[mirror_ubuntu-jammy-kernel.git] / net / xdp / xsk.c

// SPDX-License-Identifier: GPL-2.0
/* XDP sockets
 *
 * AF_XDP sockets allows a channel between XDP programs and userspace
 * applications.
 * Copyright(c) 2018 Intel Corporation.
 *
 * This program is free software; you can redistribute it and/or modify it
 * under the terms and conditions of the GNU General Public License,
 * version 2, as published by the Free Software Foundation.
 *
 * This program is distributed in the hope 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.
 *
 * Author(s): Björn Töpel <bjorn.topel@intel.com>
 *	      Magnus Karlsson <magnus.karlsson@intel.com>
 */

#define pr_fmt(fmt) "AF_XDP: %s: " fmt, __func__

#include <linux/if_xdp.h>
#include <linux/init.h>
#include <linux/sched/mm.h>
#include <linux/sched/signal.h>
#include <linux/sched/task.h>
#include <linux/socket.h>
#include <linux/file.h>
#include <linux/uaccess.h>
#include <linux/net.h>
#include <linux/netdevice.h>
#include <net/xdp_sock.h>
#include <net/xdp.h>

#include "xsk_queue.h"
#include "xdp_umem.h"

static struct xdp_sock *xdp_sk(struct sock *sk)
{
	return (struct xdp_sock *)sk;
}

bool xsk_is_setup_for_bpf_map(struct xdp_sock *xs)
{
	return !!xs->rx;
}

static int __xsk_rcv(struct xdp_sock *xs, struct xdp_buff *xdp)
{
	u32 *id, len = xdp->data_end - xdp->data;
	void *buffer;
	int err = 0;

	if (xs->dev != xdp->rxq->dev || xs->queue_id != xdp->rxq->queue_index)
		return -EINVAL;

	id = xskq_peek_id(xs->umem->fq);
	if (!id)
		return -ENOSPC;

	buffer = xdp_umem_get_data_with_headroom(xs->umem, *id);
	memcpy(buffer, xdp->data, len);
	err = xskq_produce_batch_desc(xs->rx, *id, len,
				      xs->umem->frame_headroom);
	if (!err)
		xskq_discard_id(xs->umem->fq);

	return err;
}

int xsk_rcv(struct xdp_sock *xs, struct xdp_buff *xdp)
{
	int err;

	err = __xsk_rcv(xs, xdp);
	if (likely(!err))
		xdp_return_buff(xdp);
	else
		xs->rx_dropped++;

	return err;
}

void xsk_flush(struct xdp_sock *xs)
{
	xskq_produce_flush_desc(xs->rx);
	xs->sk.sk_data_ready(&xs->sk);
}

int xsk_generic_rcv(struct xdp_sock *xs, struct xdp_buff *xdp)
{
	int err;

	err = __xsk_rcv(xs, xdp);
	if (!err)
		xsk_flush(xs);
	else
		xs->rx_dropped++;

	return err;
}

static unsigned int xsk_poll(struct file *file, struct socket *sock,
			     struct poll_table_struct *wait)
{
	unsigned int mask = datagram_poll(file, sock, wait);
	struct sock *sk = sock->sk;
	struct xdp_sock *xs = xdp_sk(sk);

	if (xs->rx && !xskq_empty_desc(xs->rx))
		mask |= POLLIN | POLLRDNORM;

	return mask;
}

static int xsk_init_queue(u32 entries, struct xsk_queue **queue,
			  bool umem_queue)
{
	struct xsk_queue *q;

	if (entries == 0 || *queue || !is_power_of_2(entries))
		return -EINVAL;

	q = xskq_create(entries, umem_queue);
	if (!q)
		return -ENOMEM;

	*queue = q;
	return 0;
}

static void __xsk_release(struct xdp_sock *xs)
{
	/* Wait for driver to stop using the xdp socket. */
	synchronize_net();

	dev_put(xs->dev);
}

static int xsk_release(struct socket *sock)
{
	struct sock *sk = sock->sk;
	struct xdp_sock *xs = xdp_sk(sk);
	struct net *net;

	if (!sk)
		return 0;

	net = sock_net(sk);

	local_bh_disable();
	sock_prot_inuse_add(net, sk->sk_prot, -1);
	local_bh_enable();

	if (xs->dev) {
		__xsk_release(xs);
		xs->dev = NULL;
	}

	sock_orphan(sk);
	sock->sk = NULL;

	sk_refcnt_debug_release(sk);
	sock_put(sk);

	return 0;
}

static struct socket *xsk_lookup_xsk_from_fd(int fd)
{
	struct socket *sock;
	int err;

	sock = sockfd_lookup(fd, &err);
	if (!sock)
		return ERR_PTR(-ENOTSOCK);

	if (sock->sk->sk_family != PF_XDP) {
		sockfd_put(sock);
		return ERR_PTR(-ENOPROTOOPT);
	}

	return sock;
}

static int xsk_bind(struct socket *sock, struct sockaddr *addr, int addr_len)
{
	struct sockaddr_xdp *sxdp = (struct sockaddr_xdp *)addr;
	struct sock *sk = sock->sk;
	struct net_device *dev, *dev_curr;
	struct xdp_sock *xs = xdp_sk(sk);
	struct xdp_umem *old_umem = NULL;
	int err = 0;

	if (addr_len < sizeof(struct sockaddr_xdp))
		return -EINVAL;
	if (sxdp->sxdp_family != AF_XDP)
		return -EINVAL;

	mutex_lock(&xs->mutex);
	dev_curr = xs->dev;
	dev = dev_get_by_index(sock_net(sk), sxdp->sxdp_ifindex);
	if (!dev) {
		err = -ENODEV;
		goto out_release;
	}

	if (!xs->rx) {
		err = -EINVAL;
		goto out_unlock;
	}

	if (sxdp->sxdp_queue_id >= dev->num_rx_queues) {
		err = -EINVAL;
		goto out_unlock;
	}

	if (sxdp->sxdp_flags & XDP_SHARED_UMEM) {
		struct xdp_sock *umem_xs;
		struct socket *sock;

		if (xs->umem) {
			/* We have already our own. */
			err = -EINVAL;
			goto out_unlock;
		}

		sock = xsk_lookup_xsk_from_fd(sxdp->sxdp_shared_umem_fd);
		if (IS_ERR(sock)) {
			err = PTR_ERR(sock);
			goto out_unlock;
		}

		umem_xs = xdp_sk(sock->sk);
		if (!umem_xs->umem) {
			/* No umem to inherit. */
			err = -EBADF;
			sockfd_put(sock);
			goto out_unlock;
		} else if (umem_xs->dev != dev ||
			   umem_xs->queue_id != sxdp->sxdp_queue_id) {
			err = -EINVAL;
			sockfd_put(sock);
			goto out_unlock;
		}

		xdp_get_umem(umem_xs->umem);
		old_umem = xs->umem;
		xs->umem = umem_xs->umem;
		sockfd_put(sock);
	} else if (!xs->umem || !xdp_umem_validate_queues(xs->umem)) {
		err = -EINVAL;
		goto out_unlock;
	} else {
		/* This xsk has its own umem. */
		xskq_set_umem(xs->umem->fq, &xs->umem->props);
		xskq_set_umem(xs->umem->cq, &xs->umem->props);
	}

	/* Rebind? */
	if (dev_curr && (dev_curr != dev ||
			 xs->queue_id != sxdp->sxdp_queue_id)) {
		__xsk_release(xs);
		if (old_umem)
			xdp_put_umem(old_umem);
	}

	xs->dev = dev;
	xs->queue_id = sxdp->sxdp_queue_id;

	xskq_set_umem(xs->rx, &xs->umem->props);

out_unlock:
	if (err)
		dev_put(dev);
out_release:
	mutex_unlock(&xs->mutex);
	return err;
}

static int xsk_setsockopt(struct socket *sock, int level, int optname,
			  char __user *optval, unsigned int optlen)
{
	struct sock *sk = sock->sk;
	struct xdp_sock *xs = xdp_sk(sk);
	int err;

	if (level != SOL_XDP)
		return -ENOPROTOOPT;

	switch (optname) {
	case XDP_RX_RING:
	{
		struct xsk_queue **q;
		int entries;

		if (optlen < sizeof(entries))
			return -EINVAL;
		if (copy_from_user(&entries, optval, sizeof(entries)))
			return -EFAULT;

		mutex_lock(&xs->mutex);
		q = &xs->rx;
		err = xsk_init_queue(entries, q, false);
		mutex_unlock(&xs->mutex);
		return err;
	}
	case XDP_UMEM_REG:
	{
		struct xdp_umem_reg mr;
		struct xdp_umem *umem;

		if (xs->umem)
			return -EBUSY;

		if (copy_from_user(&mr, optval, sizeof(mr)))
			return -EFAULT;

		mutex_lock(&xs->mutex);
		err = xdp_umem_create(&umem);

		err = xdp_umem_reg(umem, &mr);
		if (err) {
			kfree(umem);
			mutex_unlock(&xs->mutex);
			return err;
		}

		/* Make sure umem is ready before it can be seen by others */
		smp_wmb();

		xs->umem = umem;
		mutex_unlock(&xs->mutex);
		return 0;
	}
	case XDP_UMEM_FILL_RING:
	case XDP_UMEM_COMPLETION_RING:
	{
		struct xsk_queue **q;
		int entries;

		if (!xs->umem)
			return -EINVAL;

		if (copy_from_user(&entries, optval, sizeof(entries)))
			return -EFAULT;

		mutex_lock(&xs->mutex);
		q = (optname == XDP_UMEM_FILL_RING) ? &xs->umem->fq :
			&xs->umem->cq;
		err = xsk_init_queue(entries, q, true);
		mutex_unlock(&xs->mutex);
		return err;
	}
	default:
		break;
	}

	return -ENOPROTOOPT;
}

static int xsk_mmap(struct file *file, struct socket *sock,
		    struct vm_area_struct *vma)
{
	unsigned long offset = vma->vm_pgoff << PAGE_SHIFT;
	unsigned long size = vma->vm_end - vma->vm_start;
	struct xdp_sock *xs = xdp_sk(sock->sk);
	struct xsk_queue *q = NULL;
	unsigned long pfn;
	struct page *qpg;

	if (offset == XDP_PGOFF_RX_RING) {
		q = xs->rx;
	} else {
		if (!xs->umem)
			return -EINVAL;

		if (offset == XDP_UMEM_PGOFF_FILL_RING)
			q = xs->umem->fq;
		else if (offset == XDP_UMEM_PGOFF_COMPLETION_RING)
			q = xs->umem->cq;
		else
			return -EINVAL;
	}

	if (!q)
		return -EINVAL;

	qpg = virt_to_head_page(q->ring);
	if (size > (PAGE_SIZE << compound_order(qpg)))
		return -EINVAL;

	pfn = virt_to_phys(q->ring) >> PAGE_SHIFT;
	return remap_pfn_range(vma, vma->vm_start, pfn,
			       size, vma->vm_page_prot);
}

static struct proto xsk_proto = {
	.name =		"XDP",
	.owner =	THIS_MODULE,
	.obj_size =	sizeof(struct xdp_sock),
};

static const struct proto_ops xsk_proto_ops = {
	.family =	PF_XDP,
	.owner =	THIS_MODULE,
	.release =	xsk_release,
	.bind =		xsk_bind,
	.connect =	sock_no_connect,
	.socketpair =	sock_no_socketpair,
	.accept =	sock_no_accept,
	.getname =	sock_no_getname,
	.poll =		xsk_poll,
	.ioctl =	sock_no_ioctl,
	.listen =	sock_no_listen,
	.shutdown =	sock_no_shutdown,
	.setsockopt =	xsk_setsockopt,
	.getsockopt =	sock_no_getsockopt,
	.sendmsg =	sock_no_sendmsg,
	.recvmsg =	sock_no_recvmsg,
	.mmap =		xsk_mmap,
	.sendpage =	sock_no_sendpage,
};

static void xsk_destruct(struct sock *sk)
{
	struct xdp_sock *xs = xdp_sk(sk);

	if (!sock_flag(sk, SOCK_DEAD))
		return;

	xskq_destroy(xs->rx);
	xdp_put_umem(xs->umem);

	sk_refcnt_debug_dec(sk);
}

static int xsk_create(struct net *net, struct socket *sock, int protocol,
		      int kern)
{
	struct sock *sk;
	struct xdp_sock *xs;

	if (!ns_capable(net->user_ns, CAP_NET_RAW))
		return -EPERM;
	if (sock->type != SOCK_RAW)
		return -ESOCKTNOSUPPORT;

	if (protocol)
		return -EPROTONOSUPPORT;

	sock->state = SS_UNCONNECTED;

	sk = sk_alloc(net, PF_XDP, GFP_KERNEL, &xsk_proto, kern);
	if (!sk)
		return -ENOBUFS;

	sock->ops = &xsk_proto_ops;

	sock_init_data(sock, sk);

	sk->sk_family = PF_XDP;

	sk->sk_destruct = xsk_destruct;
	sk_refcnt_debug_inc(sk);

	xs = xdp_sk(sk);
	mutex_init(&xs->mutex);

	local_bh_disable();
	sock_prot_inuse_add(net, &xsk_proto, 1);
	local_bh_enable();

	return 0;
}

static const struct net_proto_family xsk_family_ops = {
	.family = PF_XDP,
	.create = xsk_create,
	.owner	= THIS_MODULE,
};

static int __init xsk_init(void)
{
	int err;

	err = proto_register(&xsk_proto, 0 /* no slab */);
	if (err)
		goto out;

	err = sock_register(&xsk_family_ops);
	if (err)
		goto out_proto;

	return 0;

out_proto:
	proto_unregister(&xsk_proto);
out:
	return err;
}

fs_initcall(xsk_init);
Commit	Line	Data
c0c77d8f BT	1	// SPDX-License-Identifier: GPL-2.0
	2	/* XDP sockets
	3	*
	4	* AF_XDP sockets allows a channel between XDP programs and userspace
	5	* applications.
	6	* Copyright(c) 2018 Intel Corporation.
	7	*
	8	* This program is free software; you can redistribute it and/or modify it
	9	* under the terms and conditions of the GNU General Public License,
	10	* version 2, as published by the Free Software Foundation.
	11	*
	12	* This program is distributed in the hope it will be useful, but WITHOUT
	13	* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
	14	* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
	15	* more details.
	16	*
	17	* Author(s): Björn Töpel <bjorn.topel@intel.com>
	18	* Magnus Karlsson <magnus.karlsson@intel.com>
	19	*/
	20
	21	#define pr_fmt(fmt) "AF_XDP: %s: " fmt, __func__
	22
	23	#include <linux/if_xdp.h>
	24	#include <linux/init.h>
	25	#include <linux/sched/mm.h>
	26	#include <linux/sched/signal.h>
	27	#include <linux/sched/task.h>
	28	#include <linux/socket.h>
	29	#include <linux/file.h>
	30	#include <linux/uaccess.h>
	31	#include <linux/net.h>
	32	#include <linux/netdevice.h>
	33	#include <net/xdp_sock.h>
b9b6b68e	34	#include <net/xdp.h>
c0c77d8f	35
423f3832	36	#include "xsk_queue.h"
c0c77d8f BT	37	#include "xdp_umem.h"
	38
	39	static struct xdp_sock xdp_sk(struct sock sk)
	40	{
	41	return (struct xdp_sock *)sk;
	42	}
	43
fbfc504a BT	44	bool xsk_is_setup_for_bpf_map(struct xdp_sock *xs)
	45	{
	46	return !!xs->rx;
	47	}
	48
c497176c BT	49	static int __xsk_rcv(struct xdp_sock xs, struct xdp_buff xdp)
	50	{
	51	u32 *id, len = xdp->data_end - xdp->data;
	52	void *buffer;
	53	int err = 0;
	54
	55	if (xs->dev != xdp->rxq->dev \|\| xs->queue_id != xdp->rxq->queue_index)
	56	return -EINVAL;
	57
	58	id = xskq_peek_id(xs->umem->fq);
	59	if (!id)
	60	return -ENOSPC;
	61
	62	buffer = xdp_umem_get_data_with_headroom(xs->umem, *id);
	63	memcpy(buffer, xdp->data, len);
	64	err = xskq_produce_batch_desc(xs->rx, *id, len,
	65	xs->umem->frame_headroom);
	66	if (!err)
	67	xskq_discard_id(xs->umem->fq);
	68
	69	return err;
	70	}
	71
	72	int xsk_rcv(struct xdp_sock xs, struct xdp_buff xdp)
	73	{
	74	int err;
	75
	76	err = __xsk_rcv(xs, xdp);
	77	if (likely(!err))
	78	xdp_return_buff(xdp);
	79	else
	80	xs->rx_dropped++;
	81
	82	return err;
	83	}
	84
	85	void xsk_flush(struct xdp_sock *xs)
	86	{
	87	xskq_produce_flush_desc(xs->rx);
	88	xs->sk.sk_data_ready(&xs->sk);
	89	}
	90
	91	int xsk_generic_rcv(struct xdp_sock xs, struct xdp_buff xdp)
	92	{
	93	int err;
	94
	95	err = __xsk_rcv(xs, xdp);
	96	if (!err)
	97	xsk_flush(xs);
	98	else
	99	xs->rx_dropped++;
	100
	101	return err;
	102	}
	103
	104	static unsigned int xsk_poll(struct file file, struct socket sock,
	105	struct poll_table_struct *wait)
	106	{
	107	unsigned int mask = datagram_poll(file, sock, wait);
	108	struct sock *sk = sock->sk;
	109	struct xdp_sock *xs = xdp_sk(sk);
	110
	111	if (xs->rx && !xskq_empty_desc(xs->rx))
	112	mask \|= POLLIN \| POLLRDNORM;
113
114	return mask;
115	}
116
b9b6b68e BT	117	static int xsk_init_queue(u32 entries, struct xsk_queue **queue,
b9b6b68e BT	118	bool umem_queue)
423f3832 MK	119	{
	120	struct xsk_queue *q;
	121
	122	if (entries == 0 \|\| *queue \|\| !is_power_of_2(entries))
	123	return -EINVAL;
	124
b9b6b68e	125	q = xskq_create(entries, umem_queue);
423f3832 MK	126	if (!q)
	127	return -ENOMEM;
	128
	129	*queue = q;
	130	return 0;
	131	}
	132
965a9909 MK	133	static void __xsk_release(struct xdp_sock *xs)
	134	{
	135	/* Wait for driver to stop using the xdp socket. */
	136	synchronize_net();
	137
	138	dev_put(xs->dev);
	139	}
	140
c0c77d8f BT	141	static int xsk_release(struct socket *sock)
	142	{
	143	struct sock *sk = sock->sk;
965a9909	144	struct xdp_sock *xs = xdp_sk(sk);
c0c77d8f BT	145	struct net *net;
	146
	147	if (!sk)
	148	return 0;
	149
	150	net = sock_net(sk);
	151
	152	local_bh_disable();
	153	sock_prot_inuse_add(net, sk->sk_prot, -1);
	154	local_bh_enable();
	155
965a9909 MK	156	if (xs->dev) {
	157	__xsk_release(xs);
	158	xs->dev = NULL;
	159	}
	160
c0c77d8f BT	161	sock_orphan(sk);
	162	sock->sk = NULL;
	163
	164	sk_refcnt_debug_release(sk);
	165	sock_put(sk);
	166
	167	return 0;
	168	}
	169
965a9909 MK	170	static struct socket *xsk_lookup_xsk_from_fd(int fd)
	171	{
	172	struct socket *sock;
	173	int err;
	174
	175	sock = sockfd_lookup(fd, &err);
	176	if (!sock)
	177	return ERR_PTR(-ENOTSOCK);
	178
	179	if (sock->sk->sk_family != PF_XDP) {
	180	sockfd_put(sock);
	181	return ERR_PTR(-ENOPROTOOPT);
	182	}
	183
	184	return sock;
	185	}
	186
	187	static int xsk_bind(struct socket sock, struct sockaddr addr, int addr_len)
	188	{
	189	struct sockaddr_xdp sxdp = (struct sockaddr_xdp )addr;
	190	struct sock *sk = sock->sk;
	191	struct net_device dev, dev_curr;
	192	struct xdp_sock *xs = xdp_sk(sk);
	193	struct xdp_umem *old_umem = NULL;
	194	int err = 0;
	195
	196	if (addr_len < sizeof(struct sockaddr_xdp))
	197	return -EINVAL;
	198	if (sxdp->sxdp_family != AF_XDP)
	199	return -EINVAL;
	200
	201	mutex_lock(&xs->mutex);
	202	dev_curr = xs->dev;
	203	dev = dev_get_by_index(sock_net(sk), sxdp->sxdp_ifindex);
	204	if (!dev) {
	205	err = -ENODEV;
	206	goto out_release;
	207	}
	208
	209	if (!xs->rx) {
	210	err = -EINVAL;
	211	goto out_unlock;
	212	}
	213
	214	if (sxdp->sxdp_queue_id >= dev->num_rx_queues) {
	215	err = -EINVAL;
	216	goto out_unlock;
	217	}
	218
	219	if (sxdp->sxdp_flags & XDP_SHARED_UMEM) {
	220	struct xdp_sock *umem_xs;
	221	struct socket *sock;
	222
	223	if (xs->umem) {
	224	/* We have already our own. */
	225	err = -EINVAL;
	226	goto out_unlock;
	227	}
	228
	229	sock = xsk_lookup_xsk_from_fd(sxdp->sxdp_shared_umem_fd);
	230	if (IS_ERR(sock)) {
	231	err = PTR_ERR(sock);
	232	goto out_unlock;
	233	}
234
235	umem_xs = xdp_sk(sock->sk);
236	if (!umem_xs->umem) {
237	/* No umem to inherit. */
238	err = -EBADF;
239	sockfd_put(sock);
240	goto out_unlock;
241	} else if (umem_xs->dev != dev \|\|
242	umem_xs->queue_id != sxdp->sxdp_queue_id) {
243	err = -EINVAL;
244	sockfd_put(sock);
245	goto out_unlock;
246	}
247
248	xdp_get_umem(umem_xs->umem);
249	old_umem = xs->umem;
250	xs->umem = umem_xs->umem;
251	sockfd_put(sock);
252	} else if (!xs->umem \|\| !xdp_umem_validate_queues(xs->umem)) {
253	err = -EINVAL;
254	goto out_unlock;
c497176c BT	255	} else {
	256	/* This xsk has its own umem. */
	257	xskq_set_umem(xs->umem->fq, &xs->umem->props);
fe230832	258	xskq_set_umem(xs->umem->cq, &xs->umem->props);
965a9909 MK	259	}
	260
	261	/* Rebind? */
	262	if (dev_curr && (dev_curr != dev \|\|
	263	xs->queue_id != sxdp->sxdp_queue_id)) {
	264	__xsk_release(xs);
	265	if (old_umem)
	266	xdp_put_umem(old_umem);
	267	}
	268
	269	xs->dev = dev;
	270	xs->queue_id = sxdp->sxdp_queue_id;
	271
	272	xskq_set_umem(xs->rx, &xs->umem->props);
	273
	274	out_unlock:
	275	if (err)
	276	dev_put(dev);
	277	out_release:
	278	mutex_unlock(&xs->mutex);
	279	return err;
	280	}
	281
c0c77d8f BT	282	static int xsk_setsockopt(struct socket *sock, int level, int optname,
	283	char __user *optval, unsigned int optlen)
	284	{
	285	struct sock *sk = sock->sk;
	286	struct xdp_sock *xs = xdp_sk(sk);
	287	int err;
	288
	289	if (level != SOL_XDP)
	290	return -ENOPROTOOPT;
	291
	292	switch (optname) {
b9b6b68e BT	293	case XDP_RX_RING:
	294	{
	295	struct xsk_queue **q;
	296	int entries;
	297
	298	if (optlen < sizeof(entries))
	299	return -EINVAL;
	300	if (copy_from_user(&entries, optval, sizeof(entries)))
	301	return -EFAULT;
	302
	303	mutex_lock(&xs->mutex);
	304	q = &xs->rx;
	305	err = xsk_init_queue(entries, q, false);
	306	mutex_unlock(&xs->mutex);
	307	return err;
	308	}
c0c77d8f BT	309	case XDP_UMEM_REG:
	310	{
	311	struct xdp_umem_reg mr;
	312	struct xdp_umem *umem;
	313
	314	if (xs->umem)
	315	return -EBUSY;
	316
	317	if (copy_from_user(&mr, optval, sizeof(mr)))
	318	return -EFAULT;
	319
	320	mutex_lock(&xs->mutex);
	321	err = xdp_umem_create(&umem);
	322
	323	err = xdp_umem_reg(umem, &mr);
	324	if (err) {
	325	kfree(umem);
	326	mutex_unlock(&xs->mutex);
	327	return err;
	328	}
	329
	330	/* Make sure umem is ready before it can be seen by others */
	331	smp_wmb();
	332
	333	xs->umem = umem;
	334	mutex_unlock(&xs->mutex);
	335	return 0;
	336	}
423f3832	337	case XDP_UMEM_FILL_RING:
fe230832	338	case XDP_UMEM_COMPLETION_RING:
423f3832 MK	339	{
	340	struct xsk_queue **q;
	341	int entries;
	342
	343	if (!xs->umem)
	344	return -EINVAL;
	345
	346	if (copy_from_user(&entries, optval, sizeof(entries)))
	347	return -EFAULT;
	348
	349	mutex_lock(&xs->mutex);
fe230832 MK	350	q = (optname == XDP_UMEM_FILL_RING) ? &xs->umem->fq :
fe230832 MK	351	&xs->umem->cq;
b9b6b68e	352	err = xsk_init_queue(entries, q, true);
423f3832 MK	353	mutex_unlock(&xs->mutex);
	354	return err;
	355	}
c0c77d8f BT	356	default:
	357	break;
	358	}
	359
	360	return -ENOPROTOOPT;
	361	}
	362
423f3832 MK	363	static int xsk_mmap(struct file file, struct socket sock,
	364	struct vm_area_struct *vma)
	365	{
	366	unsigned long offset = vma->vm_pgoff << PAGE_SHIFT;
	367	unsigned long size = vma->vm_end - vma->vm_start;
	368	struct xdp_sock *xs = xdp_sk(sock->sk);
	369	struct xsk_queue *q = NULL;
	370	unsigned long pfn;
	371	struct page *qpg;
	372
b9b6b68e BT	373	if (offset == XDP_PGOFF_RX_RING) {
	374	q = xs->rx;
	375	} else {
	376	if (!xs->umem)
	377	return -EINVAL;
423f3832	378
b9b6b68e BT	379	if (offset == XDP_UMEM_PGOFF_FILL_RING)
b9b6b68e BT	380	q = xs->umem->fq;
fe230832 MK	381	else if (offset == XDP_UMEM_PGOFF_COMPLETION_RING)
fe230832 MK	382	q = xs->umem->cq;
b9b6b68e BT	383	else
	384	return -EINVAL;
	385	}
423f3832 MK	386
	387	if (!q)
	388	return -EINVAL;
	389
	390	qpg = virt_to_head_page(q->ring);
	391	if (size > (PAGE_SIZE << compound_order(qpg)))
	392	return -EINVAL;
	393
	394	pfn = virt_to_phys(q->ring) >> PAGE_SHIFT;
	395	return remap_pfn_range(vma, vma->vm_start, pfn,
	396	size, vma->vm_page_prot);
	397	}
	398
c0c77d8f BT	399	static struct proto xsk_proto = {
	400	.name = "XDP",
	401	.owner = THIS_MODULE,
	402	.obj_size = sizeof(struct xdp_sock),
	403	};
	404
	405	static const struct proto_ops xsk_proto_ops = {
	406	.family = PF_XDP,
	407	.owner = THIS_MODULE,
	408	.release = xsk_release,
965a9909	409	.bind = xsk_bind,
c0c77d8f BT	410	.connect = sock_no_connect,
	411	.socketpair = sock_no_socketpair,
	412	.accept = sock_no_accept,
	413	.getname = sock_no_getname,
c497176c	414	.poll = xsk_poll,
c0c77d8f BT	415	.ioctl = sock_no_ioctl,
	416	.listen = sock_no_listen,
	417	.shutdown = sock_no_shutdown,
	418	.setsockopt = xsk_setsockopt,
	419	.getsockopt = sock_no_getsockopt,
	420	.sendmsg = sock_no_sendmsg,
	421	.recvmsg = sock_no_recvmsg,
423f3832	422	.mmap = xsk_mmap,
c0c77d8f BT	423	.sendpage = sock_no_sendpage,
	424	};
	425
	426	static void xsk_destruct(struct sock *sk)
	427	{
	428	struct xdp_sock *xs = xdp_sk(sk);
	429
	430	if (!sock_flag(sk, SOCK_DEAD))
	431	return;
	432
b9b6b68e	433	xskq_destroy(xs->rx);
c0c77d8f BT	434	xdp_put_umem(xs->umem);
	435
	436	sk_refcnt_debug_dec(sk);
	437	}
	438
	439	static int xsk_create(struct net net, struct socket sock, int protocol,
	440	int kern)
	441	{
	442	struct sock *sk;
	443	struct xdp_sock *xs;
	444
	445	if (!ns_capable(net->user_ns, CAP_NET_RAW))
	446	return -EPERM;
	447	if (sock->type != SOCK_RAW)
	448	return -ESOCKTNOSUPPORT;
	449
	450	if (protocol)
	451	return -EPROTONOSUPPORT;
	452
	453	sock->state = SS_UNCONNECTED;
	454
	455	sk = sk_alloc(net, PF_XDP, GFP_KERNEL, &xsk_proto, kern);
	456	if (!sk)
	457	return -ENOBUFS;
	458
	459	sock->ops = &xsk_proto_ops;
	460
	461	sock_init_data(sock, sk);
	462
	463	sk->sk_family = PF_XDP;
	464
	465	sk->sk_destruct = xsk_destruct;
	466	sk_refcnt_debug_inc(sk);
	467
	468	xs = xdp_sk(sk);
	469	mutex_init(&xs->mutex);
	470
	471	local_bh_disable();
	472	sock_prot_inuse_add(net, &xsk_proto, 1);
	473	local_bh_enable();
	474
	475	return 0;
	476	}
	477
	478	static const struct net_proto_family xsk_family_ops = {
	479	.family = PF_XDP,
	480	.create = xsk_create,
	481	.owner = THIS_MODULE,
	482	};
	483
	484	static int __init xsk_init(void)
	485	{
	486	int err;
	487
	488	err = proto_register(&xsk_proto, 0 /* no slab */);
	489	if (err)
	490	goto out;
	491
	492	err = sock_register(&xsk_family_ops);
	493	if (err)
	494	goto out_proto;
	495
	496	return 0;
	497
498	out_proto:
499	proto_unregister(&xsk_proto);
500	out:
501	return err;
502	}
503
504	fs_initcall(xsk_init);