Browse Source

Update LuaJIT to LuaJIT/LuaJIT@f73e649.

Miku AuahDark 11 months ago
parent
commit
064aeeb377
91 changed files with 671 additions and 856 deletions
  1. 1 1
      CMakeLists.txt
  2. 1 1
      libs/LuaJIT/.relver
  3. 6 4
      libs/LuaJIT/Makefile
  4. 35 122
      libs/LuaJIT/android/arm64-v8a/jit/bcsave.lua
  5. 20 0
      libs/LuaJIT/android/arm64-v8a/jit/dis_arm64.lua
  6. 6 1
      libs/LuaJIT/android/arm64-v8a/jit/dump.lua
  7. 1 3
      libs/LuaJIT/android/arm64-v8a/jit/p.lua
  8. 7 2
      libs/LuaJIT/android/arm64-v8a/jit/v.lua
  9. 2 2
      libs/LuaJIT/android/arm64-v8a/jit/vmdef.lua
  10. BIN
      libs/LuaJIT/android/arm64-v8a/libluajit.a
  11. BIN
      libs/LuaJIT/android/arm64-v8a/libluajit.so
  12. 2 2
      libs/LuaJIT/android/arm64-v8a/luajit.h
  13. 35 122
      libs/LuaJIT/android/armeabi-v7a/jit/bcsave.lua
  14. 20 0
      libs/LuaJIT/android/armeabi-v7a/jit/dis_arm64.lua
  15. 6 1
      libs/LuaJIT/android/armeabi-v7a/jit/dump.lua
  16. 1 3
      libs/LuaJIT/android/armeabi-v7a/jit/p.lua
  17. 7 2
      libs/LuaJIT/android/armeabi-v7a/jit/v.lua
  18. 2 2
      libs/LuaJIT/android/armeabi-v7a/jit/vmdef.lua
  19. BIN
      libs/LuaJIT/android/armeabi-v7a/libluajit.a
  20. BIN
      libs/LuaJIT/android/armeabi-v7a/libluajit.so
  21. 2 2
      libs/LuaJIT/android/armeabi-v7a/luajit.h
  22. 35 122
      libs/LuaJIT/android/x86/jit/bcsave.lua
  23. 20 0
      libs/LuaJIT/android/x86/jit/dis_arm64.lua
  24. 6 1
      libs/LuaJIT/android/x86/jit/dump.lua
  25. 1 3
      libs/LuaJIT/android/x86/jit/p.lua
  26. 7 2
      libs/LuaJIT/android/x86/jit/v.lua
  27. 2 2
      libs/LuaJIT/android/x86/jit/vmdef.lua
  28. BIN
      libs/LuaJIT/android/x86/libluajit.a
  29. BIN
      libs/LuaJIT/android/x86/libluajit.so
  30. 2 2
      libs/LuaJIT/android/x86/luajit.h
  31. 35 122
      libs/LuaJIT/android/x86_64/jit/bcsave.lua
  32. 20 0
      libs/LuaJIT/android/x86_64/jit/dis_arm64.lua
  33. 6 1
      libs/LuaJIT/android/x86_64/jit/dump.lua
  34. 1 3
      libs/LuaJIT/android/x86_64/jit/p.lua
  35. 7 2
      libs/LuaJIT/android/x86_64/jit/v.lua
  36. 2 2
      libs/LuaJIT/android/x86_64/jit/vmdef.lua
  37. BIN
      libs/LuaJIT/android/x86_64/libluajit.a
  38. BIN
      libs/LuaJIT/android/x86_64/libluajit.so
  39. 2 2
      libs/LuaJIT/android/x86_64/luajit.h
  40. 3 3
      libs/LuaJIT/doc/ext_buffer.html
  41. 13 0
      libs/LuaJIT/doc/ext_ffi_semantics.html
  42. 5 1
      libs/LuaJIT/doc/extensions.html
  43. 5 2
      libs/LuaJIT/src/Makefile
  44. 3 2
      libs/LuaJIT/src/Makefile.dep
  45. 4 0
      libs/LuaJIT/src/host/buildvm_asm.c
  46. 35 122
      libs/LuaJIT/src/jit/bcsave.lua
  47. 20 0
      libs/LuaJIT/src/jit/dis_arm64.lua
  48. 6 1
      libs/LuaJIT/src/jit/dump.lua
  49. 1 3
      libs/LuaJIT/src/jit/p.lua
  50. 7 2
      libs/LuaJIT/src/jit/v.lua
  51. 1 1
      libs/LuaJIT/src/lib_bit.c
  52. 3 19
      libs/LuaJIT/src/lib_ffi.c
  53. 5 1
      libs/LuaJIT/src/lib_math.c
  54. 1 1
      libs/LuaJIT/src/lj_alloc.c
  55. 1 1
      libs/LuaJIT/src/lj_arch.h
  56. 2 2
      libs/LuaJIT/src/lj_asm.c
  57. 1 1
      libs/LuaJIT/src/lj_asm_arm64.h
  58. 1 1
      libs/LuaJIT/src/lj_asm_mips.h
  59. 1 1
      libs/LuaJIT/src/lj_bcread.c
  60. 1 3
      libs/LuaJIT/src/lj_buf.c
  61. 1 3
      libs/LuaJIT/src/lj_carith.c
  62. 1 1
      libs/LuaJIT/src/lj_cdata.c
  63. 27 14
      libs/LuaJIT/src/lj_crecord.c
  64. 12 0
      libs/LuaJIT/src/lj_ctype.c
  65. 1 1
      libs/LuaJIT/src/lj_ctype.h
  66. 25 6
      libs/LuaJIT/src/lj_emit_arm64.h
  67. 0 6
      libs/LuaJIT/src/lj_err.c
  68. 1 0
      libs/LuaJIT/src/lj_errmsg.h
  69. 5 1
      libs/LuaJIT/src/lj_ffrecord.c
  70. 18 33
      libs/LuaJIT/src/lj_gc.c
  71. 2 2
      libs/LuaJIT/src/lj_jit.h
  72. 10 9
      libs/LuaJIT/src/lj_load.c
  73. 4 1
      libs/LuaJIT/src/lj_obj.h
  74. 19 6
      libs/LuaJIT/src/lj_opt_fold.c
  75. 2 1
      libs/LuaJIT/src/lj_opt_narrow.c
  76. 6 4
      libs/LuaJIT/src/lj_parse.c
  77. 50 12
      libs/LuaJIT/src/lj_record.c
  78. 6 3
      libs/LuaJIT/src/lj_snap.c
  79. 3 0
      libs/LuaJIT/src/lj_state.c
  80. 1 8
      libs/LuaJIT/src/lj_tab.c
  81. 0 3
      libs/LuaJIT/src/lj_tab.h
  82. 2 0
      libs/LuaJIT/src/lj_target_arm64.h
  83. 1 1
      libs/LuaJIT/src/lj_traceerr.h
  84. 5 3
      libs/LuaJIT/src/luaconf.h
  85. 22 11
      libs/LuaJIT/src/msvcbuild.bat
  86. 7 6
      libs/LuaJIT/src/nxbuild.bat
  87. 12 1
      libs/LuaJIT/src/vm_arm.dasc
  88. 6 15
      libs/LuaJIT/src/vm_arm64.dasc
  89. 3 2
      libs/LuaJIT/src/vm_mips64.dasc
  90. 0 1
      libs/LuaJIT/src/vm_x64.dasc
  91. 0 1
      libs/LuaJIT/src/vm_x86.dasc

+ 1 - 1
CMakeLists.txt

@@ -221,7 +221,7 @@ endif()
 
 set(MEGA_ZLIB_VER "1.3.1")
 set(MEGA_LUA51_VER "5.1.5")
-set(MEGA_LUAJIT_VER "2.1.1710088188")
+set(MEGA_LUAJIT_VER "2.1.1734355927")
 set(MEGA_LIBOGG_VER "1.3.2")
 set(MEGA_LIBVORBIS_VER "1.3.5")
 set(MEGA_LIBTHEORA_VER "1.1.1")

+ 1 - 1
libs/LuaJIT/.relver

@@ -1 +1 @@
-1710088188
+1734355927

+ 6 - 4
libs/LuaJIT/Makefile

@@ -37,12 +37,13 @@ export MULTILIB= lib
 DPREFIX= $(DESTDIR)$(PREFIX)
 INSTALL_BIN=   $(DPREFIX)/bin
 INSTALL_LIB=   $(DPREFIX)/$(MULTILIB)
-INSTALL_SHARE= $(DPREFIX)/share
+INSTALL_SHARE_= $(PREFIX)/share
+INSTALL_SHARE= $(DESTDIR)$(INSTALL_SHARE_)
 INSTALL_DEFINC= $(DPREFIX)/include/luajit-$(MMVERSION)
 INSTALL_INC=   $(INSTALL_DEFINC)
 
-INSTALL_LJLIBD= $(INSTALL_SHARE)/luajit-$(MMVERSION)
-INSTALL_JITLIB= $(INSTALL_LJLIBD)/jit
+export INSTALL_LJLIBD= $(INSTALL_SHARE_)/luajit-$(MMVERSION)
+INSTALL_JITLIB= $(DESTDIR)$(INSTALL_LJLIBD)/jit
 INSTALL_LMODD= $(INSTALL_SHARE)/lua
 INSTALL_LMOD= $(INSTALL_LMODD)/$(ABIVER)
 INSTALL_CMODD= $(INSTALL_LIB)/lua
@@ -71,7 +72,7 @@ INSTALL_PC= $(INSTALL_PKGCONFIG)/$(INSTALL_PCNAME)
 
 INSTALL_DIRS= $(INSTALL_BIN) $(INSTALL_LIB) $(INSTALL_INC) $(INSTALL_MAN) \
   $(INSTALL_PKGCONFIG) $(INSTALL_JITLIB) $(INSTALL_LMOD) $(INSTALL_CMOD)
-UNINSTALL_DIRS= $(INSTALL_JITLIB) $(INSTALL_LJLIBD) $(INSTALL_INC) \
+UNINSTALL_DIRS= $(INSTALL_JITLIB) $(DESTDIR)$(INSTALL_LJLIBD) $(INSTALL_INC) \
   $(INSTALL_LMOD) $(INSTALL_LMODD) $(INSTALL_CMOD) $(INSTALL_CMODD)
 
 RM= rm -f
@@ -114,6 +115,7 @@ ifeq (Darwin,$(TARGET_SYS))
   INSTALL_SOSHORT1= $(INSTALL_DYLIBSHORT1)
   INSTALL_SOSHORT2= $(INSTALL_DYLIBSHORT2)
   LDCONFIG= :
+  SED_PC+= -e "s| -Wl,-E||"
 endif
 
 ##############################################################################

+ 35 - 122
libs/LuaJIT/android/arm64-v8a/jit/bcsave.lua

@@ -439,24 +439,12 @@ typedef struct
 {
   mach_header; uint32_t reserved;
 } mach_header_64;
-typedef struct {
-  uint32_t cmd, cmdsize;
-  char segname[16];
-  uint32_t vmaddr, vmsize, fileoff, filesize;
-  uint32_t maxprot, initprot, nsects, flags;
-} mach_segment_command;
 typedef struct {
   uint32_t cmd, cmdsize;
   char segname[16];
   uint64_t vmaddr, vmsize, fileoff, filesize;
   uint32_t maxprot, initprot, nsects, flags;
 } mach_segment_command_64;
-typedef struct {
-  char sectname[16], segname[16];
-  uint32_t addr, size;
-  uint32_t offset, align, reloff, nreloc, flags;
-  uint32_t reserved1, reserved2;
-} mach_section;
 typedef struct {
   char sectname[16], segname[16];
   uint64_t addr, size;
@@ -466,139 +454,64 @@ typedef struct {
 typedef struct {
   uint32_t cmd, cmdsize, symoff, nsyms, stroff, strsize;
 } mach_symtab_command;
-typedef struct {
-  int32_t strx;
-  uint8_t type, sect;
-  int16_t desc;
-  uint32_t value;
-} mach_nlist;
 typedef struct {
   int32_t strx;
   uint8_t type, sect;
   uint16_t desc;
   uint64_t value;
 } mach_nlist_64;
-typedef struct
-{
-  int32_t magic, nfat_arch;
-} mach_fat_header;
-typedef struct
-{
-  int32_t cputype, cpusubtype, offset, size, align;
-} mach_fat_arch;
-typedef struct {
-  struct {
-    mach_header hdr;
-    mach_segment_command seg;
-    mach_section sec;
-    mach_symtab_command sym;
-  } arch[1];
-  mach_nlist sym_entry;
-  uint8_t space[4096];
-} mach_obj;
 typedef struct {
-  struct {
-    mach_header_64 hdr;
-    mach_segment_command_64 seg;
-    mach_section_64 sec;
-    mach_symtab_command sym;
-  } arch[1];
+  mach_header_64 hdr;
+  mach_segment_command_64 seg;
+  mach_section_64 sec;
+  mach_symtab_command sym;
   mach_nlist_64 sym_entry;
   uint8_t space[4096];
 } mach_obj_64;
-typedef struct {
-  mach_fat_header fat;
-  mach_fat_arch fat_arch[2];
-  struct {
-    mach_header hdr;
-    mach_segment_command seg;
-    mach_section sec;
-    mach_symtab_command sym;
-  } arch[2];
-  mach_nlist sym_entry;
-  uint8_t space[4096];
-} mach_fat_obj;
-typedef struct {
-  mach_fat_header fat;
-  mach_fat_arch fat_arch[2];
-  struct {
-    mach_header_64 hdr;
-    mach_segment_command_64 seg;
-    mach_section_64 sec;
-    mach_symtab_command sym;
-  } arch[2];
-  mach_nlist_64 sym_entry;
-  uint8_t space[4096];
-} mach_fat_obj_64;
 ]]
   local symname = '_'..LJBC_PREFIX..ctx.modname
-  local isfat, is64, align, mobj = false, false, 4, "mach_obj"
-  if ctx.arch == "x64" then
-    is64, align, mobj = true, 8, "mach_obj_64"
-  elseif ctx.arch == "arm" then
-    isfat, mobj = true, "mach_fat_obj"
-  elseif ctx.arch == "arm64" then
-    is64, align, isfat, mobj = true, 8, true, "mach_fat_obj_64"
-  else
-    check(ctx.arch == "x86", "unsupported architecture for OSX")
+  local cputype, cpusubtype = 0x01000007, 3
+  if ctx.arch ~= "x64" then
+    check(ctx.arch == "arm64", "unsupported architecture for OSX")
+    cputype, cpusubtype = 0x0100000c, 0
   end
   local function aligned(v, a) return bit.band(v+a-1, -a) end
-  local be32 = bit.bswap -- Mach-O FAT is BE, supported archs are LE.
 
   -- Create Mach-O object and fill in header.
-  local o = ffi.new(mobj)
-  local mach_size = aligned(ffi.offsetof(o, "space")+#symname+2, align)
-  local cputype = ({ x86={7}, x64={0x01000007}, arm={7,12}, arm64={0x01000007,0x0100000c} })[ctx.arch]
-  local cpusubtype = ({ x86={3}, x64={3}, arm={3,9}, arm64={3,0} })[ctx.arch]
-  if isfat then
-    o.fat.magic = be32(0xcafebabe)
-    o.fat.nfat_arch = be32(#cpusubtype)
-  end
+  local o = ffi.new("mach_obj_64")
+  local mach_size = aligned(ffi.offsetof(o, "space")+#symname+2, 8)
 
   -- Fill in sections and symbols.
-  for i=0,#cpusubtype-1 do
-    local ofs = 0
-    if isfat then
-      local a = o.fat_arch[i]
-      a.cputype = be32(cputype[i+1])
-      a.cpusubtype = be32(cpusubtype[i+1])
-      -- Subsequent slices overlap each other to share data.
-      ofs = ffi.offsetof(o, "arch") + i*ffi.sizeof(o.arch[0])
-      a.offset = be32(ofs)
-      a.size = be32(mach_size-ofs+#s)
-    end
-    local a = o.arch[i]
-    a.hdr.magic = is64 and 0xfeedfacf or 0xfeedface
-    a.hdr.cputype = cputype[i+1]
-    a.hdr.cpusubtype = cpusubtype[i+1]
-    a.hdr.filetype = 1
-    a.hdr.ncmds = 2
-    a.hdr.sizeofcmds = ffi.sizeof(a.seg)+ffi.sizeof(a.sec)+ffi.sizeof(a.sym)
-    a.seg.cmd = is64 and 0x19 or 0x1
-    a.seg.cmdsize = ffi.sizeof(a.seg)+ffi.sizeof(a.sec)
-    a.seg.vmsize = #s
-    a.seg.fileoff = mach_size-ofs
-    a.seg.filesize = #s
-    a.seg.maxprot = 1
-    a.seg.initprot = 1
-    a.seg.nsects = 1
-    ffi.copy(a.sec.sectname, "__data")
-    ffi.copy(a.sec.segname, "__DATA")
-    a.sec.size = #s
-    a.sec.offset = mach_size-ofs
-    a.sym.cmd = 2
-    a.sym.cmdsize = ffi.sizeof(a.sym)
-    a.sym.symoff = ffi.offsetof(o, "sym_entry")-ofs
-    a.sym.nsyms = 1
-    a.sym.stroff = ffi.offsetof(o, "sym_entry")+ffi.sizeof(o.sym_entry)-ofs
-    a.sym.strsize = aligned(#symname+2, align)
-  end
+  o.hdr.magic = 0xfeedfacf
+  o.hdr.cputype = cputype
+  o.hdr.cpusubtype = cpusubtype
+  o.hdr.filetype = 1
+  o.hdr.ncmds = 2
+  o.hdr.sizeofcmds = ffi.sizeof(o.seg)+ffi.sizeof(o.sec)+ffi.sizeof(o.sym)
+  o.seg.cmd = 0x19
+  o.seg.cmdsize = ffi.sizeof(o.seg)+ffi.sizeof(o.sec)
+  o.seg.vmsize = #s
+  o.seg.fileoff = mach_size
+  o.seg.filesize = #s
+  o.seg.maxprot = 1
+  o.seg.initprot = 1
+  o.seg.nsects = 1
+  ffi.copy(o.sec.sectname, "__data")
+  ffi.copy(o.sec.segname, "__DATA")
+  o.sec.size = #s
+  o.sec.offset = mach_size
+  o.sym.cmd = 2
+  o.sym.cmdsize = ffi.sizeof(o.sym)
+  o.sym.symoff = ffi.offsetof(o, "sym_entry")
+  o.sym.nsyms = 1
+  o.sym.stroff = ffi.offsetof(o, "sym_entry")+ffi.sizeof(o.sym_entry)
+  o.sym.strsize = aligned(#symname+2, 8)
   o.sym_entry.type = 0xf
   o.sym_entry.sect = 1
   o.sym_entry.strx = 1
   ffi.copy(o.space+1, symname)
 
-  -- Write Macho-O object file.
+  -- Write Mach-O object file.
   local fp = savefile(output, "wb")
   fp:write(ffi.string(o, mach_size))
   bcsave_tail(fp, output, s)

+ 20 - 0
libs/LuaJIT/android/arm64-v8a/jit/dis_arm64.lua

@@ -658,6 +658,10 @@ local map_datafp = { -- Data processing, SIMD and FP.
 	}
       }
     }
+  },
+  { -- 010
+    shift = 0, mask = 0x81f8fc00,
+    [0x100e400] = "moviDdG"
   }
 }
 
@@ -832,6 +836,20 @@ local function parse_fpimm8(op)
   return sign * frac * 2^exp
 end
 
+local function decode_fpmovi(op)
+  local lo = rshift(op, 5)
+  local hi = rshift(op, 9)
+  lo = bor(band(lo, 1) * 0xff, band(lo, 2) * 0x7f80, band(lo, 4) * 0x3fc000,
+	   band(lo, 8) * 0x1fe00000)
+  hi = bor(band(hi, 1) * 0xff, band(hi, 0x80) * 0x1fe,
+	   band(hi, 0x100) * 0xff00, band(hi, 0x200) * 0x7f8000)
+  if hi ~= 0 then
+    return fmt_hex32(hi)..tohex(lo)
+  else
+    return fmt_hex32(lo)
+  end
+end
+
 local function prefer_bfx(sf, uns, imms, immr)
   if imms < immr or imms == 31 or imms == 63 then
     return false
@@ -1131,6 +1149,8 @@ local function disass_ins(ctx)
       x = 0
     elseif p == "F" then
       x = parse_fpimm8(op)
+    elseif p == "G" then
+      x = "#0x"..decode_fpmovi(op)
     elseif p == "g" or p == "f" or p == "x" or p == "w" or
 	   p == "d" or p == "s" then
       -- These are handled in D/N/M/A.

+ 6 - 1
libs/LuaJIT/android/arm64-v8a/jit/dump.lua

@@ -552,7 +552,12 @@ local recdepth = 0
 local function fmterr(err, info)
   if type(err) == "number" then
     if type(info) == "function" then info = fmtfunc(info) end
-    err = format(vmdef.traceerr[err], info)
+    local fmt = vmdef.traceerr[err]
+    if fmt == "NYI: bytecode %s" then
+      local oidx = 6 * info
+      info = sub(vmdef.bcnames, oidx+1, oidx+6)
+    end
+    err = format(fmt, info)
   end
   return err
 end

+ 1 - 3
libs/LuaJIT/android/arm64-v8a/jit/p.lua

@@ -227,9 +227,7 @@ local function prof_finish()
     local samples = prof_samples
     if samples == 0 then
       if prof_raw ~= true then out:write("[No samples collected]\n") end
-      return
-    end
-    if prof_ann then
+    elseif prof_ann then
       prof_annotate(prof_count1, samples)
     else
       prof_top(prof_count1, prof_count2, samples, "")

+ 7 - 2
libs/LuaJIT/android/arm64-v8a/jit/v.lua

@@ -62,7 +62,7 @@ local jit = require("jit")
 local jutil = require("jit.util")
 local vmdef = require("jit.vmdef")
 local funcinfo, traceinfo = jutil.funcinfo, jutil.traceinfo
-local type, format = type, string.format
+local type, sub, format = type, string.sub, string.format
 local stdout, stderr = io.stdout, io.stderr
 
 -- Active flag and output file handle.
@@ -89,7 +89,12 @@ end
 local function fmterr(err, info)
   if type(err) == "number" then
     if type(info) == "function" then info = fmtfunc(info) end
-    err = format(vmdef.traceerr[err], info)
+    local fmt = vmdef.traceerr[err]
+    if fmt == "NYI: bytecode %s" then
+      local oidx = 6 * info
+      info = sub(vmdef.bcnames, oidx+1, oidx+6)
+    end
+    err = format(fmt, info)
   end
   return err
 end

+ 2 - 2
libs/LuaJIT/android/arm64-v8a/jit/vmdef.lua

@@ -1,6 +1,6 @@
 -- This is a generated file. DO NOT EDIT!
 
-assert(require("jit").version == "LuaJIT 2.1.1710088188", "LuaJIT core/library version mismatch")
+assert(require("jit").version == "LuaJIT 2.1.1734355927", "LuaJIT core/library version mismatch")
 
 return {
 
@@ -136,7 +136,7 @@ traceerr = {
 "too many snapshots",
 "blacklisted",
 "retry recording",
-"NYI: bytecode %d",
+"NYI: bytecode %s",
 "leaving loop in root trace",
 "inner loop in root trace",
 "loop unroll limit reached",

BIN
libs/LuaJIT/android/arm64-v8a/libluajit.a


BIN
libs/LuaJIT/android/arm64-v8a/libluajit.so


+ 2 - 2
libs/LuaJIT/android/arm64-v8a/luajit.h

@@ -30,9 +30,9 @@
 
 #include "lua.h"
 
-#define LUAJIT_VERSION		"LuaJIT 2.1.1710088188"
+#define LUAJIT_VERSION		"LuaJIT 2.1.1734355927"
 #define LUAJIT_VERSION_NUM	20199  /* Deprecated. */
-#define LUAJIT_VERSION_SYM	luaJIT_version_2_1_1710088188
+#define LUAJIT_VERSION_SYM	luaJIT_version_2_1_1734355927
 #define LUAJIT_COPYRIGHT	"Copyright (C) 2005-2023 Mike Pall"
 #define LUAJIT_URL		"https://luajit.org/"
 

+ 35 - 122
libs/LuaJIT/android/armeabi-v7a/jit/bcsave.lua

@@ -439,24 +439,12 @@ typedef struct
 {
   mach_header; uint32_t reserved;
 } mach_header_64;
-typedef struct {
-  uint32_t cmd, cmdsize;
-  char segname[16];
-  uint32_t vmaddr, vmsize, fileoff, filesize;
-  uint32_t maxprot, initprot, nsects, flags;
-} mach_segment_command;
 typedef struct {
   uint32_t cmd, cmdsize;
   char segname[16];
   uint64_t vmaddr, vmsize, fileoff, filesize;
   uint32_t maxprot, initprot, nsects, flags;
 } mach_segment_command_64;
-typedef struct {
-  char sectname[16], segname[16];
-  uint32_t addr, size;
-  uint32_t offset, align, reloff, nreloc, flags;
-  uint32_t reserved1, reserved2;
-} mach_section;
 typedef struct {
   char sectname[16], segname[16];
   uint64_t addr, size;
@@ -466,139 +454,64 @@ typedef struct {
 typedef struct {
   uint32_t cmd, cmdsize, symoff, nsyms, stroff, strsize;
 } mach_symtab_command;
-typedef struct {
-  int32_t strx;
-  uint8_t type, sect;
-  int16_t desc;
-  uint32_t value;
-} mach_nlist;
 typedef struct {
   int32_t strx;
   uint8_t type, sect;
   uint16_t desc;
   uint64_t value;
 } mach_nlist_64;
-typedef struct
-{
-  int32_t magic, nfat_arch;
-} mach_fat_header;
-typedef struct
-{
-  int32_t cputype, cpusubtype, offset, size, align;
-} mach_fat_arch;
-typedef struct {
-  struct {
-    mach_header hdr;
-    mach_segment_command seg;
-    mach_section sec;
-    mach_symtab_command sym;
-  } arch[1];
-  mach_nlist sym_entry;
-  uint8_t space[4096];
-} mach_obj;
 typedef struct {
-  struct {
-    mach_header_64 hdr;
-    mach_segment_command_64 seg;
-    mach_section_64 sec;
-    mach_symtab_command sym;
-  } arch[1];
+  mach_header_64 hdr;
+  mach_segment_command_64 seg;
+  mach_section_64 sec;
+  mach_symtab_command sym;
   mach_nlist_64 sym_entry;
   uint8_t space[4096];
 } mach_obj_64;
-typedef struct {
-  mach_fat_header fat;
-  mach_fat_arch fat_arch[2];
-  struct {
-    mach_header hdr;
-    mach_segment_command seg;
-    mach_section sec;
-    mach_symtab_command sym;
-  } arch[2];
-  mach_nlist sym_entry;
-  uint8_t space[4096];
-} mach_fat_obj;
-typedef struct {
-  mach_fat_header fat;
-  mach_fat_arch fat_arch[2];
-  struct {
-    mach_header_64 hdr;
-    mach_segment_command_64 seg;
-    mach_section_64 sec;
-    mach_symtab_command sym;
-  } arch[2];
-  mach_nlist_64 sym_entry;
-  uint8_t space[4096];
-} mach_fat_obj_64;
 ]]
   local symname = '_'..LJBC_PREFIX..ctx.modname
-  local isfat, is64, align, mobj = false, false, 4, "mach_obj"
-  if ctx.arch == "x64" then
-    is64, align, mobj = true, 8, "mach_obj_64"
-  elseif ctx.arch == "arm" then
-    isfat, mobj = true, "mach_fat_obj"
-  elseif ctx.arch == "arm64" then
-    is64, align, isfat, mobj = true, 8, true, "mach_fat_obj_64"
-  else
-    check(ctx.arch == "x86", "unsupported architecture for OSX")
+  local cputype, cpusubtype = 0x01000007, 3
+  if ctx.arch ~= "x64" then
+    check(ctx.arch == "arm64", "unsupported architecture for OSX")
+    cputype, cpusubtype = 0x0100000c, 0
   end
   local function aligned(v, a) return bit.band(v+a-1, -a) end
-  local be32 = bit.bswap -- Mach-O FAT is BE, supported archs are LE.
 
   -- Create Mach-O object and fill in header.
-  local o = ffi.new(mobj)
-  local mach_size = aligned(ffi.offsetof(o, "space")+#symname+2, align)
-  local cputype = ({ x86={7}, x64={0x01000007}, arm={7,12}, arm64={0x01000007,0x0100000c} })[ctx.arch]
-  local cpusubtype = ({ x86={3}, x64={3}, arm={3,9}, arm64={3,0} })[ctx.arch]
-  if isfat then
-    o.fat.magic = be32(0xcafebabe)
-    o.fat.nfat_arch = be32(#cpusubtype)
-  end
+  local o = ffi.new("mach_obj_64")
+  local mach_size = aligned(ffi.offsetof(o, "space")+#symname+2, 8)
 
   -- Fill in sections and symbols.
-  for i=0,#cpusubtype-1 do
-    local ofs = 0
-    if isfat then
-      local a = o.fat_arch[i]
-      a.cputype = be32(cputype[i+1])
-      a.cpusubtype = be32(cpusubtype[i+1])
-      -- Subsequent slices overlap each other to share data.
-      ofs = ffi.offsetof(o, "arch") + i*ffi.sizeof(o.arch[0])
-      a.offset = be32(ofs)
-      a.size = be32(mach_size-ofs+#s)
-    end
-    local a = o.arch[i]
-    a.hdr.magic = is64 and 0xfeedfacf or 0xfeedface
-    a.hdr.cputype = cputype[i+1]
-    a.hdr.cpusubtype = cpusubtype[i+1]
-    a.hdr.filetype = 1
-    a.hdr.ncmds = 2
-    a.hdr.sizeofcmds = ffi.sizeof(a.seg)+ffi.sizeof(a.sec)+ffi.sizeof(a.sym)
-    a.seg.cmd = is64 and 0x19 or 0x1
-    a.seg.cmdsize = ffi.sizeof(a.seg)+ffi.sizeof(a.sec)
-    a.seg.vmsize = #s
-    a.seg.fileoff = mach_size-ofs
-    a.seg.filesize = #s
-    a.seg.maxprot = 1
-    a.seg.initprot = 1
-    a.seg.nsects = 1
-    ffi.copy(a.sec.sectname, "__data")
-    ffi.copy(a.sec.segname, "__DATA")
-    a.sec.size = #s
-    a.sec.offset = mach_size-ofs
-    a.sym.cmd = 2
-    a.sym.cmdsize = ffi.sizeof(a.sym)
-    a.sym.symoff = ffi.offsetof(o, "sym_entry")-ofs
-    a.sym.nsyms = 1
-    a.sym.stroff = ffi.offsetof(o, "sym_entry")+ffi.sizeof(o.sym_entry)-ofs
-    a.sym.strsize = aligned(#symname+2, align)
-  end
+  o.hdr.magic = 0xfeedfacf
+  o.hdr.cputype = cputype
+  o.hdr.cpusubtype = cpusubtype
+  o.hdr.filetype = 1
+  o.hdr.ncmds = 2
+  o.hdr.sizeofcmds = ffi.sizeof(o.seg)+ffi.sizeof(o.sec)+ffi.sizeof(o.sym)
+  o.seg.cmd = 0x19
+  o.seg.cmdsize = ffi.sizeof(o.seg)+ffi.sizeof(o.sec)
+  o.seg.vmsize = #s
+  o.seg.fileoff = mach_size
+  o.seg.filesize = #s
+  o.seg.maxprot = 1
+  o.seg.initprot = 1
+  o.seg.nsects = 1
+  ffi.copy(o.sec.sectname, "__data")
+  ffi.copy(o.sec.segname, "__DATA")
+  o.sec.size = #s
+  o.sec.offset = mach_size
+  o.sym.cmd = 2
+  o.sym.cmdsize = ffi.sizeof(o.sym)
+  o.sym.symoff = ffi.offsetof(o, "sym_entry")
+  o.sym.nsyms = 1
+  o.sym.stroff = ffi.offsetof(o, "sym_entry")+ffi.sizeof(o.sym_entry)
+  o.sym.strsize = aligned(#symname+2, 8)
   o.sym_entry.type = 0xf
   o.sym_entry.sect = 1
   o.sym_entry.strx = 1
   ffi.copy(o.space+1, symname)
 
-  -- Write Macho-O object file.
+  -- Write Mach-O object file.
   local fp = savefile(output, "wb")
   fp:write(ffi.string(o, mach_size))
   bcsave_tail(fp, output, s)

+ 20 - 0
libs/LuaJIT/android/armeabi-v7a/jit/dis_arm64.lua

@@ -658,6 +658,10 @@ local map_datafp = { -- Data processing, SIMD and FP.
 	}
       }
     }
+  },
+  { -- 010
+    shift = 0, mask = 0x81f8fc00,
+    [0x100e400] = "moviDdG"
   }
 }
 
@@ -832,6 +836,20 @@ local function parse_fpimm8(op)
   return sign * frac * 2^exp
 end
 
+local function decode_fpmovi(op)
+  local lo = rshift(op, 5)
+  local hi = rshift(op, 9)
+  lo = bor(band(lo, 1) * 0xff, band(lo, 2) * 0x7f80, band(lo, 4) * 0x3fc000,
+	   band(lo, 8) * 0x1fe00000)
+  hi = bor(band(hi, 1) * 0xff, band(hi, 0x80) * 0x1fe,
+	   band(hi, 0x100) * 0xff00, band(hi, 0x200) * 0x7f8000)
+  if hi ~= 0 then
+    return fmt_hex32(hi)..tohex(lo)
+  else
+    return fmt_hex32(lo)
+  end
+end
+
 local function prefer_bfx(sf, uns, imms, immr)
   if imms < immr or imms == 31 or imms == 63 then
     return false
@@ -1131,6 +1149,8 @@ local function disass_ins(ctx)
       x = 0
     elseif p == "F" then
       x = parse_fpimm8(op)
+    elseif p == "G" then
+      x = "#0x"..decode_fpmovi(op)
     elseif p == "g" or p == "f" or p == "x" or p == "w" or
 	   p == "d" or p == "s" then
       -- These are handled in D/N/M/A.

+ 6 - 1
libs/LuaJIT/android/armeabi-v7a/jit/dump.lua

@@ -552,7 +552,12 @@ local recdepth = 0
 local function fmterr(err, info)
   if type(err) == "number" then
     if type(info) == "function" then info = fmtfunc(info) end
-    err = format(vmdef.traceerr[err], info)
+    local fmt = vmdef.traceerr[err]
+    if fmt == "NYI: bytecode %s" then
+      local oidx = 6 * info
+      info = sub(vmdef.bcnames, oidx+1, oidx+6)
+    end
+    err = format(fmt, info)
   end
   return err
 end

+ 1 - 3
libs/LuaJIT/android/armeabi-v7a/jit/p.lua

@@ -227,9 +227,7 @@ local function prof_finish()
     local samples = prof_samples
     if samples == 0 then
       if prof_raw ~= true then out:write("[No samples collected]\n") end
-      return
-    end
-    if prof_ann then
+    elseif prof_ann then
       prof_annotate(prof_count1, samples)
     else
       prof_top(prof_count1, prof_count2, samples, "")

+ 7 - 2
libs/LuaJIT/android/armeabi-v7a/jit/v.lua

@@ -62,7 +62,7 @@ local jit = require("jit")
 local jutil = require("jit.util")
 local vmdef = require("jit.vmdef")
 local funcinfo, traceinfo = jutil.funcinfo, jutil.traceinfo
-local type, format = type, string.format
+local type, sub, format = type, string.sub, string.format
 local stdout, stderr = io.stdout, io.stderr
 
 -- Active flag and output file handle.
@@ -89,7 +89,12 @@ end
 local function fmterr(err, info)
   if type(err) == "number" then
     if type(info) == "function" then info = fmtfunc(info) end
-    err = format(vmdef.traceerr[err], info)
+    local fmt = vmdef.traceerr[err]
+    if fmt == "NYI: bytecode %s" then
+      local oidx = 6 * info
+      info = sub(vmdef.bcnames, oidx+1, oidx+6)
+    end
+    err = format(fmt, info)
   end
   return err
 end

+ 2 - 2
libs/LuaJIT/android/armeabi-v7a/jit/vmdef.lua

@@ -1,6 +1,6 @@
 -- This is a generated file. DO NOT EDIT!
 
-assert(require("jit").version == "LuaJIT 2.1.1710088188", "LuaJIT core/library version mismatch")
+assert(require("jit").version == "LuaJIT 2.1.1734355927", "LuaJIT core/library version mismatch")
 
 return {
 
@@ -136,7 +136,7 @@ traceerr = {
 "too many snapshots",
 "blacklisted",
 "retry recording",
-"NYI: bytecode %d",
+"NYI: bytecode %s",
 "leaving loop in root trace",
 "inner loop in root trace",
 "loop unroll limit reached",

BIN
libs/LuaJIT/android/armeabi-v7a/libluajit.a


BIN
libs/LuaJIT/android/armeabi-v7a/libluajit.so


+ 2 - 2
libs/LuaJIT/android/armeabi-v7a/luajit.h

@@ -30,9 +30,9 @@
 
 #include "lua.h"
 
-#define LUAJIT_VERSION		"LuaJIT 2.1.1710088188"
+#define LUAJIT_VERSION		"LuaJIT 2.1.1734355927"
 #define LUAJIT_VERSION_NUM	20199  /* Deprecated. */
-#define LUAJIT_VERSION_SYM	luaJIT_version_2_1_1710088188
+#define LUAJIT_VERSION_SYM	luaJIT_version_2_1_1734355927
 #define LUAJIT_COPYRIGHT	"Copyright (C) 2005-2023 Mike Pall"
 #define LUAJIT_URL		"https://luajit.org/"
 

+ 35 - 122
libs/LuaJIT/android/x86/jit/bcsave.lua

@@ -439,24 +439,12 @@ typedef struct
 {
   mach_header; uint32_t reserved;
 } mach_header_64;
-typedef struct {
-  uint32_t cmd, cmdsize;
-  char segname[16];
-  uint32_t vmaddr, vmsize, fileoff, filesize;
-  uint32_t maxprot, initprot, nsects, flags;
-} mach_segment_command;
 typedef struct {
   uint32_t cmd, cmdsize;
   char segname[16];
   uint64_t vmaddr, vmsize, fileoff, filesize;
   uint32_t maxprot, initprot, nsects, flags;
 } mach_segment_command_64;
-typedef struct {
-  char sectname[16], segname[16];
-  uint32_t addr, size;
-  uint32_t offset, align, reloff, nreloc, flags;
-  uint32_t reserved1, reserved2;
-} mach_section;
 typedef struct {
   char sectname[16], segname[16];
   uint64_t addr, size;
@@ -466,139 +454,64 @@ typedef struct {
 typedef struct {
   uint32_t cmd, cmdsize, symoff, nsyms, stroff, strsize;
 } mach_symtab_command;
-typedef struct {
-  int32_t strx;
-  uint8_t type, sect;
-  int16_t desc;
-  uint32_t value;
-} mach_nlist;
 typedef struct {
   int32_t strx;
   uint8_t type, sect;
   uint16_t desc;
   uint64_t value;
 } mach_nlist_64;
-typedef struct
-{
-  int32_t magic, nfat_arch;
-} mach_fat_header;
-typedef struct
-{
-  int32_t cputype, cpusubtype, offset, size, align;
-} mach_fat_arch;
-typedef struct {
-  struct {
-    mach_header hdr;
-    mach_segment_command seg;
-    mach_section sec;
-    mach_symtab_command sym;
-  } arch[1];
-  mach_nlist sym_entry;
-  uint8_t space[4096];
-} mach_obj;
 typedef struct {
-  struct {
-    mach_header_64 hdr;
-    mach_segment_command_64 seg;
-    mach_section_64 sec;
-    mach_symtab_command sym;
-  } arch[1];
+  mach_header_64 hdr;
+  mach_segment_command_64 seg;
+  mach_section_64 sec;
+  mach_symtab_command sym;
   mach_nlist_64 sym_entry;
   uint8_t space[4096];
 } mach_obj_64;
-typedef struct {
-  mach_fat_header fat;
-  mach_fat_arch fat_arch[2];
-  struct {
-    mach_header hdr;
-    mach_segment_command seg;
-    mach_section sec;
-    mach_symtab_command sym;
-  } arch[2];
-  mach_nlist sym_entry;
-  uint8_t space[4096];
-} mach_fat_obj;
-typedef struct {
-  mach_fat_header fat;
-  mach_fat_arch fat_arch[2];
-  struct {
-    mach_header_64 hdr;
-    mach_segment_command_64 seg;
-    mach_section_64 sec;
-    mach_symtab_command sym;
-  } arch[2];
-  mach_nlist_64 sym_entry;
-  uint8_t space[4096];
-} mach_fat_obj_64;
 ]]
   local symname = '_'..LJBC_PREFIX..ctx.modname
-  local isfat, is64, align, mobj = false, false, 4, "mach_obj"
-  if ctx.arch == "x64" then
-    is64, align, mobj = true, 8, "mach_obj_64"
-  elseif ctx.arch == "arm" then
-    isfat, mobj = true, "mach_fat_obj"
-  elseif ctx.arch == "arm64" then
-    is64, align, isfat, mobj = true, 8, true, "mach_fat_obj_64"
-  else
-    check(ctx.arch == "x86", "unsupported architecture for OSX")
+  local cputype, cpusubtype = 0x01000007, 3
+  if ctx.arch ~= "x64" then
+    check(ctx.arch == "arm64", "unsupported architecture for OSX")
+    cputype, cpusubtype = 0x0100000c, 0
   end
   local function aligned(v, a) return bit.band(v+a-1, -a) end
-  local be32 = bit.bswap -- Mach-O FAT is BE, supported archs are LE.
 
   -- Create Mach-O object and fill in header.
-  local o = ffi.new(mobj)
-  local mach_size = aligned(ffi.offsetof(o, "space")+#symname+2, align)
-  local cputype = ({ x86={7}, x64={0x01000007}, arm={7,12}, arm64={0x01000007,0x0100000c} })[ctx.arch]
-  local cpusubtype = ({ x86={3}, x64={3}, arm={3,9}, arm64={3,0} })[ctx.arch]
-  if isfat then
-    o.fat.magic = be32(0xcafebabe)
-    o.fat.nfat_arch = be32(#cpusubtype)
-  end
+  local o = ffi.new("mach_obj_64")
+  local mach_size = aligned(ffi.offsetof(o, "space")+#symname+2, 8)
 
   -- Fill in sections and symbols.
-  for i=0,#cpusubtype-1 do
-    local ofs = 0
-    if isfat then
-      local a = o.fat_arch[i]
-      a.cputype = be32(cputype[i+1])
-      a.cpusubtype = be32(cpusubtype[i+1])
-      -- Subsequent slices overlap each other to share data.
-      ofs = ffi.offsetof(o, "arch") + i*ffi.sizeof(o.arch[0])
-      a.offset = be32(ofs)
-      a.size = be32(mach_size-ofs+#s)
-    end
-    local a = o.arch[i]
-    a.hdr.magic = is64 and 0xfeedfacf or 0xfeedface
-    a.hdr.cputype = cputype[i+1]
-    a.hdr.cpusubtype = cpusubtype[i+1]
-    a.hdr.filetype = 1
-    a.hdr.ncmds = 2
-    a.hdr.sizeofcmds = ffi.sizeof(a.seg)+ffi.sizeof(a.sec)+ffi.sizeof(a.sym)
-    a.seg.cmd = is64 and 0x19 or 0x1
-    a.seg.cmdsize = ffi.sizeof(a.seg)+ffi.sizeof(a.sec)
-    a.seg.vmsize = #s
-    a.seg.fileoff = mach_size-ofs
-    a.seg.filesize = #s
-    a.seg.maxprot = 1
-    a.seg.initprot = 1
-    a.seg.nsects = 1
-    ffi.copy(a.sec.sectname, "__data")
-    ffi.copy(a.sec.segname, "__DATA")
-    a.sec.size = #s
-    a.sec.offset = mach_size-ofs
-    a.sym.cmd = 2
-    a.sym.cmdsize = ffi.sizeof(a.sym)
-    a.sym.symoff = ffi.offsetof(o, "sym_entry")-ofs
-    a.sym.nsyms = 1
-    a.sym.stroff = ffi.offsetof(o, "sym_entry")+ffi.sizeof(o.sym_entry)-ofs
-    a.sym.strsize = aligned(#symname+2, align)
-  end
+  o.hdr.magic = 0xfeedfacf
+  o.hdr.cputype = cputype
+  o.hdr.cpusubtype = cpusubtype
+  o.hdr.filetype = 1
+  o.hdr.ncmds = 2
+  o.hdr.sizeofcmds = ffi.sizeof(o.seg)+ffi.sizeof(o.sec)+ffi.sizeof(o.sym)
+  o.seg.cmd = 0x19
+  o.seg.cmdsize = ffi.sizeof(o.seg)+ffi.sizeof(o.sec)
+  o.seg.vmsize = #s
+  o.seg.fileoff = mach_size
+  o.seg.filesize = #s
+  o.seg.maxprot = 1
+  o.seg.initprot = 1
+  o.seg.nsects = 1
+  ffi.copy(o.sec.sectname, "__data")
+  ffi.copy(o.sec.segname, "__DATA")
+  o.sec.size = #s
+  o.sec.offset = mach_size
+  o.sym.cmd = 2
+  o.sym.cmdsize = ffi.sizeof(o.sym)
+  o.sym.symoff = ffi.offsetof(o, "sym_entry")
+  o.sym.nsyms = 1
+  o.sym.stroff = ffi.offsetof(o, "sym_entry")+ffi.sizeof(o.sym_entry)
+  o.sym.strsize = aligned(#symname+2, 8)
   o.sym_entry.type = 0xf
   o.sym_entry.sect = 1
   o.sym_entry.strx = 1
   ffi.copy(o.space+1, symname)
 
-  -- Write Macho-O object file.
+  -- Write Mach-O object file.
   local fp = savefile(output, "wb")
   fp:write(ffi.string(o, mach_size))
   bcsave_tail(fp, output, s)

+ 20 - 0
libs/LuaJIT/android/x86/jit/dis_arm64.lua

@@ -658,6 +658,10 @@ local map_datafp = { -- Data processing, SIMD and FP.
 	}
       }
     }
+  },
+  { -- 010
+    shift = 0, mask = 0x81f8fc00,
+    [0x100e400] = "moviDdG"
   }
 }
 
@@ -832,6 +836,20 @@ local function parse_fpimm8(op)
   return sign * frac * 2^exp
 end
 
+local function decode_fpmovi(op)
+  local lo = rshift(op, 5)
+  local hi = rshift(op, 9)
+  lo = bor(band(lo, 1) * 0xff, band(lo, 2) * 0x7f80, band(lo, 4) * 0x3fc000,
+	   band(lo, 8) * 0x1fe00000)
+  hi = bor(band(hi, 1) * 0xff, band(hi, 0x80) * 0x1fe,
+	   band(hi, 0x100) * 0xff00, band(hi, 0x200) * 0x7f8000)
+  if hi ~= 0 then
+    return fmt_hex32(hi)..tohex(lo)
+  else
+    return fmt_hex32(lo)
+  end
+end
+
 local function prefer_bfx(sf, uns, imms, immr)
   if imms < immr or imms == 31 or imms == 63 then
     return false
@@ -1131,6 +1149,8 @@ local function disass_ins(ctx)
       x = 0
     elseif p == "F" then
       x = parse_fpimm8(op)
+    elseif p == "G" then
+      x = "#0x"..decode_fpmovi(op)
     elseif p == "g" or p == "f" or p == "x" or p == "w" or
 	   p == "d" or p == "s" then
       -- These are handled in D/N/M/A.

+ 6 - 1
libs/LuaJIT/android/x86/jit/dump.lua

@@ -552,7 +552,12 @@ local recdepth = 0
 local function fmterr(err, info)
   if type(err) == "number" then
     if type(info) == "function" then info = fmtfunc(info) end
-    err = format(vmdef.traceerr[err], info)
+    local fmt = vmdef.traceerr[err]
+    if fmt == "NYI: bytecode %s" then
+      local oidx = 6 * info
+      info = sub(vmdef.bcnames, oidx+1, oidx+6)
+    end
+    err = format(fmt, info)
   end
   return err
 end

+ 1 - 3
libs/LuaJIT/android/x86/jit/p.lua

@@ -227,9 +227,7 @@ local function prof_finish()
     local samples = prof_samples
     if samples == 0 then
       if prof_raw ~= true then out:write("[No samples collected]\n") end
-      return
-    end
-    if prof_ann then
+    elseif prof_ann then
       prof_annotate(prof_count1, samples)
     else
       prof_top(prof_count1, prof_count2, samples, "")

+ 7 - 2
libs/LuaJIT/android/x86/jit/v.lua

@@ -62,7 +62,7 @@ local jit = require("jit")
 local jutil = require("jit.util")
 local vmdef = require("jit.vmdef")
 local funcinfo, traceinfo = jutil.funcinfo, jutil.traceinfo
-local type, format = type, string.format
+local type, sub, format = type, string.sub, string.format
 local stdout, stderr = io.stdout, io.stderr
 
 -- Active flag and output file handle.
@@ -89,7 +89,12 @@ end
 local function fmterr(err, info)
   if type(err) == "number" then
     if type(info) == "function" then info = fmtfunc(info) end
-    err = format(vmdef.traceerr[err], info)
+    local fmt = vmdef.traceerr[err]
+    if fmt == "NYI: bytecode %s" then
+      local oidx = 6 * info
+      info = sub(vmdef.bcnames, oidx+1, oidx+6)
+    end
+    err = format(fmt, info)
   end
   return err
 end

+ 2 - 2
libs/LuaJIT/android/x86/jit/vmdef.lua

@@ -1,6 +1,6 @@
 -- This is a generated file. DO NOT EDIT!
 
-assert(require("jit").version == "LuaJIT 2.1.1710088188", "LuaJIT core/library version mismatch")
+assert(require("jit").version == "LuaJIT 2.1.1734355927", "LuaJIT core/library version mismatch")
 
 return {
 
@@ -136,7 +136,7 @@ traceerr = {
 "too many snapshots",
 "blacklisted",
 "retry recording",
-"NYI: bytecode %d",
+"NYI: bytecode %s",
 "leaving loop in root trace",
 "inner loop in root trace",
 "loop unroll limit reached",

BIN
libs/LuaJIT/android/x86/libluajit.a


BIN
libs/LuaJIT/android/x86/libluajit.so


+ 2 - 2
libs/LuaJIT/android/x86/luajit.h

@@ -30,9 +30,9 @@
 
 #include "lua.h"
 
-#define LUAJIT_VERSION		"LuaJIT 2.1.1710088188"
+#define LUAJIT_VERSION		"LuaJIT 2.1.1734355927"
 #define LUAJIT_VERSION_NUM	20199  /* Deprecated. */
-#define LUAJIT_VERSION_SYM	luaJIT_version_2_1_1710088188
+#define LUAJIT_VERSION_SYM	luaJIT_version_2_1_1734355927
 #define LUAJIT_COPYRIGHT	"Copyright (C) 2005-2023 Mike Pall"
 #define LUAJIT_URL		"https://luajit.org/"
 

+ 35 - 122
libs/LuaJIT/android/x86_64/jit/bcsave.lua

@@ -439,24 +439,12 @@ typedef struct
 {
   mach_header; uint32_t reserved;
 } mach_header_64;
-typedef struct {
-  uint32_t cmd, cmdsize;
-  char segname[16];
-  uint32_t vmaddr, vmsize, fileoff, filesize;
-  uint32_t maxprot, initprot, nsects, flags;
-} mach_segment_command;
 typedef struct {
   uint32_t cmd, cmdsize;
   char segname[16];
   uint64_t vmaddr, vmsize, fileoff, filesize;
   uint32_t maxprot, initprot, nsects, flags;
 } mach_segment_command_64;
-typedef struct {
-  char sectname[16], segname[16];
-  uint32_t addr, size;
-  uint32_t offset, align, reloff, nreloc, flags;
-  uint32_t reserved1, reserved2;
-} mach_section;
 typedef struct {
   char sectname[16], segname[16];
   uint64_t addr, size;
@@ -466,139 +454,64 @@ typedef struct {
 typedef struct {
   uint32_t cmd, cmdsize, symoff, nsyms, stroff, strsize;
 } mach_symtab_command;
-typedef struct {
-  int32_t strx;
-  uint8_t type, sect;
-  int16_t desc;
-  uint32_t value;
-} mach_nlist;
 typedef struct {
   int32_t strx;
   uint8_t type, sect;
   uint16_t desc;
   uint64_t value;
 } mach_nlist_64;
-typedef struct
-{
-  int32_t magic, nfat_arch;
-} mach_fat_header;
-typedef struct
-{
-  int32_t cputype, cpusubtype, offset, size, align;
-} mach_fat_arch;
-typedef struct {
-  struct {
-    mach_header hdr;
-    mach_segment_command seg;
-    mach_section sec;
-    mach_symtab_command sym;
-  } arch[1];
-  mach_nlist sym_entry;
-  uint8_t space[4096];
-} mach_obj;
 typedef struct {
-  struct {
-    mach_header_64 hdr;
-    mach_segment_command_64 seg;
-    mach_section_64 sec;
-    mach_symtab_command sym;
-  } arch[1];
+  mach_header_64 hdr;
+  mach_segment_command_64 seg;
+  mach_section_64 sec;
+  mach_symtab_command sym;
   mach_nlist_64 sym_entry;
   uint8_t space[4096];
 } mach_obj_64;
-typedef struct {
-  mach_fat_header fat;
-  mach_fat_arch fat_arch[2];
-  struct {
-    mach_header hdr;
-    mach_segment_command seg;
-    mach_section sec;
-    mach_symtab_command sym;
-  } arch[2];
-  mach_nlist sym_entry;
-  uint8_t space[4096];
-} mach_fat_obj;
-typedef struct {
-  mach_fat_header fat;
-  mach_fat_arch fat_arch[2];
-  struct {
-    mach_header_64 hdr;
-    mach_segment_command_64 seg;
-    mach_section_64 sec;
-    mach_symtab_command sym;
-  } arch[2];
-  mach_nlist_64 sym_entry;
-  uint8_t space[4096];
-} mach_fat_obj_64;
 ]]
   local symname = '_'..LJBC_PREFIX..ctx.modname
-  local isfat, is64, align, mobj = false, false, 4, "mach_obj"
-  if ctx.arch == "x64" then
-    is64, align, mobj = true, 8, "mach_obj_64"
-  elseif ctx.arch == "arm" then
-    isfat, mobj = true, "mach_fat_obj"
-  elseif ctx.arch == "arm64" then
-    is64, align, isfat, mobj = true, 8, true, "mach_fat_obj_64"
-  else
-    check(ctx.arch == "x86", "unsupported architecture for OSX")
+  local cputype, cpusubtype = 0x01000007, 3
+  if ctx.arch ~= "x64" then
+    check(ctx.arch == "arm64", "unsupported architecture for OSX")
+    cputype, cpusubtype = 0x0100000c, 0
   end
   local function aligned(v, a) return bit.band(v+a-1, -a) end
-  local be32 = bit.bswap -- Mach-O FAT is BE, supported archs are LE.
 
   -- Create Mach-O object and fill in header.
-  local o = ffi.new(mobj)
-  local mach_size = aligned(ffi.offsetof(o, "space")+#symname+2, align)
-  local cputype = ({ x86={7}, x64={0x01000007}, arm={7,12}, arm64={0x01000007,0x0100000c} })[ctx.arch]
-  local cpusubtype = ({ x86={3}, x64={3}, arm={3,9}, arm64={3,0} })[ctx.arch]
-  if isfat then
-    o.fat.magic = be32(0xcafebabe)
-    o.fat.nfat_arch = be32(#cpusubtype)
-  end
+  local o = ffi.new("mach_obj_64")
+  local mach_size = aligned(ffi.offsetof(o, "space")+#symname+2, 8)
 
   -- Fill in sections and symbols.
-  for i=0,#cpusubtype-1 do
-    local ofs = 0
-    if isfat then
-      local a = o.fat_arch[i]
-      a.cputype = be32(cputype[i+1])
-      a.cpusubtype = be32(cpusubtype[i+1])
-      -- Subsequent slices overlap each other to share data.
-      ofs = ffi.offsetof(o, "arch") + i*ffi.sizeof(o.arch[0])
-      a.offset = be32(ofs)
-      a.size = be32(mach_size-ofs+#s)
-    end
-    local a = o.arch[i]
-    a.hdr.magic = is64 and 0xfeedfacf or 0xfeedface
-    a.hdr.cputype = cputype[i+1]
-    a.hdr.cpusubtype = cpusubtype[i+1]
-    a.hdr.filetype = 1
-    a.hdr.ncmds = 2
-    a.hdr.sizeofcmds = ffi.sizeof(a.seg)+ffi.sizeof(a.sec)+ffi.sizeof(a.sym)
-    a.seg.cmd = is64 and 0x19 or 0x1
-    a.seg.cmdsize = ffi.sizeof(a.seg)+ffi.sizeof(a.sec)
-    a.seg.vmsize = #s
-    a.seg.fileoff = mach_size-ofs
-    a.seg.filesize = #s
-    a.seg.maxprot = 1
-    a.seg.initprot = 1
-    a.seg.nsects = 1
-    ffi.copy(a.sec.sectname, "__data")
-    ffi.copy(a.sec.segname, "__DATA")
-    a.sec.size = #s
-    a.sec.offset = mach_size-ofs
-    a.sym.cmd = 2
-    a.sym.cmdsize = ffi.sizeof(a.sym)
-    a.sym.symoff = ffi.offsetof(o, "sym_entry")-ofs
-    a.sym.nsyms = 1
-    a.sym.stroff = ffi.offsetof(o, "sym_entry")+ffi.sizeof(o.sym_entry)-ofs
-    a.sym.strsize = aligned(#symname+2, align)
-  end
+  o.hdr.magic = 0xfeedfacf
+  o.hdr.cputype = cputype
+  o.hdr.cpusubtype = cpusubtype
+  o.hdr.filetype = 1
+  o.hdr.ncmds = 2
+  o.hdr.sizeofcmds = ffi.sizeof(o.seg)+ffi.sizeof(o.sec)+ffi.sizeof(o.sym)
+  o.seg.cmd = 0x19
+  o.seg.cmdsize = ffi.sizeof(o.seg)+ffi.sizeof(o.sec)
+  o.seg.vmsize = #s
+  o.seg.fileoff = mach_size
+  o.seg.filesize = #s
+  o.seg.maxprot = 1
+  o.seg.initprot = 1
+  o.seg.nsects = 1
+  ffi.copy(o.sec.sectname, "__data")
+  ffi.copy(o.sec.segname, "__DATA")
+  o.sec.size = #s
+  o.sec.offset = mach_size
+  o.sym.cmd = 2
+  o.sym.cmdsize = ffi.sizeof(o.sym)
+  o.sym.symoff = ffi.offsetof(o, "sym_entry")
+  o.sym.nsyms = 1
+  o.sym.stroff = ffi.offsetof(o, "sym_entry")+ffi.sizeof(o.sym_entry)
+  o.sym.strsize = aligned(#symname+2, 8)
   o.sym_entry.type = 0xf
   o.sym_entry.sect = 1
   o.sym_entry.strx = 1
   ffi.copy(o.space+1, symname)
 
-  -- Write Macho-O object file.
+  -- Write Mach-O object file.
   local fp = savefile(output, "wb")
   fp:write(ffi.string(o, mach_size))
   bcsave_tail(fp, output, s)

+ 20 - 0
libs/LuaJIT/android/x86_64/jit/dis_arm64.lua

@@ -658,6 +658,10 @@ local map_datafp = { -- Data processing, SIMD and FP.
 	}
       }
     }
+  },
+  { -- 010
+    shift = 0, mask = 0x81f8fc00,
+    [0x100e400] = "moviDdG"
   }
 }
 
@@ -832,6 +836,20 @@ local function parse_fpimm8(op)
   return sign * frac * 2^exp
 end
 
+local function decode_fpmovi(op)
+  local lo = rshift(op, 5)
+  local hi = rshift(op, 9)
+  lo = bor(band(lo, 1) * 0xff, band(lo, 2) * 0x7f80, band(lo, 4) * 0x3fc000,
+	   band(lo, 8) * 0x1fe00000)
+  hi = bor(band(hi, 1) * 0xff, band(hi, 0x80) * 0x1fe,
+	   band(hi, 0x100) * 0xff00, band(hi, 0x200) * 0x7f8000)
+  if hi ~= 0 then
+    return fmt_hex32(hi)..tohex(lo)
+  else
+    return fmt_hex32(lo)
+  end
+end
+
 local function prefer_bfx(sf, uns, imms, immr)
   if imms < immr or imms == 31 or imms == 63 then
     return false
@@ -1131,6 +1149,8 @@ local function disass_ins(ctx)
       x = 0
     elseif p == "F" then
       x = parse_fpimm8(op)
+    elseif p == "G" then
+      x = "#0x"..decode_fpmovi(op)
     elseif p == "g" or p == "f" or p == "x" or p == "w" or
 	   p == "d" or p == "s" then
       -- These are handled in D/N/M/A.

+ 6 - 1
libs/LuaJIT/android/x86_64/jit/dump.lua

@@ -552,7 +552,12 @@ local recdepth = 0
 local function fmterr(err, info)
   if type(err) == "number" then
     if type(info) == "function" then info = fmtfunc(info) end
-    err = format(vmdef.traceerr[err], info)
+    local fmt = vmdef.traceerr[err]
+    if fmt == "NYI: bytecode %s" then
+      local oidx = 6 * info
+      info = sub(vmdef.bcnames, oidx+1, oidx+6)
+    end
+    err = format(fmt, info)
   end
   return err
 end

+ 1 - 3
libs/LuaJIT/android/x86_64/jit/p.lua

@@ -227,9 +227,7 @@ local function prof_finish()
     local samples = prof_samples
     if samples == 0 then
       if prof_raw ~= true then out:write("[No samples collected]\n") end
-      return
-    end
-    if prof_ann then
+    elseif prof_ann then
       prof_annotate(prof_count1, samples)
     else
       prof_top(prof_count1, prof_count2, samples, "")

+ 7 - 2
libs/LuaJIT/android/x86_64/jit/v.lua

@@ -62,7 +62,7 @@ local jit = require("jit")
 local jutil = require("jit.util")
 local vmdef = require("jit.vmdef")
 local funcinfo, traceinfo = jutil.funcinfo, jutil.traceinfo
-local type, format = type, string.format
+local type, sub, format = type, string.sub, string.format
 local stdout, stderr = io.stdout, io.stderr
 
 -- Active flag and output file handle.
@@ -89,7 +89,12 @@ end
 local function fmterr(err, info)
   if type(err) == "number" then
     if type(info) == "function" then info = fmtfunc(info) end
-    err = format(vmdef.traceerr[err], info)
+    local fmt = vmdef.traceerr[err]
+    if fmt == "NYI: bytecode %s" then
+      local oidx = 6 * info
+      info = sub(vmdef.bcnames, oidx+1, oidx+6)
+    end
+    err = format(fmt, info)
   end
   return err
 end

+ 2 - 2
libs/LuaJIT/android/x86_64/jit/vmdef.lua

@@ -1,6 +1,6 @@
 -- This is a generated file. DO NOT EDIT!
 
-assert(require("jit").version == "LuaJIT 2.1.1710088188", "LuaJIT core/library version mismatch")
+assert(require("jit").version == "LuaJIT 2.1.1734355927", "LuaJIT core/library version mismatch")
 
 return {
 
@@ -136,7 +136,7 @@ traceerr = {
 "too many snapshots",
 "blacklisted",
 "retry recording",
-"NYI: bytecode %d",
+"NYI: bytecode %s",
 "leaving loop in root trace",
 "inner loop in root trace",
 "loop unroll limit reached",

BIN
libs/LuaJIT/android/x86_64/libluajit.a


BIN
libs/LuaJIT/android/x86_64/libluajit.so


+ 2 - 2
libs/LuaJIT/android/x86_64/luajit.h

@@ -30,9 +30,9 @@
 
 #include "lua.h"
 
-#define LUAJIT_VERSION		"LuaJIT 2.1.1710088188"
+#define LUAJIT_VERSION		"LuaJIT 2.1.1734355927"
 #define LUAJIT_VERSION_NUM	20199  /* Deprecated. */
-#define LUAJIT_VERSION_SYM	luaJIT_version_2_1_1710088188
+#define LUAJIT_VERSION_SYM	luaJIT_version_2_1_1734355927
 #define LUAJIT_COPYRIGHT	"Copyright (C) 2005-2023 Mike Pall"
 #define LUAJIT_URL		"https://luajit.org/"
 

+ 3 - 3
libs/LuaJIT/doc/ext_buffer.html

@@ -85,7 +85,7 @@ operations.
 </p>
 <p>
 The string buffer library also includes a high-performance
-<a href="serialize">serializer</a> for Lua objects.
+<a href="#serialize">serializer</a> for Lua objects.
 </p>
 
 <h2 id="use">Using the String Buffer Library</h2>
@@ -588,9 +588,9 @@ num       → 0x07 double.L
 tab       → 0x08                                   // Empty table
           | 0x09 h.U h*{object object}          // Key/value hash
           | 0x0a a.U a*object                    // 0-based array
-          | 0x0b a.U a*object h.U h*{object object}      // Mixed
+          | 0x0b a.U h.U a*object h*{object object}      // Mixed
           | 0x0c a.U (a-1)*object                // 1-based array
-          | 0x0d a.U (a-1)*object h.U h*{object object}  // Mixed
+          | 0x0d a.U h.U (a-1)*object h*{object object}  // Mixed
 tab_mt    → 0x0e (index-1).U tab          // Metatable dict entry
 
 int64     → 0x10 int.L                             // FFI int64_t

+ 13 - 0
libs/LuaJIT/doc/ext_ffi_semantics.html

@@ -440,6 +440,19 @@ If you don't do this, the default Lua number &rarr; <tt>double</tt>
 conversion rule applies. A vararg C&nbsp;function expecting an integer
 will see a garbled or uninitialized value.
 </p>
+<p>
+Note: this is the only place where creating a boxed scalar number type is
+actually useful. <b>Never use <tt>ffi.new("int")</tt>, <tt>ffi.new("float")</tt>
+etc. anywhere else!</b>
+</p>
+<p style="font-size: 8pt;">
+Ditto for <tt>ffi.cast()</tt>. Explicitly boxing scalars <b>does not</b>
+improve performance or force <tt>int</tt> or <tt>float</tt> arithmetic! It
+just adds costly boxing, unboxing and conversions steps. And it may lead
+to surprise results, because
+<a href="#cdata_arith">cdata arithmetic on scalar numbers</a>
+is always performed on 64 bit integers.
+</p>
 
 <h2 id="init">Initializers</h2>
 <p>

+ 5 - 1
libs/LuaJIT/doc/extensions.html

@@ -265,7 +265,7 @@ and let the GC do its work.
 LuaJIT uses a Tausworthe PRNG with period 2^223 to implement
 <tt>math.random()</tt> and <tt>math.randomseed()</tt>. The quality of
 the PRNG results is much superior compared to the standard Lua
-implementation, which uses the platform-specific ANSI rand().
+implementation, which uses the platform-specific ANSI <tt>rand()</tt>.
 </p>
 <p>
 The PRNG generates the same sequences from the same seeds on all
@@ -276,6 +276,10 @@ It's correctly scaled up and rounded for <tt>math.random(n&nbsp;[,m])</tt> to
 preserve uniformity.
 </p>
 <p>
+Call <tt>math.randomseed()</tt> without any arguments to seed it from
+system entropy.
+</p>
+<p>
 Important: Neither this nor any other PRNG based on the simplistic
 <tt>math.random()</tt> API is suitable for cryptographic use.
 </p>

+ 5 - 2
libs/LuaJIT/src/Makefile

@@ -299,6 +299,9 @@ endif
 ifneq (,$(LMULTILIB))
   TARGET_XCFLAGS+= -DLUA_LMULTILIB=\"$(LMULTILIB)\"
 endif
+ifneq (,$(INSTALL_LJLIBD))
+  TARGET_XCFLAGS+= -DLUA_LJDIR=\"$(INSTALL_LJLIBD)\"
+endif
 
 ##############################################################################
 # Target system detection.
@@ -320,13 +323,13 @@ ifeq (Darwin,$(TARGET_SYS))
   endif
   TARGET_STRIP+= -x
   TARGET_XCFLAGS+= -DLUAJIT_UNWIND_EXTERNAL
-  TARGET_XSHLDFLAGS= -dynamiclib -single_module -undefined dynamic_lookup -fPIC
+  TARGET_XSHLDFLAGS= -dynamiclib -undefined dynamic_lookup -fPIC
   TARGET_DYNXLDOPTS=
   TARGET_XSHLDFLAGS+= -install_name $(TARGET_DYLIBPATH) -compatibility_version $(MAJVER).$(MINVER) -current_version $(MAJVER).$(MINVER).255
 else
 ifeq (iOS,$(TARGET_SYS))
   TARGET_STRIP+= -x
-  TARGET_XSHLDFLAGS= -dynamiclib -single_module -undefined dynamic_lookup -fPIC
+  TARGET_XSHLDFLAGS= -dynamiclib -undefined dynamic_lookup -fPIC
   TARGET_DYNXLDOPTS=
   TARGET_XSHLDFLAGS+= -install_name $(TARGET_DYLIBPATH) -compatibility_version $(MAJVER).$(MINVER) -current_version $(MAJVER).$(MINVER).255
   ifeq (arm64,$(TARGET_LJARCH))

+ 3 - 2
libs/LuaJIT/src/Makefile.dep

@@ -32,7 +32,8 @@ lib_jit.o: lib_jit.c lua.h luaconf.h lauxlib.h lualib.h lj_obj.h lj_def.h \
  lj_target.h lj_target_*.h lj_trace.h lj_dispatch.h lj_traceerr.h \
  lj_vm.h lj_vmevent.h lj_lib.h luajit.h lj_libdef.h
 lib_math.o: lib_math.c lua.h luaconf.h lauxlib.h lualib.h lj_obj.h \
- lj_def.h lj_arch.h lj_lib.h lj_vm.h lj_prng.h lj_libdef.h
+ lj_def.h lj_arch.h lj_err.h lj_errmsg.h lj_lib.h lj_vm.h lj_prng.h \
+ lj_libdef.h
 lib_os.o: lib_os.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_lib.h \
  lj_libdef.h
@@ -97,7 +98,7 @@ lj_crecord.o: lj_crecord.c lj_obj.h lua.h luaconf.h lj_def.h lj_arch.h \
  lj_cdata.h lj_cparse.h lj_cconv.h lj_carith.h lj_clib.h lj_ccall.h \
  lj_ff.h lj_ffdef.h lj_ir.h lj_jit.h lj_ircall.h lj_iropt.h lj_trace.h \
  lj_dispatch.h lj_traceerr.h lj_record.h lj_ffrecord.h lj_snap.h \
- lj_crecord.h lj_strfmt.h
+ lj_crecord.h lj_strfmt.h lj_strscan.h
 lj_ctype.o: lj_ctype.c lj_obj.h lua.h luaconf.h lj_def.h lj_arch.h \
  lj_gc.h lj_err.h lj_errmsg.h lj_str.h lj_tab.h lj_strfmt.h lj_ctype.h \
  lj_ccallback.h lj_buf.h

+ 4 - 0
libs/LuaJIT/src/host/buildvm_asm.c

@@ -339,6 +339,10 @@ void emit_asm(BuildCtx *ctx)
     fprintf(ctx->fp, "\t.ident \"%s\"\n", ctx->dasm_ident);
     break;
   case BUILD_machasm:
+#if defined(__apple_build_version__) && __apple_build_version__ >= 15000000 && __apple_build_version__ < 15000300
+    /* Workaround for XCode 15.0 - 15.2. */
+    fprintf(ctx->fp, "\t.subsections_via_symbols\n");
+#endif
     fprintf(ctx->fp,
       "\t.cstring\n"
       "\t.ascii \"%s\\0\"\n", ctx->dasm_ident);

+ 35 - 122
libs/LuaJIT/src/jit/bcsave.lua

@@ -439,24 +439,12 @@ typedef struct
 {
   mach_header; uint32_t reserved;
 } mach_header_64;
-typedef struct {
-  uint32_t cmd, cmdsize;
-  char segname[16];
-  uint32_t vmaddr, vmsize, fileoff, filesize;
-  uint32_t maxprot, initprot, nsects, flags;
-} mach_segment_command;
 typedef struct {
   uint32_t cmd, cmdsize;
   char segname[16];
   uint64_t vmaddr, vmsize, fileoff, filesize;
   uint32_t maxprot, initprot, nsects, flags;
 } mach_segment_command_64;
-typedef struct {
-  char sectname[16], segname[16];
-  uint32_t addr, size;
-  uint32_t offset, align, reloff, nreloc, flags;
-  uint32_t reserved1, reserved2;
-} mach_section;
 typedef struct {
   char sectname[16], segname[16];
   uint64_t addr, size;
@@ -466,139 +454,64 @@ typedef struct {
 typedef struct {
   uint32_t cmd, cmdsize, symoff, nsyms, stroff, strsize;
 } mach_symtab_command;
-typedef struct {
-  int32_t strx;
-  uint8_t type, sect;
-  int16_t desc;
-  uint32_t value;
-} mach_nlist;
 typedef struct {
   int32_t strx;
   uint8_t type, sect;
   uint16_t desc;
   uint64_t value;
 } mach_nlist_64;
-typedef struct
-{
-  int32_t magic, nfat_arch;
-} mach_fat_header;
-typedef struct
-{
-  int32_t cputype, cpusubtype, offset, size, align;
-} mach_fat_arch;
-typedef struct {
-  struct {
-    mach_header hdr;
-    mach_segment_command seg;
-    mach_section sec;
-    mach_symtab_command sym;
-  } arch[1];
-  mach_nlist sym_entry;
-  uint8_t space[4096];
-} mach_obj;
 typedef struct {
-  struct {
-    mach_header_64 hdr;
-    mach_segment_command_64 seg;
-    mach_section_64 sec;
-    mach_symtab_command sym;
-  } arch[1];
+  mach_header_64 hdr;
+  mach_segment_command_64 seg;
+  mach_section_64 sec;
+  mach_symtab_command sym;
   mach_nlist_64 sym_entry;
   uint8_t space[4096];
 } mach_obj_64;
-typedef struct {
-  mach_fat_header fat;
-  mach_fat_arch fat_arch[2];
-  struct {
-    mach_header hdr;
-    mach_segment_command seg;
-    mach_section sec;
-    mach_symtab_command sym;
-  } arch[2];
-  mach_nlist sym_entry;
-  uint8_t space[4096];
-} mach_fat_obj;
-typedef struct {
-  mach_fat_header fat;
-  mach_fat_arch fat_arch[2];
-  struct {
-    mach_header_64 hdr;
-    mach_segment_command_64 seg;
-    mach_section_64 sec;
-    mach_symtab_command sym;
-  } arch[2];
-  mach_nlist_64 sym_entry;
-  uint8_t space[4096];
-} mach_fat_obj_64;
 ]]
   local symname = '_'..LJBC_PREFIX..ctx.modname
-  local isfat, is64, align, mobj = false, false, 4, "mach_obj"
-  if ctx.arch == "x64" then
-    is64, align, mobj = true, 8, "mach_obj_64"
-  elseif ctx.arch == "arm" then
-    isfat, mobj = true, "mach_fat_obj"
-  elseif ctx.arch == "arm64" then
-    is64, align, isfat, mobj = true, 8, true, "mach_fat_obj_64"
-  else
-    check(ctx.arch == "x86", "unsupported architecture for OSX")
+  local cputype, cpusubtype = 0x01000007, 3
+  if ctx.arch ~= "x64" then
+    check(ctx.arch == "arm64", "unsupported architecture for OSX")
+    cputype, cpusubtype = 0x0100000c, 0
   end
   local function aligned(v, a) return bit.band(v+a-1, -a) end
-  local be32 = bit.bswap -- Mach-O FAT is BE, supported archs are LE.
 
   -- Create Mach-O object and fill in header.
-  local o = ffi.new(mobj)
-  local mach_size = aligned(ffi.offsetof(o, "space")+#symname+2, align)
-  local cputype = ({ x86={7}, x64={0x01000007}, arm={7,12}, arm64={0x01000007,0x0100000c} })[ctx.arch]
-  local cpusubtype = ({ x86={3}, x64={3}, arm={3,9}, arm64={3,0} })[ctx.arch]
-  if isfat then
-    o.fat.magic = be32(0xcafebabe)
-    o.fat.nfat_arch = be32(#cpusubtype)
-  end
+  local o = ffi.new("mach_obj_64")
+  local mach_size = aligned(ffi.offsetof(o, "space")+#symname+2, 8)
 
   -- Fill in sections and symbols.
-  for i=0,#cpusubtype-1 do
-    local ofs = 0
-    if isfat then
-      local a = o.fat_arch[i]
-      a.cputype = be32(cputype[i+1])
-      a.cpusubtype = be32(cpusubtype[i+1])
-      -- Subsequent slices overlap each other to share data.
-      ofs = ffi.offsetof(o, "arch") + i*ffi.sizeof(o.arch[0])
-      a.offset = be32(ofs)
-      a.size = be32(mach_size-ofs+#s)
-    end
-    local a = o.arch[i]
-    a.hdr.magic = is64 and 0xfeedfacf or 0xfeedface
-    a.hdr.cputype = cputype[i+1]
-    a.hdr.cpusubtype = cpusubtype[i+1]
-    a.hdr.filetype = 1
-    a.hdr.ncmds = 2
-    a.hdr.sizeofcmds = ffi.sizeof(a.seg)+ffi.sizeof(a.sec)+ffi.sizeof(a.sym)
-    a.seg.cmd = is64 and 0x19 or 0x1
-    a.seg.cmdsize = ffi.sizeof(a.seg)+ffi.sizeof(a.sec)
-    a.seg.vmsize = #s
-    a.seg.fileoff = mach_size-ofs
-    a.seg.filesize = #s
-    a.seg.maxprot = 1
-    a.seg.initprot = 1
-    a.seg.nsects = 1
-    ffi.copy(a.sec.sectname, "__data")
-    ffi.copy(a.sec.segname, "__DATA")
-    a.sec.size = #s
-    a.sec.offset = mach_size-ofs
-    a.sym.cmd = 2
-    a.sym.cmdsize = ffi.sizeof(a.sym)
-    a.sym.symoff = ffi.offsetof(o, "sym_entry")-ofs
-    a.sym.nsyms = 1
-    a.sym.stroff = ffi.offsetof(o, "sym_entry")+ffi.sizeof(o.sym_entry)-ofs
-    a.sym.strsize = aligned(#symname+2, align)
-  end
+  o.hdr.magic = 0xfeedfacf
+  o.hdr.cputype = cputype
+  o.hdr.cpusubtype = cpusubtype
+  o.hdr.filetype = 1
+  o.hdr.ncmds = 2
+  o.hdr.sizeofcmds = ffi.sizeof(o.seg)+ffi.sizeof(o.sec)+ffi.sizeof(o.sym)
+  o.seg.cmd = 0x19
+  o.seg.cmdsize = ffi.sizeof(o.seg)+ffi.sizeof(o.sec)
+  o.seg.vmsize = #s
+  o.seg.fileoff = mach_size
+  o.seg.filesize = #s
+  o.seg.maxprot = 1
+  o.seg.initprot = 1
+  o.seg.nsects = 1
+  ffi.copy(o.sec.sectname, "__data")
+  ffi.copy(o.sec.segname, "__DATA")
+  o.sec.size = #s
+  o.sec.offset = mach_size
+  o.sym.cmd = 2
+  o.sym.cmdsize = ffi.sizeof(o.sym)
+  o.sym.symoff = ffi.offsetof(o, "sym_entry")
+  o.sym.nsyms = 1
+  o.sym.stroff = ffi.offsetof(o, "sym_entry")+ffi.sizeof(o.sym_entry)
+  o.sym.strsize = aligned(#symname+2, 8)
   o.sym_entry.type = 0xf
   o.sym_entry.sect = 1
   o.sym_entry.strx = 1
   ffi.copy(o.space+1, symname)
 
-  -- Write Macho-O object file.
+  -- Write Mach-O object file.
   local fp = savefile(output, "wb")
   fp:write(ffi.string(o, mach_size))
   bcsave_tail(fp, output, s)

+ 20 - 0
libs/LuaJIT/src/jit/dis_arm64.lua

@@ -658,6 +658,10 @@ local map_datafp = { -- Data processing, SIMD and FP.
 	}
       }
     }
+  },
+  { -- 010
+    shift = 0, mask = 0x81f8fc00,
+    [0x100e400] = "moviDdG"
   }
 }
 
@@ -832,6 +836,20 @@ local function parse_fpimm8(op)
   return sign * frac * 2^exp
 end
 
+local function decode_fpmovi(op)
+  local lo = rshift(op, 5)
+  local hi = rshift(op, 9)
+  lo = bor(band(lo, 1) * 0xff, band(lo, 2) * 0x7f80, band(lo, 4) * 0x3fc000,
+	   band(lo, 8) * 0x1fe00000)
+  hi = bor(band(hi, 1) * 0xff, band(hi, 0x80) * 0x1fe,
+	   band(hi, 0x100) * 0xff00, band(hi, 0x200) * 0x7f8000)
+  if hi ~= 0 then
+    return fmt_hex32(hi)..tohex(lo)
+  else
+    return fmt_hex32(lo)
+  end
+end
+
 local function prefer_bfx(sf, uns, imms, immr)
   if imms < immr or imms == 31 or imms == 63 then
     return false
@@ -1131,6 +1149,8 @@ local function disass_ins(ctx)
       x = 0
     elseif p == "F" then
       x = parse_fpimm8(op)
+    elseif p == "G" then
+      x = "#0x"..decode_fpmovi(op)
     elseif p == "g" or p == "f" or p == "x" or p == "w" or
 	   p == "d" or p == "s" then
       -- These are handled in D/N/M/A.

+ 6 - 1
libs/LuaJIT/src/jit/dump.lua

@@ -552,7 +552,12 @@ local recdepth = 0
 local function fmterr(err, info)
   if type(err) == "number" then
     if type(info) == "function" then info = fmtfunc(info) end
-    err = format(vmdef.traceerr[err], info)
+    local fmt = vmdef.traceerr[err]
+    if fmt == "NYI: bytecode %s" then
+      local oidx = 6 * info
+      info = sub(vmdef.bcnames, oidx+1, oidx+6)
+    end
+    err = format(fmt, info)
   end
   return err
 end

+ 1 - 3
libs/LuaJIT/src/jit/p.lua

@@ -227,9 +227,7 @@ local function prof_finish()
     local samples = prof_samples
     if samples == 0 then
       if prof_raw ~= true then out:write("[No samples collected]\n") end
-      return
-    end
-    if prof_ann then
+    elseif prof_ann then
       prof_annotate(prof_count1, samples)
     else
       prof_top(prof_count1, prof_count2, samples, "")

+ 7 - 2
libs/LuaJIT/src/jit/v.lua

@@ -62,7 +62,7 @@ local jit = require("jit")
 local jutil = require("jit.util")
 local vmdef = require("jit.vmdef")
 local funcinfo, traceinfo = jutil.funcinfo, jutil.traceinfo
-local type, format = type, string.format
+local type, sub, format = type, string.sub, string.format
 local stdout, stderr = io.stdout, io.stderr
 
 -- Active flag and output file handle.
@@ -89,7 +89,12 @@ end
 local function fmterr(err, info)
   if type(err) == "number" then
     if type(info) == "function" then info = fmtfunc(info) end
-    err = format(vmdef.traceerr[err], info)
+    local fmt = vmdef.traceerr[err]
+    if fmt == "NYI: bytecode %s" then
+      local oidx = 6 * info
+      info = sub(vmdef.bcnames, oidx+1, oidx+6)
+    end
+    err = format(fmt, info)
   end
   return err
 end

+ 1 - 1
libs/LuaJIT/src/lib_bit.c

@@ -98,7 +98,7 @@ LJLIB_ASM(bit_lshift)		LJLIB_REC(bit_shift IR_BSHL)
     x = lj_carith_shift64(x, sh, curr_func(L)->c.ffid - (int)FF_bit_lshift);
     return bit_result64(L, id, x);
   }
-  if (id2) setintV(L->base+1, sh);
+  setintV(L->base+1, sh);
   return FFH_RETRY;
 #else
   lj_lib_checknumber(L, 1);

+ 3 - 19
libs/LuaJIT/src/lib_ffi.c

@@ -305,7 +305,7 @@ LJLIB_CF(ffi_meta___tostring)
       p = *(void **)p;
     } else if (ctype_isenum(ct->info)) {
       msg = "cdata<%s>: %d";
-      p = (void *)(uintptr_t)*(uint32_t **)p;
+      p = (void *)(uintptr_t)*(uint32_t *)p;
     } else {
       if (ctype_isptr(ct->info)) {
 	p = cdata_getptr(p, ct->size);
@@ -513,7 +513,7 @@ LJLIB_CF(ffi_new)	LJLIB_REC(.)
     /* Handle ctype __gc metamethod. Use the fast lookup here. */
     cTValue *tv = lj_tab_getinth(cts->miscmap, -(int32_t)id);
     if (tv && tvistab(tv) && (tv = lj_meta_fast(L, tabV(tv), MM_gc))) {
-      GCtab *t = cts->finalizer;
+      GCtab *t = tabref(G(L)->gcroot[GCROOT_FFI_FIN]);
       if (gcref(t->metatable)) {
 	/* Add to finalizer table, if still enabled. */
 	copyTV(L, lj_tab_set(L, t, o-1), tv);
@@ -765,7 +765,7 @@ LJLIB_CF(ffi_abi)	LJLIB_REC(.)
   return 1;
 }
 
-LJLIB_PUSH(top-8) LJLIB_SET(!)  /* Store reference to miscmap table. */
+LJLIB_PUSH(top-7) LJLIB_SET(!)  /* Store reference to miscmap table. */
 
 LJLIB_CF(ffi_metatype)
 {
@@ -791,8 +791,6 @@ LJLIB_CF(ffi_metatype)
   return 1;
 }
 
-LJLIB_PUSH(top-7) LJLIB_SET(!)  /* Store reference to finalizer table. */
-
 LJLIB_CF(ffi_gc)	LJLIB_REC(.)
 {
   GCcdata *cd = ffi_checkcdata(L, 1);
@@ -825,19 +823,6 @@ LJLIB_PUSH(top-2) LJLIB_SET(arch)
 
 /* ------------------------------------------------------------------------ */
 
-/* Create special weak-keyed finalizer table. */
-static GCtab *ffi_finalizer(lua_State *L)
-{
-  /* NOBARRIER: The table is new (marked white). */
-  GCtab *t = lj_tab_new(L, 0, 1);
-  settabV(L, L->top++, t);
-  setgcref(t->metatable, obj2gco(t));
-  setstrV(L, lj_tab_setstr(L, t, lj_str_newlit(L, "__mode")),
-	  lj_str_newlit(L, "k"));
-  t->nomm = (uint8_t)(~(1u<<MM_mode));
-  return t;
-}
-
 /* Register FFI module as loaded. */
 static void ffi_register_module(lua_State *L)
 {
@@ -853,7 +838,6 @@ LUALIB_API int luaopen_ffi(lua_State *L)
 {
   CTState *cts = lj_ctype_init(L);
   settabV(L, L->top++, (cts->miscmap = lj_tab_new(L, 0, 1)));
-  cts->finalizer = ffi_finalizer(L);
   LJ_LIB_REG(L, NULL, ffi_meta);
   /* NOBARRIER: basemt is a GC root. */
   setgcref(basemt_it(G(L), LJ_TCDATA), obj2gco(tabV(L->top-1)));

+ 5 - 1
libs/LuaJIT/src/lib_math.c

@@ -13,6 +13,7 @@
 #include "lualib.h"
 
 #include "lj_obj.h"
+#include "lj_err.h"
 #include "lj_lib.h"
 #include "lj_vm.h"
 #include "lj_prng.h"
@@ -183,7 +184,10 @@ LJLIB_PUSH(top-2)  /* Upvalue holds userdata with PRNGState. */
 LJLIB_CF(math_randomseed)
 {
   PRNGState *rs = (PRNGState *)(uddata(udataV(lj_lib_upvalue(L, 1))));
-  random_seed(rs, lj_lib_checknum(L, 1));
+  if (L->base != L->top)
+    random_seed(rs, lj_lib_checknum(L, 1));
+  else if (!lj_prng_seed_secure(rs))
+    lj_err_caller(L, LJ_ERR_PRNGSD);
   return 0;
 }
 

+ 1 - 1
libs/LuaJIT/src/lj_alloc.c

@@ -1057,7 +1057,7 @@ static size_t release_unused_segments(mstate m)
       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) {
+      if (!cinuse(p) && (char *)p + psize == (char *)mem2chunk(sp)) {
 	tchunkptr tp = (tchunkptr)p;
 	if (p == m->dv) {
 	  m->dv = 0;

+ 1 - 1
libs/LuaJIT/src/lj_arch.h

@@ -124,7 +124,7 @@
 #define LJ_TARGET_POSIX		(LUAJIT_OS > LUAJIT_OS_WINDOWS)
 #define LJ_TARGET_DLOPEN	LJ_TARGET_POSIX
 
-#if TARGET_OS_IPHONE
+#if defined(TARGET_OS_IPHONE) && TARGET_OS_IPHONE
 #define LJ_TARGET_IOS		1
 #else
 #define LJ_TARGET_IOS		0

+ 2 - 2
libs/LuaJIT/src/lj_asm.c

@@ -949,11 +949,11 @@ static int asm_sunk_store(ASMState *as, IRIns *ira, IRIns *irs)
 static void asm_snap_alloc1(ASMState *as, IRRef ref)
 {
   IRIns *ir = IR(ref);
-  if (!irref_isk(ref) && ir->r != RID_SUNK) {
+  if (!irref_isk(ref)) {
     bloomset(as->snapfilt1, ref);
     bloomset(as->snapfilt2, hashrot(ref, ref + HASH_BIAS));
     if (ra_used(ir)) return;
-    if (ir->r == RID_SINK) {
+    if (ir->r == RID_SINK || ir->r == RID_SUNK) {
       ir->r = RID_SUNK;
 #if LJ_HASFFI
       if (ir->o == IR_CNEWI) {  /* Allocate CNEWI value. */

+ 1 - 1
libs/LuaJIT/src/lj_asm_arm64.h

@@ -787,7 +787,7 @@ static void asm_href(ASMState *as, IRIns *ir, IROp merge)
   int destused = ra_used(ir);
   Reg dest = ra_dest(as, ir, allow);
   Reg tab = ra_alloc1(as, ir->op1, rset_clear(allow, dest));
-  Reg tmp = RID_TMP, type = RID_NONE, key, tkey;
+  Reg tmp = RID_TMP, type = RID_NONE, key = RID_NONE, tkey;
   IRRef refkey = ir->op2;
   IRIns *irkey = IR(refkey);
   int isk = irref_isk(refkey);

+ 1 - 1
libs/LuaJIT/src/lj_asm_mips.h

@@ -456,7 +456,7 @@ static void asm_retf(ASMState *as, IRIns *ir)
   emit_addptr(as, base, -8*delta);
   asm_guard(as, MIPSI_BNE, RID_TMP,
 	    ra_allock(as, igcptr(pc), rset_exclude(RSET_GPR, base)));
-  emit_tsi(as, MIPSI_AL, RID_TMP, base, -8);
+  emit_tsi(as, MIPSI_AL, RID_TMP, base, (LJ_BE || LJ_FR2) ? -8 : -4);
 }
 
 /* -- Buffer operations --------------------------------------------------- */

+ 1 - 1
libs/LuaJIT/src/lj_bcread.c

@@ -408,7 +408,7 @@ static int bcread_header(LexState *ls)
 #endif
   }
   if ((flags & BCDUMP_F_STRIP)) {
-    ls->chunkname = lj_str_newz(ls->L, ls->chunkarg);
+    ls->chunkname = lj_str_newz(ls->L, *ls->chunkarg == BCDUMP_HEAD1 ? "=?" : ls->chunkarg);
   } else {
     MSize len = bcread_uleb128(ls);
     bcread_need(ls, len);

+ 1 - 3
libs/LuaJIT/src/lj_buf.c

@@ -92,10 +92,8 @@ void LJ_FASTCALL lj_buf_shrink(lua_State *L, SBuf *sb)
   char *b = sb->b;
   MSize osz = (MSize)(sb->e - b);
   if (osz > 2*LJ_MIN_SBUF) {
-    MSize n = (MSize)(sb->w - b);
     b = lj_mem_realloc(L, b, osz, (osz >> 1));
-    sb->b = b;
-    sb->w = b + n;
+    sb->w = sb->b = b;  /* Not supposed to keep data across shrinks. */
     sb->e = b + (osz >> 1);
   }
   lj_assertG_(G(sbufL(sb)), !sbufisext(sb), "YAGNI shrink SBufExt");

+ 1 - 3
libs/LuaJIT/src/lj_carith.c

@@ -349,9 +349,7 @@ uint64_t lj_carith_check64(lua_State *L, int narg, CTypeID *id)
   if (LJ_LIKELY(tvisint(o))) {
     return (uint32_t)intV(o);
   } else {
-    int32_t i = lj_num2bit(numV(o));
-    if (LJ_DUALNUM) setintV(o, i);
-    return (uint32_t)i;
+    return (uint32_t)lj_num2bit(numV(o));
   }
 }
 

+ 1 - 1
libs/LuaJIT/src/lj_cdata.c

@@ -86,7 +86,7 @@ void LJ_FASTCALL lj_cdata_free(global_State *g, GCcdata *cd)
 
 void lj_cdata_setfin(lua_State *L, GCcdata *cd, GCobj *obj, uint32_t it)
 {
-  GCtab *t = ctype_ctsG(G(L))->finalizer;
+  GCtab *t = tabref(G(L)->gcroot[GCROOT_FFI_FIN]);
   if (gcref(t->metatable)) {
     /* Add cdata to finalizer table, if still enabled. */
     TValue *tv, tmp;

+ 27 - 14
libs/LuaJIT/src/lj_crecord.c

@@ -32,6 +32,7 @@
 #include "lj_crecord.h"
 #include "lj_dispatch.h"
 #include "lj_strfmt.h"
+#include "lj_strscan.h"
 
 /* Some local macros to save typing. Undef'd at the end. */
 #define IR(ref)			(&J->cur.ir[(ref)])
@@ -1498,7 +1499,8 @@ static TRef crec_arith_meta(jit_State *J, TRef *sp, CType **s, CTState *cts,
 
 void LJ_FASTCALL recff_cdata_arith(jit_State *J, RecordFFData *rd)
 {
-  CTState *cts = ctype_ctsG(J2G(J));
+  CTState *cts = ctype_cts(J->L);
+  MMS mm = (MMS)rd->data;
   TRef sp[2];
   CType *s[2];
   MSize i;
@@ -1548,6 +1550,8 @@ void LJ_FASTCALL recff_cdata_arith(jit_State *J, RecordFFData *rd)
 	}
       }
     } else if (tref_isnil(tr)) {
+      if (!(mm == MM_len || mm == MM_eq || mm == MM_lt || mm == MM_le))
+	lj_trace_err(J, LJ_TRERR_BADTYPE);
       tr = lj_ir_kptr(J, NULL);
       ct = ctype_get(cts, CTID_P_VOID);
     } else if (tref_isinteger(tr)) {
@@ -1566,12 +1570,12 @@ void LJ_FASTCALL recff_cdata_arith(jit_State *J, RecordFFData *rd)
 	  ct = ctype_child(cts, cct);
 	  tr = lj_ir_kint(J, (int32_t)ofs);
 	} else {  /* Interpreter will throw or return false. */
-	  ct = ctype_get(cts, CTID_P_VOID);
+	  lj_trace_err(J, LJ_TRERR_BADTYPE);
 	}
       } else if (ctype_isptr(ct->info)) {
 	tr = emitir(IRT(IR_ADD, IRT_PTR), tr, lj_ir_kintp(J, sizeof(GCstr)));
       } else {
-	ct = ctype_get(cts, CTID_P_VOID);
+	lj_trace_err(J, LJ_TRERR_BADTYPE);
       }
     } else if (!tref_isnum(tr)) {
       tr = 0;
@@ -1583,7 +1587,6 @@ void LJ_FASTCALL recff_cdata_arith(jit_State *J, RecordFFData *rd)
   }
   {
     TRef tr;
-    MMS mm = (MMS)rd->data;
     if ((mm == MM_len || mm == MM_concat ||
 	 (!(tr = crec_arith_int64(J, sp, s, mm)) &&
 	  !(tr = crec_arith_ptr(J, sp, s, mm)))) &&
@@ -1805,11 +1808,21 @@ static CTypeID crec_bit64_type(CTState *cts, cTValue *tv)
   return 0;  /* Use regular 32 bit ops. */
 }
 
+static TRef crec_bit64_arg(jit_State *J, CType *d, TRef sp, TValue *sval)
+{
+  if (LJ_UNLIKELY(tref_isstr(sp))) {
+    if (lj_strscan_num(strV(sval), sval)) {
+      sp = emitir(IRTG(IR_STRTO, IRT_NUM), sp, 0);
+    }  /* else: interpreter will throw. */
+  }
+  return crec_ct_tv(J, d, 0, sp, sval);
+}
+
 void LJ_FASTCALL recff_bit64_tobit(jit_State *J, RecordFFData *rd)
 {
   CTState *cts = ctype_ctsG(J2G(J));
-  TRef tr = crec_ct_tv(J, ctype_get(cts, CTID_INT64), 0,
-		       J->base[0], &rd->argv[0]);
+  TRef tr = crec_bit64_arg(J, ctype_get(cts, CTID_INT64),
+			   J->base[0], &rd->argv[0]);
   if (!tref_isinteger(tr))
     tr = emitconv(tr, IRT_INT, tref_type(tr), 0);
   J->base[0] = tr;
@@ -1820,7 +1833,7 @@ int LJ_FASTCALL recff_bit64_unary(jit_State *J, RecordFFData *rd)
   CTState *cts = ctype_ctsG(J2G(J));
   CTypeID id = crec_bit64_type(cts, &rd->argv[0]);
   if (id) {
-    TRef tr = crec_ct_tv(J, ctype_get(cts, id), 0, J->base[0], &rd->argv[0]);
+    TRef tr = crec_bit64_arg(J, ctype_get(cts, id), J->base[0], &rd->argv[0]);
     tr = emitir(IRT(rd->data, id-CTID_INT64+IRT_I64), tr, 0);
     J->base[0] = emitir(IRTG(IR_CNEWI, IRT_CDATA), lj_ir_kint(J, id), tr);
     return 1;
@@ -1840,9 +1853,9 @@ int LJ_FASTCALL recff_bit64_nary(jit_State *J, RecordFFData *rd)
   if (id) {
     CType *ct = ctype_get(cts, id);
     uint32_t ot = IRT(rd->data, id-CTID_INT64+IRT_I64);
-    TRef tr = crec_ct_tv(J, ct, 0, J->base[0], &rd->argv[0]);
+    TRef tr = crec_bit64_arg(J, ct, J->base[0], &rd->argv[0]);
     for (i = 1; J->base[i] != 0; i++) {
-      TRef tr2 = crec_ct_tv(J, ct, 0, J->base[i], &rd->argv[i]);
+      TRef tr2 = crec_bit64_arg(J, ct, J->base[i], &rd->argv[i]);
       tr = emitir(ot, tr, tr2);
     }
     J->base[0] = emitir(IRTG(IR_CNEWI, IRT_CDATA), lj_ir_kint(J, id), tr);
@@ -1857,15 +1870,15 @@ int LJ_FASTCALL recff_bit64_shift(jit_State *J, RecordFFData *rd)
   CTypeID id;
   TRef tsh = 0;
   if (J->base[0] && tref_iscdata(J->base[1])) {
-    tsh = crec_ct_tv(J, ctype_get(cts, CTID_INT64), 0,
-		     J->base[1], &rd->argv[1]);
+    tsh = crec_bit64_arg(J, ctype_get(cts, CTID_INT64),
+			 J->base[1], &rd->argv[1]);
     if (!tref_isinteger(tsh))
       tsh = emitconv(tsh, IRT_INT, tref_type(tsh), 0);
     J->base[1] = tsh;
   }
   id = crec_bit64_type(cts, &rd->argv[0]);
   if (id) {
-    TRef tr = crec_ct_tv(J, ctype_get(cts, id), 0, J->base[0], &rd->argv[0]);
+    TRef tr = crec_bit64_arg(J, ctype_get(cts, id), J->base[0], &rd->argv[0]);
     uint32_t op = rd->data;
     if (!tsh) tsh = lj_opt_narrow_tobit(J, J->base[1]);
     if (!(op < IR_BROL ? LJ_TARGET_MASKSHIFT : LJ_TARGET_MASKROT) &&
@@ -1895,7 +1908,7 @@ TRef recff_bit64_tohex(jit_State *J, RecordFFData *rd, TRef hdr)
     CTypeID id2 = 0;
     n = (int32_t)lj_carith_check64(J->L, 2, &id2);
     if (id2)
-      trsf = crec_ct_tv(J, ctype_get(cts, CTID_INT32), 0, trsf, &rd->argv[1]);
+      trsf = crec_bit64_arg(J, ctype_get(cts, CTID_INT32), trsf, &rd->argv[1]);
     else
       trsf = lj_opt_narrow_tobit(J, trsf);
     emitir(IRTGI(IR_EQ), trsf, lj_ir_kint(J, n));  /* Specialize to n. */
@@ -1906,7 +1919,7 @@ TRef recff_bit64_tohex(jit_State *J, RecordFFData *rd, TRef hdr)
   if ((uint32_t)n > 254) n = 254;
   sf |= ((SFormat)((n+1)&255) << STRFMT_SH_PREC);
   if (id) {
-    tr = crec_ct_tv(J, ctype_get(cts, id), 0, J->base[0], &rd->argv[0]);
+    tr = crec_bit64_arg(J, ctype_get(cts, id), J->base[0], &rd->argv[0]);
     if (n < 16)
       tr = emitir(IRT(IR_BAND, IRT_U64), tr,
 		  lj_ir_kint64(J, ((uint64_t)1 << 4*n)-1));

+ 12 - 0
libs/LuaJIT/src/lj_ctype.c

@@ -643,6 +643,18 @@ CTState *lj_ctype_init(lua_State *L)
   return cts;
 }
 
+/* Create special weak-keyed finalizer table. */
+void lj_ctype_initfin(lua_State *L)
+{
+  /* NOBARRIER: The table is new (marked white). */
+  GCtab *t = lj_tab_new(L, 0, 1);
+  setgcref(t->metatable, obj2gco(t));
+  setstrV(L, lj_tab_setstr(L, t, lj_str_newlit(L, "__mode")),
+	  lj_str_newlit(L, "k"));
+  t->nomm = (uint8_t)(~(1u<<MM_mode));
+  setgcref(G(L)->gcroot[GCROOT_FFI_FIN], obj2gco(t));
+}
+
 /* Free C type table and state. */
 void lj_ctype_freestate(global_State *g)
 {

+ 1 - 1
libs/LuaJIT/src/lj_ctype.h

@@ -177,7 +177,6 @@ typedef struct CTState {
   MSize sizetab;	/* Size of C type table. */
   lua_State *L;		/* Lua state (needed for errors and allocations). */
   global_State *g;	/* Global state. */
-  GCtab *finalizer;	/* Map of cdata to finalizer. */
   GCtab *miscmap;	/* Map of -CTypeID to metatable and cb slot to func. */
   CCallback cb;		/* Temporary callback state. */
   CTypeID1 hash[CTHASH_SIZE];  /* Hash anchors for C type table. */
@@ -476,6 +475,7 @@ LJ_FUNC GCstr *lj_ctype_repr(lua_State *L, CTypeID id, GCstr *name);
 LJ_FUNC GCstr *lj_ctype_repr_int64(lua_State *L, uint64_t n, int isunsigned);
 LJ_FUNC GCstr *lj_ctype_repr_complex(lua_State *L, void *sp, CTSize size);
 LJ_FUNC CTState *lj_ctype_init(lua_State *L);
+LJ_FUNC void lj_ctype_initfin(lua_State *L);
 LJ_FUNC void lj_ctype_freestate(global_State *g);
 
 #endif

+ 25 - 6
libs/LuaJIT/src/lj_emit_arm64.h

@@ -66,6 +66,17 @@ static uint32_t emit_isfpk64(uint64_t n)
   return ~0u;
 }
 
+static uint32_t emit_isfpmovi(uint64_t n)
+{
+  /* Is every byte either 0x00 or 0xff? */
+  if ((n & U64x(01010101,01010101)) * 0xff != n) return 0;
+  /* Form 8-bit value by taking one bit from each byte. */
+  n &= U64x(80402010,08040201);
+  n = (n * U64x(01010101,01010101)) >> 56;
+  /* Split into the format expected by movi. */
+  return ((n & 0xe0) << 6) | 0x700 | (n & 0x1f);
+}
+
 /* -- Emit basic instructions --------------------------------------------- */
 
 static void emit_dnma(ASMState *as, A64Ins ai, Reg rd, Reg rn, Reg rm, Reg ra)
@@ -98,6 +109,11 @@ static void emit_d(ASMState *as, A64Ins ai, Reg rd)
   *--as->mcp = ai | A64F_D(rd);
 }
 
+static void emit_dl(ASMState *as, A64Ins ai, Reg rd, uint32_t l)
+{
+  *--as->mcp = ai | A64F_D(rd) | A64F_S19(l >> 2);
+}
+
 static void emit_n(ASMState *as, A64Ins ai, Reg rn)
 {
   *--as->mcp = ai | A64F_N(rn);
@@ -215,7 +231,7 @@ static int emit_kadrp(ASMState *as, Reg rd, uint64_t k)
       emit_dn(as, (A64I_ADDx^A64I_K12)|A64F_U12(k - kpage), rd, rd);
     ai = A64I_ADRP;
   }
-  emit_d(as, ai|(((uint32_t)ofs&3)<<29)|A64F_S19(ofs>>2), rd);
+  emit_dl(as, ai|(((uint32_t)ofs&3)<<29), rd, ofs);
   return 1;
 }
 
@@ -280,7 +296,7 @@ static void emit_lsptr(ASMState *as, A64Ins ai, Reg r, void *p)
     /* GL + offset, might subsequently fuse to LDP/STP. */
   } else if (ai == A64I_LDRx && checkmcpofs(as, p)) {
     /* IP + offset is cheaper than allock, but address must be in range. */
-    emit_d(as, A64I_LDRLx | A64F_S19(mcpofs(as, p)>>2), r);
+    emit_dl(as, A64I_LDRLx, r, mcpofs(as, p));
     return;
   } else {  /* Split up into base reg + offset. */
     int64_t i64 = i64ptr(p);
@@ -300,21 +316,24 @@ static void emit_loadk64(ASMState *as, Reg r, IRIns *ir)
     if (fpk != ~0u) {
       emit_d(as, A64I_FMOV_DI | A64F_FP8(fpk), (r & 31));
       return;
+    } else if ((fpk = emit_isfpmovi(*k))) {
+      emit_d(as, A64I_MOVI_DI | (fpk << 5), (r & 31));
+      return;
     }
   }
   ofs = glofs(as, k);
   if (emit_checkofs(A64I_LDRx, ofs)) {
     emit_lso(as, r >= RID_MAX_GPR ? A64I_LDRd : A64I_LDRx,
 	     (r & 31), RID_GL, ofs);
+  } else if (checkmcpofs(as, k)) {
+    emit_dl(as, r >= RID_MAX_GPR ? A64I_LDRLd : A64I_LDRLx,
+	    (r & 31), mcpofs(as, k));
   } else {
     if (r >= RID_MAX_GPR) {
       emit_dn(as, A64I_FMOV_D_R, (r & 31), RID_TMP);
       r = RID_TMP;
     }
-    if (checkmcpofs(as, k))
-      emit_d(as, A64I_LDRLx | A64F_S19(mcpofs(as, k)>>2), r);
-    else
-      emit_loadu64(as, r, *k);
+    emit_loadu64(as, r, *k);
   }
 }
 

+ 0 - 6
libs/LuaJIT/src/lj_err.c

@@ -239,12 +239,6 @@ typedef struct UndocumentedDispatcherContext {
 /* Another wild guess. */
 extern void __DestructExceptionObject(EXCEPTION_RECORD *rec, int nothrow);
 
-#if LJ_TARGET_X64 && defined(MINGW_SDK_INIT)
-/* Workaround for broken MinGW64 declaration. */
-VOID RtlUnwindEx_FIXED(PVOID,PVOID,PVOID,PVOID,PVOID,PVOID) asm("RtlUnwindEx");
-#define RtlUnwindEx RtlUnwindEx_FIXED
-#endif
-
 #define LJ_MSVC_EXCODE		((DWORD)0xe06d7363)
 #define LJ_GCC_EXCODE		((DWORD)0x20474343)
 

+ 1 - 0
libs/LuaJIT/src/lj_errmsg.h

@@ -79,6 +79,7 @@ ERRDEF(SETFENV,	LUA_QL("setfenv") " cannot change environment of given object")
 ERRDEF(CORUN,	"cannot resume running coroutine")
 ERRDEF(CODEAD,	"cannot resume dead coroutine")
 ERRDEF(COSUSP,	"cannot resume non-suspended coroutine")
+ERRDEF(PRNGSD,	"PRNG seeding failed")
 ERRDEF(TABINS,	"wrong number of arguments to " LUA_QL("insert"))
 ERRDEF(TABCAT,	"invalid value (%s) at index %d in table for " LUA_QL("concat"))
 ERRDEF(TABSORT,	"invalid order function for sorting")

+ 5 - 1
libs/LuaJIT/src/lj_ffrecord.c

@@ -152,6 +152,8 @@ static void recff_stitch(jit_State *J)
   if (errcode) {
     if (errcode == LUA_ERRRUN)
       copyTV(L, L->top-1, L->top + (1 + LJ_FR2));
+    else
+      setintV(L->top-1, (int32_t)LJ_TRERR_RECERR);
     lj_err_throw(L, errcode);  /* Propagate errors. */
   }
 }
@@ -257,7 +259,7 @@ static void LJ_FASTCALL recff_setmetatable(jit_State *J, RecordFFData *rd)
     mtref = tref_isnil(mt) ? lj_ir_knull(J, IRT_TAB) : mt;
     emitir(IRT(IR_FSTORE, IRT_TAB), fref, mtref);
     if (!tref_isnil(mt))
-      emitir(IRT(IR_TBAR, IRT_TAB), tr, 0);
+      emitir(IRT(IR_TBAR, IRT_NIL), tr, 0);
     J->base[0] = tr;
     J->needsnap = 1;
   }  /* else: Interpreter will throw. */
@@ -1003,6 +1005,7 @@ static void recff_format(jit_State *J, RecordFFData *rd, TRef hdr, int sbufx)
   GCstr *fmt = argv2str(J, &rd->argv[arg]);
   FormatState fs;
   SFormat sf;
+  int nfmt = 0;
   /* Specialize to the format string. */
   emitir(IRTG(IR_EQ, IRT_STR), trfmt, lj_ir_kstr(J, fmt));
   lj_strfmt_init(&fs, strdata(fmt), fmt->len);
@@ -1080,6 +1083,7 @@ static void recff_format(jit_State *J, RecordFFData *rd, TRef hdr, int sbufx)
       recff_nyiu(J, rd);
       return;
     }
+    if (++nfmt > 100) lj_trace_err(J, LJ_TRERR_TRACEOV);
   }
   if (sbufx) {
     emitir(IRT(IR_USE, IRT_NIL), tr, 0);

+ 18 - 33
libs/LuaJIT/src/lj_gc.c

@@ -108,9 +108,6 @@ static void gc_mark_start(global_State *g)
   gc_markobj(g, tabref(mainthread(g)->env));
   gc_marktv(g, &g->registrytv);
   gc_mark_gcroot(g);
-#if LJ_HASFFI
-  if (ctype_ctsG(g)) gc_markobj(g, ctype_ctsG(g)->finalizer);
-#endif
   g->gc.state = GCSpropagate;
 }
 
@@ -190,8 +187,7 @@ static int gc_traverse_tab(global_State *g, GCtab *t)
     }
     if (weak) {  /* Weak tables are cleared in the atomic phase. */
 #if LJ_HASFFI
-      CTState *cts = ctype_ctsG(g);
-      if (cts && cts->finalizer == t) {
+      if (gcref(g->gcroot[GCROOT_FFI_FIN]) == obj2gco(t)) {
 	weak = (int)(~0u & ~LJ_GC_WEAKVAL);
       } else
 #endif
@@ -556,9 +552,8 @@ static void gc_finalize(lua_State *L)
     o->gch.marked &= (uint8_t)~LJ_GC_CDATA_FIN;
     /* Resolve finalizer. */
     setcdataV(L, &tmp, gco2cd(o));
-    tv = lj_tab_set(L, ctype_ctsG(g)->finalizer, &tmp);
+    tv = lj_tab_set(L, tabref(g->gcroot[GCROOT_FFI_FIN]), &tmp);
     if (!tvisnil(tv)) {
-      g->gc.nocdatafin = 0;
       copyTV(L, &tmp, tv);
       setnilV(tv);  /* Clear entry in finalizer table. */
       gc_call_finalizer(g, L, &tmp, o);
@@ -588,35 +583,31 @@ void lj_gc_finalize_udata(lua_State *L)
 void lj_gc_finalize_cdata(lua_State *L)
 {
   global_State *g = G(L);
-  CTState *cts = ctype_ctsG(g);
-  if (cts) {
-    GCtab *t = cts->finalizer;
-    Node *node = noderef(t->node);
-    ptrdiff_t i;
-    setgcrefnull(t->metatable);  /* Mark finalizer table as disabled. */
-    for (i = (ptrdiff_t)t->hmask; i >= 0; i--)
-      if (!tvisnil(&node[i].val) && tviscdata(&node[i].key)) {
-	GCobj *o = gcV(&node[i].key);
-	TValue tmp;
-	makewhite(g, o);
-	o->gch.marked &= (uint8_t)~LJ_GC_CDATA_FIN;
-	copyTV(L, &tmp, &node[i].val);
-	setnilV(&node[i].val);
-	gc_call_finalizer(g, L, &tmp, o);
-      }
-  }
+  GCtab *t = tabref(g->gcroot[GCROOT_FFI_FIN]);
+  Node *node = noderef(t->node);
+  ptrdiff_t i;
+  setgcrefnull(t->metatable);  /* Mark finalizer table as disabled. */
+  for (i = (ptrdiff_t)t->hmask; i >= 0; i--)
+    if (!tvisnil(&node[i].val) && tviscdata(&node[i].key)) {
+      GCobj *o = gcV(&node[i].key);
+      TValue tmp;
+      makewhite(g, o);
+      o->gch.marked &= (uint8_t)~LJ_GC_CDATA_FIN;
+      copyTV(L, &tmp, &node[i].val);
+      setnilV(&node[i].val);
+      gc_call_finalizer(g, L, &tmp, o);
+    }
 }
 #endif
 
 /* Free all remaining GC objects. */
 void lj_gc_freeall(global_State *g)
 {
-  MSize i, strmask;
+  MSize i;
   /* Free everything, except super-fixed objects (the main thread). */
   g->gc.currentwhite = LJ_GC_WHITES | LJ_GC_SFIXED;
   gc_fullsweep(g, &g->gc.root);
-  strmask = g->str.mask;
-  for (i = 0; i <= strmask; i++)  /* Free all string hash chains. */
+  for (i = g->str.mask; i != ~(MSize)0; i--)  /* Free all string hash chains. */
     gc_sweepstr(g, &g->str.tab[i]);
 }
 
@@ -697,9 +688,6 @@ static size_t gc_onestep(lua_State *L)
 	lj_str_resize(L, g->str.mask >> 1);  /* Shrink string table. */
       if (gcref(g->gc.mmudata)) {  /* Need any finalizations? */
 	g->gc.state = GCSfinalize;
-#if LJ_HASFFI
-	g->gc.nocdatafin = 1;
-#endif
       } else {  /* Otherwise skip this phase to help the JIT. */
 	g->gc.state = GCSpause;  /* End of GC cycle. */
 	g->gc.debt = 0;
@@ -719,9 +707,6 @@ static size_t gc_onestep(lua_State *L)
 	g->gc.estimate -= GCFINALIZECOST;
       return GCFINALIZECOST;
     }
-#if LJ_HASFFI
-    if (!g->gc.nocdatafin) lj_tab_rehash(L, ctype_ctsG(g)->finalizer);
-#endif
     g->gc.state = GCSpause;  /* End of GC cycle. */
     g->gc.debt = 0;
     return 0;

+ 2 - 2
libs/LuaJIT/src/lj_jit.h

@@ -460,8 +460,8 @@ typedef struct jit_State {
 #endif
 
   IRIns *irbuf;		/* Temp. IR instruction buffer. Biased with REF_BIAS. */
-  IRRef irtoplim;	/* Upper limit of instuction buffer (biased). */
-  IRRef irbotlim;	/* Lower limit of instuction buffer (biased). */
+  IRRef irtoplim;	/* Upper limit of instruction buffer (biased). */
+  IRRef irbotlim;	/* Lower limit of instruction buffer (biased). */
   IRRef loopref;	/* Last loop reference or ref of final LOOP (or 0). */
 
   MSize sizesnap;	/* Size of temp. snapshot buffer. */

+ 10 - 9
libs/LuaJIT/src/lj_load.c

@@ -101,29 +101,30 @@ LUALIB_API int luaL_loadfilex(lua_State *L, const char *filename,
   FileReaderCtx ctx;
   int status;
   const char *chunkname;
+  int err = 0;
   if (filename) {
+    chunkname = lua_pushfstring(L, "@%s", filename);
     ctx.fp = fopen(filename, "rb");
     if (ctx.fp == NULL) {
+      L->top--;
       lua_pushfstring(L, "cannot open %s: %s", filename, strerror(errno));
       return LUA_ERRFILE;
     }
-    chunkname = lua_pushfstring(L, "@%s", filename);
   } else {
     ctx.fp = stdin;
     chunkname = "=stdin";
   }
   status = lua_loadx(L, reader_file, &ctx, chunkname, mode);
-  if (ferror(ctx.fp)) {
-    L->top -= filename ? 2 : 1;
-    lua_pushfstring(L, "cannot read %s: %s", chunkname+1, strerror(errno));
-    if (filename)
-      fclose(ctx.fp);
-    return LUA_ERRFILE;
-  }
+  if (ferror(ctx.fp)) err = errno;
   if (filename) {
+    fclose(ctx.fp);
     L->top--;
     copyTV(L, L->top-1, L->top);
-    fclose(ctx.fp);
+  }
+  if (err) {
+    L->top--;
+    lua_pushfstring(L, "cannot read %s: %s", chunkname+1, strerror(err));
+    return LUA_ERRFILE;
   }
   return status;
 }

+ 4 - 1
libs/LuaJIT/src/lj_obj.h

@@ -579,6 +579,9 @@ typedef enum {
   GCROOT_BASEMT_NUM = GCROOT_BASEMT + ~LJ_TNUMX,
   GCROOT_IO_INPUT,	/* Userdata for default I/O input file. */
   GCROOT_IO_OUTPUT,	/* Userdata for default I/O output file. */
+#if LJ_HASFFI
+  GCROOT_FFI_FIN,	/* FFI finalizer table. */
+#endif
   GCROOT_MAX
 } GCRootID;
 
@@ -592,7 +595,7 @@ typedef struct GCState {
   GCSize threshold;	/* Memory threshold. */
   uint8_t currentwhite;	/* Current white color. */
   uint8_t state;	/* GC state. */
-  uint8_t nocdatafin;	/* No cdata finalizer called. */
+  uint8_t unused0;
 #if LJ_64
   uint8_t lightudnum;	/* Number of lightuserdata segments - 1. */
 #else

+ 19 - 6
libs/LuaJIT/src/lj_opt_fold.c

@@ -1135,7 +1135,7 @@ LJFOLDF(shortcut_conv_num_int)
 }
 
 LJFOLD(CONV CONV IRCONV_INT_NUM)  /* _INT */
-LJFOLD(CONV CONV IRCONV_U32_NUM)  /* _U32*/
+LJFOLD(CONV CONV IRCONV_U32_NUM)  /* _U32 */
 LJFOLDF(simplify_conv_int_num)
 {
   /* Fold even across PHI to avoid expensive num->int conversions in loop. */
@@ -1170,8 +1170,10 @@ LJFOLDF(simplify_conv_i64_num)
 
 LJFOLD(CONV CONV IRCONV_INT_I64)  /* _INT or _U32 */
 LJFOLD(CONV CONV IRCONV_INT_U64)  /* _INT or _U32 */
+LJFOLD(CONV CONV IRCONV_INT_U32)  /* _INT or _U32 */
 LJFOLD(CONV CONV IRCONV_U32_I64)  /* _INT or _U32 */
 LJFOLD(CONV CONV IRCONV_U32_U64)  /* _INT or _U32 */
+LJFOLD(CONV CONV IRCONV_U32_INT)  /* _INT or _U32 */
 LJFOLDF(simplify_conv_int_i64)
 {
   int src;
@@ -1216,14 +1218,13 @@ LJFOLDF(simplify_tobit_conv)
   return NEXTFOLD;
 }
 
-/* Shortcut floor/ceil/round + IRT_NUM <- IRT_INT/IRT_U32 conversion. */
+/* Shortcut floor/ceil/trunc + IRT_NUM <- integer conversion. */
 LJFOLD(FPMATH CONV IRFPM_FLOOR)
 LJFOLD(FPMATH CONV IRFPM_CEIL)
 LJFOLD(FPMATH CONV IRFPM_TRUNC)
 LJFOLDF(simplify_floor_conv)
 {
-  if ((fleft->op2 & IRCONV_SRCMASK) == IRT_INT ||
-      (fleft->op2 & IRCONV_SRCMASK) == IRT_U32)
+  if ((uint32_t)(fleft->op2 & IRCONV_SRCMASK) - (uint32_t)IRT_I8 <= (uint32_t)(IRT_U64 - IRT_U8))
     return LEFTFOLD;
   return NEXTFOLD;
 }
@@ -1935,9 +1936,10 @@ LJFOLDF(abc_k)
 LJFOLD(ABC any any)
 LJFOLDF(abc_invar)
 {
-  /* Invariant ABC marked as PTR. Drop if op1 is invariant, too. */
+  /* Invariant ABC marked as P32 or U32. Drop if op1 is invariant too. */
   if (!irt_isint(fins->t) && fins->op1 < J->chain[IR_LOOP] &&
-      !irt_isphi(IR(fins->op1)->t))
+      (irt_isu32(fins->t) ||
+       (!irref_isk(fins->op1) && !irt_isphi(IR(fins->op1)->t))))
     return DROPFOLD;
   return NEXTFOLD;
 }
@@ -2485,6 +2487,17 @@ LJFOLD(CNEW any any)
 LJFOLD(XSNEW any any)
 LJFOLDX(lj_ir_emit)
 
+/* -- Miscellaneous ------------------------------------------------------- */
+
+LJFOLD(CARG any any)
+LJFOLDF(cse_carg)
+{
+  TRef tr = lj_opt_cse(J);
+  if (tref_ref(tr) < J->chain[IR_LOOP])  /* CSE across loop? */
+    return EMITFOLD;  /* Raw emit. Assumes fins is left intact by CSE. */
+  return tr;
+}
+
 /* ------------------------------------------------------------------------ */
 
 /* Every entry in the generated hash table is a 32 bit pattern:

+ 2 - 1
libs/LuaJIT/src/lj_opt_narrow.c

@@ -341,7 +341,8 @@ static int narrow_conv_backprop(NarrowConv *nc, IRRef ref, int depth)
       NarrowIns *savesp = nc->sp;
       int count = narrow_conv_backprop(nc, ir->op1, depth);
       count += narrow_conv_backprop(nc, ir->op2, depth);
-      if (count <= 1) {  /* Limit total number of conversions. */
+      /* Limit total number of conversions. */
+      if (count <= 1 && nc->sp < nc->maxsp) {
 	*nc->sp++ = NARROWINS(IRT(ir->o, nc->t), ref);
 	return count;
       }

+ 6 - 4
libs/LuaJIT/src/lj_parse.c

@@ -2339,11 +2339,15 @@ static void parse_return(LexState *ls)
     BCReg nret = expr_list(ls, &e);
     if (nret == 1) {  /* Return one result. */
       if (e.k == VCALL) {  /* Check for tail call. */
+#ifdef LUAJIT_DISABLE_TAILCALL
+	goto notailcall;
+#else
 	BCIns *ip = bcptr(fs, &e);
 	/* It doesn't pay off to add BC_VARGT just for 'return ...'. */
 	if (bc_op(*ip) == BC_VARG) goto notailcall;
 	fs->pc--;
 	ins = BCINS_AD(bc_op(*ip)-BC_CALL+BC_CALLT, bc_a(*ip), bc_c(*ip));
+#endif
       } else {  /* Can return the result from any register. */
 	ins = BCINS_AD(BC_RET1, expr_toanyreg(fs, &e), 2);
       }
@@ -2517,11 +2521,9 @@ static void parse_for_num(LexState *ls, GCstr *varname, BCLine line)
 */
 static int predict_next(LexState *ls, FuncState *fs, BCPos pc)
 {
-  BCIns ins;
+  BCIns ins = fs->bcbase[pc].ins;
   GCstr *name;
   cTValue *o;
-  if (pc >= fs->bclim) return 0;
-  ins = fs->bcbase[pc].ins;
   switch (bc_op(ins)) {
   case BC_MOV:
     if (bc_d(ins) >= fs->nactvar) return 0;
@@ -2570,7 +2572,7 @@ static void parse_for_iter(LexState *ls, GCstr *indexname)
   assign_adjust(ls, 3, expr_list(ls, &e), &e);
   /* The iterator needs another 3 [4] slots (func [pc] | state ctl). */
   bcreg_bump(fs, 3+ls->fr2);
-  isnext = (nvars <= 5 && predict_next(ls, fs, exprpc));
+  isnext = (nvars <= 5 && fs->pc > exprpc && predict_next(ls, fs, exprpc));
   var_add(ls, 3);  /* Hidden control variables. */
   lex_check(ls, TK_do);
   loop = bcemit_AJ(fs, isnext ? BC_ISNEXT : BC_JMP, base, NO_JMP);

+ 50 - 12
libs/LuaJIT/src/lj_record.c

@@ -1025,7 +1025,9 @@ void lj_record_ret(jit_State *J, BCReg rbase, ptrdiff_t gotresults)
 	J->L->base = b + baseadj;
 	copyTV(J->L, b-(2<<LJ_FR2), &save);
       }
-      if (tr) {  /* Store final result. */
+      if (tr >= 0xffffff00) {
+	lj_err_throw(J->L, -(int32_t)tr);  /* Propagate errors. */
+      } else if (tr) {  /* Store final result. */
 	BCReg dst = bc_a(*(frame_contpc(frame)-1));
 	J->base[dst] = tr;
 	if (dst >= J->maxslot) {
@@ -1072,6 +1074,7 @@ int lj_record_mm_lookup(jit_State *J, RecordIndex *ix, MMS mm)
   } else if (tref_isudata(ix->tab)) {
     int udtype = udataV(&ix->tabv)->udtype;
     mt = tabref(udataV(&ix->tabv)->metatable);
+    mix.tab = emitir(IRT(IR_FLOAD, IRT_TAB), ix->tab, IRFL_UDATA_META);
     /* The metatables of special userdata objects are treated as immutable. */
     if (udtype != UDTYPE_USERDATA) {
       cTValue *mo;
@@ -1085,6 +1088,8 @@ int lj_record_mm_lookup(jit_State *J, RecordIndex *ix, MMS mm)
       }
   immutable_mt:
       mo = lj_tab_getstr(mt, mmname_str(J2G(J), mm));
+      ix->mt = mix.tab;
+      ix->mtv = mt;
       if (!mo || tvisnil(mo))
 	return 0;  /* No metamethod. */
       /* Treat metamethod or index table as immutable, too. */
@@ -1092,11 +1097,8 @@ int lj_record_mm_lookup(jit_State *J, RecordIndex *ix, MMS mm)
 	lj_trace_err(J, LJ_TRERR_BADTYPE);
       copyTV(J->L, &ix->mobjv, mo);
       ix->mobj = lj_ir_kgc(J, gcV(mo), tvisfunc(mo) ? IRT_FUNC : IRT_TAB);
-      ix->mtv = mt;
-      ix->mt = TREF_NIL;  /* Dummy value for comparison semantics. */
       return 1;  /* Got metamethod or index table. */
     }
-    mix.tab = emitir(IRT(IR_FLOAD, IRT_TAB), ix->tab, IRFL_UDATA_META);
   } else {
     /* Specialize to base metatable. Must flush mcode in lua_setmetatable(). */
     mt = tabref(basemt_obj(J2G(J), &ix->tabv));
@@ -1393,12 +1395,13 @@ static void rec_idx_abc(jit_State *J, TRef asizeref, TRef ikey, uint32_t asize)
       /* Runtime value for stop of loop is within bounds? */
       if ((uint64_t)stop + ofs < (uint64_t)asize) {
 	/* Emit invariant bounds check for stop. */
-	emitir(IRTG(IR_ABC, IRT_P32), asizeref, ofs == 0 ? J->scev.stop :
+	uint32_t abc = IRTG(IR_ABC, tref_isk(asizeref) ? IRT_U32 : IRT_P32);
+	emitir(abc, asizeref, ofs == 0 ? J->scev.stop :
 	       emitir(IRTI(IR_ADD), J->scev.stop, ofsref));
 	/* Emit invariant bounds check for start, if not const or negative. */
 	if (!(J->scev.dir && J->scev.start &&
 	      (int64_t)IR(J->scev.start)->i + ofs >= 0))
-	  emitir(IRTG(IR_ABC, IRT_P32), asizeref, ikey);
+	  emitir(abc, asizeref, ikey);
 	return;
       }
     }
@@ -2075,12 +2078,21 @@ static TRef rec_tnew(jit_State *J, uint32_t ah)
 
 /* -- Concatenation ------------------------------------------------------- */
 
-static TRef rec_cat(jit_State *J, BCReg baseslot, BCReg topslot)
+typedef struct RecCatDataCP {
+  jit_State *J;
+  BCReg baseslot, topslot;
+  TRef tr;
+} RecCatDataCP;
+
+static TValue *rec_mm_concat_cp(lua_State *L, lua_CFunction dummy, void *ud)
 {
+  RecCatDataCP *rcd = (RecCatDataCP *)ud;
+  jit_State *J = rcd->J;
+  BCReg baseslot = rcd->baseslot, topslot = rcd->topslot;
   TRef *top = &J->base[topslot];
-  TValue savetv[5+LJ_FR2];
   BCReg s;
   RecordIndex ix;
+  UNUSED(L); UNUSED(dummy);
   lj_assertJ(baseslot < topslot, "bad CAT arg");
   for (s = baseslot; s <= topslot; s++)
     (void)getslot(J, s);  /* Ensure all arguments have a reference. */
@@ -2102,7 +2114,10 @@ static TRef rec_cat(jit_State *J, BCReg baseslot, BCReg topslot)
     } while (trp <= top);
     tr = emitir(IRTG(IR_BUFSTR, IRT_STR), tr, hdr);
     J->maxslot = (BCReg)(xbase - J->base);
-    if (xbase == base) return tr;  /* Return simple concatenation result. */
+    if (xbase == base) {
+      rcd->tr = tr;  /* Return simple concatenation result. */
+      return NULL;
+    }
     /* Pass partial result. */
     topslot = J->maxslot--;
     *xbase = tr;
@@ -2115,10 +2130,31 @@ static TRef rec_cat(jit_State *J, BCReg baseslot, BCReg topslot)
   copyTV(J->L, &ix.tabv, &J->L->base[topslot-1]);
   ix.tab = top[-1];
   ix.key = top[0];
-  memcpy(savetv, &J->L->base[topslot-1], sizeof(savetv));  /* Save slots. */
   rec_mm_arith(J, &ix, MM_concat);  /* Call __concat metamethod. */
-  memcpy(&J->L->base[topslot-1], savetv, sizeof(savetv));  /* Restore slots. */
-  return 0;  /* No result yet. */
+  rcd->tr = 0;  /* No result yet. */
+  return NULL;
+}
+
+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;
+  RecCatDataCP rcd;
+  int errcode;
+  rcd.J = J;
+  rcd.baseslot = baseslot;
+  rcd.topslot = topslot;
+  memcpy(savetv, &L->base[topslot-1], sizeof(savetv));  /* Save slots. */
+  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. */
+  if (errcode) {
+    L->top = L->base + delta;
+    copyTV(L, L->top++, &errobj);
+    return (TRef)(-errcode);
+  }
+  return rcd.tr;
 }
 
 /* -- Record bytecode ops ------------------------------------------------- */
@@ -2440,6 +2476,8 @@ void lj_record_ins(jit_State *J)
 
   case BC_CAT:
     rc = rec_cat(J, rb, rc);
+    if (rc >= 0xffffff00)
+      lj_err_throw(J->L, -(int32_t)rc);  /* Propagate errors. */
     break;
 
   /* -- Constant and move ops --------------------------------------------- */

+ 6 - 3
libs/LuaJIT/src/lj_snap.c

@@ -801,7 +801,6 @@ static void snap_restoredata(jit_State *J, GCtrace *T, ExitState *ex,
 	*(lua_Number *)dst = (lua_Number)*(int32_t *)dst;
 	return;
       }
-      src = (int32_t *)&ex->gpr[r-RID_MIN_GPR];
 #if !LJ_SOFTFP
       if (r >= RID_MAX_GPR) {
 	src = (int32_t *)&ex->fpr[r-RID_MIN_FPR];
@@ -815,7 +814,10 @@ static void snap_restoredata(jit_State *J, GCtrace *T, ExitState *ex,
 #endif
       } else
 #endif
-      if (LJ_64 && LJ_BE && sz == 4) src++;
+      {
+	src = (int32_t *)&ex->gpr[r-RID_MIN_GPR];
+	if (LJ_64 && LJ_BE && sz == 4) src++;
+      }
     }
   }
   lj_assertJ(sz == 1 || sz == 2 || sz == 4 || sz == 8,
@@ -955,7 +957,8 @@ const BCIns *lj_snap_restore(jit_State *J, void *exptr)
   lua_State *L = J->L;
 
   /* Set interpreter PC to the next PC to get correct error messages. */
-  setcframe_pc(cframe_raw(L->cframe), pc+1);
+  setcframe_pc(L->cframe, pc+1);
+  setcframe_pc(cframe_raw(cframe_prev(L->cframe)), pc);
 
   /* Make sure the stack is big enough for the slots from the snapshot. */
   if (LJ_UNLIKELY(L->base + snap->topslot >= tvref(L->maxstack))) {

+ 3 - 0
libs/LuaJIT/src/lj_state.c

@@ -196,6 +196,9 @@ static TValue *cpluaopen(lua_State *L, lua_CFunction dummy, void *ud)
   lj_lex_init(L);
   fixstring(lj_err_str(L, LJ_ERR_ERRMEM));  /* Preallocate memory error msg. */
   g->gc.threshold = 4*g->gc.total;
+#if LJ_HASFFI
+  lj_ctype_initfin(L);
+#endif
   lj_trace_initstate(g);
   lj_err_verify();
   return NULL;

+ 1 - 8
libs/LuaJIT/src/lj_tab.c

@@ -367,13 +367,6 @@ static void rehashtab(lua_State *L, GCtab *t, cTValue *ek)
   lj_tab_resize(L, t, asize, hsize2hbits(total));
 }
 
-#if LJ_HASFFI
-void lj_tab_rehash(lua_State *L, GCtab *t)
-{
-  rehashtab(L, t, niltv(L));
-}
-#endif
-
 void lj_tab_reasize(lua_State *L, GCtab *t, uint32_t nasize)
 {
   lj_tab_resize(L, t, nasize+1, t->hmask > 0 ? lj_fls(t->hmask)+1 : 0);
@@ -642,7 +635,7 @@ LJ_NOINLINE static MSize tab_len_slow(GCtab *t, size_t hi)
   while ((tv = lj_tab_getint(t, (int32_t)hi)) && !tvisnil(tv)) {
     lo = hi;
     hi += hi;
-    if (hi > (size_t)(INT_MAX-2)) {  /* Punt and do a linear search. */
+    if (hi > (size_t)(0x7fffffff - 2)) {  /* Punt and do a linear search. */
       lo = 1;
       while ((tv = lj_tab_getint(t, (int32_t)lo)) && !tvisnil(tv)) lo++;
       return (MSize)(lo - 1);

+ 0 - 3
libs/LuaJIT/src/lj_tab.h

@@ -60,9 +60,6 @@ LJ_FUNC GCtab * LJ_FASTCALL lj_tab_new1(lua_State *L, uint32_t ahsize);
 LJ_FUNCA GCtab * LJ_FASTCALL lj_tab_dup(lua_State *L, const GCtab *kt);
 LJ_FUNC void LJ_FASTCALL lj_tab_clear(GCtab *t);
 LJ_FUNC void LJ_FASTCALL lj_tab_free(global_State *g, GCtab *t);
-#if LJ_HASFFI
-LJ_FUNC void lj_tab_rehash(lua_State *L, GCtab *t);
-#endif
 LJ_FUNC void lj_tab_resize(lua_State *L, GCtab *t, uint32_t asize, uint32_t hbits);
 LJ_FUNCA void lj_tab_reasize(lua_State *L, GCtab *t, uint32_t nasize);
 

+ 2 - 0
libs/LuaJIT/src/lj_target_arm64.h

@@ -281,6 +281,7 @@ typedef enum A64Ins {
   A64I_FSQRTd = 0x1e61c000,
   A64I_LDRs = 0xbd400000,
   A64I_LDRd = 0xfd400000,
+  A64I_LDRLd = 0x5c000000,
   A64I_STRs = 0xbd000000,
   A64I_STRd = 0xfd000000,
   A64I_LDPs = 0x2d400000,
@@ -320,6 +321,7 @@ typedef enum A64Ins {
   A64I_FMOV_R_D = 0x9e660000,
   A64I_FMOV_D_R = 0x9e670000,
   A64I_FMOV_DI = 0x1e601000,
+  A64I_MOVI_DI = 0x2f000400,
 } A64Ins;
 
 #define A64I_BR_AUTH	(LJ_ABI_PAUTH ? A64I_BRAAZ : A64I_BR)

+ 1 - 1
libs/LuaJIT/src/lj_traceerr.h

@@ -13,7 +13,7 @@ TREDEF(STACKOV,	"trace too deep")
 TREDEF(SNAPOV,	"too many snapshots")
 TREDEF(BLACKL,	"blacklisted")
 TREDEF(RETRY,	"retry recording")
-TREDEF(NYIBC,	"NYI: bytecode %d")
+TREDEF(NYIBC,	"NYI: bytecode %s")
 
 /* Recording loop ops. */
 TREDEF(LLEAVE,	"leaving loop in root trace")

+ 5 - 3
libs/LuaJIT/src/luaconf.h

@@ -9,7 +9,6 @@
 #ifndef WINVER
 #define WINVER 0x0501
 #endif
-#include <limits.h>
 #include <stddef.h>
 
 /* Default path for loading Lua and C modules with require(). */
@@ -37,7 +36,6 @@
 #endif
 #define LUA_LROOT	"/usr/local"
 #define LUA_LUADIR	"/lua/5.1/"
-#define LUA_LJDIR	"/luajit-2.1/"
 
 #ifdef LUA_ROOT
 #define LUA_JROOT	LUA_ROOT
@@ -51,7 +49,11 @@
 #define LUA_RCPATH
 #endif
 
-#define LUA_JPATH	";" LUA_JROOT "/share" LUA_LJDIR "?.lua"
+#ifndef LUA_LJDIR
+#define LUA_LJDIR	LUA_JROOT "/share/luajit-2.1"
+#endif
+
+#define LUA_JPATH	";" LUA_LJDIR "/?.lua"
 #define LUA_LLDIR	LUA_LROOT "/share" LUA_LUADIR
 #define LUA_LCDIR	LUA_LROOT "/" LUA_LMULTILIB LUA_LUADIR
 #define LUA_LLPATH	";" LUA_LLDIR "?.lua;" LUA_LLDIR "?/init.lua"

+ 22 - 11
libs/LuaJIT/src/msvcbuild.bat

@@ -13,10 +13,15 @@
 @if not defined INCLUDE goto :FAIL
 
 @setlocal
-@rem Add more debug flags here, e.g. DEBUGCFLAGS=/DLUA_USE_APICHECK
+@rem Add more debug flags here, e.g. DEBUGCFLAGS=/DLUA_USE_ASSERT
 @set DEBUGCFLAGS=
 @set LJCOMPILE=cl /nologo /c /O2 /W3 /D_CRT_SECURE_NO_DEPRECATE /D_CRT_STDIO_INLINE=__declspec(dllexport)__inline
-@set LJDYNBUILD=/MD /DLUA_BUILD_AS_DLL
+@set LJDYNBUILD=/DLUA_BUILD_AS_DLL /MD
+@set LJDYNBUILD_DEBUG=/DLUA_BUILD_AS_DLL /MDd 
+@set LJCOMPILETARGET=/Zi
+@set LJLINKTYPE=/DEBUG /RELEASE
+@set LJLINKTYPE_DEBUG=/DEBUG
+@set LJLINKTARGET=/OPT:REF /OPT:ICF /INCREMENTAL:NO
 @set LJLINK=link /nologo
 @set LJMT=mt /nologo
 @set LJLIB=lib /nologo /nodefaultlib
@@ -25,7 +30,6 @@
 @set DASC=vm_x64.dasc
 @set LJDLLNAME=lua51.dll
 @set LJLIBNAME=lua51.lib
-@set BUILDTYPE=release
 @set ALL_LIB=lib_base.c lib_math.c lib_bit.c lib_string.c lib_table.c lib_io.c lib_os.c lib_package.c lib_debug.c lib_jit.c lib_ffi.c lib_buffer.c
 
 @setlocal
@@ -92,17 +96,17 @@ buildvm -m folddef -o lj_folddef.h lj_opt_fold.c
 
 @if "%1" neq "debug" goto :NODEBUG
 @shift
-@set BUILDTYPE=debug
-@set LJCOMPILE=%LJCOMPILE% /Zi %DEBUGCFLAGS%
-@set LJDYNBUILD=/MD /DLUA_BUILD_AS_DLL
-@set LJLINK=%LJLINK% /opt:ref /opt:icf /incremental:no
+@set LJCOMPILE=%LJCOMPILE% %DEBUGCFLAGS%
+@set LJDYNBUILD=%LJDYNBUILD_DEBUG%
+@set LJLINKTYPE=%LJLINKTYPE_DEBUG%
 :NODEBUG
-@set LJLINK=%LJLINK% /%BUILDTYPE%
+@set LJCOMPILE=%LJCOMPILE% %LJCOMPILETARGET%
+@set LJLINK=%LJLINK% %LJLINKTYPE% %LJLINKTARGET%
 @if "%1"=="amalg" goto :AMALGDLL
 @if "%1"=="static" goto :STATIC
 %LJCOMPILE% %LJDYNBUILD% lj_*.c lib_*.c
 @if errorlevel 1 goto :BAD
-%LJLINK% /DLL /out:%LJDLLNAME% lj_*.obj lib_*.obj
+%LJLINK% /DLL /OUT:%LJDLLNAME% lj_*.obj lib_*.obj
 @if errorlevel 1 goto :BAD
 @goto :MTDLL
 :STATIC
@@ -112,9 +116,16 @@ buildvm -m folddef -o lj_folddef.h lj_opt_fold.c
 @if errorlevel 1 goto :BAD
 @goto :MTDLL
 :AMALGDLL
+@if "%2"=="static" goto :AMALGSTATIC
 %LJCOMPILE% %LJDYNBUILD% ljamalg.c
 @if errorlevel 1 goto :BAD
-%LJLINK% /DLL /out:%LJDLLNAME% ljamalg.obj lj_vm.obj
+%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
 @if errorlevel 1 goto :BAD
 :MTDLL
 if exist %LJDLLNAME%.manifest^
@@ -122,7 +133,7 @@ if exist %LJDLLNAME%.manifest^
 
 %LJCOMPILE% luajit.c
 @if errorlevel 1 goto :BAD
-%LJLINK% /out:luajit.exe luajit.obj %LJLIBNAME%
+%LJLINK% /OUT:luajit.exe luajit.obj %LJLIBNAME%
 @if errorlevel 1 goto :BAD
 if exist luajit.exe.manifest^
   %LJMT% -manifest luajit.exe.manifest -outputresource:luajit.exe

+ 7 - 6
libs/LuaJIT/src/nxbuild.bat

@@ -99,20 +99,21 @@ buildvm -m folddef -o lj_folddef.h lj_opt_fold.c
 @if errorlevel 1 goto :BAD
 
 @rem ---- Cross compiler ----
+@set NXCOMPILER_ROOT="%NINTENDO_SDK_ROOT%\Compilers\NintendoClang"
 @if "%platform%" neq "x64" goto :NX32_CROSSBUILD
-@set LJCOMPILE="%NINTENDO_SDK_ROOT%\Compilers\NX\nx\aarch64\bin\clang" -Wall -I%NINTENDO_SDK_ROOT%\Include %DASMTARGET% -DLUAJIT_OS=LUAJIT_OS_OTHER -DLUAJIT_DISABLE_JIT -DLUAJIT_DISABLE_FFI -DLUAJIT_USE_SYSMALLOC -c
-@set LJLIB="%NINTENDO_SDK_ROOT%\Compilers\NX\nx\aarch64\bin\aarch64-nintendo-nx-elf-ar" rc
+@set LJCOMPILE="%NXCOMPILER_ROOT%\bin\clang" --target=aarch64-nintendo-nx-elf -Wall -I%NINTENDO_SDK_ROOT%\Include %DASMTARGET% -DLUAJIT_OS=LUAJIT_OS_OTHER -DLUAJIT_DISABLE_JIT -DLUAJIT_DISABLE_FFI -DLUAJIT_USE_SYSMALLOC -c
+@set LJLIB="%NXCOMPILER_ROOT%\bin\llvm-ar" rc
 @set TARGETLIB_SUFFIX=nx64
 
-%NINTENDO_SDK_ROOT%\Compilers\NX\nx\aarch64\bin\aarch64-nintendo-nx-elf-as -o lj_vm.o lj_vm.s
+%NXCOMPILER_ROOT%\bin\clang --target=aarch64-nintendo-nx-elf -o lj_vm.o -c lj_vm.s
 goto :DEBUGCHECK
 
 :NX32_CROSSBUILD
-@set LJCOMPILE="%NINTENDO_SDK_ROOT%\Compilers\NX\nx\armv7l\bin\clang" -Wall -I%NINTENDO_SDK_ROOT%\Include %DASMTARGET% -DLUAJIT_OS=LUAJIT_OS_OTHER -DLUAJIT_DISABLE_JIT -DLUAJIT_DISABLE_FFI -DLUAJIT_USE_SYSMALLOC -c
-@set LJLIB="%NINTENDO_SDK_ROOT%\Compilers\NX\nx\armv7l\bin\armv7l-nintendo-nx-eabihf-ar" rc
+@set LJCOMPILE="%NXCOMPILER_ROOT%\bin\clang" --target=armv7l-nintendo-nx-eabihf -Wall -I%NINTENDO_SDK_ROOT%\Include %DASMTARGET% -DLUAJIT_OS=LUAJIT_OS_OTHER -DLUAJIT_DISABLE_JIT -DLUAJIT_DISABLE_FFI -DLUAJIT_USE_SYSMALLOC -c
+@set LJLIB="%NXCOMPILER_ROOT%\bin\llvm-ar" rc
 @set TARGETLIB_SUFFIX=nx32
 
-%NINTENDO_SDK_ROOT%\Compilers\NX\nx\armv7l\bin\armv7l-nintendo-nx-eabihf-as -o lj_vm.o lj_vm.s
+%NXCOMPILER_ROOT%\bin\clang --target=armv7l-nintendo-nx-eabihf -o lj_vm.o -c lj_vm.s
 :DEBUGCHECK
 
 @if "%1" neq "debug" goto :NODEBUG

+ 12 - 1
libs/LuaJIT/src/vm_arm.dasc

@@ -1823,9 +1823,19 @@ static void build_subroutines(BuildCtx *ctx)
   |
   |//-- Bit library --------------------------------------------------------
   |
-  |// FP number to bit conversion for soft-float. Clobbers r0-r3.
   |->vm_tobit_fb:
   |  bhi ->fff_fallback
+  |.if FPU
+  |// FP number to bit conversion for hard-float. Clobbers r0, d0-d1.
+  |  vldr d1, >9
+  |  vmov d0, CARG1, CARG2
+  |  vadd.f64 d0, d0, d1
+  |  vmov CARG1, s0
+  |  bx lr
+  |9:
+  |  .long 0, 0x43380000		// (double)(2^52 + 2^51).
+  |.else
+  |// FP number to bit conversion for soft-float. Clobbers r0-r3.
   |->vm_tobit:
   |  lsl RB, CARG2, #1
   |  adds RB, RB, #0x00200000
@@ -1850,6 +1860,7 @@ static void build_subroutines(BuildCtx *ctx)
   |  orr CARG1, CARG4, CARG1, lsl RB
   |   rsblt CARG1, CARG1, #0
   |  bx lr
+  |.endif
   |
   |.macro .ffunc_bit, name
   |  .ffunc_1 bit_..name

+ 6 - 15
libs/LuaJIT/src/vm_arm64.dasc

@@ -1658,22 +1658,14 @@ static void build_subroutines(BuildCtx *ctx)
   |
   |//-- Bit library --------------------------------------------------------
   |
-  |// FP number to bit conversion for soft-float. Clobbers CARG1-CARG3
+  |// FP number to bit conversion. Clobbers CARG1-CARG2, FARG1-FARG2.
   |->vm_tobit_fb:
   |  bls ->fff_fallback
-  |  add CARG2, CARG1, CARG1
-  |  mov CARG3, #1076
-  |  sub CARG3, CARG3, CARG2, lsr #53
-  |  cmp CARG3, #53
-  |  bhi >1
-  |  and CARG2, CARG2, #U64x(001fffff,ffffffff)
-  |  orr CARG2, CARG2, #U64x(00200000,00000000)
-  |   cmp CARG1, #0
-  |  lsr CARG2, CARG2, CARG3
-  |   cneg CARG1w, CARG2w, mi
-  |  br lr
-  |1:
-  |  mov CARG1w, #0
+  |  fmov FARG1, CARG1
+  |  movz CARG2, #0x4338, lsl #48
+  |  fmov FARG2, CARG2
+  |  fadd FARG1, FARG1, FARG2
+  |  fmov CARG1w, s0
   |  br lr
   |
   |.macro .ffunc_bit, name
@@ -4220,7 +4212,6 @@ static void emit_asm_debug(BuildCtx *ctx)
 	"LEFDEY:\n\n", fcsize);
     }
 #endif
-    fprintf(ctx->fp, ".subsections_via_symbols\n");
     }
     break;
 #endif

+ 3 - 2
libs/LuaJIT/src/vm_mips64.dasc

@@ -1420,9 +1420,10 @@ static void build_subroutines(BuildCtx *ctx)
   |  sltu AT, TMP1, TMP2
   |  bnez AT, ->fff_fallback
   |.  lbu TMP3, DISPATCH_GL(hookmask)(DISPATCH)
-  |  daddiu NARGS8:RC, NARGS8:RC, -8
-  |  bltz NARGS8:RC, ->fff_fallback
+  |  daddiu NARGS8:TMP0, NARGS8:RC, -8
+  |  bltz NARGS8:TMP0, ->fff_fallback
   |.   move TMP2, BASE
+  |  move NARGS8:RC, NARGS8:TMP0
   |   daddiu BASE, BASE, 16
   |  // Remember active hook before pcall.
   |  srl TMP3, TMP3, HOOK_ACTIVE_SHIFT

+ 0 - 1
libs/LuaJIT/src/vm_x64.dasc

@@ -4941,7 +4941,6 @@ static void emit_asm_debug(BuildCtx *ctx)
 	  "LEFDEY:\n\n", fcsize);
     }
 #endif
-    fprintf(ctx->fp, ".subsections_via_symbols\n");
     }
     break;
 #endif

+ 0 - 1
libs/LuaJIT/src/vm_x86.dasc

@@ -5841,7 +5841,6 @@ static void emit_asm_debug(BuildCtx *ctx)
 	  fprintf(ctx->fp, "L_%s$stub:\n\t.indirect_symbol _%s\n\t.ascii \"\\364\\364\\364\\364\\364\"\n", *xn, *xn);
     }
 #endif
-    fprintf(ctx->fp, ".subsections_via_symbols\n");
     }
     break;
 #endif