count, 0, MLX4_CMD_UNMAP_ICM,
+			MLX4_CMD_TIME_CLASS_B);
+}
+
+int mlx4_MAP_ICM_page(struct mlx4_dev *dev, u64 dma_addr, u64 virt)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	__be64 *inbox;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	inbox = mailbox->buf;
+
+	inbox[0] = cpu_to_be64(virt);
+	inbox[1] = cpu_to_be64(dma_addr);
+
+	err = mlx4_cmd(dev, mailbox->dma, 1, 0, MLX4_CMD_MAP_ICM,
+		       MLX4_CMD_TIME_CLASS_B);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+	if (!err)
+		mlx4_dbg(dev, "Mapped page at %llx to %llx for ICM.\n",
+			  (unsigned long long) dma_addr, (unsigned long long) virt);
+
+	return err;
+}
+
+int mlx4_MAP_ICM_AUX(struct mlx4_dev *dev, struct mlx4_icm *icm)
+{
+	return mlx4_map_cmd(dev, MLX4_CMD_MAP_ICM_AUX, icm, -1);
+}
+
+int mlx4_UNMAP_ICM_AUX(struct mlx4_dev *dev)
+{
+	return mlx4_cmd(dev, 0, 0, 0, MLX4_CMD_UNMAP_ICM_AUX, MLX4_CMD_TIME_CLASS_B);
+}
+
+int mlx4_table_get(struct mlx4_dev *dev, struct mlx4_icm_table *table, int obj)
+{
+	int i = (obj & (table->num_obj - 1)) / (MLX4_TABLE_CHUNK_SIZE / table->obj_size);
+	int ret = 0;
+
+	mutex_lock(&table->mutex);
+
+	if (table->icm[i]) {
+		++table->icm[i]->refcount;
+		goto out;
+	}
+
+	table->icm[i] = mlx4_alloc_icm(dev, MLX4_TABLE_CHUNK_SIZE >> PAGE_SHIFT,
+				       (table->lowmem ? GFP_KERNEL : GFP_HIGHUSER) |
+				       __GFP_NOWARN, table->coherent);
+	if (!table->icm[i]) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	if (mlx4_MAP_ICM(dev, table->icm[i], table->virt +
+			 (u64) i * MLX4_TABLE_CHUNK_SIZE)) {
+		mlx4_free_icm(dev, table->icm[i], table->coherent);
+		table->icm[i] = NULL;
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	++table->icm[i]->refcount;
+
+out:
+	mutex_unlock(&table->mutex);
+	return ret;
+}
+
+void mlx4_table_put(struct mlx4_dev *dev, struct mlx4_icm_table *table, int obj)
+{
+	int i;
+
+	i = (obj & (table->num_obj - 1)) / (MLX4_TABLE_CHUNK_SIZE / table->obj_size);
+
+	mutex_lock(&table->mutex);
+
+	if (--table->icm[i]->refcount == 0) {
+		mlx4_UNMAP_ICM(dev, table->virt + i * MLX4_TABLE_CHUNK_SIZE,
+			       MLX4_TABLE_CHUNK_SIZE / MLX4_ICM_PAGE_SIZE);
+		mlx4_free_icm(dev, table->icm[i], table->coherent);
+		table->icm[i] = NULL;
+	}
+
+	mutex_unlock(&table->mutex);
+}
+
+void *mlx4_table_find(struct mlx4_icm_table *table, int obj, dma_addr_t *dma_handle)
+{
+	int idx, offset, dma_offset, i;
+	struct mlx4_icm_chunk *chunk;
+	struct mlx4_icm *icm;
+	struct page *page = NULL;
+
+	if (!table->lowmem)
+		return NULL;
+
+	mutex_lock(&table->mutex);
+
+	idx = (obj & (table->num_obj - 1)) * table->obj_size;
+	icm = table->icm[idx / MLX4_TABLE_CHUNK_SIZE];
+	dma_offset = offset = idx % MLX4_TABLE_CHUNK_SIZE;
+
+	if (!icm)
+		goto out;
+
+	list_for_each_entry(chunk, &icm->chunk_list, list) {
+		for (i = 0; i < chunk->npages; ++i) {
+			if (dma_handle && dma_offset >= 0) {
+				if (sg_dma_len(&chunk->mem[i]) > dma_offset)
+					*dma_handle = sg_dma_address(&chunk->mem[i]) +
+						dma_offset;
+				dma_offset -= sg_dma_len(&chunk->mem[i]);
+			}
+			/*
+			 * DMA mapping can merge pages but not split them,
+			 * so if we found the page, dma_handle has already
+			 * been assigned to.
+			 */
+			if (chunk->mem[i].length > offset) {
+				page = sg_page(&chunk->mem[i]);
+				goto out;
+			}
+			offset -= chunk->mem[i].length;
+		}
+	}
+
+out:
+	mutex_unlock(&table->mutex);
+	return page ? lowmem_page_address(page) + offset : NULL;
+}
+
+int mlx4_table_get_range(struct mlx4_dev *dev, struct mlx4_icm_table *table,
+			 int start, int end)
+{
+	int inc = MLX4_TABLE_CHUNK_SIZE / table->obj_size;
+	int i, err;
+
+	for (i = start; i <= end; i += inc) {
+		err = mlx4_table_get(dev, table, i);
+		if (err)
+			goto fail;
+	}
+
+	return 0;
+
+fail:
+	while (i > start) {
+		i -= inc;
+		mlx4_table_put(dev, table, i);
+	}
+
+	return err;
+}
+
+void mlx4_table_put_range(struct mlx4_dev *dev, struct mlx4_icm_table *table,
+			  int start, int end)
+{
+	int i;
+
+	for (i = start; i <= end; i += MLX4_TABLE_CHUNK_SIZE / table->obj_size)
+		mlx4_table_put(dev, table, i);
+}
+
+int mlx4_init_icm_table(struct mlx4_dev *dev, struct mlx4_icm_table *table,
+			u64 virt, int obj_size,	int nobj, int reserved,
+			int use_lowmem, int use_coherent)
+{
+	int obj_per_chunk;
+	int num_icm;
+	unsigned chunk_size;
+	int i;
+
+	obj_per_chunk = MLX4_TABLE_CHUNK_SIZE / obj_size;
+	num_icm = (nobj + obj_per_chunk - 1) / obj_per_chunk;
+
+	table->icm      = kcalloc(num_icm, sizeof *table->icm, GFP_KERNEL);
+	if (!table->icm)
+		return -ENOMEM;
+	table->virt     = virt;
+	table->num_icm  = num_icm;
+	table->num_obj  = nobj;
+	table->obj_size = obj_size;
+	table->lowmem   = use_lowmem;
+	table->coherent = use_coherent;
+	mutex_init(&table->mutex);
+
+	for (i = 0; i * MLX4_TABLE_CHUNK_SIZE < reserved * obj_size; ++i) {
+		chunk_size = MLX4_TABLE_CHUNK_SIZE;
+		if ((i + 1) * MLX4_TABLE_CHUNK_SIZE > nobj * obj_size)
+			chunk_size = PAGE_ALIGN(nobj * obj_size - i * MLX4_TABLE_CHUNK_SIZE);
+
+		table->icm[i] = mlx4_alloc_icm(dev, chunk_size >> PAGE_SHIFT,
+					       (use_lowmem ? GFP_KERNEL : GFP_HIGHUSER) |
+					       __GFP_NOWARN, use_coherent);
+		if (!table->icm[i])
+			goto err;
+		if (mlx4_MAP_ICM(dev, table->icm[i], virt + i * MLX4_TABLE_CHUNK_SIZE)) {
+			mlx4_free_icm(dev, table->icm[i], use_coherent);
+			table->icm[i] = NULL;
+			goto err;
+		}
+
+		/*
+		 * Add a reference to this ICM chunk so that it never
+		 * gets freed (since it contains reserved firmware objects).
+		 */
+		++table->icm[i]->refcount;
+	}
+
+	return 0;
+
+err:
+	for (i = 0; i < num_icm; ++i)
+		if (table->icm[i]) {
+			mlx4_UNMAP_ICM(dev, virt + i * MLX4_TABLE_CHUNK_SIZE,
+				       MLX4_TABLE_CHUNK_SIZE / MLX4_ICM_PAGE_SIZE);
+			mlx4_free_icm(dev, table->icm[i], use_coherent);
+		}
+
+	return -ENOMEM;
+}
+
+void mlx4_cleanup_icm_table(struct mlx4_dev *dev, struct mlx4_icm_table *table)
+{
+	int i;
+
+	for (i = 0; i < table->num_icm; ++i)
+		if (table->icm[i]) {
+			mlx4_UNMAP_ICM(dev, table->virt + i * MLX4_TABLE_CHUNK_SIZE,
+				       MLX4_TABLE_CHUNK_SIZE / MLX4_ICM_PAGE_SIZE);
+			mlx4_free_icm(dev, table->icm[i], table->coherent);
+		}
+
+	kfree(table->icm);
+}

kernel/drivers/net/mlx4/icm.h

@@ -0,0 +1,136 @@
+/*
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2006, 2007 Cisco Systems, 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 MLX4_ICM_H
+#define MLX4_ICM_H
+
+#include <linux/list.h>
+#include <linux/pci.h>
+#include <linux/mutex.h>
+
+#define MLX4_ICM_CHUNK_LEN						\
+	((256 - sizeof (struct list_head) - 2 * sizeof (int)) /		\
+	 (sizeof (struct scatterlist)))
+
+enum {
+	MLX4_ICM_PAGE_SHIFT	= 12,
+	MLX4_ICM_PAGE_SIZE	= 1 << MLX4_ICM_PAGE_SHIFT,
+};
+
+struct mlx4_icm_chunk {
+	struct list_head	list;
+	int			npages;
+	int			nsg;
+	struct scatterlist	mem[MLX4_ICM_CHUNK_LEN];
+};
+
+struct mlx4_icm {
+	struct list_head	chunk_list;
+	int			refcount;
+};
+
+struct mlx4_icm_iter {
+	struct mlx4_icm	       *icm;
+	struct mlx4_icm_chunk  *chunk;
+	int			page_idx;
+};
+
+struct mlx4_dev;
+
+struct mlx4_icm *mlx4_alloc_icm(struct mlx4_dev *dev, int npages,
+				gfp_t gfp_mask, int coherent);
+void mlx4_free_icm(struct mlx4_dev *dev, struct mlx4_icm *icm, int coherent);
+
+int mlx4_table_get(struct mlx4_dev *dev, struct mlx4_icm_table *table, int obj);
+void mlx4_table_put(struct mlx4_dev *dev, struct mlx4_icm_table *table, int obj);
+int mlx4_table_get_range(struct mlx4_dev *dev, struct mlx4_icm_table *table,
+			 int start, int end);
+void mlx4_table_put_range(struct mlx4_dev *dev, struct mlx4_icm_table *table,
+			  int start, int end);
+int mlx4_init_icm_table(struct mlx4_dev *dev, struct mlx4_icm_table *table,
+			u64 virt, int obj_size,	int nobj, int reserved,
+			int use_lowmem, int use_coherent);
+void mlx4_cleanup_icm_table(struct mlx4_dev *dev, struct mlx4_icm_table *table);
+int mlx4_table_get(struct mlx4_dev *dev, struct mlx4_icm_table *table, int obj);
+void mlx4_table_put(struct mlx4_dev *dev, struct mlx4_icm_table *table, int obj);
+void *mlx4_table_find(struct mlx4_icm_table *table, int obj, dma_addr_t *dma_handle);
+int mlx4_table_get_range(struct mlx4_dev *dev, struct mlx4_icm_table *table,
+			 int start, int end);
+void mlx4_table_put_range(struct mlx4_dev *dev, struct mlx4_icm_table *table,
+			  int start, int end);
+
+static inline void mlx4_icm_first(struct mlx4_icm *icm,
+				  struct mlx4_icm_iter *iter)
+{
+	iter->icm      = icm;
+	iter->chunk    = list_empty(&icm->chunk_list) ?
+		NULL : list_entry(icm->chunk_list.next,
+				  struct mlx4_icm_chunk, list);
+	iter->page_idx = 0;
+}
+
+static inline int mlx4_icm_last(struct mlx4_icm_iter *iter)
+{
+	return !iter->chunk;
+}
+
+static inline void mlx4_icm_next(struct mlx4_icm_iter *iter)
+{
+	if (++iter->page_idx >= iter->chunk->nsg) {
+		if (iter->chunk->list.next == &iter->icm->chunk_list) {
+			iter->chunk = NULL;
+			return;
+		}
+
+		iter->chunk = list_entry(iter->chunk->list.next,
+					 struct mlx4_icm_chunk, list);
+		iter->page_idx = 0;
+	}
+}
+
+static inline dma_addr_t mlx4_icm_addr(struct mlx4_icm_iter *iter)
+{
+	return sg_dma_address(&iter->chunk->mem[iter->page_idx]);
+}
+
+static inline unsigned long mlx4_icm_size(struct mlx4_icm_iter *iter)
+{
+	return sg_dma_len(&iter->chunk->mem[iter->page_idx]);
+}
+
+int mlx4_UNMAP_ICM(struct mlx4_dev *dev, u64 virt, u32 page_count);
+int mlx4_MAP_ICM_page(struct mlx4_dev *dev, u64 dma_addr, u64 virt);
+int mlx4_MAP_ICM_AUX(struct mlx4_dev *dev, struct mlx4_icm *icm);
+int mlx4_UNMAP_ICM_AUX(struct mlx4_dev *dev);
+
+#endif /* MLX4_ICM_H */

kernel/drivers/net/mlx4/intf.c

@@ -0,0 +1,161 @@
+/*
+ * Copyright (c) 2006, 2007 Cisco Systems, Inc. All rights reserved.
+ * Copyright (c) 2007, 2008 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
+ * 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 "mlx4.h"
+
+struct mlx4_device_context {
+	struct list_head	list;
+	struct mlx4_interface  *intf;
+	void		       *context;
+};
+
+static LIST_HEAD(intf_list);
+static LIST_HEAD(dev_list);
+static DEFINE_MUTEX(intf_mutex);
+
+static void mlx4_add_device(struct mlx4_interface *intf, struct mlx4_priv *priv)
+{
+	struct mlx4_device_context *dev_ctx;
+
+	dev_ctx = kmalloc(sizeof *dev_ctx, GFP_KERNEL);
+	if (!dev_ctx)
+		return;
+
+	dev_ctx->intf    = intf;
+	dev_ctx->context = intf->add(&priv->dev);
+
+	if (dev_ctx->context) {
+		spin_lock_irq(&priv->ctx_lock);
+		list_add_tail(&dev_ctx->list, &priv->ctx_list);
+		spin_unlock_irq(&priv->ctx_lock);
+	} else
+		kfree(dev_ctx);
+}
+
+static void mlx4_remove_device(struct mlx4_interface *intf, struct mlx4_priv *priv)
+{
+	struct mlx4_device_context *dev_ctx;
+
+	list_for_each_entry(dev_ctx, &priv->ctx_list, list)
+		if (dev_ctx->intf == intf) {
+			spin_lock_irq(&priv->ctx_lock);
+			list_del(&dev_ctx->list);
+			spin_unlock_irq(&priv->ctx_lock);
+
+			intf->remove(&priv->dev, dev_ctx->context);
+			kfree(dev_ctx);
+			return;
+		}
+}
+
+int mlx4_register_interface(struct mlx4_interface *intf)
+{
+	struct mlx4_priv *priv;
+
+	if (!intf->add || !intf->remove)
+		return -EINVAL;
+
+	mutex_lock(&intf_mutex);
+
+	list_add_tail(&intf->list, &intf_list);
+	list_for_each_entry(priv, &dev_list, dev_list)
+		mlx4_add_device(intf, priv);
+
+	mutex_unlock(&intf_mutex);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_register_interface);
+
+void mlx4_unregister_interface(struct mlx4_interface *intf)
+{
+	struct mlx4_priv *priv;
+
+	mutex_lock(&intf_mutex);
+
+	list_for_each_entry(priv, &dev_list, dev_list)
+		mlx4_remove_device(intf, priv);
+
+	list_del(&intf->list);
+
+	mutex_unlock(&intf_mutex);
+}
+EXPORT_SYMBOL_GPL(mlx4_unregister_interface);
+
+void mlx4_dispatch_event(struct mlx4_dev *dev, enum mlx4_dev_event type, int port)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_device_context *dev_ctx;
+	unsigned long flags;
+
+	spin_lock_irqsave(&priv->ctx_lock, flags);
+
+	list_for_each_entry(dev_ctx, &priv->ctx_list, list)
+		if (dev_ctx->intf->event)
+			dev_ctx->intf->event(dev, dev_ctx->context, type, port);
+
+	spin_unlock_irqrestore(&priv->ctx_lock, flags);
+}
+
+int mlx4_register_device(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_interface *intf;
+
+	mutex_lock(&intf_mutex);
+
+	list_add_tail(&priv->dev_list, &dev_list);
+	list_for_each_entry(intf, &intf_list, list)
+		mlx4_add_device(intf, priv);
+
+	mutex_unlock(&intf_mutex);
+	mlx4_start_catas_poll(dev);
+
+	return 0;
+}
+
+void mlx4_unregister_device(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_interface *intf;
+
+	mlx4_stop_catas_poll(dev);
+	mutex_lock(&intf_mutex);
+
+	list_for_each_entry(intf, &intf_list, list)
+		mlx4_remove_device(intf, priv);
+
+	list_del(&priv->dev_list);
+
+	mutex_unlock(&intf_mutex);
+}

kernel/drivers/net/mlx4/mcg.c

@@ -0,0 +1,364 @@
+/*
+ * Copyright (c) 2006, 2007 Cisco Systems, Inc.  All rights reserved.
+ * Copyright (c) 2007, 2008 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
+ * 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/string.h>
+#include <linux/slab.h>
+
+#include <linux/mlx4/cmd.h>
+
+#include "mlx4.h"
+
+#define MGM_QPN_MASK       0x00FFFFFF
+#define MGM_BLCK_LB_BIT    30
+
+struct mlx4_mgm {
+	__be32			next_gid_index;
+	__be32			members_count;
+	u32			reserved[2];
+	u8			gid[16];
+	__be32			qp[MLX4_QP_PER_MGM];
+};
+
+static const u8 zero_gid[16];	/* automatically initialized to 0 */
+
+static int mlx4_READ_MCG(struct mlx4_dev *dev, int index,
+			 struct mlx4_cmd_mailbox *mailbox)
+{
+	return mlx4_cmd_box(dev, 0, mailbox->dma, index, 0, MLX4_CMD_READ_MCG,
+			    MLX4_CMD_TIME_CLASS_A);
+}
+
+static int mlx4_WRITE_MCG(struct mlx4_dev *dev, int index,
+			  struct mlx4_cmd_mailbox *mailbox)
+{
+	return mlx4_cmd(dev, mailbox->dma, index, 0, MLX4_CMD_WRITE_MCG,
+			MLX4_CMD_TIME_CLASS_A);
+}
+
+static int mlx4_MGID_HASH(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *mailbox,
+			  u16 *hash)
+{
+	u64 imm;
+	int err;
+
+	err = mlx4_cmd_imm(dev, mailbox->dma, &imm, 0, 0, MLX4_CMD_MGID_HASH,
+			   MLX4_CMD_TIME_CLASS_A);
+
+	if (!err)
+		*hash = imm;
+
+	return err;
+}
+
+/*
+ * Caller must hold MCG table semaphore.  gid and mgm parameters must
+ * be properly aligned for command interface.
+ *
+ *  Returns 0 unless a firmware command error occurs.
+ *
+ * If GID is found in MGM or MGM is empty, *index = *hash, *prev = -1
+ * and *mgm holds MGM entry.
+ *
+ * if GID is found in AMGM, *index = index in AMGM, *prev = index of
+ * previous entry in hash chain and *mgm holds AMGM entry.
+ *
+ * If no AMGM exists for given gid, *index = -1, *prev = index of last
+ * entry in hash chain and *mgm holds end of hash chain.
+ */
+static int find_mgm(struct mlx4_dev *dev,
+		    u8 *gid, struct mlx4_cmd_mailbox *mgm_mailbox,
+		    u16 *hash, int *prev, int *index)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_mgm *mgm = mgm_mailbox->buf;
+	u8 *mgid;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return -ENOMEM;
+	mgid = mailbox->buf;
+
+	memcpy(mgid, gid, 16);
+
+	err = mlx4_MGID_HASH(dev, mailbox, hash);
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	if (err)
+		return err;
+
+	if (0)
+		mlx4_dbg(dev, "Hash for %pI6 is %04x\n", gid, *hash);
+
+	*index = *hash;
+	*prev  = -1;
+
+	do {
+		err = mlx4_READ_MCG(dev, *index, mgm_mailbox);
+		if (err)
+			return err;
+
+		if (!memcmp(mgm->gid, zero_gid, 16)) {
+			if (*index != *hash) {
+				mlx4_err(dev, "Found zero MGID in AMGM.\n");
+				err = -EINVAL;
+			}
+			return err;
+		}
+
+		if (!memcmp(mgm->gid, gid, 16))
+			return err;
+
+		*prev = *index;
+		*index = be32_to_cpu(mgm->next_gid_index) >> 6;
+	} while (*index);
+
+	*index = -1;
+	return err;
+}
+
+int mlx4_multicast_attach(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16],
+			  int block_mcast_loopback)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_mgm *mgm;
+	u32 members_count;
+	u16 hash;
+	int index, prev;
+	int link = 0;
+	int i;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	mgm = mailbox->buf;
+
+	mutex_lock(&priv->mcg_table.mutex);
+
+	err = find_mgm(dev, gid, mailbox, &hash, &prev, &index);
+	if (err)
+		goto out;
+
+	if (index != -1) {
+		if (!memcmp(mgm->gid, zero_gid, 16))
+			memcpy(mgm->gid, gid, 16);
+	} else {
+		link = 1;
+
+		index = mlx4_bitmap_alloc(&priv->mcg_table.bitmap);
+		if (index == -1) {
+			mlx4_err(dev, "No AMGM entries left\n");
+			err = -ENOMEM;
+			goto out;
+		}
+		index += dev->caps.num_mgms;
+
+		memset(mgm, 0, sizeof *mgm);
+		memcpy(mgm->gid, gid, 16);
+	}
+
+	members_count = be32_to_cpu(mgm->members_count);
+	if (members_count == MLX4_QP_PER_MGM) {
+		mlx4_err(dev, "MGM at index %x is full.\n", index);
+		err = -ENOMEM;
+		goto out;
+	}
+
+	for (i = 0; i < members_count; ++i)
+		if ((be32_to_cpu(mgm->qp[i]) & MGM_QPN_MASK) == qp->qpn) {
+			mlx4_dbg(dev, "QP %06x already a member of MGM\n", qp->qpn);
+			err = 0;
+			goto out;
+		}
+
+	if (block_mcast_loopback)
+		mgm->qp[members_count++] = cpu_to_be32((qp->qpn & MGM_QPN_MASK) |
+						       (1U << MGM_BLCK_LB_BIT));
+	else
+		mgm->qp[members_count++] = cpu_to_be32(qp->qpn & MGM_QPN_MASK);
+
+	mgm->members_count       = cpu_to_be32(members_count);
+
+	err = mlx4_WRITE_MCG(dev, index, mailbox);
+	if (err)
+		goto out;
+
+	if (!link)
+		goto out;
+
+	err = mlx4_READ_MCG(dev, prev, mailbox);
+	if (err)
+		goto out;
+
+	mgm->next_gid_index = cpu_to_be32(index << 6);
+
+	err = mlx4_WRITE_MCG(dev, prev, mailbox);
+	if (err)
+		goto out;
+
+out:
+	if (err && link && index != -1) {
+		if (index < dev->caps.num_mgms)
+			mlx4_warn(dev, "Got AMGM index %d < %d",
+				  index, dev->caps.num_mgms);
+		else
+			mlx4_bitmap_free(&priv->mcg_table.bitmap,
+					 index - dev->caps.num_mgms);
+	}
+	mutex_unlock(&priv->mcg_table.mutex);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_multicast_attach);
+
+int mlx4_multicast_detach(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16])
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_mgm *mgm;
+	u32 members_count;
+	u16 hash;
+	int prev, index;
+	int i, loc;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	mgm = mailbox->buf;
+
+	mutex_lock(&priv->mcg_table.mutex);
+
+	err = find_mgm(dev, gid, mailbox, &hash, &prev, &index);
+	if (err)
+		goto out;
+
+	if (index == -1) {
+		mlx4_err(dev, "MGID %pI6 not found\n", gid);
+		err = -EINVAL;
+		goto out;
+	}
+
+	members_count = be32_to_cpu(mgm->members_count);
+	for (loc = -1, i = 0; i < members_count; ++i)
+		if ((be32_to_cpu(mgm->qp[i]) & MGM_QPN_MASK) == qp->qpn)
+			loc = i;
+
+	if (loc == -1) {
+		mlx4_err(dev, "QP %06x not found in MGM\n", qp->qpn);
+		err = -EINVAL;
+		goto out;
+	}
+
+
+	mgm->members_count = cpu_to_be32(--members_count);
+	mgm->qp[loc]       = mgm->qp[i - 1];
+	mgm->qp[i - 1]     = 0;
+
+	if (i != 1) {
+		err = mlx4_WRITE_MCG(dev, index, mailbox);
+		goto out;
+	}
+
+	if (prev == -1) {
+		/* Remove entry from MGM */
+		int amgm_index = be32_to_cpu(mgm->next_gid_index) >> 6;
+		if (amgm_index) {
+			err = mlx4_READ_MCG(dev, amgm_index, mailbox);
+			if (err)
+				goto out;
+		} else
+			memset(mgm->gid, 0, 16);
+
+		err = mlx4_WRITE_MCG(dev, index, mailbox);
+		if (err)
+			goto out;
+
+		if (amgm_index) {
+			if (amgm_index < dev->caps.num_mgms)
+				mlx4_warn(dev, "MGM entry %d had AMGM index %d < %d",
+					  index, amgm_index, dev->caps.num_mgms);
+			else
+				mlx4_bitmap_free(&priv->mcg_table.bitmap,
+						 amgm_index - dev->caps.num_mgms);
+		}
+	} else {
+		/* Remove entry from AMGM */
+		int cur_next_index = be32_to_cpu(mgm->next_gid_index) >> 6;
+		err = mlx4_READ_MCG(dev, prev, mailbox);
+		if (err)
+			goto out;
+
+		mgm->next_gid_index = cpu_to_be32(cur_next_index << 6);
+
+		err = mlx4_WRITE_MCG(dev, prev, mailbox);
+		if (err)
+			goto out;
+
+		if (index < dev->caps.num_mgms)
+			mlx4_warn(dev, "entry %d had next AMGM index %d < %d",
+				  prev, index, dev->caps.num_mgms);
+		else
+			mlx4_bitmap_free(&priv->mcg_table.bitmap,
+					 index - dev->caps.num_mgms);
+	}
+
+out:
+	mutex_unlock(&priv->mcg_table.mutex);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_multicast_detach);
+
+int mlx4_init_mcg_table(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int err;
+
+	err = mlx4_bitmap_init(&priv->mcg_table.bitmap, dev->caps.num_amgms,
+			       dev->caps.num_amgms - 1, 0, 0);
+	if (err)
+		return err;
+
+	mutex_init(&priv->mcg_table.mutex);
+
+	return 0;
+}
+
+void mlx4_cleanup_mcg_table(struct mlx4_dev *dev)
+{
+	mlx4_bitmap_cleanup(&mlx4_priv(dev)->mcg_table.bitmap);
+}

kernel/drivers/net/mlx4/mlx4.h

@@ -0,0 +1,406 @@
+/*
+ * Copyright (c) 2004, 2005 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved.
+ * Copyright (c) 2005, 2006, 2007 Cisco Systems.  All rights reserved.
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2004 Voltaire, 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 MLX4_H
+#define MLX4_H
+
+#include <linux/mutex.h>
+#include <linux/radix-tree.h>
+#include <linux/timer.h>
+#include <linux/workqueue.h>
+
+#include <linux/mlx4/device.h>
+#include <linux/mlx4/driver.h>
+#include <linux/mlx4/doorbell.h>
+
+#define DRV_NAME	"mlx4_core"
+#define PFX		DRV_NAME ": "
+#define DRV_VERSION	"0.01"
+#define DRV_RELDATE	"May 1, 2007"
+
+enum {
+	MLX4_HCR_BASE		= 0x80680,
+	MLX4_HCR_SIZE		= 0x0001c,
+	MLX4_CLR_INT_SIZE	= 0x00008
+};
+
+enum {
+	MLX4_MGM_ENTRY_SIZE	=  0x100,
+	MLX4_QP_PER_MGM		= 4 * (MLX4_MGM_ENTRY_SIZE / 16 - 2),
+	MLX4_MTT_ENTRY_PER_SEG	= 8
+};
+
+enum {
+	MLX4_NUM_PDS		= 1 << 15
+};
+
+enum {
+	MLX4_CMPT_TYPE_QP	= 0,
+	MLX4_CMPT_TYPE_SRQ	= 1,
+	MLX4_CMPT_TYPE_CQ	= 2,
+	MLX4_CMPT_TYPE_EQ	= 3,
+	MLX4_CMPT_NUM_TYPE
+};
+
+enum {
+	MLX4_CMPT_SHIFT		= 24,
+	MLX4_NUM_CMPTS		= MLX4_CMPT_NUM_TYPE << MLX4_CMPT_SHIFT
+};
+
+#ifdef CONFIG_MLX4_DEBUG
+extern int mlx4_debug_level;
+#else /* CONFIG_MLX4_DEBUG */
+#define mlx4_debug_level	(0)
+#endif /* CONFIG_MLX4_DEBUG */
+
+#define mlx4_dbg(mdev, format, arg...)					\
+	do {								\
+		if (mlx4_debug_level)					\
+			dev_printk(KERN_DEBUG, &mdev->pdev->dev, format, ## arg); \
+	} while (0)
+
+#define mlx4_err(mdev, format, arg...) \
+	dev_err(&mdev->pdev->dev, format, ## arg)
+#define mlx4_info(mdev, format, arg...) \
+	dev_info(&mdev->pdev->dev, format, ## arg)
+#define mlx4_warn(mdev, format, arg...) \
+	dev_warn(&mdev->pdev->dev, format, ## arg)
+
+struct mlx4_bitmap {
+	u32			last;
+	u32			top;
+	u32			max;
+	u32                     reserved_top;
+	u32			mask;
+	spinlock_t		lock;
+	unsigned long	       *table;
+};
+
+struct mlx4_buddy {
+	unsigned long	      **bits;
+	unsigned int	       *num_free;
+	int			max_order;
+	spinlock_t		lock;
+};
+
+struct mlx4_icm;
+
+struct mlx4_icm_table {
+	u64			virt;
+	int			num_icm;
+	int			num_obj;
+	int			obj_size;
+	int			lowmem;
+	int			coherent;
+	struct mutex		mutex;
+	struct mlx4_icm	      **icm;
+};
+
+struct mlx4_eq {
+	struct mlx4_dev	       *dev;
+	void __iomem	       *doorbell;
+	int			eqn;
+	u32			cons_index;
+	u16			irq;
+	u16			have_irq;
+	int			nent;
+	struct mlx4_buf_list   *page_list;
+	struct mlx4_mtt		mtt;
+};
+
+struct mlx4_profile {
+	int			num_qp;
+	int			rdmarc_per_qp;
+	int			num_srq;
+	int			num_cq;
+	int			num_mcg;
+	int			num_mpt;
+	int			num_mtt;
+};
+
+struct mlx4_fw {
+	u64			clr_int_base;
+	u64			catas_offset;
+	struct mlx4_icm	       *fw_icm;
+	struct mlx4_icm	       *aux_icm;
+	u32			catas_size;
+	u16			fw_pages;
+	u8			clr_int_bar;
+	u8			catas_bar;
+};
+
+struct mlx4_cmd {
+	struct pci_pool	       *pool;
+	void __iomem	       *hcr;
+	struct mutex		hcr_mutex;
+	struct semaphore	poll_sem;
+	struct semaphore	event_sem;
+	int			max_cmds;
+	spinlock_t		context_lock;
+	int			free_head;
+	struct mlx4_cmd_context *context;
+	u16			token_mask;
+	u8			use_events;
+	u8			toggle;
+};
+
+struct mlx4_uar_table {
+	struct mlx4_bitmap	bitmap;
+};
+
+struct mlx4_mr_table {
+	struct mlx4_bitmap	mpt_bitmap;
+	struct mlx4_buddy	mtt_buddy;
+	u64			mtt_base;
+	u64			mpt_base;
+	struct mlx4_icm_table	mtt_table;
+	struct mlx4_icm_table	dmpt_table;
+};
+
+struct mlx4_cq_table {
+	struct mlx4_bitmap	bitmap;
+	spinlock_t		lock;
+	struct radix_tree_root	tree;
+	struct mlx4_icm_table	table;
+	struct mlx4_icm_table	cmpt_table;
+};
+
+struct mlx4_eq_table {
+	struct mlx4_bitmap	bitmap;
+	char		       *irq_names;
+	void __iomem	       *clr_int;
+	void __iomem	      **uar_map;
+	u32			clr_mask;
+	struct mlx4_eq	       *eq;
+	struct mlx4_icm_table	table;
+	struct mlx4_icm_table	cmpt_table;
+	int			have_irq;
+	u8			inta_pin;
+};
+
+struct mlx4_srq_table {
+	struct mlx4_bitmap	bitmap;
+	spinlock_t		lock;
+	struct radix_tree_root	tree;
+	struct mlx4_icm_table	table;
+	struct mlx4_icm_table	cmpt_table;
+};
+
+struct mlx4_qp_table {
+	struct mlx4_bitmap	bitmap;
+	u32			rdmarc_base;
+	int			rdmarc_shift;
+	spinlock_t		lock;
+	struct mlx4_icm_table	qp_table;
+	struct mlx4_icm_table	auxc_table;
+	struct mlx4_icm_table	altc_table;
+	struct mlx4_icm_table	rdmarc_table;
+	struct mlx4_icm_table	cmpt_table;
+};
+
+struct mlx4_mcg_table {
+	struct mutex		mutex;
+	struct mlx4_bitmap	bitmap;
+	struct mlx4_icm_table	table;
+};
+
+struct mlx4_catas_err {
+	u32 __iomem	       *map;
+	struct timer_list	timer;
+	struct list_head	list;
+};
+
+#define MLX4_MAX_MAC_NUM	128
+#define MLX4_MAC_TABLE_SIZE	(MLX4_MAX_MAC_NUM << 3)
+
+struct mlx4_mac_table {
+	__be64			entries[MLX4_MAX_MAC_NUM];
+	int			refs[MLX4_MAX_MAC_NUM];
+	struct mutex		mutex;
+	int			total;
+	int			max;
+};
+
+#define MLX4_MAX_VLAN_NUM	128
+#define MLX4_VLAN_TABLE_SIZE	(MLX4_MAX_VLAN_NUM << 2)
+
+struct mlx4_vlan_table {
+	__be32			entries[MLX4_MAX_VLAN_NUM];
+	int			refs[MLX4_MAX_VLAN_NUM];
+	struct mutex		mutex;
+	int			total;
+	int			max;
+};
+
+struct mlx4_port_info {
+	struct mlx4_dev	       *dev;
+	int			port;
+	char			dev_name[16];
+	struct device_attribute port_attr;
+	enum mlx4_port_type	tmp_type;
+	struct mlx4_mac_table	mac_table;
+	struct mlx4_vlan_table	vlan_table;
+};
+
+struct mlx4_sense {
+	struct mlx4_dev		*dev;
+	u8			do_sense_port[MLX4_MAX_PORTS + 1];
+	u8			sense_allowed[MLX4_MAX_PORTS + 1];
+	struct delayed_work	sense_poll;
+};
+
+struct mlx4_priv {
+	struct mlx4_dev		dev;
+
+	struct list_head	dev_list;
+	struct list_head	ctx_list;
+	spinlock_t		ctx_lock;
+
+	struct list_head        pgdir_list;
+	struct mutex            pgdir_mutex;
+
+	struct mlx4_fw		fw;
+	struct mlx4_cmd		cmd;
+
+	struct mlx4_bitmap	pd_bitmap;
+	struct mlx4_uar_table	uar_table;
+	struct mlx4_mr_table	mr_table;
+	struct mlx4_cq_table	cq_table;
+	struct mlx4_eq_table	eq_table;
+	struct mlx4_srq_table	srq_table;
+	struct mlx4_qp_table	qp_table;
+	struct mlx4_mcg_table	mcg_table;
+
+	struct mlx4_catas_err	catas_err;
+
+	void __iomem	       *clr_base;
+
+	struct mlx4_uar		driver_uar;
+	void __iomem	       *kar;
+	struct mlx4_port_info	port[MLX4_MAX_PORTS + 1];
+	struct mlx4_sense       sense;
+	struct mutex		port_mutex;
+};
+
+static inline struct mlx4_priv *mlx4_priv(struct mlx4_dev *dev)
+{
+	return container_of(dev, struct mlx4_priv, dev);
+}
+
+#define MLX4_SENSE_RANGE	(HZ * 3)
+
+extern struct workqueue_struct *mlx4_wq;
+
+u32 mlx4_bitmap_alloc(struct mlx4_bitmap *bitmap);
+void mlx4_bitmap_free(struct mlx4_bitmap *bitmap, u32 obj);
+u32 mlx4_bitmap_alloc_range(struct mlx4_bitmap *bitmap, int cnt, int align);
+void mlx4_bitmap_free_range(struct mlx4_bitmap *bitmap, u32 obj, int cnt);
+int mlx4_bitmap_init(struct mlx4_bitmap *bitmap, u32 num, u32 mask,
+		     u32 reserved_bot, u32 resetrved_top);
+void mlx4_bitmap_cleanup(struct mlx4_bitmap *bitmap);
+
+int mlx4_reset(struct mlx4_dev *dev);
+
+int mlx4_alloc_eq_table(struct mlx4_dev *dev);
+void mlx4_free_eq_table(struct mlx4_dev *dev);
+
+int mlx4_init_pd_table(struct mlx4_dev *dev);
+int mlx4_init_uar_table(struct mlx4_dev *dev);
+int mlx4_init_mr_table(struct mlx4_dev *dev);
+int mlx4_init_eq_table(struct mlx4_dev *dev);
+int mlx4_init_cq_table(struct mlx4_dev *dev);
+int mlx4_init_qp_table(struct mlx4_dev *dev);
+int mlx4_init_srq_table(struct mlx4_dev *dev);
+int mlx4_init_mcg_table(struct mlx4_dev *dev);
+
+void mlx4_cleanup_pd_table(struct mlx4_dev *dev);
+void mlx4_cleanup_uar_table(struct mlx4_dev *dev);
+void mlx4_cleanup_mr_table(struct mlx4_dev *dev);
+void mlx4_cleanup_eq_table(struct mlx4_dev *dev);
+void mlx4_cleanup_cq_table(struct mlx4_dev *dev);
+void mlx4_cleanup_qp_table(struct mlx4_dev *dev);
+void mlx4_cleanup_srq_table(struct mlx4_dev *dev);
+void mlx4_cleanup_mcg_table(struct mlx4_dev *dev);
+
+void mlx4_start_catas_poll(struct mlx4_dev *dev);
+void mlx4_stop_catas_poll(struct mlx4_dev *dev);
+void mlx4_catas_init(void);
+int mlx4_restart_one(struct pci_dev *pdev);
+int mlx4_register_device(struct mlx4_dev *dev);
+void mlx4_unregister_device(struct mlx4_dev *dev);
+void mlx4_dispatch_event(struct mlx4_dev *dev, enum mlx4_dev_event type, int port);
+
+struct mlx4_dev_cap;
+struct mlx4_init_hca_param;
+
+u64 mlx4_make_profile(struct mlx4_dev *dev,
+		      struct mlx4_profile *request,
+		      struct mlx4_dev_cap *dev_cap,
+		      struct mlx4_init_hca_param *init_hca);
+
+int mlx4_cmd_init(struct mlx4_dev *dev);
+void mlx4_cmd_cleanup(struct mlx4_dev *dev);
+void mlx4_cmd_event(struct mlx4_dev *dev, u16 token, u8 status, u64 out_param);
+int mlx4_cmd_use_events(struct mlx4_dev *dev);
+void mlx4_cmd_use_polling(struct mlx4_dev *dev);
+
+void mlx4_cq_completion(struct mlx4_dev *dev, u32 cqn);
+void mlx4_cq_event(struct mlx4_dev *dev, u32 cqn, int event_type);
+
+void mlx4_qp_event(struct mlx4_dev *dev, u32 qpn, int event_type);
+
+void mlx4_srq_event(struct mlx4_dev *dev, u32 srqn, int event_type);
+
+void mlx4_handle_catas_err(struct mlx4_dev *dev);
+
+void mlx4_do_sense_ports(struct mlx4_dev *dev,
+			 enum mlx4_port_type *stype,
+			 enum mlx4_port_type *defaults);
+void mlx4_start_sense(struct mlx4_dev *dev);
+void mlx4_stop_sense(struct mlx4_dev *dev);
+void mlx4_sense_init(struct mlx4_dev *dev);
+int mlx4_check_port_params(struct mlx4_dev *dev,
+			   enum mlx4_port_type *port_type);
+int mlx4_change_port_types(struct mlx4_dev *dev,
+			   enum mlx4_port_type *port_types);
+
+void mlx4_init_mac_table(struct mlx4_dev *dev, struct mlx4_mac_table *table);
+void mlx4_init_vlan_table(struct mlx4_dev *dev, struct mlx4_vlan_table *table);
+
+int mlx4_SET_PORT(struct mlx4_dev *dev, u8 port);
+int mlx4_get_port_ib_caps(struct mlx4_dev *dev, u8 port, __be32 *caps);
+
+#endif /* MLX4_H */

kernel/drivers/net/mlx4/mlx4_en.h

@@ -0,0 +1,570 @@
+/*
+ * Copyright (c) 2007 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
+ * 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 _MLX4_EN_H_
+#define _MLX4_EN_H_
+
+#include <linux/compiler.h>
+#include <linux/list.h>
+#include <linux/mutex.h>
+#include <linux/netdevice.h>
+#include <linux/inet_lro.h>
+
+#include <linux/mlx4/device.h>
+#include <linux/mlx4/qp.h>
+#include <linux/mlx4/cq.h>
+#include <linux/mlx4/srq.h>
+#include <linux/mlx4/doorbell.h>
+
+#include "en_port.h"
+
+#define DRV_NAME	"mlx4_en"
+#define DRV_VERSION	"1.4.1.1"
+#define DRV_RELDATE	"June 2009"
+
+
+#define MLX4_EN_MSG_LEVEL	(NETIF_MSG_LINK | NETIF_MSG_IFDOWN)
+
+#define en_print(level, priv, format, arg...)			\
+	{							\
+	if ((priv)->registered)					\
+		printk(level "%s: %s: " format, DRV_NAME,	\
+			(priv->dev)->name, ## arg);		\
+	else							\
+		printk(level "%s: %s: Port %d: " format,	\
+			DRV_NAME, dev_name(&priv->mdev->pdev->dev), \
+			(priv)->port, ## arg);			\
+	}
+
+#define en_dbg(mlevel, priv, format, arg...)			\
+	{							\
+	if (NETIF_MSG_##mlevel & priv->msg_enable)		\
+		en_print(KERN_DEBUG, priv, format, ## arg)	\
+	}
+#define en_warn(priv, format, arg...)				\
+	en_print(KERN_WARNING, priv, format, ## arg)
+#define en_err(priv, format, arg...)				\
+	en_print(KERN_ERR, priv, format, ## arg)
+
+#define mlx4_err(mdev, format, arg...) \
+	printk(KERN_ERR "%s %s: " format , DRV_NAME ,\
+		dev_name(&mdev->pdev->dev) , ## arg)
+#define mlx4_info(mdev, format, arg...) \
+	printk(KERN_INFO "%s %s: " format , DRV_NAME ,\
+		dev_name(&mdev->pdev->dev) , ## arg)
+#define mlx4_warn(mdev, format, arg...) \
+	printk(KERN_WARNING "%s %s: " format , DRV_NAME ,\
+		dev_name(&mdev->pdev->dev) , ## arg)
+
+/*
+ * Device constants
+ */
+
+
+#define MLX4_EN_PAGE_SHIFT	12
+#define MLX4_EN_PAGE_SIZE	(1 << MLX4_EN_PAGE_SHIFT)
+#define MAX_TX_RINGS		16
+#define MAX_RX_RINGS		16
+#define TXBB_SIZE		64
+#define HEADROOM		(2048 / TXBB_SIZE + 1)
+#define STAMP_STRIDE		64
+#define STAMP_DWORDS		(STAMP_STRIDE / 4)
+#define STAMP_SHIFT		31
+#define STAMP_VAL		0x7fffffff
+#define STATS_DELAY		(HZ / 4)
+
+/* Typical TSO descriptor with 16 gather entries is 352 bytes... */
+#define MAX_DESC_SIZE		512
+#define MAX_DESC_TXBBS		(MAX_DESC_SIZE / TXBB_SIZE)
+
+/*
+ * OS related constants and tunables
+ */
+
+#define MLX4_EN_WATCHDOG_TIMEOUT	(15 * HZ)
+
+#define MLX4_EN_ALLOC_ORDER	2
+#define MLX4_EN_ALLOC_SIZE	(PAGE_SIZE << MLX4_EN_ALLOC_ORDER)
+
+#define MLX4_EN_MAX_LRO_DESCRIPTORS	32
+
+/* Receive fragment sizes; we use at most 4 fragments (for 9600 byte MTU
+ * and 4K allocations) */
+enum {
+	FRAG_SZ0 = 512 - NET_IP_ALIGN,
+	FRAG_SZ1 = 1024,
+	FRAG_SZ2 = 4096,
+	FRAG_SZ3 = MLX4_EN_ALLOC_SIZE
+};
+#define MLX4_EN_MAX_RX_FRAGS	4
+
+/* Maximum ring sizes */
+#define MLX4_EN_MAX_TX_SIZE	8192
+#define MLX4_EN_MAX_RX_SIZE	8192
+
+/* Minimum ring size for our page-allocation sceme to work */
+#define MLX4_EN_MIN_RX_SIZE	(MLX4_EN_ALLOC_SIZE / SMP_CACHE_BYTES)
+#define MLX4_EN_MIN_TX_SIZE	(4096 / TXBB_SIZE)
+
+#define MLX4_EN_SMALL_PKT_SIZE		64
+#define MLX4_EN_NUM_TX_RINGS		8
+#define MLX4_EN_NUM_PPP_RINGS		8
+#define MLX4_EN_DEF_TX_RING_SIZE	512
+#define MLX4_EN_DEF_RX_RING_SIZE  	1024
+
+/* Target number of packets to coalesce with interrupt moderation */
+#define MLX4_EN_RX_COAL_TARGET	44
+#define MLX4_EN_RX_COAL_TIME	0x10
+
+#define MLX4_EN_TX_COAL_PKTS	5
+#define MLX4_EN_TX_COAL_TIME	0x80
+
+#define MLX4_EN_RX_RATE_LOW		400000
+#define MLX4_EN_RX_COAL_TIME_LOW	0
+#define MLX4_EN_RX_RATE_HIGH		450000
+#define MLX4_EN_RX_COAL_TIME_HIGH	128
+#define MLX4_EN_RX_SIZE_THRESH		1024
+#define MLX4_EN_RX_RATE_THRESH		(1000000 / MLX4_EN_RX_COAL_TIME_HIGH)
+#define MLX4_EN_SAMPLE_INTERVAL		0
+
+#define MLX4_EN_AUTO_CONF	0xffff
+
+#define MLX4_EN_DEF_RX_PAUSE	1
+#define MLX4_EN_DEF_TX_PAUSE	1
+
+/* Interval between sucessive polls in the Tx routine when polling is used
+   instead of interrupts (in per-core Tx rings) - should be power of 2 */
+#define MLX4_EN_TX_POLL_MODER	16
+#define MLX4_EN_TX_POLL_TIMEOUT	(HZ / 4)
+
+#define ETH_LLC_SNAP_SIZE	8
+
+#define SMALL_PACKET_SIZE      (256 - NET_IP_ALIGN)
+#define HEADER_COPY_SIZE       (128 - NET_IP_ALIGN)
+
+#define MLX4_EN_MIN_MTU		46
+#define ETH_BCAST		0xffffffffffffULL
+
+#ifdef MLX4_EN_PERF_STAT
+/* Number of samples to 'average' */
+#define AVG_SIZE			128
+#define AVG_FACTOR			1024
+#define NUM_PERF_STATS			NUM_PERF_COUNTERS
+
+#define INC_PERF_COUNTER(cnt)		(++(cnt))
+#define ADD_PERF_COUNTER(cnt, add)	((cnt) += (add))
+#define AVG_PERF_COUNTER(cnt, sample) \
+	((cnt) = ((cnt) * (AVG_SIZE - 1) + (sample) * AVG_FACTOR) / AVG_SIZE)
+#define GET_PERF_COUNTER(cnt)		(cnt)
+#define GET_AVG_PERF_COUNTER(cnt)	((cnt) / AVG_FACTOR)
+
+#else
+
+#define NUM_PERF_STATS			0
+#define INC_PERF_COUNTER(cnt)		do {} while (0)
+#define ADD_PERF_COUNTER(cnt, add)	do {} while (0)
+#define AVG_PERF_COUNTER(cnt, sample)	do {} while (0)
+#define GET_PERF_COUNTER(cnt)		(0)
+#define GET_AVG_PERF_COUNTER(cnt)	(0)
+#endif /* MLX4_EN_PERF_STAT */
+
+/*
+ * Configurables
+ */
+
+enum cq_type {
+	RX = 0,
+	TX = 1,
+};
+
+
+/*
+ * Useful macros
+ */
+#define ROUNDUP_LOG2(x)		ilog2(roundup_pow_of_two(x))
+#define XNOR(x, y)		(!(x) == !(y))
+#define ILLEGAL_MAC(addr)	(addr == 0xffffffffffffULL || addr == 0x0)
+
+
+struct mlx4_en_tx_info {
+	struct sk_buff *skb;
+	u32 nr_txbb;
+	u8 linear;
+	u8 data_offset;
+	u8 inl;
+};
+
+
+#define MLX4_EN_BIT_DESC_OWN	0x80000000
+#define CTRL_SIZE	sizeof(struct mlx4_wqe_ctrl_seg)
+#define MLX4_EN_MEMTYPE_PAD	0x100
+#define DS_SIZE		sizeof(struct mlx4_wqe_data_seg)
+
+
+struct mlx4_en_tx_desc {
+	struct mlx4_wqe_ctrl_seg ctrl;
+	union {
+		struct mlx4_wqe_data_seg data; /* at least one data segment */
+		struct mlx4_wqe_lso_seg lso;
+		struct mlx4_wqe_inline_seg inl;
+	};
+};
+
+#define MLX4_EN_USE_SRQ		0x01000000
+
+struct mlx4_en_rx_alloc {
+	struct page *page;
+	u16 offset;
+};
+
+struct mlx4_en_tx_ring {
+	struct mlx4_hwq_resources wqres;
+	u32 size ; /* number of TXBBs */
+	u32 size_mask;
+	u16 stride;
+	u16 cqn;	/* index of port CQ associated with this ring */
+	u32 prod;
+	u32 cons;
+	u32 buf_size;
+	u32 doorbell_qpn;
+	void *buf;
+	u16 poll_cnt;
+	int blocked;
+	struct mlx4_en_tx_info *tx_info;
+	u8 *bounce_buf;
+	u32 last_nr_txbb;
+	struct mlx4_qp qp;
+	struct mlx4_qp_context context;
+	int qpn;
+	enum mlx4_qp_state qp_state;
+	struct mlx4_srq dummy;
+	unsigned long bytes;
+	unsigned long packets;
+	spinlock_t comp_lock;
+};
+
+struct mlx4_en_rx_desc {
+	/* actual number of entries depends on rx ring stride */
+	struct mlx4_wqe_data_seg data[0];
+};
+
+struct mlx4_en_rx_ring {
+	struct mlx4_hwq_resources wqres;
+	struct mlx4_en_rx_alloc page_alloc[MLX4_EN_MAX_RX_FRAGS];
+	struct net_lro_mgr lro;
+	u32 size ;	/* number of Rx descs*/
+	u32 actual_size;
+	u32 size_mask;
+	u16 stride;
+	u16 log_stride;
+	u16 cqn;	/* index of port CQ associated with this ring */
+	u32 prod;
+	u32 cons;
+	u32 buf_size;
+	void *buf;
+	void *rx_info;
+	unsigned long bytes;
+	unsigned long packets;
+};
+
+
+static inline int mlx4_en_can_lro(__be16 status)
+{
+	return (status & cpu_to_be16(MLX4_CQE_STATUS_IPV4	|
+				     MLX4_CQE_STATUS_IPV4F	|
+				     MLX4_CQE_STATUS_IPV6	|
+				     MLX4_CQE_STATUS_IPV4OPT	|
+				     MLX4_CQE_STATUS_TCP	|
+				     MLX4_CQE_STATUS_UDP	|
+				     MLX4_CQE_STATUS_IPOK)) ==
+		cpu_to_be16(MLX4_CQE_STATUS_IPV4 |
+			    MLX4_CQE_STATUS_IPOK |
+			    MLX4_CQE_STATUS_TCP);
+}
+
+struct mlx4_en_cq {
+	struct mlx4_cq          mcq;
+	struct mlx4_hwq_resources wqres;
+	int                     ring;
+	spinlock_t              lock;
+	struct net_device      *dev;
+	struct napi_struct	napi;
+	/* Per-core Tx cq processing support */
+	struct timer_list timer;
+	int size;
+	int buf_size;
+	unsigned vector;
+	enum cq_type is_tx;
+	u16 moder_time;
+	u16 moder_cnt;
+	struct mlx4_cqe *buf;
+#define MLX4_EN_OPCODE_ERROR	0x1e
+};
+
+struct mlx4_en_port_profile {
+	u32 flags;
+	u32 tx_ring_num;
+	u32 rx_ring_num;
+	u32 tx_ring_size;
+	u32 rx_ring_size;
+	u8 rx_pause;
+	u8 rx_ppp;
+	u8 tx_pause;
+	u8 tx_ppp;
+};
+
+struct mlx4_en_profile {
+	int rss_xor;
+	int num_lro;
+	u8 rss_mask;
+	u32 active_ports;
+	u32 small_pkt_int;
+	u8 no_reset;
+	struct mlx4_en_port_profile prof[MLX4_MAX_PORTS + 1];
+};
+
+struct mlx4_en_dev {
+	struct mlx4_dev         *dev;
+	struct pci_dev		*pdev;
+	struct mutex		state_lock;
+	struct net_device       *pndev[MLX4_MAX_PORTS + 1];
+	u32                     port_cnt;
+	bool			device_up;
+	struct mlx4_en_profile  profile;
+	u32			LSO_support;
+	struct workqueue_struct *workqueue;
+	struct device           *dma_device;
+	void __iomem            *uar_map;
+	struct mlx4_uar         priv_uar;
+	struct mlx4_mr		mr;
+	u32                     priv_pdn;
+	spinlock_t              uar_lock;
+};
+
+
+struct mlx4_en_rss_map {
+	int base_qpn;
+	struct mlx4_qp qps[MAX_RX_RINGS];
+	enum mlx4_qp_state state[MAX_RX_RINGS];
+	struct mlx4_qp indir_qp;
+	enum mlx4_qp_state indir_state;
+};
+
+struct mlx4_en_rss_context {
+	__be32 base_qpn;
+	__be32 default_qpn;
+	u16 reserved;
+	u8 hash_fn;
+	u8 flags;
+	__be32 rss_key[10];
+};
+
+struct mlx4_en_pkt_stats {
+	unsigned long broadcast;
+	unsigned long rx_prio[8];
+	unsigned long tx_prio[8];
+#define NUM_PKT_STATS		17
+};
+
+struct mlx4_en_port_stats {
+	unsigned long lro_aggregated;
+	unsigned long lro_flushed;
+	unsigned long lro_no_desc;
+	unsigned long tso_packets;
+	unsigned long queue_stopped;
+	unsigned long wake_queue;
+	unsigned long tx_timeout;
+	unsigned long rx_alloc_failed;
+	unsigned long rx_chksum_good;
+	unsigned long rx_chksum_none;
+	unsigned long tx_chksum_offload;
+#define NUM_PORT_STATS		11
+};
+
+struct mlx4_en_perf_stats {
+	u32 tx_poll;
+	u64 tx_pktsz_avg;
+	u32 inflight_avg;
+	u16 tx_coal_avg;
+	u16 rx_coal_avg;
+	u32 napi_quota;
+#define NUM_PERF_COUNTERS		6
+};
+
+struct mlx4_en_frag_info {
+	u16 frag_size;
+	u16 frag_prefix_size;
+	u16 frag_stride;
+	u16 frag_align;
+	u16 last_offset;
+
+};
+
+struct mlx4_en_priv {
+	struct mlx4_en_dev *mdev;
+	struct mlx4_en_port_profile *prof;
+	struct net_device *dev;
+	struct vlan_group *vlgrp;
+	struct net_device_stats stats;
+	struct net_device_stats ret_stats;
+	spinlock_t stats_lock;
+
+	unsigned long last_moder_packets;
+	unsigned long last_moder_tx_packets;
+	unsigned long last_moder_bytes;
+	unsigned long last_moder_jiffies;
+	int last_moder_time;
+	u16 rx_usecs;
+	u16 rx_frames;
+	u16 tx_usecs;
+	u16 tx_frames;
+	u32 pkt_rate_low;
+	u16 rx_usecs_low;
+	u32 pkt_rate_high;
+	u16 rx_usecs_high;
+	u16 sample_interval;
+	u16 adaptive_rx_coal;
+	u32 msg_enable;
+
+	struct mlx4_hwq_resources res;
+	int link_state;
+	int last_link_state;
+	bool port_up;
+	int port;
+	int registered;
+	int allocated;
+	int stride;
+	int rx_csum;
+	u64 mac;
+	int mac_index;
+	unsigned max_mtu;
+	int base_qpn;
+
+	struct mlx4_en_rss_map rss_map;
+	u32 flags;
+#define MLX4_EN_FLAG_PROMISC	0x1
+	u32 tx_ring_num;
+	u32 rx_ring_num;
+	u32 rx_skb_size;
+	struct mlx4_en_frag_info frag_info[MLX4_EN_MAX_RX_FRAGS];
+	u16 num_frags;
+	u16 log_rx_info;
+
+	struct mlx4_en_tx_ring tx_ring[MAX_TX_RINGS];
+	struct mlx4_en_rx_ring rx_ring[MAX_RX_RINGS];
+	struct mlx4_en_cq tx_cq[MAX_TX_RINGS];
+	struct mlx4_en_cq rx_cq[MAX_RX_RINGS];
+	struct work_struct mcast_task;
+	struct work_struct mac_task;
+	struct work_struct watchdog_task;
+	struct work_struct linkstate_task;
+	struct delayed_work stats_task;
+	struct mlx4_en_perf_stats pstats;
+	struct mlx4_en_pkt_stats pkstats;
+	struct mlx4_en_port_stats port_stats;
+	struct dev_mc_list *mc_list;
+	struct mlx4_en_stat_out_mbox hw_stats;
+};
+
+
+void mlx4_en_destroy_netdev(struct net_device *dev);
+int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port,
+			struct mlx4_en_port_profile *prof);
+
+int mlx4_en_start_port(struct net_device *dev);
+void mlx4_en_stop_port(struct net_device *dev);
+
+void mlx4_en_free_resources(struct mlx4_en_priv *priv);
+int mlx4_en_alloc_resources(struct mlx4_en_priv *priv);
+
+int mlx4_en_create_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq,
+		      int entries, int ring, enum cq_type mode);
+void mlx4_en_destroy_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq);
+int mlx4_en_activate_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq);
+void mlx4_en_deactivate_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq);
+int mlx4_en_set_cq_moder(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq);
+int mlx4_en_arm_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq);
+
+void mlx4_en_poll_tx_cq(unsigned long data);
+void mlx4_en_tx_irq(struct mlx4_cq *mcq);
+u16 mlx4_en_select_queue(struct net_device *dev, struct sk_buff *skb);
+netdev_tx_t mlx4_en_xmit(struct sk_buff *skb, struct net_device *dev);
+
+int mlx4_en_create_tx_ring(struct mlx4_en_priv *priv, struct mlx4_en_tx_ring *ring,
+			   u32 size, u16 stride);
+void mlx4_en_destroy_tx_ring(struct mlx4_en_priv *priv, struct mlx4_en_tx_ring *ring);
+int mlx4_en_activate_tx_ring(struct mlx4_en_priv *priv,
+			     struct mlx4_en_tx_ring *ring,
+			     int cq);
+void mlx4_en_deactivate_tx_ring(struct mlx4_en_priv *priv,
+				struct mlx4_en_tx_ring *ring);
+
+int mlx4_en_create_rx_ring(struct mlx4_en_priv *priv,
+			   struct mlx4_en_rx_ring *ring,
+			   u32 size, u16 stride);
+void mlx4_en_destroy_rx_ring(struct mlx4_en_priv *priv,
+			     struct mlx4_en_rx_ring *ring);
+int mlx4_en_activate_rx_rings(struct mlx4_en_priv *priv);
+void mlx4_en_deactivate_rx_ring(struct mlx4_en_priv *priv,
+				struct mlx4_en_rx_ring *ring);
+int mlx4_en_process_rx_cq(struct net_device *dev,
+			  struct mlx4_en_cq *cq,
+			  int budget);
+int mlx4_en_poll_rx_cq(struct napi_struct *napi, int budget);
+void mlx4_en_fill_qp_context(struct mlx4_en_priv *priv, int size, int stride,
+			     int is_tx, int rss, int qpn, int cqn,
+			     struct mlx4_qp_context *context);
+void mlx4_en_sqp_event(struct mlx4_qp *qp, enum mlx4_event event);
+int mlx4_en_map_buffer(struct mlx4_buf *buf);
+void mlx4_en_unmap_buffer(struct mlx4_buf *buf);
+
+void mlx4_en_calc_rx_buf(struct net_device *dev);
+int mlx4_en_config_rss_steer(struct mlx4_en_priv *priv);
+void mlx4_en_release_rss_steer(struct mlx4_en_priv *priv);
+int mlx4_en_free_tx_buf(struct net_device *dev, struct mlx4_en_tx_ring *ring);
+void mlx4_en_rx_irq(struct mlx4_cq *mcq);
+
+int mlx4_SET_MCAST_FLTR(struct mlx4_dev *dev, u8 port, u64 mac, u64 clear, u8 mode);
+int mlx4_SET_VLAN_FLTR(struct mlx4_dev *dev, u8 port, struct vlan_group *grp);
+int mlx4_SET_PORT_general(struct mlx4_dev *dev, u8 port, int mtu,
+			  u8 pptx, u8 pfctx, u8 pprx, u8 pfcrx);
+int mlx4_SET_PORT_qpn_calc(struct mlx4_dev *dev, u8 port, u32 base_qpn,
+			   u8 promisc);
+
+int mlx4_en_DUMP_ETH_STATS(struct mlx4_en_dev *mdev, u8 port, u8 reset);
+
+/*
+ * Globals
+ */
+extern const struct ethtool_ops mlx4_en_ethtool_ops;
+#endif

kernel/drivers/net/mlx4/mr.c

@@ -0,0 +1,666 @@
+/*
+ * Copyright (c) 2004 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2006, 2007 Cisco Systems, 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/errno.h>
+
+#include <linux/mlx4/cmd.h>
+
+#include "mlx4.h"
+#include "icm.h"
+
+/*
+ * Must be packed because mtt_seg is 64 bits but only aligned to 32 bits.
+ */
+struct mlx4_mpt_entry {
+	__be32 flags;
+	__be32 qpn;
+	__be32 key;
+	__be32 pd_flags;
+	__be64 start;
+	__be64 length;
+	__be32 lkey;
+	__be32 win_cnt;
+	u8	reserved1[3];
+	u8	mtt_rep;
+	__be64 mtt_seg;
+	__be32 mtt_sz;
+	__be32 entity_size;
+	__be32 first_byte_offset;
+} __attribute__((packed));
+
+#define MLX4_MPT_FLAG_SW_OWNS	    (0xfUL << 28)
+#define MLX4_MPT_FLAG_FREE	    (0x3UL << 28)
+#define MLX4_MPT_FLAG_MIO	    (1 << 17)
+#define MLX4_MPT_FLAG_BIND_ENABLE   (1 << 15)
+#define MLX4_MPT_FLAG_PHYSICAL	    (1 <<  9)
+#define MLX4_MPT_FLAG_REGION	    (1 <<  8)
+
+#define MLX4_MPT_PD_FLAG_FAST_REG   (1 << 27)
+#define MLX4_MPT_PD_FLAG_RAE	    (1 << 28)
+#define MLX4_MPT_PD_FLAG_EN_INV	    (3 << 24)
+
+#define MLX4_MPT_STATUS_SW		0xF0
+#define MLX4_MPT_STATUS_HW		0x00
+
+static u32 mlx4_buddy_alloc(struct mlx4_buddy *buddy, int order)
+{
+	int o;
+	int m;
+	u32 seg;
+
+	spin_lock(&buddy->lock);
+
+	for (o = order; o <= buddy->max_order; ++o)
+		if (buddy->num_free[o]) {
+			m = 1 << (buddy->max_order - o);
+			seg = find_first_bit(buddy->bits[o], m);
+			if (seg < m)
+				goto found;
+		}
+
+	spin_unlock(&buddy->lock);
+	return -1;
+
+ found:
+	clear_bit(seg, buddy->bits[o]);
+	--buddy->num_free[o];
+
+	while (o > order) {
+		--o;
+		seg <<= 1;
+		set_bit(seg ^ 1, buddy->bits[o]);
+		++buddy->num_free[o];
+	}
+
+	spin_unlock(&buddy->lock);
+
+	seg <<= order;
+
+	return seg;
+}
+
+static void mlx4_buddy_free(struct mlx4_buddy *buddy, u32 seg, int order)
+{
+	seg >>= order;
+
+	spin_lock(&buddy->lock);
+
+	while (test_bit(seg ^ 1, buddy->bits[order])) {
+		clear_bit(seg ^ 1, buddy->bits[order]);
+		--buddy->num_free[order];
+		seg >>= 1;
+		++order;
+	}
+
+	set_bit(seg, buddy->bits[order]);
+	++buddy->num_free[order];
+
+	spin_unlock(&buddy->lock);
+}
+
+static int mlx4_buddy_init(struct mlx4_buddy *buddy, int max_order)
+{
+	int i, s;
+
+	buddy->max_order = max_order;
+	spin_lock_init(&buddy->lock);
+
+	buddy->bits = kzalloc((buddy->max_order + 1) * sizeof (long *),
+			      GFP_KERNEL);
+	buddy->num_free = kzalloc((buddy->max_order + 1) * sizeof (int *),
+				  GFP_KERNEL);
+	if (!buddy->bits || !buddy->num_free)
+		goto err_out;
+
+	for (i = 0; i <= buddy->max_order; ++i) {
+		s = BITS_TO_LONGS(1 << (buddy->max_order - i));
+		buddy->bits[i] = kmalloc(s * sizeof (long), GFP_KERNEL);
+		if (!buddy->bits[i])
+			goto err_out_free;
+		bitmap_zero(buddy->bits[i], 1 << (buddy->max_order - i));
+	}
+
+	set_bit(0, buddy->bits[buddy->max_order]);
+	buddy->num_free[buddy->max_order] = 1;
+
+	return 0;
+
+err_out_free:
+	for (i = 0; i <= buddy->max_order; ++i)
+		kfree(buddy->bits[i]);
+
+err_out:
+	kfree(buddy->bits);
+	kfree(buddy->num_free);
+
+	return -ENOMEM;
+}
+
+static void mlx4_buddy_cleanup(struct mlx4_buddy *buddy)
+{
+	int i;
+
+	for (i = 0; i <= buddy->max_order; ++i)
+		kfree(buddy->bits[i]);
+
+	kfree(buddy->bits);
+	kfree(buddy->num_free);
+}
+
+static u32 mlx4_alloc_mtt_range(struct mlx4_dev *dev, int order)
+{
+	struct mlx4_mr_table *mr_table = &mlx4_priv(dev)->mr_table;
+	u32 seg;
+
+	seg = mlx4_buddy_alloc(&mr_table->mtt_buddy, order);
+	if (seg == -1)
+		return -1;
+
+	if (mlx4_table_get_range(dev, &mr_table->mtt_table, seg,
+				 seg + (1 << order) - 1)) {
+		mlx4_buddy_free(&mr_table->mtt_buddy, seg, order);
+		return -1;
+	}
+
+	return seg;
+}
+
+int mlx4_mtt_init(struct mlx4_dev *dev, int npages, int page_shift,
+		  struct mlx4_mtt *mtt)
+{
+	int i;
+
+	if (!npages) {
+		mtt->order      = -1;
+		mtt->page_shift = MLX4_ICM_PAGE_SHIFT;
+		return 0;
+	} else
+		mtt->page_shift = page_shift;
+
+	for (mtt->order = 0, i = dev->caps.mtts_per_seg; i < npages; i <<= 1)
+		++mtt->order;
+
+	mtt->first_seg = mlx4_alloc_mtt_range(dev, mtt->order);
+	if (mtt->first_seg == -1)
+		return -ENOMEM;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_mtt_init);
+
+void mlx4_mtt_cleanup(struct mlx4_dev *dev, struct mlx4_mtt *mtt)
+{
+	struct mlx4_mr_table *mr_table = &mlx4_priv(dev)->mr_table;
+
+	if (mtt->order < 0)
+		return;
+
+	mlx4_buddy_free(&mr_table->mtt_buddy, mtt->first_seg, mtt->order);
+	mlx4_table_put_range(dev, &mr_table->mtt_table, mtt->first_seg,
+			     mtt->first_seg + (1 << mtt->order) - 1);
+}
+EXPORT_SYMBOL_GPL(mlx4_mtt_cleanup);
+
+u64 mlx4_mtt_addr(struct mlx4_dev *dev, struct mlx4_mtt *mtt)
+{
+	return (u64) mtt->first_seg * dev->caps.mtt_entry_sz;
+}
+EXPORT_SYMBOL_GPL(mlx4_mtt_addr);
+
+static u32 hw_index_to_key(u32 ind)
+{
+	return (ind >> 24) | (ind << 8);
+}
+
+static u32 key_to_hw_index(u32 key)
+{
+	return (key << 24) | (key >> 8);
+}
+
+static int mlx4_SW2HW_MPT(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *mailbox,
+			  int mpt_index)
+{
+	return mlx4_cmd(dev, mailbox->dma, mpt_index, 0, MLX4_CMD_SW2HW_MPT,
+			MLX4_CMD_TIME_CLASS_B);
+}
+
+static int mlx4_HW2SW_MPT(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *mailbox,
+			  int mpt_index)
+{
+	return mlx4_cmd_box(dev, 0, mailbox ? mailbox->dma : 0, mpt_index,
+			    !mailbox, MLX4_CMD_HW2SW_MPT, MLX4_CMD_TIME_CLASS_B);
+}
+
+int mlx4_mr_alloc(struct mlx4_dev *dev, u32 pd, u64 iova, u64 size, u32 access,
+		  int npages, int page_shift, struct mlx4_mr *mr)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	u32 index;
+	int err;
+
+	index = mlx4_bitmap_alloc(&priv->mr_table.mpt_bitmap);
+	if (index == -1)
+		return -ENOMEM;
+
+	mr->iova       = iova;
+	mr->size       = size;
+	mr->pd	       = pd;
+	mr->access     = access;
+	mr->enabled    = 0;
+	mr->key	       = hw_index_to_key(index);
+
+	err = mlx4_mtt_init(dev, npages, page_shift, &mr->mtt);
+	if (err)
+		mlx4_bitmap_free(&priv->mr_table.mpt_bitmap, index);
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_mr_alloc);
+
+void mlx4_mr_free(struct mlx4_dev *dev, struct mlx4_mr *mr)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int err;
+
+	if (mr->enabled) {
+		err = mlx4_HW2SW_MPT(dev, NULL,
+				     key_to_hw_index(mr->key) &
+				     (dev->caps.num_mpts - 1));
+		if (err)
+			mlx4_warn(dev, "HW2SW_MPT failed (%d)\n", err);
+	}
+
+	mlx4_mtt_cleanup(dev, &mr->mtt);
+	mlx4_bitmap_free(&priv->mr_table.mpt_bitmap, key_to_hw_index(mr->key));
+}
+EXPORT_SYMBOL_GPL(mlx4_mr_free);
+
+int mlx4_mr_enable(struct mlx4_dev *dev, struct mlx4_mr *mr)
+{
+	struct mlx4_mr_table *mr_table = &mlx4_priv(dev)->mr_table;
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_mpt_entry *mpt_entry;
+	int err;
+
+	err = mlx4_table_get(dev, &mr_table->dmpt_table, key_to_hw_index(mr->key));
+	if (err)
+		return err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox)) {
+		err = PTR_ERR(mailbox);
+		goto err_table;
+	}
+	mpt_entry = mailbox->buf;
+
+	memset(mpt_entry, 0, sizeof *mpt_entry);
+
+	mpt_entry->flags = cpu_to_be32(MLX4_MPT_FLAG_MIO	 |
+				       MLX4_MPT_FLAG_REGION	 |
+				       mr->access);
+
+	mpt_entry->key	       = cpu_to_be32(key_to_hw_index(mr->key));
+	mpt_entry->pd_flags    = cpu_to_be32(mr->pd | MLX4_MPT_PD_FLAG_EN_INV);
+	mpt_entry->start       = cpu_to_be64(mr->iova);
+	mpt_entry->length      = cpu_to_be64(mr->size);
+	mpt_entry->entity_size = cpu_to_be32(mr->mtt.page_shift);
+
+	if (mr->mtt.order < 0) {
+		mpt_entry->flags |= cpu_to_be32(MLX4_MPT_FLAG_PHYSICAL);
+		mpt_entry->mtt_seg = 0;
+	} else {
+		mpt_entry->mtt_seg = cpu_to_be64(mlx4_mtt_addr(dev, &mr->mtt));
+	}
+
+	if (mr->mtt.order >= 0 && mr->mtt.page_shift == 0) {
+		/* fast register MR in free state */
+		mpt_entry->flags    |= cpu_to_be32(MLX4_MPT_FLAG_FREE);
+		mpt_entry->pd_flags |= cpu_to_be32(MLX4_MPT_PD_FLAG_FAST_REG |
+						   MLX4_MPT_PD_FLAG_RAE);
+		mpt_entry->mtt_sz    = cpu_to_be32((1 << mr->mtt.order) *
+						   dev->caps.mtts_per_seg);
+	} else {
+		mpt_entry->flags    |= cpu_to_be32(MLX4_MPT_FLAG_SW_OWNS);
+	}
+
+	err = mlx4_SW2HW_MPT(dev, mailbox,
+			     key_to_hw_index(mr->key) & (dev->caps.num_mpts - 1));
+	if (err) {
+		mlx4_warn(dev, "SW2HW_MPT failed (%d)\n", err);
+		goto err_cmd;
+	}
+
+	mr->enabled = 1;
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+	return 0;
+
+err_cmd:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+err_table:
+	mlx4_table_put(dev, &mr_table->dmpt_table, key_to_hw_index(mr->key));
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_mr_enable);
+
+static int mlx4_write_mtt_chunk(struct mlx4_dev *dev, struct mlx4_mtt *mtt,
+				int start_index, int npages, u64 *page_list)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	__be64 *mtts;
+	dma_addr_t dma_handle;
+	int i;
+	int s = start_index * sizeof (u64);
+
+	/* All MTTs must fit in the same page */
+	if (start_index / (PAGE_SIZE / sizeof (u64)) !=
+	    (start_index + npages - 1) / (PAGE_SIZE / sizeof (u64)))
+		return -EINVAL;
+
+	if (start_index & (dev->caps.mtts_per_seg - 1))
+		return -EINVAL;
+
+	mtts = mlx4_table_find(&priv->mr_table.mtt_table, mtt->first_seg +
+				s / dev->caps.mtt_entry_sz, &dma_handle);
+	if (!mtts)
+		return -ENOMEM;
+
+	dma_sync_single_for_cpu(&dev->pdev->dev, dma_handle,
+				npages * sizeof (u64), DMA_TO_DEVICE);
+
+	for (i = 0; i < npages; ++i)
+		mtts[i] = cpu_to_be64(page_list[i] | MLX4_MTT_FLAG_PRESENT);
+
+	dma_sync_single_for_device(&dev->pdev->dev, dma_handle,
+				   npages * sizeof (u64), DMA_TO_DEVICE);
+
+	return 0;
+}
+
+int mlx4_write_mtt(struct mlx4_dev *dev, struct mlx4_mtt *mtt,
+		   int start_index, int npages, u64 *page_list)
+{
+	int chunk;
+	int err;
+
+	if (mtt->order < 0)
+		return -EINVAL;
+
+	while (npages > 0) {
+		chunk = min_t(int, PAGE_SIZE / sizeof(u64), npages);
+		err = mlx4_write_mtt_chunk(dev, mtt, start_index, chunk, page_list);
+		if (err)
+			return err;
+
+		npages      -= chunk;
+		start_index += chunk;
+		page_list   += chunk;
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_write_mtt);
+
+int mlx4_buf_write_mtt(struct mlx4_dev *dev, struct mlx4_mtt *mtt,
+		       struct mlx4_buf *buf)
+{
+	u64 *page_list;
+	int err;
+	int i;
+
+	page_list = kmalloc(buf->npages * sizeof *page_list, GFP_KERNEL);
+	if (!page_list)
+		return -ENOMEM;
+
+	for (i = 0; i < buf->npages; ++i)
+		if (buf->nbufs == 1)
+			page_list[i] = buf->direct.map + (i << buf->page_shift);
+		else
+			page_list[i] = buf->page_list[i].map;
+
+	err = mlx4_write_mtt(dev, mtt, 0, buf->npages, page_list);
+
+	kfree(page_list);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_buf_write_mtt);
+
+int mlx4_init_mr_table(struct mlx4_dev *dev)
+{
+	struct mlx4_mr_table *mr_table = &mlx4_priv(dev)->mr_table;
+	int err;
+
+	err = mlx4_bitmap_init(&mr_table->mpt_bitmap, dev->caps.num_mpts,
+			       ~0, dev->caps.reserved_mrws, 0);
+	if (err)
+		return err;
+
+	err = mlx4_buddy_init(&mr_table->mtt_buddy,
+			      ilog2(dev->caps.num_mtt_segs));
+	if (err)
+		goto err_buddy;
+
+	if (dev->caps.reserved_mtts) {
+		if (mlx4_alloc_mtt_range(dev, fls(dev->caps.reserved_mtts - 1)) == -1) {
+			mlx4_warn(dev, "MTT table of order %d is too small.\n",
+				  mr_table->mtt_buddy.max_order);
+			err = -ENOMEM;
+			goto err_reserve_mtts;
+		}
+	}
+
+	return 0;
+
+err_reserve_mtts:
+	mlx4_buddy_cleanup(&mr_table->mtt_buddy);
+
+err_buddy:
+	mlx4_bitmap_cleanup(&mr_table->mpt_bitmap);
+
+	return err;
+}
+
+void mlx4_cleanup_mr_table(struct mlx4_dev *dev)
+{
+	struct mlx4_mr_table *mr_table = &mlx4_priv(dev)->mr_table;
+
+	mlx4_buddy_cleanup(&mr_table->mtt_buddy);
+	mlx4_bitmap_cleanup(&mr_table->mpt_bitmap);
+}
+
+static inline int mlx4_check_fmr(struct mlx4_fmr *fmr, u64 *page_list,
+				  int npages, u64 iova)
+{
+	int i, page_mask;
+
+	if (npages > fmr->max_pages)
+		return -EINVAL;
+
+	page_mask = (1 << fmr->page_shift) - 1;
+
+	/* We are getting page lists, so va must be page aligned. */
+	if (iova & page_mask)
+		return -EINVAL;
+
+	/* Trust the user not to pass misaligned data in page_list */
+	if (0)
+		for (i = 0; i < npages; ++i) {
+			if (page_list[i] & ~page_mask)
+				return -EINVAL;
+		}
+
+	if (fmr->maps >= fmr->max_maps)
+		return -EINVAL;
+
+	return 0;
+}
+
+int mlx4_map_phys_fmr(struct mlx4_dev *dev, struct mlx4_fmr *fmr, u64 *page_list,
+		      int npages, u64 iova, u32 *lkey, u32 *rkey)
+{
+	u32 key;
+	int i, err;
+
+	err = mlx4_check_fmr(fmr, page_list, npages, iova);
+	if (err)
+		return err;
+
+	++fmr->maps;
+
+	key = key_to_hw_index(fmr->mr.key);
+	key += dev->caps.num_mpts;
+	*lkey = *rkey = fmr->mr.key = hw_index_to_key(key);
+
+	*(u8 *) fmr->mpt = MLX4_MPT_STATUS_SW;
+
+	/* Make sure MPT status is visible before writing MTT entries */
+	wmb();
+
+	dma_sync_single_for_cpu(&dev->pdev->dev, fmr->dma_handle,
+				npages * sizeof(u64), DMA_TO_DEVICE);
+
+	for (i = 0; i < npages; ++i)
+		fmr->mtts[i] = cpu_to_be64(page_list[i] | MLX4_MTT_FLAG_PRESENT);
+
+	dma_sync_single_for_device(&dev->pdev->dev, fmr->dma_handle,
+				   npages * sizeof(u64), DMA_TO_DEVICE);
+
+	fmr->mpt->key    = cpu_to_be32(key);
+	fmr->mpt->lkey   = cpu_to_be32(key);
+	fmr->mpt->length = cpu_to_be64(npages * (1ull << fmr->page_shift));
+	fmr->mpt->start  = cpu_to_be64(iova);
+
+	/* Make MTT entries are visible before setting MPT status */
+	wmb();
+
+	*(u8 *) fmr->mpt = MLX4_MPT_STATUS_HW;
+
+	/* Make sure MPT status is visible before consumer can use FMR */
+	wmb();
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_map_phys_fmr);
+
+int mlx4_fmr_alloc(struct mlx4_dev *dev, u32 pd, u32 access, int max_pages,
+		   int max_maps, u8 page_shift, struct mlx4_fmr *fmr)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	u64 mtt_seg;
+	int err = -ENOMEM;
+
+	if (page_shift < (ffs(dev->caps.page_size_cap) - 1) || page_shift >= 32)
+		return -EINVAL;
+
+	/* All MTTs must fit in the same page */
+	if (max_pages * sizeof *fmr->mtts > PAGE_SIZE)
+		return -EINVAL;
+
+	fmr->page_shift = page_shift;
+	fmr->max_pages  = max_pages;
+	fmr->max_maps   = max_maps;
+	fmr->maps = 0;
+
+	err = mlx4_mr_alloc(dev, pd, 0, 0, access, max_pages,
+			    page_shift, &fmr->mr);
+	if (err)
+		return err;
+
+	mtt_seg = fmr->mr.mtt.first_seg * dev->caps.mtt_entry_sz;
+
+	fmr->mtts = mlx4_table_find(&priv->mr_table.mtt_table,
+				    fmr->mr.mtt.first_seg,
+				    &fmr->dma_handle);
+	if (!fmr->mtts) {
+		err = -ENOMEM;
+		goto err_free;
+	}
+
+	return 0;
+
+err_free:
+	mlx4_mr_free(dev, &fmr->mr);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_fmr_alloc);
+
+int mlx4_fmr_enable(struct mlx4_dev *dev, struct mlx4_fmr *fmr)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int err;
+
+	err = mlx4_mr_enable(dev, &fmr->mr);
+	if (err)
+		return err;
+
+	fmr->mpt = mlx4_table_find(&priv->mr_table.dmpt_table,
+				    key_to_hw_index(fmr->mr.key), NULL);
+	if (!fmr->mpt)
+		return -ENOMEM;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_fmr_enable);
+
+void mlx4_fmr_unmap(struct mlx4_dev *dev, struct mlx4_fmr *fmr,
+		    u32 *lkey, u32 *rkey)
+{
+	if (!fmr->maps)
+		return;
+
+	fmr->maps = 0;
+
+	*(u8 *) fmr->mpt = MLX4_MPT_STATUS_SW;
+}
+EXPORT_SYMBOL_GPL(mlx4_fmr_unmap);
+
+int mlx4_fmr_free(struct mlx4_dev *dev, struct mlx4_fmr *fmr)
+{
+	if (fmr->maps)
+		return -EBUSY;
+
+	fmr->mr.enabled = 0;
+	mlx4_mr_free(dev, &fmr->mr);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_fmr_free);
+
+int mlx4_SYNC_TPT(struct mlx4_dev *dev)
+{
+	return mlx4_cmd(dev, 0, 0, 0, MLX4_CMD_SYNC_TPT, 1000);
+}
+EXPORT_SYMBOL_GPL(mlx4_SYNC_TPT);

kernel/drivers/net/mlx4/pd.c

@@ -0,0 +1,108 @@
+/*
+ * Copyright (c) 2006, 2007 Cisco Systems, Inc.  All rights reserved.
+ * Copyright (c) 2005 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
+ * 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/errno.h>
+
+#include <asm/page.h>
+
+#include "mlx4.h"
+#include "icm.h"
+
+int mlx4_pd_alloc(struct mlx4_dev *dev, u32 *pdn)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	*pdn = mlx4_bitmap_alloc(&priv->pd_bitmap);
+	if (*pdn == -1)
+		return -ENOMEM;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_pd_alloc);
+
+void mlx4_pd_free(struct mlx4_dev *dev, u32 pdn)
+{
+	mlx4_bitmap_free(&mlx4_priv(dev)->pd_bitmap, pdn);
+}
+EXPORT_SYMBOL_GPL(mlx4_pd_free);
+
+int mlx4_init_pd_table(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	return mlx4_bitmap_init(&priv->pd_bitmap, dev->caps.num_pds,
+				(1 << 24) - 1, dev->caps.reserved_pds, 0);
+}
+
+void mlx4_cleanup_pd_table(struct mlx4_dev *dev)
+{
+	mlx4_bitmap_cleanup(&mlx4_priv(dev)->pd_bitmap);
+}
+
+
+int mlx4_uar_alloc(struct mlx4_dev *dev, struct mlx4_uar *uar)
+{
+	uar->index = mlx4_bitmap_alloc(&mlx4_priv(dev)->uar_table.bitmap);
+	if (uar->index == -1)
+		return -ENOMEM;
+
+	uar->pfn = (pci_resource_start(dev->pdev, 2) >> PAGE_SHIFT) + uar->index;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_uar_alloc);
+
+void mlx4_uar_free(struct mlx4_dev *dev, struct mlx4_uar *uar)
+{
+	mlx4_bitmap_free(&mlx4_priv(dev)->uar_table.bitmap, uar->index);
+}
+EXPORT_SYMBOL_GPL(mlx4_uar_free);
+
+int mlx4_init_uar_table(struct mlx4_dev *dev)
+{
+	if (dev->caps.num_uars <= 128) {
+		mlx4_err(dev, "Only %d UAR pages (need more than 128)\n",
+			 dev->caps.num_uars);
+		mlx4_err(dev, "Increase firmware log2_uar_bar_megabytes?\n");
+		return -ENODEV;
+	}
+
+	return mlx4_bitmap_init(&mlx4_priv(dev)->uar_table.bitmap,
+				dev->caps.num_uars, dev->caps.num_uars - 1,
+				max(128, dev->caps.reserved_uars), 0);
+}
+
+void mlx4_cleanup_uar_table(struct mlx4_dev *dev)
+{
+	mlx4_bitmap_cleanup(&mlx4_priv(dev)->uar_table.bitmap);
+}

kernel/drivers/net/mlx4/port.c

@@ -0,0 +1,317 @@
+/*
+ * Copyright (c) 2007 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
+ * 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/errno.h>
+#include <linux/if_ether.h>
+
+#include <linux/mlx4/cmd.h>
+
+#include "mlx4.h"
+
+#define MLX4_MAC_VALID		(1ull << 63)
+#define MLX4_MAC_MASK		0xffffffffffffULL
+
+#define MLX4_VLAN_VALID		(1u << 31)
+#define MLX4_VLAN_MASK		0xfff
+
+void mlx4_init_mac_table(struct mlx4_dev *dev, struct mlx4_mac_table *table)
+{
+	int i;
+
+	mutex_init(&table->mutex);
+	for (i = 0; i < MLX4_MAX_MAC_NUM; i++) {
+		table->entries[i] = 0;
+		table->refs[i]	 = 0;
+	}
+	table->max   = 1 << dev->caps.log_num_macs;
+	table->total = 0;
+}
+
+void mlx4_init_vlan_table(struct mlx4_dev *dev, struct mlx4_vlan_table *table)
+{
+	int i;
+
+	mutex_init(&table->mutex);
+	for (i = 0; i < MLX4_MAX_VLAN_NUM; i++) {
+		table->entries[i] = 0;
+		table->refs[i]	 = 0;
+	}
+	table->max   = 1 << dev->caps.log_num_vlans;
+	table->total = 0;
+}
+
+static int mlx4_set_port_mac_table(struct mlx4_dev *dev, u8 port,
+				   __be64 *entries)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	u32 in_mod;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	memcpy(mailbox->buf, entries, MLX4_MAC_TABLE_SIZE);
+
+	in_mod = MLX4_SET_PORT_MAC_TABLE << 8 | port;
+	err = mlx4_cmd(dev, mailbox->dma, in_mod, 1, MLX4_CMD_SET_PORT,
+		       MLX4_CMD_TIME_CLASS_B);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+int mlx4_register_mac(struct mlx4_dev *dev, u8 port, u64 mac, int *index)
+{
+	struct mlx4_mac_table *table = &mlx4_priv(dev)->port[port].mac_table;
+	int i, err = 0;
+	int free = -1;
+
+	mlx4_dbg(dev, "Registering MAC: 0x%llx\n", (unsigned long long) mac);
+	mutex_lock(&table->mutex);
+	for (i = 0; i < MLX4_MAX_MAC_NUM - 1; i++) {
+		if (free < 0 && !table->refs[i]) {
+			free = i;
+			continue;
+		}
+
+		if (mac == (MLX4_MAC_MASK & be64_to_cpu(table->entries[i]))) {
+			/* MAC already registered, increase refernce count */
+			*index = i;
+			++table->refs[i];
+			goto out;
+		}
+	}
+	mlx4_dbg(dev, "Free MAC index is %d\n", free);
+
+	if (table->total == table->max) {
+		/* No free mac entries */
+		err = -ENOSPC;
+		goto out;
+	}
+
+	/* Register new MAC */
+	table->refs[free] = 1;
+	table->entries[free] = cpu_to_be64(mac | MLX4_MAC_VALID);
+
+	err = mlx4_set_port_mac_table(dev, port, table->entries);
+	if (unlikely(err)) {
+		mlx4_err(dev, "Failed adding MAC: 0x%llx\n", (unsigned long long) mac);
+		table->refs[free] = 0;
+		table->entries[free] = 0;
+		goto out;
+	}
+
+	*index = free;
+	++table->total;
+out:
+	mutex_unlock(&table->mutex);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_register_mac);
+
+void mlx4_unregister_mac(struct mlx4_dev *dev, u8 port, int index)
+{
+	struct mlx4_mac_table *table = &mlx4_priv(dev)->port[port].mac_table;
+
+	mutex_lock(&table->mutex);
+	if (!table->refs[index]) {
+		mlx4_warn(dev, "No MAC entry for index %d\n", index);
+		goto out;
+	}
+	if (--table->refs[index]) {
+		mlx4_warn(dev, "Have more references for index %d,"
+			  "no need to modify MAC table\n", index);
+		goto out;
+	}
+	table->entries[index] = 0;
+	mlx4_set_port_mac_table(dev, port, table->entries);
+	--table->total;
+out:
+	mutex_unlock(&table->mutex);
+}
+EXPORT_SYMBOL_GPL(mlx4_unregister_mac);
+
+static int mlx4_set_port_vlan_table(struct mlx4_dev *dev, u8 port,
+				    __be32 *entries)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	u32 in_mod;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	memcpy(mailbox->buf, entries, MLX4_VLAN_TABLE_SIZE);
+	in_mod = MLX4_SET_PORT_VLAN_TABLE << 8 | port;
+	err = mlx4_cmd(dev, mailbox->dma, in_mod, 1, MLX4_CMD_SET_PORT,
+		       MLX4_CMD_TIME_CLASS_B);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+	return err;
+}
+
+int mlx4_register_vlan(struct mlx4_dev *dev, u8 port, u16 vlan, int *index)
+{
+	struct mlx4_vlan_table *table = &mlx4_priv(dev)->port[port].vlan_table;
+	int i, err = 0;
+	int free = -1;
+
+	mutex_lock(&table->mutex);
+	for (i = MLX4_VLAN_REGULAR; i < MLX4_MAX_VLAN_NUM; i++) {
+		if (free < 0 && (table->refs[i] == 0)) {
+			free = i;
+			continue;
+		}
+
+		if (table->refs[i] &&
+		    (vlan == (MLX4_VLAN_MASK &
+			      be32_to_cpu(table->entries[i])))) {
+			/* Vlan already registered, increase refernce count */
+			*index = i;
+			++table->refs[i];
+			goto out;
+		}
+	}
+
+	if (table->total == table->max) {
+		/* No free vlan entries */
+		err = -ENOSPC;
+		goto out;
+	}
+
+	/* Register new MAC */
+	table->refs[free] = 1;
+	table->entries[free] = cpu_to_be32(vlan | MLX4_VLAN_VALID);
+
+	err = mlx4_set_port_vlan_table(dev, port, table->entries);
+	if (unlikely(err)) {
+		mlx4_warn(dev, "Failed adding vlan: %u\n", vlan);
+		table->refs[free] = 0;
+		table->entries[free] = 0;
+		goto out;
+	}
+
+	*index = free;
+	++table->total;
+out:
+	mutex_unlock(&table->mutex);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_register_vlan);
+
+void mlx4_unregister_vlan(struct mlx4_dev *dev, u8 port, int index)
+{
+	struct mlx4_vlan_table *table = &mlx4_priv(dev)->port[port].vlan_table;
+
+	if (index < MLX4_VLAN_REGULAR) {
+		mlx4_warn(dev, "Trying to free special vlan index %d\n", index);
+		return;
+	}
+
+	mutex_lock(&table->mutex);
+	if (!table->refs[index]) {
+		mlx4_warn(dev, "No vlan entry for index %d\n", index);
+		goto out;
+	}
+	if (--table->refs[index]) {
+		mlx4_dbg(dev, "Have more references for index %d,"
+			 "no need to modify vlan table\n", index);
+		goto out;
+	}
+	table->entries[index] = 0;
+	mlx4_set_port_vlan_table(dev, port, table->entries);
+	--table->total;
+out:
+	mutex_unlock(&table->mutex);
+}
+EXPORT_SYMBOL_GPL(mlx4_unregister_vlan);
+
+int mlx4_get_port_ib_caps(struct mlx4_dev *dev, u8 port, __be32 *caps)
+{
+	struct mlx4_cmd_mailbox *inmailbox, *outmailbox;
+	u8 *inbuf, *outbuf;
+	int err;
+
+	inmailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(inmailbox))
+		return PTR_ERR(inmailbox);
+
+	outmailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(outmailbox)) {
+		mlx4_free_cmd_mailbox(dev, inmailbox);
+		return PTR_ERR(outmailbox);
+	}
+
+	inbuf = inmailbox->buf;
+	outbuf = outmailbox->buf;
+	memset(inbuf, 0, 256);
+	memset(outbuf, 0, 256);
+	inbuf[0] = 1;
+	inbuf[1] = 1;
+	inbuf[2] = 1;
+	inbuf[3] = 1;
+	*(__be16 *) (&inbuf[16]) = cpu_to_be16(0x0015);
+	*(__be32 *) (&inbuf[20]) = cpu_to_be32(port);
+
+	err = mlx4_cmd_box(dev, inmailbox->dma, outmailbox->dma, port, 3,
+			   MLX4_CMD_MAD_IFC, MLX4_CMD_TIME_CLASS_C);
+	if (!err)
+		*caps = *(__be32 *) (outbuf + 84);
+	mlx4_free_cmd_mailbox(dev, inmailbox);
+	mlx4_free_cmd_mailbox(dev, outmailbox);
+	return err;
+}
+
+int mlx4_SET_PORT(struct mlx4_dev *dev, u8 port)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	int err;
+
+	if (dev->caps.port_type[port] == MLX4_PORT_TYPE_ETH)
+		return 0;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	memset(mailbox->buf, 0, 256);
+
+	((__be32 *) mailbox->buf)[1] = dev->caps.ib_port_def_cap[port];
+	err = mlx4_cmd(dev, mailbox->dma, port, 0, MLX4_CMD_SET_PORT,
+		       MLX4_CMD_TIME_CLASS_B);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}

kernel/drivers/net/mlx4/profile.c

@@ -0,0 +1,238 @@
+/*
+ * Copyright (c) 2004, 2005 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2006, 2007 Cisco Systems, 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 "mlx4.h"
+#include "fw.h"
+
+enum {
+	MLX4_RES_QP,
+	MLX4_RES_RDMARC,
+	MLX4_RES_ALTC,
+	MLX4_RES_AUXC,
+	MLX4_RES_SRQ,
+	MLX4_RES_CQ,
+	MLX4_RES_EQ,
+	MLX4_RES_DMPT,
+	MLX4_RES_CMPT,
+	MLX4_RES_MTT,
+	MLX4_RES_MCG,
+	MLX4_RES_NUM
+};
+
+static const char *res_name[] = {
+	[MLX4_RES_QP]		= "QP",
+	[MLX4_RES_RDMARC]	= "RDMARC",
+	[MLX4_RES_ALTC]		= "ALTC",
+	[MLX4_RES_AUXC]		= "AUXC",
+	[MLX4_RES_SRQ]		= "SRQ",
+	[MLX4_RES_CQ]		= "CQ",
+	[MLX4_RES_EQ]		= "EQ",
+	[MLX4_RES_DMPT]		= "DMPT",
+	[MLX4_RES_CMPT]		= "CMPT",
+	[MLX4_RES_MTT]		= "MTT",
+	[MLX4_RES_MCG]		= "MCG",
+};
+
+u64 mlx4_make_profile(struct mlx4_dev *dev,
+		      struct mlx4_profile *request,
+		      struct mlx4_dev_cap *dev_cap,
+		      struct mlx4_init_hca_param *init_hca)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource {
+		u64 size;
+		u64 start;
+		int type;
+		int num;
+		int log_num;
+	};
+
+	u64 total_size = 0;
+	struct mlx4_resource *profile;
+	struct mlx4_resource tmp;
+	int i, j;
+
+	profile = kzalloc(MLX4_RES_NUM * sizeof *profile, GFP_KERNEL);
+	if (!profile)
+		return -ENOMEM;
+
+	profile[MLX4_RES_QP].size     = dev_cap->qpc_entry_sz;
+	profile[MLX4_RES_RDMARC].size = dev_cap->rdmarc_entry_sz;
+	profile[MLX4_RES_ALTC].size   = dev_cap->altc_entry_sz;
+	profile[MLX4_RES_AUXC].size   = dev_cap->aux_entry_sz;
+	profile[MLX4_RES_SRQ].size    = dev_cap->srq_entry_sz;
+	profile[MLX4_RES_CQ].size     = dev_cap->cqc_entry_sz;
+	profile[MLX4_RES_EQ].size     = dev_cap->eqc_entry_sz;
+	profile[MLX4_RES_DMPT].size   = dev_cap->dmpt_entry_sz;
+	profile[MLX4_RES_CMPT].size   = dev_cap->cmpt_entry_sz;
+	profile[MLX4_RES_MTT].size    = dev->caps.mtts_per_seg * dev_cap->mtt_entry_sz;
+	profile[MLX4_RES_MCG].size    = MLX4_MGM_ENTRY_SIZE;
+
+	profile[MLX4_RES_QP].num      = request->num_qp;
+	profile[MLX4_RES_RDMARC].num  = request->num_qp * request->rdmarc_per_qp;
+	profile[MLX4_RES_ALTC].num    = request->num_qp;
+	profile[MLX4_RES_AUXC].num    = request->num_qp;
+	profile[MLX4_RES_SRQ].num     = request->num_srq;
+	profile[MLX4_RES_CQ].num      = request->num_cq;
+	profile[MLX4_RES_EQ].num      = min_t(unsigned, dev_cap->max_eqs,
+					      dev_cap->reserved_eqs +
+					      num_possible_cpus() + 1);
+	profile[MLX4_RES_DMPT].num    = request->num_mpt;
+	profile[MLX4_RES_CMPT].num    = MLX4_NUM_CMPTS;
+	profile[MLX4_RES_MTT].num     = request->num_mtt;
+	profile[MLX4_RES_MCG].num     = request->num_mcg;
+
+	for (i = 0; i < MLX4_RES_NUM; ++i) {
+		profile[i].type     = i;
+		profile[i].num      = roundup_pow_of_two(profile[i].num);
+		profile[i].log_num  = ilog2(profile[i].num);
+		profile[i].size    *= profile[i].num;
+		profile[i].size     = max(profile[i].size, (u64) PAGE_SIZE);
+	}
+
+	/*
+	 * Sort the resources in decreasing order of size.  Since they
+	 * all have sizes that are powers of 2, we'll be able to keep
+	 * resources aligned to their size and pack them without gaps
+	 * using the sorted order.
+	 */
+	for (i = MLX4_RES_NUM; i > 0; --i)
+		for (j = 1; j < i; ++j) {
+			if (profile[j].size > profile[j - 1].size) {
+				tmp	       = profile[j];
+				profile[j]     = profile[j - 1];
+				profile[j - 1] = tmp;
+			}
+		}
+
+	for (i = 0; i < MLX4_RES_NUM; ++i) {
+		if (profile[i].size) {
+			profile[i].start = total_size;
+			total_size	+= profile[i].size;
+		}
+
+		if (total_size > dev_cap->max_icm_sz) {
+			mlx4_err(dev, "Profile requires 0x%llx bytes; "
+				  "won't fit in 0x%llx bytes of context memory.\n",
+				  (unsigned long long) total_size,
+				  (unsigned long long) dev_cap->max_icm_sz);
+			kfree(profile);
+			return -ENOMEM;
+		}
+
+		if (profile[i].size)
+			mlx4_dbg(dev, "  profile[%2d] (%6s): 2^%02d entries @ 0x%10llx, "
+				  "size 0x%10llx\n",
+				 i, res_name[profile[i].type], profile[i].log_num,
+				 (unsigned long long) profile[i].start,
+				 (unsigned long long) profile[i].size);
+	}
+
+	mlx4_dbg(dev, "HCA context memory: reserving %d KB\n",
+		 (int) (total_size >> 10));
+
+	for (i = 0; i < MLX4_RES_NUM; ++i) {
+		switch (profile[i].type) {
+		case MLX4_RES_QP:
+			dev->caps.num_qps     = profile[i].num;
+			init_hca->qpc_base    = profile[i].start;
+			init_hca->log_num_qps = profile[i].log_num;
+			break;
+		case MLX4_RES_RDMARC:
+			for (priv->qp_table.rdmarc_shift = 0;
+			     request->num_qp << priv->qp_table.rdmarc_shift < profile[i].num;
+			     ++priv->qp_table.rdmarc_shift)
+				; /* nothing */
+			dev->caps.max_qp_dest_rdma = 1 << priv->qp_table.rdmarc_shift;
+			priv->qp_table.rdmarc_base   = (u32) profile[i].start;
+			init_hca->rdmarc_base	     = profile[i].start;
+			init_hca->log_rd_per_qp	     = priv->qp_table.rdmarc_shift;
+			break;
+		case MLX4_RES_ALTC:
+			init_hca->altc_base = profile[i].start;
+			break;
+		case MLX4_RES_AUXC:
+			init_hca->auxc_base = profile[i].start;
+			break;
+		case MLX4_RES_SRQ:
+			dev->caps.num_srqs     = profile[i].num;
+			init_hca->srqc_base    = profile[i].start;
+			init_hca->log_num_srqs = profile[i].log_num;
+			break;
+		case MLX4_RES_CQ:
+			dev->caps.num_cqs     = profile[i].num;
+			init_hca->cqc_base    = profile[i].start;
+			init_hca->log_num_cqs = profile[i].log_num;
+			break;
+		case MLX4_RES_EQ:
+			dev->caps.num_eqs     = profile[i].num;
+			init_hca->eqc_base    = profile[i].start;
+			init_hca->log_num_eqs = profile[i].log_num;
+			break;
+		case MLX4_RES_DMPT:
+			dev->caps.num_mpts	= profile[i].num;
+			priv->mr_table.mpt_base = profile[i].start;
+			init_hca->dmpt_base	= profile[i].start;
+			init_hca->log_mpt_sz	= profile[i].log_num;
+			break;
+		case MLX4_RES_CMPT:
+			init_hca->cmpt_base	 = profile[i].start;
+			break;
+		case MLX4_RES_MTT:
+			dev->caps.num_mtt_segs	 = profile[i].num;
+			priv->mr_table.mtt_base	 = profile[i].start;
+			init_hca->mtt_base	 = profile[i].start;
+			break;
+		case MLX4_RES_MCG:
+			dev->caps.num_mgms	  = profile[i].num >> 1;
+			dev->caps.num_amgms	  = profile[i].num >> 1;
+			init_hca->mc_base	  = profile[i].start;
+			init_hca->log_mc_entry_sz = ilog2(MLX4_MGM_ENTRY_SIZE);
+			init_hca->log_mc_table_sz = profile[i].log_num;
+			init_hca->log_mc_hash_sz  = profile[i].log_num - 1;
+			break;
+		default:
+			break;
+		}
+	}
+
+	/*
+	 * PDs don't take any HCA memory, but we assign them as part
+	 * of the HCA profile anyway.
+	 */
+	dev->caps.num_pds = MLX4_NUM_PDS;
+
+	kfree(profile);
+	return total_size;
+}

kernel/drivers/net/mlx4/qp.c

@@ -0,0 +1,379 @@
+/*
+ * Copyright (c) 2004 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005, 2006, 2007 Cisco Systems, Inc. All rights reserved.
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2004 Voltaire, 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/mlx4/cmd.h>
+#include <linux/mlx4/qp.h>
+
+#include "mlx4.h"
+#include "icm.h"
+
+void mlx4_qp_event(struct mlx4_dev *dev, u32 qpn, int event_type)
+{
+	struct mlx4_qp_table *qp_table = &mlx4_priv(dev)->qp_table;
+	struct mlx4_qp *qp;
+
+	spin_lock(&qp_table->lock);
+
+	qp = __mlx4_qp_lookup(dev, qpn);
+	if (qp)
+		atomic_inc(&qp->refcount);
+
+	spin_unlock(&qp_table->lock);
+
+	if (!qp) {
+		mlx4_warn(dev, "Async event for bogus QP %08x\n", qpn);
+		return;
+	}
+
+	qp->event(qp, event_type);
+
+	if (atomic_dec_and_test(&qp->refcount))
+		complete(&qp->free);
+}
+
+int mlx4_qp_modify(struct mlx4_dev *dev, struct mlx4_mtt *mtt,
+		   enum mlx4_qp_state cur_state, enum mlx4_qp_state new_state,
+		   struct mlx4_qp_context *context, enum mlx4_qp_optpar optpar,
+		   int sqd_event, struct mlx4_qp *qp)
+{
+	static const u16 op[MLX4_QP_NUM_STATE][MLX4_QP_NUM_STATE] = {
+		[MLX4_QP_STATE_RST] = {
+			[MLX4_QP_STATE_RST]	= MLX4_CMD_2RST_QP,
+			[MLX4_QP_STATE_ERR]	= MLX4_CMD_2ERR_QP,
+			[MLX4_QP_STATE_INIT]	= MLX4_CMD_RST2INIT_QP,
+		},
+		[MLX4_QP_STATE_INIT]  = {
+			[MLX4_QP_STATE_RST]	= MLX4_CMD_2RST_QP,
+			[MLX4_QP_STATE_ERR]	= MLX4_CMD_2ERR_QP,
+			[MLX4_QP_STATE_INIT]	= MLX4_CMD_INIT2INIT_QP,
+			[MLX4_QP_STATE_RTR]	= MLX4_CMD_INIT2RTR_QP,
+		},
+		[MLX4_QP_STATE_RTR]   = {
+			[MLX4_QP_STATE_RST]	= MLX4_CMD_2RST_QP,
+			[MLX4_QP_STATE_ERR]	= MLX4_CMD_2ERR_QP,
+			[MLX4_QP_STATE_RTS]	= MLX4_CMD_RTR2RTS_QP,
+		},
+		[MLX4_QP_STATE_RTS]   = {
+			[MLX4_QP_STATE_RST]	= MLX4_CMD_2RST_QP,
+			[MLX4_QP_STATE_ERR]	= MLX4_CMD_2ERR_QP,
+			[MLX4_QP_STATE_RTS]	= MLX4_CMD_RTS2RTS_QP,
+			[MLX4_QP_STATE_SQD]	= MLX4_CMD_RTS2SQD_QP,
+		},
+		[MLX4_QP_STATE_SQD] = {
+			[MLX4_QP_STATE_RST]	= MLX4_CMD_2RST_QP,
+			[MLX4_QP_STATE_ERR]	= MLX4_CMD_2ERR_QP,
+			[MLX4_QP_STATE_RTS]	= MLX4_CMD_SQD2RTS_QP,
+			[MLX4_QP_STATE_SQD]	= MLX4_CMD_SQD2SQD_QP,
+		},
+		[MLX4_QP_STATE_SQER] = {
+			[MLX4_QP_STATE_RST]	= MLX4_CMD_2RST_QP,
+			[MLX4_QP_STATE_ERR]	= MLX4_CMD_2ERR_QP,
+			[MLX4_QP_STATE_RTS]	= MLX4_CMD_SQERR2RTS_QP,
+		},
+		[MLX4_QP_STATE_ERR] = {
+			[MLX4_QP_STATE_RST]	= MLX4_CMD_2RST_QP,
+			[MLX4_QP_STATE_ERR]	= MLX4_CMD_2ERR_QP,
+		}
+	};
+
+	struct mlx4_cmd_mailbox *mailbox;
+	int ret = 0;
+
+	if (cur_state >= MLX4_QP_NUM_STATE || new_state >= MLX4_QP_NUM_STATE ||
+	    !op[cur_state][new_state])
+		return -EINVAL;
+
+	if (op[cur_state][new_state] == MLX4_CMD_2RST_QP)
+		return mlx4_cmd(dev, 0, qp->qpn, 2,
+				MLX4_CMD_2RST_QP, MLX4_CMD_TIME_CLASS_A);
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	if (cur_state == MLX4_QP_STATE_RST && new_state == MLX4_QP_STATE_INIT) {
+		u64 mtt_addr = mlx4_mtt_addr(dev, mtt);
+		context->mtt_base_addr_h = mtt_addr >> 32;
+		context->mtt_base_addr_l = cpu_to_be32(mtt_addr & 0xffffffff);
+		context->log_page_size   = mtt->page_shift - MLX4_ICM_PAGE_SHIFT;
+	}
+
+	*(__be32 *) mailbox->buf = cpu_to_be32(optpar);
+	memcpy(mailbox->buf + 8, context, sizeof *context);
+
+	((struct mlx4_qp_context *) (mailbox->buf + 8))->local_qpn =
+		cpu_to_be32(qp->qpn);
+
+	ret = mlx4_cmd(dev, mailbox->dma, qp->qpn | (!!sqd_event << 31),
+		       new_state == MLX4_QP_STATE_RST ? 2 : 0,
+		       op[cur_state][new_state], MLX4_CMD_TIME_CLASS_C);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(mlx4_qp_modify);
+
+int mlx4_qp_reserve_range(struct mlx4_dev *dev, int cnt, int align, int *base)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_qp_table *qp_table = &priv->qp_table;
+	int qpn;
+
+	qpn = mlx4_bitmap_alloc_range(&qp_table->bitmap, cnt, align);
+	if (qpn == -1)
+		return -ENOMEM;
+
+	*base = qpn;
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_qp_reserve_range);
+
+void mlx4_qp_release_range(struct mlx4_dev *dev, int base_qpn, int cnt)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_qp_table *qp_table = &priv->qp_table;
+	if (base_qpn < dev->caps.sqp_start + 8)
+		return;
+
+	mlx4_bitmap_free_range(&qp_table->bitmap, base_qpn, cnt);
+}
+EXPORT_SYMBOL_GPL(mlx4_qp_release_range);
+
+int mlx4_qp_alloc(struct mlx4_dev *dev, int qpn, struct mlx4_qp *qp)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_qp_table *qp_table = &priv->qp_table;
+	int err;
+
+	if (!qpn)
+		return -EINVAL;
+
+	qp->qpn = qpn;
+
+	err = mlx4_table_get(dev, &qp_table->qp_table, qp->qpn);
+	if (err)
+		goto err_out;
+
+	err = mlx4_table_get(dev, &qp_table->auxc_table, qp->qpn);
+	if (err)
+		goto err_put_qp;
+
+	err = mlx4_table_get(dev, &qp_table->altc_table, qp->qpn);
+	if (err)
+		goto err_put_auxc;
+
+	err = mlx4_table_get(dev, &qp_table->rdmarc_table, qp->qpn);
+	if (err)
+		goto err_put_altc;
+
+	err = mlx4_table_get(dev, &qp_table->cmpt_table, qp->qpn);
+	if (err)
+		goto err_put_rdmarc;
+
+	spin_lock_irq(&qp_table->lock);
+	err = radix_tree_insert(&dev->qp_table_tree, qp->qpn & (dev->caps.num_qps - 1), qp);
+	spin_unlock_irq(&qp_table->lock);
+	if (err)
+		goto err_put_cmpt;
+
+	atomic_set(&qp->refcount, 1);
+	init_completion(&qp->free);
+
+	return 0;
+
+err_put_cmpt:
+	mlx4_table_put(dev, &qp_table->cmpt_table, qp->qpn);
+
+err_put_rdmarc:
+	mlx4_table_put(dev, &qp_table->rdmarc_table, qp->qpn);
+
+err_put_altc:
+	mlx4_table_put(dev, &qp_table->altc_table, qp->qpn);
+
+err_put_auxc:
+	mlx4_table_put(dev, &qp_table->auxc_table, qp->qpn);
+
+err_put_qp:
+	mlx4_table_put(dev, &qp_table->qp_table, qp->qpn);
+
+err_out:
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_qp_alloc);
+
+void mlx4_qp_remove(struct mlx4_dev *dev, struct mlx4_qp *qp)
+{
+	struct mlx4_qp_table *qp_table = &mlx4_priv(dev)->qp_table;
+	unsigned long flags;
+
+	spin_lock_irqsave(&qp_table->lock, flags);
+	radix_tree_delete(&dev->qp_table_tree, qp->qpn & (dev->caps.num_qps - 1));
+	spin_unlock_irqrestore(&qp_table->lock, flags);
+}
+EXPORT_SYMBOL_GPL(mlx4_qp_remove);
+
+void mlx4_qp_free(struct mlx4_dev *dev, struct mlx4_qp *qp)
+{
+	struct mlx4_qp_table *qp_table = &mlx4_priv(dev)->qp_table;
+
+	if (atomic_dec_and_test(&qp->refcount))
+		complete(&qp->free);
+	wait_for_completion(&qp->free);
+
+	mlx4_table_put(dev, &qp_table->cmpt_table, qp->qpn);
+	mlx4_table_put(dev, &qp_table->rdmarc_table, qp->qpn);
+	mlx4_table_put(dev, &qp_table->altc_table, qp->qpn);
+	mlx4_table_put(dev, &qp_table->auxc_table, qp->qpn);
+	mlx4_table_put(dev, &qp_table->qp_table, qp->qpn);
+}
+EXPORT_SYMBOL_GPL(mlx4_qp_free);
+
+static int mlx4_CONF_SPECIAL_QP(struct mlx4_dev *dev, u32 base_qpn)
+{
+	return mlx4_cmd(dev, 0, base_qpn, 0, MLX4_CMD_CONF_SPECIAL_QP,
+			MLX4_CMD_TIME_CLASS_B);
+}
+
+int mlx4_init_qp_table(struct mlx4_dev *dev)
+{
+	struct mlx4_qp_table *qp_table = &mlx4_priv(dev)->qp_table;
+	int err;
+	int reserved_from_top = 0;
+
+	spin_lock_init(&qp_table->lock);
+	INIT_RADIX_TREE(&dev->qp_table_tree, GFP_ATOMIC);
+
+	/*
+	 * We reserve 2 extra QPs per port for the special QPs.  The
+	 * block of special QPs must be aligned to a multiple of 8, so
+	 * round up.
+	 */
+	dev->caps.sqp_start =
+		ALIGN(dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW], 8);
+
+	{
+		int sort[MLX4_NUM_QP_REGION];
+		int i, j, tmp;
+		int last_base = dev->caps.num_qps;
+
+		for (i = 1; i < MLX4_NUM_QP_REGION; ++i)
+			sort[i] = i;
+
+		for (i = MLX4_NUM_QP_REGION; i > 0; --i) {
+			for (j = 2; j < i; ++j) {
+				if (dev->caps.reserved_qps_cnt[sort[j]] >
+				    dev->caps.reserved_qps_cnt[sort[j - 1]]) {
+					tmp             = sort[j];
+					sort[j]         = sort[j - 1];
+					sort[j - 1]     = tmp;
+				}
+			}
+		}
+
+		for (i = 1; i < MLX4_NUM_QP_REGION; ++i) {
+			last_base -= dev->caps.reserved_qps_cnt[sort[i]];
+			dev->caps.reserved_qps_base[sort[i]] = last_base;
+			reserved_from_top +=
+				dev->caps.reserved_qps_cnt[sort[i]];
+		}
+
+	}
+
+	err = mlx4_bitmap_init(&qp_table->bitmap, dev->caps.num_qps,
+			       (1 << 23) - 1, dev->caps.sqp_start + 8,
+			       reserved_from_top);
+	if (err)
+		return err;
+
+	return mlx4_CONF_SPECIAL_QP(dev, dev->caps.sqp_start);
+}
+
+void mlx4_cleanup_qp_table(struct mlx4_dev *dev)
+{
+	mlx4_CONF_SPECIAL_QP(dev, 0);
+	mlx4_bitmap_cleanup(&mlx4_priv(dev)->qp_table.bitmap);
+}
+
+int mlx4_qp_query(struct mlx4_dev *dev, struct mlx4_qp *qp,
+		  struct mlx4_qp_context *context)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	err = mlx4_cmd_box(dev, 0, mailbox->dma, qp->qpn, 0,
+			   MLX4_CMD_QUERY_QP, MLX4_CMD_TIME_CLASS_A);
+	if (!err)
+		memcpy(context, mailbox->buf + 8, sizeof *context);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_qp_query);
+
+int mlx4_qp_to_ready(struct mlx4_dev *dev, struct mlx4_mtt *mtt,
+		     struct mlx4_qp_context *context,
+		     struct mlx4_qp *qp, enum mlx4_qp_state *qp_state)
+{
+	int err;
+	int i;
+	enum mlx4_qp_state states[] = {
+		MLX4_QP_STATE_RST,
+		MLX4_QP_STATE_INIT,
+		MLX4_QP_STATE_RTR,
+		MLX4_QP_STATE_RTS
+	};
+
+	for (i = 0; i < ARRAY_SIZE(states) - 1; i++) {
+		context->flags &= cpu_to_be32(~(0xf << 28));
+		context->flags |= cpu_to_be32(states[i + 1] << 28);
+		err = mlx4_qp_modify(dev, mtt, states[i], states[i + 1],
+				     context, 0, 0, qp);
+		if (err) {
+			mlx4_err(dev, "Failed to bring QP to state: "
+				 "%d with error: %d\n",
+				 states[i + 1], err);
+			return err;
+		}
+
+		*qp_state = states[i + 1];
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_qp_to_ready);

kernel/drivers/net/mlx4/reset.c

@@ -0,0 +1,185 @@
+/*
+ * Copyright (c) 2006, 2007 Cisco Systems, Inc.  All rights reserved.
+ * Copyright (c) 2007, 2008 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
+ * 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/errno.h>
+#include <linux/pci.h>
+#include <linux/delay.h>
+#include <linux/slab.h>
+#include <linux/jiffies.h>
+
+#include "mlx4.h"
+
+int mlx4_reset(struct mlx4_dev *dev)
+{
+	void __iomem *reset;
+	u32 *hca_header = NULL;
+	int pcie_cap;
+	u16 devctl;
+	u16 linkctl;
+	u16 vendor;
+	unsigned long end;
+	u32 sem;
+	int i;
+	int err = 0;
+
+#define MLX4_RESET_BASE		0xf0000
+#define MLX4_RESET_SIZE		  0x400
+#define MLX4_SEM_OFFSET		  0x3fc
+#define MLX4_RESET_OFFSET	   0x10
+#define MLX4_RESET_VALUE	swab32(1)
+
+#define MLX4_SEM_TIMEOUT_JIFFIES	(10 * HZ)
+#define MLX4_RESET_TIMEOUT_JIFFIES	(2 * HZ)
+
+	/*
+	 * Reset the chip.  This is somewhat ugly because we have to
+	 * save off the PCI header before reset and then restore it
+	 * after the chip reboots.  We skip config space offsets 22
+	 * and 23 since those have a special meaning.
+	 */
+
+	/* Do we need to save off the full 4K PCI Express header?? */
+	hca_header = kmalloc(256, GFP_KERNEL);
+	if (!hca_header) {
+		err = -ENOMEM;
+		mlx4_err(dev, "Couldn't allocate memory to save HCA "
+			  "PCI header, aborting.\n");
+		goto out;
+	}
+
+	pcie_cap = pci_find_capability(dev->pdev, PCI_CAP_ID_EXP);
+
+	for (i = 0; i < 64; ++i) {
+		if (i == 22 || i == 23)
+			continue;
+		if (pci_read_config_dword(dev->pdev, i * 4, hca_header + i)) {
+			err = -ENODEV;
+			mlx4_err(dev, "Couldn't save HCA "
+				  "PCI header, aborting.\n");
+			goto out;
+		}
+	}
+
+	reset = ioremap(pci_resource_start(dev->pdev, 0) + MLX4_RESET_BASE,
+			MLX4_RESET_SIZE);
+	if (!reset) {
+		err = -ENOMEM;
+		mlx4_err(dev, "Couldn't map HCA reset register, aborting.\n");
+		goto out;
+	}
+
+	/* grab HW semaphore to lock out flash updates */
+	end = jiffies + MLX4_SEM_TIMEOUT_JIFFIES;
+	do {
+		sem = readl(reset + MLX4_SEM_OFFSET);
+		if (!sem)
+			break;
+
+		msleep(1);
+	} while (time_before(jiffies, end));
+
+	if (sem) {
+		mlx4_err(dev, "Failed to obtain HW semaphore, aborting\n");
+		err = -EAGAIN;
+		iounmap(reset);
+		goto out;
+	}
+
+	/* actually hit reset */
+	writel(MLX4_RESET_VALUE, reset + MLX4_RESET_OFFSET);
+	iounmap(reset);
+
+	/* Docs say to wait one second before accessing device */
+	msleep(1000);
+
+	end = jiffies + MLX4_RESET_TIMEOUT_JIFFIES;
+	do {
+		if (!pci_read_config_word(dev->pdev, PCI_VENDOR_ID, &vendor) &&
+		    vendor != 0xffff)
+			break;
+
+		msleep(1);
+	} while (time_before(jiffies, end));
+
+	if (vendor == 0xffff) {
+		err = -ENODEV;
+		mlx4_err(dev, "PCI device did not come back after reset, "
+			  "aborting.\n");
+		goto out;
+	}
+
+	/* Now restore the PCI headers */
+	if (pcie_cap) {
+		devctl = hca_header[(pcie_cap + PCI_EXP_DEVCTL) / 4];
+		if (pci_write_config_word(dev->pdev, pcie_cap + PCI_EXP_DEVCTL,
+					   devctl)) {
+			err = -ENODEV;
+			mlx4_err(dev, "Couldn't restore HCA PCI Express "
+				 "Device Control register, aborting.\n");
+			goto out;
+		}
+		linkctl = hca_header[(pcie_cap + PCI_EXP_LNKCTL) / 4];
+		if (pci_write_config_word(dev->pdev, pcie_cap + PCI_EXP_LNKCTL,
+					   linkctl)) {
+			err = -ENODEV;
+			mlx4_err(dev, "Couldn't restore HCA PCI Express "
+				 "Link control register, aborting.\n");
+			goto out;
+		}
+	}
+
+	for (i = 0; i < 16; ++i) {
+		if (i * 4 == PCI_COMMAND)
+			continue;
+
+		if (pci_write_config_dword(dev->pdev, i * 4, hca_header[i])) {
+			err = -ENODEV;
+			mlx4_err(dev, "Couldn't restore HCA reg %x, "
+				  "aborting.\n", i);
+			goto out;
+		}
+	}
+
+	if (pci_write_config_dword(dev->pdev, PCI_COMMAND,
+				   hca_header[PCI_COMMAND / 4])) {
+		err = -ENODEV;
+		mlx4_err(dev, "Couldn't restore HCA COMMAND, "
+			  "aborting.\n");
+		goto out;
+	}
+
+out:
+	kfree(hca_header);
+
+	return err;
+}

kernel/drivers/net/mlx4/sense.c

@@ -0,0 +1,156 @@
+/*
+ * Copyright (c) 2007 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
+ * 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/errno.h>
+#include <linux/if_ether.h>
+
+#include <linux/mlx4/cmd.h>
+
+#include "mlx4.h"
+
+static int mlx4_SENSE_PORT(struct mlx4_dev *dev, int port,
+			   enum mlx4_port_type *type)
+{
+	u64 out_param;
+	int err = 0;
+
+	err = mlx4_cmd_imm(dev, 0, &out_param, port, 0,
+			   MLX4_CMD_SENSE_PORT, MLX4_CMD_TIME_CLASS_B);
+	if (err) {
+		mlx4_err(dev, "Sense command failed for port: %d\n", port);
+		return err;
+	}
+
+	if (out_param > 2) {
+		mlx4_err(dev, "Sense returned illegal value: 0x%llx\n", out_param);
+		return EINVAL;
+	}
+
+	*type = out_param;
+	return 0;
+}
+
+void mlx4_do_sense_ports(struct mlx4_dev *dev,
+			 enum mlx4_port_type *stype,
+			 enum mlx4_port_type *defaults)
+{
+	struct mlx4_sense *sense = &mlx4_priv(dev)->sense;
+	int err;
+	int i;
+
+	for (i = 1; i <= dev->caps.num_ports; i++) {
+		stype[i - 1] = 0;
+		if (sense->do_sense_port[i] && sense->sense_allowed[i] &&
+		    dev->caps.possible_type[i] == MLX4_PORT_TYPE_AUTO) {
+			err = mlx4_SENSE_PORT(dev, i, &stype[i - 1]);
+			if (err)
+				stype[i - 1] = defaults[i - 1];
+		} else
+			stype[i - 1] = defaults[i - 1];
+	}
+
+	/*
+	 * Adjust port configuration:
+	 * If port 1 sensed nothing and port 2 is IB, set both as IB
+	 * If port 2 sensed nothing and port 1 is Eth, set both as Eth
+	 */
+	if (stype[0] == MLX4_PORT_TYPE_ETH) {
+		for (i = 1; i < dev->caps.num_ports; i++)
+			stype[i] = stype[i] ? stype[i] : MLX4_PORT_TYPE_ETH;
+	}
+	if (stype[dev->caps.num_ports - 1] == MLX4_PORT_TYPE_IB) {
+		for (i = 0; i < dev->caps.num_ports - 1; i++)
+			stype[i] = stype[i] ? stype[i] : MLX4_PORT_TYPE_IB;
+	}
+
+	/*
+	 * If sensed nothing, remain in current configuration.
+	 */
+	for (i = 0; i < dev->caps.num_ports; i++)
+		stype[i] = stype[i] ? stype[i] : defaults[i];
+
+}
+
+static void mlx4_sense_port(struct work_struct *work)
+{
+	struct delayed_work *delay = to_delayed_work(work);
+	struct mlx4_sense *sense = container_of(delay, struct mlx4_sense,
+						sense_poll);
+	struct mlx4_dev *dev = sense->dev;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	enum mlx4_port_type stype[MLX4_MAX_PORTS];
+
+	mutex_lock(&priv->port_mutex);
+	mlx4_do_sense_ports(dev, stype, &dev->caps.port_type[1]);
+
+	if (mlx4_check_port_params(dev, stype))
+		goto sense_again;
+
+	if (mlx4_change_port_types(dev, stype))
+		mlx4_err(dev, "Failed to change port_types\n");
+
+sense_again:
+	mutex_unlock(&priv->port_mutex);
+	queue_delayed_work(mlx4_wq , &sense->sense_poll,
+			   round_jiffies_relative(MLX4_SENSE_RANGE));
+}
+
+void mlx4_start_sense(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_sense *sense = &priv->sense;
+
+	if (!(dev->caps.flags & MLX4_DEV_CAP_FLAG_DPDP))
+		return;
+
+	queue_delayed_work(mlx4_wq , &sense->sense_poll,
+			   round_jiffies_relative(MLX4_SENSE_RANGE));
+}
+
+void mlx4_stop_sense(struct mlx4_dev *dev)
+{
+	cancel_delayed_work_sync(&mlx4_priv(dev)->sense.sense_poll);
+}
+
+void  mlx4_sense_init(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_sense *sense = &priv->sense;
+	int port;
+
+	sense->dev = dev;
+	for (port = 1; port <= dev->caps.num_ports; port++)
+		sense->do_sense_port[port] = 1;
+
+	INIT_DELAYED_WORK_DEFERRABLE(&sense->sense_poll, mlx4_sense_port);
+}

kernel/drivers/net/mlx4/srq.c

@@ -0,0 +1,256 @@
+/*
+ * Copyright (c) 2006, 2007 Cisco Systems, Inc. All rights reserved.
+ * Copyright (c) 2007, 2008 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
+ * 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/mlx4/cmd.h>
+
+#include "mlx4.h"
+#include "icm.h"
+
+struct mlx4_srq_context {
+	__be32			state_logsize_srqn;
+	u8			logstride;
+	u8			reserved1[3];
+	u8			pg_offset;
+	u8			reserved2[3];
+	u32			reserved3;
+	u8			log_page_size;
+	u8			reserved4[2];
+	u8			mtt_base_addr_h;
+	__be32			mtt_base_addr_l;
+	__be32			pd;
+	__be16			limit_watermark;
+	__be16			wqe_cnt;
+	u16			reserved5;
+	__be16			wqe_counter;
+	u32			reserved6;
+	__be64			db_rec_addr;
+};
+
+void mlx4_srq_event(struct mlx4_dev *dev, u32 srqn, int event_type)
+{
+	struct mlx4_srq_table *srq_table = &mlx4_priv(dev)->srq_table;
+	struct mlx4_srq *srq;
+
+	spin_lock(&srq_table->lock);
+
+	srq = radix_tree_lookup(&srq_table->tree, srqn & (dev->caps.num_srqs - 1));
+	if (srq)
+		atomic_inc(&srq->refcount);
+
+	spin_unlock(&srq_table->lock);
+
+	if (!srq) {
+		mlx4_warn(dev, "Async event for bogus SRQ %08x\n", srqn);
+		return;
+	}
+
+	srq->event(srq, event_type);
+
+	if (atomic_dec_and_test(&srq->refcount))
+		complete(&srq->free);
+}
+
+static int mlx4_SW2HW_SRQ(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *mailbox,
+			  int srq_num)
+{
+	return mlx4_cmd(dev, mailbox->dma, srq_num, 0, MLX4_CMD_SW2HW_SRQ,
+			MLX4_CMD_TIME_CLASS_A);
+}
+
+static int mlx4_HW2SW_SRQ(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *mailbox,
+			  int srq_num)
+{
+	return mlx4_cmd_box(dev, 0, mailbox ? mailbox->dma : 0, srq_num,
+			    mailbox ? 0 : 1, MLX4_CMD_HW2SW_SRQ,
+			    MLX4_CMD_TIME_CLASS_A);
+}
+
+static int mlx4_ARM_SRQ(struct mlx4_dev *dev, int srq_num, int limit_watermark)
+{
+	return mlx4_cmd(dev, limit_watermark, srq_num, 0, MLX4_CMD_ARM_SRQ,
+			MLX4_CMD_TIME_CLASS_B);
+}
+
+static int mlx4_QUERY_SRQ(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *mailbox,
+			  int srq_num)
+{
+	return mlx4_cmd_box(dev, 0, mailbox->dma, srq_num, 0, MLX4_CMD_QUERY_SRQ,
+			    MLX4_CMD_TIME_CLASS_A);
+}
+
+int mlx4_srq_alloc(struct mlx4_dev *dev, u32 pdn, struct mlx4_mtt *mtt,
+		   u64 db_rec, struct mlx4_srq *srq)
+{
+	struct mlx4_srq_table *srq_table = &mlx4_priv(dev)->srq_table;
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_srq_context *srq_context;
+	u64 mtt_addr;
+	int err;
+
+	srq->srqn = mlx4_bitmap_alloc(&srq_table->bitmap);
+	if (srq->srqn == -1)
+		return -ENOMEM;
+
+	err = mlx4_table_get(dev, &srq_table->table, srq->srqn);
+	if (err)
+		goto err_out;
+
+	err = mlx4_table_get(dev, &srq_table->cmpt_table, srq->srqn);
+	if (err)
+		goto err_put;
+
+	spin_lock_irq(&srq_table->lock);
+	err = radix_tree_insert(&srq_table->tree, srq->srqn, srq);
+	spin_unlock_irq(&srq_table->lock);
+	if (err)
+		goto err_cmpt_put;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox)) {
+		err = PTR_ERR(mailbox);
+		goto err_radix;
+	}
+
+	srq_context = mailbox->buf;
+	memset(srq_context, 0, sizeof *srq_context);
+
+	srq_context->state_logsize_srqn = cpu_to_be32((ilog2(srq->max) << 24) |
+						      srq->srqn);
+	srq_context->logstride          = srq->wqe_shift - 4;
+	srq_context->log_page_size      = mtt->page_shift - MLX4_ICM_PAGE_SHIFT;
+
+	mtt_addr = mlx4_mtt_addr(dev, mtt);
+	srq_context->mtt_base_addr_h    = mtt_addr >> 32;
+	srq_context->mtt_base_addr_l    = cpu_to_be32(mtt_addr & 0xffffffff);
+	srq_context->pd			= cpu_to_be32(pdn);
+	srq_context->db_rec_addr        = cpu_to_be64(db_rec);
+
+	err = mlx4_SW2HW_SRQ(dev, mailbox, srq->srqn);
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	if (err)
+		goto err_radix;
+
+	atomic_set(&srq->refcount, 1);
+	init_completion(&srq->free);
+
+	return 0;
+
+err_radix:
+	spin_lock_irq(&srq_table->lock);
+	radix_tree_delete(&srq_table->tree, srq->srqn);
+	spin_unlock_irq(&srq_table->lock);
+
+err_cmpt_put:
+	mlx4_table_put(dev, &srq_table->cmpt_table, srq->srqn);
+
+err_put:
+	mlx4_table_put(dev, &srq_table->table, srq->srqn);
+
+err_out:
+	mlx4_bitmap_free(&srq_table->bitmap, srq->srqn);
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_srq_alloc);
+
+void mlx4_srq_free(struct mlx4_dev *dev, struct mlx4_srq *srq)
+{
+	struct mlx4_srq_table *srq_table = &mlx4_priv(dev)->srq_table;
+	int err;
+
+	err = mlx4_HW2SW_SRQ(dev, NULL, srq->srqn);
+	if (err)
+		mlx4_warn(dev, "HW2SW_SRQ failed (%d) for SRQN %06x\n", err, srq->srqn);
+
+	spin_lock_irq(&srq_table->lock);
+	radix_tree_delete(&srq_table->tree, srq->srqn);
+	spin_unlock_irq(&srq_table->lock);
+
+	if (atomic_dec_and_test(&srq->refcount))
+		complete(&srq->free);
+	wait_for_completion(&srq->free);
+
+	mlx4_table_put(dev, &srq_table->table, srq->srqn);
+	mlx4_bitmap_free(&srq_table->bitmap, srq->srqn);
+}
+EXPORT_SYMBOL_GPL(mlx4_srq_free);
+
+int mlx4_srq_arm(struct mlx4_dev *dev, struct mlx4_srq *srq, int limit_watermark)
+{
+	return mlx4_ARM_SRQ(dev, srq->srqn, limit_watermark);
+}
+EXPORT_SYMBOL_GPL(mlx4_srq_arm);
+
+int mlx4_srq_query(struct mlx4_dev *dev, struct mlx4_srq *srq, int *limit_watermark)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_srq_context *srq_context;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	srq_context = mailbox->buf;
+
+	err = mlx4_QUERY_SRQ(dev, mailbox, srq->srqn);
+	if (err)
+		goto err_out;
+	*limit_watermark = be16_to_cpu(srq_context->limit_watermark);
+
+err_out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_srq_query);
+
+int mlx4_init_srq_table(struct mlx4_dev *dev)
+{
+	struct mlx4_srq_table *srq_table = &mlx4_priv(dev)->srq_table;
+	int err;
+
+	spin_lock_init(&srq_table->lock);
+	INIT_RADIX_TREE(&srq_table->tree, GFP_ATOMIC);
+
+	err = mlx4_bitmap_init(&srq_table->bitmap, dev->caps.num_srqs,
+			       dev->caps.num_srqs - 1, dev->caps.reserved_srqs, 0);
+	if (err)
+		return err;
+
+	return 0;
+}
+
+void mlx4_cleanup_srq_table(struct mlx4_dev *dev)
+{
+	mlx4_bitmap_cleanup(&mlx4_priv(dev)->srq_table.bitmap);
+}