Implement two pass encoding

First pass:
 - collect information for numbers, registers and which instructions
   contain label references
 - encode all instructions that don't contain label references
 - Set (temporary) addresses for each instruction

Second pass:
 - Collect information about label references (address, offset, size)
 - encode all instructions that contain label references
 - Update (if necessary) addresses for each instruction

 The second pass is iterated 10 times or until no instructions change
 size, whichever comes first.

src/ast.h

@@ -75,7 +75,7 @@ typedef struct register_ {
 } register_t;
 
 typedef struct opcode_encoding {
-    uint8_t encoding[32];
+    uint8_t buffer[32];
     size_t len;
 } opcode_encoding_t;
 
@@ -89,7 +89,19 @@ struct ast_node {
     union {
         register_t reg;
         number_t number;
-        opcode_encoding_t encoding;
+        struct {
+            bool has_reference;
+            opcode_encoding_t encoding;
+            int64_t address;
+        } instruction;
+        struct {
+            int64_t offset;
+            int64_t address;
+            operand_size_t size;
+        } reference;
+        struct {
+            int64_t address;
+        } label;
     } value;
 };
 

src/encoder/encoder.c

@@ -6,6 +6,31 @@
 #include <errno.h>
 #include <string.h>
 
+/**
+ * General encoder flow:
+ *
+ * There are 2 major passes the encoder does:
+ *
+ * First pass:
+ *   - Run through the AST and collect information:
+ *     - Set register values
+ *     - Parse/set number values
+ *     - Mark all instructions that use label references
+ *   - Encode all instructions that don't use label references
+ *   - Update addresses of all labels and instructions. Use an estimated
+ *     instruction size for those instructions that use label references.
+ *
+ * Second pass:
+ *   - Run through the AST for all instructions that use label references and
+ *     collect size information using the estimated addresses from pass 1
+ *   - Encode label references with the estimated addresses, this fixes their
+ *     size.
+ *   - Update all addresses
+ *
+ * Iteration:
+ *   - Repeat the second pass until addresses converge
+ */
+
 error_t *const err_encoder_invalid_register =
     &(error_t){.message = "Invalid register"};
 error_t *const err_encoder_number_overflow =
@@ -23,13 +48,15 @@ error_t *const err_encoder_not_implemented =
 error_t *const err_encoder_unexpected_length =
     &(error_t){.message = "Unexpectedly long encoding"};
 
-error_t *encoder_alloc(encoder_t **output) {
+error_t *encoder_alloc(encoder_t **output, ast_node_t *ast) {
     *output = nullptr;
     encoder_t *encoder = calloc(1, sizeof(encoder_t));
 
     if (encoder == nullptr)
         return err_allocation_failed;
 
+    encoder->ast = ast;
+
     error_t *err = symbol_table_alloc(&encoder->symbols);
     if (err) {
         free(encoder);
@@ -213,16 +240,15 @@ static inline uint8_t modrm_rm(uint8_t modrm, register_id_t id) {
     return (modrm & ~modrm_rm_mask) | (id & 0b111);
 }
 
-/**
- * Perform the initial pass over the AST. Records all symbols and sets the
- * values of registers and numbers.
- */
-error_t *encoder_first_pass(encoder_t *encoder, ast_node_t *node) {
+error_t *encoder_collect_info(encoder_t *encoder, ast_node_t *node,
+                              ast_node_t *statement) {
     error_t *err = nullptr;
 
-    if (encoder_is_symbols_node(node))
-        err = symbol_table_update(encoder->symbols, node);
-    else if (node->id == NODE_NUMBER)
+    if (encoder_is_symbols_node(node)) {
+        err = symbol_table_update(encoder->symbols, node, statement);
+        if (statement->id == NODE_INSTRUCTION)
+            statement->value.instruction.has_reference = true;
+    } else if (node->id == NODE_NUMBER)
         err = encoder_set_number_value(node);
     else if (node->id == NODE_REGISTER)
         err = encoder_set_register_value(node);
@@ -230,7 +256,8 @@ error_t *encoder_first_pass(encoder_t *encoder, ast_node_t *node) {
         return err;
 
     for (size_t i = 0; i < node->len; ++i) {
-        error_t *err = encoder_first_pass(encoder, node->children[i]);
+        error_t *err =
+            encoder_collect_info(encoder, node->children[i], statement);
         if (err)
             return err;
     }
@@ -456,7 +483,7 @@ error_t *encoder_encode_instruction(encoder_t *encoder,
         return err;
 
     // produce the actual encoding output in the NODE_INSTRUCTION value
-    uint8_t *output = instruction->value.encoding.encoding;
+    uint8_t *output = instruction->value.instruction.encoding.buffer;
     size_t output_len = 0;
 
     // Handle prefixes
@@ -475,24 +502,136 @@ error_t *encoder_encode_instruction(encoder_t *encoder,
     memcpy(output + output_len, encoding->buffer, encoding->len);
     output_len += encoding->len;
 
-    instruction->value.encoding.len = output_len;
+    instruction->value.instruction.encoding.len = output_len;
 
     return nullptr;
 }
 
 /**
- * Perform the second pass that performs actual encoding. Will use
- * placeholder values for label references because instruction size has not
- * yet been determined.
+ * Perform the initial pass over the AST.
+ *
+ * - Collect information about the operands
+ *   - parse and set number values
+ *   - set the register values
+ *   - determine if label references are used by an instruction
+ * - encode instructions that don't use label references
+ * - determine estimated addresses of each statement
+ *
  */
-error_t *encoder_encoding_pass(encoder_t *encoder, ast_node_t *root) {
+constexpr size_t instruction_size_estimate = 10;
+error_t *encoder_first_pass(encoder_t *encoder) {
+    ast_node_t *root = encoder->ast;
+    assert(root->id == NODE_PROGRAM);
+
+    uintptr_t address = 0;
+
     for (size_t i = 0; i < root->len; ++i) {
-        if (root->children[i]->id != NODE_INSTRUCTION)
-            continue;
-        ast_node_t *instruction = root->children[i];
-        error_t *err = encoder_encode_instruction(encoder, instruction);
+        ast_node_t *statement = root->children[i];
+        error_t *err = encoder_collect_info(encoder, statement, statement);
         if (err)
             return err;
+
+        if (statement->id == NODE_INSTRUCTION &&
+            statement->value.instruction.has_reference == false) {
+            err = encoder_encode_instruction(encoder, statement);
+            if (err)
+                return err;
+            statement->value.instruction.address = address;
+            address += statement->value.instruction.encoding.len;
+        } else if (statement->id == NODE_INSTRUCTION) {
+            statement->value.instruction.encoding.len =
+                instruction_size_estimate;
+            statement->value.instruction.address = address;
+            address += instruction_size_estimate;
+        } else if (statement->id == NODE_LABEL) {
+            statement->value.instruction.address = address;
+        }
+    }
+
+    return nullptr;
+}
+
+operand_size_t signed_to_size_mask(int64_t value) {
+    operand_size_t size = OPERAND_SIZE_64;
+
+    if (value >= INT8_MIN && value <= INT8_MAX)
+        size |= OPERAND_SIZE_8;
+
+    if (value >= INT16_MIN && value <= INT16_MAX)
+        size |= OPERAND_SIZE_16;
+
+    if (value >= INT32_MIN && value <= INT32_MAX)
+        size |= OPERAND_SIZE_32;
+
+    return size;
+}
+
+int64_t statement_offset(ast_node_t *from, ast_node_t *to) {
+    assert(from->id == NODE_LABEL || from->id == NODE_INSTRUCTION);
+    assert(to->id == NODE_LABEL || to->id == NODE_INSTRUCTION);
+
+    int64_t from_addr =
+        from->value.instruction.address + from->value.instruction.encoding.len;
+    int64_t to_addr = to->value.instruction.address;
+
+    return to_addr - from_addr;
+}
+
+error_t *encoder_collect_label_info(encoder_t *encoder, ast_node_t *node,
+                                    ast_node_t *statement) {
+    assert(statement->id == NODE_INSTRUCTION);
+    if (node->id == NODE_LABEL_REFERENCE) {
+        const char *name = node->token_entry->token.value;
+        symbol_t *symbol = symbol_table_lookup(encoder->symbols, name);
+        assert(symbol && symbol->statement &&
+               symbol->statement->id == NODE_LABEL);
+
+        int64_t offset = statement_offset(statement, symbol->statement);
+        int64_t absolute = symbol->statement->value.instruction.address;
+        operand_size_t size = signed_to_size_mask(offset);
+
+        node->value.reference.address = absolute;
+        node->value.reference.offset = offset;
+        node->value.reference.size = size;
+    }
+
+    return nullptr;
+}
+
+/**
+ * Perform the second pass. Updates the label info and encodes all instructions
+ * that have a label reference.that performs actual encoding.
+ */
+error_t *encoder_second_pass(encoder_t *encoder, bool *did_update) {
+    ast_node_t *root = encoder->ast;
+
+    *did_update = false;
+    int64_t address = 0;
+    for (size_t i = 0; i < root->len; ++i) {
+        ast_node_t *statement = root->children[i];
+
+        if (statement->id == NODE_INSTRUCTION &&
+            statement->value.instruction.has_reference) {
+            statement->value.instruction.address = address;
+            size_t before = statement->value.instruction.encoding.len;
+            error_t *err =
+                encoder_collect_label_info(encoder, statement, statement);
+            if (err)
+                return err;
+            err = encoder_encode_instruction(encoder, statement);
+            if (err)
+                return err;
+            size_t after = statement->value.instruction.encoding.len;
+            address += after;
+            *did_update = *did_update || (before != after);
+        } else if (statement->id == NODE_INSTRUCTION &&
+                   statement->value.instruction.has_reference) {
+            statement->value.instruction.address = address;
+            address += statement->value.instruction.encoding.len;
+        } else if (statement->id == NODE_LABEL) {
+            statement->value.label.address = address;
+        }
+        address += statement->value.instruction.encoding.len;
     }
     return nullptr;
 }
@@ -515,12 +654,19 @@ error_t *encoder_check_symbols(encoder_t *encoder) {
     return nullptr;
 }
 
-error_t *encoder_encode(encoder_t *encoder, ast_node_t *ast) {
-    error_t *err = encoder_first_pass(encoder, ast);
+error_t *encoder_encode(encoder_t *encoder) {
+    error_t *err = encoder_first_pass(encoder);
     if (err)
         return err;
     err = encoder_check_symbols(encoder);
     if (err)
         return err;
-    return encoder_encoding_pass(encoder, ast);
+
+    bool did_update = true;
+    for (int i = 0; i < 10 && did_update; ++i) {
+        err = encoder_second_pass(encoder, &did_update);
+        if (err)
+            return err;
+    }
+    return nullptr;
 }

src/encoder/encoder.h

@@ -5,6 +5,7 @@
 
 typedef struct encoder {
     symbol_table_t *symbols;
+    ast_node_t *ast;
 } encoder_t;
 
 constexpr uint8_t modrm_mod_memory = 0b00'000'000;
@@ -16,8 +17,8 @@ constexpr uint8_t modrm_reg_mask = 0b00'111'000;
 constexpr uint8_t modrm_rm_mask = 0b00'000'111;
 constexpr uint8_t modrm_mod_mask = 0b11'000'000;
 
-error_t *encoder_alloc(encoder_t **encoder);
-error_t *encoder_encode(encoder_t *encoder, ast_node_t *ast);
+error_t *encoder_alloc(encoder_t **encoder, ast_node_t *ast);
+error_t *encoder_encode(encoder_t *encoder);
 void encoder_free(encoder_t *encoder);
 
 extern error_t *const err_encoder_invalid_register;

src/encoder/symbols.c

@@ -92,7 +92,7 @@ EXPORT       |           |          |   ERR    |          |
 -------------|-----------|----------|----------|----------|
 */
 
-bool symbol_table_should_update(symbol_kind_t old, symbol_kind_t new) {
+bool symbol_table_should_upgrade(symbol_kind_t old, symbol_kind_t new) {
     if (old == SYMBOL_REFERENCE)
         return new != SYMBOL_REFERENCE;
     if (old == SYMBOL_LOCAL)
@@ -112,7 +112,7 @@ bool symbol_table_should_error(symbol_kind_t old, symbol_kind_t new) {
  * @pre The symbol _must not_ already be in the table.
  */
 error_t *symbol_table_add(symbol_table_t *table, char *name, symbol_kind_t kind,
-                          ast_node_t *node) {
+                          ast_node_t *statement) {
     if (table->len >= table->cap) {
         error_t *err = symbol_table_grow_cap(table);
         if (err)
@@ -122,7 +122,7 @@ error_t *symbol_table_add(symbol_table_t *table, char *name, symbol_kind_t kind,
     table->symbols[table->len] = (symbol_t){
         .name = name,
         .kind = kind,
-        .node = node,
+        .statement = statement,
     };
 
     table->len += 1;
@@ -130,23 +130,29 @@ error_t *symbol_table_add(symbol_table_t *table, char *name, symbol_kind_t kind,
     return nullptr;
 }
 
-error_t *symbol_table_update(symbol_table_t *table, ast_node_t *node) {
+error_t *symbol_table_update(symbol_table_t *table, ast_node_t *node,
+                             ast_node_t *statement) {
     char *name;
     symbol_kind_t kind;
     error_t *err = symbol_table_get_node_info(node, &kind, &name);
     if (err)
         return err;
 
+    if (kind != SYMBOL_LOCAL)
+        statement = nullptr;
+
     symbol_t *symbol = symbol_table_lookup(table, name);
     if (!symbol)
-        return symbol_table_add(table, name, kind, node);
+        return symbol_table_add(table, name, kind, statement);
     if (symbol_table_should_error(symbol->kind, kind))
         return err_symbol_table_incompatible_symbols;
-    if (symbol_table_should_update(symbol->kind, kind)) {
-        symbol->name = name;
+    if (symbol_table_should_upgrade(symbol->kind, kind)) {
         symbol->kind = kind;
-        symbol->node = node;
     }
+
+    if (kind == SYMBOL_LOCAL && symbol->statement == nullptr)
+        symbol->statement = statement;
+
     return nullptr;
 }
 

src/encoder/symbols.h

@@ -29,7 +29,7 @@ typedef enum symbol_kind {
 typedef struct symbol {
     char *name;
     symbol_kind_t kind;
-    ast_node_t *node;
+    ast_node_t *statement;
 } symbol_t;
 
 typedef struct symbol_table {
@@ -40,7 +40,8 @@ typedef struct symbol_table {
 
 error_t *symbol_table_alloc(symbol_table_t **table);
 void symbol_table_free(symbol_table_t *table);
-error_t *symbol_table_update(symbol_table_t *table, ast_node_t *node);
+error_t *symbol_table_update(symbol_table_t *table, ast_node_t *node,
+                             ast_node_t *statement);
 symbol_t *symbol_table_lookup(symbol_table_t *table, const char *name);
 
 #endif // INCLUDE_ENCODER_SYMBOLS_H_

src/main.c

@@ -74,11 +74,11 @@ error_t *print_encoding(tokenlist_t *list) {
         return result.err;
 
     encoder_t *encoder;
-    error_t *err = encoder_alloc(&encoder);
+    error_t *err = encoder_alloc(&encoder, result.node);
     if (err)
         goto cleanup_ast;
 
-    err = encoder_encode(encoder, result.node);
+    err = encoder_encode(encoder);
     if (err)
         goto cleanup_ast;
 
@@ -88,7 +88,8 @@ error_t *print_encoding(tokenlist_t *list) {
         if (node->id != NODE_INSTRUCTION)
             continue;
 
-        print_hex(node->value.encoding.len, node->value.encoding.encoding);
+        print_hex(node->value.instruction.encoding.len,
+                  node->value.instruction.encoding.buffer);
     }
 
     encoder_free(encoder);