user_mad.c

来自「LINUX 2.6.17.4的源码」· C语言 代码 · 共 1,071 行 · 第 1/2 页

/*
 * Copyright (c) 2004 Topspin Communications.  All rights reserved.
 * Copyright (c) 2005 Voltaire, Inc. All rights reserved. 
 * Copyright (c) 2005 Sun Microsystems, 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.
 *
 * $Id: user_mad.c 5596 2006-03-03 01:00:07Z sean.hefty $
 */

#include <linux/module.h>
#include <linux/init.h>
#include <linux/device.h>
#include <linux/err.h>
#include <linux/fs.h>
#include <linux/cdev.h>
#include <linux/pci.h>
#include <linux/dma-mapping.h>
#include <linux/poll.h>
#include <linux/rwsem.h>
#include <linux/kref.h>

#include <asm/uaccess.h>
#include <asm/semaphore.h>

#include <rdma/ib_mad.h>
#include <rdma/ib_user_mad.h>

MODULE_AUTHOR("Roland Dreier");
MODULE_DESCRIPTION("InfiniBand userspace MAD packet access");
MODULE_LICENSE("Dual BSD/GPL");

enum {
	IB_UMAD_MAX_PORTS  = 64,
	IB_UMAD_MAX_AGENTS = 32,

	IB_UMAD_MAJOR      = 231,
	IB_UMAD_MINOR_BASE = 0
};

/*
 * Our lifetime rules for these structs are the following: each time a
 * device special file is opened, we look up the corresponding struct
 * ib_umad_port by minor in the umad_port[] table while holding the
 * port_lock.  If this lookup succeeds, we take a reference on the
 * ib_umad_port's struct ib_umad_device while still holding the
 * port_lock; if the lookup fails, we fail the open().  We drop these
 * references in the corresponding close().
 *
 * In addition to references coming from open character devices, there
 * is one more reference to each ib_umad_device representing the
 * module's reference taken when allocating the ib_umad_device in
 * ib_umad_add_one().
 *
 * When destroying an ib_umad_device, we clear all of its
 * ib_umad_ports from umad_port[] while holding port_lock before
 * dropping the module's reference to the ib_umad_device.  This is
 * always safe because any open() calls will either succeed and obtain
 * a reference before we clear the umad_port[] entries, or fail after
 * we clear the umad_port[] entries.
 */

struct ib_umad_port {
	struct cdev           *dev;
	struct class_device   *class_dev;

	struct cdev           *sm_dev;
	struct class_device   *sm_class_dev;
	struct semaphore       sm_sem;

	struct rw_semaphore    mutex;
	struct list_head       file_list;

	struct ib_device      *ib_dev;
	struct ib_umad_device *umad_dev;
	int                    dev_num;
	u8                     port_num;
};

struct ib_umad_device {
	int                  start_port, end_port;
	struct kref          ref;
	struct ib_umad_port  port[0];
};

struct ib_umad_file {
	struct ib_umad_port    *port;
	struct list_head	recv_list;
	struct list_head	port_list;
	spinlock_t		recv_lock;
	wait_queue_head_t	recv_wait;
	struct ib_mad_agent    *agent[IB_UMAD_MAX_AGENTS];
	int			agents_dead;
};

struct ib_umad_packet {
	struct ib_mad_send_buf *msg;
	struct ib_mad_recv_wc  *recv_wc;
	struct list_head   list;
	int		   length;
	struct ib_user_mad mad;
};

static struct class *umad_class;

static const dev_t base_dev = MKDEV(IB_UMAD_MAJOR, IB_UMAD_MINOR_BASE);

static DEFINE_SPINLOCK(port_lock);
static struct ib_umad_port *umad_port[IB_UMAD_MAX_PORTS];
static DECLARE_BITMAP(dev_map, IB_UMAD_MAX_PORTS * 2);

static void ib_umad_add_one(struct ib_device *device);
static void ib_umad_remove_one(struct ib_device *device);

static void ib_umad_release_dev(struct kref *ref)
{
	struct ib_umad_device *dev =
		container_of(ref, struct ib_umad_device, ref);

	kfree(dev);
}

/* caller must hold port->mutex at least for reading */
static struct ib_mad_agent *__get_agent(struct ib_umad_file *file, int id)
{
	return file->agents_dead ? NULL : file->agent[id];
}

static int queue_packet(struct ib_umad_file *file,
			struct ib_mad_agent *agent,
			struct ib_umad_packet *packet)
{
	int ret = 1;

	down_read(&file->port->mutex);

	for (packet->mad.hdr.id = 0;
	     packet->mad.hdr.id < IB_UMAD_MAX_AGENTS;
	     packet->mad.hdr.id++)
		if (agent == __get_agent(file, packet->mad.hdr.id)) {
			spin_lock_irq(&file->recv_lock);
			list_add_tail(&packet->list, &file->recv_list);
			spin_unlock_irq(&file->recv_lock);
			wake_up_interruptible(&file->recv_wait);
			ret = 0;
			break;
		}

	up_read(&file->port->mutex);

	return ret;
}

static void send_handler(struct ib_mad_agent *agent,
			 struct ib_mad_send_wc *send_wc)
{
	struct ib_umad_file *file = agent->context;
	struct ib_umad_packet *packet = send_wc->send_buf->context[0];

	ib_destroy_ah(packet->msg->ah);
	ib_free_send_mad(packet->msg);

	if (send_wc->status == IB_WC_RESP_TIMEOUT_ERR) {
		packet->length = IB_MGMT_MAD_HDR;
		packet->mad.hdr.status = ETIMEDOUT;
		if (!queue_packet(file, agent, packet))
			return;
	}
	kfree(packet);
}

static void recv_handler(struct ib_mad_agent *agent,
			 struct ib_mad_recv_wc *mad_recv_wc)
{
	struct ib_umad_file *file = agent->context;
	struct ib_umad_packet *packet;

	if (mad_recv_wc->wc->status != IB_WC_SUCCESS)
		goto err1;

	packet = kzalloc(sizeof *packet, GFP_KERNEL);
	if (!packet)
		goto err1;

	packet->length = mad_recv_wc->mad_len;
	packet->recv_wc = mad_recv_wc;

	packet->mad.hdr.status    = 0;
	packet->mad.hdr.length    = sizeof (struct ib_user_mad) +
				    mad_recv_wc->mad_len;
	packet->mad.hdr.qpn 	  = cpu_to_be32(mad_recv_wc->wc->src_qp);
	packet->mad.hdr.lid 	  = cpu_to_be16(mad_recv_wc->wc->slid);
	packet->mad.hdr.sl  	  = mad_recv_wc->wc->sl;
	packet->mad.hdr.path_bits = mad_recv_wc->wc->dlid_path_bits;
	packet->mad.hdr.grh_present = !!(mad_recv_wc->wc->wc_flags & IB_WC_GRH);
	if (packet->mad.hdr.grh_present) {
		/* XXX parse GRH */
		packet->mad.hdr.gid_index 	= 0;
		packet->mad.hdr.hop_limit 	= 0;
		packet->mad.hdr.traffic_class	= 0;
		memset(packet->mad.hdr.gid, 0, 16);
		packet->mad.hdr.flow_label	= 0;
	}

	if (queue_packet(file, agent, packet))
		goto err2;
	return;

err2:
	kfree(packet);
err1:
	ib_free_recv_mad(mad_recv_wc);
}

static ssize_t copy_recv_mad(char __user *buf, struct ib_umad_packet *packet,
			     size_t count)
{
	struct ib_mad_recv_buf *recv_buf;
	int left, seg_payload, offset, max_seg_payload;

	/* We need enough room to copy the first (or only) MAD segment. */
	recv_buf = &packet->recv_wc->recv_buf;
	if ((packet->length <= sizeof (*recv_buf->mad) &&
	     count < sizeof (packet->mad) + packet->length) ||
	    (packet->length > sizeof (*recv_buf->mad) &&
	     count < sizeof (packet->mad) + sizeof (*recv_buf->mad)))
		return -EINVAL;

	if (copy_to_user(buf, &packet->mad, sizeof (packet->mad)))
		return -EFAULT;

	buf += sizeof (packet->mad);
	seg_payload = min_t(int, packet->length, sizeof (*recv_buf->mad));
	if (copy_to_user(buf, recv_buf->mad, seg_payload))
		return -EFAULT;

	if (seg_payload < packet->length) {
		/*
		 * Multipacket RMPP MAD message. Copy remainder of message.
		 * Note that last segment may have a shorter payload.
		 */
		if (count < sizeof (packet->mad) + packet->length) {
			/*
			 * The buffer is too small, return the first RMPP segment,
			 * which includes the RMPP message length.
			 */
			return -ENOSPC;
		}
		offset = ib_get_mad_data_offset(recv_buf->mad->mad_hdr.mgmt_class);
		max_seg_payload = sizeof (struct ib_mad) - offset;

		for (left = packet->length - seg_payload, buf += seg_payload;
		     left; left -= seg_payload, buf += seg_payload) {
			recv_buf = container_of(recv_buf->list.next,
						struct ib_mad_recv_buf, list);
			seg_payload = min(left, max_seg_payload);
			if (copy_to_user(buf, ((void *) recv_buf->mad) + offset,
					 seg_payload))
				return -EFAULT;
		}
	}
	return sizeof (packet->mad) + packet->length;
}

static ssize_t copy_send_mad(char __user *buf, struct ib_umad_packet *packet,
			     size_t count)
{
	ssize_t size = sizeof (packet->mad) + packet->length;

	if (count < size)
		return -EINVAL;

	if (copy_to_user(buf, &packet->mad, size))
		return -EFAULT;

	return size;
}

static ssize_t ib_umad_read(struct file *filp, char __user *buf,
			    size_t count, loff_t *pos)
{
	struct ib_umad_file *file = filp->private_data;
	struct ib_umad_packet *packet;
	ssize_t ret;

	if (count < sizeof (struct ib_user_mad))
		return -EINVAL;

	spin_lock_irq(&file->recv_lock);

	while (list_empty(&file->recv_list)) {
		spin_unlock_irq(&file->recv_lock);

		if (filp->f_flags & O_NONBLOCK)
			return -EAGAIN;

		if (wait_event_interruptible(file->recv_wait,
					     !list_empty(&file->recv_list)))
			return -ERESTARTSYS;

		spin_lock_irq(&file->recv_lock);
	}

	packet = list_entry(file->recv_list.next, struct ib_umad_packet, list);
	list_del(&packet->list);

	spin_unlock_irq(&file->recv_lock);

	if (packet->recv_wc)
		ret = copy_recv_mad(buf, packet, count);
	else
		ret = copy_send_mad(buf, packet, count);

	if (ret < 0) {
		/* Requeue packet */
		spin_lock_irq(&file->recv_lock);
		list_add(&packet->list, &file->recv_list);
		spin_unlock_irq(&file->recv_lock);
	} else {
		if (packet->recv_wc)
			ib_free_recv_mad(packet->recv_wc);
		kfree(packet);
	}
	return ret;
}

static int copy_rmpp_mad(struct ib_mad_send_buf *msg, const char __user *buf)
{
	int left, seg;

	/* Copy class specific header */
	if ((msg->hdr_len > IB_MGMT_RMPP_HDR) &&
	    copy_from_user(msg->mad + IB_MGMT_RMPP_HDR, buf + IB_MGMT_RMPP_HDR,
			   msg->hdr_len - IB_MGMT_RMPP_HDR))
		return -EFAULT;

	/* All headers are in place.  Copy data segments. */
	for (seg = 1, left = msg->data_len, buf += msg->hdr_len; left > 0;
	     seg++, left -= msg->seg_size, buf += msg->seg_size) {
		if (copy_from_user(ib_get_rmpp_segment(msg, seg), buf,
				   min(left, msg->seg_size)))
			return -EFAULT;
	}
	return 0;
}

static ssize_t ib_umad_write(struct file *filp, const char __user *buf,
			     size_t count, loff_t *pos)
{
	struct ib_umad_file *file = filp->private_data;
	struct ib_umad_packet *packet;
	struct ib_mad_agent *agent;
	struct ib_ah_attr ah_attr;
	struct ib_ah *ah;
	struct ib_rmpp_mad *rmpp_mad;
	u8 method;
	__be64 *tid;
	int ret, data_len, hdr_len, copy_offset, rmpp_active;

	if (count < sizeof (struct ib_user_mad) + IB_MGMT_RMPP_HDR)
		return -EINVAL;

	packet = kzalloc(sizeof *packet + IB_MGMT_RMPP_HDR, GFP_KERNEL);
	if (!packet)
		return -ENOMEM;

	if (copy_from_user(&packet->mad, buf,
			    sizeof (struct ib_user_mad) + IB_MGMT_RMPP_HDR)) {
		ret = -EFAULT;
		goto err;
	}

	if (packet->mad.hdr.id < 0 ||
	    packet->mad.hdr.id >= IB_UMAD_MAX_AGENTS) {
		ret = -EINVAL;
		goto err;
	}

	down_read(&file->port->mutex);

	agent = __get_agent(file, packet->mad.hdr.id);
	if (!agent) {
		ret = -EINVAL;
		goto err_up;
	}

	memset(&ah_attr, 0, sizeof ah_attr);
	ah_attr.dlid          = be16_to_cpu(packet->mad.hdr.lid);
	ah_attr.sl            = packet->mad.hdr.sl;
	ah_attr.src_path_bits = packet->mad.hdr.path_bits;
	ah_attr.port_num      = file->port->port_num;
	if (packet->mad.hdr.grh_present) {
		ah_attr.ah_flags = IB_AH_GRH;
		memcpy(ah_attr.grh.dgid.raw, packet->mad.hdr.gid, 16);
		ah_attr.grh.flow_label 	   = be32_to_cpu(packet->mad.hdr.flow_label);
		ah_attr.grh.hop_limit  	   = packet->mad.hdr.hop_limit;
		ah_attr.grh.traffic_class  = packet->mad.hdr.traffic_class;
	}

	ah = ib_create_ah(agent->qp->pd, &ah_attr);
	if (IS_ERR(ah)) {
		ret = PTR_ERR(ah);
		goto err_up;
	}

	rmpp_mad = (struct ib_rmpp_mad *) packet->mad.data;
	hdr_len = ib_get_mad_data_offset(rmpp_mad->mad_hdr.mgmt_class);
	if (!ib_is_mad_class_rmpp(rmpp_mad->mad_hdr.mgmt_class)) {
		copy_offset = IB_MGMT_MAD_HDR;
		rmpp_active = 0;
	} else {
		copy_offset = IB_MGMT_RMPP_HDR;
		rmpp_active = ib_get_rmpp_flags(&rmpp_mad->rmpp_hdr) &
			      IB_MGMT_RMPP_FLAG_ACTIVE;
	}

	data_len = count - sizeof (struct ib_user_mad) - hdr_len;
	packet->msg = ib_create_send_mad(agent,
					 be32_to_cpu(packet->mad.hdr.qpn),
					 0, rmpp_active, hdr_len,
					 data_len, GFP_KERNEL);
	if (IS_ERR(packet->msg)) {
		ret = PTR_ERR(packet->msg);
		goto err_ah;
	}

	packet->msg->ah 	= ah;
	packet->msg->timeout_ms = packet->mad.hdr.timeout_ms;
	packet->msg->retries 	= packet->mad.hdr.retries;
	packet->msg->context[0] = packet;

	/* Copy MAD header.  Any RMPP header is already in place. */
	memcpy(packet->msg->mad, packet->mad.data, IB_MGMT_MAD_HDR);
	buf += sizeof (struct ib_user_mad);

	if (!rmpp_active) {
		if (copy_from_user(packet->msg->mad + copy_offset,
				   buf + copy_offset,
				   hdr_len + data_len - copy_offset)) {
			ret = -EFAULT;
			goto err_msg;
		}
	} else {
		ret = copy_rmpp_mad(packet->msg, buf);
		if (ret)
			goto err_msg;
	}

	/*
	 * If userspace is generating a request that will generate a
	 * response, we need to make sure the high-order part of the
	 * transaction ID matches the agent being used to send the
	 * MAD.
	 */
	method = ((struct ib_mad_hdr *) packet->msg->mad)->method;

	if (!(method & IB_MGMT_METHOD_RESP)       &&
	    method != IB_MGMT_METHOD_TRAP_REPRESS &&
	    method != IB_MGMT_METHOD_SEND) {
		tid = &((struct ib_mad_hdr *) packet->msg->mad)->tid;
		*tid = cpu_to_be64(((u64) agent->hi_tid) << 32 |
				   (be64_to_cpup(tid) & 0xffffffff));
	}

	ret = ib_post_send_mad(packet->msg, NULL);
	if (ret)
		goto err_msg;

	up_read(&file->port->mutex);
	return count;

err_msg:
	ib_free_send_mad(packet->msg);
err_ah:
	ib_destroy_ah(ah);
err_up:
	up_read(&file->port->mutex);
err:
	kfree(packet);
	return ret;
}

static unsigned int ib_umad_poll(struct file *filp, struct poll_table_struct *wait)
{
	struct ib_umad_file *file = filp->private_data;

	/* we will always be able to post a MAD send */
	unsigned int mask = POLLOUT | POLLWRNORM;

	poll_wait(filp, &file->recv_wait, wait);

	if (!list_empty(&file->recv_list))
		mask |= POLLIN | POLLRDNORM;

	return mask;
}

static int ib_umad_reg_agent(struct ib_umad_file *file, unsigned long arg)
{
	struct ib_user_mad_reg_req ureq;
	struct ib_mad_reg_req req;
	struct ib_mad_agent *agent;
	int agent_id;
	int ret;

	down_write(&file->port->mutex);

	if (!file->port->ib_dev) {
		ret = -EPIPE;
		goto out;