Resguardo inicial

linux-7.2.3/arch/x86/hyperv/.gitignore

@@ -0,0 +1 @@
+mshv-asm-offsets.h

linux-7.2.3/arch/x86/hyperv/Makefile

@@ -0,0 +1,22 @@
+# SPDX-License-Identifier: GPL-2.0-only
+obj-y			:= hv_init.o mmu.o nested.o irqdomain.o ivm.o
+obj-$(CONFIG_X86_64)	+= hv_apic.o
+obj-$(CONFIG_HYPERV_VTL_MODE)	+= hv_vtl.o mshv_vtl_asm.o
+
+$(obj)/mshv_vtl_asm.o: $(obj)/mshv-asm-offsets.h
+
+$(obj)/mshv-asm-offsets.h: $(obj)/mshv-asm-offsets.s FORCE
+	$(call filechk,offsets,__MSHV_ASM_OFFSETS_H__)
+
+ifdef CONFIG_X86_64
+obj-$(CONFIG_PARAVIRT_SPINLOCKS)	+= hv_spinlock.o
+
+ ifdef CONFIG_MSHV_ROOT
+  CFLAGS_REMOVE_hv_trampoline.o += -pg
+  CFLAGS_hv_trampoline.o        += -fno-stack-protector
+  obj-$(CONFIG_CRASH_DUMP)      += hv_crash.o hv_trampoline.o
+ endif
+endif
+
+targets += mshv-asm-offsets.s
+clean-files += mshv-asm-offsets.h

linux-7.2.3/arch/x86/hyperv/hv_apic.c

@@ -0,0 +1,341 @@
+// SPDX-License-Identifier: GPL-2.0
+
+/*
+ * Hyper-V specific APIC code.
+ *
+ * Copyright (C) 2018, Microsoft, Inc.
+ *
+ * Author : K. Y. Srinivasan <kys@microsoft.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or
+ * NON INFRINGEMENT.  See the GNU General Public License for more
+ * details.
+ *
+ */
+
+#include <linux/types.h>
+#include <linux/vmalloc.h>
+#include <linux/mm.h>
+#include <linux/clockchips.h>
+#include <linux/slab.h>
+#include <linux/cpuhotplug.h>
+#include <asm/hypervisor.h>
+#include <asm/mshyperv.h>
+#include <asm/apic.h>
+#include <asm/msr.h>
+
+#include <asm/trace/hyperv.h>
+
+static struct apic orig_apic;
+
+static u64 hv_apic_icr_read(void)
+{
+	u64 reg_val;
+
+	rdmsrq(HV_X64_MSR_ICR, reg_val);
+	return reg_val;
+}
+
+static void hv_apic_icr_write(u32 low, u32 id)
+{
+	u64 reg_val;
+
+	reg_val = SET_XAPIC_DEST_FIELD(id);
+	reg_val = reg_val << 32;
+	reg_val |= low;
+
+	wrmsrq(HV_X64_MSR_ICR, reg_val);
+}
+
+void hv_enable_coco_interrupt(unsigned int cpu, unsigned int vector, bool set)
+{
+	apic_update_vector(cpu, vector, set);
+}
+
+static u32 hv_apic_read(u32 reg)
+{
+	u32 reg_val, hi;
+
+	switch (reg) {
+	case APIC_EOI:
+		rdmsr(HV_X64_MSR_EOI, reg_val, hi);
+		(void)hi;
+		return reg_val;
+	case APIC_TASKPRI:
+		rdmsr(HV_X64_MSR_TPR, reg_val, hi);
+		(void)hi;
+		return reg_val;
+
+	default:
+		return native_apic_mem_read(reg);
+	}
+}
+
+static void hv_apic_write(u32 reg, u32 val)
+{
+	switch (reg) {
+	case APIC_EOI:
+		wrmsrq(HV_X64_MSR_EOI, val);
+		break;
+	case APIC_TASKPRI:
+		wrmsrq(HV_X64_MSR_TPR, val);
+		break;
+	default:
+		native_apic_mem_write(reg, val);
+	}
+}
+
+static void hv_apic_eoi_write(void)
+{
+	struct hv_vp_assist_page *hvp = hv_vp_assist_page[smp_processor_id()];
+
+	if (hvp && (xchg(&hvp->apic_assist, 0) & 0x1))
+		return;
+
+	wrmsrq(HV_X64_MSR_EOI, APIC_EOI_ACK);
+}
+
+static bool cpu_is_self(int cpu)
+{
+	return cpu == smp_processor_id();
+}
+
+/*
+ * IPI implementation on Hyper-V.
+ */
+static bool __send_ipi_mask_ex(const struct cpumask *mask, int vector,
+			       bool exclude_self)
+{
+	struct hv_send_ipi_ex *ipi_arg;
+	unsigned long flags;
+	int nr_bank = 0;
+	u64 status = HV_STATUS_INVALID_PARAMETER;
+
+	if (!(ms_hyperv.hints & HV_X64_EX_PROCESSOR_MASKS_RECOMMENDED))
+		return false;
+
+	local_irq_save(flags);
+	ipi_arg = *this_cpu_ptr(hyperv_pcpu_input_arg);
+
+	if (unlikely(!ipi_arg))
+		goto ipi_mask_ex_done;
+
+	ipi_arg->vector = vector;
+	ipi_arg->reserved = 0;
+	ipi_arg->vp_set.valid_bank_mask = 0;
+
+	/*
+	 * Use HV_GENERIC_SET_ALL and avoid converting cpumask to VP_SET
+	 * when the IPI is sent to all currently present CPUs.
+	 */
+	if (!cpumask_equal(mask, cpu_present_mask) || exclude_self) {
+		ipi_arg->vp_set.format = HV_GENERIC_SET_SPARSE_4K;
+
+		nr_bank = cpumask_to_vpset_skip(&ipi_arg->vp_set, mask,
+						exclude_self ? cpu_is_self : NULL);
+
+		/*
+		 * 'nr_bank <= 0' means some CPUs in cpumask can't be
+		 * represented in VP_SET. Return an error and fall back to
+		 * native (architectural) method of sending IPIs.
+		 */
+		if (nr_bank <= 0)
+			goto ipi_mask_ex_done;
+	} else {
+		ipi_arg->vp_set.format = HV_GENERIC_SET_ALL;
+	}
+
+	/*
+	 * For this hypercall, Hyper-V treats the valid_bank_mask field
+	 * of ipi_arg->vp_set as part of the fixed size input header.
+	 * So the variable input header size is equal to nr_bank.
+	 */
+	status = hv_do_rep_hypercall(HVCALL_SEND_IPI_EX, 0, nr_bank,
+				     ipi_arg, NULL);
+
+ipi_mask_ex_done:
+	local_irq_restore(flags);
+	return hv_result_success(status);
+}
+
+static bool __send_ipi_mask(const struct cpumask *mask, int vector,
+			    bool exclude_self)
+{
+	int cur_cpu, vcpu, this_cpu = smp_processor_id();
+	struct hv_send_ipi ipi_arg;
+	u64 status;
+	unsigned int weight;
+
+	trace_hyperv_send_ipi_mask(mask, vector);
+
+	weight = cpumask_weight(mask);
+
+	/*
+	 * Do nothing if
+	 *   1. the mask is empty
+	 *   2. the mask only contains self when exclude_self is true
+	 */
+	if (weight == 0 ||
+	    (exclude_self && weight == 1 && cpumask_test_cpu(this_cpu, mask)))
+		return true;
+
+	/* A fully enlightened TDX VM uses GHCI rather than hv_hypercall_pg. */
+	if (!hv_hypercall_pg) {
+		if (ms_hyperv.paravisor_present || !hv_isolation_type_tdx())
+			return false;
+	}
+
+	if (vector < HV_IPI_LOW_VECTOR || vector > HV_IPI_HIGH_VECTOR)
+		return false;
+
+	/*
+	 * From the supplied CPU set we need to figure out if we can get away
+	 * with cheaper HVCALL_SEND_IPI hypercall. This is possible when the
+	 * highest VP number in the set is < 64. As VP numbers are usually in
+	 * ascending order and match Linux CPU ids, here is an optimization:
+	 * we check the VP number for the highest bit in the supplied set first
+	 * so we can quickly find out if using HVCALL_SEND_IPI_EX hypercall is
+	 * a must. We will also check all VP numbers when walking the supplied
+	 * CPU set to remain correct in all cases.
+	 */
+	if (hv_cpu_number_to_vp_number(cpumask_last(mask)) >= 64)
+		goto do_ex_hypercall;
+
+	ipi_arg.vector = vector;
+	ipi_arg.cpu_mask = 0;
+
+	for_each_cpu(cur_cpu, mask) {
+		if (exclude_self && cur_cpu == this_cpu)
+			continue;
+		vcpu = hv_cpu_number_to_vp_number(cur_cpu);
+		if (vcpu == VP_INVAL)
+			return false;
+
+		/*
+		 * This particular version of the IPI hypercall can
+		 * only target up to 64 CPUs.
+		 */
+		if (vcpu >= 64)
+			goto do_ex_hypercall;
+
+		__set_bit(vcpu, (unsigned long *)&ipi_arg.cpu_mask);
+	}
+
+	status = hv_do_fast_hypercall16(HVCALL_SEND_IPI, ipi_arg.vector,
+					ipi_arg.cpu_mask);
+	return hv_result_success(status);
+
+do_ex_hypercall:
+	return __send_ipi_mask_ex(mask, vector, exclude_self);
+}
+
+static bool __send_ipi_one(int cpu, int vector)
+{
+	int vp = hv_cpu_number_to_vp_number(cpu);
+	u64 status;
+
+	trace_hyperv_send_ipi_one(cpu, vector);
+
+	if (vp == VP_INVAL)
+		return false;
+
+	/* A fully enlightened TDX VM uses GHCI rather than hv_hypercall_pg. */
+	if (!hv_hypercall_pg) {
+		if (ms_hyperv.paravisor_present || !hv_isolation_type_tdx())
+			return false;
+	}
+
+	if (vector < HV_IPI_LOW_VECTOR || vector > HV_IPI_HIGH_VECTOR)
+		return false;
+
+	if (vp >= 64)
+		return __send_ipi_mask_ex(cpumask_of(cpu), vector, false);
+
+	status = hv_do_fast_hypercall16(HVCALL_SEND_IPI, vector, BIT_ULL(vp));
+	return hv_result_success(status);
+}
+
+static void hv_send_ipi(int cpu, int vector)
+{
+	if (!__send_ipi_one(cpu, vector))
+		orig_apic.send_IPI(cpu, vector);
+}
+
+static void hv_send_ipi_mask(const struct cpumask *mask, int vector)
+{
+	if (!__send_ipi_mask(mask, vector, false))
+		orig_apic.send_IPI_mask(mask, vector);
+}
+
+static void hv_send_ipi_mask_allbutself(const struct cpumask *mask, int vector)
+{
+	if (!__send_ipi_mask(mask, vector, true))
+		orig_apic.send_IPI_mask_allbutself(mask, vector);
+}
+
+static void hv_send_ipi_allbutself(int vector)
+{
+	hv_send_ipi_mask_allbutself(cpu_online_mask, vector);
+}
+
+static void hv_send_ipi_all(int vector)
+{
+	if (!__send_ipi_mask(cpu_online_mask, vector, false))
+		orig_apic.send_IPI_all(vector);
+}
+
+static void hv_send_ipi_self(int vector)
+{
+	if (!__send_ipi_one(smp_processor_id(), vector))
+		orig_apic.send_IPI_self(vector);
+}
+
+void __init hv_apic_init(void)
+{
+	if (cc_platform_has(CC_ATTR_SNP_SECURE_AVIC))
+		return;
+
+	if (ms_hyperv.hints & HV_X64_CLUSTER_IPI_RECOMMENDED) {
+		pr_info("Hyper-V: Using IPI hypercalls\n");
+		/*
+		 * Set the IPI entry points.
+		 */
+		orig_apic = *apic;
+
+		apic_update_callback(send_IPI, hv_send_ipi);
+		apic_update_callback(send_IPI_mask, hv_send_ipi_mask);
+		apic_update_callback(send_IPI_mask_allbutself, hv_send_ipi_mask_allbutself);
+		apic_update_callback(send_IPI_allbutself, hv_send_ipi_allbutself);
+		apic_update_callback(send_IPI_all, hv_send_ipi_all);
+		apic_update_callback(send_IPI_self, hv_send_ipi_self);
+	}
+
+	if (ms_hyperv.hints & HV_X64_APIC_ACCESS_RECOMMENDED) {
+		pr_info("Hyper-V: Using enlightened APIC (%s mode)",
+			x2apic_enabled() ? "x2apic" : "xapic");
+		/*
+		 * When in x2apic mode, don't use the Hyper-V specific APIC
+		 * accessors since the field layout in the ICR register is
+		 * different in x2apic mode. Furthermore, the architectural
+		 * x2apic MSRs function just as well as the Hyper-V
+		 * synthetic APIC MSRs, so there's no benefit in having
+		 * separate Hyper-V accessors for x2apic mode. The only
+		 * exception is hv_apic_eoi_write, because it benefits from
+		 * lazy EOI when available, but the same accessor works for
+		 * both xapic and x2apic because the field layout is the same.
+		 */
+		apic_update_callback(eoi, hv_apic_eoi_write);
+		if (!x2apic_enabled()) {
+			apic_update_callback(read, hv_apic_read);
+			apic_update_callback(write, hv_apic_write);
+			apic_update_callback(icr_write, hv_apic_icr_write);
+			apic_update_callback(icr_read, hv_apic_icr_read);
+		}
+	}
+}

linux-7.2.3/arch/x86/hyperv/hv_crash.c

@@ -0,0 +1,647 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * X86 specific Hyper-V root partition kdump/crash support module
+ *
+ * Copyright (C) 2025, Microsoft, Inc.
+ *
+ * This module implements hypervisor RAM collection into vmcore for both
+ * cases of the hypervisor crash and Linux root crash. Hyper-V implements
+ * a disable hypercall with a 32bit protected mode ABI callback. This
+ * mechanism must be used to unlock hypervisor RAM. Since the hypervisor RAM
+ * is already mapped in Linux, it is automatically collected into Linux vmcore,
+ * and can be examined by the crash command (raw RAM dump) or windbg.
+ *
+ * At a high level:
+ *
+ *  Hypervisor Crash:
+ *    Upon crash, hypervisor goes into an emergency minimal dispatch loop, a
+ *    restrictive mode with very limited hypercall and MSR support. Each cpu
+ *    then injects NMIs into root vcpus. A shared page is used to check
+ *    by Linux in the NMI handler if the hypervisor has crashed. This shared
+ *    page is setup in hv_root_crash_init during boot.
+ *
+ *  Linux Crash:
+ *    In case of Linux crash, the callback hv_crash_stop_other_cpus will send
+ *    NMIs to all cpus, then proceed to the crash_nmi_callback where it waits
+ *    for all cpus to be in NMI.
+ *
+ *  NMI Handler (upon quorum):
+ *    Eventually, in both cases, all cpus will end up in the NMI handler.
+ *    Hyper-V requires the disable hypervisor must be done from the BSP. So
+ *    the BSP NMI handler saves current context, does some fixups and makes
+ *    the hypercall to disable the hypervisor, ie, devirtualize. Hypervisor
+ *    at that point will suspend all vcpus (except the BSP), unlock all its
+ *    RAM, and return to Linux at the 32bit mode entry RIP.
+ *
+ *  Linux 32bit entry trampoline will then restore long mode and call C
+ *  function here to restore context and continue execution to crash kexec.
+ */
+
+#include <linux/delay.h>
+#include <linux/kexec.h>
+#include <linux/crash_dump.h>
+#include <linux/panic.h>
+#include <asm/apic.h>
+#include <asm/desc.h>
+#include <asm/page.h>
+#include <asm/pgalloc.h>
+#include <asm/mshyperv.h>
+#include <asm/nmi.h>
+#include <asm/idtentry.h>
+#include <asm/reboot.h>
+#include <asm/intel_pt.h>
+
+bool hv_crash_enabled;
+EXPORT_SYMBOL_GPL(hv_crash_enabled);
+
+struct hv_crash_ctxt {
+	ulong rsp;
+	ulong cr0;
+	ulong cr2;
+	ulong cr4;
+	ulong cr8;
+
+	u16 cs;
+	u16 ss;
+	u16 ds;
+	u16 es;
+	u16 fs;
+	u16 gs;
+
+	u16 gdt_fill;
+	struct desc_ptr gdtr;
+	char idt_fill[6];
+	struct desc_ptr idtr;
+
+	u64 gsbase;
+	u64 efer;
+	u64 pat;
+};
+static struct hv_crash_ctxt hv_crash_ctxt;
+
+/* Shared hypervisor page that contains crash dump area we peek into.
+ * NB: windbg looks for "hv_cda" symbol so don't change it.
+ */
+static struct hv_crashdump_area *hv_cda;
+
+static u32 trampoline_pa, devirt_arg;
+static atomic_t crash_cpus_wait;
+static void *hv_crash_ptpgs[4];
+static bool hv_has_crashed, lx_has_crashed;
+
+static void __noreturn hv_panic_timeout_reboot(void)
+{
+	#define PANIC_TIMER_STEP 100
+
+	if (panic_timeout > 0) {
+		int i;
+
+		for (i = 0; i < panic_timeout * 1000; i += PANIC_TIMER_STEP)
+			mdelay(PANIC_TIMER_STEP);
+	}
+
+	if (panic_timeout)
+		native_wrmsrq(HV_X64_MSR_RESET, 1);    /* get hyp to reboot */
+
+	for (;;)
+		cpu_relax();
+}
+
+static void hv_crash_restore_tss(void)
+{
+	load_TR_desc();
+}
+
+static void hv_crash_clear_kernpt(void)
+{
+	pgd_t *pgd;
+	p4d_t *p4d;
+
+	/* Clear entry so it's not confusing to someone looking at the core */
+	pgd = pgd_offset_k(trampoline_pa);
+	p4d = p4d_offset(pgd, trampoline_pa);
+	native_p4d_clear(p4d);
+}
+
+
+static void __noreturn hv_crash_handle(void)
+{
+	hv_crash_restore_tss();
+	hv_crash_clear_kernpt();
+
+	/* we are now fully in devirtualized normal kernel mode */
+	__crash_kexec(NULL);
+
+	hv_panic_timeout_reboot();
+}
+
+/*
+ * __naked functions do not permit function calls, not even to __always_inline
+ * functions that only contain asm() blocks themselves. So use a macro instead.
+ */
+#define hv_wrmsr(msr, val) \
+	asm volatile("wrmsr" :: "c"(msr), "a"((u32)val), "d"((u32)(val >> 32)) : "memory")
+
+/*
+ * This is the C entry point from the asm glue code after the disable hypercall.
+ * We enter here in IA32-e long mode, ie, full 64bit mode running on kernel
+ * page tables with our below 4G page identity mapped, but using a temporary
+ * GDT. ds/fs/gs/es are null. ss is not usable. bp is null. stack is not
+ * available. We restore kernel GDT, and rest of the context, and continue
+ * to kexec.
+ */
+static void __naked hv_crash_c_entry(void)
+{
+	/* first thing, restore kernel gdt */
+	asm volatile("lgdt %0" : : "m" (hv_crash_ctxt.gdtr));
+
+	asm volatile("movw %0, %%ss\n\t"
+		     "movq %1, %%rsp"
+		     :: "m"(hv_crash_ctxt.ss), "m"(hv_crash_ctxt.rsp));
+
+	asm volatile("movw %0, %%ds" : : "m"(hv_crash_ctxt.ds));
+	asm volatile("movw %0, %%es" : : "m"(hv_crash_ctxt.es));
+	asm volatile("movw %0, %%fs" : : "m"(hv_crash_ctxt.fs));
+	asm volatile("movw %0, %%gs" : : "m"(hv_crash_ctxt.gs));
+
+	hv_wrmsr(MSR_IA32_CR_PAT, hv_crash_ctxt.pat);
+	asm volatile("movq %0, %%cr0" : : "r"(hv_crash_ctxt.cr0));
+
+	asm volatile("movq %0, %%cr8" : : "r"(hv_crash_ctxt.cr8));
+	asm volatile("movq %0, %%cr4" : : "r"(hv_crash_ctxt.cr4));
+	asm volatile("movq %0, %%cr2" : : "r"(hv_crash_ctxt.cr2));
+
+	asm volatile("lidt %0" : : "m" (hv_crash_ctxt.idtr));
+	hv_wrmsr(MSR_GS_BASE, hv_crash_ctxt.gsbase);
+	hv_wrmsr(MSR_EFER, hv_crash_ctxt.efer);
+
+	/* restore the original kernel CS now via far return */
+	asm volatile("pushq %q0\n\t"
+		     "pushq %q1\n\t"
+		     "lretq"
+		     :: "r"(hv_crash_ctxt.cs), "r"(hv_crash_handle));
+}
+/* Tell objtool we are using lretq long jump in the above function intentionally */
+STACK_FRAME_NON_STANDARD(hv_crash_c_entry);
+
+static void hv_mark_tss_not_busy(void)
+{
+	struct desc_struct *desc = get_current_gdt_rw();
+	tss_desc tss;
+
+	memcpy(&tss, &desc[GDT_ENTRY_TSS], sizeof(tss_desc));
+	tss.type = 0x9;        /* available 64-bit TSS. 0xB is busy TSS */
+	write_gdt_entry(desc, GDT_ENTRY_TSS, &tss, DESC_TSS);
+}
+
+/* Save essential context */
+static void hv_hvcrash_ctxt_save(void)
+{
+	struct hv_crash_ctxt *ctxt = &hv_crash_ctxt;
+
+	ctxt->rsp = current_stack_pointer;
+
+	ctxt->cr0 = native_read_cr0();
+	ctxt->cr4 = native_read_cr4();
+
+	asm volatile("movq %%cr2, %0" : "=r"(ctxt->cr2));
+	asm volatile("movq %%cr8, %0" : "=r"(ctxt->cr8));
+
+	asm volatile("movw %%cs, %0" : "=m"(ctxt->cs));
+	asm volatile("movw %%ss, %0" : "=m"(ctxt->ss));
+	asm volatile("movw %%ds, %0" : "=m"(ctxt->ds));
+	asm volatile("movw %%es, %0" : "=m"(ctxt->es));
+	asm volatile("movw %%fs, %0" : "=m"(ctxt->fs));
+	asm volatile("movw %%gs, %0" : "=m"(ctxt->gs));
+
+	native_store_gdt(&ctxt->gdtr);
+	store_idt(&ctxt->idtr);
+
+	ctxt->gsbase = __rdmsr(MSR_GS_BASE);
+	ctxt->efer = __rdmsr(MSR_EFER);
+	ctxt->pat = __rdmsr(MSR_IA32_CR_PAT);
+}
+
+/* Add trampoline page to the kernel pagetable for transition to kernel PT */
+static void hv_crash_fixup_kernpt(void)
+{
+	pgd_t *pgd;
+	p4d_t *p4d;
+
+	pgd = pgd_offset_k(trampoline_pa);
+	p4d = p4d_offset(pgd, trampoline_pa);
+
+	/* trampoline_pa is below 4G, so no pre-existing entry to clobber */
+	p4d_populate(&init_mm, p4d, (pud_t *)hv_crash_ptpgs[1]);
+	p4d->p4d = p4d->p4d & ~(_PAGE_NX);    /* enable execute */
+}
+
+/*
+ * Notify the hyp that Linux has crashed. This will cause the hyp to quiesce
+ * and suspend all guest VPs.
+ */
+static void hv_notify_prepare_hyp(void)
+{
+	u64 status;
+	struct hv_input_notify_partition_event *input;
+	struct hv_partition_event_root_crashdump_input *cda;
+
+	input = *this_cpu_ptr(hyperv_pcpu_input_arg);
+	cda = &input->input.crashdump_input;
+	memset(input, 0, sizeof(*input));
+	input->event = HV_PARTITION_EVENT_ROOT_CRASHDUMP;
+
+	cda->crashdump_action = HV_CRASHDUMP_ENTRY;
+	status = hv_do_hypercall(HVCALL_NOTIFY_PARTITION_EVENT, input, NULL);
+	if (!hv_result_success(status))
+		return;
+
+	cda->crashdump_action = HV_CRASHDUMP_SUSPEND_ALL_VPS;
+	hv_do_hypercall(HVCALL_NOTIFY_PARTITION_EVENT, input, NULL);
+}
+
+/*
+ * Common function for all cpus before devirtualization.
+ *
+ * Hypervisor crash: all cpus get here in NMI context.
+ * Linux crash: the panicing cpu gets here at base level, all others in NMI
+ *		context. Note, panicing cpu may not be the BSP.
+ *
+ * The function is not inlined so it will show on the stack. It is named so
+ * because the crash cmd looks for certain well known function names on the
+ * stack before looking into the cpu saved note in the elf section, and
+ * that work is currently incomplete.
+ *
+ * Notes:
+ *  Hypervisor crash:
+ *    - the hypervisor is in a very restrictive mode at this point and any
+ *	vmexit it cannot handle would result in reboot. So, no mumbo jumbo,
+ *	just get to kexec as quickly as possible.
+ *
+ *  Devirtualization is supported from the BSP only at present.
+ */
+static noinline __noclone void crash_nmi_callback(struct pt_regs *regs)
+{
+	struct hv_input_disable_hyp_ex *input;
+	int msecs = 1000, ccpu = smp_processor_id();
+
+	if (ccpu == 0) {
+		/* crash_save_cpu() will be done in the kexec path */
+		cpu_emergency_stop_pt();	/* disable performance trace */
+		atomic_inc(&crash_cpus_wait);
+	} else {
+		crash_save_cpu(regs, ccpu);
+		cpu_emergency_stop_pt();	/* disable performance trace */
+		atomic_inc(&crash_cpus_wait);
+		for (;;)
+			cpu_relax();
+	}
+
+	while (atomic_read(&crash_cpus_wait) < num_online_cpus() && msecs--)
+		mdelay(1);
+
+	stop_nmi();
+	if (!hv_has_crashed)
+		hv_notify_prepare_hyp();
+
+	if (crashing_cpu == -1)
+		crashing_cpu = ccpu;		/* crash cmd uses this */
+
+	hv_hvcrash_ctxt_save();
+	hv_mark_tss_not_busy();
+	hv_crash_fixup_kernpt();
+
+	input = *this_cpu_ptr(hyperv_pcpu_input_arg);
+	memset(input, 0, sizeof(*input));
+	input->rip = trampoline_pa;
+	input->arg = devirt_arg;
+
+	(void)hv_do_hypercall(HVCALL_DISABLE_HYP_EX, input, NULL);
+
+	hv_panic_timeout_reboot();
+}
+
+
+static DEFINE_SPINLOCK(hv_crash_reboot_lk);
+
+/*
+ * Generic NMI callback handler: could be called without any crash also.
+ *   hv crash: hypervisor injects NMI's into all cpus
+ *   lx crash: panicing cpu sends NMI to all but self via crash_stop_other_cpus
+ */
+static int hv_crash_nmi_local(unsigned int cmd, struct pt_regs *regs)
+{
+	if (!hv_has_crashed && hv_cda && hv_cda->cda_valid)
+		hv_has_crashed = true;
+
+	if (!hv_has_crashed && !lx_has_crashed)
+		return NMI_DONE;	/* ignore the NMI */
+
+	if (hv_has_crashed && !kexec_crash_loaded()) {
+		if (spin_trylock(&hv_crash_reboot_lk))
+			hv_panic_timeout_reboot();
+		else
+			for (;;)
+				cpu_relax();
+	}
+
+	crash_nmi_callback(regs);
+
+	return NMI_DONE;
+}
+
+/*
+ * hv_crash_stop_other_cpus() == smp_ops.crash_stop_other_cpus
+ *
+ * On normal Linux panic, this is called twice: first from panic and then again
+ * from native_machine_crash_shutdown.
+ *
+ * In case of hyperv, 3 ways to get here:
+ *  1. hv crash (only BSP will get here):
+ *	BSP : NMI callback -> DisableHv -> hv_crash_asm32 -> hv_crash_c_entry
+ *		  -> __crash_kexec -> native_machine_crash_shutdown
+ *		  -> crash_smp_send_stop -> smp_ops.crash_stop_other_cpus
+ *  Linux panic:
+ *	2. panic cpu x: panic() -> crash_smp_send_stop
+ *				     -> smp_ops.crash_stop_other_cpus
+ *	3. BSP: native_machine_crash_shutdown -> crash_smp_send_stop
+ *
+ * NB: noclone and non standard stack because of call to crash_setup_regs().
+ */
+static void __noclone hv_crash_stop_other_cpus(void)
+{
+	static bool crash_stop_done;
+	struct pt_regs lregs;
+	int ccpu = smp_processor_id();
+
+	if (hv_has_crashed)
+		return;		/* all cpus already in NMI handler path */
+
+	if (!kexec_crash_loaded()) {
+		hv_notify_prepare_hyp();
+		hv_panic_timeout_reboot();	/* no return */
+	}
+
+	/* If the hv crashes also, we could come here again before cpus_stopped
+	 * is set in crash_smp_send_stop(). So use our own check.
+	 */
+	if (crash_stop_done)
+		return;
+	crash_stop_done = true;
+
+	/* Linux has crashed: hv is healthy, we can IPI safely */
+	lx_has_crashed = true;
+	wmb();			/* NMI handlers look at lx_has_crashed */
+
+	apic->send_IPI_allbutself(NMI_VECTOR);
+
+	if (crashing_cpu == -1)
+		crashing_cpu = ccpu;		/* crash cmd uses this */
+
+	/* crash_setup_regs() happens in kexec also, but for the kexec cpu which
+	 * is the BSP. We could be here on non-BSP cpu, collect regs if so.
+	 */
+	if (ccpu)
+		crash_setup_regs(&lregs, NULL);
+
+	crash_nmi_callback(&lregs);
+}
+STACK_FRAME_NON_STANDARD(hv_crash_stop_other_cpus);
+
+/* This GDT is accessed in IA32-e compat mode which uses 32bits addresses */
+struct hv_gdtreg_32 {
+	u16 fill;
+	u16 limit;
+	u32 address;
+} __packed;
+
+/* We need a CS with L bit to goto IA32-e long mode from 32bit compat mode */
+struct hv_crash_tramp_gdt {
+	u64 null;	/* index 0, selector 0, null selector */
+	u64 cs64;	/* index 1, selector 8, cs64 selector */
+} __packed;
+
+/* No stack, so jump via far ptr in memory to load the 64bit CS */
+struct hv_cs_jmptgt {
+	u32 address;
+	u16 csval;
+	u16 fill;
+} __packed;
+
+/* Linux use only, hypervisor doesn't look at this struct */
+struct hv_crash_tramp_data {
+	u64 tramp32_cr3;
+	u64 kernel_cr3;
+	struct hv_gdtreg_32 gdtr32;
+	struct hv_crash_tramp_gdt tramp_gdt;
+	struct hv_cs_jmptgt cs_jmptgt;
+	u64 c_entry_addr;
+} __packed;
+
+/*
+ * Setup a temporary gdt to allow the asm code to switch to the long mode.
+ * Since the asm code is relocated/copied to a below 4G page, it cannot use rip
+ * relative addressing, hence we must use trampoline_pa here. Also, save other
+ * info like jmp and C entry targets for same reasons.
+ *
+ * Returns: 0 on success, -1 on error
+ */
+static int hv_crash_setup_trampdata(u64 trampoline_va)
+{
+	int size, offs;
+	void *dest;
+	struct hv_crash_tramp_data *tramp;
+
+	/* These must match exactly the ones in the corresponding asm file */
+	BUILD_BUG_ON(offsetof(struct hv_crash_tramp_data, tramp32_cr3) != 0);
+	BUILD_BUG_ON(offsetof(struct hv_crash_tramp_data, kernel_cr3) != 8);
+	BUILD_BUG_ON(offsetof(struct hv_crash_tramp_data, gdtr32.limit) != 18);
+	BUILD_BUG_ON(offsetof(struct hv_crash_tramp_data,
+						     cs_jmptgt.address) != 40);
+	BUILD_BUG_ON(offsetof(struct hv_crash_tramp_data, c_entry_addr) != 48);
+
+	/* hv_crash_asm_end is beyond last byte by 1 */
+	size = &hv_crash_asm_end - &hv_crash_asm32;
+	if (size + sizeof(struct hv_crash_tramp_data) > PAGE_SIZE) {
+		pr_err("%s: trampoline page overflow\n", __func__);
+		return -1;
+	}
+
+	dest = (void *)trampoline_va;
+	memcpy(dest, &hv_crash_asm32, size);
+
+	dest += size;
+	dest = (void *)round_up((ulong)dest, 16);
+	tramp = (struct hv_crash_tramp_data *)dest;
+
+	/* see MAX_ASID_AVAILABLE in tlb.c: "PCID 0 is reserved for use by
+	 * non-PCID-aware users". Build cr3 with pcid 0
+	 */
+	tramp->tramp32_cr3 = __sme_pa(hv_crash_ptpgs[0]);
+
+	/* Note, when restoring X86_CR4_PCIDE, cr3[11:0] must be zero */
+	tramp->kernel_cr3 = __sme_pa(init_mm.pgd);
+
+	tramp->gdtr32.limit = sizeof(struct hv_crash_tramp_gdt);
+	tramp->gdtr32.address = trampoline_pa +
+				   (ulong)&tramp->tramp_gdt - trampoline_va;
+
+	 /* base:0 limit:0xfffff type:b dpl:0 P:1 L:1 D:0 avl:0 G:1 */
+	tramp->tramp_gdt.cs64 = 0x00af9a000000ffff;
+
+	tramp->cs_jmptgt.csval = 0x8;
+	offs = (ulong)&hv_crash_asm64 - (ulong)&hv_crash_asm32;
+	tramp->cs_jmptgt.address = trampoline_pa + offs;
+
+	tramp->c_entry_addr = (u64)&hv_crash_c_entry;
+
+	devirt_arg = trampoline_pa + (ulong)dest - trampoline_va;
+
+	return 0;
+}
+
+/*
+ * Build 32bit trampoline page table for transition from protected mode
+ * non-paging to long-mode paging. This transition needs pagetables below 4G.
+ */
+static void hv_crash_build_tramp_pt(void)
+{
+	p4d_t *p4d;
+	pud_t *pud;
+	pmd_t *pmd;
+	pte_t *pte;
+	u64 pa, addr = trampoline_pa;
+
+	p4d = hv_crash_ptpgs[0] + pgd_index(addr) * sizeof(p4d);
+	pa = virt_to_phys(hv_crash_ptpgs[1]);
+	set_p4d(p4d, __p4d(_PAGE_TABLE | pa));
+	p4d->p4d &= ~(_PAGE_NX);	/* enable execute */
+
+	pud = hv_crash_ptpgs[1] + pud_index(addr) * sizeof(pud);
+	pa = virt_to_phys(hv_crash_ptpgs[2]);
+	set_pud(pud, __pud(_PAGE_TABLE | pa));
+
+	pmd = hv_crash_ptpgs[2] + pmd_index(addr) * sizeof(pmd);
+	pa = virt_to_phys(hv_crash_ptpgs[3]);
+	set_pmd(pmd, __pmd(_PAGE_TABLE | pa));
+
+	pte = hv_crash_ptpgs[3] + pte_index(addr) * sizeof(pte);
+	set_pte(pte, pfn_pte(addr >> PAGE_SHIFT, PAGE_KERNEL_EXEC));
+}
+
+/*
+ * Setup trampoline for devirtualization:
+ *  - a page below 4G, ie 32bit addr containing asm glue code that hyp jmps to
+ *    in protected mode.
+ *  - 4 pages for a temporary page table that asm code uses to turn paging on
+ *  - a temporary gdt to use in the compat mode.
+ *
+ *  Returns: 0 on success
+ */
+static int hv_crash_trampoline_setup(void)
+{
+	int i, rc, order;
+	struct page *page;
+	u64 trampoline_va;
+	gfp_t flags32 = GFP_KERNEL | GFP_DMA32 | __GFP_ZERO;
+
+	/* page for 32bit trampoline assembly code + hv_crash_tramp_data */
+	page = alloc_page(flags32);
+	if (page == NULL) {
+		pr_err("%s: failed to alloc asm stub page\n", __func__);
+		return -1;
+	}
+
+	trampoline_va = (u64)page_to_virt(page);
+	trampoline_pa = (u32)page_to_phys(page);
+
+	order = 2;	   /* alloc 2^2 pages */
+	page = alloc_pages(flags32, order);
+	if (page == NULL) {
+		pr_err("%s: failed to alloc pt pages\n", __func__);
+		free_page(trampoline_va);
+		return -1;
+	}
+
+	for (i = 0; i < 4; i++, page++)
+		hv_crash_ptpgs[i] = page_to_virt(page);
+
+	hv_crash_build_tramp_pt();
+
+	rc = hv_crash_setup_trampdata(trampoline_va);
+	if (rc)
+		goto errout;
+
+	return 0;
+
+errout:
+	free_page(trampoline_va);
+	free_pages((ulong)hv_crash_ptpgs[0], order);
+
+	return rc;
+}
+
+/* Setup for kdump kexec to collect hypervisor RAM when running as root */
+void hv_root_crash_init(void)
+{
+	int rc;
+	struct hv_input_get_system_property *input;
+	struct hv_output_get_system_property *output;
+	unsigned long flags;
+	u64 status;
+	union hv_pfn_range cda_info;
+
+	if (pgtable_l5_enabled()) {
+		pr_err("Hyper-V: crash dump not yet supported on 5level PTs\n");
+		return;
+	}
+
+	rc = register_nmi_handler(NMI_LOCAL, hv_crash_nmi_local, NMI_FLAG_FIRST,
+				  "hv_crash_nmi");
+	if (rc) {
+		pr_err("Hyper-V: failed to register crash nmi handler\n");
+		return;
+	}
+
+	local_irq_save(flags);
+	input = *this_cpu_ptr(hyperv_pcpu_input_arg);
+	output = *this_cpu_ptr(hyperv_pcpu_output_arg);
+
+	memset(input, 0, sizeof(*input));
+	input->property_id = HV_SYSTEM_PROPERTY_CRASHDUMPAREA;
+
+	status = hv_do_hypercall(HVCALL_GET_SYSTEM_PROPERTY, input, output);
+	cda_info.as_uint64 = output->hv_cda_info.as_uint64;
+	local_irq_restore(flags);
+
+	if (!hv_result_success(status)) {
+		pr_err("Hyper-V: %s: property:%d %s\n", __func__,
+		       input->property_id, hv_result_to_string(status));
+		goto err_out;
+	}
+
+	if (cda_info.base_pfn == 0) {
+		pr_err("Hyper-V: hypervisor crash dump area pfn is 0\n");
+		goto err_out;
+	}
+
+	hv_cda = phys_to_virt(cda_info.base_pfn << HV_HYP_PAGE_SHIFT);
+
+	rc = hv_crash_trampoline_setup();
+	if (rc)
+		goto err_out;
+
+#ifdef CONFIG_SMP
+	smp_ops.crash_stop_other_cpus = hv_crash_stop_other_cpus;
+#endif
+
+	crash_kexec_post_notifiers = true;
+	hv_crash_enabled = true;
+	pr_info("Hyper-V: both linux and hypervisor kdump support enabled\n");
+
+	return;
+
+err_out:
+	unregister_nmi_handler(NMI_LOCAL, "hv_crash_nmi");
+	pr_err("Hyper-V: only linux root kdump support enabled\n");
+}

linux-7.2.3/arch/x86/hyperv/hv_init.c

@@ -0,0 +1,745 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * X86 specific Hyper-V initialization code.
+ *
+ * Copyright (C) 2016, Microsoft, Inc.
+ *
+ * Author : K. Y. Srinivasan <kys@microsoft.com>
+ */
+
+#define pr_fmt(fmt)  "Hyper-V: " fmt
+
+#include <linux/efi.h>
+#include <linux/types.h>
+#include <linux/bitfield.h>
+#include <linux/io.h>
+#include <asm/apic.h>
+#include <asm/desc.h>
+#include <asm/e820/api.h>
+#include <asm/sev.h>
+#include <asm/hypervisor.h>
+#include <hyperv/hvhdk.h>
+#include <asm/mshyperv.h>
+#include <asm/msr.h>
+#include <asm/idtentry.h>
+#include <asm/set_memory.h>
+#include <linux/kexec.h>
+#include <linux/version.h>
+#include <linux/vmalloc.h>
+#include <linux/mm.h>
+#include <linux/slab.h>
+#include <linux/kernel.h>
+#include <linux/cpuhotplug.h>
+#include <linux/syscore_ops.h>
+#include <clocksource/hyperv_timer.h>
+#include <linux/highmem.h>
+#include <linux/export.h>
+
+void *hv_hypercall_pg;
+
+#ifdef CONFIG_X86_64
+static u64 __hv_hyperfail(u64 control, u64 param1, u64 param2)
+{
+	return U64_MAX;
+}
+
+DEFINE_STATIC_CALL(__hv_hypercall, __hv_hyperfail);
+
+u64 hv_std_hypercall(u64 control, u64 param1, u64 param2)
+{
+	u64 hv_status;
+
+	register u64 __r8 asm("r8") = param2;
+	asm volatile ("call " STATIC_CALL_TRAMP_STR(__hv_hypercall)
+		      : "=a" (hv_status), ASM_CALL_CONSTRAINT,
+		        "+c" (control), "+d" (param1), "+r" (__r8)
+		      : : "cc", "memory", "r9", "r10", "r11");
+
+	return hv_status;
+}
+
+typedef u64 (*hv_hypercall_f)(u64 control, u64 param1, u64 param2);
+
+static inline void hv_set_hypercall_pg(void *ptr)
+{
+	hv_hypercall_pg = ptr;
+
+	if (!ptr)
+		ptr = &__hv_hyperfail;
+	static_call_update(__hv_hypercall, (hv_hypercall_f)ptr);
+}
+#else
+static inline void hv_set_hypercall_pg(void *ptr)
+{
+	hv_hypercall_pg = ptr;
+}
+EXPORT_SYMBOL_GPL(hv_hypercall_pg);
+#endif
+
+union hv_ghcb * __percpu *hv_ghcb_pg;
+
+/* Storage to save the hypercall page temporarily for hibernation */
+static void *hv_hypercall_pg_saved;
+
+struct hv_vp_assist_page **hv_vp_assist_page;
+EXPORT_SYMBOL_GPL(hv_vp_assist_page);
+
+static int hyperv_init_ghcb(void)
+{
+	u64 ghcb_gpa;
+	void *ghcb_va;
+	void **ghcb_base;
+
+	if (!ms_hyperv.paravisor_present || !hv_isolation_type_snp())
+		return 0;
+
+	if (!hv_ghcb_pg)
+		return -EINVAL;
+
+	/*
+	 * GHCB page is allocated by paravisor. The address
+	 * returned by MSR_AMD64_SEV_ES_GHCB is above shared
+	 * memory boundary and map it here.
+	 */
+	rdmsrq(MSR_AMD64_SEV_ES_GHCB, ghcb_gpa);
+
+	/* Mask out vTOM bit and map as decrypted */
+	ghcb_gpa &= ~ms_hyperv.shared_gpa_boundary;
+	ghcb_va = memremap(ghcb_gpa, HV_HYP_PAGE_SIZE, MEMREMAP_WB | MEMREMAP_DEC);
+	if (!ghcb_va)
+		return -ENOMEM;
+
+	ghcb_base = (void **)this_cpu_ptr(hv_ghcb_pg);
+	*ghcb_base = ghcb_va;
+
+	return 0;
+}
+
+static int hv_cpu_init(unsigned int cpu)
+{
+	union hv_vp_assist_msr_contents msr = { 0 };
+	struct hv_vp_assist_page **hvp;
+	int ret;
+
+	ret = hv_common_cpu_init(cpu);
+	if (ret)
+		return ret;
+
+	if (!hv_vp_assist_page)
+		return 0;
+
+	hvp = &hv_vp_assist_page[cpu];
+	if (hv_root_partition()) {
+		/*
+		 * For root partition we get the hypervisor provided VP assist
+		 * page, instead of allocating a new page.
+		 */
+		rdmsrq(HV_X64_MSR_VP_ASSIST_PAGE, msr.as_uint64);
+		*hvp = memremap(msr.pfn << HV_X64_MSR_VP_ASSIST_PAGE_ADDRESS_SHIFT,
+				PAGE_SIZE, MEMREMAP_WB);
+	} else {
+		/*
+		 * The VP assist page is an "overlay" page (see Hyper-V TLFS's
+		 * Section 5.2.1 "GPA Overlay Pages"). Here it must be zeroed
+		 * out to make sure we always write the EOI MSR in
+		 * hv_apic_eoi_write() *after* the EOI optimization is disabled
+		 * in hv_cpu_die(), otherwise a CPU may not be stopped in the
+		 * case of CPU offlining and the VM will hang.
+		 */
+		if (!*hvp) {
+			*hvp = __vmalloc(PAGE_SIZE, GFP_KERNEL | __GFP_ZERO);
+
+			/*
+			 * Hyper-V should never specify a VM that is a Confidential
+			 * VM and also running in the root partition. Root partition
+			 * is blocked to run in Confidential VM. So only decrypt assist
+			 * page in non-root partition here.
+			 */
+			if (*hvp && !ms_hyperv.paravisor_present && hv_isolation_type_snp()) {
+				WARN_ON_ONCE(set_memory_decrypted((unsigned long)(*hvp), 1));
+				memset(*hvp, 0, PAGE_SIZE);
+			}
+		}
+
+		if (*hvp)
+			msr.pfn = vmalloc_to_pfn(*hvp);
+
+	}
+	if (!WARN_ON(!(*hvp))) {
+		msr.enable = 1;
+		wrmsrq(HV_X64_MSR_VP_ASSIST_PAGE, msr.as_uint64);
+	}
+
+	/* Allow Hyper-V stimer vector to be injected from Hypervisor. */
+	if (ms_hyperv.misc_features & HV_STIMER_DIRECT_MODE_AVAILABLE)
+		apic_update_vector(cpu, HYPERV_STIMER0_VECTOR, true);
+
+	return hyperv_init_ghcb();
+}
+
+static void (*hv_reenlightenment_cb)(void);
+
+static void hv_reenlightenment_notify(struct work_struct *dummy)
+{
+	struct hv_tsc_emulation_status emu_status;
+
+	rdmsrq(HV_X64_MSR_TSC_EMULATION_STATUS, *(u64 *)&emu_status);
+
+	/* Don't issue the callback if TSC accesses are not emulated */
+	if (hv_reenlightenment_cb && emu_status.inprogress)
+		hv_reenlightenment_cb();
+}
+static DECLARE_DELAYED_WORK(hv_reenlightenment_work, hv_reenlightenment_notify);
+
+void hyperv_stop_tsc_emulation(void)
+{
+	u64 freq;
+	struct hv_tsc_emulation_status emu_status;
+
+	rdmsrq(HV_X64_MSR_TSC_EMULATION_STATUS, *(u64 *)&emu_status);
+	emu_status.inprogress = 0;
+	wrmsrq(HV_X64_MSR_TSC_EMULATION_STATUS, *(u64 *)&emu_status);
+
+	rdmsrq(HV_X64_MSR_TSC_FREQUENCY, freq);
+	tsc_khz = div64_u64(freq, 1000);
+}
+EXPORT_SYMBOL_GPL(hyperv_stop_tsc_emulation);
+
+static inline bool hv_reenlightenment_available(void)
+{
+	/*
+	 * Check for required features and privileges to make TSC frequency
+	 * change notifications work.
+	 */
+	return ms_hyperv.features & HV_ACCESS_FREQUENCY_MSRS &&
+		ms_hyperv.misc_features & HV_FEATURE_FREQUENCY_MSRS_AVAILABLE &&
+		ms_hyperv.features & HV_ACCESS_REENLIGHTENMENT;
+}
+
+DEFINE_IDTENTRY_SYSVEC(sysvec_hyperv_reenlightenment)
+{
+	apic_eoi();
+	inc_irq_stat(HYPERV_REENLIGHTENMENT);
+	schedule_delayed_work(&hv_reenlightenment_work, HZ/10);
+}
+
+void set_hv_tscchange_cb(void (*cb)(void))
+{
+	struct hv_reenlightenment_control re_ctrl = {
+		.vector = HYPERV_REENLIGHTENMENT_VECTOR,
+		.enabled = 1,
+	};
+	struct hv_tsc_emulation_control emu_ctrl = {.enabled = 1};
+
+	if (!hv_reenlightenment_available()) {
+		pr_warn("reenlightenment support is unavailable\n");
+		return;
+	}
+
+	if (!hv_vp_index)
+		return;
+
+	hv_reenlightenment_cb = cb;
+
+	/* Make sure callback is registered before we write to MSRs */
+	wmb();
+
+	re_ctrl.target_vp = hv_vp_index[get_cpu()];
+
+	wrmsrq(HV_X64_MSR_REENLIGHTENMENT_CONTROL, *((u64 *)&re_ctrl));
+	wrmsrq(HV_X64_MSR_TSC_EMULATION_CONTROL, *((u64 *)&emu_ctrl));
+
+	put_cpu();
+}
+EXPORT_SYMBOL_GPL(set_hv_tscchange_cb);
+
+void clear_hv_tscchange_cb(void)
+{
+	struct hv_reenlightenment_control re_ctrl;
+
+	if (!hv_reenlightenment_available())
+		return;
+
+	rdmsrq(HV_X64_MSR_REENLIGHTENMENT_CONTROL, *(u64 *)&re_ctrl);
+	re_ctrl.enabled = 0;
+	wrmsrq(HV_X64_MSR_REENLIGHTENMENT_CONTROL, *(u64 *)&re_ctrl);
+
+	hv_reenlightenment_cb = NULL;
+}
+EXPORT_SYMBOL_GPL(clear_hv_tscchange_cb);
+
+static int hv_cpu_die(unsigned int cpu)
+{
+	struct hv_reenlightenment_control re_ctrl;
+	unsigned int new_cpu;
+	void **ghcb_va;
+
+	if (hv_ghcb_pg) {
+		ghcb_va = (void **)this_cpu_ptr(hv_ghcb_pg);
+		if (*ghcb_va)
+			memunmap(*ghcb_va);
+		*ghcb_va = NULL;
+	}
+
+	if (ms_hyperv.misc_features & HV_STIMER_DIRECT_MODE_AVAILABLE)
+		apic_update_vector(cpu, HYPERV_STIMER0_VECTOR, false);
+
+	hv_common_cpu_die(cpu);
+
+	if (hv_vp_assist_page && hv_vp_assist_page[cpu]) {
+		union hv_vp_assist_msr_contents msr = { 0 };
+		if (hv_root_partition()) {
+			/*
+			 * For root partition the VP assist page is mapped to
+			 * hypervisor provided page, and thus we unmap the
+			 * page here and nullify it, so that in future we have
+			 * correct page address mapped in hv_cpu_init.
+			 */
+			memunmap(hv_vp_assist_page[cpu]);
+			hv_vp_assist_page[cpu] = NULL;
+			rdmsrq(HV_X64_MSR_VP_ASSIST_PAGE, msr.as_uint64);
+			msr.enable = 0;
+		}
+		wrmsrq(HV_X64_MSR_VP_ASSIST_PAGE, msr.as_uint64);
+	}
+
+	if (hv_reenlightenment_cb == NULL)
+		return 0;
+
+	rdmsrq(HV_X64_MSR_REENLIGHTENMENT_CONTROL, *((u64 *)&re_ctrl));
+	if (re_ctrl.target_vp == hv_vp_index[cpu]) {
+		/*
+		 * Reassign reenlightenment notifications to some other online
+		 * CPU or just disable the feature if there are no online CPUs
+		 * left (happens on hibernation).
+		 */
+		new_cpu = cpumask_any_but(cpu_online_mask, cpu);
+
+		if (new_cpu < nr_cpu_ids)
+			re_ctrl.target_vp = hv_vp_index[new_cpu];
+		else
+			re_ctrl.enabled = 0;
+
+		wrmsrq(HV_X64_MSR_REENLIGHTENMENT_CONTROL, *((u64 *)&re_ctrl));
+	}
+
+	return 0;
+}
+
+static int __init hv_pci_init(void)
+{
+	bool gen2vm = efi_enabled(EFI_BOOT);
+
+	/*
+	 * A Generation-2 VM doesn't support legacy PCI/PCIe, so both
+	 * raw_pci_ops and raw_pci_ext_ops are NULL, and pci_subsys_init() ->
+	 * pcibios_init() doesn't call pcibios_resource_survey() ->
+	 * e820__reserve_resources_late(); as a result, any emulated persistent
+	 * memory of E820_TYPE_PRAM (12) via the kernel parameter
+	 * memmap=nn[KMG]!ss is not added into iomem_resource and hence can't be
+	 * detected by register_e820_pmem(). Fix this by directly calling
+	 * e820__reserve_resources_late() here: e820__reserve_resources_late()
+	 * depends on e820__reserve_resources(), which has been called earlier
+	 * from setup_arch(). Note: e820__reserve_resources_late() also adds
+	 * any memory of E820_TYPE_PMEM (7) into iomem_resource, and
+	 * acpi_nfit_register_region() -> acpi_nfit_insert_resource() ->
+	 * region_intersects() returns REGION_INTERSECTS, so the memory of
+	 * E820_TYPE_PMEM won't get added twice.
+	 *
+	 * We return 0 here so that pci_arch_init() won't print the warning:
+	 * "PCI: Fatal: No config space access function found"
+	 */
+	if (gen2vm) {
+		e820__reserve_resources_late();
+		return 0;
+	}
+
+	/* For Generation-1 VM, we'll proceed in pci_arch_init().  */
+	return 1;
+}
+
+static int hv_suspend(void *data)
+{
+	union hv_x64_msr_hypercall_contents hypercall_msr;
+	int ret;
+
+	if (hv_root_partition())
+		return -EPERM;
+
+	/*
+	 * Reset the hypercall page as it is going to be invalidated
+	 * across hibernation. Setting hv_hypercall_pg to NULL ensures
+	 * that any subsequent hypercall operation fails safely instead of
+	 * crashing due to an access of an invalid page. The hypercall page
+	 * pointer is restored on resume.
+	 */
+	hv_hypercall_pg_saved = hv_hypercall_pg;
+	hv_set_hypercall_pg(NULL);
+
+	/* Disable the hypercall page in the hypervisor */
+	rdmsrq(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);
+	hypercall_msr.enable = 0;
+	wrmsrq(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);
+
+	ret = hv_cpu_die(0);
+	return ret;
+}
+
+static void hv_resume(void *data)
+{
+	union hv_x64_msr_hypercall_contents hypercall_msr;
+	int ret;
+
+	ret = hv_cpu_init(0);
+	WARN_ON(ret);
+
+	/* Re-enable the hypercall page */
+	rdmsrq(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);
+	hypercall_msr.enable = 1;
+	hypercall_msr.guest_physical_address =
+		vmalloc_to_pfn(hv_hypercall_pg_saved);
+	wrmsrq(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);
+
+	hv_set_hypercall_pg(hv_hypercall_pg_saved);
+	hv_hypercall_pg_saved = NULL;
+
+	/*
+	 * Reenlightenment notifications are disabled by hv_cpu_die(0),
+	 * reenable them here if hv_reenlightenment_cb was previously set.
+	 */
+	if (hv_reenlightenment_cb)
+		set_hv_tscchange_cb(hv_reenlightenment_cb);
+}
+
+/* Note: when the ops are called, only CPU0 is online and IRQs are disabled. */
+static const struct syscore_ops hv_syscore_ops = {
+	.suspend	= hv_suspend,
+	.resume		= hv_resume,
+};
+
+static struct syscore hv_syscore = {
+	.ops = &hv_syscore_ops,
+};
+
+static void (* __initdata old_setup_percpu_clockev)(void);
+
+static void __init hv_stimer_setup_percpu_clockev(void)
+{
+	/*
+	 * Ignore any errors in setting up stimer clockevents
+	 * as we can run with the LAPIC timer as a fallback.
+	 */
+	(void)hv_stimer_alloc(false);
+
+	/*
+	 * Still register the LAPIC timer, because the direct-mode STIMER is
+	 * not supported by old versions of Hyper-V. This also allows users
+	 * to switch to LAPIC timer via /sys, if they want to.
+	 */
+	if (old_setup_percpu_clockev)
+		old_setup_percpu_clockev();
+}
+
+/*
+ * This function is to be invoked early in the boot sequence after the
+ * hypervisor has been detected.
+ *
+ * 1. Setup the hypercall page.
+ * 2. Register Hyper-V specific clocksource.
+ * 3. Setup Hyper-V specific APIC entry points.
+ */
+void __init hyperv_init(void)
+{
+	u64 guest_id;
+	union hv_x64_msr_hypercall_contents hypercall_msr;
+	int cpuhp;
+
+	if (x86_hyper_type != X86_HYPER_MS_HYPERV)
+		return;
+
+	if (hv_common_init())
+		return;
+
+	/*
+	 * The VP assist page is useless to a TDX guest: the only use we
+	 * would have for it is lazy EOI, which can not be used with TDX.
+	 */
+	if (hv_isolation_type_tdx())
+		hv_vp_assist_page = NULL;
+	else
+		hv_vp_assist_page = kzalloc_objs(*hv_vp_assist_page, nr_cpu_ids);
+	if (!hv_vp_assist_page) {
+		ms_hyperv.hints &= ~HV_X64_ENLIGHTENED_VMCS_RECOMMENDED;
+
+		if (!hv_isolation_type_tdx())
+			goto common_free;
+	}
+
+	if (ms_hyperv.paravisor_present && hv_isolation_type_snp()) {
+		/* Negotiate GHCB Version. */
+		if (!hv_ghcb_negotiate_protocol())
+			hv_ghcb_terminate(SEV_TERM_SET_GEN,
+					  GHCB_SEV_ES_PROT_UNSUPPORTED);
+
+		hv_ghcb_pg = alloc_percpu(union hv_ghcb *);
+		if (!hv_ghcb_pg)
+			goto free_vp_assist_page;
+	}
+
+	cpuhp = cpuhp_setup_state(CPUHP_AP_HYPERV_ONLINE, "x86/hyperv_init:online",
+				  hv_cpu_init, hv_cpu_die);
+	if (cpuhp < 0)
+		goto free_ghcb_page;
+
+	/*
+	 * Setup the hypercall page and enable hypercalls.
+	 * 1. Register the guest ID
+	 * 2. Enable the hypercall and register the hypercall page
+	 *
+	 * A TDX VM with no paravisor only uses TDX GHCI rather than hv_hypercall_pg:
+	 * when the hypercall input is a page, such a VM must pass a decrypted
+	 * page to Hyper-V, e.g. hv_post_message() uses the per-CPU page
+	 * hyperv_pcpu_input_arg, which is decrypted if no paravisor is present.
+	 *
+	 * A TDX VM with the paravisor uses hv_hypercall_pg for most hypercalls,
+	 * which are handled by the paravisor and the VM must use an encrypted
+	 * input page: in such a VM, the hyperv_pcpu_input_arg is encrypted and
+	 * used in the hypercalls, e.g. see hv_mark_gpa_visibility() and
+	 * hv_arch_irq_unmask(). Such a VM uses TDX GHCI for two hypercalls:
+	 * 1. HVCALL_SIGNAL_EVENT: see vmbus_set_event() and _hv_do_fast_hypercall8().
+	 * 2. HVCALL_POST_MESSAGE: the input page must be a decrypted page, i.e.
+	 * hv_post_message() in such a VM can't use the encrypted hyperv_pcpu_input_arg;
+	 * instead, hv_post_message() uses the post_msg_page, which is decrypted
+	 * in such a VM and is only used in such a VM.
+	 */
+	guest_id = hv_generate_guest_id(LINUX_VERSION_CODE);
+	wrmsrq(HV_X64_MSR_GUEST_OS_ID, guest_id);
+
+	/* With the paravisor, the VM must also write the ID via GHCB/GHCI */
+	hv_ivm_msr_write(HV_X64_MSR_GUEST_OS_ID, guest_id);
+
+	/* A TDX VM with no paravisor only uses TDX GHCI rather than hv_hypercall_pg */
+	if (hv_isolation_type_tdx() && !ms_hyperv.paravisor_present)
+		goto skip_hypercall_pg_init;
+
+	hv_hypercall_pg = __vmalloc_node_range(PAGE_SIZE, 1, MODULES_VADDR,
+			MODULES_END, GFP_KERNEL, PAGE_KERNEL_ROX,
+			VM_FLUSH_RESET_PERMS, NUMA_NO_NODE,
+			__builtin_return_address(0));
+	if (hv_hypercall_pg == NULL)
+		goto clean_guest_os_id;
+
+	rdmsrq(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);
+	hypercall_msr.enable = 1;
+
+	if (hv_root_partition()) {
+		struct page *pg;
+		void *src;
+
+		/*
+		 * For the root partition, the hypervisor will set up its
+		 * hypercall page. The hypervisor guarantees it will not show
+		 * up in the root's address space. The root can't change the
+		 * location of the hypercall page.
+		 *
+		 * Order is important here. We must enable the hypercall page
+		 * so it is populated with code, then copy the code to an
+		 * executable page.
+		 */
+		wrmsrq(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);
+
+		pg = vmalloc_to_page(hv_hypercall_pg);
+		src = memremap(hypercall_msr.guest_physical_address << PAGE_SHIFT, PAGE_SIZE,
+				MEMREMAP_WB);
+		BUG_ON(!src);
+		memcpy_to_page(pg, 0, src, HV_HYP_PAGE_SIZE);
+		memunmap(src);
+
+		hv_remap_tsc_clocksource();
+		hv_sleep_notifiers_register();
+	} else {
+		hypercall_msr.guest_physical_address = vmalloc_to_pfn(hv_hypercall_pg);
+		wrmsrq(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);
+	}
+
+	hv_set_hypercall_pg(hv_hypercall_pg);
+
+	if (hv_root_partition())        /* after set hypercall pg */
+		hv_root_crash_init();
+
+skip_hypercall_pg_init:
+	/*
+	 * hyperv_init() is called before LAPIC is initialized: see
+	 * apic_intr_mode_init() -> x86_platform.apic_post_init() and
+	 * apic_bsp_setup() -> setup_local_APIC(). The direct-mode STIMER
+	 * depends on LAPIC, so hv_stimer_alloc() should be called from
+	 * x86_init.timers.setup_percpu_clockev.
+	 */
+	old_setup_percpu_clockev = x86_init.timers.setup_percpu_clockev;
+	x86_init.timers.setup_percpu_clockev = hv_stimer_setup_percpu_clockev;
+
+	hv_apic_init();
+
+	x86_init.pci.arch_init = hv_pci_init;
+
+	register_syscore(&hv_syscore);
+
+	if (ms_hyperv.priv_high & HV_ACCESS_PARTITION_ID)
+		hv_get_partition_id();
+
+#ifdef CONFIG_PCI_MSI
+	/*
+	 * If we're running as root, we want to create our own PCI MSI domain.
+	 * We can't set this in hv_pci_init because that would be too late.
+	 */
+	if (hv_root_partition())
+		x86_init.irqs.create_pci_msi_domain = hv_create_pci_msi_domain;
+#endif
+
+	/* Query the VMs extended capability once, so that it can be cached. */
+	hv_query_ext_cap(0);
+
+	/* Find the VTL */
+	ms_hyperv.vtl = get_vtl();
+
+	if (ms_hyperv.vtl > 0) /* non default VTL */
+		hv_vtl_early_init();
+
+	return;
+
+clean_guest_os_id:
+	wrmsrq(HV_X64_MSR_GUEST_OS_ID, 0);
+	hv_ivm_msr_write(HV_X64_MSR_GUEST_OS_ID, 0);
+	cpuhp_remove_state(CPUHP_AP_HYPERV_ONLINE);
+free_ghcb_page:
+	free_percpu(hv_ghcb_pg);
+free_vp_assist_page:
+	kfree(hv_vp_assist_page);
+	hv_vp_assist_page = NULL;
+common_free:
+	hv_common_free();
+}
+
+/*
+ * This routine is called before kexec/kdump, it does the required cleanup.
+ */
+void hyperv_cleanup(void)
+{
+	union hv_x64_msr_hypercall_contents hypercall_msr;
+	union hv_reference_tsc_msr tsc_msr;
+
+	/* Reset our OS id */
+	wrmsrq(HV_X64_MSR_GUEST_OS_ID, 0);
+	hv_ivm_msr_write(HV_X64_MSR_GUEST_OS_ID, 0);
+
+	/*
+	 * Reset hv_hypercall_pg before resetting it in the hypervisor.
+	 * hv_set_hypercall_pg(NULL) is not used because at this point in the
+	 * panic path other CPUs have been stopped, causing static_call_update()
+	 * to hang. So resetting hv_hypercall_pg to cause hypercalls to fail
+	 * cleanly is only operative on 32-bit builds. But this is OK as it is
+	 * just a preventative measure to ease detecting a hypercall being made
+	 * after this point, which shouldn't be happening anyway.
+	 */
+	hv_hypercall_pg = NULL;
+
+	/* Reset the hypercall page */
+	hypercall_msr.as_uint64 = hv_get_msr(HV_X64_MSR_HYPERCALL);
+	hypercall_msr.enable = 0;
+	hv_set_msr(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);
+
+	/* Reset the TSC page */
+	tsc_msr.as_uint64 = hv_get_msr(HV_X64_MSR_REFERENCE_TSC);
+	tsc_msr.enable = 0;
+	hv_set_msr(HV_X64_MSR_REFERENCE_TSC, tsc_msr.as_uint64);
+}
+
+void hyperv_report_panic(struct pt_regs *regs, long err, bool in_die)
+{
+	static bool panic_reported;
+	u64 guest_id;
+
+	if (in_die && !panic_on_oops)
+		return;
+
+	/*
+	 * We prefer to report panic on 'die' chain as we have proper
+	 * registers to report, but if we miss it (e.g. on BUG()) we need
+	 * to report it on 'panic'.
+	 */
+	if (panic_reported)
+		return;
+	panic_reported = true;
+
+	rdmsrq(HV_X64_MSR_GUEST_OS_ID, guest_id);
+
+	wrmsrq(HV_X64_MSR_CRASH_P0, err);
+	wrmsrq(HV_X64_MSR_CRASH_P1, guest_id);
+	wrmsrq(HV_X64_MSR_CRASH_P2, regs->ip);
+	wrmsrq(HV_X64_MSR_CRASH_P3, regs->ax);
+	wrmsrq(HV_X64_MSR_CRASH_P4, regs->sp);
+
+	/*
+	 * Let Hyper-V know there is crash data available
+	 */
+	wrmsrq(HV_X64_MSR_CRASH_CTL, HV_CRASH_CTL_CRASH_NOTIFY);
+}
+EXPORT_SYMBOL_GPL(hyperv_report_panic);
+
+bool hv_is_hyperv_initialized(void)
+{
+	union hv_x64_msr_hypercall_contents hypercall_msr;
+
+	/*
+	 * Ensure that we're really on Hyper-V, and not a KVM or Xen
+	 * emulation of Hyper-V
+	 */
+	if (x86_hyper_type != X86_HYPER_MS_HYPERV)
+		return false;
+
+	/* A TDX VM with no paravisor uses TDX GHCI call rather than hv_hypercall_pg */
+	if (hv_isolation_type_tdx() && !ms_hyperv.paravisor_present)
+		return true;
+	/*
+	 * Verify that earlier initialization succeeded by checking
+	 * that the hypercall page is setup
+	 */
+	hypercall_msr.as_uint64 = 0;
+	rdmsrq(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);
+
+	return hypercall_msr.enable;
+}
+EXPORT_SYMBOL_GPL(hv_is_hyperv_initialized);
+
+int hv_apicid_to_vp_index(u32 apic_id)
+{
+	u64 control;
+	u64 status;
+	unsigned long irq_flags;
+	struct hv_get_vp_from_apic_id_in *input;
+	u32 *output, ret;
+
+	local_irq_save(irq_flags);
+
+	input = *this_cpu_ptr(hyperv_pcpu_input_arg);
+	memset(input, 0, sizeof(*input));
+	input->partition_id = HV_PARTITION_ID_SELF;
+	input->apic_ids[0] = apic_id;
+
+	output = *this_cpu_ptr(hyperv_pcpu_output_arg);
+
+	control = HV_HYPERCALL_REP_COMP_1 | HVCALL_GET_VP_INDEX_FROM_APIC_ID;
+	status = hv_do_hypercall(control, input, output);
+	ret = output[0];
+
+	local_irq_restore(irq_flags);
+
+	if (!hv_result_success(status)) {
+		pr_err("failed to get vp index from apic id %d, status %#llx\n",
+		       apic_id, status);
+		return -EINVAL;
+	}
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(hv_apicid_to_vp_index);

linux-7.2.3/arch/x86/hyperv/hv_spinlock.c

@@ -0,0 +1,93 @@
+// SPDX-License-Identifier: GPL-2.0
+
+/*
+ * Hyper-V specific spinlock code.
+ *
+ * Copyright (C) 2018, Intel, Inc.
+ *
+ * Author : Yi Sun <yi.y.sun@intel.com>
+ */
+
+#define pr_fmt(fmt) "Hyper-V: " fmt
+
+#include <linux/spinlock.h>
+
+#include <asm/mshyperv.h>
+#include <asm/apic.h>
+#include <asm/msr.h>
+
+static bool hv_pvspin __initdata = true;
+
+static void hv_qlock_kick(int cpu)
+{
+	__apic_send_IPI(cpu, X86_PLATFORM_IPI_VECTOR);
+}
+
+static void hv_qlock_wait(u8 *byte, u8 val)
+{
+	unsigned long flags;
+
+	if (in_nmi())
+		return;
+
+	/*
+	 * Reading HV_X64_MSR_GUEST_IDLE MSR tells the hypervisor that the
+	 * vCPU can be put into 'idle' state. This 'idle' state is
+	 * terminated by an IPI, usually from hv_qlock_kick(), even if
+	 * interrupts are disabled on the vCPU.
+	 *
+	 * To prevent a race against the unlock path it is required to
+	 * disable interrupts before accessing the HV_X64_MSR_GUEST_IDLE
+	 * MSR. Otherwise, if the IPI from hv_qlock_kick() arrives between
+	 * the lock value check and the rdmsrq() then the vCPU might be put
+	 * into 'idle' state by the hypervisor and kept in that state for
+	 * an unspecified amount of time.
+	 */
+	local_irq_save(flags);
+	/*
+	 * Only issue the rdmsrq() when the lock state has not changed.
+	 */
+	if (READ_ONCE(*byte) == val) {
+		unsigned long msr_val;
+
+		rdmsrq(HV_X64_MSR_GUEST_IDLE, msr_val);
+
+		(void)msr_val;
+	}
+	local_irq_restore(flags);
+}
+
+/*
+ * Hyper-V does not support this so far.
+ */
+__visible bool hv_vcpu_is_preempted(int vcpu)
+{
+	return false;
+}
+
+PV_CALLEE_SAVE_REGS_THUNK(hv_vcpu_is_preempted);
+
+void __init hv_init_spinlocks(void)
+{
+	if (!hv_pvspin || !apic ||
+	    !(ms_hyperv.hints & HV_X64_CLUSTER_IPI_RECOMMENDED) ||
+	    !(ms_hyperv.features & HV_MSR_GUEST_IDLE_AVAILABLE)) {
+		pr_info("PV spinlocks disabled\n");
+		return;
+	}
+	pr_info("PV spinlocks enabled\n");
+
+	__pv_init_lock_hash();
+	pv_ops_lock.queued_spin_lock_slowpath = __pv_queued_spin_lock_slowpath;
+	pv_ops_lock.queued_spin_unlock = PV_CALLEE_SAVE(__pv_queued_spin_unlock);
+	pv_ops_lock.wait = hv_qlock_wait;
+	pv_ops_lock.kick = hv_qlock_kick;
+	pv_ops_lock.vcpu_is_preempted = PV_CALLEE_SAVE(hv_vcpu_is_preempted);
+}
+
+static __init int hv_parse_nopvspin(char *arg)
+{
+	hv_pvspin = false;
+	return 0;
+}
+early_param("hv_nopvspin", hv_parse_nopvspin);

linux-7.2.3/arch/x86/hyperv/hv_trampoline.S

@@ -0,0 +1,101 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * X86 specific Hyper-V kdump/crash related code.
+ *
+ * Copyright (C) 2025, Microsoft, Inc.
+ *
+ */
+#include <linux/linkage.h>
+#include <asm/alternative.h>
+#include <asm/msr.h>
+#include <asm/processor-flags.h>
+#include <asm/nospec-branch.h>
+
+/*
+ * void noreturn hv_crash_asm32(arg1)
+ *    arg1 == edi == 32bit PA of struct hv_crash_tramp_data
+ *
+ * The hypervisor jumps here upon devirtualization in protected mode. This
+ * code gets copied to a page in the low 4G ie, 32bit space so it can run
+ * in the protected mode. Hence we cannot use any compile/link time offsets or
+ * addresses. It restores long mode via temporary gdt and page tables and
+ * eventually jumps to kernel code entry at HV_CRASHDATA_OFFS_C_entry.
+ *
+ * PreCondition (ie, Hypervisor call back ABI):
+ *  o CR0 is set to 0x0021: PE(prot mode) and NE are set, paging is disabled
+ *  o CR4 is set to 0x0
+ *  o IA32_EFER is set to 0x901 (SCE and NXE are set)
+ *  o EDI is set to the Arg passed to HVCALL_DISABLE_HYP_EX.
+ *  o CS, DS, ES, FS, GS are all initialized with a base of 0 and limit 0xFFFF
+ *  o IDTR, TR and GDTR are initialized with a base of 0 and limit of 0xFFFF
+ *  o LDTR is initialized as invalid (limit of 0)
+ *  o MSR PAT is power on default.
+ *  o Other state/registers are cleared. All TLBs flushed.
+ */
+
+#define HV_CRASHDATA_OFFS_TRAMPCR3    0x0    /*  0 */
+#define HV_CRASHDATA_OFFS_KERNCR3     0x8    /*  8 */
+#define HV_CRASHDATA_OFFS_GDTRLIMIT  0x12    /* 18 */
+#define HV_CRASHDATA_OFFS_CS_JMPTGT  0x28    /* 40 */
+#define HV_CRASHDATA_OFFS_C_entry    0x30    /* 48 */
+
+	.text
+	.code32
+
+SYM_CODE_START(hv_crash_asm32)
+	UNWIND_HINT_UNDEFINED
+	ENDBR
+	movl	$X86_CR4_PAE, %ecx
+	movl	%ecx, %cr4
+
+	movl %edi, %ebx
+	add $HV_CRASHDATA_OFFS_TRAMPCR3, %ebx
+	movl %cs:(%ebx), %eax
+	movl %eax, %cr3
+
+	/* Setup EFER for long mode now */
+	movl	$MSR_EFER, %ecx
+	rdmsr
+	btsl	$_EFER_LME, %eax
+	wrmsr
+
+	/* Turn paging on using the temp 32bit trampoline page table */
+	movl %cr0, %eax
+	orl $(X86_CR0_PG), %eax
+	movl %eax, %cr0
+
+	/* since kernel cr3 could be above 4G, we need to be in the long mode
+	 * before we can load 64bits of the kernel cr3. We use a temp gdt for
+	 * that with CS.L=1 and CS.D=0 */
+	mov %edi, %eax
+	add $HV_CRASHDATA_OFFS_GDTRLIMIT, %eax
+	lgdtl %cs:(%eax)
+
+	/* not done yet, restore CS now to switch to CS.L=1 */
+	mov %edi, %eax
+	add $HV_CRASHDATA_OFFS_CS_JMPTGT, %eax
+	ljmp %cs:*(%eax)
+SYM_CODE_END(hv_crash_asm32)
+
+	/* we now run in full 64bit IA32-e long mode, CS.L=1 and CS.D=0 */
+	.code64
+	.balign 8
+SYM_CODE_START(hv_crash_asm64)
+	UNWIND_HINT_UNDEFINED
+	ENDBR
+	/* restore kernel page tables so we can jump to kernel code */
+	mov %edi, %eax
+	add $HV_CRASHDATA_OFFS_KERNCR3, %eax
+	movq %cs:(%eax), %rbx
+	movq %rbx, %cr3
+
+	mov %edi, %eax
+	add $HV_CRASHDATA_OFFS_C_entry, %eax
+	movq %cs:(%eax), %rbx
+	ANNOTATE_RETPOLINE_SAFE
+	jmp *%rbx
+
+	int $3
+
+SYM_INNER_LABEL(hv_crash_asm_end, SYM_L_GLOBAL)
+SYM_CODE_END(hv_crash_asm64)

linux-7.2.3/arch/x86/hyperv/hv_vtl.c

@@ -0,0 +1,313 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2023, Microsoft Corporation.
+ *
+ * Author:
+ *   Saurabh Sengar <ssengar@microsoft.com>
+ */
+
+#include <linux/acpi.h>
+
+#include <asm/acpi.h>
+#include <asm/apic.h>
+#include <asm/boot.h>
+#include <asm/desc.h>
+#include <asm/fpu/api.h>
+#include <asm/fpu/types.h>
+#include <asm/i8259.h>
+#include <asm/mshyperv.h>
+#include <asm/msr.h>
+#include <asm/realmode.h>
+#include <asm/reboot.h>
+#include <asm/smap.h>
+#include <linux/export.h>
+#include <../kernel/smpboot.h>
+#include "../../kernel/fpu/legacy.h"
+
+extern struct boot_params boot_params;
+static struct real_mode_header hv_vtl_real_mode_header;
+
+static bool __init hv_vtl_msi_ext_dest_id(void)
+{
+	return true;
+}
+
+/*
+ * The `native_machine_emergency_restart` function from `reboot.c` writes
+ * to the physical address 0x472 to indicate the type of reboot for the
+ * firmware. We cannot have that in VSM as the memory composition might
+ * be more generic, and such write effectively corrupts the memory thus
+ * making diagnostics harder at the very least.
+ */
+static void  __noreturn hv_vtl_emergency_restart(void)
+{
+	/*
+	 * Cause a triple fault and the immediate reset. Here the code does not run
+	 * on the top of any firmware, whereby cannot reach out to its services.
+	 * The inifinite loop is for the improbable case that the triple fault does
+	 * not work and have to preserve the state intact for debugging.
+	 */
+	for (;;) {
+		idt_invalidate();
+		__asm__ __volatile__("int3");
+	}
+}
+
+/*
+ * The only way to restart in the VTL mode is to triple fault as the kernel runs
+ * as firmware.
+ */
+static void  __noreturn hv_vtl_restart(char __maybe_unused *cmd)
+{
+	hv_vtl_emergency_restart();
+}
+
+static inline bool within_page(u64 addr, u64 start)
+{
+	return addr >= start && addr < (start + PAGE_SIZE);
+}
+
+static bool hv_vtl_is_private_mmio_tdx(u64 addr)
+{
+	u64 mb_addr = acpi_get_mp_wakeup_mailbox_paddr();
+
+	return mb_addr && within_page(addr, mb_addr);
+}
+
+void __init hv_vtl_init_platform(void)
+{
+	/*
+	 * This function is a no-op if the VTL mode is not enabled.
+	 * If it is, this function runs if and only the kernel boots in
+	 * VTL2 which the x86 hv initialization path makes sure of.
+	 */
+	pr_info("Linux runs in Hyper-V Virtual Trust Level %d\n", ms_hyperv.vtl);
+
+	/* There is no paravisor present if we are here. */
+	if (hv_isolation_type_tdx()) {
+		x86_init.resources.realmode_limit = SZ_4G;
+		x86_platform.hyper.is_private_mmio = hv_vtl_is_private_mmio_tdx;
+
+	} else {
+		x86_platform.realmode_reserve = x86_init_noop;
+		x86_platform.realmode_init = x86_init_noop;
+		real_mode_header = &hv_vtl_real_mode_header;
+	}
+	x86_init.irqs.pre_vector_init = x86_init_noop;
+	x86_init.timers.timer_init = x86_init_noop;
+	x86_init.resources.probe_roms = x86_init_noop;
+
+	/* Avoid searching for BIOS MP tables */
+	x86_init.mpparse.find_mptable = x86_init_noop;
+	x86_init.mpparse.early_parse_smp_cfg = x86_init_noop;
+
+	x86_platform.get_wallclock = get_rtc_noop;
+	x86_platform.set_wallclock = set_rtc_noop;
+	x86_platform.get_nmi_reason = hv_get_nmi_reason;
+
+	x86_platform.legacy.i8042 = X86_LEGACY_I8042_PLATFORM_ABSENT;
+	x86_platform.legacy.rtc = 0;
+	x86_platform.legacy.warm_reset = 0;
+	x86_platform.legacy.reserve_bios_regions = 0;
+	x86_platform.legacy.devices.pnpbios = 0;
+
+	x86_init.hyper.msi_ext_dest_id = hv_vtl_msi_ext_dest_id;
+}
+
+static inline u64 hv_vtl_system_desc_base(struct ldttss_desc *desc)
+{
+	return ((u64)desc->base3 << 32) | ((u64)desc->base2 << 24) |
+		(desc->base1 << 16) | desc->base0;
+}
+
+static inline u32 hv_vtl_system_desc_limit(struct ldttss_desc *desc)
+{
+	return ((u32)desc->limit1 << 16) | (u32)desc->limit0;
+}
+
+typedef void (*secondary_startup_64_fn)(void*, void*);
+static void hv_vtl_ap_entry(void)
+{
+	((secondary_startup_64_fn)secondary_startup_64)(&boot_params, &boot_params);
+}
+
+static int hv_vtl_bringup_vcpu(u32 target_vp_index, int cpu, u64 eip_ignored)
+{
+	u64 status, rsp, rip;
+	int ret = 0;
+	struct hv_enable_vp_vtl *input;
+	unsigned long irq_flags;
+
+	struct desc_ptr gdt_ptr;
+	struct desc_ptr idt_ptr;
+
+	struct ldttss_desc *tss;
+	struct ldttss_desc *ldt;
+	struct desc_struct *gdt;
+
+	struct task_struct *idle = idle_thread_get(cpu);
+	if (IS_ERR(idle))
+		return PTR_ERR(idle);
+
+	rsp = (unsigned long)idle->thread.sp;
+	rip = (u64)&hv_vtl_ap_entry;
+
+	native_store_gdt(&gdt_ptr);
+	store_idt(&idt_ptr);
+
+	gdt = (struct desc_struct *)((void *)(gdt_ptr.address));
+	tss = (struct ldttss_desc *)(gdt + GDT_ENTRY_TSS);
+	ldt = (struct ldttss_desc *)(gdt + GDT_ENTRY_LDT);
+
+	local_irq_save(irq_flags);
+
+	input = *this_cpu_ptr(hyperv_pcpu_input_arg);
+	memset(input, 0, sizeof(*input));
+
+	input->partition_id = HV_PARTITION_ID_SELF;
+	input->vp_index = target_vp_index;
+	input->target_vtl.target_vtl = HV_VTL_MGMT;
+
+	/*
+	 * The x86_64 Linux kernel follows the 16-bit -> 32-bit -> 64-bit
+	 * mode transition sequence after waking up an AP with SIPI whose
+	 * vector points to the 16-bit AP startup trampoline code. Here in
+	 * VTL2, we can't perform that sequence as the AP has to start in
+	 * the 64-bit mode.
+	 *
+	 * To make this happen, we tell the hypervisor to load a valid 64-bit
+	 * context (most of which is just magic numbers from the CPU manual)
+	 * so that AP jumps right to the 64-bit entry of the kernel, and the
+	 * control registers are loaded with values that let the AP fetch the
+	 * code and data and carry on with work it gets assigned.
+	 */
+
+	input->vp_context.rip = rip;
+	input->vp_context.rsp = rsp;
+	input->vp_context.rflags = 0x0000000000000002;
+	input->vp_context.efer = native_rdmsrq(MSR_EFER);
+	input->vp_context.cr0 = native_read_cr0();
+	input->vp_context.cr3 = __native_read_cr3();
+	input->vp_context.cr4 = native_read_cr4();
+	input->vp_context.msr_cr_pat = native_rdmsrq(MSR_IA32_CR_PAT);
+	input->vp_context.idtr.limit = idt_ptr.size;
+	input->vp_context.idtr.base = idt_ptr.address;
+	input->vp_context.gdtr.limit = gdt_ptr.size;
+	input->vp_context.gdtr.base = gdt_ptr.address;
+
+	/* Non-system desc (64bit), long, code, present */
+	input->vp_context.cs.selector = __KERNEL_CS;
+	input->vp_context.cs.base = 0;
+	input->vp_context.cs.limit = 0xffffffff;
+	input->vp_context.cs.attributes = 0xa09b;
+	/* Non-system desc (64bit), data, present, granularity, default */
+	input->vp_context.ss.selector = __KERNEL_DS;
+	input->vp_context.ss.base = 0;
+	input->vp_context.ss.limit = 0xffffffff;
+	input->vp_context.ss.attributes = 0xc093;
+
+	/* System desc (128bit), present, LDT */
+	input->vp_context.ldtr.selector = GDT_ENTRY_LDT * 8;
+	input->vp_context.ldtr.base = hv_vtl_system_desc_base(ldt);
+	input->vp_context.ldtr.limit = hv_vtl_system_desc_limit(ldt);
+	input->vp_context.ldtr.attributes = 0x82;
+
+	/* System desc (128bit), present, TSS, 0x8b - busy, 0x89 -- default */
+	input->vp_context.tr.selector = GDT_ENTRY_TSS * 8;
+	input->vp_context.tr.base = hv_vtl_system_desc_base(tss);
+	input->vp_context.tr.limit = hv_vtl_system_desc_limit(tss);
+	input->vp_context.tr.attributes = 0x8b;
+
+	status = hv_do_hypercall(HVCALL_ENABLE_VP_VTL, input, NULL);
+
+	if (!hv_result_success(status) &&
+	    hv_result(status) != HV_STATUS_VTL_ALREADY_ENABLED) {
+		pr_err("HVCALL_ENABLE_VP_VTL failed for VP : %d ! [Err: %#llx\n]",
+		       target_vp_index, status);
+		ret = -EINVAL;
+		goto free_lock;
+	}
+
+	status = hv_do_hypercall(HVCALL_START_VP, input, NULL);
+
+	if (!hv_result_success(status)) {
+		pr_err("HVCALL_START_VP failed for VP : %d ! [Err: %#llx]\n",
+		       target_vp_index, status);
+		ret = -EINVAL;
+	}
+
+free_lock:
+	local_irq_restore(irq_flags);
+
+	return ret;
+}
+
+static int hv_vtl_wakeup_secondary_cpu(u32 apicid, unsigned long start_eip, unsigned int cpu)
+{
+	int vp_index;
+
+	pr_debug("Bringing up CPU with APIC ID %d in VTL2...\n", apicid);
+	vp_index = hv_apicid_to_vp_index(apicid);
+
+	if (vp_index < 0) {
+		pr_err("Couldn't find CPU with APIC ID %d\n", apicid);
+		return -EINVAL;
+	}
+	if (vp_index > ms_hyperv.max_vp_index) {
+		pr_err("Invalid CPU id %d for APIC ID %d\n", vp_index, apicid);
+		return -EINVAL;
+	}
+
+	return hv_vtl_bringup_vcpu(vp_index, cpu, start_eip);
+}
+
+int __init hv_vtl_early_init(void)
+{
+	machine_ops.emergency_restart = hv_vtl_emergency_restart;
+	machine_ops.restart = hv_vtl_restart;
+
+	/*
+	 * `boot_cpu_has` returns the runtime feature support,
+	 * and here is the earliest it can be used.
+	 */
+	if (cpu_feature_enabled(X86_FEATURE_XSAVE))
+		panic("XSAVE has to be disabled as it is not supported by this module.\n"
+			  "Please add 'noxsave' to the kernel command line.\n");
+
+	/*
+	 * TDX confidential VMs do not trust the hypervisor and cannot use it to
+	 * boot secondary CPUs. Instead, they will be booted using the wakeup
+	 * mailbox if detected during boot. See setup_arch().
+	 *
+	 * There is no paravisor present if we are here.
+	 */
+	if (!hv_isolation_type_tdx())
+		apic_update_callback(wakeup_secondary_cpu_64, hv_vtl_wakeup_secondary_cpu);
+
+	return 0;
+}
+
+DEFINE_STATIC_CALL_NULL(__mshv_vtl_return_hypercall, void (*)(void));
+
+void mshv_vtl_return_call_init(u64 vtl_return_offset)
+{
+	static_call_update(__mshv_vtl_return_hypercall,
+			   (void *)((u8 *)hv_hypercall_pg + vtl_return_offset));
+}
+EXPORT_SYMBOL(mshv_vtl_return_call_init);
+
+void mshv_vtl_return_call(struct mshv_vtl_cpu_context *vtl0)
+{
+	struct hv_vp_assist_page *hvp;
+
+	hvp = hv_vp_assist_page[smp_processor_id()];
+	hvp->vtl_ret_x64rax = vtl0->rax;
+	hvp->vtl_ret_x64rcx = vtl0->rcx;
+
+	kernel_fpu_begin_mask(0);
+	fxrstor(&vtl0->fx_state);
+	__mshv_vtl_return_call(vtl0);
+	fxsave(&vtl0->fx_state);
+	kernel_fpu_end();
+}
+EXPORT_SYMBOL(mshv_vtl_return_call);

linux-7.2.3/arch/x86/hyperv/irqdomain.c

@@ -0,0 +1,428 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Irqdomain for Linux to run as the root partition on Microsoft Hypervisor.
+ *
+ * Authors:
+ *  Sunil Muthuswamy <sunilmut@microsoft.com>
+ *  Wei Liu <wei.liu@kernel.org>
+ */
+
+#include <linux/pci.h>
+#include <linux/irq.h>
+#include <linux/export.h>
+#include <linux/irqchip/irq-msi-lib.h>
+#include <asm/mshyperv.h>
+
+static int hv_map_interrupt(union hv_device_id hv_devid, bool level,
+		int cpu, int vector, struct hv_interrupt_entry *ret_entry)
+{
+	struct hv_input_map_device_interrupt *input;
+	struct hv_output_map_device_interrupt *output;
+	struct hv_device_interrupt_descriptor *intr_desc;
+	unsigned long flags;
+	u64 status;
+	int nr_bank, var_size;
+
+	local_irq_save(flags);
+
+	input = *this_cpu_ptr(hyperv_pcpu_input_arg);
+	output = *this_cpu_ptr(hyperv_pcpu_output_arg);
+
+	intr_desc = &input->interrupt_descriptor;
+	memset(input, 0, sizeof(*input));
+	input->partition_id = hv_current_partition_id;
+	input->device_id = hv_devid.as_uint64;
+	intr_desc->interrupt_type = HV_X64_INTERRUPT_TYPE_FIXED;
+	intr_desc->vector_count = 1;
+	intr_desc->target.vector = vector;
+
+	if (level)
+		intr_desc->trigger_mode = HV_INTERRUPT_TRIGGER_MODE_LEVEL;
+	else
+		intr_desc->trigger_mode = HV_INTERRUPT_TRIGGER_MODE_EDGE;
+
+	intr_desc->target.vp_set.valid_bank_mask = 0;
+	intr_desc->target.vp_set.format = HV_GENERIC_SET_SPARSE_4K;
+	nr_bank = cpumask_to_vpset(&intr_desc->target.vp_set, cpumask_of(cpu));
+	if (nr_bank < 0) {
+		local_irq_restore(flags);
+		pr_err("%s: unable to generate VP set\n", __func__);
+		return -EINVAL;
+	}
+	intr_desc->target.flags = HV_DEVICE_INTERRUPT_TARGET_PROCESSOR_SET;
+
+	/*
+	 * var-sized hypercall, var-size starts after vp_mask (thus
+	 * vp_set.format does not count, but vp_set.valid_bank_mask
+	 * does).
+	 */
+	var_size = nr_bank + 1;
+
+	status = hv_do_rep_hypercall(HVCALL_MAP_DEVICE_INTERRUPT, 0, var_size,
+			input, output);
+	*ret_entry = output->interrupt_entry;
+
+	local_irq_restore(flags);
+
+	if (!hv_result_success(status))
+		hv_status_err(status, "\n");
+
+	return hv_result_to_errno(status);
+}
+
+static int hv_unmap_interrupt(u64 id, struct hv_interrupt_entry *irq_entry)
+{
+	unsigned long flags;
+	struct hv_input_unmap_device_interrupt *input;
+	u64 status;
+
+	local_irq_save(flags);
+	input = *this_cpu_ptr(hyperv_pcpu_input_arg);
+
+	memset(input, 0, sizeof(*input));
+	input->partition_id = hv_current_partition_id;
+	input->device_id = id;
+	input->interrupt_entry = *irq_entry;
+
+	status = hv_do_hypercall(HVCALL_UNMAP_DEVICE_INTERRUPT, input, NULL);
+	local_irq_restore(flags);
+
+	if (!hv_result_success(status))
+		hv_status_err(status, "\n");
+
+	return hv_result_to_errno(status);
+}
+
+#ifdef CONFIG_PCI_MSI
+struct rid_data {
+	struct pci_dev *bridge;
+	u32 rid;
+};
+
+static int get_rid_cb(struct pci_dev *pdev, u16 alias, void *data)
+{
+	struct rid_data *rd = data;
+	u8 bus = PCI_BUS_NUM(rd->rid);
+
+	if (pdev->bus->number != bus || PCI_BUS_NUM(alias) != bus) {
+		rd->bridge = pdev;
+		rd->rid = alias;
+	}
+
+	return 0;
+}
+
+static union hv_device_id hv_build_devid_type_pci(struct pci_dev *pdev)
+{
+	int pos;
+	union hv_device_id hv_devid;
+	struct rid_data data = {
+		.bridge = NULL,
+		.rid = PCI_DEVID(pdev->bus->number, pdev->devfn)
+	};
+
+	pci_for_each_dma_alias(pdev, get_rid_cb, &data);
+
+	hv_devid.as_uint64 = 0;
+	hv_devid.device_type = HV_DEVICE_TYPE_PCI;
+	hv_devid.pci.segment = pci_domain_nr(pdev->bus);
+
+	hv_devid.pci.bdf.bus = PCI_BUS_NUM(data.rid);
+	hv_devid.pci.bdf.device = PCI_SLOT(data.rid);
+	hv_devid.pci.bdf.function = PCI_FUNC(data.rid);
+	hv_devid.pci.source_shadow = HV_SOURCE_SHADOW_NONE;
+
+	if (data.bridge == NULL)
+		goto out;
+
+	/*
+	 * Microsoft Hypervisor requires a bus range when the bridge is
+	 * running in PCI-X mode.
+	 *
+	 * To distinguish conventional vs PCI-X bridge, we can check
+	 * the bridge's PCI-X Secondary Status Register, Secondary Bus
+	 * Mode and Frequency bits. See PCI Express to PCI/PCI-X Bridge
+	 * Specification Revision 1.0 5.2.2.1.3.
+	 *
+	 * Value zero means it is in conventional mode, otherwise it is
+	 * in PCI-X mode.
+	 */
+
+	pos = pci_find_capability(data.bridge, PCI_CAP_ID_PCIX);
+	if (pos) {
+		u16 status;
+
+		pci_read_config_word(data.bridge, pos + PCI_X_BRIDGE_SSTATUS,
+				     &status);
+
+		if (status & PCI_X_SSTATUS_FREQ) {
+			/* Non-zero, PCI-X mode */
+			u8 sec_bus, sub_bus;
+
+			hv_devid.pci.source_shadow =
+					     HV_SOURCE_SHADOW_BRIDGE_BUS_RANGE;
+
+			pci_read_config_byte(data.bridge, PCI_SECONDARY_BUS,
+					     &sec_bus);
+			hv_devid.pci.shadow_bus_range.secondary_bus = sec_bus;
+			pci_read_config_byte(data.bridge, PCI_SUBORDINATE_BUS,
+					     &sub_bus);
+			hv_devid.pci.shadow_bus_range.subordinate_bus = sub_bus;
+		}
+	}
+
+out:
+	return hv_devid;
+}
+
+/*
+ * hv_map_msi_interrupt() - Map the MSI IRQ in the hypervisor.
+ * @data:      Describes the IRQ
+ * @out_entry: Hypervisor (MSI) interrupt entry (can be NULL)
+ *
+ * Map the IRQ in the hypervisor by issuing a MAP_DEVICE_INTERRUPT hypercall.
+ *
+ * Return: 0 on success, -errno on failure
+ */
+int hv_map_msi_interrupt(struct irq_data *data,
+			 struct hv_interrupt_entry *out_entry)
+{
+	struct irq_cfg *cfg = irqd_cfg(data);
+	struct hv_interrupt_entry dummy;
+	union hv_device_id hv_devid;
+	struct msi_desc *msidesc;
+	struct pci_dev *pdev;
+	int cpu;
+
+	msidesc = irq_data_get_msi_desc(data);
+	pdev = msi_desc_to_pci_dev(msidesc);
+	hv_devid = hv_build_devid_type_pci(pdev);
+	cpu = cpumask_first(irq_data_get_effective_affinity_mask(data));
+
+	return hv_map_interrupt(hv_devid, false, cpu, cfg->vector,
+				out_entry ? out_entry : &dummy);
+}
+EXPORT_SYMBOL_GPL(hv_map_msi_interrupt);
+
+static void entry_to_msi_msg(struct hv_interrupt_entry *entry,
+			     struct msi_msg *msg)
+{
+	/* High address is always 0 */
+	msg->address_hi = 0;
+	msg->address_lo = entry->msi_entry.address.as_uint32;
+	msg->data = entry->msi_entry.data.as_uint32;
+}
+
+static int hv_unmap_msi_interrupt(struct pci_dev *pdev,
+				  struct hv_interrupt_entry *irq_entry);
+
+static void hv_irq_compose_msi_msg(struct irq_data *data, struct msi_msg *msg)
+{
+	struct hv_interrupt_entry *stored_entry;
+	struct irq_cfg *cfg = irqd_cfg(data);
+	struct msi_desc *msidesc;
+	struct pci_dev *pdev;
+	int ret;
+
+	msidesc = irq_data_get_msi_desc(data);
+	pdev = msi_desc_to_pci_dev(msidesc);
+
+	if (!cfg) {
+		pr_debug("%s: cfg is NULL", __func__);
+		return;
+	}
+
+	if (data->chip_data) {
+		/*
+		 * This interrupt is already mapped. Let's unmap first.
+		 *
+		 * We don't use retarget interrupt hypercalls here because
+		 * Microsoft Hypervisor doesn't allow root to change the vector
+		 * or specify VPs outside of the set that is initially used
+		 * during mapping.
+		 */
+		stored_entry = data->chip_data;
+		data->chip_data = NULL;
+
+		ret = hv_unmap_msi_interrupt(pdev, stored_entry);
+
+		kfree(stored_entry);
+
+		if (ret)
+			return;
+	}
+
+	stored_entry = kzalloc_obj(*stored_entry, GFP_ATOMIC);
+	if (!stored_entry)
+		return;
+
+	ret = hv_map_msi_interrupt(data, stored_entry);
+	if (ret) {
+		kfree(stored_entry);
+		return;
+	}
+
+	data->chip_data = stored_entry;
+	entry_to_msi_msg(data->chip_data, msg);
+}
+
+static int hv_unmap_msi_interrupt(struct pci_dev *pdev,
+				  struct hv_interrupt_entry *irq_entry)
+{
+	union hv_device_id hv_devid;
+
+	hv_devid = hv_build_devid_type_pci(pdev);
+	return hv_unmap_interrupt(hv_devid.as_uint64, irq_entry);
+}
+
+/* NB: during map, hv_interrupt_entry is saved via data->chip_data */
+static void hv_teardown_msi_irq(struct pci_dev *pdev, struct irq_data *irqd)
+{
+	struct hv_interrupt_entry irq_entry;
+	struct msi_msg msg;
+
+	if (!irqd->chip_data) {
+		pr_debug("%s: no chip data\n!", __func__);
+		return;
+	}
+
+	irq_entry = *(struct hv_interrupt_entry *)irqd->chip_data;
+	entry_to_msi_msg(&irq_entry, &msg);
+
+	kfree(irqd->chip_data);
+	irqd->chip_data = NULL;
+
+	(void)hv_unmap_msi_interrupt(pdev, &irq_entry);
+}
+
+/*
+ * IRQ Chip for MSI PCI/PCI-X/PCI-Express Devices,
+ * which implement the MSI or MSI-X Capability Structure.
+ */
+static struct irq_chip hv_pci_msi_controller = {
+	.name			= "HV-PCI-MSI",
+	.irq_ack		= irq_chip_ack_parent,
+	.irq_compose_msi_msg	= hv_irq_compose_msi_msg,
+	.irq_set_affinity	= irq_chip_set_affinity_parent,
+};
+
+static bool hv_init_dev_msi_info(struct device *dev, struct irq_domain *domain,
+				 struct irq_domain *real_parent,
+				 struct msi_domain_info *info)
+{
+	struct irq_chip *chip = info->chip;
+
+	if (!msi_lib_init_dev_msi_info(dev, domain, real_parent, info))
+		return false;
+
+	chip->flags |= IRQCHIP_SKIP_SET_WAKE | IRQCHIP_MOVE_DEFERRED;
+
+	info->ops->msi_prepare = pci_msi_prepare;
+
+	return true;
+}
+
+#define HV_MSI_FLAGS_SUPPORTED	(MSI_GENERIC_FLAGS_MASK | MSI_FLAG_PCI_MSIX)
+#define HV_MSI_FLAGS_REQUIRED	(MSI_FLAG_USE_DEF_DOM_OPS |	\
+				 MSI_FLAG_USE_DEF_CHIP_OPS)
+
+static struct msi_parent_ops hv_msi_parent_ops = {
+	.supported_flags	= HV_MSI_FLAGS_SUPPORTED,
+	.required_flags		= HV_MSI_FLAGS_REQUIRED,
+	.bus_select_token	= DOMAIN_BUS_NEXUS,
+	.bus_select_mask	= MATCH_PCI_MSI,
+	.chip_flags		= MSI_CHIP_FLAG_SET_ACK,
+	.prefix			= "HV-",
+	.init_dev_msi_info	= hv_init_dev_msi_info,
+};
+
+/* Allocate nr_irqs IRQs for the given irq domain */
+static int hv_msi_domain_alloc(struct irq_domain *d, unsigned int virq,
+			       unsigned int nr_irqs, void *arg)
+{
+	/*
+	 * TODO: The allocation bits of hv_irq_compose_msi_msg(), i.e.
+	 *	 everything except entry_to_msi_msg() should be in here.
+	 */
+	int ret;
+
+	ret = irq_domain_alloc_irqs_parent(d, virq, nr_irqs, arg);
+	if (ret)
+		return ret;
+
+	for (int i = 0; i < nr_irqs; ++i) {
+		irq_domain_set_info(d, virq + i, 0, &hv_pci_msi_controller,
+				    NULL, handle_edge_irq, NULL, "edge");
+	}
+
+	return 0;
+}
+
+static void hv_msi_domain_free(struct irq_domain *d, unsigned int virq,
+			       unsigned int nr_irqs)
+{
+	for (int i = 0; i < nr_irqs; ++i) {
+		struct irq_data *irqd = irq_domain_get_irq_data(d, virq);
+		struct msi_desc *desc;
+
+		desc = irq_data_get_msi_desc(irqd);
+		if (!desc || !desc->irq || WARN_ON_ONCE(!dev_is_pci(desc->dev)))
+			continue;
+
+		hv_teardown_msi_irq(to_pci_dev(desc->dev), irqd);
+	}
+
+	irq_domain_free_irqs_top(d, virq, nr_irqs);
+}
+
+static const struct irq_domain_ops hv_msi_domain_ops = {
+	.select	= msi_lib_irq_domain_select,
+	.alloc	= hv_msi_domain_alloc,
+	.free	= hv_msi_domain_free,
+};
+
+struct irq_domain * __init hv_create_pci_msi_domain(void)
+{
+	struct irq_domain *d = NULL;
+
+	struct irq_domain_info info = {
+		.fwnode		= irq_domain_alloc_named_fwnode("HV-PCI-MSI"),
+		.ops		= &hv_msi_domain_ops,
+		.parent		= x86_vector_domain,
+	};
+
+	if (info.fwnode)
+		d = msi_create_parent_irq_domain(&info, &hv_msi_parent_ops);
+
+	/* No point in going further if we can't get an irq domain */
+	BUG_ON(!d);
+
+	return d;
+}
+
+#endif /* CONFIG_PCI_MSI */
+
+int hv_unmap_ioapic_interrupt(int ioapic_id, struct hv_interrupt_entry *entry)
+{
+	union hv_device_id hv_devid;
+
+	hv_devid.as_uint64 = 0;
+	hv_devid.device_type = HV_DEVICE_TYPE_IOAPIC;
+	hv_devid.ioapic.ioapic_id = (u8)ioapic_id;
+
+	return hv_unmap_interrupt(hv_devid.as_uint64, entry);
+}
+EXPORT_SYMBOL_GPL(hv_unmap_ioapic_interrupt);
+
+int hv_map_ioapic_interrupt(int ioapic_id, bool level, int cpu, int vector,
+		struct hv_interrupt_entry *entry)
+{
+	union hv_device_id hv_devid;
+
+	hv_devid.as_uint64 = 0;
+	hv_devid.device_type = HV_DEVICE_TYPE_IOAPIC;
+	hv_devid.ioapic.ioapic_id = (u8)ioapic_id;
+
+	return hv_map_interrupt(hv_devid, level, cpu, vector, entry);
+}
+EXPORT_SYMBOL_GPL(hv_map_ioapic_interrupt);

linux-7.2.3/arch/x86/hyperv/ivm.c

@@ -0,0 +1,946 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Hyper-V Isolation VM interface with paravisor and hypervisor
+ *
+ * Author:
+ *  Tianyu Lan <Tianyu.Lan@microsoft.com>
+ */
+
+#include <linux/bitfield.h>
+#include <linux/types.h>
+#include <linux/slab.h>
+#include <linux/cpu.h>
+#include <linux/export.h>
+#include <asm/svm.h>
+#include <asm/sev.h>
+#include <asm/io.h>
+#include <asm/coco.h>
+#include <asm/mem_encrypt.h>
+#include <asm/set_memory.h>
+#include <asm/mshyperv.h>
+#include <asm/hypervisor.h>
+#include <asm/mtrr.h>
+#include <asm/io_apic.h>
+#include <asm/realmode.h>
+#include <asm/e820/api.h>
+#include <asm/desc.h>
+#include <asm/msr.h>
+#include <asm/segment.h>
+#include <uapi/asm/vmx.h>
+
+#ifdef CONFIG_AMD_MEM_ENCRYPT
+
+#define GHCB_USAGE_HYPERV_CALL	1
+
+union hv_ghcb {
+	struct ghcb ghcb;
+	struct {
+		u64 hypercalldata[509];
+		u64 outputgpa;
+		union {
+			union {
+				struct {
+					u32 callcode        : 16;
+					u32 isfast          : 1;
+					u32 reserved1       : 14;
+					u32 isnested        : 1;
+					u32 countofelements : 12;
+					u32 reserved2       : 4;
+					u32 repstartindex   : 12;
+					u32 reserved3       : 4;
+				};
+				u64 asuint64;
+			} hypercallinput;
+			union {
+				struct {
+					u16 callstatus;
+					u16 reserved1;
+					u32 elementsprocessed : 12;
+					u32 reserved2         : 20;
+				};
+				u64 asunit64;
+			} hypercalloutput;
+		};
+		u64 reserved2;
+	} hypercall;
+} __packed __aligned(HV_HYP_PAGE_SIZE);
+
+/* Only used in an SNP VM with the paravisor */
+static u16 hv_ghcb_version __ro_after_init;
+
+/* Functions only used in an SNP VM with the paravisor go here. */
+u64 hv_ghcb_hypercall(u64 control, void *input, void *output, u32 input_size)
+{
+	union hv_ghcb *hv_ghcb;
+	void **ghcb_base;
+	unsigned long flags;
+	u64 status;
+
+	if (!hv_ghcb_pg)
+		return -EFAULT;
+
+	WARN_ON(in_nmi());
+
+	local_irq_save(flags);
+	ghcb_base = (void **)this_cpu_ptr(hv_ghcb_pg);
+	hv_ghcb = (union hv_ghcb *)*ghcb_base;
+	if (!hv_ghcb) {
+		local_irq_restore(flags);
+		return -EFAULT;
+	}
+
+	hv_ghcb->ghcb.protocol_version = GHCB_PROTOCOL_MAX;
+	hv_ghcb->ghcb.ghcb_usage = GHCB_USAGE_HYPERV_CALL;
+
+	hv_ghcb->hypercall.outputgpa = (u64)output;
+	hv_ghcb->hypercall.hypercallinput.asuint64 = 0;
+	hv_ghcb->hypercall.hypercallinput.callcode = control;
+
+	if (input_size)
+		memcpy(hv_ghcb->hypercall.hypercalldata, input, input_size);
+
+	VMGEXIT();
+
+	hv_ghcb->ghcb.ghcb_usage = 0xffffffff;
+	memset(hv_ghcb->ghcb.save.valid_bitmap, 0,
+	       sizeof(hv_ghcb->ghcb.save.valid_bitmap));
+
+	status = hv_ghcb->hypercall.hypercalloutput.callstatus;
+
+	local_irq_restore(flags);
+
+	return status;
+}
+
+static inline u64 rd_ghcb_msr(void)
+{
+	return native_rdmsrq(MSR_AMD64_SEV_ES_GHCB);
+}
+
+static inline void wr_ghcb_msr(u64 val)
+{
+	native_wrmsrq(MSR_AMD64_SEV_ES_GHCB, val);
+}
+
+static enum es_result hv_ghcb_hv_call(struct ghcb *ghcb, u64 exit_code,
+				   u64 exit_info_1, u64 exit_info_2)
+{
+	/* Fill in protocol and format specifiers */
+	ghcb->protocol_version = hv_ghcb_version;
+	ghcb->ghcb_usage       = GHCB_DEFAULT_USAGE;
+
+	ghcb_set_sw_exit_code(ghcb, exit_code);
+	ghcb_set_sw_exit_info_1(ghcb, exit_info_1);
+	ghcb_set_sw_exit_info_2(ghcb, exit_info_2);
+
+	VMGEXIT();
+
+	if (ghcb->save.sw_exit_info_1 & GENMASK_ULL(31, 0))
+		return ES_VMM_ERROR;
+	else
+		return ES_OK;
+}
+
+void __noreturn hv_ghcb_terminate(unsigned int set, unsigned int reason)
+{
+	u64 val = GHCB_MSR_TERM_REQ;
+
+	/* Tell the hypervisor what went wrong. */
+	val |= GHCB_SEV_TERM_REASON(set, reason);
+
+	/* Request Guest Termination from Hypervisor */
+	wr_ghcb_msr(val);
+	VMGEXIT();
+
+	while (true)
+		asm volatile("hlt\n" : : : "memory");
+}
+
+bool hv_ghcb_negotiate_protocol(void)
+{
+	u64 ghcb_gpa;
+	u64 val;
+
+	/* Save ghcb page gpa. */
+	ghcb_gpa = rd_ghcb_msr();
+
+	/* Do the GHCB protocol version negotiation */
+	wr_ghcb_msr(GHCB_MSR_SEV_INFO_REQ);
+	VMGEXIT();
+	val = rd_ghcb_msr();
+
+	if (GHCB_MSR_INFO(val) != GHCB_MSR_SEV_INFO_RESP)
+		return false;
+
+	if (GHCB_MSR_PROTO_MAX(val) < GHCB_PROTOCOL_MIN ||
+	    GHCB_MSR_PROTO_MIN(val) > GHCB_PROTOCOL_MAX)
+		return false;
+
+	hv_ghcb_version = min_t(size_t, GHCB_MSR_PROTO_MAX(val),
+			     GHCB_PROTOCOL_MAX);
+
+	/* Write ghcb page back after negotiating protocol. */
+	wr_ghcb_msr(ghcb_gpa);
+	VMGEXIT();
+
+	return true;
+}
+
+static void hv_ghcb_msr_write(u64 msr, u64 value)
+{
+	union hv_ghcb *hv_ghcb;
+	void **ghcb_base;
+	unsigned long flags;
+
+	if (!hv_ghcb_pg)
+		return;
+
+	WARN_ON(in_nmi());
+
+	local_irq_save(flags);
+	ghcb_base = (void **)this_cpu_ptr(hv_ghcb_pg);
+	hv_ghcb = (union hv_ghcb *)*ghcb_base;
+	if (!hv_ghcb) {
+		local_irq_restore(flags);
+		return;
+	}
+
+	ghcb_set_rcx(&hv_ghcb->ghcb, msr);
+	ghcb_set_rax(&hv_ghcb->ghcb, lower_32_bits(value));
+	ghcb_set_rdx(&hv_ghcb->ghcb, upper_32_bits(value));
+
+	if (hv_ghcb_hv_call(&hv_ghcb->ghcb, SVM_EXIT_MSR, 1, 0))
+		pr_warn("Fail to write msr via ghcb %llx.\n", msr);
+
+	local_irq_restore(flags);
+}
+
+static void hv_ghcb_msr_read(u64 msr, u64 *value)
+{
+	union hv_ghcb *hv_ghcb;
+	void **ghcb_base;
+	unsigned long flags;
+
+	/* Check size of union hv_ghcb here. */
+	BUILD_BUG_ON(sizeof(union hv_ghcb) != HV_HYP_PAGE_SIZE);
+
+	if (!hv_ghcb_pg)
+		return;
+
+	WARN_ON(in_nmi());
+
+	local_irq_save(flags);
+	ghcb_base = (void **)this_cpu_ptr(hv_ghcb_pg);
+	hv_ghcb = (union hv_ghcb *)*ghcb_base;
+	if (!hv_ghcb) {
+		local_irq_restore(flags);
+		return;
+	}
+
+	ghcb_set_rcx(&hv_ghcb->ghcb, msr);
+	if (hv_ghcb_hv_call(&hv_ghcb->ghcb, SVM_EXIT_MSR, 0, 0))
+		pr_warn("Fail to read msr via ghcb %llx.\n", msr);
+	else
+		*value = (u64)lower_32_bits(hv_ghcb->ghcb.save.rax)
+			| ((u64)lower_32_bits(hv_ghcb->ghcb.save.rdx) << 32);
+	local_irq_restore(flags);
+}
+
+/* Only used in a fully enlightened SNP VM, i.e. without the paravisor */
+static u8 ap_start_input_arg[PAGE_SIZE] __bss_decrypted __aligned(PAGE_SIZE);
+static u8 ap_start_stack[PAGE_SIZE] __aligned(PAGE_SIZE);
+static DEFINE_PER_CPU(struct sev_es_save_area *, hv_sev_vmsa);
+
+/* Functions only used in an SNP VM without the paravisor go here. */
+
+#define hv_populate_vmcb_seg(seg, gdtr_base)			\
+do {								\
+	if (seg.selector) {					\
+		seg.base = 0;					\
+		seg.limit = HV_AP_SEGMENT_LIMIT;		\
+		seg.attrib = *(u16 *)(gdtr_base + seg.selector + 5);	\
+		seg.attrib = (seg.attrib & 0xFF) | ((seg.attrib >> 4) & 0xF00); \
+	}							\
+} while (0)							\
+
+static int snp_set_vmsa(void *va, bool vmsa)
+{
+	u64 attrs;
+
+	/*
+	 * Running at VMPL0 allows the kernel to change the VMSA bit for a page
+	 * using the RMPADJUST instruction. However, for the instruction to
+	 * succeed it must target the permissions of a lesser privileged
+	 * (higher numbered) VMPL level, so use VMPL1 (refer to the RMPADJUST
+	 * instruction in the AMD64 APM Volume 3).
+	 */
+	attrs = 1;
+	if (vmsa)
+		attrs |= RMPADJUST_VMSA_PAGE_BIT;
+
+	return rmpadjust((unsigned long)va, RMP_PG_SIZE_4K, attrs);
+}
+
+static void snp_cleanup_vmsa(struct sev_es_save_area *vmsa)
+{
+	int err;
+
+	err = snp_set_vmsa(vmsa, false);
+	if (err)
+		pr_err("clear VMSA page failed (%u), leaking page\n", err);
+	else
+		free_page((unsigned long)vmsa);
+}
+
+int hv_snp_boot_ap(u32 apic_id, unsigned long start_ip, unsigned int cpu)
+{
+	struct sev_es_save_area *vmsa = (struct sev_es_save_area *)
+		__get_free_page(GFP_KERNEL | __GFP_ZERO);
+	struct sev_es_save_area *cur_vmsa;
+	struct desc_ptr gdtr;
+	u64 ret, retry = 5;
+	struct hv_enable_vp_vtl *start_vp_input;
+	unsigned long flags;
+	int vp_index;
+
+	if (!vmsa)
+		return -ENOMEM;
+
+	/* Find the Hyper-V VP index which might be not the same as APIC ID */
+	vp_index = hv_apicid_to_vp_index(apic_id);
+	if (vp_index < 0 || vp_index > ms_hyperv.max_vp_index)
+		return -EINVAL;
+
+	native_store_gdt(&gdtr);
+
+	vmsa->gdtr.base = gdtr.address;
+	vmsa->gdtr.limit = gdtr.size;
+
+	savesegment(es, vmsa->es.selector);
+	hv_populate_vmcb_seg(vmsa->es, vmsa->gdtr.base);
+
+	savesegment(cs, vmsa->cs.selector);
+	hv_populate_vmcb_seg(vmsa->cs, vmsa->gdtr.base);
+
+	savesegment(ss, vmsa->ss.selector);
+	hv_populate_vmcb_seg(vmsa->ss, vmsa->gdtr.base);
+
+	savesegment(ds, vmsa->ds.selector);
+	hv_populate_vmcb_seg(vmsa->ds, vmsa->gdtr.base);
+
+	vmsa->efer = native_read_msr(MSR_EFER);
+
+	vmsa->cr4 = native_read_cr4();
+	vmsa->cr3 = __native_read_cr3();
+	vmsa->cr0 = native_read_cr0();
+
+	vmsa->xcr0 = 1;
+	vmsa->g_pat = HV_AP_INIT_GPAT_DEFAULT;
+	vmsa->rip = (u64)secondary_startup_64_no_verify;
+	vmsa->rsp = (u64)&ap_start_stack[PAGE_SIZE];
+
+	/*
+	 * Set the SNP-specific fields for this VMSA:
+	 *   VMPL level
+	 *   SEV_FEATURES (matches the SEV STATUS MSR right shifted 2 bits)
+	 */
+	vmsa->vmpl = 0;
+	vmsa->sev_features = sev_status >> 2;
+
+	ret = snp_set_vmsa(vmsa, true);
+	if (ret) {
+		pr_err("RMPADJUST(%llx) failed: %llx\n", (u64)vmsa, ret);
+		free_page((u64)vmsa);
+		return ret;
+	}
+
+	local_irq_save(flags);
+	start_vp_input = (struct hv_enable_vp_vtl *)ap_start_input_arg;
+	memset(start_vp_input, 0, sizeof(*start_vp_input));
+	start_vp_input->partition_id = -1;
+	start_vp_input->vp_index = vp_index;
+	start_vp_input->target_vtl.target_vtl = ms_hyperv.vtl;
+	*(u64 *)&start_vp_input->vp_context = __pa(vmsa) | 1;
+
+	do {
+		ret = hv_do_hypercall(HVCALL_START_VP,
+				      start_vp_input, NULL);
+	} while (hv_result(ret) == HV_STATUS_TIME_OUT && retry--);
+
+	local_irq_restore(flags);
+
+	if (!hv_result_success(ret)) {
+		pr_err("HvCallStartVirtualProcessor failed: %llx\n", ret);
+		snp_cleanup_vmsa(vmsa);
+		vmsa = NULL;
+	}
+
+	cur_vmsa = per_cpu(hv_sev_vmsa, cpu);
+	/* Free up any previous VMSA page */
+	if (cur_vmsa)
+		snp_cleanup_vmsa(cur_vmsa);
+
+	/* Record the current VMSA page */
+	per_cpu(hv_sev_vmsa, cpu) = vmsa;
+
+	return ret;
+}
+
+u64 hv_snp_hypercall(u64 control, u64 param1, u64 param2)
+{
+	u64 hv_status;
+
+	register u64 __r8 asm("r8") = param2;
+	asm volatile("vmmcall"
+		     : "=a" (hv_status),
+		       "+c" (control), "+d" (param1), "+r" (__r8)
+		     : : "cc", "memory", "r9", "r10", "r11");
+
+	return hv_status;
+}
+
+#else
+static inline void hv_ghcb_msr_write(u64 msr, u64 value) {}
+static inline void hv_ghcb_msr_read(u64 msr, u64 *value) {}
+u64 hv_snp_hypercall(u64 control, u64 param1, u64 param2) { return U64_MAX; }
+#endif /* CONFIG_AMD_MEM_ENCRYPT */
+
+#ifdef CONFIG_INTEL_TDX_GUEST
+static void hv_tdx_msr_write(u64 msr, u64 val)
+{
+	struct tdx_module_args args = {
+		.r10 = TDX_HYPERCALL_STANDARD,
+		.r11 = EXIT_REASON_MSR_WRITE,
+		.r12 = msr,
+		.r13 = val,
+	};
+
+	u64 ret = __tdx_hypercall(&args);
+
+	WARN_ONCE(ret, "Failed to emulate MSR write: %lld\n", ret);
+}
+
+static void hv_tdx_msr_read(u64 msr, u64 *val)
+{
+	struct tdx_module_args args = {
+		.r10 = TDX_HYPERCALL_STANDARD,
+		.r11 = EXIT_REASON_MSR_READ,
+		.r12 = msr,
+	};
+
+	u64 ret = __tdx_hypercall(&args);
+
+	if (WARN_ONCE(ret, "Failed to emulate MSR read: %lld\n", ret))
+		*val = 0;
+	else
+		*val = args.r11;
+}
+
+u64 hv_tdx_hypercall(u64 control, u64 param1, u64 param2)
+{
+	struct tdx_module_args args = { };
+
+	args.r10 = control;
+	args.rdx = param1;
+	args.r8  = param2;
+
+	(void)__tdx_hypercall(&args);
+
+	return args.r11;
+}
+
+#else
+static inline void hv_tdx_msr_write(u64 msr, u64 value) {}
+static inline void hv_tdx_msr_read(u64 msr, u64 *value) {}
+u64 hv_tdx_hypercall(u64 control, u64 param1, u64 param2) { return U64_MAX; }
+#endif /* CONFIG_INTEL_TDX_GUEST */
+
+#if defined(CONFIG_AMD_MEM_ENCRYPT) || defined(CONFIG_INTEL_TDX_GUEST)
+void hv_ivm_msr_write(u64 msr, u64 value)
+{
+	if (!ms_hyperv.paravisor_present)
+		return;
+
+	if (hv_isolation_type_tdx())
+		hv_tdx_msr_write(msr, value);
+	else if (hv_isolation_type_snp())
+		hv_ghcb_msr_write(msr, value);
+}
+
+void hv_ivm_msr_read(u64 msr, u64 *value)
+{
+	if (!ms_hyperv.paravisor_present)
+		return;
+
+	if (hv_isolation_type_tdx())
+		hv_tdx_msr_read(msr, value);
+	else if (hv_isolation_type_snp())
+		hv_ghcb_msr_read(msr, value);
+}
+
+/*
+ * Keep track of the PFN regions which were shared with the host. The access
+ * must be revoked upon kexec/kdump (see hv_ivm_clear_host_access()).
+ */
+struct hv_enc_pfn_region {
+	struct list_head list;
+	u64 pfn;
+	int count;
+};
+
+static LIST_HEAD(hv_list_enc);
+static DEFINE_RAW_SPINLOCK(hv_list_enc_lock);
+
+static int hv_list_enc_add(const u64 *pfn_list, int count)
+{
+	struct hv_enc_pfn_region *ent;
+	unsigned long flags;
+	u64 pfn;
+	int i;
+
+	for (i = 0; i < count; i++) {
+		pfn = pfn_list[i];
+
+		raw_spin_lock_irqsave(&hv_list_enc_lock, flags);
+		/* Check if the PFN already exists in some region first */
+		list_for_each_entry(ent, &hv_list_enc, list) {
+			if ((ent->pfn <= pfn) && (ent->pfn + ent->count - 1 >= pfn))
+				/* Nothing to do - pfn is already in the list */
+				goto unlock_done;
+		}
+
+		/*
+		 * Check if the PFN is adjacent to an existing region. Growing
+		 * a region can make it adjacent to another one but merging is
+		 * not (yet) implemented for simplicity. A PFN cannot be added
+		 * to two regions to keep the logic in hv_list_enc_remove()
+		 * correct.
+		 */
+		list_for_each_entry(ent, &hv_list_enc, list) {
+			if (ent->pfn + ent->count == pfn) {
+				/* Grow existing region up */
+				ent->count++;
+				goto unlock_done;
+			} else if (pfn + 1 == ent->pfn) {
+				/* Grow existing region down */
+				ent->pfn--;
+				ent->count++;
+				goto unlock_done;
+			}
+		}
+		raw_spin_unlock_irqrestore(&hv_list_enc_lock, flags);
+
+		/* No adjacent region found -- create a new one */
+		ent = kzalloc_obj(struct hv_enc_pfn_region);
+		if (!ent)
+			return -ENOMEM;
+
+		ent->pfn = pfn;
+		ent->count = 1;
+
+		raw_spin_lock_irqsave(&hv_list_enc_lock, flags);
+		list_add(&ent->list, &hv_list_enc);
+
+unlock_done:
+		raw_spin_unlock_irqrestore(&hv_list_enc_lock, flags);
+	}
+
+	return 0;
+}
+
+static int hv_list_enc_remove(const u64 *pfn_list, int count)
+{
+	struct hv_enc_pfn_region *ent, *t;
+	struct hv_enc_pfn_region new_region;
+	unsigned long flags;
+	u64 pfn;
+	int i;
+
+	for (i = 0; i < count; i++) {
+		pfn = pfn_list[i];
+
+		raw_spin_lock_irqsave(&hv_list_enc_lock, flags);
+		list_for_each_entry_safe(ent, t, &hv_list_enc, list) {
+			if (pfn == ent->pfn + ent->count - 1) {
+				/* Removing tail pfn */
+				ent->count--;
+				if (!ent->count) {
+					list_del(&ent->list);
+					kfree(ent);
+				}
+				goto unlock_done;
+			} else if (pfn == ent->pfn) {
+				/* Removing head pfn */
+				ent->count--;
+				ent->pfn++;
+				if (!ent->count) {
+					list_del(&ent->list);
+					kfree(ent);
+				}
+				goto unlock_done;
+			} else if (pfn > ent->pfn && pfn < ent->pfn + ent->count - 1) {
+				/*
+				 * Removing a pfn in the middle. Cut off the tail
+				 * of the existing region and create a template for
+				 * the new one.
+				 */
+				new_region.pfn = pfn + 1;
+				new_region.count = ent->count - (pfn - ent->pfn + 1);
+				ent->count = pfn - ent->pfn;
+				goto unlock_split;
+			}
+
+		}
+unlock_done:
+		raw_spin_unlock_irqrestore(&hv_list_enc_lock, flags);
+		continue;
+
+unlock_split:
+		raw_spin_unlock_irqrestore(&hv_list_enc_lock, flags);
+
+		ent = kzalloc_obj(struct hv_enc_pfn_region);
+		if (!ent)
+			return -ENOMEM;
+
+		ent->pfn = new_region.pfn;
+		ent->count = new_region.count;
+
+		raw_spin_lock_irqsave(&hv_list_enc_lock, flags);
+		list_add(&ent->list, &hv_list_enc);
+		raw_spin_unlock_irqrestore(&hv_list_enc_lock, flags);
+	}
+
+	return 0;
+}
+
+/* Stop new private<->shared conversions */
+static void hv_vtom_kexec_begin(void)
+{
+	if (!IS_ENABLED(CONFIG_KEXEC_CORE))
+		return;
+
+	/*
+	 * Crash kernel reaches here with interrupts disabled: can't wait for
+	 * conversions to finish.
+	 *
+	 * If race happened, just report and proceed.
+	 */
+	if (!set_memory_enc_stop_conversion())
+		pr_warn("Failed to stop shared<->private conversions\n");
+}
+
+static void hv_vtom_kexec_finish(void)
+{
+	struct hv_gpa_range_for_visibility *input;
+	struct hv_enc_pfn_region *ent;
+	unsigned long flags;
+	u64 hv_status;
+	int cur, i;
+
+	local_irq_save(flags);
+	input = *this_cpu_ptr(hyperv_pcpu_input_arg);
+
+	if (unlikely(!input))
+		goto out;
+
+	list_for_each_entry(ent, &hv_list_enc, list) {
+		for (i = 0, cur = 0; i < ent->count; i++) {
+			input->gpa_page_list[cur] = ent->pfn + i;
+			cur++;
+
+			if (cur == HV_MAX_MODIFY_GPA_REP_COUNT || i == ent->count - 1) {
+				input->partition_id = HV_PARTITION_ID_SELF;
+				input->host_visibility = VMBUS_PAGE_NOT_VISIBLE;
+				input->reserved0 = 0;
+				input->reserved1 = 0;
+				hv_status = hv_do_rep_hypercall(
+					HVCALL_MODIFY_SPARSE_GPA_PAGE_HOST_VISIBILITY,
+					cur, 0, input, NULL);
+				WARN_ON_ONCE(!hv_result_success(hv_status));
+				cur = 0;
+			}
+		}
+
+	}
+
+out:
+	local_irq_restore(flags);
+}
+
+/*
+ * hv_mark_gpa_visibility - Set pages visible to host via hvcall.
+ *
+ * In Isolation VM, all guest memory is encrypted from host and guest
+ * needs to set memory visible to host via hvcall before sharing memory
+ * with host.
+ */
+static int hv_mark_gpa_visibility(u16 count, const u64 pfn[],
+			   enum hv_mem_host_visibility visibility)
+{
+	struct hv_gpa_range_for_visibility *input;
+	u64 hv_status;
+	unsigned long flags;
+	int ret;
+
+	/* no-op if partition isolation is not enabled */
+	if (!hv_is_isolation_supported())
+		return 0;
+
+	if (count > HV_MAX_MODIFY_GPA_REP_COUNT) {
+		pr_err("Hyper-V: GPA count:%d exceeds supported:%lu\n", count,
+			HV_MAX_MODIFY_GPA_REP_COUNT);
+		return -EINVAL;
+	}
+
+	if (visibility == VMBUS_PAGE_NOT_VISIBLE)
+		ret = hv_list_enc_remove(pfn, count);
+	else
+		ret = hv_list_enc_add(pfn, count);
+	if (ret)
+		return ret;
+
+	local_irq_save(flags);
+	input = *this_cpu_ptr(hyperv_pcpu_input_arg);
+
+	if (unlikely(!input)) {
+		local_irq_restore(flags);
+		return -EINVAL;
+	}
+
+	input->partition_id = HV_PARTITION_ID_SELF;
+	input->host_visibility = visibility;
+	input->reserved0 = 0;
+	input->reserved1 = 0;
+	memcpy((void *)input->gpa_page_list, pfn, count * sizeof(*pfn));
+	hv_status = hv_do_rep_hypercall(
+			HVCALL_MODIFY_SPARSE_GPA_PAGE_HOST_VISIBILITY, count,
+			0, input, NULL);
+	local_irq_restore(flags);
+
+	if (hv_result_success(hv_status))
+		return 0;
+
+	if (visibility == VMBUS_PAGE_NOT_VISIBLE)
+		ret = hv_list_enc_add(pfn, count);
+	else
+		ret = hv_list_enc_remove(pfn, count);
+	/*
+	 * There's no good way to recover from -ENOMEM here, the accounting is
+	 * wrong either way.
+	 */
+	WARN_ON_ONCE(ret);
+
+	return -EFAULT;
+}
+
+/*
+ * When transitioning memory between encrypted and decrypted, the caller
+ * of set_memory_encrypted() or set_memory_decrypted() is responsible for
+ * ensuring that the memory isn't in use and isn't referenced while the
+ * transition is in progress.  The transition has multiple steps, and the
+ * memory is in an inconsistent state until all steps are complete. A
+ * reference while the state is inconsistent could result in an exception
+ * that can't be cleanly fixed up.
+ *
+ * But the Linux kernel load_unaligned_zeropad() mechanism could cause a
+ * stray reference that can't be prevented by the caller, so Linux has
+ * specific code to handle this case. But when the #VC and #VE exceptions
+ * routed to a paravisor, the specific code doesn't work. To avoid this
+ * problem, mark the pages as "not present" while the transition is in
+ * progress. If load_unaligned_zeropad() causes a stray reference, a normal
+ * page fault is generated instead of #VC or #VE, and the page-fault-based
+ * handlers for load_unaligned_zeropad() resolve the reference.  When the
+ * transition is complete, hv_vtom_set_host_visibility() marks the pages
+ * as "present" again.
+ */
+static int hv_vtom_clear_present(unsigned long kbuffer, int pagecount, bool enc)
+{
+	return set_memory_np(kbuffer, pagecount);
+}
+
+/*
+ * hv_vtom_set_host_visibility - Set specified memory visible to host.
+ *
+ * In Isolation VM, all guest memory is encrypted from host and guest
+ * needs to set memory visible to host via hvcall before sharing memory
+ * with host. This function works as wrap of hv_mark_gpa_visibility()
+ * with memory base and size.
+ */
+static int hv_vtom_set_host_visibility(unsigned long kbuffer, int pagecount, bool enc)
+{
+	enum hv_mem_host_visibility visibility = enc ?
+			VMBUS_PAGE_NOT_VISIBLE : VMBUS_PAGE_VISIBLE_READ_WRITE;
+	u64 *pfn_array;
+	phys_addr_t paddr;
+	int i, pfn, err;
+	void *vaddr;
+	int ret = 0;
+
+	pfn_array = kmalloc(HV_HYP_PAGE_SIZE, GFP_KERNEL);
+	if (!pfn_array) {
+		ret = -ENOMEM;
+		goto err_set_memory_p;
+	}
+
+	for (i = 0, pfn = 0; i < pagecount; i++) {
+		/*
+		 * Use slow_virt_to_phys() because the PRESENT bit has been
+		 * temporarily cleared in the PTEs.  slow_virt_to_phys() works
+		 * without the PRESENT bit while virt_to_hvpfn() or similar
+		 * does not.
+		 */
+		vaddr = (void *)kbuffer + (i * HV_HYP_PAGE_SIZE);
+		paddr = slow_virt_to_phys(vaddr);
+		pfn_array[pfn] = paddr >> HV_HYP_PAGE_SHIFT;
+		pfn++;
+
+		if (pfn == HV_MAX_MODIFY_GPA_REP_COUNT || i == pagecount - 1) {
+			ret = hv_mark_gpa_visibility(pfn, pfn_array,
+						     visibility);
+			if (ret)
+				goto err_free_pfn_array;
+			pfn = 0;
+		}
+	}
+
+err_free_pfn_array:
+	kfree(pfn_array);
+
+err_set_memory_p:
+	/*
+	 * Set the PTE PRESENT bits again to revert what hv_vtom_clear_present()
+	 * did. Do this even if there is an error earlier in this function in
+	 * order to avoid leaving the memory range in a "broken" state. Setting
+	 * the PRESENT bits shouldn't fail, but return an error if it does.
+	 */
+	err = set_memory_p(kbuffer, pagecount);
+	if (err && !ret)
+		ret = err;
+
+	return ret;
+}
+
+static bool hv_vtom_tlb_flush_required(bool private)
+{
+	/*
+	 * Since hv_vtom_clear_present() marks the PTEs as "not present"
+	 * and flushes the TLB, they can't be in the TLB. That makes the
+	 * flush controlled by this function redundant, so return "false".
+	 */
+	return false;
+}
+
+static bool hv_vtom_cache_flush_required(void)
+{
+	return false;
+}
+
+static bool hv_is_private_mmio(u64 addr)
+{
+	/*
+	 * Hyper-V always provides a single IO-APIC in a guest VM.
+	 * When a paravisor is used, it is emulated by the paravisor
+	 * in the guest context and must be mapped private.
+	 */
+	if (addr >= HV_IOAPIC_BASE_ADDRESS &&
+	    addr < (HV_IOAPIC_BASE_ADDRESS + PAGE_SIZE))
+		return true;
+
+	/* Same with a vTPM */
+	if (addr >= VTPM_BASE_ADDRESS &&
+	    addr < (VTPM_BASE_ADDRESS + PAGE_SIZE))
+		return true;
+
+	return false;
+}
+
+void __init hv_vtom_init(void)
+{
+	enum hv_isolation_type type = hv_get_isolation_type();
+
+	switch (type) {
+	case HV_ISOLATION_TYPE_VBS:
+		fallthrough;
+	/*
+	 * By design, a VM using vTOM doesn't see the SEV setting,
+	 * so SEV initialization is bypassed and sev_status isn't set.
+	 * Set it here to indicate a vTOM VM.
+	 *
+	 * Note: if CONFIG_AMD_MEM_ENCRYPT is not set, sev_status is
+	 * defined as 0ULL, to which we can't assigned a value.
+	 */
+#ifdef CONFIG_AMD_MEM_ENCRYPT
+	case HV_ISOLATION_TYPE_SNP:
+		sev_status = MSR_AMD64_SNP_VTOM;
+		cc_vendor = CC_VENDOR_AMD;
+		break;
+#endif
+
+	case HV_ISOLATION_TYPE_TDX:
+		cc_vendor = CC_VENDOR_INTEL;
+		break;
+
+	default:
+		panic("hv_vtom_init: unsupported isolation type %d\n", type);
+	}
+
+	cc_set_mask(ms_hyperv.shared_gpa_boundary);
+	physical_mask &= ms_hyperv.shared_gpa_boundary - 1;
+
+	x86_platform.hyper.is_private_mmio = hv_is_private_mmio;
+	x86_platform.guest.enc_cache_flush_required = hv_vtom_cache_flush_required;
+	x86_platform.guest.enc_tlb_flush_required = hv_vtom_tlb_flush_required;
+	x86_platform.guest.enc_status_change_prepare = hv_vtom_clear_present;
+	x86_platform.guest.enc_status_change_finish = hv_vtom_set_host_visibility;
+	x86_platform.guest.enc_kexec_begin = hv_vtom_kexec_begin;
+	x86_platform.guest.enc_kexec_finish = hv_vtom_kexec_finish;
+
+	/* Set WB as the default cache mode. */
+	guest_force_mtrr_state(NULL, 0, MTRR_TYPE_WRBACK);
+}
+
+#endif /* defined(CONFIG_AMD_MEM_ENCRYPT) || defined(CONFIG_INTEL_TDX_GUEST) */
+
+enum hv_isolation_type hv_get_isolation_type(void)
+{
+	if (!(ms_hyperv.priv_high & HV_ISOLATION))
+		return HV_ISOLATION_TYPE_NONE;
+	return FIELD_GET(HV_ISOLATION_TYPE, ms_hyperv.isolation_config_b);
+}
+EXPORT_SYMBOL_GPL(hv_get_isolation_type);
+
+/*
+ * hv_is_isolation_supported - Check system runs in the Hyper-V
+ * isolation VM.
+ */
+bool hv_is_isolation_supported(void)
+{
+	if (!cpu_feature_enabled(X86_FEATURE_HYPERVISOR))
+		return false;
+
+	if (!hypervisor_is_type(X86_HYPER_MS_HYPERV))
+		return false;
+
+	return hv_get_isolation_type() != HV_ISOLATION_TYPE_NONE;
+}
+
+DEFINE_STATIC_KEY_FALSE(isolation_type_snp);
+
+/*
+ * hv_isolation_type_snp - Check if the system runs in an AMD SEV-SNP based
+ * isolation VM.
+ */
+bool hv_isolation_type_snp(void)
+{
+	return static_branch_unlikely(&isolation_type_snp);
+}
+
+DEFINE_STATIC_KEY_FALSE(isolation_type_tdx);
+/*
+ * hv_isolation_type_tdx - Check if the system runs in an Intel TDX based
+ * isolated VM.
+ */
+bool hv_isolation_type_tdx(void)
+{
+	return static_branch_unlikely(&isolation_type_tdx);
+}

linux-7.2.3/arch/x86/hyperv/mmu.c

@@ -0,0 +1,246 @@
+#define pr_fmt(fmt)  "Hyper-V: " fmt
+
+#include <linux/log2.h>
+#include <linux/slab.h>
+#include <linux/types.h>
+
+#include <asm/fpu/api.h>
+#include <asm/mshyperv.h>
+#include <asm/msr.h>
+#include <asm/tlbflush.h>
+#include <asm/tlb.h>
+
+#define CREATE_TRACE_POINTS
+#include <asm/trace/hyperv.h>
+
+/* Each gva in gva_list encodes up to 4096 pages to flush */
+#define HV_TLB_FLUSH_UNIT (4096 * PAGE_SIZE)
+
+static u64 hyperv_flush_tlb_others_ex(const struct cpumask *cpus,
+				      const struct flush_tlb_info *info);
+
+/*
+ * Fills in gva_list starting from offset. Returns the number of items added.
+ */
+static inline int fill_gva_list(u64 gva_list[], int offset,
+				unsigned long start, unsigned long end)
+{
+	int gva_n = offset;
+	unsigned long cur = start, diff;
+
+	do {
+		diff = end > cur ? end - cur : 0;
+
+		gva_list[gva_n] = cur & PAGE_MASK;
+		/*
+		 * Lower 12 bits encode the number of additional
+		 * pages to flush (in addition to the 'cur' page).
+		 */
+		if (diff >= HV_TLB_FLUSH_UNIT) {
+			gva_list[gva_n] |= ~PAGE_MASK;
+			cur += HV_TLB_FLUSH_UNIT;
+		}  else if (diff) {
+			gva_list[gva_n] |= (diff - 1) >> PAGE_SHIFT;
+			cur = end;
+		}
+
+		gva_n++;
+
+	} while (cur < end);
+
+	return gva_n - offset;
+}
+
+static bool cpu_is_lazy(int cpu)
+{
+	return per_cpu(cpu_tlbstate_shared.is_lazy, cpu);
+}
+
+static void hyperv_flush_tlb_multi(const struct cpumask *cpus,
+				   const struct flush_tlb_info *info)
+{
+	int cpu, vcpu, gva_n, max_gvas;
+	struct hv_tlb_flush *flush;
+	u64 status;
+	unsigned long flags;
+	bool do_lazy = !info->freed_tables;
+
+	trace_hyperv_mmu_flush_tlb_multi(cpus, info);
+
+	if (!hv_hypercall_pg)
+		goto do_native;
+
+	local_irq_save(flags);
+
+	flush = *this_cpu_ptr(hyperv_pcpu_input_arg);
+
+	if (unlikely(!flush)) {
+		local_irq_restore(flags);
+		goto do_native;
+	}
+
+	if (info->mm) {
+		/*
+		 * AddressSpace argument must match the CR3 with PCID bits
+		 * stripped out.
+		 */
+		flush->address_space = virt_to_phys(info->mm->pgd);
+		flush->address_space &= CR3_ADDR_MASK;
+		flush->flags = 0;
+	} else {
+		flush->address_space = 0;
+		flush->flags = HV_FLUSH_ALL_VIRTUAL_ADDRESS_SPACES;
+	}
+
+	flush->processor_mask = 0;
+	if (cpumask_equal(cpus, cpu_present_mask)) {
+		flush->flags |= HV_FLUSH_ALL_PROCESSORS;
+	} else {
+		/*
+		 * From the supplied CPU set we need to figure out if we can get
+		 * away with cheaper HVCALL_FLUSH_VIRTUAL_ADDRESS_{LIST,SPACE}
+		 * hypercalls. This is possible when the highest VP number in
+		 * the set is < 64. As VP numbers are usually in ascending order
+		 * and match Linux CPU ids, here is an optimization: we check
+		 * the VP number for the highest bit in the supplied set first
+		 * so we can quickly find out if using *_EX hypercalls is a
+		 * must. We will also check all VP numbers when walking the
+		 * supplied CPU set to remain correct in all cases.
+		 */
+		cpu = cpumask_last(cpus);
+
+		if (cpu < nr_cpumask_bits && hv_cpu_number_to_vp_number(cpu) >= 64)
+			goto do_ex_hypercall;
+
+		for_each_cpu(cpu, cpus) {
+			if (do_lazy && cpu_is_lazy(cpu))
+				continue;
+			vcpu = hv_cpu_number_to_vp_number(cpu);
+			if (vcpu == VP_INVAL) {
+				local_irq_restore(flags);
+				goto do_native;
+			}
+
+			if (vcpu >= 64)
+				goto do_ex_hypercall;
+
+			__set_bit(vcpu, (unsigned long *)
+				  &flush->processor_mask);
+		}
+
+		/* nothing to flush if 'processor_mask' ends up being empty */
+		if (!flush->processor_mask) {
+			local_irq_restore(flags);
+			return;
+		}
+	}
+
+	/*
+	 * We can flush not more than max_gvas with one hypercall. Flush the
+	 * whole address space if we were asked to do more.
+	 */
+	max_gvas = (PAGE_SIZE - sizeof(*flush)) / sizeof(flush->gva_list[0]);
+
+	if (info->end == TLB_FLUSH_ALL) {
+		flush->flags |= HV_FLUSH_NON_GLOBAL_MAPPINGS_ONLY;
+		status = hv_do_hypercall(HVCALL_FLUSH_VIRTUAL_ADDRESS_SPACE,
+					 flush, NULL);
+	} else if (info->end &&
+		   ((info->end - info->start)/HV_TLB_FLUSH_UNIT) > max_gvas) {
+		status = hv_do_hypercall(HVCALL_FLUSH_VIRTUAL_ADDRESS_SPACE,
+					 flush, NULL);
+	} else {
+		gva_n = fill_gva_list(flush->gva_list, 0,
+				      info->start, info->end);
+		status = hv_do_rep_hypercall(HVCALL_FLUSH_VIRTUAL_ADDRESS_LIST,
+					     gva_n, 0, flush, NULL);
+	}
+	goto check_status;
+
+do_ex_hypercall:
+	status = hyperv_flush_tlb_others_ex(cpus, info);
+
+check_status:
+	local_irq_restore(flags);
+
+	if (hv_result_success(status))
+		return;
+do_native:
+	native_flush_tlb_multi(cpus, info);
+}
+
+static u64 hyperv_flush_tlb_others_ex(const struct cpumask *cpus,
+				      const struct flush_tlb_info *info)
+{
+	int nr_bank = 0, max_gvas, gva_n;
+	struct hv_tlb_flush_ex *flush;
+	u64 status;
+
+	if (!(ms_hyperv.hints & HV_X64_EX_PROCESSOR_MASKS_RECOMMENDED))
+		return HV_STATUS_INVALID_PARAMETER;
+
+	flush = *this_cpu_ptr(hyperv_pcpu_input_arg);
+
+	if (info->mm) {
+		/*
+		 * AddressSpace argument must match the CR3 with PCID bits
+		 * stripped out.
+		 */
+		flush->address_space = virt_to_phys(info->mm->pgd);
+		flush->address_space &= CR3_ADDR_MASK;
+		flush->flags = 0;
+	} else {
+		flush->address_space = 0;
+		flush->flags = HV_FLUSH_ALL_VIRTUAL_ADDRESS_SPACES;
+	}
+
+	flush->hv_vp_set.valid_bank_mask = 0;
+
+	flush->hv_vp_set.format = HV_GENERIC_SET_SPARSE_4K;
+	nr_bank = cpumask_to_vpset_skip(&flush->hv_vp_set, cpus,
+			info->freed_tables ? NULL : cpu_is_lazy);
+	if (nr_bank < 0)
+		return HV_STATUS_INVALID_PARAMETER;
+
+	/*
+	 * We can flush not more than max_gvas with one hypercall. Flush the
+	 * whole address space if we were asked to do more.
+	 *
+	 * For these hypercalls, Hyper-V treats the valid_bank_mask field
+	 * of flush->hv_vp_set as part of the fixed size input header.
+	 * So the variable input header size is equal to nr_bank.
+	 */
+	max_gvas =
+		(PAGE_SIZE - sizeof(*flush) - nr_bank *
+		 sizeof(flush->hv_vp_set.bank_contents[0])) /
+		sizeof(flush->gva_list[0]);
+
+	if (info->end == TLB_FLUSH_ALL) {
+		flush->flags |= HV_FLUSH_NON_GLOBAL_MAPPINGS_ONLY;
+		status = hv_do_rep_hypercall(
+			HVCALL_FLUSH_VIRTUAL_ADDRESS_SPACE_EX,
+			0, nr_bank, flush, NULL);
+	} else if (info->end &&
+		   ((info->end - info->start)/HV_TLB_FLUSH_UNIT) > max_gvas) {
+		status = hv_do_rep_hypercall(
+			HVCALL_FLUSH_VIRTUAL_ADDRESS_SPACE_EX,
+			0, nr_bank, flush, NULL);
+	} else {
+		gva_n = fill_gva_list(flush->gva_list, nr_bank,
+				      info->start, info->end);
+		status = hv_do_rep_hypercall(
+			HVCALL_FLUSH_VIRTUAL_ADDRESS_LIST_EX,
+			gva_n, nr_bank, flush, NULL);
+	}
+
+	return status;
+}
+
+void hyperv_setup_mmu_ops(void)
+{
+	if (!(ms_hyperv.hints & HV_X64_REMOTE_TLB_FLUSH_RECOMMENDED))
+		return;
+
+	pr_info("Using hypercall for remote TLB flush\n");
+	pv_ops.mmu.flush_tlb_multi = hyperv_flush_tlb_multi;
+}

linux-7.2.3/arch/x86/hyperv/mshv-asm-offsets.c

@@ -0,0 +1,37 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Generate definitions needed by assembly language modules.
+ * This code generates raw asm output which is post-processed to extract
+ * and format the required data.
+ *
+ * Copyright (c) 2025, Microsoft Corporation.
+ *
+ * Author:
+ *   Naman Jain <namjain@microsoft.com>
+ */
+#define COMPILE_OFFSETS
+
+#include <linux/kbuild.h>
+#include <asm/mshyperv.h>
+
+static void __used common(void)
+{
+	if (IS_ENABLED(CONFIG_HYPERV_VTL_MODE)) {
+		OFFSET(MSHV_VTL_CPU_CONTEXT_rax, mshv_vtl_cpu_context, rax);
+		OFFSET(MSHV_VTL_CPU_CONTEXT_rcx, mshv_vtl_cpu_context, rcx);
+		OFFSET(MSHV_VTL_CPU_CONTEXT_rdx, mshv_vtl_cpu_context, rdx);
+		OFFSET(MSHV_VTL_CPU_CONTEXT_rbx, mshv_vtl_cpu_context, rbx);
+		OFFSET(MSHV_VTL_CPU_CONTEXT_rbp, mshv_vtl_cpu_context, rbp);
+		OFFSET(MSHV_VTL_CPU_CONTEXT_rsi, mshv_vtl_cpu_context, rsi);
+		OFFSET(MSHV_VTL_CPU_CONTEXT_rdi, mshv_vtl_cpu_context, rdi);
+		OFFSET(MSHV_VTL_CPU_CONTEXT_r8,  mshv_vtl_cpu_context, r8);
+		OFFSET(MSHV_VTL_CPU_CONTEXT_r9,  mshv_vtl_cpu_context, r9);
+		OFFSET(MSHV_VTL_CPU_CONTEXT_r10, mshv_vtl_cpu_context, r10);
+		OFFSET(MSHV_VTL_CPU_CONTEXT_r11, mshv_vtl_cpu_context, r11);
+		OFFSET(MSHV_VTL_CPU_CONTEXT_r12, mshv_vtl_cpu_context, r12);
+		OFFSET(MSHV_VTL_CPU_CONTEXT_r13, mshv_vtl_cpu_context, r13);
+		OFFSET(MSHV_VTL_CPU_CONTEXT_r14, mshv_vtl_cpu_context, r14);
+		OFFSET(MSHV_VTL_CPU_CONTEXT_r15, mshv_vtl_cpu_context, r15);
+		OFFSET(MSHV_VTL_CPU_CONTEXT_cr2, mshv_vtl_cpu_context, cr2);
+	}
+}

linux-7.2.3/arch/x86/hyperv/mshv_vtl_asm.S

@@ -0,0 +1,116 @@
+/* SPDX-License-Identifier: GPL-2.0
+ *
+ * Assembly level code for mshv_vtl VTL transition
+ *
+ * Copyright (c) 2025, Microsoft Corporation.
+ *
+ * Author:
+ *   Naman Jain <namjain@microsoft.com>
+ */
+
+#include <linux/linkage.h>
+#include <linux/static_call_types.h>
+#include <asm/asm.h>
+#include <asm/asm-offsets.h>
+#include <asm/frame.h>
+#include "mshv-asm-offsets.h"
+
+	.text
+	.section .noinstr.text, "ax"
+/*
+ * void __mshv_vtl_return_call(struct mshv_vtl_cpu_context *vtl0)
+ *
+ * This function is used to context switch between different Virtual Trust Levels.
+ * It is marked as 'noinstr' to prevent against instrumentation and debugging facilities.
+ * NMIs aren't a problem because the NMI handler saves/restores CR2 specifically to guard
+ * against #PFs in NMI context clobbering the guest state.
+ */
+SYM_FUNC_START(__mshv_vtl_return_call)
+	/* Push callee save registers */
+	pushq %rbp
+	mov %rsp, %rbp
+	pushq %r12
+	pushq %r13
+	pushq %r14
+	pushq %r15
+	pushq %rbx
+
+	/* register switch to VTL0 clobbers all registers except rax/rcx */
+	mov %_ASM_ARG1, %rax
+
+	/* grab rbx/rbp/rsi/rdi/r8-r15 */
+	mov MSHV_VTL_CPU_CONTEXT_rbx(%rax), %rbx
+	mov MSHV_VTL_CPU_CONTEXT_rbp(%rax), %rbp
+	mov MSHV_VTL_CPU_CONTEXT_rsi(%rax), %rsi
+	mov MSHV_VTL_CPU_CONTEXT_rdi(%rax), %rdi
+	mov MSHV_VTL_CPU_CONTEXT_r8(%rax), %r8
+	mov MSHV_VTL_CPU_CONTEXT_r9(%rax), %r9
+	mov MSHV_VTL_CPU_CONTEXT_r10(%rax), %r10
+	mov MSHV_VTL_CPU_CONTEXT_r11(%rax), %r11
+	mov MSHV_VTL_CPU_CONTEXT_r12(%rax), %r12
+	mov MSHV_VTL_CPU_CONTEXT_r13(%rax), %r13
+	mov MSHV_VTL_CPU_CONTEXT_r14(%rax), %r14
+	mov MSHV_VTL_CPU_CONTEXT_r15(%rax), %r15
+
+	mov MSHV_VTL_CPU_CONTEXT_cr2(%rax), %rdx
+	mov %rdx, %cr2
+	mov MSHV_VTL_CPU_CONTEXT_rdx(%rax), %rdx
+
+	/* stash host registers on stack */
+	pushq %rax
+	pushq %rcx
+
+	xor %ecx, %ecx
+
+	/* make a hypercall to switch VTL */
+	call STATIC_CALL_TRAMP_STR(__mshv_vtl_return_hypercall)
+
+	/* stash guest registers on stack, restore saved host copies */
+	pushq %rax
+	pushq %rcx
+	mov 16(%rsp), %rcx
+	mov 24(%rsp), %rax
+
+	mov %rdx, MSHV_VTL_CPU_CONTEXT_rdx(%rax)
+	mov %cr2, %rdx
+	mov %rdx, MSHV_VTL_CPU_CONTEXT_cr2(%rax)
+	pop MSHV_VTL_CPU_CONTEXT_rcx(%rax)
+	pop MSHV_VTL_CPU_CONTEXT_rax(%rax)
+	add $16, %rsp
+
+	/* save rbx/rbp/rsi/rdi/r8-r15 */
+	mov %rbx, MSHV_VTL_CPU_CONTEXT_rbx(%rax)
+	mov %rbp, MSHV_VTL_CPU_CONTEXT_rbp(%rax)
+	mov %rsi, MSHV_VTL_CPU_CONTEXT_rsi(%rax)
+	mov %rdi, MSHV_VTL_CPU_CONTEXT_rdi(%rax)
+	mov %r8,  MSHV_VTL_CPU_CONTEXT_r8(%rax)
+	mov %r9,  MSHV_VTL_CPU_CONTEXT_r9(%rax)
+	mov %r10, MSHV_VTL_CPU_CONTEXT_r10(%rax)
+	mov %r11, MSHV_VTL_CPU_CONTEXT_r11(%rax)
+	mov %r12, MSHV_VTL_CPU_CONTEXT_r12(%rax)
+	mov %r13, MSHV_VTL_CPU_CONTEXT_r13(%rax)
+	mov %r14, MSHV_VTL_CPU_CONTEXT_r14(%rax)
+	mov %r15, MSHV_VTL_CPU_CONTEXT_r15(%rax)
+
+	/* pop callee-save registers r12-r15, rbx */
+	pop %rbx
+	pop %r15
+	pop %r14
+	pop %r13
+	pop %r12
+
+	pop %rbp
+	RET
+SYM_FUNC_END(__mshv_vtl_return_call)
+/*
+ * Make sure that static_call_key symbol: __SCK____mshv_vtl_return_hypercall is accessible here.
+ * Below code is inspired from __ADDRESSABLE(sym) macro. Symbol name is kept simple, to avoid
+ * naming it something like "__UNIQUE_ID_addressable___SCK____mshv_vtl_return_hypercall_662.0"
+ * which would otherwise have been generated by the macro.
+ */
+	.section	.discard.addressable,"aw"
+	.align 8
+	.type	mshv_vtl_return_sym, @object
+	.size	mshv_vtl_return_sym, 8
+mshv_vtl_return_sym:
+	.quad	__SCK____mshv_vtl_return_hypercall

linux-7.2.3/arch/x86/hyperv/nested.c

@@ -0,0 +1,130 @@
+// SPDX-License-Identifier: GPL-2.0
+
+/*
+ * Hyper-V nested virtualization code.
+ *
+ * Copyright (C) 2018, Microsoft, Inc.
+ *
+ * Author : Lan Tianyu <Tianyu.Lan@microsoft.com>
+ */
+#define pr_fmt(fmt)  "Hyper-V: " fmt
+
+
+#include <linux/types.h>
+#include <linux/export.h>
+#include <hyperv/hvhdk.h>
+#include <asm/mshyperv.h>
+#include <asm/tlbflush.h>
+
+#include <asm/trace/hyperv.h>
+
+int hyperv_flush_guest_mapping(u64 as)
+{
+	struct hv_guest_mapping_flush *flush;
+	u64 status;
+	unsigned long flags;
+	int ret = -ENOTSUPP;
+
+	if (!hv_hypercall_pg)
+		goto fault;
+
+	local_irq_save(flags);
+
+	flush = *this_cpu_ptr(hyperv_pcpu_input_arg);
+
+	if (unlikely(!flush)) {
+		local_irq_restore(flags);
+		goto fault;
+	}
+
+	flush->address_space = as;
+	flush->flags = 0;
+
+	status = hv_do_hypercall(HVCALL_FLUSH_GUEST_PHYSICAL_ADDRESS_SPACE,
+				 flush, NULL);
+	local_irq_restore(flags);
+
+	if (hv_result_success(status))
+		ret = 0;
+
+fault:
+	trace_hyperv_nested_flush_guest_mapping(as, ret);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(hyperv_flush_guest_mapping);
+
+int hyperv_fill_flush_guest_mapping_list(
+		struct hv_guest_mapping_flush_list *flush,
+		u64 start_gfn, u64 pages)
+{
+	u64 cur = start_gfn;
+	u64 additional_pages;
+	int gpa_n = 0;
+
+	do {
+		/*
+		 * If flush requests exceed max flush count, go back to
+		 * flush tlbs without range.
+		 */
+		if (gpa_n >= HV_MAX_FLUSH_REP_COUNT)
+			return -ENOSPC;
+
+		additional_pages = min_t(u64, pages, HV_MAX_FLUSH_PAGES) - 1;
+
+		flush->gpa_list[gpa_n].page.additional_pages = additional_pages;
+		flush->gpa_list[gpa_n].page.largepage = false;
+		flush->gpa_list[gpa_n].page.basepfn = cur;
+
+		pages -= additional_pages + 1;
+		cur += additional_pages + 1;
+		gpa_n++;
+	} while (pages > 0);
+
+	return gpa_n;
+}
+EXPORT_SYMBOL_GPL(hyperv_fill_flush_guest_mapping_list);
+
+int hyperv_flush_guest_mapping_range(u64 as,
+		hyperv_fill_flush_list_func fill_flush_list_func, void *data)
+{
+	struct hv_guest_mapping_flush_list *flush;
+	u64 status;
+	unsigned long flags;
+	int ret = -ENOTSUPP;
+	int gpa_n = 0;
+
+	if (!hv_hypercall_pg || !fill_flush_list_func)
+		goto fault;
+
+	local_irq_save(flags);
+
+	flush = *this_cpu_ptr(hyperv_pcpu_input_arg);
+
+	if (unlikely(!flush)) {
+		local_irq_restore(flags);
+		goto fault;
+	}
+
+	flush->address_space = as;
+	flush->flags = 0;
+
+	gpa_n = fill_flush_list_func(flush, data);
+	if (gpa_n < 0) {
+		local_irq_restore(flags);
+		goto fault;
+	}
+
+	status = hv_do_rep_hypercall(HVCALL_FLUSH_GUEST_PHYSICAL_ADDRESS_LIST,
+				     gpa_n, 0, flush, NULL);
+
+	local_irq_restore(flags);
+
+	if (hv_result_success(status))
+		ret = 0;
+	else
+		ret = hv_result(status);
+fault:
+	trace_hyperv_nested_flush_guest_mapping_range(as, ret);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(hyperv_flush_guest_mapping_range);