libffi/ChangeLog

   1 2012-02-27  Mikael Pettersson  <mikpe@it.uu.se>
   2
   3         PR libffi/52223
   4         * Makefile.am (FLAGS_TO_PASS): Define.
   5         * Makefile.in: Regenerate.
   6
   7 2012-02-23  Kai Tietz  <ktietz@redhat.com>
   8
   9         PR libffi/52221
  10         * src/x86/ffi.c (ffi_closure_raw_THISCALL): New
  11         prototype.
  12         (ffi_prep_raw_closure_loc): Use ffi_closure_raw_THISCALL for
  13         thiscall-convention.
  14         (ffi_raw_call): Use ffi_prep_args_raw.
  15         * src/x86/win32.S (ffi_closure_raw_THISCALL): Add
  16         implementation for stub.
  17
  18 2012-02-13  Kai Tietz  <ktietz@redhat.com>
  19
  20         PR libffi/52221
  21         * src/x86/ffi.c (ffi_prep_raw_closure_loc): Add thiscall
  22         support for X86_WIN32.
  23         (FFI_INIT_TRAMPOLINE_THISCALL): Fix displacement.
  24
  25 2012-02-11  Eric Botcazou  <ebotcazou@adacore.com>
  26
  27         * src/sparc/v9.S (STACKFRAME): Bump to 176.
  28
  29 2012-02-10  Kai Tietz  <ktietz@redhat.com>
  30
  31         * configure.ac (AM_LTLDFLAGS): Add -no-undefine for x64
  32         windows target.
  33         * configure: Regenerated.
  34
  35 2012-02-08  Kai Tietz  <ktietz@redhat.com>
  36
  37         * src/prep_cif.c (ffi_prep_cif): Allow for X86_WIN32
  38         also FFI_THISCALL.
  39         * src/x86/ffi.c (ffi_closure_THISCALL): Add prototype.
  40         (FFI_INIT_TRAMPOLINE_THISCALL): New trampoline code.
  41         (ffi_prep_closure_loc): Add FFI_THISCALL support.
  42         * src/x86/ffitarget.h (FFI_TRAMPOLINE_SIZE): Adjust size.
  43         * src/x86/win32.S (ffi_closure_THISCALL): New closure code
  44         for thiscall-calling convention.
  45         * testsuite/libffi.call/closure_thiscall.c: New test.
  46
  47 2012-01-28  Kai Tietz  <ktietz@redhat.com>
  48
  49         * src/libffi/src/x86/ffi.c (ffi_call_win32): Add new
  50         argument to prototype for specify calling-convention.
  51         (ffi_call): Add support for stdcall/thiscall convention.
  52         (ffi_prep_args): Likewise.
  53         (ffi_raw_call): Likewise.
  54         * src/x86/ffitarget.h (ffi_abi): Add FFI_THISCALL and
  55         FFI_FASTCALL.
  56         * src/x86/win32.S (_ffi_call_win32): Add support for
  57         fastcall/thiscall calling-convention calls.
  58         * testsuite/libffi.call/fastthis1_win32.c: New test.
  59         * testsuite/libffi.call/fastthis2_win32.c: New test.
  60         * testsuite/libffi.call/fastthis3_win32.c: New test.
  61         * testsuite/libffi.call/strlen2_win32.c: New test.
  62         * testsuite/libffi.call/many2_win32.c: New test.
  63         * testsuite/libffi.call/struct1_win32.c: New test.
  64         * testsuite/libffi.call/struct2_win32.c: New test.
  65
  66 2012-01-23  Andreas Schwab  <schwab@linux-m68k.org>
  67
  68         * src/m68k/sysv.S (ffi_call_SYSV): Properly test for plain
  69         mc68000.  Test for __HAVE_68881__ in addition to __MC68881__.
  70
  71 2012-01-19  Jakub Jelinek  <jakub@redhat.com>
  72
  73         PR rtl-optimization/48496
  74         * src/ia64/ffi.c (ffi_call): Fix up aliasing violations.
  75
  76 2012-01-09  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
  77
  78         * configure.ac (i?86-*-*): Set TARGET to X86_64.
  79         * configure: Regenerate.
  80
  81 2011-12-07  Andrew Pinski  <apinski@cavium.com>
  82
  83         PR libffi/50051
  84         * src/mips/n32.S: Add ".set mips4".
  85
  86 2011-11-21  Andreas Tobler  <andreast@fgznet.ch>
  87
  88         * configure: Regenerate.
  89
  90 2011-11-10  Richard Henderson  <rth@redhat.com>
  91
  92         * configure.ac (GCC_AS_CFI_PSEUDO_OP): Use it instead of inline check.
  93         * configure, aclocal.m4: Rebuild.
  94
  95 2011-09-04  Iain Sandoe  <iains@gcc.gnu.org>
  96
  97         PR libffi/49594
  98         * src/powerpc/darwin_closure.S (stubs): Make the stub binding
  99         helper reference track the architecture pointer size.
 100
 101 2011-08-25  Andrew Haley  <aph@redhat.com>
 102
 103         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Remove hard-coded assembly
 104         instructions.
 105         * src/arm/sysv.S (ffi_arm_trampoline): Put them here instead.
 106
 107 2011-07-11  Andrew Haley  <aph@redhat.com>
 108
 109         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Clear icache.
 110
 111 2011-06-29  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 112
 113         * testsuite/libffi.call/cls_double_va.c: Move PR number to comment.
 114         * testsuite/libffi.call/cls_longdouble_va.c: Likewise.
 115
 116 2011-06-29  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 117
 118         PR libffi/46660
 119         * testsuite/libffi.call/cls_double_va.c: xfail dg-output on
 120         mips-sgi-irix6*.
 121         * testsuite/libffi.call/cls_longdouble_va.c: Likewise.
 122
 123 2011-06-14  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 124
 125         * testsuite/libffi.call/huge_struct.c (test_large_fn): Use PRIu8,
 126         PRId8 instead of %hhu, %hhd.
 127         * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRId8,
 128         PRIu8): Define.
 129         [__sgi__] (PRId8, PRIu8): Define.
 130
 131 2011-04-29  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 132
 133         * src/alpha/osf.S (UA_SI, FDE_ENCODING, FDE_ENCODE, FDE_ARANGE):
 134         Define.
 135         Use them to handle ELF vs. ECOFF differences.
 136         [__osf__] (_GLOBAL__F_ffi_call_osf): Define.
 137
 138 2011-02-13  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 139
 140         * configure: Regenerate.
 141
 142 2011-02-09  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 143
 144         PR libffi/46661
 145         * testsuite/libffi.call/cls_pointer.c (main): Cast void * to
 146         uintptr_t first.
 147         * testsuite/libffi.call/cls_pointer_stack.c (main): Likewise.
 148
 149 2011-02-07  Joel Sherrill <joel.sherrill@oarcorp.com>
 150
 151         * libffi/src/m68k/ffi.c: Add RTEMS support for cache flushing.
 152         Handle case when CPU variant does not have long double support.
 153         * libffi/src/m68k/sysv.S: Add support for mc68000, Coldfire,
 154         and cores with soft floating point.
 155
 156 2011-02-07  Joel Sherrill <joel.sherrill@oarcorp.com>
 157
 158         * configure.ac: Add mips*-*-rtems* support.
 159         * configure: Regenerate.
 160         * src/mips/ffitarget.h: Ensure needed constants are available
 161         for targets which do not have sgidefs.h.
 162
 163 2011-01-26  Dave Korn  <dave.korn.cygwin@gmail.com>
 164
 165         PR target/40125
 166         * configure.ac (AM_LTLDFLAGS): Add -bindir option for windows DLLs.
 167         * configure: Regenerate.
 168
 169 2010-12-18  Iain Sandoe  <iains@gcc.gnu.org>
 170
 171         PR libffi/29152
 172         PR libffi/42378
 173         * src/powerpc/darwin_closure.S: Provide Darwin64 implementation,
 174         update comments.
 175         * src/powerpc/ffitarget.h (POWERPC_DARWIN64): New,
 176         (FFI_TRAMPOLINE_SIZE): Update for Darwin64.
 177         * src/powerpc/darwin.S: Provide Darwin64 implementation,
 178         update comments.
 179         * src/powerpc/ffi_darwin.c: Likewise.
 180
 181 2010-12-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 182
 183         * configure.ac (libffi_cv_as_ascii_pseudo_op): Use double
 184         backslashes.
 185         (libffi_cv_as_string_pseudo_op): Likewise.
 186         * configure: Regenerate.
 187
 188 2010-12-03  Chung-Lin Tang  <cltang@codesourcery.com>
 189
 190         * src/arm/sysv.S (ffi_closure_SYSV): Add UNWIND to .pad directive.
 191         (ffi_closure_VFP): Same.
 192         (ffi_call_VFP): Move down to before ffi_closure_VFP. Add '.fpu vfp'
 193         directive.
 194
 195 2010-12-01  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 196
 197         * testsuite/libffi.call/ffitest.h [__sgi] (PRId64, PRIu64): Define.
 198         (PRIuPTR): Define.
 199
 200 2010-11-29  Richard Henderson  <rth@redhat.com>
 201             Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 202
 203         * src/x86/sysv.S (FDE_ENCODING, FDE_ENCODE): Define.
 204         (.eh_frame): Use FDE_ENCODING.
 205         (.LASFDE1, .LASFDE2, LASFDE3): Simplify with FDE_ENCODE.
 206
 207 2010-11-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 208
 209         * testsuite/lib/libffi-dg.exp: Rename ...
 210         * testsuite/lib/libffi.exp: ... to this.
 211         * libffi/testsuite/libffi.call/call.exp: Don't load libffi-dg.exp.
 212         * libffi/testsuite/libffi.special/special.exp: Likewise.
 213
 214 2010-10-28  Chung-Lin Tang  <cltang@codesourcery.com>
 215
 216         * src/arm/ffi.c (ffi_prep_args): Add VFP register argument handling
 217         code, new parameter, and return value. Update comments.
 218         (ffi_prep_cif_machdep): Add case for VFP struct return values. Add
 219         call to layout_vfp_args().
 220         (ffi_call_SYSV): Update declaration.
 221         (ffi_call_VFP): New declaration.
 222         (ffi_call): Add VFP struct return conditions. Call ffi_call_VFP()
 223         when ABI is FFI_VFP.
 224         (ffi_closure_VFP): New declaration.
 225         (ffi_closure_SYSV_inner): Add new vfp_args parameter, update call to
 226         ffi_prep_incoming_args_SYSV().
 227         (ffi_prep_incoming_args_SYSV): Update parameters. Add VFP argument
 228         case handling.
 229         (ffi_prep_closure_loc): Pass ffi_closure_VFP to trampoline
 230         construction under VFP hard-float.
 231         (rec_vfp_type_p): New function.
 232         (vfp_type_p): Same.
 233         (place_vfp_arg): Same.
 234         (layout_vfp_args): Same.
 235         * src/arm/ffitarget.h (ffi_abi): Add FFI_VFP. Define FFI_DEFAULT_ABI
 236         based on __ARM_PCS_VFP.
 237         (FFI_EXTRA_CIF_FIELDS): Define for adding VFP hard-float specific
 238         fields.
 239         (FFI_TYPE_STRUCT_VFP_FLOAT): Define internally used type code.
 240         (FFI_TYPE_STRUCT_VFP_DOUBLE): Same.
 241         * src/arm/sysv.S (ffi_call_SYSV): Change call of ffi_prep_args() to
 242         direct call. Move function pointer load upwards.
 243         (ffi_call_VFP): New function.
 244         (ffi_closure_VFP): Same.
 245
 246         * testsuite/lib/libffi-dg.exp (check-flags): New function.
 247         (dg-skip-if): New function.
 248         * testsuite/libffi.call/cls_double_va.c: Skip if target is arm*-*-*
 249         and compiler options include -mfloat-abi=hard.
 250         * testsuite/libffi.call/cls_longdouble_va.c: Same.
 251
 252 2010-10-01  Jakub Jelinek  <jakub@redhat.com>
 253
 254         PR libffi/45677
 255         * src/x86/ffi64.c (ffi_prep_cif_machdep): Ensure cif->bytes is
 256         a multiple of 8.
 257         * testsuite/libffi.call/many2.c: New test.
 258
 259 2010-08-20  Mark Wielaard  <mjw@redhat.com>
 260
 261         * src/closures.c (open_temp_exec_file_mnt): Check if getmntent_r
 262         returns NULL.
 263
 264 2010-08-09  Andreas Tobler  <andreast@fgznet.ch>
 265
 266         * configure.ac: Add target powerpc64-*-freebsd*.
 267         * configure: Regenerate.
 268         * testsuite/libffi.call/cls_align_longdouble_split.c: Pass
 269         -mlong-double-128 only to linux targets.
 270         * testsuite/libffi.call/cls_align_longdouble_split2.c: Likewise.
 271         * testsuite/libffi.call/cls_longdouble.c: Likewise.
 272         * testsuite/libffi.call/huge_struct.c: Likewise.
 273
 274 2010-07-10  Evan Phoenix  <evan@fallingsnow.net>
 275
 276         * src/closures.c (selinux_enabled_check): Fix strncmp usage bug.
 277
 278 2010-07-07  Dan Horák <dan@danny.cz>
 279
 280         * include/ffi.h.in: Protect #define with #ifndef.
 281         * src/powerpc/ffitarget.h: Ditto.
 282         * src/s390/ffitarget.h: Ditto.
 283         * src/sparc/ffitarget.h: Ditto.
 284
 285 2010-07-07   Neil Roberts <neil@linux.intel.com>
 286
 287         * src/x86/sysv.S (ffi_call_SYSV): Align the stack pointer to
 288         16-bytes.
 289
 290 2010-07-02  Jakub Jelinek  <jakub@redhat.com>
 291
 292         * Makefile.am (AM_MAKEFLAGS): Pass also mandir to submakes.
 293         * Makefile.in: Regenerated.
 294
 295 2010-05-19  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 296
 297         * configure.ac (libffi_cv_as_x86_pcrel): Check for illegal in as
 298         output, too.
 299         (libffi_cv_as_ascii_pseudo_op): Check for .ascii.
 300         (libffi_cv_as_string_pseudo_op): Check for .string.
 301         * configure: Regenerate.
 302         * fficonfig.h.in: Regenerate.
 303         * src/x86/sysv.S (.eh_frame): Use .ascii, .string or error.
 304
 305 2010-05-05  Michael Kohler <michaelkohler@live.com>
 306
 307         * src/dlmalloc.c (dlfree): Fix spelling.
 308         * src/ia64/ffi.c (ffi_prep_cif_machdep): Ditto.
 309         * configure.ac: Ditto.
 310         * configure: Rebuilt.
 311
 312 2010-04-13  Dan Witte  <dwitte@mozilla.com>
 313
 314         * msvcc.sh: Build with -W3 instead of -Wall.
 315         * src/powerpc/ffi_darwin.c: Remove build warnings.
 316         * src/x86/ffi.c: Ditto.
 317         * src/x86/ffitarget.h: Ditto.
 318
 319 2010-04-12  Dan Witte  <dwitte@mozilla.com>
 320             Walter Meinl <wuno@lsvw.de>
 321
 322         * configure.ac: Add OS/2 support.
 323         * configure: Rebuilt.
 324         * src/closures.c: Ditto.
 325         * src/dlmalloc.c: Ditto.
 326         * src/x86/win32.S: Ditto.
 327
 328 2010-04-07  Jakub Jelinek  <jakub@redhat.com>
 329
 330         * testsuite/libffi.call/err_bad_abi.c: Remove unused args variable.
 331
 332 2010-04-02  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 333
 334         * Makefile.in: Regenerate.
 335         * aclocal.m4: Regenerate.
 336         * include/Makefile.in: Regenerate.
 337         * man/Makefile.in: Regenerate.
 338         * testsuite/Makefile.in: Regenerate.
 339
 340 2010-03-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 341
 342         * configure.ac (libffi_cv_as_x86_64_unwind_section_type): New test.
 343         * configure: Regenerate.
 344         * fficonfig.h.in: Regenerate.
 345         * libffi/src/x86/unix64.S (.eh_frame)
 346         [HAVE_AS_X86_64_UNWIND_SECTION_TYPE]: Use @unwind section type.
 347
 348 2010-03-14  Matthias Klose  <doko@ubuntu.com>
 349
 350         * src/x86/ffi64.c: Fix typo in comment.
 351         * src/x86/ffi.c: Use /* ... */ comment style.
 352
 353 2010-02-24  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 354
 355         * doc/libffi.texi (The Closure API): Fix typo.
 356         * doc/libffi.info: Remove.
 357
 358 2010-02-15  Matthias Klose  <doko@ubuntu.com>
 359
 360         * src/arm/sysv.S (__ARM_ARCH__): Define for processor
 361         __ARM_ARCH_7EM__.
 362
 363 2010-01-15  Anthony Green  <green@redhat.com>
 364
 365         * README: Add notes on building with Microsoft Visual C++.
 366
 367 2010-01-15  Daniel Witte  <dwitte@mozilla.com>
 368
 369         * msvcc.sh: New file.
 370
 371         * src/x86/win32.S: Port assembly routines to MSVC and #ifdef.
 372         * src/x86/ffi.c: Tweak function declaration and remove excess
 373         parens.
 374         * include/ffi.h.in: Add __declspec(align(8)) to typedef struct
 375         ffi_closure.
 376
 377         * src/x86/ffi.c: Merge ffi_call_SYSV and ffi_call_STDCALL into new
 378         function ffi_call_win32 on X86_WIN32.
 379         * src/x86/win32.S (ffi_call_SYSV): Rename to ffi_call_win32.
 380         (ffi_call_STDCALL): Remove.
 381
 382         * src/prep_cif.c (ffi_prep_cif): Move stack space allocation code
 383         to ffi_prep_cif_machdep for x86.
 384         * src/x86/ffi.c (ffi_prep_cif_machdep): To here.
 385
 386 2010-01-15  Oliver Kiddle  <okiddle@yahoo.co.uk>
 387
 388         * src/x86/ffitarget.h (ffi_abi): Check for __i386 and __amd64 for
 389         Sun Studio compiler compatibility.
 390
 391 2010-01-12  Conrad Irwin <conrad.irwin@gmail.com>
 392
 393         * doc/libffi.texi: Add closure example.
 394
 395 2010-01-07  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 396
 397         PR libffi/40701
 398         * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRIdLL,
 399         PRIuLL, PRId64, PRIu64, PRIuPTR): Define.
 400         * testsuite/libffi.call/cls_align_sint64.c: Add -Wno-format on
 401         alpha*-dec-osf*.
 402         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
 403         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
 404         * testsuite/libffi.call/return_ll1.c: Likewise.
 405         * testsuite/libffi.call/stret_medium2.c: Likewise.
 406         * testsuite/libffi.special/ffitestcxx.h (allocate_mmap): Cast
 407         MAP_FAILED to char *.
 408
 409 2010-01-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 410
 411         * src/mips/n32.S: Use .abicalls and .eh_frame with __GNUC__.
 412
 413 2009-12-31  Anthony Green  <green@redhat.com>
 414
 415         * README: Update for libffi 3.0.9.
 416
 417 2009-12-27  Matthias Klose  <doko@ubuntu.com>
 418
 419         * configure.ac (HAVE_LONG_DOUBLE): Define for mips when
 420         appropriate.
 421         * configure: Rebuilt.
 422
 423 2009-12-26  Anthony Green  <green@redhat.com>
 424
 425         * testsuite/libffi.call/cls_longdouble_va.c: Mark as xfail for
 426         avr32*-*-*.
 427         * testsuite/libffi.call/cls_double_va.c: Ditto.
 428
 429 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
 430
 431         * testsuite/libffi.call/ffitest.h: Conditionally include stdint.h
 432         and inttypes.h.
 433         * testsuite/libffi.special/unwindtest.cc: Ditto.
 434
 435 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
 436
 437         * configure.ac: Add amd64-*-openbsd*.
 438         * configure: Rebuilt.
 439         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Link
 440         openbsd programs with -lpthread.
 441
 442 2009-12-26  Anthony Green  <green@redhat.com>
 443
 444         * testsuite/libffi.call/cls_double_va.c,
 445         testsuite/libffi.call/cls_longdouble.c,
 446         testsuite/libffi.call/cls_longdouble_va.c,
 447         testsuite/libffi.call/cls_pointer.c,
 448         testsuite/libffi.call/cls_pointer_stack.c: Remove xfail for
 449         mips*-*-* and arm*-*-*.
 450         * testsuite/libffi.call/cls_align_longdouble_split.c,
 451         testsuite/libffi.call/cls_align_longdouble_split2.c,
 452         testsuite/libffi.call/stret_medium2.c,
 453         testsuite/libffi.call/stret_medium.c,
 454         testsuite/libffi.call/stret_large.c,
 455         testsuite/libffi.call/stret_large2.c: Remove xfail for arm*-*-*.
 456
 457 2009-12-31  Kay Tietz  <ktietz70@googlemail.com>
 458
 459         * testsuite/libffi.call/ffitest.h,
 460         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRuLL): Fix
 461         definitions.
 462
 463 2009-12-31  Carlo Bramini  <carlo.bramix@libero.it>
 464
 465         * configure.ac (AM_LTLDFLAGS): Define for windows hosts.
 466         * Makefile.am (libffi_la_LDFLAGS): Add AM_LTLDFLAGS.
 467         * configure: Rebuilt.
 468         * Makefile.in: Rebuilt.
 469
 470 2009-12-31  Anthony Green  <green@redhat.com>
 471             Blake Chaffin.
 472
 473         * testsuite/libffi.call/huge_struct.c: New test case from Blake
 474         Chaffin @ Apple.
 475
 476 2009-12-28  David Edelsohn  <edelsohn@gnu.org>
 477
 478         * src/powerpc/ffi_darwin.c (ffi_prep_args): Copy abi and nargs to
 479         local variables.
 480         (aix_adjust_aggregate_sizes): New function.
 481         (ffi_prep_cif_machdep): Call it.
 482
 483 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
 484
 485         * configure.ac: Define FFI_MMAP_EXEC_WRIT for the given targets.
 486         * configure: Regenerate.
 487         * fficonfig.h.in: Likewise.
 488         * src/closures.c: Remove the FFI_MMAP_EXEC_WRIT definition for
 489         Solaris/x86.
 490
 491 2009-12-26  Andreas Schwab  <schwab@linux-m68k.org>
 492
 493         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Advance intarg_count
 494         when a float arguments is passed in memory.
 495         (ffi_closure_helper_SYSV): Mark general registers as used up when
 496         a 64bit or soft-float long double argument is passed in memory.
 497
 498 2009-12-25  Matthias Klose  <doko@ubuntu.com>
 499
 500         * man/ffi_call.3: Fix #include in examples.
 501         * doc/libffi.texi: Add dircategory.
 502
 503 2009-12-25  Frank Everdij <f.p.x.everdij@tudelft.nl>
 504
 505         * include/ffi.h.in: Placed '__GNUC__' ifdef around
 506         '__attribute__((aligned(8)))' in ffi_closure, fixes compile for
 507         IRIX MIPSPro c99.
 508         * include/ffi_common.h: Added '__sgi' define to non
 509         '__attribute__((__mode__()))' integer typedefs.
 510         * src/mips/ffi.c (ffi_call, ffi_closure_mips_inner_O32,
 511         ffi_closure_mips_inner_N32): Added 'defined(_MIPSEB)' to BE check.
 512         (ffi_closure_mips_inner_O32, ffi_closure_mips_inner_N32): Added
 513         FFI_LONGDOUBLE support and alignment(N32 only).
 514         * src/mips/ffitarget.h: Corrected '#include <sgidefs.h>' for IRIX and
 515         fixed non '__attribute__((__mode__()))' integer typedefs.
 516         * src/mips/n32.S: Put '#ifdef linux' around '.abicalls' and '.eh_frame'
 517         since they are Linux/GNU Assembler specific.
 518
 519 2009-12-25  Bradley Smith  <brad@brad-smith.co.uk>
 520
 521         * configure.ac, Makefile.am, src/avr32/ffi.c,
 522         src/avr32/ffitarget.h,
 523         src/avr32/sysv.S: Add AVR32 port.
 524         * configure, Makefile.in: Rebuilt.
 525
 526 2009-12-21  Andreas Tobler  <a.tobler@schweiz.org>
 527
 528         * configure.ac: Make i?86 build on FreeBSD and OpenBSD.
 529         * configure: Regenerate.
 530
 531 2009-12-15  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
 532
 533         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on PA HP-UX.
 534
 535 2009-12-13  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
 536
 537         * src/pa/ffi.c (ffi_closure_inner_pa32): Handle FFI_TYPE_LONGDOUBLE
 538         type on HP-UX.
 539
 540 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
 541
 542         * src/sparc/ffi.c (ffi_closure_sparc_inner_v9): Properly align 'long
 543         double' arguments.
 544
 545 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
 546
 547         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on Solaris < 10.
 548
 549 2009-12-10  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 550
 551         PR libffi/40700
 552         * src/closures.c [X86_64 && __sun__ && __svr4__]
 553         (FFI_MMAP_EXEC_WRIT): Define.
 554
 555 2009-12-08  David Daney  <ddaney@caviumnetworks.com>
 556
 557         * testsuite/libffi.call/stret_medium.c: Remove xfail for mips*-*-*
 558         * testsuite/libffi.call/cls_align_longdouble_split2.c: Same.
 559         * testsuite/libffi.call/stret_large.c: Same.
 560         * testsuite/libffi.call/cls_align_longdouble_split.c: Same.
 561         * testsuite/libffi.call/stret_large2.c: Same.
 562         * testsuite/libffi.call/stret_medium2.c: Same.
 563
 564 2009-12-07  David Edelsohn  <edelsohn@gnu.org>
 565
 566         * src/powerpc/aix_closure.S (libffi_closure_ASM): Fix tablejump
 567         typo.
 568
 569 2009-12-05  David Edelsohn  <edelsohn@gnu.org>
 570
 571         * src/powerpc/aix.S: Update AIX32 code to be consistent with AIX64
 572         code.
 573         * src/powerpc/aix_closure.S: Same.
 574
 575 2009-12-05  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 576
 577         * Makefile.in: Regenerate.
 578         * configure: Regenerate.
 579         * include/Makefile.in: Regenerate.
 580         * man/Makefile.in: Regenerate.
 581         * testsuite/Makefile.in: Regenerate.
 582
 583 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
 584
 585         * src/powerpc/aix_closure.S: Reorganize 64-bit code to match
 586         linux64_closure.S.
 587
 588 2009-12-04  Uros Bizjak  <ubizjak@gmail.com>
 589
 590         PR libffi/41908
 591         * src/x86/ffi64.c (classify_argument): Update from
 592         gcc/config/i386/i386.c.
 593         (ffi_closure_unix64_inner): Do not use the address of two consecutive
 594         SSE registers directly.
 595         * testsuite/libffi.call/cls_dbls_struct.c (main): Remove xfail
 596         for x86_64 linux targets.
 597
 598 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
 599
 600         * src/powerpc/ffi_darwin.c (ffi_closure_helper_DARWIN): Increment
 601         pfr for long double split between fpr13 and stack.
 602
 603 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
 604
 605         * src/powerpc/ffi_darwin.c (ffi_prep_args): Increment next_arg and
 606         fparg_count twice for long double.
 607
 608 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
 609
 610         PR libffi/42243
 611         * src/powerpc/ffi_darwin.c (ffi_prep_args): Remove extra parentheses.
 612
 613 2009-12-03  Uros Bizjak  <ubizjak@gmail.com>
 614
 615         * testsuite/libffi.call/cls_longdouble_va.c (main): Fix format string.
 616         Remove xfails for x86 linux targets.
 617
 618 2009-12-02  David Edelsohn  <edelsohn@gnu.org>
 619
 620         * src/powerpc/ffi_darwin.c (ffi_prep_args): Fix typo in INT64
 621         case.
 622
 623 2009-12-01  David Edelsohn  <edelsohn@gnu.org>
 624
 625         * src/powerpc/aix.S (ffi_call_AIX): Convert to more standard
 626         register usage.  Call ffi_prep_args directly.  Add long double
 627         return value support.
 628         * src/powerpc/ffi_darwin.c (ffi_prep_args): Double arg increment
 629         applies to FFI_TYPE_DOUBLE.  Correct fpr_base increment typo.
 630         Separate FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases.
 631         (ffi_prep_cif_machdep): Only 16 byte stack alignment in 64 bit
 632         mode.
 633         (ffi_closure_helper_DARWIN): Remove nf and ng counters.  Move temp
 634         into case.
 635         * src/powerpc/aix_closure.S: Maintain 16 byte stack alignment.
 636         Allocate result area between params and FPRs.
 637
 638 2009-11-30  David Edelsohn  <edelsohn@gnu.org>
 639
 640         PR target/35484
 641         * src/powerpc/ffitarget.h (POWERPC64): Define for PPC64 Linux and
 642         AIX64.
 643         * src/powerpc/aix.S: Implement AIX64 version.
 644         * src/powerpc/aix_closure.S: Implement AIX64 version.
 645         (ffi_closure_ASM): Use extsb, lha and displament addresses.
 646         * src/powerpc/ffi_darwin.c (ffi_prep_args): Implement AIX64
 647         support.
 648         (ffi_prep_cif_machdep): Same.
 649         (ffi_call): Same.
 650         (ffi_closure_helper_DARWIN): Same.
 651
 652 2009-11-02  Andreas Tobler  <a.tobler@schweiz.org>
 653
 654         PR libffi/41908
 655         * testsuite/libffi.call/testclosure.c: New test.
 656
 657 2009-09-28  Kai Tietz  <kai.tietz@onevision.com>
 658
 659         * src/x86/win64.S (_ffi_call_win64 stack): Remove for gnu
 660         assembly version use of ___chkstk.
 661
 662 2009-09-23  Matthias Klose  <doko@ubuntu.com>
 663
 664         PR libffi/40242, PR libffi/41443
 665         * src/arm/sysv.S (__ARM_ARCH__): Define for processors
 666         __ARM_ARCH_6T2__, __ARM_ARCH_6M__, __ARM_ARCH_7__,
 667         __ARM_ARCH_7A__, __ARM_ARCH_7R__, __ARM_ARCH_7M__.
 668         Change the conditionals to __SOFTFP__ || __ARM_EABI__
 669         for -mfloat-abi=softfp to work.
 670
 671 2009-09-17  Loren J. Rittle  <ljrittle@acm.org>
 672
 673         PR testsuite/32843 (strikes again)
 674         * src/x86/ffi.c (ffi_prep_cif_machdep): Add X86_FREEBSD to
 675         enable proper extension on char and short.
 676
 677 2009-09-15  David Daney  <ddaney@caviumnetworks.com>
 678
 679         * src/java_raw_api.c (ffi_java_raw_to_rvalue): Remove special
 680         handling for FFI_TYPE_POINTER.
 681         * src/mips/ffitarget.h (FFI_TYPE_STRUCT_D_SOFT,
 682         FFI_TYPE_STRUCT_F_SOFT, FFI_TYPE_STRUCT_DD_SOFT,
 683         FFI_TYPE_STRUCT_FF_SOFT, FFI_TYPE_STRUCT_FD_SOFT,
 684         FFI_TYPE_STRUCT_DF_SOFT, FFI_TYPE_STRUCT_SOFT): New defines.
 685         (FFI_N32_SOFT_FLOAT, FFI_N64_SOFT_FLOAT): New ffi_abi enumerations.
 686         (enum ffi_abi): Set FFI_DEFAULT_ABI for soft-float.
 687         * src/mips/n32.S (ffi_call_N32): Add handling for soft-float
 688         structure and pointer returns.
 689         (ffi_closure_N32): Add handling for pointer returns.
 690         * src/mips/ffi.c (ffi_prep_args, calc_n32_struct_flags,
 691         calc_n32_return_struct_flags): Handle soft-float.
 692         (ffi_prep_cif_machdep):  Handle soft-float, fix pointer handling.
 693         (ffi_call_N32): Declare proper argument types.
 694         (ffi_call, copy_struct_N32, ffi_closure_mips_inner_N32): Handle
 695         soft-float.
 696
 697 2009-08-24  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 698
 699         * configure.ac (AC_PREREQ): Bump to 2.64.
 700
 701 2009-08-22  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 702
 703         * Makefile.am (install-html, install-pdf): Remove.
 704         * Makefile.in: Regenerate.
 705
 706         * Makefile.in: Regenerate.
 707         * aclocal.m4: Regenerate.
 708         * configure: Regenerate.
 709         * fficonfig.h.in: Regenerate.
 710         * include/Makefile.in: Regenerate.
 711         * man/Makefile.in: Regenerate.
 712         * testsuite/Makefile.in: Regenerate.
 713
 714 2009-07-30  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 715
 716         * configure.ac (_AC_ARG_VAR_PRECIOUS): Use m4_rename_force.
 717
 718 2009-07-24  Dave Korn  <dave.korn.cygwin@gmail.com>
 719
 720         PR libffi/40807
 721         * src/x86/ffi.c (ffi_prep_cif_machdep): Also use sign/zero-extending
 722         return types for X86_WIN32.
 723         * src/x86/win32.S (_ffi_call_SYSV): Handle omitted return types.
 724         (_ffi_call_STDCALL, _ffi_closure_SYSV, _ffi_closure_raw_SYSV,
 725         _ffi_closure_STDCALL): Likewise.
 726
 727         * src/closures.c (is_selinux_enabled): Define to const 0 for Cygwin.
 728         (dlmmap, dlmunmap): Also use these functions on Cygwin.
 729
 730 2009-07-11  Richard Sandiford  <rdsandiford@googlemail.com>
 731
 732         PR testsuite/40699
 733         PR testsuite/40707
 734         PR testsuite/40709
 735         * testsuite/lib/libffi-dg.exp: Revert 2009-07-02, 2009-07-01 and
 736         2009-06-30 commits.
 737
 738 2009-07-01  Richard Sandiford  <r.sandiford@uk.ibm.com>
 739
 740         * testsuite/lib/libffi-dg.exp (libffi-init): Set ld_library_path
 741         to "" before adding paths.  (This reinstates an assignment that
 742         was removed by my 2009-06-30 commit, but changes the initial
 743         value from "." to "".)
 744
 745 2009-07-01  H.J. Lu  <hongjiu.lu@intel.com>
 746
 747         PR testsuite/40601
 748         * testsuite/lib/libffi-dg.exp (libffi-init): Properly set
 749         gccdir.  Adjust ld_library_path for gcc only if gccdir isn't
 750         empty.
 751
 752 2009-06-30  Richard Sandiford  <r.sandiford@uk.ibm.com>
 753
 754         * testsuite/lib/libffi-dg.exp (libffi-init): Don't add "."
 755         to ld_library_path.  Use add_path.  Add just find_libgcc_s
 756         to ld_library_path, not every libgcc multilib directory.
 757
 758 2009-06-16  Wim Lewis  <wiml@hhhh.org>
 759
 760         * src/powerpc/ffi.c: Avoid clobbering cr3 and cr4, which are
 761         supposed to be callee-saved.
 762         * src/powerpc/sysv.S (small_struct_return_value): Fix overrun of
 763         return buffer for odd-size structs.
 764
 765 2009-06-16  Andreas Tobler  <a.tobler@schweiz.org>
 766
 767         PR libffi/40444
 768         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Add
 769         allow_stack_execute for Darwin.
 770
 771 2009-06-16  Andrew Haley  <aph@redhat.com>
 772
 773         * configure.ac (TARGETDIR): Add missing blank lines.
 774         * configure: Regenerate.
 775
 776 2009-06-16  Andrew Haley  <aph@redhat.com>
 777
 778         * testsuite/libffi.call/cls_align_sint64.c,
 779         testsuite/libffi.call/cls_align_uint64.c,
 780         testsuite/libffi.call/cls_longdouble_va.c,
 781         testsuite/libffi.call/cls_ulonglong.c,
 782         testsuite/libffi.call/return_ll1.c,
 783         testsuite/libffi.call/stret_medium2.c: Fix printf format
 784         specifiers.
 785         * testsuite/libffi.call/ffitest.h,
 786         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRIuLL): Define.
 787
 788 2009-06-15  Andrew Haley  <aph@redhat.com>
 789
 790         * testsuite/libffi.call/err_bad_typedef.c: xfail everywhere.
 791         * testsuite/libffi.call/err_bad_abi.c: Likewise.
 792
 793 2009-06-12  Andrew Haley  <aph@redhat.com>
 794
 795         * Makefile.am: Remove info_TEXINFOS.
 796
 797 2009-06-12  Andrew Haley  <aph@redhat.com>
 798
 799         * ChangeLog.libffi: testsuite/libffi.call/cls_align_sint64.c,
 800         testsuite/libffi.call/cls_align_uint64.c,
 801         testsuite/libffi.call/cls_ulonglong.c,
 802         testsuite/libffi.call/return_ll1.c,
 803         testsuite/libffi.call/stret_medium2.c: Fix printf format
 804         specifiers.
 805         testsuite/libffi.special/unwindtest.cc: include stdint.h.
 806
 807 2009-06-11  Timothy Wall  <twall@users.sf.net>
 808
 809         * Makefile.am,
 810         configure.ac,
 811         include/ffi.h.in,
 812         include/ffi_common.h,
 813         src/closures.c,
 814         src/dlmalloc.c,
 815         src/x86/ffi.c,
 816         src/x86/ffitarget.h,
 817         src/x86/win64.S (new),
 818         README: Added win64 support (mingw or MSVC)
 819         * Makefile.in,
 820         include/Makefile.in,
 821         man/Makefile.in,
 822         testsuite/Makefile.in,
 823         configure,
 824         aclocal.m4: Regenerated
 825         * ltcf-c.sh: properly escape cygwin/w32 path
 826         * man/ffi_call.3: Clarify size requirements for return value.
 827         * src/x86/ffi64.c: Fix filename in comment.
 828         * src/x86/win32.S: Remove unused extern.
 829
 830         * testsuite/libffi.call/closure_fn0.c,
 831         testsuite/libffi.call/closure_fn1.c,
 832         testsuite/libffi.call/closure_fn2.c,
 833         testsuite/libffi.call/closure_fn3.c,
 834         testsuite/libffi.call/closure_fn4.c,
 835         testsuite/libffi.call/closure_fn5.c,
 836         testsuite/libffi.call/closure_fn6.c,
 837         testsuite/libffi.call/closure_stdcall.c,
 838         testsuite/libffi.call/cls_12byte.c,
 839         testsuite/libffi.call/cls_16byte.c,
 840         testsuite/libffi.call/cls_18byte.c,
 841         testsuite/libffi.call/cls_19byte.c,
 842         testsuite/libffi.call/cls_1_1byte.c,
 843         testsuite/libffi.call/cls_20byte.c,
 844         testsuite/libffi.call/cls_20byte1.c,
 845         testsuite/libffi.call/cls_24byte.c,
 846         testsuite/libffi.call/cls_2byte.c,
 847         testsuite/libffi.call/cls_3_1byte.c,
 848         testsuite/libffi.call/cls_3byte1.c,
 849         testsuite/libffi.call/cls_3byte2.c,
 850         testsuite/libffi.call/cls_4_1byte.c,
 851         testsuite/libffi.call/cls_4byte.c,
 852         testsuite/libffi.call/cls_5_1_byte.c,
 853         testsuite/libffi.call/cls_5byte.c,
 854         testsuite/libffi.call/cls_64byte.c,
 855         testsuite/libffi.call/cls_6_1_byte.c,
 856         testsuite/libffi.call/cls_6byte.c,
 857         testsuite/libffi.call/cls_7_1_byte.c,
 858         testsuite/libffi.call/cls_7byte.c,
 859         testsuite/libffi.call/cls_8byte.c,
 860         testsuite/libffi.call/cls_9byte1.c,
 861         testsuite/libffi.call/cls_9byte2.c,
 862         testsuite/libffi.call/cls_align_double.c,
 863         testsuite/libffi.call/cls_align_float.c,
 864         testsuite/libffi.call/cls_align_longdouble.c,
 865         testsuite/libffi.call/cls_align_longdouble_split.c,
 866         testsuite/libffi.call/cls_align_longdouble_split2.c,
 867         testsuite/libffi.call/cls_align_pointer.c,
 868         testsuite/libffi.call/cls_align_sint16.c,
 869         testsuite/libffi.call/cls_align_sint32.c,
 870         testsuite/libffi.call/cls_align_sint64.c,
 871         testsuite/libffi.call/cls_align_uint16.c,
 872         testsuite/libffi.call/cls_align_uint32.c,
 873         testsuite/libffi.call/cls_align_uint64.c,
 874         testsuite/libffi.call/cls_dbls_struct.c,
 875         testsuite/libffi.call/cls_double.c,
 876         testsuite/libffi.call/cls_double_va.c,
 877         testsuite/libffi.call/cls_float.c,
 878         testsuite/libffi.call/cls_longdouble.c,
 879         testsuite/libffi.call/cls_longdouble_va.c,
 880         testsuite/libffi.call/cls_multi_schar.c,
 881         testsuite/libffi.call/cls_multi_sshort.c,
 882         testsuite/libffi.call/cls_multi_sshortchar.c,
 883         testsuite/libffi.call/cls_multi_uchar.c,
 884         testsuite/libffi.call/cls_multi_ushort.c,
 885         testsuite/libffi.call/cls_multi_ushortchar.c,
 886         testsuite/libffi.call/cls_pointer.c,
 887         testsuite/libffi.call/cls_pointer_stack.c,
 888         testsuite/libffi.call/cls_schar.c,
 889         testsuite/libffi.call/cls_sint.c,
 890         testsuite/libffi.call/cls_sshort.c,
 891         testsuite/libffi.call/cls_uchar.c,
 892         testsuite/libffi.call/cls_uint.c,
 893         testsuite/libffi.call/cls_ulonglong.c,
 894         testsuite/libffi.call/cls_ushort.c,
 895         testsuite/libffi.call/err_bad_abi.c,
 896         testsuite/libffi.call/err_bad_typedef.c,
 897         testsuite/libffi.call/float2.c,
 898         testsuite/libffi.call/huge_struct.c,
 899         testsuite/libffi.call/nested_struct.c,
 900         testsuite/libffi.call/nested_struct1.c,
 901         testsuite/libffi.call/nested_struct10.c,
 902         testsuite/libffi.call/nested_struct2.c,
 903         testsuite/libffi.call/nested_struct3.c,
 904         testsuite/libffi.call/nested_struct4.c,
 905         testsuite/libffi.call/nested_struct5.c,
 906         testsuite/libffi.call/nested_struct6.c,
 907         testsuite/libffi.call/nested_struct7.c,
 908         testsuite/libffi.call/nested_struct8.c,
 909         testsuite/libffi.call/nested_struct9.c,
 910         testsuite/libffi.call/problem1.c,
 911         testsuite/libffi.call/return_ldl.c,
 912         testsuite/libffi.call/return_ll1.c,
 913         testsuite/libffi.call/stret_large.c,
 914         testsuite/libffi.call/stret_large2.c,
 915         testsuite/libffi.call/stret_medium.c,
 916         testsuite/libffi.call/stret_medium2.c,
 917         testsuite/libffi.special/unwindtest.cc: use ffi_closure_alloc instead
 918         of checking for MMAP.  Use intptr_t instead of long casts.
 919
 920 2009-06-11  Kaz Kojima  <kkojima@gcc.gnu.org>
 921
 922         * testsuite/libffi.call/cls_longdouble_va.c: Add xfail sh*-*-linux-*.
 923         * testsuite/libffi.call/err_bad_abi.c: Add xfail sh*-*-*.
 924         * testsuite/libffi.call/err_bad_typedef.c: Likewise.
 925
 926 2009-06-09  Andrew Haley  <aph@redhat.com>
 927
 928         * src/x86/freebsd.S: Add missing file.
 929
 930 2009-06-08  Andrew Haley  <aph@redhat.com>
 931
 932         Import from libffi 3.0.8:
 933
 934         * doc/libffi.texi: New file.
 935         * doc/libffi.info: Likewise.
 936         * doc/stamp-vti: Likewise.
 937         * man/Makefile.am: New file.
 938         * man/ffi_call.3: New file.
 939
 940         * Makefile.am (EXTRA_DIST): Add src/x86/darwin64.S,
 941         src/dlmalloc.c.
 942         (nodist_libffi_la_SOURCES): Add X86_FREEBSD.
 943
 944         * configure.ac: Bump version to 3.0.8.
 945         parisc*-*-linux*: Add.
 946         i386-*-freebsd* | i386-*-openbsd*: Add.
 947         powerpc-*-beos*: Add.
 948         AM_CONDITIONAL X86_FREEBSD: Add.
 949         AC_CONFIG_FILES: Add man/Makefile.
 950
 951         * include/ffi.h.in (FFI_FN): Change void (*)() to void (*)(void).
 952
 953 2009-06-08  Andrew Haley  <aph@redhat.com>
 954
 955         * README: Import from libffi 3.0.8.
 956
 957 2009-06-08  Andrew Haley  <aph@redhat.com>
 958
 959         * testsuite/libffi.call/err_bad_abi.c: Add xfails.
 960         * testsuite/libffi.call/cls_longdouble_va.c: Add xfails.
 961         * testsuite/libffi.call/cls_dbls_struct.c: Add xfail x86_64-*-linux-*.
 962         * testsuite/libffi.call/err_bad_typedef.c: Add xfails.
 963
 964         * testsuite/libffi.call/stret_medium2.c: Add __UNUSED__ to args.
 965         * testsuite/libffi.call/stret_medium.c: Likewise.
 966         * testsuite/libffi.call/stret_large2.c: Likewise.
 967         * testsuite/libffi.call/stret_large.c:  Likewise.
 968
 969 2008-12-26  Timothy Wall  <twall@users.sf.net>
 970
 971         * testsuite/libffi.call/cls_longdouble.c,
 972         testsuite/libffi.call/cls_longdouble_va.c,
 973         testsuite/libffi.call/cls_align_longdouble.c,
 974         testsuite/libffi.call/cls_align_longdouble_split.c,
 975         testsuite/libffi.call/cls_align_longdouble_split2.c: mark expected
 976         failures on x86_64 cygwin/mingw.
 977
 978 2008-12-22  Timothy Wall  <twall@users.sf.net>
 979
 980         * testsuite/libffi.call/closure_fn0.c,
 981         testsuite/libffi.call/closure_fn1.c,
 982         testsuite/libffi.call/closure_fn2.c,
 983         testsuite/libffi.call/closure_fn3.c,
 984         testsuite/libffi.call/closure_fn4.c,
 985         testsuite/libffi.call/closure_fn5.c,
 986         testsuite/libffi.call/closure_fn6.c,
 987         testsuite/libffi.call/closure_loc_fn0.c,
 988         testsuite/libffi.call/closure_stdcall.c,
 989         testsuite/libffi.call/cls_align_pointer.c,
 990         testsuite/libffi.call/cls_pointer.c,
 991         testsuite/libffi.call/cls_pointer_stack.c: use portable cast from
 992         pointer to integer (intptr_t).
 993         * testsuite/libffi.call/cls_longdouble.c: disable for win64.
 994
 995 2008-07-24  Anthony Green  <green@redhat.com>
 996
 997         * testsuite/libffi.call/cls_dbls_struct.c,
 998         testsuite/libffi.call/cls_double_va.c,
 999         testsuite/libffi.call/cls_longdouble.c,
1000         testsuite/libffi.call/cls_longdouble_va.c,
1001         testsuite/libffi.call/cls_pointer.c,
1002         testsuite/libffi.call/cls_pointer_stack.c,
1003         testsuite/libffi.call/err_bad_abi.c: Clean up failures from
1004         compiler warnings.
1005
1006 2008-03-04  Anthony Green  <green@redhat.com>
1007             Blake Chaffin
1008             hos@tamanegi.org
1009
1010         * testsuite/libffi.call/cls_align_longdouble_split2.c
1011           testsuite/libffi.call/cls_align_longdouble_split.c
1012           testsuite/libffi.call/cls_dbls_struct.c
1013           testsuite/libffi.call/cls_double_va.c
1014           testsuite/libffi.call/cls_longdouble.c
1015           testsuite/libffi.call/cls_longdouble_va.c
1016           testsuite/libffi.call/cls_pointer.c
1017           testsuite/libffi.call/cls_pointer_stack.c
1018           testsuite/libffi.call/err_bad_abi.c
1019           testsuite/libffi.call/err_bad_typedef.c
1020           testsuite/libffi.call/stret_large2.c
1021           testsuite/libffi.call/stret_large.c
1022           testsuite/libffi.call/stret_medium2.c
1023           testsuite/libffi.call/stret_medium.c: New tests from Apple.
1024
1025 2009-06-05  Andrew Haley  <aph@redhat.com>
1026
1027         * src/x86/ffitarget.h, src/x86/ffi.c: Merge stdcall changes from
1028         libffi.
1029
1030 2009-06-04  Andrew Haley  <aph@redhat.com>
1031
1032         * src/x86/ffitarget.h, src/x86/win32.S, src/x86/ffi.c: Back out
1033         stdcall changes.
1034
1035 2008-02-26  Anthony Green  <green@redhat.com>
1036             Thomas Heller  <theller@ctypes.org>
1037
1038         * src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C
1039         comment.
1040
1041 2008-02-03  Timothy Wall  <twall@users.sf.net>
1042
1043         * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return
1044           offset based on code pointer, not data pointer.
1045
1046 2008-01-31  Timothy Wall <twall@users.sf.net>
1047
1048         * testsuite/libffi.call/closure_stdcall.c: Add test for stdcall
1049         closures.
1050         * src/x86/ffitarget.h: Increase size of trampoline for stdcall
1051         closures.
1052         * src/x86/win32.S: Add assembly for stdcall closure.
1053         * src/x86/ffi.c: Initialize stdcall closure trampoline.
1054
1055 2009-06-04  Andrew Haley  <aph@redhat.com>
1056
1057         * include/ffi.h.in: Change void (*)() to void (*)(void).
1058         * src/x86/ffi.c: Likewise.
1059
1060 2009-06-04  Andrew Haley  <aph@redhat.com>
1061
1062         * src/powerpc/ppc_closure.S: Insert licence header.
1063         * src/powerpc/linux64_closure.S: Likewise.
1064         * src/m68k/sysv.S: Likewise.
1065
1066         * src/sh64/ffi.c: Change void (*)() to void (*)(void).
1067         * src/powerpc/ffi.c: Likewise.
1068         * src/powerpc/ffi_darwin.c: Likewise.
1069         * src/m32r/ffi.c: Likewise.
1070         * src/sh64/ffi.c: Likewise.
1071         * src/x86/ffi64.c: Likewise.
1072         * src/alpha/ffi.c: Likewise.
1073         * src/alpha/osf.S: Likewise.
1074         * src/frv/ffi.c: Likewise.
1075         * src/s390/ffi.c: Likewise.
1076         * src/pa/ffi.c: Likewise.
1077         * src/pa/hpux32.S: Likewise.
1078         * src/ia64/unix.S: Likewise.
1079         * src/ia64/ffi.c: Likewise.
1080         * src/sparc/ffi.c: Likewise.
1081         * src/mips/ffi.c: Likewise.
1082         * src/sh/ffi.c: Likewise.
1083
1084 2008-02-15  David Daney  <ddaney@avtrex.com>
1085
1086         * src/mips/ffi.c (USE__BUILTIN___CLEAR_CACHE):
1087         Define (conditionally), and use it to include cachectl.h.
1088         (ffi_prep_closure_loc): Fix cache flushing.
1089         * src/mips/ffitarget.h (_ABIN32, _ABI64, _ABIO32): Define.
1090
1091 2009-06-04  Andrew Haley  <aph@redhat.com>
1092
1093         include/ffi.h.in,
1094         src/arm/ffitarget.h,
1095         src/arm/ffi.c,
1096         src/arm/sysv.S,
1097         src/powerpc/ffitarget.h,
1098         src/closures.c,
1099         src/sh64/ffitarget.h,
1100         src/sh64/ffi.c,
1101         src/sh64/sysv.S,
1102         src/types.c,
1103         src/x86/ffi64.c,
1104         src/x86/ffitarget.h,
1105         src/x86/win32.S,
1106         src/x86/darwin.S,
1107         src/x86/ffi.c,
1108         src/x86/sysv.S,
1109         src/x86/unix64.S,
1110         src/alpha/ffitarget.h,
1111         src/alpha/ffi.c,
1112         src/alpha/osf.S,
1113         src/m68k/ffitarget.h,
1114         src/frv/ffitarget.h,
1115         src/frv/ffi.c,
1116         src/s390/ffitarget.h,
1117         src/s390/sysv.S,
1118         src/cris/ffitarget.h,
1119         src/pa/linux.S,
1120         src/pa/ffitarget.h,
1121         src/pa/ffi.c,
1122         src/raw_api.c,
1123         src/ia64/ffitarget.h,
1124         src/ia64/unix.S,
1125         src/ia64/ffi.c,
1126         src/ia64/ia64_flags.h,
1127         src/java_raw_api.c,
1128         src/debug.c,
1129         src/sparc/v9.S,
1130         src/sparc/ffitarget.h,
1131         src/sparc/ffi.c,
1132         src/sparc/v8.S,
1133         src/mips/ffitarget.h,
1134         src/mips/n32.S,
1135         src/mips/o32.S,
1136         src/mips/ffi.c,
1137         src/prep_cif.c,
1138         src/sh/ffitarget.h,
1139         src/sh/ffi.c,
1140         src/sh/sysv.S: Update license text.
1141
1142 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
1143
1144         * src/x86/win32.S (_ffi_closure_STDCALL):  New function.
1145         (.eh_frame):  Add FDE for it.
1146
1147 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
1148
1149         * configure.ac:  Also check if assembler supports pc-relative
1150         relocs on X86_WIN32 targets.
1151         * configure:  Regenerate.
1152         * src/x86/win32.S (ffi_prep_args):  Declare extern, not global.
1153         (_ffi_call_SYSV):  Add missing function type symbol .def and
1154         add EH markup labels.
1155         (_ffi_call_STDCALL):  Likewise.
1156         (_ffi_closure_SYSV):  Likewise.
1157         (_ffi_closure_raw_SYSV):  Likewise.
1158         (.eh_frame):  Add hand-crafted EH data.
1159
1160 2009-04-09  Jakub Jelinek  <jakub@redhat.com>
1161
1162         * testsuite/lib/libffi-dg.exp: Change copyright header to refer to
1163         version 3 of the GNU General Public License and to point readers
1164         at the COPYING3 file and the FSF's license web page.
1165         * testsuite/libffi.call/call.exp: Likewise.
1166         * testsuite/libffi.special/special.exp: Likewise.
1167
1168 2009-03-01  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1169
1170         * configure: Regenerate.
1171
1172 2008-12-18  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1173
1174         PR libffi/26048
1175         * configure.ac (HAVE_AS_X86_PCREL): New test.
1176         * configure: Regenerate.
1177         * fficonfig.h.in: Regenerate.
1178         * src/x86/sysv.S [!FFI_NO_RAW_API]: Precalculate
1179         RAW_CLOSURE_CIF_OFFSET, RAW_CLOSURE_FUN_OFFSET,
1180         RAW_CLOSURE_USER_DATA_OFFSET for the Solaris 10/x86 assembler.
1181         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
1182         * src/x86/unix64.S (.Lstore_table): Move to .text section.
1183         (.Lload_table): Likewise.
1184         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
1185
1186 2008-12-18  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1187
1188         * configure: Regenerate.
1189
1190 2008-11-21  Eric Botcazou  <ebotcazou@adacore.com>
1191
1192         * src/sparc/ffi.c (ffi_prep_cif_machdep): Add support for
1193         signed/unsigned int8/16 return values.
1194         * src/sparc/v8.S (ffi_call_v8): Likewise.
1195         (ffi_closure_v8): Likewise.
1196
1197 2008-09-26  Peter O'Gorman  <pogma@thewrittenword.com>
1198             Steve Ellcey  <sje@cup.hp.com>
1199
1200         * configure: Regenerate for new libtool.
1201         * Makefile.in: Ditto.
1202         * include/Makefile.in: Ditto.
1203         * aclocal.m4: Ditto.
1204
1205 2008-08-25  Andreas Tobler  <a.tobler@schweiz.org>
1206
1207         * src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and
1208         FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum.
1209         Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT.
1210         Adjust copyright notice.
1211         * src/powerpc/ffi.c: Add two new flags to indicate if we have one
1212         register or two register to use for FFI_SYSV structs.
1213         (ffi_prep_cif_machdep): Pass the right register flag introduced above.
1214         (ffi_closure_helper_SYSV): Fix the return type for
1215         FFI_SYSV_TYPE_SMALL_STRUCT. Comment.
1216         Adjust copyright notice.
1217
1218 2008-07-16  Kaz Kojima  <kkojima@gcc.gnu.org>
1219
1220         * src/sh/ffi.c (ffi_prep_closure_loc): Turn INSN into an unsigned
1221         int.
1222
1223 2008-06-17  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1224
1225         * configure: Regenerate.
1226         * include/Makefile.in: Regenerate.
1227         * testsuite/Makefile.in: Regenerate.
1228
1229 2008-06-07  Joseph Myers  <joseph@codesourcery.com>
1230
1231         * configure.ac (parisc*-*-linux*, powerpc-*-sysv*,
1232         powerpc-*-beos*): Remove.
1233         * configure: Regenerate.
1234
1235 2008-05-09  Julian Brown  <julian@codesourcery.com>
1236
1237         * Makefile.am (LTLDFLAGS): New.
1238         (libffi_la_LDFLAGS): Use above.
1239         * Makefile.in: Regenerate.
1240
1241 2008-04-18  Paolo Bonzini  <bonzini@gnu.org>
1242
1243         PR bootstrap/35457
1244         * aclocal.m4: Regenerate.
1245         * configure: Regenerate.
1246
1247 2008-03-26  Kaz Kojima  <kkojima@gcc.gnu.org>
1248
1249         * src/sh/sysv.S: Add .note.GNU-stack on Linux.
1250         * src/sh64/sysv.S: Likewise.
1251
1252 2008-03-26  Daniel Jacobowitz  <dan@debian.org>
1253
1254         * src/arm/sysv.S: Fix ARM comment marker.
1255
1256 2008-03-26  Jakub Jelinek  <jakub@redhat.com>
1257
1258         * src/alpha/osf.S: Add .note.GNU-stack on Linux.
1259         * src/s390/sysv.S: Likewise.
1260         * src/powerpc/ppc_closure.S: Likewise.
1261         * src/powerpc/sysv.S: Likewise.
1262         * src/x86/unix64.S: Likewise.
1263         * src/x86/sysv.S: Likewise.
1264         * src/sparc/v8.S: Likewise.
1265         * src/sparc/v9.S: Likewise.
1266         * src/m68k/sysv.S: Likewise.
1267         * src/arm/sysv.S: Likewise.
1268
1269 2008-03-16  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1270
1271         * aclocal.m4: Regenerate.
1272         * configure: Likewise.
1273         * Makefile.in: Likewise.
1274         * include/Makefile.in: Likewise.
1275         * testsuite/Makefile.in: Likewise.
1276
1277 2008-02-12  Bjoern Koenig  <bkoenig@alpha-tierchen.de>
1278             Andreas Tobler  <a.tobler@schweiz.org>
1279
1280         * configure.ac: Add amd64-*-freebsd* target.
1281         * configure: Regenerate.
1282
1283 2008-01-30  H.J. Lu  <hongjiu.lu@intel.com>
1284
1285         PR libffi/34612
1286         * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when
1287         returning struct.
1288
1289         * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer"
1290         tests.
1291
1292 2008-01-24  David Edelsohn  <edelsohn@gnu.org>
1293
1294         * configure: Regenerate.
1295
1296 2008-01-06  Andreas Tobler  <a.tobler@schweiz.org>
1297
1298         * src/x86/ffi.c (ffi_prep_cif_machdep): Fix thinko.
1299
1300 2008-01-05  Andreas Tobler  <a.tobler@schweiz.org>
1301
1302         PR testsuite/32843
1303         * src/x86/ffi.c (ffi_prep_cif_machdep): Add code for
1304         signed/unsigned int8/16 for X86_DARWIN.
1305         Updated copyright info.
1306         Handle one and two byte structs with special cif->flags.
1307         * src/x86/ffitarget.h: Add special types for one and two byte structs.
1308         Updated copyright info.
1309         * src/x86/darwin.S (ffi_call_SYSV): Rewrite to use a jump table like
1310         sysv.S
1311         Remove code to pop args from the stack after call.
1312         Special-case signed/unsigned for int8/16, one and two byte structs.
1313         (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1314         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1315         FFI_TYPE_SINT32.
1316         Updated copyright info.
1317
1318 2007-12-08  David Daney  <ddaney@avtrex.com>
1319
1320         * src/mips/n32.S (ffi_call_N32):  Replace dadd with ADDU, dsub with
1321         SUBU, add with ADDU and use smaller code sequences.
1322
1323 2007-12-07  David Daney  <ddaney@avtrex.com>
1324
1325         * src/mips/ffi.c (ffi_prep_cif_machdep): Handle long double return
1326         type.
1327
1328 2007-12-06  David Daney  <ddaney@avtrex.com>
1329
1330         * include/ffi.h.in (FFI_SIZEOF_JAVA_RAW): Define if not already
1331         defined.
1332         (ffi_java_raw): New typedef.
1333         (ffi_java_raw_call, ffi_java_ptrarray_to_raw,
1334         ffi_java_raw_to_ptrarray): Change parameter types from ffi_raw to
1335         ffi_java_raw.
1336         (ffi_java_raw_closure) : Same.
1337         (ffi_prep_java_raw_closure, ffi_prep_java_raw_closure_loc): Change
1338         parameter types.
1339         * src/java_raw_api.c (ffi_java_raw_size):  Replace FFI_SIZEOF_ARG with
1340         FFI_SIZEOF_JAVA_RAW.
1341         (ffi_java_raw_to_ptrarray): Change type of raw to ffi_java_raw.
1342         Replace FFI_SIZEOF_ARG with FFI_SIZEOF_JAVA_RAW. Use
1343         sizeof(ffi_java_raw) for alignment calculations.
1344         (ffi_java_ptrarray_to_raw): Same.
1345         (ffi_java_rvalue_to_raw): Add special handling for FFI_TYPE_POINTER
1346         if FFI_SIZEOF_JAVA_RAW == 4.
1347         (ffi_java_raw_to_rvalue): Same.
1348         (ffi_java_raw_call): Change type of raw to ffi_java_raw.
1349         (ffi_java_translate_args): Same.
1350         (ffi_prep_java_raw_closure_loc, ffi_prep_java_raw_closure): Change
1351         parameter types.
1352         * src/mips/ffitarget.h (FFI_SIZEOF_JAVA_RAW): Define for N32 ABI.
1353
1354 2007-12-06  David Daney  <ddaney@avtrex.com>
1355
1356         * src/mips/n32.S (ffi_closure_N32): Use 64-bit add instruction on
1357         pointer values.
1358
1359 2007-12-01  Andreas Tobler  <a.tobler@schweiz.org>
1360
1361         PR libffi/31937
1362         * src/powerpc/ffitarget.h: Introduce new ABI FFI_LINUX_SOFT_FLOAT.
1363         Add local FFI_TYPE_UINT128 to handle soft-float long-double-128.
1364         * src/powerpc/ffi.c: Distinguish between __NO_FPRS__ and not and
1365         set the NUM_FPR_ARG_REGISTERS according to.
1366         Add support for potential soft-float support under hard-float
1367         architecture.
1368         (ffi_prep_args_SYSV): Set NUM_FPR_ARG_REGISTERS to 0 in case of
1369         FFI_LINUX_SOFT_FLOAT, handle float, doubles and long-doubles according
1370         to the FFI_LINUX_SOFT_FLOAT ABI.
1371         (ffi_prep_cif_machdep): Likewise.
1372         (ffi_closure_helper_SYSV): Likewise.
1373         * src/powerpc/ppc_closure.S: Make sure not to store float/double
1374         on archs where __NO_FPRS__ is true.
1375         Add FFI_TYPE_UINT128 support.
1376         * src/powerpc/sysv.S: Add support for soft-float long-double-128.
1377         Adjust copyright notice.
1378
1379 2007-11-25  Andreas Tobler  <a.tobler@schweiz.org>
1380
1381         * src/closures.c: Move defintion of MAYBE_UNUSED from here to ...
1382         * include/ffi_common.h: ... here.
1383         Update copyright.
1384
1385 2007-11-17  Andreas Tobler  <a.tobler@schweiz.org>
1386
1387         * src/powerpc/sysv.S: Load correct cr to compare if we have long double.
1388         * src/powerpc/linux64.S: Likewise.
1389         * src/powerpc/ffi.c: Add a comment to show which part goes into cr6.
1390         * testsuite/libffi.call/return_ldl.c: New test.
1391
1392 2007-09-04    <aph@redhat.com>
1393
1394         * src/arm/sysv.S (UNWIND): New.
1395         (Whole file): Conditionally compile unwinder directives.
1396         * src/arm/sysv.S: Add unwinder directives.
1397
1398         * src/arm/ffi.c (ffi_prep_args): Align structs by at least 4 bytes.
1399         Only treat r0 as a struct address if we're actually returning a
1400         struct by address.
1401         Only copy the bytes that are actually within a struct.
1402         (ffi_prep_cif_machdep): A Composite Type not larger than 4 bytes
1403         is returned in r0, not passed by address.
1404         (ffi_call): Allocate a word-sized temporary for the case where
1405         a composite is returned in r0.
1406         (ffi_prep_incoming_args_SYSV): Align as necessary.
1407
1408 2007-08-05  Steven Newbury  <s_j_newbury@yahoo.co.uk>
1409
1410         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Use __clear_cache instead of
1411         directly using the sys_cacheflush syscall.
1412
1413 2007-07-27  Andrew Haley  <aph@redhat.com>
1414
1415         * src/arm/sysv.S (ffi_closure_SYSV): Add soft-float.
1416
1417 2007-09-03  Maciej W. Rozycki  <macro@linux-mips.org>
1418
1419         * Makefile.am: Unify MIPS_IRIX and MIPS_LINUX into MIPS.
1420         * configure.ac: Likewise.
1421         * Makefile.in: Regenerate.
1422         * include/Makefile.in: Likewise.
1423         * testsuite/Makefile.in: Likewise.
1424         * configure: Likewise.
1425
1426 2007-08-24  David Daney  <ddaney@avtrex.com>
1427
1428         * testsuite/libffi.call/return_sl.c: New test.
1429
1430 2007-08-10  David Daney  <ddaney@avtrex.com>
1431
1432         * testsuite/libffi.call/cls_multi_ushort.c,
1433         testsuite/libffi.call/cls_align_uint16.c,
1434         testsuite/libffi.call/nested_struct1.c,
1435         testsuite/libffi.call/nested_struct3.c,
1436         testsuite/libffi.call/cls_7_1_byte.c,
1437         testsuite/libffi.call/nested_struct5.c,
1438         testsuite/libffi.call/cls_double.c,
1439         testsuite/libffi.call/nested_struct7.c,
1440         testsuite/libffi.call/cls_sint.c,
1441         testsuite/libffi.call/nested_struct9.c,
1442         testsuite/libffi.call/cls_20byte1.c,
1443         testsuite/libffi.call/cls_multi_sshortchar.c,
1444         testsuite/libffi.call/cls_align_sint64.c,
1445         testsuite/libffi.call/cls_3byte2.c,
1446         testsuite/libffi.call/cls_multi_schar.c,
1447         testsuite/libffi.call/cls_multi_uchar.c,
1448         testsuite/libffi.call/cls_19byte.c,
1449         testsuite/libffi.call/cls_9byte1.c,
1450         testsuite/libffi.call/cls_align_float.c,
1451         testsuite/libffi.call/closure_fn1.c,
1452         testsuite/libffi.call/problem1.c,
1453         testsuite/libffi.call/closure_fn3.c,
1454         testsuite/libffi.call/cls_sshort.c,
1455         testsuite/libffi.call/closure_fn5.c,
1456         testsuite/libffi.call/cls_align_double.c,
1457         testsuite/libffi.call/nested_struct.c,
1458         testsuite/libffi.call/cls_2byte.c,
1459         testsuite/libffi.call/nested_struct10.c,
1460         testsuite/libffi.call/cls_4byte.c,
1461         testsuite/libffi.call/cls_6byte.c,
1462         testsuite/libffi.call/cls_8byte.c,
1463         testsuite/libffi.call/cls_multi_sshort.c,
1464         testsuite/libffi.call/cls_align_sint16.c,
1465         testsuite/libffi.call/cls_align_uint32.c,
1466         testsuite/libffi.call/cls_20byte.c,
1467         testsuite/libffi.call/cls_float.c,
1468         testsuite/libffi.call/nested_struct2.c,
1469         testsuite/libffi.call/cls_5_1_byte.c,
1470         testsuite/libffi.call/nested_struct4.c,
1471         testsuite/libffi.call/cls_24byte.c,
1472         testsuite/libffi.call/nested_struct6.c,
1473         testsuite/libffi.call/cls_64byte.c,
1474         testsuite/libffi.call/nested_struct8.c,
1475         testsuite/libffi.call/cls_uint.c,
1476         testsuite/libffi.call/cls_multi_ushortchar.c,
1477         testsuite/libffi.call/cls_schar.c,
1478         testsuite/libffi.call/cls_uchar.c,
1479         testsuite/libffi.call/cls_align_uint64.c,
1480         testsuite/libffi.call/cls_ulonglong.c,
1481         testsuite/libffi.call/cls_align_longdouble.c,
1482         testsuite/libffi.call/cls_1_1byte.c,
1483         testsuite/libffi.call/cls_12byte.c,
1484         testsuite/libffi.call/cls_3_1byte.c,
1485         testsuite/libffi.call/cls_3byte1.c,
1486         testsuite/libffi.call/cls_4_1byte.c,
1487         testsuite/libffi.call/cls_6_1_byte.c,
1488         testsuite/libffi.call/cls_16byte.c,
1489         testsuite/libffi.call/cls_18byte.c,
1490         testsuite/libffi.call/closure_fn0.c,
1491         testsuite/libffi.call/cls_9byte2.c,
1492         testsuite/libffi.call/closure_fn2.c,
1493         testsuite/libffi.call/closure_fn4.c,
1494         testsuite/libffi.call/cls_ushort.c,
1495         testsuite/libffi.call/closure_fn6.c,
1496         testsuite/libffi.call/cls_5byte.c,
1497         testsuite/libffi.call/cls_align_pointer.c,
1498         testsuite/libffi.call/cls_7byte.c,
1499         testsuite/libffi.call/cls_align_sint32.c,
1500         testsuite/libffi.special/unwindtest_ffi_call.cc,
1501         testsuite/libffi.special/unwindtest.cc: Remove xfail for mips64*-*-*.
1502
1503 2007-08-10  David Daney  <ddaney@avtrex.com>
1504
1505         PR libffi/28313
1506         * configure.ac: Don't treat mips64 as a special case.
1507         * Makefile.am (nodist_libffi_la_SOURCES): Add n32.S.
1508         * configure: Regenerate
1509         * Makefile.in: Ditto.
1510         * fficonfig.h.in: Ditto.
1511         * src/mips/ffitarget.h (REG_L, REG_S, SUBU, ADDU, SRL, LI): Indent.
1512         (LA, EH_FRAME_ALIGN, FDE_ADDR_BYTES): New preprocessor macros.
1513         (FFI_DEFAULT_ABI): Set for n64 case.
1514         (FFI_CLOSURES, FFI_TRAMPOLINE_SIZE): Define for n32 and n64 cases.
1515         * src/mips/n32.S (ffi_call_N32): Add debug macros and labels for FDE.
1516         (ffi_closure_N32): New function.
1517         (.eh_frame): New section
1518         * src/mips/o32.S: Clean up comments.
1519         (ffi_closure_O32): Pass ffi_closure parameter in $12.
1520         * src/mips/ffi.c: Use FFI_MIPS_N32 instead of
1521         _MIPS_SIM == _ABIN32 throughout.
1522         (FFI_MIPS_STOP_HERE): New, use in place of
1523         ffi_stop_here.
1524         (ffi_prep_args): Use unsigned long to hold pointer values.  Rewrite
1525         to support n32/n64 ABIs.
1526         (calc_n32_struct_flags): Rewrite.
1527         (calc_n32_return_struct_flags): Remove unused variable.  Reverse
1528         position of flag bits.
1529         (ffi_prep_cif_machdep): Rewrite n32 portion.
1530         (ffi_call): Enable for n64.  Add special handling for small structure
1531         return values.
1532         (ffi_prep_closure_loc): Add n32 and n64 support.
1533         (ffi_closure_mips_inner_O32): Add cast to silence warning.
1534         (copy_struct_N32, ffi_closure_mips_inner_N32): New functions.
1535
1536 2007-08-08  David Daney  <ddaney@avtrex.com>
1537
1538         * testsuite/libffi.call/ffitest.h (ffi_type_mylong): Remove definition.
1539         * testsuite/libffi.call/cls_align_uint16.c (main): Use correct type
1540         specifiers.
1541         * testsuite/libffi.call/nested_struct1.c (main): Ditto.
1542         * testsuite/libffi.call/cls_sint.c (main): Ditto.
1543         * testsuite/libffi.call/nested_struct9.c (main): Ditto.
1544         * testsuite/libffi.call/cls_20byte1.c (main): Ditto.
1545         * testsuite/libffi.call/cls_9byte1.c (main): Ditto.
1546         * testsuite/libffi.call/closure_fn1.c (main): Ditto.
1547         * testsuite/libffi.call/closure_fn3.c (main): Ditto.
1548         * testsuite/libffi.call/return_dbl2.c (main): Ditto.
1549         * testsuite/libffi.call/cls_sshort.c (main): Ditto.
1550         * testsuite/libffi.call/return_fl3.c (main): Ditto.
1551         * testsuite/libffi.call/closure_fn5.c (main): Ditto.
1552         * testsuite/libffi.call/nested_struct.c (main): Ditto.
1553         * testsuite/libffi.call/nested_struct10.c (main): Ditto.
1554         * testsuite/libffi.call/return_ll1.c (main): Ditto.
1555         * testsuite/libffi.call/cls_8byte.c (main): Ditto.
1556         * testsuite/libffi.call/cls_align_uint32.c (main): Ditto.
1557         * testsuite/libffi.call/cls_align_sint16.c (main): Ditto.
1558         * testsuite/libffi.call/cls_20byte.c (main): Ditto.
1559         * testsuite/libffi.call/nested_struct2.c (main): Ditto.
1560         * testsuite/libffi.call/cls_24byte.c (main): Ditto.
1561         * testsuite/libffi.call/nested_struct6.c (main): Ditto.
1562         * testsuite/libffi.call/cls_uint.c (main): Ditto.
1563         * testsuite/libffi.call/cls_12byte.c (main): Ditto.
1564         * testsuite/libffi.call/cls_16byte.c (main): Ditto.
1565         * testsuite/libffi.call/closure_fn0.c (main): Ditto.
1566         * testsuite/libffi.call/cls_9byte2.c (main): Ditto.
1567         * testsuite/libffi.call/closure_fn2.c (main): Ditto.
1568         * testsuite/libffi.call/return_dbl1.c (main): Ditto.
1569         * testsuite/libffi.call/closure_fn4.c (main): Ditto.
1570         * testsuite/libffi.call/closure_fn6.c (main): Ditto.
1571         * testsuite/libffi.call/cls_align_sint32.c (main): Ditto.
1572
1573 2007-08-07  Andrew Haley  <aph@redhat.com>
1574
1575         * src/x86/sysv.S (ffi_closure_raw_SYSV): Fix typo in previous
1576         checkin.
1577
1578 2007-08-06  Andrew Haley  <aph@redhat.com>
1579
1580         PR testsuite/32843
1581         * src/x86/sysv.S (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1582         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1583         FFI_TYPE_SINT32.
1584
1585 2007-08-02  David Daney  <ddaney@avtrex.com>
1586
1587         * testsuite/libffi.call/return_ul.c (main): Define return type as
1588         ffi_arg.  Use proper printf conversion specifier.
1589
1590 2007-07-30  Andrew Haley  <aph@redhat.com>
1591
1592         PR testsuite/32843
1593         * src/x86/ffi.c (ffi_prep_cif_machdep): in x86 case, add code for
1594         signed/unsigned int8/16.
1595         * src/x86/sysv.S (ffi_call_SYSV): Rewrite to:
1596         Use a jump table.
1597         Remove code to pop args from the stack after call.
1598         Special-case signed/unsigned int8/16.
1599         * testsuite/libffi.call/return_sc.c (main): Revert.
1600
1601 2007-07-26  Richard Guenther  <rguenther@suse.de>
1602
1603         PR testsuite/32843
1604         * testsuite/libffi.call/return_sc.c (main): Verify call
1605         result as signed char, not ffi_arg.
1606
1607 2007-07-16  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1608
1609         * configure.ac (i?86-*-solaris2.1[0-9]): Set TARGET to X86_64.
1610         * configure: Regenerate.
1611
1612 2007-07-11  David Daney  <ddaney@avtrex.com>
1613
1614         * src/mips/ffi.c: Don't include sys/cachectl.h.
1615         (ffi_prep_closure_loc): Use __builtin___clear_cache() instead of
1616         cacheflush().
1617
1618 2007-05-18  Aurelien Jarno  <aurelien@aurel32.net>
1619
1620         * src/arm/ffi.c (ffi_prep_closure_loc): Renamed and ajusted
1621         from (ffi_prep_closure): ... this.
1622         (FFI_INIT_TRAMPOLINE): Adjust.
1623
1624 2005-12-31  Phil Blundell  <pb@reciva.com>
1625
1626         * src/arm/ffi.c (ffi_prep_incoming_args_SYSV,
1627         ffi_closure_SYSV_inner, ffi_prep_closure): New, add closure support.
1628         * src/arm/sysv.S(ffi_closure_SYSV): Likewise.
1629         * src/arm/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1630         (FFI_CLOSURES): Enable closure support.
1631
1632 2007-07-03  Andrew Haley  <aph@hedges.billgatliff.com>
1633
1634         * testsuite/libffi.call/cls_multi_ushort.c,
1635         testsuite/libffi.call/cls_align_uint16.c,
1636         testsuite/libffi.call/nested_struct1.c,
1637         testsuite/libffi.call/nested_struct3.c,
1638         testsuite/libffi.call/cls_7_1_byte.c,
1639         testsuite/libffi.call/cls_double.c,
1640         testsuite/libffi.call/nested_struct5.c,
1641         testsuite/libffi.call/nested_struct7.c,
1642         testsuite/libffi.call/cls_sint.c,
1643         testsuite/libffi.call/nested_struct9.c,
1644         testsuite/libffi.call/cls_20byte1.c,
1645         testsuite/libffi.call/cls_multi_sshortchar.c,
1646         testsuite/libffi.call/cls_align_sint64.c,
1647         testsuite/libffi.call/cls_3byte2.c,
1648         testsuite/libffi.call/cls_multi_schar.c,
1649         testsuite/libffi.call/cls_multi_uchar.c,
1650         testsuite/libffi.call/cls_19byte.c,
1651         testsuite/libffi.call/cls_9byte1.c,
1652         testsuite/libffi.call/cls_align_float.c,
1653         testsuite/libffi.call/closure_fn1.c,
1654         testsuite/libffi.call/problem1.c,
1655         testsuite/libffi.call/closure_fn3.c,
1656         testsuite/libffi.call/cls_sshort.c,
1657         testsuite/libffi.call/closure_fn5.c,
1658         testsuite/libffi.call/cls_align_double.c,
1659         testsuite/libffi.call/cls_2byte.c,
1660         testsuite/libffi.call/nested_struct.c,
1661         testsuite/libffi.call/nested_struct10.c,
1662         testsuite/libffi.call/cls_4byte.c,
1663         testsuite/libffi.call/cls_6byte.c,
1664         testsuite/libffi.call/cls_8byte.c,
1665         testsuite/libffi.call/cls_multi_sshort.c,
1666         testsuite/libffi.call/cls_align_uint32.c,
1667         testsuite/libffi.call/cls_align_sint16.c,
1668         testsuite/libffi.call/cls_float.c,
1669         testsuite/libffi.call/cls_20byte.c,
1670         testsuite/libffi.call/cls_5_1_byte.c,
1671         testsuite/libffi.call/nested_struct2.c,
1672         testsuite/libffi.call/cls_24byte.c,
1673         testsuite/libffi.call/nested_struct4.c,
1674         testsuite/libffi.call/nested_struct6.c,
1675         testsuite/libffi.call/cls_64byte.c,
1676         testsuite/libffi.call/nested_struct8.c,
1677         testsuite/libffi.call/cls_uint.c,
1678         testsuite/libffi.call/cls_multi_ushortchar.c,
1679         testsuite/libffi.call/cls_schar.c,
1680         testsuite/libffi.call/cls_uchar.c,
1681         testsuite/libffi.call/cls_align_uint64.c,
1682         testsuite/libffi.call/cls_ulonglong.c,
1683         testsuite/libffi.call/cls_align_longdouble.c,
1684         testsuite/libffi.call/cls_1_1byte.c,
1685         testsuite/libffi.call/cls_12byte.c,
1686         testsuite/libffi.call/cls_3_1byte.c,
1687         testsuite/libffi.call/cls_3byte1.c,
1688         testsuite/libffi.call/cls_4_1byte.c,
1689         testsuite/libffi.call/cls_6_1_byte.c,
1690         testsuite/libffi.call/cls_16byte.c,
1691         testsuite/libffi.call/cls_18byte.c,
1692         testsuite/libffi.call/closure_fn0.c,
1693         testsuite/libffi.call/cls_9byte2.c,
1694         testsuite/libffi.call/closure_fn2.c,
1695         testsuite/libffi.call/closure_fn4.c,
1696         testsuite/libffi.call/cls_ushort.c,
1697         testsuite/libffi.call/closure_fn6.c,
1698         testsuite/libffi.call/cls_5byte.c,
1699         testsuite/libffi.call/cls_align_pointer.c,
1700         testsuite/libffi.call/cls_7byte.c,
1701         testsuite/libffi.call/cls_align_sint32.c,
1702         testsuite/libffi.special/unwindtest_ffi_call.cc,
1703         testsuite/libffi.special/unwindtest.cc: Enable for ARM.
1704
1705 2007-07-05  H.J. Lu  <hongjiu.lu@intel.com>
1706
1707         * aclocal.m4: Regenerated.
1708
1709 2007-06-02  Paolo Bonzini  <bonzini@gnu.org>
1710
1711         * configure: Regenerate.
1712
1713 2007-05-23  Steve Ellcey  <sje@cup.hp.com>
1714
1715         * Makefile.in: Regenerate.
1716         * configure: Regenerate.
1717         * aclocal.m4: Regenerate.
1718         * include/Makefile.in: Regenerate.
1719         * testsuite/Makefile.in: Regenerate.
1720
1721 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1722
1723         * src/m68k/ffi.c (ffi_prep_incoming_args_SYSV,
1724         ffi_closure_SYSV_inner,ffi_prep_closure): New, add closure support.
1725         * src/m68k/sysv.S(ffi_closure_SYSV,ffi_closure_struct_SYSV): Likewise.
1726         * src/m68k/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1727         (FFI_CLOSURES): Enable closure support.
1728
1729 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1730
1731         * configure.ac (HAVE_AS_CFI_PSEUDO_OP): New test.
1732         * configure: Regenerate.
1733         * fficonfig.h.in: Regenerate.
1734         * src/m68k/sysv.S (CFI_STARTPROC,CFI_ENDPROC,
1735         CFI_OFFSET,CFI_DEF_CFA): New macros.
1736         (ffi_call_SYSV): Add callframe annotation.
1737
1738 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1739
1740         * src/m68k/ffi.c (ffi_prep_args,ffi_prep_cif_machdep): Fix
1741         numerous test suite failures.
1742         * src/m68k/sysv.S (ffi_call_SYSV): Likewise.
1743
1744 2007-04-11  Paolo Bonzini  <bonzini@gnu.org>
1745
1746         * Makefile.am (EXTRA_DIST): Bring up to date.
1747         * Makefile.in: Regenerate.
1748         * src/frv/eabi.S: Remove RCS keyword.
1749
1750 2007-04-06  Richard Henderson  <rth@redhat.com>
1751
1752         * configure.ac: Tidy target case.
1753         (HAVE_LONG_DOUBLE): Allow the target to override.
1754         * configure: Regenerate.
1755         * include/ffi.h.in: Don't define ffi_type_foo if
1756         LIBFFI_HIDE_BASIC_TYPES is defined.
1757         (ffi_type_longdouble): If not HAVE_LONG_DOUBLE, define
1758         to ffi_type_double.
1759         * types.c (LIBFFI_HIDE_BASIC_TYPES): Define.
1760         (FFI_TYPEDEF, ffi_type_void): Mark the data const.
1761         (ffi_type_longdouble): Special case for Alpha.  Don't define
1762         if long double == double.
1763
1764         * src/alpha/ffi.c (FFI_TYPE_LONGDOUBLE): Assert unique value.
1765         (ffi_prep_cif_machdep): Handle it as the 128-bit type.
1766         (ffi_call, ffi_closure_osf_inner): Likewise.
1767         (ffi_closure_osf_inner): Likewise.  Mark hidden.
1768         (ffi_call_osf, ffi_closure_osf): Mark hidden.
1769         * src/alpha/ffitarget.h (FFI_LAST_ABI): Tidy definition.
1770         * src/alpha/osf.S (ffi_call_osf, ffi_closure_osf): Mark hidden.
1771         (load_table): Handle 128-bit long double.
1772
1773         * testsuite/libffi.call/float4.c: Add -mieee for alpha.
1774
1775 2007-04-06  Tom Tromey  <tromey@redhat.com>
1776
1777         PR libffi/31491:
1778         * README: Fixed bug in example.
1779
1780 2007-04-03  Jakub Jelinek  <jakub@redhat.com>
1781
1782         * src/closures.c: Include sys/statfs.h.
1783         (_GNU_SOURCE): Define on Linux.
1784         (FFI_MMAP_EXEC_SELINUX): Define.
1785         (selinux_enabled): New variable.
1786         (selinux_enabled_check): New function.
1787         (is_selinux_enabled): Define.
1788         (dlmmap): Use it.
1789
1790 2007-03-24  Uros Bizjak  <ubizjak@gmail.com>
1791
1792         * testsuite/libffi.call/return_fl2.c (return_fl): Mark as static.
1793         Use 'volatile float sum' to create sum of floats to avoid false
1794         negative due to excess precision on ix86 targets.
1795         (main): Ditto.
1796
1797 2007-03-08  Alexandre Oliva  <aoliva@redhat.com>
1798
1799         * src/powerpc/ffi.c (flush_icache): Fix left-over from previous
1800         patch.
1801         (ffi_prep_closure_loc): Remove unneeded casts.  Add needed ones.
1802
1803 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1804
1805         * include/ffi.h.in (ffi_closure_alloc, ffi_closure_free): New.
1806         (ffi_prep_closure_loc): New.
1807         (ffi_prep_raw_closure_loc): New.
1808         (ffi_prep_java_raw_closure_loc): New.
1809         * src/closures.c: New file.
1810         * src/dlmalloc.c [FFI_MMAP_EXEC_WRIT] (struct malloc_segment):
1811         Replace sflags with exec_offset.
1812         [FFI_MMAP_EXEC_WRIT] (mmap_exec_offset, add_segment_exec_offset,
1813         sub_segment_exec_offset): New macros.
1814         (get_segment_flags, set_segment_flags, check_segment_merge): New
1815         macros.
1816         (is_mmapped_segment, is_extern_segment): Use get_segment_flags.
1817         (add_segment, sys_alloc, create_mspace, create_mspace_with_base,
1818         destroy_mspace): Use new macros.
1819         (sys_alloc): Silence warning.
1820         * Makefile.am (libffi_la_SOURCES): Add src/closures.c.
1821         * Makefile.in: Rebuilt.
1822         * src/prep_cif [FFI_CLOSURES] (ffi_prep_closure): Implement in
1823         terms of ffi_prep_closure_loc.
1824         * src/raw_api.c (ffi_prep_raw_closure_loc): Renamed and adjusted
1825         from...
1826         (ffi_prep_raw_closure): ... this.  Re-implement in terms of the
1827         renamed version.
1828         * src/java_raw_api (ffi_prep_java_raw_closure_loc): Renamed and
1829         adjusted from...
1830         (ffi_prep_java_raw_closure): ... this.  Re-implement in terms of
1831         the renamed version.
1832         * src/alpha/ffi.c (ffi_prep_closure_loc): Renamed from
1833         (ffi_prep_closure): ... this.
1834         * src/pa/ffi.c: Likewise.
1835         * src/cris/ffi.c: Likewise.  Adjust.
1836         * src/frv/ffi.c: Likewise.
1837         * src/ia64/ffi.c: Likewise.
1838         * src/mips/ffi.c: Likewise.
1839         * src/powerpc/ffi_darwin.c: Likewise.
1840         * src/s390/ffi.c: Likewise.
1841         * src/sh/ffi.c: Likewise.
1842         * src/sh64/ffi.c: Likewise.
1843         * src/sparc/ffi.c: Likewise.
1844         * src/x86/ffi64.c: Likewise.
1845         * src/x86/ffi.c: Likewise.
1846         (FFI_INIT_TRAMPOLINE): Adjust.
1847         (ffi_prep_raw_closure_loc): Renamed and adjusted from...
1848         (ffi_prep_raw_closure): ... this.
1849         * src/powerpc/ffi.c (ffi_prep_closure_loc): Renamed from
1850         (ffi_prep_closure): ... this.
1851         (flush_icache): Adjust.
1852
1853 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1854
1855         * src/dlmalloc.c: New file, imported version 2.8.3 of Doug
1856         Lea's malloc.
1857
1858 2007-03-01  Brooks Moses  <brooks.moses@codesourcery.com>
1859
1860         * Makefile.am: Add dummy install-pdf target.
1861         * Makefile.in: Regenerate
1862
1863 2007-02-13  Andreas Krebbel  <krebbel1@de.ibm.com>
1864
1865         * src/s390/ffi.c (ffi_prep_args, ffi_prep_cif_machdep,
1866         ffi_closure_helper_SYSV): Add long double handling.
1867
1868 2007-02-02  Jakub Jelinek  <jakub@redhat.com>
1869
1870         * src/powerpc/linux64.S (ffi_call_LINUX64): Move restore of r2
1871         immediately after bctrl instruction.
1872
1873 2007-01-18  Alexandre Oliva  <aoliva@redhat.com>
1874
1875         * Makefile.am (all-recursive, install-recursive,
1876         mostlyclean-recursive, clean-recursive, distclean-recursive,
1877         maintainer-clean-recursive): Add missing targets.
1878         * Makefile.in: Rebuilt.
1879
1880 2006-12-14  Andreas Tobler  <a.tobler@schweiz.org>
1881
1882         * configure.ac: Add TARGET for x86_64-*-darwin*.
1883         * Makefile.am (nodist_libffi_la_SOURCES): Add rules for 64-bit sources
1884         for X86_DARWIN.
1885         * src/x86/ffitarget.h: Set trampoline size for x86_64-*-darwin*.
1886         * src/x86/darwin64.S: New file for x86_64-*-darwin* support.
1887         * configure: Regenerate.
1888         * Makefile.in: Regenerate.
1889         * include/Makefile.in: Regenerate.
1890         * testsuite/Makefile.in: Regenerate.
1891         * testsuite/libffi.special/unwindtest_ffi_call.cc: New test case for
1892         ffi_call only.
1893
1894 2006-12-13  Andreas Tobler <a.tobler@schweiz.org>
1895
1896         * aclocal.m4: Regenerate with aclocal -I .. as written in the
1897         Makefile.am.
1898
1899 2006-10-31  Geoffrey Keating  <geoffk@apple.com>
1900
1901         * src/powerpc/ffi_darwin.c (darwin_adjust_aggregate_sizes): New.
1902         (ffi_prep_cif_machdep): Call darwin_adjust_aggregate_sizes for
1903         Darwin.
1904         * testsuite/libffi.call/nested_struct4.c: Remove Darwin XFAIL.
1905         * testsuite/libffi.call/nested_struct6.c: Remove Darwin XFAIL.
1906
1907 2006-10-10  Paolo Bonzini  <bonzini@gnu.org>
1908             Sandro Tolaini  <tolaini@libero.it>
1909
1910         * configure.ac [i*86-*-darwin*]: Set X86_DARWIN symbol and
1911         conditional.
1912         * configure: Regenerated.
1913         * Makefile.am (nodist_libffi_la_SOURCES) [X86_DARWIN]: New case.
1914         (EXTRA_DIST): Add src/x86/darwin.S.
1915         * Makefile.in: Regenerated.
1916         * include/Makefile.in: Regenerated.
1917         * testsuite/Makefile.in: Regenerated.
1918
1919         * src/x86/ffi.c (ffi_prep_cif_machdep) [X86_DARWIN]: Treat like
1920         X86_WIN32, and additionally align stack to 16 bytes.
1921         * src/x86/darwin.S: New, based on sysv.S.
1922         * src/prep_cif.c (ffi_prep_cif) [X86_DARWIN]: Align > 8-byte structs.
1923
1924 2006-09-12  David Daney  <ddaney@avtrex.com>
1925
1926         PR libffi/23935
1927         * include/Makefile.am: Install both ffi.h and ffitarget.h in
1928         $(libdir)/gcc/$(target_alias)/$(gcc_version)/include.
1929         * aclocal.m4: Regenerated for automake 1.9.6.
1930         * Makefile.in: Regenerated.
1931         * include/Makefile.in: Regenerated.
1932         * testsuite/Makefile.in: Regenerated.
1933
1934 2006-08-17  Andreas Tobler  <a.tobler@schweiz.ch>
1935
1936         * include/ffi_common.h (struct): Revert accidental commit.
1937
1938 2006-08-15  Andreas Tobler  <a.tobler@schweiz.ch>
1939
1940         * include/ffi_common.h: Remove lint directives.
1941         * include/ffi.h.in: Likewise.
1942
1943 2006-07-25  Torsten Schoenfeld  <kaffeetisch@gmx.de>
1944
1945         * include/ffi.h.in (ffi_type_ulong, ffi_type_slong): Define correctly
1946         for 32-bit architectures.
1947         * testsuite/libffi.call/return_ul.c: New test case.
1948
1949 2006-07-19  David Daney  <ddaney@avtrex.com>
1950
1951         * testsuite/libffi.call/closure_fn6.c: Remove xfail for mips,
1952         xfail remains for mips64.
1953
1954 2006-05-23  Carlos O'Donell  <carlos@codesourcery.com>
1955
1956         * Makefile.am: Add install-html target. Add install-html to .PHONY
1957         * Makefile.in: Regenerate.
1958         * aclocal.m4: Regenerate.
1959         * include/Makefile.in: Regenerate.
1960         * testsuite/Makefile.in: Regenerate.
1961
1962 2006-05-18  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1963
1964         * pa/ffi.c (ffi_prep_args_pa32): Load floating point arguments from
1965         stack slot.
1966
1967 2006-04-22  Andreas Tobler  <a.tobler@schweiz.ch>
1968
1969         * README: Remove notice about 'Crazy Comments'.
1970         * src/debug.c: Remove lint directives. Cleanup white spaces.
1971         * src/java_raw_api.c: Likewise.
1972         * src/prep_cif.c: Likewise.
1973         * src/raw_api.c: Likewise.
1974         * src/ffitest.c: Delete. No longer needed, all test cases migrated
1975         to the testsuite.
1976         * src/arm/ffi.c: Remove lint directives.
1977         * src/m32r/ffi.c: Likewise.
1978         * src/pa/ffi.c: Likewise.
1979         * src/powerpc/ffi.c: Likewise.
1980         * src/powerpc/ffi_darwin.c: Likewise.
1981         * src/sh/ffi.c: Likewise.
1982         * src/sh64/ffi.c: Likewise.
1983         * src/x86/ffi.c: Likewise.
1984         * testsuite/libffi.call/float2.c: Likewise.
1985         * testsuite/libffi.call/promotion.c: Likewise.
1986         * testsuite/libffi.call/struct1.c: Likewise.
1987
1988 2006-04-13  Andreas Tobler  <a.tobler@schweiz.ch>
1989
1990         * src/pa/hpux32.S: Correct unwind offset calculation for
1991         ffi_closure_pa32.
1992         * src/pa/linux.S: Likewise.
1993
1994 2006-04-12  James E Wilson  <wilson@specifix.com>
1995
1996         PR libgcj/26483
1997         * src/ia64/ffi.c (stf_spill, ldf_fill): Rewrite as macros.
1998         (hfa_type_load): Call stf_spill.
1999         (hfa_type_store): Call ldf_fill.
2000         (ffi_call): Adjust calls to above routines.  Add local temps for
2001         macro result.
2002
2003 2006-04-10  Matthias Klose  <doko@debian.org>
2004
2005         * testsuite/lib/libffi-dg.exp (libffi-init): Recognize multilib
2006         directory names containing underscores.
2007
2008 2006-04-07  James E Wilson  <wilson@specifix.com>
2009
2010         * testsuite/libffi.call/float4.c: New testcase.
2011
2012 2006-04-05  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2013             Andreas Tobler  <a.tobler@schweiz.ch>
2014
2015         * Makefile.am: Add PA_HPUX port.
2016         * Makefile.in: Regenerate.
2017         * include/Makefile.in: Likewise.
2018         * testsuite/Makefile.in: Likewise.
2019         * configure.ac: Add PA_HPUX rules.
2020         * configure: Regenerate.
2021         * src/pa/ffitarget.h: Rename linux target to PA_LINUX.
2022         Add PA_HPUX and PA64_HPUX.
2023         Rename FFI_LINUX ABI to FFI_PA32 ABI.
2024         (FFI_TRAMPOLINE_SIZE): Define for 32-bit HP-UX targets.
2025         (FFI_TYPE_SMALL_STRUCT2): Define.
2026         (FFI_TYPE_SMALL_STRUCT4): Likewise.
2027         (FFI_TYPE_SMALL_STRUCT8): Likewise.
2028         (FFI_TYPE_SMALL_STRUCT3): Redefine.
2029         (FFI_TYPE_SMALL_STRUCT5): Likewise.
2030         (FFI_TYPE_SMALL_STRUCT6): Likewise.
2031         (FFI_TYPE_SMALL_STRUCT7): Likewise.
2032         * src/pa/ffi.c (ROUND_DOWN): Delete.
2033         (fldw, fstw, fldd, fstd): Use '__asm__'.
2034         (ffi_struct_type): Add support for FFI_TYPE_SMALL_STRUCT2,
2035         FFI_TYPE_SMALL_STRUCT4 and FFI_TYPE_SMALL_STRUCT8.
2036         (ffi_prep_args_LINUX): Rename to ffi_prep_args_pa32. Update comment.
2037         Simplify incrementing of stack slot variable. Change type of local
2038         'n' to unsigned int.
2039         (ffi_size_stack_LINUX): Rename to ffi_size_stack_pa32. Handle long
2040         double on PA_HPUX.
2041         (ffi_prep_cif_machdep): Likewise.
2042         (ffi_call): Likewise.
2043         (ffi_closure_inner_LINUX): Rename to ffi_closure_inner_pa32. Change
2044         return type to ffi_status. Simplify incrementing of stack slot
2045         variable. Only copy floating point argument registers when PA_LINUX
2046         is true. Reformat debug statement.
2047         Add support for FFI_TYPE_SMALL_STRUCT2, FFI_TYPE_SMALL_STRUCT4 and
2048         FFI_TYPE_SMALL_STRUCT8.
2049         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Add 'extern' to
2050         declaration.
2051         (ffi_prep_closure): Make linux trampoline conditional on PA_LINUX.
2052         Add nops to cache flush.  Add trampoline for PA_HPUX.
2053         * src/pa/hpux32.S: New file.
2054         * src/pa/linux.S (ffi_call_LINUX): Rename to ffi_call_pa32. Rename
2055         ffi_prep_args_LINUX to ffi_prep_args_pa32.
2056         Localize labels. Add support for 2, 4 and 8-byte small structs. Handle
2057         unaligned destinations in 3, 5, 6 and 7-byte small structs. Order
2058         argument type checks so that common argument types appear first.
2059         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Rename
2060         ffi_closure_inner_LINUX to ffi_closure_inner_pa32.
2061
2062 2006-03-24  Alan Modra  <amodra@bigpond.net.au>
2063
2064         * src/powerpc/ffitarget.h (enum ffi_abi): Add FFI_LINUX.  Default
2065         for 32-bit using IBM extended double format.  Fix FFI_LAST_ABI.
2066         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Handle linux variant of
2067         FFI_TYPE_LONGDOUBLE.
2068         (ffi_prep_args64): Assert using IBM extended double.
2069         (ffi_prep_cif_machdep): Don't munge FFI_TYPE_LONGDOUBLE type.
2070         Handle FFI_LINUX FFI_TYPE_LONGDOUBLE return and args.
2071         (ffi_call): Handle FFI_LINUX.
2072         (ffi_closure_helper_SYSV): Non FFI_LINUX long double return needs
2073         gpr3 return pointer as for struct return.  Handle FFI_LINUX
2074         FFI_TYPE_LONGDOUBLE return and args.  Don't increment "nf"
2075         unnecessarily.
2076         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Load both f1 and f2
2077         for FFI_TYPE_LONGDOUBLE.  Move epilogue insns into case table.
2078         Don't use r6 as pointer to results, instead use sp offset.  Don't
2079         make a special call to load lr with case table address, instead
2080         use offset from previous call.
2081         * src/powerpc/sysv.S (ffi_call_SYSV): Save long double return.
2082         * src/powerpc/linux64.S (ffi_call_LINUX64): Simplify long double
2083         return.
2084
2085 2006-03-15  Kaz Kojima  <kkojima@gcc.gnu.org>
2086
2087         * src/sh64/ffi.c (ffi_prep_cif_machdep): Handle float arguments
2088         passed with FP registers correctly.
2089         (ffi_closure_helper_SYSV): Likewise.
2090         * src/sh64/sysv.S: Likewise.
2091
2092 2006-03-01  Andreas Tobler  <a.tobler@schweiz.ch>
2093
2094         * testsuite/libffi.special/unwindtest.cc (closure_test_fn): Mark cif,
2095         args and userdata unused.
2096         (closure_test_fn1): Mark cif and userdata unused.
2097         (main): Remove unused res.
2098
2099 2006-02-28  Andreas Tobler  <a.tobler@schweiz.ch>
2100
2101         * testsuite/libffi.call/call.exp: Adjust FSF address. Add test runs for
2102         -O2, -O3, -Os and the warning flags -W -Wall.
2103         * testsuite/libffi.special/special.exp: Likewise.
2104         * testsuite/libffi.call/ffitest.h: Add an __UNUSED__ macro to mark
2105         unused parameter unused for gcc or else do nothing.
2106         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2107         * testsuite/libffi.call/cls_12byte.c (cls_struct_12byte_gn): Mark cif
2108         and userdata unused.
2109         * testsuite/libffi.call/cls_16byte.c (cls_struct_16byte_gn): Likewise.
2110         * testsuite/libffi.call/cls_18byte.c (cls_struct_18byte_gn): Likewise.
2111         * testsuite/libffi.call/cls_19byte.c (cls_struct_19byte_gn): Likewise.
2112         * testsuite/libffi.call/cls_1_1byte.c (cls_struct_1_1byte_gn): Likewise.
2113         * testsuite/libffi.call/cls_20byte.c (cls_struct_20byte_gn): Likewise.
2114         * testsuite/libffi.call/cls_20byte1.c (cls_struct_20byte_gn): Likewise.
2115         * testsuite/libffi.call/cls_24byte.c (cls_struct_24byte_gn): Likewise.
2116         * testsuite/libffi.call/cls_2byte.c (cls_struct_2byte_gn): Likewise.
2117         * testsuite/libffi.call/cls_3_1byte.c (cls_struct_3_1byte_gn): Likewise.
2118         * testsuite/libffi.call/cls_3byte1.c (cls_struct_3byte_gn): Likewise.
2119         * testsuite/libffi.call/cls_3byte2.c (cls_struct_3byte_gn1): Likewise.
2120         * testsuite/libffi.call/cls_4_1byte.c (cls_struct_4_1byte_gn): Likewise.
2121         * testsuite/libffi.call/cls_4byte.c (cls_struct_4byte_gn): Likewise.
2122         * testsuite/libffi.call/cls_5_1_byte.c (cls_struct_5byte_gn): Likewise.
2123         * testsuite/libffi.call/cls_5byte.c (cls_struct_5byte_gn): Likewise.
2124         * testsuite/libffi.call/cls_64byte.c (cls_struct_64byte_gn): Likewise.
2125         * testsuite/libffi.call/cls_6_1_byte.c (cls_struct_6byte_gn): Likewise.
2126         * testsuite/libffi.call/cls_6byte.c (cls_struct_6byte_gn): Likewise.
2127         * testsuite/libffi.call/cls_7_1_byte.c (cls_struct_7byte_gn): Likewise.
2128         * testsuite/libffi.call/cls_7byte.c (cls_struct_7byte_gn): Likewise.
2129         * testsuite/libffi.call/cls_8byte.c (cls_struct_8byte_gn): Likewise.
2130         * testsuite/libffi.call/cls_9byte1.c (cls_struct_9byte_gn): Likewise.
2131         * testsuite/libffi.call/cls_9byte2.c (cls_struct_9byte_gn): Likewise.
2132         * testsuite/libffi.call/cls_align_double.c (cls_struct_align_gn):
2133         Likewise.
2134         * testsuite/libffi.call/cls_align_float.c (cls_struct_align_gn):
2135         Likewise.
2136         * testsuite/libffi.call/cls_align_longdouble.c (cls_struct_align_gn):
2137         Likewise.
2138         * testsuite/libffi.call/cls_align_pointer.c (cls_struct_align_fn): Cast
2139         void* to avoid compiler warning.
2140         (main): Likewise.
2141         (cls_struct_align_gn): Mark cif and userdata unused.
2142         * testsuite/libffi.call/cls_align_sint16.c (cls_struct_align_gn):
2143         Likewise.
2144         * testsuite/libffi.call/cls_align_sint32.c (cls_struct_align_gn):
2145         Likewise.
2146         * testsuite/libffi.call/cls_align_sint64.c (cls_struct_align_gn):
2147         Likewise.
2148         * testsuite/libffi.call/cls_align_uint16.c (cls_struct_align_gn):
2149         Likewise.
2150         * testsuite/libffi.call/cls_align_uint32.c (cls_struct_align_gn):
2151         Likewise.
2152         * testsuite/libffi.call/cls_double.c (cls_ret_double_fn): Likewise.
2153         * testsuite/libffi.call/cls_float.c (cls_ret_float_fn): Likewise.
2154         * testsuite/libffi.call/cls_multi_schar.c (test_func_gn): Mark cif and
2155         data unused.
2156         (main): Cast res_call to silence gcc.
2157         * testsuite/libffi.call/cls_multi_sshort.c (test_func_gn): Mark cif and
2158         data unused.
2159         (main): Cast res_call to silence gcc.
2160         * testsuite/libffi.call/cls_multi_sshortchar.c (test_func_gn): Mark cif
2161         and data unused.
2162         (main): Cast res_call to silence gcc.
2163         * testsuite/libffi.call/cls_multi_uchar.c (test_func_gn): Mark cif and
2164         data unused.
2165         (main): Cast res_call to silence gcc.
2166         * testsuite/libffi.call/cls_multi_ushort.c (test_func_gn): Mark cif and
2167         data unused.
2168         (main): Cast res_call to silence gcc.
2169         * testsuite/libffi.call/cls_multi_ushortchar.c (test_func_gn): Mark cif
2170         and data unused.
2171         (main): Cast res_call to silence gcc.
2172         * testsuite/libffi.call/cls_schar.c (cls_ret_schar_fn): Mark cif and
2173         userdata unused.
2174         (cls_ret_schar_fn): Cast printf parameter to silence gcc.
2175         * testsuite/libffi.call/cls_sint.c (cls_ret_sint_fn): Mark cif and
2176         userdata unused.
2177         (cls_ret_sint_fn): Cast printf parameter to silence gcc.
2178         * testsuite/libffi.call/cls_sshort.c (cls_ret_sshort_fn): Mark cif and
2179         userdata unused.
2180         (cls_ret_sshort_fn): Cast printf parameter to silence gcc.
2181         * testsuite/libffi.call/cls_uchar.c (cls_ret_uchar_fn):  Mark cif and
2182         userdata unused.
2183         (cls_ret_uchar_fn): Cast printf parameter to silence gcc.
2184         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Mark cif and
2185         userdata unused.
2186         (cls_ret_uint_fn): Cast printf parameter to silence gcc.
2187         * testsuite/libffi.call/cls_ulonglong.c (cls_ret_ulonglong_fn): Mark cif
2188         and userdata unused.
2189         * testsuite/libffi.call/cls_ushort.c (cls_ret_ushort_fn): Mark cif and
2190         userdata unused.
2191         (cls_ret_ushort_fn): Cast printf parameter to silence gcc.
2192         * testsuite/libffi.call/float.c (floating): Remove unused parameter e.
2193         * testsuite/libffi.call/float1.c (main): Remove unused variable i.
2194         Cleanup white spaces.
2195         * testsuite/libffi.call/negint.c (checking): Remove unused variable i.
2196         * testsuite/libffi.call/nested_struct.c (cls_struct_combined_gn): Mark
2197         cif and userdata unused.
2198         * testsuite/libffi.call/nested_struct1.c (cls_struct_combined_gn):
2199         Likewise.
2200         * testsuite/libffi.call/nested_struct10.c (B_gn): Likewise.
2201         * testsuite/libffi.call/nested_struct2.c (B_fn): Adjust printf
2202         formatters to silence gcc.
2203         (B_gn): Mark cif and userdata unused.
2204         * testsuite/libffi.call/nested_struct3.c (B_gn): Mark cif and userdata
2205         unused.
2206         * testsuite/libffi.call/nested_struct4.c: Mention related PR.
2207         (B_gn): Mark cif and userdata unused.
2208         * testsuite/libffi.call/nested_struct5.c (B_gn): Mark cif and userdata
2209         unused.
2210         * testsuite/libffi.call/nested_struct6.c: Mention related PR.
2211         (B_gn): Mark cif and userdata unused.
2212         * testsuite/libffi.call/nested_struct7.c (B_gn): Mark cif and userdata
2213         unused.
2214         * testsuite/libffi.call/nested_struct8.c (B_gn): Likewise.
2215         * testsuite/libffi.call/nested_struct9.c (B_gn): Likewise.
2216         * testsuite/libffi.call/problem1.c (stub): Likewise.
2217         * testsuite/libffi.call/pyobjc-tc.c (main): Cast the result to silence
2218         gcc.
2219         * testsuite/libffi.call/return_fl2.c (return_fl): Add the note mentioned
2220         in the last commit for this test case in the test case itself.
2221         * testsuite/libffi.call/closure_fn0.c (closure_test_fn0): Mark cif as
2222         unused.
2223         * testsuite/libffi.call/closure_fn1.c (closure_test_fn1): Likewise.
2224         * testsuite/libffi.call/closure_fn2.c (closure_test_fn2): Likewise.
2225         * testsuite/libffi.call/closure_fn3.c (closure_test_fn3): Likewise.
2226         * testsuite/libffi.call/closure_fn4.c (closure_test_fn0): Likewise.
2227         * testsuite/libffi.call/closure_fn5.c (closure_test_fn5): Likewise.
2228         * testsuite/libffi.call/closure_fn6.c (closure_test_fn0): Likewise.
2229
2230 2006-02-22  Kaz Kojima  <kkojima@gcc.gnu.org>
2231
2232         * src/sh/sysv.S: Fix register numbers in the FDE for
2233         ffi_closure_SYSV.
2234
2235 2006-02-20  Andreas Tobler  <a.tobler@schweiz.ch>
2236
2237         * testsuite/libffi.call/return_fl2.c (return_fl): Remove static
2238         declaration to avoid a false negative on ix86. See PR323.
2239
2240 2006-02-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2241
2242         * src/sh/ffi.c (ffi_closure_helper_SYSV): Remove unused variable
2243         and cast integer to void * if needed.  Update the pointer to
2244         the FP register saved area correctly.
2245
2246 2006-02-17  Andreas Tobler  <a.tobler@schweiz.ch>
2247
2248         * testsuite/libffi.call/nested_struct6.c: XFAIL this test until PR25630
2249         is fixed.
2250         * testsuite/libffi.call/nested_struct4.c: Likewise.
2251
2252 2006-02-16  Andreas Tobler  <a.tobler@schweiz.ch>
2253
2254         * testsuite/libffi.call/return_dbl.c: New test case.
2255         * testsuite/libffi.call/return_dbl1.c: Likewise.
2256         * testsuite/libffi.call/return_dbl2.c: Likewise.
2257         * testsuite/libffi.call/return_fl.c: Likewise.
2258         * testsuite/libffi.call/return_fl1.c: Likewise.
2259         * testsuite/libffi.call/return_fl2.c: Likewise.
2260         * testsuite/libffi.call/return_fl3.c: Likewise.
2261         * testsuite/libffi.call/closure_fn6.c: Likewise.
2262
2263         * testsuite/libffi.call/nested_struct2.c: Remove ffi_type_mylong
2264         definition.
2265         * testsuite/libffi.call/ffitest.h: Add ffi_type_mylong definition
2266         here to be used by other test cases too.
2267
2268         * testsuite/libffi.call/nested_struct10.c: New test case.
2269         * testsuite/libffi.call/nested_struct9.c: Likewise.
2270         * testsuite/libffi.call/nested_struct8.c: Likewise.
2271         * testsuite/libffi.call/nested_struct7.c: Likewise.
2272         * testsuite/libffi.call/nested_struct6.c: Likewise.
2273         * testsuite/libffi.call/nested_struct5.c: Likewise.
2274         * testsuite/libffi.call/nested_struct4.c: Likewise.
2275
2276 2006-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
2277
2278         * configure.ac: Enable libffi for sparc64-*-freebsd*.
2279         * configure: Rebuilt.
2280
2281 2006-01-18  Jakub Jelinek  <jakub@redhat.com>
2282
2283         * src/powerpc/sysv.S (smst_two_register): Don't call __ashldi3,
2284         instead do the shifting inline.
2285         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't compute %r5
2286         shift count unconditionally.  Simplify load sequences for 1, 2, 3, 4
2287         and 8 byte structs, for the remaining struct sizes don't call
2288         __lshrdi3, instead do the shifting inline.
2289
2290 2005-12-07  Thiemo Seufer  <ths@networkno.de>
2291
2292         * src/mips/ffitarget.h: Remove obsolete sgidefs.h include. Add
2293         missing parentheses.
2294         * src/mips/o32.S (ffi_call_O32): Code formatting. Define
2295         and use A3_OFF, FP_OFF, RA_OFF. Micro-optimizations.
2296         (ffi_closure_O32): Likewise, but with newly defined A3_OFF2,
2297         A2_OFF2, A1_OFF2, A0_OFF2, RA_OFF2, FP_OFF2, S0_OFF2, GP_OFF2,
2298         V1_OFF2, V0_OFF2, FA_1_1_OFF2, FA_1_0_OFF2, FA_0_1_OFF2,
2299         FA_0_0_OFF2.
2300         * src/mips/ffi.c (ffi_prep_args): Code formatting. Fix
2301         endianness bugs.
2302         (ffi_prep_closure): Improve trampoline instruction scheduling.
2303         (ffi_closure_mips_inner_O32): Fix endianness bugs.
2304
2305 2005-12-03  Alan Modra  <amodra@bigpond.net.au>
2306
2307         * src/powerpc/ffi.c: Formatting.
2308         (ffi_prep_args_SYSV): Avoid possible aliasing problems by using unions.
2309         (ffi_prep_args64): Likewise.
2310
2311 2005-09-30  Geoffrey Keating  <geoffk@apple.com>
2312
2313         * testsuite/lib/libffi-dg.exp (libffi_target_compile): For
2314         darwin, use -shared-libgcc not -lgcc_s, and explain why.
2315
2316 2005-09-26  Tom Tromey  <tromey@redhat.com>
2317
2318         * testsuite/libffi.call/float1.c (value_type): New typedef.
2319         (CANARY): New define.
2320         (main): Check for result buffer overflow.
2321         * src/powerpc/linux64.S: Handle linux64 long double returns.
2322         * src/powerpc/ffi.c (FLAG_RETURNS_128BITS): New constant.
2323         (ffi_prep_cif_machdep): Handle linux64 long double returns.
2324
2325 2005-08-25  Alan Modra  <amodra@bigpond.net.au>
2326
2327         PR target/23404
2328         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Correct placement of stack
2329         homed fp args.
2330         (ffi_status ffi_prep_cif_machdep): Correct stack sizing for same.
2331
2332 2005-08-11  Jakub Jelinek  <jakub@redhat.com>
2333
2334         * configure.ac (HAVE_HIDDEN_VISIBILITY_ATTRIBUTE): New test.
2335         (AH_BOTTOM): Add FFI_HIDDEN definition.
2336         * configure: Rebuilt.
2337         * fficonfig.h.in: Rebuilt.
2338         * src/powerpc/ffi.c (hidden): Remove.
2339         (ffi_closure_LINUX64, ffi_prep_args64, ffi_call_LINUX64,
2340         ffi_closure_helper_LINUX64): Use FFI_HIDDEN instead of hidden.
2341         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64,
2342         .ffi_closure_LINUX64): Use FFI_HIDDEN instead of .hidden.
2343         * src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV): Remove,
2344         add FFI_HIDDEN to its prototype.
2345         (ffi_closure_SYSV_inner): New.
2346         * src/x86/sysv.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2347         * src/x86/win32.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2348
2349 2005-08-10  Alfred M. Szmidt  <ams@gnu.org>
2350
2351         PR libffi/21819:
2352         * configure: Rebuilt.
2353         * configure.ac: Handle i*86-*-gnu*.
2354
2355 2005-08-09  Jakub Jelinek  <jakub@redhat.com>
2356
2357         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Use
2358         DW_CFA_offset_extended_sf rather than
2359         DW_CFA_GNU_negative_offset_extended.
2360         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise.
2361
2362 2005-07-22  SUGIOKA Toshinobu  <sugioka@itonet.co.jp>
2363
2364         * src/sh/sysv.S (ffi_call_SYSV): Stop argument popping correctly
2365         on sh3.
2366         (ffi_closure_SYSV): Change the stack layout for sh3 struct argument.
2367         * src/sh/ffi.c (ffi_prep_args): Fix sh3 argument copy, when it is
2368         partially on register.
2369         (ffi_closure_helper_SYSV): Likewise.
2370         (ffi_prep_cif_machdep): Don't set too many cif->flags.
2371
2372 2005-07-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2373
2374         * src/sh/ffi.c (ffi_call): Handle small structures correctly.
2375         Remove empty line.
2376         * src/sh64/ffi.c (simple_type): Remove.
2377         (return_type): Handle small structures correctly.
2378         (ffi_prep_args): Likewise.
2379         (ffi_call): Likewise.
2380         (ffi_closure_helper_SYSV): Likewise.
2381         * src/sh64/sysv.S (ffi_call_SYSV): Handle 1, 2 and 4-byte return.
2382         Emit position independent code if PIC and remove wrong datalabel
2383         prefixes from EH data.
2384
2385 2005-07-19  Andreas Tobler  <a.tobler@schweiz.ch>
2386
2387         * Makefile.am (nodist_libffi_la_SOURCES): Add POWERPC_FREEBSD.
2388         * Makefile.in: Regenerate.
2389         * include/Makefile.in: Likewise.
2390         * testsuite/Makefile.in: Likewise.
2391         * configure.ac: Add POWERPC_FREEBSD rules.
2392         * configure: Regenerate.
2393         * src/powerpc/ffitarget.h: Add POWERPC_FREEBSD rules.
2394         (FFI_SYSV_TYPE_SMALL_STRUCT): Define.
2395         * src/powerpc/ffi.c: Add flags to handle small structure returns
2396         in ffi_call_SYSV.
2397         (ffi_prep_cif_machdep): Handle small structures for SYSV 4 ABI.
2398         Aka FFI_SYSV.
2399         (ffi_closure_helper_SYSV): Likewise.
2400         * src/powerpc/ppc_closure.S: Add return types for small structures.
2401         * src/powerpc/sysv.S: Add bits to handle small structures for
2402         final SYSV 4 ABI.
2403
2404 2005-07-10  Andreas Tobler  <a.tobler@schweiz.ch>
2405
2406         * testsuite/libffi.call/cls_5_1_byte.c: New test file.
2407         * testsuite/libffi.call/cls_6_1_byte.c: Likewise.
2408         * testsuite/libffi.call/cls_7_1_byte.c: Likewise.
2409
2410 2005-07-05  Randolph Chung  <tausq@debian.org>
2411
2412         * src/pa/ffi.c (ffi_struct_type): Rename FFI_TYPE_SMALL_STRUCT1
2413         as FFI_TYPE_SMALL_STRUCT3.  Break out handling for 5-7 byte
2414         structures.  Kill compilation warnings.
2415         (ffi_closure_inner_LINUX): Print return values as hex in debug
2416         message.  Rename FFI_TYPE_SMALL_STRUCT1 as FFI_TYPE_SMALL_STRUCT3.
2417         Properly handle 5-7 byte structure returns.
2418         * src/pa/ffitarget.h (FFI_TYPE_SMALL_STRUCT1)
2419         (FFI_TYPE_SMALL_STRUCT2): Remove.
2420         (FFI_TYPE_SMALL_STRUCT3, FFI_TYPE_SMALL_STRUCT5)
2421         (FFI_TYPE_SMALL_STRUCT6, FFI_TYPE_SMALL_STRUCT7): Define.
2422         * src/pa/linux.S: Mark source file as using PA1.1 assembly.
2423         (checksmst1, checksmst2): Remove.
2424         (checksmst3): Optimize handling of 3-byte struct returns.
2425         (checksmst567): Properly handle 5-7 byte struct returns.
2426
2427 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2428
2429         PR libgcj/21943
2430         * src/mips/n32.S: Enforce PIC code.
2431         * src/mips/o32.S: Likewise.
2432
2433 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2434
2435         * configure.ac: Treat i*86-*-solaris2.10 and up as X86_64.
2436         * configure: Regenerate.
2437
2438 2005-06-01  Alan Modra  <amodra@bigpond.net.au>
2439
2440         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't use JUMPTARGET
2441         to call ffi_closure_helper_SYSV.  Append @local instead.
2442         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise for ffi_prep_args_SYSV.
2443
2444 2005-05-17  Kelley Cook  <kcook@gcc.gnu.org>
2445
2446         * configure.ac: Use AC_C_BIGENDIAN instead of AC_C_BIGENDIAN_CROSS.
2447         Use AC_CHECK_SIZEOF instead of AC_COMPILE_CHECK_SIZEOF.
2448         * Makefile.am (ACLOCAL_AMFLAGS): Remove -I ../config.
2449         * aclocal.m4, configure, fficonfig.h.in, Makefile.in,
2450         include/Makefile.in, testsuite/Makefile.in: Regenerate.
2451
2452 2005-05-09  Mike Stump  <mrs@apple.com>
2453
2454         * configure: Regenerate.
2455
2456 2005-05-08  Richard Henderson  <rth@redhat.com>
2457
2458         PR libffi/21285
2459         * src/alpha/osf.S: Update unwind into to match code.
2460
2461 2005-05-04  Andreas Degert <ad@papyrus-gmbh.de>
2462             Richard Henderson  <rth@redhat.com>
2463
2464         * src/x86/ffi64.c (ffi_prep_cif_machdep): Save sse-used flag in
2465         bit 11 of flags.
2466         (ffi_call): Mask return type field.  Pass ssecount to ffi_call_unix64.
2467         (ffi_prep_closure): Set carry bit if sse-used flag set.
2468         * src/x86/unix64.S (ffi_call_unix64): Add ssecount argument.
2469         Only load sse registers if ssecount non-zero.
2470         (ffi_closure_unix64): Only save sse registers if carry set on entry.
2471
2472 2005-04-29  Ralf Corsepius  <ralf.corsepius@rtems.org>
2473
2474         * configure.ac: Add i*86-*-rtems*, sparc*-*-rtems*,
2475         powerpc-*rtems*, arm*-*-rtems*, sh-*-rtems*.
2476         * configure: Regenerate.
2477
2478 2005-04-20  Hans-Peter Nilsson  <hp@axis.com>
2479
2480         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): In regsub use,
2481         have Tcl8.3-compatible intermediate variable.
2482
2483 2005-04-18  Simon Posnjak <simon.posnjak@siol.net>
2484             Hans-Peter Nilsson  <hp@axis.com>
2485
2486         * Makefile.am: Add CRIS support.
2487         * configure.ac: Likewise.
2488         * Makefile.in, configure, testsuite/Makefile.in,
2489         include/Makefile.in: Regenerate.
2490         * src/cris: New directory.
2491         * src/cris/ffi.c, src/cris/sysv.S, src/cris/ffitarget.h: New files.
2492         * src/prep_cif.c (ffi_prep_cif): Wrap in #ifndef __CRIS__.
2493
2494         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): Replace \n with
2495         \r?\n in output tests.
2496
2497 2005-04-12  Mike Stump  <mrs@apple.com>
2498
2499         * configure: Regenerate.
2500
2501 2005-03-30  Hans Boehm  <Hans.Boehm@hp.com>
2502
2503         * src/ia64/ffitarget.h (ffi_arg): Use long long instead of DI.
2504
2505 2005-03-30  Steve Ellcey  <sje@cup.hp.com>
2506
2507         * src/ia64/ffitarget.h (ffi_arg) ADD DI attribute.
2508         (ffi_sarg) Ditto.
2509         * src/ia64/unix.S (ffi_closure_unix): Extend gp
2510         to 64 bits in ILP32 mode.
2511         Load 64 bits even for short data.
2512
2513 2005-03-23  Mike Stump  <mrs@apple.com>
2514
2515         * src/powerpc/darwin.S: Update for -m64 multilib.
2516         * src/powerpc/darwin_closure.S: Likewise.
2517
2518 2005-03-21  Zack Weinberg  <zack@codesourcery.com>
2519
2520         * configure.ac: Do not invoke TL_AC_GCC_VERSION.
2521         Do not set tool_include_dir.
2522         * aclocal.m4, configure, Makefile.in, testsuite/Makefile.in:
2523         Regenerate.
2524         * include/Makefile.am: Set gcc_version and toollibffidir.
2525         * include/Makefile.in: Regenerate.
2526
2527 2005-02-22  Andrew Haley  <aph@redhat.com>
2528
2529         * src/powerpc/ffi.c (ffi_prep_cif_machdep): Bump alignment to
2530         odd-numbered register pairs for 64-bit integer types.
2531
2532 2005-02-23  Andreas Tobler  <a.tobler@schweiz.ch>
2533
2534         PR libffi/20104
2535         * testsuite/libffi.call/return_ll1.c: New test case.
2536
2537 2005-02-11  Janis Johnson  <janis187@us.ibm.com>
2538
2539         * testsuite/libffi.call/cls_align_longdouble.c: Remove dg-options.
2540         * testsuite/libffi.call/float.c: Ditto.
2541         * testsuite/libffi.call/float2.c: Ditto.
2542         * testsuite/libffi.call/float3.c: Ditto.
2543
2544 2005-02-08  Andreas Tobler  <a.tobler@schweiz.ch>
2545
2546         * src/frv/ffitarget.h: Remove PPC stuff which does not belong to frv.
2547
2548 2005-01-12  Eric Botcazou  <ebotcazou@libertysurf.fr>
2549
2550         * testsuite/libffi.special/special.exp (cxx_options): Add
2551         -shared-libgcc.
2552
2553 2004-12-31  Richard Henderson  <rth@redhat.com>
2554
2555         * src/types.c (FFI_AGGREGATE_TYPEDEF): Remove.
2556         (FFI_TYPEDEF): Rename from FFI_INTEGRAL_TYPEDEF.  Replace size and
2557         offset parameters with a type parameter; deduce size and structure
2558         alignment.  Update all users.
2559
2560 2004-12-31  Richard Henderson  <rth@redhat.com>
2561
2562         * src/types.c (FFI_TYPE_POINTER): Define with sizeof.
2563         (FFI_TYPE_LONGDOUBLE): Fix for ia64.
2564         * src/ia64/ffitarget.h (struct ffi_ia64_trampoline_struct): Move
2565         into ffi_prep_closure.
2566         * src/ia64/ia64_flags.h, src/ia64/ffi.c, src/ia64/unix.S: Rewrite
2567         from scratch.
2568
2569 2004-12-27  Richard Henderson  <rth@redhat.com>
2570
2571         * src/x86/unix64.S: Fix typo in unwind info.
2572
2573 2004-12-25  Richard Henderson  <rth@redhat.com>
2574
2575         * src/x86/ffi64.c (struct register_args): Rename from stackLayout.
2576         (enum x86_64_reg_class): Add X86_64_COMPLEX_X87_CLASS.
2577         (merge_classes): Check for it.
2578         (SSE_CLASS_P): New.
2579         (classify_argument): Pass byte_offset by value; perform all updates
2580         inside struct case.
2581         (examine_argument): Add classes argument; handle
2582         X86_64_COMPLEX_X87_CLASS.
2583         (ffi_prep_args): Merge into ...
2584         (ffi_call): ... here.  Share stack frame with ffi_call_unix64.
2585         (ffi_prep_cif_machdep): Setup cif->flags for proper structure return.
2586         (ffi_fill_return_value): Remove.
2587         (ffi_prep_closure): Remove dead assert.
2588         (ffi_closure_unix64_inner): Rename from ffi_closure_UNIX64_inner.
2589         Rewrite to use struct register_args instead of va_list.  Create
2590         flags for handling structure returns.
2591         * src/x86/unix64.S: Remove dead strings.
2592         (ffi_call_unix64): Rename from ffi_call_UNIX64.  Rewrite to share
2593         stack frame with ffi_call.  Handle structure returns properly.
2594         (float2sse, floatfloat2sse, double2sse): Remove.
2595         (sse2float, sse2double, sse2floatfloat): Remove.
2596         (ffi_closure_unix64): Rename from ffi_closure_UNIX64.  Rewrite
2597         to handle structure returns properly.
2598
2599 2004-12-08  David Edelsohn  <edelsohn@gnu.org>
2600
2601         * Makefile.am (AM_MAKEFLAGS): Remove duplicate LIBCFLAGS and
2602         PICFLAG.
2603         * Makefile.in: Regenerated.
2604
2605 2004-12-02  Richard Sandiford  <rsandifo@redhat.com>
2606
2607         * configure.ac: Use TL_AC_GCC_VERSION to set gcc_version.
2608         * configure, aclocal.m4, Makefile.in: Regenerate.
2609         * include/Makefile.in, testsuite/Makefile.in: Regenerate.
2610
2611 2004-11-29  Kelley Cook  <kcook@gcc.gnu.org>
2612
2613         * configure: Regenerate for libtool change.
2614
2615 2004-11-25  Kelley Cook  <kcook@gcc.gnu.org>
2616
2617         * configure: Regenerate for libtool reversion.
2618
2619 2004-11-24  Kelley Cook  <kcook@gcc.gnu.org>
2620
2621         * configure: Regenerate for libtool change.
2622
2623 2004-11-23  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2624
2625         * testsuite/lib/libffi-dg.exp: Use new procs in target-libpath.exp.
2626
2627 2004-11-23  Richard Sandiford  <rsandifo@redhat.com>
2628
2629         * src/mips/o32.S (ffi_call_O32, ffi_closure_O32): Use jalr instead
2630         of jal.  Use an absolute encoding for the frame information.
2631
2632 2004-11-23  Kelley Cook  <kcook@gcc.gnu.org>
2633
2634         * Makefile.am: Remove no-dependencies.  Add ACLOCAL_AMFLAGS.
2635         * acinclude.m4: Delete logic for sincludes.
2636         * aclocal.m4, Makefile.in, configure: Regenerate.
2637         * include/Makefile: Likewise.
2638         * testsuite/Makefile: Likewise.
2639
2640 2004-11-22  Eric Botcazou  <ebotcazou@libertysurf.fr>
2641
2642         * src/sparc/ffi.c (ffi_prep_closure): Align doubles and 64-bit integers
2643         on a 8-byte boundary.
2644         * src/sparc/v8.S (ffi_closure_v8): Reserve frame space for arguments.
2645
2646 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2647
2648         * src/arm/ffi.c (ffi_prep_cif_machdep): Handle functions that return
2649         long long values.  Round stack allocation to a multiple of 8 bytes
2650         for ATPCS compatibility.
2651         * src/arm/sysv.S (ffi_call_SYSV): Rework to avoid use of APCS register
2652         names.  Handle returning long long types.  Add Thumb and interworking
2653         support.  Improve soft-float code.
2654
2655 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2656
2657         * testsuite/lib/libffi-db.exp (load_gcc_lib): New function.
2658         (libffi_exit): New function.
2659         (libffi_init): Build the testglue wrapper if needed.
2660
2661 2004-10-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2662
2663         PR other/18138
2664         * testsuite/lib/libffi-dg.exp: Accept more than one multilib libgcc.
2665
2666 2004-10-25  Kazuhiro Inaoka <inaoka.kazuhiro@renesas.com>
2667
2668         * src/m32r/libffitarget.h (FFI_CLOSURES): Set to 0.
2669
2670 2004-10-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2671
2672         * src/sh/sysv.S (ffi_call_SYSV): Don't align for double data.
2673         * testsuite/libffi.call/float3.c: New test case.
2674
2675 2004-10-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2676
2677         * src/sh/ffi.c (ffi_prep_closure): Set T bit in trampoline for
2678         the function returning a structure pointed with R2.
2679         * src/sh/sysv.S (ffi_closure_SYSV): Use R2 as the pointer to
2680         the structure return value if T bit set.  Emit position
2681         independent code and EH data if PIC.
2682
2683 2004-10-13  Kazuhiro Inaoka  <inaoka.kazuhiro@renesas.com>
2684
2685         * Makefile.am: Add m32r support.
2686         * configure.ac: Likewise.
2687         * Makefile.in: Regenerate.
2688         * confiugre: Regenerate.
2689         * src/types.c: Add m32r port to FFI_INTERNAL_TYPEDEF
2690         (uint64, sint64, double, longdouble)
2691         * src/m32r: New directory.
2692         * src/m32r/ffi.c: New file.
2693         * src/m32r/sysv.S: Likewise.
2694         * src/m32r/ffitarget.h: Likewise.
2695
2696 2004-10-02  Kaz Kojima  <kkojima@gcc.gnu.org>
2697
2698         * testsuite/libffi.call/negint.c: New test case.
2699
2700 2004-09-14  H.J. Lu  <hongjiu.lu@intel.com>
2701
2702         PR libgcj/17465
2703         * testsuite/lib/libffi-dg.exp: Don't use global ld_library_path.
2704         Set up LD_LIBRARY_PATH, SHLIB_PATH, LD_LIBRARYN32_PATH,
2705         LD_LIBRARY64_PATH, LD_LIBRARY_PATH_32, LD_LIBRARY_PATH_64 and
2706         DYLD_LIBRARY_PATH.
2707
2708 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2709
2710         * testsuite/libffi.call/many_win32.c: Remove whitespaces.
2711         * testsuite/libffi.call/promotion.c: Likewise.
2712         * testsuite/libffi.call/return_ll.c: Remove unused var. Cleanup
2713         whitespaces.
2714         * testsuite/libffi.call/return_sc.c: Likewise.
2715         * testsuite/libffi.call/return_uc.c: Likewise.
2716
2717 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2718
2719         * src/powerpc/darwin.S: Fix comments and identation.
2720         * src/powerpc/darwin_closure.S: Likewise.
2721
2722 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2723
2724         * src/powerpc/ffi_darwin.c: Add flag for longdouble return values.
2725         (ffi_prep_args): Handle longdouble arguments.
2726         (ffi_prep_cif_machdep): Set flags for longdouble. Calculate space for
2727         longdouble.
2728         (ffi_closure_helper_DARWIN): Add closure handling for longdouble.
2729         * src/powerpc/darwin.S (_ffi_call_DARWIN): Add handling of longdouble
2730         values.
2731         * src/powerpc/darwin_closure.S (_ffi_closure_ASM): Likewise.
2732         * src/types.c: Defined longdouble size and alignment for darwin.
2733
2734 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2735
2736         * src/powerpc/aix.S: Remove whitespaces.
2737         * src/powerpc/aix_closure.S: Likewise.
2738         * src/powerpc/asm.h: Likewise.
2739         * src/powerpc/ffi.c: Likewise.
2740         * src/powerpc/ffitarget.h: Likewise.
2741         * src/powerpc/linux64.S: Likewise.
2742         * src/powerpc/linux64_closure.S: Likewise.
2743         * src/powerpc/ppc_closure.S: Likewise.
2744         * src/powerpc/sysv.S: Likewise.
2745
2746 2004-08-30  Anthony Green  <green@redhat.com>
2747
2748         * Makefile.am: Add frv support.
2749         * Makefile.in, testsuite/Makefile.in: Rebuilt.
2750         * configure.ac: Read configure.host.
2751         * configure.in: Read configure.host.
2752         * configure.host: New file.  frv-elf needs libgloss.
2753         * include/ffi.h.in: Force ffi_closure to have a nice big (8)
2754         alignment.  This is needed to frv and shouldn't harm the others.
2755         * include/ffi_common.h (ALIGN_DOWN): New macro.
2756         * src/frv/ffi.c, src/frv/ffitarget.h, src/frv/eabi.S: New files.
2757
2758 2004-08-24  David Daney  <daney@avtrex.com>
2759
2760         * testsuite/libffi.call/closure_fn0.c: Xfail mips64* instead of mips*.
2761         * testsuite/libffi.call/closure_fn1.c: Likewise.
2762         * testsuite/libffi.call/closure_fn2.c  Likewise.
2763         * testsuite/libffi.call/closure_fn3.c: Likewise.
2764         * testsuite/libffi.call/closure_fn4.c: Likewise.
2765         * testsuite/libffi.call/closure_fn5.c: Likewise.
2766         * testsuite/libffi.call/cls_18byte.c: Likewise.
2767         * testsuite/libffi.call/cls_19byte.c: Likewise.
2768         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2769         * testsuite/libffi.call/cls_20byte.c: Likewise.
2770         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2771         * testsuite/libffi.call/cls_24byte.c: Likewise.
2772         * testsuite/libffi.call/cls_2byte.c: Likewise.
2773         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2774         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2775         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2776         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2777         * testsuite/libffi.call/cls_4byte.c: Likewise.
2778         * testsuite/libffi.call/cls_64byte.c: Likewise.
2779         * testsuite/libffi.call/cls_6byte.c: Likewise.
2780         * testsuite/libffi.call/cls_7byte.c: Likewise.
2781         * testsuite/libffi.call/cls_8byte.c: Likewise.
2782         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2783         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2784         * testsuite/libffi.call/cls_align_double.c: Likewise.
2785         * testsuite/libffi.call/cls_align_float.c: Likewise.
2786         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2787         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2788         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2789         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2790         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2791         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2792         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2793         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
2794         * testsuite/libffi.call/cls_double.c: Likewise.
2795         * testsuite/libffi.call/cls_float.c: Likewise.
2796         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2797         * testsuite/libffi.call/cls_multi_sshort.c: Likewise.
2798         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2799         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2800         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2801         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2802         * testsuite/libffi.call/cls_schar.c: Likewise.
2803         * testsuite/libffi.call/cls_sint.c: Likewise.
2804         * testsuite/libffi.call/cls_sshort.c: Likewise.
2805         * testsuite/libffi.call/cls_uchar.c: Likewise.
2806         * testsuite/libffi.call/cls_uint.c: Likewise.
2807         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2808         * testsuite/libffi.call/cls_ushort.c: Likewise.
2809         * testsuite/libffi.call/nested_struct.c: Likewise.
2810         * testsuite/libffi.call/nested_struct1.c: Likewise.
2811         * testsuite/libffi.call/nested_struct2.c: Likewise.
2812         * testsuite/libffi.call/nested_struct3.c: Likewise.
2813         * testsuite/libffi.call/problem1.c: Likewise.
2814         * testsuite/libffi.special/unwindtest.cc: Likewise.
2815         * testsuite/libffi.call/cls_12byte.c: Likewise and set return value
2816         to zero.
2817         * testsuite/libffi.call/cls_16byte.c: Likewise.
2818         * testsuite/libffi.call/cls_5byte.c: Likewise.
2819
2820 2004-08-23  David Daney <daney@avtrex.com>
2821
2822         PR libgcj/13141
2823         * src/mips/ffitarget.h (FFI_O32_SOFT_FLOAT): New ABI.
2824         * src/mips/ffi.c (ffi_prep_args): Fix alignment calculation.
2825         (ffi_prep_cif_machdep): Handle FFI_O32_SOFT_FLOAT floating point
2826         parameters and return types.
2827         (ffi_call): Handle FFI_O32_SOFT_FLOAT ABI.
2828         (ffi_prep_closure): Ditto.
2829         (ffi_closure_mips_inner_O32): Handle FFI_O32_SOFT_FLOAT ABI, fix
2830         alignment calculations.
2831         * src/mips/o32.S (ffi_closure_O32): Don't use floating point
2832         instructions if FFI_O32_SOFT_FLOAT, make stack frame ABI compliant.
2833
2834 2004-08-14  Casey Marshall <csm@gnu.org>
2835
2836         * src/mips/ffi.c (ffi_pref_cif_machdep): set `cif->flags' to
2837         contain `FFI_TYPE_UINT64' as return type for any 64-bit
2838         integer (O32 ABI only).
2839         (ffi_prep_closure): new function.
2840         (ffi_closure_mips_inner_O32): new function.
2841         * src/mips/ffitarget.h: Define `FFI_CLOSURES' and
2842         `FFI_TRAMPOLINE_SIZE' appropriately if the ABI is o32.
2843         * src/mips/o32.S (ffi_call_O32): add labels for .eh_frame. Return
2844         64 bit integers correctly.
2845         (ffi_closure_O32): new function.
2846         Added DWARF-2 unwind info for both functions.
2847
2848 2004-08-10  Andrew Haley  <aph@redhat.com>
2849
2850         * src/x86/ffi64.c (ffi_prep_args ): 8-align all stack arguments.
2851
2852 2004-08-01  Robert Millan  <robertmh@gnu.org>
2853
2854         * configure.ac: Detect knetbsd-gnu and kfreebsd-gnu.
2855         * configure: Regenerate.
2856
2857 2004-07-30  Maciej W. Rozycki  <macro@linux-mips.org>
2858
2859         * acinclude.m4 (AC_FUNC_MMAP_BLACKLIST): Check for <sys/mman.h>
2860         and mmap() explicitly instead of relying on preset autoconf cache
2861         variables.
2862         * aclocal.m4: Regenerate.
2863         * configure: Regenerate.
2864
2865 2004-07-11  Ulrich Weigand  <uweigand@de.ibm.com>
2866
2867         * src/s390/ffi.c (ffi_prep_args): Fix C aliasing violation.
2868         (ffi_check_float_struct): Remove unused prototype.
2869
2870 2004-06-30  Geoffrey Keating  <geoffk@apple.com>
2871
2872         * src/powerpc/ffi_darwin.c (flush_icache): ';' is a comment
2873         character on Darwin, use '\n\t' instead.
2874
2875 2004-06-26  Matthias Klose  <doko@debian.org>
2876
2877         * libtool-version: Fix typo in revision/age.
2878
2879 2004-06-17  Matthias Klose  <doko@debian.org>
2880
2881         * libtool-version: New.
2882         * Makefile.am (libffi_la_LDFLAGS): Use -version-info for soname.
2883         * Makefile.in: Regenerate.
2884
2885 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2886
2887         * Makefile.am: Remove useless multilib rules.
2888         * Makefile.in: Regenerate.
2889         * aclocal.m4: Regenerate with automake 1.8.5.
2890         * configure.ac: Remove useless multilib configury.
2891         * configure: Regenerate.
2892
2893 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2894
2895         * .cvsignore: New file.
2896
2897 2004-06-10  Jakub Jelinek  <jakub@redhat.com>
2898
2899         * src/ia64/unix.S (ffi_call_unix): Insert group barrier break
2900         fp_done.
2901         (ffi_closure_UNIX): Fix f14/f15 adjustment if FLOAT_SZ is ever
2902         changed from 8.
2903
2904 2004-06-06  Sean McNeil  <sean@mcneil.com>
2905
2906         * configure.ac: Add x86_64-*-freebsd* support.
2907         * configure: Regenerate.
2908
2909 2004-04-26  Joe Buck <jbuck@welsh-buck.org>
2910
2911         Bug 15093
2912         * configure.ac: Test for existence of mmap and sys/mman.h before
2913         checking blacklist.  Fix suggested by Jim Wilson.
2914         * configure: Regenerate.
2915
2916 2004-04-26  Matt Austern  <austern@apple.com>
2917
2918         * src/powerpc/darwin.S: Go through a non-lazy pointer for initial
2919         FDE location.
2920         * src/powerpc/darwin_closure.S: Likewise.
2921
2922 2004-04-24  Andreas Tobler  <a.tobler@schweiz.ch>
2923
2924         * testsuite/libffi.call/cls_multi_schar.c (main): Fix initialization
2925         error. Reported by Thomas Heller <theller@python.net>.
2926         * testsuite/libffi.call/cls_multi_sshort.c (main): Likewise.
2927         * testsuite/libffi.call/cls_multi_ushort.c (main): Likewise.
2928
2929 2004-03-20  Matthias Klose  <doko@debian.org>
2930
2931         * src/pa/linux.S: Fix typo.
2932
2933 2004-03-19  Matthias Klose  <doko@debian.org>
2934
2935         * Makefile.am: Update.
2936         * Makefile.in: Regenerate.
2937         * src/pa/ffi.h.in: Remove.
2938         * src/pa/ffitarget.h: New file.
2939
2940 2004-02-10  Randolph Chung  <tausq@debian.org>
2941
2942         * Makefile.am: Add PA support.
2943         * Makefile.in: Regenerate.
2944         * include/Makefile.in: Regenerate.
2945         * configure.ac: Add PA target.
2946         * configure: Regenerate.
2947         * src/pa/ffi.c: New file.
2948         * src/pa/ffi.h.in: Add PA support.
2949         * src/pa/linux.S: New file.
2950         * prep_cif.c: Add PA support.
2951
2952 2004-03-16  Hosaka Yuji  <hos@tamanegi.org>
2953
2954         * src/types.c: Fix alignment size of X86_WIN32 case int64 and
2955         double.
2956         * src/x86/ffi.c (ffi_prep_args): Replace ecif->cif->rtype->type
2957         with ecif->cif->flags.
2958         (ffi_call, ffi_prep_incoming_args_SYSV): Replace cif->rtype->type
2959         with cif->flags.
2960         (ffi_prep_cif_machdep): Add X86_WIN32 struct case.
2961         (ffi_closure_SYSV): Add 1 or 2-bytes struct case for X86_WIN32.
2962         * src/x86/win32.S (retstruct1b, retstruct2b, sc_retstruct1b,
2963         sc_retstruct2b): Add for 1 or 2-bytes struct case.
2964
2965 2004-03-15 Kelley Cook <kcook@gcc.gnu.org>
2966
2967         * configure.in: Rename file to ...
2968         * configure.ac: ... this.
2969         * fficonfig.h.in: Regenerate.
2970         * Makefile.in: Regenerate.
2971         * include/Makefile.in: Regenerate.
2972         * testsuite/Makefile.in: Regenerate.
2973
2974 2004-03-12  Matt Austern  <austern@apple.com>
2975
2976         * src/powerpc/darwin.S: Fix EH information so it corresponds to
2977         changes in EH format resulting from addition of linkonce support.
2978         * src/powerpc/darwin_closure.S: Likewise.
2979
2980 2004-03-11  Andreas Tobler  <a.tobler@schweiz.ch>
2981             Paolo Bonzini  <bonzini@gnu.org>
2982
2983         * Makefile.am (AUTOMAKE_OPTIONS): Set them.
2984         Remove VPATH. Remove rules for object files. Remove multilib support.
2985         (AM_CCASFLAGS): Add.
2986         * configure.in (AC_CONFIG_HEADERS): Relace AM_CONFIG_HEADER.
2987         (AC_PREREQ): Bump version to 2.59.
2988         (AC_INIT): Fill with version info and bug address.
2989         (ORIGINAL_LD_FOR_MULTILIBS): Remove.
2990         (AM_ENABLE_MULTILIB): Use this instead of AC_ARG_ENABLE.
2991         De-precious CC so that the right flags are passed down to multilibs.
2992         (AC_MSG_ERROR): Replace obsolete macro AC_ERROR.
2993         (AC_CONFIG_FILES): Replace obsolete macro AC_LINK_FILES.
2994         (AC_OUTPUT): Reorganize the output with AC_CONFIG_COMMANDS.
2995         * configure: Rebuilt.
2996         * aclocal.m4: Likewise.
2997         * Makefile.in, include/Makefile.in, testsuite/Makefile.in: Likewise.
2998         * fficonfig.h.in: Likewise.
2999
3000 2004-03-11  Andreas Schwab  <schwab@suse.de>
3001
3002         * src/ia64/ffi.c (ffi_prep_incoming_args_UNIX): Get floating point
3003         arguments from fp registers only for the first 8 parameter slots.
3004         Don't convert a float parameter when passed in memory.
3005
3006 2004-03-09  Hans-Peter Nilsson  <hp@axis.com>
3007
3008         * configure: Regenerate for config/accross.m4 correction.
3009
3010 2004-02-25  Matt Kraai  <kraai@alumni.cmu.edu>
3011
3012         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Change
3013         ecif->cif->bytes to bytes.
3014         (ffi_prep_cif_machdep): Add braces around nested if statement.
3015
3016 2004-02-09  Alan Modra  <amodra@bigpond.net.au>
3017
3018         * src/types.c (pointer): POWERPC64 has 8 byte pointers.
3019
3020         * src/powerpc/ffi.c (ffi_prep_args64): Correct long double handling.
3021         (ffi_closure_helper_LINUX64): Fix typo.
3022         * testsuite/libffi.call/cls_align_longdouble.c: Pass -mlong-double-128
3023         for powerpc64-*-*.
3024         * testsuite/libffi.call/float.c: Likewise.
3025         * testsuite/libffi.call/float2.c: Likewise.
3026
3027 2004-02-08  Alan Modra  <amodra@bigpond.net.au>
3028
3029         * src/powerpc/ffi.c (ffi_prep_cif_machdep <FFI_LINUX64>): Correct
3030         long double function return and long double arg handling.
3031         (ffi_closure_helper_LINUX64): Formatting.  Delete unused "ng" var.
3032         Use "end_pfr" instead of "nf".  Correct long double handling.
3033         Localise "temp".
3034         * src/powerpc/linux64.S (ffi_call_LINUX64): Save f2 long double
3035         return value.
3036         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Allocate
3037         space for long double return value.  Adjust stack frame and offsets.
3038         Load f2 long double return.
3039
3040 2004-02-07  Alan Modra  <amodra@bigpond.net.au>
3041
3042         * src/types.c: Use 16 byte long double for POWERPC64.
3043
3044 2004-01-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
3045
3046         * src/sparc/ffi.c (ffi_prep_args_v9): Shift the parameter array
3047         when the structure return address is passed in %o0.
3048         (ffi_V9_return_struct): Rename into ffi_v9_layout_struct.
3049         (ffi_v9_layout_struct): Align the field following a nested structure
3050         on a word boundary.  Use memmove instead of memcpy.
3051         (ffi_call): Update call to ffi_V9_return_struct.
3052         (ffi_prep_closure): Define 'ctx' only for V8.
3053         (ffi_closure_sparc_inner): Clone into ffi_closure_sparc_inner_v8
3054         and ffi_closure_sparc_inner_v9.
3055         (ffi_closure_sparc_inner_v8): Return long doubles by reference.
3056         Always skip the structure return address.  For structures and long
3057         doubles, copy the argument directly.
3058         (ffi_closure_sparc_inner_v9): Skip the structure return address only
3059         if required.  Shift the maximum floating-point slot accordingly.  For
3060         big structures, copy the argument directly; otherwise, left-justify the
3061         argument and call ffi_v9_layout_struct to lay out the structure on
3062         the stack.
3063         * src/sparc/v8.S: Undef STACKFRAME before defining it.
3064         (ffi_closure_v8): Pass the structure return address.  Update call to
3065         ffi_closure_sparc_inner_v8.  Short-circuit FFI_TYPE_INT handling.
3066         Skip the 'unimp' insn when returning long doubles and structures.
3067         * src/sparc/v9.S: Undef STACKFRAME before defining it.
3068         (ffi_closure_v9): Increase the frame size by 2 words.  Short-circuit
3069         FFI_TYPE_INT handling.  Load structures both in integers and
3070         floating-point registers on return.
3071         * README: Update status of the SPARC port.
3072
3073 2004-01-24  Andreas Tobler  <a.tobler@schweiz.ch>
3074
3075         * testsuite/libffi.call/pyobjc-tc.c (main): Treat result value
3076         as of type ffi_arg.
3077         * testsuite/libffi.call/struct3.c (main): Fix CHECK.
3078
3079 2004-01-22  Ulrich Weigand  <uweigand@de.ibm.com>
3080
3081         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Treat result
3082         value as of type ffi_arg, not unsigned int.
3083
3084 2004-01-21  Michael Ritzert  <ritzert@t-online.de>
3085
3086         * ffi64.c (ffi_prep_args): Cast the RHS of an assignment instead
3087         of the LHS.
3088
3089 2004-01-12  Andreas Tobler  <a.tobler@schweiz.ch>
3090
3091         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_32 for
3092         Solaris.
3093
3094 2004-01-08  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3095
3096         * testsuite/libffi.call/ffitest.h (allocate_mmap): Cast MAP_FAILED
3097         to void *.
3098
3099 2003-12-10  Richard Henderson  <rth@redhat.com>
3100
3101         * testsuite/libffi.call/cls_align_pointer.c: Cast pointers to
3102         size_t instead of int.
3103
3104 2003-12-04  Hosaka Yuji  <hos@tamanegi.org>
3105
3106         * testsuite/libffi.call/many_win32.c: Include <float.h>.
3107         * testsuite/libffi.call/many_win32.c (main): Replace variable
3108         int i with unsigned long ul.
3109
3110         * testsuite/libffi.call/cls_align_uint64.c: New test case.
3111         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
3112         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
3113         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
3114         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
3115         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
3116         * testsuite/libffi.call/cls_align_float.c: Likewise.
3117         * testsuite/libffi.call/cls_align_double.c: Likewise.
3118         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
3119         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
3120
3121 2003-12-02  Hosaka Yuji  <hos@tamanegi.org>
3122
3123         PR other/13221
3124         * src/x86/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV):
3125         Align arguments to 32 bits.
3126
3127 2003-12-01  Andreas Tobler  <a.tobler@schweiz.ch>
3128
3129         PR other/13221
3130         * testsuite/libffi.call/cls_multi_sshort.c: New test case.
3131         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
3132         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
3133         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
3134         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
3135         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
3136
3137         * testsuite/libffi.special/unwindtest.cc: Cosmetics.
3138
3139 2003-11-26  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
3140
3141         * testsuite/libffi.call/ffitest.h: Include <fcntl.h>.
3142         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3143
3144 2003-11-22  Andreas Tobler  <a.tobler@schweiz.ch>
3145
3146         * Makefile.in: Rebuilt.
3147         * configure: Likewise.
3148         * testsuite/libffi.special/unwindtest.cc: Convert the mmap to
3149         the right type.
3150
3151 2003-11-21  Andreas Jaeger  <aj@suse.de>
3152             Andreas Tobler  <a.tobler@schweiz.ch>
3153
3154         * acinclude.m4: Add AC_FUNC_MMAP_BLACKLIST.
3155         * configure.in: Call AC_FUNC_MMAP_BLACKLIST.
3156         * Makefile.in: Rebuilt.
3157         * aclocal.m4: Likewise.
3158         * configure: Likewise.
3159         * fficonfig.h.in: Likewise.
3160         * testsuite/lib/libffi-dg.exp: Add include dir.
3161         * testsuite/libffi.call/ffitest.h: Add MMAP definitions.
3162         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3163         * testsuite/libffi.call/closure_fn0.c: Use MMAP functionality
3164         for ffi_closure if available.
3165         * testsuite/libffi.call/closure_fn1.c: Likewise.
3166         * testsuite/libffi.call/closure_fn2.c: Likewise.
3167         * testsuite/libffi.call/closure_fn3.c: Likewise.
3168         * testsuite/libffi.call/closure_fn4.c: Likewise.
3169         * testsuite/libffi.call/closure_fn5.c: Likewise.
3170         * testsuite/libffi.call/cls_12byte.c: Likewise.
3171         * testsuite/libffi.call/cls_16byte.c: Likewise.
3172         * testsuite/libffi.call/cls_18byte.c: Likewise.
3173         * testsuite/libffi.call/cls_19byte.c: Likewise.
3174         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3175         * testsuite/libffi.call/cls_20byte.c: Likewise.
3176         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3177         * testsuite/libffi.call/cls_24byte.c: Likewise.
3178         * testsuite/libffi.call/cls_2byte.c: Likewise.
3179         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3180         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3181         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3182         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3183         * testsuite/libffi.call/cls_4byte.c: Likewise.
3184         * testsuite/libffi.call/cls_5byte.c: Likewise.
3185         * testsuite/libffi.call/cls_64byte.c: Likewise.
3186         * testsuite/libffi.call/cls_6byte.c: Likewise.
3187         * testsuite/libffi.call/cls_7byte.c: Likewise.
3188         * testsuite/libffi.call/cls_8byte.c: Likewise.
3189         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3190         * testsuite/libffi.call/cls_9byte2.c: Likewise.
3191         * testsuite/libffi.call/cls_double.c: Likewise.
3192         * testsuite/libffi.call/cls_float.c: Likewise.
3193         * testsuite/libffi.call/cls_schar.c: Likewise.
3194         * testsuite/libffi.call/cls_sint.c: Likewise.
3195         * testsuite/libffi.call/cls_sshort.c: Likewise.
3196         * testsuite/libffi.call/cls_uchar.c: Likewise.
3197         * testsuite/libffi.call/cls_uint.c: Likewise.
3198         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3199         * testsuite/libffi.call/cls_ushort.c: Likewise.
3200         * testsuite/libffi.call/nested_struct.c: Likewise.
3201         * testsuite/libffi.call/nested_struct1.c: Likewise.
3202         * testsuite/libffi.call/nested_struct2.c: Likewise.
3203         * testsuite/libffi.call/nested_struct3.c: Likewise.
3204         * testsuite/libffi.call/problem1.c: Likewise.
3205         * testsuite/libffi.special/unwindtest.cc: Likewise.
3206
3207 2003-11-20  Andreas Tobler  <a.tobler@schweiz.ch>
3208
3209         * testsuite/lib/libffi-dg.exp: Make the -lgcc_s conditional.
3210
3211 2003-11-19  Andreas Tobler  <a.tobler@schweiz.ch>
3212
3213         * testsuite/lib/libffi-dg.exp: Add DYLD_LIBRARY_PATH for darwin.
3214         Add -lgcc_s to additional flags.
3215
3216 2003-11-12  Andreas Tobler  <a.tobler@schweiz.ch>
3217
3218         * configure.in, include/Makefile.am: PR libgcj/11147, install
3219         the ffitarget.h header file in a gcc versioned and target
3220         dependent place.
3221         * configure: Regenerated.
3222         * Makefile.in, include/Makefile.in: Likewise.
3223         * testsuite/Makefile.in: Likewise.
3224
3225 2003-11-09  Andreas Tobler  <a.tobler@schweiz.ch>
3226
3227         * testsuite/libffi.call/closure_fn0.c: Print result and check
3228         with dg-output to make debugging easier.
3229         * testsuite/libffi.call/closure_fn1.c: Likewise.
3230         * testsuite/libffi.call/closure_fn2.c: Likewise.
3231         * testsuite/libffi.call/closure_fn3.c: Likewise.
3232         * testsuite/libffi.call/closure_fn4.c: Likewise.
3233         * testsuite/libffi.call/closure_fn5.c: Likewise.
3234         * testsuite/libffi.call/cls_12byte.c: Likewise.
3235         * testsuite/libffi.call/cls_16byte.c: Likewise.
3236         * testsuite/libffi.call/cls_18byte.c: Likewise.
3237         * testsuite/libffi.call/cls_19byte.c: Likewise.
3238         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3239         * testsuite/libffi.call/cls_20byte.c: Likewise.
3240         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3241         * testsuite/libffi.call/cls_24byte.c: Likewise.
3242         * testsuite/libffi.call/cls_2byte.c: Likewise.
3243         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3244         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3245         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3246         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3247         * testsuite/libffi.call/cls_4byte.c: Likewise.
3248         * testsuite/libffi.call/cls_5byte.c: Likewise.
3249         * testsuite/libffi.call/cls_64byte.c: Likewise.
3250         * testsuite/libffi.call/cls_6byte.c: Likewise.
3251         * testsuite/libffi.call/cls_7byte.c: Likewise.
3252         * testsuite/libffi.call/cls_8byte.c: Likewise.
3253         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3254         * testsuite/libffi.call/cls_9byte2.c: Likewise.
3255         * testsuite/libffi.call/cls_double.c: Likewise.
3256         * testsuite/libffi.call/cls_float.c: Likewise.
3257         * testsuite/libffi.call/cls_schar.c: Likewise.
3258         * testsuite/libffi.call/cls_sint.c: Likewise.
3259         * testsuite/libffi.call/cls_sshort.c: Likewise.
3260         * testsuite/libffi.call/cls_uchar.c: Likewise.
3261         * testsuite/libffi.call/cls_uint.c: Likewise.
3262         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3263         * testsuite/libffi.call/cls_ushort.c: Likewise.
3264         * testsuite/libffi.call/problem1.c: Likewise.
3265
3266         * testsuite/libffi.special/unwindtest.cc: Make ffi_closure
3267         static.
3268
3269 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
3270
3271         * testsuite/libffi.call/cls_9byte2.c: New test case.
3272         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3273         * testsuite/libffi.call/cls_64byte.c: Likewise.
3274         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3275         * testsuite/libffi.call/cls_19byte.c: Likewise.
3276         * testsuite/libffi.call/cls_18byte.c: Likewise.
3277         * testsuite/libffi.call/closure_fn4.c: Likewise.
3278         * testsuite/libffi.call/closure_fn5.c: Likewise.
3279         * testsuite/libffi.call/cls_schar.c: Likewise.
3280         * testsuite/libffi.call/cls_sint.c: Likewise.
3281         * testsuite/libffi.call/cls_sshort.c: Likewise.
3282         * testsuite/libffi.call/nested_struct2.c: Likewise.
3283         * testsuite/libffi.call/nested_struct3.c: Likewise.
3284
3285 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
3286
3287         * testsuite/libffi.call/cls_double.c: Do a check on the result.
3288         * testsuite/libffi.call/cls_uchar.c: Likewise.
3289         * testsuite/libffi.call/cls_uint.c: Likewise.
3290         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3291         * testsuite/libffi.call/cls_ushort.c: Likewise.
3292         * testsuite/libffi.call/return_sc.c: Cleanup whitespaces.
3293
3294 2003-11-06  Andreas Tobler  <a.tobler@schweiz.ch>
3295
3296         * src/prep_cif.c (ffi_prep_cif): Move the validity check after
3297         the initialization.
3298
3299 2003-10-23  Andreas Tobler  <a.tobler@schweiz.ch>
3300
3301         * src/java_raw_api.c (ffi_java_ptrarray_to_raw): Replace
3302         FFI_ASSERT(FALSE) with FFI_ASSERT(0).
3303
3304 2003-10-22  David Daney  <ddaney@avtrex.com>
3305
3306         * src/mips/ffitarget.h: Replace undefined UINT32 and friends with
3307         __attribute__((__mode__(__SI__))) and friends.
3308
3309 2003-10-22  Andreas Schwab  <schwab@suse.de>
3310
3311         * src/ia64/ffi.c: Replace FALSE/TRUE with false/true.
3312
3313 2003-10-21  Andreas Tobler  <a.tobler@schweiz.ch>
3314
3315         * configure.in: AC_LINK_FILES(ffitarget.h).
3316         * configure: Regenerate.
3317         * Makefile.in: Likewise.
3318         * include/Makefile.in: Likewise.
3319         * testsuite/Makefile.in: Likewise.
3320         * fficonfig.h.in: Likewise.
3321
3322 2003-10-21  Paolo Bonzini  <bonzini@gnu.org>
3323             Richard Henderson  <rth@redhat.com>
3324
3325         Avoid that ffi.h includes fficonfig.h.
3326
3327         * Makefile.am (EXTRA_DIST): Include ffitarget.h files
3328         (TARGET_SRC_MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3329         (TARGET_SRC_MIPS_SGI): Removed.
3330         (MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3331         (MIPS_SGI): Removed.
3332         (CLEANFILES): Removed.
3333         (mostlyclean-am, clean-am, mostlyclean-sub, clean-sub): New
3334         targets.
3335         * acconfig.h: Removed.
3336         * configure.in: Compute sizeofs only for double and long double.
3337         Use them to define and subst HAVE_LONG_DOUBLE.  Include comments
3338         into AC_DEFINE instead of using acconfig.h.  Create
3339         include/ffitarget.h instead of include/fficonfig.h.  Rename
3340         MIPS_GCC to MIPS_IRIX, drop MIPS_SGI since we are in gcc's tree.
3341         AC_DEFINE EH_FRAME_FLAGS.
3342         * include/Makefile.am (DISTCLEANFILES): New automake macro.
3343         (hack_DATA): Add ffitarget.h.
3344         * include/ffi.h.in: Remove all system specific definitions.
3345         Declare raw API even if it is not installed, why bother?
3346         Use limits.h instead of SIZEOF_* to define ffi_type_*.  Do
3347         not define EH_FRAME_FLAGS, it is in fficonfig.h now.  Include
3348         ffitarget.h instead of fficonfig.h.  Remove ALIGN macro.
3349         (UINT_ARG, INT_ARG): Removed, use ffi_arg and ffi_sarg instead.
3350         * include/ffi_common.h (bool): Do not define.
3351         (ffi_assert): Accept failed assertion.
3352         (ffi_type_test): Return void and accept file/line.
3353         (FFI_ASSERT): Pass stringized failed assertion.
3354         (FFI_ASSERT_AT): New macro.
3355         (FFI_ASSERT_VALID_TYPE): New macro.
3356         (UINT8, SINT8, UINT16, SINT16, UINT32, SINT32,
3357         UINT64, SINT64): Define here with gcc's __attribute__ macro
3358         instead of in ffi.h
3359         (FLOAT32, ALIGN): Define here instead of in ffi.h
3360         * include/ffi-mips.h: Removed.  Its content moved to
3361         src/mips/ffitarget.h after separating assembly and C sections.
3362         * src/alpha/ffi.c, src/alpha/ffi.c, src/java_raw_api.c
3363         src/prep_cif.c, src/raw_api.c, src/ia64/ffi.c,
3364         src/mips/ffi.c, src/mips/n32.S, src/mips/o32.S,
3365         src/mips/ffitarget.h, src/sparc/ffi.c, src/x86/ffi64.c:
3366         SIZEOF_ARG -> FFI_SIZEOF_ARG.
3367         * src/ia64/ffi.c: Include stdbool.h (provided by GCC 2.95+).
3368         * src/debug.c (ffi_assert): Accept stringized failed assertion.
3369         (ffi_type_test): Rewritten.
3370         * src/prep-cif.c (initialize_aggregate, ffi_prep_cif): Call
3371         FFI_ASSERT_VALID_TYPE.
3372         * src/alpha/ffitarget.h, src/arm/ffitarget.h,
3373         src/ia64/ffitarget.h, src/m68k/ffitarget.h,
3374         src/mips/ffitarget.h, src/powerpc/ffitarget.h,
3375         src/s390/ffitarget.h, src/sh/ffitarget.h,
3376         src/sh64/ffitarget.h, src/sparc/ffitarget.h,
3377         src/x86/ffitarget.h: New files.
3378         * src/alpha/osf.S, src/arm/sysv.S, src/ia64/unix.S,
3379         src/m68k/sysv.S, src/mips/n32.S, src/mips/o32.S,
3380         src/powerpc/aix.S, src/powerpc/darwin.S,
3381         src/powerpc/ffi_darwin.c, src/powerpc/linux64.S,
3382         src/powerpc/linux64_closure.S, src/powerpc/ppc_closure.S,
3383         src/powerpc/sysv.S, src/s390/sysv.S, src/sh/sysv.S,
3384         src/sh64/sysv.S, src/sparc/v8.S, src/sparc/v9.S,
3385         src/x86/sysv.S, src/x86/unix64.S, src/x86/win32.S:
3386         include fficonfig.h
3387
3388 2003-10-20  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3389
3390         * src/mips/ffi.c: Use _ABIN32, _ABIO32 instead of external
3391         _MIPS_SIM_NABI32, _MIPS_SIM_ABI32.
3392
3393 2003-10-19  Andreas Tobler  <a.tobler@schweiz.ch>
3394
3395         * src/powerpc/ffi_darwin.c (ffi_prep_args): Declare bytes again.
3396         Used when FFI_DEBUG = 1.
3397
3398 2003-10-14  Alan Modra  <amodra@bigpond.net.au>
3399
3400         * src/types.c (double, longdouble): Default POWERPC64 to 8 byte size
3401         and align.
3402
3403 2003-10-06  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3404
3405         * include/ffi_mips.h: Define FFI_MIPS_N32 for N32/N64 ABIs,
3406         FFI_MIPS_O32 for O32 ABI.
3407
3408 2003-10-01  Andreas Tobler  <a.tobler@schweiz.ch>
3409
3410         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_64 for
3411         SPARC64. Cleanup whitespaces.
3412
3413 2003-09-19  Andreas Tobler  <a.tobler@schweiz.ch>
3414
3415         * testsuite/libffi.call/closure_fn0.c: Xfail mips, arm,
3416         strongarm, xscale. Cleanup whitespaces.
3417         * testsuite/libffi.call/closure_fn1.c: Likewise.
3418         * testsuite/libffi.call/closure_fn2.c: Likewise.
3419         * testsuite/libffi.call/closure_fn3.c: Likewise.
3420         * testsuite/libffi.call/cls_12byte.c: Likewise.
3421         * testsuite/libffi.call/cls_16byte.c: Likewise.
3422         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3423         * testsuite/libffi.call/cls_20byte.c: Likewise.
3424         * testsuite/libffi.call/cls_24byte.c: Likewise.
3425         * testsuite/libffi.call/cls_2byte.c: Likewise.
3426         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3427         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3428         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3429         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3430         * testsuite/libffi.call/cls_4byte.c: Likewise.
3431         * testsuite/libffi.call/cls_5byte.c: Likewise.
3432         * testsuite/libffi.call/cls_6byte.c: Likewise.
3433         * testsuite/libffi.call/cls_7byte.c: Likewise.
3434         * testsuite/libffi.call/cls_8byte.c: Likewise.
3435         * testsuite/libffi.call/cls_double.c: Likewise.
3436         * testsuite/libffi.call/cls_float.c: Likewise.
3437         * testsuite/libffi.call/cls_uchar.c: Likewise.
3438         * testsuite/libffi.call/cls_uint.c: Likewise.
3439         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3440         * testsuite/libffi.call/cls_ushort.c: Likewise.
3441         * testsuite/libffi.call/nested_struct.c: Likewise.
3442         * testsuite/libffi.call/nested_struct1.c: Likewise.
3443         * testsuite/libffi.call/problem1.c: Likewise.
3444         * testsuite/libffi.special/unwindtest.cc: Likewise.
3445         * testsuite/libffi.call/pyobjc-tc.c: Cleanup whitespaces.
3446
3447 2003-09-18  David Edelsohn  <edelsohn@gnu.org>