diff --git a/yjit_asm.c b/yjit_asm.c
index 001856395f..4b7205e14a 100644
--- a/yjit_asm.c
+++ b/yjit_asm.c
@@ -128,7 +128,7 @@ x86opnd_t const_ptr_opnd(const void *ptr)
 }
 
 // Align the current write position to a multiple of bytes
-static uint8_t* align_ptr(uint8_t* ptr, uint32_t multiple)
+static uint8_t *align_ptr(uint8_t *ptr, uint32_t multiple)
 {
     // Compute the pointer modulo the given alignment boundary
     uint32_t rem = ((uint32_t)(uintptr_t)ptr) % multiple;
@@ -144,16 +144,16 @@ static uint8_t* align_ptr(uint8_t* ptr, uint32_t multiple)
 }
 
 // Allocate a block of executable memory
-uint8_t* alloc_exec_mem(uint32_t mem_size)
+uint8_t *alloc_exec_mem(uint32_t mem_size)
 {
 #ifndef _WIN32
-    uint8_t* mem_block;
+    uint8_t *mem_block;
 
     // On Linux
     #if defined(MAP_FIXED_NOREPLACE) && defined(_SC_PAGESIZE)
         // Align the requested address to page size
         uint32_t page_size = (uint32_t)sysconf(_SC_PAGESIZE);
-        uint8_t* req_addr = align_ptr((uint8_t*)&alloc_exec_mem, page_size);
+        uint8_t *req_addr = align_ptr((uint8_t*)&alloc_exec_mem, page_size);
 
         while (req_addr < (uint8_t*)&alloc_exec_mem + INT32_MAX)
         {
@@ -223,16 +223,16 @@ uint8_t* alloc_exec_mem(uint32_t mem_size)
 code_page_t *freelist = NULL;
 
 // Allocate a single code page from a pool of free pages
-code_page_t* alloc_code_page()
+code_page_t *alloc_code_page()
 {
     // If the free list is empty
     if (!freelist) {
         // Allocate many pages at once
-        uint8_t* code_chunk = alloc_exec_mem(PAGES_PER_ALLOC * CODE_PAGE_SIZE);
+        uint8_t *code_chunk = alloc_exec_mem(PAGES_PER_ALLOC * CODE_PAGE_SIZE);
 
         // Do this in reverse order so we allocate our pages in order
         for (int i = PAGES_PER_ALLOC - 1; i >= 0; --i) {
-            code_page_t* code_page = malloc(sizeof(code_page_t));
+            code_page_t *code_page = malloc(sizeof(code_page_t));
             code_page->mem_block = code_chunk + i * CODE_PAGE_SIZE;
             assert ((intptr_t)code_page->mem_block % CODE_PAGE_SIZE == 0);
             code_page->page_size = CODE_PAGE_SIZE;
@@ -241,21 +241,21 @@ code_page_t* alloc_code_page()
         }
     }
 
-    code_page_t* free_page = freelist;
+    code_page_t *free_page = freelist;
     freelist = freelist->_next;
 
     return free_page;
 }
 
 // Put a code page back into the allocation pool
-void free_code_page(code_page_t* code_page)
+void free_code_page(code_page_t *code_page)
 {
     code_page->_next = freelist;
     freelist = code_page;
 }
 
 // Initialize a code block object
-void cb_init(codeblock_t* cb, uint8_t* mem_block, uint32_t mem_size)
+void cb_init(codeblock_t *cb, uint8_t *mem_block, uint32_t mem_size)
 {
     assert (mem_block);
     cb->mem_block = mem_block;
@@ -266,11 +266,11 @@ void cb_init(codeblock_t* cb, uint8_t* mem_block, uint32_t mem_size)
 }
 
 // Align the current write position to a multiple of bytes
-void cb_align_pos(codeblock_t* cb, uint32_t multiple)
+void cb_align_pos(codeblock_t *cb, uint32_t multiple)
 {
     // Compute the pointer modulo the given alignment boundary
-    uint8_t* ptr = &cb->mem_block[cb->write_pos];
-    uint8_t* aligned_ptr = align_ptr(ptr, multiple);
+    uint8_t *ptr = &cb->mem_block[cb->write_pos];
+    uint8_t *aligned_ptr = align_ptr(ptr, multiple);
 
     // Pad the pointer by the necessary amount to align it
     ptrdiff_t pad = aligned_ptr - ptr;
@@ -278,14 +278,14 @@ void cb_align_pos(codeblock_t* cb, uint32_t multiple)
 }
 
 // Set the current write position
-void cb_set_pos(codeblock_t* cb, uint32_t pos)
+void cb_set_pos(codeblock_t *cb, uint32_t pos)
 {
     assert (pos < cb->mem_size);
     cb->write_pos = pos;
 }
 
 // Set the current write position from a pointer
-void cb_set_write_ptr(codeblock_t* cb, uint8_t* code_ptr)
+void cb_set_write_ptr(codeblock_t *cb, uint8_t *code_ptr)
 {
     intptr_t pos = code_ptr - cb->mem_block;
     assert (pos < cb->mem_size);
@@ -293,20 +293,20 @@ void cb_set_write_ptr(codeblock_t* cb, uint8_t* code_ptr)
 }
 
 // Get a direct pointer into the executable memory block
-uint8_t* cb_get_ptr(codeblock_t* cb, uint32_t index)
+uint8_t *cb_get_ptr(codeblock_t *cb, uint32_t index)
 {
     assert (index < cb->mem_size);
     return &cb->mem_block[index];
 }
 
 // Get a direct pointer to the current write position
-uint8_t* cb_get_write_ptr(codeblock_t* cb)
+uint8_t *cb_get_write_ptr(codeblock_t *cb)
 {
     return cb_get_ptr(cb, cb->write_pos);
 }
 
 // Write a byte at the current position
-void cb_write_byte(codeblock_t* cb, uint8_t byte)
+void cb_write_byte(codeblock_t *cb, uint8_t byte)
 {
     assert (cb->mem_block);
     assert (cb->write_pos + 1 <= cb->mem_size);
@@ -314,7 +314,7 @@ void cb_write_byte(codeblock_t* cb, uint8_t byte)
 }
 
 // Write multiple bytes starting from the current position
-void cb_write_bytes(codeblock_t* cb, uint32_t num_bytes, ...)
+void cb_write_bytes(codeblock_t *cb, uint32_t num_bytes, ...)
 {
     va_list va;
     va_start(va, num_bytes);
@@ -329,7 +329,7 @@ void cb_write_bytes(codeblock_t* cb, uint32_t num_bytes, ...)
 }
 
 // Write a signed integer over a given number of bits at the current position
-void cb_write_int(codeblock_t* cb, uint64_t val, uint32_t num_bits)
+void cb_write_int(codeblock_t *cb, uint64_t val, uint32_t num_bits)
 {
     assert (num_bits > 0);
     assert (num_bits % 8 == 0);
@@ -378,7 +378,7 @@ void cb_write_int(codeblock_t* cb, uint64_t val, uint32_t num_bits)
 }
 
 // Allocate a new label with a given name
-uint32_t cb_new_label(codeblock_t* cb, const char* name)
+uint32_t cb_new_label(codeblock_t *cb, const char *name)
 {
     //if (hasASM)
     //    writeString(to!string(label) ~ ":");
@@ -396,14 +396,14 @@ uint32_t cb_new_label(codeblock_t* cb, const char* name)
 }
 
 // Write a label at the current address
-void cb_write_label(codeblock_t* cb, uint32_t label_idx)
+void cb_write_label(codeblock_t *cb, uint32_t label_idx)
 {
     assert (label_idx < MAX_LABELS);
     cb->label_addrs[label_idx] = cb->write_pos;
 }
 
 // Add a label reference at the current write position
-void cb_label_ref(codeblock_t* cb, uint32_t label_idx)
+void cb_label_ref(codeblock_t *cb, uint32_t label_idx)
 {
     assert (label_idx < MAX_LABELS);
     assert (cb->num_refs < MAX_LABEL_REFS);
@@ -414,7 +414,7 @@ void cb_label_ref(codeblock_t* cb, uint32_t label_idx)
 }
 
 // Link internal label references
-void cb_link_labels(codeblock_t* cb)
+void cb_link_labels(codeblock_t *cb)
 {
     uint32_t orig_pos = cb->write_pos;
 
@@ -516,7 +516,7 @@ uint32_t disp_size(x86opnd_t opnd)
 
 // Write the REX byte
 static void cb_write_rex(
-    codeblock_t* cb,
+    codeblock_t *cb,
     bool w_flag,
     uint8_t reg_no,
     uint8_t idx_reg_no,
@@ -539,7 +539,7 @@ static void cb_write_rex(
 }
 
 // Write an opcode byte with an embedded register operand
-static void cb_write_opcode(codeblock_t* cb, uint8_t opcode, x86opnd_t reg)
+static void cb_write_opcode(codeblock_t *cb, uint8_t opcode, x86opnd_t reg)
 {
     // Write the reg field into the opcode byte
     uint8_t op_byte = opcode | (reg.as.reg.reg_no & 7);
@@ -548,7 +548,7 @@ static void cb_write_opcode(codeblock_t* cb, uint8_t opcode, x86opnd_t reg)
 
 // Encode an RM instruction
 void cb_write_rm(
-    codeblock_t* cb,
+    codeblock_t *cb,
     bool szPref,
     bool rexW,
     x86opnd_t r_opnd,
@@ -709,8 +709,8 @@ void cb_write_rm(
 
 // Encode a mul-like single-operand RM instruction
 void write_rm_unary(
-    codeblock_t* cb,
-    const char* mnem,
+    codeblock_t *cb,
+    const char *mnem,
     uint8_t opMemReg8,
     uint8_t opMemRegPref,
     uint8_t opExt,
@@ -738,8 +738,8 @@ void write_rm_unary(
 
 // Encode an add-like RM instruction with multiple possible encodings
 void cb_write_rm_multi(
-    codeblock_t* cb,
-    const char* mnem,
+    codeblock_t *cb,
+    const char *mnem,
     uint8_t opMemReg8,
     uint8_t opMemRegPref,
     uint8_t opRegMem8,
@@ -837,8 +837,8 @@ void cb_write_rm_multi(
 
 // Encode a single-operand shift instruction
 void cb_write_shift(
-    codeblock_t* cb,
-    const char* mnem,
+    codeblock_t *cb,
+    const char *mnem,
     uint8_t opMemOnePref,
     uint8_t opMemClPref,
     uint8_t opMemImmPref,
@@ -887,7 +887,7 @@ void cb_write_shift(
 
 // Encode a relative jump to a label (direct or conditional)
 // Note: this always encodes a 32-bit offset
-void cb_write_jcc(codeblock_t* cb, const char* mnem, uint8_t op0, uint8_t op1, uint32_t label_idx)
+void cb_write_jcc(codeblock_t *cb, const char *mnem, uint8_t op0, uint8_t op1, uint32_t label_idx)
 {
     //cb.writeASM(mnem, label);
 
@@ -904,7 +904,7 @@ void cb_write_jcc(codeblock_t* cb, const char* mnem, uint8_t op0, uint8_t op1, u
 }
 
 // Encode a relative jump to a pointer at a 32-bit offset (direct or conditional)
-void cb_write_jcc_ptr(codeblock_t* cb, const char* mnem, uint8_t op0, uint8_t op1, uint8_t* dst_ptr)
+void cb_write_jcc_ptr(codeblock_t *cb, const char *mnem, uint8_t op0, uint8_t op1, uint8_t *dst_ptr)
 {
     //cb.writeASM(mnem, label);
 
@@ -914,7 +914,7 @@ void cb_write_jcc_ptr(codeblock_t* cb, const char* mnem, uint8_t op0, uint8_t op
     cb_write_byte(cb, op1);
 
     // Pointer to the end of this jump instruction
-    uint8_t* end_ptr = &cb->mem_block[cb->write_pos] + 4;
+    uint8_t *end_ptr = &cb->mem_block[cb->write_pos] + 4;
 
     // Compute the jump offset
     int64_t rel64 = (int64_t)(dst_ptr - end_ptr);
@@ -925,7 +925,7 @@ void cb_write_jcc_ptr(codeblock_t* cb, const char* mnem, uint8_t op0, uint8_t op
 }
 
 // Encode a conditional move instruction
-void cb_write_cmov(codeblock_t* cb, const char* mnem, uint8_t opcode1, x86opnd_t dst, x86opnd_t src)
+void cb_write_cmov(codeblock_t *cb, const char *mnem, uint8_t opcode1, x86opnd_t dst, x86opnd_t src)
 {
     //cb.writeASM(mnem, dst, src);
 
@@ -940,7 +940,7 @@ void cb_write_cmov(codeblock_t* cb, const char* mnem, uint8_t opcode1, x86opnd_t
 }
 
 // add - Integer addition
-void add(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1)
+void add(codeblock_t *cb, x86opnd_t opnd0, x86opnd_t opnd1)
 {
     cb_write_rm_multi(
         cb,
@@ -959,7 +959,7 @@ void add(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1)
 }
 
 /// and - Bitwise AND
-void and(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1)
+void and(codeblock_t *cb, x86opnd_t opnd0, x86opnd_t opnd1)
 {
     cb_write_rm_multi(
         cb,
@@ -978,7 +978,7 @@ void and(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1)
 }
 
 // call - Call to a pointer with a 32-bit displacement offset
-void call_rel32(codeblock_t* cb, int32_t rel32)
+void call_rel32(codeblock_t *cb, int32_t rel32)
 {
     //cb.writeASM("call", rel32);
 
@@ -990,12 +990,12 @@ void call_rel32(codeblock_t* cb, int32_t rel32)
 }
 
 // call - Call a pointer, encode with a 32-bit offset if possible
-void call_ptr(codeblock_t* cb, x86opnd_t scratch_reg, uint8_t* dst_ptr)
+void call_ptr(codeblock_t *cb, x86opnd_t scratch_reg, uint8_t *dst_ptr)
 {
     assert (scratch_reg.type == OPND_REG);
 
     // Pointer to the end of this call instruction
-    uint8_t* end_ptr = &cb->mem_block[cb->write_pos] + 5;
+    uint8_t *end_ptr = &cb->mem_block[cb->write_pos] + 5;
 
     // Compute the jump offset
     int64_t rel64 = (int64_t)(dst_ptr - end_ptr);
@@ -1013,7 +1013,7 @@ void call_ptr(codeblock_t* cb, x86opnd_t scratch_reg, uint8_t* dst_ptr)
 }
 
 /// call - Call to label with 32-bit offset
-void call_label(codeblock_t* cb, uint32_t label_idx)
+void call_label(codeblock_t *cb, uint32_t label_idx)
 {
     //cb.writeASM("call", label);
 
@@ -1028,46 +1028,46 @@ void call_label(codeblock_t* cb, uint32_t label_idx)
 }
 
 /// call - Indirect call with an R/M operand
-void call(codeblock_t* cb, x86opnd_t opnd)
+void call(codeblock_t *cb, x86opnd_t opnd)
 {
     //cb.writeASM("call", opnd);
     cb_write_rm(cb, false, false, NO_OPND, opnd, 2, 1, 0xFF);
 }
 
 /// cmovcc - Conditional move
-void cmova(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmova", 0x47, dst, src); }
-void cmovae(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovae", 0x43, dst, src); }
-void cmovb(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovb", 0x42, dst, src); }
-void cmovbe(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovbe", 0x46, dst, src); }
-void cmovc(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovc", 0x42, dst, src); }
-void cmove(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmove", 0x44, dst, src); }
-void cmovg(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovg", 0x4F, dst, src); }
-void cmovge(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovge", 0x4D, dst, src); }
-void cmovl(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovl", 0x4C, dst, src); }
-void cmovle(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovle", 0x4E, dst, src); }
-void cmovna(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovna", 0x46, dst, src); }
-void cmovnae(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovnae", 0x42, dst, src); }
-void cmovnb(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovnb", 0x43, dst, src); }
-void cmovnbe(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovnbe", 0x47, dst, src); }
-void cmovnc(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovnc", 0x43, dst, src); }
-void cmovne(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovne", 0x45, dst, src); }
-void cmovng(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovng", 0x4E, dst, src); }
-void cmovnge(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovnge", 0x4C, dst, src); }
-void cmovnl(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovnl" , 0x4D, dst, src); }
-void cmovnle(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovnle", 0x4F, dst, src); }
-void cmovno(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovno", 0x41, dst, src); }
-void cmovnp(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovnp", 0x4B, dst, src); }
-void cmovns(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovns", 0x49, dst, src); }
-void cmovnz(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovnz", 0x45, dst, src); }
-void cmovo(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovo", 0x40, dst, src); }
-void cmovp(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovp", 0x4A, dst, src); }
-void cmovpe(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovpe", 0x4A, dst, src); }
-void cmovpo(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovpo", 0x4B, dst, src); }
-void cmovs(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovs", 0x48, dst, src); }
-void cmovz(codeblock_t* cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovz", 0x44, dst, src); }
+void cmova(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmova", 0x47, dst, src); }
+void cmovae(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovae", 0x43, dst, src); }
+void cmovb(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovb", 0x42, dst, src); }
+void cmovbe(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovbe", 0x46, dst, src); }
+void cmovc(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovc", 0x42, dst, src); }
+void cmove(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmove", 0x44, dst, src); }
+void cmovg(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovg", 0x4F, dst, src); }
+void cmovge(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovge", 0x4D, dst, src); }
+void cmovl(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovl", 0x4C, dst, src); }
+void cmovle(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovle", 0x4E, dst, src); }
+void cmovna(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovna", 0x46, dst, src); }
+void cmovnae(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovnae", 0x42, dst, src); }
+void cmovnb(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovnb", 0x43, dst, src); }
+void cmovnbe(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovnbe", 0x47, dst, src); }
+void cmovnc(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovnc", 0x43, dst, src); }
+void cmovne(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovne", 0x45, dst, src); }
+void cmovng(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovng", 0x4E, dst, src); }
+void cmovnge(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovnge", 0x4C, dst, src); }
+void cmovnl(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovnl" , 0x4D, dst, src); }
+void cmovnle(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovnle", 0x4F, dst, src); }
+void cmovno(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovno", 0x41, dst, src); }
+void cmovnp(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovnp", 0x4B, dst, src); }
+void cmovns(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovns", 0x49, dst, src); }
+void cmovnz(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovnz", 0x45, dst, src); }
+void cmovo(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovo", 0x40, dst, src); }
+void cmovp(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovp", 0x4A, dst, src); }
+void cmovpe(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovpe", 0x4A, dst, src); }
+void cmovpo(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovpo", 0x4B, dst, src); }
+void cmovs(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovs", 0x48, dst, src); }
+void cmovz(codeblock_t *cb, x86opnd_t dst, x86opnd_t src) { cb_write_cmov(cb, "cmovz", 0x44, dst, src); }
 
 /// cmp - Compare and set flags
-void cmp(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1)
+void cmp(codeblock_t *cb, x86opnd_t opnd0, x86opnd_t opnd1)
 {
     cb_write_rm_multi(
         cb,
@@ -1086,21 +1086,21 @@ void cmp(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1)
 }
 
 /// cdq - Convert doubleword to quadword
-void cdq(codeblock_t* cb)
+void cdq(codeblock_t *cb)
 {
     //cb.writeASM("cdq");
     cb_write_byte(cb, 0x99);
 }
 
 /// cqo - Convert quadword to octaword
-void cqo(codeblock_t* cb)
+void cqo(codeblock_t *cb)
 {
     //cb.writeASM("cqo");
     cb_write_bytes(cb, 2, 0x48, 0x99);
 }
 
 /// Interrupt 3 - trap to debugger
-void int3(codeblock_t* cb)
+void int3(codeblock_t *cb)
 {
     //cb.writeASM("INT 3");
     cb_write_byte(cb, 0xCC);
@@ -1205,80 +1205,80 @@ void imul(CodeBlock cb, X86Opnd opnd0, X86Opnd opnd1, X86Opnd opnd2)
 */
 
 /// jcc - relative jumps to a label
-void ja_label  (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "ja"  , 0x0F, 0x87, label_idx); }
-void jae_label (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jae" , 0x0F, 0x83, label_idx); }
-void jb_label  (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jb"  , 0x0F, 0x82, label_idx); }
-void jbe_label (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jbe" , 0x0F, 0x86, label_idx); }
-void jc_label  (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jc"  , 0x0F, 0x82, label_idx); }
-void je_label  (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "je"  , 0x0F, 0x84, label_idx); }
-void jg_label  (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jg"  , 0x0F, 0x8F, label_idx); }
-void jge_label (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jge" , 0x0F, 0x8D, label_idx); }
-void jl_label  (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jl"  , 0x0F, 0x8C, label_idx); }
-void jle_label (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jle" , 0x0F, 0x8E, label_idx); }
-void jna_label (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jna" , 0x0F, 0x86, label_idx); }
-void jnae_label(codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jnae", 0x0F, 0x82, label_idx); }
-void jnb_label (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jnb" , 0x0F, 0x83, label_idx); }
-void jnbe_label(codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jnbe", 0x0F, 0x87, label_idx); }
-void jnc_label (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jnc" , 0x0F, 0x83, label_idx); }
-void jne_label (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jne" , 0x0F, 0x85, label_idx); }
-void jng_label (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jng" , 0x0F, 0x8E, label_idx); }
-void jnge_label(codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jnge", 0x0F, 0x8C, label_idx); }
-void jnl_label (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jnl" , 0x0F, 0x8D, label_idx); }
-void jnle_label(codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jnle", 0x0F, 0x8F, label_idx); }
-void jno_label (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jno" , 0x0F, 0x81, label_idx); }
-void jnp_label (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jnp" , 0x0F, 0x8b, label_idx); }
-void jns_label (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jns" , 0x0F, 0x89, label_idx); }
-void jnz_label (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jnz" , 0x0F, 0x85, label_idx); }
-void jo_label  (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jo"  , 0x0F, 0x80, label_idx); }
-void jp_label  (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jp"  , 0x0F, 0x8A, label_idx); }
-void jpe_label (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jpe" , 0x0F, 0x8A, label_idx); }
-void jpo_label (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jpo" , 0x0F, 0x8B, label_idx); }
-void js_label  (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "js"  , 0x0F, 0x88, label_idx); }
-void jz_label  (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jz"  , 0x0F, 0x84, label_idx); }
-void jmp_label (codeblock_t* cb, uint32_t label_idx) { cb_write_jcc(cb, "jmp" , 0xFF, 0xE9, label_idx); }
+void ja_label  (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "ja"  , 0x0F, 0x87, label_idx); }
+void jae_label (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jae" , 0x0F, 0x83, label_idx); }
+void jb_label  (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jb"  , 0x0F, 0x82, label_idx); }
+void jbe_label (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jbe" , 0x0F, 0x86, label_idx); }
+void jc_label  (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jc"  , 0x0F, 0x82, label_idx); }
+void je_label  (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "je"  , 0x0F, 0x84, label_idx); }
+void jg_label  (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jg"  , 0x0F, 0x8F, label_idx); }
+void jge_label (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jge" , 0x0F, 0x8D, label_idx); }
+void jl_label  (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jl"  , 0x0F, 0x8C, label_idx); }
+void jle_label (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jle" , 0x0F, 0x8E, label_idx); }
+void jna_label (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jna" , 0x0F, 0x86, label_idx); }
+void jnae_label(codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jnae", 0x0F, 0x82, label_idx); }
+void jnb_label (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jnb" , 0x0F, 0x83, label_idx); }
+void jnbe_label(codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jnbe", 0x0F, 0x87, label_idx); }
+void jnc_label (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jnc" , 0x0F, 0x83, label_idx); }
+void jne_label (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jne" , 0x0F, 0x85, label_idx); }
+void jng_label (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jng" , 0x0F, 0x8E, label_idx); }
+void jnge_label(codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jnge", 0x0F, 0x8C, label_idx); }
+void jnl_label (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jnl" , 0x0F, 0x8D, label_idx); }
+void jnle_label(codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jnle", 0x0F, 0x8F, label_idx); }
+void jno_label (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jno" , 0x0F, 0x81, label_idx); }
+void jnp_label (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jnp" , 0x0F, 0x8b, label_idx); }
+void jns_label (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jns" , 0x0F, 0x89, label_idx); }
+void jnz_label (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jnz" , 0x0F, 0x85, label_idx); }
+void jo_label  (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jo"  , 0x0F, 0x80, label_idx); }
+void jp_label  (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jp"  , 0x0F, 0x8A, label_idx); }
+void jpe_label (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jpe" , 0x0F, 0x8A, label_idx); }
+void jpo_label (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jpo" , 0x0F, 0x8B, label_idx); }
+void js_label  (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "js"  , 0x0F, 0x88, label_idx); }
+void jz_label  (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jz"  , 0x0F, 0x84, label_idx); }
+void jmp_label (codeblock_t *cb, uint32_t label_idx) { cb_write_jcc(cb, "jmp" , 0xFF, 0xE9, label_idx); }
 
 /// jcc - relative jumps to a pointer (32-bit offset)
-void ja_ptr  (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "ja"  , 0x0F, 0x87, ptr); }
-void jae_ptr (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jae" , 0x0F, 0x83, ptr); }
-void jb_ptr  (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jb"  , 0x0F, 0x82, ptr); }
-void jbe_ptr (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jbe" , 0x0F, 0x86, ptr); }
-void jc_ptr  (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jc"  , 0x0F, 0x82, ptr); }
-void je_ptr  (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "je"  , 0x0F, 0x84, ptr); }
-void jg_ptr  (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jg"  , 0x0F, 0x8F, ptr); }
-void jge_ptr (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jge" , 0x0F, 0x8D, ptr); }
-void jl_ptr  (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jl"  , 0x0F, 0x8C, ptr); }
-void jle_ptr (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jle" , 0x0F, 0x8E, ptr); }
-void jna_ptr (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jna" , 0x0F, 0x86, ptr); }
-void jnae_ptr(codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jnae", 0x0F, 0x82, ptr); }
-void jnb_ptr (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jnb" , 0x0F, 0x83, ptr); }
-void jnbe_ptr(codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jnbe", 0x0F, 0x87, ptr); }
-void jnc_ptr (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jnc" , 0x0F, 0x83, ptr); }
-void jne_ptr (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jne" , 0x0F, 0x85, ptr); }
-void jng_ptr (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jng" , 0x0F, 0x8E, ptr); }
-void jnge_ptr(codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jnge", 0x0F, 0x8C, ptr); }
-void jnl_ptr (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jnl" , 0x0F, 0x8D, ptr); }
-void jnle_ptr(codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jnle", 0x0F, 0x8F, ptr); }
-void jno_ptr (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jno" , 0x0F, 0x81, ptr); }
-void jnp_ptr (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jnp" , 0x0F, 0x8b, ptr); }
-void jns_ptr (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jns" , 0x0F, 0x89, ptr); }
-void jnz_ptr (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jnz" , 0x0F, 0x85, ptr); }
-void jo_ptr  (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jo"  , 0x0F, 0x80, ptr); }
-void jp_ptr  (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jp"  , 0x0F, 0x8A, ptr); }
-void jpe_ptr (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jpe" , 0x0F, 0x8A, ptr); }
-void jpo_ptr (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jpo" , 0x0F, 0x8B, ptr); }
-void js_ptr  (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "js"  , 0x0F, 0x88, ptr); }
-void jz_ptr  (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jz"  , 0x0F, 0x84, ptr); }
-void jmp_ptr (codeblock_t* cb, uint8_t* ptr) { cb_write_jcc_ptr(cb, "jmp" , 0xFF, 0xE9, ptr); }
+void ja_ptr  (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "ja"  , 0x0F, 0x87, ptr); }
+void jae_ptr (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jae" , 0x0F, 0x83, ptr); }
+void jb_ptr  (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jb"  , 0x0F, 0x82, ptr); }
+void jbe_ptr (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jbe" , 0x0F, 0x86, ptr); }
+void jc_ptr  (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jc"  , 0x0F, 0x82, ptr); }
+void je_ptr  (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "je"  , 0x0F, 0x84, ptr); }
+void jg_ptr  (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jg"  , 0x0F, 0x8F, ptr); }
+void jge_ptr (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jge" , 0x0F, 0x8D, ptr); }
+void jl_ptr  (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jl"  , 0x0F, 0x8C, ptr); }
+void jle_ptr (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jle" , 0x0F, 0x8E, ptr); }
+void jna_ptr (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jna" , 0x0F, 0x86, ptr); }
+void jnae_ptr(codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jnae", 0x0F, 0x82, ptr); }
+void jnb_ptr (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jnb" , 0x0F, 0x83, ptr); }
+void jnbe_ptr(codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jnbe", 0x0F, 0x87, ptr); }
+void jnc_ptr (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jnc" , 0x0F, 0x83, ptr); }
+void jne_ptr (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jne" , 0x0F, 0x85, ptr); }
+void jng_ptr (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jng" , 0x0F, 0x8E, ptr); }
+void jnge_ptr(codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jnge", 0x0F, 0x8C, ptr); }
+void jnl_ptr (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jnl" , 0x0F, 0x8D, ptr); }
+void jnle_ptr(codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jnle", 0x0F, 0x8F, ptr); }
+void jno_ptr (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jno" , 0x0F, 0x81, ptr); }
+void jnp_ptr (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jnp" , 0x0F, 0x8b, ptr); }
+void jns_ptr (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jns" , 0x0F, 0x89, ptr); }
+void jnz_ptr (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jnz" , 0x0F, 0x85, ptr); }
+void jo_ptr  (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jo"  , 0x0F, 0x80, ptr); }
+void jp_ptr  (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jp"  , 0x0F, 0x8A, ptr); }
+void jpe_ptr (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jpe" , 0x0F, 0x8A, ptr); }
+void jpo_ptr (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jpo" , 0x0F, 0x8B, ptr); }
+void js_ptr  (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "js"  , 0x0F, 0x88, ptr); }
+void jz_ptr  (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jz"  , 0x0F, 0x84, ptr); }
+void jmp_ptr (codeblock_t *cb, uint8_t *ptr) { cb_write_jcc_ptr(cb, "jmp" , 0xFF, 0xE9, ptr); }
 
 /// jmp - Indirect jump near to an R/M operand
-void jmp_rm(codeblock_t* cb, x86opnd_t opnd)
+void jmp_rm(codeblock_t *cb, x86opnd_t opnd)
 {
     //cb.writeASM("jmp", opnd);
     cb_write_rm(cb, false, false, NO_OPND, opnd, 4, 1, 0xFF);
 }
 
 // jmp - Jump with relative 32-bit offset
-void jmp32(codeblock_t* cb, int32_t offset)
+void jmp32(codeblock_t *cb, int32_t offset)
 {
     //cb.writeASM("jmp", ((offset > 0)? "+":"-") ~ to!string(offset));
     cb_write_byte(cb, 0xE9);
@@ -1286,7 +1286,7 @@ void jmp32(codeblock_t* cb, int32_t offset)
 }
 
 /// lea - Load Effective Address
-void lea(codeblock_t* cb, x86opnd_t dst, x86opnd_t src)
+void lea(codeblock_t *cb, x86opnd_t dst, x86opnd_t src)
 {
     //cb.writeASM("lea", dst, src);
     assert (dst.num_bits == 64);
@@ -1294,7 +1294,7 @@ void lea(codeblock_t* cb, x86opnd_t dst, x86opnd_t src)
 }
 
 /// mov - Data move operation
-void mov(codeblock_t* cb, x86opnd_t dst, x86opnd_t src)
+void mov(codeblock_t *cb, x86opnd_t dst, x86opnd_t src)
 {
     // R/M + Imm
     if (src.type == OPND_IMM)
@@ -1357,7 +1357,7 @@ void mov(codeblock_t* cb, x86opnd_t dst, x86opnd_t src)
 }
 
 /// movsx - Move with sign extension (signed integers)
-void movsx(codeblock_t* cb, x86opnd_t dst, x86opnd_t src)
+void movsx(codeblock_t *cb, x86opnd_t dst, x86opnd_t src)
 {
     assert (dst.type == OPND_REG);
     assert (src.type == OPND_REG || src.type == OPND_MEM);
@@ -1385,7 +1385,7 @@ void movsx(codeblock_t* cb, x86opnd_t dst, x86opnd_t src)
 
 /*
 /// movzx - Move with zero extension (unsigned values)
-void movzx(codeblock_t* cb, x86opnd_t dst, x86opnd_t src)
+void movzx(codeblock_t *cb, x86opnd_t dst, x86opnd_t src)
 {
     cb.writeASM("movzx", dst, src);
 
@@ -1424,7 +1424,7 @@ void movzx(codeblock_t* cb, x86opnd_t dst, x86opnd_t src)
 */
 
 // neg - Integer negation (multiplication by -1)
-void neg(codeblock_t* cb, x86opnd_t opnd)
+void neg(codeblock_t *cb, x86opnd_t opnd)
 {
     write_rm_unary(
         cb,
@@ -1437,7 +1437,7 @@ void neg(codeblock_t* cb, x86opnd_t opnd)
 }
 
 // nop - Noop, one or multiple bytes long
-void nop(codeblock_t* cb, uint32_t length)
+void nop(codeblock_t *cb, uint32_t length)
 {
     switch (length)
     {
@@ -1504,7 +1504,7 @@ void nop(codeblock_t* cb, uint32_t length)
 }
 
 // not - Bitwise NOT
-void not(codeblock_t* cb, x86opnd_t opnd)
+void not(codeblock_t *cb, x86opnd_t opnd)
 {
     write_rm_unary(
         cb,
@@ -1517,7 +1517,7 @@ void not(codeblock_t* cb, x86opnd_t opnd)
 }
 
 /// or - Bitwise OR
-void or(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1)
+void or(codeblock_t *cb, x86opnd_t opnd0, x86opnd_t opnd1)
 {
     cb_write_rm_multi(
         cb,
@@ -1536,7 +1536,7 @@ void or(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1)
 }
 
 /// pop - Pop a register off the stack
-void pop(codeblock_t* cb, x86opnd_t opnd)
+void pop(codeblock_t *cb, x86opnd_t opnd)
 {
     assert (opnd.num_bits == 64);
 
@@ -1554,7 +1554,7 @@ void pop(codeblock_t* cb, x86opnd_t opnd)
 }
 
 /// popfq - Pop the flags register (64-bit)
-void popfq(codeblock_t* cb)
+void popfq(codeblock_t *cb)
 {
     //cb.writeASM("popfq");
 
@@ -1563,7 +1563,7 @@ void popfq(codeblock_t* cb)
 }
 
 /// push - Push an operand on the stack
-void push(codeblock_t* cb, x86opnd_t opnd)
+void push(codeblock_t *cb, x86opnd_t opnd)
 {
     assert (opnd.num_bits == 64);
 
@@ -1581,21 +1581,21 @@ void push(codeblock_t* cb, x86opnd_t opnd)
 }
 
 /// pushfq - Push the flags register (64-bit)
-void pushfq(codeblock_t* cb)
+void pushfq(codeblock_t *cb)
 {
     //cb.writeASM("pushfq");
     cb_write_byte(cb, 0x9C);
 }
 
 /// ret - Return from call, popping only the return address
-void ret(codeblock_t* cb)
+void ret(codeblock_t *cb)
 {
     //cb.writeASM("ret");
     cb_write_byte(cb, 0xC3);
 }
 
 // sal - Shift arithmetic left
-void sal(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1)
+void sal(codeblock_t *cb, x86opnd_t opnd0, x86opnd_t opnd1)
 {
     cb_write_shift(
         cb,
@@ -1610,7 +1610,7 @@ void sal(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1)
 }
 
 /// sar - Shift arithmetic right (signed)
-void sar(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1)
+void sar(codeblock_t *cb, x86opnd_t opnd0, x86opnd_t opnd1)
 {
     cb_write_shift(
         cb,
@@ -1624,7 +1624,7 @@ void sar(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1)
     );
 }
 // shl - Shift logical left
-void shl(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1)
+void shl(codeblock_t *cb, x86opnd_t opnd0, x86opnd_t opnd1)
 {
     cb_write_shift(
         cb,
@@ -1639,7 +1639,7 @@ void shl(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1)
 }
 
 /// shr - Shift logical right (unsigned)
-void shr(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1)
+void shr(codeblock_t *cb, x86opnd_t opnd0, x86opnd_t opnd1)
 {
     cb_write_shift(
         cb,
@@ -1654,7 +1654,7 @@ void shr(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1)
 }
 
 /// sub - Integer subtraction
-void sub(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1)
+void sub(codeblock_t *cb, x86opnd_t opnd0, x86opnd_t opnd1)
 {
     cb_write_rm_multi(
         cb,
@@ -1673,7 +1673,7 @@ void sub(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1)
 }
 
 /// test - Logical Compare
-void test(codeblock_t* cb, x86opnd_t rm_opnd, x86opnd_t test_opnd)
+void test(codeblock_t *cb, x86opnd_t rm_opnd, x86opnd_t test_opnd)
 {
     assert (rm_opnd.type == OPND_REG || rm_opnd.type == OPND_MEM);
     assert (test_opnd.type == OPND_REG || test_opnd.type == OPND_IMM);
@@ -1728,13 +1728,13 @@ void test(codeblock_t* cb, x86opnd_t rm_opnd, x86opnd_t test_opnd)
 }
 
 /// Undefined opcode
-void ud2(codeblock_t* cb)
+void ud2(codeblock_t *cb)
 {
     cb_write_bytes(cb, 2, 0x0F, 0x0B);
 }
 
 /// xchg - Exchange Register/Memory with Register
-void xchg(codeblock_t* cb, x86opnd_t rm_opnd, x86opnd_t r_opnd)
+void xchg(codeblock_t *cb, x86opnd_t rm_opnd, x86opnd_t r_opnd)
 {
     assert (rm_opnd.num_bits == 64);
     assert (r_opnd.num_bits == 64);
@@ -1757,7 +1757,7 @@ void xchg(codeblock_t* cb, x86opnd_t rm_opnd, x86opnd_t r_opnd)
 }
 
 /// xor - Exclusive bitwise OR
-void xor(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1)
+void xor(codeblock_t *cb, x86opnd_t opnd0, x86opnd_t opnd1)
 {
     cb_write_rm_multi(
         cb,
@@ -1776,7 +1776,7 @@ void xor(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1)
 }
 
 // LOCK - lock prefix for atomic shared memory operations
-void cb_write_lock_prefix(codeblock_t* cb)
+void cb_write_lock_prefix(codeblock_t *cb)
 {
     cb_write_byte(cb, 0xF0);
 }
diff --git a/yjit_asm.h b/yjit_asm.h
index 58f9a849ad..a9796b8584 100644
--- a/yjit_asm.h
+++ b/yjit_asm.h
@@ -32,7 +32,7 @@ typedef struct LabelRef
 typedef struct CodeBlock
 {
     // Memory block
-    uint8_t* mem_block;
+    uint8_t *mem_block;
 
     // Memory block size
     uint32_t mem_size;
@@ -45,7 +45,7 @@ typedef struct CodeBlock
 
     // Table of registered label names
     // Note that these should be constant strings only
-    const char* label_names[MAX_LABELS];
+    const char *label_names[MAX_LABELS];
 
     // References to labels
     labelref_t label_refs[MAX_LABEL_REFS];
@@ -141,13 +141,13 @@ typedef struct X86Opnd
 typedef struct code_page_struct
 {
     // Chunk of executable memory
-    uint8_t* mem_block;
+    uint8_t *mem_block;
 
     // Size of the executable memory chunk
     uint32_t page_size;
 
     // Next node in the free list (private)
-    struct code_page_struct* _next;
+    struct code_page_struct *_next;
 
 } code_page_t;
 
@@ -261,150 +261,150 @@ x86opnd_t const_ptr_opnd(const void *ptr);
 )
 
 // Machine code allocation
-uint8_t* alloc_exec_mem(uint32_t mem_size);
-code_page_t* alloc_code_page(void);
-void free_code_page(code_page_t* code_page);
+uint8_t *alloc_exec_mem(uint32_t mem_size);
+code_page_t *alloc_code_page(void);
+void free_code_page(code_page_t *code_page);
 
 // Code block methods
-void cb_init(codeblock_t* cb, uint8_t* mem_block, uint32_t mem_size);
-void cb_align_pos(codeblock_t* cb, uint32_t multiple);
-void cb_set_pos(codeblock_t* cb, uint32_t pos);
-void cb_set_write_ptr(codeblock_t* cb, uint8_t* code_ptr);
-uint8_t* cb_get_ptr(codeblock_t* cb, uint32_t index);
-uint8_t* cb_get_write_ptr(codeblock_t* cb);
-void cb_write_byte(codeblock_t* cb, uint8_t byte);
-void cb_write_bytes(codeblock_t* cb, uint32_t num_bytes, ...);
-void cb_write_int(codeblock_t* cb, uint64_t val, uint32_t num_bits);
-uint32_t cb_new_label(codeblock_t* cb, const char* name);
-void cb_write_label(codeblock_t* cb, uint32_t label_idx);
-void cb_label_ref(codeblock_t* cb, uint32_t label_idx);
-void cb_link_labels(codeblock_t* cb);
+void cb_init(codeblock_t *cb, uint8_t *mem_block, uint32_t mem_size);
+void cb_align_pos(codeblock_t *cb, uint32_t multiple);
+void cb_set_pos(codeblock_t *cb, uint32_t pos);
+void cb_set_write_ptr(codeblock_t *cb, uint8_t *code_ptr);
+uint8_t *cb_get_ptr(codeblock_t *cb, uint32_t index);
+uint8_t *cb_get_write_ptr(codeblock_t *cb);
+void cb_write_byte(codeblock_t *cb, uint8_t byte);
+void cb_write_bytes(codeblock_t *cb, uint32_t num_bytes, ...);
+void cb_write_int(codeblock_t *cb, uint64_t val, uint32_t num_bits);
+uint32_t cb_new_label(codeblock_t *cb, const char *name);
+void cb_write_label(codeblock_t *cb, uint32_t label_idx);
+void cb_label_ref(codeblock_t *cb, uint32_t label_idx);
+void cb_link_labels(codeblock_t *cb);
 
 // Encode individual instructions into a code block
-void add(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1);
-void and(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1);
-void call_ptr(codeblock_t* cb, x86opnd_t scratch_reg, uint8_t* dst_ptr);
-void call_label(codeblock_t* cb, uint32_t label_idx);
-void call(codeblock_t* cb, x86opnd_t opnd);
-void cmova(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovae(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovb(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovbe(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovc(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmove(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovg(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovge(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovl(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovle(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovna(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovnae(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovnb(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovnbe(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovnc(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovne(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovng(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovnge(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovnl(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovnle(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovno(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovnp(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovns(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovnz(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovo(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovp(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovpe(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovpo(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovs(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmovz(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void cmp(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1);
-void cdq(codeblock_t* cb);
-void cqo(codeblock_t* cb);
-void int3(codeblock_t* cb);
-void ja_label(codeblock_t* cb, uint32_t label_idx);
-void jae_label(codeblock_t* cb, uint32_t label_idx);
-void jb_label(codeblock_t* cb, uint32_t label_idx);
-void jbe_label(codeblock_t* cb, uint32_t label_idx);
-void jc_label(codeblock_t* cb, uint32_t label_idx);
-void je_label(codeblock_t* cb, uint32_t label_idx);
-void jg_label(codeblock_t* cb, uint32_t label_idx);
-void jge_label(codeblock_t* cb, uint32_t label_idx);
-void jl_label(codeblock_t* cb, uint32_t label_idx);
-void jle_label(codeblock_t* cb, uint32_t label_idx);
-void jna_label(codeblock_t* cb, uint32_t label_idx);
-void jnae_label(codeblock_t* cb, uint32_t label_idx);
-void jnb_label(codeblock_t* cb, uint32_t label_idx);
-void jnbe_label(codeblock_t* cb, uint32_t label_idx);
-void jnc_label(codeblock_t* cb, uint32_t label_idx);
-void jne_label(codeblock_t* cb, uint32_t label_idx);
-void jng_label(codeblock_t* cb, uint32_t label_idx);
-void jnge_label(codeblock_t* cb, uint32_t label_idx);
-void jnl_label(codeblock_t* cb, uint32_t label_idx);
-void jnle_label(codeblock_t* cb, uint32_t label_idx);
-void jno_label(codeblock_t* cb, uint32_t label_idx);
-void jnp_label(codeblock_t* cb, uint32_t label_idx);
-void jns_label(codeblock_t* cb, uint32_t label_idx);
-void jnz_label(codeblock_t* cb, uint32_t label_idx);
-void jo_label(codeblock_t* cb, uint32_t label_idx);
-void jp_label(codeblock_t* cb, uint32_t label_idx);
-void jpe_label(codeblock_t* cb, uint32_t label_idx);
-void jpo_label(codeblock_t* cb, uint32_t label_idx);
-void js_label(codeblock_t* cb, uint32_t label_idx);
-void jz_label(codeblock_t* cb, uint32_t label_idx);
-void ja_ptr(codeblock_t* cb, uint8_t* ptr);
-void jae_ptr(codeblock_t* cb, uint8_t* ptr);
-void jb_ptr(codeblock_t* cb, uint8_t* ptr);
-void jbe_ptr(codeblock_t* cb, uint8_t* ptr);
-void jc_ptr(codeblock_t* cb, uint8_t* ptr);
-void je_ptr(codeblock_t* cb, uint8_t* ptr);
-void jg_ptr(codeblock_t* cb, uint8_t* ptr);
-void jge_ptr(codeblock_t* cb, uint8_t* ptr);
-void jl_ptr(codeblock_t* cb, uint8_t* ptr);
-void jle_ptr(codeblock_t* cb, uint8_t* ptr);
-void jna_ptr(codeblock_t* cb, uint8_t* ptr);
-void jnae_ptr(codeblock_t* cb, uint8_t* ptr);
-void jnb_ptr(codeblock_t* cb, uint8_t* ptr);
-void jnbe_ptr(codeblock_t* cb, uint8_t* ptr);
-void jnc_ptr(codeblock_t* cb, uint8_t* ptr);
-void jne_ptr(codeblock_t* cb, uint8_t* ptr);
-void jng_ptr(codeblock_t* cb, uint8_t* ptr);
-void jnge_ptr(codeblock_t* cb, uint8_t* ptr);
-void jnl_ptr(codeblock_t* cb, uint8_t* ptr);
-void jnle_ptr(codeblock_t* cb, uint8_t* ptr);
-void jno_ptr(codeblock_t* cb, uint8_t* ptr);
-void jnp_ptr(codeblock_t* cb, uint8_t* ptr);
-void jns_ptr(codeblock_t* cb, uint8_t* ptr);
-void jnz_ptr(codeblock_t* cb, uint8_t* ptr);
-void jo_ptr(codeblock_t* cb, uint8_t* ptr);
-void jp_ptr(codeblock_t* cb, uint8_t* ptr);
-void jpe_ptr(codeblock_t* cb, uint8_t* ptr);
-void jpo_ptr(codeblock_t* cb, uint8_t* ptr);
-void js_ptr(codeblock_t* cb, uint8_t* ptr);
-void jz_ptr(codeblock_t* cb, uint8_t* ptr);
-void jmp_label(codeblock_t* cb, uint32_t label_idx);
-void jmp_ptr(codeblock_t* cb, uint8_t* ptr);
-void jmp_rm(codeblock_t* cb, x86opnd_t opnd);
-void jmp32(codeblock_t* cb, int32_t offset);
-void lea(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void mov(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void movsx(codeblock_t* cb, x86opnd_t dst, x86opnd_t src);
-void neg(codeblock_t* cb, x86opnd_t opnd);
-void nop(codeblock_t* cb, uint32_t length);
-void not(codeblock_t* cb, x86opnd_t opnd);
-void or(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1);
-void pop(codeblock_t* cb, x86opnd_t reg);
-void popfq(codeblock_t* cb);
-void push(codeblock_t* cb, x86opnd_t opnd);
-void pushfq(codeblock_t* cb);
-void ret(codeblock_t* cb);
-void sal(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1);
-void sar(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1);
-void shl(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1);
-void shr(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1);
-void sub(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1);
-void test(codeblock_t* cb, x86opnd_t rm_opnd, x86opnd_t test_opnd);
-void ud2(codeblock_t* cb);
-void xchg(codeblock_t* cb, x86opnd_t rm_opnd, x86opnd_t r_opnd);
-void xor(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1);
-void cb_write_lock_prefix(codeblock_t* cb);
+void add(codeblock_t *cb, x86opnd_t opnd0, x86opnd_t opnd1);
+void and(codeblock_t *cb, x86opnd_t opnd0, x86opnd_t opnd1);
+void call_ptr(codeblock_t *cb, x86opnd_t scratch_reg, uint8_t *dst_ptr);
+void call_label(codeblock_t *cb, uint32_t label_idx);
+void call(codeblock_t *cb, x86opnd_t opnd);
+void cmova(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovae(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovb(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovbe(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovc(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmove(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovg(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovge(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovl(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovle(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovna(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovnae(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovnb(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovnbe(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovnc(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovne(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovng(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovnge(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovnl(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovnle(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovno(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovnp(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovns(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovnz(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovo(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovp(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovpe(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovpo(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovs(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmovz(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void cmp(codeblock_t *cb, x86opnd_t opnd0, x86opnd_t opnd1);
+void cdq(codeblock_t *cb);
+void cqo(codeblock_t *cb);
+void int3(codeblock_t *cb);
+void ja_label(codeblock_t *cb, uint32_t label_idx);
+void jae_label(codeblock_t *cb, uint32_t label_idx);
+void jb_label(codeblock_t *cb, uint32_t label_idx);
+void jbe_label(codeblock_t *cb, uint32_t label_idx);
+void jc_label(codeblock_t *cb, uint32_t label_idx);
+void je_label(codeblock_t *cb, uint32_t label_idx);
+void jg_label(codeblock_t *cb, uint32_t label_idx);
+void jge_label(codeblock_t *cb, uint32_t label_idx);
+void jl_label(codeblock_t *cb, uint32_t label_idx);
+void jle_label(codeblock_t *cb, uint32_t label_idx);
+void jna_label(codeblock_t *cb, uint32_t label_idx);
+void jnae_label(codeblock_t *cb, uint32_t label_idx);
+void jnb_label(codeblock_t *cb, uint32_t label_idx);
+void jnbe_label(codeblock_t *cb, uint32_t label_idx);
+void jnc_label(codeblock_t *cb, uint32_t label_idx);
+void jne_label(codeblock_t *cb, uint32_t label_idx);
+void jng_label(codeblock_t *cb, uint32_t label_idx);
+void jnge_label(codeblock_t *cb, uint32_t label_idx);
+void jnl_label(codeblock_t *cb, uint32_t label_idx);
+void jnle_label(codeblock_t *cb, uint32_t label_idx);
+void jno_label(codeblock_t *cb, uint32_t label_idx);
+void jnp_label(codeblock_t *cb, uint32_t label_idx);
+void jns_label(codeblock_t *cb, uint32_t label_idx);
+void jnz_label(codeblock_t *cb, uint32_t label_idx);
+void jo_label(codeblock_t *cb, uint32_t label_idx);
+void jp_label(codeblock_t *cb, uint32_t label_idx);
+void jpe_label(codeblock_t *cb, uint32_t label_idx);
+void jpo_label(codeblock_t *cb, uint32_t label_idx);
+void js_label(codeblock_t *cb, uint32_t label_idx);
+void jz_label(codeblock_t *cb, uint32_t label_idx);
+void ja_ptr(codeblock_t *cb, uint8_t *ptr);
+void jae_ptr(codeblock_t *cb, uint8_t *ptr);
+void jb_ptr(codeblock_t *cb, uint8_t *ptr);
+void jbe_ptr(codeblock_t *cb, uint8_t *ptr);
+void jc_ptr(codeblock_t *cb, uint8_t *ptr);
+void je_ptr(codeblock_t *cb, uint8_t *ptr);
+void jg_ptr(codeblock_t *cb, uint8_t *ptr);
+void jge_ptr(codeblock_t *cb, uint8_t *ptr);
+void jl_ptr(codeblock_t *cb, uint8_t *ptr);
+void jle_ptr(codeblock_t *cb, uint8_t *ptr);
+void jna_ptr(codeblock_t *cb, uint8_t *ptr);
+void jnae_ptr(codeblock_t *cb, uint8_t *ptr);
+void jnb_ptr(codeblock_t *cb, uint8_t *ptr);
+void jnbe_ptr(codeblock_t *cb, uint8_t *ptr);
+void jnc_ptr(codeblock_t *cb, uint8_t *ptr);
+void jne_ptr(codeblock_t *cb, uint8_t *ptr);
+void jng_ptr(codeblock_t *cb, uint8_t *ptr);
+void jnge_ptr(codeblock_t *cb, uint8_t *ptr);
+void jnl_ptr(codeblock_t *cb, uint8_t *ptr);
+void jnle_ptr(codeblock_t *cb, uint8_t *ptr);
+void jno_ptr(codeblock_t *cb, uint8_t *ptr);
+void jnp_ptr(codeblock_t *cb, uint8_t *ptr);
+void jns_ptr(codeblock_t *cb, uint8_t *ptr);
+void jnz_ptr(codeblock_t *cb, uint8_t *ptr);
+void jo_ptr(codeblock_t *cb, uint8_t *ptr);
+void jp_ptr(codeblock_t *cb, uint8_t *ptr);
+void jpe_ptr(codeblock_t *cb, uint8_t *ptr);
+void jpo_ptr(codeblock_t *cb, uint8_t *ptr);
+void js_ptr(codeblock_t *cb, uint8_t *ptr);
+void jz_ptr(codeblock_t *cb, uint8_t *ptr);
+void jmp_label(codeblock_t *cb, uint32_t label_idx);
+void jmp_ptr(codeblock_t *cb, uint8_t *ptr);
+void jmp_rm(codeblock_t *cb, x86opnd_t opnd);
+void jmp32(codeblock_t *cb, int32_t offset);
+void lea(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void mov(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void movsx(codeblock_t *cb, x86opnd_t dst, x86opnd_t src);
+void neg(codeblock_t *cb, x86opnd_t opnd);
+void nop(codeblock_t *cb, uint32_t length);
+void not(codeblock_t *cb, x86opnd_t opnd);
+void or(codeblock_t *cb, x86opnd_t opnd0, x86opnd_t opnd1);
+void pop(codeblock_t *cb, x86opnd_t reg);
+void popfq(codeblock_t *cb);
+void push(codeblock_t *cb, x86opnd_t opnd);
+void pushfq(codeblock_t *cb);
+void ret(codeblock_t *cb);
+void sal(codeblock_t *cb, x86opnd_t opnd0, x86opnd_t opnd1);
+void sar(codeblock_t *cb, x86opnd_t opnd0, x86opnd_t opnd1);
+void shl(codeblock_t *cb, x86opnd_t opnd0, x86opnd_t opnd1);
+void shr(codeblock_t *cb, x86opnd_t opnd0, x86opnd_t opnd1);
+void sub(codeblock_t *cb, x86opnd_t opnd0, x86opnd_t opnd1);
+void test(codeblock_t *cb, x86opnd_t rm_opnd, x86opnd_t test_opnd);
+void ud2(codeblock_t *cb);
+void xchg(codeblock_t *cb, x86opnd_t rm_opnd, x86opnd_t r_opnd);
+void xor(codeblock_t *cb, x86opnd_t opnd0, x86opnd_t opnd1);
+void cb_write_lock_prefix(codeblock_t *cb);
 
 #endif
diff --git a/yjit_codegen.c b/yjit_codegen.c
index 7e765b2503..7d1d7e31b5 100644
--- a/yjit_codegen.c
+++ b/yjit_codegen.c
@@ -30,11 +30,11 @@ static st_table *yjit_method_codegen_table = NULL;
 
 // Code block into which we write machine code
 static codeblock_t block;
-codeblock_t* cb = NULL;
+codeblock_t *cb = NULL;
 
 // Code block into which we write out-of-line machine code
 static codeblock_t outline_block;
-codeblock_t* ocb = NULL;
+codeblock_t *ocb = NULL;
 
 // Code for exiting back to the interpreter from the leave insn
 static void *leave_exit_code;
@@ -61,7 +61,7 @@ uint32_t yjit_codepage_frozen_bytes = 0;
 // Print the current source location for debugging purposes
 RBIMPL_ATTR_MAYBE_UNUSED()
 static void
-jit_print_loc(jitstate_t* jit, const char* msg)
+jit_print_loc(jitstate_t *jit, const char *msg)
 {
     char *ptr;
     long len;
@@ -82,21 +82,21 @@ jit_obj_info_dump(codeblock_t *cb, x86opnd_t opnd) {
 
 // Get the current instruction's opcode
 static int
-jit_get_opcode(jitstate_t* jit)
+jit_get_opcode(jitstate_t *jit)
 {
     return jit->opcode;
 }
 
 // Get the index of the next instruction
 static uint32_t
-jit_next_idx(jitstate_t* jit)
+jit_next_insn_idx(jitstate_t *jit)
 {
     return jit->insn_idx + insn_len(jit_get_opcode(jit));
 }
 
 // Get an instruction argument by index
 static VALUE
-jit_get_arg(jitstate_t* jit, size_t arg_idx)
+jit_get_arg(jitstate_t *jit, size_t arg_idx)
 {
     RUBY_ASSERT(arg_idx + 1 < (size_t)insn_len(jit_get_opcode(jit)));
     return *(jit->pc + arg_idx + 1);
@@ -104,7 +104,7 @@ jit_get_arg(jitstate_t* jit, size_t arg_idx)
 
 // Load a VALUE into a register and keep track of the reference if it is on the GC heap.
 static void
-jit_mov_gc_ptr(jitstate_t* jit, codeblock_t* cb, x86opnd_t reg, VALUE ptr)
+jit_mov_gc_ptr(jitstate_t *jit, codeblock_t *cb, x86opnd_t reg, VALUE ptr)
 {
     RUBY_ASSERT(reg.type == OPND_REG && reg.num_bits == 64);
 
@@ -124,16 +124,16 @@ jit_mov_gc_ptr(jitstate_t* jit, codeblock_t* cb, x86opnd_t reg, VALUE ptr)
 // Check if we are compiling the instruction at the stub PC
 // Meaning we are compiling the instruction that is next to execute
 static bool
-jit_at_current_insn(jitstate_t* jit)
+jit_at_current_insn(jitstate_t *jit)
 {
-    const VALUE* ec_pc = jit->ec->cfp->pc;
+    const VALUE *ec_pc = jit->ec->cfp->pc;
     return (ec_pc == jit->pc);
 }
 
 // Peek at the nth topmost value on the Ruby stack.
 // Returns the topmost value when n == 0.
 static VALUE
-jit_peek_at_stack(jitstate_t* jit, ctx_t* ctx, int n)
+jit_peek_at_stack(jitstate_t *jit, ctx_t *ctx, int n)
 {
     RUBY_ASSERT(jit_at_current_insn(jit));
 
@@ -168,9 +168,9 @@ jit_peek_at_local(jitstate_t *jit, ctx_t *ctx, int n)
 // Save the incremented PC on the CFP
 // This is necessary when calleees can raise or allocate
 static void
-jit_save_pc(jitstate_t* jit, x86opnd_t scratch_reg)
+jit_save_pc(jitstate_t *jit, x86opnd_t scratch_reg)
 {
-    codeblock_t* cb = jit->cb;
+    codeblock_t *cb = jit->cb;
     mov(cb, scratch_reg, const_ptr_opnd(jit->pc + insn_len(jit->opcode)));
     mov(cb, mem_opnd(64, REG_CFP, offsetof(rb_control_frame_t, pc)), scratch_reg);
 }
@@ -180,11 +180,11 @@ jit_save_pc(jitstate_t* jit, x86opnd_t scratch_reg)
 // Note: this will change the current value of REG_SP,
 //       which could invalidate memory operands
 static void
-jit_save_sp(jitstate_t* jit, ctx_t* ctx)
+jit_save_sp(jitstate_t *jit, ctx_t *ctx)
 {
     if (ctx->sp_offset != 0) {
         x86opnd_t stack_pointer = ctx_sp_opnd(ctx, 0);
-        codeblock_t* cb = jit->cb;
+        codeblock_t *cb = jit->cb;
         lea(cb, REG_SP, stack_pointer);
         mov(cb, member_opnd(REG_CFP, rb_control_frame_t, sp), REG_SP);
         ctx->sp_offset = 0;
@@ -213,13 +213,13 @@ record_global_inval_patch(const codeblock_t *cb, uint32_t outline_block_target_p
     if (!rb_darray_append(&global_inval_patches, patch_point)) rb_bug("allocation failed");
 }
 
-static bool jit_guard_known_klass(jitstate_t *jit, ctx_t* ctx, VALUE known_klass, insn_opnd_t insn_opnd, VALUE sample_instance, const int max_chain_depth, uint8_t *side_exit);
+static bool jit_guard_known_klass(jitstate_t *jit, ctx_t *ctx, VALUE known_klass, insn_opnd_t insn_opnd, VALUE sample_instance, const int max_chain_depth, uint8_t *side_exit);
 
 #if YJIT_STATS
 
 // Add a comment at the current position in the code block
 static void
-_add_comment(codeblock_t* cb, const char* comment_str)
+_add_comment(codeblock_t *cb, const char *comment_str)
 {
     // We can't add comments to the outlined code block
     if (cb == ocb)
@@ -417,7 +417,7 @@ static uint8_t *
 yjit_side_exit(jitstate_t *jit, ctx_t *ctx)
 {
     if (!jit->side_exit_for_pc) {
-        codeblock_t* ocb = jit->ocb;
+        codeblock_t *ocb = jit->ocb;
         uint32_t pos = yjit_gen_exit(jit->pc, ctx, ocb);
         jit->side_exit_for_pc = cb_get_ptr(ocb, pos);
     }
@@ -431,7 +431,7 @@ yjit_side_exit(jitstate_t *jit, ctx_t *ctx)
 // PC for the method isn't necessarily 0, but we always generated code that
 // assumes the entry point is 0.
 static void
-yjit_pc_guard(codeblock_t* cb, const rb_iseq_t *iseq)
+yjit_pc_guard(codeblock_t *cb, const rb_iseq_t *iseq)
 {
     RUBY_ASSERT(cb != NULL);
 
@@ -520,7 +520,7 @@ Compile an interpreter entry block to be inserted into an iseq
 Returns `NULL` if compilation fails.
 */
 uint8_t *
-yjit_entry_prologue(codeblock_t* cb, const rb_iseq_t *iseq)
+yjit_entry_prologue(codeblock_t *cb, const rb_iseq_t *iseq)
 {
     RUBY_ASSERT(cb != NULL);
 
@@ -566,7 +566,7 @@ yjit_entry_prologue(codeblock_t* cb, const rb_iseq_t *iseq)
 // Generate code to check for interrupts and take a side-exit.
 // Warning: this function clobbers REG0
 static void
-yjit_check_ints(codeblock_t* cb, uint8_t* side_exit)
+yjit_check_ints(codeblock_t *cb, uint8_t *side_exit)
 {
     // Check for interrupts
     // see RUBY_VM_CHECK_INTS(ec) macro
@@ -614,7 +614,7 @@ yjit_gen_block(block_t *block, rb_execution_context_t *ec)
 
     // Copy the block's context to avoid mutating it
     ctx_t ctx_copy = block->ctx;
-    ctx_t* ctx = &ctx_copy;
+    ctx_t *ctx = &ctx_copy;
 
     const rb_iseq_t *iseq = block->blockid.iseq;
     uint32_t insn_idx = block->blockid.idx;
@@ -745,17 +745,17 @@ yjit_gen_block(block_t *block, rb_execution_context_t *ec)
     }
 }
 
-static codegen_status_t gen_opt_send_without_block(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb);
+static codegen_status_t gen_opt_send_without_block(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb);
 
 static codegen_status_t
-gen_nop(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_nop(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     // Do nothing
     return YJIT_KEEP_COMPILING;
 }
 
 static codegen_status_t
-gen_dup(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_dup(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     // Get the top value and its type
     x86opnd_t dup_val = ctx_stack_pop(ctx, 0);
@@ -771,7 +771,7 @@ gen_dup(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 
 // duplicate stack top n elements
 static codegen_status_t
-gen_dupn(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_dupn(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     rb_num_t n = (rb_num_t)jit_get_arg(jit, 0);
 
@@ -798,7 +798,7 @@ gen_dupn(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 
 // Swap top 2 stack entries
 static codegen_status_t
-gen_swap(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_swap(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     x86opnd_t opnd0 = ctx_stack_opnd(ctx, 0);
     x86opnd_t opnd1 = ctx_stack_opnd(ctx, 1);
@@ -818,7 +818,7 @@ gen_swap(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 
 // set Nth stack entry to stack top
 static codegen_status_t
-gen_setn(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_setn(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     rb_num_t n = (rb_num_t)jit_get_arg(jit, 0);
 
@@ -836,7 +836,7 @@ gen_setn(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 
 // get nth stack value, then push it
 static codegen_status_t
-gen_topn(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_topn(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     int32_t n = (int32_t)jit_get_arg(jit, 0);
 
@@ -852,7 +852,7 @@ gen_topn(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_pop(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_pop(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     // Decrement SP
     ctx_stack_pop(ctx, 1);
@@ -861,7 +861,7 @@ gen_pop(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 
 // Pop n values off the stack
 static codegen_status_t
-gen_adjuststack(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_adjuststack(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     rb_num_t n = (rb_num_t)jit_get_arg(jit, 0);
     ctx_stack_pop(ctx, n);
@@ -870,7 +870,7 @@ gen_adjuststack(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 
 // new array initialized from top N values
 static codegen_status_t
-gen_newarray(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_newarray(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     rb_num_t n = (rb_num_t)jit_get_arg(jit, 0);
 
@@ -894,7 +894,7 @@ gen_newarray(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 
 // dup array
 static codegen_status_t
-gen_duparray(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_duparray(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     VALUE ary = jit_get_arg(jit, 0);
 
@@ -915,7 +915,7 @@ VALUE rb_vm_splat_array(VALUE flag, VALUE ary);
 
 // call to_a on the array on the stack
 static codegen_status_t
-gen_splatarray(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_splatarray(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     VALUE flag = (VALUE) jit_get_arg(jit, 0);
 
@@ -939,7 +939,7 @@ gen_splatarray(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 
 // new range initialized from top 2 values
 static codegen_status_t
-gen_newrange(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_newrange(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     rb_num_t flag = (rb_num_t)jit_get_arg(jit, 0);
 
@@ -990,7 +990,7 @@ guard_object_is_array(codeblock_t *cb, x86opnd_t object_opnd, x86opnd_t flags_op
 
 // push enough nils onto the stack to fill out an array
 static codegen_status_t
-gen_expandarray(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_expandarray(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     int flag = (int) jit_get_arg(jit, 1);
 
@@ -1072,7 +1072,7 @@ gen_expandarray(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 
 // new hash initialized from top N values
 static codegen_status_t
-gen_newhash(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_newhash(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     rb_num_t n = (rb_num_t)jit_get_arg(jit, 0);
 
@@ -1093,7 +1093,7 @@ gen_newhash(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_putnil(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_putnil(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     // Write constant at SP
     x86opnd_t stack_top = ctx_stack_push(ctx, TYPE_NIL);
@@ -1102,7 +1102,7 @@ gen_putnil(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_putobject(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_putobject(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     VALUE arg = jit_get_arg(jit, 0);
 
@@ -1146,7 +1146,7 @@ gen_putobject(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_putstring(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_putstring(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     VALUE put_val = jit_get_arg(jit, 0);
 
@@ -1164,7 +1164,7 @@ gen_putstring(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_putobject_int2fix(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_putobject_int2fix(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     int opcode = jit_get_opcode(jit);
     int cst_val = (opcode == BIN(putobject_INT2FIX_0_))? 0:1;
@@ -1177,7 +1177,7 @@ gen_putobject_int2fix(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_putself(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_putself(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     // Load self from CFP
     mov(cb, REG0, member_opnd(REG_CFP, rb_control_frame_t, self));
@@ -1190,7 +1190,7 @@ gen_putself(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_putspecialobject(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_putspecialobject(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     enum vm_special_object_type type = (enum vm_special_object_type)jit_get_arg(jit, 0);
 
@@ -1216,7 +1216,7 @@ gen_get_ep(codeblock_t *cb, x86opnd_t reg, uint32_t level)
     while (level--) {
         // Get the previous EP from the current EP
         // See GET_PREV_EP(ep) macro
-        // VALUE* prev_ep = ((VALUE *)((ep)[VM_ENV_DATA_INDEX_SPECVAL] & ~0x03))
+        // VALUE *prev_ep = ((VALUE *)((ep)[VM_ENV_DATA_INDEX_SPECVAL] & ~0x03))
         mov(cb, reg, mem_opnd(64, REG0, SIZEOF_VALUE * VM_ENV_DATA_INDEX_SPECVAL));
         and(cb, reg, imm_opnd(~0x03));
     }
@@ -1235,7 +1235,7 @@ slot_to_local_idx(const rb_iseq_t *iseq, int32_t slot_idx)
 }
 
 static codegen_status_t
-gen_getlocal_wc0(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_getlocal_wc0(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     // Compute the offset from BP to the local
     int32_t slot_idx = (int32_t)jit_get_arg(jit, 0);
@@ -1256,7 +1256,7 @@ gen_getlocal_wc0(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_getlocal_generic(ctx_t* ctx, uint32_t local_idx, uint32_t level)
+gen_getlocal_generic(ctx_t *ctx, uint32_t local_idx, uint32_t level)
 {
     gen_get_ep(cb, REG0, level);
 
@@ -1273,7 +1273,7 @@ gen_getlocal_generic(ctx_t* ctx, uint32_t local_idx, uint32_t level)
 }
 
 static codegen_status_t
-gen_getlocal(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_getlocal(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     int32_t idx = (int32_t)jit_get_arg(jit, 0);
     int32_t level = (int32_t)jit_get_arg(jit, 1);
@@ -1281,14 +1281,14 @@ gen_getlocal(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_getlocal_wc1(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_getlocal_wc1(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     int32_t idx = (int32_t)jit_get_arg(jit, 0);
     return gen_getlocal_generic(ctx, idx, 1);
 }
 
 static codegen_status_t
-gen_setlocal_wc0(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_setlocal_wc0(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     /*
     vm_env_write(const VALUE *ep, int index, VALUE v)
@@ -1335,7 +1335,7 @@ gen_setlocal_wc0(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_setlocal_generic(jitstate_t *jit, ctx_t* ctx, uint32_t local_idx, uint32_t level)
+gen_setlocal_generic(jitstate_t *jit, ctx_t *ctx, uint32_t local_idx, uint32_t level)
 {
     // Load environment pointer EP at level
     gen_get_ep(cb, REG0, level);
@@ -1362,7 +1362,7 @@ gen_setlocal_generic(jitstate_t *jit, ctx_t* ctx, uint32_t local_idx, uint32_t l
 }
 
 static codegen_status_t
-gen_setlocal(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_setlocal(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     int32_t idx = (int32_t)jit_get_arg(jit, 0);
     int32_t level = (int32_t)jit_get_arg(jit, 1);
@@ -1370,7 +1370,7 @@ gen_setlocal(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_setlocal_wc1(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_setlocal_wc1(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     int32_t idx = (int32_t)jit_get_arg(jit, 0);
     return gen_setlocal_generic(jit, ctx, idx, 1);
@@ -1686,7 +1686,7 @@ gen_get_ivar(jitstate_t *jit, ctx_t *ctx, const int max_chain_depth, VALUE compt
 }
 
 static codegen_status_t
-gen_getinstancevariable(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_getinstancevariable(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     // Defer compilation so we can specialize on a runtime `self`
     if (!jit_at_current_insn(jit)) {
@@ -1714,7 +1714,7 @@ gen_getinstancevariable(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 void rb_vm_setinstancevariable(const rb_iseq_t *iseq, VALUE obj, ID id, VALUE val, IVC ic);
 
 static codegen_status_t
-gen_setinstancevariable(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_setinstancevariable(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     ID id = (ID)jit_get_arg(jit, 0);
     IVC ic = (IVC)jit_get_arg(jit, 1);
@@ -1740,7 +1740,7 @@ gen_setinstancevariable(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 bool rb_vm_defined(rb_execution_context_t *ec, rb_control_frame_t *reg_cfp, rb_num_t op_type, VALUE obj, VALUE v);
 
 static codegen_status_t
-gen_defined(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_defined(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     rb_num_t op_type = (rb_num_t)jit_get_arg(jit, 0);
     VALUE obj = (VALUE)jit_get_arg(jit, 1);
@@ -1778,7 +1778,7 @@ gen_defined(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_checktype(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_checktype(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     enum ruby_value_type type_val = (enum ruby_value_type)jit_get_arg(jit, 0);
     // Only three types are emitted by compile.c
@@ -1837,7 +1837,7 @@ gen_checktype(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_concatstrings(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_concatstrings(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     rb_num_t n = (rb_num_t)jit_get_arg(jit, 0);
 
@@ -1859,7 +1859,7 @@ gen_concatstrings(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static void
-guard_two_fixnums(ctx_t* ctx, uint8_t* side_exit)
+guard_two_fixnums(ctx_t *ctx, uint8_t *side_exit)
 {
     // Get the stack operand types
     val_type_t arg1_type = ctx_get_opnd_type(ctx, OPND_STACK(0));
@@ -1907,10 +1907,10 @@ guard_two_fixnums(ctx_t* ctx, uint8_t* side_exit)
 }
 
 // Conditional move operation used by comparison operators
-typedef void (*cmov_fn)(codeblock_t* cb, x86opnd_t opnd0, x86opnd_t opnd1);
+typedef void (*cmov_fn)(codeblock_t *cb, x86opnd_t opnd0, x86opnd_t opnd1);
 
 static codegen_status_t
-gen_fixnum_cmp(jitstate_t* jit, ctx_t* ctx, cmov_fn cmov_op)
+gen_fixnum_cmp(jitstate_t *jit, ctx_t *ctx, cmov_fn cmov_op)
 {
     // Defer compilation so we can specialize base on a runtime receiver
     if (!jit_at_current_insn(jit)) {
@@ -1924,7 +1924,7 @@ gen_fixnum_cmp(jitstate_t* jit, ctx_t* ctx, cmov_fn cmov_op)
     if (FIXNUM_P(comptime_a) && FIXNUM_P(comptime_b)) {
         // Create a size-exit to fall back to the interpreter
         // Note: we generate the side-exit before popping operands from the stack
-        uint8_t* side_exit = yjit_side_exit(jit, ctx);
+        uint8_t *side_exit = yjit_side_exit(jit, ctx);
 
         if (!assume_bop_not_redefined(jit->block, INTEGER_REDEFINED_OP_FLAG, BOP_LT)) {
             return YJIT_CANT_COMPILE;
@@ -1955,25 +1955,25 @@ gen_fixnum_cmp(jitstate_t* jit, ctx_t* ctx, cmov_fn cmov_op)
 }
 
 static codegen_status_t
-gen_opt_lt(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_lt(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     return gen_fixnum_cmp(jit, ctx, cmovl);
 }
 
 static codegen_status_t
-gen_opt_le(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_le(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     return gen_fixnum_cmp(jit, ctx, cmovle);
 }
 
 static codegen_status_t
-gen_opt_ge(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_ge(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     return gen_fixnum_cmp(jit, ctx, cmovge);
 }
 
 static codegen_status_t
-gen_opt_gt(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_gt(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     return gen_fixnum_cmp(jit, ctx, cmovg);
 }
@@ -1981,7 +1981,7 @@ gen_opt_gt(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 // Implements specialized equality for either two fixnum or two strings
 // Returns true if code was generated, otherwise false
 bool
-gen_equality_specialized(jitstate_t* jit, ctx_t* ctx, uint8_t *side_exit)
+gen_equality_specialized(jitstate_t *jit, ctx_t *ctx, uint8_t *side_exit)
 {
     VALUE comptime_a = jit_peek_at_stack(jit, ctx, 1);
     VALUE comptime_b = jit_peek_at_stack(jit, ctx, 0);
@@ -2056,7 +2056,7 @@ gen_equality_specialized(jitstate_t* jit, ctx_t* ctx, uint8_t *side_exit)
 }
 
 static codegen_status_t
-gen_opt_eq(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_eq(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     // Defer compilation so we can specialize base on a runtime receiver
     if (!jit_at_current_insn(jit)) {
@@ -2078,7 +2078,7 @@ gen_opt_eq(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 static codegen_status_t gen_send_general(jitstate_t *jit, ctx_t *ctx, struct rb_call_data *cd, rb_iseq_t *block);
 
 static codegen_status_t
-gen_opt_neq(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_neq(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     // opt_neq is passed two rb_call_data as arguments:
     // first for ==, second for !=
@@ -2087,7 +2087,7 @@ gen_opt_neq(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_opt_aref(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_aref(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     struct rb_call_data * cd = (struct rb_call_data *)jit_get_arg(jit, 0);
     int32_t argc = (int32_t)vm_ci_argc(cd->ci);
@@ -2201,7 +2201,7 @@ gen_opt_aref(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_opt_aset(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_aset(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     // Defer compilation so we can specialize on a runtime `self`
     if (!jit_at_current_insn(jit)) {
@@ -2218,7 +2218,7 @@ gen_opt_aset(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
     x86opnd_t val = ctx_stack_opnd(ctx, 0);
 
     if (CLASS_OF(comptime_recv) == rb_cArray && FIXNUM_P(comptime_key)) {
-        uint8_t* side_exit = yjit_side_exit(jit, ctx);
+        uint8_t *side_exit = yjit_side_exit(jit, ctx);
 
         // Guard receiver is an Array
         mov(cb, REG0, recv);
@@ -2251,7 +2251,7 @@ gen_opt_aset(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
         jit_jump_to_next_insn(jit, ctx);
         return YJIT_END_BLOCK;
     } else if (CLASS_OF(comptime_recv) == rb_cHash) {
-        uint8_t* side_exit = yjit_side_exit(jit, ctx);
+        uint8_t *side_exit = yjit_side_exit(jit, ctx);
 
         // Guard receiver is a Hash
         mov(cb, REG0, recv);
@@ -2280,7 +2280,7 @@ gen_opt_aset(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_opt_and(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_and(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     // Defer compilation so we can specialize on a runtime `self`
     if (!jit_at_current_insn(jit)) {
@@ -2294,7 +2294,7 @@ gen_opt_and(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
     if (FIXNUM_P(comptime_a) && FIXNUM_P(comptime_b)) {
         // Create a size-exit to fall back to the interpreter
         // Note: we generate the side-exit before popping operands from the stack
-        uint8_t* side_exit = yjit_side_exit(jit, ctx);
+        uint8_t *side_exit = yjit_side_exit(jit, ctx);
 
         if (!assume_bop_not_redefined(jit->block, INTEGER_REDEFINED_OP_FLAG, BOP_AND)) {
             return YJIT_CANT_COMPILE;
@@ -2323,7 +2323,7 @@ gen_opt_and(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_opt_or(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_or(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     // Defer compilation so we can specialize on a runtime `self`
     if (!jit_at_current_insn(jit)) {
@@ -2337,7 +2337,7 @@ gen_opt_or(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
     if (FIXNUM_P(comptime_a) && FIXNUM_P(comptime_b)) {
         // Create a size-exit to fall back to the interpreter
         // Note: we generate the side-exit before popping operands from the stack
-        uint8_t* side_exit = yjit_side_exit(jit, ctx);
+        uint8_t *side_exit = yjit_side_exit(jit, ctx);
 
         if (!assume_bop_not_redefined(jit->block, INTEGER_REDEFINED_OP_FLAG, BOP_OR)) {
             return YJIT_CANT_COMPILE;
@@ -2366,7 +2366,7 @@ gen_opt_or(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_opt_minus(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_minus(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     // Defer compilation so we can specialize on a runtime `self`
     if (!jit_at_current_insn(jit)) {
@@ -2380,7 +2380,7 @@ gen_opt_minus(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
     if (FIXNUM_P(comptime_a) && FIXNUM_P(comptime_b)) {
         // Create a size-exit to fall back to the interpreter
         // Note: we generate the side-exit before popping operands from the stack
-        uint8_t* side_exit = yjit_side_exit(jit, ctx);
+        uint8_t *side_exit = yjit_side_exit(jit, ctx);
 
         if (!assume_bop_not_redefined(jit->block, INTEGER_REDEFINED_OP_FLAG, BOP_MINUS)) {
             return YJIT_CANT_COMPILE;
@@ -2411,7 +2411,7 @@ gen_opt_minus(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_opt_plus(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_plus(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     // Defer compilation so we can specialize on a runtime `self`
     if (!jit_at_current_insn(jit)) {
@@ -2425,7 +2425,7 @@ gen_opt_plus(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
     if (FIXNUM_P(comptime_a) && FIXNUM_P(comptime_b)) {
         // Create a size-exit to fall back to the interpreter
         // Note: we generate the side-exit before popping operands from the stack
-        uint8_t* side_exit = yjit_side_exit(jit, ctx);
+        uint8_t *side_exit = yjit_side_exit(jit, ctx);
 
         if (!assume_bop_not_redefined(jit->block, INTEGER_REDEFINED_OP_FLAG, BOP_PLUS)) {
             return YJIT_CANT_COMPILE;
@@ -2456,14 +2456,14 @@ gen_opt_plus(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_opt_mult(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_mult(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     // Delegate to send, call the method on the recv
     return gen_opt_send_without_block(jit, ctx, cb);
 }
 
 static codegen_status_t
-gen_opt_div(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_div(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     // Delegate to send, call the method on the recv
     return gen_opt_send_without_block(jit, ctx, cb);
@@ -2472,13 +2472,13 @@ gen_opt_div(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 VALUE rb_vm_opt_mod(VALUE recv, VALUE obj);
 
 static codegen_status_t
-gen_opt_mod(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_mod(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     // Save the PC and SP because the callee may allocate bignums
     // Note that this modifies REG_SP, which is why we do it first
     jit_prepare_routine_call(jit, ctx, REG0);
 
-    uint8_t* side_exit = yjit_side_exit(jit, ctx);
+    uint8_t *side_exit = yjit_side_exit(jit, ctx);
 
     // Get the operands from the stack
     x86opnd_t arg1 = ctx_stack_pop(ctx, 1);
@@ -2501,28 +2501,28 @@ gen_opt_mod(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_opt_ltlt(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_ltlt(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     // Delegate to send, call the method on the recv
     return gen_opt_send_without_block(jit, ctx, cb);
 }
 
 static codegen_status_t
-gen_opt_nil_p(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_nil_p(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     // Delegate to send, call the method on the recv
     return gen_opt_send_without_block(jit, ctx, cb);
 }
 
 static codegen_status_t
-gen_opt_empty_p(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_empty_p(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     // Delegate to send, call the method on the recv
     return gen_opt_send_without_block(jit, ctx, cb);
 }
 
 static codegen_status_t
-gen_opt_str_freeze(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_str_freeze(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     if (!assume_bop_not_redefined(jit->block, STRING_REDEFINED_OP_FLAG, BOP_FREEZE)) {
         return YJIT_CANT_COMPILE;
@@ -2539,7 +2539,7 @@ gen_opt_str_freeze(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_opt_str_uminus(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_str_uminus(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     if (!assume_bop_not_redefined(jit->block, STRING_REDEFINED_OP_FLAG, BOP_UMINUS)) {
         return YJIT_CANT_COMPILE;
@@ -2556,31 +2556,31 @@ gen_opt_str_uminus(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_opt_not(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_not(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     return gen_opt_send_without_block(jit, ctx, cb);
 }
 
 static codegen_status_t
-gen_opt_size(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_size(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     return gen_opt_send_without_block(jit, ctx, cb);
 }
 
 static codegen_status_t
-gen_opt_length(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_length(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     return gen_opt_send_without_block(jit, ctx, cb);
 }
 
 static codegen_status_t
-gen_opt_regexpmatch2(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_regexpmatch2(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     return gen_opt_send_without_block(jit, ctx, cb);
 }
 
 static codegen_status_t
-gen_opt_case_dispatch(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_case_dispatch(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     // Normally this instruction would lookup the key in a hash and jump to an
     // offset based on that.
@@ -2596,7 +2596,7 @@ gen_opt_case_dispatch(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 void
-gen_branchif_branch(codeblock_t* cb, uint8_t* target0, uint8_t* target1, uint8_t shape)
+gen_branchif_branch(codeblock_t *cb, uint8_t *target0, uint8_t *target1, uint8_t shape)
 {
     switch (shape)
     {
@@ -2616,13 +2616,13 @@ gen_branchif_branch(codeblock_t* cb, uint8_t* target0, uint8_t* target1, uint8_t
 }
 
 static codegen_status_t
-gen_branchif(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_branchif(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     int32_t jump_offset = (int32_t)jit_get_arg(jit, 0);
 
     // Check for interrupts, but only on backward branches that may create loops
     if (jump_offset < 0) {
-        uint8_t* side_exit = yjit_side_exit(jit, ctx);
+        uint8_t *side_exit = yjit_side_exit(jit, ctx);
         yjit_check_ints(cb, side_exit);
     }
 
@@ -2653,7 +2653,7 @@ gen_branchif(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 void
-gen_branchunless_branch(codeblock_t* cb, uint8_t* target0, uint8_t* target1, uint8_t shape)
+gen_branchunless_branch(codeblock_t *cb, uint8_t *target0, uint8_t *target1, uint8_t shape)
 {
     switch (shape)
     {
@@ -2673,13 +2673,13 @@ gen_branchunless_branch(codeblock_t* cb, uint8_t* target0, uint8_t* target1, uin
 }
 
 static codegen_status_t
-gen_branchunless(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_branchunless(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     int32_t jump_offset = (int32_t)jit_get_arg(jit, 0);
 
     // Check for interrupts, but only on backward branches that may create loops
     if (jump_offset < 0) {
-        uint8_t* side_exit = yjit_side_exit(jit, ctx);
+        uint8_t *side_exit = yjit_side_exit(jit, ctx);
         yjit_check_ints(cb, side_exit);
     }
 
@@ -2710,7 +2710,7 @@ gen_branchunless(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 void
-gen_branchnil_branch(codeblock_t* cb, uint8_t* target0, uint8_t* target1, uint8_t shape)
+gen_branchnil_branch(codeblock_t *cb, uint8_t *target0, uint8_t *target1, uint8_t shape)
 {
     switch (shape)
     {
@@ -2730,13 +2730,13 @@ gen_branchnil_branch(codeblock_t* cb, uint8_t* target0, uint8_t* target1, uint8_
 }
 
 static codegen_status_t
-gen_branchnil(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_branchnil(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     int32_t jump_offset = (int32_t)jit_get_arg(jit, 0);
 
     // Check for interrupts, but only on backward branches that may create loops
     if (jump_offset < 0) {
-        uint8_t* side_exit = yjit_side_exit(jit, ctx);
+        uint8_t *side_exit = yjit_side_exit(jit, ctx);
         yjit_check_ints(cb, side_exit);
     }
 
@@ -2766,13 +2766,13 @@ gen_branchnil(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_jump(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_jump(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     int32_t jump_offset = (int32_t)jit_get_arg(jit, 0);
 
     // Check for interrupts, but only on backward branches that may create loops
     if (jump_offset < 0) {
-        uint8_t* side_exit = yjit_side_exit(jit, ctx);
+        uint8_t *side_exit = yjit_side_exit(jit, ctx);
         yjit_check_ints(cb, side_exit);
     }
 
@@ -3301,7 +3301,7 @@ gen_send_cfunc(jitstate_t *jit, ctx_t *ctx, const struct rb_callinfo *ci, const
 }
 
 static void
-gen_return_branch(codeblock_t* cb, uint8_t* target0, uint8_t* target1, uint8_t shape)
+gen_return_branch(codeblock_t *cb, uint8_t *target0, uint8_t *target1, uint8_t shape)
 {
     switch (shape)
     {
@@ -3742,14 +3742,14 @@ gen_send_general(jitstate_t *jit, ctx_t *ctx, struct rb_call_data *cd, rb_iseq_t
 }
 
 static codegen_status_t
-gen_opt_send_without_block(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_send_without_block(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     struct rb_call_data *cd = (struct rb_call_data *)jit_get_arg(jit, 0);
     return gen_send_general(jit, ctx, cd, NULL);
 }
 
 static codegen_status_t
-gen_send(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_send(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     struct rb_call_data *cd = (struct rb_call_data *)jit_get_arg(jit, 0);
     rb_iseq_t *block = (rb_iseq_t *)jit_get_arg(jit, 1);
@@ -3757,7 +3757,7 @@ gen_send(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_invokesuper(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_invokesuper(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     struct rb_call_data *cd = (struct rb_call_data *)jit_get_arg(jit, 0);
     rb_iseq_t *block = (rb_iseq_t *)jit_get_arg(jit, 1);
@@ -3893,13 +3893,13 @@ gen_invokesuper(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_leave(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_leave(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     // Only the return value should be on the stack
     RUBY_ASSERT(ctx->stack_size == 1);
 
     // Create a size-exit to fall back to the interpreter
-    uint8_t* side_exit = yjit_side_exit(jit, ctx);
+    uint8_t *side_exit = yjit_side_exit(jit, ctx);
 
     // Load environment pointer EP from CFP
     mov(cb, REG1, member_opnd(REG_CFP, rb_control_frame_t, ep));
@@ -3931,7 +3931,7 @@ gen_leave(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 RUBY_EXTERN rb_serial_t ruby_vm_global_constant_state;
 
 static codegen_status_t
-gen_getglobal(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_getglobal(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     ID gid = jit_get_arg(jit, 0);
 
@@ -3949,7 +3949,7 @@ gen_getglobal(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_setglobal(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_setglobal(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     ID gid = jit_get_arg(jit, 0);
 
@@ -3969,7 +3969,7 @@ gen_setglobal(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_tostring(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_tostring(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     // Save the PC and SP because we might make a Ruby call for
     // Kernel#set_trace_var
@@ -3991,7 +3991,7 @@ gen_tostring(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_toregexp(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_toregexp(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     rb_num_t opt = jit_get_arg(jit, 0);
     rb_num_t cnt = jit_get_arg(jit, 1);
@@ -4031,7 +4031,7 @@ gen_toregexp(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_getspecial(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_getspecial(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     // This takes two arguments, key and type
     // key is only used when type == 0
@@ -4102,7 +4102,7 @@ gen_getspecial(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_opt_getinlinecache(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_getinlinecache(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     VALUE jump_offset = jit_get_arg(jit, 0);
     VALUE const_cache_as_value = jit_get_arg(jit, 1);
@@ -4168,7 +4168,7 @@ gen_opt_getinlinecache(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 // interpreter's scheme for avoiding Proc allocations when delegating
 // explict block parameters.
 static codegen_status_t
-gen_getblockparamproxy(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_getblockparamproxy(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     // A mirror of the interpreter code. Checking for the case
     // where it's pushing rb_block_param_proxy.
@@ -4205,7 +4205,7 @@ gen_getblockparamproxy(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 }
 
 static codegen_status_t
-gen_invokebuiltin(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_invokebuiltin(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     const struct rb_builtin_function *bf = (struct rb_builtin_function *)jit_get_arg(jit, 0);
 
@@ -4242,7 +4242,7 @@ gen_invokebuiltin(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
 // invokebuiltin does, but instead of taking arguments from the top of the
 // stack uses the argument locals (and self) from the current method.
 static codegen_status_t
-gen_opt_invokebuiltin_delegate(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb)
+gen_opt_invokebuiltin_delegate(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb)
 {
     const struct rb_builtin_function *bf = (struct rb_builtin_function *)jit_get_arg(jit, 0);
     int32_t start_index = (int32_t)jit_get_arg(jit, 1);
diff --git a/yjit_codegen.h b/yjit_codegen.h
index 6d5fc27479..cbb3640dea 100644
--- a/yjit_codegen.h
+++ b/yjit_codegen.h
@@ -16,11 +16,11 @@ typedef enum codegen_status {
 } codegen_status_t;
 
 // Code generation function signature
-typedef codegen_status_t (*codegen_fn)(jitstate_t* jit, ctx_t* ctx, codeblock_t* cb);
+typedef codegen_status_t (*codegen_fn)(jitstate_t *jit, ctx_t *ctx, codeblock_t *cb);
 
-uint8_t* yjit_entry_prologue(codeblock_t* cb, const rb_iseq_t* iseq);
+uint8_t *yjit_entry_prologue(codeblock_t *cb, const rb_iseq_t *iseq);
 
-void yjit_gen_block(block_t* block, rb_execution_context_t* ec);
+void yjit_gen_block(block_t *block, rb_execution_context_t *ec);
 
 void yjit_init_codegen(void);
 
diff --git a/yjit_core.c b/yjit_core.c
index 64f7b21156..f3bae6cc7d 100644
--- a/yjit_core.c
+++ b/yjit_core.c
@@ -14,7 +14,7 @@
 Get an operand for the adjusted stack pointer address
 */
 x86opnd_t
-ctx_sp_opnd(ctx_t* ctx, int32_t offset_bytes)
+ctx_sp_opnd(ctx_t *ctx, int32_t offset_bytes)
 {
     int32_t offset = (ctx->sp_offset * sizeof(VALUE)) + offset_bytes;
     return mem_opnd(64, REG_SP, offset);
@@ -25,7 +25,7 @@ Push one new value on the temp stack with an explicit mapping
 Return a pointer to the new stack top
 */
 x86opnd_t
-ctx_stack_push_mapping(ctx_t* ctx, temp_type_mapping_t mapping)
+ctx_stack_push_mapping(ctx_t *ctx, temp_type_mapping_t mapping)
 {
     // Keep track of the type and mapping of the value
     if (ctx->stack_size < MAX_TEMP_TYPES) {
@@ -51,7 +51,7 @@ Push one new value on the temp stack
 Return a pointer to the new stack top
 */
 x86opnd_t
-ctx_stack_push(ctx_t* ctx, val_type_t type)
+ctx_stack_push(ctx_t *ctx, val_type_t type)
 {
     temp_type_mapping_t mapping = { MAP_STACK, type };
     return ctx_stack_push_mapping(ctx, mapping);
@@ -61,7 +61,7 @@ ctx_stack_push(ctx_t* ctx, val_type_t type)
 Push the self value on the stack
 */
 x86opnd_t
-ctx_stack_push_self(ctx_t* ctx)
+ctx_stack_push_self(ctx_t *ctx)
 {
     temp_type_mapping_t mapping = { MAP_SELF, TYPE_UNKNOWN };
     return ctx_stack_push_mapping(ctx, mapping);
@@ -71,7 +71,7 @@ ctx_stack_push_self(ctx_t* ctx)
 Push a local variable on the stack
 */
 x86opnd_t
-ctx_stack_push_local(ctx_t* ctx, size_t local_idx)
+ctx_stack_push_local(ctx_t *ctx, size_t local_idx)
 {
     if (local_idx >= MAX_LOCAL_TYPES) {
         return ctx_stack_push(ctx, TYPE_UNKNOWN);
@@ -89,7 +89,7 @@ Pop N values off the stack
 Return a pointer to the stack top before the pop operation
 */
 x86opnd_t
-ctx_stack_pop(ctx_t* ctx, size_t n)
+ctx_stack_pop(ctx_t *ctx, size_t n)
 {
     RUBY_ASSERT(n <= ctx->stack_size);
 
@@ -117,7 +117,7 @@ ctx_stack_pop(ctx_t* ctx, size_t n)
 Get an operand pointing to a slot on the temp stack
 */
 x86opnd_t
-ctx_stack_opnd(ctx_t* ctx, int32_t idx)
+ctx_stack_opnd(ctx_t *ctx, int32_t idx)
 {
     // SP points just above the topmost value
     int32_t offset = (ctx->sp_offset - 1 - idx) * sizeof(VALUE);
@@ -130,7 +130,7 @@ ctx_stack_opnd(ctx_t* ctx, int32_t idx)
 Get the type of an instruction operand
 */
 val_type_t
-ctx_get_opnd_type(const ctx_t* ctx, insn_opnd_t opnd)
+ctx_get_opnd_type(const ctx_t *ctx, insn_opnd_t opnd)
 {
     if (opnd.is_self)
         return ctx->self_type;
@@ -172,7 +172,7 @@ This value must be compatible and at least as specific as the previously known t
 If this value originated from self, or an lvar, the learned type will be
 propagated back to its source.
 */
-void ctx_upgrade_opnd_type(ctx_t* ctx, insn_opnd_t opnd, val_type_t type)
+void ctx_upgrade_opnd_type(ctx_t *ctx, insn_opnd_t opnd, val_type_t type)
 {
     if (opnd.is_self) {
         UPGRADE_TYPE(ctx->self_type, type);
@@ -211,7 +211,7 @@ This is can be used with ctx_stack_push_mapping or ctx_set_opnd_mapping to copy
 a stack value's type while maintaining the mapping.
 */
 temp_type_mapping_t
-ctx_get_opnd_mapping(const ctx_t* ctx, insn_opnd_t opnd)
+ctx_get_opnd_mapping(const ctx_t *ctx, insn_opnd_t opnd)
 {
     temp_type_mapping_t type_mapping;
     type_mapping.type = ctx_get_opnd_type(ctx, opnd);
@@ -240,7 +240,7 @@ ctx_get_opnd_mapping(const ctx_t* ctx, insn_opnd_t opnd)
 Overwrite both the type and mapping of a stack operand.
 */
 void
-ctx_set_opnd_mapping(ctx_t* ctx, insn_opnd_t opnd, temp_type_mapping_t type_mapping)
+ctx_set_opnd_mapping(ctx_t *ctx, insn_opnd_t opnd, temp_type_mapping_t type_mapping)
 {
     // self is always MAP_SELF
     RUBY_ASSERT(!opnd.is_self);
@@ -261,7 +261,7 @@ ctx_set_opnd_mapping(ctx_t* ctx, insn_opnd_t opnd, temp_type_mapping_t type_mapp
 /**
 Set the type of a local variable
 */
-void ctx_set_local_type(ctx_t* ctx, size_t idx, val_type_t type)
+void ctx_set_local_type(ctx_t *ctx, size_t idx, val_type_t type)
 {
     if (idx >= MAX_LOCAL_TYPES)
         return;
@@ -280,7 +280,7 @@ void ctx_set_local_type(ctx_t* ctx, size_t idx, val_type_t type)
 
 // Erase local variable type information
 // eg: because of a call we can't track
-void ctx_clear_local_types(ctx_t* ctx)
+void ctx_clear_local_types(ctx_t *ctx)
 {
     // When clearing local types we must detach any stack mappings to those
     // locals. Even if local values may have changed, stack values will not.
@@ -412,7 +412,7 @@ Returns 0 if the two contexts are the same
 Returns > 0 if different but compatible
 Returns INT_MAX if incompatible
 */
-int ctx_diff(const ctx_t* src, const ctx_t* dst)
+int ctx_diff(const ctx_t *src, const ctx_t *dst)
 {
     // Can only lookup the first version in the chain
     if (dst->chain_depth != 0)
@@ -504,7 +504,7 @@ static size_t get_num_versions(blockid_t blockid)
 
 // Keep track of a block version. Block should be fully constructed.
 static void
-add_block_version(blockid_t blockid, block_t* block)
+add_block_version(blockid_t blockid, block_t *block)
 {
     const rb_iseq_t *iseq = block->blockid.iseq;
     struct rb_iseq_constant_body *body = iseq->body;
@@ -565,12 +565,12 @@ add_block_version(blockid_t blockid, block_t* block)
 
 // Create a new outgoing branch entry for a block
 static branch_t*
-make_branch_entry(block_t* block, const ctx_t* src_ctx, branchgen_fn gen_fn)
+make_branch_entry(block_t *block, const ctx_t *src_ctx, branchgen_fn gen_fn)
 {
     RUBY_ASSERT(block != NULL);
 
     // Allocate and zero-initialize
-    branch_t* branch = calloc(1, sizeof(branch_t));
+    branch_t *branch = calloc(1, sizeof(branch_t));
 
     branch->block = block;
     branch->src_ctx = *src_ctx;
@@ -584,12 +584,12 @@ make_branch_entry(block_t* block, const ctx_t* src_ctx, branchgen_fn gen_fn)
 }
 
 // Retrieve a basic block version for an (iseq, idx) tuple
-block_t* find_block_version(blockid_t blockid, const ctx_t* ctx)
+block_t *find_block_version(blockid_t blockid, const ctx_t *ctx)
 {
     rb_yjit_block_array_t versions = yjit_get_version_array(blockid.iseq, blockid.idx);
 
     // Best match found
-    block_t* best_version = NULL;
+    block_t *best_version = NULL;
     int best_diff = INT_MAX;
 
     // For each version matching the blockid
@@ -619,7 +619,7 @@ block_t* find_block_version(blockid_t blockid, const ctx_t* ctx)
 
 // Produce a generic context when the block version limit is hit for a blockid
 // Note that this will mutate the ctx argument
-void limit_block_versions(blockid_t blockid, ctx_t* ctx)
+void limit_block_versions(blockid_t blockid, ctx_t *ctx)
 {
     // Guard chains implement limits separately, do nothing
     if (ctx->chain_depth > 0)
@@ -641,15 +641,15 @@ void limit_block_versions(blockid_t blockid, ctx_t* ctx)
 }
 
 // Compile a new block version immediately
-block_t* gen_block_version(blockid_t blockid, const ctx_t* start_ctx, rb_execution_context_t* ec)
+block_t *gen_block_version(blockid_t blockid, const ctx_t *start_ctx, rb_execution_context_t *ec)
 {
     // Allocate a new block version object
-    block_t* block = calloc(1, sizeof(block_t));
+    block_t *block = calloc(1, sizeof(block_t));
     block->blockid = blockid;
     memcpy(&block->ctx, start_ctx, sizeof(ctx_t));
 
     // Store a pointer to the first block (returned by this function)
-    block_t* first_block = block;
+    block_t *first_block = block;
 
     // Limit the number of specialized versions for this block
     limit_block_versions(block->blockid, &block->ctx);
@@ -668,7 +668,7 @@ block_t* gen_block_version(blockid_t blockid, const ctx_t* start_ctx, rb_executi
         }
 
         // Get the last outgoing branch from the previous block
-        branch_t* last_branch = rb_darray_back(block->outgoing);
+        branch_t *last_branch = rb_darray_back(block->outgoing);
 
         // If there is no next block to compile, stop
         if (last_branch->dst_addrs[0] || last_branch->dst_addrs[1]) {
@@ -707,7 +707,7 @@ block_t* gen_block_version(blockid_t blockid, const ctx_t* start_ctx, rb_executi
 }
 
 // Generate a block version that is an entry point inserted into an iseq
-uint8_t* gen_entry_point(const rb_iseq_t *iseq, uint32_t insn_idx, rb_execution_context_t *ec)
+uint8_t *gen_entry_point(const rb_iseq_t *iseq, uint32_t insn_idx, rb_execution_context_t *ec)
 {
     // If we aren't at PC 0, don't generate code
     // See yjit_pc_guard
@@ -719,10 +719,10 @@ uint8_t* gen_entry_point(const rb_iseq_t *iseq, uint32_t insn_idx, rb_execution_
     blockid_t blockid = { iseq, insn_idx };
 
     // Write the interpreter entry prologue
-    uint8_t* code_ptr = yjit_entry_prologue(cb, iseq);
+    uint8_t *code_ptr = yjit_entry_prologue(cb, iseq);
 
     // Try to generate code for the entry block
-    block_t* block = gen_block_version(blockid, &DEFAULT_CTX, ec);
+    block_t *block = gen_block_version(blockid, &DEFAULT_CTX, ec);
 
     // If we couldn't generate any code
     if (block->end_idx == insn_idx)
@@ -736,9 +736,9 @@ uint8_t* gen_entry_point(const rb_iseq_t *iseq, uint32_t insn_idx, rb_execution_
 // Called by the generated code when a branch stub is executed
 // Triggers compilation of branches and code patching
 static uint8_t *
-branch_stub_hit(branch_t* branch, const uint32_t target_idx, rb_execution_context_t* ec)
+branch_stub_hit(branch_t *branch, const uint32_t target_idx, rb_execution_context_t *ec)
 {
-    uint8_t* dst_addr;
+    uint8_t *dst_addr;
 
     // Stop other ractors since we are going to patch machine code.
     // This is how the GC does it.
@@ -748,7 +748,7 @@ branch_stub_hit(branch_t* branch, const uint32_t target_idx, rb_execution_contex
     RUBY_ASSERT(branch != NULL);
     RUBY_ASSERT(target_idx < 2);
     blockid_t target = branch->targets[target_idx];
-    const ctx_t* target_ctx = &branch->target_ctxs[target_idx];
+    const ctx_t *target_ctx = &branch->target_ctxs[target_idx];
 
     // If this branch has already been patched, return the dst address
     // Note: ractors can cause the same stub to be hit multiple times
@@ -775,7 +775,7 @@ branch_stub_hit(branch_t* branch, const uint32_t target_idx, rb_execution_contex
         ec->cfp->pc = yjit_iseq_pc_at_idx(target.iseq, target.idx);
 
         // Try to find an existing compiled version of this block
-        block_t* p_block = find_block_version(target, target_ctx);
+        block_t *p_block = find_block_version(target, target_ctx);
 
         // If this block hasn't yet been compiled
         if (!p_block) {
@@ -832,16 +832,16 @@ branch_stub_hit(branch_t* branch, const uint32_t target_idx, rb_execution_contex
 }
 
 // Get a version or stub corresponding to a branch target
-uint8_t* get_branch_target(
+uint8_t *get_branch_target(
     blockid_t target,
-    const ctx_t* ctx,
-    branch_t* branch,
+    const ctx_t *ctx,
+    branch_t *branch,
     uint32_t target_idx
 )
 {
     //fprintf(stderr, "get_branch_target, block (%p, %d)\n", target.iseq, target.idx);
 
-    block_t* p_block = find_block_version(target, ctx);
+    block_t *p_block = find_block_version(target, ctx);
 
     // If the block already exists
     if (p_block)
@@ -855,7 +855,7 @@ uint8_t* get_branch_target(
     }
 
     // Generate an outlined stub that will call branch_stub_hit()
-    uint8_t* stub_addr = cb_get_ptr(ocb, ocb->write_pos);
+    uint8_t *stub_addr = cb_get_ptr(ocb, ocb->write_pos);
 
     // Call branch_stub_hit(branch_idx, target_idx, ec)
     mov(ocb, C_ARG_REGS[2], REG_EC);
@@ -871,18 +871,18 @@ uint8_t* get_branch_target(
 }
 
 void gen_branch(
-    jitstate_t* jit,
-    const ctx_t* src_ctx,
+    jitstate_t *jit,
+    const ctx_t *src_ctx,
     blockid_t target0,
-    const ctx_t* ctx0,
+    const ctx_t *ctx0,
     blockid_t target1,
-    const ctx_t* ctx1,
+    const ctx_t *ctx1,
     branchgen_fn gen_fn
 )
 {
     RUBY_ASSERT(target0.iseq != NULL);
 
-    branch_t* branch = make_branch_entry(jit->block, src_ctx, gen_fn);
+    branch_t *branch = make_branch_entry(jit->block, src_ctx, gen_fn);
     branch->targets[0] = target0;
     branch->targets[1] = target1;
     branch->target_ctxs[0] = *ctx0;
@@ -899,7 +899,7 @@ void gen_branch(
 }
 
 void
-gen_jump_branch(codeblock_t* cb, uint8_t* target0, uint8_t* target1, uint8_t shape)
+gen_jump_branch(codeblock_t *cb, uint8_t *target0, uint8_t *target1, uint8_t shape)
 {
     switch (shape)
     {
@@ -917,18 +917,18 @@ gen_jump_branch(codeblock_t* cb, uint8_t* target0, uint8_t* target1, uint8_t sha
 }
 
 void gen_direct_jump(
-    jitstate_t* jit,
-    const ctx_t* ctx,
+    jitstate_t *jit,
+    const ctx_t *ctx,
     blockid_t target0
 )
 {
     RUBY_ASSERT(target0.iseq != NULL);
 
-    branch_t* branch = make_branch_entry(jit->block, ctx, gen_jump_branch);
+    branch_t *branch = make_branch_entry(jit->block, ctx, gen_jump_branch);
     branch->targets[0] = target0;
     branch->target_ctxs[0] = *ctx;
 
-    block_t* p_block = find_block_version(target0, ctx);
+    block_t *p_block = find_block_version(target0, ctx);
 
     // If the version already exists
     if (p_block) {
@@ -955,8 +955,8 @@ void gen_direct_jump(
 
 // Create a stub to force the code up to this point to be executed
 void defer_compilation(
-    jitstate_t* jit,
-    ctx_t* cur_ctx
+    jitstate_t *jit,
+    ctx_t *cur_ctx
 )
 {
     //fprintf(stderr, "defer compilation at (%p, %d) depth=%d\n", block->blockid.iseq, insn_idx, cur_ctx->chain_depth);
@@ -973,7 +973,7 @@ void defer_compilation(
 
     next_ctx.chain_depth += 1;
 
-    branch_t* branch = make_branch_entry(jit->block, cur_ctx, gen_jump_branch);
+    branch_t *branch = make_branch_entry(jit->block, cur_ctx, gen_jump_branch);
 
     // Get the branch targets or stubs
     branch->target_ctxs[0] = next_ctx;
@@ -981,7 +981,7 @@ void defer_compilation(
     branch->dst_addrs[0] = get_branch_target(branch->targets[0], &next_ctx, branch, 0);
 
     // Call the branch generation function
-    codeblock_t* cb = jit->cb;
+    codeblock_t *cb = jit->cb;
     branch->start_pos = cb->write_pos;
     gen_jump_branch(cb, branch->dst_addrs[0], NULL, SHAPE_DEFAULT);
     branch->end_pos = cb->write_pos;
@@ -997,7 +997,7 @@ yjit_free_block(block_t *block)
     // Remove this block from the predecessor's targets
     rb_darray_for(block->incoming, incoming_idx) {
         // Branch from the predecessor to us
-        branch_t* pred_branch = rb_darray_get(block->incoming, incoming_idx);
+        branch_t *pred_branch = rb_darray_get(block->incoming, incoming_idx);
 
         // If this is us, nullify the target block
         for (size_t succ_idx = 0; succ_idx < 2; succ_idx++) {
@@ -1009,18 +1009,18 @@ yjit_free_block(block_t *block)
 
     // For each outgoing branch
     rb_darray_for(block->outgoing, branch_idx) {
-        branch_t* out_branch = rb_darray_get(block->outgoing, branch_idx);
+        branch_t *out_branch = rb_darray_get(block->outgoing, branch_idx);
 
         // For each successor block
         for (size_t succ_idx = 0; succ_idx < 2; succ_idx++) {
-            block_t* succ = out_branch->blocks[succ_idx];
+            block_t *succ = out_branch->blocks[succ_idx];
 
             if (succ == NULL)
                 continue;
 
             // Remove this block from the successor's incoming list
             rb_darray_for(succ->incoming, incoming_idx) {
-                branch_t* pred_branch = rb_darray_get(succ->incoming, incoming_idx);
+                branch_t *pred_branch = rb_darray_get(succ->incoming, incoming_idx);
                 if (pred_branch == out_branch) {
                     rb_darray_remove_unordered(succ->incoming, incoming_idx);
                     break;
@@ -1072,11 +1072,11 @@ invalidate_block_version(block_t *block)
     block_array_remove(versions, block);
 
     // Get a pointer to the generated code for this block
-    uint8_t* code_ptr = cb_get_ptr(cb, block->start_pos);
+    uint8_t *code_ptr = cb_get_ptr(cb, block->start_pos);
 
     // For each incoming branch
     rb_darray_for(block->incoming, incoming_idx) {
-        branch_t* branch = rb_darray_get(block->incoming, incoming_idx);
+        branch_t *branch = rb_darray_get(block->incoming, incoming_idx);
         uint32_t target_idx = (branch->dst_addrs[0] == code_ptr)? 0:1;
         RUBY_ASSERT(branch->dst_addrs[target_idx] == code_ptr);
         RUBY_ASSERT(branch->blocks[target_idx] == block);
diff --git a/yjit_iface.c b/yjit_iface.c
index be4480b9d6..7361a01452 100644
--- a/yjit_iface.c
+++ b/yjit_iface.c
@@ -486,7 +486,7 @@ rb_yjit_compile_iseq(const rb_iseq_t *iseq, rb_execution_context_t *ec)
     // TODO: I think we need to stop all other ractors here
 
     // Compile a block version starting at the first instruction
-    uint8_t* code_ptr = gen_entry_point(iseq, 0, ec);
+    uint8_t *code_ptr = gen_entry_point(iseq, 0, ec);
 
     if (code_ptr)
     {
@@ -541,7 +541,7 @@ block_address(VALUE self)
 {
     block_t * block;
     TypedData_Get_Struct(self, block_t, &yjit_block_type, block);
-    uint8_t* code_addr = cb_get_ptr(cb, block->start_pos);
+    uint8_t *code_addr = cb_get_ptr(cb, block->start_pos);
     return LONG2NUM((intptr_t)code_addr);
 }
 
@@ -883,7 +883,7 @@ rb_yjit_iseq_mark(const struct rb_iseq_constant_body *body)
 
             // Mark outgoing branch entries
             rb_darray_for(block->outgoing, branch_idx) {
-                branch_t* branch = rb_darray_get(block->outgoing, branch_idx);
+                branch_t *branch = rb_darray_get(block->outgoing, branch_idx);
                 for (int i = 0; i < 2; ++i) {
                     rb_gc_mark_movable((VALUE)branch->targets[i].iseq);
                 }
@@ -925,7 +925,7 @@ rb_yjit_iseq_update_references(const struct rb_iseq_constant_body *body)
 
             // Update outgoing branch entries
             rb_darray_for(block->outgoing, branch_idx) {
-                branch_t* branch = rb_darray_get(block->outgoing, branch_idx);
+                branch_t *branch = rb_darray_get(block->outgoing, branch_idx);
                 for (int i = 0; i < 2; ++i) {
                     branch->targets[i].iseq = (const void *)rb_gc_location((VALUE)branch->targets[i].iseq);
                 }
@@ -986,7 +986,7 @@ static const rb_data_type_t yjit_code_page_type = {
 // Allocate a code page and wrap it into a Ruby object owned by the GC
 VALUE rb_yjit_code_page_alloc(void)
 {
-    code_page_t* code_page = alloc_code_page();
+    code_page_t *code_page = alloc_code_page();
     VALUE cp_obj = TypedData_Wrap_Struct(0, &yjit_code_page_type, code_page);
 
     // Write a pointer to the wrapper object at the beginning of the code page
@@ -1004,21 +1004,21 @@ code_page_t *rb_yjit_code_page_unwrap(VALUE cp_obj)
 }
 
 // Get the code page wrapper object for a code pointer
-VALUE rb_yjit_code_page_from_ptr(uint8_t* code_ptr)
+VALUE rb_yjit_code_page_from_ptr(uint8_t *code_ptr)
 {
-    VALUE* page_start = (VALUE*)((intptr_t)code_ptr & ~(CODE_PAGE_SIZE - 1));
+    VALUE *page_start = (VALUE*)((intptr_t)code_ptr & ~(CODE_PAGE_SIZE - 1));
     VALUE wrapper = *page_start;
     return wrapper;
 }
 
 // Get the inline code block corresponding to a code pointer
-void rb_yjit_get_cb(codeblock_t* cb, uint8_t* code_ptr)
+void rb_yjit_get_cb(codeblock_t *cb, uint8_t *code_ptr)
 {
     VALUE page_wrapper = rb_yjit_code_page_from_ptr(code_ptr);
     code_page_t *code_page = rb_yjit_code_page_unwrap(page_wrapper);
 
     // A pointer to the page wrapper object is written at the start of the code page
-    uint8_t* mem_block = code_page->mem_block + sizeof(VALUE);
+    uint8_t *mem_block = code_page->mem_block + sizeof(VALUE);
     uint32_t mem_size = (code_page->page_size/2) - sizeof(VALUE);
     RUBY_ASSERT(mem_block);
 
@@ -1027,13 +1027,13 @@ void rb_yjit_get_cb(codeblock_t* cb, uint8_t* code_ptr)
 }
 
 // Get the outlined code block corresponding to a code pointer
-void rb_yjit_get_ocb(codeblock_t* cb, uint8_t* code_ptr)
+void rb_yjit_get_ocb(codeblock_t *cb, uint8_t *code_ptr)
 {
     VALUE page_wrapper = rb_yjit_code_page_from_ptr(code_ptr);
     code_page_t *code_page = rb_yjit_code_page_unwrap(page_wrapper);
 
     // A pointer to the page wrapper object is written at the start of the code page
-    uint8_t* mem_block = code_page->mem_block + (code_page->page_size/2);
+    uint8_t *mem_block = code_page->mem_block + (code_page->page_size/2);
     uint32_t mem_size = code_page->page_size/2;
     RUBY_ASSERT(mem_block);
 
@@ -1086,10 +1086,10 @@ outgoing_ids(VALUE self)
     VALUE ids = rb_ary_new();
 
     rb_darray_for(block->outgoing, branch_idx) {
-        branch_t* out_branch = rb_darray_get(block->outgoing, branch_idx);
+        branch_t *out_branch = rb_darray_get(block->outgoing, branch_idx);
 
         for (size_t succ_idx = 0; succ_idx < 2; succ_idx++) {
-            block_t* succ = out_branch->blocks[succ_idx];
+            block_t *succ = out_branch->blocks[succ_idx];
 
             if (succ == NULL)
                 continue;
diff --git a/yjit_iface.h b/yjit_iface.h
index 7820cafc99..de6c94da9a 100644
--- a/yjit_iface.h
+++ b/yjit_iface.h
@@ -139,7 +139,7 @@ void yjit_block_assumptions_free(block_t *block);
 
 VALUE rb_yjit_code_page_alloc(void);
 code_page_t *rb_yjit_code_page_unwrap(VALUE cp_obj);
-void rb_yjit_get_cb(codeblock_t* cb, uint8_t* code_ptr);
-void rb_yjit_get_ocb(codeblock_t* cb, uint8_t* code_ptr);
+void rb_yjit_get_cb(codeblock_t *cb, uint8_t *code_ptr);
+void rb_yjit_get_ocb(codeblock_t *cb, uint8_t *code_ptr);
 
 #endif // #ifndef YJIT_IFACE_H
diff --git a/yjit_utils.c b/yjit_utils.c
index 4e764f0aea..57bd027b89 100644
--- a/yjit_utils.c
+++ b/yjit_utils.c
@@ -5,7 +5,7 @@
 #include "yjit_asm.h"
 
 // Save caller-save registers on the stack before a C call
-void push_regs(codeblock_t* cb)
+void push_regs(codeblock_t *cb)
 {
     push(cb, RAX);
     push(cb, RCX);
@@ -20,7 +20,7 @@ void push_regs(codeblock_t* cb)
 }
 
 // Restore caller-save registers from the after a C call
-void pop_regs(codeblock_t* cb)
+void pop_regs(codeblock_t *cb)
 {
     popfq(cb);
     pop(cb, R11);
@@ -39,7 +39,7 @@ static void print_int_cfun(int64_t val)
     fprintf(stderr, "%lld\n", (long long int)val);
 }
 
-void print_int(codeblock_t* cb, x86opnd_t opnd)
+void print_int(codeblock_t *cb, x86opnd_t opnd)
 {
     push_regs(cb);
 
@@ -55,12 +55,12 @@ void print_int(codeblock_t* cb, x86opnd_t opnd)
     pop_regs(cb);
 }
 
-static void print_ptr_cfun(void* val)
+static void print_ptr_cfun(void *val)
 {
     fprintf(stderr, "%p\n", val);
 }
 
-void print_ptr(codeblock_t* cb, x86opnd_t opnd)
+void print_ptr(codeblock_t *cb, x86opnd_t opnd)
 {
     assert (opnd.num_bits == 64);
 
@@ -73,13 +73,13 @@ void print_ptr(codeblock_t* cb, x86opnd_t opnd)
     pop_regs(cb);
 }
 
-static void print_str_cfun(const char* str)
+static void print_str_cfun(const char *str)
 {
     fprintf(stderr, "%s\n", str);
 }
 
 // Print a constant string to stdout
-void print_str(codeblock_t* cb, const char* str)
+void print_str(codeblock_t *cb, const char *str)
 {
     //as.comment("printStr(\"" ~ str ~ "\")");
     size_t len = strlen(str);
diff --git a/yjit_utils.h b/yjit_utils.h
index cbe3e11863..2a8605aab2 100644
--- a/yjit_utils.h
+++ b/yjit_utils.h
@@ -6,10 +6,10 @@
 #include <stdbool.h>
 #include "yjit_asm.h"
 
-void push_regs(codeblock_t* cb);
-void pop_regs(codeblock_t* cb);
-void print_int(codeblock_t* cb, x86opnd_t opnd);
-void print_ptr(codeblock_t* cb, x86opnd_t opnd);
-void print_str(codeblock_t* cb, const char* str);
+void push_regs(codeblock_t *cb);
+void pop_regs(codeblock_t *cb);
+void print_int(codeblock_t *cb, x86opnd_t opnd);
+void print_ptr(codeblock_t *cb, x86opnd_t opnd);
+void print_str(codeblock_t *cb, const char *str);
 
 #endif // #ifndef YJIT_UTILS_H