[people/ms/linux.git] / net / 9p / trans_xen.c

// SPDX-License-Identifier: GPL-2.0-only
/*
 * linux/fs/9p/trans_xen
 *
 * Xen transport layer.
 *
 * Copyright (C) 2017 by Stefano Stabellini <stefano@aporeto.com>
 */

#include <xen/events.h>
#include <xen/grant_table.h>
#include <xen/xen.h>
#include <xen/xenbus.h>
#include <xen/interface/io/9pfs.h>

#include <linux/module.h>
#include <linux/spinlock.h>
#include <net/9p/9p.h>
#include <net/9p/client.h>
#include <net/9p/transport.h>

#define XEN_9PFS_NUM_RINGS 2
#define XEN_9PFS_RING_ORDER 9
#define XEN_9PFS_RING_SIZE(ring)  XEN_FLEX_RING_SIZE(ring->intf->ring_order)

struct xen_9pfs_header {
	uint32_t size;
	uint8_t id;
	uint16_t tag;

	/* uint8_t sdata[]; */
} __attribute__((packed));

/* One per ring, more than one per 9pfs share */
struct xen_9pfs_dataring {
	struct xen_9pfs_front_priv *priv;

	struct xen_9pfs_data_intf *intf;
	grant_ref_t ref;
	int evtchn;
	int irq;
	/* protect a ring from concurrent accesses */
	spinlock_t lock;

	struct xen_9pfs_data data;
	wait_queue_head_t wq;
	struct work_struct work;
};

/* One per 9pfs share */
struct xen_9pfs_front_priv {
	struct list_head list;
	struct xenbus_device *dev;
	char *tag;
	struct p9_client *client;

	int num_rings;
	struct xen_9pfs_dataring *rings;
};

static LIST_HEAD(xen_9pfs_devs);
static DEFINE_RWLOCK(xen_9pfs_lock);

/* We don't currently allow canceling of requests */
static int p9_xen_cancel(struct p9_client *client, struct p9_req_t *req)
{
	return 1;
}

static int p9_xen_create(struct p9_client *client, const char *addr, char *args)
{
	struct xen_9pfs_front_priv *priv;

	if (addr == NULL)
		return -EINVAL;

	read_lock(&xen_9pfs_lock);
	list_for_each_entry(priv, &xen_9pfs_devs, list) {
		if (!strcmp(priv->tag, addr)) {
			priv->client = client;
			read_unlock(&xen_9pfs_lock);
			return 0;
		}
	}
	read_unlock(&xen_9pfs_lock);
	return -EINVAL;
}

static void p9_xen_close(struct p9_client *client)
{
	struct xen_9pfs_front_priv *priv;

	read_lock(&xen_9pfs_lock);
	list_for_each_entry(priv, &xen_9pfs_devs, list) {
		if (priv->client == client) {
			priv->client = NULL;
			read_unlock(&xen_9pfs_lock);
			return;
		}
	}
	read_unlock(&xen_9pfs_lock);
}

static bool p9_xen_write_todo(struct xen_9pfs_dataring *ring, RING_IDX size)
{
	RING_IDX cons, prod;

	cons = ring->intf->out_cons;
	prod = ring->intf->out_prod;
	virt_mb();

	return XEN_9PFS_RING_SIZE(ring) -
		xen_9pfs_queued(prod, cons, XEN_9PFS_RING_SIZE(ring)) >= size;
}

static int p9_xen_request(struct p9_client *client, struct p9_req_t *p9_req)
{
	struct xen_9pfs_front_priv *priv;
	RING_IDX cons, prod, masked_cons, masked_prod;
	unsigned long flags;
	u32 size = p9_req->tc.size;
	struct xen_9pfs_dataring *ring;
	int num;

	read_lock(&xen_9pfs_lock);
	list_for_each_entry(priv, &xen_9pfs_devs, list) {
		if (priv->client == client)
			break;
	}
	read_unlock(&xen_9pfs_lock);
	if (list_entry_is_head(priv, &xen_9pfs_devs, list))
		return -EINVAL;

	num = p9_req->tc.tag % priv->num_rings;
	ring = &priv->rings[num];

again:
	while (wait_event_killable(ring->wq,
				   p9_xen_write_todo(ring, size)) != 0)
		;

	spin_lock_irqsave(&ring->lock, flags);
	cons = ring->intf->out_cons;
	prod = ring->intf->out_prod;
	virt_mb();

	if (XEN_9PFS_RING_SIZE(ring) -
	    xen_9pfs_queued(prod, cons, XEN_9PFS_RING_SIZE(ring)) < size) {
		spin_unlock_irqrestore(&ring->lock, flags);
		goto again;
	}

	masked_prod = xen_9pfs_mask(prod, XEN_9PFS_RING_SIZE(ring));
	masked_cons = xen_9pfs_mask(cons, XEN_9PFS_RING_SIZE(ring));

	xen_9pfs_write_packet(ring->data.out, p9_req->tc.sdata, size,
			      &masked_prod, masked_cons,
			      XEN_9PFS_RING_SIZE(ring));

	p9_req->status = REQ_STATUS_SENT;
	virt_wmb();			/* write ring before updating pointer */
	prod += size;
	ring->intf->out_prod = prod;
	spin_unlock_irqrestore(&ring->lock, flags);
	notify_remote_via_irq(ring->irq);
	p9_req_put(p9_req);

	return 0;
}

static void p9_xen_response(struct work_struct *work)
{
	struct xen_9pfs_front_priv *priv;
	struct xen_9pfs_dataring *ring;
	RING_IDX cons, prod, masked_cons, masked_prod;
	struct xen_9pfs_header h;
	struct p9_req_t *req;
	int status;

	ring = container_of(work, struct xen_9pfs_dataring, work);
	priv = ring->priv;

	while (1) {
		cons = ring->intf->in_cons;
		prod = ring->intf->in_prod;
		virt_rmb();

		if (xen_9pfs_queued(prod, cons, XEN_9PFS_RING_SIZE(ring)) <
		    sizeof(h)) {
			notify_remote_via_irq(ring->irq);
			return;
		}

		masked_prod = xen_9pfs_mask(prod, XEN_9PFS_RING_SIZE(ring));
		masked_cons = xen_9pfs_mask(cons, XEN_9PFS_RING_SIZE(ring));

		/* First, read just the header */
		xen_9pfs_read_packet(&h, ring->data.in, sizeof(h),
				     masked_prod, &masked_cons,
				     XEN_9PFS_RING_SIZE(ring));

		req = p9_tag_lookup(priv->client, h.tag);
		if (!req || req->status != REQ_STATUS_SENT) {
			dev_warn(&priv->dev->dev, "Wrong req tag=%x\n", h.tag);
			cons += h.size;
			virt_mb();
			ring->intf->in_cons = cons;
			continue;
		}

		memcpy(&req->rc, &h, sizeof(h));
		req->rc.offset = 0;

		masked_cons = xen_9pfs_mask(cons, XEN_9PFS_RING_SIZE(ring));
		/* Then, read the whole packet (including the header) */
		xen_9pfs_read_packet(req->rc.sdata, ring->data.in, h.size,
				     masked_prod, &masked_cons,
				     XEN_9PFS_RING_SIZE(ring));

		virt_mb();
		cons += h.size;
		ring->intf->in_cons = cons;

		status = (req->status != REQ_STATUS_ERROR) ?
			REQ_STATUS_RCVD : REQ_STATUS_ERROR;

		p9_client_cb(priv->client, req, status);
	}
}

static irqreturn_t xen_9pfs_front_event_handler(int irq, void *r)
{
	struct xen_9pfs_dataring *ring = r;

	if (!ring || !ring->priv->client) {
		/* ignore spurious interrupt */
		return IRQ_HANDLED;
	}

	wake_up_interruptible(&ring->wq);
	schedule_work(&ring->work);

	return IRQ_HANDLED;
}

static struct p9_trans_module p9_xen_trans = {
	.name = "xen",
	.maxsize = 1 << (XEN_9PFS_RING_ORDER + XEN_PAGE_SHIFT - 2),
	.def = 1,
	.create = p9_xen_create,
	.close = p9_xen_close,
	.request = p9_xen_request,
	.cancel = p9_xen_cancel,
	.owner = THIS_MODULE,
};

static const struct xenbus_device_id xen_9pfs_front_ids[] = {
	{ "9pfs" },
	{ "" }
};

static void xen_9pfs_front_free(struct xen_9pfs_front_priv *priv)
{
	int i, j;

	write_lock(&xen_9pfs_lock);
	list_del(&priv->list);
	write_unlock(&xen_9pfs_lock);

	for (i = 0; i < priv->num_rings; i++) {
		if (!priv->rings[i].intf)
			break;
		if (priv->rings[i].irq > 0)
			unbind_from_irqhandler(priv->rings[i].irq, priv->dev);
		if (priv->rings[i].data.in) {
			for (j = 0;
			     j < (1 << priv->rings[i].intf->ring_order);
			     j++) {
				grant_ref_t ref;

				ref = priv->rings[i].intf->ref[j];
				gnttab_end_foreign_access(ref, NULL);
			}
			free_pages_exact(priv->rings[i].data.in,
				   1UL << (priv->rings[i].intf->ring_order +
					   XEN_PAGE_SHIFT));
		}
		gnttab_end_foreign_access(priv->rings[i].ref, NULL);
		free_page((unsigned long)priv->rings[i].intf);
	}
	kfree(priv->rings);
	kfree(priv->tag);
	kfree(priv);
}

static int xen_9pfs_front_remove(struct xenbus_device *dev)
{
	struct xen_9pfs_front_priv *priv = dev_get_drvdata(&dev->dev);

	dev_set_drvdata(&dev->dev, NULL);
	xen_9pfs_front_free(priv);
	return 0;
}

static int xen_9pfs_front_alloc_dataring(struct xenbus_device *dev,
					 struct xen_9pfs_dataring *ring,
					 unsigned int order)
{
	int i = 0;
	int ret = -ENOMEM;
	void *bytes = NULL;

	init_waitqueue_head(&ring->wq);
	spin_lock_init(&ring->lock);
	INIT_WORK(&ring->work, p9_xen_response);

	ring->intf = (struct xen_9pfs_data_intf *)get_zeroed_page(GFP_KERNEL);
	if (!ring->intf)
		return ret;
	ret = gnttab_grant_foreign_access(dev->otherend_id,
					  virt_to_gfn(ring->intf), 0);
	if (ret < 0)
		goto out;
	ring->ref = ret;
	bytes = alloc_pages_exact(1UL << (order + XEN_PAGE_SHIFT),
				  GFP_KERNEL | __GFP_ZERO);
	if (!bytes) {
		ret = -ENOMEM;
		goto out;
	}
	for (; i < (1 << order); i++) {
		ret = gnttab_grant_foreign_access(
				dev->otherend_id, virt_to_gfn(bytes) + i, 0);
		if (ret < 0)
			goto out;
		ring->intf->ref[i] = ret;
	}
	ring->intf->ring_order = order;
	ring->data.in = bytes;
	ring->data.out = bytes + XEN_FLEX_RING_SIZE(order);

	ret = xenbus_alloc_evtchn(dev, &ring->evtchn);
	if (ret)
		goto out;
	ring->irq = bind_evtchn_to_irqhandler(ring->evtchn,
					      xen_9pfs_front_event_handler,
					      0, "xen_9pfs-frontend", ring);
	if (ring->irq >= 0)
		return 0;

	xenbus_free_evtchn(dev, ring->evtchn);
	ret = ring->irq;
out:
	if (bytes) {
		for (i--; i >= 0; i--)
			gnttab_end_foreign_access(ring->intf->ref[i], NULL);
		free_pages_exact(bytes, 1UL << (order + XEN_PAGE_SHIFT));
	}
	gnttab_end_foreign_access(ring->ref, NULL);
	free_page((unsigned long)ring->intf);
	return ret;
}

static int xen_9pfs_front_probe(struct xenbus_device *dev,
				const struct xenbus_device_id *id)
{
	int ret, i;
	struct xenbus_transaction xbt;
	struct xen_9pfs_front_priv *priv = NULL;
	char *versions;
	unsigned int max_rings, max_ring_order, len = 0;

	versions = xenbus_read(XBT_NIL, dev->otherend, "versions", &len);
	if (IS_ERR(versions))
		return PTR_ERR(versions);
	if (strcmp(versions, "1")) {
		kfree(versions);
		return -EINVAL;
	}
	kfree(versions);
	max_rings = xenbus_read_unsigned(dev->otherend, "max-rings", 0);
	if (max_rings < XEN_9PFS_NUM_RINGS)
		return -EINVAL;
	max_ring_order = xenbus_read_unsigned(dev->otherend,
					      "max-ring-page-order", 0);
	if (max_ring_order > XEN_9PFS_RING_ORDER)
		max_ring_order = XEN_9PFS_RING_ORDER;
	if (p9_xen_trans.maxsize > XEN_FLEX_RING_SIZE(max_ring_order))
		p9_xen_trans.maxsize = XEN_FLEX_RING_SIZE(max_ring_order) / 2;

	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
	if (!priv)
		return -ENOMEM;

	priv->dev = dev;
	priv->num_rings = XEN_9PFS_NUM_RINGS;
	priv->rings = kcalloc(priv->num_rings, sizeof(*priv->rings),
			      GFP_KERNEL);
	if (!priv->rings) {
		kfree(priv);
		return -ENOMEM;
	}

	for (i = 0; i < priv->num_rings; i++) {
		priv->rings[i].priv = priv;
		ret = xen_9pfs_front_alloc_dataring(dev, &priv->rings[i],
						    max_ring_order);
		if (ret < 0)
			goto error;
	}

 again:
	ret = xenbus_transaction_start(&xbt);
	if (ret) {
		xenbus_dev_fatal(dev, ret, "starting transaction");
		goto error;
	}
	ret = xenbus_printf(xbt, dev->nodename, "version", "%u", 1);
	if (ret)
		goto error_xenbus;
	ret = xenbus_printf(xbt, dev->nodename, "num-rings", "%u",
			    priv->num_rings);
	if (ret)
		goto error_xenbus;
	for (i = 0; i < priv->num_rings; i++) {
		char str[16];

		BUILD_BUG_ON(XEN_9PFS_NUM_RINGS > 9);
		sprintf(str, "ring-ref%d", i);
		ret = xenbus_printf(xbt, dev->nodename, str, "%d",
				    priv->rings[i].ref);
		if (ret)
			goto error_xenbus;

		sprintf(str, "event-channel-%d", i);
		ret = xenbus_printf(xbt, dev->nodename, str, "%u",
				    priv->rings[i].evtchn);
		if (ret)
			goto error_xenbus;
	}
	priv->tag = xenbus_read(xbt, dev->nodename, "tag", NULL);
	if (IS_ERR(priv->tag)) {
		ret = PTR_ERR(priv->tag);
		goto error_xenbus;
	}
	ret = xenbus_transaction_end(xbt, 0);
	if (ret) {
		if (ret == -EAGAIN)
			goto again;
		xenbus_dev_fatal(dev, ret, "completing transaction");
		goto error;
	}

	write_lock(&xen_9pfs_lock);
	list_add_tail(&priv->list, &xen_9pfs_devs);
	write_unlock(&xen_9pfs_lock);
	dev_set_drvdata(&dev->dev, priv);
	xenbus_switch_state(dev, XenbusStateInitialised);

	return 0;

 error_xenbus:
	xenbus_transaction_end(xbt, 1);
	xenbus_dev_fatal(dev, ret, "writing xenstore");
 error:
	dev_set_drvdata(&dev->dev, NULL);
	xen_9pfs_front_free(priv);
	return ret;
}

static int xen_9pfs_front_resume(struct xenbus_device *dev)
{
	dev_warn(&dev->dev, "suspend/resume unsupported\n");
	return 0;
}

static void xen_9pfs_front_changed(struct xenbus_device *dev,
				   enum xenbus_state backend_state)
{
	switch (backend_state) {
	case XenbusStateReconfiguring:
	case XenbusStateReconfigured:
	case XenbusStateInitialising:
	case XenbusStateInitialised:
	case XenbusStateUnknown:
		break;

	case XenbusStateInitWait:
		break;

	case XenbusStateConnected:
		xenbus_switch_state(dev, XenbusStateConnected);
		break;

	case XenbusStateClosed:
		if (dev->state == XenbusStateClosed)
			break;
		fallthrough;	/* Missed the backend's CLOSING state */
	case XenbusStateClosing:
		xenbus_frontend_closed(dev);
		break;
	}
}

static struct xenbus_driver xen_9pfs_front_driver = {
	.ids = xen_9pfs_front_ids,
	.probe = xen_9pfs_front_probe,
	.remove = xen_9pfs_front_remove,
	.resume = xen_9pfs_front_resume,
	.otherend_changed = xen_9pfs_front_changed,
};

static int p9_trans_xen_init(void)
{
	int rc;

	if (!xen_domain())
		return -ENODEV;

	pr_info("Initialising Xen transport for 9pfs\n");

	v9fs_register_trans(&p9_xen_trans);
	rc = xenbus_register_frontend(&xen_9pfs_front_driver);
	if (rc)
		v9fs_unregister_trans(&p9_xen_trans);

	return rc;
}
module_init(p9_trans_xen_init);
MODULE_ALIAS_9P("xen");

static void p9_trans_xen_exit(void)
{
	v9fs_unregister_trans(&p9_xen_trans);
	return xenbus_unregister_driver(&xen_9pfs_front_driver);
}
module_exit(p9_trans_xen_exit);

MODULE_ALIAS("xen:9pfs");
MODULE_AUTHOR("Stefano Stabellini <stefano@aporeto.com>");
MODULE_DESCRIPTION("Xen Transport for 9P");
MODULE_LICENSE("GPL");
Commit	Line	Data
024b7d6a	1	// SPDX-License-Identifier: GPL-2.0-only
868eb122 SS	2	/*
	3	* linux/fs/9p/trans_xen
	4	*
	5	* Xen transport layer.
	6	*
	7	* Copyright (C) 2017 by Stefano Stabellini <stefano@aporeto.com>
868eb122 SS	8	*/
	9
	10	#include <xen/events.h>
	11	#include <xen/grant_table.h>
	12	#include <xen/xen.h>
	13	#include <xen/xenbus.h>
	14	#include <xen/interface/io/9pfs.h>
	15
	16	#include <linux/module.h>
71ebd719	17	#include <linux/spinlock.h>
868eb122 SS	18	#include <net/9p/9p.h>
	19	#include <net/9p/client.h>
	20	#include <net/9p/transport.h>
	21
71ebd719	22	#define XEN_9PFS_NUM_RINGS 2
36f99675 SS	23	#define XEN_9PFS_RING_ORDER 9
36f99675 SS	24	#define XEN_9PFS_RING_SIZE(ring) XEN_FLEX_RING_SIZE(ring->intf->ring_order)
71ebd719 SS	25
	26	struct xen_9pfs_header {
	27	uint32_t size;
	28	uint8_t id;
	29	uint16_t tag;
	30
	31	/* uint8_t sdata[]; */
	32	} __attribute__((packed));
	33
	34	/* One per ring, more than one per 9pfs share */
	35	struct xen_9pfs_dataring {
	36	struct xen_9pfs_front_priv *priv;
	37
	38	struct xen_9pfs_data_intf *intf;
	39	grant_ref_t ref;
	40	int evtchn;
	41	int irq;
	42	/* protect a ring from concurrent accesses */
	43	spinlock_t lock;
	44
	45	struct xen_9pfs_data data;
	46	wait_queue_head_t wq;
	47	struct work_struct work;
	48	};
	49
	50	/* One per 9pfs share */
	51	struct xen_9pfs_front_priv {
	52	struct list_head list;
	53	struct xenbus_device *dev;
	54	char *tag;
	55	struct p9_client *client;
	56
	57	int num_rings;
	58	struct xen_9pfs_dataring *rings;
	59	};
	60
	61	static LIST_HEAD(xen_9pfs_devs);
	62	static DEFINE_RWLOCK(xen_9pfs_lock);
	63
f023f18d	64	/* We don't currently allow canceling of requests */
868eb122 SS	65	static int p9_xen_cancel(struct p9_client client, struct p9_req_t req)
868eb122 SS	66	{
f023f18d	67	return 1;
868eb122 SS	68	}
	69
	70	static int p9_xen_create(struct p9_client client, const char addr, char *args)
	71	{
f023f18d SS	72	struct xen_9pfs_front_priv *priv;
f023f18d SS	73
10aa1452 TB	74	if (addr == NULL)
	75	return -EINVAL;
	76
f023f18d SS	77	read_lock(&xen_9pfs_lock);
	78	list_for_each_entry(priv, &xen_9pfs_devs, list) {
	79	if (!strcmp(priv->tag, addr)) {
	80	priv->client = client;
	81	read_unlock(&xen_9pfs_lock);
	82	return 0;
	83	}
	84	}
	85	read_unlock(&xen_9pfs_lock);
	86	return -EINVAL;
868eb122 SS	87	}
	88
	89	static void p9_xen_close(struct p9_client *client)
	90	{
f023f18d SS	91	struct xen_9pfs_front_priv *priv;
	92
	93	read_lock(&xen_9pfs_lock);
	94	list_for_each_entry(priv, &xen_9pfs_devs, list) {
	95	if (priv->client == client) {
	96	priv->client = NULL;
	97	read_unlock(&xen_9pfs_lock);
	98	return;
	99	}
	100	}
	101	read_unlock(&xen_9pfs_lock);
	102	}
	103
	104	static bool p9_xen_write_todo(struct xen_9pfs_dataring *ring, RING_IDX size)
	105	{
	106	RING_IDX cons, prod;
	107
	108	cons = ring->intf->out_cons;
	109	prod = ring->intf->out_prod;
	110	virt_mb();
	111
36f99675 SS	112	return XEN_9PFS_RING_SIZE(ring) -
36f99675 SS	113	xen_9pfs_queued(prod, cons, XEN_9PFS_RING_SIZE(ring)) >= size;
868eb122 SS	114	}
	115
	116	static int p9_xen_request(struct p9_client client, struct p9_req_t p9_req)
	117	{
732b33d0	118	struct xen_9pfs_front_priv *priv;
f023f18d SS	119	RING_IDX cons, prod, masked_cons, masked_prod;
f023f18d SS	120	unsigned long flags;
523adb6c	121	u32 size = p9_req->tc.size;
f023f18d SS	122	struct xen_9pfs_dataring *ring;
	123	int num;
	124
	125	read_lock(&xen_9pfs_lock);
	126	list_for_each_entry(priv, &xen_9pfs_devs, list) {
	127	if (priv->client == client)
	128	break;
	129	}
	130	read_unlock(&xen_9pfs_lock);
732b33d0	131	if (list_entry_is_head(priv, &xen_9pfs_devs, list))
f023f18d SS	132	return -EINVAL;
f023f18d SS	133
523adb6c	134	num = p9_req->tc.tag % priv->num_rings;
f023f18d SS	135	ring = &priv->rings[num];
	136
	137	again:
9523feac TT	138	while (wait_event_killable(ring->wq,
9523feac TT	139	p9_xen_write_todo(ring, size)) != 0)
f023f18d SS	140	;
	141
	142	spin_lock_irqsave(&ring->lock, flags);
	143	cons = ring->intf->out_cons;
	144	prod = ring->intf->out_prod;
	145	virt_mb();
	146
36f99675 SS	147	if (XEN_9PFS_RING_SIZE(ring) -
36f99675 SS	148	xen_9pfs_queued(prod, cons, XEN_9PFS_RING_SIZE(ring)) < size) {
f023f18d SS	149	spin_unlock_irqrestore(&ring->lock, flags);
	150	goto again;
	151	}
	152
36f99675 SS	153	masked_prod = xen_9pfs_mask(prod, XEN_9PFS_RING_SIZE(ring));
36f99675 SS	154	masked_cons = xen_9pfs_mask(cons, XEN_9PFS_RING_SIZE(ring));
f023f18d	155
523adb6c	156	xen_9pfs_write_packet(ring->data.out, p9_req->tc.sdata, size,
36f99675 SS	157	&masked_prod, masked_cons,
36f99675 SS	158	XEN_9PFS_RING_SIZE(ring));
f023f18d SS	159
	160	p9_req->status = REQ_STATUS_SENT;
	161	virt_wmb(); /* write ring before updating pointer */
	162	prod += size;
	163	ring->intf->out_prod = prod;
	164	spin_unlock_irqrestore(&ring->lock, flags);
	165	notify_remote_via_irq(ring->irq);
728356de	166	p9_req_put(p9_req);
f023f18d	167
868eb122 SS	168	return 0;
	169	}
	170
71ebd719 SS	171	static void p9_xen_response(struct work_struct *work)
71ebd719 SS	172	{
f66c72be SS	173	struct xen_9pfs_front_priv *priv;
	174	struct xen_9pfs_dataring *ring;
	175	RING_IDX cons, prod, masked_cons, masked_prod;
	176	struct xen_9pfs_header h;
	177	struct p9_req_t *req;
	178	int status;
	179
	180	ring = container_of(work, struct xen_9pfs_dataring, work);
	181	priv = ring->priv;
	182
	183	while (1) {
	184	cons = ring->intf->in_cons;
	185	prod = ring->intf->in_prod;
	186	virt_rmb();
	187
36f99675	188	if (xen_9pfs_queued(prod, cons, XEN_9PFS_RING_SIZE(ring)) <
f66c72be SS	189	sizeof(h)) {
	190	notify_remote_via_irq(ring->irq);
	191	return;
	192	}
	193
36f99675 SS	194	masked_prod = xen_9pfs_mask(prod, XEN_9PFS_RING_SIZE(ring));
36f99675 SS	195	masked_cons = xen_9pfs_mask(cons, XEN_9PFS_RING_SIZE(ring));
f66c72be SS	196
	197	/* First, read just the header */
	198	xen_9pfs_read_packet(&h, ring->data.in, sizeof(h),
	199	masked_prod, &masked_cons,
36f99675	200	XEN_9PFS_RING_SIZE(ring));
f66c72be SS	201
	202	req = p9_tag_lookup(priv->client, h.tag);
	203	if (!req \|\| req->status != REQ_STATUS_SENT) {
	204	dev_warn(&priv->dev->dev, "Wrong req tag=%x\n", h.tag);
	205	cons += h.size;
	206	virt_mb();
	207	ring->intf->in_cons = cons;
	208	continue;
	209	}
	210
523adb6c DM	211	memcpy(&req->rc, &h, sizeof(h));
523adb6c DM	212	req->rc.offset = 0;
f66c72be	213
36f99675	214	masked_cons = xen_9pfs_mask(cons, XEN_9PFS_RING_SIZE(ring));
f66c72be	215	/* Then, read the whole packet (including the header) */
523adb6c	216	xen_9pfs_read_packet(req->rc.sdata, ring->data.in, h.size,
f66c72be	217	masked_prod, &masked_cons,
36f99675	218	XEN_9PFS_RING_SIZE(ring));
f66c72be SS	219
	220	virt_mb();
	221	cons += h.size;
	222	ring->intf->in_cons = cons;
	223
	224	status = (req->status != REQ_STATUS_ERROR) ?
	225	REQ_STATUS_RCVD : REQ_STATUS_ERROR;
	226
	227	p9_client_cb(priv->client, req, status);
	228	}
71ebd719 SS	229	}
	230
	231	static irqreturn_t xen_9pfs_front_event_handler(int irq, void *r)
	232	{
	233	struct xen_9pfs_dataring *ring = r;
	234
	235	if (!ring \|\| !ring->priv->client) {
	236	/* ignore spurious interrupt */
	237	return IRQ_HANDLED;
	238	}
	239
	240	wake_up_interruptible(&ring->wq);
	241	schedule_work(&ring->work);
	242
	243	return IRQ_HANDLED;
	244	}
	245
868eb122 SS	246	static struct p9_trans_module p9_xen_trans = {
868eb122 SS	247	.name = "xen",
36f99675	248	.maxsize = 1 << (XEN_9PFS_RING_ORDER + XEN_PAGE_SHIFT - 2),
868eb122 SS	249	.def = 1,
	250	.create = p9_xen_create,
	251	.close = p9_xen_close,
	252	.request = p9_xen_request,
	253	.cancel = p9_xen_cancel,
	254	.owner = THIS_MODULE,
	255	};
	256
	257	static const struct xenbus_device_id xen_9pfs_front_ids[] = {
	258	{ "9pfs" },
	259	{ "" }
	260	};
	261
71ebd719 SS	262	static void xen_9pfs_front_free(struct xen_9pfs_front_priv *priv)
	263	{
	264	int i, j;
	265
	266	write_lock(&xen_9pfs_lock);
	267	list_del(&priv->list);
	268	write_unlock(&xen_9pfs_lock);
	269
	270	for (i = 0; i < priv->num_rings; i++) {
	271	if (!priv->rings[i].intf)
	272	break;
	273	if (priv->rings[i].irq > 0)
	274	unbind_from_irqhandler(priv->rings[i].irq, priv->dev);
	275	if (priv->rings[i].data.in) {
36f99675 SS	276	for (j = 0;
	277	j < (1 << priv->rings[i].intf->ring_order);
	278	j++) {
71ebd719 SS	279	grant_ref_t ref;
	280
	281	ref = priv->rings[i].intf->ref[j];
49f8b459	282	gnttab_end_foreign_access(ref, NULL);
71ebd719	283	}
5cadd4bb JG	284	free_pages_exact(priv->rings[i].data.in,
	285	1UL << (priv->rings[i].intf->ring_order +
	286	XEN_PAGE_SHIFT));
71ebd719	287	}
49f8b459	288	gnttab_end_foreign_access(priv->rings[i].ref, NULL);
71ebd719 SS	289	free_page((unsigned long)priv->rings[i].intf);
	290	}
	291	kfree(priv->rings);
	292	kfree(priv->tag);
	293	kfree(priv);
	294	}
	295
868eb122 SS	296	static int xen_9pfs_front_remove(struct xenbus_device *dev)
868eb122 SS	297	{
71ebd719 SS	298	struct xen_9pfs_front_priv *priv = dev_get_drvdata(&dev->dev);
	299
	300	dev_set_drvdata(&dev->dev, NULL);
	301	xen_9pfs_front_free(priv);
868eb122 SS	302	return 0;
	303	}
	304
71ebd719	305	static int xen_9pfs_front_alloc_dataring(struct xenbus_device *dev,
36f99675 SS	306	struct xen_9pfs_dataring *ring,
36f99675 SS	307	unsigned int order)
71ebd719 SS	308	{
	309	int i = 0;
	310	int ret = -ENOMEM;
	311	void *bytes = NULL;
	312
	313	init_waitqueue_head(&ring->wq);
	314	spin_lock_init(&ring->lock);
	315	INIT_WORK(&ring->work, p9_xen_response);
	316
	317	ring->intf = (struct xen_9pfs_data_intf *)get_zeroed_page(GFP_KERNEL);
	318	if (!ring->intf)
	319	return ret;
	320	ret = gnttab_grant_foreign_access(dev->otherend_id,
	321	virt_to_gfn(ring->intf), 0);
	322	if (ret < 0)
	323	goto out;
	324	ring->ref = ret;
5cadd4bb JG	325	bytes = alloc_pages_exact(1UL << (order + XEN_PAGE_SHIFT),
5cadd4bb JG	326	GFP_KERNEL \| __GFP_ZERO);
71ebd719 SS	327	if (!bytes) {
	328	ret = -ENOMEM;
	329	goto out;
	330	}
36f99675	331	for (; i < (1 << order); i++) {
71ebd719 SS	332	ret = gnttab_grant_foreign_access(
	333	dev->otherend_id, virt_to_gfn(bytes) + i, 0);
	334	if (ret < 0)
	335	goto out;
	336	ring->intf->ref[i] = ret;
	337	}
36f99675	338	ring->intf->ring_order = order;
71ebd719	339	ring->data.in = bytes;
36f99675	340	ring->data.out = bytes + XEN_FLEX_RING_SIZE(order);
71ebd719 SS	341
	342	ret = xenbus_alloc_evtchn(dev, &ring->evtchn);
	343	if (ret)
	344	goto out;
	345	ring->irq = bind_evtchn_to_irqhandler(ring->evtchn,
	346	xen_9pfs_front_event_handler,
	347	0, "xen_9pfs-frontend", ring);
	348	if (ring->irq >= 0)
	349	return 0;
	350
	351	xenbus_free_evtchn(dev, ring->evtchn);
	352	ret = ring->irq;
	353	out:
	354	if (bytes) {
	355	for (i--; i >= 0; i--)
49f8b459	356	gnttab_end_foreign_access(ring->intf->ref[i], NULL);
5cadd4bb	357	free_pages_exact(bytes, 1UL << (order + XEN_PAGE_SHIFT));
71ebd719	358	}
49f8b459	359	gnttab_end_foreign_access(ring->ref, NULL);
71ebd719 SS	360	free_page((unsigned long)ring->intf);
	361	return ret;
	362	}
	363
868eb122 SS	364	static int xen_9pfs_front_probe(struct xenbus_device *dev,
	365	const struct xenbus_device_id *id)
	366	{
71ebd719 SS	367	int ret, i;
	368	struct xenbus_transaction xbt;
	369	struct xen_9pfs_front_priv *priv = NULL;
	370	char *versions;
31d47266	371	unsigned int max_rings, max_ring_order, len = 0;
71ebd719 SS	372
71ebd719 SS	373	versions = xenbus_read(XBT_NIL, dev->otherend, "versions", &len);
2f9ad0ac DM	374	if (IS_ERR(versions))
2f9ad0ac DM	375	return PTR_ERR(versions);
71ebd719 SS	376	if (strcmp(versions, "1")) {
	377	kfree(versions);
	378	return -EINVAL;
	379	}
	380	kfree(versions);
	381	max_rings = xenbus_read_unsigned(dev->otherend, "max-rings", 0);
	382	if (max_rings < XEN_9PFS_NUM_RINGS)
	383	return -EINVAL;
	384	max_ring_order = xenbus_read_unsigned(dev->otherend,
	385	"max-ring-page-order", 0);
36f99675 SS	386	if (max_ring_order > XEN_9PFS_RING_ORDER)
	387	max_ring_order = XEN_9PFS_RING_ORDER;
	388	if (p9_xen_trans.maxsize > XEN_FLEX_RING_SIZE(max_ring_order))
	389	p9_xen_trans.maxsize = XEN_FLEX_RING_SIZE(max_ring_order) / 2;
71ebd719 SS	390
	391	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
	392	if (!priv)
	393	return -ENOMEM;
	394
	395	priv->dev = dev;
	396	priv->num_rings = XEN_9PFS_NUM_RINGS;
	397	priv->rings = kcalloc(priv->num_rings, sizeof(*priv->rings),
	398	GFP_KERNEL);
	399	if (!priv->rings) {
	400	kfree(priv);
	401	return -ENOMEM;
	402	}
	403
	404	for (i = 0; i < priv->num_rings; i++) {
	405	priv->rings[i].priv = priv;
36f99675 SS	406	ret = xen_9pfs_front_alloc_dataring(dev, &priv->rings[i],
36f99675 SS	407	max_ring_order);
71ebd719 SS	408	if (ret < 0)
	409	goto error;
	410	}
	411
	412	again:
	413	ret = xenbus_transaction_start(&xbt);
	414	if (ret) {
	415	xenbus_dev_fatal(dev, ret, "starting transaction");
	416	goto error;
	417	}
	418	ret = xenbus_printf(xbt, dev->nodename, "version", "%u", 1);
	419	if (ret)
	420	goto error_xenbus;
	421	ret = xenbus_printf(xbt, dev->nodename, "num-rings", "%u",
	422	priv->num_rings);
	423	if (ret)
	424	goto error_xenbus;
	425	for (i = 0; i < priv->num_rings; i++) {
	426	char str[16];
	427
	428	BUILD_BUG_ON(XEN_9PFS_NUM_RINGS > 9);
316a1bef	429	sprintf(str, "ring-ref%d", i);
71ebd719 SS	430	ret = xenbus_printf(xbt, dev->nodename, str, "%d",
	431	priv->rings[i].ref);
	432	if (ret)
	433	goto error_xenbus;
	434
316a1bef	435	sprintf(str, "event-channel-%d", i);
71ebd719 SS	436	ret = xenbus_printf(xbt, dev->nodename, str, "%u",
	437	priv->rings[i].evtchn);
	438	if (ret)
	439	goto error_xenbus;
	440	}
	441	priv->tag = xenbus_read(xbt, dev->nodename, "tag", NULL);
14e3995e WY	442	if (IS_ERR(priv->tag)) {
14e3995e WY	443	ret = PTR_ERR(priv->tag);
71ebd719 SS	444	goto error_xenbus;
	445	}
	446	ret = xenbus_transaction_end(xbt, 0);
	447	if (ret) {
	448	if (ret == -EAGAIN)
	449	goto again;
	450	xenbus_dev_fatal(dev, ret, "completing transaction");
	451	goto error;
	452	}
	453
	454	write_lock(&xen_9pfs_lock);
	455	list_add_tail(&priv->list, &xen_9pfs_devs);
	456	write_unlock(&xen_9pfs_lock);
	457	dev_set_drvdata(&dev->dev, priv);
	458	xenbus_switch_state(dev, XenbusStateInitialised);
	459
868eb122	460	return 0;
71ebd719 SS	461
	462	error_xenbus:
	463	xenbus_transaction_end(xbt, 1);
	464	xenbus_dev_fatal(dev, ret, "writing xenstore");
	465	error:
	466	dev_set_drvdata(&dev->dev, NULL);
	467	xen_9pfs_front_free(priv);
	468	return ret;
868eb122 SS	469	}
	470
	471	static int xen_9pfs_front_resume(struct xenbus_device *dev)
	472	{
680a2846	473	dev_warn(&dev->dev, "suspend/resume unsupported\n");
868eb122 SS	474	return 0;
	475	}
	476
	477	static void xen_9pfs_front_changed(struct xenbus_device *dev,
	478	enum xenbus_state backend_state)
	479	{
71ebd719 SS	480	switch (backend_state) {
	481	case XenbusStateReconfiguring:
	482	case XenbusStateReconfigured:
	483	case XenbusStateInitialising:
	484	case XenbusStateInitialised:
	485	case XenbusStateUnknown:
	486	break;
	487
	488	case XenbusStateInitWait:
	489	break;
	490
	491	case XenbusStateConnected:
	492	xenbus_switch_state(dev, XenbusStateConnected);
	493	break;
	494
	495	case XenbusStateClosed:
	496	if (dev->state == XenbusStateClosed)
	497	break;
df561f66	498	fallthrough; /* Missed the backend's CLOSING state */
71ebd719 SS	499	case XenbusStateClosing:
	500	xenbus_frontend_closed(dev);
	501	break;
	502	}
868eb122 SS	503	}
	504
	505	static struct xenbus_driver xen_9pfs_front_driver = {
	506	.ids = xen_9pfs_front_ids,
	507	.probe = xen_9pfs_front_probe,
	508	.remove = xen_9pfs_front_remove,
	509	.resume = xen_9pfs_front_resume,
	510	.otherend_changed = xen_9pfs_front_changed,
	511	};
	512
aaf0475a	513	static int p9_trans_xen_init(void)
868eb122	514	{
80a316ff Y	515	int rc;
80a316ff Y	516
868eb122 SS	517	if (!xen_domain())
	518	return -ENODEV;
	519
	520	pr_info("Initialising Xen transport for 9pfs\n");
	521
	522	v9fs_register_trans(&p9_xen_trans);
80a316ff Y	523	rc = xenbus_register_frontend(&xen_9pfs_front_driver);
	524	if (rc)
	525	v9fs_unregister_trans(&p9_xen_trans);
	526
	527	return rc;
868eb122 SS	528	}
868eb122 SS	529	module_init(p9_trans_xen_init);
4cd82a5b	530	MODULE_ALIAS_9P("xen");
868eb122	531
aaf0475a	532	static void p9_trans_xen_exit(void)
868eb122 SS	533	{
	534	v9fs_unregister_trans(&p9_xen_trans);
	535	return xenbus_unregister_driver(&xen_9pfs_front_driver);
	536	}
	537	module_exit(p9_trans_xen_exit);
d542296a	538
99aa673e	539	MODULE_ALIAS("xen:9pfs");
d542296a SH	540	MODULE_AUTHOR("Stefano Stabellini <stefano@aporeto.com>");
	541	MODULE_DESCRIPTION("Xen Transport for 9P");
	542	MODULE_LICENSE("GPL");