xref: /linux/arch/x86/kernel/alternative.c (revision 5c00eca95a9a20e662bd290c3ef3f2e07dfa9baa)

// SPDX-License-Identifier: GPL-2.0-only
#define pr_fmt(fmt) "SMP alternatives: " fmt

#include <linux/mmu_context.h>
#include <linux/perf_event.h>
#include <linux/vmalloc.h>
#include <linux/memory.h>
#include <linux/execmem.h>

#include <asm/text-patching.h>
#include <asm/insn.h>
#include <asm/ibt.h>
#include <asm/set_memory.h>
#include <asm/nmi.h>

int __read_mostly alternatives_patched;

EXPORT_SYMBOL_GPL(alternatives_patched);

#define MAX_PATCH_LEN (255-1)

#define DA_ALL		(~0)
#define DA_ALT		0x01
#define DA_RET		0x02
#define DA_RETPOLINE	0x04
#define DA_ENDBR	0x08
#define DA_SMP		0x10

static unsigned int debug_alternative;

static int __init debug_alt(char *str)
{
	if (str && *str == '=')
		str++;

	if (!str || kstrtouint(str, 0, &debug_alternative))
		debug_alternative = DA_ALL;

	return 1;
}
__setup("debug-alternative", debug_alt);

static int noreplace_smp;

static int __init setup_noreplace_smp(char *str)
{
	noreplace_smp = 1;
	return 1;
}
__setup("noreplace-smp", setup_noreplace_smp);

#define DPRINTK(type, fmt, args...)					\
do {									\
	if (debug_alternative & DA_##type)				\
		printk(KERN_DEBUG pr_fmt(fmt) "\n", ##args);		\
} while (0)

#define DUMP_BYTES(type, buf, len, fmt, args...)			\
do {									\
	if (unlikely(debug_alternative & DA_##type)) {			\
		int j;							\
									\
		if (!(len))						\
			break;						\
									\
		printk(KERN_DEBUG pr_fmt(fmt), ##args);			\
		for (j = 0; j < (len) - 1; j++)				\
			printk(KERN_CONT "%02hhx ", buf[j]);		\
		printk(KERN_CONT "%02hhx\n", buf[j]);			\
	}								\
} while (0)

static const unsigned char x86nops[] =
{
	BYTES_NOP1,
	BYTES_NOP2,
	BYTES_NOP3,
	BYTES_NOP4,
	BYTES_NOP5,
	BYTES_NOP6,
	BYTES_NOP7,
	BYTES_NOP8,
#ifdef CONFIG_64BIT
	BYTES_NOP9,
	BYTES_NOP10,
	BYTES_NOP11,
#endif
};

const unsigned char * const x86_nops[ASM_NOP_MAX+1] =
{
	NULL,
	x86nops,
	x86nops + 1,
	x86nops + 1 + 2,
	x86nops + 1 + 2 + 3,
	x86nops + 1 + 2 + 3 + 4,
	x86nops + 1 + 2 + 3 + 4 + 5,
	x86nops + 1 + 2 + 3 + 4 + 5 + 6,
	x86nops + 1 + 2 + 3 + 4 + 5 + 6 + 7,
#ifdef CONFIG_64BIT
	x86nops + 1 + 2 + 3 + 4 + 5 + 6 + 7 + 8,
	x86nops + 1 + 2 + 3 + 4 + 5 + 6 + 7 + 8 + 9,
	x86nops + 1 + 2 + 3 + 4 + 5 + 6 + 7 + 8 + 9 + 10,
#endif
};

#ifdef CONFIG_FINEIBT
static bool cfi_paranoid __ro_after_init;
#endif

#ifdef CONFIG_MITIGATION_ITS

#ifdef CONFIG_MODULES
static struct module *its_mod;
#endif
static void *its_page;
static unsigned int its_offset;
struct its_array its_pages;

static void *__its_alloc(struct its_array *pages)
{
	void *page __free(execmem) = execmem_alloc(EXECMEM_MODULE_TEXT, PAGE_SIZE);
	if (!page)
		return NULL;

	void *tmp = krealloc(pages->pages, (pages->num+1) * sizeof(void *),
			     GFP_KERNEL);
	if (!tmp)
		return NULL;

	pages->pages = tmp;
	pages->pages[pages->num++] = page;

	return no_free_ptr(page);
}

/* Initialize a thunk with the "jmp *reg; int3" instructions. */
static void *its_init_thunk(void *thunk, int reg)
{
	u8 *bytes = thunk;
	int offset = 0;
	int i = 0;

#ifdef CONFIG_FINEIBT
	if (cfi_paranoid) {
		/*
		 * When ITS uses indirect branch thunk the fineibt_paranoid
		 * caller sequence doesn't fit in the caller site. So put the
		 * remaining part of the sequence (<ea> + JNE) into the ITS
		 * thunk.
		 */
		bytes[i++] = 0xea; /* invalid instruction */
		bytes[i++] = 0x75; /* JNE */
		bytes[i++] = 0xfd;

		offset = 1;
	}
#endif

	if (reg >= 8) {
		bytes[i++] = 0x41; /* REX.B prefix */
		reg -= 8;
	}
	bytes[i++] = 0xff;
	bytes[i++] = 0xe0 + reg; /* jmp *reg */
	bytes[i++] = 0xcc;

	return thunk + offset;
}

static void its_pages_protect(struct its_array *pages)
{
	for (int i = 0; i < pages->num; i++) {
		void *page = pages->pages[i];
		execmem_restore_rox(page, PAGE_SIZE);
	}
}

static void its_fini_core(void)
{
	if (IS_ENABLED(CONFIG_STRICT_KERNEL_RWX))
		its_pages_protect(&its_pages);
	kfree(its_pages.pages);
}

#ifdef CONFIG_MODULES
void its_init_mod(struct module *mod)
{
	if (!cpu_feature_enabled(X86_FEATURE_INDIRECT_THUNK_ITS))
		return;

	mutex_lock(&text_mutex);
	its_mod = mod;
	its_page = NULL;
}

void its_fini_mod(struct module *mod)
{
	if (!cpu_feature_enabled(X86_FEATURE_INDIRECT_THUNK_ITS))
		return;

	WARN_ON_ONCE(its_mod != mod);

	its_mod = NULL;
	its_page = NULL;
	mutex_unlock(&text_mutex);

	if (IS_ENABLED(CONFIG_STRICT_MODULE_RWX))
		its_pages_protect(&mod->arch.its_pages);
}

void its_free_mod(struct module *mod)
{
	if (!cpu_feature_enabled(X86_FEATURE_INDIRECT_THUNK_ITS))
		return;

	for (int i = 0; i < mod->arch.its_pages.num; i++) {
		void *page = mod->arch.its_pages.pages[i];
		execmem_free(page);
	}
	kfree(mod->arch.its_pages.pages);
}
#endif /* CONFIG_MODULES */

static void *its_alloc(void)
{
	struct its_array *pages = &its_pages;
	void *page;

#ifdef CONFIG_MODULES
	if (its_mod)
		pages = &its_mod->arch.its_pages;
#endif

	page = __its_alloc(pages);
	if (!page)
		return NULL;

	execmem_make_temp_rw(page, PAGE_SIZE);
	if (pages == &its_pages)
		set_memory_x((unsigned long)page, 1);

	return page;
}

static void *its_allocate_thunk(int reg)
{
	int size = 3 + (reg / 8);
	void *thunk;

#ifdef CONFIG_FINEIBT
	/*
	 * The ITS thunk contains an indirect jump and an int3 instruction so
	 * its size is 3 or 4 bytes depending on the register used. If CFI
	 * paranoid is used then 3 extra bytes are added in the ITS thunk to
	 * complete the fineibt_paranoid caller sequence.
	 */
	if (cfi_paranoid)
		size += 3;
#endif

	if (!its_page || (its_offset + size - 1) >= PAGE_SIZE) {
		its_page = its_alloc();
		if (!its_page) {
			pr_err("ITS page allocation failed\n");
			return NULL;
		}
		memset(its_page, INT3_INSN_OPCODE, PAGE_SIZE);
		its_offset = 32;
	}

	/*
	 * If the indirect branch instruction will be in the lower half
	 * of a cacheline, then update the offset to reach the upper half.
	 */
	if ((its_offset + size - 1) % 64 < 32)
		its_offset = ((its_offset - 1) | 0x3F) + 33;

	thunk = its_page + its_offset;
	its_offset += size;

	return its_init_thunk(thunk, reg);
}

u8 *its_static_thunk(int reg)
{
	u8 *thunk = __x86_indirect_its_thunk_array[reg];

#ifdef CONFIG_FINEIBT
	/* Paranoid thunk starts 2 bytes before */
	if (cfi_paranoid)
		return thunk - 2;
#endif
	return thunk;
}

#else
static inline void its_fini_core(void) {}
#endif /* CONFIG_MITIGATION_ITS */

/*
 * Nomenclature for variable names to simplify and clarify this code and ease
 * any potential staring at it:
 *
 * @instr: source address of the original instructions in the kernel text as
 * generated by the compiler.
 *
 * @buf: temporary buffer on which the patching operates. This buffer is
 * eventually text-poked into the kernel image.
 *
 * @replacement/@repl: pointer to the opcodes which are replacing @instr, located
 * in the .altinstr_replacement section.
 */

/*
 * Fill the buffer with a single effective instruction of size @len.
 *
 * In order not to issue an ORC stack depth tracking CFI entry (Call Frame Info)
 * for every single-byte NOP, try to generate the maximally available NOP of
 * size <= ASM_NOP_MAX such that only a single CFI entry is generated (vs one for
 * each single-byte NOPs). If @len to fill out is > ASM_NOP_MAX, pad with INT3 and
 * *jump* over instead of executing long and daft NOPs.
 */
static void add_nop(u8 *buf, unsigned int len)
{
	u8 *target = buf + len;

	if (!len)
		return;

	if (len <= ASM_NOP_MAX) {
		memcpy(buf, x86_nops[len], len);
		return;
	}

	if (len < 128) {
		__text_gen_insn(buf, JMP8_INSN_OPCODE, buf, target, JMP8_INSN_SIZE);
		buf += JMP8_INSN_SIZE;
	} else {
		__text_gen_insn(buf, JMP32_INSN_OPCODE, buf, target, JMP32_INSN_SIZE);
		buf += JMP32_INSN_SIZE;
	}

	for (;buf < target; buf++)
		*buf = INT3_INSN_OPCODE;
}

/*
 * Matches NOP and NOPL, not any of the other possible NOPs.
 */
static bool insn_is_nop(struct insn *insn)
{
	/* Anything NOP, but no REP NOP */
	if (insn->opcode.bytes[0] == 0x90 &&
	    (!insn->prefixes.nbytes || insn->prefixes.bytes[0] != 0xF3))
		return true;

	/* NOPL */
	if (insn->opcode.bytes[0] == 0x0F && insn->opcode.bytes[1] == 0x1F)
		return true;

	/* TODO: more nops */

	return false;
}

/*
 * Find the offset of the first non-NOP instruction starting at @offset
 * but no further than @len.
 */
static int skip_nops(u8 *buf, int offset, int len)
{
	struct insn insn;

	for (; offset < len; offset += insn.length) {
		if (insn_decode_kernel(&insn, &buf[offset]))
			break;

		if (!insn_is_nop(&insn))
			break;
	}

	return offset;
}

/*
 * "noinline" to cause control flow change and thus invalidate I$ and
 * cause refetch after modification.
 */
static void noinline optimize_nops(const u8 * const instr, u8 *buf, size_t len)
{
	for (int next, i = 0; i < len; i = next) {
		struct insn insn;

		if (insn_decode_kernel(&insn, &buf[i]))
			return;

		next = i + insn.length;

		if (insn_is_nop(&insn)) {
			int nop = i;

			/* Has the NOP already been optimized? */
			if (i + insn.length == len)
				return;

			next = skip_nops(buf, next, len);

			add_nop(buf + nop, next - nop);
			DUMP_BYTES(ALT, buf, len, "%px: [%d:%d) optimized NOPs: ", instr, nop, next);
		}
	}
}

/*
 * In this context, "source" is where the instructions are placed in the
 * section .altinstr_replacement, for example during kernel build by the
 * toolchain.
 * "Destination" is where the instructions are being patched in by this
 * machinery.
 *
 * The source offset is:
 *
 *   src_imm = target - src_next_ip                  (1)
 *
 * and the target offset is:
 *
 *   dst_imm = target - dst_next_ip                  (2)
 *
 * so rework (1) as an expression for target like:
 *
 *   target = src_imm + src_next_ip                  (1a)
 *
 * and substitute in (2) to get:
 *
 *   dst_imm = (src_imm + src_next_ip) - dst_next_ip (3)
 *
 * Now, since the instruction stream is 'identical' at src and dst (it
 * is being copied after all) it can be stated that:
 *
 *   src_next_ip = src + ip_offset
 *   dst_next_ip = dst + ip_offset                   (4)
 *
 * Substitute (4) in (3) and observe ip_offset being cancelled out to
 * obtain:
 *
 *   dst_imm = src_imm + (src + ip_offset) - (dst + ip_offset)
 *           = src_imm + src - dst + ip_offset - ip_offset
 *           = src_imm + src - dst                   (5)
 *
 * IOW, only the relative displacement of the code block matters.
 */

#define apply_reloc_n(n_, p_, d_)				\
	do {							\
		s32 v = *(s##n_ *)(p_);				\
		v += (d_);					\
		BUG_ON((v >> 31) != (v >> (n_-1)));		\
		*(s##n_ *)(p_) = (s##n_)v;			\
	} while (0)


static __always_inline
void apply_reloc(int n, void *ptr, uintptr_t diff)
{
	switch (n) {
	case 1: apply_reloc_n(8, ptr, diff); break;
	case 2: apply_reloc_n(16, ptr, diff); break;
	case 4: apply_reloc_n(32, ptr, diff); break;
	default: BUG();
	}
}

static __always_inline
bool need_reloc(unsigned long offset, u8 *src, size_t src_len)
{
	u8 *target = src + offset;
	/*
	 * If the target is inside the patched block, it's relative to the
	 * block itself and does not need relocation.
	 */
	return (target < src || target > src + src_len);
}

static void __apply_relocation(u8 *buf, const u8 * const instr, size_t instrlen, u8 *repl, size_t repl_len)
{
	for (int next, i = 0; i < instrlen; i = next) {
		struct insn insn;

		if (WARN_ON_ONCE(insn_decode_kernel(&insn, &buf[i])))
			return;

		next = i + insn.length;

		switch (insn.opcode.bytes[0]) {
		case 0x0f:
			if (insn.opcode.bytes[1] < 0x80 ||
			    insn.opcode.bytes[1] > 0x8f)
				break;

			fallthrough;	/* Jcc.d32 */
		case 0x70 ... 0x7f:	/* Jcc.d8 */
		case JMP8_INSN_OPCODE:
		case JMP32_INSN_OPCODE:
		case CALL_INSN_OPCODE:
			if (need_reloc(next + insn.immediate.value, repl, repl_len)) {
				apply_reloc(insn.immediate.nbytes,
					    buf + i + insn_offset_immediate(&insn),
					    repl - instr);
			}

			/*
			 * Where possible, convert JMP.d32 into JMP.d8.
			 */
			if (insn.opcode.bytes[0] == JMP32_INSN_OPCODE) {
				s32 imm = insn.immediate.value;
				imm += repl - instr;
				imm += JMP32_INSN_SIZE - JMP8_INSN_SIZE;
				if ((imm >> 31) == (imm >> 7)) {
					buf[i+0] = JMP8_INSN_OPCODE;
					buf[i+1] = (s8)imm;

					memset(&buf[i+2], INT3_INSN_OPCODE, insn.length - 2);
				}
			}
			break;
		}

		if (insn_rip_relative(&insn)) {
			if (need_reloc(next + insn.displacement.value, repl, repl_len)) {
				apply_reloc(insn.displacement.nbytes,
					    buf + i + insn_offset_displacement(&insn),
					    repl - instr);
			}
		}
	}
}

void text_poke_apply_relocation(u8 *buf, const u8 * const instr, size_t instrlen, u8 *repl, size_t repl_len)
{
	__apply_relocation(buf, instr, instrlen, repl, repl_len);
	optimize_nops(instr, buf, instrlen);
}

/* Low-level backend functions usable from alternative code replacements. */
DEFINE_ASM_FUNC(nop_func, "", .entry.text);
EXPORT_SYMBOL_GPL(nop_func);

noinstr void BUG_func(void)
{
	BUG();
}
EXPORT_SYMBOL(BUG_func);

#define CALL_RIP_REL_OPCODE	0xff
#define CALL_RIP_REL_MODRM	0x15

/*
 * Rewrite the "call BUG_func" replacement to point to the target of the
 * indirect pv_ops call "call *disp(%ip)".
 */
static int alt_replace_call(u8 *instr, u8 *insn_buff, struct alt_instr *a)
{
	void *target, *bug = &BUG_func;
	s32 disp;

	if (a->replacementlen != 5 || insn_buff[0] != CALL_INSN_OPCODE) {
		pr_err("ALT_FLAG_DIRECT_CALL set for a non-call replacement instruction\n");
		BUG();
	}

	if (a->instrlen != 6 ||
	    instr[0] != CALL_RIP_REL_OPCODE ||
	    instr[1] != CALL_RIP_REL_MODRM) {
		pr_err("ALT_FLAG_DIRECT_CALL set for unrecognized indirect call\n");
		BUG();
	}

	/* Skip CALL_RIP_REL_OPCODE and CALL_RIP_REL_MODRM */
	disp = *(s32 *)(instr + 2);
#ifdef CONFIG_X86_64
	/* ff 15 00 00 00 00   call   *0x0(%rip) */
	/* target address is stored at "next instruction + disp". */
	target = *(void **)(instr + a->instrlen + disp);
#else
	/* ff 15 00 00 00 00   call   *0x0 */
	/* target address is stored at disp. */
	target = *(void **)disp;
#endif
	if (!target)
		target = bug;

	/* (BUG_func - .) + (target - BUG_func) := target - . */
	*(s32 *)(insn_buff + 1) += target - bug;

	if (target == &nop_func)
		return 0;

	return 5;
}

static inline u8 * instr_va(struct alt_instr *i)
{
	return (u8 *)&i->instr_offset + i->instr_offset;
}

/*
 * Replace instructions with better alternatives for this CPU type. This runs
 * before SMP is initialized to avoid SMP problems with self modifying code.
 * This implies that asymmetric systems where APs have less capabilities than
 * the boot processor are not handled. Tough. Make sure you disable such
 * features by hand.
 *
 * Marked "noinline" to cause control flow change and thus insn cache
 * to refetch changed I$ lines.
 */
void __init_or_module noinline apply_alternatives(struct alt_instr *start,
						  struct alt_instr *end)
{
	u8 insn_buff[MAX_PATCH_LEN];
	u8 *instr, *replacement;
	struct alt_instr *a, *b;

	DPRINTK(ALT, "alt table %px, -> %px", start, end);

	/*
	 * KASAN_SHADOW_START is defined using
	 * cpu_feature_enabled(X86_FEATURE_LA57) and is therefore patched here.
	 * During the process, KASAN becomes confused seeing partial LA57
	 * conversion and triggers a false-positive out-of-bound report.
	 *
	 * Disable KASAN until the patching is complete.
	 */
	kasan_disable_current();

	/*
	 * The scan order should be from start to end. A later scanned
	 * alternative code can overwrite previously scanned alternative code.
	 * Some kernel functions (e.g. memcpy, memset, etc) use this order to
	 * patch code.
	 *
	 * So be careful if you want to change the scan order to any other
	 * order.
	 */
	for (a = start; a < end; a++) {
		int insn_buff_sz = 0;

		/*
		 * In case of nested ALTERNATIVE()s the outer alternative might
		 * add more padding. To ensure consistent patching find the max
		 * padding for all alt_instr entries for this site (nested
		 * alternatives result in consecutive entries).
		 */
		for (b = a+1; b < end && instr_va(b) == instr_va(a); b++) {
			u8 len = max(a->instrlen, b->instrlen);
			a->instrlen = b->instrlen = len;
		}

		instr = instr_va(a);
		replacement = (u8 *)&a->repl_offset + a->repl_offset;
		BUG_ON(a->instrlen > sizeof(insn_buff));
		BUG_ON(a->cpuid >= (NCAPINTS + NBUGINTS) * 32);

		/*
		 * Patch if either:
		 * - feature is present
		 * - feature not present but ALT_FLAG_NOT is set to mean,
		 *   patch if feature is *NOT* present.
		 */
		if (!boot_cpu_has(a->cpuid) == !(a->flags & ALT_FLAG_NOT)) {
			memcpy(insn_buff, instr, a->instrlen);
			optimize_nops(instr, insn_buff, a->instrlen);
			text_poke_early(instr, insn_buff, a->instrlen);
			continue;
		}

		DPRINTK(ALT, "feat: %d*32+%d, old: (%pS (%px) len: %d), repl: (%px, len: %d) flags: 0x%x",
			a->cpuid >> 5,
			a->cpuid & 0x1f,
			instr, instr, a->instrlen,
			replacement, a->replacementlen, a->flags);

		memcpy(insn_buff, replacement, a->replacementlen);
		insn_buff_sz = a->replacementlen;

		if (a->flags & ALT_FLAG_DIRECT_CALL) {
			insn_buff_sz = alt_replace_call(instr, insn_buff, a);
			if (insn_buff_sz < 0)
				continue;
		}

		for (; insn_buff_sz < a->instrlen; insn_buff_sz++)
			insn_buff[insn_buff_sz] = 0x90;

		text_poke_apply_relocation(insn_buff, instr, a->instrlen, replacement, a->replacementlen);

		DUMP_BYTES(ALT, instr, a->instrlen, "%px:   old_insn: ", instr);
		DUMP_BYTES(ALT, replacement, a->replacementlen, "%px:   rpl_insn: ", replacement);
		DUMP_BYTES(ALT, insn_buff, insn_buff_sz, "%px: final_insn: ", instr);

		text_poke_early(instr, insn_buff, insn_buff_sz);
	}

	kasan_enable_current();
}

static inline bool is_jcc32(struct insn *insn)
{
	/* Jcc.d32 second opcode byte is in the range: 0x80-0x8f */
	return insn->opcode.bytes[0] == 0x0f && (insn->opcode.bytes[1] & 0xf0) == 0x80;
}

#if defined(CONFIG_MITIGATION_RETPOLINE) && defined(CONFIG_OBJTOOL)

/*
 * CALL/JMP *%\reg
 */
static int emit_indirect(int op, int reg, u8 *bytes)
{
	int i = 0;
	u8 modrm;

	switch (op) {
	case CALL_INSN_OPCODE:
		modrm = 0x10; /* Reg = 2; CALL r/m */
		break;

	case JMP32_INSN_OPCODE:
		modrm = 0x20; /* Reg = 4; JMP r/m */
		break;

	default:
		WARN_ON_ONCE(1);
		return -1;
	}

	if (reg >= 8) {
		bytes[i++] = 0x41; /* REX.B prefix */
		reg -= 8;
	}

	modrm |= 0xc0; /* Mod = 3 */
	modrm += reg;

	bytes[i++] = 0xff; /* opcode */
	bytes[i++] = modrm;

	return i;
}

static int __emit_trampoline(void *addr, struct insn *insn, u8 *bytes,
			     void *call_dest, void *jmp_dest)
{
	u8 op = insn->opcode.bytes[0];
	int i = 0;

	/*
	 * Clang does 'weird' Jcc __x86_indirect_thunk_r11 conditional
	 * tail-calls. Deal with them.
	 */
	if (is_jcc32(insn)) {
		bytes[i++] = op;
		op = insn->opcode.bytes[1];
		goto clang_jcc;
	}

	if (insn->length == 6)
		bytes[i++] = 0x2e; /* CS-prefix */

	switch (op) {
	case CALL_INSN_OPCODE:
		__text_gen_insn(bytes+i, op, addr+i,
				call_dest,
				CALL_INSN_SIZE);
		i += CALL_INSN_SIZE;
		break;

	case JMP32_INSN_OPCODE:
clang_jcc:
		__text_gen_insn(bytes+i, op, addr+i,
				jmp_dest,
				JMP32_INSN_SIZE);
		i += JMP32_INSN_SIZE;
		break;

	default:
		WARN(1, "%pS %px %*ph\n", addr, addr, 6, addr);
		return -1;
	}

	WARN_ON_ONCE(i != insn->length);

	return i;
}

static int emit_call_track_retpoline(void *addr, struct insn *insn, int reg, u8 *bytes)
{
	return __emit_trampoline(addr, insn, bytes,
				 __x86_indirect_call_thunk_array[reg],
				 __x86_indirect_jump_thunk_array[reg]);
}

#ifdef CONFIG_MITIGATION_ITS
static int emit_its_trampoline(void *addr, struct insn *insn, int reg, u8 *bytes)
{
	u8 *thunk = __x86_indirect_its_thunk_array[reg];
	u8 *tmp = its_allocate_thunk(reg);

	if (tmp)
		thunk = tmp;

	return __emit_trampoline(addr, insn, bytes, thunk, thunk);
}

/* Check if an indirect branch is at ITS-unsafe address */
static bool cpu_wants_indirect_its_thunk_at(unsigned long addr, int reg)
{
	if (!cpu_feature_enabled(X86_FEATURE_INDIRECT_THUNK_ITS))
		return false;

	/* Indirect branch opcode is 2 or 3 bytes depending on reg */
	addr += 1 + reg / 8;

	/* Lower-half of the cacheline? */
	return !(addr & 0x20);
}
#else /* CONFIG_MITIGATION_ITS */

#ifdef CONFIG_FINEIBT
static bool cpu_wants_indirect_its_thunk_at(unsigned long addr, int reg)
{
	return false;
}
#endif

#endif /* CONFIG_MITIGATION_ITS */

/*
 * Rewrite the compiler generated retpoline thunk calls.
 *
 * For spectre_v2=off (!X86_FEATURE_RETPOLINE), rewrite them into immediate
 * indirect instructions, avoiding the extra indirection.
 *
 * For example, convert:
 *
 *   CALL __x86_indirect_thunk_\reg
 *
 * into:
 *
 *   CALL *%\reg
 *
 * It also tries to inline spectre_v2=retpoline,lfence when size permits.
 */
static int patch_retpoline(void *addr, struct insn *insn, u8 *bytes)
{
	retpoline_thunk_t *target;
	int reg, ret, i = 0;
	u8 op, cc;

	target = addr + insn->length + insn->immediate.value;
	reg = target - __x86_indirect_thunk_array;

	if (WARN_ON_ONCE(reg & ~0xf))
		return -1;

	/* If anyone ever does: CALL/JMP *%rsp, we're in deep trouble. */
	BUG_ON(reg == 4);

	if (cpu_feature_enabled(X86_FEATURE_RETPOLINE) &&
	    !cpu_feature_enabled(X86_FEATURE_RETPOLINE_LFENCE)) {
		if (cpu_feature_enabled(X86_FEATURE_CALL_DEPTH))
			return emit_call_track_retpoline(addr, insn, reg, bytes);

		return -1;
	}

	op = insn->opcode.bytes[0];

	/*
	 * Convert:
	 *
	 *   Jcc.d32 __x86_indirect_thunk_\reg
	 *
	 * into:
	 *
	 *   Jncc.d8 1f
	 *   [ LFENCE ]
	 *   JMP *%\reg
	 *   [ NOP ]
	 * 1:
	 */
	if (is_jcc32(insn)) {
		cc = insn->opcode.bytes[1] & 0xf;
		cc ^= 1; /* invert condition */

		bytes[i++] = 0x70 + cc;        /* Jcc.d8 */
		bytes[i++] = insn->length - 2; /* sizeof(Jcc.d8) == 2 */

		/* Continue as if: JMP.d32 __x86_indirect_thunk_\reg */
		op = JMP32_INSN_OPCODE;
	}

	/*
	 * For RETPOLINE_LFENCE: prepend the indirect CALL/JMP with an LFENCE.
	 */
	if (cpu_feature_enabled(X86_FEATURE_RETPOLINE_LFENCE)) {
		bytes[i++] = 0x0f;
		bytes[i++] = 0xae;
		bytes[i++] = 0xe8; /* LFENCE */
	}

#ifdef CONFIG_MITIGATION_ITS
	/*
	 * Check if the address of last byte of emitted-indirect is in
	 * lower-half of the cacheline. Such branches need ITS mitigation.
	 */
	if (cpu_wants_indirect_its_thunk_at((unsigned long)addr + i, reg))
		return emit_its_trampoline(addr, insn, reg, bytes);
#endif

	ret = emit_indirect(op, reg, bytes + i);
	if (ret < 0)
		return ret;
	i += ret;

	/*
	 * The compiler is supposed to EMIT an INT3 after every unconditional
	 * JMP instruction due to AMD BTC. However, if the compiler is too old
	 * or MITIGATION_SLS isn't enabled, we still need an INT3 after
	 * indirect JMPs even on Intel.
	 */
	if (op == JMP32_INSN_OPCODE && i < insn->length)
		bytes[i++] = INT3_INSN_OPCODE;

	for (; i < insn->length;)
		bytes[i++] = BYTES_NOP1;

	return i;
}

/*
 * Generated by 'objtool --retpoline'.
 */
void __init_or_module noinline apply_retpolines(s32 *start, s32 *end)
{
	s32 *s;

	for (s = start; s < end; s++) {
		void *addr = (void *)s + *s;
		struct insn insn;
		int len, ret;
		u8 bytes[16];
		u8 op1, op2;
		u8 *dest;

		ret = insn_decode_kernel(&insn, addr);
		if (WARN_ON_ONCE(ret < 0))
			continue;

		op1 = insn.opcode.bytes[0];
		op2 = insn.opcode.bytes[1];

		switch (op1) {
		case 0x70 ... 0x7f:	/* Jcc.d8 */
			/* See cfi_paranoid. */
			WARN_ON_ONCE(cfi_mode != CFI_FINEIBT);
			continue;

		case CALL_INSN_OPCODE:
		case JMP32_INSN_OPCODE:
			/* Check for cfi_paranoid + ITS */
			dest = addr + insn.length + insn.immediate.value;
			if (dest[-1] == 0xea && (dest[0] & 0xf0) == 0x70) {
				WARN_ON_ONCE(cfi_mode != CFI_FINEIBT);
				continue;
			}
			break;

		case 0x0f: /* escape */
			if (op2 >= 0x80 && op2 <= 0x8f)
				break;
			fallthrough;
		default:
			WARN_ON_ONCE(1);
			continue;
		}

		DPRINTK(RETPOLINE, "retpoline at: %pS (%px) len: %d to: %pS",
			addr, addr, insn.length,
			addr + insn.length + insn.immediate.value);

		len = patch_retpoline(addr, &insn, bytes);
		if (len == insn.length) {
			optimize_nops(addr, bytes, len);
			DUMP_BYTES(RETPOLINE, ((u8*)addr),  len, "%px: orig: ", addr);
			DUMP_BYTES(RETPOLINE, ((u8*)bytes), len, "%px: repl: ", addr);
			text_poke_early(addr, bytes, len);
		}
	}
}

#ifdef CONFIG_MITIGATION_RETHUNK

bool cpu_wants_rethunk(void)
{
	return cpu_feature_enabled(X86_FEATURE_RETHUNK);
}

bool cpu_wants_rethunk_at(void *addr)
{
	if (!cpu_feature_enabled(X86_FEATURE_RETHUNK))
		return false;
	if (x86_return_thunk != its_return_thunk)
		return true;

	return !((unsigned long)addr & 0x20);
}

/*
 * Rewrite the compiler generated return thunk tail-calls.
 *
 * For example, convert:
 *
 *   JMP __x86_return_thunk
 *
 * into:
 *
 *   RET
 */
static int patch_return(void *addr, struct insn *insn, u8 *bytes)
{
	int i = 0;

	/* Patch the custom return thunks... */
	if (cpu_wants_rethunk_at(addr)) {
		i = JMP32_INSN_SIZE;
		__text_gen_insn(bytes, JMP32_INSN_OPCODE, addr, x86_return_thunk, i);
	} else {
		/* ... or patch them out if not needed. */
		bytes[i++] = RET_INSN_OPCODE;
	}

	for (; i < insn->length;)
		bytes[i++] = INT3_INSN_OPCODE;
	return i;
}

void __init_or_module noinline apply_returns(s32 *start, s32 *end)
{
	s32 *s;

	if (cpu_wants_rethunk())
		static_call_force_reinit();

	for (s = start; s < end; s++) {
		void *dest = NULL, *addr = (void *)s + *s;
		struct insn insn;
		int len, ret;
		u8 bytes[16];
		u8 op;

		ret = insn_decode_kernel(&insn, addr);
		if (WARN_ON_ONCE(ret < 0))
			continue;

		op = insn.opcode.bytes[0];
		if (op == JMP32_INSN_OPCODE)
			dest = addr + insn.length + insn.immediate.value;

		if (__static_call_fixup(addr, op, dest) ||
		    WARN_ONCE(dest != &__x86_return_thunk,
			      "missing return thunk: %pS-%pS: %*ph",
			      addr, dest, 5, addr))
			continue;

		DPRINTK(RET, "return thunk at: %pS (%px) len: %d to: %pS",
			addr, addr, insn.length,
			addr + insn.length + insn.immediate.value);

		len = patch_return(addr, &insn, bytes);
		if (len == insn.length) {
			DUMP_BYTES(RET, ((u8*)addr),  len, "%px: orig: ", addr);
			DUMP_BYTES(RET, ((u8*)bytes), len, "%px: repl: ", addr);
			text_poke_early(addr, bytes, len);
		}
	}
}
#else /* !CONFIG_MITIGATION_RETHUNK: */
void __init_or_module noinline apply_returns(s32 *start, s32 *end) { }
#endif /* !CONFIG_MITIGATION_RETHUNK */

#else /* !CONFIG_MITIGATION_RETPOLINE || !CONFIG_OBJTOOL */

void __init_or_module noinline apply_retpolines(s32 *start, s32 *end) { }
void __init_or_module noinline apply_returns(s32 *start, s32 *end) { }

#endif /* !CONFIG_MITIGATION_RETPOLINE || !CONFIG_OBJTOOL */

#ifdef CONFIG_X86_KERNEL_IBT

__noendbr bool is_endbr(u32 *val)
{
	u32 endbr;

	__get_kernel_nofault(&endbr, val, u32, Efault);
	return __is_endbr(endbr);

Efault:
	return false;
}

#ifdef CONFIG_FINEIBT

static __noendbr bool exact_endbr(u32 *val)
{
	u32 endbr;

	__get_kernel_nofault(&endbr, val, u32, Efault);
	return endbr == gen_endbr();

Efault:
	return false;
}

#endif

static void poison_cfi(void *addr);

static void __init_or_module poison_endbr(void *addr)
{
	u32 poison = gen_endbr_poison();

	if (WARN_ON_ONCE(!is_endbr(addr)))
		return;

	DPRINTK(ENDBR, "ENDBR at: %pS (%px)", addr, addr);

	/*
	 * When we have IBT, the lack of ENDBR will trigger #CP
	 */
	DUMP_BYTES(ENDBR, ((u8*)addr), 4, "%px: orig: ", addr);
	DUMP_BYTES(ENDBR, ((u8*)&poison), 4, "%px: repl: ", addr);
	text_poke_early(addr, &poison, 4);
}

/*
 * Generated by: objtool --ibt
 *
 * Seal the functions for indirect calls by clobbering the ENDBR instructions
 * and the kCFI hash value.
 */
void __init_or_module noinline apply_seal_endbr(s32 *start, s32 *end)
{
	s32 *s;

	for (s = start; s < end; s++) {
		void *addr = (void *)s + *s;

		poison_endbr(addr);
		if (IS_ENABLED(CONFIG_FINEIBT))
			poison_cfi(addr - 16);
	}
}

#else /* !CONFIG_X86_KERNEL_IBT: */

void __init_or_module apply_seal_endbr(s32 *start, s32 *end) { }

#endif /* !CONFIG_X86_KERNEL_IBT */

#ifdef CONFIG_CFI_AUTO_DEFAULT
# define __CFI_DEFAULT CFI_AUTO
#elif defined(CONFIG_CFI_CLANG)
# define __CFI_DEFAULT CFI_KCFI
#else
# define __CFI_DEFAULT CFI_OFF
#endif

enum cfi_mode cfi_mode __ro_after_init = __CFI_DEFAULT;

#ifdef CONFIG_FINEIBT_BHI
bool cfi_bhi __ro_after_init = false;
#endif

#ifdef CONFIG_CFI_CLANG
struct bpf_insn;

/* Must match bpf_func_t / DEFINE_BPF_PROG_RUN() */
extern unsigned int __bpf_prog_runX(const void *ctx,
				    const struct bpf_insn *insn);

KCFI_REFERENCE(__bpf_prog_runX);

/* u32 __ro_after_init cfi_bpf_hash = __kcfi_typeid___bpf_prog_runX; */
asm (
"	.pushsection	.data..ro_after_init,\"aw\",@progbits	\n"
"	.type	cfi_bpf_hash,@object				\n"
"	.globl	cfi_bpf_hash					\n"
"	.p2align	2, 0x0					\n"
"cfi_bpf_hash:							\n"
"	.long	__kcfi_typeid___bpf_prog_runX			\n"
"	.size	cfi_bpf_hash, 4					\n"
"	.popsection						\n"
);

/* Must match bpf_callback_t */
extern u64 __bpf_callback_fn(u64, u64, u64, u64, u64);

KCFI_REFERENCE(__bpf_callback_fn);

/* u32 __ro_after_init cfi_bpf_subprog_hash = __kcfi_typeid___bpf_callback_fn; */
asm (
"	.pushsection	.data..ro_after_init,\"aw\",@progbits	\n"
"	.type	cfi_bpf_subprog_hash,@object			\n"
"	.globl	cfi_bpf_subprog_hash				\n"
"	.p2align	2, 0x0					\n"
"cfi_bpf_subprog_hash:						\n"
"	.long	__kcfi_typeid___bpf_callback_fn			\n"
"	.size	cfi_bpf_subprog_hash, 4				\n"
"	.popsection						\n"
);

u32 cfi_get_func_hash(void *func)
{
	u32 hash;

	func -= cfi_get_offset();
	switch (cfi_mode) {
	case CFI_FINEIBT:
		func += 7;
		break;
	case CFI_KCFI:
		func += 1;
		break;
	default:
		return 0;
	}

	if (get_kernel_nofault(hash, func))
		return 0;

	return hash;
}

int cfi_get_func_arity(void *func)
{
	bhi_thunk *target;
	s32 disp;

	if (cfi_mode != CFI_FINEIBT && !cfi_bhi)
		return 0;

	if (get_kernel_nofault(disp, func - 4))
		return 0;

	target = func + disp;
	return target - __bhi_args;
}
#endif

#ifdef CONFIG_FINEIBT

static bool cfi_rand __ro_after_init = true;
static u32  cfi_seed __ro_after_init;

/*
 * Re-hash the CFI hash with a boot-time seed while making sure the result is
 * not a valid ENDBR instruction.
 */
static u32 cfi_rehash(u32 hash)
{
	hash ^= cfi_seed;
	while (unlikely(__is_endbr(hash) || __is_endbr(-hash))) {
		bool lsb = hash & 1;
		hash >>= 1;
		if (lsb)
			hash ^= 0x80200003;
	}
	return hash;
}

static __init int cfi_parse_cmdline(char *str)
{
	if (!str)
		return -EINVAL;

	while (str) {
		char *next = strchr(str, ',');
		if (next) {
			*next = 0;
			next++;
		}

		if (!strcmp(str, "auto")) {
			cfi_mode = CFI_AUTO;
		} else if (!strcmp(str, "off")) {
			cfi_mode = CFI_OFF;
			cfi_rand = false;
		} else if (!strcmp(str, "kcfi")) {
			cfi_mode = CFI_KCFI;
		} else if (!strcmp(str, "fineibt")) {
			cfi_mode = CFI_FINEIBT;
		} else if (!strcmp(str, "norand")) {
			cfi_rand = false;
		} else if (!strcmp(str, "warn")) {
			pr_alert("CFI mismatch non-fatal!\n");
			cfi_warn = true;
		} else if (!strcmp(str, "paranoid")) {
			if (cfi_mode == CFI_FINEIBT) {
				cfi_paranoid = true;
			} else {
				pr_err("Ignoring paranoid; depends on fineibt.\n");
			}
		} else if (!strcmp(str, "bhi")) {
#ifdef CONFIG_FINEIBT_BHI
			if (cfi_mode == CFI_FINEIBT) {
				cfi_bhi = true;
			} else {
				pr_err("Ignoring bhi; depends on fineibt.\n");
			}
#else
			pr_err("Ignoring bhi; depends on FINEIBT_BHI=y.\n");
#endif
		} else {
			pr_err("Ignoring unknown cfi option (%s).", str);
		}

		str = next;
	}

	return 0;
}
early_param("cfi", cfi_parse_cmdline);

/*
 * kCFI						FineIBT
 *
 * __cfi_\func:					__cfi_\func:
 *	movl   $0x12345678,%eax		// 5	     endbr64			// 4
 *	nop					     subl   $0x12345678,%r10d   // 7
 *	nop					     jne    __cfi_\func+6	// 2
 *	nop					     nop3			// 3
 *	nop
 *	nop
 *	nop
 *	nop
 *	nop
 *	nop
 *	nop
 *	nop
 *
 *
 * caller:					caller:
 *	movl	$(-0x12345678),%r10d	 // 6	     movl   $0x12345678,%r10d	// 6
 *	addl	$-15(%r11),%r10d	 // 4	     lea    -0x10(%r11),%r11	// 4
 *	je	1f			 // 2	     nop4			// 4
 *	ud2				 // 2
 * 1:	cs call	__x86_indirect_thunk_r11 // 6	     call   *%r11; nop3;	// 6
 *
 */

/*
 * <fineibt_preamble_start>:
 *  0:   f3 0f 1e fa             endbr64
 *  4:   41 81 <ea> 78 56 34 12  sub    $0x12345678, %r10d
 *  b:   75 f9                   jne    6 <fineibt_preamble_start+0x6>
 *  d:   0f 1f 00                nopl   (%rax)
 *
 * Note that the JNE target is the 0xEA byte inside the SUB, this decodes as
 * (bad) on x86_64 and raises #UD.
 */
asm(	".pushsection .rodata				\n"
	"fineibt_preamble_start:			\n"
	"	endbr64					\n"
	"	subl	$0x12345678, %r10d		\n"
	"fineibt_preamble_bhi:				\n"
	"	jne	fineibt_preamble_start+6	\n"
	ASM_NOP3
	"fineibt_preamble_end:				\n"
	".popsection\n"
);

extern u8 fineibt_preamble_start[];
extern u8 fineibt_preamble_bhi[];
extern u8 fineibt_preamble_end[];

#define fineibt_preamble_size (fineibt_preamble_end - fineibt_preamble_start)
#define fineibt_preamble_bhi  (fineibt_preamble_bhi - fineibt_preamble_start)
#define fineibt_preamble_ud   6
#define fineibt_preamble_hash 7

/*
 * <fineibt_caller_start>:
 *  0:   41 ba 78 56 34 12       mov    $0x12345678, %r10d
 *  6:   4d 8d 5b f0             lea    -0x10(%r11), %r11
 *  a:   0f 1f 40 00             nopl   0x0(%rax)
 */
asm(	".pushsection .rodata			\n"
	"fineibt_caller_start:			\n"
	"	movl	$0x12345678, %r10d	\n"
	"	lea	-0x10(%r11), %r11	\n"
	ASM_NOP4
	"fineibt_caller_end:			\n"
	".popsection				\n"
);

extern u8 fineibt_caller_start[];
extern u8 fineibt_caller_end[];

#define fineibt_caller_size (fineibt_caller_end - fineibt_caller_start)
#define fineibt_caller_hash 2

#define fineibt_caller_jmp (fineibt_caller_size - 2)

/*
 * Since FineIBT does hash validation on the callee side it is prone to
 * circumvention attacks where a 'naked' ENDBR instruction exists that
 * is not part of the fineibt_preamble sequence.
 *
 * Notably the x86 entry points must be ENDBR and equally cannot be
 * fineibt_preamble.
 *
 * The fineibt_paranoid caller sequence adds additional caller side
 * hash validation. This stops such circumvention attacks dead, but at the cost
 * of adding a load.
 *
 * <fineibt_paranoid_start>:
 *  0:   41 ba 78 56 34 12       mov    $0x12345678, %r10d
 *  6:   45 3b 53 f7             cmp    -0x9(%r11), %r10d
 *  a:   4d 8d 5b <f0>           lea    -0x10(%r11), %r11
 *  e:   75 fd                   jne    d <fineibt_paranoid_start+0xd>
 * 10:   41 ff d3                call   *%r11
 * 13:   90                      nop
 *
 * Notably LEA does not modify flags and can be reordered with the CMP,
 * avoiding a dependency. Again, using a non-taken (backwards) branch
 * for the failure case, abusing LEA's immediate 0xf0 as LOCK prefix for the
 * Jcc.d8, causing #UD.
 */
asm(	".pushsection .rodata				\n"
	"fineibt_paranoid_start:			\n"
	"	movl	$0x12345678, %r10d		\n"
	"	cmpl	-9(%r11), %r10d			\n"
	"	lea	-0x10(%r11), %r11		\n"
	"	jne	fineibt_paranoid_start+0xd	\n"
	"fineibt_paranoid_ind:				\n"
	"	call	*%r11				\n"
	"	nop					\n"
	"fineibt_paranoid_end:				\n"
	".popsection					\n"
);

extern u8 fineibt_paranoid_start[];
extern u8 fineibt_paranoid_ind[];
extern u8 fineibt_paranoid_end[];

#define fineibt_paranoid_size (fineibt_paranoid_end - fineibt_paranoid_start)
#define fineibt_paranoid_ind  (fineibt_paranoid_ind - fineibt_paranoid_start)
#define fineibt_paranoid_ud   0xd

static u32 decode_preamble_hash(void *addr, int *reg)
{
	u8 *p = addr;

	/* b8+reg 78 56 34 12          movl    $0x12345678,\reg */
	if (p[0] >= 0xb8 && p[0] < 0xc0) {
		if (reg)
			*reg = p[0] - 0xb8;
		return *(u32 *)(addr + 1);
	}

	return 0; /* invalid hash value */
}

static u32 decode_caller_hash(void *addr)
{
	u8 *p = addr;

	/* 41 ba 88 a9 cb ed       mov    $(-0x12345678),%r10d */
	if (p[0] == 0x41 && p[1] == 0xba)
		return -*(u32 *)(addr + 2);

	/* e8 0c 88 a9 cb ed	   jmp.d8  +12 */
	if (p[0] == JMP8_INSN_OPCODE && p[1] == fineibt_caller_jmp)
		return -*(u32 *)(addr + 2);

	return 0; /* invalid hash value */
}

/* .retpoline_sites */
static int cfi_disable_callers(s32 *start, s32 *end)
{
	/*
	 * Disable kCFI by patching in a JMP.d8, this leaves the hash immediate
	 * in tact for later usage. Also see decode_caller_hash() and
	 * cfi_rewrite_callers().
	 */
	const u8 jmp[] = { JMP8_INSN_OPCODE, fineibt_caller_jmp };
	s32 *s;

	for (s = start; s < end; s++) {
		void *addr = (void *)s + *s;
		u32 hash;

		addr -= fineibt_caller_size;
		hash = decode_caller_hash(addr);
		if (!hash) /* nocfi callers */
			continue;

		text_poke_early(addr, jmp, 2);
	}

	return 0;
}

static int cfi_enable_callers(s32 *start, s32 *end)
{
	/*
	 * Re-enable kCFI, undo what cfi_disable_callers() did.
	 */
	const u8 mov[] = { 0x41, 0xba };
	s32 *s;

	for (s = start; s < end; s++) {
		void *addr = (void *)s + *s;
		u32 hash;

		addr -= fineibt_caller_size;
		hash = decode_caller_hash(addr);
		if (!hash) /* nocfi callers */
			continue;

		text_poke_early(addr, mov, 2);
	}

	return 0;
}

/* .cfi_sites */
static int cfi_rand_preamble(s32 *start, s32 *end)
{
	s32 *s;

	for (s = start; s < end; s++) {
		void *addr = (void *)s + *s;
		u32 hash;

		hash = decode_preamble_hash(addr, NULL);
		if (WARN(!hash, "no CFI hash found at: %pS %px %*ph\n",
			 addr, addr, 5, addr))
			return -EINVAL;

		hash = cfi_rehash(hash);
		text_poke_early(addr + 1, &hash, 4);
	}

	return 0;
}

static void cfi_fineibt_bhi_preamble(void *addr, int arity)
{
	if (!arity)
		return;

	if (!cfi_warn && arity == 1) {
		/*
		 * Crazy scheme to allow arity-1 inline:
		 *
		 * __cfi_foo:
		 *  0: f3 0f 1e fa             endbr64
		 *  4: 41 81 <ea> 78 56 34 12  sub     0x12345678, %r10d
		 *  b: 49 0f 45 fa             cmovne  %r10, %rdi
		 *  f: 75 f5                   jne     __cfi_foo+6
		 * 11: 0f 1f 00                nopl    (%rax)
		 *
		 * Code that direct calls to foo()+0, decodes the tail end as:
		 *
		 * foo:
		 *  0: f5                      cmc
		 *  1: 0f 1f 00                nopl    (%rax)
		 *
		 * which clobbers CF, but does not affect anything ABI
		 * wise.
		 *
		 * Notably, this scheme is incompatible with permissive CFI
		 * because the CMOVcc is unconditional and RDI will have been
		 * clobbered.
		 */
		const u8 magic[9] = {
			0x49, 0x0f, 0x45, 0xfa,
			0x75, 0xf5,
			BYTES_NOP3,
		};

		text_poke_early(addr + fineibt_preamble_bhi, magic, 9);

		return;
	}

	text_poke_early(addr + fineibt_preamble_bhi,
			text_gen_insn(CALL_INSN_OPCODE,
				      addr + fineibt_preamble_bhi,
				      __bhi_args[arity]),
			CALL_INSN_SIZE);
}

static int cfi_rewrite_preamble(s32 *start, s32 *end)
{
	s32 *s;

	for (s = start; s < end; s++) {
		void *addr = (void *)s + *s;
		int arity;
		u32 hash;

		/*
		 * When the function doesn't start with ENDBR the compiler will
		 * have determined there are no indirect calls to it and we
		 * don't need no CFI either.
		 */
		if (!is_endbr(addr + 16))
			continue;

		hash = decode_preamble_hash(addr, &arity);
		if (WARN(!hash, "no CFI hash found at: %pS %px %*ph\n",
			 addr, addr, 5, addr))
			return -EINVAL;

		text_poke_early(addr, fineibt_preamble_start, fineibt_preamble_size);
		WARN_ON(*(u32 *)(addr + fineibt_preamble_hash) != 0x12345678);
		text_poke_early(addr + fineibt_preamble_hash, &hash, 4);

		WARN_ONCE(!IS_ENABLED(CONFIG_FINEIBT_BHI) && arity,
			  "kCFI preamble has wrong register at: %pS %*ph\n",
			  addr, 5, addr);

		if (cfi_bhi)
			cfi_fineibt_bhi_preamble(addr, arity);
	}

	return 0;
}

static void cfi_rewrite_endbr(s32 *start, s32 *end)
{
	s32 *s;

	for (s = start; s < end; s++) {
		void *addr = (void *)s + *s;

		if (!exact_endbr(addr + 16))
			continue;

		poison_endbr(addr + 16);
	}
}

/* .retpoline_sites */
static int cfi_rand_callers(s32 *start, s32 *end)
{
	s32 *s;

	for (s = start; s < end; s++) {
		void *addr = (void *)s + *s;
		u32 hash;

		addr -= fineibt_caller_size;
		hash = decode_caller_hash(addr);
		if (hash) {
			hash = -cfi_rehash(hash);
			text_poke_early(addr + 2, &hash, 4);
		}
	}

	return 0;
}

static int emit_paranoid_trampoline(void *addr, struct insn *insn, int reg, u8 *bytes)
{
	u8 *thunk = (void *)__x86_indirect_its_thunk_array[reg] - 2;

#ifdef CONFIG_MITIGATION_ITS
	u8 *tmp = its_allocate_thunk(reg);
	if (tmp)
		thunk = tmp;
#endif

	return __emit_trampoline(addr, insn, bytes, thunk, thunk);
}

static int cfi_rewrite_callers(s32 *start, s32 *end)
{
	s32 *s;

	BUG_ON(fineibt_paranoid_size != 20);

	for (s = start; s < end; s++) {
		void *addr = (void *)s + *s;
		struct insn insn;
		u8 bytes[20];
		u32 hash;
		int ret;
		u8 op;

		addr -= fineibt_caller_size;
		hash = decode_caller_hash(addr);
		if (!hash)
			continue;

		if (!cfi_paranoid) {
			text_poke_early(addr, fineibt_caller_start, fineibt_caller_size);
			WARN_ON(*(u32 *)(addr + fineibt_caller_hash) != 0x12345678);
			text_poke_early(addr + fineibt_caller_hash, &hash, 4);
			/* rely on apply_retpolines() */
			continue;
		}

		/* cfi_paranoid */
		ret = insn_decode_kernel(&insn, addr + fineibt_caller_size);
		if (WARN_ON_ONCE(ret < 0))
			continue;

		op = insn.opcode.bytes[0];
		if (op != CALL_INSN_OPCODE && op != JMP32_INSN_OPCODE) {
			WARN_ON_ONCE(1);
			continue;
		}

		memcpy(bytes, fineibt_paranoid_start, fineibt_paranoid_size);
		memcpy(bytes + fineibt_caller_hash, &hash, 4);

		if (cpu_wants_indirect_its_thunk_at((unsigned long)addr + fineibt_paranoid_ind, 11)) {
			emit_paranoid_trampoline(addr + fineibt_caller_size,
						 &insn, 11, bytes + fineibt_caller_size);
		} else {
			ret = emit_indirect(op, 11, bytes + fineibt_paranoid_ind);
			if (WARN_ON_ONCE(ret != 3))
				continue;
		}

		text_poke_early(addr, bytes, fineibt_paranoid_size);
	}

	return 0;
}

static void __apply_fineibt(s32 *start_retpoline, s32 *end_retpoline,
			    s32 *start_cfi, s32 *end_cfi, bool builtin)
{
	int ret;

	if (WARN_ONCE(fineibt_preamble_size != 16,
		      "FineIBT preamble wrong size: %ld", fineibt_preamble_size))
		return;

	if (cfi_mode == CFI_AUTO) {
		cfi_mode = CFI_KCFI;
		if (HAS_KERNEL_IBT && cpu_feature_enabled(X86_FEATURE_IBT)) {
			/*
			 * FRED has much saner context on exception entry and
			 * is less easy to take advantage of.
			 */
			if (!cpu_feature_enabled(X86_FEATURE_FRED))
				cfi_paranoid = true;
			cfi_mode = CFI_FINEIBT;
		}
	}

	/*
	 * Rewrite the callers to not use the __cfi_ stubs, such that we might
	 * rewrite them. This disables all CFI. If this succeeds but any of the
	 * later stages fails, we're without CFI.
	 */
	ret = cfi_disable_callers(start_retpoline, end_retpoline);
	if (ret)
		goto err;

	if (cfi_rand) {
		if (builtin) {
			cfi_seed = get_random_u32();
			cfi_bpf_hash = cfi_rehash(cfi_bpf_hash);
			cfi_bpf_subprog_hash = cfi_rehash(cfi_bpf_subprog_hash);
		}

		ret = cfi_rand_preamble(start_cfi, end_cfi);
		if (ret)
			goto err;

		ret = cfi_rand_callers(start_retpoline, end_retpoline);
		if (ret)
			goto err;
	}

	switch (cfi_mode) {
	case CFI_OFF:
		if (builtin)
			pr_info("Disabling CFI\n");
		return;

	case CFI_KCFI:
		ret = cfi_enable_callers(start_retpoline, end_retpoline);
		if (ret)
			goto err;

		if (builtin)
			pr_info("Using kCFI\n");
		return;

	case CFI_FINEIBT:
		/* place the FineIBT preamble at func()-16 */
		ret = cfi_rewrite_preamble(start_cfi, end_cfi);
		if (ret)
			goto err;

		/* rewrite the callers to target func()-16 */
		ret = cfi_rewrite_callers(start_retpoline, end_retpoline);
		if (ret)
			goto err;

		/* now that nobody targets func()+0, remove ENDBR there */
		cfi_rewrite_endbr(start_cfi, end_cfi);

		if (builtin) {
			pr_info("Using %sFineIBT%s CFI\n",
				cfi_paranoid ? "paranoid " : "",
				cfi_bhi ? "+BHI" : "");
		}
		return;

	default:
		break;
	}

err:
	pr_err("Something went horribly wrong trying to rewrite the CFI implementation.\n");
}

static inline void poison_hash(void *addr)
{
	*(u32 *)addr = 0;
}

static void poison_cfi(void *addr)
{
	/*
	 * Compilers manage to be inconsistent with ENDBR vs __cfi prefixes,
	 * some (static) functions for which they can determine the address
	 * is never taken do not get a __cfi prefix, but *DO* get an ENDBR.
	 *
	 * As such, these functions will get sealed, but we need to be careful
	 * to not unconditionally scribble the previous function.
	 */
	switch (cfi_mode) {
	case CFI_FINEIBT:
		/*
		 * FineIBT prefix should start with an ENDBR.
		 */
		if (!is_endbr(addr))
			break;

		/*
		 * __cfi_\func:
		 *	osp nopl (%rax)
		 *	subl	$0, %r10d
		 *	jz	1f
		 *	ud2
		 * 1:	nop
		 */
		poison_endbr(addr);
		poison_hash(addr + fineibt_preamble_hash);
		break;

	case CFI_KCFI:
		/*
		 * kCFI prefix should start with a valid hash.
		 */
		if (!decode_preamble_hash(addr, NULL))
			break;

		/*
		 * __cfi_\func:
		 *	movl	$0, %eax
		 *	.skip	11, 0x90
		 */
		poison_hash(addr + 1);
		break;

	default:
		break;
	}
}

/*
 * When regs->ip points to a 0xEA byte in the FineIBT preamble,
 * return true and fill out target and type.
 *
 * We check the preamble by checking for the ENDBR instruction relative to the
 * 0xEA instruction.
 */
static bool decode_fineibt_preamble(struct pt_regs *regs, unsigned long *target, u32 *type)
{
	unsigned long addr = regs->ip - fineibt_preamble_ud;
	u32 hash;

	if (!exact_endbr((void *)addr))
		return false;

	*target = addr + fineibt_preamble_size;

	__get_kernel_nofault(&hash, addr + fineibt_preamble_hash, u32, Efault);
	*type = (u32)regs->r10 + hash;

	/*
	 * Since regs->ip points to the middle of an instruction; it cannot
	 * continue with the normal fixup.
	 */
	regs->ip = *target;

	return true;

Efault:
	return false;
}

/*
 * regs->ip points to one of the UD2 in __bhi_args[].
 */
static bool decode_fineibt_bhi(struct pt_regs *regs, unsigned long *target, u32 *type)
{
	unsigned long addr;
	u32 hash;

	if (!cfi_bhi)
		return false;

	if (regs->ip < (unsigned long)__bhi_args ||
	    regs->ip >= (unsigned long)__bhi_args_end)
		return false;

	/*
	 * Fetch the return address from the stack, this points to the
	 * FineIBT preamble. Since the CALL instruction is in the 5 last
	 * bytes of the preamble, the return address is in fact the target
	 * address.
	 */
	__get_kernel_nofault(&addr, regs->sp, unsigned long, Efault);
	*target = addr;

	addr -= fineibt_preamble_size;
	if (!exact_endbr((void *)addr))
		return false;

	__get_kernel_nofault(&hash, addr + fineibt_preamble_hash, u32, Efault);
	*type = (u32)regs->r10 + hash;

	/*
	 * The UD2 sites are constructed with a RET immediately following,
	 * as such the non-fatal case can use the regular fixup.
	 */
	return true;

Efault:
	return false;
}

static bool is_paranoid_thunk(unsigned long addr)
{
	u32 thunk;

	__get_kernel_nofault(&thunk, (u32 *)addr, u32, Efault);
	return (thunk & 0x00FFFFFF) == 0xfd75ea;

Efault:
	return false;
}

/*
 * regs->ip points to a LOCK Jcc.d8 instruction from the fineibt_paranoid_start[]
 * sequence, or to an invalid instruction (0xea) + Jcc.d8 for cfi_paranoid + ITS
 * thunk.
 */
static bool decode_fineibt_paranoid(struct pt_regs *regs, unsigned long *target, u32 *type)
{
	unsigned long addr = regs->ip - fineibt_paranoid_ud;

	if (!cfi_paranoid)
		return false;

	if (is_cfi_trap(addr + fineibt_caller_size - LEN_UD2)) {
		*target = regs->r11 + fineibt_preamble_size;
		*type = regs->r10;

		/*
		 * Since the trapping instruction is the exact, but LOCK prefixed,
		 * Jcc.d8 that got us here, the normal fixup will work.
		 */
		return true;
	}

	/*
	 * The cfi_paranoid + ITS thunk combination results in:
	 *
	 *  0:   41 ba 78 56 34 12       mov    $0x12345678, %r10d
	 *  6:   45 3b 53 f7             cmp    -0x9(%r11), %r10d
	 *  a:   4d 8d 5b f0             lea    -0x10(%r11), %r11
	 *  e:   2e e8 XX XX XX XX	 cs call __x86_indirect_paranoid_thunk_r11
	 *
	 * Where the paranoid_thunk looks like:
	 *
	 *  1d:  <ea>                    (bad)
	 *  __x86_indirect_paranoid_thunk_r11:
	 *  1e:  75 fd                   jne 1d
	 *  __x86_indirect_its_thunk_r11:
	 *  20:  41 ff eb                jmp *%r11
	 *  23:  cc                      int3
	 *
	 */
	if (is_paranoid_thunk(regs->ip)) {
		*target = regs->r11 + fineibt_preamble_size;
		*type = regs->r10;

		regs->ip = *target;
		return true;
	}

	return false;
}

bool decode_fineibt_insn(struct pt_regs *regs, unsigned long *target, u32 *type)
{
	if (decode_fineibt_paranoid(regs, target, type))
		return true;

	if (decode_fineibt_bhi(regs, target, type))
		return true;

	return decode_fineibt_preamble(regs, target, type);
}

#else /* !CONFIG_FINEIBT: */

static void __apply_fineibt(s32 *start_retpoline, s32 *end_retpoline,
			    s32 *start_cfi, s32 *end_cfi, bool builtin)
{
}

#ifdef CONFIG_X86_KERNEL_IBT
static void poison_cfi(void *addr) { }
#endif

#endif /* !CONFIG_FINEIBT */

void apply_fineibt(s32 *start_retpoline, s32 *end_retpoline,
		   s32 *start_cfi, s32 *end_cfi)
{
	return __apply_fineibt(start_retpoline, end_retpoline,
			       start_cfi, end_cfi,
			       /* .builtin = */ false);
}

#ifdef CONFIG_SMP
static void alternatives_smp_lock(const s32 *start, const s32 *end,
				  u8 *text, u8 *text_end)
{
	const s32 *poff;

	for (poff = start; poff < end; poff++) {
		u8 *ptr = (u8 *)poff + *poff;

		if (!*poff || ptr < text || ptr >= text_end)
			continue;
		/* turn DS segment override prefix into lock prefix */
		if (*ptr == 0x3e)
			text_poke(ptr, ((unsigned char []){0xf0}), 1);
	}
}

static void alternatives_smp_unlock(const s32 *start, const s32 *end,
				    u8 *text, u8 *text_end)
{
	const s32 *poff;

	for (poff = start; poff < end; poff++) {
		u8 *ptr = (u8 *)poff + *poff;

		if (!*poff || ptr < text || ptr >= text_end)
			continue;
		/* turn lock prefix into DS segment override prefix */
		if (*ptr == 0xf0)
			text_poke(ptr, ((unsigned char []){0x3E}), 1);
	}
}

struct smp_alt_module {
	/* what is this ??? */
	struct module	*mod;
	char		*name;

	/* ptrs to lock prefixes */
	const s32	*locks;
	const s32	*locks_end;

	/* .text segment, needed to avoid patching init code ;) */
	u8		*text;
	u8		*text_end;

	struct list_head next;
};
static LIST_HEAD(smp_alt_modules);
static bool uniproc_patched = false;	/* protected by text_mutex */

void __init_or_module alternatives_smp_module_add(struct module *mod,
						  char *name,
						  void *locks, void *locks_end,
						  void *text,  void *text_end)
{
	struct smp_alt_module *smp;

	mutex_lock(&text_mutex);
	if (!uniproc_patched)
		goto unlock;

	if (num_possible_cpus() == 1)
		/* Don't bother remembering, we'll never have to undo it. */
		goto smp_unlock;

	smp = kzalloc(sizeof(*smp), GFP_KERNEL);
	if (NULL == smp)
		/* we'll run the (safe but slow) SMP code then ... */
		goto unlock;

	smp->mod	= mod;
	smp->name	= name;
	smp->locks	= locks;
	smp->locks_end	= locks_end;
	smp->text	= text;
	smp->text_end	= text_end;
	DPRINTK(SMP, "locks %p -> %p, text %p -> %p, name %s\n",
		smp->locks, smp->locks_end,
		smp->text, smp->text_end, smp->name);

	list_add_tail(&smp->next, &smp_alt_modules);
smp_unlock:
	alternatives_smp_unlock(locks, locks_end, text, text_end);
unlock:
	mutex_unlock(&text_mutex);
}

void __init_or_module alternatives_smp_module_del(struct module *mod)
{
	struct smp_alt_module *item;

	mutex_lock(&text_mutex);
	list_for_each_entry(item, &smp_alt_modules, next) {
		if (mod != item->mod)
			continue;
		list_del(&item->next);
		kfree(item);
		break;
	}
	mutex_unlock(&text_mutex);
}

void alternatives_enable_smp(void)
{
	struct smp_alt_module *mod;

	/* Why bother if there are no other CPUs? */
	BUG_ON(num_possible_cpus() == 1);

	mutex_lock(&text_mutex);

	if (uniproc_patched) {
		pr_info("switching to SMP code\n");
		BUG_ON(num_online_cpus() != 1);
		clear_cpu_cap(&boot_cpu_data, X86_FEATURE_UP);
		clear_cpu_cap(&cpu_data(0), X86_FEATURE_UP);
		list_for_each_entry(mod, &smp_alt_modules, next)
			alternatives_smp_lock(mod->locks, mod->locks_end,
					      mod->text, mod->text_end);
		uniproc_patched = false;
	}
	mutex_unlock(&text_mutex);
}

/*
 * Return 1 if the address range is reserved for SMP-alternatives.
 * Must hold text_mutex.
 */
int alternatives_text_reserved(void *start, void *end)
{
	struct smp_alt_module *mod;
	const s32 *poff;
	u8 *text_start = start;
	u8 *text_end = end;

	lockdep_assert_held(&text_mutex);

	list_for_each_entry(mod, &smp_alt_modules, next) {
		if (mod->text > text_end || mod->text_end < text_start)
			continue;
		for (poff = mod->locks; poff < mod->locks_end; poff++) {
			const u8 *ptr = (const u8 *)poff + *poff;

			if (text_start <= ptr && text_end > ptr)
				return 1;
		}
	}

	return 0;
}
#endif /* CONFIG_SMP */

/*
 * Self-test for the INT3 based CALL emulation code.
 *
 * This exercises int3_emulate_call() to make sure INT3 pt_regs are set up
 * properly and that there is a stack gap between the INT3 frame and the
 * previous context. Without this gap doing a virtual PUSH on the interrupted
 * stack would corrupt the INT3 IRET frame.
 *
 * See entry_{32,64}.S for more details.
 */

/*
 * We define the int3_magic() function in assembly to control the calling
 * convention such that we can 'call' it from assembly.
 */

extern void int3_magic(unsigned int *ptr); /* defined in asm */

asm (
"	.pushsection	.init.text, \"ax\", @progbits\n"
"	.type		int3_magic, @function\n"
"int3_magic:\n"
	ANNOTATE_NOENDBR
"	movl	$1, (%" _ASM_ARG1 ")\n"
	ASM_RET
"	.size		int3_magic, .-int3_magic\n"
"	.popsection\n"
);

extern void int3_selftest_ip(void); /* defined in asm below */

static int __init
int3_exception_notify(struct notifier_block *self, unsigned long val, void *data)
{
	unsigned long selftest = (unsigned long)&int3_selftest_ip;
	struct die_args *args = data;
	struct pt_regs *regs = args->regs;

	OPTIMIZER_HIDE_VAR(selftest);

	if (!regs || user_mode(regs))
		return NOTIFY_DONE;

	if (val != DIE_INT3)
		return NOTIFY_DONE;

	if (regs->ip - INT3_INSN_SIZE != selftest)
		return NOTIFY_DONE;

	int3_emulate_call(regs, (unsigned long)&int3_magic);
	return NOTIFY_STOP;
}

/* Must be noinline to ensure uniqueness of int3_selftest_ip. */
static noinline void __init int3_selftest(void)
{
	static __initdata struct notifier_block int3_exception_nb = {
		.notifier_call	= int3_exception_notify,
		.priority	= INT_MAX-1, /* last */
	};
	unsigned int val = 0;

	BUG_ON(register_die_notifier(&int3_exception_nb));

	/*
	 * Basically: int3_magic(&val); but really complicated :-)
	 *
	 * INT3 padded with NOP to CALL_INSN_SIZE. The int3_exception_nb
	 * notifier above will emulate CALL for us.
	 */
	asm volatile ("int3_selftest_ip:\n\t"
		      ANNOTATE_NOENDBR
		      "    int3; nop; nop; nop; nop\n\t"
		      : ASM_CALL_CONSTRAINT
		      : __ASM_SEL_RAW(a, D) (&val)
		      : "memory");

	BUG_ON(val != 1);

	unregister_die_notifier(&int3_exception_nb);
}

static __initdata int __alt_reloc_selftest_addr;

extern void __init __alt_reloc_selftest(void *arg);
__visible noinline void __init __alt_reloc_selftest(void *arg)
{
	WARN_ON(arg != &__alt_reloc_selftest_addr);
}

static noinline void __init alt_reloc_selftest(void)
{
	/*
	 * Tests text_poke_apply_relocation().
	 *
	 * This has a relative immediate (CALL) in a place other than the first
	 * instruction and additionally on x86_64 we get a RIP-relative LEA:
	 *
	 *   lea    0x0(%rip),%rdi  # 5d0: R_X86_64_PC32    .init.data+0x5566c
	 *   call   +0              # 5d5: R_X86_64_PLT32   __alt_reloc_selftest-0x4
	 *
	 * Getting this wrong will either crash and burn or tickle the WARN
	 * above.
	 */
	asm_inline volatile (
		ALTERNATIVE("", "lea %[mem], %%" _ASM_ARG1 "; call __alt_reloc_selftest;", X86_FEATURE_ALWAYS)
		: ASM_CALL_CONSTRAINT
		: [mem] "m" (__alt_reloc_selftest_addr)
		: _ASM_ARG1
	);
}

void __init alternative_instructions(void)
{
	u64 ibt;

	int3_selftest();

	/*
	 * The patching is not fully atomic, so try to avoid local
	 * interruptions that might execute the to be patched code.
	 * Other CPUs are not running.
	 */
	stop_nmi();

	/*
	 * Don't stop machine check exceptions while patching.
	 * MCEs only happen when something got corrupted and in this
	 * case we must do something about the corruption.
	 * Ignoring it is worse than an unlikely patching race.
	 * Also machine checks tend to be broadcast and if one CPU
	 * goes into machine check the others follow quickly, so we don't
	 * expect a machine check to cause undue problems during to code
	 * patching.
	 */

	/*
	 * Make sure to set (artificial) features depending on used paravirt
	 * functions which can later influence alternative patching.
	 */
	paravirt_set_cap();

	/* Keep CET-IBT disabled until caller/callee are patched */
	ibt = ibt_save(/*disable*/ true);

	__apply_fineibt(__retpoline_sites, __retpoline_sites_end,
			__cfi_sites, __cfi_sites_end, true);

	/*
	 * Rewrite the retpolines, must be done before alternatives since
	 * those can rewrite the retpoline thunks.
	 */
	apply_retpolines(__retpoline_sites, __retpoline_sites_end);
	apply_returns(__return_sites, __return_sites_end);

	its_fini_core();

	/*
	 * Adjust all CALL instructions to point to func()-10, including
	 * those in .altinstr_replacement.
	 */
	callthunks_patch_builtin_calls();

	apply_alternatives(__alt_instructions, __alt_instructions_end);

	/*
	 * Seal all functions that do not have their address taken.
	 */
	apply_seal_endbr(__ibt_endbr_seal, __ibt_endbr_seal_end);

	ibt_restore(ibt);

#ifdef CONFIG_SMP
	/* Patch to UP if other cpus not imminent. */
	if (!noreplace_smp && (num_present_cpus() == 1 || setup_max_cpus <= 1)) {
		uniproc_patched = true;
		alternatives_smp_module_add(NULL, "core kernel",
					    __smp_locks, __smp_locks_end,
					    _text, _etext);
	}

	if (!uniproc_patched || num_possible_cpus() == 1) {
		free_init_pages("SMP alternatives",
				(unsigned long)__smp_locks,
				(unsigned long)__smp_locks_end);
	}
#endif

	restart_nmi();
	alternatives_patched = 1;

	alt_reloc_selftest();
}

/**
 * text_poke_early - Update instructions on a live kernel at boot time
 * @addr: address to modify
 * @opcode: source of the copy
 * @len: length to copy
 *
 * When you use this code to patch more than one byte of an instruction
 * you need to make sure that other CPUs cannot execute this code in parallel.
 * Also no thread must be currently preempted in the middle of these
 * instructions. And on the local CPU you need to be protected against NMI or
 * MCE handlers seeing an inconsistent instruction while you patch.
 */
void __init_or_module text_poke_early(void *addr, const void *opcode,
				      size_t len)
{
	unsigned long flags;

	if (boot_cpu_has(X86_FEATURE_NX) &&
	    is_module_text_address((unsigned long)addr)) {
		/*
		 * Modules text is marked initially as non-executable, so the
		 * code cannot be running and speculative code-fetches are
		 * prevented. Just change the code.
		 */
		memcpy(addr, opcode, len);
	} else {
		local_irq_save(flags);
		memcpy(addr, opcode, len);
		sync_core();
		local_irq_restore(flags);

		/*
		 * Could also do a CLFLUSH here to speed up CPU recovery; but
		 * that causes hangs on some VIA CPUs.
		 */
	}
}

__ro_after_init struct mm_struct *text_poke_mm;
__ro_after_init unsigned long text_poke_mm_addr;

static void text_poke_memcpy(void *dst, const void *src, size_t len)
{
	memcpy(dst, src, len);
}

static void text_poke_memset(void *dst, const void *src, size_t len)
{
	int c = *(const int *)src;

	memset(dst, c, len);
}

typedef void text_poke_f(void *dst, const void *src, size_t len);

static void *__text_poke(text_poke_f func, void *addr, const void *src, size_t len)
{
	bool cross_page_boundary = offset_in_page(addr) + len > PAGE_SIZE;
	struct page *pages[2] = {NULL};
	struct mm_struct *prev_mm;
	unsigned long flags;
	pte_t pte, *ptep;
	spinlock_t *ptl;
	pgprot_t pgprot;

	/*
	 * While boot memory allocator is running we cannot use struct pages as
	 * they are not yet initialized. There is no way to recover.
	 */
	BUG_ON(!after_bootmem);

	if (!core_kernel_text((unsigned long)addr)) {
		pages[0] = vmalloc_to_page(addr);
		if (cross_page_boundary)
			pages[1] = vmalloc_to_page(addr + PAGE_SIZE);
	} else {
		pages[0] = virt_to_page(addr);
		WARN_ON(!PageReserved(pages[0]));
		if (cross_page_boundary)
			pages[1] = virt_to_page(addr + PAGE_SIZE);
	}
	/*
	 * If something went wrong, crash and burn since recovery paths are not
	 * implemented.
	 */
	BUG_ON(!pages[0] || (cross_page_boundary && !pages[1]));

	/*
	 * Map the page without the global bit, as TLB flushing is done with
	 * flush_tlb_mm_range(), which is intended for non-global PTEs.
	 */
	pgprot = __pgprot(pgprot_val(PAGE_KERNEL) & ~_PAGE_GLOBAL);

	/*
	 * The lock is not really needed, but this allows to avoid open-coding.
	 */
	ptep = get_locked_pte(text_poke_mm, text_poke_mm_addr, &ptl);

	/*
	 * This must not fail; preallocated in poking_init().
	 */
	VM_BUG_ON(!ptep);

	local_irq_save(flags);

	pte = mk_pte(pages[0], pgprot);
	set_pte_at(text_poke_mm, text_poke_mm_addr, ptep, pte);

	if (cross_page_boundary) {
		pte = mk_pte(pages[1], pgprot);
		set_pte_at(text_poke_mm, text_poke_mm_addr + PAGE_SIZE, ptep + 1, pte);
	}

	/*
	 * Loading the temporary mm behaves as a compiler barrier, which
	 * guarantees that the PTE will be set at the time memcpy() is done.
	 */
	prev_mm = use_temporary_mm(text_poke_mm);

	kasan_disable_current();
	func((u8 *)text_poke_mm_addr + offset_in_page(addr), src, len);
	kasan_enable_current();

	/*
	 * Ensure that the PTE is only cleared after the instructions of memcpy
	 * were issued by using a compiler barrier.
	 */
	barrier();

	pte_clear(text_poke_mm, text_poke_mm_addr, ptep);
	if (cross_page_boundary)
		pte_clear(text_poke_mm, text_poke_mm_addr + PAGE_SIZE, ptep + 1);

	/*
	 * Loading the previous page-table hierarchy requires a serializing
	 * instruction that already allows the core to see the updated version.
	 * Xen-PV is assumed to serialize execution in a similar manner.
	 */
	unuse_temporary_mm(prev_mm);

	/*
	 * Flushing the TLB might involve IPIs, which would require enabled
	 * IRQs, but not if the mm is not used, as it is in this point.
	 */
	flush_tlb_mm_range(text_poke_mm, text_poke_mm_addr, text_poke_mm_addr +
			   (cross_page_boundary ? 2 : 1) * PAGE_SIZE,
			   PAGE_SHIFT, false);

	if (func == text_poke_memcpy) {
		/*
		 * If the text does not match what we just wrote then something is
		 * fundamentally screwy; there's nothing we can really do about that.
		 */
		BUG_ON(memcmp(addr, src, len));
	}

	local_irq_restore(flags);
	pte_unmap_unlock(ptep, ptl);
	return addr;
}

/**
 * text_poke - Update instructions on a live kernel
 * @addr: address to modify
 * @opcode: source of the copy
 * @len: length to copy
 *
 * Only atomic text poke/set should be allowed when not doing early patching.
 * It means the size must be writable atomically and the address must be aligned
 * in a way that permits an atomic write. It also makes sure we fit on a single
 * page.
 *
 * Note that the caller must ensure that if the modified code is part of a
 * module, the module would not be removed during poking. This can be achieved
 * by registering a module notifier, and ordering module removal and patching
 * through a mutex.
 */
void *text_poke(void *addr, const void *opcode, size_t len)
{
	lockdep_assert_held(&text_mutex);

	return __text_poke(text_poke_memcpy, addr, opcode, len);
}

/**
 * text_poke_kgdb - Update instructions on a live kernel by kgdb
 * @addr: address to modify
 * @opcode: source of the copy
 * @len: length to copy
 *
 * Only atomic text poke/set should be allowed when not doing early patching.
 * It means the size must be writable atomically and the address must be aligned
 * in a way that permits an atomic write. It also makes sure we fit on a single
 * page.
 *
 * Context: should only be used by kgdb, which ensures no other core is running,
 *	    despite the fact it does not hold the text_mutex.
 */
void *text_poke_kgdb(void *addr, const void *opcode, size_t len)
{
	return __text_poke(text_poke_memcpy, addr, opcode, len);
}

void *text_poke_copy_locked(void *addr, const void *opcode, size_t len,
			    bool core_ok)
{
	unsigned long start = (unsigned long)addr;
	size_t patched = 0;

	if (WARN_ON_ONCE(!core_ok && core_kernel_text(start)))
		return NULL;

	while (patched < len) {
		unsigned long ptr = start + patched;
		size_t s;

		s = min_t(size_t, PAGE_SIZE * 2 - offset_in_page(ptr), len - patched);

		__text_poke(text_poke_memcpy, (void *)ptr, opcode + patched, s);
		patched += s;
	}
	return addr;
}

/**
 * text_poke_copy - Copy instructions into (an unused part of) RX memory
 * @addr: address to modify
 * @opcode: source of the copy
 * @len: length to copy, could be more than 2x PAGE_SIZE
 *
 * Not safe against concurrent execution; useful for JITs to dump
 * new code blocks into unused regions of RX memory. Can be used in
 * conjunction with synchronize_rcu_tasks() to wait for existing
 * execution to quiesce after having made sure no existing functions
 * pointers are live.
 */
void *text_poke_copy(void *addr, const void *opcode, size_t len)
{
	mutex_lock(&text_mutex);
	addr = text_poke_copy_locked(addr, opcode, len, false);
	mutex_unlock(&text_mutex);
	return addr;
}

/**
 * text_poke_set - memset into (an unused part of) RX memory
 * @addr: address to modify
 * @c: the byte to fill the area with
 * @len: length to copy, could be more than 2x PAGE_SIZE
 *
 * This is useful to overwrite unused regions of RX memory with illegal
 * instructions.
 */
void *text_poke_set(void *addr, int c, size_t len)
{
	unsigned long start = (unsigned long)addr;
	size_t patched = 0;

	if (WARN_ON_ONCE(core_kernel_text(start)))
		return NULL;

	mutex_lock(&text_mutex);
	while (patched < len) {
		unsigned long ptr = start + patched;
		size_t s;

		s = min_t(size_t, PAGE_SIZE * 2 - offset_in_page(ptr), len - patched);

		__text_poke(text_poke_memset, (void *)ptr, (void *)&c, s);
		patched += s;
	}
	mutex_unlock(&text_mutex);
	return addr;
}

static void do_sync_core(void *info)
{
	sync_core();
}

void smp_text_poke_sync_each_cpu(void)
{
	on_each_cpu(do_sync_core, NULL, 1);
}

/*
 * NOTE: crazy scheme to allow patching Jcc.d32 but not increase the size of
 * this thing. When len == 6 everything is prefixed with 0x0f and we map
 * opcode to Jcc.d8, using len to distinguish.
 */
struct smp_text_poke_loc {
	/* addr := _stext + rel_addr */
	s32 rel_addr;
	s32 disp;
	u8 len;
	u8 opcode;
	const u8 text[TEXT_POKE_MAX_OPCODE_SIZE];
	/* see smp_text_poke_batch_finish() */
	u8 old;
};

#define TEXT_POKE_ARRAY_MAX (PAGE_SIZE / sizeof(struct smp_text_poke_loc))

static struct smp_text_poke_array {
	struct smp_text_poke_loc vec[TEXT_POKE_ARRAY_MAX];
	int nr_entries;
} text_poke_array;

static DEFINE_PER_CPU(atomic_t, text_poke_array_refs);

/*
 * These four __always_inline annotations imply noinstr, necessary
 * due to smp_text_poke_int3_handler() being noinstr:
 */

static __always_inline bool try_get_text_poke_array(void)
{
	atomic_t *refs = this_cpu_ptr(&text_poke_array_refs);

	if (!raw_atomic_inc_not_zero(refs))
		return false;

	return true;
}

static __always_inline void put_text_poke_array(void)
{
	atomic_t *refs = this_cpu_ptr(&text_poke_array_refs);

	smp_mb__before_atomic();
	raw_atomic_dec(refs);
}

static __always_inline void *text_poke_addr(const struct smp_text_poke_loc *tpl)
{
	return _stext + tpl->rel_addr;
}

static __always_inline int patch_cmp(const void *tpl_a, const void *tpl_b)
{
	if (tpl_a < text_poke_addr(tpl_b))
		return -1;
	if (tpl_a > text_poke_addr(tpl_b))
		return 1;
	return 0;
}

noinstr int smp_text_poke_int3_handler(struct pt_regs *regs)
{
	struct smp_text_poke_loc *tpl;
	int ret = 0;
	void *ip;

	if (user_mode(regs))
		return 0;

	/*
	 * Having observed our INT3 instruction, we now must observe
	 * text_poke_array with non-zero refcount:
	 *
	 *	text_poke_array_refs = 1		INT3
	 *	WMB			RMB
	 *	write INT3		if (text_poke_array_refs != 0)
	 */
	smp_rmb();

	if (!try_get_text_poke_array())
		return 0;

	/*
	 * Discount the INT3. See smp_text_poke_batch_finish().
	 */
	ip = (void *) regs->ip - INT3_INSN_SIZE;

	/*
	 * Skip the binary search if there is a single member in the vector.
	 */
	if (unlikely(text_poke_array.nr_entries > 1)) {
		tpl = __inline_bsearch(ip, text_poke_array.vec, text_poke_array.nr_entries,
				      sizeof(struct smp_text_poke_loc),
				      patch_cmp);
		if (!tpl)
			goto out_put;
	} else {
		tpl = text_poke_array.vec;
		if (text_poke_addr(tpl) != ip)
			goto out_put;
	}

	ip += tpl->len;

	switch (tpl->opcode) {
	case INT3_INSN_OPCODE:
		/*
		 * Someone poked an explicit INT3, they'll want to handle it,
		 * do not consume.
		 */
		goto out_put;

	case RET_INSN_OPCODE:
		int3_emulate_ret(regs);
		break;

	case CALL_INSN_OPCODE:
		int3_emulate_call(regs, (long)ip + tpl->disp);
		break;

	case JMP32_INSN_OPCODE:
	case JMP8_INSN_OPCODE:
		int3_emulate_jmp(regs, (long)ip + tpl->disp);
		break;

	case 0x70 ... 0x7f: /* Jcc */
		int3_emulate_jcc(regs, tpl->opcode & 0xf, (long)ip, tpl->disp);
		break;

	default:
		BUG();
	}

	ret = 1;

out_put:
	put_text_poke_array();
	return ret;
}

/**
 * smp_text_poke_batch_finish() -- update instructions on live kernel on SMP
 *
 * Input state:
 *  text_poke_array.vec: vector of instructions to patch
 *  text_poke_array.nr_entries: number of entries in the vector
 *
 * Modify multi-byte instructions by using INT3 breakpoints on SMP.
 * We completely avoid using stop_machine() here, and achieve the
 * synchronization using INT3 breakpoints and SMP cross-calls.
 *
 * The way it is done:
 *	- For each entry in the vector:
 *		- add an INT3 trap to the address that will be patched
 *	- SMP sync all CPUs
 *	- For each entry in the vector:
 *		- update all but the first byte of the patched range
 *	- SMP sync all CPUs
 *	- For each entry in the vector:
 *		- replace the first byte (INT3) by the first byte of the
 *		  replacing opcode
 *	- SMP sync all CPUs
 */
void smp_text_poke_batch_finish(void)
{
	unsigned char int3 = INT3_INSN_OPCODE;
	unsigned int i;
	int do_sync;

	if (!text_poke_array.nr_entries)
		return;

	lockdep_assert_held(&text_mutex);

	/*
	 * Corresponds to the implicit memory barrier in try_get_text_poke_array() to
	 * ensure reading a non-zero refcount provides up to date text_poke_array data.
	 */
	for_each_possible_cpu(i)
		atomic_set_release(per_cpu_ptr(&text_poke_array_refs, i), 1);

	/*
	 * Function tracing can enable thousands of places that need to be
	 * updated. This can take quite some time, and with full kernel debugging
	 * enabled, this could cause the softlockup watchdog to trigger.
	 * This function gets called every 256 entries added to be patched.
	 * Call cond_resched() here to make sure that other tasks can get scheduled
	 * while processing all the functions being patched.
	 */
	cond_resched();

	/*
	 * Corresponding read barrier in INT3 notifier for making sure the
	 * text_poke_array.nr_entries and handler are correctly ordered wrt. patching.
	 */
	smp_wmb();

	/*
	 * First step: add a INT3 trap to the address that will be patched.
	 */
	for (i = 0; i < text_poke_array.nr_entries; i++) {
		text_poke_array.vec[i].old = *(u8 *)text_poke_addr(&text_poke_array.vec[i]);
		text_poke(text_poke_addr(&text_poke_array.vec[i]), &int3, INT3_INSN_SIZE);
	}

	smp_text_poke_sync_each_cpu();

	/*
	 * Second step: update all but the first byte of the patched range.
	 */
	for (do_sync = 0, i = 0; i < text_poke_array.nr_entries; i++) {
		u8 old[TEXT_POKE_MAX_OPCODE_SIZE+1] = { text_poke_array.vec[i].old, };
		u8 _new[TEXT_POKE_MAX_OPCODE_SIZE+1];
		const u8 *new = text_poke_array.vec[i].text;
		int len = text_poke_array.vec[i].len;

		if (len - INT3_INSN_SIZE > 0) {
			memcpy(old + INT3_INSN_SIZE,
			       text_poke_addr(&text_poke_array.vec[i]) + INT3_INSN_SIZE,
			       len - INT3_INSN_SIZE);

			if (len == 6) {
				_new[0] = 0x0f;
				memcpy(_new + 1, new, 5);
				new = _new;
			}

			text_poke(text_poke_addr(&text_poke_array.vec[i]) + INT3_INSN_SIZE,
				  new + INT3_INSN_SIZE,
				  len - INT3_INSN_SIZE);

			do_sync++;
		}

		/*
		 * Emit a perf event to record the text poke, primarily to
		 * support Intel PT decoding which must walk the executable code
		 * to reconstruct the trace. The flow up to here is:
		 *   - write INT3 byte
		 *   - IPI-SYNC
		 *   - write instruction tail
		 * At this point the actual control flow will be through the
		 * INT3 and handler and not hit the old or new instruction.
		 * Intel PT outputs FUP/TIP packets for the INT3, so the flow
		 * can still be decoded. Subsequently:
		 *   - emit RECORD_TEXT_POKE with the new instruction
		 *   - IPI-SYNC
		 *   - write first byte
		 *   - IPI-SYNC
		 * So before the text poke event timestamp, the decoder will see
		 * either the old instruction flow or FUP/TIP of INT3. After the
		 * text poke event timestamp, the decoder will see either the
		 * new instruction flow or FUP/TIP of INT3. Thus decoders can
		 * use the timestamp as the point at which to modify the
		 * executable code.
		 * The old instruction is recorded so that the event can be
		 * processed forwards or backwards.
		 */
		perf_event_text_poke(text_poke_addr(&text_poke_array.vec[i]), old, len, new, len);
	}

	if (do_sync) {
		/*
		 * According to Intel, this core syncing is very likely
		 * not necessary and we'd be safe even without it. But
		 * better safe than sorry (plus there's not only Intel).
		 */
		smp_text_poke_sync_each_cpu();
	}

	/*
	 * Third step: replace the first byte (INT3) by the first byte of the
	 * replacing opcode.
	 */
	for (do_sync = 0, i = 0; i < text_poke_array.nr_entries; i++) {
		u8 byte = text_poke_array.vec[i].text[0];

		if (text_poke_array.vec[i].len == 6)
			byte = 0x0f;

		if (byte == INT3_INSN_OPCODE)
			continue;

		text_poke(text_poke_addr(&text_poke_array.vec[i]), &byte, INT3_INSN_SIZE);
		do_sync++;
	}

	if (do_sync)
		smp_text_poke_sync_each_cpu();

	/*
	 * Remove and wait for refs to be zero.
	 *
	 * Notably, if after step-3 above the INT3 got removed, then the
	 * smp_text_poke_sync_each_cpu() will have serialized against any running INT3
	 * handlers and the below spin-wait will not happen.
	 *
	 * IOW. unless the replacement instruction is INT3, this case goes
	 * unused.
	 */
	for_each_possible_cpu(i) {
		atomic_t *refs = per_cpu_ptr(&text_poke_array_refs, i);

		if (unlikely(!atomic_dec_and_test(refs)))
			atomic_cond_read_acquire(refs, !VAL);
	}

	/* They are all completed: */
	text_poke_array.nr_entries = 0;
}

static void __smp_text_poke_batch_add(void *addr, const void *opcode, size_t len, const void *emulate)
{
	struct smp_text_poke_loc *tpl;
	struct insn insn;
	int ret, i = 0;

	tpl = &text_poke_array.vec[text_poke_array.nr_entries++];

	if (len == 6)
		i = 1;
	memcpy((void *)tpl->text, opcode+i, len-i);
	if (!emulate)
		emulate = opcode;

	ret = insn_decode_kernel(&insn, emulate);
	BUG_ON(ret < 0);

	tpl->rel_addr = addr - (void *)_stext;
	tpl->len = len;
	tpl->opcode = insn.opcode.bytes[0];

	if (is_jcc32(&insn)) {
		/*
		 * Map Jcc.d32 onto Jcc.d8 and use len to distinguish.
		 */
		tpl->opcode = insn.opcode.bytes[1] - 0x10;
	}

	switch (tpl->opcode) {
	case RET_INSN_OPCODE:
	case JMP32_INSN_OPCODE:
	case JMP8_INSN_OPCODE:
		/*
		 * Control flow instructions without implied execution of the
		 * next instruction can be padded with INT3.
		 */
		for (i = insn.length; i < len; i++)
			BUG_ON(tpl->text[i] != INT3_INSN_OPCODE);
		break;

	default:
		BUG_ON(len != insn.length);
	}

	switch (tpl->opcode) {
	case INT3_INSN_OPCODE:
	case RET_INSN_OPCODE:
		break;

	case CALL_INSN_OPCODE:
	case JMP32_INSN_OPCODE:
	case JMP8_INSN_OPCODE:
	case 0x70 ... 0x7f: /* Jcc */
		tpl->disp = insn.immediate.value;
		break;

	default: /* assume NOP */
		switch (len) {
		case 2: /* NOP2 -- emulate as JMP8+0 */
			BUG_ON(memcmp(emulate, x86_nops[len], len));
			tpl->opcode = JMP8_INSN_OPCODE;
			tpl->disp = 0;
			break;

		case 5: /* NOP5 -- emulate as JMP32+0 */
			BUG_ON(memcmp(emulate, x86_nops[len], len));
			tpl->opcode = JMP32_INSN_OPCODE;
			tpl->disp = 0;
			break;

		default: /* unknown instruction */
			BUG();
		}
		break;
	}
}

/*
 * We hard rely on the text_poke_array.vec being ordered; ensure this is so by flushing
 * early if needed.
 */
static bool text_poke_addr_ordered(void *addr)
{
	WARN_ON_ONCE(!addr);

	if (!text_poke_array.nr_entries)
		return true;

	/*
	 * If the last current entry's address is higher than the
	 * new entry's address we'd like to add, then ordering
	 * is violated and we must first flush all pending patching
	 * requests:
	 */
	if (text_poke_addr(text_poke_array.vec + text_poke_array.nr_entries-1) > addr)
		return false;

	return true;
}

/**
 * smp_text_poke_batch_add() -- update instruction on live kernel on SMP, batched
 * @addr:	address to patch
 * @opcode:	opcode of new instruction
 * @len:	length to copy
 * @emulate:	instruction to be emulated
 *
 * Add a new instruction to the current queue of to-be-patched instructions
 * the kernel maintains. The patching request will not be executed immediately,
 * but becomes part of an array of patching requests, optimized for batched
 * execution. All pending patching requests will be executed on the next
 * smp_text_poke_batch_finish() call.
 */
void __ref smp_text_poke_batch_add(void *addr, const void *opcode, size_t len, const void *emulate)
{
	if (text_poke_array.nr_entries == TEXT_POKE_ARRAY_MAX || !text_poke_addr_ordered(addr))
		smp_text_poke_batch_finish();
	__smp_text_poke_batch_add(addr, opcode, len, emulate);
}

/**
 * smp_text_poke_single() -- update instruction on live kernel on SMP immediately
 * @addr:	address to patch
 * @opcode:	opcode of new instruction
 * @len:	length to copy
 * @emulate:	instruction to be emulated
 *
 * Update a single instruction with the vector in the stack, avoiding
 * dynamically allocated memory. This function should be used when it is
 * not possible to allocate memory for a vector. The single instruction
 * is patched in immediately.
 */
void __ref smp_text_poke_single(void *addr, const void *opcode, size_t len, const void *emulate)
{
	smp_text_poke_batch_add(addr, opcode, len, emulate);
	smp_text_poke_batch_finish();
}