Update LuaJIT to LuaJIT/LuaJIT@03080b7

2026-08-21 13:10:24 +02:00 · 2022-08-25 18:33:56 +08:00
parent 27276a031d
commit 65e6adcbd1
69 changed files with 992 additions and 553 deletions
@@ -33,7 +33,8 @@ DPREFIX= $(DESTDIR)$(PREFIX)
 INSTALL_BIN=   $(DPREFIX)/bin
 INSTALL_LIB=   $(DPREFIX)/$(MULTILIB)
 INSTALL_SHARE= $(DPREFIX)/share
-INSTALL_INC=   $(DPREFIX)/include/luajit-$(MAJVER).$(MINVER)
+INSTALL_DEFINC= $(DPREFIX)/include/luajit-$(MAJVER).$(MINVER)
+INSTALL_INC=   $(INSTALL_DEFINC)

 INSTALL_LJLIBD= $(INSTALL_SHARE)/luajit-$(VERSION)
 INSTALL_JITLIB= $(INSTALL_LJLIBD)/jit
@@ -78,6 +79,9 @@ UNINSTALL= $(RM)
 LDCONFIG= ldconfig -n 2>/dev/null
 SED_PC= sed -e "s|^prefix=.*|prefix=$(PREFIX)|" \
            -e "s|^multilib=.*|multilib=$(MULTILIB)|"
+ifneq ($(INSTALL_DEFINC),$(INSTALL_INC))
+  SED_PC+= -e "s|^includedir=.*|includedir=$(INSTALL_INC)|"
+endif

 FILE_T= luajit
 FILE_A= libluajit.a
@@ -86,7 +86,7 @@ overhead. In conjunction with the FFI library, they allow zero-copy
 operations.
 </p>
 <p>
-The string buffer libary also includes a high-performance
+The string buffer library also includes a high-performance
 <a href="serialize">serializer</a> for Lua objects.
 </p>

@@ -434,8 +434,8 @@ unsupported object types, circular references or deeply nested tables.
 <h3 id="buffer_decode"><tt>obj = buffer.decode(str)<br>
 obj = buf:decode()</tt></h3>
 <p>
-The stand-alone function de-serializes (decodes) the string
-<tt>str</tt>, the buffer method de-serializes one object from the
+The stand-alone function deserializes (decodes) the string
+<tt>str</tt>, the buffer method deserializes one object from the
 buffer. Both return a Lua object <tt>obj</tt>.
 </p>
 <p>
@@ -449,7 +449,7 @@ data after decoding a single top-level object. The buffer method leaves
 any left-over data in the buffer.
 </p>
 <p>
-Attempting to de-serialize an FFI type will throw an error, if the FFI
+Attempting to deserialize an FFI type will throw an error, if the FFI
 library is not built-in or has not been loaded, yet.
 </p>

@@ -462,7 +462,7 @@ the following members (all optional):
 <li>
 <tt>dict</tt> is a Lua table holding a <b>dictionary of strings</b> that
 commonly occur as table keys of objects you are serializing. These keys
-are compactly encoded as indexes during serialization. A well chosen
+are compactly encoded as indexes during serialization. A well-chosen
 dictionary saves space and improves serialization performance.
 </li>
 <li>
@@ -624,7 +624,7 @@ errors are best caught with an outer wrapper for larger parts of code.
 There's not much one can do after that, anyway.
 </p>
 <p>
-OTOH you may want to catch some errors individually. Buffer methods need
+OTOH, you may want to catch some errors individually. Buffer methods need
 to receive the buffer object as the first argument. The Lua colon-syntax
 <tt>obj:method()</tt> does that implicitly. But to wrap a method with
 <tt>pcall()</tt>, the arguments need to be passed like this:
@@ -107,7 +107,7 @@ Turn the whole JIT compiler on or off or flush the whole cache of compiled code.
 This sets the mode for the function at the stack index <tt>idx</tt> or
 the parent of the calling function (<tt>idx = 0</tt>). It either
 enables JIT compilation for a function, disables it and flushes any
-already compiled code or only flushes already compiled code. This
+already compiled code, or only flushes already compiled code. This
 applies recursively to all sub-functions of the function with
 <tt>LUAJIT_MODE_ALLFUNC</tt> or only to the sub-functions with
 <tt>LUAJIT_MODE_ALLSUBFUNC</tt>.
@@ -126,7 +126,7 @@ traces which link to it.
 This mode defines a wrapper function for calls to C functions. If
 called with <tt>LUAJIT_MODE_ON</tt>, the stack index at <tt>idx</tt>
 must be a <tt>lightuserdata</tt> object holding a pointer to the wrapper
-function. From now on all C functions are called through the wrapper
+function. From now on, all C functions are called through the wrapper
 function. If called with <tt>LUAJIT_MODE_OFF</tt> this mode is turned
 off and all C functions are directly called.
 </p>
@@ -157,7 +157,7 @@ call the binding function. Phew!
 <h2 id="cdata">Motivating Example: Using C Data Structures</h2>
 <p>
 The FFI library allows you to create and access C&nbsp;data
-structures. Of course the main use for this is for interfacing with
+structures. Of course, the main use for this is for interfacing with
 C&nbsp;functions. But they can be used stand-alone, too.
 </p>
 <p>
@@ -169,7 +169,7 @@ implemented with a big table holding lots of tiny tables. This imposes
 both a substantial memory overhead as well as a performance overhead.
 </p>
 <p>
-Here's a sketch of a library that operates on color images plus a
+Here's a sketch of a library that operates on color images, plus a
 simple benchmark. First, the plain Lua version:
 </p>
 <pre class="code">
@@ -184,7 +184,7 @@ local function image_ramp_green(n)
  return img
 end

-local function image_to_grey(img, n)
+local function image_to_gray(img, n)
  for i=1,n do
    local y = floor(0.3*img[i].red + 0.59*img[i].green + 0.11*img[i].blue)
    img[i].red = y; img[i].green = y; img[i].blue = y
@@ -194,14 +194,14 @@ end
 local N = 400*400
 local img = image_ramp_green(N)
 for i=1,1000 do
-  image_to_grey(img, N)
+  image_to_gray(img, N)
 end
 </pre>
 <p>
 This creates a table with 160.000 pixels, each of which is a table
-holding four number values in the range of 0-255. First an image with
+holding four number values in the range of 0-255. First, an image with
 a green ramp is created (1D for simplicity), then the image is
-converted to greyscale 1000 times. Yes, that's silly, but I was in
+converted to grayscale 1000 times. Yes, that's silly, but I was in
 need of a simple example ...
 </p>
 <p>
@@ -308,7 +308,7 @@ be more compact and faster. This is certainly true (by a factor of
 ~1.7x). Switching to a struct-of-arrays would help, too.
 </p>
 <p style="font-size: 8pt;">
-However the resulting code would be less idiomatic and rather
+However, the resulting code would be less idiomatic and rather
 error-prone. And it still doesn't get even close to the performance of
 the FFI version of the code. Also, high-level data structures cannot
 be easily passed to other C&nbsp;functions, especially I/O functions,
@@ -121,7 +121,7 @@ separated by semicolons. The trailing semicolon for a single
 declaration may be omitted.
 </p>
 <p>
-Please note that external symbols are only <em>declared</em>, but they
+Please note, that external symbols are only <em>declared</em>, but they
 are <em>not bound</em> to any specific address, yet. Binding is
 achieved with C&nbsp;library namespaces (see below).
 </p>
@@ -209,7 +209,7 @@ parse the cdecl only once and get its ctype with
 <tt>ffi.typeof()</tt>. Then use the ctype as a constructor repeatedly.
 </p>
 <p style="font-size: 8pt;">
-Please note that an anonymous <tt>struct</tt> declaration implicitly
+Please note, that an anonymous <tt>struct</tt> declaration implicitly
 creates a new and distinguished ctype every time you use it for
 <tt>ffi.new()</tt>. This is probably <b>not</b> what you want,
 especially if you create more than one cdata object. Different anonymous
@@ -256,12 +256,12 @@ afterwards. Neither the contents of the <tt>metatable</tt> nor the
 contents of an <tt>__index</tt> table (if any) may be modified
 afterwards. The associated metatable automatically applies to all uses
 of this type, no matter how the objects are created or where they
-originate from. Note that pre-defined operations on types have
+originate from. Note that predefined operations on types have
 precedence (e.g. declared field names cannot be overridden).
 </p>
 <p>
 All standard Lua metamethods are implemented. These are called directly,
-without shortcuts and on any mix of types. For binary operations, the
+without shortcuts, and on any mix of types. For binary operations, the
 left operand is checked first for a valid ctype metamethod. The
 <tt>__gc</tt> metamethod only applies to <tt>struct</tt>/<tt>union</tt>
 types and performs an implicit <a href="#ffi_gc"><tt>ffi.gc()</tt></a>
@@ -492,7 +492,7 @@ have some extra methods:
 <p>
 Free the resources associated with a callback. The associated Lua
 function is unanchored and may be garbage collected. The callback
-function pointer is no longer valid and must not be called anymore
+function pointer is no longer valid and must not be called again
 (it may be reused by a subsequently created callback).
 </p>

@@ -88,7 +88,7 @@ footprint. It's used by the <a href="ext_ffi_api.html">ffi.* library
 functions</a> to declare C&nbsp;types or external symbols.
 </p>
 <p>
-It's only purpose is to parse C&nbsp;declarations, as found e.g. in
+Its only purpose is to parse C&nbsp;declarations, as found e.g. in
 C&nbsp;header files. Although it does evaluate constant expressions,
 it's <em>not</em> a C&nbsp;compiler. The body of <tt>inline</tt>
 C&nbsp;function definitions is simply ignored.
@@ -165,7 +165,7 @@ function declarations.</li>

 </ul>
 <p>
-The following C&nbsp;types are pre-defined by the C&nbsp;parser (like
+The following C&nbsp;types are predefined by the C&nbsp;parser (like
 a <tt>typedef</tt>, except re-declarations will be ignored):
 </p>
 <ul>
@@ -583,9 +583,9 @@ ffi.new("struct nested", {x=1,y={2,3}}) --> x = 1, y.a = 2, y.b = 3

 <h2 id="cdata_ops">Operations on cdata Objects</h2>
 <p>
-All of the standard Lua operators can be applied to cdata objects or a
+All standard Lua operators can be applied to cdata objects or a
 mix of a cdata object and another Lua object. The following list shows
-the pre-defined operations.
+the predefined operations.
 </p>
 <p>
 Reference types are dereferenced <em>before</em> performing each of
@@ -593,7 +593,7 @@ the operations below &mdash; the operation is applied to the
 C&nbsp;type pointed to by the reference.
 </p>
 <p>
-The pre-defined operations are always tried first before deferring to a
+The predefined operations are always tried first before deferring to a
 metamethod or index table (if any) for the corresponding ctype (except
 for <tt>__new</tt>). An error is raised if the metamethod lookup or
 index table lookup fails.
@@ -643,7 +643,7 @@ assigning to an index of a vector raises an error.</li>
 </ul>
 <p>
 A ctype object can be indexed with a string key, too. The only
-pre-defined operation is reading scoped constants of
+predefined operation is reading scoped constants of
 <tt>struct</tt>/<tt>union</tt> types. All other accesses defer
 to the corresponding metamethods or index tables (if any).
 </p>
@@ -656,7 +656,7 @@ certain optimizations.
 <p>
 As a consequence, the <em>elements</em> of complex numbers and
 vectors are immutable. But the elements of an aggregate holding these
-types <em>may</em> be modified of course. I.e. you cannot assign to
+types <em>may</em> be modified, of course. I.e. you cannot assign to
 <tt>foo.c.im</tt>, but you can assign a (newly created) complex number
 to <tt>foo.c</tt>.
 </p>
@@ -675,8 +675,8 @@ through unions is explicitly detected and allowed.
 to <tt>ffi.new(ct, ...)</tt>, unless a <tt>__new</tt> metamethod is
 defined. The <tt>__new</tt> metamethod is called with the ctype object
 plus any other arguments passed to the constructor. Note that you have to
-use <tt>ffi.new</tt> inside of it, since calling <tt>ct(...)</tt> would
-cause infinite recursion.</li>
+use <tt>ffi.new</tt> inside the metamethod, since calling <tt>ct(...)</tt>
+would cause infinite recursion.</li>

 <li><b>C&nbsp;function call</b>: a cdata function or cdata function
 pointer can be called. The passed arguments are
@@ -687,7 +687,7 @@ variable argument part of vararg C&nbsp;function use
 C&nbsp;function is called and the return value (if any) is
 <a href="#convert_tolua">converted to a Lua object</a>.<br>
 On Windows/x86 systems, <tt>__stdcall</tt> functions are automatically
-detected and a function declared as <tt>__cdecl</tt> (the default) is
+detected, and a function declared as <tt>__cdecl</tt> (the default) is
 silently fixed up after the first call.</li>

 </ul>
@@ -697,7 +697,7 @@ silently fixed up after the first call.</li>

 <li><b>Pointer arithmetic</b>: a cdata pointer/array and a cdata
 number or a Lua number can be added or subtracted. The number must be
-on the right hand side for a subtraction. The result is a pointer of
+on the right-hand side for a subtraction. The result is a pointer of
 the same type with an address plus or minus the number value
 multiplied by the element size in bytes. An error is raised if the
 element size is undefined.</li>
@@ -712,7 +712,7 @@ operators (<tt>+&nbsp;-&nbsp;*&nbsp;/&nbsp;%&nbsp;^</tt> and unary
 minus) can be applied to two cdata numbers, or a cdata number and a
 Lua number. If one of them is an <tt>uint64_t</tt>, the other side is
 converted to an <tt>uint64_t</tt> and an unsigned arithmetic operation
-is performed. Otherwise both sides are converted to an
+is performed. Otherwise, both sides are converted to an
 <tt>int64_t</tt> and a signed arithmetic operation is performed. The
 result is a boxed 64&nbsp;bit cdata object.<br>

@@ -759,7 +759,7 @@ which is compatible with any other pointer type.</li>
 <li><b>64&nbsp;bit integer comparison</b>: two cdata numbers, or a
 cdata number and a Lua number can be compared with each other. If one
 of them is an <tt>uint64_t</tt>, the other side is converted to an
-<tt>uint64_t</tt> and an unsigned comparison is performed. Otherwise
+<tt>uint64_t</tt> and an unsigned comparison is performed. Otherwise,
 both sides are converted to an <tt>int64_t</tt> and a signed
 comparison is performed.<br>

@@ -784,9 +784,9 @@ keys!</b>
 A cdata object is treated like any other garbage-collected object and
 is hashed and compared by its address for table indexing. Since
 there's no interning for cdata value types, the same value may be
-boxed in different cdata objects with different addresses. Thus
+boxed in different cdata objects with different addresses. Thus,
 <tt>t[1LL+1LL]</tt> and <tt>t[2LL]</tt> usually <b>do not</b> point to
-the same hash slot and they certainly <b>do not</b> point to the same
+the same hash slot, and they certainly <b>do not</b> point to the same
 hash slot as <tt>t[2]</tt>.
 </p>
 <p>
@@ -808,7 +808,7 @@ the resulting Lua number as a key when indexing tables.<br>
 One obvious benefit: <tt>t[tonumber(2LL)]</tt> <b>does</b> point to
 the same slot as <tt>t[2]</tt>.</li>

-<li>Otherwise use either <tt>tostring()</tt> on 64&nbsp;bit integers
+<li>Otherwise, use either <tt>tostring()</tt> on 64&nbsp;bit integers
 or complex numbers or combine multiple fields of a cdata aggregate to
 a Lua string (e.g. with
 <a href="ext_ffi_api.html#ffi_string"><tt>ffi.string()</tt></a>). Then
@@ -816,7 +816,7 @@ use the resulting Lua string as a key when indexing tables.</li>

 <li>Create your own specialized hash table implementation using the
 C&nbsp;types provided by the FFI library, just like you would in
-C&nbsp;code. Ultimately this may give much better performance than the
+C&nbsp;code. Ultimately, this may give much better performance than the
 other alternatives or what a generic by-value hash table could
 possibly provide.</li>

@@ -882,7 +882,7 @@ garbage collector will automatically free the memory used by it (at
 the end of the next GC cycle).
 </p>
 <p>
-Please note that pointers themselves are cdata objects, however they
+Please note, that pointers themselves are cdata objects, however they
 are <b>not</b> followed by the garbage collector. So e.g. if you
 assign a cdata array to a pointer, you must keep the cdata object
 holding the array alive as long as the pointer is still in use:
@@ -931,18 +931,18 @@ of the function pointer and the Lua function object (closure).
 </p>
 <p>
 This can happen implicitly due to the usual conversions, e.g. when
-passing a Lua function to a function pointer argument. Or you can use
+passing a Lua function to a function pointer argument. Or, you can use
 <tt>ffi.cast()</tt> to explicitly cast a Lua function to a
 C&nbsp;function pointer.
 </p>
 <p>
-Currently only certain C&nbsp;function types can be used as callback
+Currently, only certain C&nbsp;function types can be used as callback
 functions. Neither C&nbsp;vararg functions nor functions with
 pass-by-value aggregate argument or result types are supported. There
-are no restrictions for the kind of Lua functions that can be called
+are no restrictions on the kind of Lua functions that can be called
 from the callback &mdash; no checks for the proper number of arguments
 are made. The return value of the Lua function will be converted to the
-result type and an error will be thrown for invalid conversions.
+result type, and an error will be thrown for invalid conversions.
 </p>
 <p>
 It's allowed to throw errors across a callback invocation, but it's not
@@ -1003,7 +1003,7 @@ convention cannot be automatically detected, unlike for
 <tt>__stdcall</tt> calls <em>to</em> Windows functions.
 </p>
 <p>
-For some use cases it's necessary to free up the resources or to
+For some use cases, it's necessary to free up the resources or to
 dynamically redirect callbacks. Use an explicit cast to a
 C&nbsp;function pointer and keep the resulting cdata object. Then use
 the <a href="ext_ffi_api.html#callback_free"><tt>cb:free()</tt></a>
@@ -1056,7 +1056,7 @@ GUI application, which waits for user input most of the time, anyway.
 </p>
 <p>
 For new designs <b>avoid push-style APIs</b>: a C&nbsp;function repeatedly
-calling a callback for each result. Instead <b>use pull-style APIs</b>:
+calling a callback for each result. Instead, <b>use pull-style APIs</b>:
 call a C&nbsp;function repeatedly to get a new result. Calls from Lua
 to C via the FFI are much faster than the other way round. Most well-designed
 libraries already use pull-style APIs (read/write, get/put).
@@ -1075,7 +1075,7 @@ function.
 </p>
 <p>
 Indexing a C&nbsp;library namespace object with a symbol name (a Lua
-string) automatically binds it to the library. First the symbol type
+string) automatically binds it to the library. First, the symbol type
 is resolved &mdash; it must have been declared with
 <a href="ext_ffi_api.html#ffi_cdef"><tt>ffi.cdef</tt></a>. Then the
 symbol address is resolved by searching for the symbol name in the
@@ -1130,7 +1130,7 @@ Performance notice: the JIT compiler specializes to the identity of
 namespace objects and to the strings used to index it. This
 effectively turns function cdata objects into constants. It's not
 useful and actually counter-productive to explicitly cache these
-function objects, e.g. <tt>local strlen = ffi.C.strlen</tt>. OTOH it
+function objects, e.g. <tt>local strlen = ffi.C.strlen</tt>. OTOH, it
 <em>is</em> useful to cache the namespace itself, e.g. <tt>local C =
 ffi.C</tt>.
 </p>
@@ -1155,14 +1155,14 @@ This behavior is inevitable, since the goal is to provide full
 interoperability with C&nbsp;code. Adding extra safety measures, like
 bounds checks, would be futile. There's no way to detect
 misdeclarations of C&nbsp;functions, since shared libraries only
-provide symbol names, but no type information. Likewise there's no way
+provide symbol names, but no type information. Likewise, there's no way
 to infer the valid range of indexes for a returned pointer.
 </p>
 <p>
 Again: the FFI library is a low-level library. This implies it needs
 to be used with care, but it's flexibility and performance often
 outweigh this concern. If you're a C or C++ developer, it'll be easy
-to apply your existing knowledge. OTOH writing code for the FFI
+to apply your existing knowledge. OTOH, writing code for the FFI
 library is not for the faint of heart and probably shouldn't be the
 first exercise for someone with little experience in Lua, C or C++.
 </p>
@@ -1190,7 +1190,7 @@ currently incomplete:
 <li>C&nbsp;declarations are not passed through a C&nbsp;pre-processor,
 yet.</li>
 <li>The C&nbsp;parser is able to evaluate most constant expressions
-commonly found in C&nbsp;header files. However it doesn't handle the
+commonly found in C&nbsp;header files. However, it doesn't handle the
 full range of C&nbsp;expression semantics and may fail for some
 obscure constructs.</li>
 <li><tt>static const</tt> declarations only work for integer types
@@ -85,7 +85,7 @@ of its functions:
 local ffi = require("ffi")
 </pre>
 <p>
-Please note this doesn't define an <tt>ffi</tt> variable in the table
+Please note, this doesn't define an <tt>ffi</tt> variable in the table
 of globals &mdash; you really need to use the local variable. The
 <tt>require</tt> function ensures the library is only loaded once.
 </p>
@@ -194,7 +194,7 @@ don't need to declare them as such.
 <span class="mark">&#9316;</span> The <tt>poll()</tt>
 function takes a couple more arguments we're not going to use. You can
 simply use <tt>nil</tt> to pass a <tt>NULL</tt> pointer and <tt>0</tt>
-for the <tt>nfds</tt> parameter. Please note that the
+for the <tt>nfds</tt> parameter. Please note, that the
 number&nbsp;<tt>0</tt> <em>does not convert to a pointer value</em>,
 unlike in C++. You really have to pass pointers to pointer arguments
 and numbers to number arguments.
@@ -291,12 +291,12 @@ Here's the step-by-step explanation:
 <p>
 <span class="mark">&#9312;</span> This defines some of the
 C&nbsp;functions provided by zlib. For the sake of this example, some
-type indirections have been reduced and it uses the pre-defined
+type indirections have been reduced and it uses the predefined
 fixed-size integer types, while still adhering to the zlib API/ABI.
 </p>
 <p>
 <span class="mark">&#9313;</span> This loads the zlib shared
-library. On POSIX systems it's named <tt>libz.so</tt> and usually
+library. On POSIX systems, it's named <tt>libz.so</tt> and usually
 comes pre-installed. Since <tt>ffi.load()</tt> automatically adds any
 missing standard prefixes/suffixes, we can simply load the
 <tt>"z"</tt> library. On Windows it's named <tt>zlib1.dll</tt> and
@@ -324,7 +324,7 @@ actual length that was used.
 <p>
 In C you'd pass in the address of a local variable
 (<tt>&amp;buflen</tt>). But since there's no address-of operator in
-Lua, we'll just pass in a one-element array. Conveniently it can be
+Lua, we'll just pass in a one-element array. Conveniently, it can be
 initialized with the maximum buffer size in one step. Calling the
 actual <tt>zlib.compress2</tt> function is then straightforward.
 </p>
@@ -348,7 +348,7 @@ for garbage collection and string interning.
 <span class="mark">&#9317;</span> The <tt>uncompress</tt>
 functions does the exact opposite of the <tt>compress</tt> function.
 The compressed data doesn't include the size of the original string,
-so this needs to be passed in. Otherwise no surprises here.
+so this needs to be passed in. Otherwise, no surprises here.
 </p>
 <p>
 <span class="mark">&#9318;</span> The code, that makes use
@@ -382,7 +382,7 @@ Ok, so the <tt>ffi.*</tt> functions generally accept cdata objects
 wherever you'd want to use a number. That's why we get a away with
 passing <tt>n</tt> to <tt>ffi.string()</tt> above. But other Lua
 library functions or modules don't know how to deal with this. So for
-maximum portability one needs to use <tt>tonumber()</tt> on returned
+maximum portability, one needs to use <tt>tonumber()</tt> on returned
 <tt>long</tt> results before passing them on. Otherwise the
 application might work on some systems, but would fail in a POSIX/x64
 environment.
@@ -454,7 +454,7 @@ the origin.
 </p>
 <p>
 <span class="mark">&#9315;</span> If we run out of operators, we can
-define named methods, too. Here the <tt>__index</tt> table defines an
+define named methods, too. Here, the <tt>__index</tt> table defines an
 <tt>area</tt> function. For custom indexing needs, one might want to
 define <tt>__index</tt> and <tt>__newindex</tt> <em>functions</em> instead.
 </p>
@@ -468,13 +468,13 @@ be used e.g. to create an array of points. The metamethods automatically
 apply to any and all uses of this type.
 </p>
 <p>
-Please note that the association with a metatable is permanent and
+Please note, that the association with a metatable is permanent and
 <b>the metatable must not be modified afterwards!</b> Ditto for the
 <tt>__index</tt> table.
 </p>
 <p>
 <span class="mark">&#9317;</span> Here are some simple usage examples
-for the point type and their expected results. The pre-defined
+for the point type and their expected results. The predefined
 operations (such as <tt>a.x</tt>) can be freely mixed with the newly
 defined metamethods. Note that <tt>area</tt> is a method and must be
 called with the Lua syntax for methods: <tt>a:area()</tt>, not
@@ -483,7 +483,7 @@ called with the Lua syntax for methods: <tt>a:area()</tt>, not
 <p>
 The C&nbsp;type metamethod mechanism is most useful when used in
 conjunction with C&nbsp;libraries that are written in an object-oriented
-style. Creators return a pointer to a new instance and methods take an
+style. Creators return a pointer to a new instance, and methods take an
 instance pointer as the first argument. Sometimes you can just point
 <tt>__index</tt> to the library namespace and <tt>__gc</tt> to the
 destructor and you're done. But often enough you'll want to add
@@ -569,7 +569,7 @@ end
 </pre>
 <p>
 This turns them into indirect calls and generates bigger and slower
-machine code. Instead you'll want to cache the namespace itself and
+machine code. Instead, you'll want to cache the namespace itself and
 rely on the JIT compiler to eliminate the lookups:
 </p>
 <pre class="code">
@@ -154,7 +154,7 @@ Contains the target architecture name:

 <h2 id="jit_opt"><tt>jit.opt.*</tt> &mdash; JIT compiler optimization control</h2>
 <p>
-This sub-module provides the backend for the <tt>-O</tt> command line
+This submodule provides the backend for the <tt>-O</tt> command line
 option.
 </p>
 <p>
@@ -174,7 +174,7 @@ which was one of the ways to enable optimization.

 <h2 id="jit_util"><tt>jit.util.*</tt> &mdash; JIT compiler introspection</h2>
 <p>
-This sub-module holds functions to introspect the bytecode, generated
+This submodule holds functions to introspect the bytecode, generated
 traces, the IR and the generated machine code. The functionality
 provided by this module is still in flux and therefore undocumented.
 </p>
@@ -158,7 +158,7 @@ To see how much time is spent in different VM states or
 Combinations of <tt>v/z</tt> with <tt>f/F/l</tt> produce two-level
 views, e.g. <tt>-jp=vf</tt> or <tt>-jp=fv</tt>. This shows the time
 spent in a VM state or zone vs. hotspots. This can be used to answer
-questions like "Which time consuming functions are only interpreted?" or
+questions like "Which time-consuming functions are only interpreted?" or
 "What's the garbage collector overhead for a specific function?".
 </p>
 <p>
@@ -217,7 +217,7 @@ local profile = require("jit.profile")
 This module can be used to implement your own higher-level profiler.
 A typical profiling run starts the profiler, captures stack dumps in
 the profiler callback, adds them to a hash table to aggregate the number
-of samples, stops the profiler and then analyzes all of the captured
+of samples, stops the profiler and then analyzes all captured
 stack dumps. Other parameters can be sampled in the profiler callback,
 too. But it's important not to spend too much time in the callback,
 since this may skew the statistics.
@@ -271,9 +271,9 @@ returns a string with a stack dump for the <tt>thread</tt> (coroutine),
 formatted according to the <tt>fmt</tt> argument:
 </p>
 <ul>
-<li><tt>p</tt> &mdash; Preserve the full path for module names. Otherwise
+<li><tt>p</tt> &mdash; Preserve the full path for module names. Otherwise,
 only the file name is used.</li>
-<li><tt>f</tt> &mdash; Dump the function name if it can be derived. Otherwise
+<li><tt>f</tt> &mdash; Dump the function name if it can be derived. Otherwise,
 use module:line.</li>
 <li><tt>F</tt> &mdash; Ditto, but dump module:name.</li>
 <li><tt>l</tt> &mdash; Dump module:line.</li>
@@ -88,7 +88,7 @@ or LuaJIT.
 </p>
 <p>
 LuaJIT extends the standard Lua VM with new functionality and adds
-several extension modules. Please note this page is only about
+several extension modules. Please note, this page is only about
 <em>functional</em> enhancements and not about performance enhancements,
 such as the optimized VM, the faster interpreter or the JIT compiler.
 </p>
@@ -197,7 +197,7 @@ usage. See also the
 </p>
 <p>
 The generated bytecode is portable and can be loaded on any architecture
-that LuaJIT supports, independent of word size or endianess. However the
+that LuaJIT supports, independent of word size or endianess. However, the
 bytecode compatibility versions must match. Bytecode stays compatible
 for dot releases (x.y.0 &rarr; x.y.1), but may change with major or
 minor releases (2.0 &rarr; 2.1) or between any beta release. Foreign
@@ -229,7 +229,7 @@ avoids managing backlinks, saves an allocation and the overhead of
 incremental array/hash part growth.
 </p>
 <p>
-Please note this function is meant for very specific situations. In most
+Please note, this function is meant for very specific situations. In most
 cases it's better to replace the (usually single) link with a new table
 and let the GC do its work.
 </p>
@@ -239,7 +239,7 @@ and let the GC do its work.
 LuaJIT uses a Tausworthe PRNG with period 2^223 to implement
 <tt>math.random()</tt> and <tt>math.randomseed()</tt>. The quality of
 the PRNG results is much superior compared to the standard Lua
-implementation which uses the platform-specific ANSI rand().
+implementation, which uses the platform-specific ANSI rand().
 </p>
 <p>
 The PRNG generates the same sequences from the same seeds on all
@@ -257,7 +257,7 @@ Important: Neither this nor any other PRNG based on the simplistic
 <h3 id="io"><tt>io.*</tt> functions handle 64&nbsp;bit file offsets</h3>
 <p>
 The file I/O functions in the standard <tt>io.*</tt> library handle
-64&nbsp;bit file offsets. In particular this means it's possible
+64&nbsp;bit file offsets. In particular, this means it's possible
 to open files larger than 2&nbsp;Gigabytes and to reposition or obtain
 the current file position for offsets beyond 2&nbsp;GB
 (<tt>fp:seek()</tt> method).
@@ -120,7 +120,7 @@ Direct3D version 10 or higher do not show this behavior anymore.
 Consider testing your application with older versions, too.<br>

 Similarly, the Borland/Delphi runtime modifies the FPU control word and
-enables FP exceptions. Of course this violates the Windows ABI, too.
+enables FP exceptions. Of course, this violates the Windows ABI, too.
 Please check the Delphi docs for the Set8087CW method.</dd>
 </dl>

@@ -130,7 +130,7 @@ Please check the Delphi docs for the Set8087CW method.</dd>
 ignored by compiled code. If your program is running in a tight loop
 and never falls back to the interpreter, the debug hook never runs and
 can't throw the "interrupted!" error.<br>
-You have to press Ctrl-C twice to get stop your program. That's similar
+You have to press Ctrl-C twice to stop your program. That's similar
 to when it's stuck running inside a C function under the Lua interpreter.</dd>
 </dl>

@@ -150,7 +150,7 @@ so it doesn't rely on the key order. Or sort the table keys, if you must.</dd>
 <dl id="sandbox">
 <dt>Q: Can Lua code be safely sandboxed?</dt>
 <dd>
-Maybe for an extremly restricted subset of Lua and if you relentlessly
+Maybe for an extremely restricted subset of Lua and if you relentlessly
 scrutinize every single interface function you offer to the untrusted code.<br>

 Although Lua provides some sandboxing functionality (<tt>setfenv()</tt>, hooks),
@@ -14,25 +14,20 @@ table.compat {
 }
 table.compat td {
  border: 1px solid #bfcfff;
-  height: 2.5em;
+  height: 1.5em;
 }
 table.compat tr.compathead td {
  font-weight: bold;
  border-bottom: 2px solid #bfcfff;
 }
-tr.compathead td.compatos {
-  vertical-align: top;
+td.compatname, td.compatver {
+  width: 10%;
 }
-table.compat td.compatcpu {
-  width: 18%;
-  border-right: 2px solid #bfcfff;
+td.compatbits {
+  width: 5%;
 }
-td.compatos {
+td.compatx {
  width: 21%;
-  vertical-align: middle;
-}
-td.compatno {
-  background-color: #d0d0d0;
 }
 </style>
 </head>
@@ -95,62 +90,204 @@ For the impatient (on POSIX systems):
 <pre class="code">
 make &amp;&amp; sudo make install
 </pre>
+
+<h2 id="req">Requirements</h2>
+<h3 id="systems">Systems</h3>
 <p>
-LuaJIT currently builds out-of-the box on most systems.
-Here's the compatibility matrix for the supported combinations of
-operating systems, CPUs and compilers:
+LuaJIT currently builds out-of-the box on most systems:
 </p>
 <table class="compat">
 <tr class="compathead">
-<td class="compatcpu">CPU / OS</td>
-<td class="compatos"><a href="#posix">Linux</a> or<br><a href="#android">Android</a></td>
-<td class="compatos"><a href="#posix">*BSD, Other</a></td>
-<td class="compatos"><a href="#posix">macOS 10.4+</a> or<br><a href="#ios">iOS 3.0+</a></td>
-<td class="compatos"><a href="#windows">Windows 7<br>or later</a></td>
+<td class="compatname">OS</td>
+<td class="compatver">Min. Version</td>
+<td class="compatx">Requirements</td>
+<td class="compatx">LuaJIT Versions</td>
 </tr>
 <tr class="odd separate">
-<td class="compatcpu">x86 (32 bit)</td>
-<td class="compatos">GCC 4.2+</td>
-<td class="compatos">GCC 4.2+</td>
-<td class="compatos">XCode 5.0+<br>Clang</td>
-<td class="compatos">MSVC<br>MinGW, Cygwin</td>
+<td class="compatname"><a href="#windows">Windows</a></td>
+<td class="compatver">7</td>
+<td class="compatx">x86 or x64, ARM64: TBA</td>
+<td class="compatx">v2.0 &ndash;</td>
 </tr>
 <tr class="even">
-<td class="compatcpu">x64 (64 bit)</td>
-<td class="compatos">GCC 4.2+</td>
-<td class="compatos">GCC 4.2+<br>ORBIS (<a href="#ps4">PS4</a>)</td>
-<td class="compatos">XCode 5.0+<br>Clang</td>
-<td class="compatos">MSVC<br>Durango (<a href="#xboxone">Xbox One</a>)</td>
+<td class="compatname"><a href="#posix">Linux</a></td>
+<td class="compatver">&nbsp;</td>
+<td class="compatx">&nbsp;</td>
+<td class="compatx">v2.0 &ndash;</td>
 </tr>
 <tr class="odd">
-<td class="compatcpu"><a href="#cross2">ARMv5+<br>ARM9E+</a></td>
-<td class="compatos">GCC 4.2+</td>
-<td class="compatos">GCC 4.2+<br>PSP2 (<a href="#psvita">PS VITA</a>)</td>
-<td class="compatos">XCode 5.0+<br>Clang</td>
-<td class="compatos compatno">&nbsp;</td>
+<td class="compatname"><a href="#posix">*BSD</a></td>
+<td class="compatver">&nbsp;</td>
+<td class="compatx">&nbsp;</td>
+<td class="compatx">v2.0 &ndash;</td>
 </tr>
 <tr class="even">
-<td class="compatcpu"><a href="#cross2">ARM64<br>ARM64be</a></td>
-<td class="compatos">GCC 4.8+</td>
-<td class="compatos compatno">&nbsp;</td>
-<td class="compatos">XCode 6.0+<br>Clang 3.5+</td>
-<td class="compatos compatno">&nbsp;</td>
+<td class="compatname"><a href="#posix">macOS (OSX)</a></td>
+<td class="compatver">10.4</td>
+<td class="compatx">&nbsp;</td>
+<td class="compatx">v2.1 &ndash;</td>
 </tr>
 <tr class="odd">
-<td class="compatcpu"><a href="#cross2">PPC</a></td>
-<td class="compatos">GCC 4.3+</td>
-<td class="compatos">GCC 4.3+<br>GCC 4.1 (<a href="#ps3">PS3</a>)</td>
-<td class="compatos compatno">&nbsp;</td>
-<td class="compatos">XEDK (<a href="#xbox360">Xbox 360</a>)</td>
+<td class="compatname"><a href="#posix">POSIX</a></td>
+<td class="compatver">&nbsp;</td>
+<td class="compatx">mmap, dlopen</td>
+<td class="compatx">v2.0 &ndash;</td>
+</tr>
+<tr class="even separate">
+<td class="compatname"><a href="#android">Android</a></td>
+<td class="compatver">4.0</td>
+<td class="compatx">Recent Android NDK</td>
+<td class="compatx">v2.0 &ndash;</td>
+</tr>
+<tr class="odd">
+<td class="compatname"><a href="#ios">iOS</a></td>
+<td class="compatver">3.0</td>
+<td class="compatx">Xcode iOS SDK</td>
+<td class="compatx">v2.1 &ndash;</td>
+</tr>
+<tr class="even separate">
+<td class="compatname"><a href="#consoles">PS3</a></td>
+<td class="compatver">&nbsp;</td>
+<td class="compatx">PS3 SDK</td>
+<td class="compatx">v2.0 &ndash; v2.1 EOL</td>
+</tr>
+<tr class="odd">
+<td class="compatname"><a href="#consoles">PS4</a></td>
+<td class="compatver">&nbsp;</td>
+<td class="compatx">PS4 SDK (ORBIS)</td>
+<td class="compatx">v2.0 &ndash;</td>
 </tr>
 <tr class="even">
-<td class="compatcpu"><a href="#cross2">MIPS32<br>MIPS64<br>MIPS64r6</a></td>
-<td class="compatos">GCC 4.3+</td>
-<td class="compatos">GCC 4.3+</td>
-<td class="compatos compatno">&nbsp;</td>
-<td class="compatos compatno">&nbsp;</td>
+<td class="compatname"><a href="#consoles">PS5</a></td>
+<td class="compatver">&nbsp;</td>
+<td class="compatx">PS5 SDK (PROSPERO)</td>
+<td class="compatx">v2.1 &ndash;</td>
+</tr>
+<tr class="odd">
+<td class="compatname"><a href="#consoles">PS Vita</a></td>
+<td class="compatver">&nbsp;</td>
+<td class="compatx">PS Vita SDK (PSP2)</td>
+<td class="compatx">v2.0 &ndash; v2.1 EOL</td>
+</tr>
+<tr class="even">
+<td class="compatname"><a href="#consoles">Xbox 360</a></td>
+<td class="compatver">&nbsp;</td>
+<td class="compatx">Xbox 360 SDK (XEDK)</td>
+<td class="compatx">v2.0 &ndash; v2.1 EOL</td>
+</tr>
+<tr class="odd">
+<td class="compatname"><a href="#consoles">Xbox One</a></td>
+<td class="compatver">&nbsp;</td>
+<td class="compatx">Xbox One SDK (DURANGO)</td>
+<td class="compatx">v2.1 &ndash;</td>
+</tr>
+<tr class="even">
+<td class="compatname"><a href="#consoles">Nintendo Switch</a></td>
+<td class="compatver">&nbsp;</td>
+<td class="compatx">NintendoSDK + NX Addon</td>
+<td class="compatx">v2.1 &ndash;</td>
 </tr>
 </table>
+<p>
+The codebase has compatibility defines for some more systems, but
+without official support.
+</p>
+<h3 id="toolchains">Toolchains</h3>
+<p>
+Building LuaJIT requires a recent toolchain based on GCC, Clang/LLVM or
+MSVC++.
+</p>
+<p>
+The Makefile-based build system requires GNU Make and supports
+cross-builds. Batch files are provided for MSVC++ builds and console
+cross-builds.
+</p>
+<h3 id="architectures">CPU Architectures</h3>
+<table class="compat">
+<tr class="compathead">
+<td class="compatname">CPU</td>
+<td class="compatbits">Bits</td>
+<td class="compatx">Requirements</td>
+<td class="compatx">Variants</td>
+<td class="compatx">LuaJIT Versions</td>
+</tr>
+<tr class="odd separate">
+<td class="compatname">x86</td>
+<td class="compatbits">32</td>
+<td class="compatx">v2.1+: SSE2</td>
+<td class="compatx">&nbsp;</td>
+<td class="compatx">v2.0 &ndash;</td>
+</tr>
+<tr class="even">
+<td class="compatname">x64</td>
+<td class="compatbits">64</td>
+<td class="compatx">&nbsp;</td>
+<td class="compatx">&nbsp;</td>
+<td class="compatx">v2.0 &ndash;</td>
+</tr>
+<tr class="odd">
+<td class="compatname">ARM</td>
+<td class="compatbits">32</td>
+<td class="compatx">ARMv5+, ARM9E+</td>
+<td class="compatx">hard-fp + soft-fp</td>
+<td class="compatx">v2.0 &ndash;</td>
+</tr>
+<tr class="even">
+<td class="compatname">ARM64</td>
+<td class="compatbits">64</td>
+<td class="compatx">&nbsp;</td>
+<td class="compatx">ARM64le + ARM64be</td>
+<td class="compatx">v2.1 &ndash;</td>
+</tr>
+<tr class="odd">
+<td class="compatname">PPC32</td>
+<td class="compatbits">32</td>
+<td class="compatx">&nbsp;</td>
+<td class="compatx">hard-fp + soft-fp</td>
+<td class="compatx">v2.0 &ndash; v2.1 EOL</td>
+</tr>
+<tr class="even">
+<td class="compatname">PPC/e500</td>
+<td class="compatbits">32</td>
+<td class="compatx">e500v2</td>
+<td class="compatx">&nbsp;</td>
+<td class="compatx">v2.0 EOL</td>
+</tr>
+<tr class="odd">
+<td class="compatname">MIPS32</td>
+<td class="compatbits">32</td>
+<td class="compatx">MIPS32r1 &ndash; r5</td>
+<td class="compatx">hard-fp + soft-fp</td>
+<td class="compatx">v2.0 &ndash;</td>
+</tr>
+<tr class="even">
+<td class="compatname">MIPS64</td>
+<td class="compatbits">64</td>
+<td class="compatx">MIPS64r1 &ndash; r5</td>
+<td class="compatx">hard-fp + soft-fp</td>
+<td class="compatx">v2.1 &ndash;</td>
+</tr>
+<tr class="odd">
+<td class="compatname">MIPS64</td>
+<td class="compatbits">64</td>
+<td class="compatx">MIPS64r6</td>
+<td class="compatx">hard-fp + soft-fp</td>
+<td class="compatx">v2.1 EOL</td>
+</tr>
+<tr class="even">
+<td class="compatname">RISC-V</td>
+<td class="compatbits">64</td>
+<td class="compatx">RVA22+</td>
+<td class="compatx">&nbsp;</td>
+<td class="compatx">TBA</td>
+</tr>
+</table>
+<p>
+There are no plans to add historic architectures or to continue support
+for end-of-life (EOL) architectures, for which no new CPUs are commonly
+available anymore. Likewise, there are no plans to support marginal
+and/or de-facto-dead architectures.
+</p>

 <h2>Configuring LuaJIT</h2>
 <p>
@@ -191,7 +328,7 @@ The recommended way to fetch the latest version is to do a pull from
 the git repository.
 </p>
 <p>
-Alternatively download the latest source package of LuaJIT (pick the .tar.gz).
+Alternatively, download the latest source package of LuaJIT (pick the .tar.gz).
 Move it to a directory of your choice, open a terminal window and change
 to this directory. Now unpack the archive and change to the newly created
 directory (replace XX.YY.ZZ with the version you downloaded):
@@ -411,7 +548,7 @@ NDKCROSS=$NDKBIN/aarch64-linux-android-
 NDKCC=$NDKBIN/aarch64-linux-android21-clang
 make CROSS=$NDKCROSS \
     STATIC_CC=$NDKCC DYNAMIC_CC="$NDKCC -fPIC" \
-     TARGET_LD=$NDKCC TARGET_AR=$NDKBIN/llvm-ar
+     TARGET_LD=$NDKCC TARGET_AR="$NDKBIN/llvm-ar rcus" \
     TARGET_STRIP=$NDKBIN/llvm-strip

 # Android/ARM, armeabi-v7a (ARMv7 VFP), Android 4.1+ (JB)
@@ -421,7 +558,7 @@ NDKCROSS=$NDKBIN/arm-linux-androideabi-
 NDKCC=$NDKBIN/armv7a-linux-androideabi16-clang
 make HOST_CC="gcc -m32" CROSS=$NDKCROSS \
     STATIC_CC=$NDKCC DYNAMIC_CC="$NDKCC -fPIC" \
-     TARGET_LD=$NDKCC TARGET_AR=$NDKBIN/llvm-ar
+     TARGET_LD=$NDKCC TARGET_AR="$NDKBIN/llvm-ar rcus" \
     TARGET_STRIP=$NDKBIN/llvm-strip
 </pre>
 <p>
@@ -446,8 +583,7 @@ make DEFAULT_CC=clang CROSS="$(dirname $ICC)/" \
 <h3 id="consoles">Cross-compiling for consoles</h3>
 <p>
 Building LuaJIT for consoles requires both a supported host compiler
-(x86 or x64) and a cross-compiler (to PPC or ARM) from the official
-console SDK.
+(x86 or x64) and a cross-compiler from the official console SDK.
 </p>
 <p>
 Due to restrictions on consoles, the JIT compiler is disabled and only
@@ -468,45 +604,58 @@ To cross-compile for <b id="ps3">PS3</b> from a Linux host (requires
 make HOST_CC="gcc -m32" CROSS=ppu-lv2-
 </pre>
 <p>
-To cross-compile for <b id="ps4">PS4</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:
+To cross-compile for the other consoles from a Windows host, open a
+"Native Tools Command Prompt for VS". You need to choose either the 32
+or the 64&nbsp;bit version of the host compiler to match the target.
+Then <tt>cd</tt> to the <tt>src</tt> directory below where you've
+unpacked the sources and run the build command given in the table:
 </p>
-<pre class="code">
-cd src
-ps4build
-</pre>
+<table class="compat">
+<tr class="compathead">
+<td class="compatname">Console</td>
+<td class="compatbits">Bits</td>
+<td class="compatx">Build Command</td>
+</tr>
+<tr class="odd separate">
+<td class="compatname"><b id="ps4">PS4</b></td>
+<td class="compatbits">64</td>
+<td class="compatx"><tt>ps4build</tt></td>
+</tr>
+<tr class="even">
+<td class="compatname"><b id="ps5">PS5</b></td>
+<td class="compatbits">64</td>
+<td class="compatx"><tt>ps5build</tt></td>
+</tr>
+<tr class="odd">
+<td class="compatname"><b id="psvita">PS Vita</b></td>
+<td class="compatbits">32</td>
+<td class="compatx"><tt>psvitabuild</tt></td>
+</tr>
+<tr class="even">
+<td class="compatname"><b id="xbox360">Xbox 360</b></td>
+<td class="compatbits">32</td>
+<td class="compatx"><tt>xedkbuild</tt></td>
+</tr>
+<tr class="odd">
+<td class="compatname"><b id="xboxone">Xbox One</b></td>
+<td class="compatbits">64</td>
+<td class="compatx"><tt>xb1build</tt></td>
+</tr>
+<tr class="even">
+<td class="compatname"><b id="nx32">Nintendo Switch NX32</b></td>
+<td class="compatbits">32</td>
+<td class="compatx"><tt>nxbuild</tt></td>
+</tr>
+<tr class="odd">
+<td class="compatname"><b id="nx64">Nintendo Switch NX64</b></td>
+<td class="compatbits">64</td>
+<td class="compatx"><tt>nxbuild</tt></td>
+</tr>
+</table>
 <p>
-To cross-compile for <b id="psvita">PS Vita</b> from a Windows host,
-open a "Visual Studio .NET Command Prompt" (32&nbsp;bit host compiler),
-<tt>cd</tt> to the directory where you've unpacked the sources and
-run the following commands:
+Please check out the comments in the corresponding <tt>*.bat</tt>
+file for more options.
 </p>
-<pre class="code">
-cd src
-psvitabuild
-</pre>
-<p>
-To cross-compile for <b id="xbox360">Xbox 360</b> from a Windows host,
-open a "Visual Studio .NET Command Prompt" (32&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
-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>
@@ -548,7 +697,7 @@ allocator from your system (no support for this on 64&nbsp;bit architectures).</
 of calling <tt>luaopen_base</tt> etc. directly.</li>
 <li>To change or extend the list of standard libraries to load, copy
 <tt>src/lib_init.c</tt> to your project and modify it accordingly.
-Make sure the <tt>jit</tt> library is loaded or the JIT compiler
+Make sure the <tt>jit</tt> library is loaded, or the JIT compiler
 will not be activated.</li>
 <li>The <tt>bit.*</tt> module for bitwise operations
 is already built-in. There's no need to statically link
@@ -567,7 +716,7 @@ in unspeakable ways.
 There should be absolutely no need to patch <tt>luaconf.h</tt> or any
 of the Makefiles. And please do not hand-pick files for your packages &mdash;
 simply use whatever <tt>make install</tt> creates. There's a reason
-for all of the files <em>and</em> directories it creates.
+for all the files <em>and</em> directories it creates.
 </p>
 <p>
 The build system uses GNU make and auto-detects most settings based on
@@ -162,7 +162,7 @@ LuaJIT is Copyright &copy; 2005-2022 Mike Pall, released under the
 <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><td>Xbox One</td></tr>
+<tr><td>PS3</td><td>PS4<br>PS5</td><td>PS Vita</td><td>Xbox 360</td><td>Xbox One</td><td>Nintendo<br>Switch</td></tr>
 </table>
 <table class="feature compiler">
 <tr><td>GCC</td><td>Clang<br>LLVM</td><td>MSVC</td></tr>
@@ -111,6 +111,7 @@ are accepted:
 <li><tt>-t type</tt> &mdash; Set output file type (default: auto-detect from output name).</li>
 <li><tt>-a arch</tt> &mdash; Override architecture for object files (default: native).</li>
 <li><tt>-o os</tt> &mdash; Override OS for object files (default: native).</li>
+<li><tt>-F name</tt> &mdash; Override filename (default: input filename).</li>
 <li><tt>-e chunk</tt> &mdash; Use chunk string as input.</li>
 <li><tt>-</tt> (a single minus sign) &mdash; Use stdin as input and/or stdout as output.</li>
 </ul>
@@ -184,7 +185,7 @@ written in Lua. They are mainly used for debugging the JIT compiler
 itself. For a description of their options and output format, please
 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
+distribution or installed under the <tt>jit</tt> directory. By default,
 this is <tt>/usr/local/share/luajit-XX.YY.ZZ>/jit</tt> on POSIX
 systems (replace XX.YY.ZZ by the installed version).
 </p>
@@ -216,7 +217,7 @@ to a specific value.
 You can either use this option multiple times (like <tt>-Ocse
 -O-dce -Ohotloop=10</tt>) or separate several settings with a comma
 (like <tt>-O+cse,-dce,hotloop=10</tt>). The settings are applied from
-left to right and later settings override earlier ones. You can freely
+left to right, and later settings override earlier ones. You can freely
 mix the three forms, but note that setting an optimization level
 overrides all earlier flags.
 </p>
@@ -83,7 +83,7 @@ Known incompatibilities and issues in LuaJIT&nbsp;2.0:
 <ul>
 <li>
 There are some differences in <b>implementation-defined</b> behavior.
-These either have a good reason, are arbitrary design choices
+These either have a good reason, are arbitrary design choices,
 or are due to quirks in the VM. The latter cases may get fixed if a
 demonstrable need is shown.
 </li>
@@ -158,10 +158,10 @@ void dasm_setup(Dst_DECL, const void *actionlist)
 #ifdef DASM_CHECKS
 #define CK(x, st) \
  do { if (!(x)) { \
-    D->status = DASM_S_##st|(p-D->actionlist-1); return; } } while (0)
+    D->status = DASM_S_##st|(int)(p-D->actionlist-1); return; } } while (0)
 #define CKPL(kind, st) \
  do { if ((size_t)((char *)pl-(char *)D->kind##labels) >= D->kind##size) { \
-    D->status = DASM_S_RANGE_##st|(p-D->actionlist-1); return; } } while (0)
+    D->status = DASM_S_RANGE_##st|(int)(p-D->actionlist-1); return; } } while (0)
 #else
 #define CK(x, st)	((void)0)
 #define CKPL(kind, st)	((void)0)
@@ -190,7 +190,9 @@ static int dasm_imm13(int lo, int hi)
  unsigned long long n = (((unsigned long long)hi) << 32) | (unsigned int)lo;
  unsigned long long m = 1ULL, a, b, c;
  if (n & 1) { n = ~n; inv = 1; }
-  a = n & -n; b = (n+a)&-(n+a); c = (n+a-b)&-(n+a-b);
+  a = n & (unsigned long long)-(long long)n;
+  b = (n+a)&(unsigned long long)-(long long)(n+a);
+  c = (n+a-b)&(unsigned long long)-(long long)(n+a-b);
  xa = dasm_ffs(a); xb = dasm_ffs(b);
  if (c) {
    w = dasm_ffs(c) - xa;
@@ -406,7 +408,7 @@ int dasm_link(Dst_DECL, size_t *szp)

 #ifdef DASM_CHECKS
 #define CK(x, st) \
-  do { if (!(x)) return DASM_S_##st|(p-D->actionlist-1); } while (0)
+  do { if (!(x)) return DASM_S_##st|(int)(p-D->actionlist-1); } while (0)
 #else
 #define CK(x, st)	((void)0)
 #endif
@@ -438,7 +440,7 @@ int dasm_encode(Dst_DECL, void *buffer)
 	  n = DASM_EXTERN(Dst, (unsigned char *)cp, (ins&2047), !(ins&2048));
 	  goto patchrel;
 	case DASM_ALIGN:
-	  ins &= 255; while ((((char *)cp - base) & ins)) *cp++ = 0xe1a00000;
+	  ins &= 255; while ((((char *)cp - base) & ins)) *cp++ = 0xd503201f;
 	  break;
 	case DASM_REL_LG:
 	  if (n < 0) {
@@ -554,7 +556,7 @@ int dasm_checkstep(Dst_DECL, int secmatch)
  }
  if (D->status == DASM_S_OK && secmatch >= 0 &&
      D->section != &D->sections[secmatch])
-    D->status = DASM_S_MATCH_SEC|(D->section-D->sections);
+    D->status = DASM_S_MATCH_SEC|(int)(D->section-D->sections);
  return D->status;
 }
 #endif
@@ -248,7 +248,7 @@ local map_cond = {

 local parse_reg_type

-local function parse_reg(expr, shift)
+local function parse_reg(expr, shift, no_vreg)
  if not expr then werror("expected register name") end
  local tname, ovreg = match(expr, "^([%w_]+):(@?%l%d+)$")
  if not tname then
@@ -281,7 +281,7 @@ local function parse_reg(expr, shift)
    elseif parse_reg_type ~= vrt then
      werror("register size mismatch")
    end
-    if shift then waction("VREG", shift, vreg) end
+    if not no_vreg then waction("VREG", shift, vreg) end
    return 0
  end
  werror("bad register name `"..expr.."'")
@@ -638,7 +638,7 @@ local function alias_bfx(p)
 end

 local function alias_bfiz(p)
-  parse_reg(p[1], 0)
+  parse_reg(p[1], 0, true)
  if parse_reg_type == "w" then
    p[3] = "#(32-("..p[3]:sub(2).."))%32"
    p[4] = "#("..p[4]:sub(2)..")-1"
@@ -649,7 +649,7 @@ local function alias_bfiz(p)
 end

 local alias_lslimm = op_alias("ubfm_4", function(p)
-  parse_reg(p[1], 0)
+  parse_reg(p[1], 0, true)
  local sh = p[3]:sub(2)
  if parse_reg_type == "w" then
    p[3] = "#(32-("..sh.."))%32"
@@ -155,10 +155,10 @@ void dasm_setup(Dst_DECL, const void *actionlist)
 #ifdef DASM_CHECKS
 #define CK(x, st) \
  do { if (!(x)) { \
-    D->status = DASM_S_##st|(p-D->actionlist-1); return; } } while (0)
+    D->status = DASM_S_##st|(int)(p-D->actionlist-1); return; } } while (0)
 #define CKPL(kind, st) \
  do { if ((size_t)((char *)pl-(char *)D->kind##labels) >= D->kind##size) { \
-    D->status = DASM_S_RANGE_##st|(p-D->actionlist-1); return; } } while (0)
+    D->status = DASM_S_RANGE_##st|(int)(p-D->actionlist-1); return; } } while (0)
 #else
 #define CK(x, st)	((void)0)
 #define CKPL(kind, st)	((void)0)
@@ -314,7 +314,7 @@ int dasm_link(Dst_DECL, size_t *szp)

 #ifdef DASM_CHECKS
 #define CK(x, st) \
-  do { if (!(x)) return DASM_S_##st|(p-D->actionlist-1); } while (0)
+  do { if (!(x)) return DASM_S_##st|(int)(p-D->actionlist-1); } while (0)
 #else
 #define CK(x, st)	((void)0)
 #endif
@@ -417,7 +417,7 @@ int dasm_checkstep(Dst_DECL, int secmatch)
  }
  if (D->status == DASM_S_OK && secmatch >= 0 &&
      D->section != &D->sections[secmatch])
-    D->status = DASM_S_MATCH_SEC|(D->section-D->sections);
+    D->status = DASM_S_MATCH_SEC|(int)(D->section-D->sections);
  return D->status;
 }
 #endif
@@ -18,8 +18,10 @@
 #include "lj_obj.h"
 #include "lj_gc.h"
 #include "lj_bc.h"
+#if LJ_HASJIT
 #include "lj_ir.h"
 #include "lj_ircall.h"
+#endif
 #include "lj_frame.h"
 #include "lj_dispatch.h"
 #if LJ_HASFFI
@@ -250,6 +252,7 @@ BCDEF(BCNAME)
  NULL
 };

+#if LJ_HASJIT
 const char *const ir_names[] = {
 #define IRNAME(name, m, m1, m2)	#name,
 IRDEF(IRNAME)
@@ -290,7 +293,9 @@ static const char *const trace_errors[] = {
 #include "lj_traceerr.h"
  NULL
 };
+#endif

+#if LJ_HASJIT
 static const char *lower(char *buf, const char *s)
 {
  char *p = buf;
@@ -301,6 +306,7 @@ static const char *lower(char *buf, const char *s)
  *p = '\0';
  return buf;
 }
+#endif

 /* Emit C source code for bytecode-related definitions. */
 static void emit_bcdef(BuildCtx *ctx)
@@ -318,7 +324,9 @@ static void emit_bcdef(BuildCtx *ctx)
 /* Emit VM definitions as Lua code for debug modules. */
 static void emit_vmdef(BuildCtx *ctx)
 {
+#if LJ_HASJIT
  char buf[80];
+#endif
  int i;
  fprintf(ctx->fp, "-- This is a generated file. DO NOT EDIT!\n\n");
  fprintf(ctx->fp, "return {\n\n");
@@ -327,6 +335,7 @@ static void emit_vmdef(BuildCtx *ctx)
  for (i = 0; bc_names[i]; i++) fprintf(ctx->fp, "%-6s", bc_names[i]);
  fprintf(ctx->fp, "\",\n\n");

+#if LJ_HASJIT
  fprintf(ctx->fp, "irnames = \"");
  for (i = 0; ir_names[i]; i++) fprintf(ctx->fp, "%-6s", ir_names[i]);
  fprintf(ctx->fp, "\",\n\n");
@@ -355,6 +364,7 @@ static void emit_vmdef(BuildCtx *ctx)
  for (i = 0; trace_errors[i]; i++)
    fprintf(ctx->fp, "\"%s\",\n", trace_errors[i]);
  fprintf(ctx->fp, "},\n\n");
+#endif
 }

 /* -- Argument parsing ---------------------------------------------------- */
@@ -5,6 +5,7 @@

 #include "buildvm.h"
 #include "lj_obj.h"
+#if LJ_HASJIT
 #include "lj_ir.h"

 /* Context for the folding hash table generator. */
@@ -226,4 +227,10 @@ void emit_fold(BuildCtx *ctx)

  makehash(ctx);
 }
+#else
+void emit_fold(BuildCtx *ctx)
+{
+  UNUSED(ctx);
+}
+#endif

@@ -4,42 +4,67 @@ static const int libbc_endian = 0;

 static const uint8_t libbc_code[] = {
 #if LJ_FR2
-0,1,2,0,0,1,2,24,1,0,0,76,1,2,0,241,135,158,166,3,220,203,178,130,4,0,1,2,0,
-0,1,2,24,1,0,0,76,1,2,0,243,244,148,165,20,198,190,199,252,3,0,1,2,0,0,0,3,
-16,0,5,0,21,1,0,0,76,1,2,0,0,2,10,0,0,0,15,16,0,12,0,16,1,9,0,41,2,1,0,21,3,
-0,0,41,4,1,0,77,2,8,128,18,6,1,0,18,8,5,0,59,9,5,0,66,6,3,2,10,6,0,0,88,7,1,
-128,76,6,2,0,79,2,248,127,75,0,1,0,0,2,11,0,0,0,16,16,0,12,0,16,1,9,0,43,2,
-0,0,18,3,0,0,41,4,0,0,88,5,7,128,18,7,1,0,18,9,5,0,18,10,6,0,66,7,3,2,10,7,
-0,0,88,8,1,128,76,7,2,0,70,5,3,3,82,5,247,127,75,0,1,0,0,1,2,0,0,0,3,16,0,12,
-0,21,1,0,0,76,1,2,0,0,2,10,0,0,2,30,16,0,12,0,21,2,0,0,11,1,0,0,88,3,7,128,
-8,2,0,0,88,3,23,128,59,3,2,0,43,4,0,0,64,4,2,0,76,3,2,0,88,3,18,128,16,1,14,
-0,41,3,1,0,3,3,1,0,88,3,14,128,3,1,2,0,88,3,12,128,59,3,1,0,22,4,1,1,18,5,2,
-0,41,6,1,0,77,4,4,128,23,8,1,7,59,9,7,0,64,9,8,0,79,4,252,127,43,4,0,0,64,4,
-2,0,76,3,2,0,75,0,1,0,0,2,0,5,12,0,0,0,35,16,0,12,0,16,1,14,0,16,2,14,0,16,
-3,14,0,11,4,0,0,88,5,1,128,18,4,0,0,16,4,12,0,3,1,2,0,88,5,24,128,33,5,1,3,
-0,2,3,0,88,6,4,128,2,3,1,0,88,6,2,128,4,4,0,0,88,6,9,128,18,6,1,0,18,7,2,0,
-41,8,1,0,77,6,4,128,32,10,5,9,59,11,9,0,64,11,10,4,79,6,252,127,88,6,8,128,
-18,6,2,0,18,7,1,0,41,8,255,255,77,6,4,128,32,10,5,9,59,11,9,0,64,11,10,4,79,
-6,252,127,76,4,2,0,0
+/* math.deg */ 0,1,2,0,0,1,2,BC_MULVN,1,0,0,BC_RET1,1,2,0,241,135,158,166,3,
+220,203,178,130,4,
+/* math.rad */ 0,1,2,0,0,1,2,BC_MULVN,1,0,0,BC_RET1,1,2,0,243,244,148,165,20,
+198,190,199,252,3,
+/* string.len */ 0,1,2,0,0,0,3,BC_ISTYPE,0,5,0,BC_LEN,1,0,0,BC_RET1,1,2,0,
+/* table.foreachi */ 0,2,10,0,0,0,15,BC_ISTYPE,0,12,0,BC_ISTYPE,1,9,0,
+BC_KSHORT,2,1,0,BC_LEN,3,0,0,BC_KSHORT,4,1,0,BC_FORI,2,8,128,BC_MOV,6,1,0,
+BC_MOV,8,5,0,BC_TGETR,9,5,0,BC_CALL,6,3,2,BC_ISEQP,6,0,0,BC_JMP,7,1,128,
+BC_RET1,6,2,0,BC_FORL,2,248,127,BC_RET0,0,1,0,
+/* table.foreach */ 0,2,11,0,0,1,16,BC_ISTYPE,0,12,0,BC_ISTYPE,1,9,0,BC_KPRI,
+2,0,0,BC_MOV,3,0,0,BC_KNUM,4,0,0,BC_JMP,5,7,128,BC_MOV,7,1,0,BC_MOV,9,5,0,
+BC_MOV,10,6,0,BC_CALL,7,3,2,BC_ISEQP,7,0,0,BC_JMP,8,1,128,BC_RET1,7,2,0,
+BC_ITERN,5,3,3,BC_ITERL,5,247,127,BC_RET0,0,1,0,1,255,255,249,255,15,
+/* table.getn */ 0,1,2,0,0,0,3,BC_ISTYPE,0,12,0,BC_LEN,1,0,0,BC_RET1,1,2,0,
+/* table.remove */ 0,2,10,0,0,2,30,BC_ISTYPE,0,12,0,BC_LEN,2,0,0,BC_ISNEP,1,0,
+0,BC_JMP,3,7,128,BC_ISEQN,2,0,0,BC_JMP,3,23,128,BC_TGETR,3,2,0,BC_KPRI,4,0,0,
+BC_TSETR,4,2,0,BC_RET1,3,2,0,BC_JMP,3,18,128,BC_ISTYPE,1,14,0,BC_KSHORT,3,1,0,
+BC_ISGT,3,1,0,BC_JMP,3,14,128,BC_ISGT,1,2,0,BC_JMP,3,12,128,BC_TGETR,3,1,0,
+BC_ADDVN,4,1,1,BC_MOV,5,2,0,BC_KSHORT,6,1,0,BC_FORI,4,4,128,BC_SUBVN,8,1,7,
+BC_TGETR,9,7,0,BC_TSETR,9,8,0,BC_FORL,4,252,127,BC_KPRI,4,0,0,BC_TSETR,4,2,0,
+BC_RET1,3,2,0,BC_RET0,0,1,0,0,2,
+/* table.move */ 0,5,12,0,0,0,35,BC_ISTYPE,0,12,0,BC_ISTYPE,1,14,0,BC_ISTYPE,
+2,14,0,BC_ISTYPE,3,14,0,BC_ISNEP,4,0,0,BC_JMP,5,1,128,BC_MOV,4,0,0,BC_ISTYPE,
+4,12,0,BC_ISGT,1,2,0,BC_JMP,5,24,128,BC_SUBVV,5,1,3,BC_ISLT,2,3,0,BC_JMP,6,4,
+128,BC_ISLE,3,1,0,BC_JMP,6,2,128,BC_ISEQV,4,0,0,BC_JMP,6,9,128,BC_MOV,6,1,0,
+BC_MOV,7,2,0,BC_KSHORT,8,1,0,BC_FORI,6,4,128,BC_ADDVV,10,5,9,BC_TGETR,11,9,0,
+BC_TSETR,11,10,4,BC_FORL,6,252,127,BC_JMP,6,8,128,BC_MOV,6,2,0,BC_MOV,7,1,0,
+BC_KSHORT,8,255,255,BC_FORI,6,4,128,BC_ADDVV,10,5,9,BC_TGETR,11,9,0,BC_TSETR,
+11,10,4,BC_FORL,6,252,127,BC_RET1,4,2,0,
 #else
-0,1,2,0,0,1,2,24,1,0,0,76,1,2,0,241,135,158,166,3,220,203,178,130,4,0,1,2,0,
-0,1,2,24,1,0,0,76,1,2,0,243,244,148,165,20,198,190,199,252,3,0,1,2,0,0,0,3,
-16,0,5,0,21,1,0,0,76,1,2,0,0,2,9,0,0,0,15,16,0,12,0,16,1,9,0,41,2,1,0,21,3,
-0,0,41,4,1,0,77,2,8,128,18,6,1,0,18,7,5,0,59,8,5,0,66,6,3,2,10,6,0,0,88,7,1,
-128,76,6,2,0,79,2,248,127,75,0,1,0,0,2,10,0,0,0,16,16,0,12,0,16,1,9,0,43,2,
-0,0,18,3,0,0,41,4,0,0,88,5,7,128,18,7,1,0,18,8,5,0,18,9,6,0,66,7,3,2,10,7,0,
-0,88,8,1,128,76,7,2,0,70,5,3,3,82,5,247,127,75,0,1,0,0,1,2,0,0,0,3,16,0,12,
-0,21,1,0,0,76,1,2,0,0,2,10,0,0,2,30,16,0,12,0,21,2,0,0,11,1,0,0,88,3,7,128,
-8,2,0,0,88,3,23,128,59,3,2,0,43,4,0,0,64,4,2,0,76,3,2,0,88,3,18,128,16,1,14,
-0,41,3,1,0,3,3,1,0,88,3,14,128,3,1,2,0,88,3,12,128,59,3,1,0,22,4,1,1,18,5,2,
-0,41,6,1,0,77,4,4,128,23,8,1,7,59,9,7,0,64,9,8,0,79,4,252,127,43,4,0,0,64,4,
-2,0,76,3,2,0,75,0,1,0,0,2,0,5,12,0,0,0,35,16,0,12,0,16,1,14,0,16,2,14,0,16,
-3,14,0,11,4,0,0,88,5,1,128,18,4,0,0,16,4,12,0,3,1,2,0,88,5,24,128,33,5,1,3,
-0,2,3,0,88,6,4,128,2,3,1,0,88,6,2,128,4,4,0,0,88,6,9,128,18,6,1,0,18,7,2,0,
-41,8,1,0,77,6,4,128,32,10,5,9,59,11,9,0,64,11,10,4,79,6,252,127,88,6,8,128,
-18,6,2,0,18,7,1,0,41,8,255,255,77,6,4,128,32,10,5,9,59,11,9,0,64,11,10,4,79,
-6,252,127,76,4,2,0,0
+/* math.deg */ 0,1,2,0,0,1,2,BC_MULVN,1,0,0,BC_RET1,1,2,0,241,135,158,166,3,
+220,203,178,130,4,
+/* math.rad */ 0,1,2,0,0,1,2,BC_MULVN,1,0,0,BC_RET1,1,2,0,243,244,148,165,20,
+198,190,199,252,3,
+/* string.len */ 0,1,2,0,0,0,3,BC_ISTYPE,0,5,0,BC_LEN,1,0,0,BC_RET1,1,2,0,
+/* table.foreachi */ 0,2,9,0,0,0,15,BC_ISTYPE,0,12,0,BC_ISTYPE,1,9,0,
+BC_KSHORT,2,1,0,BC_LEN,3,0,0,BC_KSHORT,4,1,0,BC_FORI,2,8,128,BC_MOV,6,1,0,
+BC_MOV,7,5,0,BC_TGETR,8,5,0,BC_CALL,6,3,2,BC_ISEQP,6,0,0,BC_JMP,7,1,128,
+BC_RET1,6,2,0,BC_FORL,2,248,127,BC_RET0,0,1,0,
+/* table.foreach */ 0,2,10,0,0,1,16,BC_ISTYPE,0,12,0,BC_ISTYPE,1,9,0,BC_KPRI,
+2,0,0,BC_MOV,3,0,0,BC_KNUM,4,0,0,BC_JMP,5,7,128,BC_MOV,7,1,0,BC_MOV,8,5,0,
+BC_MOV,9,6,0,BC_CALL,7,3,2,BC_ISEQP,7,0,0,BC_JMP,8,1,128,BC_RET1,7,2,0,
+BC_ITERN,5,3,3,BC_ITERL,5,247,127,BC_RET0,0,1,0,1,255,255,249,255,15,
+/* table.getn */ 0,1,2,0,0,0,3,BC_ISTYPE,0,12,0,BC_LEN,1,0,0,BC_RET1,1,2,0,
+/* table.remove */ 0,2,10,0,0,2,30,BC_ISTYPE,0,12,0,BC_LEN,2,0,0,BC_ISNEP,1,0,
+0,BC_JMP,3,7,128,BC_ISEQN,2,0,0,BC_JMP,3,23,128,BC_TGETR,3,2,0,BC_KPRI,4,0,0,
+BC_TSETR,4,2,0,BC_RET1,3,2,0,BC_JMP,3,18,128,BC_ISTYPE,1,14,0,BC_KSHORT,3,1,0,
+BC_ISGT,3,1,0,BC_JMP,3,14,128,BC_ISGT,1,2,0,BC_JMP,3,12,128,BC_TGETR,3,1,0,
+BC_ADDVN,4,1,1,BC_MOV,5,2,0,BC_KSHORT,6,1,0,BC_FORI,4,4,128,BC_SUBVN,8,1,7,
+BC_TGETR,9,7,0,BC_TSETR,9,8,0,BC_FORL,4,252,127,BC_KPRI,4,0,0,BC_TSETR,4,2,0,
+BC_RET1,3,2,0,BC_RET0,0,1,0,0,2,
+/* table.move */ 0,5,12,0,0,0,35,BC_ISTYPE,0,12,0,BC_ISTYPE,1,14,0,BC_ISTYPE,
+2,14,0,BC_ISTYPE,3,14,0,BC_ISNEP,4,0,0,BC_JMP,5,1,128,BC_MOV,4,0,0,BC_ISTYPE,
+4,12,0,BC_ISGT,1,2,0,BC_JMP,5,24,128,BC_SUBVV,5,1,3,BC_ISLT,2,3,0,BC_JMP,6,4,
+128,BC_ISLE,3,1,0,BC_JMP,6,2,128,BC_ISEQV,4,0,0,BC_JMP,6,9,128,BC_MOV,6,1,0,
+BC_MOV,7,2,0,BC_KSHORT,8,1,0,BC_FORI,6,4,128,BC_ADDVV,10,5,9,BC_TGETR,11,9,0,
+BC_TSETR,11,10,4,BC_FORL,6,252,127,BC_JMP,6,8,128,BC_MOV,6,2,0,BC_MOV,7,1,0,
+BC_KSHORT,8,255,255,BC_FORI,6,4,128,BC_ADDVV,10,5,9,BC_TGETR,11,9,0,BC_TSETR,
+11,10,4,BC_FORL,6,252,127,BC_RET1,4,2,0,
 #endif
+0
 };

 static const struct { const char *name; int ofs; } libbc_map[] = {
@@ -48,9 +73,9 @@ static const struct { const char *name; int ofs; } libbc_map[] = {
 {"string_len",50},
 {"table_foreachi",69},
 {"table_foreach",136},
-{"table_getn",207},
-{"table_remove",226},
-{"table_move",355},
-{NULL,502}
+{"table_getn",213},
+{"table_remove",232},
+{"table_move",361},
+{NULL,508}
 };

@@ -55,7 +55,7 @@ local function transform_lua(code)
  end)
  code = string.gsub(code, "PAIRS%((.-)%)", function(var)
    fixup.PAIRS = true
-    return format("nil, %s, 0", var)
+    return format("nil, %s, 0x4dp80", var)
  end)
  return "return "..code, fixup
 end
@@ -79,9 +79,11 @@ local name2itype = {
  str = 5, func = 9, tab = 12, int = 14, num = 15
 }

-local BC = {}
+local BC, BCN = {}, {}
 for i=0,#bcnames/6-1 do
-  BC[string.gsub(string.sub(bcnames, i*6+1, i*6+6), " ", "")] = i
+  local name = bcnames:sub(i*6+1, i*6+6):gsub(" ", "")
+  BC[name] = i
+  BCN[i] = name
 end
 local xop, xra = isbe and 3 or 0, isbe and 2 or 1
 local xrc, xrb = isbe and 1 or 2, isbe and 0 or 3
@@ -96,6 +98,7 @@ local function fixup_dump(dump, fixup)
  p = read_uleb128(p)
  p = read_uleb128(p)
  p, sizebc = read_uleb128(p)
+  local startbc = tonumber(p - start)
  local rawtab = {}
  for i=0,sizebc-1 do
    local op = p[xop]
@@ -129,7 +132,10 @@ local function fixup_dump(dump, fixup)
    end
    p = p + 4
  end
-  return ffi.string(start, n)
+  local ndump = ffi.string(start, n)
+  -- Fixup hi-part of 0x4dp80 to LJ_KEYINDEX.
+  ndump = ndump:gsub("\x80\x80\xcd\xaa\x04", "\xff\xff\xf9\xff\x0f")
+  return { dump = ndump, startbc = startbc, sizebc = sizebc }
 end

 local function find_defs(src)
@@ -149,24 +155,46 @@ local function gen_header(defs)
  local function w(x) t[#t+1] = x end
  w("/* This is a generated file. DO NOT EDIT! */\n\n")
  w("static const int libbc_endian = ") w(isbe and 1 or 0) w(";\n\n")
-  local s = ""
-  for _,name in ipairs(defs) do
-    s = s .. defs[name]
+  local s, sb = "", ""
+  for i,name in ipairs(defs) do
+    local d = defs[name]
+    s = s .. d.dump
+    sb = sb .. string.char(i) .. ("\0"):rep(d.startbc - 1)
+	    .. (isbe and "\0\0\0\255" or "\255\0\0\0"):rep(d.sizebc)
+	    .. ("\0"):rep(#d.dump - d.startbc - d.sizebc*4)
  end
  w("static const uint8_t libbc_code[] = {\n")
  local n = 0
  for i=1,#s do
    local x = string.byte(s, i)
-    w(x); w(",")
-    n = n + (x < 10 and 2 or (x < 100 and 3 or 4))
-    if n >= 75 then n = 0; w("\n") end
+    local xb = string.byte(sb, i)
+    if xb == 255 then
+      local name = BCN[x]
+      local m = #name + 4
+      if n + m > 78 then n = 0; w("\n") end
+      n = n + m
+      w("BC_"); w(name)
+    else
+      local m = x < 10 and 2 or (x < 100 and 3 or 4)
+      if xb == 0 then
+	if n + m > 78 then n = 0; w("\n") end
+      else
+	local name = defs[xb]:gsub("_", ".")
+	if n ~= 0 then w("\n") end
+	w("/* "); w(name); w(" */ ")
+	n = #name + 7
      end
-  w("0\n};\n\n")
+      n = n + m
+      w(x)
+    end
+    w(",")
+  end
+  w("\n0\n};\n\n")
  w("static const struct { const char *name; int ofs; } libbc_map[] = {\n")
  local m = 0
  for _,name in ipairs(defs) do
    w('{"'); w(name); w('",'); w(m) w('},\n')
-    m = m + #defs[name]
+    m = m + #defs[name].dump
  end
  w("{NULL,"); w(m); w("}\n};\n\n")
  return table.concat(t)
@@ -327,6 +327,12 @@ local function rename_tokens2(src)
  return gsub(src, "ZY([%w_]+)", "union %1")
 end

+local function fix_bugs_and_warnings(src)
+ src = gsub(src, "(luaD_checkstack%(L,p%->maxstacksize)%)", "%1+p->numparams)")
+ src = gsub(src, "if%(sep==%-1%)(return'%[';)\nelse (luaX_lexerror%b();)", "if (sep!=-1)%2\n%1")
+ return gsub(src, "(default:{\nNode%*n=mainposition)", "/*fallthrough*/\n%1")
+end
+
 local function func_gather(src)
  local nodes, list = {}, {}
  local pos, len = 1, #src
@@ -425,5 +431,6 @@ src = rename_tokens1(src)
 src = func_collect(src)
 src = rename_tokens2(src)
 src = restore_strings(src)
+src = fix_bugs_and_warnings(src)
 src = merge_header(src, license)
 io.write(src)
@@ -1639,6 +1639,7 @@ lua_number2int(k,n);
 if(luai_numeq(cast_num(k),nvalue(key)))
 return luaH_getnum(t,k);
 }
+/*fallthrough*/
 default:{
 Node*n=mainposition(t,key);
 do{
@@ -2905,8 +2906,8 @@ if(sep>=0){
 read_long_string(ls,seminfo,sep);
 return TK_STRING;
 }
-else if(sep==-1)return'[';
-else luaX_lexerror(ls,"invalid long string delimiter",TK_STRING);
+else if (sep!=-1)luaX_lexerror(ls,"invalid long string delimiter",TK_STRING);
+return'[';
 }
 case'=':{
 next(ls);
@@ -33,6 +33,7 @@ Save LuaJIT bytecode: luajit -b[options] input output
  -t type   Set output file type (default: auto-detect from output name).
  -a arch   Override architecture for object files (default: native).
  -o os     Override OS for object files (default: native).
+  -F name   Override filename (default: input filename).
  -e chunk  Use chunk string as input.
  --        Stop handling options.
  -         Use stdin as input and/or stdout as output.
@@ -49,11 +50,23 @@ local function check(ok, ...)
  os.exit(1)
 end

-local function readfile(input)
+local function readfile(ctx, input)
  if type(input) == "function" then return input end
+  if ctx.filename then
+    local data
+    if input == "-" then
+      data = io.stdin:read("*a")
+    else
+      local fp = assert(io.open(input, "rb"))
+      data = assert(fp:read("*a"))
+      assert(fp:close())
+    end
+    return check(load(data, ctx.filename))
+  else
    if input == "-" then input = nil end
    return check(loadfile(input))
  end
+end

 local function savefile(name, mode)
  if name == "-" then return io.stdout end
@@ -456,18 +469,18 @@ typedef struct {
  uint32_t value;
 } mach_nlist;
 typedef struct {
-  uint32_t strx;
+  int32_t strx;
  uint8_t type, sect;
  uint16_t desc;
  uint64_t value;
 } mach_nlist_64;
 typedef struct
 {
-  uint32_t magic, nfat_arch;
+  int32_t magic, nfat_arch;
 } mach_fat_header;
 typedef struct
 {
-  uint32_t cputype, cpusubtype, offset, size, align;
+  int32_t cputype, cpusubtype, offset, size, align;
 } mach_fat_arch;
 typedef struct {
  struct {
@@ -501,6 +514,18 @@ typedef struct {
  mach_nlist sym_entry;
  uint8_t space[4096];
 } mach_fat_obj;
+typedef struct {
+  mach_fat_header fat;
+  mach_fat_arch fat_arch[2];
+  struct {
+    mach_header_64 hdr;
+    mach_segment_command_64 seg;
+    mach_section_64 sec;
+    mach_symtab_command sym;
+  } arch[2];
+  mach_nlist_64 sym_entry;
+  uint8_t space[4096];
+} mach_fat_obj_64;
 ]]
  local symname = '_'..LJBC_PREFIX..ctx.modname
  local isfat, is64, align, mobj = false, false, 4, "mach_obj"
@@ -509,7 +534,7 @@ typedef struct {
  elseif ctx.arch == "arm" then
    isfat, mobj = true, "mach_fat_obj"
  elseif ctx.arch == "arm64" then
-    is64, align, isfat, mobj = true, 8, true, "mach_fat_obj"
+    is64, align, isfat, mobj = true, 8, true, "mach_fat_obj_64"
  else
    check(ctx.arch == "x86", "unsupported architecture for OSX")
  end
@@ -592,13 +617,13 @@ end

 ------------------------------------------------------------------------------

-local function bclist(input, output)
-  local f = readfile(input)
+local function bclist(ctx, input, output)
+  local f = readfile(ctx, input)
  require("jit.bc").dump(f, savefile(output, "w"), true)
 end

 local function bcsave(ctx, input, output)
-  local f = readfile(input)
+  local f = readfile(ctx, input)
  local s = string.dump(f, ctx.strip)
  local t = ctx.type
  if not t then
@@ -651,6 +676,8 @@ local function docmd(...)
 	    ctx.arch = checkarg(tremove(arg, n), map_arch, "architecture")
 	  elseif opt == "o" then
 	    ctx.os = checkarg(tremove(arg, n), map_os, "OS name")
+	  elseif opt == "F" then
+	    ctx.filename = "@"..tremove(arg, n)
 	  else
 	    usage()
 	  end
@@ -662,7 +689,7 @@ local function docmd(...)
  end
  if list then
    if #arg == 0 or #arg > 2 then usage() end
-    bclist(arg[1], arg[2] or "-")
+    bclist(ctx, arg[1], arg[2] or "-")
  else
    if #arg ~= 2 then usage() end
    bcsave(ctx, arg[1], arg[2])
@@ -76,6 +76,8 @@ LJLIB_CF(buffer_method_skip)		LJLIB_REC(.)
  MSize len = sbufxlen(sbx);
  if (n < len) {
    sbx->r += n;
+  } else if (sbufiscow(sbx)) {
+    sbx->r = sbx->w;
  } else {
    sbx->r = sbx->w = sbx->b;
  }
@@ -173,7 +175,7 @@ LJLIB_CF(buffer_method_get)		LJLIB_REC(.)
    setstrV(L, o, lj_str_new(L, sbx->r, n));
    sbx->r += n;
  }
-  if (sbx->r == sbx->w) sbx->r = sbx->w = sbx->b;
+  if (sbx->r == sbx->w && !sbufiscow(sbx)) sbx->r = sbx->w = sbx->b;
  lj_gc_check(L);
  return narg-1;
 }
@@ -321,6 +323,7 @@ LJLIB_CF(buffer_new)
  setgcref(sbx->dict_str, obj2gco(dict_str));
  setgcref(sbx->dict_mt, obj2gco(dict_mt));
  if (sz > 0) lj_buf_need2((SBuf *)sbx, sz);
+  lj_gc_check(L);
  return 1;
 }

@@ -337,6 +340,7 @@ LJLIB_CF(buffer_decode)			LJLIB_REC(.)
  GCstr *str = lj_lib_checkstrx(L, 1);
  setnilV(L->top++);
  lj_serialize_decode(L, L->top-1, str);
+  lj_gc_check(L);
  return 1;
 }

@@ -639,7 +639,7 @@ LJLIB_CF(ffi_alignof)	LJLIB_REC(ffi_xof FF_ffi_alignof)
  CTState *cts = ctype_cts(L);
  CTypeID id = ffi_checkctype(L, cts, NULL);
  CTSize sz = 0;
-  CTInfo info = lj_ctype_info(cts, id, &sz);
+  CTInfo info = lj_ctype_info_raw(cts, id, &sz);
  setintV(L->top-1, 1 << ctype_align(info));
  return 1;
 }
@@ -770,7 +770,7 @@ LJLIB_CF(ffi_metatype)
  CTypeID id = ffi_checkctype(L, cts, NULL);
  GCtab *mt = lj_lib_checktab(L, 2);
  GCtab *t = cts->miscmap;
-  CType *ct = ctype_get(cts, id);  /* Only allow raw types. */
+  CType *ct = ctype_raw(cts, id);
  TValue *tv;
  GCcdata *cd;
  if (!(ctype_isstruct(ct->info) || ctype_iscomplex(ct->info) ||
@@ -439,7 +439,7 @@ LJLIB_CF(io_popen)
 LJLIB_CF(io_tmpfile)
 {
  IOFileUD *iof = io_file_new(L);
-#if LJ_TARGET_PS3 || LJ_TARGET_PS4 || LJ_TARGET_PSVITA
+#if LJ_TARGET_PS3 || LJ_TARGET_PS4 || LJ_TARGET_PS5 || LJ_TARGET_PSVITA || LJ_TARGET_NX
  iof->fp = NULL; errno = ENOSYS;
 #else
  iof->fp = tmpfile();
@@ -76,7 +76,7 @@ LJLIB_CF(os_rename)

 LJLIB_CF(os_tmpname)
 {
-#if LJ_TARGET_PS3 || LJ_TARGET_PS4 || LJ_TARGET_PSVITA
+#if LJ_TARGET_PS3 || LJ_TARGET_PS4 || LJ_TARGET_PS5 || LJ_TARGET_PSVITA || LJ_TARGET_NX
  lj_err_caller(L, LJ_ERR_OSUNIQF);
  return 0;
 #else
@@ -57,7 +57,7 @@ static lua_CFunction ll_sym(lua_State *L, void *lib, const char *sym)

 static const char *ll_bcsym(void *lib, const char *sym)
 {
-#if defined(RTLD_DEFAULT)
+#if defined(RTLD_DEFAULT) && !defined(NO_RTLD_DEFAULT)
  if (lib == NULL) lib = RTLD_DEFAULT;
 #elif LJ_TARGET_OSX || LJ_TARGET_BSD
  if (lib == NULL) lib = (void *)(intptr_t)-2;
@@ -330,7 +330,7 @@ static void *mmap_plain(size_t size)
 #define CALL_MMAP(prng, size)	mmap_plain(size)
 #endif

-#if LJ_64 && !LJ_GC64 && ((defined(__FreeBSD__) && __FreeBSD__ < 10) || defined(__FreeBSD_kernel__)) && !LJ_TARGET_PS4
+#if LJ_64 && !LJ_GC64 && ((defined(__FreeBSD__) && __FreeBSD__ < 10) || defined(__FreeBSD_kernel__)) && !LJ_TARGET_PS4 && !LJ_TARGET_PS5

 #include <sys/resource.h>

@@ -779,7 +779,7 @@ LUA_API void lua_concat(lua_State *L, int n)
 	L->top -= n;
 	break;
      }
-      n -= (int)(L->top - top);
+      n -= (int)(L->top - (top - 2*LJ_FR2));
      L->top = top+2;
      lj_vm_call(L, top, 1+1);
      L->top -= 1+LJ_FR2;
@@ -83,7 +83,7 @@
 #define LUAJIT_OS	LUAJIT_OS_OSX
 #elif (defined(__FreeBSD__) || defined(__FreeBSD_kernel__) || \
       defined(__NetBSD__) || defined(__OpenBSD__) || \
-       defined(__DragonFly__)) && !defined(__ORBIS__)
+       defined(__DragonFly__)) && !defined(__ORBIS__) && !defined(__PROSPERO__)
 #define LUAJIT_OS	LUAJIT_OS_BSD
 #elif (defined(__sun__) && defined(__svr4__))
 #define LJ_TARGET_SOLARIS	1
@@ -139,6 +139,13 @@
 #define NULL ((void*)0)
 #endif

+#ifdef __PROSPERO__
+#define LJ_TARGET_PS5		1
+#define LJ_TARGET_CONSOLE	1
+#undef NULL
+#define NULL ((void*)0)
+#endif
+
 #ifdef __psp2__
 #define LJ_TARGET_PSVITA	1
 #define LJ_TARGET_CONSOLE	1
@@ -155,6 +162,13 @@
 #define LJ_TARGET_GC64		1
 #endif

+#ifdef __NX__
+#define LJ_TARGET_NX		1
+#define LJ_TARGET_CONSOLE	1
+#undef NULL
+#define NULL ((void*)0)
+#endif
+
 #ifdef _UWP
 #define LJ_TARGET_UWP		1
 #if LUAJIT_TARGET == LUAJIT_ARCH_X64
@@ -634,7 +648,7 @@ extern void *LJ_WIN_LOADLIBA(const char *path);
 #endif
 #endif

-#if defined(LUAJIT_NO_UNWIND) || __GNU_COMPACT_EH__ || defined(__symbian__) || LJ_TARGET_IOS || LJ_TARGET_PS3 || LJ_TARGET_PS4
+#if defined(LUAJIT_NO_UNWIND) || __GNU_COMPACT_EH__ || defined(__symbian__) || LJ_TARGET_IOS || LJ_TARGET_PS3 || LJ_TARGET_PS4 || LJ_TARGET_PS5
 #define LJ_NO_UNWIND		1
 #endif

@@ -847,7 +847,8 @@ static void ra_destpair(ASMState *as, IRIns *ir)
  if (destlo == RID_RETHI) {
    if (desthi == RID_RETLO) {
 #if LJ_TARGET_X86ORX64
-      *--as->mcp = REX_64IR(irx, XI_XCHGa + RID_RETHI);
+      *--as->mcp = XI_XCHGa + RID_RETHI;
+      if (LJ_64 && irt_is64(irx->t)) *--as->mcp = 0x48;
 #else
      emit_movrr(as, irx, RID_RETHI, RID_TMP);
      emit_movrr(as, irx, RID_RETLO, RID_RETHI);
@@ -1670,7 +1671,6 @@ static void asm_loop(ASMState *as)
 #if !LJ_SOFTFP32
 #if !LJ_TARGET_X86ORX64
 #define asm_ldexp(as, ir)	asm_callid(as, ir, IRCALL_ldexp)
-#define asm_fppowi(as, ir)	asm_callid(as, ir, IRCALL_lj_vm_powi)
 #endif

 static void asm_pow(ASMState *as, IRIns *ir)
@@ -1681,10 +1681,7 @@ static void asm_pow(ASMState *as, IRIns *ir)
 					  IRCALL_lj_carith_powu64);
  else
 #endif
-  if (irt_isnum(IR(ir->op2)->t))
  asm_callid(as, ir, IRCALL_pow);
-  else
-    asm_fppowi(as, ir);
 }

 static void asm_div(ASMState *as, IRIns *ir)
@@ -1250,8 +1250,13 @@ dotypecheck:
      }
    }
    asm_guardcc(as, t == IRT_NUM ? CC_HS : CC_NE);
+    if ((ir->op2 & IRSLOAD_KEYINDEX)) {
+      emit_n(as, ARMI_CMN|ARMI_K12|1, type);
+      emit_dn(as, ARMI_EOR^emit_isk12(ARMI_EOR, ~LJ_KEYINDEX), type, type);
+    } else {
      emit_n(as, ARMI_CMN|ARMI_K12|-irt_toitype_(t), type);
    }
+  }
  if (ra_hasreg(dest)) {
 #if !LJ_SOFTFP
    if (t == IRT_NUM) {
@@ -1209,7 +1209,7 @@ dotypecheck:
      lj_assertA(irt_isinteger(t) || irt_isnum(t),
 		 "bad SLOAD type %d", irt_type(t));
      emit_nm(as, A64I_CMPx | A64F_SH(A64SH_LSR, 32),
-	      ra_allock(as, LJ_TISNUM << 15, allow), tmp);
+	      ra_allock(as, (ir->op2 & IRSLOAD_KEYINDEX) ? LJ_KEYINDEX : (LJ_TISNUM << 15), allow), tmp);
    } else if (irt_isnil(t)) {
      emit_n(as, (A64I_CMNx^A64I_K12) | A64F_U12(1), tmp);
    } else if (irt_ispri(t)) {
@@ -1577,7 +1577,7 @@ dotypecheck:
      asm_guard(as, MIPSI_BEQ, RID_TMP, RID_ZERO);
      emit_tsi(as, MIPSI_SLTIU, RID_TMP, type, (int32_t)LJ_TISNUM);
    } else {
-      Reg ktype = ra_allock(as, irt_toitype(t), allow);
+      Reg ktype = ra_allock(as, (ir->op2 & IRSLOAD_KEYINDEX) ? LJ_KEYINDEX : irt_toitype(t), allow);
      asm_guard(as, MIPSI_BNE, type, ktype);
    }
  }
@@ -1595,6 +1595,10 @@ dotypecheck:
    if (irt_ispri(t)) {
      asm_guard(as, MIPSI_BNE, type,
 		ra_allock(as, ~((int64_t)~irt_toitype(t) << 47) , allow));
+    } else if ((ir->op2 & IRSLOAD_KEYINDEX)) {
+      asm_guard(as, MIPSI_BNE, RID_TMP,
+		ra_allock(as, (int32_t)LJ_KEYINDEX, allow));
+      emit_dta(as, MIPSI_DSRA32, RID_TMP, type, 0);
    } else {
      if (irt_isnum(t)) {
 	asm_guard(as, MIPSI_BEQ, RID_TMP, RID_ZERO);
@@ -2568,7 +2572,22 @@ static void asm_stack_restore(ASMState *as, SnapShot *snap)
      }
      emit_tsi(as, MIPSI_SW, type, RID_BASE, ofs+(LJ_BE?0:4));
 #else
+      if ((sn & SNAP_KEYINDEX)) {
+	RegSet allow = rset_exclude(RSET_GPR, RID_BASE);
+	int64_t kki = (int64_t)LJ_KEYINDEX << 32;
+	if (irref_isk(ref)) {
+	  emit_tsi(as, MIPSI_SD,
+		   ra_allock(as, kki | (int64_t)(uint32_t)ir->i, allow),
+		   RID_BASE, ofs);
+	} else {
+	  Reg src = ra_alloc1(as, ref, allow);
+	  Reg rki = ra_allock(as, kki, rset_exclude(allow, src));
+	  emit_tsi(as, MIPSI_SD, RID_TMP, RID_BASE, ofs);
+	  emit_dst(as, MIPSI_DADDU, RID_TMP, src, rki);
+	}
+      } else {
 	asm_tvstore64(as, RID_BASE, ofs, ref);
+      }
 #endif
    }
    checkmclim(as);
@@ -1169,7 +1169,12 @@ dotypecheck:
  } else {
    if ((ir->op2 & IRSLOAD_TYPECHECK)) {
      asm_guardcc(as, CC_NE);
+      if ((ir->op2 & IRSLOAD_KEYINDEX)) {
+	emit_ai(as, PPCI_CMPWI, RID_TMP, (LJ_KEYINDEX & 0xffff));
+	emit_asi(as, PPCI_XORIS, RID_TMP, RID_TMP, (LJ_KEYINDEX >> 16));
+      } else {
 	emit_ai(as, PPCI_CMPWI, RID_TMP, irt_toitype(t));
+      }
      type = RID_TMP;
    }
    if (ra_hasreg(dest)) emit_tai(as, PPCI_LWZ, dest, base, ofs);
@@ -485,7 +485,8 @@ static Reg asm_fuseload(ASMState *as, IRRef ref, RegSet allow)
 	asm_fusexref(as, ir->op1, xallow);
 	return RID_MRM;
      }
-    } else if (ir->o == IR_VLOAD && !(LJ_GC64 && irt_isaddr(ir->t))) {
+    } else if (ir->o == IR_VLOAD && IR(ir->op1)->o == IR_AREF &&
+	       !(LJ_GC64 && irt_isaddr(ir->t))) {
      asm_fuseahuref(as, ir->op1, xallow);
      as->mrm.ofs += 8 * ir->op2;
      return RID_MRM;
@@ -1768,14 +1769,11 @@ static void asm_sload(ASMState *as, IRIns *ir)
  if ((ir->op2 & IRSLOAD_TYPECHECK)) {
    /* Need type check, even if the load result is unused. */
    asm_guardcc(as, irt_isnum(t) ? CC_AE : CC_NE);
-    if (LJ_64 && irt_type(t) >= IRT_NUM) {
+    if ((LJ_64 && irt_type(t) >= IRT_NUM) || (ir->op2 & IRSLOAD_KEYINDEX)) {
      lj_assertA(irt_isinteger(t) || irt_isnum(t),
 		 "bad SLOAD type %d", irt_type(t));
-#if LJ_GC64
-      emit_u32(as, LJ_TISNUM << 15);
-#else
-      emit_u32(as, LJ_TISNUM);
-#endif
+      emit_u32(as, (ir->op2 & IRSLOAD_KEYINDEX) ? LJ_KEYINDEX :
+		   LJ_GC64 ? (LJ_TISNUM << 15) : LJ_TISNUM);
      emit_rmro(as, XO_ARITHi, XOg_CMP, base, ofs+4);
 #if LJ_GC64
    } else if (irt_isnil(t)) {
@@ -2017,19 +2015,6 @@ static void asm_ldexp(ASMState *as, IRIns *ir)
  asm_x87load(as, ir->op2);
 }

-static void asm_fppowi(ASMState *as, IRIns *ir)
-{
-  /* The modified regs must match with the *.dasc implementation. */
-  RegSet drop = RSET_RANGE(RID_XMM0, RID_XMM1+1)|RID2RSET(RID_EAX);
-  if (ra_hasreg(ir->r))
-    rset_clear(drop, ir->r);  /* Dest reg handled below. */
-  ra_evictset(as, drop);
-  ra_destreg(as, ir, RID_XMM0);
-  emit_call(as, lj_vm_powi_sse);
-  ra_left(as, RID_XMM0, ir->op1);
-  ra_left(as, RID_EAX, ir->op2);
-}
-
 static int asm_swapops(ASMState *as, IRIns *ir)
 {
  IRIns *irl = IR(ir->op1);
@@ -66,7 +66,7 @@ LJ_NOINLINE char *LJ_FASTCALL lj_buf_more2(SBuf *sb, MSize sz)
      lj_err_mem(sbufL(sbx));
    if (len + sz > sbufsz(sbx)) {  /* Must grow. */
      buf_grow((SBuf *)sbx, len + sz);
-    } else if (sbufxslack(sbx) < (sbufsz(sbx) >> 3)) {
+    } else if (sbufiscow(sb) || sbufxslack(sbx) < (sbufsz(sbx) >> 3)) {
      /* Also grow to avoid excessive compactions, if slack < size/8. */
      buf_grow((SBuf *)sbx, sbuflen(sbx) + sz);  /* Not sbufxlen! */
      return sbx->w;
@@ -25,7 +25,7 @@
 #include <dlfcn.h>
 #include <stdio.h>

-#if defined(RTLD_DEFAULT)
+#if defined(RTLD_DEFAULT) && !defined(NO_RTLD_DEFAULT)
 #define CLIB_DEFHANDLE	RTLD_DEFAULT
 #elif LJ_TARGET_OSX || LJ_TARGET_BSD
 #define CLIB_DEFHANDLE	((void *)(intptr_t)-2)
@@ -468,7 +468,7 @@ static void cp_expr_sizeof(CPState *cp, CPValue *k, int wantsz)
  } else {
    cp_expr_unary(cp, k);
  }
-  info = lj_ctype_info(cp->cts, k->id, &sz);
+  info = lj_ctype_info_raw(cp->cts, k->id, &sz);
  if (wantsz) {
    if (sz != CTSIZE_INVALID)
      k->u32 = sz;
@@ -333,6 +333,14 @@ CTInfo lj_ctype_info(CTState *cts, CTypeID id, CTSize *szp)
  return qual;
 }

+/* Ditto, but follow a reference. */
+CTInfo lj_ctype_info_raw(CTState *cts, CTypeID id, CTSize *szp)
+{
+  CType *ct = ctype_get(cts, id);
+  if (ctype_isref(ct->info)) id = ctype_cid(ct->info);
+  return lj_ctype_info(cts, id, szp);
+}
+
 /* Get ctype metamethod. */
 cTValue *lj_ctype_meta(CTState *cts, CTypeID id, MMS mm)
 {
@@ -468,6 +468,7 @@ LJ_FUNC CType *lj_ctype_rawref(CTState *cts, CTypeID id);
 LJ_FUNC CTSize lj_ctype_size(CTState *cts, CTypeID id);
 LJ_FUNC CTSize lj_ctype_vlsize(CTState *cts, CType *ct, CTSize nelem);
 LJ_FUNC CTInfo lj_ctype_info(CTState *cts, CTypeID id, CTSize *szp);
+LJ_FUNC CTInfo lj_ctype_info_raw(CTState *cts, CTypeID id, CTSize *szp);
 LJ_FUNC cTValue *lj_ctype_meta(CTState *cts, CTypeID id, MMS mm);
 LJ_FUNC GCstr *lj_ctype_repr(lua_State *L, CTypeID id, GCstr *name);
 LJ_FUNC GCstr *lj_ctype_repr_int64(lua_State *L, uint64_t n, int isunsigned);
@@ -89,7 +89,7 @@ typedef uint16_t HotCount;
 typedef struct GG_State {
  lua_State L;				/* Main thread. */
  global_State g;			/* Global state. */
-#if LJ_TARGET_ARM
+#if LJ_TARGET_ARM && !LJ_TARGET_NX
  /* Make g reachable via K12 encoded DISPATCH-relative addressing. */
  uint8_t align1[(16-sizeof(global_State))&15];
 #endif
@@ -99,7 +99,7 @@ typedef struct GG_State {
 #if LJ_HASJIT
  jit_State J;				/* JIT state. */
  HotCount hotcount[HOTCOUNT_SIZE];	/* Hot counters. */
-#if LJ_TARGET_ARM
+#if LJ_TARGET_ARM && !LJ_TARGET_NX
  /* Ditto for J. */
  uint8_t align2[(16-sizeof(jit_State)-sizeof(HotCount)*HOTCOUNT_SIZE)&15];
 #endif
@@ -638,8 +638,8 @@ static void LJ_FASTCALL recff_math_call(jit_State *J, RecordFFData *rd)

 static void LJ_FASTCALL recff_math_pow(jit_State *J, RecordFFData *rd)
 {
-  J->base[0] = lj_opt_narrow_pow(J, J->base[0], J->base[1],
-				 &rd->argv[0], &rd->argv[1]);
+  J->base[0] = lj_opt_narrow_arith(J, J->base[0], J->base[1],
+				   &rd->argv[0], &rd->argv[1], IR_POW);
  UNUSED(rd);
 }

@@ -1222,6 +1222,12 @@ static void LJ_FASTCALL recff_buffer_method_put(jit_State *J, RecordFFData *rd)
  TRef tr;
  ptrdiff_t arg;
  if (!J->base[1]) return;
+  for (arg = 1; (tr = J->base[arg]); arg++) {
+    if (tref_isudata(tr)) {
+      TRef ud2 = recff_sbufx_check(J, rd, arg);
+      emitir(IRTG(IR_NE, IRT_PGC), ud, ud2);
+    }
+  }
  for (arg = 1; (tr = J->base[arg]); arg++) {
    if (tref_isstr(tr)) {
      trbuf = emitir(IRTG(IR_BUFPUT, IRT_PGC), trbuf, tr);
@@ -1230,11 +1236,9 @@ static void LJ_FASTCALL recff_buffer_method_put(jit_State *J, RecordFFData *rd)
 		     emitir(IRT(IR_TOSTR, IRT_STR), tr,
 			    tref_isnum(tr) ? IRTOSTR_NUM : IRTOSTR_INT));
    } else if (tref_isudata(tr)) {
-      TRef ud2 = recff_sbufx_check(J, rd, arg);
-      TRef trr = recff_sbufx_get_ptr(J, ud2, IRFL_SBUF_R);
-      TRef trw = recff_sbufx_get_ptr(J, ud2, IRFL_SBUF_W);
+      TRef trr = recff_sbufx_get_ptr(J, tr, IRFL_SBUF_R);
+      TRef trw = recff_sbufx_get_ptr(J, tr, IRFL_SBUF_W);
      TRef len = recff_sbufx_len(J, trr, trw);
-      emitir(IRTG(IR_NE, IRT_PGC), ud, ud2);
      trbuf = lj_ir_call(J, IRCALL_lj_buf_putmem, trbuf, trr, len);
    } else {
      recff_nyiu(J, rd);
@@ -1258,15 +1262,19 @@ static void LJ_FASTCALL recff_buffer_method_get(jit_State *J, RecordFFData *rd)
  TRef tr;
  ptrdiff_t arg;
  if (!J->base[1]) { J->base[1] = TREF_NIL; J->base[2] = 0; }
+  for (arg = 0; (tr = J->base[arg+1]); arg++) {
+    if (!tref_isnil(tr)) {
+      J->base[arg+1] = recff_sbufx_checkint(J, rd, arg+1);
+    }
+  }
  for (arg = 0; (tr = J->base[arg+1]); arg++) {
    TRef len = recff_sbufx_len(J, trr, trw);
    if (tref_isnil(tr)) {
      J->base[arg] = emitir(IRT(IR_XSNEW, IRT_STR), trr, len);
      trr = trw;
    } else {
-      TRef trn = recff_sbufx_checkint(J, rd, arg+1);
      TRef tru;
-      len = emitir(IRTI(IR_MIN), len, trn);
+      len = emitir(IRTI(IR_MIN), len, tr);
      tru = emitir(IRT(IR_ADD, IRT_PTR), trr, len);
      J->base[arg] = emitir(IRT(IR_XSNEW, IRT_STR), trr, len);
      trr = tru;  /* Doing the ADD before the SNEW generates better code. */
@@ -700,9 +700,12 @@ static size_t gc_onestep(lua_State *L)
    }
  case GCSfinalize:
    if (gcref(g->gc.mmudata) != NULL) {
+      GCSize old = g->gc.total;
      if (tvref(g->jit_base))  /* Don't call finalizers on trace. */
 	return LJ_MAX_MEM;
      gc_finalize(L);  /* Finalize one userdata object. */
+      if (old >= g->gc.total && g->gc.estimate > old - g->gc.total)
+	g->gc.estimate -= old - g->gc.total;
      if (g->gc.estimate > GCFINALIZECOST)
 	g->gc.estimate -= GCFINALIZECOST;
      return GCFINALIZECOST;
@@ -217,7 +217,6 @@ typedef struct CCallInfo {
  _(FPMATH,	sqrt,			1,   N, NUM, XA_FP) \
  _(ANY,	log,			1,   N, NUM, XA_FP) \
  _(ANY,	lj_vm_log2,		1,   N, NUM, XA_FP) \
-  _(ANY,	lj_vm_powi,		2,   N, NUM, XA_FP) \
  _(ANY,	pow,			2,   N, NUM, XA2_FP) \
  _(ANY,	atan2,			2,   N, NUM, XA2_FP) \
  _(ANY,	ldexp,			2,   N, NUM, XA_FP) \
@@ -145,7 +145,6 @@ LJ_FUNC TRef lj_opt_narrow_arith(jit_State *J, TRef rb, TRef rc,
 				 TValue *vb, TValue *vc, IROp op);
 LJ_FUNC TRef lj_opt_narrow_unm(jit_State *J, TRef rc, TValue *vc);
 LJ_FUNC TRef lj_opt_narrow_mod(jit_State *J, TRef rb, TRef rc, TValue *vb, TValue *vc);
-LJ_FUNC TRef lj_opt_narrow_pow(jit_State *J, TRef rb, TRef rc, TValue *vb, TValue *vc);
 LJ_FUNC IRType lj_opt_narrow_forl(jit_State *J, cTValue *forbase);

 /* Optimization passes. */
@@ -7,6 +7,7 @@
 #define _LJ_JIT_H

 #include "lj_obj.h"
+#if LJ_HASJIT
 #include "lj_ir.h"

 /* -- JIT engine flags ---------------------------------------------------- */
@@ -371,6 +372,7 @@ enum {
 #endif
  LJ_K64__MAX,
 };
+#define LJ_K64__USED	(LJ_TARGET_X86ORX64 || LJ_TARGET_MIPS)

 enum {
 #if LJ_TARGET_X86ORX64
@@ -389,6 +391,7 @@ enum {
 #endif
  LJ_K32__MAX
 };
+#define LJ_K32__USED	(LJ_TARGET_X86ORX64 || LJ_TARGET_PPC || LJ_TARGET_MIPS)

 /* Get 16 byte aligned pointer to SIMD constant. */
 #define LJ_KSIMD(J, n) \
@@ -443,9 +446,13 @@ typedef struct jit_State {
  int32_t framedepth;	/* Current frame depth. */
  int32_t retdepth;	/* Return frame depth (count of RETF). */

+#if LJ_K32__USED
  uint32_t k32[LJ_K32__MAX];  /* Common 4 byte constants used by backends. */
+#endif
  TValue ksimd[LJ_KSIMD__MAX*2+1];  /* 16 byte aligned SIMD constants. */
+#if LJ_K64__USED
  TValue k64[LJ_K64__MAX];  /* Common 8 byte constants. */
+#endif

  IRIns *irbuf;		/* Temp. IR instruction buffer. Biased with REF_BIAS. */
  IRRef irtoplim;	/* Upper limit of instuction buffer (biased). */
@@ -516,5 +523,6 @@ typedef struct jit_State {
 #else
 #define lj_assertJ(c, ...)	((void)J)
 #endif
+#endif

 #endif
@@ -168,7 +168,7 @@ static void mcode_protect(jit_State *J, int prot)
 #define MCPROT_RUN	MCPROT_RX

 /* Protection twiddling failed. Probably due to kernel security. */
-static LJ_NOINLINE void mcode_protfail(jit_State *J)
+static LJ_NORET LJ_NOINLINE void mcode_protfail(jit_State *J)
 {
  lua_CFunction panic = J2G(J)->panic;
  if (panic) {
@@ -176,6 +176,7 @@ static LJ_NOINLINE void mcode_protfail(jit_State *J)
    setstrV(L, L->top++, lj_err_str(L, LJ_ERR_JITPROT));
    panic(L);
  }
+  exit(EXIT_FAILURE);
 }

 /* Change protection of MCode area. */
@@ -511,7 +511,7 @@ typedef struct GCtab {
 } GCtab;

 #define sizetabcolo(n)	((n)*sizeof(TValue) + sizeof(GCtab))
-#define tabref(r)	(&gcref((r))->tab)
+#define tabref(r)	((GCtab *)gcref((r)))
 #define noderef(r)	(mref((r), Node))
 #define nextnode(n)	(mref((n)->next, Node))
 #if LJ_GC64
@@ -845,6 +845,7 @@ static LJ_AINLINE void *lightudV(global_State *g, cTValue *o)
  uint64_t seg = lightudseg(u);
  uint32_t *segmap = mref(g->gc.lightudseg, uint32_t);
  lj_assertG(tvislightud(o), "lightuserdata expected");
+  if (seg == (1 << LJ_LIGHTUD_BITS_SEG)-1) return NULL;
  lj_assertG(seg <= g->gc.lightudnum, "bad lightuserdata segment %d", seg);
  return (void *)(((uint64_t)segmap[seg] << 32) | lightudlo(u));
 }
@@ -236,14 +236,10 @@ LJFOLDF(kfold_fpcall2)
  return NEXTFOLD;
 }

-LJFOLD(POW KNUM KINT)
 LJFOLD(POW KNUM KNUM)
 LJFOLDF(kfold_numpow)
 {
-  lua_Number a = knumleft;
-  lua_Number b = fright->o == IR_KINT ? (lua_Number)fright->i : knumright;
-  lua_Number y = lj_vm_foldarith(a, b, IR_POW - IR_ADD);
-  return lj_ir_knum(J, y);
+  return lj_ir_knum(J, lj_vm_foldarith(knumleft, knumright, IR_POW - IR_ADD));
 }

 /* Must not use kfold_kref for numbers (could be NaN). */
@@ -598,14 +594,15 @@ LJFOLDF(bufput_bufstr)
    if (fleft->o == IR_BUFHDR && fleft->op2 == IRBUFHDR_RESET &&
 	fleft->prev == hdr &&
 	fleft->op1 == IR(hdr)->op1 &&
-	!(irt_isphi(fright->t) && IR(hdr)->prev)) {
+	!(irt_isphi(fright->t) && IR(hdr)->prev) &&
+	(!LJ_HASBUFFER || J->chain[IR_CALLA] < hdr)) {
      IRRef ref = fins->op1;
      IR(ref)->op2 = IRBUFHDR_APPEND;  /* Modify BUFHDR. */
      IR(ref)->op1 = fright->op1;
      return ref;
    }
    /* Replay puts to global temporary buffer. */
-    if (IR(hdr)->op2 == IRBUFHDR_RESET) {
+    if (IR(hdr)->op2 == IRBUFHDR_RESET && !irt_isphi(fright->t)) {
      IRIns *ir = IR(fright->op1);
      /* For now only handle single string.reverse .lower .upper .rep. */
      if (ir->o == IR_CALLL &&
@@ -1112,60 +1109,16 @@ LJFOLDF(simplify_nummuldiv_negneg)
  return RETRYFOLD;
 }

-LJFOLD(POW any KINT)
-LJFOLDF(simplify_numpow_xkint)
-{
-  int32_t k = fright->i;
-  TRef ref = fins->op1;
-  if (k == 0)  /* x ^ 0 ==> 1 */
-    return lj_ir_knum_one(J);  /* Result must be a number, not an int. */
-  if (k == 1)  /* x ^ 1 ==> x */
-    return LEFTFOLD;
-  if ((uint32_t)(k+65536) > 2*65536u)  /* Limit code explosion. */
-    return NEXTFOLD;
-  if (k < 0) {  /* x ^ (-k) ==> (1/x) ^ k. */
-    ref = emitir(IRTN(IR_DIV), lj_ir_knum_one(J), ref);
-    k = -k;
-  }
-  /* Unroll x^k for 1 <= k <= 65536. */
-  for (; (k & 1) == 0; k >>= 1)  /* Handle leading zeros. */
-    ref = emitir(IRTN(IR_MUL), ref, ref);
-  if ((k >>= 1) != 0) {  /* Handle trailing bits. */
-    TRef tmp = emitir(IRTN(IR_MUL), ref, ref);
-    for (; k != 1; k >>= 1) {
-      if (k & 1)
-	ref = emitir(IRTN(IR_MUL), ref, tmp);
-      tmp = emitir(IRTN(IR_MUL), tmp, tmp);
-    }
-    ref = emitir(IRTN(IR_MUL), ref, tmp);
-  }
-  return ref;
-}
-
 LJFOLD(POW any KNUM)
-LJFOLDF(simplify_numpow_xknum)
+LJFOLDF(simplify_numpow_k)
 {
-  if (knumright == 0.5)  /* x ^ 0.5 ==> sqrt(x) */
-    return emitir(IRTN(IR_FPMATH), fins->op1, IRFPM_SQRT);
-  return NEXTFOLD;
-}
-
-LJFOLD(POW KNUM any)
-LJFOLDF(simplify_numpow_kx)
-{
-  lua_Number n = knumleft;
-  if (n == 2.0 && irt_isint(fright->t)) {  /* 2.0 ^ i ==> ldexp(1.0, i) */
-#if LJ_TARGET_X86ORX64
-    /* Different IR_LDEXP calling convention on x86/x64 requires conversion. */
-    fins->o = IR_CONV;
-    fins->op1 = fins->op2;
-    fins->op2 = IRCONV_NUM_INT;
-    fins->op2 = (IRRef1)lj_opt_fold(J);
-#endif
-    fins->op1 = (IRRef1)lj_ir_knum_one(J);
-    fins->o = IR_LDEXP;
-    return RETRYFOLD;
-  }
+  if (knumright == 0.0)  /* x ^ 0 ==> 1 */
+    return lj_ir_knum_one(J);  /* Result must be a number, not an int. */
+  else if (knumright == 1.0)  /* x ^ 1 ==> x */
+    return LEFTFOLD;
+  else if (knumright == 2.0)  /* x ^ 2 ==> x * x */
+    return emitir(IRTN(IR_MUL), fins->op1, fins->op1);
+  else
    return NEXTFOLD;
 }

@@ -432,7 +432,7 @@ TRef LJ_FASTCALL lj_opt_fwd_alen(jit_State *J)
 	    fins->op2 = aref->op2;  /* Set ALEN hint. */
 	  }
 	  goto doemit;  /* Conflicting store, possibly giving a hint. */
-	} else if (aa_table(J, tab, fref->op1) == ALIAS_NO) {
+	} else if (aa_table(J, tab, fref->op1) != ALIAS_NO) {
 	  goto doemit;  /* Conflicting store. */
 	}
 	sref = store->prev;
@@ -584,36 +584,6 @@ TRef lj_opt_narrow_mod(jit_State *J, TRef rb, TRef rc, TValue *vb, TValue *vc)
  return emitir(IRTN(IR_SUB), rb, tmp);
 }

-/* Narrowing of power operator or math.pow. */
-TRef lj_opt_narrow_pow(jit_State *J, TRef rb, TRef rc, TValue *vb, TValue *vc)
-{
-  rb = conv_str_tonum(J, rb, vb);
-  rb = lj_ir_tonum(J, rb);  /* Left arg is always treated as an FP number. */
-  rc = conv_str_tonum(J, rc, vc);
-  /* Narrowing must be unconditional to preserve (-x)^i semantics. */
-  if (tvisint(vc) || numisint(numV(vc))) {
-    int checkrange = 0;
-    /* pow() is faster for bigger exponents. But do this only for (+k)^i. */
-    if (tref_isk(rb) && (int32_t)ir_knum(IR(tref_ref(rb)))->u32.hi >= 0) {
-      int32_t k = numberVint(vc);
-      if (!(k >= -65536 && k <= 65536)) goto force_pow_num;
-      checkrange = 1;
-    }
-    if (!tref_isinteger(rc)) {
-      /* Guarded conversion to integer! */
-      rc = emitir(IRTGI(IR_CONV), rc, IRCONV_INT_NUM|IRCONV_CHECK);
-    }
-    if (checkrange && !tref_isk(rc)) {  /* Range guard: -65536 <= i <= 65536 */
-      TRef tmp = emitir(IRTI(IR_ADD), rc, lj_ir_kint(J, 65536));
-      emitir(IRTGI(IR_ULE), tmp, lj_ir_kint(J, 2*65536));
-    }
-  } else {
-force_pow_num:
-    rc = lj_ir_tonum(J, rc);  /* Want POW(num, num), not POW(num, int). */
-  }
-  return emitir(IRTN(IR_POW), rb, rc);
-}
-
 /* -- Predictive narrowing of induction variables ------------------------- */

 /* Narrow a single runtime value. */
@@ -400,7 +400,7 @@ static void split_ir(jit_State *J)
 	hi = split_call_ll(J, hisubst, oir, ir, IRCALL_softfp_div);
 	break;
      case IR_POW:
-	hi = split_call_li(J, hisubst, oir, ir, IRCALL_lj_vm_powi);
+	hi = split_call_ll(J, hisubst, oir, ir, IRCALL_pow);
 	break;
      case IR_FPMATH:
 	hi = split_call_l(J, hisubst, oir, ir, IRCALL_lj_vm_floor + ir->op2);
@@ -1554,7 +1554,7 @@ static void fs_fixup_ret(FuncState *fs)
 	/* Replace with UCLO plus branch. */
 	fs->bcbase[pc].ins = BCINS_AD(BC_UCLO, 0, offset);
 	break;
-      case BC_UCLO:
+      case BC_FNEW:
 	return;  /* We're done. */
      default:
 	break;
@@ -83,10 +83,14 @@ extern int XNetRandom(void *buf, unsigned int len);

 extern int sys_get_random_number(void *buf, uint64_t len);

-#elif LJ_TARGET_PS4 || LJ_TARGET_PSVITA
+#elif LJ_TARGET_PS4 || LJ_TARGET_PS5 || LJ_TARGET_PSVITA

 extern int sceRandomGetRandomNumber(void *buf, size_t len);

+#elif LJ_TARGET_NX
+
+#include <unistd.h>
+
 #elif LJ_TARGET_WINDOWS || LJ_TARGET_XBOXONE

 #define WIN32_LEAN_AND_MEAN
@@ -126,7 +130,7 @@ static PRGR libfunc_rgr;
 #endif

 #if LJ_TARGET_HAS_GETENTROPY
-extern int getentropy(void *buf, size_t len);
+extern int getentropy(void *buf, size_t len)
 #ifdef __ELF__
  __attribute__((weak))
 #endif
@@ -171,11 +175,16 @@ int LJ_FASTCALL lj_prng_seed_secure(PRNGState *rs)
  if (sys_get_random_number(rs->u, sizeof(rs->u)) == 0)
    goto ok;

-#elif LJ_TARGET_PS4 || LJ_TARGET_PSVITA
+#elif LJ_TARGET_PS4 || LJ_TARGET_PS5 || LJ_TARGET_PSVITA

  if (sceRandomGetRandomNumber(rs->u, sizeof(rs->u)) == 0)
    goto ok;

+#elif LJ_TARGET_NX
+
+  if (getentropy(rs->u, sizeof(rs->u)) == 0)
+    goto ok;
+
 #elif LJ_TARGET_UWP || LJ_TARGET_XBOXONE

  if (BCryptGenRandom(NULL, (PUCHAR)(rs->u), (ULONG)sizeof(rs->u),
@@ -664,17 +664,24 @@ static LoopEvent rec_itern(jit_State *J, BCReg ra, BCReg rb)
  RecordIndex ix;
  /* Since ITERN is recorded at the start, we need our own loop detection. */
  if (J->pc == J->startpc &&
-      (J->cur.nins > REF_FIRST+1 ||
-       (J->cur.nins == REF_FIRST+1 && J->cur.ir[REF_FIRST].o != IR_PROF)) &&
      J->framedepth + J->retdepth == 0 && J->parent == 0 && J->exitno == 0) {
+    IRRef ref = REF_FIRST + LJ_HASPROFILE;
+#ifdef LUAJIT_ENABLE_CHECKHOOK
+    ref += 3;
+#endif
+    if (J->cur.nins > ref ||
+       (LJ_HASPROFILE && J->cur.nins == ref && J->cur.ir[ref-1].o != IR_PROF)) {
+      J->instunroll = 0;  /* Cannot continue unrolling across an ITERN. */
      lj_record_stop(J, LJ_TRLINK_LOOP, J->cur.traceno);  /* Looping trace. */
      return LOOPEV_ENTER;
    }
+  }
  J->maxslot = ra;
  lj_snap_add(J);  /* Required to make JLOOP the first ins in a side-trace. */
  ix.tab = getslot(J, ra-2);
  ix.key = J->base[ra-1] ? J->base[ra-1] :
-	   sloadt(J, (int32_t)(ra-1), IRT_INT, IRSLOAD_KEYINDEX);
+	   sloadt(J, (int32_t)(ra-1), IRT_GUARD|IRT_INT,
+		  IRSLOAD_TYPECHECK|IRSLOAD_KEYINDEX);
  copyTV(J->L, &ix.tabv, &J->L->base[ra-2]);
  copyTV(J->L, &ix.keyv, &J->L->base[ra-1]);
  ix.idxchain = (rb < 3);  /* Omit value type check, if unused. */
@@ -1446,16 +1453,16 @@ static TRef rec_idx_key(jit_State *J, RecordIndex *ix, IRRef *rbref,
    key = emitir(IRTN(IR_CONV), key, IRCONV_NUM_INT);
  if (tref_isk(key)) {
    /* Optimize lookup of constant hash keys. */
-    MSize hslot = (MSize)((char *)ix->oldv - (char *)&noderef(t->node)[0].val);
-    if (t->hmask > 0 && hslot <= t->hmask*(MSize)sizeof(Node) &&
-	hslot <= 65535*(MSize)sizeof(Node)) {
+    GCSize hslot = (GCSize)((char *)ix->oldv-(char *)&noderef(t->node)[0].val);
+    if (hslot <= t->hmask*(GCSize)sizeof(Node) &&
+	hslot <= 65535*(GCSize)sizeof(Node)) {
      TRef node, kslot, hm;
      *rbref = J->cur.nins;  /* Mark possible rollback point. */
      *rbguard = J->guardemit;
      hm = emitir(IRTI(IR_FLOAD), ix->tab, IRFL_TAB_HMASK);
      emitir(IRTGI(IR_EQ), hm, lj_ir_kint(J, (int32_t)t->hmask));
      node = emitir(IRT(IR_FLOAD, IRT_PGC), ix->tab, IRFL_TAB_NODE);
-      kslot = lj_ir_kslot(J, key, hslot / sizeof(Node));
+      kslot = lj_ir_kslot(J, key, (IRRef)(hslot / sizeof(Node)));
      return emitir(IRTG(IR_HREFK, IRT_PGC), node, kslot);
    }
  }
@@ -1954,7 +1961,7 @@ static void rec_varg(jit_State *J, BCReg dst, ptrdiff_t nresults)
 	  emitir(IRTGI(IR_EQ), fr,
 		 lj_ir_kint(J, (int32_t)frame_ftsz(J->L->base-1)));
 	vbase = emitir(IRT(IR_SUB, IRT_IGC), REF_BASE, fr);
-	vbase = emitir(IRT(IR_ADD, IRT_PGC), vbase, lj_ir_kint(J, frofs-8));
+	vbase = emitir(IRT(IR_ADD, IRT_PGC), vbase, lj_ir_kint(J, frofs-8*(1+LJ_FR2)));
 	for (i = 0; i < nload; i++) {
 	  IRType t = itype2irt(&J->L->base[i-1-LJ_FR2-nvararg]);
 	  J->base[dst+i] = lj_record_vload(J, vbase, i, t);
@@ -2041,7 +2048,7 @@ static TRef rec_tnew(jit_State *J, uint32_t ah)
 static TRef rec_cat(jit_State *J, BCReg baseslot, BCReg topslot)
 {
  TRef *top = &J->base[topslot];
-  TValue savetv[5];
+  TValue savetv[5+LJ_FR2];
  BCReg s;
  RecordIndex ix;
  lj_assertJ(baseslot < topslot, "bad CAT arg");
@@ -2231,6 +2238,7 @@ void lj_record_ins(jit_State *J)
  case BCMpri: setpriV(rcv, ~rc); ix.key = rc = TREF_PRI(IRT_NIL+rc); break;
  case BCMnum: { cTValue *tv = proto_knumtv(J->pt, rc);
    copyTV(J->L, rcv, tv); ix.key = rc = tvisint(tv) ? lj_ir_kint(J, intV(tv)) :
+    tv->u32.hi == LJ_KEYINDEX ? (lj_ir_kint(J, 0) | TREF_KEYINDEX) :
    lj_ir_knumint(J, numV(tv)); } break;
  case BCMstr: { GCstr *s = gco2str(proto_kgc(J->pt, ~(ptrdiff_t)rc));
    setstrV(J->L, rcv, s); ix.key = rc = lj_ir_kstr(J, s); } break;
@@ -2393,7 +2401,7 @@ void lj_record_ins(jit_State *J)

  case BC_POW:
    if (tref_isnumber_str(rb) && tref_isnumber_str(rc))
-      rc = lj_opt_narrow_pow(J, rb, rc, rbv, rcv);
+      rc = lj_opt_narrow_arith(J, rb, rc, rbv, rcv, IR_POW);
    else
      rc = rec_mm_arith(J, &ix, MM_pow);
    break;
@@ -2657,6 +2665,8 @@ static const BCIns *rec_setup_root(jit_State *J)
    J->bc_min = pc;
    break;
  case BC_ITERL:
+    if (bc_op(pc[-1]) == BC_JLOOP)
+      lj_trace_err(J, LJ_TRERR_LINNER);
    lj_assertJ(bc_op(pc[-1]) == BC_ITERC, "no ITERC before ITERL");
    J->maxslot = ra + bc_b(pc[-1]) - 1;
    J->bc_extent = (MSize)(-bc_j(ins))*sizeof(BCIns);
@@ -313,7 +313,7 @@ static BCReg snap_usedef(jit_State *J, uint8_t *udf,
 }

 /* Mark slots used by upvalues of child prototypes as used. */
-void snap_useuv(GCproto *pt, uint8_t *udf)
+static void snap_useuv(GCproto *pt, uint8_t *udf)
 {
  /* This is a coarse check, because it's difficult to correlate the lifetime
  ** of slots and closures. But the number of false positives is quite low.
@@ -672,9 +672,11 @@ static TValue *trace_state(lua_State *L, lua_CFunction dummy, void *ud)
      trace_pendpatch(J, 0);
      setvmstate(J2G(J), RECORD);
      lj_vmevent_send_(L, RECORD,
-	/* Save/restore tmptv state for trace recorder. */
+	/* Save/restore state for trace recorder. */
 	TValue savetv = J2G(J)->tmptv;
 	TValue savetv2 = J2G(J)->tmptv2;
+	TraceNo parent = J->parent;
+	ExitNo exitno = J->exitno;
 	setintV(L->top++, J->cur.traceno);
 	setfuncV(L, L->top++, J->fn);
 	setintV(L->top++, J->pt ? (int32_t)proto_bcpos(J->pt, J->pc) : -1);
@@ -682,6 +684,8 @@ static TValue *trace_state(lua_State *L, lua_CFunction dummy, void *ud)
      ,
 	J2G(J)->tmptv = savetv;
 	J2G(J)->tmptv2 = savetv2;
+	J->parent = parent;
+	J->exitno = exitno;
      );
      lj_record_ins(J);
      break;
@@ -83,10 +83,6 @@ LJ_ASMF int32_t LJ_FASTCALL lj_vm_modi(int32_t, int32_t);
 LJ_ASMF void lj_vm_floor_sse(void);
 LJ_ASMF void lj_vm_ceil_sse(void);
 LJ_ASMF void lj_vm_trunc_sse(void);
-LJ_ASMF void lj_vm_powi_sse(void);
-#define lj_vm_powi	NULL
-#else
-LJ_ASMF double lj_vm_powi(double, int32_t);
 #endif
 #if LJ_TARGET_PPC || LJ_TARGET_ARM64
 #define lj_vm_trunc	trunc
@@ -34,7 +34,7 @@ LJ_FUNCA double lj_wrap_pow(double x, double y) { return pow(x, y); }
 LJ_FUNCA double lj_wrap_fmod(double x, double y) { return fmod(x, y); }
 #endif

-/* -- Helper functions for generated machine code ------------------------- */
+/* -- Helper functions ---------------------------------------------------- */

 double lj_vm_foldarith(double x, double y, int op)
 {
@@ -56,6 +56,8 @@ double lj_vm_foldarith(double x, double y, int op)
  }
 }

+/* -- Helper functions for generated machine code ------------------------- */
+
 #if (LJ_HASJIT && !(LJ_TARGET_ARM || LJ_TARGET_ARM64 || LJ_TARGET_PPC)) || LJ_TARGET_MIPS
 int32_t LJ_FASTCALL lj_vm_modi(int32_t a, int32_t b)
 {
@@ -80,40 +82,6 @@ double lj_vm_log2(double a)
 }
 #endif

-#if !LJ_TARGET_X86ORX64
-/* Unsigned x^k. */
-static double lj_vm_powui(double x, uint32_t k)
-{
-  double y;
-  lj_assertX(k != 0, "pow with zero exponent");
-  for (; (k & 1) == 0; k >>= 1) x *= x;
-  y = x;
-  if ((k >>= 1) != 0) {
-    for (;;) {
-      x *= x;
-      if (k == 1) break;
-      if (k & 1) y *= x;
-      k >>= 1;
-    }
-    y *= x;
-  }
-  return y;
-}
-
-/* Signed x^k. */
-double lj_vm_powi(double x, int32_t k)
-{
-  if (k > 1)
-    return lj_vm_powui(x, (uint32_t)k);
-  else if (k == 1)
-    return x;
-  else if (k == 0)
-    return 1.0;
-  else
-    return 1.0 / lj_vm_powui(x, (uint32_t)-k);
-}
-#endif
-
 /* Computes fpm(x) for extended math functions. */
 double lj_vm_foldfpm(double x, int fpm)
 {
@@ -39,6 +39,7 @@

 static lua_State *globalL = NULL;
 static const char *progname = LUA_PROGNAME;
+static char *empty_argv[2] = { NULL, NULL };

 #if !LJ_TARGET_CONSOLE
 static void lstop(lua_State *L, lua_Debug *ar)
@@ -78,9 +79,9 @@ static void print_usage(void)
  fflush(stderr);
 }

-static void l_message(const char *pname, const char *msg)
+static void l_message(const char *msg)
 {
-  if (pname) { fputs(pname, stderr); fputc(':', stderr); fputc(' ', stderr); }
+  if (progname) { fputs(progname, stderr); fputc(':', stderr); fputc(' ', stderr); }
  fputs(msg, stderr); fputc('\n', stderr);
  fflush(stderr);
 }
@@ -90,7 +91,7 @@ static int report(lua_State *L, int status)
  if (status && !lua_isnil(L, -1)) {
    const char *msg = lua_tostring(L, -1);
    if (msg == NULL) msg = "(error object is not a string)";
-    l_message(progname, msg);
+    l_message(msg);
    lua_pop(L, 1);
  }
  return status;
@@ -256,8 +257,7 @@ static void dotty(lua_State *L)
      lua_getglobal(L, "print");
      lua_insert(L, 1);
      if (lua_pcall(L, lua_gettop(L)-1, 0, 0) != 0)
-	l_message(progname,
-	  lua_pushfstring(L, "error calling " LUA_QL("print") " (%s)",
+	l_message(lua_pushfstring(L, "error calling " LUA_QL("print") " (%s)",
 				  lua_tostring(L, -1)));
    }
  }
@@ -310,8 +310,7 @@ static int loadjitmodule(lua_State *L)
  lua_getfield(L, -1, "start");
  if (lua_isnil(L, -1)) {
  nomodule:
-    l_message(progname,
-	      "unknown luaJIT command or jit.* modules not installed");
+    l_message("unknown luaJIT command or jit.* modules not installed");
    return 1;
  }
  lua_remove(L, -2);  /* Drop module table. */
@@ -516,8 +515,6 @@ static int pmain(lua_State *L)
  int argn;
  int flags = 0;
  globalL = L;
-  if (argv[0] && argv[0][0]) progname = argv[0];
-
  LUAJIT_VERSION_SYM();  /* Linker-enforced version check. */

  argn = collectargs(argv, &flags);
@@ -572,9 +569,11 @@ static int pmain(lua_State *L)
 int main(int argc, char **argv)
 {
  int status;
-  lua_State *L = lua_open();
+  lua_State *L;
+  if (!argv[0]) argv = empty_argv; else if (argv[0][0]) progname = argv[0];
+  L = lua_open();
  if (L == NULL) {
-    l_message(argv[0], "cannot create state: not enough memory");
+    l_message("cannot create state: not enough memory");
    return EXIT_FAILURE;
  }
  smain.argc = argc;
@@ -0,0 +1,159 @@
+@rem Script to build LuaJIT with NintendoSDK + NX Addon.
+@rem Donated to the public domain by Swyter.
+@rem
+@rem To run this script you must open a "Native Tools Command Prompt for VS".
+@rem
+@rem Either the x86 version for NX32, or x64 for the NX64 target.
+@rem This is because the pointer size of the LuaJIT host tools (buildvm.exe)
+@rem must match the cross-compiled target (32 or 64 bits).
+@rem
+@rem Then cd to this directory and run this script.
+@rem
+@rem Recommended invocation:
+@rem
+@rem nxbuild            # release build, amalgamated
+@rem nxbuild debug      # debug build, amalgamated
+@rem
+@rem Additional command-line options (not generally recommended):
+@rem
+@rem noamalg            # (after debug) non-amalgamated build
+
+@if not defined INCLUDE goto :FAIL
+@if not defined NINTENDO_SDK_ROOT goto :FAIL
+@if not defined PLATFORM goto :FAIL
+
+@if "%platform%" == "x86" goto :DO_NX32
+@if "%platform%" == "x64" goto :DO_NX64
+
+@echo Error: Current host platform is %platform%!
+@echo.
+@goto :FAIL
+
+@setlocal
+
+:DO_NX32
+@set DASC=vm_arm.dasc
+@set DASMFLAGS= -D HFABI -D FPU
+@set DASMTARGET= -D LUAJIT_TARGET=LUAJIT_ARCH_ARM
+@set HOST_PTR_SIZE=4
+goto :BEGIN
+
+:DO_NX64
+@set DASC=vm_arm64.dasc
+@set DASMFLAGS= -D ENDIAN_LE
+@set DASMTARGET= -D LUAJIT_TARGET=LUAJIT_ARCH_ARM64
+@set HOST_PTR_SIZE=8
+
+:BEGIN
+@rem ---- Host compiler ----
+@set LJCOMPILE=cl /nologo /c /MD /O2 /W3 /wo4146 /wo4244 /D_CRT_SECURE_NO_DEPRECATE
+@set LJLINK=link /nologo
+@set LJMT=mt /nologo
+@set DASMDIR=..\dynasm
+@set DASM=%DASMDIR%\dynasm.lua
+@set ALL_LIB=lib_base.c lib_math.c lib_bit.c lib_string.c lib_table.c lib_io.c lib_os.c lib_package.c lib_debug.c lib_jit.c lib_ffi.c lib_buffer.c
+
+%LJCOMPILE% host\minilua.c
+@if errorlevel 1 goto :BAD
+%LJLINK% /out:minilua.exe minilua.obj
+@if errorlevel 1 goto :BAD
+if exist minilua.exe.manifest^
+  %LJMT% -manifest minilua.exe.manifest -outputresource:minilua.exe
+
+@rem Check that we have the right 32/64 bit host compiler to generate the right virtual machine files.
+@minilua
+@if "%ERRORLEVEL%" == "%HOST_PTR_SIZE%" goto :PASSED_PTR_CHECK
+
+@echo The pointer size of the host in bytes (%HOST_PTR_SIZE%) does not match the expected value (%errorlevel%).
+@echo Check that the script is being ran under the correct x86/x64 VS prompt.
+@goto :BAD
+
+:PASSED_PTR_CHECK
+@set DASMFLAGS=%DASMFLAGS% %DASMTARGET% -D LJ_TARGET_NX -D LUAJIT_OS=LUAJIT_OS_OTHER -D LUAJIT_DISABLE_JIT -D LUAJIT_DISABLE_FFI
+minilua %DASM% -LN %DASMFLAGS% -o host\buildvm_arch.h %DASC%
+@if errorlevel 1 goto :BAD
+%LJCOMPILE% /I "." /I %DASMDIR% %DASMTARGET% -D LJ_TARGET_NX -DLUAJIT_OS=LUAJIT_OS_OTHER -DLUAJIT_DISABLE_JIT -DLUAJIT_DISABLE_FFI host\buildvm*.c
+@if errorlevel 1 goto :BAD
+%LJLINK% /out:buildvm.exe buildvm*.obj
+@if errorlevel 1 goto :BAD
+if exist buildvm.exe.manifest^
+  %LJMT% -manifest buildvm.exe.manifest -outputresource:buildvm.exe
+
+buildvm -m elfasm -o lj_vm.s
+@if errorlevel 1 goto :BAD
+buildvm -m bcdef -o lj_bcdef.h %ALL_LIB%
+@if errorlevel 1 goto :BAD
+buildvm -m ffdef -o lj_ffdef.h %ALL_LIB%
+@if errorlevel 1 goto :BAD
+buildvm -m libdef -o lj_libdef.h %ALL_LIB%
+@if errorlevel 1 goto :BAD
+buildvm -m recdef -o lj_recdef.h %ALL_LIB%
+@if errorlevel 1 goto :BAD
+buildvm -m vmdef -o jit\vmdef.lua %ALL_LIB%
+@if errorlevel 1 goto :BAD
+buildvm -m folddef -o lj_folddef.h lj_opt_fold.c
+@if errorlevel 1 goto :BAD
+
+@rem ---- Cross compiler ----
+@if "%platform%" neq "x64" goto :NX32_CROSSBUILD
+@set LJCOMPILE="%NINTENDO_SDK_ROOT%\Compilers\NX\nx\aarch64\bin\clang" -Wall -I%NINTENDO_SDK_ROOT%\Include %DASMTARGET% -DLUAJIT_OS=LUAJIT_OS_OTHER -DLUAJIT_DISABLE_JIT -DLUAJIT_DISABLE_FFI -DLUAJIT_USE_SYSMALLOC -c
+@set LJLIB="%NINTENDO_SDK_ROOT%\Compilers\NX\nx\aarch64\bin\aarch64-nintendo-nx-elf-ar" rc
+@set TARGETLIB_SUFFIX=nx64
+
+%NINTENDO_SDK_ROOT%\Compilers\NX\nx\aarch64\bin\aarch64-nintendo-nx-elf-as -o lj_vm.o lj_vm.s
+goto :DEBUGCHECK
+
+:NX32_CROSSBUILD
+@set LJCOMPILE="%NINTENDO_SDK_ROOT%\Compilers\NX\nx\armv7l\bin\clang" -Wall -I%NINTENDO_SDK_ROOT%\Include %DASMTARGET% -DLUAJIT_OS=LUAJIT_OS_OTHER -DLUAJIT_DISABLE_JIT -DLUAJIT_DISABLE_FFI -DLUAJIT_USE_SYSMALLOC -c
+@set LJLIB="%NINTENDO_SDK_ROOT%\Compilers\NX\nx\armv7l\bin\armv7l-nintendo-nx-eabihf-ar" rc
+@set TARGETLIB_SUFFIX=nx32
+
+%NINTENDO_SDK_ROOT%\Compilers\NX\nx\armv7l\bin\armv7l-nintendo-nx-eabihf-as -o lj_vm.o lj_vm.s
+:DEBUGCHECK
+
+@if "%1" neq "debug" goto :NODEBUG
+@shift
+@set LJCOMPILE=%LJCOMPILE% -DNN_SDK_BUILD_DEBUG -g -O0
+@set TARGETLIB=libluajitD_%TARGETLIB_SUFFIX%.a
+goto :BUILD
+:NODEBUG
+@set LJCOMPILE=%LJCOMPILE% -DNN_SDK_BUILD_RELEASE -O3
+@set TARGETLIB=libluajit_%TARGETLIB_SUFFIX%.a
+:BUILD
+del %TARGETLIB%
+@if "%1" neq "noamalg" goto :AMALG
+for %%f in (lj_*.c lib_*.c) do (
+  %LJCOMPILE% %%f
+  @if errorlevel 1 goto :BAD
+)
+
+%LJLIB% %TARGETLIB% lj_*.o lib_*.o
+@if errorlevel 1 goto :BAD
+@goto :NOAMALG
+:AMALG
+%LJCOMPILE% ljamalg.c
+@if errorlevel 1 goto :BAD
+%LJLIB% %TARGETLIB% ljamalg.o lj_vm.o
+@if errorlevel 1 goto :BAD
+:NOAMALG
+
+@del *.o *.obj *.manifest minilua.exe buildvm.exe
+@echo.
+@echo === Successfully built LuaJIT for Nintendo Switch (%TARGETLIB_SUFFIX%) ===
+
+@goto :END
+:BAD
+@echo.
+@echo *******************************************************
+@echo *** Build FAILED -- Please check the error messages ***
+@echo *******************************************************
+@goto :END
+:FAIL
+@echo To run this script you must open a "Native Tools Command Prompt for VS".
+@echo.
+@echo Either the x86 version for NX32, or x64 for the NX64 target.
+@echo This is because the pointer size of the LuaJIT host tools (buildvm.exe)
+@echo must match the cross-compiled target (32 or 64 bits).
+@echo.
+@echo Keep in mind that NintendoSDK + NX Addon must be installed, too.
+:END
@@ -0,0 +1,123 @@
+@rem Script to build LuaJIT with the PS5 SDK.
+@rem Donated to the public domain.
+@rem
+@rem Open a "Visual Studio .NET Command Prompt" (64 bit host compiler)
+@rem or "VS20xx x64 Native Tools Command Prompt".
+@rem
+@rem Then cd to this directory and run this script.
+@rem
+@rem Recommended invocation:
+@rem
+@rem ps5build        release build, amalgamated, 64-bit GC
+@rem ps5build debug    debug build, amalgamated, 64-bit GC
+@rem
+@rem Additional command-line options (not generally recommended):
+@rem
+@rem gc32 (before debug)    32-bit GC
+@rem noamalg (after debug)  non-amalgamated build
+
+@if not defined INCLUDE goto :FAIL
+@if not defined SCE_PROSPERO_SDK_DIR goto :FAIL
+
+@setlocal
+@rem ---- Host compiler ----
+@set LJCOMPILE=cl /nologo /c /MD /O2 /W3 /D_CRT_SECURE_NO_DEPRECATE
+@set LJLINK=link /nologo
+@set LJMT=mt /nologo
+@set DASMDIR=..\dynasm
+@set DASM=%DASMDIR%\dynasm.lua
+@set ALL_LIB=lib_base.c lib_math.c lib_bit.c lib_string.c lib_table.c lib_io.c lib_os.c lib_package.c lib_debug.c lib_jit.c lib_ffi.c lib_buffer.c
+@set GC64=
+@set DASC=vm_x64.dasc
+
+@if "%1" neq "gc32" goto :NOGC32
+@shift
+@set GC64=-DLUAJIT_DISABLE_GC64
+@set DASC=vm_x86.dasc
+:NOGC32
+
+%LJCOMPILE% host\minilua.c
+@if errorlevel 1 goto :BAD
+%LJLINK% /out:minilua.exe minilua.obj
+@if errorlevel 1 goto :BAD
+if exist minilua.exe.manifest^
+  %LJMT% -manifest minilua.exe.manifest -outputresource:minilua.exe
+
+@rem Check for 64 bit host compiler.
+@minilua
+@if not errorlevel 8 goto :FAIL
+
+@set DASMFLAGS=-D P64 -D NO_UNWIND
+minilua %DASM% -LN %DASMFLAGS% -o host\buildvm_arch.h %DASC%
+@if errorlevel 1 goto :BAD
+
+%LJCOMPILE% /I "." /I %DASMDIR% %GC64% -DLUAJIT_TARGET=LUAJIT_ARCH_X64 -DLUAJIT_OS=LUAJIT_OS_OTHER -DLUAJIT_DISABLE_JIT -DLUAJIT_DISABLE_FFI -DLUAJIT_NO_UNWIND host\buildvm*.c
+@if errorlevel 1 goto :BAD
+%LJLINK% /out:buildvm.exe buildvm*.obj
+@if errorlevel 1 goto :BAD
+if exist buildvm.exe.manifest^
+  %LJMT% -manifest buildvm.exe.manifest -outputresource:buildvm.exe
+
+buildvm -m elfasm -o lj_vm.s
+@if errorlevel 1 goto :BAD
+buildvm -m bcdef -o lj_bcdef.h %ALL_LIB%
+@if errorlevel 1 goto :BAD
+buildvm -m ffdef -o lj_ffdef.h %ALL_LIB%
+@if errorlevel 1 goto :BAD
+buildvm -m libdef -o lj_libdef.h %ALL_LIB%
+@if errorlevel 1 goto :BAD
+buildvm -m recdef -o lj_recdef.h %ALL_LIB%
+@if errorlevel 1 goto :BAD
+buildvm -m vmdef -o jit\vmdef.lua %ALL_LIB%
+@if errorlevel 1 goto :BAD
+buildvm -m folddef -o lj_folddef.h lj_opt_fold.c
+@if errorlevel 1 goto :BAD
+
+@rem ---- Cross compiler ----
+@set LJCOMPILE="%SCE_PROSPERO_SDK_DIR%\host_tools\bin\prospero-clang" -c -Wall -DLUAJIT_DISABLE_FFI %GC64%
+@set LJLIB="%SCE_PROSPERO_SDK_DIR%\host_tools\bin\prospero-llvm-ar" rcus
+@set INCLUDE=""
+
+%SCE_PROSPERO_SDK_DIR%\host_tools\bin\prospero-llvm-as -o lj_vm.o lj_vm.s
+
+@if "%1" neq "debug" goto :NODEBUG
+@shift
+@set LJCOMPILE=%LJCOMPILE% -g -O0
+@set TARGETLIB=libluajitD_ps5.a
+goto :BUILD
+:NODEBUG
+@set LJCOMPILE=%LJCOMPILE% -O2
+@set TARGETLIB=libluajit_ps5.a
+:BUILD
+del %TARGETLIB%
+@if "%1" neq "noamalg" goto :AMALG
+for %%f in (lj_*.c lib_*.c) do (
+  %LJCOMPILE% %%f
+  @if errorlevel 1 goto :BAD
+)
+
+%LJLIB% %TARGETLIB% lj_*.o lib_*.o
+@if errorlevel 1 goto :BAD
+@goto :NOAMALG
+:AMALG
+%LJCOMPILE% ljamalg.c
+@if errorlevel 1 goto :BAD
+%LJLIB% %TARGETLIB% ljamalg.o lj_vm.o
+@if errorlevel 1 goto :BAD
+:NOAMALG
+
+@del *.o *.obj *.manifest minilua.exe buildvm.exe
+@echo.
+@echo === Successfully built LuaJIT for PS5 ===
+
+@goto :END
+:BAD
+@echo.
+@echo *******************************************************
+@echo *** Build FAILED -- Please check the error messages ***
+@echo *******************************************************
+@goto :END
+:FAIL
+@echo To run this script you must open a "Visual Studio .NET Command Prompt"
+@echo (64 bit host compiler). The PS5 Prospero SDK must be installed, too.
+:END
@@ -3988,6 +3988,7 @@ static void emit_asm_debug(BuildCtx *ctx)
 	"\t.align 3\n"
 	".LEFDE1:\n\n", (int)ctx->codesz - fcofs);
 #endif
+#if !LJ_NO_UNWIND
    fprintf(ctx->fp, "\t.section .eh_frame,\"a\",%%progbits\n");
    fprintf(ctx->fp,
 	".Lframe1:\n"
@@ -4055,6 +4056,7 @@ static void emit_asm_debug(BuildCtx *ctx)
 	"\t.byte 0x94\n\t.uleb128 4\n"		/* offset x20 */
 	"\t.align 3\n"
 	".LEFDE3:\n\n", (int)ctx->codesz - fcofs);
+#endif
 #endif
    break;
 #if !LJ_NO_UNWIND
@@ -359,9 +359,6 @@
 |.macro sseconst_1, reg, tmp		// Synthesize 1.0.
 |  sseconst_hi reg, tmp, 3ff00000
 |.endmacro
-|.macro sseconst_m1, reg, tmp		// Synthesize -1.0.
-|  sseconst_hi reg, tmp, bff00000
-|.endmacro
 |.macro sseconst_2p52, reg, tmp		// Synthesize 2^52.
 |  sseconst_hi reg, tmp, 43300000
 |.endmacro
@@ -2530,15 +2527,17 @@ static void build_subroutines(BuildCtx *ctx)
  |  addsd xmm1, xmm3			// (|x| + 2^52) - 2^52
  |  subsd xmm1, xmm3
  |  orpd xmm1, xmm2			// Merge sign bit back in.
+  |  sseconst_1 xmm3, RD
  |  .if mode == 1		// ceil(x)?
-  |    sseconst_m1 xmm2, RD		// Must subtract -1 to preserve -0.
  |    cmpsd xmm0, xmm1, 6		// x > result?
+  |    andpd xmm0, xmm3
+  |    addsd xmm1, xmm0			// If yes, add 1.
+  |    orpd xmm1, xmm2			// Merge sign bit back in (again).
  |  .else			// floor(x)?
-  |    sseconst_1 xmm2, RD
  |    cmpsd xmm0, xmm1, 1		// x < result?
+  |    andpd xmm0, xmm3
+  |    subsd xmm1, xmm0			// If yes, subtract 1.
  |  .endif
-  |  andpd xmm0, xmm2
-  |  subsd xmm1, xmm0			// If yes, subtract +-1.
  |.endif
  |  movaps xmm0, xmm1
  |1:
@@ -2579,41 +2578,6 @@ static void build_subroutines(BuildCtx *ctx)
  |  subsd xmm0, xmm1
  |  ret
  |
-  |// Args in xmm0/eax. Ret in xmm0. xmm0-xmm1 and eax modified.
-  |->vm_powi_sse:
-  |  cmp eax, 1; jle >6			// i<=1?
-  |  // Now 1 < (unsigned)i <= 0x80000000.
-  |1:  // Handle leading zeros.
-  |  test eax, 1; jnz >2
-  |  mulsd xmm0, xmm0
-  |  shr eax, 1
-  |  jmp <1
-  |2:
-  |  shr eax, 1; jz >5
-  |  movaps xmm1, xmm0
-  |3:  // Handle trailing bits.
-  |  mulsd xmm0, xmm0
-  |  shr eax, 1; jz >4
-  |  jnc <3
-  |  mulsd xmm1, xmm0
-  |  jmp <3
-  |4:
-  |  mulsd xmm0, xmm1
-  |5:
-  |  ret
-  |6:
-  |  je <5				// x^1 ==> x
-  |  jb >7				// x^0 ==> 1
-  |  neg eax
-  |  call <1
-  |  sseconst_1 xmm1, RD
-  |  divsd xmm1, xmm0
-  |  movaps xmm0, xmm1
-  |  ret
-  |7:
-  |  sseconst_1 xmm0, RD
-  |  ret
-  |
  |//-----------------------------------------------------------------------
  |//-- Miscellaneous functions --------------------------------------------
  |//-----------------------------------------------------------------------
@@ -464,9 +464,6 @@
 |.macro sseconst_1, reg, tmp		// Synthesize 1.0.
 |  sseconst_hi reg, tmp, 3ff00000
 |.endmacro
-|.macro sseconst_m1, reg, tmp		// Synthesize -1.0.
-|  sseconst_hi reg, tmp, bff00000
-|.endmacro
 |.macro sseconst_2p52, reg, tmp		// Synthesize 2^52.
 |  sseconst_hi reg, tmp, 43300000
 |.endmacro
@@ -2989,15 +2986,17 @@ static void build_subroutines(BuildCtx *ctx)
  |  addsd xmm1, xmm3			// (|x| + 2^52) - 2^52
  |  subsd xmm1, xmm3
  |  orpd xmm1, xmm2			// Merge sign bit back in.
+  |  sseconst_1 xmm3, RDa
  |  .if mode == 1		// ceil(x)?
-  |    sseconst_m1 xmm2, RDa		// Must subtract -1 to preserve -0.
  |    cmpsd xmm0, xmm1, 6		// x > result?
+  |    andpd xmm0, xmm3
+  |    addsd xmm1, xmm0			// If yes, add 1.
+  |    orpd xmm1, xmm2			// Merge sign bit back in (again).
  |  .else			// floor(x)?
-  |    sseconst_1 xmm2, RDa
  |    cmpsd xmm0, xmm1, 1		// x < result?
+  |    andpd xmm0, xmm3
+  |    subsd xmm1, xmm0			// If yes, subtract 1.
  |  .endif
-  |  andpd xmm0, xmm2
-  |  subsd xmm1, xmm0			// If yes, subtract +-1.
  |.endif
  |  movaps xmm0, xmm1
  |1:
@@ -3038,41 +3037,6 @@ static void build_subroutines(BuildCtx *ctx)
  |  subsd xmm0, xmm1
  |  ret
  |
-  |// Args in xmm0/eax. Ret in xmm0. xmm0-xmm1 and eax modified.
-  |->vm_powi_sse:
-  |  cmp eax, 1; jle >6			// i<=1?
-  |  // Now 1 < (unsigned)i <= 0x80000000.
-  |1:  // Handle leading zeros.
-  |  test eax, 1; jnz >2
-  |  mulsd xmm0, xmm0
-  |  shr eax, 1
-  |  jmp <1
-  |2:
-  |  shr eax, 1; jz >5
-  |  movaps xmm1, xmm0
-  |3:  // Handle trailing bits.
-  |  mulsd xmm0, xmm0
-  |  shr eax, 1; jz >4
-  |  jnc <3
-  |  mulsd xmm1, xmm0
-  |  jmp <3
-  |4:
-  |  mulsd xmm0, xmm1
-  |5:
-  |  ret
-  |6:
-  |  je <5				// x^1 ==> x
-  |  jb >7				// x^0 ==> 1
-  |  neg eax
-  |  call <1
-  |  sseconst_1 xmm1, RDa
-  |  divsd xmm1, xmm0
-  |  movaps xmm0, xmm1
-  |  ret
-  |7:
-  |  sseconst_1 xmm0, RDa
-  |  ret
-  |
  |//-----------------------------------------------------------------------
  |//-- Miscellaneous functions --------------------------------------------
  |//-----------------------------------------------------------------------