Merge branch 'v2.1' into v2.1-agentzh

doc/changes.html

@@ -65,7 +65,7 @@ <h1>LuaJIT Change History</h1>
 <div id="main">
 <p>
 This is a list of changes between the released versions of LuaJIT.<br>
-The current <span style="color: #0000c0;">stable version</span> is <strong>LuaJIT&nbsp;2.0.3</strong>.<br>
+The current <span style="color: #0000c0;">stable version</span> is <strong>LuaJIT&nbsp;2.0.4</strong>.<br>
 </p>
 <p>
 Please check the
@@ -74,6 +74,56 @@ <h1>LuaJIT Change History</h1>
 </p>
 
 <div class="major" style="background: #d0d0ff;">
+<h2 id="LuaJIT-2.0.4">LuaJIT 2.0.4 &mdash; 2015-05-14</h2>
+<ul>
+<li>Fix stack check in narrowing optimization.</li>
+<li>Fix Lua/C API typecheck error for special indexes.</li>
+<li>Fix string to number conversion.</li>
+<li>Fix lexer error for chunks without tokens.</li>
+<li>Don't compile <tt>IR_RETF</tt> after <tt>CALLT</tt> to ff with-side effects.</li>
+<li>Fix <tt>BC_UCLO</tt>/<tt>BC_JMP</tt> join optimization in Lua parser.</li>
+<li>Fix corner case in string to number conversion.</li>
+<li>Gracefully handle <tt>lua_error()</tt> for a suspended coroutine.</li>
+<li>Avoid error messages when building with Clang.</li>
+<li>Fix snapshot #0 handling for traces with a stack check on entry.</li>
+<li>Fix fused constant loads under high register pressure.</li>
+<li>Invalidate backpropagation cache after DCE.</li>
+<li>Fix ABC elimination.</li>
+<li>Fix debug info for main chunk of stripped bytecode.</li>
+<li>Fix FOLD rule for <tt>string.sub(s, ...) == k</tt>.</li>
+<li>Fix FOLD rule for <tt>STRREF</tt> of <tt>SNEW</tt>.</li>
+<li>Fix frame traversal while searching for error function.</li>
+<li>Prevent GC estimate miscalculation due to buffer growth.</li>
+<li>Prevent adding side traces for stack checks.</li>
+<li>Fix top slot calculation for snapshots with continuations.</li>
+<li>Fix check for reuse of SCEV results in <tt>FORL</tt>.</li>
+<li>Add PS Vita port.</li>
+<li>Fix compatibility issues with Illumos.</li>
+<li>Fix DragonFly build (unsupported).</li>
+<li>OpenBSD/x86: Better executable memory allocation for W^X mode.</li>
+<li>x86: Fix argument checks for <tt>ipairs()</tt> iterator.</li>
+<li>x86: <tt>lj_math_random_step()</tt> clobbers XMM regs on OSX Clang.</li>
+<li>x86: Fix code generation for unused result of <tt>math.random()</tt>.</li>
+<li>x64: Allow building with <tt>LUAJIT_USE_SYSMALLOC</tt> and <tt>LUAJIT_USE_VALGRIND</tt>.</li>
+<li>x86/x64: Fix argument check for bit shifts.</li>
+<li>x86/x64: Fix code generation for fused test/arith ops.</li>
+<li>ARM: Fix write barrier check in <tt>BC_USETS</tt>.</li>
+<li>PPC: Fix red zone overflow in machine code generation.</li>
+<li>PPC: Don't use <tt>mcrxr</tt> on PPE.</li>
+<li>Various archs: Fix excess stack growth in interpreter.</li>
+<li>FFI: Fix FOLD rule for <tt>TOBIT</tt> + <tt>CONV num.u32</tt>.</li>
+<li>FFI: Prevent DSE across <tt>ffi.string()</tt>.</li>
+<li>FFI: No meta fallback when indexing pointer to incomplete struct.</li>
+<li>FFI: Fix initialization of unions of subtypes.</li>
+<li>FFI: Fix cdata vs. non-cdata arithmetic and comparisons.</li>
+<li>FFI: Fix <tt>__index</tt>/<tt>__newindex</tt> metamethod resolution for ctypes.</li>
+<li>FFI: Fix compilation of reference field access.</li>
+<li>FFI: Fix frame traversal for backtraces with FFI callbacks.</li>
+<li>FFI: Fix recording of indexing a struct pointer ctype object itself.</li>
+<li>FFI: Allow non-scalar cdata to be compared for equality by address.</li>
+<li>FFI: Fix pseudo type conversions for type punning.</li>
+</ul>
+
 <h2 id="LuaJIT-2.0.3">LuaJIT 2.0.3 &mdash; 2014-03-12</h2>
 <ul>
 <li>Add PS4 port.</li>

doc/install.html

@@ -124,7 +124,7 @@ <h1>Installation</h1>
 <td class="compatos">GCC 4.x</td>
 <td class="compatos">ORBIS (<a href="#ps4">PS4</a>)</td>
 <td class="compatos">GCC 4.x</td>
-<td class="compatos">MSVC + SDK v7.0<br>WinSDK v7.0</td>
+<td class="compatos">MSVC + SDK v7.0<br>WinSDK v7.0<br>Durango (<a href="#xboxone">Xbox One</a>)</td>
 </tr>
 <tr class="odd">
 <td class="compatcpu"><a href="#cross2">ARMv5+<br>ARM9E+</a></td>
@@ -190,8 +190,8 @@ <h3>Prerequisites</h3>
 and change to the newly created directory:
 </p>
 <pre class="code">
-tar zxf LuaJIT-2.0.3.tar.gz
-cd LuaJIT-2.0.3</pre>
+tar zxf LuaJIT-2.0.4.tar.gz
+cd LuaJIT-2.0.4</pre>
 <h3>Building LuaJIT</h3>
 <p>
 The supplied Makefiles try to auto-detect the settings needed for your
@@ -516,6 +516,16 @@ <h3 id="consoles">Cross-compiling for consoles</h3>
 cd src
 xedkbuild
 </pre>
+<p>
+To cross-compile for <b id="xboxone">Xbox One</b> from a Windows host,
+open a "Visual Studio .NET Command Prompt" (64&nbsp;bit host compiler),
+<tt>cd</tt> to the directory where you've unpacked the sources and run
+the following commands:
+</p>
+<pre class="code">
+cd src
+xb1build
+</pre>
 
 <h2 id="embed">Embedding LuaJIT</h2>
 <p>

doc/luajit.html

@@ -166,7 +166,7 @@ <h2>Compatibility</h2>
 <tr><td><span style="font-size:90%;">Embedded</span></td><td>Android</td><td>iOS</td></tr>
 </table>
 <table class="feature os os3">
-<tr><td>PS3</td><td>PS4</td><td>PS Vita</td><td>Xbox 360</td></tr>
+<tr><td>PS3</td><td>PS4</td><td>PS Vita</td><td>Xbox 360</td><td>Xbox One</td></tr>
 </table>
 <table class="feature compiler">
 <tr><td>GCC</td><td>CLANG<br>LLVM</td><td>MSVC</td></tr>

doc/running.html

@@ -189,7 +189,7 @@ <h3 id="opt_j"><tt>-j cmd[=arg[,arg...]]</tt></h3>
 read the comment block at the start of their source.
 They can be found in the <tt>lib</tt> directory of the source
 distribution or installed under the <tt>jit</tt> directory. By default
-this is <tt>/usr/local/share/luajit-2.0.3/jit</tt> on POSIX
+this is <tt>/usr/local/share/luajit-2.0.4/jit</tt> on POSIX
 systems.
 </p>
 

dynasm/dasm_ppc.h

@@ -21,7 +21,7 @@ enum {
   /* The following actions need a buffer position. */
   DASM_ALIGN, DASM_REL_LG, DASM_LABEL_LG,
   /* The following actions also have an argument. */
-  DASM_REL_PC, DASM_LABEL_PC, DASM_IMM,
+  DASM_REL_PC, DASM_LABEL_PC, DASM_IMM, DASM_IMMSH,
   DASM__MAX
 };
 
@@ -244,6 +244,10 @@ void dasm_put(Dst_DECL, int start, ...)
 #endif
 	b[pos++] = n;
 	break;
+      case DASM_IMMSH:
+	CK((n >> 6) == 0, RANGE_I);
+	b[pos++] = n;
+	break;
       }
     }
   }
@@ -299,7 +303,7 @@ int dasm_link(Dst_DECL, size_t *szp)
 	case DASM_ALIGN: ofs -= (b[pos++] + ofs) & (ins & 255); break;
 	case DASM_REL_LG: case DASM_REL_PC: pos++; break;
 	case DASM_LABEL_LG: case DASM_LABEL_PC: b[pos++] += ofs; break;
-	case DASM_IMM: pos++; break;
+	case DASM_IMM: case DASM_IMMSH: pos++; break;
 	}
       }
       stop: (void)0;
@@ -366,6 +370,9 @@ int dasm_encode(Dst_DECL, void *buffer)
 	case DASM_IMM:
 	  cp[-1] |= (n & ((1<<((ins>>5)&31))-1)) << (ins&31);
 	  break;
+	case DASM_IMMSH:
+	  cp[-1] |= (ins & 1) ? ((n&31)<<11)|((n&32)>>4) : ((n&31)<<6)|(n&32);
+	  break;
 	default: *cp++ = ins; break;
 	}
       }

dynasm/dasm_ppc.lua

@@ -41,7 +41,7 @@ local wline, werror, wfatal, wwarn
 local action_names = {
   "STOP", "SECTION", "ESC", "REL_EXT",
   "ALIGN", "REL_LG", "LABEL_LG",
-  "REL_PC", "LABEL_PC", "IMM",
+  "REL_PC", "LABEL_PC", "IMM", "IMMSH"
 }
 
 -- Maximum number of section buffer positions for dasm_put().
@@ -230,8 +230,18 @@ local map_cond = {
 
 ------------------------------------------------------------------------------
 
+local map_op, op_template
+
+local function op_alias(opname, f)
+  return function(params, nparams)
+    if not params then return "-> "..opname:sub(1, -3) end
+    f(params, nparams)
+    op_template(params, map_op[opname], nparams)
+  end
+end
+
 -- Template strings for PPC instructions.
-local map_op = {
+map_op = {
   tdi_3 =	"08000000ARI",
   twi_3 =	"0c000000ARI",
   mulli_3 =	"1c000000RRI",
@@ -299,6 +309,33 @@ local map_op = {
   std_2 =	"f8000000RD",
   stdu_2 =	"f8000001RD",
 
+  subi_3 =	op_alias("addi_3", function(p) p[3] = "-("..p[3]..")" end),
+  subis_3 =	op_alias("addis_3", function(p) p[3] = "-("..p[3]..")" end),
+  subic_3 =	op_alias("addic_3", function(p) p[3] = "-("..p[3]..")" end),
+  ["subic._3"] = op_alias("addic._3", function(p) p[3] = "-("..p[3]..")" end),
+
+  rotlwi_3 =	op_alias("rlwinm_5", function(p)
+    p[4] = "0"; p[5] = "31"
+  end),
+  rotrwi_3 =	op_alias("rlwinm_5", function(p)
+    p[3] = "32-("..p[3]..")"; p[4] = "0"; p[5] = "31"
+  end),
+  rotlw_3 =	op_alias("rlwnm_5", function(p)
+    p[4] = "0"; p[5] = "31"
+  end),
+  slwi_3 =	op_alias("rlwinm_5", function(p)
+    p[5] = "31-("..p[3]..")"; p[4] = "0"
+  end),
+  srwi_3 =	op_alias("rlwinm_5", function(p)
+    p[4] = p[3]; p[3] = "32-("..p[3]..")"; p[5] = "31"
+  end),
+  clrlwi_3 =	op_alias("rlwinm_5", function(p)
+    p[4] = p[3]; p[3] = "0"; p[5] = "31"
+  end),
+  clrrwi_3 =	op_alias("rlwinm_5", function(p)
+    p[5] = "31-("..p[3]..")"; p[3] = "0"; p[4] = "0"
+  end),
+
   -- Primary opcode 4:
   mulhhwu_3 =		"10000010RRR.",
   machhwu_3 =		"10000018RRR.",
@@ -790,6 +827,28 @@ local map_op = {
   rldcl_4 =	"78000010RR~RM.",
   rldcr_4 =	"78000012RR~RM.",
 
+  rotldi_3 =	op_alias("rldicl_4", function(p)
+    p[4] = "0"
+  end),
+  rotrdi_3 =	op_alias("rldicl_4", function(p)
+    p[3] = "64-("..p[3]..")"; p[4] = "0"
+  end),
+  rotld_3 =	op_alias("rldcl_4", function(p)
+    p[4] = "0"
+  end),
+  sldi_3 =	op_alias("rldicr_4", function(p)
+    p[4] = "63-("..p[3]..")"
+  end),
+  srdi_3 =	op_alias("rldicl_4", function(p)
+    p[4] = p[3]; p[3] = "64-("..p[3]..")"
+  end),
+  clrldi_3 =	op_alias("rldicl_4", function(p)
+    p[4] = p[3]; p[3] = "0"
+  end),
+  clrrdi_3 =	op_alias("rldicr_4", function(p)
+    p[4] = "63-("..p[3]..")"; p[3] = "0"
+  end),
+
   -- Primary opcode 56:
   lq_2 =	"e0000000R:D", -- NYI: displacement must be divisible by 8.
 
@@ -1358,7 +1417,7 @@ local map_op = {
 do
   local t = {}
   for k,v in pairs(map_op) do
-    if sub(v, -1) == "." then
+    if type(v) == "string" and sub(v, -1) == "." then
       local v2 = sub(v, 1, 7)..char(byte(v, 8)+1)..sub(v, 9, -2)
       t[sub(k, 1, -3).."."..sub(k, -2)] = v2
     end
@@ -1454,8 +1513,30 @@ local function parse_cond(expr)
   werror("bad condition bit name `"..expr.."'")
 end
 
+local parse_ctx = {}
+
+local loadenv = setfenv and function(s)
+  local code = loadstring(s, "")
+  if code then setfenv(code, parse_ctx) end
+  return code
+end or function(s)
+  return load(s, "", nil, parse_ctx)
+end
+
+-- Try to parse simple arithmetic, too, since some basic ops are aliases.
+local function parse_number(n)
+  local x = tonumber(n)
+  if x then return x end
+  local code = loadenv("return "..n)
+  if code then
+    local ok, y = pcall(code)
+    if ok then return y end
+  end
+  return nil
+end
+
 local function parse_imm(imm, bits, shift, scale, signed)
-  local n = tonumber(imm)
+  local n = parse_number(imm)
   if n then
     local m = sar(n, scale)
     if shl(m, scale) == n then
@@ -1479,19 +1560,20 @@ local function parse_imm(imm, bits, shift, scale, signed)
 end
 
 local function parse_shiftmask(imm, isshift)
-  local n = tonumber(imm)
+  local n = parse_number(imm)
   if n then
     if shr(n, 6) == 0 then
-      local lsb = band(imm, 31)
-      local msb = imm - lsb
+      local lsb = band(n, 31)
+      local msb = n - lsb
       return isshift and (shl(lsb, 11)+shr(msb, 4)) or (shl(lsb, 6)+msb)
     end
     werror("out of range immediate `"..imm.."'")
   elseif match(imm, "^r([1-3]?[0-9])$") or
 	 match(imm, "^([%w_]+):(r[1-3]?[0-9])$") then
     werror("expected immediate operand, got register")
   else
-    werror("NYI: parameterized 64 bit shift/mask")
+    waction("IMMSH", isshift and 1 or 0, imm)
+    return 0;
   end
 end
 
@@ -1566,7 +1648,7 @@ end
 ------------------------------------------------------------------------------
 
 -- Handle opcodes defined with template strings.
-map_op[".template__"] = function(params, template, nparams)
+op_template = function(params, template, nparams)
   if not params then return sub(template, 9) end
   local op = tonumber(sub(template, 1, 8), 16)
   local n, rs = 1, 26
@@ -1668,6 +1750,8 @@ map_op[".template__"] = function(params, template, nparams)
   wputpos(pos, op)
 end
 
+map_op[".template__"] = op_template
+
 ------------------------------------------------------------------------------
 
 -- Pseudo-opcode to mark the position where the action list is to be emitted.

src/Makefile

@@ -221,6 +221,11 @@ ifneq (,$(findstring LJ_TARGET_ARM64 ,$(TARGET_TESTARCH)))
   TARGET_LJARCH= arm64
 else
 ifneq (,$(findstring LJ_TARGET_PPC ,$(TARGET_TESTARCH)))
+  ifneq (,$(findstring LJ_LE 1,$(TARGET_TESTARCH)))
+    TARGET_ARCH= -DLJ_ARCH_ENDIAN=LUAJIT_LE
+  else
+    TARGET_ARCH= -DLJ_ARCH_ENDIAN=LUAJIT_BE
+  endif
   TARGET_LJARCH= ppc
 else
 ifneq (,$(findstring LJ_TARGET_MIPS ,$(TARGET_TESTARCH)))
@@ -243,9 +248,6 @@ ifneq (,$(findstring LJ_TARGET_PS3 1,$(TARGET_TESTARCH)))
   TARGET_XCFLAGS+= -DLUAJIT_USE_SYSMALLOC
   TARGET_XLIBS+= -lpthread
 endif
-ifneq (,$(findstring LJ_NO_UNWIND 1,$(TARGET_TESTARCH)))
-  TARGET_ARCH+= -DLUAJIT_NO_UNWIND
-endif
 
 TARGET_XCFLAGS+= $(CCOPT_$(TARGET_LJARCH))
 TARGET_ARCH+= $(patsubst %,-DLUAJIT_TARGET=LUAJIT_ARCH_%,$(TARGET_LJARCH))
@@ -400,6 +402,10 @@ ifeq (,$(findstring LJ_ABI_SOFTFP 1,$(TARGET_TESTARCH)))
 else
   TARGET_ARCH+= -DLJ_ABI_SOFTFP=1
 endif
+ifneq (,$(findstring LJ_NO_UNWIND 1,$(TARGET_TESTARCH)))
+  DASM_AFLAGS+= -D NO_UNWIND
+  TARGET_ARCH+= -DLUAJIT_NO_UNWIND
+endif
 DASM_AFLAGS+= -D VER=$(subst LJ_ARCH_VERSION_,,$(filter LJ_ARCH_VERSION_%,$(subst LJ_ARCH_VERSION ,LJ_ARCH_VERSION_,$(TARGET_TESTARCH))))
 ifeq (Windows,$(TARGET_SYS))
   DASM_AFLAGS+= -D WIN

src/lib_base.c

@@ -428,7 +428,7 @@ LJLIB_CF(dofile)
 
 LJLIB_CF(gcinfo)
 {
-  setintV(L->top++, (G(L)->gc.total >> 10));
+  setintV(L->top++, (int32_t)(G(L)->gc.total >> 10));
   return 1;
 }
 

src/lib_io.c

@@ -99,7 +99,7 @@ static int io_file_close(lua_State *L, IOFileUD *iof)
     int stat = -1;
 #if LJ_TARGET_POSIX
     stat = pclose(iof->fp);
-#elif LJ_TARGET_WINDOWS
+#elif LJ_TARGET_WINDOWS && !LJ_TARGET_XBOXONE
     stat = _pclose(iof->fp);
 #else
     lua_assert(0);
@@ -400,7 +400,7 @@ LJLIB_CF(io_open)
 
 LJLIB_CF(io_popen)
 {
-#if LJ_TARGET_POSIX || LJ_TARGET_WINDOWS
+#if LJ_TARGET_POSIX || (LJ_TARGET_WINDOWS && !LJ_TARGET_XBOXONE)
   const char *fname = strdata(lj_lib_checkstr(L, 1));
   GCstr *s = lj_lib_optstr(L, 2);
   const char *mode = s ? strdata(s) : "r";