virtio-pci.c 49.6 KB
Newer Older
P
Paul Brook 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13
/*
 * Virtio PCI Bindings
 *
 * Copyright IBM, Corp. 2007
 * Copyright (c) 2009 CodeSourcery
 *
 * Authors:
 *  Anthony Liguori   <aliguori@us.ibm.com>
 *  Paul Brook        <paul@codesourcery.com>
 *
 * This work is licensed under the terms of the GNU GPL, version 2.  See
 * the COPYING file in the top-level directory.
 *
14 15
 * Contributions after 2012-01-13 are licensed under the terms of the
 * GNU GPL, version 2 or (at your option) any later version.
P
Paul Brook 已提交
16 17 18 19
 */

#include <inttypes.h>

P
Paolo Bonzini 已提交
20 21 22 23 24 25
#include "hw/virtio/virtio.h"
#include "hw/virtio/virtio-blk.h"
#include "hw/virtio/virtio-net.h"
#include "hw/virtio/virtio-serial.h"
#include "hw/virtio/virtio-scsi.h"
#include "hw/virtio/virtio-balloon.h"
26
#include "hw/pci/pci.h"
27
#include "qemu/error-report.h"
28 29 30
#include "hw/pci/msi.h"
#include "hw/pci/msix.h"
#include "hw/loader.h"
31 32
#include "sysemu/kvm.h"
#include "sysemu/blockdev.h"
33
#include "virtio-pci.h"
34
#include "qemu/range.h"
P
Paolo Bonzini 已提交
35
#include "hw/virtio/virtio-bus.h"
36
#include "qapi/visitor.h"
P
Paul Brook 已提交
37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

/* from Linux's linux/virtio_pci.h */

/* A 32-bit r/o bitmask of the features supported by the host */
#define VIRTIO_PCI_HOST_FEATURES        0

/* A 32-bit r/w bitmask of features activated by the guest */
#define VIRTIO_PCI_GUEST_FEATURES       4

/* A 32-bit r/w PFN for the currently selected queue */
#define VIRTIO_PCI_QUEUE_PFN            8

/* A 16-bit r/o queue size for the currently selected queue */
#define VIRTIO_PCI_QUEUE_NUM            12

/* A 16-bit r/w queue selector */
#define VIRTIO_PCI_QUEUE_SEL            14

/* A 16-bit r/w queue notifier */
#define VIRTIO_PCI_QUEUE_NOTIFY         16

/* An 8-bit device status register.  */
#define VIRTIO_PCI_STATUS               18

/* An 8-bit r/o interrupt status register.  Reading the value will return the
 * current contents of the ISR and will also clear it.  This is effectively
 * a read-and-acknowledge. */
#define VIRTIO_PCI_ISR                  19

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83
/* MSI-X registers: only enabled if MSI-X is enabled. */
/* A 16-bit vector for configuration changes. */
#define VIRTIO_MSI_CONFIG_VECTOR        20
/* A 16-bit vector for selected queue notifications. */
#define VIRTIO_MSI_QUEUE_VECTOR         22

/* Config space size */
#define VIRTIO_PCI_CONFIG_NOMSI         20
#define VIRTIO_PCI_CONFIG_MSI           24
#define VIRTIO_PCI_REGION_SIZE(dev)     (msix_present(dev) ? \
                                         VIRTIO_PCI_CONFIG_MSI : \
                                         VIRTIO_PCI_CONFIG_NOMSI)

/* The remaining space is defined by each driver as the per-driver
 * configuration space */
#define VIRTIO_PCI_CONFIG(dev)          (msix_enabled(dev) ? \
                                         VIRTIO_PCI_CONFIG_MSI : \
                                         VIRTIO_PCI_CONFIG_NOMSI)
P
Paul Brook 已提交
84 85 86 87 88

/* How many bits to shift physical queue address written to QUEUE_PFN.
 * 12 is historical, and due to x86 page size. */
#define VIRTIO_PCI_QUEUE_ADDR_SHIFT    12

89 90
/* Flags track per-device state like workarounds for quirks in older guests. */
#define VIRTIO_PCI_FLAG_BUS_MASTER_BUG  (1 << 0)
91

92 93 94
/* HACK for virtio to determine if it's running a big endian guest */
bool virtio_is_big_endian(void);

95 96
static void virtio_pci_bus_new(VirtioBusState *bus, VirtIOPCIProxy *dev);

P
Paul Brook 已提交
97
/* virtio device */
98 99 100 101 102
/* DeviceState to VirtIOPCIProxy. For use off data-path. TODO: use QOM. */
static inline VirtIOPCIProxy *to_virtio_pci_proxy(DeviceState *d)
{
    return container_of(d, VirtIOPCIProxy, pci_dev.qdev);
}
P
Paul Brook 已提交
103

104 105 106 107
/* DeviceState to VirtIOPCIProxy. Note: used on datapath,
 * be careful and test performance if you change this.
 */
static inline VirtIOPCIProxy *to_virtio_pci_proxy_fast(DeviceState *d)
P
Paul Brook 已提交
108
{
109 110 111 112 113 114
    return container_of(d, VirtIOPCIProxy, pci_dev.qdev);
}

static void virtio_pci_notify(DeviceState *d, uint16_t vector)
{
    VirtIOPCIProxy *proxy = to_virtio_pci_proxy_fast(d);
115 116 117 118
    if (msix_enabled(&proxy->pci_dev))
        msix_notify(&proxy->pci_dev, vector);
    else
        qemu_set_irq(proxy->pci_dev.irq[0], proxy->vdev->isr & 1);
P
Paul Brook 已提交
119 120
}

121
static void virtio_pci_save_config(DeviceState *d, QEMUFile *f)
122
{
123
    VirtIOPCIProxy *proxy = to_virtio_pci_proxy(d);
124 125 126 127 128 129
    pci_device_save(&proxy->pci_dev, f);
    msix_save(&proxy->pci_dev, f);
    if (msix_present(&proxy->pci_dev))
        qemu_put_be16(f, proxy->vdev->config_vector);
}

130
static void virtio_pci_save_queue(DeviceState *d, int n, QEMUFile *f)
131
{
132
    VirtIOPCIProxy *proxy = to_virtio_pci_proxy(d);
133 134 135 136
    if (msix_present(&proxy->pci_dev))
        qemu_put_be16(f, virtio_queue_vector(proxy->vdev, n));
}

137
static int virtio_pci_load_config(DeviceState *d, QEMUFile *f)
138
{
139
    VirtIOPCIProxy *proxy = to_virtio_pci_proxy(d);
140 141
    int ret;
    ret = pci_device_load(&proxy->pci_dev, f);
142
    if (ret) {
143
        return ret;
144
    }
145
    msix_unuse_all_vectors(&proxy->pci_dev);
146
    msix_load(&proxy->pci_dev, f);
147
    if (msix_present(&proxy->pci_dev)) {
148
        qemu_get_be16s(f, &proxy->vdev->config_vector);
149 150 151 152 153 154
    } else {
        proxy->vdev->config_vector = VIRTIO_NO_VECTOR;
    }
    if (proxy->vdev->config_vector != VIRTIO_NO_VECTOR) {
        return msix_vector_use(&proxy->pci_dev, proxy->vdev->config_vector);
    }
155 156 157
    return 0;
}

158
static int virtio_pci_load_queue(DeviceState *d, int n, QEMUFile *f)
159
{
160
    VirtIOPCIProxy *proxy = to_virtio_pci_proxy(d);
161
    uint16_t vector;
162 163 164 165 166
    if (msix_present(&proxy->pci_dev)) {
        qemu_get_be16s(f, &vector);
    } else {
        vector = VIRTIO_NO_VECTOR;
    }
167
    virtio_queue_set_vector(proxy->vdev, n, vector);
168 169 170
    if (vector != VIRTIO_NO_VECTOR) {
        return msix_vector_use(&proxy->pci_dev, vector);
    }
171 172 173
    return 0;
}

174
static int virtio_pci_set_host_notifier_internal(VirtIOPCIProxy *proxy,
P
Paolo Bonzini 已提交
175
                                                 int n, bool assign, bool set_handler)
176 177 178
{
    VirtQueue *vq = virtio_get_queue(proxy->vdev, n);
    EventNotifier *notifier = virtio_queue_get_host_notifier(vq);
A
Avi Kivity 已提交
179 180
    int r = 0;

181 182 183
    if (assign) {
        r = event_notifier_init(notifier, 1);
        if (r < 0) {
184 185
            error_report("%s: unable to init event notifier: %d",
                         __func__, r);
186 187
            return r;
        }
P
Paolo Bonzini 已提交
188
        virtio_queue_set_host_notifier_fd_handler(vq, true, set_handler);
A
Avi Kivity 已提交
189
        memory_region_add_eventfd(&proxy->bar, VIRTIO_PCI_QUEUE_NOTIFY, 2,
190
                                  true, n, notifier);
191
    } else {
A
Avi Kivity 已提交
192
        memory_region_del_eventfd(&proxy->bar, VIRTIO_PCI_QUEUE_NOTIFY, 2,
193
                                  true, n, notifier);
P
Paolo Bonzini 已提交
194
        virtio_queue_set_host_notifier_fd_handler(vq, false, false);
195 196 197 198 199
        event_notifier_cleanup(notifier);
    }
    return r;
}

200
static void virtio_pci_start_ioeventfd(VirtIOPCIProxy *proxy)
201 202 203 204 205 206
{
    int n, r;

    if (!(proxy->flags & VIRTIO_PCI_FLAG_USE_IOEVENTFD) ||
        proxy->ioeventfd_disabled ||
        proxy->ioeventfd_started) {
207
        return;
208 209 210 211 212 213 214
    }

    for (n = 0; n < VIRTIO_PCI_QUEUE_MAX; n++) {
        if (!virtio_queue_get_num(proxy->vdev, n)) {
            continue;
        }

P
Paolo Bonzini 已提交
215
        r = virtio_pci_set_host_notifier_internal(proxy, n, true, true);
216 217 218 219 220
        if (r < 0) {
            goto assign_error;
        }
    }
    proxy->ioeventfd_started = true;
221
    return;
222 223 224 225 226 227 228

assign_error:
    while (--n >= 0) {
        if (!virtio_queue_get_num(proxy->vdev, n)) {
            continue;
        }

P
Paolo Bonzini 已提交
229
        r = virtio_pci_set_host_notifier_internal(proxy, n, false, false);
230
        assert(r >= 0);
231 232
    }
    proxy->ioeventfd_started = false;
233
    error_report("%s: failed. Fallback to a userspace (slower).", __func__);
234 235
}

236
static void virtio_pci_stop_ioeventfd(VirtIOPCIProxy *proxy)
237
{
238
    int r;
239 240 241
    int n;

    if (!proxy->ioeventfd_started) {
242
        return;
243 244 245 246 247 248 249
    }

    for (n = 0; n < VIRTIO_PCI_QUEUE_MAX; n++) {
        if (!virtio_queue_get_num(proxy->vdev, n)) {
            continue;
        }

P
Paolo Bonzini 已提交
250
        r = virtio_pci_set_host_notifier_internal(proxy, n, false, false);
251
        assert(r >= 0);
252 253 254 255
    }
    proxy->ioeventfd_started = false;
}

P
Paul Brook 已提交
256 257 258 259
static void virtio_ioport_write(void *opaque, uint32_t addr, uint32_t val)
{
    VirtIOPCIProxy *proxy = opaque;
    VirtIODevice *vdev = proxy->vdev;
A
Avi Kivity 已提交
260
    hwaddr pa;
P
Paul Brook 已提交
261 262 263

    switch (addr) {
    case VIRTIO_PCI_GUEST_FEATURES:
264 265 266 267
        /* Guest does not negotiate properly?  We have to assume nothing. */
        if (val & (1 << VIRTIO_F_BAD_FEATURE)) {
            val = virtio_bus_get_vdev_bad_features(&proxy->bus);
        }
268
        virtio_set_features(vdev, val);
P
Paul Brook 已提交
269 270
        break;
    case VIRTIO_PCI_QUEUE_PFN:
A
Avi Kivity 已提交
271
        pa = (hwaddr)val << VIRTIO_PCI_QUEUE_ADDR_SHIFT;
272
        if (pa == 0) {
273
            virtio_pci_stop_ioeventfd(proxy);
274 275 276
            virtio_reset(proxy->vdev);
            msix_unuse_all_vectors(&proxy->pci_dev);
        }
277 278
        else
            virtio_queue_set_addr(vdev, vdev->queue_sel, pa);
P
Paul Brook 已提交
279 280 281 282 283 284
        break;
    case VIRTIO_PCI_QUEUE_SEL:
        if (val < VIRTIO_PCI_QUEUE_MAX)
            vdev->queue_sel = val;
        break;
    case VIRTIO_PCI_QUEUE_NOTIFY:
285 286 287
        if (val < VIRTIO_PCI_QUEUE_MAX) {
            virtio_queue_notify(vdev, val);
        }
P
Paul Brook 已提交
288 289
        break;
    case VIRTIO_PCI_STATUS:
290 291 292 293
        if (!(val & VIRTIO_CONFIG_S_DRIVER_OK)) {
            virtio_pci_stop_ioeventfd(proxy);
        }

294
        virtio_set_status(vdev, val & 0xFF);
295 296 297 298 299

        if (val & VIRTIO_CONFIG_S_DRIVER_OK) {
            virtio_pci_start_ioeventfd(proxy);
        }

300 301 302 303
        if (vdev->status == 0) {
            virtio_reset(proxy->vdev);
            msix_unuse_all_vectors(&proxy->pci_dev);
        }
304 305 306 307 308 309

        /* Linux before 2.6.34 sets the device as OK without enabling
           the PCI device bus master bit. In this case we need to disable
           some safety checks. */
        if ((val & VIRTIO_CONFIG_S_DRIVER_OK) &&
            !(proxy->pci_dev.config[PCI_COMMAND] & PCI_COMMAND_MASTER)) {
310
            proxy->flags |= VIRTIO_PCI_FLAG_BUS_MASTER_BUG;
311
        }
P
Paul Brook 已提交
312
        break;
313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328
    case VIRTIO_MSI_CONFIG_VECTOR:
        msix_vector_unuse(&proxy->pci_dev, vdev->config_vector);
        /* Make it possible for guest to discover an error took place. */
        if (msix_vector_use(&proxy->pci_dev, val) < 0)
            val = VIRTIO_NO_VECTOR;
        vdev->config_vector = val;
        break;
    case VIRTIO_MSI_QUEUE_VECTOR:
        msix_vector_unuse(&proxy->pci_dev,
                          virtio_queue_vector(vdev, vdev->queue_sel));
        /* Make it possible for guest to discover an error took place. */
        if (msix_vector_use(&proxy->pci_dev, val) < 0)
            val = VIRTIO_NO_VECTOR;
        virtio_queue_set_vector(vdev, vdev->queue_sel, val);
        break;
    default:
329 330
        error_report("%s: unexpected address 0x%x value 0x%x",
                     __func__, addr, val);
331
        break;
P
Paul Brook 已提交
332 333 334
    }
}

335
static uint32_t virtio_ioport_read(VirtIOPCIProxy *proxy, uint32_t addr)
P
Paul Brook 已提交
336 337 338 339 340 341
{
    VirtIODevice *vdev = proxy->vdev;
    uint32_t ret = 0xFFFFFFFF;

    switch (addr) {
    case VIRTIO_PCI_HOST_FEATURES:
342
        ret = proxy->host_features;
P
Paul Brook 已提交
343 344
        break;
    case VIRTIO_PCI_GUEST_FEATURES:
345
        ret = vdev->guest_features;
P
Paul Brook 已提交
346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363
        break;
    case VIRTIO_PCI_QUEUE_PFN:
        ret = virtio_queue_get_addr(vdev, vdev->queue_sel)
              >> VIRTIO_PCI_QUEUE_ADDR_SHIFT;
        break;
    case VIRTIO_PCI_QUEUE_NUM:
        ret = virtio_queue_get_num(vdev, vdev->queue_sel);
        break;
    case VIRTIO_PCI_QUEUE_SEL:
        ret = vdev->queue_sel;
        break;
    case VIRTIO_PCI_STATUS:
        ret = vdev->status;
        break;
    case VIRTIO_PCI_ISR:
        /* reading from the ISR also clears it. */
        ret = vdev->isr;
        vdev->isr = 0;
364
        qemu_set_irq(proxy->pci_dev.irq[0], 0);
P
Paul Brook 已提交
365
        break;
366 367 368 369 370 371
    case VIRTIO_MSI_CONFIG_VECTOR:
        ret = vdev->config_vector;
        break;
    case VIRTIO_MSI_QUEUE_VECTOR:
        ret = virtio_queue_vector(vdev, vdev->queue_sel);
        break;
P
Paul Brook 已提交
372 373 374 375 376 377 378
    default:
        break;
    }

    return ret;
}

379 380
static uint64_t virtio_pci_config_read(void *opaque, hwaddr addr,
                                       unsigned size)
P
Paul Brook 已提交
381 382
{
    VirtIOPCIProxy *proxy = opaque;
383
    uint32_t config = VIRTIO_PCI_CONFIG(&proxy->pci_dev);
384
    uint64_t val = 0;
385
    if (addr < config) {
386
        return virtio_ioport_read(proxy, addr);
387 388
    }
    addr -= config;
P
Paul Brook 已提交
389

390 391 392 393 394 395
    switch (size) {
    case 1:
        val = virtio_config_readb(proxy->vdev, addr);
        break;
    case 2:
        val = virtio_config_readw(proxy->vdev, addr);
396 397 398
        if (virtio_is_big_endian()) {
            val = bswap16(val);
        }
399 400 401
        break;
    case 4:
        val = virtio_config_readl(proxy->vdev, addr);
402 403 404
        if (virtio_is_big_endian()) {
            val = bswap32(val);
        }
405
        break;
406
    }
407
    return val;
P
Paul Brook 已提交
408 409
}

410 411
static void virtio_pci_config_write(void *opaque, hwaddr addr,
                                    uint64_t val, unsigned size)
P
Paul Brook 已提交
412 413
{
    VirtIOPCIProxy *proxy = opaque;
414 415 416 417 418 419
    uint32_t config = VIRTIO_PCI_CONFIG(&proxy->pci_dev);
    if (addr < config) {
        virtio_ioport_write(proxy, addr, val);
        return;
    }
    addr -= config;
420 421 422 423 424 425 426 427 428
    /*
     * Virtio-PCI is odd. Ioports are LE but config space is target native
     * endian.
     */
    switch (size) {
    case 1:
        virtio_config_writeb(proxy->vdev, addr, val);
        break;
    case 2:
429 430 431
        if (virtio_is_big_endian()) {
            val = bswap16(val);
        }
432 433 434
        virtio_config_writew(proxy->vdev, addr, val);
        break;
    case 4:
435 436 437
        if (virtio_is_big_endian()) {
            val = bswap32(val);
        }
438 439
        virtio_config_writel(proxy->vdev, addr, val);
        break;
440
    }
P
Paul Brook 已提交
441 442
}

A
Avi Kivity 已提交
443
static const MemoryRegionOps virtio_pci_config_ops = {
444 445 446 447 448 449
    .read = virtio_pci_config_read,
    .write = virtio_pci_config_write,
    .impl = {
        .min_access_size = 1,
        .max_access_size = 4,
    },
450
    .endianness = DEVICE_LITTLE_ENDIAN,
A
Avi Kivity 已提交
451
};
452 453 454 455

static void virtio_write_config(PCIDevice *pci_dev, uint32_t address,
                                uint32_t val, int len)
{
456 457
    VirtIOPCIProxy *proxy = DO_UPCAST(VirtIOPCIProxy, pci_dev, pci_dev);

458 459 460 461 462 463 464 465
    pci_default_write_config(pci_dev, address, val, len);

    if (range_covers_byte(address, len, PCI_COMMAND) &&
        !(pci_dev->config[PCI_COMMAND] & PCI_COMMAND_MASTER) &&
        !(proxy->flags & VIRTIO_PCI_FLAG_BUS_MASTER_BUG)) {
        virtio_pci_stop_ioeventfd(proxy);
        virtio_set_status(proxy->vdev,
                          proxy->vdev->status & ~VIRTIO_CONFIG_S_DRIVER_OK);
466
    }
P
Paul Brook 已提交
467 468
}

469
static unsigned virtio_pci_get_features(DeviceState *d)
470
{
471
    VirtIOPCIProxy *proxy = to_virtio_pci_proxy(d);
472
    return proxy->host_features;
473 474
}

475 476 477 478 479 480
static int kvm_virtio_pci_vq_vector_use(VirtIOPCIProxy *proxy,
                                        unsigned int queue_no,
                                        unsigned int vector,
                                        MSIMessage msg)
{
    VirtIOIRQFD *irqfd = &proxy->vector_irqfd[vector];
481
    int ret;
482 483 484 485 486 487 488 489 490 491 492 493 494 495

    if (irqfd->users == 0) {
        ret = kvm_irqchip_add_msi_route(kvm_state, msg);
        if (ret < 0) {
            return ret;
        }
        irqfd->virq = ret;
    }
    irqfd->users++;
    return 0;
}

static void kvm_virtio_pci_vq_vector_release(VirtIOPCIProxy *proxy,
                                             unsigned int vector)
496 497 498 499 500 501 502
{
    VirtIOIRQFD *irqfd = &proxy->vector_irqfd[vector];
    if (--irqfd->users == 0) {
        kvm_irqchip_release_virq(kvm_state, irqfd->virq);
    }
}

503 504 505 506 507 508 509 510 511 512 513 514 515 516 517
static int kvm_virtio_pci_irqfd_use(VirtIOPCIProxy *proxy,
                                 unsigned int queue_no,
                                 unsigned int vector)
{
    VirtIOIRQFD *irqfd = &proxy->vector_irqfd[vector];
    VirtQueue *vq = virtio_get_queue(proxy->vdev, queue_no);
    EventNotifier *n = virtio_queue_get_guest_notifier(vq);
    int ret;
    ret = kvm_irqchip_add_irqfd_notifier(kvm_state, n, irqfd->virq);
    return ret;
}

static void kvm_virtio_pci_irqfd_release(VirtIOPCIProxy *proxy,
                                      unsigned int queue_no,
                                      unsigned int vector)
518 519
{
    VirtQueue *vq = virtio_get_queue(proxy->vdev, queue_no);
520
    EventNotifier *n = virtio_queue_get_guest_notifier(vq);
521
    VirtIOIRQFD *irqfd = &proxy->vector_irqfd[vector];
522
    int ret;
523

J
Jan Kiszka 已提交
524
    ret = kvm_irqchip_remove_irqfd_notifier(kvm_state, n, irqfd->virq);
525
    assert(ret == 0);
526
}
527

528 529 530 531
static int kvm_virtio_pci_vector_use(VirtIOPCIProxy *proxy, int nvqs)
{
    PCIDevice *dev = &proxy->pci_dev;
    VirtIODevice *vdev = proxy->vdev;
532
    VirtioDeviceClass *k = VIRTIO_DEVICE_GET_CLASS(vdev);
533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548
    unsigned int vector;
    int ret, queue_no;
    MSIMessage msg;

    for (queue_no = 0; queue_no < nvqs; queue_no++) {
        if (!virtio_queue_get_num(vdev, queue_no)) {
            break;
        }
        vector = virtio_queue_vector(vdev, queue_no);
        if (vector >= msix_nr_vectors_allocated(dev)) {
            continue;
        }
        msg = msix_get_message(dev, vector);
        ret = kvm_virtio_pci_vq_vector_use(proxy, queue_no, vector, msg);
        if (ret < 0) {
            goto undo;
549
        }
550 551 552
        /* If guest supports masking, set up irqfd now.
         * Otherwise, delay until unmasked in the frontend.
         */
553
        if (k->guest_notifier_mask) {
554 555 556 557 558 559
            ret = kvm_virtio_pci_irqfd_use(proxy, queue_no, vector);
            if (ret < 0) {
                kvm_virtio_pci_vq_vector_release(proxy, vector);
                goto undo;
            }
        }
560 561
    }
    return 0;
562 563 564 565 566 567 568

undo:
    while (--queue_no >= 0) {
        vector = virtio_queue_vector(vdev, queue_no);
        if (vector >= msix_nr_vectors_allocated(dev)) {
            continue;
        }
569
        if (k->guest_notifier_mask) {
570
            kvm_virtio_pci_irqfd_release(proxy, queue_no, vector);
571
        }
572 573 574
        kvm_virtio_pci_vq_vector_release(proxy, vector);
    }
    return ret;
575 576
}

577 578 579 580 581 582
static void kvm_virtio_pci_vector_release(VirtIOPCIProxy *proxy, int nvqs)
{
    PCIDevice *dev = &proxy->pci_dev;
    VirtIODevice *vdev = proxy->vdev;
    unsigned int vector;
    int queue_no;
583
    VirtioDeviceClass *k = VIRTIO_DEVICE_GET_CLASS(vdev);
584 585 586 587 588 589 590 591 592

    for (queue_no = 0; queue_no < nvqs; queue_no++) {
        if (!virtio_queue_get_num(vdev, queue_no)) {
            break;
        }
        vector = virtio_queue_vector(vdev, queue_no);
        if (vector >= msix_nr_vectors_allocated(dev)) {
            continue;
        }
593 594 595
        /* If guest supports masking, clean up irqfd now.
         * Otherwise, it was cleaned when masked in the frontend.
         */
596
        if (k->guest_notifier_mask) {
597
            kvm_virtio_pci_irqfd_release(proxy, queue_no, vector);
598
        }
599 600 601 602
        kvm_virtio_pci_vq_vector_release(proxy, vector);
    }
}

603 604 605 606
static int virtio_pci_vq_vector_unmask(VirtIOPCIProxy *proxy,
                                       unsigned int queue_no,
                                       unsigned int vector,
                                       MSIMessage msg)
607
{
608
    VirtioDeviceClass *k = VIRTIO_DEVICE_GET_CLASS(proxy->vdev);
609 610
    VirtQueue *vq = virtio_get_queue(proxy->vdev, queue_no);
    EventNotifier *n = virtio_queue_get_guest_notifier(vq);
611
    VirtIOIRQFD *irqfd;
612
    int ret = 0;
613

614 615 616 617 618 619 620
    if (proxy->vector_irqfd) {
        irqfd = &proxy->vector_irqfd[vector];
        if (irqfd->msg.data != msg.data || irqfd->msg.address != msg.address) {
            ret = kvm_irqchip_update_msi_route(kvm_state, irqfd->virq, msg);
            if (ret < 0) {
                return ret;
            }
621 622 623
        }
    }

624 625 626
    /* If guest supports masking, irqfd is already setup, unmask it.
     * Otherwise, set it up now.
     */
627 628
    if (k->guest_notifier_mask) {
        k->guest_notifier_mask(proxy->vdev, queue_no, false);
629
        /* Test after unmasking to avoid losing events. */
630 631
        if (k->guest_notifier_pending &&
            k->guest_notifier_pending(proxy->vdev, queue_no)) {
632 633 634 635
            event_notifier_set(n);
        }
    } else {
        ret = kvm_virtio_pci_irqfd_use(proxy, queue_no, vector);
636
    }
637
    return ret;
638 639
}

640
static void virtio_pci_vq_vector_mask(VirtIOPCIProxy *proxy,
641 642 643
                                             unsigned int queue_no,
                                             unsigned int vector)
{
644 645
    VirtioDeviceClass *k = VIRTIO_DEVICE_GET_CLASS(proxy->vdev);

646 647 648
    /* If guest supports masking, keep irqfd but mask it.
     * Otherwise, clean it up now.
     */ 
649 650
    if (k->guest_notifier_mask) {
        k->guest_notifier_mask(proxy->vdev, queue_no, true);
651
    } else {
652
        kvm_virtio_pci_irqfd_release(proxy, queue_no, vector);
653
    }
654 655
}

656 657
static int virtio_pci_vector_unmask(PCIDevice *dev, unsigned vector,
                                    MSIMessage msg)
658 659 660 661 662
{
    VirtIOPCIProxy *proxy = container_of(dev, VirtIOPCIProxy, pci_dev);
    VirtIODevice *vdev = proxy->vdev;
    int ret, queue_no;

663
    for (queue_no = 0; queue_no < proxy->nvqs_with_notifiers; queue_no++) {
664 665 666 667 668 669
        if (!virtio_queue_get_num(vdev, queue_no)) {
            break;
        }
        if (virtio_queue_vector(vdev, queue_no) != vector) {
            continue;
        }
670
        ret = virtio_pci_vq_vector_unmask(proxy, queue_no, vector, msg);
671 672 673 674 675 676 677 678 679 680 681
        if (ret < 0) {
            goto undo;
        }
    }
    return 0;

undo:
    while (--queue_no >= 0) {
        if (virtio_queue_vector(vdev, queue_no) != vector) {
            continue;
        }
682
        virtio_pci_vq_vector_mask(proxy, queue_no, vector);
683 684 685 686
    }
    return ret;
}

687
static void virtio_pci_vector_mask(PCIDevice *dev, unsigned vector)
688 689 690 691 692
{
    VirtIOPCIProxy *proxy = container_of(dev, VirtIOPCIProxy, pci_dev);
    VirtIODevice *vdev = proxy->vdev;
    int queue_no;

693
    for (queue_no = 0; queue_no < proxy->nvqs_with_notifiers; queue_no++) {
694 695 696 697 698 699
        if (!virtio_queue_get_num(vdev, queue_no)) {
            break;
        }
        if (virtio_queue_vector(vdev, queue_no) != vector) {
            continue;
        }
700
        virtio_pci_vq_vector_mask(proxy, queue_no, vector);
701 702 703
    }
}

704 705 706
static void virtio_pci_vector_poll(PCIDevice *dev,
                                   unsigned int vector_start,
                                   unsigned int vector_end)
707 708 709
{
    VirtIOPCIProxy *proxy = container_of(dev, VirtIOPCIProxy, pci_dev);
    VirtIODevice *vdev = proxy->vdev;
710
    VirtioDeviceClass *k = VIRTIO_DEVICE_GET_CLASS(vdev);
711 712 713 714 715
    int queue_no;
    unsigned int vector;
    EventNotifier *notifier;
    VirtQueue *vq;

716
    for (queue_no = 0; queue_no < proxy->nvqs_with_notifiers; queue_no++) {
717 718 719 720 721 722 723 724 725 726
        if (!virtio_queue_get_num(vdev, queue_no)) {
            break;
        }
        vector = virtio_queue_vector(vdev, queue_no);
        if (vector < vector_start || vector >= vector_end ||
            !msix_is_masked(dev, vector)) {
            continue;
        }
        vq = virtio_get_queue(vdev, queue_no);
        notifier = virtio_queue_get_guest_notifier(vq);
727 728
        if (k->guest_notifier_pending) {
            if (k->guest_notifier_pending(vdev, queue_no)) {
729 730 731
                msix_set_pending(dev, vector);
            }
        } else if (event_notifier_test_and_clear(notifier)) {
732 733 734 735 736 737 738
            msix_set_pending(dev, vector);
        }
    }
}

static int virtio_pci_set_guest_notifier(DeviceState *d, int n, bool assign,
                                         bool with_irqfd)
739
{
740
    VirtIOPCIProxy *proxy = to_virtio_pci_proxy(d);
741
    VirtioDeviceClass *vdc = VIRTIO_DEVICE_GET_CLASS(proxy->vdev);
742 743 744 745 746 747 748 749
    VirtQueue *vq = virtio_get_queue(proxy->vdev, n);
    EventNotifier *notifier = virtio_queue_get_guest_notifier(vq);

    if (assign) {
        int r = event_notifier_init(notifier, 0);
        if (r < 0) {
            return r;
        }
750
        virtio_queue_set_guest_notifier_fd_handler(vq, true, with_irqfd);
751
    } else {
752
        virtio_queue_set_guest_notifier_fd_handler(vq, false, with_irqfd);
753 754 755
        event_notifier_cleanup(notifier);
    }

756 757 758 759
    if (!msix_enabled(&proxy->pci_dev) && vdc->guest_notifier_mask) {
        vdc->guest_notifier_mask(proxy->vdev, n, !assign);
    }

760 761 762
    return 0;
}

763
static bool virtio_pci_query_guest_notifiers(DeviceState *d)
764
{
765
    VirtIOPCIProxy *proxy = to_virtio_pci_proxy(d);
766 767 768
    return msix_enabled(&proxy->pci_dev);
}

769
static int virtio_pci_set_guest_notifiers(DeviceState *d, int nvqs, bool assign)
770
{
771
    VirtIOPCIProxy *proxy = to_virtio_pci_proxy(d);
772
    VirtIODevice *vdev = proxy->vdev;
773
    VirtioDeviceClass *k = VIRTIO_DEVICE_GET_CLASS(vdev);
774
    int r, n;
775 776
    bool with_irqfd = msix_enabled(&proxy->pci_dev) &&
        kvm_msi_via_irqfd_enabled();
777

778 779 780 781 782 783 784 785 786
    nvqs = MIN(nvqs, VIRTIO_PCI_QUEUE_MAX);

    /* When deassigning, pass a consistent nvqs value
     * to avoid leaking notifiers.
     */
    assert(assign || nvqs == proxy->nvqs_with_notifiers);

    proxy->nvqs_with_notifiers = nvqs;

787
    /* Must unset vector notifier while guest notifier is still assigned */
788
    if ((proxy->vector_irqfd || k->guest_notifier_mask) && !assign) {
789
        msix_unset_vector_notifiers(&proxy->pci_dev);
790 791 792 793 794
        if (proxy->vector_irqfd) {
            kvm_virtio_pci_vector_release(proxy, nvqs);
            g_free(proxy->vector_irqfd);
            proxy->vector_irqfd = NULL;
        }
795 796
    }

797
    for (n = 0; n < nvqs; n++) {
798 799 800 801
        if (!virtio_queue_get_num(vdev, n)) {
            break;
        }

802
        r = virtio_pci_set_guest_notifier(d, n, assign, with_irqfd);
803 804 805 806 807
        if (r < 0) {
            goto assign_error;
        }
    }

808
    /* Must set vector notifier after guest notifier has been assigned */
809
    if ((with_irqfd || k->guest_notifier_mask) && assign) {
810 811 812 813 814 815 816 817
        if (with_irqfd) {
            proxy->vector_irqfd =
                g_malloc0(sizeof(*proxy->vector_irqfd) *
                          msix_nr_vectors_allocated(&proxy->pci_dev));
            r = kvm_virtio_pci_vector_use(proxy, nvqs);
            if (r < 0) {
                goto assign_error;
            }
818
        }
819
        r = msix_set_vector_notifiers(&proxy->pci_dev,
820 821 822
                                      virtio_pci_vector_unmask,
                                      virtio_pci_vector_mask,
                                      virtio_pci_vector_poll);
823
        if (r < 0) {
824
            goto notifiers_error;
825 826 827
        }
    }

828 829
    return 0;

830
notifiers_error:
831 832 833 834
    if (with_irqfd) {
        assert(assign);
        kvm_virtio_pci_vector_release(proxy, nvqs);
    }
835

836 837
assign_error:
    /* We get here on assignment failure. Recover by undoing for VQs 0 .. n. */
838
    assert(assign);
839
    while (--n >= 0) {
840
        virtio_pci_set_guest_notifier(d, n, !assign, with_irqfd);
841 842 843 844
    }
    return r;
}

845
static int virtio_pci_set_host_notifier(DeviceState *d, int n, bool assign)
846
{
847
    VirtIOPCIProxy *proxy = to_virtio_pci_proxy(d);
848 849 850 851 852

    /* Stop using ioeventfd for virtqueue kick if the device starts using host
     * notifiers.  This makes it easy to avoid stepping on each others' toes.
     */
    proxy->ioeventfd_disabled = assign;
853
    if (assign) {
854 855 856 857 858 859
        virtio_pci_stop_ioeventfd(proxy);
    }
    /* We don't need to start here: it's not needed because backend
     * currently only stops on status change away from ok,
     * reset, vmstop and such. If we do add code to start here,
     * need to check vmstate, device state etc. */
P
Paolo Bonzini 已提交
860
    return virtio_pci_set_host_notifier_internal(proxy, n, assign, false);
861 862
}

863
static void virtio_pci_vmstate_change(DeviceState *d, bool running)
864
{
865
    VirtIOPCIProxy *proxy = to_virtio_pci_proxy(d);
866 867

    if (running) {
868 869 870 871 872 873
        /* Try to find out if the guest has bus master disabled, but is
           in ready state. Then we have a buggy guest OS. */
        if ((proxy->vdev->status & VIRTIO_CONFIG_S_DRIVER_OK) &&
            !(proxy->pci_dev.config[PCI_COMMAND] & PCI_COMMAND_MASTER)) {
            proxy->flags |= VIRTIO_PCI_FLAG_BUS_MASTER_BUG;
        }
874
        virtio_pci_start_ioeventfd(proxy);
875
    } else {
876
        virtio_pci_stop_ioeventfd(proxy);
877 878 879
    }
}

880
#ifdef CONFIG_VIRTFS
881
static int virtio_9p_init_pci(VirtIOPCIProxy *vpci_dev)
882
{
883 884
    V9fsPCIState *dev = VIRTIO_9P_PCI(vpci_dev);
    DeviceState *vdev = DEVICE(&dev->vdev);
885

886 887 888 889
    qdev_set_parent_bus(vdev, BUS(&vpci_dev->bus));
    if (qdev_init(vdev) < 0) {
        return -1;
    }
890 891 892
    return 0;
}

893 894 895
static Property virtio_9p_pci_properties[] = {
    DEFINE_PROP_BIT("ioeventfd", VirtIOPCIProxy, flags,
                    VIRTIO_PCI_FLAG_USE_IOEVENTFD_BIT, true),
896 897
    DEFINE_PROP_UINT32("vectors", VirtIOPCIProxy, nvectors, 2),
    DEFINE_VIRTIO_COMMON_FEATURES(VirtIOPCIProxy, host_features),
898
    DEFINE_VIRTIO_9P_PROPERTIES(V9fsPCIState, vdev.fsconf),
899 900 901
    DEFINE_PROP_END_OF_LIST(),
};

902
static void virtio_9p_pci_class_init(ObjectClass *klass, void *data)
903 904
{
    DeviceClass *dc = DEVICE_CLASS(klass);
905 906
    PCIDeviceClass *pcidev_k = PCI_DEVICE_CLASS(klass);
    VirtioPCIClass *k = VIRTIO_PCI_CLASS(klass);
907 908

    k->init = virtio_9p_init_pci;
909 910 911 912
    pcidev_k->vendor_id = PCI_VENDOR_ID_REDHAT_QUMRANET;
    pcidev_k->device_id = PCI_DEVICE_ID_VIRTIO_9P;
    pcidev_k->revision = VIRTIO_PCI_ABI_VERSION;
    pcidev_k->class_id = 0x2;
913
    set_bit(DEVICE_CATEGORY_STORAGE, dc->categories);
914
    dc->props = virtio_9p_pci_properties;
915 916
}

917 918 919 920 921 922 923 924 925 926 927 928 929
static void virtio_9p_pci_instance_init(Object *obj)
{
    V9fsPCIState *dev = VIRTIO_9P_PCI(obj);
    object_initialize(OBJECT(&dev->vdev), TYPE_VIRTIO_9P);
    object_property_add_child(obj, "virtio-backend", OBJECT(&dev->vdev), NULL);
}

static const TypeInfo virtio_9p_pci_info = {
    .name          = TYPE_VIRTIO_9P_PCI,
    .parent        = TYPE_VIRTIO_PCI,
    .instance_size = sizeof(V9fsPCIState),
    .instance_init = virtio_9p_pci_instance_init,
    .class_init    = virtio_9p_pci_class_init,
930
};
931
#endif /* CONFIG_VIRTFS */
932

933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968
/*
 * virtio-pci: This is the PCIDevice which has a virtio-pci-bus.
 */

/* This is called by virtio-bus just after the device is plugged. */
static void virtio_pci_device_plugged(DeviceState *d)
{
    VirtIOPCIProxy *proxy = VIRTIO_PCI(d);
    VirtioBusState *bus = &proxy->bus;
    uint8_t *config;
    uint32_t size;

    proxy->vdev = bus->vdev;

    config = proxy->pci_dev.config;
    if (proxy->class_code) {
        pci_config_set_class(config, proxy->class_code);
    }
    pci_set_word(config + PCI_SUBSYSTEM_VENDOR_ID,
                 pci_get_word(config + PCI_VENDOR_ID));
    pci_set_word(config + PCI_SUBSYSTEM_ID, virtio_bus_get_vdev_id(bus));
    config[PCI_INTERRUPT_PIN] = 1;

    if (proxy->nvectors &&
        msix_init_exclusive_bar(&proxy->pci_dev, proxy->nvectors, 1)) {
        proxy->nvectors = 0;
    }

    proxy->pci_dev.config_write = virtio_write_config;

    size = VIRTIO_PCI_REGION_SIZE(&proxy->pci_dev)
         + virtio_bus_get_vdev_config_len(bus);
    if (size & (size - 1)) {
        size = 1 << qemu_fls(size);
    }

969 970
    memory_region_init_io(&proxy->bar, OBJECT(proxy), &virtio_pci_config_ops,
                          proxy, "virtio-pci", size);
971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997
    pci_register_bar(&proxy->pci_dev, 0, PCI_BASE_ADDRESS_SPACE_IO,
                     &proxy->bar);

    if (!kvm_has_many_ioeventfds()) {
        proxy->flags &= ~VIRTIO_PCI_FLAG_USE_IOEVENTFD;
    }

    proxy->host_features |= 0x1 << VIRTIO_F_NOTIFY_ON_EMPTY;
    proxy->host_features |= 0x1 << VIRTIO_F_BAD_FEATURE;
    proxy->host_features = virtio_bus_get_vdev_features(bus,
                                                      proxy->host_features);
}

static int virtio_pci_init(PCIDevice *pci_dev)
{
    VirtIOPCIProxy *dev = VIRTIO_PCI(pci_dev);
    VirtioPCIClass *k = VIRTIO_PCI_GET_CLASS(pci_dev);
    virtio_pci_bus_new(&dev->bus, dev);
    if (k->init != NULL) {
        return k->init(dev);
    }
    return 0;
}

static void virtio_pci_exit(PCIDevice *pci_dev)
{
    VirtIOPCIProxy *proxy = VIRTIO_PCI(pci_dev);
K
KONRAD Frederic 已提交
998
    virtio_pci_stop_ioeventfd(proxy);
K
KONRAD Frederic 已提交
999 1000
    memory_region_destroy(&proxy->bar);
    msix_uninit_exclusive_bar(pci_dev);
1001 1002
}

1003
static void virtio_pci_reset(DeviceState *qdev)
1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022
{
    VirtIOPCIProxy *proxy = VIRTIO_PCI(qdev);
    VirtioBusState *bus = VIRTIO_BUS(&proxy->bus);
    virtio_pci_stop_ioeventfd(proxy);
    virtio_bus_reset(bus);
    msix_unuse_all_vectors(&proxy->pci_dev);
    proxy->flags &= ~VIRTIO_PCI_FLAG_BUS_MASTER_BUG;
}

static void virtio_pci_class_init(ObjectClass *klass, void *data)
{
    DeviceClass *dc = DEVICE_CLASS(klass);
    PCIDeviceClass *k = PCI_DEVICE_CLASS(klass);

    k->init = virtio_pci_init;
    k->exit = virtio_pci_exit;
    k->vendor_id = PCI_VENDOR_ID_REDHAT_QUMRANET;
    k->revision = VIRTIO_PCI_ABI_VERSION;
    k->class_id = PCI_CLASS_OTHERS;
1023
    dc->reset = virtio_pci_reset;
1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034
}

static const TypeInfo virtio_pci_info = {
    .name          = TYPE_VIRTIO_PCI,
    .parent        = TYPE_PCI_DEVICE,
    .instance_size = sizeof(VirtIOPCIProxy),
    .class_init    = virtio_pci_class_init,
    .class_size    = sizeof(VirtioPCIClass),
    .abstract      = true,
};

1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067
/* virtio-blk-pci */

static Property virtio_blk_pci_properties[] = {
    DEFINE_PROP_HEX32("class", VirtIOPCIProxy, class_code, 0),
    DEFINE_PROP_BIT("ioeventfd", VirtIOPCIProxy, flags,
                    VIRTIO_PCI_FLAG_USE_IOEVENTFD_BIT, true),
    DEFINE_PROP_UINT32("vectors", VirtIOPCIProxy, nvectors, 2),
#ifdef CONFIG_VIRTIO_BLK_DATA_PLANE
    DEFINE_PROP_BIT("x-data-plane", VirtIOBlkPCI, blk.data_plane, 0, false),
#endif
    DEFINE_VIRTIO_BLK_FEATURES(VirtIOPCIProxy, host_features),
    DEFINE_VIRTIO_BLK_PROPERTIES(VirtIOBlkPCI, blk),
    DEFINE_PROP_END_OF_LIST(),
};

static int virtio_blk_pci_init(VirtIOPCIProxy *vpci_dev)
{
    VirtIOBlkPCI *dev = VIRTIO_BLK_PCI(vpci_dev);
    DeviceState *vdev = DEVICE(&dev->vdev);
    virtio_blk_set_conf(vdev, &(dev->blk));
    qdev_set_parent_bus(vdev, BUS(&vpci_dev->bus));
    if (qdev_init(vdev) < 0) {
        return -1;
    }
    return 0;
}

static void virtio_blk_pci_class_init(ObjectClass *klass, void *data)
{
    DeviceClass *dc = DEVICE_CLASS(klass);
    VirtioPCIClass *k = VIRTIO_PCI_CLASS(klass);
    PCIDeviceClass *pcidev_k = PCI_DEVICE_CLASS(klass);

1068
    set_bit(DEVICE_CATEGORY_STORAGE, dc->categories);
1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091
    dc->props = virtio_blk_pci_properties;
    k->init = virtio_blk_pci_init;
    pcidev_k->vendor_id = PCI_VENDOR_ID_REDHAT_QUMRANET;
    pcidev_k->device_id = PCI_DEVICE_ID_VIRTIO_BLOCK;
    pcidev_k->revision = VIRTIO_PCI_ABI_VERSION;
    pcidev_k->class_id = PCI_CLASS_STORAGE_SCSI;
}

static void virtio_blk_pci_instance_init(Object *obj)
{
    VirtIOBlkPCI *dev = VIRTIO_BLK_PCI(obj);
    object_initialize(OBJECT(&dev->vdev), TYPE_VIRTIO_BLK);
    object_property_add_child(obj, "virtio-backend", OBJECT(&dev->vdev), NULL);
}

static const TypeInfo virtio_blk_pci_info = {
    .name          = TYPE_VIRTIO_BLK_PCI,
    .parent        = TYPE_VIRTIO_PCI,
    .instance_size = sizeof(VirtIOBlkPCI),
    .instance_init = virtio_blk_pci_instance_init,
    .class_init    = virtio_blk_pci_class_init,
};

1092 1093 1094 1095 1096 1097 1098 1099
/* virtio-scsi-pci */

static Property virtio_scsi_pci_properties[] = {
    DEFINE_PROP_BIT("ioeventfd", VirtIOPCIProxy, flags,
                    VIRTIO_PCI_FLAG_USE_IOEVENTFD_BIT, true),
    DEFINE_PROP_UINT32("vectors", VirtIOPCIProxy, nvectors,
                       DEV_NVECTORS_UNSPECIFIED),
    DEFINE_VIRTIO_SCSI_FEATURES(VirtIOPCIProxy, host_features),
1100
    DEFINE_VIRTIO_SCSI_PROPERTIES(VirtIOSCSIPCI, vdev.parent_obj.conf),
1101 1102 1103 1104 1105 1106 1107
    DEFINE_PROP_END_OF_LIST(),
};

static int virtio_scsi_pci_init_pci(VirtIOPCIProxy *vpci_dev)
{
    VirtIOSCSIPCI *dev = VIRTIO_SCSI_PCI(vpci_dev);
    DeviceState *vdev = DEVICE(&dev->vdev);
1108
    VirtIOSCSICommon *vs = VIRTIO_SCSI_COMMON(vdev);
1109 1110
    DeviceState *proxy = DEVICE(vpci_dev);
    char *bus_name;
1111 1112

    if (vpci_dev->nvectors == DEV_NVECTORS_UNSPECIFIED) {
1113
        vpci_dev->nvectors = vs->conf.num_queues + 3;
1114 1115
    }

1116 1117 1118 1119 1120 1121 1122 1123 1124 1125
    /*
     * For command line compatibility, this sets the virtio-scsi-device bus
     * name as before.
     */
    if (proxy->id) {
        bus_name = g_strdup_printf("%s.0", proxy->id);
        virtio_device_set_child_bus_name(VIRTIO_DEVICE(vdev), bus_name);
        g_free(bus_name);
    }

1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138
    qdev_set_parent_bus(vdev, BUS(&vpci_dev->bus));
    if (qdev_init(vdev) < 0) {
        return -1;
    }
    return 0;
}

static void virtio_scsi_pci_class_init(ObjectClass *klass, void *data)
{
    DeviceClass *dc = DEVICE_CLASS(klass);
    VirtioPCIClass *k = VIRTIO_PCI_CLASS(klass);
    PCIDeviceClass *pcidev_k = PCI_DEVICE_CLASS(klass);
    k->init = virtio_scsi_pci_init_pci;
1139
    set_bit(DEVICE_CATEGORY_STORAGE, dc->categories);
1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161
    dc->props = virtio_scsi_pci_properties;
    pcidev_k->vendor_id = PCI_VENDOR_ID_REDHAT_QUMRANET;
    pcidev_k->device_id = PCI_DEVICE_ID_VIRTIO_SCSI;
    pcidev_k->revision = 0x00;
    pcidev_k->class_id = PCI_CLASS_STORAGE_SCSI;
}

static void virtio_scsi_pci_instance_init(Object *obj)
{
    VirtIOSCSIPCI *dev = VIRTIO_SCSI_PCI(obj);
    object_initialize(OBJECT(&dev->vdev), TYPE_VIRTIO_SCSI);
    object_property_add_child(obj, "virtio-backend", OBJECT(&dev->vdev), NULL);
}

static const TypeInfo virtio_scsi_pci_info = {
    .name          = TYPE_VIRTIO_SCSI_PCI,
    .parent        = TYPE_VIRTIO_PCI,
    .instance_size = sizeof(VirtIOSCSIPCI),
    .instance_init = virtio_scsi_pci_instance_init,
    .class_init    = virtio_scsi_pci_class_init,
};

1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195
/* vhost-scsi-pci */

#ifdef CONFIG_VHOST_SCSI
static Property vhost_scsi_pci_properties[] = {
    DEFINE_PROP_UINT32("vectors", VirtIOPCIProxy, nvectors,
                       DEV_NVECTORS_UNSPECIFIED),
    DEFINE_VIRTIO_COMMON_FEATURES(VirtIOPCIProxy, host_features),
    DEFINE_VHOST_SCSI_PROPERTIES(VHostSCSIPCI, vdev.parent_obj.conf),
    DEFINE_PROP_END_OF_LIST(),
};

static int vhost_scsi_pci_init_pci(VirtIOPCIProxy *vpci_dev)
{
    VHostSCSIPCI *dev = VHOST_SCSI_PCI(vpci_dev);
    DeviceState *vdev = DEVICE(&dev->vdev);
    VirtIOSCSICommon *vs = VIRTIO_SCSI_COMMON(vdev);

    if (vpci_dev->nvectors == DEV_NVECTORS_UNSPECIFIED) {
        vpci_dev->nvectors = vs->conf.num_queues + 3;
    }

    qdev_set_parent_bus(vdev, BUS(&vpci_dev->bus));
    if (qdev_init(vdev) < 0) {
        return -1;
    }
    return 0;
}

static void vhost_scsi_pci_class_init(ObjectClass *klass, void *data)
{
    DeviceClass *dc = DEVICE_CLASS(klass);
    VirtioPCIClass *k = VIRTIO_PCI_CLASS(klass);
    PCIDeviceClass *pcidev_k = PCI_DEVICE_CLASS(klass);
    k->init = vhost_scsi_pci_init_pci;
1196
    set_bit(DEVICE_CATEGORY_STORAGE, dc->categories);
1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219
    dc->props = vhost_scsi_pci_properties;
    pcidev_k->vendor_id = PCI_VENDOR_ID_REDHAT_QUMRANET;
    pcidev_k->device_id = PCI_DEVICE_ID_VIRTIO_SCSI;
    pcidev_k->revision = 0x00;
    pcidev_k->class_id = PCI_CLASS_STORAGE_SCSI;
}

static void vhost_scsi_pci_instance_init(Object *obj)
{
    VHostSCSIPCI *dev = VHOST_SCSI_PCI(obj);
    object_initialize(OBJECT(&dev->vdev), TYPE_VHOST_SCSI);
    object_property_add_child(obj, "virtio-backend", OBJECT(&dev->vdev), NULL);
}

static const TypeInfo vhost_scsi_pci_info = {
    .name          = TYPE_VHOST_SCSI_PCI,
    .parent        = TYPE_VIRTIO_PCI,
    .instance_size = sizeof(VHostSCSIPCI),
    .instance_init = vhost_scsi_pci_instance_init,
    .class_init    = vhost_scsi_pci_class_init,
};
#endif

1220 1221
/* virtio-balloon-pci */

1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247
static void balloon_pci_stats_get_all(Object *obj, struct Visitor *v,
                                      void *opaque, const char *name,
                                      Error **errp)
{
    VirtIOBalloonPCI *dev = opaque;
    object_property_get(OBJECT(&dev->vdev), v, "guest-stats", errp);
}

static void balloon_pci_stats_get_poll_interval(Object *obj, struct Visitor *v,
                                                void *opaque, const char *name,
                                                Error **errp)
{
    VirtIOBalloonPCI *dev = opaque;
    object_property_get(OBJECT(&dev->vdev), v, "guest-stats-polling-interval",
                        errp);
}

static void balloon_pci_stats_set_poll_interval(Object *obj, struct Visitor *v,
                                                void *opaque, const char *name,
                                                Error **errp)
{
    VirtIOBalloonPCI *dev = opaque;
    object_property_set(OBJECT(&dev->vdev), v, "guest-stats-polling-interval",
                        errp);
}

1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276
static Property virtio_balloon_pci_properties[] = {
    DEFINE_VIRTIO_COMMON_FEATURES(VirtIOPCIProxy, host_features),
    DEFINE_PROP_HEX32("class", VirtIOPCIProxy, class_code, 0),
    DEFINE_PROP_END_OF_LIST(),
};

static int virtio_balloon_pci_init(VirtIOPCIProxy *vpci_dev)
{
    VirtIOBalloonPCI *dev = VIRTIO_BALLOON_PCI(vpci_dev);
    DeviceState *vdev = DEVICE(&dev->vdev);

    if (vpci_dev->class_code != PCI_CLASS_OTHERS &&
        vpci_dev->class_code != PCI_CLASS_MEMORY_RAM) { /* qemu < 1.1 */
        vpci_dev->class_code = PCI_CLASS_OTHERS;
    }

    qdev_set_parent_bus(vdev, BUS(&vpci_dev->bus));
    if (qdev_init(vdev) < 0) {
        return -1;
    }
    return 0;
}

static void virtio_balloon_pci_class_init(ObjectClass *klass, void *data)
{
    DeviceClass *dc = DEVICE_CLASS(klass);
    VirtioPCIClass *k = VIRTIO_PCI_CLASS(klass);
    PCIDeviceClass *pcidev_k = PCI_DEVICE_CLASS(klass);
    k->init = virtio_balloon_pci_init;
1277
    set_bit(DEVICE_CATEGORY_MISC, dc->categories);
1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289
    dc->props = virtio_balloon_pci_properties;
    pcidev_k->vendor_id = PCI_VENDOR_ID_REDHAT_QUMRANET;
    pcidev_k->device_id = PCI_DEVICE_ID_VIRTIO_BALLOON;
    pcidev_k->revision = VIRTIO_PCI_ABI_VERSION;
    pcidev_k->class_id = PCI_CLASS_OTHERS;
}

static void virtio_balloon_pci_instance_init(Object *obj)
{
    VirtIOBalloonPCI *dev = VIRTIO_BALLOON_PCI(obj);
    object_initialize(OBJECT(&dev->vdev), TYPE_VIRTIO_BALLOON);
    object_property_add_child(obj, "virtio-backend", OBJECT(&dev->vdev), NULL);
1290 1291 1292 1293 1294 1295 1296 1297 1298

    object_property_add(obj, "guest-stats", "guest statistics",
                        balloon_pci_stats_get_all, NULL, NULL, dev,
                        NULL);

    object_property_add(obj, "guest-stats-polling-interval", "int",
                        balloon_pci_stats_get_poll_interval,
                        balloon_pci_stats_set_poll_interval,
                        NULL, dev, NULL);
1299 1300 1301 1302 1303 1304 1305 1306 1307 1308
}

static const TypeInfo virtio_balloon_pci_info = {
    .name          = TYPE_VIRTIO_BALLOON_PCI,
    .parent        = TYPE_VIRTIO_PCI,
    .instance_size = sizeof(VirtIOBalloonPCI),
    .instance_init = virtio_balloon_pci_instance_init,
    .class_init    = virtio_balloon_pci_class_init,
};

1309 1310 1311 1312 1313 1314
/* virtio-serial-pci */

static int virtio_serial_pci_init(VirtIOPCIProxy *vpci_dev)
{
    VirtIOSerialPCI *dev = VIRTIO_SERIAL_PCI(vpci_dev);
    DeviceState *vdev = DEVICE(&dev->vdev);
1315 1316
    DeviceState *proxy = DEVICE(vpci_dev);
    char *bus_name;
1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329

    if (vpci_dev->class_code != PCI_CLASS_COMMUNICATION_OTHER &&
        vpci_dev->class_code != PCI_CLASS_DISPLAY_OTHER && /* qemu 0.10 */
        vpci_dev->class_code != PCI_CLASS_OTHERS) {        /* qemu-kvm  */
            vpci_dev->class_code = PCI_CLASS_COMMUNICATION_OTHER;
    }

    /* backwards-compatibility with machines that were created with
       DEV_NVECTORS_UNSPECIFIED */
    if (vpci_dev->nvectors == DEV_NVECTORS_UNSPECIFIED) {
        vpci_dev->nvectors = dev->vdev.serial.max_virtserial_ports + 1;
    }

1330 1331 1332 1333 1334 1335 1336 1337 1338 1339
    /*
     * For command line compatibility, this sets the virtio-serial-device bus
     * name as before.
     */
    if (proxy->id) {
        bus_name = g_strdup_printf("%s.0", proxy->id);
        virtio_device_set_child_bus_name(VIRTIO_DEVICE(vdev), bus_name);
        g_free(bus_name);
    }

1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362
    qdev_set_parent_bus(vdev, BUS(&vpci_dev->bus));
    if (qdev_init(vdev) < 0) {
        return -1;
    }
    return 0;
}

static Property virtio_serial_pci_properties[] = {
    DEFINE_PROP_BIT("ioeventfd", VirtIOPCIProxy, flags,
                    VIRTIO_PCI_FLAG_USE_IOEVENTFD_BIT, true),
    DEFINE_PROP_UINT32("vectors", VirtIOPCIProxy, nvectors, 2),
    DEFINE_PROP_HEX32("class", VirtIOPCIProxy, class_code, 0),
    DEFINE_VIRTIO_COMMON_FEATURES(VirtIOPCIProxy, host_features),
    DEFINE_VIRTIO_SERIAL_PROPERTIES(VirtIOSerialPCI, vdev.serial),
    DEFINE_PROP_END_OF_LIST(),
};

static void virtio_serial_pci_class_init(ObjectClass *klass, void *data)
{
    DeviceClass *dc = DEVICE_CLASS(klass);
    VirtioPCIClass *k = VIRTIO_PCI_CLASS(klass);
    PCIDeviceClass *pcidev_k = PCI_DEVICE_CLASS(klass);
    k->init = virtio_serial_pci_init;
1363
    set_bit(DEVICE_CATEGORY_INPUT, dc->categories);
1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385
    dc->props = virtio_serial_pci_properties;
    pcidev_k->vendor_id = PCI_VENDOR_ID_REDHAT_QUMRANET;
    pcidev_k->device_id = PCI_DEVICE_ID_VIRTIO_CONSOLE;
    pcidev_k->revision = VIRTIO_PCI_ABI_VERSION;
    pcidev_k->class_id = PCI_CLASS_COMMUNICATION_OTHER;
}

static void virtio_serial_pci_instance_init(Object *obj)
{
    VirtIOSerialPCI *dev = VIRTIO_SERIAL_PCI(obj);
    object_initialize(OBJECT(&dev->vdev), TYPE_VIRTIO_SERIAL);
    object_property_add_child(obj, "virtio-backend", OBJECT(&dev->vdev), NULL);
}

static const TypeInfo virtio_serial_pci_info = {
    .name          = TYPE_VIRTIO_SERIAL_PCI,
    .parent        = TYPE_VIRTIO_PCI,
    .instance_size = sizeof(VirtIOSerialPCI),
    .instance_init = virtio_serial_pci_instance_init,
    .class_init    = virtio_serial_pci_class_init,
};

1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399
/* virtio-net-pci */

static Property virtio_net_properties[] = {
    DEFINE_PROP_BIT("ioeventfd", VirtIOPCIProxy, flags,
                    VIRTIO_PCI_FLAG_USE_IOEVENTFD_BIT, false),
    DEFINE_PROP_UINT32("vectors", VirtIOPCIProxy, nvectors, 3),
    DEFINE_VIRTIO_NET_FEATURES(VirtIOPCIProxy, host_features),
    DEFINE_NIC_PROPERTIES(VirtIONetPCI, vdev.nic_conf),
    DEFINE_VIRTIO_NET_PROPERTIES(VirtIONetPCI, vdev.net_conf),
    DEFINE_PROP_END_OF_LIST(),
};

static int virtio_net_pci_init(VirtIOPCIProxy *vpci_dev)
{
1400
    DeviceState *qdev = DEVICE(vpci_dev);
1401 1402 1403 1404
    VirtIONetPCI *dev = VIRTIO_NET_PCI(vpci_dev);
    DeviceState *vdev = DEVICE(&dev->vdev);

    virtio_net_set_config_size(&dev->vdev, vpci_dev->host_features);
1405 1406
    virtio_net_set_netclient_name(&dev->vdev, qdev->id,
                                  object_get_typename(OBJECT(qdev)));
1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424
    qdev_set_parent_bus(vdev, BUS(&vpci_dev->bus));
    if (qdev_init(vdev) < 0) {
        return -1;
    }
    return 0;
}

static void virtio_net_pci_class_init(ObjectClass *klass, void *data)
{
    DeviceClass *dc = DEVICE_CLASS(klass);
    PCIDeviceClass *k = PCI_DEVICE_CLASS(klass);
    VirtioPCIClass *vpciklass = VIRTIO_PCI_CLASS(klass);

    k->romfile = "efi-virtio.rom";
    k->vendor_id = PCI_VENDOR_ID_REDHAT_QUMRANET;
    k->device_id = PCI_DEVICE_ID_VIRTIO_NET;
    k->revision = VIRTIO_PCI_ABI_VERSION;
    k->class_id = PCI_CLASS_NETWORK_ETHERNET;
1425
    set_bit(DEVICE_CATEGORY_NETWORK, dc->categories);
1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444
    dc->props = virtio_net_properties;
    vpciklass->init = virtio_net_pci_init;
}

static void virtio_net_pci_instance_init(Object *obj)
{
    VirtIONetPCI *dev = VIRTIO_NET_PCI(obj);
    object_initialize(OBJECT(&dev->vdev), TYPE_VIRTIO_NET);
    object_property_add_child(obj, "virtio-backend", OBJECT(&dev->vdev), NULL);
}

static const TypeInfo virtio_net_pci_info = {
    .name          = TYPE_VIRTIO_NET_PCI,
    .parent        = TYPE_VIRTIO_PCI,
    .instance_size = sizeof(VirtIONetPCI),
    .instance_init = virtio_net_pci_instance_init,
    .class_init    = virtio_net_pci_class_init,
};

1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463
/* virtio-rng-pci */

static Property virtio_rng_pci_properties[] = {
    DEFINE_VIRTIO_COMMON_FEATURES(VirtIOPCIProxy, host_features),
    DEFINE_VIRTIO_RNG_PROPERTIES(VirtIORngPCI, vdev.conf),
    DEFINE_PROP_END_OF_LIST(),
};

static int virtio_rng_pci_init(VirtIOPCIProxy *vpci_dev)
{
    VirtIORngPCI *vrng = VIRTIO_RNG_PCI(vpci_dev);
    DeviceState *vdev = DEVICE(&vrng->vdev);

    qdev_set_parent_bus(vdev, BUS(&vpci_dev->bus));
    if (qdev_init(vdev) < 0) {
        return -1;
    }

    object_property_set_link(OBJECT(vrng),
1464
                             OBJECT(vrng->vdev.conf.rng), "rng",
1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476
                             NULL);

    return 0;
}

static void virtio_rng_pci_class_init(ObjectClass *klass, void *data)
{
    DeviceClass *dc = DEVICE_CLASS(klass);
    VirtioPCIClass *k = VIRTIO_PCI_CLASS(klass);
    PCIDeviceClass *pcidev_k = PCI_DEVICE_CLASS(klass);

    k->init = virtio_rng_pci_init;
1477
    set_bit(DEVICE_CATEGORY_MISC, dc->categories);
1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503
    dc->props = virtio_rng_pci_properties;

    pcidev_k->vendor_id = PCI_VENDOR_ID_REDHAT_QUMRANET;
    pcidev_k->device_id = PCI_DEVICE_ID_VIRTIO_RNG;
    pcidev_k->revision = VIRTIO_PCI_ABI_VERSION;
    pcidev_k->class_id = PCI_CLASS_OTHERS;
}

static void virtio_rng_initfn(Object *obj)
{
    VirtIORngPCI *dev = VIRTIO_RNG_PCI(obj);
    object_initialize(OBJECT(&dev->vdev), TYPE_VIRTIO_RNG);
    object_property_add_child(obj, "virtio-backend", OBJECT(&dev->vdev), NULL);
    object_property_add_link(obj, "rng", TYPE_RNG_BACKEND,
                             (Object **)&dev->vdev.conf.rng, NULL);

}

static const TypeInfo virtio_rng_pci_info = {
    .name          = TYPE_VIRTIO_RNG_PCI,
    .parent        = TYPE_VIRTIO_PCI,
    .instance_size = sizeof(VirtIORngPCI),
    .instance_init = virtio_rng_initfn,
    .class_init    = virtio_rng_pci_class_init,
};

1504 1505
/* virtio-pci-bus */

1506
static void virtio_pci_bus_new(VirtioBusState *bus, VirtIOPCIProxy *dev)
1507 1508 1509
{
    DeviceState *qdev = DEVICE(dev);
    BusState *qbus;
1510 1511 1512 1513
    char virtio_bus_name[] = "virtio-bus";

    qbus_create_inplace((BusState *)bus, TYPE_VIRTIO_PCI_BUS, qdev,
                        virtio_bus_name);
1514
    qbus = BUS(bus);
1515
    qbus->allow_hotplug = 1;
1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532
}

static void virtio_pci_bus_class_init(ObjectClass *klass, void *data)
{
    BusClass *bus_class = BUS_CLASS(klass);
    VirtioBusClass *k = VIRTIO_BUS_CLASS(klass);
    bus_class->max_dev = 1;
    k->notify = virtio_pci_notify;
    k->save_config = virtio_pci_save_config;
    k->load_config = virtio_pci_load_config;
    k->save_queue = virtio_pci_save_queue;
    k->load_queue = virtio_pci_load_queue;
    k->get_features = virtio_pci_get_features;
    k->query_guest_notifiers = virtio_pci_query_guest_notifiers;
    k->set_host_notifier = virtio_pci_set_host_notifier;
    k->set_guest_notifiers = virtio_pci_set_guest_notifiers;
    k->vmstate_change = virtio_pci_vmstate_change;
1533
    k->device_plugged = virtio_pci_device_plugged;
1534 1535 1536 1537 1538 1539 1540 1541 1542
}

static const TypeInfo virtio_pci_bus_info = {
    .name          = TYPE_VIRTIO_PCI_BUS,
    .parent        = TYPE_VIRTIO_BUS,
    .instance_size = sizeof(VirtioPCIBusState),
    .class_init    = virtio_pci_bus_class_init,
};

A
Andreas Färber 已提交
1543
static void virtio_pci_register_types(void)
P
Paul Brook 已提交
1544
{
1545
    type_register_static(&virtio_rng_pci_info);
1546
    type_register_static(&virtio_pci_bus_info);
1547
    type_register_static(&virtio_pci_info);
1548
#ifdef CONFIG_VIRTFS
1549
    type_register_static(&virtio_9p_pci_info);
1550
#endif
1551
    type_register_static(&virtio_blk_pci_info);
1552
    type_register_static(&virtio_scsi_pci_info);
1553
    type_register_static(&virtio_balloon_pci_info);
1554
    type_register_static(&virtio_serial_pci_info);
1555
    type_register_static(&virtio_net_pci_info);
1556 1557 1558
#ifdef CONFIG_VHOST_SCSI
    type_register_static(&vhost_scsi_pci_info);
#endif
P
Paul Brook 已提交
1559 1560
}

A
Andreas Färber 已提交
1561
type_init(virtio_pci_register_types)