initial_commit

6 files changed, 1211 insertions, 0 deletions

diff --git a/drivers/xen/xenfs/Makefile b/drivers/xen/xenfs/Makefile
new file mode 100644
index 00000000..4fde9440
--- /dev/null
+++ b/drivers/xen/xenfs/Makefile
@@ -0,0 +1,4 @@
+obj-$(CONFIG_XENFS) += xenfs.o
+
+xenfs-y			  = super.o xenbus.o privcmd.o
+xenfs-$(CONFIG_XEN_DOM0) += xenstored.o
diff --git a/drivers/xen/xenfs/privcmd.c b/drivers/xen/xenfs/privcmd.c
new file mode 100644
index 00000000..dbd3b16f
--- /dev/null
+++ b/drivers/xen/xenfs/privcmd.c
@@ -0,0 +1,400 @@
+/******************************************************************************
+ * privcmd.c
+ *
+ * Interface to privileged domain-0 commands.
+ *
+ * Copyright (c) 2002-2004, K A Fraser, B Dragovic
+ */
+
+#include <linux/kernel.h>
+#include <linux/sched.h>
+#include <linux/slab.h>
+#include <linux/string.h>
+#include <linux/errno.h>
+#include <linux/mm.h>
+#include <linux/mman.h>
+#include <linux/uaccess.h>
+#include <linux/swap.h>
+#include <linux/highmem.h>
+#include <linux/pagemap.h>
+#include <linux/seq_file.h>
+
+#include <asm/pgalloc.h>
+#include <asm/pgtable.h>
+#include <asm/tlb.h>
+#include <asm/xen/hypervisor.h>
+#include <asm/xen/hypercall.h>
+
+#include <xen/xen.h>
+#include <xen/privcmd.h>
+#include <xen/interface/xen.h>
+#include <xen/features.h>
+#include <xen/page.h>
+#include <xen/xen-ops.h>
+
+#ifndef HAVE_ARCH_PRIVCMD_MMAP
+static int privcmd_enforce_singleshot_mapping(struct vm_area_struct *vma);
+#endif
+
+static long privcmd_ioctl_hypercall(void __user *udata)
+{
+	struct privcmd_hypercall hypercall;
+	long ret;
+
+	if (copy_from_user(&hypercall, udata, sizeof(hypercall)))
+		return -EFAULT;
+
+	ret = privcmd_call(hypercall.op,
+			   hypercall.arg[0], hypercall.arg[1],
+			   hypercall.arg[2], hypercall.arg[3],
+			   hypercall.arg[4]);
+
+	return ret;
+}
+
+static void free_page_list(struct list_head *pages)
+{
+	struct page *p, *n;
+
+	list_for_each_entry_safe(p, n, pages, lru)
+		__free_page(p);
+
+	INIT_LIST_HEAD(pages);
+}
+
+/*
+ * Given an array of items in userspace, return a list of pages
+ * containing the data.  If copying fails, either because of memory
+ * allocation failure or a problem reading user memory, return an
+ * error code; its up to the caller to dispose of any partial list.
+ */
+static int gather_array(struct list_head *pagelist,
+			unsigned nelem, size_t size,
+			void __user *data)
+{
+	unsigned pageidx;
+	void *pagedata;
+	int ret;
+
+	if (size > PAGE_SIZE)
+		return 0;
+
+	pageidx = PAGE_SIZE;
+	pagedata = NULL;	/* quiet, gcc */
+	while (nelem--) {
+		if (pageidx > PAGE_SIZE-size) {
+			struct page *page = alloc_page(GFP_KERNEL);
+
+			ret = -ENOMEM;
+			if (page == NULL)
+				goto fail;
+
+			pagedata = page_address(page);
+
+			list_add_tail(&page->lru, pagelist);
+			pageidx = 0;
+		}
+
+		ret = -EFAULT;
+		if (copy_from_user(pagedata + pageidx, data, size))
+			goto fail;
+
+		data += size;
+		pageidx += size;
+	}
+
+	ret = 0;
+
+fail:
+	return ret;
+}
+
+/*
+ * Call function "fn" on each element of the array fragmented
+ * over a list of pages.
+ */
+static int traverse_pages(unsigned nelem, size_t size,
+			  struct list_head *pos,
+			  int (*fn)(void *data, void *state),
+			  void *state)
+{
+	void *pagedata;
+	unsigned pageidx;
+	int ret = 0;
+
+	BUG_ON(size > PAGE_SIZE);
+
+	pageidx = PAGE_SIZE;
+	pagedata = NULL;	/* hush, gcc */
+
+	while (nelem--) {
+		if (pageidx > PAGE_SIZE-size) {
+			struct page *page;
+			pos = pos->next;
+			page = list_entry(pos, struct page, lru);
+			pagedata = page_address(page);
+			pageidx = 0;
+		}
+
+		ret = (*fn)(pagedata + pageidx, state);
+		if (ret)
+			break;
+		pageidx += size;
+	}
+
+	return ret;
+}
+
+struct mmap_mfn_state {
+	unsigned long va;
+	struct vm_area_struct *vma;
+	domid_t domain;
+};
+
+static int mmap_mfn_range(void *data, void *state)
+{
+	struct privcmd_mmap_entry *msg = data;
+	struct mmap_mfn_state *st = state;
+	struct vm_area_struct *vma = st->vma;
+	int rc;
+
+	/* Do not allow range to wrap the address space. */
+	if ((msg->npages > (LONG_MAX >> PAGE_SHIFT)) ||
+	    ((unsigned long)(msg->npages << PAGE_SHIFT) >= -st->va))
+		return -EINVAL;
+
+	/* Range chunks must be contiguous in va space. */
+	if ((msg->va != st->va) ||
+	    ((msg->va+(msg->npages<<PAGE_SHIFT)) > vma->vm_end))
+		return -EINVAL;
+
+	rc = xen_remap_domain_mfn_range(vma,
+					msg->va & PAGE_MASK,
+					msg->mfn, msg->npages,
+					vma->vm_page_prot,
+					st->domain);
+	if (rc < 0)
+		return rc;
+
+	st->va += msg->npages << PAGE_SHIFT;
+
+	return 0;
+}
+
+static long privcmd_ioctl_mmap(void __user *udata)
+{
+	struct privcmd_mmap mmapcmd;
+	struct mm_struct *mm = current->mm;
+	struct vm_area_struct *vma;
+	int rc;
+	LIST_HEAD(pagelist);
+	struct mmap_mfn_state state;
+
+	if (!xen_initial_domain())
+		return -EPERM;
+
+	if (copy_from_user(&mmapcmd, udata, sizeof(mmapcmd)))
+		return -EFAULT;
+
+	rc = gather_array(&pagelist,
+			  mmapcmd.num, sizeof(struct privcmd_mmap_entry),
+			  mmapcmd.entry);
+
+	if (rc || list_empty(&pagelist))
+		goto out;
+
+	down_write(&mm->mmap_sem);
+
+	{
+		struct page *page = list_first_entry(&pagelist,
+						     struct page, lru);
+		struct privcmd_mmap_entry *msg = page_address(page);
+
+		vma = find_vma(mm, msg->va);
+		rc = -EINVAL;
+
+		if (!vma || (msg->va != vma->vm_start) ||
+		    !privcmd_enforce_singleshot_mapping(vma))
+			goto out_up;
+	}
+
+	state.va = vma->vm_start;
+	state.vma = vma;
+	state.domain = mmapcmd.dom;
+
+	rc = traverse_pages(mmapcmd.num, sizeof(struct privcmd_mmap_entry),
+			    &pagelist,
+			    mmap_mfn_range, &state);
+
+
+out_up:
+	up_write(&mm->mmap_sem);
+
+out:
+	free_page_list(&pagelist);
+
+	return rc;
+}
+
+struct mmap_batch_state {
+	domid_t domain;
+	unsigned long va;
+	struct vm_area_struct *vma;
+	int err;
+
+	xen_pfn_t __user *user;
+};
+
+static int mmap_batch_fn(void *data, void *state)
+{
+	xen_pfn_t *mfnp = data;
+	struct mmap_batch_state *st = state;
+
+	if (xen_remap_domain_mfn_range(st->vma, st->va & PAGE_MASK, *mfnp, 1,
+				       st->vma->vm_page_prot, st->domain) < 0) {
+		*mfnp |= 0xf0000000U;
+		st->err++;
+	}
+	st->va += PAGE_SIZE;
+
+	return 0;
+}
+
+static int mmap_return_errors(void *data, void *state)
+{
+	xen_pfn_t *mfnp = data;
+	struct mmap_batch_state *st = state;
+
+	return put_user(*mfnp, st->user++);
+}
+
+static struct vm_operations_struct privcmd_vm_ops;
+
+static long privcmd_ioctl_mmap_batch(void __user *udata)
+{
+	int ret;
+	struct privcmd_mmapbatch m;
+	struct mm_struct *mm = current->mm;
+	struct vm_area_struct *vma;
+	unsigned long nr_pages;
+	LIST_HEAD(pagelist);
+	struct mmap_batch_state state;
+
+	if (!xen_initial_domain())
+		return -EPERM;
+
+	if (copy_from_user(&m, udata, sizeof(m)))
+		return -EFAULT;
+
+	nr_pages = m.num;
+	if ((m.num <= 0) || (nr_pages > (LONG_MAX >> PAGE_SHIFT)))
+		return -EINVAL;
+
+	ret = gather_array(&pagelist, m.num, sizeof(xen_pfn_t),
+			   m.arr);
+
+	if (ret || list_empty(&pagelist))
+		goto out;
+
+	down_write(&mm->mmap_sem);
+
+	vma = find_vma(mm, m.addr);
+	ret = -EINVAL;
+	if (!vma ||
+	    vma->vm_ops != &privcmd_vm_ops ||
+	    (m.addr != vma->vm_start) ||
+	    ((m.addr + (nr_pages << PAGE_SHIFT)) != vma->vm_end) ||
+	    !privcmd_enforce_singleshot_mapping(vma)) {
+		up_write(&mm->mmap_sem);
+		goto out;
+	}
+
+	state.domain = m.dom;
+	state.vma = vma;
+	state.va = m.addr;
+	state.err = 0;
+
+	ret = traverse_pages(m.num, sizeof(xen_pfn_t),
+			     &pagelist, mmap_batch_fn, &state);
+
+	up_write(&mm->mmap_sem);
+
+	if (state.err > 0) {
+		state.user = m.arr;
+		ret = traverse_pages(m.num, sizeof(xen_pfn_t),
+			       &pagelist,
+			       mmap_return_errors, &state);
+	}
+
+out:
+	free_page_list(&pagelist);
+
+	return ret;
+}
+
+static long privcmd_ioctl(struct file *file,
+			  unsigned int cmd, unsigned long data)
+{
+	int ret = -ENOSYS;
+	void __user *udata = (void __user *) data;
+
+	switch (cmd) {
+	case IOCTL_PRIVCMD_HYPERCALL:
+		ret = privcmd_ioctl_hypercall(udata);
+		break;
+
+	case IOCTL_PRIVCMD_MMAP:
+		ret = privcmd_ioctl_mmap(udata);
+		break;
+
+	case IOCTL_PRIVCMD_MMAPBATCH:
+		ret = privcmd_ioctl_mmap_batch(udata);
+		break;
+
+	default:
+		ret = -EINVAL;
+		break;
+	}
+
+	return ret;
+}
+
+#ifndef HAVE_ARCH_PRIVCMD_MMAP
+static int privcmd_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
+{
+	printk(KERN_DEBUG "privcmd_fault: vma=%p %lx-%lx, pgoff=%lx, uv=%p\n",
+	       vma, vma->vm_start, vma->vm_end,
+	       vmf->pgoff, vmf->virtual_address);
+
+	return VM_FAULT_SIGBUS;
+}
+
+static struct vm_operations_struct privcmd_vm_ops = {
+	.fault = privcmd_fault
+};
+
+static int privcmd_mmap(struct file *file, struct vm_area_struct *vma)
+{
+	/* Unsupported for auto-translate guests. */
+	if (xen_feature(XENFEAT_auto_translated_physmap))
+		return -ENOSYS;
+
+	/* DONTCOPY is essential for Xen because copy_page_range doesn't know
+	 * how to recreate these mappings */
+	vma->vm_flags |= VM_RESERVED | VM_IO | VM_DONTCOPY | VM_PFNMAP;
+	vma->vm_ops = &privcmd_vm_ops;
+	vma->vm_private_data = NULL;
+
+	return 0;
+}
+
+static int privcmd_enforce_singleshot_mapping(struct vm_area_struct *vma)
+{
+	return (xchg(&vma->vm_private_data, (void *)1) == NULL);
+}
+#endif
+
+const struct file_operations privcmd_file_ops = {
+	.unlocked_ioctl = privcmd_ioctl,
+	.mmap = privcmd_mmap,
+};
diff --git a/drivers/xen/xenfs/super.c b/drivers/xen/xenfs/super.c
new file mode 100644
index 00000000..1aa38971
--- /dev/null
+++ b/drivers/xen/xenfs/super.c
@@ -0,0 +1,137 @@
+/*
+ *  xenfs.c - a filesystem for passing info between the a domain and
+ *  the hypervisor.
+ *
+ * 2008-10-07  Alex Zeffertt    Replaced /proc/xen/xenbus with xenfs filesystem
+ *                              and /proc/xen compatibility mount point.
+ *                              Turned xenfs into a loadable module.
+ */
+
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/module.h>
+#include <linux/fs.h>
+#include <linux/magic.h>
+
+#include <xen/xen.h>
+
+#include "xenfs.h"
+
+#include <asm/xen/hypervisor.h>
+
+MODULE_DESCRIPTION("Xen filesystem");
+MODULE_LICENSE("GPL");
+
+static struct inode *xenfs_make_inode(struct super_block *sb, int mode)
+{
+	struct inode *ret = new_inode(sb);
+
+	if (ret) {
+		ret->i_mode = mode;
+		ret->i_uid = ret->i_gid = 0;
+		ret->i_blocks = 0;
+		ret->i_atime = ret->i_mtime = ret->i_ctime = CURRENT_TIME;
+	}
+	return ret;
+}
+
+static struct dentry *xenfs_create_file(struct super_block *sb,
+					struct dentry *parent,
+					const char *name,
+					const struct file_operations *fops,
+					void *data,
+					int mode)
+{
+	struct dentry *dentry;
+	struct inode *inode;
+
+	dentry = d_alloc_name(parent, name);
+	if (!dentry)
+		return NULL;
+
+	inode = xenfs_make_inode(sb, S_IFREG | mode);
+	if (!inode) {
+		dput(dentry);
+		return NULL;
+	}
+
+	inode->i_fop = fops;
+	inode->i_private = data;
+
+	d_add(dentry, inode);
+	return dentry;
+}
+
+static ssize_t capabilities_read(struct file *file, char __user *buf,
+				 size_t size, loff_t *off)
+{
+	char *tmp = "";
+
+	if (xen_initial_domain())
+		tmp = "control_d\n";
+
+	return simple_read_from_buffer(buf, size, off, tmp, strlen(tmp));
+}
+
+static const struct file_operations capabilities_file_ops = {
+	.read = capabilities_read,
+	.llseek = default_llseek,
+};
+
+static int xenfs_fill_super(struct super_block *sb, void *data, int silent)
+{
+	static struct tree_descr xenfs_files[] = {
+		[1] = {},
+		{ "xenbus", &xenbus_file_ops, S_IRUSR|S_IWUSR },
+		{ "capabilities", &capabilities_file_ops, S_IRUGO },
+		{ "privcmd", &privcmd_file_ops, S_IRUSR|S_IWUSR },
+		{""},
+	};
+	int rc;
+
+	rc = simple_fill_super(sb, XENFS_SUPER_MAGIC, xenfs_files);
+	if (rc < 0)
+		return rc;
+
+	if (xen_initial_domain()) {
+		xenfs_create_file(sb, sb->s_root, "xsd_kva",
+				  &xsd_kva_file_ops, NULL, S_IRUSR|S_IWUSR);
+		xenfs_create_file(sb, sb->s_root, "xsd_port",
+				  &xsd_port_file_ops, NULL, S_IRUSR|S_IWUSR);
+	}
+
+	return rc;
+}
+
+static struct dentry *xenfs_mount(struct file_system_type *fs_type,
+				  int flags, const char *dev_name,
+				  void *data)
+{
+	return mount_single(fs_type, flags, data, xenfs_fill_super);
+}
+
+static struct file_system_type xenfs_type = {
+	.owner =	THIS_MODULE,
+	.name =		"xenfs",
+	.mount =	xenfs_mount,
+	.kill_sb =	kill_litter_super,
+};
+
+static int __init xenfs_init(void)
+{
+	if (xen_domain())
+		return register_filesystem(&xenfs_type);
+
+	printk(KERN_INFO "XENFS: not registering filesystem on non-xen platform\n");
+	return 0;
+}
+
+static void __exit xenfs_exit(void)
+{
+	if (xen_domain())
+		unregister_filesystem(&xenfs_type);
+}
+
+module_init(xenfs_init);
+module_exit(xenfs_exit);
+
diff --git a/drivers/xen/xenfs/xenbus.c b/drivers/xen/xenfs/xenbus.c
new file mode 100644
index 00000000..bbd000f8
--- /dev/null
+++ b/drivers/xen/xenfs/xenbus.c
@@ -0,0 +1,593 @@
+/*
+ * Driver giving user-space access to the kernel's xenbus connection
+ * to xenstore.
+ *
+ * Copyright (c) 2005, Christian Limpach
+ * 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.
+ *
+ * Changes:
+ * 2008-10-07  Alex Zeffertt    Replaced /proc/xen/xenbus with xenfs filesystem
+ *                              and /proc/xen compatibility mount point.
+ *                              Turned xenfs into a loadable module.
+ */
+
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/uio.h>
+#include <linux/notifier.h>
+#include <linux/wait.h>
+#include <linux/fs.h>
+#include <linux/poll.h>
+#include <linux/mutex.h>
+#include <linux/sched.h>
+#include <linux/spinlock.h>
+#include <linux/mount.h>
+#include <linux/pagemap.h>
+#include <linux/uaccess.h>
+#include <linux/init.h>
+#include <linux/namei.h>
+#include <linux/string.h>
+#include <linux/slab.h>
+
+#include "xenfs.h"
+#include "../xenbus/xenbus_comms.h"
+
+#include <xen/xenbus.h>
+#include <asm/xen/hypervisor.h>
+
+/*
+ * An element of a list of outstanding transactions, for which we're
+ * still waiting a reply.
+ */
+struct xenbus_transaction_holder {
+	struct list_head list;
+	struct xenbus_transaction handle;
+};
+
+/*
+ * A buffer of data on the queue.
+ */
+struct read_buffer {
+	struct list_head list;
+	unsigned int cons;
+	unsigned int len;
+	char msg[];
+};
+
+struct xenbus_file_priv {
+	/*
+	 * msgbuffer_mutex is held while partial requests are built up
+	 * and complete requests are acted on.  It therefore protects
+	 * the "transactions" and "watches" lists, and the partial
+	 * request length and buffer.
+	 *
+	 * reply_mutex protects the reply being built up to return to
+	 * usermode.  It nests inside msgbuffer_mutex but may be held
+	 * alone during a watch callback.
+	 */
+	struct mutex msgbuffer_mutex;
+
+	/* In-progress transactions */
+	struct list_head transactions;
+
+	/* Active watches. */
+	struct list_head watches;
+
+	/* Partial request. */
+	unsigned int len;
+	union {
+		struct xsd_sockmsg msg;
+		char buffer[PAGE_SIZE];
+	} u;
+
+	/* Response queue. */
+	struct mutex reply_mutex;
+	struct list_head read_buffers;
+	wait_queue_head_t read_waitq;
+
+};
+
+/* Read out any raw xenbus messages queued up. */
+static ssize_t xenbus_file_read(struct file *filp,
+			       char __user *ubuf,
+			       size_t len, loff_t *ppos)
+{
+	struct xenbus_file_priv *u = filp->private_data;
+	struct read_buffer *rb;
+	unsigned i;
+	int ret;
+
+	mutex_lock(&u->reply_mutex);
+again:
+	while (list_empty(&u->read_buffers)) {
+		mutex_unlock(&u->reply_mutex);
+		if (filp->f_flags & O_NONBLOCK)
+			return -EAGAIN;
+
+		ret = wait_event_interruptible(u->read_waitq,
+					       !list_empty(&u->read_buffers));
+		if (ret)
+			return ret;
+		mutex_lock(&u->reply_mutex);
+	}
+
+	rb = list_entry(u->read_buffers.next, struct read_buffer, list);
+	i = 0;
+	while (i < len) {
+		unsigned sz = min((unsigned)len - i, rb->len - rb->cons);
+
+		ret = copy_to_user(ubuf + i, &rb->msg[rb->cons], sz);
+
+		i += sz - ret;
+		rb->cons += sz - ret;
+
+		if (ret != 0) {
+			if (i == 0)
+				i = -EFAULT;
+			goto out;
+		}
+
+		/* Clear out buffer if it has been consumed */
+		if (rb->cons == rb->len) {
+			list_del(&rb->list);
+			kfree(rb);
+			if (list_empty(&u->read_buffers))
+				break;
+			rb = list_entry(u->read_buffers.next,
+					struct read_buffer, list);
+		}
+	}
+	if (i == 0)
+		goto again;
+
+out:
+	mutex_unlock(&u->reply_mutex);
+	return i;
+}
+
+/*
+ * Add a buffer to the queue.  Caller must hold the appropriate lock
+ * if the queue is not local.  (Commonly the caller will build up
+ * multiple queued buffers on a temporary local list, and then add it
+ * to the appropriate list under lock once all the buffers have een
+ * successfully allocated.)
+ */
+static int queue_reply(struct list_head *queue, const void *data, size_t len)
+{
+	struct read_buffer *rb;
+
+	if (len == 0)
+		return 0;
+
+	rb = kmalloc(sizeof(*rb) + len, GFP_KERNEL);
+	if (rb == NULL)
+		return -ENOMEM;
+
+	rb->cons = 0;
+	rb->len = len;
+
+	memcpy(rb->msg, data, len);
+
+	list_add_tail(&rb->list, queue);
+	return 0;
+}
+
+/*
+ * Free all the read_buffer s on a list.
+ * Caller must have sole reference to list.
+ */
+static void queue_cleanup(struct list_head *list)
+{
+	struct read_buffer *rb;
+
+	while (!list_empty(list)) {
+		rb = list_entry(list->next, struct read_buffer, list);
+		list_del(list->next);
+		kfree(rb);
+	}
+}
+
+struct watch_adapter {
+	struct list_head list;
+	struct xenbus_watch watch;
+	struct xenbus_file_priv *dev_data;
+	char *token;
+};
+
+static void free_watch_adapter(struct watch_adapter *watch)
+{
+	kfree(watch->watch.node);
+	kfree(watch->token);
+	kfree(watch);
+}
+
+static struct watch_adapter *alloc_watch_adapter(const char *path,
+						 const char *token)
+{
+	struct watch_adapter *watch;
+
+	watch = kzalloc(sizeof(*watch), GFP_KERNEL);
+	if (watch == NULL)
+		goto out_fail;
+
+	watch->watch.node = kstrdup(path, GFP_KERNEL);
+	if (watch->watch.node == NULL)
+		goto out_free;
+
+	watch->token = kstrdup(token, GFP_KERNEL);
+	if (watch->token == NULL)
+		goto out_free;
+
+	return watch;
+
+out_free:
+	free_watch_adapter(watch);
+
+out_fail:
+	return NULL;
+}
+
+static void watch_fired(struct xenbus_watch *watch,
+			const char **vec,
+			unsigned int len)
+{
+	struct watch_adapter *adap;
+	struct xsd_sockmsg hdr;
+	const char *path, *token;
+	int path_len, tok_len, body_len, data_len = 0;
+	int ret;
+	LIST_HEAD(staging_q);
+
+	adap = container_of(watch, struct watch_adapter, watch);
+
+	path = vec[XS_WATCH_PATH];
+	token = adap->token;
+
+	path_len = strlen(path) + 1;
+	tok_len = strlen(token) + 1;
+	if (len > 2)
+		data_len = vec[len] - vec[2] + 1;
+	body_len = path_len + tok_len + data_len;
+
+	hdr.type = XS_WATCH_EVENT;
+	hdr.len = body_len;
+
+	mutex_lock(&adap->dev_data->reply_mutex);
+
+	ret = queue_reply(&staging_q, &hdr, sizeof(hdr));
+	if (!ret)
+		ret = queue_reply(&staging_q, path, path_len);
+	if (!ret)
+		ret = queue_reply(&staging_q, token, tok_len);
+	if (!ret && len > 2)
+		ret = queue_reply(&staging_q, vec[2], data_len);
+
+	if (!ret) {
+		/* success: pass reply list onto watcher */
+		list_splice_tail(&staging_q, &adap->dev_data->read_buffers);
+		wake_up(&adap->dev_data->read_waitq);
+	} else
+		queue_cleanup(&staging_q);
+
+	mutex_unlock(&adap->dev_data->reply_mutex);
+}
+
+static int xenbus_write_transaction(unsigned msg_type,
+				    struct xenbus_file_priv *u)
+{
+	int rc;
+	void *reply;
+	struct xenbus_transaction_holder *trans = NULL;
+	LIST_HEAD(staging_q);
+
+	if (msg_type == XS_TRANSACTION_START) {
+		trans = kmalloc(sizeof(*trans), GFP_KERNEL);
+		if (!trans) {
+			rc = -ENOMEM;
+			goto out;
+		}
+	}
+
+	reply = xenbus_dev_request_and_reply(&u->u.msg);
+	if (IS_ERR(reply)) {
+		kfree(trans);
+		rc = PTR_ERR(reply);
+		goto out;
+	}
+
+	if (msg_type == XS_TRANSACTION_START) {
+		trans->handle.id = simple_strtoul(reply, NULL, 0);
+
+		list_add(&trans->list, &u->transactions);
+	} else if (msg_type == XS_TRANSACTION_END) {
+		list_for_each_entry(trans, &u->transactions, list)
+			if (trans->handle.id == u->u.msg.tx_id)
+				break;
+		BUG_ON(&trans->list == &u->transactions);
+		list_del(&trans->list);
+
+		kfree(trans);
+	}
+
+	mutex_lock(&u->reply_mutex);
+	rc = queue_reply(&staging_q, &u->u.msg, sizeof(u->u.msg));
+	if (!rc)
+		rc = queue_reply(&staging_q, reply, u->u.msg.len);
+	if (!rc) {
+		list_splice_tail(&staging_q, &u->read_buffers);
+		wake_up(&u->read_waitq);
+	} else {
+		queue_cleanup(&staging_q);
+	}
+	mutex_unlock(&u->reply_mutex);
+
+	kfree(reply);
+
+out:
+	return rc;
+}
+
+static int xenbus_write_watch(unsigned msg_type, struct xenbus_file_priv *u)
+{
+	struct watch_adapter *watch, *tmp_watch;
+	char *path, *token;
+	int err, rc;
+	LIST_HEAD(staging_q);
+
+	path = u->u.buffer + sizeof(u->u.msg);
+	token = memchr(path, 0, u->u.msg.len);
+	if (token == NULL) {
+		rc = -EILSEQ;
+		goto out;
+	}
+	token++;
+
+	if (msg_type == XS_WATCH) {
+		watch = alloc_watch_adapter(path, token);
+		if (watch == NULL) {
+			rc = -ENOMEM;
+			goto out;
+		}
+
+		watch->watch.callback = watch_fired;
+		watch->dev_data = u;
+
+		err = register_xenbus_watch(&watch->watch);
+		if (err) {
+			free_watch_adapter(watch);
+			rc = err;
+			goto out;
+		}
+		list_add(&watch->list, &u->watches);
+	} else {
+		list_for_each_entry_safe(watch, tmp_watch, &u->watches, list) {
+			if (!strcmp(watch->token, token) &&
+			    !strcmp(watch->watch.node, path)) {
+				unregister_xenbus_watch(&watch->watch);
+				list_del(&watch->list);
+				free_watch_adapter(watch);
+				break;
+			}
+		}
+	}
+
+	/* Success.  Synthesize a reply to say all is OK. */
+	{
+		struct {
+			struct xsd_sockmsg hdr;
+			char body[3];
+		} __packed reply = {
+			{
+				.type = msg_type,
+				.len = sizeof(reply.body)
+			},
+			"OK"
+		};
+
+		mutex_lock(&u->reply_mutex);
+		rc = queue_reply(&u->read_buffers, &reply, sizeof(reply));
+		wake_up(&u->read_waitq);
+		mutex_unlock(&u->reply_mutex);
+	}
+
+out:
+	return rc;
+}
+
+static ssize_t xenbus_file_write(struct file *filp,
+				const char __user *ubuf,
+				size_t len, loff_t *ppos)
+{
+	struct xenbus_file_priv *u = filp->private_data;
+	uint32_t msg_type;
+	int rc = len;
+	int ret;
+	LIST_HEAD(staging_q);
+
+	/*
+	 * We're expecting usermode to be writing properly formed
+	 * xenbus messages.  If they write an incomplete message we
+	 * buffer it up.  Once it is complete, we act on it.
+	 */
+
+	/*
+	 * Make sure concurrent writers can't stomp all over each
+	 * other's messages and make a mess of our partial message
+	 * buffer.  We don't make any attemppt to stop multiple
+	 * writers from making a mess of each other's incomplete
+	 * messages; we're just trying to guarantee our own internal
+	 * consistency and make sure that single writes are handled
+	 * atomically.
+	 */
+	mutex_lock(&u->msgbuffer_mutex);
+
+	/* Get this out of the way early to avoid confusion */
+	if (len == 0)
+		goto out;
+
+	/* Can't write a xenbus message larger we can buffer */
+	if ((len + u->len) > sizeof(u->u.buffer)) {
+		/* On error, dump existing buffer */
+		u->len = 0;
+		rc = -EINVAL;
+		goto out;
+	}
+
+	ret = copy_from_user(u->u.buffer + u->len, ubuf, len);
+
+	if (ret != 0) {
+		rc = -EFAULT;
+		goto out;
+	}
+
+	/* Deal with a partial copy. */
+	len -= ret;
+	rc = len;
+
+	u->len += len;
+
+	/* Return if we haven't got a full message yet */
+	if (u->len < sizeof(u->u.msg))
+		goto out;	/* not even the header yet */
+
+	/* If we're expecting a message that's larger than we can
+	   possibly send, dump what we have and return an error. */
+	if ((sizeof(u->u.msg) + u->u.msg.len) > sizeof(u->u.buffer)) {
+		rc = -E2BIG;
+		u->len = 0;
+		goto out;
+	}
+
+	if (u->len < (sizeof(u->u.msg) + u->u.msg.len))
+		goto out;	/* incomplete data portion */
+
+	/*
+	 * OK, now we have a complete message.  Do something with it.
+	 */
+
+	msg_type = u->u.msg.type;
+
+	switch (msg_type) {
+	case XS_WATCH:
+	case XS_UNWATCH:
+		/* (Un)Ask for some path to be watched for changes */
+		ret = xenbus_write_watch(msg_type, u);
+		break;
+
+	default:
+		/* Send out a transaction */
+		ret = xenbus_write_transaction(msg_type, u);
+		break;
+	}
+	if (ret != 0)
+		rc = ret;
+
+	/* Buffered message consumed */
+	u->len = 0;
+
+ out:
+	mutex_unlock(&u->msgbuffer_mutex);
+	return rc;
+}
+
+static int xenbus_file_open(struct inode *inode, struct file *filp)
+{
+	struct xenbus_file_priv *u;
+
+	if (xen_store_evtchn == 0)
+		return -ENOENT;
+
+	nonseekable_open(inode, filp);
+
+	u = kzalloc(sizeof(*u), GFP_KERNEL);
+	if (u == NULL)
+		return -ENOMEM;
+
+	INIT_LIST_HEAD(&u->transactions);
+	INIT_LIST_HEAD(&u->watches);
+	INIT_LIST_HEAD(&u->read_buffers);
+	init_waitqueue_head(&u->read_waitq);
+
+	mutex_init(&u->reply_mutex);
+	mutex_init(&u->msgbuffer_mutex);
+
+	filp->private_data = u;
+
+	return 0;
+}
+
+static int xenbus_file_release(struct inode *inode, struct file *filp)
+{
+	struct xenbus_file_priv *u = filp->private_data;
+	struct xenbus_transaction_holder *trans, *tmp;
+	struct watch_adapter *watch, *tmp_watch;
+	struct read_buffer *rb, *tmp_rb;
+
+	/*
+	 * No need for locking here because there are no other users,
+	 * by definition.
+	 */
+
+	list_for_each_entry_safe(trans, tmp, &u->transactions, list) {
+		xenbus_transaction_end(trans->handle, 1);
+		list_del(&trans->list);
+		kfree(trans);
+	}
+
+	list_for_each_entry_safe(watch, tmp_watch, &u->watches, list) {
+		unregister_xenbus_watch(&watch->watch);
+		list_del(&watch->list);
+		free_watch_adapter(watch);
+	}
+
+	list_for_each_entry_safe(rb, tmp_rb, &u->read_buffers, list) {
+		list_del(&rb->list);
+		kfree(rb);
+	}
+	kfree(u);
+
+	return 0;
+}
+
+static unsigned int xenbus_file_poll(struct file *file, poll_table *wait)
+{
+	struct xenbus_file_priv *u = file->private_data;
+
+	poll_wait(file, &u->read_waitq, wait);
+	if (!list_empty(&u->read_buffers))
+		return POLLIN | POLLRDNORM;
+	return 0;
+}
+
+const struct file_operations xenbus_file_ops = {
+	.read = xenbus_file_read,
+	.write = xenbus_file_write,
+	.open = xenbus_file_open,
+	.release = xenbus_file_release,
+	.poll = xenbus_file_poll,
+	.llseek = no_llseek,
+};
diff --git a/drivers/xen/xenfs/xenfs.h b/drivers/xen/xenfs/xenfs.h
new file mode 100644
index 00000000..b68aa620
--- /dev/null
+++ b/drivers/xen/xenfs/xenfs.h
@@ -0,0 +1,9 @@
+#ifndef _XENFS_XENBUS_H
+#define _XENFS_XENBUS_H
+
+extern const struct file_operations xenbus_file_ops;
+extern const struct file_operations privcmd_file_ops;
+extern const struct file_operations xsd_kva_file_ops;
+extern const struct file_operations xsd_port_file_ops;
+
+#endif	/* _XENFS_XENBUS_H */
diff --git a/drivers/xen/xenfs/xenstored.c b/drivers/xen/xenfs/xenstored.c
new file mode 100644
index 00000000..fef20dbc
--- /dev/null
+++ b/drivers/xen/xenfs/xenstored.c
@@ -0,0 +1,68 @@
+#include <linux/slab.h>
+#include <linux/types.h>
+#include <linux/mm.h>
+#include <linux/fs.h>
+
+#include <xen/page.h>
+
+#include "xenfs.h"
+#include "../xenbus/xenbus_comms.h"
+
+static ssize_t xsd_read(struct file *file, char __user *buf,
+			    size_t size, loff_t *off)
+{
+	const char *str = (const char *)file->private_data;
+	return simple_read_from_buffer(buf, size, off, str, strlen(str));
+}
+
+static int xsd_release(struct inode *inode, struct file *file)
+{
+	kfree(file->private_data);
+	return 0;
+}
+
+static int xsd_kva_open(struct inode *inode, struct file *file)
+{
+	file->private_data = (void *)kasprintf(GFP_KERNEL, "0x%p",
+					       xen_store_interface);
+	if (!file->private_data)
+		return -ENOMEM;
+	return 0;
+}
+
+static int xsd_kva_mmap(struct file *file, struct vm_area_struct *vma)
+{
+	size_t size = vma->vm_end - vma->vm_start;
+
+	if ((size > PAGE_SIZE) || (vma->vm_pgoff != 0))
+		return -EINVAL;
+
+	if (remap_pfn_range(vma, vma->vm_start,
+			    virt_to_pfn(xen_store_interface),
+			    size, vma->vm_page_prot))
+		return -EAGAIN;
+
+	return 0;
+}
+
+const struct file_operations xsd_kva_file_ops = {
+	.open = xsd_kva_open,
+	.mmap = xsd_kva_mmap,
+	.read = xsd_read,
+	.release = xsd_release,
+};
+
+static int xsd_port_open(struct inode *inode, struct file *file)
+{
+	file->private_data = (void *)kasprintf(GFP_KERNEL, "%d",
+					       xen_store_evtchn);
+	if (!file->private_data)
+		return -ENOMEM;
+	return 0;
+}
+
+const struct file_operations xsd_port_file_ops = {
+	.open = xsd_port_open,
+	.read = xsd_read,
+	.release = xsd_release,
+};