xen: add blkback support

Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>

Conflicts:

	drivers/xen/Makefile

drivers/xen/Kconfig

@@ -37,6 +37,14 @@ config XEN_BACKEND
 	  Support for backend device drivers that provide I/O services
 	  to other virtual machines.
 
+config XEN_BLKDEV_BACKEND
+	tristate "Block-device backend driver"
+	depends on XEN_BACKEND && BLOCK
+	help
+	  The block-device backend driver allows the kernel to export its
+	  block devices to other guests via a high-performance shared-memory
+	  interface.
+
 config XENFS
 	tristate "Xen filesystem"
 	default y

drivers/xen/Makefile

@@ -11,6 +11,7 @@ obj-$(CONFIG_XEN_BALLOON)	+= xen-balloon.o
 obj-$(CONFIG_XEN_DEV_EVTCHN)	+= xen-evtchn.o
 obj-$(CONFIG_XEN_GNTDEV)	+= xen-gntdev.o
 obj-$(CONFIG_XEN_GRANT_DEV_ALLOC)	+= xen-gntalloc.o
+obj-$(CONFIG_XEN_BLKDEV_BACKEND)	+= blkback/
 obj-$(CONFIG_XENFS)		+= xenfs/
 obj-$(CONFIG_XEN_SYS_HYPERVISOR)	+= sys-hypervisor.o
 obj-$(CONFIG_XEN_PLATFORM_PCI)	+= xen-platform-pci.o

drivers/xen/blkback/Makefile

@@ -0,0 +1,3 @@
+obj-$(CONFIG_XEN_BLKDEV_BACKEND) := blkbk.o
+
+blkbk-y	:= blkback.o xenbus.o interface.o vbd.o

drivers/xen/blkback/blkback.c

@@ -0,0 +1,656 @@
+/******************************************************************************
+ * arch/xen/drivers/blkif/backend/main.c
+ *
+ * Back-end of the driver for virtual block devices. This portion of the
+ * driver exports a 'unified' block-device interface that can be accessed
+ * by any operating system that implements a compatible front end. A
+ * reference front-end implementation can be found in:
+ *  arch/xen/drivers/blkif/frontend
+ *
+ * Copyright (c) 2003-2004, Keir Fraser & Steve Hand
+ * Copyright (c) 2005, Christopher Clark
+ *
+ * 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/spinlock.h>
+#include <linux/kthread.h>
+#include <linux/list.h>
+#include <linux/delay.h>
+#include <xen/balloon.h>
+#include <asm/hypervisor.h>
+#include "common.h"
+
+/*
+ * These are rather arbitrary. They are fairly large because adjacent requests
+ * pulled from a communication ring are quite likely to end up being part of
+ * the same scatter/gather request at the disc.
+ *
+ * ** TRY INCREASING 'blkif_reqs' IF WRITE SPEEDS SEEM TOO LOW **
+ *
+ * This will increase the chances of being able to write whole tracks.
+ * 64 should be enough to keep us competitive with Linux.
+ */
+static int blkif_reqs = 64;
+module_param_named(reqs, blkif_reqs, int, 0);
+MODULE_PARM_DESC(reqs, "Number of blkback requests to allocate");
+
+/* Run-time switchable: /sys/module/blkback/parameters/ */
+static unsigned int log_stats = 0;
+static unsigned int debug_lvl = 0;
+module_param(log_stats, int, 0644);
+module_param(debug_lvl, int, 0644);
+
+/*
+ * Each outstanding request that we've passed to the lower device layers has a
+ * 'pending_req' allocated to it. Each buffer_head that completes decrements
+ * the pendcnt towards zero. When it hits zero, the specified domain has a
+ * response queued for it, with the saved 'id' passed back.
+ */
+typedef struct {
+	blkif_t       *blkif;
+	u64            id;
+	int            nr_pages;
+	atomic_t       pendcnt;
+	unsigned short operation;
+	int            status;
+	struct list_head free_list;
+} pending_req_t;
+
+static pending_req_t *pending_reqs;
+static struct list_head pending_free;
+static DEFINE_SPINLOCK(pending_free_lock);
+static DECLARE_WAIT_QUEUE_HEAD(pending_free_wq);
+
+#define BLKBACK_INVALID_HANDLE (~0)
+
+static struct page **pending_pages;
+static grant_handle_t *pending_grant_handles;
+
+static inline int vaddr_pagenr(pending_req_t *req, int seg)
+{
+	return (req - pending_reqs) * BLKIF_MAX_SEGMENTS_PER_REQUEST + seg;
+}
+
+static inline unsigned long vaddr(pending_req_t *req, int seg)
+{
+	unsigned long pfn = page_to_pfn(pending_pages[vaddr_pagenr(req, seg)]);
+	return (unsigned long)pfn_to_kaddr(pfn);
+}
+
+#define pending_handle(_req, _seg) \
+	(pending_grant_handles[vaddr_pagenr(_req, _seg)])
+
+
+static int do_block_io_op(blkif_t *blkif);
+static void dispatch_rw_block_io(blkif_t *blkif,
+				 blkif_request_t *req,
+				 pending_req_t *pending_req);
+static void make_response(blkif_t *blkif, u64 id,
+			  unsigned short op, int st);
+
+/******************************************************************
+ * misc small helpers
+ */
+static pending_req_t* alloc_req(void)
+{
+	pending_req_t *req = NULL;
+	unsigned long flags;
+
+	spin_lock_irqsave(&pending_free_lock, flags);
+	if (!list_empty(&pending_free)) {
+		req = list_entry(pending_free.next, pending_req_t, free_list);
+		list_del(&req->free_list);
+	}
+	spin_unlock_irqrestore(&pending_free_lock, flags);
+	return req;
+}
+
+static void free_req(pending_req_t *req)
+{
+	unsigned long flags;
+	int was_empty;
+
+	spin_lock_irqsave(&pending_free_lock, flags);
+	was_empty = list_empty(&pending_free);
+	list_add(&req->free_list, &pending_free);
+	spin_unlock_irqrestore(&pending_free_lock, flags);
+	if (was_empty)
+		wake_up(&pending_free_wq);
+}
+
+static void unplug_queue(blkif_t *blkif)
+{
+	if (blkif->plug == NULL)
+		return;
+	if (blkif->plug->unplug_fn)
+		blkif->plug->unplug_fn(blkif->plug);
+	blk_put_queue(blkif->plug);
+	blkif->plug = NULL;
+}
+
+static void plug_queue(blkif_t *blkif, struct block_device *bdev)
+{
+	request_queue_t *q = bdev_get_queue(bdev);
+
+	if (q == blkif->plug)
+		return;
+	unplug_queue(blkif);
+	blk_get_queue(q);
+	blkif->plug = q;
+}
+
+static void fast_flush_area(pending_req_t *req)
+{
+	struct gnttab_unmap_grant_ref unmap[BLKIF_MAX_SEGMENTS_PER_REQUEST];
+	unsigned int i, invcount = 0;
+	grant_handle_t handle;
+	int ret;
+
+	for (i = 0; i < req->nr_pages; i++) {
+		handle = pending_handle(req, i);
+		if (handle == BLKBACK_INVALID_HANDLE)
+			continue;
+		gnttab_set_unmap_op(&unmap[invcount], vaddr(req, i),
+				    GNTMAP_host_map, handle);
+		pending_handle(req, i) = BLKBACK_INVALID_HANDLE;
+		invcount++;
+	}
+
+	ret = HYPERVISOR_grant_table_op(
+		GNTTABOP_unmap_grant_ref, unmap, invcount);
+	BUG_ON(ret);
+}
+
+/******************************************************************
+ * SCHEDULER FUNCTIONS
+ */
+
+static void print_stats(blkif_t *blkif)
+{
+	printk(KERN_DEBUG "%s: oo %3d  |  rd %4d  |  wr %4d  |  br %4d\n",
+	       current->comm, blkif->st_oo_req,
+	       blkif->st_rd_req, blkif->st_wr_req, blkif->st_br_req);
+	blkif->st_print = jiffies + msecs_to_jiffies(10 * 1000);
+	blkif->st_rd_req = 0;
+	blkif->st_wr_req = 0;
+	blkif->st_oo_req = 0;
+}
+
+int blkif_schedule(void *arg)
+{
+	blkif_t *blkif = arg;
+
+	blkif_get(blkif);
+
+	if (debug_lvl)
+		printk(KERN_DEBUG "%s: started\n", current->comm);
+
+	while (!kthread_should_stop()) {
+		if (try_to_freeze())
+			continue;
+
+		wait_event_interruptible(
+			blkif->wq,
+			blkif->waiting_reqs || kthread_should_stop());
+		wait_event_interruptible(
+			pending_free_wq,
+			!list_empty(&pending_free) || kthread_should_stop());
+
+		blkif->waiting_reqs = 0;
+		smp_mb(); /* clear flag *before* checking for work */
+
+		if (do_block_io_op(blkif))
+			blkif->waiting_reqs = 1;
+		unplug_queue(blkif);
+
+		if (log_stats && time_after(jiffies, blkif->st_print))
+			print_stats(blkif);
+	}
+
+	if (log_stats)
+		print_stats(blkif);
+	if (debug_lvl)
+		printk(KERN_DEBUG "%s: exiting\n", current->comm);
+
+	blkif->xenblkd = NULL;
+	blkif_put(blkif);
+
+	return 0;
+}
+
+/******************************************************************
+ * COMPLETION CALLBACK -- Called as bh->b_end_io()
+ */
+
+static void __end_block_io_op(pending_req_t *pending_req, int error)
+{
+	/* An error fails the entire request. */
+	if ((pending_req->operation == BLKIF_OP_WRITE_BARRIER) &&
+	    (error == -EOPNOTSUPP)) {
+		DPRINTK("blkback: write barrier op failed, not supported\n");
+		blkback_barrier(XBT_NIL, pending_req->blkif->be, 0);
+		pending_req->status = BLKIF_RSP_EOPNOTSUPP;
+	} else if (error) {
+		DPRINTK("Buffer not up-to-date at end of operation, "
+			"error=%d\n", error);
+		pending_req->status = BLKIF_RSP_ERROR;
+	}
+
+	if (atomic_dec_and_test(&pending_req->pendcnt)) {
+		fast_flush_area(pending_req);
+		make_response(pending_req->blkif, pending_req->id,
+			      pending_req->operation, pending_req->status);
+		blkif_put(pending_req->blkif);
+		free_req(pending_req);
+	}
+}
+
+static int end_block_io_op(struct bio *bio, unsigned int done, int error)
+{
+	if (bio->bi_size != 0)
+		return 1;
+	__end_block_io_op(bio->bi_private, error);
+	bio_put(bio);
+	return error;
+}
+
+
+/******************************************************************************
+ * NOTIFICATION FROM GUEST OS.
+ */
+
+static void blkif_notify_work(blkif_t *blkif)
+{
+	blkif->waiting_reqs = 1;
+	wake_up(&blkif->wq);
+}
+
+irqreturn_t blkif_be_int(int irq, void *dev_id, struct pt_regs *regs)
+{
+	blkif_notify_work(dev_id);
+	return IRQ_HANDLED;
+}
+
+
+
+/******************************************************************
+ * DOWNWARD CALLS -- These interface with the block-device layer proper.
+ */
+
+static int do_block_io_op(blkif_t *blkif)
+{
+	blkif_back_rings_t *blk_rings = &blkif->blk_rings;
+	blkif_request_t req;
+	pending_req_t *pending_req;
+	RING_IDX rc, rp;
+	int more_to_do = 0;
+
+	rc = blk_rings->common.req_cons;
+	rp = blk_rings->common.sring->req_prod;
+	rmb(); /* Ensure we see queued requests up to 'rp'. */
+
+	while (rc != rp) {
+
+		if (RING_REQUEST_CONS_OVERFLOW(&blk_rings->common, rc))
+			break;
+
+		pending_req = alloc_req();
+		if (NULL == pending_req) {
+			blkif->st_oo_req++;
+			more_to_do = 1;
+			break;
+		}
+
+		if (kthread_should_stop()) {
+			more_to_do = 1;
+			break;
+		}
+
+		switch (blkif->blk_protocol) {
+		case BLKIF_PROTOCOL_NATIVE:
+			memcpy(&req, RING_GET_REQUEST(&blk_rings->native, rc), sizeof(req));
+			break;
+		case BLKIF_PROTOCOL_X86_32:
+			blkif_get_x86_32_req(&req, RING_GET_REQUEST(&blk_rings->x86_32, rc));
+			break;
+		case BLKIF_PROTOCOL_X86_64:
+			blkif_get_x86_64_req(&req, RING_GET_REQUEST(&blk_rings->x86_64, rc));
+			break;
+		default:
+			BUG();
+		}
+		blk_rings->common.req_cons = ++rc; /* before make_response() */
+
+		/* Apply all sanity checks to /private copy/ of request. */
+		barrier();
+
+		switch (req.operation) {
+		case BLKIF_OP_READ:
+			blkif->st_rd_req++;
+			dispatch_rw_block_io(blkif, &req, pending_req);
+			break;
+		case BLKIF_OP_WRITE_BARRIER:
+			blkif->st_br_req++;
+			/* fall through */
+		case BLKIF_OP_WRITE:
+			blkif->st_wr_req++;
+			dispatch_rw_block_io(blkif, &req, pending_req);
+			break;
+		default:
+			/* A good sign something is wrong: sleep for a while to
+			 * avoid excessive CPU consumption by a bad guest. */
+			msleep(1);
+			DPRINTK("error: unknown block io operation [%d]\n",
+				req.operation);
+			make_response(blkif, req.id, req.operation,
+				      BLKIF_RSP_ERROR);
+			free_req(pending_req);
+			break;
+		}
+
+		/* Yield point for this unbounded loop. */
+		cond_resched();
+	}
+
+	return more_to_do;
+}
+
+static void dispatch_rw_block_io(blkif_t *blkif,
+				 blkif_request_t *req,
+				 pending_req_t *pending_req)
+{
+	extern void ll_rw_block(int rw, int nr, struct buffer_head * bhs[]);
+	struct gnttab_map_grant_ref map[BLKIF_MAX_SEGMENTS_PER_REQUEST];
+	struct phys_req preq;
+	struct {
+		unsigned long buf; unsigned int nsec;
+	} seg[BLKIF_MAX_SEGMENTS_PER_REQUEST];
+	unsigned int nseg;
+	struct bio *bio = NULL;
+	int ret, i;
+	int operation;
+
+	switch (req->operation) {
+	case BLKIF_OP_READ:
+		operation = READ;
+		break;
+	case BLKIF_OP_WRITE:
+		operation = WRITE;
+		break;
+	case BLKIF_OP_WRITE_BARRIER:
+		operation = WRITE_BARRIER;
+		break;
+	default:
+		operation = 0; /* make gcc happy */
+		BUG();
+	}
+
+	/* Check that number of segments is sane. */
+	nseg = req->nr_segments;
+	if (unlikely(nseg == 0 && operation != WRITE_BARRIER) ||
+	    unlikely(nseg > BLKIF_MAX_SEGMENTS_PER_REQUEST)) {
+		DPRINTK("Bad number of segments in request (%d)\n", nseg);
+		goto fail_response;
+	}
+
+	preq.dev           = req->handle;
+	preq.sector_number = req->sector_number;
+	preq.nr_sects      = 0;
+
+	pending_req->blkif     = blkif;
+	pending_req->id        = req->id;
+	pending_req->operation = req->operation;
+	pending_req->status    = BLKIF_RSP_OKAY;
+	pending_req->nr_pages  = nseg;
+
+	for (i = 0; i < nseg; i++) {
+		uint32_t flags;
+
+		seg[i].nsec = req->seg[i].last_sect -
+			req->seg[i].first_sect + 1;
+
+		if ((req->seg[i].last_sect >= (PAGE_SIZE >> 9)) ||
+		    (req->seg[i].last_sect < req->seg[i].first_sect))
+			goto fail_response;
+		preq.nr_sects += seg[i].nsec;
+
+		flags = GNTMAP_host_map;
+		if (operation != READ)
+			flags |= GNTMAP_readonly;
+		gnttab_set_map_op(&map[i], vaddr(pending_req, i), flags,
+				  req->seg[i].gref, blkif->domid);
+	}
+
+	ret = HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref, map, nseg);
+	BUG_ON(ret);
+
+	for (i = 0; i < nseg; i++) {
+		if (unlikely(map[i].status != 0)) {
+			DPRINTK("invalid buffer -- could not remap it\n");
+			map[i].handle = BLKBACK_INVALID_HANDLE;
+			ret |= 1;
+		}
+
+		pending_handle(pending_req, i) = map[i].handle;
+
+		if (ret)
+			continue;
+
+		set_phys_to_machine(__pa(vaddr(
+			pending_req, i)) >> PAGE_SHIFT,
+			FOREIGN_FRAME(map[i].dev_bus_addr >> PAGE_SHIFT));
+		seg[i].buf  = map[i].dev_bus_addr |
+			(req->seg[i].first_sect << 9);
+	}
+
+	if (ret)
+		goto fail_flush;
+
+	if (vbd_translate(&preq, blkif, operation) != 0) {
+		DPRINTK("access denied: %s of [%llu,%llu] on dev=%04x\n",
+			operation == READ ? "read" : "write",
+			preq.sector_number,
+			preq.sector_number + preq.nr_sects, preq.dev);
+		goto fail_flush;
+	}
+
+	plug_queue(blkif, preq.bdev);
+	atomic_set(&pending_req->pendcnt, 1);
+	blkif_get(blkif);
+
+	for (i = 0; i < nseg; i++) {
+		if (((int)preq.sector_number|(int)seg[i].nsec) &
+		    ((bdev_hardsect_size(preq.bdev) >> 9) - 1)) {
+			DPRINTK("Misaligned I/O request from domain %d",
+				blkif->domid);
+			goto fail_put_bio;
+		}
+
+		while ((bio == NULL) ||
+		       (bio_add_page(bio,
+				     virt_to_page(vaddr(pending_req, i)),
+				     seg[i].nsec << 9,
+				     seg[i].buf & ~PAGE_MASK) == 0)) {
+			if (bio) {
+				atomic_inc(&pending_req->pendcnt);
+				submit_bio(operation, bio);
+			}
+
+			bio = bio_alloc(GFP_KERNEL, nseg-i);
+			if (unlikely(bio == NULL))
+				goto fail_put_bio;
+
+			bio->bi_bdev    = preq.bdev;
+			bio->bi_private = pending_req;
+			bio->bi_end_io  = end_block_io_op;
+			bio->bi_sector  = preq.sector_number;
+		}
+
+		preq.sector_number += seg[i].nsec;
+	}
+
+	if (!bio) {
+		BUG_ON(operation != WRITE_BARRIER);
+		bio = bio_alloc(GFP_KERNEL, 0);
+		if (unlikely(bio == NULL))
+			goto fail_put_bio;
+
+		bio->bi_bdev    = preq.bdev;
+		bio->bi_private = pending_req;
+		bio->bi_end_io  = end_block_io_op;
+		bio->bi_sector  = -1;
+	}
+
+	submit_bio(operation, bio);
+
+	if (operation == READ)
+		blkif->st_rd_sect += preq.nr_sects;
+	else if (operation == WRITE || operation == WRITE_BARRIER)
+		blkif->st_wr_sect += preq.nr_sects;
+
+	return;
+
+ fail_flush:
+	fast_flush_area(pending_req);
+ fail_response:
+	make_response(blkif, req->id, req->operation, BLKIF_RSP_ERROR);
+	free_req(pending_req);
+	msleep(1); /* back off a bit */
+	return;
+
+ fail_put_bio:
+	__end_block_io_op(pending_req, -EINVAL);
+	if (bio)
+		bio_put(bio);
+	unplug_queue(blkif);
+	msleep(1); /* back off a bit */
+	return;
+}
+
+
+
+/******************************************************************
+ * MISCELLANEOUS SETUP / TEARDOWN / DEBUGGING
+ */
+
+
+static void make_response(blkif_t *blkif, u64 id,
+			  unsigned short op, int st)
+{
+	blkif_response_t  resp;
+	unsigned long     flags;
+	blkif_back_rings_t *blk_rings = &blkif->blk_rings;
+	int more_to_do = 0;
+	int notify;
+
+	resp.id        = id;
+	resp.operation = op;
+	resp.status    = st;
+
+	spin_lock_irqsave(&blkif->blk_ring_lock, flags);
+	/* Place on the response ring for the relevant domain. */
+	switch (blkif->blk_protocol) {
+	case BLKIF_PROTOCOL_NATIVE:
+		memcpy(RING_GET_RESPONSE(&blk_rings->native, blk_rings->native.rsp_prod_pvt),
+		       &resp, sizeof(resp));
+		break;
+	case BLKIF_PROTOCOL_X86_32:
+		memcpy(RING_GET_RESPONSE(&blk_rings->x86_32, blk_rings->x86_32.rsp_prod_pvt),
+		       &resp, sizeof(resp));
+		break;
+	case BLKIF_PROTOCOL_X86_64:
+		memcpy(RING_GET_RESPONSE(&blk_rings->x86_64, blk_rings->x86_64.rsp_prod_pvt),
+		       &resp, sizeof(resp));
+		break;
+	default:
+		BUG();
+	}
+	blk_rings->common.rsp_prod_pvt++;
+	RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(&blk_rings->common, notify);
+	if (blk_rings->common.rsp_prod_pvt == blk_rings->common.req_cons) {
+		/*
+		 * Tail check for pending requests. Allows frontend to avoid
+		 * notifications if requests are already in flight (lower
+		 * overheads and promotes batching).
+		 */
+		RING_FINAL_CHECK_FOR_REQUESTS(&blk_rings->common, more_to_do);
+
+	} else if (RING_HAS_UNCONSUMED_REQUESTS(&blk_rings->common)) {
+		more_to_do = 1;
+	}
+
+	spin_unlock_irqrestore(&blkif->blk_ring_lock, flags);
+
+	if (more_to_do)
+		blkif_notify_work(blkif);
+	if (notify)
+		notify_remote_via_irq(blkif->irq);
+}
+
+static int __init blkif_init(void)
+{
+	int i, mmap_pages;
+
+	if (!is_running_on_xen())
+		return -ENODEV;
+
+	mmap_pages = blkif_reqs * BLKIF_MAX_SEGMENTS_PER_REQUEST;
+
+	pending_reqs          = kmalloc(sizeof(pending_reqs[0]) *
+					blkif_reqs, GFP_KERNEL);
+	pending_grant_handles = kmalloc(sizeof(pending_grant_handles[0]) *
+					mmap_pages, GFP_KERNEL);
+	pending_pages         = alloc_empty_pages_and_pagevec(mmap_pages);
+
+	if (!pending_reqs || !pending_grant_handles || !pending_pages)
+		goto out_of_memory;
+
+	for (i = 0; i < mmap_pages; i++)
+		pending_grant_handles[i] = BLKBACK_INVALID_HANDLE;
+
+	blkif_interface_init();
+
+	memset(pending_reqs, 0, sizeof(pending_reqs));
+	INIT_LIST_HEAD(&pending_free);
+
+	for (i = 0; i < blkif_reqs; i++)
+		list_add_tail(&pending_reqs[i].free_list, &pending_free);
+
+	blkif_xenbus_init();
+
+	return 0;
+
+ out_of_memory:
+	kfree(pending_reqs);
+	kfree(pending_grant_handles);
+	free_empty_pages_and_pagevec(pending_pages, mmap_pages);
+	printk("%s: out of memory\n", __FUNCTION__);
+	return -ENOMEM;
+}
+
+module_init(blkif_init);
+
+MODULE_LICENSE("Dual BSD/GPL");

drivers/xen/blkback/common.h

@@ -0,0 +1,139 @@
+/*
+ * 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.
+ */
+
+#ifndef __BLKIF__BACKEND__COMMON_H__
+#define __BLKIF__BACKEND__COMMON_H__
+
+#include <linux/version.h>
+#include <linux/module.h>
+#include <linux/interrupt.h>
+#include <linux/slab.h>
+#include <linux/blkdev.h>
+#include <linux/vmalloc.h>
+#include <linux/wait.h>
+#include <asm/io.h>
+#include <asm/setup.h>
+#include <asm/pgalloc.h>
+#include <xen/evtchn.h>
+#include <asm/hypervisor.h>
+#include <xen/blkif.h>
+#include <xen/gnttab.h>
+#include <xen/driver_util.h>
+#include <xen/xenbus.h>
+
+#define DPRINTK(_f, _a...)			\
+	pr_debug("(file=%s, line=%d) " _f,	\
+		 __FILE__ , __LINE__ , ## _a )
+
+struct vbd {
+	blkif_vdev_t   handle;      /* what the domain refers to this vbd as */
+	unsigned char  readonly;    /* Non-zero -> read-only */
+	unsigned char  type;        /* VDISK_xxx */
+	u32            pdevice;     /* phys device that this vbd maps to */
+	struct block_device *bdev;
+};
+
+struct backend_info;
+
+typedef struct blkif_st {
+	/* Unique identifier for this interface. */
+	domid_t           domid;
+	unsigned int      handle;
+	/* Physical parameters of the comms window. */
+	unsigned int      irq;
+	/* Comms information. */
+	enum blkif_protocol blk_protocol;
+	blkif_back_rings_t blk_rings;
+	struct vm_struct *blk_ring_area;
+	/* The VBD attached to this interface. */
+	struct vbd        vbd;
+	/* Back pointer to the backend_info. */
+	struct backend_info *be;
+	/* Private fields. */
+	spinlock_t       blk_ring_lock;
+	atomic_t         refcnt;
+
+	wait_queue_head_t   wq;
+	struct task_struct  *xenblkd;
+	unsigned int        waiting_reqs;
+	request_queue_t     *plug;
+
+	/* statistics */
+	unsigned long       st_print;
+	int                 st_rd_req;
+	int                 st_wr_req;
+	int                 st_oo_req;
+	int                 st_br_req;
+	int                 st_rd_sect;
+	int                 st_wr_sect;
+
+	wait_queue_head_t waiting_to_free;
+
+	grant_handle_t shmem_handle;
+	grant_ref_t    shmem_ref;
+} blkif_t;
+
+blkif_t *blkif_alloc(domid_t domid);
+void blkif_disconnect(blkif_t *blkif);
+void blkif_free(blkif_t *blkif);
+int blkif_map(blkif_t *blkif, unsigned long shared_page, unsigned int evtchn);
+
+#define blkif_get(_b) (atomic_inc(&(_b)->refcnt))
+#define blkif_put(_b)					\
+	do {						\
+		if (atomic_dec_and_test(&(_b)->refcnt))	\
+			wake_up(&(_b)->waiting_to_free);\
+	} while (0)
+
+/* Create a vbd. */
+int vbd_create(blkif_t *blkif, blkif_vdev_t vdevice, unsigned major,
+	       unsigned minor, int readonly, int cdrom);
+void vbd_free(struct vbd *vbd);
+
+unsigned long long vbd_size(struct vbd *vbd);
+unsigned int vbd_info(struct vbd *vbd);
+unsigned long vbd_secsize(struct vbd *vbd);
+
+struct phys_req {
+	unsigned short       dev;
+	unsigned short       nr_sects;
+	struct block_device *bdev;
+	blkif_sector_t       sector_number;
+};
+
+int vbd_translate(struct phys_req *req, blkif_t *blkif, int operation);
+
+void blkif_interface_init(void);
+
+void blkif_xenbus_init(void);
+
+irqreturn_t blkif_be_int(int irq, void *dev_id, struct pt_regs *regs);
+int blkif_schedule(void *arg);
+
+int blkback_barrier(struct xenbus_transaction xbt,
+		    struct backend_info *be, int state);
+
+#endif /* __BLKIF__BACKEND__COMMON_H__ */

drivers/xen/blkback/interface.c

@@ -0,0 +1,181 @@
+/******************************************************************************
+ * arch/xen/drivers/blkif/backend/interface.c
+ *
+ * Block-device interface management.
+ *
+ * Copyright (c) 2004, Keir Fraser
+ *
+ * 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 "common.h"
+#include <xen/evtchn.h>
+#include <linux/kthread.h>
+
+static kmem_cache_t *blkif_cachep;
+
+blkif_t *blkif_alloc(domid_t domid)
+{
+	blkif_t *blkif;
+
+	blkif = kmem_cache_alloc(blkif_cachep, GFP_KERNEL);
+	if (!blkif)
+		return ERR_PTR(-ENOMEM);
+
+	memset(blkif, 0, sizeof(*blkif));
+	blkif->domid = domid;
+	spin_lock_init(&blkif->blk_ring_lock);
+	atomic_set(&blkif->refcnt, 1);
+	init_waitqueue_head(&blkif->wq);
+	blkif->st_print = jiffies;
+	init_waitqueue_head(&blkif->waiting_to_free);
+
+	return blkif;
+}
+
+static int map_frontend_page(blkif_t *blkif, unsigned long shared_page)
+{
+	struct gnttab_map_grant_ref op;
+
+	gnttab_set_map_op(&op, (unsigned long)blkif->blk_ring_area->addr,
+			  GNTMAP_host_map, shared_page, blkif->domid);
+
+	if (HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref, &op, 1))
+		BUG();
+
+	if (op.status) {
+		DPRINTK(" Grant table operation failure !\n");
+		return op.status;
+	}
+
+	blkif->shmem_ref = shared_page;
+	blkif->shmem_handle = op.handle;
+
+	return 0;
+}
+
+static void unmap_frontend_page(blkif_t *blkif)
+{
+	struct gnttab_unmap_grant_ref op;
+
+	gnttab_set_unmap_op(&op, (unsigned long)blkif->blk_ring_area->addr,
+			    GNTMAP_host_map, blkif->shmem_handle);
+
+	if (HYPERVISOR_grant_table_op(GNTTABOP_unmap_grant_ref, &op, 1))
+		BUG();
+}
+
+int blkif_map(blkif_t *blkif, unsigned long shared_page, unsigned int evtchn)
+{
+	int err;
+
+	/* Already connected through? */
+	if (blkif->irq)
+		return 0;
+
+	if ( (blkif->blk_ring_area = alloc_vm_area(PAGE_SIZE)) == NULL )
+		return -ENOMEM;
+
+	err = map_frontend_page(blkif, shared_page);
+	if (err) {
+		free_vm_area(blkif->blk_ring_area);
+		return err;
+	}
+
+	switch (blkif->blk_protocol) {
+	case BLKIF_PROTOCOL_NATIVE:
+	{
+		blkif_sring_t *sring;
+		sring = (blkif_sring_t *)blkif->blk_ring_area->addr;
+		BACK_RING_INIT(&blkif->blk_rings.native, sring, PAGE_SIZE);
+		break;
+	}
+	case BLKIF_PROTOCOL_X86_32:
+	{
+		blkif_x86_32_sring_t *sring_x86_32;
+		sring_x86_32 = (blkif_x86_32_sring_t *)blkif->blk_ring_area->addr;
+		BACK_RING_INIT(&blkif->blk_rings.x86_32, sring_x86_32, PAGE_SIZE);
+		break;
+	}
+	case BLKIF_PROTOCOL_X86_64:
+	{
+		blkif_x86_64_sring_t *sring_x86_64;
+		sring_x86_64 = (blkif_x86_64_sring_t *)blkif->blk_ring_area->addr;
+		BACK_RING_INIT(&blkif->blk_rings.x86_64, sring_x86_64, PAGE_SIZE);
+		break;
+	}
+	default:
+		BUG();
+	}
+
+	err = bind_interdomain_evtchn_to_irqhandler(
+		blkif->domid, evtchn, blkif_be_int, 0, "blkif-backend", blkif);
+	if (err < 0)
+	{
+		unmap_frontend_page(blkif);
+		free_vm_area(blkif->blk_ring_area);
+		blkif->blk_rings.common.sring = NULL;
+		return err;
+	}
+	blkif->irq = err;
+
+	return 0;
+}
+
+void blkif_disconnect(blkif_t *blkif)
+{
+	if (blkif->xenblkd) {
+		kthread_stop(blkif->xenblkd);
+		blkif->xenblkd = NULL;
+	}
+
+	atomic_dec(&blkif->refcnt);
+	wait_event(blkif->waiting_to_free, atomic_read(&blkif->refcnt) == 0);
+	atomic_inc(&blkif->refcnt);
+
+	if (blkif->irq) {
+		unbind_from_irqhandler(blkif->irq, blkif);
+		blkif->irq = 0;
+	}
+
+	if (blkif->blk_rings.common.sring) {
+		unmap_frontend_page(blkif);
+		free_vm_area(blkif->blk_ring_area);
+		blkif->blk_rings.common.sring = NULL;
+	}
+}
+
+void blkif_free(blkif_t *blkif)
+{
+	if (!atomic_dec_and_test(&blkif->refcnt))
+		BUG();
+	kmem_cache_free(blkif_cachep, blkif);
+}
+
+void __init blkif_interface_init(void)
+{
+	blkif_cachep = kmem_cache_create("blkif_cache", sizeof(blkif_t),
+					 0, 0, NULL, NULL);
+}

drivers/xen/blkback/vbd.c

@@ -0,0 +1,118 @@
+/******************************************************************************
+ * blkback/vbd.c
+ *
+ * Routines for managing virtual block devices (VBDs).
+ *
+ * Copyright (c) 2003-2005, Keir Fraser & Steve Hand
+ *
+ * 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 "common.h"
+
+#define vbd_sz(_v)   ((_v)->bdev->bd_part ?				\
+	(_v)->bdev->bd_part->nr_sects : (_v)->bdev->bd_disk->capacity)
+
+unsigned long long vbd_size(struct vbd *vbd)
+{
+	return vbd_sz(vbd);
+}
+
+unsigned int vbd_info(struct vbd *vbd)
+{
+	return vbd->type | (vbd->readonly?VDISK_READONLY:0);
+}
+
+unsigned long vbd_secsize(struct vbd *vbd)
+{
+	return bdev_hardsect_size(vbd->bdev);
+}
+
+int vbd_create(blkif_t *blkif, blkif_vdev_t handle, unsigned major,
+	       unsigned minor, int readonly, int cdrom)
+{
+	struct vbd *vbd;
+	struct block_device *bdev;
+
+	vbd = &blkif->vbd;
+	vbd->handle   = handle;
+	vbd->readonly = readonly;
+	vbd->type     = 0;
+
+	vbd->pdevice  = MKDEV(major, minor);
+
+	bdev = open_by_devnum(vbd->pdevice,
+			      vbd->readonly ? FMODE_READ : FMODE_WRITE);
+
+	if (IS_ERR(bdev)) {
+		DPRINTK("vbd_creat: device %08x could not be opened.\n",
+			vbd->pdevice);
+		return -ENOENT;
+	}
+
+	vbd->bdev = bdev;
+
+	if (vbd->bdev->bd_disk == NULL) {
+		DPRINTK("vbd_creat: device %08x doesn't exist.\n",
+			vbd->pdevice);
+		vbd_free(vbd);
+		return -ENOENT;
+	}
+
+	if (vbd->bdev->bd_disk->flags & GENHD_FL_CD || cdrom)
+		vbd->type |= VDISK_CDROM;
+	if (vbd->bdev->bd_disk->flags & GENHD_FL_REMOVABLE)
+		vbd->type |= VDISK_REMOVABLE;
+
+	DPRINTK("Successful creation of handle=%04x (dom=%u)\n",
+		handle, blkif->domid);
+	return 0;
+}
+
+void vbd_free(struct vbd *vbd)
+{
+	if (vbd->bdev)
+		blkdev_put(vbd->bdev);
+	vbd->bdev = NULL;
+}
+
+int vbd_translate(struct phys_req *req, blkif_t *blkif, int operation)
+{
+	struct vbd *vbd = &blkif->vbd;
+	int rc = -EACCES;
+
+	if ((operation != READ) && vbd->readonly)
+		goto out;
+
+	if (unlikely((req->sector_number + req->nr_sects) > vbd_sz(vbd)))
+		goto out;
+
+	req->dev  = vbd->pdevice;
+	req->bdev = vbd->bdev;
+	rc = 0;
+
+ out:
+	return rc;
+}

drivers/xen/blkback/xenbus.c

@@ -0,0 +1,541 @@
+/*  Xenbus code for blkif backend
+    Copyright (C) 2005 Rusty Russell <rusty@rustcorp.com.au>
+    Copyright (C) 2005 XenSource Ltd
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License
+    along with this program; if not, write to the Free Software
+    Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+*/
+
+#include <stdarg.h>
+#include <linux/module.h>
+#include <linux/kthread.h>
+#include "common.h"
+
+#undef DPRINTK
+#define DPRINTK(fmt, args...)				\
+	pr_debug("blkback/xenbus (%s:%d) " fmt ".\n",	\
+		 __FUNCTION__, __LINE__, ##args)
+
+struct backend_info
+{
+	struct xenbus_device *dev;
+	blkif_t *blkif;
+	struct xenbus_watch backend_watch;
+	unsigned major;
+	unsigned minor;
+	char *mode;
+};
+
+static void connect(struct backend_info *);
+static int connect_ring(struct backend_info *);
+static void backend_changed(struct xenbus_watch *, const char **,
+			    unsigned int);
+
+static int blkback_name(blkif_t *blkif, char *buf)
+{
+	char *devpath, *devname;
+	struct xenbus_device *dev = blkif->be->dev;
+
+	devpath = xenbus_read(XBT_NIL, dev->nodename, "dev", NULL);
+	if (IS_ERR(devpath))
+		return PTR_ERR(devpath);
+
+	if ((devname = strstr(devpath, "/dev/")) != NULL)
+		devname += strlen("/dev/");
+	else
+		devname  = devpath;
+
+	snprintf(buf, TASK_COMM_LEN, "blkback.%d.%s", blkif->domid, devname);
+	kfree(devpath);
+
+	return 0;
+}
+
+static void update_blkif_status(blkif_t *blkif)
+{
+	int err;
+	char name[TASK_COMM_LEN];
+
+	/* Not ready to connect? */
+	if (!blkif->irq || !blkif->vbd.bdev)
+		return;
+
+	/* Already connected? */
+	if (blkif->be->dev->state == XenbusStateConnected)
+		return;
+
+	/* Attempt to connect: exit if we fail to. */
+	connect(blkif->be);
+	if (blkif->be->dev->state != XenbusStateConnected)
+		return;
+
+	err = blkback_name(blkif, name);
+	if (err) {
+		xenbus_dev_error(blkif->be->dev, err, "get blkback dev name");
+		return;
+	}
+
+	blkif->xenblkd = kthread_run(blkif_schedule, blkif, name);
+	if (IS_ERR(blkif->xenblkd)) {
+		err = PTR_ERR(blkif->xenblkd);
+		blkif->xenblkd = NULL;
+		xenbus_dev_error(blkif->be->dev, err, "start xenblkd");
+	}
+}
+
+
+/****************************************************************
+ *  sysfs interface for VBD I/O requests
+ */
+
+#define VBD_SHOW(name, format, args...)					\
+	static ssize_t show_##name(struct device *_dev,			\
+				   struct device_attribute *attr,	\
+				   char *buf)				\
+	{								\
+		struct xenbus_device *dev = to_xenbus_device(_dev);	\
+		struct backend_info *be = dev->dev.driver_data;		\
+									\
+		return sprintf(buf, format, ##args);			\
+	}								\
+	static DEVICE_ATTR(name, S_IRUGO, show_##name, NULL)
+
+VBD_SHOW(oo_req,  "%d\n", be->blkif->st_oo_req);
+VBD_SHOW(rd_req,  "%d\n", be->blkif->st_rd_req);
+VBD_SHOW(wr_req,  "%d\n", be->blkif->st_wr_req);
+VBD_SHOW(br_req,  "%d\n", be->blkif->st_br_req);
+VBD_SHOW(rd_sect, "%d\n", be->blkif->st_rd_sect);
+VBD_SHOW(wr_sect, "%d\n", be->blkif->st_wr_sect);
+
+static struct attribute *vbdstat_attrs[] = {
+	&dev_attr_oo_req.attr,
+	&dev_attr_rd_req.attr,
+	&dev_attr_wr_req.attr,
+	&dev_attr_br_req.attr,
+	&dev_attr_rd_sect.attr,
+	&dev_attr_wr_sect.attr,
+	NULL
+};
+
+static struct attribute_group vbdstat_group = {
+	.name = "statistics",
+	.attrs = vbdstat_attrs,
+};
+
+VBD_SHOW(physical_device, "%x:%x\n", be->major, be->minor);
+VBD_SHOW(mode, "%s\n", be->mode);
+
+int xenvbd_sysfs_addif(struct xenbus_device *dev)
+{
+	int error;
+
+	error = device_create_file(&dev->dev, &dev_attr_physical_device);
+ 	if (error)
+		goto fail1;
+
+	error = device_create_file(&dev->dev, &dev_attr_mode);
+	if (error)
+		goto fail2;
+
+	error = sysfs_create_group(&dev->dev.kobj, &vbdstat_group);
+	if (error)
+		goto fail3;
+
+	return 0;
+
+fail3:	sysfs_remove_group(&dev->dev.kobj, &vbdstat_group);
+fail2:	device_remove_file(&dev->dev, &dev_attr_mode);
+fail1:	device_remove_file(&dev->dev, &dev_attr_physical_device);
+	return error;
+}
+
+void xenvbd_sysfs_delif(struct xenbus_device *dev)
+{
+	sysfs_remove_group(&dev->dev.kobj, &vbdstat_group);
+	device_remove_file(&dev->dev, &dev_attr_mode);
+	device_remove_file(&dev->dev, &dev_attr_physical_device);
+}
+
+static int blkback_remove(struct xenbus_device *dev)
+{
+	struct backend_info *be = dev->dev.driver_data;
+
+	DPRINTK("");
+
+	if (be->major || be->minor)
+		xenvbd_sysfs_delif(dev);
+
+	if (be->backend_watch.node) {
+		unregister_xenbus_watch(&be->backend_watch);
+		kfree(be->backend_watch.node);
+		be->backend_watch.node = NULL;
+	}
+
+	if (be->blkif) {
+		blkif_disconnect(be->blkif);
+		vbd_free(&be->blkif->vbd);
+		blkif_free(be->blkif);
+		be->blkif = NULL;
+	}
+
+	kfree(be);
+	dev->dev.driver_data = NULL;
+	return 0;
+}
+
+int blkback_barrier(struct xenbus_transaction xbt,
+		    struct backend_info *be, int state)
+{
+	struct xenbus_device *dev = be->dev;
+	int err;
+
+	err = xenbus_printf(xbt, dev->nodename, "feature-barrier",
+			    "%d", state);
+	if (err)
+		xenbus_dev_fatal(dev, err, "writing feature-barrier");
+
+	return err;
+}
+
+/**
+ * Entry point to this code when a new device is created.  Allocate the basic
+ * structures, and watch the store waiting for the hotplug scripts to tell us
+ * the device's physical major and minor numbers.  Switch to InitWait.
+ */
+static int blkback_probe(struct xenbus_device *dev,
+			 const struct xenbus_device_id *id)
+{
+	int err;
+	struct backend_info *be = kzalloc(sizeof(struct backend_info),
+					  GFP_KERNEL);
+	if (!be) {
+		xenbus_dev_fatal(dev, -ENOMEM,
+				 "allocating backend structure");
+		return -ENOMEM;
+	}
+	be->dev = dev;
+	dev->dev.driver_data = be;
+
+	be->blkif = blkif_alloc(dev->otherend_id);
+	if (IS_ERR(be->blkif)) {
+		err = PTR_ERR(be->blkif);
+		be->blkif = NULL;
+		xenbus_dev_fatal(dev, err, "creating block interface");
+		goto fail;
+	}
+
+	/* setup back pointer */
+	be->blkif->be = be;
+
+	err = xenbus_watch_path2(dev, dev->nodename, "physical-device",
+				 &be->backend_watch, backend_changed);
+	if (err)
+		goto fail;
+
+	err = xenbus_switch_state(dev, XenbusStateInitWait);
+	if (err)
+		goto fail;
+
+	return 0;
+
+fail:
+	DPRINTK("failed");
+	blkback_remove(dev);
+	return err;
+}
+
+
+/**
+ * Callback received when the hotplug scripts have placed the physical-device
+ * node.  Read it and the mode node, and create a vbd.  If the frontend is
+ * ready, connect.
+ */
+static void backend_changed(struct xenbus_watch *watch,
+			    const char **vec, unsigned int len)
+{
+	int err;
+	unsigned major;
+	unsigned minor;
+	struct backend_info *be
+		= container_of(watch, struct backend_info, backend_watch);
+	struct xenbus_device *dev = be->dev;
+	int cdrom = 0;
+	char *device_type;
+
+	DPRINTK("");
+
+	err = xenbus_scanf(XBT_NIL, dev->nodename, "physical-device", "%x:%x",
+			   &major, &minor);
+	if (XENBUS_EXIST_ERR(err)) {
+		/* Since this watch will fire once immediately after it is
+		   registered, we expect this.  Ignore it, and wait for the
+		   hotplug scripts. */
+		return;
+	}
+	if (err != 2) {
+		xenbus_dev_fatal(dev, err, "reading physical-device");
+		return;
+	}
+
+	if ((be->major || be->minor) &&
+	    ((be->major != major) || (be->minor != minor))) {
+		printk(KERN_WARNING
+		       "blkback: changing physical device (from %x:%x to "
+		       "%x:%x) not supported.\n", be->major, be->minor,
+		       major, minor);
+		return;
+	}
+
+	be->mode = xenbus_read(XBT_NIL, dev->nodename, "mode", NULL);
+	if (IS_ERR(be->mode)) {
+		err = PTR_ERR(be->mode);
+		be->mode = NULL;
+		xenbus_dev_fatal(dev, err, "reading mode");
+		return;
+	}
+
+	device_type = xenbus_read(XBT_NIL, dev->otherend, "device-type", NULL);
+	if (!IS_ERR(device_type)) {
+		cdrom = strcmp(device_type, "cdrom") == 0;
+		kfree(device_type);
+	}
+
+	if (be->major == 0 && be->minor == 0) {
+		/* Front end dir is a number, which is used as the handle. */
+
+		char *p = strrchr(dev->otherend, '/') + 1;
+		long handle = simple_strtoul(p, NULL, 0);
+
+		be->major = major;
+		be->minor = minor;
+
+		err = vbd_create(be->blkif, handle, major, minor,
+				 (NULL == strchr(be->mode, 'w')), cdrom);
+		if (err) {
+			be->major = be->minor = 0;
+			xenbus_dev_fatal(dev, err, "creating vbd structure");
+			return;
+		}
+
+		err = xenvbd_sysfs_addif(dev);
+		if (err) {
+			vbd_free(&be->blkif->vbd);
+			be->major = be->minor = 0;
+			xenbus_dev_fatal(dev, err, "creating sysfs entries");
+			return;
+		}
+
+		/* We're potentially connected now */
+		update_blkif_status(be->blkif);
+	}
+}
+
+
+/**
+ * Callback received when the frontend's state changes.
+ */
+static void frontend_changed(struct xenbus_device *dev,
+			     enum xenbus_state frontend_state)
+{
+	struct backend_info *be = dev->dev.driver_data;
+	int err;
+
+	DPRINTK("%s", xenbus_strstate(frontend_state));
+
+	switch (frontend_state) {
+	case XenbusStateInitialising:
+		if (dev->state == XenbusStateClosed) {
+			printk(KERN_INFO "%s: %s: prepare for reconnect\n",
+			       __FUNCTION__, dev->nodename);
+			xenbus_switch_state(dev, XenbusStateInitWait);
+		}
+		break;
+
+	case XenbusStateInitialised:
+	case XenbusStateConnected:
+		/* Ensure we connect even when two watches fire in
+		   close successsion and we miss the intermediate value
+		   of frontend_state. */
+		if (dev->state == XenbusStateConnected)
+			break;
+
+		err = connect_ring(be);
+		if (err)
+			break;
+		update_blkif_status(be->blkif);
+		break;
+
+	case XenbusStateClosing:
+		blkif_disconnect(be->blkif);
+		xenbus_switch_state(dev, XenbusStateClosing);
+		break;
+
+	case XenbusStateClosed:
+		xenbus_switch_state(dev, XenbusStateClosed);
+		if (xenbus_dev_is_online(dev))
+			break;
+		/* fall through if not online */
+	case XenbusStateUnknown:
+		device_unregister(&dev->dev);
+		break;
+
+	default:
+		xenbus_dev_fatal(dev, -EINVAL, "saw state %d at frontend",
+				 frontend_state);
+		break;
+	}
+}
+
+
+/* ** Connection ** */
+
+
+/**
+ * Write the physical details regarding the block device to the store, and
+ * switch to Connected state.
+ */
+static void connect(struct backend_info *be)
+{
+	struct xenbus_transaction xbt;
+	int err;
+	struct xenbus_device *dev = be->dev;
+
+	DPRINTK("%s", dev->otherend);
+
+	/* Supply the information about the device the frontend needs */
+again:
+	err = xenbus_transaction_start(&xbt);
+	if (err) {
+		xenbus_dev_fatal(dev, err, "starting transaction");
+		return;
+	}
+
+	err = blkback_barrier(xbt, be, 1);
+	if (err)
+		goto abort;
+
+	err = xenbus_printf(xbt, dev->nodename, "sectors", "%llu",
+			    vbd_size(&be->blkif->vbd));
+	if (err) {
+		xenbus_dev_fatal(dev, err, "writing %s/sectors",
+				 dev->nodename);
+		goto abort;
+	}
+
+	/* FIXME: use a typename instead */
+	err = xenbus_printf(xbt, dev->nodename, "info", "%u",
+			    vbd_info(&be->blkif->vbd));
+	if (err) {
+		xenbus_dev_fatal(dev, err, "writing %s/info",
+				 dev->nodename);
+		goto abort;
+	}
+	err = xenbus_printf(xbt, dev->nodename, "sector-size", "%lu",
+			    vbd_secsize(&be->blkif->vbd));
+	if (err) {
+		xenbus_dev_fatal(dev, err, "writing %s/sector-size",
+				 dev->nodename);
+		goto abort;
+	}
+
+	err = xenbus_transaction_end(xbt, 0);
+	if (err == -EAGAIN)
+		goto again;
+	if (err)
+		xenbus_dev_fatal(dev, err, "ending transaction");
+
+	err = xenbus_switch_state(dev, XenbusStateConnected);
+	if (err)
+		xenbus_dev_fatal(dev, err, "switching to Connected state",
+				 dev->nodename);
+
+	return;
+ abort:
+	xenbus_transaction_end(xbt, 1);
+}
+
+
+static int connect_ring(struct backend_info *be)
+{
+	struct xenbus_device *dev = be->dev;
+	unsigned long ring_ref;
+	unsigned int evtchn;
+	char protocol[64] = "";
+	int err;
+
+	DPRINTK("%s", dev->otherend);
+
+	err = xenbus_gather(XBT_NIL, dev->otherend, "ring-ref", "%lu", &ring_ref,
+			    "event-channel", "%u", &evtchn, NULL);
+	if (err) {
+		xenbus_dev_fatal(dev, err,
+				 "reading %s/ring-ref and event-channel",
+				 dev->otherend);
+		return err;
+	}
+
+	be->blkif->blk_protocol = BLKIF_PROTOCOL_NATIVE;
+	err = xenbus_gather(XBT_NIL, dev->otherend, "protocol",
+			    "%63s", protocol, NULL);
+	if (err)
+		strcpy(protocol, "unspecified, assuming native");
+	else if (0 == strcmp(protocol, XEN_IO_PROTO_ABI_NATIVE))
+		be->blkif->blk_protocol = BLKIF_PROTOCOL_NATIVE;
+	else if (0 == strcmp(protocol, XEN_IO_PROTO_ABI_X86_32))
+		be->blkif->blk_protocol = BLKIF_PROTOCOL_X86_32;
+	else if (0 == strcmp(protocol, XEN_IO_PROTO_ABI_X86_64))
+		be->blkif->blk_protocol = BLKIF_PROTOCOL_X86_64;
+	else {
+		xenbus_dev_fatal(dev, err, "unknown fe protocol %s", protocol);
+		return -1;
+	}
+	printk(KERN_INFO
+	       "blkback: ring-ref %ld, event-channel %d, protocol %d (%s)\n",
+	       ring_ref, evtchn, be->blkif->blk_protocol, protocol);
+
+	/* Map the shared frame, irq etc. */
+	err = blkif_map(be->blkif, ring_ref, evtchn);
+	if (err) {
+		xenbus_dev_fatal(dev, err, "mapping ring-ref %lu port %u",
+				 ring_ref, evtchn);
+		return err;
+	}
+
+	return 0;
+}
+
+
+/* ** Driver Registration ** */
+
+
+static const struct xenbus_device_id blkback_ids[] = {
+	{ "vbd" },
+	{ "" }
+};
+
+
+static struct xenbus_driver blkback = {
+	.name = "vbd",
+	.owner = THIS_MODULE,
+	.ids = blkback_ids,
+	.probe = blkback_probe,
+	.remove = blkback_remove,
+	.otherend_changed = frontend_changed
+};
+
+
+void blkif_xenbus_init(void)
+{
+	xenbus_register_backend(&blkback);
+}

include/xen/blkif.h

@@ -0,0 +1,123 @@
+/*
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (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.
+ */
+
+#ifndef __XEN_BLKIF_H__
+#define __XEN_BLKIF_H__
+
+#include <xen/interface/io/ring.h>
+#include <xen/interface/io/blkif.h>
+#include <xen/interface/io/protocols.h>
+
+/* Not a real protocol.  Used to generate ring structs which contain
+ * the elements common to all protocols only.  This way we get a
+ * compiler-checkable way to use common struct elements, so we can
+ * avoid using switch(protocol) in a number of places.  */
+struct blkif_common_request {
+	char dummy;
+};
+struct blkif_common_response {
+	char dummy;
+};
+
+/* i386 protocol version */
+#pragma pack(push, 4)
+struct blkif_x86_32_request {
+	uint8_t        operation;    /* BLKIF_OP_???                         */
+	uint8_t        nr_segments;  /* number of segments                   */
+	blkif_vdev_t   handle;       /* only for read/write requests         */
+	uint64_t       id;           /* private guest value, echoed in resp  */
+	blkif_sector_t sector_number;/* start sector idx on disk (r/w only)  */
+	struct blkif_request_segment seg[BLKIF_MAX_SEGMENTS_PER_REQUEST];
+};
+struct blkif_x86_32_response {
+	uint64_t        id;              /* copied from request */
+	uint8_t         operation;       /* copied from request */
+	int16_t         status;          /* BLKIF_RSP_???       */
+};
+typedef struct blkif_x86_32_request blkif_x86_32_request_t;
+typedef struct blkif_x86_32_response blkif_x86_32_response_t;
+#pragma pack(pop)
+
+/* x86_64 protocol version */
+struct blkif_x86_64_request {
+	uint8_t        operation;    /* BLKIF_OP_???                         */
+	uint8_t        nr_segments;  /* number of segments                   */
+	blkif_vdev_t   handle;       /* only for read/write requests         */
+	uint64_t       __attribute__((__aligned__(8))) id;
+	blkif_sector_t sector_number;/* start sector idx on disk (r/w only)  */
+	struct blkif_request_segment seg[BLKIF_MAX_SEGMENTS_PER_REQUEST];
+};
+struct blkif_x86_64_response {
+	uint64_t       __attribute__((__aligned__(8))) id;
+	uint8_t         operation;       /* copied from request */
+	int16_t         status;          /* BLKIF_RSP_???       */
+};
+typedef struct blkif_x86_64_request blkif_x86_64_request_t;
+typedef struct blkif_x86_64_response blkif_x86_64_response_t;
+
+DEFINE_RING_TYPES(blkif_common, struct blkif_common_request, struct blkif_common_response);
+DEFINE_RING_TYPES(blkif_x86_32, struct blkif_x86_32_request, struct blkif_x86_32_response);
+DEFINE_RING_TYPES(blkif_x86_64, struct blkif_x86_64_request, struct blkif_x86_64_response);
+
+union blkif_back_rings {
+	blkif_back_ring_t        native;
+	blkif_common_back_ring_t common;
+	blkif_x86_32_back_ring_t x86_32;
+	blkif_x86_64_back_ring_t x86_64;
+};
+typedef union blkif_back_rings blkif_back_rings_t;
+
+enum blkif_protocol {
+	BLKIF_PROTOCOL_NATIVE = 1,
+	BLKIF_PROTOCOL_X86_32 = 2,
+	BLKIF_PROTOCOL_X86_64 = 3,
+};
+
+static void inline blkif_get_x86_32_req(blkif_request_t *dst, blkif_x86_32_request_t *src)
+{
+	int i, n = BLKIF_MAX_SEGMENTS_PER_REQUEST;
+	dst->operation = src->operation;
+	dst->nr_segments = src->nr_segments;
+	dst->handle = src->handle;
+	dst->id = src->id;
+	dst->sector_number = src->sector_number;
+	barrier();
+	if (n > dst->nr_segments)
+		n = dst->nr_segments;
+	for (i = 0; i < n; i++)
+		dst->seg[i] = src->seg[i];
+}
+
+static void inline blkif_get_x86_64_req(blkif_request_t *dst, blkif_x86_64_request_t *src)
+{
+	int i, n = BLKIF_MAX_SEGMENTS_PER_REQUEST;
+	dst->operation = src->operation;
+	dst->nr_segments = src->nr_segments;
+	dst->handle = src->handle;
+	dst->id = src->id;
+	dst->sector_number = src->sector_number;
+	barrier();
+	if (n > dst->nr_segments)
+		n = dst->nr_segments;
+	for (i = 0; i < n; i++)
+		dst->seg[i] = src->seg[i];
+}
+
+#endif /* __XEN_BLKIF_H__ */