// Copyright (C) 2020-2025 Free Software Foundation, Inc. // This file is part of GCC. // GCC is free software; you can redistribute it and/or modify it under // the terms of the GNU General Public License as published by the Free // Software Foundation; either version 3, or (at your option) any later // version. // GCC is distributed in the hope that it will be useful, but WITHOUT ANY // WARRANTY; without even the implied warranty of MERCHANTABILITY or // FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License // for more details. // You should have received a copy of the GNU General Public License // along with GCC; see the file COPYING3. If not see // . #include "rust-macro-builtins-asm.h" namespace Rust { std::map InlineAsmOptionMap{ {AST::InlineAsmOption::PURE, "pure"}, {AST::InlineAsmOption::NOMEM, "nomem"}, {AST::InlineAsmOption::READONLY, "readonly"}, {AST::InlineAsmOption::PRESERVES_FLAGS, "preserves_flags"}, {AST::InlineAsmOption::NORETURN, "noreturn"}, {AST::InlineAsmOption::NOSTACK, "nostack"}, {AST::InlineAsmOption::MAY_UNWIND, "may_unwind"}, {AST::InlineAsmOption::ATT_SYNTAX, "att_syntax"}, {AST::InlineAsmOption::RAW, "raw"}, }; int parseDirSpec (Parser &parser, TokenId last_token_id) { return 0; } int parse_clobber_abi (Parser &parser, TokenId last_token_id, InlineAsmContext &inlineAsmCtx) { // clobber_abi := "clobber_abi(" *("," ) [","] ")" // PARSE EVERYTHING COMMITTEDLY IN THIS FUNCTION, WE CONFIRMED VIA clobber_abi // identifier keyword auto &inlineAsm = inlineAsmCtx.inlineAsm; auto token = parser.peek_current_token (); if (!parser.skip_token (LEFT_PAREN)) { // TODO: Raise error exactly like rustc if left parenthesis is not // encountered. token = parser.peek_current_token (); // TODO: Error reporting shifted to the left 1 character, I'm not sure // why. if (token->get_id () == last_token_id) { rust_error_at (parser.peek_current_token ()->get_locus (), "expected `(`, found end of macro arguments"); return -1; } else { rust_error_at ( parser.peek_current_token ()->get_locus (), "expected `(`, found `%s`", parser.peek_current_token ()->get_token_description ()); } return -1; } if (parser.skip_token (RIGHT_PAREN)) { // TODO: We encountered a "clobber_abi()", which should be illegal? // https://github.com/rust-lang/rust/blob/c00957a3e269219413041a4e3565f33b1f9d0779/compiler/rustc_builtin_macros/src/asm.rs#L381 rust_error_at ( parser.peek_current_token ()->get_locus (), "at least one abi must be provided as an argument to `clobber_abi`"); return -1; } std::vector new_abis; token = parser.peek_current_token (); while (token->get_id () != last_token_id && token->get_id () != RIGHT_PAREN) { // Check if it is a string literal or not, codename: in ABNF if (token->get_id () == STRING_LITERAL) { // TODO: Caring for span in here. new_abis.push_back ({token->as_string (), token->get_locus ()}); } else { // TODO: We encountered something that is not string literal, which // should be illegal, please emit the correct error // https://github.com/rust-lang/rust/blob/b92758a9aef1cef7b79e2b72c3d8ba113e547f89/compiler/rustc_builtin_macros/src/asm.rs#L387 } if (parser.skip_token (RIGHT_PAREN)) { break; } if (!parser.skip_token (COMMA)) { // TODO: If the skip of comma is unsuccessful, which should be // illegal, pleaes emit the correct error. return -1; } token = parser.peek_current_token (); } // Done processing the local clobber abis, push that to the main Args in // argument for (auto abi : new_abis) { inlineAsm.clobber_abi.push_back (abi); } return 0; } tl::optional parse_reg (Parser &parser, TokenId last_token_id, InlineAsmContext &inlineAsmCtx) { using RegType = AST::InlineAsmRegOrRegClass::Type; if (!parser.skip_token (LEFT_PAREN)) { // TODO: we expect a left parenthesis here, please return the correct // error. return tl::nullopt; } // after successful left parenthesis parsing, we should return ast of // InlineAsmRegOrRegClass of reg or reg class auto token = parser.peek_current_token (); auto tok_id = token->get_id (); AST::InlineAsmRegOrRegClass regClass; if (parser.skip_token (IDENTIFIER)) { // construct a InlineAsmRegOrRegClass regClass.type = RegType::RegClass; regClass.regClass.Symbol = token->as_string (); } else if (tok_id == STRING_LITERAL) { // TODO: there is STRING_LITERAL, and BYTE_STRING_LITERAL, should we check // for both? // construct a InlineAsmRegOrRegClass // parse_format_string regClass.type = RegType::Reg; inlineAsmCtx.is_explicit = true; regClass.regClass.Symbol = token->as_string (); } else { // TODO: This should emit error // return // Err(p.dcx().create_err(errors::ExpectedRegisterClassOrExplicitRegister // { // span: p.token.span, // })); } if (!parser.skip_token (RIGHT_PAREN)) { // TODO: we expect a left parenthesis here, please return the correct // error. return tl::nullopt; } return regClass; } int parse_operand (Parser &parser, TokenId last_token_id, InlineAsmContext &inlineAsmCtx) { return 0; } // From rustc tl::optional parse_reg_operand (Parser &parser, TokenId last_token_id, InlineAsmContext &inlineAsmCtx) { // let name = if p.token.is_ident() && p.look_ahead(1, |t| *t == token::Eq) { // let (ident, _) = p.token.ident().unwrap(); // p.bump(); // p.expect(&token::Eq)?; // allow_templates = false; // Some(ident.name) // } else { // None // }; using RegisterType = AST::InlineAsmOperand::RegisterType; AST::InlineAsmOperand reg_operand; auto token = parser.peek_current_token (); auto iden_token = parser.peek_current_token (); auto &inlineAsm = inlineAsmCtx.inlineAsm; if (check_identifier (parser, "")) { auto equal_token = parser.peek_current_token (); if (!parser.skip_token (EQUAL)) { // TODO: Please handle this. rust_unreachable (); } } token = parser.peek_current_token (); bool is_explicit_reg = false; bool is_global_asm = inlineAsm.is_global_asm; if (!is_global_asm && check_identifier (parser, "in")) { return tl::nullopt; } else if (!is_global_asm && check_identifier (parser, "out")) { return tl::nullopt; } else if (!is_global_asm && check_identifier (parser, "lateout")) { return tl::nullopt; } else if (!is_global_asm && check_identifier (parser, "inout")) { return tl::nullopt; } else if (!is_global_asm && check_identifier (parser, "inlateout")) { return tl::nullopt; } else if (parser.peek_current_token ()->get_id () == CONST) { rust_unreachable (); // todo: Please handle const return tl::nullopt; } else if (false && check_identifier (parser, "sym")) { // todo: Please handle sym return tl::nullopt; } else if (false && check_identifier (parser, "label")) { // todo: Please handle label return tl::nullopt; } else { return tl::nullopt; } return reg_operand; } void check_and_set (Parser &parser, InlineAsmContext &inlineAsmCtx, AST::InlineAsmOption option) { auto &inlineAsm = inlineAsmCtx.inlineAsm; if (inlineAsm.options.count (option) != 0) { // TODO: report an error of duplication rust_error_at (parser.peek_current_token ()->get_locus (), "the `%s` option was already provided", InlineAsmOptionMap[option].c_str ()); return; } else { inlineAsm.options.insert (option); } } int parse_options (Parser &parser, TokenId last_token_id, InlineAsmContext &inlineAsmCtx) { bool is_global_asm = inlineAsmCtx.inlineAsm.is_global_asm; // Parse everything commitedly if (!parser.skip_token (LEFT_PAREN)) { // We have shifted `options` to search for the left parenthesis next, we // should error out if this is not possible. // TODO: report some error. return -1; } auto token = parser.peek_current_token (); while (token->get_id () != last_token_id && token->get_id () != RIGHT_PAREN) { if (!is_global_asm && check_identifier (parser, "pure")) { check_and_set (parser, inlineAsmCtx, AST::InlineAsmOption::PURE); } else if (!is_global_asm && check_identifier (parser, "nomem")) { check_and_set (parser, inlineAsmCtx, AST::InlineAsmOption::NOMEM); } else if (!is_global_asm && check_identifier (parser, "readonly")) { check_and_set (parser, inlineAsmCtx, AST::InlineAsmOption::READONLY); } else if (!is_global_asm && check_identifier (parser, "preserves_flags")) { check_and_set (parser, inlineAsmCtx, AST::InlineAsmOption::PRESERVES_FLAGS); } else if (!is_global_asm && check_identifier (parser, "noreturn")) { check_and_set (parser, inlineAsmCtx, AST::InlineAsmOption::NORETURN); } else if (!is_global_asm && check_identifier (parser, "nostack")) { check_and_set (parser, inlineAsmCtx, AST::InlineAsmOption::NOSTACK); } else if (!is_global_asm && check_identifier (parser, "may_unwind")) { check_and_set (parser, inlineAsmCtx, AST::InlineAsmOption::MAY_UNWIND); } else if (check_identifier (parser, "att_syntax")) { check_and_set (parser, inlineAsmCtx, AST::InlineAsmOption::ATT_SYNTAX); } else if (check_identifier (parser, "raw")) { check_and_set (parser, inlineAsmCtx, AST::InlineAsmOption::RAW); } else { // TODO: Unexpected error, please return the correct error rust_error_at (token->get_locus (), "Unexpected token encountered in parse_options"); } if (parser.skip_token (RIGHT_PAREN)) { break; } // Parse comma as optional if (parser.skip_token (COMMA)) { continue; } else { std::cout << "Sum wrong here, check it out" << std::endl; token = parser.peek_current_token (); return -1; } } // TODO: Per rust asm.rs regarding options_spans // I'm guessing this has to do with some error reporting. // let new_span = span_start.to(p.prev_token.span); // args.options_spans.push(new_span); return 0; } bool check_identifier (Parser &p, std::string ident) { auto token = p.peek_current_token (); if (token->get_id () == IDENTIFIER && (token->as_string () == ident || ident == "")) { p.skip_token (); return true; } else { return false; } } tl::optional parse_format_string (Parser &parser, TokenId last_token_id, InlineAsmContext &inlineAsmCtx) { auto token = parser.peek_current_token (); if (token->get_id () != last_token_id && token->get_id () == STRING_LITERAL) { // very nice, we got a supposedly formatted string. parser.skip_token (); return token->as_string (); } else { return tl::nullopt; } } tl::optional MacroBuiltin::asm_handler (location_t invoc_locus, AST::MacroInvocData &invoc, bool is_global_asm) { return parse_asm (invoc_locus, invoc, is_global_asm); } int parse_asm_arg (Parser &parser, TokenId last_token_id, InlineAsmContext &inlineAsmCtx) { auto token = parser.peek_current_token (); tl::optional fm_string; while (token->get_id () != last_token_id) { token = parser.peek_current_token (); // We accept a comma token here. if (token->get_id () != COMMA && inlineAsmCtx.consumed_comma_without_formatted_string) { // if it is not a comma, but we consumed it previously, this is fine // but we have to set it to false tho. inlineAsmCtx.consumed_comma_without_formatted_string = false; } else if (token->get_id () == COMMA && !inlineAsmCtx.consumed_comma_without_formatted_string) { inlineAsmCtx.consumed_comma_without_formatted_string = false; parser.skip_token (); } else { // TODO: we consumed comma, and there happens to also be a comma // error should be: expected expression, found `,` break; } // And if that token comma is also the trailing comma, we break // TODO: Check with mentor see what last_token_id means token = parser.peek_current_token (); if (token->get_id () == COMMA && token->get_id () == last_token_id) { parser.skip_token (); break; } // Ok after the left paren is good, we better be parsing correctly // everything in here, which is operand in ABNF // TODO: Parse clobber abi, eat the identifier named "clobber_abi" if true if (check_identifier (parser, "clobber_abi")) { parse_clobber_abi (parser, last_token_id, inlineAsmCtx); continue; } // TODO: Parse options if (check_identifier (parser, "options")) { parse_options (parser, last_token_id, inlineAsmCtx); continue; } // Ok after we have check that neither clobber_abi nor options works, the // only other logical choice is reg_operand // std::cout << "reg_operand" << std::endl; fm_string = parse_format_string (parser, last_token_id, inlineAsmCtx); } return 0; } tl::optional parse_asm (location_t invoc_locus, AST::MacroInvocData &invoc, bool is_global_asm) { // From the rule of asm. // We first peek and see if it is a format string or not. // If yes, we process the first ever format string, and move on to the // recurrent of format string Else we exit out // After that, we peek and see if it is a reoccuring stream of format string // or not. If it is, keep on going to do this format string. Else, move on // After that, we peek and see if it is a reoccuring stream of operands or not // If it is, keep on going to do this operand thingy. // Else, move on // We check if there is an optional "," at the end, per ABNF spec. // If it is, consume it. // Done MacroInvocLexer lex (invoc.get_delim_tok_tree ().to_token_stream ()); Parser parser (lex); auto last_token_id = macro_end_token (invoc.get_delim_tok_tree (), parser); AST::InlineAsm inlineAsm (invoc_locus, is_global_asm); auto inlineAsmCtx = InlineAsmContext (inlineAsm); // Parse the first ever formatted string, success or not, will skip 1 token auto fm_string = parse_format_string (parser, last_token_id, inlineAsmCtx); if (fm_string == tl::nullopt) { rust_error_at (parser.peek_current_token ()->get_locus (), "asm template must be a string literal"); return tl::nullopt; } // formatted string stream while (parser.peek_current_token ()->get_id () != last_token_id) { if (!parser.skip_token (COMMA)) { break; } // Ok after the comma is good, we better be parsing correctly everything // in here, which is formatted string in ABNF inlineAsmCtx.consumed_comma_without_formatted_string = false; fm_string = parse_format_string (parser, last_token_id, inlineAsmCtx); if (fm_string == tl::nullopt) { inlineAsmCtx.consumed_comma_without_formatted_string = true; break; } } // operands stream, also handles the optional "," parse_asm_arg (parser, last_token_id, inlineAsmCtx); AST::SingleASTNode single = AST::SingleASTNode (inlineAsmCtx.inlineAsm.clone_expr_without_block ()); std::vector single_vec = {single}; AST::Fragment fragment_ast = AST::Fragment (single_vec, std::vector> ()); return fragment_ast; } } // namespace Rust