diff third_party/luajit/dynasm/dasm_x86.lua @ 178:94705b5986b3

[ThirdParty] Added WRK and luajit for load testing.
author MrJuneJune <me@mrjunejune.com>
date Thu, 22 Jan 2026 20:10:30 -0800
parents
children
line wrap: on
line diff
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/third_party/luajit/dynasm/dasm_x86.lua	Thu Jan 22 20:10:30 2026 -0800
@@ -0,0 +1,2388 @@
+------------------------------------------------------------------------------
+-- DynASM x86/x64 module.
+--
+-- Copyright (C) 2005-2023 Mike Pall. All rights reserved.
+-- See dynasm.lua for full copyright notice.
+------------------------------------------------------------------------------
+
+local x64 = x64
+
+-- Module information:
+local _info = {
+  arch =	x64 and "x64" or "x86",
+  description =	"DynASM x86/x64 module",
+  version =	"1.5.0",
+  vernum =	 10500,
+  release =	"2021-05-02",
+  author =	"Mike Pall",
+  license =	"MIT",
+}
+
+-- Exported glue functions for the arch-specific module.
+local _M = { _info = _info }
+
+-- Cache library functions.
+local type, tonumber, pairs, ipairs = type, tonumber, pairs, ipairs
+local assert, unpack, setmetatable = assert, unpack or table.unpack, setmetatable
+local _s = string
+local sub, format, byte, char = _s.sub, _s.format, _s.byte, _s.char
+local find, match, gmatch, gsub = _s.find, _s.match, _s.gmatch, _s.gsub
+local concat, sort, remove = table.concat, table.sort, table.remove
+local bit = bit or require("bit")
+local band, bxor, shl, shr = bit.band, bit.bxor, bit.lshift, bit.rshift
+
+-- Inherited tables and callbacks.
+local g_opt, g_arch
+local wline, werror, wfatal, wwarn
+
+-- Action name list.
+-- CHECK: Keep this in sync with the C code!
+local action_names = {
+  -- int arg, 1 buffer pos:
+  "DISP",  "IMM_S", "IMM_B", "IMM_W", "IMM_D",  "IMM_WB", "IMM_DB",
+  -- action arg (1 byte), int arg, 1 buffer pos (reg/num):
+  "VREG", "SPACE",
+  -- ptrdiff_t arg, 1 buffer pos (address): !x64
+  "SETLABEL", "REL_A",
+  -- action arg (1 byte) or int arg, 2 buffer pos (link, offset):
+  "REL_LG", "REL_PC",
+  -- action arg (1 byte) or int arg, 1 buffer pos (link):
+  "IMM_LG", "IMM_PC",
+  -- action arg (1 byte) or int arg, 1 buffer pos (offset):
+  "LABEL_LG", "LABEL_PC",
+  -- action arg (1 byte), 1 buffer pos (offset):
+  "ALIGN",
+  -- action args (2 bytes), no buffer pos.
+  "EXTERN",
+  -- action arg (1 byte), no buffer pos.
+  "ESC",
+  -- no action arg, no buffer pos.
+  "MARK",
+  -- action arg (1 byte), no buffer pos, terminal action:
+  "SECTION",
+  -- no args, no buffer pos, terminal action:
+  "STOP"
+}
+
+-- Maximum number of section buffer positions for dasm_put().
+-- CHECK: Keep this in sync with the C code!
+local maxsecpos = 25 -- Keep this low, to avoid excessively long C lines.
+
+-- Action name -> action number (dynamically generated below).
+local map_action = {}
+-- First action number. Everything below does not need to be escaped.
+local actfirst = 256-#action_names
+
+-- Action list buffer and string (only used to remove dupes).
+local actlist = {}
+local actstr = ""
+
+-- Argument list for next dasm_put(). Start with offset 0 into action list.
+local actargs = { 0 }
+
+-- Current number of section buffer positions for dasm_put().
+local secpos = 1
+
+-- VREG kind encodings, pre-shifted by 5 bits.
+local map_vreg = {
+  ["modrm.rm.m"] = 0x00,
+  ["modrm.rm.r"] = 0x20,
+  ["opcode"] =     0x20,
+  ["sib.base"] =   0x20,
+  ["sib.index"] =  0x40,
+  ["modrm.reg"] =  0x80,
+  ["vex.v"] =      0xa0,
+  ["imm.hi"] =     0xc0,
+}
+
+-- Current number of VREG actions contributing to REX/VEX shrinkage.
+local vreg_shrink_count = 0
+
+------------------------------------------------------------------------------
+
+-- Compute action numbers for action names.
+for n,name in ipairs(action_names) do
+  local num = actfirst + n - 1
+  map_action[name] = num
+end
+
+-- Dump action names and numbers.
+local function dumpactions(out)
+  out:write("DynASM encoding engine action codes:\n")
+  for n,name in ipairs(action_names) do
+    local num = map_action[name]
+    out:write(format("  %-10s %02X  %d\n", name, num, num))
+  end
+  out:write("\n")
+end
+
+-- Write action list buffer as a huge static C array.
+local function writeactions(out, name)
+  local nn = #actlist
+  local last = actlist[nn] or 255
+  actlist[nn] = nil -- Remove last byte.
+  if nn == 0 then nn = 1 end
+  out:write("static const unsigned char ", name, "[", nn, "] = {\n")
+  local s = "  "
+  for n,b in ipairs(actlist) do
+    s = s..b..","
+    if #s >= 75 then
+      assert(out:write(s, "\n"))
+      s = "  "
+    end
+  end
+  out:write(s, last, "\n};\n\n") -- Add last byte back.
+end
+
+------------------------------------------------------------------------------
+
+-- Add byte to action list.
+local function wputxb(n)
+  assert(n >= 0 and n <= 255 and n % 1 == 0, "byte out of range")
+  actlist[#actlist+1] = n
+end
+
+-- Add action to list with optional arg. Advance buffer pos, too.
+local function waction(action, a, num)
+  wputxb(assert(map_action[action], "bad action name `"..action.."'"))
+  if a then actargs[#actargs+1] = a end
+  if a or num then secpos = secpos + (num or 1) end
+end
+
+-- Optionally add a VREG action.
+local function wvreg(kind, vreg, psz, sk, defer)
+  if not vreg then return end
+  waction("VREG", vreg)
+  local b = assert(map_vreg[kind], "bad vreg kind `"..vreg.."'")
+  if b < (sk or 0) then
+    vreg_shrink_count = vreg_shrink_count + 1
+  end
+  if not defer then
+    b = b + vreg_shrink_count * 8
+    vreg_shrink_count = 0
+  end
+  wputxb(b + (psz or 0))
+end
+
+-- Add call to embedded DynASM C code.
+local function wcall(func, args)
+  wline(format("dasm_%s(Dst, %s);", func, concat(args, ", ")), true)
+end
+
+-- Delete duplicate action list chunks. A tad slow, but so what.
+local function dedupechunk(offset)
+  local al, as = actlist, actstr
+  local chunk = char(unpack(al, offset+1, #al))
+  local orig = find(as, chunk, 1, true)
+  if orig then
+    actargs[1] = orig-1 -- Replace with original offset.
+    for i=offset+1,#al do al[i] = nil end -- Kill dupe.
+  else
+    actstr = as..chunk
+  end
+end
+
+-- Flush action list (intervening C code or buffer pos overflow).
+local function wflush(term)
+  local offset = actargs[1]
+  if #actlist == offset then return end -- Nothing to flush.
+  if not term then waction("STOP") end -- Terminate action list.
+  dedupechunk(offset)
+  wcall("put", actargs) -- Add call to dasm_put().
+  actargs = { #actlist } -- Actionlist offset is 1st arg to next dasm_put().
+  secpos = 1 -- The actionlist offset occupies a buffer position, too.
+end
+
+-- Put escaped byte.
+local function wputb(n)
+  if n >= actfirst then waction("ESC") end -- Need to escape byte.
+  wputxb(n)
+end
+
+------------------------------------------------------------------------------
+
+-- Global label name -> global label number. With auto assignment on 1st use.
+local next_global = 10
+local map_global = setmetatable({}, { __index = function(t, name)
+  if not match(name, "^[%a_][%w_@]*$") then werror("bad global label") end
+  local n = next_global
+  if n > 246 then werror("too many global labels") end
+  next_global = n + 1
+  t[name] = n
+  return n
+end})
+
+-- Dump global labels.
+local function dumpglobals(out, lvl)
+  local t = {}
+  for name, n in pairs(map_global) do t[n] = name end
+  out:write("Global labels:\n")
+  for i=10,next_global-1 do
+    out:write(format("  %s\n", t[i]))
+  end
+  out:write("\n")
+end
+
+-- Write global label enum.
+local function writeglobals(out, prefix)
+  local t = {}
+  for name, n in pairs(map_global) do t[n] = name end
+  out:write("enum {\n")
+  for i=10,next_global-1 do
+    out:write("  ", prefix, gsub(t[i], "@.*", ""), ",\n")
+  end
+  out:write("  ", prefix, "_MAX\n};\n")
+end
+
+-- Write global label names.
+local function writeglobalnames(out, name)
+  local t = {}
+  for name, n in pairs(map_global) do t[n] = name end
+  out:write("static const char *const ", name, "[] = {\n")
+  for i=10,next_global-1 do
+    out:write("  \"", t[i], "\",\n")
+  end
+  out:write("  (const char *)0\n};\n")
+end
+
+------------------------------------------------------------------------------
+
+-- Extern label name -> extern label number. With auto assignment on 1st use.
+local next_extern = -1
+local map_extern = setmetatable({}, { __index = function(t, name)
+  -- No restrictions on the name for now.
+  local n = next_extern
+  if n < -256 then werror("too many extern labels") end
+  next_extern = n - 1
+  t[name] = n
+  return n
+end})
+
+-- Dump extern labels.
+local function dumpexterns(out, lvl)
+  local t = {}
+  for name, n in pairs(map_extern) do t[-n] = name end
+  out:write("Extern labels:\n")
+  for i=1,-next_extern-1 do
+    out:write(format("  %s\n", t[i]))
+  end
+  out:write("\n")
+end
+
+-- Write extern label names.
+local function writeexternnames(out, name)
+  local t = {}
+  for name, n in pairs(map_extern) do t[-n] = name end
+  out:write("static const char *const ", name, "[] = {\n")
+  for i=1,-next_extern-1 do
+    out:write("  \"", t[i], "\",\n")
+  end
+  out:write("  (const char *)0\n};\n")
+end
+
+------------------------------------------------------------------------------
+
+-- Arch-specific maps.
+local map_archdef = {}		-- Ext. register name -> int. name.
+local map_reg_rev = {}		-- Int. register name -> ext. name.
+local map_reg_num = {}		-- Int. register name -> register number.
+local map_reg_opsize = {}	-- Int. register name -> operand size.
+local map_reg_valid_base = {}	-- Int. register name -> valid base register?
+local map_reg_valid_index = {}	-- Int. register name -> valid index register?
+local map_reg_needrex = {}	-- Int. register name -> need rex vs. no rex.
+local reg_list = {}		-- Canonical list of int. register names.
+
+local map_type = {}		-- Type name -> { ctype, reg }
+local ctypenum = 0		-- Type number (for _PTx macros).
+
+local addrsize = x64 and "q" or "d"	-- Size for address operands.
+
+-- Helper functions to fill register maps.
+local function mkrmap(sz, cl, names)
+  local cname = format("@%s", sz)
+  reg_list[#reg_list+1] = cname
+  map_archdef[cl] = cname
+  map_reg_rev[cname] = cl
+  map_reg_num[cname] = -1
+  map_reg_opsize[cname] = sz
+  if sz == addrsize or sz == "d" then
+    map_reg_valid_base[cname] = true
+    map_reg_valid_index[cname] = true
+  end
+  if names then
+    for n,name in ipairs(names) do
+      local iname = format("@%s%x", sz, n-1)
+      reg_list[#reg_list+1] = iname
+      map_archdef[name] = iname
+      map_reg_rev[iname] = name
+      map_reg_num[iname] = n-1
+      map_reg_opsize[iname] = sz
+      if sz == "b" and n > 4 then map_reg_needrex[iname] = false end
+      if sz == addrsize or sz == "d" then
+	map_reg_valid_base[iname] = true
+	map_reg_valid_index[iname] = true
+      end
+    end
+  end
+  for i=0,(x64 and sz ~= "f") and 15 or 7 do
+    local needrex = sz == "b" and i > 3
+    local iname = format("@%s%x%s", sz, i, needrex and "R" or "")
+    if needrex then map_reg_needrex[iname] = true end
+    local name
+    if sz == "o" or sz == "y" then name = format("%s%d", cl, i)
+    elseif sz == "f" then name = format("st%d", i)
+    else name = format("r%d%s", i, sz == addrsize and "" or sz) end
+    map_archdef[name] = iname
+    if not map_reg_rev[iname] then
+      reg_list[#reg_list+1] = iname
+      map_reg_rev[iname] = name
+      map_reg_num[iname] = i
+      map_reg_opsize[iname] = sz
+      if sz == addrsize or sz == "d" then
+	map_reg_valid_base[iname] = true
+	map_reg_valid_index[iname] = true
+      end
+    end
+  end
+  reg_list[#reg_list+1] = ""
+end
+
+-- Integer registers (qword, dword, word and byte sized).
+if x64 then
+  mkrmap("q", "Rq", {"rax", "rcx", "rdx", "rbx", "rsp", "rbp", "rsi", "rdi"})
+end
+mkrmap("d", "Rd", {"eax", "ecx", "edx", "ebx", "esp", "ebp", "esi", "edi"})
+mkrmap("w", "Rw", {"ax", "cx", "dx", "bx", "sp", "bp", "si", "di"})
+mkrmap("b", "Rb", {"al", "cl", "dl", "bl", "ah", "ch", "dh", "bh"})
+map_reg_valid_index[map_archdef.esp] = false
+if x64 then map_reg_valid_index[map_archdef.rsp] = false end
+if x64 then map_reg_needrex[map_archdef.Rb] = true end
+map_archdef["Ra"] = "@"..addrsize
+
+-- FP registers (internally tword sized, but use "f" as operand size).
+mkrmap("f", "Rf")
+
+-- SSE registers (oword sized, but qword and dword accessible).
+mkrmap("o", "xmm")
+
+-- AVX registers (yword sized, but oword, qword and dword accessible).
+mkrmap("y", "ymm")
+
+-- Operand size prefixes to codes.
+local map_opsize = {
+  byte = "b", word = "w", dword = "d", qword = "q", oword = "o", yword = "y",
+  tword = "t", aword = addrsize,
+}
+
+-- Operand size code to number.
+local map_opsizenum = {
+  b = 1, w = 2, d = 4, q = 8, o = 16, y = 32, t = 10,
+}
+
+-- Operand size code to name.
+local map_opsizename = {
+  b = "byte", w = "word", d = "dword", q = "qword", o = "oword", y = "yword",
+  t = "tword", f = "fpword",
+}
+
+-- Valid index register scale factors.
+local map_xsc = {
+  ["1"] = 0, ["2"] = 1, ["4"] = 2, ["8"] = 3,
+}
+
+-- Condition codes.
+local map_cc = {
+  o = 0, no = 1, b = 2, nb = 3, e = 4, ne = 5, be = 6, nbe = 7,
+  s = 8, ns = 9, p = 10, np = 11, l = 12, nl = 13, le = 14, nle = 15,
+  c = 2, nae = 2, nc = 3, ae = 3, z = 4, nz = 5, na = 6, a = 7,
+  pe = 10, po = 11, nge = 12, ge = 13, ng = 14, g = 15,
+}
+
+
+-- Reverse defines for registers.
+function _M.revdef(s)
+  return gsub(s, "@%w+", map_reg_rev)
+end
+
+-- Dump register names and numbers
+local function dumpregs(out)
+  out:write("Register names, sizes and internal numbers:\n")
+  for _,reg in ipairs(reg_list) do
+    if reg == "" then
+      out:write("\n")
+    else
+      local name = map_reg_rev[reg]
+      local num = map_reg_num[reg]
+      local opsize = map_opsizename[map_reg_opsize[reg]]
+      out:write(format("  %-5s %-8s %s\n", name, opsize,
+		       num < 0 and "(variable)" or num))
+    end
+  end
+end
+
+------------------------------------------------------------------------------
+
+-- Put action for label arg (IMM_LG, IMM_PC, REL_LG, REL_PC).
+local function wputlabel(aprefix, imm, num)
+  if type(imm) == "number" then
+    if imm < 0 then
+      waction("EXTERN")
+      wputxb(aprefix == "IMM_" and 0 or 1)
+      imm = -imm-1
+    else
+      waction(aprefix.."LG", nil, num);
+    end
+    wputxb(imm)
+  else
+    waction(aprefix.."PC", imm, num)
+  end
+end
+
+-- Put signed byte or arg.
+local function wputsbarg(n)
+  if type(n) == "number" then
+    if n < -128 or n > 127 then
+      werror("signed immediate byte out of range")
+    end
+    if n < 0 then n = n + 256 end
+    wputb(n)
+  else waction("IMM_S", n) end
+end
+
+-- Put unsigned byte or arg.
+local function wputbarg(n)
+  if type(n) == "number" then
+    if n < 0 or n > 255 then
+      werror("unsigned immediate byte out of range")
+    end
+    wputb(n)
+  else waction("IMM_B", n) end
+end
+
+-- Put unsigned word or arg.
+local function wputwarg(n)
+  if type(n) == "number" then
+    if shr(n, 16) ~= 0 then
+      werror("unsigned immediate word out of range")
+    end
+    wputb(band(n, 255)); wputb(shr(n, 8));
+  else waction("IMM_W", n) end
+end
+
+-- Put signed or unsigned dword or arg.
+local function wputdarg(n)
+  local tn = type(n)
+  if tn == "number" then
+    wputb(band(n, 255))
+    wputb(band(shr(n, 8), 255))
+    wputb(band(shr(n, 16), 255))
+    wputb(shr(n, 24))
+  elseif tn == "table" then
+    wputlabel("IMM_", n[1], 1)
+  else
+    waction("IMM_D", n)
+  end
+end
+
+-- Put signed or unsigned qword or arg.
+local function wputqarg(n)
+  local tn = type(n)
+  if tn == "number" then -- This is only used for numbers from -2^31..2^32-1.
+    wputb(band(n, 255))
+    wputb(band(shr(n, 8), 255))
+    wputb(band(shr(n, 16), 255))
+    wputb(shr(n, 24))
+    local sign = n < 0 and 255 or 0
+    wputb(sign); wputb(sign); wputb(sign); wputb(sign)
+  else
+    waction("IMM_D", format("(unsigned int)(%s)", n))
+    waction("IMM_D", format("(unsigned int)((unsigned long long)(%s)>>32)", n))
+  end
+end
+
+-- Put operand-size dependent number or arg (defaults to dword).
+local function wputszarg(sz, n)
+  if not sz or sz == "d" or sz == "q" then wputdarg(n)
+  elseif sz == "w" then wputwarg(n)
+  elseif sz == "b" then wputbarg(n)
+  elseif sz == "s" then wputsbarg(n)
+  else werror("bad operand size") end
+end
+
+-- Put multi-byte opcode with operand-size dependent modifications.
+local function wputop(sz, op, rex, vex, vregr, vregxb)
+  local psz, sk = 0, nil
+  if vex then
+    local tail
+    if vex.m == 1 and band(rex, 11) == 0 then
+      if x64 and vregxb then
+	sk = map_vreg["modrm.reg"]
+      else
+	wputb(0xc5)
+      tail = shl(bxor(band(rex, 4), 4), 5)
+      psz = 3
+      end
+    end
+    if not tail then
+      wputb(0xc4)
+      wputb(shl(bxor(band(rex, 7), 7), 5) + vex.m)
+      tail = shl(band(rex, 8), 4)
+      psz = 4
+    end
+    local reg, vreg = 0, nil
+    if vex.v then
+      reg = vex.v.reg
+      if not reg then werror("bad vex operand") end
+      if reg < 0 then reg = 0; vreg = vex.v.vreg end
+    end
+    if sz == "y" or vex.l then tail = tail + 4 end
+    wputb(tail + shl(bxor(reg, 15), 3) + vex.p)
+    wvreg("vex.v", vreg)
+    rex = 0
+    if op >= 256 then werror("bad vex opcode") end
+  else
+    if rex ~= 0 then
+      if not x64 then werror("bad operand size") end
+    elseif (vregr or vregxb) and x64 then
+      rex = 0x10
+      sk = map_vreg["vex.v"]
+    end
+  end
+  local r
+  if sz == "w" then wputb(102) end
+  -- Needs >32 bit numbers, but only for crc32 eax, word [ebx]
+  if op >= 4294967296 then r = op%4294967296 wputb((op-r)/4294967296) op = r end
+  if op >= 16777216 then wputb(shr(op, 24)); op = band(op, 0xffffff) end
+  if op >= 65536 then
+    if rex ~= 0 then
+      local opc3 = band(op, 0xffff00)
+      if opc3 == 0x0f3a00 or opc3 == 0x0f3800 then
+	wputb(64 + band(rex, 15)); rex = 0; psz = 2
+      end
+    end
+    wputb(shr(op, 16)); op = band(op, 0xffff); psz = psz + 1
+  end
+  if op >= 256 then
+    local b = shr(op, 8)
+    if b == 15 and rex ~= 0 then wputb(64 + band(rex, 15)); rex = 0; psz = 2 end
+    wputb(b); op = band(op, 255); psz = psz + 1
+  end
+  if rex ~= 0 then wputb(64 + band(rex, 15)); psz = 2 end
+  if sz == "b" then op = op - 1 end
+  wputb(op)
+  return psz, sk
+end
+
+-- Put ModRM or SIB formatted byte.
+local function wputmodrm(m, s, rm, vs, vrm)
+  assert(m < 4 and s < 16 and rm < 16, "bad modrm operands")
+  wputb(shl(m, 6) + shl(band(s, 7), 3) + band(rm, 7))
+end
+
+-- Put ModRM/SIB plus optional displacement.
+local function wputmrmsib(t, imark, s, vsreg, psz, sk)
+  local vreg, vxreg
+  local reg, xreg = t.reg, t.xreg
+  if reg and reg < 0 then reg = 0; vreg = t.vreg end
+  if xreg and xreg < 0 then xreg = 0; vxreg = t.vxreg end
+  if s < 0 then s = 0 end
+
+  -- Register mode.
+  if sub(t.mode, 1, 1) == "r" then
+    wputmodrm(3, s, reg)
+    wvreg("modrm.reg", vsreg, psz+1, sk, vreg)
+    wvreg("modrm.rm.r", vreg, psz+1, sk)
+    return
+  end
+
+  local disp = t.disp
+  local tdisp = type(disp)
+  -- No base register?
+  if not reg then
+    local riprel = false
+    if xreg then
+      -- Indexed mode with index register only.
+      -- [xreg*xsc+disp] -> (0, s, esp) (xsc, xreg, ebp)
+      wputmodrm(0, s, 4)
+      if imark == "I" then waction("MARK") end
+      wvreg("modrm.reg", vsreg, psz+1, sk, vxreg)
+      wputmodrm(t.xsc, xreg, 5)
+      wvreg("sib.index", vxreg, psz+2, sk)
+    else
+      -- Pure 32 bit displacement.
+      if x64 and tdisp ~= "table" then
+	wputmodrm(0, s, 4) -- [disp] -> (0, s, esp) (0, esp, ebp)
+	wvreg("modrm.reg", vsreg, psz+1, sk)
+	if imark == "I" then waction("MARK") end
+	wputmodrm(0, 4, 5)
+      else
+	riprel = x64
+	wputmodrm(0, s, 5) -- [disp|rip-label] -> (0, s, ebp)
+	wvreg("modrm.reg", vsreg, psz+1, sk)
+	if imark == "I" then waction("MARK") end
+      end
+    end
+    if riprel then -- Emit rip-relative displacement.
+      if match("UWSiI", imark) then
+	werror("NYI: rip-relative displacement followed by immediate")
+      end
+      -- The previous byte in the action buffer cannot be 0xe9 or 0x80-0x8f.
+      wputlabel("REL_", disp[1], 2)
+    else
+      wputdarg(disp)
+    end
+    return
+  end
+
+  local m
+  if tdisp == "number" then -- Check displacement size at assembly time.
+    if disp == 0 and band(reg, 7) ~= 5 then -- [ebp] -> [ebp+0] (in SIB, too)
+      if not vreg then m = 0 end -- Force DISP to allow [Rd(5)] -> [ebp+0]
+    elseif disp >= -128 and disp <= 127 then m = 1
+    else m = 2 end
+  elseif tdisp == "table" then
+    m = 2
+  end
+
+  -- Index register present or esp as base register: need SIB encoding.
+  if xreg or band(reg, 7) == 4 then
+    wputmodrm(m or 2, s, 4) -- ModRM.
+    if m == nil or imark == "I" then waction("MARK") end
+    wvreg("modrm.reg", vsreg, psz+1, sk, vxreg or vreg)
+    wputmodrm(t.xsc or 0, xreg or 4, reg) -- SIB.
+    wvreg("sib.index", vxreg, psz+2, sk, vreg)
+    wvreg("sib.base", vreg, psz+2, sk)
+  else
+    wputmodrm(m or 2, s, reg) -- ModRM.
+    if (imark == "I" and (m == 1 or m == 2)) or
+       (m == nil and (vsreg or vreg)) then waction("MARK") end
+    wvreg("modrm.reg", vsreg, psz+1, sk, vreg)
+    wvreg("modrm.rm.m", vreg, psz+1, sk)
+  end
+
+  -- Put displacement.
+  if m == 1 then wputsbarg(disp)
+  elseif m == 2 then wputdarg(disp)
+  elseif m == nil then waction("DISP", disp) end
+end
+
+------------------------------------------------------------------------------
+
+-- Return human-readable operand mode string.
+local function opmodestr(op, args)
+  local m = {}
+  for i=1,#args do
+    local a = args[i]
+    m[#m+1] = sub(a.mode, 1, 1)..(a.opsize or "?")
+  end
+  return op.." "..concat(m, ",")
+end
+
+-- Convert number to valid integer or nil.
+local function toint(expr, isqword)
+  local n = tonumber(expr)
+  if n then
+    if n % 1 ~= 0 then
+      werror("not an integer number `"..expr.."'")
+    elseif isqword then
+      if n < -2147483648 or n > 2147483647 then
+	n = nil -- Handle it as an expression to avoid precision loss.
+      end
+    elseif n < -2147483648 or n > 4294967295 then
+      werror("bad integer number `"..expr.."'")
+    end
+    return n
+  end
+end
+
+-- Parse immediate expression.
+local function immexpr(expr)
+  -- &expr (pointer)
+  if sub(expr, 1, 1) == "&" then
+    return "iPJ", format("(ptrdiff_t)(%s)", sub(expr,2))
+  end
+
+  local prefix = sub(expr, 1, 2)
+  -- =>expr (pc label reference)
+  if prefix == "=>" then
+    return "iJ", sub(expr, 3)
+  end
+  -- ->name (global label reference)
+  if prefix == "->" then
+    return "iJ", map_global[sub(expr, 3)]
+  end
+
+  -- [<>][1-9] (local label reference)
+  local dir, lnum = match(expr, "^([<>])([1-9])$")
+  if dir then -- Fwd: 247-255, Bkwd: 1-9.
+    return "iJ", lnum + (dir == ">" and 246 or 0)
+  end
+
+  local extname = match(expr, "^extern%s+(%S+)$")
+  if extname then
+    return "iJ", map_extern[extname]
+  end
+
+  -- expr (interpreted as immediate)
+  return "iI", expr
+end
+
+-- Parse displacement expression: +-num, +-expr, +-opsize*num
+local function dispexpr(expr)
+  local disp = expr == "" and 0 or toint(expr)
+  if disp then return disp end
+  local c, dispt = match(expr, "^([+-])%s*(.+)$")
+  if c == "+" then
+    expr = dispt
+  elseif not c then
+    werror("bad displacement expression `"..expr.."'")
+  end
+  local opsize, tailops = match(dispt, "^(%w+)%s*%*%s*(.+)$")
+  local ops, imm = map_opsize[opsize], toint(tailops)
+  if ops and imm then
+    if c == "-" then imm = -imm end
+    return imm*map_opsizenum[ops]
+  end
+  local mode, iexpr = immexpr(dispt)
+  if mode == "iJ" then
+    if c == "-" then werror("cannot invert label reference") end
+    return { iexpr }
+  end
+  return expr -- Need to return original signed expression.
+end
+
+-- Parse register or type expression.
+local function rtexpr(expr)
+  if not expr then return end
+  local tname, ovreg = match(expr, "^([%w_]+):(@[%w_]+)$")
+  local tp = map_type[tname or expr]
+  if tp then
+    local reg = ovreg or tp.reg
+    local rnum = map_reg_num[reg]
+    if not rnum then
+      werror("type `"..(tname or expr).."' needs a register override")
+    end
+    if not map_reg_valid_base[reg] then
+      werror("bad base register override `"..(map_reg_rev[reg] or reg).."'")
+    end
+    return reg, rnum, tp
+  end
+  return expr, map_reg_num[expr]
+end
+
+-- Parse operand and return { mode, opsize, reg, xreg, xsc, disp, imm }.
+local function parseoperand(param, isqword)
+  local t = {}
+
+  local expr = param
+  local opsize, tailops = match(param, "^(%w+)%s*(.+)$")
+  if opsize then
+    t.opsize = map_opsize[opsize]
+    if t.opsize then expr = tailops end
+  end
+
+  local br = match(expr, "^%[%s*(.-)%s*%]$")
+  repeat
+    if br then
+      t.mode = "xm"
+
+      -- [disp]
+      t.disp = toint(br)
+      if t.disp then
+	t.mode = x64 and "xm" or "xmO"
+	break
+      end
+
+      -- [reg...]
+      local tp
+      local reg, tailr = match(br, "^([@%w_:]+)%s*(.*)$")
+      reg, t.reg, tp = rtexpr(reg)
+      if not t.reg then
+	-- [expr]
+	t.mode = x64 and "xm" or "xmO"
+	t.disp = dispexpr("+"..br)
+	break
+      end
+
+      if t.reg == -1 then
+	t.vreg, tailr = match(tailr, "^(%b())(.*)$")
+	if not t.vreg then werror("bad variable register expression") end
+      end
+
+      -- [xreg*xsc] or [xreg*xsc+-disp] or [xreg*xsc+-expr]
+      local xsc, tailsc = match(tailr, "^%*%s*([1248])%s*(.*)$")
+      if xsc then
+	if not map_reg_valid_index[reg] then
+	  werror("bad index register `"..map_reg_rev[reg].."'")
+	end
+	t.xsc = map_xsc[xsc]
+	t.xreg = t.reg
+	t.vxreg = t.vreg
+	t.reg = nil
+	t.vreg = nil
+	t.disp = dispexpr(tailsc)
+	break
+      end
+      if not map_reg_valid_base[reg] then
+	werror("bad base register `"..map_reg_rev[reg].."'")
+      end
+
+      -- [reg] or [reg+-disp]
+      t.disp = toint(tailr) or (tailr == "" and 0)
+      if t.disp then break end
+
+      -- [reg+xreg...]
+      local xreg, tailx = match(tailr, "^%+%s*([@%w_:]+)%s*(.*)$")
+      xreg, t.xreg, tp = rtexpr(xreg)
+      if not t.xreg then
+	-- [reg+-expr]
+	t.disp = dispexpr(tailr)
+	break
+      end
+      if not map_reg_valid_index[xreg] then
+	werror("bad index register `"..map_reg_rev[xreg].."'")
+      end
+
+      if t.xreg == -1 then
+	t.vxreg, tailx = match(tailx, "^(%b())(.*)$")
+	if not t.vxreg then werror("bad variable register expression") end
+      end
+
+      -- [reg+xreg*xsc...]
+      local xsc, tailsc = match(tailx, "^%*%s*([1248])%s*(.*)$")
+      if xsc then
+	t.xsc = map_xsc[xsc]
+	tailx = tailsc
+      end
+
+      -- [...] or [...+-disp] or [...+-expr]
+      t.disp = dispexpr(tailx)
+    else
+      -- imm or opsize*imm
+      local imm = toint(expr, isqword)
+      if not imm and sub(expr, 1, 1) == "*" and t.opsize then
+	imm = toint(sub(expr, 2))
+	if imm then
+	  imm = imm * map_opsizenum[t.opsize]
+	  t.opsize = nil
+	end
+      end
+      if imm then
+	if t.opsize then werror("bad operand size override") end
+	local m = "i"
+	if imm == 1 then m = m.."1" end
+	if imm >= 4294967168 and imm <= 4294967295 then imm = imm-4294967296 end
+	if imm >= -128 and imm <= 127 then m = m.."S" end
+	t.imm = imm
+	t.mode = m
+	break
+      end
+
+      local tp
+      local reg, tailr = match(expr, "^([@%w_:]+)%s*(.*)$")
+      reg, t.reg, tp = rtexpr(reg)
+      if t.reg then
+	if t.reg == -1 then
+	  t.vreg, tailr = match(tailr, "^(%b())(.*)$")
+	  if not t.vreg then werror("bad variable register expression") end
+	end
+	-- reg
+	if tailr == "" then
+	  if t.opsize then werror("bad operand size override") end
+	  t.opsize = map_reg_opsize[reg]
+	  if t.opsize == "f" then
+	    t.mode = t.reg == 0 and "fF" or "f"
+	  else
+	    if reg == "@w4" or (x64 and reg == "@d4") then
+	      wwarn("bad idea, try again with `"..(x64 and "rsp'" or "esp'"))
+	    end
+	    t.mode = t.reg == 0 and "rmR" or (reg == "@b1" and "rmC" or "rm")
+	  end
+	  t.needrex = map_reg_needrex[reg]
+	  break
+	end
+
+	-- type[idx], type[idx].field, type->field -> [reg+offset_expr]
+	if not tp then werror("bad operand `"..param.."'") end
+	t.mode = "xm"
+	t.disp = format(tp.ctypefmt, tailr)
+      else
+	t.mode, t.imm = immexpr(expr)
+	if sub(t.mode, -1) == "J" then
+	  if t.opsize and t.opsize ~= addrsize then
+	    werror("bad operand size override")
+	  end
+	  t.opsize = addrsize
+	end
+      end
+    end
+  until true
+  return t
+end
+
+------------------------------------------------------------------------------
+-- x86 Template String Description
+-- ===============================
+--
+-- Each template string is a list of [match:]pattern pairs,
+-- separated by "|". The first match wins. No match means a
+-- bad or unsupported combination of operand modes or sizes.
+--
+-- The match part and the ":" is omitted if the operation has
+-- no operands. Otherwise the first N characters are matched
+-- against the mode strings of each of the N operands.
+--
+-- The mode string for each operand type is (see parseoperand()):
+--   Integer register: "rm", +"R" for eax, ax, al, +"C" for cl
+--   FP register:      "f",  +"F" for st0
+--   Index operand:    "xm", +"O" for [disp] (pure offset)
+--   Immediate:        "i",  +"S" for signed 8 bit, +"1" for 1,
+--                     +"I" for arg, +"P" for pointer
+--   Any:              +"J" for valid jump targets
+--
+-- So a match character "m" (mixed) matches both an integer register
+-- and an index operand (to be encoded with the ModRM/SIB scheme).
+-- But "r" matches only a register and "x" only an index operand
+-- (e.g. for FP memory access operations).
+--
+-- The operand size match string starts right after the mode match
+-- characters and ends before the ":". "dwb" or "qdwb" is assumed, if empty.
+-- The effective data size of the operation is matched against this list.
+--
+-- If only the regular "b", "w", "d", "q", "t" operand sizes are
+-- present, then all operands must be the same size. Unspecified sizes
+-- are ignored, but at least one operand must have a size or the pattern
+-- won't match (use the "byte", "word", "dword", "qword", "tword"
+-- operand size overrides. E.g.: mov dword [eax], 1).
+--
+-- If the list has a "1" or "2" prefix, the operand size is taken
+-- from the respective operand and any other operand sizes are ignored.
+-- If the list contains only ".", all operand sizes are ignored.
+-- If the list has a "/" prefix, the concatenated (mixed) operand sizes
+-- are compared to the match.
+--
+-- E.g. "rrdw" matches for either two dword registers or two word
+-- registers. "Fx2dq" matches an st0 operand plus an index operand
+-- pointing to a dword (float) or qword (double).
+--
+-- Every character after the ":" is part of the pattern string:
+--   Hex chars are accumulated to form the opcode (left to right).
+--   "n"       disables the standard opcode mods
+--             (otherwise: -1 for "b", o16 prefix for "w", rex.w for "q")
+--   "X"       Force REX.W.
+--   "r"/"R"   adds the reg. number from the 1st/2nd operand to the opcode.
+--   "m"/"M"   generates ModRM/SIB from the 1st/2nd operand.
+--             The spare 3 bits are either filled with the last hex digit or
+--             the result from a previous "r"/"R". The opcode is restored.
+--   "u"       Use VEX encoding, vvvv unused.
+--   "v"/"V"   Use VEX encoding, vvvv from 1st/2nd operand (the operand is
+--             removed from the list used by future characters).
+--   "w"       Use VEX encoding, vvvv from 3rd operand.
+--   "L"       Force VEX.L
+--
+-- All of the following characters force a flush of the opcode:
+--   "o"/"O"   stores a pure 32 bit disp (offset) from the 1st/2nd operand.
+--   "s"       stores a 4 bit immediate from the last register operand,
+--             followed by 4 zero bits.
+--   "S"       stores a signed 8 bit immediate from the last operand.
+--   "U"       stores an unsigned 8 bit immediate from the last operand.
+--   "W"       stores an unsigned 16 bit immediate from the last operand.
+--   "i"       stores an operand sized immediate from the last operand.
+--   "I"       dito, but generates an action code to optionally modify
+--             the opcode (+2) for a signed 8 bit immediate.
+--   "J"       generates one of the REL action codes from the last operand.
+--
+------------------------------------------------------------------------------
+
+-- Template strings for x86 instructions. Ordered by first opcode byte.
+-- Unimplemented opcodes (deliberate omissions) are marked with *.
+local map_op = {
+  -- 00-05: add...
+  -- 06: *push es
+  -- 07: *pop es
+  -- 08-0D: or...
+  -- 0E: *push cs
+  -- 0F: two byte opcode prefix
+  -- 10-15: adc...
+  -- 16: *push ss
+  -- 17: *pop ss
+  -- 18-1D: sbb...
+  -- 1E: *push ds
+  -- 1F: *pop ds
+  -- 20-25: and...
+  es_0 =	"26",
+  -- 27: *daa
+  -- 28-2D: sub...
+  cs_0 =	"2E",
+  -- 2F: *das
+  -- 30-35: xor...
+  ss_0 =	"36",
+  -- 37: *aaa
+  -- 38-3D: cmp...
+  ds_0 =	"3E",
+  -- 3F: *aas
+  inc_1 =	x64 and "m:FF0m" or "rdw:40r|m:FF0m",
+  dec_1 =	x64 and "m:FF1m" or "rdw:48r|m:FF1m",
+  push_1 =	(x64 and "rq:n50r|rw:50r|mq:nFF6m|mw:FF6m" or
+			 "rdw:50r|mdw:FF6m").."|S.:6AS|ib:n6Ai|i.:68i",
+  pop_1 =	x64 and "rq:n58r|rw:58r|mq:n8F0m|mw:8F0m" or "rdw:58r|mdw:8F0m",
+  -- 60: *pusha, *pushad, *pushaw
+  -- 61: *popa, *popad, *popaw
+  -- 62: *bound rdw,x
+  -- 63: x86: *arpl mw,rw
+  movsxd_2 =	x64 and "rm/qd:63rM",
+  fs_0 =	"64",
+  gs_0 =	"65",
+  o16_0 =	"66",
+  a16_0 =	not x64 and "67" or nil,
+  a32_0 =	x64 and "67",
+  -- 68: push idw
+  -- 69: imul rdw,mdw,idw
+  -- 6A: push ib
+  -- 6B: imul rdw,mdw,S
+  -- 6C: *insb
+  -- 6D: *insd, *insw
+  -- 6E: *outsb
+  -- 6F: *outsd, *outsw
+  -- 70-7F: jcc lb
+  -- 80: add... mb,i
+  -- 81: add... mdw,i
+  -- 82: *undefined
+  -- 83: add... mdw,S
+  test_2 =	"mr:85Rm|rm:85rM|Ri:A9ri|mi:F70mi",
+  -- 86: xchg rb,mb
+  -- 87: xchg rdw,mdw
+  -- 88: mov mb,r
+  -- 89: mov mdw,r
+  -- 8A: mov r,mb
+  -- 8B: mov r,mdw
+  -- 8C: *mov mdw,seg
+  lea_2 =	"rx1dq:8DrM",
+  -- 8E: *mov seg,mdw
+  -- 8F: pop mdw
+  nop_0 =	"90",
+  xchg_2 =	"Rrqdw:90R|rRqdw:90r|rm:87rM|mr:87Rm",
+  cbw_0 =	"6698",
+  cwde_0 =	"98",
+  cdqe_0 =	"4898",
+  cwd_0 =	"6699",
+  cdq_0 =	"99",
+  cqo_0 =	"4899",
+  -- 9A: *call iw:idw
+  wait_0 =	"9B",
+  fwait_0 =	"9B",
+  pushf_0 =	"9C",
+  pushfd_0 =	not x64 and "9C",
+  pushfq_0 =	x64 and "9C",
+  popf_0 =	"9D",
+  popfd_0 =	not x64 and "9D",
+  popfq_0 =	x64 and "9D",
+  sahf_0 =	"9E",
+  lahf_0 =	"9F",
+  mov_2 =	"OR:A3o|RO:A1O|mr:89Rm|rm:8BrM|rib:nB0ri|ridw:B8ri|mi:C70mi",
+  movsb_0 =	"A4",
+  movsw_0 =	"66A5",
+  movsd_0 =	"A5",
+  cmpsb_0 =	"A6",
+  cmpsw_0 =	"66A7",
+  cmpsd_0 =	"A7",
+  -- A8: test Rb,i
+  -- A9: test Rdw,i
+  stosb_0 =	"AA",
+  stosw_0 =	"66AB",
+  stosd_0 =	"AB",
+  lodsb_0 =	"AC",
+  lodsw_0 =	"66AD",
+  lodsd_0 =	"AD",
+  scasb_0 =	"AE",
+  scasw_0 =	"66AF",
+  scasd_0 =	"AF",
+  -- B0-B7: mov rb,i
+  -- B8-BF: mov rdw,i
+  -- C0: rol... mb,i
+  -- C1: rol... mdw,i
+  ret_1 =	"i.:nC2W",
+  ret_0 =	"C3",
+  -- C4: *les rdw,mq
+  -- C5: *lds rdw,mq
+  -- C6: mov mb,i
+  -- C7: mov mdw,i
+  -- C8: *enter iw,ib
+  leave_0 =	"C9",
+  -- CA: *retf iw
+  -- CB: *retf
+  int3_0 =	"CC",
+  int_1 =	"i.:nCDU",
+  into_0 =	"CE",
+  -- CF: *iret
+  -- D0: rol... mb,1
+  -- D1: rol... mdw,1
+  -- D2: rol... mb,cl
+  -- D3: rol... mb,cl
+  -- D4: *aam ib
+  -- D5: *aad ib
+  -- D6: *salc
+  -- D7: *xlat
+  -- D8-DF: floating point ops
+  -- E0: *loopne
+  -- E1: *loope
+  -- E2: *loop
+  -- E3: *jcxz, *jecxz
+  -- E4: *in Rb,ib
+  -- E5: *in Rdw,ib
+  -- E6: *out ib,Rb
+  -- E7: *out ib,Rdw
+  call_1 =	x64 and "mq:nFF2m|J.:E8nJ" or "md:FF2m|J.:E8J",
+  jmp_1 =	x64 and "mq:nFF4m|J.:E9nJ" or "md:FF4m|J.:E9J", -- short: EB
+  -- EA: *jmp iw:idw
+  -- EB: jmp ib
+  -- EC: *in Rb,dx
+  -- ED: *in Rdw,dx
+  -- EE: *out dx,Rb
+  -- EF: *out dx,Rdw
+  lock_0 =	"F0",
+  int1_0 =	"F1",
+  repne_0 =	"F2",
+  repnz_0 =	"F2",
+  rep_0 =	"F3",
+  repe_0 =	"F3",
+  repz_0 =	"F3",
+  -- F4: *hlt
+  cmc_0 =	"F5",
+  -- F6: test... mb,i; div... mb
+  -- F7: test... mdw,i; div... mdw
+  clc_0 =	"F8",
+  stc_0 =	"F9",
+  -- FA: *cli
+  cld_0 =	"FC",
+  std_0 =	"FD",
+  -- FE: inc... mb
+  -- FF: inc... mdw
+
+  -- misc ops
+  not_1 =	"m:F72m",
+  neg_1 =	"m:F73m",
+  mul_1 =	"m:F74m",
+  imul_1 =	"m:F75m",
+  div_1 =	"m:F76m",
+  idiv_1 =	"m:F77m",
+
+  imul_2 =	"rmqdw:0FAFrM|rIqdw:69rmI|rSqdw:6BrmS|riqdw:69rmi",
+  imul_3 =	"rmIqdw:69rMI|rmSqdw:6BrMS|rmiqdw:69rMi",
+
+  movzx_2 =	"rm/db:0FB6rM|rm/qb:|rm/wb:0FB6rM|rm/dw:0FB7rM|rm/qw:",
+  movsx_2 =	"rm/db:0FBErM|rm/qb:|rm/wb:0FBErM|rm/dw:0FBFrM|rm/qw:",
+
+  bswap_1 =	"rqd:0FC8r",
+  bsf_2 =	"rmqdw:0FBCrM",
+  bsr_2 =	"rmqdw:0FBDrM",
+  bt_2 =	"mrqdw:0FA3Rm|miqdw:0FBA4mU",
+  btc_2 =	"mrqdw:0FBBRm|miqdw:0FBA7mU",
+  btr_2 =	"mrqdw:0FB3Rm|miqdw:0FBA6mU",
+  bts_2 =	"mrqdw:0FABRm|miqdw:0FBA5mU",
+
+  shld_3 =	"mriqdw:0FA4RmU|mrC/qq:0FA5Rm|mrC/dd:|mrC/ww:",
+  shrd_3 =	"mriqdw:0FACRmU|mrC/qq:0FADRm|mrC/dd:|mrC/ww:",
+
+  rdtsc_0 =	"0F31", -- P1+
+  rdpmc_0 =	"0F33", -- P6+
+  cpuid_0 =	"0FA2", -- P1+
+
+  -- floating point ops
+  fst_1 =	"ff:DDD0r|xd:D92m|xq:nDD2m",
+  fstp_1 =	"ff:DDD8r|xd:D93m|xq:nDD3m|xt:DB7m",
+  fld_1 =	"ff:D9C0r|xd:D90m|xq:nDD0m|xt:DB5m",
+
+  fpop_0 =	"DDD8", -- Alias for fstp st0.
+
+  fist_1 =	"xw:nDF2m|xd:DB2m",
+  fistp_1 =	"xw:nDF3m|xd:DB3m|xq:nDF7m",
+  fild_1 =	"xw:nDF0m|xd:DB0m|xq:nDF5m",
+
+  fxch_0 =	"D9C9",
+  fxch_1 =	"ff:D9C8r",
+  fxch_2 =	"fFf:D9C8r|Fff:D9C8R",
+
+  fucom_1 =	"ff:DDE0r",
+  fucom_2 =	"Fff:DDE0R",
+  fucomp_1 =	"ff:DDE8r",
+  fucomp_2 =	"Fff:DDE8R",
+  fucomi_1 =	"ff:DBE8r", -- P6+
+  fucomi_2 =	"Fff:DBE8R", -- P6+
+  fucomip_1 =	"ff:DFE8r", -- P6+
+  fucomip_2 =	"Fff:DFE8R", -- P6+
+  fcomi_1 =	"ff:DBF0r", -- P6+
+  fcomi_2 =	"Fff:DBF0R", -- P6+
+  fcomip_1 =	"ff:DFF0r", -- P6+
+  fcomip_2 =	"Fff:DFF0R", -- P6+
+  fucompp_0 =	"DAE9",
+  fcompp_0 =	"DED9",
+
+  fldenv_1 =	"x.:D94m",
+  fnstenv_1 =	"x.:D96m",
+  fstenv_1 =	"x.:9BD96m",
+  fldcw_1 =	"xw:nD95m",
+  fstcw_1 =	"xw:n9BD97m",
+  fnstcw_1 =	"xw:nD97m",
+  fstsw_1 =	"Rw:n9BDFE0|xw:n9BDD7m",
+  fnstsw_1 =	"Rw:nDFE0|xw:nDD7m",
+  fclex_0 =	"9BDBE2",
+  fnclex_0 =	"DBE2",
+
+  fnop_0 =	"D9D0",
+  -- D9D1-D9DF: unassigned
+
+  fchs_0 =	"D9E0",
+  fabs_0 =	"D9E1",
+  -- D9E2: unassigned
+  -- D9E3: unassigned
+  ftst_0 =	"D9E4",
+  fxam_0 =	"D9E5",
+  -- D9E6: unassigned
+  -- D9E7: unassigned
+  fld1_0 =	"D9E8",
+  fldl2t_0 =	"D9E9",
+  fldl2e_0 =	"D9EA",
+  fldpi_0 =	"D9EB",
+  fldlg2_0 =	"D9EC",
+  fldln2_0 =	"D9ED",
+  fldz_0 =	"D9EE",
+  -- D9EF: unassigned
+
+  f2xm1_0 =	"D9F0",
+  fyl2x_0 =	"D9F1",
+  fptan_0 =	"D9F2",
+  fpatan_0 =	"D9F3",
+  fxtract_0 =	"D9F4",
+  fprem1_0 =	"D9F5",
+  fdecstp_0 =	"D9F6",
+  fincstp_0 =	"D9F7",
+  fprem_0 =	"D9F8",
+  fyl2xp1_0 =	"D9F9",
+  fsqrt_0 =	"D9FA",
+  fsincos_0 =	"D9FB",
+  frndint_0 =	"D9FC",
+  fscale_0 =	"D9FD",
+  fsin_0 =	"D9FE",
+  fcos_0 =	"D9FF",
+
+  -- SSE, SSE2
+  andnpd_2 =	"rmo:660F55rM",
+  andnps_2 =	"rmo:0F55rM",
+  andpd_2 =	"rmo:660F54rM",
+  andps_2 =	"rmo:0F54rM",
+  clflush_1 =	"x.:0FAE7m",
+  cmppd_3 =	"rmio:660FC2rMU",
+  cmpps_3 =	"rmio:0FC2rMU",
+  cmpsd_3 =	"rrio:F20FC2rMU|rxi/oq:",
+  cmpss_3 =	"rrio:F30FC2rMU|rxi/od:",
+  comisd_2 =	"rro:660F2FrM|rx/oq:",
+  comiss_2 =	"rro:0F2FrM|rx/od:",
+  cvtdq2pd_2 =	"rro:F30FE6rM|rx/oq:",
+  cvtdq2ps_2 =	"rmo:0F5BrM",
+  cvtpd2dq_2 =	"rmo:F20FE6rM",
+  cvtpd2ps_2 =	"rmo:660F5ArM",
+  cvtpi2pd_2 =	"rx/oq:660F2ArM",
+  cvtpi2ps_2 =	"rx/oq:0F2ArM",
+  cvtps2dq_2 =	"rmo:660F5BrM",
+  cvtps2pd_2 =	"rro:0F5ArM|rx/oq:",
+  cvtsd2si_2 =	"rr/do:F20F2DrM|rr/qo:|rx/dq:|rxq:",
+  cvtsd2ss_2 =	"rro:F20F5ArM|rx/oq:",
+  cvtsi2sd_2 =	"rm/od:F20F2ArM|rm/oq:F20F2ArXM",
+  cvtsi2ss_2 =	"rm/od:F30F2ArM|rm/oq:F30F2ArXM",
+  cvtss2sd_2 =	"rro:F30F5ArM|rx/od:",
+  cvtss2si_2 =	"rr/do:F30F2DrM|rr/qo:|rxd:|rx/qd:",
+  cvttpd2dq_2 =	"rmo:660FE6rM",
+  cvttps2dq_2 =	"rmo:F30F5BrM",
+  cvttsd2si_2 =	"rr/do:F20F2CrM|rr/qo:|rx/dq:|rxq:",
+  cvttss2si_2 =	"rr/do:F30F2CrM|rr/qo:|rxd:|rx/qd:",
+  fxsave_1 =	"x.:0FAE0m",
+  fxrstor_1 =	"x.:0FAE1m",
+  ldmxcsr_1 =	"xd:0FAE2m",
+  lfence_0 =	"0FAEE8",
+  maskmovdqu_2 = "rro:660FF7rM",
+  mfence_0 =	"0FAEF0",
+  movapd_2 =	"rmo:660F28rM|mro:660F29Rm",
+  movaps_2 =	"rmo:0F28rM|mro:0F29Rm",
+  movd_2 =	"rm/od:660F6ErM|rm/oq:660F6ErXM|mr/do:660F7ERm|mr/qo:",
+  movdqa_2 =	"rmo:660F6FrM|mro:660F7FRm",
+  movdqu_2 =	"rmo:F30F6FrM|mro:F30F7FRm",
+  movhlps_2 =	"rro:0F12rM",
+  movhpd_2 =	"rx/oq:660F16rM|xr/qo:n660F17Rm",
+  movhps_2 =	"rx/oq:0F16rM|xr/qo:n0F17Rm",
+  movlhps_2 =	"rro:0F16rM",
+  movlpd_2 =	"rx/oq:660F12rM|xr/qo:n660F13Rm",
+  movlps_2 =	"rx/oq:0F12rM|xr/qo:n0F13Rm",
+  movmskpd_2 =	"rr/do:660F50rM",
+  movmskps_2 =	"rr/do:0F50rM",
+  movntdq_2 =	"xro:660FE7Rm",
+  movnti_2 =	"xrqd:0FC3Rm",
+  movntpd_2 =	"xro:660F2BRm",
+  movntps_2 =	"xro:0F2BRm",
+  movq_2 =	"rro:F30F7ErM|rx/oq:|xr/qo:n660FD6Rm",
+  movsd_2 =	"rro:F20F10rM|rx/oq:|xr/qo:nF20F11Rm",
+  movss_2 =	"rro:F30F10rM|rx/od:|xr/do:F30F11Rm",
+  movupd_2 =	"rmo:660F10rM|mro:660F11Rm",
+  movups_2 =	"rmo:0F10rM|mro:0F11Rm",
+  orpd_2 =	"rmo:660F56rM",
+  orps_2 =	"rmo:0F56rM",
+  pause_0 =	"F390",
+  pextrw_3 =	"rri/do:660FC5rMU|xri/wo:660F3A15nRmU", -- Mem op: SSE4.1 only.
+  pinsrw_3 =	"rri/od:660FC4rMU|rxi/ow:",
+  pmovmskb_2 =	"rr/do:660FD7rM",
+  prefetchnta_1 = "xb:n0F180m",
+  prefetcht0_1 = "xb:n0F181m",
+  prefetcht1_1 = "xb:n0F182m",
+  prefetcht2_1 = "xb:n0F183m",
+  pshufd_3 =	"rmio:660F70rMU",
+  pshufhw_3 =	"rmio:F30F70rMU",
+  pshuflw_3 =	"rmio:F20F70rMU",
+  pslld_2 =	"rmo:660FF2rM|rio:660F726mU",
+  pslldq_2 =	"rio:660F737mU",
+  psllq_2 =	"rmo:660FF3rM|rio:660F736mU",
+  psllw_2 =	"rmo:660FF1rM|rio:660F716mU",
+  psrad_2 =	"rmo:660FE2rM|rio:660F724mU",
+  psraw_2 =	"rmo:660FE1rM|rio:660F714mU",
+  psrld_2 =	"rmo:660FD2rM|rio:660F722mU",
+  psrldq_2 =	"rio:660F733mU",
+  psrlq_2 =	"rmo:660FD3rM|rio:660F732mU",
+  psrlw_2 =	"rmo:660FD1rM|rio:660F712mU",
+  rcpps_2 =	"rmo:0F53rM",
+  rcpss_2 =	"rro:F30F53rM|rx/od:",
+  rsqrtps_2 =	"rmo:0F52rM",
+  rsqrtss_2 =	"rmo:F30F52rM",
+  sfence_0 =	"0FAEF8",
+  shufpd_3 =	"rmio:660FC6rMU",
+  shufps_3 =	"rmio:0FC6rMU",
+  stmxcsr_1 =   "xd:0FAE3m",
+  ucomisd_2 =	"rro:660F2ErM|rx/oq:",
+  ucomiss_2 =	"rro:0F2ErM|rx/od:",
+  unpckhpd_2 =	"rmo:660F15rM",
+  unpckhps_2 =	"rmo:0F15rM",
+  unpcklpd_2 =	"rmo:660F14rM",
+  unpcklps_2 =	"rmo:0F14rM",
+  xorpd_2 =	"rmo:660F57rM",
+  xorps_2 =	"rmo:0F57rM",
+
+  -- SSE3 ops
+  fisttp_1 =	"xw:nDF1m|xd:DB1m|xq:nDD1m",
+  addsubpd_2 =	"rmo:660FD0rM",
+  addsubps_2 =	"rmo:F20FD0rM",
+  haddpd_2 =	"rmo:660F7CrM",
+  haddps_2 =	"rmo:F20F7CrM",
+  hsubpd_2 =	"rmo:660F7DrM",
+  hsubps_2 =	"rmo:F20F7DrM",
+  lddqu_2 =	"rxo:F20FF0rM",
+  movddup_2 =	"rmo:F20F12rM",
+  movshdup_2 =	"rmo:F30F16rM",
+  movsldup_2 =	"rmo:F30F12rM",
+
+  -- SSSE3 ops
+  pabsb_2 =	"rmo:660F381CrM",
+  pabsd_2 =	"rmo:660F381ErM",
+  pabsw_2 =	"rmo:660F381DrM",
+  palignr_3 =	"rmio:660F3A0FrMU",
+  phaddd_2 =	"rmo:660F3802rM",
+  phaddsw_2 =	"rmo:660F3803rM",
+  phaddw_2 =	"rmo:660F3801rM",
+  phsubd_2 =	"rmo:660F3806rM",
+  phsubsw_2 =	"rmo:660F3807rM",
+  phsubw_2 =	"rmo:660F3805rM",
+  pmaddubsw_2 =	"rmo:660F3804rM",
+  pmulhrsw_2 =	"rmo:660F380BrM",
+  pshufb_2 =	"rmo:660F3800rM",
+  psignb_2 =	"rmo:660F3808rM",
+  psignd_2 =	"rmo:660F380ArM",
+  psignw_2 =	"rmo:660F3809rM",
+
+  -- SSE4.1 ops
+  blendpd_3 =	"rmio:660F3A0DrMU",
+  blendps_3 =	"rmio:660F3A0CrMU",
+  blendvpd_3 =	"rmRo:660F3815rM",
+  blendvps_3 =	"rmRo:660F3814rM",
+  dppd_3 =	"rmio:660F3A41rMU",
+  dpps_3 =	"rmio:660F3A40rMU",
+  extractps_3 =	"mri/do:660F3A17RmU|rri/qo:660F3A17RXmU",
+  insertps_3 =	"rrio:660F3A41rMU|rxi/od:",
+  movntdqa_2 =	"rxo:660F382ArM",
+  mpsadbw_3 =	"rmio:660F3A42rMU",
+  packusdw_2 =	"rmo:660F382BrM",
+  pblendvb_3 =	"rmRo:660F3810rM",
+  pblendw_3 =	"rmio:660F3A0ErMU",
+  pcmpeqq_2 =	"rmo:660F3829rM",
+  pextrb_3 =	"rri/do:660F3A14nRmU|rri/qo:|xri/bo:",
+  pextrd_3 =	"mri/do:660F3A16RmU",
+  pextrq_3 =	"mri/qo:660F3A16RmU",
+  -- pextrw is SSE2, mem operand is SSE4.1 only
+  phminposuw_2 = "rmo:660F3841rM",
+  pinsrb_3 =	"rri/od:660F3A20nrMU|rxi/ob:",
+  pinsrd_3 =	"rmi/od:660F3A22rMU",
+  pinsrq_3 =	"rmi/oq:660F3A22rXMU",
+  pmaxsb_2 =	"rmo:660F383CrM",
+  pmaxsd_2 =	"rmo:660F383DrM",
+  pmaxud_2 =	"rmo:660F383FrM",
+  pmaxuw_2 =	"rmo:660F383ErM",
+  pminsb_2 =	"rmo:660F3838rM",
+  pminsd_2 =	"rmo:660F3839rM",
+  pminud_2 =	"rmo:660F383BrM",
+  pminuw_2 =	"rmo:660F383ArM",
+  pmovsxbd_2 =	"rro:660F3821rM|rx/od:",
+  pmovsxbq_2 =	"rro:660F3822rM|rx/ow:",
+  pmovsxbw_2 =	"rro:660F3820rM|rx/oq:",
+  pmovsxdq_2 =	"rro:660F3825rM|rx/oq:",
+  pmovsxwd_2 =	"rro:660F3823rM|rx/oq:",
+  pmovsxwq_2 =	"rro:660F3824rM|rx/od:",
+  pmovzxbd_2 =	"rro:660F3831rM|rx/od:",
+  pmovzxbq_2 =	"rro:660F3832rM|rx/ow:",
+  pmovzxbw_2 =	"rro:660F3830rM|rx/oq:",
+  pmovzxdq_2 =	"rro:660F3835rM|rx/oq:",
+  pmovzxwd_2 =	"rro:660F3833rM|rx/oq:",
+  pmovzxwq_2 =	"rro:660F3834rM|rx/od:",
+  pmuldq_2 =	"rmo:660F3828rM",
+  pmulld_2 =	"rmo:660F3840rM",
+  ptest_2 =	"rmo:660F3817rM",
+  roundpd_3 =	"rmio:660F3A09rMU",
+  roundps_3 =	"rmio:660F3A08rMU",
+  roundsd_3 =	"rrio:660F3A0BrMU|rxi/oq:",
+  roundss_3 =	"rrio:660F3A0ArMU|rxi/od:",
+
+  -- SSE4.2 ops
+  crc32_2 =	"rmqd:F20F38F1rM|rm/dw:66F20F38F1rM|rm/db:F20F38F0rM|rm/qb:",
+  pcmpestri_3 =	"rmio:660F3A61rMU",
+  pcmpestrm_3 =	"rmio:660F3A60rMU",
+  pcmpgtq_2 =	"rmo:660F3837rM",
+  pcmpistri_3 =	"rmio:660F3A63rMU",
+  pcmpistrm_3 =	"rmio:660F3A62rMU",
+  popcnt_2 =	"rmqdw:F30FB8rM",
+
+  -- SSE4a
+  extrq_2 =	"rro:660F79rM",
+  extrq_3 =	"riio:660F780mUU",
+  insertq_2 =	"rro:F20F79rM",
+  insertq_4 =	"rriio:F20F78rMUU",
+  lzcnt_2 =	"rmqdw:F30FBDrM",
+  movntsd_2 =	"xr/qo:nF20F2BRm",
+  movntss_2 =	"xr/do:F30F2BRm",
+  -- popcnt is also in SSE4.2
+
+  -- AES-NI
+  aesdec_2 =	"rmo:660F38DErM",
+  aesdeclast_2 = "rmo:660F38DFrM",
+  aesenc_2 =	"rmo:660F38DCrM",
+  aesenclast_2 = "rmo:660F38DDrM",
+  aesimc_2 =	"rmo:660F38DBrM",
+  aeskeygenassist_3 = "rmio:660F3ADFrMU",
+  pclmulqdq_3 =	"rmio:660F3A44rMU",
+
+   -- AVX FP ops
+  vaddsubpd_3 =	"rrmoy:660FVD0rM",
+  vaddsubps_3 =	"rrmoy:F20FVD0rM",
+  vandpd_3 =	"rrmoy:660FV54rM",
+  vandps_3 =	"rrmoy:0FV54rM",
+  vandnpd_3 =	"rrmoy:660FV55rM",
+  vandnps_3 =	"rrmoy:0FV55rM",
+  vblendpd_4 =	"rrmioy:660F3AV0DrMU",
+  vblendps_4 =	"rrmioy:660F3AV0CrMU",
+  vblendvpd_4 =	"rrmroy:660F3AV4BrMs",
+  vblendvps_4 =	"rrmroy:660F3AV4ArMs",
+  vbroadcastf128_2 = "rx/yo:660F38u1ArM",
+  vcmppd_4 =	"rrmioy:660FVC2rMU",
+  vcmpps_4 =	"rrmioy:0FVC2rMU",
+  vcmpsd_4 =	"rrrio:F20FVC2rMU|rrxi/ooq:",
+  vcmpss_4 =	"rrrio:F30FVC2rMU|rrxi/ood:",
+  vcomisd_2 =	"rro:660Fu2FrM|rx/oq:",
+  vcomiss_2 =	"rro:0Fu2FrM|rx/od:",
+  vcvtdq2pd_2 =	"rro:F30FuE6rM|rx/oq:|rm/yo:",
+  vcvtdq2ps_2 =	"rmoy:0Fu5BrM",
+  vcvtpd2dq_2 =	"rmoy:F20FuE6rM",
+  vcvtpd2ps_2 =	"rmoy:660Fu5ArM",
+  vcvtps2dq_2 =	"rmoy:660Fu5BrM",
+  vcvtps2pd_2 =	"rro:0Fu5ArM|rx/oq:|rm/yo:",
+  vcvtsd2si_2 =	"rr/do:F20Fu2DrM|rx/dq:|rr/qo:|rxq:",
+  vcvtsd2ss_3 =	"rrro:F20FV5ArM|rrx/ooq:",
+  vcvtsi2sd_3 =	"rrm/ood:F20FV2ArM|rrm/ooq:F20FVX2ArM",
+  vcvtsi2ss_3 =	"rrm/ood:F30FV2ArM|rrm/ooq:F30FVX2ArM",
+  vcvtss2sd_3 =	"rrro:F30FV5ArM|rrx/ood:",
+  vcvtss2si_2 =	"rr/do:F30Fu2DrM|rxd:|rr/qo:|rx/qd:",
+  vcvttpd2dq_2 = "rmo:660FuE6rM|rm/oy:660FuLE6rM",
+  vcvttps2dq_2 = "rmoy:F30Fu5BrM",
+  vcvttsd2si_2 = "rr/do:F20Fu2CrM|rx/dq:|rr/qo:|rxq:",
+  vcvttss2si_2 = "rr/do:F30Fu2CrM|rxd:|rr/qo:|rx/qd:",
+  vdppd_4 =	"rrmio:660F3AV41rMU",
+  vdpps_4 =	"rrmioy:660F3AV40rMU",
+  vextractf128_3 = "mri/oy:660F3AuL19RmU",
+  vextractps_3 = "mri/do:660F3Au17RmU",
+  vhaddpd_3 =	"rrmoy:660FV7CrM",
+  vhaddps_3 =	"rrmoy:F20FV7CrM",
+  vhsubpd_3 =	"rrmoy:660FV7DrM",
+  vhsubps_3 =	"rrmoy:F20FV7DrM",
+  vinsertf128_4 = "rrmi/yyo:660F3AV18rMU",
+  vinsertps_4 =	"rrrio:660F3AV21rMU|rrxi/ood:",
+  vldmxcsr_1 =	"xd:0FuAE2m",
+  vmaskmovps_3 = "rrxoy:660F38V2CrM|xrroy:660F38V2ERm",
+  vmaskmovpd_3 = "rrxoy:660F38V2DrM|xrroy:660F38V2FRm",
+  vmovapd_2 =	"rmoy:660Fu28rM|mroy:660Fu29Rm",
+  vmovaps_2 =	"rmoy:0Fu28rM|mroy:0Fu29Rm",
+  vmovd_2 =	"rm/od:660Fu6ErM|rm/oq:660FuX6ErM|mr/do:660Fu7ERm|mr/qo:",
+  vmovq_2 =	"rro:F30Fu7ErM|rx/oq:|xr/qo:660FuD6Rm",
+  vmovddup_2 =	"rmy:F20Fu12rM|rro:|rx/oq:",
+  vmovhlps_3 =	"rrro:0FV12rM",
+  vmovhpd_2 =	"xr/qo:660Fu17Rm",
+  vmovhpd_3 =	"rrx/ooq:660FV16rM",
+  vmovhps_2 =	"xr/qo:0Fu17Rm",
+  vmovhps_3 =	"rrx/ooq:0FV16rM",
+  vmovlhps_3 =	"rrro:0FV16rM",
+  vmovlpd_2 =	"xr/qo:660Fu13Rm",
+  vmovlpd_3 =	"rrx/ooq:660FV12rM",
+  vmovlps_2 =	"xr/qo:0Fu13Rm",
+  vmovlps_3 =	"rrx/ooq:0FV12rM",
+  vmovmskpd_2 =	"rr/do:660Fu50rM|rr/dy:660FuL50rM",
+  vmovmskps_2 =	"rr/do:0Fu50rM|rr/dy:0FuL50rM",
+  vmovntpd_2 =	"xroy:660Fu2BRm",
+  vmovntps_2 =	"xroy:0Fu2BRm",
+  vmovsd_2 =	"rx/oq:F20Fu10rM|xr/qo:F20Fu11Rm",
+  vmovsd_3 =	"rrro:F20FV10rM",
+  vmovshdup_2 =	"rmoy:F30Fu16rM",
+  vmovsldup_2 =	"rmoy:F30Fu12rM",
+  vmovss_2 =	"rx/od:F30Fu10rM|xr/do:F30Fu11Rm",
+  vmovss_3 =	"rrro:F30FV10rM",
+  vmovupd_2 =	"rmoy:660Fu10rM|mroy:660Fu11Rm",
+  vmovups_2 =	"rmoy:0Fu10rM|mroy:0Fu11Rm",
+  vorpd_3 =	"rrmoy:660FV56rM",
+  vorps_3 =	"rrmoy:0FV56rM",
+  vpermilpd_3 =	"rrmoy:660F38V0DrM|rmioy:660F3Au05rMU",
+  vpermilps_3 =	"rrmoy:660F38V0CrM|rmioy:660F3Au04rMU",
+  vperm2f128_4 = "rrmiy:660F3AV06rMU",
+  vptestpd_2 =	"rmoy:660F38u0FrM",
+  vptestps_2 =	"rmoy:660F38u0ErM",
+  vrcpps_2 =	"rmoy:0Fu53rM",
+  vrcpss_3 =	"rrro:F30FV53rM|rrx/ood:",
+  vrsqrtps_2 =	"rmoy:0Fu52rM",
+  vrsqrtss_3 =	"rrro:F30FV52rM|rrx/ood:",
+  vroundpd_3 =	"rmioy:660F3Au09rMU",
+  vroundps_3 =	"rmioy:660F3Au08rMU",
+  vroundsd_4 =	"rrrio:660F3AV0BrMU|rrxi/ooq:",
+  vroundss_4 =	"rrrio:660F3AV0ArMU|rrxi/ood:",
+  vshufpd_4 =	"rrmioy:660FVC6rMU",
+  vshufps_4 =	"rrmioy:0FVC6rMU",
+  vsqrtps_2 =	"rmoy:0Fu51rM",
+  vsqrtss_2 =	"rro:F30Fu51rM|rx/od:",
+  vsqrtpd_2 =	"rmoy:660Fu51rM",
+  vsqrtsd_2 =	"rro:F20Fu51rM|rx/oq:",
+  vstmxcsr_1 =	"xd:0FuAE3m",
+  vucomisd_2 =	"rro:660Fu2ErM|rx/oq:",
+  vucomiss_2 =	"rro:0Fu2ErM|rx/od:",
+  vunpckhpd_3 =	"rrmoy:660FV15rM",
+  vunpckhps_3 =	"rrmoy:0FV15rM",
+  vunpcklpd_3 =	"rrmoy:660FV14rM",
+  vunpcklps_3 =	"rrmoy:0FV14rM",
+  vxorpd_3 =	"rrmoy:660FV57rM",
+  vxorps_3 =	"rrmoy:0FV57rM",
+  vzeroall_0 =	"0FuL77",
+  vzeroupper_0 = "0Fu77",
+
+  -- AVX2 FP ops
+  vbroadcastss_2 = "rx/od:660F38u18rM|rx/yd:|rro:|rr/yo:",
+  vbroadcastsd_2 = "rx/yq:660F38u19rM|rr/yo:",
+  -- *vgather* (!vsib)
+  vpermpd_3 =	"rmiy:660F3AuX01rMU",
+  vpermps_3 =	"rrmy:660F38V16rM",
+
+  -- AVX, AVX2 integer ops
+  -- In general, xmm requires AVX, ymm requires AVX2.
+  vaesdec_3 =  "rrmo:660F38VDErM",
+  vaesdeclast_3 = "rrmo:660F38VDFrM",
+  vaesenc_3 =  "rrmo:660F38VDCrM",
+  vaesenclast_3 = "rrmo:660F38VDDrM",
+  vaesimc_2 =  "rmo:660F38uDBrM",
+  vaeskeygenassist_3 = "rmio:660F3AuDFrMU",
+  vlddqu_2 =	"rxoy:F20FuF0rM",
+  vmaskmovdqu_2 = "rro:660FuF7rM",
+  vmovdqa_2 =	"rmoy:660Fu6FrM|mroy:660Fu7FRm",
+  vmovdqu_2 =	"rmoy:F30Fu6FrM|mroy:F30Fu7FRm",
+  vmovntdq_2 =	"xroy:660FuE7Rm",
+  vmovntdqa_2 =	"rxoy:660F38u2ArM",
+  vmpsadbw_4 =	"rrmioy:660F3AV42rMU",
+  vpabsb_2 =	"rmoy:660F38u1CrM",
+  vpabsd_2 =	"rmoy:660F38u1ErM",
+  vpabsw_2 =	"rmoy:660F38u1DrM",
+  vpackusdw_3 =	"rrmoy:660F38V2BrM",
+  vpalignr_4 =	"rrmioy:660F3AV0FrMU",
+  vpblendvb_4 =	"rrmroy:660F3AV4CrMs",
+  vpblendw_4 =	"rrmioy:660F3AV0ErMU",
+  vpclmulqdq_4 = "rrmio:660F3AV44rMU",
+  vpcmpeqq_3 =	"rrmoy:660F38V29rM",
+  vpcmpestri_3 = "rmio:660F3Au61rMU",
+  vpcmpestrm_3 = "rmio:660F3Au60rMU",
+  vpcmpgtq_3 =	"rrmoy:660F38V37rM",
+  vpcmpistri_3 = "rmio:660F3Au63rMU",
+  vpcmpistrm_3 = "rmio:660F3Au62rMU",
+  vpextrb_3 =	"rri/do:660F3Au14nRmU|rri/qo:|xri/bo:",
+  vpextrw_3 =	"rri/do:660FuC5rMU|xri/wo:660F3Au15nRmU",
+  vpextrd_3 =	"mri/do:660F3Au16RmU",
+  vpextrq_3 =	"mri/qo:660F3Au16RmU",
+  vphaddw_3 =	"rrmoy:660F38V01rM",
+  vphaddd_3 =	"rrmoy:660F38V02rM",
+  vphaddsw_3 =	"rrmoy:660F38V03rM",
+  vphminposuw_2 = "rmo:660F38u41rM",
+  vphsubw_3 =	"rrmoy:660F38V05rM",
+  vphsubd_3 =	"rrmoy:660F38V06rM",
+  vphsubsw_3 =	"rrmoy:660F38V07rM",
+  vpinsrb_4 =	"rrri/ood:660F3AV20rMU|rrxi/oob:",
+  vpinsrw_4 =	"rrri/ood:660FVC4rMU|rrxi/oow:",
+  vpinsrd_4 =	"rrmi/ood:660F3AV22rMU",
+  vpinsrq_4 =	"rrmi/ooq:660F3AVX22rMU",
+  vpmaddubsw_3 = "rrmoy:660F38V04rM",
+  vpmaxsb_3 =	"rrmoy:660F38V3CrM",
+  vpmaxsd_3 =	"rrmoy:660F38V3DrM",
+  vpmaxuw_3 =	"rrmoy:660F38V3ErM",
+  vpmaxud_3 =	"rrmoy:660F38V3FrM",
+  vpminsb_3 =	"rrmoy:660F38V38rM",
+  vpminsd_3 =	"rrmoy:660F38V39rM",
+  vpminuw_3 =	"rrmoy:660F38V3ArM",
+  vpminud_3 =	"rrmoy:660F38V3BrM",
+  vpmovmskb_2 =	"rr/do:660FuD7rM|rr/dy:660FuLD7rM",
+  vpmovsxbw_2 =	"rroy:660F38u20rM|rx/oq:|rx/yo:",
+  vpmovsxbd_2 =	"rroy:660F38u21rM|rx/od:|rx/yq:",
+  vpmovsxbq_2 =	"rroy:660F38u22rM|rx/ow:|rx/yd:",
+  vpmovsxwd_2 =	"rroy:660F38u23rM|rx/oq:|rx/yo:",
+  vpmovsxwq_2 =	"rroy:660F38u24rM|rx/od:|rx/yq:",
+  vpmovsxdq_2 =	"rroy:660F38u25rM|rx/oq:|rx/yo:",
+  vpmovzxbw_2 =	"rroy:660F38u30rM|rx/oq:|rx/yo:",
+  vpmovzxbd_2 =	"rroy:660F38u31rM|rx/od:|rx/yq:",
+  vpmovzxbq_2 =	"rroy:660F38u32rM|rx/ow:|rx/yd:",
+  vpmovzxwd_2 =	"rroy:660F38u33rM|rx/oq:|rx/yo:",
+  vpmovzxwq_2 =	"rroy:660F38u34rM|rx/od:|rx/yq:",
+  vpmovzxdq_2 =	"rroy:660F38u35rM|rx/oq:|rx/yo:",
+  vpmuldq_3 =	"rrmoy:660F38V28rM",
+  vpmulhrsw_3 =	"rrmoy:660F38V0BrM",
+  vpmulld_3 =	"rrmoy:660F38V40rM",
+  vpshufb_3 =	"rrmoy:660F38V00rM",
+  vpshufd_3 =	"rmioy:660Fu70rMU",
+  vpshufhw_3 =	"rmioy:F30Fu70rMU",
+  vpshuflw_3 =	"rmioy:F20Fu70rMU",
+  vpsignb_3 =	"rrmoy:660F38V08rM",
+  vpsignw_3 =	"rrmoy:660F38V09rM",
+  vpsignd_3 =	"rrmoy:660F38V0ArM",
+  vpslldq_3 =	"rrioy:660Fv737mU",
+  vpsllw_3 =	"rrmoy:660FVF1rM|rrioy:660Fv716mU",
+  vpslld_3 =	"rrmoy:660FVF2rM|rrioy:660Fv726mU",
+  vpsllq_3 =	"rrmoy:660FVF3rM|rrioy:660Fv736mU",
+  vpsraw_3 =	"rrmoy:660FVE1rM|rrioy:660Fv714mU",
+  vpsrad_3 =	"rrmoy:660FVE2rM|rrioy:660Fv724mU",
+  vpsrldq_3 =	"rrioy:660Fv733mU",
+  vpsrlw_3 =	"rrmoy:660FVD1rM|rrioy:660Fv712mU",
+  vpsrld_3 =	"rrmoy:660FVD2rM|rrioy:660Fv722mU",
+  vpsrlq_3 =	"rrmoy:660FVD3rM|rrioy:660Fv732mU",
+  vptest_2 =	"rmoy:660F38u17rM",
+
+  -- AVX2 integer ops
+  vbroadcasti128_2 = "rx/yo:660F38u5ArM",
+  vinserti128_4 = "rrmi/yyo:660F3AV38rMU",
+  vextracti128_3 = "mri/oy:660F3AuL39RmU",
+  vpblendd_4 =	"rrmioy:660F3AV02rMU",
+  vpbroadcastb_2 = "rro:660F38u78rM|rx/ob:|rr/yo:|rx/yb:",
+  vpbroadcastw_2 = "rro:660F38u79rM|rx/ow:|rr/yo:|rx/yw:",
+  vpbroadcastd_2 = "rro:660F38u58rM|rx/od:|rr/yo:|rx/yd:",
+  vpbroadcastq_2 = "rro:660F38u59rM|rx/oq:|rr/yo:|rx/yq:",
+  vpermd_3 =	"rrmy:660F38V36rM",
+  vpermq_3 =	"rmiy:660F3AuX00rMU",
+  -- *vpgather* (!vsib)
+  vperm2i128_4 = "rrmiy:660F3AV46rMU",
+  vpmaskmovd_3 = "rrxoy:660F38V8CrM|xrroy:660F38V8ERm",
+  vpmaskmovq_3 = "rrxoy:660F38VX8CrM|xrroy:660F38VX8ERm",
+  vpsllvd_3 =	"rrmoy:660F38V47rM",
+  vpsllvq_3 =	"rrmoy:660F38VX47rM",
+  vpsravd_3 =	"rrmoy:660F38V46rM",
+  vpsrlvd_3 =	"rrmoy:660F38V45rM",
+  vpsrlvq_3 =	"rrmoy:660F38VX45rM",
+
+  -- Intel ADX
+  adcx_2 =	"rmqd:660F38F6rM",
+  adox_2 =	"rmqd:F30F38F6rM",
+
+  -- BMI1
+  andn_3 =	"rrmqd:0F38VF2rM",
+  bextr_3 =	"rmrqd:0F38wF7rM",
+  blsi_2 =	"rmqd:0F38vF33m",
+  blsmsk_2 =	"rmqd:0F38vF32m",
+  blsr_2 =	"rmqd:0F38vF31m",
+  tzcnt_2 =	"rmqdw:F30FBCrM",
+
+  -- BMI2
+  bzhi_3 =	"rmrqd:0F38wF5rM",
+  mulx_3 =	"rrmqd:F20F38VF6rM",
+  pdep_3 =	"rrmqd:F20F38VF5rM",
+  pext_3 =	"rrmqd:F30F38VF5rM",
+  rorx_3 =	"rmSqd:F20F3AuF0rMS",
+  sarx_3 =	"rmrqd:F30F38wF7rM",
+  shrx_3 =	"rmrqd:F20F38wF7rM",
+  shlx_3 =	"rmrqd:660F38wF7rM",
+
+  -- FMA3
+  vfmaddsub132pd_3 = "rrmoy:660F38VX96rM",
+  vfmaddsub132ps_3 = "rrmoy:660F38V96rM",
+  vfmaddsub213pd_3 = "rrmoy:660F38VXA6rM",
+  vfmaddsub213ps_3 = "rrmoy:660F38VA6rM",
+  vfmaddsub231pd_3 = "rrmoy:660F38VXB6rM",
+  vfmaddsub231ps_3 = "rrmoy:660F38VB6rM",
+
+  vfmsubadd132pd_3 = "rrmoy:660F38VX97rM",
+  vfmsubadd132ps_3 = "rrmoy:660F38V97rM",
+  vfmsubadd213pd_3 = "rrmoy:660F38VXA7rM",
+  vfmsubadd213ps_3 = "rrmoy:660F38VA7rM",
+  vfmsubadd231pd_3 = "rrmoy:660F38VXB7rM",
+  vfmsubadd231ps_3 = "rrmoy:660F38VB7rM",
+
+  vfmadd132pd_3 = "rrmoy:660F38VX98rM",
+  vfmadd132ps_3 = "rrmoy:660F38V98rM",
+  vfmadd132sd_3 = "rrro:660F38VX99rM|rrx/ooq:",
+  vfmadd132ss_3 = "rrro:660F38V99rM|rrx/ood:",
+  vfmadd213pd_3 = "rrmoy:660F38VXA8rM",
+  vfmadd213ps_3 = "rrmoy:660F38VA8rM",
+  vfmadd213sd_3 = "rrro:660F38VXA9rM|rrx/ooq:",
+  vfmadd213ss_3 = "rrro:660F38VA9rM|rrx/ood:",
+  vfmadd231pd_3 = "rrmoy:660F38VXB8rM",
+  vfmadd231ps_3 = "rrmoy:660F38VB8rM",
+  vfmadd231sd_3 = "rrro:660F38VXB9rM|rrx/ooq:",
+  vfmadd231ss_3 = "rrro:660F38VB9rM|rrx/ood:",
+
+  vfmsub132pd_3 = "rrmoy:660F38VX9ArM",
+  vfmsub132ps_3 = "rrmoy:660F38V9ArM",
+  vfmsub132sd_3 = "rrro:660F38VX9BrM|rrx/ooq:",
+  vfmsub132ss_3 = "rrro:660F38V9BrM|rrx/ood:",
+  vfmsub213pd_3 = "rrmoy:660F38VXAArM",
+  vfmsub213ps_3 = "rrmoy:660F38VAArM",
+  vfmsub213sd_3 = "rrro:660F38VXABrM|rrx/ooq:",
+  vfmsub213ss_3 = "rrro:660F38VABrM|rrx/ood:",
+  vfmsub231pd_3 = "rrmoy:660F38VXBArM",
+  vfmsub231ps_3 = "rrmoy:660F38VBArM",
+  vfmsub231sd_3 = "rrro:660F38VXBBrM|rrx/ooq:",
+  vfmsub231ss_3 = "rrro:660F38VBBrM|rrx/ood:",
+
+  vfnmadd132pd_3 = "rrmoy:660F38VX9CrM",
+  vfnmadd132ps_3 = "rrmoy:660F38V9CrM",
+  vfnmadd132sd_3 = "rrro:660F38VX9DrM|rrx/ooq:",
+  vfnmadd132ss_3 = "rrro:660F38V9DrM|rrx/ood:",
+  vfnmadd213pd_3 = "rrmoy:660F38VXACrM",
+  vfnmadd213ps_3 = "rrmoy:660F38VACrM",
+  vfnmadd213sd_3 = "rrro:660F38VXADrM|rrx/ooq:",
+  vfnmadd213ss_3 = "rrro:660F38VADrM|rrx/ood:",
+  vfnmadd231pd_3 = "rrmoy:660F38VXBCrM",
+  vfnmadd231ps_3 = "rrmoy:660F38VBCrM",
+  vfnmadd231sd_3 = "rrro:660F38VXBDrM|rrx/ooq:",
+  vfnmadd231ss_3 = "rrro:660F38VBDrM|rrx/ood:",
+
+  vfnmsub132pd_3 = "rrmoy:660F38VX9ErM",
+  vfnmsub132ps_3 = "rrmoy:660F38V9ErM",
+  vfnmsub132sd_3 = "rrro:660F38VX9FrM|rrx/ooq:",
+  vfnmsub132ss_3 = "rrro:660F38V9FrM|rrx/ood:",
+  vfnmsub213pd_3 = "rrmoy:660F38VXAErM",
+  vfnmsub213ps_3 = "rrmoy:660F38VAErM",
+  vfnmsub213sd_3 = "rrro:660F38VXAFrM|rrx/ooq:",
+  vfnmsub213ss_3 = "rrro:660F38VAFrM|rrx/ood:",
+  vfnmsub231pd_3 = "rrmoy:660F38VXBErM",
+  vfnmsub231ps_3 = "rrmoy:660F38VBErM",
+  vfnmsub231sd_3 = "rrro:660F38VXBFrM|rrx/ooq:",
+  vfnmsub231ss_3 = "rrro:660F38VBFrM|rrx/ood:",
+}
+
+------------------------------------------------------------------------------
+
+-- Arithmetic ops.
+for name,n in pairs{ add = 0, ["or"] = 1, adc = 2, sbb = 3,
+		     ["and"] = 4, sub = 5, xor = 6, cmp = 7 } do
+  local n8 = shl(n, 3)
+  map_op[name.."_2"] = format(
+    "mr:%02XRm|rm:%02XrM|mI1qdw:81%XmI|mS1qdw:83%XmS|Ri1qdwb:%02Xri|mi1qdwb:81%Xmi",
+    1+n8, 3+n8, n, n, 5+n8, n)
+end
+
+-- Shift ops.
+for name,n in pairs{ rol = 0, ror = 1, rcl = 2, rcr = 3,
+		     shl = 4, shr = 5,          sar = 7, sal = 4 } do
+  map_op[name.."_2"] = format("m1:D1%Xm|mC1qdwb:D3%Xm|mi:C1%XmU", n, n, n)
+end
+
+-- Conditional ops.
+for cc,n in pairs(map_cc) do
+  map_op["j"..cc.."_1"] = format("J.:n0F8%XJ", n) -- short: 7%X
+  map_op["set"..cc.."_1"] = format("mb:n0F9%X2m", n)
+  map_op["cmov"..cc.."_2"] = format("rmqdw:0F4%XrM", n) -- P6+
+end
+
+-- FP arithmetic ops.
+for name,n in pairs{ add = 0, mul = 1, com = 2, comp = 3,
+		     sub = 4, subr = 5, div = 6, divr = 7 } do
+  local nc = 0xc0 + shl(n, 3)
+  local nr = nc + (n < 4 and 0 or (n % 2 == 0 and 8 or -8))
+  local fn = "f"..name
+  map_op[fn.."_1"] = format("ff:D8%02Xr|xd:D8%Xm|xq:nDC%Xm", nc, n, n)
+  if n == 2 or n == 3 then
+    map_op[fn.."_2"] = format("Fff:D8%02XR|Fx2d:D8%XM|Fx2q:nDC%XM", nc, n, n)
+  else
+    map_op[fn.."_2"] = format("Fff:D8%02XR|fFf:DC%02Xr|Fx2d:D8%XM|Fx2q:nDC%XM", nc, nr, n, n)
+    map_op[fn.."p_1"] = format("ff:DE%02Xr", nr)
+    map_op[fn.."p_2"] = format("fFf:DE%02Xr", nr)
+  end
+  map_op["fi"..name.."_1"] = format("xd:DA%Xm|xw:nDE%Xm", n, n)
+end
+
+-- FP conditional moves.
+for cc,n in pairs{ b=0, e=1, be=2, u=3, nb=4, ne=5, nbe=6, nu=7 } do
+  local nc = 0xdac0 + shl(band(n, 3), 3) + shl(band(n, 4), 6)
+  map_op["fcmov"..cc.."_1"] = format("ff:%04Xr", nc) -- P6+
+  map_op["fcmov"..cc.."_2"] = format("Fff:%04XR", nc) -- P6+
+end
+
+-- SSE / AVX FP arithmetic ops.
+for name,n in pairs{ sqrt = 1, add = 8, mul = 9,
+		     sub = 12, min = 13, div = 14, max = 15 } do
+  map_op[name.."ps_2"] = format("rmo:0F5%XrM", n)
+  map_op[name.."ss_2"] = format("rro:F30F5%XrM|rx/od:", n)
+  map_op[name.."pd_2"] = format("rmo:660F5%XrM", n)
+  map_op[name.."sd_2"] = format("rro:F20F5%XrM|rx/oq:", n)
+  if n ~= 1 then
+    map_op["v"..name.."ps_3"] = format("rrmoy:0FV5%XrM", n)
+    map_op["v"..name.."ss_3"] = format("rrro:F30FV5%XrM|rrx/ood:", n)
+    map_op["v"..name.."pd_3"] = format("rrmoy:660FV5%XrM", n)
+    map_op["v"..name.."sd_3"] = format("rrro:F20FV5%XrM|rrx/ooq:", n)
+  end
+end
+
+-- SSE2 / AVX / AVX2 integer arithmetic ops (66 0F leaf).
+for name,n in pairs{
+  paddb = 0xFC, paddw = 0xFD, paddd = 0xFE, paddq = 0xD4,
+  paddsb = 0xEC, paddsw = 0xED, packssdw = 0x6B,
+  packsswb = 0x63, packuswb = 0x67, paddusb = 0xDC,
+  paddusw = 0xDD, pand = 0xDB, pandn = 0xDF, pavgb = 0xE0,
+  pavgw = 0xE3, pcmpeqb = 0x74, pcmpeqd = 0x76,
+  pcmpeqw = 0x75, pcmpgtb = 0x64, pcmpgtd = 0x66,
+  pcmpgtw = 0x65, pmaddwd = 0xF5, pmaxsw = 0xEE,
+  pmaxub = 0xDE, pminsw = 0xEA, pminub = 0xDA,
+  pmulhuw = 0xE4, pmulhw = 0xE5, pmullw = 0xD5,
+  pmuludq = 0xF4, por = 0xEB, psadbw = 0xF6, psubb = 0xF8,
+  psubw = 0xF9, psubd = 0xFA, psubq = 0xFB, psubsb = 0xE8,
+  psubsw = 0xE9, psubusb = 0xD8, psubusw = 0xD9,
+  punpckhbw = 0x68, punpckhwd = 0x69, punpckhdq = 0x6A,
+  punpckhqdq = 0x6D, punpcklbw = 0x60, punpcklwd = 0x61,
+  punpckldq = 0x62, punpcklqdq = 0x6C, pxor = 0xEF
+} do
+  map_op[name.."_2"] = format("rmo:660F%02XrM", n)
+  map_op["v"..name.."_3"] = format("rrmoy:660FV%02XrM", n)
+end
+
+------------------------------------------------------------------------------
+
+local map_vexarg = { u = false, v = 1, V = 2, w = 3 }
+
+-- Process pattern string.
+local function dopattern(pat, args, sz, op, needrex)
+  local digit, addin, vex
+  local opcode = 0
+  local szov = sz
+  local narg = 1
+  local rex = 0
+
+  -- Limit number of section buffer positions used by a single dasm_put().
+  -- A single opcode needs a maximum of 6 positions.
+  if secpos+6 > maxsecpos then wflush() end
+
+  -- Process each character.
+  for c in gmatch(pat.."|", ".") do
+    if match(c, "%x") then	-- Hex digit.
+      digit = byte(c) - 48
+      if digit > 48 then digit = digit - 39
+      elseif digit > 16 then digit = digit - 7 end
+      opcode = opcode*16 + digit
+      addin = nil
+    elseif c == "n" then	-- Disable operand size mods for opcode.
+      szov = nil
+    elseif c == "X" then	-- Force REX.W.
+      rex = 8
+    elseif c == "L" then	-- Force VEX.L.
+      vex.l = true
+    elseif c == "r" then	-- Merge 1st operand regno. into opcode.
+      addin = args[1]; opcode = opcode + (addin.reg % 8)
+      if narg < 2 then narg = 2 end
+    elseif c == "R" then	-- Merge 2nd operand regno. into opcode.
+      addin = args[2]; opcode = opcode + (addin.reg % 8)
+      narg = 3
+    elseif c == "m" or c == "M" then	-- Encode ModRM/SIB.
+      local s
+      if addin then
+	s = addin.reg
+	opcode = opcode - band(s, 7)	-- Undo regno opcode merge.
+      else
+	s = band(opcode, 15)	-- Undo last digit.
+	opcode = shr(opcode, 4)
+      end
+      local nn = c == "m" and 1 or 2
+      local t = args[nn]
+      if narg <= nn then narg = nn + 1 end
+      if szov == "q" and rex == 0 then rex = rex + 8 end
+      if t.reg and t.reg > 7 then rex = rex + 1 end
+      if t.xreg and t.xreg > 7 then rex = rex + 2 end
+      if s > 7 then rex = rex + 4 end
+      if needrex then rex = rex + 16 end
+      local psz, sk = wputop(szov, opcode, rex, vex, s < 0, t.vreg or t.vxreg)
+      opcode = nil
+      local imark = sub(pat, -1) -- Force a mark (ugly).
+      -- Put ModRM/SIB with regno/last digit as spare.
+      wputmrmsib(t, imark, s, addin and addin.vreg, psz, sk)
+      addin = nil
+    elseif map_vexarg[c] ~= nil then -- Encode using VEX prefix
+      local b = band(opcode, 255); opcode = shr(opcode, 8)
+      local m = 1
+      if b == 0x38 then m = 2
+      elseif b == 0x3a then m = 3 end
+      if m ~= 1 then b = band(opcode, 255); opcode = shr(opcode, 8) end
+      if b ~= 0x0f then
+	werror("expected `0F', `0F38', or `0F3A' to precede `"..c..
+	  "' in pattern `"..pat.."' for `"..op.."'")
+      end
+      local v = map_vexarg[c]
+      if v then v = remove(args, v) end
+      b = band(opcode, 255)
+      local p = 0
+      if b == 0x66 then p = 1
+      elseif b == 0xf3 then p = 2
+      elseif b == 0xf2 then p = 3 end
+      if p ~= 0 then opcode = shr(opcode, 8) end
+      if opcode ~= 0 then wputop(nil, opcode, 0); opcode = 0 end
+      vex = { m = m, p = p, v = v }
+    else
+      if opcode then -- Flush opcode.
+	if szov == "q" and rex == 0 then rex = rex + 8 end
+	if needrex then rex = rex + 16 end
+	if addin and addin.reg == -1 then
+	  local psz, sk = wputop(szov, opcode - 7, rex, vex, true)
+	  wvreg("opcode", addin.vreg, psz, sk)
+	else
+	  if addin and addin.reg > 7 then rex = rex + 1 end
+	  wputop(szov, opcode, rex, vex)
+	end
+	opcode = nil
+      end
+      if c == "|" then break end
+      if c == "o" then -- Offset (pure 32 bit displacement).
+	wputdarg(args[1].disp); if narg < 2 then narg = 2 end
+      elseif c == "O" then
+	wputdarg(args[2].disp); narg = 3
+      else
+	-- Anything else is an immediate operand.
+	local a = args[narg]
+	narg = narg + 1
+	local mode, imm = a.mode, a.imm
+	if mode == "iJ" and not match(x64 and "J" or "iIJ", c) then
+	  werror("bad operand size for label")
+	end
+	if c == "S" then
+	  wputsbarg(imm)
+	elseif c == "U" then
+	  wputbarg(imm)
+	elseif c == "W" then
+	  wputwarg(imm)
+	elseif c == "i" or c == "I" then
+	  if mode == "iJ" then
+	    wputlabel("IMM_", imm, 1)
+	  elseif mode == "iI" and c == "I" then
+	    waction(sz == "w" and "IMM_WB" or "IMM_DB", imm)
+	  else
+	    wputszarg(sz, imm)
+	  end
+	elseif c == "J" then
+	  if mode == "iPJ" then
+	    waction("REL_A", imm) -- !x64 (secpos)
+	  else
+	    wputlabel("REL_", imm, 2)
+	  end
+	elseif c == "s" then
+	  local reg = a.reg
+	  if reg < 0 then
+	    wputb(0)
+	    wvreg("imm.hi", a.vreg)
+	  else
+	    wputb(shl(reg, 4))
+	  end
+	else
+	  werror("bad char `"..c.."' in pattern `"..pat.."' for `"..op.."'")
+	end
+      end
+    end
+  end
+end
+
+------------------------------------------------------------------------------
+
+-- Mapping of operand modes to short names. Suppress output with '#'.
+local map_modename = {
+  r = "reg", R = "eax", C = "cl", x = "mem", m = "mrm", i = "imm",
+  f = "stx", F = "st0", J = "lbl", ["1"] = "1",
+  I = "#", S = "#", O = "#",
+}
+
+-- Return a table/string showing all possible operand modes.
+local function templatehelp(template, nparams)
+  if nparams == 0 then return "" end
+  local t = {}
+  for tm in gmatch(template, "[^%|]+") do
+    local s = map_modename[sub(tm, 1, 1)]
+    s = s..gsub(sub(tm, 2, nparams), ".", function(c)
+      return ", "..map_modename[c]
+    end)
+    if not match(s, "#") then t[#t+1] = s end
+  end
+  return t
+end
+
+-- Match operand modes against mode match part of template.
+local function matchtm(tm, args)
+  for i=1,#args do
+    if not match(args[i].mode, sub(tm, i, i)) then return end
+  end
+  return true
+end
+
+-- Handle opcodes defined with template strings.
+map_op[".template__"] = function(params, template, nparams)
+  if not params then return templatehelp(template, nparams) end
+  local args = {}
+
+  -- Zero-operand opcodes have no match part.
+  if #params == 0 then
+    dopattern(template, args, "d", params.op, nil)
+    return
+  end
+
+  -- Determine common operand size (coerce undefined size) or flag as mixed.
+  local sz, szmix, needrex
+  for i,p in ipairs(params) do
+    args[i] = parseoperand(p)
+    local nsz = args[i].opsize
+    if nsz then
+      if sz and sz ~= nsz then szmix = true else sz = nsz end
+    end
+    local nrex = args[i].needrex
+    if nrex ~= nil then
+      if needrex == nil then
+	needrex = nrex
+      elseif needrex ~= nrex then
+	werror("bad mix of byte-addressable registers")
+      end
+    end
+  end
+
+  -- Try all match:pattern pairs (separated by '|').
+  local gotmatch, lastpat
+  for tm in gmatch(template, "[^%|]+") do
+    -- Split off size match (starts after mode match) and pattern string.
+    local szm, pat = match(tm, "^(.-):(.*)$", #args+1)
+    if pat == "" then pat = lastpat else lastpat = pat end
+    if matchtm(tm, args) then
+      local prefix = sub(szm, 1, 1)
+      if prefix == "/" then -- Exactly match leading operand sizes.
+	for i = #szm,1,-1 do
+	  if i == 1 then
+	    dopattern(pat, args, sz, params.op, needrex) -- Process pattern.
+	    return
+	  elseif args[i-1].opsize ~= sub(szm, i, i) then
+	    break
+	  end
+	end
+      else -- Match common operand size.
+	local szp = sz
+	if szm == "" then szm = x64 and "qdwb" or "dwb" end -- Default sizes.
+	if prefix == "1" then szp = args[1].opsize; szmix = nil
+	elseif prefix == "2" then szp = args[2].opsize; szmix = nil end
+	if not szmix and (prefix == "." or match(szm, szp or "#")) then
+	  dopattern(pat, args, szp, params.op, needrex) -- Process pattern.
+	  return
+	end
+      end
+      gotmatch = true
+    end
+  end
+
+  local msg = "bad operand mode"
+  if gotmatch then
+    if szmix then
+      msg = "mixed operand size"
+    else
+      msg = sz and "bad operand size" or "missing operand size"
+    end
+  end
+
+  werror(msg.." in `"..opmodestr(params.op, args).."'")
+end
+
+------------------------------------------------------------------------------
+
+-- x64-specific opcode for 64 bit immediates and displacements.
+if x64 then
+  function map_op.mov64_2(params)
+    if not params then return { "reg, imm", "reg, [disp]", "[disp], reg" } end
+    if secpos+2 > maxsecpos then wflush() end
+    local opcode, op64, sz, rex, vreg
+    local op64 = match(params[1], "^%[%s*(.-)%s*%]$")
+    if op64 then
+      local a = parseoperand(params[2])
+      if a.mode ~= "rmR" then werror("bad operand mode") end
+      sz = a.opsize
+      rex = sz == "q" and 8 or 0
+      opcode = 0xa3
+    else
+      op64 = match(params[2], "^%[%s*(.-)%s*%]$")
+      local a = parseoperand(params[1])
+      if op64 then
+	if a.mode ~= "rmR" then werror("bad operand mode") end
+	sz = a.opsize
+	rex = sz == "q" and 8 or 0
+	opcode = 0xa1
+      else
+	if sub(a.mode, 1, 1) ~= "r" or a.opsize ~= "q" then
+	  werror("bad operand mode")
+	end
+	op64 = params[2]
+	if a.reg == -1 then
+	  vreg = a.vreg
+	  opcode = 0xb8
+	else
+	  opcode = 0xb8 + band(a.reg, 7)
+	end
+	rex = a.reg > 7 and 9 or 8
+      end
+    end
+    local psz, sk = wputop(sz, opcode, rex, nil, vreg)
+    wvreg("opcode", vreg, psz, sk)
+    waction("IMM_D", format("(unsigned int)(%s)", op64))
+    waction("IMM_D", format("(unsigned int)((%s)>>32)", op64))
+  end
+end
+
+------------------------------------------------------------------------------
+
+-- Pseudo-opcodes for data storage.
+local function op_data(params)
+  if not params then return "imm..." end
+  local sz = sub(params.op, 2, 2)
+  if sz == "l" then sz = "d" elseif sz == "a" then sz = addrsize end
+  for _,p in ipairs(params) do
+    local a = parseoperand(p, sz == "q")
+    if sub(a.mode, 1, 1) ~= "i" or (a.opsize and a.opsize ~= sz) then
+      werror("bad mode or size in `"..p.."'")
+    end
+    if a.mode == "iJ" then
+      wputlabel("IMM_", a.imm, 1)
+    elseif sz == "q" then
+      wputqarg(a.imm)
+    else
+      wputszarg(sz, a.imm)
+    end
+    if secpos+2 > maxsecpos then wflush() end
+  end
+end
+
+map_op[".byte_*"] = op_data
+map_op[".sbyte_*"] = op_data
+map_op[".word_*"] = op_data
+map_op[".dword_*"] = op_data
+map_op[".qword_*"] = op_data
+map_op[".aword_*"] = op_data
+map_op[".long_*"] = op_data
+map_op[".quad_*"] = op_data
+map_op[".addr_*"] = op_data
+
+------------------------------------------------------------------------------
+
+-- Pseudo-opcode to mark the position where the action list is to be emitted.
+map_op[".actionlist_1"] = function(params)
+  if not params then return "cvar" end
+  local name = params[1] -- No syntax check. You get to keep the pieces.
+  wline(function(out) writeactions(out, name) end)
+end
+
+-- Pseudo-opcode to mark the position where the global enum is to be emitted.
+map_op[".globals_1"] = function(params)
+  if not params then return "prefix" end
+  local prefix = params[1] -- No syntax check. You get to keep the pieces.
+  wline(function(out) writeglobals(out, prefix) end)
+end
+
+-- Pseudo-opcode to mark the position where the global names are to be emitted.
+map_op[".globalnames_1"] = function(params)
+  if not params then return "cvar" end
+  local name = params[1] -- No syntax check. You get to keep the pieces.
+  wline(function(out) writeglobalnames(out, name) end)
+end
+
+-- Pseudo-opcode to mark the position where the extern names are to be emitted.
+map_op[".externnames_1"] = function(params)
+  if not params then return "cvar" end
+  local name = params[1] -- No syntax check. You get to keep the pieces.
+  wline(function(out) writeexternnames(out, name) end)
+end
+
+------------------------------------------------------------------------------
+
+-- Label pseudo-opcode (converted from trailing colon form).
+map_op[".label_2"] = function(params)
+  if not params then return "[1-9] | ->global | =>pcexpr  [, addr]" end
+  if secpos+2 > maxsecpos then wflush() end
+  local a = parseoperand(params[1])
+  local mode, imm = a.mode, a.imm
+  if type(imm) == "number" and (mode == "iJ" or (imm >= 1 and imm <= 9)) then
+    -- Local label (1: ... 9:) or global label (->global:).
+    waction("LABEL_LG", nil, 1)
+    wputxb(imm)
+  elseif mode == "iJ" then
+    -- PC label (=>pcexpr:).
+    waction("LABEL_PC", imm)
+  else
+    werror("bad label definition")
+  end
+  -- SETLABEL must immediately follow LABEL_LG/LABEL_PC.
+  local addr = params[2]
+  if addr then
+    local a = parseoperand(addr)
+    if a.mode == "iPJ" then
+      waction("SETLABEL", a.imm)
+    else
+      werror("bad label assignment")
+    end
+  end
+end
+map_op[".label_1"] = map_op[".label_2"]
+
+------------------------------------------------------------------------------
+
+-- Alignment pseudo-opcode.
+map_op[".align_1"] = function(params)
+  if not params then return "numpow2" end
+  if secpos+1 > maxsecpos then wflush() end
+  local align = tonumber(params[1]) or map_opsizenum[map_opsize[params[1]]]
+  if align then
+    local x = align
+    -- Must be a power of 2 in the range (2 ... 256).
+    for i=1,8 do
+      x = x / 2
+      if x == 1 then
+	waction("ALIGN", nil, 1)
+	wputxb(align-1) -- Action byte is 2**n-1.
+	return
+      end
+    end
+  end
+  werror("bad alignment")
+end
+
+-- Spacing pseudo-opcode.
+map_op[".space_2"] = function(params)
+  if not params then return "num [, filler]" end
+  if secpos+1 > maxsecpos then wflush() end
+  waction("SPACE", params[1])
+  local fill = params[2]
+  if fill then
+    fill = tonumber(fill)
+    if not fill or fill < 0 or fill > 255 then werror("bad filler") end
+  end
+  wputxb(fill or 0)
+end
+map_op[".space_1"] = map_op[".space_2"]
+
+------------------------------------------------------------------------------
+
+-- Pseudo-opcode for (primitive) type definitions (map to C types).
+map_op[".type_3"] = function(params, nparams)
+  if not params then
+    return nparams == 2 and "name, ctype" or "name, ctype, reg"
+  end
+  local name, ctype, reg = params[1], params[2], params[3]
+  if not match(name, "^[%a_][%w_]*$") then
+    werror("bad type name `"..name.."'")
+  end
+  local tp = map_type[name]
+  if tp then
+    werror("duplicate type `"..name.."'")
+  end
+  if reg and not map_reg_valid_base[reg] then
+    werror("bad base register `"..(map_reg_rev[reg] or reg).."'")
+  end
+  -- Add #type to defines. A bit unclean to put it in map_archdef.
+  map_archdef["#"..name] = "sizeof("..ctype..")"
+  -- Add new type and emit shortcut define.
+  local num = ctypenum + 1
+  map_type[name] = {
+    ctype = ctype,
+    ctypefmt = format("Dt%X(%%s)", num),
+    reg = reg,
+  }
+  wline(format("#define Dt%X(_V) (int)(ptrdiff_t)&(((%s *)0)_V)", num, ctype))
+  ctypenum = num
+end
+map_op[".type_2"] = map_op[".type_3"]
+
+-- Dump type definitions.
+local function dumptypes(out, lvl)
+  local t = {}
+  for name in pairs(map_type) do t[#t+1] = name end
+  sort(t)
+  out:write("Type definitions:\n")
+  for _,name in ipairs(t) do
+    local tp = map_type[name]
+    local reg = tp.reg and map_reg_rev[tp.reg] or ""
+    out:write(format("  %-20s %-20s %s\n", name, tp.ctype, reg))
+  end
+  out:write("\n")
+end
+
+------------------------------------------------------------------------------
+
+-- Set the current section.
+function _M.section(num)
+  waction("SECTION")
+  wputxb(num)
+  wflush(true) -- SECTION is a terminal action.
+end
+
+------------------------------------------------------------------------------
+
+-- Dump architecture description.
+function _M.dumparch(out)
+  out:write(format("DynASM %s version %s, released %s\n\n",
+    _info.arch, _info.version, _info.release))
+  dumpregs(out)
+  dumpactions(out)
+end
+
+-- Dump all user defined elements.
+function _M.dumpdef(out, lvl)
+  dumptypes(out, lvl)
+  dumpglobals(out, lvl)
+  dumpexterns(out, lvl)
+end
+
+------------------------------------------------------------------------------
+
+-- Pass callbacks from/to the DynASM core.
+function _M.passcb(wl, we, wf, ww)
+  wline, werror, wfatal, wwarn = wl, we, wf, ww
+  return wflush
+end
+
+-- Setup the arch-specific module.
+function _M.setup(arch, opt)
+  g_arch, g_opt = arch, opt
+end
+
+-- Merge the core maps and the arch-specific maps.
+function _M.mergemaps(map_coreop, map_def)
+  setmetatable(map_op, { __index = map_coreop })
+  setmetatable(map_def, { __index = map_archdef })
+  return map_op, map_def
+end
+
+return _M
+
+------------------------------------------------------------------------------
+