1#ifndef _LINUX_VIRTIO_RING_H
2#define _LINUX_VIRTIO_RING_H
3/* An interface for efficient virtio implementation, currently for use by KVM,
4 * 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#include <stdint.h>
35#include "standard-headers/linux/types.h"
36#include "standard-headers/linux/virtio_types.h"
37
38/* This marks a buffer as continuing via the next field. */
39#define VRING_DESC_F_NEXT 1
40/* This marks a buffer as write-only (otherwise read-only). */
41#define VRING_DESC_F_WRITE 2
42/* This means the buffer contains a list of buffer descriptors. */
43#define VRING_DESC_F_INDIRECT 4
44
45/*
46 * Mark a descriptor as available or used in packed ring.
47 * Notice: they are defined as shifts instead of shifted values.
48 */
49#define VRING_PACKED_DESC_F_AVAIL 7
50#define VRING_PACKED_DESC_F_USED 15
51
52/* The Host uses this in used->flags to advise the Guest: don't kick me when
53 * you add a buffer. It's unreliable, so it's simply an optimization. Guest
54 * will still kick if it's out of buffers. */
55#define VRING_USED_F_NO_NOTIFY 1
56/* The Guest uses this in avail->flags to advise the Host: don't interrupt me
57 * when you consume a buffer. It's unreliable, so it's simply an
58 * optimization. */
59#define VRING_AVAIL_F_NO_INTERRUPT 1
60
61/* Enable events in packed ring. */
62#define VRING_PACKED_EVENT_FLAG_ENABLE 0x0
63/* Disable events in packed ring. */
64#define VRING_PACKED_EVENT_FLAG_DISABLE 0x1
65/*
66 * Enable events for a specific descriptor in packed ring.
67 * (as specified by Descriptor Ring Change Event Offset/Wrap Counter).
68 * Only valid if VIRTIO_RING_F_EVENT_IDX has been negotiated.
69 */
70#define VRING_PACKED_EVENT_FLAG_DESC 0x2
71
72/*
73 * Wrap counter bit shift in event suppression structure
74 * of packed ring.
75 */
76#define VRING_PACKED_EVENT_F_WRAP_CTR 15
77
78/* We support indirect buffer descriptors */
79#define VIRTIO_RING_F_INDIRECT_DESC 28
80
81/* The Guest publishes the used index for which it expects an interrupt
82 * at the end of the avail ring. Host should ignore the avail->flags field. */
83/* The Host publishes the avail index for which it expects a kick
84 * at the end of the used ring. Guest should ignore the used->flags field. */
85#define VIRTIO_RING_F_EVENT_IDX 29
86
87/* Virtio ring descriptors: 16 bytes. These can chain together via "next". */
88struct vring_desc {
89 /* Address (guest-physical). */
90 __virtio64 addr;
91 /* Length. */
92 __virtio32 len;
93 /* The flags as indicated above. */
94 __virtio16 flags;
95 /* We chain unused descriptors via this, too */
96 __virtio16 next;
97};
98
99struct vring_avail {
100 __virtio16 flags;
101 __virtio16 idx;
102 __virtio16 ring[];
103};
104
105/* uint32_t is used here for ids for padding reasons. */
106struct vring_used_elem {
107 /* Index of start of used descriptor chain. */
108 __virtio32 id;
109 /* Total length of the descriptor chain which was used (written to) */
110 __virtio32 len;
111};
112
113struct vring_used {
114 __virtio16 flags;
115 __virtio16 idx;
116 struct vring_used_elem ring[];
117};
118
119struct vring {
120 unsigned int num;
121
122 struct vring_desc *desc;
123
124 struct vring_avail *avail;
125
126 struct vring_used *used;
127};
128
129/* Alignment requirements for vring elements.
130 * When using pre-virtio 1.0 layout, these fall out naturally.
131 */
132#define VRING_AVAIL_ALIGN_SIZE 2
133#define VRING_USED_ALIGN_SIZE 4
134#define VRING_DESC_ALIGN_SIZE 16
135
136/* The standard layout for the ring is a continuous chunk of memory which looks
137 * like this. We assume num is a power of 2.
138 *
139 * struct vring
140 * {
141 * // The actual descriptors (16 bytes each)
142 * struct vring_desc desc[num];
143 *
144 * // A ring of available descriptor heads with free-running index.
145 * __virtio16 avail_flags;
146 * __virtio16 avail_idx;
147 * __virtio16 available[num];
148 * __virtio16 used_event_idx;
149 *
150 * // Padding to the next align boundary.
151 * char pad[];
152 *
153 * // A ring of used descriptor heads with free-running index.
154 * __virtio16 used_flags;
155 * __virtio16 used_idx;
156 * struct vring_used_elem used[num];
157 * __virtio16 avail_event_idx;
158 * };
159 */
160/* We publish the used event index at the end of the available ring, and vice
161 * versa. They are at the end for backwards compatibility. */
162#define vring_used_event(vr) ((vr)->avail->ring[(vr)->num])
163#define vring_avail_event(vr) (*(__virtio16 *)&(vr)->used->ring[(vr)->num])
164
165static inline void vring_init(struct vring *vr, unsigned int num, void *p,
166 unsigned long align)
167{
168 vr->num = num;
169 vr->desc = p;
170 vr->avail = p + num*sizeof(struct vring_desc);
171 vr->used = (void *)(((uintptr_t)&vr->avail->ring[num] + sizeof(__virtio16)
172 + align-1) & ~(align - 1));
173}
174
175static inline unsigned vring_size(unsigned int num, unsigned long align)
176{
177 return ((sizeof(struct vring_desc) * num + sizeof(__virtio16) * (3 + num)
178 + align - 1) & ~(align - 1))
179 + sizeof(__virtio16) * 3 + sizeof(struct vring_used_elem) * num;
180}
181
182/* The following is used with USED_EVENT_IDX and AVAIL_EVENT_IDX */
183/* Assuming a given event_idx value from the other side, if
184 * we have just incremented index from old to new_idx,
185 * should we trigger an event? */
186static inline int vring_need_event(uint16_t event_idx, uint16_t new_idx, uint16_t old)
187{
188 /* Note: Xen has similar logic for notification hold-off
189 * in include/xen/interface/io/ring.h with req_event and req_prod
190 * corresponding to event_idx + 1 and new_idx respectively.
191 * Note also that req_event and req_prod in Xen start at 1,
192 * event indexes in virtio start at 0. */
193 return (uint16_t)(new_idx - event_idx - 1) < (uint16_t)(new_idx - old);
194}
195
196struct vring_packed_desc_event {
197 /* Descriptor Ring Change Event Offset/Wrap Counter. */
198 uint16_t off_wrap;
199 /* Descriptor Ring Change Event Flags. */
200 uint16_t flags;
201};
202
203struct vring_packed_desc {
204 /* Buffer Address. */
205 uint64_t addr;
206 /* Buffer Length. */
207 uint32_t len;
208 /* Buffer ID. */
209 uint16_t id;
210 /* The flags depending on descriptor type. */
211 uint16_t flags;
212};
213
214#endif /* _LINUX_VIRTIO_RING_H */
215