Resguardo inicial

linux-7.2.3/arch/arm64/kvm/hyp/nvhe/.gitignore

@@ -0,0 +1,4 @@
+# SPDX-License-Identifier: GPL-2.0-only
+gen-hyprel
+hyp.lds
+hyp-reloc.S

linux-7.2.3/arch/arm64/kvm/hyp/nvhe/Makefile

@@ -0,0 +1,112 @@
+# SPDX-License-Identifier: GPL-2.0
+#
+# Makefile for Kernel-based Virtual Machine module, HYP/nVHE part
+#
+
+asflags-y := -D__KVM_NVHE_HYPERVISOR__ -D__DISABLE_EXPORTS
+
+# Tracepoint and MMIO logging symbols should not be visible at nVHE KVM as
+# there is no way to execute them and any such MMIO access from nVHE KVM
+# will explode instantly (Words of Marc Zyngier). So introduce a generic flag
+# __DISABLE_TRACE_MMIO__ to disable MMIO tracing for nVHE KVM.
+ccflags-y := -D__KVM_NVHE_HYPERVISOR__ -D__DISABLE_EXPORTS -D__DISABLE_TRACE_MMIO__
+ccflags-y += -fno-stack-protector	\
+	     -DDISABLE_BRANCH_PROFILING	\
+	     $(DISABLE_KSTACK_ERASE)
+
+hostprogs := gen-hyprel
+HOST_EXTRACFLAGS += -I$(objtree)/include
+
+lib-objs := clear_page.o copy_page.o memcpy.o memset.o tishift.o
+lib-objs := $(addprefix ../../../lib/, $(lib-objs))
+
+CFLAGS_switch.nvhe.o += -Wno-override-init
+
+hyp-obj-y := timer-sr.o sysreg-sr.o debug-sr.o switch.o tlb.o hyp-init.o host.o \
+	 hyp-main.o hyp-smp.o psci-relay.o early_alloc.o page_alloc.o \
+	 cache.o setup.o mm.o mem_protect.o sys_regs.o pkvm.o stacktrace.o ffa.o
+hyp-obj-y += ../vgic-v3-sr.o ../aarch32.o ../vgic-v2-cpuif-proxy.o ../entry.o \
+	 ../hyp-entry.o ../exception.o ../pgtable.o ../vgic-v5-sr.o
+hyp-obj-y += ../../../kernel/smccc-call.o
+hyp-obj-$(CONFIG_LIST_HARDENED) += list_debug.o
+hyp-obj-$(CONFIG_NVHE_EL2_TRACING) += clock.o trace.o events.o
+hyp-obj-y += $(lib-objs)
+
+# Path to simple_ring_buffer.c
+CFLAGS_trace.nvhe.o += -I$(srctree)/kernel/trace/
+
+##
+## Build rules for compiling nVHE hyp code
+## Output of this folder is `kvm_nvhe.o`, a partially linked object
+## file containing all nVHE hyp code and data.
+##
+
+hyp-obj := $(patsubst %.o,%.nvhe.o,$(hyp-obj-y))
+obj-y := kvm_nvhe.o
+targets += $(hyp-obj) kvm_nvhe.tmp.o kvm_nvhe.rel.o hyp.lds hyp-reloc.S hyp-reloc.o
+
+# 1) Compile all source files to `.nvhe.o` object files. The file extension
+#    avoids file name clashes for files shared with VHE.
+$(obj)/%.nvhe.o: $(src)/%.c FORCE
+	$(call if_changed_rule,cc_o_c)
+$(obj)/%.nvhe.o: $(src)/%.S FORCE
+	$(call if_changed_rule,as_o_S)
+
+# 2) Compile linker script.
+$(obj)/hyp.lds: $(src)/hyp.lds.S FORCE
+	$(call if_changed_dep,cpp_lds_S)
+
+# 3) Partially link all '.nvhe.o' files and apply the linker script.
+#    Prefixes names of ELF sections with '.hyp', eg. '.hyp.text'.
+#    Note: The following rule assumes that the 'ld' rule puts LDFLAGS before
+#          the list of dependencies to form '-T $(obj)/hyp.lds'. This is to
+#          keep the dependency on the target while avoiding an error from
+#          GNU ld if the linker script is passed to it twice.
+LDFLAGS_kvm_nvhe.tmp.o := -r -T
+$(obj)/kvm_nvhe.tmp.o: $(obj)/hyp.lds $(addprefix $(obj)/,$(hyp-obj)) FORCE
+	$(call if_changed,ld)
+
+# 4) Generate list of hyp code/data positions that need to be relocated at
+#    runtime. Because the hypervisor is part of the kernel binary, relocations
+#    produce a kernel VA. We enumerate relocations targeting hyp at build time
+#    and convert the kernel VAs at those positions to hyp VAs.
+$(obj)/hyp-reloc.S: $(obj)/kvm_nvhe.tmp.o $(obj)/gen-hyprel FORCE
+	$(call if_changed,hyprel)
+
+# 5) Compile hyp-reloc.S and link it into the existing partially linked object.
+#    The object file now contains a section with pointers to hyp positions that
+#    will contain kernel VAs at runtime. These pointers have relocations on them
+#    so that they get updated as the hyp object is linked into `vmlinux`.
+LDFLAGS_kvm_nvhe.rel.o := -r
+$(obj)/kvm_nvhe.rel.o: $(obj)/kvm_nvhe.tmp.o $(obj)/hyp-reloc.o FORCE
+	$(call if_changed,ld)
+
+# 6) Produce the final 'kvm_nvhe.o', ready to be linked into 'vmlinux'.
+#    Prefixes names of ELF symbols with '__kvm_nvhe_'.
+$(obj)/kvm_nvhe.o: $(obj)/kvm_nvhe.rel.o FORCE
+	$(call if_changed,hypcopy)
+
+# The HYPREL command calls `gen-hyprel` to generate an assembly file with
+# a list of relocations targeting hyp code/data.
+quiet_cmd_hyprel = HYPREL  $@
+      cmd_hyprel = $(obj)/gen-hyprel $< > $@
+
+# The HYPCOPY command uses `objcopy` to prefix all ELF symbol names
+# to avoid clashes with VHE code/data.
+quiet_cmd_hypcopy = HYPCOPY $@
+      cmd_hypcopy = $(OBJCOPY) --prefix-symbols=__kvm_nvhe_ $< $@
+
+# Remove ftrace and Shadow Call Stack CFLAGS.
+# This is equivalent to the 'notrace' and '__noscs' annotations.
+KBUILD_CFLAGS := $(filter-out $(CC_FLAGS_FTRACE) $(CC_FLAGS_SCS), $(KBUILD_CFLAGS))
+# Starting from 13.0.0 llvm emits SHT_REL section '.llvm.call-graph-profile'
+# when profile optimization is applied. gen-hyprel does not support SHT_REL and
+# causes a build failure. Remove profile optimization flags.
+KBUILD_CFLAGS := $(filter-out -fprofile-sample-use=% -fprofile-use=%, $(KBUILD_CFLAGS))
+KBUILD_CFLAGS += -fno-asynchronous-unwind-tables -fno-unwind-tables
+
+ifeq ($(CONFIG_UBSAN_KVM_EL2),y)
+UBSAN_SANITIZE := y
+# Always use brk and not hooks
+ccflags-y += $(CFLAGS_UBSAN_TRAP)
+endif

linux-7.2.3/arch/arm64/kvm/hyp/nvhe/cache.S

@@ -0,0 +1,25 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Code copied from arch/arm64/mm/cache.S.
+ */
+
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+#include <asm/alternative.h>
+
+SYM_FUNC_START(__pi_dcache_clean_inval_poc)
+	dcache_by_line_op civac, sy, x0, x1, x2, x3
+	ret
+SYM_FUNC_END(__pi_dcache_clean_inval_poc)
+SYM_FUNC_ALIAS(dcache_clean_inval_poc, __pi_dcache_clean_inval_poc)
+
+SYM_FUNC_START(__pi_icache_inval_pou)
+alternative_if ARM64_HAS_CACHE_DIC
+	isb
+	ret
+alternative_else_nop_endif
+
+	invalidate_icache_by_line x0, x1, x2, x3
+	ret
+SYM_FUNC_END(__pi_icache_inval_pou)
+SYM_FUNC_ALIAS(icache_inval_pou, __pi_icache_inval_pou)

linux-7.2.3/arch/arm64/kvm/hyp/nvhe/clock.c

@@ -0,0 +1,68 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (C) 2025 Google LLC
+ * Author: Vincent Donnefort <vdonnefort@google.com>
+ */
+
+#include <nvhe/clock.h>
+
+#include <asm/arch_timer.h>
+#include <asm/div64.h>
+
+static struct clock_data {
+	struct {
+		u32 mult;
+		u32 shift;
+		u64 epoch_ns;
+		u64 epoch_cyc;
+		u64 cyc_overflow64;
+	} data[2];
+	u64 cur;
+} trace_clock_data;
+
+static u64 __clock_mult_uint128(u64 cyc, u32 mult, u32 shift)
+{
+	__uint128_t ns = (__uint128_t)cyc * mult;
+
+	ns >>= shift;
+
+	return (u64)ns;
+}
+
+/* Does not guarantee no reader on the modified bank. */
+void trace_clock_update(u32 mult, u32 shift, u64 epoch_ns, u64 epoch_cyc)
+{
+	struct clock_data *clock = &trace_clock_data;
+	u64 bank = clock->cur ^ 1;
+
+	if (!mult || shift >= 64)
+		return;
+
+	clock->data[bank].mult			= mult;
+	clock->data[bank].shift			= shift;
+	clock->data[bank].epoch_ns		= epoch_ns;
+	clock->data[bank].epoch_cyc		= epoch_cyc;
+	clock->data[bank].cyc_overflow64	= ULONG_MAX / mult;
+
+	smp_store_release(&clock->cur, bank);
+}
+
+/* Use untrusted host data */
+u64 trace_clock(void)
+{
+	struct clock_data *clock = &trace_clock_data;
+	u64 bank = smp_load_acquire(&clock->cur);
+	u64 cyc, ns;
+
+	cyc = __arch_counter_get_cntvct() - clock->data[bank].epoch_cyc;
+
+	if (likely(cyc < clock->data[bank].cyc_overflow64)) {
+		ns = cyc * clock->data[bank].mult;
+		ns >>= clock->data[bank].shift;
+	} else {
+		ns = __clock_mult_uint128(cyc, clock->data[bank].mult,
+					  clock->data[bank].shift);
+	}
+
+	return (u64)ns + clock->data[bank].epoch_ns;
+}

linux-7.2.3/arch/arm64/kvm/hyp/nvhe/debug-sr.c

@@ -0,0 +1,221 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2015 - ARM Ltd
+ * Author: Marc Zyngier <marc.zyngier@arm.com>
+ */
+
+#include <hyp/debug-sr.h>
+
+#include <linux/compiler.h>
+#include <linux/kvm_host.h>
+
+#include <asm/debug-monitors.h>
+#include <asm/kvm_asm.h>
+#include <asm/kvm_hyp.h>
+#include <asm/kvm_mmu.h>
+
+static void __debug_save_spe(void)
+{
+	u64 *pmscr_el1, *pmblimitr_el1;
+
+	pmscr_el1 = host_data_ptr(host_debug_state.pmscr_el1);
+	pmblimitr_el1 = host_data_ptr(host_debug_state.pmblimitr_el1);
+
+	/*
+	 * At this point, we know that this CPU implements
+	 * SPE and is available to the host.
+	 * Check if the host is actually using it ?
+	 */
+	*pmblimitr_el1 = read_sysreg_s(SYS_PMBLIMITR_EL1);
+	if (!(*pmblimitr_el1 & BIT(PMBLIMITR_EL1_E_SHIFT)))
+		return;
+
+	/* Yes; save the control register and disable data generation */
+	*pmscr_el1 = read_sysreg_el1(SYS_PMSCR);
+	write_sysreg_el1(0, SYS_PMSCR);
+	isb();
+
+	/* Now drain all buffered data to memory */
+	psb_csync();
+	dsb(nsh);
+
+	/* And disable the profiling buffer */
+	write_sysreg_s(0, SYS_PMBLIMITR_EL1);
+	isb();
+}
+
+static void __debug_restore_spe(void)
+{
+	u64 pmblimitr_el1 = *host_data_ptr(host_debug_state.pmblimitr_el1);
+
+	if (!(pmblimitr_el1 & BIT(PMBLIMITR_EL1_E_SHIFT)))
+		return;
+
+	/* The host page table is installed, but not yet synchronised */
+	isb();
+
+	/* Re-enable the profiling buffer. */
+	write_sysreg_s(pmblimitr_el1, SYS_PMBLIMITR_EL1);
+	isb();
+
+	/* Re-enable data generation */
+	write_sysreg_el1(*host_data_ptr(host_debug_state.pmscr_el1), SYS_PMSCR);
+}
+
+static void __trace_do_switch(u64 *saved_trfcr, u64 new_trfcr)
+{
+	*saved_trfcr = read_sysreg_el1(SYS_TRFCR);
+	write_sysreg_el1(new_trfcr, SYS_TRFCR);
+}
+
+static void __trace_drain_and_disable(void)
+{
+	u64 *trblimitr_el1 = host_data_ptr(host_debug_state.trblimitr_el1);
+	bool needs_drain = is_protected_kvm_enabled() ?
+			   host_data_test_flag(HAS_TRBE) :
+			   host_data_test_flag(TRBE_ENABLED);
+
+	if (!needs_drain) {
+		*trblimitr_el1 = 0;
+		return;
+	}
+
+	*trblimitr_el1 = read_sysreg_s(SYS_TRBLIMITR_EL1);
+	if (*trblimitr_el1 & TRBLIMITR_EL1_E) {
+		/*
+		 * The host has enabled the Trace Buffer Unit so we have
+		 * to beat the CPU with a stick until it stops accessing
+		 * memory.
+		 */
+
+		/* First, ensure that our prior write to TRFCR has stuck. */
+		isb();
+
+		/* Now synchronise with the trace and drain the buffer. */
+		tsb_csync();
+		dsb(nsh);
+
+		/*
+		 * With no more trace being generated, we can disable the
+		 * Trace Buffer Unit.
+		 */
+		write_sysreg_s(0, SYS_TRBLIMITR_EL1);
+		if (cpus_have_final_cap(ARM64_WORKAROUND_2064142)) {
+			/*
+			 * Some CPUs are so good, we have to drain 'em
+			 * twice.
+			 */
+			tsb_csync();
+			dsb(nsh);
+		}
+
+		/*
+		 * Ensure that the Trace Buffer Unit is disabled before
+		 * we start mucking with the stage-2 and trap
+		 * configuration.
+		 */
+		isb();
+	}
+}
+
+static bool __trace_needs_switch(void)
+{
+	return host_data_test_flag(TRBE_ENABLED) ||
+	       host_data_test_flag(EL1_TRACING_CONFIGURED);
+}
+
+static void __trace_switch_to_guest(void)
+{
+	/* Unsupported with TRBE so disable */
+	if (host_data_test_flag(TRBE_ENABLED))
+		*host_data_ptr(trfcr_while_in_guest) = 0;
+
+	__trace_do_switch(host_data_ptr(host_debug_state.trfcr_el1),
+			  *host_data_ptr(trfcr_while_in_guest));
+	__trace_drain_and_disable();
+}
+
+static void __trace_switch_to_host(void)
+{
+	u64 trblimitr_el1 = *host_data_ptr(host_debug_state.trblimitr_el1);
+
+	if (trblimitr_el1 & TRBLIMITR_EL1_E) {
+		/* Re-enable the Trace Buffer Unit for the host. */
+		write_sysreg_s(trblimitr_el1, SYS_TRBLIMITR_EL1);
+		isb();
+		if (cpus_have_final_cap(ARM64_WORKAROUND_2038923)) {
+			/*
+			 * Make sure the unit is re-enabled before we
+			 * poke TRFCR.
+			 */
+			isb();
+		}
+	}
+
+	__trace_do_switch(host_data_ptr(trfcr_while_in_guest),
+			  *host_data_ptr(host_debug_state.trfcr_el1));
+}
+
+static void __debug_save_brbe(void)
+{
+	u64 *brbcr_el1 = host_data_ptr(host_debug_state.brbcr_el1);
+
+	*brbcr_el1 = 0;
+
+	/* Check if the BRBE is enabled */
+	if (!(read_sysreg_el1(SYS_BRBCR) & (BRBCR_ELx_E0BRE | BRBCR_ELx_ExBRE)))
+		return;
+
+	/*
+	 * Prohibit branch record generation while we are in guest.
+	 * Since access to BRBCR_EL1 is trapped, the guest can't
+	 * modify the filtering set by the host.
+	 */
+	*brbcr_el1 = read_sysreg_el1(SYS_BRBCR);
+	write_sysreg_el1(0, SYS_BRBCR);
+}
+
+static void __debug_restore_brbe(void)
+{
+	u64 brbcr_el1 = *host_data_ptr(host_debug_state.brbcr_el1);
+
+	if (!brbcr_el1)
+		return;
+
+	/* Restore BRBE controls */
+	write_sysreg_el1(brbcr_el1, SYS_BRBCR);
+}
+
+void __debug_save_host_buffers_nvhe(struct kvm_vcpu *vcpu)
+{
+	/* Disable and flush SPE data generation */
+	if (host_data_test_flag(HAS_SPE))
+		__debug_save_spe();
+
+	/* Disable BRBE branch records */
+	if (host_data_test_flag(HAS_BRBE))
+		__debug_save_brbe();
+
+	if (__trace_needs_switch())
+		__trace_switch_to_guest();
+}
+
+void __debug_switch_to_guest(struct kvm_vcpu *vcpu)
+{
+	__debug_switch_to_guest_common(vcpu);
+}
+
+void __debug_restore_host_buffers_nvhe(struct kvm_vcpu *vcpu)
+{
+	if (host_data_test_flag(HAS_SPE))
+		__debug_restore_spe();
+	if (host_data_test_flag(HAS_BRBE))
+		__debug_restore_brbe();
+	if (__trace_needs_switch())
+		__trace_switch_to_host();
+}
+
+void __debug_switch_to_host(struct kvm_vcpu *vcpu)
+{
+	__debug_switch_to_host_common(vcpu);
+}

linux-7.2.3/arch/arm64/kvm/hyp/nvhe/early_alloc.c

@@ -0,0 +1,59 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2020 Google LLC
+ * Author: Quentin Perret <qperret@google.com>
+ */
+
+#include <asm/kvm_pgtable.h>
+
+#include <nvhe/early_alloc.h>
+#include <nvhe/memory.h>
+
+struct kvm_pgtable_mm_ops hyp_early_alloc_mm_ops;
+s64 __ro_after_init hyp_physvirt_offset;
+
+static unsigned long base;
+static unsigned long end;
+static unsigned long cur;
+
+unsigned long hyp_early_alloc_nr_used_pages(void)
+{
+	return (cur - base) >> PAGE_SHIFT;
+}
+
+void *hyp_early_alloc_contig(unsigned int nr_pages)
+{
+	unsigned long size = (nr_pages << PAGE_SHIFT);
+	void *ret = (void *)cur;
+
+	if (!nr_pages)
+		return NULL;
+
+	if (end - cur < size)
+		return NULL;
+
+	cur += size;
+	memset(ret, 0, size);
+
+	return ret;
+}
+
+void *hyp_early_alloc_page(void *arg)
+{
+	return hyp_early_alloc_contig(1);
+}
+
+static void hyp_early_alloc_get_page(void *addr) { }
+static void hyp_early_alloc_put_page(void *addr) { }
+
+void hyp_early_alloc_init(void *virt, unsigned long size)
+{
+	base = cur = (unsigned long)virt;
+	end = base + size;
+
+	hyp_early_alloc_mm_ops.zalloc_page = hyp_early_alloc_page;
+	hyp_early_alloc_mm_ops.phys_to_virt = hyp_phys_to_virt;
+	hyp_early_alloc_mm_ops.virt_to_phys = hyp_virt_to_phys;
+	hyp_early_alloc_mm_ops.get_page = hyp_early_alloc_get_page;
+	hyp_early_alloc_mm_ops.put_page = hyp_early_alloc_put_page;
+}

linux-7.2.3/arch/arm64/kvm/hyp/nvhe/events.c

@@ -0,0 +1,25 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2025 Google LLC
+ * Author: Vincent Donnefort <vdonnefort@google.com>
+ */
+
+#include <nvhe/mm.h>
+#include <nvhe/trace.h>
+
+#include <nvhe/define_events.h>
+
+int __tracing_enable_event(unsigned short id, bool enable)
+{
+	struct hyp_event_id *event_id = &__hyp_event_ids_start[id];
+	atomic_t *enabled;
+
+	if (event_id >= __hyp_event_ids_end)
+		return -EINVAL;
+
+	enabled = hyp_fixmap_map(__hyp_pa(&event_id->enabled));
+	atomic_set(enabled, enable);
+	hyp_fixmap_unmap();
+
+	return 0;
+}

linux-7.2.3/arch/arm64/kvm/hyp/nvhe/gen-hyprel.c

@@ -0,0 +1,462 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2020 - Google LLC
+ * Author: David Brazdil <dbrazdil@google.com>
+ *
+ * Generates relocation information used by the kernel to convert
+ * absolute addresses in hyp data from kernel VAs to hyp VAs.
+ *
+ * This is necessary because hyp code is linked into the same binary
+ * as the kernel but executes under different memory mappings.
+ * If the compiler used absolute addressing, those addresses need to
+ * be converted before they are used by hyp code.
+ *
+ * The input of this program is the relocatable ELF object containing
+ * all hyp code/data, not yet linked into vmlinux. Hyp section names
+ * should have been prefixed with `.hyp` at this point.
+ *
+ * The output (printed to stdout) is an assembly file containing
+ * an array of 32-bit integers and static relocations that instruct
+ * the linker of `vmlinux` to populate the array entries with offsets
+ * to positions in the kernel binary containing VAs used by hyp code.
+ *
+ * Note that dynamic relocations could be used for the same purpose.
+ * However, those are only generated if CONFIG_RELOCATABLE=y.
+ */
+
+#include <elf.h>
+#include <endian.h>
+#include <errno.h>
+#include <fcntl.h>
+#include <stdbool.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/mman.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <unistd.h>
+
+#include <generated/autoconf.h>
+
+#define HYP_SECTION_PREFIX		".hyp"
+#define HYP_RELOC_SECTION		".hyp.reloc"
+#define HYP_SECTION_SYMBOL_PREFIX	"__hyp_section_"
+
+/*
+ * AArch64 relocation type constants.
+ * Included in case these are not defined in the host toolchain.
+ */
+#ifndef R_AARCH64_ABS64
+#define R_AARCH64_ABS64			257
+#endif
+#ifndef R_AARCH64_ABS32
+#define R_AARCH64_ABS32			258
+#endif
+#ifndef R_AARCH64_PREL64
+#define R_AARCH64_PREL64		260
+#endif
+#ifndef R_AARCH64_PREL32
+#define R_AARCH64_PREL32		261
+#endif
+#ifndef R_AARCH64_PREL16
+#define R_AARCH64_PREL16		262
+#endif
+#ifndef R_AARCH64_PLT32
+#define R_AARCH64_PLT32			314
+#endif
+#ifndef R_AARCH64_LD_PREL_LO19
+#define R_AARCH64_LD_PREL_LO19		273
+#endif
+#ifndef R_AARCH64_ADR_PREL_LO21
+#define R_AARCH64_ADR_PREL_LO21		274
+#endif
+#ifndef R_AARCH64_ADR_PREL_PG_HI21
+#define R_AARCH64_ADR_PREL_PG_HI21	275
+#endif
+#ifndef R_AARCH64_ADR_PREL_PG_HI21_NC
+#define R_AARCH64_ADR_PREL_PG_HI21_NC	276
+#endif
+#ifndef R_AARCH64_ADD_ABS_LO12_NC
+#define R_AARCH64_ADD_ABS_LO12_NC	277
+#endif
+#ifndef R_AARCH64_LDST8_ABS_LO12_NC
+#define R_AARCH64_LDST8_ABS_LO12_NC	278
+#endif
+#ifndef R_AARCH64_TSTBR14
+#define R_AARCH64_TSTBR14		279
+#endif
+#ifndef R_AARCH64_CONDBR19
+#define R_AARCH64_CONDBR19		280
+#endif
+#ifndef R_AARCH64_JUMP26
+#define R_AARCH64_JUMP26		282
+#endif
+#ifndef R_AARCH64_CALL26
+#define R_AARCH64_CALL26		283
+#endif
+#ifndef R_AARCH64_LDST16_ABS_LO12_NC
+#define R_AARCH64_LDST16_ABS_LO12_NC	284
+#endif
+#ifndef R_AARCH64_LDST32_ABS_LO12_NC
+#define R_AARCH64_LDST32_ABS_LO12_NC	285
+#endif
+#ifndef R_AARCH64_LDST64_ABS_LO12_NC
+#define R_AARCH64_LDST64_ABS_LO12_NC	286
+#endif
+#ifndef R_AARCH64_MOVW_PREL_G0
+#define R_AARCH64_MOVW_PREL_G0		287
+#endif
+#ifndef R_AARCH64_MOVW_PREL_G0_NC
+#define R_AARCH64_MOVW_PREL_G0_NC	288
+#endif
+#ifndef R_AARCH64_MOVW_PREL_G1
+#define R_AARCH64_MOVW_PREL_G1		289
+#endif
+#ifndef R_AARCH64_MOVW_PREL_G1_NC
+#define R_AARCH64_MOVW_PREL_G1_NC	290
+#endif
+#ifndef R_AARCH64_MOVW_PREL_G2
+#define R_AARCH64_MOVW_PREL_G2		291
+#endif
+#ifndef R_AARCH64_MOVW_PREL_G2_NC
+#define R_AARCH64_MOVW_PREL_G2_NC	292
+#endif
+#ifndef R_AARCH64_MOVW_PREL_G3
+#define R_AARCH64_MOVW_PREL_G3		293
+#endif
+#ifndef R_AARCH64_LDST128_ABS_LO12_NC
+#define R_AARCH64_LDST128_ABS_LO12_NC	299
+#endif
+
+/* Global state of the processed ELF. */
+static struct {
+	const char	*path;
+	char		*begin;
+	size_t		size;
+	Elf64_Ehdr	*ehdr;
+	Elf64_Shdr	*sh_table;
+	const char	*sh_string;
+} elf;
+
+#if defined(CONFIG_CPU_LITTLE_ENDIAN)
+
+#define elf16toh(x)	le16toh(x)
+#define elf32toh(x)	le32toh(x)
+#define elf64toh(x)	le64toh(x)
+
+#define ELFENDIAN	ELFDATA2LSB
+
+#elif defined(CONFIG_CPU_BIG_ENDIAN)
+
+#define elf16toh(x)	be16toh(x)
+#define elf32toh(x)	be32toh(x)
+#define elf64toh(x)	be64toh(x)
+
+#define ELFENDIAN	ELFDATA2MSB
+
+#else
+
+#error PDP-endian sadly unsupported...
+
+#endif
+
+#define fatal_error(fmt, ...)						\
+	({								\
+		fprintf(stderr, "error: %s: " fmt "\n",			\
+			elf.path, ## __VA_ARGS__);			\
+		exit(EXIT_FAILURE);					\
+		__builtin_unreachable();				\
+	})
+
+#define fatal_perror(msg)						\
+	({								\
+		fprintf(stderr, "error: %s: " msg ": %s\n",		\
+			elf.path, strerror(errno));			\
+		exit(EXIT_FAILURE);					\
+		__builtin_unreachable();				\
+	})
+
+#define assert_op(lhs, rhs, fmt, op)					\
+	({								\
+		typeof(lhs) _lhs = (lhs);				\
+		typeof(rhs) _rhs = (rhs);				\
+									\
+		if (!(_lhs op _rhs)) {					\
+			fatal_error("assertion " #lhs " " #op " " #rhs	\
+				" failed (lhs=" fmt ", rhs=" fmt	\
+				", line=%d)", _lhs, _rhs, __LINE__);	\
+		}							\
+	})
+
+#define assert_eq(lhs, rhs, fmt)	assert_op(lhs, rhs, fmt, ==)
+#define assert_ne(lhs, rhs, fmt)	assert_op(lhs, rhs, fmt, !=)
+#define assert_lt(lhs, rhs, fmt)	assert_op(lhs, rhs, fmt, <)
+#define assert_ge(lhs, rhs, fmt)	assert_op(lhs, rhs, fmt, >=)
+
+/*
+ * Return a pointer of a given type at a given offset from
+ * the beginning of the ELF file.
+ */
+#define elf_ptr(type, off) ((type *)(elf.begin + (off)))
+
+/* Iterate over all sections in the ELF. */
+#define for_each_section(var) \
+	for (var = elf.sh_table; var < elf.sh_table + elf16toh(elf.ehdr->e_shnum); ++var)
+
+/* Iterate over all Elf64_Rela relocations in a given section. */
+#define for_each_rela(shdr, var)					\
+	for (var = elf_ptr(Elf64_Rela, elf64toh(shdr->sh_offset));	\
+	     var < elf_ptr(Elf64_Rela, elf64toh(shdr->sh_offset) + elf64toh(shdr->sh_size)); var++)
+
+/* True if a string starts with a given prefix. */
+static inline bool starts_with(const char *str, const char *prefix)
+{
+	return memcmp(str, prefix, strlen(prefix)) == 0;
+}
+
+/* Returns a string containing the name of a given section. */
+static inline const char *section_name(Elf64_Shdr *shdr)
+{
+	return elf.sh_string + elf32toh(shdr->sh_name);
+}
+
+/* Returns a pointer to the first byte of section data. */
+static inline const char *section_begin(Elf64_Shdr *shdr)
+{
+	return elf_ptr(char, elf64toh(shdr->sh_offset));
+}
+
+/* Find a section by its offset from the beginning of the file. */
+static inline Elf64_Shdr *section_by_off(Elf64_Off off)
+{
+	assert_ne(off, 0UL, "%lu");
+	return elf_ptr(Elf64_Shdr, off);
+}
+
+/* Find a section by its index. */
+static inline Elf64_Shdr *section_by_idx(uint16_t idx)
+{
+	assert_ne(idx, SHN_UNDEF, "%u");
+	return &elf.sh_table[idx];
+}
+
+/*
+ * Memory-map the given ELF file, perform sanity checks, and
+ * populate global state.
+ */
+static void init_elf(const char *path)
+{
+	int fd, ret;
+	struct stat stat;
+
+	/* Store path in the global struct for error printing. */
+	elf.path = path;
+
+	/* Open the ELF file. */
+	fd = open(path, O_RDONLY);
+	if (fd < 0)
+		fatal_perror("Could not open ELF file");
+
+	/* Get status of ELF file to obtain its size. */
+	ret = fstat(fd, &stat);
+	if (ret < 0) {
+		close(fd);
+		fatal_perror("Could not get status of ELF file");
+	}
+
+	/* mmap() the entire ELF file read-only at an arbitrary address. */
+	elf.begin = mmap(0, stat.st_size, PROT_READ, MAP_PRIVATE, fd, 0);
+	if (elf.begin == MAP_FAILED) {
+		close(fd);
+		fatal_perror("Could not mmap ELF file");
+	}
+
+	/* mmap() was successful, close the FD. */
+	close(fd);
+
+	/* Get pointer to the ELF header. */
+	assert_ge(stat.st_size, sizeof(*elf.ehdr), "%lu");
+	elf.ehdr = elf_ptr(Elf64_Ehdr, 0);
+
+	/* Check the ELF magic. */
+	assert_eq(elf.ehdr->e_ident[EI_MAG0], ELFMAG0, "0x%x");
+	assert_eq(elf.ehdr->e_ident[EI_MAG1], ELFMAG1, "0x%x");
+	assert_eq(elf.ehdr->e_ident[EI_MAG2], ELFMAG2, "0x%x");
+	assert_eq(elf.ehdr->e_ident[EI_MAG3], ELFMAG3, "0x%x");
+
+	/* Sanity check that this is an ELF64 relocatable object for AArch64. */
+	assert_eq(elf.ehdr->e_ident[EI_CLASS], ELFCLASS64, "%u");
+	assert_eq(elf.ehdr->e_ident[EI_DATA], ELFENDIAN, "%u");
+	assert_eq(elf16toh(elf.ehdr->e_type), ET_REL, "%u");
+	assert_eq(elf16toh(elf.ehdr->e_machine), EM_AARCH64, "%u");
+
+	/* Populate fields of the global struct. */
+	elf.sh_table = section_by_off(elf64toh(elf.ehdr->e_shoff));
+	elf.sh_string = section_begin(section_by_idx(elf16toh(elf.ehdr->e_shstrndx)));
+}
+
+/* Print the prologue of the output ASM file. */
+static void emit_prologue(void)
+{
+	printf(".data\n"
+	       ".pushsection " HYP_RELOC_SECTION ", \"a\"\n");
+}
+
+/* Print ASM statements needed as a prologue to a processed hyp section. */
+static void emit_section_prologue(const char *sh_orig_name)
+{
+	/* Declare the hyp section symbol. */
+	printf(".global %s%s\n", HYP_SECTION_SYMBOL_PREFIX, sh_orig_name);
+}
+
+/*
+ * Print ASM statements to create a hyp relocation entry for a given
+ * R_AARCH64_ABS64 relocation.
+ *
+ * The linker of vmlinux will populate the position given by `rela` with
+ * an absolute 64-bit kernel VA. If the kernel is relocatable, it will
+ * also generate a dynamic relocation entry so that the kernel can shift
+ * the address at runtime for KASLR.
+ *
+ * Emit a 32-bit offset from the current address to the position given
+ * by `rela`. This way the kernel can iterate over all kernel VAs used
+ * by hyp at runtime and convert them to hyp VAs. However, that offset
+ * will not be known until linking of `vmlinux`, so emit a PREL32
+ * relocation referencing a symbol that the hyp linker script put at
+ * the beginning of the relocated section + the offset from `rela`.
+ */
+static void emit_rela_abs64(Elf64_Rela *rela, const char *sh_orig_name)
+{
+	/* Offset of this reloc from the beginning of HYP_RELOC_SECTION. */
+	static size_t reloc_offset;
+
+	/* Create storage for the 32-bit offset. */
+	printf(".word 0\n");
+
+	/*
+	 * Create a PREL32 relocation which instructs the linker of `vmlinux`
+	 * to insert offset to position <base> + <offset>, where <base> is
+	 * a symbol at the beginning of the relocated section, and <offset>
+	 * is `rela->r_offset`.
+	 */
+	printf(".reloc %lu, R_AARCH64_PREL32, %s%s + 0x%lx\n",
+	       reloc_offset, HYP_SECTION_SYMBOL_PREFIX, sh_orig_name,
+	       elf64toh(rela->r_offset));
+
+	reloc_offset += 4;
+}
+
+/* Print the epilogue of the output ASM file. */
+static void emit_epilogue(void)
+{
+	printf(".popsection\n");
+}
+
+/*
+ * Iterate over all RELA relocations in a given section and emit
+ * hyp relocation data for all absolute addresses in hyp code/data.
+ *
+ * Static relocations that generate PC-relative-addressing are ignored.
+ * Failure is reported for unexpected relocation types.
+ */
+static void emit_rela_section(Elf64_Shdr *sh_rela)
+{
+	Elf64_Shdr *sh_orig = &elf.sh_table[elf32toh(sh_rela->sh_info)];
+	const char *sh_orig_name = section_name(sh_orig);
+	Elf64_Rela *rela;
+
+	/* Skip all non-hyp sections. */
+	if (!starts_with(sh_orig_name, HYP_SECTION_PREFIX))
+		return;
+
+	emit_section_prologue(sh_orig_name);
+
+	for_each_rela(sh_rela, rela) {
+		uint32_t type = (uint32_t)elf64toh(rela->r_info);
+
+		/* Check that rela points inside the relocated section. */
+		assert_lt(elf64toh(rela->r_offset), elf64toh(sh_orig->sh_size), "0x%lx");
+
+		switch (type) {
+		/*
+		 * Data relocations to generate absolute addressing.
+		 * Emit a hyp relocation.
+		 */
+		case R_AARCH64_ABS64:
+			emit_rela_abs64(rela, sh_orig_name);
+			break;
+		/* Allow 32-bit absolute relocation, for kCFI type hashes. */
+		case R_AARCH64_ABS32:
+			break;
+		/* Allow position-relative data relocations. */
+		case R_AARCH64_PREL64:
+		case R_AARCH64_PREL32:
+		case R_AARCH64_PREL16:
+		case R_AARCH64_PLT32:
+			break;
+		/* Allow relocations to generate PC-relative addressing. */
+		case R_AARCH64_LD_PREL_LO19:
+		case R_AARCH64_ADR_PREL_LO21:
+		case R_AARCH64_ADR_PREL_PG_HI21:
+		case R_AARCH64_ADR_PREL_PG_HI21_NC:
+		case R_AARCH64_ADD_ABS_LO12_NC:
+		case R_AARCH64_LDST8_ABS_LO12_NC:
+		case R_AARCH64_LDST16_ABS_LO12_NC:
+		case R_AARCH64_LDST32_ABS_LO12_NC:
+		case R_AARCH64_LDST64_ABS_LO12_NC:
+		case R_AARCH64_LDST128_ABS_LO12_NC:
+			break;
+		/* Allow relative relocations for control-flow instructions. */
+		case R_AARCH64_TSTBR14:
+		case R_AARCH64_CONDBR19:
+		case R_AARCH64_JUMP26:
+		case R_AARCH64_CALL26:
+			break;
+		/* Allow group relocations to create PC-relative offset inline. */
+		case R_AARCH64_MOVW_PREL_G0:
+		case R_AARCH64_MOVW_PREL_G0_NC:
+		case R_AARCH64_MOVW_PREL_G1:
+		case R_AARCH64_MOVW_PREL_G1_NC:
+		case R_AARCH64_MOVW_PREL_G2:
+		case R_AARCH64_MOVW_PREL_G2_NC:
+		case R_AARCH64_MOVW_PREL_G3:
+			break;
+		default:
+			fatal_error("Unexpected RELA type %u", type);
+		}
+	}
+}
+
+/* Iterate over all sections and emit hyp relocation data for RELA sections. */
+static void emit_all_relocs(void)
+{
+	Elf64_Shdr *shdr;
+
+	for_each_section(shdr) {
+		switch (elf32toh(shdr->sh_type)) {
+		case SHT_REL:
+			fatal_error("Unexpected SHT_REL section \"%s\"",
+				section_name(shdr));
+		case SHT_RELA:
+			emit_rela_section(shdr);
+			break;
+		}
+	}
+}
+
+int main(int argc, const char **argv)
+{
+	if (argc != 2) {
+		fprintf(stderr, "Usage: %s <elf_input>\n", argv[0]);
+		return EXIT_FAILURE;
+	}
+
+	init_elf(argv[1]);
+
+	emit_prologue();
+	emit_all_relocs();
+	emit_epilogue();
+
+	return EXIT_SUCCESS;
+}

linux-7.2.3/arch/arm64/kvm/hyp/nvhe/host.S

@@ -0,0 +1,292 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (C) 2020 - Google Inc
+ * Author: Andrew Scull <ascull@google.com>
+ */
+
+#include <linux/linkage.h>
+
+#include <asm/assembler.h>
+#include <asm/kvm_arm.h>
+#include <asm/kvm_asm.h>
+#include <asm/kvm_mmu.h>
+#include <asm/kvm_ptrauth.h>
+
+	.text
+
+SYM_FUNC_START(__host_exit)
+	get_host_ctxt	x0, x1
+
+	/* Store the host regs x2 and x3 */
+	stp	x2, x3,   [x0, #CPU_XREG_OFFSET(2)]
+
+	/* Retrieve the host regs x0-x1 from the stack */
+	ldp	x2, x3, [sp], #16	// x0, x1
+
+	/* Store the host regs x0-x1 and x4-x17 */
+	stp	x2, x3,   [x0, #CPU_XREG_OFFSET(0)]
+	stp	x4, x5,   [x0, #CPU_XREG_OFFSET(4)]
+	stp	x6, x7,   [x0, #CPU_XREG_OFFSET(6)]
+	stp	x8, x9,   [x0, #CPU_XREG_OFFSET(8)]
+	stp	x10, x11, [x0, #CPU_XREG_OFFSET(10)]
+	stp	x12, x13, [x0, #CPU_XREG_OFFSET(12)]
+	stp	x14, x15, [x0, #CPU_XREG_OFFSET(14)]
+	stp	x16, x17, [x0, #CPU_XREG_OFFSET(16)]
+
+	/* Store the host regs x18-x29, lr */
+	save_callee_saved_regs x0
+
+	/* Save the host context pointer in x29 across the function call */
+	mov	x29, x0
+
+#ifdef CONFIG_ARM64_PTR_AUTH_KERNEL
+alternative_if_not ARM64_HAS_ADDRESS_AUTH
+b __skip_pauth_save
+alternative_else_nop_endif
+
+alternative_if ARM64_KVM_PROTECTED_MODE
+	/* Save kernel ptrauth keys. */
+	add x18, x29, #CPU_APIAKEYLO_EL1
+	ptrauth_save_state x18, x19, x20
+
+	/* Use hyp keys. */
+	adr_this_cpu x18, kvm_hyp_ctxt, x19
+	add x18, x18, #CPU_APIAKEYLO_EL1
+	ptrauth_restore_state x18, x19, x20
+	isb
+alternative_else_nop_endif
+__skip_pauth_save:
+#endif /* CONFIG_ARM64_PTR_AUTH_KERNEL */
+
+	bl	handle_trap
+
+__host_enter_restore_full:
+	/* Restore kernel keys. */
+#ifdef CONFIG_ARM64_PTR_AUTH_KERNEL
+alternative_if_not ARM64_HAS_ADDRESS_AUTH
+b __skip_pauth_restore
+alternative_else_nop_endif
+
+alternative_if ARM64_KVM_PROTECTED_MODE
+	add x18, x29, #CPU_APIAKEYLO_EL1
+	ptrauth_restore_state x18, x19, x20
+alternative_else_nop_endif
+__skip_pauth_restore:
+#endif /* CONFIG_ARM64_PTR_AUTH_KERNEL */
+
+	/* Restore host regs x0-x17 */
+	ldp	x0, x1,   [x29, #CPU_XREG_OFFSET(0)]
+	ldp	x2, x3,   [x29, #CPU_XREG_OFFSET(2)]
+	ldp	x4, x5,   [x29, #CPU_XREG_OFFSET(4)]
+	ldp	x6, x7,   [x29, #CPU_XREG_OFFSET(6)]
+
+	/* x0-7 are use for panic arguments */
+__host_enter_for_panic:
+	ldp	x8, x9,   [x29, #CPU_XREG_OFFSET(8)]
+	ldp	x10, x11, [x29, #CPU_XREG_OFFSET(10)]
+	ldp	x12, x13, [x29, #CPU_XREG_OFFSET(12)]
+	ldp	x14, x15, [x29, #CPU_XREG_OFFSET(14)]
+	ldp	x16, x17, [x29, #CPU_XREG_OFFSET(16)]
+
+	/* Restore host regs x18-x29, lr */
+	restore_callee_saved_regs x29
+
+	/* Do not touch any register after this! */
+__host_enter_without_restoring:
+	eret
+	sb
+SYM_FUNC_END(__host_exit)
+
+/*
+ * void __noreturn __host_enter(struct kvm_cpu_context *host_ctxt);
+ */
+SYM_FUNC_START(__host_enter)
+	mov	x29, x0
+	b	__host_enter_restore_full
+SYM_FUNC_END(__host_enter)
+
+/*
+ * void __noreturn __hyp_do_panic(struct kvm_cpu_context *host_ctxt, u64 spsr,
+ * 				  u64 elr, u64 par);
+ */
+SYM_FUNC_START(__hyp_do_panic)
+	/* Prepare and exit to the host's panic function. */
+	mov	lr, #(PSR_F_BIT | PSR_I_BIT | PSR_A_BIT | PSR_D_BIT |\
+		      PSR_MODE_EL1h)
+	msr	spsr_el2, lr
+	adr_l	lr, nvhe_hyp_panic_handler
+	hyp_kimg_va lr, x6
+	msr	elr_el2, lr
+
+	mov	x29, x0
+
+#ifdef CONFIG_PKVM_DISABLE_STAGE2_ON_PANIC
+	/* Ensure host stage-2 is disabled */
+	mrs	x0, hcr_el2
+	bic	x0, x0, #HCR_VM
+	msr_hcr_el2 x0
+	tlbi	vmalls12e1
+	dsb	nsh
+#endif
+
+	/* Load the panic arguments into x0-7 */
+	mrs	x0, esr_el2
+	mov	x4, x3
+	mov	x3, x2
+	hyp_pa	x3, x6
+	get_vcpu_ptr x5, x6
+	mrs	x6, far_el2
+	mrs	x7, hpfar_el2
+
+	/* Enter the host, conditionally restoring the host context. */
+	cbz	x29, __host_enter_without_restoring
+	b	__host_enter_for_panic
+SYM_FUNC_END(__hyp_do_panic)
+
+SYM_FUNC_START(__host_hvc)
+	ldp	x0, x1, [sp]		// Don't fixup the stack yet
+
+	/* No stub for you, sonny Jim */
+alternative_if ARM64_KVM_PROTECTED_MODE
+	b	__host_exit
+alternative_else_nop_endif
+
+	/* Check for a stub HVC call */
+	cmp	x0, #HVC_STUB_HCALL_NR
+	b.hs	__host_exit
+
+	add	sp, sp, #16
+	/*
+	 * Compute the idmap address of __kvm_handle_stub_hvc and
+	 * jump there.
+	 *
+	 * Preserve x0-x4, which may contain stub parameters.
+	 */
+	adr_l	x5, __kvm_handle_stub_hvc
+	hyp_pa	x5, x6
+	br	x5
+SYM_FUNC_END(__host_hvc)
+
+.macro host_el1_sync_vect
+	.align 7
+.L__vect_start\@:
+	stp	x0, x1, [sp, #-16]!
+	mrs	x0, esr_el2
+	ubfx	x0, x0, #ESR_ELx_EC_SHIFT, #ESR_ELx_EC_WIDTH
+	cmp	x0, #ESR_ELx_EC_HVC64
+	b.eq	__host_hvc
+	b	__host_exit
+.L__vect_end\@:
+.if ((.L__vect_end\@ - .L__vect_start\@) > 0x80)
+	.error "host_el1_sync_vect larger than vector entry"
+.endif
+.endm
+
+.macro invalid_host_el2_vect
+	.align 7
+
+	/*
+	 * Test whether the SP has overflowed, without corrupting a GPR.
+	 * nVHE hypervisor stacks are aligned so that the NVHE_STACK_SHIFT bit
+	 * of SP should always be 1.
+	 */
+	add	sp, sp, x0			// sp' = sp + x0
+	sub	x0, sp, x0			// x0' = sp' - x0 = (sp + x0) - x0 = sp
+	tbz	x0, #NVHE_STACK_SHIFT, .L__hyp_sp_overflow\@
+	sub	x0, sp, x0			// x0'' = sp' - x0' = (sp + x0) - sp = x0
+	sub	sp, sp, x0			// sp'' = sp' - x0 = (sp + x0) - x0 = sp
+
+	/*
+	 * The panic may not be clean if the exception is taken before the host
+	 * context has been saved by __host_exit or after the hyp context has
+	 * been partially clobbered by __host_enter.
+	 */
+	b	hyp_panic
+
+.L__hyp_sp_overflow\@:
+	/* Switch to the overflow stack */
+	adr_this_cpu sp, overflow_stack + OVERFLOW_STACK_SIZE, x0
+
+	b	hyp_panic_bad_stack
+	ASM_BUG()
+.endm
+
+.macro invalid_host_el1_vect
+	.align 7
+	mov	x0, xzr		/* restore_host = false */
+	mrs	x1, spsr_el2
+	mrs	x2, elr_el2
+	mrs	x3, par_el1
+	b	__hyp_do_panic
+.endm
+
+/*
+ * The host vector does not use an ESB instruction in order to avoid consuming
+ * SErrors that should only be consumed by the host. Guest entry is deferred by
+ * __guest_enter if there are any pending asynchronous exceptions so hyp will
+ * always return to the host without having consumerd host SErrors.
+ *
+ * CONFIG_KVM_INDIRECT_VECTORS is not applied to the host vectors because the
+ * host knows about the EL2 vectors already, and there is no point in hiding
+ * them.
+ */
+	.align 11
+SYM_CODE_START(__kvm_hyp_host_vector)
+	invalid_host_el2_vect			// Synchronous EL2t
+	invalid_host_el2_vect			// IRQ EL2t
+	invalid_host_el2_vect			// FIQ EL2t
+	invalid_host_el2_vect			// Error EL2t
+
+	invalid_host_el2_vect			// Synchronous EL2h
+	invalid_host_el2_vect			// IRQ EL2h
+	invalid_host_el2_vect			// FIQ EL2h
+	invalid_host_el2_vect			// Error EL2h
+
+	host_el1_sync_vect			// Synchronous 64-bit EL1/EL0
+	invalid_host_el1_vect			// IRQ 64-bit EL1/EL0
+	invalid_host_el1_vect			// FIQ 64-bit EL1/EL0
+	invalid_host_el1_vect			// Error 64-bit EL1/EL0
+
+	host_el1_sync_vect			// Synchronous 32-bit EL1/EL0
+	invalid_host_el1_vect			// IRQ 32-bit EL1/EL0
+	invalid_host_el1_vect			// FIQ 32-bit EL1/EL0
+	invalid_host_el1_vect			// Error 32-bit EL1/EL0
+SYM_CODE_END(__kvm_hyp_host_vector)
+
+/*
+ * Forward SMC with arguments in struct kvm_cpu_context, and
+ * store the result into the same struct. Assumes SMCCC 1.2 or older.
+ *
+ * x0: struct kvm_cpu_context*
+ */
+SYM_CODE_START(__kvm_hyp_host_forward_smc)
+	/*
+	 * Use x18 to keep the pointer to the host context because
+	 * x18 is callee-saved in SMCCC but not in AAPCS64.
+	 */
+	mov	x18, x0
+
+	ldp	x0, x1,   [x18, #CPU_XREG_OFFSET(0)]
+	ldp	x2, x3,   [x18, #CPU_XREG_OFFSET(2)]
+	ldp	x4, x5,   [x18, #CPU_XREG_OFFSET(4)]
+	ldp	x6, x7,   [x18, #CPU_XREG_OFFSET(6)]
+	ldp	x8, x9,   [x18, #CPU_XREG_OFFSET(8)]
+	ldp	x10, x11, [x18, #CPU_XREG_OFFSET(10)]
+	ldp	x12, x13, [x18, #CPU_XREG_OFFSET(12)]
+	ldp	x14, x15, [x18, #CPU_XREG_OFFSET(14)]
+	ldp	x16, x17, [x18, #CPU_XREG_OFFSET(16)]
+
+	smc	#0
+
+	stp	x0, x1,   [x18, #CPU_XREG_OFFSET(0)]
+	stp	x2, x3,   [x18, #CPU_XREG_OFFSET(2)]
+	stp	x4, x5,   [x18, #CPU_XREG_OFFSET(4)]
+	stp	x6, x7,   [x18, #CPU_XREG_OFFSET(6)]
+	stp	x8, x9,   [x18, #CPU_XREG_OFFSET(8)]
+	stp	x10, x11, [x18, #CPU_XREG_OFFSET(10)]
+	stp	x12, x13, [x18, #CPU_XREG_OFFSET(12)]
+	stp	x14, x15, [x18, #CPU_XREG_OFFSET(14)]
+	stp	x16, x17, [x18, #CPU_XREG_OFFSET(16)]
+
+	ret
+SYM_CODE_END(__kvm_hyp_host_forward_smc)

linux-7.2.3/arch/arm64/kvm/hyp/nvhe/hyp-init.S

@@ -0,0 +1,297 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (C) 2012,2013 - ARM Ltd
+ * Author: Marc Zyngier <marc.zyngier@arm.com>
+ */
+
+#include <linux/arm-smccc.h>
+#include <linux/cfi_types.h>
+#include <linux/linkage.h>
+
+#include <asm/alternative.h>
+#include <asm/assembler.h>
+#include <asm/el2_setup.h>
+#include <asm/kvm_arm.h>
+#include <asm/kvm_asm.h>
+#include <asm/kvm_mmu.h>
+#include <asm/pgtable-hwdef.h>
+#include <asm/sysreg.h>
+#include <asm/virt.h>
+
+	.text
+	.pushsection	.idmap.text, "ax"
+
+	.align	11
+
+SYM_CODE_START(__kvm_hyp_init)
+	ventry	.			// Synchronous EL2t
+	ventry	.			// IRQ EL2t
+	ventry	.			// FIQ EL2t
+	ventry	.			// Error EL2t
+
+	ventry	.			// Synchronous EL2h
+	ventry	.			// IRQ EL2h
+	ventry	.			// FIQ EL2h
+	ventry	.			// Error EL2h
+
+	ventry	__do_hyp_init		// Synchronous 64-bit EL1
+	ventry	.			// IRQ 64-bit EL1
+	ventry	.			// FIQ 64-bit EL1
+	ventry	.			// Error 64-bit EL1
+
+	ventry	.			// Synchronous 32-bit EL1
+	ventry	.			// IRQ 32-bit EL1
+	ventry	.			// FIQ 32-bit EL1
+	ventry	.			// Error 32-bit EL1
+
+	/*
+	 * Only uses x0..x3 so as to not clobber callee-saved SMCCC registers.
+	 *
+	 * x0: SMCCC function ID
+	 * x1: struct kvm_nvhe_init_params PA
+	 */
+__do_hyp_init:
+	/* Check for a stub HVC call */
+	cmp	x0, #HVC_STUB_HCALL_NR
+	b.lo	__kvm_handle_stub_hvc
+
+	bic	x0, x0, #ARM_SMCCC_CALL_HINTS
+	mov	x3, #KVM_HOST_SMCCC_FUNC(__kvm_hyp_init)
+	cmp	x0, x3
+	b.eq	1f
+
+	mov	x0, #SMCCC_RET_NOT_SUPPORTED
+	eret
+
+1:	mov	x0, x1
+	mov	x3, lr
+	bl	___kvm_hyp_init			// Clobbers x0..x2
+	mov	lr, x3
+
+	/* Hello, World! */
+	mov	x0, #SMCCC_RET_SUCCESS
+	eret
+SYM_CODE_END(__kvm_hyp_init)
+
+/*
+ * Initialize EL2 CPU state to sane values.
+ *
+ * HCR_EL2.E2H must have been initialized already.
+ */
+SYM_CODE_START_LOCAL(__kvm_init_el2_state)
+	init_el2_state				// Clobbers x0..x2
+	finalise_el2_state
+	ret
+SYM_CODE_END(__kvm_init_el2_state)
+
+/*
+ * Initialize the hypervisor in EL2.
+ *
+ * Only uses x0..x2 so as to not clobber callee-saved SMCCC registers
+ * and leave x3 for the caller.
+ *
+ * x0: struct kvm_nvhe_init_params PA
+ */
+SYM_CODE_START_LOCAL(___kvm_hyp_init)
+	ldr	x1, [x0, #NVHE_INIT_STACK_HYP_VA]
+	mov	sp, x1
+
+	ldr	x1, [x0, #NVHE_INIT_MAIR_EL2]
+	msr	mair_el2, x1
+
+	ldr	x1, [x0, #NVHE_INIT_HCR_EL2]
+	msr_hcr_el2 x1
+
+	mov	x2, #HCR_E2H
+	and	x2, x1, x2
+	cbz	x2, 1f
+
+	// hVHE: Replay the EL2 setup to account for the E2H bit
+	// TPIDR_EL2 is used to preserve x0 across the macro maze...
+	isb
+	msr	tpidr_el2, x0
+	str	lr, [x0, #NVHE_INIT_TMP]
+
+	bl	__kvm_init_el2_state
+
+	mrs	x0, tpidr_el2
+	ldr	lr, [x0, #NVHE_INIT_TMP]
+
+1:
+	ldr	x1, [x0, #NVHE_INIT_TPIDR_EL2]
+	msr	tpidr_el2, x1
+
+	ldr	x1, [x0, #NVHE_INIT_VTTBR]
+	msr	vttbr_el2, x1
+
+	ldr	x1, [x0, #NVHE_INIT_VTCR]
+	msr	vtcr_el2, x1
+
+	ldr	x1, [x0, #NVHE_INIT_PGD_PA]
+	phys_to_ttbr x2, x1
+alternative_if ARM64_HAS_CNP
+	orr	x2, x2, #TTBRx_EL1_CnP
+alternative_else_nop_endif
+	msr	ttbr0_el2, x2
+
+	ldr	x0, [x0, #NVHE_INIT_TCR_EL2]
+	msr	tcr_el2, x0
+
+	isb
+
+	/* Invalidate the stale TLBs from Bootloader */
+	tlbi	alle2
+	tlbi	alle1
+	dsb	sy
+
+	mov_q	x0, INIT_SCTLR_EL2_MMU_ON
+alternative_if ARM64_HAS_ADDRESS_AUTH
+	mov_q	x1, (SCTLR_ELx_ENIA | SCTLR_ELx_ENIB | \
+		     SCTLR_ELx_ENDA | SCTLR_ELx_ENDB)
+	orr	x0, x0, x1
+alternative_else_nop_endif
+
+#ifdef CONFIG_ARM64_BTI_KERNEL
+alternative_if ARM64_BTI
+	orr	x0, x0, #SCTLR_EL2_BT
+alternative_else_nop_endif
+#endif /* CONFIG_ARM64_BTI_KERNEL */
+
+	msr	sctlr_el2, x0
+	isb
+
+	/* Set the host vector */
+	ldr	x0, =__kvm_hyp_host_vector
+	msr	vbar_el2, x0
+
+	ret
+SYM_CODE_END(___kvm_hyp_init)
+
+/*
+ * PSCI CPU_ON entry point
+ *
+ * x0: struct kvm_nvhe_init_params PA
+ */
+SYM_CODE_START(kvm_hyp_cpu_entry)
+	ldr	x29, =__kvm_host_psci_cpu_on_entry
+	b	__kvm_hyp_init_cpu
+
+/*
+ * PSCI CPU_SUSPEND / SYSTEM_SUSPEND entry point
+ *
+ * x0: struct kvm_nvhe_init_params PA
+ */
+SYM_CODE_START(kvm_hyp_cpu_resume)
+	ldr	x29, =__kvm_host_psci_cpu_resume_entry
+
+SYM_INNER_LABEL(__kvm_hyp_init_cpu, SYM_L_LOCAL)
+	mov	x28, x0				// Stash arguments
+
+	/* Check that the core was booted in EL2. */
+	mrs	x0, CurrentEL
+	cmp	x0, #CurrentEL_EL2
+	b.ne	1f
+
+	msr	SPsel, #1			// We want to use SP_EL2
+
+	init_el2_hcr	0
+
+	bl	__kvm_init_el2_state
+
+	/* Enable MMU, set vectors and stack. */
+	mov	x0, x28
+	bl	___kvm_hyp_init			// Clobbers x0..x2
+
+	/* Leave idmap -- using BLR is OK, LR is restored from host context */
+	blr	x29
+
+	// The core booted in EL1, or the C code unexpectedly returned.
+	// Either way, KVM cannot be initialized on it.
+1:	wfe
+	wfi
+	b	1b
+SYM_CODE_END(kvm_hyp_cpu_resume)
+SYM_CODE_END(kvm_hyp_cpu_entry)
+
+SYM_CODE_START(__kvm_handle_stub_hvc)
+	/*
+	 * __kvm_handle_stub_hvc called from __host_hvc through branch instruction(br) so
+	 * we need bti j at beginning.
+	 */
+	bti j
+	cmp	x0, #HVC_SOFT_RESTART
+	b.ne	1f
+
+	/* This is where we're about to jump, staying at EL2 */
+	msr	elr_el2, x1
+	mov	x0, #(PSR_F_BIT | PSR_I_BIT | PSR_A_BIT | PSR_D_BIT | PSR_MODE_EL2h)
+	msr	spsr_el2, x0
+
+	/* Shuffle the arguments, and don't come back */
+	mov	x0, x2
+	mov	x1, x3
+	mov	x2, x4
+	b	reset
+
+1:	cmp	x0, #HVC_RESET_VECTORS
+	b.ne	1f
+
+	/*
+	 * Set the HVC_RESET_VECTORS return code before entering the common
+	 * path so that we do not clobber x0-x2 in case we are coming via
+	 * HVC_SOFT_RESTART.
+	 */
+	mov	x0, xzr
+reset:
+	/* Reset kvm back to the hyp stub. */
+	mov_q	x5, INIT_SCTLR_EL2_MMU_OFF
+	pre_disable_mmu_workaround
+	msr	sctlr_el2, x5
+	isb
+
+	/* Install stub vectors */
+	adr_l	x5, __hyp_stub_vectors
+	msr	vbar_el2, x5
+	eret
+
+1:	/* Bad stub call */
+	mov_q	x0, HVC_STUB_ERR
+	eret
+
+SYM_CODE_END(__kvm_handle_stub_hvc)
+
+/*
+ * void __pkvm_init_switch_pgd(phys_addr_t pgd, unsigned long sp,
+ *                             void (*fn)(void));
+ *
+ * SYM_TYPED_FUNC_START() allows C to call this ID-mapped function indirectly
+ * using a physical pointer without triggering a kCFI failure.
+ */
+SYM_TYPED_FUNC_START(__pkvm_init_switch_pgd)
+	/* Turn the MMU off */
+	pre_disable_mmu_workaround
+	mrs	x3, sctlr_el2
+	bic	x4, x3, #SCTLR_ELx_M
+	msr	sctlr_el2, x4
+	isb
+
+	tlbi	alle2
+
+	/* Install the new pgtables */
+	phys_to_ttbr x5, x0
+alternative_if ARM64_HAS_CNP
+	orr	x5, x5, #TTBRx_EL1_CnP
+alternative_else_nop_endif
+	msr	ttbr0_el2, x5
+
+	/* Set the new stack pointer */
+	mov	sp, x1
+
+	/* And turn the MMU back on! */
+	dsb	nsh
+	isb
+	set_sctlr_el2	x3
+	ret	x2
+SYM_FUNC_END(__pkvm_init_switch_pgd)
+
+	.popsection

linux-7.2.3/arch/arm64/kvm/hyp/nvhe/hyp-main.c

@@ -0,0 +1,944 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2020 - Google Inc
+ * Author: Andrew Scull <ascull@google.com>
+ */
+
+#include <hyp/adjust_pc.h>
+#include <hyp/switch.h>
+
+#include <asm/pgtable-types.h>
+#include <asm/kvm_asm.h>
+#include <asm/kvm_emulate.h>
+#include <asm/kvm_host.h>
+#include <asm/kvm_hyp.h>
+#include <asm/kvm_hypevents.h>
+#include <asm/kvm_mmu.h>
+
+#include <nvhe/ffa.h>
+#include <nvhe/mem_protect.h>
+#include <nvhe/mm.h>
+#include <nvhe/pkvm.h>
+#include <nvhe/trace.h>
+#include <nvhe/trap_handler.h>
+
+DEFINE_PER_CPU(struct kvm_nvhe_init_params, kvm_init_params);
+
+/* Number of implemented GICv3 LRs. Used by flush_hyp_vcpu(). */
+unsigned int hyp_gicv3_nr_lr;
+
+void __kvm_hyp_host_forward_smc(struct kvm_cpu_context *host_ctxt);
+
+static void __hyp_sve_save_guest(struct kvm_vcpu *vcpu)
+{
+	__vcpu_assign_sys_reg(vcpu, ZCR_EL1, read_sysreg_el1(SYS_ZCR));
+	/*
+	 * On saving/restoring guest sve state, always use the maximum VL for
+	 * the guest. The layout of the data when saving the sve state depends
+	 * on the VL, so use a consistent (i.e., the maximum) guest VL.
+	 */
+	sve_cond_update_zcr_vq(vcpu_sve_max_vq(vcpu) - 1, SYS_ZCR_EL2);
+	sve_save_state(kern_hyp_va(vcpu->arch.sve_state), true);
+	fpsimd_save_common(&vcpu->arch.ctxt.fp_regs);
+	write_sysreg_s(sve_vq_from_vl(kvm_host_sve_max_vl) - 1, SYS_ZCR_EL2);
+}
+
+static void __hyp_sve_restore_host(void)
+{
+	struct kvm_cpu_context *hctxt = host_data_ptr(host_ctxt);
+	struct arm64_sve_state *sve_regs = *host_data_ptr(sve_regs);
+
+	/*
+	 * On saving/restoring host sve state, always use the maximum VL for
+	 * the host. The layout of the data when saving the sve state depends
+	 * on the VL, so use a consistent (i.e., the maximum) host VL.
+	 *
+	 * Note that this constrains the PE to the maximum shared VL
+	 * that was discovered, if we wish to use larger VLs this will
+	 * need to be revisited.
+	 */
+	write_sysreg_s(sve_vq_from_vl(kvm_host_sve_max_vl) - 1, SYS_ZCR_EL2);
+	sve_load_state(sve_regs, true);
+	fpsimd_load_common(&hctxt->fp_regs);
+	write_sysreg_el1(ctxt_sys_reg(hctxt, ZCR_EL1), SYS_ZCR);
+}
+
+static void fpsimd_sve_flush(void)
+{
+	*host_data_ptr(fp_owner) = FP_STATE_HOST_OWNED;
+}
+
+static void fpsimd_sve_sync(struct kvm_vcpu *vcpu)
+{
+	struct kvm_cpu_context *hctxt = host_data_ptr(host_ctxt);
+	bool has_fpmr;
+
+	if (!guest_owns_fp_regs())
+		return;
+
+	/*
+	 * Traps have been disabled by __deactivate_cptr_traps(), but there
+	 * hasn't necessarily been a context synchronization event yet.
+	 */
+	isb();
+
+	if (vcpu_has_sve(vcpu))
+		__hyp_sve_save_guest(vcpu);
+	else
+		fpsimd_save_state(&vcpu->arch.ctxt.fp_regs);
+
+	has_fpmr = kvm_has_fpmr(kern_hyp_va(vcpu->kvm));
+	if (has_fpmr)
+		__vcpu_assign_sys_reg(vcpu, FPMR, read_sysreg_s(SYS_FPMR));
+
+	if (system_supports_sve())
+		__hyp_sve_restore_host();
+	else
+		fpsimd_load_state(&hctxt->fp_regs);
+
+	if (has_fpmr)
+		write_sysreg_s(ctxt_sys_reg(hctxt, FPMR), SYS_FPMR);
+
+	*host_data_ptr(fp_owner) = FP_STATE_HOST_OWNED;
+}
+
+static void flush_debug_state(struct pkvm_hyp_vcpu *hyp_vcpu)
+{
+	struct kvm_vcpu *host_vcpu = hyp_vcpu->host_vcpu;
+
+	hyp_vcpu->vcpu.arch.debug_owner = host_vcpu->arch.debug_owner;
+
+	if (kvm_guest_owns_debug_regs(&hyp_vcpu->vcpu))
+		hyp_vcpu->vcpu.arch.vcpu_debug_state = host_vcpu->arch.vcpu_debug_state;
+	else if (kvm_host_owns_debug_regs(&hyp_vcpu->vcpu))
+		hyp_vcpu->vcpu.arch.external_debug_state = host_vcpu->arch.external_debug_state;
+}
+
+static void sync_debug_state(struct pkvm_hyp_vcpu *hyp_vcpu)
+{
+	struct kvm_vcpu *host_vcpu = hyp_vcpu->host_vcpu;
+
+	if (kvm_guest_owns_debug_regs(&hyp_vcpu->vcpu))
+		host_vcpu->arch.vcpu_debug_state = hyp_vcpu->vcpu.arch.vcpu_debug_state;
+	else if (kvm_host_owns_debug_regs(&hyp_vcpu->vcpu))
+		host_vcpu->arch.external_debug_state = hyp_vcpu->vcpu.arch.external_debug_state;
+}
+
+static void flush_hyp_vcpu(struct pkvm_hyp_vcpu *hyp_vcpu)
+{
+	struct kvm_vcpu *host_vcpu = hyp_vcpu->host_vcpu;
+
+	fpsimd_sve_flush();
+	flush_debug_state(hyp_vcpu);
+
+	hyp_vcpu->vcpu.arch.ctxt	= host_vcpu->arch.ctxt;
+
+	/* __hyp_running_vcpu must be NULL in a guest context. */
+	hyp_vcpu->vcpu.arch.ctxt.__hyp_running_vcpu = NULL;
+
+	hyp_vcpu->vcpu.arch.mdcr_el2	= host_vcpu->arch.mdcr_el2;
+	/*
+	 * HCR_EL2.VSE is host-owned (a pending virtual SError to inject), not a
+	 * trap-control bit, so it must flow to the hyp vCPU alongside TWI/TWE
+	 * for the vSError to be delivered. sync_hyp_vcpu() reflects it back.
+	 */
+	hyp_vcpu->vcpu.arch.hcr_el2 &= ~(HCR_TWI | HCR_TWE | HCR_VSE);
+	hyp_vcpu->vcpu.arch.hcr_el2 |= READ_ONCE(host_vcpu->arch.hcr_el2) &
+						 (HCR_TWI | HCR_TWE | HCR_VSE);
+
+	hyp_vcpu->vcpu.arch.iflags	= host_vcpu->arch.iflags;
+
+	hyp_vcpu->vcpu.arch.vsesr_el2	= host_vcpu->arch.vsesr_el2;
+
+	hyp_vcpu->vcpu.arch.vgic_cpu.vgic_v3 = host_vcpu->arch.vgic_cpu.vgic_v3;
+
+	/* Bound used_lrs by the number of implemented list registers. */
+	hyp_vcpu->vcpu.arch.vgic_cpu.vgic_v3.used_lrs =
+		min_t(unsigned int,
+		      hyp_vcpu->vcpu.arch.vgic_cpu.vgic_v3.used_lrs,
+		      hyp_gicv3_nr_lr);
+
+	hyp_vcpu->vcpu.arch.pid = host_vcpu->arch.pid;
+}
+
+static void sync_hyp_vcpu(struct pkvm_hyp_vcpu *hyp_vcpu)
+{
+	struct kvm_vcpu *host_vcpu = hyp_vcpu->host_vcpu;
+	struct vgic_v3_cpu_if *hyp_cpu_if = &hyp_vcpu->vcpu.arch.vgic_cpu.vgic_v3;
+	struct vgic_v3_cpu_if *host_cpu_if = &host_vcpu->arch.vgic_cpu.vgic_v3;
+	unsigned int i;
+
+	fpsimd_sve_sync(&hyp_vcpu->vcpu);
+	sync_debug_state(hyp_vcpu);
+
+	host_vcpu->arch.ctxt		= hyp_vcpu->vcpu.arch.ctxt;
+
+	host_vcpu->arch.hcr_el2		= hyp_vcpu->vcpu.arch.hcr_el2;
+
+	host_vcpu->arch.fault		= hyp_vcpu->vcpu.arch.fault;
+
+	host_vcpu->arch.iflags		= hyp_vcpu->vcpu.arch.iflags;
+
+	host_cpu_if->vgic_hcr		= hyp_cpu_if->vgic_hcr;
+	host_cpu_if->vgic_vmcr		= hyp_cpu_if->vgic_vmcr;
+	for (i = 0; i < hyp_cpu_if->used_lrs; ++i)
+		host_cpu_if->vgic_lr[i] = hyp_cpu_if->vgic_lr[i];
+}
+
+static void handle___pkvm_vcpu_load(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(pkvm_handle_t, handle, host_ctxt, 1);
+	DECLARE_REG(unsigned int, vcpu_idx, host_ctxt, 2);
+	DECLARE_REG(u64, hcr_el2, host_ctxt, 3);
+	struct pkvm_hyp_vcpu *hyp_vcpu;
+
+	hyp_vcpu = pkvm_load_hyp_vcpu(handle, vcpu_idx);
+	if (!hyp_vcpu)
+		return;
+
+	if (pkvm_hyp_vcpu_is_protected(hyp_vcpu)) {
+		/* Propagate WFx trapping flags */
+		hyp_vcpu->vcpu.arch.hcr_el2 &= ~(HCR_TWE | HCR_TWI);
+		hyp_vcpu->vcpu.arch.hcr_el2 |= hcr_el2 & (HCR_TWE | HCR_TWI);
+	} else {
+		memcpy(&hyp_vcpu->vcpu.arch.fgt, hyp_vcpu->host_vcpu->arch.fgt,
+		       sizeof(hyp_vcpu->vcpu.arch.fgt));
+	}
+}
+
+static void handle___pkvm_vcpu_put(struct kvm_cpu_context *host_ctxt)
+{
+	struct pkvm_hyp_vcpu *hyp_vcpu = pkvm_get_loaded_hyp_vcpu();
+
+	if (hyp_vcpu)
+		pkvm_put_hyp_vcpu(hyp_vcpu);
+}
+
+static void handle___kvm_vcpu_run(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(struct kvm_vcpu *, host_vcpu, host_ctxt, 1);
+	int ret;
+
+	if (unlikely(is_protected_kvm_enabled())) {
+		struct pkvm_hyp_vcpu *hyp_vcpu = pkvm_get_loaded_hyp_vcpu();
+
+		/*
+		 * KVM (and pKVM) doesn't support SME guests for now, and
+		 * ensures that SME features aren't enabled in pstate when
+		 * loading a vcpu. Therefore, if SME features enabled the host
+		 * is misbehaving.
+		 */
+		if (unlikely(system_supports_sme() && read_sysreg_s(SYS_SVCR))) {
+			ret = -EINVAL;
+			goto out;
+		}
+
+		if (!hyp_vcpu) {
+			ret = -EINVAL;
+			goto out;
+		}
+
+		flush_hyp_vcpu(hyp_vcpu);
+
+		ret = __kvm_vcpu_run(&hyp_vcpu->vcpu);
+
+		sync_hyp_vcpu(hyp_vcpu);
+	} else {
+		struct kvm_vcpu *vcpu = kern_hyp_va(host_vcpu);
+
+		/* The host is fully trusted, run its vCPU directly. */
+		fpsimd_lazy_switch_to_guest(vcpu);
+		ret = __kvm_vcpu_run(vcpu);
+		fpsimd_lazy_switch_to_host(vcpu);
+	}
+out:
+	cpu_reg(host_ctxt, 1) =  ret;
+}
+
+static int pkvm_refill_memcache(struct pkvm_hyp_vcpu *hyp_vcpu)
+{
+	struct kvm_vcpu *host_vcpu = hyp_vcpu->host_vcpu;
+
+	return refill_memcache(&hyp_vcpu->vcpu.arch.pkvm_memcache,
+			       host_vcpu->arch.pkvm_memcache.nr_pages,
+			       &host_vcpu->arch.pkvm_memcache);
+}
+
+static void handle___pkvm_host_donate_guest(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(u64, pfn, host_ctxt, 1);
+	DECLARE_REG(u64, gfn, host_ctxt, 2);
+	struct pkvm_hyp_vcpu *hyp_vcpu;
+	int ret = -EINVAL;
+
+	hyp_vcpu = pkvm_get_loaded_hyp_vcpu();
+	if (!hyp_vcpu || !pkvm_hyp_vcpu_is_protected(hyp_vcpu))
+		goto out;
+
+	ret = pkvm_refill_memcache(hyp_vcpu);
+	if (ret)
+		goto out;
+
+	ret = __pkvm_host_donate_guest(pfn, gfn, hyp_vcpu);
+out:
+	cpu_reg(host_ctxt, 1) =  ret;
+}
+
+static void handle___pkvm_host_share_guest(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(u64, pfn, host_ctxt, 1);
+	DECLARE_REG(u64, gfn, host_ctxt, 2);
+	DECLARE_REG(u64, nr_pages, host_ctxt, 3);
+	DECLARE_REG(enum kvm_pgtable_prot, prot, host_ctxt, 4);
+	struct pkvm_hyp_vcpu *hyp_vcpu;
+	int ret = -EINVAL;
+
+	hyp_vcpu = pkvm_get_loaded_hyp_vcpu();
+	if (!hyp_vcpu || pkvm_hyp_vcpu_is_protected(hyp_vcpu))
+		goto out;
+
+	ret = pkvm_refill_memcache(hyp_vcpu);
+	if (ret)
+		goto out;
+
+	ret = __pkvm_host_share_guest(pfn, gfn, nr_pages, hyp_vcpu, prot);
+out:
+	cpu_reg(host_ctxt, 1) =  ret;
+}
+
+static void handle___pkvm_host_unshare_guest(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(pkvm_handle_t, handle, host_ctxt, 1);
+	DECLARE_REG(u64, gfn, host_ctxt, 2);
+	DECLARE_REG(u64, nr_pages, host_ctxt, 3);
+	struct pkvm_hyp_vm *hyp_vm;
+	int ret = -EINVAL;
+
+	hyp_vm = get_np_pkvm_hyp_vm(handle);
+	if (!hyp_vm)
+		goto out;
+
+	ret = __pkvm_host_unshare_guest(gfn, nr_pages, hyp_vm);
+	put_pkvm_hyp_vm(hyp_vm);
+out:
+	cpu_reg(host_ctxt, 1) =  ret;
+}
+
+static void handle___pkvm_host_relax_perms_guest(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(u64, gfn, host_ctxt, 1);
+	DECLARE_REG(enum kvm_pgtable_prot, prot, host_ctxt, 2);
+	struct pkvm_hyp_vcpu *hyp_vcpu;
+	int ret = -EINVAL;
+
+	hyp_vcpu = pkvm_get_loaded_hyp_vcpu();
+	if (!hyp_vcpu || pkvm_hyp_vcpu_is_protected(hyp_vcpu))
+		goto out;
+
+	ret = __pkvm_host_relax_perms_guest(gfn, hyp_vcpu, prot);
+out:
+	cpu_reg(host_ctxt, 1) = ret;
+}
+
+static void handle___pkvm_host_wrprotect_guest(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(pkvm_handle_t, handle, host_ctxt, 1);
+	DECLARE_REG(u64, gfn, host_ctxt, 2);
+	DECLARE_REG(u64, nr_pages, host_ctxt, 3);
+	struct pkvm_hyp_vm *hyp_vm;
+	int ret = -EINVAL;
+
+	hyp_vm = get_np_pkvm_hyp_vm(handle);
+	if (!hyp_vm)
+		goto out;
+
+	ret = __pkvm_host_wrprotect_guest(gfn, nr_pages, hyp_vm);
+	put_pkvm_hyp_vm(hyp_vm);
+out:
+	cpu_reg(host_ctxt, 1) = ret;
+}
+
+static void handle___pkvm_host_test_clear_young_guest(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(pkvm_handle_t, handle, host_ctxt, 1);
+	DECLARE_REG(u64, gfn, host_ctxt, 2);
+	DECLARE_REG(u64, nr_pages, host_ctxt, 3);
+	DECLARE_REG(bool, mkold, host_ctxt, 4);
+	struct pkvm_hyp_vm *hyp_vm;
+	int ret = -EINVAL;
+
+	hyp_vm = get_np_pkvm_hyp_vm(handle);
+	if (!hyp_vm)
+		goto out;
+
+	ret = __pkvm_host_test_clear_young_guest(gfn, nr_pages, mkold, hyp_vm);
+	put_pkvm_hyp_vm(hyp_vm);
+out:
+	cpu_reg(host_ctxt, 1) = ret;
+}
+
+static void handle___pkvm_host_mkyoung_guest(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(u64, gfn, host_ctxt, 1);
+	struct pkvm_hyp_vcpu *hyp_vcpu;
+	int ret = -EINVAL;
+
+	hyp_vcpu = pkvm_get_loaded_hyp_vcpu();
+	if (!hyp_vcpu || pkvm_hyp_vcpu_is_protected(hyp_vcpu))
+		goto out;
+
+	ret = __pkvm_host_mkyoung_guest(gfn, hyp_vcpu);
+out:
+	cpu_reg(host_ctxt, 1) =  ret;
+}
+
+static void handle___kvm_adjust_pc(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(struct kvm_vcpu *, vcpu, host_ctxt, 1);
+
+	__kvm_adjust_pc(kern_hyp_va(vcpu));
+}
+
+static void handle___kvm_flush_vm_context(struct kvm_cpu_context *host_ctxt)
+{
+	__kvm_flush_vm_context();
+}
+
+static void handle___kvm_tlb_flush_vmid_ipa(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(struct kvm_s2_mmu *, mmu, host_ctxt, 1);
+	DECLARE_REG(phys_addr_t, ipa, host_ctxt, 2);
+	DECLARE_REG(int, level, host_ctxt, 3);
+
+	__kvm_tlb_flush_vmid_ipa(kern_hyp_va(mmu), ipa, level);
+}
+
+static void handle___kvm_tlb_flush_vmid_ipa_nsh(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(struct kvm_s2_mmu *, mmu, host_ctxt, 1);
+	DECLARE_REG(phys_addr_t, ipa, host_ctxt, 2);
+	DECLARE_REG(int, level, host_ctxt, 3);
+
+	__kvm_tlb_flush_vmid_ipa_nsh(kern_hyp_va(mmu), ipa, level);
+}
+
+static void
+handle___kvm_tlb_flush_vmid_range(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(struct kvm_s2_mmu *, mmu, host_ctxt, 1);
+	DECLARE_REG(phys_addr_t, start, host_ctxt, 2);
+	DECLARE_REG(unsigned long, pages, host_ctxt, 3);
+
+	__kvm_tlb_flush_vmid_range(kern_hyp_va(mmu), start, pages);
+}
+
+static void handle___kvm_tlb_flush_vmid(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(struct kvm_s2_mmu *, mmu, host_ctxt, 1);
+
+	__kvm_tlb_flush_vmid(kern_hyp_va(mmu));
+}
+
+static void handle___pkvm_tlb_flush_vmid(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(pkvm_handle_t, handle, host_ctxt, 1);
+	struct pkvm_hyp_vm *hyp_vm = get_np_pkvm_hyp_vm(handle);
+
+	if (!hyp_vm)
+		return;
+
+	__kvm_tlb_flush_vmid(&hyp_vm->kvm.arch.mmu);
+	put_pkvm_hyp_vm(hyp_vm);
+}
+
+static void handle___kvm_flush_cpu_context(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(struct kvm_s2_mmu *, mmu, host_ctxt, 1);
+
+	__kvm_flush_cpu_context(kern_hyp_va(mmu));
+}
+
+static void handle___kvm_timer_set_cntvoff(struct kvm_cpu_context *host_ctxt)
+{
+	__kvm_timer_set_cntvoff(cpu_reg(host_ctxt, 1));
+}
+
+static void handle___kvm_enable_ssbs(struct kvm_cpu_context *host_ctxt)
+{
+	u64 tmp;
+
+	tmp = read_sysreg_el2(SYS_SCTLR);
+	tmp |= SCTLR_ELx_DSSBS;
+	write_sysreg_el2(tmp, SYS_SCTLR);
+}
+
+static void handle___vgic_v3_get_gic_config(struct kvm_cpu_context *host_ctxt)
+{
+	cpu_reg(host_ctxt, 1) = __vgic_v3_get_gic_config();
+}
+
+static void handle___vgic_v3_init_lrs(struct kvm_cpu_context *host_ctxt)
+{
+	__vgic_v3_init_lrs();
+}
+
+static void handle___vgic_v3_save_aprs(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(struct vgic_v3_cpu_if *, cpu_if, host_ctxt, 1);
+
+	__vgic_v3_save_aprs(kern_hyp_va(cpu_if));
+}
+
+static void handle___vgic_v3_restore_vmcr_aprs(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(struct vgic_v3_cpu_if *, cpu_if, host_ctxt, 1);
+
+	__vgic_v3_restore_vmcr_aprs(kern_hyp_va(cpu_if));
+}
+
+static void handle___pkvm_init(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(phys_addr_t, phys, host_ctxt, 1);
+	DECLARE_REG(unsigned long, size, host_ctxt, 2);
+	DECLARE_REG(unsigned long *, per_cpu_base, host_ctxt, 3);
+	DECLARE_REG(u32, hyp_va_bits, host_ctxt, 4);
+
+	/*
+	 * __pkvm_init() will return only if an error occurred, otherwise it
+	 * will tail-call in __pkvm_init_finalise() which will have to deal
+	 * with the host context directly.
+	 */
+	cpu_reg(host_ctxt, 1) = __pkvm_init(phys, size, per_cpu_base, hyp_va_bits);
+}
+
+static void handle___pkvm_cpu_set_vector(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(enum arm64_hyp_spectre_vector, slot, host_ctxt, 1);
+
+	cpu_reg(host_ctxt, 1) = pkvm_cpu_set_vector(slot);
+}
+
+static void handle___pkvm_host_share_hyp(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(u64, pfn, host_ctxt, 1);
+
+	cpu_reg(host_ctxt, 1) = __pkvm_host_share_hyp(pfn);
+}
+
+static void handle___pkvm_host_unshare_hyp(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(u64, pfn, host_ctxt, 1);
+
+	cpu_reg(host_ctxt, 1) = __pkvm_host_unshare_hyp(pfn);
+}
+
+static void handle___pkvm_create_private_mapping(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(phys_addr_t, phys, host_ctxt, 1);
+	DECLARE_REG(size_t, size, host_ctxt, 2);
+	DECLARE_REG(enum kvm_pgtable_prot, prot, host_ctxt, 3);
+
+	/*
+	 * __pkvm_create_private_mapping() populates a pointer with the
+	 * hypervisor start address of the allocation.
+	 *
+	 * However, handle___pkvm_create_private_mapping() hypercall crosses the
+	 * EL1/EL2 boundary so the pointer would not be valid in this context.
+	 *
+	 * Instead pass the allocation address as the return value (or return
+	 * ERR_PTR() on failure).
+	 */
+	unsigned long haddr;
+	int err = __pkvm_create_private_mapping(phys, size, prot, &haddr);
+
+	if (err)
+		haddr = (unsigned long)ERR_PTR(err);
+
+	cpu_reg(host_ctxt, 1) = haddr;
+}
+
+static void handle___pkvm_prot_finalize(struct kvm_cpu_context *host_ctxt)
+{
+	cpu_reg(host_ctxt, 1) = __pkvm_prot_finalize();
+}
+
+static void handle___pkvm_reserve_vm(struct kvm_cpu_context *host_ctxt)
+{
+	cpu_reg(host_ctxt, 1) = __pkvm_reserve_vm();
+}
+
+static void handle___pkvm_unreserve_vm(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(pkvm_handle_t, handle, host_ctxt, 1);
+
+	__pkvm_unreserve_vm(handle);
+}
+
+static void handle___pkvm_init_vm(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(struct kvm *, host_kvm, host_ctxt, 1);
+	DECLARE_REG(unsigned long, vm_hva, host_ctxt, 2);
+	DECLARE_REG(unsigned long, pgd_hva, host_ctxt, 3);
+
+	host_kvm = kern_hyp_va(host_kvm);
+	cpu_reg(host_ctxt, 1) = __pkvm_init_vm(host_kvm, vm_hva, pgd_hva);
+}
+
+static void handle___pkvm_init_vcpu(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(pkvm_handle_t, handle, host_ctxt, 1);
+	DECLARE_REG(struct kvm_vcpu *, host_vcpu, host_ctxt, 2);
+	DECLARE_REG(unsigned long, vcpu_hva, host_ctxt, 3);
+
+	host_vcpu = kern_hyp_va(host_vcpu);
+	cpu_reg(host_ctxt, 1) = __pkvm_init_vcpu(handle, host_vcpu, vcpu_hva);
+}
+
+static void handle___pkvm_vcpu_in_poison_fault(struct kvm_cpu_context *host_ctxt)
+{
+	int ret;
+	struct pkvm_hyp_vcpu *hyp_vcpu = pkvm_get_loaded_hyp_vcpu();
+
+	ret = hyp_vcpu ? __pkvm_vcpu_in_poison_fault(hyp_vcpu) : -EINVAL;
+	cpu_reg(host_ctxt, 1) = ret;
+}
+
+static void handle___pkvm_force_reclaim_guest_page(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(phys_addr_t, phys, host_ctxt, 1);
+
+	cpu_reg(host_ctxt, 1) = __pkvm_host_force_reclaim_page_guest(phys);
+}
+
+static void handle___pkvm_reclaim_dying_guest_page(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(pkvm_handle_t, handle, host_ctxt, 1);
+	DECLARE_REG(u64, gfn, host_ctxt, 2);
+
+	cpu_reg(host_ctxt, 1) = __pkvm_reclaim_dying_guest_page(handle, gfn);
+}
+
+static void handle___pkvm_start_teardown_vm(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(pkvm_handle_t, handle, host_ctxt, 1);
+
+	cpu_reg(host_ctxt, 1) = __pkvm_start_teardown_vm(handle);
+}
+
+static void handle___pkvm_finalize_teardown_vm(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(pkvm_handle_t, handle, host_ctxt, 1);
+
+	cpu_reg(host_ctxt, 1) = __pkvm_finalize_teardown_vm(handle);
+}
+
+static void handle___tracing_load(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(unsigned long, desc_hva, host_ctxt, 1);
+	DECLARE_REG(size_t, desc_size, host_ctxt, 2);
+
+	cpu_reg(host_ctxt, 1) = __tracing_load(desc_hva, desc_size);
+}
+
+static void handle___tracing_unload(struct kvm_cpu_context *host_ctxt)
+{
+	__tracing_unload();
+}
+
+static void handle___tracing_enable(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(bool, enable, host_ctxt, 1);
+
+	cpu_reg(host_ctxt, 1) = __tracing_enable(enable);
+}
+
+static void handle___tracing_swap_reader(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(unsigned int, cpu, host_ctxt, 1);
+
+	cpu_reg(host_ctxt, 1) = __tracing_swap_reader(cpu);
+}
+
+static void handle___tracing_update_clock(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(u32, mult, host_ctxt, 1);
+	DECLARE_REG(u32, shift, host_ctxt, 2);
+	DECLARE_REG(u64, epoch_ns, host_ctxt, 3);
+	DECLARE_REG(u64, epoch_cyc, host_ctxt, 4);
+
+	__tracing_update_clock(mult, shift, epoch_ns, epoch_cyc);
+}
+
+static void handle___tracing_reset(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(unsigned int, cpu, host_ctxt, 1);
+
+	cpu_reg(host_ctxt, 1) = __tracing_reset(cpu);
+}
+
+static void handle___tracing_enable_event(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(unsigned short, id, host_ctxt, 1);
+	DECLARE_REG(bool, enable, host_ctxt, 2);
+
+	cpu_reg(host_ctxt, 1) = __tracing_enable_event(id, enable);
+}
+
+static void handle___tracing_write_event(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(u64, id, host_ctxt, 1);
+
+	trace_selftest(id);
+}
+
+static void handle___vgic_v5_save_apr(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(struct vgic_v5_cpu_if *, cpu_if, host_ctxt, 1);
+
+	__vgic_v5_save_apr(kern_hyp_va(cpu_if));
+}
+
+static void handle___vgic_v5_restore_vmcr_apr(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(struct vgic_v5_cpu_if *, cpu_if, host_ctxt, 1);
+
+	__vgic_v5_restore_vmcr_apr(kern_hyp_va(cpu_if));
+}
+
+typedef void (*hcall_t)(struct kvm_cpu_context *);
+
+#define HANDLE_FUNC(x)	[__KVM_HOST_SMCCC_FUNC_##x] = (hcall_t)handle_##x
+
+static const hcall_t host_hcall[] = {
+	/* ___kvm_hyp_init */
+	HANDLE_FUNC(__pkvm_init),
+	HANDLE_FUNC(__pkvm_create_private_mapping),
+	HANDLE_FUNC(__pkvm_cpu_set_vector),
+	HANDLE_FUNC(__kvm_enable_ssbs),
+	HANDLE_FUNC(__vgic_v3_init_lrs),
+	HANDLE_FUNC(__vgic_v3_get_gic_config),
+	HANDLE_FUNC(__pkvm_prot_finalize),
+
+	HANDLE_FUNC(__kvm_adjust_pc),
+	HANDLE_FUNC(__kvm_vcpu_run),
+	HANDLE_FUNC(__kvm_flush_vm_context),
+	HANDLE_FUNC(__kvm_tlb_flush_vmid_ipa),
+	HANDLE_FUNC(__kvm_tlb_flush_vmid_ipa_nsh),
+	HANDLE_FUNC(__kvm_tlb_flush_vmid),
+	HANDLE_FUNC(__kvm_tlb_flush_vmid_range),
+	HANDLE_FUNC(__kvm_flush_cpu_context),
+	HANDLE_FUNC(__kvm_timer_set_cntvoff),
+	HANDLE_FUNC(__tracing_load),
+	HANDLE_FUNC(__tracing_unload),
+	HANDLE_FUNC(__tracing_enable),
+	HANDLE_FUNC(__tracing_swap_reader),
+	HANDLE_FUNC(__tracing_update_clock),
+	HANDLE_FUNC(__tracing_reset),
+	HANDLE_FUNC(__tracing_enable_event),
+	HANDLE_FUNC(__tracing_write_event),
+	HANDLE_FUNC(__vgic_v3_save_aprs),
+	HANDLE_FUNC(__vgic_v3_restore_vmcr_aprs),
+	HANDLE_FUNC(__vgic_v5_save_apr),
+	HANDLE_FUNC(__vgic_v5_restore_vmcr_apr),
+
+	HANDLE_FUNC(__pkvm_host_share_hyp),
+	HANDLE_FUNC(__pkvm_host_unshare_hyp),
+	HANDLE_FUNC(__pkvm_host_donate_guest),
+	HANDLE_FUNC(__pkvm_host_share_guest),
+	HANDLE_FUNC(__pkvm_host_unshare_guest),
+	HANDLE_FUNC(__pkvm_host_relax_perms_guest),
+	HANDLE_FUNC(__pkvm_host_wrprotect_guest),
+	HANDLE_FUNC(__pkvm_host_test_clear_young_guest),
+	HANDLE_FUNC(__pkvm_host_mkyoung_guest),
+	HANDLE_FUNC(__pkvm_reserve_vm),
+	HANDLE_FUNC(__pkvm_unreserve_vm),
+	HANDLE_FUNC(__pkvm_init_vm),
+	HANDLE_FUNC(__pkvm_init_vcpu),
+	HANDLE_FUNC(__pkvm_vcpu_in_poison_fault),
+	HANDLE_FUNC(__pkvm_force_reclaim_guest_page),
+	HANDLE_FUNC(__pkvm_reclaim_dying_guest_page),
+	HANDLE_FUNC(__pkvm_start_teardown_vm),
+	HANDLE_FUNC(__pkvm_finalize_teardown_vm),
+	HANDLE_FUNC(__pkvm_vcpu_load),
+	HANDLE_FUNC(__pkvm_vcpu_put),
+	HANDLE_FUNC(__pkvm_tlb_flush_vmid),
+};
+
+static void handle_host_hcall(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(unsigned long, id, host_ctxt, 0);
+	unsigned long hcall_min = 0, hcall_max = __KVM_HOST_SMCCC_FUNC_MAX;
+	hcall_t hfn;
+
+	BUILD_BUG_ON(ARRAY_SIZE(host_hcall) != __KVM_HOST_SMCCC_FUNC_MAX);
+
+	/*
+	 * If pKVM has been initialised then reject any calls to the
+	 * early "privileged" hypercalls. Note that we cannot reject
+	 * calls to __pkvm_prot_finalize for two reasons: (1) The static
+	 * key used to determine initialisation must be toggled prior to
+	 * finalisation and (2) finalisation is performed on a per-CPU
+	 * basis. This is all fine, however, since __pkvm_prot_finalize
+	 * returns -EPERM after the first call for a given CPU.
+	 */
+	if (static_branch_unlikely(&kvm_protected_mode_initialized)) {
+		hcall_min = __KVM_HOST_SMCCC_FUNC_MIN_PKVM;
+	} else {
+		hcall_max = __KVM_HOST_SMCCC_FUNC_PKVM_ONLY;
+	}
+
+	id &= ~ARM_SMCCC_CALL_HINTS;
+	id -= KVM_HOST_SMCCC_ID(0);
+
+	if (unlikely(id < hcall_min || id >= hcall_max))
+		goto inval;
+
+	hfn = host_hcall[id];
+	if (unlikely(!hfn))
+		goto inval;
+
+	cpu_reg(host_ctxt, 0) = SMCCC_RET_SUCCESS;
+	hfn(host_ctxt);
+
+	return;
+inval:
+	cpu_reg(host_ctxt, 0) = SMCCC_RET_NOT_SUPPORTED;
+}
+
+static void default_host_smc_handler(struct kvm_cpu_context *host_ctxt)
+{
+	trace_hyp_exit(host_ctxt, HYP_REASON_SMC);
+	__kvm_hyp_host_forward_smc(host_ctxt);
+	trace_hyp_enter(host_ctxt, HYP_REASON_SMC);
+}
+
+static void handle_host_smc(struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(u64, func_id, host_ctxt, 0);
+	u64 esr = read_sysreg_el2(SYS_ESR);
+	bool handled;
+
+	if (esr & ESR_ELx_xVC_IMM_MASK) {
+		cpu_reg(host_ctxt, 0) = SMCCC_RET_NOT_SUPPORTED;
+		goto exit_skip_instr;
+	}
+
+	func_id &= ~ARM_SMCCC_CALL_HINTS;
+	if (upper_32_bits(func_id)) {
+		cpu_reg(host_ctxt, 0) = SMCCC_RET_NOT_SUPPORTED;
+		goto exit_skip_instr;
+	}
+
+	handled = kvm_host_psci_handler(host_ctxt, func_id);
+	if (!handled)
+		handled = kvm_host_ffa_handler(host_ctxt, func_id);
+	if (!handled)
+		default_host_smc_handler(host_ctxt);
+
+exit_skip_instr:
+	/* SMC was trapped, move ELR past the current PC. */
+	kvm_skip_host_instr();
+}
+
+void inject_host_exception(u64 esr)
+{
+	u64 sctlr, spsr_el1, spsr_el2, exc_offset = except_type_sync;
+	const u64 spsr_mask = PSR_N_BIT | PSR_Z_BIT | PSR_C_BIT |
+			      PSR_V_BIT | PSR_DIT_BIT | PSR_PAN_BIT;
+
+	spsr_el1 = spsr_el2 = read_sysreg_el2(SYS_SPSR);
+	switch (spsr_el1 & (PSR_MODE_MASK | PSR_MODE32_BIT)) {
+	case PSR_MODE_EL0t:
+		exc_offset += LOWER_EL_AArch64_VECTOR;
+		break;
+	case PSR_MODE_EL0t | PSR_MODE32_BIT:
+		exc_offset += LOWER_EL_AArch32_VECTOR;
+		break;
+	default:
+		exc_offset += CURRENT_EL_SP_ELx_VECTOR;
+	}
+
+	spsr_el2 &= spsr_mask;
+	spsr_el2 |= PSR_D_BIT | PSR_A_BIT | PSR_I_BIT | PSR_F_BIT |
+		    PSR_MODE_EL1h;
+
+	sctlr = read_sysreg_el1(SYS_SCTLR);
+	if (!(sctlr & SCTLR_EL1_SPAN))
+		spsr_el2 |= PSR_PAN_BIT;
+
+	if (sctlr & SCTLR_ELx_DSSBS)
+		spsr_el2 |= PSR_SSBS_BIT;
+
+	if (system_supports_mte())
+		spsr_el2 |= PSR_TCO_BIT;
+
+	if (esr_fsc_is_translation_fault(esr))
+		write_sysreg_el1(read_sysreg_el2(SYS_FAR), SYS_FAR);
+
+	write_sysreg_el1(esr, SYS_ESR);
+	write_sysreg_el1(read_sysreg_el2(SYS_ELR), SYS_ELR);
+	write_sysreg_el1(spsr_el1, SYS_SPSR);
+	write_sysreg_el2(read_sysreg_el1(SYS_VBAR) + exc_offset, SYS_ELR);
+	write_sysreg_el2(spsr_el2, SYS_SPSR);
+}
+
+static void inject_host_undef64(void)
+{
+	inject_host_exception((ESR_ELx_EC_UNKNOWN << ESR_ELx_EC_SHIFT) |
+			       ESR_ELx_IL);
+}
+
+static bool handle_host_mte(u64 esr)
+{
+	switch (esr_sys64_to_sysreg(esr)) {
+	case SYS_RGSR_EL1:
+	case SYS_GCR_EL1:
+	case SYS_TFSR_EL1:
+	case SYS_TFSRE0_EL1:
+		/* If we're here for any reason other than MTE, it's a bug. */
+		if (read_sysreg(HCR_EL2) & HCR_ATA)
+			return false;
+		break;
+	case SYS_GMID_EL1:
+		/* If we're here for any reason other than MTE, it's a bug. */
+		if (!(read_sysreg(HCR_EL2) & HCR_TID5))
+			return false;
+		break;
+	default:
+		return false;
+	}
+
+	inject_host_undef64();
+	return true;
+}
+
+void handle_trap(struct kvm_cpu_context *host_ctxt)
+{
+	u64 esr = read_sysreg_el2(SYS_ESR);
+
+
+	switch (ESR_ELx_EC(esr)) {
+	case ESR_ELx_EC_HVC64:
+		trace_hyp_enter(host_ctxt, HYP_REASON_HVC);
+		handle_host_hcall(host_ctxt);
+		break;
+	case ESR_ELx_EC_SMC64:
+		trace_hyp_enter(host_ctxt, HYP_REASON_SMC);
+		handle_host_smc(host_ctxt);
+		break;
+	case ESR_ELx_EC_IABT_LOW:
+	case ESR_ELx_EC_DABT_LOW:
+		trace_hyp_enter(host_ctxt, HYP_REASON_HOST_ABORT);
+		handle_host_mem_abort(host_ctxt);
+		break;
+	case ESR_ELx_EC_SYS64:
+		trace_hyp_enter(host_ctxt, HYP_REASON_SYS);
+		if (handle_host_mte(esr))
+			break;
+		fallthrough;
+	default:
+		BUG();
+	}
+
+	trace_hyp_exit(host_ctxt, HYP_REASON_ERET_HOST);
+}

linux-7.2.3/arch/arm64/kvm/hyp/nvhe/hyp-smp.c

@@ -0,0 +1,40 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2020 - Google LLC
+ * Author: David Brazdil <dbrazdil@google.com>
+ */
+
+#include <asm/kvm_asm.h>
+#include <asm/kvm_hyp.h>
+#include <asm/kvm_mmu.h>
+
+/*
+ * nVHE copy of data structures tracking available CPU cores.
+ * Only entries for CPUs that were online at KVM init are populated.
+ * Other CPUs should not be allowed to boot because their features were
+ * not checked against the finalized system capabilities.
+ */
+u64 __ro_after_init hyp_cpu_logical_map[NR_CPUS] = { [0 ... NR_CPUS-1] = INVALID_HWID };
+
+u64 cpu_logical_map(unsigned int cpu)
+{
+	BUG_ON(cpu >= ARRAY_SIZE(hyp_cpu_logical_map));
+
+	return hyp_cpu_logical_map[cpu];
+}
+
+unsigned long __ro_after_init kvm_arm_hyp_percpu_base[NR_CPUS];
+
+unsigned long __hyp_per_cpu_offset(unsigned int cpu)
+{
+	unsigned long *cpu_base_array;
+	unsigned long this_cpu_base;
+	unsigned long elf_base;
+
+	BUG_ON(cpu >= ARRAY_SIZE(kvm_arm_hyp_percpu_base));
+
+	cpu_base_array = (unsigned long *)&kvm_arm_hyp_percpu_base;
+	this_cpu_base = kern_hyp_va(cpu_base_array[cpu]);
+	elf_base = (unsigned long)&__per_cpu_start;
+	return this_cpu_base - elf_base;
+}

linux-7.2.3/arch/arm64/kvm/hyp/nvhe/hyp.lds.S

@@ -0,0 +1,37 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (C) 2020 Google LLC.
+ * Written by David Brazdil <dbrazdil@google.com>
+ *
+ * Linker script used for partial linking of nVHE EL2 object files.
+ */
+
+#include <asm/hyp_image.h>
+#include <asm-generic/vmlinux.lds.h>
+#include <asm/cache.h>
+#include <asm/memory.h>
+
+SECTIONS {
+	HYP_SECTION(.idmap.text)
+	HYP_SECTION(.text)
+	HYP_SECTION(.data..ro_after_init)
+	HYP_SECTION(.rodata)
+#ifdef CONFIG_NVHE_EL2_TRACING
+	. = ALIGN(PAGE_SIZE);
+	BEGIN_HYP_SECTION(.event_ids)
+		*(SORT(.hyp.event_ids.*))
+	END_HYP_SECTION
+#endif
+
+	/*
+	 * .hyp..data..percpu needs to be page aligned to maintain the same
+	 * alignment for when linking into vmlinux.
+	 */
+	. = ALIGN(PAGE_SIZE);
+	BEGIN_HYP_SECTION(.data..percpu)
+		PERCPU_INPUT(L1_CACHE_BYTES)
+	END_HYP_SECTION
+
+	HYP_SECTION(.bss)
+	HYP_SECTION(.data)
+}

linux-7.2.3/arch/arm64/kvm/hyp/nvhe/list_debug.c

@@ -0,0 +1,56 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2022 - Google LLC
+ * Author: Keir Fraser <keirf@google.com>
+ */
+
+#include <linux/list.h>
+#include <linux/bug.h>
+
+static inline __must_check bool nvhe_check_data_corruption(bool v)
+{
+	return v;
+}
+
+#define NVHE_CHECK_DATA_CORRUPTION(condition)				 \
+	nvhe_check_data_corruption(({					 \
+		bool corruption = unlikely(condition);			 \
+		if (corruption) {					 \
+			if (IS_ENABLED(CONFIG_BUG_ON_DATA_CORRUPTION)) { \
+				BUG();				 	 \
+			} else						 \
+				WARN_ON(1);				 \
+		}							 \
+		corruption;						 \
+	}))
+
+/* The predicates checked here are taken from lib/list_debug.c. */
+
+__list_valid_slowpath
+bool __list_add_valid_or_report(struct list_head *new, struct list_head *prev,
+				struct list_head *next)
+{
+	if (NVHE_CHECK_DATA_CORRUPTION(next->prev != prev) ||
+	    NVHE_CHECK_DATA_CORRUPTION(prev->next != next) ||
+	    NVHE_CHECK_DATA_CORRUPTION(new == prev || new == next))
+		return false;
+
+	return true;
+}
+
+__list_valid_slowpath
+bool __list_del_entry_valid_or_report(struct list_head *entry)
+{
+	struct list_head *prev, *next;
+
+	prev = entry->prev;
+	next = entry->next;
+
+	if (NVHE_CHECK_DATA_CORRUPTION(next == LIST_POISON1) ||
+	    NVHE_CHECK_DATA_CORRUPTION(prev == LIST_POISON2) ||
+	    NVHE_CHECK_DATA_CORRUPTION(prev->next != entry) ||
+	    NVHE_CHECK_DATA_CORRUPTION(next->prev != entry))
+		return false;
+
+	return true;
+}

linux-7.2.3/arch/arm64/kvm/hyp/nvhe/mm.c

@@ -0,0 +1,504 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2020 Google LLC
+ * Author: Quentin Perret <qperret@google.com>
+ */
+
+#include <linux/kvm_host.h>
+#include <asm/kvm_hyp.h>
+#include <asm/kvm_mmu.h>
+#include <asm/kvm_pgtable.h>
+#include <asm/kvm_pkvm.h>
+#include <asm/spectre.h>
+
+#include <nvhe/early_alloc.h>
+#include <nvhe/gfp.h>
+#include <nvhe/memory.h>
+#include <nvhe/mem_protect.h>
+#include <nvhe/mm.h>
+#include <nvhe/spinlock.h>
+
+struct kvm_pgtable pkvm_pgtable;
+hyp_spinlock_t pkvm_pgd_lock;
+
+struct memblock_region hyp_memory[HYP_MEMBLOCK_REGIONS];
+unsigned int hyp_memblock_nr;
+
+static u64 __io_map_base;
+
+struct hyp_fixmap_slot {
+	u64 addr;
+	kvm_pte_t *ptep;
+};
+static DEFINE_PER_CPU(struct hyp_fixmap_slot, fixmap_slots);
+
+static int __pkvm_create_mappings(unsigned long start, unsigned long size,
+				  unsigned long phys, enum kvm_pgtable_prot prot)
+{
+	int err;
+
+	hyp_spin_lock(&pkvm_pgd_lock);
+	err = kvm_pgtable_hyp_map(&pkvm_pgtable, start, size, phys, prot);
+	hyp_spin_unlock(&pkvm_pgd_lock);
+
+	return err;
+}
+
+static int __pkvm_alloc_private_va_range(unsigned long start, size_t size)
+{
+	unsigned long cur;
+
+	hyp_assert_lock_held(&pkvm_pgd_lock);
+
+	if (!start || start < __io_map_base)
+		return -EINVAL;
+
+	/* The allocated size is always a multiple of PAGE_SIZE */
+	cur = start + PAGE_ALIGN(size);
+
+	/* Are we overflowing on the vmemmap ? */
+	if (cur > __hyp_vmemmap)
+		return -ENOMEM;
+
+	__io_map_base = cur;
+
+	return 0;
+}
+
+/**
+ * pkvm_alloc_private_va_range - Allocates a private VA range.
+ * @size:	The size of the VA range to reserve.
+ * @haddr:	The hypervisor virtual start address of the allocation.
+ *
+ * The private virtual address (VA) range is allocated above __io_map_base
+ * and aligned based on the order of @size.
+ *
+ * Return: 0 on success or negative error code on failure.
+ */
+int pkvm_alloc_private_va_range(size_t size, unsigned long *haddr)
+{
+	unsigned long addr;
+	int ret;
+
+	hyp_spin_lock(&pkvm_pgd_lock);
+	addr = __io_map_base;
+	ret = __pkvm_alloc_private_va_range(addr, size);
+	hyp_spin_unlock(&pkvm_pgd_lock);
+
+	*haddr = addr;
+
+	return ret;
+}
+
+int __pkvm_create_private_mapping(phys_addr_t phys, size_t size,
+				  enum kvm_pgtable_prot prot,
+				  unsigned long *haddr)
+{
+	unsigned long addr;
+	int err;
+
+	size = PAGE_ALIGN(size + offset_in_page(phys));
+	err = pkvm_alloc_private_va_range(size, &addr);
+	if (err)
+		return err;
+
+	err = __pkvm_create_mappings(addr, size, phys, prot);
+	if (err)
+		return err;
+
+	*haddr = addr + offset_in_page(phys);
+	return err;
+}
+
+int pkvm_create_mappings_locked(void *from, void *to, enum kvm_pgtable_prot prot)
+{
+	unsigned long start = (unsigned long)from;
+	unsigned long end = (unsigned long)to;
+	unsigned long virt_addr;
+	phys_addr_t phys;
+
+	hyp_assert_lock_held(&pkvm_pgd_lock);
+
+	start = start & PAGE_MASK;
+	end = PAGE_ALIGN(end);
+
+	for (virt_addr = start; virt_addr < end; virt_addr += PAGE_SIZE) {
+		int err;
+
+		phys = hyp_virt_to_phys((void *)virt_addr);
+		err = kvm_pgtable_hyp_map(&pkvm_pgtable, virt_addr, PAGE_SIZE,
+					  phys, prot);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+int pkvm_create_mappings(void *from, void *to, enum kvm_pgtable_prot prot)
+{
+	int ret;
+
+	hyp_spin_lock(&pkvm_pgd_lock);
+	ret = pkvm_create_mappings_locked(from, to, prot);
+	hyp_spin_unlock(&pkvm_pgd_lock);
+
+	return ret;
+}
+
+int hyp_back_vmemmap(phys_addr_t back)
+{
+	unsigned long i, start, size, end = 0;
+	int ret;
+
+	for (i = 0; i < hyp_memblock_nr; i++) {
+		start = hyp_memory[i].base;
+		start = ALIGN_DOWN((u64)hyp_phys_to_page(start), PAGE_SIZE);
+		/*
+		 * The beginning of the hyp_vmemmap region for the current
+		 * memblock may already be backed by the page backing the end
+		 * the previous region, so avoid mapping it twice.
+		 */
+		start = max(start, end);
+
+		end = hyp_memory[i].base + hyp_memory[i].size;
+		end = PAGE_ALIGN((u64)hyp_phys_to_page(end));
+		if (start >= end)
+			continue;
+
+		size = end - start;
+		ret = __pkvm_create_mappings(start, size, back, PAGE_HYP);
+		if (ret)
+			return ret;
+
+		memset(hyp_phys_to_virt(back), 0, size);
+		back += size;
+	}
+
+	return 0;
+}
+
+static void *__hyp_bp_vect_base;
+int pkvm_cpu_set_vector(enum arm64_hyp_spectre_vector slot)
+{
+	void *vector;
+
+	switch (slot) {
+	case HYP_VECTOR_DIRECT: {
+		vector = __kvm_hyp_vector;
+		break;
+	}
+	case HYP_VECTOR_SPECTRE_DIRECT: {
+		vector = __bp_harden_hyp_vecs;
+		break;
+	}
+	case HYP_VECTOR_INDIRECT:
+	case HYP_VECTOR_SPECTRE_INDIRECT: {
+		vector = (void *)__hyp_bp_vect_base;
+		break;
+	}
+	default:
+		return -EINVAL;
+	}
+
+	vector = __kvm_vector_slot2addr(vector, slot);
+	*this_cpu_ptr(&kvm_hyp_vector) = (unsigned long)vector;
+
+	return 0;
+}
+
+int hyp_map_vectors(void)
+{
+	phys_addr_t phys;
+	unsigned long bp_base;
+	int ret;
+
+	if (!kvm_system_needs_idmapped_vectors()) {
+		__hyp_bp_vect_base = __bp_harden_hyp_vecs;
+		return 0;
+	}
+
+	phys = __hyp_pa(__bp_harden_hyp_vecs);
+	ret = __pkvm_create_private_mapping(phys, __BP_HARDEN_HYP_VECS_SZ,
+					    PAGE_HYP_EXEC, &bp_base);
+	if (ret)
+		return ret;
+
+	__hyp_bp_vect_base = (void *)bp_base;
+
+	return 0;
+}
+
+static void *fixmap_map_slot(struct hyp_fixmap_slot *slot, phys_addr_t phys)
+{
+	kvm_pte_t pte, *ptep = slot->ptep;
+
+	pte = *ptep;
+	pte &= ~kvm_phys_to_pte(KVM_PHYS_INVALID);
+	pte |= kvm_phys_to_pte(phys) | KVM_PTE_VALID;
+	WRITE_ONCE(*ptep, pte);
+	dsb(ishst);
+
+	return (void *)slot->addr;
+}
+
+void *hyp_fixmap_map(phys_addr_t phys)
+{
+	return fixmap_map_slot(this_cpu_ptr(&fixmap_slots), phys) + offset_in_page(phys);
+}
+
+static void fixmap_clear_slot(struct hyp_fixmap_slot *slot)
+{
+	kvm_pte_t *ptep = slot->ptep;
+	u64 addr = slot->addr;
+	u32 level;
+
+	if (FIELD_GET(KVM_PTE_TYPE, *ptep) == KVM_PTE_TYPE_PAGE)
+		level = KVM_PGTABLE_LAST_LEVEL;
+	else
+		level = KVM_PGTABLE_LAST_LEVEL - 1; /* create_fixblock() guarantees PMD level */
+
+	WRITE_ONCE(*ptep, *ptep & ~KVM_PTE_VALID);
+
+	/*
+	 * Irritatingly, the architecture requires that we use inner-shareable
+	 * broadcast TLB invalidation here in case another CPU speculates
+	 * through our fixmap and decides to create an "amalagamation of the
+	 * values held in the TLB" due to the apparent lack of a
+	 * break-before-make sequence.
+	 *
+	 * https://lore.kernel.org/kvm/20221017115209.2099-1-will@kernel.org/T/#mf10dfbaf1eaef9274c581b81c53758918c1d0f03
+	 */
+	dsb(ishst);
+	__tlbi_level(vale2is, addr, level);
+	__tlbi_sync_s1ish_hyp();
+	isb();
+}
+
+void hyp_fixmap_unmap(void)
+{
+	fixmap_clear_slot(this_cpu_ptr(&fixmap_slots));
+}
+
+static int __create_fixmap_slot_cb(const struct kvm_pgtable_visit_ctx *ctx,
+				   enum kvm_pgtable_walk_flags visit)
+{
+	struct hyp_fixmap_slot *slot = (struct hyp_fixmap_slot *)ctx->arg;
+
+	if (!kvm_pte_valid(ctx->old) || (ctx->end - ctx->start) != kvm_granule_size(ctx->level))
+		return -EINVAL;
+
+	slot->addr = ctx->addr;
+	slot->ptep = ctx->ptep;
+
+	/*
+	 * Clear the PTE, but keep the page-table page refcount elevated to
+	 * prevent it from ever being freed. This lets us manipulate the PTEs
+	 * by hand safely without ever needing to allocate memory.
+	 */
+	fixmap_clear_slot(slot);
+
+	return 0;
+}
+
+static int create_fixmap_slot(u64 addr, u64 cpu)
+{
+	struct kvm_pgtable_walker walker = {
+		.cb	= __create_fixmap_slot_cb,
+		.flags	= KVM_PGTABLE_WALK_LEAF,
+		.arg	= per_cpu_ptr(&fixmap_slots, cpu),
+	};
+
+	return kvm_pgtable_walk(&pkvm_pgtable, addr, PAGE_SIZE, &walker);
+}
+
+#if PAGE_SHIFT < 16
+#define HAS_FIXBLOCK
+static struct hyp_fixmap_slot hyp_fixblock_slot;
+static DEFINE_HYP_SPINLOCK(hyp_fixblock_lock);
+#endif
+
+static int create_fixblock(void)
+{
+#ifdef HAS_FIXBLOCK
+	struct kvm_pgtable_walker walker = {
+		.cb	= __create_fixmap_slot_cb,
+		.flags	= KVM_PGTABLE_WALK_LEAF,
+		.arg	= &hyp_fixblock_slot,
+	};
+	unsigned long addr;
+	phys_addr_t phys;
+	int ret, i;
+
+	/* Find a RAM phys address, PMD aligned */
+	for (i = 0; i < hyp_memblock_nr; i++) {
+		phys = ALIGN(hyp_memory[i].base, PMD_SIZE);
+		if (phys + PMD_SIZE < (hyp_memory[i].base + hyp_memory[i].size))
+			break;
+	}
+
+	if (i >= hyp_memblock_nr)
+		return -EINVAL;
+
+	hyp_spin_lock(&pkvm_pgd_lock);
+	addr = ALIGN(__io_map_base, PMD_SIZE);
+	ret = __pkvm_alloc_private_va_range(addr, PMD_SIZE);
+	if (ret)
+		goto unlock;
+
+	ret = kvm_pgtable_hyp_map(&pkvm_pgtable, addr, PMD_SIZE, phys, PAGE_HYP);
+	if (ret)
+		goto unlock;
+
+	ret = kvm_pgtable_walk(&pkvm_pgtable, addr, PMD_SIZE, &walker);
+
+unlock:
+	hyp_spin_unlock(&pkvm_pgd_lock);
+
+	return ret;
+#else
+	return 0;
+#endif
+}
+
+void *hyp_fixblock_map(phys_addr_t phys, size_t *size)
+{
+#ifdef HAS_FIXBLOCK
+	*size = PMD_SIZE;
+	hyp_spin_lock(&hyp_fixblock_lock);
+	return fixmap_map_slot(&hyp_fixblock_slot, phys) + offset_in_page(phys);
+#else
+	*size = PAGE_SIZE;
+	return hyp_fixmap_map(phys);
+#endif
+}
+
+void hyp_fixblock_unmap(void)
+{
+#ifdef HAS_FIXBLOCK
+	fixmap_clear_slot(&hyp_fixblock_slot);
+	hyp_spin_unlock(&hyp_fixblock_lock);
+#else
+	hyp_fixmap_unmap();
+#endif
+}
+
+int hyp_create_fixmap(void)
+{
+	unsigned long addr, i;
+	int ret;
+
+	for (i = 0; i < hyp_nr_cpus; i++) {
+		ret = pkvm_alloc_private_va_range(PAGE_SIZE, &addr);
+		if (ret)
+			return ret;
+
+		ret = kvm_pgtable_hyp_map(&pkvm_pgtable, addr, PAGE_SIZE,
+					  __hyp_pa(__hyp_bss_start), PAGE_HYP);
+		if (ret)
+			return ret;
+
+		ret = create_fixmap_slot(addr, i);
+		if (ret)
+			return ret;
+	}
+
+	return create_fixblock();
+}
+
+int hyp_create_idmap(u32 hyp_va_bits)
+{
+	unsigned long start, end;
+
+	start = hyp_virt_to_phys((void *)__hyp_idmap_text_start);
+	start = ALIGN_DOWN(start, PAGE_SIZE);
+
+	end = hyp_virt_to_phys((void *)__hyp_idmap_text_end);
+	end = ALIGN(end, PAGE_SIZE);
+
+	/*
+	 * One half of the VA space is reserved to linearly map portions of
+	 * memory -- see va_layout.c for more details. The other half of the VA
+	 * space contains the trampoline page, and needs some care. Split that
+	 * second half in two and find the quarter of VA space not conflicting
+	 * with the idmap to place the IOs and the vmemmap. IOs use the lower
+	 * half of the quarter and the vmemmap the upper half.
+	 */
+	__io_map_base = start & BIT(hyp_va_bits - 2);
+	__io_map_base ^= BIT(hyp_va_bits - 2);
+	__hyp_vmemmap = __io_map_base | BIT(hyp_va_bits - 3);
+
+	return __pkvm_create_mappings(start, end - start, start, PAGE_HYP_EXEC);
+}
+
+int pkvm_create_stack(phys_addr_t phys, unsigned long *haddr)
+{
+	unsigned long addr, prev_base;
+	size_t size;
+	int ret;
+
+	hyp_spin_lock(&pkvm_pgd_lock);
+
+	prev_base = __io_map_base;
+	/*
+	 * Efficient stack verification using the NVHE_STACK_SHIFT bit implies
+	 * an alignment of our allocation on the order of the size.
+	 */
+	size = NVHE_STACK_SIZE * 2;
+	addr = ALIGN(__io_map_base, size);
+
+	ret = __pkvm_alloc_private_va_range(addr, size);
+	if (!ret) {
+		/*
+		 * Since the stack grows downwards, map the stack to the page
+		 * at the higher address and leave the lower guard page
+		 * unbacked.
+		 *
+		 * Any valid stack address now has the NVHE_STACK_SHIFT bit as 1
+		 * and addresses corresponding to the guard page have the
+		 * NVHE_STACK_SHIFT bit as 0 - this is used for overflow detection.
+		 */
+		ret = kvm_pgtable_hyp_map(&pkvm_pgtable, addr + NVHE_STACK_SIZE,
+					  NVHE_STACK_SIZE, phys, PAGE_HYP);
+		if (ret)
+			__io_map_base = prev_base;
+	}
+	hyp_spin_unlock(&pkvm_pgd_lock);
+
+	*haddr = addr + size;
+
+	return ret;
+}
+
+static void *admit_host_page(void *arg)
+{
+	struct kvm_hyp_memcache *host_mc = arg;
+
+	if (!host_mc->nr_pages)
+		return NULL;
+
+	/*
+	 * The host still owns the pages in its memcache, so we need to go
+	 * through a full host-to-hyp donation cycle to change it. Fortunately,
+	 * __pkvm_host_donate_hyp() takes care of races for us, so if it
+	 * succeeds we're good to go.
+	 */
+	if (__pkvm_host_donate_hyp(hyp_phys_to_pfn(host_mc->head), 1))
+		return NULL;
+
+	return pop_hyp_memcache(host_mc, hyp_phys_to_virt);
+}
+
+/* Refill our local memcache by popping pages from the one provided by the host. */
+int refill_memcache(struct kvm_hyp_memcache *mc, unsigned long min_pages,
+		    struct kvm_hyp_memcache *host_mc)
+{
+	struct kvm_hyp_memcache tmp = *host_mc;
+	int ret;
+
+	ret =  __topup_hyp_memcache(mc, min_pages, admit_host_page,
+				    hyp_virt_to_phys, &tmp);
+	*host_mc = tmp;
+
+	return ret;
+}

linux-7.2.3/arch/arm64/kvm/hyp/nvhe/page_alloc.c

@@ -0,0 +1,259 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2020 Google LLC
+ * Author: Quentin Perret <qperret@google.com>
+ */
+
+#include <asm/kvm_hyp.h>
+#include <nvhe/gfp.h>
+
+u64 __hyp_vmemmap;
+
+/*
+ * Index the hyp_vmemmap to find a potential buddy page, but make no assumption
+ * about its current state.
+ *
+ * Example buddy-tree for a 4-pages physically contiguous pool:
+ *
+ *                 o : Page 3
+ *                /
+ *               o-o : Page 2
+ *              /
+ *             /   o : Page 1
+ *            /   /
+ *           o---o-o : Page 0
+ *    Order  2   1 0
+ *
+ * Example of requests on this pool:
+ *   __find_buddy_nocheck(pool, page 0, order 0) => page 1
+ *   __find_buddy_nocheck(pool, page 0, order 1) => page 2
+ *   __find_buddy_nocheck(pool, page 1, order 0) => page 0
+ *   __find_buddy_nocheck(pool, page 2, order 0) => page 3
+ */
+static struct hyp_page *__find_buddy_nocheck(struct hyp_pool *pool,
+					     struct hyp_page *p,
+					     u8 order)
+{
+	phys_addr_t addr = hyp_page_to_phys(p);
+
+	addr ^= (PAGE_SIZE << order);
+
+	/*
+	 * Don't return a page outside the pool range -- it belongs to
+	 * something else and may not be mapped in hyp_vmemmap.
+	 */
+	if (addr < pool->range_start || addr >= pool->range_end)
+		return NULL;
+
+	return hyp_phys_to_page(addr);
+}
+
+/* Find a buddy page currently available for allocation */
+static struct hyp_page *__find_buddy_avail(struct hyp_pool *pool,
+					   struct hyp_page *p,
+					   u8 order)
+{
+	struct hyp_page *buddy = __find_buddy_nocheck(pool, p, order);
+
+	if (!buddy || buddy->order != order || buddy->refcount)
+		return NULL;
+
+	return buddy;
+
+}
+
+/*
+ * Pages that are available for allocation are tracked in free-lists, so we use
+ * the pages themselves to store the list nodes to avoid wasting space. As the
+ * allocator always returns zeroed pages (which are zeroed on the hyp_put_page()
+ * path to optimize allocation speed), we also need to clean-up the list node in
+ * each page when we take it out of the list.
+ */
+static inline void page_remove_from_list(struct hyp_page *p)
+{
+	struct list_head *node = hyp_page_to_virt(p);
+
+	__list_del_entry(node);
+	memset(node, 0, sizeof(*node));
+}
+
+static inline void page_add_to_list(struct hyp_page *p, struct list_head *head)
+{
+	struct list_head *node = hyp_page_to_virt(p);
+
+	INIT_LIST_HEAD(node);
+	list_add_tail(node, head);
+}
+
+static inline struct hyp_page *node_to_page(struct list_head *node)
+{
+	return hyp_virt_to_page(node);
+}
+
+static void __hyp_attach_page(struct hyp_pool *pool,
+			      struct hyp_page *p)
+{
+	phys_addr_t phys = hyp_page_to_phys(p);
+	struct hyp_page *buddy;
+	bool coalesce = true;
+	u8 order = p->order;
+
+	/*
+	 * 'external' pages are never coalesced and their ->order field
+	 * untrusted as they bypass hyp_pool_init(). Enforce order-0.
+	 */
+	if (phys < pool->range_start || phys >= pool->range_end) {
+		order = 0;
+		coalesce = false;
+	}
+
+	memset(hyp_page_to_virt(p), 0, PAGE_SIZE << order);
+
+	if (!coalesce)
+		goto insert;
+
+	/*
+	 * Only the first struct hyp_page of a high-order page (otherwise known
+	 * as the 'head') should have p->order set. The non-head pages should
+	 * have p->order = HYP_NO_ORDER. Here @p may no longer be the head
+	 * after coalescing, so make sure to mark it HYP_NO_ORDER proactively.
+	 */
+	p->order = HYP_NO_ORDER;
+	for (; (order + 1) <= pool->max_order; order++) {
+		buddy = __find_buddy_avail(pool, p, order);
+		if (!buddy)
+			break;
+
+		/* Take the buddy out of its list, and coalesce with @p */
+		page_remove_from_list(buddy);
+		buddy->order = HYP_NO_ORDER;
+		p = min(p, buddy);
+	}
+
+insert:
+	/* Mark the new head, and insert it */
+	p->order = order;
+	page_add_to_list(p, &pool->free_area[order]);
+}
+
+static struct hyp_page *__hyp_extract_page(struct hyp_pool *pool,
+					   struct hyp_page *p,
+					   u8 order)
+{
+	struct hyp_page *buddy;
+
+	page_remove_from_list(p);
+	while (p->order > order) {
+		/*
+		 * The buddy of order n - 1 currently has HYP_NO_ORDER as it
+		 * is covered by a higher-level page (whose head is @p). Use
+		 * __find_buddy_nocheck() to find it and inject it in the
+		 * free_list[n - 1], effectively splitting @p in half.
+		 */
+		p->order--;
+		buddy = __find_buddy_nocheck(pool, p, p->order);
+		buddy->order = p->order;
+		page_add_to_list(buddy, &pool->free_area[buddy->order]);
+	}
+
+	return p;
+}
+
+static void __hyp_put_page(struct hyp_pool *pool, struct hyp_page *p)
+{
+	if (hyp_page_ref_dec_and_test(p))
+		__hyp_attach_page(pool, p);
+}
+
+/*
+ * Changes to the buddy tree and page refcounts must be done with the hyp_pool
+ * lock held. If a refcount change requires an update to the buddy tree (e.g.
+ * hyp_put_page()), both operations must be done within the same critical
+ * section to guarantee transient states (e.g. a page with null refcount but
+ * not yet attached to a free list) can't be observed by well-behaved readers.
+ */
+void hyp_put_page(struct hyp_pool *pool, void *addr)
+{
+	struct hyp_page *p = hyp_virt_to_page(addr);
+
+	hyp_spin_lock(&pool->lock);
+	__hyp_put_page(pool, p);
+	hyp_spin_unlock(&pool->lock);
+}
+
+void hyp_get_page(struct hyp_pool *pool, void *addr)
+{
+	struct hyp_page *p = hyp_virt_to_page(addr);
+
+	hyp_spin_lock(&pool->lock);
+	hyp_page_ref_inc(p);
+	hyp_spin_unlock(&pool->lock);
+}
+
+void hyp_split_page(struct hyp_page *p)
+{
+	u8 order = p->order;
+	unsigned int i;
+
+	p->order = 0;
+	for (i = 1; i < (1 << order); i++) {
+		struct hyp_page *tail = p + i;
+
+		tail->order = 0;
+		hyp_set_page_refcounted(tail);
+	}
+}
+
+void *hyp_alloc_pages(struct hyp_pool *pool, u8 order)
+{
+	struct hyp_page *p;
+	u8 i = order;
+
+	hyp_spin_lock(&pool->lock);
+
+	/* Look for a high-enough-order page */
+	while (i <= pool->max_order && list_empty(&pool->free_area[i]))
+		i++;
+	if (i > pool->max_order) {
+		hyp_spin_unlock(&pool->lock);
+		return NULL;
+	}
+
+	/* Extract it from the tree at the right order */
+	p = node_to_page(pool->free_area[i].next);
+	p = __hyp_extract_page(pool, p, order);
+
+	hyp_set_page_refcounted(p);
+	hyp_spin_unlock(&pool->lock);
+
+	return hyp_page_to_virt(p);
+}
+
+int hyp_pool_init(struct hyp_pool *pool, u64 pfn, unsigned int nr_pages,
+		  unsigned int reserved_pages)
+{
+	phys_addr_t phys = hyp_pfn_to_phys(pfn);
+	struct hyp_page *p;
+	int i;
+
+	hyp_spin_lock_init(&pool->lock);
+	pool->max_order = min(MAX_PAGE_ORDER,
+			      get_order(nr_pages << PAGE_SHIFT));
+	for (i = 0; i <= pool->max_order; i++)
+		INIT_LIST_HEAD(&pool->free_area[i]);
+	pool->range_start = phys;
+	pool->range_end = phys + (nr_pages << PAGE_SHIFT);
+
+	/* Init the vmemmap portion */
+	p = hyp_phys_to_page(phys);
+	for (i = 0; i < nr_pages; i++) {
+		hyp_set_page_refcounted(&p[i]);
+		p[i].order = 0;
+	}
+
+	/* Attach the unused pages to the buddy tree */
+	for (i = reserved_pages; i < nr_pages; i++)
+		__hyp_put_page(pool, &p[i]);
+
+	return 0;
+}

linux-7.2.3/arch/arm64/kvm/hyp/nvhe/psci-relay.c

@@ -0,0 +1,321 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2020 - Google LLC
+ * Author: David Brazdil <dbrazdil@google.com>
+ */
+
+#include <asm/kvm_asm.h>
+#include <asm/kvm_hyp.h>
+#include <asm/kvm_hypevents.h>
+#include <asm/kvm_mmu.h>
+#include <linux/kvm_host.h>
+#include <uapi/linux/psci.h>
+
+#include <nvhe/arm-smccc.h>
+#include <nvhe/memory.h>
+#include <nvhe/trap_handler.h>
+
+void kvm_hyp_cpu_entry(unsigned long r0);
+void kvm_hyp_cpu_resume(unsigned long r0);
+
+void __noreturn __host_enter(struct kvm_cpu_context *host_ctxt);
+
+/* Config options set by the host. */
+struct kvm_host_psci_config __ro_after_init kvm_host_psci_config;
+
+#define INVALID_CPU_ID	UINT_MAX
+
+struct psci_boot_args {
+	atomic_t lock;
+	unsigned long pc;
+	unsigned long r0;
+};
+
+#define PSCI_BOOT_ARGS_UNLOCKED		0
+#define PSCI_BOOT_ARGS_LOCKED		1
+
+#define PSCI_BOOT_ARGS_INIT					\
+	((struct psci_boot_args){				\
+		.lock = ATOMIC_INIT(PSCI_BOOT_ARGS_UNLOCKED),	\
+	})
+
+static DEFINE_PER_CPU(struct psci_boot_args, cpu_on_args) = PSCI_BOOT_ARGS_INIT;
+static DEFINE_PER_CPU(struct psci_boot_args, suspend_args) = PSCI_BOOT_ARGS_INIT;
+
+#define	is_psci_0_1(what, func_id)					\
+	(kvm_host_psci_config.psci_0_1_ ## what ## _implemented &&	\
+	 (func_id) == kvm_host_psci_config.function_ids_0_1.what)
+
+static bool is_psci_0_1_call(u64 func_id)
+{
+	return (is_psci_0_1(cpu_suspend, func_id) ||
+		is_psci_0_1(cpu_on, func_id) ||
+		is_psci_0_1(cpu_off, func_id) ||
+		is_psci_0_1(migrate, func_id));
+}
+
+static bool is_psci_0_2_call(u64 func_id)
+{
+	/* SMCCC reserves IDs 0x00-1F with the given 32/64-bit base for PSCI. */
+	return (PSCI_0_2_FN(0) <= func_id && func_id <= PSCI_0_2_FN(31)) ||
+	       (PSCI_0_2_FN64(0) <= func_id && func_id <= PSCI_0_2_FN64(31));
+}
+
+static unsigned long psci_call(unsigned long fn, unsigned long arg0,
+			       unsigned long arg1, unsigned long arg2)
+{
+	struct arm_smccc_res res;
+
+	hyp_smccc_1_1_smc(fn, arg0, arg1, arg2, &res);
+	return res.a0;
+}
+
+static unsigned long psci_forward(struct kvm_cpu_context *host_ctxt)
+{
+	return psci_call(cpu_reg(host_ctxt, 0), cpu_reg(host_ctxt, 1),
+			 cpu_reg(host_ctxt, 2), cpu_reg(host_ctxt, 3));
+}
+
+static unsigned int find_cpu_id(u64 mpidr)
+{
+	unsigned int i;
+
+	/* Reject invalid MPIDRs */
+	if (mpidr & ~MPIDR_HWID_BITMASK)
+		return INVALID_CPU_ID;
+
+	for (i = 0; i < NR_CPUS; i++) {
+		if (cpu_logical_map(i) == mpidr)
+			return i;
+	}
+
+	return INVALID_CPU_ID;
+}
+
+static __always_inline bool try_acquire_boot_args(struct psci_boot_args *args)
+{
+	return atomic_cmpxchg_acquire(&args->lock,
+				      PSCI_BOOT_ARGS_UNLOCKED,
+				      PSCI_BOOT_ARGS_LOCKED) ==
+		PSCI_BOOT_ARGS_UNLOCKED;
+}
+
+static __always_inline void release_boot_args(struct psci_boot_args *args)
+{
+	atomic_set_release(&args->lock, PSCI_BOOT_ARGS_UNLOCKED);
+}
+
+static int psci_cpu_on(u64 func_id, struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(u64, mpidr, host_ctxt, 1);
+	DECLARE_REG(unsigned long, pc, host_ctxt, 2);
+	DECLARE_REG(unsigned long, r0, host_ctxt, 3);
+
+	unsigned int cpu_id;
+	struct psci_boot_args *boot_args;
+	struct kvm_nvhe_init_params *init_params;
+	int ret;
+
+	/*
+	 * Find the logical CPU ID for the given MPIDR. The search set is
+	 * the set of CPUs that were online at the point of KVM initialization.
+	 * Booting other CPUs is rejected because their cpufeatures were not
+	 * checked against the finalized capabilities. This could be relaxed
+	 * by doing the feature checks in hyp.
+	 */
+	cpu_id = find_cpu_id(mpidr);
+	if (cpu_id == INVALID_CPU_ID)
+		return PSCI_RET_INVALID_PARAMS;
+
+	boot_args = per_cpu_ptr(&cpu_on_args, cpu_id);
+	init_params = per_cpu_ptr(&kvm_init_params, cpu_id);
+
+	/* Check if the target CPU is already being booted. */
+	if (!try_acquire_boot_args(boot_args))
+		return PSCI_RET_ALREADY_ON;
+
+	boot_args->pc = pc;
+	boot_args->r0 = r0;
+	wmb();
+
+	ret = psci_call(func_id, mpidr,
+			__hyp_pa(&kvm_hyp_cpu_entry),
+			__hyp_pa(init_params));
+
+	/* If successful, the lock will be released by the target CPU. */
+	if (ret != PSCI_RET_SUCCESS)
+		release_boot_args(boot_args);
+
+	return ret;
+}
+
+static int psci_cpu_suspend(u64 func_id, struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(u64, power_state, host_ctxt, 1);
+	DECLARE_REG(unsigned long, pc, host_ctxt, 2);
+	DECLARE_REG(unsigned long, r0, host_ctxt, 3);
+
+	struct psci_boot_args *boot_args;
+	struct kvm_nvhe_init_params *init_params;
+
+	boot_args = this_cpu_ptr(&suspend_args);
+	init_params = this_cpu_ptr(&kvm_init_params);
+
+	/*
+	 * No need to acquire a lock before writing to boot_args because a core
+	 * can only suspend itself. Racy CPU_ON calls use a separate struct.
+	 */
+	boot_args->pc = pc;
+	boot_args->r0 = r0;
+
+	/*
+	 * Will either return if shallow sleep state, or wake up into the entry
+	 * point if it is a deep sleep state.
+	 */
+	return psci_call(func_id, power_state,
+			 __hyp_pa(&kvm_hyp_cpu_resume),
+			 __hyp_pa(init_params));
+}
+
+static int psci_system_suspend(u64 func_id, struct kvm_cpu_context *host_ctxt)
+{
+	DECLARE_REG(unsigned long, pc, host_ctxt, 1);
+	DECLARE_REG(unsigned long, r0, host_ctxt, 2);
+
+	struct psci_boot_args *boot_args;
+	struct kvm_nvhe_init_params *init_params;
+
+	boot_args = this_cpu_ptr(&suspend_args);
+	init_params = this_cpu_ptr(&kvm_init_params);
+
+	/*
+	 * No need to acquire a lock before writing to boot_args because a core
+	 * can only suspend itself. Racy CPU_ON calls use a separate struct.
+	 */
+	boot_args->pc = pc;
+	boot_args->r0 = r0;
+
+	/* Will only return on error. */
+	return psci_call(func_id,
+			 __hyp_pa(&kvm_hyp_cpu_resume),
+			 __hyp_pa(init_params), 0);
+}
+
+static void __noreturn __kvm_host_psci_cpu_entry(unsigned long pc, unsigned long r0)
+{
+	struct kvm_cpu_context *host_ctxt = host_data_ptr(host_ctxt);
+
+	trace_hyp_enter(host_ctxt, HYP_REASON_PSCI);
+
+	cpu_reg(host_ctxt, 0) = r0;
+	write_sysreg_el2(pc, SYS_ELR);
+
+	write_sysreg_el1(INIT_SCTLR_EL1_MMU_OFF, SYS_SCTLR);
+	write_sysreg(INIT_PSTATE_EL1, SPSR_EL2);
+
+	trace_hyp_exit(host_ctxt, HYP_REASON_PSCI);
+	__host_enter(host_ctxt);
+}
+
+asmlinkage void __noreturn __kvm_host_psci_cpu_on_entry(void)
+{
+	struct psci_boot_args *boot_args = this_cpu_ptr(&cpu_on_args);
+	unsigned long pc, r0;
+
+	pc = READ_ONCE(boot_args->pc);
+	r0 = READ_ONCE(boot_args->r0);
+
+	release_boot_args(boot_args);
+
+	__kvm_host_psci_cpu_entry(pc, r0);
+}
+
+asmlinkage void __noreturn __kvm_host_psci_cpu_resume_entry(void)
+{
+	struct psci_boot_args *boot_args = this_cpu_ptr(&suspend_args);
+
+	__kvm_host_psci_cpu_entry(boot_args->pc, boot_args->r0);
+}
+
+static unsigned long psci_0_1_handler(u64 func_id, struct kvm_cpu_context *host_ctxt)
+{
+	if (is_psci_0_1(cpu_off, func_id) || is_psci_0_1(migrate, func_id))
+		return psci_forward(host_ctxt);
+	if (is_psci_0_1(cpu_on, func_id))
+		return psci_cpu_on(func_id, host_ctxt);
+	if (is_psci_0_1(cpu_suspend, func_id))
+		return psci_cpu_suspend(func_id, host_ctxt);
+
+	return PSCI_RET_NOT_SUPPORTED;
+}
+
+static unsigned long psci_0_2_handler(u64 func_id, struct kvm_cpu_context *host_ctxt)
+{
+	switch (func_id) {
+	case PSCI_0_2_FN_PSCI_VERSION:
+	case PSCI_0_2_FN_CPU_OFF:
+	case PSCI_0_2_FN64_AFFINITY_INFO:
+	case PSCI_0_2_FN64_MIGRATE:
+	case PSCI_0_2_FN_MIGRATE_INFO_TYPE:
+	case PSCI_0_2_FN64_MIGRATE_INFO_UP_CPU:
+		return psci_forward(host_ctxt);
+	/*
+	 * SYSTEM_OFF/RESET should not return according to the spec.
+	 * Allow it so as to stay robust to broken firmware.
+	 */
+	case PSCI_0_2_FN_SYSTEM_OFF:
+	case PSCI_0_2_FN_SYSTEM_RESET:
+		return psci_forward(host_ctxt);
+	case PSCI_0_2_FN64_CPU_SUSPEND:
+		return psci_cpu_suspend(func_id, host_ctxt);
+	case PSCI_0_2_FN64_CPU_ON:
+		return psci_cpu_on(func_id, host_ctxt);
+	default:
+		return PSCI_RET_NOT_SUPPORTED;
+	}
+}
+
+static unsigned long psci_1_0_handler(u64 func_id, struct kvm_cpu_context *host_ctxt)
+{
+	switch (func_id) {
+	case PSCI_1_0_FN_PSCI_FEATURES:
+	case PSCI_1_0_FN_SET_SUSPEND_MODE:
+	case PSCI_1_1_FN64_SYSTEM_RESET2:
+	case PSCI_1_3_FN_SYSTEM_OFF2:
+	case PSCI_1_3_FN64_SYSTEM_OFF2:
+		return psci_forward(host_ctxt);
+	case PSCI_1_0_FN64_SYSTEM_SUSPEND:
+		return psci_system_suspend(func_id, host_ctxt);
+	default:
+		return psci_0_2_handler(func_id, host_ctxt);
+	}
+}
+
+bool kvm_host_psci_handler(struct kvm_cpu_context *host_ctxt, u32 func_id)
+{
+	unsigned long ret;
+
+	switch (kvm_host_psci_config.version) {
+	case PSCI_VERSION(0, 1):
+		if (!is_psci_0_1_call(func_id))
+			return false;
+		ret = psci_0_1_handler(func_id, host_ctxt);
+		break;
+	case PSCI_VERSION(0, 2):
+		if (!is_psci_0_2_call(func_id))
+			return false;
+		ret = psci_0_2_handler(func_id, host_ctxt);
+		break;
+	default:
+		if (!is_psci_0_2_call(func_id))
+			return false;
+		ret = psci_1_0_handler(func_id, host_ctxt);
+		break;
+	}
+
+	cpu_reg(host_ctxt, 0) = ret;
+	cpu_reg(host_ctxt, 1) = 0;
+	cpu_reg(host_ctxt, 2) = 0;
+	cpu_reg(host_ctxt, 3) = 0;
+	return true;
+}

linux-7.2.3/arch/arm64/kvm/hyp/nvhe/setup.c

@@ -0,0 +1,374 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2020 Google LLC
+ * Author: Quentin Perret <qperret@google.com>
+ */
+
+#include <linux/kvm_host.h>
+#include <asm/kvm_hyp.h>
+#include <asm/kvm_mmu.h>
+#include <asm/kvm_pgtable.h>
+#include <asm/kvm_pkvm.h>
+
+#include <nvhe/early_alloc.h>
+#include <nvhe/ffa.h>
+#include <nvhe/gfp.h>
+#include <nvhe/memory.h>
+#include <nvhe/mem_protect.h>
+#include <nvhe/mm.h>
+#include <nvhe/pkvm.h>
+#include <nvhe/trap_handler.h>
+
+unsigned long hyp_nr_cpus;
+
+#define hyp_percpu_size ((unsigned long)__per_cpu_end - \
+			 (unsigned long)__per_cpu_start)
+
+static void *vmemmap_base;
+static void *vm_table_base;
+static void *hyp_pgt_base;
+static void *host_s2_pgt_base;
+static void *selftest_base;
+static void *ffa_proxy_pages;
+static struct kvm_pgtable_mm_ops pkvm_pgtable_mm_ops;
+static struct hyp_pool hpool;
+
+static int divide_memory_pool(void *virt, unsigned long size)
+{
+	unsigned long nr_pages;
+
+	hyp_early_alloc_init(virt, size);
+
+	nr_pages = pkvm_selftest_pages();
+	selftest_base = hyp_early_alloc_contig(nr_pages);
+	if (nr_pages && !selftest_base)
+		return -ENOMEM;
+
+	nr_pages = hyp_vmemmap_pages(sizeof(struct hyp_page));
+	vmemmap_base = hyp_early_alloc_contig(nr_pages);
+	if (!vmemmap_base)
+		return -ENOMEM;
+
+	nr_pages = hyp_vm_table_pages();
+	vm_table_base = hyp_early_alloc_contig(nr_pages);
+	if (!vm_table_base)
+		return -ENOMEM;
+
+	nr_pages = hyp_s1_pgtable_pages();
+	hyp_pgt_base = hyp_early_alloc_contig(nr_pages);
+	if (!hyp_pgt_base)
+		return -ENOMEM;
+
+	nr_pages = host_s2_pgtable_pages();
+	host_s2_pgt_base = hyp_early_alloc_contig(nr_pages);
+	if (!host_s2_pgt_base)
+		return -ENOMEM;
+
+	nr_pages = hyp_ffa_proxy_pages();
+	ffa_proxy_pages = hyp_early_alloc_contig(nr_pages);
+	if (!ffa_proxy_pages)
+		return -ENOMEM;
+
+	return 0;
+}
+
+static int pkvm_create_host_sve_mappings(void)
+{
+	void *start, *end;
+	int ret, i;
+
+	if (!system_supports_sve())
+		return 0;
+
+	for (i = 0; i < hyp_nr_cpus; i++) {
+		struct kvm_host_data *host_data = per_cpu_ptr(&kvm_host_data, i);
+		struct arm64_sve_state *sve_regs = host_data->sve_regs;
+
+		start = kern_hyp_va(sve_regs);
+		end = start + PAGE_ALIGN(pkvm_host_sve_state_size());
+		ret = pkvm_create_mappings(start, end, PAGE_HYP);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static int recreate_hyp_mappings(phys_addr_t phys, unsigned long size,
+				 unsigned long *per_cpu_base,
+				 u32 hyp_va_bits)
+{
+	void *start, *end, *virt = hyp_phys_to_virt(phys);
+	unsigned long pgt_size = hyp_s1_pgtable_pages() << PAGE_SHIFT;
+	int ret, i;
+
+	/* Recreate the hyp page-table using the early page allocator */
+	hyp_early_alloc_init(hyp_pgt_base, pgt_size);
+	ret = kvm_pgtable_hyp_init(&pkvm_pgtable, hyp_va_bits,
+				   &hyp_early_alloc_mm_ops);
+	if (ret)
+		return ret;
+
+	ret = hyp_create_idmap(hyp_va_bits);
+	if (ret)
+		return ret;
+
+	ret = hyp_map_vectors();
+	if (ret)
+		return ret;
+
+	ret = hyp_back_vmemmap(hyp_virt_to_phys(vmemmap_base));
+	if (ret)
+		return ret;
+
+	ret = pkvm_create_mappings(__hyp_text_start, __hyp_text_end, PAGE_HYP_EXEC);
+	if (ret)
+		return ret;
+
+	ret = pkvm_create_mappings(__hyp_data_start, __hyp_data_end, PAGE_HYP);
+	if (ret)
+		return ret;
+
+	ret = pkvm_create_mappings(__hyp_rodata_start, __hyp_rodata_end, PAGE_HYP_RO);
+	if (ret)
+		return ret;
+
+	ret = pkvm_create_mappings(__hyp_bss_start, __hyp_bss_end, PAGE_HYP);
+	if (ret)
+		return ret;
+
+	ret = pkvm_create_mappings(virt, virt + size, PAGE_HYP);
+	if (ret)
+		return ret;
+
+	for (i = 0; i < hyp_nr_cpus; i++) {
+		struct kvm_nvhe_init_params *params = per_cpu_ptr(&kvm_init_params, i);
+
+		start = (void *)kern_hyp_va(per_cpu_base[i]);
+		end = start + PAGE_ALIGN(hyp_percpu_size);
+		ret = pkvm_create_mappings(start, end, PAGE_HYP);
+		if (ret)
+			return ret;
+
+		ret = pkvm_create_stack(params->stack_pa, &params->stack_hyp_va);
+		if (ret)
+			return ret;
+	}
+
+	return pkvm_create_host_sve_mappings();
+}
+
+static void update_nvhe_init_params(void)
+{
+	struct kvm_nvhe_init_params *params;
+	unsigned long i;
+
+	for (i = 0; i < hyp_nr_cpus; i++) {
+		params = per_cpu_ptr(&kvm_init_params, i);
+		params->pgd_pa = __hyp_pa(pkvm_pgtable.pgd);
+		dcache_clean_inval_poc((unsigned long)params,
+				    (unsigned long)params + sizeof(*params));
+	}
+}
+
+static void *hyp_zalloc_hyp_page(void *arg)
+{
+	return hyp_alloc_pages(&hpool, 0);
+}
+
+static void hpool_get_page(void *addr)
+{
+	hyp_get_page(&hpool, addr);
+}
+
+static void hpool_put_page(void *addr)
+{
+	hyp_put_page(&hpool, addr);
+}
+
+static int fix_host_ownership_walker(const struct kvm_pgtable_visit_ctx *ctx,
+				     enum kvm_pgtable_walk_flags visit)
+{
+	enum pkvm_page_state state;
+	struct hyp_page *page;
+	phys_addr_t phys;
+	enum kvm_pgtable_prot prot;
+
+	if (!kvm_pte_valid(ctx->old))
+		return 0;
+
+	if (ctx->level != KVM_PGTABLE_LAST_LEVEL)
+		return -EINVAL;
+
+	phys = kvm_pte_to_phys(ctx->old);
+	if (!addr_is_memory(phys))
+		return -EINVAL;
+
+	page = hyp_phys_to_page(phys);
+
+	/*
+	 * Adjust the host stage-2 mappings to match the ownership attributes
+	 * configured in the hypervisor stage-1, and make sure to propagate them
+	 * to the hyp_vmemmap state.
+	 */
+	prot = kvm_pgtable_hyp_pte_prot(ctx->old);
+	state = pkvm_getstate(prot);
+	switch (state) {
+	case PKVM_PAGE_OWNED:
+		set_hyp_state(page, PKVM_PAGE_OWNED);
+		/* hyp text is RO in the host stage-2 to be inspected on panic. */
+		if (prot == PAGE_HYP_EXEC) {
+			set_host_state(page, PKVM_NOPAGE);
+			return host_stage2_idmap_locked(phys, PAGE_SIZE, KVM_PGTABLE_PROT_R);
+		} else {
+			return host_stage2_set_owner_locked(phys, PAGE_SIZE, PKVM_ID_HYP);
+		}
+	case PKVM_PAGE_SHARED_OWNED:
+		set_hyp_state(page, PKVM_PAGE_SHARED_OWNED);
+		set_host_state(page, PKVM_PAGE_SHARED_BORROWED);
+		break;
+	case PKVM_PAGE_SHARED_BORROWED:
+		set_hyp_state(page, PKVM_PAGE_SHARED_BORROWED);
+		set_host_state(page, PKVM_PAGE_SHARED_OWNED);
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int fix_hyp_pgtable_refcnt_walker(const struct kvm_pgtable_visit_ctx *ctx,
+					 enum kvm_pgtable_walk_flags visit)
+{
+	/*
+	 * Fix-up the refcount for the page-table pages as the early allocator
+	 * was unable to access the hyp_vmemmap and so the buddy allocator has
+	 * initialised the refcount to '1'.
+	 */
+	if (kvm_pte_valid(ctx->old))
+		ctx->mm_ops->get_page(ctx->ptep);
+
+	return 0;
+}
+
+static int fix_host_ownership(void)
+{
+	struct kvm_pgtable_walker walker = {
+		.cb	= fix_host_ownership_walker,
+		.flags	= KVM_PGTABLE_WALK_LEAF,
+	};
+	int i, ret;
+
+	for (i = 0; i < hyp_memblock_nr; i++) {
+		struct memblock_region *reg = &hyp_memory[i];
+		u64 start = (u64)hyp_phys_to_virt(reg->base);
+
+		ret = kvm_pgtable_walk(&pkvm_pgtable, start, reg->size, &walker);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static int fix_hyp_pgtable_refcnt(void)
+{
+	struct kvm_pgtable_walker walker = {
+		.cb	= fix_hyp_pgtable_refcnt_walker,
+		.flags	= KVM_PGTABLE_WALK_LEAF | KVM_PGTABLE_WALK_TABLE_POST,
+		.arg	= pkvm_pgtable.mm_ops,
+	};
+
+	return kvm_pgtable_walk(&pkvm_pgtable, 0, BIT(pkvm_pgtable.ia_bits),
+				&walker);
+}
+
+void __noreturn __pkvm_init_finalise(void)
+{
+	struct kvm_cpu_context *host_ctxt = host_data_ptr(host_ctxt);
+	unsigned long nr_pages, reserved_pages, pfn;
+	int ret;
+
+	/* Now that the vmemmap is backed, install the full-fledged allocator */
+	pfn = hyp_virt_to_pfn(hyp_pgt_base);
+	nr_pages = hyp_s1_pgtable_pages();
+	reserved_pages = hyp_early_alloc_nr_used_pages();
+	ret = hyp_pool_init(&hpool, pfn, nr_pages, reserved_pages);
+	if (ret)
+		goto out;
+
+	ret = kvm_host_prepare_stage2(host_s2_pgt_base);
+	if (ret)
+		goto out;
+
+	pkvm_pgtable_mm_ops = (struct kvm_pgtable_mm_ops) {
+		.zalloc_page = hyp_zalloc_hyp_page,
+		.phys_to_virt = hyp_phys_to_virt,
+		.virt_to_phys = hyp_virt_to_phys,
+		.get_page = hpool_get_page,
+		.put_page = hpool_put_page,
+		.page_count = hyp_page_count,
+	};
+	pkvm_pgtable.mm_ops = &pkvm_pgtable_mm_ops;
+
+	ret = fix_hyp_pgtable_refcnt();
+	if (ret)
+		goto out;
+
+	ret = hyp_create_fixmap();
+	if (ret)
+		goto out;
+
+	ret = fix_host_ownership();
+	if (ret)
+		goto out;
+
+	ret = hyp_ffa_init(ffa_proxy_pages);
+	if (ret)
+		goto out;
+
+	pkvm_hyp_vm_table_init(vm_table_base);
+
+	pkvm_ownership_selftest(selftest_base);
+out:
+	/*
+	 * We tail-called to here from handle___pkvm_init() and will not return,
+	 * so make sure to propagate the return value to the host.
+	 */
+	cpu_reg(host_ctxt, 1) = ret;
+
+	__host_enter(host_ctxt);
+}
+
+int __pkvm_init(phys_addr_t phys, unsigned long size, unsigned long *per_cpu_base, u32 hyp_va_bits)
+{
+	struct kvm_nvhe_init_params *params;
+	void *virt = hyp_phys_to_virt(phys);
+	typeof(__pkvm_init_switch_pgd) *fn;
+	int ret;
+
+	BUG_ON(kvm_check_pvm_sysreg_table());
+
+	if (!PAGE_ALIGNED(phys) || !PAGE_ALIGNED(size))
+		return -EINVAL;
+
+	hyp_spin_lock_init(&pkvm_pgd_lock);
+
+	ret = divide_memory_pool(virt, size);
+	if (ret)
+		return ret;
+
+	ret = recreate_hyp_mappings(phys, size, per_cpu_base, hyp_va_bits);
+	if (ret)
+		return ret;
+
+	update_nvhe_init_params();
+
+	/* Jump in the idmap page to switch to the new page-tables */
+	params = this_cpu_ptr(&kvm_init_params);
+	fn = (typeof(fn))__hyp_pa(__pkvm_init_switch_pgd);
+	fn(params->pgd_pa, params->stack_hyp_va, __pkvm_init_finalise);
+
+	unreachable();
+}

linux-7.2.3/arch/arm64/kvm/hyp/nvhe/stacktrace.c

@@ -0,0 +1,158 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * KVM nVHE hypervisor stack tracing support.
+ *
+ * Copyright (C) 2022 Google LLC
+ */
+#include <asm/kvm_asm.h>
+#include <asm/kvm_hyp.h>
+#include <asm/memory.h>
+#include <asm/percpu.h>
+
+DEFINE_PER_CPU(unsigned long [OVERFLOW_STACK_SIZE/sizeof(long)], overflow_stack)
+	__aligned(16);
+
+DEFINE_PER_CPU(struct kvm_nvhe_stacktrace_info, kvm_stacktrace_info);
+
+/*
+ * hyp_prepare_backtrace - Prepare non-protected nVHE backtrace.
+ *
+ * @fp : frame pointer at which to start the unwinding.
+ * @pc : program counter at which to start the unwinding.
+ *
+ * Save the information needed by the host to unwind the non-protected
+ * nVHE hypervisor stack in EL1.
+ */
+static void hyp_prepare_backtrace(unsigned long fp, unsigned long pc)
+{
+	struct kvm_nvhe_stacktrace_info *stacktrace_info = this_cpu_ptr(&kvm_stacktrace_info);
+	struct kvm_nvhe_init_params *params = this_cpu_ptr(&kvm_init_params);
+
+	stacktrace_info->stack_base = (unsigned long)(params->stack_hyp_va - NVHE_STACK_SIZE);
+	stacktrace_info->overflow_stack_base = (unsigned long)this_cpu_ptr(overflow_stack);
+	stacktrace_info->fp = fp;
+	stacktrace_info->pc = pc;
+}
+
+#ifdef CONFIG_PKVM_STACKTRACE
+#include <asm/stacktrace/nvhe.h>
+
+DEFINE_PER_CPU(unsigned long [NVHE_STACKTRACE_SIZE/sizeof(long)], pkvm_stacktrace);
+
+static struct stack_info stackinfo_get_overflow(void)
+{
+	unsigned long low = (unsigned long)this_cpu_ptr(overflow_stack);
+	unsigned long high = low + OVERFLOW_STACK_SIZE;
+
+	return (struct stack_info) {
+		.low = low,
+		.high = high,
+	};
+}
+
+static struct stack_info stackinfo_get_hyp(void)
+{
+	struct kvm_nvhe_init_params *params = this_cpu_ptr(&kvm_init_params);
+	unsigned long high = params->stack_hyp_va;
+	unsigned long low = high - NVHE_STACK_SIZE;
+
+	return (struct stack_info) {
+		.low = low,
+		.high = high,
+	};
+}
+
+static int unwind_next(struct unwind_state *state)
+{
+	return unwind_next_frame_record(state);
+}
+
+static void notrace unwind(struct unwind_state *state,
+			   stack_trace_consume_fn consume_entry,
+			   void *cookie)
+{
+	while (1) {
+		int ret;
+
+		if (!consume_entry(cookie, state->pc))
+			break;
+		ret = unwind_next(state);
+		if (ret < 0)
+			break;
+	}
+}
+
+/*
+ * pkvm_save_backtrace_entry - Saves a protected nVHE HYP stacktrace entry
+ *
+ * @arg    : index of the entry in the stacktrace buffer
+ * @where  : the program counter corresponding to the stack frame
+ *
+ * Save the return address of a stack frame to the shared stacktrace buffer.
+ * The host can access this shared buffer from EL1 to dump the backtrace.
+ */
+static bool pkvm_save_backtrace_entry(void *arg, unsigned long where)
+{
+	unsigned long *stacktrace = this_cpu_ptr(pkvm_stacktrace);
+	int *idx = (int *)arg;
+
+	/*
+	 * Need 2 free slots: 1 for current entry and 1 for the
+	 * delimiter.
+	 */
+	if (*idx > ARRAY_SIZE(pkvm_stacktrace) - 2)
+		return false;
+
+	stacktrace[*idx] = where;
+	stacktrace[++*idx] = 0UL;
+
+	return true;
+}
+
+/*
+ * pkvm_save_backtrace - Saves the protected nVHE HYP stacktrace
+ *
+ * @fp : frame pointer at which to start the unwinding.
+ * @pc : program counter at which to start the unwinding.
+ *
+ * Save the unwinded stack addresses to the shared stacktrace buffer.
+ * The host can access this shared buffer from EL1 to dump the backtrace.
+ */
+static void pkvm_save_backtrace(unsigned long fp, unsigned long pc)
+{
+	struct stack_info stacks[] = {
+		stackinfo_get_overflow(),
+		stackinfo_get_hyp(),
+	};
+	struct unwind_state state = {
+		.stacks = stacks,
+		.nr_stacks = ARRAY_SIZE(stacks),
+	};
+	int idx = 0;
+
+	kvm_nvhe_unwind_init(&state, fp, pc);
+
+	unwind(&state, pkvm_save_backtrace_entry, &idx);
+}
+#else /* !CONFIG_PKVM_STACKTRACE */
+static void pkvm_save_backtrace(unsigned long fp, unsigned long pc)
+{
+}
+#endif /* CONFIG_PKVM_STACKTRACE */
+
+/*
+ * kvm_nvhe_prepare_backtrace - prepare to dump the nVHE backtrace
+ *
+ * @fp : frame pointer at which to start the unwinding.
+ * @pc : program counter at which to start the unwinding.
+ *
+ * Saves the information needed by the host to dump the nVHE hypervisor
+ * backtrace.
+ */
+void kvm_nvhe_prepare_backtrace(unsigned long fp, unsigned long pc)
+{
+	if (is_protected_kvm_enabled())
+		pkvm_save_backtrace(fp, pc);
+	else
+		hyp_prepare_backtrace(fp, pc);
+}

linux-7.2.3/arch/arm64/kvm/hyp/nvhe/switch.c

@@ -0,0 +1,409 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2015 - ARM Ltd
+ * Author: Marc Zyngier <marc.zyngier@arm.com>
+ */
+
+#include <hyp/switch.h>
+#include <hyp/sysreg-sr.h>
+
+#include <linux/kvm_host.h>
+#include <linux/types.h>
+#include <linux/jump_label.h>
+#include <uapi/linux/psci.h>
+
+#include <kvm/arm_psci.h>
+
+#include <asm/barrier.h>
+#include <asm/cpufeature.h>
+#include <asm/kprobes.h>
+#include <asm/kvm_asm.h>
+#include <asm/kvm_emulate.h>
+#include <asm/kvm_hyp.h>
+#include <asm/kvm_hypevents.h>
+#include <asm/kvm_mmu.h>
+#include <asm/fpsimd.h>
+#include <asm/debug-monitors.h>
+#include <asm/processor.h>
+
+#include <nvhe/mem_protect.h>
+
+/* Non-VHE specific context */
+DEFINE_PER_CPU(struct kvm_host_data, kvm_host_data);
+DEFINE_PER_CPU(struct kvm_cpu_context, kvm_hyp_ctxt);
+DEFINE_PER_CPU(unsigned long, kvm_hyp_vector);
+
+struct fgt_masks hfgrtr_masks;
+struct fgt_masks hfgwtr_masks;
+struct fgt_masks hfgitr_masks;
+struct fgt_masks hdfgrtr_masks;
+struct fgt_masks hdfgwtr_masks;
+struct fgt_masks hafgrtr_masks;
+struct fgt_masks hfgrtr2_masks;
+struct fgt_masks hfgwtr2_masks;
+struct fgt_masks hfgitr2_masks;
+struct fgt_masks hdfgrtr2_masks;
+struct fgt_masks hdfgwtr2_masks;
+struct fgt_masks ich_hfgrtr_masks;
+struct fgt_masks ich_hfgwtr_masks;
+struct fgt_masks ich_hfgitr_masks;
+
+extern void kvm_nvhe_prepare_backtrace(unsigned long fp, unsigned long pc);
+
+static void __activate_traps(struct kvm_vcpu *vcpu)
+{
+	___activate_traps(vcpu, vcpu->arch.hcr_el2);
+
+	*host_data_ptr(host_debug_state.mdcr_el2) = read_sysreg(mdcr_el2);
+	write_sysreg(vcpu->arch.mdcr_el2, mdcr_el2);
+
+	__activate_traps_common(vcpu);
+	__activate_cptr_traps(vcpu);
+
+	write_sysreg(__this_cpu_read(kvm_hyp_vector), vbar_el2);
+
+	if (cpus_have_final_cap(ARM64_WORKAROUND_SPECULATIVE_AT)) {
+		struct kvm_cpu_context *ctxt = &vcpu->arch.ctxt;
+
+		isb();
+		/*
+		 * At this stage, and thanks to the above isb(), S2 is
+		 * configured and enabled. We can now restore the guest's S1
+		 * configuration: SCTLR, and only then TCR.
+		 */
+		write_sysreg_el1(ctxt_sys_reg(ctxt, SCTLR_EL1),	SYS_SCTLR);
+		isb();
+		write_sysreg_el1(ctxt_sys_reg(ctxt, TCR_EL1),	SYS_TCR);
+	}
+}
+
+static void __deactivate_traps(struct kvm_vcpu *vcpu)
+{
+	extern char __kvm_hyp_host_vector[];
+
+	___deactivate_traps(vcpu);
+
+	if (cpus_have_final_cap(ARM64_WORKAROUND_SPECULATIVE_AT)) {
+		u64 val;
+
+		/*
+		 * Set the TCR and SCTLR registers in the exact opposite
+		 * sequence as __activate_traps (first prevent walks,
+		 * then force the MMU on). A generous sprinkling of isb()
+		 * ensure that things happen in this exact order.
+		 */
+		val = read_sysreg_el1(SYS_TCR);
+		write_sysreg_el1(val | TCR_EPD1_MASK | TCR_EPD0_MASK, SYS_TCR);
+		isb();
+		val = read_sysreg_el1(SYS_SCTLR);
+		write_sysreg_el1(val | SCTLR_ELx_M, SYS_SCTLR);
+		isb();
+	}
+
+	write_sysreg(*host_data_ptr(host_debug_state.mdcr_el2), mdcr_el2);
+
+	__deactivate_traps_common(vcpu);
+
+	write_sysreg_hcr(this_cpu_ptr(&kvm_init_params)->hcr_el2);
+
+	__deactivate_cptr_traps(vcpu);
+	write_sysreg(__kvm_hyp_host_vector, vbar_el2);
+}
+
+/* Save VGICv3 state on non-VHE systems */
+static void __hyp_vgic_save_state(struct kvm_vcpu *vcpu)
+{
+	if (vgic_is_v5(kern_hyp_va(vcpu->kvm))) {
+		__vgic_v5_save_state(&vcpu->arch.vgic_cpu.vgic_v5);
+		__vgic_v5_save_ppi_state(&vcpu->arch.vgic_cpu.vgic_v5);
+		return;
+	}
+
+	if (static_branch_unlikely(&kvm_vgic_global_state.gicv3_cpuif)) {
+		__vgic_v3_save_state(&vcpu->arch.vgic_cpu.vgic_v3);
+		__vgic_v3_deactivate_traps(&vcpu->arch.vgic_cpu.vgic_v3);
+	}
+}
+
+/* Restore VGICv3 state on non-VHE systems */
+static void __hyp_vgic_restore_state(struct kvm_vcpu *vcpu)
+{
+	if (vgic_is_v5(kern_hyp_va(vcpu->kvm))) {
+		__vgic_v5_restore_state(&vcpu->arch.vgic_cpu.vgic_v5);
+		__vgic_v5_restore_ppi_state(&vcpu->arch.vgic_cpu.vgic_v5);
+		return;
+	}
+
+	if (static_branch_unlikely(&kvm_vgic_global_state.gicv3_cpuif)) {
+		__vgic_v3_activate_traps(&vcpu->arch.vgic_cpu.vgic_v3);
+		__vgic_v3_restore_state(&vcpu->arch.vgic_cpu.vgic_v3);
+	}
+}
+
+/*
+ * Disable host events, enable guest events
+ */
+#ifdef CONFIG_HW_PERF_EVENTS
+static bool __pmu_switch_to_guest(struct kvm_vcpu *vcpu)
+{
+	struct kvm_pmu_events *pmu = &vcpu->arch.pmu.events;
+
+	if (pmu->events_host)
+		write_sysreg(pmu->events_host, pmcntenclr_el0);
+
+	if (pmu->events_guest)
+		write_sysreg(pmu->events_guest, pmcntenset_el0);
+
+	return (pmu->events_host || pmu->events_guest);
+}
+
+/*
+ * Disable guest events, enable host events
+ */
+static void __pmu_switch_to_host(struct kvm_vcpu *vcpu)
+{
+	struct kvm_pmu_events *pmu = &vcpu->arch.pmu.events;
+
+	if (pmu->events_guest)
+		write_sysreg(pmu->events_guest, pmcntenclr_el0);
+
+	if (pmu->events_host)
+		write_sysreg(pmu->events_host, pmcntenset_el0);
+}
+#else
+#define __pmu_switch_to_guest(v)	({ false; })
+#define __pmu_switch_to_host(v)		do {} while (0)
+#endif
+
+/*
+ * Handler for protected VM MSR, MRS or System instruction execution in AArch64.
+ *
+ * Returns true if the hypervisor has handled the exit, and control should go
+ * back to the guest, or false if it hasn't.
+ */
+static bool kvm_handle_pvm_sys64(struct kvm_vcpu *vcpu, u64 *exit_code)
+{
+	/*
+	 * Make sure we handle the exit for workarounds before the pKVM
+	 * handling, as the latter could decide to UNDEF.
+	 */
+	return (kvm_hyp_handle_sysreg(vcpu, exit_code) ||
+		kvm_handle_pvm_sysreg(vcpu, exit_code));
+}
+
+static const exit_handler_fn hyp_exit_handlers[] = {
+	[0 ... ESR_ELx_EC_MAX]		= NULL,
+	[ESR_ELx_EC_CP15_32]		= kvm_hyp_handle_cp15_32,
+	[ESR_ELx_EC_SYS64]		= kvm_hyp_handle_sysreg,
+	[ESR_ELx_EC_SVE]		= kvm_hyp_handle_fpsimd,
+	[ESR_ELx_EC_FP_ASIMD]		= kvm_hyp_handle_fpsimd,
+	[ESR_ELx_EC_IABT_LOW]		= kvm_hyp_handle_iabt_low,
+	[ESR_ELx_EC_DABT_LOW]		= kvm_hyp_handle_dabt_low,
+	[ESR_ELx_EC_WATCHPT_LOW]	= kvm_hyp_handle_watchpt_low,
+	[ESR_ELx_EC_MOPS]		= kvm_hyp_handle_mops,
+};
+
+static const exit_handler_fn pvm_exit_handlers[] = {
+	[0 ... ESR_ELx_EC_MAX]		= NULL,
+	[ESR_ELx_EC_HVC64]		= kvm_handle_pvm_hvc64,
+	[ESR_ELx_EC_SYS64]		= kvm_handle_pvm_sys64,
+	[ESR_ELx_EC_SVE]		= kvm_handle_pvm_restricted,
+	[ESR_ELx_EC_FP_ASIMD]		= kvm_hyp_handle_fpsimd,
+	[ESR_ELx_EC_IABT_LOW]		= kvm_hyp_handle_iabt_low,
+	[ESR_ELx_EC_DABT_LOW]		= kvm_hyp_handle_dabt_low,
+	[ESR_ELx_EC_WATCHPT_LOW]	= kvm_hyp_handle_watchpt_low,
+	[ESR_ELx_EC_MOPS]		= kvm_hyp_handle_mops,
+};
+
+static const exit_handler_fn *kvm_get_exit_handler_array(struct kvm_vcpu *vcpu)
+{
+	if (unlikely(vcpu_is_protected(vcpu)))
+		return pvm_exit_handlers;
+
+	return hyp_exit_handlers;
+}
+
+static inline bool fixup_guest_exit(struct kvm_vcpu *vcpu, u64 *exit_code)
+{
+	const exit_handler_fn *handlers = kvm_get_exit_handler_array(vcpu);
+
+	synchronize_vcpu_pstate(vcpu);
+
+	/*
+	 * Some guests (e.g., protected VMs) are not be allowed to run in
+	 * AArch32.  The ARMv8 architecture does not give the hypervisor a
+	 * mechanism to prevent a guest from dropping to AArch32 EL0 if
+	 * implemented by the CPU. If the hypervisor spots a guest in such a
+	 * state ensure it is handled, and don't trust the host to spot or fix
+	 * it.  The check below is based on the one in
+	 * kvm_arch_vcpu_ioctl_run().
+	 */
+	if (unlikely(vcpu_is_protected(vcpu) && vcpu_mode_is_32bit(vcpu))) {
+		/*
+		 * As we have caught the guest red-handed, decide that it isn't
+		 * fit for purpose anymore by making the vcpu invalid. The VMM
+		 * can try and fix it by re-initializing the vcpu with
+		 * KVM_ARM_VCPU_INIT, however, this is likely not possible for
+		 * protected VMs.
+		 */
+		vcpu_clear_flag(vcpu, VCPU_INITIALIZED);
+		*exit_code &= BIT(ARM_EXIT_WITH_SERROR_BIT);
+		*exit_code |= ARM_EXCEPTION_IL;
+	}
+
+	return __fixup_guest_exit(vcpu, exit_code, handlers);
+}
+
+/* Switch to the guest for legacy non-VHE systems */
+int __kvm_vcpu_run(struct kvm_vcpu *vcpu)
+{
+	struct kvm_cpu_context *host_ctxt;
+	struct kvm_cpu_context *guest_ctxt;
+	struct kvm_s2_mmu *mmu;
+	bool pmu_switch_needed;
+	u64 exit_code;
+
+	/*
+	 * Having IRQs masked via PMR when entering the guest means the GIC
+	 * will not signal the CPU of interrupts of lower priority, and the
+	 * only way to get out will be via guest exceptions.
+	 * Naturally, we want to avoid this.
+	 */
+	if (system_uses_irq_prio_masking()) {
+		gic_write_pmr(GIC_PRIO_IRQON | GIC_PRIO_PSR_I_SET);
+		pmr_sync();
+	}
+
+	host_ctxt = host_data_ptr(host_ctxt);
+	host_ctxt->__hyp_running_vcpu = vcpu;
+	guest_ctxt = &vcpu->arch.ctxt;
+
+	pmu_switch_needed = __pmu_switch_to_guest(vcpu);
+
+	__sysreg_save_state_nvhe(host_ctxt);
+	/*
+	 * We must flush and disable the SPE buffer for nVHE, as
+	 * the translation regime(EL1&0) is going to be loaded with
+	 * that of the guest. And we must do this before we change the
+	 * translation regime to EL2 (via MDCR_EL2_E2PB == 0) and
+	 * before we load guest Stage1.
+	 */
+	__debug_save_host_buffers_nvhe(vcpu);
+
+	/*
+	 * We're about to restore some new MMU state. Make sure
+	 * ongoing page-table walks that have started before we
+	 * trapped to EL2 have completed. This also synchronises the
+	 * above disabling of BRBE.
+	 *
+	 * See DDI0487I.a D8.1.5 "Out-of-context translation regimes",
+	 * rule R_LFHQG and subsequent information statements.
+	 */
+	dsb(nsh);
+
+	__kvm_adjust_pc(vcpu);
+
+	/*
+	 * We must restore the 32-bit state before the sysregs, thanks
+	 * to erratum #852523 (Cortex-A57) or #853709 (Cortex-A72).
+	 *
+	 * Also, and in order to be able to deal with erratum #1319537 (A57)
+	 * and #1319367 (A72), we must ensure that all VM-related sysreg are
+	 * restored before we enable S2 translation.
+	 */
+	__sysreg32_restore_state(vcpu);
+	__sysreg_restore_state_nvhe(guest_ctxt);
+
+	mmu = kern_hyp_va(vcpu->arch.hw_mmu);
+	__load_stage2(mmu);
+	__activate_traps(vcpu);
+
+	__hyp_vgic_restore_state(vcpu);
+	__timer_enable_traps(vcpu);
+
+	__debug_switch_to_guest(vcpu);
+
+	do {
+		trace_hyp_exit(host_ctxt, HYP_REASON_ERET_GUEST);
+
+		/* Jump in the fire! */
+		exit_code = __guest_enter(vcpu);
+
+		/* And we're baaack! */
+		trace_hyp_enter(host_ctxt, HYP_REASON_GUEST_EXIT);
+	} while (fixup_guest_exit(vcpu, &exit_code));
+
+	__sysreg_save_state_nvhe(guest_ctxt);
+	__sysreg32_save_state(vcpu);
+	__timer_disable_traps(vcpu);
+	__hyp_vgic_save_state(vcpu);
+
+	/*
+	 * Same thing as before the guest run: we're about to switch
+	 * the MMU context, so let's make sure we don't have any
+	 * ongoing EL1&0 translations.
+	 */
+	dsb(nsh);
+
+	__deactivate_traps(vcpu);
+	__load_host_stage2();
+
+	__sysreg_restore_state_nvhe(host_ctxt);
+
+	if (guest_owns_fp_regs())
+		__fpsimd_save_fpexc32(vcpu);
+
+	__debug_switch_to_host(vcpu);
+	/*
+	 * This must come after restoring the host sysregs, since a non-VHE
+	 * system may enable SPE here and make use of the TTBRs.
+	 */
+	__debug_restore_host_buffers_nvhe(vcpu);
+
+	if (pmu_switch_needed)
+		__pmu_switch_to_host(vcpu);
+
+	/* Returning to host will clear PSR.I, remask PMR if needed */
+	if (system_uses_irq_prio_masking())
+		gic_write_pmr(GIC_PRIO_IRQOFF);
+
+	host_ctxt->__hyp_running_vcpu = NULL;
+
+	return exit_code;
+}
+
+asmlinkage void __noreturn hyp_panic(void)
+{
+	u64 spsr = read_sysreg_el2(SYS_SPSR);
+	u64 elr = read_sysreg_el2(SYS_ELR);
+	u64 par = read_sysreg_par();
+	struct kvm_cpu_context *host_ctxt;
+	struct kvm_vcpu *vcpu;
+
+	host_ctxt = host_data_ptr(host_ctxt);
+	vcpu = host_ctxt->__hyp_running_vcpu;
+
+	if (vcpu) {
+		__timer_disable_traps(vcpu);
+		__deactivate_traps(vcpu);
+		__load_host_stage2();
+		__sysreg_restore_state_nvhe(host_ctxt);
+	}
+
+	/* Prepare to dump kvm nvhe hyp stacktrace */
+	kvm_nvhe_prepare_backtrace((unsigned long)__builtin_frame_address(0),
+				   _THIS_IP_);
+
+	__hyp_do_panic(host_ctxt, spsr, elr, par);
+	unreachable();
+}
+
+asmlinkage void __noreturn hyp_panic_bad_stack(void)
+{
+	hyp_panic();
+}
+
+asmlinkage void kvm_unexpected_el2_exception(void)
+{
+	__kvm_unexpected_el2_exception();
+}

linux-7.2.3/arch/arm64/kvm/hyp/nvhe/sys_regs.c

@@ -0,0 +1,616 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2021 Google LLC
+ * Author: Fuad Tabba <tabba@google.com>
+ */
+
+#include <linux/irqchip/arm-gic-v3.h>
+
+#include <asm/kvm_asm.h>
+#include <asm/kvm_mmu.h>
+
+#include <hyp/adjust_pc.h>
+
+#include <nvhe/pkvm.h>
+
+#include "../../sys_regs.h"
+
+/*
+ * Copies of the host's CPU features registers holding sanitized values at hyp.
+ */
+u64 id_aa64pfr0_el1_sys_val;
+u64 id_aa64pfr1_el1_sys_val;
+u64 id_aa64pfr2_el1_sys_val;
+u64 id_aa64isar0_el1_sys_val;
+u64 id_aa64isar1_el1_sys_val;
+u64 id_aa64isar2_el1_sys_val;
+u64 id_aa64mmfr0_el1_sys_val;
+u64 id_aa64mmfr1_el1_sys_val;
+u64 id_aa64mmfr2_el1_sys_val;
+u64 id_aa64smfr0_el1_sys_val;
+
+struct pvm_ftr_bits {
+	bool		sign;
+	u8		shift;
+	u8		width;
+	u8		max_val;
+	bool (*vm_supported)(const struct kvm *kvm);
+};
+
+#define __MAX_FEAT_FUNC(id, fld, max, func, sgn)				\
+	{									\
+		.sign = sgn,							\
+		.shift = id##_##fld##_SHIFT,					\
+		.width = id##_##fld##_WIDTH,					\
+		.max_val = id##_##fld##_##max,					\
+		.vm_supported = func,						\
+	}
+
+#define MAX_FEAT_FUNC(id, fld, max, func)					\
+	__MAX_FEAT_FUNC(id, fld, max, func, id##_##fld##_SIGNED)
+
+#define MAX_FEAT(id, fld, max)							\
+	MAX_FEAT_FUNC(id, fld, max, NULL)
+
+#define MAX_FEAT_ENUM(id, fld, max)						\
+	__MAX_FEAT_FUNC(id, fld, max, NULL, false)
+
+#define FEAT_END {	.width = 0,	}
+
+static bool vm_has_ptrauth(const struct kvm *kvm)
+{
+	if (!IS_ENABLED(CONFIG_ARM64_PTR_AUTH))
+		return false;
+
+	return (cpus_have_final_cap(ARM64_HAS_ADDRESS_AUTH) ||
+		cpus_have_final_cap(ARM64_HAS_GENERIC_AUTH)) &&
+		kvm_vcpu_has_feature(kvm, KVM_ARM_VCPU_PTRAUTH_GENERIC);
+}
+
+static bool vm_has_sve(const struct kvm *kvm)
+{
+	return system_supports_sve() && kvm_vcpu_has_feature(kvm, KVM_ARM_VCPU_SVE);
+}
+
+/*
+ * Definitions for features to be allowed or restricted for protected guests.
+ *
+ * Each field in the masks represents the highest supported value for the
+ * feature. If a feature field is not present, it is not supported. Moreover,
+ * these are used to generate the guest's view of the feature registers.
+ *
+ * The approach for protected VMs is to at least support features that are:
+ * - Needed by common Linux distributions (e.g., floating point)
+ * - Trivial to support, e.g., supporting the feature does not introduce or
+ * require tracking of additional state in KVM
+ * - Cannot be trapped or prevent the guest from using anyway
+ */
+
+static const struct pvm_ftr_bits pvmid_aa64pfr0[] = {
+	MAX_FEAT(ID_AA64PFR0_EL1, EL0, IMP),
+	MAX_FEAT(ID_AA64PFR0_EL1, EL1, IMP),
+	MAX_FEAT(ID_AA64PFR0_EL1, EL2, IMP),
+	MAX_FEAT(ID_AA64PFR0_EL1, EL3, IMP),
+	MAX_FEAT(ID_AA64PFR0_EL1, FP, FP16),
+	MAX_FEAT(ID_AA64PFR0_EL1, AdvSIMD, FP16),
+	MAX_FEAT(ID_AA64PFR0_EL1, GIC, IMP),
+	MAX_FEAT_FUNC(ID_AA64PFR0_EL1, SVE, IMP, vm_has_sve),
+	MAX_FEAT(ID_AA64PFR0_EL1, RAS, IMP),
+	MAX_FEAT(ID_AA64PFR0_EL1, DIT, IMP),
+	MAX_FEAT(ID_AA64PFR0_EL1, CSV2, IMP),
+	MAX_FEAT(ID_AA64PFR0_EL1, CSV3, IMP),
+	FEAT_END
+};
+
+static const struct pvm_ftr_bits pvmid_aa64pfr1[] = {
+	MAX_FEAT(ID_AA64PFR1_EL1, BT, IMP),
+	MAX_FEAT(ID_AA64PFR1_EL1, SSBS, SSBS2),
+	MAX_FEAT_ENUM(ID_AA64PFR1_EL1, MTE_frac, NI),
+	FEAT_END
+};
+
+static const struct pvm_ftr_bits pvmid_aa64pfr2[] = {
+	MAX_FEAT(ID_AA64PFR2_EL1, GCIE, NI),
+	FEAT_END
+};
+
+static const struct pvm_ftr_bits pvmid_aa64mmfr0[] = {
+	MAX_FEAT_ENUM(ID_AA64MMFR0_EL1, PARANGE, 40),
+	MAX_FEAT_ENUM(ID_AA64MMFR0_EL1, ASIDBITS, 16),
+	MAX_FEAT(ID_AA64MMFR0_EL1, BIGEND, IMP),
+	MAX_FEAT(ID_AA64MMFR0_EL1, SNSMEM, IMP),
+	MAX_FEAT(ID_AA64MMFR0_EL1, BIGENDEL0, IMP),
+	MAX_FEAT(ID_AA64MMFR0_EL1, EXS, IMP),
+	FEAT_END
+};
+
+static const struct pvm_ftr_bits pvmid_aa64mmfr1[] = {
+	MAX_FEAT(ID_AA64MMFR1_EL1, HAFDBS, DBM),
+	MAX_FEAT_ENUM(ID_AA64MMFR1_EL1, VMIDBits, 16),
+	MAX_FEAT(ID_AA64MMFR1_EL1, HPDS, HPDS2),
+	MAX_FEAT(ID_AA64MMFR1_EL1, PAN, PAN3),
+	MAX_FEAT(ID_AA64MMFR1_EL1, SpecSEI, IMP),
+	MAX_FEAT(ID_AA64MMFR1_EL1, ETS, IMP),
+	MAX_FEAT(ID_AA64MMFR1_EL1, CMOW, IMP),
+	FEAT_END
+};
+
+static const struct pvm_ftr_bits pvmid_aa64mmfr2[] = {
+	MAX_FEAT(ID_AA64MMFR2_EL1, CnP, IMP),
+	MAX_FEAT(ID_AA64MMFR2_EL1, UAO, IMP),
+	MAX_FEAT(ID_AA64MMFR2_EL1, IESB, IMP),
+	MAX_FEAT(ID_AA64MMFR2_EL1, AT, IMP),
+	MAX_FEAT(ID_AA64MMFR2_EL1, IDS, IMP),
+	MAX_FEAT(ID_AA64MMFR2_EL1, TTL, IMP),
+	MAX_FEAT(ID_AA64MMFR2_EL1, BBM, 2),
+	MAX_FEAT(ID_AA64MMFR2_EL1, E0PD, IMP),
+	FEAT_END
+};
+
+static const struct pvm_ftr_bits pvmid_aa64isar1[] = {
+	MAX_FEAT(ID_AA64ISAR1_EL1, DPB, DPB2),
+	MAX_FEAT_FUNC(ID_AA64ISAR1_EL1, APA, PAuth, vm_has_ptrauth),
+	MAX_FEAT_FUNC(ID_AA64ISAR1_EL1, API, PAuth, vm_has_ptrauth),
+	MAX_FEAT(ID_AA64ISAR1_EL1, JSCVT, IMP),
+	MAX_FEAT(ID_AA64ISAR1_EL1, FCMA, IMP),
+	MAX_FEAT(ID_AA64ISAR1_EL1, LRCPC, LRCPC3),
+	MAX_FEAT(ID_AA64ISAR1_EL1, GPA, IMP),
+	MAX_FEAT(ID_AA64ISAR1_EL1, GPI, IMP),
+	MAX_FEAT(ID_AA64ISAR1_EL1, FRINTTS, IMP),
+	MAX_FEAT(ID_AA64ISAR1_EL1, SB, IMP),
+	MAX_FEAT(ID_AA64ISAR1_EL1, SPECRES, COSP_RCTX),
+	MAX_FEAT(ID_AA64ISAR1_EL1, BF16, EBF16),
+	MAX_FEAT(ID_AA64ISAR1_EL1, DGH, IMP),
+	MAX_FEAT(ID_AA64ISAR1_EL1, I8MM, IMP),
+	FEAT_END
+};
+
+static const struct pvm_ftr_bits pvmid_aa64isar2[] = {
+	MAX_FEAT_FUNC(ID_AA64ISAR2_EL1, GPA3, IMP, vm_has_ptrauth),
+	MAX_FEAT_FUNC(ID_AA64ISAR2_EL1, APA3, PAuth, vm_has_ptrauth),
+	MAX_FEAT(ID_AA64ISAR2_EL1, ATS1A, IMP),
+	FEAT_END
+};
+
+/*
+ * None of the features in ID_AA64DFR0_EL1 nor ID_AA64MMFR4_EL1 are supported.
+ * However, both have Not-Implemented values that are non-zero. Define them
+ * so they can be used when getting the value of these registers.
+ */
+#define ID_AA64DFR0_EL1_NONZERO_NI					\
+(									\
+	SYS_FIELD_PREP_ENUM(ID_AA64DFR0_EL1, DoubleLock, NI)	|	\
+	SYS_FIELD_PREP_ENUM(ID_AA64DFR0_EL1, MTPMU, NI)			\
+)
+
+#define ID_AA64MMFR4_EL1_NONZERO_NI					\
+	SYS_FIELD_PREP_ENUM(ID_AA64MMFR4_EL1, E2H0, NI)
+
+/*
+ * Returns the value of the feature registers based on the system register
+ * value, the vcpu support for the revelant features, and the additional
+ * restrictions for protected VMs.
+ */
+static u64 get_restricted_features(const struct kvm_vcpu *vcpu,
+				   u64 sys_reg_val,
+				   const struct pvm_ftr_bits restrictions[])
+{
+	u64 val = 0UL;
+	int i;
+
+	for (i = 0; restrictions[i].width != 0; i++) {
+		bool (*vm_supported)(const struct kvm *) = restrictions[i].vm_supported;
+		bool sign = restrictions[i].sign;
+		int shift = restrictions[i].shift;
+		int width = restrictions[i].width;
+		u64 min_signed = (1UL << width) - 1UL;
+		u64 sign_bit = 1UL << (width - 1);
+		u64 mask = GENMASK_ULL(width + shift - 1, shift);
+		u64 sys_val = (sys_reg_val & mask) >> shift;
+		u64 pvm_max = restrictions[i].max_val;
+
+		if (vm_supported && !vm_supported(vcpu->kvm))
+			val |= (sign ? min_signed : 0) << shift;
+		else if (sign && (sys_val >= sign_bit || pvm_max >= sign_bit))
+			val |= max(sys_val, pvm_max) << shift;
+		else
+			val |= min(sys_val, pvm_max) << shift;
+	}
+
+	return val;
+}
+
+static u64 pvm_calc_id_reg(const struct kvm_vcpu *vcpu, u32 id)
+{
+	switch (id) {
+	case SYS_ID_AA64PFR0_EL1:
+		return get_restricted_features(vcpu, id_aa64pfr0_el1_sys_val, pvmid_aa64pfr0);
+	case SYS_ID_AA64PFR1_EL1:
+		return get_restricted_features(vcpu, id_aa64pfr1_el1_sys_val, pvmid_aa64pfr1);
+	case SYS_ID_AA64PFR2_EL1:
+		return get_restricted_features(vcpu, id_aa64pfr2_el1_sys_val, pvmid_aa64pfr2);
+	case SYS_ID_AA64ISAR0_EL1:
+		return id_aa64isar0_el1_sys_val;
+	case SYS_ID_AA64ISAR1_EL1:
+		return get_restricted_features(vcpu, id_aa64isar1_el1_sys_val, pvmid_aa64isar1);
+	case SYS_ID_AA64ISAR2_EL1:
+		return get_restricted_features(vcpu, id_aa64isar2_el1_sys_val, pvmid_aa64isar2);
+	case SYS_ID_AA64MMFR0_EL1:
+		return get_restricted_features(vcpu, id_aa64mmfr0_el1_sys_val, pvmid_aa64mmfr0);
+	case SYS_ID_AA64MMFR1_EL1:
+		return get_restricted_features(vcpu, id_aa64mmfr1_el1_sys_val, pvmid_aa64mmfr1);
+	case SYS_ID_AA64MMFR2_EL1:
+		return get_restricted_features(vcpu, id_aa64mmfr2_el1_sys_val, pvmid_aa64mmfr2);
+	case SYS_ID_AA64DFR0_EL1:
+		return ID_AA64DFR0_EL1_NONZERO_NI;
+	case SYS_ID_AA64MMFR4_EL1:
+		return ID_AA64MMFR4_EL1_NONZERO_NI;
+	default:
+		/* Unhandled ID register, RAZ */
+		return 0;
+	}
+}
+
+static void inject_sync64(struct kvm_vcpu *vcpu, u64 esr)
+{
+	*vcpu_pc(vcpu) = read_sysreg_el2(SYS_ELR);
+	*vcpu_cpsr(vcpu) = read_sysreg_el2(SYS_SPSR);
+
+	/*
+	 * Make sure we have the latest update to VBAR_EL1, as pKVM
+	 * handles traps very early, before sysregs are resync'ed
+	 */
+	__vcpu_assign_sys_reg(vcpu, VBAR_EL1, read_sysreg_el1(SYS_VBAR));
+
+	kvm_pend_exception(vcpu, EXCEPT_AA64_EL1_SYNC);
+
+	__kvm_adjust_pc(vcpu);
+
+	write_sysreg_el1(esr, SYS_ESR);
+	write_sysreg_el1(read_sysreg_el2(SYS_ELR), SYS_ELR);
+	write_sysreg_el1(read_sysreg_el2(SYS_SPSR), SYS_SPSR);
+	write_sysreg_el2(*vcpu_pc(vcpu), SYS_ELR);
+	write_sysreg_el2(*vcpu_cpsr(vcpu), SYS_SPSR);
+}
+
+/*
+ * Inject an unknown/undefined exception to an AArch64 guest while most of its
+ * sysregs are live.
+ */
+static void inject_undef64(struct kvm_vcpu *vcpu)
+{
+	inject_sync64(vcpu, (ESR_ELx_EC_UNKNOWN << ESR_ELx_EC_SHIFT) | ESR_ELx_IL);
+}
+
+static u64 read_id_reg(const struct kvm_vcpu *vcpu,
+		       struct sys_reg_desc const *r)
+{
+	struct kvm *kvm = vcpu->kvm;
+	u32 reg = reg_to_encoding(r);
+
+	if (WARN_ON_ONCE(!test_bit(KVM_ARCH_FLAG_ID_REGS_INITIALIZED, &kvm->arch.flags)))
+		return 0;
+
+	if (reg >= sys_reg(3, 0, 0, 1, 0) && reg <= sys_reg(3, 0, 0, 7, 7))
+		return kvm->arch.id_regs[IDREG_IDX(reg)];
+
+	return 0;
+}
+
+/* Handler to RAZ/WI sysregs */
+static bool pvm_access_raz_wi(struct kvm_vcpu *vcpu, struct sys_reg_params *p,
+			      const struct sys_reg_desc *r)
+{
+	if (!p->is_write)
+		p->regval = 0;
+
+	return true;
+}
+
+/*
+ * Accessor for AArch32 feature id registers.
+ *
+ * The value of these registers is "unknown" according to the spec if AArch32
+ * isn't supported.
+ */
+static bool pvm_access_id_aarch32(struct kvm_vcpu *vcpu,
+				  struct sys_reg_params *p,
+				  const struct sys_reg_desc *r)
+{
+	if (p->is_write) {
+		inject_undef64(vcpu);
+		return false;
+	}
+
+	return pvm_access_raz_wi(vcpu, p, r);
+}
+
+/*
+ * Accessor for AArch64 feature id registers.
+ *
+ * If access is allowed, set the regval to the protected VM's view of the
+ * register and return true.
+ * Otherwise, inject an undefined exception and return false.
+ */
+static bool pvm_access_id_aarch64(struct kvm_vcpu *vcpu,
+				  struct sys_reg_params *p,
+				  const struct sys_reg_desc *r)
+{
+	if (p->is_write) {
+		inject_undef64(vcpu);
+		return false;
+	}
+
+	p->regval = read_id_reg(vcpu, r);
+	return true;
+}
+
+static bool pvm_gic_read_sre(struct kvm_vcpu *vcpu,
+			     struct sys_reg_params *p,
+			     const struct sys_reg_desc *r)
+{
+	/* pVMs only support GICv3. 'nuf said. */
+	if (!p->is_write)
+		p->regval = ICC_SRE_EL1_DIB | ICC_SRE_EL1_DFB | ICC_SRE_EL1_SRE;
+
+	return true;
+}
+
+static bool pvm_idst_access(struct kvm_vcpu *vcpu,
+			    struct sys_reg_params *p,
+			    const struct sys_reg_desc *r)
+{
+	if (kvm_has_feat(vcpu->kvm, ID_AA64MMFR2_EL1, IDS, IMP))
+		inject_sync64(vcpu, kvm_vcpu_get_esr(vcpu));
+	else
+		inject_undef64(vcpu);
+
+	return false;
+}
+
+/* Mark the specified system register as an AArch32 feature id register. */
+#define AARCH32(REG) { SYS_DESC(REG), .access = pvm_access_id_aarch32 }
+
+/* Mark the specified system register as an AArch64 feature id register. */
+#define AARCH64(REG) { SYS_DESC(REG), .access = pvm_access_id_aarch64 }
+
+/*
+ * sys_reg_desc initialiser for architecturally unallocated cpufeature ID
+ * register with encoding Op0=3, Op1=0, CRn=0, CRm=crm, Op2=op2
+ * (1 <= crm < 8, 0 <= Op2 < 8).
+ */
+#define ID_UNALLOCATED(crm, op2) {			\
+	Op0(3), Op1(0), CRn(0), CRm(crm), Op2(op2),	\
+	.access = pvm_access_id_aarch64,		\
+}
+
+/* Mark the specified system register as Read-As-Zero/Write-Ignored */
+#define RAZ_WI(REG) { SYS_DESC(REG), .access = pvm_access_raz_wi }
+
+/* Mark the specified system register as not being handled in hyp. */
+#define HOST_HANDLED(REG) { SYS_DESC(REG), .access = NULL }
+
+/*
+ * Architected system registers.
+ * Important: Must be sorted ascending by Op0, Op1, CRn, CRm, Op2
+ *
+ * NOTE: Anything not explicitly listed here is *restricted by default*, i.e.,
+ * it will lead to injecting an exception into the guest.
+ */
+static const struct sys_reg_desc pvm_sys_reg_descs[] = {
+	/* Cache maintenance by set/way operations are restricted. */
+
+	/* Debug and Trace Registers are restricted. */
+	RAZ_WI(SYS_DBGBVRn_EL1(0)),
+	RAZ_WI(SYS_DBGBCRn_EL1(0)),
+	RAZ_WI(SYS_DBGWVRn_EL1(0)),
+	RAZ_WI(SYS_DBGWCRn_EL1(0)),
+	RAZ_WI(SYS_MDSCR_EL1),
+	RAZ_WI(SYS_OSLAR_EL1),
+	RAZ_WI(SYS_OSLSR_EL1),
+	RAZ_WI(SYS_OSDLR_EL1),
+
+	/* Group 1 ID registers */
+	HOST_HANDLED(SYS_REVIDR_EL1),
+
+	/* AArch64 mappings of the AArch32 ID registers */
+	/* CRm=1 */
+	AARCH32(SYS_ID_PFR0_EL1),
+	AARCH32(SYS_ID_PFR1_EL1),
+	AARCH32(SYS_ID_DFR0_EL1),
+	AARCH32(SYS_ID_AFR0_EL1),
+	AARCH32(SYS_ID_MMFR0_EL1),
+	AARCH32(SYS_ID_MMFR1_EL1),
+	AARCH32(SYS_ID_MMFR2_EL1),
+	AARCH32(SYS_ID_MMFR3_EL1),
+
+	/* CRm=2 */
+	AARCH32(SYS_ID_ISAR0_EL1),
+	AARCH32(SYS_ID_ISAR1_EL1),
+	AARCH32(SYS_ID_ISAR2_EL1),
+	AARCH32(SYS_ID_ISAR3_EL1),
+	AARCH32(SYS_ID_ISAR4_EL1),
+	AARCH32(SYS_ID_ISAR5_EL1),
+	AARCH32(SYS_ID_MMFR4_EL1),
+	AARCH32(SYS_ID_ISAR6_EL1),
+
+	/* CRm=3 */
+	AARCH32(SYS_MVFR0_EL1),
+	AARCH32(SYS_MVFR1_EL1),
+	AARCH32(SYS_MVFR2_EL1),
+	ID_UNALLOCATED(3,3),
+	AARCH32(SYS_ID_PFR2_EL1),
+	AARCH32(SYS_ID_DFR1_EL1),
+	AARCH32(SYS_ID_MMFR5_EL1),
+	ID_UNALLOCATED(3,7),
+
+	/* AArch64 ID registers */
+	/* CRm=4 */
+	AARCH64(SYS_ID_AA64PFR0_EL1),
+	AARCH64(SYS_ID_AA64PFR1_EL1),
+	AARCH64(SYS_ID_AA64PFR2_EL1),
+	ID_UNALLOCATED(4,3),
+	AARCH64(SYS_ID_AA64ZFR0_EL1),
+	ID_UNALLOCATED(4,5),
+	ID_UNALLOCATED(4,6),
+	ID_UNALLOCATED(4,7),
+	AARCH64(SYS_ID_AA64DFR0_EL1),
+	AARCH64(SYS_ID_AA64DFR1_EL1),
+	ID_UNALLOCATED(5,2),
+	ID_UNALLOCATED(5,3),
+	AARCH64(SYS_ID_AA64AFR0_EL1),
+	AARCH64(SYS_ID_AA64AFR1_EL1),
+	ID_UNALLOCATED(5,6),
+	ID_UNALLOCATED(5,7),
+	AARCH64(SYS_ID_AA64ISAR0_EL1),
+	AARCH64(SYS_ID_AA64ISAR1_EL1),
+	AARCH64(SYS_ID_AA64ISAR2_EL1),
+	ID_UNALLOCATED(6,3),
+	ID_UNALLOCATED(6,4),
+	ID_UNALLOCATED(6,5),
+	ID_UNALLOCATED(6,6),
+	ID_UNALLOCATED(6,7),
+	AARCH64(SYS_ID_AA64MMFR0_EL1),
+	AARCH64(SYS_ID_AA64MMFR1_EL1),
+	AARCH64(SYS_ID_AA64MMFR2_EL1),
+	ID_UNALLOCATED(7,3),
+	ID_UNALLOCATED(7,4),
+	ID_UNALLOCATED(7,5),
+	ID_UNALLOCATED(7,6),
+	ID_UNALLOCATED(7,7),
+
+	/* Scalable Vector Registers are restricted. */
+
+	HOST_HANDLED(SYS_ICC_PMR_EL1),
+
+	RAZ_WI(SYS_ERRIDR_EL1),
+	RAZ_WI(SYS_ERRSELR_EL1),
+	RAZ_WI(SYS_ERXFR_EL1),
+	RAZ_WI(SYS_ERXCTLR_EL1),
+	RAZ_WI(SYS_ERXSTATUS_EL1),
+	RAZ_WI(SYS_ERXADDR_EL1),
+	RAZ_WI(SYS_ERXMISC0_EL1),
+	RAZ_WI(SYS_ERXMISC1_EL1),
+
+	/* Performance Monitoring Registers are restricted. */
+
+	/* Limited Ordering Regions Registers are restricted. */
+
+	HOST_HANDLED(SYS_ICC_DIR_EL1),
+	HOST_HANDLED(SYS_ICC_RPR_EL1),
+	HOST_HANDLED(SYS_ICC_SGI1R_EL1),
+	HOST_HANDLED(SYS_ICC_ASGI1R_EL1),
+	HOST_HANDLED(SYS_ICC_SGI0R_EL1),
+	HOST_HANDLED(SYS_ICC_CTLR_EL1),
+	{ SYS_DESC(SYS_ICC_SRE_EL1), .access = pvm_gic_read_sre, },
+
+	HOST_HANDLED(SYS_CCSIDR_EL1),
+	HOST_HANDLED(SYS_CLIDR_EL1),
+	{ SYS_DESC(SYS_CCSIDR2_EL1), .access = pvm_idst_access },
+	{ SYS_DESC(SYS_GMID_EL1), .access = pvm_idst_access },
+	{ SYS_DESC(SYS_SMIDR_EL1), .access = pvm_idst_access },
+	HOST_HANDLED(SYS_AIDR_EL1),
+	HOST_HANDLED(SYS_CSSELR_EL1),
+	HOST_HANDLED(SYS_CTR_EL0),
+
+	/* Performance Monitoring Registers are restricted. */
+
+	/* Activity Monitoring Registers are restricted. */
+
+	HOST_HANDLED(SYS_CNTP_TVAL_EL0),
+	HOST_HANDLED(SYS_CNTP_CTL_EL0),
+	HOST_HANDLED(SYS_CNTP_CVAL_EL0),
+
+	/* Performance Monitoring Registers are restricted. */
+};
+
+/*
+ * Initializes feature registers for protected vms.
+ */
+void kvm_init_pvm_id_regs(struct kvm_vcpu *vcpu)
+{
+	struct kvm *kvm = vcpu->kvm;
+	struct kvm_arch *ka = &kvm->arch;
+	u32 r;
+
+	hyp_assert_lock_held(&vm_table_lock);
+
+	if (test_bit(KVM_ARCH_FLAG_ID_REGS_INITIALIZED, &kvm->arch.flags))
+		return;
+
+	/*
+	 * Initialize only AArch64 id registers since AArch32 isn't supported
+	 * for protected VMs.
+	 */
+	for (r = sys_reg(3, 0, 0, 4, 0); r <= sys_reg(3, 0, 0, 7, 7); r += sys_reg(0, 0, 0, 0, 1))
+		ka->id_regs[IDREG_IDX(r)] = pvm_calc_id_reg(vcpu, r);
+
+	set_bit(KVM_ARCH_FLAG_ID_REGS_INITIALIZED, &kvm->arch.flags);
+}
+
+/*
+ * Checks that the sysreg table is unique and in-order.
+ *
+ * Returns 0 if the table is consistent, or 1 otherwise.
+ */
+int kvm_check_pvm_sysreg_table(void)
+{
+	unsigned int i;
+
+	for (i = 1; i < ARRAY_SIZE(pvm_sys_reg_descs); i++) {
+		if (cmp_sys_reg(&pvm_sys_reg_descs[i-1], &pvm_sys_reg_descs[i]) >= 0)
+			return 1;
+	}
+
+	return 0;
+}
+
+/*
+ * Handler for protected VM MSR, MRS or System instruction execution.
+ *
+ * Returns true if the hypervisor has handled the exit, and control should go
+ * back to the guest, or false if it hasn't, to be handled by the host.
+ */
+bool kvm_handle_pvm_sysreg(struct kvm_vcpu *vcpu, u64 *exit_code)
+{
+	const struct sys_reg_desc *r;
+	struct sys_reg_params params;
+	unsigned long esr = kvm_vcpu_get_esr(vcpu);
+	int Rt = kvm_vcpu_sys_get_rt(vcpu);
+
+	params = esr_sys64_to_params(esr);
+	params.regval = vcpu_get_reg(vcpu, Rt);
+
+	r = find_reg(&params, pvm_sys_reg_descs, ARRAY_SIZE(pvm_sys_reg_descs));
+
+	/* Undefined (RESTRICTED). */
+	if (r == NULL) {
+		inject_undef64(vcpu);
+		return true;
+	}
+
+	/* Handled by the host (HOST_HANDLED) */
+	if (r->access == NULL)
+		return false;
+
+	/* Handled by hyp: skip instruction if instructed to do so. */
+	if (r->access(vcpu, &params, r))
+		__kvm_skip_instr(vcpu);
+
+	if (!params.is_write)
+		vcpu_set_reg(vcpu, Rt, params.regval);
+
+	return true;
+}
+
+/*
+ * Handler for protected VM restricted exceptions.
+ *
+ * Inject an undefined exception into the guest and return true to indicate that
+ * the hypervisor has handled the exit, and control should go back to the guest.
+ */
+bool kvm_handle_pvm_restricted(struct kvm_vcpu *vcpu, u64 *exit_code)
+{
+	inject_undef64(vcpu);
+	return true;
+}

linux-7.2.3/arch/arm64/kvm/hyp/nvhe/sysreg-sr.c

@@ -0,0 +1,37 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2012-2015 - ARM Ltd
+ * Author: Marc Zyngier <marc.zyngier@arm.com>
+ */
+
+#include <hyp/sysreg-sr.h>
+
+#include <linux/compiler.h>
+#include <linux/kvm_host.h>
+
+#include <asm/kprobes.h>
+#include <asm/kvm_asm.h>
+#include <asm/kvm_emulate.h>
+#include <asm/kvm_hyp.h>
+
+/*
+ * Non-VHE: Both host and guest must save everything.
+ */
+
+void __sysreg_save_state_nvhe(struct kvm_cpu_context *ctxt)
+{
+	__sysreg_save_el1_state(ctxt);
+	__sysreg_save_common_state(ctxt);
+	__sysreg_save_user_state(ctxt);
+	__sysreg_save_el2_return_state(ctxt);
+}
+
+void __sysreg_restore_state_nvhe(struct kvm_cpu_context *ctxt)
+{
+	u64 midr = ctxt_midr_el1(ctxt);
+
+	__sysreg_restore_el1_state(ctxt, midr, ctxt_sys_reg(ctxt, MPIDR_EL1));
+	__sysreg_restore_common_state(ctxt);
+	__sysreg_restore_user_state(ctxt);
+	__sysreg_restore_el2_return_state(ctxt);
+}

linux-7.2.3/arch/arm64/kvm/hyp/nvhe/timer-sr.c

@@ -0,0 +1,70 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2012-2015 - ARM Ltd
+ * Author: Marc Zyngier <marc.zyngier@arm.com>
+ */
+
+#include <clocksource/arm_arch_timer.h>
+#include <linux/compiler.h>
+#include <linux/kvm_host.h>
+
+#include <asm/kvm_hyp.h>
+#include <asm/kvm_mmu.h>
+
+void __kvm_timer_set_cntvoff(u64 cntvoff)
+{
+	write_sysreg(cntvoff, cntvoff_el2);
+}
+
+/*
+ * Should only be called on non-VHE or hVHE setups.
+ * VHE systems use EL2 timers and configure EL1 timers in kvm_timer_init_vhe().
+ */
+void __timer_disable_traps(struct kvm_vcpu *vcpu)
+{
+	u64 set, clr, shift = 0;
+
+	if (has_hvhe())
+		shift = 10;
+
+	/* Allow physical timer/counter access for the host */
+	set = (CNTHCTL_EL1PCTEN | CNTHCTL_EL1PCEN) << shift;
+	clr = CNTHCTL_EL1TVT | CNTHCTL_EL1TVCT;
+
+	sysreg_clear_set(cnthctl_el2, clr, set);
+}
+
+/*
+ * Should only be called on non-VHE or hVHE setups.
+ * VHE systems use EL2 timers and configure EL1 timers in kvm_timer_init_vhe().
+ */
+void __timer_enable_traps(struct kvm_vcpu *vcpu)
+{
+	u64 clr = 0, set = 0;
+
+	/*
+	 * Disallow physical timer access for the guest
+	 * Physical counter access is allowed if no offset is enforced
+	 * or running protected (we don't offset anything in this case).
+	 */
+	clr = CNTHCTL_EL1PCEN;
+	if (is_protected_kvm_enabled() ||
+	    !kern_hyp_va(vcpu->kvm)->arch.timer_data.poffset)
+		set |= CNTHCTL_EL1PCTEN;
+	else
+		clr |= CNTHCTL_EL1PCTEN;
+
+	if (has_hvhe()) {
+		clr <<= 10;
+		set <<= 10;
+	}
+
+	/*
+	 * Trap the virtual counter/timer if we have a broken cntvoff
+	 * implementation.
+	 */
+	if (has_broken_cntvoff())
+		set |= CNTHCTL_EL1TVT | CNTHCTL_EL1TVCT;
+
+	sysreg_clear_set(cnthctl_el2, clr, set);
+}

linux-7.2.3/arch/arm64/kvm/hyp/nvhe/tlb.c

@@ -0,0 +1,268 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2015 - ARM Ltd
+ * Author: Marc Zyngier <marc.zyngier@arm.com>
+ */
+
+#include <asm/kvm_hyp.h>
+#include <asm/kvm_mmu.h>
+#include <asm/tlbflush.h>
+
+#include <nvhe/mem_protect.h>
+
+struct tlb_inv_context {
+	struct kvm_s2_mmu	*mmu;
+	u64			tcr;
+	u64			sctlr;
+};
+
+static void enter_vmid_context(struct kvm_s2_mmu *mmu,
+			       struct tlb_inv_context *cxt,
+			       bool nsh)
+{
+	struct kvm_s2_mmu *host_s2_mmu = &host_mmu.arch.mmu;
+	struct kvm_cpu_context *host_ctxt;
+	struct kvm_vcpu *vcpu;
+
+	host_ctxt = &this_cpu_ptr(&kvm_host_data)->host_ctxt;
+	vcpu = host_ctxt->__hyp_running_vcpu;
+	cxt->mmu = NULL;
+
+	/*
+	 * We have two requirements:
+	 *
+	 * - ensure that the page table updates are visible to all
+	 *   CPUs, for which a dsb(DOMAIN-st) is what we need, DOMAIN
+	 *   being either ish or nsh, depending on the invalidation
+	 *   type.
+	 *
+	 * - complete any speculative page table walk started before
+	 *   we trapped to EL2 so that we can mess with the MM
+	 *   registers out of context, for which dsb(nsh) is enough
+	 *
+	 * The composition of these two barriers is a dsb(DOMAIN), and
+	 * the 'nsh' parameter tracks the distinction between
+	 * Inner-Shareable and Non-Shareable, as specified by the
+	 * callers.
+	 */
+	if (nsh)
+		dsb(nsh);
+	else
+		dsb(ish);
+
+	/*
+	 * If we're already in the desired context, then there's nothing to do.
+	 */
+	if (vcpu) {
+		/*
+		 * We're in guest context. However, for this to work, this needs
+		 * to be called from within __kvm_vcpu_run(), which ensures that
+		 * __hyp_running_vcpu is set to the current guest vcpu.
+		 */
+		if (mmu == vcpu->arch.hw_mmu || WARN_ON(mmu != host_s2_mmu))
+			return;
+
+		cxt->mmu = vcpu->arch.hw_mmu;
+	} else {
+		/* We're in host context. */
+		if (mmu == host_s2_mmu)
+			return;
+
+		cxt->mmu = host_s2_mmu;
+	}
+
+	if (cpus_have_final_cap(ARM64_WORKAROUND_SPECULATIVE_AT)) {
+		u64 val;
+
+		/*
+		 * For CPUs that are affected by ARM 1319367, we need to
+		 * avoid a Stage-1 walk with the old VMID while we have
+		 * the new VMID set in the VTTBR in order to invalidate TLBs.
+		 * We're guaranteed that the host S1 MMU is enabled, so
+		 * we can simply set the EPD bits to avoid any further
+		 * TLB fill. For guests, we ensure that the S1 MMU is
+		 * temporarily enabled in the next context.
+		 */
+		val = cxt->tcr = read_sysreg_el1(SYS_TCR);
+		val |= TCR_EPD1_MASK | TCR_EPD0_MASK;
+		write_sysreg_el1(val, SYS_TCR);
+		isb();
+
+		if (vcpu) {
+			val = cxt->sctlr = read_sysreg_el1(SYS_SCTLR);
+			if (!(val & SCTLR_ELx_M)) {
+				val |= SCTLR_ELx_M;
+				write_sysreg_el1(val, SYS_SCTLR);
+				isb();
+			}
+		} else {
+			/* The host S1 MMU is always enabled. */
+			cxt->sctlr = SCTLR_ELx_M;
+		}
+	}
+
+	/*
+	 * __load_stage2() includes an ISB only when the AT
+	 * workaround is applied. Take care of the opposite condition,
+	 * ensuring that we always have an ISB, but not two ISBs back
+	 * to back.
+	 */
+	if (vcpu)
+		__load_host_stage2();
+	else
+		__load_stage2(mmu);
+
+	asm(ALTERNATIVE("isb", "nop", ARM64_WORKAROUND_SPECULATIVE_AT));
+}
+
+static void exit_vmid_context(struct tlb_inv_context *cxt)
+{
+	struct kvm_s2_mmu *mmu = cxt->mmu;
+	struct kvm_cpu_context *host_ctxt;
+	struct kvm_vcpu *vcpu;
+
+	host_ctxt = &this_cpu_ptr(&kvm_host_data)->host_ctxt;
+	vcpu = host_ctxt->__hyp_running_vcpu;
+
+	if (!mmu)
+		return;
+
+	if (vcpu)
+		__load_stage2(mmu);
+	else
+		__load_host_stage2();
+
+	/* Ensure write of the old VMID */
+	isb();
+
+	if (cpus_have_final_cap(ARM64_WORKAROUND_SPECULATIVE_AT)) {
+		if (!(cxt->sctlr & SCTLR_ELx_M)) {
+			write_sysreg_el1(cxt->sctlr, SYS_SCTLR);
+			isb();
+		}
+
+		write_sysreg_el1(cxt->tcr, SYS_TCR);
+	}
+}
+
+void __kvm_tlb_flush_vmid_ipa(struct kvm_s2_mmu *mmu,
+			      phys_addr_t ipa, int level)
+{
+	struct tlb_inv_context cxt;
+
+	/* Switch to requested VMID */
+	enter_vmid_context(mmu, &cxt, false);
+
+	/*
+	 * We could do so much better if we had the VA as well.
+	 * Instead, we invalidate Stage-2 for this IPA, and the
+	 * whole of Stage-1. Weep...
+	 */
+	__tlbi_level(ipas2e1is, ipa, level);
+
+	/*
+	 * We have to ensure completion of the invalidation at Stage-2,
+	 * since a table walk on another CPU could refill a TLB with a
+	 * complete (S1 + S2) walk based on the old Stage-2 mapping if
+	 * the Stage-1 invalidation happened first.
+	 */
+	dsb(ish);
+	__tlbi(vmalle1is);
+	__tlbi_sync_s1ish_hyp();
+	isb();
+
+	exit_vmid_context(&cxt);
+}
+
+void __kvm_tlb_flush_vmid_ipa_nsh(struct kvm_s2_mmu *mmu,
+				  phys_addr_t ipa, int level)
+{
+	struct tlb_inv_context cxt;
+
+	/* Switch to requested VMID */
+	enter_vmid_context(mmu, &cxt, true);
+
+	/*
+	 * We could do so much better if we had the VA as well.
+	 * Instead, we invalidate Stage-2 for this IPA, and the
+	 * whole of Stage-1. Weep...
+	 */
+	__tlbi_level(ipas2e1, ipa, level);
+
+	/*
+	 * We have to ensure completion of the invalidation at Stage-2,
+	 * since a table walk on another CPU could refill a TLB with a
+	 * complete (S1 + S2) walk based on the old Stage-2 mapping if
+	 * the Stage-1 invalidation happened first.
+	 */
+	dsb(nsh);
+	__tlbi(vmalle1);
+	dsb(nsh);
+	isb();
+
+	exit_vmid_context(&cxt);
+}
+
+void __kvm_tlb_flush_vmid_range(struct kvm_s2_mmu *mmu,
+				phys_addr_t start, unsigned long pages)
+{
+	struct tlb_inv_context cxt;
+	unsigned long stride;
+
+	/*
+	 * Since the range of addresses may not be mapped at
+	 * the same level, assume the worst case as PAGE_SIZE
+	 */
+	stride = PAGE_SIZE;
+	start = round_down(start, stride);
+
+	/* Switch to requested VMID */
+	enter_vmid_context(mmu, &cxt, false);
+
+	__flush_s2_tlb_range_op(ipas2e1is, start, pages, stride,
+				TLBI_TTL_UNKNOWN);
+
+	dsb(ish);
+	__tlbi(vmalle1is);
+	__tlbi_sync_s1ish_hyp();
+	isb();
+
+	exit_vmid_context(&cxt);
+}
+
+void __kvm_tlb_flush_vmid(struct kvm_s2_mmu *mmu)
+{
+	struct tlb_inv_context cxt;
+
+	/* Switch to requested VMID */
+	enter_vmid_context(mmu, &cxt, false);
+
+	__tlbi(vmalls12e1is);
+	__tlbi_sync_s1ish_hyp();
+	isb();
+
+	exit_vmid_context(&cxt);
+}
+
+void __kvm_flush_cpu_context(struct kvm_s2_mmu *mmu)
+{
+	struct tlb_inv_context cxt;
+
+	/* Switch to requested VMID */
+	enter_vmid_context(mmu, &cxt, false);
+
+	__tlbi(vmalle1);
+	asm volatile("ic iallu");
+	dsb(nsh);
+	isb();
+
+	exit_vmid_context(&cxt);
+}
+
+void __kvm_flush_vm_context(void)
+{
+	/* Same remark as in enter_vmid_context() */
+	dsb(ish);
+	__tlbi(alle1is);
+	__tlbi_sync_s1ish_hyp();
+}

linux-7.2.3/arch/arm64/kvm/hyp/nvhe/trace.c

@@ -0,0 +1,311 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2025 Google LLC
+ * Author: Vincent Donnefort <vdonnefort@google.com>
+ */
+
+#include <nvhe/clock.h>
+#include <nvhe/mem_protect.h>
+#include <nvhe/mm.h>
+#include <nvhe/trace.h>
+
+#include <asm/percpu.h>
+#include <asm/kvm_mmu.h>
+#include <asm/local.h>
+
+#include "simple_ring_buffer.c"
+
+static DEFINE_PER_CPU(struct simple_rb_per_cpu, __simple_rbs);
+
+static struct hyp_trace_buffer {
+	struct simple_rb_per_cpu __percpu	*simple_rbs;
+	void					*bpages_backing_start;
+	size_t					bpages_backing_size;
+	hyp_spinlock_t				lock;
+} trace_buffer = {
+	.simple_rbs = &__simple_rbs,
+	.lock = __HYP_SPIN_LOCK_UNLOCKED,
+};
+
+static bool hyp_trace_buffer_loaded(struct hyp_trace_buffer *trace_buffer)
+{
+	return trace_buffer->bpages_backing_size > 0;
+}
+
+void *tracing_reserve_entry(unsigned long length)
+{
+	return simple_ring_buffer_reserve(this_cpu_ptr(trace_buffer.simple_rbs), length,
+					  trace_clock());
+}
+
+void tracing_commit_entry(void)
+{
+	simple_ring_buffer_commit(this_cpu_ptr(trace_buffer.simple_rbs));
+}
+
+static int __admit_host_mem(void *start, u64 size)
+{
+	if (!PAGE_ALIGNED(start) || !PAGE_ALIGNED(size) || !size)
+		return -EINVAL;
+
+	if (!is_protected_kvm_enabled())
+		return 0;
+
+	return __pkvm_host_donate_hyp(hyp_virt_to_pfn(start), size >> PAGE_SHIFT);
+}
+
+static void __release_host_mem(void *start, u64 size)
+{
+	if (!is_protected_kvm_enabled())
+		return;
+
+	WARN_ON(__pkvm_hyp_donate_host(hyp_virt_to_pfn(start), size >> PAGE_SHIFT));
+}
+
+static int hyp_trace_buffer_load_bpage_backing(struct hyp_trace_buffer *trace_buffer,
+					       struct hyp_trace_desc *desc)
+{
+	void *start = (void *)kern_hyp_va(desc->bpages_backing_start);
+	size_t size = desc->bpages_backing_size;
+	int ret;
+
+	ret = __admit_host_mem(start, size);
+	if (ret)
+		return ret;
+
+	memset(start, 0, size);
+
+	trace_buffer->bpages_backing_start = start;
+	trace_buffer->bpages_backing_size = size;
+
+	return 0;
+}
+
+static void hyp_trace_buffer_unload_bpage_backing(struct hyp_trace_buffer *trace_buffer)
+{
+	void *start = trace_buffer->bpages_backing_start;
+	size_t size = trace_buffer->bpages_backing_size;
+
+	if (!size)
+		return;
+
+	memset(start, 0, size);
+
+	__release_host_mem(start, size);
+
+	trace_buffer->bpages_backing_start = 0;
+	trace_buffer->bpages_backing_size = 0;
+}
+
+static void *__pin_shared_page(unsigned long kern_va)
+{
+	void *va = kern_hyp_va((void *)kern_va);
+
+	if (!is_protected_kvm_enabled())
+		return va;
+
+	return hyp_pin_shared_mem(va, va + PAGE_SIZE) ? NULL : va;
+}
+
+static void __unpin_shared_page(void *va)
+{
+	if (!is_protected_kvm_enabled())
+		return;
+
+	hyp_unpin_shared_mem(va, va + PAGE_SIZE);
+}
+
+static void hyp_trace_buffer_unload(struct hyp_trace_buffer *trace_buffer)
+{
+	int cpu;
+
+	hyp_assert_lock_held(&trace_buffer->lock);
+
+	if (!hyp_trace_buffer_loaded(trace_buffer))
+		return;
+
+	for (cpu = 0; cpu < hyp_nr_cpus; cpu++)
+		simple_ring_buffer_unload_mm(per_cpu_ptr(trace_buffer->simple_rbs, cpu),
+					     __unpin_shared_page);
+
+	hyp_trace_buffer_unload_bpage_backing(trace_buffer);
+}
+
+static int hyp_trace_buffer_load(struct hyp_trace_buffer *trace_buffer,
+				 struct hyp_trace_desc *desc)
+{
+	struct simple_buffer_page *bpages;
+	struct ring_buffer_desc *rb_desc;
+	int ret, cpu;
+
+	hyp_assert_lock_held(&trace_buffer->lock);
+
+	if (hyp_trace_buffer_loaded(trace_buffer))
+		return -EINVAL;
+
+	ret = hyp_trace_buffer_load_bpage_backing(trace_buffer, desc);
+	if (ret)
+		return ret;
+
+	bpages = trace_buffer->bpages_backing_start;
+	for_each_ring_buffer_desc(rb_desc, cpu, &desc->trace_buffer_desc) {
+		ret = simple_ring_buffer_init_mm(per_cpu_ptr(trace_buffer->simple_rbs, cpu),
+						 bpages, rb_desc, __pin_shared_page,
+						__unpin_shared_page);
+		if (ret)
+			break;
+
+		bpages += rb_desc->nr_page_va;
+	}
+
+	if (ret)
+		hyp_trace_buffer_unload(trace_buffer);
+
+	return ret;
+}
+
+static bool hyp_trace_desc_is_valid(struct hyp_trace_desc *desc, size_t desc_size)
+{
+	struct ring_buffer_desc *rb_desc;
+	unsigned int cpu;
+	size_t nr_bpages;
+	void *desc_end;
+
+	if (!is_protected_kvm_enabled())
+		return true;
+
+	/*
+	 * Both desc_size and bpages_backing_size are untrusted host-provided
+	 * values. We rely on __pkvm_host_donate_hyp() to enforce their validity.
+	 */
+	desc_end = (void *)desc + desc_size;
+	nr_bpages = desc->bpages_backing_size / sizeof(struct simple_buffer_page);
+
+	for_each_ring_buffer_desc(rb_desc, cpu, &desc->trace_buffer_desc) {
+		/* Can we read nr_page_va? */
+		if ((void *)rb_desc + struct_size(rb_desc, page_va, 0) > desc_end)
+			return false;
+
+		/* Overflow desc? */
+		if ((void *)rb_desc + struct_size(rb_desc, page_va, rb_desc->nr_page_va) > desc_end)
+			return false;
+
+		/* Overflow bpages backing memory? */
+		if (nr_bpages < rb_desc->nr_page_va)
+			return false;
+
+		if (cpu >= hyp_nr_cpus)
+			return false;
+
+		if (cpu != rb_desc->cpu)
+			return false;
+
+		nr_bpages -= rb_desc->nr_page_va;
+	}
+
+	return true;
+}
+
+int __tracing_load(unsigned long desc_hva, size_t desc_size)
+{
+	struct hyp_trace_desc *desc = (struct hyp_trace_desc *)kern_hyp_va(desc_hva);
+	int ret;
+
+	ret = __admit_host_mem(desc, desc_size);
+	if (ret)
+		return ret;
+
+	if (!hyp_trace_desc_is_valid(desc, desc_size)) {
+		ret = -EINVAL;
+		goto err_release_desc;
+	}
+
+	hyp_spin_lock(&trace_buffer.lock);
+
+	ret = hyp_trace_buffer_load(&trace_buffer, desc);
+
+	hyp_spin_unlock(&trace_buffer.lock);
+
+err_release_desc:
+	__release_host_mem(desc, desc_size);
+	return ret;
+}
+
+void __tracing_unload(void)
+{
+	hyp_spin_lock(&trace_buffer.lock);
+	hyp_trace_buffer_unload(&trace_buffer);
+	hyp_spin_unlock(&trace_buffer.lock);
+}
+
+int __tracing_enable(bool enable)
+{
+	int cpu, ret = enable ? -EINVAL : 0;
+
+	hyp_spin_lock(&trace_buffer.lock);
+
+	if (!hyp_trace_buffer_loaded(&trace_buffer))
+		goto unlock;
+
+	for (cpu = 0; cpu < hyp_nr_cpus; cpu++)
+		simple_ring_buffer_enable_tracing(per_cpu_ptr(trace_buffer.simple_rbs, cpu),
+						  enable);
+
+	ret = 0;
+
+unlock:
+	hyp_spin_unlock(&trace_buffer.lock);
+
+	return ret;
+}
+
+int __tracing_swap_reader(unsigned int cpu)
+{
+	int ret = -ENODEV;
+
+	if (cpu >= hyp_nr_cpus)
+		return -EINVAL;
+
+	hyp_spin_lock(&trace_buffer.lock);
+
+	if (hyp_trace_buffer_loaded(&trace_buffer))
+		ret = simple_ring_buffer_swap_reader_page(
+				per_cpu_ptr(trace_buffer.simple_rbs, cpu));
+
+	hyp_spin_unlock(&trace_buffer.lock);
+
+	return ret;
+}
+
+void __tracing_update_clock(u32 mult, u32 shift, u64 epoch_ns, u64 epoch_cyc)
+{
+	int cpu;
+
+	/* After this loop, all CPUs are observing the new bank... */
+	for (cpu = 0; cpu < hyp_nr_cpus; cpu++) {
+		struct simple_rb_per_cpu *simple_rb = per_cpu_ptr(trace_buffer.simple_rbs, cpu);
+
+		while (READ_ONCE(simple_rb->status) == SIMPLE_RB_WRITING)
+			;
+	}
+
+	/* ...we can now override the old one and swap. */
+	trace_clock_update(mult, shift, epoch_ns, epoch_cyc);
+}
+
+int __tracing_reset(unsigned int cpu)
+{
+	int ret = -ENODEV;
+
+	if (cpu >= hyp_nr_cpus)
+		return -EINVAL;
+
+	hyp_spin_lock(&trace_buffer.lock);
+
+	if (hyp_trace_buffer_loaded(&trace_buffer))
+		ret = simple_ring_buffer_reset(per_cpu_ptr(trace_buffer.simple_rbs, cpu));
+
+	hyp_spin_unlock(&trace_buffer.lock);
+
+	return ret;
+}