You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@trafficserver.apache.org by zw...@apache.org on 2015/07/23 13:14:27 UTC

[30/43] trafficserver git commit: TS-3783 TS-3030 Add luajit v2.0.4 as a subtree

http://git-wip-us.apache.org/repos/asf/trafficserver/blob/1f27b840/lib/luajit/src/lj_alloc.c
----------------------------------------------------------------------
diff --git a/lib/luajit/src/lj_alloc.c b/lib/luajit/src/lj_alloc.c
new file mode 100644
index 0000000..7c7ec67
--- /dev/null
+++ b/lib/luajit/src/lj_alloc.c
@@ -0,0 +1,1396 @@
+/*
+** Bundled memory allocator.
+**
+** Beware: this is a HEAVILY CUSTOMIZED version of dlmalloc.
+** The original bears the following remark:
+**
+**   This is a version (aka dlmalloc) of malloc/free/realloc written by
+**   Doug Lea and released to the public domain, as explained at
+**   http://creativecommons.org/licenses/publicdomain.
+**
+**   * Version pre-2.8.4 Wed Mar 29 19:46:29 2006    (dl at gee)
+**
+** No additional copyright is claimed over the customizations.
+** Please do NOT bother the original author about this version here!
+**
+** If you want to use dlmalloc in another project, you should get
+** the original from: ftp://gee.cs.oswego.edu/pub/misc/
+** For thread-safe derivatives, take a look at:
+** - ptmalloc: http://www.malloc.de/
+** - nedmalloc: http://www.nedprod.com/programs/portable/nedmalloc/
+*/
+
+#define lj_alloc_c
+#define LUA_CORE
+
+/* To get the mremap prototype. Must be defined before any system includes. */
+#if defined(__linux__) && !defined(_GNU_SOURCE)
+#define _GNU_SOURCE
+#endif
+
+#include "lj_def.h"
+#include "lj_arch.h"
+#include "lj_alloc.h"
+
+#ifndef LUAJIT_USE_SYSMALLOC
+
+#define MAX_SIZE_T		(~(size_t)0)
+#define MALLOC_ALIGNMENT	((size_t)8U)
+
+#define DEFAULT_GRANULARITY	((size_t)128U * (size_t)1024U)
+#define DEFAULT_TRIM_THRESHOLD	((size_t)2U * (size_t)1024U * (size_t)1024U)
+#define DEFAULT_MMAP_THRESHOLD	((size_t)128U * (size_t)1024U)
+#define MAX_RELEASE_CHECK_RATE	255
+
+/* ------------------- size_t and alignment properties -------------------- */
+
+/* The byte and bit size of a size_t */
+#define SIZE_T_SIZE		(sizeof(size_t))
+#define SIZE_T_BITSIZE		(sizeof(size_t) << 3)
+
+/* Some constants coerced to size_t */
+/* Annoying but necessary to avoid errors on some platforms */
+#define SIZE_T_ZERO		((size_t)0)
+#define SIZE_T_ONE		((size_t)1)
+#define SIZE_T_TWO		((size_t)2)
+#define TWO_SIZE_T_SIZES	(SIZE_T_SIZE<<1)
+#define FOUR_SIZE_T_SIZES	(SIZE_T_SIZE<<2)
+#define SIX_SIZE_T_SIZES	(FOUR_SIZE_T_SIZES+TWO_SIZE_T_SIZES)
+
+/* The bit mask value corresponding to MALLOC_ALIGNMENT */
+#define CHUNK_ALIGN_MASK	(MALLOC_ALIGNMENT - SIZE_T_ONE)
+
+/* the number of bytes to offset an address to align it */
+#define align_offset(A)\
+ ((((size_t)(A) & CHUNK_ALIGN_MASK) == 0)? 0 :\
+  ((MALLOC_ALIGNMENT - ((size_t)(A) & CHUNK_ALIGN_MASK)) & CHUNK_ALIGN_MASK))
+
+/* -------------------------- MMAP support ------------------------------- */
+
+#define MFAIL			((void *)(MAX_SIZE_T))
+#define CMFAIL			((char *)(MFAIL)) /* defined for convenience */
+
+#define IS_DIRECT_BIT		(SIZE_T_ONE)
+
+#if LJ_TARGET_WINDOWS
+
+#define WIN32_LEAN_AND_MEAN
+#include <windows.h>
+
+#if LJ_64
+
+/* Undocumented, but hey, that's what we all love so much about Windows. */
+typedef long (*PNTAVM)(HANDLE handle, void **addr, ULONG zbits,
+		       size_t *size, ULONG alloctype, ULONG prot);
+static PNTAVM ntavm;
+
+/* Number of top bits of the lower 32 bits of an address that must be zero.
+** Apparently 0 gives us full 64 bit addresses and 1 gives us the lower 2GB.
+*/
+#define NTAVM_ZEROBITS		1
+
+static void INIT_MMAP(void)
+{
+  ntavm = (PNTAVM)GetProcAddress(GetModuleHandleA("ntdll.dll"),
+				 "NtAllocateVirtualMemory");
+}
+
+/* Win64 32 bit MMAP via NtAllocateVirtualMemory. */
+static LJ_AINLINE void *CALL_MMAP(size_t size)
+{
+  DWORD olderr = GetLastError();
+  void *ptr = NULL;
+  long st = ntavm(INVALID_HANDLE_VALUE, &ptr, NTAVM_ZEROBITS, &size,
+		  MEM_RESERVE|MEM_COMMIT, PAGE_READWRITE);
+  SetLastError(olderr);
+  return st == 0 ? ptr : MFAIL;
+}
+
+/* For direct MMAP, use MEM_TOP_DOWN to minimize interference */
+static LJ_AINLINE void *DIRECT_MMAP(size_t size)
+{
+  DWORD olderr = GetLastError();
+  void *ptr = NULL;
+  long st = ntavm(INVALID_HANDLE_VALUE, &ptr, NTAVM_ZEROBITS, &size,
+		  MEM_RESERVE|MEM_COMMIT|MEM_TOP_DOWN, PAGE_READWRITE);
+  SetLastError(olderr);
+  return st == 0 ? ptr : MFAIL;
+}
+
+#else
+
+#define INIT_MMAP()		((void)0)
+
+/* Win32 MMAP via VirtualAlloc */
+static LJ_AINLINE void *CALL_MMAP(size_t size)
+{
+  DWORD olderr = GetLastError();
+  void *ptr = VirtualAlloc(0, size, MEM_RESERVE|MEM_COMMIT, PAGE_READWRITE);
+  SetLastError(olderr);
+  return ptr ? ptr : MFAIL;
+}
+
+/* For direct MMAP, use MEM_TOP_DOWN to minimize interference */
+static LJ_AINLINE void *DIRECT_MMAP(size_t size)
+{
+  DWORD olderr = GetLastError();
+  void *ptr = VirtualAlloc(0, size, MEM_RESERVE|MEM_COMMIT|MEM_TOP_DOWN,
+			   PAGE_READWRITE);
+  SetLastError(olderr);
+  return ptr ? ptr : MFAIL;
+}
+
+#endif
+
+/* This function supports releasing coalesed segments */
+static LJ_AINLINE int CALL_MUNMAP(void *ptr, size_t size)
+{
+  DWORD olderr = GetLastError();
+  MEMORY_BASIC_INFORMATION minfo;
+  char *cptr = (char *)ptr;
+  while (size) {
+    if (VirtualQuery(cptr, &minfo, sizeof(minfo)) == 0)
+      return -1;
+    if (minfo.BaseAddress != cptr || minfo.AllocationBase != cptr ||
+	minfo.State != MEM_COMMIT || minfo.RegionSize > size)
+      return -1;
+    if (VirtualFree(cptr, 0, MEM_RELEASE) == 0)
+      return -1;
+    cptr += minfo.RegionSize;
+    size -= minfo.RegionSize;
+  }
+  SetLastError(olderr);
+  return 0;
+}
+
+#else
+
+#include <errno.h>
+#include <sys/mman.h>
+
+#define MMAP_PROT		(PROT_READ|PROT_WRITE)
+#if !defined(MAP_ANONYMOUS) && defined(MAP_ANON)
+#define MAP_ANONYMOUS		MAP_ANON
+#endif
+#define MMAP_FLAGS		(MAP_PRIVATE|MAP_ANONYMOUS)
+
+#if LJ_64
+/* 64 bit mode needs special support for allocating memory in the lower 2GB. */
+
+#if defined(MAP_32BIT)
+
+#if defined(__sun__)
+#define MMAP_REGION_START	((uintptr_t)0x1000)
+#else
+/* Actually this only gives us max. 1GB in current Linux kernels. */
+#define MMAP_REGION_START	((uintptr_t)0)
+#endif
+
+static LJ_AINLINE void *CALL_MMAP(size_t size)
+{
+  int olderr = errno;
+  void *ptr = mmap((void *)MMAP_REGION_START, size, MMAP_PROT, MAP_32BIT|MMAP_FLAGS, -1, 0);
+  errno = olderr;
+  return ptr;
+}
+
+#elif LJ_TARGET_OSX || LJ_TARGET_PS4 || defined(__FreeBSD__) || defined(__FreeBSD_kernel__) || defined(__NetBSD__) || defined(__OpenBSD__) || defined(__DragonFly__) || defined(__sun__)
+
+/* OSX and FreeBSD mmap() use a naive first-fit linear search.
+** That's perfect for us. Except that -pagezero_size must be set for OSX,
+** otherwise the lower 4GB are blocked. And the 32GB RLIMIT_DATA needs
+** to be reduced to 250MB on FreeBSD.
+*/
+#if LJ_TARGET_OSX || defined(__DragonFly__)
+#define MMAP_REGION_START	((uintptr_t)0x10000)
+#elif LJ_TARGET_PS4
+#define MMAP_REGION_START	((uintptr_t)0x4000)
+#else
+#define MMAP_REGION_START	((uintptr_t)0x10000000)
+#endif
+#define MMAP_REGION_END		((uintptr_t)0x80000000)
+
+#if (defined(__FreeBSD__) || defined(__FreeBSD_kernel__)) && !LJ_TARGET_PS4
+#include <sys/resource.h>
+#endif
+
+static LJ_AINLINE void *CALL_MMAP(size_t size)
+{
+  int olderr = errno;
+  /* Hint for next allocation. Doesn't need to be thread-safe. */
+  static uintptr_t alloc_hint = MMAP_REGION_START;
+  int retry = 0;
+#if (defined(__FreeBSD__) || defined(__FreeBSD_kernel__)) && !LJ_TARGET_PS4
+  static int rlimit_modified = 0;
+  if (LJ_UNLIKELY(rlimit_modified == 0)) {
+    struct rlimit rlim;
+    rlim.rlim_cur = rlim.rlim_max = MMAP_REGION_START;
+    setrlimit(RLIMIT_DATA, &rlim);  /* Ignore result. May fail below. */
+    rlimit_modified = 1;
+  }
+#endif
+  for (;;) {
+    void *p = mmap((void *)alloc_hint, size, MMAP_PROT, MMAP_FLAGS, -1, 0);
+    if ((uintptr_t)p >= MMAP_REGION_START &&
+	(uintptr_t)p + size < MMAP_REGION_END) {
+      alloc_hint = (uintptr_t)p + size;
+      errno = olderr;
+      return p;
+    }
+    if (p != CMFAIL) munmap(p, size);
+#if defined(__sun__) || defined(__DragonFly__)
+    alloc_hint += 0x1000000;  /* Need near-exhaustive linear scan. */
+    if (alloc_hint + size < MMAP_REGION_END) continue;
+#endif
+    if (retry) break;
+    retry = 1;
+    alloc_hint = MMAP_REGION_START;
+  }
+  errno = olderr;
+  return CMFAIL;
+}
+
+#else
+
+#error "NYI: need an equivalent of MAP_32BIT for this 64 bit OS"
+
+#endif
+
+#else
+
+/* 32 bit mode is easy. */
+static LJ_AINLINE void *CALL_MMAP(size_t size)
+{
+  int olderr = errno;
+  void *ptr = mmap(NULL, size, MMAP_PROT, MMAP_FLAGS, -1, 0);
+  errno = olderr;
+  return ptr;
+}
+
+#endif
+
+#define INIT_MMAP()		((void)0)
+#define DIRECT_MMAP(s)		CALL_MMAP(s)
+
+static LJ_AINLINE int CALL_MUNMAP(void *ptr, size_t size)
+{
+  int olderr = errno;
+  int ret = munmap(ptr, size);
+  errno = olderr;
+  return ret;
+}
+
+#if LJ_TARGET_LINUX
+/* Need to define _GNU_SOURCE to get the mremap prototype. */
+static LJ_AINLINE void *CALL_MREMAP_(void *ptr, size_t osz, size_t nsz,
+				     int flags)
+{
+  int olderr = errno;
+  ptr = mremap(ptr, osz, nsz, flags);
+  errno = olderr;
+  return ptr;
+}
+
+#define CALL_MREMAP(addr, osz, nsz, mv) CALL_MREMAP_((addr), (osz), (nsz), (mv))
+#define CALL_MREMAP_NOMOVE	0
+#define CALL_MREMAP_MAYMOVE	1
+#if LJ_64
+#define CALL_MREMAP_MV		CALL_MREMAP_NOMOVE
+#else
+#define CALL_MREMAP_MV		CALL_MREMAP_MAYMOVE
+#endif
+#endif
+
+#endif
+
+#ifndef CALL_MREMAP
+#define CALL_MREMAP(addr, osz, nsz, mv) ((void)osz, MFAIL)
+#endif
+
+/* -----------------------  Chunk representations ------------------------ */
+
+struct malloc_chunk {
+  size_t               prev_foot;  /* Size of previous chunk (if free).  */
+  size_t               head;       /* Size and inuse bits. */
+  struct malloc_chunk *fd;         /* double links -- used only if free. */
+  struct malloc_chunk *bk;
+};
+
+typedef struct malloc_chunk  mchunk;
+typedef struct malloc_chunk *mchunkptr;
+typedef struct malloc_chunk *sbinptr;  /* The type of bins of chunks */
+typedef size_t bindex_t;               /* Described below */
+typedef unsigned int binmap_t;         /* Described below */
+typedef unsigned int flag_t;           /* The type of various bit flag sets */
+
+/* ------------------- Chunks sizes and alignments ----------------------- */
+
+#define MCHUNK_SIZE		(sizeof(mchunk))
+
+#define CHUNK_OVERHEAD		(SIZE_T_SIZE)
+
+/* Direct chunks need a second word of overhead ... */
+#define DIRECT_CHUNK_OVERHEAD	(TWO_SIZE_T_SIZES)
+/* ... and additional padding for fake next-chunk at foot */
+#define DIRECT_FOOT_PAD		(FOUR_SIZE_T_SIZES)
+
+/* The smallest size we can malloc is an aligned minimal chunk */
+#define MIN_CHUNK_SIZE\
+  ((MCHUNK_SIZE + CHUNK_ALIGN_MASK) & ~CHUNK_ALIGN_MASK)
+
+/* conversion from malloc headers to user pointers, and back */
+#define chunk2mem(p)		((void *)((char *)(p) + TWO_SIZE_T_SIZES))
+#define mem2chunk(mem)		((mchunkptr)((char *)(mem) - TWO_SIZE_T_SIZES))
+/* chunk associated with aligned address A */
+#define align_as_chunk(A)	(mchunkptr)((A) + align_offset(chunk2mem(A)))
+
+/* Bounds on request (not chunk) sizes. */
+#define MAX_REQUEST		((~MIN_CHUNK_SIZE+1) << 2)
+#define MIN_REQUEST		(MIN_CHUNK_SIZE - CHUNK_OVERHEAD - SIZE_T_ONE)
+
+/* pad request bytes into a usable size */
+#define pad_request(req) \
+   (((req) + CHUNK_OVERHEAD + CHUNK_ALIGN_MASK) & ~CHUNK_ALIGN_MASK)
+
+/* pad request, checking for minimum (but not maximum) */
+#define request2size(req) \
+  (((req) < MIN_REQUEST)? MIN_CHUNK_SIZE : pad_request(req))
+
+/* ------------------ Operations on head and foot fields ----------------- */
+
+#define PINUSE_BIT		(SIZE_T_ONE)
+#define CINUSE_BIT		(SIZE_T_TWO)
+#define INUSE_BITS		(PINUSE_BIT|CINUSE_BIT)
+
+/* Head value for fenceposts */
+#define FENCEPOST_HEAD		(INUSE_BITS|SIZE_T_SIZE)
+
+/* extraction of fields from head words */
+#define cinuse(p)		((p)->head & CINUSE_BIT)
+#define pinuse(p)		((p)->head & PINUSE_BIT)
+#define chunksize(p)		((p)->head & ~(INUSE_BITS))
+
+#define clear_pinuse(p)		((p)->head &= ~PINUSE_BIT)
+#define clear_cinuse(p)		((p)->head &= ~CINUSE_BIT)
+
+/* Treat space at ptr +/- offset as a chunk */
+#define chunk_plus_offset(p, s)		((mchunkptr)(((char *)(p)) + (s)))
+#define chunk_minus_offset(p, s)	((mchunkptr)(((char *)(p)) - (s)))
+
+/* Ptr to next or previous physical malloc_chunk. */
+#define next_chunk(p)	((mchunkptr)(((char *)(p)) + ((p)->head & ~INUSE_BITS)))
+#define prev_chunk(p)	((mchunkptr)(((char *)(p)) - ((p)->prev_foot) ))
+
+/* extract next chunk's pinuse bit */
+#define next_pinuse(p)	((next_chunk(p)->head) & PINUSE_BIT)
+
+/* Get/set size at footer */
+#define get_foot(p, s)	(((mchunkptr)((char *)(p) + (s)))->prev_foot)
+#define set_foot(p, s)	(((mchunkptr)((char *)(p) + (s)))->prev_foot = (s))
+
+/* Set size, pinuse bit, and foot */
+#define set_size_and_pinuse_of_free_chunk(p, s)\
+  ((p)->head = (s|PINUSE_BIT), set_foot(p, s))
+
+/* Set size, pinuse bit, foot, and clear next pinuse */
+#define set_free_with_pinuse(p, s, n)\
+  (clear_pinuse(n), set_size_and_pinuse_of_free_chunk(p, s))
+
+#define is_direct(p)\
+  (!((p)->head & PINUSE_BIT) && ((p)->prev_foot & IS_DIRECT_BIT))
+
+/* Get the internal overhead associated with chunk p */
+#define overhead_for(p)\
+ (is_direct(p)? DIRECT_CHUNK_OVERHEAD : CHUNK_OVERHEAD)
+
+/* ---------------------- Overlaid data structures ----------------------- */
+
+struct malloc_tree_chunk {
+  /* The first four fields must be compatible with malloc_chunk */
+  size_t                    prev_foot;
+  size_t                    head;
+  struct malloc_tree_chunk *fd;
+  struct malloc_tree_chunk *bk;
+
+  struct malloc_tree_chunk *child[2];
+  struct malloc_tree_chunk *parent;
+  bindex_t                  index;
+};
+
+typedef struct malloc_tree_chunk  tchunk;
+typedef struct malloc_tree_chunk *tchunkptr;
+typedef struct malloc_tree_chunk *tbinptr; /* The type of bins of trees */
+
+/* A little helper macro for trees */
+#define leftmost_child(t) ((t)->child[0] != 0? (t)->child[0] : (t)->child[1])
+
+/* ----------------------------- Segments -------------------------------- */
+
+struct malloc_segment {
+  char        *base;             /* base address */
+  size_t       size;             /* allocated size */
+  struct malloc_segment *next;   /* ptr to next segment */
+};
+
+typedef struct malloc_segment  msegment;
+typedef struct malloc_segment *msegmentptr;
+
+/* ---------------------------- malloc_state ----------------------------- */
+
+/* Bin types, widths and sizes */
+#define NSMALLBINS		(32U)
+#define NTREEBINS		(32U)
+#define SMALLBIN_SHIFT		(3U)
+#define SMALLBIN_WIDTH		(SIZE_T_ONE << SMALLBIN_SHIFT)
+#define TREEBIN_SHIFT		(8U)
+#define MIN_LARGE_SIZE		(SIZE_T_ONE << TREEBIN_SHIFT)
+#define MAX_SMALL_SIZE		(MIN_LARGE_SIZE - SIZE_T_ONE)
+#define MAX_SMALL_REQUEST  (MAX_SMALL_SIZE - CHUNK_ALIGN_MASK - CHUNK_OVERHEAD)
+
+struct malloc_state {
+  binmap_t   smallmap;
+  binmap_t   treemap;
+  size_t     dvsize;
+  size_t     topsize;
+  mchunkptr  dv;
+  mchunkptr  top;
+  size_t     trim_check;
+  size_t     release_checks;
+  mchunkptr  smallbins[(NSMALLBINS+1)*2];
+  tbinptr    treebins[NTREEBINS];
+  msegment   seg;
+};
+
+typedef struct malloc_state *mstate;
+
+#define is_initialized(M)	((M)->top != 0)
+
+/* -------------------------- system alloc setup ------------------------- */
+
+/* page-align a size */
+#define page_align(S)\
+ (((S) + (LJ_PAGESIZE - SIZE_T_ONE)) & ~(LJ_PAGESIZE - SIZE_T_ONE))
+
+/* granularity-align a size */
+#define granularity_align(S)\
+  (((S) + (DEFAULT_GRANULARITY - SIZE_T_ONE))\
+   & ~(DEFAULT_GRANULARITY - SIZE_T_ONE))
+
+#if LJ_TARGET_WINDOWS
+#define mmap_align(S)	granularity_align(S)
+#else
+#define mmap_align(S)	page_align(S)
+#endif
+
+/*  True if segment S holds address A */
+#define segment_holds(S, A)\
+  ((char *)(A) >= S->base && (char *)(A) < S->base + S->size)
+
+/* Return segment holding given address */
+static msegmentptr segment_holding(mstate m, char *addr)
+{
+  msegmentptr sp = &m->seg;
+  for (;;) {
+    if (addr >= sp->base && addr < sp->base + sp->size)
+      return sp;
+    if ((sp = sp->next) == 0)
+      return 0;
+  }
+}
+
+/* Return true if segment contains a segment link */
+static int has_segment_link(mstate m, msegmentptr ss)
+{
+  msegmentptr sp = &m->seg;
+  for (;;) {
+    if ((char *)sp >= ss->base && (char *)sp < ss->base + ss->size)
+      return 1;
+    if ((sp = sp->next) == 0)
+      return 0;
+  }
+}
+
+/*
+  TOP_FOOT_SIZE is padding at the end of a segment, including space
+  that may be needed to place segment records and fenceposts when new
+  noncontiguous segments are added.
+*/
+#define TOP_FOOT_SIZE\
+  (align_offset(chunk2mem(0))+pad_request(sizeof(struct malloc_segment))+MIN_CHUNK_SIZE)
+
+/* ---------------------------- Indexing Bins ---------------------------- */
+
+#define is_small(s)		(((s) >> SMALLBIN_SHIFT) < NSMALLBINS)
+#define small_index(s)		((s)  >> SMALLBIN_SHIFT)
+#define small_index2size(i)	((i)  << SMALLBIN_SHIFT)
+#define MIN_SMALL_INDEX		(small_index(MIN_CHUNK_SIZE))
+
+/* addressing by index. See above about smallbin repositioning */
+#define smallbin_at(M, i)	((sbinptr)((char *)&((M)->smallbins[(i)<<1])))
+#define treebin_at(M,i)		(&((M)->treebins[i]))
+
+/* assign tree index for size S to variable I */
+#define compute_tree_index(S, I)\
+{\
+  unsigned int X = (unsigned int)(S >> TREEBIN_SHIFT);\
+  if (X == 0) {\
+    I = 0;\
+  } else if (X > 0xFFFF) {\
+    I = NTREEBINS-1;\
+  } else {\
+    unsigned int K = lj_fls(X);\
+    I =  (bindex_t)((K << 1) + ((S >> (K + (TREEBIN_SHIFT-1)) & 1)));\
+  }\
+}
+
+/* Bit representing maximum resolved size in a treebin at i */
+#define bit_for_tree_index(i) \
+   (i == NTREEBINS-1)? (SIZE_T_BITSIZE-1) : (((i) >> 1) + TREEBIN_SHIFT - 2)
+
+/* Shift placing maximum resolved bit in a treebin at i as sign bit */
+#define leftshift_for_tree_index(i) \
+   ((i == NTREEBINS-1)? 0 : \
+    ((SIZE_T_BITSIZE-SIZE_T_ONE) - (((i) >> 1) + TREEBIN_SHIFT - 2)))
+
+/* The size of the smallest chunk held in bin with index i */
+#define minsize_for_tree_index(i) \
+   ((SIZE_T_ONE << (((i) >> 1) + TREEBIN_SHIFT)) |  \
+   (((size_t)((i) & SIZE_T_ONE)) << (((i) >> 1) + TREEBIN_SHIFT - 1)))
+
+/* ------------------------ Operations on bin maps ----------------------- */
+
+/* bit corresponding to given index */
+#define idx2bit(i)		((binmap_t)(1) << (i))
+
+/* Mark/Clear bits with given index */
+#define mark_smallmap(M,i)	((M)->smallmap |=  idx2bit(i))
+#define clear_smallmap(M,i)	((M)->smallmap &= ~idx2bit(i))
+#define smallmap_is_marked(M,i)	((M)->smallmap &   idx2bit(i))
+
+#define mark_treemap(M,i)	((M)->treemap  |=  idx2bit(i))
+#define clear_treemap(M,i)	((M)->treemap  &= ~idx2bit(i))
+#define treemap_is_marked(M,i)	((M)->treemap  &   idx2bit(i))
+
+/* mask with all bits to left of least bit of x on */
+#define left_bits(x)		((x<<1) | (~(x<<1)+1))
+
+/* Set cinuse bit and pinuse bit of next chunk */
+#define set_inuse(M,p,s)\
+  ((p)->head = (((p)->head & PINUSE_BIT)|s|CINUSE_BIT),\
+  ((mchunkptr)(((char *)(p)) + (s)))->head |= PINUSE_BIT)
+
+/* Set cinuse and pinuse of this chunk and pinuse of next chunk */
+#define set_inuse_and_pinuse(M,p,s)\
+  ((p)->head = (s|PINUSE_BIT|CINUSE_BIT),\
+  ((mchunkptr)(((char *)(p)) + (s)))->head |= PINUSE_BIT)
+
+/* Set size, cinuse and pinuse bit of this chunk */
+#define set_size_and_pinuse_of_inuse_chunk(M, p, s)\
+  ((p)->head = (s|PINUSE_BIT|CINUSE_BIT))
+
+/* ----------------------- Operations on smallbins ----------------------- */
+
+/* Link a free chunk into a smallbin  */
+#define insert_small_chunk(M, P, S) {\
+  bindex_t I = small_index(S);\
+  mchunkptr B = smallbin_at(M, I);\
+  mchunkptr F = B;\
+  if (!smallmap_is_marked(M, I))\
+    mark_smallmap(M, I);\
+  else\
+    F = B->fd;\
+  B->fd = P;\
+  F->bk = P;\
+  P->fd = F;\
+  P->bk = B;\
+}
+
+/* Unlink a chunk from a smallbin  */
+#define unlink_small_chunk(M, P, S) {\
+  mchunkptr F = P->fd;\
+  mchunkptr B = P->bk;\
+  bindex_t I = small_index(S);\
+  if (F == B) {\
+    clear_smallmap(M, I);\
+  } else {\
+    F->bk = B;\
+    B->fd = F;\
+  }\
+}
+
+/* Unlink the first chunk from a smallbin */
+#define unlink_first_small_chunk(M, B, P, I) {\
+  mchunkptr F = P->fd;\
+  if (B == F) {\
+    clear_smallmap(M, I);\
+  } else {\
+    B->fd = F;\
+    F->bk = B;\
+  }\
+}
+
+/* Replace dv node, binning the old one */
+/* Used only when dvsize known to be small */
+#define replace_dv(M, P, S) {\
+  size_t DVS = M->dvsize;\
+  if (DVS != 0) {\
+    mchunkptr DV = M->dv;\
+    insert_small_chunk(M, DV, DVS);\
+  }\
+  M->dvsize = S;\
+  M->dv = P;\
+}
+
+/* ------------------------- Operations on trees ------------------------- */
+
+/* Insert chunk into tree */
+#define insert_large_chunk(M, X, S) {\
+  tbinptr *H;\
+  bindex_t I;\
+  compute_tree_index(S, I);\
+  H = treebin_at(M, I);\
+  X->index = I;\
+  X->child[0] = X->child[1] = 0;\
+  if (!treemap_is_marked(M, I)) {\
+    mark_treemap(M, I);\
+    *H = X;\
+    X->parent = (tchunkptr)H;\
+    X->fd = X->bk = X;\
+  } else {\
+    tchunkptr T = *H;\
+    size_t K = S << leftshift_for_tree_index(I);\
+    for (;;) {\
+      if (chunksize(T) != S) {\
+	tchunkptr *C = &(T->child[(K >> (SIZE_T_BITSIZE-SIZE_T_ONE)) & 1]);\
+	K <<= 1;\
+	if (*C != 0) {\
+	  T = *C;\
+	} else {\
+	  *C = X;\
+	  X->parent = T;\
+	  X->fd = X->bk = X;\
+	  break;\
+	}\
+      } else {\
+	tchunkptr F = T->fd;\
+	T->fd = F->bk = X;\
+	X->fd = F;\
+	X->bk = T;\
+	X->parent = 0;\
+	break;\
+      }\
+    }\
+  }\
+}
+
+#define unlink_large_chunk(M, X) {\
+  tchunkptr XP = X->parent;\
+  tchunkptr R;\
+  if (X->bk != X) {\
+    tchunkptr F = X->fd;\
+    R = X->bk;\
+    F->bk = R;\
+    R->fd = F;\
+  } else {\
+    tchunkptr *RP;\
+    if (((R = *(RP = &(X->child[1]))) != 0) ||\
+	((R = *(RP = &(X->child[0]))) != 0)) {\
+      tchunkptr *CP;\
+      while ((*(CP = &(R->child[1])) != 0) ||\
+	     (*(CP = &(R->child[0])) != 0)) {\
+	R = *(RP = CP);\
+      }\
+      *RP = 0;\
+    }\
+  }\
+  if (XP != 0) {\
+    tbinptr *H = treebin_at(M, X->index);\
+    if (X == *H) {\
+      if ((*H = R) == 0) \
+	clear_treemap(M, X->index);\
+    } else {\
+      if (XP->child[0] == X) \
+	XP->child[0] = R;\
+      else \
+	XP->child[1] = R;\
+    }\
+    if (R != 0) {\
+      tchunkptr C0, C1;\
+      R->parent = XP;\
+      if ((C0 = X->child[0]) != 0) {\
+	R->child[0] = C0;\
+	C0->parent = R;\
+      }\
+      if ((C1 = X->child[1]) != 0) {\
+	R->child[1] = C1;\
+	C1->parent = R;\
+      }\
+    }\
+  }\
+}
+
+/* Relays to large vs small bin operations */
+
+#define insert_chunk(M, P, S)\
+  if (is_small(S)) { insert_small_chunk(M, P, S)\
+  } else { tchunkptr TP = (tchunkptr)(P); insert_large_chunk(M, TP, S); }
+
+#define unlink_chunk(M, P, S)\
+  if (is_small(S)) { unlink_small_chunk(M, P, S)\
+  } else { tchunkptr TP = (tchunkptr)(P); unlink_large_chunk(M, TP); }
+
+/* -----------------------  Direct-mmapping chunks ----------------------- */
+
+static void *direct_alloc(size_t nb)
+{
+  size_t mmsize = mmap_align(nb + SIX_SIZE_T_SIZES + CHUNK_ALIGN_MASK);
+  if (LJ_LIKELY(mmsize > nb)) {     /* Check for wrap around 0 */
+    char *mm = (char *)(DIRECT_MMAP(mmsize));
+    if (mm != CMFAIL) {
+      size_t offset = align_offset(chunk2mem(mm));
+      size_t psize = mmsize - offset - DIRECT_FOOT_PAD;
+      mchunkptr p = (mchunkptr)(mm + offset);
+      p->prev_foot = offset | IS_DIRECT_BIT;
+      p->head = psize|CINUSE_BIT;
+      chunk_plus_offset(p, psize)->head = FENCEPOST_HEAD;
+      chunk_plus_offset(p, psize+SIZE_T_SIZE)->head = 0;
+      return chunk2mem(p);
+    }
+  }
+  return NULL;
+}
+
+static mchunkptr direct_resize(mchunkptr oldp, size_t nb)
+{
+  size_t oldsize = chunksize(oldp);
+  if (is_small(nb)) /* Can't shrink direct regions below small size */
+    return NULL;
+  /* Keep old chunk if big enough but not too big */
+  if (oldsize >= nb + SIZE_T_SIZE &&
+      (oldsize - nb) <= (DEFAULT_GRANULARITY >> 1)) {
+    return oldp;
+  } else {
+    size_t offset = oldp->prev_foot & ~IS_DIRECT_BIT;
+    size_t oldmmsize = oldsize + offset + DIRECT_FOOT_PAD;
+    size_t newmmsize = mmap_align(nb + SIX_SIZE_T_SIZES + CHUNK_ALIGN_MASK);
+    char *cp = (char *)CALL_MREMAP((char *)oldp - offset,
+				   oldmmsize, newmmsize, CALL_MREMAP_MV);
+    if (cp != CMFAIL) {
+      mchunkptr newp = (mchunkptr)(cp + offset);
+      size_t psize = newmmsize - offset - DIRECT_FOOT_PAD;
+      newp->head = psize|CINUSE_BIT;
+      chunk_plus_offset(newp, psize)->head = FENCEPOST_HEAD;
+      chunk_plus_offset(newp, psize+SIZE_T_SIZE)->head = 0;
+      return newp;
+    }
+  }
+  return NULL;
+}
+
+/* -------------------------- mspace management -------------------------- */
+
+/* Initialize top chunk and its size */
+static void init_top(mstate m, mchunkptr p, size_t psize)
+{
+  /* Ensure alignment */
+  size_t offset = align_offset(chunk2mem(p));
+  p = (mchunkptr)((char *)p + offset);
+  psize -= offset;
+
+  m->top = p;
+  m->topsize = psize;
+  p->head = psize | PINUSE_BIT;
+  /* set size of fake trailing chunk holding overhead space only once */
+  chunk_plus_offset(p, psize)->head = TOP_FOOT_SIZE;
+  m->trim_check = DEFAULT_TRIM_THRESHOLD; /* reset on each update */
+}
+
+/* Initialize bins for a new mstate that is otherwise zeroed out */
+static void init_bins(mstate m)
+{
+  /* Establish circular links for smallbins */
+  bindex_t i;
+  for (i = 0; i < NSMALLBINS; i++) {
+    sbinptr bin = smallbin_at(m,i);
+    bin->fd = bin->bk = bin;
+  }
+}
+
+/* Allocate chunk and prepend remainder with chunk in successor base. */
+static void *prepend_alloc(mstate m, char *newbase, char *oldbase, size_t nb)
+{
+  mchunkptr p = align_as_chunk(newbase);
+  mchunkptr oldfirst = align_as_chunk(oldbase);
+  size_t psize = (size_t)((char *)oldfirst - (char *)p);
+  mchunkptr q = chunk_plus_offset(p, nb);
+  size_t qsize = psize - nb;
+  set_size_and_pinuse_of_inuse_chunk(m, p, nb);
+
+  /* consolidate remainder with first chunk of old base */
+  if (oldfirst == m->top) {
+    size_t tsize = m->topsize += qsize;
+    m->top = q;
+    q->head = tsize | PINUSE_BIT;
+  } else if (oldfirst == m->dv) {
+    size_t dsize = m->dvsize += qsize;
+    m->dv = q;
+    set_size_and_pinuse_of_free_chunk(q, dsize);
+  } else {
+    if (!cinuse(oldfirst)) {
+      size_t nsize = chunksize(oldfirst);
+      unlink_chunk(m, oldfirst, nsize);
+      oldfirst = chunk_plus_offset(oldfirst, nsize);
+      qsize += nsize;
+    }
+    set_free_with_pinuse(q, qsize, oldfirst);
+    insert_chunk(m, q, qsize);
+  }
+
+  return chunk2mem(p);
+}
+
+/* Add a segment to hold a new noncontiguous region */
+static void add_segment(mstate m, char *tbase, size_t tsize)
+{
+  /* Determine locations and sizes of segment, fenceposts, old top */
+  char *old_top = (char *)m->top;
+  msegmentptr oldsp = segment_holding(m, old_top);
+  char *old_end = oldsp->base + oldsp->size;
+  size_t ssize = pad_request(sizeof(struct malloc_segment));
+  char *rawsp = old_end - (ssize + FOUR_SIZE_T_SIZES + CHUNK_ALIGN_MASK);
+  size_t offset = align_offset(chunk2mem(rawsp));
+  char *asp = rawsp + offset;
+  char *csp = (asp < (old_top + MIN_CHUNK_SIZE))? old_top : asp;
+  mchunkptr sp = (mchunkptr)csp;
+  msegmentptr ss = (msegmentptr)(chunk2mem(sp));
+  mchunkptr tnext = chunk_plus_offset(sp, ssize);
+  mchunkptr p = tnext;
+
+  /* reset top to new space */
+  init_top(m, (mchunkptr)tbase, tsize - TOP_FOOT_SIZE);
+
+  /* Set up segment record */
+  set_size_and_pinuse_of_inuse_chunk(m, sp, ssize);
+  *ss = m->seg; /* Push current record */
+  m->seg.base = tbase;
+  m->seg.size = tsize;
+  m->seg.next = ss;
+
+  /* Insert trailing fenceposts */
+  for (;;) {
+    mchunkptr nextp = chunk_plus_offset(p, SIZE_T_SIZE);
+    p->head = FENCEPOST_HEAD;
+    if ((char *)(&(nextp->head)) < old_end)
+      p = nextp;
+    else
+      break;
+  }
+
+  /* Insert the rest of old top into a bin as an ordinary free chunk */
+  if (csp != old_top) {
+    mchunkptr q = (mchunkptr)old_top;
+    size_t psize = (size_t)(csp - old_top);
+    mchunkptr tn = chunk_plus_offset(q, psize);
+    set_free_with_pinuse(q, psize, tn);
+    insert_chunk(m, q, psize);
+  }
+}
+
+/* -------------------------- System allocation -------------------------- */
+
+static void *alloc_sys(mstate m, size_t nb)
+{
+  char *tbase = CMFAIL;
+  size_t tsize = 0;
+
+  /* Directly map large chunks */
+  if (LJ_UNLIKELY(nb >= DEFAULT_MMAP_THRESHOLD)) {
+    void *mem = direct_alloc(nb);
+    if (mem != 0)
+      return mem;
+  }
+
+  {
+    size_t req = nb + TOP_FOOT_SIZE + SIZE_T_ONE;
+    size_t rsize = granularity_align(req);
+    if (LJ_LIKELY(rsize > nb)) { /* Fail if wraps around zero */
+      char *mp = (char *)(CALL_MMAP(rsize));
+      if (mp != CMFAIL) {
+	tbase = mp;
+	tsize = rsize;
+      }
+    }
+  }
+
+  if (tbase != CMFAIL) {
+    msegmentptr sp = &m->seg;
+    /* Try to merge with an existing segment */
+    while (sp != 0 && tbase != sp->base + sp->size)
+      sp = sp->next;
+    if (sp != 0 && segment_holds(sp, m->top)) { /* append */
+      sp->size += tsize;
+      init_top(m, m->top, m->topsize + tsize);
+    } else {
+      sp = &m->seg;
+      while (sp != 0 && sp->base != tbase + tsize)
+	sp = sp->next;
+      if (sp != 0) {
+	char *oldbase = sp->base;
+	sp->base = tbase;
+	sp->size += tsize;
+	return prepend_alloc(m, tbase, oldbase, nb);
+      } else {
+	add_segment(m, tbase, tsize);
+      }
+    }
+
+    if (nb < m->topsize) { /* Allocate from new or extended top space */
+      size_t rsize = m->topsize -= nb;
+      mchunkptr p = m->top;
+      mchunkptr r = m->top = chunk_plus_offset(p, nb);
+      r->head = rsize | PINUSE_BIT;
+      set_size_and_pinuse_of_inuse_chunk(m, p, nb);
+      return chunk2mem(p);
+    }
+  }
+
+  return NULL;
+}
+
+/* -----------------------  system deallocation -------------------------- */
+
+/* Unmap and unlink any mmapped segments that don't contain used chunks */
+static size_t release_unused_segments(mstate m)
+{
+  size_t released = 0;
+  size_t nsegs = 0;
+  msegmentptr pred = &m->seg;
+  msegmentptr sp = pred->next;
+  while (sp != 0) {
+    char *base = sp->base;
+    size_t size = sp->size;
+    msegmentptr next = sp->next;
+    nsegs++;
+    {
+      mchunkptr p = align_as_chunk(base);
+      size_t psize = chunksize(p);
+      /* Can unmap if first chunk holds entire segment and not pinned */
+      if (!cinuse(p) && (char *)p + psize >= base + size - TOP_FOOT_SIZE) {
+	tchunkptr tp = (tchunkptr)p;
+	if (p == m->dv) {
+	  m->dv = 0;
+	  m->dvsize = 0;
+	} else {
+	  unlink_large_chunk(m, tp);
+	}
+	if (CALL_MUNMAP(base, size) == 0) {
+	  released += size;
+	  /* unlink obsoleted record */
+	  sp = pred;
+	  sp->next = next;
+	} else { /* back out if cannot unmap */
+	  insert_large_chunk(m, tp, psize);
+	}
+      }
+    }
+    pred = sp;
+    sp = next;
+  }
+  /* Reset check counter */
+  m->release_checks = nsegs > MAX_RELEASE_CHECK_RATE ?
+		      nsegs : MAX_RELEASE_CHECK_RATE;
+  return released;
+}
+
+static int alloc_trim(mstate m, size_t pad)
+{
+  size_t released = 0;
+  if (pad < MAX_REQUEST && is_initialized(m)) {
+    pad += TOP_FOOT_SIZE; /* ensure enough room for segment overhead */
+
+    if (m->topsize > pad) {
+      /* Shrink top space in granularity-size units, keeping at least one */
+      size_t unit = DEFAULT_GRANULARITY;
+      size_t extra = ((m->topsize - pad + (unit - SIZE_T_ONE)) / unit -
+		      SIZE_T_ONE) * unit;
+      msegmentptr sp = segment_holding(m, (char *)m->top);
+
+      if (sp->size >= extra &&
+	  !has_segment_link(m, sp)) { /* can't shrink if pinned */
+	size_t newsize = sp->size - extra;
+	/* Prefer mremap, fall back to munmap */
+	if ((CALL_MREMAP(sp->base, sp->size, newsize, CALL_MREMAP_NOMOVE) != MFAIL) ||
+	    (CALL_MUNMAP(sp->base + newsize, extra) == 0)) {
+	  released = extra;
+	}
+      }
+
+      if (released != 0) {
+	sp->size -= released;
+	init_top(m, m->top, m->topsize - released);
+      }
+    }
+
+    /* Unmap any unused mmapped segments */
+    released += release_unused_segments(m);
+
+    /* On failure, disable autotrim to avoid repeated failed future calls */
+    if (released == 0 && m->topsize > m->trim_check)
+      m->trim_check = MAX_SIZE_T;
+  }
+
+  return (released != 0)? 1 : 0;
+}
+
+/* ---------------------------- malloc support --------------------------- */
+
+/* allocate a large request from the best fitting chunk in a treebin */
+static void *tmalloc_large(mstate m, size_t nb)
+{
+  tchunkptr v = 0;
+  size_t rsize = ~nb+1; /* Unsigned negation */
+  tchunkptr t;
+  bindex_t idx;
+  compute_tree_index(nb, idx);
+
+  if ((t = *treebin_at(m, idx)) != 0) {
+    /* Traverse tree for this bin looking for node with size == nb */
+    size_t sizebits = nb << leftshift_for_tree_index(idx);
+    tchunkptr rst = 0;  /* The deepest untaken right subtree */
+    for (;;) {
+      tchunkptr rt;
+      size_t trem = chunksize(t) - nb;
+      if (trem < rsize) {
+	v = t;
+	if ((rsize = trem) == 0)
+	  break;
+      }
+      rt = t->child[1];
+      t = t->child[(sizebits >> (SIZE_T_BITSIZE-SIZE_T_ONE)) & 1];
+      if (rt != 0 && rt != t)
+	rst = rt;
+      if (t == 0) {
+	t = rst; /* set t to least subtree holding sizes > nb */
+	break;
+      }
+      sizebits <<= 1;
+    }
+  }
+
+  if (t == 0 && v == 0) { /* set t to root of next non-empty treebin */
+    binmap_t leftbits = left_bits(idx2bit(idx)) & m->treemap;
+    if (leftbits != 0)
+      t = *treebin_at(m, lj_ffs(leftbits));
+  }
+
+  while (t != 0) { /* find smallest of tree or subtree */
+    size_t trem = chunksize(t) - nb;
+    if (trem < rsize) {
+      rsize = trem;
+      v = t;
+    }
+    t = leftmost_child(t);
+  }
+
+  /*  If dv is a better fit, return NULL so malloc will use it */
+  if (v != 0 && rsize < (size_t)(m->dvsize - nb)) {
+    mchunkptr r = chunk_plus_offset(v, nb);
+    unlink_large_chunk(m, v);
+    if (rsize < MIN_CHUNK_SIZE) {
+      set_inuse_and_pinuse(m, v, (rsize + nb));
+    } else {
+      set_size_and_pinuse_of_inuse_chunk(m, v, nb);
+      set_size_and_pinuse_of_free_chunk(r, rsize);
+      insert_chunk(m, r, rsize);
+    }
+    return chunk2mem(v);
+  }
+  return NULL;
+}
+
+/* allocate a small request from the best fitting chunk in a treebin */
+static void *tmalloc_small(mstate m, size_t nb)
+{
+  tchunkptr t, v;
+  mchunkptr r;
+  size_t rsize;
+  bindex_t i = lj_ffs(m->treemap);
+
+  v = t = *treebin_at(m, i);
+  rsize = chunksize(t) - nb;
+
+  while ((t = leftmost_child(t)) != 0) {
+    size_t trem = chunksize(t) - nb;
+    if (trem < rsize) {
+      rsize = trem;
+      v = t;
+    }
+  }
+
+  r = chunk_plus_offset(v, nb);
+  unlink_large_chunk(m, v);
+  if (rsize < MIN_CHUNK_SIZE) {
+    set_inuse_and_pinuse(m, v, (rsize + nb));
+  } else {
+    set_size_and_pinuse_of_inuse_chunk(m, v, nb);
+    set_size_and_pinuse_of_free_chunk(r, rsize);
+    replace_dv(m, r, rsize);
+  }
+  return chunk2mem(v);
+}
+
+/* ----------------------------------------------------------------------- */
+
+void *lj_alloc_create(void)
+{
+  size_t tsize = DEFAULT_GRANULARITY;
+  char *tbase;
+  INIT_MMAP();
+  tbase = (char *)(CALL_MMAP(tsize));
+  if (tbase != CMFAIL) {
+    size_t msize = pad_request(sizeof(struct malloc_state));
+    mchunkptr mn;
+    mchunkptr msp = align_as_chunk(tbase);
+    mstate m = (mstate)(chunk2mem(msp));
+    memset(m, 0, msize);
+    msp->head = (msize|PINUSE_BIT|CINUSE_BIT);
+    m->seg.base = tbase;
+    m->seg.size = tsize;
+    m->release_checks = MAX_RELEASE_CHECK_RATE;
+    init_bins(m);
+    mn = next_chunk(mem2chunk(m));
+    init_top(m, mn, (size_t)((tbase + tsize) - (char *)mn) - TOP_FOOT_SIZE);
+    return m;
+  }
+  return NULL;
+}
+
+void lj_alloc_destroy(void *msp)
+{
+  mstate ms = (mstate)msp;
+  msegmentptr sp = &ms->seg;
+  while (sp != 0) {
+    char *base = sp->base;
+    size_t size = sp->size;
+    sp = sp->next;
+    CALL_MUNMAP(base, size);
+  }
+}
+
+static LJ_NOINLINE void *lj_alloc_malloc(void *msp, size_t nsize)
+{
+  mstate ms = (mstate)msp;
+  void *mem;
+  size_t nb;
+  if (nsize <= MAX_SMALL_REQUEST) {
+    bindex_t idx;
+    binmap_t smallbits;
+    nb = (nsize < MIN_REQUEST)? MIN_CHUNK_SIZE : pad_request(nsize);
+    idx = small_index(nb);
+    smallbits = ms->smallmap >> idx;
+
+    if ((smallbits & 0x3U) != 0) { /* Remainderless fit to a smallbin. */
+      mchunkptr b, p;
+      idx += ~smallbits & 1;       /* Uses next bin if idx empty */
+      b = smallbin_at(ms, idx);
+      p = b->fd;
+      unlink_first_small_chunk(ms, b, p, idx);
+      set_inuse_and_pinuse(ms, p, small_index2size(idx));
+      mem = chunk2mem(p);
+      return mem;
+    } else if (nb > ms->dvsize) {
+      if (smallbits != 0) { /* Use chunk in next nonempty smallbin */
+	mchunkptr b, p, r;
+	size_t rsize;
+	binmap_t leftbits = (smallbits << idx) & left_bits(idx2bit(idx));
+	bindex_t i = lj_ffs(leftbits);
+	b = smallbin_at(ms, i);
+	p = b->fd;
+	unlink_first_small_chunk(ms, b, p, i);
+	rsize = small_index2size(i) - nb;
+	/* Fit here cannot be remainderless if 4byte sizes */
+	if (SIZE_T_SIZE != 4 && rsize < MIN_CHUNK_SIZE) {
+	  set_inuse_and_pinuse(ms, p, small_index2size(i));
+	} else {
+	  set_size_and_pinuse_of_inuse_chunk(ms, p, nb);
+	  r = chunk_plus_offset(p, nb);
+	  set_size_and_pinuse_of_free_chunk(r, rsize);
+	  replace_dv(ms, r, rsize);
+	}
+	mem = chunk2mem(p);
+	return mem;
+      } else if (ms->treemap != 0 && (mem = tmalloc_small(ms, nb)) != 0) {
+	return mem;
+      }
+    }
+  } else if (nsize >= MAX_REQUEST) {
+    nb = MAX_SIZE_T; /* Too big to allocate. Force failure (in sys alloc) */
+  } else {
+    nb = pad_request(nsize);
+    if (ms->treemap != 0 && (mem = tmalloc_large(ms, nb)) != 0) {
+      return mem;
+    }
+  }
+
+  if (nb <= ms->dvsize) {
+    size_t rsize = ms->dvsize - nb;
+    mchunkptr p = ms->dv;
+    if (rsize >= MIN_CHUNK_SIZE) { /* split dv */
+      mchunkptr r = ms->dv = chunk_plus_offset(p, nb);
+      ms->dvsize = rsize;
+      set_size_and_pinuse_of_free_chunk(r, rsize);
+      set_size_and_pinuse_of_inuse_chunk(ms, p, nb);
+    } else { /* exhaust dv */
+      size_t dvs = ms->dvsize;
+      ms->dvsize = 0;
+      ms->dv = 0;
+      set_inuse_and_pinuse(ms, p, dvs);
+    }
+    mem = chunk2mem(p);
+    return mem;
+  } else if (nb < ms->topsize) { /* Split top */
+    size_t rsize = ms->topsize -= nb;
+    mchunkptr p = ms->top;
+    mchunkptr r = ms->top = chunk_plus_offset(p, nb);
+    r->head = rsize | PINUSE_BIT;
+    set_size_and_pinuse_of_inuse_chunk(ms, p, nb);
+    mem = chunk2mem(p);
+    return mem;
+  }
+  return alloc_sys(ms, nb);
+}
+
+static LJ_NOINLINE void *lj_alloc_free(void *msp, void *ptr)
+{
+  if (ptr != 0) {
+    mchunkptr p = mem2chunk(ptr);
+    mstate fm = (mstate)msp;
+    size_t psize = chunksize(p);
+    mchunkptr next = chunk_plus_offset(p, psize);
+    if (!pinuse(p)) {
+      size_t prevsize = p->prev_foot;
+      if ((prevsize & IS_DIRECT_BIT) != 0) {
+	prevsize &= ~IS_DIRECT_BIT;
+	psize += prevsize + DIRECT_FOOT_PAD;
+	CALL_MUNMAP((char *)p - prevsize, psize);
+	return NULL;
+      } else {
+	mchunkptr prev = chunk_minus_offset(p, prevsize);
+	psize += prevsize;
+	p = prev;
+	/* consolidate backward */
+	if (p != fm->dv) {
+	  unlink_chunk(fm, p, prevsize);
+	} else if ((next->head & INUSE_BITS) == INUSE_BITS) {
+	  fm->dvsize = psize;
+	  set_free_with_pinuse(p, psize, next);
+	  return NULL;
+	}
+      }
+    }
+    if (!cinuse(next)) {  /* consolidate forward */
+      if (next == fm->top) {
+	size_t tsize = fm->topsize += psize;
+	fm->top = p;
+	p->head = tsize | PINUSE_BIT;
+	if (p == fm->dv) {
+	  fm->dv = 0;
+	  fm->dvsize = 0;
+	}
+	if (tsize > fm->trim_check)
+	  alloc_trim(fm, 0);
+	return NULL;
+      } else if (next == fm->dv) {
+	size_t dsize = fm->dvsize += psize;
+	fm->dv = p;
+	set_size_and_pinuse_of_free_chunk(p, dsize);
+	return NULL;
+      } else {
+	size_t nsize = chunksize(next);
+	psize += nsize;
+	unlink_chunk(fm, next, nsize);
+	set_size_and_pinuse_of_free_chunk(p, psize);
+	if (p == fm->dv) {
+	  fm->dvsize = psize;
+	  return NULL;
+	}
+      }
+    } else {
+      set_free_with_pinuse(p, psize, next);
+    }
+
+    if (is_small(psize)) {
+      insert_small_chunk(fm, p, psize);
+    } else {
+      tchunkptr tp = (tchunkptr)p;
+      insert_large_chunk(fm, tp, psize);
+      if (--fm->release_checks == 0)
+	release_unused_segments(fm);
+    }
+  }
+  return NULL;
+}
+
+static LJ_NOINLINE void *lj_alloc_realloc(void *msp, void *ptr, size_t nsize)
+{
+  if (nsize >= MAX_REQUEST) {
+    return NULL;
+  } else {
+    mstate m = (mstate)msp;
+    mchunkptr oldp = mem2chunk(ptr);
+    size_t oldsize = chunksize(oldp);
+    mchunkptr next = chunk_plus_offset(oldp, oldsize);
+    mchunkptr newp = 0;
+    size_t nb = request2size(nsize);
+
+    /* Try to either shrink or extend into top. Else malloc-copy-free */
+    if (is_direct(oldp)) {
+      newp = direct_resize(oldp, nb);  /* this may return NULL. */
+    } else if (oldsize >= nb) { /* already big enough */
+      size_t rsize = oldsize - nb;
+      newp = oldp;
+      if (rsize >= MIN_CHUNK_SIZE) {
+	mchunkptr rem = chunk_plus_offset(newp, nb);
+	set_inuse(m, newp, nb);
+	set_inuse(m, rem, rsize);
+	lj_alloc_free(m, chunk2mem(rem));
+      }
+    } else if (next == m->top && oldsize + m->topsize > nb) {
+      /* Expand into top */
+      size_t newsize = oldsize + m->topsize;
+      size_t newtopsize = newsize - nb;
+      mchunkptr newtop = chunk_plus_offset(oldp, nb);
+      set_inuse(m, oldp, nb);
+      newtop->head = newtopsize |PINUSE_BIT;
+      m->top = newtop;
+      m->topsize = newtopsize;
+      newp = oldp;
+    }
+
+    if (newp != 0) {
+      return chunk2mem(newp);
+    } else {
+      void *newmem = lj_alloc_malloc(m, nsize);
+      if (newmem != 0) {
+	size_t oc = oldsize - overhead_for(oldp);
+	memcpy(newmem, ptr, oc < nsize ? oc : nsize);
+	lj_alloc_free(m, ptr);
+      }
+      return newmem;
+    }
+  }
+}
+
+void *lj_alloc_f(void *msp, void *ptr, size_t osize, size_t nsize)
+{
+  (void)osize;
+  if (nsize == 0) {
+    return lj_alloc_free(msp, ptr);
+  } else if (ptr == NULL) {
+    return lj_alloc_malloc(msp, nsize);
+  } else {
+    return lj_alloc_realloc(msp, ptr, nsize);
+  }
+}
+
+#endif

http://git-wip-us.apache.org/repos/asf/trafficserver/blob/1f27b840/lib/luajit/src/lj_alloc.h
----------------------------------------------------------------------
diff --git a/lib/luajit/src/lj_alloc.h b/lib/luajit/src/lj_alloc.h
new file mode 100644
index 0000000..f87a7cf
--- /dev/null
+++ b/lib/luajit/src/lj_alloc.h
@@ -0,0 +1,17 @@
+/*
+** Bundled memory allocator.
+** Donated to the public domain.
+*/
+
+#ifndef _LJ_ALLOC_H
+#define _LJ_ALLOC_H
+
+#include "lj_def.h"
+
+#ifndef LUAJIT_USE_SYSMALLOC
+LJ_FUNC void *lj_alloc_create(void);
+LJ_FUNC void lj_alloc_destroy(void *msp);
+LJ_FUNC void *lj_alloc_f(void *msp, void *ptr, size_t osize, size_t nsize);
+#endif
+
+#endif

http://git-wip-us.apache.org/repos/asf/trafficserver/blob/1f27b840/lib/luajit/src/lj_api.c
----------------------------------------------------------------------
diff --git a/lib/luajit/src/lj_api.c b/lib/luajit/src/lj_api.c
new file mode 100644
index 0000000..3bedb39
--- /dev/null
+++ b/lib/luajit/src/lj_api.c
@@ -0,0 +1,1200 @@
+/*
+** Public Lua/C API.
+** Copyright (C) 2005-2015 Mike Pall. See Copyright Notice in luajit.h
+**
+** Major portions taken verbatim or adapted from the Lua interpreter.
+** Copyright (C) 1994-2008 Lua.org, PUC-Rio. See Copyright Notice in lua.h
+*/
+
+#define lj_api_c
+#define LUA_CORE
+
+#include "lj_obj.h"
+#include "lj_gc.h"
+#include "lj_err.h"
+#include "lj_debug.h"
+#include "lj_str.h"
+#include "lj_tab.h"
+#include "lj_func.h"
+#include "lj_udata.h"
+#include "lj_meta.h"
+#include "lj_state.h"
+#include "lj_bc.h"
+#include "lj_frame.h"
+#include "lj_trace.h"
+#include "lj_vm.h"
+#include "lj_strscan.h"
+
+/* -- Common helper functions --------------------------------------------- */
+
+#define api_checknelems(L, n)		api_check(L, (n) <= (L->top - L->base))
+#define api_checkvalidindex(L, i)	api_check(L, (i) != niltv(L))
+
+static TValue *index2adr(lua_State *L, int idx)
+{
+  if (idx > 0) {
+    TValue *o = L->base + (idx - 1);
+    return o < L->top ? o : niltv(L);
+  } else if (idx > LUA_REGISTRYINDEX) {
+    api_check(L, idx != 0 && -idx <= L->top - L->base);
+    return L->top + idx;
+  } else if (idx == LUA_GLOBALSINDEX) {
+    TValue *o = &G(L)->tmptv;
+    settabV(L, o, tabref(L->env));
+    return o;
+  } else if (idx == LUA_REGISTRYINDEX) {
+    return registry(L);
+  } else {
+    GCfunc *fn = curr_func(L);
+    api_check(L, fn->c.gct == ~LJ_TFUNC && !isluafunc(fn));
+    if (idx == LUA_ENVIRONINDEX) {
+      TValue *o = &G(L)->tmptv;
+      settabV(L, o, tabref(fn->c.env));
+      return o;
+    } else {
+      idx = LUA_GLOBALSINDEX - idx;
+      return idx <= fn->c.nupvalues ? &fn->c.upvalue[idx-1] : niltv(L);
+    }
+  }
+}
+
+static TValue *stkindex2adr(lua_State *L, int idx)
+{
+  if (idx > 0) {
+    TValue *o = L->base + (idx - 1);
+    return o < L->top ? o : niltv(L);
+  } else {
+    api_check(L, idx != 0 && -idx <= L->top - L->base);
+    return L->top + idx;
+  }
+}
+
+static GCtab *getcurrenv(lua_State *L)
+{
+  GCfunc *fn = curr_func(L);
+  return fn->c.gct == ~LJ_TFUNC ? tabref(fn->c.env) : tabref(L->env);
+}
+
+/* -- Miscellaneous API functions ----------------------------------------- */
+
+LUA_API int lua_status(lua_State *L)
+{
+  return L->status;
+}
+
+LUA_API int lua_checkstack(lua_State *L, int size)
+{
+  if (size > LUAI_MAXCSTACK || (L->top - L->base + size) > LUAI_MAXCSTACK) {
+    return 0;  /* Stack overflow. */
+  } else if (size > 0) {
+    lj_state_checkstack(L, (MSize)size);
+  }
+  return 1;
+}
+
+LUALIB_API void luaL_checkstack(lua_State *L, int size, const char *msg)
+{
+  if (!lua_checkstack(L, size))
+    lj_err_callerv(L, LJ_ERR_STKOVM, msg);
+}
+
+LUA_API void lua_xmove(lua_State *from, lua_State *to, int n)
+{
+  TValue *f, *t;
+  if (from == to) return;
+  api_checknelems(from, n);
+  api_check(from, G(from) == G(to));
+  lj_state_checkstack(to, (MSize)n);
+  f = from->top;
+  t = to->top = to->top + n;
+  while (--n >= 0) copyTV(to, --t, --f);
+  from->top = f;
+}
+
+/* -- Stack manipulation -------------------------------------------------- */
+
+LUA_API int lua_gettop(lua_State *L)
+{
+  return (int)(L->top - L->base);
+}
+
+LUA_API void lua_settop(lua_State *L, int idx)
+{
+  if (idx >= 0) {
+    api_check(L, idx <= tvref(L->maxstack) - L->base);
+    if (L->base + idx > L->top) {
+      if (L->base + idx >= tvref(L->maxstack))
+	lj_state_growstack(L, (MSize)idx - (MSize)(L->top - L->base));
+      do { setnilV(L->top++); } while (L->top < L->base + idx);
+    } else {
+      L->top = L->base + idx;
+    }
+  } else {
+    api_check(L, -(idx+1) <= (L->top - L->base));
+    L->top += idx+1;  /* Shrinks top (idx < 0). */
+  }
+}
+
+LUA_API void lua_remove(lua_State *L, int idx)
+{
+  TValue *p = stkindex2adr(L, idx);
+  api_checkvalidindex(L, p);
+  while (++p < L->top) copyTV(L, p-1, p);
+  L->top--;
+}
+
+LUA_API void lua_insert(lua_State *L, int idx)
+{
+  TValue *q, *p = stkindex2adr(L, idx);
+  api_checkvalidindex(L, p);
+  for (q = L->top; q > p; q--) copyTV(L, q, q-1);
+  copyTV(L, p, L->top);
+}
+
+LUA_API void lua_replace(lua_State *L, int idx)
+{
+  api_checknelems(L, 1);
+  if (idx == LUA_GLOBALSINDEX) {
+    api_check(L, tvistab(L->top-1));
+    /* NOBARRIER: A thread (i.e. L) is never black. */
+    setgcref(L->env, obj2gco(tabV(L->top-1)));
+  } else if (idx == LUA_ENVIRONINDEX) {
+    GCfunc *fn = curr_func(L);
+    if (fn->c.gct != ~LJ_TFUNC)
+      lj_err_msg(L, LJ_ERR_NOENV);
+    api_check(L, tvistab(L->top-1));
+    setgcref(fn->c.env, obj2gco(tabV(L->top-1)));
+    lj_gc_barrier(L, fn, L->top-1);
+  } else {
+    TValue *o = index2adr(L, idx);
+    api_checkvalidindex(L, o);
+    copyTV(L, o, L->top-1);
+    if (idx < LUA_GLOBALSINDEX)  /* Need a barrier for upvalues. */
+      lj_gc_barrier(L, curr_func(L), L->top-1);
+  }
+  L->top--;
+}
+
+LUA_API void lua_pushvalue(lua_State *L, int idx)
+{
+  copyTV(L, L->top, index2adr(L, idx));
+  incr_top(L);
+}
+
+/* -- Stack getters ------------------------------------------------------- */
+
+LUA_API int lua_type(lua_State *L, int idx)
+{
+  cTValue *o = index2adr(L, idx);
+  if (tvisnumber(o)) {
+    return LUA_TNUMBER;
+#if LJ_64
+  } else if (tvislightud(o)) {
+    return LUA_TLIGHTUSERDATA;
+#endif
+  } else if (o == niltv(L)) {
+    return LUA_TNONE;
+  } else {  /* Magic internal/external tag conversion. ORDER LJ_T */
+    uint32_t t = ~itype(o);
+#if LJ_64
+    int tt = (int)((U64x(75a06,98042110) >> 4*t) & 15u);
+#else
+    int tt = (int)(((t < 8 ? 0x98042110u : 0x75a06u) >> 4*(t&7)) & 15u);
+#endif
+    lua_assert(tt != LUA_TNIL || tvisnil(o));
+    return tt;
+  }
+}
+
+LUALIB_API void luaL_checktype(lua_State *L, int idx, int tt)
+{
+  if (lua_type(L, idx) != tt)
+    lj_err_argt(L, idx, tt);
+}
+
+LUALIB_API void luaL_checkany(lua_State *L, int idx)
+{
+  if (index2adr(L, idx) == niltv(L))
+    lj_err_arg(L, idx, LJ_ERR_NOVAL);
+}
+
+LUA_API const char *lua_typename(lua_State *L, int t)
+{
+  UNUSED(L);
+  return lj_obj_typename[t+1];
+}
+
+LUA_API int lua_iscfunction(lua_State *L, int idx)
+{
+  cTValue *o = index2adr(L, idx);
+  return tvisfunc(o) && !isluafunc(funcV(o));
+}
+
+LUA_API int lua_isnumber(lua_State *L, int idx)
+{
+  cTValue *o = index2adr(L, idx);
+  TValue tmp;
+  return (tvisnumber(o) || (tvisstr(o) && lj_strscan_number(strV(o), &tmp)));
+}
+
+LUA_API int lua_isstring(lua_State *L, int idx)
+{
+  cTValue *o = index2adr(L, idx);
+  return (tvisstr(o) || tvisnumber(o));
+}
+
+LUA_API int lua_isuserdata(lua_State *L, int idx)
+{
+  cTValue *o = index2adr(L, idx);
+  return (tvisudata(o) || tvislightud(o));
+}
+
+LUA_API int lua_rawequal(lua_State *L, int idx1, int idx2)
+{
+  cTValue *o1 = index2adr(L, idx1);
+  cTValue *o2 = index2adr(L, idx2);
+  return (o1 == niltv(L) || o2 == niltv(L)) ? 0 : lj_obj_equal(o1, o2);
+}
+
+LUA_API int lua_equal(lua_State *L, int idx1, int idx2)
+{
+  cTValue *o1 = index2adr(L, idx1);
+  cTValue *o2 = index2adr(L, idx2);
+  if (tvisint(o1) && tvisint(o2)) {
+    return intV(o1) == intV(o2);
+  } else if (tvisnumber(o1) && tvisnumber(o2)) {
+    return numberVnum(o1) == numberVnum(o2);
+  } else if (itype(o1) != itype(o2)) {
+    return 0;
+  } else if (tvispri(o1)) {
+    return o1 != niltv(L) && o2 != niltv(L);
+#if LJ_64
+  } else if (tvislightud(o1)) {
+    return o1->u64 == o2->u64;
+#endif
+  } else if (gcrefeq(o1->gcr, o2->gcr)) {
+    return 1;
+  } else if (!tvistabud(o1)) {
+    return 0;
+  } else {
+    TValue *base = lj_meta_equal(L, gcV(o1), gcV(o2), 0);
+    if ((uintptr_t)base <= 1) {
+      return (int)(uintptr_t)base;
+    } else {
+      L->top = base+2;
+      lj_vm_call(L, base, 1+1);
+      L->top -= 2;
+      return tvistruecond(L->top+1);
+    }
+  }
+}
+
+LUA_API int lua_lessthan(lua_State *L, int idx1, int idx2)
+{
+  cTValue *o1 = index2adr(L, idx1);
+  cTValue *o2 = index2adr(L, idx2);
+  if (o1 == niltv(L) || o2 == niltv(L)) {
+    return 0;
+  } else if (tvisint(o1) && tvisint(o2)) {
+    return intV(o1) < intV(o2);
+  } else if (tvisnumber(o1) && tvisnumber(o2)) {
+    return numberVnum(o1) < numberVnum(o2);
+  } else {
+    TValue *base = lj_meta_comp(L, o1, o2, 0);
+    if ((uintptr_t)base <= 1) {
+      return (int)(uintptr_t)base;
+    } else {
+      L->top = base+2;
+      lj_vm_call(L, base, 1+1);
+      L->top -= 2;
+      return tvistruecond(L->top+1);
+    }
+  }
+}
+
+LUA_API lua_Number lua_tonumber(lua_State *L, int idx)
+{
+  cTValue *o = index2adr(L, idx);
+  TValue tmp;
+  if (LJ_LIKELY(tvisnumber(o)))
+    return numberVnum(o);
+  else if (tvisstr(o) && lj_strscan_num(strV(o), &tmp))
+    return numV(&tmp);
+  else
+    return 0;
+}
+
+LUALIB_API lua_Number luaL_checknumber(lua_State *L, int idx)
+{
+  cTValue *o = index2adr(L, idx);
+  TValue tmp;
+  if (LJ_LIKELY(tvisnumber(o)))
+    return numberVnum(o);
+  else if (!(tvisstr(o) && lj_strscan_num(strV(o), &tmp)))
+    lj_err_argt(L, idx, LUA_TNUMBER);
+  return numV(&tmp);
+}
+
+LUALIB_API lua_Number luaL_optnumber(lua_State *L, int idx, lua_Number def)
+{
+  cTValue *o = index2adr(L, idx);
+  TValue tmp;
+  if (LJ_LIKELY(tvisnumber(o)))
+    return numberVnum(o);
+  else if (tvisnil(o))
+    return def;
+  else if (!(tvisstr(o) && lj_strscan_num(strV(o), &tmp)))
+    lj_err_argt(L, idx, LUA_TNUMBER);
+  return numV(&tmp);
+}
+
+LUA_API lua_Integer lua_tointeger(lua_State *L, int idx)
+{
+  cTValue *o = index2adr(L, idx);
+  TValue tmp;
+  lua_Number n;
+  if (LJ_LIKELY(tvisint(o))) {
+    return intV(o);
+  } else if (LJ_LIKELY(tvisnum(o))) {
+    n = numV(o);
+  } else {
+    if (!(tvisstr(o) && lj_strscan_number(strV(o), &tmp)))
+      return 0;
+    if (tvisint(&tmp))
+      return (lua_Integer)intV(&tmp);
+    n = numV(&tmp);
+  }
+#if LJ_64
+  return (lua_Integer)n;
+#else
+  return lj_num2int(n);
+#endif
+}
+
+LUALIB_API lua_Integer luaL_checkinteger(lua_State *L, int idx)
+{
+  cTValue *o = index2adr(L, idx);
+  TValue tmp;
+  lua_Number n;
+  if (LJ_LIKELY(tvisint(o))) {
+    return intV(o);
+  } else if (LJ_LIKELY(tvisnum(o))) {
+    n = numV(o);
+  } else {
+    if (!(tvisstr(o) && lj_strscan_number(strV(o), &tmp)))
+      lj_err_argt(L, idx, LUA_TNUMBER);
+    if (tvisint(&tmp))
+      return (lua_Integer)intV(&tmp);
+    n = numV(&tmp);
+  }
+#if LJ_64
+  return (lua_Integer)n;
+#else
+  return lj_num2int(n);
+#endif
+}
+
+LUALIB_API lua_Integer luaL_optinteger(lua_State *L, int idx, lua_Integer def)
+{
+  cTValue *o = index2adr(L, idx);
+  TValue tmp;
+  lua_Number n;
+  if (LJ_LIKELY(tvisint(o))) {
+    return intV(o);
+  } else if (LJ_LIKELY(tvisnum(o))) {
+    n = numV(o);
+  } else if (tvisnil(o)) {
+    return def;
+  } else {
+    if (!(tvisstr(o) && lj_strscan_number(strV(o), &tmp)))
+      lj_err_argt(L, idx, LUA_TNUMBER);
+    if (tvisint(&tmp))
+      return (lua_Integer)intV(&tmp);
+    n = numV(&tmp);
+  }
+#if LJ_64
+  return (lua_Integer)n;
+#else
+  return lj_num2int(n);
+#endif
+}
+
+LUA_API int lua_toboolean(lua_State *L, int idx)
+{
+  cTValue *o = index2adr(L, idx);
+  return tvistruecond(o);
+}
+
+LUA_API const char *lua_tolstring(lua_State *L, int idx, size_t *len)
+{
+  TValue *o = index2adr(L, idx);
+  GCstr *s;
+  if (LJ_LIKELY(tvisstr(o))) {
+    s = strV(o);
+  } else if (tvisnumber(o)) {
+    lj_gc_check(L);
+    o = index2adr(L, idx);  /* GC may move the stack. */
+    s = lj_str_fromnumber(L, o);
+    setstrV(L, o, s);
+  } else {
+    if (len != NULL) *len = 0;
+    return NULL;
+  }
+  if (len != NULL) *len = s->len;
+  return strdata(s);
+}
+
+LUALIB_API const char *luaL_checklstring(lua_State *L, int idx, size_t *len)
+{
+  TValue *o = index2adr(L, idx);
+  GCstr *s;
+  if (LJ_LIKELY(tvisstr(o))) {
+    s = strV(o);
+  } else if (tvisnumber(o)) {
+    lj_gc_check(L);
+    o = index2adr(L, idx);  /* GC may move the stack. */
+    s = lj_str_fromnumber(L, o);
+    setstrV(L, o, s);
+  } else {
+    lj_err_argt(L, idx, LUA_TSTRING);
+  }
+  if (len != NULL) *len = s->len;
+  return strdata(s);
+}
+
+LUALIB_API const char *luaL_optlstring(lua_State *L, int idx,
+				       const char *def, size_t *len)
+{
+  TValue *o = index2adr(L, idx);
+  GCstr *s;
+  if (LJ_LIKELY(tvisstr(o))) {
+    s = strV(o);
+  } else if (tvisnil(o)) {
+    if (len != NULL) *len = def ? strlen(def) : 0;
+    return def;
+  } else if (tvisnumber(o)) {
+    lj_gc_check(L);
+    o = index2adr(L, idx);  /* GC may move the stack. */
+    s = lj_str_fromnumber(L, o);
+    setstrV(L, o, s);
+  } else {
+    lj_err_argt(L, idx, LUA_TSTRING);
+  }
+  if (len != NULL) *len = s->len;
+  return strdata(s);
+}
+
+LUALIB_API int luaL_checkoption(lua_State *L, int idx, const char *def,
+				const char *const lst[])
+{
+  ptrdiff_t i;
+  const char *s = lua_tolstring(L, idx, NULL);
+  if (s == NULL && (s = def) == NULL)
+    lj_err_argt(L, idx, LUA_TSTRING);
+  for (i = 0; lst[i]; i++)
+    if (strcmp(lst[i], s) == 0)
+      return (int)i;
+  lj_err_argv(L, idx, LJ_ERR_INVOPTM, s);
+}
+
+LUA_API size_t lua_objlen(lua_State *L, int idx)
+{
+  TValue *o = index2adr(L, idx);
+  if (tvisstr(o)) {
+    return strV(o)->len;
+  } else if (tvistab(o)) {
+    return (size_t)lj_tab_len(tabV(o));
+  } else if (tvisudata(o)) {
+    return udataV(o)->len;
+  } else if (tvisnumber(o)) {
+    GCstr *s = lj_str_fromnumber(L, o);
+    setstrV(L, o, s);
+    return s->len;
+  } else {
+    return 0;
+  }
+}
+
+LUA_API lua_CFunction lua_tocfunction(lua_State *L, int idx)
+{
+  cTValue *o = index2adr(L, idx);
+  if (tvisfunc(o)) {
+    BCOp op = bc_op(*mref(funcV(o)->c.pc, BCIns));
+    if (op == BC_FUNCC || op == BC_FUNCCW)
+      return funcV(o)->c.f;
+  }
+  return NULL;
+}
+
+LUA_API void *lua_touserdata(lua_State *L, int idx)
+{
+  cTValue *o = index2adr(L, idx);
+  if (tvisudata(o))
+    return uddata(udataV(o));
+  else if (tvislightud(o))
+    return lightudV(o);
+  else
+    return NULL;
+}
+
+LUA_API lua_State *lua_tothread(lua_State *L, int idx)
+{
+  cTValue *o = index2adr(L, idx);
+  return (!tvisthread(o)) ? NULL : threadV(o);
+}
+
+LUA_API const void *lua_topointer(lua_State *L, int idx)
+{
+  cTValue *o = index2adr(L, idx);
+  if (tvisudata(o))
+    return uddata(udataV(o));
+  else if (tvislightud(o))
+    return lightudV(o);
+  else if (tviscdata(o))
+    return cdataptr(cdataV(o));
+  else if (tvisgcv(o))
+    return gcV(o);
+  else
+    return NULL;
+}
+
+/* -- Stack setters (object creation) ------------------------------------- */
+
+LUA_API void lua_pushnil(lua_State *L)
+{
+  setnilV(L->top);
+  incr_top(L);
+}
+
+LUA_API void lua_pushnumber(lua_State *L, lua_Number n)
+{
+  setnumV(L->top, n);
+  if (LJ_UNLIKELY(tvisnan(L->top)))
+    setnanV(L->top);  /* Canonicalize injected NaNs. */
+  incr_top(L);
+}
+
+LUA_API void lua_pushinteger(lua_State *L, lua_Integer n)
+{
+  setintptrV(L->top, n);
+  incr_top(L);
+}
+
+LUA_API void lua_pushlstring(lua_State *L, const char *str, size_t len)
+{
+  GCstr *s;
+  lj_gc_check(L);
+  s = lj_str_new(L, str, len);
+  setstrV(L, L->top, s);
+  incr_top(L);
+}
+
+LUA_API void lua_pushstring(lua_State *L, const char *str)
+{
+  if (str == NULL) {
+    setnilV(L->top);
+  } else {
+    GCstr *s;
+    lj_gc_check(L);
+    s = lj_str_newz(L, str);
+    setstrV(L, L->top, s);
+  }
+  incr_top(L);
+}
+
+LUA_API const char *lua_pushvfstring(lua_State *L, const char *fmt,
+				     va_list argp)
+{
+  lj_gc_check(L);
+  return lj_str_pushvf(L, fmt, argp);
+}
+
+LUA_API const char *lua_pushfstring(lua_State *L, const char *fmt, ...)
+{
+  const char *ret;
+  va_list argp;
+  lj_gc_check(L);
+  va_start(argp, fmt);
+  ret = lj_str_pushvf(L, fmt, argp);
+  va_end(argp);
+  return ret;
+}
+
+LUA_API void lua_pushcclosure(lua_State *L, lua_CFunction f, int n)
+{
+  GCfunc *fn;
+  lj_gc_check(L);
+  api_checknelems(L, n);
+  fn = lj_func_newC(L, (MSize)n, getcurrenv(L));
+  fn->c.f = f;
+  L->top -= n;
+  while (n--)
+    copyTV(L, &fn->c.upvalue[n], L->top+n);
+  setfuncV(L, L->top, fn);
+  lua_assert(iswhite(obj2gco(fn)));
+  incr_top(L);
+}
+
+LUA_API void lua_pushboolean(lua_State *L, int b)
+{
+  setboolV(L->top, (b != 0));
+  incr_top(L);
+}
+
+LUA_API void lua_pushlightuserdata(lua_State *L, void *p)
+{
+  setlightudV(L->top, checklightudptr(L, p));
+  incr_top(L);
+}
+
+LUA_API void lua_createtable(lua_State *L, int narray, int nrec)
+{
+  GCtab *t;
+  lj_gc_check(L);
+  t = lj_tab_new(L, (uint32_t)(narray > 0 ? narray+1 : 0), hsize2hbits(nrec));
+  settabV(L, L->top, t);
+  incr_top(L);
+}
+
+LUALIB_API int luaL_newmetatable(lua_State *L, const char *tname)
+{
+  GCtab *regt = tabV(registry(L));
+  TValue *tv = lj_tab_setstr(L, regt, lj_str_newz(L, tname));
+  if (tvisnil(tv)) {
+    GCtab *mt = lj_tab_new(L, 0, 1);
+    settabV(L, tv, mt);
+    settabV(L, L->top++, mt);
+    lj_gc_anybarriert(L, regt);
+    return 1;
+  } else {
+    copyTV(L, L->top++, tv);
+    return 0;
+  }
+}
+
+LUA_API int lua_pushthread(lua_State *L)
+{
+  setthreadV(L, L->top, L);
+  incr_top(L);
+  return (mainthread(G(L)) == L);
+}
+
+LUA_API lua_State *lua_newthread(lua_State *L)
+{
+  lua_State *L1;
+  lj_gc_check(L);
+  L1 = lj_state_new(L);
+  setthreadV(L, L->top, L1);
+  incr_top(L);
+  return L1;
+}
+
+LUA_API void *lua_newuserdata(lua_State *L, size_t size)
+{
+  GCudata *ud;
+  lj_gc_check(L);
+  if (size > LJ_MAX_UDATA)
+    lj_err_msg(L, LJ_ERR_UDATAOV);
+  ud = lj_udata_new(L, (MSize)size, getcurrenv(L));
+  setudataV(L, L->top, ud);
+  incr_top(L);
+  return uddata(ud);
+}
+
+LUA_API void lua_concat(lua_State *L, int n)
+{
+  api_checknelems(L, n);
+  if (n >= 2) {
+    n--;
+    do {
+      TValue *top = lj_meta_cat(L, L->top-1, -n);
+      if (top == NULL) {
+	L->top -= n;
+	break;
+      }
+      n -= (int)(L->top - top);
+      L->top = top+2;
+      lj_vm_call(L, top, 1+1);
+      L->top--;
+      copyTV(L, L->top-1, L->top);
+    } while (--n > 0);
+  } else if (n == 0) {  /* Push empty string. */
+    setstrV(L, L->top, &G(L)->strempty);
+    incr_top(L);
+  }
+  /* else n == 1: nothing to do. */
+}
+
+/* -- Object getters ------------------------------------------------------ */
+
+LUA_API void lua_gettable(lua_State *L, int idx)
+{
+  cTValue *v, *t = index2adr(L, idx);
+  api_checkvalidindex(L, t);
+  v = lj_meta_tget(L, t, L->top-1);
+  if (v == NULL) {
+    L->top += 2;
+    lj_vm_call(L, L->top-2, 1+1);
+    L->top -= 2;
+    v = L->top+1;
+  }
+  copyTV(L, L->top-1, v);
+}
+
+LUA_API void lua_getfield(lua_State *L, int idx, const char *k)
+{
+  cTValue *v, *t = index2adr(L, idx);
+  TValue key;
+  api_checkvalidindex(L, t);
+  setstrV(L, &key, lj_str_newz(L, k));
+  v = lj_meta_tget(L, t, &key);
+  if (v == NULL) {
+    L->top += 2;
+    lj_vm_call(L, L->top-2, 1+1);
+    L->top -= 2;
+    v = L->top+1;
+  }
+  copyTV(L, L->top, v);
+  incr_top(L);
+}
+
+LUA_API void lua_rawget(lua_State *L, int idx)
+{
+  cTValue *t = index2adr(L, idx);
+  api_check(L, tvistab(t));
+  copyTV(L, L->top-1, lj_tab_get(L, tabV(t), L->top-1));
+}
+
+LUA_API void lua_rawgeti(lua_State *L, int idx, int n)
+{
+  cTValue *v, *t = index2adr(L, idx);
+  api_check(L, tvistab(t));
+  v = lj_tab_getint(tabV(t), n);
+  if (v) {
+    copyTV(L, L->top, v);
+  } else {
+    setnilV(L->top);
+  }
+  incr_top(L);
+}
+
+LUA_API int lua_getmetatable(lua_State *L, int idx)
+{
+  cTValue *o = index2adr(L, idx);
+  GCtab *mt = NULL;
+  if (tvistab(o))
+    mt = tabref(tabV(o)->metatable);
+  else if (tvisudata(o))
+    mt = tabref(udataV(o)->metatable);
+  else
+    mt = tabref(basemt_obj(G(L), o));
+  if (mt == NULL)
+    return 0;
+  settabV(L, L->top, mt);
+  incr_top(L);
+  return 1;
+}
+
+LUALIB_API int luaL_getmetafield(lua_State *L, int idx, const char *field)
+{
+  if (lua_getmetatable(L, idx)) {
+    cTValue *tv = lj_tab_getstr(tabV(L->top-1), lj_str_newz(L, field));
+    if (tv && !tvisnil(tv)) {
+      copyTV(L, L->top-1, tv);
+      return 1;
+    }
+    L->top--;
+  }
+  return 0;
+}
+
+LUA_API void lua_getfenv(lua_State *L, int idx)
+{
+  cTValue *o = index2adr(L, idx);
+  api_checkvalidindex(L, o);
+  if (tvisfunc(o)) {
+    settabV(L, L->top, tabref(funcV(o)->c.env));
+  } else if (tvisudata(o)) {
+    settabV(L, L->top, tabref(udataV(o)->env));
+  } else if (tvisthread(o)) {
+    settabV(L, L->top, tabref(threadV(o)->env));
+  } else {
+    setnilV(L->top);
+  }
+  incr_top(L);
+}
+
+LUA_API int lua_next(lua_State *L, int idx)
+{
+  cTValue *t = index2adr(L, idx);
+  int more;
+  api_check(L, tvistab(t));
+  more = lj_tab_next(L, tabV(t), L->top-1);
+  if (more) {
+    incr_top(L);  /* Return new key and value slot. */
+  } else {  /* End of traversal. */
+    L->top--;  /* Remove key slot. */
+  }
+  return more;
+}
+
+LUA_API const char *lua_getupvalue(lua_State *L, int idx, int n)
+{
+  TValue *val;
+  const char *name = lj_debug_uvnamev(index2adr(L, idx), (uint32_t)(n-1), &val);
+  if (name) {
+    copyTV(L, L->top, val);
+    incr_top(L);
+  }
+  return name;
+}
+
+LUA_API void *lua_upvalueid(lua_State *L, int idx, int n)
+{
+  GCfunc *fn = funcV(index2adr(L, idx));
+  n--;
+  api_check(L, (uint32_t)n < fn->l.nupvalues);
+  return isluafunc(fn) ? (void *)gcref(fn->l.uvptr[n]) :
+			 (void *)&fn->c.upvalue[n];
+}
+
+LUA_API void lua_upvaluejoin(lua_State *L, int idx1, int n1, int idx2, int n2)
+{
+  GCfunc *fn1 = funcV(index2adr(L, idx1));
+  GCfunc *fn2 = funcV(index2adr(L, idx2));
+  n1--; n2--;
+  api_check(L, isluafunc(fn1) && (uint32_t)n1 < fn1->l.nupvalues);
+  api_check(L, isluafunc(fn2) && (uint32_t)n2 < fn2->l.nupvalues);
+  setgcrefr(fn1->l.uvptr[n1], fn2->l.uvptr[n2]);
+  lj_gc_objbarrier(L, fn1, gcref(fn1->l.uvptr[n1]));
+}
+
+LUALIB_API void *luaL_checkudata(lua_State *L, int idx, const char *tname)
+{
+  cTValue *o = index2adr(L, idx);
+  if (tvisudata(o)) {
+    GCudata *ud = udataV(o);
+    cTValue *tv = lj_tab_getstr(tabV(registry(L)), lj_str_newz(L, tname));
+    if (tv && tvistab(tv) && tabV(tv) == tabref(ud->metatable))
+      return uddata(ud);
+  }
+  lj_err_argtype(L, idx, tname);
+  return NULL;  /* unreachable */
+}
+
+/* -- Object setters ------------------------------------------------------ */
+
+LUA_API void lua_settable(lua_State *L, int idx)
+{
+  TValue *o;
+  cTValue *t = index2adr(L, idx);
+  api_checknelems(L, 2);
+  api_checkvalidindex(L, t);
+  o = lj_meta_tset(L, t, L->top-2);
+  if (o) {
+    /* NOBARRIER: lj_meta_tset ensures the table is not black. */
+    copyTV(L, o, L->top-1);
+    L->top -= 2;
+  } else {
+    L->top += 3;
+    copyTV(L, L->top-1, L->top-6);
+    lj_vm_call(L, L->top-3, 0+1);
+    L->top -= 3;
+  }
+}
+
+LUA_API void lua_setfield(lua_State *L, int idx, const char *k)
+{
+  TValue *o;
+  TValue key;
+  cTValue *t = index2adr(L, idx);
+  api_checknelems(L, 1);
+  api_checkvalidindex(L, t);
+  setstrV(L, &key, lj_str_newz(L, k));
+  o = lj_meta_tset(L, t, &key);
+  if (o) {
+    L->top--;
+    /* NOBARRIER: lj_meta_tset ensures the table is not black. */
+    copyTV(L, o, L->top);
+  } else {
+    L->top += 3;
+    copyTV(L, L->top-1, L->top-6);
+    lj_vm_call(L, L->top-3, 0+1);
+    L->top -= 2;
+  }
+}
+
+LUA_API void lua_rawset(lua_State *L, int idx)
+{
+  GCtab *t = tabV(index2adr(L, idx));
+  TValue *dst, *key;
+  api_checknelems(L, 2);
+  key = L->top-2;
+  dst = lj_tab_set(L, t, key);
+  copyTV(L, dst, key+1);
+  lj_gc_anybarriert(L, t);
+  L->top = key;
+}
+
+LUA_API void lua_rawseti(lua_State *L, int idx, int n)
+{
+  GCtab *t = tabV(index2adr(L, idx));
+  TValue *dst, *src;
+  api_checknelems(L, 1);
+  dst = lj_tab_setint(L, t, n);
+  src = L->top-1;
+  copyTV(L, dst, src);
+  lj_gc_barriert(L, t, dst);
+  L->top = src;
+}
+
+LUA_API int lua_setmetatable(lua_State *L, int idx)
+{
+  global_State *g;
+  GCtab *mt;
+  cTValue *o = index2adr(L, idx);
+  api_checknelems(L, 1);
+  api_checkvalidindex(L, o);
+  if (tvisnil(L->top-1)) {
+    mt = NULL;
+  } else {
+    api_check(L, tvistab(L->top-1));
+    mt = tabV(L->top-1);
+  }
+  g = G(L);
+  if (tvistab(o)) {
+    setgcref(tabV(o)->metatable, obj2gco(mt));
+    if (mt)
+      lj_gc_objbarriert(L, tabV(o), mt);
+  } else if (tvisudata(o)) {
+    setgcref(udataV(o)->metatable, obj2gco(mt));
+    if (mt)
+      lj_gc_objbarrier(L, udataV(o), mt);
+  } else {
+    /* Flush cache, since traces specialize to basemt. But not during __gc. */
+    if (lj_trace_flushall(L))
+      lj_err_caller(L, LJ_ERR_NOGCMM);
+    if (tvisbool(o)) {
+      /* NOBARRIER: basemt is a GC root. */
+      setgcref(basemt_it(g, LJ_TTRUE), obj2gco(mt));
+      setgcref(basemt_it(g, LJ_TFALSE), obj2gco(mt));
+    } else {
+      /* NOBARRIER: basemt is a GC root. */
+      setgcref(basemt_obj(g, o), obj2gco(mt));
+    }
+  }
+  L->top--;
+  return 1;
+}
+
+LUA_API int lua_setfenv(lua_State *L, int idx)
+{
+  cTValue *o = index2adr(L, idx);
+  GCtab *t;
+  api_checknelems(L, 1);
+  api_checkvalidindex(L, o);
+  api_check(L, tvistab(L->top-1));
+  t = tabV(L->top-1);
+  if (tvisfunc(o)) {
+    setgcref(funcV(o)->c.env, obj2gco(t));
+  } else if (tvisudata(o)) {
+    setgcref(udataV(o)->env, obj2gco(t));
+  } else if (tvisthread(o)) {
+    setgcref(threadV(o)->env, obj2gco(t));
+  } else {
+    L->top--;
+    return 0;
+  }
+  lj_gc_objbarrier(L, gcV(o), t);
+  L->top--;
+  return 1;
+}
+
+LUA_API const char *lua_setupvalue(lua_State *L, int idx, int n)
+{
+  cTValue *f = index2adr(L, idx);
+  TValue *val;
+  const char *name;
+  api_checknelems(L, 1);
+  name = lj_debug_uvnamev(f, (uint32_t)(n-1), &val);
+  if (name) {
+    L->top--;
+    copyTV(L, val, L->top);
+    lj_gc_barrier(L, funcV(f), L->top);
+  }
+  return name;
+}
+
+/* -- Calls --------------------------------------------------------------- */
+
+LUA_API void lua_call(lua_State *L, int nargs, int nresults)
+{
+  api_check(L, L->status == 0 || L->status == LUA_ERRERR);
+  api_checknelems(L, nargs+1);
+  lj_vm_call(L, L->top - nargs, nresults+1);
+}
+
+LUA_API int lua_pcall(lua_State *L, int nargs, int nresults, int errfunc)
+{
+  global_State *g = G(L);
+  uint8_t oldh = hook_save(g);
+  ptrdiff_t ef;
+  int status;
+  api_check(L, L->status == 0 || L->status == LUA_ERRERR);
+  api_checknelems(L, nargs+1);
+  if (errfunc == 0) {
+    ef = 0;
+  } else {
+    cTValue *o = stkindex2adr(L, errfunc);
+    api_checkvalidindex(L, o);
+    ef = savestack(L, o);
+  }
+  status = lj_vm_pcall(L, L->top - nargs, nresults+1, ef);
+  if (status) hook_restore(g, oldh);
+  return status;
+}
+
+static TValue *cpcall(lua_State *L, lua_CFunction func, void *ud)
+{
+  GCfunc *fn = lj_func_newC(L, 0, getcurrenv(L));
+  fn->c.f = func;
+  setfuncV(L, L->top, fn);
+  setlightudV(L->top+1, checklightudptr(L, ud));
+  cframe_nres(L->cframe) = 1+0;  /* Zero results. */
+  L->top += 2;
+  return L->top-1;  /* Now call the newly allocated C function. */
+}
+
+LUA_API int lua_cpcall(lua_State *L, lua_CFunction func, void *ud)
+{
+  global_State *g = G(L);
+  uint8_t oldh = hook_save(g);
+  int status;
+  api_check(L, L->status == 0 || L->status == LUA_ERRERR);
+  status = lj_vm_cpcall(L, func, ud, cpcall);
+  if (status) hook_restore(g, oldh);
+  return status;
+}
+
+LUALIB_API int luaL_callmeta(lua_State *L, int idx, const char *field)
+{
+  if (luaL_getmetafield(L, idx, field)) {
+    TValue *base = L->top--;
+    copyTV(L, base, index2adr(L, idx));
+    L->top = base+1;
+    lj_vm_call(L, base, 1+1);
+    return 1;
+  }
+  return 0;
+}
+
+/* -- Coroutine yield and resume ------------------------------------------ */
+
+LUA_API int lua_yield(lua_State *L, int nresults)
+{
+  void *cf = L->cframe;
+  global_State *g = G(L);
+  if (cframe_canyield(cf)) {
+    cf = cframe_raw(cf);
+    if (!hook_active(g)) {  /* Regular yield: move results down if needed. */
+      cTValue *f = L->top - nresults;
+      if (f > L->base) {
+	TValue *t = L->base;
+	while (--nresults >= 0) copyTV(L, t++, f++);
+	L->top = t;
+      }
+      L->cframe = NULL;
+      L->status = LUA_YIELD;
+      return -1;
+    } else {  /* Yield from hook: add a pseudo-frame. */
+      TValue *top = L->top;
+      hook_leave(g);
+      top->u64 = cframe_multres(cf);
+      setcont(top+1, lj_cont_hook);
+      setframe_pc(top+1, cframe_pc(cf)-1);
+      setframe_gc(top+2, obj2gco(L));
+      setframe_ftsz(top+2, (int)((char *)(top+3)-(char *)L->base)+FRAME_CONT);
+      L->top = L->base = top+3;
+#if LJ_TARGET_X64
+      lj_err_throw(L, LUA_YIELD);
+#else
+      L->cframe = NULL;
+      L->status = LUA_YIELD;
+      lj_vm_unwind_c(cf, LUA_YIELD);
+#endif
+    }
+  }
+  lj_err_msg(L, LJ_ERR_CYIELD);
+  return 0;  /* unreachable */
+}
+
+LUA_API int lua_resume(lua_State *L, int nargs)
+{
+  if (L->cframe == NULL && L->status <= LUA_YIELD)
+    return lj_vm_resume(L, L->top - nargs, 0, 0);
+  L->top = L->base;
+  setstrV(L, L->top, lj_err_str(L, LJ_ERR_COSUSP));
+  incr_top(L);
+  return LUA_ERRRUN;
+}
+
+/* -- GC and memory management -------------------------------------------- */
+
+LUA_API int lua_gc(lua_State *L, int what, int data)
+{
+  global_State *g = G(L);
+  int res = 0;
+  switch (what) {
+  case LUA_GCSTOP:
+    g->gc.threshold = LJ_MAX_MEM;
+    break;
+  case LUA_GCRESTART:
+    g->gc.threshold = data == -1 ? (g->gc.total/100)*g->gc.pause : g->gc.total;
+    break;
+  case LUA_GCCOLLECT:
+    lj_gc_fullgc(L);
+    break;
+  case LUA_GCCOUNT:
+    res = (int)(g->gc.total >> 10);
+    break;
+  case LUA_GCCOUNTB:
+    res = (int)(g->gc.total & 0x3ff);
+    break;
+  case LUA_GCSTEP: {
+    MSize a = (MSize)data << 10;
+    g->gc.threshold = (a <= g->gc.total) ? (g->gc.total - a) : 0;
+    while (g->gc.total >= g->gc.threshold)
+      if (lj_gc_step(L) > 0) {
+	res = 1;
+	break;
+      }
+    break;
+  }
+  case LUA_GCSETPAUSE:
+    res = (int)(g->gc.pause);
+    g->gc.pause = (MSize)data;
+    break;
+  case LUA_GCSETSTEPMUL:
+    res = (int)(g->gc.stepmul);
+    g->gc.stepmul = (MSize)data;
+    break;
+  default:
+    res = -1;  /* Invalid option. */
+  }
+  return res;
+}
+
+LUA_API lua_Alloc lua_getallocf(lua_State *L, void **ud)
+{
+  global_State *g = G(L);
+  if (ud) *ud = g->allocd;
+  return g->allocf;
+}
+
+LUA_API void lua_setallocf(lua_State *L, lua_Alloc f, void *ud)
+{
+  global_State *g = G(L);
+  g->allocd = ud;
+  g->allocf = f;
+}
+

http://git-wip-us.apache.org/repos/asf/trafficserver/blob/1f27b840/lib/luajit/src/lj_arch.h
----------------------------------------------------------------------
diff --git a/lib/luajit/src/lj_arch.h b/lib/luajit/src/lj_arch.h
new file mode 100644
index 0000000..d3a9d57
--- /dev/null
+++ b/lib/luajit/src/lj_arch.h
@@ -0,0 +1,437 @@
+/*
+** Target architecture selection.
+** Copyright (C) 2005-2015 Mike Pall. See Copyright Notice in luajit.h
+*/
+
+#ifndef _LJ_ARCH_H
+#define _LJ_ARCH_H
+
+#include "lua.h"
+
+/* Target endianess. */
+#define LUAJIT_LE	0
+#define LUAJIT_BE	1
+
+/* Target architectures. */
+#define LUAJIT_ARCH_X86		1
+#define LUAJIT_ARCH_x86		1
+#define LUAJIT_ARCH_X64		2
+#define LUAJIT_ARCH_x64		2
+#define LUAJIT_ARCH_ARM		3
+#define LUAJIT_ARCH_arm		3
+#define LUAJIT_ARCH_PPC		4
+#define LUAJIT_ARCH_ppc		4
+#define LUAJIT_ARCH_PPCSPE	5
+#define LUAJIT_ARCH_ppcspe	5
+#define LUAJIT_ARCH_MIPS	6
+#define LUAJIT_ARCH_mips	6
+
+/* Target OS. */
+#define LUAJIT_OS_OTHER		0
+#define LUAJIT_OS_WINDOWS	1
+#define LUAJIT_OS_LINUX		2
+#define LUAJIT_OS_OSX		3
+#define LUAJIT_OS_BSD		4
+#define LUAJIT_OS_POSIX		5
+
+/* Select native target if no target defined. */
+#ifndef LUAJIT_TARGET
+
+#if defined(__i386) || defined(__i386__) || defined(_M_IX86)
+#define LUAJIT_TARGET	LUAJIT_ARCH_X86
+#elif defined(__x86_64__) || defined(__x86_64) || defined(_M_X64) || defined(_M_AMD64)
+#define LUAJIT_TARGET	LUAJIT_ARCH_X64
+#elif defined(__arm__) || defined(__arm) || defined(__ARM__) || defined(__ARM)
+#define LUAJIT_TARGET	LUAJIT_ARCH_ARM
+#elif defined(__ppc__) || defined(__ppc) || defined(__PPC__) || defined(__PPC) || defined(__powerpc__) || defined(__powerpc) || defined(__POWERPC__) || defined(__POWERPC) || defined(_M_PPC)
+#ifdef __NO_FPRS__
+#define LUAJIT_TARGET	LUAJIT_ARCH_PPCSPE
+#else
+#define LUAJIT_TARGET	LUAJIT_ARCH_PPC
+#endif
+#elif defined(__mips__) || defined(__mips) || defined(__MIPS__) || defined(__MIPS)
+#define LUAJIT_TARGET	LUAJIT_ARCH_MIPS
+#else
+#error "No support for this architecture (yet)"
+#endif
+
+#endif
+
+/* Select native OS if no target OS defined. */
+#ifndef LUAJIT_OS
+
+#if defined(_WIN32) && !defined(_XBOX_VER)
+#define LUAJIT_OS	LUAJIT_OS_WINDOWS
+#elif defined(__linux__)
+#define LUAJIT_OS	LUAJIT_OS_LINUX
+#elif defined(__MACH__) && defined(__APPLE__)
+#define LUAJIT_OS	LUAJIT_OS_OSX
+#elif (defined(__FreeBSD__) || defined(__FreeBSD_kernel__) || \
+       defined(__NetBSD__) || defined(__OpenBSD__) || \
+       defined(__DragonFly__)) && !defined(__ORBIS__)
+#define LUAJIT_OS	LUAJIT_OS_BSD
+#elif (defined(__sun__) && defined(__svr4__)) || defined(__CYGWIN__)
+#define LUAJIT_OS	LUAJIT_OS_POSIX
+#else
+#define LUAJIT_OS	LUAJIT_OS_OTHER
+#endif
+
+#endif
+
+/* Set target OS properties. */
+#if LUAJIT_OS == LUAJIT_OS_WINDOWS
+#define LJ_OS_NAME	"Windows"
+#elif LUAJIT_OS == LUAJIT_OS_LINUX
+#define LJ_OS_NAME	"Linux"
+#elif LUAJIT_OS == LUAJIT_OS_OSX
+#define LJ_OS_NAME	"OSX"
+#elif LUAJIT_OS == LUAJIT_OS_BSD
+#define LJ_OS_NAME	"BSD"
+#elif LUAJIT_OS == LUAJIT_OS_POSIX
+#define LJ_OS_NAME	"POSIX"
+#else
+#define LJ_OS_NAME	"Other"
+#endif
+
+#define LJ_TARGET_WINDOWS	(LUAJIT_OS == LUAJIT_OS_WINDOWS)
+#define LJ_TARGET_LINUX		(LUAJIT_OS == LUAJIT_OS_LINUX)
+#define LJ_TARGET_OSX		(LUAJIT_OS == LUAJIT_OS_OSX)
+#define LJ_TARGET_IOS		(LJ_TARGET_OSX && LUAJIT_TARGET == LUAJIT_ARCH_ARM)
+#define LJ_TARGET_POSIX		(LUAJIT_OS > LUAJIT_OS_WINDOWS)
+#define LJ_TARGET_DLOPEN	LJ_TARGET_POSIX
+
+#ifdef __CELLOS_LV2__
+#define LJ_TARGET_PS3		1
+#define LJ_TARGET_CONSOLE	1
+#endif
+
+#ifdef __ORBIS__
+#define LJ_TARGET_PS4		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
+#endif
+
+#if _XBOX_VER >= 200
+#define LJ_TARGET_XBOX360	1
+#define LJ_TARGET_CONSOLE	1
+#endif
+
+#define LJ_NUMMODE_SINGLE	0	/* Single-number mode only. */
+#define LJ_NUMMODE_SINGLE_DUAL	1	/* Default to single-number mode. */
+#define LJ_NUMMODE_DUAL		2	/* Dual-number mode only. */
+#define LJ_NUMMODE_DUAL_SINGLE	3	/* Default to dual-number mode. */
+
+/* Set target architecture properties. */
+#if LUAJIT_TARGET == LUAJIT_ARCH_X86
+
+#define LJ_ARCH_NAME		"x86"
+#define LJ_ARCH_BITS		32
+#define LJ_ARCH_ENDIAN		LUAJIT_LE
+#if LJ_TARGET_WINDOWS || __CYGWIN__
+#define LJ_ABI_WIN		1
+#else
+#define LJ_ABI_WIN		0
+#endif
+#define LJ_TARGET_X86		1
+#define LJ_TARGET_X86ORX64	1
+#define LJ_TARGET_EHRETREG	0
+#define LJ_TARGET_MASKSHIFT	1
+#define LJ_TARGET_MASKROT	1
+#define LJ_TARGET_UNALIGNED	1
+#define LJ_ARCH_NUMMODE		LJ_NUMMODE_SINGLE_DUAL
+
+#elif LUAJIT_TARGET == LUAJIT_ARCH_X64
+
+#define LJ_ARCH_NAME		"x64"
+#define LJ_ARCH_BITS		64
+#define LJ_ARCH_ENDIAN		LUAJIT_LE
+#define LJ_ABI_WIN		LJ_TARGET_WINDOWS
+#define LJ_TARGET_X64		1
+#define LJ_TARGET_X86ORX64	1
+#define LJ_TARGET_EHRETREG	0
+#define LJ_TARGET_JUMPRANGE	31	/* +-2^31 = +-2GB */
+#define LJ_TARGET_MASKSHIFT	1
+#define LJ_TARGET_MASKROT	1
+#define LJ_TARGET_UNALIGNED	1
+#define LJ_ARCH_NUMMODE		LJ_NUMMODE_SINGLE_DUAL
+
+#elif LUAJIT_TARGET == LUAJIT_ARCH_ARM
+
+#define LJ_ARCH_NAME		"arm"
+#define LJ_ARCH_BITS		32
+#define LJ_ARCH_ENDIAN		LUAJIT_LE
+#if !defined(LJ_ARCH_HASFPU) && __SOFTFP__
+#define LJ_ARCH_HASFPU		0
+#endif
+#if !defined(LJ_ABI_SOFTFP) && !__ARM_PCS_VFP
+#define LJ_ABI_SOFTFP		1
+#endif
+#define LJ_ABI_EABI		1
+#define LJ_TARGET_ARM		1
+#define LJ_TARGET_EHRETREG	0
+#define LJ_TARGET_JUMPRANGE	25	/* +-2^25 = +-32MB */
+#define LJ_TARGET_MASKSHIFT	0
+#define LJ_TARGET_MASKROT	1
+#define LJ_TARGET_UNIFYROT	2	/* Want only IR_BROR. */
+#define LJ_ARCH_NUMMODE		LJ_NUMMODE_DUAL
+
+#if __ARM_ARCH____ARM_ARCH_8__ || __ARM_ARCH_8A__
+#define LJ_ARCH_VERSION		80
+#elif __ARM_ARCH_7__ || __ARM_ARCH_7A__ || __ARM_ARCH_7R__ || __ARM_ARCH_7S__ || __ARM_ARCH_7VE__
+#define LJ_ARCH_VERSION		70
+#elif __ARM_ARCH_6T2__
+#define LJ_ARCH_VERSION		61
+#elif __ARM_ARCH_6__ || __ARM_ARCH_6J__ || __ARM_ARCH_6K__ || __ARM_ARCH_6Z__ || __ARM_ARCH_6ZK__
+#define LJ_ARCH_VERSION		60
+#else
+#define LJ_ARCH_VERSION		50
+#endif
+
+#elif LUAJIT_TARGET == LUAJIT_ARCH_PPC
+
+#define LJ_ARCH_NAME		"ppc"
+#if _LP64
+#define LJ_ARCH_BITS		64
+#else
+#define LJ_ARCH_BITS		32
+#endif
+#define LJ_ARCH_ENDIAN		LUAJIT_BE
+#define LJ_TARGET_PPC		1
+#define LJ_TARGET_EHRETREG	3
+#define LJ_TARGET_JUMPRANGE	25	/* +-2^25 = +-32MB */
+#define LJ_TARGET_MASKSHIFT	0
+#define LJ_TARGET_MASKROT	1
+#define LJ_TARGET_UNIFYROT	1	/* Want only IR_BROL. */
+#define LJ_ARCH_NUMMODE		LJ_NUMMODE_DUAL_SINGLE
+
+#if _ARCH_PWR7
+#define LJ_ARCH_VERSION		70
+#elif _ARCH_PWR6
+#define LJ_ARCH_VERSION		60
+#elif _ARCH_PWR5X
+#define LJ_ARCH_VERSION		51
+#elif _ARCH_PWR5
+#define LJ_ARCH_VERSION		50
+#elif _ARCH_PWR4
+#define LJ_ARCH_VERSION		40
+#else
+#define LJ_ARCH_VERSION		0
+#endif
+#if __PPC64__ || __powerpc64__ || LJ_TARGET_CONSOLE
+#define LJ_ARCH_PPC64		1
+#define LJ_ARCH_NOFFI		1
+#endif
+#if _ARCH_PPCSQ
+#define LJ_ARCH_SQRT		1
+#endif
+#if _ARCH_PWR5X
+#define LJ_ARCH_ROUND		1
+#endif
+#if __PPU__
+#define LJ_ARCH_CELL		1
+#endif
+#if LJ_TARGET_XBOX360
+#define LJ_ARCH_XENON		1
+#endif
+
+#elif LUAJIT_TARGET == LUAJIT_ARCH_PPCSPE
+
+#define LJ_ARCH_NAME		"ppcspe"
+#define LJ_ARCH_BITS		32
+#define LJ_ARCH_ENDIAN		LUAJIT_BE
+#ifndef LJ_ABI_SOFTFP
+#define LJ_ABI_SOFTFP		1
+#endif
+#define LJ_ABI_EABI		1
+#define LJ_TARGET_PPCSPE	1
+#define LJ_TARGET_EHRETREG	3
+#define LJ_TARGET_JUMPRANGE	25	/* +-2^25 = +-32MB */
+#define LJ_TARGET_MASKSHIFT	0
+#define LJ_TARGET_MASKROT	1
+#define LJ_TARGET_UNIFYROT	1	/* Want only IR_BROL. */
+#define LJ_ARCH_NUMMODE		LJ_NUMMODE_SINGLE
+#define LJ_ARCH_NOFFI		1	/* NYI: comparisons, calls. */
+#define LJ_ARCH_NOJIT		1
+
+#elif LUAJIT_TARGET == LUAJIT_ARCH_MIPS
+
+#if defined(__MIPSEL__) || defined(__MIPSEL) || defined(_MIPSEL)
+#define LJ_ARCH_NAME		"mipsel"
+#define LJ_ARCH_ENDIAN		LUAJIT_LE
+#else
+#define LJ_ARCH_NAME		"mips"
+#define LJ_ARCH_ENDIAN		LUAJIT_BE
+#endif
+#define LJ_ARCH_BITS		32
+#define LJ_TARGET_MIPS		1
+#define LJ_TARGET_EHRETREG	4
+#define LJ_TARGET_JUMPRANGE	27	/* 2*2^27 = 256MB-aligned region */
+#define LJ_TARGET_MASKSHIFT	1
+#define LJ_TARGET_MASKROT	1
+#define LJ_TARGET_UNIFYROT	2	/* Want only IR_BROR. */
+#define LJ_ARCH_NUMMODE		LJ_NUMMODE_SINGLE
+
+#if _MIPS_ARCH_MIPS32R2
+#define LJ_ARCH_VERSION		20
+#else
+#define LJ_ARCH_VERSION		10
+#endif
+
+#else
+#error "No target architecture defined"
+#endif
+
+#ifndef LJ_PAGESIZE
+#define LJ_PAGESIZE		4096
+#endif
+
+/* Check for minimum required compiler versions. */
+#if defined(__GNUC__)
+#if LJ_TARGET_X86
+#if (__GNUC__ < 3) || ((__GNUC__ == 3) && __GNUC_MINOR__ < 4)
+#error "Need at least GCC 3.4 or newer"
+#endif
+#elif LJ_TARGET_X64
+#if __GNUC__ < 4
+#error "Need at least GCC 4.0 or newer"
+#endif
+#elif LJ_TARGET_ARM
+#if (__GNUC__ < 4) || ((__GNUC__ == 4) && __GNUC_MINOR__ < 2)
+#error "Need at least GCC 4.2 or newer"
+#endif
+#elif !LJ_TARGET_PS3
+#if (__GNUC__ < 4) || ((__GNUC__ == 4) && __GNUC_MINOR__ < 3)
+#error "Need at least GCC 4.3 or newer"
+#endif
+#endif
+#endif
+
+/* Check target-specific constraints. */
+#ifndef _BUILDVM_H
+#if LJ_TARGET_X64
+#if __USING_SJLJ_EXCEPTIONS__
+#error "Need a C compiler with native exception handling on x64"
+#endif
+#elif LJ_TARGET_ARM
+#if defined(__ARMEB__)
+#error "No support for big-endian ARM"
+#endif
+#if __ARM_ARCH_6M__ || __ARM_ARCH_7M__ || __ARM_ARCH_7EM__
+#error "No support for Cortex-M CPUs"
+#endif
+#if !(__ARM_EABI__ || LJ_TARGET_IOS)
+#error "Only ARM EABI or iOS 3.0+ ABI is supported"
+#endif
+#elif LJ_TARGET_PPC || LJ_TARGET_PPCSPE
+#if defined(_SOFT_FLOAT) || defined(_SOFT_DOUBLE)
+#error "No support for PowerPC CPUs without double-precision FPU"
+#endif
+#if defined(_LITTLE_ENDIAN)
+#error "No support for little-endian PowerPC"
+#endif
+#if defined(_LP64)
+#error "No support for PowerPC 64 bit mode"
+#endif
+#elif LJ_TARGET_MIPS
+#if defined(__mips_soft_float)
+#error "No support for MIPS CPUs without FPU"
+#endif
+#if defined(_LP64)
+#error "No support for MIPS64"
+#endif
+#endif
+#endif
+
+/* Enable or disable the dual-number mode for the VM. */
+#if (LJ_ARCH_NUMMODE == LJ_NUMMODE_SINGLE && LUAJIT_NUMMODE == 2) || \
+    (LJ_ARCH_NUMMODE == LJ_NUMMODE_DUAL && LUAJIT_NUMMODE == 1)
+#error "No support for this number mode on this architecture"
+#endif
+#if LJ_ARCH_NUMMODE == LJ_NUMMODE_DUAL || \
+    (LJ_ARCH_NUMMODE == LJ_NUMMODE_DUAL_SINGLE && LUAJIT_NUMMODE != 1) || \
+    (LJ_ARCH_NUMMODE == LJ_NUMMODE_SINGLE_DUAL && LUAJIT_NUMMODE == 2)
+#define LJ_DUALNUM		1
+#else
+#define LJ_DUALNUM		0
+#endif
+
+#if LJ_TARGET_IOS || LJ_TARGET_CONSOLE
+/* Runtime code generation is restricted on iOS. Complain to Apple, not me. */
+/* Ditto for the consoles. Complain to Sony or MS, not me. */
+#ifndef LUAJIT_ENABLE_JIT
+#define LJ_OS_NOJIT		1
+#endif
+#endif
+
+/* Disable or enable the JIT compiler. */
+#if defined(LUAJIT_DISABLE_JIT) || defined(LJ_ARCH_NOJIT) || defined(LJ_OS_NOJIT)
+#define LJ_HASJIT		0
+#else
+#define LJ_HASJIT		1
+#endif
+
+/* Disable or enable the FFI extension. */
+#if defined(LUAJIT_DISABLE_FFI) || defined(LJ_ARCH_NOFFI)
+#define LJ_HASFFI		0
+#else
+#define LJ_HASFFI		1
+#endif
+
+#ifndef LJ_ARCH_HASFPU
+#define LJ_ARCH_HASFPU		1
+#endif
+#ifndef LJ_ABI_SOFTFP
+#define LJ_ABI_SOFTFP		0
+#endif
+#define LJ_SOFTFP		(!LJ_ARCH_HASFPU)
+
+#if LJ_ARCH_ENDIAN == LUAJIT_BE
+#define LJ_LE			0
+#define LJ_BE			1
+#define LJ_ENDIAN_SELECT(le, be)	be
+#define LJ_ENDIAN_LOHI(lo, hi)		hi lo
+#else
+#define LJ_LE			1
+#define LJ_BE			0
+#define LJ_ENDIAN_SELECT(le, be)	le
+#define LJ_ENDIAN_LOHI(lo, hi)		lo hi
+#endif
+
+#if LJ_ARCH_BITS == 32
+#define LJ_32			1
+#define LJ_64			0
+#else
+#define LJ_32			0
+#define LJ_64			1
+#endif
+
+#ifndef LJ_TARGET_UNALIGNED
+#define LJ_TARGET_UNALIGNED	0
+#endif
+
+/* Various workarounds for embedded operating systems. */
+#if (defined(__ANDROID__) && !defined(LJ_TARGET_X86ORX64)) || defined(__symbian__) || LJ_TARGET_XBOX360
+#define LUAJIT_NO_LOG2
+#endif
+#if defined(__symbian__)
+#define LUAJIT_NO_EXP2
+#endif
+
+#if defined(LUAJIT_NO_UNWIND) || defined(__symbian__) || LJ_TARGET_IOS || LJ_TARGET_PS3
+#define LJ_NO_UNWIND		1
+#endif
+
+/* Compatibility with Lua 5.1 vs. 5.2. */
+#ifdef LUAJIT_ENABLE_LUA52COMPAT
+#define LJ_52			1
+#else
+#define LJ_52			0
+#endif
+
+#endif