libffi/ChangeLog

   1 2012-02-11  Eric Botcazou  <ebotcazou@adacore.com>
   2
   3         * src/sparc/v9.S (STACKFRAME): Bump to 176.
   4
   5 2012-02-10  Kai Tietz  <ktietz@redhat.com>
   6
   7         * configure.ac (AM_LTLDFLAGS): Add -no-undefine for x64
   8         windows target.
   9         * configure: Regenerated.
  10
  11 2012-02-08  Kai Tietz  <ktietz@redhat.com>
  12
  13         * src/prep_cif.c (ffi_prep_cif): Allow for X86_WIN32
  14         also FFI_THISCALL.
  15         * src/x86/ffi.c (ffi_closure_THISCALL): Add prototype.
  16         (FFI_INIT_TRAMPOLINE_THISCALL): New trampoline code.
  17         (ffi_prep_closure_loc): Add FFI_THISCALL support.
  18         * src/x86/ffitarget.h (FFI_TRAMPOLINE_SIZE): Adjust size.
  19         * src/x86/win32.S (ffi_closure_THISCALL): New closure code
  20         for thiscall-calling convention.
  21         * testsuite/libffi.call/closure_thiscall.c: New test.
  22
  23 2012-01-28  Kai Tietz  <ktietz@redhat.com>
  24
  25         * src/libffi/src/x86/ffi.c (ffi_call_win32): Add new
  26         argument to prototype for specify calling-convention.
  27         (ffi_call): Add support for stdcall/thiscall convention.
  28         (ffi_prep_args): Likewise.
  29         (ffi_raw_call): Likewise.
  30         * src/x86/ffitarget.h (ffi_abi): Add FFI_THISCALL and
  31         FFI_FASTCALL.
  32         * src/x86/win32.S (_ffi_call_win32): Add support for
  33         fastcall/thiscall calling-convention calls.
  34         * testsuite/libffi.call/fastthis1_win32.c: New test.
  35         * testsuite/libffi.call/fastthis2_win32.c: New test.
  36         * testsuite/libffi.call/fastthis3_win32.c: New test.
  37         * testsuite/libffi.call/strlen2_win32.c: New test.
  38         * testsuite/libffi.call/many2_win32.c: New test.
  39         * testsuite/libffi.call/struct1_win32.c: New test.
  40         * testsuite/libffi.call/struct2_win32.c: New test.
  41
  42 2012-01-23  Andreas Schwab  <schwab@linux-m68k.org>
  43
  44         * src/m68k/sysv.S (ffi_call_SYSV): Properly test for plain
  45         mc68000.  Test for __HAVE_68881__ in addition to __MC68881__.
  46
  47 2012-01-19  Jakub Jelinek  <jakub@redhat.com>
  48
  49         PR rtl-optimization/48496
  50         * src/ia64/ffi.c (ffi_call): Fix up aliasing violations.
  51
  52 2012-01-09  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
  53
  54         * configure.ac (i?86-*-*): Set TARGET to X86_64.
  55         * configure: Regenerate.
  56
  57 2011-12-07  Andrew Pinski  <apinski@cavium.com>
  58
  59         PR libffi/50051
  60         * src/mips/n32.S: Add ".set mips4".
  61
  62 2011-11-21  Andreas Tobler  <andreast@fgznet.ch>
  63
  64         * configure: Regenerate.
  65
  66 2011-11-10  Richard Henderson  <rth@redhat.com>
  67
  68         * configure.ac (GCC_AS_CFI_PSEUDO_OP): Use it instead of inline check.
  69         * configure, aclocal.m4: Rebuild.
  70
  71 2011-09-04  Iain Sandoe  <iains@gcc.gnu.org>
  72
  73         PR libffi/49594
  74         * src/powerpc/darwin_closure.S (stubs): Make the stub binding
  75         helper reference track the architecture pointer size.
  76
  77 2011-08-25  Andrew Haley  <aph@redhat.com>
  78
  79         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Remove hard-coded assembly
  80         instructions.
  81         * src/arm/sysv.S (ffi_arm_trampoline): Put them here instead.
  82
  83 2011-07-11  Andrew Haley  <aph@redhat.com>
  84
  85         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Clear icache.
  86
  87 2011-06-29  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
  88
  89         * testsuite/libffi.call/cls_double_va.c: Move PR number to comment.
  90         * testsuite/libffi.call/cls_longdouble_va.c: Likewise.
  91
  92 2011-06-29  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
  93
  94         PR libffi/46660
  95         * testsuite/libffi.call/cls_double_va.c: xfail dg-output on
  96         mips-sgi-irix6*.
  97         * testsuite/libffi.call/cls_longdouble_va.c: Likewise.
  98
  99 2011-06-14  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 100
 101         * testsuite/libffi.call/huge_struct.c (test_large_fn): Use PRIu8,
 102         PRId8 instead of %hhu, %hhd.
 103         * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRId8,
 104         PRIu8): Define.
 105         [__sgi__] (PRId8, PRIu8): Define.
 106
 107 2011-04-29  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 108
 109         * src/alpha/osf.S (UA_SI, FDE_ENCODING, FDE_ENCODE, FDE_ARANGE):
 110         Define.
 111         Use them to handle ELF vs. ECOFF differences.
 112         [__osf__] (_GLOBAL__F_ffi_call_osf): Define.
 113
 114 2011-02-13  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 115
 116         * configure: Regenerate.
 117
 118 2011-02-09  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 119
 120         PR libffi/46661
 121         * testsuite/libffi.call/cls_pointer.c (main): Cast void * to
 122         uintptr_t first.
 123         * testsuite/libffi.call/cls_pointer_stack.c (main): Likewise.
 124
 125 2011-02-07  Joel Sherrill <joel.sherrill@oarcorp.com>
 126
 127         * libffi/src/m68k/ffi.c: Add RTEMS support for cache flushing.
 128         Handle case when CPU variant does not have long double support.
 129         * libffi/src/m68k/sysv.S: Add support for mc68000, Coldfire,
 130         and cores with soft floating point.
 131
 132 2011-02-07  Joel Sherrill <joel.sherrill@oarcorp.com>
 133
 134         * configure.ac: Add mips*-*-rtems* support.
 135         * configure: Regenerate.
 136         * src/mips/ffitarget.h: Ensure needed constants are available
 137         for targets which do not have sgidefs.h.
 138
 139 2011-01-26  Dave Korn  <dave.korn.cygwin@gmail.com>
 140
 141         PR target/40125
 142         * configure.ac (AM_LTLDFLAGS): Add -bindir option for windows DLLs.
 143         * configure: Regenerate.
 144
 145 2010-12-18  Iain Sandoe  <iains@gcc.gnu.org>
 146
 147         PR libffi/29152
 148         PR libffi/42378
 149         * src/powerpc/darwin_closure.S: Provide Darwin64 implementation,
 150         update comments.
 151         * src/powerpc/ffitarget.h (POWERPC_DARWIN64): New,
 152         (FFI_TRAMPOLINE_SIZE): Update for Darwin64.
 153         * src/powerpc/darwin.S: Provide Darwin64 implementation,
 154         update comments.
 155         * src/powerpc/ffi_darwin.c: Likewise.
 156
 157 2010-12-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 158
 159         * configure.ac (libffi_cv_as_ascii_pseudo_op): Use double
 160         backslashes.
 161         (libffi_cv_as_string_pseudo_op): Likewise.
 162         * configure: Regenerate.
 163
 164 2010-12-03  Chung-Lin Tang  <cltang@codesourcery.com>
 165
 166         * src/arm/sysv.S (ffi_closure_SYSV): Add UNWIND to .pad directive.
 167         (ffi_closure_VFP): Same.
 168         (ffi_call_VFP): Move down to before ffi_closure_VFP. Add '.fpu vfp'
 169         directive.
 170
 171 2010-12-01  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 172
 173         * testsuite/libffi.call/ffitest.h [__sgi] (PRId64, PRIu64): Define.
 174         (PRIuPTR): Define.
 175
 176 2010-11-29  Richard Henderson  <rth@redhat.com>
 177             Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 178
 179         * src/x86/sysv.S (FDE_ENCODING, FDE_ENCODE): Define.
 180         (.eh_frame): Use FDE_ENCODING.
 181         (.LASFDE1, .LASFDE2, LASFDE3): Simplify with FDE_ENCODE.
 182
 183 2010-11-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 184
 185         * testsuite/lib/libffi-dg.exp: Rename ...
 186         * testsuite/lib/libffi.exp: ... to this.
 187         * libffi/testsuite/libffi.call/call.exp: Don't load libffi-dg.exp.
 188         * libffi/testsuite/libffi.special/special.exp: Likewise.
 189
 190 2010-10-28  Chung-Lin Tang  <cltang@codesourcery.com>
 191
 192         * src/arm/ffi.c (ffi_prep_args): Add VFP register argument handling
 193         code, new parameter, and return value. Update comments.
 194         (ffi_prep_cif_machdep): Add case for VFP struct return values. Add
 195         call to layout_vfp_args().
 196         (ffi_call_SYSV): Update declaration.
 197         (ffi_call_VFP): New declaration.
 198         (ffi_call): Add VFP struct return conditions. Call ffi_call_VFP()
 199         when ABI is FFI_VFP.
 200         (ffi_closure_VFP): New declaration.
 201         (ffi_closure_SYSV_inner): Add new vfp_args parameter, update call to
 202         ffi_prep_incoming_args_SYSV().
 203         (ffi_prep_incoming_args_SYSV): Update parameters. Add VFP argument
 204         case handling.
 205         (ffi_prep_closure_loc): Pass ffi_closure_VFP to trampoline
 206         construction under VFP hard-float.
 207         (rec_vfp_type_p): New function.
 208         (vfp_type_p): Same.
 209         (place_vfp_arg): Same.
 210         (layout_vfp_args): Same.
 211         * src/arm/ffitarget.h (ffi_abi): Add FFI_VFP. Define FFI_DEFAULT_ABI
 212         based on __ARM_PCS_VFP.
 213         (FFI_EXTRA_CIF_FIELDS): Define for adding VFP hard-float specific
 214         fields.
 215         (FFI_TYPE_STRUCT_VFP_FLOAT): Define internally used type code.
 216         (FFI_TYPE_STRUCT_VFP_DOUBLE): Same.
 217         * src/arm/sysv.S (ffi_call_SYSV): Change call of ffi_prep_args() to
 218         direct call. Move function pointer load upwards.
 219         (ffi_call_VFP): New function.
 220         (ffi_closure_VFP): Same.
 221
 222         * testsuite/lib/libffi-dg.exp (check-flags): New function.
 223         (dg-skip-if): New function.
 224         * testsuite/libffi.call/cls_double_va.c: Skip if target is arm*-*-*
 225         and compiler options include -mfloat-abi=hard.
 226         * testsuite/libffi.call/cls_longdouble_va.c: Same.
 227
 228 2010-10-01  Jakub Jelinek  <jakub@redhat.com>
 229
 230         PR libffi/45677
 231         * src/x86/ffi64.c (ffi_prep_cif_machdep): Ensure cif->bytes is
 232         a multiple of 8.
 233         * testsuite/libffi.call/many2.c: New test.
 234
 235 2010-08-20  Mark Wielaard  <mjw@redhat.com>
 236
 237         * src/closures.c (open_temp_exec_file_mnt): Check if getmntent_r
 238         returns NULL.
 239
 240 2010-08-09  Andreas Tobler  <andreast@fgznet.ch>
 241
 242         * configure.ac: Add target powerpc64-*-freebsd*.
 243         * configure: Regenerate.
 244         * testsuite/libffi.call/cls_align_longdouble_split.c: Pass
 245         -mlong-double-128 only to linux targets.
 246         * testsuite/libffi.call/cls_align_longdouble_split2.c: Likewise.
 247         * testsuite/libffi.call/cls_longdouble.c: Likewise.
 248         * testsuite/libffi.call/huge_struct.c: Likewise.
 249
 250 2010-07-10  Evan Phoenix  <evan@fallingsnow.net>
 251
 252         * src/closures.c (selinux_enabled_check): Fix strncmp usage bug.
 253
 254 2010-07-07  Dan Horák <dan@danny.cz>
 255
 256         * include/ffi.h.in: Protect #define with #ifndef.
 257         * src/powerpc/ffitarget.h: Ditto.
 258         * src/s390/ffitarget.h: Ditto.
 259         * src/sparc/ffitarget.h: Ditto.
 260
 261 2010-07-07   Neil Roberts <neil@linux.intel.com>
 262
 263         * src/x86/sysv.S (ffi_call_SYSV): Align the stack pointer to
 264         16-bytes.
 265
 266 2010-07-02  Jakub Jelinek  <jakub@redhat.com>
 267
 268         * Makefile.am (AM_MAKEFLAGS): Pass also mandir to submakes.
 269         * Makefile.in: Regenerated.
 270
 271 2010-05-19  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 272
 273         * configure.ac (libffi_cv_as_x86_pcrel): Check for illegal in as
 274         output, too.
 275         (libffi_cv_as_ascii_pseudo_op): Check for .ascii.
 276         (libffi_cv_as_string_pseudo_op): Check for .string.
 277         * configure: Regenerate.
 278         * fficonfig.h.in: Regenerate.
 279         * src/x86/sysv.S (.eh_frame): Use .ascii, .string or error.
 280
 281 2010-05-05  Michael Kohler <michaelkohler@live.com>
 282
 283         * src/dlmalloc.c (dlfree): Fix spelling.
 284         * src/ia64/ffi.c (ffi_prep_cif_machdep): Ditto.
 285         * configure.ac: Ditto.
 286         * configure: Rebuilt.
 287
 288 2010-04-13  Dan Witte  <dwitte@mozilla.com>
 289
 290         * msvcc.sh: Build with -W3 instead of -Wall.
 291         * src/powerpc/ffi_darwin.c: Remove build warnings.
 292         * src/x86/ffi.c: Ditto.
 293         * src/x86/ffitarget.h: Ditto.
 294
 295 2010-04-12  Dan Witte  <dwitte@mozilla.com>
 296             Walter Meinl <wuno@lsvw.de>
 297
 298         * configure.ac: Add OS/2 support.
 299         * configure: Rebuilt.
 300         * src/closures.c: Ditto.
 301         * src/dlmalloc.c: Ditto.
 302         * src/x86/win32.S: Ditto.
 303
 304 2010-04-07  Jakub Jelinek  <jakub@redhat.com>
 305
 306         * testsuite/libffi.call/err_bad_abi.c: Remove unused args variable.
 307
 308 2010-04-02  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 309
 310         * Makefile.in: Regenerate.
 311         * aclocal.m4: Regenerate.
 312         * include/Makefile.in: Regenerate.
 313         * man/Makefile.in: Regenerate.
 314         * testsuite/Makefile.in: Regenerate.
 315
 316 2010-03-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 317
 318         * configure.ac (libffi_cv_as_x86_64_unwind_section_type): New test.
 319         * configure: Regenerate.
 320         * fficonfig.h.in: Regenerate.
 321         * libffi/src/x86/unix64.S (.eh_frame)
 322         [HAVE_AS_X86_64_UNWIND_SECTION_TYPE]: Use @unwind section type.
 323
 324 2010-03-14  Matthias Klose  <doko@ubuntu.com>
 325
 326         * src/x86/ffi64.c: Fix typo in comment.
 327         * src/x86/ffi.c: Use /* ... */ comment style.
 328
 329 2010-02-24  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 330
 331         * doc/libffi.texi (The Closure API): Fix typo.
 332         * doc/libffi.info: Remove.
 333
 334 2010-02-15  Matthias Klose  <doko@ubuntu.com>
 335
 336         * src/arm/sysv.S (__ARM_ARCH__): Define for processor
 337         __ARM_ARCH_7EM__.
 338
 339 2010-01-15  Anthony Green  <green@redhat.com>
 340
 341         * README: Add notes on building with Microsoft Visual C++.
 342
 343 2010-01-15  Daniel Witte  <dwitte@mozilla.com>
 344
 345         * msvcc.sh: New file.
 346
 347         * src/x86/win32.S: Port assembly routines to MSVC and #ifdef.
 348         * src/x86/ffi.c: Tweak function declaration and remove excess
 349         parens.
 350         * include/ffi.h.in: Add __declspec(align(8)) to typedef struct
 351         ffi_closure.
 352
 353         * src/x86/ffi.c: Merge ffi_call_SYSV and ffi_call_STDCALL into new
 354         function ffi_call_win32 on X86_WIN32.
 355         * src/x86/win32.S (ffi_call_SYSV): Rename to ffi_call_win32.
 356         (ffi_call_STDCALL): Remove.
 357
 358         * src/prep_cif.c (ffi_prep_cif): Move stack space allocation code
 359         to ffi_prep_cif_machdep for x86.
 360         * src/x86/ffi.c (ffi_prep_cif_machdep): To here.
 361
 362 2010-01-15  Oliver Kiddle  <okiddle@yahoo.co.uk>
 363
 364         * src/x86/ffitarget.h (ffi_abi): Check for __i386 and __amd64 for
 365         Sun Studio compiler compatibility.
 366
 367 2010-01-12  Conrad Irwin <conrad.irwin@gmail.com>
 368
 369         * doc/libffi.texi: Add closure example.
 370
 371 2010-01-07  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 372
 373         PR libffi/40701
 374         * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRIdLL,
 375         PRIuLL, PRId64, PRIu64, PRIuPTR): Define.
 376         * testsuite/libffi.call/cls_align_sint64.c: Add -Wno-format on
 377         alpha*-dec-osf*.
 378         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
 379         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
 380         * testsuite/libffi.call/return_ll1.c: Likewise.
 381         * testsuite/libffi.call/stret_medium2.c: Likewise.
 382         * testsuite/libffi.special/ffitestcxx.h (allocate_mmap): Cast
 383         MAP_FAILED to char *.
 384
 385 2010-01-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 386
 387         * src/mips/n32.S: Use .abicalls and .eh_frame with __GNUC__.
 388
 389 2009-12-31  Anthony Green  <green@redhat.com>
 390
 391         * README: Update for libffi 3.0.9.
 392
 393 2009-12-27  Matthias Klose  <doko@ubuntu.com>
 394
 395         * configure.ac (HAVE_LONG_DOUBLE): Define for mips when
 396         appropriate.
 397         * configure: Rebuilt.
 398
 399 2009-12-26  Anthony Green  <green@redhat.com>
 400
 401         * testsuite/libffi.call/cls_longdouble_va.c: Mark as xfail for
 402         avr32*-*-*.
 403         * testsuite/libffi.call/cls_double_va.c: Ditto.
 404
 405 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
 406
 407         * testsuite/libffi.call/ffitest.h: Conditionally include stdint.h
 408         and inttypes.h.
 409         * testsuite/libffi.special/unwindtest.cc: Ditto.
 410
 411 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
 412
 413         * configure.ac: Add amd64-*-openbsd*.
 414         * configure: Rebuilt.
 415         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Link
 416         openbsd programs with -lpthread.
 417
 418 2009-12-26  Anthony Green  <green@redhat.com>
 419
 420         * testsuite/libffi.call/cls_double_va.c,
 421         testsuite/libffi.call/cls_longdouble.c,
 422         testsuite/libffi.call/cls_longdouble_va.c,
 423         testsuite/libffi.call/cls_pointer.c,
 424         testsuite/libffi.call/cls_pointer_stack.c: Remove xfail for
 425         mips*-*-* and arm*-*-*.
 426         * testsuite/libffi.call/cls_align_longdouble_split.c,
 427         testsuite/libffi.call/cls_align_longdouble_split2.c,
 428         testsuite/libffi.call/stret_medium2.c,
 429         testsuite/libffi.call/stret_medium.c,
 430         testsuite/libffi.call/stret_large.c,
 431         testsuite/libffi.call/stret_large2.c: Remove xfail for arm*-*-*.
 432
 433 2009-12-31  Kay Tietz  <ktietz70@googlemail.com>
 434
 435         * testsuite/libffi.call/ffitest.h,
 436         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRuLL): Fix
 437         definitions.
 438
 439 2009-12-31  Carlo Bramini  <carlo.bramix@libero.it>
 440
 441         * configure.ac (AM_LTLDFLAGS): Define for windows hosts.
 442         * Makefile.am (libffi_la_LDFLAGS): Add AM_LTLDFLAGS.
 443         * configure: Rebuilt.
 444         * Makefile.in: Rebuilt.
 445
 446 2009-12-31  Anthony Green  <green@redhat.com>
 447             Blake Chaffin.
 448
 449         * testsuite/libffi.call/huge_struct.c: New test case from Blake
 450         Chaffin @ Apple.
 451
 452 2009-12-28  David Edelsohn  <edelsohn@gnu.org>
 453
 454         * src/powerpc/ffi_darwin.c (ffi_prep_args): Copy abi and nargs to
 455         local variables.
 456         (aix_adjust_aggregate_sizes): New function.
 457         (ffi_prep_cif_machdep): Call it.
 458
 459 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
 460
 461         * configure.ac: Define FFI_MMAP_EXEC_WRIT for the given targets.
 462         * configure: Regenerate.
 463         * fficonfig.h.in: Likewise.
 464         * src/closures.c: Remove the FFI_MMAP_EXEC_WRIT definition for
 465         Solaris/x86.
 466
 467 2009-12-26  Andreas Schwab  <schwab@linux-m68k.org>
 468
 469         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Advance intarg_count
 470         when a float arguments is passed in memory.
 471         (ffi_closure_helper_SYSV): Mark general registers as used up when
 472         a 64bit or soft-float long double argument is passed in memory.
 473
 474 2009-12-25  Matthias Klose  <doko@ubuntu.com>
 475
 476         * man/ffi_call.3: Fix #include in examples.
 477         * doc/libffi.texi: Add dircategory.
 478
 479 2009-12-25  Frank Everdij <f.p.x.everdij@tudelft.nl>
 480
 481         * include/ffi.h.in: Placed '__GNUC__' ifdef around
 482         '__attribute__((aligned(8)))' in ffi_closure, fixes compile for
 483         IRIX MIPSPro c99.
 484         * include/ffi_common.h: Added '__sgi' define to non
 485         '__attribute__((__mode__()))' integer typedefs.
 486         * src/mips/ffi.c (ffi_call, ffi_closure_mips_inner_O32,
 487         ffi_closure_mips_inner_N32): Added 'defined(_MIPSEB)' to BE check.
 488         (ffi_closure_mips_inner_O32, ffi_closure_mips_inner_N32): Added
 489         FFI_LONGDOUBLE support and alignment(N32 only).
 490         * src/mips/ffitarget.h: Corrected '#include <sgidefs.h>' for IRIX and
 491         fixed non '__attribute__((__mode__()))' integer typedefs.
 492         * src/mips/n32.S: Put '#ifdef linux' around '.abicalls' and '.eh_frame'
 493         since they are Linux/GNU Assembler specific.
 494
 495 2009-12-25  Bradley Smith  <brad@brad-smith.co.uk>
 496
 497         * configure.ac, Makefile.am, src/avr32/ffi.c,
 498         src/avr32/ffitarget.h,
 499         src/avr32/sysv.S: Add AVR32 port.
 500         * configure, Makefile.in: Rebuilt.
 501
 502 2009-12-21  Andreas Tobler  <a.tobler@schweiz.org>
 503
 504         * configure.ac: Make i?86 build on FreeBSD and OpenBSD.
 505         * configure: Regenerate.
 506
 507 2009-12-15  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
 508
 509         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on PA HP-UX.
 510
 511 2009-12-13  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
 512
 513         * src/pa/ffi.c (ffi_closure_inner_pa32): Handle FFI_TYPE_LONGDOUBLE
 514         type on HP-UX.
 515
 516 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
 517
 518         * src/sparc/ffi.c (ffi_closure_sparc_inner_v9): Properly align 'long
 519         double' arguments.
 520
 521 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
 522
 523         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on Solaris < 10.
 524
 525 2009-12-10  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 526
 527         PR libffi/40700
 528         * src/closures.c [X86_64 && __sun__ && __svr4__]
 529         (FFI_MMAP_EXEC_WRIT): Define.
 530
 531 2009-12-08  David Daney  <ddaney@caviumnetworks.com>
 532
 533         * testsuite/libffi.call/stret_medium.c: Remove xfail for mips*-*-*
 534         * testsuite/libffi.call/cls_align_longdouble_split2.c: Same.
 535         * testsuite/libffi.call/stret_large.c: Same.
 536         * testsuite/libffi.call/cls_align_longdouble_split.c: Same.
 537         * testsuite/libffi.call/stret_large2.c: Same.
 538         * testsuite/libffi.call/stret_medium2.c: Same.
 539
 540 2009-12-07  David Edelsohn  <edelsohn@gnu.org>
 541
 542         * src/powerpc/aix_closure.S (libffi_closure_ASM): Fix tablejump
 543         typo.
 544
 545 2009-12-05  David Edelsohn  <edelsohn@gnu.org>
 546
 547         * src/powerpc/aix.S: Update AIX32 code to be consistent with AIX64
 548         code.
 549         * src/powerpc/aix_closure.S: Same.
 550
 551 2009-12-05  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 552
 553         * Makefile.in: Regenerate.
 554         * configure: Regenerate.
 555         * include/Makefile.in: Regenerate.
 556         * man/Makefile.in: Regenerate.
 557         * testsuite/Makefile.in: Regenerate.
 558
 559 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
 560
 561         * src/powerpc/aix_closure.S: Reorganize 64-bit code to match
 562         linux64_closure.S.
 563
 564 2009-12-04  Uros Bizjak  <ubizjak@gmail.com>
 565
 566         PR libffi/41908
 567         * src/x86/ffi64.c (classify_argument): Update from
 568         gcc/config/i386/i386.c.
 569         (ffi_closure_unix64_inner): Do not use the address of two consecutive
 570         SSE registers directly.
 571         * testsuite/libffi.call/cls_dbls_struct.c (main): Remove xfail
 572         for x86_64 linux targets.
 573
 574 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
 575
 576         * src/powerpc/ffi_darwin.c (ffi_closure_helper_DARWIN): Increment
 577         pfr for long double split between fpr13 and stack.
 578
 579 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
 580
 581         * src/powerpc/ffi_darwin.c (ffi_prep_args): Increment next_arg and
 582         fparg_count twice for long double.
 583
 584 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
 585
 586         PR libffi/42243
 587         * src/powerpc/ffi_darwin.c (ffi_prep_args): Remove extra parentheses.
 588
 589 2009-12-03  Uros Bizjak  <ubizjak@gmail.com>
 590
 591         * testsuite/libffi.call/cls_longdouble_va.c (main): Fix format string.
 592         Remove xfails for x86 linux targets.
 593
 594 2009-12-02  David Edelsohn  <edelsohn@gnu.org>
 595
 596         * src/powerpc/ffi_darwin.c (ffi_prep_args): Fix typo in INT64
 597         case.
 598
 599 2009-12-01  David Edelsohn  <edelsohn@gnu.org>
 600
 601         * src/powerpc/aix.S (ffi_call_AIX): Convert to more standard
 602         register usage.  Call ffi_prep_args directly.  Add long double
 603         return value support.
 604         * src/powerpc/ffi_darwin.c (ffi_prep_args): Double arg increment
 605         applies to FFI_TYPE_DOUBLE.  Correct fpr_base increment typo.
 606         Separate FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases.
 607         (ffi_prep_cif_machdep): Only 16 byte stack alignment in 64 bit
 608         mode.
 609         (ffi_closure_helper_DARWIN): Remove nf and ng counters.  Move temp
 610         into case.
 611         * src/powerpc/aix_closure.S: Maintain 16 byte stack alignment.
 612         Allocate result area between params and FPRs.
 613
 614 2009-11-30  David Edelsohn  <edelsohn@gnu.org>
 615
 616         PR target/35484
 617         * src/powerpc/ffitarget.h (POWERPC64): Define for PPC64 Linux and
 618         AIX64.
 619         * src/powerpc/aix.S: Implement AIX64 version.
 620         * src/powerpc/aix_closure.S: Implement AIX64 version.
 621         (ffi_closure_ASM): Use extsb, lha and displament addresses.
 622         * src/powerpc/ffi_darwin.c (ffi_prep_args): Implement AIX64
 623         support.
 624         (ffi_prep_cif_machdep): Same.
 625         (ffi_call): Same.
 626         (ffi_closure_helper_DARWIN): Same.
 627
 628 2009-11-02  Andreas Tobler  <a.tobler@schweiz.org>
 629
 630         PR libffi/41908
 631         * testsuite/libffi.call/testclosure.c: New test.
 632
 633 2009-09-28  Kai Tietz  <kai.tietz@onevision.com>
 634
 635         * src/x86/win64.S (_ffi_call_win64 stack): Remove for gnu
 636         assembly version use of ___chkstk.
 637
 638 2009-09-23  Matthias Klose  <doko@ubuntu.com>
 639
 640         PR libffi/40242, PR libffi/41443
 641         * src/arm/sysv.S (__ARM_ARCH__): Define for processors
 642         __ARM_ARCH_6T2__, __ARM_ARCH_6M__, __ARM_ARCH_7__,
 643         __ARM_ARCH_7A__, __ARM_ARCH_7R__, __ARM_ARCH_7M__.
 644         Change the conditionals to __SOFTFP__ || __ARM_EABI__
 645         for -mfloat-abi=softfp to work.
 646
 647 2009-09-17  Loren J. Rittle  <ljrittle@acm.org>
 648
 649         PR testsuite/32843 (strikes again)
 650         * src/x86/ffi.c (ffi_prep_cif_machdep): Add X86_FREEBSD to
 651         enable proper extension on char and short.
 652
 653 2009-09-15  David Daney  <ddaney@caviumnetworks.com>
 654
 655         * src/java_raw_api.c (ffi_java_raw_to_rvalue): Remove special
 656         handling for FFI_TYPE_POINTER.
 657         * src/mips/ffitarget.h (FFI_TYPE_STRUCT_D_SOFT,
 658         FFI_TYPE_STRUCT_F_SOFT, FFI_TYPE_STRUCT_DD_SOFT,
 659         FFI_TYPE_STRUCT_FF_SOFT, FFI_TYPE_STRUCT_FD_SOFT,
 660         FFI_TYPE_STRUCT_DF_SOFT, FFI_TYPE_STRUCT_SOFT): New defines.
 661         (FFI_N32_SOFT_FLOAT, FFI_N64_SOFT_FLOAT): New ffi_abi enumerations.
 662         (enum ffi_abi): Set FFI_DEFAULT_ABI for soft-float.
 663         * src/mips/n32.S (ffi_call_N32): Add handling for soft-float
 664         structure and pointer returns.
 665         (ffi_closure_N32): Add handling for pointer returns.
 666         * src/mips/ffi.c (ffi_prep_args, calc_n32_struct_flags,
 667         calc_n32_return_struct_flags): Handle soft-float.
 668         (ffi_prep_cif_machdep):  Handle soft-float, fix pointer handling.
 669         (ffi_call_N32): Declare proper argument types.
 670         (ffi_call, copy_struct_N32, ffi_closure_mips_inner_N32): Handle
 671         soft-float.
 672
 673 2009-08-24  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 674
 675         * configure.ac (AC_PREREQ): Bump to 2.64.
 676
 677 2009-08-22  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 678
 679         * Makefile.am (install-html, install-pdf): Remove.
 680         * Makefile.in: Regenerate.
 681
 682         * Makefile.in: Regenerate.
 683         * aclocal.m4: Regenerate.
 684         * configure: Regenerate.
 685         * fficonfig.h.in: Regenerate.
 686         * include/Makefile.in: Regenerate.
 687         * man/Makefile.in: Regenerate.
 688         * testsuite/Makefile.in: Regenerate.
 689
 690 2009-07-30  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 691
 692         * configure.ac (_AC_ARG_VAR_PRECIOUS): Use m4_rename_force.
 693
 694 2009-07-24  Dave Korn  <dave.korn.cygwin@gmail.com>
 695
 696         PR libffi/40807
 697         * src/x86/ffi.c (ffi_prep_cif_machdep): Also use sign/zero-extending
 698         return types for X86_WIN32.
 699         * src/x86/win32.S (_ffi_call_SYSV): Handle omitted return types.
 700         (_ffi_call_STDCALL, _ffi_closure_SYSV, _ffi_closure_raw_SYSV,
 701         _ffi_closure_STDCALL): Likewise.
 702
 703         * src/closures.c (is_selinux_enabled): Define to const 0 for Cygwin.
 704         (dlmmap, dlmunmap): Also use these functions on Cygwin.
 705
 706 2009-07-11  Richard Sandiford  <rdsandiford@googlemail.com>
 707
 708         PR testsuite/40699
 709         PR testsuite/40707
 710         PR testsuite/40709
 711         * testsuite/lib/libffi-dg.exp: Revert 2009-07-02, 2009-07-01 and
 712         2009-06-30 commits.
 713
 714 2009-07-01  Richard Sandiford  <r.sandiford@uk.ibm.com>
 715
 716         * testsuite/lib/libffi-dg.exp (libffi-init): Set ld_library_path
 717         to "" before adding paths.  (This reinstates an assignment that
 718         was removed by my 2009-06-30 commit, but changes the initial
 719         value from "." to "".)
 720
 721 2009-07-01  H.J. Lu  <hongjiu.lu@intel.com>
 722
 723         PR testsuite/40601
 724         * testsuite/lib/libffi-dg.exp (libffi-init): Properly set
 725         gccdir.  Adjust ld_library_path for gcc only if gccdir isn't
 726         empty.
 727
 728 2009-06-30  Richard Sandiford  <r.sandiford@uk.ibm.com>
 729
 730         * testsuite/lib/libffi-dg.exp (libffi-init): Don't add "."
 731         to ld_library_path.  Use add_path.  Add just find_libgcc_s
 732         to ld_library_path, not every libgcc multilib directory.
 733
 734 2009-06-16  Wim Lewis  <wiml@hhhh.org>
 735
 736         * src/powerpc/ffi.c: Avoid clobbering cr3 and cr4, which are
 737         supposed to be callee-saved.
 738         * src/powerpc/sysv.S (small_struct_return_value): Fix overrun of
 739         return buffer for odd-size structs.
 740
 741 2009-06-16  Andreas Tobler  <a.tobler@schweiz.org>
 742
 743         PR libffi/40444
 744         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Add
 745         allow_stack_execute for Darwin.
 746
 747 2009-06-16  Andrew Haley  <aph@redhat.com>
 748
 749         * configure.ac (TARGETDIR): Add missing blank lines.
 750         * configure: Regenerate.
 751
 752 2009-06-16  Andrew Haley  <aph@redhat.com>
 753
 754         * testsuite/libffi.call/cls_align_sint64.c,
 755         testsuite/libffi.call/cls_align_uint64.c,
 756         testsuite/libffi.call/cls_longdouble_va.c,
 757         testsuite/libffi.call/cls_ulonglong.c,
 758         testsuite/libffi.call/return_ll1.c,
 759         testsuite/libffi.call/stret_medium2.c: Fix printf format
 760         specifiers.
 761         * testsuite/libffi.call/ffitest.h,
 762         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRIuLL): Define.
 763
 764 2009-06-15  Andrew Haley  <aph@redhat.com>
 765
 766         * testsuite/libffi.call/err_bad_typedef.c: xfail everywhere.
 767         * testsuite/libffi.call/err_bad_abi.c: Likewise.
 768
 769 2009-06-12  Andrew Haley  <aph@redhat.com>
 770
 771         * Makefile.am: Remove info_TEXINFOS.
 772
 773 2009-06-12  Andrew Haley  <aph@redhat.com>
 774
 775         * ChangeLog.libffi: testsuite/libffi.call/cls_align_sint64.c,
 776         testsuite/libffi.call/cls_align_uint64.c,
 777         testsuite/libffi.call/cls_ulonglong.c,
 778         testsuite/libffi.call/return_ll1.c,
 779         testsuite/libffi.call/stret_medium2.c: Fix printf format
 780         specifiers.
 781         testsuite/libffi.special/unwindtest.cc: include stdint.h.
 782
 783 2009-06-11  Timothy Wall  <twall@users.sf.net>
 784
 785         * Makefile.am,
 786         configure.ac,
 787         include/ffi.h.in,
 788         include/ffi_common.h,
 789         src/closures.c,
 790         src/dlmalloc.c,
 791         src/x86/ffi.c,
 792         src/x86/ffitarget.h,
 793         src/x86/win64.S (new),
 794         README: Added win64 support (mingw or MSVC)
 795         * Makefile.in,
 796         include/Makefile.in,
 797         man/Makefile.in,
 798         testsuite/Makefile.in,
 799         configure,
 800         aclocal.m4: Regenerated
 801         * ltcf-c.sh: properly escape cygwin/w32 path
 802         * man/ffi_call.3: Clarify size requirements for return value.
 803         * src/x86/ffi64.c: Fix filename in comment.
 804         * src/x86/win32.S: Remove unused extern.
 805
 806         * testsuite/libffi.call/closure_fn0.c,
 807         testsuite/libffi.call/closure_fn1.c,
 808         testsuite/libffi.call/closure_fn2.c,
 809         testsuite/libffi.call/closure_fn3.c,
 810         testsuite/libffi.call/closure_fn4.c,
 811         testsuite/libffi.call/closure_fn5.c,
 812         testsuite/libffi.call/closure_fn6.c,
 813         testsuite/libffi.call/closure_stdcall.c,
 814         testsuite/libffi.call/cls_12byte.c,
 815         testsuite/libffi.call/cls_16byte.c,
 816         testsuite/libffi.call/cls_18byte.c,
 817         testsuite/libffi.call/cls_19byte.c,
 818         testsuite/libffi.call/cls_1_1byte.c,
 819         testsuite/libffi.call/cls_20byte.c,
 820         testsuite/libffi.call/cls_20byte1.c,
 821         testsuite/libffi.call/cls_24byte.c,
 822         testsuite/libffi.call/cls_2byte.c,
 823         testsuite/libffi.call/cls_3_1byte.c,
 824         testsuite/libffi.call/cls_3byte1.c,
 825         testsuite/libffi.call/cls_3byte2.c,
 826         testsuite/libffi.call/cls_4_1byte.c,
 827         testsuite/libffi.call/cls_4byte.c,
 828         testsuite/libffi.call/cls_5_1_byte.c,
 829         testsuite/libffi.call/cls_5byte.c,
 830         testsuite/libffi.call/cls_64byte.c,
 831         testsuite/libffi.call/cls_6_1_byte.c,
 832         testsuite/libffi.call/cls_6byte.c,
 833         testsuite/libffi.call/cls_7_1_byte.c,
 834         testsuite/libffi.call/cls_7byte.c,
 835         testsuite/libffi.call/cls_8byte.c,
 836         testsuite/libffi.call/cls_9byte1.c,
 837         testsuite/libffi.call/cls_9byte2.c,
 838         testsuite/libffi.call/cls_align_double.c,
 839         testsuite/libffi.call/cls_align_float.c,
 840         testsuite/libffi.call/cls_align_longdouble.c,
 841         testsuite/libffi.call/cls_align_longdouble_split.c,
 842         testsuite/libffi.call/cls_align_longdouble_split2.c,
 843         testsuite/libffi.call/cls_align_pointer.c,
 844         testsuite/libffi.call/cls_align_sint16.c,
 845         testsuite/libffi.call/cls_align_sint32.c,
 846         testsuite/libffi.call/cls_align_sint64.c,
 847         testsuite/libffi.call/cls_align_uint16.c,
 848         testsuite/libffi.call/cls_align_uint32.c,
 849         testsuite/libffi.call/cls_align_uint64.c,
 850         testsuite/libffi.call/cls_dbls_struct.c,
 851         testsuite/libffi.call/cls_double.c,
 852         testsuite/libffi.call/cls_double_va.c,
 853         testsuite/libffi.call/cls_float.c,
 854         testsuite/libffi.call/cls_longdouble.c,
 855         testsuite/libffi.call/cls_longdouble_va.c,
 856         testsuite/libffi.call/cls_multi_schar.c,
 857         testsuite/libffi.call/cls_multi_sshort.c,
 858         testsuite/libffi.call/cls_multi_sshortchar.c,
 859         testsuite/libffi.call/cls_multi_uchar.c,
 860         testsuite/libffi.call/cls_multi_ushort.c,
 861         testsuite/libffi.call/cls_multi_ushortchar.c,
 862         testsuite/libffi.call/cls_pointer.c,
 863         testsuite/libffi.call/cls_pointer_stack.c,
 864         testsuite/libffi.call/cls_schar.c,
 865         testsuite/libffi.call/cls_sint.c,
 866         testsuite/libffi.call/cls_sshort.c,
 867         testsuite/libffi.call/cls_uchar.c,
 868         testsuite/libffi.call/cls_uint.c,
 869         testsuite/libffi.call/cls_ulonglong.c,
 870         testsuite/libffi.call/cls_ushort.c,
 871         testsuite/libffi.call/err_bad_abi.c,
 872         testsuite/libffi.call/err_bad_typedef.c,
 873         testsuite/libffi.call/float2.c,
 874         testsuite/libffi.call/huge_struct.c,
 875         testsuite/libffi.call/nested_struct.c,
 876         testsuite/libffi.call/nested_struct1.c,
 877         testsuite/libffi.call/nested_struct10.c,
 878         testsuite/libffi.call/nested_struct2.c,
 879         testsuite/libffi.call/nested_struct3.c,
 880         testsuite/libffi.call/nested_struct4.c,
 881         testsuite/libffi.call/nested_struct5.c,
 882         testsuite/libffi.call/nested_struct6.c,
 883         testsuite/libffi.call/nested_struct7.c,
 884         testsuite/libffi.call/nested_struct8.c,
 885         testsuite/libffi.call/nested_struct9.c,
 886         testsuite/libffi.call/problem1.c,
 887         testsuite/libffi.call/return_ldl.c,
 888         testsuite/libffi.call/return_ll1.c,
 889         testsuite/libffi.call/stret_large.c,
 890         testsuite/libffi.call/stret_large2.c,
 891         testsuite/libffi.call/stret_medium.c,
 892         testsuite/libffi.call/stret_medium2.c,
 893         testsuite/libffi.special/unwindtest.cc: use ffi_closure_alloc instead
 894         of checking for MMAP.  Use intptr_t instead of long casts.
 895
 896 2009-06-11  Kaz Kojima  <kkojima@gcc.gnu.org>
 897
 898         * testsuite/libffi.call/cls_longdouble_va.c: Add xfail sh*-*-linux-*.
 899         * testsuite/libffi.call/err_bad_abi.c: Add xfail sh*-*-*.
 900         * testsuite/libffi.call/err_bad_typedef.c: Likewise.
 901
 902 2009-06-09  Andrew Haley  <aph@redhat.com>
 903
 904         * src/x86/freebsd.S: Add missing file.
 905
 906 2009-06-08  Andrew Haley  <aph@redhat.com>
 907
 908         Import from libffi 3.0.8:
 909
 910         * doc/libffi.texi: New file.
 911         * doc/libffi.info: Likewise.
 912         * doc/stamp-vti: Likewise.
 913         * man/Makefile.am: New file.
 914         * man/ffi_call.3: New file.
 915
 916         * Makefile.am (EXTRA_DIST): Add src/x86/darwin64.S,
 917         src/dlmalloc.c.
 918         (nodist_libffi_la_SOURCES): Add X86_FREEBSD.
 919
 920         * configure.ac: Bump version to 3.0.8.
 921         parisc*-*-linux*: Add.
 922         i386-*-freebsd* | i386-*-openbsd*: Add.
 923         powerpc-*-beos*: Add.
 924         AM_CONDITIONAL X86_FREEBSD: Add.
 925         AC_CONFIG_FILES: Add man/Makefile.
 926
 927         * include/ffi.h.in (FFI_FN): Change void (*)() to void (*)(void).
 928
 929 2009-06-08  Andrew Haley  <aph@redhat.com>
 930
 931         * README: Import from libffi 3.0.8.
 932
 933 2009-06-08  Andrew Haley  <aph@redhat.com>
 934
 935         * testsuite/libffi.call/err_bad_abi.c: Add xfails.
 936         * testsuite/libffi.call/cls_longdouble_va.c: Add xfails.
 937         * testsuite/libffi.call/cls_dbls_struct.c: Add xfail x86_64-*-linux-*.
 938         * testsuite/libffi.call/err_bad_typedef.c: Add xfails.
 939
 940         * testsuite/libffi.call/stret_medium2.c: Add __UNUSED__ to args.
 941         * testsuite/libffi.call/stret_medium.c: Likewise.
 942         * testsuite/libffi.call/stret_large2.c: Likewise.
 943         * testsuite/libffi.call/stret_large.c:  Likewise.
 944
 945 2008-12-26  Timothy Wall  <twall@users.sf.net>
 946
 947         * testsuite/libffi.call/cls_longdouble.c,
 948         testsuite/libffi.call/cls_longdouble_va.c,
 949         testsuite/libffi.call/cls_align_longdouble.c,
 950         testsuite/libffi.call/cls_align_longdouble_split.c,
 951         testsuite/libffi.call/cls_align_longdouble_split2.c: mark expected
 952         failures on x86_64 cygwin/mingw.
 953
 954 2008-12-22  Timothy Wall  <twall@users.sf.net>
 955
 956         * testsuite/libffi.call/closure_fn0.c,
 957         testsuite/libffi.call/closure_fn1.c,
 958         testsuite/libffi.call/closure_fn2.c,
 959         testsuite/libffi.call/closure_fn3.c,
 960         testsuite/libffi.call/closure_fn4.c,
 961         testsuite/libffi.call/closure_fn5.c,
 962         testsuite/libffi.call/closure_fn6.c,
 963         testsuite/libffi.call/closure_loc_fn0.c,
 964         testsuite/libffi.call/closure_stdcall.c,
 965         testsuite/libffi.call/cls_align_pointer.c,
 966         testsuite/libffi.call/cls_pointer.c,
 967         testsuite/libffi.call/cls_pointer_stack.c: use portable cast from
 968         pointer to integer (intptr_t).
 969         * testsuite/libffi.call/cls_longdouble.c: disable for win64.
 970
 971 2008-07-24  Anthony Green  <green@redhat.com>
 972
 973         * testsuite/libffi.call/cls_dbls_struct.c,
 974         testsuite/libffi.call/cls_double_va.c,
 975         testsuite/libffi.call/cls_longdouble.c,
 976         testsuite/libffi.call/cls_longdouble_va.c,
 977         testsuite/libffi.call/cls_pointer.c,
 978         testsuite/libffi.call/cls_pointer_stack.c,
 979         testsuite/libffi.call/err_bad_abi.c: Clean up failures from
 980         compiler warnings.
 981
 982 2008-03-04  Anthony Green  <green@redhat.com>
 983             Blake Chaffin
 984             hos@tamanegi.org
 985
 986         * testsuite/libffi.call/cls_align_longdouble_split2.c
 987           testsuite/libffi.call/cls_align_longdouble_split.c
 988           testsuite/libffi.call/cls_dbls_struct.c
 989           testsuite/libffi.call/cls_double_va.c
 990           testsuite/libffi.call/cls_longdouble.c
 991           testsuite/libffi.call/cls_longdouble_va.c
 992           testsuite/libffi.call/cls_pointer.c
 993           testsuite/libffi.call/cls_pointer_stack.c
 994           testsuite/libffi.call/err_bad_abi.c
 995           testsuite/libffi.call/err_bad_typedef.c
 996           testsuite/libffi.call/stret_large2.c
 997           testsuite/libffi.call/stret_large.c
 998           testsuite/libffi.call/stret_medium2.c
 999           testsuite/libffi.call/stret_medium.c: New tests from Apple.
1000
1001 2009-06-05  Andrew Haley  <aph@redhat.com>
1002
1003         * src/x86/ffitarget.h, src/x86/ffi.c: Merge stdcall changes from
1004         libffi.
1005
1006 2009-06-04  Andrew Haley  <aph@redhat.com>
1007
1008         * src/x86/ffitarget.h, src/x86/win32.S, src/x86/ffi.c: Back out
1009         stdcall changes.
1010
1011 2008-02-26  Anthony Green  <green@redhat.com>
1012             Thomas Heller  <theller@ctypes.org>
1013
1014         * src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C
1015         comment.
1016
1017 2008-02-03  Timothy Wall  <twall@users.sf.net>
1018
1019         * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return
1020           offset based on code pointer, not data pointer.
1021
1022 2008-01-31  Timothy Wall <twall@users.sf.net>
1023
1024         * testsuite/libffi.call/closure_stdcall.c: Add test for stdcall
1025         closures.
1026         * src/x86/ffitarget.h: Increase size of trampoline for stdcall
1027         closures.
1028         * src/x86/win32.S: Add assembly for stdcall closure.
1029         * src/x86/ffi.c: Initialize stdcall closure trampoline.
1030
1031 2009-06-04  Andrew Haley  <aph@redhat.com>
1032
1033         * include/ffi.h.in: Change void (*)() to void (*)(void).
1034         * src/x86/ffi.c: Likewise.
1035
1036 2009-06-04  Andrew Haley  <aph@redhat.com>
1037
1038         * src/powerpc/ppc_closure.S: Insert licence header.
1039         * src/powerpc/linux64_closure.S: Likewise.
1040         * src/m68k/sysv.S: Likewise.
1041
1042         * src/sh64/ffi.c: Change void (*)() to void (*)(void).
1043         * src/powerpc/ffi.c: Likewise.
1044         * src/powerpc/ffi_darwin.c: Likewise.
1045         * src/m32r/ffi.c: Likewise.
1046         * src/sh64/ffi.c: Likewise.
1047         * src/x86/ffi64.c: Likewise.
1048         * src/alpha/ffi.c: Likewise.
1049         * src/alpha/osf.S: Likewise.
1050         * src/frv/ffi.c: Likewise.
1051         * src/s390/ffi.c: Likewise.
1052         * src/pa/ffi.c: Likewise.
1053         * src/pa/hpux32.S: Likewise.
1054         * src/ia64/unix.S: Likewise.
1055         * src/ia64/ffi.c: Likewise.
1056         * src/sparc/ffi.c: Likewise.
1057         * src/mips/ffi.c: Likewise.
1058         * src/sh/ffi.c: Likewise.
1059
1060 2008-02-15  David Daney  <ddaney@avtrex.com>
1061
1062         * src/mips/ffi.c (USE__BUILTIN___CLEAR_CACHE):
1063         Define (conditionally), and use it to include cachectl.h.
1064         (ffi_prep_closure_loc): Fix cache flushing.
1065         * src/mips/ffitarget.h (_ABIN32, _ABI64, _ABIO32): Define.
1066
1067 2009-06-04  Andrew Haley  <aph@redhat.com>
1068
1069         include/ffi.h.in,
1070         src/arm/ffitarget.h,
1071         src/arm/ffi.c,
1072         src/arm/sysv.S,
1073         src/powerpc/ffitarget.h,
1074         src/closures.c,
1075         src/sh64/ffitarget.h,
1076         src/sh64/ffi.c,
1077         src/sh64/sysv.S,
1078         src/types.c,
1079         src/x86/ffi64.c,
1080         src/x86/ffitarget.h,
1081         src/x86/win32.S,
1082         src/x86/darwin.S,
1083         src/x86/ffi.c,
1084         src/x86/sysv.S,
1085         src/x86/unix64.S,
1086         src/alpha/ffitarget.h,
1087         src/alpha/ffi.c,
1088         src/alpha/osf.S,
1089         src/m68k/ffitarget.h,
1090         src/frv/ffitarget.h,
1091         src/frv/ffi.c,
1092         src/s390/ffitarget.h,
1093         src/s390/sysv.S,
1094         src/cris/ffitarget.h,
1095         src/pa/linux.S,
1096         src/pa/ffitarget.h,
1097         src/pa/ffi.c,
1098         src/raw_api.c,
1099         src/ia64/ffitarget.h,
1100         src/ia64/unix.S,
1101         src/ia64/ffi.c,
1102         src/ia64/ia64_flags.h,
1103         src/java_raw_api.c,
1104         src/debug.c,
1105         src/sparc/v9.S,
1106         src/sparc/ffitarget.h,
1107         src/sparc/ffi.c,
1108         src/sparc/v8.S,
1109         src/mips/ffitarget.h,
1110         src/mips/n32.S,
1111         src/mips/o32.S,
1112         src/mips/ffi.c,
1113         src/prep_cif.c,
1114         src/sh/ffitarget.h,
1115         src/sh/ffi.c,
1116         src/sh/sysv.S: Update license text.
1117
1118 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
1119
1120         * src/x86/win32.S (_ffi_closure_STDCALL):  New function.
1121         (.eh_frame):  Add FDE for it.
1122
1123 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
1124
1125         * configure.ac:  Also check if assembler supports pc-relative
1126         relocs on X86_WIN32 targets.
1127         * configure:  Regenerate.
1128         * src/x86/win32.S (ffi_prep_args):  Declare extern, not global.
1129         (_ffi_call_SYSV):  Add missing function type symbol .def and
1130         add EH markup labels.
1131         (_ffi_call_STDCALL):  Likewise.
1132         (_ffi_closure_SYSV):  Likewise.
1133         (_ffi_closure_raw_SYSV):  Likewise.
1134         (.eh_frame):  Add hand-crafted EH data.
1135
1136 2009-04-09  Jakub Jelinek  <jakub@redhat.com>
1137
1138         * testsuite/lib/libffi-dg.exp: Change copyright header to refer to
1139         version 3 of the GNU General Public License and to point readers
1140         at the COPYING3 file and the FSF's license web page.
1141         * testsuite/libffi.call/call.exp: Likewise.
1142         * testsuite/libffi.special/special.exp: Likewise.
1143
1144 2009-03-01  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1145
1146         * configure: Regenerate.
1147
1148 2008-12-18  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1149
1150         PR libffi/26048
1151         * configure.ac (HAVE_AS_X86_PCREL): New test.
1152         * configure: Regenerate.
1153         * fficonfig.h.in: Regenerate.
1154         * src/x86/sysv.S [!FFI_NO_RAW_API]: Precalculate
1155         RAW_CLOSURE_CIF_OFFSET, RAW_CLOSURE_FUN_OFFSET,
1156         RAW_CLOSURE_USER_DATA_OFFSET for the Solaris 10/x86 assembler.
1157         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
1158         * src/x86/unix64.S (.Lstore_table): Move to .text section.
1159         (.Lload_table): Likewise.
1160         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
1161
1162 2008-12-18  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1163
1164         * configure: Regenerate.
1165
1166 2008-11-21  Eric Botcazou  <ebotcazou@adacore.com>
1167
1168         * src/sparc/ffi.c (ffi_prep_cif_machdep): Add support for
1169         signed/unsigned int8/16 return values.
1170         * src/sparc/v8.S (ffi_call_v8): Likewise.
1171         (ffi_closure_v8): Likewise.
1172
1173 2008-09-26  Peter O'Gorman  <pogma@thewrittenword.com>
1174             Steve Ellcey  <sje@cup.hp.com>
1175
1176         * configure: Regenerate for new libtool.
1177         * Makefile.in: Ditto.
1178         * include/Makefile.in: Ditto.
1179         * aclocal.m4: Ditto.
1180
1181 2008-08-25  Andreas Tobler  <a.tobler@schweiz.org>
1182
1183         * src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and
1184         FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum.
1185         Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT.
1186         Adjust copyright notice.
1187         * src/powerpc/ffi.c: Add two new flags to indicate if we have one
1188         register or two register to use for FFI_SYSV structs.
1189         (ffi_prep_cif_machdep): Pass the right register flag introduced above.
1190         (ffi_closure_helper_SYSV): Fix the return type for
1191         FFI_SYSV_TYPE_SMALL_STRUCT. Comment.
1192         Adjust copyright notice.
1193
1194 2008-07-16  Kaz Kojima  <kkojima@gcc.gnu.org>
1195
1196         * src/sh/ffi.c (ffi_prep_closure_loc): Turn INSN into an unsigned
1197         int.
1198
1199 2008-06-17  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1200
1201         * configure: Regenerate.
1202         * include/Makefile.in: Regenerate.
1203         * testsuite/Makefile.in: Regenerate.
1204
1205 2008-06-07  Joseph Myers  <joseph@codesourcery.com>
1206
1207         * configure.ac (parisc*-*-linux*, powerpc-*-sysv*,
1208         powerpc-*-beos*): Remove.
1209         * configure: Regenerate.
1210
1211 2008-05-09  Julian Brown  <julian@codesourcery.com>
1212
1213         * Makefile.am (LTLDFLAGS): New.
1214         (libffi_la_LDFLAGS): Use above.
1215         * Makefile.in: Regenerate.
1216
1217 2008-04-18  Paolo Bonzini  <bonzini@gnu.org>
1218
1219         PR bootstrap/35457
1220         * aclocal.m4: Regenerate.
1221         * configure: Regenerate.
1222
1223 2008-03-26  Kaz Kojima  <kkojima@gcc.gnu.org>
1224
1225         * src/sh/sysv.S: Add .note.GNU-stack on Linux.
1226         * src/sh64/sysv.S: Likewise.
1227
1228 2008-03-26  Daniel Jacobowitz  <dan@debian.org>
1229
1230         * src/arm/sysv.S: Fix ARM comment marker.
1231
1232 2008-03-26  Jakub Jelinek  <jakub@redhat.com>
1233
1234         * src/alpha/osf.S: Add .note.GNU-stack on Linux.
1235         * src/s390/sysv.S: Likewise.
1236         * src/powerpc/ppc_closure.S: Likewise.
1237         * src/powerpc/sysv.S: Likewise.
1238         * src/x86/unix64.S: Likewise.
1239         * src/x86/sysv.S: Likewise.
1240         * src/sparc/v8.S: Likewise.
1241         * src/sparc/v9.S: Likewise.
1242         * src/m68k/sysv.S: Likewise.
1243         * src/arm/sysv.S: Likewise.
1244
1245 2008-03-16  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1246
1247         * aclocal.m4: Regenerate.
1248         * configure: Likewise.
1249         * Makefile.in: Likewise.
1250         * include/Makefile.in: Likewise.
1251         * testsuite/Makefile.in: Likewise.
1252
1253 2008-02-12  Bjoern Koenig  <bkoenig@alpha-tierchen.de>
1254             Andreas Tobler  <a.tobler@schweiz.org>
1255
1256         * configure.ac: Add amd64-*-freebsd* target.
1257         * configure: Regenerate.
1258
1259 2008-01-30  H.J. Lu  <hongjiu.lu@intel.com>
1260
1261         PR libffi/34612
1262         * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when
1263         returning struct.
1264
1265         * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer"
1266         tests.
1267
1268 2008-01-24  David Edelsohn  <edelsohn@gnu.org>
1269
1270         * configure: Regenerate.
1271
1272 2008-01-06  Andreas Tobler  <a.tobler@schweiz.org>
1273
1274         * src/x86/ffi.c (ffi_prep_cif_machdep): Fix thinko.
1275
1276 2008-01-05  Andreas Tobler  <a.tobler@schweiz.org>
1277
1278         PR testsuite/32843
1279         * src/x86/ffi.c (ffi_prep_cif_machdep): Add code for
1280         signed/unsigned int8/16 for X86_DARWIN.
1281         Updated copyright info.
1282         Handle one and two byte structs with special cif->flags.
1283         * src/x86/ffitarget.h: Add special types for one and two byte structs.
1284         Updated copyright info.
1285         * src/x86/darwin.S (ffi_call_SYSV): Rewrite to use a jump table like
1286         sysv.S
1287         Remove code to pop args from the stack after call.
1288         Special-case signed/unsigned for int8/16, one and two byte structs.
1289         (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1290         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1291         FFI_TYPE_SINT32.
1292         Updated copyright info.
1293
1294 2007-12-08  David Daney  <ddaney@avtrex.com>
1295
1296         * src/mips/n32.S (ffi_call_N32):  Replace dadd with ADDU, dsub with
1297         SUBU, add with ADDU and use smaller code sequences.
1298
1299 2007-12-07  David Daney  <ddaney@avtrex.com>
1300
1301         * src/mips/ffi.c (ffi_prep_cif_machdep): Handle long double return
1302         type.
1303
1304 2007-12-06  David Daney  <ddaney@avtrex.com>
1305
1306         * include/ffi.h.in (FFI_SIZEOF_JAVA_RAW): Define if not already
1307         defined.
1308         (ffi_java_raw): New typedef.
1309         (ffi_java_raw_call, ffi_java_ptrarray_to_raw,
1310         ffi_java_raw_to_ptrarray): Change parameter types from ffi_raw to
1311         ffi_java_raw.
1312         (ffi_java_raw_closure) : Same.
1313         (ffi_prep_java_raw_closure, ffi_prep_java_raw_closure_loc): Change
1314         parameter types.
1315         * src/java_raw_api.c (ffi_java_raw_size):  Replace FFI_SIZEOF_ARG with
1316         FFI_SIZEOF_JAVA_RAW.
1317         (ffi_java_raw_to_ptrarray): Change type of raw to ffi_java_raw.
1318         Replace FFI_SIZEOF_ARG with FFI_SIZEOF_JAVA_RAW. Use
1319         sizeof(ffi_java_raw) for alignment calculations.
1320         (ffi_java_ptrarray_to_raw): Same.
1321         (ffi_java_rvalue_to_raw): Add special handling for FFI_TYPE_POINTER
1322         if FFI_SIZEOF_JAVA_RAW == 4.
1323         (ffi_java_raw_to_rvalue): Same.
1324         (ffi_java_raw_call): Change type of raw to ffi_java_raw.
1325         (ffi_java_translate_args): Same.
1326         (ffi_prep_java_raw_closure_loc, ffi_prep_java_raw_closure): Change
1327         parameter types.
1328         * src/mips/ffitarget.h (FFI_SIZEOF_JAVA_RAW): Define for N32 ABI.
1329
1330 2007-12-06  David Daney  <ddaney@avtrex.com>
1331
1332         * src/mips/n32.S (ffi_closure_N32): Use 64-bit add instruction on
1333         pointer values.
1334
1335 2007-12-01  Andreas Tobler  <a.tobler@schweiz.org>
1336
1337         PR libffi/31937
1338         * src/powerpc/ffitarget.h: Introduce new ABI FFI_LINUX_SOFT_FLOAT.
1339         Add local FFI_TYPE_UINT128 to handle soft-float long-double-128.
1340         * src/powerpc/ffi.c: Distinguish between __NO_FPRS__ and not and
1341         set the NUM_FPR_ARG_REGISTERS according to.
1342         Add support for potential soft-float support under hard-float
1343         architecture.
1344         (ffi_prep_args_SYSV): Set NUM_FPR_ARG_REGISTERS to 0 in case of
1345         FFI_LINUX_SOFT_FLOAT, handle float, doubles and long-doubles according
1346         to the FFI_LINUX_SOFT_FLOAT ABI.
1347         (ffi_prep_cif_machdep): Likewise.
1348         (ffi_closure_helper_SYSV): Likewise.
1349         * src/powerpc/ppc_closure.S: Make sure not to store float/double
1350         on archs where __NO_FPRS__ is true.
1351         Add FFI_TYPE_UINT128 support.
1352         * src/powerpc/sysv.S: Add support for soft-float long-double-128.
1353         Adjust copyright notice.
1354
1355 2007-11-25  Andreas Tobler  <a.tobler@schweiz.org>
1356
1357         * src/closures.c: Move defintion of MAYBE_UNUSED from here to ...
1358         * include/ffi_common.h: ... here.
1359         Update copyright.
1360
1361 2007-11-17  Andreas Tobler  <a.tobler@schweiz.org>
1362
1363         * src/powerpc/sysv.S: Load correct cr to compare if we have long double.
1364         * src/powerpc/linux64.S: Likewise.
1365         * src/powerpc/ffi.c: Add a comment to show which part goes into cr6.
1366         * testsuite/libffi.call/return_ldl.c: New test.
1367
1368 2007-09-04    <aph@redhat.com>
1369
1370         * src/arm/sysv.S (UNWIND): New.
1371         (Whole file): Conditionally compile unwinder directives.
1372         * src/arm/sysv.S: Add unwinder directives.
1373
1374         * src/arm/ffi.c (ffi_prep_args): Align structs by at least 4 bytes.
1375         Only treat r0 as a struct address if we're actually returning a
1376         struct by address.
1377         Only copy the bytes that are actually within a struct.
1378         (ffi_prep_cif_machdep): A Composite Type not larger than 4 bytes
1379         is returned in r0, not passed by address.
1380         (ffi_call): Allocate a word-sized temporary for the case where
1381         a composite is returned in r0.
1382         (ffi_prep_incoming_args_SYSV): Align as necessary.
1383
1384 2007-08-05  Steven Newbury  <s_j_newbury@yahoo.co.uk>
1385
1386         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Use __clear_cache instead of
1387         directly using the sys_cacheflush syscall.
1388
1389 2007-07-27  Andrew Haley  <aph@redhat.com>
1390
1391         * src/arm/sysv.S (ffi_closure_SYSV): Add soft-float.
1392
1393 2007-09-03  Maciej W. Rozycki  <macro@linux-mips.org>
1394
1395         * Makefile.am: Unify MIPS_IRIX and MIPS_LINUX into MIPS.
1396         * configure.ac: Likewise.
1397         * Makefile.in: Regenerate.
1398         * include/Makefile.in: Likewise.
1399         * testsuite/Makefile.in: Likewise.
1400         * configure: Likewise.
1401
1402 2007-08-24  David Daney  <ddaney@avtrex.com>
1403
1404         * testsuite/libffi.call/return_sl.c: New test.
1405
1406 2007-08-10  David Daney  <ddaney@avtrex.com>
1407
1408         * testsuite/libffi.call/cls_multi_ushort.c,
1409         testsuite/libffi.call/cls_align_uint16.c,
1410         testsuite/libffi.call/nested_struct1.c,
1411         testsuite/libffi.call/nested_struct3.c,
1412         testsuite/libffi.call/cls_7_1_byte.c,
1413         testsuite/libffi.call/nested_struct5.c,
1414         testsuite/libffi.call/cls_double.c,
1415         testsuite/libffi.call/nested_struct7.c,
1416         testsuite/libffi.call/cls_sint.c,
1417         testsuite/libffi.call/nested_struct9.c,
1418         testsuite/libffi.call/cls_20byte1.c,
1419         testsuite/libffi.call/cls_multi_sshortchar.c,
1420         testsuite/libffi.call/cls_align_sint64.c,
1421         testsuite/libffi.call/cls_3byte2.c,
1422         testsuite/libffi.call/cls_multi_schar.c,
1423         testsuite/libffi.call/cls_multi_uchar.c,
1424         testsuite/libffi.call/cls_19byte.c,
1425         testsuite/libffi.call/cls_9byte1.c,
1426         testsuite/libffi.call/cls_align_float.c,
1427         testsuite/libffi.call/closure_fn1.c,
1428         testsuite/libffi.call/problem1.c,
1429         testsuite/libffi.call/closure_fn3.c,
1430         testsuite/libffi.call/cls_sshort.c,
1431         testsuite/libffi.call/closure_fn5.c,
1432         testsuite/libffi.call/cls_align_double.c,
1433         testsuite/libffi.call/nested_struct.c,
1434         testsuite/libffi.call/cls_2byte.c,
1435         testsuite/libffi.call/nested_struct10.c,
1436         testsuite/libffi.call/cls_4byte.c,
1437         testsuite/libffi.call/cls_6byte.c,
1438         testsuite/libffi.call/cls_8byte.c,
1439         testsuite/libffi.call/cls_multi_sshort.c,
1440         testsuite/libffi.call/cls_align_sint16.c,
1441         testsuite/libffi.call/cls_align_uint32.c,
1442         testsuite/libffi.call/cls_20byte.c,
1443         testsuite/libffi.call/cls_float.c,
1444         testsuite/libffi.call/nested_struct2.c,
1445         testsuite/libffi.call/cls_5_1_byte.c,
1446         testsuite/libffi.call/nested_struct4.c,
1447         testsuite/libffi.call/cls_24byte.c,
1448         testsuite/libffi.call/nested_struct6.c,
1449         testsuite/libffi.call/cls_64byte.c,
1450         testsuite/libffi.call/nested_struct8.c,
1451         testsuite/libffi.call/cls_uint.c,
1452         testsuite/libffi.call/cls_multi_ushortchar.c,
1453         testsuite/libffi.call/cls_schar.c,
1454         testsuite/libffi.call/cls_uchar.c,
1455         testsuite/libffi.call/cls_align_uint64.c,
1456         testsuite/libffi.call/cls_ulonglong.c,
1457         testsuite/libffi.call/cls_align_longdouble.c,
1458         testsuite/libffi.call/cls_1_1byte.c,
1459         testsuite/libffi.call/cls_12byte.c,
1460         testsuite/libffi.call/cls_3_1byte.c,
1461         testsuite/libffi.call/cls_3byte1.c,
1462         testsuite/libffi.call/cls_4_1byte.c,
1463         testsuite/libffi.call/cls_6_1_byte.c,
1464         testsuite/libffi.call/cls_16byte.c,
1465         testsuite/libffi.call/cls_18byte.c,
1466         testsuite/libffi.call/closure_fn0.c,
1467         testsuite/libffi.call/cls_9byte2.c,
1468         testsuite/libffi.call/closure_fn2.c,
1469         testsuite/libffi.call/closure_fn4.c,
1470         testsuite/libffi.call/cls_ushort.c,
1471         testsuite/libffi.call/closure_fn6.c,
1472         testsuite/libffi.call/cls_5byte.c,
1473         testsuite/libffi.call/cls_align_pointer.c,
1474         testsuite/libffi.call/cls_7byte.c,
1475         testsuite/libffi.call/cls_align_sint32.c,
1476         testsuite/libffi.special/unwindtest_ffi_call.cc,
1477         testsuite/libffi.special/unwindtest.cc: Remove xfail for mips64*-*-*.
1478
1479 2007-08-10  David Daney  <ddaney@avtrex.com>
1480
1481         PR libffi/28313
1482         * configure.ac: Don't treat mips64 as a special case.
1483         * Makefile.am (nodist_libffi_la_SOURCES): Add n32.S.
1484         * configure: Regenerate
1485         * Makefile.in: Ditto.
1486         * fficonfig.h.in: Ditto.
1487         * src/mips/ffitarget.h (REG_L, REG_S, SUBU, ADDU, SRL, LI): Indent.
1488         (LA, EH_FRAME_ALIGN, FDE_ADDR_BYTES): New preprocessor macros.
1489         (FFI_DEFAULT_ABI): Set for n64 case.
1490         (FFI_CLOSURES, FFI_TRAMPOLINE_SIZE): Define for n32 and n64 cases.
1491         * src/mips/n32.S (ffi_call_N32): Add debug macros and labels for FDE.
1492         (ffi_closure_N32): New function.
1493         (.eh_frame): New section
1494         * src/mips/o32.S: Clean up comments.
1495         (ffi_closure_O32): Pass ffi_closure parameter in $12.
1496         * src/mips/ffi.c: Use FFI_MIPS_N32 instead of
1497         _MIPS_SIM == _ABIN32 throughout.
1498         (FFI_MIPS_STOP_HERE): New, use in place of
1499         ffi_stop_here.
1500         (ffi_prep_args): Use unsigned long to hold pointer values.  Rewrite
1501         to support n32/n64 ABIs.
1502         (calc_n32_struct_flags): Rewrite.
1503         (calc_n32_return_struct_flags): Remove unused variable.  Reverse
1504         position of flag bits.
1505         (ffi_prep_cif_machdep): Rewrite n32 portion.
1506         (ffi_call): Enable for n64.  Add special handling for small structure
1507         return values.
1508         (ffi_prep_closure_loc): Add n32 and n64 support.
1509         (ffi_closure_mips_inner_O32): Add cast to silence warning.
1510         (copy_struct_N32, ffi_closure_mips_inner_N32): New functions.
1511
1512 2007-08-08  David Daney  <ddaney@avtrex.com>
1513
1514         * testsuite/libffi.call/ffitest.h (ffi_type_mylong): Remove definition.
1515         * testsuite/libffi.call/cls_align_uint16.c (main): Use correct type
1516         specifiers.
1517         * testsuite/libffi.call/nested_struct1.c (main): Ditto.
1518         * testsuite/libffi.call/cls_sint.c (main): Ditto.
1519         * testsuite/libffi.call/nested_struct9.c (main): Ditto.
1520         * testsuite/libffi.call/cls_20byte1.c (main): Ditto.
1521         * testsuite/libffi.call/cls_9byte1.c (main): Ditto.
1522         * testsuite/libffi.call/closure_fn1.c (main): Ditto.
1523         * testsuite/libffi.call/closure_fn3.c (main): Ditto.
1524         * testsuite/libffi.call/return_dbl2.c (main): Ditto.
1525         * testsuite/libffi.call/cls_sshort.c (main): Ditto.
1526         * testsuite/libffi.call/return_fl3.c (main): Ditto.
1527         * testsuite/libffi.call/closure_fn5.c (main): Ditto.
1528         * testsuite/libffi.call/nested_struct.c (main): Ditto.
1529         * testsuite/libffi.call/nested_struct10.c (main): Ditto.
1530         * testsuite/libffi.call/return_ll1.c (main): Ditto.
1531         * testsuite/libffi.call/cls_8byte.c (main): Ditto.
1532         * testsuite/libffi.call/cls_align_uint32.c (main): Ditto.
1533         * testsuite/libffi.call/cls_align_sint16.c (main): Ditto.
1534         * testsuite/libffi.call/cls_20byte.c (main): Ditto.
1535         * testsuite/libffi.call/nested_struct2.c (main): Ditto.
1536         * testsuite/libffi.call/cls_24byte.c (main): Ditto.
1537         * testsuite/libffi.call/nested_struct6.c (main): Ditto.
1538         * testsuite/libffi.call/cls_uint.c (main): Ditto.
1539         * testsuite/libffi.call/cls_12byte.c (main): Ditto.
1540         * testsuite/libffi.call/cls_16byte.c (main): Ditto.
1541         * testsuite/libffi.call/closure_fn0.c (main): Ditto.
1542         * testsuite/libffi.call/cls_9byte2.c (main): Ditto.
1543         * testsuite/libffi.call/closure_fn2.c (main): Ditto.
1544         * testsuite/libffi.call/return_dbl1.c (main): Ditto.
1545         * testsuite/libffi.call/closure_fn4.c (main): Ditto.
1546         * testsuite/libffi.call/closure_fn6.c (main): Ditto.
1547         * testsuite/libffi.call/cls_align_sint32.c (main): Ditto.
1548
1549 2007-08-07  Andrew Haley  <aph@redhat.com>
1550
1551         * src/x86/sysv.S (ffi_closure_raw_SYSV): Fix typo in previous
1552         checkin.
1553
1554 2007-08-06  Andrew Haley  <aph@redhat.com>
1555
1556         PR testsuite/32843
1557         * src/x86/sysv.S (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1558         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1559         FFI_TYPE_SINT32.
1560
1561 2007-08-02  David Daney  <ddaney@avtrex.com>
1562
1563         * testsuite/libffi.call/return_ul.c (main): Define return type as
1564         ffi_arg.  Use proper printf conversion specifier.
1565
1566 2007-07-30  Andrew Haley  <aph@redhat.com>
1567
1568         PR testsuite/32843
1569         * src/x86/ffi.c (ffi_prep_cif_machdep): in x86 case, add code for
1570         signed/unsigned int8/16.
1571         * src/x86/sysv.S (ffi_call_SYSV): Rewrite to:
1572         Use a jump table.
1573         Remove code to pop args from the stack after call.
1574         Special-case signed/unsigned int8/16.
1575         * testsuite/libffi.call/return_sc.c (main): Revert.
1576
1577 2007-07-26  Richard Guenther  <rguenther@suse.de>
1578
1579         PR testsuite/32843
1580         * testsuite/libffi.call/return_sc.c (main): Verify call
1581         result as signed char, not ffi_arg.
1582
1583 2007-07-16  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1584
1585         * configure.ac (i?86-*-solaris2.1[0-9]): Set TARGET to X86_64.
1586         * configure: Regenerate.
1587
1588 2007-07-11  David Daney  <ddaney@avtrex.com>
1589
1590         * src/mips/ffi.c: Don't include sys/cachectl.h.
1591         (ffi_prep_closure_loc): Use __builtin___clear_cache() instead of
1592         cacheflush().
1593
1594 2007-05-18  Aurelien Jarno  <aurelien@aurel32.net>
1595
1596         * src/arm/ffi.c (ffi_prep_closure_loc): Renamed and ajusted
1597         from (ffi_prep_closure): ... this.
1598         (FFI_INIT_TRAMPOLINE): Adjust.
1599
1600 2005-12-31  Phil Blundell  <pb@reciva.com>
1601
1602         * src/arm/ffi.c (ffi_prep_incoming_args_SYSV,
1603         ffi_closure_SYSV_inner, ffi_prep_closure): New, add closure support.
1604         * src/arm/sysv.S(ffi_closure_SYSV): Likewise.
1605         * src/arm/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1606         (FFI_CLOSURES): Enable closure support.
1607
1608 2007-07-03  Andrew Haley  <aph@hedges.billgatliff.com>
1609
1610         * testsuite/libffi.call/cls_multi_ushort.c,
1611         testsuite/libffi.call/cls_align_uint16.c,
1612         testsuite/libffi.call/nested_struct1.c,
1613         testsuite/libffi.call/nested_struct3.c,
1614         testsuite/libffi.call/cls_7_1_byte.c,
1615         testsuite/libffi.call/cls_double.c,
1616         testsuite/libffi.call/nested_struct5.c,
1617         testsuite/libffi.call/nested_struct7.c,
1618         testsuite/libffi.call/cls_sint.c,
1619         testsuite/libffi.call/nested_struct9.c,
1620         testsuite/libffi.call/cls_20byte1.c,
1621         testsuite/libffi.call/cls_multi_sshortchar.c,
1622         testsuite/libffi.call/cls_align_sint64.c,
1623         testsuite/libffi.call/cls_3byte2.c,
1624         testsuite/libffi.call/cls_multi_schar.c,
1625         testsuite/libffi.call/cls_multi_uchar.c,
1626         testsuite/libffi.call/cls_19byte.c,
1627         testsuite/libffi.call/cls_9byte1.c,
1628         testsuite/libffi.call/cls_align_float.c,
1629         testsuite/libffi.call/closure_fn1.c,
1630         testsuite/libffi.call/problem1.c,
1631         testsuite/libffi.call/closure_fn3.c,
1632         testsuite/libffi.call/cls_sshort.c,
1633         testsuite/libffi.call/closure_fn5.c,
1634         testsuite/libffi.call/cls_align_double.c,
1635         testsuite/libffi.call/cls_2byte.c,
1636         testsuite/libffi.call/nested_struct.c,
1637         testsuite/libffi.call/nested_struct10.c,
1638         testsuite/libffi.call/cls_4byte.c,
1639         testsuite/libffi.call/cls_6byte.c,
1640         testsuite/libffi.call/cls_8byte.c,
1641         testsuite/libffi.call/cls_multi_sshort.c,
1642         testsuite/libffi.call/cls_align_uint32.c,
1643         testsuite/libffi.call/cls_align_sint16.c,
1644         testsuite/libffi.call/cls_float.c,
1645         testsuite/libffi.call/cls_20byte.c,
1646         testsuite/libffi.call/cls_5_1_byte.c,
1647         testsuite/libffi.call/nested_struct2.c,
1648         testsuite/libffi.call/cls_24byte.c,
1649         testsuite/libffi.call/nested_struct4.c,
1650         testsuite/libffi.call/nested_struct6.c,
1651         testsuite/libffi.call/cls_64byte.c,
1652         testsuite/libffi.call/nested_struct8.c,
1653         testsuite/libffi.call/cls_uint.c,
1654         testsuite/libffi.call/cls_multi_ushortchar.c,
1655         testsuite/libffi.call/cls_schar.c,
1656         testsuite/libffi.call/cls_uchar.c,
1657         testsuite/libffi.call/cls_align_uint64.c,
1658         testsuite/libffi.call/cls_ulonglong.c,
1659         testsuite/libffi.call/cls_align_longdouble.c,
1660         testsuite/libffi.call/cls_1_1byte.c,
1661         testsuite/libffi.call/cls_12byte.c,
1662         testsuite/libffi.call/cls_3_1byte.c,
1663         testsuite/libffi.call/cls_3byte1.c,
1664         testsuite/libffi.call/cls_4_1byte.c,
1665         testsuite/libffi.call/cls_6_1_byte.c,
1666         testsuite/libffi.call/cls_16byte.c,
1667         testsuite/libffi.call/cls_18byte.c,
1668         testsuite/libffi.call/closure_fn0.c,
1669         testsuite/libffi.call/cls_9byte2.c,
1670         testsuite/libffi.call/closure_fn2.c,
1671         testsuite/libffi.call/closure_fn4.c,
1672         testsuite/libffi.call/cls_ushort.c,
1673         testsuite/libffi.call/closure_fn6.c,
1674         testsuite/libffi.call/cls_5byte.c,
1675         testsuite/libffi.call/cls_align_pointer.c,
1676         testsuite/libffi.call/cls_7byte.c,
1677         testsuite/libffi.call/cls_align_sint32.c,
1678         testsuite/libffi.special/unwindtest_ffi_call.cc,
1679         testsuite/libffi.special/unwindtest.cc: Enable for ARM.
1680
1681 2007-07-05  H.J. Lu  <hongjiu.lu@intel.com>
1682
1683         * aclocal.m4: Regenerated.
1684
1685 2007-06-02  Paolo Bonzini  <bonzini@gnu.org>
1686
1687         * configure: Regenerate.
1688
1689 2007-05-23  Steve Ellcey  <sje@cup.hp.com>
1690
1691         * Makefile.in: Regenerate.
1692         * configure: Regenerate.
1693         * aclocal.m4: Regenerate.
1694         * include/Makefile.in: Regenerate.
1695         * testsuite/Makefile.in: Regenerate.
1696
1697 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1698
1699         * src/m68k/ffi.c (ffi_prep_incoming_args_SYSV,
1700         ffi_closure_SYSV_inner,ffi_prep_closure): New, add closure support.
1701         * src/m68k/sysv.S(ffi_closure_SYSV,ffi_closure_struct_SYSV): Likewise.
1702         * src/m68k/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1703         (FFI_CLOSURES): Enable closure support.
1704
1705 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1706
1707         * configure.ac (HAVE_AS_CFI_PSEUDO_OP): New test.
1708         * configure: Regenerate.
1709         * fficonfig.h.in: Regenerate.
1710         * src/m68k/sysv.S (CFI_STARTPROC,CFI_ENDPROC,
1711         CFI_OFFSET,CFI_DEF_CFA): New macros.
1712         (ffi_call_SYSV): Add callframe annotation.
1713
1714 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1715
1716         * src/m68k/ffi.c (ffi_prep_args,ffi_prep_cif_machdep): Fix
1717         numerous test suite failures.
1718         * src/m68k/sysv.S (ffi_call_SYSV): Likewise.
1719
1720 2007-04-11  Paolo Bonzini  <bonzini@gnu.org>
1721
1722         * Makefile.am (EXTRA_DIST): Bring up to date.
1723         * Makefile.in: Regenerate.
1724         * src/frv/eabi.S: Remove RCS keyword.
1725
1726 2007-04-06  Richard Henderson  <rth@redhat.com>
1727
1728         * configure.ac: Tidy target case.
1729         (HAVE_LONG_DOUBLE): Allow the target to override.
1730         * configure: Regenerate.
1731         * include/ffi.h.in: Don't define ffi_type_foo if
1732         LIBFFI_HIDE_BASIC_TYPES is defined.
1733         (ffi_type_longdouble): If not HAVE_LONG_DOUBLE, define
1734         to ffi_type_double.
1735         * types.c (LIBFFI_HIDE_BASIC_TYPES): Define.
1736         (FFI_TYPEDEF, ffi_type_void): Mark the data const.
1737         (ffi_type_longdouble): Special case for Alpha.  Don't define
1738         if long double == double.
1739
1740         * src/alpha/ffi.c (FFI_TYPE_LONGDOUBLE): Assert unique value.
1741         (ffi_prep_cif_machdep): Handle it as the 128-bit type.
1742         (ffi_call, ffi_closure_osf_inner): Likewise.
1743         (ffi_closure_osf_inner): Likewise.  Mark hidden.
1744         (ffi_call_osf, ffi_closure_osf): Mark hidden.
1745         * src/alpha/ffitarget.h (FFI_LAST_ABI): Tidy definition.
1746         * src/alpha/osf.S (ffi_call_osf, ffi_closure_osf): Mark hidden.
1747         (load_table): Handle 128-bit long double.
1748
1749         * testsuite/libffi.call/float4.c: Add -mieee for alpha.
1750
1751 2007-04-06  Tom Tromey  <tromey@redhat.com>
1752
1753         PR libffi/31491:
1754         * README: Fixed bug in example.
1755
1756 2007-04-03  Jakub Jelinek  <jakub@redhat.com>
1757
1758         * src/closures.c: Include sys/statfs.h.
1759         (_GNU_SOURCE): Define on Linux.
1760         (FFI_MMAP_EXEC_SELINUX): Define.
1761         (selinux_enabled): New variable.
1762         (selinux_enabled_check): New function.
1763         (is_selinux_enabled): Define.
1764         (dlmmap): Use it.
1765
1766 2007-03-24  Uros Bizjak  <ubizjak@gmail.com>
1767
1768         * testsuite/libffi.call/return_fl2.c (return_fl): Mark as static.
1769         Use 'volatile float sum' to create sum of floats to avoid false
1770         negative due to excess precision on ix86 targets.
1771         (main): Ditto.
1772
1773 2007-03-08  Alexandre Oliva  <aoliva@redhat.com>
1774
1775         * src/powerpc/ffi.c (flush_icache): Fix left-over from previous
1776         patch.
1777         (ffi_prep_closure_loc): Remove unneeded casts.  Add needed ones.
1778
1779 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1780
1781         * include/ffi.h.in (ffi_closure_alloc, ffi_closure_free): New.
1782         (ffi_prep_closure_loc): New.
1783         (ffi_prep_raw_closure_loc): New.
1784         (ffi_prep_java_raw_closure_loc): New.
1785         * src/closures.c: New file.
1786         * src/dlmalloc.c [FFI_MMAP_EXEC_WRIT] (struct malloc_segment):
1787         Replace sflags with exec_offset.
1788         [FFI_MMAP_EXEC_WRIT] (mmap_exec_offset, add_segment_exec_offset,
1789         sub_segment_exec_offset): New macros.
1790         (get_segment_flags, set_segment_flags, check_segment_merge): New
1791         macros.
1792         (is_mmapped_segment, is_extern_segment): Use get_segment_flags.
1793         (add_segment, sys_alloc, create_mspace, create_mspace_with_base,
1794         destroy_mspace): Use new macros.
1795         (sys_alloc): Silence warning.
1796         * Makefile.am (libffi_la_SOURCES): Add src/closures.c.
1797         * Makefile.in: Rebuilt.
1798         * src/prep_cif [FFI_CLOSURES] (ffi_prep_closure): Implement in
1799         terms of ffi_prep_closure_loc.
1800         * src/raw_api.c (ffi_prep_raw_closure_loc): Renamed and adjusted
1801         from...
1802         (ffi_prep_raw_closure): ... this.  Re-implement in terms of the
1803         renamed version.
1804         * src/java_raw_api (ffi_prep_java_raw_closure_loc): Renamed and
1805         adjusted from...
1806         (ffi_prep_java_raw_closure): ... this.  Re-implement in terms of
1807         the renamed version.
1808         * src/alpha/ffi.c (ffi_prep_closure_loc): Renamed from
1809         (ffi_prep_closure): ... this.
1810         * src/pa/ffi.c: Likewise.
1811         * src/cris/ffi.c: Likewise.  Adjust.
1812         * src/frv/ffi.c: Likewise.
1813         * src/ia64/ffi.c: Likewise.
1814         * src/mips/ffi.c: Likewise.
1815         * src/powerpc/ffi_darwin.c: Likewise.
1816         * src/s390/ffi.c: Likewise.
1817         * src/sh/ffi.c: Likewise.
1818         * src/sh64/ffi.c: Likewise.
1819         * src/sparc/ffi.c: Likewise.
1820         * src/x86/ffi64.c: Likewise.
1821         * src/x86/ffi.c: Likewise.
1822         (FFI_INIT_TRAMPOLINE): Adjust.
1823         (ffi_prep_raw_closure_loc): Renamed and adjusted from...
1824         (ffi_prep_raw_closure): ... this.
1825         * src/powerpc/ffi.c (ffi_prep_closure_loc): Renamed from
1826         (ffi_prep_closure): ... this.
1827         (flush_icache): Adjust.
1828
1829 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1830
1831         * src/dlmalloc.c: New file, imported version 2.8.3 of Doug
1832         Lea's malloc.
1833
1834 2007-03-01  Brooks Moses  <brooks.moses@codesourcery.com>
1835
1836         * Makefile.am: Add dummy install-pdf target.
1837         * Makefile.in: Regenerate
1838
1839 2007-02-13  Andreas Krebbel  <krebbel1@de.ibm.com>
1840
1841         * src/s390/ffi.c (ffi_prep_args, ffi_prep_cif_machdep,
1842         ffi_closure_helper_SYSV): Add long double handling.
1843
1844 2007-02-02  Jakub Jelinek  <jakub@redhat.com>
1845
1846         * src/powerpc/linux64.S (ffi_call_LINUX64): Move restore of r2
1847         immediately after bctrl instruction.
1848
1849 2007-01-18  Alexandre Oliva  <aoliva@redhat.com>
1850
1851         * Makefile.am (all-recursive, install-recursive,
1852         mostlyclean-recursive, clean-recursive, distclean-recursive,
1853         maintainer-clean-recursive): Add missing targets.
1854         * Makefile.in: Rebuilt.
1855
1856 2006-12-14  Andreas Tobler  <a.tobler@schweiz.org>
1857
1858         * configure.ac: Add TARGET for x86_64-*-darwin*.
1859         * Makefile.am (nodist_libffi_la_SOURCES): Add rules for 64-bit sources
1860         for X86_DARWIN.
1861         * src/x86/ffitarget.h: Set trampoline size for x86_64-*-darwin*.
1862         * src/x86/darwin64.S: New file for x86_64-*-darwin* support.
1863         * configure: Regenerate.
1864         * Makefile.in: Regenerate.
1865         * include/Makefile.in: Regenerate.
1866         * testsuite/Makefile.in: Regenerate.
1867         * testsuite/libffi.special/unwindtest_ffi_call.cc: New test case for
1868         ffi_call only.
1869
1870 2006-12-13  Andreas Tobler <a.tobler@schweiz.org>
1871
1872         * aclocal.m4: Regenerate with aclocal -I .. as written in the
1873         Makefile.am.
1874
1875 2006-10-31  Geoffrey Keating  <geoffk@apple.com>
1876
1877         * src/powerpc/ffi_darwin.c (darwin_adjust_aggregate_sizes): New.
1878         (ffi_prep_cif_machdep): Call darwin_adjust_aggregate_sizes for
1879         Darwin.
1880         * testsuite/libffi.call/nested_struct4.c: Remove Darwin XFAIL.
1881         * testsuite/libffi.call/nested_struct6.c: Remove Darwin XFAIL.
1882
1883 2006-10-10  Paolo Bonzini  <bonzini@gnu.org>
1884             Sandro Tolaini  <tolaini@libero.it>
1885
1886         * configure.ac [i*86-*-darwin*]: Set X86_DARWIN symbol and
1887         conditional.
1888         * configure: Regenerated.
1889         * Makefile.am (nodist_libffi_la_SOURCES) [X86_DARWIN]: New case.
1890         (EXTRA_DIST): Add src/x86/darwin.S.
1891         * Makefile.in: Regenerated.
1892         * include/Makefile.in: Regenerated.
1893         * testsuite/Makefile.in: Regenerated.
1894
1895         * src/x86/ffi.c (ffi_prep_cif_machdep) [X86_DARWIN]: Treat like
1896         X86_WIN32, and additionally align stack to 16 bytes.
1897         * src/x86/darwin.S: New, based on sysv.S.
1898         * src/prep_cif.c (ffi_prep_cif) [X86_DARWIN]: Align > 8-byte structs.
1899
1900 2006-09-12  David Daney  <ddaney@avtrex.com>
1901
1902         PR libffi/23935
1903         * include/Makefile.am: Install both ffi.h and ffitarget.h in
1904         $(libdir)/gcc/$(target_alias)/$(gcc_version)/include.
1905         * aclocal.m4: Regenerated for automake 1.9.6.
1906         * Makefile.in: Regenerated.
1907         * include/Makefile.in: Regenerated.
1908         * testsuite/Makefile.in: Regenerated.
1909
1910 2006-08-17  Andreas Tobler  <a.tobler@schweiz.ch>
1911
1912         * include/ffi_common.h (struct): Revert accidental commit.
1913
1914 2006-08-15  Andreas Tobler  <a.tobler@schweiz.ch>
1915
1916         * include/ffi_common.h: Remove lint directives.
1917         * include/ffi.h.in: Likewise.
1918
1919 2006-07-25  Torsten Schoenfeld  <kaffeetisch@gmx.de>
1920
1921         * include/ffi.h.in (ffi_type_ulong, ffi_type_slong): Define correctly
1922         for 32-bit architectures.
1923         * testsuite/libffi.call/return_ul.c: New test case.
1924
1925 2006-07-19  David Daney  <ddaney@avtrex.com>
1926
1927         * testsuite/libffi.call/closure_fn6.c: Remove xfail for mips,
1928         xfail remains for mips64.
1929
1930 2006-05-23  Carlos O'Donell  <carlos@codesourcery.com>
1931
1932         * Makefile.am: Add install-html target. Add install-html to .PHONY
1933         * Makefile.in: Regenerate.
1934         * aclocal.m4: Regenerate.
1935         * include/Makefile.in: Regenerate.
1936         * testsuite/Makefile.in: Regenerate.
1937
1938 2006-05-18  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1939
1940         * pa/ffi.c (ffi_prep_args_pa32): Load floating point arguments from
1941         stack slot.
1942
1943 2006-04-22  Andreas Tobler  <a.tobler@schweiz.ch>
1944
1945         * README: Remove notice about 'Crazy Comments'.
1946         * src/debug.c: Remove lint directives. Cleanup white spaces.
1947         * src/java_raw_api.c: Likewise.
1948         * src/prep_cif.c: Likewise.
1949         * src/raw_api.c: Likewise.
1950         * src/ffitest.c: Delete. No longer needed, all test cases migrated
1951         to the testsuite.
1952         * src/arm/ffi.c: Remove lint directives.
1953         * src/m32r/ffi.c: Likewise.
1954         * src/pa/ffi.c: Likewise.
1955         * src/powerpc/ffi.c: Likewise.
1956         * src/powerpc/ffi_darwin.c: Likewise.
1957         * src/sh/ffi.c: Likewise.
1958         * src/sh64/ffi.c: Likewise.
1959         * src/x86/ffi.c: Likewise.
1960         * testsuite/libffi.call/float2.c: Likewise.
1961         * testsuite/libffi.call/promotion.c: Likewise.
1962         * testsuite/libffi.call/struct1.c: Likewise.
1963
1964 2006-04-13  Andreas Tobler  <a.tobler@schweiz.ch>
1965
1966         * src/pa/hpux32.S: Correct unwind offset calculation for
1967         ffi_closure_pa32.
1968         * src/pa/linux.S: Likewise.
1969
1970 2006-04-12  James E Wilson  <wilson@specifix.com>
1971
1972         PR libgcj/26483
1973         * src/ia64/ffi.c (stf_spill, ldf_fill): Rewrite as macros.
1974         (hfa_type_load): Call stf_spill.
1975         (hfa_type_store): Call ldf_fill.
1976         (ffi_call): Adjust calls to above routines.  Add local temps for
1977         macro result.
1978
1979 2006-04-10  Matthias Klose  <doko@debian.org>
1980
1981         * testsuite/lib/libffi-dg.exp (libffi-init): Recognize multilib
1982         directory names containing underscores.
1983
1984 2006-04-07  James E Wilson  <wilson@specifix.com>
1985
1986         * testsuite/libffi.call/float4.c: New testcase.
1987
1988 2006-04-05  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1989             Andreas Tobler  <a.tobler@schweiz.ch>
1990
1991         * Makefile.am: Add PA_HPUX port.
1992         * Makefile.in: Regenerate.
1993         * include/Makefile.in: Likewise.
1994         * testsuite/Makefile.in: Likewise.
1995         * configure.ac: Add PA_HPUX rules.
1996         * configure: Regenerate.
1997         * src/pa/ffitarget.h: Rename linux target to PA_LINUX.
1998         Add PA_HPUX and PA64_HPUX.
1999         Rename FFI_LINUX ABI to FFI_PA32 ABI.
2000         (FFI_TRAMPOLINE_SIZE): Define for 32-bit HP-UX targets.
2001         (FFI_TYPE_SMALL_STRUCT2): Define.
2002         (FFI_TYPE_SMALL_STRUCT4): Likewise.
2003         (FFI_TYPE_SMALL_STRUCT8): Likewise.
2004         (FFI_TYPE_SMALL_STRUCT3): Redefine.
2005         (FFI_TYPE_SMALL_STRUCT5): Likewise.
2006         (FFI_TYPE_SMALL_STRUCT6): Likewise.
2007         (FFI_TYPE_SMALL_STRUCT7): Likewise.
2008         * src/pa/ffi.c (ROUND_DOWN): Delete.
2009         (fldw, fstw, fldd, fstd): Use '__asm__'.
2010         (ffi_struct_type): Add support for FFI_TYPE_SMALL_STRUCT2,
2011         FFI_TYPE_SMALL_STRUCT4 and FFI_TYPE_SMALL_STRUCT8.
2012         (ffi_prep_args_LINUX): Rename to ffi_prep_args_pa32. Update comment.
2013         Simplify incrementing of stack slot variable. Change type of local
2014         'n' to unsigned int.
2015         (ffi_size_stack_LINUX): Rename to ffi_size_stack_pa32. Handle long
2016         double on PA_HPUX.
2017         (ffi_prep_cif_machdep): Likewise.
2018         (ffi_call): Likewise.
2019         (ffi_closure_inner_LINUX): Rename to ffi_closure_inner_pa32. Change
2020         return type to ffi_status. Simplify incrementing of stack slot
2021         variable. Only copy floating point argument registers when PA_LINUX
2022         is true. Reformat debug statement.
2023         Add support for FFI_TYPE_SMALL_STRUCT2, FFI_TYPE_SMALL_STRUCT4 and
2024         FFI_TYPE_SMALL_STRUCT8.
2025         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Add 'extern' to
2026         declaration.
2027         (ffi_prep_closure): Make linux trampoline conditional on PA_LINUX.
2028         Add nops to cache flush.  Add trampoline for PA_HPUX.
2029         * src/pa/hpux32.S: New file.
2030         * src/pa/linux.S (ffi_call_LINUX): Rename to ffi_call_pa32. Rename
2031         ffi_prep_args_LINUX to ffi_prep_args_pa32.
2032         Localize labels. Add support for 2, 4 and 8-byte small structs. Handle
2033         unaligned destinations in 3, 5, 6 and 7-byte small structs. Order
2034         argument type checks so that common argument types appear first.
2035         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Rename
2036         ffi_closure_inner_LINUX to ffi_closure_inner_pa32.
2037
2038 2006-03-24  Alan Modra  <amodra@bigpond.net.au>
2039
2040         * src/powerpc/ffitarget.h (enum ffi_abi): Add FFI_LINUX.  Default
2041         for 32-bit using IBM extended double format.  Fix FFI_LAST_ABI.
2042         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Handle linux variant of
2043         FFI_TYPE_LONGDOUBLE.
2044         (ffi_prep_args64): Assert using IBM extended double.
2045         (ffi_prep_cif_machdep): Don't munge FFI_TYPE_LONGDOUBLE type.
2046         Handle FFI_LINUX FFI_TYPE_LONGDOUBLE return and args.
2047         (ffi_call): Handle FFI_LINUX.
2048         (ffi_closure_helper_SYSV): Non FFI_LINUX long double return needs
2049         gpr3 return pointer as for struct return.  Handle FFI_LINUX
2050         FFI_TYPE_LONGDOUBLE return and args.  Don't increment "nf"
2051         unnecessarily.
2052         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Load both f1 and f2
2053         for FFI_TYPE_LONGDOUBLE.  Move epilogue insns into case table.
2054         Don't use r6 as pointer to results, instead use sp offset.  Don't
2055         make a special call to load lr with case table address, instead
2056         use offset from previous call.
2057         * src/powerpc/sysv.S (ffi_call_SYSV): Save long double return.
2058         * src/powerpc/linux64.S (ffi_call_LINUX64): Simplify long double
2059         return.
2060
2061 2006-03-15  Kaz Kojima  <kkojima@gcc.gnu.org>
2062
2063         * src/sh64/ffi.c (ffi_prep_cif_machdep): Handle float arguments
2064         passed with FP registers correctly.
2065         (ffi_closure_helper_SYSV): Likewise.
2066         * src/sh64/sysv.S: Likewise.
2067
2068 2006-03-01  Andreas Tobler  <a.tobler@schweiz.ch>
2069
2070         * testsuite/libffi.special/unwindtest.cc (closure_test_fn): Mark cif,
2071         args and userdata unused.
2072         (closure_test_fn1): Mark cif and userdata unused.
2073         (main): Remove unused res.
2074
2075 2006-02-28  Andreas Tobler  <a.tobler@schweiz.ch>
2076
2077         * testsuite/libffi.call/call.exp: Adjust FSF address. Add test runs for
2078         -O2, -O3, -Os and the warning flags -W -Wall.
2079         * testsuite/libffi.special/special.exp: Likewise.
2080         * testsuite/libffi.call/ffitest.h: Add an __UNUSED__ macro to mark
2081         unused parameter unused for gcc or else do nothing.
2082         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2083         * testsuite/libffi.call/cls_12byte.c (cls_struct_12byte_gn): Mark cif
2084         and userdata unused.
2085         * testsuite/libffi.call/cls_16byte.c (cls_struct_16byte_gn): Likewise.
2086         * testsuite/libffi.call/cls_18byte.c (cls_struct_18byte_gn): Likewise.
2087         * testsuite/libffi.call/cls_19byte.c (cls_struct_19byte_gn): Likewise.
2088         * testsuite/libffi.call/cls_1_1byte.c (cls_struct_1_1byte_gn): Likewise.
2089         * testsuite/libffi.call/cls_20byte.c (cls_struct_20byte_gn): Likewise.
2090         * testsuite/libffi.call/cls_20byte1.c (cls_struct_20byte_gn): Likewise.
2091         * testsuite/libffi.call/cls_24byte.c (cls_struct_24byte_gn): Likewise.
2092         * testsuite/libffi.call/cls_2byte.c (cls_struct_2byte_gn): Likewise.
2093         * testsuite/libffi.call/cls_3_1byte.c (cls_struct_3_1byte_gn): Likewise.
2094         * testsuite/libffi.call/cls_3byte1.c (cls_struct_3byte_gn): Likewise.
2095         * testsuite/libffi.call/cls_3byte2.c (cls_struct_3byte_gn1): Likewise.
2096         * testsuite/libffi.call/cls_4_1byte.c (cls_struct_4_1byte_gn): Likewise.
2097         * testsuite/libffi.call/cls_4byte.c (cls_struct_4byte_gn): Likewise.
2098         * testsuite/libffi.call/cls_5_1_byte.c (cls_struct_5byte_gn): Likewise.
2099         * testsuite/libffi.call/cls_5byte.c (cls_struct_5byte_gn): Likewise.
2100         * testsuite/libffi.call/cls_64byte.c (cls_struct_64byte_gn): Likewise.
2101         * testsuite/libffi.call/cls_6_1_byte.c (cls_struct_6byte_gn): Likewise.
2102         * testsuite/libffi.call/cls_6byte.c (cls_struct_6byte_gn): Likewise.
2103         * testsuite/libffi.call/cls_7_1_byte.c (cls_struct_7byte_gn): Likewise.
2104         * testsuite/libffi.call/cls_7byte.c (cls_struct_7byte_gn): Likewise.
2105         * testsuite/libffi.call/cls_8byte.c (cls_struct_8byte_gn): Likewise.
2106         * testsuite/libffi.call/cls_9byte1.c (cls_struct_9byte_gn): Likewise.
2107         * testsuite/libffi.call/cls_9byte2.c (cls_struct_9byte_gn): Likewise.
2108         * testsuite/libffi.call/cls_align_double.c (cls_struct_align_gn):
2109         Likewise.
2110         * testsuite/libffi.call/cls_align_float.c (cls_struct_align_gn):
2111         Likewise.
2112         * testsuite/libffi.call/cls_align_longdouble.c (cls_struct_align_gn):
2113         Likewise.
2114         * testsuite/libffi.call/cls_align_pointer.c (cls_struct_align_fn): Cast
2115         void* to avoid compiler warning.
2116         (main): Likewise.
2117         (cls_struct_align_gn): Mark cif and userdata unused.
2118         * testsuite/libffi.call/cls_align_sint16.c (cls_struct_align_gn):
2119         Likewise.
2120         * testsuite/libffi.call/cls_align_sint32.c (cls_struct_align_gn):
2121         Likewise.
2122         * testsuite/libffi.call/cls_align_sint64.c (cls_struct_align_gn):
2123         Likewise.
2124         * testsuite/libffi.call/cls_align_uint16.c (cls_struct_align_gn):
2125         Likewise.
2126         * testsuite/libffi.call/cls_align_uint32.c (cls_struct_align_gn):
2127         Likewise.
2128         * testsuite/libffi.call/cls_double.c (cls_ret_double_fn): Likewise.
2129         * testsuite/libffi.call/cls_float.c (cls_ret_float_fn): Likewise.
2130         * testsuite/libffi.call/cls_multi_schar.c (test_func_gn): Mark cif and
2131         data unused.
2132         (main): Cast res_call to silence gcc.
2133         * testsuite/libffi.call/cls_multi_sshort.c (test_func_gn): Mark cif and
2134         data unused.
2135         (main): Cast res_call to silence gcc.
2136         * testsuite/libffi.call/cls_multi_sshortchar.c (test_func_gn): Mark cif
2137         and data unused.
2138         (main): Cast res_call to silence gcc.
2139         * testsuite/libffi.call/cls_multi_uchar.c (test_func_gn): Mark cif and
2140         data unused.
2141         (main): Cast res_call to silence gcc.
2142         * testsuite/libffi.call/cls_multi_ushort.c (test_func_gn): Mark cif and
2143         data unused.
2144         (main): Cast res_call to silence gcc.
2145         * testsuite/libffi.call/cls_multi_ushortchar.c (test_func_gn): Mark cif
2146         and data unused.
2147         (main): Cast res_call to silence gcc.
2148         * testsuite/libffi.call/cls_schar.c (cls_ret_schar_fn): Mark cif and
2149         userdata unused.
2150         (cls_ret_schar_fn): Cast printf parameter to silence gcc.
2151         * testsuite/libffi.call/cls_sint.c (cls_ret_sint_fn): Mark cif and
2152         userdata unused.
2153         (cls_ret_sint_fn): Cast printf parameter to silence gcc.
2154         * testsuite/libffi.call/cls_sshort.c (cls_ret_sshort_fn): Mark cif and
2155         userdata unused.
2156         (cls_ret_sshort_fn): Cast printf parameter to silence gcc.
2157         * testsuite/libffi.call/cls_uchar.c (cls_ret_uchar_fn):  Mark cif and
2158         userdata unused.
2159         (cls_ret_uchar_fn): Cast printf parameter to silence gcc.
2160         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Mark cif and
2161         userdata unused.
2162         (cls_ret_uint_fn): Cast printf parameter to silence gcc.
2163         * testsuite/libffi.call/cls_ulonglong.c (cls_ret_ulonglong_fn): Mark cif
2164         and userdata unused.
2165         * testsuite/libffi.call/cls_ushort.c (cls_ret_ushort_fn): Mark cif and
2166         userdata unused.
2167         (cls_ret_ushort_fn): Cast printf parameter to silence gcc.
2168         * testsuite/libffi.call/float.c (floating): Remove unused parameter e.
2169         * testsuite/libffi.call/float1.c (main): Remove unused variable i.
2170         Cleanup white spaces.
2171         * testsuite/libffi.call/negint.c (checking): Remove unused variable i.
2172         * testsuite/libffi.call/nested_struct.c (cls_struct_combined_gn): Mark
2173         cif and userdata unused.
2174         * testsuite/libffi.call/nested_struct1.c (cls_struct_combined_gn):
2175         Likewise.
2176         * testsuite/libffi.call/nested_struct10.c (B_gn): Likewise.
2177         * testsuite/libffi.call/nested_struct2.c (B_fn): Adjust printf
2178         formatters to silence gcc.
2179         (B_gn): Mark cif and userdata unused.
2180         * testsuite/libffi.call/nested_struct3.c (B_gn): Mark cif and userdata
2181         unused.
2182         * testsuite/libffi.call/nested_struct4.c: Mention related PR.
2183         (B_gn): Mark cif and userdata unused.
2184         * testsuite/libffi.call/nested_struct5.c (B_gn): Mark cif and userdata
2185         unused.
2186         * testsuite/libffi.call/nested_struct6.c: Mention related PR.
2187         (B_gn): Mark cif and userdata unused.
2188         * testsuite/libffi.call/nested_struct7.c (B_gn): Mark cif and userdata
2189         unused.
2190         * testsuite/libffi.call/nested_struct8.c (B_gn): Likewise.
2191         * testsuite/libffi.call/nested_struct9.c (B_gn): Likewise.
2192         * testsuite/libffi.call/problem1.c (stub): Likewise.
2193         * testsuite/libffi.call/pyobjc-tc.c (main): Cast the result to silence
2194         gcc.
2195         * testsuite/libffi.call/return_fl2.c (return_fl): Add the note mentioned
2196         in the last commit for this test case in the test case itself.
2197         * testsuite/libffi.call/closure_fn0.c (closure_test_fn0): Mark cif as
2198         unused.
2199         * testsuite/libffi.call/closure_fn1.c (closure_test_fn1): Likewise.
2200         * testsuite/libffi.call/closure_fn2.c (closure_test_fn2): Likewise.
2201         * testsuite/libffi.call/closure_fn3.c (closure_test_fn3): Likewise.
2202         * testsuite/libffi.call/closure_fn4.c (closure_test_fn0): Likewise.
2203         * testsuite/libffi.call/closure_fn5.c (closure_test_fn5): Likewise.
2204         * testsuite/libffi.call/closure_fn6.c (closure_test_fn0): Likewise.
2205
2206 2006-02-22  Kaz Kojima  <kkojima@gcc.gnu.org>
2207
2208         * src/sh/sysv.S: Fix register numbers in the FDE for
2209         ffi_closure_SYSV.
2210
2211 2006-02-20  Andreas Tobler  <a.tobler@schweiz.ch>
2212
2213         * testsuite/libffi.call/return_fl2.c (return_fl): Remove static
2214         declaration to avoid a false negative on ix86. See PR323.
2215
2216 2006-02-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2217
2218         * src/sh/ffi.c (ffi_closure_helper_SYSV): Remove unused variable
2219         and cast integer to void * if needed.  Update the pointer to
2220         the FP register saved area correctly.
2221
2222 2006-02-17  Andreas Tobler  <a.tobler@schweiz.ch>
2223
2224         * testsuite/libffi.call/nested_struct6.c: XFAIL this test until PR25630
2225         is fixed.
2226         * testsuite/libffi.call/nested_struct4.c: Likewise.
2227
2228 2006-02-16  Andreas Tobler  <a.tobler@schweiz.ch>
2229
2230         * testsuite/libffi.call/return_dbl.c: New test case.
2231         * testsuite/libffi.call/return_dbl1.c: Likewise.
2232         * testsuite/libffi.call/return_dbl2.c: Likewise.
2233         * testsuite/libffi.call/return_fl.c: Likewise.
2234         * testsuite/libffi.call/return_fl1.c: Likewise.
2235         * testsuite/libffi.call/return_fl2.c: Likewise.
2236         * testsuite/libffi.call/return_fl3.c: Likewise.
2237         * testsuite/libffi.call/closure_fn6.c: Likewise.
2238
2239         * testsuite/libffi.call/nested_struct2.c: Remove ffi_type_mylong
2240         definition.
2241         * testsuite/libffi.call/ffitest.h: Add ffi_type_mylong definition
2242         here to be used by other test cases too.
2243
2244         * testsuite/libffi.call/nested_struct10.c: New test case.
2245         * testsuite/libffi.call/nested_struct9.c: Likewise.
2246         * testsuite/libffi.call/nested_struct8.c: Likewise.
2247         * testsuite/libffi.call/nested_struct7.c: Likewise.
2248         * testsuite/libffi.call/nested_struct6.c: Likewise.
2249         * testsuite/libffi.call/nested_struct5.c: Likewise.
2250         * testsuite/libffi.call/nested_struct4.c: Likewise.
2251
2252 2006-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
2253
2254         * configure.ac: Enable libffi for sparc64-*-freebsd*.
2255         * configure: Rebuilt.
2256
2257 2006-01-18  Jakub Jelinek  <jakub@redhat.com>
2258
2259         * src/powerpc/sysv.S (smst_two_register): Don't call __ashldi3,
2260         instead do the shifting inline.
2261         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't compute %r5
2262         shift count unconditionally.  Simplify load sequences for 1, 2, 3, 4
2263         and 8 byte structs, for the remaining struct sizes don't call
2264         __lshrdi3, instead do the shifting inline.
2265
2266 2005-12-07  Thiemo Seufer  <ths@networkno.de>
2267
2268         * src/mips/ffitarget.h: Remove obsolete sgidefs.h include. Add
2269         missing parentheses.
2270         * src/mips/o32.S (ffi_call_O32): Code formatting. Define
2271         and use A3_OFF, FP_OFF, RA_OFF. Micro-optimizations.
2272         (ffi_closure_O32): Likewise, but with newly defined A3_OFF2,
2273         A2_OFF2, A1_OFF2, A0_OFF2, RA_OFF2, FP_OFF2, S0_OFF2, GP_OFF2,
2274         V1_OFF2, V0_OFF2, FA_1_1_OFF2, FA_1_0_OFF2, FA_0_1_OFF2,
2275         FA_0_0_OFF2.
2276         * src/mips/ffi.c (ffi_prep_args): Code formatting. Fix
2277         endianness bugs.
2278         (ffi_prep_closure): Improve trampoline instruction scheduling.
2279         (ffi_closure_mips_inner_O32): Fix endianness bugs.
2280
2281 2005-12-03  Alan Modra  <amodra@bigpond.net.au>
2282
2283         * src/powerpc/ffi.c: Formatting.
2284         (ffi_prep_args_SYSV): Avoid possible aliasing problems by using unions.
2285         (ffi_prep_args64): Likewise.
2286
2287 2005-09-30  Geoffrey Keating  <geoffk@apple.com>
2288
2289         * testsuite/lib/libffi-dg.exp (libffi_target_compile): For
2290         darwin, use -shared-libgcc not -lgcc_s, and explain why.
2291
2292 2005-09-26  Tom Tromey  <tromey@redhat.com>
2293
2294         * testsuite/libffi.call/float1.c (value_type): New typedef.
2295         (CANARY): New define.
2296         (main): Check for result buffer overflow.
2297         * src/powerpc/linux64.S: Handle linux64 long double returns.
2298         * src/powerpc/ffi.c (FLAG_RETURNS_128BITS): New constant.
2299         (ffi_prep_cif_machdep): Handle linux64 long double returns.
2300
2301 2005-08-25  Alan Modra  <amodra@bigpond.net.au>
2302
2303         PR target/23404
2304         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Correct placement of stack
2305         homed fp args.
2306         (ffi_status ffi_prep_cif_machdep): Correct stack sizing for same.
2307
2308 2005-08-11  Jakub Jelinek  <jakub@redhat.com>
2309
2310         * configure.ac (HAVE_HIDDEN_VISIBILITY_ATTRIBUTE): New test.
2311         (AH_BOTTOM): Add FFI_HIDDEN definition.
2312         * configure: Rebuilt.
2313         * fficonfig.h.in: Rebuilt.
2314         * src/powerpc/ffi.c (hidden): Remove.
2315         (ffi_closure_LINUX64, ffi_prep_args64, ffi_call_LINUX64,
2316         ffi_closure_helper_LINUX64): Use FFI_HIDDEN instead of hidden.
2317         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64,
2318         .ffi_closure_LINUX64): Use FFI_HIDDEN instead of .hidden.
2319         * src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV): Remove,
2320         add FFI_HIDDEN to its prototype.
2321         (ffi_closure_SYSV_inner): New.
2322         * src/x86/sysv.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2323         * src/x86/win32.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2324
2325 2005-08-10  Alfred M. Szmidt  <ams@gnu.org>
2326
2327         PR libffi/21819:
2328         * configure: Rebuilt.
2329         * configure.ac: Handle i*86-*-gnu*.
2330
2331 2005-08-09  Jakub Jelinek  <jakub@redhat.com>
2332
2333         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Use
2334         DW_CFA_offset_extended_sf rather than
2335         DW_CFA_GNU_negative_offset_extended.
2336         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise.
2337
2338 2005-07-22  SUGIOKA Toshinobu  <sugioka@itonet.co.jp>
2339
2340         * src/sh/sysv.S (ffi_call_SYSV): Stop argument popping correctly
2341         on sh3.
2342         (ffi_closure_SYSV): Change the stack layout for sh3 struct argument.
2343         * src/sh/ffi.c (ffi_prep_args): Fix sh3 argument copy, when it is
2344         partially on register.
2345         (ffi_closure_helper_SYSV): Likewise.
2346         (ffi_prep_cif_machdep): Don't set too many cif->flags.
2347
2348 2005-07-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2349
2350         * src/sh/ffi.c (ffi_call): Handle small structures correctly.
2351         Remove empty line.
2352         * src/sh64/ffi.c (simple_type): Remove.
2353         (return_type): Handle small structures correctly.
2354         (ffi_prep_args): Likewise.
2355         (ffi_call): Likewise.
2356         (ffi_closure_helper_SYSV): Likewise.
2357         * src/sh64/sysv.S (ffi_call_SYSV): Handle 1, 2 and 4-byte return.
2358         Emit position independent code if PIC and remove wrong datalabel
2359         prefixes from EH data.
2360
2361 2005-07-19  Andreas Tobler  <a.tobler@schweiz.ch>
2362
2363         * Makefile.am (nodist_libffi_la_SOURCES): Add POWERPC_FREEBSD.
2364         * Makefile.in: Regenerate.
2365         * include/Makefile.in: Likewise.
2366         * testsuite/Makefile.in: Likewise.
2367         * configure.ac: Add POWERPC_FREEBSD rules.
2368         * configure: Regenerate.
2369         * src/powerpc/ffitarget.h: Add POWERPC_FREEBSD rules.
2370         (FFI_SYSV_TYPE_SMALL_STRUCT): Define.
2371         * src/powerpc/ffi.c: Add flags to handle small structure returns
2372         in ffi_call_SYSV.
2373         (ffi_prep_cif_machdep): Handle small structures for SYSV 4 ABI.
2374         Aka FFI_SYSV.
2375         (ffi_closure_helper_SYSV): Likewise.
2376         * src/powerpc/ppc_closure.S: Add return types for small structures.
2377         * src/powerpc/sysv.S: Add bits to handle small structures for
2378         final SYSV 4 ABI.
2379
2380 2005-07-10  Andreas Tobler  <a.tobler@schweiz.ch>
2381
2382         * testsuite/libffi.call/cls_5_1_byte.c: New test file.
2383         * testsuite/libffi.call/cls_6_1_byte.c: Likewise.
2384         * testsuite/libffi.call/cls_7_1_byte.c: Likewise.
2385
2386 2005-07-05  Randolph Chung  <tausq@debian.org>
2387
2388         * src/pa/ffi.c (ffi_struct_type): Rename FFI_TYPE_SMALL_STRUCT1
2389         as FFI_TYPE_SMALL_STRUCT3.  Break out handling for 5-7 byte
2390         structures.  Kill compilation warnings.
2391         (ffi_closure_inner_LINUX): Print return values as hex in debug
2392         message.  Rename FFI_TYPE_SMALL_STRUCT1 as FFI_TYPE_SMALL_STRUCT3.
2393         Properly handle 5-7 byte structure returns.
2394         * src/pa/ffitarget.h (FFI_TYPE_SMALL_STRUCT1)
2395         (FFI_TYPE_SMALL_STRUCT2): Remove.
2396         (FFI_TYPE_SMALL_STRUCT3, FFI_TYPE_SMALL_STRUCT5)
2397         (FFI_TYPE_SMALL_STRUCT6, FFI_TYPE_SMALL_STRUCT7): Define.
2398         * src/pa/linux.S: Mark source file as using PA1.1 assembly.
2399         (checksmst1, checksmst2): Remove.
2400         (checksmst3): Optimize handling of 3-byte struct returns.
2401         (checksmst567): Properly handle 5-7 byte struct returns.
2402
2403 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2404
2405         PR libgcj/21943
2406         * src/mips/n32.S: Enforce PIC code.
2407         * src/mips/o32.S: Likewise.
2408
2409 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2410
2411         * configure.ac: Treat i*86-*-solaris2.10 and up as X86_64.
2412         * configure: Regenerate.
2413
2414 2005-06-01  Alan Modra  <amodra@bigpond.net.au>
2415
2416         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't use JUMPTARGET
2417         to call ffi_closure_helper_SYSV.  Append @local instead.
2418         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise for ffi_prep_args_SYSV.
2419
2420 2005-05-17  Kelley Cook  <kcook@gcc.gnu.org>
2421
2422         * configure.ac: Use AC_C_BIGENDIAN instead of AC_C_BIGENDIAN_CROSS.
2423         Use AC_CHECK_SIZEOF instead of AC_COMPILE_CHECK_SIZEOF.
2424         * Makefile.am (ACLOCAL_AMFLAGS): Remove -I ../config.
2425         * aclocal.m4, configure, fficonfig.h.in, Makefile.in,
2426         include/Makefile.in, testsuite/Makefile.in: Regenerate.
2427
2428 2005-05-09  Mike Stump  <mrs@apple.com>
2429
2430         * configure: Regenerate.
2431
2432 2005-05-08  Richard Henderson  <rth@redhat.com>
2433
2434         PR libffi/21285
2435         * src/alpha/osf.S: Update unwind into to match code.
2436
2437 2005-05-04  Andreas Degert <ad@papyrus-gmbh.de>
2438             Richard Henderson  <rth@redhat.com>
2439
2440         * src/x86/ffi64.c (ffi_prep_cif_machdep): Save sse-used flag in
2441         bit 11 of flags.
2442         (ffi_call): Mask return type field.  Pass ssecount to ffi_call_unix64.
2443         (ffi_prep_closure): Set carry bit if sse-used flag set.
2444         * src/x86/unix64.S (ffi_call_unix64): Add ssecount argument.
2445         Only load sse registers if ssecount non-zero.
2446         (ffi_closure_unix64): Only save sse registers if carry set on entry.
2447
2448 2005-04-29  Ralf Corsepius  <ralf.corsepius@rtems.org>
2449
2450         * configure.ac: Add i*86-*-rtems*, sparc*-*-rtems*,
2451         powerpc-*rtems*, arm*-*-rtems*, sh-*-rtems*.
2452         * configure: Regenerate.
2453
2454 2005-04-20  Hans-Peter Nilsson  <hp@axis.com>
2455
2456         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): In regsub use,
2457         have Tcl8.3-compatible intermediate variable.
2458
2459 2005-04-18  Simon Posnjak <simon.posnjak@siol.net>
2460             Hans-Peter Nilsson  <hp@axis.com>
2461
2462         * Makefile.am: Add CRIS support.
2463         * configure.ac: Likewise.
2464         * Makefile.in, configure, testsuite/Makefile.in,
2465         include/Makefile.in: Regenerate.
2466         * src/cris: New directory.
2467         * src/cris/ffi.c, src/cris/sysv.S, src/cris/ffitarget.h: New files.
2468         * src/prep_cif.c (ffi_prep_cif): Wrap in #ifndef __CRIS__.
2469
2470         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): Replace \n with
2471         \r?\n in output tests.
2472
2473 2005-04-12  Mike Stump  <mrs@apple.com>
2474
2475         * configure: Regenerate.
2476
2477 2005-03-30  Hans Boehm  <Hans.Boehm@hp.com>
2478
2479         * src/ia64/ffitarget.h (ffi_arg): Use long long instead of DI.
2480
2481 2005-03-30  Steve Ellcey  <sje@cup.hp.com>
2482
2483         * src/ia64/ffitarget.h (ffi_arg) ADD DI attribute.
2484         (ffi_sarg) Ditto.
2485         * src/ia64/unix.S (ffi_closure_unix): Extend gp
2486         to 64 bits in ILP32 mode.
2487         Load 64 bits even for short data.
2488
2489 2005-03-23  Mike Stump  <mrs@apple.com>
2490
2491         * src/powerpc/darwin.S: Update for -m64 multilib.
2492         * src/powerpc/darwin_closure.S: Likewise.
2493
2494 2005-03-21  Zack Weinberg  <zack@codesourcery.com>
2495
2496         * configure.ac: Do not invoke TL_AC_GCC_VERSION.
2497         Do not set tool_include_dir.
2498         * aclocal.m4, configure, Makefile.in, testsuite/Makefile.in:
2499         Regenerate.
2500         * include/Makefile.am: Set gcc_version and toollibffidir.
2501         * include/Makefile.in: Regenerate.
2502
2503 2005-02-22  Andrew Haley  <aph@redhat.com>
2504
2505         * src/powerpc/ffi.c (ffi_prep_cif_machdep): Bump alignment to
2506         odd-numbered register pairs for 64-bit integer types.
2507
2508 2005-02-23  Andreas Tobler  <a.tobler@schweiz.ch>
2509
2510         PR libffi/20104
2511         * testsuite/libffi.call/return_ll1.c: New test case.
2512
2513 2005-02-11  Janis Johnson  <janis187@us.ibm.com>
2514
2515         * testsuite/libffi.call/cls_align_longdouble.c: Remove dg-options.
2516         * testsuite/libffi.call/float.c: Ditto.
2517         * testsuite/libffi.call/float2.c: Ditto.
2518         * testsuite/libffi.call/float3.c: Ditto.
2519
2520 2005-02-08  Andreas Tobler  <a.tobler@schweiz.ch>
2521
2522         * src/frv/ffitarget.h: Remove PPC stuff which does not belong to frv.
2523
2524 2005-01-12  Eric Botcazou  <ebotcazou@libertysurf.fr>
2525
2526         * testsuite/libffi.special/special.exp (cxx_options): Add
2527         -shared-libgcc.
2528
2529 2004-12-31  Richard Henderson  <rth@redhat.com>
2530
2531         * src/types.c (FFI_AGGREGATE_TYPEDEF): Remove.
2532         (FFI_TYPEDEF): Rename from FFI_INTEGRAL_TYPEDEF.  Replace size and
2533         offset parameters with a type parameter; deduce size and structure
2534         alignment.  Update all users.
2535
2536 2004-12-31  Richard Henderson  <rth@redhat.com>
2537
2538         * src/types.c (FFI_TYPE_POINTER): Define with sizeof.
2539         (FFI_TYPE_LONGDOUBLE): Fix for ia64.
2540         * src/ia64/ffitarget.h (struct ffi_ia64_trampoline_struct): Move
2541         into ffi_prep_closure.
2542         * src/ia64/ia64_flags.h, src/ia64/ffi.c, src/ia64/unix.S: Rewrite
2543         from scratch.
2544
2545 2004-12-27  Richard Henderson  <rth@redhat.com>
2546
2547         * src/x86/unix64.S: Fix typo in unwind info.
2548
2549 2004-12-25  Richard Henderson  <rth@redhat.com>
2550
2551         * src/x86/ffi64.c (struct register_args): Rename from stackLayout.
2552         (enum x86_64_reg_class): Add X86_64_COMPLEX_X87_CLASS.
2553         (merge_classes): Check for it.
2554         (SSE_CLASS_P): New.
2555         (classify_argument): Pass byte_offset by value; perform all updates
2556         inside struct case.
2557         (examine_argument): Add classes argument; handle
2558         X86_64_COMPLEX_X87_CLASS.
2559         (ffi_prep_args): Merge into ...
2560         (ffi_call): ... here.  Share stack frame with ffi_call_unix64.
2561         (ffi_prep_cif_machdep): Setup cif->flags for proper structure return.
2562         (ffi_fill_return_value): Remove.
2563         (ffi_prep_closure): Remove dead assert.
2564         (ffi_closure_unix64_inner): Rename from ffi_closure_UNIX64_inner.
2565         Rewrite to use struct register_args instead of va_list.  Create
2566         flags for handling structure returns.
2567         * src/x86/unix64.S: Remove dead strings.
2568         (ffi_call_unix64): Rename from ffi_call_UNIX64.  Rewrite to share
2569         stack frame with ffi_call.  Handle structure returns properly.
2570         (float2sse, floatfloat2sse, double2sse): Remove.
2571         (sse2float, sse2double, sse2floatfloat): Remove.
2572         (ffi_closure_unix64): Rename from ffi_closure_UNIX64.  Rewrite
2573         to handle structure returns properly.
2574
2575 2004-12-08  David Edelsohn  <edelsohn@gnu.org>
2576
2577         * Makefile.am (AM_MAKEFLAGS): Remove duplicate LIBCFLAGS and
2578         PICFLAG.
2579         * Makefile.in: Regenerated.
2580
2581 2004-12-02  Richard Sandiford  <rsandifo@redhat.com>
2582
2583         * configure.ac: Use TL_AC_GCC_VERSION to set gcc_version.
2584         * configure, aclocal.m4, Makefile.in: Regenerate.
2585         * include/Makefile.in, testsuite/Makefile.in: Regenerate.
2586
2587 2004-11-29  Kelley Cook  <kcook@gcc.gnu.org>
2588
2589         * configure: Regenerate for libtool change.
2590
2591 2004-11-25  Kelley Cook  <kcook@gcc.gnu.org>
2592
2593         * configure: Regenerate for libtool reversion.
2594
2595 2004-11-24  Kelley Cook  <kcook@gcc.gnu.org>
2596
2597         * configure: Regenerate for libtool change.
2598
2599 2004-11-23  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2600
2601         * testsuite/lib/libffi-dg.exp: Use new procs in target-libpath.exp.
2602
2603 2004-11-23  Richard Sandiford  <rsandifo@redhat.com>
2604
2605         * src/mips/o32.S (ffi_call_O32, ffi_closure_O32): Use jalr instead
2606         of jal.  Use an absolute encoding for the frame information.
2607
2608 2004-11-23  Kelley Cook  <kcook@gcc.gnu.org>
2609
2610         * Makefile.am: Remove no-dependencies.  Add ACLOCAL_AMFLAGS.
2611         * acinclude.m4: Delete logic for sincludes.
2612         * aclocal.m4, Makefile.in, configure: Regenerate.
2613         * include/Makefile: Likewise.
2614         * testsuite/Makefile: Likewise.
2615
2616 2004-11-22  Eric Botcazou  <ebotcazou@libertysurf.fr>
2617
2618         * src/sparc/ffi.c (ffi_prep_closure): Align doubles and 64-bit integers
2619         on a 8-byte boundary.
2620         * src/sparc/v8.S (ffi_closure_v8): Reserve frame space for arguments.
2621
2622 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2623
2624         * src/arm/ffi.c (ffi_prep_cif_machdep): Handle functions that return
2625         long long values.  Round stack allocation to a multiple of 8 bytes
2626         for ATPCS compatibility.
2627         * src/arm/sysv.S (ffi_call_SYSV): Rework to avoid use of APCS register
2628         names.  Handle returning long long types.  Add Thumb and interworking
2629         support.  Improve soft-float code.
2630
2631 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2632
2633         * testsuite/lib/libffi-db.exp (load_gcc_lib): New function.
2634         (libffi_exit): New function.
2635         (libffi_init): Build the testglue wrapper if needed.
2636
2637 2004-10-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2638
2639         PR other/18138
2640         * testsuite/lib/libffi-dg.exp: Accept more than one multilib libgcc.
2641
2642 2004-10-25  Kazuhiro Inaoka <inaoka.kazuhiro@renesas.com>
2643
2644         * src/m32r/libffitarget.h (FFI_CLOSURES): Set to 0.
2645
2646 2004-10-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2647
2648         * src/sh/sysv.S (ffi_call_SYSV): Don't align for double data.
2649         * testsuite/libffi.call/float3.c: New test case.
2650
2651 2004-10-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2652
2653         * src/sh/ffi.c (ffi_prep_closure): Set T bit in trampoline for
2654         the function returning a structure pointed with R2.
2655         * src/sh/sysv.S (ffi_closure_SYSV): Use R2 as the pointer to
2656         the structure return value if T bit set.  Emit position
2657         independent code and EH data if PIC.
2658
2659 2004-10-13  Kazuhiro Inaoka  <inaoka.kazuhiro@renesas.com>
2660
2661         * Makefile.am: Add m32r support.
2662         * configure.ac: Likewise.
2663         * Makefile.in: Regenerate.
2664         * confiugre: Regenerate.
2665         * src/types.c: Add m32r port to FFI_INTERNAL_TYPEDEF
2666         (uint64, sint64, double, longdouble)
2667         * src/m32r: New directory.
2668         * src/m32r/ffi.c: New file.
2669         * src/m32r/sysv.S: Likewise.
2670         * src/m32r/ffitarget.h: Likewise.
2671
2672 2004-10-02  Kaz Kojima  <kkojima@gcc.gnu.org>
2673
2674         * testsuite/libffi.call/negint.c: New test case.
2675
2676 2004-09-14  H.J. Lu  <hongjiu.lu@intel.com>
2677
2678         PR libgcj/17465
2679         * testsuite/lib/libffi-dg.exp: Don't use global ld_library_path.
2680         Set up LD_LIBRARY_PATH, SHLIB_PATH, LD_LIBRARYN32_PATH,
2681         LD_LIBRARY64_PATH, LD_LIBRARY_PATH_32, LD_LIBRARY_PATH_64 and
2682         DYLD_LIBRARY_PATH.
2683
2684 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2685
2686         * testsuite/libffi.call/many_win32.c: Remove whitespaces.
2687         * testsuite/libffi.call/promotion.c: Likewise.
2688         * testsuite/libffi.call/return_ll.c: Remove unused var. Cleanup
2689         whitespaces.
2690         * testsuite/libffi.call/return_sc.c: Likewise.
2691         * testsuite/libffi.call/return_uc.c: Likewise.
2692
2693 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2694
2695         * src/powerpc/darwin.S: Fix comments and identation.
2696         * src/powerpc/darwin_closure.S: Likewise.
2697
2698 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2699
2700         * src/powerpc/ffi_darwin.c: Add flag for longdouble return values.
2701         (ffi_prep_args): Handle longdouble arguments.
2702         (ffi_prep_cif_machdep): Set flags for longdouble. Calculate space for
2703         longdouble.
2704         (ffi_closure_helper_DARWIN): Add closure handling for longdouble.
2705         * src/powerpc/darwin.S (_ffi_call_DARWIN): Add handling of longdouble
2706         values.
2707         * src/powerpc/darwin_closure.S (_ffi_closure_ASM): Likewise.
2708         * src/types.c: Defined longdouble size and alignment for darwin.
2709
2710 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2711
2712         * src/powerpc/aix.S: Remove whitespaces.
2713         * src/powerpc/aix_closure.S: Likewise.
2714         * src/powerpc/asm.h: Likewise.
2715         * src/powerpc/ffi.c: Likewise.
2716         * src/powerpc/ffitarget.h: Likewise.
2717         * src/powerpc/linux64.S: Likewise.
2718         * src/powerpc/linux64_closure.S: Likewise.
2719         * src/powerpc/ppc_closure.S: Likewise.
2720         * src/powerpc/sysv.S: Likewise.
2721
2722 2004-08-30  Anthony Green  <green@redhat.com>
2723
2724         * Makefile.am: Add frv support.
2725         * Makefile.in, testsuite/Makefile.in: Rebuilt.
2726         * configure.ac: Read configure.host.
2727         * configure.in: Read configure.host.
2728         * configure.host: New file.  frv-elf needs libgloss.
2729         * include/ffi.h.in: Force ffi_closure to have a nice big (8)
2730         alignment.  This is needed to frv and shouldn't harm the others.
2731         * include/ffi_common.h (ALIGN_DOWN): New macro.
2732         * src/frv/ffi.c, src/frv/ffitarget.h, src/frv/eabi.S: New files.
2733
2734 2004-08-24  David Daney  <daney@avtrex.com>
2735
2736         * testsuite/libffi.call/closure_fn0.c: Xfail mips64* instead of mips*.
2737         * testsuite/libffi.call/closure_fn1.c: Likewise.
2738         * testsuite/libffi.call/closure_fn2.c  Likewise.
2739         * testsuite/libffi.call/closure_fn3.c: Likewise.
2740         * testsuite/libffi.call/closure_fn4.c: Likewise.
2741         * testsuite/libffi.call/closure_fn5.c: Likewise.
2742         * testsuite/libffi.call/cls_18byte.c: Likewise.
2743         * testsuite/libffi.call/cls_19byte.c: Likewise.
2744         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2745         * testsuite/libffi.call/cls_20byte.c: Likewise.
2746         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2747         * testsuite/libffi.call/cls_24byte.c: Likewise.
2748         * testsuite/libffi.call/cls_2byte.c: Likewise.
2749         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2750         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2751         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2752         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2753         * testsuite/libffi.call/cls_4byte.c: Likewise.
2754         * testsuite/libffi.call/cls_64byte.c: Likewise.
2755         * testsuite/libffi.call/cls_6byte.c: Likewise.
2756         * testsuite/libffi.call/cls_7byte.c: Likewise.
2757         * testsuite/libffi.call/cls_8byte.c: Likewise.
2758         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2759         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2760         * testsuite/libffi.call/cls_align_double.c: Likewise.
2761         * testsuite/libffi.call/cls_align_float.c: Likewise.
2762         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2763         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2764         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2765         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2766         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2767         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2768         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2769         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
2770         * testsuite/libffi.call/cls_double.c: Likewise.
2771         * testsuite/libffi.call/cls_float.c: Likewise.
2772         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2773         * testsuite/libffi.call/cls_multi_sshort.c: Likewise.
2774         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2775         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2776         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2777         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2778         * testsuite/libffi.call/cls_schar.c: Likewise.
2779         * testsuite/libffi.call/cls_sint.c: Likewise.
2780         * testsuite/libffi.call/cls_sshort.c: Likewise.
2781         * testsuite/libffi.call/cls_uchar.c: Likewise.
2782         * testsuite/libffi.call/cls_uint.c: Likewise.
2783         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2784         * testsuite/libffi.call/cls_ushort.c: Likewise.
2785         * testsuite/libffi.call/nested_struct.c: Likewise.
2786         * testsuite/libffi.call/nested_struct1.c: Likewise.
2787         * testsuite/libffi.call/nested_struct2.c: Likewise.
2788         * testsuite/libffi.call/nested_struct3.c: Likewise.
2789         * testsuite/libffi.call/problem1.c: Likewise.
2790         * testsuite/libffi.special/unwindtest.cc: Likewise.
2791         * testsuite/libffi.call/cls_12byte.c: Likewise and set return value
2792         to zero.
2793         * testsuite/libffi.call/cls_16byte.c: Likewise.
2794         * testsuite/libffi.call/cls_5byte.c: Likewise.
2795
2796 2004-08-23  David Daney <daney@avtrex.com>
2797
2798         PR libgcj/13141
2799         * src/mips/ffitarget.h (FFI_O32_SOFT_FLOAT): New ABI.
2800         * src/mips/ffi.c (ffi_prep_args): Fix alignment calculation.
2801         (ffi_prep_cif_machdep): Handle FFI_O32_SOFT_FLOAT floating point
2802         parameters and return types.
2803         (ffi_call): Handle FFI_O32_SOFT_FLOAT ABI.
2804         (ffi_prep_closure): Ditto.
2805         (ffi_closure_mips_inner_O32): Handle FFI_O32_SOFT_FLOAT ABI, fix
2806         alignment calculations.
2807         * src/mips/o32.S (ffi_closure_O32): Don't use floating point
2808         instructions if FFI_O32_SOFT_FLOAT, make stack frame ABI compliant.
2809
2810 2004-08-14  Casey Marshall <csm@gnu.org>
2811
2812         * src/mips/ffi.c (ffi_pref_cif_machdep): set `cif->flags' to
2813         contain `FFI_TYPE_UINT64' as return type for any 64-bit
2814         integer (O32 ABI only).
2815         (ffi_prep_closure): new function.
2816         (ffi_closure_mips_inner_O32): new function.
2817         * src/mips/ffitarget.h: Define `FFI_CLOSURES' and
2818         `FFI_TRAMPOLINE_SIZE' appropriately if the ABI is o32.
2819         * src/mips/o32.S (ffi_call_O32): add labels for .eh_frame. Return
2820         64 bit integers correctly.
2821         (ffi_closure_O32): new function.
2822         Added DWARF-2 unwind info for both functions.
2823
2824 2004-08-10  Andrew Haley  <aph@redhat.com>
2825
2826         * src/x86/ffi64.c (ffi_prep_args ): 8-align all stack arguments.
2827
2828 2004-08-01  Robert Millan  <robertmh@gnu.org>
2829
2830         * configure.ac: Detect knetbsd-gnu and kfreebsd-gnu.
2831         * configure: Regenerate.
2832
2833 2004-07-30  Maciej W. Rozycki  <macro@linux-mips.org>
2834
2835         * acinclude.m4 (AC_FUNC_MMAP_BLACKLIST): Check for <sys/mman.h>
2836         and mmap() explicitly instead of relying on preset autoconf cache
2837         variables.
2838         * aclocal.m4: Regenerate.
2839         * configure: Regenerate.
2840
2841 2004-07-11  Ulrich Weigand  <uweigand@de.ibm.com>
2842
2843         * src/s390/ffi.c (ffi_prep_args): Fix C aliasing violation.
2844         (ffi_check_float_struct): Remove unused prototype.
2845
2846 2004-06-30  Geoffrey Keating  <geoffk@apple.com>
2847
2848         * src/powerpc/ffi_darwin.c (flush_icache): ';' is a comment
2849         character on Darwin, use '\n\t' instead.
2850
2851 2004-06-26  Matthias Klose  <doko@debian.org>
2852
2853         * libtool-version: Fix typo in revision/age.
2854
2855 2004-06-17  Matthias Klose  <doko@debian.org>
2856
2857         * libtool-version: New.
2858         * Makefile.am (libffi_la_LDFLAGS): Use -version-info for soname.
2859         * Makefile.in: Regenerate.
2860
2861 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2862
2863         * Makefile.am: Remove useless multilib rules.
2864         * Makefile.in: Regenerate.
2865         * aclocal.m4: Regenerate with automake 1.8.5.
2866         * configure.ac: Remove useless multilib configury.
2867         * configure: Regenerate.
2868
2869 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2870
2871         * .cvsignore: New file.
2872
2873 2004-06-10  Jakub Jelinek  <jakub@redhat.com>
2874
2875         * src/ia64/unix.S (ffi_call_unix): Insert group barrier break
2876         fp_done.
2877         (ffi_closure_UNIX): Fix f14/f15 adjustment if FLOAT_SZ is ever
2878         changed from 8.
2879
2880 2004-06-06  Sean McNeil  <sean@mcneil.com>
2881
2882         * configure.ac: Add x86_64-*-freebsd* support.
2883         * configure: Regenerate.
2884
2885 2004-04-26  Joe Buck <jbuck@welsh-buck.org>
2886
2887         Bug 15093
2888         * configure.ac: Test for existence of mmap and sys/mman.h before
2889         checking blacklist.  Fix suggested by Jim Wilson.
2890         * configure: Regenerate.
2891
2892 2004-04-26  Matt Austern  <austern@apple.com>
2893
2894         * src/powerpc/darwin.S: Go through a non-lazy pointer for initial
2895         FDE location.
2896         * src/powerpc/darwin_closure.S: Likewise.
2897
2898 2004-04-24  Andreas Tobler  <a.tobler@schweiz.ch>
2899
2900         * testsuite/libffi.call/cls_multi_schar.c (main): Fix initialization
2901         error. Reported by Thomas Heller <theller@python.net>.
2902         * testsuite/libffi.call/cls_multi_sshort.c (main): Likewise.
2903         * testsuite/libffi.call/cls_multi_ushort.c (main): Likewise.
2904
2905 2004-03-20  Matthias Klose  <doko@debian.org>
2906
2907         * src/pa/linux.S: Fix typo.
2908
2909 2004-03-19  Matthias Klose  <doko@debian.org>
2910
2911         * Makefile.am: Update.
2912         * Makefile.in: Regenerate.
2913         * src/pa/ffi.h.in: Remove.
2914         * src/pa/ffitarget.h: New file.
2915
2916 2004-02-10  Randolph Chung  <tausq@debian.org>
2917
2918         * Makefile.am: Add PA support.
2919         * Makefile.in: Regenerate.
2920         * include/Makefile.in: Regenerate.
2921         * configure.ac: Add PA target.
2922         * configure: Regenerate.
2923         * src/pa/ffi.c: New file.
2924         * src/pa/ffi.h.in: Add PA support.
2925         * src/pa/linux.S: New file.
2926         * prep_cif.c: Add PA support.
2927
2928 2004-03-16  Hosaka Yuji  <hos@tamanegi.org>
2929
2930         * src/types.c: Fix alignment size of X86_WIN32 case int64 and
2931         double.
2932         * src/x86/ffi.c (ffi_prep_args): Replace ecif->cif->rtype->type
2933         with ecif->cif->flags.
2934         (ffi_call, ffi_prep_incoming_args_SYSV): Replace cif->rtype->type
2935         with cif->flags.
2936         (ffi_prep_cif_machdep): Add X86_WIN32 struct case.
2937         (ffi_closure_SYSV): Add 1 or 2-bytes struct case for X86_WIN32.
2938         * src/x86/win32.S (retstruct1b, retstruct2b, sc_retstruct1b,
2939         sc_retstruct2b): Add for 1 or 2-bytes struct case.
2940
2941 2004-03-15 Kelley Cook <kcook@gcc.gnu.org>
2942
2943         * configure.in: Rename file to ...
2944         * configure.ac: ... this.
2945         * fficonfig.h.in: Regenerate.
2946         * Makefile.in: Regenerate.
2947         * include/Makefile.in: Regenerate.
2948         * testsuite/Makefile.in: Regenerate.
2949
2950 2004-03-12  Matt Austern  <austern@apple.com>
2951
2952         * src/powerpc/darwin.S: Fix EH information so it corresponds to
2953         changes in EH format resulting from addition of linkonce support.
2954         * src/powerpc/darwin_closure.S: Likewise.
2955
2956 2004-03-11  Andreas Tobler  <a.tobler@schweiz.ch>
2957             Paolo Bonzini  <bonzini@gnu.org>
2958
2959         * Makefile.am (AUTOMAKE_OPTIONS): Set them.
2960         Remove VPATH. Remove rules for object files. Remove multilib support.
2961         (AM_CCASFLAGS): Add.
2962         * configure.in (AC_CONFIG_HEADERS): Relace AM_CONFIG_HEADER.
2963         (AC_PREREQ): Bump version to 2.59.
2964         (AC_INIT): Fill with version info and bug address.
2965         (ORIGINAL_LD_FOR_MULTILIBS): Remove.
2966         (AM_ENABLE_MULTILIB): Use this instead of AC_ARG_ENABLE.
2967         De-precious CC so that the right flags are passed down to multilibs.
2968         (AC_MSG_ERROR): Replace obsolete macro AC_ERROR.
2969         (AC_CONFIG_FILES): Replace obsolete macro AC_LINK_FILES.
2970         (AC_OUTPUT): Reorganize the output with AC_CONFIG_COMMANDS.
2971         * configure: Rebuilt.
2972         * aclocal.m4: Likewise.
2973         * Makefile.in, include/Makefile.in, testsuite/Makefile.in: Likewise.
2974         * fficonfig.h.in: Likewise.
2975
2976 2004-03-11  Andreas Schwab  <schwab@suse.de>
2977
2978         * src/ia64/ffi.c (ffi_prep_incoming_args_UNIX): Get floating point
2979         arguments from fp registers only for the first 8 parameter slots.
2980         Don't convert a float parameter when passed in memory.
2981
2982 2004-03-09  Hans-Peter Nilsson  <hp@axis.com>
2983
2984         * configure: Regenerate for config/accross.m4 correction.
2985
2986 2004-02-25  Matt Kraai  <kraai@alumni.cmu.edu>
2987
2988         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Change
2989         ecif->cif->bytes to bytes.
2990         (ffi_prep_cif_machdep): Add braces around nested if statement.
2991
2992 2004-02-09  Alan Modra  <amodra@bigpond.net.au>
2993
2994         * src/types.c (pointer): POWERPC64 has 8 byte pointers.
2995
2996         * src/powerpc/ffi.c (ffi_prep_args64): Correct long double handling.
2997         (ffi_closure_helper_LINUX64): Fix typo.
2998         * testsuite/libffi.call/cls_align_longdouble.c: Pass -mlong-double-128
2999         for powerpc64-*-*.
3000         * testsuite/libffi.call/float.c: Likewise.
3001         * testsuite/libffi.call/float2.c: Likewise.
3002
3003 2004-02-08  Alan Modra  <amodra@bigpond.net.au>
3004
3005         * src/powerpc/ffi.c (ffi_prep_cif_machdep <FFI_LINUX64>): Correct
3006         long double function return and long double arg handling.
3007         (ffi_closure_helper_LINUX64): Formatting.  Delete unused "ng" var.
3008         Use "end_pfr" instead of "nf".  Correct long double handling.
3009         Localise "temp".
3010         * src/powerpc/linux64.S (ffi_call_LINUX64): Save f2 long double
3011         return value.
3012         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Allocate
3013         space for long double return value.  Adjust stack frame and offsets.
3014         Load f2 long double return.
3015
3016 2004-02-07  Alan Modra  <amodra@bigpond.net.au>
3017
3018         * src/types.c: Use 16 byte long double for POWERPC64.
3019
3020 2004-01-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
3021
3022         * src/sparc/ffi.c (ffi_prep_args_v9): Shift the parameter array
3023         when the structure return address is passed in %o0.
3024         (ffi_V9_return_struct): Rename into ffi_v9_layout_struct.
3025         (ffi_v9_layout_struct): Align the field following a nested structure
3026         on a word boundary.  Use memmove instead of memcpy.
3027         (ffi_call): Update call to ffi_V9_return_struct.
3028         (ffi_prep_closure): Define 'ctx' only for V8.
3029         (ffi_closure_sparc_inner): Clone into ffi_closure_sparc_inner_v8
3030         and ffi_closure_sparc_inner_v9.
3031         (ffi_closure_sparc_inner_v8): Return long doubles by reference.
3032         Always skip the structure return address.  For structures and long
3033         doubles, copy the argument directly.
3034         (ffi_closure_sparc_inner_v9): Skip the structure return address only
3035         if required.  Shift the maximum floating-point slot accordingly.  For
3036         big structures, copy the argument directly; otherwise, left-justify the
3037         argument and call ffi_v9_layout_struct to lay out the structure on
3038         the stack.
3039         * src/sparc/v8.S: Undef STACKFRAME before defining it.
3040         (ffi_closure_v8): Pass the structure return address.  Update call to
3041         ffi_closure_sparc_inner_v8.  Short-circuit FFI_TYPE_INT handling.
3042         Skip the 'unimp' insn when returning long doubles and structures.
3043         * src/sparc/v9.S: Undef STACKFRAME before defining it.
3044         (ffi_closure_v9): Increase the frame size by 2 words.  Short-circuit
3045         FFI_TYPE_INT handling.  Load structures both in integers and
3046         floating-point registers on return.
3047         * README: Update status of the SPARC port.
3048
3049 2004-01-24  Andreas Tobler  <a.tobler@schweiz.ch>
3050
3051         * testsuite/libffi.call/pyobjc-tc.c (main): Treat result value
3052         as of type ffi_arg.
3053         * testsuite/libffi.call/struct3.c (main): Fix CHECK.
3054
3055 2004-01-22  Ulrich Weigand  <uweigand@de.ibm.com>
3056
3057         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Treat result
3058         value as of type ffi_arg, not unsigned int.
3059
3060 2004-01-21  Michael Ritzert  <ritzert@t-online.de>
3061
3062         * ffi64.c (ffi_prep_args): Cast the RHS of an assignment instead
3063         of the LHS.
3064
3065 2004-01-12  Andreas Tobler  <a.tobler@schweiz.ch>
3066
3067         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_32 for
3068         Solaris.
3069
3070 2004-01-08  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3071
3072         * testsuite/libffi.call/ffitest.h (allocate_mmap): Cast MAP_FAILED
3073         to void *.
3074
3075 2003-12-10  Richard Henderson  <rth@redhat.com>
3076
3077         * testsuite/libffi.call/cls_align_pointer.c: Cast pointers to
3078         size_t instead of int.
3079
3080 2003-12-04  Hosaka Yuji  <hos@tamanegi.org>
3081
3082         * testsuite/libffi.call/many_win32.c: Include <float.h>.
3083         * testsuite/libffi.call/many_win32.c (main): Replace variable
3084         int i with unsigned long ul.
3085
3086         * testsuite/libffi.call/cls_align_uint64.c: New test case.
3087         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
3088         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
3089         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
3090         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
3091         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
3092         * testsuite/libffi.call/cls_align_float.c: Likewise.
3093         * testsuite/libffi.call/cls_align_double.c: Likewise.
3094         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
3095         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
3096
3097 2003-12-02  Hosaka Yuji  <hos@tamanegi.org>
3098
3099         PR other/13221
3100         * src/x86/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV):
3101         Align arguments to 32 bits.
3102
3103 2003-12-01  Andreas Tobler  <a.tobler@schweiz.ch>
3104
3105         PR other/13221
3106         * testsuite/libffi.call/cls_multi_sshort.c: New test case.
3107         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
3108         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
3109         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
3110         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
3111         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
3112
3113         * testsuite/libffi.special/unwindtest.cc: Cosmetics.
3114
3115 2003-11-26  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
3116
3117         * testsuite/libffi.call/ffitest.h: Include <fcntl.h>.
3118         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3119
3120 2003-11-22  Andreas Tobler  <a.tobler@schweiz.ch>
3121
3122         * Makefile.in: Rebuilt.
3123         * configure: Likewise.
3124         * testsuite/libffi.special/unwindtest.cc: Convert the mmap to
3125         the right type.
3126
3127 2003-11-21  Andreas Jaeger  <aj@suse.de>
3128             Andreas Tobler  <a.tobler@schweiz.ch>
3129
3130         * acinclude.m4: Add AC_FUNC_MMAP_BLACKLIST.
3131         * configure.in: Call AC_FUNC_MMAP_BLACKLIST.
3132         * Makefile.in: Rebuilt.
3133         * aclocal.m4: Likewise.
3134         * configure: Likewise.
3135         * fficonfig.h.in: Likewise.
3136         * testsuite/lib/libffi-dg.exp: Add include dir.
3137         * testsuite/libffi.call/ffitest.h: Add MMAP definitions.
3138         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3139         * testsuite/libffi.call/closure_fn0.c: Use MMAP functionality
3140         for ffi_closure if available.
3141         * testsuite/libffi.call/closure_fn1.c: Likewise.
3142         * testsuite/libffi.call/closure_fn2.c: Likewise.
3143         * testsuite/libffi.call/closure_fn3.c: Likewise.
3144         * testsuite/libffi.call/closure_fn4.c: Likewise.
3145         * testsuite/libffi.call/closure_fn5.c: Likewise.
3146         * testsuite/libffi.call/cls_12byte.c: Likewise.
3147         * testsuite/libffi.call/cls_16byte.c: Likewise.
3148         * testsuite/libffi.call/cls_18byte.c: Likewise.
3149         * testsuite/libffi.call/cls_19byte.c: Likewise.
3150         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3151         * testsuite/libffi.call/cls_20byte.c: Likewise.
3152         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3153         * testsuite/libffi.call/cls_24byte.c: Likewise.
3154         * testsuite/libffi.call/cls_2byte.c: Likewise.
3155         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3156         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3157         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3158         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3159         * testsuite/libffi.call/cls_4byte.c: Likewise.
3160         * testsuite/libffi.call/cls_5byte.c: Likewise.
3161         * testsuite/libffi.call/cls_64byte.c: Likewise.
3162         * testsuite/libffi.call/cls_6byte.c: Likewise.
3163         * testsuite/libffi.call/cls_7byte.c: Likewise.
3164         * testsuite/libffi.call/cls_8byte.c: Likewise.
3165         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3166         * testsuite/libffi.call/cls_9byte2.c: Likewise.
3167         * testsuite/libffi.call/cls_double.c: Likewise.
3168         * testsuite/libffi.call/cls_float.c: Likewise.
3169         * testsuite/libffi.call/cls_schar.c: Likewise.
3170         * testsuite/libffi.call/cls_sint.c: Likewise.
3171         * testsuite/libffi.call/cls_sshort.c: Likewise.
3172         * testsuite/libffi.call/cls_uchar.c: Likewise.
3173         * testsuite/libffi.call/cls_uint.c: Likewise.
3174         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3175         * testsuite/libffi.call/cls_ushort.c: Likewise.
3176         * testsuite/libffi.call/nested_struct.c: Likewise.
3177         * testsuite/libffi.call/nested_struct1.c: Likewise.
3178         * testsuite/libffi.call/nested_struct2.c: Likewise.
3179         * testsuite/libffi.call/nested_struct3.c: Likewise.
3180         * testsuite/libffi.call/problem1.c: Likewise.
3181         * testsuite/libffi.special/unwindtest.cc: Likewise.
3182
3183 2003-11-20  Andreas Tobler  <a.tobler@schweiz.ch>
3184
3185         * testsuite/lib/libffi-dg.exp: Make the -lgcc_s conditional.
3186
3187 2003-11-19  Andreas Tobler  <a.tobler@schweiz.ch>
3188
3189         * testsuite/lib/libffi-dg.exp: Add DYLD_LIBRARY_PATH for darwin.
3190         Add -lgcc_s to additional flags.
3191
3192 2003-11-12  Andreas Tobler  <a.tobler@schweiz.ch>
3193
3194         * configure.in, include/Makefile.am: PR libgcj/11147, install
3195         the ffitarget.h header file in a gcc versioned and target
3196         dependent place.
3197         * configure: Regenerated.
3198         * Makefile.in, include/Makefile.in: Likewise.
3199         * testsuite/Makefile.in: Likewise.
3200
3201 2003-11-09  Andreas Tobler  <a.tobler@schweiz.ch>
3202
3203         * testsuite/libffi.call/closure_fn0.c: Print result and check
3204         with dg-output to make debugging easier.
3205         * testsuite/libffi.call/closure_fn1.c: Likewise.
3206         * testsuite/libffi.call/closure_fn2.c: Likewise.
3207         * testsuite/libffi.call/closure_fn3.c: Likewise.
3208         * testsuite/libffi.call/closure_fn4.c: Likewise.
3209         * testsuite/libffi.call/closure_fn5.c: Likewise.
3210         * testsuite/libffi.call/cls_12byte.c: Likewise.
3211         * testsuite/libffi.call/cls_16byte.c: Likewise.
3212         * testsuite/libffi.call/cls_18byte.c: Likewise.
3213         * testsuite/libffi.call/cls_19byte.c: Likewise.
3214         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3215         * testsuite/libffi.call/cls_20byte.c: Likewise.
3216         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3217         * testsuite/libffi.call/cls_24byte.c: Likewise.
3218         * testsuite/libffi.call/cls_2byte.c: Likewise.
3219         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3220         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3221         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3222         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3223         * testsuite/libffi.call/cls_4byte.c: Likewise.
3224         * testsuite/libffi.call/cls_5byte.c: Likewise.
3225         * testsuite/libffi.call/cls_64byte.c: Likewise.
3226         * testsuite/libffi.call/cls_6byte.c: Likewise.
3227         * testsuite/libffi.call/cls_7byte.c: Likewise.
3228         * testsuite/libffi.call/cls_8byte.c: Likewise.
3229         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3230         * testsuite/libffi.call/cls_9byte2.c: Likewise.
3231         * testsuite/libffi.call/cls_double.c: Likewise.
3232         * testsuite/libffi.call/cls_float.c: Likewise.
3233         * testsuite/libffi.call/cls_schar.c: Likewise.
3234         * testsuite/libffi.call/cls_sint.c: Likewise.
3235         * testsuite/libffi.call/cls_sshort.c: Likewise.
3236         * testsuite/libffi.call/cls_uchar.c: Likewise.
3237         * testsuite/libffi.call/cls_uint.c: Likewise.
3238         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3239         * testsuite/libffi.call/cls_ushort.c: Likewise.
3240         * testsuite/libffi.call/problem1.c: Likewise.
3241
3242         * testsuite/libffi.special/unwindtest.cc: Make ffi_closure
3243         static.
3244
3245 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
3246
3247         * testsuite/libffi.call/cls_9byte2.c: New test case.
3248         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3249         * testsuite/libffi.call/cls_64byte.c: Likewise.
3250         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3251         * testsuite/libffi.call/cls_19byte.c: Likewise.
3252         * testsuite/libffi.call/cls_18byte.c: Likewise.
3253         * testsuite/libffi.call/closure_fn4.c: Likewise.
3254         * testsuite/libffi.call/closure_fn5.c: Likewise.
3255         * testsuite/libffi.call/cls_schar.c: Likewise.
3256         * testsuite/libffi.call/cls_sint.c: Likewise.
3257         * testsuite/libffi.call/cls_sshort.c: Likewise.
3258         * testsuite/libffi.call/nested_struct2.c: Likewise.
3259         * testsuite/libffi.call/nested_struct3.c: Likewise.
3260
3261 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
3262
3263         * testsuite/libffi.call/cls_double.c: Do a check on the result.
3264         * testsuite/libffi.call/cls_uchar.c: Likewise.
3265         * testsuite/libffi.call/cls_uint.c: Likewise.
3266         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3267         * testsuite/libffi.call/cls_ushort.c: Likewise.
3268         * testsuite/libffi.call/return_sc.c: Cleanup whitespaces.
3269
3270 2003-11-06  Andreas Tobler  <a.tobler@schweiz.ch>
3271
3272         * src/prep_cif.c (ffi_prep_cif): Move the validity check after
3273         the initialization.
3274
3275 2003-10-23  Andreas Tobler  <a.tobler@schweiz.ch>
3276
3277         * src/java_raw_api.c (ffi_java_ptrarray_to_raw): Replace
3278         FFI_ASSERT(FALSE) with FFI_ASSERT(0).
3279
3280 2003-10-22  David Daney  <ddaney@avtrex.com>
3281
3282         * src/mips/ffitarget.h: Replace undefined UINT32 and friends with
3283         __attribute__((__mode__(__SI__))) and friends.
3284
3285 2003-10-22  Andreas Schwab  <schwab@suse.de>
3286
3287         * src/ia64/ffi.c: Replace FALSE/TRUE with false/true.
3288
3289 2003-10-21  Andreas Tobler  <a.tobler@schweiz.ch>
3290
3291         * configure.in: AC_LINK_FILES(ffitarget.h).
3292         * configure: Regenerate.
3293         * Makefile.in: Likewise.
3294         * include/Makefile.in: Likewise.
3295         * testsuite/Makefile.in: Likewise.
3296         * fficonfig.h.in: Likewise.
3297
3298 2003-10-21  Paolo Bonzini  <bonzini@gnu.org>
3299             Richard Henderson  <rth@redhat.com>
3300
3301         Avoid that ffi.h includes fficonfig.h.
3302
3303         * Makefile.am (EXTRA_DIST): Include ffitarget.h files
3304         (TARGET_SRC_MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3305         (TARGET_SRC_MIPS_SGI): Removed.
3306         (MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3307         (MIPS_SGI): Removed.
3308         (CLEANFILES): Removed.
3309         (mostlyclean-am, clean-am, mostlyclean-sub, clean-sub): New
3310         targets.
3311         * acconfig.h: Removed.
3312         * configure.in: Compute sizeofs only for double and long double.
3313         Use them to define and subst HAVE_LONG_DOUBLE.  Include comments
3314         into AC_DEFINE instead of using acconfig.h.  Create
3315         include/ffitarget.h instead of include/fficonfig.h.  Rename
3316         MIPS_GCC to MIPS_IRIX, drop MIPS_SGI since we are in gcc's tree.
3317         AC_DEFINE EH_FRAME_FLAGS.
3318         * include/Makefile.am (DISTCLEANFILES): New automake macro.
3319         (hack_DATA): Add ffitarget.h.
3320         * include/ffi.h.in: Remove all system specific definitions.
3321         Declare raw API even if it is not installed, why bother?
3322         Use limits.h instead of SIZEOF_* to define ffi_type_*.  Do
3323         not define EH_FRAME_FLAGS, it is in fficonfig.h now.  Include
3324         ffitarget.h instead of fficonfig.h.  Remove ALIGN macro.
3325         (UINT_ARG, INT_ARG): Removed, use ffi_arg and ffi_sarg instead.
3326         * include/ffi_common.h (bool): Do not define.
3327         (ffi_assert): Accept failed assertion.
3328         (ffi_type_test): Return void and accept file/line.
3329         (FFI_ASSERT): Pass stringized failed assertion.
3330         (FFI_ASSERT_AT): New macro.
3331         (FFI_ASSERT_VALID_TYPE): New macro.
3332         (UINT8, SINT8, UINT16, SINT16, UINT32, SINT32,
3333         UINT64, SINT64): Define here with gcc's __attribute__ macro
3334         instead of in ffi.h
3335         (FLOAT32, ALIGN): Define here instead of in ffi.h
3336         * include/ffi-mips.h: Removed.  Its content moved to
3337         src/mips/ffitarget.h after separating assembly and C sections.
3338         * src/alpha/ffi.c, src/alpha/ffi.c, src/java_raw_api.c
3339         src/prep_cif.c, src/raw_api.c, src/ia64/ffi.c,
3340         src/mips/ffi.c, src/mips/n32.S, src/mips/o32.S,
3341         src/mips/ffitarget.h, src/sparc/ffi.c, src/x86/ffi64.c:
3342         SIZEOF_ARG -> FFI_SIZEOF_ARG.
3343         * src/ia64/ffi.c: Include stdbool.h (provided by GCC 2.95+).
3344         * src/debug.c (ffi_assert): Accept stringized failed assertion.
3345         (ffi_type_test): Rewritten.
3346         * src/prep-cif.c (initialize_aggregate, ffi_prep_cif): Call
3347         FFI_ASSERT_VALID_TYPE.
3348         * src/alpha/ffitarget.h, src/arm/ffitarget.h,
3349         src/ia64/ffitarget.h, src/m68k/ffitarget.h,
3350         src/mips/ffitarget.h, src/powerpc/ffitarget.h,
3351         src/s390/ffitarget.h, src/sh/ffitarget.h,
3352         src/sh64/ffitarget.h, src/sparc/ffitarget.h,
3353         src/x86/ffitarget.h: New files.
3354         * src/alpha/osf.S, src/arm/sysv.S, src/ia64/unix.S,
3355         src/m68k/sysv.S, src/mips/n32.S, src/mips/o32.S,
3356         src/powerpc/aix.S, src/powerpc/darwin.S,
3357         src/powerpc/ffi_darwin.c, src/powerpc/linux64.S,
3358         src/powerpc/linux64_closure.S, src/powerpc/ppc_closure.S,
3359         src/powerpc/sysv.S, src/s390/sysv.S, src/sh/sysv.S,
3360         src/sh64/sysv.S, src/sparc/v8.S, src/sparc/v9.S,
3361         src/x86/sysv.S, src/x86/unix64.S, src/x86/win32.S:
3362         include fficonfig.h
3363
3364 2003-10-20  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3365
3366         * src/mips/ffi.c: Use _ABIN32, _ABIO32 instead of external
3367         _MIPS_SIM_NABI32, _MIPS_SIM_ABI32.
3368
3369 2003-10-19  Andreas Tobler  <a.tobler@schweiz.ch>
3370
3371         * src/powerpc/ffi_darwin.c (ffi_prep_args): Declare bytes again.
3372         Used when FFI_DEBUG = 1.
3373
3374 2003-10-14  Alan Modra  <amodra@bigpond.net.au>
3375
3376         * src/types.c (double, longdouble): Default POWERPC64 to 8 byte size
3377         and align.
3378
3379 2003-10-06  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3380
3381         * include/ffi_mips.h: Define FFI_MIPS_N32 for N32/N64 ABIs,
3382         FFI_MIPS_O32 for O32 ABI.
3383
3384 2003-10-01  Andreas Tobler  <a.tobler@schweiz.ch>
3385
3386         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_64 for
3387         SPARC64. Cleanup whitespaces.
3388
3389 2003-09-19  Andreas Tobler  <a.tobler@schweiz.ch>
3390
3391         * testsuite/libffi.call/closure_fn0.c: Xfail mips, arm,
3392         strongarm, xscale. Cleanup whitespaces.
3393         * testsuite/libffi.call/closure_fn1.c: Likewise.
3394         * testsuite/libffi.call/closure_fn2.c: Likewise.
3395         * testsuite/libffi.call/closure_fn3.c: Likewise.
3396         * testsuite/libffi.call/cls_12byte.c: Likewise.
3397         * testsuite/libffi.call/cls_16byte.c: Likewise.
3398         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3399         * testsuite/libffi.call/cls_20byte.c: Likewise.
3400         * testsuite/libffi.call/cls_24byte.c: Likewise.
3401         * testsuite/libffi.call/cls_2byte.c: Likewise.
3402         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3403         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3404         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3405         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3406         * testsuite/libffi.call/cls_4byte.c: Likewise.
3407         * testsuite/libffi.call/cls_5byte.c: Likewise.
3408         * testsuite/libffi.call/cls_6byte.c: Likewise.
3409         * testsuite/libffi.call/cls_7byte.c: Likewise.
3410         * testsuite/libffi.call/cls_8byte.c: Likewise.
3411         * testsuite/libffi.call/cls_double.c: Likewise.
3412         * testsuite/libffi.call/cls_float.c: Likewise.
3413         * testsuite/libffi.call/cls_uchar.c: Likewise.
3414         * testsuite/libffi.call/cls_uint.c: Likewise.
3415         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3416         * testsuite/libffi.call/cls_ushort.c: Likewise.
3417         * testsuite/libffi.call/nested_struct.c: Likewise.
3418         * testsuite/libffi.call/nested_struct1.c: Likewise.
3419         * testsuite/libffi.call/problem1.c: Likewise.
3420         * testsuite/libffi.special/unwindtest.cc: Likewise.
3421         * testsuite/libffi.call/pyobjc-tc.c: Cleanup whitespaces.
3422
3423 2003-09-18  David Edelsohn  <edelsohn@gnu.org>
3424
3425         * src/powerpc/aix.S: Cleanup whitespaces.
3426         * src/powerpc/aix_closure.S: Likewise.
3427
3428 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3429
3430         * src/powerpc/darwin.S: Cleanup whitespaces, comment formatting.
3431         * src/powerpc/darwin_closure.S: Likewise.
3432         * src/powerpc/ffi_darwin.c: Likewise.
3433
3434 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3435             David Edelsohn  <edelsohn@gnu.org>
3436
3437         * src/types.c (double): Add AIX and Darwin to the right TYPEDEF.
3438         * src/powerpc/aix_closure.S: Remove the pointer to the outgoing
3439         parameter stack.
3440         * src/powerpc/darwin_closure.S: Likewise.
3441         * src/powerpc/ffi_darwin.c (ffi_prep_args): Handle structures
3442         according to the Darwin/AIX ABI.
3443         (ffi_prep_cif_machdep): Likewise.
3444         (ffi_closure_helper_DARWIN): Likewise.
3445         Remove the outgoing parameter stack logic. Simplify the evaluation
3446         of the differ