py/bc: Replace big opcode format table with simple macro.

5 files changed, 33 insertions, 214 deletions

diff --git a/py/bc.c b/py/bc.c
index 5625da8cf..6887dc438 100644
--- a/py/bc.c
+++ b/py/bc.c
@@ -292,106 +292,17 @@ continue2:;
 #if MICROPY_PERSISTENT_CODE_LOAD || MICROPY_PERSISTENT_CODE_SAVE
 
 // The following table encodes the number of bytes that a specific opcode
-// takes up.  There are 4 special opcodes that always have an extra byte:
-//     MP_BC_UNWIND_JUMP
-//     MP_BC_MAKE_CLOSURE
-//     MP_BC_MAKE_CLOSURE_DEFARGS
-//     MP_BC_RAISE_VARARGS
+// takes up.  Some opcodes have an extra byte, defined by MP_BC_MASK_EXTRA_BYTE.
 // There are 4 special opcodes that have an extra byte only when
 // MICROPY_OPT_CACHE_MAP_LOOKUP_IN_BYTECODE is enabled (and they take a qstr):
 //     MP_BC_LOAD_NAME
 //     MP_BC_LOAD_GLOBAL
 //     MP_BC_LOAD_ATTR
 //     MP_BC_STORE_ATTR
-#define OC4(a, b, c, d) (a | (b << 2) | (c << 4) | (d << 6))
-#define U (0) // undefined opcode
-#define B (MP_OPCODE_BYTE) // single byte
-#define Q (MP_OPCODE_QSTR) // single byte plus 2-byte qstr
-#define V (MP_OPCODE_VAR_UINT) // single byte plus variable encoded unsigned int
-#define O (MP_OPCODE_OFFSET) // single byte plus 2-byte bytecode offset
-STATIC const byte opcode_format_table[64] = {
-    OC4(U, U, U, U), // 0x00-0x03
-    OC4(U, U, U, U), // 0x04-0x07
-    OC4(U, U, U, U), // 0x08-0x0b
-    OC4(U, U, U, U), // 0x0c-0x0f
-    OC4(B, B, B, U), // 0x10-0x13
-    OC4(V, U, Q, V), // 0x14-0x17
-    OC4(B, V, V, Q), // 0x18-0x1b
-    OC4(Q, Q, Q, Q), // 0x1c-0x1f
-    OC4(B, B, V, V), // 0x20-0x23
-    OC4(Q, Q, Q, B), // 0x24-0x27
-    OC4(V, V, Q, Q), // 0x28-0x2b
-    OC4(U, U, U, U), // 0x2c-0x2f
-    OC4(B, B, B, B), // 0x30-0x33
-    OC4(B, O, O, O), // 0x34-0x37
-    OC4(O, O, U, U), // 0x38-0x3b
-    OC4(U, O, B, O), // 0x3c-0x3f
-    OC4(O, B, B, O), // 0x40-0x43
-    OC4(O, U, O, B), // 0x44-0x47
-    OC4(U, U, U, U), // 0x48-0x4b
-    OC4(U, U, U, U), // 0x4c-0x4f
-    OC4(V, V, U, V), // 0x50-0x53
-    OC4(B, U, V, V), // 0x54-0x57
-    OC4(V, V, V, B), // 0x58-0x5b
-    OC4(B, B, B, U), // 0x5c-0x5f
-    OC4(V, V, V, V), // 0x60-0x63
-    OC4(V, V, V, V), // 0x64-0x67
-    OC4(Q, Q, B, U), // 0x68-0x6b
-    OC4(U, U, U, U), // 0x6c-0x6f
-
-    OC4(B, B, B, B), // 0x70-0x73
-    OC4(B, B, B, B), // 0x74-0x77
-    OC4(B, B, B, B), // 0x78-0x7b
-    OC4(B, B, B, B), // 0x7c-0x7f
-    OC4(B, B, B, B), // 0x80-0x83
-    OC4(B, B, B, B), // 0x84-0x87
-    OC4(B, B, B, B), // 0x88-0x8b
-    OC4(B, B, B, B), // 0x8c-0x8f
-    OC4(B, B, B, B), // 0x90-0x93
-    OC4(B, B, B, B), // 0x94-0x97
-    OC4(B, B, B, B), // 0x98-0x9b
-    OC4(B, B, B, B), // 0x9c-0x9f
-    OC4(B, B, B, B), // 0xa0-0xa3
-    OC4(B, B, B, B), // 0xa4-0xa7
-    OC4(B, B, B, B), // 0xa8-0xab
-    OC4(B, B, B, B), // 0xac-0xaf
-
-    OC4(B, B, B, B), // 0xb0-0xb3
-    OC4(B, B, B, B), // 0xb4-0xb7
-    OC4(B, B, B, B), // 0xb8-0xbb
-    OC4(B, B, B, B), // 0xbc-0xbf
-
-    OC4(B, B, B, B), // 0xc0-0xc3
-    OC4(B, B, B, B), // 0xc4-0xc7
-    OC4(B, B, B, B), // 0xc8-0xcb
-    OC4(B, B, B, B), // 0xcc-0xcf
-
-    OC4(B, B, B, B), // 0xd0-0xd3
-    OC4(U, U, U, B), // 0xd4-0xd7
-    OC4(B, B, B, B), // 0xd8-0xdb
-    OC4(B, B, B, B), // 0xdc-0xdf
-
-    OC4(B, B, B, B), // 0xe0-0xe3
-    OC4(B, B, B, B), // 0xe4-0xe7
-    OC4(B, B, B, B), // 0xe8-0xeb
-    OC4(B, B, B, B), // 0xec-0xef
-
-    OC4(B, B, B, B), // 0xf0-0xf3
-    OC4(B, B, B, B), // 0xf4-0xf7
-    OC4(U, U, U, U), // 0xf8-0xfb
-    OC4(U, U, U, U), // 0xfc-0xff
-};
-#undef OC4
-#undef U
-#undef B
-#undef Q
-#undef V
-#undef O
-
 uint mp_opcode_format(const byte *ip, size_t *opcode_size, bool count_var_uint) {
-    uint f = (opcode_format_table[*ip >> 2] >> (2 * (*ip & 3))) & 3;
+    uint f = MP_BC_FORMAT(*ip);
     const byte *ip_start = ip;
-    if (f == MP_OPCODE_QSTR) {
+    if (f == MP_BC_FORMAT_QSTR) {
         if (MICROPY_OPT_CACHE_MAP_LOOKUP_IN_BYTECODE_DYNAMIC) {
             if (*ip == MP_BC_LOAD_NAME
                 || *ip == MP_BC_LOAD_GLOBAL
@@ -402,19 +313,14 @@ uint mp_opcode_format(const byte *ip, size_t *opcode_size, bool count_var_uint)
         }
         ip += 3;
     } else {
-        int extra_byte = (
-            *ip == MP_BC_UNWIND_JUMP
-            || *ip == MP_BC_RAISE_VARARGS
-            || *ip == MP_BC_MAKE_CLOSURE
-            || *ip == MP_BC_MAKE_CLOSURE_DEFARGS
-        );
+        int extra_byte = (*ip & MP_BC_MASK_EXTRA_BYTE) == 0;
         ip += 1;
-        if (f == MP_OPCODE_VAR_UINT) {
+        if (f == MP_BC_FORMAT_VAR_UINT) {
             if (count_var_uint) {
                 while ((*ip++ & 0x80) != 0) {
                 }
             }
-        } else if (f == MP_OPCODE_OFFSET) {
+        } else if (f == MP_BC_FORMAT_OFFSET) {
             ip += 2;
         }
         ip += extra_byte;
diff --git a/py/bc.h b/py/bc.h
index ac9105a55..e1b39c03d 100644
--- a/py/bc.h
+++ b/py/bc.h
@@ -128,11 +128,6 @@ const byte *mp_bytecode_print_str(const byte *ip);
 
 #if MICROPY_PERSISTENT_CODE_LOAD || MICROPY_PERSISTENT_CODE_SAVE
 
-#define MP_OPCODE_BYTE (0)
-#define MP_OPCODE_QSTR (1)
-#define MP_OPCODE_VAR_UINT (2)
-#define MP_OPCODE_OFFSET (3)
-
 uint mp_opcode_format(const byte *ip, size_t *opcode_size, bool count_var_uint);
 
 #endif
diff --git a/py/bc0.h b/py/bc0.h
index 857b75db9..96f9f2136 100644
--- a/py/bc0.h
+++ b/py/bc0.h
@@ -31,6 +31,14 @@
 #define MP_BC_MASK_FORMAT                   (0xf0)
 #define MP_BC_MASK_EXTRA_BYTE               (0x9e)
 
+#define MP_BC_FORMAT_BYTE                   (0)
+#define MP_BC_FORMAT_QSTR                   (1)
+#define MP_BC_FORMAT_VAR_UINT               (2)
+#define MP_BC_FORMAT_OFFSET                 (3)
+
+// Nibbles in magic number are: BB BB BB BB BB BO VV QU
+#define MP_BC_FORMAT(op) ((0x000003a4 >> (2 * ((op) >> 4))) & 3)
+
 // Load, Store, Delete, Import, Make, Build, Unpack, Call, Jump, Exception, For, sTack, Return, Yield, Op
 #define MP_BC_BASE_RESERVED                 (0x00) // ----------------
 #define MP_BC_BASE_QSTR_O                   (0x10) // LLLLLLSSSDDII---
diff --git a/py/persistentcode.c b/py/persistentcode.c
index 5e970b553..3b59746ba 100644
--- a/py/persistentcode.c
+++ b/py/persistentcode.c
@@ -32,7 +32,7 @@
 #include "py/reader.h"
 #include "py/emitglue.h"
 #include "py/persistentcode.h"
-#include "py/bc.h"
+#include "py/bc0.h"
 
 #if MICROPY_PERSISTENT_CODE_LOAD || MICROPY_PERSISTENT_CODE_SAVE
 
@@ -307,12 +307,12 @@ STATIC void load_bytecode(mp_reader_t *reader, qstr_window_t *qw, byte *ip, byte
         uint f = mp_opcode_format(ip, &sz, false);
         ++ip;
         --sz;
-        if (f == MP_OPCODE_QSTR) {
+        if (f == MP_BC_FORMAT_QSTR) {
             qstr qst = load_qstr(reader, qw);
             *ip++ = qst;
             *ip++ = qst >> 8;
             sz -= 2;
-        } else if (f == MP_OPCODE_VAR_UINT) {
+        } else if (f == MP_BC_FORMAT_VAR_UINT) {
             while ((*ip++ = read_byte(reader)) & 0x80) {
             }
         }
@@ -600,7 +600,7 @@ STATIC void save_bytecode(mp_print_t *print, qstr_window_t *qw, const byte *ip,
     while (ip < ip_top) {
         size_t sz;
         uint f = mp_opcode_format(ip, &sz, true);
-        if (f == MP_OPCODE_QSTR) {
+        if (f == MP_BC_FORMAT_QSTR) {
             mp_print_bytes(print, ip, 1);
             qstr qst = ip[1] | (ip[2] << 8);
             save_qstr(print, qw, qst);
diff --git a/tools/mpy-tool.py b/tools/mpy-tool.py
index d49789620..557ee98ff 100755
--- a/tools/mpy-tool.py
+++ b/tools/mpy-tool.py
@@ -103,110 +103,25 @@ MP_NATIVE_ARCH_ARMV7EMSP = 7
 MP_NATIVE_ARCH_ARMV7EMDP = 8
 MP_NATIVE_ARCH_XTENSA = 9
 
-MP_OPCODE_BYTE = 0
-MP_OPCODE_QSTR = 1
-MP_OPCODE_VAR_UINT = 2
-MP_OPCODE_OFFSET = 3
-
-# extra bytes:
-MP_BC_UNWIND_JUMP = 0x40
-MP_BC_MAKE_CLOSURE = 0x20
-MP_BC_MAKE_CLOSURE_DEFARGS = 0x21
-MP_BC_RAISE_VARARGS = 0x60
+MP_BC_MASK_EXTRA_BYTE = 0x9e
+
+MP_BC_FORMAT_BYTE = 0
+MP_BC_FORMAT_QSTR = 1
+MP_BC_FORMAT_VAR_UINT = 2
+MP_BC_FORMAT_OFFSET = 3
+
 # extra byte if caching enabled:
 MP_BC_LOAD_NAME = 0x11
 MP_BC_LOAD_GLOBAL = 0x12
 MP_BC_LOAD_ATTR = 0x13
 MP_BC_STORE_ATTR = 0x18
 
-def make_opcode_format():
-    def OC4(a, b, c, d):
-        return a | (b << 2) | (c << 4) | (d << 6)
-    U = 0
-    B = 0
-    Q = 1
-    V = 2
-    O = 3
-    return bytes_cons((
-    # this table is taken verbatim from py/bc.c
-    OC4(U, U, U, U), # 0x00-0x03
-    OC4(U, U, U, U), # 0x04-0x07
-    OC4(U, U, U, U), # 0x08-0x0b
-    OC4(U, U, U, U), # 0x0c-0x0f
-    OC4(B, B, B, U), # 0x10-0x13
-    OC4(V, U, Q, V), # 0x14-0x17
-    OC4(B, V, V, Q), # 0x18-0x1b
-    OC4(Q, Q, Q, Q), # 0x1c-0x1f
-    OC4(B, B, V, V), # 0x20-0x23
-    OC4(Q, Q, Q, B), # 0x24-0x27
-    OC4(V, V, Q, Q), # 0x28-0x2b
-    OC4(U, U, U, U), # 0x2c-0x2f
-    OC4(B, B, B, B), # 0x30-0x33
-    OC4(B, O, O, O), # 0x34-0x37
-    OC4(O, O, U, U), # 0x38-0x3b
-    OC4(U, O, B, O), # 0x3c-0x3f
-    OC4(O, B, B, O), # 0x40-0x43
-    OC4(O, U, O, B), # 0x44-0x47
-    OC4(U, U, U, U), # 0x48-0x4b
-    OC4(U, U, U, U), # 0x4c-0x4f
-    OC4(V, V, U, V), # 0x50-0x53
-    OC4(B, U, V, V), # 0x54-0x57
-    OC4(V, V, V, B), # 0x58-0x5b
-    OC4(B, B, B, U), # 0x5c-0x5f
-    OC4(V, V, V, V), # 0x60-0x63
-    OC4(V, V, V, V), # 0x64-0x67
-    OC4(Q, Q, B, U), # 0x68-0x6b
-    OC4(U, U, U, U), # 0x6c-0x6f
-
-    OC4(B, B, B, B), # 0x70-0x73
-    OC4(B, B, B, B), # 0x74-0x77
-    OC4(B, B, B, B), # 0x78-0x7b
-    OC4(B, B, B, B), # 0x7c-0x7f
-    OC4(B, B, B, B), # 0x80-0x83
-    OC4(B, B, B, B), # 0x84-0x87
-    OC4(B, B, B, B), # 0x88-0x8b
-    OC4(B, B, B, B), # 0x8c-0x8f
-    OC4(B, B, B, B), # 0x90-0x93
-    OC4(B, B, B, B), # 0x94-0x97
-    OC4(B, B, B, B), # 0x98-0x9b
-    OC4(B, B, B, B), # 0x9c-0x9f
-    OC4(B, B, B, B), # 0xa0-0xa3
-    OC4(B, B, B, B), # 0xa4-0xa7
-    OC4(B, B, B, B), # 0xa8-0xab
-    OC4(B, B, B, B), # 0xac-0xaf
-
-    OC4(B, B, B, B), # 0xb0-0xb3
-    OC4(B, B, B, B), # 0xb4-0xb7
-    OC4(B, B, B, B), # 0xb8-0xbb
-    OC4(B, B, B, B), # 0xbc-0xbf
-
-    OC4(B, B, B, B), # 0xc0-0xc3
-    OC4(B, B, B, B), # 0xc4-0xc7
-    OC4(B, B, B, B), # 0xc8-0xcb
-    OC4(B, B, B, B), # 0xcc-0xcf
-
-    OC4(B, B, B, B), # 0xd0-0xd3
-    OC4(U, U, U, B), # 0xd4-0xd7
-    OC4(B, B, B, B), # 0xd8-0xdb
-    OC4(B, B, B, B), # 0xdc-0xdf
-
-    OC4(B, B, B, B), # 0xe0-0xe3
-    OC4(B, B, B, B), # 0xe4-0xe7
-    OC4(B, B, B, B), # 0xe8-0xeb
-    OC4(B, B, B, B), # 0xec-0xef
-
-    OC4(B, B, B, B), # 0xf0-0xf3
-    OC4(B, B, B, B), # 0xf4-0xf7
-    OC4(U, U, U, U), # 0xf8-0xfb
-    OC4(U, U, U, U), # 0xfc-0xff
-    ))
-
 # this function mirrors that in py/bc.c
-def mp_opcode_format(bytecode, ip, count_var_uint, opcode_format=make_opcode_format()):
+def mp_opcode_format(bytecode, ip, count_var_uint):
     opcode = bytecode[ip]
     ip_start = ip
-    f = (opcode_format[opcode >> 2] >> (2 * (opcode & 3))) & 3
-    if f == MP_OPCODE_QSTR:
+    f = ((0x000003a4 >> (2 * ((opcode) >> 4))) & 3)
+    if f == MP_BC_FORMAT_QSTR:
         if config.MICROPY_OPT_CACHE_MAP_LOOKUP_IN_BYTECODE:
             if (opcode == MP_BC_LOAD_NAME
                 or opcode == MP_BC_LOAD_GLOBAL
@@ -215,19 +130,14 @@ def mp_opcode_format(bytecode, ip, count_var_uint, opcode_format=make_opcode_for
                 ip += 1
         ip += 3
     else:
-        extra_byte = (
-            opcode == MP_BC_UNWIND_JUMP
-            or opcode == MP_BC_RAISE_VARARGS
-            or opcode == MP_BC_MAKE_CLOSURE
-            or opcode == MP_BC_MAKE_CLOSURE_DEFARGS
-        )
+        extra_byte = (opcode & MP_BC_MASK_EXTRA_BYTE) == 0
         ip += 1
-        if f == MP_OPCODE_VAR_UINT:
+        if f == MP_BC_FORMAT_VAR_UINT:
             if count_var_uint:
                 while bytecode[ip] & 0x80 != 0:
                     ip += 1
                 ip += 1
-        elif f == MP_OPCODE_OFFSET:
+        elif f == MP_BC_FORMAT_OFFSET:
             ip += 2
         ip += extra_byte
     return f, ip - ip_start
@@ -672,10 +582,10 @@ def read_bytecode(file, bytecode, qstr_win):
         op = read_byte(file, bytecode)
         f, sz = mp_opcode_format(bytecode.buf, bytecode.idx - 1, False)
         sz -= 1
-        if f == MP_OPCODE_QSTR:
+        if f == MP_BC_FORMAT_QSTR:
             read_qstr_and_pack(file, bytecode, qstr_win)
             sz -= 2
-        elif f == MP_OPCODE_VAR_UINT:
+        elif f == MP_BC_FORMAT_VAR_UINT:
             while read_byte(file, bytecode) & 0x80:
                 pass
         for _ in range(sz):