1#ifndef _UAPI_LINUX_VIRTIO_RING_H
2#define _UAPI_LINUX_VIRTIO_RING_H
3/* An interface for efficient virtio implementation, currently for use by KVM
4 * and lguest, but hopefully others soon.  Do NOT change this since it will
5 * break existing servers and clients.
6 *
7 * This header is BSD licensed so anyone can use the definitions to implement
8 * compatible drivers/servers.
9 *
10 * Redistribution and use in source and binary forms, with or without
11 * modification, are permitted provided that the following conditions
12 * are met:
13 * 1. Redistributions of source code must retain the above copyright
14 *    notice, this list of conditions and the following disclaimer.
15 * 2. Redistributions in binary form must reproduce the above copyright
16 *    notice, this list of conditions and the following disclaimer in the
17 *    documentation and/or other materials provided with the distribution.
18 * 3. Neither the name of IBM nor the names of its contributors
19 *    may be used to endorse or promote products derived from this software
20 *    without specific prior written permission.
21 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS ``AS IS'' AND
22 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
23 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
24 * ARE DISCLAIMED.  IN NO EVENT SHALL IBM OR CONTRIBUTORS BE LIABLE
25 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
26 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
27 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
28 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
29 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
30 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
31 * SUCH DAMAGE.
32 *
33 * Copyright Rusty Russell IBM Corporation 2007. */
34#ifndef __KERNEL__
35#include <stdint.h>
36#endif
37#include <linux/types.h>
38#include <linux/virtio_types.h>
39
40/* This marks a buffer as continuing via the next field. */
41#define VRING_DESC_F_NEXT	1
42/* This marks a buffer as write-only (otherwise read-only). */
43#define VRING_DESC_F_WRITE	2
44/* This means the buffer contains a list of buffer descriptors. */
45#define VRING_DESC_F_INDIRECT	4
46
47/* The Host uses this in used->flags to advise the Guest: don't kick me when
48 * you add a buffer.  It's unreliable, so it's simply an optimization.  Guest
49 * will still kick if it's out of buffers. */
50#define VRING_USED_F_NO_NOTIFY	1
51/* The Guest uses this in avail->flags to advise the Host: don't interrupt me
52 * when you consume a buffer.  It's unreliable, so it's simply an
53 * optimization.  */
54#define VRING_AVAIL_F_NO_INTERRUPT	1
55
56/* We support indirect buffer descriptors */
57#define VIRTIO_RING_F_INDIRECT_DESC	28
58
59/* The Guest publishes the used index for which it expects an interrupt
60 * at the end of the avail ring. Host should ignore the avail->flags field. */
61/* The Host publishes the avail index for which it expects a kick
62 * at the end of the used ring. Guest should ignore the used->flags field. */
63#define VIRTIO_RING_F_EVENT_IDX		29
64
65/* Virtio ring descriptors: 16 bytes.  These can chain together via "next". */
66struct vring_desc {
67	/* Address (guest-physical). */
68	__virtio64 addr;
69	/* Length. */
70	__virtio32 len;
71	/* The flags as indicated above. */
72	__virtio16 flags;
73	/* We chain unused descriptors via this, too */
74	__virtio16 next;
75};
76
77struct vring_avail {
78	__virtio16 flags;
79	__virtio16 idx;
80	__virtio16 ring[];
81};
82
83/* u32 is used here for ids for padding reasons. */
84struct vring_used_elem {
85	/* Index of start of used descriptor chain. */
86	__virtio32 id;
87	/* Total length of the descriptor chain which was used (written to) */
88	__virtio32 len;
89};
90
91struct vring_used {
92	__virtio16 flags;
93	__virtio16 idx;
94	struct vring_used_elem ring[];
95};
96
97struct vring {
98	unsigned int num;
99
100	struct vring_desc *desc;
101
102	struct vring_avail *avail;
103
104	struct vring_used *used;
105};
106
107/* Alignment requirements for vring elements.
108 * When using pre-virtio 1.0 layout, these fall out naturally.
109 */
110#define VRING_AVAIL_ALIGN_SIZE 2
111#define VRING_USED_ALIGN_SIZE 4
112#define VRING_DESC_ALIGN_SIZE 16
113
114/* The standard layout for the ring is a continuous chunk of memory which looks
115 * like this.  We assume num is a power of 2.
116 *
117 * struct vring
118 * {
119 *	// The actual descriptors (16 bytes each)
120 *	struct vring_desc desc[num];
121 *
122 *	// A ring of available descriptor heads with free-running index.
123 *	__virtio16 avail_flags;
124 *	__virtio16 avail_idx;
125 *	__virtio16 available[num];
126 *	__virtio16 used_event_idx;
127 *
128 *	// Padding to the next align boundary.
129 *	char pad[];
130 *
131 *	// A ring of used descriptor heads with free-running index.
132 *	__virtio16 used_flags;
133 *	__virtio16 used_idx;
134 *	struct vring_used_elem used[num];
135 *	__virtio16 avail_event_idx;
136 * };
137 */
138/* We publish the used event index at the end of the available ring, and vice
139 * versa. They are at the end for backwards compatibility. */
140#define vring_used_event(vr) ((vr)->avail->ring[(vr)->num])
141#define vring_avail_event(vr) (*(__virtio16 *)&(vr)->used->ring[(vr)->num])
142
143static inline void vring_init(struct vring *vr, unsigned int num, void *p,
144			      unsigned long align)
145{
146	vr->num = num;
147	vr->desc = p;
148	vr->avail = p + num*sizeof(struct vring_desc);
149	vr->used = (void *)(((uintptr_t)&vr->avail->ring[num] + sizeof(__virtio16)
150		+ align-1) & ~(align - 1));
151}
152
153static inline unsigned vring_size(unsigned int num, unsigned long align)
154{
155	return ((sizeof(struct vring_desc) * num + sizeof(__virtio16) * (3 + num)
156		 + align - 1) & ~(align - 1))
157		+ sizeof(__virtio16) * 3 + sizeof(struct vring_used_elem) * num;
158}
159
160/* The following is used with USED_EVENT_IDX and AVAIL_EVENT_IDX */
161/* Assuming a given event_idx value from the other side, if
162 * we have just incremented index from old to new_idx,
163 * should we trigger an event? */
164static inline int vring_need_event(__u16 event_idx, __u16 new_idx, __u16 old)
165{
166	/* Note: Xen has similar logic for notification hold-off
167	 * in include/xen/interface/io/ring.h with req_event and req_prod
168	 * corresponding to event_idx + 1 and new_idx respectively.
169	 * Note also that req_event and req_prod in Xen start at 1,
170	 * event indexes in virtio start at 0. */
171	return (__u16)(new_idx - event_idx - 1) < (__u16)(new_idx - old);
172}
173
174#endif /* _UAPI_LINUX_VIRTIO_RING_H */
175