Gabe Black has uploaded this change for review. (
https://gem5-review.googlesource.com/c/public/gem5/+/56336 )
Change subject: arch,arch-x86: Separate out the block parser for ucode asm.
......................................................................
arch,arch-x86: Separate out the block parser for ucode asm.
This separation makes it possible to enter the block parser recursively,
which will be necessary to allow expanding macros.
Change-Id: I7fd45f8b03d7ac22fcc6d2b8715845d1b10af12d
---
M src/arch/SConscript
M src/arch/micro_asm.py
M src/arch/micro_asm_test.py
A src/arch/ucasmlib/__init__.py
A src/arch/ucasmlib/assembler.py
A src/arch/ucasmlib/block.py
A src/arch/ucasmlib/parser.py
M src/arch/x86/isa/macroop.isa
M src/arch/x86/isa/microasm.isa
M src/arch/x86/isa/rom.isa
10 files changed, 652 insertions(+), 433 deletions(-)
diff --git a/src/arch/SConscript b/src/arch/SConscript
index 3034dac..8a48e17 100644
--- a/src/arch/SConscript
+++ b/src/arch/SConscript
@@ -106,6 +106,7 @@
parser_files = Glob('isa_parser/*.py')
micro_asm_py = File('micro_asm.py')
+ucasmlib_py = Glob('ucasmlib/*.py')
# import ply here because SCons screws with sys.path when performing
actions.
import ply
@@ -209,7 +210,7 @@
source_gen('generic_cpu_exec_%d.cc' % i)
# Actually create the builder.
- sources = [desc, micro_asm_py] + parser_files
+ sources = [desc, ucasmlib_py, micro_asm_py] + parser_files
IsaDescBuilder(target=gen, source=sources, env=env)
return gen
diff --git a/src/arch/micro_asm.py b/src/arch/micro_asm.py
index 6e3a3a0..80ec697 100644
--- a/src/arch/micro_asm.py
+++ b/src/arch/micro_asm.py
@@ -24,431 +24,4 @@
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
-import os
-import sys
-import re
-import traceback
-# get type names
-from types import *
-
-from ply import lex
-from ply import yacc
-
-##########################################################################
-#
-# Base classes for use outside of the assembler
-#
-##########################################################################
-
-class MicroContainer:
- def __init__(self, name):
- self.microops = []
- self.name = name
- self.directives = {}
- self.micro_classes = {}
- self.labels = {}
-
- def add_microop(self, name, microop):
- microop.mnemonic = name
- microop.micropc = len(self.microops)
- self.microops.append(microop)
-
- def add_label(self, label):
- self.labels[label.name] = len(self.microops)
-
- def __str__(self):
- string = "%s:\n" % self.name
- for microop in self.microops:
- string += " %s\n" % microop
- return string
-
-class CombinationalMacroop(MicroContainer):
- pass
-
-class RomMacroop:
- def __init__(self, name, target):
- self.name = name
- self.target = target
-
- def __str__(self):
- return "%s: %s\n" % (self.name, self.target)
-
-class Rom(MicroContainer):
- def __init__(self, name):
- super().__init__(name)
- self.externs = {}
-
- def add_label(self, label):
- super().add_label(label)
- if label.is_extern:
- self.externs[label.name] = len(self.microops)
-
-##########################################################################
-#
-# Support classes
-#
-##########################################################################
-
-class Block:
- def __init__(self):
- self.statements = []
-
-class Statement:
- def __init__(self, name, is_directive=False):
- self.name = name
- self.is_directive = is_directive
-
-class Microop(Statement):
- def __init__(self, name, params):
- super().__init__(name)
- self.params = params
-
- def handle(self, assembler, container):
- microop = assembler.microops.get(self.name, None)
- if not microop:
- raise Exception(f'Unrecongized mnemonic: "{self.name}"')
-
- try:
- microop = eval(f'_cls({self.params})',
- {'_cls': microop}, assembler.symbols)
- except:
- print_error(f'Error instantiating microop "{self.name}"')
- raise
-
- container.add_microop(self.name, microop)
-
-class Directive(Statement):
- def __init__(self, name, params):
- super().__init__(name, True)
- self.params = params
-
- def handle(self, assembler, container):
- directive = container.directives.get(self.name, None)
- if not directive:
- raise Exception(f'Unrecognized directive: "{self.name}"')
- local = {'_dir': directive}
- try:
- eval(f'_dir({self.params})',
- {'_dir': directive}, assembler.symbols)
- except:
- print_error(f'Error executing directive "{self.name}"')
- raise
-
-class Label(Statement):
- def __init__(self, name, is_extern):
- super().__init__(name)
- self.is_extern = is_extern
-
- def handle(self, assembler, container):
- container.add_label(self)
-
-##########################################################################
-#
-# Functions that handle common tasks
-#
-##########################################################################
-
-def print_error(message):
- print()
- print("*** %s" % message)
- print()
-
-class MicroAssembler:
- ######################################################################
- #
- # Lexer specification
- #
- ######################################################################
-
- # Error handler. Just call exit. Output formatted to work under
- # Emacs compile-mode. Optional 'print_traceback' arg, if set to True,
- # prints a Python stack backtrace too (can be handy when trying to
- # debug the parser itself).
- def error(self, lineno, string, print_traceback = False):
- # Print a Python stack backtrace if requested.
- if (print_traceback):
- traceback.print_exc()
- if lineno != 0:
- line_str = "%d:" % lineno
- else:
- line_str = ""
- sys.exit("%s %s" % (line_str, string))
-
- reserved = ('DEF', 'MACROOP', 'ROM', 'EXTERN')
-
- tokens = reserved + (
- # identifier
- 'ID',
- # arguments for microops and directives
- 'PARAMS',
-
- 'LPAREN', 'RPAREN',
- 'LBRACE', 'RBRACE',
- 'COLON', 'SEMI', 'DOT',
- 'NEWLINE'
- )
-
- # New lines are ignored at the top level, but they end statements in
the
- # assembler
- states = (
- ('asm', 'exclusive'),
- ('params', 'exclusive'),
- ('header', 'exclusive'),
- )
-
- reserved_map = { }
- for r in reserved:
- reserved_map[r.lower()] = r
-
- # Ignore comments
- def t_ANY_COMMENT(self, t):
- r'\#[^\n]*(?=\n)'
-
- def t_ANY_MULTILINECOMMENT(self, t):
- r'/\*([^/]|((?<!\*)/))*\*/'
-
- # A colon marks the end of a label. It should follow an ID which will
- # put the lexer in the "params" state. Seeing the colon will put it
back
- # in the "asm" state since it knows it saw a label and not a mnemonic.
- def t_params_COLON(self, t):
- r':'
- t.lexer.pop_state()
- return t
-
- # Parameters are a string of text which don't contain an unescaped
- # statement statement terminator, ie a newline or semi colon.
- def t_params_PARAMS(self, t):
- r'([^\n;\\]|(\\[\n;\\]))+'
- t.lineno += t.value.count('\n')
- unescapeParamsRE = re.compile(r'(\\[\n;\\])')
- def unescapeParams(mo):
- val = mo.group(0)
- return val[1]
- t.value = unescapeParamsRE.sub(unescapeParams, t.value)
- t.lexer.pop_state()
- return t
-
- # An "ID" in the micro assembler is either a label, directive, or
mnemonic
- # If it's either a directive or a mnemonic, it will be optionally
followed
- # by parameters. If it's a label, the following colon will make the
lexer
- # stop looking for parameters.
- def t_asm_ID(self, t):
- r'[A-Za-z_]\w*'
- t.type = self.reserved_map.get(t.value, 'ID')
- # If the ID is really "extern", we shouldn't start looking for
- # parameters yet. The real ID, the label itself, is coming up.
- if t.type != 'EXTERN':
- t.lexer.push_state('params')
- return t
-
- def t_header_ID(self, t):
- r'[A-Za-z_]\w*'
- return t
-
- # If there is a label and you're -not- in the assembler (which would be
- # caught above), don't start looking for parameters.
- def t_ANY_ID(self, t):
- r'[A-Za-z_]\w*'
- t.type = self.reserved_map.get(t.value, 'ID')
- if t.type == 'MACROOP':
- t.lexer.push_state('asm')
- t.lexer.push_state('header')
- elif t.type == 'ROM':
- t.lexer.push_state('asm')
- t.lexer.push_state('header')
- return t
-
- # Braces enter and exit micro assembly
- def t_header_LBRACE(self, t):
- r'\{'
- t.lexer.pop_state()
- return t
-
- def t_asm_RBRACE(self, t):
- r'\}'
- t.lexer.pop_state()
- return t
-
- # In the micro assembler, do line counting but also return a token. The
- # token is needed by the parser to detect the end of a statement.
- def t_asm_NEWLINE(self, t):
- r'\n+'
- t.lineno += t.value.count('\n')
- return t
-
- # A newline or semi colon when looking for params signals that the
- # statement is over and the lexer should go back to looking for regular
- # assembly.
- def t_params_NEWLINE(self, t):
- r'\n+'
- t.lineno += t.value.count('\n')
- t.lexer.pop_state()
- return t
-
- def t_params_SEMI(self, t):
- r';'
- t.lexer.pop_state()
- return t
-
- # Unless handled specially above, track newlines only for line
counting.
- def t_ANY_NEWLINE(self, t):
- r'\n+'
- t.lineno += t.value.count('\n')
-
- # Basic regular expressions to pick out simple tokens
- t_ANY_LPAREN = r'\('
- t_ANY_RPAREN = r'\)'
- t_ANY_SEMI = r';'
- t_ANY_DOT = r'\.'
-
- t_ANY_ignore = ' \t\x0c'
-
- def t_ANY_error(self, t):
- error(t.lineno, "illegal character '%s'" % t.value[0])
- t.skip(1)
-
- ######################################################################
- #
- # Parser specification
- #
- ######################################################################
-
- # A file which may have one or more "object" defined in it.
- def p_file(self, t):
- 'file : opt_objects'
-
- # The objects are optional.
- def p_opt_objects(self, t):
- '''opt_objects : objects
- |'''
-
- # One or more objects.
- def p_objects(self, t):
- '''objects : objects object
- | object'''
-
- # Objects can be of various types.
- def p_object(self, t):
- '''object : rom_block
- | macroop_def'''
-
- # Defines a section of microcode that should go in the current ROM.
- def p_rom_block(self, t):
- 'rom_block : DEF ROM block SEMI'
- if not self.rom:
- print_error("Rom block found, but no Rom object specified.")
- raise TypeError("Rom block found, but nowhere to put it.")
- for statement in t[3].statements:
- statement.handle(self, self.rom)
- t[0] = self.rom
-
- # Defines a macroop that jumps to an external label in the ROM.
- def p_macroop_jump(self, t):
- 'macroop_def : DEF MACROOP ID LPAREN ID RPAREN SEMI'
- if not self.rom_macroop_type:
- print_error("ROM based macroop found, but no ROM macroop " +
- "class was specified.")
- raise TypeError("ROM based macroop found, but no ROM macroop "
+
- "class was specified.")
- macroop = self.rom_macroop_type(t[3], t[5])
- self.macroops[t[3]] = macroop
-
-
- # Defines a macroop that is combinationally generated.
- def p_macroop_def(self, t):
- 'macroop_def : DEF MACROOP ID block SEMI'
- try:
- curop = self.macro_type(t[3])
- except TypeError:
- print_error("Error creating macroop object.")
- raise
- for statement in t[4].statements:
- statement.handle(self, curop)
- self.macroops[t[3]] = curop
-
- # A block of statements
- def p_block(self, t):
- 'block : LBRACE statements RBRACE'
- block = Block()
- block.statements = t[2]
- t[0] = block
-
- # One or more statements
- def p_statements_0(self, t):
- 'statements : statement'
- t[0] = [t[1]] if t[1] else []
-
- def p_statements_1(self, t):
- 'statements : statements statement'
- if t[2]:
- t[1].append(t[2])
- t[0] = t[1]
-
- # A statement can be of various types.
- def p_statement(self, t):
- '''statement : empty_statement
- | label
- | microop
- | directive'''
- t[0] = t[1]
-
- # Parameters are optional.
- def p_opt_params_0(self, t):
- 'opt_params : PARAMS'
- t[0] = t[1]
-
- def p_opt_params_1(self, t):
- 'opt_params :'
- t[0] = ""
-
- # Statements are often ended by newlines or a semi colon.
- def p_end_of_statement(self, t):
- '''end_of_statement : NEWLINE
- | SEMI'''
- pass
-
- # Ignore empty statements.
- def p_empty_statement(self, t):
- 'empty_statement : end_of_statement'
- pass
-
- # A label in the microcode.
- def p_label_0(self, t):
- 'label : EXTERN ID COLON'
- t[0] = Label(t[2], True)
-
- def p_label_1(self, t):
- 'label : ID COLON'
- t[0] = Label(t[1], False)
-
- # A microop with optional parameters.
- def p_microop(self, t):
- 'microop : ID opt_params end_of_statement'
- t[0] = Microop(t[1], t[2])
-
- # Directives for the macroop.
- def p_directive(self, t):
- 'directive : DOT ID opt_params end_of_statement'
- t[0] = Directive(t[2], t[3])
-
- # Parse error handler. Note that the argument here is the offending
- # *token*, not a grammar symbol (hence the need to use t.value)
- def p_error(self, t):
- if t:
- error(t.lineno, "syntax error at '%s'" % t.value)
- else:
- error(0, "unknown syntax error", True)
-
- def __init__(self, macro_type, microops, rom=None,
rom_macroop_type=None):
- self.lexer = lex.lex(object=self)
- self.parser = yacc.yacc(module=self)
- self.macro_type = macro_type
- self.macroops = {}
- self.microops = microops
- self.rom = rom
- self.rom_macroop_type = rom_macroop_type
- self.symbols = {}
-
- def assemble(self, asm):
- self.parser.parse(asm, lexer=self.lexer)
- return self.macroops
+from .ucasmlib.assembler import MicroAssembler
diff --git a/src/arch/micro_asm_test.py b/src/arch/micro_asm_test.py
index 08c2412..45dab1e 100755
--- a/src/arch/micro_asm_test.py
+++ b/src/arch/micro_asm_test.py
@@ -24,7 +24,8 @@
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
-from micro_asm import MicroAssembler, CombinationalMacroop, RomMacroop, Rom
+from ucasmlib import CombinationalMacroop, RomMacroop, Rom
+from ucasmlib.assembler import MicroAssembler
class Bah(object):
def __init__(self):
diff --git a/src/arch/ucasmlib/__init__.py b/src/arch/ucasmlib/__init__.py
new file mode 100644
index 0000000..983f813
--- /dev/null
+++ b/src/arch/ucasmlib/__init__.py
@@ -0,0 +1,74 @@
+# Copyright (c) 2003-2005 The Regents of The University of Michigan
+# All rights reserved.
+#
+# Redistribution and use in source and binary forms, with or without
+# modification, are permitted provided that the following conditions are
+# met: redistributions of source code must retain the above copyright
+# notice, this list of conditions and the following disclaimer;
+# redistributions in binary form must reproduce the above copyright
+# notice, this list of conditions and the following disclaimer in the
+# documentation and/or other materials provided with the distribution;
+# neither the name of the copyright holders nor the names of its
+# contributors may be used to endorse or promote products derived from
+# this software without specific prior written permission.
+#
+# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+##########################################################################
+#
+# Base classes for use outside of the assembler
+#
+##########################################################################
+
+class MicroContainer:
+ def __init__(self, name):
+ self.microops = []
+ self.name = name
+ self.directives = {}
+ self.micro_classes = {}
+ self.labels = {}
+
+ def add_microop(self, name, microop):
+ microop.mnemonic = name
+ microop.micropc = len(self.microops)
+ self.microops.append(microop)
+
+ def add_label(self, label):
+ self.labels[label.name] = len(self.microops)
+
+ def __str__(self):
+ string = "%s:\n" % self.name
+ for microop in self.microops:
+ string += " %s\n" % microop
+ return string
+
+class CombinationalMacroop(MicroContainer):
+ pass
+
+class RomMacroop:
+ def __init__(self, name, target):
+ self.name = name
+ self.target = target
+
+ def __str__(self):
+ return "%s: %s\n" % (self.name, self.target)
+
+class Rom(MicroContainer):
+ def __init__(self, name):
+ super().__init__(name)
+ self.externs = {}
+
+ def add_label(self, label):
+ super().add_label(label)
+ if label.is_extern:
+ self.externs[label.name] = len(self.microops)
diff --git a/src/arch/ucasmlib/assembler.py b/src/arch/ucasmlib/assembler.py
new file mode 100644
index 0000000..c5bf506
--- /dev/null
+++ b/src/arch/ucasmlib/assembler.py
@@ -0,0 +1,220 @@
+# Copyright (c) 2003-2005 The Regents of The University of Michigan
+# All rights reserved.
+#
+# Redistribution and use in source and binary forms, with or without
+# modification, are permitted provided that the following conditions are
+# met: redistributions of source code must retain the above copyright
+# notice, this list of conditions and the following disclaimer;
+# redistributions in binary form must reproduce the above copyright
+# notice, this list of conditions and the following disclaimer in the
+# documentation and/or other materials provided with the distribution;
+# neither the name of the copyright holders nor the names of its
+# contributors may be used to endorse or promote products derived from
+# this software without specific prior written permission.
+#
+# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+from .block import BlockParser
+from .parser import ParserBase
+
+from ply import lex
+from ply import yacc
+
+class MicroAssembler(ParserBase):
+ ######################################################################
+ #
+ # Lexer specification
+ #
+ ######################################################################
+
+ reserved = ('DEF', 'MACROOP', 'ROM')
+
+ tokens = reserved + (
+ # identifier
+ 'ID',
+ # the body of a block
+ 'BODY',
+
+ 'LPAREN', 'RPAREN',
+ 'LBRACE', 'RBRACE',
+ 'SEMI'
+ )
+
+ # New lines are ignored at the top level, but they end statements in
the
+ # assembler
+ states = (
+ ('asm', 'exclusive'),
+ ('header', 'exclusive'),
+ )
+
+ reserved_map = { }
+ for r in reserved:
+ reserved_map[r.lower()] = r
+
+ def t_ANY_ID(self, t):
+ r'[A-Za-z_]\w*'
+ t.type = self.reserved_map.get(t.value, 'ID')
+ if t.type == 'MACROOP':
+ t.lexer.push_state('asm')
+ t.lexer.push_state('header')
+ elif t.type == 'ROM':
+ t.lexer.push_state('asm')
+ t.lexer.push_state('header')
+ return t
+
+ def t_header_ID(self, t):
+ r'[A-Za-z_]\w*'
+ return t
+
+ # Braces enter and exit blocks.
+ def t_header_LBRACE(self, t):
+ r'\{'
+ t.lexer.pop_state()
+ return t
+
+
+ #
+ # Blocks have lines in them which are terminated with unescaped
newlines,
+ # and end when a line starts with a }.
+ #
+
+ #
+ # The regular expression to match a single line of a body.
+ #
+
+ # Leading whitespace.
+ body_line_re = r'\s*'
+ # First non-whitespace, non } character.
+ body_line_re += r'[^\s\}]'
+ # Non-newline or escaped characters.
+ body_line_re += r'([^\n\\]|(\\.))*'
+
+ #
+ # The regular expression to match an entire body.
+ #
+
+ # Leading whitespace for entire body.
+ body_re = r'\s*'
+ # Any number of body lines followed by a new line.
+ body_re += r'(' + body_line_re + r'\n)*'
+ # One body line which is not followed by a new line.
+ body_re += r'(' + body_line_re + r')'
+ # Trailing whitespace for entire body.
+ body_re += r'\s*'
+
+ @lex.TOKEN(body_re)
+ def t_asm_BODY(self, t):
+ t.value = (t.lineno, t.value)
+ t.lineno += t.value[1].count('\n')
+ return t
+
+
+ def t_asm_RBRACE(self, t):
+ r'\}'
+ t.lexer.pop_state()
+ return t
+
+ # Unless handled specially above, track newlines only for line
counting.
+ def t_ANY_NEWLINE(self, t):
+ r'\n+'
+ t.lineno += t.value.count('\n')
+
+ # Basic regular expressions to pick out simple tokens
+ t_ANY_LPAREN = r'\('
+ t_ANY_RPAREN = r'\)'
+ t_ANY_SEMI = r';'
+
+ t_ANY_ignore = ' \t\x0c'
+
+ ######################################################################
+ #
+ # Parser specification
+ #
+ ######################################################################
+
+ # A file which may have one or more "object" defined in it.
+ def p_file(self, t):
+ 'file : opt_objects'
+
+ # The objects are optional.
+ def p_opt_objects(self, t):
+ '''opt_objects : objects
+ |'''
+
+ # One or more objects.
+ def p_objects(self, t):
+ '''objects : objects object
+ | object'''
+
+ # Objects can be of various types.
+ def p_object(self, t):
+ '''object : rom_block
+ | macroop_def'''
+
+ # Defines a section of microcode that should go in the current ROM.
+ def p_rom_block(self, t):
+ 'rom_block : DEF ROM block SEMI'
+ if not self.rom:
+ self.print_error("Rom block found, but no Rom object
specified.")
+ raise TypeError("Rom block found, but nowhere to put it.")
+ for statement in t[3]:
+ statement.handle(self, self.rom)
+ t[0] = self.rom
+
+ # Defines a macroop that jumps to an external label in the ROM.
+ def p_macroop_jump(self, t):
+ 'macroop_def : DEF MACROOP ID LPAREN ID RPAREN SEMI'
+ if not self.rom_macroop_type:
+ self.print_error("ROM based macroop found, but no ROM macroop "
+ "class was specified.")
+ raise TypeError("ROM based macroop found, but no ROM macroop "
+ "class was specified.")
+ macroop = self.rom_macroop_type(t[3], t[5])
+ self.macroops[t[3]] = macroop
+
+
+ # Defines a macroop that is combinationally generated.
+ def p_macroop_def(self, t):
+ 'macroop_def : DEF MACROOP ID block SEMI'
+ try:
+ curop = self.macro_type(t[3])
+ except TypeError:
+ self.print_error("Error creating macroop object.")
+ raise
+ for statement in t[4]:
+ statement.handle(self, curop)
+ self.macroops[t[3]] = curop
+
+ # A block of statements
+ def p_block_0(self, t):
+ 'block : LBRACE BODY RBRACE'
+ starting_line, text = t[2]
+ t[0] = BlockParser().parse(starting_line, text)
+
+ def p_block_1(self, t):
+ 'block : LBRACE RBRACE'
+ t[0] = Block()
+
+ def __init__(self, macro_type, microops, rom=None,
rom_macroop_type=None):
+ self.lexer = lex.lex(object=self)
+ self.parser = yacc.yacc(module=self)
+ self.macro_type = macro_type
+ self.macroops = {}
+ self.microops = microops
+ self.rom = rom
+ self.rom_macroop_type = rom_macroop_type
+ self.symbols = {}
+
+ def assemble(self, asm):
+ self.parser.parse(asm, lexer=self.lexer)
+ return self.macroops
diff --git a/src/arch/ucasmlib/block.py b/src/arch/ucasmlib/block.py
new file mode 100644
index 0000000..c3e8b02
--- /dev/null
+++ b/src/arch/ucasmlib/block.py
@@ -0,0 +1,264 @@
+# Copyright (c) 2003-2005 The Regents of The University of Michigan
+# All rights reserved.
+#
+# Redistribution and use in source and binary forms, with or without
+# modification, are permitted provided that the following conditions are
+# met: redistributions of source code must retain the above copyright
+# notice, this list of conditions and the following disclaimer;
+# redistributions in binary form must reproduce the above copyright
+# notice, this list of conditions and the following disclaimer in the
+# documentation and/or other materials provided with the distribution;
+# neither the name of the copyright holders nor the names of its
+# contributors may be used to endorse or promote products derived from
+# this software without specific prior written permission.
+#
+# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+from .parser import ParserBase
+
+import re
+
+from ply import lex
+from ply import yacc
+
+##########################################################################
+#
+# Support classes
+#
+##########################################################################
+
+class Statement:
+ def __init__(self, name, is_directive=False):
+ self.name = name
+ self.is_directive = is_directive
+
+class Microop(Statement):
+ def __init__(self, name, params):
+ super().__init__(name)
+ self.params = params
+
+ def handle(self, assembler, container):
+ microop = assembler.microops.get(self.name, None)
+ if not microop:
+ raise Exception(f'Unrecognized mnemonic: "{self.name}"')
+
+ try:
+ microop = eval(f'_cls({self.params})',
+ {'_cls': microop}, assembler.symbols)
+ except:
+ assembler.print_error(f'Error instantiating
microop "{self.name}"')
+ raise
+
+ container.add_microop(self.name, microop)
+
+class Directive(Statement):
+ def __init__(self, name, params):
+ super().__init__(name, True)
+ self.params = params
+
+ def handle(self, assembler, container):
+ directive = container.directives.get(self.name, None)
+ if not directive:
+ raise Exception(f'Unrecognized directive: "{self.name}"')
+ local = {'_dir': directive}
+ try:
+ eval(f'_dir({self.params})',
+ {'_dir': directive}, assembler.symbols)
+ except:
+ assembler.print_error(f'Error executing
directive "{self.name}"')
+ raise
+
+class Label(Statement):
+ def __init__(self, name, is_extern):
+ super().__init__(name)
+ self.is_extern = is_extern
+
+ def handle(self, assembler, container):
+ container.add_label(self)
+
+class BlockParser(ParserBase):
+ ######################################################################
+ #
+ # Lexer specification
+ #
+ ######################################################################
+
+ reserved = ('EXTERN',)
+
+ tokens = reserved + (
+ # identifier
+ 'ID',
+ # arguments for microops and directives
+ 'PARAMS',
+
+ 'DOT', 'COLON',
+ 'NEWLINE', 'SEMI',
+ )
+
+ # New lines are ignored at the top level, but they end statements in
the
+ # assembler
+ states = (
+ ('params', 'exclusive'),
+ )
+
+ reserved_map = { }
+ for r in reserved:
+ reserved_map[r.lower()] = r
+
+ # A colon marks the end of a label. It should follow an ID which will
+ # put the lexer in the "params" state. Seeing the colon will put it
back
+ # in the INITIAL state since it knows it saw a label and not a
mnemonic.
+ def t_params_COLON(self, t):
+ r':'
+ t.lexer.pop_state()
+ return t
+
+ # Parameters are a string of text which don't contain an unescaped
+ # statement statement terminator, ie a newline or semi colon.
+ def t_params_PARAMS(self, t):
+ r'([^\n;\\]|(\\.))+'
+ t.lineno += t.value.count('\n')
+ unescapeParamsRE = re.compile(r'(\\[\n;\\])')
+ def unescapeParams(mo):
+ val = mo.group(0)
+ return val[1]
+ t.value = unescapeParamsRE.sub(unescapeParams, t.value)
+ t.lexer.pop_state()
+ return t
+
+ # An "ID" in the micro assembler is either a label, directive, or
mnemonic
+ # If it's either a directive or a mnemonic, it will be optionally
followed
+ # by parameters. If it's a label, the following colon will make the
lexer
+ # stop looking for parameters.
+ def t_INITIAL_ID(self, t):
+ r'[A-Za-z_]\w*'
+ t.type = self.reserved_map.get(t.value, 'ID')
+ # If the ID is really "extern", we shouldn't start looking for
+ # parameters yet. The real ID, the label itself, is coming up.
+ if t.type != 'EXTERN':
+ t.lexer.push_state('params')
+ return t
+
+ # In the micro assembler, do line counting but also return a token. The
+ # token is needed by the parser to detect the end of a statement.
+ def t_INITIAL_NEWLINE(self, t):
+ r'\n+'
+ t.lineno += t.value.count('\n')
+ return t
+
+ # A newline or semi colon when looking for params signals that the
+ # statement is over and the lexer should go back to looking for regular
+ # assembly.
+ def t_params_NEWLINE(self, t):
+ r'\n+'
+ t.lineno += t.value.count('\n')
+ t.lexer.pop_state()
+ return t
+
+ def t_params_SEMI(self, t):
+ r';'
+ t.lexer.pop_state()
+ return t
+
+ # Basic regular expressions to pick out simple tokens
+ t_ANY_SEMI = r';'
+ t_ANY_DOT = r'\.'
+
+ t_ANY_ignore = ' \t\x0c'
+
+ ######################################################################
+ #
+ # Parser specification
+ #
+ ######################################################################
+
+ # A block of statements
+ def p_block(self, t):
+ 'block : opt_statements'
+ t[0] = t[1]
+
+ # Having statements is optional.
+ def p_opt_statements_0(self, t):
+ 'opt_statements : statements'
+ t[0] = t[1]
+
+ def p_opt_statements_1(self, t):
+ 'opt_statements :'
+ t[0] = []
+
+ # One or more statements.
+ def p_statements_0(self, t):
+ 'statements : statement'
+ t[0] = [t[1]] if t[1] else []
+
+ def p_statements_1(self, t):
+ 'statements : statements statement'
+ if t[2]:
+ t[1].append(t[2])
+ t[0] = t[1]
+
+ # A statement can be of various types.
+ def p_statement(self, t):
+ '''statement : empty_statement
+ | label
+ | microop
+ | directive'''
+ t[0] = t[1]
+
+ # Parameters are optional.
+ def p_opt_params_0(self, t):
+ 'opt_params : PARAMS'
+ t[0] = t[1]
+
+ def p_opt_params_1(self, t):
+ 'opt_params :'
+ t[0] = ""
+
+ # Statements are often ended by newlines or a semi colon.
+ def p_end_of_statement(self, t):
+ '''end_of_statement : NEWLINE
+ | SEMI'''
+ pass
+
+ # Ignore empty statements.
+ def p_empty_statement(self, t):
+ 'empty_statement : end_of_statement'
+ pass
+
+ # A label in the microcode.
+ def p_label_0(self, t):
+ 'label : EXTERN ID COLON'
+ t[0] = Label(t[2], True)
+
+ def p_label_1(self, t):
+ 'label : ID COLON'
+ t[0] = Label(t[1], False)
+
+ # A microop with optional parameters.
+ def p_microop(self, t):
+ 'microop : ID opt_params end_of_statement'
+ t[0] = Microop(t[1], t[2])
+
+ # Directives for the macroop.
+ def p_directive(self, t):
+ 'directive : DOT ID opt_params end_of_statement'
+ t[0] = Directive(t[2], t[3])
+
+ def __init__(self):
+ self.lexer = lex.lex(object=self)
+ self.parser = yacc.yacc(module=self)
+ self.statements = []
+
+ def parse(self, starting_lineno, text):
+ self.lexer.lineno = starting_lineno
+ return self.parser.parse(text, lexer=self.lexer)
diff --git a/src/arch/ucasmlib/parser.py b/src/arch/ucasmlib/parser.py
new file mode 100644
index 0000000..43e10d2
--- /dev/null
+++ b/src/arch/ucasmlib/parser.py
@@ -0,0 +1,73 @@
+# Copyright (c) 2003-2005 The Regents of The University of Michigan
+# All rights reserved.
+#
+# Redistribution and use in source and binary forms, with or without
+# modification, are permitted provided that the following conditions are
+# met: redistributions of source code must retain the above copyright
+# notice, this list of conditions and the following disclaimer;
+# redistributions in binary form must reproduce the above copyright
+# notice, this list of conditions and the following disclaimer in the
+# documentation and/or other materials provided with the distribution;
+# neither the name of the copyright holders nor the names of its
+# contributors may be used to endorse or promote products derived from
+# this software without specific prior written permission.
+#
+# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+import sys
+import traceback
+
+class ParserBase:
+ def print_error(message):
+ print()
+ print("*** %s" % message)
+ print()
+
+ ######################################################################
+ #
+ # Lexer specification
+ #
+ ######################################################################
+
+ # Error handler. Just call exit. Output formatted to work under
+ # Emacs compile-mode. Optional 'print_traceback' arg, if set to True,
+ # prints a Python stack backtrace too (can be handy when trying to
+ # debug the parser itself).
+ def error(self, lineno, string, print_traceback = False):
+ # Print a Python stack backtrace if requested.
+ if (print_traceback):
+ traceback.print_exc()
+ if lineno != 0:
+ line_str = "%d:" % lineno
+ else:
+ line_str = ""
+ sys.exit("%s %s" % (line_str, string))
+
+ # Ignore comments
+ def t_ANY_COMMENT(self, t):
+ r'\#[^\n]*(?=\n)'
+
+ def t_ANY_MULTILINECOMMENT(self, t):
+ r'/\*([^/]|((?<!\*)/))*\*/'
+
+ def t_ANY_error(self, t):
+ self.error(t.lineno, "illegal character '%s'" % t.value[0])
+ t.skip(1)
+
+ # Parse error handler. Note that the argument here is the offending
+ # *token*, not a grammar symbol (hence the need to use t.value)
+ def p_error(self, t):
+ if t:
+ self.error(t.lineno, f"syntax error at '{t.value}'")
+ else:
+ self.error(0, "unknown syntax error", True)
diff --git a/src/arch/x86/isa/macroop.isa b/src/arch/x86/isa/macroop.isa
index 07f8a02..6073756 100644
--- a/src/arch/x86/isa/macroop.isa
+++ b/src/arch/x86/isa/macroop.isa
@@ -131,7 +131,7 @@
}};
let {{
- from micro_asm import CombinationalMacroop, RomMacroop
+ from ucasmlib import CombinationalMacroop, RomMacroop
class X86Macroop(CombinationalMacroop):
def setAdjustEnv(self, val):
self.adjust_env = val
diff --git a/src/arch/x86/isa/microasm.isa b/src/arch/x86/isa/microasm.isa
index 7d621b7..88ecfbc 100644
--- a/src/arch/x86/isa/microasm.isa
+++ b/src/arch/x86/isa/microasm.isa
@@ -52,7 +52,8 @@
sys.path[0:0] = ["src/arch/x86/isa/"]
from insts import microcode
# print microcode
- from micro_asm import MicroAssembler, RomMacroop
+ from ucasmlib import RomMacroop
+ from ucasmlib.assembler import MicroAssembler
mainRom = X86MicrocodeRom('main ROM')
assembler = MicroAssembler(X86Macroop, microopClasses, mainRom,
RomMacroop)
diff --git a/src/arch/x86/isa/rom.isa b/src/arch/x86/isa/rom.isa
index d53c9de..289efc8 100644
--- a/src/arch/x86/isa/rom.isa
+++ b/src/arch/x86/isa/rom.isa
@@ -25,7 +25,7 @@
// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
let {{
- from micro_asm import Rom
+ from ucasmlib import Rom
class X86MicrocodeRom(Rom):
def getDeclaration(self):
--
To view, visit https://gem5-review.googlesource.com/c/public/gem5/+/56336
To unsubscribe, or for help writing mail filters, visit
https://gem5-review.googlesource.com/settings
Gerrit-Project: public/gem5
Gerrit-Branch: develop
Gerrit-Change-Id: I7fd45f8b03d7ac22fcc6d2b8715845d1b10af12d
Gerrit-Change-Number: 56336
Gerrit-PatchSet: 1
Gerrit-Owner: Gabe Black <gabe.bl...@gmail.com>
Gerrit-MessageType: newchange
_______________________________________________
gem5-dev mailing list -- gem5-dev@gem5.org
To unsubscribe send an email to gem5-dev-le...@gem5.org
%(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s
