Edit
IABSD.fr/src/sys/dev/pv/hyperv.c
jsg
- sys/dev/pv/hyperv.c
-
/*- * Copyright (c) 2009-2012 Microsoft Corp. * Copyright (c) 2012 NetApp Inc. * Copyright (c) 2012 Citrix Inc. * Copyright (c) 2016 Mike Belopuhov <mike@esdenera.com> * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice unmodified, this list of conditions, and the following * disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ /* * The OpenBSD port was done under funding by Esdenera Networks GmbH. */ #include <sys/param.h> /* Hyperv requires locked atomic operations */ #ifndef MULTIPROCESSOR #define _HYPERVMPATOMICS #define MULTIPROCESSOR #endif #include <sys/atomic.h> #ifdef _HYPERVMPATOMICS #undef MULTIPROCESSOR #undef _HYPERVMPATOMICS #endif #include <sys/systm.h> #include <sys/malloc.h> #include <sys/device.h> #include <sys/timetc.h> #include <sys/task.h> #include <machine/bus.h> #include <machine/cpu.h> #include <machine/cpufunc.h> #include <uvm/uvm_extern.h> #include <machine/i82489var.h> #include <dev/pv/pvvar.h> #include <dev/pv/hypervreg.h> #include <dev/pv/hypervvar.h> /* Command submission flags */ #define HCF_SLEEPOK 0x0001 /* M_WAITOK */ #define HCF_NOSLEEP 0x0002 /* M_NOWAIT */ #define HCF_NOREPLY 0x0004 struct hv_softc *hv_sc; int hv_match(struct device *, void *, void *); void hv_attach(struct device *, struct device *, void *); void hv_set_version(struct hv_softc *); u_int hv_gettime(struct timecounter *); int hv_init_hypercall(struct hv_softc *); uint64_t hv_hypercall(struct hv_softc *, uint64_t, void *, void *); int hv_init_interrupts(struct hv_softc *); int hv_init_synic(struct hv_softc *); int hv_cmd(struct hv_softc *, void *, size_t, void *, size_t, int); int hv_start(struct hv_softc *, struct hv_msg *); int hv_reply(struct hv_softc *, struct hv_msg *); void hv_wait(struct hv_softc *, int (*done)(struct hv_softc *, struct hv_msg *), struct hv_msg *, void *, const char *); uint16_t hv_intr_signal(struct hv_softc *, void *); void hv_intr(void); void hv_event_intr(struct hv_softc *); void hv_message_intr(struct hv_softc *); int hv_vmbus_connect(struct hv_softc *); void hv_channel_response(struct hv_softc *, struct vmbus_chanmsg_hdr *); void hv_channel_offer(struct hv_softc *, struct vmbus_chanmsg_hdr *); void hv_channel_rescind(struct hv_softc *, struct vmbus_chanmsg_hdr *); void hv_channel_delivered(struct hv_softc *, struct vmbus_chanmsg_hdr *); int hv_channel_scan(struct hv_softc *); void hv_process_offer(struct hv_softc *, struct hv_offer *); struct hv_channel * hv_channel_lookup(struct hv_softc *, uint32_t); int hv_channel_ring_create(struct hv_channel *, uint32_t); void hv_channel_ring_destroy(struct hv_channel *); void hv_channel_pause(struct hv_channel *); uint hv_channel_unpause(struct hv_channel *); uint hv_channel_ready(struct hv_channel *); extern void hv_attach_icdevs(struct hv_softc *); int hv_attach_devices(struct hv_softc *); struct { int hmd_response; int hmd_request; void (*hmd_handler)(struct hv_softc *, struct vmbus_chanmsg_hdr *); } hv_msg_dispatch[] = { { 0, 0, NULL }, { VMBUS_CHANMSG_CHOFFER, 0, hv_channel_offer }, { VMBUS_CHANMSG_CHRESCIND, 0, hv_channel_rescind }, { VMBUS_CHANMSG_CHREQUEST, VMBUS_CHANMSG_CHOFFER, NULL }, { VMBUS_CHANMSG_CHOFFER_DONE, 0, hv_channel_delivered }, { VMBUS_CHANMSG_CHOPEN, 0, NULL }, { VMBUS_CHANMSG_CHOPEN_RESP, VMBUS_CHANMSG_CHOPEN, hv_channel_response }, { VMBUS_CHANMSG_CHCLOSE, 0, NULL }, { VMBUS_CHANMSG_GPADL_CONN, 0, NULL }, { VMBUS_CHANMSG_GPADL_SUBCONN, 0, NULL }, { VMBUS_CHANMSG_GPADL_CONNRESP, VMBUS_CHANMSG_GPADL_CONN, hv_channel_response }, { VMBUS_CHANMSG_GPADL_DISCONN, 0, NULL }, { VMBUS_CHANMSG_GPADL_DISCONNRESP, VMBUS_CHANMSG_GPADL_DISCONN, hv_channel_response }, { VMBUS_CHANMSG_CHFREE, 0, NULL }, { VMBUS_CHANMSG_CONNECT, 0, NULL }, { VMBUS_CHANMSG_CONNECT_RESP, VMBUS_CHANMSG_CONNECT, hv_channel_response }, { VMBUS_CHANMSG_DISCONNECT, 0, NULL }, }; struct timecounter hv_timecounter = { .tc_get_timecount = hv_gettime, .tc_counter_mask = 0xffffffff, .tc_frequency = 10000000, .tc_name = "hyperv", .tc_quality = 9001, .tc_priv = NULL, .tc_user = 0, }; struct cfdriver hyperv_cd = { NULL, "hyperv", DV_DULL }; const struct cfattach hyperv_ca = { sizeof(struct hv_softc), hv_match, hv_attach }; const struct hv_guid hv_guid_network = { { 0x63, 0x51, 0x61, 0xf8, 0x3e, 0xdf, 0xc5, 0x46, 0x91, 0x3f, 0xf2, 0xd2, 0xf9, 0x65, 0xed, 0x0e } }; const struct hv_guid hv_guid_ide = { { 0x32, 0x26, 0x41, 0x32, 0xcb, 0x86, 0xa2, 0x44, 0x9b, 0x5c, 0x50, 0xd1, 0x41, 0x73, 0x54, 0xf5 } }; const struct hv_guid hv_guid_scsi = { { 0xd9, 0x63, 0x61, 0xba, 0xa1, 0x04, 0x29, 0x4d, 0xb6, 0x05, 0x72, 0xe2, 0xff, 0xb1, 0xdc, 0x7f } }; const struct hv_guid hv_guid_shutdown = { { 0x31, 0x60, 0x0b, 0x0e, 0x13, 0x52, 0x34, 0x49, 0x81, 0x8b, 0x38, 0xd9, 0x0c, 0xed, 0x39, 0xdb } }; const struct hv_guid hv_guid_timesync = { { 0x30, 0xe6, 0x27, 0x95, 0xae, 0xd0, 0x7b, 0x49, 0xad, 0xce, 0xe8, 0x0a, 0xb0, 0x17, 0x5c, 0xaf } }; const struct hv_guid hv_guid_heartbeat = { { 0x39, 0x4f, 0x16, 0x57, 0x15, 0x91, 0x78, 0x4e, 0xab, 0x55, 0x38, 0x2f, 0x3b, 0xd5, 0x42, 0x2d } }; const struct hv_guid hv_guid_kvp = { { 0xe7, 0xf4, 0xa0, 0xa9, 0x45, 0x5a, 0x96, 0x4d, 0xb8, 0x27, 0x8a, 0x84, 0x1e, 0x8c, 0x03, 0xe6 } }; #ifdef HYPERV_DEBUG const struct hv_guid hv_guid_vss = { { 0x29, 0x2e, 0xfa, 0x35, 0x23, 0xea, 0x36, 0x42, 0x96, 0xae, 0x3a, 0x6e, 0xba, 0xcb, 0xa4, 0x40 } }; const struct hv_guid hv_guid_dynmem = { { 0xdc, 0x74, 0x50, 0x52, 0x85, 0x89, 0xe2, 0x46, 0x80, 0x57, 0xa3, 0x07, 0xdc, 0x18, 0xa5, 0x02 } }; const struct hv_guid hv_guid_mouse = { { 0x9e, 0xb6, 0xa8, 0xcf, 0x4a, 0x5b, 0xc0, 0x4c, 0xb9, 0x8b, 0x8b, 0xa1, 0xa1, 0xf3, 0xf9, 0x5a } }; const struct hv_guid hv_guid_kbd = { { 0x6d, 0xad, 0x12, 0xf9, 0x17, 0x2b, 0xea, 0x48, 0xbd, 0x65, 0xf9, 0x27, 0xa6, 0x1c, 0x76, 0x84 } }; const struct hv_guid hv_guid_video = { { 0x02, 0x78, 0x0a, 0xda, 0x77, 0xe3, 0xac, 0x4a, 0x8e, 0x77, 0x05, 0x58, 0xeb, 0x10, 0x73, 0xf8 } }; const struct hv_guid hv_guid_fc = { { 0x4a, 0xcc, 0x9b, 0x2f, 0x69, 0x00, 0xf3, 0x4a, 0xb7, 0x6b, 0x6f, 0xd0, 0xbe, 0x52, 0x8c, 0xda } }; const struct hv_guid hv_guid_fcopy = { { 0xe3, 0x4b, 0xd1, 0x34, 0xe4, 0xde, 0xc8, 0x41, 0x9a, 0xe7, 0x6b, 0x17, 0x49, 0x77, 0xc1, 0x92 } }; const struct hv_guid hv_guid_pcie = { { 0x1d, 0xf6, 0xc4, 0x44, 0x44, 0x44, 0x00, 0x44, 0x9d, 0x52, 0x80, 0x2e, 0x27, 0xed, 0xe1, 0x9f } }; const struct hv_guid hv_guid_netdir = { { 0x3d, 0xaf, 0x2e, 0x8c, 0xa7, 0x32, 0x09, 0x4b, 0xab, 0x99, 0xbd, 0x1f, 0x1c, 0x86, 0xb5, 0x01 } }; const struct hv_guid hv_guid_rdesktop = { { 0xf4, 0xac, 0x6a, 0x27, 0x15, 0xac, 0x6c, 0x42, 0x98, 0xdd, 0x75, 0x21, 0xad, 0x3f, 0x01, 0xfe } }; /* Automatic Virtual Machine Activation (AVMA) Services */ const struct hv_guid hv_guid_avma1 = { { 0x55, 0xb2, 0x87, 0x44, 0x8c, 0xb8, 0x3f, 0x40, 0xbb, 0x51, 0xd1, 0xf6, 0x9c, 0xf1, 0x7f, 0x87 } }; const struct hv_guid hv_guid_avma2 = { { 0xf4, 0xba, 0x75, 0x33, 0x15, 0x9e, 0x30, 0x4b, 0xb7, 0x65, 0x67, 0xac, 0xb1, 0x0d, 0x60, 0x7b } }; const struct hv_guid hv_guid_avma3 = { { 0xa0, 0x1f, 0x22, 0x99, 0xad, 0x24, 0xe2, 0x11, 0xbe, 0x98, 0x00, 0x1a, 0xa0, 0x1b, 0xbf, 0x6e } }; const struct hv_guid hv_guid_avma4 = { { 0x16, 0x57, 0xe6, 0xf8, 0xb3, 0x3c, 0x06, 0x4a, 0x9a, 0x60, 0x18, 0x89, 0xc5, 0xcc, 0xca, 0xb5 } }; #endif /* HYPERV_DEBUG */ int hv_match(struct device *parent, void *match, void *aux) { struct pv_attach_args *pva = aux; struct pvbus_hv *hv = &pva->pva_hv[PVBUS_HYPERV]; if ((hv->hv_major == 0 && hv->hv_minor == 0) || hv->hv_base == 0) return (0); return (1); } void hv_attach(struct device *parent, struct device *self, void *aux) { struct hv_softc *sc = (struct hv_softc *)self; struct pv_attach_args *pva = aux; struct pvbus_hv *hv = &pva->pva_hv[PVBUS_HYPERV]; sc->sc_pvbus = hv; sc->sc_dmat = pva->pva_dmat; if (!(hv->hv_features & CPUID_HV_MSR_HYPERCALL) || !(hv->hv_features & CPUID_HV_MSR_SYNIC)) { printf(": not functional\n"); return; } DPRINTF("\n"); hv_set_version(sc); if (hv->hv_features & CPUID_HV_MSR_TIME_REFCNT) tc_init(&hv_timecounter); if (hv_init_hypercall(sc)) return; /* Wire it up to the global */ hv_sc = sc; if (hv_init_interrupts(sc)) return; if (hv_vmbus_connect(sc)) return; DPRINTF("%s", sc->sc_dev.dv_xname); printf(": protocol %d.%d, features %#x\n", VMBUS_VERSION_MAJOR(sc->sc_proto), VMBUS_VERSION_MINOR(sc->sc_proto), hv->hv_features); if (hv_channel_scan(sc)) return; /* Attach heartbeat, KVP and other "internal" services */ hv_attach_icdevs(sc); /* Attach devices with external drivers */ hv_attach_devices(sc); } void hv_set_version(struct hv_softc *sc) { uint64_t ver; /* OpenBSD build date */ ver = MSR_HV_GUESTID_OSTYPE_OPENBSD; ver |= (uint64_t)OpenBSD << MSR_HV_GUESTID_VERSION_SHIFT; wrmsr(MSR_HV_GUEST_OS_ID, ver); } u_int hv_gettime(struct timecounter *tc) { u_int now = rdmsr(MSR_HV_TIME_REF_COUNT); return (now); } void hv_delay(int usecs) { uint64_t interval, start; /* 10 MHz fixed frequency */ interval = (uint64_t)usecs * 10; start = rdmsr(MSR_HV_TIME_REF_COUNT); while (rdmsr(MSR_HV_TIME_REF_COUNT) - start < interval) CPU_BUSY_CYCLE(); } int hv_init_hypercall(struct hv_softc *sc) { extern void *hv_hypercall_page; uint64_t msr; paddr_t pa; sc->sc_hc = &hv_hypercall_page; if (!pmap_extract(pmap_kernel(), (vaddr_t)sc->sc_hc, &pa)) { printf(": hypercall page PA extraction failed\n"); return (-1); } msr = (atop(pa) << MSR_HV_HYPERCALL_PGSHIFT) | MSR_HV_HYPERCALL_ENABLE; wrmsr(MSR_HV_HYPERCALL, msr); if (!(rdmsr(MSR_HV_HYPERCALL) & MSR_HV_HYPERCALL_ENABLE)) { printf(": failed to set up a hypercall page\n"); return (-1); } return (0); } uint64_t hv_hypercall(struct hv_softc *sc, uint64_t control, void *input, void *output) { paddr_t input_pa = 0, output_pa = 0; uint64_t status = 0; if (input != NULL && pmap_extract(pmap_kernel(), (vaddr_t)input, &input_pa) == 0) { printf("%s: hypercall input PA extraction failed\n", sc->sc_dev.dv_xname); return (~HYPERCALL_STATUS_SUCCESS); } if (output != NULL && pmap_extract(pmap_kernel(), (vaddr_t)output, &output_pa) == 0) { printf("%s: hypercall output PA extraction failed\n", sc->sc_dev.dv_xname); return (~HYPERCALL_STATUS_SUCCESS); } #ifdef __amd64__ extern uint64_t hv_hypercall_trampoline(uint64_t, paddr_t, paddr_t); status = hv_hypercall_trampoline(control, input_pa, output_pa); #else /* __i386__ */ { uint32_t control_hi = control >> 32; uint32_t control_lo = control & 0xfffffffff; uint32_t status_hi = 1; uint32_t status_lo = 1; __asm__ volatile ("call *%8" : "=d" (status_hi), "=a"(status_lo) : "d" (control_hi), "a" (control_lo), "b" (0), "c" (input_pa), "D" (0), "S" (output_pa), "m" (sc->sc_hc)); status = status_lo | ((uint64_t)status_hi << 32); } #endif /* __amd64__ */ return (status); } int hv_init_interrupts(struct hv_softc *sc) { struct cpu_info *ci = curcpu(); int cpu = CPU_INFO_UNIT(ci); sc->sc_idtvec = LAPIC_HYPERV_VECTOR; TAILQ_INIT(&sc->sc_reqs); mtx_init(&sc->sc_reqlck, IPL_NET); TAILQ_INIT(&sc->sc_rsps); mtx_init(&sc->sc_rsplck, IPL_NET); sc->sc_simp[cpu] = km_alloc(PAGE_SIZE, &kv_any, &kp_zero, &kd_nowait); if (sc->sc_simp[cpu] == NULL) { printf(": failed to allocate SIMP\n"); return (-1); } sc->sc_siep[cpu] = km_alloc(PAGE_SIZE, &kv_any, &kp_zero, &kd_nowait); if (sc->sc_siep[cpu] == NULL) { printf(": failed to allocate SIEP\n"); km_free(sc->sc_simp[cpu], PAGE_SIZE, &kv_any, &kp_zero); return (-1); } sc->sc_proto = VMBUS_VERSION_WS2008; return (hv_init_synic(sc)); } int hv_init_synic(struct hv_softc *sc) { struct cpu_info *ci = curcpu(); int cpu = CPU_INFO_UNIT(ci); uint64_t simp, siefp, sctrl, sint; paddr_t pa; /* * Setup the Synic's message page */ if (!pmap_extract(pmap_kernel(), (vaddr_t)sc->sc_simp[cpu], &pa)) { printf(": SIMP PA extraction failed\n"); return (-1); } simp = rdmsr(MSR_HV_SIMP); simp &= (1 << MSR_HV_SIMP_PGSHIFT) - 1; simp |= (atop(pa) << MSR_HV_SIMP_PGSHIFT); simp |= MSR_HV_SIMP_ENABLE; wrmsr(MSR_HV_SIMP, simp); /* * Setup the Synic's event page */ if (!pmap_extract(pmap_kernel(), (vaddr_t)sc->sc_siep[cpu], &pa)) { printf(": SIEP PA extraction failed\n"); return (-1); } siefp = rdmsr(MSR_HV_SIEFP); siefp &= (1<<MSR_HV_SIEFP_PGSHIFT) - 1; siefp |= (atop(pa) << MSR_HV_SIEFP_PGSHIFT); siefp |= MSR_HV_SIEFP_ENABLE; wrmsr(MSR_HV_SIEFP, siefp); /* * Configure and unmask SINT for message and event flags */ sint = rdmsr(MSR_HV_SINT0 + VMBUS_SINT_MESSAGE); sint = sc->sc_idtvec | MSR_HV_SINT_AUTOEOI | (sint & MSR_HV_SINT_RSVD_MASK); wrmsr(MSR_HV_SINT0 + VMBUS_SINT_MESSAGE, sint); /* Enable the global synic bit */ sctrl = rdmsr(MSR_HV_SCONTROL); sctrl |= MSR_HV_SCTRL_ENABLE; wrmsr(MSR_HV_SCONTROL, sctrl); sc->sc_vcpus[cpu] = rdmsr(MSR_HV_VP_INDEX); DPRINTF("vcpu%u: SIMP %#llx SIEFP %#llx SCTRL %#llx\n", sc->sc_vcpus[cpu], simp, siefp, sctrl); return (0); } int hv_cmd(struct hv_softc *sc, void *cmd, size_t cmdlen, void *rsp, size_t rsplen, int flags) { struct hv_msg msg; int rv; if (cmdlen > VMBUS_MSG_DSIZE_MAX) { printf("%s: payload too large (%lu)\n", sc->sc_dev.dv_xname, cmdlen); return (EMSGSIZE); } memset(&msg, 0, sizeof(msg)); msg.msg_req.hc_dsize = cmdlen; memcpy(msg.msg_req.hc_data, cmd, cmdlen); if (!(flags & HCF_NOREPLY)) { msg.msg_rsp = rsp; msg.msg_rsplen = rsplen; } else msg.msg_flags |= MSGF_NOQUEUE; if (flags & HCF_NOSLEEP) msg.msg_flags |= MSGF_NOSLEEP; if ((rv = hv_start(sc, &msg)) != 0) return (rv); return (hv_reply(sc, &msg)); } int hv_start(struct hv_softc *sc, struct hv_msg *msg) { const int delays[] = { 100, 100, 100, 500, 500, 5000, 5000, 5000 }; const char *wchan = "hvstart"; uint16_t status; int i, s; msg->msg_req.hc_connid = VMBUS_CONNID_MESSAGE; msg->msg_req.hc_msgtype = 1; if (!(msg->msg_flags & MSGF_NOQUEUE)) { mtx_enter(&sc->sc_reqlck); TAILQ_INSERT_TAIL(&sc->sc_reqs, msg, msg_entry); mtx_leave(&sc->sc_reqlck); } for (i = 0; i < nitems(delays); i++) { status = hv_hypercall(sc, HYPERCALL_POST_MESSAGE, &msg->msg_req, NULL); if (status == HYPERCALL_STATUS_SUCCESS) break; if (msg->msg_flags & MSGF_NOSLEEP) { delay(delays[i]); s = splnet(); hv_intr(); splx(s); } else { tsleep_nsec(wchan, PRIBIO, wchan, USEC_TO_NSEC(delays[i])); } } if (status != 0) { printf("%s: posting vmbus message failed with %d\n", sc->sc_dev.dv_xname, status); if (!(msg->msg_flags & MSGF_NOQUEUE)) { mtx_enter(&sc->sc_reqlck); TAILQ_REMOVE(&sc->sc_reqs, msg, msg_entry); mtx_leave(&sc->sc_reqlck); } return (EIO); } return (0); } static int hv_reply_done(struct hv_softc *sc, struct hv_msg *msg) { struct hv_msg *m; mtx_enter(&sc->sc_rsplck); TAILQ_FOREACH(m, &sc->sc_rsps, msg_entry) { if (m == msg) { mtx_leave(&sc->sc_rsplck); return (1); } } mtx_leave(&sc->sc_rsplck); return (0); } int hv_reply(struct hv_softc *sc, struct hv_msg *msg) { if (msg->msg_flags & MSGF_NOQUEUE) return (0); hv_wait(sc, hv_reply_done, msg, msg, "hvreply"); mtx_enter(&sc->sc_rsplck); TAILQ_REMOVE(&sc->sc_rsps, msg, msg_entry); mtx_leave(&sc->sc_rsplck); return (0); } void hv_wait(struct hv_softc *sc, int (*cond)(struct hv_softc *, struct hv_msg *), struct hv_msg *msg, void *wchan, const char *wmsg) { int s; KASSERT(cold ? msg->msg_flags & MSGF_NOSLEEP : 1); while (!cond(sc, msg)) { if (msg->msg_flags & MSGF_NOSLEEP) { delay(1000); s = splnet(); hv_intr(); splx(s); } else { tsleep_nsec(wchan, PRIBIO, wmsg ? wmsg : "hvwait", USEC_TO_NSEC(1000)); } } } uint16_t hv_intr_signal(struct hv_softc *sc, void *con) { uint64_t status; status = hv_hypercall(sc, HYPERCALL_SIGNAL_EVENT, con, NULL); return ((uint16_t)status); } void hv_intr(void) { struct hv_softc *sc = hv_sc; hv_event_intr(sc); hv_message_intr(sc); } void hv_event_intr(struct hv_softc *sc) { struct vmbus_evtflags *evt; struct cpu_info *ci = curcpu(); int cpu = CPU_INFO_UNIT(ci); int bit, row, maxrow, chanid; struct hv_channel *ch; u_long *revents, pending; evt = (struct vmbus_evtflags *)sc->sc_siep[cpu] + VMBUS_SINT_MESSAGE; if ((sc->sc_proto == VMBUS_VERSION_WS2008) || (sc->sc_proto == VMBUS_VERSION_WIN7)) { if (!test_bit(0, &evt->evt_flags[0])) return; clear_bit(0, &evt->evt_flags[0]); maxrow = VMBUS_CHAN_MAX_COMPAT / VMBUS_EVTFLAG_LEN; /* * receive size is 1/2 page and divide that by 4 bytes */ revents = sc->sc_revents; } else { maxrow = nitems(evt->evt_flags); /* * On Host with Win8 or above, the event page can be * checked directly to get the id of the channel * that has the pending interrupt. */ revents = &evt->evt_flags[0]; } for (row = 0; row < maxrow; row++) { if (revents[row] == 0) continue; pending = atomic_swap_ulong(&revents[row], 0); for (bit = 0; pending > 0; pending >>= 1, bit++) { if ((pending & 1) == 0) continue; chanid = (row * LONG_BIT) + bit; /* vmbus channel protocol message */ if (chanid == 0) continue; ch = hv_channel_lookup(sc, chanid); if (ch == NULL) { printf("%s: unhandled event on %d\n", sc->sc_dev.dv_xname, chanid); continue; } if (ch->ch_state != HV_CHANSTATE_OPENED) { printf("%s: channel %d is not active\n", sc->sc_dev.dv_xname, chanid); continue; } ch->ch_evcnt.ec_count++; hv_channel_schedule(ch); } } } void hv_message_intr(struct hv_softc *sc) { struct vmbus_message *msg; struct vmbus_chanmsg_hdr *hdr; struct cpu_info *ci = curcpu(); int cpu = CPU_INFO_UNIT(ci); for (;;) { msg = (struct vmbus_message *)sc->sc_simp[cpu] + VMBUS_SINT_MESSAGE; if (msg->msg_type == VMBUS_MSGTYPE_NONE) break; hdr = (struct vmbus_chanmsg_hdr *)msg->msg_data; if (hdr->chm_type >= VMBUS_CHANMSG_COUNT) { printf("%s: unhandled message type %u flags %#x\n", sc->sc_dev.dv_xname, hdr->chm_type, msg->msg_flags); goto skip; } if (hv_msg_dispatch[hdr->chm_type].hmd_handler) hv_msg_dispatch[hdr->chm_type].hmd_handler(sc, hdr); else printf("%s: unhandled message type %u\n", sc->sc_dev.dv_xname, hdr->chm_type); skip: msg->msg_type = VMBUS_MSGTYPE_NONE; virtio_membar_sync(); if (msg->msg_flags & VMBUS_MSGFLAG_PENDING) wrmsr(MSR_HV_EOM, 0); } } void hv_channel_response(struct hv_softc *sc, struct vmbus_chanmsg_hdr *rsphdr) { struct hv_msg *msg; struct vmbus_chanmsg_hdr *reqhdr; int req; req = hv_msg_dispatch[rsphdr->chm_type].hmd_request; mtx_enter(&sc->sc_reqlck); TAILQ_FOREACH(msg, &sc->sc_reqs, msg_entry) { reqhdr = (struct vmbus_chanmsg_hdr *)&msg->msg_req.hc_data; if (reqhdr->chm_type == req) { TAILQ_REMOVE(&sc->sc_reqs, msg, msg_entry); break; } } mtx_leave(&sc->sc_reqlck); if (msg != NULL) { memcpy(msg->msg_rsp, rsphdr, msg->msg_rsplen); mtx_enter(&sc->sc_rsplck); TAILQ_INSERT_TAIL(&sc->sc_rsps, msg, msg_entry); mtx_leave(&sc->sc_rsplck); wakeup(msg); } } void hv_channel_offer(struct hv_softc *sc, struct vmbus_chanmsg_hdr *hdr) { struct hv_offer *co; co = malloc(sizeof(*co), M_DEVBUF, M_NOWAIT | M_ZERO); if (co == NULL) { printf("%s: failed to allocate an offer object\n", sc->sc_dev.dv_xname); return; } memcpy(&co->co_chan, hdr, sizeof(co->co_chan)); mtx_enter(&sc->sc_offerlck); SIMPLEQ_INSERT_TAIL(&sc->sc_offers, co, co_entry); mtx_leave(&sc->sc_offerlck); } void hv_channel_rescind(struct hv_softc *sc, struct vmbus_chanmsg_hdr *hdr) { const struct vmbus_chanmsg_chrescind *cmd; cmd = (const struct vmbus_chanmsg_chrescind *)hdr; printf("%s: revoking channel %u\n", sc->sc_dev.dv_xname, cmd->chm_chanid); } void hv_channel_delivered(struct hv_softc *sc, struct vmbus_chanmsg_hdr *hdr) { atomic_setbits_int(&sc->sc_flags, HSF_OFFERS_DELIVERED); wakeup(&sc->sc_offers); } int hv_vmbus_connect(struct hv_softc *sc) { const uint32_t versions[] = { VMBUS_VERSION_WIN10, VMBUS_VERSION_WIN8_1, VMBUS_VERSION_WIN8, VMBUS_VERSION_WIN7, VMBUS_VERSION_WS2008 }; struct vmbus_chanmsg_connect cmd; struct vmbus_chanmsg_connect_resp rsp; paddr_t epa, mpa1, mpa2; int i; sc->sc_events = km_alloc(PAGE_SIZE, &kv_any, &kp_zero, &kd_nowait); if (sc->sc_events == NULL) { printf(": failed to allocate channel port events page\n"); goto errout; } if (!pmap_extract(pmap_kernel(), (vaddr_t)sc->sc_events, &epa)) { printf(": channel port events page PA extraction failed\n"); goto errout; } sc->sc_wevents = (u_long *)sc->sc_events; sc->sc_revents = (u_long *)((caddr_t)sc->sc_events + (PAGE_SIZE >> 1)); sc->sc_monitor[0] = km_alloc(PAGE_SIZE, &kv_any, &kp_zero, &kd_nowait); if (sc->sc_monitor[0] == NULL) { printf(": failed to allocate monitor page 1\n"); goto errout; } if (!pmap_extract(pmap_kernel(), (vaddr_t)sc->sc_monitor[0], &mpa1)) { printf(": monitor page 1 PA extraction failed\n"); goto errout; } sc->sc_monitor[1] = km_alloc(PAGE_SIZE, &kv_any, &kp_zero, &kd_nowait); if (sc->sc_monitor[1] == NULL) { printf(": failed to allocate monitor page 2\n"); goto errout; } if (!pmap_extract(pmap_kernel(), (vaddr_t)sc->sc_monitor[1], &mpa2)) { printf(": monitor page 2 PA extraction failed\n"); goto errout; } memset(&cmd, 0, sizeof(cmd)); cmd.chm_hdr.chm_type = VMBUS_CHANMSG_CONNECT; cmd.chm_evtflags = (uint64_t)epa; cmd.chm_mnf1 = (uint64_t)mpa1; cmd.chm_mnf2 = (uint64_t)mpa2; memset(&rsp, 0, sizeof(rsp)); for (i = 0; i < nitems(versions); i++) { cmd.chm_ver = versions[i]; if (hv_cmd(sc, &cmd, sizeof(cmd), &rsp, sizeof(rsp), HCF_NOSLEEP)) { DPRINTF("%s: CONNECT failed\n", sc->sc_dev.dv_xname); goto errout; } if (rsp.chm_done) { sc->sc_flags |= HSF_CONNECTED; sc->sc_proto = versions[i]; sc->sc_handle = VMBUS_GPADL_START; break; } } if (i == nitems(versions)) { printf("%s: failed to negotiate protocol version\n", sc->sc_dev.dv_xname); goto errout; } return (0); errout: if (sc->sc_events) { km_free(sc->sc_events, PAGE_SIZE, &kv_any, &kp_zero); sc->sc_events = NULL; sc->sc_wevents = NULL; sc->sc_revents = NULL; } if (sc->sc_monitor[0]) { km_free(sc->sc_monitor[0], PAGE_SIZE, &kv_any, &kp_zero); sc->sc_monitor[0] = NULL; } if (sc->sc_monitor[1]) { km_free(sc->sc_monitor[1], PAGE_SIZE, &kv_any, &kp_zero); sc->sc_monitor[1] = NULL; } return (-1); } #ifdef HYPERV_DEBUG static inline char * guidprint(struct hv_guid *a) { /* 3 0 5 4 7 6 8 9 10 15 */ /* 33221100-5544-7766-9988-FFEEDDCCBBAA */ static char buf[16 * 2 + 4 + 1]; int i, j = 0; for (i = 3; i != -1; i -= 1, j += 2) snprintf(&buf[j], 3, "%02x", (uint8_t)a->data[i]); buf[j++] = '-'; for (i = 5; i != 3; i -= 1, j += 2) snprintf(&buf[j], 3, "%02x", (uint8_t)a->data[i]); buf[j++] = '-'; for (i = 7; i != 5; i -= 1, j += 2) snprintf(&buf[j], 3, "%02x", (uint8_t)a->data[i]); buf[j++] = '-'; for (i = 8; i < 10; i += 1, j += 2) snprintf(&buf[j], 3, "%02x", (uint8_t)a->data[i]); buf[j++] = '-'; for (i = 10; i < 16; i += 1, j += 2) snprintf(&buf[j], 3, "%02x", (uint8_t)a->data[i]); return (&buf[0]); } #endif /* HYPERV_DEBUG */ void hv_guid_sprint(struct hv_guid *guid, char *str, size_t size) { const struct { const struct hv_guid *guid; const char *ident; } map[] = { { &hv_guid_network, "network" }, { &hv_guid_ide, "ide" }, { &hv_guid_scsi, "scsi" }, { &hv_guid_shutdown, "shutdown" }, { &hv_guid_timesync, "timesync" }, { &hv_guid_heartbeat, "heartbeat" }, { &hv_guid_kvp, "kvp" }, #ifdef HYPERV_DEBUG { &hv_guid_vss, "vss" }, { &hv_guid_dynmem, "dynamic-memory" }, { &hv_guid_mouse, "mouse" }, { &hv_guid_kbd, "keyboard" }, { &hv_guid_video, "video" }, { &hv_guid_fc, "fiber-channel" }, { &hv_guid_fcopy, "file-copy" }, { &hv_guid_pcie, "pcie-passthrough" }, { &hv_guid_netdir, "network-direct" }, { &hv_guid_rdesktop, "remote-desktop" }, { &hv_guid_avma1, "avma-1" }, { &hv_guid_avma2, "avma-2" }, { &hv_guid_avma3, "avma-3" }, { &hv_guid_avma4, "avma-4" }, #endif }; int i; for (i = 0; i < nitems(map); i++) { if (memcmp(guid, map[i].guid, sizeof(*guid)) == 0) { strlcpy(str, map[i].ident, size); return; } } #ifdef HYPERV_DEBUG strlcpy(str, guidprint(guid), size); #endif } static int hv_channel_scan_done(struct hv_softc *sc, struct hv_msg *msg __unused) { return (sc->sc_flags & HSF_OFFERS_DELIVERED); } int hv_channel_scan(struct hv_softc *sc) { struct vmbus_chanmsg_hdr hdr; struct vmbus_chanmsg_choffer rsp; struct hv_offer *co; SIMPLEQ_INIT(&sc->sc_offers); mtx_init(&sc->sc_offerlck, IPL_NET); memset(&hdr, 0, sizeof(hdr)); hdr.chm_type = VMBUS_CHANMSG_CHREQUEST; if (hv_cmd(sc, &hdr, sizeof(hdr), &rsp, sizeof(rsp), HCF_NOSLEEP | HCF_NOREPLY)) { DPRINTF("%s: CHREQUEST failed\n", sc->sc_dev.dv_xname); return (-1); } hv_wait(sc, hv_channel_scan_done, (struct hv_msg *)&hdr, &sc->sc_offers, "hvscan"); TAILQ_INIT(&sc->sc_channels); mtx_init(&sc->sc_channelck, IPL_NET); mtx_enter(&sc->sc_offerlck); while (!SIMPLEQ_EMPTY(&sc->sc_offers)) { co = SIMPLEQ_FIRST(&sc->sc_offers); SIMPLEQ_REMOVE_HEAD(&sc->sc_offers, co_entry); mtx_leave(&sc->sc_offerlck); hv_process_offer(sc, co); free(co, M_DEVBUF, sizeof(*co)); mtx_enter(&sc->sc_offerlck); } mtx_leave(&sc->sc_offerlck); return (0); } void hv_process_offer(struct hv_softc *sc, struct hv_offer *co) { struct hv_channel *ch, *nch; nch = malloc(sizeof(*nch), M_DEVBUF, M_ZERO | M_NOWAIT); if (nch == NULL) { printf("%s: failed to allocate memory for the channel\n", sc->sc_dev.dv_xname); return; } nch->ch_sc = sc; hv_guid_sprint(&co->co_chan.chm_chtype, nch->ch_ident, sizeof(nch->ch_ident)); /* * By default we setup state to enable batched reading. * A specific service can choose to disable this prior * to opening the channel. */ nch->ch_flags |= CHF_BATCHED; KASSERT((((vaddr_t)&nch->ch_monprm) & 0x7) == 0); memset(&nch->ch_monprm, 0, sizeof(nch->ch_monprm)); nch->ch_monprm.mp_connid = VMBUS_CONNID_EVENT; if (sc->sc_proto != VMBUS_VERSION_WS2008) nch->ch_monprm.mp_connid = co->co_chan.chm_connid; if (co->co_chan.chm_flags1 & VMBUS_CHOFFER_FLAG1_HASMNF) { nch->ch_mgroup = co->co_chan.chm_montrig / VMBUS_MONTRIG_LEN; nch->ch_mindex = co->co_chan.chm_montrig % VMBUS_MONTRIG_LEN; nch->ch_flags |= CHF_MONITOR; } nch->ch_id = co->co_chan.chm_chanid; memcpy(&nch->ch_type, &co->co_chan.chm_chtype, sizeof(ch->ch_type)); memcpy(&nch->ch_inst, &co->co_chan.chm_chinst, sizeof(ch->ch_inst)); mtx_enter(&sc->sc_channelck); TAILQ_FOREACH(ch, &sc->sc_channels, ch_entry) { if (!memcmp(&ch->ch_type, &nch->ch_type, sizeof(ch->ch_type)) && !memcmp(&ch->ch_inst, &nch->ch_inst, sizeof(ch->ch_inst))) break; } if (ch != NULL) { if (co->co_chan.chm_subidx == 0) { printf("%s: unknown offer \"%s\"\n", sc->sc_dev.dv_xname, nch->ch_ident); mtx_leave(&sc->sc_channelck); free(nch, M_DEVBUF, sizeof(*nch)); return; } #ifdef HYPERV_DEBUG printf("%s: subchannel %u for \"%s\"\n", sc->sc_dev.dv_xname, co->co_chan.chm_subidx, ch->ch_ident); #endif mtx_leave(&sc->sc_channelck); free(nch, M_DEVBUF, sizeof(*nch)); return; } nch->ch_state = HV_CHANSTATE_OFFERED; TAILQ_INSERT_TAIL(&sc->sc_channels, nch, ch_entry); mtx_leave(&sc->sc_channelck); #ifdef HYPERV_DEBUG printf("%s: channel %u: \"%s\"", sc->sc_dev.dv_xname, nch->ch_id, nch->ch_ident); if (nch->ch_flags & CHF_MONITOR) printf(", monitor %u\n", co->co_chan.chm_montrig); else printf("\n"); #endif } struct hv_channel * hv_channel_lookup(struct hv_softc *sc, uint32_t relid) { struct hv_channel *ch; TAILQ_FOREACH(ch, &sc->sc_channels, ch_entry) { if (ch->ch_id == relid) return (ch); } return (NULL); } int hv_channel_ring_create(struct hv_channel *ch, uint32_t buflen) { struct hv_softc *sc = ch->ch_sc; buflen = roundup(buflen, PAGE_SIZE) + sizeof(struct vmbus_bufring); ch->ch_ring = km_alloc(2 * buflen, &kv_any, &kp_zero, cold ? &kd_nowait : &kd_waitok); if (ch->ch_ring == NULL) { printf("%s: failed to allocate channel ring\n", sc->sc_dev.dv_xname); return (-1); } ch->ch_ring_size = 2 * buflen; memset(&ch->ch_wrd, 0, sizeof(ch->ch_wrd)); ch->ch_wrd.rd_ring = (struct vmbus_bufring *)ch->ch_ring; ch->ch_wrd.rd_size = buflen; ch->ch_wrd.rd_dsize = buflen - sizeof(struct vmbus_bufring); mtx_init(&ch->ch_wrd.rd_lock, IPL_NET); memset(&ch->ch_rrd, 0, sizeof(ch->ch_rrd)); ch->ch_rrd.rd_ring = (struct vmbus_bufring *)((uint8_t *)ch->ch_ring + buflen); ch->ch_rrd.rd_size = buflen; ch->ch_rrd.rd_dsize = buflen - sizeof(struct vmbus_bufring); mtx_init(&ch->ch_rrd.rd_lock, IPL_NET); if (hv_handle_alloc(ch, ch->ch_ring, 2 * buflen, &ch->ch_ring_gpadl)) { printf("%s: failed to obtain a PA handle for the ring\n", sc->sc_dev.dv_xname); hv_channel_ring_destroy(ch); return (-1); } return (0); } void hv_channel_ring_destroy(struct hv_channel *ch) { km_free(ch->ch_ring, ch->ch_ring_size, &kv_any, &kp_zero); ch->ch_ring = NULL; hv_handle_free(ch, ch->ch_ring_gpadl); memset(&ch->ch_wrd, 0, sizeof(ch->ch_wrd)); memset(&ch->ch_rrd, 0, sizeof(ch->ch_rrd)); } int hv_channel_open(struct hv_channel *ch, size_t buflen, void *udata, size_t udatalen, void (*handler)(void *), void *arg) { struct hv_softc *sc = ch->ch_sc; struct vmbus_chanmsg_chopen cmd; struct vmbus_chanmsg_chopen_resp rsp; int rv; if (ch->ch_ring == NULL && hv_channel_ring_create(ch, buflen)) { DPRINTF("%s: failed to create channel ring\n", sc->sc_dev.dv_xname); return (-1); } memset(&cmd, 0, sizeof(cmd)); cmd.chm_hdr.chm_type = VMBUS_CHANMSG_CHOPEN; cmd.chm_openid = ch->ch_id; cmd.chm_chanid = ch->ch_id; cmd.chm_gpadl = ch->ch_ring_gpadl; cmd.chm_txbr_pgcnt = ch->ch_wrd.rd_size >> PAGE_SHIFT; cmd.chm_vcpuid = ch->ch_vcpu; if (udata && udatalen > 0) memcpy(cmd.chm_udata, udata, udatalen); memset(&rsp, 0, sizeof(rsp)); ch->ch_handler = handler; ch->ch_ctx = arg; ch->ch_state = HV_CHANSTATE_OPENED; rv = hv_cmd(sc, &cmd, sizeof(cmd), &rsp, sizeof(rsp), cold ? HCF_NOSLEEP : HCF_SLEEPOK); if (rv) { hv_channel_ring_destroy(ch); DPRINTF("%s: CHOPEN failed with %d\n", sc->sc_dev.dv_xname, rv); ch->ch_handler = NULL; ch->ch_ctx = NULL; ch->ch_state = HV_CHANSTATE_OFFERED; return (-1); } return (0); } int hv_channel_close(struct hv_channel *ch) { struct hv_softc *sc = ch->ch_sc; struct vmbus_chanmsg_chclose cmd; int rv; memset(&cmd, 0, sizeof(cmd)); cmd.chm_hdr.chm_type = VMBUS_CHANMSG_CHCLOSE; cmd.chm_chanid = ch->ch_id; ch->ch_state = HV_CHANSTATE_CLOSING; rv = hv_cmd(sc, &cmd, sizeof(cmd), NULL, 0, HCF_NOREPLY); if (rv) { DPRINTF("%s: CHCLOSE failed with %d\n", sc->sc_dev.dv_xname, rv); return (-1); } ch->ch_state = HV_CHANSTATE_CLOSED; hv_channel_ring_destroy(ch); return (0); } static inline void hv_channel_setevent(struct hv_softc *sc, struct hv_channel *ch) { struct vmbus_mon_trig *mtg; /* Each uint32_t represents 32 channels */ set_bit(ch->ch_id, sc->sc_wevents); if (ch->ch_flags & CHF_MONITOR) { mtg = &sc->sc_monitor[1]->mnf_trigs[ch->ch_mgroup]; set_bit(ch->ch_mindex, &mtg->mt_pending); } else hv_intr_signal(sc, &ch->ch_monprm); } void hv_channel_intr(void *arg) { struct hv_channel *ch = arg; if (hv_channel_ready(ch)) ch->ch_handler(ch->ch_ctx); if (hv_channel_unpause(ch) == 0) return; hv_channel_pause(ch); hv_channel_schedule(ch); } int hv_channel_setdeferred(struct hv_channel *ch, const char *name) { ch->ch_taskq = taskq_create(name, 1, IPL_NET, TASKQ_MPSAFE); if (ch->ch_taskq == NULL) return (-1); task_set(&ch->ch_task, hv_channel_intr, ch); return (0); } void hv_channel_schedule(struct hv_channel *ch) { if (ch->ch_handler) { if (!cold && (ch->ch_flags & CHF_BATCHED)) { hv_channel_pause(ch); task_add(ch->ch_taskq, &ch->ch_task); } else ch->ch_handler(ch->ch_ctx); } } static inline void hv_ring_put(struct hv_ring_data *wrd, uint8_t *data, uint32_t datalen) { int left = MIN(datalen, wrd->rd_dsize - wrd->rd_prod); memcpy(&wrd->rd_ring->br_data[wrd->rd_prod], data, left); memcpy(&wrd->rd_ring->br_data[0], data + left, datalen - left); wrd->rd_prod += datalen; if (wrd->rd_prod >= wrd->rd_dsize) wrd->rd_prod -= wrd->rd_dsize; } static inline void hv_ring_get(struct hv_ring_data *rrd, uint8_t *data, uint32_t datalen, int peek) { int left = MIN(datalen, rrd->rd_dsize - rrd->rd_cons); memcpy(data, &rrd->rd_ring->br_data[rrd->rd_cons], left); memcpy(data + left, &rrd->rd_ring->br_data[0], datalen - left); if (!peek) { rrd->rd_cons += datalen; if (rrd->rd_cons >= rrd->rd_dsize) rrd->rd_cons -= rrd->rd_dsize; } } static inline void hv_ring_avail(struct hv_ring_data *rd, uint32_t *towrite, uint32_t *toread) { uint32_t ridx = rd->rd_ring->br_rindex; uint32_t widx = rd->rd_ring->br_windex; uint32_t r, w; if (widx >= ridx) w = rd->rd_dsize - (widx - ridx); else w = ridx - widx; r = rd->rd_dsize - w; if (towrite) *towrite = w; if (toread) *toread = r; } int hv_ring_write(struct hv_ring_data *wrd, struct iovec *iov, int iov_cnt, int *needsig) { uint64_t indices = 0; uint32_t avail, oprod, datalen = sizeof(indices); int i; for (i = 0; i < iov_cnt; i++) datalen += iov[i].iov_len; KASSERT(datalen <= wrd->rd_dsize); hv_ring_avail(wrd, &avail, NULL); if (avail <= datalen) { DPRINTF("%s: avail %u datalen %u\n", __func__, avail, datalen); return (EAGAIN); } oprod = wrd->rd_prod; for (i = 0; i < iov_cnt; i++) hv_ring_put(wrd, iov[i].iov_base, iov[i].iov_len); indices = (uint64_t)oprod << 32; hv_ring_put(wrd, (uint8_t *)&indices, sizeof(indices)); virtio_membar_sync(); wrd->rd_ring->br_windex = wrd->rd_prod; virtio_membar_sync(); /* Signal when the ring transitions from being empty to non-empty */ if (wrd->rd_ring->br_imask == 0 && wrd->rd_ring->br_rindex == oprod) *needsig = 1; else *needsig = 0; return (0); } int hv_channel_send(struct hv_channel *ch, void *data, uint32_t datalen, uint64_t rid, int type, uint32_t flags) { struct hv_softc *sc = ch->ch_sc; struct vmbus_chanpkt cp; struct iovec iov[3]; uint32_t pktlen, pktlen_aligned; uint64_t zeropad = 0; int rv, needsig = 0; pktlen = sizeof(cp) + datalen; pktlen_aligned = roundup(pktlen, sizeof(uint64_t)); cp.cp_hdr.cph_type = type; cp.cp_hdr.cph_flags = flags; VMBUS_CHANPKT_SETLEN(cp.cp_hdr.cph_hlen, sizeof(cp)); VMBUS_CHANPKT_SETLEN(cp.cp_hdr.cph_tlen, pktlen_aligned); cp.cp_hdr.cph_tid = rid; iov[0].iov_base = &cp; iov[0].iov_len = sizeof(cp); iov[1].iov_base = data; iov[1].iov_len = datalen; iov[2].iov_base = &zeropad; iov[2].iov_len = pktlen_aligned - pktlen; mtx_enter(&ch->ch_wrd.rd_lock); rv = hv_ring_write(&ch->ch_wrd, iov, 3, &needsig); mtx_leave(&ch->ch_wrd.rd_lock); if (rv == 0 && needsig) hv_channel_setevent(sc, ch); return (rv); } int hv_channel_send_sgl(struct hv_channel *ch, struct vmbus_gpa *sgl, uint32_t nsge, void *data, uint32_t datalen, uint64_t rid) { struct hv_softc *sc = ch->ch_sc; struct vmbus_chanpkt_sglist cp; struct iovec iov[4]; uint32_t buflen, pktlen, pktlen_aligned; uint64_t zeropad = 0; int rv, needsig = 0; buflen = sizeof(struct vmbus_gpa) * nsge; pktlen = sizeof(cp) + datalen + buflen; pktlen_aligned = roundup(pktlen, sizeof(uint64_t)); cp.cp_hdr.cph_type = VMBUS_CHANPKT_TYPE_GPA; cp.cp_hdr.cph_flags = VMBUS_CHANPKT_FLAG_RC; VMBUS_CHANPKT_SETLEN(cp.cp_hdr.cph_hlen, sizeof(cp) + buflen); VMBUS_CHANPKT_SETLEN(cp.cp_hdr.cph_tlen, pktlen_aligned); cp.cp_hdr.cph_tid = rid; cp.cp_gpa_cnt = nsge; cp.cp_rsvd = 0; iov[0].iov_base = &cp; iov[0].iov_len = sizeof(cp); iov[1].iov_base = sgl; iov[1].iov_len = buflen; iov[2].iov_base = data; iov[2].iov_len = datalen; iov[3].iov_base = &zeropad; iov[3].iov_len = pktlen_aligned - pktlen; mtx_enter(&ch->ch_wrd.rd_lock); rv = hv_ring_write(&ch->ch_wrd, iov, 4, &needsig); mtx_leave(&ch->ch_wrd.rd_lock); if (rv == 0 && needsig) hv_channel_setevent(sc, ch); return (rv); } int hv_channel_send_prpl(struct hv_channel *ch, struct vmbus_gpa_range *prpl, uint32_t nprp, void *data, uint32_t datalen, uint64_t rid) { struct hv_softc *sc = ch->ch_sc; struct vmbus_chanpkt_prplist cp; struct iovec iov[4]; uint32_t buflen, pktlen, pktlen_aligned; uint64_t zeropad = 0; int rv, needsig = 0; buflen = sizeof(struct vmbus_gpa_range) * (nprp + 1); pktlen = sizeof(cp) + datalen + buflen; pktlen_aligned = roundup(pktlen, sizeof(uint64_t)); cp.cp_hdr.cph_type = VMBUS_CHANPKT_TYPE_GPA; cp.cp_hdr.cph_flags = VMBUS_CHANPKT_FLAG_RC; VMBUS_CHANPKT_SETLEN(cp.cp_hdr.cph_hlen, sizeof(cp) + buflen); VMBUS_CHANPKT_SETLEN(cp.cp_hdr.cph_tlen, pktlen_aligned); cp.cp_hdr.cph_tid = rid; cp.cp_range_cnt = 1; cp.cp_rsvd = 0; iov[0].iov_base = &cp; iov[0].iov_len = sizeof(cp); iov[1].iov_base = prpl; iov[1].iov_len = buflen; iov[2].iov_base = data; iov[2].iov_len = datalen; iov[3].iov_base = &zeropad; iov[3].iov_len = pktlen_aligned - pktlen; mtx_enter(&ch->ch_wrd.rd_lock); rv = hv_ring_write(&ch->ch_wrd, iov, 4, &needsig); mtx_leave(&ch->ch_wrd.rd_lock); if (rv == 0 && needsig) hv_channel_setevent(sc, ch); return (rv); } int hv_ring_peek(struct hv_ring_data *rrd, void *data, uint32_t datalen) { uint32_t avail; KASSERT(datalen <= rrd->rd_dsize); hv_ring_avail(rrd, NULL, &avail); if (avail < datalen) return (EAGAIN); hv_ring_get(rrd, (uint8_t *)data, datalen, 1); return (0); } int hv_ring_read(struct hv_ring_data *rrd, void *data, uint32_t datalen, uint32_t offset) { uint64_t indices; uint32_t avail; KASSERT(datalen <= rrd->rd_dsize); hv_ring_avail(rrd, NULL, &avail); if (avail < datalen) { DPRINTF("%s: avail %u datalen %u\n", __func__, avail, datalen); return (EAGAIN); } if (offset) { rrd->rd_cons += offset; if (rrd->rd_cons >= rrd->rd_dsize) rrd->rd_cons -= rrd->rd_dsize; } hv_ring_get(rrd, (uint8_t *)data, datalen, 0); hv_ring_get(rrd, (uint8_t *)&indices, sizeof(indices), 0); virtio_membar_sync(); rrd->rd_ring->br_rindex = rrd->rd_cons; return (0); } int hv_channel_recv(struct hv_channel *ch, void *data, uint32_t datalen, uint32_t *rlen, uint64_t *rid, int raw) { struct vmbus_chanpkt_hdr cph; uint32_t offset, pktlen; int rv; *rlen = 0; mtx_enter(&ch->ch_rrd.rd_lock); if ((rv = hv_ring_peek(&ch->ch_rrd, &cph, sizeof(cph))) != 0) { mtx_leave(&ch->ch_rrd.rd_lock); return (rv); } offset = raw ? 0 : VMBUS_CHANPKT_GETLEN(cph.cph_hlen); pktlen = VMBUS_CHANPKT_GETLEN(cph.cph_tlen) - offset; if (pktlen > datalen) { mtx_leave(&ch->ch_rrd.rd_lock); printf("%s: pktlen %u datalen %u\n", __func__, pktlen, datalen); return (EINVAL); } rv = hv_ring_read(&ch->ch_rrd, data, pktlen, offset); if (rv == 0) { *rlen = pktlen; *rid = cph.cph_tid; } mtx_leave(&ch->ch_rrd.rd_lock); return (rv); } static inline void hv_ring_mask(struct hv_ring_data *rd) { virtio_membar_sync(); rd->rd_ring->br_imask = 1; virtio_membar_sync(); } static inline void hv_ring_unmask(struct hv_ring_data *rd) { virtio_membar_sync(); rd->rd_ring->br_imask = 0; virtio_membar_sync(); } void hv_channel_pause(struct hv_channel *ch) { hv_ring_mask(&ch->ch_rrd); } uint hv_channel_unpause(struct hv_channel *ch) { uint32_t avail; hv_ring_unmask(&ch->ch_rrd); hv_ring_avail(&ch->ch_rrd, NULL, &avail); return (avail); } uint hv_channel_ready(struct hv_channel *ch) { uint32_t avail; hv_ring_avail(&ch->ch_rrd, NULL, &avail); return (avail); } /* How many PFNs can be referenced by the header */ #define HV_NPFNHDR ((VMBUS_MSG_DSIZE_MAX - \ sizeof(struct vmbus_chanmsg_gpadl_conn)) / sizeof(uint64_t)) /* How many PFNs can be referenced by the body */ #define HV_NPFNBODY ((VMBUS_MSG_DSIZE_MAX - \ sizeof(struct vmbus_chanmsg_gpadl_subconn)) / sizeof(uint64_t)) int hv_handle_alloc(struct hv_channel *ch, void *buffer, uint32_t buflen, uint32_t *handle) { struct hv_softc *sc = ch->ch_sc; struct vmbus_chanmsg_gpadl_conn *hdr; struct vmbus_chanmsg_gpadl_subconn *cmd; struct vmbus_chanmsg_gpadl_connresp rsp; struct hv_msg *msg; int i, j, last, left, rv; int bodylen = 0, ncmds = 0, pfn = 0; int waitflag = cold ? M_NOWAIT : M_WAITOK; uint64_t *frames; paddr_t pa; caddr_t body; /* Total number of pages to reference */ int total = atop(buflen); /* Number of pages that will fit the header */ int inhdr = MIN(total, HV_NPFNHDR); KASSERT((buflen & (PAGE_SIZE - 1)) == 0); if ((msg = malloc(sizeof(*msg), M_DEVBUF, M_ZERO | waitflag)) == NULL) return (ENOMEM); /* Prepare array of frame addresses */ if ((frames = mallocarray(total, sizeof(*frames), M_DEVBUF, M_ZERO | waitflag)) == NULL) { free(msg, M_DEVBUF, sizeof(*msg)); return (ENOMEM); } for (i = 0; i < total; i++) { if (!pmap_extract(pmap_kernel(), (vaddr_t)buffer + PAGE_SIZE * i, &pa)) { free(msg, M_DEVBUF, sizeof(*msg)); free(frames, M_DEVBUF, total * sizeof(*frames)); return (EFAULT); } frames[i] = atop(pa); } msg->msg_req.hc_dsize = sizeof(struct vmbus_chanmsg_gpadl_conn) + inhdr * sizeof(uint64_t); hdr = (struct vmbus_chanmsg_gpadl_conn *)msg->msg_req.hc_data; msg->msg_rsp = &rsp; msg->msg_rsplen = sizeof(rsp); if (waitflag == M_NOWAIT) msg->msg_flags = MSGF_NOSLEEP; left = total - inhdr; /* Allocate additional gpadl_body structures if required */ if (left > 0) { ncmds = MAX(1, left / HV_NPFNBODY + left % HV_NPFNBODY); bodylen = ncmds * VMBUS_MSG_DSIZE_MAX; body = malloc(bodylen, M_DEVBUF, M_ZERO | waitflag); if (body == NULL) { free(msg, M_DEVBUF, sizeof(*msg)); free(frames, M_DEVBUF, atop(buflen) * sizeof(*frames)); return (ENOMEM); } } *handle = atomic_inc_int_nv(&sc->sc_handle); hdr->chm_hdr.chm_type = VMBUS_CHANMSG_GPADL_CONN; hdr->chm_chanid = ch->ch_id; hdr->chm_gpadl = *handle; /* Single range for a contiguous buffer */ hdr->chm_range_cnt = 1; hdr->chm_range_len = sizeof(struct vmbus_gpa_range) + total * sizeof(uint64_t); hdr->chm_range.gpa_ofs = 0; hdr->chm_range.gpa_len = buflen; /* Fit as many pages as possible into the header */ for (i = 0; i < inhdr; i++) hdr->chm_range.gpa_page[i] = frames[pfn++]; for (i = 0; i < ncmds; i++) { cmd = (struct vmbus_chanmsg_gpadl_subconn *)(body + VMBUS_MSG_DSIZE_MAX * i); cmd->chm_hdr.chm_type = VMBUS_CHANMSG_GPADL_SUBCONN; cmd->chm_gpadl = *handle; last = MIN(left, HV_NPFNBODY); for (j = 0; j < last; j++) cmd->chm_gpa_page[j] = frames[pfn++]; left -= last; } rv = hv_start(sc, msg); if (rv != 0) { DPRINTF("%s: GPADL_CONN failed\n", sc->sc_dev.dv_xname); goto out; } for (i = 0; i < ncmds; i++) { int cmdlen = sizeof(*cmd); cmd = (struct vmbus_chanmsg_gpadl_subconn *)(body + VMBUS_MSG_DSIZE_MAX * i); /* Last element can be short */ if (i == ncmds - 1) cmdlen += last * sizeof(uint64_t); else cmdlen += HV_NPFNBODY * sizeof(uint64_t); rv = hv_cmd(sc, cmd, cmdlen, NULL, 0, waitflag | HCF_NOREPLY); if (rv != 0) { DPRINTF("%s: GPADL_SUBCONN (iteration %d/%d) failed " "with %d\n", sc->sc_dev.dv_xname, i, ncmds, rv); goto out; } } rv = hv_reply(sc, msg); if (rv != 0) DPRINTF("%s: GPADL allocation failed with %d\n", sc->sc_dev.dv_xname, rv); out: free(msg, M_DEVBUF, sizeof(*msg)); free(frames, M_DEVBUF, total * sizeof(*frames)); if (bodylen > 0) free(body, M_DEVBUF, bodylen); if (rv != 0) return (rv); KASSERT(*handle == rsp.chm_gpadl); return (0); } void hv_handle_free(struct hv_channel *ch, uint32_t handle) { struct hv_softc *sc = ch->ch_sc; struct vmbus_chanmsg_gpadl_disconn cmd; struct vmbus_chanmsg_gpadl_disconn rsp; int rv; memset(&cmd, 0, sizeof(cmd)); cmd.chm_hdr.chm_type = VMBUS_CHANMSG_GPADL_DISCONN; cmd.chm_chanid = ch->ch_id; cmd.chm_gpadl = handle; rv = hv_cmd(sc, &cmd, sizeof(cmd), &rsp, sizeof(rsp), cold ? HCF_NOSLEEP : 0); if (rv) DPRINTF("%s: GPADL_DISCONN failed with %d\n", sc->sc_dev.dv_xname, rv); } static int hv_attach_print(void *aux, const char *name) { struct hv_attach_args *aa = aux; if (name) printf("\"%s\" at %s", aa->aa_ident, name); return (UNCONF); } int hv_attach_devices(struct hv_softc *sc) { struct hv_dev *dv; struct hv_channel *ch; SLIST_INIT(&sc->sc_devs); mtx_init(&sc->sc_devlck, IPL_NET); TAILQ_FOREACH(ch, &sc->sc_channels, ch_entry) { if (ch->ch_state != HV_CHANSTATE_OFFERED) continue; if (!(ch->ch_flags & CHF_MONITOR)) continue; dv = malloc(sizeof(*dv), M_DEVBUF, M_ZERO | M_NOWAIT); if (dv == NULL) { printf("%s: failed to allocate device object\n", sc->sc_dev.dv_xname); return (-1); } dv->dv_aa.aa_parent = sc; dv->dv_aa.aa_type = &ch->ch_type; dv->dv_aa.aa_inst = &ch->ch_inst; dv->dv_aa.aa_ident = ch->ch_ident; dv->dv_aa.aa_chan = ch; dv->dv_aa.aa_dmat = sc->sc_dmat; mtx_enter(&sc->sc_devlck); SLIST_INSERT_HEAD(&sc->sc_devs, dv, dv_entry); mtx_leave(&sc->sc_devlck); config_found((struct device *)sc, &dv->dv_aa, hv_attach_print); } return (0); } void hv_evcount_attach(struct hv_channel *ch, const char *name) { struct hv_softc *sc = ch->ch_sc; evcount_attach(&ch->ch_evcnt, name, &sc->sc_idtvec); }