gh-87092: change assembler to use instruction sequence instead of CFG (…

…#103933)

Include/internal/pycore_compile.h

@@ -19,6 +19,7 @@ PyAPI_FUNC(PyCodeObject*) _PyAST_Compile(
     int optimize,
     struct _arena *arena);
 
+static const _PyCompilerSrcLocation NO_LOCATION = {-1, -1, -1, -1};
 
 typedef struct {
     int optimize;
@@ -33,15 +34,21 @@ extern int _PyAST_Optimize(
     struct _arena *arena,
     _PyASTOptimizeState *state);
 
+typedef struct {
+    int h_offset;
+    int h_startdepth;
+    int h_preserve_lasti;
+} _PyCompile_ExceptHandlerInfo;
 
 typedef struct {
     int i_opcode;
     int i_oparg;
     _PyCompilerSrcLocation i_loc;
-} _PyCompilerInstruction;
+    _PyCompile_ExceptHandlerInfo i_except_handler_info;
+} _PyCompile_Instruction;
 
 typedef struct {
-    _PyCompilerInstruction *s_instrs;
+    _PyCompile_Instruction *s_instrs;
     int s_allocated;
     int s_used;
 
@@ -82,6 +89,8 @@ int _PyCompile_EnsureArrayLargeEnough(
 
 int _PyCompile_ConstCacheMergeOne(PyObject *const_cache, PyObject **obj);
 
+int _PyCompile_InstrSize(int opcode, int oparg);
+
 /* Access compiler internals for unit testing */
 
 PyAPI_FUNC(PyObject*) _PyCompile_CodeGen(

Include/internal/pycore_flowgraph.h

@@ -11,7 +11,6 @@ extern "C" {
 #include "pycore_opcode_utils.h"
 #include "pycore_compile.h"
 
-static const _PyCompilerSrcLocation NO_LOCATION = {-1, -1, -1, -1};
 
 typedef struct {
     int i_opcode;
@@ -97,7 +96,6 @@ int _PyCfg_OptimizeCodeUnit(_PyCfgBuilder *g, PyObject *consts, PyObject *const_
 int _PyCfg_Stackdepth(_PyCfgBasicblock *entryblock, int code_flags);
 void _PyCfg_ConvertExceptionHandlersToNops(_PyCfgBasicblock *entryblock);
 int _PyCfg_ResolveJumps(_PyCfgBuilder *g);
-int _PyCfg_InstrSize(_PyCfgInstruction *instruction);
 
 
 static inline int
@@ -113,7 +111,7 @@ basicblock_nofallthrough(const _PyCfgBasicblock *b) {
 
 PyCodeObject *
 _PyAssemble_MakeCodeObject(_PyCompile_CodeUnitMetadata *u, PyObject *const_cache,
-                           PyObject *consts, int maxdepth, _PyCfgBasicblock *entryblock,
+                           PyObject *consts, int maxdepth, _PyCompile_InstructionSequence *instrs,
                            int nlocalsplus, int code_flags, PyObject *filename);
 
 #ifdef __cplusplus

Python/assemble.c

@@ -1,10 +1,10 @@
 #include <stdbool.h>
 
 #include "Python.h"
-#include "pycore_flowgraph.h"
+#include "pycore_code.h"          // write_location_entry_start()
 #include "pycore_compile.h"
+#include "pycore_opcode.h"        // _PyOpcode_Caches[] and opcode category macros
 #include "pycore_pymem.h"         // _PyMem_IsPtrFreed()
-#include "pycore_code.h"          // write_location_entry_start()
 
 
 #define DEFAULT_CODE_SIZE 128
@@ -22,8 +22,8 @@
     }
 
 typedef _PyCompilerSrcLocation location;
-typedef _PyCfgInstruction cfg_instr;
-typedef _PyCfgBasicblock basicblock;
+typedef _PyCompile_Instruction instruction;
+typedef _PyCompile_InstructionSequence instr_sequence;
 
 static inline bool
 same_location(location a, location b)
@@ -117,21 +117,22 @@ assemble_emit_exception_table_item(struct assembler *a, int value, int msb)
 #define MAX_SIZE_OF_ENTRY 20
 
 static int
-assemble_emit_exception_table_entry(struct assembler *a, int start, int end, basicblock *handler)
+assemble_emit_exception_table_entry(struct assembler *a, int start, int end,
+                                    _PyCompile_ExceptHandlerInfo *handler)
 {
     Py_ssize_t len = PyBytes_GET_SIZE(a->a_except_table);
     if (a->a_except_table_off + MAX_SIZE_OF_ENTRY >= len) {
         RETURN_IF_ERROR(_PyBytes_Resize(&a->a_except_table, len * 2));
     }
     int size = end-start;
     assert(end > start);
-    int target = handler->b_offset;
-    int depth = handler->b_startdepth - 1;
-    if (handler->b_preserve_lasti) {
+    int target = handler->h_offset;
+    int depth = handler->h_startdepth - 1;
+    if (handler->h_preserve_lasti) {
         depth -= 1;
     }
     assert(depth >= 0);
-    int depth_lasti = (depth<<1) | handler->b_preserve_lasti;
+    int depth_lasti = (depth<<1) | handler->h_preserve_lasti;
     assemble_emit_exception_table_item(a, start, (1<<7));
     assemble_emit_exception_table_item(a, size, 0);
     assemble_emit_exception_table_item(a, target, 0);
@@ -140,29 +141,26 @@ assemble_emit_exception_table_entry(struct assembler *a, int start, int end, bas
 }
 
 static int
-assemble_exception_table(struct assembler *a, basicblock *entryblock)
+assemble_exception_table(struct assembler *a, instr_sequence *instrs)
 {
-    basicblock *b;
     int ioffset = 0;
-    basicblock *handler = NULL;
+    _PyCompile_ExceptHandlerInfo handler;
+    handler.h_offset = -1;
     int start = -1;
-    for (b = entryblock; b != NULL; b = b->b_next) {
-        ioffset = b->b_offset;
-        for (int i = 0; i < b->b_iused; i++) {
-            cfg_instr *instr = &b->b_instr[i];
-            if (instr->i_except != handler) {
-                if (handler != NULL) {
-                    RETURN_IF_ERROR(
-                        assemble_emit_exception_table_entry(a, start, ioffset, handler));
-                }
-                start = ioffset;
-                handler = instr->i_except;
+    for (int i = 0; i < instrs->s_used; i++) {
+        instruction *instr = &instrs->s_instrs[i];
+        if (instr->i_except_handler_info.h_offset != handler.h_offset) {
+            if (handler.h_offset >= 0) {
+                RETURN_IF_ERROR(
+                    assemble_emit_exception_table_entry(a, start, ioffset, &handler));
             }
-            ioffset += _PyCfg_InstrSize(instr);
+            start = ioffset;
+            handler = instr->i_except_handler_info;
         }
+        ioffset += _PyCompile_InstrSize(instr->i_opcode, instr->i_oparg);
     }
-    if (handler != NULL) {
-        RETURN_IF_ERROR(assemble_emit_exception_table_entry(a, start, ioffset, handler));
+    if (handler.h_offset >= 0) {
+        RETURN_IF_ERROR(assemble_emit_exception_table_entry(a, start, ioffset, &handler));
     }
     return SUCCESS;
 }
@@ -316,31 +314,31 @@ assemble_emit_location(struct assembler* a, location loc, int isize)
 }
 
 static int
-assemble_location_info(struct assembler *a, basicblock *entryblock, int firstlineno)
+assemble_location_info(struct assembler *a, instr_sequence *instrs,
+                       int firstlineno)
 {
     a->a_lineno = firstlineno;
     location loc = NO_LOCATION;
     int size = 0;
-    for (basicblock *b = entryblock; b != NULL; b = b->b_next) {
-        for (int j = 0; j < b->b_iused; j++) {
-            if (!same_location(loc, b->b_instr[j].i_loc)) {
+    for (int i = 0; i < instrs->s_used; i++) {
+        instruction *instr = &instrs->s_instrs[i];
+        if (!same_location(loc, instr->i_loc)) {
                 RETURN_IF_ERROR(assemble_emit_location(a, loc, size));
-                loc = b->b_instr[j].i_loc;
+                loc = instr->i_loc;
                 size = 0;
-            }
-            size += _PyCfg_InstrSize(&b->b_instr[j]);
         }
+        size += _PyCompile_InstrSize(instr->i_opcode, instr->i_oparg);
     }
     RETURN_IF_ERROR(assemble_emit_location(a, loc, size));
     return SUCCESS;
 }
 
 static void
-write_instr(_Py_CODEUNIT *codestr, cfg_instr *instruction, int ilen)
+write_instr(_Py_CODEUNIT *codestr, instruction *instr, int ilen)
 {
-    int opcode = instruction->i_opcode;
+    int opcode = instr->i_opcode;
     assert(!IS_PSEUDO_OPCODE(opcode));
-    int oparg = instruction->i_oparg;
+    int oparg = instr->i_oparg;
     assert(HAS_ARG(opcode) || oparg == 0);
     int caches = _PyOpcode_Caches[opcode];
     switch (ilen - caches) {
@@ -380,12 +378,12 @@ write_instr(_Py_CODEUNIT *codestr, cfg_instr *instruction, int ilen)
 */
 
 static int
-assemble_emit_instr(struct assembler *a, cfg_instr *i)
+assemble_emit_instr(struct assembler *a, instruction *instr)
 {
     Py_ssize_t len = PyBytes_GET_SIZE(a->a_bytecode);
     _Py_CODEUNIT *code;
 
-    int size = _PyCfg_InstrSize(i);
+    int size = _PyCompile_InstrSize(instr->i_opcode, instr->i_oparg);
     if (a->a_offset + size >= len / (int)sizeof(_Py_CODEUNIT)) {
         if (len > PY_SSIZE_T_MAX / 2) {
             return ERROR;
@@ -394,25 +392,24 @@ assemble_emit_instr(struct assembler *a, cfg_instr *i)
     }
     code = (_Py_CODEUNIT *)PyBytes_AS_STRING(a->a_bytecode) + a->a_offset;
     a->a_offset += size;
-    write_instr(code, i, size);
+    write_instr(code, instr, size);
     return SUCCESS;
 }
 
 static int
-assemble_emit(struct assembler *a, basicblock *entryblock, int first_lineno,
-              PyObject *const_cache)
+assemble_emit(struct assembler *a, instr_sequence *instrs,
+              int first_lineno, PyObject *const_cache)
 {
     RETURN_IF_ERROR(assemble_init(a, first_lineno));
 
-    for (basicblock *b = entryblock; b != NULL; b = b->b_next) {
-        for (int j = 0; j < b->b_iused; j++) {
-            RETURN_IF_ERROR(assemble_emit_instr(a, &b->b_instr[j]));
-        }
+    for (int i = 0; i < instrs->s_used; i++) {
+        instruction *instr = &instrs->s_instrs[i];
+        RETURN_IF_ERROR(assemble_emit_instr(a, instr));
     }
 
-    RETURN_IF_ERROR(assemble_location_info(a, entryblock, a->a_lineno));
+    RETURN_IF_ERROR(assemble_location_info(a, instrs, a->a_lineno));
 
-    RETURN_IF_ERROR(assemble_exception_table(a, entryblock));
+    RETURN_IF_ERROR(assemble_exception_table(a, instrs));
 
     RETURN_IF_ERROR(_PyBytes_Resize(&a->a_except_table, a->a_except_table_off));
     RETURN_IF_ERROR(_PyCompile_ConstCacheMergeOne(const_cache, &a->a_except_table));
@@ -586,13 +583,13 @@ makecode(_PyCompile_CodeUnitMetadata *umd, struct assembler *a, PyObject *const_
 
 PyCodeObject *
 _PyAssemble_MakeCodeObject(_PyCompile_CodeUnitMetadata *umd, PyObject *const_cache,
-                           PyObject *consts, int maxdepth, basicblock *entryblock,
+                           PyObject *consts, int maxdepth, instr_sequence *instrs,
                            int nlocalsplus, int code_flags, PyObject *filename)
 {
     PyCodeObject *co = NULL;
 
     struct assembler a;
-    int res = assemble_emit(&a, entryblock, umd->u_firstlineno, const_cache);
+    int res = assemble_emit(&a, instrs, umd->u_firstlineno, const_cache);
     if (res == SUCCESS) {
         co = makecode(umd, &a, const_cache, consts, maxdepth, nlocalsplus,
                       code_flags, filename);

Python/compile.c

@@ -149,7 +149,18 @@ enum {
     COMPILER_SCOPE_COMPREHENSION,
 };
 
-typedef _PyCompilerInstruction instruction;
+
+int
+_PyCompile_InstrSize(int opcode, int oparg)
+{
+    assert(!IS_PSEUDO_OPCODE(opcode));
+    assert(HAS_ARG(opcode) || oparg == 0);
+    int extended_args = (0xFFFFFF < oparg) + (0xFFFF < oparg) + (0xFF < oparg);
+    int caches = _PyOpcode_Caches[opcode];
+    return extended_args + 1 + caches;
+}
+
+typedef _PyCompile_Instruction instruction;
 typedef _PyCompile_InstructionSequence instr_sequence;
 
 #define INITIAL_INSTR_SEQUENCE_SIZE 100
@@ -6968,10 +6979,6 @@ optimize_and_assemble_code_unit(struct compiler_unit *u, PyObject *const_cache,
         goto error;
     }
 
-    if (cfg_to_instr_sequence(&g, &optimized_instrs) < 0) {
-        goto error;
-    }
-
     /** Assembly **/
     int nlocalsplus = prepare_localsplus(u, &g, code_flags);
     if (nlocalsplus < 0) {
@@ -6990,15 +6997,15 @@ optimize_and_assemble_code_unit(struct compiler_unit *u, PyObject *const_cache,
     if (_PyCfg_ResolveJumps(&g) < 0) {
         goto error;
     }
+
+    /* Can't modify the bytecode after computing jump offsets. */
+
     if (cfg_to_instr_sequence(&g, &optimized_instrs) < 0) {
         goto error;
     }
 
-
-    /* Can't modify the bytecode after computing jump offsets. */
-
     co = _PyAssemble_MakeCodeObject(&u->u_metadata, const_cache, consts,
-                                    maxdepth, g.g_entryblock, nlocalsplus,
+                                    maxdepth, &optimized_instrs, nlocalsplus,
                                     code_flags, filename);
 
 error:
@@ -7039,11 +7046,18 @@ cfg_to_instr_sequence(cfg_builder *g, instr_sequence *seq)
         RETURN_IF_ERROR(instr_sequence_use_label(seq, b->b_label.id));
         for (int i = 0; i < b->b_iused; i++) {
             cfg_instr *instr = &b->b_instr[i];
-            int arg = HAS_TARGET(instr->i_opcode) ?
-                      instr->i_target->b_label.id :
-                      instr->i_oparg;
             RETURN_IF_ERROR(
-                instr_sequence_addop(seq, instr->i_opcode, arg, instr->i_loc));
+                instr_sequence_addop(seq, instr->i_opcode, instr->i_oparg, instr->i_loc));
+
+            _PyCompile_ExceptHandlerInfo *hi = &seq->s_instrs[seq->s_used-1].i_except_handler_info;
+            if (instr->i_except != NULL) {
+                hi->h_offset = instr->i_except->b_offset;
+                hi->h_startdepth = instr->i_except->b_startdepth;
+                hi->h_preserve_lasti = instr->i_except->b_preserve_lasti;
+            }
+            else {
+                hi->h_offset = -1;
+            }
         }
     }
     return SUCCESS;

Python/flowgraph.c

@@ -166,24 +166,18 @@ _PyBasicblock_InsertInstruction(basicblock *block, int pos, cfg_instr *instr) {
     return SUCCESS;
 }
 
-int
-_PyCfg_InstrSize(cfg_instr *instruction)
+static int
+instr_size(cfg_instr *instruction)
 {
-    int opcode = instruction->i_opcode;
-    assert(!IS_PSEUDO_OPCODE(opcode));
-    int oparg = instruction->i_oparg;
-    assert(HAS_ARG(opcode) || oparg == 0);
-    int extended_args = (0xFFFFFF < oparg) + (0xFFFF < oparg) + (0xFF < oparg);
-    int caches = _PyOpcode_Caches[opcode];
-    return extended_args + 1 + caches;
+    return _PyCompile_InstrSize(instruction->i_opcode, instruction->i_oparg);
 }
 
 static int
 blocksize(basicblock *b)
 {
     int size = 0;
     for (int i = 0; i < b->b_iused; i++) {
-        size += _PyCfg_InstrSize(&b->b_instr[i]);
+        size += instr_size(&b->b_instr[i]);
     }
     return size;
 }
@@ -492,7 +486,7 @@ resolve_jump_offsets(basicblock *entryblock)
             bsize = b->b_offset;
             for (int i = 0; i < b->b_iused; i++) {
                 cfg_instr *instr = &b->b_instr[i];
-                int isize = _PyCfg_InstrSize(instr);
+                int isize = instr_size(instr);
                 /* jump offsets are computed relative to
                  * the instruction pointer after fetching
                  * the jump instruction.
@@ -508,7 +502,7 @@ resolve_jump_offsets(basicblock *entryblock)
                         assert(!IS_BACKWARDS_JUMP_OPCODE(instr->i_opcode));
                         instr->i_oparg -= bsize;
                     }
-                    if (_PyCfg_InstrSize(instr) != isize) {
+                    if (instr_size(instr) != isize) {
                         extended_arg_recompile = 1;
                     }
                 }
@@ -520,7 +514,7 @@ resolve_jump_offsets(basicblock *entryblock)
         with a better solution.
 
         The issue is that in the first loop blocksize() is called
-        which calls _PyCfg_InstrSize() which requires i_oparg be set
+        which calls instr_size() which requires i_oparg be set
         appropriately. There is a bootstrap problem because
         i_oparg is calculated in the second loop above.