Merge 035f133798 into e3c70a7d81

Reported by Christian Clason. #1334

Makefile

@@ -110,7 +110,7 @@ else
 endif
 TARGET_SYS?= $(HOST_SYS)
 
-ifeq (Darwin,$(TARGET_SYS))
+ifneq (,$(filter $(TARGET_SYS),Darwin iOS))
   INSTALL_SONAME= $(INSTALL_DYLIBNAME)
   INSTALL_SOSHORT1= $(INSTALL_DYLIBSHORT1)
   INSTALL_SOSHORT2= $(INSTALL_DYLIBSHORT2)

doc/install.html

@@ -117,7 +117,7 @@ hold all user-configurable settings:
 <li><tt>Makefile</tt> has settings for <b>installing</b> LuaJIT (POSIX
 only).</li>
 <li><tt>src/Makefile</tt> has settings for <b>compiling</b> LuaJIT
-under POSIX, MinGW or Cygwin.</li>
+under POSIX or MinGW.</li>
 <li><tt>src/msvcbuild.bat</tt> has settings for compiling LuaJIT with
 MSVC (Visual Studio).</li>
 </ul>
@@ -195,10 +195,8 @@ Obviously the prefixes given during build and installation need to be the same.
 <h2 id="windows">Windows Systems</h2>
 <h3>Prerequisites</h3>
 <p>
-Either install one of the open source SDKs
-(<a href="http://mingw.org/"><span class="ext">&raquo;</span>&nbsp;MinGW</a> or
-<a href="https://www.cygwin.com/"><span class="ext">&raquo;</span>&nbsp;Cygwin</a>), which come with a modified
-GCC plus the required development headers.
+Either install the open source SDK <a href="http://mingw.org/"><span class="ext">&raquo;</span>&nbsp;MinGW</a>,
+which comes with a modified GCC plus the required development headers.
 Or install Microsoft's Visual Studio (MSVC).
 </p>
 <h3>Building with MSVC</h3>
@@ -217,9 +215,9 @@ Then follow the installation instructions below.
 <p>
 For an x64 to ARM64 cross-build run this first: <tt>vcvarsall.bat x64_arm64</tt>
 </p>
-<h3>Building with MinGW or Cygwin</h3>
+<h3>Building with MinGW</h3>
 <p>
-Open a command prompt window and make sure the MinGW or Cygwin programs
+Open a command prompt window and make sure the MinGW programs
 are in your path. Then <tt>cd</tt> to the directory of the git repository.
 Then run this command for MinGW:
 </p>
@@ -227,12 +225,6 @@ Then run this command for MinGW:
 mingw32-make
 </pre>
 <p>
-Or this command for Cygwin:
-</p>
-<pre class="code">
-make
-</pre>
-<p>
 Then follow the installation instructions below.
 </p>
 <h3>Installing LuaJIT</h3>
@@ -249,6 +241,19 @@ absolute path names &mdash; all modules are loaded relative to the
 directory where <tt>luajit.exe</tt> is installed
 (see <tt>src/luaconf.h</tt>).
 </p>
+<p>
+The final directory layout should look like this:
+</p>
+<pre class="code">
+├── luajit.exe
+├── lua51.dll
+├── <- put your own classic Lua/C API modules (*.dll) here
+└── lua
+    ├── <- put your own Lua modules (*.lua) here
+    └── jit
+        ├── bc.lua
+        └── (etc …)
+</pre>
 
 <h2 id="cross">Cross-compiling LuaJIT</h2>
 <p>

dynasm/dasm_s390x.h

@@ -0,0 +1,546 @@
+/*
+** DynASM s390x encoding engine.
+** Copyright (C) 2005-2016 Mike Pall. All rights reserved.
+** Released under the MIT license. See dynasm.lua for full copyright notice.
+*/
+
+#include <stddef.h>
+#include <stdarg.h>
+#include <string.h>
+#include <stdlib.h>
+
+#define DASM_ARCH		"s390x"
+
+#ifndef DASM_EXTERN
+#define DASM_EXTERN(a,b,c,d)	0
+#endif
+
+/* Action definitions. */
+enum {
+  DASM_STOP, DASM_SECTION, DASM_ESC, DASM_REL_EXT,
+  /* The following actions need a buffer position. */
+  DASM_ALIGN, DASM_REL_LG, DASM_LABEL_LG,
+  /* The following actions also have an argument. */
+  DASM_REL_PC, DASM_LABEL_PC,
+  DASM_DISP12, DASM_DISP20,
+  DASM_IMM8, DASM_IMM16, DASM_IMM32,
+  DASM_LEN8R,DASM_LEN4HR,DASM_LEN4LR,
+  DASM__MAX
+};
+
+/* Maximum number of section buffer positions for a single dasm_put() call. */
+#define DASM_MAXSECPOS		25
+
+/* DynASM encoder status codes. Action list offset or number are or'ed in. */
+#define DASM_S_OK		0x00000000
+#define DASM_S_NOMEM		0x01000000
+#define DASM_S_PHASE		0x02000000
+#define DASM_S_MATCH_SEC	0x03000000
+#define DASM_S_RANGE_I		0x11000000
+#define DASM_S_RANGE_SEC	0x12000000
+#define DASM_S_RANGE_LG		0x13000000
+#define DASM_S_RANGE_PC		0x14000000
+#define DASM_S_RANGE_REL	0x15000000
+#define DASM_S_UNDEF_LG		0x21000000
+#define DASM_S_UNDEF_PC		0x22000000
+
+/* Macros to convert positions (8 bit section + 24 bit index). */
+#define DASM_POS2IDX(pos)	((pos)&0x00ffffff)
+#define DASM_POS2BIAS(pos)	((pos)&0xff000000)
+#define DASM_SEC2POS(sec)	((sec)<<24)
+#define DASM_POS2SEC(pos)	((pos)>>24)
+#define DASM_POS2PTR(D, pos)	(D->sections[DASM_POS2SEC(pos)].rbuf + (pos))
+
+/* Action list type. */
+typedef const unsigned short *dasm_ActList;
+
+/* Per-section structure. */
+typedef struct dasm_Section {
+  int *rbuf;                    /* Biased buffer pointer (negative section bias). */
+  int *buf;                     /* True buffer pointer. */
+  size_t bsize;                 /* Buffer size in bytes. */
+  int pos;                      /* Biased buffer position. */
+  int epos;                     /* End of biased buffer position - max single put. */
+  int ofs;                      /* Byte offset into section. */
+} dasm_Section;
+
+/* Core structure holding the DynASM encoding state. */
+struct dasm_State {
+  size_t psize;                 /* Allocated size of this structure. */
+  dasm_ActList actionlist;      /* Current actionlist pointer. */
+  int *lglabels;                /* Local/global chain/pos ptrs. */
+  size_t lgsize;
+  int *pclabels;                /* PC label chains/pos ptrs. */
+  size_t pcsize;
+  void **globals;               /* Array of globals (bias -10). */
+  dasm_Section *section;        /* Pointer to active section. */
+  size_t codesize;              /* Total size of all code sections. */
+  int maxsection;               /* 0 <= sectionidx < maxsection. */
+  int status;                   /* Status code. */
+  dasm_Section sections[1];     /* All sections. Alloc-extended. */
+};
+
+/* The size of the core structure depends on the max. number of sections. */
+#define DASM_PSZ(ms)	(sizeof(dasm_State)+(ms-1)*sizeof(dasm_Section))
+
+
+/* Initialize DynASM state. */
+void dasm_init(Dst_DECL, int maxsection)
+{
+  dasm_State *D;
+  size_t psz = 0;
+  int i;
+  Dst_REF = NULL;
+  DASM_M_GROW(Dst, struct dasm_State, Dst_REF, psz, DASM_PSZ(maxsection));
+  D = Dst_REF;
+  D->psize = psz;
+  D->lglabels = NULL;
+  D->lgsize = 0;
+  D->pclabels = NULL;
+  D->pcsize = 0;
+  D->globals = NULL;
+  D->maxsection = maxsection;
+  for (i = 0; i < maxsection; i++) {
+    D->sections[i].buf = NULL;  /* Need this for pass3. */
+    D->sections[i].rbuf = D->sections[i].buf - DASM_SEC2POS(i);
+    D->sections[i].bsize = 0;
+    D->sections[i].epos = 0;    /* Wrong, but is recalculated after resize. */
+  }
+}
+
+/* Free DynASM state. */
+void dasm_free(Dst_DECL)
+{
+  dasm_State *D = Dst_REF;
+  int i;
+  for (i = 0; i < D->maxsection; i++)
+    if (D->sections[i].buf)
+      DASM_M_FREE(Dst, D->sections[i].buf, D->sections[i].bsize);
+  if (D->pclabels)
+    DASM_M_FREE(Dst, D->pclabels, D->pcsize);
+  if (D->lglabels)
+    DASM_M_FREE(Dst, D->lglabels, D->lgsize);
+  DASM_M_FREE(Dst, D, D->psize);
+}
+
+/* Setup global label array. Must be called before dasm_setup(). */
+void dasm_setupglobal(Dst_DECL, void **gl, unsigned int maxgl)
+{
+  dasm_State *D = Dst_REF;
+  D->globals = gl - 10;         /* Negative bias to compensate for locals. */
+  DASM_M_GROW(Dst, int, D->lglabels, D->lgsize, (10 + maxgl) * sizeof(int));
+}
+
+/* Grow PC label array. Can be called after dasm_setup(), too. */
+void dasm_growpc(Dst_DECL, unsigned int maxpc)
+{
+  dasm_State *D = Dst_REF;
+  size_t osz = D->pcsize;
+  DASM_M_GROW(Dst, int, D->pclabels, D->pcsize, maxpc * sizeof(int));
+  memset((void *)(((unsigned char *)D->pclabels) + osz), 0, D->pcsize - osz);
+}
+
+/* Setup encoder. */
+void dasm_setup(Dst_DECL, const void *actionlist)
+{
+  dasm_State *D = Dst_REF;
+  int i;
+  D->actionlist = (dasm_ActList) actionlist;
+  D->status = DASM_S_OK;
+  D->section = &D->sections[0];
+  memset((void *)D->lglabels, 0, D->lgsize);
+  if (D->pclabels)
+    memset((void *)D->pclabels, 0, D->pcsize);
+  for (i = 0; i < D->maxsection; i++) {
+    D->sections[i].pos = DASM_SEC2POS(i);
+    D->sections[i].ofs = 0;
+  }
+}
+
+
+#ifdef DASM_CHECKS
+#define CK(x, st) \
+  do { if (!(x)) { \
+    D->status = DASM_S_##st|(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)
+#else
+#define CK(x, st)	((void)0)
+#define CKPL(kind, st)	((void)0)
+#endif
+
+/* Pass 1: Store actions and args, link branches/labels, estimate offsets. */
+void dasm_put(Dst_DECL, int start, ...)
+{
+  va_list ap;
+  dasm_State *D = Dst_REF;
+  dasm_ActList p = D->actionlist + start;
+  dasm_Section *sec = D->section;
+  int pos = sec->pos, ofs = sec->ofs;
+  int *b;
+
+  if (pos >= sec->epos) {
+    DASM_M_GROW(Dst, int, sec->buf, sec->bsize,
+                sec->bsize + 2 * DASM_MAXSECPOS * sizeof(int));
+    sec->rbuf = sec->buf - DASM_POS2BIAS(pos);
+    sec->epos =
+      (int)sec->bsize / sizeof(int) - DASM_MAXSECPOS + DASM_POS2BIAS(pos);
+  }
+
+  b = sec->rbuf;
+  b[pos++] = start;
+
+  va_start(ap, start);
+  while (1) {
+    unsigned short ins = *p++;
+    unsigned short action = ins;
+    if (action >= DASM__MAX) {
+      ofs += 2;
+      continue;
+    }
+
+    int *pl, n = action >= DASM_REL_PC ? va_arg(ap, int) : 0;
+    switch (action) {
+    case DASM_STOP:
+      goto stop;
+    case DASM_SECTION:
+      n = *p++ & 255;
+      CK(n < D->maxsection, RANGE_SEC);
+      D->section = &D->sections[n];
+      goto stop;
+    case DASM_ESC:
+      p++;
+      ofs += 2;
+      break;
+    case DASM_REL_EXT:
+      p++;
+      ofs += 4;
+      break;
+    case DASM_ALIGN:
+      ofs += *p++;
+      b[pos++] = ofs;
+      break;
+    case DASM_REL_LG:
+      if (p[-2] >> 12 == 0xc) { /* RIL instruction needs 32-bit immediate. */
+        ofs += 2;
+      }
+      n = *p++ - 10;
+      pl = D->lglabels + n;
+      /* Bkwd rel or global. */
+      if (n >= 0) {
+        CK(n >= 10 || *pl < 0, RANGE_LG);
+        CKPL(lg, LG);
+        goto putrel;
+      }
+      pl += 10;
+      n = *pl;
+      if (n < 0)
+        n = 0;                  /* Start new chain for fwd rel if label exists. */
+      goto linkrel;
+    case DASM_REL_PC:
+      if (p[-2] >> 12 == 0xc) { /* RIL instruction needs 32-bit immediate. */
+        ofs += 2;
+      }
+      pl = D->pclabels + n;
+      CKPL(pc, PC);
+    putrel:
+      n = *pl;
+      if (n < 0) {              /* Label exists. Get label pos and store it. */
+        b[pos] = -n;
+      } else {
+      linkrel:
+        b[pos] = n;             /* Else link to rel chain, anchored at label. */
+        *pl = pos;
+      }
+      ofs += 2;
+      pos++;
+      break;
+    case DASM_LABEL_LG:
+      pl = D->lglabels + *p++ - 10;
+      CKPL(lg, LG);
+      goto putlabel;
+    case DASM_LABEL_PC:
+      pl = D->pclabels + n;
+      CKPL(pc, PC);
+    putlabel:
+      n = *pl;                  /* n > 0: Collapse rel chain and replace with label pos. */
+      while (n > 0) {
+        int *pb = DASM_POS2PTR(D, n);
+        n = *pb;
+        *pb = pos;
+      }
+      *pl = -pos;               /* Label exists now. */
+      b[pos++] = ofs;           /* Store pass1 offset estimate. */
+      break;
+    case DASM_IMM8:
+      b[pos++] = n;
+      break;
+    case DASM_IMM16:
+      CK(((short)n) == n || ((unsigned short)n) == n, RANGE_I);     /* TODO: is this the right way to handle unsigned immediates? */
+      ofs += 2;
+      b[pos++] = n;
+      break;
+    case DASM_IMM32:
+      ofs += 4;
+      b[pos++] = n;
+      break;
+    case DASM_DISP20:
+      CK(-(1 << 19) <= n && n < (1 << 19), RANGE_I);
+      b[pos++] = n;
+      break;
+    case DASM_DISP12:
+      CK((n >> 12) == 0, RANGE_I);
+      b[pos++] = n;
+      break;
+    case DASM_LEN8R:
+      CK(n >= 1 && n <= 256, RANGE_I);
+      b[pos++] = n;
+      break;
+    case DASM_LEN4HR:
+    case DASM_LEN4LR:
+      CK(n >= 1 && n <= 128, RANGE_I);
+      b[pos++] = n;
+      break;
+    }
+  }
+stop:
+  va_end(ap);
+  sec->pos = pos;
+  sec->ofs = ofs;
+}
+
+#undef CK
+
+/* Pass 2: Link sections, shrink aligns, fix label offsets. */
+int dasm_link(Dst_DECL, size_t * szp)
+{
+  dasm_State *D = Dst_REF;
+  int secnum;
+  int ofs = 0;
+
+#ifdef DASM_CHECKS
+  *szp = 0;
+  if (D->status != DASM_S_OK)
+    return D->status;
+  {
+    int pc;
+    for (pc = 0; pc * sizeof(int) < D->pcsize; pc++)
+      if (D->pclabels[pc] > 0)
+        return DASM_S_UNDEF_PC | pc;
+  }
+#endif
+
+  {                             /* Handle globals not defined in this translation unit. */
+    int idx;
+    for (idx = 20; idx * sizeof(int) < D->lgsize; idx++) {
+      int n = D->lglabels[idx];
+      /* Undefined label: Collapse rel chain and replace with marker (< 0). */
+      while (n > 0) {
+        int *pb = DASM_POS2PTR(D, n);
+        n = *pb;
+        *pb = -idx;
+      }
+    }
+  }
+
+  /* Combine all code sections. No support for data sections (yet). */
+  for (secnum = 0; secnum < D->maxsection; secnum++) {
+    dasm_Section *sec = D->sections + secnum;
+    int *b = sec->rbuf;
+    int pos = DASM_SEC2POS(secnum);
+    int lastpos = sec->pos;
+
+    while (pos != lastpos) {
+      dasm_ActList p = D->actionlist + b[pos++];
+      while (1) {
+        unsigned short ins = *p++;
+        unsigned short action = ins;
+        switch (action) {
+        case DASM_STOP:
+        case DASM_SECTION:
+          goto stop;
+        case DASM_ESC:
+          p++;
+          break;
+        case DASM_REL_EXT:
+          p++;
+          break;
+        case DASM_ALIGN:
+          ofs -= (b[pos++] + ofs) & *p++;
+          break;
+        case DASM_REL_LG:
+        case DASM_REL_PC:
+          p++;
+          pos++;
+          break;
+        case DASM_LABEL_LG:
+        case DASM_LABEL_PC:
+          p++;
+          b[pos++] += ofs;
+          break;
+        case DASM_IMM8:
+        case DASM_IMM16:
+        case DASM_IMM32:
+        case DASM_DISP20:
+        case DASM_DISP12:
+        case DASM_LEN8R:
+        case DASM_LEN4HR:
+        case DASM_LEN4LR:
+          pos++;
+          break;
+        }
+      }
+    stop:(void)0;
+    }
+    ofs += sec->ofs;            /* Next section starts right after current section. */
+  }
+
+  D->codesize = ofs;            /* Total size of all code sections */
+  *szp = ofs;
+  return DASM_S_OK;
+}
+
+#ifdef DASM_CHECKS
+#define CK(x, st) \
+  do { if (!(x)) return DASM_S_##st|(p-D->actionlist-1); } while (0)
+#else
+#define CK(x, st)	((void)0)
+#endif
+
+/* Pass 3: Encode sections. */
+int dasm_encode(Dst_DECL, void *buffer)
+{
+  dasm_State *D = Dst_REF;
+  char *base = (char *)buffer;
+  unsigned short *cp = (unsigned short *)buffer;
+  int secnum;
+
+  /* Encode all code sections. No support for data sections (yet). */
+  for (secnum = 0; secnum < D->maxsection; secnum++) {
+    dasm_Section *sec = D->sections + secnum;
+    int *b = sec->buf;
+    int *endb = sec->rbuf + sec->pos;
+
+    while (b != endb) {
+      dasm_ActList p = D->actionlist + *b++;
+      while (1) {
+        unsigned short ins = *p++;
+        unsigned short action = ins;
+        int n = (action >= DASM_ALIGN && action < DASM__MAX) ? *b++ : 0;
+        switch (action) {
+        case DASM_STOP:
+        case DASM_SECTION:
+          goto stop;
+        case DASM_ESC:
+          *cp++ = *p++;
+          break;
+        case DASM_REL_EXT:
+          n = DASM_EXTERN(Dst, (unsigned char *)cp, *p++, 1) - 4;
+          goto patchrel;
+        case DASM_ALIGN:
+          ins = *p++;
+          /* TODO: emit 4-byte noprs instead of 2-byte nops where possible. */
+          while ((((char *)cp - base) & ins))
+            *cp++ = 0x0700;     /* nop */
+          break;
+        case DASM_REL_LG:
+          CK(n >= 0, UNDEF_LG);
+        case DASM_REL_PC:
+          CK(n >= 0, UNDEF_PC);
+          n = *DASM_POS2PTR(D, n) - (int)((char *)cp - base);
+          p++;                  /* skip argument */
+        patchrel:
+          /* Offsets are halfword aligned (so need to be halved). */
+          n += 2;               /* Offset is relative to start of instruction. */
+          if (cp[-1] >> 12 == 0xc) {
+            *cp++ = n >> 17;
+          } else {
+            CK(-(1 << 16) <= n && n < (1 << 16) && (n & 1) == 0, RANGE_LG);
+          }
+          *cp++ = n >> 1;
+          break;
+        case DASM_LABEL_LG:
+          ins = *p++;
+          if (ins >= 20)
+            D->globals[ins - 10] = (void *)(base + n);
+          break;
+        case DASM_LABEL_PC:
+          break;
+        case DASM_IMM8:
+          cp[-1] |= n & 0xff;
+          break;
+        case DASM_IMM16:
+          *cp++ = n;
+          break;
+        case DASM_IMM32:
+          *cp++ = n >> 16;
+          *cp++ = n;
+          break;
+        case DASM_DISP20:
+          cp[-2] |= n & 0xfff;
+          cp[-1] |= (n >> 4) & 0xff00;
+          break;
+        case DASM_DISP12:
+          cp[-1] |= n & 0xfff;
+          break;
+        case DASM_LEN8R:
+          cp[-1] |= (n - 1) & 0xff;
+          break;
+        case DASM_LEN4HR:
+          cp[-1] |= ((n - 1) << 4) & 0xf0;
+          break;
+        case DASM_LEN4LR:
+          cp[-1] |= (n - 1) & 0x0f;
+          break;
+        default:
+          *cp++ = ins;
+          break;
+        }
+      }
+    stop:(void)0;
+    }
+  }
+
+  if (base + D->codesize != (char *)cp) /* Check for phase errors. */
+    return DASM_S_PHASE;
+  return DASM_S_OK;
+}
+
+#undef CK
+
+/* Get PC label offset. */
+int dasm_getpclabel(Dst_DECL, unsigned int pc)
+{
+  dasm_State *D = Dst_REF;
+  if (pc * sizeof(int) < D->pcsize) {
+    int pos = D->pclabels[pc];
+    if (pos < 0)
+      return *DASM_POS2PTR(D, -pos);
+    if (pos > 0)
+      return -1;                /* Undefined. */
+  }
+  return -2;                    /* Unused or out of range. */
+}
+
+#ifdef DASM_CHECKS
+/* Optional sanity checker to call between isolated encoding steps. */
+int dasm_checkstep(Dst_DECL, int secmatch)
+{
+  dasm_State *D = Dst_REF;
+  if (D->status == DASM_S_OK) {
+    int i;
+    for (i = 1; i <= 9; i++) {
+      if (D->lglabels[i] > 0) {
+        D->status = DASM_S_UNDEF_LG | i;
+        break;
+      }
+      D->lglabels[i] = 0;
+    }
+  }
+  if (D->status == DASM_S_OK && secmatch >= 0 &&
+      D->section != &D->sections[secmatch])
+    D->status = DASM_S_MATCH_SEC | (D->section - D->sections);
+  return D->status;
+}
+#endif

src/Makefile

@@ -243,6 +243,9 @@ else
 ifneq (,$(findstring LJ_TARGET_ARM ,$(TARGET_TESTARCH)))
   TARGET_LJARCH= arm
 else
+ifneq (,$(findstring LJ_TARGET_S390X ,$(TARGET_TESTARCH)))
+  TARGET_LJARCH= s390x
+else
 ifneq (,$(findstring LJ_TARGET_ARM64 ,$(TARGET_TESTARCH)))
   ifneq (,$(findstring __AARCH64EB__ ,$(TARGET_TESTARCH)))
     TARGET_ARCH= -D__AARCH64EB__=1
@@ -274,6 +277,7 @@ endif
 endif
 endif
 endif
+endif
 
 ifneq (,$(findstring LJ_TARGET_PS3 1,$(TARGET_TESTARCH)))
   TARGET_SYS= PS3

src/Makefile.dep

@@ -25,7 +25,7 @@ lib_ffi.o: lib_ffi.c lua.h luaconf.h lauxlib.h lualib.h lj_obj.h lj_def.h \
 lib_init.o: lib_init.c lua.h luaconf.h lauxlib.h lualib.h lj_arch.h
 lib_io.o: lib_io.c lua.h luaconf.h lauxlib.h lualib.h lj_obj.h lj_def.h \
  lj_arch.h lj_gc.h lj_err.h lj_errmsg.h lj_buf.h lj_str.h lj_state.h \
- lj_strfmt.h lj_ff.h lj_ffdef.h lj_lib.h lj_libdef.h
+ lj_strfmt.h lj_ff.h lj_ffdef.h lj_lib.h lj_strscan.h lj_libdef.h
 lib_jit.o: lib_jit.c lua.h luaconf.h lauxlib.h lualib.h lj_obj.h lj_def.h \
  lj_arch.h lj_gc.h lj_err.h lj_errmsg.h lj_debug.h lj_str.h lj_tab.h \
  lj_state.h lj_bc.h lj_ctype.h lj_ir.h lj_jit.h lj_ircall.h lj_iropt.h \

src/host/buildvm.c

@@ -67,6 +67,8 @@ static int collect_reloc(BuildCtx *ctx, uint8_t *addr, int idx, int type);
 #include "../dynasm/dasm_ppc.h"
 #elif LJ_TARGET_MIPS
 #include "../dynasm/dasm_mips.h"
+#elif LJ_TARGET_S390X
+#include "../dynasm/dasm_s390x.h"
 #else
 #error "No support for this architecture (yet)"
 #endif

src/host/buildvm_asm.c

@@ -87,6 +87,54 @@ err:
   }
   fprintf(ctx->fp, "\t%s %s\n", opname, sym);
 }
+#elif LJ_TARGET_S390X
+/* Emit halfwords piecewise as assembler text. */
+static void emit_asm_halfwords(BuildCtx *ctx, uint8_t *p, int n)
+{
+  uint16_t *cp = (uint16_t*)p;
+  n /= 2;
+  int i;
+  for (i = 0; i < n; i++) {
+    if ((i & 7) == 0)
+      fprintf(ctx->fp, "\t.hword 0x%hx", cp[i]);
+    else
+      fprintf(ctx->fp, ",0x%hx", cp[i]);
+    if ((i & 7) == 7) putc('\n', ctx->fp);
+  }
+  if ((n & 7) != 0) putc('\n', ctx->fp);
+}
+
+/* Emit s390x text relocations. */
+static void emit_asm_reloc_text(BuildCtx *ctx, uint8_t *cp, int n,
+				const char *sym)
+{
+  if (n & 1 || n < 2) {
+    fprintf(stderr, "Error: instruction stream length invalid: %d.\n", n);
+    exit(1);
+  }
+  n -= 2;
+  const char *opname = NULL;
+  const char *argt = ""; /* Inserted before argument. */
+  int opcode = *(uint16_t*)(&cp[n]);
+  int arg = (opcode>>4) & 0xf;
+  switch (opcode & 0xff0f) {
+  case 0xa705: opname = "bras"; argt = "%r"; break;
+  case 0xc005: opname = "brasl"; argt = "%r"; break;
+  case 0xa704: opname = "brc"; break;
+  case 0xc004: opname = "brcl"; break;
+  default:
+    fprintf(stderr, "Error: unsupported opcode for %s symbol relocation.\n",
+	    sym);
+    exit(1);
+  }
+  emit_asm_halfwords(ctx, cp, n);
+  if (strncmp(sym+(*sym == '_'), LABEL_PREFIX, sizeof(LABEL_PREFIX)-1)) {
+    /* Various fixups for external symbols outside of our binary. */
+    fprintf(ctx->fp, "\t%s %s%d, %s@PLT\n", opname, argt, arg, sym);
+    return;
+  }
+  fprintf(ctx->fp, "\t%s %s%d, %s\n", opname, argt, arg, sym);
+}
 #else
 /* Emit words piecewise as assembler text. */
 static void emit_asm_words(BuildCtx *ctx, uint8_t *p, int n)
@@ -302,6 +350,9 @@ void emit_asm(BuildCtx *ctx)
 	emit_asm_reloc(ctx, r->type, ctx->relocsym[r->sym]);
       }
       ofs += n+4;
+#elif LJ_TARGET_S390X
+      emit_asm_reloc_text(ctx, ctx->code+ofs, n, ctx->relocsym[r->sym]);
+      ofs += n+4;
 #else
       emit_asm_wordreloc(ctx, ctx->code+ofs, n, ctx->relocsym[r->sym]);
       ofs += n;
@@ -310,6 +361,8 @@ void emit_asm(BuildCtx *ctx)
     }
 #if LJ_TARGET_X86ORX64
     emit_asm_bytes(ctx, ctx->code+ofs, next-ofs);
+#elif LJ_TARGET_S390X
+    emit_asm_halfwords(ctx, ctx->code+ofs, next-ofs);
 #else
     emit_asm_words(ctx, ctx->code+ofs, next-ofs);
 #endif

src/jit/bcsave.lua

@@ -101,6 +101,7 @@ local map_arch = {
   mips64el =	{ e = "le", b = 64, m = 8, f = 0x80000007, },
   mips64r6 =	{ e = "be", b = 64, m = 8, f = 0xa0000407, },
   mips64r6el =	{ e = "le", b = 64, m = 8, f = 0xa0000407, },
+  s390x =	{ e = "be", b = 64, m = 22, },
 }
 
 local map_os = {

src/jit/dis_s390x.lua

@@ -0,0 +1 @@
+-- Not yet implemented.

src/lib_io.c

@@ -25,6 +25,7 @@
 #include "lj_strfmt.h"
 #include "lj_ff.h"
 #include "lj_lib.h"
+#include "lj_strscan.h"
 
 /* Userdata payload for I/O file. */
 typedef struct IOFileUD {
@@ -323,13 +324,14 @@ LJLIB_CF(io_method_seek)
   FILE *fp = io_tofile(L)->fp;
   int opt = lj_lib_checkopt(L, 2, 1, "\3set\3cur\3end");
   int64_t ofs = 0;
-  cTValue *o;
+  TValue *o;
   int res;
   if (opt == 0) opt = SEEK_SET;
   else if (opt == 1) opt = SEEK_CUR;
   else if (opt == 2) opt = SEEK_END;
   o = L->base+2;
   if (o < L->top) {
+    if (tvisstr(o)) lj_strscan_num(strV(o), o);
     if (tvisint(o))
       ofs = (int64_t)intV(o);
     else if (tvisnum(o))

src/lib_jit.c

@@ -702,6 +702,8 @@ static uint32_t jit_cpudetect(void)
   }
 #endif
 
+#elif LJ_TARGET_S390X
+  /* No optional CPU features to detect (for now). */
 #else
 #error "Missing CPU detection for this architecture"
 #endif

src/lj_arch.h

@@ -31,6 +31,8 @@
 #define LUAJIT_ARCH_mips32	6
 #define LUAJIT_ARCH_MIPS64	7
 #define LUAJIT_ARCH_mips64	7
+#define LUAJIT_ARCH_S390X	8
+#define LUAJIT_ARCH_s390x	8
 
 /* Target OS. */
 #define LUAJIT_OS_OTHER		0
@@ -59,6 +61,8 @@
 #define LUAJIT_TARGET	LUAJIT_ARCH_ARM
 #elif defined(__aarch64__) || defined(_M_ARM64)
 #define LUAJIT_TARGET	LUAJIT_ARCH_ARM64
+#elif defined(__s390x__) || defined(__s390x)
+#define LUAJIT_TARGET	LUAJIT_ARCH_S390X
 #elif defined(__ppc__) || defined(__ppc) || defined(__PPC__) || defined(__PPC) || defined(__powerpc__) || defined(__powerpc) || defined(__POWERPC__) || defined(__POWERPC) || defined(_M_PPC)
 #define LUAJIT_TARGET	LUAJIT_ARCH_PPC
 #elif defined(__mips64__) || defined(__mips64) || defined(__MIPS64__) || defined(__MIPS64)
@@ -439,6 +443,21 @@
 #define LJ_ARCH_VERSION		10
 #endif
 
+#elif LUAJIT_TARGET == LUAJIT_ARCH_S390X
+
+#define LJ_ARCH_NAME		"s390x"
+#define LJ_ARCH_BITS		64
+#define LJ_ARCH_ENDIAN		LUAJIT_BE
+#define LJ_TARGET_S390X		1
+#define LJ_TARGET_EHRETREG	0xe
+#define LJ_TARGET_JUMPRANGE	32	/* +-2^32 = +-4GB (32-bit, halfword aligned) */
+#define LJ_TARGET_MASKSHIFT	1
+#define LJ_TARGET_MASKROT	1
+#define LJ_TARGET_UNALIGNED	1
+#define LJ_ARCH_NUMMODE		LJ_NUMMODE_DUAL
+#define LJ_TARGET_GC64		1
+#define LJ_ARCH_NOJIT		1	/* NYI */
+
 #else
 #error "No target architecture defined"
 #endif

src/lj_asm.c

@@ -1708,6 +1708,8 @@ static void asm_loop(ASMState *as)
 #include "lj_asm_ppc.h"
 #elif LJ_TARGET_MIPS
 #include "lj_asm_mips.h"
+#elif LJ_TARGET_S390X
+#include "lj_asm_s390x.h"
 #else
 #error "Missing assembler for target CPU"
 #endif

src/lj_bcread.c

@@ -179,7 +179,7 @@ static const void *bcread_varinfo(GCproto *pt)
 }
 
 /* Read a single constant key/value of a template table. */
-static void bcread_ktabk(LexState *ls, TValue *o)
+static void bcread_ktabk(LexState *ls, TValue *o, GCtab *t)
 {
   MSize tp = bcread_uleb128(ls);
   if (tp >= BCDUMP_KTAB_STR) {
@@ -191,6 +191,8 @@ static void bcread_ktabk(LexState *ls, TValue *o)
   } else if (tp == BCDUMP_KTAB_NUM) {
     o->u32.lo = bcread_uleb128(ls);
     o->u32.hi = bcread_uleb128(ls);
+  } else if (tp == BCDUMP_KTAB_NIL) { /* Restore nil value marker. */
+    settabV(ls->L, o, t);
   } else {
     lj_assertLS(tp <= BCDUMP_KTAB_TRUE, "bad constant type %d", tp);
     setpriV(o, ~tp);
@@ -207,15 +209,15 @@ static GCtab *bcread_ktab(LexState *ls)
     MSize i;
     TValue *o = tvref(t->array);
     for (i = 0; i < narray; i++, o++)
-      bcread_ktabk(ls, o);
+      bcread_ktabk(ls, o, t);
   }
   if (nhash) {  /* Read hash entries. */
     MSize i;
     for (i = 0; i < nhash; i++) {
       TValue key;
-      bcread_ktabk(ls, &key);
+      bcread_ktabk(ls, &key, t);
       lj_assertLS(!tvisnil(&key), "nil key");
-      bcread_ktabk(ls, lj_tab_set(ls->L, t, &key));
+      bcread_ktabk(ls, lj_tab_set(ls->L, t, &key), t);
     }
   }
   return t;

src/lj_bcwrite.c

@@ -71,6 +71,8 @@ static void bcwrite_ktabk(BCWriteCtx *ctx, cTValue *o, int narrow)
     *p++ = BCDUMP_KTAB_NUM;
     p = lj_strfmt_wuleb128(p, o->u32.lo);
     p = lj_strfmt_wuleb128(p, o->u32.hi);
+  } else if (tvistab(o)) { /* Write the nil value marker as a nil. */
+    *p++ = BCDUMP_KTAB_NIL;
   } else {
     lj_assertBCW(tvispri(o), "unhandled type %d", itype(o));
     *p++ = BCDUMP_KTAB_NIL+~itype(o);
@@ -133,7 +135,7 @@ static void bcwrite_ktab_sorted_hash(BCWriteCtx *ctx, Node *node, MSize nhash)
   TValue **heap = ctx->heap;
   MSize i = nhash;
   for (;; node--) {  /* Build heap. */
-    if (!tvisnil(&node->key)) {
+    if (!tvisnil(&node->val)) {
       bcwrite_ktabk_heap_insert(heap, --i, nhash, &node->key);
       if (i == 0) break;
     }
@@ -163,7 +165,7 @@ static void bcwrite_ktab(BCWriteCtx *ctx, char *p, const GCtab *t)
     MSize i, hmask = t->hmask;
     Node *node = noderef(t->node);
     for (i = 0; i <= hmask; i++)
-      nhash += !tvisnil(&node[i].key);
+      nhash += !tvisnil(&node[i].val);
   }
   /* Write number of array slots and hash slots. */
   p = lj_strfmt_wuleb128(p, narray);
@@ -184,7 +186,7 @@ static void bcwrite_ktab(BCWriteCtx *ctx, char *p, const GCtab *t)
     } else {
       MSize i = nhash;
       for (;; node--)
-	if (!tvisnil(&node->key)) {
+	if (!tvisnil(&node->val)) {
 	  bcwrite_ktabk(ctx, &node->key, 0);
 	  bcwrite_ktabk(ctx, &node->val, 1);
 	  if (--i == 0) break;

src/lj_ccall.c

@@ -575,6 +575,40 @@
     goto done; \
   }
 
+#elif LJ_TARGET_S390X
+/* -- POSIX/s390x calling conventions --------------------------------------- */
+
+#define CCALL_HANDLE_STRUCTRET \
+  cc->retref = 1;  /* Return all structs by reference. */ \
+  cc->gpr[ngpr++] = (GPRArg)dp;
+
+#define CCALL_HANDLE_COMPLEXRET \
+  cc->retref = 1;  /* Return all complex values by reference. */ \
+  cc->gpr[ngpr++] = (GPRArg)dp;
+
+#define CCALL_HANDLE_COMPLEXRET2 \
+  UNUSED(dp); /* Nothing to do. */
+
+#define CCALL_HANDLE_STRUCTARG \
+  /* Pass structs of size 1, 2, 4 or 8 in a GPR by value. */ \
+  if (!(sz == 1 || sz == 2 || sz == 4 || sz == 8)) { \
+    rp = cdataptr(lj_cdata_new(cts, did, sz)); \
+    sz = CTSIZE_PTR;  /* Pass all other structs by reference. */ \
+  }
+
+#define CCALL_HANDLE_COMPLEXARG \
+  /* Pass complex numbers by reference. */ \
+  /* TODO: not sure why this is different to structs. */ \
+  rp = cdataptr(lj_cdata_new(cts, did, sz)); \
+  sz = CTSIZE_PTR; \
+
+#define CCALL_HANDLE_REGARG \
+  if (isfp) { \
+    if (nfpr < CCALL_NARG_FPR) { dp = &cc->fpr[nfpr++]; goto done; } \
+  } else { \
+    if (ngpr < maxgpr) { dp = &cc->gpr[ngpr++]; goto done; } \
+  }
+
 #else
 #error "Missing calling convention definitions for this architecture"
 #endif
@@ -999,6 +1033,9 @@ static int ccall_set_args(lua_State *L, CTState *cts, CType *ct,
     CType *d;
     CTSize sz;
     MSize n, isfp = 0, isva = 0;
+#if LJ_TARGET_S390X
+    MSize onstack = 0;
+#endif
     void *dp, *rp = NULL;
 
     if (fid) {  /* Get argument type from field. */
@@ -1037,6 +1074,9 @@ static int ccall_set_args(lua_State *L, CTState *cts, CType *ct,
     CCALL_HANDLE_REGARG  /* Handle register arguments. */
 
     /* Otherwise pass argument on stack. */
+#if LJ_TARGET_S390X
+    onstack = 1;
+#endif
     if (CCALL_ALIGN_STACKARG) {  /* Align argument on stack. */
       MSize align = (1u << ctype_align(d->info)) - 1;
       if (rp || (CCALL_PACK_STACKARG && isva && align < CTSIZE_PTR-1))
@@ -1086,6 +1126,16 @@ static int ccall_set_args(lua_State *L, CTState *cts, CType *ct,
       *(int64_t *)dp = (int64_t)*(int32_t *)dp;  /* Sign-extend to 64 bit. */
     }
 #endif
+#if LJ_TARGET_S390X
+    /* Arguments need to be sign-/zero-extended to 64-bits. */
+    if ((ctype_isinteger_or_bool(d->info) || ctype_isenum(d->info) ||
+          (isfp && onstack)) && d->size <= 4) {
+      if (d->info & CTF_UNSIGNED || isfp)
+        *(uint64_t *)dp = (uint64_t)*(uint32_t *)dp;
+      else
+        *(int64_t *)dp = (int64_t)*(int32_t *)dp;
+    }
+#endif
 #if LJ_TARGET_X64 && LJ_ABI_WIN
     if (isva) {  /* Windows/x64 mirrors varargs in both register sets. */
       if (nfpr == ngpr)

src/lj_ccall.h

@@ -129,6 +129,21 @@ typedef union FPRArg {
   struct { LJ_ENDIAN_LOHI(float f; , float g;) };
 } FPRArg;
 
+#elif LJ_TARGET_S390X
+
+#define CCALL_NARG_GPR		5	/* GPR 2,3,4,5,6 */
+#define CCALL_NARG_FPR		4	/* FPR 0,2,4,8 */
+#define CCALL_NRET_GPR		1	/* GPR 2 */
+#define CCALL_NRET_FPR		1	/* FPR 0 */
+#define CCALL_SPS_EXTRA		20	/* 160-byte callee save area (not sure if this is the right place) */
+#define CCALL_SPS_FREE		0
+
+typedef intptr_t GPRArg;
+typedef union FPRArg {
+  double d;
+  float f;
+} FPRArg;
+
 #else
 #error "Missing calling convention definitions for this architecture"
 #endif

src/lj_ccallback.c

@@ -262,6 +262,14 @@ static void *callback_mcode_init(global_State *g, uint32_t *page)
 #define CCPROT_CREATE	0
 #endif
 
+/* Check for macOS hardened runtime. */
+#if LUAJIT_SECURITY_MCODE != 0 && defined(MAP_JIT) && __ENVIRONMENT_MAC_OS_X_VERSION_MIN_REQUIRED__ >= 110000
+#include <pthread.h>
+#define CCMAP_CREATE	MAP_JIT
+#else
+#define CCMAP_CREATE	0
+#endif
+
 #endif
 
 /* Allocate and initialize area for callback function pointers. */
@@ -276,10 +284,13 @@ static void callback_mcode_new(CTState *cts)
   if (!p)
     lj_err_caller(cts->L, LJ_ERR_FFI_CBACKOV);
 #elif LJ_TARGET_POSIX
-  p = mmap(NULL, sz, (PROT_READ|PROT_WRITE|CCPROT_CREATE), MAP_PRIVATE|MAP_ANONYMOUS,
-	   -1, 0);
+  p = mmap(NULL, sz, PROT_READ|PROT_WRITE|CCPROT_CREATE,
+	   MAP_PRIVATE|MAP_ANONYMOUS|CCMAP_CREATE, -1, 0);
   if (p == MAP_FAILED)
     lj_err_caller(cts->L, LJ_ERR_FFI_CBACKOV);
+#if CCMAP_CREATE
+  pthread_jit_write_protect_np(0);
+#endif
 #else
   /* Fallback allocator. Fails if memory is not executable by default. */
   p = lj_mem_new(cts->L, sz);
@@ -296,8 +307,12 @@ static void callback_mcode_new(CTState *cts)
     LJ_WIN_VPROTECT(p, sz, PAGE_EXECUTE_READ, &oprot);
   }
 #elif LJ_TARGET_POSIX
+#if CCMAP_CREATE
+  pthread_jit_write_protect_np(1);
+#else
   mprotect(p, sz, (PROT_READ|PROT_EXEC));
 #endif
+#endif
 }
 
 /* Free area for callback function pointers. */
@@ -516,6 +531,15 @@ void lj_ccallback_mcode_free(CTState *cts)
   if (ctype_isfp(ctr->info) && ctr->size == sizeof(float)) \
     ((float *)dp)[1] = *(float *)dp;
 
+#elif LJ_TARGET_S390X
+
+#define CALLBACK_HANDLE_REGARG \
+  if (isfp) { \
+    if (nfpr < CCALL_NARG_FPR) { sp = &cts->cb.fpr[nfpr++]; goto done; } \
+  } else { \
+    if (ngpr < maxgpr) { sp = &cts->cb.gpr[ngpr++]; goto done; } \
+  }
+
 #else
 #error "Missing calling convention definitions for this architecture"
 #endif

src/lj_ctype.c

@@ -33,10 +33,12 @@
   _("int16_t",			INT16) \
   _("int32_t",			INT32) \
   _("int64_t",			INT64) \
+  _("int128_t",			INT128) \
   _("uint8_t",			UINT8) \
   _("uint16_t",			UINT16) \
   _("uint32_t",			UINT32) \
   _("uint64_t",			UINT64) \
+  _("uint128_t",		UINT128) \
   _("intptr_t",			INT_PSZ) \
   _("uintptr_t",		UINT_PSZ) \
   /* From POSIX. */ \
@@ -55,6 +57,7 @@
   _("__int16",		2,	CTOK_INT) \
   _("__int32",		4,	CTOK_INT) \
   _("__int64",		8,	CTOK_INT) \
+  _("__int128",		16,	CTOK_INT) \
   _("float",		4,	CTOK_FP) \
   _("double",		8,	CTOK_FP) \
   _("long",		0,	CTOK_LONG) \

src/lj_ctype.h

@@ -292,6 +292,8 @@ typedef struct CTState {
   _(UINT32,		4,	CT_NUM, CTF_UNSIGNED|CTALIGN(2)) \
   _(INT64,		8,	CT_NUM, CTF_LONG_IF8|CTALIGN(3)) \
   _(UINT64,		8,	CT_NUM, CTF_UNSIGNED|CTF_LONG_IF8|CTALIGN(3)) \
+  _(INT128,		16,	CT_NUM, CTALIGN(4)) \
+  _(UINT128,		16,	CT_NUM, CTF_UNSIGNED|CTALIGN(4)) \
   _(FLOAT,		4,	CT_NUM, CTF_FP|CTALIGN(2)) \
   _(DOUBLE,		8,	CT_NUM, CTF_FP|CTALIGN(3)) \
   _(COMPLEX_FLOAT,	8,	CT_ARRAY, CTF_COMPLEX|CTALIGN(2)|CTID_FLOAT) \

src/lj_err.c

@@ -442,6 +442,9 @@ LJ_FUNCA int lj_err_unwind_dwarf(int version, int actions,
   if (version != 1)
     return _URC_FATAL_PHASE1_ERROR;
   cf = (void *)_Unwind_GetCFA(ctx);
+#ifdef LJ_TARGET_S390X
+  cf -= 160; /* CFA points 160 bytes above r15. */
+#endif
   L = cframe_L(cf);
   if ((actions & _UA_SEARCH_PHASE)) {
 #if LJ_UNWIND_EXT

src/lj_frame.h

@@ -264,6 +264,20 @@ enum { LJ_CONT_TAILCALL, LJ_CONT_FFI_CALLBACK };  /* Special continuations. */
 #endif
 #define CFRAME_OFS_MULTRES	0
 #define CFRAME_SHIFT_MULTRES	3
+#elif LJ_TARGET_S390X
+#define CFRAME_OFS_ERRF		280
+#define CFRAME_OFS_NRES		272
+#define CFRAME_OFS_PREV		264
+#define CFRAME_OFS_L		256
+#define CFRAME_OFS_PC		168
+#define CFRAME_OFS_MULTRES	160
+#define CFRAME_SIZE		240
+/*
+** TODO: it would be good if we always decoded param*8 like
+** the RISC architectures do. If so then SHIFT_MULTRES will
+** need to change to 3.
+*/
+#define CFRAME_SHIFT_MULTRES	0
 #else
 #error "Missing CFRAME_* definitions for this architecture"
 #endif

src/lj_mcode.c

@@ -98,6 +98,14 @@ static int mcode_setprot(void *p, size_t sz, DWORD prot)
 #define MAP_ANONYMOUS	MAP_ANON
 #endif
 
+/* Check for macOS hardened runtime. */
+#if LUAJIT_SECURITY_MCODE != 0 && defined(MAP_JIT) && __ENVIRONMENT_MAC_OS_X_VERSION_MIN_REQUIRED__ >= 110000
+#include <pthread.h>
+#define MCMAP_CREATE	MAP_JIT
+#else
+#define MCMAP_CREATE	0
+#endif
+
 #define MCPROT_RW	(PROT_READ|PROT_WRITE)
 #define MCPROT_RX	(PROT_READ|PROT_EXEC)
 #define MCPROT_RWX	(PROT_READ|PROT_WRITE|PROT_EXEC)
@@ -109,10 +117,14 @@ static int mcode_setprot(void *p, size_t sz, DWORD prot)
 
 static void *mcode_alloc_at(jit_State *J, uintptr_t hint, size_t sz, int prot)
 {
-  void *p = mmap((void *)hint, sz, prot|MCPROT_CREATE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0);
+  void *p = mmap((void *)hint, sz, prot|MCPROT_CREATE, MAP_PRIVATE|MAP_ANONYMOUS|MCMAP_CREATE, -1, 0);
   if (p == MAP_FAILED) {
     if (!hint) lj_trace_err(J, LJ_TRERR_MCODEAL);
     p = NULL;
+#if MCMAP_CREATE
+  } else {
+    pthread_jit_write_protect_np(0);
+#endif
   }
   return p;
 }
@@ -125,7 +137,12 @@ static void mcode_free(jit_State *J, void *p, size_t sz)
 
 static int mcode_setprot(void *p, size_t sz, int prot)
 {
+#if MCMAP_CREATE
+  pthread_jit_write_protect_np((prot & PROT_EXEC));
+  return 0;
+#else
   return mprotect(p, sz, prot);
+#endif
 }
 
 #else

src/lj_opt_fold.c

@@ -2217,9 +2217,11 @@ LJFOLD(HREF TDUP KNUM)
 LJFOLDF(fwd_href_tdup)
 {
   TValue keyv;
+  cTValue *val;
   lj_ir_kvalue(J->L, &keyv, fright);
-  if (lj_tab_get(J->L, ir_ktab(IR(fleft->op1)), &keyv) == niltvg(J2G(J)) &&
-      lj_opt_fwd_href_nokey(J))
+  val = lj_tab_get(J->L, ir_ktab(IR(fleft->op1)), &keyv);
+  /* Check for either nil or the nil value marker in the template table. */
+  if ((tvisnil(val) || tvistab(val)) && lj_opt_fwd_href_nokey(J))
     return lj_ir_kkptr(J, niltvg(J2G(J)));
   return NEXTFOLD;
 }

src/lj_opt_mem.c

@@ -233,7 +233,9 @@ static TRef fwd_ahload(jit_State *J, IRRef xref)
 	  return lj_ir_knum_u64(J, tv->u64);
 	else if (tvisint(tv))
 	  return lj_ir_kint(J, intV(tv));
-	else if (tvisgcv(tv))
+	else if (tvistab(tv)) /* Template table nil value marker. */
+	  return TREF_NIL;
+	else if (tvisstr(tv))
 	  return lj_ir_kstr(J, strV(tv));
       }
       /* Othwerwise: don't intern as a constant. */

src/lj_parse.c

@@ -1725,7 +1725,7 @@ static void expr_table(LexState *ls, ExpDesc *e)
   FuncState *fs = ls->fs;
   BCLine line = ls->linenumber;
   GCtab *t = NULL;
-  int vcall = 0, needarr = 0, fixt = 0;
+  int vcall = 0, needarr = 0;
   uint32_t narr = 1;  /* First array index. */
   uint32_t nhash = 0;  /* Number of hash entries. */
   BCReg freg = fs->freereg;
@@ -1769,9 +1769,10 @@ static void expr_table(LexState *ls, ExpDesc *e)
       lj_gc_anybarriert(fs->L, t);
       if (expr_isk_nojump(&val)) {  /* Add const key/value to template table. */
 	expr_kvalue(fs, v, &val);
-      } else {  /* Otherwise create dummy string key (avoids lj_tab_newkey). */
-	settabV(fs->L, v, t);  /* Preserve key with table itself as value. */
-	fixt = 1;   /* Fix this later, after all resizes. */
+	/* Mark nil value with table value itself to preserve the key. */
+	if (key.k == VKSTR && tvisnil(v)) settabV(fs->L, v, t);
+      } else {  /* Preserve the key for the following non-const store.  */
+	settabV(fs->L, v, t);
 	goto nonconst;
       }
     } else {
@@ -1813,17 +1814,6 @@ static void expr_table(LexState *ls, ExpDesc *e)
   } else {
     if (needarr && t->asize < narr)
       lj_tab_reasize(fs->L, t, narr-1);
-    if (fixt) {  /* Fix value for dummy keys in template table. */
-      Node *node = noderef(t->node);
-      uint32_t i, hmask = t->hmask;
-      for (i = 0; i <= hmask; i++) {
-	Node *n = &node[i];
-	if (tvistab(&n->val)) {
-	  lj_assertFS(tabV(&n->val) == t, "bad dummy key in template table");
-	  setnilV(&n->val);  /* Turn value into nil. */
-	}
-      }
-    }
     lj_gc_check(fs->L);
   }
 }

src/lj_record.c

@@ -2079,6 +2079,7 @@ static TRef rec_tnew(jit_State *J, uint32_t ah)
 /* -- Concatenation ------------------------------------------------------- */
 
 typedef struct RecCatDataCP {
+  TValue savetv[5+LJ_FR2];
   jit_State *J;
   BCReg baseslot, topslot;
   TRef tr;
@@ -2119,7 +2120,9 @@ static TValue *rec_mm_concat_cp(lua_State *L, lua_CFunction dummy, void *ud)
       return NULL;
     }
     /* Pass partial result. */
-    topslot = J->maxslot--;
+    rcd->topslot = topslot = J->maxslot--;
+    /* Save updated range of slots. */
+    memcpy(rcd->savetv, &L->base[topslot-1], sizeof(rcd->savetv));
     *xbase = tr;
     top = xbase;
     setstrV(J->L, &ix.keyv, &J2G(J)->strempty);  /* Simulate string result. */
@@ -2139,16 +2142,18 @@ static TRef rec_cat(jit_State *J, BCReg baseslot, BCReg topslot)
 {
   lua_State *L = J->L;
   ptrdiff_t delta = L->top - L->base;
-  TValue savetv[5+LJ_FR2], errobj;
+  TValue errobj;
   RecCatDataCP rcd;
   int errcode;
   rcd.J = J;
   rcd.baseslot = baseslot;
   rcd.topslot = topslot;
-  memcpy(savetv, &L->base[topslot-1], sizeof(savetv));  /* Save slots. */
+  /* Save slots. */
+  memcpy(rcd.savetv, &L->base[topslot-1], sizeof(rcd.savetv));
   errcode = lj_vm_cpcall(L, NULL, &rcd, rec_mm_concat_cp);
   if (errcode) copyTV(L, &errobj, L->top-1);
-  memcpy(&L->base[topslot-1], savetv, sizeof(savetv));  /* Restore slots. */
+  /* Restore slots. */
+  memcpy(&L->base[rcd.topslot-1], rcd.savetv, sizeof(rcd.savetv));
   if (errcode) {
     L->top = L->base + delta;
     copyTV(L, L->top++, &errobj);

src/lj_tab.c

@@ -194,6 +194,7 @@ GCtab * LJ_FASTCALL lj_tab_dup(lua_State *L, const GCtab *kt)
       Node *next = nextnode(kn);
       /* Don't use copyTV here, since it asserts on a copy of a dead key. */
       n->val = kn->val; n->key = kn->key;
+      if (tvistab(&n->val)) setnilV(&n->val); /* Replace nil value marker. */
       setmref(n->next, next == NULL? next : (Node *)((char *)next + d));
     }
   }

src/lj_target.h

@@ -143,6 +143,8 @@ typedef uint32_t RegCost;
 #include "lj_target_ppc.h"
 #elif LJ_TARGET_MIPS
 #include "lj_target_mips.h"
+#elif LJ_TARGET_S390X
+#include "lj_target_s390x.h"
 #else
 #error "Missing include for target CPU"
 #endif

src/lj_target_s390x.h

@@ -0,0 +1,80 @@
+/*
+** Definitions for IBM z/Architecture (s390x) CPUs.
+** Copyright (C) 2005-2016 Mike Pall. See Copyright Notice in luajit.h
+*/
+
+#ifndef _LJ_TARGET_S390X_H
+#define _LJ_TARGET_S390X_H
+
+/* -- Registers IDs ------------------------------------------------------- */
+
+#define GPRDEF(_) \
+  _(R0) _(R1) _(R2) _(R3) _(R4) _(R5) _(R6) _(R7) \
+  _(R8) _(R9) _(R10) _(R11) _(R12) _(R13) _(R14) _(R15)
+#define FPRDEF(_) \
+  _(F0) _(F1) _(F2) _(F3) \
+  _(F4) _(F5) _(F6) _(F7) \
+  _(F8) _(F9) _(F10) _(F11) \
+  _(F12) _(F13) _(F14) _(F15)
+// TODO: VREG?
+
+#define RIDENUM(name)	RID_##name,
+
+enum {
+  GPRDEF(RIDENUM)		/* General-purpose registers (GPRs). */
+  FPRDEF(RIDENUM)		/* Floating-point registers (FPRs). */
+  RID_MAX,
+
+  /* Calling conventions. */
+  RID_SP = RID_R15,
+  RID_RET = RID_R2,
+  RID_FPRET = RID_F0,
+
+  /* These definitions must match with the *.dasc file(s): */
+  RID_BASE = RID_R7,		/* Interpreter BASE. */
+  RID_LPC = RID_R9,		/* Interpreter PC. */
+  RID_DISPATCH = RID_R10,	/* Interpreter DISPATCH table. */
+
+  /* Register ranges [min, max) and number of registers. */
+  RID_MIN_GPR = RID_R0,
+  RID_MIN_FPR = RID_F0,
+  RID_MAX_GPR = RID_MIN_FPR,
+  RID_MAX_FPR = RID_MAX,
+  RID_NUM_GPR = RID_MAX_GPR - RID_MIN_GPR,
+  RID_NUM_FPR = RID_MAX_FPR - RID_MIN_FPR,
+};
+
+/* -- Register sets ------------------------------------------------------- */
+
+/* -- Spill slots --------------------------------------------------------- */
+
+/* Spill slots are 32 bit wide. An even/odd pair is used for FPRs.
+**
+** SPS_FIXED: Available fixed spill slots in interpreter frame.
+** This definition must match with the *.dasc file(s).
+**
+** SPS_FIRST: First spill slot for general use. Reserve min. two 32 bit slots.
+*/
+#define SPS_FIXED	2
+#define SPS_FIRST	2
+
+#define SPOFS_TMP	0
+
+#define sps_scale(slot)		(4 * (int32_t)(slot))
+#define sps_align(slot)		(((slot) - SPS_FIXED + 1) & ~1)
+
+/* -- Exit state ---------------------------------------------------------- */
+
+/* This definition must match with the *.dasc file(s). */
+typedef struct {
+  lua_Number fpr[RID_NUM_FPR];	/* Floating-point registers. */
+  int32_t gpr[RID_NUM_GPR];	/* General-purpose registers. */
+  int32_t spill[256];		/* Spill slots. */
+} ExitState;
+
+#define EXITSTUB_SPACING        4
+#define EXITSTUBS_PER_GROUP     32
+
+/* -- Instructions -------------------------------------------------------- */
+
+#endif

src/lj_trace.c

@@ -222,14 +222,6 @@ static void trace_unpatch(jit_State *J, GCtrace *T)
 	       bc_isret(op), "bad original bytecode %d", op);
     *pc = T->startins;
     break;
-  case BC_JMP:
-    lj_assertJ(op == BC_ITERL, "bad original bytecode %d", op);
-    pc += bc_j(*pc)+2;
-    if (bc_op(*pc) == BC_JITERL) {
-      lj_assertJ(traceref(J, bc_d(*pc)) == T, "JITERL references other trace");
-      *pc = T->startins;
-    }
-    break;
   case BC_JFUNCF:
     lj_assertJ(op == BC_FUNCF, "bad original bytecode %d", op);
     *pc = T->startins;
@@ -245,18 +237,19 @@ static void trace_flushroot(jit_State *J, GCtrace *T)
   GCproto *pt = &gcref(T->startpt)->pt;
   lj_assertJ(T->root == 0, "not a root trace");
   lj_assertJ(pt != NULL, "trace has no prototype");
-  /* First unpatch any modified bytecode. */
-  trace_unpatch(J, T);
   /* Unlink root trace from chain anchored in prototype. */
   if (pt->trace == T->traceno) {  /* Trace is first in chain. Easy. */
     pt->trace = T->nextroot;
+unpatch:
+    /* Unpatch modified bytecode only if the trace has not been flushed. */
+    trace_unpatch(J, T);
   } else if (pt->trace) {  /* Otherwise search in chain of root traces. */
     GCtrace *T2 = traceref(J, pt->trace);
     if (T2) {
       for (; T2->nextroot; T2 = traceref(J, T2->nextroot))
 	if (T2->nextroot == T->traceno) {
 	  T2->nextroot = T->nextroot;  /* Unlink from chain. */
-	  break;
+	  goto unpatch;
 	}
     }
   }

src/luajit.c

@@ -35,6 +35,21 @@
 
 #if !LJ_TARGET_CONSOLE
 #include <signal.h>
+
+#if LJ_TARGET_POSIX
+/* Improve signal handling on POSIX. Try CTRL-C on: luajit -e 'io.read()' */
+static void signal_set(int sig, void (*h)(int))
+{
+  struct sigaction sa;
+  memset(&sa, 0, sizeof(sa));
+  sa.sa_handler = h;
+  sigemptyset(&sa.sa_mask);
+  sigaction(sig, &sa, NULL);
+}
+#else
+#define signal_set	signal
+#endif
+
 #endif
 
 static lua_State *globalL = NULL;
@@ -54,8 +69,8 @@ static void lstop(lua_State *L, lua_Debug *ar)
 
 static void laction(int i)
 {
-  signal(i, SIG_DFL); /* if another SIGINT happens before lstop,
-			 terminate process (default action) */
+  /* Terminate process if another SIGINT happens (double CTRL-C). */
+  signal_set(i, SIG_DFL);
   lua_sethook(globalL, lstop, LUA_MASKCALL | LUA_MASKRET | LUA_MASKCOUNT, 1);
 }
 #endif
@@ -117,11 +132,11 @@ static int docall(lua_State *L, int narg, int clear)
   lua_pushcfunction(L, traceback);  /* push traceback function */
   lua_insert(L, base);  /* put it under chunk and args */
 #if !LJ_TARGET_CONSOLE
-  signal(SIGINT, laction);
+  signal_set(SIGINT, laction);
 #endif
   status = lua_pcall(L, narg, (clear ? 0 : LUA_MULTRET), base);
 #if !LJ_TARGET_CONSOLE
-  signal(SIGINT, SIG_DFL);
+  signal_set(SIGINT, SIG_DFL);
 #endif
   lua_remove(L, base);  /* remove traceback function */
   /* force a complete garbage collection in case of errors */

src/msvcbuild.bat

@@ -8,7 +8,8 @@
 @rem   nogc64   disable LJ_GC64 mode for x64
 @rem   debug    emit debug symbols
 @rem   amalg    amalgamated build
-@rem   static   static linkage
+@rem   static   create static lib to statically link into your project
+@rem   mixed    create static lib to build a DLL in your project
 
 @if not defined INCLUDE goto :FAIL
 
@@ -106,12 +107,14 @@ buildvm -m folddef -o lj_folddef.h lj_opt_fold.c
 @if "%1"=="static" goto :STATIC
 %LJCOMPILE% %LJDYNBUILD% lj_*.c lib_*.c
 @if errorlevel 1 goto :BAD
+@if "%1"=="mixed" goto :STATICLIB
 %LJLINK% /DLL /OUT:%LJDLLNAME% lj_*.obj lib_*.obj
 @if errorlevel 1 goto :BAD
 @goto :MTDLL
 :STATIC
 %LJCOMPILE% lj_*.c lib_*.c
 @if errorlevel 1 goto :BAD
+:STATICLIB
 %LJLIB% /OUT:%LJLIBNAME% lj_*.obj lib_*.obj
 @if errorlevel 1 goto :BAD
 @goto :MTDLL
@@ -119,13 +122,15 @@ buildvm -m folddef -o lj_folddef.h lj_opt_fold.c
 @if "%2"=="static" goto :AMALGSTATIC
 %LJCOMPILE% %LJDYNBUILD% ljamalg.c
 @if errorlevel 1 goto :BAD
+@if "%2"=="mixed" goto :AMALGSTATICLIB
 %LJLINK% /DLL /OUT:%LJDLLNAME% ljamalg.obj lj_vm.obj
 @if errorlevel 1 goto :BAD
 @goto :MTDLL
 :AMALGSTATIC
 %LJCOMPILE% ljamalg.c
 @if errorlevel 1 goto :BAD
-%LJLINK% /OUT:%LJDLLNAME% ljamalg.obj lj_vm.obj
+:AMALGSTATICLIB
+%LJLIB% /OUT:%LJLIBNAME% ljamalg.obj lj_vm.obj
 @if errorlevel 1 goto :BAD
 :MTDLL
 if exist %LJDLLNAME%.manifest^