diff options
Diffstat (limited to 'hw')
-rw-r--r-- | hw/audio/intel-hda.c | 6 | ||||
-rw-r--r-- | hw/block/vhost-user-blk.c | 45 | ||||
-rw-r--r-- | hw/display/qxl-logger.c | 22 | ||||
-rw-r--r-- | hw/display/qxl-render.c | 20 | ||||
-rw-r--r-- | hw/display/qxl.c | 42 | ||||
-rw-r--r-- | hw/display/qxl.h | 23 | ||||
-rw-r--r-- | hw/net/vhost_net.c | 8 | ||||
-rw-r--r-- | hw/scsi/vhost-scsi-common.c | 4 | ||||
-rw-r--r-- | hw/usb/hcd-xhci-pci.c | 2 | ||||
-rw-r--r-- | hw/usb/hcd-xhci-sysbus.c | 2 | ||||
-rw-r--r-- | hw/usb/hcd-xhci.c | 2 | ||||
-rw-r--r-- | hw/usb/redirect.c | 3 | ||||
-rw-r--r-- | hw/virtio/trace-events | 4 | ||||
-rw-r--r-- | hw/virtio/vhost-user-fs.c | 4 | ||||
-rw-r--r-- | hw/virtio/vhost-user-gpio.c | 26 | ||||
-rw-r--r-- | hw/virtio/vhost-user-i2c.c | 4 | ||||
-rw-r--r-- | hw/virtio/vhost-user-rng.c | 4 | ||||
-rw-r--r-- | hw/virtio/vhost-user.c | 71 | ||||
-rw-r--r-- | hw/virtio/vhost-vsock-common.c | 4 | ||||
-rw-r--r-- | hw/virtio/vhost.c | 44 |
20 files changed, 244 insertions, 96 deletions
diff --git a/hw/audio/intel-hda.c b/hw/audio/intel-hda.c index f38117057b..b9ed231fe8 100644 --- a/hw/audio/intel-hda.c +++ b/hw/audio/intel-hda.c @@ -220,8 +220,6 @@ struct IntelHDAReg { void (*rhandler)(IntelHDAState *d, const IntelHDAReg *reg); }; -static void intel_hda_reset(DeviceState *dev); - /* --------------------------------------------------------------------- */ static hwaddr intel_hda_addr(uint32_t lbase, uint32_t ubase) @@ -516,7 +514,7 @@ static void intel_hda_notify_codecs(IntelHDAState *d, uint32_t stream, bool runn static void intel_hda_set_g_ctl(IntelHDAState *d, const IntelHDAReg *reg, uint32_t old) { if ((d->g_ctl & ICH6_GCTL_RESET) == 0) { - intel_hda_reset(DEVICE(d)); + device_cold_reset(DEVICE(d)); } } @@ -1083,11 +1081,9 @@ static void intel_hda_reset(DeviceState *dev) intel_hda_regs_reset(d); d->wall_base_ns = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL); - /* reset codecs */ QTAILQ_FOREACH(kid, &d->codecs.qbus.children, sibling) { DeviceState *qdev = kid->child; cdev = HDA_CODEC_DEVICE(qdev); - device_legacy_reset(DEVICE(cdev)); d->state_sts |= (1 << cdev->cad); } intel_hda_update_irq(d); diff --git a/hw/block/vhost-user-blk.c b/hw/block/vhost-user-blk.c index 0d5190accf..aff4d2b8cb 100644 --- a/hw/block/vhost-user-blk.c +++ b/hw/block/vhost-user-blk.c @@ -178,7 +178,7 @@ static int vhost_user_blk_start(VirtIODevice *vdev, Error **errp) } s->dev.vq_index_end = s->dev.nvqs; - ret = vhost_dev_start(&s->dev, vdev); + ret = vhost_dev_start(&s->dev, vdev, true); if (ret < 0) { error_setg_errno(errp, -ret, "Error starting vhost"); goto err_guest_notifiers; @@ -213,7 +213,7 @@ static void vhost_user_blk_stop(VirtIODevice *vdev) return; } - vhost_dev_stop(&s->dev, vdev); + vhost_dev_stop(&s->dev, vdev, true); ret = k->set_guest_notifiers(qbus->parent, s->dev.nvqs, false); if (ret < 0) { @@ -369,17 +369,10 @@ static void vhost_user_blk_disconnect(DeviceState *dev) vhost_user_blk_stop(vdev); vhost_dev_cleanup(&s->dev); -} -static void vhost_user_blk_chr_closed_bh(void *opaque) -{ - DeviceState *dev = opaque; - VirtIODevice *vdev = VIRTIO_DEVICE(dev); - VHostUserBlk *s = VHOST_USER_BLK(vdev); - - vhost_user_blk_disconnect(dev); + /* Re-instate the event handler for new connections */ qemu_chr_fe_set_handlers(&s->chardev, NULL, NULL, vhost_user_blk_event, - NULL, opaque, NULL, true); + NULL, dev, NULL, true); } static void vhost_user_blk_event(void *opaque, QEMUChrEvent event) @@ -398,33 +391,9 @@ static void vhost_user_blk_event(void *opaque, QEMUChrEvent event) } break; case CHR_EVENT_CLOSED: - if (!runstate_check(RUN_STATE_SHUTDOWN)) { - /* - * A close event may happen during a read/write, but vhost - * code assumes the vhost_dev remains setup, so delay the - * stop & clear. - */ - AioContext *ctx = qemu_get_current_aio_context(); - - qemu_chr_fe_set_handlers(&s->chardev, NULL, NULL, NULL, NULL, - NULL, NULL, false); - aio_bh_schedule_oneshot(ctx, vhost_user_blk_chr_closed_bh, opaque); - - /* - * Move vhost device to the stopped state. The vhost-user device - * will be clean up and disconnected in BH. This can be useful in - * the vhost migration code. If disconnect was caught there is an - * option for the general vhost code to get the dev state without - * knowing its type (in this case vhost-user). - * - * FIXME: this is sketchy to be reaching into vhost_dev - * now because we are forcing something that implies we - * have executed vhost_dev_stop() but that won't happen - * until vhost_user_blk_stop() gets called from the bh. - * Really this state check should be tracked locally. - */ - s->dev.started = false; - } + /* defer close until later to avoid circular close */ + vhost_user_async_close(dev, &s->chardev, &s->dev, + vhost_user_blk_disconnect); break; case CHR_EVENT_BREAK: case CHR_EVENT_MUX_IN: diff --git a/hw/display/qxl-logger.c b/hw/display/qxl-logger.c index 68bfa47568..35c38f6252 100644 --- a/hw/display/qxl-logger.c +++ b/hw/display/qxl-logger.c @@ -106,7 +106,7 @@ static int qxl_log_image(PCIQXLDevice *qxl, QXLPHYSICAL addr, int group_id) QXLImage *image; QXLImageDescriptor *desc; - image = qxl_phys2virt(qxl, addr, group_id); + image = qxl_phys2virt(qxl, addr, group_id, sizeof(QXLImage)); if (!image) { return 1; } @@ -214,7 +214,8 @@ int qxl_log_cmd_cursor(PCIQXLDevice *qxl, QXLCursorCmd *cmd, int group_id) cmd->u.set.position.y, cmd->u.set.visible ? "yes" : "no", cmd->u.set.shape); - cursor = qxl_phys2virt(qxl, cmd->u.set.shape, group_id); + cursor = qxl_phys2virt(qxl, cmd->u.set.shape, group_id, + sizeof(QXLCursor)); if (!cursor) { return 1; } @@ -236,6 +237,7 @@ int qxl_log_command(PCIQXLDevice *qxl, const char *ring, QXLCommandExt *ext) { bool compat = ext->flags & QXL_COMMAND_FLAG_COMPAT; void *data; + size_t datasz; int ret; if (!qxl->cmdlog) { @@ -247,7 +249,20 @@ int qxl_log_command(PCIQXLDevice *qxl, const char *ring, QXLCommandExt *ext) qxl_name(qxl_type, ext->cmd.type), compat ? "(compat)" : ""); - data = qxl_phys2virt(qxl, ext->cmd.data, ext->group_id); + switch (ext->cmd.type) { + case QXL_CMD_DRAW: + datasz = compat ? sizeof(QXLCompatDrawable) : sizeof(QXLDrawable); + break; + case QXL_CMD_SURFACE: + datasz = sizeof(QXLSurfaceCmd); + break; + case QXL_CMD_CURSOR: + datasz = sizeof(QXLCursorCmd); + break; + default: + goto out; + } + data = qxl_phys2virt(qxl, ext->cmd.data, ext->group_id, datasz); if (!data) { return 1; } @@ -269,6 +284,7 @@ int qxl_log_command(PCIQXLDevice *qxl, const char *ring, QXLCommandExt *ext) qxl_log_cmd_cursor(qxl, data, ext->group_id); break; } +out: fprintf(stderr, "\n"); return 0; } diff --git a/hw/display/qxl-render.c b/hw/display/qxl-render.c index ca217004bf..fcfd40c3ac 100644 --- a/hw/display/qxl-render.c +++ b/hw/display/qxl-render.c @@ -107,7 +107,9 @@ static void qxl_render_update_area_unlocked(PCIQXLDevice *qxl) qxl->guest_primary.resized = 0; qxl->guest_primary.data = qxl_phys2virt(qxl, qxl->guest_primary.surface.mem, - MEMSLOT_GROUP_GUEST); + MEMSLOT_GROUP_GUEST, + qxl->guest_primary.abs_stride + * height); if (!qxl->guest_primary.data) { goto end; } @@ -228,7 +230,8 @@ static void qxl_unpack_chunks(void *dest, size_t size, PCIQXLDevice *qxl, if (offset == size) { return; } - chunk = qxl_phys2virt(qxl, chunk->next_chunk, group_id); + chunk = qxl_phys2virt(qxl, chunk->next_chunk, group_id, + sizeof(QXLDataChunk) + chunk->data_size); if (!chunk) { return; } @@ -295,7 +298,8 @@ fail: /* called from spice server thread context only */ int qxl_render_cursor(PCIQXLDevice *qxl, QXLCommandExt *ext) { - QXLCursorCmd *cmd = qxl_phys2virt(qxl, ext->cmd.data, ext->group_id); + QXLCursorCmd *cmd = qxl_phys2virt(qxl, ext->cmd.data, ext->group_id, + sizeof(QXLCursorCmd)); QXLCursor *cursor; QEMUCursor *c; @@ -314,7 +318,15 @@ int qxl_render_cursor(PCIQXLDevice *qxl, QXLCommandExt *ext) } switch (cmd->type) { case QXL_CURSOR_SET: - cursor = qxl_phys2virt(qxl, cmd->u.set.shape, ext->group_id); + /* First read the QXLCursor to get QXLDataChunk::data_size ... */ + cursor = qxl_phys2virt(qxl, cmd->u.set.shape, ext->group_id, + sizeof(QXLCursor)); + if (!cursor) { + return 1; + } + /* Then read including the chunked data following QXLCursor. */ + cursor = qxl_phys2virt(qxl, cmd->u.set.shape, ext->group_id, + sizeof(QXLCursor) + cursor->chunk.data_size); if (!cursor) { return 1; } diff --git a/hw/display/qxl.c b/hw/display/qxl.c index 5b10f697f1..6772849dec 100644 --- a/hw/display/qxl.c +++ b/hw/display/qxl.c @@ -274,7 +274,8 @@ static void qxl_spice_monitors_config_async(PCIQXLDevice *qxl, int replay) QXL_IO_MONITORS_CONFIG_ASYNC)); } - cfg = qxl_phys2virt(qxl, qxl->guest_monitors_config, MEMSLOT_GROUP_GUEST); + cfg = qxl_phys2virt(qxl, qxl->guest_monitors_config, MEMSLOT_GROUP_GUEST, + sizeof(QXLMonitorsConfig)); if (cfg != NULL && cfg->count == 1) { qxl->guest_primary.resized = 1; qxl->guest_head0_width = cfg->heads[0].width; @@ -459,7 +460,8 @@ static int qxl_track_command(PCIQXLDevice *qxl, struct QXLCommandExt *ext) switch (le32_to_cpu(ext->cmd.type)) { case QXL_CMD_SURFACE: { - QXLSurfaceCmd *cmd = qxl_phys2virt(qxl, ext->cmd.data, ext->group_id); + QXLSurfaceCmd *cmd = qxl_phys2virt(qxl, ext->cmd.data, ext->group_id, + sizeof(QXLSurfaceCmd)); if (!cmd) { return 1; @@ -494,7 +496,8 @@ static int qxl_track_command(PCIQXLDevice *qxl, struct QXLCommandExt *ext) } case QXL_CMD_CURSOR: { - QXLCursorCmd *cmd = qxl_phys2virt(qxl, ext->cmd.data, ext->group_id); + QXLCursorCmd *cmd = qxl_phys2virt(qxl, ext->cmd.data, ext->group_id, + sizeof(QXLCursorCmd)); if (!cmd) { return 1; @@ -1381,6 +1384,7 @@ static int qxl_add_memslot(PCIQXLDevice *d, uint32_t slot_id, uint64_t delta, qxl_set_guest_bug(d, "%s: pci_region = %d", __func__, pci_region); return 1; } + assert(guest_end - pci_start <= memory_region_size(mr)); virt_start = (intptr_t)memory_region_get_ram_ptr(mr); memslot.slot_id = slot_id; @@ -1421,11 +1425,13 @@ static void qxl_reset_surfaces(PCIQXLDevice *d) /* can be also called from spice server thread context */ static bool qxl_get_check_slot_offset(PCIQXLDevice *qxl, QXLPHYSICAL pqxl, - uint32_t *s, uint64_t *o) + uint32_t *s, uint64_t *o, + size_t size_requested) { uint64_t phys = le64_to_cpu(pqxl); uint32_t slot = (phys >> (64 - 8)) & 0xff; uint64_t offset = phys & 0xffffffffffff; + uint64_t size_available; if (slot >= NUM_MEMSLOTS) { qxl_set_guest_bug(qxl, "slot too large %d >= %d", slot, @@ -1449,6 +1455,23 @@ static bool qxl_get_check_slot_offset(PCIQXLDevice *qxl, QXLPHYSICAL pqxl, slot, offset, qxl->guest_slots[slot].size); return false; } + size_available = memory_region_size(qxl->guest_slots[slot].mr); + if (qxl->guest_slots[slot].offset + offset >= size_available) { + qxl_set_guest_bug(qxl, + "slot %d offset %"PRIu64" > region size %"PRIu64"\n", + slot, qxl->guest_slots[slot].offset + offset, + size_available); + return false; + } + size_available -= qxl->guest_slots[slot].offset + offset; + if (size_requested > size_available) { + qxl_set_guest_bug(qxl, + "slot %d offset %"PRIu64" size %zu: " + "overrun by %"PRIu64" bytes\n", + slot, offset, size_requested, + size_requested - size_available); + return false; + } *s = slot; *o = offset; @@ -1456,7 +1479,8 @@ static bool qxl_get_check_slot_offset(PCIQXLDevice *qxl, QXLPHYSICAL pqxl, } /* can be also called from spice server thread context */ -void *qxl_phys2virt(PCIQXLDevice *qxl, QXLPHYSICAL pqxl, int group_id) +void *qxl_phys2virt(PCIQXLDevice *qxl, QXLPHYSICAL pqxl, int group_id, + size_t size) { uint64_t offset; uint32_t slot; @@ -1467,7 +1491,7 @@ void *qxl_phys2virt(PCIQXLDevice *qxl, QXLPHYSICAL pqxl, int group_id) offset = le64_to_cpu(pqxl) & 0xffffffffffff; return (void *)(intptr_t)offset; case MEMSLOT_GROUP_GUEST: - if (!qxl_get_check_slot_offset(qxl, pqxl, &slot, &offset)) { + if (!qxl_get_check_slot_offset(qxl, pqxl, &slot, &offset, size)) { return NULL; } ptr = memory_region_get_ram_ptr(qxl->guest_slots[slot].mr); @@ -1933,9 +1957,9 @@ static void qxl_dirty_one_surface(PCIQXLDevice *qxl, QXLPHYSICAL pqxl, uint32_t slot; bool rc; - rc = qxl_get_check_slot_offset(qxl, pqxl, &slot, &offset); - assert(rc == true); size = (uint64_t)height * abs(stride); + rc = qxl_get_check_slot_offset(qxl, pqxl, &slot, &offset, size); + assert(rc == true); trace_qxl_surfaces_dirty(qxl->id, offset, size); qxl_set_dirty(qxl->guest_slots[slot].mr, qxl->guest_slots[slot].offset + offset, @@ -1964,7 +1988,7 @@ static void qxl_dirty_surfaces(PCIQXLDevice *qxl) } cmd = qxl_phys2virt(qxl, qxl->guest_surfaces.cmds[i], - MEMSLOT_GROUP_GUEST); + MEMSLOT_GROUP_GUEST, sizeof(QXLSurfaceCmd)); assert(cmd); assert(cmd->type == QXL_SURFACE_CMD_CREATE); qxl_dirty_one_surface(qxl, cmd->u.surface_create.data, diff --git a/hw/display/qxl.h b/hw/display/qxl.h index e74de9579d..7894bd5134 100644 --- a/hw/display/qxl.h +++ b/hw/display/qxl.h @@ -147,7 +147,28 @@ OBJECT_DECLARE_SIMPLE_TYPE(PCIQXLDevice, PCI_QXL) #define QXL_DEFAULT_REVISION (QXL_REVISION_STABLE_V12 + 1) /* qxl.c */ -void *qxl_phys2virt(PCIQXLDevice *qxl, QXLPHYSICAL phys, int group_id); +/** + * qxl_phys2virt: Get a pointer within a PCI VRAM memory region. + * + * @qxl: QXL device + * @phys: physical offset of buffer within the VRAM + * @group_id: memory slot group + * @size: size of the buffer + * + * Returns a host pointer to a buffer placed at offset @phys within the + * active slot @group_id of the PCI VGA RAM memory region associated with + * the @qxl device. If the slot is inactive, or the offset + size are out + * of the memory region, returns NULL. + * + * Use with care; by the time this function returns, the returned pointer is + * not protected by RCU anymore. If the caller is not within an RCU critical + * section and does not hold the iothread lock, it must have other means of + * protecting the pointer, such as a reference to the region that includes + * the incoming ram_addr_t. + * + */ +void *qxl_phys2virt(PCIQXLDevice *qxl, QXLPHYSICAL phys, int group_id, + size_t size); void qxl_set_guest_bug(PCIQXLDevice *qxl, const char *msg, ...) G_GNUC_PRINTF(2, 3); diff --git a/hw/net/vhost_net.c b/hw/net/vhost_net.c index 26e4930676..043058ff43 100644 --- a/hw/net/vhost_net.c +++ b/hw/net/vhost_net.c @@ -259,7 +259,7 @@ static int vhost_net_start_one(struct vhost_net *net, goto fail_notifiers; } - r = vhost_dev_start(&net->dev, dev); + r = vhost_dev_start(&net->dev, dev, false); if (r < 0) { goto fail_start; } @@ -308,7 +308,7 @@ fail: if (net->nc->info->poll) { net->nc->info->poll(net->nc, true); } - vhost_dev_stop(&net->dev, dev); + vhost_dev_stop(&net->dev, dev, false); fail_start: vhost_dev_disable_notifiers(&net->dev, dev); fail_notifiers: @@ -329,7 +329,7 @@ static void vhost_net_stop_one(struct vhost_net *net, if (net->nc->info->poll) { net->nc->info->poll(net->nc, true); } - vhost_dev_stop(&net->dev, dev); + vhost_dev_stop(&net->dev, dev, false); if (net->nc->info->stop) { net->nc->info->stop(net->nc); } @@ -606,7 +606,7 @@ err_start: assert(r >= 0); } - vhost_dev_stop(&net->dev, vdev); + vhost_dev_stop(&net->dev, vdev, false); return r; } diff --git a/hw/scsi/vhost-scsi-common.c b/hw/scsi/vhost-scsi-common.c index 767f827e55..18ea5dcfa1 100644 --- a/hw/scsi/vhost-scsi-common.c +++ b/hw/scsi/vhost-scsi-common.c @@ -68,7 +68,7 @@ int vhost_scsi_common_start(VHostSCSICommon *vsc) goto err_guest_notifiers; } - ret = vhost_dev_start(&vsc->dev, vdev); + ret = vhost_dev_start(&vsc->dev, vdev, true); if (ret < 0) { error_report("Error start vhost dev"); goto err_guest_notifiers; @@ -101,7 +101,7 @@ void vhost_scsi_common_stop(VHostSCSICommon *vsc) VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(qbus); int ret = 0; - vhost_dev_stop(&vsc->dev, vdev); + vhost_dev_stop(&vsc->dev, vdev, true); if (k->set_guest_notifiers) { ret = k->set_guest_notifiers(qbus->parent, vsc->dev.nvqs, false); diff --git a/hw/usb/hcd-xhci-pci.c b/hw/usb/hcd-xhci-pci.c index e934b1a5b1..643d4643e4 100644 --- a/hw/usb/hcd-xhci-pci.c +++ b/hw/usb/hcd-xhci-pci.c @@ -85,7 +85,7 @@ static void xhci_pci_reset(DeviceState *dev) { XHCIPciState *s = XHCI_PCI(dev); - device_legacy_reset(DEVICE(&s->xhci)); + device_cold_reset(DEVICE(&s->xhci)); } static int xhci_pci_vmstate_post_load(void *opaque, int version_id) diff --git a/hw/usb/hcd-xhci-sysbus.c b/hw/usb/hcd-xhci-sysbus.c index a14e438196..faf57b4797 100644 --- a/hw/usb/hcd-xhci-sysbus.c +++ b/hw/usb/hcd-xhci-sysbus.c @@ -29,7 +29,7 @@ void xhci_sysbus_reset(DeviceState *dev) { XHCISysbusState *s = XHCI_SYSBUS(dev); - device_legacy_reset(DEVICE(&s->xhci)); + device_cold_reset(DEVICE(&s->xhci)); } static void xhci_sysbus_realize(DeviceState *dev, Error **errp) diff --git a/hw/usb/hcd-xhci.c b/hw/usb/hcd-xhci.c index 8299f35e66..b89b618ec2 100644 --- a/hw/usb/hcd-xhci.c +++ b/hw/usb/hcd-xhci.c @@ -796,7 +796,7 @@ static int xhci_ring_chain_length(XHCIState *xhci, const XHCIRing *ring) */ } while (length < TRB_LINK_LIMIT * 65536 / TRB_SIZE); - qemu_log_mask(LOG_GUEST_ERROR, "%s: exceeded maximum tranfer ring size!\n", + qemu_log_mask(LOG_GUEST_ERROR, "%s: exceeded maximum transfer ring size!\n", __func__); return -1; diff --git a/hw/usb/redirect.c b/hw/usb/redirect.c index 1bd30efc3e..fd7df599bc 100644 --- a/hw/usb/redirect.c +++ b/hw/usb/redirect.c @@ -1280,8 +1280,7 @@ static void usbredir_create_parser(USBRedirDevice *dev) } #endif - if (runstate_check(RUN_STATE_INMIGRATE) || - runstate_check(RUN_STATE_PRELAUNCH)) { + if (runstate_check(RUN_STATE_INMIGRATE)) { flags |= usbredirparser_fl_no_hello; } usbredirparser_init(dev->parser, VERSION, caps, USB_REDIR_CAPS_SIZE, diff --git a/hw/virtio/trace-events b/hw/virtio/trace-events index 820dadc26c..14fc5b9bb2 100644 --- a/hw/virtio/trace-events +++ b/hw/virtio/trace-events @@ -9,8 +9,8 @@ vhost_section(const char *name) "%s" vhost_reject_section(const char *name, int d) "%s:%d" vhost_iotlb_miss(void *dev, int step) "%p step %d" vhost_dev_cleanup(void *dev) "%p" -vhost_dev_start(void *dev, const char *name) "%p:%s" -vhost_dev_stop(void *dev, const char *name) "%p:%s" +vhost_dev_start(void *dev, const char *name, bool vrings) "%p:%s vrings:%d" +vhost_dev_stop(void *dev, const char *name, bool vrings) "%p:%s vrings:%d" # vhost-user.c diff --git a/hw/virtio/vhost-user-fs.c b/hw/virtio/vhost-user-fs.c index dc4014cdef..d97b179e6f 100644 --- a/hw/virtio/vhost-user-fs.c +++ b/hw/virtio/vhost-user-fs.c @@ -76,7 +76,7 @@ static void vuf_start(VirtIODevice *vdev) } fs->vhost_dev.acked_features = vdev->guest_features; - ret = vhost_dev_start(&fs->vhost_dev, vdev); + ret = vhost_dev_start(&fs->vhost_dev, vdev, true); if (ret < 0) { error_report("Error starting vhost: %d", -ret); goto err_guest_notifiers; @@ -110,7 +110,7 @@ static void vuf_stop(VirtIODevice *vdev) return; } - vhost_dev_stop(&fs->vhost_dev, vdev); + vhost_dev_stop(&fs->vhost_dev, vdev, true); ret = k->set_guest_notifiers(qbus->parent, fs->vhost_dev.nvqs, false); if (ret < 0) { diff --git a/hw/virtio/vhost-user-gpio.c b/hw/virtio/vhost-user-gpio.c index 5851cb3bc9..b7b82a1099 100644 --- a/hw/virtio/vhost-user-gpio.c +++ b/hw/virtio/vhost-user-gpio.c @@ -81,11 +81,12 @@ static int vu_gpio_start(VirtIODevice *vdev) */ vhost_ack_features(&gpio->vhost_dev, feature_bits, vdev->guest_features); - ret = vhost_dev_start(&gpio->vhost_dev, vdev); + ret = vhost_dev_start(&gpio->vhost_dev, vdev, false); if (ret < 0) { error_report("Error starting vhost-user-gpio: %d", ret); goto err_guest_notifiers; } + gpio->started_vu = true; /* * guest_notifier_mask/pending not used yet, so just unmask @@ -126,20 +127,16 @@ static void vu_gpio_stop(VirtIODevice *vdev) struct vhost_dev *vhost_dev = &gpio->vhost_dev; int ret; - if (!k->set_guest_notifiers) { + if (!gpio->started_vu) { return; } + gpio->started_vu = false; - /* - * We can call vu_gpio_stop multiple times, for example from - * vm_state_notify and the final object finalisation. Check we - * aren't already stopped before doing so. - */ - if (!vhost_dev_is_started(vhost_dev)) { + if (!k->set_guest_notifiers) { return; } - vhost_dev_stop(vhost_dev, vdev); + vhost_dev_stop(vhost_dev, vdev, false); ret = k->set_guest_notifiers(qbus->parent, vhost_dev->nvqs, false); if (ret < 0) { @@ -236,6 +233,8 @@ static int vu_gpio_connect(DeviceState *dev, Error **errp) return 0; } +static void vu_gpio_event(void *opaque, QEMUChrEvent event); + static void vu_gpio_disconnect(DeviceState *dev) { VirtIODevice *vdev = VIRTIO_DEVICE(dev); @@ -248,6 +247,11 @@ static void vu_gpio_disconnect(DeviceState *dev) vu_gpio_stop(vdev); vhost_dev_cleanup(&gpio->vhost_dev); + + /* Re-instate the event handler for new connections */ + qemu_chr_fe_set_handlers(&gpio->chardev, + NULL, NULL, vu_gpio_event, + NULL, dev, NULL, true); } static void vu_gpio_event(void *opaque, QEMUChrEvent event) @@ -265,7 +269,9 @@ static void vu_gpio_event(void *opaque, QEMUChrEvent event) } break; case CHR_EVENT_CLOSED: - vu_gpio_disconnect(dev); + /* defer close until later to avoid circular close */ + vhost_user_async_close(dev, &gpio->chardev, &gpio->vhost_dev, + vu_gpio_disconnect); break; case CHR_EVENT_BREAK: case CHR_EVENT_MUX_IN: diff --git a/hw/virtio/vhost-user-i2c.c b/hw/virtio/vhost-user-i2c.c index 1c9f3d20dc..dc5c828ba6 100644 --- a/hw/virtio/vhost-user-i2c.c +++ b/hw/virtio/vhost-user-i2c.c @@ -46,7 +46,7 @@ static void vu_i2c_start(VirtIODevice *vdev) i2c->vhost_dev.acked_features = vdev->guest_features; - ret = vhost_dev_start(&i2c->vhost_dev, vdev); + ret = vhost_dev_start(&i2c->vhost_dev, vdev, true); if (ret < 0) { error_report("Error starting vhost-user-i2c: %d", -ret); goto err_guest_notifiers; @@ -80,7 +80,7 @@ static void vu_i2c_stop(VirtIODevice *vdev) return; } - vhost_dev_stop(&i2c->vhost_dev, vdev); + vhost_dev_stop(&i2c->vhost_dev, vdev, true); ret = k->set_guest_notifiers(qbus->parent, i2c->vhost_dev.nvqs, false); if (ret < 0) { diff --git a/hw/virtio/vhost-user-rng.c b/hw/virtio/vhost-user-rng.c index f9084cde58..201a39e220 100644 --- a/hw/virtio/vhost-user-rng.c +++ b/hw/virtio/vhost-user-rng.c @@ -47,7 +47,7 @@ static void vu_rng_start(VirtIODevice *vdev) } rng->vhost_dev.acked_features = vdev->guest_features; - ret = vhost_dev_start(&rng->vhost_dev, vdev); + ret = vhost_dev_start(&rng->vhost_dev, vdev, true); if (ret < 0) { error_report("Error starting vhost-user-rng: %d", -ret); goto err_guest_notifiers; @@ -81,7 +81,7 @@ static void vu_rng_stop(VirtIODevice *vdev) return; } - vhost_dev_stop(&rng->vhost_dev, vdev); + vhost_dev_stop(&rng->vhost_dev, vdev, true); ret = k->set_guest_notifiers(qbus->parent, rng->vhost_dev.nvqs, false); if (ret < 0) { diff --git a/hw/virtio/vhost-user.c b/hw/virtio/vhost-user.c index abe23d4ebe..8f635844af 100644 --- a/hw/virtio/vhost-user.c +++ b/hw/virtio/vhost-user.c @@ -21,6 +21,7 @@ #include "qemu/error-report.h" #include "qemu/main-loop.h" #include "qemu/sockets.h" +#include "sysemu/runstate.h" #include "sysemu/cryptodev.h" #include "migration/migration.h" #include "migration/postcopy-ram.h" @@ -2670,6 +2671,76 @@ void vhost_user_cleanup(VhostUserState *user) user->chr = NULL; } + +typedef struct { + vu_async_close_fn cb; + DeviceState *dev; + CharBackend *cd; + struct vhost_dev *vhost; +} VhostAsyncCallback; + +static void vhost_user_async_close_bh(void *opaque) +{ + VhostAsyncCallback *data = opaque; + struct vhost_dev *vhost = data->vhost; + + /* + * If the vhost_dev has been cleared in the meantime there is + * nothing left to do as some other path has completed the + * cleanup. + */ + if (vhost->vdev) { + data->cb(data->dev); + } + + g_free(data); +} + +/* + * We only schedule the work if the machine is running. If suspended + * we want to keep all the in-flight data as is for migration + * purposes. + */ +void vhost_user_async_close(DeviceState *d, + CharBackend *chardev, struct vhost_dev *vhost, + vu_async_close_fn cb) +{ + if (!runstate_check(RUN_STATE_SHUTDOWN)) { + /* + * A close event may happen during a read/write, but vhost + * code assumes the vhost_dev remains setup, so delay the + * stop & clear. + */ + AioContext *ctx = qemu_get_current_aio_context(); + VhostAsyncCallback *data = g_new0(VhostAsyncCallback, 1); + + /* Save data for the callback */ + data->cb = cb; + data->dev = d; + data->cd = chardev; + data->vhost = vhost; + + /* Disable any further notifications on the chardev */ + qemu_chr_fe_set_handlers(chardev, + NULL, NULL, NULL, NULL, NULL, NULL, + false); + + aio_bh_schedule_oneshot(ctx, vhost_user_async_close_bh, data); + + /* + * Move vhost device to the stopped state. The vhost-user device + * will be clean up and disconnected in BH. This can be useful in + * the vhost migration code. If disconnect was caught there is an + * option for the general vhost code to get the dev state without + * knowing its type (in this case vhost-user). + * + * Note if the vhost device is fully cleared by the time we + * execute the bottom half we won't continue with the cleanup. + */ + vhost->started = false; + } +} + static int vhost_user_dev_start(struct vhost_dev *dev, bool started) { if (!virtio_has_feature(dev->protocol_features, diff --git a/hw/virtio/vhost-vsock-common.c b/hw/virtio/vhost-vsock-common.c index a67a275de2..d21c72b401 100644 --- a/hw/virtio/vhost-vsock-common.c +++ b/hw/virtio/vhost-vsock-common.c @@ -70,7 +70,7 @@ int vhost_vsock_common_start(VirtIODevice *vdev) } vvc->vhost_dev.acked_features = vdev->guest_features; - ret = vhost_dev_start(&vvc->vhost_dev, vdev); + ret = vhost_dev_start(&vvc->vhost_dev, vdev, true); if (ret < 0) { error_report("Error starting vhost: %d", -ret); goto err_guest_notifiers; @@ -105,7 +105,7 @@ void vhost_vsock_common_stop(VirtIODevice *vdev) return; } - vhost_dev_stop(&vvc->vhost_dev, vdev); + vhost_dev_stop(&vvc->vhost_dev, vdev, true); ret = k->set_guest_notifiers(qbus->parent, vvc->vhost_dev.nvqs, false); if (ret < 0) { diff --git a/hw/virtio/vhost.c b/hw/virtio/vhost.c index d1c4c20b8c..7fb008bc9e 100644 --- a/hw/virtio/vhost.c +++ b/hw/virtio/vhost.c @@ -1777,15 +1777,36 @@ int vhost_dev_get_inflight(struct vhost_dev *dev, uint16_t queue_size, return 0; } +static int vhost_dev_set_vring_enable(struct vhost_dev *hdev, int enable) +{ + if (!hdev->vhost_ops->vhost_set_vring_enable) { + return 0; + } + + /* + * For vhost-user devices, if VHOST_USER_F_PROTOCOL_FEATURES has not + * been negotiated, the rings start directly in the enabled state, and + * .vhost_set_vring_enable callback will fail since + * VHOST_USER_SET_VRING_ENABLE is not supported. + */ + if (hdev->vhost_ops->backend_type == VHOST_BACKEND_TYPE_USER && + !virtio_has_feature(hdev->backend_features, + VHOST_USER_F_PROTOCOL_FEATURES)) { + return 0; + } + + return hdev->vhost_ops->vhost_set_vring_enable(hdev, enable); +} + /* Host notifiers must be enabled at this point. */ -int vhost_dev_start(struct vhost_dev *hdev, VirtIODevice *vdev) +int vhost_dev_start(struct vhost_dev *hdev, VirtIODevice *vdev, bool vrings) { int i, r; /* should only be called after backend is connected */ assert(hdev->vhost_ops); - trace_vhost_dev_start(hdev, vdev->name); + trace_vhost_dev_start(hdev, vdev->name, vrings); vdev->vhost_started = true; hdev->started = true; @@ -1830,10 +1851,16 @@ int vhost_dev_start(struct vhost_dev *hdev, VirtIODevice *vdev) goto fail_log; } } + if (vrings) { + r = vhost_dev_set_vring_enable(hdev, true); + if (r) { + goto fail_log; + } + } if (hdev->vhost_ops->vhost_dev_start) { r = hdev->vhost_ops->vhost_dev_start(hdev, true); if (r) { - goto fail_log; + goto fail_start; } } if (vhost_dev_has_iommu(hdev) && @@ -1848,6 +1875,10 @@ int vhost_dev_start(struct vhost_dev *hdev, VirtIODevice *vdev) } } return 0; +fail_start: + if (vrings) { + vhost_dev_set_vring_enable(hdev, false); + } fail_log: vhost_log_put(hdev, false); fail_vq: @@ -1866,18 +1897,21 @@ fail_features: } /* Host notifiers must be enabled at this point. */ -void vhost_dev_stop(struct vhost_dev *hdev, VirtIODevice *vdev) +void vhost_dev_stop(struct vhost_dev *hdev, VirtIODevice *vdev, bool vrings) { int i; /* should only be called after backend is connected */ assert(hdev->vhost_ops); - trace_vhost_dev_stop(hdev, vdev->name); + trace_vhost_dev_stop(hdev, vdev->name, vrings); if (hdev->vhost_ops->vhost_dev_start) { hdev->vhost_ops->vhost_dev_start(hdev, false); } + if (vrings) { + vhost_dev_set_vring_enable(hdev, false); + } for (i = 0; i < hdev->nvqs; ++i) { vhost_virtqueue_stop(hdev, vdev, |