cleanup a little

.gitignore

@@ -0,0 +1 @@
+*:Zone.Identifier

Makefile

@@ -232,8 +232,8 @@ $$(dll_$1_$2_$3) $$(implib_$1_$2_$3):	$$(obj_$1_$2_$3) $$(abslib_$1_$2_$3) \
 	$(ccache) $$(toolchain_$1)gcc -shared $$(srcdir_$3)/$3.def \
 		-o $$(dll_$1_$2_$3) -Wl,--out-implib,$$(implib_$1_$2_$3) \
 		$$^ $$(ldflags_$3) $(optflags_$1)
-	strip -s $$(dll_$1_$2_$3)
-	ranlib $$(implib_$1_$2_$3)
+	$$(toolchain_$1)strip -s $$(dll_$1_$2_$3)
+	$$(toolchain_$1)ranlib $$(implib_$1_$2_$3)
 
 endef
 
@@ -248,7 +248,7 @@ exe_$1_$2_$3	:= $$(bindir_$1_$2)/$3.exe
 $$(exe_$1_$2_$3): $$(obj_$1_$2_$3) $$(abslib_$1_$2_$3) $$(absdpl_$1_$2_$3) \
 		| $$(bindir_$1_$2)
 	$(ccache) $$(toolchain_$1)gcc -o $$@ $$^ $$(ldflags_$3) $(optflags_$1)
-	strip -s $$@
+	$$(toolchain_$1)strip -s $$@
 
 endef
 
@@ -259,7 +259,7 @@ define t_import
 impdef_$1_$2_$3	?= imports/import_$1_$2_$3.def
 
 $$(bindir_$1_$2)/lib$3.a: $$(impdef_$1_$2_$3) | $$(bindir_$1_$2)
-	dlltool -l $$@ -d $$<
+	$$(toolchain_$1)dlltool -l $$@ -d $$<
 
 endef
 

Module.mk

@@ -6,6 +6,7 @@ avsvers_64      := 1700 1509
 imps            += avs avs-ea3
 
 include util/Module.mk
+include libdisasm/Module.mk
 include minhook/Module.mk
 include popnhax/Module.mk
 
@@ -14,17 +15,15 @@ include popnhax/Module.mk
 #
 
 zipdir          := $(BUILDDIR)/zip
+popnhax_version := $(shell grep "define PROGRAM_VERSION" popnhax/dllmain.cc | cut -d'"' -f2)
 
-$(zipdir)/:
-	mkdir -p $@
+$(BUILDDIR)/popnhax_v$(popnhax_version).zip: \
+	build/bin/avs2_1508-32/popnhax.dll
+	@echo ... $@
+	@mkdir -p $(zipdir)
+	@cp -a -p build/bin/avs2_1508-32/popnhax.dll $(zipdir)
+	@cp -r -a -p dist/popnhax/* $(zipdir)
+	@cd $(zipdir) \
+	&& zip -r ../popnhax_v$(popnhax_version).zip ./*
 
-$(BUILDDIR)/popnhax.zip: \
-		build/bin/avs2_1508-32/popnhax.dll \
-		dist/popnhax/popnhax.xml \
-		dist/popnhax/D3d9.dll \
-		dist/popnhax/ifs_hook.dll \
-		| $(zipdir)/
-	echo ... $@
-	zip -j $@ $^
-
-all: $(BUILDDIR)/popnhax.zip
+all: $(BUILDDIR)/popnhax_v$(popnhax_version).zip

README.md

@@ -2,10 +2,32 @@
 
 # popnhax
 
-Arcade game patcher.
+Patcher for pop'n music arcade game.
+
+Featuring pfree, instant retire, visual and audio offset adjust, 1000Hz input polling for true ms-based timing, unlimited favorites, auto hi-speed selection, iidx-like hard gauge and more..
 
 Based on [bemanihax](https://github.com/windyfairy/bemanihax) whose an updated version was included with omnimix v1
 
+### Features
+
+Refer to [popnhax.xml](https://github.com/CrazyRedMachine/popnhax/blob/main/dist/popnhax/popnhax.xml) for complete list and explanations
+
+More info could be found in [Release Notes](https://github.com/CrazyRedMachine/popnhax/releases) or in the [popnhax_tools](https://github.com/CrazyRedMachine/popnhax_tools) repository.
+
+### Run Instructions
+
+- Extract all files directly in the `contents` folder of your install.
+
+(**Note**: if you're running your dlls from `modules` subfolder, please rather copy them back into `contents` folder).
+
+- Edit `popnhax.xml` with a text editor and set your desired options.
+
+- Add `popnhax.dll` as an inject dll to your gamestart command or option menu.
+
+eg. modify your gamestart.bat to add `-k popnhax.dll` or `-K popnhax.dll` depending on the launcher you use.
+
+Some launchers also feature an option menu (accessible by pressing F4 ingame), in which case you can locate the "Inject Hook" setting in option tab. Enter `popnhax.dll` there.
+
 ### Build Instructions
 
-Should be working out of the box with MSYS2/MinGW32. Just run `make`.
+Using WSL is the recommended method. Just run `make`.

dist/popnhax/popnhax.xml

@@ -5,6 +5,22 @@
     <patch_db __type="bool">0</patch_db>
     <!-- Force unlock music, charts, characters, and deco parts when applicable -->
     <force_unlocks __type="bool">0</force_unlocks>
+    <!-- Put customs into their own category (0: no, 1: simple category with all customs, 2: subcategories by folder name) -->
+    <custom_categ __type="u8">2</custom_categ>
+    <!-- Prevent customs from showing up in version folders -->
+    <custom_exclude_from_version __type="bool">1</custom_exclude_from_version>
+    <!-- Prevent customs from showing up in level folders -->
+    <custom_exclude_from_level __type="bool">0</custom_exclude_from_level>
+
+  <!-- Other categories -->
+    <!-- Bring back score challenge in the game for servers supporting it (only for kaimei onwards) -->
+    <score_challenge __type="bool">0</score_challenge>
+    <!-- Handle favorites through data_mods/<game>.<friendID>.fav files (Note: allows UNLIMITED favorites as well as favorites without an account/server) -->
+    <local_favorites __type="bool">0</local_favorites>
+
+  <!-- Framerate -->
+    <!-- Fix animation speed at higher framerate for a smoother experience (lower visual latency) -->
+    <high_framerate __type="bool">0</high_framerate>
 
   <!-- Classic patches -->
     <!-- Prevent crash on boot when using a different default audio source (aka HDMI audio patch) -->
@@ -26,10 +42,8 @@
     <!-- Press numpad 9 in song to retire instantly, or on result screen to leave game session (thank you for playing) -->
     <!-- quick_retire with pfree also enables quick retry: press numpad 8 during song or on result screen to retry (keep holding to skip option select) -->
     <quick_retire __type="bool">0</quick_retire>
-
-  <!-- Network features -->
-    <!-- Bring back score challenge in the game for servers supporting it (only for kaimei onwards) -->
-    <score_challenge __type="bool">0</score_challenge>
+    <!-- Press 9 on option select screen to go back to song selection (requires quick_retire) -->
+    <back_to_song_select __type="bool">0</back_to_song_select>
 
   <!-- Audio offset -->
     <!-- Offset the audio by x ms (negative plays audio earlier). This will disable keysounds -->
@@ -49,7 +63,7 @@
     <hispeed_default_bpm __type="u16">0</hispeed_default_bpm>
 
     <!-- IIDX-like hard gauge (start with full gauge, instant fail if gauge drops to 0) -->
-    <!-- Gauge details: increment: +0.1% for each cool/great/good (like spicy gauge), decrement: -9% for each bad, or -4.5% if gauge <=30% ) -->
+    <!-- Gauge details: increment: +0.1% for each cool/great/good (like spicy gauge), decrement: -9% for each bad, or -4.5% if gauge <=30% (like IIDX) -->
     <iidx_hard_gauge __type="bool">0</iidx_hard_gauge>
     <!-- Force full options by default (useful when no numpad is available) -->
     <force_full_opt __type="bool">0</force_full_opt>
@@ -86,32 +100,28 @@
     <!-- Enable practice mode menu, (r-ran, constant chart speed, slow motion, etc...) -->
     <!-- PLEASE USE OFFLINE ONLY -->
     <practice_mode __type="bool">0</practice_mode>
-    <!-- Press 9 on option select screen to go back to song selection (requires quick_retire) -->
-    <!-- Note: causes issues with sounds on song select -->
-    <back_to_song_select __type="bool">0</back_to_song_select>
-
 
   <!-- ========================================================================================
        DEBUG OPTIONS FOLLOW (GENERALLY SHOULD NOT BE CHANGED UNLESS YOU KNOW WHAT YOU'RE DOING)
        ======================================================================================== -->
 
   <!-- Datecode and Multiboot -->
-    <!-- Force a different datecode than the one found in ea3-config (yyyymmdd00) -->
-    <force_datecode __type="str"></force_datecode>
+    <!-- Force a different datecode than the one found in ea3-config (yyyymmdd00), or use "auto" to let music limit decide for you if patch_db is on -->
+    <force_datecode __type="str">auto</force_datecode>
     <!-- Also apply force_datecode to network packets -->
     <network_datecode __type="bool">1</network_datecode>
-    <!-- Disable multiboot auto conf tuning (which takes place when using popn22_yyyymmddrr.dll format and an xml without force_datecode option) -->
+    <!-- Disable multiboot auto conf tuning (which takes place when using popn22_yyyymmddrr.dll format and an xml without another datecode in force_datecode) -->
     <disable_multiboot __type="bool">0</disable_multiboot>
 
   <!-- Timing and lanes -->
+    <!-- Base visual offset (value will be added to the base SD (-60) and base HD (-76) values) -->
+    <base_offset __type="s8">0</base_offset>
     <!-- Automatically play keysounds during songs -->
     <disable_keysounds __type="bool">0</disable_keysounds>
     <!-- Offset the keysounds by x ms (negative is earlier). With disable_keysounds, becomes an audio offset -->
     <keysound_offset __type="s8">0</keysound_offset>
     <!-- Adjust pop-kun and beam brightness (won't affect long popkuns) -->
     <beam_brightness __type="s8">0</beam_brightness>
-    <!-- Disable the builtin frame limiter (faster/smoother animations at 120fps+) -->
-    <fps_uncap __type="bool">0</fps_uncap>
     <!-- 1000Hz polling thread priority (for enhanced_polling only, might cause crashes on some systems if set too high)
          values THREAD_PRIORITY_LOWEST        -2
                 THREAD_PRIORITY_BELOW_NORMAL  -1
@@ -121,28 +131,40 @@
                 THREAD_PRIORITY_TIME_CRITICAL  15                             -->
     <enhanced_polling_priority __type="s8">1</enhanced_polling_priority>
 
-  <!-- Song db patches -->
-    <!-- Auto select patch file from data_mods folder (will detect datecode from ea3-config or force_datecode option) -->
+  <!-- Framerate -->
+    <!-- Reference fps value for high_framerate animation speed (0: use monitor refreshrate and disable frame limiter) -->
+    <high_framerate_fps __type="u16">0</high_framerate_fps>
+    <!-- Tune the builtin frame limiter the closest to high_framerate_fps -->
+    <high_framerate_limiter __type="bool">1</high_framerate_limiter>
+    <!-- Disable the builtin frame limiter -->
+    <fps_uncap __type="bool">0</fps_uncap>
+
+  <!-- Song db patches (requires patch_db) -->
+    <!-- Auto select patch file from data_mods folder based on music limit, or datecode otherwise (will detect datecode from ea3-config or force_datecode option) -->
     <patch_xml_auto __type="bool">1</patch_xml_auto>
     <!-- Manually set XML file containing patches (requires patch_xml_auto to be disabled) -->
     <patch_xml_filename __type="str"></patch_xml_filename>
-    <!-- Force the newly created buffers to be the same size as the original buffers -->
-    <disable_expansions __type="bool">0</disable_expansions>
-    <!-- Copy the new table information over top the old tables (automatically enables disable_expansions) -->
-    <disable_redirection __type="bool">0</disable_redirection>
+    <!-- Do not perform music limit checks for patch_xml_auto (not recommended) -->
+    <ignore_music_limit __type="bool">0</ignore_music_limit>
+
+  <!-- Custom category options (requires patch_db) -->
+    <!-- Also exclude omnimix (song id < 3000) tracks from version/level (requires custom_exclude_from_level or custom_exclude_from_version) -->
+    <exclude_omni __type="bool">0</exclude_omni>
+    <!-- Any new chart added to an existing song moves the song to the customs folder -->
+    <partial_entries __type="bool">0</partial_entries>
+    <!-- Minimum songid for a song to be seen as "custom" (e.g. use 4000 for real customs only) -->
+    <custom_categ_min_songid __type="u16">0</custom_categ_min_songid>
+    <!-- Category title for customs -->
+    <custom_category_title __type="str">Customs</custom_category_title>
+    <!-- Format used for category title (in BM2DXFontScript format, refer to BM2DXFontScript.md in popnhax_tools repo)  -->
+    <custom_category_format __type="str">[ol:4][olc:d92f0d]%s</custom_category_format>
+    <!-- Format used for custom song titles on song select (Note: colors not working for kaimei and above, but rotation does, e.g. "* [rz:3]%s[/rz]" ) -->
+    <custom_track_title_format __type="str"></custom_track_title_format>
+    <!-- Optional secondary format used for older games only (Full colors supported, e.g. "[ol:4][olc:d92f0d]%s") -->
+    <custom_track_title_format2 __type="str"></custom_track_title_format2>
 
   <!-- Translation -->
     <!-- Disable .dict string replacements and .ips patches -->
     <disable_translation __type="bool">0</disable_translation>
-    <!-- Dump applied translations and dll runtime data to help debug translation -->
-    <translation_debug __type="bool">0</translation_debug>
-
-  <!-- HARD Gauge alteration modes -->
-    <!-- Turn hard gauge into survival gauge (1: former COURSE gauge (-3.42%), 2: NORMAL gauge decrease rate (-4.99%), 3: IIDX HARD gauge decrease rate (-9%), 4: HARD gauge decrease rate (-19.94%)) -->
-    <survival_gauge __type="u8">0</survival_gauge>
-    <!-- Survival gauge has IIDX-like decrease rate adjustment (decrease amount halved when gauge is at 30% or less) -->
-    <survival_iidx __type="bool">0</survival_iidx>
-    <!-- Set survival gauge increase rate to spicy (+01/+01 instead of +02/+01) -->
-    <survival_spicy __type="bool">0</survival_spicy>
 
 </popnhax>

libdisasm/Makefile.am

@@ -0,0 +1,43 @@
+include_HEADERS = libdis.h
+lib_LTLIBRARIES = libdisasm.la
+libdisasm_la_SOURCES = \
+	ia32_implicit.c \
+	ia32_implicit.h \
+	ia32_insn.c \
+	ia32_insn.h \
+	ia32_invariant.c \
+	ia32_invariant.h \
+	ia32_modrm.c \
+	ia32_modrm.h \
+	ia32_opcode_tables.c \
+	ia32_opcode_tables.h \
+	ia32_operand.c \
+	ia32_operand.h \
+	ia32_reg.c \
+	ia32_reg.h \
+	ia32_settings.c \
+	ia32_settings.h \
+	libdis.h \
+	qword.h \
+	x86_disasm.c \
+	x86_format.c \
+	x86_imm.c \
+	x86_imm.h \
+	x86_insn.c \
+	x86_misc.c \
+	x86_operand_list.c \
+	x86_operand_list.h
+
+# Cheat to get non-autoconf swig into tarball,
+# even if it doesn't build by default.
+EXTRA_DIST = \
+swig/Makefile \
+swig/libdisasm.i \
+swig/libdisasm_oop.i \
+swig/python/Makefile-swig \
+swig/perl/Makefile-swig \
+swig/perl/Makefile.PL \
+swig/ruby/Makefile-swig \
+swig/ruby/extconf.rb \
+swig/tcl/Makefile-swig \
+swig/README

libdisasm/Makefile.in

@@ -0,0 +1,544 @@
+# Makefile.in generated by automake 1.10 from Makefile.am.
+# @configure_input@
+
+# Copyright (C) 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002,
+# 2003, 2004, 2005, 2006  Free Software Foundation, Inc.
+# This Makefile.in is free software; the Free Software Foundation
+# gives unlimited permission to copy and/or distribute it,
+# with or without modifications, as long as this notice is preserved.
+
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY, to the extent permitted by law; without
+# even the implied warranty of MERCHANTABILITY or FITNESS FOR A
+# PARTICULAR PURPOSE.
+
+@SET_MAKE@
+
+
+VPATH = @srcdir@
+pkgdatadir = $(datadir)/@PACKAGE@
+pkglibdir = $(libdir)/@PACKAGE@
+pkgincludedir = $(includedir)/@PACKAGE@
+am__cd = CDPATH="$${ZSH_VERSION+.}$(PATH_SEPARATOR)" && cd
+install_sh_DATA = $(install_sh) -c -m 644
+install_sh_PROGRAM = $(install_sh) -c
+install_sh_SCRIPT = $(install_sh) -c
+INSTALL_HEADER = $(INSTALL_DATA)
+transform = $(program_transform_name)
+NORMAL_INSTALL = :
+PRE_INSTALL = :
+POST_INSTALL = :
+NORMAL_UNINSTALL = :
+PRE_UNINSTALL = :
+POST_UNINSTALL = :
+build_triplet = @build@
+host_triplet = @host@
+subdir = libdisasm
+DIST_COMMON = $(include_HEADERS) $(srcdir)/Makefile.am \
+	$(srcdir)/Makefile.in TODO
+ACLOCAL_M4 = $(top_srcdir)/aclocal.m4
+am__aclocal_m4_deps = $(top_srcdir)/configure.ac
+am__configure_deps = $(am__aclocal_m4_deps) $(CONFIGURE_DEPENDENCIES) \
+	$(ACLOCAL_M4)
+mkinstalldirs = $(install_sh) -d
+CONFIG_HEADER = $(top_builddir)/config.h
+CONFIG_CLEAN_FILES =
+am__vpath_adj_setup = srcdirstrip=`echo "$(srcdir)" | sed 's|.|.|g'`;
+am__vpath_adj = case $$p in \
+    $(srcdir)/*) f=`echo "$$p" | sed "s|^$$srcdirstrip/||"`;; \
+    *) f=$$p;; \
+  esac;
+am__strip_dir = `echo $$p | sed -e 's|^.*/||'`;
+am__installdirs = "$(DESTDIR)$(libdir)" "$(DESTDIR)$(includedir)"
+libLTLIBRARIES_INSTALL = $(INSTALL)
+LTLIBRARIES = $(lib_LTLIBRARIES)
+libdisasm_la_LIBADD =
+am_libdisasm_la_OBJECTS = ia32_implicit.lo ia32_insn.lo \
+	ia32_invariant.lo ia32_modrm.lo ia32_opcode_tables.lo \
+	ia32_operand.lo ia32_reg.lo ia32_settings.lo x86_disasm.lo \
+	x86_format.lo x86_imm.lo x86_insn.lo x86_misc.lo \
+	x86_operand_list.lo
+libdisasm_la_OBJECTS = $(am_libdisasm_la_OBJECTS)
+DEFAULT_INCLUDES = -I. -I$(top_builddir)@am__isrc@
+depcomp = $(SHELL) $(top_srcdir)/depcomp
+am__depfiles_maybe = depfiles
+COMPILE = $(CC) $(DEFS) $(DEFAULT_INCLUDES) $(INCLUDES) $(AM_CPPFLAGS) \
+	$(CPPFLAGS) $(AM_CFLAGS) $(CFLAGS)
+LTCOMPILE = $(LIBTOOL) --tag=CC $(AM_LIBTOOLFLAGS) $(LIBTOOLFLAGS) \
+	--mode=compile $(CC) $(DEFS) $(DEFAULT_INCLUDES) $(INCLUDES) \
+	$(AM_CPPFLAGS) $(CPPFLAGS) $(AM_CFLAGS) $(CFLAGS)
+CCLD = $(CC)
+LINK = $(LIBTOOL) --tag=CC $(AM_LIBTOOLFLAGS) $(LIBTOOLFLAGS) \
+	--mode=link $(CCLD) $(AM_CFLAGS) $(CFLAGS) $(AM_LDFLAGS) \
+	$(LDFLAGS) -o $@
+SOURCES = $(libdisasm_la_SOURCES)
+DIST_SOURCES = $(libdisasm_la_SOURCES)
+includeHEADERS_INSTALL = $(INSTALL_HEADER)
+HEADERS = $(include_HEADERS)
+ETAGS = etags
+CTAGS = ctags
+DISTFILES = $(DIST_COMMON) $(DIST_SOURCES) $(TEXINFOS) $(EXTRA_DIST)
+ACLOCAL = @ACLOCAL@
+AMTAR = @AMTAR@
+AR = @AR@
+AUTOCONF = @AUTOCONF@
+AUTOHEADER = @AUTOHEADER@
+AUTOMAKE = @AUTOMAKE@
+AWK = @AWK@
+CC = @CC@
+CCDEPMODE = @CCDEPMODE@
+CFLAGS = @CFLAGS@
+CPP = @CPP@
+CPPFLAGS = @CPPFLAGS@
+CXX = @CXX@
+CXXCPP = @CXXCPP@
+CXXDEPMODE = @CXXDEPMODE@
+CXXFLAGS = @CXXFLAGS@
+CYGPATH_W = @CYGPATH_W@
+DEFS = @DEFS@
+DEPDIR = @DEPDIR@
+ECHO = @ECHO@
+ECHO_C = @ECHO_C@
+ECHO_N = @ECHO_N@
+ECHO_T = @ECHO_T@
+EGREP = @EGREP@
+EXEEXT = @EXEEXT@
+F77 = @F77@
+FFLAGS = @FFLAGS@
+GREP = @GREP@
+INSTALL = @INSTALL@
+INSTALL_DATA = @INSTALL_DATA@
+INSTALL_PROGRAM = @INSTALL_PROGRAM@
+INSTALL_SCRIPT = @INSTALL_SCRIPT@
+INSTALL_STRIP_PROGRAM = @INSTALL_STRIP_PROGRAM@
+LDFLAGS = @LDFLAGS@
+LIBOBJS = @LIBOBJS@
+LIBS = @LIBS@
+LIBTOOL = @LIBTOOL@
+LN_S = @LN_S@
+LTLIBOBJS = @LTLIBOBJS@
+MAINT = @MAINT@
+MAKEINFO = @MAKEINFO@
+MKDIR_P = @MKDIR_P@
+OBJEXT = @OBJEXT@
+PACKAGE = @PACKAGE@
+PACKAGE_BUGREPORT = @PACKAGE_BUGREPORT@
+PACKAGE_NAME = @PACKAGE_NAME@
+PACKAGE_STRING = @PACKAGE_STRING@
+PACKAGE_TARNAME = @PACKAGE_TARNAME@
+PACKAGE_VERSION = @PACKAGE_VERSION@
+PATH_SEPARATOR = @PATH_SEPARATOR@
+RANLIB = @RANLIB@
+SED = @SED@
+SET_MAKE = @SET_MAKE@
+SHELL = @SHELL@
+STRIP = @STRIP@
+VERSION = @VERSION@
+abs_builddir = @abs_builddir@
+abs_srcdir = @abs_srcdir@
+abs_top_builddir = @abs_top_builddir@
+abs_top_srcdir = @abs_top_srcdir@
+ac_ct_CC = @ac_ct_CC@
+ac_ct_CXX = @ac_ct_CXX@
+ac_ct_F77 = @ac_ct_F77@
+am__include = @am__include@
+am__leading_dot = @am__leading_dot@
+am__quote = @am__quote@
+am__tar = @am__tar@
+am__untar = @am__untar@
+bindir = @bindir@
+build = @build@
+build_alias = @build_alias@
+build_cpu = @build_cpu@
+build_os = @build_os@
+build_vendor = @build_vendor@
+builddir = @builddir@
+datadir = @datadir@
+datarootdir = @datarootdir@
+docdir = @docdir@
+dvidir = @dvidir@
+exec_prefix = @exec_prefix@
+host = @host@
+host_alias = @host_alias@
+host_cpu = @host_cpu@
+host_os = @host_os@
+host_vendor = @host_vendor@
+htmldir = @htmldir@
+includedir = @includedir@
+infodir = @infodir@
+install_sh = @install_sh@
+libdir = @libdir@
+libexecdir = @libexecdir@
+localedir = @localedir@
+localstatedir = @localstatedir@
+mandir = @mandir@
+mkdir_p = @mkdir_p@
+oldincludedir = @oldincludedir@
+pdfdir = @pdfdir@
+prefix = @prefix@
+program_transform_name = @program_transform_name@
+psdir = @psdir@
+sbindir = @sbindir@
+sharedstatedir = @sharedstatedir@
+srcdir = @srcdir@
+sysconfdir = @sysconfdir@
+target_alias = @target_alias@
+top_builddir = @top_builddir@
+top_srcdir = @top_srcdir@
+include_HEADERS = libdis.h
+lib_LTLIBRARIES = libdisasm.la
+libdisasm_la_SOURCES = \
+	ia32_implicit.c \
+	ia32_implicit.h \
+	ia32_insn.c \
+	ia32_insn.h \
+	ia32_invariant.c \
+	ia32_invariant.h \
+	ia32_modrm.c \
+	ia32_modrm.h \
+	ia32_opcode_tables.c \
+	ia32_opcode_tables.h \
+	ia32_operand.c \
+	ia32_operand.h \
+	ia32_reg.c \
+	ia32_reg.h \
+	ia32_settings.c \
+	ia32_settings.h \
+	libdis.h \
+	qword.h \
+	x86_disasm.c \
+	x86_format.c \
+	x86_imm.c \
+	x86_imm.h \
+	x86_insn.c \
+	x86_misc.c \
+	x86_operand_list.c \
+	x86_operand_list.h
+
+
+# Cheat to get non-autoconf swig into tarball,
+# even if it doesn't build by default.
+EXTRA_DIST = \
+swig/Makefile \
+swig/libdisasm.i \
+swig/libdisasm_oop.i \
+swig/python/Makefile-swig \
+swig/perl/Makefile-swig \
+swig/perl/Makefile.PL \
+swig/ruby/Makefile-swig \
+swig/ruby/extconf.rb \
+swig/tcl/Makefile-swig \
+swig/README
+
+all: all-am
+
+.SUFFIXES:
+.SUFFIXES: .c .lo .o .obj
+$(srcdir)/Makefile.in: @MAINTAINER_MODE_TRUE@ $(srcdir)/Makefile.am  $(am__configure_deps)
+	@for dep in $?; do \
+	  case '$(am__configure_deps)' in \
+	    *$$dep*) \
+	      cd $(top_builddir) && $(MAKE) $(AM_MAKEFLAGS) am--refresh \
+		&& exit 0; \
+	      exit 1;; \
+	  esac; \
+	done; \
+	echo ' cd $(top_srcdir) && $(AUTOMAKE) --gnu  libdisasm/Makefile'; \
+	cd $(top_srcdir) && \
+	  $(AUTOMAKE) --gnu  libdisasm/Makefile
+.PRECIOUS: Makefile
+Makefile: $(srcdir)/Makefile.in $(top_builddir)/config.status
+	@case '$?' in \
+	  *config.status*) \
+	    cd $(top_builddir) && $(MAKE) $(AM_MAKEFLAGS) am--refresh;; \
+	  *) \
+	    echo ' cd $(top_builddir) && $(SHELL) ./config.status $(subdir)/$@ $(am__depfiles_maybe)'; \
+	    cd $(top_builddir) && $(SHELL) ./config.status $(subdir)/$@ $(am__depfiles_maybe);; \
+	esac;
+
+$(top_builddir)/config.status: $(top_srcdir)/configure $(CONFIG_STATUS_DEPENDENCIES)
+	cd $(top_builddir) && $(MAKE) $(AM_MAKEFLAGS) am--refresh
+
+$(top_srcdir)/configure: @MAINTAINER_MODE_TRUE@ $(am__configure_deps)
+	cd $(top_builddir) && $(MAKE) $(AM_MAKEFLAGS) am--refresh
+$(ACLOCAL_M4): @MAINTAINER_MODE_TRUE@ $(am__aclocal_m4_deps)
+	cd $(top_builddir) && $(MAKE) $(AM_MAKEFLAGS) am--refresh
+install-libLTLIBRARIES: $(lib_LTLIBRARIES)
+	@$(NORMAL_INSTALL)
+	test -z "$(libdir)" || $(MKDIR_P) "$(DESTDIR)$(libdir)"
+	@list='$(lib_LTLIBRARIES)'; for p in $$list; do \
+	  if test -f $$p; then \
+	    f=$(am__strip_dir) \
+	    echo " $(LIBTOOL) --mode=install $(libLTLIBRARIES_INSTALL) $(INSTALL_STRIP_FLAG) '$$p' '$(DESTDIR)$(libdir)/$$f'"; \
+	    $(LIBTOOL) --mode=install $(libLTLIBRARIES_INSTALL) $(INSTALL_STRIP_FLAG) "$$p" "$(DESTDIR)$(libdir)/$$f"; \
+	  else :; fi; \
+	done
+
+uninstall-libLTLIBRARIES:
+	@$(NORMAL_UNINSTALL)
+	@list='$(lib_LTLIBRARIES)'; for p in $$list; do \
+	  p=$(am__strip_dir) \
+	  echo " $(LIBTOOL) --mode=uninstall rm -f '$(DESTDIR)$(libdir)/$$p'"; \
+	  $(LIBTOOL) --mode=uninstall rm -f "$(DESTDIR)$(libdir)/$$p"; \
+	done
+
+clean-libLTLIBRARIES:
+	-test -z "$(lib_LTLIBRARIES)" || rm -f $(lib_LTLIBRARIES)
+	@list='$(lib_LTLIBRARIES)'; for p in $$list; do \
+	  dir="`echo $$p | sed -e 's|/[^/]*$$||'`"; \
+	  test "$$dir" != "$$p" || dir=.; \
+	  echo "rm -f \"$${dir}/so_locations\""; \
+	  rm -f "$${dir}/so_locations"; \
+	done
+libdisasm.la: $(libdisasm_la_OBJECTS) $(libdisasm_la_DEPENDENCIES) 
+	$(LINK) -rpath $(libdir) $(libdisasm_la_OBJECTS) $(libdisasm_la_LIBADD) $(LIBS)
+
+mostlyclean-compile:
+	-rm -f *.$(OBJEXT)
+
+distclean-compile:
+	-rm -f *.tab.c
+
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/ia32_implicit.Plo@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/ia32_insn.Plo@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/ia32_invariant.Plo@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/ia32_modrm.Plo@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/ia32_opcode_tables.Plo@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/ia32_operand.Plo@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/ia32_reg.Plo@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/ia32_settings.Plo@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/x86_disasm.Plo@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/x86_format.Plo@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/x86_imm.Plo@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/x86_insn.Plo@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/x86_misc.Plo@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/x86_operand_list.Plo@am__quote@
+
+.c.o:
+@am__fastdepCC_TRUE@	$(COMPILE) -MT $@ -MD -MP -MF $(DEPDIR)/$*.Tpo -c -o $@ $<
+@am__fastdepCC_TRUE@	mv -f $(DEPDIR)/$*.Tpo $(DEPDIR)/$*.Po
+@AMDEP_TRUE@@am__fastdepCC_FALSE@	source='$<' object='$@' libtool=no @AMDEPBACKSLASH@
+@AMDEP_TRUE@@am__fastdepCC_FALSE@	DEPDIR=$(DEPDIR) $(CCDEPMODE) $(depcomp) @AMDEPBACKSLASH@
+@am__fastdepCC_FALSE@	$(COMPILE) -c $<
+
+.c.obj:
+@am__fastdepCC_TRUE@	$(COMPILE) -MT $@ -MD -MP -MF $(DEPDIR)/$*.Tpo -c -o $@ `$(CYGPATH_W) '$<'`
+@am__fastdepCC_TRUE@	mv -f $(DEPDIR)/$*.Tpo $(DEPDIR)/$*.Po
+@AMDEP_TRUE@@am__fastdepCC_FALSE@	source='$<' object='$@' libtool=no @AMDEPBACKSLASH@
+@AMDEP_TRUE@@am__fastdepCC_FALSE@	DEPDIR=$(DEPDIR) $(CCDEPMODE) $(depcomp) @AMDEPBACKSLASH@
+@am__fastdepCC_FALSE@	$(COMPILE) -c `$(CYGPATH_W) '$<'`
+
+.c.lo:
+@am__fastdepCC_TRUE@	$(LTCOMPILE) -MT $@ -MD -MP -MF $(DEPDIR)/$*.Tpo -c -o $@ $<
+@am__fastdepCC_TRUE@	mv -f $(DEPDIR)/$*.Tpo $(DEPDIR)/$*.Plo
+@AMDEP_TRUE@@am__fastdepCC_FALSE@	source='$<' object='$@' libtool=yes @AMDEPBACKSLASH@
+@AMDEP_TRUE@@am__fastdepCC_FALSE@	DEPDIR=$(DEPDIR) $(CCDEPMODE) $(depcomp) @AMDEPBACKSLASH@
+@am__fastdepCC_FALSE@	$(LTCOMPILE) -c -o $@ $<
+
+mostlyclean-libtool:
+	-rm -f *.lo
+
+clean-libtool:
+	-rm -rf .libs _libs
+install-includeHEADERS: $(include_HEADERS)
+	@$(NORMAL_INSTALL)
+	test -z "$(includedir)" || $(MKDIR_P) "$(DESTDIR)$(includedir)"
+	@list='$(include_HEADERS)'; for p in $$list; do \
+	  if test -f "$$p"; then d=; else d="$(srcdir)/"; fi; \
+	  f=$(am__strip_dir) \
+	  echo " $(includeHEADERS_INSTALL) '$$d$$p' '$(DESTDIR)$(includedir)/$$f'"; \
+	  $(includeHEADERS_INSTALL) "$$d$$p" "$(DESTDIR)$(includedir)/$$f"; \
+	done
+
+uninstall-includeHEADERS:
+	@$(NORMAL_UNINSTALL)
+	@list='$(include_HEADERS)'; for p in $$list; do \
+	  f=$(am__strip_dir) \
+	  echo " rm -f '$(DESTDIR)$(includedir)/$$f'"; \
+	  rm -f "$(DESTDIR)$(includedir)/$$f"; \
+	done
+
+ID: $(HEADERS) $(SOURCES) $(LISP) $(TAGS_FILES)
+	list='$(SOURCES) $(HEADERS) $(LISP) $(TAGS_FILES)'; \
+	unique=`for i in $$list; do \
+	    if test -f "$$i"; then echo $$i; else echo $(srcdir)/$$i; fi; \
+	  done | \
+	  $(AWK) '    { files[$$0] = 1; } \
+	       END { for (i in files) print i; }'`; \
+	mkid -fID $$unique
+tags: TAGS
+
+TAGS:  $(HEADERS) $(SOURCES)  $(TAGS_DEPENDENCIES) \
+		$(TAGS_FILES) $(LISP)
+	tags=; \
+	here=`pwd`; \
+	list='$(SOURCES) $(HEADERS)  $(LISP) $(TAGS_FILES)'; \
+	unique=`for i in $$list; do \
+	    if test -f "$$i"; then echo $$i; else echo $(srcdir)/$$i; fi; \
+	  done | \
+	  $(AWK) '    { files[$$0] = 1; } \
+	       END { for (i in files) print i; }'`; \
+	if test -z "$(ETAGS_ARGS)$$tags$$unique"; then :; else \
+	  test -n "$$unique" || unique=$$empty_fix; \
+	  $(ETAGS) $(ETAGSFLAGS) $(AM_ETAGSFLAGS) $(ETAGS_ARGS) \
+	    $$tags $$unique; \
+	fi
+ctags: CTAGS
+CTAGS:  $(HEADERS) $(SOURCES)  $(TAGS_DEPENDENCIES) \
+		$(TAGS_FILES) $(LISP)
+	tags=; \
+	here=`pwd`; \
+	list='$(SOURCES) $(HEADERS)  $(LISP) $(TAGS_FILES)'; \
+	unique=`for i in $$list; do \
+	    if test -f "$$i"; then echo $$i; else echo $(srcdir)/$$i; fi; \
+	  done | \
+	  $(AWK) '    { files[$$0] = 1; } \
+	       END { for (i in files) print i; }'`; \
+	test -z "$(CTAGS_ARGS)$$tags$$unique" \
+	  || $(CTAGS) $(CTAGSFLAGS) $(AM_CTAGSFLAGS) $(CTAGS_ARGS) \
+	     $$tags $$unique
+
+GTAGS:
+	here=`$(am__cd) $(top_builddir) && pwd` \
+	  && cd $(top_srcdir) \
+	  && gtags -i $(GTAGS_ARGS) $$here
+
+distclean-tags:
+	-rm -f TAGS ID GTAGS GRTAGS GSYMS GPATH tags
+
+distdir: $(DISTFILES)
+	@srcdirstrip=`echo "$(srcdir)" | sed 's/[].[^$$\\*]/\\\\&/g'`; \
+	topsrcdirstrip=`echo "$(top_srcdir)" | sed 's/[].[^$$\\*]/\\\\&/g'`; \
+	list='$(DISTFILES)'; \
+	  dist_files=`for file in $$list; do echo $$file; done | \
+	  sed -e "s|^$$srcdirstrip/||;t" \
+	      -e "s|^$$topsrcdirstrip/|$(top_builddir)/|;t"`; \
+	case $$dist_files in \
+	  */*) $(MKDIR_P) `echo "$$dist_files" | \
+			   sed '/\//!d;s|^|$(distdir)/|;s,/[^/]*$$,,' | \
+			   sort -u` ;; \
+	esac; \
+	for file in $$dist_files; do \
+	  if test -f $$file || test -d $$file; then d=.; else d=$(srcdir); fi; \
+	  if test -d $$d/$$file; then \
+	    dir=`echo "/$$file" | sed -e 's,/[^/]*$$,,'`; \
+	    if test -d $(srcdir)/$$file && test $$d != $(srcdir); then \
+	      cp -pR $(srcdir)/$$file $(distdir)$$dir || exit 1; \
+	    fi; \
+	    cp -pR $$d/$$file $(distdir)$$dir || exit 1; \
+	  else \
+	    test -f $(distdir)/$$file \
+	    || cp -p $$d/$$file $(distdir)/$$file \
+	    || exit 1; \
+	  fi; \
+	done
+check-am: all-am
+check: check-am
+all-am: Makefile $(LTLIBRARIES) $(HEADERS)
+installdirs:
+	for dir in "$(DESTDIR)$(libdir)" "$(DESTDIR)$(includedir)"; do \
+	  test -z "$$dir" || $(MKDIR_P) "$$dir"; \
+	done
+install: install-am
+install-exec: install-exec-am
+install-data: install-data-am
+uninstall: uninstall-am
+
+install-am: all-am
+	@$(MAKE) $(AM_MAKEFLAGS) install-exec-am install-data-am
+
+installcheck: installcheck-am
+install-strip:
+	$(MAKE) $(AM_MAKEFLAGS) INSTALL_PROGRAM="$(INSTALL_STRIP_PROGRAM)" \
+	  install_sh_PROGRAM="$(INSTALL_STRIP_PROGRAM)" INSTALL_STRIP_FLAG=-s \
+	  `test -z '$(STRIP)' || \
+	    echo "INSTALL_PROGRAM_ENV=STRIPPROG='$(STRIP)'"` install
+mostlyclean-generic:
+
+clean-generic:
+
+distclean-generic:
+	-test -z "$(CONFIG_CLEAN_FILES)" || rm -f $(CONFIG_CLEAN_FILES)
+
+maintainer-clean-generic:
+	@echo "This command is intended for maintainers to use"
+	@echo "it deletes files that may require special tools to rebuild."
+clean: clean-am
+
+clean-am: clean-generic clean-libLTLIBRARIES clean-libtool \
+	mostlyclean-am
+
+distclean: distclean-am
+	-rm -rf ./$(DEPDIR)
+	-rm -f Makefile
+distclean-am: clean-am distclean-compile distclean-generic \
+	distclean-tags
+
+dvi: dvi-am
+
+dvi-am:
+
+html: html-am
+
+info: info-am
+
+info-am:
+
+install-data-am: install-includeHEADERS
+
+install-dvi: install-dvi-am
+
+install-exec-am: install-libLTLIBRARIES
+
+install-html: install-html-am
+
+install-info: install-info-am
+
+install-man:
+
+install-pdf: install-pdf-am
+
+install-ps: install-ps-am
+
+installcheck-am:
+
+maintainer-clean: maintainer-clean-am
+	-rm -rf ./$(DEPDIR)
+	-rm -f Makefile
+maintainer-clean-am: distclean-am maintainer-clean-generic
+
+mostlyclean: mostlyclean-am
+
+mostlyclean-am: mostlyclean-compile mostlyclean-generic \
+	mostlyclean-libtool
+
+pdf: pdf-am
+
+pdf-am:
+
+ps: ps-am
+
+ps-am:
+
+uninstall-am: uninstall-includeHEADERS uninstall-libLTLIBRARIES
+
+.MAKE: install-am install-strip
+
+.PHONY: CTAGS GTAGS all all-am check check-am clean clean-generic \
+	clean-libLTLIBRARIES clean-libtool ctags distclean \
+	distclean-compile distclean-generic distclean-libtool \
+	distclean-tags distdir dvi dvi-am html html-am info info-am \
+	install install-am install-data install-data-am install-dvi \
+	install-dvi-am install-exec install-exec-am install-html \
+	install-html-am install-includeHEADERS install-info \
+	install-info-am install-libLTLIBRARIES install-man install-pdf \
+	install-pdf-am install-ps install-ps-am install-strip \
+	installcheck installcheck-am installdirs maintainer-clean \
+	maintainer-clean-generic mostlyclean mostlyclean-compile \
+	mostlyclean-generic mostlyclean-libtool pdf pdf-am ps ps-am \
+	tags uninstall uninstall-am uninstall-includeHEADERS \
+	uninstall-libLTLIBRARIES
+
+# Tell versions [3.59,3.63) of GNU make to not export all variables.
+# Otherwise a system limit (for SysV at least) may be exceeded.
+.NOEXPORT:

libdisasm/Module.mk

@@ -0,0 +1,19 @@
+libs             += libdisasm
+
+cflags += -w
+
+src_libdisasm      :=  \
+	ia32_implicit.c \
+	ia32_insn.c \
+	ia32_invariant.c \
+	ia32_modrm.c \
+	ia32_opcode_tables.c \
+	ia32_operand.c \
+	ia32_reg.c \
+	ia32_settings.c \
+	x86_disasm.c \
+	x86_format.c \
+	x86_imm.c \
+	x86_insn.c \
+	x86_misc.c \
+	x86_operand_list.c

libdisasm/TODO

@@ -0,0 +1,43 @@
+x86_format.c
+------------
+intel: jmpf -> jmp, callf -> call
+att: jmpf -> ljmp, callf -> lcall
+
+opcode table
+------------
+finish typing instructions
+fix flag clear/set/toggle types
+
+ix64 stuff
+----------
+document output file formats in web page
+features doc: register aliases, implicit operands, stack mods,
+ring0 flags, eflags, cpu model/isa
+
+ia32_handle_* implementation
+
+fix operand 0F C2
+CMPPS
+
+* sysenter, sysexit as CALL types -- preceded by MSR writes
+* SYSENTER/SYSEXIT stack : overwrites SS, ESP
+* stos, cmps, scas, movs, ins, outs, lods -> OP_PTR
+* OP_SIZE in implicit operands
+* use OP_SIZE to choose reg sizes!
+
+DONE?? :
+implicit operands: provide action ?
+e.g. add/inc for stach, write, etc 
+replace table numbers in opcodes.dat with
+#defines for table names
+
+replace 0  with INSN_INVALID   [or maybe FF for imnvalid and 00 for Not Applicable */
+no wait that is only for prefix tables -- n/p
+
+if ( prefx) only use if insn != invalid
+
+these should cover all the wacky disasm exceptions 
+
+for the rep one we can chet, match only a 0x90
+
+todo: privilege | ring

libdisasm/ia32_implicit.c

@@ -0,0 +1,422 @@
+#include <stdlib.h>
+
+#include "ia32_implicit.h"
+#include "ia32_insn.h"
+#include "ia32_reg.h"
+#include "x86_operand_list.h"
+
+/* Conventions: Register operands which are aliases of another register 
+ *   operand (e.g. AX in one operand and AL in another) assume that the
+ *   operands are different registers and that alias tracking will resolve
+ *   data flow. This means that something like
+ *   	mov ax, al
+ *   would have 'write only' access for AX and 'read only' access for AL,
+ *   even though both AL and AX are read and written */
+typedef struct {
+	uint32_t type;
+	uint32_t operand;
+} op_implicit_list_t;
+
+static op_implicit_list_t list_aaa[] = 
+	/* 37 : AAA : rw AL */
+	/* 3F : AAS : rw AL */
+	{{ OP_R | OP_W, REG_BYTE_OFFSET }, {0}};	/* aaa */
+
+static op_implicit_list_t list_aad[] = 
+	/* D5 0A, D5 (ib) : AAD : rw AX */
+	/* D4 0A, D4 (ib) : AAM : rw AX */
+	{{ OP_R | OP_W, REG_WORD_OFFSET }, {0}};	/* aad */
+
+static op_implicit_list_t list_call[] = 
+	/* E8, FF, 9A, FF : CALL : rw ESP, rw EIP */
+	/* C2, C3, CA, CB : RET  : rw ESP, rw EIP */
+	{{ OP_R | OP_W, REG_EIP_INDEX }, 
+	 { OP_R | OP_W, REG_ESP_INDEX }, {0}};	/* call, ret */
+
+static op_implicit_list_t list_cbw[] = 
+	/* 98 : CBW : r AL, rw AX */
+	{{ OP_R | OP_W, REG_WORD_OFFSET },
+	 { OP_R, REG_BYTE_OFFSET}, {0}};		/* cbw */
+
+static op_implicit_list_t list_cwde[] = 
+	/* 98 : CWDE : r AX, rw EAX */
+	{{ OP_R | OP_W, REG_DWORD_OFFSET },
+	 { OP_R, REG_WORD_OFFSET }, {0}};		/* cwde */
+
+static op_implicit_list_t list_clts[] = 
+	/* 0F 06 : CLTS : rw CR0 */
+	{{ OP_R | OP_W, REG_CTRL_OFFSET}, {0}};	/* clts */
+
+static op_implicit_list_t list_cmpxchg[] = 
+	/* 0F B0 : CMPXCHG : rw AL */
+	{{ OP_R | OP_W, REG_BYTE_OFFSET }, {0}};	/* cmpxchg */
+
+static op_implicit_list_t list_cmpxchgb[] = 
+	/* 0F B1 : CMPXCHG : rw EAX */
+	{{ OP_R | OP_W, REG_DWORD_OFFSET }, {0}};	/* cmpxchg */
+
+static op_implicit_list_t list_cmpxchg8b[] = 
+	/* 0F C7 : CMPXCHG8B : rw EDX, rw EAX, r ECX, r EBX */
+	{{ OP_R | OP_W, REG_DWORD_OFFSET }, 
+	 { OP_R | OP_W, REG_DWORD_OFFSET + 2 }, 
+	 { OP_R, REG_DWORD_OFFSET + 1 }, 
+	 { OP_R, REG_DWORD_OFFSET + 3 }, {0}};	/* cmpxchg8b */
+
+static op_implicit_list_t list_cpuid[] = 
+	/* 0F A2 : CPUID : rw EAX, w EBX, w ECX, w EDX */
+	{{ OP_R | OP_W, REG_DWORD_OFFSET }, 
+	 { OP_W, REG_DWORD_OFFSET + 1 }, 
+	 { OP_W, REG_DWORD_OFFSET + 2 }, 
+	 { OP_W, REG_DWORD_OFFSET + 3 }, {0}};	/* cpuid */
+
+static op_implicit_list_t list_cwd[] = 
+	/* 99 : CWD/CWQ : rw EAX, w EDX */
+	{{ OP_R | OP_W, REG_DWORD_OFFSET }, 
+	 { OP_W, REG_DWORD_OFFSET + 2 }, {0}};	/* cwd */
+
+static op_implicit_list_t list_daa[] = 
+	/* 27 : DAA : rw AL */
+	/* 2F : DAS : rw AL */
+	{{ OP_R | OP_W, REG_BYTE_OFFSET }, {0}};	/* daa */
+
+static op_implicit_list_t list_idiv[] = 
+	/* F6 : DIV, IDIV : r AX, w AL, w AH */
+	/* FIXED: first op was EAX, not Aw. TODO: verify! */
+	{{ OP_R, REG_WORD_OFFSET }, 
+	  { OP_W, REG_BYTE_OFFSET },
+	  { OP_W, REG_BYTE_OFFSET + 4 }, {0}};	/* div */
+
+static op_implicit_list_t list_div[] = 
+	/* F7 : DIV, IDIV : rw EDX, rw EAX */
+	{{ OP_R | OP_W, REG_DWORD_OFFSET + 2 }, 
+	  { OP_R | OP_W, REG_DWORD_OFFSET }, {0}};	/* div */
+
+static op_implicit_list_t list_enter[] = 
+	/* C8 : ENTER : rw ESP w EBP */
+	{{ OP_R | OP_W, REG_DWORD_OFFSET + 4 }, 
+	 { OP_R, REG_DWORD_OFFSET + 5 }, {0}};	/* enter */
+
+static op_implicit_list_t list_f2xm1[] = 
+	/* D9 F0 : F2XM1 : rw ST(0) */
+	/* D9 E1 : FABS : rw ST(0) */
+	/* D9 E0 : FCHS : rw ST(0) */
+	/* D9 FF : FCOS : rw ST(0)*/
+	/* D8, DA : FDIV : rw ST(0) */
+	/* D8, DA : FDIVR : rw ST(0) */
+	/* D9 F2 : FPTAN : rw ST(0) */
+	/* D9 FC : FRNDINT : rw ST(0) */
+	/* D9 FB : FSINCOS : rw ST(0) */
+	/* D9 FE : FSIN : rw ST(0) */
+	/* D9 FA : FSQRT : rw ST(0) */
+	/* D9 F4 : FXTRACT : rw ST(0) */
+	{{ OP_R | OP_W, REG_FPU_OFFSET }, {0}};	/* f2xm1 */
+
+static op_implicit_list_t list_fcom[] = 
+	/* D8, DC, DE D9 : FCOM : r ST(0) */
+	/* DE, DA : FICOM : r ST(0) */
+	/* DF, D8 : FIST : r ST(0) */
+	/* D9 E4 : FTST : r ST(0) */
+	/* D9 E5 : FXAM : r ST(0) */
+	{{ OP_R, REG_FPU_OFFSET }, {0}};		/* fcom */
+
+static op_implicit_list_t list_fpatan[] = 
+	/* D9 F3 : FPATAN : r ST(0), rw ST(1) */
+	{{ OP_R, REG_FPU_OFFSET }, {0}};		/* fpatan */
+
+static op_implicit_list_t list_fprem[] = 
+	/* D9 F8, D9 F5 : FPREM : rw ST(0) r ST(1) */
+	/* D9 FD : FSCALE : rw ST(0), r ST(1) */
+	{{ OP_R | OP_W, REG_FPU_OFFSET }, 	
+	 { OP_R, REG_FPU_OFFSET + 1 }, {0}};	/* fprem */
+
+static op_implicit_list_t list_faddp[] = 
+	/* DE C1 : FADDP : r ST(0), rw ST(1) */
+	/* DE E9 : FSUBP : r ST(0), rw ST(1) */
+	/* D9 F1 : FYL2X : r ST(0), rw ST(1) */
+	/* D9 F9 : FYL2XP1 : r ST(0), rw ST(1) */
+	{{ OP_R, REG_FPU_OFFSET },
+	 { OP_R | OP_W, REG_FPU_OFFSET + 1 }, {0}};	/* faddp */
+
+static op_implicit_list_t list_fucompp[] = 
+	/* DA E9 : FUCOMPP : r ST(0), r ST(1) */
+	{{ OP_R, REG_FPU_OFFSET },
+	 { OP_R, REG_FPU_OFFSET + 1 }, {0}};	/* fucompp */
+
+static op_implicit_list_t list_imul[] = 
+	/* F6 : IMUL : r AL, w AX */
+	/* F6 : MUL : r AL, w AX */
+	{{ OP_R, REG_BYTE_OFFSET },
+	 { OP_W, REG_WORD_OFFSET }, {0}};		/* imul */
+
+static op_implicit_list_t list_mul[] = 
+	/* F7 : IMUL : rw EAX, w EDX */
+	/* F7 : MUL : rw EAX, w EDX */
+	{{ OP_R | OP_W, REG_DWORD_OFFSET },
+	 { OP_W, REG_DWORD_OFFSET + 2 }, {0}};	/* imul */
+
+static op_implicit_list_t list_lahf[] = 
+	/* 9F : LAHF : r EFLAGS, w AH */
+	{{ OP_R, REG_FLAGS_INDEX },
+	 { OP_W, REG_BYTE_OFFSET + 4 }, {0}};	/* lahf */
+
+static op_implicit_list_t list_ldmxcsr[] = 
+	/* 0F AE : LDMXCSR : w MXCSR SSE Control Status Reg */
+	{{ OP_W, REG_MXCSG_INDEX }, {0}};		/* ldmxcsr */
+
+static op_implicit_list_t list_leave[] = 
+	/* C9 : LEAVE :  rw ESP, w EBP */
+	{{ OP_R | OP_W, REG_ESP_INDEX },
+	 { OP_W, REG_DWORD_OFFSET + 5 }, {0}};	/* leave */
+
+static op_implicit_list_t list_lgdt[] = 
+	/* 0F 01 : LGDT : w GDTR */
+	{{ OP_W, REG_GDTR_INDEX }, {0}};		/* lgdt */
+
+static op_implicit_list_t list_lidt[] = 
+	/* 0F 01 : LIDT : w IDTR */
+	{{ OP_W, REG_IDTR_INDEX }, {0}};		/* lidt */
+
+static op_implicit_list_t list_lldt[] = 
+	/* 0F 00 : LLDT : w LDTR */
+	{{ OP_W, REG_LDTR_INDEX }, {0}};		/* lldt */
+
+static op_implicit_list_t list_lmsw[] = 
+	/* 0F 01 : LMSW : w CR0 */
+	{{ OP_W, REG_CTRL_OFFSET }, {0}};		/* lmsw */
+
+static op_implicit_list_t list_loop[] = 
+	/* E0, E1, E2 : LOOP : rw ECX */
+	{{ OP_R | OP_W, REG_DWORD_OFFSET + 1 }, {0}};/* loop */
+
+static op_implicit_list_t list_ltr[] = 
+	/* 0F 00 : LTR : w Task Register */
+	{{ OP_W, REG_TR_INDEX }, {0}};		/* ltr */
+
+static op_implicit_list_t list_pop[] = 
+	/* 8F, 58, 1F, 07, 17, 0F A1, 0F A9 : POP : rw ESP */
+	/* FF, 50, 6A, 68, 0E, 16, 1E, 06, 0F A0, 0F A8 : PUSH : rw ESP */
+	{{ OP_R | OP_W, REG_ESP_INDEX }, {0}};	/* pop, push */
+
+static op_implicit_list_t list_popad[] = 
+	/* 61 : POPAD : rw esp, w edi esi ebp ebx edx ecx eax */
+	{{ OP_R | OP_W, REG_ESP_INDEX },
+	 { OP_W, REG_DWORD_OFFSET + 7 },
+	 { OP_W, REG_DWORD_OFFSET + 6 },
+	 { OP_W, REG_DWORD_OFFSET + 5 },
+	 { OP_W, REG_DWORD_OFFSET + 3 },
+	 { OP_W, REG_DWORD_OFFSET + 2 },
+	 { OP_W, REG_DWORD_OFFSET + 1 },
+	 { OP_W, REG_DWORD_OFFSET }, {0}};		/* popad */
+
+static op_implicit_list_t list_popfd[] = 
+	/* 9D : POPFD : rw esp, w eflags */
+	{{ OP_R | OP_W, REG_ESP_INDEX },
+	 { OP_W, REG_FLAGS_INDEX }, {0}};		/* popfd */
+
+static op_implicit_list_t list_pushad[] = 
+	/* FF, 50, 6A, 68, 0E, 16, 1E, 06, 0F A0, 0F A8 : PUSH : rw ESP */
+	/* 60 : PUSHAD : rw esp, r eax ecx edx ebx esp ebp esi edi */
+	{{ OP_R | OP_W, REG_ESP_INDEX },
+	 { OP_R, REG_DWORD_OFFSET },
+	 { OP_R, REG_DWORD_OFFSET + 1 },
+	 { OP_R, REG_DWORD_OFFSET + 2 },
+	 { OP_R, REG_DWORD_OFFSET + 3 },
+	 { OP_R, REG_DWORD_OFFSET + 5 },
+	 { OP_R, REG_DWORD_OFFSET + 6 },
+	 { OP_R, REG_DWORD_OFFSET + 7 }, {0}};	/* pushad */
+
+static op_implicit_list_t list_pushfd[] = 
+	/* 9C : PUSHFD : rw esp, r eflags */
+	{{ OP_R | OP_W, REG_ESP_INDEX },
+	 { OP_R, REG_FLAGS_INDEX }, {0}};		/* pushfd */
+
+static op_implicit_list_t list_rdmsr[] = 
+	/* 0F 32 : RDMSR : r ECX, w EDX, w EAX */
+	{{ OP_R, REG_DWORD_OFFSET + 1 },
+	 { OP_W, REG_DWORD_OFFSET + 2 },
+	 { OP_W, REG_DWORD_OFFSET }, {0}};	/* rdmsr */
+
+static op_implicit_list_t list_rdpmc[] = 
+	/* 0F 33 : RDPMC : r ECX, w EDX, w EAX */
+	{{ OP_R, REG_DWORD_OFFSET + 1 },
+	 { OP_W, REG_DWORD_OFFSET + 2 },
+	 { OP_W, REG_DWORD_OFFSET }, {0}};		/* rdpmc */
+
+static op_implicit_list_t list_rdtsc[] = 
+	/* 0F 31 : RDTSC : rw EDX, rw EAX */
+	{{ OP_R | OP_W, REG_DWORD_OFFSET + 2 },
+	 { OP_R | OP_W, REG_DWORD_OFFSET }, {0}};	/* rdtsc */
+
+static op_implicit_list_t list_rep[] = 
+	/* F3, F2 ... : REP : rw ECX */
+	{{ OP_R | OP_W, REG_DWORD_OFFSET + 1 }, {0}};/* rep */
+
+static op_implicit_list_t list_rsm[] = 
+	/* 0F AA : RSM : r CR4, r CR0 */
+	{{ OP_R, REG_CTRL_OFFSET + 4 }, 
+	 { OP_R, REG_CTRL_OFFSET }, {0}};		/* rsm */
+
+static op_implicit_list_t list_sahf[] = 
+	/* 9E : SAHF : r ah, rw eflags (set SF ZF AF PF CF) */
+	{{ OP_R, REG_DWORD_OFFSET }, {0}};		/* sahf */
+
+static op_implicit_list_t list_sgdt[] = 
+	/* 0F : SGDT : r gdtr */
+	/* TODO: finish this! */
+	{{ OP_R, REG_DWORD_OFFSET }, {0}};		/* sgdt */
+
+static op_implicit_list_t list_sidt[] = 
+	/* 0F : SIDT : r idtr */
+	/* TODO: finish this! */
+	{{ OP_R, REG_DWORD_OFFSET }, {0}};		/* sidt */
+
+static op_implicit_list_t list_sldt[] = 
+	/* 0F : SLDT : r ldtr */
+	/* TODO: finish this! */
+	{{ OP_R, REG_DWORD_OFFSET }, {0}};		/* sldt */
+
+static op_implicit_list_t list_smsw[] = 
+	/* 0F : SMSW : r CR0 */
+	/* TODO: finish this! */
+	{{ OP_R, REG_DWORD_OFFSET }, {0}};		/* smsw */
+
+static op_implicit_list_t list_stmxcsr[] = 
+	/* 0F AE : STMXCSR : r MXCSR */
+	/* TODO: finish this! */
+	{{ OP_R, REG_DWORD_OFFSET }, {0}};		/* stmxcsr */
+
+static op_implicit_list_t list_str[] = 
+	/* 0F 00 : STR : r TR (task register) */
+	/* TODO: finish this! */
+	{{ OP_R, REG_DWORD_OFFSET }, {0}};		/* str */
+
+static op_implicit_list_t list_sysenter[] = 
+	/* 0F 34 : SYSENTER : w cs, w eip, w ss, w esp, r CR0, w eflags
+	 *         r sysenter_cs_msr, sysenter_esp_msr, sysenter_eip_msr */
+	/* TODO: finish this! */
+	{{ OP_R, REG_DWORD_OFFSET }, {0}};		/* sysenter */
+
+static op_implicit_list_t list_sysexit[] = 
+	/* 0F 35 : SYSEXIT : r edx, r ecx, w cs, w eip, w ss, w esp
+	 * 	   r sysenter_cs_msr */
+	/* TODO: finish this! */
+	{{ OP_R, REG_DWORD_OFFSET }, {0}};		/* sysexit */
+
+static op_implicit_list_t list_wrmsr[] = 
+	/* 0F 30 : WRMST : r edx, r eax, r ecx */
+	/* TODO: finish this! */
+	{{ OP_R, REG_DWORD_OFFSET }, {0}};		/* wrmsr */
+
+static op_implicit_list_t list_xlat[] = 
+	/* D7 : XLAT : rw al r ebx (ptr) */
+	/* TODO: finish this! */
+	{{ OP_R, REG_DWORD_OFFSET }, {0}};		/* xlat */
+/* TODO:
+ * monitor 0f 01 c8 eax OP_R ecx OP_R edx OP_R
+ * mwait 0f 01 c9 eax OP_R ecx OP_R
+ */
+static op_implicit_list_t list_monitor[] = 
+	{{ OP_R, REG_DWORD_OFFSET }, {0}};		/* monitor */
+static op_implicit_list_t list_mwait[] = 
+	{{ OP_R, REG_DWORD_OFFSET }, {0}};		/* mwait */
+
+op_implicit_list_t *op_implicit_list[] = {
+	/* This is a list of implicit operands which are read/written by
+	 * various x86 instructions. Note that modifications to the stack
+	 * register are mentioned here, but that additional information on
+	 * the effect an instruction has on the stack is contained in the
+	 * x86_insn_t 'stack_mod' and 'stack_mod_val' fields. Use of the
+	 * eflags register, i.e. setting, clearing, and testing flags, is
+	 * not recorded here but rather in the flags_set and flags_tested
+	 * fields of the x86_insn_t.*/
+	NULL,
+	list_aaa, list_aad, list_call, list_cbw,		/* 1 - 4 */
+	list_cwde, list_clts, list_cmpxchg, list_cmpxchgb,	/* 5 - 8 */
+	list_cmpxchg8b, list_cpuid, list_cwd, list_daa,		/* 9 - 12 */
+	list_idiv, list_div, list_enter, list_f2xm1,		/* 13 - 16 */
+	list_fcom, list_fpatan, list_fprem, list_faddp,		/* 17 - 20 */
+	list_fucompp, list_imul, list_mul, list_lahf,		/* 21 - 24 */
+	list_ldmxcsr, list_leave, list_lgdt, list_lidt,		/* 25 - 28 */
+	list_lldt, list_lmsw, list_loop, list_ltr,		/* 29 - 32 */
+	list_pop, list_popad, list_popfd, list_pushad,		/* 33 - 36 */
+	list_pushfd, list_rdmsr, list_rdpmc, list_rdtsc,	/* 37 - 40 */
+	/* NOTE: 'REP' is a hack since it is a prefix: if its position
+	 * in the table changes, then change IDX_IMPLICIT_REP in the .h */
+	list_rep, list_rsm, list_sahf, list_sgdt,		/* 41 - 44 */
+	list_sidt, list_sldt, list_smsw, list_stmxcsr,		/* 45 - 48 */
+	list_str, list_sysenter, list_sysexit, list_wrmsr,	/* 49 - 52 */
+	list_xlat, list_monitor, list_mwait,			/* 53 - 55*/
+	NULL						/* end of list */
+ };
+
+#define LAST_IMPL_IDX 55
+
+static void handle_impl_reg( x86_op_t *op, uint32_t val ) {
+	x86_reg_t *reg = &op->data.reg;
+	op->type = op_register;
+	ia32_handle_register( reg, (unsigned int) val );
+	switch (reg->size) {
+		case 1:
+			op->datatype = op_byte; break;
+		case 2:
+			op->datatype = op_word; break;
+		case 4:
+			op->datatype = op_dword; break;
+		case 8:
+			op->datatype = op_qword; break;
+		case 10:
+			op->datatype = op_extreal; break;
+		case 16:
+			op->datatype = op_dqword; break;
+	}
+	return;
+}
+
+/* 'impl_idx' is the value from the opcode table: between 1 and LAST_IMPL_IDX */
+/* returns number of operands added */
+unsigned int ia32_insn_implicit_ops( x86_insn_t *insn, unsigned int impl_idx ) {
+	op_implicit_list_t *list;
+	x86_op_t *op;
+	unsigned int num = 0;
+
+	if (! impl_idx || impl_idx > LAST_IMPL_IDX ) {
+		return 0;
+	}
+
+	for ( list = op_implicit_list[impl_idx]; list->type; list++, num++ ) {
+		enum x86_op_access access = (enum x86_op_access) OP_PERM(list->type);
+		enum x86_op_flags  flags  = (enum x86_op_flags) (OP_FLAGS(list->type) >> 12);
+
+		op = NULL;
+		/* In some cases (MUL), EAX is an implicit operand hardcoded in
+                 * the instruction without being explicitly listed in assembly.
+                 * For this situation, find the hardcoded operand and add the
+                 * implied flag rather than adding a new implicit operand. */
+		x86_oplist_t * existing;
+		if (ia32_true_register_id(list->operand) == REG_DWORD_OFFSET) {
+			for ( existing = insn->operands; existing; existing = existing->next ) {
+				if (existing->op.type == op_register &&
+	                            existing->op.data.reg.id == list->operand) {
+					op = &existing->op;
+					break;
+				}
+			}
+		}
+		if (!op) {
+			op = x86_operand_new( insn );
+			/* all implicit operands are registers */
+			handle_impl_reg( op, list->operand );
+			/* decrement the 'explicit count' incremented by default in
+			 * x86_operand_new */
+			insn->explicit_count = insn->explicit_count -1;
+		}
+		if (!op) {
+			return num;	/* gah! return early */
+		}
+		op->access |= access;
+		op->flags |= flags;
+		op->flags |= op_implied;
+	}
+	
+	return num;
+}

libdisasm/ia32_implicit.h

@@ -0,0 +1,13 @@
+#ifndef IA32_IMPLICIT_H
+#define IA32_IMPLICIT_H
+
+#include "libdis.h"
+
+/* OK, this is a hack to deal with prefixes having implicit operands...
+ * thought I had removed all the old hackishness ;( */
+
+#define IDX_IMPLICIT_REP 41	/* change this if the table changes! */
+
+unsigned int ia32_insn_implicit_ops( x86_insn_t *insn, unsigned int impl_idx );
+
+#endif

libdisasm/ia32_insn.c

@@ -0,0 +1,625 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include "qword.h"
+
+#include "ia32_insn.h"
+#include "ia32_opcode_tables.h"
+
+#include "ia32_reg.h"
+#include "ia32_operand.h"
+#include "ia32_implicit.h"
+#include "ia32_settings.h"
+
+#include "libdis.h"
+
+extern ia32_table_desc_t ia32_tables[];
+extern ia32_settings_t ia32_settings;
+
+#define IS_SP( op )  (op->type == op_register && 	\
+		(op->data.reg.id == REG_ESP_INDEX || 	\
+		 op->data.reg.alias == REG_ESP_INDEX) )
+#define IS_IMM( op ) (op->type == op_immediate )
+
+#ifdef WIN32
+#  define INLINE 
+#else
+#  define INLINE inline
+#endif
+
+/* for calculating stack modification based on an operand */
+static INLINE int32_t long_from_operand( x86_op_t *op ) {
+
+	if (! IS_IMM(op) ) {
+		return 0L;
+	}
+
+	switch ( op->datatype ) {
+		case op_byte:
+			return (int32_t) op->data.sbyte;
+		case op_word:
+			return (int32_t) op->data.sword;
+		case op_qword:
+			return (int32_t) op->data.sqword;
+		case op_dword:
+			return op->data.sdword;
+		default:
+			/* these are not used in stack insn */
+			break;
+	}
+
+	return 0L;
+}
+		
+
+/* determine what this insn does to the stack */
+static void ia32_stack_mod(x86_insn_t *insn) {
+	x86_op_t *dest, *src = NULL;
+
+	if (! insn || ! insn->operands ) {
+		return;
+	}
+       
+	dest = &insn->operands->op;
+	if ( dest ) {
+		src = &insn->operands->next->op;
+	}
+
+	insn->stack_mod = 0; 
+	insn->stack_mod_val = 0;
+
+	switch ( insn->type ) {
+		case insn_call:
+		case insn_callcc:
+			insn->stack_mod = 1;
+			insn->stack_mod_val = insn->addr_size * -1;
+			break;
+		case insn_push:
+			insn->stack_mod = 1;
+			insn->stack_mod_val = insn->addr_size * -1;
+			break;
+		case insn_return:
+			insn->stack_mod = 1;
+			insn->stack_mod_val = insn->addr_size;
+		case insn_int: case insn_intcc:
+		case insn_iret:
+			break;
+		case insn_pop:
+			insn->stack_mod = 1;
+			if (! IS_SP( dest ) ) {
+				insn->stack_mod_val = insn->op_size;
+			} /* else we don't know the stack change in a pop esp */
+			break;
+		case insn_enter:
+			insn->stack_mod = 1;
+			insn->stack_mod_val = 0; /* TODO : FIX */
+			break;
+		case insn_leave:
+			insn->stack_mod = 1;
+			insn->stack_mod_val = 0; /* TODO : FIX */
+			break;
+		case insn_pushregs:
+			insn->stack_mod = 1;
+			insn->stack_mod_val = 0; /* TODO : FIX */
+			break;
+		case insn_popregs:
+			insn->stack_mod = 1;
+			insn->stack_mod_val = 0; /* TODO : FIX */
+			break;
+		case insn_pushflags:
+			insn->stack_mod = 1;
+			insn->stack_mod_val = 0; /* TODO : FIX */
+			break;
+		case insn_popflags:
+			insn->stack_mod = 1;
+			insn->stack_mod_val = 0; /* TODO : FIX */
+			break;
+		case insn_add:
+			if ( IS_SP( dest ) ) {
+				insn->stack_mod = 1;
+				insn->stack_mod_val = long_from_operand( src ); 
+			}
+			break;
+		case insn_sub:
+			if ( IS_SP( dest ) ) {
+				insn->stack_mod = 1;
+				insn->stack_mod_val = long_from_operand( src ); 
+				insn->stack_mod_val *= -1;
+			}
+			break;
+		case insn_inc:
+			if ( IS_SP( dest ) ) {
+				insn->stack_mod = 1;
+				insn->stack_mod_val = 1;
+			}
+			break;
+		case insn_dec:
+			if ( IS_SP( dest ) ) {
+				insn->stack_mod = 1;
+				insn->stack_mod_val = 1;
+			}
+			break;
+		case insn_mov: case insn_movcc:
+		case insn_xchg: case insn_xchgcc:
+		case insn_mul: case insn_div:
+		case insn_shl: case insn_shr:
+		case insn_rol: case insn_ror:
+		case insn_and: case insn_or:
+		case insn_not: case insn_neg:
+		case insn_xor:
+			if ( IS_SP( dest ) ) {
+				insn->stack_mod = 1;
+			}
+			break;
+		default:
+			break;
+	}
+	if (! strcmp("enter", insn->mnemonic) ) {
+		insn->stack_mod = 1;
+	} else if (! strcmp("leave", insn->mnemonic) ) {
+		insn->stack_mod = 1;
+	}
+
+	/* for mov, etc we return 0 -- unknown stack mod */
+
+	return;
+}
+
+/* get the cpu details for this insn from cpu flags int */
+static void ia32_handle_cpu( x86_insn_t *insn, unsigned int cpu ) {
+	insn->cpu = (enum x86_insn_cpu) CPU_MODEL(cpu);
+	insn->isa = (enum x86_insn_isa) (ISA_SUBSET(cpu)) >> 16;
+	return;
+}
+
+/* handle mnemonic type and group */
+static void ia32_handle_mnemtype(x86_insn_t *insn, unsigned int mnemtype) {
+	unsigned int type = mnemtype & ~INS_FLAG_MASK;
+        insn->group = (enum x86_insn_group) (INS_GROUP(type)) >> 12;
+        insn->type = (enum x86_insn_type) INS_TYPE(type);
+
+	return;
+}
+
+static void ia32_handle_notes(x86_insn_t *insn, unsigned int notes) {
+	insn->note = (enum x86_insn_note) notes;
+	return;
+}
+
+static void ia32_handle_eflags( x86_insn_t *insn, unsigned int eflags) {
+        unsigned int flags;
+
+        /* handle flags effected */
+        flags = INS_FLAGS_TEST(eflags);
+        /* handle weird OR cases */
+        /* these are either JLE (ZF | SF<>OF) or JBE (CF | ZF) */
+        if (flags & INS_TEST_OR) {
+                flags &= ~INS_TEST_OR;
+                if ( flags & INS_TEST_ZERO ) {
+                        flags &= ~INS_TEST_ZERO;
+                        if ( flags & INS_TEST_CARRY ) {
+                                flags &= ~INS_TEST_CARRY ;
+                                flags |= (int)insn_carry_or_zero_set;
+                        } else if ( flags & INS_TEST_SFNEOF ) {
+                                flags &= ~INS_TEST_SFNEOF;
+                                flags |= (int)insn_zero_set_or_sign_ne_oflow;
+                        }
+                }
+        }
+        insn->flags_tested = (enum x86_flag_status) flags;
+
+        insn->flags_set = (enum x86_flag_status) INS_FLAGS_SET(eflags) >> 16;
+
+	return;
+}
+
+static void ia32_handle_prefix( x86_insn_t *insn, unsigned int prefixes ) {
+
+        insn->prefix = (enum x86_insn_prefix) prefixes & PREFIX_MASK; // >> 20;
+        if (! (insn->prefix & PREFIX_PRINT_MASK) ) {
+		/* no printable prefixes */
+                insn->prefix = insn_no_prefix;
+        }
+
+        /* concat all prefix strings */
+        if ( (unsigned int)insn->prefix & PREFIX_LOCK ) {
+                strncat(insn->prefix_string, "lock ", 32 - 
+				strlen(insn->prefix_string));
+        }
+
+        if ( (unsigned int)insn->prefix & PREFIX_REPNZ ) {
+                strncat(insn->prefix_string, "repnz ", 32  - 
+				strlen(insn->prefix_string));
+        } else if ( (unsigned int)insn->prefix & PREFIX_REPZ ) {
+                strncat(insn->prefix_string, "repz ", 32 - 
+				strlen(insn->prefix_string));
+        }
+
+        return;
+}
+
+
+static void reg_32_to_16( x86_op_t *op, x86_insn_t *insn, void *arg ) {
+
+	/* if this is a 32-bit register and it is a general register ... */
+	if ( op->type == op_register && op->data.reg.size == 4 && 
+	     (op->data.reg.type & reg_gen) ) {
+		/* WORD registers are 8 indices off from DWORD registers */
+		ia32_handle_register( &(op->data.reg), 
+				op->data.reg.id + 8 );
+	}
+}
+
+static void handle_insn_metadata( x86_insn_t *insn, ia32_insn_t *raw_insn ) {
+	ia32_handle_mnemtype( insn, raw_insn->mnem_flag );
+	ia32_handle_notes( insn, raw_insn->notes );
+	ia32_handle_eflags( insn, raw_insn->flags_effected );
+	ia32_handle_cpu( insn, raw_insn->cpu );
+	ia32_stack_mod( insn );
+}
+
+static size_t ia32_decode_insn( unsigned char *buf, size_t buf_len, 
+			   ia32_insn_t *raw_insn, x86_insn_t *insn,
+			   unsigned int prefixes ) {
+	size_t size, op_size;
+	unsigned char modrm;
+
+	/* this should never happen, but just in case... */
+	if ( raw_insn->mnem_flag == INS_INVALID ) {
+		return 0;
+	}
+
+	if (ia32_settings.options & opt_16_bit) {
+		insn->op_size = ( prefixes & PREFIX_OP_SIZE ) ? 4 : 2;
+		insn->addr_size = ( prefixes & PREFIX_ADDR_SIZE ) ? 4 : 2;
+	} else {
+		insn->op_size = ( prefixes & PREFIX_OP_SIZE ) ? 2 : 4;
+		insn->addr_size = ( prefixes & PREFIX_ADDR_SIZE ) ? 2 : 4;
+	}
+
+
+	/*  ++++   1. Copy mnemonic and mnemonic-flags to CODE struct */
+	if ((ia32_settings.options & opt_att_mnemonics) && raw_insn->mnemonic_att[0]) {
+		strncpy( insn->mnemonic, raw_insn->mnemonic_att, 16 );
+	}
+	else {
+		strncpy( insn->mnemonic, raw_insn->mnemonic, 16 );
+	}
+	ia32_handle_prefix( insn, prefixes );
+
+	handle_insn_metadata( insn, raw_insn );
+
+	/* prefetch the next byte in case it is a modr/m byte -- saves
+	 * worrying about whether the 'mod/rm' operand or the 'reg' operand
+	 * occurs first */
+	modrm = GET_BYTE( buf, buf_len );
+
+	/*  ++++   2. Decode Explicit Operands */
+	/* Intel uses up to 3 explicit operands in its instructions;
+	 * the first is 'dest', the second is 'src', and the third
+	 * is an additional source value (usually an immediate value,
+	 * e.g. in the MUL instructions). These three explicit operands
+	 * are encoded in the opcode tables, even if they are not used 
+	 * by the instruction. Additional implicit operands are stored
+	 * in a supplemental table and are handled later. */
+
+	op_size = ia32_decode_operand( buf, buf_len, insn, raw_insn->dest, 
+					raw_insn->dest_flag, prefixes, modrm );
+	/* advance buffer, increase size if necessary */
+	buf += op_size;
+	buf_len -= op_size;
+	size = op_size;
+
+	op_size = ia32_decode_operand( buf, buf_len, insn, raw_insn->src, 
+					raw_insn->src_flag, prefixes, modrm );
+	buf += op_size;
+	buf_len -= op_size;
+	size += op_size;
+
+	op_size = ia32_decode_operand( buf, buf_len, insn, raw_insn->aux, 
+					raw_insn->aux_flag, prefixes, modrm );
+	size += op_size;
+
+
+	/*  ++++   3. Decode Implicit Operands */
+	/* apply implicit operands */
+	ia32_insn_implicit_ops( insn, raw_insn->implicit_ops );
+	/* we have one small inelegant hack here, to deal with 
+	 * the two prefixes that have implicit operands. If Intel
+	 * adds more, we'll change the algorithm to suit :) */
+	if ( (prefixes & PREFIX_REPZ) || (prefixes & PREFIX_REPNZ) ) {
+		ia32_insn_implicit_ops( insn, IDX_IMPLICIT_REP );
+	}
+
+
+	/* 16-bit hack: foreach operand, if 32-bit reg, make 16-bit reg */
+	if ( insn->op_size == 2 ) {
+		x86_operand_foreach( insn, reg_32_to_16, NULL, op_any );
+	}
+
+	return size;
+}
+
+
+/* convenience routine */
+#define USES_MOD_RM(flag) \
+	(flag == ADDRMETH_E || flag == ADDRMETH_M || flag == ADDRMETH_Q || \
+	 flag == ADDRMETH_W || flag == ADDRMETH_R)
+
+static int uses_modrm_flag( unsigned int flag ) {
+	unsigned int meth;
+	if ( flag == ARG_NONE ) {
+		return 0;
+	}
+	meth = (flag & ADDRMETH_MASK);
+	if ( USES_MOD_RM(meth) ) {
+		return 1;
+	}
+
+	return 0;
+}
+
+/* This routine performs the actual byte-by-byte opcode table lookup.
+ * Originally it was pretty simple: get a byte, adjust it to a proper
+ * index into the table, then check the table row at that index to
+ * determine what to do next. But is anything that simple with Intel?
+ * This is now a huge, convoluted mess, mostly of bitter comments. */
+/* buf: pointer to next byte to read from stream 
+ * buf_len: length of buf
+ * table: index of table to use for lookups
+ * raw_insn: output pointer that receives opcode definition
+ * prefixes: output integer that is encoded with prefixes in insn 
+ * returns : number of bytes consumed from stream during lookup */ 
+size_t ia32_table_lookup( unsigned char *buf, size_t buf_len,
+				 unsigned int table, ia32_insn_t **raw_insn,
+				 unsigned int *prefixes ) {
+	unsigned char *next, op = buf[0];	/* byte value -- 'opcode' */
+	size_t size = 1, sub_size = 0, next_len;
+	ia32_table_desc_t *table_desc;
+	unsigned int subtable, prefix = 0, recurse_table = 0;
+
+	table_desc = &ia32_tables[table];
+
+	op = GET_BYTE( buf, buf_len );
+
+	if ( table_desc->type == tbl_fpu && op > table_desc->maxlim) {
+		/* one of the fucking FPU tables out of the 00-BH range */
+		/* OK,. this is a bit of a hack -- the proper way would
+		 * have been to use subtables in the 00-BF FPU opcode tables,
+		 * but that is rather wasteful of space... */
+		table_desc = &ia32_tables[table +1];
+	}
+
+	/* PERFORM TABLE LOOKUP */
+
+	/* ModR/M trick: shift extension bits into lowest bits of byte */
+	/* Note: non-ModR/M tables have a shift value of 0 */
+	op >>= table_desc->shift;
+
+	/* ModR/M trick: mask out high bits to turn extension into an index */
+	/* Note: non-ModR/M tables have a mask value of 0xFF */
+	op &= table_desc->mask;
+
+
+	/* Sparse table trick: check that byte is <= max value */
+	/* Note: full (256-entry) tables have a maxlim of 155 */
+	if ( op > table_desc->maxlim ) {
+		/* this is a partial table, truncated at the tail,
+		   and op is out of range! */
+		return INVALID_INSN;
+	}
+
+	/* Sparse table trick: check that byte is >= min value */
+	/* Note: full (256-entry) tables have a minlim of 0 */
+	if ( table_desc->minlim > op ) {
+		/* this is a partial table, truncated at the head,
+		   and op is out of range! */
+		return INVALID_INSN;
+	}
+	/* adjust op to be an offset from table index 0 */
+	op -= table_desc->minlim;
+
+	/* Yay! 'op' is now fully adjusted to be an index into 'table' */
+	*raw_insn = &(table_desc->table[op]);
+	//printf("BYTE %X TABLE %d OP %X\n", buf[0], table, op ); 
+
+	if ( (*raw_insn)->mnem_flag & INS_FLAG_PREFIX ) {
+		prefix = (*raw_insn)->mnem_flag & PREFIX_MASK;
+	}
+
+
+	/* handle escape to a multibyte/coproc/extension/etc table */
+	/* NOTE: if insn is a prefix and has a subtable, then we
+	 *       only recurse if this is the first prefix byte --
+	 *       that is, if *prefixes is 0. 
+	 * NOTE also that suffix tables are handled later */
+	subtable = (*raw_insn)->table;
+
+	if ( subtable && ia32_tables[subtable].type != tbl_suffix &&
+	     (! prefix || ! *prefixes) ) {
+
+	     	if ( ia32_tables[subtable].type == tbl_ext_ext ||
+	     	     ia32_tables[subtable].type == tbl_fpu_ext ) {
+			/* opcode extension: reuse current byte in buffer */
+			next = buf;
+			next_len = buf_len;
+		} else {
+			/* "normal" opcode: advance to next byte in buffer */
+			if ( buf_len > 1 ) {
+				next = &buf[1];
+				next_len = buf_len - 1;
+			}
+			else {
+				// buffer is truncated 
+				return INVALID_INSN;
+			}
+		}
+		/* we encountered a multibyte opcode: recurse using the
+		 * table specified in the opcode definition */
+		sub_size = ia32_table_lookup( next, next_len, subtable, 
+				raw_insn, prefixes );
+
+		/* SSE/prefix hack: if the original opcode def was a 
+		 * prefix that specified a subtable, and the subtable
+		 * lookup returned a valid insn, then we have encountered
+		 * an SSE opcode definition; otherwise, we pretend we
+		 * never did the subtable lookup, and deal with the 
+		 * prefix normally later */
+		if ( prefix && ( sub_size == INVALID_INSN  ||
+		       INS_TYPE((*raw_insn)->mnem_flag) == INS_INVALID ) ) {
+			/* this is a prefix, not an SSE insn :
+			 * lookup next byte in main table,
+			 * subsize will be reset during the
+			 * main table lookup */
+			recurse_table = 1;
+		} else {
+			/* this is either a subtable (two-byte) insn
+			 * or an invalid insn: either way, set prefix
+			 * to NULL and end the opcode lookup */
+			prefix = 0;
+			// short-circuit lookup on invalid insn
+			if (sub_size == INVALID_INSN) return INVALID_INSN;
+		}
+	} else if ( prefix ) {
+		recurse_table = 1;
+	}
+
+	/* by default, we assume that we have the opcode definition,
+	 * and there is no need to recurse on the same table, but
+	 * if we do then a prefix was encountered... */
+	if ( recurse_table ) {
+		/* this must have been a prefix: use the same table for
+		 * lookup of the next byte */
+		sub_size = ia32_table_lookup( &buf[1], buf_len - 1, table, 
+				raw_insn, prefixes );
+
+		// short-circuit lookup on invalid insn
+		if (sub_size == INVALID_INSN) return INVALID_INSN;
+
+		/* a bit of a hack for branch hints */
+		if ( prefix & BRANCH_HINT_MASK ) {
+			if ( INS_GROUP((*raw_insn)->mnem_flag) == INS_EXEC ) {
+				/* segment override prefixes are invalid for
+			 	* all branch instructions, so delete them */
+				prefix &= ~PREFIX_REG_MASK;
+			} else {
+				prefix &= ~BRANCH_HINT_MASK;
+			}
+		}
+
+		/* apply prefix to instruction */
+
+		/* TODO: implement something enforcing prefix groups */
+		(*prefixes) |= prefix;
+	}
+
+	/* if this lookup was in a ModR/M table, then an opcode byte is 
+	 * NOT consumed: subtract accordingly. NOTE that if none of the
+	 * operands used the ModR/M, then we need to consume the byte
+	 * here, but ONLY in the 'top-level' opcode extension table */
+
+	if ( table_desc->type == tbl_ext_ext ) {
+		/* extensions-to-extensions never consume a byte */
+		--size;
+	} else if ( (table_desc->type == tbl_extension || 
+	       	     table_desc->type == tbl_fpu ||
+		     table_desc->type == tbl_fpu_ext ) && 
+		/* extensions that have an operand encoded in ModR/M
+		 * never consume a byte */
+	      	    (uses_modrm_flag((*raw_insn)->dest_flag) || 
+	             uses_modrm_flag((*raw_insn)->src_flag) )  	) {
+		--size;
+	}
+
+	size += sub_size;
+
+	return size;
+}
+
+static size_t handle_insn_suffix( unsigned char *buf, size_t buf_len,
+			   ia32_insn_t *raw_insn, x86_insn_t * insn ) {
+	ia32_table_desc_t *table_desc;
+	ia32_insn_t *sfx_insn;
+	size_t size;
+	unsigned int prefixes = 0;
+
+	table_desc = &ia32_tables[raw_insn->table]; 
+	size = ia32_table_lookup( buf, buf_len, raw_insn->table, &sfx_insn,
+				 &prefixes );
+	if (size == INVALID_INSN || sfx_insn->mnem_flag == INS_INVALID ) {
+		return 0;
+	}
+
+	strncpy( insn->mnemonic, sfx_insn->mnemonic, 16 );
+	handle_insn_metadata( insn, sfx_insn );
+
+	return 1;
+}
+
+/* invalid instructions are handled by returning 0 [error] from the
+ * function, setting the size of the insn to 1 byte, and copying
+ * the byte at the start of the invalid insn into the x86_insn_t.
+ * if the caller is saving the x86_insn_t for invalid instructions,
+ * instead of discarding them, this will maintain a consistent
+ * address space in the x86_insn_ts */
+
+/* this function is called by the controlling disassembler, so its name and
+ * calling convention cannot be changed */
+/*    buf   points to the loc of the current opcode (start of the 
+ *          instruction) in the instruction stream. The instruction 
+ *          stream is assumed to be a buffer of bytes read directly 
+ *          from the file for the purpose of disassembly; a mem-mapped 
+ *          file is ideal for *        this.
+ *    insn points to a code structure to be filled by instr_decode
+ *    returns the size of the decoded instruction in bytes */
+size_t ia32_disasm_addr( unsigned char * buf, size_t buf_len, 
+		x86_insn_t *insn ) {
+	ia32_insn_t *raw_insn = NULL;
+	unsigned int prefixes = 0;
+	size_t size, sfx_size;
+	
+	if ( (ia32_settings.options & opt_ignore_nulls) && buf_len > 3 &&
+	    !buf[0] && !buf[1] && !buf[2] && !buf[3]) {
+		/* IF IGNORE_NULLS is set AND
+		 * first 4 bytes in the intruction stream are NULL
+		 * THEN return 0 (END_OF_DISASSEMBLY) */
+		/* TODO: set errno */
+		MAKE_INVALID( insn, buf );
+		return 0;	/* 4 00 bytes in a row? This isn't code! */
+	}
+
+	/* Perform recursive table lookup starting with main table (0) */
+	size = ia32_table_lookup(buf, buf_len, idx_Main, &raw_insn, &prefixes);
+	if ( size == INVALID_INSN || size > buf_len || raw_insn->mnem_flag == INS_INVALID ) {
+		MAKE_INVALID( insn, buf );
+		/* TODO: set errno */
+		return 0;
+	}
+
+	/* We now have the opcode itself figured out: we can decode
+	 * the rest of the instruction. */
+	size += ia32_decode_insn( &buf[size], buf_len - size, raw_insn, insn, 
+				  prefixes );
+	if ( raw_insn->mnem_flag & INS_FLAG_SUFFIX ) {
+		/* AMD 3DNow! suffix -- get proper operand type here */
+		sfx_size = handle_insn_suffix( &buf[size], buf_len - size,
+				raw_insn, insn );
+		if (! sfx_size ) {
+			/* TODO: set errno */
+			MAKE_INVALID( insn, buf );
+			return 0;
+		}
+
+		size += sfx_size;
+	}
+
+	if (! size ) {
+		/* invalid insn */
+		MAKE_INVALID( insn, buf );
+		return 0;
+	}
+
+
+	insn->size = size;
+	return size;		/* return size of instruction in bytes */
+}

libdisasm/ia32_insn.h

@@ -0,0 +1,506 @@
+#ifndef IA32_INSN_H
+#define IA32_INSN_H
+/* this file contains the structure of opcode definitions and the
+ * constants they use */
+
+#include <sys/types.h>
+#include "libdis.h"
+
+
+#define GET_BYTE( buf, buf_len ) buf_len ? *buf : 0
+
+#define OP_SIZE_16	1
+#define OP_SIZE_32	2
+#define ADDR_SIZE_16	4
+#define ADDR_SIZE_32	8
+
+#define MAX_INSTRUCTION_SIZE 20
+
+/* invalid instructions are handled by returning 0 [error] from the
+ * function, setting the size of the insn to 1 byte, and copying
+ * the byte at the start of the invalid insn into the x86_insn_t.
+ * if the caller is saving the x86_insn_t for invalid instructions,
+ * instead of discarding them, this will maintain a consistent
+ * address space in the x86_insn_ts */
+
+#define INVALID_INSN ((size_t) -1)	/* return value for invalid insn */
+#define MAKE_INVALID( i, buf )                          \
+                strcpy( i->mnemonic, "invalid" );       \
+                x86_oplist_free( i );                   \
+                i->size = 1;                            \
+                i->group = insn_none;                   \
+                i->type = insn_invalid;                 \
+                memcpy( i->bytes, buf, 1 );
+
+
+size_t ia32_disasm_addr( unsigned char * buf, size_t buf_len, 
+		x86_insn_t *insn);
+
+
+/* --------------------------------------------------------- Table Lookup */
+/* IA32 Instruction defintion for ia32_opcodes.c */
+typedef struct {
+   unsigned int table;          /* escape to this sub-table */
+   unsigned int mnem_flag;      /* Flags referring to mnemonic */
+   unsigned int notes;          /* Notes for this instruction */
+   unsigned int dest_flag, src_flag, aux_flag; /* and for specific operands */
+   unsigned int cpu;            /* minimumCPU [AND with clocks?? */
+   char mnemonic[16];           /* buffers for building instruction */
+   char mnemonic_att[16];       /* at&t style mnemonic name */
+   int32_t dest;
+   int32_t src;
+   int32_t aux;
+   unsigned int flags_effected;
+   unsigned int implicit_ops;	/* implicit operands */
+} ia32_insn_t;
+
+
+
+/* --------------------------------------------------------- Prefixes */
+/* Prefix Flags */
+/* Prefixes, same order as in the manual */
+/* had to reverse the values of the first three as they were entered into
+ * libdis.h incorrectly. */
+#define PREFIX_LOCK       0x0004
+#define PREFIX_REPNZ      0x0002
+#define PREFIX_REPZ       0x0001
+#define PREFIX_OP_SIZE    0x0010
+#define PREFIX_ADDR_SIZE  0x0020
+#define PREFIX_CS         0x0100
+#define PREFIX_SS         0x0200
+#define PREFIX_DS         0x0300
+#define PREFIX_ES         0x0400
+#define PREFIX_FS         0x0500
+#define PREFIX_GS         0x0600
+#define PREFIX_TAKEN      0x1000	/* branch taken */
+#define PREFIX_NOTTAKEN   0x2000	/* branch not taken */
+#define PREFIX_REG_MASK   0x0F00
+#define BRANCH_HINT_MASK  0x3000 
+#define PREFIX_PRINT_MASK 0x000F	/* printable prefixes */
+#define PREFIX_MASK       0xFFFF
+
+/* ---------------------------------------------------------- CPU Type */
+
+#define cpu_8086         0x0001
+#define cpu_80286        0x0002
+#define cpu_80386        0x0003
+#define cpu_80387        0x0004 /* originally these were a co-proc */
+#define cpu_80486        0x0005
+#define cpu_PENTIUM      0x0006
+#define cpu_PENTPRO      0x0007
+#define cpu_PENTIUM2     0x0008
+#define cpu_PENTIUM3     0x0009
+#define cpu_PENTIUM4     0x000A
+#define cpu_K6		 0x0010
+#define cpu_K7		 0x0020
+#define cpu_ATHLON	 0x0030
+#define CPU_MODEL_MASK	 0xFFFF
+#define CPU_MODEL(cpu)	 (cpu & CPU_MODEL_MASK)
+/* intel instruction subsets */
+#define isa_GP		 0x10000	/* General Purpose Instructions */
+#define isa_FPU		 0x20000	/* FPU instructions */
+#define isa_FPUMGT	 0x30000	/* FPU/SIMD Management */
+#define isa_MMX		 0x40000	/* MMX */
+#define isa_SSE1	 0x50000	/* SSE */
+#define isa_SSE2	 0x60000	/* SSE 2 */
+#define isa_SSE3	 0x70000	/* SSE 3 */
+#define isa_3DNOW	 0x80000	/* AMD 3d Now */
+#define isa_SYS		 0x90000	/* System Instructions */
+#define ISA_SUBSET_MASK	 0xFFFF0000
+#define ISA_SUBSET(isa)	(isa & ISA_SUBSET_MASK)
+
+
+/* ------------------------------------------------------ Operand Decoding */
+#define ARG_NONE         0
+
+/* Using a mask allows us to store info such as OP_SIGNED in the
+ * operand flags field */
+#define   OPFLAGS_MASK 	0x0000FFFF
+
+/* Operand Addressing Methods, per intel manual */
+#define   ADDRMETH_MASK	0x00FF0000
+
+/* note: for instructions with implied operands, use no ADDRMETH */
+#define   ADDRMETH_A  	0x00010000   
+#define   ADDRMETH_C   	0x00020000
+#define   ADDRMETH_D   	0x00030000
+#define   ADDRMETH_E   	0x00040000
+#define   ADDRMETH_F   	0x00050000
+#define   ADDRMETH_G   	0x00060000
+#define   ADDRMETH_I   	0x00070000
+#define   ADDRMETH_J   	0x00080000
+#define   ADDRMETH_M   	0x00090000
+#define   ADDRMETH_O   	0x000A0000
+#define   ADDRMETH_P   	0x000B0000
+#define   ADDRMETH_Q   	0x000C0000
+#define   ADDRMETH_R   	0x000D0000
+#define   ADDRMETH_S   	0x000E0000
+#define   ADDRMETH_T   	0x000F0000
+#define   ADDRMETH_V   	0x00100000
+#define   ADDRMETH_W   	0x00110000
+#define   ADDRMETH_X   	0x00120000
+#define   ADDRMETH_Y   	0x00130000
+#define	  ADDRMETH_RR  	0x00140000	/* gen reg hard-coded in opcode */
+#define	  ADDRMETH_RS  	0x00150000	/* seg reg hard-coded in opcode */
+#define	  ADDRMETH_RT  	0x00160000	/* test reg hard-coded in opcode */
+#define	  ADDRMETH_RF  	0x00170000	/* fpu reg hard-coded in opcode */
+#define	  ADDRMETH_II  	0x00180000	/* immediate hard-coded in opcode */
+#define   ADDRMETH_PP   0x00190000	/* mm reg ONLY in modr/m field */
+#define   ADDRMETH_VV   0x001A0000	/* xmm reg ONLY in mod/rm field */
+
+/* Operand Types, per intel manual */
+#define OPTYPE_MASK	0xFF000000
+
+#define OPTYPE_a	0x01000000 /* BOUND: h:h or w:w */
+#define OPTYPE_b   	0x02000000 /* byte */
+#define OPTYPE_c   	0x03000000 /* byte or word */
+#define OPTYPE_d   	0x04000000 /* word */
+#define OPTYPE_dq   	0x05000000 /* qword */
+#define OPTYPE_p   	0x06000000 /* 16:16 or 16:32 pointer */
+#define OPTYPE_pi   	0x07000000 /* dword MMX reg */
+#define OPTYPE_ps   	0x08000000 /* 128-bit single fp */
+#define OPTYPE_q   	0x09000000 /* dword */
+#define OPTYPE_s   	0x0A000000 /* 6-byte descriptor */
+#define OPTYPE_ss   	0x0B000000 /* scalar of 128-bit single fp */
+#define OPTYPE_si   	0x0C000000 /* word general register */
+#define OPTYPE_v   	0x0D000000 /* hword or word */
+#define OPTYPE_w   	0x0E000000 /* hword */
+#define OPTYPE_m   	0x0F000000	/* to handle LEA */
+#define OPTYPE_none 0xFF000000 /* no valid operand size, INVLPG */
+
+/* custom ones for FPU instructions */
+#define OPTYPE_fs	0x10000000	/* pointer to single-real*/
+#define OPTYPE_fd	0x20000000	/* pointer to double real */
+#define OPTYPE_fe	0x30000000	/* pointer to extended real */
+#define OPTYPE_fb	0x40000000	/* pointer to packed BCD */
+#define OPTYPE_fv	0x50000000	/* pointer to FPU env: 14|28-bytes */
+#define OPTYPE_ft	0x60000000	/* pointer to FPU state: 94|108-bytes */
+#define OPTYPE_fx       0x70000000      /* pointer to FPU regs: 512 bites */
+#define OPTYPE_fp       0x80000000      /* general fpu register: dbl ext */
+
+/* SSE2 operand types */
+#define OPTYPE_sd	0x90000000	/* scalar of 128-bit double fp */
+#define OPTYPE_pd	0xA0000000	/* 128-bit double fp */
+
+
+
+/* ---------------------------------------------- Opcode Table Descriptions */
+/* the table type describes how to handle byte/size increments before 
+ * and after lookup. Some tables re-use the current byte, others
+ * consume a byte only if the ModR/M encodes no operands, etc */
+enum ia32_tbl_type_id {
+	tbl_opcode = 0,	/* standard opcode table: no surprises */
+	tbl_prefix,	/* Prefix Override, e.g. 66/F2/F3 */
+	tbl_suffix,	/* 3D Now style */
+	tbl_extension,	/* ModR/M extension: 00-FF -> 00-07 */
+	tbl_ext_ext,	/* extension of modr/m using R/M field */
+	tbl_fpu,	/* fpu table: 00-BF -> 00-0F */
+	tbl_fpu_ext	/* fpu extension : C0-FF -> 00-1F */
+ };
+
+/* How it works:
+ * Bytes are 'consumed' if the next table lookup requires that the byte
+ * pointer be advanced in the instruction stream. 'Does not consume' means
+ * that, when the lookup function recurses, the same byte it re-used in the
+ * new table. It also means that size is not decremented, for example when
+ * a ModR/M byte is used. Note that tbl_extension (ModR/M) instructions that
+ * do not increase the size of an insn with their operands have a forced
+ 3 size increase in the lookup algo. Weird, yes, confusing, yes, welcome
+ * to the Intel ISA. Another note: tbl_prefix is used as an override, so an
+ * empty insn in a prefix table causes the instruction in the original table
+ * to be used, rather than an invalid insn being generated.
+ * 	tbl_opcode uses current byte and consumes it
+ * 	tbl_prefix uses current byte but does not consume it
+ * 	tbl_suffix uses and consumes last byte in insn
+ * 	tbl_extension uses current byte but does not consume it
+ * 	tbl_ext_ext uses current byte but does not consume it
+ * 	tbl_fpu uses current byte and consumes it
+ * 	tbl_fpu_ext uses current byte but does not consume it 
+ */
+
+/* Convenience struct for opcode tables : these will be stored in a 
+ * 'table of tables' so we can use a table index instead of a pointer */
+typedef struct {		/* Assembly instruction tables */
+   ia32_insn_t *table;		/* Pointer to table of instruction encodings */
+   enum ia32_tbl_type_id type;
+   unsigned char shift;		/* amount to shift modrm byte */
+   unsigned char mask;		/* bit mask for look up */
+   unsigned char minlim,maxlim;	/* limits on min/max entries. */
+} ia32_table_desc_t;
+
+
+/* ---------------------------------------------- 'Cooked' Operand Type Info */
+/*                   Permissions: */
+#define OP_R         0x001      /* operand is READ */
+#define OP_W         0x002      /* operand is WRITTEN */
+#define OP_RW        0x003	/* (OP_R|OP_W): convenience macro */
+#define OP_X         0x004      /* operand is EXECUTED */
+
+#define OP_PERM_MASK 0x0000007  /* perms are NOT mutually exclusive */
+#define OP_PERM( type )       (type & OP_PERM_MASK)
+
+/* Flags */
+#define OP_SIGNED    0x010   	/* operand is signed */
+
+#define OP_FLAG_MASK  0x0F0  /* mods are NOT mutually exclusive */
+#define OP_FLAGS( type )        (type & OP_FLAG_MASK)
+
+#define OP_REG_MASK    0x0000FFFF /* lower WORD is register ID */
+#define OP_REGTBL_MASK 0xFFFF0000 /* higher word is register type [gen/dbg] */
+#define OP_REGID( type )      (type & OP_REG_MASK)
+#define OP_REGTYPE( type )    (type & OP_REGTBL_MASK)
+
+/* ------------------------------------------'Cooked' Instruction Type Info */
+/* high-bit opcode types/insn meta-types */
+#define INS_FLAG_PREFIX		0x10000000	/* insn is a prefix */
+#define INS_FLAG_SUFFIX		0x20000000	/* followed by a suffix byte */
+#define INS_FLAG_MASK    	0xFF000000
+
+/* insn notes */
+#define INS_NOTE_RING0		0x00000001	/* insn is privileged */
+#define INS_NOTE_SMM		0x00000002	/* Sys Mgt Mode only */
+#define INS_NOTE_SERIAL		0x00000004	/* serializes */
+#define INS_NOTE_NONSWAP    0x00000008  /* insn is not swapped in att format */ // could be separate field?
+#define INS_NOTE_NOSUFFIX   0x00000010  /* insn has no size suffix in att format */ // could be separate field?
+//#define INS_NOTE_NMI		
+
+#define INS_INVALID 	0
+
+/* instruction groups */
+#define INS_EXEC	0x1000
+#define INS_ARITH	0x2000
+#define INS_LOGIC	0x3000
+#define INS_STACK	0x4000
+#define INS_COND	0x5000
+#define INS_LOAD	0x6000
+#define INS_ARRAY	0x7000
+#define INS_BIT		0x8000
+#define INS_FLAG	0x9000
+#define INS_FPU		0xA000
+#define INS_TRAPS	0xD000
+#define INS_SYSTEM	0xE000
+#define INS_OTHER	0xF000
+
+#define INS_GROUP_MASK	0xF000
+#define INS_GROUP( type )     ( type & INS_GROUP_MASK )
+
+/* INS_EXEC group */
+#define INS_BRANCH	(INS_EXEC | 0x01)	/* Unconditional branch */
+#define INS_BRANCHCC	(INS_EXEC | 0x02)	/* Conditional branch */
+#define INS_CALL	(INS_EXEC | 0x03)	/* Jump to subroutine */
+#define INS_CALLCC	(INS_EXEC | 0x04)	/* Jump to subroutine */
+#define INS_RET		(INS_EXEC | 0x05)	/* Return from subroutine */
+
+/* INS_ARITH group */
+#define INS_ADD 	(INS_ARITH | 0x01)
+#define INS_SUB		(INS_ARITH | 0x02)
+#define INS_MUL		(INS_ARITH | 0x03)
+#define INS_DIV		(INS_ARITH | 0x04)
+#define INS_INC		(INS_ARITH | 0x05)	/* increment */
+#define INS_DEC		(INS_ARITH | 0x06)	/* decrement */
+#define INS_SHL		(INS_ARITH | 0x07)	/* shift right */
+#define INS_SHR		(INS_ARITH | 0x08)	/* shift left */
+#define INS_ROL		(INS_ARITH | 0x09)	/* rotate left */
+#define INS_ROR		(INS_ARITH | 0x0A)	/* rotate right */
+#define INS_MIN		(INS_ARITH | 0x0B)	/* min func */
+#define INS_MAX		(INS_ARITH | 0x0C)	/* max func */
+#define INS_AVG		(INS_ARITH | 0x0D)	/* avg func */
+#define INS_FLR		(INS_ARITH | 0x0E)	/* floor func */
+#define INS_CEIL	(INS_ARITH | 0x0F)	/* ceiling func */
+
+/* INS_LOGIC group */
+#define INS_AND		(INS_LOGIC | 0x01)
+#define INS_OR		(INS_LOGIC | 0x02)
+#define INS_XOR		(INS_LOGIC | 0x03)
+#define INS_NOT		(INS_LOGIC | 0x04)
+#define INS_NEG		(INS_LOGIC | 0x05)
+#define INS_NAND	(INS_LOGIC | 0x06)
+
+/* INS_STACK group */
+#define INS_PUSH	(INS_STACK | 0x01)
+#define INS_POP		(INS_STACK | 0x02)
+#define INS_PUSHREGS	(INS_STACK | 0x03)	/* push register context */
+#define INS_POPREGS	(INS_STACK | 0x04)	/* pop register context */
+#define INS_PUSHFLAGS	(INS_STACK | 0x05)	/* push all flags */
+#define INS_POPFLAGS	(INS_STACK | 0x06)	/* pop all flags */
+#define INS_ENTER	(INS_STACK | 0x07)	/* enter stack frame */
+#define INS_LEAVE	(INS_STACK | 0x08)	/* leave stack frame */
+
+/* INS_COND group */
+#define INS_TEST	(INS_COND | 0x01)
+#define INS_CMP		(INS_COND | 0x02)
+
+/* INS_LOAD group */
+#define INS_MOV		(INS_LOAD | 0x01)
+#define INS_MOVCC	(INS_LOAD | 0x02)
+#define INS_XCHG	(INS_LOAD | 0x03)
+#define INS_XCHGCC	(INS_LOAD | 0x04)
+#define INS_CONV	(INS_LOAD | 0x05)	/* move and convert type */
+
+/* INS_ARRAY group */
+#define INS_STRCMP	(INS_ARRAY | 0x01)
+#define INS_STRLOAD	(INS_ARRAY | 0x02)
+#define INS_STRMOV	(INS_ARRAY | 0x03)
+#define INS_STRSTOR	(INS_ARRAY | 0x04)
+#define INS_XLAT	(INS_ARRAY | 0x05)
+
+/* INS_BIT group */
+#define INS_BITTEST	(INS_BIT | 0x01)
+#define INS_BITSET	(INS_BIT | 0x02)
+#define INS_BITCLR	(INS_BIT | 0x03)
+
+/* INS_FLAG group */
+#define INS_CLEARCF	(INS_FLAG | 0x01)	/* clear Carry flag */
+#define INS_CLEARZF	(INS_FLAG | 0x02)	/* clear Zero flag */
+#define INS_CLEAROF	(INS_FLAG | 0x03)	/* clear Overflow flag */
+#define INS_CLEARDF	(INS_FLAG | 0x04)	/* clear Direction flag */
+#define INS_CLEARSF	(INS_FLAG | 0x05)	/* clear Sign flag */
+#define INS_CLEARPF	(INS_FLAG | 0x06)	/* clear Parity flag */
+#define INS_SETCF	(INS_FLAG | 0x07)
+#define INS_SETZF	(INS_FLAG | 0x08)
+#define INS_SETOF	(INS_FLAG | 0x09)
+#define INS_SETDF	(INS_FLAG | 0x0A)
+#define INS_SETSF	(INS_FLAG | 0x0B)
+#define INS_SETPF	(INS_FLAG | 0x0C)
+#define INS_TOGCF	(INS_FLAG | 0x10)	/* toggle */
+#define INS_TOGZF	(INS_FLAG | 0x20)
+#define INS_TOGOF	(INS_FLAG | 0x30)
+#define INS_TOGDF	(INS_FLAG | 0x40)
+#define INS_TOGSF	(INS_FLAG | 0x50)
+#define INS_TOGPF	(INS_FLAG | 0x60)
+
+/* INS_FPU */
+#define INS_FMOV       (INS_FPU | 0x1)
+#define INS_FMOVCC     (INS_FPU | 0x2)
+#define INS_FNEG       (INS_FPU | 0x3)
+#define INS_FABS       (INS_FPU | 0x4)
+#define INS_FADD       (INS_FPU | 0x5)
+#define INS_FSUB       (INS_FPU | 0x6)
+#define INS_FMUL       (INS_FPU | 0x7)
+#define INS_FDIV       (INS_FPU | 0x8)
+#define INS_FSQRT      (INS_FPU | 0x9)
+#define INS_FCMP       (INS_FPU | 0xA)
+#define INS_FCOS       (INS_FPU | 0xC)               /* cosine */
+#define INS_FLDPI      (INS_FPU | 0xD)               /* load pi */
+#define INS_FLDZ       (INS_FPU | 0xE)               /* load 0 */
+#define INS_FTAN       (INS_FPU | 0xF)               /* tanget */
+#define INS_FSINE      (INS_FPU | 0x10)              /* sine */
+#define INS_FSYS       (INS_FPU | 0x20)              /* misc */
+
+/* INS_TRAP */
+#define INS_TRAP	(INS_TRAPS | 0x01)	/* generate trap */
+#define INS_TRAPCC	(INS_TRAPS | 0x02)	/* conditional trap gen */
+#define INS_TRET	(INS_TRAPS | 0x03)	/* return from trap */
+#define INS_BOUNDS	(INS_TRAPS | 0x04)	/* gen bounds trap */
+#define INS_DEBUG	(INS_TRAPS | 0x05)	/* gen breakpoint trap */
+#define INS_TRACE	(INS_TRAPS | 0x06)	/* gen single step trap */
+#define INS_INVALIDOP	(INS_TRAPS | 0x07)	/* gen invalid insn */
+#define INS_OFLOW	(INS_TRAPS | 0x08)	/* gen overflow trap */
+#define INS_ICEBP	(INS_TRAPS | 0x09)	/* ICE breakpoint */
+
+/* INS_SYSTEM */
+#define INS_HALT	(INS_SYSTEM | 0x01)	/* halt machine */
+#define INS_IN		(INS_SYSTEM | 0x02)	/* input form port */
+#define INS_OUT		(INS_SYSTEM | 0x03)	/* output to port */
+#define INS_CPUID	(INS_SYSTEM | 0x04)	/* identify cpu */
+
+/* INS_OTHER */
+#define INS_NOP		(INS_OTHER | 0x01)
+#define INS_BCDCONV	(INS_OTHER | 0x02)	/* convert to/from BCD */
+#define INS_SZCONV	(INS_OTHER | 0x03)	/* convert size of operand */
+#define INS_SALC	(INS_OTHER | 0x04)	/* set %al on carry */
+#define INS_UNKNOWN	(INS_OTHER | 0x05)
+ 
+
+#define INS_TYPE_MASK	0xFFFF
+#define INS_TYPE( type )      ( type & INS_TYPE_MASK )
+
+   /* flags effected by instruction */
+#define INS_TEST_CARRY        0x01    /* carry */
+#define INS_TEST_ZERO         0x02    /* zero/equal */
+#define INS_TEST_OFLOW        0x04    /* overflow */
+#define INS_TEST_DIR          0x08    /* direction */
+#define INS_TEST_SIGN         0x10    /* negative */
+#define INS_TEST_PARITY       0x20    /* parity */
+#define INS_TEST_OR           0x40    /* used in jle */
+#define INS_TEST_NCARRY       0x100	/* ! carry */
+#define INS_TEST_NZERO        0x200	/* ! zero */
+#define INS_TEST_NOFLOW       0x400	/* ! oflow */
+#define INS_TEST_NDIR         0x800	/* ! dir */
+#define INS_TEST_NSIGN        0x100	/* ! sign */
+#define INS_TEST_NPARITY      0x2000	/* ! parity */
+/* SF == OF */
+#define INS_TEST_SFEQOF       0x4000
+/* SF != OF */
+#define INS_TEST_SFNEOF       0x8000
+
+#define INS_TEST_ALL		INS_TEST_CARRY | INS_TEST_ZERO | \
+				INS_TEST_OFLOW | INS_TEST_SIGN | \
+				INS_TEST_PARITY
+
+#define INS_SET_CARRY        0x010000    /* carry */
+#define INS_SET_ZERO         0x020000    /* zero/equal */
+#define INS_SET_OFLOW        0x040000    /* overflow */
+#define INS_SET_DIR          0x080000    /* direction */
+#define INS_SET_SIGN         0x100000    /* negative */
+#define INS_SET_PARITY       0x200000    /* parity */
+#define INS_SET_NCARRY       0x1000000 
+#define INS_SET_NZERO        0x2000000
+#define INS_SET_NOFLOW       0x4000000
+#define INS_SET_NDIR         0x8000000
+#define INS_SET_NSIGN        0x10000000
+#define INS_SET_NPARITY      0x20000000
+#define INS_SET_SFEQOF       0x40000000
+#define INS_SET_SFNEOF       0x80000000
+
+#define INS_SET_ALL		INS_SET_CARRY | INS_SET_ZERO | \
+				INS_SET_OFLOW | INS_SET_SIGN | \
+				INS_SET_PARITY
+
+#define INS_TEST_MASK          0x0000FFFF
+#define INS_FLAGS_TEST(x)      (x & INS_TEST_MASK)
+#define INS_SET_MASK           0xFFFF0000
+#define INS_FLAGS_SET(x)       (x & INS_SET_MASK)
+
+#if 0
+/* TODO: actually start using these */
+#define X86_PAIR_NP	1		/* not pairable; execs in U */
+#define X86_PAIR_PU	2		/* pairable in U pipe */
+#define X86_PAIR_PV	3		/* pairable in V pipe */
+#define X86_PAIR_UV	4		/* pairable in UV pipe */
+#define X86_PAIR_FX	5		/* pairable with FXCH */
+
+#define X86_EXEC_PORT_0	1
+#define X86_EXEC_PORT_1	2
+#define X86_EXEC_PORT_2	4
+#define X86_EXEC_PORT_3	8
+#define X86_EXEC_PORT_4	16
+
+#define X86_EXEC_UNITS
+
+typedef struct {	/* representation of an insn during decoding */
+	uint32_t flags;		/* runtime settings */
+	/* instruction prefixes and other foolishness */
+	uint32_t prefix;		/* encoding of prefix */
+	char prefix_str[16];		/* mnemonics for prefix */
+	uint32_t branch_hint;	/* gah! */
+	unsigned int cpu_ver;		/* TODO: cpu version */
+	unsigned int clocks;		/* TODO: clock cycles: min/max */
+	unsigned char last_prefix;
+	/* runtime intruction decoding helpers */
+	unsigned char mode;		/* 16, 32, 64 */
+	unsigned char gen_regs;		/* offset of default general reg set */
+	unsigned char sz_operand;	/* operand size for insn */
+	unsigned char sz_address;	/* address size for insn */
+	unsigned char uops;		/* uops per insn */
+	unsigned char pairing;		/* np,pu,pv.lv */
+	unsigned char exec_unit;
+	unsigned char exec_port;
+	unsigned char latency;
+} ia32_info_t;
+#define MODE_32 0	/* default */
+#define MODE_16 1
+#define MODE_64 2
+#endif
+
+#endif

libdisasm/ia32_invariant.c

@@ -0,0 +1,313 @@
+#include <stdlib.h>
+#include <string.h>
+
+#include "ia32_invariant.h"
+#include "ia32_insn.h"
+#include "ia32_settings.h"
+
+extern ia32_table_desc_t *ia32_tables;
+extern ia32_settings_t ia32_settings;
+
+extern size_t ia32_table_lookup( unsigned char *buf, size_t buf_len,
+		unsigned int table, ia32_insn_t **raw_insn,
+		 unsigned int *prefixes );
+
+
+/* -------------------------------- ModR/M, SIB */
+/* Convenience flags */
+#define MODRM_EA  1                     /* ModR/M is an effective addr */
+#define MODRM_reg 2                     /* ModR/M is a register */
+
+/* ModR/M flags */
+#define MODRM_RM_SIB            0x04    /* R/M == 100 */
+#define MODRM_RM_NOREG          0x05    /* R/B == 101 */
+/* if (MODRM.MOD_NODISP && MODRM.RM_NOREG) then just disp32 */
+#define MODRM_MOD_NODISP        0x00    /* mod == 00 */
+#define MODRM_MOD_DISP8         0x01    /* mod == 01 */
+#define MODRM_MOD_DISP32        0x02    /* mod == 10 */
+#define MODRM_MOD_NOEA          0x03    /* mod == 11 */
+/* 16-bit modrm flags */
+#define MOD16_MOD_NODISP      0
+#define MOD16_MOD_DISP8       1
+#define MOD16_MOD_DISP16      2
+#define MOD16_MOD_REG         3
+
+#define MOD16_RM_BXSI         0
+#define MOD16_RM_BXDI         1
+#define MOD16_RM_BPSI         2
+#define MOD16_RM_BPDI         3
+#define MOD16_RM_SI           4
+#define MOD16_RM_DI           5
+#define MOD16_RM_BP           6
+#define MOD16_RM_BX           7
+
+/* SIB flags */
+#define SIB_INDEX_NONE       0x04
+#define SIB_BASE_EBP       0x05
+#define SIB_SCALE_NOBASE    0x00
+
+/* Convenience struct for modR/M bitfield */
+struct modRM_byte {  
+   unsigned int mod : 2;
+   unsigned int reg : 3;
+   unsigned int rm  : 3; 
+};
+
+/* Convenience struct for SIB bitfield */
+struct SIB_byte {
+   unsigned int scale : 2;
+   unsigned int index : 3;
+   unsigned int base  : 3;
+};
+
+#ifdef WIN32
+static void byte_decode(unsigned char b, struct modRM_byte *modrm) {
+#else
+static inline void byte_decode(unsigned char b, struct modRM_byte *modrm) {
+#endif
+	/* generic bitfield-packing routine */
+
+	modrm->mod = b >> 6;	/* top 2 bits */
+	modrm->reg = (b & 56) >> 3;	/* middle 3 bits */
+	modrm->rm = b & 7;	/* bottom 3 bits */
+}
+static int ia32_invariant_modrm( unsigned char *in, unsigned char *out,
+				 unsigned int mode_16, x86_invariant_op_t *op) {
+	struct modRM_byte modrm;
+	struct SIB_byte sib;
+	unsigned char *c, *cin;
+	unsigned short *s;
+	unsigned int *i;
+	int size = 0;	/* modrm byte is already counted */
+
+
+	byte_decode(*in, &modrm);	/* get bitfields */
+
+	out[0] = in[0];	/* save modrm byte */
+	cin = &in[1];
+	c = &out[1];
+	s = (unsigned short *)&out[1];
+	i = (unsigned int *)&out[1];
+
+	op->type = op_expression;
+	op->flags |= op_pointer;
+	if ( ! mode_16 && modrm.rm == MODRM_RM_SIB && 
+			      modrm.mod != MODRM_MOD_NOEA ) {
+		size ++;
+		byte_decode(*cin, (struct modRM_byte *)(void*)&sib);
+
+		out[1] = in[1];	/* save sib byte */
+		cin = &in[2];
+		c = &out[2];
+		s = (unsigned short *)&out[2];
+		i = (unsigned int *)&out[2];
+
+		if ( sib.base == SIB_BASE_EBP && ! modrm.mod ) {
+			/* disp 32 is variant! */
+			memset( i, X86_WILDCARD_BYTE, 4 );
+			size += 4;
+		}
+	}
+
+	if (! modrm.mod && modrm.rm == 101) {
+		if ( mode_16 ) {	/* straight RVA in disp */
+			memset( s, X86_WILDCARD_BYTE, 2 );
+			size += 2;
+		} else {
+			memset( i, X86_WILDCARD_BYTE, 2 );
+			size += 4;
+		}
+	} else if (modrm.mod && modrm.mod < 3) {
+		if (modrm.mod == MODRM_MOD_DISP8) {	 /* offset in disp */
+			*c = *cin;	
+			size += 1;
+		} else if ( mode_16 ) {
+			*s = (* ((unsigned short *) cin));
+			size += 2;
+		} else {
+			*i = (*((unsigned int *) cin));
+			size += 4;
+		}
+	} else if ( modrm.mod == 3 ) {
+		op->type = op_register;
+		op->flags &= ~op_pointer;
+	}
+
+	return (size);
+}
+
+
+static int ia32_decode_invariant( unsigned char *buf, size_t buf_len, 
+				ia32_insn_t *t, unsigned char *out, 
+				unsigned int prefixes, x86_invariant_t *inv) {
+
+	unsigned int addr_size, op_size, mode_16;
+	unsigned int op_flags[3] = { t->dest_flag, t->src_flag, t->aux_flag };
+	int x, type, bytes = 0, size = 0, modrm = 0;
+
+	/* set addressing mode */
+	if (ia32_settings.options & opt_16_bit) {
+		op_size = ( prefixes & PREFIX_OP_SIZE ) ? 4 : 2;
+		addr_size = ( prefixes & PREFIX_ADDR_SIZE ) ? 4 : 2;
+		mode_16 = ( prefixes & PREFIX_ADDR_SIZE ) ? 0 : 1;
+	} else {
+		op_size = ( prefixes & PREFIX_OP_SIZE ) ? 2 : 4;
+		addr_size = ( prefixes & PREFIX_ADDR_SIZE ) ? 2 : 4;
+		mode_16 = ( prefixes & PREFIX_ADDR_SIZE ) ? 1 : 0;
+	}
+
+	for (x = 0; x < 3; x++) {
+		inv->operands[x].access = (enum x86_op_access) 
+						OP_PERM(op_flags[x]);
+		inv->operands[x].flags = (enum x86_op_flags) 
+						(OP_FLAGS(op_flags[x]) >> 12);
+
+		switch (op_flags[x] & OPTYPE_MASK) {
+			case OPTYPE_c:
+				size = (op_size == 4) ? 2 : 1;
+				break;
+			case OPTYPE_a: case OPTYPE_v:
+				size = (op_size == 4) ? 4 : 2;
+				break;
+			case OPTYPE_p:
+				size = (op_size == 4) ? 6 : 4;
+				break;
+			case OPTYPE_b:
+				size = 1;
+				break;
+			case OPTYPE_w:
+				size = 2;
+				break;
+			case OPTYPE_d: case OPTYPE_fs: case OPTYPE_fd:
+			case OPTYPE_fe: case OPTYPE_fb: case OPTYPE_fv:
+			case OPTYPE_si: case OPTYPE_fx:
+				size = 4;
+				break;
+			case OPTYPE_s:
+				size = 6;
+				break;
+			case OPTYPE_q: case OPTYPE_pi:
+				size = 8;
+				break;
+			case OPTYPE_dq: case OPTYPE_ps: case OPTYPE_ss:
+			case OPTYPE_pd: case OPTYPE_sd:
+				size = 16;
+				break;
+			case OPTYPE_m:	
+				size = (addr_size == 4) ? 4 : 2;
+				break;
+			default:
+				break;
+		}
+
+		type = op_flags[x] & ADDRMETH_MASK;
+		switch (type) {
+			case ADDRMETH_E: case ADDRMETH_M: case ADDRMETH_Q:
+			case ADDRMETH_R: case ADDRMETH_W:
+				modrm = 1;	
+				bytes += ia32_invariant_modrm( buf, out, 
+						mode_16, &inv->operands[x]);
+				break;
+			case ADDRMETH_C: case ADDRMETH_D: case ADDRMETH_G:
+			case ADDRMETH_P: case ADDRMETH_S: case ADDRMETH_T:
+			case ADDRMETH_V:
+				inv->operands[x].type = op_register;
+				modrm = 1;
+				break;
+			case ADDRMETH_A: case ADDRMETH_O:
+				/* pad with xF4's */
+				memset( &out[bytes + modrm], X86_WILDCARD_BYTE, 
+					size );
+				bytes += size;
+				inv->operands[x].type = op_offset;
+				if ( type == ADDRMETH_O ) {
+					inv->operands[x].flags |= op_signed |
+								  op_pointer;
+				}
+				break;
+			case ADDRMETH_I: case ADDRMETH_J:
+				/* grab imm value */
+				if ((op_flags[x] & OPTYPE_MASK) == OPTYPE_v) {
+					/* assume this is an address */
+					memset( &out[bytes + modrm], 
+						X86_WILDCARD_BYTE, size );
+				} else {
+					memcpy( &out[bytes + modrm], 
+						&buf[bytes + modrm], size );
+				}
+					
+				bytes += size;
+				if ( type == ADDRMETH_J ) {
+					if ( size == 1 ) {
+						inv->operands[x].type = 
+							op_relative_near;
+					} else {
+						inv->operands[x].type = 
+							op_relative_far;
+					}
+					inv->operands[x].flags |= op_signed;
+				} else {
+					inv->operands[x].type = op_immediate;
+				}
+				break;
+			case ADDRMETH_F:
+				inv->operands[x].type = op_register;
+				break;
+			case ADDRMETH_X:
+				inv->operands[x].flags |= op_signed |
+					  op_pointer | op_ds_seg | op_string;
+				break;
+			case ADDRMETH_Y:
+				inv->operands[x].flags |= op_signed |
+					  op_pointer | op_es_seg | op_string;
+				break;
+			case ADDRMETH_RR:	
+				inv->operands[x].type = op_register;
+				break;
+			case ADDRMETH_II:	
+				inv->operands[x].type = op_immediate;
+				break;
+			default:
+				inv->operands[x].type = op_unused;
+				break;
+		}
+	}
+
+	return (bytes + modrm);
+}
+
+size_t ia32_disasm_invariant( unsigned char * buf, size_t buf_len, 
+		x86_invariant_t *inv ) {
+	ia32_insn_t *raw_insn = NULL;
+	unsigned int prefixes;
+	unsigned int type;
+	size_t size;
+	
+	/* Perform recursive table lookup starting with main table (0) */
+	size = ia32_table_lookup( buf, buf_len, 0, &raw_insn, &prefixes );
+	if ( size == INVALID_INSN || size > buf_len ) {
+		/* TODO: set errno */
+		return 0;
+	}
+
+	/* copy opcode bytes to buffer */
+	memcpy( inv->bytes, buf, size );
+
+	/* set mnemonic type and group */
+	type = raw_insn->mnem_flag & ~INS_FLAG_MASK;
+        inv->group = (enum x86_insn_group) (INS_GROUP(type)) >> 12;
+        inv->type = (enum x86_insn_type) INS_TYPE(type);
+
+	/* handle operands */
+	size += ia32_decode_invariant( buf + size, buf_len - size, raw_insn, 
+					&buf[size - 1], prefixes, inv );
+
+	inv->size = size;
+
+	return size;		/* return size of instruction in bytes */
+}
+
+size_t ia32_disasm_size( unsigned char *buf, size_t buf_len ) {
+	x86_invariant_t inv = { {0} };
+	return( ia32_disasm_invariant( buf, buf_len, &inv ) );
+}

libdisasm/ia32_invariant.h

@@ -0,0 +1,11 @@
+#ifndef IA32_INVARIANT_H
+#define IA32_INVARIANT_H
+
+#include "libdis.h"
+
+size_t ia32_disasm_invariant( unsigned char *buf, size_t buf_len, 
+		x86_invariant_t *inv);
+
+size_t ia32_disasm_size( unsigned char *buf, size_t buf_len );
+
+#endif

libdisasm/ia32_modrm.c

@@ -0,0 +1,310 @@
+#include "ia32_modrm.h"
+#include "ia32_reg.h"
+#include "x86_imm.h"
+
+/* NOTE: when decoding ModR/M and SIB, we have to add 1 to all register
+ * values obtained from decoding the ModR/M or SIB byte, since they
+ * are encoded with eAX = 0 and the tables in ia32_reg.c use eAX = 1.
+ * ADDENDUM: this is only the case when the register value is used
+ * directly as an index into the register table, not when it is added to
+ * a genregs offset. */
+
+/* -------------------------------- ModR/M, SIB */
+/* ModR/M flags */
+#define MODRM_RM_SIB            0x04    /* R/M == 100 */
+#define MODRM_RM_NOREG          0x05    /* R/B == 101 */
+
+/* if (MODRM.MOD_NODISP && MODRM.RM_NOREG) then just disp32 */
+#define MODRM_MOD_NODISP        0x00    /* mod == 00 */
+#define MODRM_MOD_DISP8         0x01    /* mod == 01 */
+#define MODRM_MOD_DISP32        0x02    /* mod == 10 */
+#define MODRM_MOD_NOEA          0x03    /* mod == 11 */
+
+/* 16-bit modrm flags */
+#define MOD16_MOD_NODISP      0
+#define MOD16_MOD_DISP8       1
+#define MOD16_MOD_DISP16      2
+#define MOD16_MOD_REG         3
+
+#define MOD16_RM_BXSI         0
+#define MOD16_RM_BXDI         1
+#define MOD16_RM_BPSI         2
+#define MOD16_RM_BPDI         3
+#define MOD16_RM_SI           4
+#define MOD16_RM_DI           5
+#define MOD16_RM_BP           6
+#define MOD16_RM_BX           7
+
+/* SIB flags */
+#define SIB_INDEX_NONE       0x04
+#define SIB_BASE_EBP       0x05
+#define SIB_SCALE_NOBASE    0x00
+
+/* Convenience struct for modR/M bitfield */
+struct modRM_byte {  
+   unsigned int mod : 2;
+   unsigned int reg : 3;
+   unsigned int rm  : 3; 
+};
+
+/* Convenience struct for SIB bitfield */
+struct SIB_byte {
+   unsigned int scale : 2;
+   unsigned int index : 3;
+   unsigned int base  : 3;
+};
+
+
+#if 0
+int modrm_rm[] = {0,1,2,3,MODRM_RM_SIB,MODRM_MOD_DISP32,6,7};
+int modrm_reg[] = {0, 1, 2, 3, 4, 5, 6, 7};
+int modrm_mod[]  = {0, MODRM_MOD_DISP8, MODRM_MOD_DISP32, MODRM_MOD_NOEA};
+int sib_scl[] = {0, 2, 4, 8};
+int sib_idx[] = {0, 1, 2, 3, SIB_INDEX_NONE, 5, 6, 7 };
+int sib_bas[] = {0, 1, 2, 3, 4, SIB_SCALE_NOBASE, 6, 7 };
+#endif
+
+/* this is needed to replace x86_imm_signsized() which does not sign-extend
+ * to dest */
+static unsigned int imm32_signsized( unsigned char *buf, size_t buf_len,
+				     int32_t *dest, unsigned int size ) {
+	if ( size > buf_len ) {
+		return 0;
+	}
+
+	switch (size) {
+		case 1:
+			*dest = *((signed char *) buf);
+			break;
+		case 2:
+			*dest = *((signed short *) buf);
+			break;
+		case 4:
+		default:
+			*dest = *((signed int *) buf);
+			break;
+	}
+
+	return size;
+}
+
+
+
+static void byte_decode(unsigned char b, struct modRM_byte *modrm) {
+	/* generic bitfield-packing routine */
+
+	modrm->mod = b >> 6;	/* top 2 bits */
+	modrm->reg = (b & 56) >> 3;	/* middle 3 bits */
+	modrm->rm = b & 7;	/* bottom 3 bits */
+}
+
+
+static size_t sib_decode( unsigned char *buf, size_t buf_len, x86_ea_t *ea, 
+			  unsigned int mod ) {
+	/* set Address Expression fields (scale, index, base, disp) 
+	 * according to the contents of the SIB byte.
+	 *  b points to the SIB byte in the instruction-stream buffer; the
+	 *    byte after b[0] is therefore the byte after the SIB
+	 *  returns number of bytes 'used', including the SIB byte */
+	size_t size = 1;		/* start at 1 for SIB byte */
+	struct SIB_byte sib;
+
+	if ( buf_len < 1 ) {
+		return 0;
+	}
+
+	byte_decode( *buf, (struct modRM_byte *)(void*)&sib );  /* get bit-fields */
+
+	if ( sib.base == SIB_BASE_EBP && ! mod ) {  /* if base == 101 (ebp) */
+	    /* IF BASE == EBP, deal with exception */
+		/* IF (ModR/M did not create a Disp */
+		/* ... create a 32-bit Displacement */
+		imm32_signsized( &buf[1], buf_len, &ea->disp, sizeof(int32_t));
+		ea->disp_size = sizeof(int32_t);
+		ea->disp_sign = (ea->disp < 0) ? 1 : 0;
+		size += 4;	/* add sizeof disp to count */
+
+	} else {
+		/* ELSE BASE refers to a General Register */
+		ia32_handle_register( &ea->base, sib.base + 1 );
+	}
+
+	/* set scale to 1, 2, 4, 8 */
+	ea->scale = 1 << sib.scale;
+
+	if (sib.index != SIB_INDEX_NONE) {
+		/* IF INDEX is not 'ESP' (100) */
+		ia32_handle_register( &ea->index, sib.index + 1 );
+	}
+
+	return (size);		/* return number of bytes processed */
+}
+
+static size_t modrm_decode16( unsigned char *buf, unsigned int buf_len,
+			    x86_op_t *op, struct modRM_byte *modrm ) {
+	/* 16-bit mode: hackish, but not as hackish as 32-bit mode ;) */
+	size_t size = 1; /* # of bytes decoded [1 for modR/M byte] */
+	x86_ea_t * ea = &op->data.expression;
+
+	switch( modrm->rm ) {
+		case MOD16_RM_BXSI:
+			ia32_handle_register(&ea->base, REG_WORD_OFFSET + 3);
+			ia32_handle_register(&ea->index, REG_WORD_OFFSET + 6);
+			break;
+		case MOD16_RM_BXDI:
+			ia32_handle_register(&ea->base, REG_WORD_OFFSET + 3);
+			ia32_handle_register(&ea->index, REG_WORD_OFFSET + 7);
+		case MOD16_RM_BPSI:
+			op->flags |= op_ss_seg;
+			ia32_handle_register(&ea->base, REG_WORD_OFFSET + 5);
+			ia32_handle_register(&ea->index, REG_WORD_OFFSET + 6);
+			break;
+		case MOD16_RM_BPDI:
+			op->flags |= op_ss_seg;
+			ia32_handle_register(&ea->base, REG_WORD_OFFSET + 5);
+			ia32_handle_register(&ea->index, REG_WORD_OFFSET + 7);
+			break;
+		case MOD16_RM_SI:
+			ia32_handle_register(&ea->base, REG_WORD_OFFSET + 6);
+			break;
+		case MOD16_RM_DI:
+			ia32_handle_register(&ea->base, REG_WORD_OFFSET + 7);
+			break;
+		case MOD16_RM_BP:
+			if ( modrm->mod != MOD16_MOD_NODISP ) {
+				op->flags |= op_ss_seg;
+				ia32_handle_register(&ea->base, 
+						     REG_WORD_OFFSET + 5);
+			}
+			break;
+		case MOD16_RM_BX:
+			ia32_handle_register(&ea->base, REG_WORD_OFFSET + 3);
+			break;
+	}
+
+	/* move to byte after ModR/M */
+	++buf;
+	--buf_len;
+
+	if ( modrm->mod == MOD16_MOD_DISP8 ) {
+		imm32_signsized( buf, buf_len, &ea->disp, sizeof(char) );
+		ea->disp_sign = (ea->disp < 0) ? 1 : 0;
+		ea->disp_size = sizeof(char);
+		size += sizeof(char);
+	} else if ( modrm->mod == MOD16_MOD_DISP16 ) {
+		imm32_signsized( buf, buf_len, &ea->disp, sizeof(short) );
+		ea->disp_sign = (ea->disp < 0) ? 1 : 0;
+		ea->disp_size = sizeof(short);
+		size += sizeof(short);
+	} 
+
+	return size;
+}
+
+/* TODO : Mark index modes
+    Use addressing mode flags to imply arrays (index), structure (disp),
+    two-dimensional arrays [disp + index], classes [ea reg], and so on.
+*/
+size_t ia32_modrm_decode( unsigned char *buf, unsigned int buf_len,
+			    x86_op_t *op, x86_insn_t *insn, size_t gen_regs ) {
+	/* create address expression and/or fill operand based on value of
+	 * ModR/M byte. Calls sib_decode as appropriate.
+	 *    flags specifies whether Reg or mod+R/M fields are being decoded
+	 *  returns the number of bytes in the instruction, including modR/M */
+	struct modRM_byte modrm;
+	size_t size = 1;	/* # of bytes decoded [1 for modR/M byte] */
+	x86_ea_t * ea;
+
+
+	byte_decode(*buf, &modrm);	/* get bitfields */
+
+	/* first, handle the case where the mod field is a register only */
+	if ( modrm.mod == MODRM_MOD_NOEA ) {
+		op->type = op_register;
+		ia32_handle_register(&op->data.reg, modrm.rm + gen_regs);
+                /* increase insn size by 1 for modrm byte */
+ 		return 1;
+ 	}
+ 
+	/* then deal with cases where there is an effective address */
+	ea = &op->data.expression;
+	op->type = op_expression;
+	op->flags |= op_pointer;
+
+	if ( insn->addr_size == 2 ) {
+		/* gah! 16 bit mode! */
+		return modrm_decode16( buf, buf_len, op, &modrm);
+	}
+
+	/* move to byte after ModR/M */
+	++buf;
+	--buf_len;
+
+	if (modrm.mod == MODRM_MOD_NODISP) {	/* if mod == 00 */
+
+		/* IF MOD == No displacement, just Indirect Register */
+		if (modrm.rm == MODRM_RM_NOREG) {	/* if r/m == 101 */
+			/* IF RM == No Register, just Displacement */
+			/* This is an Intel Moronic Exception TM */
+			imm32_signsized( buf, buf_len, &ea->disp, 
+					sizeof(int32_t) );
+			ea->disp_size = sizeof(int32_t);
+			ea->disp_sign = (ea->disp < 0) ? 1 : 0;
+			size += 4;	/* add sizeof disp to count */
+
+		} else if (modrm.rm == MODRM_RM_SIB) {	/* if r/m == 100 */
+			/* ELSE IF an SIB byte is present */
+			/* TODO: check for 0 retval */
+			size += sib_decode( buf, buf_len, ea, modrm.mod);
+			/* move to byte after SIB for displacement */
+			++buf;
+			--buf_len;
+		} else {	/* modR/M specifies base register */
+			/* ELSE RM encodes a general register */
+			ia32_handle_register( &ea->base, modrm.rm + 1 );
+		}
+	} else { 					/* mod is 01 or 10 */
+		if (modrm.rm == MODRM_RM_SIB) {	/* rm == 100 */
+			/* IF base is an AddrExpr specified by an SIB byte */
+			/* TODO: check for 0 retval */
+			size += sib_decode( buf, buf_len, ea, modrm.mod);
+			/* move to byte after SIB for displacement */
+			++buf;
+			--buf_len;
+		} else {
+			/* ELSE base is a general register */
+			ia32_handle_register( &ea->base, modrm.rm + 1 );
+		}
+
+		/* ELSE mod + r/m specify a disp##[base] or disp##(SIB) */
+		if (modrm.mod == MODRM_MOD_DISP8) {		/* mod == 01 */
+			/* If this is an 8-bit displacement */
+			imm32_signsized( buf, buf_len, &ea->disp, 
+					sizeof(char));
+			ea->disp_size = sizeof(char);
+			ea->disp_sign = (ea->disp < 0) ? 1 : 0;
+			size += 1;	/* add sizeof disp to count */
+
+		} else {
+			/* Displacement is dependent on address size */
+			imm32_signsized( buf, buf_len, &ea->disp, 
+					insn->addr_size);
+			ea->disp_size = insn->addr_size;
+			ea->disp_sign = (ea->disp < 0) ? 1 : 0;
+			size += 4;
+		}
+	}
+
+	return size;		/* number of bytes found in instruction */
+}
+
+void ia32_reg_decode( unsigned char byte, x86_op_t *op, size_t gen_regs ) {
+	struct modRM_byte modrm;
+	byte_decode( byte, &modrm );	/* get bitfields */
+
+ 	/* set operand to register ID */
+	op->type = op_register;
+	ia32_handle_register(&op->data.reg, modrm.reg + gen_regs);
+
+	return;
+}

libdisasm/ia32_modrm.h

@@ -0,0 +1,13 @@
+#ifndef IA32_MODRM_H
+#define IA32_MODRM_H
+
+#include "libdis.h"
+#include "ia32_insn.h"
+
+size_t ia32_modrm_decode( unsigned char *buf, unsigned int buf_len,
+			    x86_op_t *op, x86_insn_t *insn,
+			    size_t gen_regs );
+
+void ia32_reg_decode( unsigned char byte, x86_op_t *op, size_t gen_regs );
+
+#endif

libdisasm/ia32_opcode_tables.h

@@ -0,0 +1,57 @@
+#define idx_Main 0
+#define idx_66 1
+#define idx_F2 2
+#define idx_F3 3
+#define idx_0F 4
+#define idx_660F 5
+#define idx_F20F 6
+#define idx_F30F 7
+#define idx_0F00 8
+#define idx_0F01 9
+#define idx_0F0111 10
+#define idx_0F12 11
+#define idx_0F16 12
+#define idx_0F18 13
+#define idx_0F71 14
+#define idx_660F71 15
+#define idx_0F72 16
+#define idx_660F72 17
+#define idx_0F73 18
+#define idx_660F73 19
+#define idx_0FAE 20
+#define idx_0FBA 21
+#define idx_0FC7 22
+#define idx_0FB9 23
+#define idx_C6 24
+#define idx_C7 25
+#define idx_80 26
+#define idx_81 27
+#define idx_82 28
+#define idx_83 29
+#define idx_C0 30
+#define idx_C1 31
+#define idx_D0 32
+#define idx_D1 33
+#define idx_D2 34
+#define idx_D3 35
+#define idx_F6 36
+#define idx_F7 37
+#define idx_FE 38
+#define idx_FF 39
+#define idx_D8 40
+#define idx_D8C0 41
+#define idx_D9 42
+#define idx_D9C0 43
+#define idx_DA 44
+#define idx_DAC0 45
+#define idx_DB 46
+#define idx_DBC0 47
+#define idx_DC 48
+#define idx_DCC0 49
+#define idx_DD 50
+#define idx_DDC0 51
+#define idx_DE 52
+#define idx_DEC0 53
+#define idx_DF 54
+#define idx_DFC0 55
+#define idx_0F0F 56

libdisasm/ia32_operand.c

@@ -0,0 +1,425 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include "libdis.h"
+#include "ia32_insn.h"
+#include "ia32_operand.h"
+#include "ia32_modrm.h"
+#include "ia32_reg.h"
+#include "x86_imm.h"
+#include "x86_operand_list.h"
+
+
+
+/* apply segment override to memory operand in insn */
+static void apply_seg( x86_op_t *op, unsigned int prefixes ) {
+	if (! prefixes ) return;
+
+	/* apply overrides from prefix */
+	switch ( prefixes & PREFIX_REG_MASK ) {
+		case PREFIX_CS:
+			op->flags |= op_cs_seg; break;
+		case PREFIX_SS:
+			op->flags |= op_ss_seg; break;
+		case PREFIX_DS:
+			op->flags |= op_ds_seg; break;
+		case PREFIX_ES:
+			op->flags |= op_es_seg; break;
+		case PREFIX_FS:
+			op->flags |= op_fs_seg; break;
+		case PREFIX_GS:
+			op->flags |= op_gs_seg; break;
+	}
+
+	return;
+}
+
+static size_t decode_operand_value( unsigned char *buf, size_t buf_len,
+			    x86_op_t *op, x86_insn_t *insn, 
+			    unsigned int addr_meth, size_t op_size, 
+			    unsigned int op_value, unsigned char modrm, 
+			    size_t gen_regs ) {
+	size_t size = 0;
+
+	/* ++ Do Operand Addressing Method / Decode operand ++ */
+	switch (addr_meth) {
+		/* This sets the operand Size based on the Intel Opcode Map
+		 * (Vol 2, Appendix A). Letter encodings are from section
+		 * A.1.1, 'Codes for Addressing Method' */
+
+		/* ---------------------- Addressing Method -------------- */
+		/* Note that decoding mod ModR/M operand adjusts the size of
+		 * the instruction, but decoding the reg operand does not. 
+		 * This should not cause any problems, as every 'reg' operand 
+		 * has an associated 'mod' operand. 
+		 * Goddamn-Intel-Note:
+		 *   Some Intel addressing methods [M, R] specify that modR/M
+		 *   byte may only refer to a memory address/may only refer to
+		 *   a register -- however Intel provides no clues on what to do
+		 *   if, say, the modR/M for an M opcode decodes to a register
+		 *   rather than a memory address ... returning 0 is out of the
+		 *   question, as this would be an Immediate or a RelOffset, so
+		 *   instead these modR/Ms are decoded with total disregard to 
+		 *   the M, R constraints. */
+
+		/* MODRM -- mod operand. sets size to at least 1! */
+		case ADDRMETH_E:	/* ModR/M present, Gen reg or memory  */
+			size = ia32_modrm_decode( buf, buf_len, op, insn, 
+						  gen_regs );
+			break;
+		case ADDRMETH_M:	/* ModR/M only refers to memory */
+			size = ia32_modrm_decode( buf, buf_len, op, insn, 
+						  gen_regs );
+			break;
+		case ADDRMETH_Q:	/* ModR/M present, MMX or Memory */
+			size = ia32_modrm_decode( buf, buf_len, op, insn, 
+						  REG_MMX_OFFSET );
+			break;
+		case ADDRMETH_R:	/* ModR/M mod == gen reg */
+			size = ia32_modrm_decode( buf, buf_len, op, insn, 
+						  gen_regs );
+			break;
+		case ADDRMETH_W:	/* ModR/M present, mem or SIMD reg */
+			size = ia32_modrm_decode( buf, buf_len, op, insn, 
+						  REG_SIMD_OFFSET );
+			break;
+
+		/* MODRM -- reg operand. does not effect size! */
+		case ADDRMETH_C:	/* ModR/M reg == control reg */
+			ia32_reg_decode( modrm, op, REG_CTRL_OFFSET );
+			break;
+		case ADDRMETH_D:	/* ModR/M reg == debug reg */
+			ia32_reg_decode( modrm, op, REG_DEBUG_OFFSET );
+			break;
+		case ADDRMETH_G:	/* ModR/M reg == gen-purpose reg */
+			ia32_reg_decode( modrm, op, gen_regs );
+			break;
+		case ADDRMETH_P:	/* ModR/M reg == qword MMX reg */
+			ia32_reg_decode( modrm, op, REG_MMX_OFFSET );
+			break;
+		case ADDRMETH_S:	/* ModR/M reg == segment reg */
+			ia32_reg_decode( modrm, op, REG_SEG_OFFSET );
+			break;
+		case ADDRMETH_T:	/* ModR/M reg == test reg */
+			ia32_reg_decode( modrm, op, REG_TEST_OFFSET );
+			break;
+		case ADDRMETH_V:	/* ModR/M reg == SIMD reg */
+			ia32_reg_decode( modrm, op, REG_SIMD_OFFSET );
+			break;
+
+		/* No MODRM : note these set operand type explicitly */
+		case ADDRMETH_A:	/* No modR/M -- direct addr */
+			op->type = op_absolute;
+
+			/* segment:offset address used in far calls */
+			x86_imm_sized( buf, buf_len, 
+				       &op->data.absolute.segment, 2 );
+			if ( insn->addr_size == 4 ) {
+				x86_imm_sized( buf, buf_len, 
+				    &op->data.absolute.offset.off32, 4 );
+				size = 6;
+			} else {
+				x86_imm_sized( buf, buf_len, 
+				    &op->data.absolute.offset.off16, 2 );
+				size = 4;
+			}
+
+			break;
+		case ADDRMETH_I:	/* Immediate val */
+			op->type = op_immediate;
+			/* if it ever becomes legal to have imm as dest and
+			 * there is a src ModR/M operand, we are screwed! */
+			if ( op->flags & op_signed ) {
+				x86_imm_signsized(buf, buf_len, &op->data.byte, 
+						op_size);
+			} else {
+				x86_imm_sized(buf, buf_len, &op->data.byte, 
+						op_size);
+			}
+			size = op_size;
+			break;
+		case ADDRMETH_J:	/* Rel offset to add to IP [jmp] */
+			/* this fills op->data.near_offset or
+			   op->data.far_offset depending on the size of
+			   the operand */
+			op->flags |= op_signed;
+			if ( op_size == 1 ) {
+				/* one-byte near offset */
+				op->type = op_relative_near;
+				x86_imm_signsized(buf, buf_len, 
+						&op->data.relative_near, 1);
+			} else {
+				/* far offset...is this truly signed? */
+				op->type = op_relative_far;
+				x86_imm_signsized(buf, buf_len, 
+					&op->data.relative_far, op_size );
+			}
+			size = op_size;
+			break;
+		case ADDRMETH_O:	/* No ModR/M; op is word/dword offset */
+			/* NOTE: these are actually RVAs not offsets to seg!! */
+			/* note bene: 'O' ADDR_METH uses addr_size  to
+			   determine operand size */
+			op->type = op_offset;
+			op->flags |= op_pointer;
+			x86_imm_sized( buf, buf_len, &op->data.offset, 
+					insn->addr_size );
+
+			size = insn->addr_size;
+			break;
+
+		/* Hard-coded: these are specified in the insn definition */
+		case ADDRMETH_F:	/* EFLAGS register */
+			op->type = op_register;
+			op->flags |= op_hardcode;
+			ia32_handle_register( &op->data.reg, REG_FLAGS_INDEX );
+			break;
+		case ADDRMETH_X:	/* Memory addressed by DS:SI [string] */
+			op->type = op_expression;
+			op->flags |= op_hardcode;
+			op->flags |= op_ds_seg | op_pointer | op_string;
+			ia32_handle_register( &op->data.expression.base, 
+					     REG_DWORD_OFFSET + 6 );
+			break;
+		case ADDRMETH_Y:	/* Memory addressed by ES:DI [string] */
+			op->type = op_expression;
+			op->flags |= op_hardcode;
+			op->flags |= op_es_seg | op_pointer | op_string;
+			ia32_handle_register( &op->data.expression.base, 
+					     REG_DWORD_OFFSET + 7 );
+			break;
+		case ADDRMETH_RR:	/* Gen Register hard-coded in opcode */
+			op->type = op_register;
+			op->flags |= op_hardcode;
+			ia32_handle_register( &op->data.reg, 
+						op_value + gen_regs );
+			break;
+		case ADDRMETH_RS:	/* Seg Register hard-coded in opcode */
+			op->type = op_register;
+			op->flags |= op_hardcode;
+			ia32_handle_register( &op->data.reg, 
+						op_value + REG_SEG_OFFSET );
+			break;
+		case ADDRMETH_RF:	/* FPU Register hard-coded in opcode */
+			op->type = op_register;
+			op->flags |= op_hardcode;
+			ia32_handle_register( &op->data.reg, 
+						op_value + REG_FPU_OFFSET );
+			break;
+		case ADDRMETH_RT:	/* TST Register hard-coded in opcode */
+			op->type = op_register;
+			op->flags |= op_hardcode;
+			ia32_handle_register( &op->data.reg, 
+						op_value + REG_TEST_OFFSET );
+			break;
+		case ADDRMETH_II:	/* Immediate hard-coded in opcode */
+			op->type = op_immediate;
+			op->data.dword = op_value;
+			op->flags |= op_hardcode;
+			break;
+
+		case 0:	/* Operand is not used */
+		default:
+			/* ignore -- operand not used in this insn */
+			op->type = op_unused;	/* this shouldn't happen! */
+			break;
+	}
+
+	return size;
+}
+
+static size_t decode_operand_size( unsigned int op_type, x86_insn_t *insn, 
+				   x86_op_t *op ){
+	size_t size;
+
+	/* ++ Do Operand Type ++ */
+	switch (op_type) {
+		/* This sets the operand Size based on the Intel Opcode Map
+		 * (Vol 2, Appendix A). Letter encodings are from section
+		 * A.1.2, 'Codes for Operand Type' */
+		/* NOTE: in this routines, 'size' refers to the size
+		 *       of the operand in the raw (encoded) instruction;
+		 *       'datatype' stores the actual size and datatype
+		 *       of the operand */
+
+		/* ------------------------ Operand Type ----------------- */
+		case OPTYPE_c:	/* byte or word [op size attr] */
+			size = (insn->op_size == 4) ? 2 : 1;
+			op->datatype = (size == 4) ? op_word : op_byte;
+			break;
+		case OPTYPE_a:	/* 2 word or 2 dword [op size attr] */
+			/* pointer to a 16:16 or 32:32 BOUNDS operand */
+			size = (insn->op_size == 4) ? 8 : 4;
+			op->datatype = (size == 4) ? op_bounds32 : op_bounds16;
+			break;
+		case OPTYPE_v:	/* word or dword [op size attr] */
+			size = (insn->op_size == 4) ? 4 : 2;
+			op->datatype = (size == 4) ? op_dword : op_word;
+			break;
+		case OPTYPE_p:	/* 32/48-bit ptr [op size attr] */
+			/* technically these flags are not accurate: the
+			 * value s a 16:16 pointer or a 16:32 pointer, where
+			 * the first '16' is a segment */
+			size = (insn->addr_size == 4) ? 6 : 4;
+			op->datatype = (size == 4) ? op_descr32 : op_descr16;
+			break;
+		case OPTYPE_b:	/* byte, ignore op-size */
+			size = 1;
+			op->datatype = op_byte;
+			break;
+		case OPTYPE_w:	/* word, ignore op-size */
+			size = 2;
+			op->datatype = op_word;
+			break;
+		case OPTYPE_d:	/* dword , ignore op-size */
+			size = 4;
+			op->datatype = op_dword;
+			break;
+		case OPTYPE_s:	/* 6-byte psuedo-descriptor */
+			/* ptr to 6-byte value which is 32:16 in 32-bit
+			 * mode, or 8:24:16 in 16-bit mode. The high byte
+			 * is ignored in 16-bit mode. */
+			size = 6;
+			op->datatype = (insn->addr_size == 4) ? 
+				op_pdescr32 : op_pdescr16;
+			break;
+		case OPTYPE_q:	/* qword, ignore op-size */
+			size = 8;
+			op->datatype = op_qword;
+			break;
+		case OPTYPE_dq:	/* d-qword, ignore op-size */
+			size = 16;
+			op->datatype = op_dqword;
+			break;
+		case OPTYPE_ps:	/* 128-bit FP data */
+			size = 16;
+			/* really this is 4 packed SP FP values */
+			op->datatype = op_ssimd;
+			break;
+		case OPTYPE_pd:	/* 128-bit FP data */
+			size = 16;
+			/* really this is 2 packed DP FP values */
+			op->datatype = op_dsimd;
+			break;
+		case OPTYPE_ss:	/* Scalar elem of 128-bit FP data */
+			size = 16;
+			/* this only looks at the low dword (4 bytes)
+			 * of the xmmm register passed as a param. 
+			 * This is a 16-byte register where only 4 bytes
+			 * are used in the insn. Painful, ain't it? */
+			op->datatype = op_sssimd;
+			break;
+		case OPTYPE_sd:	/* Scalar elem of 128-bit FP data */
+			size = 16;
+			/* this only looks at the low qword (8 bytes)
+			 * of the xmmm register passed as a param. 
+			 * This is a 16-byte register where only 8 bytes
+			 * are used in the insn. Painful, again... */
+			op->datatype = op_sdsimd;
+			break;
+		case OPTYPE_pi:	/* qword mmx register */
+			size = 8;
+			op->datatype = op_qword;
+			break;
+		case OPTYPE_si:	/* dword integer register */
+			size = 4;
+			op->datatype = op_dword;
+			break;
+		case OPTYPE_fs:	/* single-real */
+			size = 4;
+			op->datatype = op_sreal;
+			break;
+		case OPTYPE_fd:	/* double real */
+			size = 8;
+			op->datatype = op_dreal;
+			break;
+		case OPTYPE_fe:	/* extended real */
+			size = 10;
+			op->datatype = op_extreal;
+			break;
+		case OPTYPE_fb:	/* packed BCD */
+			size = 10;
+			op->datatype = op_bcd;
+			break;
+		case OPTYPE_fv:	/* pointer to FPU env: 14 or 28-bytes */
+			size = (insn->addr_size == 4)? 28 : 14;
+			op->datatype = (size == 28)?  op_fpuenv32: op_fpuenv16;
+			break;
+		case OPTYPE_ft:	/* pointer to FPU env: 94 or 108 bytes */
+			size = (insn->addr_size == 4)? 108 : 94;
+			op->datatype = (size == 108)? 
+				op_fpustate32: op_fpustate16;
+			break;
+		case OPTYPE_fx:	/* 512-byte register stack */
+			size = 512;
+			op->datatype = op_fpregset;
+			break;
+		case OPTYPE_fp:	/* floating point register */
+			size = 10;	/* double extended precision */
+			op->datatype = op_fpreg;
+			break;
+		case OPTYPE_m:	/* fake operand type used for "lea Gv, M" */
+			size = insn->addr_size;
+			op->datatype = (size == 4) ?  op_dword : op_word;
+			break;
+		case OPTYPE_none: /* handle weird instructions that have no encoding but use a dword datatype, like invlpg */
+			size = 0;
+			op->datatype = op_none;
+			break;
+		case 0:
+		default:
+			size = insn->op_size;
+			op->datatype = (size == 4) ? op_dword : op_word;
+			break;
+		}
+	return size;
+}
+
+size_t ia32_decode_operand( unsigned char *buf, size_t buf_len,
+			      x86_insn_t *insn, unsigned int raw_op, 
+			      unsigned int raw_flags, unsigned int prefixes, 
+			      unsigned char modrm ) {
+	unsigned int addr_meth, op_type, op_size, gen_regs;
+	x86_op_t *op;
+	size_t size;
+
+	/* ++ Yank optype and addr mode out of operand flags */
+	addr_meth = raw_flags & ADDRMETH_MASK;
+	op_type = raw_flags & OPTYPE_MASK;
+
+	if ( raw_flags == ARG_NONE ) {
+		/* operand is not used in this instruction */
+		return 0;
+	}
+
+	/* allocate a new operand */
+	op = x86_operand_new( insn );
+
+	/* ++ Copy flags from opcode table to x86_insn_t */
+	op->access = (enum x86_op_access) OP_PERM(raw_flags);
+	op->flags = (enum x86_op_flags) (OP_FLAGS(raw_flags) >> 12);
+
+	/* Get size (for decoding)  and datatype of operand */
+	op_size = decode_operand_size(op_type, insn, op);
+
+	/* override default register set based on Operand Type */
+	/* this allows mixing of 8, 16, and 32 bit regs in insn */
+	if (op_size == 1) {
+		gen_regs = REG_BYTE_OFFSET;
+	} else if (op_size == 2) {
+		gen_regs = REG_WORD_OFFSET;
+	} else {
+		gen_regs = REG_DWORD_OFFSET;
+	}
+
+	size = decode_operand_value( buf, buf_len, op, insn, addr_meth, 
+				      op_size, raw_op, modrm, gen_regs );
+
+	/* if operand is an address, apply any segment override prefixes */
+	if ( op->type == op_expression || op->type == op_offset ) {
+		apply_seg(op, prefixes);
+	}
+
+	return size;		/* return number of bytes in instruction */
+}

libdisasm/ia32_operand.h

@@ -0,0 +1,11 @@
+#ifndef IA32_OPERAND_H
+#define IA32_OPERAND_H
+
+#include "libdis.h"
+#include "ia32_insn.h"
+
+size_t ia32_decode_operand( unsigned char *buf, size_t buf_len,
+			      x86_insn_t *insn, unsigned int raw_op, 
+			      unsigned int raw_flags, unsigned int prefixes,
+			      unsigned char modrm );
+#endif

libdisasm/ia32_reg.c

@@ -0,0 +1,234 @@
+#include <stdlib.h>
+#include <string.h>
+
+#include "ia32_reg.h"
+#include "ia32_insn.h"
+
+#define NUM_X86_REGS	92
+
+/* register sizes */
+#define REG_DWORD_SIZE 4
+#define REG_WORD_SIZE 2
+#define REG_BYTE_SIZE 1
+#define REG_MMX_SIZE 8
+#define REG_SIMD_SIZE 16
+#define REG_DEBUG_SIZE 4
+#define REG_CTRL_SIZE 4
+#define REG_TEST_SIZE 4
+#define REG_SEG_SIZE 2
+#define REG_FPU_SIZE 10
+#define REG_FLAGS_SIZE 4
+#define REG_FPCTRL_SIZE 2
+#define REG_FPSTATUS_SIZE 2
+#define REG_FPTAG_SIZE 2
+#define REG_EIP_SIZE 4
+#define REG_IP_SIZE 2
+
+/* REGISTER ALIAS TABLE:
+ *
+ * NOTE: the MMX register mapping is fixed to the physical registers
+ * used by the FPU. The floating FP stack does not effect the location
+ * of the MMX registers, so this aliasing is not 100% accurate.
+ * */
+static struct {
+	unsigned char alias;	/* id of register this is an alias for */
+	unsigned char shift;	/* # of bits register must be shifted */
+} ia32_reg_aliases[] = {
+	{ 0,0 },
+	{ REG_DWORD_OFFSET,	0 }, 	/* al :  1 */
+	{ REG_DWORD_OFFSET,	8 }, 	/* ah :  2 */
+	{ REG_DWORD_OFFSET,	0 }, 	/* ax :  3 */
+	{ REG_DWORD_OFFSET + 1,	0 }, 	/* cl :  4 */
+	{ REG_DWORD_OFFSET + 1,	8 }, 	/* ch :  5 */
+	{ REG_DWORD_OFFSET + 1,	0 }, 	/* cx :  6 */
+	{ REG_DWORD_OFFSET + 2,	0 }, 	/* dl :  7 */
+	{ REG_DWORD_OFFSET + 2,	8 }, 	/* dh :  8 */
+	{ REG_DWORD_OFFSET + 2,	0 }, 	/* dx :  9 */
+	{ REG_DWORD_OFFSET + 3,	0 }, 	/* bl : 10 */
+	{ REG_DWORD_OFFSET + 3,	8 }, 	/* bh : 11 */
+	{ REG_DWORD_OFFSET + 3,	0 }, 	/* bx : 12 */
+	{ REG_DWORD_OFFSET + 4,	0 }, 	/* sp : 13 */
+	{ REG_DWORD_OFFSET + 5,	0 }, 	/* bp : 14 */
+	{ REG_DWORD_OFFSET + 6,	0 }, 	/* si : 15 */
+	{ REG_DWORD_OFFSET + 7,	0 }, 	/* di : 16 */
+	{ REG_EIP_INDEX,	0 }, 	/* ip : 17 */
+	{ REG_FPU_OFFSET,	0 }, 	/* mm0 : 18 */
+	{ REG_FPU_OFFSET + 1,	0 }, 	/* mm1 : 19 */
+	{ REG_FPU_OFFSET + 2,	0 }, 	/* mm2 : 20 */
+	{ REG_FPU_OFFSET + 3,	0 }, 	/* mm3 : 21 */
+	{ REG_FPU_OFFSET + 4,	0 }, 	/* mm4 : 22 */
+	{ REG_FPU_OFFSET + 5,	0 }, 	/* mm5 : 23 */
+	{ REG_FPU_OFFSET + 6,	0 }, 	/* mm6 : 24 */
+	{ REG_FPU_OFFSET + 7,	0 } 	/* mm7 : 25 */
+ };
+
+/* REGISTER TABLE: size, type, and name of every register in the 
+ *                 CPU. Does not include MSRs since the are, after all,
+ *                 model specific. */
+static struct {
+	unsigned int size;
+	enum x86_reg_type type;
+	unsigned int alias;
+	char mnemonic[8];
+} ia32_reg_table[NUM_X86_REGS + 2] = {
+	{ 0, 0, 0, "" },
+	/* REG_DWORD_OFFSET */
+	{ REG_DWORD_SIZE, reg_gen | reg_ret, 0, "eax" },
+	{ REG_DWORD_SIZE, reg_gen | reg_count, 0, "ecx" },
+	{ REG_DWORD_SIZE, reg_gen, 0, "edx" },
+	{ REG_DWORD_SIZE, reg_gen, 0, "ebx" },
+	/* REG_ESP_INDEX */
+	{ REG_DWORD_SIZE, reg_gen | reg_sp, 0, "esp" },
+	{ REG_DWORD_SIZE, reg_gen | reg_fp, 0, "ebp" },
+	{ REG_DWORD_SIZE, reg_gen | reg_src, 0, "esi" },
+	{ REG_DWORD_SIZE, reg_gen | reg_dest, 0, "edi" },
+	/* REG_WORD_OFFSET */
+	{ REG_WORD_SIZE, reg_gen | reg_ret, 3, "ax" },
+	{ REG_WORD_SIZE, reg_gen | reg_count, 6, "cx" },
+	{ REG_WORD_SIZE, reg_gen, 9, "dx" },
+	{ REG_WORD_SIZE, reg_gen, 12, "bx" },
+	{ REG_WORD_SIZE, reg_gen | reg_sp, 13, "sp" },
+	{ REG_WORD_SIZE, reg_gen | reg_fp, 14, "bp" },
+	{ REG_WORD_SIZE, reg_gen | reg_src, 15, "si" },
+	{ REG_WORD_SIZE, reg_gen | reg_dest, 16, "di" },
+	/* REG_BYTE_OFFSET */
+	{ REG_BYTE_SIZE, reg_gen, 1, "al" },
+	{ REG_BYTE_SIZE, reg_gen, 4, "cl" },
+	{ REG_BYTE_SIZE, reg_gen, 7, "dl" },
+	{ REG_BYTE_SIZE, reg_gen, 10, "bl" },
+	{ REG_BYTE_SIZE, reg_gen, 2, "ah" },
+	{ REG_BYTE_SIZE, reg_gen, 5, "ch" },
+	{ REG_BYTE_SIZE, reg_gen, 8, "dh" },
+	{ REG_BYTE_SIZE, reg_gen, 11, "bh" },
+	/* REG_MMX_OFFSET */
+	{ REG_MMX_SIZE, reg_simd, 18, "mm0" },
+	{ REG_MMX_SIZE, reg_simd, 19, "mm1" },
+	{ REG_MMX_SIZE, reg_simd, 20, "mm2" },
+	{ REG_MMX_SIZE, reg_simd, 21, "mm3" },
+	{ REG_MMX_SIZE, reg_simd, 22, "mm4" },
+	{ REG_MMX_SIZE, reg_simd, 23, "mm5" },
+	{ REG_MMX_SIZE, reg_simd, 24, "mm6" },
+	{ REG_MMX_SIZE, reg_simd, 25, "mm7" },
+	/* REG_SIMD_OFFSET */
+	{ REG_SIMD_SIZE, reg_simd, 0, "xmm0" },
+	{ REG_SIMD_SIZE, reg_simd, 0, "xmm1" },
+	{ REG_SIMD_SIZE, reg_simd, 0, "xmm2" },
+	{ REG_SIMD_SIZE, reg_simd, 0, "xmm3" },
+	{ REG_SIMD_SIZE, reg_simd, 0, "xmm4" },
+	{ REG_SIMD_SIZE, reg_simd, 0, "xmm5" },
+	{ REG_SIMD_SIZE, reg_simd, 0, "xmm6" },
+	{ REG_SIMD_SIZE, reg_simd, 0, "xmm7" },
+	/* REG_DEBUG_OFFSET */
+	{ REG_DEBUG_SIZE, reg_sys, 0, "dr0" },
+	{ REG_DEBUG_SIZE, reg_sys, 0, "dr1" },
+	{ REG_DEBUG_SIZE, reg_sys, 0, "dr2" },
+	{ REG_DEBUG_SIZE, reg_sys, 0, "dr3" },
+	{ REG_DEBUG_SIZE, reg_sys, 0, "dr4" },
+	{ REG_DEBUG_SIZE, reg_sys, 0, "dr5" },
+	{ REG_DEBUG_SIZE, reg_sys, 0, "dr6" },
+	{ REG_DEBUG_SIZE, reg_sys, 0, "dr7" },
+	/* REG_CTRL_OFFSET */
+	{ REG_CTRL_SIZE, reg_sys, 0, "cr0" },
+	{ REG_CTRL_SIZE, reg_sys, 0, "cr1" },
+	{ REG_CTRL_SIZE, reg_sys, 0, "cr2" },
+	{ REG_CTRL_SIZE, reg_sys, 0, "cr3" },
+	{ REG_CTRL_SIZE, reg_sys, 0, "cr4" },
+	{ REG_CTRL_SIZE, reg_sys, 0, "cr5" },
+	{ REG_CTRL_SIZE, reg_sys, 0, "cr6" },
+	{ REG_CTRL_SIZE, reg_sys, 0, "cr7" },
+	/* REG_TEST_OFFSET */
+	{ REG_TEST_SIZE, reg_sys, 0, "tr0" },
+	{ REG_TEST_SIZE, reg_sys, 0, "tr1" },
+	{ REG_TEST_SIZE, reg_sys, 0, "tr2" },
+	{ REG_TEST_SIZE, reg_sys, 0, "tr3" },
+	{ REG_TEST_SIZE, reg_sys, 0, "tr4" },
+	{ REG_TEST_SIZE, reg_sys, 0, "tr5" },
+	{ REG_TEST_SIZE, reg_sys, 0, "tr6" },
+	{ REG_TEST_SIZE, reg_sys, 0, "tr7" },
+	/* REG_SEG_OFFSET */
+	{ REG_SEG_SIZE, reg_seg, 0, "es" },
+	{ REG_SEG_SIZE, reg_seg, 0, "cs" },
+	{ REG_SEG_SIZE, reg_seg, 0, "ss" },
+	{ REG_SEG_SIZE, reg_seg, 0, "ds" },
+	{ REG_SEG_SIZE, reg_seg, 0, "fs" },
+	{ REG_SEG_SIZE, reg_seg, 0, "gs" },
+	/* REG_LDTR_INDEX */
+	{ REG_DWORD_SIZE, reg_sys, 0, "ldtr" },
+	/* REG_GDTR_INDEX */
+	{ REG_DWORD_SIZE, reg_sys, 0, "gdtr" },
+	/* REG_FPU_OFFSET */
+	{ REG_FPU_SIZE, reg_fpu, 0, "st(0)" },
+	{ REG_FPU_SIZE, reg_fpu, 0, "st(1)" },
+	{ REG_FPU_SIZE, reg_fpu, 0, "st(2)" },
+	{ REG_FPU_SIZE, reg_fpu, 0, "st(3)" },
+	{ REG_FPU_SIZE, reg_fpu, 0, "st(4)" },
+	{ REG_FPU_SIZE, reg_fpu, 0, "st(5)" },
+	{ REG_FPU_SIZE, reg_fpu, 0, "st(6)" },
+	{ REG_FPU_SIZE, reg_fpu, 0, "st(7)" },
+	/* REG_FLAGS_INDEX : 81 */
+	{ REG_FLAGS_SIZE, reg_cond, 0, "eflags" }, 
+	/* REG_FPCTRL_INDEX  : 82*/
+	{ REG_FPCTRL_SIZE, reg_fpu | reg_sys, 0, "fpctrl" }, 
+	/* REG_FPSTATUS_INDEX : 83*/
+	{ REG_FPSTATUS_SIZE, reg_fpu | reg_sys, 0, "fpstat" },
+	/* REG_FPTAG_INDEX : 84 */
+	{ REG_FPTAG_SIZE, reg_fpu | reg_sys, 0, "fptag" }, 
+	/* REG_EIP_INDEX : 85 */
+	{ REG_EIP_SIZE, reg_pc, 0, "eip" },
+	/* REG_IP_INDEX : 86 */
+	{ REG_IP_SIZE, reg_pc, 17, "ip" },
+	/* REG_IDTR_INDEX : 87 */
+	{ REG_DWORD_SIZE, reg_sys, 0, "idtr" },
+	/* REG_MXCSG_INDEX : SSE Control Reg : 88 */
+	{ REG_DWORD_SIZE, reg_sys | reg_simd, 0, "mxcsr" },
+	/* REG_TR_INDEX : Task Register : 89 */
+	{ 16 + 64, reg_sys, 0, "tr" },
+	/* REG_CSMSR_INDEX : SYSENTER_CS_MSR : 90 */
+	{ REG_DWORD_SIZE, reg_sys, 0, "cs_msr" },
+	/* REG_ESPMSR_INDEX : SYSENTER_ESP_MSR : 91 */
+	{ REG_DWORD_SIZE, reg_sys, 0, "esp_msr" },
+	/* REG_EIPMSR_INDEX : SYSENTER_EIP_MSR : 92 */
+	{ REG_DWORD_SIZE, reg_sys, 0, "eip_msr" },
+	{ 0 }
+ };
+
+
+static size_t sz_regtable = NUM_X86_REGS + 1;
+
+
+void ia32_handle_register( x86_reg_t *reg, size_t id ) {
+	unsigned int alias;
+	if (! id || id > sz_regtable ) {
+		return;
+	}
+
+	memset( reg, 0, sizeof(x86_reg_t) );
+
+        strncpy( reg->name, ia32_reg_table[id].mnemonic, MAX_REGNAME );
+
+        reg->type = ia32_reg_table[id].type;
+        reg->size = ia32_reg_table[id].size;
+
+	alias = ia32_reg_table[id].alias;
+	if ( alias ) {
+		reg->alias = ia32_reg_aliases[alias].alias;
+		reg->shift = ia32_reg_aliases[alias].shift;
+	}
+        reg->id = id;
+
+	return;
+}
+
+size_t ia32_true_register_id( size_t id ) {
+	size_t reg;
+
+	if (! id || id > sz_regtable ) {
+		return 0;
+	}
+
+	reg = id;
+	if (ia32_reg_table[reg].alias) {
+		reg = ia32_reg_aliases[ia32_reg_table[reg].alias].alias;
+	}
+	return reg;
+}

libdisasm/ia32_reg.h

@@ -0,0 +1,41 @@
+#ifndef IA32_REG_H
+#define IA32_REG_H
+
+#include <sys/types.h>	/* for size_t */
+#include "libdis.h"	/* for x86_reg_t */
+
+/* NOTE these are used in opcode tables for hard-coded registers */
+#define REG_DWORD_OFFSET 	 1	/* 0 + 1 */
+#define REG_ECX_INDEX		 2	/* 0 + 1 + 1 */
+#define REG_ESP_INDEX		 5	/* 0 + 4 + 1 */
+#define REG_EBP_INDEX		 6	/* 0 + 5 + 1 */
+#define REG_ESI_INDEX		 7	/* 0 + 6 + 1 */
+#define REG_EDI_INDEX		 8	/* 0 + 7 + 1 */
+#define REG_WORD_OFFSET 	 9	/* 1 * 8 + 1 */
+#define REG_BYTE_OFFSET 	17	/* 2 * 8 + 1 */
+#define REG_MMX_OFFSET 		25	/* 3 * 8 + 1 */
+#define REG_SIMD_OFFSET 	33	/* 4 * 8 + 1 */
+#define REG_DEBUG_OFFSET 	41	/* 5 * 8 + 1 */
+#define REG_CTRL_OFFSET 	49	/* 6 * 8 + 1 */
+#define REG_TEST_OFFSET 	57	/* 7 * 8 + 1 */
+#define REG_SEG_OFFSET 		65	/* 8 * 8 + 1 */
+#define REG_LDTR_INDEX		71	/* 8 * 8 + 1 + 1 */
+#define REG_GDTR_INDEX		72	/* 8 * 8 + 2 + 1 */
+#define REG_FPU_OFFSET 		73	/* 9 * 8 + 1 */
+#define REG_FLAGS_INDEX 	81	/* 10 * 8 + 1 */
+#define REG_FPCTRL_INDEX 	82	/* 10 * 8 + 1 + 1 */
+#define REG_FPSTATUS_INDEX 	83	/* 10 * 8 + 2 + 1 */
+#define REG_FPTAG_INDEX 	84	/* 10 * 8 + 3 + 1 */
+#define REG_EIP_INDEX 		85	/* 10 * 8 + 4 + 1 */
+#define REG_IP_INDEX 		86	/* 10 * 8 + 5 + 1 */
+#define REG_IDTR_INDEX		87	/* 10 * 8 + 6 + 1 */
+#define REG_MXCSG_INDEX		88	/* 10 * 8 + 7 + 1 */
+#define REG_TR_INDEX		89	/* 10 * 8 + 8 + 1 */
+#define REG_CSMSR_INDEX		90	/* 10 * 8 + 9 + 1 */
+#define REG_ESPMSR_INDEX	91	/* 10 * 8 + 10 + 1 */
+#define REG_EIPMSR_INDEX	92	/* 10 * 8 + 11 + 1 */
+
+void ia32_handle_register( x86_reg_t *reg, size_t id );
+size_t ia32_true_register_id( size_t id );
+
+#endif

libdisasm/ia32_settings.c

@@ -0,0 +1,13 @@
+#include "libdis.h"
+#include "ia32_settings.h"
+#include "ia32_reg.h"
+#include "ia32_insn.h"
+
+ia32_settings_t ia32_settings = {
+	1, 0xF4, 
+	MAX_INSTRUCTION_SIZE,
+	4, 4, 8, 4, 8,
+	REG_ESP_INDEX, REG_EBP_INDEX, REG_EIP_INDEX, REG_FLAGS_INDEX,
+	REG_DWORD_OFFSET, REG_SEG_OFFSET, REG_FPU_OFFSET,
+	opt_none
+};

libdisasm/ia32_settings.h

@@ -0,0 +1,27 @@
+#ifndef IA32_SETTINGS_H
+#define IA32_SETTINGS_H
+
+#include "libdis.h"
+
+typedef struct {
+	/* options */
+	unsigned char endian,		/* 0 = big, 1 = little */
+		      wc_byte,		/* wildcard byte */
+		      max_insn,		/* max insn size */
+		      sz_addr,		/* default address size */
+		      sz_oper,		/* default operand size */
+		      sz_byte,		/* # bits in byte */
+		      sz_word,		/* # bytes in machine word */
+		      sz_dword;		/* # bytes in machine dword */
+	unsigned int id_sp_reg,		/* id of stack pointer */
+		     id_fp_reg,		/* id of frame pointer */
+		     id_ip_reg,		/* id of instruction pointer */
+		     id_flag_reg,	/* id of flags register */
+		     offset_gen_regs,	/* start of general regs */
+		     offset_seg_regs,	/* start of segment regs */
+		     offset_fpu_regs;	/* start of floating point regs */
+	/* user-controlled settings */
+	enum x86_options options;
+} ia32_settings_t;
+
+#endif

libdisasm/libdis.h

@@ -0,0 +1,836 @@
+#ifndef LIBDISASM_H
+#define LIBDISASM_H
+
+#ifdef WIN32
+#include <windows.h>
+#endif
+
+#include <stdint.h>
+
+/* 'NEW" types
+ * __________________________________________________________________________*/
+#ifndef LIBDISASM_QWORD_H       /* do not interfere with qword.h */
+        #define LIBDISASM_QWORD_H
+        #ifdef _MSC_VER
+                typedef __int64         qword_t;
+        #else
+                typedef int64_t         qword_t;
+        #endif
+#endif
+
+#include <sys/types.h>
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/* 'NEW" x86 API
+ * __________________________________________________________________________*/
+
+
+/* ========================================= Error Reporting */
+/* REPORT CODES
+ *      These are passed to a reporter function passed at initialization.
+ *      Each code determines the type of the argument passed to the reporter;
+ *      this allows the report to recover from errors, or just log them.
+ */
+enum x86_report_codes {
+        report_disasm_bounds,   /* RVA OUT OF BOUNDS : The disassembler could
+                                   not disassemble the supplied RVA as it is
+                                   out of the range of the buffer. The
+                                   application should store the address and
+                                   attempt to determine what section of the
+                                   binary it is in, then disassemble the
+                                   address from the bytes in that section.
+                                        data: uint32_t rva */
+        report_insn_bounds,     /* INSTRUCTION OUT OF BOUNDS: The disassembler
+                                   could not disassemble the instruction as
+                                   the instruction would require bytes beyond
+                                   the end of the current buffer. This usually
+                                   indicated garbage bytes at the end of a
+                                   buffer, or an incorrectly-sized buffer.
+                                        data: uint32_t rva */
+        report_invalid_insn,    /* INVALID INSTRUCTION: The disassembler could
+                                   not disassemble the instruction as it has an
+                                   invalid combination of opcodes and operands.
+                                   This will stop automated disassembly; the
+                                   application can restart the disassembly
+                                   after the invalid instruction.
+                                        data: uint32_t rva */
+        report_unknown
+};
+
+/* 'arg' is optional arbitrary data provided by the code passing the 
+ *       callback -- for example, it could be 'this' or 'self' in OOP code.
+ * 'code' is provided by libdisasm, it is one of the above
+ * 'data' is provided by libdisasm and is context-specific, per the enums */
+typedef void (*DISASM_REPORTER)( enum x86_report_codes code, 
+				 void *data, void *arg );
+
+
+/* x86_report_error : Call the register reporter to report an error */
+void x86_report_error( enum x86_report_codes code, void *data );
+
+/* ========================================= Libdisasm Management Routines */
+enum x86_options {		/* these can be ORed together */
+        opt_none= 0,
+        opt_ignore_nulls=1,     /* ignore sequences of > 4 NULL bytes */
+        opt_16_bit=2,           /* 16-bit/DOS disassembly */
+        opt_att_mnemonics=4,    /* use AT&T syntax names for alternate opcode mnemonics */
+};
+
+/* management routines */
+/* 'arg' is caller-specific data which is passed as the first argument
+ * to the reporter callback routine */
+int x86_init( enum x86_options options, DISASM_REPORTER reporter, void *arg);
+void x86_set_reporter( DISASM_REPORTER reporter, void *arg);
+void x86_set_options( enum x86_options options );
+enum x86_options x86_get_options( void );
+int x86_cleanup(void);
+
+
+/* ========================================= Instruction Representation */
+/* these defines are only intended for use in the array decl's */
+#define MAX_REGNAME 8
+
+#define MAX_PREFIX_STR 32
+#define MAX_MNEM_STR 16
+#define MAX_INSN_SIZE 20        /* same as in i386.h */
+#define MAX_OP_STRING 32        /* max possible operand size in string form */
+#define MAX_OP_RAW_STRING 64    /* max possible operand size in raw form */
+#define MAX_OP_XML_STRING 256   /* max possible operand size in xml form */
+#define MAX_NUM_OPERANDS 8	/* max # implicit and explicit operands */
+/* in these, the '2 *' is arbitrary: the max # of operands should require
+ * more space than the rest of the insn */
+#define MAX_INSN_STRING 512        /* 2 * 8 * MAX_OP_STRING */
+#define MAX_INSN_RAW_STRING 1024   /* 2 * 8 * MAX_OP_RAW_STRING */
+#define MAX_INSN_XML_STRING 4096   /* 2 * 8 * MAX_OP_XML_STRING */
+
+enum x86_reg_type {     /* NOTE: these may be ORed together */
+        reg_gen         = 0x00001,      /* general purpose */
+        reg_in          = 0x00002,      /* incoming args, ala RISC */
+        reg_out         = 0x00004,      /* args to calls, ala RISC */
+        reg_local       = 0x00008,      /* local vars, ala RISC */
+        reg_fpu         = 0x00010,      /* FPU data register */
+        reg_seg         = 0x00020,      /* segment register */
+        reg_simd        = 0x00040,      /* SIMD/MMX reg */
+        reg_sys         = 0x00080,      /* restricted/system register */
+        reg_sp          = 0x00100,      /* stack pointer */
+        reg_fp          = 0x00200,      /* frame pointer */
+        reg_pc          = 0x00400,      /* program counter */
+        reg_retaddr     = 0x00800,      /* return addr for func */
+        reg_cond        = 0x01000,      /* condition code / flags */
+        reg_zero        = 0x02000,      /* zero register, ala RISC */
+        reg_ret         = 0x04000,      /* return value */
+        reg_src         = 0x10000,      /* array/rep source */
+        reg_dest        = 0x20000,      /* array/rep destination */
+        reg_count       = 0x40000       /* array/rep/loop counter */
+};
+
+/* x86_reg_t : an X86 CPU register */
+typedef struct {
+        char name[MAX_REGNAME];
+        enum x86_reg_type type;         /* what register is used for */
+        unsigned int size;              /* size of register in bytes */
+        unsigned int id;                /* register ID #, for quick compares */
+	unsigned int alias;		/* ID of reg this is an alias for */
+	unsigned int shift;		/* amount to shift aliased reg by */
+} x86_reg_t;
+
+/* x86_ea_t : an X86 effective address (address expression) */
+typedef struct {
+        unsigned int     scale;         /* scale factor */
+        x86_reg_t        index, base;   /* index, base registers */
+        int32_t          disp;          /* displacement */
+        char             disp_sign;     /* is negative? 1/0 */
+        char             disp_size;     /* 0, 1, 2, 4 */
+} x86_ea_t;
+
+/* x86_absolute_t : an X86 segment:offset address (descriptor) */
+typedef struct {
+	unsigned short	segment;	/* loaded directly into CS */
+	union {
+		unsigned short	off16;	/* loaded directly into IP */
+		uint32_t		off32;	/* loaded directly into EIP */
+	} offset;	
+} x86_absolute_t;
+
+enum x86_op_type {      /* mutually exclusive */
+        op_unused = 0,          /* empty/unused operand: should never occur */
+        op_register = 1,        /* CPU register */
+        op_immediate = 2,       /* Immediate Value */
+        op_relative_near = 3,   /* Relative offset from IP */
+        op_relative_far = 4,    /* Relative offset from IP */
+        op_absolute = 5,        /* Absolute address (ptr16:32) */
+        op_expression = 6,      /* Address expression (scale/index/base/disp) */
+        op_offset = 7,          /* Offset from start of segment (m32) */
+        op_unknown
+};
+
+#define x86_optype_is_address( optype ) \
+	( optype == op_absolute || optype == op_offset )
+#define x86_optype_is_relative( optype ) \
+	( optype == op_relative_near || optype == op_relative_far )
+#define x86_optype_is_memory( optype ) \
+	( optype > op_immediate && optype < op_unknown )
+
+enum x86_op_datatype {          /* these use Intel's lame terminology */
+        op_byte = 1,            /* 1 byte integer */
+        op_word = 2,            /* 2 byte integer */
+        op_dword = 3,           /* 4 byte integer */
+        op_qword = 4,           /* 8 byte integer */
+        op_dqword = 5,          /* 16 byte integer */
+        op_sreal = 6,           /* 4 byte real (single real) */
+        op_dreal = 7,           /* 8 byte real (double real) */
+        op_extreal = 8,         /* 10 byte real (extended real) */
+        op_bcd = 9,             /* 10 byte binary-coded decimal */
+        op_ssimd = 10,          /* 16 byte : 4 packed single FP (SIMD, MMX) */
+        op_dsimd = 11,          /* 16 byte : 2 packed double FP (SIMD, MMX) */
+        op_sssimd = 12,         /* 4 byte : scalar single FP (SIMD, MMX) */
+        op_sdsimd = 13,         /* 8 byte : scalar double FP (SIMD, MMX) */
+	op_descr32 = 14,	/* 6 byte Intel descriptor 2:4 */
+	op_descr16 = 15,	/* 4 byte Intel descriptor 2:2 */
+	op_pdescr32 = 16,	/* 6 byte Intel pseudo-descriptor 32:16 */
+	op_pdescr16 = 17,	/* 6 byte Intel pseudo-descriptor 8:24:16 */
+	op_bounds16 = 18,	/* signed 16:16 lower:upper bounds */
+	op_bounds32 = 19,	/* signed 32:32 lower:upper bounds */
+        op_fpuenv16 = 20,	/* 14 byte FPU control/environment data */
+        op_fpuenv32 = 21,	/* 28 byte FPU control/environment data */
+	op_fpustate16 = 22,	/* 94 byte FPU state (env & reg stack) */
+	op_fpustate32 = 23,	/* 108 byte FPU state (env & reg stack) */
+	op_fpregset = 24,	/* 512 bytes: register set */
+	op_fpreg = 25,		/* FPU register */
+    op_none = 0xFF,     /* operand without a datatype (INVLPG) */
+};
+
+enum x86_op_access {    /* ORed together */
+        op_read = 1,
+        op_write = 2,
+        op_execute = 4
+};
+
+enum x86_op_flags {     /* ORed together, but segs are mutually exclusive */
+        op_signed = 1,          /* signed integer */
+        op_string = 2,          /* possible string or array */
+        op_constant = 4,        /* symbolic constant */
+        op_pointer = 8,         /* operand points to a memory address */
+	op_sysref = 0x010,	/* operand is a syscall number */
+	op_implied = 0x020,	/* operand is implicit in the insn */
+	op_hardcode = 0x40,	/* operand is hardcoded in insn definition */
+	/* NOTE: an 'implied' operand is one which can be considered a side
+	 * effect of the insn, e.g. %esp being modified by PUSH or POP. A
+	 * 'hard-coded' operand is one which is specified in the instruction
+	 * definition, e.g. %es:%edi in MOVSB or 1 in ROL Eb, 1. The difference
+	 * is that hard-coded operands are printed by disassemblers and are
+	 * required to re-assemble, while implicit operands are invisible. */
+        op_es_seg = 0x100,      /* ES segment override */
+        op_cs_seg = 0x200,      /* CS segment override */
+        op_ss_seg = 0x300,      /* SS segment override */
+        op_ds_seg = 0x400,      /* DS segment override */
+        op_fs_seg = 0x500,      /* FS segment override */
+        op_gs_seg = 0x600       /* GS segment override */
+};
+
+/* x86_op_t : an X86 instruction operand */
+typedef struct {
+        enum x86_op_type        type;           /* operand type */
+        enum x86_op_datatype    datatype;       /* operand size */
+        enum x86_op_access      access;         /* operand access [RWX] */
+        enum x86_op_flags       flags;          /* misc flags */
+        union {
+		/* sizeof will have to work on these union members! */
+                /* immediate values */
+                char            sbyte;
+                short           sword;
+                int32_t         sdword;
+                qword_t         sqword;
+                unsigned char   byte;
+                unsigned short  word;
+                uint32_t        dword;
+                qword_t         qword;
+                float           sreal;
+                double          dreal;
+                /* misc large/non-native types */
+                unsigned char   extreal[10];
+                unsigned char   bcd[10];
+                qword_t         dqword[2];
+                unsigned char   simd[16];
+                unsigned char   fpuenv[28];
+                /* offset from segment */
+                uint32_t        offset;
+                /* ID of CPU register */
+                x86_reg_t       reg;
+                /* offsets from current insn */
+                char            relative_near;
+                int32_t         relative_far;
+		/* segment:offset */
+		x86_absolute_t	absolute;
+                /* effective address [expression] */
+                x86_ea_t        expression;
+        } data;
+	/* this is needed to make formatting operands more sane */
+	void * insn;		/* pointer to x86_insn_t owning operand */
+} x86_op_t;
+
+/* Linked list of x86_op_t; provided for manual traversal of the operand
+ * list in an insn. Users wishing to add operands to this list, e.g. to add
+ * implicit operands, should use x86_operand_new in x86_operand_list.h */
+typedef struct x86_operand_list {
+	x86_op_t op;
+	struct x86_operand_list *next;
+} x86_oplist_t;
+
+enum x86_insn_group {
+	insn_none = 0,		/* invalid instruction */
+        insn_controlflow = 1,
+        insn_arithmetic = 2,
+        insn_logic = 3,
+        insn_stack = 4,
+        insn_comparison = 5,
+        insn_move = 6,
+        insn_string = 7,
+        insn_bit_manip = 8,
+        insn_flag_manip = 9,
+        insn_fpu = 10,
+        insn_interrupt = 13,
+        insn_system = 14,
+        insn_other = 15
+};
+
+enum x86_insn_type {
+	insn_invalid = 0,	/* invalid instruction */
+        /* insn_controlflow */
+        insn_jmp = 0x1001,
+        insn_jcc = 0x1002,
+        insn_call = 0x1003,
+        insn_callcc = 0x1004,
+        insn_return = 0x1005,
+        /* insn_arithmetic */
+        insn_add = 0x2001,
+        insn_sub = 0x2002,
+        insn_mul = 0x2003,
+        insn_div = 0x2004,
+        insn_inc = 0x2005,
+        insn_dec = 0x2006,
+        insn_shl = 0x2007,
+        insn_shr = 0x2008,
+        insn_rol = 0x2009,
+        insn_ror = 0x200A,
+        /* insn_logic */
+        insn_and = 0x3001,
+        insn_or = 0x3002,
+        insn_xor = 0x3003,
+        insn_not = 0x3004,
+        insn_neg = 0x3005,
+        /* insn_stack */
+        insn_push = 0x4001,
+        insn_pop = 0x4002,
+        insn_pushregs = 0x4003,
+        insn_popregs = 0x4004,
+        insn_pushflags = 0x4005,
+        insn_popflags = 0x4006,
+        insn_enter = 0x4007,
+        insn_leave = 0x4008,
+        /* insn_comparison */
+        insn_test = 0x5001,
+        insn_cmp = 0x5002,
+        /* insn_move */
+        insn_mov = 0x6001,      /* move */
+        insn_movcc = 0x6002,    /* conditional move */
+        insn_xchg = 0x6003,     /* exchange */
+        insn_xchgcc = 0x6004,   /* conditional exchange */
+        /* insn_string */
+        insn_strcmp = 0x7001,
+        insn_strload = 0x7002,
+        insn_strmov = 0x7003,
+        insn_strstore = 0x7004,
+        insn_translate = 0x7005,        /* xlat */
+        /* insn_bit_manip */
+        insn_bittest = 0x8001,
+        insn_bitset = 0x8002,
+        insn_bitclear = 0x8003,
+        /* insn_flag_manip */
+        insn_clear_carry = 0x9001,
+        insn_clear_zero = 0x9002,
+        insn_clear_oflow = 0x9003,
+        insn_clear_dir = 0x9004,
+        insn_clear_sign = 0x9005,
+        insn_clear_parity = 0x9006,
+        insn_set_carry = 0x9007,
+        insn_set_zero = 0x9008,
+        insn_set_oflow = 0x9009,
+        insn_set_dir = 0x900A,
+        insn_set_sign = 0x900B,
+        insn_set_parity = 0x900C,
+        insn_tog_carry = 0x9010,
+        insn_tog_zero = 0x9020,
+        insn_tog_oflow = 0x9030,
+        insn_tog_dir = 0x9040,
+        insn_tog_sign = 0x9050,
+        insn_tog_parity = 0x9060,
+        /* insn_fpu */
+        insn_fmov = 0xA001,
+        insn_fmovcc = 0xA002,
+        insn_fneg = 0xA003,
+        insn_fabs = 0xA004,
+        insn_fadd = 0xA005,
+        insn_fsub = 0xA006,
+        insn_fmul = 0xA007,
+        insn_fdiv = 0xA008,
+        insn_fsqrt = 0xA009,
+        insn_fcmp = 0xA00A,
+        insn_fcos = 0xA00C,
+        insn_fldpi = 0xA00D,
+        insn_fldz = 0xA00E,
+        insn_ftan = 0xA00F,
+        insn_fsine = 0xA010,
+        insn_fsys = 0xA020,
+        /* insn_interrupt */
+        insn_int = 0xD001,
+        insn_intcc = 0xD002,    /* not present in x86 ISA */
+        insn_iret = 0xD003,
+        insn_bound = 0xD004,
+        insn_debug = 0xD005,
+        insn_trace = 0xD006,
+        insn_invalid_op = 0xD007,
+        insn_oflow = 0xD008,
+        /* insn_system */
+        insn_halt = 0xE001,
+        insn_in = 0xE002,       /* input from port/bus */
+        insn_out = 0xE003,      /* output to port/bus */
+        insn_cpuid = 0xE004,
+        /* insn_other */
+        insn_nop = 0xF001,
+        insn_bcdconv = 0xF002,  /* convert to or from BCD */
+        insn_szconv = 0xF003    /* change size of operand */
+};
+
+/* These flags specify special characteristics of the instruction, such as
+ * whether the inatruction is privileged or whether it serializes the
+ * pipeline.
+ * NOTE : These may not be accurate for all instructions; updates to the
+ * opcode tables have not been completed. */
+enum x86_insn_note {
+	insn_note_ring0		= 1,	/* Only available in ring 0 */
+	insn_note_smm		= 2,	/* "" in System Management Mode */
+	insn_note_serial	= 4,	/* Serializing instruction */
+	insn_note_nonswap	= 8,	/* Does not swap arguments in att-style formatting */
+	insn_note_nosuffix  = 16,	/* Does not have size suffix in att-style formatting */
+};
+
+/* This specifies what effects the instruction has on the %eflags register */
+enum x86_flag_status {
+        insn_carry_set = 0x1,			/* CF */
+        insn_zero_set = 0x2,			/* ZF */
+        insn_oflow_set = 0x4,			/* OF */
+        insn_dir_set = 0x8,			/* DF */
+        insn_sign_set = 0x10,			/* SF */
+        insn_parity_set = 0x20,			/* PF */
+        insn_carry_or_zero_set = 0x40,
+        insn_zero_set_or_sign_ne_oflow = 0x80,
+        insn_carry_clear = 0x100,
+        insn_zero_clear = 0x200,
+        insn_oflow_clear = 0x400,
+        insn_dir_clear = 0x800,
+        insn_sign_clear = 0x1000,
+        insn_parity_clear = 0x2000,
+        insn_sign_eq_oflow = 0x4000,
+        insn_sign_ne_oflow = 0x8000
+};
+
+/* The CPU model in which the insturction first appeared; this can be used
+ * to mask out instructions appearing in earlier or later models or to
+ * check the portability of a binary.
+ * NOTE : These may not be accurate for all instructions; updates to the
+ * opcode tables have not been completed. */
+enum x86_insn_cpu {
+	cpu_8086 	= 1,	/* Intel */
+	cpu_80286	= 2,
+	cpu_80386	= 3,
+	cpu_80387	= 4,
+	cpu_80486	= 5,
+	cpu_pentium	= 6,
+	cpu_pentiumpro	= 7,
+	cpu_pentium2	= 8,
+	cpu_pentium3	= 9,
+	cpu_pentium4	= 10,
+	cpu_k6		= 16,	/* AMD */
+	cpu_k7		= 32,
+	cpu_athlon	= 48
+};
+
+/* CPU ISA subsets: These are derived from the Instruction Groups in
+ * Intel Vol 1 Chapter 5; they represent subsets of the IA32 ISA but
+ * do not reflect the 'type' of the instruction in the same way that
+ * x86_insn_group does. In short, these are AMD/Intel's somewhat useless 
+ * designations.
+ * NOTE : These may not be accurate for all instructions; updates to the
+ * opcode tables have not been completed. */
+enum x86_insn_isa {
+	isa_gp		= 1,	/* general purpose */
+	isa_fp		= 2,	/* floating point */
+	isa_fpumgt	= 3,	/* FPU/SIMD management */
+	isa_mmx		= 4,	/* Intel MMX */
+	isa_sse1	= 5,	/* Intel SSE SIMD */
+	isa_sse2	= 6,	/* Intel SSE2 SIMD */
+	isa_sse3	= 7,	/* Intel SSE3 SIMD */
+	isa_3dnow	= 8,	/* AMD 3DNow! SIMD */
+	isa_sys		= 9	/* system instructions */
+};
+
+enum x86_insn_prefix {
+        insn_no_prefix = 0,
+        insn_rep_zero = 1,	/* REPZ and REPE */
+        insn_rep_notzero = 2,	/* REPNZ and REPNZ */
+        insn_lock = 4		/* LOCK: */
+};
+
+/* TODO: maybe provide insn_new/free(), and have disasm return new insn_t */
+/* x86_insn_t : an X86 instruction */
+typedef struct {
+        /* information about the instruction */
+        uint32_t addr;             /* load address */
+        uint32_t offset;           /* offset into file/buffer */
+        enum x86_insn_group group;      /* meta-type, e.g. INS_EXEC */
+        enum x86_insn_type type;        /* type, e.g. INS_BRANCH */
+	enum x86_insn_note note;	/* note, e.g. RING0 */
+        unsigned char bytes[MAX_INSN_SIZE];
+        unsigned char size;             /* size of insn in bytes */
+	/* 16/32-bit mode settings */
+	unsigned char addr_size;	/* default address size : 2 or 4 */
+	unsigned char op_size;		/* default operand size : 2 or 4 */
+	/* CPU/instruction set */
+	enum x86_insn_cpu cpu;
+	enum x86_insn_isa isa;
+	/* flags */
+        enum x86_flag_status flags_set; /* flags set or tested by insn */
+        enum x86_flag_status flags_tested;
+	/* stack */
+	unsigned char stack_mod;	/* 0 or 1 : is the stack modified? */
+	int32_t stack_mod_val;		/* val stack is modified by if known */
+
+        /* the instruction proper */
+        enum x86_insn_prefix prefix;	/* prefixes ORed together */
+        char prefix_string[MAX_PREFIX_STR]; /* prefixes [might be truncated] */
+        char mnemonic[MAX_MNEM_STR];
+        x86_oplist_t *operands;		/* list of explicit/implicit operands */
+	size_t operand_count;		/* total number of operands */
+	size_t explicit_count;		/* number of explicit operands */
+        /* convenience fields for user */
+        void *block;                    /* code block containing this insn */
+        void *function;                 /* function containing this insn */
+        int tag;			/* tag the insn as seen/processed */
+} x86_insn_t;
+
+
+/* returns 0 if an instruction is invalid, 1 if valid */
+int x86_insn_is_valid( x86_insn_t *insn );
+
+/* DISASSEMBLY ROUTINES
+ *      Canonical order of arguments is
+ *        (buf, buf_len, buf_rva, offset, len, insn, func, arg, resolve_func)
+ *      ...but of course all of these are not used at the same time.
+ */
+
+
+/* Function prototype for caller-supplied callback routine
+ *      These callbacks are intended to process 'insn' further, e.g. by
+ *      adding it to a linked list, database, etc */
+typedef void (*DISASM_CALLBACK)( x86_insn_t *insn, void * arg );
+
+/* Function prototype for caller-supplied address resolver.
+ *      This routine is used to determine the rva to disassemble next, given
+ *      the 'dest' operand of a jump/call. This allows the caller to resolve
+ *      jump/call targets stored in a register or on the stack, and also allows
+ *      the caller to prevent endless loops by checking if an address has
+ *      already been disassembled. If an address cannot be resolved from the
+ *      operand, or if the address has already been disassembled, this routine
+ *      should return -1; in all other cases the RVA to be disassembled next
+ *      should be returned. */
+typedef int32_t (*DISASM_RESOLVER)( x86_op_t *op, x86_insn_t * current_insn,
+				 void *arg );
+
+
+/* x86_disasm: Disassemble a single instruction from a buffer of bytes.
+ *             Returns size of instruction in bytes.
+ *             Caller is responsible for calling x86_oplist_free() on
+ *             a reused "insn" to avoid leaking memory when calling this
+ *             function repeatedly.
+ *      buf     : Buffer of bytes to disassemble
+ *      buf_len : Length of the buffer
+ *      buf_rva : Load address of the start of the buffer
+ *      offset  : Offset in buffer to disassemble
+ *      insn    : Structure to fill with disassembled instruction
+ */
+unsigned int x86_disasm( unsigned char *buf, unsigned int buf_len,
+                	 uint32_t buf_rva, unsigned int offset,
+                	 x86_insn_t * insn );
+
+/* x86_disasm_range: Sequential disassembly of a range of bytes in a buffer,
+ *                   invoking a callback function each time an instruction
+ *                   is successfully disassembled. The 'range' refers to the
+ *                   bytes between 'offset' and 'offset + len' in the buffer;
+ *                   'len' is assumed to be less than the length of the buffer.
+ *                   Returns number of instructions processed.
+ *      buf     : Buffer of bytes to disassemble (e.g. .text section)
+ *      buf_rva : Load address of buffer (e.g. ELF Virtual Address)
+ *      offset  : Offset in buffer to start disassembly at
+ *      len     : Number of bytes to disassemble
+ *      func    : Callback function to invoke (may be NULL)
+ *      arg     : Arbitrary data to pass to callback (may be NULL)
+ */
+unsigned int x86_disasm_range( unsigned char *buf, uint32_t buf_rva,
+	                       unsigned int offset, unsigned int len,
+	                       DISASM_CALLBACK func, void *arg );
+
+/* x86_disasm_forward: Flow-of-execution disassembly of the bytes in a buffer,
+ *                     invoking a callback function each time an instruction
+ *                     is successfully disassembled.
+ *      buf     : Buffer to disassemble (e.g. .text section)
+ *      buf_len : Number of bytes in buffer
+ *      buf_rva : Load address of buffer (e.g. ELF Virtual Address)
+ *      offset  : Offset in buffer to start disassembly at (e.g. entry point)
+ *      func    : Callback function to invoke (may be NULL)
+ *      arg     : Arbitrary data to pass to callback (may be NULL)
+ *      resolver: Caller-supplied address resolver. If no resolver is
+ *                supplied, a default internal one is used -- however the
+ *                internal resolver does NOT catch loops and could end up
+ *                disassembling forever..
+ *      r_arg	: Arbitrary data to pass to resolver (may be NULL)
+ */
+unsigned int x86_disasm_forward( unsigned char *buf, unsigned int buf_len,
+	                         uint32_t buf_rva, unsigned int offset,
+	                         DISASM_CALLBACK func, void *arg,
+	                         DISASM_RESOLVER resolver, void *r_arg );
+
+/* Instruction operands: these are stored as a list of explicit and
+ * implicit operands. It is recommended that the 'foreach' routines
+ * be used to when examining operands for purposes of data flow analysis */
+
+/* Operand FOREACH callback: 'arg' is an abritrary parameter passed to the
+ * foreach routine, 'insn' is the x86_insn_t whose operands are being
+ * iterated over, and 'op' is the current x86_op_t */
+typedef void (*x86_operand_fn)(x86_op_t *op, x86_insn_t *insn, void *arg);
+
+/* FOREACH types: these are used to limit the foreach results to 
+ * operands which match a certain "type" (implicit or explicit)
+ * or which are accessed in certain ways (e.g. read or write). Note
+ * that this operates on the operand list of single instruction, so
+ * specifying the 'real' operand type (register, memory, etc) is not
+ * useful. Note also that by definition Execute Access implies Read
+ * Access and implies Not Write Access.
+ * The "type" (implicit or explicit) and the access method can
+ * be ORed together, e.g. op_wo | op_explicit */
+enum x86_op_foreach_type {
+	op_any 	= 0,		/* ALL operands (explicit, implicit, rwx) */
+	op_dest = 1,		/* operands with Write access */
+	op_src 	= 2,		/* operands with Read access */
+	op_ro 	= 3,		/* operands with Read but not Write access */
+	op_wo 	= 4,		/* operands with Write but not Read access */
+	op_xo 	= 5,		/* operands with Execute access */
+	op_rw 	= 6,		/* operands with Read AND Write access */
+	op_implicit = 0x10,	/* operands that are implied by the opcode */
+	op_explicit = 0x20	/* operands that are not side-effects */
+};
+
+
+/* free the operand list associated with an instruction -- useful for
+ * preventing memory leaks when free()ing an x86_insn_t */
+void x86_oplist_free( x86_insn_t *insn );
+
+/* Operand foreach: invokes 'func' with 'insn' and 'arg' as arguments. The
+ * 'type' parameter is used to select only operands matching specific
+ * criteria. */
+int x86_operand_foreach( x86_insn_t *insn, x86_operand_fn func, void *arg,
+	       	  	 enum x86_op_foreach_type type);
+
+/* convenience routine: returns count of operands matching 'type' */
+size_t x86_operand_count( x86_insn_t *insn, enum x86_op_foreach_type type );
+
+/* accessor functions for the operands */
+x86_op_t * x86_operand_1st( x86_insn_t *insn );
+x86_op_t * x86_operand_2nd( x86_insn_t *insn );
+x86_op_t * x86_operand_3rd( x86_insn_t *insn );
+
+/* these allow libdisasm 2.0 accessor functions to still be used */
+#define x86_get_dest_operand( insn ) x86_operand_1st( insn )
+#define x86_get_src_operand( insn ) x86_operand_2nd( insn )
+#define x86_get_imm_operand( insn ) x86_operand_3rd( insn )
+
+/* get size of operand data in bytes */
+unsigned int x86_operand_size( x86_op_t *op );
+
+/* Operand Convenience Routines: the following three routines are common
+ * operations on operands, intended to ease the burden of the programmer. */
+
+/* Get Address: return the value of an offset operand, or the offset of
+ * a segment:offset absolute address */
+uint32_t x86_get_address( x86_insn_t *insn );
+
+/* Get Relative Offset: return as a sign-extended int32_t the near or far
+ * relative offset operand, or 0 if there is none. There can be only one
+ * relaive offset operand in an instruction. */
+int32_t x86_get_rel_offset( x86_insn_t *insn );
+
+/* Get Branch Target: return the x86_op_t containing the target of
+ * a jump or call operand, or NULL if there is no branch target. 
+ * Internally, a 'branch target' is defined as any operand with
+ * Execute Access set. There can be only one branch target per instruction. */
+x86_op_t * x86_get_branch_target( x86_insn_t *insn );
+
+/* Get Immediate: return the x86_op_t containing the immediate operand
+ * for this instruction, or NULL if there is no immediate operand. There
+ * can be only one immediate operand per instruction */
+x86_op_t * x86_get_imm( x86_insn_t *insn );
+
+/* Get Raw Immediate Data: returns a pointer to the immediate data encoded
+ * in the instruction. This is useful for large data types [>32 bits] currently
+ * not supported by libdisasm, or for determining if the disassembler
+ * screwed up the conversion of the immediate data. Note that 'imm' in this
+ * context refers to immediate data encoded at the end of an instruction as
+ * detailed in the Intel Manual Vol II Chapter 2; it does not refer to the
+ * 'op_imm' operand (the third operand in instructions like 'mul' */
+unsigned char * x86_get_raw_imm( x86_insn_t *insn );
+
+
+/* More accessor fuctions, this time for user-defined info... */
+/* set the address (usually RVA) of the insn */
+void x86_set_insn_addr( x86_insn_t *insn, uint32_t addr );
+
+/* set the offset (usually offset into file) of the insn */
+void x86_set_insn_offset( x86_insn_t *insn, unsigned int offset );
+
+/* set a pointer to the function owning the instruction. The 
+ * type of 'func' is user-defined; libdisasm does not use the func field. */
+void x86_set_insn_function( x86_insn_t *insn, void * func );
+
+/* set a pointer to the block of code owning the instruction. The 
+ * type of 'block' is user-defined; libdisasm does not use the block field. */
+void x86_set_insn_block( x86_insn_t *insn, void * block );
+
+/* instruction tagging: these routines allow the programmer to mark
+ * instructions as "seen" in a DFS, for example. libdisasm does not use
+ * the tag field.*/ 
+/* set insn->tag to 1 */
+void x86_tag_insn( x86_insn_t *insn );
+/* set insn->tag to 0 */
+void x86_untag_insn( x86_insn_t *insn );
+/* return insn->tag */
+int x86_insn_is_tagged( x86_insn_t *insn );
+
+
+/* Disassembly formats:
+ *      AT&T is standard AS/GAS-style: "mnemonic\tsrc, dest, imm"
+ *      Intel is standard MASM/NASM/TASM: "mnemonic\tdest,src, imm"
+ *      Native is tab-delimited: "RVA\tbytes\tmnemonic\tdest\tsrc\timm"
+ *      XML is your typical <insn> ... </insn>
+ *      Raw is addr|offset|size|bytes|prefix... see libdisasm_formats.7
+ */
+enum x86_asm_format { 
+	unknown_syntax = 0,		/* never use! */
+	native_syntax, 			/* header: 35 bytes */
+	intel_syntax, 			/* header: 23 bytes */
+	att_syntax,  			/* header: 23 bytes */
+	xml_syntax,			/* header: 679 bytes */
+	raw_syntax			/* header: 172 bytes */
+};
+
+/* format (sprintf) an operand into 'buf' using specified syntax */
+int x86_format_operand(x86_op_t *op, char *buf, int len,
+                  enum x86_asm_format format);
+
+/* format (sprintf) an instruction mnemonic into 'buf' using specified syntax */
+int x86_format_mnemonic(x86_insn_t *insn, char *buf, int len,
+                        enum x86_asm_format format);
+
+/* format (sprintf) an instruction into 'buf' using specified syntax;
+ * this includes formatting all operands */
+int x86_format_insn(x86_insn_t *insn, char *buf, int len, enum x86_asm_format);
+
+/* fill 'buf' with a description of the format's syntax */
+int x86_format_header( char *buf, int len, enum x86_asm_format format);
+
+/* Endianness of an x86 CPU : 0 is big, 1 is little; always returns 1 */
+unsigned int x86_endian(void);
+
+/* Default address and operand size in bytes */
+unsigned int x86_addr_size(void);
+unsigned int x86_op_size(void);
+
+/* Size of a machine word in bytes */
+unsigned int x86_word_size(void);
+
+/* maximum size of a code instruction */
+#define x86_max_inst_size(x) x86_max_insn_size(x)
+unsigned int x86_max_insn_size(void);
+
+/* register IDs of Stack, Frame, Instruction pointer and Flags register */
+unsigned int x86_sp_reg(void);
+unsigned int x86_fp_reg(void);
+unsigned int x86_ip_reg(void);
+unsigned int x86_flag_reg(void);
+
+/* fill 'reg' struct with details of register 'id' */
+void x86_reg_from_id( unsigned int id, x86_reg_t * reg );
+
+/* convenience macro demonstrating how to get an aliased register; proto is
+ *   void x86_get_aliased_reg( x86_reg_t *alias_reg, x86_reg_t *output_reg )
+ * where 'alias_reg' is a reg operand and 'output_reg' is filled with the
+ * register that the operand is an alias for */
+#define x86_get_aliased_reg( alias_reg, output_reg )			\
+	x86_reg_from_id( alias_reg->alias, output_reg )
+
+
+/* ================================== Invariant Instruction Representation */
+/* Invariant instructions are used for generating binary signatures; 
+ * the instruction is modified so that all variant bytes in an instruction
+ * are replaced with a wildcard byte. 
+ *
+ * A 'variant byte' is one that is expected to be modified by either the 
+ * static or the dynamic linker: for example, an address encoded in an 
+ * instruction. 
+ *
+ * By comparing the invariant representation of one instruction [or of a
+ * sequence of instructions] with the invariant representation of another,
+ * one determine whether the two invariant representations are from the same
+ * relocatable object [.o] file. Thus one can use binary signatures [which
+ * are just sequences of invariant instruction representations] to look for
+ * library routines which have been statically-linked into a binary.
+ *
+ * The invariant routines are faster and smaller than the disassembly
+ * routines; they can be used to determine the size of an instruction 
+ * without all of the overhead of a full instruction disassembly.
+ */
+
+/* This byte is used to replace variant bytes */
+#define X86_WILDCARD_BYTE 0xF4
+
+typedef struct {
+        enum x86_op_type        type;           /* operand type */
+        enum x86_op_datatype    datatype;       /* operand size */
+        enum x86_op_access      access;         /* operand access [RWX] */
+        enum x86_op_flags       flags;          /* misc flags */
+} x86_invariant_op_t;
+
+typedef struct {
+	unsigned char bytes[64];	/* invariant representation */
+	unsigned int  size;		/* number of bytes in insn */
+        enum x86_insn_group group;      /* meta-type, e.g. INS_EXEC */
+        enum x86_insn_type type;        /* type, e.g. INS_BRANCH */
+	x86_invariant_op_t operands[3];	/* operands: dest, src, imm */
+} x86_invariant_t;
+ 
+
+/* return a version of the instruction with the variant bytes masked out */
+size_t x86_invariant_disasm( unsigned char *buf, int buf_len, 
+			  x86_invariant_t *inv );
+/* return the size in bytes of the intruction pointed to by 'buf';
+ * this used x86_invariant_disasm since it faster than x86_disasm */
+size_t x86_size_disasm( unsigned char *buf, unsigned int buf_len );
+
+#ifdef __cplusplus
+}
+#endif
+
+
+#endif

libdisasm/qword.h

@@ -0,0 +1,14 @@
+#ifndef LIBDISASM_QWORD_H
+#define LIBDISASM_QWORD_H
+
+#include <stdint.h>
+
+/* platform independent data types */
+
+#ifdef _MSC_VER
+	typedef __int64         qword_t;
+#else
+	typedef int64_t         qword_t;
+#endif
+
+#endif

libdisasm/swig/Makefile

@@ -0,0 +1,70 @@
+# change these values if you need to
+SWIG		=	swig	# apt-get install swig !
+GCC		=	gcc
+
+CC_FLAGS	=	-c -fPIC
+LD_FLAGS	=	-shared -L../.. -ldisasm
+
+BASE_NAME	=	x86disasm
+
+export INTERFACE_FILE BASE_NAME SWIG GCC CC_FLAGS LD_FLAGS
+
+#====================================================
+# TARGETS
+
+all: swig
+dummy: swig swig-python swig-ruby swig-perl swig-tcl install uninstall clean
+
+swig: swig-python swig-perl
+# swig-rub swig-tcl
+
+swig-python: 
+	cd python && make -f Makefile-swig
+
+swig-ruby:
+	cd ruby && make -f Makefile-swig
+
+swig-perl:
+	cd perl && make -f Makefile-swig
+
+swig-tcl:
+	cd tcl && make -f Makefile-swig
+
+# ==================================================================
+install: install-python install-perl
+# install-ruby install-tcl
+
+install-python:
+	cd python && sudo make -f Makefile-swig install
+
+install-ruby:
+	cd ruby && sudo make -f Makefile-swig install
+
+install-perl:
+	cd perl && sudo make -f Makefile-swig install
+
+install-tcl:
+	cd tcl && sudo make -f Makefile-swig install
+
+# ==================================================================
+uninstall: uninstall-python
+#uninstall-ruby uninstall-perl uninstall-tcl
+
+uninstall-python:
+	cd python && sudo make -f Makefile-swig uninstall
+
+uninstall-ruby:
+	cd ruby && sudo make -f Makefile-swig uninstall
+
+uninstall-perl:
+	cd perl && sudo make -f Makefile-swig uninstall
+
+uninstall-tcl:
+	cd tcl && sudo make -f Makefile-swig uninstall
+
+# ==================================================================
+clean:
+	cd python && make -f Makefile-swig clean
+	cd ruby && make -f Makefile-swig clean
+	cd perl && make -f Makefile-swig clean
+	cd tcl && make -f Makefile-swig clean

libdisasm/swig/README

@@ -0,0 +1,128 @@
+			Libdisasm SWIG README
+
+The SWIG utility (www.swig.org) can be used to generate 
+
+
+Building SWIG Modules
+---------------------
+
+	make
+	make install
+
+Make and Install both build Python, Perl, Ruby, and Tcl modules. If you
+do not have one of these languages installed, comment out the relevant
+target in the main Makefile.
+
+Install uses 'sudo' to put files in the correct locations; if you
+do not have sudo installed, change the install targets.
+
+The Module API
+--------------
+
+The OOP API
+-----------
+
+
+The Python Module
+-----------------
+
+To test that the module loads:
+
+	bash# python
+	>>> import x86disasm
+	>>> x86disasm.version_string()
+	'0.21-pre'
+	>>>^D
+	bash#
+
+	>>> import x86disasm
+	>>> import array
+	>>> disasm = x86disasm.X86_Disasm( )
+	>>> tgt = open( "/tmp/a.out", "rb" )
+	>>> tgt.seek( 0, 2 )
+	>>> size = tgt.tell()
+	>>> tgt.seek( 0, 0 )
+	>>> buf = array.array( 'B' )
+	>>> buf.fromfile( tgt, size )
+	>>> tgt.close()
+	>>> data = x86disasm.byteArray( size )
+	>>> for i in range( size ):
+	...     data[i] = buf.pop(0)
+	...
+	>>> del buf
+	>>> del tgt
+	>>> insn = disasm.disasm( data, size - 1, 0, 0 )
+	>>> insn.format( x86disasm.att_syntax )
+	 'jg\t0x00000047'
+	>>> insn.format( x86disasm.raw_syntax )
+	'0x00000000|0x00000000|2|7F 45 |||controlflow|jcc|jg|80386|General Purpose|||zero_clear sign_eq_oflow |0|0|relative|sbyte|00000047|'
+	>>> ops = insn.operand_list()
+	>>> node = ops.first()
+	>>> while node is not None:
+	...     s = node.op.format(x86disasm.raw_syntax)
+	...     print s
+	...     node = ops.next()
+	... 
+	relative|sbyte|00000047|
+
+
+
+
+
+
+The Perl Module
+---------------
+
+To test that the module loads:
+
+	bash# perl
+	use x86disasm;
+	print x86disasm::version_string() . "\n";
+	^D
+	0.21-pre
+	bash#
+
+The Ruby Module
+---------------
+
+To test that the module loads:
+
+	bash# irb
+	irb(main):001:0> require 'x86disasm'
+	=> true
+	irb(main):002:0> X86disasm.version_string()
+	=> "0.21-pre"
+	irb(main):003:0> x = X86disasm::X86_Disasm.new
+	=> #<X86disasm::X86_Disasm:0xb7d624a4>
+	irb(main):004:0> x.max_register_string()
+	=> 8
+	irb(main):003:0> ^D
+	bash#
+
+The Tcl Module
+---------------
+
+To test that the module loads:
+
+	bash# tclsh
+	% load /usr/lib/tcl8.3/x86disasm.so X86disasm
+	% version_string
+	0.21-pre
+	% ^D
+	bash#
+
+	% x86_init 0 NULL NULL
+		OR
+	% x86disasm dis
+	_486b0708_p_x86disasm
+	%  puts "[dis cget -last_error]"
+	0
+
+
+
+
+The Interface Files
+-------------------
+
+	libdisasm.i	-- interface file without shadow classes
+	libdisasm_oop.i	-- interface file with shadow classes

libdisasm/swig/libdisasm.i

@@ -0,0 +1,508 @@
+%module x86disasm
+%{
+#include "../../libdis.h"
+#include "../../../config.h"
+%}
+
+%rename(version_string) x86_version_string;
+%include "../../libdis.h"
+#include "../../../config.h"
+
+%inline %{
+	const char * x86_version_string( void ) {
+		return PACKAGE_VERSION;
+	}
+%}
+
+%rename(report_codes) x86_report_codes;
+%rename(report_error) x86_report_error;
+%rename(options) x86_options;
+%rename(init) x86_init;
+%rename(set_reporter) x86_set_reporter;
+%rename(set_options) x86_set_options;
+%rename(options) x86_get_options;
+%rename(cleanup) x86_cleanup;
+%rename(reg_type) x86_reg_type;
+%rename(reg) x86_reg_t;
+%rename(eaddr) x86_ea_t;
+%rename(op_type) x86_op_type;
+%rename(optype_is_address) x86_optype_is_address;
+%rename(optype_is_relative) x86_optype_is_relative;
+%rename(op_datatype) x86_op_datatype;
+%rename(op_access) x86_op_access;
+%rename(op_flags) x86_op_flags;
+%rename(operand) x86_op_t;
+%rename(insn_group) x86_insn_group; 
+%rename(insn_type) x86_insn_type;
+%rename(insn_note) x86_insn_note ;
+%rename(flag_status) x86_flag_status; 
+%rename(insn_cpu) x86_insn_cpu ;
+%rename(insn_isa) x86_insn_isa ;
+%rename(insn_prefix) x86_insn_prefix ;
+%rename(insn) x86_insn_t;
+%rename(insn_is_valid) x86_insn_is_valid;
+%rename(i_disasm) x86_disasm;
+%rename(i_disasm_range) x86_disasm_range;
+%rename(i_disasm_forward) x86_disasm_forward;
+%rename(insn_operand_count) x86_operand_count;
+%rename(insn_operand_1st) x86_operand_1st;
+%rename(insn_operand_2nd) x86_operand_2nd;
+%rename(insn_operand_3rd) x86_operand_3rd;
+%rename(insn_dest_operand) x86_get_dest_operand;
+%rename(insn_src_operand) x86_get_src_operand;
+%rename(insn_imm_operand) x86_get_imm_operand;
+%rename(operand_size) x86_operand_size;
+%rename(insn_rel_offset) x86_get_rel_offset;
+%rename(insn_branch_target) x86_get_branch_target;
+%rename(insn_imm) x86_get_imm;
+%rename(insn_raw_imm) x86_get_raw_imm;
+%rename(insn_set_addr) x86_set_insn_addr;
+%rename(insn_set_offset) x86_set_insn_offset;
+%rename(insn_set_function) x86_set_insn_function;
+%rename(insn_set_block) x86_set_insn_block;
+%rename(insn_tag) x86_tag_insn;
+%rename(insn_untag) x86_untag_insn;
+%rename(insn_is_tagged) x86_insn_is_tagged;
+%rename(asm_format) x86_asm_format;
+%rename(operand_format) x86_format_operand;
+%rename(insn_format_mnemonic) x86_format_mnemonic;
+%rename(insn_format) x86_format_insn;
+%rename(header_format) x86_format_header;
+%rename(endian) x86_endian;
+%rename(size_default_address) x86_addr_size;
+%rename(size_default_operand) x86_op_size;
+%rename(size_machine_word) x86_word_size;
+%rename(size_max_insn) x86_max_insn_size;
+%rename(reg_sp) x86_sp_reg;
+%rename(reg_fp) x86_fp_reg;
+%rename(reg_ip) x86_ip_reg;
+%rename(reg_from_id) x86_reg_from_id;
+%rename(reg_from_alias) x86_get_aliased_reg;
+%rename(invariant_op) x86_invariant_op_t;
+%rename(invariant) x86_invariant_t;
+%rename(disasm_invariant) x86_invariant_disasm;
+%rename(disasm_size) x86_size_disasm;
+
+%include "carrays.i"
+
+%array_class( unsigned char, byteArray );
+
+
+%apply (unsigned char *STRING, int LENGTH) { 
+	(unsigned char *buf, size_t buf_len) 
+};
+
+
+%newobject x86_op_copy;
+%inline %{
+	x86_op_t * x86_op_copy( x86_op_t * src ) {
+		x86_op_t *op;
+		
+		if (! src ) {
+			return NULL;
+		}
+
+		op = (x86_op_t *) calloc( sizeof(x86_op_t), 1 );
+		if ( op ) {
+			memcpy( op, src, sizeof(x86_op_t) );
+		}
+
+		return op;
+	}
+
+	typedef struct x86_op_list_node {
+		x86_op_t *op;
+		struct x86_op_list_node *next, *prev;
+	} x86_op_list_node;
+
+	typedef struct x86_op_list {
+		size_t count;
+		x86_op_list_node *head, *tail, *curr;
+	} x86_op_list;
+
+	x86_op_list * x86_op_list_new () {
+		x86_op_list *list = (x86_op_list *) 
+				calloc( sizeof(x86_op_list), 1 );
+		list->count = 0;
+		return list;
+	}
+
+	void x86_op_list_free(x86_op_list *list) {
+		x86_op_list_node *node, *next;
+
+		node = list->head;
+		while ( node ) {
+			next = node->next;
+			/* free( node->insn ); */
+			free( node );
+			node = next;
+		}
+
+		free( list );
+	}
+
+	x86_op_list_node * x86_op_list_first(x86_op_list *list) { 
+		return list->head; 
+	}
+
+	x86_op_list_node * x86_op_list_last(x86_op_list *list) { 
+		return list->tail; 
+	}
+
+	x86_op_list_node * x86_op_list_next(x86_op_list *list) { 
+		if (! list->curr ) {
+			list->curr = list->head;
+			return list->head;
+		}
+
+		list->curr = list->curr->next;
+		return list->curr;
+	}
+
+	x86_op_list_node * x86_op_list_prev(x86_op_list *list) { 
+		if (! list->curr ) {
+			list->curr = list->tail;
+			return list->tail;
+		}
+
+		list->curr = list->curr->prev;
+		return list->curr;
+	}
+
+%}
+
+%newobject x86_op_list_append;
+
+%inline %{
+	void x86_op_list_append( x86_op_list * list, x86_op_t *op ) {
+		x86_op_list_node *node = (x86_op_list_node *)
+					calloc( sizeof(x86_op_list_node) , 1 );
+		if (! node ) {
+			return;
+		}
+
+		list->count++;
+		if ( ! list->tail ) {
+			list->head = list->tail = node;
+		} else {
+			list->tail->next = node;
+			node->prev = list->tail;
+			list->tail = node;
+		}
+
+		node->op = x86_op_copy( op );
+	}
+
+	x86_oplist_t * x86_op_list_node_copy( x86_oplist_t * list ) {
+		x86_oplist_t *ptr;
+		ptr = (x86_oplist_t *) calloc( sizeof(x86_oplist_t), 1 );
+		if ( ptr ) {
+			memcpy( &ptr->op, &list->op, sizeof(x86_op_t) );
+		}
+
+		return ptr;
+	}
+
+	x86_insn_t * x86_insn_new() {
+		x86_insn_t *insn = (x86_insn_t *)
+				   calloc( sizeof(x86_insn_t), 1 );
+		return insn;
+	}
+
+	void x86_insn_free( x86_insn_t *insn ) {
+		x86_oplist_free( insn );
+		free( insn );
+	}
+%}
+
+%newobject x86_insn_copy;
+
+%inline %{
+	x86_insn_t * x86_insn_copy( x86_insn_t *src) {
+		x86_oplist_t *ptr, *list, *last = NULL;
+		x86_insn_t *insn = (x86_insn_t *)
+				   calloc( sizeof(x86_insn_t), 1 );
+
+		if ( insn ) {
+			memcpy( insn, src, sizeof(x86_insn_t) );
+			insn->operands = NULL;
+			insn->block = NULL;
+			insn->function = NULL;
+
+			/* copy operand list */
+			for ( list = src->operands; list; list = list->next ) {
+				ptr = x86_op_list_node_copy( list );
+
+				if (! ptr ) {
+					continue;
+				}
+
+				if ( insn->operands ) {
+					last->next = ptr;
+				} else {
+					insn->operands = ptr;
+				}
+				last = ptr;
+			}
+		}
+
+		return insn;
+	}
+
+	x86_op_list * x86_insn_op_list( x86_insn_t *insn ) {
+		x86_oplist_t *list = insn->operands;
+		x86_op_list *op_list = x86_op_list_new();
+
+		for ( list = insn->operands; list; list = list->next ) {
+			x86_op_list_append( op_list, &list->op );
+		}
+
+		return op_list;
+	}
+
+	typedef struct x86_insn_list_node {
+		x86_insn_t *insn;
+		struct x86_insn_list_node *next, *prev;
+	} x86_insn_list_node;
+
+	typedef struct x86_insn_list {
+		size_t count;
+		x86_insn_list_node *head, *tail, *curr;
+	} x86_insn_list;
+
+%}
+
+%newobject x86_insn_list_new;
+
+%inline %{
+	x86_insn_list * x86_insn_list_new () {
+		x86_insn_list *list = (x86_insn_list *) 
+				calloc( sizeof(x86_insn_list), 1 );
+		list->count = 0;
+		return list;
+	}
+
+	void x86_insn_list_free( x86_insn_list * list ) {
+		x86_insn_list_node *node, *next;
+
+		if (! list ) {
+			return;
+		}
+
+		node = list->head;
+		while ( node ) {
+			next = node->next;
+			/* free( node->insn ); */
+			free( node );
+			node = next;
+		}
+
+		free( list );
+	}
+
+	x86_insn_list_node * x86_insn_list_first( x86_insn_list *list ) { 
+		if (! list ) {
+			return NULL;
+		}
+		return list->head; 
+	}
+
+	x86_insn_list_node * x86_insn_list_last( x86_insn_list *list ) { 
+		if (! list ) {
+			return NULL;
+		}
+		return list->tail; 
+	}
+
+	x86_insn_list_node * x86_insn_list_next( x86_insn_list *list ) { 
+		if (! list ) {
+			return NULL;
+		}
+		if (! list->curr ) {
+			list->curr = list->head;
+			return list->head;
+		}
+
+		list->curr = list->curr->next;
+		return list->curr;
+	}
+
+	x86_insn_list_node * x86_insn_list_prev( x86_insn_list *list ) { 
+		if (! list ) {
+			return NULL;
+		}
+		if (! list->curr ) {
+			list->curr = list->tail;
+			return list->tail;
+		}
+
+		list->curr = list->curr->prev;
+		return list->curr;
+	}
+
+%}
+
+%newobject x86_insn_list_append;
+
+%inline %{
+	void x86_insn_list_append( x86_insn_list *list, x86_insn_t *insn ) {
+		x86_insn_list_node *node;
+		if (! list ) {
+			return;
+		}
+
+		node = (x86_insn_list_node *)
+					calloc( sizeof(x86_insn_list_node) , 1 );
+
+		if (! node ) {
+			return;
+		}
+
+		list->count++;
+		if ( ! list->tail ) {
+			list->head = list->tail = node;
+		} else {
+			list->tail->next = node;
+			node->prev = list->tail;
+			list->tail = node;
+		}
+
+		node->insn = x86_insn_copy( insn );
+	}
+
+	typedef struct {
+		enum x86_report_codes last_error;
+		void * last_error_data;
+		void * disasm_callback;
+		void * disasm_resolver;
+	} x86disasm;
+
+	void x86_default_reporter( enum x86_report_codes code, 
+				   void *data, void *arg ) {
+		x86disasm *dis = (x86disasm *) arg;
+		if ( dis ) {
+			dis->last_error = code;
+			dis->last_error_data = data;
+		}
+	}
+
+	void x86_default_callback( x86_insn_t *insn, void *arg ) {
+		x86_insn_list *list = (x86_insn_list *) arg;
+		if ( list ) {
+			x86_insn_list_append( list, insn );
+		}
+	}
+
+	/* TODO: resolver stack, maybe a callback */
+	long x86_default_resolver( x86_op_t *op, x86_insn_t *insn, void *arg ) {
+		x86disasm *dis = (x86disasm *) arg;
+		if ( dis ) {
+			//return dis->resolver( op, insn );
+			return 0;
+		}
+
+		return 0;
+	}
+
+	
+%}
+
+%newobject x86disasm_new;
+
+%inline %{
+	x86disasm * x86disasm_new ( enum x86_options options ) {
+		x86disasm * dis = (x86disasm *)
+				calloc( sizeof( x86disasm ), 1 );
+		x86_init( options, x86_default_reporter, dis );
+		return dis;
+	}
+
+	void x86disasm_free( x86disasm * dis ) {
+		x86_cleanup();
+		free( dis );
+	}
+%}
+
+%newobject x86_disasm;
+
+%inline %{
+	x86_insn_t * disasm( unsigned char *buf, size_t buf_len, 
+		           unsigned long buf_rva, unsigned int offset ) {
+		x86_insn_t *insn = calloc( sizeof( x86_insn_t ), 1 );
+		x86_disasm( buf, buf_len, buf_rva, offset, insn );
+		return insn;
+	}
+
+	int disasm_range( unsigned char *buf, size_t buf_len, 
+	              unsigned long buf_rva, unsigned int offset,
+		      unsigned int len ) {
+
+		x86_insn_list *list = x86_insn_list_new();
+
+		if ( len > buf_len ) {
+			len = buf_len;
+		}
+
+		return x86_disasm_range( buf, buf_rva, offset, len, 
+				x86_default_callback, list );
+	}
+
+	int disasm_forward( unsigned char *buf, size_t buf_len,
+			    unsigned long buf_rva, unsigned int offset ) {
+		x86_insn_list *list = x86_insn_list_new();
+
+		/* use default resolver: damn SWIG callbacks! */
+		return x86_disasm_forward( buf, buf_len, buf_rva, offset,
+			                   x86_default_callback, list, 
+					   x86_default_resolver, NULL );
+	}
+
+	size_t disasm_invariant( unsigned char *buf, size_t buf_len, 
+			  x86_invariant_t *inv ) {
+		return x86_invariant_disasm( buf, buf_len, inv );
+	}
+
+	size_t disasm_size( unsigned char *buf, size_t buf_len ) {
+		return x86_size_disasm( buf, buf_len );
+	}
+
+	int x86_max_operand_string( enum x86_asm_format format ) {
+		switch ( format ) {
+			case xml_syntax:
+				return  MAX_OP_XML_STRING;
+				break;
+			case raw_syntax:
+				return MAX_OP_RAW_STRING;
+				break;
+			case native_syntax:
+			case intel_syntax:
+			case att_syntax:
+			case unknown_syntax:
+			default:
+				return MAX_OP_STRING;
+				break;
+		}
+	}
+
+
+	int x86_max_insn_string( enum x86_asm_format format ) {
+		switch ( format ) {
+			case xml_syntax:
+				return  MAX_INSN_XML_STRING;
+				break;
+			case raw_syntax:
+				return MAX_INSN_RAW_STRING;
+				break;
+			case native_syntax:
+			case intel_syntax:
+			case att_syntax:
+			case unknown_syntax:
+			default:
+				return MAX_INSN_STRING;
+				break;
+		}
+	}
+
+	int x86_max_num_operands( ) { return MAX_NUM_OPERANDS; }
+%}
+

libdisasm/swig/perl/Makefile-swig

@@ -0,0 +1,65 @@
+ifndef BASE_NAME
+BASE_NAME	=	x86disasm
+endif
+
+ifndef SWIG
+SWIG		=	swig	# apt-get install swig !
+endif
+
+ifndef GCC
+GCC		=	gcc
+endif
+
+ifndef CC_FLAGS
+CC_FLAGS	=	-c -fPIC
+endif
+
+ifndef LD_FLAGS
+LD_FLAGS	=	-shared -L.. -ldisasm
+endif
+
+INTERFACE_FILE	=	libdisasm_oop.i
+
+SWIG_INTERFACE	=	../$(INTERFACE_FILE)
+
+# PERL rules
+PERL_MOD	=	blib/arch/auto/$(BASE_NAME)/$(BASE_NAME).so
+PERL_SHADOW	=	$(BASE_NAME)_wrap.c
+PERL_SWIG	=	$(BASE_NAME).pl
+PERL_OBJ	=	$(BASE_NAME)_wrap.o
+PERL_INC	=	`perl -e 'use Config; print $$Config{archlib};'`/CORE
+PERL_CC_FLAGS	=	`perl -e 'use Config; print $$Config{ccflags};'`
+
+#====================================================
+# TARGETS
+
+all: swig-perl
+
+dummy: swig-perl install uninstall clean
+
+swig-perl: $(PERL_MOD)
+
+$(PERL_MOD): $(PERL_OBJ)
+	perl Makefile.PL
+	make
+	#$(GCC) $(LD_FLAGS) $(PERL_OBJ) -o $@
+
+$(PERL_OBJ): $(PERL_SHADOW)
+	$(GCC) $(CC_FLAGS) $(PERL_CC_FLAGS) -I$(PERL_INC) -o $@ $<
+
+$(PERL_SHADOW): $(SWIG_INTERFACE)
+	swig -perl -shadow -o $(PERL_SHADOW) -outdir . $<
+
+# ==================================================================
+install: $(PERL_MOD)
+	make install
+
+# ==================================================================
+uninstall:
+
+# ==================================================================
+clean: 
+	rm $(PERL_MOD) $(PERL_OBJ)
+	rm $(PERL_SHADOW)
+	rm -rf Makefile blib pm_to_blib
+

libdisasm/swig/perl/Makefile.PL

@@ -0,0 +1,7 @@
+use ExtUtils::MakeMaker;
+
+WriteMakefile(
+	'NAME'	=> 'x86disasm',
+	'LIBS'	=> ['-ldisasm'],
+	'OBJECT'	=> 'x86disasm_wrap.o'
+);

libdisasm/swig/python/Makefile-swig

@@ -0,0 +1,64 @@
+ifndef BASE_NAME
+BASE_NAME	=	x86disasm
+endif
+
+ifndef SWIG
+SWIG		=	swig	# apt-get install swig !
+endif
+
+ifndef GCC
+GCC		=	gcc
+endif
+
+ifndef CC_FLAGS
+CC_FLAGS	=	-c -fPIC
+endif
+
+ifndef LD_FLAGS
+LD_FLAGS	=	-shared -L.. -ldisasm
+endif
+
+INTERFACE_FILE	=	libdisasm_oop.i
+
+SWIG_INTERFACE	=	../$(INTERFACE_FILE)
+
+# PYTHON rules
+PYTHON_MOD	=	$(BASE_NAME)-python.so
+PYTHON_SHADOW	=	$(BASE_NAME)_wrap.c
+PYTHON_SWIG	=	$(BASE_NAME).py
+PYTHON_OBJ	=	$(BASE_NAME)_wrap.o
+PYTHON_INC	=	`/bin/echo -e 'import sys\nprint sys.prefix + "/include/python" + sys.version[:3]' | python`
+PYTHON_LIB	=	`/bin/echo -e 'import sys\nprint sys.prefix + "/lib/python" + sys.version[:3]' | python`
+PYTHON_DEST	=	$(PYTHON_LIB)/lib-dynload/_$(BASE_NAME).so
+
+#====================================================
+# TARGETS
+
+all: swig-python
+
+dummy: swig-python install uninstall clean
+
+swig-python: $(PYTHON_MOD)
+
+$(PYTHON_MOD): $(PYTHON_OBJ)
+	$(GCC) $(LD_FLAGS) $(PYTHON_OBJ) -o $@
+
+$(PYTHON_OBJ): $(PYTHON_SHADOW)
+	$(GCC) $(CC_FLAGS) -I$(PYTHON_INC) -I.. -o $@ $<
+
+$(PYTHON_SHADOW): $(SWIG_INTERFACE)
+	swig -python -shadow -o $(PYTHON_SHADOW) -outdir . $<
+
+# ==================================================================
+install: $(PYTHON_MOD)
+	sudo cp $(PYTHON_MOD) $(PYTHON_DEST)
+	sudo cp $(PYTHON_SWIG) $(PYTHON_LIB)
+
+# ==================================================================
+uninstall:
+
+# ==================================================================
+clean: 
+	rm $(PYTHON_MOD) $(PYTHON_SWIG) $(PYTHON_OBJ)
+	rm $(PYTHON_SHADOW)
+

libdisasm/swig/ruby/Makefile-swig

@@ -0,0 +1,68 @@
+ifndef BASE_NAME
+BASE_NAME	=	x86disasm
+endif
+
+ifndef SWIG
+SWIG		=	swig	# apt-get install swig !
+endif
+
+ifndef GCC
+GCC		=	gcc
+endif
+
+ifndef CC_FLAGS
+CC_FLAGS	=	-c -fPIC
+endif
+
+ifndef LD_FLAGS
+LD_FLAGS	=	-shared -L../.. -ldisasm
+endif
+
+LIBDISASM_DIR	=	../..
+
+INTERFACE_FILE	=	libdisasm_oop.i
+
+SWIG_INTERFACE	=	../$(INTERFACE_FILE)
+
+# RUBY rules
+RUBY_MAKEFILE	=	Makefile
+RUBY_MOD	=	$(BASE_NAME).so
+RUBY_SHADOW	=	$(BASE_NAME)_wrap.c
+#RUBY_SWIG	=	$(BASE_NAME).rb
+RUBY_OBJ	=	$(BASE_NAME)_wrap.o
+RUBY_INC	=	`ruby -e 'puts $$:.join("\n")' | tail -2 | head -1`
+#RUBY_LIB	=	
+#RUBY_DEST	=	
+
+#====================================================
+# TARGETS
+
+all: swig-ruby
+
+dummy: swig-ruby install uninstall clean
+
+swig-ruby: $(RUBY_MOD)
+
+$(RUBY_MOD): $(RUBY_MAKEFILE)
+	make
+
+$(RUBY_MAKEFILE): $(RUBY_OBJ)
+	ruby extconf.rb
+
+$(RUBY_OBJ):$(RUBY_SHADOW)
+	$(GCC) $(CC_FLAGS) -I$(RUBY_INC) -I.. -o $@ $<
+
+$(RUBY_SHADOW): $(SWIG_INTERFACE)
+	swig -ruby -o $(RUBY_SHADOW) -outdir . $<
+
+# ==================================================================
+install: $(RUBY_MOD)
+	make install
+
+# ==================================================================
+uninstall:
+
+# ==================================================================
+clean: 
+	make clean || true
+	rm $(RUBY_SHADOW) $(RUBY_MAKEFILE) $(RUBY_MOD) $(RUBY_OBJ)

libdisasm/swig/ruby/extconf.rb

@@ -0,0 +1,4 @@
+require 'mkmf'
+find_library('disasm', 'x86_init', "/usr/local/lib", "../..")
+create_makefile('x86disasm')
+

libdisasm/swig/tcl/Makefile-swig

@@ -0,0 +1,63 @@
+ifndef BASE_NAME
+BASE_NAME	=	x86disasm
+endif
+
+ifndef SWIG
+SWIG		=	swig	# apt-get install swig !
+endif
+
+ifndef GCC
+GCC		=	gcc
+endif
+
+ifndef CC_FLAGS
+CC_FLAGS	=	-c -fPIC
+endif
+
+ifndef LD_FLAGS
+LD_FLAGS	=	-shared -L../.. -ldisasm
+endif
+
+INTERFACE_FILE	=	libdisasm.i
+
+SWIG_INTERFACE	=	../$(INTERFACE_FILE)
+
+# TCL rules
+TCL_VERSION	=	8.3
+TCL_MOD		=	$(BASE_NAME)-tcl.so
+TCL_SHADOW	=	$(BASE_NAME)_wrap.c
+TCL_OBJ		=	$(BASE_NAME)_wrap.o
+TCL_INC		=	/usr/include/tcl$(TCL_VERSION)
+TCL_LIB		=	/usr/lib/tcl$(TCL_VERSION)
+TCL_DEST	=	$(TCL_LIB)/$(BASE_NAME).so
+
+#====================================================
+# TARGETS
+
+all: swig-tcl
+
+dummy: swig-tcl install uninstall clean
+
+swig-tcl: $(TCL_MOD)
+
+$(TCL_MOD): $(TCL_OBJ)
+	$(GCC) $(LD_FLAGS) $(TCL_OBJ) -o $@
+
+$(TCL_OBJ): $(TCL_SHADOW)
+	$(GCC) $(CC_FLAGS) -I$(TCL_INC) -I.. -o $@ $<
+
+$(TCL_SHADOW): $(SWIG_INTERFACE)
+	swig -tcl -o $(TCL_SHADOW) -outdir . $<
+
+# ==================================================================
+install: $(TCL_MOD)
+	sudo cp $(TCL_MOD) $(TCL_DEST)
+
+# ==================================================================
+uninstall:
+
+# ==================================================================
+clean: 
+	rm $(TCL_MOD) $(TCL_SWIG) $(TCL_OBJ)
+	rm $(TCL_SHADOW)
+

libdisasm/x86_disasm.c

@@ -0,0 +1,210 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include "libdis.h"
+#include "ia32_insn.h"
+#include "ia32_invariant.h"
+#include "x86_operand_list.h"
+
+
+#ifdef _MSC_VER
+        #define snprintf        _snprintf
+        #define inline          __inline
+#endif
+
+unsigned int x86_disasm( unsigned char *buf, unsigned int buf_len,
+                uint32_t buf_rva, unsigned int offset,
+                x86_insn_t *insn ){
+        int len, size;
+	unsigned char bytes[MAX_INSTRUCTION_SIZE];
+
+        if ( ! buf || ! insn || ! buf_len ) {
+                /* caller screwed up somehow */
+                return 0;
+        }
+
+
+	/* ensure we are all NULLed up */
+	memset( insn, 0, sizeof(x86_insn_t) );
+        insn->addr = buf_rva + offset;
+        insn->offset = offset;
+	/* default to invalid insn */
+	insn->type = insn_invalid;
+	insn->group = insn_none;
+
+        if ( offset >= buf_len ) {
+                /* another caller screwup ;) */
+                x86_report_error(report_disasm_bounds, (void*)(long)buf_rva+offset);
+                return 0;
+        }
+
+        len = buf_len - offset;
+
+	/* copy enough bytes for disassembly into buffer : this
+	 * helps prevent buffer overruns at the end of a file */
+	memset( bytes, 0, MAX_INSTRUCTION_SIZE );
+	memcpy( bytes, &buf[offset], (len < MAX_INSTRUCTION_SIZE) ? len : 
+		MAX_INSTRUCTION_SIZE );
+
+        /* actually do the disassembly */
+	/* TODO: allow switching when more disassemblers are added */
+        size = ia32_disasm_addr( bytes, len, insn);
+
+        /* check and see if we had an invalid instruction */
+        if (! size ) {
+                x86_report_error(report_invalid_insn, (void*)(long)buf_rva+offset );
+                return 0;
+        }
+
+        /* check if we overran the end of the buffer */
+        if ( size > len ) {
+                x86_report_error( report_insn_bounds, (void*)(long)buf_rva + offset );
+		MAKE_INVALID( insn, bytes );
+		return 0;
+	}
+
+        /* fill bytes field of insn */
+        memcpy( insn->bytes, bytes, size );
+
+        return size;
+}
+
+unsigned int x86_disasm_range( unsigned char *buf, uint32_t buf_rva,
+                      unsigned int offset, unsigned int len,
+                      DISASM_CALLBACK func, void *arg ) {
+        x86_insn_t insn;
+        unsigned int buf_len, size, count = 0, bytes = 0;
+
+        /* buf_len is implied by the arguments */
+        buf_len = len + offset;
+
+        while ( bytes < len ) {
+                size = x86_disasm( buf, buf_len, buf_rva, offset + bytes,
+                                   &insn );
+                if ( size ) {
+                        /* invoke callback if it exists */
+                        if ( func ) {
+                                (*func)( &insn, arg );
+                        }
+                        bytes += size;
+                        count ++;
+                } else {
+                        /* error */
+                        bytes++;        /* try next byte */
+                }
+
+		x86_oplist_free( &insn );
+        }
+
+        return( count );
+}
+
+static inline int follow_insn_dest( x86_insn_t *insn ) {
+        if ( insn->type == insn_jmp || insn->type == insn_jcc ||
+             insn->type == insn_call || insn->type == insn_callcc ) {
+                return(1);
+        }
+        return(0);
+}
+
+static inline int insn_doesnt_return( x86_insn_t *insn ) {
+        return( (insn->type == insn_jmp || insn->type == insn_return) ? 1: 0 );
+}
+
+static int32_t internal_resolver( x86_op_t *op, x86_insn_t *insn ){
+        int32_t next_addr = -1;
+        if ( x86_optype_is_address(op->type) ) {
+                next_addr = op->data.sdword;
+        } else if ( op->type == op_relative_near ) {
+		next_addr = insn->addr + insn->size + op->data.relative_near;
+        } else if ( op->type == op_relative_far ) {
+		next_addr = insn->addr + insn->size + op->data.relative_far;
+        }
+        return( next_addr );
+}
+
+unsigned int x86_disasm_forward( unsigned char *buf, unsigned int buf_len,
+                        uint32_t buf_rva, unsigned int offset,
+                        DISASM_CALLBACK func, void *arg,
+                        DISASM_RESOLVER resolver, void *r_arg ){
+        x86_insn_t insn;
+        x86_op_t *op;
+        int32_t next_addr;
+        uint32_t next_offset;
+        unsigned int size, count = 0, bytes = 0, cont = 1;
+
+        while ( cont && bytes < buf_len ) {
+                size = x86_disasm( buf, buf_len, buf_rva, offset + bytes,
+                           &insn );
+
+                if ( size ) {
+                        /* invoke callback if it exists */
+                        if ( func ) {
+                                (*func)( &insn, arg );
+                        }
+                        bytes += size;
+                        count ++;
+                } else {
+                        /* error */
+                        bytes++;        /* try next byte */
+                }
+
+                if ( follow_insn_dest(&insn) ) {
+                        op = x86_get_dest_operand( &insn );
+                        next_addr = -1;
+
+                        /* if caller supplied a resolver, use it to determine
+                         * the address to disassemble */
+                        if ( resolver ) {
+                                next_addr = resolver(op, &insn, r_arg);
+                        } else {
+                                next_addr = internal_resolver(op, &insn);
+                        }
+
+                        if (next_addr != -1 ) {
+                                next_offset = next_addr - buf_rva;
+                                /* if offset is in this buffer... */
+                                if ( next_offset >= 0 &&
+                                     next_offset < buf_len ) {
+                                        /* go ahead and disassemble */
+                                        count += x86_disasm_forward( buf,
+                                                            buf_len,
+                                                            buf_rva,
+                                                            next_offset,
+                                                            func, arg,
+                                                            resolver, r_arg );
+                                } else  {
+                                        /* report unresolved address */
+                                        x86_report_error( report_disasm_bounds,
+                                                     (void*)(long)next_addr );
+                                }
+                        }
+                } /* end follow_insn */
+
+                if ( insn_doesnt_return(&insn) ) {
+                        /* stop disassembling */
+                        cont = 0;
+                }
+
+		x86_oplist_free( &insn );
+        }
+        return( count );
+}
+
+/* invariant instruction representation */
+size_t x86_invariant_disasm( unsigned char *buf, int buf_len, 
+		x86_invariant_t *inv ){
+	if (! buf || ! buf_len || ! inv  ) {
+		return(0);
+	}
+
+	return ia32_disasm_invariant(buf, buf_len, inv);
+}
+size_t x86_size_disasm( unsigned char *buf, unsigned int buf_len ) {
+	if (! buf || ! buf_len  ) {
+		return(0);
+	}
+
+	return ia32_disasm_size(buf, buf_len);
+}

libdisasm/x86_imm.c

@@ -0,0 +1,70 @@
+#include "qword.h"
+#include "x86_imm.h"
+
+#include <stdio.h>
+
+unsigned int x86_imm_signsized( unsigned char * buf, size_t buf_len,
+				void *dest, unsigned int size ) {
+	signed char *cp = (signed char *) dest;
+	signed short *sp = (signed short *) dest;
+	int32_t *lp = (int32_t *) dest;
+	qword_t *qp = (qword_t *) dest;
+
+	if ( size > buf_len ) {
+		return 0;
+	}
+
+	/* Copy 'size' bytes from *buf to *op
+	 * return number of bytes copied */
+	switch (size) {
+		case 1:		/* BYTE */
+			*cp =  *((signed char *) buf);
+			break;
+		case 2:		/* WORD */
+			*sp =  *((signed short *) buf);
+			break;
+		case 6:
+		case 8:		/* QWORD */
+			*qp = *((qword_t *) buf);
+			break;
+		case 4:		/* DWORD */
+		default:
+			*lp = *((int32_t *) buf);
+			break;
+	}
+	return (size);
+}
+
+unsigned int x86_imm_sized( unsigned char * buf, size_t buf_len, void *dest,
+			    unsigned int size ) {
+	unsigned char *cp = (unsigned char *) dest;
+	unsigned short *sp = (unsigned short *) dest;
+	uint32_t *lp = (uint32_t *) dest;
+	qword_t *qp = (qword_t *) dest;
+
+	if ( size > buf_len ) {
+		return 0;
+	}
+
+	/* Copy 'size' bytes from *buf to *op
+	 * return number of bytes copied */
+	switch (size) {
+		case 1:		/* BYTE */
+			*cp =  *((unsigned char *) buf);
+			break;
+		case 2:		/* WORD */
+			*sp =  *((unsigned short *) buf);
+			break;
+		case 6:
+		case 8:		/* QWORD */
+			*qp = *((qword_t *) buf);
+			break;
+		case 4:		/* DWORD */
+		default:
+			*lp = *((uint32_t *) buf);
+			break;
+	}
+
+	return (size);
+}
+

libdisasm/x86_imm.h

@@ -0,0 +1,18 @@
+#ifndef x86_IMM_H
+#define x86_IMM_H
+
+#include "./qword.h"
+#include <sys/types.h>
+
+#ifdef WIN32
+#include <windows.h>
+#endif
+
+/* these are in the global x86 namespace but are not a part of the
+ * official API */
+unsigned int x86_imm_sized( unsigned char *buf, size_t buf_len, void *dest,
+			    unsigned int size );
+
+unsigned int x86_imm_signsized( unsigned char *buf, size_t buf_len, void *dest,
+				unsigned int size );
+#endif

libdisasm/x86_insn.c

@@ -0,0 +1,182 @@
+#include <stdio.h>
+#include <stdlib.h>
+
+#include "libdis.h"
+
+#ifdef _MSC_VER
+        #define snprintf        _snprintf
+        #define inline          __inline
+#endif
+
+int x86_insn_is_valid( x86_insn_t *insn ) {
+	if ( insn && insn->type != insn_invalid && insn->size > 0 ) {
+		return 1;
+	}
+
+	return 0;
+}
+
+uint32_t x86_get_address( x86_insn_t *insn ) {
+	x86_oplist_t *op_lst;
+        if (! insn || ! insn->operands ) {
+        	return 0;
+        }
+
+	for (op_lst = insn->operands; op_lst; op_lst = op_lst->next ) {
+		if ( op_lst->op.type == op_offset ) {
+			return op_lst->op.data.offset;
+		} else if ( op_lst->op.type == op_absolute ) {
+			if ( op_lst->op.datatype == op_descr16 ) {
+				return (uint32_t)
+					op_lst->op.data.absolute.offset.off16;
+			}
+			return op_lst->op.data.absolute.offset.off32;
+		}
+	}
+	
+	return 0;
+}
+
+int32_t x86_get_rel_offset( x86_insn_t *insn ) {
+	x86_oplist_t *op_lst;
+        if (! insn || ! insn->operands ) {
+        	return 0;
+        }
+
+	for (op_lst = insn->operands; op_lst; op_lst = op_lst->next ) {
+		if ( op_lst->op.type == op_relative_near ) {
+			return (int32_t) op_lst->op.data.relative_near;
+		} else if ( op_lst->op.type == op_relative_far ) {
+			return op_lst->op.data.relative_far;
+		}
+	}
+	
+	return 0;
+}
+
+x86_op_t * x86_get_branch_target( x86_insn_t *insn ) {
+	x86_oplist_t *op_lst;
+        if (! insn || ! insn->operands ) {
+        	return NULL;
+        }
+
+	for (op_lst = insn->operands; op_lst; op_lst = op_lst->next ) {
+		if ( op_lst->op.access & op_execute ) {
+			return &(op_lst->op);
+		}
+	}
+	
+	return NULL;
+}
+x86_op_t * x86_get_imm( x86_insn_t *insn ) {
+	x86_oplist_t *op_lst;
+        if (! insn || ! insn->operands ) {
+        	return NULL;
+        }
+
+	for (op_lst = insn->operands; op_lst; op_lst = op_lst->next ) {
+		if ( op_lst->op.type == op_immediate ) {
+			return &(op_lst->op);
+		}
+	}
+	
+	return NULL;
+}
+
+#define IS_PROPER_IMM( x ) \
+	x->op.type == op_immediate && ! (x->op.flags & op_hardcode)
+							   
+
+/* if there is an immediate value in the instruction, return a pointer to
+ * it */
+unsigned char * x86_get_raw_imm( x86_insn_t *insn ) {
+        int size, offset;
+        x86_op_t *op  = NULL;
+
+        if (! insn || ! insn->operands ) {
+        	return(NULL);
+        }
+
+	/* a bit inelegant, but oh well... */
+	if ( IS_PROPER_IMM( insn->operands ) ) {
+		op = &insn->operands->op;
+	} else if ( insn->operands->next ) {
+		if ( IS_PROPER_IMM( insn->operands->next ) ) {
+			op = &insn->operands->next->op;
+		} else if ( insn->operands->next->next && 
+			    IS_PROPER_IMM( insn->operands->next->next ) ) {
+			op = &insn->operands->next->next->op;
+		}
+	}
+	
+	if (! op ) {
+		return( NULL );
+	}
+
+	/* immediate data is at the end of the insn */
+	size = x86_operand_size( op );
+	offset = insn->size - size;
+	return( &insn->bytes[offset] );
+}
+
+
+unsigned int x86_operand_size( x86_op_t *op ) {
+        switch (op->datatype ) {
+                case op_byte:    return 1;
+                case op_word:    return 2;
+                case op_dword:   return 4;
+                case op_qword:   return 8;
+                case op_dqword:  return 16;
+                case op_sreal:   return 4;
+                case op_dreal:   return 8;
+                case op_extreal: return 10;
+                case op_bcd:     return 10;
+                case op_ssimd:   return 16;
+                case op_dsimd:   return 16;
+                case op_sssimd:  return 4;
+                case op_sdsimd:  return 8;
+                case op_descr32: return 6;
+                case op_descr16: return 4;
+                case op_pdescr32: return 6;
+                case op_pdescr16: return 6;
+		case op_bounds16: return 4;
+		case op_bounds32: return 8;
+                case op_fpuenv16:  return 14;
+                case op_fpuenv32:  return 28;
+                case op_fpustate16:  return 94;
+                case op_fpustate32:  return 108;
+                case op_fpregset: return 512;
+		case op_fpreg: return 10;
+		case op_none: return 0;
+        }
+        return(4);      /* default size */
+}
+
+void x86_set_insn_addr( x86_insn_t *insn, uint32_t addr ) {
+        if ( insn ) insn->addr = addr;
+}
+
+void x86_set_insn_offset( x86_insn_t *insn, unsigned int offset ){
+        if ( insn ) insn->offset = offset;
+}
+
+void x86_set_insn_function( x86_insn_t *insn, void * func ){
+        if ( insn ) insn->function = func;
+}
+
+void x86_set_insn_block( x86_insn_t *insn, void * block ){
+        if ( insn ) insn->block = block;
+}
+
+void x86_tag_insn( x86_insn_t *insn ){
+        if ( insn ) insn->tag = 1;
+}
+
+void x86_untag_insn( x86_insn_t *insn ){
+        if ( insn ) insn->tag = 0;
+}
+
+int x86_insn_is_tagged( x86_insn_t *insn ){
+        return insn->tag;
+}
+

libdisasm/x86_misc.c

@@ -0,0 +1,71 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include "libdis.h"
+#include "ia32_insn.h"
+#include "ia32_reg.h"		/* for ia32_reg wrapper */
+#include "ia32_settings.h"
+extern ia32_settings_t ia32_settings;
+
+#ifdef _MSC_VER
+        #define snprintf        _snprintf
+        #define inline          __inline
+#endif
+
+
+/* =========================================================== INIT/TERM */
+static DISASM_REPORTER __x86_reporter_func = NULL;
+static void * __x86_reporter_arg = NULL;
+
+int x86_init( enum x86_options options, DISASM_REPORTER reporter, void * arg )
+{
+        ia32_settings.options = options;
+        __x86_reporter_func = reporter;
+	__x86_reporter_arg = arg;
+
+        return 1;
+}
+
+void x86_set_reporter( DISASM_REPORTER reporter, void * arg ) {
+        __x86_reporter_func = reporter;
+	__x86_reporter_arg = arg;
+}
+
+void x86_set_options( enum x86_options options ){
+        ia32_settings.options = options;
+}
+
+enum x86_options x86_get_options( void ) {
+        return ia32_settings.options;
+}
+
+int x86_cleanup( void )
+{
+        return 1;
+}
+
+/* =========================================================== ERRORS */
+void x86_report_error( enum x86_report_codes code, void *data ) {
+        if ( __x86_reporter_func ) {
+                (*__x86_reporter_func)(code, data, __x86_reporter_arg);
+        }
+}
+
+
+/* =========================================================== MISC */
+unsigned int x86_endian(void)        { return ia32_settings.endian;  }
+unsigned int x86_addr_size(void)     { return ia32_settings.sz_addr; }
+unsigned int x86_op_size(void)       { return ia32_settings.sz_oper; }
+unsigned int x86_word_size(void)     { return ia32_settings.sz_word; }
+unsigned int x86_max_insn_size(void) { return ia32_settings.max_insn; }
+unsigned int x86_sp_reg(void)        { return ia32_settings.id_sp_reg;      }
+unsigned int x86_fp_reg(void)        { return ia32_settings.id_fp_reg;      }
+unsigned int x86_ip_reg(void)        { return ia32_settings.id_ip_reg;      }
+unsigned int x86_flag_reg(void)        { return ia32_settings.id_flag_reg;  }
+
+/* wrapper function to hide the IA32 register fn */
+void x86_reg_from_id( unsigned int id, x86_reg_t * reg ) {
+	ia32_handle_register( reg, id );
+	return;
+}

libdisasm/x86_operand_list.c

@@ -0,0 +1,191 @@
+#include <stdlib.h>
+#include "libdis.h"
+
+
+static void x86_oplist_append( x86_insn_t *insn, x86_oplist_t *op ) {
+	x86_oplist_t *list;
+
+	if (! insn ) {	
+		return;
+	}
+
+	list = insn->operands;
+	if (! list ) {
+		insn->operand_count = 1;
+		/* Note that we have no way of knowing if this is an
+		 * exlicit operand or not, since the caller fills 
+		 * the x86_op_t after we return. We increase the
+		 * explicit count automatically, and ia32_insn_implicit_ops
+		 * decrements it */
+		insn->explicit_count = 1;
+		insn->operands = op;
+		return;
+	}
+
+	/* get to end of list */
+	for ( ; list->next; list = list->next ) 
+		;
+
+	insn->operand_count = insn->operand_count + 1;
+	insn->explicit_count = insn->explicit_count + 1;
+	list->next = op;
+
+	return;
+}
+	
+x86_op_t * x86_operand_new( x86_insn_t *insn ) {
+	x86_oplist_t *op;
+
+	if (! insn ) {	
+		return(NULL);
+	}
+	op = calloc( sizeof(x86_oplist_t), 1 );
+	op->op.insn = insn;
+	x86_oplist_append( insn, op );
+	return( &(op->op) );
+}
+
+void x86_oplist_free( x86_insn_t *insn ) {
+	x86_oplist_t *op, *list;
+
+	if (! insn ) {
+		return;
+	}
+
+	for ( list = insn->operands; list; ) {
+		op = list;
+		list = list->next;
+		free(op);
+	}
+
+	insn->operands = NULL;
+	insn->operand_count = 0;
+	insn->explicit_count = 0;
+
+	return;
+}
+
+/* ================================================== LIBDISASM API */
+/* these could probably just be #defines, but that means exposing the
+   enum... yet one more confusing thing in the API */
+int x86_operand_foreach( x86_insn_t *insn, x86_operand_fn func, void *arg, 
+	       		enum x86_op_foreach_type type ){
+	x86_oplist_t *list;
+	char explicit = 1, implicit = 1;
+
+	if (! insn || ! func ) {
+		return 0;
+	}
+	
+	/* note: explicit and implicit can be ORed together to
+	 * allow an "all" limited by access type, even though the
+	 * user is stupid to do this since it is default behavior :) */
+	if ( (type & op_explicit) && ! (type & op_implicit) ) {
+		implicit = 0;
+	}
+	if ( (type & op_implicit) && ! (type & op_explicit) ) {
+		explicit = 0;
+	}
+
+	type = type & 0x0F; /* mask out explicit/implicit operands */
+
+	for ( list = insn->operands; list; list = list->next ) {
+		if (! implicit && (list->op.flags & op_implied) ) {
+			/* operand is implicit */
+			continue;
+		}
+
+		if (! explicit && ! (list->op.flags & op_implied) ) {
+			/* operand is not implicit */
+			continue;
+		}
+
+		switch ( type ) {
+			case op_any:
+				break;
+			case op_dest:
+				if (! (list->op.access & op_write) ) {
+					continue;
+				}
+				break;
+			case op_src:
+				if (! (list->op.access & op_read) ) {
+					continue;
+				}
+				break;
+			case op_ro:
+				if (! (list->op.access & op_read) ||
+				      (list->op.access & op_write ) ) {
+					continue;
+				}
+				break;
+			case op_wo:
+				if (! (list->op.access & op_write) ||
+				      (list->op.access & op_read ) ) {
+					continue;
+				}
+				break;
+			case op_xo:
+				if (! (list->op.access & op_execute) ) {
+					continue;
+				}
+				break;
+			case op_rw:
+				if (! (list->op.access & op_write) ||
+				    ! (list->op.access & op_read ) ) {
+					continue;
+				}
+				break;
+			case op_implicit: case op_explicit: /* make gcc happy */
+					  break;
+		}
+		/* any non-continue ends up here: invoke the callback */
+		(*func)( &list->op, insn, arg );
+	}
+
+	return 1;
+}
+
+static void count_operand( x86_op_t *op, x86_insn_t *insn, void *arg ) {
+	size_t * count = (size_t *) arg;
+	*count = *count + 1;
+}
+
+size_t x86_operand_count( x86_insn_t *insn, enum x86_op_foreach_type type ) {
+	size_t count = 0;
+	
+	/* save us a list traversal for common counts... */
+	if ( type == op_any ) {
+		return insn->operand_count;
+	} else if ( type == op_explicit ) {
+		return insn->explicit_count;
+	}
+
+	x86_operand_foreach( insn, count_operand, &count, type );
+	return count;
+}
+
+/* accessor functions */
+x86_op_t * x86_operand_1st( x86_insn_t *insn ) {
+	if (! insn->explicit_count ) {
+		return NULL;
+	}
+
+	return &(insn->operands->op);
+}
+
+x86_op_t * x86_operand_2nd( x86_insn_t *insn ) {
+	if ( insn->explicit_count < 2 ) {
+		return NULL;
+	}
+
+	return &(insn->operands->next->op);
+}
+
+x86_op_t * x86_operand_3rd( x86_insn_t *insn ) {
+	if ( insn->explicit_count < 3 ) {
+		return NULL;
+	}
+
+	return &(insn->operands->next->next->op);
+}

libdisasm/x86_operand_list.h

@@ -0,0 +1,8 @@
+#ifndef X86_OPERAND_LIST_H
+#define X86_OPERAND_LIST_H
+#include "libdis.h"
+
+
+x86_op_t * x86_operand_new( x86_insn_t *insn );
+
+#endif

popnhax/Module.mk

@@ -8,10 +8,14 @@ ldflags_popnhax    := \
 
 libs_popnhax       := \
 	util \
-	minhook
+	minhook \
+	libdisasm
 
 srcpp_popnhax      := \
     dllmain.cc \
+    config.cc \
     loader.cc \
     SearchFile.cc \
-    translation.cc
+    translation.cc \
+    omnimix_patch.cc \
+    custom_categs.cc

popnhax/config.cc

@@ -0,0 +1,242 @@
+#include <windows.h>
+#include <io.h>
+
+#include "util/crc32.h"
+#include "util/log.h"
+#include "config.h"
+
+#define F_OK 0
+
+static char *get_opt_filename(const char *xml_filename){
+    char *opt_filename = strdup(xml_filename);
+    if ( opt_filename != NULL )
+        strcpy(opt_filename+strlen(opt_filename)-3, "opt");
+    return opt_filename;
+}
+
+//works for .opt and .xml
+static char *get_option(FILE *handle, char *option_name){
+    if (handle == NULL || option_name == NULL)
+        return NULL;
+
+    rewind(handle);
+    char line[256];
+    while (fgets(line, 256, handle))
+    {
+        char *ptr = strchr(line, '<');
+        if ( ptr == NULL )
+           continue;
+
+        if ( strncmp(option_name, ptr+1, strlen(option_name)) == 0
+         && (*(ptr+1+strlen(option_name)) == ' ') ) //filter out common prefixes
+        {
+            rewind(handle);
+            return strdup(line);
+        }
+    }
+
+    return NULL;
+}
+
+//check .xml crc-32 and compare to .opt entry
+static bool config_check_crc(const char *filepath){
+    //retrieve xml crc
+    uint32_t xmlcrc = 0;
+    long charcnt = 0;
+    if ( crc32file(filepath, &xmlcrc, &charcnt) == false )
+    {
+        LOG("popnhax: config update: cannot compute %s CRC-32\n", filepath);
+        return false;
+    }
+    LOG("popnhax: config update: %s CRC-32 is %08X\n", filepath, xmlcrc);
+
+    //retrieve opt crc
+    uint32_t optcrc = 0;
+    char *opt_filename = get_opt_filename(filepath);
+    FILE *opt = fopen(opt_filename, "r");
+    if (opt == NULL)
+    {
+        LOG("popnhax: config update: opt file not found.\n");
+        return false;
+    }
+    char line[256];
+    fgets(line, 256, opt); //disclaimer line
+    fgets(line, 256, opt); //CRC line
+    sscanf(line, "<!-- %08X -->\n", &optcrc);
+    fclose(opt);
+    LOG("popnhax: config update: %s reference CRC is %08X (matching: %s)\n", opt_filename, optcrc, optcrc==xmlcrc? "true":"false");
+    free(opt_filename);
+
+    return (optcrc==xmlcrc);
+}
+
+static char *get_option_name(char *line)
+{
+    if ( line == NULL )
+        return NULL;
+
+    char *ptr = line;
+    while ( *ptr == ' ' || *ptr == '\t' )
+        ptr++;
+
+    if ( *ptr != '<' )
+        return NULL;
+
+    if ( *(++ptr) == '!' )
+        return NULL;
+
+    if ( strchr(ptr, ' ') == NULL )
+        return NULL;
+
+    size_t opt_len = strchr(ptr, ' ') - ptr;
+    char *option_name = (char*)malloc(opt_len+1);
+    strncpy(option_name, ptr, opt_len);
+    option_name[opt_len] = '\0';
+    return option_name;
+}
+
+//check if _update file is present
+static bool config_need_update()
+{
+    return (access("_update", F_OK) == 0);
+}
+
+//for every entry in .opt, overwrite xml value
+static bool config_update_xml(const char *xml_filename)
+{
+    char *opt_filename = get_opt_filename(xml_filename);
+    FILE *opt = fopen(opt_filename, "r");
+    if ( opt == NULL )
+    {
+        LOG("popnhax: config_update_xml: cannot open opt file\n");
+        free(opt_filename);
+        return true; //not a problem if it doesn't exist
+    }
+
+    FILE *xml = fopen(xml_filename, "r");
+    if ( xml == NULL )
+    {
+        LOG("popnhax: config_update_xml: cannot open xml file\n");
+        return false;
+    }
+
+    FILE *update = fopen("update.xml.tmp", "w");
+    if ( update == NULL )
+    {
+        LOG("popnhax: config_update_xml: create update.xml.tmp\n");
+        return false;
+    }
+
+    // read each line from xml and copy to update unless opt has a value in which case copy from opt instead
+    char line[256];
+    while (fgets(line, 256, xml))
+    {
+        char *option_name = get_option_name(line);
+        char *option_line = get_option(opt, option_name); //will return NULL if option_name is NULL
+
+        if (option_line == NULL)
+        {
+            //not an option line, copy to dest
+            fprintf(update, "%s", line);
+            if (option_name)
+                free(option_name);
+            continue;
+        }
+
+        // opt file has an entry for it, take that one
+        fprintf(update, "%s", option_line);
+        free(option_line);
+        free(option_name);
+    }
+
+    fclose(xml);
+    fclose(opt);
+    fclose(update);
+    free(opt_filename);
+
+    if ( remove(xml_filename) == -1 )
+    {
+        LOG("popnhax: config_update_xml: cannot remove %s (%d)\n", xml_filename, errno);
+        return false;
+    }
+    if ( rename("update.xml.tmp",xml_filename) == -1 )
+    {
+        LOG("popnhax: config_update_xml: cannot rename update.xml.tmp (%d)\n", errno);
+        return false;
+    }
+
+    return true;
+}
+
+//regen .opt file (comprised of disclaimer, crc, and all options from xml)
+static bool config_make_opt(const char *xml_filename){
+    uint32_t crc = 0;
+    long charcnt = 0;
+    if ( crc32file(xml_filename, &crc, &charcnt) == false )
+    {
+        LOG("Cannot compute %s CRC-32\n", xml_filename);
+        return false;
+    }
+
+    FILE *xml = fopen(xml_filename, "r");
+    if ( xml == NULL )
+    {
+        LOG("popnhax: config update: cannot open config xml file\n");
+        return false;
+    }
+
+    char *opt_filename = get_opt_filename(xml_filename);
+    FILE *opt = fopen(opt_filename, "w");
+    free(opt_filename);
+    if ( opt == NULL )
+    {
+        LOG("popnhax: config update: cannot create config opt file\n");
+        free(opt_filename);
+        fclose(xml);
+        return false;
+    }
+
+    fprintf(opt, "<!-- GENERATED BY POPNHAX. DO NOT EDIT THIS FILE -->\n");
+    fprintf(opt, "<!-- %08X -->\n", crc);
+
+    char line[256];
+    while (fgets(line, 256, xml))
+    {
+        char *ptr = line;
+        while ( *ptr == ' ' || *ptr == '\t' )
+            ptr++;
+
+        if ( *ptr != '<' )
+            continue;
+
+        if ( *(ptr+1) == '!' )
+            continue;
+
+        fprintf(opt, "%s", line);
+    }
+
+    fclose(opt);
+    fclose(xml);
+    return true;
+}
+
+// take care of updating .xml/.opt files if needed
+bool config_process(const char *xml_filepath){
+    // look for _update marker file
+    if ( config_need_update() )
+    {
+        //update xml with values from opt
+        LOG("popnhax: config update: popnhax update detected. Updating xml with your previous settings\n");
+        config_update_xml(xml_filepath);
+        remove("_update");
+    } else if ( config_check_crc(xml_filepath) ) {
+        LOG("popnhax: config update: nothing to do.\n");
+        return true;
+    }
+
+    // regen opt from xml file
+    LOG("popnhax: config update: xml has changed. Regenerating opt file\n");
+    bool res = config_make_opt(xml_filepath);
+
+    return res;
+}

popnhax/config.h

@@ -4,6 +4,7 @@
 #include <stdbool.h>
 
 struct popnhax_config {
+    uint8_t game_version;
     bool practice_mode;
     bool hidden_is_offset;
     bool iidx_hard_gauge;
@@ -14,6 +15,10 @@ struct popnhax_config {
     bool quick_retire;
     bool back_to_song_select;
     bool score_challenge;
+    uint8_t custom_categ;
+    uint16_t custom_categ_min_songid;
+    bool custom_exclude_from_version;
+    bool custom_exclude_from_level;
     bool force_hd_timing;
     uint8_t force_hd_resolution;
     bool force_unlocks;
@@ -26,12 +31,12 @@ struct popnhax_config {
     bool force_full_opt;
     bool netvs_off;
     bool guidese_off;
+    bool local_favorites;
 
     bool patch_db;
-    bool disable_expansions;
-    bool disable_redirection;
     bool disable_multiboot;
     bool patch_xml_auto;
+    bool ignore_music_limit;
     char patch_xml_filename[MAX_PATH];
     char force_datecode[11];
     bool network_datecode;
@@ -41,16 +46,32 @@ struct popnhax_config {
     int8_t beam_brightness;
     bool fps_uncap;
     bool disable_translation;
-    bool translation_debug;
     bool enhanced_polling;
     uint8_t debounce;
     bool enhanced_polling_stats;
     int8_t enhanced_polling_priority;
     uint8_t hispeed_auto;
     uint16_t hispeed_default_bpm;
+    int8_t base_offset;
+    char custom_category_title[16];
+    char custom_category_format[64];
+    char custom_track_title_format[64];
+    char custom_track_title_format2[64];
+    bool exclude_omni;
+    bool partial_entries;
+    bool high_framerate;
+    bool high_framerate_limiter;
+    uint16_t high_framerate_fps;
+
+    /* removed/hidden */
+    bool disable_expansions;
+    bool disable_redirection;
     uint8_t survival_gauge;
     bool survival_iidx;
     bool survival_spicy;
+    bool translation_debug;
 };
 
+bool config_process(const char *filepath); // take care of updating .xml/.opt files if needed
+
 #endif

popnhax/custom_categs.h

@@ -0,0 +1,24 @@
+#ifndef __CUSTOM_CATEGS_H__
+#define __CUSTOM_CATEGS_H__
+
+#include <stdint.h>
+#include "popnhax/config.h"
+#include "util/bst.h"
+
+typedef struct {
+    char *name;
+    uint32_t size;
+    uint32_t *songlist; //really is a (songlist_t *) pointer
+} subcategory_s;
+
+extern uint32_t g_max_id;
+extern bst_t *g_customs_bst;
+
+void init_subcategories();
+void add_song_to_subcateg(uint32_t songid, subcategory_s* subcateg);
+subcategory_s* get_subcateg(const char *title);
+
+bool patch_custom_categs(const char *dllFilename, struct popnhax_config *config);
+bool patch_local_favorites(const char *dllFilename, uint8_t version);
+
+#endif

popnhax/loader.cc

@@ -4,8 +4,11 @@
 #include <windows.h>
 
 #include "imports/avs.h"
+#include "util/bst.h"
 #include "util/patch.h"
+#include "util/log.h"
 #include "xmlhelper.h"
+#include "custom_categs.h"
 
 #include "tableinfo.h"
 #include "loader.h"
@@ -66,7 +69,7 @@ uint32_t add_chart(uint32_t cur_idx, uint8_t *folder, uint8_t *filename, int32_t
                    uint32_t file_type, uint16_t used_keys, bool override_idx);
 
 void parse_charadb(const char *input_filename, const char *target);
-void parse_musicdb(const char *input_filename, const char *target);
+void parse_musicdb(const char *input_filename, const char *target, struct popnhax_config *config);
 
 std::map<uint32_t, int8_t> chart_type_overrides;
 
@@ -818,12 +821,76 @@ void parse_charadb(const char *input_filename, const char *target) {
     free(config_xml);
 }
 
-void parse_musicdb(const char *input_filename, const char *target) {
+static char *get_subcateg_title(const char* path) {
+    char *categ_name = NULL;
+    char filename[64];
+
+    //try to open "folderpath/_name.txt"
+    size_t len = (size_t)(strchr(path+10, '\\')-(path));
+    strncpy(filename, path, len);
+    sprintf(filename+len, "\\_name.txt");
+
+    FILE *file = fopen(filename, "r");
+    if ( file != NULL ) {
+        //if it has a custom title, use it
+        char line[64];
+        if (fgets(line, sizeof(line), file)) {
+            //handle UTF-8 BOM since that could be common
+            categ_name = (memcmp(line, "\xEF\xBB\xBF", 3) == 0) ? strdup(line+3) : strdup(line);
+            LOG("%s sets subcategory name to %s\n", filename, categ_name);
+        }
+        fclose(file);
+    } else { // or just keep folder name by itself (cut "data_mods")
+        len = (size_t)(strchr(path+10, '\\')-(path+10));
+        categ_name = (char*) malloc(len+1);
+        strncpy(categ_name, path+10, len);
+        categ_name[len] = '\0';
+    }
+
+    return categ_name;
+}
+
+#define F_OK 0
+static bool is_excluded_folder(const char *path)
+{
+    char filename[64];
+
+    //try to open "folderpath/_exclude"
+    size_t len = (size_t)(strchr(path+10, '\\')-(path));
+    strncpy(filename, path, len);
+    sprintf(filename+len, "\\_exclude");
+    if (access(filename, F_OK) == 0)
+    {
+        LOG("%s causes folder to be excluded from customs (contents will be treated as regular songs)\n", filename);
+        return true;
+    }
+
+    if ( path[strlen("data_mods/")] == '_' )
+    {
+        filename[strchr(path+10, '\\')-path] = '\0';
+        LOG("%s starting with _ causes folder to be excluded from customs (contents will be treated as regular songs)\n", filename);
+        return true;
+    }
+
+    return false;
+}
+
+void parse_musicdb(const char *input_filename, const char *target, struct popnhax_config *config) {
     if (!file_exists(input_filename)) {
         printf("Couldn't find %s, skipping...\n", input_filename);
         return;
     }
 
+    bool excluded = (config->custom_categ && is_excluded_folder(input_filename));
+
+    char *subcateg_title = NULL;
+    subcategory_s *subcateg = NULL;
+    if (config->custom_categ == 2 && !excluded)
+    {
+        subcateg_title = get_subcateg_title(input_filename);
+        subcateg = get_subcateg(subcateg_title); //will return a new one if not found
+    }
+
     property *config_xml = load_prop_file(input_filename);
 
     if (target && strlen(target) > 0) {
@@ -859,7 +926,22 @@ void parse_musicdb(const char *input_filename, const char *target) {
             // If it doesn't exist, create a new entry in memory
             // Update the data in-place and make all parameters optional
             music_entry *m = get_music(idx);
-            bool is_fresh = m == NULL;
+            bool is_fresh = m == NULL; // ie. not part of internal songdb
+            bool is_gone = ( m != NULL && strcmp((const char*) m->title_ptr, "\x81\x5D") == 0); // removed entries all have this title (SJIS "-")
+
+            // Update customs/omni songid list
+            if ( is_fresh || is_gone || config->partial_entries )
+            {
+                if ( idx >= config->custom_categ_min_songid && !excluded && bst_search(g_customs_bst, idx) == NULL )
+                {
+                    g_customs_bst = bst_insert(g_customs_bst, idx);
+                    //LOG("%d inserted into customs bst\n", idx);
+                    if (config->custom_categ == 2)
+                    {
+                        add_song_to_subcateg(idx, subcateg);
+                    }
+                }
+            }
 
             if (is_fresh) {
                 // Default music entry
@@ -915,6 +997,16 @@ void parse_musicdb(const char *input_filename, const char *target) {
                 }
             }
 
+            if ( config->custom_categ
+              && config->custom_exclude_from_version
+              && !excluded
+              && idx >= config->custom_categ_min_songid 
+              && ( is_fresh || config->exclude_omni ) )
+            {
+                m->cs_version = 0;
+                m->folder = 0;
+            }
+
             if ((prop_chart = property_search(config_xml, prop, "charts/chart"))) {
                 for (; prop_chart != NULL; prop_chart = property_node_traversal(
                                                prop_chart, TRAVERSE_NEXT_SEARCH_RESULT)) {
@@ -990,9 +1082,14 @@ void parse_musicdb(const char *input_filename, const char *target) {
     }
 
     free(config_xml);
+
+    if (config->custom_categ == 2 && !excluded)
+    {
+        free(subcateg_title);
+    }
 }
 
-void load_databases(const char *target_datecode) {
+void load_databases(const char *target_datecode, struct popnhax_config *config) {
 
     SearchFile s;
     printf("XML db files search...\n");
@@ -1008,16 +1105,19 @@ void load_databases(const char *target_datecode) {
         parse_charadb(result[i].c_str(), target_datecode);
     }
 
+    if (config->custom_categ == 2)
+        init_subcategories();
+
     for(uint16_t i=0;i<result.size();i++)
     {
         if ( strstr(result[i].c_str(), "musicdb") == NULL )
             continue;
         printf("(musicdb) Loading %s...\n", result[i].c_str());
-        parse_musicdb(result[i].c_str(), target_datecode);
+        parse_musicdb(result[i].c_str(), target_datecode, config);
     }
 }
 
-void musichax_core_init(bool force_unlocks, bool is_expansion_allowed, bool is_redirection_allowed,
+void musichax_core_init(struct popnhax_config *config,
                         char *target_datecode,
 
                         char *base_data,
@@ -1036,6 +1136,10 @@ void musichax_core_init(bool force_unlocks, bool is_expansion_allowed, bool is_r
 
                         uint64_t chara_size, uint64_t *new_chara_size, char *orig_chara_data,
                         uint8_t **new_chara_table) {
+    bool force_unlocks = config->force_unlocks;
+    bool is_expansion_allowed = !config->disable_expansions;
+    bool is_redirection_allowed = !config->disable_redirection;
+
     if (style_size > fontstyle_table_size) {
         fontstyle_table_size = style_size;
     }
@@ -1181,7 +1285,7 @@ void musichax_core_init(bool force_unlocks, bool is_expansion_allowed, bool is_r
             cur->chara_x, cur->chara_y, cur->unk1, cur->display_bpm, cur->hold_flags, true);
     }
 
-    load_databases((const char *)target_datecode);
+    load_databases((const char *)target_datecode, config);
 
     // Add some filler charts to fix some bugs (hack)
     for (int i = 0; i < 10; i++) {

popnhax/loader.h

@@ -2,10 +2,11 @@
 #define __LOADER_H__
 
 #include <stdint.h>
+#include "popnhax/config.h"
 
 int8_t get_chart_type_override(uint8_t *, uint32_t, uint32_t);
 
-void musichax_core_init(bool force_unlocks, bool is_expansion_allowed, bool is_redirection_allowed,
+void musichax_core_init(struct popnhax_config *config,
                         char *target_datecode, char *base_data, uint64_t music_size,
                         uint64_t *new_music_size, char *orig_music_data, uint8_t **new_music_table,
                         uint64_t chart_size, uint64_t *new_chart_size, char *orig_chart_data,

popnhax/omnimix_patch.cc

@@ -0,0 +1,297 @@
+#include <cstdio>
+#include <cstring>
+#include <io.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <windows.h>
+
+#include "util/bst.h"
+#include "util/log.h"
+#include "util/patch.h"
+#include "util/search.h"
+#include "libdisasm/libdis.h"
+
+#include "omnimix_patch.h"
+
+#define LINE_SIZE 512
+
+
+#define MUSIC_IDX  0
+#define CHART_IDX  1
+#define STYLE_IDX  2
+#define FLAVOR_IDX 3
+#define CHARA_IDX  4
+#define NUM_IDX    5
+
+bst_t *g_xref_bst = NULL;
+bst_t *g_data_xref_bst = NULL;
+
+typedef struct table_s {
+	char *name;
+	uint32_t *addr;
+	uint32_t size; // size of one entry
+	uint32_t limit; // number of entries 
+} table_t;
+
+table_t g_tables[NUM_IDX]; 
+
+uint32_t *find_binary(char *data, DWORD dllSize, const char *search_pattern, uint32_t search_size, uint32_t search_head, int8_t search_idx)
+{
+	//fprintf(stderr, "find binary %.*s\n", search_size, search_pattern);
+    uint32_t *ea = 0;
+    int8_t found = -1;
+    while (true)
+    {
+        int64_t pattern_offset = search(data, dllSize -((uint32_t) ea)-1, search_pattern, search_size, ((uint32_t) ea)+1);
+
+        if (pattern_offset == -1)
+            break;
+
+        ea = (uint32_t *) (pattern_offset + search_head);
+        found++;
+
+        if (found != search_idx)
+            continue;
+
+        return (uint32_t *) ((int64_t) data + (int64_t) ea);
+    }
+
+    return NULL;
+}
+
+uint32_t *find_binary_xref(char *data, DWORD dllSize, const char *search_pattern, uint32_t search_size, uint32_t search_head, int8_t search_idx, int8_t xref_search_idx)
+{
+
+    uint32_t* ea = find_binary(data, dllSize, search_pattern, search_size, search_head, search_idx);
+
+	if (ea == NULL)
+		return NULL;
+
+    char *as_hex = (char *) &ea;
+
+    ea = find_binary(data, dllSize, as_hex, 4, 0, xref_search_idx);
+
+    return ea;
+
+}
+
+uint32_t get_table_size_by_xref(char *data, DWORD data_size, uint32_t *ea, uint32_t entry_size)
+{
+	//fprintf(stderr, "get table size by xref\n");
+	uint32_t *orig_ea = ea;
+	// Skip 10 entries because why not. We're looking for the end anyway
+	ea = (uint32_t*)((uint32_t)ea+(entry_size*10));
+	
+	//fprintf(stderr, "orig ea %p, entry size %u, ea %p\n", orig_ea, entry_size, ea);
+	bool found = false;
+	
+	while (!found)
+	{
+		uint32_t *as_int = (uint32_t*)&ea;
+		//uint32_t curr_limit = (((uint32_t)ea)-((uint32_t)orig_ea))/entry_size;
+		//fprintf(stderr, "limit candidate %u, looking for xref to %0X\n", curr_limit, *as_int);
+
+		if ( bst_search(g_xref_bst, *as_int) != NULL )
+		{
+			//fprintf(stderr, "found .text XREF in BST\n");
+			found = true;
+		} else if ( bst_search(g_data_xref_bst, *as_int) != NULL )
+		{
+			//fprintf(stderr, "found .data XREF at addr %p\n", xref);
+			found = true;
+		} /* else {
+                      //look for Xref in data section
+                       char *as_hex = (char*)&ea;
+                       uint32_t *xref = find_binary(data, data_size, as_hex, 4, 0, 0);
+                       if ( xref != NULL )
+                       {
+                                fprintf(stderr, "found .data XREF at addr %p\n", xref);
+                                found = true;
+						}
+		}*/
+
+		if (!found)
+			ea = (uint32_t*)((uint32_t)ea+entry_size);
+	}
+
+	//fprintf(stderr, "found value %u (ea = %p, orig_ea = %p, entry size = %u)\n", ((((uint32_t)ea)-((uint32_t)orig_ea))/entry_size), ea, orig_ea, entry_size);
+	return (((uint32_t)ea)-((uint32_t)orig_ea))/entry_size;
+}
+
+void print_cb(x86_insn_t *insn, void *arg)
+{
+   char line[LINE_SIZE];    /* buffer of line to print */
+   x86_format_insn(insn, line, LINE_SIZE, intel_syntax);
+   fprintf(stderr,"%s\n", line);
+}
+
+static void add_xref_to_bst(bst_t **bst, uint32_t addr)
+{
+	   if ( bst_search(*bst, addr) == NULL )
+	   {
+		   //LOG("added addr %0X to BST\n", addr);
+           *bst = bst_insert(*bst, addr);
+	   }
+}
+
+void add_xref(x86_insn_t *insn, void *arg)
+{
+   uint32_t addr = x86_get_address(insn);
+   if ( addr != 0 )
+   {
+	   add_xref_to_bst(&g_xref_bst, addr);
+       //LOG("found ref to %0X at ea %0X\n", addr, insn->addr);
+	   //fprintf(stderr, "\n");
+	   return;
+   }
+
+   // doesn't have an address directly
+   x86_op_t *op = x86_operand_1st(insn);
+   if ( op != NULL && op->type == op_expression )
+   {
+	   x86_ea_t exp = op->data.expression;
+//	   LOG("(expression) disp is %0X\n", exp.disp);
+	   add_xref_to_bst(&g_xref_bst, exp.disp);
+   }
+
+   op = x86_operand_2nd(insn);
+   if ( op != NULL && op->type == op_expression )
+   {
+	   x86_ea_t exp = op->data.expression;
+//	   LOG("(expression) disp is %0X\n", exp.disp);
+	   add_xref_to_bst(&g_xref_bst, exp.disp);
+   }
+
+   op = x86_operand_3rd(insn);
+   if ( op != NULL && op->type == op_expression )
+   {
+	   x86_ea_t exp = op->data.expression;
+//	   LOG("(expression) disp is %0X\n", exp.disp);
+	   add_xref_to_bst(&g_xref_bst, exp.disp);
+   }
+					
+}
+
+bool make_omnimix_patch(const char *dllFilename){
+
+    DWORD dllSize = 0;
+    char *buf = getDllData(dllFilename, &dllSize);
+
+    char *text = NULL;
+	DWORD text_size = 0;
+	
+	char *data = NULL;
+	DWORD data_size = 0;
+	
+	char *rdata = NULL;
+	DWORD rdata_size = 0;
+	
+PIMAGE_NT_HEADERS headers = (PIMAGE_NT_HEADERS)(buf + ((PIMAGE_DOS_HEADER)buf)->e_lfanew);
+PIMAGE_SECTION_HEADER section_header = IMAGE_FIRST_SECTION(headers);
+
+for (WORD SectionIndex = 0; SectionIndex < headers->FileHeader.NumberOfSections; SectionIndex++)
+{
+    PIMAGE_SECTION_HEADER curr_header = &section_header[SectionIndex];
+
+    if ( strcmp( (const char*)curr_header->Name, ".text" ) == 0 || curr_header->Characteristics & IMAGE_SCN_CNT_CODE )
+    {
+        text = (char*) ((int64_t)buf + curr_header->VirtualAddress);
+        text_size = curr_header->Misc.VirtualSize;
+    }
+    if ( strcmp( (const char*)curr_header->Name, ".data" ) == 0 )
+    {
+        data = (char*) ((int64_t)buf + curr_header->VirtualAddress);
+        data_size = curr_header->Misc.VirtualSize;
+    }
+    if ( strcmp( (const char*)curr_header->Name, ".rdata" ) == 0 )
+    {
+        rdata = (char*) ((int64_t)buf + curr_header->VirtualAddress);
+        rdata_size = curr_header->Misc.VirtualSize;
+    }
+}
+        LOG("buf addr is %p\n", buf);
+        LOG("text section size %lX, addr %p, beginning %02X %02X %02X %02X\n", text_size, text, text[0], text[1], text[2], text[3]);
+        LOG("data section size %lX, addr %p, beginning %02X %02X %02X %02X\n", data_size, data, data[0], data[1], data[2], data[3]);
+        LOG("rdata section size %lX, addr %p, beginning %02X %02X %02X %02X\n", rdata_size, rdata, rdata[0], rdata[1], rdata[2], rdata[3]);
+
+    //add all data xrefs to BST (take anything which looks like 0x10****** in .data section)
+	{
+        LOG("Parsing .data section... ");
+		uint32_t *data_ea = 0;
+		while (true)
+		{
+			int64_t pattern_offset = search(data, data_size -((uint32_t) data_ea)-1, "\x10", 1, ((uint32_t) data_ea)+1);
+			if (pattern_offset == -1)
+				break;
+
+			data_ea = (uint32_t *) ((uint32_t)data + pattern_offset - 3);
+			uint32_t *as_int = (uint32_t*)data_ea;
+
+			//LOG("found %X at offset %p\n", *as_int, data_ea);
+			//LOG("adding %X to data bst\n", *as_int);
+			add_xref_to_bst(&g_data_xref_bst, *as_int);
+			data_ea = (uint32_t *) (pattern_offset);
+		}
+		LOG("done\n");
+	}
+
+    // disassemble .text section and add all xrefs to BST
+    LOG("Disassemble .text section... ");
+    x86_init(opt_none, NULL, NULL);
+    unsigned int num_insn = x86_disasm_range( (unsigned char *)text, (uint32_t) buf, 0, text_size, &add_xref, NULL);
+    x86_cleanup();
+    LOG("done (%u instructions)\n", num_insn);
+
+LOG("<?xml version='1.0' encoding='shift-jis'?>\n");
+LOG("<patches target=\"20xxxxxx00\">\n");
+
+/* init tables */
+g_tables[MUSIC_IDX].name = strdup("music");
+g_tables[CHART_IDX].name = strdup("chart");
+g_tables[STYLE_IDX].name = strdup("style");
+g_tables[FLAVOR_IDX].name = strdup("flavor");
+g_tables[CHARA_IDX].name = strdup("chara");
+
+// These all reference the first entry in their respective tables
+g_tables[MUSIC_IDX].addr = find_binary_xref(buf, dllSize, "\x00\x83\x7C\x83\x62\x83\x76\x83\x58\x00", 10, 1, 0, 0);
+g_tables[CHART_IDX].addr = find_binary_xref(buf, dllSize, "\x00\x70\x6F\x70\x6E\x31\x00\x00", 8, 1, 0, 1);
+g_tables[STYLE_IDX].addr = find_binary(buf, dllSize, "\x01\x00\x00\x00\xFF\x54\x0C\x00\x1A\x00\x00\x00\x11\x00\x00\x00", 16, 0, 2);
+g_tables[FLAVOR_IDX].addr = find_binary(buf, dllSize, "\x00\x82\xBB\x82\xEA\x82\xA2\x82\xAF\x81\x5B\x00\x00\x00\x82\xA4\x82", 17, 1, 0);
+g_tables[CHARA_IDX].addr = find_binary_xref(buf, dllSize, "\x00\x62\x61\x6D\x62\x5F\x31\x61\x00", 9, 1, 0, 0);
+
+fprintf(stderr, "buffer addresses ok\n");
+// Modify the entry sizes as required
+g_tables[MUSIC_IDX].size = 0xAC;
+g_tables[CHART_IDX].size = 0x20; // Probably won't change?
+g_tables[STYLE_IDX].size = 0x10; // Unlikely to change
+g_tables[FLAVOR_IDX].size = 0x60;
+g_tables[CHARA_IDX].size = 0x4C;
+
+// buffer_addr + (buffer_entry_size * limit) should give you the very end of the array (after the last entry)
+g_tables[MUSIC_IDX].limit = get_table_size_by_xref(data, data_size, g_tables[MUSIC_IDX].addr, g_tables[MUSIC_IDX].size);
+g_tables[CHART_IDX].limit = get_table_size_by_xref(data, data_size, g_tables[CHART_IDX].addr, g_tables[CHART_IDX].size);
+g_tables[STYLE_IDX].limit = get_table_size_by_xref(data, data_size, g_tables[STYLE_IDX].addr, g_tables[STYLE_IDX].size);
+g_tables[FLAVOR_IDX].limit = get_table_size_by_xref(data, data_size, g_tables[FLAVOR_IDX].addr, g_tables[FLAVOR_IDX].size);
+g_tables[CHARA_IDX].limit = get_table_size_by_xref(data, data_size, g_tables[CHARA_IDX].addr, g_tables[CHARA_IDX].size);
+
+LOG("\t<limits>\n");
+for (int i=0; i < NUM_IDX; i++)
+{
+    //patch_target, buffer_addr, entry_size = buffer_info
+    LOG("\t\t<%s __type=\"u32\">%d</%s>\n", g_tables[i].name, g_tables[i].limit, g_tables[i].name);
+}
+LOG("\t</limits>\n");
+
+LOG("\t<buffer_base_addrs>\n");
+for (int i=0; i < NUM_IDX; i++)
+{
+    //patch_target, buffer_addr, entry_size = buffer_info
+    LOG("\t\t<%s __type=\"str\">0x%p</%s>\n", g_tables[i].name, g_tables[i].addr, g_tables[i].name);
+}
+LOG("\t</buffer_base_addrs>\n");
+
+
+	getchar();
+	exit(0);
+	return false;
+}

popnhax/omnimix_patch.h

@@ -0,0 +1,6 @@
+#ifndef __OMNIMIX_PATCH_H__
+#define __OMNIMIX_PATCH_H__
+
+bool make_omnimix_patch(const char *dllFilename);
+
+#endif

util/Module.mk

@@ -1,7 +1,8 @@
 libs                    += util
 
 srcpp_util              := \
-	fuzzy_search.cc \
+	bst.cc \
 	search.cc \
 	cmdline.cc \
 	patch.cc \
+	crc32.cc \

util/bst.cc

@@ -0,0 +1,34 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdint.h>
+
+#include "bst.h"
+
+bst_t* bst_search(bst_t *root, uint32_t val)
+{
+    if( root == NULL || root->data == val )
+        return root;
+    else if( val > (root->data) )
+        return bst_search(root->right, val);
+    else
+        return bst_search(root->left,val);
+}
+
+bst_t* bst_insert(bst_t *root, uint32_t val)
+{
+    if ( root == NULL )
+    {
+        bst_t *p;
+        p = (bst_t *)malloc(sizeof(bst_t));
+        p->data = val;
+        p->left = NULL;
+        p->right = NULL;
+        return p;
+    }
+    else if ( val > root->data )
+        root->right = bst_insert(root->right, val);
+    else
+        root->left = bst_insert(root->left, val);
+
+    return root;
+}

util/bst.h

@@ -0,0 +1,16 @@
+#ifndef __BST_H__
+#define __BST_H__
+
+#include <stdint.h>
+
+typedef struct bst_s
+{
+    uint32_t data;
+    struct bst_s *right;
+    struct bst_s *left;
+} bst_t;
+
+bst_t* bst_search(bst_t *root, uint32_t val);
+bst_t* bst_insert(bst_t *root, uint32_t val);
+
+#endif

util/crc32.cc

@@ -0,0 +1,160 @@
+/* Crc - 32 BIT ANSI X3.66 CRC checksum files */
+#include <stdio.h>
+#include "crc32.h"
+
+/**********************************************************************\
+|* Demonstration program to compute the 32-bit CRC used as the frame  *|
+|* check sequence in ADCCP (ANSI X3.66, also known as FIPS PUB 71     *|
+|* and FED-STD-1003, the U.S. versions of CCITT's X.25 link-level     *|
+|* protocol).  The 32-bit FCS was added via the Federal Register,     *|
+|* 1 June 1982, p.23798.  I presume but don't know for certain that   *|
+|* this polynomial is or will be included in CCITT V.41, which        *|
+|* defines the 16-bit CRC (often called CRC-CCITT) polynomial.  FIPS  *|
+|* PUB 78 says that the 32-bit FCS reduces otherwise undetected       *|
+|* errors by a factor of 10^-5 over 16-bit FCS.                       *|
+\**********************************************************************/
+
+/* Need an unsigned type capable of holding 32 bits; */
+
+typedef uint32_t UNS_32_BITS;
+
+/* Copyright (C) 1986 Gary S. Brown.  You may use this program, or
+   code or tables extracted from it, as desired without restriction.*/
+
+/* First, the polynomial itself and its table of feedback terms.  The  */
+/* polynomial is                                                       */
+/* X^32+X^26+X^23+X^22+X^16+X^12+X^11+X^10+X^8+X^7+X^5+X^4+X^2+X^1+X^0 */
+/* Note that we take it "backwards" and put the highest-order term in  */
+/* the lowest-order bit.  The X^32 term is "implied"; the LSB is the   */
+/* X^31 term, etc.  The X^0 term (usually shown as "+1") results in    */
+/* the MSB being 1.                                                    */
+
+/* Note that the usual hardware shift register implementation, which   */
+/* is what we're using (we're merely optimizing it by doing eight-bit  */
+/* chunks at a time) shifts bits into the lowest-order term.  In our   */
+/* implementation, that means shifting towards the right.  Why do we   */
+/* do it this way?  Because the calculated CRC must be transmitted in  */
+/* order from highest-order term to lowest-order term.  UARTs transmit */
+/* characters in order from LSB to MSB.  By storing the CRC this way,  */
+/* we hand it to the UART in the order low-byte to high-byte; the UART */
+/* sends each low-bit to hight-bit; and the result is transmission bit */
+/* by bit from highest- to lowest-order term without requiring any bit */
+/* shuffling on our part.  Reception works similarly.                  */
+
+/* The feedback terms table consists of 256, 32-bit entries.  Notes:   */
+/*                                                                     */
+/*  1. The table can be generated at runtime if desired; code to do so */
+/*     is shown later.  It might not be obvious, but the feedback      */
+/*     terms simply represent the results of eight shift/xor opera-    */
+/*     tions for all combinations of data and CRC register values.     */
+/*                                                                     */
+/*  2. The CRC accumulation logic is the same for all CRC polynomials, */
+/*     be they sixteen or thirty-two bits wide.  You simply choose the */
+/*     appropriate table.  Alternatively, because the table can be     */
+/*     generated at runtime, you can start by generating the table for */
+/*     the polynomial in question and use exactly the same "updcrc",   */
+/*     if your application needn't simultaneously handle two CRC       */
+/*     polynomials.  (Note, however, that XMODEM is strange.)          */
+/*                                                                     */
+/*  3. For 16-bit CRCs, the table entries need be only 16 bits wide;   */
+/*     of course, 32-bit entries work OK if the high 16 bits are zero. */
+/*                                                                     */
+/*  4. The values must be right-shifted by eight bits by the "updcrc"  */
+/*     logic; the shift must be unsigned (bring in zeroes).  On some   */
+/*     hardware you could probably optimize the shift in assembler by  */
+/*     using byte-swap instructions.                                   */
+
+static UNS_32_BITS crc_32_tab[] = { /* CRC polynomial 0xedb88320 */
+0x00000000, 0x77073096, 0xee0e612c, 0x990951ba, 0x076dc419, 0x706af48f,
+0xe963a535, 0x9e6495a3, 0x0edb8832, 0x79dcb8a4, 0xe0d5e91e, 0x97d2d988,
+0x09b64c2b, 0x7eb17cbd, 0xe7b82d07, 0x90bf1d91, 0x1db71064, 0x6ab020f2,
+0xf3b97148, 0x84be41de, 0x1adad47d, 0x6ddde4eb, 0xf4d4b551, 0x83d385c7,
+0x136c9856, 0x646ba8c0, 0xfd62f97a, 0x8a65c9ec, 0x14015c4f, 0x63066cd9,
+0xfa0f3d63, 0x8d080df5, 0x3b6e20c8, 0x4c69105e, 0xd56041e4, 0xa2677172,
+0x3c03e4d1, 0x4b04d447, 0xd20d85fd, 0xa50ab56b, 0x35b5a8fa, 0x42b2986c,
+0xdbbbc9d6, 0xacbcf940, 0x32d86ce3, 0x45df5c75, 0xdcd60dcf, 0xabd13d59,
+0x26d930ac, 0x51de003a, 0xc8d75180, 0xbfd06116, 0x21b4f4b5, 0x56b3c423,
+0xcfba9599, 0xb8bda50f, 0x2802b89e, 0x5f058808, 0xc60cd9b2, 0xb10be924,
+0x2f6f7c87, 0x58684c11, 0xc1611dab, 0xb6662d3d, 0x76dc4190, 0x01db7106,
+0x98d220bc, 0xefd5102a, 0x71b18589, 0x06b6b51f, 0x9fbfe4a5, 0xe8b8d433,
+0x7807c9a2, 0x0f00f934, 0x9609a88e, 0xe10e9818, 0x7f6a0dbb, 0x086d3d2d,
+0x91646c97, 0xe6635c01, 0x6b6b51f4, 0x1c6c6162, 0x856530d8, 0xf262004e,
+0x6c0695ed, 0x1b01a57b, 0x8208f4c1, 0xf50fc457, 0x65b0d9c6, 0x12b7e950,
+0x8bbeb8ea, 0xfcb9887c, 0x62dd1ddf, 0x15da2d49, 0x8cd37cf3, 0xfbd44c65,
+0x4db26158, 0x3ab551ce, 0xa3bc0074, 0xd4bb30e2, 0x4adfa541, 0x3dd895d7,
+0xa4d1c46d, 0xd3d6f4fb, 0x4369e96a, 0x346ed9fc, 0xad678846, 0xda60b8d0,
+0x44042d73, 0x33031de5, 0xaa0a4c5f, 0xdd0d7cc9, 0x5005713c, 0x270241aa,
+0xbe0b1010, 0xc90c2086, 0x5768b525, 0x206f85b3, 0xb966d409, 0xce61e49f,
+0x5edef90e, 0x29d9c998, 0xb0d09822, 0xc7d7a8b4, 0x59b33d17, 0x2eb40d81,
+0xb7bd5c3b, 0xc0ba6cad, 0xedb88320, 0x9abfb3b6, 0x03b6e20c, 0x74b1d29a,
+0xead54739, 0x9dd277af, 0x04db2615, 0x73dc1683, 0xe3630b12, 0x94643b84,
+0x0d6d6a3e, 0x7a6a5aa8, 0xe40ecf0b, 0x9309ff9d, 0x0a00ae27, 0x7d079eb1,
+0xf00f9344, 0x8708a3d2, 0x1e01f268, 0x6906c2fe, 0xf762575d, 0x806567cb,
+0x196c3671, 0x6e6b06e7, 0xfed41b76, 0x89d32be0, 0x10da7a5a, 0x67dd4acc,
+0xf9b9df6f, 0x8ebeeff9, 0x17b7be43, 0x60b08ed5, 0xd6d6a3e8, 0xa1d1937e,
+0x38d8c2c4, 0x4fdff252, 0xd1bb67f1, 0xa6bc5767, 0x3fb506dd, 0x48b2364b,
+0xd80d2bda, 0xaf0a1b4c, 0x36034af6, 0x41047a60, 0xdf60efc3, 0xa867df55,
+0x316e8eef, 0x4669be79, 0xcb61b38c, 0xbc66831a, 0x256fd2a0, 0x5268e236,
+0xcc0c7795, 0xbb0b4703, 0x220216b9, 0x5505262f, 0xc5ba3bbe, 0xb2bd0b28,
+0x2bb45a92, 0x5cb36a04, 0xc2d7ffa7, 0xb5d0cf31, 0x2cd99e8b, 0x5bdeae1d,
+0x9b64c2b0, 0xec63f226, 0x756aa39c, 0x026d930a, 0x9c0906a9, 0xeb0e363f,
+0x72076785, 0x05005713, 0x95bf4a82, 0xe2b87a14, 0x7bb12bae, 0x0cb61b38,
+0x92d28e9b, 0xe5d5be0d, 0x7cdcefb7, 0x0bdbdf21, 0x86d3d2d4, 0xf1d4e242,
+0x68ddb3f8, 0x1fda836e, 0x81be16cd, 0xf6b9265b, 0x6fb077e1, 0x18b74777,
+0x88085ae6, 0xff0f6a70, 0x66063bca, 0x11010b5c, 0x8f659eff, 0xf862ae69,
+0x616bffd3, 0x166ccf45, 0xa00ae278, 0xd70dd2ee, 0x4e048354, 0x3903b3c2,
+0xa7672661, 0xd06016f7, 0x4969474d, 0x3e6e77db, 0xaed16a4a, 0xd9d65adc,
+0x40df0b66, 0x37d83bf0, 0xa9bcae53, 0xdebb9ec5, 0x47b2cf7f, 0x30b5ffe9,
+0xbdbdf21c, 0xcabac28a, 0x53b39330, 0x24b4a3a6, 0xbad03605, 0xcdd70693,
+0x54de5729, 0x23d967bf, 0xb3667a2e, 0xc4614ab8, 0x5d681b02, 0x2a6f2b94,
+0xb40bbe37, 0xc30c8ea1, 0x5a05df1b, 0x2d02ef8d
+};
+
+uint32_t updateCRC32(unsigned char ch, uint32_t crc)
+{
+      return UPDC32(ch, crc);
+}
+
+bool crc32file(const char *name, uint32_t *crc, long *charcnt)
+{
+      register FILE *fin;
+      register uint32_t oldcrc32;
+      register int c;
+
+      oldcrc32 = 0xFFFFFFFF; *charcnt = 0;
+      if ((fin=fopen(name, "r"))==NULL)
+      {
+            perror(name);
+            return false;
+      }
+      while ((c=getc(fin))!=EOF)
+      {
+            ++*charcnt;
+            oldcrc32 = UPDC32(c, oldcrc32);
+      }
+
+      if (ferror(fin))
+      {
+            perror(name);
+            *charcnt = -1;
+      }
+      fclose(fin);
+
+      *crc = oldcrc32 = ~oldcrc32;
+
+      return true;
+}
+
+uint32_t crc32buf(char *buf, size_t len)
+{
+      register uint32_t oldcrc32;
+
+      oldcrc32 = 0xFFFFFFFF;
+
+      for ( ; len; --len, ++buf)
+      {
+            oldcrc32 = UPDC32(*buf, oldcrc32);
+      }
+
+      return ~oldcrc32;
+}

util/crc32.h

@@ -0,0 +1,16 @@
+/* taken from https://github.com/panzi/CRC-and-checksum-functions/ */
+#ifndef __CRC32_H__
+#define __CRC32_H__
+
+#include <stdlib.h>           /* For size_t                           */
+#include <stdint.h>           /* For uint8_t, uint16_t, uint32_t      */
+#include <stdbool.h>          /* For bool, true, false                */
+
+#define UPDC32(octet,crc) (crc_32_tab[((crc)\
+     ^ ((uint8_t)octet)) & 0xff] ^ ((crc) >> 8))
+
+uint32_t updateCRC32(unsigned char ch, uint32_t crc);
+bool crc32file(const char *name, uint32_t *crc, long *charcnt);
+uint32_t crc32buf(char *buf, size_t len);
+
+#endif

util/fuzzy_search.cc

@@ -1,62 +0,0 @@
-#include <memory.h>
-#include <stdint.h>
-
-#include "util/fuzzy_search.h"
-
-int find_block_core(char *haystack, size_t haystack_size, fuzzy_search_task *needle,
-                    size_t orig_offset, int dir) {
-    size_t offset = orig_offset;
-    haystack_size += orig_offset;
-    while (offset + 1 < haystack_size) {
-        size_t offset_temp = offset;
-        int found = 1;
-
-        if (needle->count <= 0) {
-            found = 0;
-        }
-
-        for (int i = 0; i < needle->count; i++) {
-            int subfound = -1;
-
-            if (needle->blocks[i].type == 1) {
-                offset_temp += needle->blocks[i].length;
-                continue;
-            }
-
-            for (int j = 0; j < needle->blocks[i].count; j++) {
-                if (haystack + offset_temp + needle->blocks[i].data[j].length <
-                        haystack + haystack_size &&
-                    memcmp(haystack + offset_temp, needle->blocks[i].data[j].block,
-                           needle->blocks[i].data[j].length) == 0) {
-                    subfound = j;
-                    break;
-                }
-            }
-
-            if (subfound == -1) {
-                found = 0;
-                break;
-            }
-
-            offset_temp += needle->blocks[i].data[subfound].length;
-        }
-
-        if (found == 1) {
-            return offset;
-        } else {
-            offset += dir;
-        }
-    }
-
-    return -1;
-}
-
-int find_block(char *haystack, size_t haystack_size, fuzzy_search_task *needle,
-               size_t orig_offset) {
-    return find_block_core(haystack, haystack_size, needle, orig_offset, 1);
-}
-
-int find_block_back(char *haystack, size_t haystack_size, fuzzy_search_task *needle,
-                    size_t orig_offset) {
-    return find_block_core(haystack, haystack_size, needle, orig_offset, -1);
-}

util/fuzzy_search.h

@@ -1,53 +0,0 @@
-#ifndef __FUZZY_SEARCH_H__
-#define __FUZZY_SEARCH_H__
-
-#include <stddef.h>
-
-#define FUZZY_START(task, _count)                                                                  \
-    {                                                                                              \
-        task.count = _count;                                                                       \
-        task.blocks = (fuzzy_search_block *)calloc(task.count, sizeof(fuzzy_search_block));        \
-    }
-
-#define FUZZY_CODE(task, id, code, len)                                                            \
-    {                                                                                              \
-        task.blocks[id].count = 1;                                                                 \
-        task.blocks[id].data[0].type = 0;                                                          \
-        task.blocks[id].data[0].length = len;                                                      \
-        task.blocks[id].data[0].block =                                                            \
-            (char *)calloc(task.blocks[id].data[0].length, sizeof(char));                          \
-        memcpy(task.blocks[id].data[0].block, code, task.blocks[id].data[0].length);               \
-    }
-
-#define FUZZY_WILDCARD(task, id, len)                                                              \
-    {                                                                                              \
-        task.blocks[id].count = 1;                                                                 \
-        task.blocks[id].type = 1;                                                                  \
-        task.blocks[id].length = len;                                                              \
-        task.blocks[id].data[0].length = len;                                                      \
-    }
-
-#define FUZZY_SIZE(task, id) (task.blocks[id].data[0].length)
-
-typedef struct fuzzy_search_block_data {
-    int type;
-    int length;
-    char *block;
-} fuzzy_search_block_data;
-
-typedef struct fuzzy_search_block {
-    int type;
-    int length;
-    int count;
-    fuzzy_search_block_data data[0x10];
-} fuzzy_search_block;
-
-typedef struct fuzzy_search_task {
-    int count;
-    fuzzy_search_block *blocks;
-} fuzzy_search_task;
-
-int find_block(char *haystack, size_t haystack_size, fuzzy_search_task *needle, size_t offset);
-int find_block_back(char *haystack, size_t haystack_size, fuzzy_search_task *needle, size_t offset);
-
-#endif

util/search.cc

@@ -26,7 +26,7 @@ static void badCharHeuristic(const unsigned char *str, int size, int* badchar) {
 
 #define DEBUG_SEARCH 0
 
-int _search(unsigned char *haystack, size_t haystack_size, const unsigned char *needle, size_t needle_size, int orig_offset, int debug) {
+int _search(unsigned char *haystack, size_t haystack_size, const unsigned char *needle, size_t needle_size, int orig_offset, bool wildcards, int debug) {
     int badchar[NO_OF_CHARS];
 
     badCharHeuristic(needle, needle_size, badchar);
@@ -46,12 +46,21 @@ int _search(unsigned char *haystack, size_t haystack_size, const unsigned char *
         LOG("pat...");
         for (size_t i = 0; i < needle_size; i++)
         {
+            if (wildcards && needle[i] == '?')
+                LOG("** ");
+            else
                 LOG("%02x ", needle[i]);
         }
         LOG("\n");
  }
-        while (j >= 0 && needle[j] == haystack[orig_offset + s + j])
+        if ( wildcards )
+        {
+            while (j >= 0 && ( needle[j] == '?' || needle[j] == haystack[orig_offset + s + j]) )
                 j--;
+        } else {
+            while (j >= 0 && ( needle[j] == haystack[orig_offset + s + j]) )
+                j--;
+        }
 
         if (j < 0) {
                 if (debug)
@@ -70,6 +79,21 @@ int _search(unsigned char *haystack, size_t haystack_size, const unsigned char *
 }
 
 int search(char *haystack, size_t haystack_size, const char *needle, size_t needle_size, size_t orig_offset) {
-    int res = _search((unsigned char*) haystack, haystack_size, (const unsigned char *)needle, needle_size, orig_offset, 0);
+    int res = _search((unsigned char*) haystack, haystack_size, (const unsigned char *)needle, needle_size, orig_offset, false, 0);
+    return res;
+}
+
+int wildcard_search(char *haystack, size_t haystack_size, const char *needle, size_t needle_size, size_t orig_offset) {
+    int res = _search((unsigned char*) haystack, haystack_size, (const unsigned char *)needle, needle_size, orig_offset, true, 0);
+    return res;
+}
+
+int search_debug(char *haystack, size_t haystack_size, const char *needle, size_t needle_size, size_t orig_offset) {
+    int res = _search((unsigned char*) haystack, haystack_size, (const unsigned char *)needle, needle_size, orig_offset, false, 2);
+    return res;
+}
+
+int wildcard_search_debug(char *haystack, size_t haystack_size, const char *needle, size_t needle_size, size_t orig_offset) {
+    int res = _search((unsigned char*) haystack, haystack_size, (const unsigned char *)needle, needle_size, orig_offset, true, 2);
     return res;
 }

util/search.h

@@ -2,5 +2,8 @@
 #define __SEARCH_H__
 
 int search(char *haystack, size_t haystack_size, const char *needle, size_t needle_size, size_t orig_offset);
+int search_debug(char *haystack, size_t haystack_size, const char *needle, size_t needle_size, size_t orig_offset);
+int wildcard_search(char *haystack, size_t haystack_size, const char *needle, size_t needle_size, size_t orig_offset);
+int wildcard_search_debug(char *haystack, size_t haystack_size, const char *needle, size_t needle_size, size_t orig_offset);
 
 #endif