diff options
Diffstat (limited to 'hw/i386')
| -rw-r--r-- | hw/i386/kvm/meson.build | 1 | ||||
| -rw-r--r-- | hw/i386/kvm/trace-events | 2 | ||||
| -rw-r--r-- | hw/i386/kvm/xen-stubs.c | 8 | ||||
| -rw-r--r-- | hw/i386/kvm/xen_evtchn.c | 151 | ||||
| -rw-r--r-- | hw/i386/kvm/xen_gnttab.c | 7 | ||||
| -rw-r--r-- | hw/i386/kvm/xen_primary_console.c | 193 | ||||
| -rw-r--r-- | hw/i386/kvm/xen_primary_console.h | 23 | ||||
| -rw-r--r-- | hw/i386/kvm/xen_xenstore.c | 23 | ||||
| -rw-r--r-- | hw/i386/pc.c | 11 | ||||
| -rw-r--r-- | hw/i386/pc_piix.c | 2 | ||||
| -rw-r--r-- | hw/i386/pc_q35.c | 2 | ||||
| -rw-r--r-- | hw/i386/xen/xen_platform.c | 77 |
12 files changed, 387 insertions, 113 deletions
diff --git a/hw/i386/kvm/meson.build b/hw/i386/kvm/meson.build index ab143d6474..a4a2e23c06 100644 --- a/hw/i386/kvm/meson.build +++ b/hw/i386/kvm/meson.build @@ -9,6 +9,7 @@ i386_kvm_ss.add(when: 'CONFIG_XEN_EMU', if_true: files( 'xen_evtchn.c', 'xen_gnttab.c', 'xen_xenstore.c', + 'xen_primary_console.c', 'xenstore_impl.c', )) diff --git a/hw/i386/kvm/trace-events b/hw/i386/kvm/trace-events index e4c82de6f3..67bf7f174e 100644 --- a/hw/i386/kvm/trace-events +++ b/hw/i386/kvm/trace-events @@ -18,3 +18,5 @@ xenstore_watch(const char *path, const char *token) "path %s token %s" xenstore_unwatch(const char *path, const char *token) "path %s token %s" xenstore_reset_watches(void) "" xenstore_watch_event(const char *path, const char *token) "path %s token %s" +xen_primary_console_create(void) "" +xen_primary_console_reset(int port) "port %u" diff --git a/hw/i386/kvm/xen-stubs.c b/hw/i386/kvm/xen-stubs.c index ae406e0b02..d03131e686 100644 --- a/hw/i386/kvm/xen-stubs.c +++ b/hw/i386/kvm/xen-stubs.c @@ -15,6 +15,7 @@ #include "qapi/qapi-commands-misc-target.h" #include "xen_evtchn.h" +#include "xen_primary_console.h" void xen_evtchn_snoop_msi(PCIDevice *dev, bool is_msix, unsigned int vector, uint64_t addr, uint32_t data, bool is_masked) @@ -30,6 +31,13 @@ bool xen_evtchn_deliver_pirq_msi(uint64_t address, uint32_t data) return false; } +void xen_primary_console_create(void) +{ +} + +void xen_primary_console_set_be_port(uint16_t port) +{ +} #ifdef TARGET_I386 EvtchnInfoList *qmp_xen_event_list(Error **errp) { diff --git a/hw/i386/kvm/xen_evtchn.c b/hw/i386/kvm/xen_evtchn.c index b2b4be9983..02b8cbf8df 100644 --- a/hw/i386/kvm/xen_evtchn.c +++ b/hw/i386/kvm/xen_evtchn.c @@ -58,7 +58,15 @@ OBJECT_DECLARE_SIMPLE_TYPE(XenEvtchnState, XEN_EVTCHN) typedef struct XenEvtchnPort { uint32_t vcpu; /* Xen/ACPI vcpu_id */ uint16_t type; /* EVTCHNSTAT_xxxx */ - uint16_t type_val; /* pirq# / virq# / remote port according to type */ + union { + uint16_t val; /* raw value for serialization etc. */ + uint16_t pirq; + uint16_t virq; + struct { + uint16_t port:15; + uint16_t to_qemu:1; /* Only two targets; qemu or loopback */ + } interdomain; + } u; } XenEvtchnPort; /* 32-bit compatibility definitions, also used natively in 32-bit build */ @@ -106,14 +114,6 @@ struct xenevtchn_handle { }; /* - * For unbound/interdomain ports there are only two possible remote - * domains; self and QEMU. Use a single high bit in type_val for that, - * and the low bits for the remote port number (or 0 for unbound). - */ -#define PORT_INFO_TYPEVAL_REMOTE_QEMU 0x8000 -#define PORT_INFO_TYPEVAL_REMOTE_PORT_MASK 0x7FFF - -/* * These 'emuirq' values are used by Xen in the LM stream... and yes, I am * insane enough to think about guest-transparent live migration from actual * Xen to QEMU, and ensuring that we can convert/consume the stream. @@ -210,16 +210,16 @@ static int xen_evtchn_post_load(void *opaque, int version_id) XenEvtchnPort *p = &s->port_table[i]; if (p->type == EVTCHNSTAT_pirq) { - assert(p->type_val); - assert(p->type_val < s->nr_pirqs); + assert(p->u.pirq); + assert(p->u.pirq < s->nr_pirqs); /* * Set the gsi to IRQ_UNBOUND; it may be changed to an actual * GSI# below, or to IRQ_MSI_EMU when the MSI table snooping * catches up with it. */ - s->pirq[p->type_val].gsi = IRQ_UNBOUND; - s->pirq[p->type_val].port = i; + s->pirq[p->u.pirq].gsi = IRQ_UNBOUND; + s->pirq[p->u.pirq].port = i; } } /* Rebuild s->pirq[].gsi mapping */ @@ -243,7 +243,7 @@ static const VMStateDescription xen_evtchn_port_vmstate = { .fields = (VMStateField[]) { VMSTATE_UINT32(vcpu, XenEvtchnPort), VMSTATE_UINT16(type, XenEvtchnPort), - VMSTATE_UINT16(type_val, XenEvtchnPort), + VMSTATE_UINT16(u.val, XenEvtchnPort), VMSTATE_END_OF_LIST() } }; @@ -605,14 +605,13 @@ static void unbind_backend_ports(XenEvtchnState *s) for (i = 1; i < s->nr_ports; i++) { p = &s->port_table[i]; - if (p->type == EVTCHNSTAT_interdomain && - (p->type_val & PORT_INFO_TYPEVAL_REMOTE_QEMU)) { - evtchn_port_t be_port = p->type_val & PORT_INFO_TYPEVAL_REMOTE_PORT_MASK; + if (p->type == EVTCHNSTAT_interdomain && p->u.interdomain.to_qemu) { + evtchn_port_t be_port = p->u.interdomain.port; if (s->be_handles[be_port]) { /* This part will be overwritten on the load anyway. */ p->type = EVTCHNSTAT_unbound; - p->type_val = PORT_INFO_TYPEVAL_REMOTE_QEMU; + p->u.interdomain.port = 0; /* Leave the backend port open and unbound too. */ if (kvm_xen_has_cap(EVTCHN_SEND)) { @@ -650,30 +649,22 @@ int xen_evtchn_status_op(struct evtchn_status *status) switch (p->type) { case EVTCHNSTAT_unbound: - if (p->type_val & PORT_INFO_TYPEVAL_REMOTE_QEMU) { - status->u.unbound.dom = DOMID_QEMU; - } else { - status->u.unbound.dom = xen_domid; - } + status->u.unbound.dom = p->u.interdomain.to_qemu ? DOMID_QEMU + : xen_domid; break; case EVTCHNSTAT_interdomain: - if (p->type_val & PORT_INFO_TYPEVAL_REMOTE_QEMU) { - status->u.interdomain.dom = DOMID_QEMU; - } else { - status->u.interdomain.dom = xen_domid; - } - - status->u.interdomain.port = p->type_val & - PORT_INFO_TYPEVAL_REMOTE_PORT_MASK; + status->u.interdomain.dom = p->u.interdomain.to_qemu ? DOMID_QEMU + : xen_domid; + status->u.interdomain.port = p->u.interdomain.port; break; case EVTCHNSTAT_pirq: - status->u.pirq = p->type_val; + status->u.pirq = p->u.pirq; break; case EVTCHNSTAT_virq: - status->u.virq = p->type_val; + status->u.virq = p->u.virq; break; } @@ -989,7 +980,7 @@ static int clear_port_pending(XenEvtchnState *s, evtchn_port_t port) static void free_port(XenEvtchnState *s, evtchn_port_t port) { s->port_table[port].type = EVTCHNSTAT_closed; - s->port_table[port].type_val = 0; + s->port_table[port].u.val = 0; s->port_table[port].vcpu = 0; if (s->nr_ports == port + 1) { @@ -1012,7 +1003,7 @@ static int allocate_port(XenEvtchnState *s, uint32_t vcpu, uint16_t type, if (s->port_table[p].type == EVTCHNSTAT_closed) { s->port_table[p].vcpu = vcpu; s->port_table[p].type = type; - s->port_table[p].type_val = val; + s->port_table[p].u.val = val; *port = p; @@ -1053,15 +1044,15 @@ static int close_port(XenEvtchnState *s, evtchn_port_t port, return -ENOENT; case EVTCHNSTAT_pirq: - s->pirq[p->type_val].port = 0; - if (s->pirq[p->type_val].is_translated) { + s->pirq[p->u.pirq].port = 0; + if (s->pirq[p->u.pirq].is_translated) { *flush_kvm_routes = true; } break; case EVTCHNSTAT_virq: - kvm_xen_set_vcpu_virq(virq_is_global(p->type_val) ? 0 : p->vcpu, - p->type_val, 0); + kvm_xen_set_vcpu_virq(virq_is_global(p->u.virq) ? 0 : p->vcpu, + p->u.virq, 0); break; case EVTCHNSTAT_ipi: @@ -1071,8 +1062,8 @@ static int close_port(XenEvtchnState *s, evtchn_port_t port, break; case EVTCHNSTAT_interdomain: - if (p->type_val & PORT_INFO_TYPEVAL_REMOTE_QEMU) { - uint16_t be_port = p->type_val & ~PORT_INFO_TYPEVAL_REMOTE_QEMU; + if (p->u.interdomain.to_qemu) { + uint16_t be_port = p->u.interdomain.port; struct xenevtchn_handle *xc = s->be_handles[be_port]; if (xc) { if (kvm_xen_has_cap(EVTCHN_SEND)) { @@ -1082,14 +1073,15 @@ static int close_port(XenEvtchnState *s, evtchn_port_t port, } } else { /* Loopback interdomain */ - XenEvtchnPort *rp = &s->port_table[p->type_val]; - if (!valid_port(p->type_val) || rp->type_val != port || + XenEvtchnPort *rp = &s->port_table[p->u.interdomain.port]; + if (!valid_port(p->u.interdomain.port) || + rp->u.interdomain.port != port || rp->type != EVTCHNSTAT_interdomain) { error_report("Inconsistent state for interdomain unbind"); } else { /* Set the other end back to unbound */ rp->type = EVTCHNSTAT_unbound; - rp->type_val = 0; + rp->u.interdomain.port = 0; } } break; @@ -1214,7 +1206,7 @@ int xen_evtchn_bind_vcpu_op(struct evtchn_bind_vcpu *vcpu) if (p->type == EVTCHNSTAT_interdomain || p->type == EVTCHNSTAT_unbound || p->type == EVTCHNSTAT_pirq || - (p->type == EVTCHNSTAT_virq && virq_is_global(p->type_val))) { + (p->type == EVTCHNSTAT_virq && virq_is_global(p->u.virq))) { /* * unmask_port() with do_unmask==false will just raise the event * on the new vCPU if the port was already pending. @@ -1359,19 +1351,15 @@ int xen_evtchn_bind_ipi_op(struct evtchn_bind_ipi *ipi) int xen_evtchn_bind_interdomain_op(struct evtchn_bind_interdomain *interdomain) { XenEvtchnState *s = xen_evtchn_singleton; - uint16_t type_val; int ret; if (!s) { return -ENOTSUP; } - if (interdomain->remote_dom == DOMID_QEMU) { - type_val = PORT_INFO_TYPEVAL_REMOTE_QEMU; - } else if (interdomain->remote_dom == DOMID_SELF || - interdomain->remote_dom == xen_domid) { - type_val = 0; - } else { + if (interdomain->remote_dom != DOMID_QEMU && + interdomain->remote_dom != DOMID_SELF && + interdomain->remote_dom != xen_domid) { return -ESRCH; } @@ -1382,8 +1370,8 @@ int xen_evtchn_bind_interdomain_op(struct evtchn_bind_interdomain *interdomain) qemu_mutex_lock(&s->port_lock); /* The newly allocated port starts out as unbound */ - ret = allocate_port(s, 0, EVTCHNSTAT_unbound, type_val, - &interdomain->local_port); + ret = allocate_port(s, 0, EVTCHNSTAT_unbound, 0, &interdomain->local_port); + if (ret) { goto out; } @@ -1408,7 +1396,8 @@ int xen_evtchn_bind_interdomain_op(struct evtchn_bind_interdomain *interdomain) assign_kernel_eventfd(lp->type, xc->guest_port, xc->fd); } lp->type = EVTCHNSTAT_interdomain; - lp->type_val = PORT_INFO_TYPEVAL_REMOTE_QEMU | interdomain->remote_port; + lp->u.interdomain.to_qemu = 1; + lp->u.interdomain.port = interdomain->remote_port; ret = 0; } else { /* Loopback */ @@ -1416,19 +1405,18 @@ int xen_evtchn_bind_interdomain_op(struct evtchn_bind_interdomain *interdomain) XenEvtchnPort *lp = &s->port_table[interdomain->local_port]; /* - * The 'remote' port for loopback must be an unbound port allocated for - * communication with the local domain (as indicated by rp->type_val - * being zero, not PORT_INFO_TYPEVAL_REMOTE_QEMU), and must *not* be - * the port that was just allocated for the local end. + * The 'remote' port for loopback must be an unbound port allocated + * for communication with the local domain, and must *not* be the + * port that was just allocated for the local end. */ if (interdomain->local_port != interdomain->remote_port && - rp->type == EVTCHNSTAT_unbound && rp->type_val == 0) { + rp->type == EVTCHNSTAT_unbound && !rp->u.interdomain.to_qemu) { rp->type = EVTCHNSTAT_interdomain; - rp->type_val = interdomain->local_port; + rp->u.interdomain.port = interdomain->local_port; lp->type = EVTCHNSTAT_interdomain; - lp->type_val = interdomain->remote_port; + lp->u.interdomain.port = interdomain->remote_port; } else { ret = -EINVAL; } @@ -1447,7 +1435,6 @@ int xen_evtchn_bind_interdomain_op(struct evtchn_bind_interdomain *interdomain) int xen_evtchn_alloc_unbound_op(struct evtchn_alloc_unbound *alloc) { XenEvtchnState *s = xen_evtchn_singleton; - uint16_t type_val; int ret; if (!s) { @@ -1458,18 +1445,20 @@ int xen_evtchn_alloc_unbound_op(struct evtchn_alloc_unbound *alloc) return -ESRCH; } - if (alloc->remote_dom == DOMID_QEMU) { - type_val = PORT_INFO_TYPEVAL_REMOTE_QEMU; - } else if (alloc->remote_dom == DOMID_SELF || - alloc->remote_dom == xen_domid) { - type_val = 0; - } else { + if (alloc->remote_dom != DOMID_QEMU && + alloc->remote_dom != DOMID_SELF && + alloc->remote_dom != xen_domid) { return -EPERM; } qemu_mutex_lock(&s->port_lock); - ret = allocate_port(s, 0, EVTCHNSTAT_unbound, type_val, &alloc->port); + ret = allocate_port(s, 0, EVTCHNSTAT_unbound, 0, &alloc->port); + + if (!ret && alloc->remote_dom == DOMID_QEMU) { + XenEvtchnPort *p = &s->port_table[alloc->port]; + p->u.interdomain.to_qemu = 1; + } qemu_mutex_unlock(&s->port_lock); @@ -1496,12 +1485,12 @@ int xen_evtchn_send_op(struct evtchn_send *send) switch (p->type) { case EVTCHNSTAT_interdomain: - if (p->type_val & PORT_INFO_TYPEVAL_REMOTE_QEMU) { + if (p->u.interdomain.to_qemu) { /* * This is an event from the guest to qemu itself, which is * serving as the driver domain. */ - uint16_t be_port = p->type_val & ~PORT_INFO_TYPEVAL_REMOTE_QEMU; + uint16_t be_port = p->u.interdomain.port; struct xenevtchn_handle *xc = s->be_handles[be_port]; if (xc) { eventfd_write(xc->fd, 1); @@ -1511,7 +1500,7 @@ int xen_evtchn_send_op(struct evtchn_send *send) } } else { /* Loopback interdomain ports; just a complex IPI */ - set_port_pending(s, p->type_val); + set_port_pending(s, p->u.interdomain.port); } break; @@ -1553,8 +1542,7 @@ int xen_evtchn_set_port(uint16_t port) /* QEMU has no business sending to anything but these */ if (p->type == EVTCHNSTAT_virq || - (p->type == EVTCHNSTAT_interdomain && - (p->type_val & PORT_INFO_TYPEVAL_REMOTE_QEMU))) { + (p->type == EVTCHNSTAT_interdomain && p->u.interdomain.to_qemu)) { set_port_pending(s, port); ret = 0; } @@ -2064,7 +2052,7 @@ int xen_be_evtchn_bind_interdomain(struct xenevtchn_handle *xc, uint32_t domid, switch (gp->type) { case EVTCHNSTAT_interdomain: /* Allow rebinding after migration, preserve port # if possible */ - be_port = gp->type_val & ~PORT_INFO_TYPEVAL_REMOTE_QEMU; + be_port = gp->u.interdomain.port; assert(be_port != 0); if (!s->be_handles[be_port]) { s->be_handles[be_port] = xc; @@ -2085,7 +2073,8 @@ int xen_be_evtchn_bind_interdomain(struct xenevtchn_handle *xc, uint32_t domid, } gp->type = EVTCHNSTAT_interdomain; - gp->type_val = be_port | PORT_INFO_TYPEVAL_REMOTE_QEMU; + gp->u.interdomain.to_qemu = 1; + gp->u.interdomain.port = be_port; xc->guest_port = guest_port; if (kvm_xen_has_cap(EVTCHN_SEND)) { assign_kernel_eventfd(gp->type, guest_port, xc->fd); @@ -2130,7 +2119,7 @@ int xen_be_evtchn_unbind(struct xenevtchn_handle *xc, evtchn_port_t port) /* This should never *not* be true */ if (gp->type == EVTCHNSTAT_interdomain) { gp->type = EVTCHNSTAT_unbound; - gp->type_val = PORT_INFO_TYPEVAL_REMOTE_QEMU; + gp->u.interdomain.port = 0; } if (kvm_xen_has_cap(EVTCHN_SEND)) { @@ -2284,11 +2273,11 @@ EvtchnInfoList *qmp_xen_event_list(Error **errp) info->type = p->type; if (p->type == EVTCHNSTAT_interdomain) { - info->remote_domain = g_strdup((p->type_val & PORT_INFO_TYPEVAL_REMOTE_QEMU) ? + info->remote_domain = g_strdup(p->u.interdomain.to_qemu ? "qemu" : "loopback"); - info->target = p->type_val & PORT_INFO_TYPEVAL_REMOTE_PORT_MASK; + info->target = p->u.interdomain.port; } else { - info->target = p->type_val; + info->target = p->u.val; /* pirq# or virq# */ } info->vcpu = p->vcpu; info->pending = test_bit(i, pending); diff --git a/hw/i386/kvm/xen_gnttab.c b/hw/i386/kvm/xen_gnttab.c index 839ec920a1..0a24f53f20 100644 --- a/hw/i386/kvm/xen_gnttab.c +++ b/hw/i386/kvm/xen_gnttab.c @@ -25,6 +25,7 @@ #include "hw/xen/xen_backend_ops.h" #include "xen_overlay.h" #include "xen_gnttab.h" +#include "xen_primary_console.h" #include "sysemu/kvm.h" #include "sysemu/kvm_xen.h" @@ -537,9 +538,13 @@ int xen_gnttab_reset(void) s->nr_frames = 0; memset(s->entries.v1, 0, XEN_PAGE_SIZE * s->max_frames); - s->entries.v1[GNTTAB_RESERVED_XENSTORE].flags = GTF_permit_access; s->entries.v1[GNTTAB_RESERVED_XENSTORE].frame = XEN_SPECIAL_PFN(XENSTORE); + if (xen_primary_console_get_pfn()) { + s->entries.v1[GNTTAB_RESERVED_CONSOLE].flags = GTF_permit_access; + s->entries.v1[GNTTAB_RESERVED_CONSOLE].frame = XEN_SPECIAL_PFN(CONSOLE); + } + return 0; } diff --git a/hw/i386/kvm/xen_primary_console.c b/hw/i386/kvm/xen_primary_console.c new file mode 100644 index 0000000000..abe79f565b --- /dev/null +++ b/hw/i386/kvm/xen_primary_console.c @@ -0,0 +1,193 @@ +/* + * QEMU Xen emulation: Primary console support + * + * Copyright © 2023 Amazon.com, Inc. or its affiliates. All Rights Reserved. + * + * Authors: David Woodhouse <dwmw2@infradead.org> + * + * This work is licensed under the terms of the GNU GPL, version 2 or later. + * See the COPYING file in the top-level directory. + */ + +#include "qemu/osdep.h" + +#include "qapi/error.h" + +#include "hw/sysbus.h" +#include "hw/xen/xen.h" +#include "hw/xen/xen_backend_ops.h" +#include "xen_evtchn.h" +#include "xen_overlay.h" +#include "xen_primary_console.h" + +#include "sysemu/kvm.h" +#include "sysemu/kvm_xen.h" + +#include "trace.h" + +#include "hw/xen/interface/event_channel.h" +#include "hw/xen/interface/grant_table.h" + +#define TYPE_XEN_PRIMARY_CONSOLE "xen-primary-console" +OBJECT_DECLARE_SIMPLE_TYPE(XenPrimaryConsoleState, XEN_PRIMARY_CONSOLE) + +struct XenPrimaryConsoleState { + /*< private >*/ + SysBusDevice busdev; + /*< public >*/ + + MemoryRegion console_page; + void *cp; + + evtchn_port_t guest_port; + evtchn_port_t be_port; + + struct xengntdev_handle *gt; + void *granted_xs; +}; + +struct XenPrimaryConsoleState *xen_primary_console_singleton; + +static void xen_primary_console_realize(DeviceState *dev, Error **errp) +{ + XenPrimaryConsoleState *s = XEN_PRIMARY_CONSOLE(dev); + + if (xen_mode != XEN_EMULATE) { + error_setg(errp, "Xen primary console support is for Xen emulation"); + return; + } + + memory_region_init_ram(&s->console_page, OBJECT(dev), "xen:console_page", + XEN_PAGE_SIZE, &error_abort); + memory_region_set_enabled(&s->console_page, true); + s->cp = memory_region_get_ram_ptr(&s->console_page); + memset(s->cp, 0, XEN_PAGE_SIZE); + + /* We can't map it this early as KVM isn't ready */ + xen_primary_console_singleton = s; +} + +static void xen_primary_console_class_init(ObjectClass *klass, void *data) +{ + DeviceClass *dc = DEVICE_CLASS(klass); + + dc->realize = xen_primary_console_realize; +} + +static const TypeInfo xen_primary_console_info = { + .name = TYPE_XEN_PRIMARY_CONSOLE, + .parent = TYPE_SYS_BUS_DEVICE, + .instance_size = sizeof(XenPrimaryConsoleState), + .class_init = xen_primary_console_class_init, +}; + + +void xen_primary_console_create(void) +{ + DeviceState *dev = sysbus_create_simple(TYPE_XEN_PRIMARY_CONSOLE, -1, NULL); + + trace_xen_primary_console_create(); + + xen_primary_console_singleton = XEN_PRIMARY_CONSOLE(dev); + + /* + * Defer the init (xen_primary_console_reset()) until KVM is set up and the + * overlay page can be mapped. + */ +} + +static void xen_primary_console_register_types(void) +{ + type_register_static(&xen_primary_console_info); +} + +type_init(xen_primary_console_register_types) + +uint16_t xen_primary_console_get_port(void) +{ + XenPrimaryConsoleState *s = xen_primary_console_singleton; + if (!s) { + return 0; + } + return s->guest_port; +} + +void xen_primary_console_set_be_port(uint16_t port) +{ + XenPrimaryConsoleState *s = xen_primary_console_singleton; + if (s) { + s->be_port = port; + } +} + +uint64_t xen_primary_console_get_pfn(void) +{ + XenPrimaryConsoleState *s = xen_primary_console_singleton; + if (!s) { + return 0; + } + return XEN_SPECIAL_PFN(CONSOLE); +} + +void *xen_primary_console_get_map(void) +{ + XenPrimaryConsoleState *s = xen_primary_console_singleton; + if (!s) { + return 0; + } + return s->cp; +} + +static void alloc_guest_port(XenPrimaryConsoleState *s) +{ + struct evtchn_alloc_unbound alloc = { + .dom = DOMID_SELF, + .remote_dom = DOMID_QEMU, + }; + + if (!xen_evtchn_alloc_unbound_op(&alloc)) { + s->guest_port = alloc.port; + } +} + +static void rebind_guest_port(XenPrimaryConsoleState *s) +{ + struct evtchn_bind_interdomain inter = { + .remote_dom = DOMID_QEMU, + .remote_port = s->be_port, + }; + + if (!xen_evtchn_bind_interdomain_op(&inter)) { + s->guest_port = inter.local_port; + } + + s->be_port = 0; +} + +int xen_primary_console_reset(void) +{ + XenPrimaryConsoleState *s = xen_primary_console_singleton; + if (!s) { + return 0; + } + + if (!memory_region_is_mapped(&s->console_page)) { + uint64_t gpa = XEN_SPECIAL_PFN(CONSOLE) << TARGET_PAGE_BITS; + xen_overlay_do_map_page(&s->console_page, gpa); + } + + if (s->be_port) { + rebind_guest_port(s); + } else { + alloc_guest_port(s); + } + + trace_xen_primary_console_reset(s->guest_port); + + s->gt = qemu_xen_gnttab_open(); + uint32_t xs_gntref = GNTTAB_RESERVED_CONSOLE; + s->granted_xs = qemu_xen_gnttab_map_refs(s->gt, 1, xen_domid, &xs_gntref, + PROT_READ | PROT_WRITE); + + return 0; +} diff --git a/hw/i386/kvm/xen_primary_console.h b/hw/i386/kvm/xen_primary_console.h new file mode 100644 index 0000000000..7e2989ea0d --- /dev/null +++ b/hw/i386/kvm/xen_primary_console.h @@ -0,0 +1,23 @@ +/* + * QEMU Xen emulation: Primary console support + * + * Copyright © 2023 Amazon.com, Inc. or its affiliates. All Rights Reserved. + * + * Authors: David Woodhouse <dwmw2@infradead.org> + * + * This work is licensed under the terms of the GNU GPL, version 2 or later. + * See the COPYING file in the top-level directory. + */ + +#ifndef QEMU_XEN_PRIMARY_CONSOLE_H +#define QEMU_XEN_PRIMARY_CONSOLE_H + +void xen_primary_console_create(void); +int xen_primary_console_reset(void); + +uint16_t xen_primary_console_get_port(void); +void xen_primary_console_set_be_port(uint16_t port); +uint64_t xen_primary_console_get_pfn(void); +void *xen_primary_console_get_map(void); + +#endif /* QEMU_XEN_PRIMARY_CONSOLE_H */ diff --git a/hw/i386/kvm/xen_xenstore.c b/hw/i386/kvm/xen_xenstore.c index 8e716a7009..6e651960b3 100644 --- a/hw/i386/kvm/xen_xenstore.c +++ b/hw/i386/kvm/xen_xenstore.c @@ -25,6 +25,7 @@ #include "hw/xen/xen_backend_ops.h" #include "xen_overlay.h" #include "xen_evtchn.h" +#include "xen_primary_console.h" #include "xen_xenstore.h" #include "sysemu/kvm.h" @@ -331,7 +332,7 @@ static void xs_error(XenXenstoreState *s, unsigned int id, const char *errstr = NULL; for (unsigned int i = 0; i < ARRAY_SIZE(xsd_errors); i++) { - struct xsd_errors *xsd_error = &xsd_errors[i]; + const struct xsd_errors *xsd_error = &xsd_errors[i]; if (xsd_error->errnum == errnum) { errstr = xsd_error->errstring; @@ -1434,6 +1435,8 @@ static void alloc_guest_port(XenXenstoreState *s) int xen_xenstore_reset(void) { XenXenstoreState *s = xen_xenstore_singleton; + int console_port; + GList *perms; int err; if (!s) { @@ -1461,6 +1464,24 @@ int xen_xenstore_reset(void) } s->be_port = err; + /* Create frontend store nodes */ + perms = g_list_append(NULL, xs_perm_as_string(XS_PERM_NONE, DOMID_QEMU)); + perms = g_list_append(perms, xs_perm_as_string(XS_PERM_READ, xen_domid)); + + relpath_printf(s, perms, "store/port", "%u", s->guest_port); + relpath_printf(s, perms, "store/ring-ref", "%lu", + XEN_SPECIAL_PFN(XENSTORE)); + + console_port = xen_primary_console_get_port(); + if (console_port) { + relpath_printf(s, perms, "console/ring-ref", "%lu", + XEN_SPECIAL_PFN(CONSOLE)); + relpath_printf(s, perms, "console/port", "%u", console_port); + relpath_printf(s, perms, "console/state", "%u", XenbusStateInitialised); + } + + g_list_free_full(perms, g_free); + /* * We don't actually access the guest's page through the grant, because * this isn't real Xen, and we can just use the page we gave it in the diff --git a/hw/i386/pc.c b/hw/i386/pc.c index 1aef21aa2c..188bc9d0f8 100644 --- a/hw/i386/pc.c +++ b/hw/i386/pc.c @@ -1261,7 +1261,7 @@ void pc_basic_device_init(struct PCMachineState *pcms, if (pcms->bus) { pci_create_simple(pcms->bus, -1, "xen-platform"); } - xen_bus_init(); + pcms->xenbus = xen_bus_init(); xen_be_init(); } #endif @@ -1289,7 +1289,8 @@ void pc_basic_device_init(struct PCMachineState *pcms, pcms->vmport != ON_OFF_AUTO_ON); } -void pc_nic_init(PCMachineClass *pcmc, ISABus *isa_bus, PCIBus *pci_bus) +void pc_nic_init(PCMachineClass *pcmc, ISABus *isa_bus, PCIBus *pci_bus, + BusState *xen_bus) { MachineClass *mc = MACHINE_CLASS(pcmc); int i; @@ -1299,7 +1300,11 @@ void pc_nic_init(PCMachineClass *pcmc, ISABus *isa_bus, PCIBus *pci_bus) NICInfo *nd = &nd_table[i]; const char *model = nd->model ? nd->model : mc->default_nic; - if (g_str_equal(model, "ne2k_isa")) { + if (xen_bus && (!nd->model || g_str_equal(model, "xen-net-device"))) { + DeviceState *dev = qdev_new("xen-net-device"); + qdev_set_nic_properties(dev, nd); + qdev_realize_and_unref(dev, xen_bus, &error_fatal); + } else if (g_str_equal(model, "ne2k_isa")) { pc_init_ne2k_isa(isa_bus, nd); } else { pci_nic_init_nofail(nd, pci_bus, model, NULL); diff --git a/hw/i386/pc_piix.c b/hw/i386/pc_piix.c index 26e161beb9..eace854335 100644 --- a/hw/i386/pc_piix.c +++ b/hw/i386/pc_piix.c @@ -342,7 +342,7 @@ static void pc_init1(MachineState *machine, pc_basic_device_init(pcms, isa_bus, x86ms->gsi, rtc_state, true, 0x4); - pc_nic_init(pcmc, isa_bus, pci_bus); + pc_nic_init(pcmc, isa_bus, pci_bus, pcms->xenbus); if (pcmc->pci_enabled) { pc_cmos_init(pcms, idebus[0], idebus[1], rtc_state); diff --git a/hw/i386/pc_q35.c b/hw/i386/pc_q35.c index 597943ff1b..4f3e5412f6 100644 --- a/hw/i386/pc_q35.c +++ b/hw/i386/pc_q35.c @@ -340,7 +340,7 @@ static void pc_q35_init(MachineState *machine) /* the rest devices to which pci devfn is automatically assigned */ pc_vga_init(isa_bus, host_bus); - pc_nic_init(pcmc, isa_bus, host_bus); + pc_nic_init(pcmc, isa_bus, host_bus, pcms->xenbus); if (machine->nvdimms_state->is_enabled) { nvdimm_init_acpi_state(machine->nvdimms_state, system_io, diff --git a/hw/i386/xen/xen_platform.c b/hw/i386/xen/xen_platform.c index 17457ff3de..ef7d3fc05f 100644 --- a/hw/i386/xen/xen_platform.c +++ b/hw/i386/xen/xen_platform.c @@ -140,9 +140,14 @@ static void unplug_nic(PCIBus *b, PCIDevice *d, void *o) /* Remove the peer of the NIC device. Normally, this would be a tap device. */ static void del_nic_peer(NICState *nic, void *opaque) { - NetClientState *nc; + NetClientState *nc = qemu_get_queue(nic); + ObjectClass *klass = module_object_class_by_name(nc->model); + + /* Only delete peers of PCI NICs that we're about to delete */ + if (!klass || !object_class_dynamic_cast(klass, TYPE_PCI_DEVICE)) { + return; + } - nc = qemu_get_queue(nic); if (nc->peer) qemu_del_net_client(nc->peer); } @@ -164,39 +169,60 @@ static void pci_unplug_nics(PCIBus *bus) * * [1] https://xenbits.xen.org/gitweb/?p=xen.git;a=blob;f=docs/misc/hvm-emulated-unplug.pandoc */ -static void pci_xen_ide_unplug(PCIDevice *d, bool aux) +struct ide_unplug_state { + bool aux; + int nr_unplugged; +}; + +static int ide_dev_unplug(DeviceState *dev, void *_st) { - DeviceState *dev = DEVICE(d); - PCIIDEState *pci_ide; - int i; + struct ide_unplug_state *st = _st; IDEDevice *idedev; IDEBus *idebus; BlockBackend *blk; + int unit; - pci_ide = PCI_IDE(dev); + idedev = IDE_DEVICE(object_dynamic_cast(OBJECT(dev), "ide-hd")); + if (!idedev) { + return 0; + } - for (i = aux ? 1 : 0; i < 4; i++) { - idebus = &pci_ide->bus[i / 2]; - blk = idebus->ifs[i % 2].blk; + idebus = IDE_BUS(qdev_get_parent_bus(dev)); - if (blk && idebus->ifs[i % 2].drive_kind != IDE_CD) { - if (!(i % 2)) { - idedev = idebus->master; - } else { - idedev = idebus->slave; - } + unit = (idedev == idebus->slave); + assert(unit || idedev == idebus->master); - blk_drain(blk); - blk_flush(blk); + if (st->aux && !unit && !strcmp(BUS(idebus)->name, "ide.0")) { + return 0; + } - blk_detach_dev(blk, DEVICE(idedev)); - idebus->ifs[i % 2].blk = NULL; - idedev->conf.blk = NULL; - monitor_remove_blk(blk); - blk_unref(blk); - } + blk = idebus->ifs[unit].blk; + if (blk) { + blk_drain(blk); + blk_flush(blk); + + blk_detach_dev(blk, DEVICE(idedev)); + idebus->ifs[unit].blk = NULL; + idedev->conf.blk = NULL; + monitor_remove_blk(blk); + blk_unref(blk); + } + + object_unparent(OBJECT(dev)); + st->nr_unplugged++; + + return 0; +} + +static void pci_xen_ide_unplug(PCIDevice *d, bool aux) +{ + struct ide_unplug_state st = { aux, 0 }; + DeviceState *dev = DEVICE(d); + + qdev_walk_children(dev, NULL, NULL, ide_dev_unplug, NULL, &st); + if (st.nr_unplugged) { + pci_device_reset(d); } - pci_device_reset(d); } static void unplug_disks(PCIBus *b, PCIDevice *d, void *opaque) @@ -211,6 +237,7 @@ static void unplug_disks(PCIBus *b, PCIDevice *d, void *opaque) switch (pci_get_word(d->config + PCI_CLASS_DEVICE)) { case PCI_CLASS_STORAGE_IDE: + case PCI_CLASS_STORAGE_SATA: pci_xen_ide_unplug(d, aux); break; |