Blame view
drivers/virtio/virtio_ring.c
13.8 KB
0a8a69dd7
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 |
/* Virtio ring implementation. * * Copyright 2007 Rusty Russell IBM Corporation * * 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., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA */ #include <linux/virtio.h> #include <linux/virtio_ring.h> |
e34f87256
|
21 |
#include <linux/virtio_config.h> |
0a8a69dd7
|
22 |
#include <linux/device.h> |
5a0e3ad6a
|
23 |
#include <linux/slab.h> |
b5a2c4f19
|
24 |
#include <linux/module.h> |
0a8a69dd7
|
25 |
|
d57ed95da
|
26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 |
/* virtio guest is communicating with a virtual "device" that actually runs on * a host processor. Memory barriers are used to control SMP effects. */ #ifdef CONFIG_SMP /* Where possible, use SMP barriers which are more lightweight than mandatory * barriers, because mandatory barriers control MMIO effects on accesses * through relaxed memory I/O windows (which virtio does not use). */ #define virtio_mb() smp_mb() #define virtio_rmb() smp_rmb() #define virtio_wmb() smp_wmb() #else /* We must force memory ordering even if guest is UP since host could be * running on another CPU, but SMP barriers are defined to barrier() in that * configuration. So fall back to mandatory barriers instead. */ #define virtio_mb() mb() #define virtio_rmb() rmb() #define virtio_wmb() wmb() #endif |
0a8a69dd7
|
43 44 |
#ifdef DEBUG /* For development, we want to crash whenever the ring is screwed. */ |
9499f5e7e
|
45 46 47 48 49 50 |
#define BAD_RING(_vq, fmt, args...) \ do { \ dev_err(&(_vq)->vq.vdev->dev, \ "%s:"fmt, (_vq)->vq.name, ##args); \ BUG(); \ } while (0) |
c5f841f17
|
51 52 53 54 |
/* Caller is supposed to guarantee no reentry. */ #define START_USE(_vq) \ do { \ if ((_vq)->in_use) \ |
9499f5e7e
|
55 56 57 |
panic("%s:in_use = %i ", \ (_vq)->vq.name, (_vq)->in_use); \ |
c5f841f17
|
58 |
(_vq)->in_use = __LINE__; \ |
9499f5e7e
|
59 |
} while (0) |
3a35ce7dc
|
60 |
#define END_USE(_vq) \ |
97a545ab6
|
61 |
do { BUG_ON(!(_vq)->in_use); (_vq)->in_use = 0; } while(0) |
0a8a69dd7
|
62 |
#else |
9499f5e7e
|
63 64 65 66 67 68 |
#define BAD_RING(_vq, fmt, args...) \ do { \ dev_err(&_vq->vq.vdev->dev, \ "%s:"fmt, (_vq)->vq.name, ##args); \ (_vq)->broken = true; \ } while (0) |
0a8a69dd7
|
69 70 71 72 73 74 75 76 77 78 79 80 81 |
#define START_USE(vq) #define END_USE(vq) #endif struct vring_virtqueue { struct virtqueue vq; /* Actual memory layout for this queue */ struct vring vring; /* Other side has made a mess, don't try any more. */ bool broken; |
9fa29b9df
|
82 83 |
/* Host supports indirect buffers */ bool indirect; |
a5c262c5f
|
84 85 |
/* Host publishes avail event idx */ bool event; |
0a8a69dd7
|
86 87 88 89 90 91 92 93 |
/* Number of free buffers */ unsigned int num_free; /* Head of free buffer list. */ unsigned int free_head; /* Number we've added since last sync. */ unsigned int num_added; /* Last used index we've seen. */ |
1bc4953ed
|
94 |
u16 last_used_idx; |
0a8a69dd7
|
95 96 97 98 99 100 101 102 103 104 105 106 107 108 |
/* How to notify other side. FIXME: commonalize hcalls! */ void (*notify)(struct virtqueue *vq); #ifdef DEBUG /* They're supposed to lock for us. */ unsigned int in_use; #endif /* Tokens for callbacks. */ void *data[]; }; #define to_vvq(_vq) container_of(_vq, struct vring_virtqueue, vq) |
9fa29b9df
|
109 110 111 112 |
/* Set up an indirect table of descriptors and add it to the queue. */ static int vring_add_indirect(struct vring_virtqueue *vq, struct scatterlist sg[], unsigned int out, |
bbd603efb
|
113 114 |
unsigned int in, gfp_t gfp) |
9fa29b9df
|
115 116 117 118 |
{ struct vring_desc *desc; unsigned head; int i; |
bbd603efb
|
119 |
desc = kmalloc((out + in) * sizeof(struct vring_desc), gfp); |
9fa29b9df
|
120 |
if (!desc) |
686d36378
|
121 |
return -ENOMEM; |
9fa29b9df
|
122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 |
/* Transfer entries from the sg list into the indirect page */ for (i = 0; i < out; i++) { desc[i].flags = VRING_DESC_F_NEXT; desc[i].addr = sg_phys(sg); desc[i].len = sg->length; desc[i].next = i+1; sg++; } for (; i < (out + in); i++) { desc[i].flags = VRING_DESC_F_NEXT|VRING_DESC_F_WRITE; desc[i].addr = sg_phys(sg); desc[i].len = sg->length; desc[i].next = i+1; sg++; } /* Last one doesn't continue. */ desc[i-1].flags &= ~VRING_DESC_F_NEXT; desc[i-1].next = 0; /* We're about to use a buffer */ vq->num_free--; /* Use a single buffer which doesn't continue */ head = vq->free_head; vq->vring.desc[head].flags = VRING_DESC_F_INDIRECT; vq->vring.desc[head].addr = virt_to_phys(desc); vq->vring.desc[head].len = i * sizeof(struct vring_desc); /* Update free pointer */ vq->free_head = vq->vring.desc[head].next; return head; } |
bbd603efb
|
157 158 159 160 161 162 |
int virtqueue_add_buf_gfp(struct virtqueue *_vq, struct scatterlist sg[], unsigned int out, unsigned int in, void *data, gfp_t gfp) |
0a8a69dd7
|
163 164 |
{ struct vring_virtqueue *vq = to_vvq(_vq); |
1fe9b6fef
|
165 166 |
unsigned int i, avail, uninitialized_var(prev); int head; |
0a8a69dd7
|
167 |
|
9fa29b9df
|
168 |
START_USE(vq); |
0a8a69dd7
|
169 |
BUG_ON(data == NULL); |
9fa29b9df
|
170 171 172 173 |
/* If the host supports indirect descriptor tables, and we have multiple * buffers, then go indirect. FIXME: tune this threshold */ if (vq->indirect && (out + in) > 1 && vq->num_free) { |
bbd603efb
|
174 |
head = vring_add_indirect(vq, sg, out, in, gfp); |
1fe9b6fef
|
175 |
if (likely(head >= 0)) |
9fa29b9df
|
176 177 |
goto add_head; } |
0a8a69dd7
|
178 179 |
BUG_ON(out + in > vq->vring.num); BUG_ON(out + in == 0); |
0a8a69dd7
|
180 181 182 183 |
if (vq->num_free < out + in) { pr_debug("Can't add buf len %i - avail = %i ", out + in, vq->num_free); |
44653eae1
|
184 185 186 187 188 |
/* FIXME: for historical reasons, we force a notify here if * there are outgoing parts to the buffer. Presumably the * host should service the ring ASAP. */ if (out) vq->notify(&vq->vq); |
0a8a69dd7
|
189 190 191 192 193 194 195 196 197 198 |
END_USE(vq); return -ENOSPC; } /* We're about to use some buffers from the free list. */ vq->num_free -= out + in; head = vq->free_head; for (i = vq->free_head; out; i = vq->vring.desc[i].next, out--) { vq->vring.desc[i].flags = VRING_DESC_F_NEXT; |
15f9c8903
|
199 |
vq->vring.desc[i].addr = sg_phys(sg); |
0a8a69dd7
|
200 201 202 203 204 205 |
vq->vring.desc[i].len = sg->length; prev = i; sg++; } for (; in; i = vq->vring.desc[i].next, in--) { vq->vring.desc[i].flags = VRING_DESC_F_NEXT|VRING_DESC_F_WRITE; |
15f9c8903
|
206 |
vq->vring.desc[i].addr = sg_phys(sg); |
0a8a69dd7
|
207 208 209 210 211 212 213 214 215 |
vq->vring.desc[i].len = sg->length; prev = i; sg++; } /* Last one doesn't continue. */ vq->vring.desc[prev].flags &= ~VRING_DESC_F_NEXT; /* Update free pointer */ vq->free_head = i; |
9fa29b9df
|
216 |
add_head: |
0a8a69dd7
|
217 218 219 220 221 222 223 224 225 226 227 |
/* Set token. */ vq->data[head] = data; /* Put entry in available array (but don't update avail->idx until they * do sync). FIXME: avoid modulus here? */ avail = (vq->vring.avail->idx + vq->num_added++) % vq->vring.num; vq->vring.avail->ring[avail] = head; pr_debug("Added buffer head %i to %p ", head, vq); END_USE(vq); |
3c1b27d50
|
228 |
|
3c1b27d50
|
229 |
return vq->num_free; |
0a8a69dd7
|
230 |
} |
bbd603efb
|
231 |
EXPORT_SYMBOL_GPL(virtqueue_add_buf_gfp); |
0a8a69dd7
|
232 |
|
7c5e9ed0c
|
233 |
void virtqueue_kick(struct virtqueue *_vq) |
0a8a69dd7
|
234 235 |
{ struct vring_virtqueue *vq = to_vvq(_vq); |
a5c262c5f
|
236 |
u16 new, old; |
0a8a69dd7
|
237 238 239 |
START_USE(vq); /* Descriptors and available array need to be set before we expose the * new available array entries. */ |
d57ed95da
|
240 |
virtio_wmb(); |
0a8a69dd7
|
241 |
|
a5c262c5f
|
242 243 |
old = vq->vring.avail->idx; new = vq->vring.avail->idx = old + vq->num_added; |
0a8a69dd7
|
244 245 246 |
vq->num_added = 0; /* Need to update avail index before checking if we should notify */ |
d57ed95da
|
247 |
virtio_mb(); |
0a8a69dd7
|
248 |
|
a5c262c5f
|
249 250 251 |
if (vq->event ? vring_need_event(vring_avail_event(&vq->vring), new, old) : !(vq->vring.used->flags & VRING_USED_F_NO_NOTIFY)) |
0a8a69dd7
|
252 253 254 255 256 |
/* Prod other side to tell it about changes. */ vq->notify(&vq->vq); END_USE(vq); } |
7c5e9ed0c
|
257 |
EXPORT_SYMBOL_GPL(virtqueue_kick); |
0a8a69dd7
|
258 259 260 261 262 263 264 265 266 267 |
static void detach_buf(struct vring_virtqueue *vq, unsigned int head) { unsigned int i; /* Clear data ptr. */ vq->data[head] = NULL; /* Put back on free list: find end */ i = head; |
9fa29b9df
|
268 269 270 271 |
/* Free the indirect table */ if (vq->vring.desc[i].flags & VRING_DESC_F_INDIRECT) kfree(phys_to_virt(vq->vring.desc[i].addr)); |
0a8a69dd7
|
272 273 274 275 276 277 278 279 280 281 |
while (vq->vring.desc[i].flags & VRING_DESC_F_NEXT) { i = vq->vring.desc[i].next; vq->num_free++; } vq->vring.desc[i].next = vq->free_head; vq->free_head = head; /* Plus final descriptor */ vq->num_free++; } |
0a8a69dd7
|
282 283 284 285 |
static inline bool more_used(const struct vring_virtqueue *vq) { return vq->last_used_idx != vq->vring.used->idx; } |
7c5e9ed0c
|
286 |
void *virtqueue_get_buf(struct virtqueue *_vq, unsigned int *len) |
0a8a69dd7
|
287 288 289 290 291 292 |
{ struct vring_virtqueue *vq = to_vvq(_vq); void *ret; unsigned int i; START_USE(vq); |
5ef827526
|
293 294 295 296 |
if (unlikely(vq->broken)) { END_USE(vq); return NULL; } |
0a8a69dd7
|
297 298 299 300 301 302 |
if (!more_used(vq)) { pr_debug("No more buffers in queue "); END_USE(vq); return NULL; } |
2d61ba950
|
303 |
/* Only get used array entries after they have been exposed by host. */ |
d57ed95da
|
304 |
virtio_rmb(); |
2d61ba950
|
305 |
|
0a8a69dd7
|
306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 |
i = vq->vring.used->ring[vq->last_used_idx%vq->vring.num].id; *len = vq->vring.used->ring[vq->last_used_idx%vq->vring.num].len; if (unlikely(i >= vq->vring.num)) { BAD_RING(vq, "id %u out of range ", i); return NULL; } if (unlikely(!vq->data[i])) { BAD_RING(vq, "id %u is not a head! ", i); return NULL; } /* detach_buf clears data, so grab it now. */ ret = vq->data[i]; detach_buf(vq, i); vq->last_used_idx++; |
a5c262c5f
|
324 325 326 327 328 329 330 |
/* If we expect an interrupt for the next entry, tell host * by writing event index and flush out the write before * the read in the next get_buf call. */ if (!(vq->vring.avail->flags & VRING_AVAIL_F_NO_INTERRUPT)) { vring_used_event(&vq->vring) = vq->last_used_idx; virtio_mb(); } |
0a8a69dd7
|
331 332 333 |
END_USE(vq); return ret; } |
7c5e9ed0c
|
334 |
EXPORT_SYMBOL_GPL(virtqueue_get_buf); |
0a8a69dd7
|
335 |
|
7c5e9ed0c
|
336 |
void virtqueue_disable_cb(struct virtqueue *_vq) |
18445c4d5
|
337 338 |
{ struct vring_virtqueue *vq = to_vvq(_vq); |
18445c4d5
|
339 |
vq->vring.avail->flags |= VRING_AVAIL_F_NO_INTERRUPT; |
18445c4d5
|
340 |
} |
7c5e9ed0c
|
341 |
EXPORT_SYMBOL_GPL(virtqueue_disable_cb); |
18445c4d5
|
342 |
|
7c5e9ed0c
|
343 |
bool virtqueue_enable_cb(struct virtqueue *_vq) |
0a8a69dd7
|
344 345 346 347 |
{ struct vring_virtqueue *vq = to_vvq(_vq); START_USE(vq); |
0a8a69dd7
|
348 349 350 |
/* We optimistically turn back on interrupts, then check if there was * more to do. */ |
a5c262c5f
|
351 352 353 |
/* Depending on the VIRTIO_RING_F_EVENT_IDX feature, we need to * either clear the flags bit or point the event index at the next * entry. Always do both to keep code simple. */ |
0a8a69dd7
|
354 |
vq->vring.avail->flags &= ~VRING_AVAIL_F_NO_INTERRUPT; |
a5c262c5f
|
355 |
vring_used_event(&vq->vring) = vq->last_used_idx; |
d57ed95da
|
356 |
virtio_mb(); |
0a8a69dd7
|
357 |
if (unlikely(more_used(vq))) { |
0a8a69dd7
|
358 359 360 361 362 363 364 |
END_USE(vq); return false; } END_USE(vq); return true; } |
7c5e9ed0c
|
365 |
EXPORT_SYMBOL_GPL(virtqueue_enable_cb); |
0a8a69dd7
|
366 |
|
7ab358c23
|
367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 |
bool virtqueue_enable_cb_delayed(struct virtqueue *_vq) { struct vring_virtqueue *vq = to_vvq(_vq); u16 bufs; START_USE(vq); /* We optimistically turn back on interrupts, then check if there was * more to do. */ /* Depending on the VIRTIO_RING_F_USED_EVENT_IDX feature, we need to * either clear the flags bit or point the event index at the next * entry. Always do both to keep code simple. */ vq->vring.avail->flags &= ~VRING_AVAIL_F_NO_INTERRUPT; /* TODO: tune this threshold */ bufs = (u16)(vq->vring.avail->idx - vq->last_used_idx) * 3 / 4; vring_used_event(&vq->vring) = vq->last_used_idx + bufs; virtio_mb(); if (unlikely((u16)(vq->vring.used->idx - vq->last_used_idx) > bufs)) { END_USE(vq); return false; } END_USE(vq); return true; } EXPORT_SYMBOL_GPL(virtqueue_enable_cb_delayed); |
7c5e9ed0c
|
393 |
void *virtqueue_detach_unused_buf(struct virtqueue *_vq) |
c021eac41
|
394 395 396 397 398 399 400 401 402 403 404 405 406 |
{ struct vring_virtqueue *vq = to_vvq(_vq); unsigned int i; void *buf; START_USE(vq); for (i = 0; i < vq->vring.num; i++) { if (!vq->data[i]) continue; /* detach_buf clears data, so grab it now. */ buf = vq->data[i]; detach_buf(vq, i); |
b3258ff1d
|
407 |
vq->vring.avail->idx--; |
c021eac41
|
408 409 410 411 412 413 414 415 416 |
END_USE(vq); return buf; } /* That should have freed everything. */ BUG_ON(vq->num_free != vq->vring.num); END_USE(vq); return NULL; } |
7c5e9ed0c
|
417 |
EXPORT_SYMBOL_GPL(virtqueue_detach_unused_buf); |
c021eac41
|
418 |
|
0a8a69dd7
|
419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 |
irqreturn_t vring_interrupt(int irq, void *_vq) { struct vring_virtqueue *vq = to_vvq(_vq); if (!more_used(vq)) { pr_debug("virtqueue interrupt with no work for %p ", vq); return IRQ_NONE; } if (unlikely(vq->broken)) return IRQ_HANDLED; pr_debug("virtqueue callback for %p (%p) ", vq, vq->vq.callback); |
18445c4d5
|
434 435 |
if (vq->vq.callback) vq->vq.callback(&vq->vq); |
0a8a69dd7
|
436 437 438 |
return IRQ_HANDLED; } |
c6fd47011
|
439 |
EXPORT_SYMBOL_GPL(vring_interrupt); |
0a8a69dd7
|
440 |
|
0a8a69dd7
|
441 |
struct virtqueue *vring_new_virtqueue(unsigned int num, |
87c7d57c1
|
442 |
unsigned int vring_align, |
0a8a69dd7
|
443 444 445 |
struct virtio_device *vdev, void *pages, void (*notify)(struct virtqueue *), |
9499f5e7e
|
446 447 |
void (*callback)(struct virtqueue *), const char *name) |
0a8a69dd7
|
448 449 450 |
{ struct vring_virtqueue *vq; unsigned int i; |
42b36cc0c
|
451 452 453 454 455 456 |
/* We assume num is a power of 2. */ if (num & (num - 1)) { dev_warn(&vdev->dev, "Bad virtqueue length %u ", num); return NULL; } |
0a8a69dd7
|
457 458 459 |
vq = kmalloc(sizeof(*vq) + sizeof(void *)*num, GFP_KERNEL); if (!vq) return NULL; |
87c7d57c1
|
460 |
vring_init(&vq->vring, num, pages, vring_align); |
0a8a69dd7
|
461 462 |
vq->vq.callback = callback; vq->vq.vdev = vdev; |
9499f5e7e
|
463 |
vq->vq.name = name; |
0a8a69dd7
|
464 465 466 467 |
vq->notify = notify; vq->broken = false; vq->last_used_idx = 0; vq->num_added = 0; |
9499f5e7e
|
468 |
list_add_tail(&vq->vq.list, &vdev->vqs); |
0a8a69dd7
|
469 470 471 |
#ifdef DEBUG vq->in_use = false; #endif |
9fa29b9df
|
472 |
vq->indirect = virtio_has_feature(vdev, VIRTIO_RING_F_INDIRECT_DESC); |
a5c262c5f
|
473 |
vq->event = virtio_has_feature(vdev, VIRTIO_RING_F_EVENT_IDX); |
9fa29b9df
|
474 |
|
0a8a69dd7
|
475 476 477 478 479 480 481 |
/* No callback? Tell other side not to bother us. */ if (!callback) vq->vring.avail->flags |= VRING_AVAIL_F_NO_INTERRUPT; /* Put everything in free lists. */ vq->num_free = num; vq->free_head = 0; |
3b8706240
|
482 |
for (i = 0; i < num-1; i++) { |
0a8a69dd7
|
483 |
vq->vring.desc[i].next = i+1; |
3b8706240
|
484 485 486 |
vq->data[i] = NULL; } vq->data[i] = NULL; |
0a8a69dd7
|
487 488 489 |
return &vq->vq; } |
c6fd47011
|
490 |
EXPORT_SYMBOL_GPL(vring_new_virtqueue); |
0a8a69dd7
|
491 492 493 |
void vring_del_virtqueue(struct virtqueue *vq) { |
9499f5e7e
|
494 |
list_del(&vq->list); |
0a8a69dd7
|
495 496 |
kfree(to_vvq(vq)); } |
c6fd47011
|
497 |
EXPORT_SYMBOL_GPL(vring_del_virtqueue); |
0a8a69dd7
|
498 |
|
e34f87256
|
499 500 501 502 503 504 505 |
/* Manipulates transport-specific feature bits. */ void vring_transport_features(struct virtio_device *vdev) { unsigned int i; for (i = VIRTIO_TRANSPORT_F_START; i < VIRTIO_TRANSPORT_F_END; i++) { switch (i) { |
9fa29b9df
|
506 507 |
case VIRTIO_RING_F_INDIRECT_DESC: break; |
a5c262c5f
|
508 509 |
case VIRTIO_RING_F_EVENT_IDX: break; |
e34f87256
|
510 511 512 513 514 515 516 |
default: /* We don't understand this bit. */ clear_bit(i, vdev->features); } } } EXPORT_SYMBOL_GPL(vring_transport_features); |
8f9f4668b
|
517 518 519 520 521 522 523 524 525 |
/* return the size of the vring within the virtqueue */ unsigned int virtqueue_get_vring_size(struct virtqueue *_vq) { struct vring_virtqueue *vq = to_vvq(_vq); return vq->vring.num; } EXPORT_SYMBOL_GPL(virtqueue_get_vring_size); |
c6fd47011
|
526 |
MODULE_LICENSE("GPL"); |