xenbus_xs.c

来自「linux 内核源代码」· C语言 代码 · 共 862 行 · 第 1/2 页

/******************************************************************************
 * xenbus_xs.c
 *
 * This is the kernel equivalent of the "xs" library.  We don't need everything
 * and we use xenbus_comms for communication.
 *
 * Copyright (C) 2005 Rusty Russell, IBM Corporation
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License version 2
 * as published by the Free Software Foundation; or, when distributed
 * separately from the Linux kernel or incorporated into other
 * software packages, subject to the following license:
 *
 * Permission is hereby granted, free of charge, to any person obtaining a copy
 * of this source file (the "Software"), to deal in the Software without
 * restriction, including without limitation the rights to use, copy, modify,
 * merge, publish, distribute, sublicense, and/or sell copies of the Software,
 * and to permit persons to whom the Software is furnished to do so, subject to
 * the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in
 * all copies or substantial portions of the Software.
 *
 * 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/unistd.h>
#include <linux/errno.h>
#include <linux/types.h>
#include <linux/uio.h>
#include <linux/kernel.h>
#include <linux/string.h>
#include <linux/err.h>
#include <linux/slab.h>
#include <linux/fcntl.h>
#include <linux/kthread.h>
#include <linux/rwsem.h>
#include <linux/module.h>
#include <linux/mutex.h>
#include <xen/xenbus.h>
#include "xenbus_comms.h"

struct xs_stored_msg {
	struct list_head list;

	struct xsd_sockmsg hdr;

	union {
		/* Queued replies. */
		struct {
			char *body;
		} reply;

		/* Queued watch events. */
		struct {
			struct xenbus_watch *handle;
			char **vec;
			unsigned int vec_size;
		} watch;
	} u;
};

struct xs_handle {
	/* A list of replies. Currently only one will ever be outstanding. */
	struct list_head reply_list;
	spinlock_t reply_lock;
	wait_queue_head_t reply_waitq;

	/*
	 * Mutex ordering: transaction_mutex -> watch_mutex -> request_mutex.
	 * response_mutex is never taken simultaneously with the other three.
	 */

	/* One request at a time. */
	struct mutex request_mutex;

	/* Protect xenbus reader thread against save/restore. */
	struct mutex response_mutex;

	/* Protect transactions against save/restore. */
	struct rw_semaphore transaction_mutex;

	/* Protect watch (de)register against save/restore. */
	struct rw_semaphore watch_mutex;
};

static struct xs_handle xs_state;

/* List of registered watches, and a lock to protect it. */
static LIST_HEAD(watches);
static DEFINE_SPINLOCK(watches_lock);

/* List of pending watch callback events, and a lock to protect it. */
static LIST_HEAD(watch_events);
static DEFINE_SPINLOCK(watch_events_lock);

/*
 * Details of the xenwatch callback kernel thread. The thread waits on the
 * watch_events_waitq for work to do (queued on watch_events list). When it
 * wakes up it acquires the xenwatch_mutex before reading the list and
 * carrying out work.
 */
static pid_t xenwatch_pid;
static DEFINE_MUTEX(xenwatch_mutex);
static DECLARE_WAIT_QUEUE_HEAD(watch_events_waitq);

static int get_error(const char *errorstring)
{
	unsigned int i;

	for (i = 0; strcmp(errorstring, xsd_errors[i].errstring) != 0; i++) {
		if (i == ARRAY_SIZE(xsd_errors) - 1) {
			printk(KERN_WARNING
			       "XENBUS xen store gave: unknown error %s",
			       errorstring);
			return EINVAL;
		}
	}
	return xsd_errors[i].errnum;
}

static void *read_reply(enum xsd_sockmsg_type *type, unsigned int *len)
{
	struct xs_stored_msg *msg;
	char *body;

	spin_lock(&xs_state.reply_lock);

	while (list_empty(&xs_state.reply_list)) {
		spin_unlock(&xs_state.reply_lock);
		/* XXX FIXME: Avoid synchronous wait for response here. */
		wait_event(xs_state.reply_waitq,
			   !list_empty(&xs_state.reply_list));
		spin_lock(&xs_state.reply_lock);
	}

	msg = list_entry(xs_state.reply_list.next,
			 struct xs_stored_msg, list);
	list_del(&msg->list);

	spin_unlock(&xs_state.reply_lock);

	*type = msg->hdr.type;
	if (len)
		*len = msg->hdr.len;
	body = msg->u.reply.body;

	kfree(msg);

	return body;
}

void *xenbus_dev_request_and_reply(struct xsd_sockmsg *msg)
{
	void *ret;
	struct xsd_sockmsg req_msg = *msg;
	int err;

	if (req_msg.type == XS_TRANSACTION_START)
		down_read(&xs_state.transaction_mutex);

	mutex_lock(&xs_state.request_mutex);

	err = xb_write(msg, sizeof(*msg) + msg->len);
	if (err) {
		msg->type = XS_ERROR;
		ret = ERR_PTR(err);
	} else
		ret = read_reply(&msg->type, &msg->len);

	mutex_unlock(&xs_state.request_mutex);

	if ((msg->type == XS_TRANSACTION_END) ||
	    ((req_msg.type == XS_TRANSACTION_START) &&
	     (msg->type == XS_ERROR)))
		up_read(&xs_state.transaction_mutex);

	return ret;
}

/* Send message to xs, get kmalloc'ed reply.  ERR_PTR() on error. */
static void *xs_talkv(struct xenbus_transaction t,
		      enum xsd_sockmsg_type type,
		      const struct kvec *iovec,
		      unsigned int num_vecs,
		      unsigned int *len)
{
	struct xsd_sockmsg msg;
	void *ret = NULL;
	unsigned int i;
	int err;

	msg.tx_id = t.id;
	msg.req_id = 0;
	msg.type = type;
	msg.len = 0;
	for (i = 0; i < num_vecs; i++)
		msg.len += iovec[i].iov_len;

	mutex_lock(&xs_state.request_mutex);

	err = xb_write(&msg, sizeof(msg));
	if (err) {
		mutex_unlock(&xs_state.request_mutex);
		return ERR_PTR(err);
	}

	for (i = 0; i < num_vecs; i++) {
		err = xb_write(iovec[i].iov_base, iovec[i].iov_len);
		if (err) {
			mutex_unlock(&xs_state.request_mutex);
			return ERR_PTR(err);
		}
	}

	ret = read_reply(&msg.type, len);

	mutex_unlock(&xs_state.request_mutex);

	if (IS_ERR(ret))
		return ret;

	if (msg.type == XS_ERROR) {
		err = get_error(ret);
		kfree(ret);
		return ERR_PTR(-err);
	}

	if (msg.type != type) {
		if (printk_ratelimit())
			printk(KERN_WARNING
			       "XENBUS unexpected type [%d], expected [%d]\n",
			       msg.type, type);
		kfree(ret);
		return ERR_PTR(-EINVAL);
	}
	return ret;
}

/* Simplified version of xs_talkv: single message. */
static void *xs_single(struct xenbus_transaction t,
		       enum xsd_sockmsg_type type,
		       const char *string,
		       unsigned int *len)
{
	struct kvec iovec;

	iovec.iov_base = (void *)string;
	iovec.iov_len = strlen(string) + 1;
	return xs_talkv(t, type, &iovec, 1, len);
}

/* Many commands only need an ack, don't care what it says. */
static int xs_error(char *reply)
{
	if (IS_ERR(reply))
		return PTR_ERR(reply);
	kfree(reply);
	return 0;
}

static unsigned int count_strings(const char *strings, unsigned int len)
{
	unsigned int num;
	const char *p;

	for (p = strings, num = 0; p < strings + len; p += strlen(p) + 1)
		num++;

	return num;
}

/* Return the path to dir with /name appended. Buffer must be kfree()'ed. */
static char *join(const char *dir, const char *name)
{
	char *buffer;

	if (strlen(name) == 0)
		buffer = kasprintf(GFP_KERNEL, "%s", dir);
	else
		buffer = kasprintf(GFP_KERNEL, "%s/%s", dir, name);
	return (!buffer) ? ERR_PTR(-ENOMEM) : buffer;
}

static char **split(char *strings, unsigned int len, unsigned int *num)
{
	char *p, **ret;

	/* Count the strings. */
	*num = count_strings(strings, len);

	/* Transfer to one big alloc for easy freeing. */
	ret = kmalloc(*num * sizeof(char *) + len, GFP_KERNEL);
	if (!ret) {
		kfree(strings);
		return ERR_PTR(-ENOMEM);
	}
	memcpy(&ret[*num], strings, len);
	kfree(strings);

	strings = (char *)&ret[*num];
	for (p = strings, *num = 0; p < strings + len; p += strlen(p) + 1)
		ret[(*num)++] = p;

	return ret;
}

char **xenbus_directory(struct xenbus_transaction t,
			const char *dir, const char *node, unsigned int *num)
{
	char *strings, *path;
	unsigned int len;

	path = join(dir, node);
	if (IS_ERR(path))
		return (char **)path;

	strings = xs_single(t, XS_DIRECTORY, path, &len);
	kfree(path);
	if (IS_ERR(strings))
		return (char **)strings;

	return split(strings, len, num);
}
EXPORT_SYMBOL_GPL(xenbus_directory);

/* Check if a path exists. Return 1 if it does. */
int xenbus_exists(struct xenbus_transaction t,
		  const char *dir, const char *node)
{
	char **d;
	int dir_n;

	d = xenbus_directory(t, dir, node, &dir_n);
	if (IS_ERR(d))
		return 0;
	kfree(d);
	return 1;
}
EXPORT_SYMBOL_GPL(xenbus_exists);

/* Get the value of a single file.
 * Returns a kmalloced value: call free() on it after use.
 * len indicates length in bytes.
 */
void *xenbus_read(struct xenbus_transaction t,
		  const char *dir, const char *node, unsigned int *len)
{
	char *path;
	void *ret;

	path = join(dir, node);
	if (IS_ERR(path))
		return (void *)path;

	ret = xs_single(t, XS_READ, path, len);
	kfree(path);
	return ret;
}
EXPORT_SYMBOL_GPL(xenbus_read);

/* Write the value of a single file.
 * Returns -err on failure.
 */
int xenbus_write(struct xenbus_transaction t,
		 const char *dir, const char *node, const char *string)
{
	const char *path;
	struct kvec iovec[2];
	int ret;

	path = join(dir, node);
	if (IS_ERR(path))
		return PTR_ERR(path);

	iovec[0].iov_base = (void *)path;
	iovec[0].iov_len = strlen(path) + 1;
	iovec[1].iov_base = (void *)string;
	iovec[1].iov_len = strlen(string);

	ret = xs_error(xs_talkv(t, XS_WRITE, iovec, ARRAY_SIZE(iovec), NULL));
	kfree(path);
	return ret;
}
EXPORT_SYMBOL_GPL(xenbus_write);

/* Create a new directory. */
int xenbus_mkdir(struct xenbus_transaction t,
		 const char *dir, const char *node)
{
	char *path;
	int ret;

	path = join(dir, node);
	if (IS_ERR(path))
		return PTR_ERR(path);

	ret = xs_error(xs_single(t, XS_MKDIR, path, NULL));
	kfree(path);
	return ret;
}
EXPORT_SYMBOL_GPL(xenbus_mkdir);

/* Destroy a file or directory (directories must be empty). */
int xenbus_rm(struct xenbus_transaction t, const char *dir, const char *node)
{
	char *path;
	int ret;

	path = join(dir, node);
	if (IS_ERR(path))
		return PTR_ERR(path);

	ret = xs_error(xs_single(t, XS_RM, path, NULL));
	kfree(path);
	return ret;
}
EXPORT_SYMBOL_GPL(xenbus_rm);

/* Start a transaction: changes by others will not be seen during this
 * transaction, and changes will not be visible to others until end.
 */
int xenbus_transaction_start(struct xenbus_transaction *t)
{