pci.c 78.6 KB
Newer Older
B
bellard 已提交
1 2 3 4
/*
 * QEMU PCI bus manager
 *
 * Copyright (c) 2004 Fabrice Bellard
5
 *
B
bellard 已提交
6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
 * Permission is hereby granted, free of charge, to any person obtaining a copy
 * of this software and associated documentation files (the "Software"), to deal
 * in the Software without restriction, including without limitation the rights
 * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
 * copies of the Software, and to permit persons to whom the Software is
 * furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in
 * all copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
 * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
 * THE SOFTWARE.
 */
P
Peter Maydell 已提交
24
#include "qemu/osdep.h"
25 26 27
#include "hw/hw.h"
#include "hw/pci/pci.h"
#include "hw/pci/pci_bridge.h"
28
#include "hw/pci/pci_bus.h"
29
#include "hw/pci/pci_host.h"
30
#include "monitor/monitor.h"
P
Paolo Bonzini 已提交
31
#include "net/net.h"
32
#include "sysemu/sysemu.h"
33
#include "hw/loader.h"
34
#include "qemu/error-report.h"
35
#include "qemu/range.h"
L
Luiz Capitulino 已提交
36
#include "qmp-commands.h"
D
Don Koch 已提交
37
#include "trace.h"
38 39
#include "hw/pci/msi.h"
#include "hw/pci/msix.h"
40
#include "exec/address-spaces.h"
41
#include "hw/hotplug.h"
42
#include "hw/boards.h"
43
#include "qemu/cutils.h"
B
bellard 已提交
44 45

//#define DEBUG_PCI
46
#ifdef DEBUG_PCI
47
# define PCI_DPRINTF(format, ...)       printf(format, ## __VA_ARGS__)
48 49 50
#else
# define PCI_DPRINTF(format, ...)       do { } while (0)
#endif
B
bellard 已提交
51

52
static void pcibus_dev_print(Monitor *mon, DeviceState *dev, int indent);
53
static char *pcibus_get_dev_path(DeviceState *dev);
54
static char *pcibus_get_fw_dev_path(DeviceState *dev);
P
Paolo Bonzini 已提交
55
static void pcibus_reset(BusState *qbus);
56

57 58 59 60 61 62 63 64
static Property pci_props[] = {
    DEFINE_PROP_PCI_DEVFN("addr", PCIDevice, devfn, -1),
    DEFINE_PROP_STRING("romfile", PCIDevice, romfile),
    DEFINE_PROP_UINT32("rombar",  PCIDevice, rom_bar, 1),
    DEFINE_PROP_BIT("multifunction", PCIDevice, cap_present,
                    QEMU_PCI_CAP_MULTIFUNCTION_BITNR, false),
    DEFINE_PROP_BIT("command_serr_enable", PCIDevice, cap_present,
                    QEMU_PCI_CAP_SERR_BITNR, true),
65 66
    DEFINE_PROP_BIT("x-pcie-lnksta-dllla", PCIDevice, cap_present,
                    QEMU_PCIE_LNKSTA_DLLLA_BITNR, true),
67 68 69
    DEFINE_PROP_END_OF_LIST()
};

70 71 72 73
static const VMStateDescription vmstate_pcibus = {
    .name = "PCIBUS",
    .version_id = 1,
    .minimum_version_id = 1,
74
    .fields = (VMStateField[]) {
75 76 77 78 79 80 81 82
        VMSTATE_INT32_EQUAL(nirq, PCIBus),
        VMSTATE_VARRAY_INT32(irq_count, PCIBus,
                             nirq, 0, vmstate_info_int32,
                             int32_t),
        VMSTATE_END_OF_LIST()
    }
};

83 84 85 86 87 88 89 90
static void pci_init_bus_master(PCIDevice *pci_dev)
{
    AddressSpace *dma_as = pci_device_iommu_address_space(pci_dev);

    memory_region_init_alias(&pci_dev->bus_master_enable_region,
                             OBJECT(pci_dev), "bus master",
                             dma_as->root, 0, memory_region_size(dma_as->root));
    memory_region_set_enabled(&pci_dev->bus_master_enable_region, false);
91 92
    memory_region_add_subregion(&pci_dev->bus_master_container_region, 0,
                                &pci_dev->bus_master_enable_region);
93 94 95 96 97 98 99 100 101 102 103 104 105 106
}

static void pcibus_machine_done(Notifier *notifier, void *data)
{
    PCIBus *bus = container_of(notifier, PCIBus, machine_done);
    int i;

    for (i = 0; i < ARRAY_SIZE(bus->devices); ++i) {
        if (bus->devices[i]) {
            pci_init_bus_master(bus->devices[i]);
        }
    }
}

107 108 109 110
static void pci_bus_realize(BusState *qbus, Error **errp)
{
    PCIBus *bus = PCI_BUS(qbus);

111 112 113
    bus->machine_done.notify = pcibus_machine_done;
    qemu_add_machine_init_done_notifier(&bus->machine_done);

114 115 116 117 118 119 120
    vmstate_register(NULL, -1, &vmstate_pcibus, bus);
}

static void pci_bus_unrealize(BusState *qbus, Error **errp)
{
    PCIBus *bus = PCI_BUS(qbus);

121 122
    qemu_remove_machine_init_done_notifier(&bus->machine_done);

123 124 125
    vmstate_unregister(NULL, &vmstate_pcibus, bus);
}

126 127 128 129 130
static bool pcibus_is_root(PCIBus *bus)
{
    return !bus->parent_dev;
}

131 132 133 134 135 136 137 138
static int pcibus_num(PCIBus *bus)
{
    if (pcibus_is_root(bus)) {
        return 0; /* pci host bridge */
    }
    return bus->parent_dev->config[PCI_SECONDARY_BUS];
}

139 140 141 142 143
static uint16_t pcibus_numa_node(PCIBus *bus)
{
    return NUMA_NODE_UNASSIGNED;
}

144 145 146
static void pci_bus_class_init(ObjectClass *klass, void *data)
{
    BusClass *k = BUS_CLASS(klass);
147
    PCIBusClass *pbc = PCI_BUS_CLASS(klass);
148 149 150 151

    k->print_dev = pcibus_dev_print;
    k->get_dev_path = pcibus_get_dev_path;
    k->get_fw_dev_path = pcibus_get_fw_dev_path;
152 153
    k->realize = pci_bus_realize;
    k->unrealize = pci_bus_unrealize;
154
    k->reset = pcibus_reset;
155 156

    pbc->is_root = pcibus_is_root;
157
    pbc->bus_num = pcibus_num;
158
    pbc->numa_node = pcibus_numa_node;
159 160 161 162 163 164
}

static const TypeInfo pci_bus_info = {
    .name = TYPE_PCI_BUS,
    .parent = TYPE_BUS,
    .instance_size = sizeof(PCIBus),
165
    .class_size = sizeof(PCIBusClass),
166
    .class_init = pci_bus_class_init,
167
};
B
bellard 已提交
168

169 170 171 172 173
static const TypeInfo pcie_bus_info = {
    .name = TYPE_PCIE_BUS,
    .parent = TYPE_PCI_BUS,
};

174
static PCIBus *pci_find_bus_nr(PCIBus *bus, int bus_num);
B
bellard 已提交
175
static void pci_update_mappings(PCIDevice *d);
176
static void pci_irq_handler(void *opaque, int irq_num, int level);
M
Markus Armbruster 已提交
177
static void pci_add_option_rom(PCIDevice *pdev, bool is_default_rom, Error **);
178
static void pci_del_option_rom(PCIDevice *pdev);
B
bellard 已提交
179

180 181
static uint16_t pci_default_sub_vendor_id = PCI_SUBVENDOR_ID_REDHAT_QUMRANET;
static uint16_t pci_default_sub_device_id = PCI_SUBDEVICE_ID_QEMU;
182

183
static QLIST_HEAD(, PCIHostState) pci_host_bridges;
184

185
int pci_bar(PCIDevice *d, int reg)
186
{
187 188 189 190 191 192 193
    uint8_t type;

    if (reg != PCI_ROM_SLOT)
        return PCI_BASE_ADDRESS_0 + reg * 4;

    type = d->config[PCI_HEADER_TYPE] & ~PCI_HEADER_TYPE_MULTI_FUNCTION;
    return type == PCI_HEADER_TYPE_BRIDGE ? PCI_ROM_ADDRESS1 : PCI_ROM_ADDRESS;
194 195
}

196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220
static inline int pci_irq_state(PCIDevice *d, int irq_num)
{
	return (d->irq_state >> irq_num) & 0x1;
}

static inline void pci_set_irq_state(PCIDevice *d, int irq_num, int level)
{
	d->irq_state &= ~(0x1 << irq_num);
	d->irq_state |= level << irq_num;
}

static void pci_change_irq_level(PCIDevice *pci_dev, int irq_num, int change)
{
    PCIBus *bus;
    for (;;) {
        bus = pci_dev->bus;
        irq_num = bus->map_irq(pci_dev, irq_num);
        if (bus->set_irq)
            break;
        pci_dev = bus->parent_dev;
    }
    bus->irq_count[irq_num] += change;
    bus->set_irq(bus->irq_opaque, irq_num, bus->irq_count[irq_num] != 0);
}

221 222 223 224 225 226 227
int pci_bus_get_irq_level(PCIBus *bus, int irq_num)
{
    assert(irq_num >= 0);
    assert(irq_num < bus->nirq);
    return !!bus->irq_count[irq_num];
}

228 229 230 231 232 233 234 235 236 237 238
/* Update interrupt status bit in config space on interrupt
 * state change. */
static void pci_update_irq_status(PCIDevice *dev)
{
    if (dev->irq_state) {
        dev->config[PCI_STATUS] |= PCI_STATUS_INTERRUPT;
    } else {
        dev->config[PCI_STATUS] &= ~PCI_STATUS_INTERRUPT;
    }
}

I
Isaku Yamahata 已提交
239 240 241 242
void pci_device_deassert_intx(PCIDevice *dev)
{
    int i;
    for (i = 0; i < PCI_NUM_PINS; ++i) {
243
        pci_irq_handler(dev, i, 0);
I
Isaku Yamahata 已提交
244 245 246
    }
}

P
Paolo Bonzini 已提交
247
static void pci_do_device_reset(PCIDevice *dev)
248
{
249
    int r;
250

I
Isaku Yamahata 已提交
251
    pci_device_deassert_intx(dev);
C
Cole Robinson 已提交
252 253
    assert(dev->irq_state == 0);

S
Stefan Weil 已提交
254
    /* Clear all writable bits */
255
    pci_word_test_and_clear_mask(dev->config + PCI_COMMAND,
256 257
                                 pci_get_word(dev->wmask + PCI_COMMAND) |
                                 pci_get_word(dev->w1cmask + PCI_COMMAND));
258 259 260
    pci_word_test_and_clear_mask(dev->config + PCI_STATUS,
                                 pci_get_word(dev->wmask + PCI_STATUS) |
                                 pci_get_word(dev->w1cmask + PCI_STATUS));
261 262 263
    dev->config[PCI_CACHE_LINE_SIZE] = 0x0;
    dev->config[PCI_INTERRUPT_LINE] = 0x0;
    for (r = 0; r < PCI_NUM_REGIONS; ++r) {
I
Isaku Yamahata 已提交
264 265
        PCIIORegion *region = &dev->io_regions[r];
        if (!region->size) {
266 267
            continue;
        }
I
Isaku Yamahata 已提交
268 269 270 271 272 273 274

        if (!(region->type & PCI_BASE_ADDRESS_SPACE_IO) &&
            region->type & PCI_BASE_ADDRESS_MEM_TYPE_64) {
            pci_set_quad(dev->config + pci_bar(dev, r), region->type);
        } else {
            pci_set_long(dev->config + pci_bar(dev, r), region->type);
        }
275 276
    }
    pci_update_mappings(dev);
277 278 279

    msi_reset(dev);
    msix_reset(dev);
280 281
}

P
Paolo Bonzini 已提交
282 283 284 285 286 287 288 289 290 291
/*
 * This function is called on #RST and FLR.
 * FLR if PCI_EXP_DEVCTL_BCR_FLR is set
 */
void pci_device_reset(PCIDevice *dev)
{
    qdev_reset_all(&dev->qdev);
    pci_do_device_reset(dev);
}

292 293
/*
 * Trigger pci bus reset under a given bus.
P
Paolo Bonzini 已提交
294 295
 * Called via qbus_reset_all on RST# assert, after the devices
 * have been reset qdev_reset_all-ed already.
296
 */
P
Paolo Bonzini 已提交
297
static void pcibus_reset(BusState *qbus)
G
Gleb Natapov 已提交
298
{
P
Paolo Bonzini 已提交
299
    PCIBus *bus = DO_UPCAST(PCIBus, qbus, qbus);
G
Gleb Natapov 已提交
300 301
    int i;

302 303
    for (i = 0; i < ARRAY_SIZE(bus->devices); ++i) {
        if (bus->devices[i]) {
P
Paolo Bonzini 已提交
304
            pci_do_device_reset(bus->devices[i]);
305
        }
G
Gleb Natapov 已提交
306
    }
307

P
Paolo Bonzini 已提交
308 309 310
    for (i = 0; i < bus->nirq; i++) {
        assert(bus->irq_count[i] == 0);
    }
311 312
}

313
static void pci_host_bus_register(DeviceState *host)
314
{
315
    PCIHostState *host_bridge = PCI_HOST_BRIDGE(host);
316 317

    QLIST_INSERT_HEAD(&pci_host_bridges, host_bridge, next);
318 319
}

D
David Gibson 已提交
320
PCIBus *pci_find_primary_bus(void)
321
{
322
    PCIBus *primary_bus = NULL;
323
    PCIHostState *host;
324

325
    QLIST_FOREACH(host, &pci_host_bridges, next) {
326 327 328
        if (primary_bus) {
            /* We have multiple root buses, refuse to select a primary */
            return NULL;
329
        }
330
        primary_bus = host->bus;
331 332
    }

333
    return primary_bus;
334 335
}

336
PCIBus *pci_device_root_bus(const PCIDevice *d)
337
{
338
    PCIBus *bus = d->bus;
339

340 341 342 343
    while (!pci_bus_is_root(bus)) {
        d = bus->parent_dev;
        assert(d != NULL);

344 345 346
        bus = d->bus;
    }

347 348 349
    return bus;
}

350
const char *pci_root_bus_path(PCIDevice *dev)
351
{
352 353 354
    PCIBus *rootbus = pci_device_root_bus(dev);
    PCIHostState *host_bridge = PCI_HOST_BRIDGE(rootbus->qbus.parent);
    PCIHostBridgeClass *hc = PCI_HOST_BRIDGE_GET_CLASS(host_bridge);
355

356 357 358 359
    assert(host_bridge->bus == rootbus);

    if (hc->root_bus_path) {
        return (*hc->root_bus_path)(host_bridge, rootbus);
360 361
    }

362
    return rootbus->qbus.name;
363 364
}

365
static void pci_bus_init(PCIBus *bus, DeviceState *parent,
366 367
                         MemoryRegion *address_space_mem,
                         MemoryRegion *address_space_io,
368
                         uint8_t devfn_min)
369
{
370
    assert(PCI_FUNC(devfn_min) == 0);
P
pbrook 已提交
371
    bus->devfn_min = devfn_min;
372 373
    bus->address_space_mem = address_space_mem;
    bus->address_space_io = address_space_io;
374 375 376

    /* host bridge */
    QLIST_INIT(&bus->child);
377

378
    pci_host_bus_register(parent);
379 380
}

381 382 383 384 385
bool pci_bus_is_express(PCIBus *bus)
{
    return object_dynamic_cast(OBJECT(bus), TYPE_PCIE_BUS);
}

386 387
bool pci_bus_is_root(PCIBus *bus)
{
388
    return PCI_BUS_GET_CLASS(bus)->is_root(bus);
389 390
}

391
void pci_bus_new_inplace(PCIBus *bus, size_t bus_size, DeviceState *parent,
392 393 394
                         const char *name,
                         MemoryRegion *address_space_mem,
                         MemoryRegion *address_space_io,
395
                         uint8_t devfn_min, const char *typename)
396
{
397
    qbus_create_inplace(bus, bus_size, typename, parent, name);
398
    pci_bus_init(bus, parent, address_space_mem, address_space_io, devfn_min);
399 400
}

401
PCIBus *pci_bus_new(DeviceState *parent, const char *name,
402 403
                    MemoryRegion *address_space_mem,
                    MemoryRegion *address_space_io,
404
                    uint8_t devfn_min, const char *typename)
405 406 407
{
    PCIBus *bus;

408
    bus = PCI_BUS(qbus_create(typename, parent, name));
409
    pci_bus_init(bus, parent, address_space_mem, address_space_io, devfn_min);
410 411 412 413 414 415 416 417 418 419
    return bus;
}

void pci_bus_irqs(PCIBus *bus, pci_set_irq_fn set_irq, pci_map_irq_fn map_irq,
                  void *irq_opaque, int nirq)
{
    bus->set_irq = set_irq;
    bus->map_irq = map_irq;
    bus->irq_opaque = irq_opaque;
    bus->nirq = nirq;
420
    bus->irq_count = g_malloc0(nirq * sizeof(bus->irq_count[0]));
421 422 423 424
}

PCIBus *pci_register_bus(DeviceState *parent, const char *name,
                         pci_set_irq_fn set_irq, pci_map_irq_fn map_irq,
425
                         void *irq_opaque,
426 427
                         MemoryRegion *address_space_mem,
                         MemoryRegion *address_space_io,
428
                         uint8_t devfn_min, int nirq, const char *typename)
429 430 431
{
    PCIBus *bus;

432
    bus = pci_bus_new(parent, name, address_space_mem,
433
                      address_space_io, devfn_min, typename);
434
    pci_bus_irqs(bus, set_irq, map_irq, irq_opaque, nirq);
435 436
    return bus;
}
B
bellard 已提交
437

P
pbrook 已提交
438 439
int pci_bus_num(PCIBus *s)
{
440
    return PCI_BUS_GET_CLASS(s)->bus_num(s);
P
pbrook 已提交
441 442
}

443 444 445
int pci_bus_numa_node(PCIBus *bus)
{
    return PCI_BUS_GET_CLASS(bus)->numa_node(bus);
P
pbrook 已提交
446 447
}

J
Jianjun Duan 已提交
448 449
static int get_pci_config_device(QEMUFile *f, void *pv, size_t size,
                                 VMStateField *field)
450
{
J
Juan Quintela 已提交
451
    PCIDevice *s = container_of(pv, PCIDevice, config);
452
    PCIDeviceClass *pc = PCI_DEVICE_GET_CLASS(s);
I
Isaku Yamahata 已提交
453
    uint8_t *config;
454 455
    int i;

I
Isaku Yamahata 已提交
456
    assert(size == pci_config_size(s));
457
    config = g_malloc(size);
I
Isaku Yamahata 已提交
458 459 460

    qemu_get_buffer(f, config, size);
    for (i = 0; i < size; ++i) {
461 462
        if ((config[i] ^ s->config[i]) &
            s->cmask[i] & ~s->wmask[i] & ~s->w1cmask[i]) {
463 464 465 466
            error_report("%s: Bad config data: i=0x%x read: %x device: %x "
                         "cmask: %x wmask: %x w1cmask:%x", __func__,
                         i, config[i], s->config[i],
                         s->cmask[i], s->wmask[i], s->w1cmask[i]);
467
            g_free(config);
468
            return -EINVAL;
I
Isaku Yamahata 已提交
469 470 471
        }
    }
    memcpy(s->config, config, size);
472

B
bellard 已提交
473
    pci_update_mappings(s);
474
    if (pc->is_bridge) {
475
        PCIBridge *b = PCI_BRIDGE(s);
476 477
        pci_bridge_update_mappings(b);
    }
478

G
Gerd Hoffmann 已提交
479 480 481 482
    memory_region_set_enabled(&s->bus_master_enable_region,
                              pci_get_word(s->config + PCI_COMMAND)
                              & PCI_COMMAND_MASTER);

483
    g_free(config);
484 485 486
    return 0;
}

J
Juan Quintela 已提交
487
/* just put buffer */
J
Jianjun Duan 已提交
488 489
static int put_pci_config_device(QEMUFile *f, void *pv, size_t size,
                                 VMStateField *field, QJSON *vmdesc)
J
Juan Quintela 已提交
490
{
491
    const uint8_t **v = pv;
I
Isaku Yamahata 已提交
492
    assert(size == pci_config_size(container_of(pv, PCIDevice, config)));
493
    qemu_put_buffer(f, *v, size);
J
Jianjun Duan 已提交
494 495

    return 0;
J
Juan Quintela 已提交
496 497 498 499 500 501 502 503
}

static VMStateInfo vmstate_info_pci_config = {
    .name = "pci config",
    .get  = get_pci_config_device,
    .put  = put_pci_config_device,
};

J
Jianjun Duan 已提交
504 505
static int get_pci_irq_state(QEMUFile *f, void *pv, size_t size,
                             VMStateField *field)
506
{
507
    PCIDevice *s = container_of(pv, PCIDevice, irq_state);
508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525
    uint32_t irq_state[PCI_NUM_PINS];
    int i;
    for (i = 0; i < PCI_NUM_PINS; ++i) {
        irq_state[i] = qemu_get_be32(f);
        if (irq_state[i] != 0x1 && irq_state[i] != 0) {
            fprintf(stderr, "irq state %d: must be 0 or 1.\n",
                    irq_state[i]);
            return -EINVAL;
        }
    }

    for (i = 0; i < PCI_NUM_PINS; ++i) {
        pci_set_irq_state(s, i, irq_state[i]);
    }

    return 0;
}

J
Jianjun Duan 已提交
526 527
static int put_pci_irq_state(QEMUFile *f, void *pv, size_t size,
                             VMStateField *field, QJSON *vmdesc)
528 529
{
    int i;
530
    PCIDevice *s = container_of(pv, PCIDevice, irq_state);
531 532 533 534

    for (i = 0; i < PCI_NUM_PINS; ++i) {
        qemu_put_be32(f, pci_irq_state(s, i));
    }
J
Jianjun Duan 已提交
535 536

    return 0;
537 538 539 540 541 542 543 544
}

static VMStateInfo vmstate_info_pci_irq_state = {
    .name = "pci irq state",
    .get  = get_pci_irq_state,
    .put  = put_pci_irq_state,
};

545 546 547 548 549 550 551 552 553 554
static bool migrate_is_pcie(void *opaque, int version_id)
{
    return pci_is_express((PCIDevice *)opaque);
}

static bool migrate_is_not_pcie(void *opaque, int version_id)
{
    return !pci_is_express((PCIDevice *)opaque);
}

J
Juan Quintela 已提交
555 556 557 558
const VMStateDescription vmstate_pci_device = {
    .name = "PCIDevice",
    .version_id = 2,
    .minimum_version_id = 1,
559
    .fields = (VMStateField[]) {
560
        VMSTATE_INT32_POSITIVE_LE(version_id, PCIDevice),
561 562 563
        VMSTATE_BUFFER_UNSAFE_INFO_TEST(config, PCIDevice,
                                   migrate_is_not_pcie,
                                   0, vmstate_info_pci_config,
I
Isaku Yamahata 已提交
564
                                   PCI_CONFIG_SPACE_SIZE),
565 566 567
        VMSTATE_BUFFER_UNSAFE_INFO_TEST(config, PCIDevice,
                                   migrate_is_pcie,
                                   0, vmstate_info_pci_config,
I
Isaku Yamahata 已提交
568
                                   PCIE_CONFIG_SPACE_SIZE),
569 570 571
        VMSTATE_BUFFER_UNSAFE_INFO(irq_state, PCIDevice, 2,
				   vmstate_info_pci_irq_state,
				   PCI_NUM_PINS * sizeof(int32_t)),
J
Juan Quintela 已提交
572 573 574 575
        VMSTATE_END_OF_LIST()
    }
};

I
Isaku Yamahata 已提交
576

J
Juan Quintela 已提交
577 578
void pci_device_save(PCIDevice *s, QEMUFile *f)
{
579 580 581 582 583
    /* Clear interrupt status bit: it is implicit
     * in irq_state which we are saving.
     * This makes us compatible with old devices
     * which never set or clear this bit. */
    s->config[PCI_STATUS] &= ~PCI_STATUS_INTERRUPT;
584
    vmstate_save_state(f, &vmstate_pci_device, s, NULL);
585 586
    /* Restore the interrupt status bit. */
    pci_update_irq_status(s);
J
Juan Quintela 已提交
587 588 589 590
}

int pci_device_load(PCIDevice *s, QEMUFile *f)
{
591
    int ret;
592
    ret = vmstate_load_state(f, &vmstate_pci_device, s, s->version_id);
593 594 595
    /* Restore the interrupt status bit. */
    pci_update_irq_status(s);
    return ret;
J
Juan Quintela 已提交
596 597
}

598
static void pci_set_default_subsystem_id(PCIDevice *pci_dev)
599
{
600 601 602 603
    pci_set_word(pci_dev->config + PCI_SUBSYSTEM_VENDOR_ID,
                 pci_default_sub_vendor_id);
    pci_set_word(pci_dev->config + PCI_SUBSYSTEM_ID,
                 pci_default_sub_device_id);
604 605
}

606
/*
607 608
 * Parse [[<domain>:]<bus>:]<slot>, return -1 on error if funcp == NULL
 *       [[<domain>:]<bus>:]<slot>.<func>, return -1 on error
609
 */
610 611
static int pci_parse_devaddr(const char *addr, int *domp, int *busp,
                             unsigned int *slotp, unsigned int *funcp)
612 613 614 615 616
{
    const char *p;
    char *e;
    unsigned long val;
    unsigned long dom = 0, bus = 0;
617 618
    unsigned int slot = 0;
    unsigned int func = 0;
619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641

    p = addr;
    val = strtoul(p, &e, 16);
    if (e == p)
	return -1;
    if (*e == ':') {
	bus = val;
	p = e + 1;
	val = strtoul(p, &e, 16);
	if (e == p)
	    return -1;
	if (*e == ':') {
	    dom = bus;
	    bus = val;
	    p = e + 1;
	    val = strtoul(p, &e, 16);
	    if (e == p)
		return -1;
	}
    }

    slot = val;

642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657
    if (funcp != NULL) {
        if (*e != '.')
            return -1;

        p = e + 1;
        val = strtoul(p, &e, 16);
        if (e == p)
            return -1;

        func = val;
    }

    /* if funcp == NULL func is 0 */
    if (dom > 0xffff || bus > 0xff || slot > 0x1f || func > 7)
	return -1;

658 659 660 661 662 663
    if (*e)
	return -1;

    *domp = dom;
    *busp = bus;
    *slotp = slot;
664 665
    if (funcp != NULL)
        *funcp = func;
666 667 668
    return 0;
}

669 670
static PCIBus *pci_get_bus_devfn(int *devfnp, PCIBus *root,
                                 const char *devaddr)
671 672 673 674
{
    int dom, bus;
    unsigned slot;

D
David Gibson 已提交
675 676 677 678 679
    if (!root) {
        fprintf(stderr, "No primary PCI bus\n");
        return NULL;
    }

680 681
    assert(!root->parent_dev);

682 683
    if (!devaddr) {
        *devfnp = -1;
D
David Gibson 已提交
684
        return pci_find_bus_nr(root, 0);
685 686
    }

687
    if (pci_parse_devaddr(devaddr, &dom, &bus, &slot, NULL) < 0) {
688 689 690
        return NULL;
    }

D
David Gibson 已提交
691 692 693 694 695
    if (dom != 0) {
        fprintf(stderr, "No support for non-zero PCI domains\n");
        return NULL;
    }

696
    *devfnp = PCI_DEVFN(slot, 0);
D
David Gibson 已提交
697
    return pci_find_bus_nr(root, bus);
698 699
}

700 701 702 703 704 705 706 707 708 709 710 711
static void pci_init_cmask(PCIDevice *dev)
{
    pci_set_word(dev->cmask + PCI_VENDOR_ID, 0xffff);
    pci_set_word(dev->cmask + PCI_DEVICE_ID, 0xffff);
    dev->cmask[PCI_STATUS] = PCI_STATUS_CAP_LIST;
    dev->cmask[PCI_REVISION_ID] = 0xff;
    dev->cmask[PCI_CLASS_PROG] = 0xff;
    pci_set_word(dev->cmask + PCI_CLASS_DEVICE, 0xffff);
    dev->cmask[PCI_HEADER_TYPE] = 0xff;
    dev->cmask[PCI_CAPABILITY_LIST] = 0xff;
}

712 713
static void pci_init_wmask(PCIDevice *dev)
{
I
Isaku Yamahata 已提交
714 715
    int config_size = pci_config_size(dev);

716 717
    dev->wmask[PCI_CACHE_LINE_SIZE] = 0xff;
    dev->wmask[PCI_INTERRUPT_LINE] = 0xff;
I
Isaku Yamahata 已提交
718
    pci_set_word(dev->wmask + PCI_COMMAND,
719 720
                 PCI_COMMAND_IO | PCI_COMMAND_MEMORY | PCI_COMMAND_MASTER |
                 PCI_COMMAND_INTX_DISABLE);
721 722 723
    if (dev->cap_present & QEMU_PCI_CAP_SERR) {
        pci_word_test_and_set_mask(dev->wmask + PCI_COMMAND, PCI_COMMAND_SERR);
    }
724 725 726

    memset(dev->wmask + PCI_CONFIG_HEADER_SIZE, 0xff,
           config_size - PCI_CONFIG_HEADER_SIZE);
727 728
}

729 730 731
static void pci_init_w1cmask(PCIDevice *dev)
{
    /*
732
     * Note: It's okay to set w1cmask even for readonly bits as
733 734 735 736 737 738 739 740
     * long as their value is hardwired to 0.
     */
    pci_set_word(dev->w1cmask + PCI_STATUS,
                 PCI_STATUS_PARITY | PCI_STATUS_SIG_TARGET_ABORT |
                 PCI_STATUS_REC_TARGET_ABORT | PCI_STATUS_REC_MASTER_ABORT |
                 PCI_STATUS_SIG_SYSTEM_ERROR | PCI_STATUS_DETECTED_PARITY);
}

741
static void pci_init_mask_bridge(PCIDevice *d)
742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761
{
    /* PCI_PRIMARY_BUS, PCI_SECONDARY_BUS, PCI_SUBORDINATE_BUS and
       PCI_SEC_LETENCY_TIMER */
    memset(d->wmask + PCI_PRIMARY_BUS, 0xff, 4);

    /* base and limit */
    d->wmask[PCI_IO_BASE] = PCI_IO_RANGE_MASK & 0xff;
    d->wmask[PCI_IO_LIMIT] = PCI_IO_RANGE_MASK & 0xff;
    pci_set_word(d->wmask + PCI_MEMORY_BASE,
                 PCI_MEMORY_RANGE_MASK & 0xffff);
    pci_set_word(d->wmask + PCI_MEMORY_LIMIT,
                 PCI_MEMORY_RANGE_MASK & 0xffff);
    pci_set_word(d->wmask + PCI_PREF_MEMORY_BASE,
                 PCI_PREF_RANGE_MASK & 0xffff);
    pci_set_word(d->wmask + PCI_PREF_MEMORY_LIMIT,
                 PCI_PREF_RANGE_MASK & 0xffff);

    /* PCI_PREF_BASE_UPPER32 and PCI_PREF_LIMIT_UPPER32 */
    memset(d->wmask + PCI_PREF_BASE_UPPER32, 0xff, 8);

762
    /* Supported memory and i/o types */
M
Michael S. Tsirkin 已提交
763 764
    d->config[PCI_IO_BASE] |= PCI_IO_RANGE_TYPE_16;
    d->config[PCI_IO_LIMIT] |= PCI_IO_RANGE_TYPE_16;
765 766 767 768 769
    pci_word_test_and_set_mask(d->config + PCI_PREF_MEMORY_BASE,
                               PCI_PREF_RANGE_TYPE_64);
    pci_word_test_and_set_mask(d->config + PCI_PREF_MEMORY_LIMIT,
                               PCI_PREF_RANGE_TYPE_64);

770 771 772 773
    /*
     * TODO: Bridges default to 10-bit VGA decoding but we currently only
     * implement 16-bit decoding (no alias support).
     */
774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789
    pci_set_word(d->wmask + PCI_BRIDGE_CONTROL,
                 PCI_BRIDGE_CTL_PARITY |
                 PCI_BRIDGE_CTL_SERR |
                 PCI_BRIDGE_CTL_ISA |
                 PCI_BRIDGE_CTL_VGA |
                 PCI_BRIDGE_CTL_VGA_16BIT |
                 PCI_BRIDGE_CTL_MASTER_ABORT |
                 PCI_BRIDGE_CTL_BUS_RESET |
                 PCI_BRIDGE_CTL_FAST_BACK |
                 PCI_BRIDGE_CTL_DISCARD |
                 PCI_BRIDGE_CTL_SEC_DISCARD |
                 PCI_BRIDGE_CTL_DISCARD_SERR);
    /* Below does not do anything as we never set this bit, put here for
     * completeness. */
    pci_set_word(d->w1cmask + PCI_BRIDGE_CONTROL,
                 PCI_BRIDGE_CTL_DISCARD_STATUS);
790
    d->cmask[PCI_IO_BASE] |= PCI_IO_RANGE_TYPE_MASK;
791
    d->cmask[PCI_IO_LIMIT] |= PCI_IO_RANGE_TYPE_MASK;
792 793
    pci_word_test_and_set_mask(d->cmask + PCI_PREF_MEMORY_BASE,
                               PCI_PREF_RANGE_TYPE_MASK);
794 795
    pci_word_test_and_set_mask(d->cmask + PCI_PREF_MEMORY_LIMIT,
                               PCI_PREF_RANGE_TYPE_MASK);
796 797
}

M
Markus Armbruster 已提交
798
static void pci_init_multifunction(PCIBus *bus, PCIDevice *dev, Error **errp)
799 800 801 802 803 804 805 806 807
{
    uint8_t slot = PCI_SLOT(dev->devfn);
    uint8_t func;

    if (dev->cap_present & QEMU_PCI_CAP_MULTIFUNCTION) {
        dev->config[PCI_HEADER_TYPE] |= PCI_HEADER_TYPE_MULTI_FUNCTION;
    }

    /*
S
Stefan Weil 已提交
808
     * multifunction bit is interpreted in two ways as follows.
809 810 811 812 813 814 815 816 817 818 819 820 821 822 823
     *   - all functions must set the bit to 1.
     *     Example: Intel X53
     *   - function 0 must set the bit, but the rest function (> 0)
     *     is allowed to leave the bit to 0.
     *     Example: PIIX3(also in qemu), PIIX4(also in qemu), ICH10,
     *
     * So OS (at least Linux) checks the bit of only function 0,
     * and doesn't see the bit of function > 0.
     *
     * The below check allows both interpretation.
     */
    if (PCI_FUNC(dev->devfn)) {
        PCIDevice *f0 = bus->devices[PCI_DEVFN(slot, 0)];
        if (f0 && !(f0->cap_present & QEMU_PCI_CAP_MULTIFUNCTION)) {
            /* function 0 should set multifunction bit */
M
Markus Armbruster 已提交
824 825 826
            error_setg(errp, "PCI: single function device can't be populated "
                       "in function %x.%x", slot, PCI_FUNC(dev->devfn));
            return;
827
        }
M
Markus Armbruster 已提交
828
        return;
829 830 831
    }

    if (dev->cap_present & QEMU_PCI_CAP_MULTIFUNCTION) {
M
Markus Armbruster 已提交
832
        return;
833 834 835 836
    }
    /* function 0 indicates single function, so function > 0 must be NULL */
    for (func = 1; func < PCI_FUNC_MAX; ++func) {
        if (bus->devices[PCI_DEVFN(slot, func)]) {
M
Markus Armbruster 已提交
837 838 839 840
            error_setg(errp, "PCI: %x.0 indicates single function, "
                       "but %x.%x is already populated.",
                       slot, slot, func);
            return;
841 842 843 844
        }
    }
}

I
Isaku Yamahata 已提交
845 846 847 848
static void pci_config_alloc(PCIDevice *pci_dev)
{
    int config_size = pci_config_size(pci_dev);

849 850 851 852 853
    pci_dev->config = g_malloc0(config_size);
    pci_dev->cmask = g_malloc0(config_size);
    pci_dev->wmask = g_malloc0(config_size);
    pci_dev->w1cmask = g_malloc0(config_size);
    pci_dev->used = g_malloc0(config_size);
I
Isaku Yamahata 已提交
854 855 856 857
}

static void pci_config_free(PCIDevice *pci_dev)
{
858 859 860 861 862
    g_free(pci_dev->config);
    g_free(pci_dev->cmask);
    g_free(pci_dev->wmask);
    g_free(pci_dev->w1cmask);
    g_free(pci_dev->used);
I
Isaku Yamahata 已提交
863 864
}

865 866 867 868 869 870 871 872
static void do_pci_unregister_device(PCIDevice *pci_dev)
{
    pci_dev->bus->devices[pci_dev->devfn] = NULL;
    pci_config_free(pci_dev);

    address_space_destroy(&pci_dev->bus_master_as);
}

P
Peter Xu 已提交
873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947
/* Extract PCIReqIDCache into BDF format */
static uint16_t pci_req_id_cache_extract(PCIReqIDCache *cache)
{
    uint8_t bus_n;
    uint16_t result;

    switch (cache->type) {
    case PCI_REQ_ID_BDF:
        result = pci_get_bdf(cache->dev);
        break;
    case PCI_REQ_ID_SECONDARY_BUS:
        bus_n = pci_bus_num(cache->dev->bus);
        result = PCI_BUILD_BDF(bus_n, 0);
        break;
    default:
        error_printf("Invalid PCI requester ID cache type: %d\n",
                     cache->type);
        exit(1);
        break;
    }

    return result;
}

/* Parse bridges up to the root complex and return requester ID
 * cache for specific device.  For full PCIe topology, the cache
 * result would be exactly the same as getting BDF of the device.
 * However, several tricks are required when system mixed up with
 * legacy PCI devices and PCIe-to-PCI bridges.
 *
 * Here we cache the proxy device (and type) not requester ID since
 * bus number might change from time to time.
 */
static PCIReqIDCache pci_req_id_cache_get(PCIDevice *dev)
{
    PCIDevice *parent;
    PCIReqIDCache cache = {
        .dev = dev,
        .type = PCI_REQ_ID_BDF,
    };

    while (!pci_bus_is_root(dev->bus)) {
        /* We are under PCI/PCIe bridges */
        parent = dev->bus->parent_dev;
        if (pci_is_express(parent)) {
            if (pcie_cap_get_type(parent) == PCI_EXP_TYPE_PCI_BRIDGE) {
                /* When we pass through PCIe-to-PCI/PCIX bridges, we
                 * override the requester ID using secondary bus
                 * number of parent bridge with zeroed devfn
                 * (pcie-to-pci bridge spec chap 2.3). */
                cache.type = PCI_REQ_ID_SECONDARY_BUS;
                cache.dev = dev;
            }
        } else {
            /* Legacy PCI, override requester ID with the bridge's
             * BDF upstream.  When the root complex connects to
             * legacy PCI devices (including buses), it can only
             * obtain requester ID info from directly attached
             * devices.  If devices are attached under bridges, only
             * the requester ID of the bridge that is directly
             * attached to the root complex can be recognized. */
            cache.type = PCI_REQ_ID_BDF;
            cache.dev = parent;
        }
        dev = parent;
    }

    return cache;
}

uint16_t pci_requester_id(PCIDevice *dev)
{
    return pci_req_id_cache_extract(&dev->requester_id_cache);
}

B
bellard 已提交
948
/* -1 for devfn means auto assign */
P
Paul Brook 已提交
949
static PCIDevice *do_pci_register_device(PCIDevice *pci_dev, PCIBus *bus,
M
Markus Armbruster 已提交
950 951
                                         const char *name, int devfn,
                                         Error **errp)
B
bellard 已提交
952
{
953 954 955
    PCIDeviceClass *pc = PCI_DEVICE_GET_CLASS(pci_dev);
    PCIConfigReadFunc *config_read = pc->config_read;
    PCIConfigWriteFunc *config_write = pc->config_write;
M
Markus Armbruster 已提交
956
    Error *local_err = NULL;
C
Cao jin 已提交
957 958 959
    DeviceState *dev = DEVICE(pci_dev);

    pci_dev->bus = bus;
960 961 962 963 964 965 966
    /* Only pci bridges can be attached to extra PCI root buses */
    if (pci_bus_is_root(bus) && bus->parent_dev && !pc->is_bridge) {
        error_setg(errp,
                   "PCI: Only PCI/PCIe bridges can be plugged into %s",
                    bus->parent_dev->name);
        return NULL;
    }
967

B
bellard 已提交
968
    if (devfn < 0) {
969
        for(devfn = bus->devfn_min ; devfn < ARRAY_SIZE(bus->devices);
970
            devfn += PCI_FUNC_MAX) {
971
            if (!bus->devices[devfn])
B
bellard 已提交
972 973
                goto found;
        }
M
Markus Armbruster 已提交
974 975
        error_setg(errp, "PCI: no slot/function available for %s, all in use",
                   name);
976
        return NULL;
B
bellard 已提交
977
    found: ;
978
    } else if (bus->devices[devfn]) {
M
Markus Armbruster 已提交
979 980 981 982
        error_setg(errp, "PCI: slot %d function %d not available for %s,"
                   " in use by %s",
                   PCI_SLOT(devfn), PCI_FUNC(devfn), name,
                   bus->devices[devfn]->name);
983
        return NULL;
C
Cao jin 已提交
984 985 986 987
    } else if (dev->hotplugged &&
               pci_get_function_0(pci_dev)) {
        error_setg(errp, "PCI: slot %d function 0 already ocuppied by %s,"
                   " new func %s cannot be exposed to guest.",
988 989
                   PCI_SLOT(pci_get_function_0(pci_dev)->devfn),
                   pci_get_function_0(pci_dev)->name,
C
Cao jin 已提交
990 991 992
                   name);

       return NULL;
B
bellard 已提交
993
    }
994

995
    pci_dev->devfn = devfn;
P
Peter Xu 已提交
996
    pci_dev->requester_id_cache = pci_req_id_cache_get(pci_dev);
997

998 999 1000 1001 1002
    memory_region_init(&pci_dev->bus_master_container_region, OBJECT(pci_dev),
                       "bus master container", UINT64_MAX);
    address_space_init(&pci_dev->bus_master_as,
                       &pci_dev->bus_master_container_region, pci_dev->name);

1003 1004 1005
    if (qdev_hotplug) {
        pci_init_bus_master(pci_dev);
    }
B
bellard 已提交
1006
    pstrcpy(pci_dev->name, sizeof(pci_dev->name), name);
1007
    pci_dev->irq_state = 0;
I
Isaku Yamahata 已提交
1008
    pci_config_alloc(pci_dev);
1009

1010 1011 1012 1013
    pci_config_set_vendor_id(pci_dev->config, pc->vendor_id);
    pci_config_set_device_id(pci_dev->config, pc->device_id);
    pci_config_set_revision(pci_dev->config, pc->revision);
    pci_config_set_class(pci_dev->config, pc->class_id);
1014

1015 1016
    if (!pc->is_bridge) {
        if (pc->subsystem_vendor_id || pc->subsystem_id) {
1017
            pci_set_word(pci_dev->config + PCI_SUBSYSTEM_VENDOR_ID,
1018
                         pc->subsystem_vendor_id);
1019
            pci_set_word(pci_dev->config + PCI_SUBSYSTEM_ID,
1020
                         pc->subsystem_id);
1021 1022 1023 1024 1025
        } else {
            pci_set_default_subsystem_id(pci_dev);
        }
    } else {
        /* subsystem_vendor_id/subsystem_id are only for header type 0 */
1026 1027
        assert(!pc->subsystem_vendor_id);
        assert(!pc->subsystem_id);
1028
    }
1029
    pci_init_cmask(pci_dev);
1030
    pci_init_wmask(pci_dev);
1031
    pci_init_w1cmask(pci_dev);
1032
    if (pc->is_bridge) {
1033
        pci_init_mask_bridge(pci_dev);
1034
    }
M
Markus Armbruster 已提交
1035 1036 1037
    pci_init_multifunction(bus, pci_dev, &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
1038
        do_pci_unregister_device(pci_dev);
1039 1040
        return NULL;
    }
1041 1042 1043 1044 1045

    if (!config_read)
        config_read = pci_default_read_config;
    if (!config_write)
        config_write = pci_default_write_config;
B
bellard 已提交
1046 1047
    pci_dev->config_read = config_read;
    pci_dev->config_write = config_write;
1048
    bus->devices[devfn] = pci_dev;
J
Juan Quintela 已提交
1049
    pci_dev->version_id = 2; /* Current pci device vmstate version */
B
bellard 已提交
1050 1051 1052
    return pci_dev;
}

1053 1054 1055 1056 1057 1058 1059
static void pci_unregister_io_regions(PCIDevice *pci_dev)
{
    PCIIORegion *r;
    int i;

    for(i = 0; i < PCI_NUM_REGIONS; i++) {
        r = &pci_dev->io_regions[i];
1060
        if (!r->size || r->addr == PCI_BAR_UNMAPPED)
1061
            continue;
1062
        memory_region_del_subregion(r->address_space, r->memory);
1063
    }
A
Alex Williamson 已提交
1064 1065

    pci_unregister_vga(pci_dev);
1066 1067
}

M
Markus Armbruster 已提交
1068
static void pci_qdev_unrealize(DeviceState *dev, Error **errp)
1069
{
1070 1071
    PCIDevice *pci_dev = PCI_DEVICE(dev);
    PCIDeviceClass *pc = PCI_DEVICE_GET_CLASS(pci_dev);
1072 1073

    pci_unregister_io_regions(pci_dev);
1074
    pci_del_option_rom(pci_dev);
1075

1076 1077 1078
    if (pc->exit) {
        pc->exit(pci_dev);
    }
1079

1080
    do_pci_unregister_device(pci_dev);
1081 1082
}

1083 1084
void pci_register_bar(PCIDevice *pci_dev, int region_num,
                      uint8_t type, MemoryRegion *memory)
B
bellard 已提交
1085 1086
{
    PCIIORegion *r;
C
Cao jin 已提交
1087
    uint32_t addr; /* offset in pci config space */
1088
    uint64_t wmask;
A
Avi Kivity 已提交
1089
    pcibus_t size = memory_region_size(memory);
1090

1091 1092
    assert(region_num >= 0);
    assert(region_num < PCI_NUM_REGIONS);
1093 1094
    if (size & (size-1)) {
        fprintf(stderr, "ERROR: PCI region size must be pow2 "
1095
                    "type=0x%x, size=0x%"FMT_PCIBUS"\n", type, size);
1096 1097 1098
        exit(1);
    }

B
bellard 已提交
1099
    r = &pci_dev->io_regions[region_num];
1100
    r->addr = PCI_BAR_UNMAPPED;
B
bellard 已提交
1101 1102
    r->size = size;
    r->type = type;
C
Cao jin 已提交
1103 1104 1105 1106
    r->memory = memory;
    r->address_space = type & PCI_BASE_ADDRESS_SPACE_IO
                        ? pci_dev->bus->address_space_io
                        : pci_dev->bus->address_space_mem;
1107 1108

    wmask = ~(size - 1);
P
pbrook 已提交
1109
    if (region_num == PCI_ROM_SLOT) {
S
Stefan Weil 已提交
1110
        /* ROM enable bit is writable */
1111
        wmask |= PCI_ROM_ADDRESS_ENABLE;
P
pbrook 已提交
1112
    }
C
Cao jin 已提交
1113 1114

    addr = pci_bar(pci_dev, region_num);
1115
    pci_set_long(pci_dev->config + addr, type);
C
Cao jin 已提交
1116

I
Isaku Yamahata 已提交
1117 1118 1119 1120 1121 1122 1123 1124
    if (!(r->type & PCI_BASE_ADDRESS_SPACE_IO) &&
        r->type & PCI_BASE_ADDRESS_MEM_TYPE_64) {
        pci_set_quad(pci_dev->wmask + addr, wmask);
        pci_set_quad(pci_dev->cmask + addr, ~0ULL);
    } else {
        pci_set_long(pci_dev->wmask + addr, wmask & 0xffffffff);
        pci_set_long(pci_dev->cmask + addr, 0xffffffff);
    }
1125 1126
}

A
Alex Williamson 已提交
1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183
static void pci_update_vga(PCIDevice *pci_dev)
{
    uint16_t cmd;

    if (!pci_dev->has_vga) {
        return;
    }

    cmd = pci_get_word(pci_dev->config + PCI_COMMAND);

    memory_region_set_enabled(pci_dev->vga_regions[QEMU_PCI_VGA_MEM],
                              cmd & PCI_COMMAND_MEMORY);
    memory_region_set_enabled(pci_dev->vga_regions[QEMU_PCI_VGA_IO_LO],
                              cmd & PCI_COMMAND_IO);
    memory_region_set_enabled(pci_dev->vga_regions[QEMU_PCI_VGA_IO_HI],
                              cmd & PCI_COMMAND_IO);
}

void pci_register_vga(PCIDevice *pci_dev, MemoryRegion *mem,
                      MemoryRegion *io_lo, MemoryRegion *io_hi)
{
    assert(!pci_dev->has_vga);

    assert(memory_region_size(mem) == QEMU_PCI_VGA_MEM_SIZE);
    pci_dev->vga_regions[QEMU_PCI_VGA_MEM] = mem;
    memory_region_add_subregion_overlap(pci_dev->bus->address_space_mem,
                                        QEMU_PCI_VGA_MEM_BASE, mem, 1);

    assert(memory_region_size(io_lo) == QEMU_PCI_VGA_IO_LO_SIZE);
    pci_dev->vga_regions[QEMU_PCI_VGA_IO_LO] = io_lo;
    memory_region_add_subregion_overlap(pci_dev->bus->address_space_io,
                                        QEMU_PCI_VGA_IO_LO_BASE, io_lo, 1);

    assert(memory_region_size(io_hi) == QEMU_PCI_VGA_IO_HI_SIZE);
    pci_dev->vga_regions[QEMU_PCI_VGA_IO_HI] = io_hi;
    memory_region_add_subregion_overlap(pci_dev->bus->address_space_io,
                                        QEMU_PCI_VGA_IO_HI_BASE, io_hi, 1);
    pci_dev->has_vga = true;

    pci_update_vga(pci_dev);
}

void pci_unregister_vga(PCIDevice *pci_dev)
{
    if (!pci_dev->has_vga) {
        return;
    }

    memory_region_del_subregion(pci_dev->bus->address_space_mem,
                                pci_dev->vga_regions[QEMU_PCI_VGA_MEM]);
    memory_region_del_subregion(pci_dev->bus->address_space_io,
                                pci_dev->vga_regions[QEMU_PCI_VGA_IO_LO]);
    memory_region_del_subregion(pci_dev->bus->address_space_io,
                                pci_dev->vga_regions[QEMU_PCI_VGA_IO_HI]);
    pci_dev->has_vga = false;
}

1184 1185 1186 1187 1188
pcibus_t pci_get_bar_addr(PCIDevice *pci_dev, int region_num)
{
    return pci_dev->io_regions[region_num].addr;
}

1189 1190 1191 1192 1193 1194
static pcibus_t pci_bar_address(PCIDevice *d,
				int reg, uint8_t type, pcibus_t size)
{
    pcibus_t new_addr, last_addr;
    int bar = pci_bar(d, reg);
    uint16_t cmd = pci_get_word(d->config + PCI_COMMAND);
1195 1196 1197 1198
    Object *machine = qdev_get_machine();
    ObjectClass *oc = object_get_class(machine);
    MachineClass *mc = MACHINE_CLASS(oc);
    bool allow_0_address = mc->pci_allow_0_address;
1199 1200 1201 1202 1203 1204 1205

    if (type & PCI_BASE_ADDRESS_SPACE_IO) {
        if (!(cmd & PCI_COMMAND_IO)) {
            return PCI_BAR_UNMAPPED;
        }
        new_addr = pci_get_long(d->config + bar) & ~(size - 1);
        last_addr = new_addr + size - 1;
1206 1207 1208
        /* Check if 32 bit BAR wraps around explicitly.
         * TODO: make priorities correct and remove this work around.
         */
1209 1210
        if (last_addr <= new_addr || last_addr >= UINT32_MAX ||
            (!allow_0_address && new_addr == 0)) {
1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233
            return PCI_BAR_UNMAPPED;
        }
        return new_addr;
    }

    if (!(cmd & PCI_COMMAND_MEMORY)) {
        return PCI_BAR_UNMAPPED;
    }
    if (type & PCI_BASE_ADDRESS_MEM_TYPE_64) {
        new_addr = pci_get_quad(d->config + bar);
    } else {
        new_addr = pci_get_long(d->config + bar);
    }
    /* the ROM slot has a specific enable bit */
    if (reg == PCI_ROM_SLOT && !(new_addr & PCI_ROM_ADDRESS_ENABLE)) {
        return PCI_BAR_UNMAPPED;
    }
    new_addr &= ~(size - 1);
    last_addr = new_addr + size - 1;
    /* NOTE: we do not support wrapping */
    /* XXX: as we cannot support really dynamic
       mappings, we handle specific values as invalid
       mappings. */
1234 1235
    if (last_addr <= new_addr || last_addr == PCI_BAR_UNMAPPED ||
        (!allow_0_address && new_addr == 0)) {
1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253
        return PCI_BAR_UNMAPPED;
    }

    /* Now pcibus_t is 64bit.
     * Check if 32 bit BAR wraps around explicitly.
     * Without this, PC ide doesn't work well.
     * TODO: remove this work around.
     */
    if  (!(type & PCI_BASE_ADDRESS_MEM_TYPE_64) && last_addr >= UINT32_MAX) {
        return PCI_BAR_UNMAPPED;
    }

    /*
     * OS is allowed to set BAR beyond its addressable
     * bits. For example, 32 bit OS can set 64bit bar
     * to >4G. Check it. TODO: we might need to support
     * it in the future for e.g. PAE.
     */
A
Avi Kivity 已提交
1254
    if (last_addr >= HWADDR_MAX) {
1255 1256 1257 1258 1259 1260
        return PCI_BAR_UNMAPPED;
    }

    return new_addr;
}

1261 1262 1263
static void pci_update_mappings(PCIDevice *d)
{
    PCIIORegion *r;
1264
    int i;
1265
    pcibus_t new_addr;
1266

1267
    for(i = 0; i < PCI_NUM_REGIONS; i++) {
1268
        r = &d->io_regions[i];
1269 1270

        /* this region isn't registered */
1271
        if (!r->size)
1272 1273
            continue;

1274
        new_addr = pci_bar_address(d, i, r->type, r->size);
1275 1276

        /* This bar isn't changed */
1277
        if (new_addr == r->addr)
1278 1279 1280 1281
            continue;

        /* now do the real mapping */
        if (r->addr != PCI_BAR_UNMAPPED) {
D
Don Koch 已提交
1282 1283
            trace_pci_update_mappings_del(d, pci_bus_num(d->bus),
                                          PCI_SLOT(d->devfn),
1284
                                          PCI_FUNC(d->devfn),
D
Don Koch 已提交
1285
                                          i, r->addr, r->size);
1286
            memory_region_del_subregion(r->address_space, r->memory);
1287
        }
1288 1289
        r->addr = new_addr;
        if (r->addr != PCI_BAR_UNMAPPED) {
D
Don Koch 已提交
1290 1291
            trace_pci_update_mappings_add(d, pci_bus_num(d->bus),
                                          PCI_SLOT(d->devfn),
1292
                                          PCI_FUNC(d->devfn),
D
Don Koch 已提交
1293
                                          i, r->addr, r->size);
1294 1295
            memory_region_add_subregion_overlap(r->address_space,
                                                r->addr, r->memory, 1);
1296
        }
1297
    }
A
Alex Williamson 已提交
1298 1299

    pci_update_vga(d);
1300 1301
}

1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320
static inline int pci_irq_disabled(PCIDevice *d)
{
    return pci_get_word(d->config + PCI_COMMAND) & PCI_COMMAND_INTX_DISABLE;
}

/* Called after interrupt disabled field update in config space,
 * assert/deassert interrupts if necessary.
 * Gets original interrupt disable bit value (before update). */
static void pci_update_irq_disabled(PCIDevice *d, int was_irq_disabled)
{
    int i, disabled = pci_irq_disabled(d);
    if (disabled == was_irq_disabled)
        return;
    for (i = 0; i < PCI_NUM_PINS; ++i) {
        int state = pci_irq_state(d, i);
        pci_change_irq_level(d, i, disabled ? -state : state);
    }
}

1321
uint32_t pci_default_read_config(PCIDevice *d,
1322
                                 uint32_t address, int len)
B
bellard 已提交
1323
{
1324
    uint32_t val = 0;
J
Jan Kiszka 已提交
1325

1326 1327
    memcpy(&val, d->config + address, len);
    return le32_to_cpu(val);
1328 1329
}

1330
void pci_default_write_config(PCIDevice *d, uint32_t addr, uint32_t val_in, int l)
1331
{
1332
    int i, was_irq_disabled = pci_irq_disabled(d);
1333
    uint32_t val = val_in;
1334

J
Jan Kiszka 已提交
1335
    for (i = 0; i < l; val >>= 8, ++i) {
1336
        uint8_t wmask = d->wmask[addr + i];
1337 1338
        uint8_t w1cmask = d->w1cmask[addr + i];
        assert(!(wmask & w1cmask));
1339
        d->config[addr + i] = (d->config[addr + i] & ~wmask) | (val & wmask);
1340
        d->config[addr + i] &= ~(val & w1cmask); /* W1C: Write 1 to Clear */
1341
    }
I
Isaku Yamahata 已提交
1342
    if (ranges_overlap(addr, l, PCI_BASE_ADDRESS_0, 24) ||
1343 1344
        ranges_overlap(addr, l, PCI_ROM_ADDRESS, 4) ||
        ranges_overlap(addr, l, PCI_ROM_ADDRESS1, 4) ||
I
Isaku Yamahata 已提交
1345
        range_covers_byte(addr, l, PCI_COMMAND))
1346
        pci_update_mappings(d);
1347

A
Avi Kivity 已提交
1348
    if (range_covers_byte(addr, l, PCI_COMMAND)) {
1349
        pci_update_irq_disabled(d, was_irq_disabled);
A
Avi Kivity 已提交
1350 1351 1352 1353
        memory_region_set_enabled(&d->bus_master_enable_region,
                                  pci_get_word(d->config + PCI_COMMAND)
                                    & PCI_COMMAND_MASTER);
    }
1354

1355 1356
    msi_write_config(d, addr, val_in, l);
    msix_write_config(d, addr, val_in, l);
B
bellard 已提交
1357 1358
}

P
pbrook 已提交
1359 1360
/***********************************************************/
/* generic PCI irq support */
1361

P
pbrook 已提交
1362
/* 0 <= irq_num <= 3. level must be 0 or 1 */
1363
static void pci_irq_handler(void *opaque, int irq_num, int level)
B
bellard 已提交
1364
{
1365
    PCIDevice *pci_dev = opaque;
P
pbrook 已提交
1366
    int change;
1367

1368
    change = level - pci_irq_state(pci_dev, irq_num);
P
pbrook 已提交
1369 1370
    if (!change)
        return;
1371

1372
    pci_set_irq_state(pci_dev, irq_num, level);
1373
    pci_update_irq_status(pci_dev);
1374 1375
    if (pci_irq_disabled(pci_dev))
        return;
1376
    pci_change_irq_level(pci_dev, irq_num, change);
B
bellard 已提交
1377 1378
}

1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396
static inline int pci_intx(PCIDevice *pci_dev)
{
    return pci_get_byte(pci_dev->config + PCI_INTERRUPT_PIN) - 1;
}

qemu_irq pci_allocate_irq(PCIDevice *pci_dev)
{
    int intx = pci_intx(pci_dev);

    return qemu_allocate_irq(pci_irq_handler, pci_dev, intx);
}

void pci_set_irq(PCIDevice *pci_dev, int level)
{
    int intx = pci_intx(pci_dev);
    pci_irq_handler(pci_dev, intx, level);
}

1397 1398 1399
/* Special hooks used by device assignment */
void pci_bus_set_route_irq_fn(PCIBus *bus, pci_route_irq_fn route_intx_to_irq)
{
1400
    assert(pci_bus_is_root(bus));
1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412
    bus->route_intx_to_irq = route_intx_to_irq;
}

PCIINTxRoute pci_device_route_intx_to_irq(PCIDevice *dev, int pin)
{
    PCIBus *bus;

    do {
         bus = dev->bus;
         pin = bus->map_irq(dev, pin);
         dev = bus->parent_dev;
    } while (dev);
1413 1414

    if (!bus->route_intx_to_irq) {
1415
        error_report("PCI: Bug - unimplemented PCI INTx routing (%s)",
1416 1417 1418 1419
                     object_get_typename(OBJECT(bus->qbus.parent)));
        return (PCIINTxRoute) { PCI_INTX_DISABLED, -1 };
    }

1420
    return bus->route_intx_to_irq(bus->irq_opaque, pin);
J
Jan Kiszka 已提交
1421 1422
}

1423 1424 1425 1426 1427
bool pci_intx_route_changed(PCIINTxRoute *old, PCIINTxRoute *new)
{
    return old->mode != new->mode || old->irq != new->irq;
}

J
Jan Kiszka 已提交
1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438
void pci_bus_fire_intx_routing_notifier(PCIBus *bus)
{
    PCIDevice *dev;
    PCIBus *sec;
    int i;

    for (i = 0; i < ARRAY_SIZE(bus->devices); ++i) {
        dev = bus->devices[i];
        if (dev && dev->intx_routing_notifier) {
            dev->intx_routing_notifier(dev);
        }
1439 1440 1441 1442
    }

    QLIST_FOREACH(sec, &bus->child, sibling) {
        pci_bus_fire_intx_routing_notifier(sec);
J
Jan Kiszka 已提交
1443 1444 1445 1446 1447 1448 1449
    }
}

void pci_device_set_intx_routing_notifier(PCIDevice *dev,
                                          PCIINTxRoutingNotifier notifier)
{
    dev->intx_routing_notifier = notifier;
B
bellard 已提交
1450 1451
}

1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469
/*
 * PCI-to-PCI bridge specification
 * 9.1: Interrupt routing. Table 9-1
 *
 * the PCI Express Base Specification, Revision 2.1
 * 2.2.8.1: INTx interrutp signaling - Rules
 *          the Implementation Note
 *          Table 2-20
 */
/*
 * 0 <= pin <= 3 0 = INTA, 1 = INTB, 2 = INTC, 3 = INTD
 * 0-origin unlike PCI interrupt pin register.
 */
int pci_swizzle_map_irq_fn(PCIDevice *pci_dev, int pin)
{
    return (pin + PCI_SLOT(pci_dev->devfn)) % PCI_NUM_PINS;
}

P
pbrook 已提交
1470 1471
/***********************************************************/
/* monitor info on PCI */
1472

1473 1474 1475
typedef struct {
    uint16_t class;
    const char *desc;
1476 1477
    const char *fw_name;
    uint16_t fw_ign_bits;
1478 1479
} pci_class_desc;

1480
static const pci_class_desc pci_class_descriptions[] =
1481
{
1482 1483 1484 1485 1486 1487
    { 0x0001, "VGA controller", "display"},
    { 0x0100, "SCSI controller", "scsi"},
    { 0x0101, "IDE controller", "ide"},
    { 0x0102, "Floppy controller", "fdc"},
    { 0x0103, "IPI controller", "ipi"},
    { 0x0104, "RAID controller", "raid"},
T
ths 已提交
1488 1489 1490
    { 0x0106, "SATA controller"},
    { 0x0107, "SAS controller"},
    { 0x0180, "Storage controller"},
1491 1492 1493 1494
    { 0x0200, "Ethernet controller", "ethernet"},
    { 0x0201, "Token Ring controller", "token-ring"},
    { 0x0202, "FDDI controller", "fddi"},
    { 0x0203, "ATM controller", "atm"},
T
ths 已提交
1495
    { 0x0280, "Network controller"},
1496
    { 0x0300, "VGA controller", "display", 0x00ff},
T
ths 已提交
1497 1498 1499
    { 0x0301, "XGA controller"},
    { 0x0302, "3D controller"},
    { 0x0380, "Display controller"},
1500 1501
    { 0x0400, "Video controller", "video"},
    { 0x0401, "Audio controller", "sound"},
T
ths 已提交
1502
    { 0x0402, "Phone"},
1503
    { 0x0403, "Audio controller", "sound"},
T
ths 已提交
1504
    { 0x0480, "Multimedia controller"},
1505 1506
    { 0x0500, "RAM controller", "memory"},
    { 0x0501, "Flash controller", "flash"},
T
ths 已提交
1507
    { 0x0580, "Memory controller"},
1508 1509 1510 1511
    { 0x0600, "Host bridge", "host"},
    { 0x0601, "ISA bridge", "isa"},
    { 0x0602, "EISA bridge", "eisa"},
    { 0x0603, "MC bridge", "mca"},
G
Gerd Hoffmann 已提交
1512
    { 0x0604, "PCI bridge", "pci-bridge"},
1513 1514 1515
    { 0x0605, "PCMCIA bridge", "pcmcia"},
    { 0x0606, "NUBUS bridge", "nubus"},
    { 0x0607, "CARDBUS bridge", "cardbus"},
T
ths 已提交
1516 1517
    { 0x0608, "RACEWAY bridge"},
    { 0x0680, "Bridge"},
1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533
    { 0x0700, "Serial port", "serial"},
    { 0x0701, "Parallel port", "parallel"},
    { 0x0800, "Interrupt controller", "interrupt-controller"},
    { 0x0801, "DMA controller", "dma-controller"},
    { 0x0802, "Timer", "timer"},
    { 0x0803, "RTC", "rtc"},
    { 0x0900, "Keyboard", "keyboard"},
    { 0x0901, "Pen", "pen"},
    { 0x0902, "Mouse", "mouse"},
    { 0x0A00, "Dock station", "dock", 0x00ff},
    { 0x0B00, "i386 cpu", "cpu", 0x00ff},
    { 0x0c00, "Fireware contorller", "fireware"},
    { 0x0c01, "Access bus controller", "access-bus"},
    { 0x0c02, "SSA controller", "ssa"},
    { 0x0c03, "USB controller", "usb"},
    { 0x0c04, "Fibre channel controller", "fibre-channel"},
J
Jan Kiszka 已提交
1534
    { 0x0c05, "SMBus"},
1535 1536 1537
    { 0, NULL}
};

1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565
static void pci_for_each_device_under_bus_reverse(PCIBus *bus,
                                                  void (*fn)(PCIBus *b,
                                                             PCIDevice *d,
                                                             void *opaque),
                                                  void *opaque)
{
    PCIDevice *d;
    int devfn;

    for (devfn = 0; devfn < ARRAY_SIZE(bus->devices); devfn++) {
        d = bus->devices[ARRAY_SIZE(bus->devices) - 1 - devfn];
        if (d) {
            fn(bus, d, opaque);
        }
    }
}

void pci_for_each_device_reverse(PCIBus *bus, int bus_num,
                         void (*fn)(PCIBus *b, PCIDevice *d, void *opaque),
                         void *opaque)
{
    bus = pci_find_bus_nr(bus, bus_num);

    if (bus) {
        pci_for_each_device_under_bus_reverse(bus, fn, opaque);
    }
}

1566
static void pci_for_each_device_under_bus(PCIBus *bus,
1567 1568 1569
                                          void (*fn)(PCIBus *b, PCIDevice *d,
                                                     void *opaque),
                                          void *opaque)
1570
{
1571 1572
    PCIDevice *d;
    int devfn;
1573

1574 1575 1576
    for(devfn = 0; devfn < ARRAY_SIZE(bus->devices); devfn++) {
        d = bus->devices[devfn];
        if (d) {
1577
            fn(bus, d, opaque);
1578 1579 1580 1581 1582
        }
    }
}

void pci_for_each_device(PCIBus *bus, int bus_num,
1583 1584
                         void (*fn)(PCIBus *b, PCIDevice *d, void *opaque),
                         void *opaque)
1585
{
1586
    bus = pci_find_bus_nr(bus, bus_num);
1587 1588

    if (bus) {
1589
        pci_for_each_device_under_bus(bus, fn, opaque);
1590 1591 1592
    }
}

L
Luiz Capitulino 已提交
1593
static const pci_class_desc *get_class_desc(int class)
1594
{
L
Luiz Capitulino 已提交
1595
    const pci_class_desc *desc;
1596

L
Luiz Capitulino 已提交
1597 1598 1599
    desc = pci_class_descriptions;
    while (desc->desc && class != desc->class) {
        desc++;
1600
    }
1601

L
Luiz Capitulino 已提交
1602 1603
    return desc;
}
I
Isaku Yamahata 已提交
1604

L
Luiz Capitulino 已提交
1605
static PciDeviceInfoList *qmp_query_pci_devices(PCIBus *bus, int bus_num);
1606

L
Luiz Capitulino 已提交
1607 1608 1609 1610
static PciMemoryRegionList *qmp_query_pci_regions(const PCIDevice *dev)
{
    PciMemoryRegionList *head = NULL, *cur_item = NULL;
    int i;
1611

L
Luiz Capitulino 已提交
1612 1613 1614 1615 1616 1617
    for (i = 0; i < PCI_NUM_REGIONS; i++) {
        const PCIIORegion *r = &dev->io_regions[i];
        PciMemoryRegionList *region;

        if (!r->size) {
            continue;
P
pbrook 已提交
1618
        }
1619

L
Luiz Capitulino 已提交
1620 1621
        region = g_malloc0(sizeof(*region));
        region->value = g_malloc0(sizeof(*region->value));
1622

L
Luiz Capitulino 已提交
1623 1624 1625 1626 1627 1628 1629 1630
        if (r->type & PCI_BASE_ADDRESS_SPACE_IO) {
            region->value->type = g_strdup("io");
        } else {
            region->value->type = g_strdup("memory");
            region->value->has_prefetch = true;
            region->value->prefetch = !!(r->type & PCI_BASE_ADDRESS_MEM_PREFETCH);
            region->value->has_mem_type_64 = true;
            region->value->mem_type_64 = !!(r->type & PCI_BASE_ADDRESS_MEM_TYPE_64);
1631
        }
1632

L
Luiz Capitulino 已提交
1633 1634 1635
        region->value->bar = i;
        region->value->address = r->addr;
        region->value->size = r->size;
1636

L
Luiz Capitulino 已提交
1637 1638 1639 1640 1641 1642
        /* XXX: waiting for the qapi to support GSList */
        if (!cur_item) {
            head = cur_item = region;
        } else {
            cur_item->next = region;
            cur_item = region;
1643
        }
P
pbrook 已提交
1644
    }
1645

L
Luiz Capitulino 已提交
1646
    return head;
1647 1648
}

L
Luiz Capitulino 已提交
1649 1650
static PciBridgeInfo *qmp_query_pci_bridge(PCIDevice *dev, PCIBus *bus,
                                           int bus_num)
1651
{
L
Luiz Capitulino 已提交
1652
    PciBridgeInfo *info;
1653
    PciMemoryRange *range;
1654

1655
    info = g_new0(PciBridgeInfo, 1);
1656

1657 1658 1659 1660
    info->bus = g_new0(PciBusInfo, 1);
    info->bus->number = dev->config[PCI_PRIMARY_BUS];
    info->bus->secondary = dev->config[PCI_SECONDARY_BUS];
    info->bus->subordinate = dev->config[PCI_SUBORDINATE_BUS];
1661

1662 1663 1664
    range = info->bus->io_range = g_new0(PciMemoryRange, 1);
    range->base = pci_bridge_get_base(dev, PCI_BASE_ADDRESS_SPACE_IO);
    range->limit = pci_bridge_get_limit(dev, PCI_BASE_ADDRESS_SPACE_IO);
1665

1666 1667 1668
    range = info->bus->memory_range = g_new0(PciMemoryRange, 1);
    range->base = pci_bridge_get_base(dev, PCI_BASE_ADDRESS_SPACE_MEMORY);
    range->limit = pci_bridge_get_limit(dev, PCI_BASE_ADDRESS_SPACE_MEMORY);
1669

1670 1671 1672
    range = info->bus->prefetchable_range = g_new0(PciMemoryRange, 1);
    range->base = pci_bridge_get_base(dev, PCI_BASE_ADDRESS_MEM_PREFETCH);
    range->limit = pci_bridge_get_limit(dev, PCI_BASE_ADDRESS_MEM_PREFETCH);
1673

L
Luiz Capitulino 已提交
1674
    if (dev->config[PCI_SECONDARY_BUS] != 0) {
1675
        PCIBus *child_bus = pci_find_bus_nr(bus, dev->config[PCI_SECONDARY_BUS]);
L
Luiz Capitulino 已提交
1676 1677 1678 1679
        if (child_bus) {
            info->has_devices = true;
            info->devices = qmp_query_pci_devices(child_bus, dev->config[PCI_SECONDARY_BUS]);
        }
1680 1681
    }

L
Luiz Capitulino 已提交
1682
    return info;
1683 1684
}

L
Luiz Capitulino 已提交
1685 1686
static PciDeviceInfo *qmp_query_pci_device(PCIDevice *dev, PCIBus *bus,
                                           int bus_num)
1687
{
L
Luiz Capitulino 已提交
1688 1689
    const pci_class_desc *desc;
    PciDeviceInfo *info;
1690
    uint8_t type;
L
Luiz Capitulino 已提交
1691
    int class;
1692

1693
    info = g_new0(PciDeviceInfo, 1);
L
Luiz Capitulino 已提交
1694 1695 1696 1697
    info->bus = bus_num;
    info->slot = PCI_SLOT(dev->devfn);
    info->function = PCI_FUNC(dev->devfn);

1698
    info->class_info = g_new0(PciDeviceClass, 1);
L
Luiz Capitulino 已提交
1699
    class = pci_get_word(dev->config + PCI_CLASS_DEVICE);
1700
    info->class_info->q_class = class;
L
Luiz Capitulino 已提交
1701 1702
    desc = get_class_desc(class);
    if (desc->desc) {
1703 1704
        info->class_info->has_desc = true;
        info->class_info->desc = g_strdup(desc->desc);
L
Luiz Capitulino 已提交
1705 1706
    }

1707 1708 1709
    info->id = g_new0(PciDeviceId, 1);
    info->id->vendor = pci_get_word(dev->config + PCI_VENDOR_ID);
    info->id->device = pci_get_word(dev->config + PCI_DEVICE_ID);
L
Luiz Capitulino 已提交
1710 1711
    info->regions = qmp_query_pci_regions(dev);
    info->qdev_id = g_strdup(dev->qdev.id ? dev->qdev.id : "");
1712 1713

    if (dev->config[PCI_INTERRUPT_PIN] != 0) {
L
Luiz Capitulino 已提交
1714 1715
        info->has_irq = true;
        info->irq = dev->config[PCI_INTERRUPT_LINE];
1716 1717
    }

1718 1719
    type = dev->config[PCI_HEADER_TYPE] & ~PCI_HEADER_TYPE_MULTI_FUNCTION;
    if (type == PCI_HEADER_TYPE_BRIDGE) {
L
Luiz Capitulino 已提交
1720 1721
        info->has_pci_bridge = true;
        info->pci_bridge = qmp_query_pci_bridge(dev, bus, bus_num);
1722 1723
    }

L
Luiz Capitulino 已提交
1724
    return info;
1725 1726
}

L
Luiz Capitulino 已提交
1727
static PciDeviceInfoList *qmp_query_pci_devices(PCIBus *bus, int bus_num)
1728
{
L
Luiz Capitulino 已提交
1729
    PciDeviceInfoList *info, *head = NULL, *cur_item = NULL;
1730
    PCIDevice *dev;
L
Luiz Capitulino 已提交
1731
    int devfn;
1732 1733 1734 1735

    for (devfn = 0; devfn < ARRAY_SIZE(bus->devices); devfn++) {
        dev = bus->devices[devfn];
        if (dev) {
L
Luiz Capitulino 已提交
1736 1737 1738 1739 1740 1741 1742 1743 1744 1745
            info = g_malloc0(sizeof(*info));
            info->value = qmp_query_pci_device(dev, bus, bus_num);

            /* XXX: waiting for the qapi to support GSList */
            if (!cur_item) {
                head = cur_item = info;
            } else {
                cur_item->next = info;
                cur_item = info;
            }
1746
        }
1747
    }
1748

L
Luiz Capitulino 已提交
1749
    return head;
1750 1751
}

L
Luiz Capitulino 已提交
1752
static PciInfo *qmp_query_pci_bus(PCIBus *bus, int bus_num)
1753
{
L
Luiz Capitulino 已提交
1754 1755
    PciInfo *info = NULL;

1756
    bus = pci_find_bus_nr(bus, bus_num);
P
pbrook 已提交
1757
    if (bus) {
L
Luiz Capitulino 已提交
1758 1759 1760
        info = g_malloc0(sizeof(*info));
        info->bus = bus_num;
        info->devices = qmp_query_pci_devices(bus, bus_num);
B
bellard 已提交
1761
    }
1762

L
Luiz Capitulino 已提交
1763
    return info;
B
bellard 已提交
1764 1765
}

L
Luiz Capitulino 已提交
1766
PciInfoList *qmp_query_pci(Error **errp)
B
bellard 已提交
1767
{
L
Luiz Capitulino 已提交
1768
    PciInfoList *info, *head = NULL, *cur_item = NULL;
1769
    PCIHostState *host_bridge;
1770

1771
    QLIST_FOREACH(host_bridge, &pci_host_bridges, next) {
L
Luiz Capitulino 已提交
1772
        info = g_malloc0(sizeof(*info));
1773 1774
        info->value = qmp_query_pci_bus(host_bridge->bus,
                                        pci_bus_num(host_bridge->bus));
L
Luiz Capitulino 已提交
1775 1776 1777 1778 1779 1780 1781

        /* XXX: waiting for the qapi to support GSList */
        if (!cur_item) {
            head = cur_item = info;
        } else {
            cur_item->next = info;
            cur_item = info;
1782
        }
1783
    }
1784

L
Luiz Capitulino 已提交
1785
    return head;
B
bellard 已提交
1786
}
1787

1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799
static const char * const pci_nic_models[] = {
    "ne2k_pci",
    "i82551",
    "i82557b",
    "i82559er",
    "rtl8139",
    "e1000",
    "pcnet",
    "virtio",
    NULL
};

P
Paul Brook 已提交
1800 1801 1802 1803 1804 1805 1806 1807
static const char * const pci_nic_names[] = {
    "ne2k_pci",
    "i82551",
    "i82557b",
    "i82559er",
    "rtl8139",
    "e1000",
    "pcnet",
P
Paul Brook 已提交
1808
    "virtio-net-pci",
1809 1810 1811
    NULL
};

1812
/* Initialize a PCI NIC.  */
1813
PCIDevice *pci_nic_init_nofail(NICInfo *nd, PCIBus *rootbus,
1814
                               const char *default_model,
1815
                               const char *default_devaddr)
1816
{
1817
    const char *devaddr = nd->devaddr ? nd->devaddr : default_devaddr;
1818
    PCIBus *bus;
1819
    PCIDevice *pci_dev;
P
Paul Brook 已提交
1820
    DeviceState *dev;
1821
    int devfn;
1822 1823
    int i;

1824 1825 1826 1827
    if (qemu_show_nic_models(nd->model, pci_nic_models)) {
        exit(0);
    }

1828
    i = qemu_find_nic_model(nd, pci_nic_models, default_model);
1829 1830 1831
    if (i < 0) {
        exit(1);
    }
1832

1833
    bus = pci_get_bus_devfn(&devfn, rootbus, devaddr);
1834
    if (!bus) {
1835 1836
        error_report("Invalid PCI device address %s for device %s",
                     devaddr, pci_nic_names[i]);
1837
        exit(1);
1838 1839
    }

1840
    pci_dev = pci_create(bus, devfn, pci_nic_names[i]);
1841
    dev = &pci_dev->qdev;
G
Gerd Hoffmann 已提交
1842
    qdev_set_nic_properties(dev, nd);
1843
    qdev_init_nofail(dev);
1844 1845

    return pci_dev;
1846 1847
}

1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858
PCIDevice *pci_vga_init(PCIBus *bus)
{
    switch (vga_interface_type) {
    case VGA_CIRRUS:
        return pci_create_simple(bus, -1, "cirrus-vga");
    case VGA_QXL:
        return pci_create_simple(bus, -1, "qxl-vga");
    case VGA_STD:
        return pci_create_simple(bus, -1, "VGA");
    case VGA_VMWARE:
        return pci_create_simple(bus, -1, "vmware-svga");
1859 1860
    case VGA_VIRTIO:
        return pci_create_simple(bus, -1, "virtio-vga");
1861 1862 1863 1864 1865 1866 1867
    case VGA_NONE:
    default: /* Other non-PCI types. Checking for unsupported types is already
                done in vl.c. */
        return NULL;
    }
}

1868 1869 1870 1871 1872 1873
/* Whether a given bus number is in range of the secondary
 * bus of the given bridge device. */
static bool pci_secondary_bus_in_range(PCIDevice *dev, int bus_num)
{
    return !(pci_get_word(dev->config + PCI_BRIDGE_CONTROL) &
             PCI_BRIDGE_CTL_BUS_RESET) /* Don't walk the bus if it's reset. */ &&
1874
        dev->config[PCI_SECONDARY_BUS] <= bus_num &&
1875 1876 1877
        bus_num <= dev->config[PCI_SUBORDINATE_BUS];
}

1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895
/* Whether a given bus number is in a range of a root bus */
static bool pci_root_bus_in_range(PCIBus *bus, int bus_num)
{
    int i;

    for (i = 0; i < ARRAY_SIZE(bus->devices); ++i) {
        PCIDevice *dev = bus->devices[i];

        if (dev && PCI_DEVICE_GET_CLASS(dev)->is_bridge) {
            if (pci_secondary_bus_in_range(dev, bus_num)) {
                return true;
            }
        }
    }

    return false;
}

1896
static PCIBus *pci_find_bus_nr(PCIBus *bus, int bus_num)
1897
{
I
Isaku Yamahata 已提交
1898
    PCIBus *sec;
1899

I
Isaku Yamahata 已提交
1900
    if (!bus) {
1901
        return NULL;
I
Isaku Yamahata 已提交
1902
    }
1903

1904 1905 1906 1907
    if (pci_bus_num(bus) == bus_num) {
        return bus;
    }

1908
    /* Consider all bus numbers in range for the host pci bridge. */
1909
    if (!pci_bus_is_root(bus) &&
1910 1911 1912 1913
        !pci_secondary_bus_in_range(bus->parent_dev, bus_num)) {
        return NULL;
    }

1914
    /* try child bus */
1915 1916
    for (; bus; bus = sec) {
        QLIST_FOREACH(sec, &bus->child, sibling) {
1917
            if (pci_bus_num(sec) == bus_num) {
1918 1919
                return sec;
            }
1920 1921 1922 1923 1924 1925 1926 1927 1928
            /* PXB buses assumed to be children of bus 0 */
            if (pci_bus_is_root(sec)) {
                if (pci_root_bus_in_range(sec, bus_num)) {
                    break;
                }
            } else {
                if (pci_secondary_bus_in_range(sec->parent_dev, bus_num)) {
                    break;
                }
B
Blue Swirl 已提交
1929
            }
1930 1931 1932 1933
        }
    }

    return NULL;
1934 1935
}

1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963
void pci_for_each_bus_depth_first(PCIBus *bus,
                                  void *(*begin)(PCIBus *bus, void *parent_state),
                                  void (*end)(PCIBus *bus, void *state),
                                  void *parent_state)
{
    PCIBus *sec;
    void *state;

    if (!bus) {
        return;
    }

    if (begin) {
        state = begin(bus, parent_state);
    } else {
        state = parent_state;
    }

    QLIST_FOREACH(sec, &bus->child, sibling) {
        pci_for_each_bus_depth_first(sec, begin, end, state);
    }

    if (end) {
        end(bus, state);
    }
}


1964
PCIDevice *pci_find_device(PCIBus *bus, int bus_num, uint8_t devfn)
1965
{
1966
    bus = pci_find_bus_nr(bus, bus_num);
1967 1968 1969 1970

    if (!bus)
        return NULL;

1971
    return bus->devices[devfn];
1972 1973
}

M
Markus Armbruster 已提交
1974
static void pci_qdev_realize(DeviceState *qdev, Error **errp)
P
Paul Brook 已提交
1975 1976
{
    PCIDevice *pci_dev = (PCIDevice *)qdev;
1977
    PCIDeviceClass *pc = PCI_DEVICE_GET_CLASS(pci_dev);
M
Markus Armbruster 已提交
1978
    Error *local_err = NULL;
P
Paul Brook 已提交
1979
    PCIBus *bus;
1980
    bool is_default_rom;
P
Paul Brook 已提交
1981

I
Isaku Yamahata 已提交
1982
    /* initialize cap_present for pci_is_express() and pci_config_size() */
1983
    if (pc->is_express) {
I
Isaku Yamahata 已提交
1984 1985 1986
        pci_dev->cap_present |= QEMU_PCI_CAP_EXPRESS;
    }

A
Andreas Färber 已提交
1987
    bus = PCI_BUS(qdev_get_parent_bus(qdev));
A
Anthony Liguori 已提交
1988 1989
    pci_dev = do_pci_register_device(pci_dev, bus,
                                     object_get_typename(OBJECT(qdev)),
M
Markus Armbruster 已提交
1990
                                     pci_dev->devfn, errp);
1991
    if (pci_dev == NULL)
M
Markus Armbruster 已提交
1992
        return;
1993

1994 1995 1996 1997
    if (pc->realize) {
        pc->realize(pci_dev, &local_err);
        if (local_err) {
            error_propagate(errp, local_err);
1998
            do_pci_unregister_device(pci_dev);
M
Markus Armbruster 已提交
1999
            return;
2000
        }
2001
    }
2002 2003

    /* rom loading */
2004
    is_default_rom = false;
2005 2006
    if (pci_dev->romfile == NULL && pc->romfile != NULL) {
        pci_dev->romfile = g_strdup(pc->romfile);
2007 2008
        is_default_rom = true;
    }
2009

M
Markus Armbruster 已提交
2010 2011 2012 2013 2014
    pci_add_option_rom(pci_dev, is_default_rom, &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
        pci_qdev_unrealize(DEVICE(pci_dev), NULL);
        return;
2015
    }
G
Gerd Hoffmann 已提交
2016 2017
}

2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029
static void pci_default_realize(PCIDevice *dev, Error **errp)
{
    PCIDeviceClass *pc = PCI_DEVICE_GET_CLASS(dev);

    if (pc->init) {
        if (pc->init(dev) < 0) {
            error_setg(errp, "Device initialization failed");
            return;
        }
    }
}

2030 2031
PCIDevice *pci_create_multifunction(PCIBus *bus, int devfn, bool multifunction,
                                    const char *name)
P
Paul Brook 已提交
2032 2033 2034
{
    DeviceState *dev;

P
Paul Brook 已提交
2035
    dev = qdev_create(&bus->qbus, name);
2036
    qdev_prop_set_int32(dev, "addr", devfn);
2037
    qdev_prop_set_bit(dev, "multifunction", multifunction);
2038
    return PCI_DEVICE(dev);
2039
}
P
Paul Brook 已提交
2040

2041 2042 2043
PCIDevice *pci_create_simple_multifunction(PCIBus *bus, int devfn,
                                           bool multifunction,
                                           const char *name)
2044
{
2045
    PCIDevice *dev = pci_create_multifunction(bus, devfn, multifunction, name);
M
Markus Armbruster 已提交
2046
    qdev_init_nofail(&dev->qdev);
2047
    return dev;
P
Paul Brook 已提交
2048
}
2049

2050 2051 2052 2053 2054 2055 2056 2057 2058 2059
PCIDevice *pci_create(PCIBus *bus, int devfn, const char *name)
{
    return pci_create_multifunction(bus, devfn, false, name);
}

PCIDevice *pci_create_simple(PCIBus *bus, int devfn, const char *name)
{
    return pci_create_simple_multifunction(bus, devfn, false, name);
}

2060
static uint8_t pci_find_space(PCIDevice *pdev, uint8_t size)
2061 2062 2063
{
    int offset = PCI_CONFIG_HEADER_SIZE;
    int i;
2064
    for (i = PCI_CONFIG_HEADER_SIZE; i < PCI_CONFIG_SPACE_SIZE; ++i) {
2065 2066 2067 2068
        if (pdev->used[i])
            offset = i + 1;
        else if (i - offset + 1 == size)
            return offset;
2069
    }
2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090
    return 0;
}

static uint8_t pci_find_capability_list(PCIDevice *pdev, uint8_t cap_id,
                                        uint8_t *prev_p)
{
    uint8_t next, prev;

    if (!(pdev->config[PCI_STATUS] & PCI_STATUS_CAP_LIST))
        return 0;

    for (prev = PCI_CAPABILITY_LIST; (next = pdev->config[prev]);
         prev = next + PCI_CAP_LIST_NEXT)
        if (pdev->config[next + PCI_CAP_LIST_ID] == cap_id)
            break;

    if (prev_p)
        *prev_p = prev;
    return next;
}

2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109
static uint8_t pci_find_capability_at_offset(PCIDevice *pdev, uint8_t offset)
{
    uint8_t next, prev, found = 0;

    if (!(pdev->used[offset])) {
        return 0;
    }

    assert(pdev->config[PCI_STATUS] & PCI_STATUS_CAP_LIST);

    for (prev = PCI_CAPABILITY_LIST; (next = pdev->config[prev]);
         prev = next + PCI_CAP_LIST_NEXT) {
        if (next <= offset && next > found) {
            found = next;
        }
    }
    return found;
}

2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165
/* Patch the PCI vendor and device ids in a PCI rom image if necessary.
   This is needed for an option rom which is used for more than one device. */
static void pci_patch_ids(PCIDevice *pdev, uint8_t *ptr, int size)
{
    uint16_t vendor_id;
    uint16_t device_id;
    uint16_t rom_vendor_id;
    uint16_t rom_device_id;
    uint16_t rom_magic;
    uint16_t pcir_offset;
    uint8_t checksum;

    /* Words in rom data are little endian (like in PCI configuration),
       so they can be read / written with pci_get_word / pci_set_word. */

    /* Only a valid rom will be patched. */
    rom_magic = pci_get_word(ptr);
    if (rom_magic != 0xaa55) {
        PCI_DPRINTF("Bad ROM magic %04x\n", rom_magic);
        return;
    }
    pcir_offset = pci_get_word(ptr + 0x18);
    if (pcir_offset + 8 >= size || memcmp(ptr + pcir_offset, "PCIR", 4)) {
        PCI_DPRINTF("Bad PCIR offset 0x%x or signature\n", pcir_offset);
        return;
    }

    vendor_id = pci_get_word(pdev->config + PCI_VENDOR_ID);
    device_id = pci_get_word(pdev->config + PCI_DEVICE_ID);
    rom_vendor_id = pci_get_word(ptr + pcir_offset + 4);
    rom_device_id = pci_get_word(ptr + pcir_offset + 6);

    PCI_DPRINTF("%s: ROM id %04x%04x / PCI id %04x%04x\n", pdev->romfile,
                vendor_id, device_id, rom_vendor_id, rom_device_id);

    checksum = ptr[6];

    if (vendor_id != rom_vendor_id) {
        /* Patch vendor id and checksum (at offset 6 for etherboot roms). */
        checksum += (uint8_t)rom_vendor_id + (uint8_t)(rom_vendor_id >> 8);
        checksum -= (uint8_t)vendor_id + (uint8_t)(vendor_id >> 8);
        PCI_DPRINTF("ROM checksum %02x / %02x\n", ptr[6], checksum);
        ptr[6] = checksum;
        pci_set_word(ptr + pcir_offset + 4, vendor_id);
    }

    if (device_id != rom_device_id) {
        /* Patch device id and checksum (at offset 6 for etherboot roms). */
        checksum += (uint8_t)rom_device_id + (uint8_t)(rom_device_id >> 8);
        checksum -= (uint8_t)device_id + (uint8_t)(device_id >> 8);
        PCI_DPRINTF("ROM checksum %02x / %02x\n", ptr[6], checksum);
        ptr[6] = checksum;
        pci_set_word(ptr + pcir_offset + 6, device_id);
    }
}

2166
/* Add an option rom for the device */
M
Markus Armbruster 已提交
2167 2168
static void pci_add_option_rom(PCIDevice *pdev, bool is_default_rom,
                               Error **errp)
2169 2170 2171 2172
{
    int size;
    char *path;
    void *ptr;
2173
    char name[32];
A
Anthony Liguori 已提交
2174
    const VMStateDescription *vmsd;
2175

2176
    if (!pdev->romfile)
M
Markus Armbruster 已提交
2177
        return;
2178
    if (strlen(pdev->romfile) == 0)
M
Markus Armbruster 已提交
2179
        return;
2180

2181 2182 2183 2184 2185 2186
    if (!pdev->rom_bar) {
        /*
         * Load rom via fw_cfg instead of creating a rom bar,
         * for 0.11 compatibility.
         */
        int class = pci_get_word(pdev->config + PCI_CLASS_DEVICE);
2187 2188 2189 2190 2191 2192

        /*
         * Hot-plugged devices can't use the option ROM
         * if the rom bar is disabled.
         */
        if (DEVICE(pdev)->hotplugged) {
M
Markus Armbruster 已提交
2193 2194 2195
            error_setg(errp, "Hot-plugged device without ROM bar"
                       " can't have an option ROM");
            return;
2196 2197
        }

2198 2199 2200
        if (class == 0x0300) {
            rom_add_vga(pdev->romfile);
        } else {
G
Gleb Natapov 已提交
2201
            rom_add_option(pdev->romfile, -1);
2202
        }
M
Markus Armbruster 已提交
2203
        return;
2204 2205
    }

2206
    path = qemu_find_file(QEMU_FILE_TYPE_BIOS, pdev->romfile);
2207
    if (path == NULL) {
2208
        path = g_strdup(pdev->romfile);
2209 2210 2211
    }

    size = get_image_size(path);
2212
    if (size < 0) {
M
Markus Armbruster 已提交
2213
        error_setg(errp, "failed to find romfile \"%s\"", pdev->romfile);
S
Stefan Hajnoczi 已提交
2214
        g_free(path);
M
Markus Armbruster 已提交
2215
        return;
S
Stefan Hajnoczi 已提交
2216
    } else if (size == 0) {
M
Markus Armbruster 已提交
2217
        error_setg(errp, "romfile \"%s\" is empty", pdev->romfile);
2218
        g_free(path);
M
Markus Armbruster 已提交
2219
        return;
2220
    }
2221
    size = pow2ceil(size);
2222

A
Anthony Liguori 已提交
2223 2224 2225 2226 2227
    vmsd = qdev_get_vmsd(DEVICE(pdev));

    if (vmsd) {
        snprintf(name, sizeof(name), "%s.rom", vmsd->name);
    } else {
2228
        snprintf(name, sizeof(name), "%s.rom", object_get_typename(OBJECT(pdev)));
A
Anthony Liguori 已提交
2229
    }
A
Avi Kivity 已提交
2230
    pdev->has_rom = true;
M
Michael S. Tsirkin 已提交
2231
    memory_region_init_rom(&pdev->rom, OBJECT(pdev), name, size, &error_fatal);
2232
    vmstate_register_ram(&pdev->rom, &pdev->qdev);
A
Avi Kivity 已提交
2233
    ptr = memory_region_get_ram_ptr(&pdev->rom);
2234
    load_image(path, ptr);
2235
    g_free(path);
2236

2237 2238 2239 2240 2241
    if (is_default_rom) {
        /* Only the default rom images will be patched (if needed). */
        pci_patch_ids(pdev, ptr, size);
    }

2242
    pci_register_bar(pdev, PCI_ROM_SLOT, 0, &pdev->rom);
2243 2244
}

2245 2246
static void pci_del_option_rom(PCIDevice *pdev)
{
A
Avi Kivity 已提交
2247
    if (!pdev->has_rom)
2248 2249
        return;

2250
    vmstate_unregister_ram(&pdev->rom, &pdev->qdev);
A
Avi Kivity 已提交
2251
    pdev->has_rom = false;
2252 2253
}

2254 2255 2256
/*
 * if offset = 0,
 * Find and reserve space and add capability to the linked list
2257 2258
 * in pci config space
 */
2259 2260
int pci_add_capability(PCIDevice *pdev, uint8_t cap_id,
                       uint8_t offset, uint8_t size)
2261 2262 2263 2264 2265 2266 2267
{
    int ret;
    Error *local_err = NULL;

    ret = pci_add_capability2(pdev, cap_id, offset, size, &local_err);
    if (local_err) {
        assert(ret < 0);
2268
        error_report_err(local_err);
2269 2270 2271 2272 2273 2274 2275 2276 2277 2278
    } else {
        /* success implies a positive offset in config space */
        assert(ret > 0);
    }
    return ret;
}

int pci_add_capability2(PCIDevice *pdev, uint8_t cap_id,
                       uint8_t offset, uint8_t size,
                       Error **errp)
2279
{
2280
    uint8_t *config;
2281 2282
    int i, overlapping_cap;

2283 2284
    if (!offset) {
        offset = pci_find_space(pdev, size);
2285 2286
        /* out of PCI config space is programming error */
        assert(offset);
2287 2288 2289 2290 2291 2292 2293 2294
    } else {
        /* Verify that capabilities don't overlap.  Note: device assignment
         * depends on this check to verify that the device is not broken.
         * Should never trigger for emulated devices, but it's helpful
         * for debugging these. */
        for (i = offset; i < offset + size; i++) {
            overlapping_cap = pci_find_capability_at_offset(pdev, i);
            if (overlapping_cap) {
2295 2296 2297 2298 2299 2300
                error_setg(errp, "%s:%02x:%02x.%x "
                           "Attempt to add PCI capability %x at offset "
                           "%x overlaps existing capability %x at offset %x",
                           pci_root_bus_path(pdev), pci_bus_num(pdev->bus),
                           PCI_SLOT(pdev->devfn), PCI_FUNC(pdev->devfn),
                           cap_id, offset, overlapping_cap, i);
2301 2302 2303
                return -EINVAL;
            }
        }
2304 2305 2306
    }

    config = pdev->config + offset;
2307 2308 2309 2310
    config[PCI_CAP_LIST_ID] = cap_id;
    config[PCI_CAP_LIST_NEXT] = pdev->config[PCI_CAPABILITY_LIST];
    pdev->config[PCI_CAPABILITY_LIST] = offset;
    pdev->config[PCI_STATUS] |= PCI_STATUS_CAP_LIST;
2311
    memset(pdev->used + offset, 0xFF, QEMU_ALIGN_UP(size, 4));
2312 2313
    /* Make capability read-only by default */
    memset(pdev->wmask + offset, 0, size);
2314 2315
    /* Check capability by default */
    memset(pdev->cmask + offset, 0xFF, size);
2316 2317 2318 2319 2320 2321 2322 2323 2324 2325
    return offset;
}

/* Unlink capability from the pci config space. */
void pci_del_capability(PCIDevice *pdev, uint8_t cap_id, uint8_t size)
{
    uint8_t prev, offset = pci_find_capability_list(pdev, cap_id, &prev);
    if (!offset)
        return;
    pdev->config[prev] = pdev->config[offset + PCI_CAP_LIST_NEXT];
S
Stefan Weil 已提交
2326
    /* Make capability writable again */
2327
    memset(pdev->wmask + offset, 0xff, size);
2328
    memset(pdev->w1cmask + offset, 0, size);
2329 2330
    /* Clear cmask as device-specific registers can't be checked */
    memset(pdev->cmask + offset, 0, size);
2331
    memset(pdev->used + offset, 0, QEMU_ALIGN_UP(size, 4));
2332 2333 2334 2335 2336 2337 2338 2339 2340

    if (!pdev->config[PCI_CAPABILITY_LIST])
        pdev->config[PCI_STATUS] &= ~PCI_STATUS_CAP_LIST;
}

uint8_t pci_find_capability(PCIDevice *pdev, uint8_t cap_id)
{
    return pci_find_capability_list(pdev, cap_id, NULL);
}
2341 2342 2343 2344 2345 2346 2347 2348 2349

static void pcibus_dev_print(Monitor *mon, DeviceState *dev, int indent)
{
    PCIDevice *d = (PCIDevice *)dev;
    const pci_class_desc *desc;
    char ctxt[64];
    PCIIORegion *r;
    int i, class;

2350
    class = pci_get_word(d->config + PCI_CLASS_DEVICE);
2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361
    desc = pci_class_descriptions;
    while (desc->desc && class != desc->class)
        desc++;
    if (desc->desc) {
        snprintf(ctxt, sizeof(ctxt), "%s", desc->desc);
    } else {
        snprintf(ctxt, sizeof(ctxt), "Class %04x", class);
    }

    monitor_printf(mon, "%*sclass %s, addr %02x:%02x.%x, "
                   "pci id %04x:%04x (sub %04x:%04x)\n",
2362
                   indent, "", ctxt, pci_bus_num(d->bus),
2363
                   PCI_SLOT(d->devfn), PCI_FUNC(d->devfn),
2364 2365 2366 2367
                   pci_get_word(d->config + PCI_VENDOR_ID),
                   pci_get_word(d->config + PCI_DEVICE_ID),
                   pci_get_word(d->config + PCI_SUBSYSTEM_VENDOR_ID),
                   pci_get_word(d->config + PCI_SUBSYSTEM_ID));
2368 2369 2370 2371
    for (i = 0; i < PCI_NUM_REGIONS; i++) {
        r = &d->io_regions[i];
        if (!r->size)
            continue;
2372 2373 2374
        monitor_printf(mon, "%*sbar %d: %s at 0x%"FMT_PCIBUS
                       " [0x%"FMT_PCIBUS"]\n",
                       indent, "",
2375
                       i, r->type & PCI_BASE_ADDRESS_SPACE_IO ? "i/o" : "mem",
2376 2377 2378
                       r->addr, r->addr + r->size - 1);
    }
}
G
Gerd Hoffmann 已提交
2379

2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418
static char *pci_dev_fw_name(DeviceState *dev, char *buf, int len)
{
    PCIDevice *d = (PCIDevice *)dev;
    const char *name = NULL;
    const pci_class_desc *desc =  pci_class_descriptions;
    int class = pci_get_word(d->config + PCI_CLASS_DEVICE);

    while (desc->desc &&
          (class & ~desc->fw_ign_bits) !=
          (desc->class & ~desc->fw_ign_bits)) {
        desc++;
    }

    if (desc->desc) {
        name = desc->fw_name;
    }

    if (name) {
        pstrcpy(buf, len, name);
    } else {
        snprintf(buf, len, "pci%04x,%04x",
                 pci_get_word(d->config + PCI_VENDOR_ID),
                 pci_get_word(d->config + PCI_DEVICE_ID));
    }

    return buf;
}

static char *pcibus_get_fw_dev_path(DeviceState *dev)
{
    PCIDevice *d = (PCIDevice *)dev;
    char path[50], name[33];
    int off;

    off = snprintf(path, sizeof(path), "%s@%x",
                   pci_dev_fw_name(dev, name, sizeof name),
                   PCI_SLOT(d->devfn));
    if (PCI_FUNC(d->devfn))
        snprintf(path + off, sizeof(path) + off, ",%x", PCI_FUNC(d->devfn));
2419
    return g_strdup(path);
2420 2421
}

2422 2423
static char *pcibus_get_dev_path(DeviceState *dev)
{
2424 2425 2426 2427 2428 2429 2430 2431
    PCIDevice *d = container_of(dev, PCIDevice, qdev);
    PCIDevice *t;
    int slot_depth;
    /* Path format: Domain:00:Slot.Function:Slot.Function....:Slot.Function.
     * 00 is added here to make this format compatible with
     * domain:Bus:Slot.Func for systems without nested PCI bridges.
     * Slot.Function list specifies the slot and function numbers for all
     * devices on the path from root to the specific device. */
2432 2433
    const char *root_bus_path;
    int root_bus_len;
M
Michael S. Tsirkin 已提交
2434 2435
    char slot[] = ":SS.F";
    int slot_len = sizeof slot - 1 /* For '\0' */;
2436 2437
    int path_len;
    char *path, *p;
M
Michael S. Tsirkin 已提交
2438
    int s;
2439

2440 2441 2442
    root_bus_path = pci_root_bus_path(d);
    root_bus_len = strlen(root_bus_path);

2443 2444 2445 2446 2447 2448
    /* Calculate # of slots on path between device and root. */;
    slot_depth = 0;
    for (t = d; t; t = t->bus->parent_dev) {
        ++slot_depth;
    }

2449
    path_len = root_bus_len + slot_len * slot_depth;
2450 2451

    /* Allocate memory, fill in the terminating null byte. */
2452
    path = g_malloc(path_len + 1 /* For '\0' */);
2453 2454
    path[path_len] = '\0';

2455
    memcpy(path, root_bus_path, root_bus_len);
2456 2457 2458 2459 2460 2461

    /* Fill in slot numbers. We walk up from device to root, so need to print
     * them in the reverse order, last to first. */
    p = path + path_len;
    for (t = d; t; t = t->bus->parent_dev) {
        p -= slot_len;
M
Michael S. Tsirkin 已提交
2462
        s = snprintf(slot, sizeof slot, ":%02x.%x",
2463
                     PCI_SLOT(t->devfn), PCI_FUNC(t->devfn));
M
Michael S. Tsirkin 已提交
2464 2465
        assert(s == slot_len);
        memcpy(p, slot, slot_len);
2466 2467 2468
    }

    return path;
2469 2470
}

2471 2472 2473 2474 2475 2476 2477 2478 2479
static int pci_qdev_find_recursive(PCIBus *bus,
                                   const char *id, PCIDevice **pdev)
{
    DeviceState *qdev = qdev_find_recursive(&bus->qbus, id);
    if (!qdev) {
        return -ENODEV;
    }

    /* roughly check if given qdev is pci device */
A
Anthony Liguori 已提交
2480
    if (object_dynamic_cast(OBJECT(qdev), TYPE_PCI_DEVICE)) {
2481
        *pdev = PCI_DEVICE(qdev);
2482 2483 2484 2485 2486 2487 2488
        return 0;
    }
    return -EINVAL;
}

int pci_qdev_find_device(const char *id, PCIDevice **pdev)
{
2489
    PCIHostState *host_bridge;
2490 2491
    int rc = -ENODEV;

2492 2493
    QLIST_FOREACH(host_bridge, &pci_host_bridges, next) {
        int tmp = pci_qdev_find_recursive(host_bridge->bus, id, pdev);
2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504
        if (!tmp) {
            rc = 0;
            break;
        }
        if (tmp != -ENODEV) {
            rc = tmp;
        }
    }

    return rc;
}
A
Avi Kivity 已提交
2505 2506 2507 2508 2509

MemoryRegion *pci_address_space(PCIDevice *dev)
{
    return dev->bus->address_space_mem;
}
2510 2511 2512 2513 2514

MemoryRegion *pci_address_space_io(PCIDevice *dev)
{
    return dev->bus->address_space_io;
}
2515

2516 2517 2518
static void pci_device_class_init(ObjectClass *klass, void *data)
{
    DeviceClass *k = DEVICE_CLASS(klass);
2519 2520
    PCIDeviceClass *pc = PCI_DEVICE_CLASS(klass);

M
Markus Armbruster 已提交
2521 2522
    k->realize = pci_qdev_realize;
    k->unrealize = pci_qdev_unrealize;
2523
    k->bus_type = TYPE_PCI_BUS;
2524
    k->props = pci_props;
2525
    pc->realize = pci_default_realize;
2526 2527
}

2528 2529 2530
AddressSpace *pci_device_iommu_address_space(PCIDevice *dev)
{
    PCIBus *bus = PCI_BUS(dev->bus);
2531
    PCIBus *iommu_bus = bus;
2532

2533 2534
    while(iommu_bus && !iommu_bus->iommu_fn && iommu_bus->parent_dev) {
        iommu_bus = PCI_BUS(iommu_bus->parent_dev->bus);
2535
    }
2536 2537
    if (iommu_bus && iommu_bus->iommu_fn) {
        return iommu_bus->iommu_fn(bus, iommu_bus->iommu_opaque, dev->devfn);
2538 2539 2540 2541
    }
    return &address_space_memory;
}

2542
void pci_setup_iommu(PCIBus *bus, PCIIOMMUFunc fn, void *opaque)
2543
{
2544 2545
    bus->iommu_fn = fn;
    bus->iommu_opaque = opaque;
2546 2547
}

2548 2549 2550 2551 2552
static void pci_dev_get_w64(PCIBus *b, PCIDevice *dev, void *opaque)
{
    Range *range = opaque;
    PCIDeviceClass *pc = PCI_DEVICE_GET_CLASS(dev);
    uint16_t cmd = pci_get_word(dev->config + PCI_COMMAND);
2553
    int i;
2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566

    if (!(cmd & PCI_COMMAND_MEMORY)) {
        return;
    }

    if (pc->is_bridge) {
        pcibus_t base = pci_bridge_get_base(dev, PCI_BASE_ADDRESS_MEM_PREFETCH);
        pcibus_t limit = pci_bridge_get_limit(dev, PCI_BASE_ADDRESS_MEM_PREFETCH);

        base = MAX(base, 0x1ULL << 32);

        if (limit >= base) {
            Range pref_range;
2567
            range_set_bounds(&pref_range, base, limit);
2568 2569 2570
            range_extend(range, &pref_range);
        }
    }
2571 2572
    for (i = 0; i < PCI_NUM_REGIONS; ++i) {
        PCIIORegion *r = &dev->io_regions[i];
2573
        pcibus_t lob, upb;
2574 2575
        Range region_range;

2576 2577 2578 2579 2580 2581
        if (!r->size ||
            (r->type & PCI_BASE_ADDRESS_SPACE_IO) ||
            !(r->type & PCI_BASE_ADDRESS_MEM_TYPE_64)) {
            continue;
        }

2582 2583 2584
        lob = pci_bar_address(dev, i, r->type, r->size);
        upb = lob + r->size - 1;
        if (lob == PCI_BAR_UNMAPPED) {
2585 2586 2587
            continue;
        }

2588
        lob = MAX(lob, 0x1ULL << 32);
2589

2590 2591
        if (upb >= lob) {
            range_set_bounds(&region_range, lob, upb);
2592 2593 2594 2595 2596 2597 2598
            range_extend(range, &region_range);
        }
    }
}

void pci_bus_get_w64_range(PCIBus *bus, Range *range)
{
2599
    range_make_empty(range);
2600 2601 2602
    pci_for_each_device_under_bus(bus, pci_dev_get_w64, range);
}

C
Cao jin 已提交
2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629
static bool pcie_has_upstream_port(PCIDevice *dev)
{
    PCIDevice *parent_dev = pci_bridge_get_device(dev->bus);

    /* Device associated with an upstream port.
     * As there are several types of these, it's easier to check the
     * parent device: upstream ports are always connected to
     * root or downstream ports.
     */
    return parent_dev &&
        pci_is_express(parent_dev) &&
        parent_dev->exp.exp_cap &&
        (pcie_cap_get_type(parent_dev) == PCI_EXP_TYPE_ROOT_PORT ||
         pcie_cap_get_type(parent_dev) == PCI_EXP_TYPE_DOWNSTREAM);
}

PCIDevice *pci_get_function_0(PCIDevice *pci_dev)
{
    if(pcie_has_upstream_port(pci_dev)) {
        /* With an upstream PCIe port, we only support 1 device at slot 0 */
        return pci_dev->bus->devices[0];
    } else {
        /* Other bus types might support multiple devices at slots 0-31 */
        return pci_dev->bus->devices[PCI_DEVFN(PCI_SLOT(pci_dev->devfn), 0)];
    }
}

2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644
MSIMessage pci_get_msi_message(PCIDevice *dev, int vector)
{
    MSIMessage msg;
    if (msix_enabled(dev)) {
        msg = msix_get_message(dev, vector);
    } else if (msi_enabled(dev)) {
        msg = msi_get_message(dev, vector);
    } else {
        /* Should never happen */
        error_report("%s: unknown interrupt type", __func__);
        abort();
    }
    return msg;
}

2645
static const TypeInfo pci_device_type_info = {
2646 2647 2648 2649 2650
    .name = TYPE_PCI_DEVICE,
    .parent = TYPE_DEVICE,
    .instance_size = sizeof(PCIDevice),
    .abstract = true,
    .class_size = sizeof(PCIDeviceClass),
2651
    .class_init = pci_device_class_init,
2652 2653
};

A
Andreas Färber 已提交
2654
static void pci_register_types(void)
2655
{
2656
    type_register_static(&pci_bus_info);
2657
    type_register_static(&pcie_bus_info);
2658 2659 2660
    type_register_static(&pci_device_type_info);
}

A
Andreas Färber 已提交
2661
type_init(pci_register_types)