tpy-lang 0.3.0.dev0__py3-none-any.whl

tpyc/_data/runtime/cpp/third_party/pcre2/src/sljit/sljitNativeARM_32.c

@@ -0,0 +1,4636 @@
+/*
+ *    Stack-less Just-In-Time compiler
+ *
+ *    Copyright Zoltan Herczeg (hzmester@freemail.hu). All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without modification, are
+ * permitted provided that the following conditions are met:
+ *
+ *   1. Redistributions of source code must retain the above copyright notice, this list of
+ *      conditions and the following disclaimer.
+ *
+ *   2. Redistributions in binary form must reproduce the above copyright notice, this list
+ *      of conditions and the following disclaimer in the documentation and/or other materials
+ *      provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDER(S) AND CONTRIBUTORS ``AS IS'' AND ANY
+ * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT
+ * SHALL THE COPYRIGHT HOLDER(S) OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED
+ * TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
+ * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
+ * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#ifdef __SOFTFP__
+#define ARM_ABI_INFO " ABI:softfp"
+#else
+#define ARM_ABI_INFO " ABI:hardfp"
+#endif
+
+SLJIT_API_FUNC_ATTRIBUTE const char* sljit_get_platform_name(void)
+{
+#if (defined SLJIT_CONFIG_ARM_V7 && SLJIT_CONFIG_ARM_V7)
+	return "ARMv7" SLJIT_CPUINFO ARM_ABI_INFO;
+#elif (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+	return "ARMv6" SLJIT_CPUINFO ARM_ABI_INFO;
+#else
+#error "Internal error: Unknown ARM architecture"
+#endif
+}
+
+/* Length of an instruction word. */
+typedef sljit_u32 sljit_ins;
+
+/* Last register + 1. */
+#define TMP_REG1	(SLJIT_NUMBER_OF_REGISTERS + 2)
+#define TMP_REG2	(SLJIT_NUMBER_OF_REGISTERS + 3)
+#define TMP_PC		(SLJIT_NUMBER_OF_REGISTERS + 4)
+
+#define TMP_FREG1	(SLJIT_NUMBER_OF_FLOAT_REGISTERS + 1)
+#define TMP_FREG2	(SLJIT_NUMBER_OF_FLOAT_REGISTERS + 2)
+
+/* In ARM instruction words.
+   Cache lines are usually 32 byte aligned. */
+#define CONST_POOL_ALIGNMENT	8
+#define CONST_POOL_EMPTY	0xffffffff
+
+#define ALIGN_INSTRUCTION(ptr) \
+	(sljit_ins*)(((sljit_ins)(ptr) + (CONST_POOL_ALIGNMENT * sizeof(sljit_ins)) - 1) & ~((CONST_POOL_ALIGNMENT * sizeof(sljit_ins)) - 1))
+#define MAX_DIFFERENCE(max_diff) \
+	(((max_diff) / (sljit_s32)sizeof(sljit_ins)) - (CONST_POOL_ALIGNMENT - 1))
+
+/* See sljit_emit_enter and sljit_emit_op0 if you want to change them. */
+static const sljit_u8 reg_map[SLJIT_NUMBER_OF_REGISTERS + 5] = {
+	0, 0, 1, 2, 3, 11, 10, 9, 8, 7, 6, 5, 4, 13, 12, 14, 15
+};
+
+static const sljit_u8 freg_map[((SLJIT_NUMBER_OF_FLOAT_REGISTERS + 2) << 1) + 1] = {
+	0,
+	0, 1, 2, 3, 4, 5, 15, 14, 13, 12, 11, 10, 9, 8,
+	7, 6,
+	0, 1, 2, 3, 4, 5, 15, 14, 13, 12, 11, 10, 9, 8,
+	7, 6
+};
+
+static const sljit_u8 freg_ebit_map[((SLJIT_NUMBER_OF_FLOAT_REGISTERS + 2) << 1) + 1] = {
+	0,
+	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+	0, 0,
+	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
+	1, 1
+};
+
+#define RM(rm) ((sljit_ins)reg_map[rm])
+#define RM8(rm) ((sljit_ins)reg_map[rm] << 8)
+#define RD(rd) ((sljit_ins)reg_map[rd] << 12)
+#define RN(rn) ((sljit_ins)reg_map[rn] << 16)
+
+#define VM(vm) (((sljit_ins)freg_map[vm]) | ((sljit_ins)freg_ebit_map[vm] << 5))
+#define VD(vd) (((sljit_ins)freg_map[vd] << 12) | ((sljit_ins)freg_ebit_map[vd] << 22))
+#define VN(vn) (((sljit_ins)freg_map[vn] << 16) | ((sljit_ins)freg_ebit_map[vn] << 7))
+
+/* --------------------------------------------------------------------- */
+/*  Instrucion forms                                                     */
+/* --------------------------------------------------------------------- */
+
+/* The instruction includes the AL condition.
+   INST_NAME - CONDITIONAL remove this flag. */
+#define COND_MASK	0xf0000000
+#define CONDITIONAL	0xe0000000
+#define PUSH_POOL	0xff000000
+
+#define ADC		0xe0a00000
+#define ADD		0xe0800000
+#define AND		0xe0000000
+#define B		0xea000000
+#define BIC		0xe1c00000
+#define BKPT		0xe1200070
+#define BL		0xeb000000
+#define BLX		0xe12fff30
+#define BX		0xe12fff10
+#define CLZ		0xe16f0f10
+#define CMN		0xe1600000
+#define CMP		0xe1400000
+#define EOR		0xe0200000
+#define LDR		0xe5100000
+#define LDR_POST	0xe4100000
+#define LDREX		0xe1900f9f
+#define LDREXB		0xe1d00f9f
+#define LDREXH		0xe1f00f9f
+#define MLA		0xe0200090
+#define MOV		0xe1a00000
+#define MUL		0xe0000090
+#define MVN		0xe1e00000
+#define NOP		0xe1a00000
+#define ORR		0xe1800000
+#define PUSH		0xe92d0000
+#define POP		0xe8bd0000
+#define REV		0xe6bf0f30
+#define REV16		0xe6bf0fb0
+#define RSB		0xe0600000
+#define RSC		0xe0e00000
+#define SBC		0xe0c00000
+#define SMULL		0xe0c00090
+#define STR		0xe5000000
+#define STREX		0xe1800f90
+#define STREXB		0xe1c00f90
+#define STREXH		0xe1e00f90
+#define SUB		0xe0400000
+#define SXTB		0xe6af0070
+#define SXTH		0xe6bf0070
+#define TST		0xe1000000
+#define UMULL		0xe0800090
+#define UXTB		0xe6ef0070
+#define UXTH		0xe6ff0070
+#define VABS_F32	0xeeb00ac0
+#define VADD_F32	0xee300a00
+#define VAND		0xf2000110
+#define VCMP_F32	0xeeb40a40
+#define VCVT_F32_S32	0xeeb80ac0
+#define VCVT_F32_U32	0xeeb80a40
+#define VCVT_F64_F32	0xeeb70ac0
+#define VCVT_S32_F32	0xeebd0ac0
+#define VDIV_F32	0xee800a00
+#define VDUP		0xee800b10
+#define VDUP_s		0xf3b00c00
+#define VEOR		0xf3000110
+#define VLD1		0xf4200000
+#define VLD1_r		0xf4a00c00
+#define VLD1_s		0xf4a00000
+#define VLDR_F32	0xed100a00
+#define VMOV_F32	0xeeb00a40
+#define VMOV		0xee000a10
+#define VMOV2		0xec400a10
+#define VMOV_i		0xf2800010
+#define VMOV_s		0xee000b10
+#define VMOVN		0xf3b20200
+#define VMRS		0xeef1fa10
+#define VMUL_F32	0xee200a00
+#define VNEG_F32	0xeeb10a40
+#define VORR		0xf2200110
+#define VPOP		0xecbd0b00
+#define VPUSH		0xed2d0b00
+#define VSHLL		0xf2800a10
+#define VSHR		0xf2800010
+#define VSRA		0xf2800110
+#define VST1		0xf4000000
+#define VST1_s		0xf4800000
+#define VSTR_F32	0xed000a00
+#define VSUB_F32	0xee300a40
+
+#if (defined SLJIT_CONFIG_ARM_V7 && SLJIT_CONFIG_ARM_V7)
+/* Arm v7 specific instructions. */
+#define MOVT		0xe3400000
+#define MOVW		0xe3000000
+#define RBIT		0xe6ff0f30
+#endif
+
+#if (defined SLJIT_ARGUMENT_CHECKS && SLJIT_ARGUMENT_CHECKS)
+
+static sljit_s32 function_check_is_freg(struct sljit_compiler *compiler, sljit_s32 fr, sljit_s32 is_32)
+{
+	if (compiler->scratches == -1)
+		return 0;
+
+	if (is_32 && fr >= SLJIT_F64_SECOND(SLJIT_FR0))
+		fr -= SLJIT_F64_SECOND(0);
+
+	return (fr >= SLJIT_FR0 && fr < (SLJIT_FR0 + compiler->fscratches))
+		|| (fr > (SLJIT_FS0 - compiler->fsaveds) && fr <= SLJIT_FS0)
+		|| (fr >= SLJIT_TMP_FREGISTER_BASE && fr < (SLJIT_TMP_FREGISTER_BASE + SLJIT_NUMBER_OF_TEMPORARY_FLOAT_REGISTERS));
+}
+
+#endif /* SLJIT_ARGUMENT_CHECKS */
+
+#if (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+
+static sljit_s32 push_cpool(struct sljit_compiler *compiler)
+{
+	/* Pushing the constant pool into the instruction stream. */
+	sljit_ins* inst;
+	sljit_uw* cpool_ptr;
+	sljit_uw* cpool_end;
+	sljit_s32 i;
+
+	/* The label could point the address after the constant pool. */
+	if (compiler->last_label && compiler->last_label->size == compiler->size)
+		compiler->last_label->size += compiler->cpool_fill + (CONST_POOL_ALIGNMENT - 1) + 1;
+
+	SLJIT_ASSERT(compiler->cpool_fill > 0 && compiler->cpool_fill <= CPOOL_SIZE);
+	inst = (sljit_ins*)ensure_buf(compiler, sizeof(sljit_ins));
+	FAIL_IF(!inst);
+	compiler->size++;
+	*inst = 0xff000000 | compiler->cpool_fill;
+
+	for (i = 0; i < CONST_POOL_ALIGNMENT - 1; i++) {
+		inst = (sljit_ins*)ensure_buf(compiler, sizeof(sljit_ins));
+		FAIL_IF(!inst);
+		compiler->size++;
+		*inst = 0;
+	}
+
+	cpool_ptr = compiler->cpool;
+	cpool_end = cpool_ptr + compiler->cpool_fill;
+	while (cpool_ptr < cpool_end) {
+		inst = (sljit_ins*)ensure_buf(compiler, sizeof(sljit_ins));
+		FAIL_IF(!inst);
+		compiler->size++;
+		*inst = *cpool_ptr++;
+	}
+	compiler->cpool_diff = CONST_POOL_EMPTY;
+	compiler->cpool_fill = 0;
+	return SLJIT_SUCCESS;
+}
+
+static sljit_s32 push_inst(struct sljit_compiler *compiler, sljit_ins inst)
+{
+	sljit_ins* ptr;
+
+	if (SLJIT_UNLIKELY(compiler->cpool_diff != CONST_POOL_EMPTY && compiler->size - compiler->cpool_diff >= MAX_DIFFERENCE(4092)))
+		FAIL_IF(push_cpool(compiler));
+
+	ptr = (sljit_ins*)ensure_buf(compiler, sizeof(sljit_ins));
+	FAIL_IF(!ptr);
+	compiler->size++;
+	*ptr = inst;
+	return SLJIT_SUCCESS;
+}
+
+static sljit_s32 push_inst_with_literal(struct sljit_compiler *compiler, sljit_ins inst, sljit_uw literal)
+{
+	sljit_ins* ptr;
+	sljit_uw cpool_index = CPOOL_SIZE;
+	sljit_uw* cpool_ptr;
+	sljit_uw* cpool_end;
+	sljit_u8* cpool_unique_ptr;
+
+	if (SLJIT_UNLIKELY(compiler->cpool_diff != CONST_POOL_EMPTY && compiler->size - compiler->cpool_diff >= MAX_DIFFERENCE(4092)))
+		FAIL_IF(push_cpool(compiler));
+	else if (compiler->cpool_fill > 0) {
+		cpool_ptr = compiler->cpool;
+		cpool_end = cpool_ptr + compiler->cpool_fill;
+		cpool_unique_ptr = compiler->cpool_unique;
+		do {
+			if ((*cpool_ptr == literal) && !(*cpool_unique_ptr)) {
+				cpool_index = (sljit_uw)(cpool_ptr - compiler->cpool);
+				break;
+			}
+			cpool_ptr++;
+			cpool_unique_ptr++;
+		} while (cpool_ptr < cpool_end);
+	}
+
+	if (cpool_index == CPOOL_SIZE) {
+		/* Must allocate a new entry in the literal pool. */
+		if (compiler->cpool_fill < CPOOL_SIZE) {
+			cpool_index = compiler->cpool_fill;
+			compiler->cpool_fill++;
+		}
+		else {
+			FAIL_IF(push_cpool(compiler));
+			cpool_index = 0;
+			compiler->cpool_fill = 1;
+		}
+	}
+
+	SLJIT_ASSERT((inst & 0xfff) == 0);
+	ptr = (sljit_ins*)ensure_buf(compiler, sizeof(sljit_ins));
+	FAIL_IF(!ptr);
+	compiler->size++;
+	*ptr = inst | cpool_index;
+
+	compiler->cpool[cpool_index] = literal;
+	compiler->cpool_unique[cpool_index] = 0;
+	if (compiler->cpool_diff == CONST_POOL_EMPTY)
+		compiler->cpool_diff = compiler->size;
+	return SLJIT_SUCCESS;
+}
+
+static sljit_s32 push_inst_with_unique_literal(struct sljit_compiler *compiler, sljit_ins inst, sljit_uw literal)
+{
+	sljit_ins* ptr;
+
+	if (SLJIT_UNLIKELY((compiler->cpool_diff != CONST_POOL_EMPTY && compiler->size - compiler->cpool_diff >= MAX_DIFFERENCE(4092)) || compiler->cpool_fill >= CPOOL_SIZE))
+		FAIL_IF(push_cpool(compiler));
+
+	SLJIT_ASSERT(compiler->cpool_fill < CPOOL_SIZE && (inst & 0xfff) == 0);
+	ptr = (sljit_ins*)ensure_buf(compiler, sizeof(sljit_ins));
+	FAIL_IF(!ptr);
+	compiler->size++;
+	*ptr = inst | compiler->cpool_fill;
+
+	compiler->cpool[compiler->cpool_fill] = literal;
+	compiler->cpool_unique[compiler->cpool_fill] = 1;
+	compiler->cpool_fill++;
+	if (compiler->cpool_diff == CONST_POOL_EMPTY)
+		compiler->cpool_diff = compiler->size;
+	return SLJIT_SUCCESS;
+}
+
+static SLJIT_INLINE sljit_s32 prepare_blx(struct sljit_compiler *compiler)
+{
+	/* Place for at least two instruction (doesn't matter whether the first has a literal). */
+	if (SLJIT_UNLIKELY(compiler->cpool_diff != CONST_POOL_EMPTY && compiler->size - compiler->cpool_diff >= MAX_DIFFERENCE(4088)))
+		return push_cpool(compiler);
+	return SLJIT_SUCCESS;
+}
+
+static SLJIT_INLINE sljit_s32 emit_blx(struct sljit_compiler *compiler)
+{
+	/* Must follow tightly the previous instruction (to be able to convert it to bl instruction). */
+	SLJIT_ASSERT(compiler->cpool_diff == CONST_POOL_EMPTY || compiler->size - compiler->cpool_diff < MAX_DIFFERENCE(4092));
+	SLJIT_ASSERT(reg_map[TMP_REG1] != 14);
+
+	return push_inst(compiler, BLX | RM(TMP_REG1));
+}
+
+static sljit_uw patch_pc_relative_loads(sljit_uw *last_pc_patch, sljit_uw *code_ptr, sljit_uw* const_pool, sljit_uw cpool_size)
+{
+	sljit_uw diff;
+	sljit_uw ind;
+	sljit_uw counter = 0;
+	sljit_uw* clear_const_pool = const_pool;
+	sljit_uw* clear_const_pool_end = const_pool + cpool_size;
+
+	SLJIT_ASSERT(const_pool - code_ptr <= CONST_POOL_ALIGNMENT);
+	/* Set unused flag for all literals in the constant pool.
+	   I.e.: unused literals can belong to branches, which can be encoded as B or BL.
+	   We can "compress" the constant pool by discarding these literals. */
+	while (clear_const_pool < clear_const_pool_end)
+		*clear_const_pool++ = (sljit_uw)(-1);
+
+	while (last_pc_patch < code_ptr) {
+		/* Data transfer instruction with Rn == r15. */
+		if ((*last_pc_patch & 0x0e0f0000) == 0x040f0000) {
+			diff = (sljit_uw)(const_pool - last_pc_patch);
+			ind = (*last_pc_patch) & 0xfff;
+
+			/* Must be a load instruction with immediate offset. */
+			SLJIT_ASSERT(ind < cpool_size && !(*last_pc_patch & (1 << 25)) && (*last_pc_patch & (1 << 20)));
+			if ((sljit_s32)const_pool[ind] < 0) {
+				const_pool[ind] = counter;
+				ind = counter;
+				counter++;
+			}
+			else
+				ind = const_pool[ind];
+
+			SLJIT_ASSERT(diff >= 1);
+			if (diff >= 2 || ind > 0) {
+				diff = (diff + (sljit_uw)ind - 2) << 2;
+				SLJIT_ASSERT(diff <= 0xfff);
+				*last_pc_patch = (*last_pc_patch & ~(sljit_uw)0xfff) | diff;
+			}
+			else
+				*last_pc_patch = (*last_pc_patch & ~(sljit_uw)(0xfff | (1 << 23))) | 0x004;
+		}
+		last_pc_patch++;
+	}
+	return counter;
+}
+
+/* In some rare ocasions we may need future patches. The probability is close to 0 in practice. */
+struct future_patch {
+	struct future_patch* next;
+	sljit_s32 index;
+	sljit_s32 value;
+};
+
+static sljit_s32 resolve_const_pool_index(struct sljit_compiler *compiler, struct future_patch **first_patch, sljit_uw cpool_current_index, sljit_uw *cpool_start_address, sljit_uw *buf_ptr)
+{
+	sljit_u32 value;
+	struct future_patch *curr_patch, *prev_patch;
+
+	SLJIT_UNUSED_ARG(compiler);
+
+	/* Using the values generated by patch_pc_relative_loads. */
+	if (!*first_patch)
+		value = cpool_start_address[cpool_current_index];
+	else {
+		curr_patch = *first_patch;
+		prev_patch = NULL;
+		while (1) {
+			if (!curr_patch) {
+				value = cpool_start_address[cpool_current_index];
+				break;
+			}
+			if ((sljit_uw)curr_patch->index == cpool_current_index) {
+				value = (sljit_uw)curr_patch->value;
+				if (prev_patch)
+					prev_patch->next = curr_patch->next;
+				else
+					*first_patch = curr_patch->next;
+				SLJIT_FREE(curr_patch, compiler->allocator_data);
+				break;
+			}
+			prev_patch = curr_patch;
+			curr_patch = curr_patch->next;
+		}
+	}
+
+	if ((sljit_sw)value >= 0) {
+		if (value > cpool_current_index) {
+			curr_patch = (struct future_patch*)SLJIT_MALLOC(sizeof(struct future_patch), compiler->allocator_data);
+			if (!curr_patch) {
+				while (*first_patch) {
+					curr_patch = *first_patch;
+					*first_patch = (*first_patch)->next;
+					SLJIT_FREE(curr_patch, compiler->allocator_data);
+				}
+				return SLJIT_ERR_ALLOC_FAILED;
+			}
+			curr_patch->next = *first_patch;
+			curr_patch->index = (sljit_sw)value;
+			curr_patch->value = (sljit_sw)cpool_start_address[value];
+			*first_patch = curr_patch;
+		}
+		cpool_start_address[value] = *buf_ptr;
+	}
+	return SLJIT_SUCCESS;
+}
+
+#else
+
+static sljit_s32 push_inst(struct sljit_compiler *compiler, sljit_ins inst)
+{
+	sljit_ins* ptr;
+
+	ptr = (sljit_ins*)ensure_buf(compiler, sizeof(sljit_ins));
+	FAIL_IF(!ptr);
+	compiler->size++;
+	*ptr = inst;
+	return SLJIT_SUCCESS;
+}
+
+static SLJIT_INLINE sljit_s32 emit_imm(struct sljit_compiler *compiler, sljit_s32 reg, sljit_sw imm)
+{
+	FAIL_IF(push_inst(compiler, MOVW | RD(reg) | ((imm << 4) & 0xf0000) | ((sljit_u32)imm & 0xfff)));
+	return push_inst(compiler, MOVT | RD(reg) | ((imm >> 12) & 0xf0000) | (((sljit_u32)imm >> 16) & 0xfff));
+}
+
+#endif
+
+static SLJIT_INLINE sljit_s32 detect_jump_type(struct sljit_jump *jump, sljit_uw *code_ptr, sljit_uw *code, sljit_sw executable_offset)
+{
+	sljit_sw diff;
+
+	if (jump->flags & SLJIT_REWRITABLE_JUMP)
+		return 0;
+
+#if (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+	if (jump->flags & IS_BL)
+		code_ptr--;
+#endif /* SLJIT_CONFIG_ARM_V6 */
+
+	if (jump->flags & JUMP_ADDR)
+		diff = ((sljit_sw)jump->u.target - (sljit_sw)(code_ptr + 2) - executable_offset);
+	else {
+		SLJIT_ASSERT(jump->u.label != NULL);
+		diff = ((sljit_sw)(code + jump->u.label->size) - (sljit_sw)(code_ptr + 2));
+	}
+
+	/* Branch to Thumb code has not been optimized yet. */
+	if (diff & 0x3)
+		return 0;
+
+#if (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+	if (jump->flags & IS_BL) {
+		if (diff <= 0x01ffffff && diff >= -0x02000000) {
+			*code_ptr = (BL - CONDITIONAL) | (*(code_ptr + 1) & COND_MASK);
+			jump->flags |= PATCH_B;
+			return 1;
+		}
+	}
+	else {
+		if (diff <= 0x01ffffff && diff >= -0x02000000) {
+			*code_ptr = (B - CONDITIONAL) | (*code_ptr & COND_MASK);
+			jump->flags |= PATCH_B;
+		}
+	}
+#else /* !SLJIT_CONFIG_ARM_V6 */
+	if (diff <= 0x01ffffff && diff >= -0x02000000) {
+		*code_ptr = ((jump->flags & IS_BL) ? (BL - CONDITIONAL) : (B - CONDITIONAL)) | (*code_ptr & COND_MASK);
+		jump->flags |= PATCH_B;
+		return 1;
+	}
+#endif /* SLJIT_CONFIG_ARM_V6 */
+	return 0;
+}
+
+static void set_jump_addr(sljit_uw jump_ptr, sljit_sw executable_offset, sljit_uw new_addr, sljit_s32 flush_cache)
+{
+#if (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+	sljit_ins *ptr = (sljit_ins*)jump_ptr;
+	sljit_ins *inst = (sljit_ins*)ptr[0];
+	sljit_ins mov_pc = ptr[1];
+	sljit_s32 bl = (mov_pc & 0x0000f000) != RD(TMP_PC);
+	sljit_sw diff = (sljit_sw)(((sljit_sw)new_addr - (sljit_sw)(inst + 2) - executable_offset) >> 2);
+
+	SLJIT_UNUSED_ARG(executable_offset);
+
+	if (diff <= 0x7fffff && diff >= -0x800000) {
+		/* Turn to branch. */
+		if (!bl) {
+			if (flush_cache) {
+				SLJIT_UPDATE_WX_FLAGS(inst, inst + 1, 0);
+			}
+			inst[0] = (mov_pc & COND_MASK) | (B - CONDITIONAL) | (diff & 0xffffff);
+			if (flush_cache) {
+				SLJIT_UPDATE_WX_FLAGS(inst, inst + 1, 1);
+				inst = (sljit_ins*)SLJIT_ADD_EXEC_OFFSET(inst, executable_offset);
+				SLJIT_CACHE_FLUSH(inst, inst + 1);
+			}
+		} else {
+			if (flush_cache) {
+				SLJIT_UPDATE_WX_FLAGS(inst, inst + 2, 0);
+			}
+			inst[0] = (mov_pc & COND_MASK) | (BL - CONDITIONAL) | (diff & 0xffffff);
+			inst[1] = NOP;
+			if (flush_cache) {
+				SLJIT_UPDATE_WX_FLAGS(inst, inst + 2, 1);
+				inst = (sljit_ins*)SLJIT_ADD_EXEC_OFFSET(inst, executable_offset);
+				SLJIT_CACHE_FLUSH(inst, inst + 2);
+			}
+		}
+	} else {
+		/* Get the position of the constant. */
+		if (mov_pc & (1 << 23))
+			ptr = inst + ((mov_pc & 0xfff) >> 2) + 2;
+		else
+			ptr = inst + 1;
+
+		if (*inst != mov_pc) {
+			if (flush_cache) {
+				SLJIT_UPDATE_WX_FLAGS(inst, inst + (!bl ? 1 : 2), 0);
+			}
+			inst[0] = mov_pc;
+			if (!bl) {
+				if (flush_cache) {
+					SLJIT_UPDATE_WX_FLAGS(inst, inst + 1, 1);
+					inst = (sljit_ins*)SLJIT_ADD_EXEC_OFFSET(inst, executable_offset);
+					SLJIT_CACHE_FLUSH(inst, inst + 1);
+				}
+			} else {
+				inst[1] = BLX | RM(TMP_REG1);
+				if (flush_cache) {
+					SLJIT_UPDATE_WX_FLAGS(inst, inst + 2, 1);
+					inst = (sljit_ins*)SLJIT_ADD_EXEC_OFFSET(inst, executable_offset);
+					SLJIT_CACHE_FLUSH(inst, inst + 2);
+				}
+			}
+		}
+
+		if (flush_cache) {
+			SLJIT_UPDATE_WX_FLAGS(ptr, ptr + 1, 0);
+		}
+
+		*ptr = new_addr;
+
+		if (flush_cache) {
+			SLJIT_UPDATE_WX_FLAGS(ptr, ptr + 1, 1);
+		}
+	}
+#else /* !SLJIT_CONFIG_ARM_V6 */
+	sljit_ins *inst = (sljit_ins*)jump_ptr;
+
+	SLJIT_UNUSED_ARG(executable_offset);
+
+	SLJIT_ASSERT((inst[0] & 0xfff00000) == MOVW && (inst[1] & 0xfff00000) == MOVT);
+
+	if (flush_cache) {
+		SLJIT_UPDATE_WX_FLAGS(inst, inst + 2, 0);
+	}
+
+	inst[0] = MOVW | (inst[0] & 0xf000) | ((new_addr << 4) & 0xf0000) | (new_addr & 0xfff);
+	inst[1] = MOVT | (inst[1] & 0xf000) | ((new_addr >> 12) & 0xf0000) | ((new_addr >> 16) & 0xfff);
+
+	if (flush_cache) {
+		SLJIT_UPDATE_WX_FLAGS(inst, inst + 2, 1);
+		inst = (sljit_ins*)SLJIT_ADD_EXEC_OFFSET(inst, executable_offset);
+		SLJIT_CACHE_FLUSH(inst, inst + 2);
+	}
+#endif /* SLJIT_CONFIG_ARM_V6 */
+}
+
+static sljit_uw get_imm(sljit_uw imm);
+static sljit_s32 load_immediate(struct sljit_compiler *compiler, sljit_s32 reg, sljit_uw imm);
+static sljit_s32 emit_op_mem(struct sljit_compiler *compiler, sljit_s32 flags, sljit_s32 reg, sljit_s32 arg, sljit_sw argw, sljit_s32 tmp_reg);
+
+static void set_const_value(sljit_uw addr, sljit_sw executable_offset, sljit_uw new_constant, sljit_s32 flush_cache)
+{
+#if (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+	sljit_ins *ptr = (sljit_ins*)addr;
+	sljit_ins *inst = (sljit_ins*)ptr[0];
+	sljit_uw ldr_literal = ptr[1];
+	sljit_uw src2;
+
+	SLJIT_UNUSED_ARG(executable_offset);
+
+	src2 = get_imm(new_constant);
+	if (src2) {
+		if (flush_cache) {
+			SLJIT_UPDATE_WX_FLAGS(inst, inst + 1, 0);
+		}
+
+		*inst = 0xe3a00000 | (ldr_literal & 0xf000) | src2;
+
+		if (flush_cache) {
+			SLJIT_UPDATE_WX_FLAGS(inst, inst + 1, 1);
+			inst = (sljit_ins*)SLJIT_ADD_EXEC_OFFSET(inst, executable_offset);
+			SLJIT_CACHE_FLUSH(inst, inst + 1);
+		}
+		return;
+	}
+
+	src2 = get_imm(~new_constant);
+	if (src2) {
+		if (flush_cache) {
+			SLJIT_UPDATE_WX_FLAGS(inst, inst + 1, 0);
+		}
+
+		*inst = 0xe3e00000 | (ldr_literal & 0xf000) | src2;
+
+		if (flush_cache) {
+			SLJIT_UPDATE_WX_FLAGS(inst, inst + 1, 1);
+			inst = (sljit_ins*)SLJIT_ADD_EXEC_OFFSET(inst, executable_offset);
+			SLJIT_CACHE_FLUSH(inst, inst + 1);
+		}
+		return;
+	}
+
+	if (ldr_literal & (1 << 23))
+		ptr = inst + ((ldr_literal & 0xfff) >> 2) + 2;
+	else
+		ptr = inst + 1;
+
+	if (*inst != ldr_literal) {
+		if (flush_cache) {
+			SLJIT_UPDATE_WX_FLAGS(inst, inst + 1, 0);
+		}
+
+		*inst = ldr_literal;
+
+		if (flush_cache) {
+			SLJIT_UPDATE_WX_FLAGS(inst, inst + 1, 1);
+			inst = (sljit_ins*)SLJIT_ADD_EXEC_OFFSET(inst, executable_offset);
+			SLJIT_CACHE_FLUSH(inst, inst + 1);
+		}
+	}
+
+	if (flush_cache) {
+		SLJIT_UPDATE_WX_FLAGS(ptr, ptr + 1, 0);
+	}
+
+	*ptr = new_constant;
+
+	if (flush_cache) {
+		SLJIT_UPDATE_WX_FLAGS(ptr, ptr + 1, 1);
+	}
+#else /* !SLJIT_CONFIG_ARM_V6 */
+	sljit_ins *inst = (sljit_ins*)addr;
+
+	SLJIT_UNUSED_ARG(executable_offset);
+
+	SLJIT_ASSERT((inst[0] & 0xfff00000) == MOVW && (inst[1] & 0xfff00000) == MOVT);
+
+	if (flush_cache) {
+		SLJIT_UPDATE_WX_FLAGS(inst, inst + 2, 0);
+	}
+
+	inst[0] = MOVW | (inst[0] & 0xf000) | ((new_constant << 4) & 0xf0000) | (new_constant & 0xfff);
+	inst[1] = MOVT | (inst[1] & 0xf000) | ((new_constant >> 12) & 0xf0000) | ((new_constant >> 16) & 0xfff);
+
+	if (flush_cache) {
+		SLJIT_UPDATE_WX_FLAGS(inst, inst + 2, 1);
+		inst = (sljit_ins*)SLJIT_ADD_EXEC_OFFSET(inst, executable_offset);
+		SLJIT_CACHE_FLUSH(inst, inst + 2);
+	}
+#endif /* SLJIT_CONFIG_ARM_V6 */
+}
+
+static SLJIT_INLINE sljit_sw mov_addr_get_length(struct sljit_jump *jump, sljit_ins *code_ptr, sljit_ins *code, sljit_sw executable_offset)
+{
+	sljit_uw addr;
+	sljit_sw diff;
+	SLJIT_UNUSED_ARG(executable_offset);
+
+	if (jump->flags & JUMP_ADDR)
+		addr = jump->u.target;
+	else
+		addr = (sljit_uw)SLJIT_ADD_EXEC_OFFSET(code + jump->u.label->size, executable_offset);
+
+	/* The pc+8 offset is represented by the 2 * SSIZE_OF(ins) below. */
+	diff = (sljit_sw)addr - (sljit_sw)SLJIT_ADD_EXEC_OFFSET(code_ptr, executable_offset);
+
+	if ((diff & 0x3) == 0 && diff <= (0x3fc + 2 * SSIZE_OF(ins)) && diff >= (-0x3fc + 2 * SSIZE_OF(ins))) {
+		jump->flags |= PATCH_B;
+		return 0;
+	}
+
+#if (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+	return 0;
+#else /* !SLJIT_CONFIG_ARM_V6 */
+	return 1;
+#endif /* SLJIT_CONFIG_ARM_V6 */
+}
+
+#if (defined SLJIT_CONFIG_ARM_V7 && SLJIT_CONFIG_ARM_V7)
+
+static void reduce_code_size(struct sljit_compiler *compiler)
+{
+	struct sljit_label *label;
+	struct sljit_jump *jump;
+	struct sljit_const *const_;
+	SLJIT_NEXT_DEFINE_TYPES;
+	sljit_uw total_size;
+	sljit_uw size_reduce = 0;
+	sljit_sw diff;
+
+	label = compiler->labels;
+	jump = compiler->jumps;
+	const_ = compiler->consts;
+	SLJIT_NEXT_INIT_TYPES();
+
+	while (1) {
+		SLJIT_GET_NEXT_MIN();
+
+		if (next_min_addr == SLJIT_MAX_ADDRESS)
+			break;
+
+		if (next_min_addr == next_label_size) {
+			label->size -= size_reduce;
+
+			label = label->next;
+			next_label_size = SLJIT_GET_NEXT_SIZE(label);
+		}
+
+		if (next_min_addr == next_const_addr) {
+			const_->addr -= size_reduce;
+			const_ = const_->next;
+			next_const_addr = SLJIT_GET_NEXT_ADDRESS(const_);
+			continue;
+		}
+
+		if (next_min_addr != next_jump_addr)
+			continue;
+
+		jump->addr -= size_reduce;
+		if (!(jump->flags & JUMP_MOV_ADDR)) {
+			total_size = JUMP_MAX_SIZE - 1;
+
+			if (!(jump->flags & (SLJIT_REWRITABLE_JUMP | JUMP_ADDR))) {
+				/* Unit size: instruction. */
+				diff = (sljit_sw)jump->u.label->size - (sljit_sw)jump->addr - 2;
+
+				if (diff <= (0x01ffffff / SSIZE_OF(ins)) && diff >= (-0x02000000 / SSIZE_OF(ins)))
+					total_size = 1 - 1;
+			}
+
+			size_reduce += JUMP_MAX_SIZE - 1 - total_size;
+		} else {
+			/* Real size minus 1. Unit size: instruction. */
+			total_size = 1;
+
+			if (!(jump->flags & JUMP_ADDR)) {
+				diff = (sljit_sw)jump->u.label->size - (sljit_sw)jump->addr;
+				if (diff <= 0xff + 2 && diff >= -0xff + 2)
+					total_size = 0;
+			}
+
+			size_reduce += 1 - total_size;
+		}
+
+		jump->flags |= total_size << JUMP_SIZE_SHIFT;
+		jump = jump->next;
+		next_jump_addr = SLJIT_GET_NEXT_ADDRESS(jump);
+	}
+
+	compiler->size -= size_reduce;
+}
+
+#endif /* SLJIT_CONFIG_ARM_V7 */
+
+SLJIT_API_FUNC_ATTRIBUTE void* sljit_generate_code(struct sljit_compiler *compiler, sljit_s32 options, void *exec_allocator_data)
+{
+	struct sljit_memory_fragment *buf;
+	sljit_ins *code;
+	sljit_ins *code_ptr;
+	sljit_ins *buf_ptr;
+	sljit_ins *buf_end;
+	sljit_uw word_count;
+	SLJIT_NEXT_DEFINE_TYPES;
+	sljit_sw executable_offset;
+	sljit_uw addr;
+	sljit_sw diff;
+#if (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+	sljit_uw cpool_size;
+	sljit_uw cpool_skip_alignment;
+	sljit_uw cpool_current_index;
+	sljit_ins *cpool_start_address;
+	sljit_ins *last_pc_patch;
+	struct future_patch *first_patch;
+#endif
+
+	struct sljit_label *label;
+	struct sljit_jump *jump;
+	struct sljit_const *const_;
+
+	CHECK_ERROR_PTR();
+	CHECK_PTR(check_sljit_generate_code(compiler));
+
+	/* Second code generation pass. */
+#if (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+	compiler->size += (compiler->patches << 1);
+	if (compiler->cpool_fill > 0)
+		compiler->size += compiler->cpool_fill + CONST_POOL_ALIGNMENT - 1;
+#else /* !SLJIT_CONFIG_ARM_V6 */
+	reduce_code_size(compiler);
+#endif /* SLJIT_CONFIG_ARM_V6 */
+	code = (sljit_ins*)allocate_executable_memory(compiler->size * sizeof(sljit_ins), options, exec_allocator_data, &executable_offset);
+	PTR_FAIL_WITH_EXEC_IF(code);
+
+	reverse_buf(compiler);
+	buf = compiler->buf;
+
+#if (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+	cpool_size = 0;
+	cpool_skip_alignment = 0;
+	cpool_current_index = 0;
+	cpool_start_address = NULL;
+	first_patch = NULL;
+	last_pc_patch = code;
+#endif /* SLJIT_CONFIG_ARM_V6 */
+
+	code_ptr = code;
+	word_count = 0;
+	label = compiler->labels;
+	jump = compiler->jumps;
+	const_ = compiler->consts;
+	SLJIT_NEXT_INIT_TYPES();
+	SLJIT_GET_NEXT_MIN();
+
+	do {
+		buf_ptr = (sljit_ins*)buf->memory;
+		buf_end = buf_ptr + (buf->used_size >> 2);
+		do {
+#if (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+			if (cpool_size > 0) {
+				if (cpool_skip_alignment > 0) {
+					buf_ptr++;
+					cpool_skip_alignment--;
+				} else {
+					if (SLJIT_UNLIKELY(resolve_const_pool_index(compiler, &first_patch, cpool_current_index, cpool_start_address, buf_ptr))) {
+						SLJIT_FREE_EXEC(code, exec_allocator_data);
+						compiler->error = SLJIT_ERR_ALLOC_FAILED;
+						return NULL;
+					}
+					buf_ptr++;
+					if (++cpool_current_index >= cpool_size) {
+						SLJIT_ASSERT(!first_patch);
+						cpool_size = 0;
+					}
+				}
+			} else if ((*buf_ptr & 0xff000000) != PUSH_POOL) {
+#endif /* SLJIT_CONFIG_ARM_V6 */
+				*code_ptr = *buf_ptr++;
+				if (next_min_addr == word_count) {
+					SLJIT_ASSERT(!label || label->size >= word_count);
+					SLJIT_ASSERT(!jump || jump->addr >= word_count);
+					SLJIT_ASSERT(!const_ || const_->addr >= word_count);
+
+					if (next_min_addr == next_label_size) {
+						label->u.addr = (sljit_uw)SLJIT_ADD_EXEC_OFFSET(code_ptr, executable_offset);
+						label->size = (sljit_uw)(code_ptr - code);
+						label = label->next;
+						next_label_size = SLJIT_GET_NEXT_SIZE(label);
+					}
+
+					/* These structures are ordered by their address. */
+					if (next_min_addr == next_jump_addr) {
+						if (!(jump->flags & JUMP_MOV_ADDR)) {
+#if (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+							if (detect_jump_type(jump, code_ptr, code, executable_offset))
+								code_ptr--;
+							jump->addr = (sljit_uw)code_ptr;
+#else /* !SLJIT_CONFIG_ARM_V6 */
+							word_count += jump->flags >> JUMP_SIZE_SHIFT;
+							jump->addr = (sljit_uw)code_ptr;
+							if (!detect_jump_type(jump, code_ptr, code, executable_offset)) {
+								code_ptr[2] = code_ptr[0];
+								addr = ((code_ptr[0] & 0xf) << 12);
+								code_ptr[0] = MOVW | addr;
+								code_ptr[1] = MOVT | addr;
+								code_ptr += 2;
+							}
+							SLJIT_ASSERT((sljit_uw)code_ptr - jump->addr <= (jump->flags >> JUMP_SIZE_SHIFT) * sizeof(sljit_ins));
+#endif /* SLJIT_CONFIG_ARM_V6 */
+						} else {
+#if (defined SLJIT_CONFIG_ARM_V7 && SLJIT_CONFIG_ARM_V7)
+							word_count += jump->flags >> JUMP_SIZE_SHIFT;
+#endif /* SLJIT_CONFIG_ARM_V7 */
+							addr = (sljit_uw)code_ptr;
+							code_ptr += mov_addr_get_length(jump, code_ptr, code, executable_offset);
+							jump->addr = addr;
+						}
+						jump = jump->next;
+						next_jump_addr = SLJIT_GET_NEXT_ADDRESS(jump);
+					} else if (next_min_addr == next_const_addr) {
+						const_->addr = (sljit_uw)code_ptr;
+						const_ = const_->next;
+						next_const_addr = SLJIT_GET_NEXT_ADDRESS(const_);
+					}
+
+					SLJIT_GET_NEXT_MIN();
+				}
+				code_ptr++;
+#if (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+			} else {
+				/* Fortunately, no need to shift. */
+				cpool_size = *buf_ptr++ & ~PUSH_POOL;
+				SLJIT_ASSERT(cpool_size > 0);
+				cpool_start_address = ALIGN_INSTRUCTION(code_ptr + 1);
+				cpool_current_index = patch_pc_relative_loads(last_pc_patch, code_ptr, cpool_start_address, cpool_size);
+				if (cpool_current_index > 0) {
+					/* Unconditional branch. */
+					*code_ptr = B | (((sljit_ins)(cpool_start_address - code_ptr) + cpool_current_index - 2) & ~PUSH_POOL);
+					code_ptr = (sljit_ins*)(cpool_start_address + cpool_current_index);
+				}
+				cpool_skip_alignment = CONST_POOL_ALIGNMENT - 1;
+				cpool_current_index = 0;
+				last_pc_patch = code_ptr;
+			}
+#endif /* SLJIT_CONFIG_ARM_V6 */
+			word_count++;
+		} while (buf_ptr < buf_end);
+		buf = buf->next;
+	} while (buf);
+
+	if (label && label->size == word_count) {
+		label->u.addr = (sljit_uw)SLJIT_ADD_EXEC_OFFSET(code_ptr, executable_offset);
+		label->size = (sljit_uw)(code_ptr - code);
+		label = label->next;
+	}
+
+	SLJIT_ASSERT(!label);
+	SLJIT_ASSERT(!jump);
+	SLJIT_ASSERT(!const_);
+
+#if (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+	SLJIT_ASSERT(cpool_size == 0);
+	if (compiler->cpool_fill > 0) {
+		cpool_start_address = ALIGN_INSTRUCTION(code_ptr);
+		cpool_current_index = patch_pc_relative_loads(last_pc_patch, code_ptr, cpool_start_address, compiler->cpool_fill);
+		if (cpool_current_index > 0)
+			code_ptr = (sljit_ins*)(cpool_start_address + cpool_current_index);
+
+		buf_ptr = compiler->cpool;
+		buf_end = buf_ptr + compiler->cpool_fill;
+		cpool_current_index = 0;
+		while (buf_ptr < buf_end) {
+			if (SLJIT_UNLIKELY(resolve_const_pool_index(compiler, &first_patch, cpool_current_index, cpool_start_address, buf_ptr))) {
+				SLJIT_FREE_EXEC(code, exec_allocator_data);
+				compiler->error = SLJIT_ERR_ALLOC_FAILED;
+				return NULL;
+			}
+			buf_ptr++;
+			cpool_current_index++;
+		}
+		SLJIT_ASSERT(!first_patch);
+	}
+#endif
+
+	jump = compiler->jumps;
+	while (jump) {
+		addr = (jump->flags & JUMP_ADDR) ? jump->u.target : jump->u.label->u.addr;
+		buf_ptr = (sljit_ins*)jump->addr;
+
+		if (jump->flags & JUMP_MOV_ADDR) {
+#if (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+			SLJIT_ASSERT((buf_ptr[0] & (sljit_ins)0xffff0000) == 0xe59f0000);
+#else /* !SLJIT_CONFIG_ARM_V6 */
+			SLJIT_ASSERT((buf_ptr[0] & ~(sljit_ins)0xf000) == 0);
+#endif /* SLJIT_CONFIG_ARM_V6 */
+
+			if (jump->flags & PATCH_B) {
+				SLJIT_ASSERT((((sljit_sw)addr - (sljit_sw)SLJIT_ADD_EXEC_OFFSET(buf_ptr + 2, executable_offset)) & 0x3) == 0);
+				diff = ((sljit_sw)addr - (sljit_sw)SLJIT_ADD_EXEC_OFFSET(buf_ptr + 2, executable_offset)) >> 2;
+
+				SLJIT_ASSERT(diff <= 0xff && diff >= -0xff);
+
+				addr = ADD;
+				if (diff < 0) {
+					diff = -diff;
+					addr = SUB;
+				}
+
+				buf_ptr[0] = addr | (buf_ptr[0] & 0xf000) | RN(TMP_PC) | (1 << 25) | (0xf << 8) | (sljit_ins)(diff & 0xff);
+			} else {
+#if (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+				buf_ptr[((buf_ptr[0] & 0xfff) >> 2) + 2] = addr;
+#else /* !SLJIT_CONFIG_ARM_V6 */
+				buf_ptr[1] = MOVT | buf_ptr[0] | ((addr >> 12) & 0xf0000) | ((addr >> 16) & 0xfff);
+				buf_ptr[0] = MOVW | buf_ptr[0] | ((addr << 4) & 0xf0000) | (addr & 0xfff);
+#endif /* SLJIT_CONFIG_ARM_V6 */
+			}
+		} else if (jump->flags & PATCH_B) {
+			diff = (sljit_sw)addr - (sljit_sw)SLJIT_ADD_EXEC_OFFSET(buf_ptr + 2, executable_offset);
+			SLJIT_ASSERT(diff <= 0x01ffffff && diff >= -0x02000000);
+			*buf_ptr |= (diff >> 2) & 0x00ffffff;
+		} else {
+#if (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+			if (jump->flags & IS_BL)
+				buf_ptr--;
+
+			if (jump->flags & SLJIT_REWRITABLE_JUMP) {
+				jump->addr = (sljit_uw)code_ptr;
+				code_ptr[0] = (sljit_ins)buf_ptr;
+				code_ptr[1] = *buf_ptr;
+				set_jump_addr((sljit_uw)code_ptr, executable_offset, addr, 0);
+				code_ptr += 2;
+			} else {
+				if (*buf_ptr & (1 << 23))
+					buf_ptr += ((*buf_ptr & 0xfff) >> 2) + 2;
+				else
+					buf_ptr += 1;
+				*buf_ptr = addr;
+			}
+#else /* !SLJIT_CONFIG_ARM_V6 */
+			set_jump_addr((sljit_uw)buf_ptr, executable_offset, addr, 0);
+#endif /* SLJIT_CONFIG_ARM_V6 */
+		}
+
+		jump = jump->next;
+	}
+
+#if (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+	const_ = compiler->consts;
+	while (const_) {
+		buf_ptr = (sljit_ins*)const_->addr;
+		const_->addr = (sljit_uw)code_ptr;
+
+		code_ptr[0] = (sljit_ins)buf_ptr;
+		code_ptr[1] = *buf_ptr;
+		if (*buf_ptr & (1 << 23))
+			buf_ptr += ((*buf_ptr & 0xfff) >> 2) + 2;
+		else
+			buf_ptr += 1;
+		/* Set the value again (can be a simple constant). */
+		set_const_value((sljit_uw)code_ptr, executable_offset, *buf_ptr, 0);
+		code_ptr += 2;
+
+		const_ = const_->next;
+	}
+#endif /* SLJIT_CONFIG_ARM_V6 */
+
+	SLJIT_ASSERT(code_ptr - code <= (sljit_s32)compiler->size);
+
+	compiler->error = SLJIT_ERR_COMPILED;
+	compiler->executable_offset = executable_offset;
+	compiler->executable_size = (sljit_uw)(code_ptr - code) * sizeof(sljit_uw);
+
+	code = (sljit_ins*)SLJIT_ADD_EXEC_OFFSET(code, executable_offset);
+	code_ptr = (sljit_ins*)SLJIT_ADD_EXEC_OFFSET(code_ptr, executable_offset);
+
+	SLJIT_CACHE_FLUSH(code, code_ptr);
+	SLJIT_UPDATE_WX_FLAGS(code, code_ptr, 1);
+	return code;
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_has_cpu_feature(sljit_s32 feature_type)
+{
+	switch (feature_type) {
+	case SLJIT_HAS_FPU:
+	case SLJIT_HAS_F64_AS_F32_PAIR:
+#ifdef SLJIT_IS_FPU_AVAILABLE
+		return (SLJIT_IS_FPU_AVAILABLE) != 0;
+#else
+		/* Available by default. */
+		return 1;
+#endif /* SLJIT_IS_FPU_AVAILABLE */
+	case SLJIT_HAS_SIMD:
+#if (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+		return 0;
+#else
+#ifdef SLJIT_IS_FPU_AVAILABLE
+		return (SLJIT_IS_FPU_AVAILABLE) != 0;
+#else
+		/* Available by default. */
+		return 1;
+#endif /* SLJIT_IS_FPU_AVAILABLE */
+#endif /* SLJIT_CONFIG_ARM_V6 */
+
+	case SLJIT_SIMD_REGS_ARE_PAIRS:
+	case SLJIT_HAS_CLZ:
+	case SLJIT_HAS_ROT:
+	case SLJIT_HAS_CMOV:
+	case SLJIT_HAS_REV:
+	case SLJIT_HAS_PREFETCH:
+	case SLJIT_HAS_COPY_F32:
+	case SLJIT_HAS_COPY_F64:
+	case SLJIT_HAS_ATOMIC:
+		return 1;
+
+	case SLJIT_HAS_CTZ:
+#if defined(SLJIT_CONFIG_ARM_V6) && SLJIT_CONFIG_ARM_V6
+		return 2;
+#else
+		return 1;
+#endif /* SLJIT_CONFIG_ARM_V6 */
+
+	default:
+		return 0;
+	}
+}
+
+/* --------------------------------------------------------------------- */
+/*  Entry, exit                                                          */
+/* --------------------------------------------------------------------- */
+
+/* Creates an index in data_transfer_insts array. */
+#define WORD_SIZE	0x00
+#define BYTE_SIZE	0x01
+#define HALF_SIZE	0x02
+#define PRELOAD		0x03
+#define SIGNED		0x04
+#define LOAD_DATA	0x08
+
+/* Flag bits for emit_op. */
+#define ALLOW_IMM		0x10
+#define ALLOW_INV_IMM		0x20
+#define ALLOW_ANY_IMM		(ALLOW_IMM | ALLOW_INV_IMM)
+#define ALLOW_NEG_IMM		0x40
+#define ALLOW_DOUBLE_IMM	0x80
+
+/* s/l - store/load (1 bit)
+   u/s - signed/unsigned (1 bit)
+   w/b/h/N - word/byte/half/NOT allowed (2 bit)
+   Storing signed and unsigned values are the same operations. */
+
+static const sljit_ins data_transfer_insts[16] = {
+/* s u w */ 0xe5000000 /* str */,
+/* s u b */ 0xe5400000 /* strb */,
+/* s u h */ 0xe10000b0 /* strh */,
+/* s u N */ 0x00000000 /* not allowed */,
+/* s s w */ 0xe5000000 /* str */,
+/* s s b */ 0xe5400000 /* strb */,
+/* s s h */ 0xe10000b0 /* strh */,
+/* s s N */ 0x00000000 /* not allowed */,
+
+/* l u w */ 0xe5100000 /* ldr */,
+/* l u b */ 0xe5500000 /* ldrb */,
+/* l u h */ 0xe11000b0 /* ldrh */,
+/* l u p */ 0xf5500000 /* preload */,
+/* l s w */ 0xe5100000 /* ldr */,
+/* l s b */ 0xe11000d0 /* ldrsb */,
+/* l s h */ 0xe11000f0 /* ldrsh */,
+/* l s N */ 0x00000000 /* not allowed */,
+};
+
+#define EMIT_DATA_TRANSFER(type, add, target_reg, base_reg, arg) \
+	(data_transfer_insts[(type) & 0xf] | ((add) << 23) | RD(target_reg) | RN(base_reg) | (sljit_ins)(arg))
+
+/* Normal ldr/str instruction.
+   Type2: ldrsb, ldrh, ldrsh */
+#define IS_TYPE1_TRANSFER(type) \
+	(data_transfer_insts[(type) & 0xf] & 0x04000000)
+#define TYPE2_TRANSFER_IMM(imm) \
+	(((imm) & 0xf) | (((imm) & 0xf0) << 4) | (1 << 22))
+
+#define EMIT_FPU_OPERATION(opcode, mode, dst, src1, src2) \
+	((sljit_ins)(opcode) | (sljit_ins)(mode) | VD(dst) | VM(src1) | VN(src2))
+
+/* Flags for emit_op: */
+  /* Arguments are swapped. */
+#define ARGS_SWAPPED	0x01
+  /* Inverted immediate. */
+#define INV_IMM		0x02
+  /* Source and destination is register. */
+#define REGISTER_OP	0x04
+  /* Unused return value. */
+#define UNUSED_RETURN	0x08
+/* SET_FLAGS must be (1 << 20) as it is also the value of S bit (can be used for optimization). */
+#define SET_FLAGS	(1 << 20)
+/* dst: reg
+   src1: reg
+   src2: reg or imm (if allowed)
+   SRC2_IMM must be (1 << 25) as it is also the value of I bit (can be used for optimization). */
+#define SRC2_IMM	(1 << 25)
+
+static sljit_s32 emit_op(struct sljit_compiler *compiler, sljit_s32 op, sljit_s32 inp_flags,
+	sljit_s32 dst, sljit_sw dstw,
+	sljit_s32 src1, sljit_sw src1w,
+	sljit_s32 src2, sljit_sw src2w);
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_enter(struct sljit_compiler *compiler,
+	sljit_s32 options, sljit_s32 arg_types, sljit_s32 scratches, sljit_s32 saveds,
+	sljit_s32 fscratches, sljit_s32 fsaveds, sljit_s32 local_size)
+{
+	sljit_uw imm, offset;
+	sljit_s32 i, tmp, size, word_arg_count;
+	sljit_s32 saved_arg_count = SLJIT_KEPT_SAVEDS_COUNT(options);
+#ifdef __SOFTFP__
+	sljit_u32 float_arg_count;
+#else
+	sljit_u32 old_offset, f32_offset;
+	sljit_u32 remap[3];
+	sljit_u32 *remap_ptr = remap;
+#endif
+
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_enter(compiler, options, arg_types, scratches, saveds, fscratches, fsaveds, local_size));
+	set_emit_enter(compiler, options, arg_types, scratches, saveds, fscratches, fsaveds, local_size);
+
+	imm = 0;
+
+	tmp = SLJIT_S0 - saveds;
+	for (i = SLJIT_S0 - saved_arg_count; i > tmp; i--)
+		imm |= (sljit_uw)1 << reg_map[i];
+
+	for (i = scratches; i >= SLJIT_FIRST_SAVED_REG; i--)
+		imm |= (sljit_uw)1 << reg_map[i];
+
+	SLJIT_ASSERT(reg_map[TMP_REG2] == 14);
+
+	/* Push saved and temporary registers
+	   multiple registers: stmdb sp!, {..., lr}
+	   single register: str reg, [sp, #-4]! */
+	if (imm != 0)
+		FAIL_IF(push_inst(compiler, PUSH | (1 << 14) | imm));
+	else
+		FAIL_IF(push_inst(compiler, 0xe52d0004 | RD(TMP_REG2)));
+
+	/* Stack must be aligned to 8 bytes: */
+	size = GET_SAVED_REGISTERS_SIZE(scratches, saveds - saved_arg_count, 1);
+
+	if (fsaveds > 0 || fscratches >= SLJIT_FIRST_SAVED_FLOAT_REG) {
+		if ((size & SSIZE_OF(sw)) != 0) {
+			FAIL_IF(push_inst(compiler, SUB | RD(SLJIT_SP) | RN(SLJIT_SP) | SRC2_IMM | sizeof(sljit_sw)));
+			size += SSIZE_OF(sw);
+		}
+
+		if (fsaveds + fscratches >= SLJIT_NUMBER_OF_FLOAT_REGISTERS) {
+			FAIL_IF(push_inst(compiler, VPUSH | VD(SLJIT_FS0) | ((sljit_ins)SLJIT_NUMBER_OF_SAVED_FLOAT_REGISTERS << 1)));
+		} else {
+			if (fsaveds > 0)
+				FAIL_IF(push_inst(compiler, VPUSH | VD(SLJIT_FS0) | ((sljit_ins)fsaveds << 1)));
+			if (fscratches >= SLJIT_FIRST_SAVED_FLOAT_REG)
+				FAIL_IF(push_inst(compiler, VPUSH | VD(fscratches) | ((sljit_ins)(fscratches - (SLJIT_FIRST_SAVED_FLOAT_REG - 1)) << 1)));
+		}
+	}
+
+	local_size = ((size + local_size + 0x7) & ~0x7) - size;
+	compiler->local_size = local_size;
+
+	if (options & SLJIT_ENTER_REG_ARG)
+		arg_types = 0;
+
+	arg_types >>= SLJIT_ARG_SHIFT;
+	word_arg_count = 0;
+	saved_arg_count = 0;
+#ifdef __SOFTFP__
+	SLJIT_COMPILE_ASSERT(SLJIT_FR0 == 1, float_register_index_start);
+
+	offset = 0;
+	float_arg_count = 0;
+
+	while (arg_types) {
+		switch (arg_types & SLJIT_ARG_MASK) {
+		case SLJIT_ARG_TYPE_F64:
+			if (offset & 0x7)
+				offset += sizeof(sljit_sw);
+
+			if (offset < 4 * sizeof(sljit_sw))
+				FAIL_IF(push_inst(compiler, VMOV2 | (offset << 10) | ((offset + sizeof(sljit_sw)) << 14) | float_arg_count));
+			else
+				FAIL_IF(push_inst(compiler, VLDR_F32 | 0x800100 | RN(SLJIT_SP)
+						| (float_arg_count << 12) | ((offset + (sljit_ins)size - 4 * sizeof(sljit_sw)) >> 2)));
+			float_arg_count++;
+			offset += sizeof(sljit_f64) - sizeof(sljit_sw);
+			break;
+		case SLJIT_ARG_TYPE_F32:
+			if (offset < 4 * sizeof(sljit_sw))
+				FAIL_IF(push_inst(compiler, VMOV | (float_arg_count << 16) | (offset << 10)));
+			else
+				FAIL_IF(push_inst(compiler, VLDR_F32 | 0x800000 | RN(SLJIT_SP)
+						| (float_arg_count << 12) | ((offset + (sljit_ins)size - 4 * sizeof(sljit_sw)) >> 2)));
+			float_arg_count++;
+			break;
+		default:
+			word_arg_count++;
+
+			if (!(arg_types & SLJIT_ARG_TYPE_SCRATCH_REG)) {
+				tmp = SLJIT_S0 - saved_arg_count;
+				saved_arg_count++;
+			} else if (word_arg_count - 1 != (sljit_s32)(offset >> 2))
+				tmp = word_arg_count;
+			else
+				break;
+
+			if (offset < 4 * sizeof(sljit_sw))
+				FAIL_IF(push_inst(compiler, MOV | RD(tmp) | (offset >> 2)));
+			else
+				FAIL_IF(push_inst(compiler, LDR | 0x800000 | RN(SLJIT_SP) | RD(tmp) | (offset + (sljit_ins)size - 4 * sizeof(sljit_sw))));
+			break;
+		}
+
+		offset += sizeof(sljit_sw);
+		arg_types >>= SLJIT_ARG_SHIFT;
+	}
+
+	compiler->args_size = offset;
+#else
+	offset = SLJIT_FR0;
+	old_offset = SLJIT_FR0;
+	f32_offset = 0;
+
+	while (arg_types) {
+		switch (arg_types & SLJIT_ARG_MASK) {
+		case SLJIT_ARG_TYPE_F64:
+			if (offset != old_offset)
+				*remap_ptr++ = EMIT_FPU_OPERATION(VMOV_F32, SLJIT_32, offset, old_offset, 0);
+			old_offset++;
+			offset++;
+			break;
+		case SLJIT_ARG_TYPE_F32:
+			if (f32_offset != 0) {
+				*remap_ptr++ = EMIT_FPU_OPERATION(VMOV_F32, 0x20, offset, f32_offset, 0);
+				f32_offset = 0;
+			} else {
+				if (offset != old_offset)
+					*remap_ptr++ = EMIT_FPU_OPERATION(VMOV_F32, 0, offset, old_offset, 0);
+				f32_offset = old_offset;
+				old_offset++;
+			}
+			offset++;
+			break;
+		default:
+			if (!(arg_types & SLJIT_ARG_TYPE_SCRATCH_REG)) {
+				FAIL_IF(push_inst(compiler, MOV | RD(SLJIT_S0 - saved_arg_count) | RM(SLJIT_R0 + word_arg_count)));
+				saved_arg_count++;
+			}
+
+			word_arg_count++;
+			break;
+		}
+		arg_types >>= SLJIT_ARG_SHIFT;
+	}
+
+	SLJIT_ASSERT((sljit_uw)(remap_ptr - remap) <= sizeof(remap));
+
+	while (remap_ptr > remap)
+		FAIL_IF(push_inst(compiler, *(--remap_ptr)));
+#endif
+
+	if (local_size > 0)
+		FAIL_IF(emit_op(compiler, SLJIT_SUB, ALLOW_IMM | ALLOW_DOUBLE_IMM, SLJIT_SP, 0, SLJIT_SP, 0, SLJIT_IMM, local_size));
+
+	return SLJIT_SUCCESS;
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_set_context(struct sljit_compiler *compiler,
+	sljit_s32 options, sljit_s32 arg_types, sljit_s32 scratches, sljit_s32 saveds,
+	sljit_s32 fscratches, sljit_s32 fsaveds, sljit_s32 local_size)
+{
+	sljit_s32 size;
+
+	CHECK_ERROR();
+	CHECK(check_sljit_set_context(compiler, options, arg_types, scratches, saveds, fscratches, fsaveds, local_size));
+	set_set_context(compiler, options, arg_types, scratches, saveds, fscratches, fsaveds, local_size);
+
+	size = GET_SAVED_REGISTERS_SIZE(scratches, saveds - SLJIT_KEPT_SAVEDS_COUNT(options), 1);
+
+	/* Doubles are saved, so alignment is unaffected. */
+	if ((size & SSIZE_OF(sw)) != 0 && (fsaveds > 0 || fscratches >= SLJIT_FIRST_SAVED_FLOAT_REG))
+		size += SSIZE_OF(sw);
+
+	compiler->local_size = ((size + local_size + 0x7) & ~0x7) - size;
+	return SLJIT_SUCCESS;
+}
+
+static sljit_s32 emit_add_sp(struct sljit_compiler *compiler, sljit_uw imm)
+{
+	sljit_uw imm2 = get_imm(imm);
+
+	if (imm2 == 0)
+		return emit_op(compiler, SLJIT_ADD, ALLOW_IMM | ALLOW_DOUBLE_IMM, SLJIT_SP, 0, SLJIT_SP, 0, SLJIT_IMM, (sljit_sw)imm);
+
+	return push_inst(compiler, ADD | RD(SLJIT_SP) | RN(SLJIT_SP) | imm2);
+}
+
+static sljit_s32 emit_stack_frame_release(struct sljit_compiler *compiler, sljit_s32 frame_size)
+{
+	sljit_s32 local_size, fscratches, fsaveds, i, tmp;
+	sljit_s32 restored_reg = 0;
+	sljit_s32 lr_dst = TMP_PC;
+	sljit_uw reg_list = 0;
+
+	SLJIT_ASSERT(reg_map[TMP_REG2] == 14 && frame_size <= 128);
+
+	local_size = compiler->local_size;
+	fscratches = compiler->fscratches;
+	fsaveds = compiler->fsaveds;
+
+	if (fsaveds > 0 || fscratches >= SLJIT_FIRST_SAVED_FLOAT_REG) {
+		if (local_size > 0)
+			FAIL_IF(emit_add_sp(compiler, (sljit_uw)local_size));
+
+		if (fsaveds + fscratches >= SLJIT_NUMBER_OF_FLOAT_REGISTERS) {
+			FAIL_IF(push_inst(compiler, VPOP | VD(SLJIT_FS0) | ((sljit_ins)SLJIT_NUMBER_OF_SAVED_FLOAT_REGISTERS << 1)));
+		} else {
+			if (fscratches >= SLJIT_FIRST_SAVED_FLOAT_REG)
+				FAIL_IF(push_inst(compiler, VPOP | VD(fscratches) | ((sljit_ins)(fscratches - (SLJIT_FIRST_SAVED_FLOAT_REG - 1)) << 1)));
+			if (fsaveds > 0)
+				FAIL_IF(push_inst(compiler, VPOP | VD(SLJIT_FS0) | ((sljit_ins)fsaveds << 1)));
+		}
+
+		local_size = GET_SAVED_REGISTERS_SIZE(compiler->scratches, compiler->saveds, 1) & 0x7;
+	}
+
+	if (frame_size < 0) {
+		lr_dst = TMP_REG2;
+		frame_size = 0;
+	} else if (frame_size > 0) {
+		SLJIT_ASSERT(frame_size == 1 || (frame_size & 0x7) == 0);
+		lr_dst = 0;
+		frame_size &= ~0x7;
+	}
+
+	if (lr_dst != 0)
+		reg_list |= (sljit_uw)1 << reg_map[lr_dst];
+
+	tmp = SLJIT_S0 - compiler->saveds;
+	i = SLJIT_S0 - SLJIT_KEPT_SAVEDS_COUNT(compiler->options);
+	if (tmp < i) {
+		restored_reg = i;
+		do {
+			reg_list |= (sljit_uw)1 << reg_map[i];
+		} while (--i > tmp);
+	}
+
+	i = compiler->scratches;
+	if (i >= SLJIT_FIRST_SAVED_REG) {
+		restored_reg = i;
+		do {
+			reg_list |= (sljit_uw)1 << reg_map[i];
+		} while (--i >= SLJIT_FIRST_SAVED_REG);
+	}
+
+	if (lr_dst == TMP_REG2 && reg_list == 0) {
+		restored_reg = TMP_REG2;
+		lr_dst = 0;
+	}
+
+	if (lr_dst == 0 && (reg_list & (reg_list - 1)) == 0) {
+		/* The local_size does not include the saved registers. */
+		tmp = 0;
+		if (reg_list != 0) {
+			tmp = 2;
+			if (local_size <= 0xfff) {
+				if (local_size == 0) {
+					SLJIT_ASSERT(restored_reg != TMP_REG2);
+					if (frame_size == 0)
+						return push_inst(compiler, LDR_POST | RN(SLJIT_SP) | RD(restored_reg) | 0x800008);
+					if (frame_size > 2 * SSIZE_OF(sw))
+						return push_inst(compiler, LDR_POST | RN(SLJIT_SP) | RD(restored_reg) | (sljit_ins)(frame_size - (2 * SSIZE_OF(sw))));
+				}
+
+				FAIL_IF(push_inst(compiler, LDR | 0x800000 | RN(SLJIT_SP) | RD(restored_reg) | (sljit_ins)local_size));
+				tmp = 1;
+			} else if (frame_size == 0) {
+				frame_size = (restored_reg == TMP_REG2) ? SSIZE_OF(sw) : 2 * SSIZE_OF(sw);
+				tmp = 3;
+			}
+
+			/* Place for the saved register. */
+			if (restored_reg != TMP_REG2)
+				local_size += SSIZE_OF(sw);
+		}
+
+		/* Place for the lr register. */
+		local_size += SSIZE_OF(sw);
+
+		if (frame_size > local_size)
+			FAIL_IF(push_inst(compiler, SUB | RD(SLJIT_SP) | RN(SLJIT_SP) | (1 << 25) | (sljit_ins)(frame_size - local_size)));
+		else if (frame_size < local_size)
+			FAIL_IF(emit_add_sp(compiler, (sljit_uw)(local_size - frame_size)));
+
+		if (tmp <= 1)
+			return SLJIT_SUCCESS;
+
+		if (tmp == 2) {
+			frame_size -= SSIZE_OF(sw);
+			if (restored_reg != TMP_REG2)
+				frame_size -= SSIZE_OF(sw);
+
+			return push_inst(compiler, LDR | 0x800000 | RN(SLJIT_SP) | RD(restored_reg) | (sljit_ins)frame_size);
+		}
+
+		tmp = (restored_reg == TMP_REG2) ? 0x800004 : 0x800008;
+		return push_inst(compiler, LDR_POST | RN(SLJIT_SP) | RD(restored_reg) | (sljit_ins)tmp);
+	}
+
+	if (local_size > 0)
+		FAIL_IF(emit_add_sp(compiler, (sljit_uw)local_size));
+
+	/* Pop saved and temporary registers
+	   multiple registers: ldmia sp!, {...}
+	   single register: ldr reg, [sp], #4 */
+	if ((reg_list & (reg_list - 1)) == 0) {
+		SLJIT_ASSERT(lr_dst != 0);
+		SLJIT_ASSERT(reg_list == (sljit_uw)1 << reg_map[lr_dst]);
+
+		return push_inst(compiler, LDR_POST | RN(SLJIT_SP) | RD(lr_dst) | 0x800004);
+	}
+
+	FAIL_IF(push_inst(compiler, POP | reg_list));
+
+	if (frame_size > 0)
+		return push_inst(compiler, SUB | RD(SLJIT_SP) | RN(SLJIT_SP) | (1 << 25) | ((sljit_ins)frame_size - sizeof(sljit_sw)));
+
+	if (lr_dst != 0)
+		return SLJIT_SUCCESS;
+
+	return push_inst(compiler, ADD | RD(SLJIT_SP) | RN(SLJIT_SP) | (1 << 25) | sizeof(sljit_sw));
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_return_void(struct sljit_compiler *compiler)
+{
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_return_void(compiler));
+
+	return emit_stack_frame_release(compiler, 0);
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_return_to(struct sljit_compiler *compiler,
+	sljit_s32 src, sljit_sw srcw)
+{
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_return_to(compiler, src, srcw));
+
+	if (src & SLJIT_MEM) {
+		FAIL_IF(emit_op_mem(compiler, WORD_SIZE | LOAD_DATA, TMP_REG1, src, srcw, TMP_REG1));
+		src = TMP_REG1;
+		srcw = 0;
+	} else if (src >= SLJIT_FIRST_SAVED_REG && src <= (SLJIT_S0 - SLJIT_KEPT_SAVEDS_COUNT(compiler->options))) {
+		FAIL_IF(push_inst(compiler, MOV | RD(TMP_REG1) | RM(src)));
+		src = TMP_REG1;
+		srcw = 0;
+	}
+
+	FAIL_IF(emit_stack_frame_release(compiler, 1));
+
+	SLJIT_SKIP_CHECKS(compiler);
+	return sljit_emit_ijump(compiler, SLJIT_JUMP, src, srcw);
+}
+
+/* --------------------------------------------------------------------- */
+/*  Operators                                                            */
+/* --------------------------------------------------------------------- */
+
+static SLJIT_INLINE sljit_s32 emit_single_op(struct sljit_compiler *compiler, sljit_s32 op, sljit_s32 flags,
+	sljit_uw dst, sljit_uw src1, sljit_uw src2)
+{
+	sljit_s32 reg, is_masked;
+	sljit_uw shift_type;
+
+	switch (op) {
+	case SLJIT_MOV:
+		SLJIT_ASSERT(src1 == TMP_REG1 && !(flags & ARGS_SWAPPED));
+		if (dst != src2) {
+			if (src2 & SRC2_IMM) {
+				return push_inst(compiler, ((flags & INV_IMM) ? MVN : MOV) | RD(dst) | src2);
+			}
+			return push_inst(compiler, MOV | RD(dst) | RM(src2));
+		}
+		return SLJIT_SUCCESS;
+
+	case SLJIT_MOV_U8:
+	case SLJIT_MOV_S8:
+		SLJIT_ASSERT(src1 == TMP_REG1 && !(flags & ARGS_SWAPPED));
+		if (flags & REGISTER_OP)
+			return push_inst(compiler, (op == SLJIT_MOV_U8 ? UXTB : SXTB) | RD(dst) | RM(src2));
+
+		if (dst != src2) {
+			SLJIT_ASSERT(src2 & SRC2_IMM);
+			return push_inst(compiler, ((flags & INV_IMM) ? MVN : MOV) | RD(dst) | src2);
+		}
+		return SLJIT_SUCCESS;
+
+	case SLJIT_MOV_U16:
+	case SLJIT_MOV_S16:
+		SLJIT_ASSERT(src1 == TMP_REG1 && !(flags & ARGS_SWAPPED));
+		if (flags & REGISTER_OP)
+			return push_inst(compiler, (op == SLJIT_MOV_U16 ? UXTH : SXTH) | RD(dst) | RM(src2));
+
+		if (dst != src2) {
+			SLJIT_ASSERT(src2 & SRC2_IMM);
+			return push_inst(compiler, ((flags & INV_IMM) ? MVN : MOV) | RD(dst) | src2);
+		}
+		return SLJIT_SUCCESS;
+
+	case SLJIT_CLZ:
+		SLJIT_ASSERT(!(flags & INV_IMM) && !(src2 & SRC2_IMM));
+		FAIL_IF(push_inst(compiler, CLZ | RD(dst) | RM(src2)));
+		return SLJIT_SUCCESS;
+
+	case SLJIT_CTZ:
+		SLJIT_ASSERT(!(flags & INV_IMM) && !(src2 & SRC2_IMM));
+		SLJIT_ASSERT(src1 == TMP_REG1 && src2 != TMP_REG2 && !(flags & ARGS_SWAPPED));
+#if (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+		FAIL_IF(push_inst(compiler, RSB | SRC2_IMM | RD(TMP_REG2) | RN(src2) | 0));
+		FAIL_IF(push_inst(compiler, AND | RD(TMP_REG1) | RN(src2) | RM(TMP_REG2)));
+		FAIL_IF(push_inst(compiler, CLZ | RD(dst) | RM(TMP_REG1)));
+		FAIL_IF(push_inst(compiler, CMP | SET_FLAGS | SRC2_IMM | RN(dst) | 32));
+		return push_inst(compiler, (EOR ^ 0xf0000000) | SRC2_IMM | RD(dst) | RN(dst) | 0x1f);
+#else /* !SLJIT_CONFIG_ARM_V6 */
+		FAIL_IF(push_inst(compiler, RBIT | RD(dst) | RM(src2)));
+		return push_inst(compiler, CLZ | RD(dst) | RM(dst));
+#endif /* SLJIT_CONFIG_ARM_V6 */
+
+	case SLJIT_REV:
+	case SLJIT_REV_U32:
+	case SLJIT_REV_S32:
+		SLJIT_ASSERT(src1 == TMP_REG1 && !(flags & ARGS_SWAPPED));
+		return push_inst(compiler, REV | RD(dst) | RM(src2));
+
+	case SLJIT_REV_U16:
+	case SLJIT_REV_S16:
+		SLJIT_ASSERT(src1 == TMP_REG1 && !(flags & ARGS_SWAPPED));
+		FAIL_IF(push_inst(compiler, REV16 | RD(dst) | RM(src2)));
+		if (!(flags & REGISTER_OP))
+			return SLJIT_SUCCESS;
+		return push_inst(compiler, (op == SLJIT_REV_U16 ? UXTH : SXTH) | RD(dst) | RM(dst));
+	case SLJIT_ADD:
+		SLJIT_ASSERT(!(flags & INV_IMM));
+
+		if ((flags & (UNUSED_RETURN | ARGS_SWAPPED)) == UNUSED_RETURN)
+			return push_inst(compiler, CMN | SET_FLAGS | RN(src1) | ((src2 & SRC2_IMM) ? src2 : RM(src2)));
+		return push_inst(compiler, ADD | (flags & SET_FLAGS) | RD(dst) | RN(src1) | ((src2 & SRC2_IMM) ? src2 : RM(src2)));
+
+	case SLJIT_ADDC:
+		SLJIT_ASSERT(!(flags & INV_IMM));
+		return push_inst(compiler, ADC | (flags & SET_FLAGS) | RD(dst) | RN(src1) | ((src2 & SRC2_IMM) ? src2 : RM(src2)));
+
+	case SLJIT_SUB:
+		SLJIT_ASSERT(!(flags & INV_IMM));
+
+		if ((flags & (UNUSED_RETURN | ARGS_SWAPPED)) == UNUSED_RETURN)
+			return push_inst(compiler, CMP | SET_FLAGS | RN(src1) | ((src2 & SRC2_IMM) ? src2 : RM(src2)));
+
+		return push_inst(compiler, (!(flags & ARGS_SWAPPED) ? SUB : RSB) | (flags & SET_FLAGS)
+			| RD(dst) | RN(src1) | ((src2 & SRC2_IMM) ? src2 : RM(src2)));
+
+	case SLJIT_SUBC:
+		SLJIT_ASSERT(!(flags & INV_IMM));
+		return push_inst(compiler, (!(flags & ARGS_SWAPPED) ? SBC : RSC) | (flags & SET_FLAGS)
+			| RD(dst) | RN(src1) | ((src2 & SRC2_IMM) ? src2 : RM(src2)));
+
+	case SLJIT_MUL:
+		SLJIT_ASSERT(!(flags & INV_IMM));
+		SLJIT_ASSERT(!(src2 & SRC2_IMM));
+		compiler->status_flags_state = 0;
+
+		if (!(flags & SET_FLAGS))
+			return push_inst(compiler, MUL | RN(dst) | RM8(src2) | RM(src1));
+
+		reg = dst == TMP_REG1 ? TMP_REG2 : TMP_REG1;
+		FAIL_IF(push_inst(compiler, SMULL | RN(reg) | RD(dst) | RM8(src2) | RM(src1)));
+
+		/* cmp TMP_REG1, dst asr #31. */
+		return push_inst(compiler, CMP | SET_FLAGS | RN(reg) | RM(dst) | 0xfc0);
+
+	case SLJIT_AND:
+		if ((flags & (UNUSED_RETURN | INV_IMM)) == UNUSED_RETURN)
+			return push_inst(compiler, TST | SET_FLAGS | RN(src1) | ((src2 & SRC2_IMM) ? src2 : RM(src2)));
+		return push_inst(compiler, (!(flags & INV_IMM) ? AND : BIC) | (flags & SET_FLAGS)
+			| RD(dst) | RN(src1) | ((src2 & SRC2_IMM) ? src2 : RM(src2)));
+
+	case SLJIT_OR:
+		SLJIT_ASSERT(!(flags & INV_IMM));
+		return push_inst(compiler, ORR | (flags & SET_FLAGS) | RD(dst) | RN(src1) | ((src2 & SRC2_IMM) ? src2 : RM(src2)));
+
+	case SLJIT_XOR:
+		if (flags & INV_IMM) {
+			SLJIT_ASSERT(src2 == SRC2_IMM);
+			return push_inst(compiler, MVN | (flags & SET_FLAGS) | RD(dst) | RM(src1));
+		}
+		return push_inst(compiler, EOR | (flags & SET_FLAGS) | RD(dst) | RN(src1) | ((src2 & SRC2_IMM) ? src2 : RM(src2)));
+
+	case SLJIT_SHL:
+	case SLJIT_MSHL:
+		shift_type = 0;
+		is_masked = op == SLJIT_MSHL;
+		break;
+
+	case SLJIT_LSHR:
+	case SLJIT_MLSHR:
+		shift_type = 1;
+		is_masked = op == SLJIT_MLSHR;
+		break;
+
+	case SLJIT_ASHR:
+	case SLJIT_MASHR:
+		shift_type = 2;
+		is_masked = op == SLJIT_MASHR;
+		break;
+
+	case SLJIT_ROTL:
+		if (compiler->shift_imm == 0x20) {
+			FAIL_IF(push_inst(compiler, RSB | SRC2_IMM | RD(TMP_REG2) | RN(src2) | 0));
+			src2 = TMP_REG2;
+		} else
+			compiler->shift_imm = (sljit_uw)(-(sljit_sw)compiler->shift_imm) & 0x1f;
+		/* fallthrough */
+
+	case SLJIT_ROTR:
+		shift_type = 3;
+		is_masked = 0;
+		break;
+
+	case SLJIT_MULADD:
+		return push_inst(compiler, MLA | RN(dst) | RD(dst) | RM8(src2) | RM(src1));
+
+	default:
+		SLJIT_UNREACHABLE();
+		return SLJIT_SUCCESS;
+	}
+
+	SLJIT_ASSERT(!(flags & ARGS_SWAPPED) && !(flags & INV_IMM) && !(src2 & SRC2_IMM));
+
+	if (compiler->shift_imm != 0x20) {
+		SLJIT_ASSERT(src1 == TMP_REG1);
+
+		if (compiler->shift_imm != 0)
+			return push_inst(compiler, MOV | (flags & SET_FLAGS) |
+				RD(dst) | (compiler->shift_imm << 7) | (shift_type << 5) | RM(src2));
+		return push_inst(compiler, MOV | (flags & SET_FLAGS) | RD(dst) | RM(src2));
+	}
+
+	SLJIT_ASSERT(src1 != TMP_REG2);
+
+	if (is_masked) {
+		FAIL_IF(push_inst(compiler, AND | RD(TMP_REG2) | RN(src2) | SRC2_IMM | 0x1f));
+		src2 = TMP_REG2;
+	}
+
+	return push_inst(compiler, MOV | (flags & SET_FLAGS) | RD(dst)
+		| RM8(src2) | (sljit_ins)(shift_type << 5) | 0x10 | RM(src1));
+}
+
+#undef EMIT_SHIFT_INS_AND_RETURN
+
+/* Tests whether the immediate can be stored in the 12 bit imm field.
+   Returns with 0 if not possible. */
+static sljit_uw get_imm(sljit_uw imm)
+{
+	sljit_u32 rol;
+
+	if (imm <= 0xff)
+		return SRC2_IMM | imm;
+
+	if (!(imm & 0xff000000)) {
+		imm <<= 8;
+		rol = 8;
+	} else {
+		imm = (imm << 24) | (imm >> 8);
+		rol = 0;
+	}
+
+	if (!(imm & 0xff000000)) {
+		imm <<= 8;
+		rol += 4;
+	}
+
+	if (!(imm & 0xf0000000)) {
+		imm <<= 4;
+		rol += 2;
+	}
+
+	if (!(imm & 0xc0000000)) {
+		imm <<= 2;
+		rol += 1;
+	}
+
+	if (!(imm & 0x00ffffff))
+		return SRC2_IMM | (imm >> 24) | (rol << 8);
+	return 0;
+}
+
+static sljit_uw compute_imm(sljit_uw imm, sljit_uw* imm2)
+{
+	sljit_uw mask;
+	sljit_uw imm1;
+	sljit_uw rol;
+
+	/* Step1: Search a zero byte (8 continous zero bit). */
+	mask = 0xff000000;
+	rol = 8;
+	while (1) {
+		if (!(imm & mask)) {
+			/* Rol imm by rol. */
+			imm = (imm << rol) | (imm >> (32 - rol));
+			/* Calculate arm rol. */
+			rol = 4 + (rol >> 1);
+			break;
+		}
+
+		rol += 2;
+		mask >>= 2;
+		if (mask & 0x3) {
+			/* rol by 8. */
+			imm = (imm << 8) | (imm >> 24);
+			mask = 0xff00;
+			rol = 24;
+			while (1) {
+				if (!(imm & mask)) {
+					/* Rol imm by rol. */
+					imm = (imm << rol) | (imm >> (32 - rol));
+					/* Calculate arm rol. */
+					rol = (rol >> 1) - 8;
+					break;
+				}
+				rol += 2;
+				mask >>= 2;
+				if (mask & 0x3)
+					return 0;
+			}
+			break;
+		}
+	}
+
+	/* The low 8 bit must be zero. */
+	SLJIT_ASSERT(!(imm & 0xff));
+
+	if (!(imm & 0xff000000)) {
+		imm1 = SRC2_IMM | ((imm >> 16) & 0xff) | (((rol + 4) & 0xf) << 8);
+		*imm2 = SRC2_IMM | ((imm >> 8) & 0xff) | (((rol + 8) & 0xf) << 8);
+	} else if (imm & 0xc0000000) {
+		imm1 = SRC2_IMM | ((imm >> 24) & 0xff) | ((rol & 0xf) << 8);
+		imm <<= 8;
+		rol += 4;
+
+		if (!(imm & 0xff000000)) {
+			imm <<= 8;
+			rol += 4;
+		}
+
+		if (!(imm & 0xf0000000)) {
+			imm <<= 4;
+			rol += 2;
+		}
+
+		if (!(imm & 0xc0000000)) {
+			imm <<= 2;
+			rol += 1;
+		}
+
+		if (!(imm & 0x00ffffff))
+			*imm2 = SRC2_IMM | (imm >> 24) | ((rol & 0xf) << 8);
+		else
+			return 0;
+	} else {
+		if (!(imm & 0xf0000000)) {
+			imm <<= 4;
+			rol += 2;
+		}
+
+		if (!(imm & 0xc0000000)) {
+			imm <<= 2;
+			rol += 1;
+		}
+
+		imm1 = SRC2_IMM | ((imm >> 24) & 0xff) | ((rol & 0xf) << 8);
+		imm <<= 8;
+		rol += 4;
+
+		if (!(imm & 0xf0000000)) {
+			imm <<= 4;
+			rol += 2;
+		}
+
+		if (!(imm & 0xc0000000)) {
+			imm <<= 2;
+			rol += 1;
+		}
+
+		if (!(imm & 0x00ffffff))
+			*imm2 = SRC2_IMM | (imm >> 24) | ((rol & 0xf) << 8);
+		else
+			return 0;
+	}
+
+	return imm1;
+}
+
+static sljit_s32 load_immediate(struct sljit_compiler *compiler, sljit_s32 reg, sljit_uw imm)
+{
+	sljit_uw tmp;
+#if (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+	sljit_uw imm1, imm2;
+#else /* !SLJIT_CONFIG_ARM_V6 */
+	if (!(imm & ~(sljit_uw)0xffff))
+		return push_inst(compiler, MOVW | RD(reg) | ((imm << 4) & 0xf0000) | (imm & 0xfff));
+#endif /* SLJIT_CONFIG_ARM_V6 */
+
+	/* Create imm by 1 inst. */
+	tmp = get_imm(imm);
+	if (tmp)
+		return push_inst(compiler, MOV | RD(reg) | tmp);
+
+	tmp = get_imm(~imm);
+	if (tmp)
+		return push_inst(compiler, MVN | RD(reg) | tmp);
+
+#if (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+	/* Create imm by 2 inst. */
+	imm1 = compute_imm(imm, &imm2);
+	if (imm1 != 0) {
+		FAIL_IF(push_inst(compiler, MOV | RD(reg) | imm1));
+		return push_inst(compiler, ORR | RD(reg) | RN(reg) | imm2);
+	}
+
+	imm1 = compute_imm(~imm, &imm2);
+	if (imm1 != 0) {
+		FAIL_IF(push_inst(compiler, MVN | RD(reg) | imm1));
+		return push_inst(compiler, BIC | RD(reg) | RN(reg) | imm2);
+	}
+
+	/* Load integer. */
+	return push_inst_with_literal(compiler, EMIT_DATA_TRANSFER(WORD_SIZE | LOAD_DATA, 1, reg, TMP_PC, 0), imm);
+#else /* !SLJIT_CONFIG_ARM_V6 */
+	FAIL_IF(push_inst(compiler, MOVW | RD(reg) | ((imm << 4) & 0xf0000) | (imm & 0xfff)));
+	if (imm <= 0xffff)
+		return SLJIT_SUCCESS;
+	return push_inst(compiler, MOVT | RD(reg) | ((imm >> 12) & 0xf0000) | ((imm >> 16) & 0xfff));
+#endif /* SLJIT_CONFIG_ARM_V6 */
+}
+
+static sljit_s32 emit_op_mem(struct sljit_compiler *compiler, sljit_s32 flags, sljit_s32 reg,
+	sljit_s32 arg, sljit_sw argw, sljit_s32 tmp_reg)
+{
+	sljit_uw imm, offset_reg, tmp;
+	sljit_sw mask = IS_TYPE1_TRANSFER(flags) ? 0xfff : 0xff;
+	sljit_sw sign = IS_TYPE1_TRANSFER(flags) ? 0x1000 : 0x100;
+
+	SLJIT_ASSERT(arg & SLJIT_MEM);
+	SLJIT_ASSERT((arg & REG_MASK) != tmp_reg || (arg == SLJIT_MEM1(tmp_reg) && argw >= -mask && argw <= mask));
+
+	if (SLJIT_UNLIKELY(!(arg & REG_MASK))) {
+		tmp = (sljit_uw)(argw & (sign | mask));
+		tmp = (sljit_uw)((argw + (tmp <= (sljit_uw)sign ? 0 : sign)) & ~mask);
+
+		FAIL_IF(load_immediate(compiler, tmp_reg, tmp));
+
+		argw -= (sljit_sw)tmp;
+		tmp = 1;
+
+		if (argw < 0) {
+			argw = -argw;
+			tmp = 0;
+		}
+
+		return push_inst(compiler, EMIT_DATA_TRANSFER(flags, tmp, reg, tmp_reg,
+			(mask == 0xff) ? TYPE2_TRANSFER_IMM(argw) : argw));
+	}
+
+	if (arg & OFFS_REG_MASK) {
+		offset_reg = OFFS_REG(arg);
+		arg &= REG_MASK;
+		argw &= 0x3;
+
+		if (argw != 0 && (mask == 0xff)) {
+			FAIL_IF(push_inst(compiler, ADD | RD(tmp_reg) | RN(arg) | RM(offset_reg) | ((sljit_ins)argw << 7)));
+			return push_inst(compiler, EMIT_DATA_TRANSFER(flags, 1, reg, tmp_reg, TYPE2_TRANSFER_IMM(0)));
+		}
+
+		/* Bit 25: RM is offset. */
+		return push_inst(compiler, EMIT_DATA_TRANSFER(flags, 1, reg, arg,
+			RM(offset_reg) | (mask == 0xff ? 0 : (1 << 25)) | ((sljit_ins)argw << 7)));
+	}
+
+	arg &= REG_MASK;
+
+	if (argw > mask) {
+		tmp = (sljit_uw)(argw & (sign | mask));
+		tmp = (sljit_uw)((argw + (tmp <= (sljit_uw)sign ? 0 : sign)) & ~mask);
+		imm = get_imm(tmp);
+
+		if (imm) {
+			FAIL_IF(push_inst(compiler, ADD | RD(tmp_reg) | RN(arg) | imm));
+			argw -= (sljit_sw)tmp;
+			arg = tmp_reg;
+
+			SLJIT_ASSERT(argw >= -mask && argw <= mask);
+		}
+	} else if (argw < -mask) {
+		tmp = (sljit_uw)(-argw & (sign | mask));
+		tmp = (sljit_uw)((-argw + (tmp <= (sljit_uw)sign ? 0 : sign)) & ~mask);
+		imm = get_imm(tmp);
+
+		if (imm) {
+			FAIL_IF(push_inst(compiler, SUB | RD(tmp_reg) | RN(arg) | imm));
+			argw += (sljit_sw)tmp;
+			arg = tmp_reg;
+
+			SLJIT_ASSERT(argw >= -mask && argw <= mask);
+		}
+	}
+
+	if (argw <= mask && argw >= -mask) {
+		if (argw >= 0) {
+			if (mask == 0xff)
+				argw = TYPE2_TRANSFER_IMM(argw);
+			return push_inst(compiler, EMIT_DATA_TRANSFER(flags, 1, reg, arg, argw));
+		}
+
+		argw = -argw;
+
+		if (mask == 0xff)
+			argw = TYPE2_TRANSFER_IMM(argw);
+
+		return push_inst(compiler, EMIT_DATA_TRANSFER(flags, 0, reg, arg, argw));
+	}
+
+	FAIL_IF(load_immediate(compiler, tmp_reg, (sljit_uw)argw));
+	return push_inst(compiler, EMIT_DATA_TRANSFER(flags, 1, reg, arg,
+		RM(tmp_reg) | (mask == 0xff ? 0 : (1 << 25))));
+}
+
+static sljit_s32 emit_op(struct sljit_compiler *compiler, sljit_s32 op, sljit_s32 inp_flags,
+	sljit_s32 dst, sljit_sw dstw,
+	sljit_s32 src1, sljit_sw src1w,
+	sljit_s32 src2, sljit_sw src2w)
+{
+	/* src1 is reg or TMP_REG1
+	   src2 is reg, TMP_REG2, or imm
+	   result goes to TMP_REG2, so put result can use TMP_REG1. */
+
+	/* We prefers register and simple consts. */
+	sljit_s32 dst_reg;
+	sljit_s32 src1_reg = 0;
+	sljit_s32 src2_reg = 0;
+	sljit_s32 src2_tmp_reg = 0;
+	sljit_s32 flags = HAS_FLAGS(op) ? SET_FLAGS : 0;
+	sljit_s32 neg_op = 0;
+	sljit_u32 imm2;
+
+	op = GET_OPCODE(op);
+
+	if (flags & SET_FLAGS)
+		inp_flags &= ~ALLOW_DOUBLE_IMM;
+
+	if (dst == TMP_REG1)
+		flags |= UNUSED_RETURN;
+
+	SLJIT_ASSERT(!(inp_flags & ALLOW_INV_IMM) || (inp_flags & ALLOW_IMM));
+
+	if (inp_flags & ALLOW_NEG_IMM) {
+		switch (op) {
+		case SLJIT_ADD:
+			compiler->status_flags_state = SLJIT_CURRENT_FLAGS_ADD;
+			neg_op = SLJIT_SUB;
+			break;
+		case SLJIT_ADDC:
+			compiler->status_flags_state = SLJIT_CURRENT_FLAGS_ADD;
+			neg_op = SLJIT_SUBC;
+			break;
+		case SLJIT_SUB:
+			compiler->status_flags_state = SLJIT_CURRENT_FLAGS_SUB;
+			neg_op = SLJIT_ADD;
+			break;
+		case SLJIT_SUBC:
+			compiler->status_flags_state = SLJIT_CURRENT_FLAGS_SUB;
+			neg_op = SLJIT_ADDC;
+			break;
+		}
+	}
+
+	do {
+		if (!(inp_flags & ALLOW_IMM))
+			break;
+
+		if (src2 == SLJIT_IMM) {
+			src2_reg = (sljit_s32)get_imm((sljit_uw)src2w);
+			if (src2_reg)
+				break;
+
+			if (inp_flags & ALLOW_INV_IMM) {
+				src2_reg = (sljit_s32)get_imm(~(sljit_uw)src2w);
+				if (src2_reg) {
+					flags |= INV_IMM;
+					break;
+				}
+			}
+
+			if (neg_op != 0) {
+				src2_reg = (sljit_s32)get_imm((neg_op == SLJIT_ADD || neg_op == SLJIT_SUB) ? (sljit_uw)-src2w : ~(sljit_uw)src2w);
+				if (src2_reg) {
+					op = neg_op | GET_ALL_FLAGS(op);
+					break;
+				}
+			}
+		}
+
+		if (src1 == SLJIT_IMM) {
+			src2_reg = (sljit_s32)get_imm((sljit_uw)src1w);
+			if (src2_reg) {
+				flags |= ARGS_SWAPPED;
+				src1 = src2;
+				src1w = src2w;
+				break;
+			}
+
+			if (inp_flags & ALLOW_INV_IMM) {
+				src2_reg = (sljit_s32)get_imm(~(sljit_uw)src1w);
+				if (src2_reg) {
+					flags |= ARGS_SWAPPED | INV_IMM;
+					src1 = src2;
+					src1w = src2w;
+					break;
+				}
+			}
+
+			if (neg_op >= SLJIT_SUB) {
+				/* Note: additive operation (commutative). */
+				SLJIT_ASSERT(op == SLJIT_ADD || op == SLJIT_ADDC);
+
+				src2_reg = (sljit_s32)get_imm((sljit_uw)-src1w);
+				if (src2_reg) {
+					src1 = src2;
+					src1w = src2w;
+					op = neg_op | GET_ALL_FLAGS(op);
+					break;
+				}
+			}
+		}
+	} while(0);
+
+	/* Destination. */
+	dst_reg = FAST_IS_REG(dst) ? dst : TMP_REG2;
+
+	if (op <= SLJIT_MOV_P) {
+		if (dst & SLJIT_MEM) {
+			if (inp_flags & BYTE_SIZE)
+				inp_flags &= ~SIGNED;
+
+			if (FAST_IS_REG(src2))
+				return emit_op_mem(compiler, inp_flags, src2, dst, dstw, TMP_REG1);
+		}
+
+		if (FAST_IS_REG(src2) && dst_reg != TMP_REG2)
+			flags |= REGISTER_OP;
+
+		src2_tmp_reg = dst_reg;
+	} else {
+		if (op == SLJIT_REV_U16 || op == SLJIT_REV_S16) {
+			if (!(dst & SLJIT_MEM) && (!(src2 & SLJIT_MEM) || op == SLJIT_REV_S16))
+				flags |= REGISTER_OP;
+		}
+
+		src2_tmp_reg = FAST_IS_REG(src1) ? TMP_REG1 : TMP_REG2;
+	}
+
+	if (src2_reg == 0 && (src2 & SLJIT_MEM)) {
+		src2_reg = src2_tmp_reg;
+		FAIL_IF(emit_op_mem(compiler, inp_flags | LOAD_DATA, src2_reg, src2, src2w, TMP_REG1));
+	}
+
+	/* Source 1. */
+	if (FAST_IS_REG(src1))
+		src1_reg = src1;
+	else if (src1 & SLJIT_MEM) {
+		FAIL_IF(emit_op_mem(compiler, inp_flags | LOAD_DATA, TMP_REG1, src1, src1w, TMP_REG1));
+		src1_reg = TMP_REG1;
+	} else if (!(inp_flags & ALLOW_DOUBLE_IMM) || src2_reg != 0 || op == SLJIT_SUB || op == SLJIT_SUBC) {
+		FAIL_IF(load_immediate(compiler, TMP_REG1, (sljit_uw)src1w));
+		src1_reg = TMP_REG1;
+	}
+
+	/* Source 2. */
+	if (src2_reg == 0) {
+		src2_reg = src2_tmp_reg;
+
+		if (FAST_IS_REG(src2))
+			src2_reg = src2;
+		else if (!(inp_flags & ALLOW_DOUBLE_IMM))
+			FAIL_IF(load_immediate(compiler, src2_reg, (sljit_uw)src2w));
+		else {
+			SLJIT_ASSERT(!(flags & SET_FLAGS));
+
+			if (src1_reg == 0) {
+				FAIL_IF(load_immediate(compiler, TMP_REG1, (sljit_uw)src1w));
+				src1_reg = TMP_REG1;
+			}
+
+			src2_reg = (sljit_s32)compute_imm((sljit_uw)src2w, &imm2);
+
+			if (src2_reg == 0 && neg_op != 0) {
+				src2_reg = (sljit_s32)compute_imm((sljit_uw)-src2w, &imm2);
+				if (src2_reg != 0)
+					op = neg_op;
+			}
+
+			if (src2_reg == 0) {
+				FAIL_IF(load_immediate(compiler, src2_tmp_reg, (sljit_uw)src2w));
+				src2_reg = src2_tmp_reg;
+			} else {
+				FAIL_IF(emit_single_op(compiler, op, flags, (sljit_uw)dst_reg, (sljit_uw)src1_reg, (sljit_uw)src2_reg));
+				src1_reg = dst_reg;
+				src2_reg = (sljit_s32)imm2;
+
+				if (op == SLJIT_ADDC)
+					op = SLJIT_ADD;
+				else if (op == SLJIT_SUBC)
+					op = SLJIT_SUB;
+			}
+		}
+	}
+
+	if (src1_reg == 0) {
+		SLJIT_ASSERT((inp_flags & ALLOW_DOUBLE_IMM) && !(flags & SET_FLAGS));
+
+		src1_reg = (sljit_s32)compute_imm((sljit_uw)src1w, &imm2);
+
+		if (src1_reg == 0 && neg_op != 0) {
+			src1_reg = (sljit_s32)compute_imm((sljit_uw)-src1w, &imm2);
+			if (src1_reg != 0)
+				op = neg_op;
+		}
+
+		if (src1_reg == 0) {
+			FAIL_IF(load_immediate(compiler, TMP_REG1, (sljit_uw)src1w));
+			src1_reg = TMP_REG1;
+		} else {
+			FAIL_IF(emit_single_op(compiler, op, flags, (sljit_uw)dst_reg, (sljit_uw)src2_reg, (sljit_uw)src1_reg));
+			src1_reg = dst_reg;
+			src2_reg = (sljit_s32)imm2;
+
+			if (op == SLJIT_ADDC)
+				op = SLJIT_ADD;
+		}
+	}
+
+	FAIL_IF(emit_single_op(compiler, op, flags, (sljit_uw)dst_reg, (sljit_uw)src1_reg, (sljit_uw)src2_reg));
+
+	if (!(dst & SLJIT_MEM))
+		return SLJIT_SUCCESS;
+
+	return emit_op_mem(compiler, inp_flags, dst_reg, dst, dstw, TMP_REG1);
+}
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#if defined(__GNUC__)
+extern unsigned int __aeabi_uidivmod(unsigned int numerator, unsigned int denominator);
+extern int __aeabi_idivmod(int numerator, int denominator);
+#else
+#error "Software divmod functions are needed"
+#endif
+
+#ifdef __cplusplus
+}
+#endif
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_op0(struct sljit_compiler *compiler, sljit_s32 op)
+{
+	sljit_uw saved_reg_list[3];
+	sljit_sw saved_reg_count;
+
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_op0(compiler, op));
+
+	op = GET_OPCODE(op);
+	switch (op) {
+	case SLJIT_BREAKPOINT:
+		FAIL_IF(push_inst(compiler, BKPT));
+		break;
+	case SLJIT_NOP:
+		FAIL_IF(push_inst(compiler, NOP));
+		break;
+	case SLJIT_LMUL_UW:
+	case SLJIT_LMUL_SW:
+		return push_inst(compiler, (op == SLJIT_LMUL_UW ? UMULL : SMULL)
+			| RN(SLJIT_R1) | RD(SLJIT_R0) | RM8(SLJIT_R0) | RM(SLJIT_R1));
+	case SLJIT_DIVMOD_UW:
+	case SLJIT_DIVMOD_SW:
+	case SLJIT_DIV_UW:
+	case SLJIT_DIV_SW:
+		SLJIT_COMPILE_ASSERT((SLJIT_DIVMOD_UW & 0x2) == 0 && SLJIT_DIV_UW - 0x2 == SLJIT_DIVMOD_UW, bad_div_opcode_assignments);
+		SLJIT_ASSERT(reg_map[2] == 1 && reg_map[3] == 2 && reg_map[4] == 3);
+
+		saved_reg_count = 0;
+		if (compiler->scratches >= 4)
+			saved_reg_list[saved_reg_count++] = 3;
+		if (compiler->scratches >= 3)
+			saved_reg_list[saved_reg_count++] = 2;
+		if (op >= SLJIT_DIV_UW)
+			saved_reg_list[saved_reg_count++] = 1;
+
+		if (saved_reg_count > 0) {
+			FAIL_IF(push_inst(compiler, STR | 0x2d0000 | (saved_reg_count >= 3 ? 16 : 8)
+						| (saved_reg_list[0] << 12) /* str rX, [sp, #-8/-16]! */));
+			if (saved_reg_count >= 2) {
+				SLJIT_ASSERT(saved_reg_list[1] < 8);
+				FAIL_IF(push_inst(compiler, STR | 0x8d0004 | (saved_reg_list[1] << 12) /* str rX, [sp, #4] */));
+			}
+			if (saved_reg_count >= 3) {
+				SLJIT_ASSERT(saved_reg_list[2] < 8);
+				FAIL_IF(push_inst(compiler, STR | 0x8d0008 | (saved_reg_list[2] << 12) /* str rX, [sp, #8] */));
+			}
+		}
+
+#if defined(__GNUC__)
+		FAIL_IF(sljit_emit_ijump(compiler, SLJIT_FAST_CALL, SLJIT_IMM,
+			((op | 0x2) == SLJIT_DIV_UW ? SLJIT_FUNC_ADDR(__aeabi_uidivmod) : SLJIT_FUNC_ADDR(__aeabi_idivmod))));
+#else
+#error "Software divmod functions are needed"
+#endif
+
+		if (saved_reg_count > 0) {
+			if (saved_reg_count >= 3) {
+				SLJIT_ASSERT(saved_reg_list[2] < 8);
+				FAIL_IF(push_inst(compiler, LDR | 0x8d0008 | (saved_reg_list[2] << 12) /* ldr rX, [sp, #8] */));
+			}
+			if (saved_reg_count >= 2) {
+				SLJIT_ASSERT(saved_reg_list[1] < 8);
+				FAIL_IF(push_inst(compiler, LDR | 0x8d0004 | (saved_reg_list[1] << 12) /* ldr rX, [sp, #4] */));
+			}
+			return push_inst(compiler, (LDR ^ (1 << 24)) | 0x8d0000 | (sljit_ins)(saved_reg_count >= 3 ? 16 : 8)
+						| (saved_reg_list[0] << 12) /* ldr rX, [sp], #8/16 */);
+		}
+		return SLJIT_SUCCESS;
+	case SLJIT_ENDBR:
+	case SLJIT_SKIP_FRAMES_BEFORE_RETURN:
+		return SLJIT_SUCCESS;
+	}
+
+	return SLJIT_SUCCESS;
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_op1(struct sljit_compiler *compiler, sljit_s32 op,
+	sljit_s32 dst, sljit_sw dstw,
+	sljit_s32 src, sljit_sw srcw)
+{
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_op1(compiler, op, dst, dstw, src, srcw));
+	ADJUST_LOCAL_OFFSET(dst, dstw);
+	ADJUST_LOCAL_OFFSET(src, srcw);
+
+	switch (GET_OPCODE(op)) {
+	case SLJIT_MOV:
+	case SLJIT_MOV_U32:
+	case SLJIT_MOV_S32:
+	case SLJIT_MOV32:
+	case SLJIT_MOV_P:
+		return emit_op(compiler, SLJIT_MOV, ALLOW_ANY_IMM, dst, dstw, TMP_REG1, 0, src, srcw);
+
+	case SLJIT_MOV_U8:
+		return emit_op(compiler, SLJIT_MOV_U8, ALLOW_ANY_IMM | BYTE_SIZE, dst, dstw, TMP_REG1, 0, src, (src == SLJIT_IMM) ? (sljit_u8)srcw : srcw);
+
+	case SLJIT_MOV_S8:
+		return emit_op(compiler, SLJIT_MOV_S8, ALLOW_ANY_IMM | SIGNED | BYTE_SIZE, dst, dstw, TMP_REG1, 0, src, (src == SLJIT_IMM) ? (sljit_s8)srcw : srcw);
+
+	case SLJIT_MOV_U16:
+		return emit_op(compiler, SLJIT_MOV_U16, ALLOW_ANY_IMM | HALF_SIZE, dst, dstw, TMP_REG1, 0, src, (src == SLJIT_IMM) ? (sljit_u16)srcw : srcw);
+
+	case SLJIT_MOV_S16:
+		return emit_op(compiler, SLJIT_MOV_S16, ALLOW_ANY_IMM | SIGNED | HALF_SIZE, dst, dstw, TMP_REG1, 0, src, (src == SLJIT_IMM) ? (sljit_s16)srcw : srcw);
+
+	case SLJIT_CLZ:
+	case SLJIT_CTZ:
+	case SLJIT_REV:
+	case SLJIT_REV_U32:
+	case SLJIT_REV_S32:
+		return emit_op(compiler, op, 0, dst, dstw, TMP_REG1, 0, src, srcw);
+
+	case SLJIT_REV_U16:
+	case SLJIT_REV_S16:
+		return emit_op(compiler, op, HALF_SIZE, dst, dstw, TMP_REG1, 0, src, srcw);
+	}
+
+	return SLJIT_SUCCESS;
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_op2(struct sljit_compiler *compiler, sljit_s32 op,
+	sljit_s32 dst, sljit_sw dstw,
+	sljit_s32 src1, sljit_sw src1w,
+	sljit_s32 src2, sljit_sw src2w)
+{
+	sljit_s32 inp_flags;
+
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_op2(compiler, op, 0, dst, dstw, src1, src1w, src2, src2w));
+	ADJUST_LOCAL_OFFSET(dst, dstw);
+	ADJUST_LOCAL_OFFSET(src1, src1w);
+	ADJUST_LOCAL_OFFSET(src2, src2w);
+
+	switch (GET_OPCODE(op)) {
+	case SLJIT_ADD:
+	case SLJIT_ADDC:
+	case SLJIT_SUB:
+	case SLJIT_SUBC:
+		return emit_op(compiler, op, ALLOW_IMM | ALLOW_NEG_IMM | ALLOW_DOUBLE_IMM, dst, dstw, src1, src1w, src2, src2w);
+
+	case SLJIT_OR:
+		return emit_op(compiler, op, ALLOW_IMM | ALLOW_DOUBLE_IMM, dst, dstw, src1, src1w, src2, src2w);
+
+	case SLJIT_XOR:
+		inp_flags = ALLOW_IMM | ALLOW_DOUBLE_IMM;
+		if ((src1 == SLJIT_IMM && src1w == -1) || (src2 == SLJIT_IMM && src2w == -1)) {
+			inp_flags |= ALLOW_INV_IMM;
+		}
+		return emit_op(compiler, op, inp_flags, dst, dstw, src1, src1w, src2, src2w);
+
+	case SLJIT_MUL:
+		return emit_op(compiler, op, 0, dst, dstw, src1, src1w, src2, src2w);
+
+	case SLJIT_AND:
+		return emit_op(compiler, op, ALLOW_ANY_IMM, dst, dstw, src1, src1w, src2, src2w);
+
+	case SLJIT_SHL:
+	case SLJIT_MSHL:
+	case SLJIT_LSHR:
+	case SLJIT_MLSHR:
+	case SLJIT_ASHR:
+	case SLJIT_MASHR:
+	case SLJIT_ROTL:
+	case SLJIT_ROTR:
+		if (src2 == SLJIT_IMM) {
+			compiler->shift_imm = src2w & 0x1f;
+			return emit_op(compiler, op, 0, dst, dstw, TMP_REG1, 0, src1, src1w);
+		} else {
+			compiler->shift_imm = 0x20;
+			return emit_op(compiler, op, 0, dst, dstw, src1, src1w, src2, src2w);
+		}
+	}
+
+	return SLJIT_SUCCESS;
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_op2u(struct sljit_compiler *compiler, sljit_s32 op,
+	sljit_s32 src1, sljit_sw src1w,
+	sljit_s32 src2, sljit_sw src2w)
+{
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_op2(compiler, op, 1, 0, 0, src1, src1w, src2, src2w));
+
+	SLJIT_SKIP_CHECKS(compiler);
+	return sljit_emit_op2(compiler, op, TMP_REG1, 0, src1, src1w, src2, src2w);
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_op2r(struct sljit_compiler *compiler, sljit_s32 op,
+	sljit_s32 dst_reg,
+	sljit_s32 src1, sljit_sw src1w,
+	sljit_s32 src2, sljit_sw src2w)
+{
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_op2r(compiler, op, dst_reg, src1, src1w, src2, src2w));
+	ADJUST_LOCAL_OFFSET(src1, src1w);
+	ADJUST_LOCAL_OFFSET(src2, src2w);
+
+	switch (GET_OPCODE(op)) {
+	case SLJIT_MULADD:
+		return emit_op(compiler, op, 0, dst_reg, 0, src1, src1w, src2, src2w);
+	}
+
+	return SLJIT_SUCCESS;
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_shift_into(struct sljit_compiler *compiler, sljit_s32 op,
+	sljit_s32 dst_reg,
+	sljit_s32 src1_reg,
+	sljit_s32 src2_reg,
+	sljit_s32 src3, sljit_sw src3w)
+{
+	sljit_s32 is_left;
+
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_shift_into(compiler, op, dst_reg, src1_reg, src2_reg, src3, src3w));
+
+	op = GET_OPCODE(op);
+	is_left = (op == SLJIT_SHL || op == SLJIT_MSHL);
+
+	if (src1_reg == src2_reg) {
+		SLJIT_SKIP_CHECKS(compiler);
+		return sljit_emit_op2(compiler, is_left ? SLJIT_ROTL : SLJIT_ROTR, dst_reg, 0, src1_reg, 0, src3, src3w);
+	}
+
+	ADJUST_LOCAL_OFFSET(src3, src3w);
+
+	/* Shift type of ROR is 3. */
+	if (src3 == SLJIT_IMM) {
+		src3w &= 0x1f;
+
+		if (src3w == 0)
+			return SLJIT_SUCCESS;
+
+		FAIL_IF(push_inst(compiler, MOV | RD(dst_reg) | RM(src1_reg) | ((sljit_ins)(is_left ? 0 : 1) << 5) | ((sljit_ins)src3w << 7)));
+		src3w = (src3w ^ 0x1f) + 1;
+		return push_inst(compiler, ORR | RD(dst_reg) | RN(dst_reg) | RM(src2_reg) | ((sljit_ins)(is_left ? 1 : 0) << 5) | ((sljit_ins)src3w << 7));
+	}
+
+	if (src3 & SLJIT_MEM) {
+		FAIL_IF(emit_op_mem(compiler, WORD_SIZE | LOAD_DATA, TMP_REG2, src3, src3w, TMP_REG2));
+		src3 = TMP_REG2;
+	}
+
+	if (op == SLJIT_MSHL || op == SLJIT_MLSHR || dst_reg == src3) {
+		FAIL_IF(push_inst(compiler, AND | SRC2_IMM | RD(TMP_REG2) | RN(src3) | 0x1f));
+		src3 = TMP_REG2;
+	}
+
+	FAIL_IF(push_inst(compiler, MOV | RD(dst_reg) | RM8(src3) | ((sljit_ins)(is_left ? 0 : 1) << 5) | 0x10 | RM(src1_reg)));
+	FAIL_IF(push_inst(compiler, MOV | RD(TMP_REG1) | RM(src2_reg) | ((sljit_ins)(is_left ? 1 : 0) << 5) | (1 << 7)));
+	FAIL_IF(push_inst(compiler, EOR | SRC2_IMM | RD(TMP_REG2) | RN(src3) | 0x1f));
+	return push_inst(compiler, ORR | RD(dst_reg) | RN(dst_reg) | RM8(TMP_REG2) | ((sljit_ins)(is_left ? 1 : 0) << 5) | 0x10 | RM(TMP_REG1));
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_op_src(struct sljit_compiler *compiler, sljit_s32 op,
+	sljit_s32 src, sljit_sw srcw)
+{
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_op_src(compiler, op, src, srcw));
+	ADJUST_LOCAL_OFFSET(src, srcw);
+
+	switch (op) {
+	case SLJIT_FAST_RETURN:
+		SLJIT_ASSERT(reg_map[TMP_REG2] == 14);
+
+		if (FAST_IS_REG(src))
+			FAIL_IF(push_inst(compiler, MOV | RD(TMP_REG2) | RM(src)));
+		else
+			FAIL_IF(emit_op_mem(compiler, WORD_SIZE | LOAD_DATA, TMP_REG2, src, srcw, TMP_REG1));
+
+		return push_inst(compiler, BX | RM(TMP_REG2));
+	case SLJIT_SKIP_FRAMES_BEFORE_FAST_RETURN:
+		return SLJIT_SUCCESS;
+	case SLJIT_PREFETCH_L1:
+	case SLJIT_PREFETCH_L2:
+	case SLJIT_PREFETCH_L3:
+	case SLJIT_PREFETCH_ONCE:
+		SLJIT_ASSERT(src & SLJIT_MEM);
+		return emit_op_mem(compiler, PRELOAD | LOAD_DATA, TMP_PC, src, srcw, TMP_REG1);
+	}
+
+	return SLJIT_SUCCESS;
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_op_dst(struct sljit_compiler *compiler, sljit_s32 op,
+	sljit_s32 dst, sljit_sw dstw)
+{
+	sljit_s32 size, dst_r;
+
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_op_dst(compiler, op, dst, dstw));
+	ADJUST_LOCAL_OFFSET(dst, dstw);
+
+	switch (op) {
+	case SLJIT_FAST_ENTER:
+		SLJIT_ASSERT(reg_map[TMP_REG2] == 14);
+
+		if (FAST_IS_REG(dst))
+			return push_inst(compiler, MOV | RD(dst) | RM(TMP_REG2));
+		break;
+	case SLJIT_GET_RETURN_ADDRESS:
+		size = GET_SAVED_REGISTERS_SIZE(compiler->scratches, compiler->saveds - SLJIT_KEPT_SAVEDS_COUNT(compiler->options), 0);
+
+		if (compiler->fsaveds > 0 || compiler->fscratches >= SLJIT_FIRST_SAVED_FLOAT_REG) {
+			/* The size of pc is not added above. */
+			if ((size & SSIZE_OF(sw)) == 0)
+				size += SSIZE_OF(sw);
+
+			size += GET_SAVED_FLOAT_REGISTERS_SIZE(compiler->fscratches, compiler->fsaveds, f64);
+		}
+
+		SLJIT_ASSERT(((compiler->local_size + size + SSIZE_OF(sw)) & 0x7) == 0);
+
+		dst_r = FAST_IS_REG(dst) ? dst : TMP_REG2;
+		FAIL_IF(emit_op_mem(compiler, WORD_SIZE | LOAD_DATA, dst_r, SLJIT_MEM1(SLJIT_SP), compiler->local_size + size, TMP_REG1));
+		break;
+	}
+
+	if (dst & SLJIT_MEM)
+		return emit_op_mem(compiler, WORD_SIZE, TMP_REG2, dst, dstw, TMP_REG1);
+
+	return SLJIT_SUCCESS;
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_get_register_index(sljit_s32 type, sljit_s32 reg)
+{
+	CHECK_REG_INDEX(check_sljit_get_register_index(type, reg));
+
+	if (type == SLJIT_GP_REGISTER)
+		return reg_map[reg];
+
+	if (type == SLJIT_FLOAT_REGISTER || type == SLJIT_SIMD_REG_64)
+		return freg_map[reg];
+
+	if (type != SLJIT_SIMD_REG_128)
+		return freg_map[reg] & ~0x1;
+
+	return -1;
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_op_custom(struct sljit_compiler *compiler,
+	void *instruction, sljit_u32 size)
+{
+	SLJIT_UNUSED_ARG(size);
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_op_custom(compiler, instruction, size));
+
+	return push_inst(compiler, *(sljit_ins*)instruction);
+}
+
+/* --------------------------------------------------------------------- */
+/*  Floating point operators                                             */
+/* --------------------------------------------------------------------- */
+
+#define FPU_LOAD (1 << 20)
+#define EMIT_FPU_DATA_TRANSFER(inst, add, base, freg, offs) \
+	((inst) | (sljit_ins)((add) << 23) | RN(base) | VD(freg) | (sljit_ins)(offs))
+
+static sljit_s32 emit_fop_mem(struct sljit_compiler *compiler, sljit_s32 flags, sljit_s32 reg, sljit_s32 arg, sljit_sw argw)
+{
+	sljit_uw imm;
+	sljit_ins inst = VSTR_F32 | (flags & (SLJIT_32 | FPU_LOAD));
+
+	SLJIT_ASSERT(arg & SLJIT_MEM);
+	arg &= ~SLJIT_MEM;
+
+	if (SLJIT_UNLIKELY(arg & OFFS_REG_MASK)) {
+		FAIL_IF(push_inst(compiler, ADD | RD(TMP_REG1) | RN(arg & REG_MASK) | RM(OFFS_REG(arg)) | (((sljit_ins)argw & 0x3) << 7)));
+		arg = TMP_REG1;
+		argw = 0;
+	}
+
+	/* Fast loads and stores. */
+	if (arg) {
+		if (!(argw & ~0x3fc))
+			return push_inst(compiler, EMIT_FPU_DATA_TRANSFER(inst, 1, arg & REG_MASK, reg, argw >> 2));
+		if (!(-argw & ~0x3fc))
+			return push_inst(compiler, EMIT_FPU_DATA_TRANSFER(inst, 0, arg & REG_MASK, reg, (-argw) >> 2));
+
+		imm = get_imm((sljit_uw)argw & ~(sljit_uw)0x3fc);
+		if (imm) {
+			FAIL_IF(push_inst(compiler, ADD | RD(TMP_REG1) | RN(arg & REG_MASK) | imm));
+			return push_inst(compiler, EMIT_FPU_DATA_TRANSFER(inst, 1, TMP_REG1, reg, (argw & 0x3fc) >> 2));
+		}
+		imm = get_imm((sljit_uw)-argw & ~(sljit_uw)0x3fc);
+		if (imm) {
+			argw = -argw;
+			FAIL_IF(push_inst(compiler, SUB | RD(TMP_REG1) | RN(arg & REG_MASK) | imm));
+			return push_inst(compiler, EMIT_FPU_DATA_TRANSFER(inst, 0, TMP_REG1, reg, (argw & 0x3fc) >> 2));
+		}
+	}
+
+	if (arg) {
+		FAIL_IF(load_immediate(compiler, TMP_REG1, (sljit_uw)argw));
+		FAIL_IF(push_inst(compiler, ADD | RD(TMP_REG1) | RN(arg & REG_MASK) | RM(TMP_REG1)));
+	}
+	else
+		FAIL_IF(load_immediate(compiler, TMP_REG1, (sljit_uw)argw));
+
+	return push_inst(compiler, EMIT_FPU_DATA_TRANSFER(inst, 1, TMP_REG1, reg, 0));
+}
+
+static SLJIT_INLINE sljit_s32 sljit_emit_fop1_conv_sw_from_f64(struct sljit_compiler *compiler, sljit_s32 op,
+	sljit_s32 dst, sljit_sw dstw,
+	sljit_s32 src, sljit_sw srcw)
+{
+	op ^= SLJIT_32;
+
+	if (src & SLJIT_MEM) {
+		FAIL_IF(emit_fop_mem(compiler, (op & SLJIT_32) | FPU_LOAD, TMP_FREG1, src, srcw));
+		src = TMP_FREG1;
+	}
+
+	FAIL_IF(push_inst(compiler, EMIT_FPU_OPERATION(VCVT_S32_F32, op & SLJIT_32, TMP_FREG1, src, 0)));
+
+	if (FAST_IS_REG(dst))
+		return push_inst(compiler, VMOV | (1 << 20) | RD(dst) | VN(TMP_FREG1));
+
+	/* Store the integer value from a VFP register. */
+	return emit_fop_mem(compiler, 0, TMP_FREG1, dst, dstw);
+}
+
+static sljit_s32 sljit_emit_fop1_conv_f64_from_w(struct sljit_compiler *compiler, sljit_ins ins,
+	sljit_s32 dst, sljit_sw dstw,
+	sljit_s32 src, sljit_sw srcw)
+{
+	sljit_s32 dst_r = FAST_IS_REG(dst) ? dst : TMP_FREG1;
+
+	if (FAST_IS_REG(src))
+		FAIL_IF(push_inst(compiler, VMOV | RD(src) | VN(TMP_FREG1)));
+	else if (src & SLJIT_MEM) {
+		/* Load the integer value into a VFP register. */
+		FAIL_IF(emit_fop_mem(compiler, FPU_LOAD, TMP_FREG1, src, srcw));
+	}
+	else {
+		FAIL_IF(load_immediate(compiler, TMP_REG1, (sljit_uw)srcw));
+		FAIL_IF(push_inst(compiler, VMOV | RD(TMP_REG1) | VN(TMP_FREG1)));
+	}
+
+	FAIL_IF(push_inst(compiler, EMIT_FPU_OPERATION(ins, ins & SLJIT_32, dst_r, TMP_FREG1, 0)));
+
+	if (dst & SLJIT_MEM)
+		return emit_fop_mem(compiler, (ins & SLJIT_32), TMP_FREG1, dst, dstw);
+	return SLJIT_SUCCESS;
+}
+
+static SLJIT_INLINE sljit_s32 sljit_emit_fop1_conv_f64_from_sw(struct sljit_compiler *compiler, sljit_s32 op,
+	sljit_s32 dst, sljit_sw dstw,
+	sljit_s32 src, sljit_sw srcw)
+{
+	return sljit_emit_fop1_conv_f64_from_w(compiler, VCVT_F32_S32 | (~op & SLJIT_32), dst, dstw, src, srcw);
+}
+
+static SLJIT_INLINE sljit_s32 sljit_emit_fop1_conv_f64_from_uw(struct sljit_compiler *compiler, sljit_s32 op,
+	sljit_s32 dst, sljit_sw dstw,
+	sljit_s32 src, sljit_sw srcw)
+{
+	return sljit_emit_fop1_conv_f64_from_w(compiler, VCVT_F32_U32 | (~op & SLJIT_32), dst, dstw, src, srcw);
+}
+
+static SLJIT_INLINE sljit_s32 sljit_emit_fop1_cmp(struct sljit_compiler *compiler, sljit_s32 op,
+	sljit_s32 src1, sljit_sw src1w,
+	sljit_s32 src2, sljit_sw src2w)
+{
+	op ^= SLJIT_32;
+
+	if (src1 & SLJIT_MEM) {
+		FAIL_IF(emit_fop_mem(compiler, (op & SLJIT_32) | FPU_LOAD, TMP_FREG1, src1, src1w));
+		src1 = TMP_FREG1;
+	}
+
+	if (src2 & SLJIT_MEM) {
+		FAIL_IF(emit_fop_mem(compiler, (op & SLJIT_32) | FPU_LOAD, TMP_FREG2, src2, src2w));
+		src2 = TMP_FREG2;
+	}
+
+	FAIL_IF(push_inst(compiler, EMIT_FPU_OPERATION(VCMP_F32, op & SLJIT_32, src1, src2, 0)));
+	FAIL_IF(push_inst(compiler, VMRS));
+
+	if (GET_FLAG_TYPE(op) != SLJIT_UNORDERED_OR_EQUAL)
+		return SLJIT_SUCCESS;
+
+	return push_inst(compiler, (CMP - CONDITIONAL) | (0x60000000 /* VS */) | SET_FLAGS | RN(TMP_REG1) | RM(TMP_REG1));
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_fop1(struct sljit_compiler *compiler, sljit_s32 op,
+	sljit_s32 dst, sljit_sw dstw,
+	sljit_s32 src, sljit_sw srcw)
+{
+	sljit_s32 dst_r;
+
+	CHECK_ERROR();
+
+	SLJIT_COMPILE_ASSERT((SLJIT_32 == 0x100), float_transfer_bit_error);
+	SELECT_FOP1_OPERATION_WITH_CHECKS(compiler, op, dst, dstw, src, srcw);
+
+	dst_r = FAST_IS_REG(dst) ? dst : TMP_FREG1;
+
+	if (GET_OPCODE(op) != SLJIT_CONV_F64_FROM_F32)
+		op ^= SLJIT_32;
+
+	if (src & SLJIT_MEM) {
+		FAIL_IF(emit_fop_mem(compiler, (op & SLJIT_32) | FPU_LOAD, dst_r, src, srcw));
+		src = dst_r;
+	}
+
+	switch (GET_OPCODE(op)) {
+	case SLJIT_MOV_F64:
+		if (src != dst_r) {
+			if (!(dst & SLJIT_MEM))
+				FAIL_IF(push_inst(compiler, EMIT_FPU_OPERATION(VMOV_F32, op & SLJIT_32, dst_r, src, 0)));
+			else
+				dst_r = src;
+		}
+		break;
+	case SLJIT_NEG_F64:
+		FAIL_IF(push_inst(compiler, EMIT_FPU_OPERATION(VNEG_F32, op & SLJIT_32, dst_r, src, 0)));
+		break;
+	case SLJIT_ABS_F64:
+		FAIL_IF(push_inst(compiler, EMIT_FPU_OPERATION(VABS_F32, op & SLJIT_32, dst_r, src, 0)));
+		break;
+	case SLJIT_CONV_F64_FROM_F32:
+		FAIL_IF(push_inst(compiler, EMIT_FPU_OPERATION(VCVT_F64_F32, op & SLJIT_32, dst_r, src, 0)));
+		op ^= SLJIT_32;
+		break;
+	}
+
+	if (dst & SLJIT_MEM)
+		return emit_fop_mem(compiler, (op & SLJIT_32), dst_r, dst, dstw);
+	return SLJIT_SUCCESS;
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_fop2(struct sljit_compiler *compiler, sljit_s32 op,
+	sljit_s32 dst, sljit_sw dstw,
+	sljit_s32 src1, sljit_sw src1w,
+	sljit_s32 src2, sljit_sw src2w)
+{
+	sljit_s32 dst_r;
+
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_fop2(compiler, op, dst, dstw, src1, src1w, src2, src2w));
+	ADJUST_LOCAL_OFFSET(dst, dstw);
+	ADJUST_LOCAL_OFFSET(src1, src1w);
+	ADJUST_LOCAL_OFFSET(src2, src2w);
+
+	op ^= SLJIT_32;
+
+	dst_r = FAST_IS_REG(dst) ? dst : TMP_FREG1;
+
+	if (src2 & SLJIT_MEM) {
+		FAIL_IF(emit_fop_mem(compiler, (op & SLJIT_32) | FPU_LOAD, TMP_FREG2, src2, src2w));
+		src2 = TMP_FREG2;
+	}
+
+	if (src1 & SLJIT_MEM) {
+		FAIL_IF(emit_fop_mem(compiler, (op & SLJIT_32) | FPU_LOAD, TMP_FREG1, src1, src1w));
+		src1 = TMP_FREG1;
+	}
+
+	switch (GET_OPCODE(op)) {
+	case SLJIT_ADD_F64:
+		FAIL_IF(push_inst(compiler, EMIT_FPU_OPERATION(VADD_F32, op & SLJIT_32, dst_r, src2, src1)));
+		break;
+	case SLJIT_SUB_F64:
+		FAIL_IF(push_inst(compiler, EMIT_FPU_OPERATION(VSUB_F32, op & SLJIT_32, dst_r, src2, src1)));
+		break;
+	case SLJIT_MUL_F64:
+		FAIL_IF(push_inst(compiler, EMIT_FPU_OPERATION(VMUL_F32, op & SLJIT_32, dst_r, src2, src1)));
+		break;
+	case SLJIT_DIV_F64:
+		FAIL_IF(push_inst(compiler, EMIT_FPU_OPERATION(VDIV_F32, op & SLJIT_32, dst_r, src2, src1)));
+		break;
+	case SLJIT_COPYSIGN_F64:
+		FAIL_IF(push_inst(compiler, VMOV | (1 << 20) | VN(src2) | RD(TMP_REG1) | ((op & SLJIT_32) ? (1 << 7) : 0)));
+		FAIL_IF(push_inst(compiler, EMIT_FPU_OPERATION(VABS_F32, op & SLJIT_32, dst_r, src1, 0)));
+		FAIL_IF(push_inst(compiler, CMP | SET_FLAGS | RN(TMP_REG1) | SRC2_IMM | 0));
+		return push_inst(compiler, EMIT_FPU_OPERATION((VNEG_F32 & ~COND_MASK) | 0xb0000000, op & SLJIT_32, dst_r, dst_r, 0));
+	}
+
+	if (dst_r != dst)
+		FAIL_IF(emit_fop_mem(compiler, (op & SLJIT_32), TMP_FREG1, dst, dstw));
+
+	return SLJIT_SUCCESS;
+}
+
+#undef EMIT_FPU_DATA_TRANSFER
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_fset32(struct sljit_compiler *compiler,
+	sljit_s32 freg, sljit_f32 value)
+{
+#if defined(__ARM_NEON) && __ARM_NEON
+	sljit_u32 exp;
+	sljit_ins ins;
+#endif /* NEON */
+	union {
+		sljit_u32 imm;
+		sljit_f32 value;
+	} u;
+
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_fset32(compiler, freg, value));
+
+	u.value = value;
+
+#if defined(__ARM_NEON) && __ARM_NEON
+	if ((u.imm << (32 - 19)) == 0) {
+		exp = (u.imm >> (23 + 2)) & 0x3f;
+
+		if (exp == 0x20 || exp == 0x1f) {
+			ins = ((u.imm >> 24) & 0x80) | ((u.imm >> 19) & 0x7f);
+			return push_inst(compiler, (VMOV_F32 ^ (1 << 6)) | ((ins & 0xf0) << 12) | VD(freg) | (ins & 0xf));
+		}
+	}
+#endif /* NEON */
+
+	FAIL_IF(load_immediate(compiler, TMP_REG1, u.imm));
+	return push_inst(compiler, VMOV | VN(freg) | RD(TMP_REG1));
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_fset64(struct sljit_compiler *compiler,
+	sljit_s32 freg, sljit_f64 value)
+{
+#if defined(__ARM_NEON) && __ARM_NEON
+	sljit_u32 exp;
+	sljit_ins ins;
+#endif /* NEON */
+	union {
+		sljit_u32 imm[2];
+		sljit_f64 value;
+	} u;
+
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_fset64(compiler, freg, value));
+
+	u.value = value;
+
+#if defined(__ARM_NEON) && __ARM_NEON
+	if (u.imm[0] == 0 && (u.imm[1] << (64 - 48)) == 0) {
+		exp = (u.imm[1] >> ((52 - 32) + 2)) & 0x1ff;
+
+		if (exp == 0x100 || exp == 0xff) {
+			ins = ((u.imm[1] >> (56 - 32)) & 0x80) | ((u.imm[1] >> (48 - 32)) & 0x7f);
+			return push_inst(compiler, (VMOV_F32 ^ (1 << 6)) | (1 << 8) | ((ins & 0xf0) << 12) | VD(freg) | (ins & 0xf));
+		}
+	}
+#endif /* NEON */
+
+	FAIL_IF(load_immediate(compiler, TMP_REG1, u.imm[0]));
+	if (u.imm[0] == u.imm[1])
+		return push_inst(compiler, VMOV2 | RN(TMP_REG1) | RD(TMP_REG1) | VM(freg));
+
+	FAIL_IF(load_immediate(compiler, TMP_REG2, u.imm[1]));
+	return push_inst(compiler, VMOV2 | RN(TMP_REG2) | RD(TMP_REG1) | VM(freg));
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_fcopy(struct sljit_compiler *compiler, sljit_s32 op,
+	sljit_s32 freg, sljit_s32 reg)
+{
+	sljit_s32 reg2;
+	sljit_ins inst;
+
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_fcopy(compiler, op, freg, reg));
+
+	if (reg & REG_PAIR_MASK) {
+		reg2 = REG_PAIR_SECOND(reg);
+		reg = REG_PAIR_FIRST(reg);
+
+		inst = VMOV2 | RN(reg) | RD(reg2) | VM(freg);
+	} else {
+		inst = VMOV | VN(freg) | RD(reg);
+
+		if (!(op & SLJIT_32))
+			inst |= 1 << 7;
+	}
+
+	if (GET_OPCODE(op) == SLJIT_COPY_FROM_F64)
+		inst |= 1 << 20;
+
+	return push_inst(compiler, inst);
+}
+
+/* --------------------------------------------------------------------- */
+/*  Conditional instructions                                             */
+/* --------------------------------------------------------------------- */
+
+static sljit_ins get_cc(struct sljit_compiler *compiler, sljit_s32 type)
+{
+	switch (type) {
+	case SLJIT_EQUAL:
+	case SLJIT_ATOMIC_STORED:
+	case SLJIT_F_EQUAL:
+	case SLJIT_ORDERED_EQUAL:
+	case SLJIT_UNORDERED_OR_EQUAL:
+		return 0x00000000;
+
+	case SLJIT_NOT_EQUAL:
+	case SLJIT_ATOMIC_NOT_STORED:
+	case SLJIT_F_NOT_EQUAL:
+	case SLJIT_UNORDERED_OR_NOT_EQUAL:
+	case SLJIT_ORDERED_NOT_EQUAL:
+		return 0x10000000;
+
+	case SLJIT_CARRY:
+		if (compiler->status_flags_state & SLJIT_CURRENT_FLAGS_ADD)
+			return 0x20000000;
+		/* fallthrough */
+
+	case SLJIT_LESS:
+		return 0x30000000;
+
+	case SLJIT_NOT_CARRY:
+		if (compiler->status_flags_state & SLJIT_CURRENT_FLAGS_ADD)
+			return 0x30000000;
+		/* fallthrough */
+
+	case SLJIT_GREATER_EQUAL:
+		return 0x20000000;
+
+	case SLJIT_GREATER:
+	case SLJIT_UNORDERED_OR_GREATER:
+		return 0x80000000;
+
+	case SLJIT_LESS_EQUAL:
+	case SLJIT_F_LESS_EQUAL:
+	case SLJIT_ORDERED_LESS_EQUAL:
+		return 0x90000000;
+
+	case SLJIT_SIG_LESS:
+	case SLJIT_UNORDERED_OR_LESS:
+		return 0xb0000000;
+
+	case SLJIT_SIG_GREATER_EQUAL:
+	case SLJIT_F_GREATER_EQUAL:
+	case SLJIT_ORDERED_GREATER_EQUAL:
+		return 0xa0000000;
+
+	case SLJIT_SIG_GREATER:
+	case SLJIT_F_GREATER:
+	case SLJIT_ORDERED_GREATER:
+		return 0xc0000000;
+
+	case SLJIT_SIG_LESS_EQUAL:
+	case SLJIT_UNORDERED_OR_LESS_EQUAL:
+		return 0xd0000000;
+
+	case SLJIT_OVERFLOW:
+		if (!(compiler->status_flags_state & (SLJIT_CURRENT_FLAGS_ADD | SLJIT_CURRENT_FLAGS_SUB)))
+			return 0x10000000;
+		/* fallthrough */
+
+	case SLJIT_UNORDERED:
+		return 0x60000000;
+
+	case SLJIT_NOT_OVERFLOW:
+		if (!(compiler->status_flags_state & (SLJIT_CURRENT_FLAGS_ADD | SLJIT_CURRENT_FLAGS_SUB)))
+			return 0x00000000;
+		/* fallthrough */
+
+	case SLJIT_ORDERED:
+		return 0x70000000;
+
+	case SLJIT_F_LESS:
+	case SLJIT_ORDERED_LESS:
+		return 0x40000000;
+
+	case SLJIT_UNORDERED_OR_GREATER_EQUAL:
+		return 0x50000000;
+
+	default:
+		SLJIT_ASSERT(type >= SLJIT_JUMP && type <= SLJIT_CALL_REG_ARG);
+		return 0xe0000000;
+	}
+}
+
+SLJIT_API_FUNC_ATTRIBUTE struct sljit_label* sljit_emit_label(struct sljit_compiler *compiler)
+{
+	struct sljit_label *label;
+
+	CHECK_ERROR_PTR();
+	CHECK_PTR(check_sljit_emit_label(compiler));
+
+	if (compiler->last_label && compiler->last_label->size == compiler->size)
+		return compiler->last_label;
+
+	label = (struct sljit_label*)ensure_abuf(compiler, sizeof(struct sljit_label));
+	PTR_FAIL_IF(!label);
+	set_label(label, compiler);
+	return label;
+}
+
+SLJIT_API_FUNC_ATTRIBUTE struct sljit_jump* sljit_emit_jump(struct sljit_compiler *compiler, sljit_s32 type)
+{
+	struct sljit_jump *jump;
+
+	CHECK_ERROR_PTR();
+	CHECK_PTR(check_sljit_emit_jump(compiler, type));
+
+	jump = (struct sljit_jump*)ensure_abuf(compiler, sizeof(struct sljit_jump));
+	PTR_FAIL_IF(!jump);
+	set_jump(jump, compiler, type & SLJIT_REWRITABLE_JUMP);
+	type &= 0xff;
+
+	SLJIT_ASSERT(reg_map[TMP_REG1] != 14);
+
+#if (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+	if (type >= SLJIT_FAST_CALL)
+		PTR_FAIL_IF(prepare_blx(compiler));
+
+	jump->addr = compiler->size;
+	PTR_FAIL_IF(push_inst_with_unique_literal(compiler, ((EMIT_DATA_TRANSFER(WORD_SIZE | LOAD_DATA, 1,
+		type <= SLJIT_JUMP ? TMP_PC : TMP_REG1, TMP_PC, 0)) & ~COND_MASK) | get_cc(compiler, type), 0));
+
+	if (jump->flags & SLJIT_REWRITABLE_JUMP)
+		compiler->patches++;
+
+	if (type >= SLJIT_FAST_CALL) {
+		jump->flags |= IS_BL;
+		jump->addr = compiler->size;
+		PTR_FAIL_IF(emit_blx(compiler));
+	}
+#else /* !SLJIT_CONFIG_ARM_V6 */
+	jump->addr = compiler->size;
+	if (type >= SLJIT_FAST_CALL)
+		jump->flags |= IS_BL;
+	PTR_FAIL_IF(push_inst(compiler, (((type <= SLJIT_JUMP ? BX : BLX) | RM(TMP_REG1)) & ~COND_MASK) | get_cc(compiler, type)));
+	compiler->size += JUMP_MAX_SIZE - 1;
+#endif /* SLJIT_CONFIG_ARM_V6 */
+	return jump;
+}
+
+#ifdef __SOFTFP__
+
+static sljit_s32 softfloat_call_with_args(struct sljit_compiler *compiler, sljit_s32 arg_types, sljit_s32 *src, sljit_u32 *extra_space)
+{
+	sljit_u32 is_tail_call = *extra_space & SLJIT_CALL_RETURN;
+	sljit_u32 offset = 0;
+	sljit_u32 word_arg_offset = 0;
+	sljit_u32 src_offset = 4 * sizeof(sljit_sw);
+	sljit_u32 float_arg_count = 0;
+	sljit_s32 types = 0;
+	sljit_u8 offsets[4];
+	sljit_u8 *offset_ptr = offsets;
+
+	if (src && FAST_IS_REG(*src))
+		src_offset = (sljit_u32)reg_map[*src] * sizeof(sljit_sw);
+
+	arg_types >>= SLJIT_ARG_SHIFT;
+
+	while (arg_types) {
+		types = (types << SLJIT_ARG_SHIFT) | (arg_types & SLJIT_ARG_MASK);
+
+		switch (arg_types & SLJIT_ARG_MASK) {
+		case SLJIT_ARG_TYPE_F64:
+			if (offset & 0x7)
+				offset += sizeof(sljit_sw);
+			*offset_ptr++ = (sljit_u8)offset;
+			offset += sizeof(sljit_f64);
+			float_arg_count++;
+			break;
+		case SLJIT_ARG_TYPE_F32:
+			*offset_ptr++ = (sljit_u8)offset;
+			offset += sizeof(sljit_f32);
+			float_arg_count++;
+			break;
+		default:
+			*offset_ptr++ = (sljit_u8)offset;
+			offset += sizeof(sljit_sw);
+			word_arg_offset += sizeof(sljit_sw);
+			break;
+		}
+
+		arg_types >>= SLJIT_ARG_SHIFT;
+	}
+
+	if (offset > 4 * sizeof(sljit_sw) && (!is_tail_call || offset > compiler->args_size)) {
+		/* Keep lr register on the stack. */
+		if (is_tail_call)
+			offset += sizeof(sljit_sw);
+
+		offset = ((offset - 4 * sizeof(sljit_sw)) + 0x7) & ~(sljit_u32)0x7;
+
+		*extra_space = offset;
+
+		if (is_tail_call)
+			FAIL_IF(emit_stack_frame_release(compiler, (sljit_s32)offset));
+		else
+			FAIL_IF(push_inst(compiler, SUB | RD(SLJIT_SP) | RN(SLJIT_SP) | SRC2_IMM | offset));
+	} else {
+		if (is_tail_call)
+			FAIL_IF(emit_stack_frame_release(compiler, -1));
+		*extra_space = 0;
+	}
+
+	/* Process arguments in reversed direction. */
+	while (types) {
+		switch (types & SLJIT_ARG_MASK) {
+		case SLJIT_ARG_TYPE_F64:
+			float_arg_count--;
+			offset = *(--offset_ptr);
+
+			SLJIT_ASSERT((offset & 0x7) == 0);
+
+			if (offset < 4 * sizeof(sljit_sw)) {
+				if (src_offset == offset || src_offset == offset + sizeof(sljit_sw)) {
+					FAIL_IF(push_inst(compiler, MOV | RD(TMP_REG1) | (src_offset >> 2)));
+					*src = TMP_REG1;
+				}
+				FAIL_IF(push_inst(compiler, VMOV2 | 0x100000 | (offset << 10) | ((offset + sizeof(sljit_sw)) << 14) | float_arg_count));
+			} else
+				FAIL_IF(push_inst(compiler, VSTR_F32 | 0x800100 | RN(SLJIT_SP)
+						| (float_arg_count << 12) | ((offset - 4 * sizeof(sljit_sw)) >> 2)));
+			break;
+		case SLJIT_ARG_TYPE_F32:
+			float_arg_count--;
+			offset = *(--offset_ptr);
+
+			if (offset < 4 * sizeof(sljit_sw)) {
+				if (src_offset == offset) {
+					FAIL_IF(push_inst(compiler, MOV | RD(TMP_REG1) | (src_offset >> 2)));
+					*src = TMP_REG1;
+				}
+				FAIL_IF(push_inst(compiler, VMOV | 0x100000 | (float_arg_count << 16) | (offset << 10)));
+			} else
+				FAIL_IF(push_inst(compiler, VSTR_F32 | 0x800000 | RN(SLJIT_SP)
+						| (float_arg_count << 12) | ((offset - 4 * sizeof(sljit_sw)) >> 2)));
+			break;
+		default:
+			word_arg_offset -= sizeof(sljit_sw);
+			offset = *(--offset_ptr);
+
+			SLJIT_ASSERT(offset >= word_arg_offset);
+
+			if (offset != word_arg_offset) {
+				if (offset < 4 * sizeof(sljit_sw)) {
+					if (src_offset == offset) {
+						FAIL_IF(push_inst(compiler, MOV | RD(TMP_REG1) | (src_offset >> 2)));
+						*src = TMP_REG1;
+					}
+					else if (src_offset == word_arg_offset) {
+						*src = (sljit_s32)(SLJIT_R0 + (offset >> 2));
+						src_offset = offset;
+					}
+					FAIL_IF(push_inst(compiler, MOV | (offset << 10) | (word_arg_offset >> 2)));
+				} else
+					FAIL_IF(push_inst(compiler, STR | 0x800000 | RN(SLJIT_SP) | (word_arg_offset << 10) | (offset - 4 * sizeof(sljit_sw))));
+			}
+			break;
+		}
+
+		types >>= SLJIT_ARG_SHIFT;
+	}
+
+	return SLJIT_SUCCESS;
+}
+
+static sljit_s32 softfloat_post_call_with_args(struct sljit_compiler *compiler, sljit_s32 arg_types)
+{
+	if ((arg_types & SLJIT_ARG_MASK) == SLJIT_ARG_TYPE_F64)
+		FAIL_IF(push_inst(compiler, VMOV2 | (1 << 16) | (0 << 12) | 0));
+	if ((arg_types & SLJIT_ARG_MASK) == SLJIT_ARG_TYPE_F32)
+		FAIL_IF(push_inst(compiler, VMOV | (0 << 16) | (0 << 12)));
+
+	return SLJIT_SUCCESS;
+}
+
+#else /* !__SOFTFP__ */
+
+static sljit_s32 hardfloat_call_with_args(struct sljit_compiler *compiler, sljit_s32 arg_types)
+{
+	sljit_u32 offset = SLJIT_FR0;
+	sljit_u32 new_offset = SLJIT_FR0;
+	sljit_u32 f32_offset = 0;
+
+	/* Remove return value. */
+	arg_types >>= SLJIT_ARG_SHIFT;
+
+	while (arg_types) {
+		switch (arg_types & SLJIT_ARG_MASK) {
+		case SLJIT_ARG_TYPE_F64:
+			if (offset != new_offset)
+				FAIL_IF(push_inst(compiler, EMIT_FPU_OPERATION(VMOV_F32,
+					SLJIT_32, new_offset, offset, 0)));
+
+			new_offset++;
+			offset++;
+			break;
+		case SLJIT_ARG_TYPE_F32:
+			if (f32_offset != 0) {
+				FAIL_IF(push_inst(compiler, EMIT_FPU_OPERATION(VMOV_F32,
+					0x400000, f32_offset, offset, 0)));
+				f32_offset = 0;
+			} else {
+				if (offset != new_offset)
+					FAIL_IF(push_inst(compiler, EMIT_FPU_OPERATION(VMOV_F32,
+						0, new_offset, offset, 0)));
+				f32_offset = new_offset;
+				new_offset++;
+			}
+			offset++;
+			break;
+		}
+		arg_types >>= SLJIT_ARG_SHIFT;
+	}
+
+	return SLJIT_SUCCESS;
+}
+
+#endif /* __SOFTFP__ */
+
+SLJIT_API_FUNC_ATTRIBUTE struct sljit_jump* sljit_emit_call(struct sljit_compiler *compiler, sljit_s32 type,
+	sljit_s32 arg_types)
+{
+#ifdef __SOFTFP__
+	struct sljit_jump *jump;
+	sljit_u32 extra_space = (sljit_u32)type;
+#endif
+
+	CHECK_ERROR_PTR();
+	CHECK_PTR(check_sljit_emit_call(compiler, type, arg_types));
+
+#ifdef __SOFTFP__
+	if ((type & 0xff) != SLJIT_CALL_REG_ARG) {
+		PTR_FAIL_IF(softfloat_call_with_args(compiler, arg_types, NULL, &extra_space));
+		SLJIT_ASSERT((extra_space & 0x7) == 0);
+
+		if ((type & SLJIT_CALL_RETURN) && extra_space == 0)
+			type = SLJIT_JUMP | (type & SLJIT_REWRITABLE_JUMP);
+
+		SLJIT_SKIP_CHECKS(compiler);
+		jump = sljit_emit_jump(compiler, type);
+		PTR_FAIL_IF(jump == NULL);
+
+		if (extra_space > 0) {
+			if (type & SLJIT_CALL_RETURN)
+				PTR_FAIL_IF(push_inst(compiler, EMIT_DATA_TRANSFER(WORD_SIZE | LOAD_DATA, 1,
+					TMP_REG2, SLJIT_SP, extra_space - sizeof(sljit_sw))));
+
+			PTR_FAIL_IF(push_inst(compiler, ADD | RD(SLJIT_SP) | RN(SLJIT_SP) | SRC2_IMM | extra_space));
+
+			if (type & SLJIT_CALL_RETURN) {
+				PTR_FAIL_IF(push_inst(compiler, BX | RM(TMP_REG2)));
+				return jump;
+			}
+		}
+
+		SLJIT_ASSERT(!(type & SLJIT_CALL_RETURN));
+		PTR_FAIL_IF(softfloat_post_call_with_args(compiler, arg_types));
+		return jump;
+	}
+#endif /* __SOFTFP__ */
+
+	if (type & SLJIT_CALL_RETURN) {
+		PTR_FAIL_IF(emit_stack_frame_release(compiler, -1));
+		type = SLJIT_JUMP | (type & SLJIT_REWRITABLE_JUMP);
+	}
+
+#ifndef __SOFTFP__
+	if ((type & 0xff) != SLJIT_CALL_REG_ARG)
+		PTR_FAIL_IF(hardfloat_call_with_args(compiler, arg_types));
+#endif /* !__SOFTFP__ */
+
+	SLJIT_SKIP_CHECKS(compiler);
+	return sljit_emit_jump(compiler, type);
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_ijump(struct sljit_compiler *compiler, sljit_s32 type, sljit_s32 src, sljit_sw srcw)
+{
+	struct sljit_jump *jump;
+
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_ijump(compiler, type, src, srcw));
+	ADJUST_LOCAL_OFFSET(src, srcw);
+
+	SLJIT_ASSERT(reg_map[TMP_REG1] != 14);
+
+	if (src != SLJIT_IMM) {
+		if (FAST_IS_REG(src)) {
+			SLJIT_ASSERT(reg_map[src] != 14);
+			return push_inst(compiler, (type <= SLJIT_JUMP ? BX : BLX) | RM(src));
+		}
+
+		SLJIT_ASSERT(src & SLJIT_MEM);
+		FAIL_IF(emit_op_mem(compiler, WORD_SIZE | LOAD_DATA, TMP_REG1, src, srcw, TMP_REG1));
+		return push_inst(compiler, (type <= SLJIT_JUMP ? BX : BLX) | RM(TMP_REG1));
+	}
+
+	/* These jumps are converted to jump/call instructions when possible. */
+	jump = (struct sljit_jump*)ensure_abuf(compiler, sizeof(struct sljit_jump));
+	FAIL_IF(!jump);
+	set_jump(jump, compiler, JUMP_ADDR | ((type >= SLJIT_FAST_CALL) ? IS_BL : 0));
+	jump->u.target = (sljit_uw)srcw;
+
+#if (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+	if (type >= SLJIT_FAST_CALL)
+		FAIL_IF(prepare_blx(compiler));
+	jump->addr = compiler->size;
+	FAIL_IF(push_inst_with_unique_literal(compiler, EMIT_DATA_TRANSFER(WORD_SIZE | LOAD_DATA, 1, type <= SLJIT_JUMP ? TMP_PC : TMP_REG1, TMP_PC, 0), 0));
+	if (type >= SLJIT_FAST_CALL) {
+		jump->addr = compiler->size;
+		FAIL_IF(emit_blx(compiler));
+	}
+#else /* !SLJIT_CONFIG_ARM_V6 */
+	jump->addr = compiler->size;
+	FAIL_IF(push_inst(compiler, (type <= SLJIT_JUMP ? BX : BLX) | RM(TMP_REG1)));
+	compiler->size += JUMP_MAX_SIZE - 1;
+#endif /* SLJIT_CONFIG_ARM_V6 */
+	return SLJIT_SUCCESS;
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_icall(struct sljit_compiler *compiler, sljit_s32 type,
+	sljit_s32 arg_types,
+	sljit_s32 src, sljit_sw srcw)
+{
+#ifdef __SOFTFP__
+	sljit_u32 extra_space = (sljit_u32)type;
+#endif
+
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_icall(compiler, type, arg_types, src, srcw));
+
+	if (src & SLJIT_MEM) {
+		FAIL_IF(emit_op_mem(compiler, WORD_SIZE | LOAD_DATA, TMP_REG1, src, srcw, TMP_REG1));
+		src = TMP_REG1;
+	}
+
+	if ((type & SLJIT_CALL_RETURN) && (src >= SLJIT_FIRST_SAVED_REG && src <= (SLJIT_S0 - SLJIT_KEPT_SAVEDS_COUNT(compiler->options)))) {
+		FAIL_IF(push_inst(compiler, MOV | RD(TMP_REG1) | RM(src)));
+		src = TMP_REG1;
+	}
+
+#ifdef __SOFTFP__
+	if ((type & 0xff) != SLJIT_CALL_REG_ARG) {
+		FAIL_IF(softfloat_call_with_args(compiler, arg_types, &src, &extra_space));
+		SLJIT_ASSERT((extra_space & 0x7) == 0);
+
+		if ((type & SLJIT_CALL_RETURN) && extra_space == 0)
+			type = SLJIT_JUMP;
+
+		SLJIT_SKIP_CHECKS(compiler);
+		FAIL_IF(sljit_emit_ijump(compiler, type, src, srcw));
+
+		if (extra_space > 0) {
+			if (type & SLJIT_CALL_RETURN)
+				FAIL_IF(push_inst(compiler, EMIT_DATA_TRANSFER(WORD_SIZE | LOAD_DATA, 1,
+					TMP_REG2, SLJIT_SP, extra_space - sizeof(sljit_sw))));
+
+			FAIL_IF(push_inst(compiler, ADD | RD(SLJIT_SP) | RN(SLJIT_SP) | SRC2_IMM | extra_space));
+
+			if (type & SLJIT_CALL_RETURN)
+				return push_inst(compiler, BX | RM(TMP_REG2));
+		}
+
+		SLJIT_ASSERT(!(type & SLJIT_CALL_RETURN));
+		return softfloat_post_call_with_args(compiler, arg_types);
+	}
+#endif /* __SOFTFP__ */
+
+	if (type & SLJIT_CALL_RETURN) {
+		FAIL_IF(emit_stack_frame_release(compiler, -1));
+		type = SLJIT_JUMP;
+	}
+
+#ifndef __SOFTFP__
+	if ((type & 0xff) != SLJIT_CALL_REG_ARG)
+		FAIL_IF(hardfloat_call_with_args(compiler, arg_types));
+#endif /* !__SOFTFP__ */
+
+	SLJIT_SKIP_CHECKS(compiler);
+	return sljit_emit_ijump(compiler, type, src, srcw);
+}
+
+#ifdef __SOFTFP__
+
+static SLJIT_INLINE sljit_s32 emit_fmov_before_return(struct sljit_compiler *compiler, sljit_s32 op, sljit_s32 src, sljit_sw srcw)
+{
+	if (compiler->options & SLJIT_ENTER_REG_ARG) {
+		if (src == SLJIT_FR0)
+			return SLJIT_SUCCESS;
+
+		SLJIT_SKIP_CHECKS(compiler);
+		return sljit_emit_fop1(compiler, op, SLJIT_RETURN_FREG, 0, src, srcw);
+	}
+
+	if (FAST_IS_REG(src)) {
+		if (op & SLJIT_32)
+			return push_inst(compiler, VMOV | (1 << 20) | RD(SLJIT_R0) | VN(src));
+		return push_inst(compiler, VMOV2 | (1 << 20) | RD(SLJIT_R0) | RN(SLJIT_R1) | VM(src));
+	}
+
+	SLJIT_SKIP_CHECKS(compiler);
+
+	if (op & SLJIT_32)
+		return sljit_emit_op1(compiler, SLJIT_MOV, SLJIT_R0, 0, src, srcw);
+	return sljit_emit_mem(compiler, SLJIT_MOV, SLJIT_REG_PAIR(SLJIT_R0, SLJIT_R1), src, srcw);
+}
+
+#endif /* __SOFTFP__ */
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_op_flags(struct sljit_compiler *compiler, sljit_s32 op,
+	sljit_s32 dst, sljit_sw dstw,
+	sljit_s32 type)
+{
+	sljit_s32 dst_reg, flags = GET_ALL_FLAGS(op);
+	sljit_ins cc, ins;
+
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_op_flags(compiler, op, dst, dstw, type));
+	ADJUST_LOCAL_OFFSET(dst, dstw);
+
+	op = GET_OPCODE(op);
+	cc = get_cc(compiler, type);
+	dst_reg = FAST_IS_REG(dst) ? dst : TMP_REG1;
+
+	if (op < SLJIT_ADD) {
+		FAIL_IF(push_inst(compiler, MOV | RD(dst_reg) | SRC2_IMM | 0));
+		FAIL_IF(push_inst(compiler, ((MOV | RD(dst_reg) | SRC2_IMM | 1) & ~COND_MASK) | cc));
+		if (dst & SLJIT_MEM)
+			return emit_op_mem(compiler, WORD_SIZE, TMP_REG1, dst, dstw, TMP_REG2);
+		return SLJIT_SUCCESS;
+	}
+
+	ins = (op == SLJIT_AND ? AND : (op == SLJIT_OR ? ORR : EOR));
+
+	if (dst & SLJIT_MEM)
+		FAIL_IF(emit_op_mem(compiler, WORD_SIZE | LOAD_DATA, TMP_REG1, dst, dstw, TMP_REG2));
+
+	FAIL_IF(push_inst(compiler, ((ins | RD(dst_reg) | RN(dst_reg) | SRC2_IMM | 1) & ~COND_MASK) | cc));
+
+	if (op == SLJIT_AND)
+		FAIL_IF(push_inst(compiler, ((ins | RD(dst_reg) | RN(dst_reg) | SRC2_IMM | 0) & ~COND_MASK) | (cc ^ 0x10000000)));
+
+	if (dst & SLJIT_MEM)
+		FAIL_IF(emit_op_mem(compiler, WORD_SIZE, TMP_REG1, dst, dstw, TMP_REG2));
+
+	if (flags & SLJIT_SET_Z)
+		return push_inst(compiler, MOV | SET_FLAGS | RD(TMP_REG2) | RM(dst_reg));
+	return SLJIT_SUCCESS;
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_select(struct sljit_compiler *compiler, sljit_s32 type,
+	sljit_s32 dst_reg,
+	sljit_s32 src1, sljit_sw src1w,
+	sljit_s32 src2_reg)
+{
+	sljit_ins cc, tmp;
+
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_select(compiler, type, dst_reg, src1, src1w, src2_reg));
+
+	ADJUST_LOCAL_OFFSET(src1, src1w);
+
+	if (src2_reg != dst_reg && src1 == dst_reg) {
+		src1 = src2_reg;
+		src1w = 0;
+		src2_reg = dst_reg;
+		type ^= 0x1;
+	}
+
+	if (src1 & SLJIT_MEM) {
+		FAIL_IF(emit_op_mem(compiler, WORD_SIZE | LOAD_DATA, (src2_reg != dst_reg) ? dst_reg : TMP_REG1, src1, src1w, TMP_REG1));
+
+		if (src2_reg != dst_reg) {
+			src1 = src2_reg;
+			src1w = 0;
+			type ^= 0x1;
+		} else {
+			src1 = TMP_REG1;
+			src1w = 0;
+		}
+	} else if (dst_reg != src2_reg)
+		FAIL_IF(push_inst(compiler, MOV | RD(dst_reg) | RM(src2_reg)));
+
+	cc = get_cc(compiler, type & ~SLJIT_32);
+
+	if (SLJIT_UNLIKELY(src1 == SLJIT_IMM)) {
+		tmp = get_imm((sljit_uw)src1w);
+		if (tmp)
+			return push_inst(compiler, ((MOV | RD(dst_reg) | tmp) & ~COND_MASK) | cc);
+
+		tmp = get_imm(~(sljit_uw)src1w);
+		if (tmp)
+			return push_inst(compiler, ((MVN | RD(dst_reg) | tmp) & ~COND_MASK) | cc);
+
+#if (defined SLJIT_CONFIG_ARM_V7 && SLJIT_CONFIG_ARM_V7)
+		tmp = (sljit_ins)src1w;
+		FAIL_IF(push_inst(compiler, (MOVW & ~COND_MASK) | cc | RD(dst_reg) | ((tmp << 4) & 0xf0000) | (tmp & 0xfff)));
+		if (tmp <= 0xffff)
+			return SLJIT_SUCCESS;
+		return push_inst(compiler, (MOVT & ~COND_MASK) | cc | RD(dst_reg) | ((tmp >> 12) & 0xf0000) | ((tmp >> 16) & 0xfff));
+#else /* !SLJIT_CONFIG_ARM_V7 */
+		FAIL_IF(load_immediate(compiler, TMP_REG1, (sljit_uw)src1w));
+		src1 = TMP_REG1;
+#endif /* SLJIT_CONFIG_ARM_V7 */
+	}
+
+	return push_inst(compiler, ((MOV | RD(dst_reg) | RM(src1)) & ~COND_MASK) | cc);
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_fselect(struct sljit_compiler *compiler, sljit_s32 type,
+	sljit_s32 dst_freg,
+	sljit_s32 src1, sljit_sw src1w,
+	sljit_s32 src2_freg)
+{
+	sljit_ins cc;
+
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_fselect(compiler, type, dst_freg, src1, src1w, src2_freg));
+
+	ADJUST_LOCAL_OFFSET(src1, src1w);
+
+	type ^= SLJIT_32;
+
+	if (dst_freg != src2_freg) {
+		if (dst_freg == src1) {
+			src1 = src2_freg;
+			src1w = 0;
+			type ^= 0x1;
+		} else
+			FAIL_IF(push_inst(compiler, EMIT_FPU_OPERATION(VMOV_F32, (type & SLJIT_32), dst_freg, src2_freg, 0)));
+	}
+
+	if (src1 & SLJIT_MEM) {
+		FAIL_IF(emit_fop_mem(compiler, (type & SLJIT_32) | FPU_LOAD, TMP_FREG2, src1, src1w));
+		src1 = TMP_FREG2;
+	}
+
+	cc = get_cc(compiler, type & ~SLJIT_32);
+	return push_inst(compiler, EMIT_FPU_OPERATION((VMOV_F32 & ~COND_MASK) | cc, (type & SLJIT_32), dst_freg, src1, 0));
+}
+
+#undef EMIT_FPU_OPERATION
+
+static sljit_s32 update_mem_addr(struct sljit_compiler *compiler, sljit_s32 *mem, sljit_sw *memw, sljit_s32 max_offset)
+{
+	sljit_s32 arg = *mem;
+	sljit_sw argw = *memw;
+	sljit_uw imm, tmp;
+	sljit_sw mask = 0xfff;
+	sljit_sw sign = 0x1000;
+
+	SLJIT_ASSERT(max_offset >= 0xf00);
+
+	*mem = TMP_REG1;
+
+	if (SLJIT_UNLIKELY(arg & OFFS_REG_MASK)) {
+		*memw = 0;
+		return push_inst(compiler, ADD | RD(TMP_REG1) | RN(arg & REG_MASK) | RM(OFFS_REG(arg)) | ((sljit_ins)(argw & 0x3) << 7));
+	}
+
+	arg &= REG_MASK;
+
+	if (arg) {
+		if (argw <= max_offset && argw >= -mask) {
+			*mem = arg;
+			return SLJIT_SUCCESS;
+		}
+
+		if (argw >= 0) {
+			tmp = (sljit_uw)(argw & (sign | mask));
+			tmp = (sljit_uw)((argw + ((tmp <= (sljit_uw)max_offset || tmp == (sljit_uw)sign) ? 0 : sign)) & ~mask);
+			imm = get_imm(tmp);
+
+			if (imm) {
+				*memw = argw - (sljit_sw)tmp;
+				SLJIT_ASSERT(*memw >= -mask && *memw <= max_offset);
+
+				return push_inst(compiler, ADD | RD(TMP_REG1) | RN(arg) | imm);
+			}
+		} else {
+			tmp = (sljit_uw)(-argw & (sign | mask));
+			tmp = (sljit_uw)((-argw + ((tmp <= (sljit_uw)((sign << 1) - max_offset - 1)) ? 0 : sign)) & ~mask);
+			imm = get_imm(tmp);
+
+			if (imm) {
+				*memw = argw + (sljit_sw)tmp;
+				SLJIT_ASSERT(*memw >= -mask && *memw <= max_offset);
+
+				return push_inst(compiler, SUB | RD(TMP_REG1) | RN(arg) | imm);
+			}
+		}
+	}
+
+	tmp = (sljit_uw)(argw & (sign | mask));
+	tmp = (sljit_uw)((argw + ((tmp <= (sljit_uw)max_offset || tmp == (sljit_uw)sign) ? 0 : sign)) & ~mask);
+	*memw = argw - (sljit_sw)tmp;
+
+	FAIL_IF(load_immediate(compiler, TMP_REG1, tmp));
+
+	if (arg == 0)
+		return SLJIT_SUCCESS;
+
+	return push_inst(compiler, ADD | RD(TMP_REG1) | RN(TMP_REG1) | RM(arg));
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_mem(struct sljit_compiler *compiler, sljit_s32 type,
+	sljit_s32 reg,
+	sljit_s32 mem, sljit_sw memw)
+{
+	sljit_s32 flags;
+
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_mem(compiler, type, reg, mem, memw));
+
+	if (!(reg & REG_PAIR_MASK))
+		return sljit_emit_mem_unaligned(compiler, type, reg, mem, memw);
+
+	ADJUST_LOCAL_OFFSET(mem, memw);
+
+	FAIL_IF(update_mem_addr(compiler, &mem, &memw, 0xfff - 4));
+
+	flags = WORD_SIZE;
+
+	if (!(type & SLJIT_MEM_STORE)) {
+		if (REG_PAIR_FIRST(reg) == (mem & REG_MASK)) {
+			FAIL_IF(emit_op_mem(compiler, WORD_SIZE | LOAD_DATA, REG_PAIR_SECOND(reg), SLJIT_MEM1(mem), memw + SSIZE_OF(sw), TMP_REG1));
+			return emit_op_mem(compiler, WORD_SIZE | LOAD_DATA, REG_PAIR_FIRST(reg), SLJIT_MEM1(mem), memw, TMP_REG1);
+		}
+
+		flags = WORD_SIZE | LOAD_DATA;
+	}
+
+	FAIL_IF(emit_op_mem(compiler, flags, REG_PAIR_FIRST(reg), SLJIT_MEM1(mem), memw, TMP_REG1));
+	return emit_op_mem(compiler, flags, REG_PAIR_SECOND(reg), SLJIT_MEM1(mem), memw + SSIZE_OF(sw), TMP_REG1);
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_mem_update(struct sljit_compiler *compiler, sljit_s32 type,
+	sljit_s32 reg,
+	sljit_s32 mem, sljit_sw memw)
+{
+	sljit_s32 flags;
+	sljit_ins is_type1_transfer, inst;
+
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_mem_update(compiler, type, reg, mem, memw));
+
+	is_type1_transfer = 1;
+
+	switch (type & 0xff) {
+	case SLJIT_MOV:
+	case SLJIT_MOV_U32:
+	case SLJIT_MOV_S32:
+	case SLJIT_MOV32:
+	case SLJIT_MOV_P:
+		flags = WORD_SIZE;
+		break;
+	case SLJIT_MOV_U8:
+		flags = BYTE_SIZE;
+		break;
+	case SLJIT_MOV_S8:
+		if (!(type & SLJIT_MEM_STORE))
+			is_type1_transfer = 0;
+		flags = BYTE_SIZE | SIGNED;
+		break;
+	case SLJIT_MOV_U16:
+		is_type1_transfer = 0;
+		flags = HALF_SIZE;
+		break;
+	case SLJIT_MOV_S16:
+		is_type1_transfer = 0;
+		flags = HALF_SIZE | SIGNED;
+		break;
+	default:
+		SLJIT_UNREACHABLE();
+		flags = WORD_SIZE;
+		break;
+	}
+
+	if (!(type & SLJIT_MEM_STORE))
+		flags |= LOAD_DATA;
+
+	SLJIT_ASSERT(is_type1_transfer == !!IS_TYPE1_TRANSFER(flags));
+
+	if (SLJIT_UNLIKELY(mem & OFFS_REG_MASK)) {
+		if (!is_type1_transfer && memw != 0)
+			return SLJIT_ERR_UNSUPPORTED;
+	} else {
+		if (is_type1_transfer) {
+			if (memw > 4095 || memw < -4095)
+				return SLJIT_ERR_UNSUPPORTED;
+		} else if (memw > 255 || memw < -255)
+			return SLJIT_ERR_UNSUPPORTED;
+	}
+
+	if (type & SLJIT_MEM_SUPP)
+		return SLJIT_SUCCESS;
+
+	if (SLJIT_UNLIKELY(mem & OFFS_REG_MASK)) {
+		memw &= 0x3;
+
+		inst = EMIT_DATA_TRANSFER(flags, 1, reg, mem & REG_MASK, RM(OFFS_REG(mem)) | ((sljit_ins)memw << 7));
+
+		if (is_type1_transfer)
+			inst |= (1 << 25);
+
+		if (type & SLJIT_MEM_POST)
+			inst ^= (1 << 24);
+		else
+			inst |= (1 << 21);
+
+		return push_inst(compiler, inst);
+	}
+
+	inst = EMIT_DATA_TRANSFER(flags, 0, reg, mem & REG_MASK, 0);
+
+	if (type & SLJIT_MEM_POST)
+		inst ^= (1 << 24);
+	else
+		inst |= (1 << 21);
+
+	if (is_type1_transfer) {
+		if (memw >= 0)
+			inst |= (1 << 23);
+		else
+			memw = -memw;
+
+		return push_inst(compiler, inst | (sljit_ins)memw);
+	}
+
+	if (memw >= 0)
+		inst |= (1 << 23);
+	else
+		memw = -memw;
+
+	return push_inst(compiler, inst | TYPE2_TRANSFER_IMM((sljit_ins)memw));
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_fmem(struct sljit_compiler *compiler, sljit_s32 type,
+	sljit_s32 freg,
+	sljit_s32 mem, sljit_sw memw)
+{
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_fmem(compiler, type, freg, mem, memw));
+
+	if (type & SLJIT_MEM_ALIGNED_32)
+		return emit_fop_mem(compiler, ((type ^ SLJIT_32) & SLJIT_32) | ((type & SLJIT_MEM_STORE) ? 0 : FPU_LOAD), freg, mem, memw);
+
+	if (type & SLJIT_MEM_STORE) {
+		FAIL_IF(push_inst(compiler, VMOV | (1 << 20) | VN(freg) | RD(TMP_REG2)));
+
+		if (type & SLJIT_32)
+			return emit_op_mem(compiler, WORD_SIZE, TMP_REG2, mem, memw, TMP_REG1);
+
+		FAIL_IF(update_mem_addr(compiler, &mem, &memw, 0xfff - 4));
+		mem |= SLJIT_MEM;
+
+		FAIL_IF(emit_op_mem(compiler, WORD_SIZE, TMP_REG2, mem, memw, TMP_REG1));
+		FAIL_IF(push_inst(compiler, VMOV | (1 << 20) | VN(freg) | 0x80 | RD(TMP_REG2)));
+		return emit_op_mem(compiler, WORD_SIZE, TMP_REG2, mem, memw + 4, TMP_REG1);
+	}
+
+	if (type & SLJIT_32) {
+		FAIL_IF(emit_op_mem(compiler, WORD_SIZE | LOAD_DATA, TMP_REG2, mem, memw, TMP_REG1));
+		return push_inst(compiler, VMOV | VN(freg) | RD(TMP_REG2));
+	}
+
+	FAIL_IF(update_mem_addr(compiler, &mem, &memw, 0xfff - 4));
+	mem |= SLJIT_MEM;
+
+	FAIL_IF(emit_op_mem(compiler, WORD_SIZE | LOAD_DATA, TMP_REG2, mem, memw, TMP_REG1));
+	FAIL_IF(emit_op_mem(compiler, WORD_SIZE | LOAD_DATA, TMP_REG1, mem, memw + 4, TMP_REG1));
+	return push_inst(compiler, VMOV2 | VM(freg) | RD(TMP_REG2) | RN(TMP_REG1));
+}
+
+static sljit_s32 sljit_emit_simd_mem_offset(struct sljit_compiler *compiler, sljit_s32 *mem_ptr, sljit_sw memw)
+{
+	sljit_s32 mem = *mem_ptr;
+	sljit_uw imm;
+
+	if (SLJIT_UNLIKELY(mem & OFFS_REG_MASK)) {
+		*mem_ptr = TMP_REG1;
+		return push_inst(compiler, ADD | RD(TMP_REG1) | RN(mem & REG_MASK) | RM(OFFS_REG(mem)) | ((sljit_ins)(memw & 0x3) << 7));
+	}
+
+	if (SLJIT_UNLIKELY(!(mem & REG_MASK))) {
+		*mem_ptr = TMP_REG1;
+		return load_immediate(compiler, TMP_REG1, (sljit_uw)memw);
+	}
+
+	mem &= REG_MASK;
+
+	if (memw == 0) {
+		*mem_ptr = mem;
+		return SLJIT_SUCCESS;
+	}
+
+	*mem_ptr = TMP_REG1;
+	imm = get_imm((sljit_uw)(memw < 0 ? -memw : memw));
+
+	if (imm != 0)
+		return push_inst(compiler, ((memw < 0) ? SUB : ADD) | RD(TMP_REG1) | RN(mem) | imm);
+
+	FAIL_IF(load_immediate(compiler, TMP_REG1, (sljit_uw)memw));
+	return push_inst(compiler, ADD | RD(TMP_REG1) | RN(TMP_REG1) | RM(mem));
+}
+
+static SLJIT_INLINE sljit_s32 simd_get_quad_reg_index(sljit_s32 freg)
+{
+	freg += freg & 0x1;
+
+	SLJIT_ASSERT((freg_map[freg] & 0x1) == (freg <= SLJIT_NUMBER_OF_SCRATCH_FLOAT_REGISTERS));
+
+	if (freg <= SLJIT_NUMBER_OF_SCRATCH_FLOAT_REGISTERS)
+		freg--;
+
+	return freg;
+}
+
+#define SLJIT_QUAD_OTHER_HALF(freg) ((((freg) & 0x1) << 1) - 1)
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_simd_mov(struct sljit_compiler *compiler, sljit_s32 type,
+	sljit_s32 freg,
+	sljit_s32 srcdst, sljit_sw srcdstw)
+{
+	sljit_s32 reg_size = SLJIT_SIMD_GET_REG_SIZE(type);
+	sljit_s32 elem_size = SLJIT_SIMD_GET_ELEM_SIZE(type);
+	sljit_s32 alignment = SLJIT_SIMD_GET_ELEM2_SIZE(type);
+	sljit_ins ins;
+
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_simd_mov(compiler, type, freg, srcdst, srcdstw));
+
+	ADJUST_LOCAL_OFFSET(srcdst, srcdstw);
+
+	if (reg_size != 3 && reg_size != 4)
+		return SLJIT_ERR_UNSUPPORTED;
+
+	if ((type & SLJIT_SIMD_FLOAT) && (elem_size < 2 || elem_size > 3))
+		return SLJIT_ERR_UNSUPPORTED;
+
+	if (type & SLJIT_SIMD_TEST)
+		return SLJIT_SUCCESS;
+
+	if (reg_size == 4)
+		freg = simd_get_quad_reg_index(freg);
+
+	if (!(srcdst & SLJIT_MEM)) {
+		if (reg_size == 4)
+			srcdst = simd_get_quad_reg_index(srcdst);
+
+		if (type & SLJIT_SIMD_STORE)
+			ins = VD(srcdst) | VN(freg) | VM(freg);
+		else
+			ins = VD(freg) | VN(srcdst) | VM(srcdst);
+
+		if (reg_size == 4)
+			ins |= (sljit_ins)1 << 6;
+
+		return push_inst(compiler, VORR | ins);
+	}
+
+	FAIL_IF(sljit_emit_simd_mem_offset(compiler, &srcdst, srcdstw));
+
+	if (elem_size > 3)
+		elem_size = 3;
+
+	ins = ((type & SLJIT_SIMD_STORE) ? VST1 : VLD1) | VD(freg)
+		| (sljit_ins)((reg_size == 3) ? (0x7 << 8) : (0xa << 8));
+
+	SLJIT_ASSERT(reg_size >= alignment);
+
+	if (alignment == 3)
+		ins |= 0x10;
+	else if (alignment >= 3)
+		ins |= 0x20;
+
+	return push_inst(compiler, ins | RN(srcdst) | ((sljit_ins)elem_size) << 6 | 0xf);
+}
+
+static sljit_ins simd_get_imm(sljit_s32 elem_size, sljit_uw value)
+{
+	sljit_ins result;
+
+	if (elem_size > 1 && (sljit_u16)value == (value >> 16)) {
+		elem_size = 1;
+		value = (sljit_u16)value;
+	}
+
+	if (elem_size == 1 && (sljit_u8)value == (value >> 8)) {
+		elem_size = 0;
+		value = (sljit_u8)value;
+	}
+
+	switch (elem_size) {
+	case 0:
+		SLJIT_ASSERT(value <= 0xff);
+		result = 0xe00;
+		break;
+	case 1:
+		SLJIT_ASSERT(value <= 0xffff);
+		result = 0;
+
+		while (1) {
+			if (value <= 0xff) {
+				result |= 0x800;
+				break;
+			}
+
+			if ((value & 0xff) == 0) {
+				value >>= 8;
+				result |= 0xa00;
+				break;
+			}
+
+			if (result != 0)
+				return ~(sljit_ins)0;
+
+			value ^= (sljit_uw)0xffff;
+			result = (1 << 5);
+		}
+		break;
+	default:
+		SLJIT_ASSERT(value <= 0xffffffff);
+		result = 0;
+
+		while (1) {
+			if (value <= 0xff) {
+				result |= 0x000;
+				break;
+			}
+
+			if ((value & ~(sljit_uw)0xff00) == 0) {
+				value >>= 8;
+				result |= 0x200;
+				break;
+			}
+
+			if ((value & ~(sljit_uw)0xff0000) == 0) {
+				value >>= 16;
+				result |= 0x400;
+				break;
+			}
+
+			if ((value & ~(sljit_uw)0xff000000) == 0) {
+				value >>= 24;
+				result |= 0x600;
+				break;
+			}
+
+			if ((value & (sljit_uw)0xff) == 0xff && (value >> 16) == 0) {
+				value >>= 8;
+				result |= 0xc00;
+				break;
+			}
+
+			if ((value & (sljit_uw)0xffff) == 0xffff && (value >> 24) == 0) {
+				value >>= 16;
+				result |= 0xd00;
+				break;
+			}
+
+			if (result != 0)
+				return ~(sljit_ins)0;
+
+			value = ~value;
+			result = (1 << 5);
+		}
+		break;
+	}
+
+	return ((sljit_ins)value & 0xf) | (((sljit_ins)value & 0x70) << 12) | (((sljit_ins)value & 0x80) << 17) | result;
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_simd_replicate(struct sljit_compiler *compiler, sljit_s32 type,
+	sljit_s32 freg,
+	sljit_s32 src, sljit_sw srcw)
+{
+	sljit_s32 reg_size = SLJIT_SIMD_GET_REG_SIZE(type);
+	sljit_s32 elem_size = SLJIT_SIMD_GET_ELEM_SIZE(type);
+	sljit_ins ins, imm;
+
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_simd_replicate(compiler, type, freg, src, srcw));
+
+	ADJUST_LOCAL_OFFSET(src, srcw);
+
+	if (reg_size != 3 && reg_size != 4)
+		return SLJIT_ERR_UNSUPPORTED;
+
+	if ((type & SLJIT_SIMD_FLOAT) ? (elem_size < 2 || elem_size > 3) : (elem_size > 2))
+		return SLJIT_ERR_UNSUPPORTED;
+
+	if (type & SLJIT_SIMD_TEST)
+		return SLJIT_SUCCESS;
+
+	if (reg_size == 4)
+		freg = simd_get_quad_reg_index(freg);
+
+	if (src == SLJIT_IMM && srcw == 0)
+		return push_inst(compiler, VMOV_i | ((reg_size == 4) ? (1 << 6) : 0) | VD(freg));
+
+	if (SLJIT_UNLIKELY(elem_size == 3)) {
+		SLJIT_ASSERT(type & SLJIT_SIMD_FLOAT);
+
+		if (src & SLJIT_MEM) {
+			FAIL_IF(emit_fop_mem(compiler, FPU_LOAD | SLJIT_32, freg, src, srcw));
+			src = freg;
+		} else if (freg != src)
+			FAIL_IF(push_inst(compiler, VORR | VD(freg) | VN(src) | VM(src)));
+
+		freg += SLJIT_QUAD_OTHER_HALF(freg);
+
+		if (freg != src)
+			return push_inst(compiler, VORR | VD(freg) | VN(src) | VM(src));
+		return SLJIT_SUCCESS;
+	}
+
+	if (src & SLJIT_MEM) {
+		FAIL_IF(sljit_emit_simd_mem_offset(compiler, &src, srcw));
+
+		ins = (sljit_ins)(elem_size << 6);
+
+		if (reg_size == 4)
+			ins |= (sljit_ins)1 << 5;
+
+		return push_inst(compiler, VLD1_r | ins | VD(freg) | RN(src) | 0xf);
+	}
+
+	if (type & SLJIT_SIMD_FLOAT) {
+		SLJIT_ASSERT(elem_size == 2);
+		ins = ((sljit_ins)freg_ebit_map[src] << (16 + 2 + 1)) | ((sljit_ins)1 << (16 + 2));
+
+		if (reg_size == 4)
+			ins |= (sljit_ins)1 << 6;
+
+		return push_inst(compiler, VDUP_s | ins | VD(freg) | (sljit_ins)freg_map[src]);
+	}
+
+	if (src == SLJIT_IMM) {
+		if (elem_size < 2)
+			srcw &= ((sljit_sw)1 << (((sljit_sw)1 << elem_size) << 3)) - 1;
+
+		imm = simd_get_imm(elem_size, (sljit_uw)srcw);
+
+		if (imm != ~(sljit_ins)0) {
+			if (reg_size == 4)
+				imm |= (sljit_ins)1 << 6;
+
+			return push_inst(compiler, VMOV_i | imm | VD(freg));
+		}
+
+		FAIL_IF(load_immediate(compiler, TMP_REG1, (sljit_uw)srcw));
+		src = TMP_REG1;
+	}
+
+	switch (elem_size) {
+	case 0:
+		ins = 1 << 22;
+		break;
+	case 1:
+		ins = 1 << 5;
+		break;
+	default:
+		ins = 0;
+		break;
+	}
+
+	if (reg_size == 4)
+		ins |= (sljit_ins)1 << 21;
+
+	return push_inst(compiler, VDUP | ins | VN(freg) | RD(src));
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_simd_lane_mov(struct sljit_compiler *compiler, sljit_s32 type,
+	sljit_s32 freg, sljit_s32 lane_index,
+	sljit_s32 srcdst, sljit_sw srcdstw)
+{
+	sljit_s32 reg_size = SLJIT_SIMD_GET_REG_SIZE(type);
+	sljit_s32 elem_size = SLJIT_SIMD_GET_ELEM_SIZE(type);
+	sljit_ins ins;
+
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_simd_lane_mov(compiler, type, freg, lane_index, srcdst, srcdstw));
+
+	ADJUST_LOCAL_OFFSET(srcdst, srcdstw);
+
+	if (reg_size != 3 && reg_size != 4)
+		return SLJIT_ERR_UNSUPPORTED;
+
+	if ((type & SLJIT_SIMD_FLOAT) ? (elem_size < 2 || elem_size > 3) : (elem_size > 2))
+		return SLJIT_ERR_UNSUPPORTED;
+
+	if (type & SLJIT_SIMD_TEST)
+		return SLJIT_SUCCESS;
+
+	if (reg_size == 4)
+		freg = simd_get_quad_reg_index(freg);
+
+	if (type & SLJIT_SIMD_LANE_ZERO) {
+		ins = (reg_size == 3) ? 0 : ((sljit_ins)1 << 6);
+
+		if (type & SLJIT_SIMD_FLOAT) {
+			if (elem_size == 3 && !(srcdst & SLJIT_MEM)) {
+				if (lane_index == 1)
+					freg += SLJIT_QUAD_OTHER_HALF(freg);
+
+				if (srcdst != freg)
+					FAIL_IF(push_inst(compiler, VORR | VD(freg) | VN(srcdst) | VM(srcdst)));
+
+				freg += SLJIT_QUAD_OTHER_HALF(freg);
+				return push_inst(compiler, VMOV_i | VD(freg));
+			}
+
+			if (srcdst == freg || (elem_size == 3 && srcdst == (freg + SLJIT_QUAD_OTHER_HALF(freg)))) {
+				FAIL_IF(push_inst(compiler, VORR | ins | VD(TMP_FREG2) | VN(freg) | VM(freg)));
+				srcdst = TMP_FREG2;
+				srcdstw = 0;
+			}
+		}
+
+		FAIL_IF(push_inst(compiler, VMOV_i | ins | VD(freg)));
+	}
+
+	if (reg_size == 4 && lane_index >= (0x8 >> elem_size)) {
+		lane_index -= (0x8 >> elem_size);
+		freg += SLJIT_QUAD_OTHER_HALF(freg);
+	}
+
+	if (srcdst & SLJIT_MEM) {
+		if (elem_size == 3)
+			return emit_fop_mem(compiler, ((type & SLJIT_SIMD_STORE) ? 0 : FPU_LOAD) | SLJIT_32, freg, srcdst, srcdstw);
+
+		FAIL_IF(sljit_emit_simd_mem_offset(compiler, &srcdst, srcdstw));
+
+		lane_index = lane_index << elem_size;
+		ins = (sljit_ins)((elem_size << 10) | (lane_index << 5));
+		return push_inst(compiler, ((type & SLJIT_SIMD_STORE) ? VST1_s : VLD1_s) | ins | VD(freg) | RN(srcdst) | 0xf);
+	}
+
+	if (type & SLJIT_SIMD_FLOAT) {
+		if (elem_size == 3) {
+			if (type & SLJIT_SIMD_STORE)
+				return push_inst(compiler, VORR | VD(srcdst) | VN(freg) | VM(freg));
+			return push_inst(compiler, VMOV_F32 | SLJIT_32 | VD(freg) | VM(srcdst));
+		}
+
+		if (type & SLJIT_SIMD_STORE) {
+			if (freg_ebit_map[freg] == 0) {
+				if (lane_index == 1)
+					freg = SLJIT_F64_SECOND(freg);
+
+				return push_inst(compiler, VMOV_F32 | VD(srcdst) | VM(freg));
+			}
+
+			FAIL_IF(push_inst(compiler, VMOV_s | (1 << 20) | ((sljit_ins)lane_index << 21) | VN(freg) | RD(TMP_REG1)));
+			return push_inst(compiler, VMOV | VN(srcdst) | RD(TMP_REG1));
+		}
+
+		FAIL_IF(push_inst(compiler, VMOV | (1 << 20) | VN(srcdst) | RD(TMP_REG1)));
+		return push_inst(compiler, VMOV_s | ((sljit_ins)lane_index << 21) | VN(freg) | RD(TMP_REG1));
+	}
+
+	if (srcdst == SLJIT_IMM) {
+		if (elem_size < 2)
+			srcdstw &= ((sljit_sw)1 << (((sljit_sw)1 << elem_size) << 3)) - 1;
+
+		FAIL_IF(load_immediate(compiler, TMP_REG1, (sljit_uw)srcdstw));
+		srcdst = TMP_REG1;
+	}
+
+	if (elem_size == 0)
+		ins = 0x400000;
+	else if (elem_size == 1)
+		ins = 0x20;
+	else
+		ins = 0;
+
+	lane_index = lane_index << elem_size;
+	ins |= (sljit_ins)(((lane_index & 0x4) << 19) | ((lane_index & 0x3) << 5));
+
+	if (type & SLJIT_SIMD_STORE) {
+		ins |= (1 << 20);
+
+		if (elem_size < 2 && !(type & SLJIT_SIMD_LANE_SIGNED))
+			ins |= (1 << 23);
+	}
+
+	return push_inst(compiler, VMOV_s | ins | VN(freg) | RD(srcdst));
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_simd_lane_replicate(struct sljit_compiler *compiler, sljit_s32 type,
+	sljit_s32 freg,
+	sljit_s32 src, sljit_s32 src_lane_index)
+{
+	sljit_s32 reg_size = SLJIT_SIMD_GET_REG_SIZE(type);
+	sljit_s32 elem_size = SLJIT_SIMD_GET_ELEM_SIZE(type);
+	sljit_ins ins;
+
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_simd_lane_replicate(compiler, type, freg, src, src_lane_index));
+
+	if (reg_size != 3 && reg_size != 4)
+		return SLJIT_ERR_UNSUPPORTED;
+
+	if ((type & SLJIT_SIMD_FLOAT) && (elem_size < 2 || elem_size > 3))
+		return SLJIT_ERR_UNSUPPORTED;
+
+	if (type & SLJIT_SIMD_TEST)
+		return SLJIT_SUCCESS;
+
+	if (reg_size == 4) {
+		freg = simd_get_quad_reg_index(freg);
+		src = simd_get_quad_reg_index(src);
+
+		if (src_lane_index >= (0x8 >> elem_size)) {
+			src_lane_index -= (0x8 >> elem_size);
+			src += SLJIT_QUAD_OTHER_HALF(src);
+		}
+	}
+
+	if (elem_size == 3) {
+		if (freg != src)
+			FAIL_IF(push_inst(compiler, VORR | VD(freg) | VN(src) | VM(src)));
+
+		freg += SLJIT_QUAD_OTHER_HALF(freg);
+
+		if (freg != src)
+			return push_inst(compiler, VORR | VD(freg) | VN(src) | VM(src));
+		return SLJIT_SUCCESS;
+	}
+
+	ins = ((((sljit_ins)src_lane_index << 1) | 1) << (16 + elem_size));
+
+	if (reg_size == 4)
+		ins |= (sljit_ins)1 << 6;
+
+	return push_inst(compiler, VDUP_s | ins | VD(freg) | VM(src));
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_simd_extend(struct sljit_compiler *compiler, sljit_s32 type,
+	sljit_s32 freg,
+	sljit_s32 src, sljit_sw srcw)
+{
+	sljit_s32 reg_size = SLJIT_SIMD_GET_REG_SIZE(type);
+	sljit_s32 elem_size = SLJIT_SIMD_GET_ELEM_SIZE(type);
+	sljit_s32 elem2_size = SLJIT_SIMD_GET_ELEM2_SIZE(type);
+	sljit_s32 dst_reg;
+
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_simd_extend(compiler, type, freg, src, srcw));
+
+	ADJUST_LOCAL_OFFSET(src, srcw);
+
+	if (reg_size != 3 && reg_size != 4)
+		return SLJIT_ERR_UNSUPPORTED;
+
+	if ((type & SLJIT_SIMD_FLOAT) && (elem_size != 2 || elem2_size != 3))
+		return SLJIT_ERR_UNSUPPORTED;
+
+	if (type & SLJIT_SIMD_TEST)
+		return SLJIT_SUCCESS;
+
+	if (reg_size == 4)
+		freg = simd_get_quad_reg_index(freg);
+
+	if (src & SLJIT_MEM) {
+		FAIL_IF(sljit_emit_simd_mem_offset(compiler, &src, srcw));
+		if (reg_size == 4 && elem2_size - elem_size == 1)
+			FAIL_IF(push_inst(compiler, VLD1 | (0x7 << 8) | VD(freg) | RN(src) | 0xf));
+		else
+			FAIL_IF(push_inst(compiler, VLD1_s | (sljit_ins)((reg_size - elem2_size + elem_size) << 10) | VD(freg) | RN(src) | 0xf));
+		src = freg;
+	} else if (reg_size == 4)
+		src = simd_get_quad_reg_index(src);
+
+	if (!(type & SLJIT_SIMD_FLOAT)) {
+		dst_reg = (reg_size == 4) ? freg : TMP_FREG2;
+
+		do {
+			FAIL_IF(push_inst(compiler, VSHLL | ((type & SLJIT_SIMD_EXTEND_SIGNED) ? 0 : (1 << 24))
+				| ((sljit_ins)1 << (19 + elem_size)) | VD(dst_reg) | VM(src)));
+			src = dst_reg;
+		} while (++elem_size < elem2_size);
+
+		if (dst_reg == TMP_FREG2)
+			return push_inst(compiler, VORR | VD(freg) | VN(TMP_FREG2) | VM(TMP_FREG2));
+		return SLJIT_SUCCESS;
+	}
+
+	/* No SIMD variant, must use VFP instead. */
+	SLJIT_ASSERT(reg_size == 4);
+
+	if (freg == src) {
+		freg += SLJIT_QUAD_OTHER_HALF(freg);
+		FAIL_IF(push_inst(compiler, VCVT_F64_F32 | VD(freg) | VM(src) | 0x20));
+		freg += SLJIT_QUAD_OTHER_HALF(freg);
+		return push_inst(compiler, VCVT_F64_F32 | VD(freg) | VM(src));
+	}
+
+	FAIL_IF(push_inst(compiler, VCVT_F64_F32 | VD(freg) | VM(src)));
+	freg += SLJIT_QUAD_OTHER_HALF(freg);
+	return push_inst(compiler, VCVT_F64_F32 | VD(freg) | VM(src) | 0x20);
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_simd_sign(struct sljit_compiler *compiler, sljit_s32 type,
+	sljit_s32 freg,
+	sljit_s32 dst, sljit_sw dstw)
+{
+	sljit_s32 reg_size = SLJIT_SIMD_GET_REG_SIZE(type);
+	sljit_s32 elem_size = SLJIT_SIMD_GET_ELEM_SIZE(type);
+	sljit_ins ins, imms;
+	sljit_s32 dst_r;
+
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_simd_sign(compiler, type, freg, dst, dstw));
+
+	ADJUST_LOCAL_OFFSET(dst, dstw);
+
+	if (reg_size != 3 && reg_size != 4)
+		return SLJIT_ERR_UNSUPPORTED;
+
+	if ((type & SLJIT_SIMD_FLOAT) && (elem_size < 2 || elem_size > 3))
+		return SLJIT_ERR_UNSUPPORTED;
+
+	if (type & SLJIT_SIMD_TEST)
+		return SLJIT_SUCCESS;
+
+	switch (elem_size) {
+	case 0:
+		imms = 0x243219;
+		ins = VSHR | (1 << 24) | (0x9 << 16);
+		break;
+	case 1:
+		imms = (reg_size == 4) ? 0x243219 : 0x2231;
+		ins = VSHR | (1 << 24) | (0x11 << 16);
+		break;
+	case 2:
+		imms = (reg_size == 4) ? 0x2231 : 0x21;
+		ins = VSHR | (1 << 24) | (0x21 << 16);
+		break;
+	default:
+		imms = 0x21;
+		ins = VSHR | (1 << 24) | (0x1 << 16) | (1 << 7);
+		break;
+	}
+
+	if (reg_size == 4) {
+		freg = simd_get_quad_reg_index(freg);
+		ins |= (sljit_ins)1 << 6;
+	}
+
+	SLJIT_ASSERT((freg_map[TMP_FREG2] & 0x1) == 0);
+	FAIL_IF(push_inst(compiler, ins | VD(TMP_FREG2) | VM(freg)));
+
+	if (reg_size == 4 && elem_size > 0)
+		FAIL_IF(push_inst(compiler, VMOVN | ((sljit_ins)(elem_size - 1) << 18) | VD(TMP_FREG2) | VM(TMP_FREG2)));
+
+	ins = (reg_size == 4 && elem_size == 0) ? (1 << 6) : 0;
+
+	while (imms >= 0x100) {
+		FAIL_IF(push_inst(compiler, VSRA | (1 << 24) | ins | ((imms & 0xff) << 16) | VD(TMP_FREG2) | VM(TMP_FREG2)));
+		imms >>= 8;
+	}
+
+	FAIL_IF(push_inst(compiler, VSRA | (1 << 24) | ins | (1 << 7) | (imms << 16) | VD(TMP_FREG2) | VM(TMP_FREG2)));
+
+	dst_r = FAST_IS_REG(dst) ? dst : TMP_REG1;
+	FAIL_IF(push_inst(compiler, VMOV_s | (1 << 20) | (1 << 23) | (0x2 << 21) | RD(dst_r) | VN(TMP_FREG2)));
+
+	if (reg_size == 4 && elem_size == 0) {
+		SLJIT_ASSERT(freg_map[TMP_FREG2] + 1 == freg_map[TMP_FREG1]);
+		FAIL_IF(push_inst(compiler, VMOV_s | (1 << 20) | (1 << 23) | (0x2 << 21) | RD(TMP_REG2) | VN(TMP_FREG1)));
+		FAIL_IF(push_inst(compiler, ORR | RD(dst_r) | RN(dst_r) | RM(TMP_REG2) | (0x8 << 7)));
+	}
+
+	if (dst_r == TMP_REG1)
+		return emit_op_mem(compiler, WORD_SIZE, TMP_REG1, dst, dstw, TMP_REG2);
+
+	return SLJIT_SUCCESS;
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_simd_op2(struct sljit_compiler *compiler, sljit_s32 type,
+	sljit_s32 dst_freg, sljit_s32 src1_freg, sljit_s32 src2_freg)
+{
+	sljit_s32 reg_size = SLJIT_SIMD_GET_REG_SIZE(type);
+	sljit_s32 elem_size = SLJIT_SIMD_GET_ELEM_SIZE(type);
+	sljit_ins ins = 0;
+
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_simd_op2(compiler, type, dst_freg, src1_freg, src2_freg));
+
+	if (reg_size != 3 && reg_size != 4)
+		return SLJIT_ERR_UNSUPPORTED;
+
+	if ((type & SLJIT_SIMD_FLOAT) && (elem_size < 2 || elem_size > 3))
+		return SLJIT_ERR_UNSUPPORTED;
+
+	switch (SLJIT_SIMD_GET_OPCODE(type)) {
+	case SLJIT_SIMD_OP2_AND:
+		ins = VAND;
+		break;
+	case SLJIT_SIMD_OP2_OR:
+		ins = VORR;
+		break;
+	case SLJIT_SIMD_OP2_XOR:
+		ins = VEOR;
+		break;
+	}
+
+	if (type & SLJIT_SIMD_TEST)
+		return SLJIT_SUCCESS;
+
+	if (reg_size == 4) {
+		dst_freg = simd_get_quad_reg_index(dst_freg);
+		src1_freg = simd_get_quad_reg_index(src1_freg);
+		src2_freg = simd_get_quad_reg_index(src2_freg);
+		ins |= (sljit_ins)1 << 6;
+	}
+
+	return push_inst(compiler, ins | VD(dst_freg) | VN(src1_freg) | VM(src2_freg));
+}
+
+#undef FPU_LOAD
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_atomic_load(struct sljit_compiler *compiler, sljit_s32 op,
+	sljit_s32 dst_reg,
+	sljit_s32 mem_reg)
+{
+	sljit_u32 ins;
+
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_atomic_load(compiler, op, dst_reg, mem_reg));
+
+	switch (GET_OPCODE(op)) {
+	case SLJIT_MOV_U8:
+		ins = LDREXB;
+		break;
+	case SLJIT_MOV_U16:
+		ins = LDREXH;
+		break;
+	default:
+		ins = LDREX;
+		break;
+	}
+
+	return push_inst(compiler, ins | RN(mem_reg) | RD(dst_reg));
+}
+
+SLJIT_API_FUNC_ATTRIBUTE sljit_s32 sljit_emit_atomic_store(struct sljit_compiler *compiler, sljit_s32 op,
+	sljit_s32 src_reg,
+	sljit_s32 mem_reg,
+	sljit_s32 temp_reg)
+{
+	sljit_u32 ins;
+
+	/* temp_reg == mem_reg is undefined so use another temp register */
+	SLJIT_UNUSED_ARG(temp_reg);
+
+	CHECK_ERROR();
+	CHECK(check_sljit_emit_atomic_store(compiler, op, src_reg, mem_reg, temp_reg));
+
+	switch (GET_OPCODE(op)) {
+	case SLJIT_MOV_U8:
+		ins = STREXB;
+		break;
+	case SLJIT_MOV_U16:
+		ins = STREXH;
+		break;
+	default:
+		ins = STREX;
+		break;
+	}
+
+	FAIL_IF(push_inst(compiler, ins | RN(mem_reg) | RD(TMP_REG1) | RM(src_reg)));
+	if (op & SLJIT_SET_ATOMIC_STORED)
+		return push_inst(compiler, CMP | SET_FLAGS | SRC2_IMM | RN(TMP_REG1));
+
+	return SLJIT_SUCCESS;
+}
+
+SLJIT_API_FUNC_ATTRIBUTE struct sljit_const* sljit_emit_const(struct sljit_compiler *compiler, sljit_s32 dst, sljit_sw dstw, sljit_sw init_value)
+{
+	struct sljit_const *const_;
+	sljit_s32 dst_r;
+
+	CHECK_ERROR_PTR();
+	CHECK_PTR(check_sljit_emit_const(compiler, dst, dstw, init_value));
+	ADJUST_LOCAL_OFFSET(dst, dstw);
+
+	const_ = (struct sljit_const*)ensure_abuf(compiler, sizeof(struct sljit_const));
+	PTR_FAIL_IF(!const_);
+	set_const(const_, compiler);
+
+	dst_r = FAST_IS_REG(dst) ? dst : TMP_REG2;
+
+#if (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+	PTR_FAIL_IF(push_inst_with_unique_literal(compiler,
+		EMIT_DATA_TRANSFER(WORD_SIZE | LOAD_DATA, 1, dst_r, TMP_PC, 0), (sljit_ins)init_value));
+	compiler->patches++;
+#else /* !SLJIT_CONFIG_ARM_V6 */
+	PTR_FAIL_IF(emit_imm(compiler, dst_r, init_value));
+#endif /* SLJIT_CONFIG_ARM_V6 */
+
+	if (dst & SLJIT_MEM)
+		PTR_FAIL_IF(emit_op_mem(compiler, WORD_SIZE, TMP_REG2, dst, dstw, TMP_REG1));
+	return const_;
+}
+
+SLJIT_API_FUNC_ATTRIBUTE struct sljit_jump* sljit_emit_mov_addr(struct sljit_compiler *compiler, sljit_s32 dst, sljit_sw dstw)
+{
+	struct sljit_jump *jump;
+	sljit_s32 dst_r;
+
+	CHECK_ERROR_PTR();
+	CHECK_PTR(check_sljit_emit_mov_addr(compiler, dst, dstw));
+	ADJUST_LOCAL_OFFSET(dst, dstw);
+
+	dst_r = FAST_IS_REG(dst) ? dst : TMP_REG2;
+
+#if (defined SLJIT_CONFIG_ARM_V6 && SLJIT_CONFIG_ARM_V6)
+	PTR_FAIL_IF(push_inst_with_unique_literal(compiler, EMIT_DATA_TRANSFER(WORD_SIZE | LOAD_DATA, 1, dst_r, TMP_PC, 0), 0));
+	compiler->patches++;
+#else /* !SLJIT_CONFIG_ARM_V6 */
+	PTR_FAIL_IF(push_inst(compiler, RD(dst_r)));
+#endif /* SLJIT_CONFIG_ARM_V6 */
+
+	jump = (struct sljit_jump*)ensure_abuf(compiler, sizeof(struct sljit_jump));
+	PTR_FAIL_IF(!jump);
+	set_mov_addr(jump, compiler, 1);
+
+#if (defined SLJIT_CONFIG_ARM_V7 && SLJIT_CONFIG_ARM_V7)
+	compiler->size += 1;
+#endif /* SLJIT_CONFIG_ARM_V7 */
+
+	if (dst & SLJIT_MEM)
+		PTR_FAIL_IF(emit_op_mem(compiler, WORD_SIZE, TMP_REG2, dst, dstw, TMP_REG1));
+	return jump;
+}
+
+SLJIT_API_FUNC_ATTRIBUTE void sljit_set_jump_addr(sljit_uw addr, sljit_uw new_target, sljit_sw executable_offset)
+{
+	set_jump_addr(addr, executable_offset, new_target, 1);
+}
+
+SLJIT_API_FUNC_ATTRIBUTE void sljit_set_const(sljit_uw addr, sljit_sw new_constant, sljit_sw executable_offset)
+{
+	set_const_value(addr, executable_offset, (sljit_uw)new_constant, 1);
+}