libffi/ChangeLog

   1 2010-04-07  Jakub Jelinek  <jakub@redhat.com>
   2
   3         * regex.c (byte_re_match_2_internal): Avoid set but not used
   4         warning.
   5
   6 2010-04-02  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
   7
   8         * Makefile.in: Regenerate.
   9         * aclocal.m4: Regenerate.
  10         * include/Makefile.in: Regenerate.
  11         * man/Makefile.in: Regenerate.
  12         * testsuite/Makefile.in: Regenerate.
  13
  14 2010-03-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
  15
  16         * configure.ac (libffi_cv_as_x86_64_unwind_section_type): New test.
  17         * configure: Regenerate.
  18         * fficonfig.h.in: Regenerate.
  19         * libffi/src/x86/unix64.S (.eh_frame)
  20         [HAVE_AS_X86_64_UNWIND_SECTION_TYPE]: Use @unwind section type.
  21
  22 2010-02-24  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
  23
  24         * doc/libffi.texi (The Closure API): Fix typo.
  25         * doc/libffi.info: Remove.
  26
  27 2010-02-15  Matthias Klose  <doko@ubuntu.com>
  28
  29         * src/arm/sysv.S (__ARM_ARCH__): Define for processor
  30         __ARM_ARCH_7EM__.
  31
  32 2010-01-07  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
  33
  34         PR libffi/40701
  35         * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRIdLL,
  36         PRIuLL, PRId64, PRIu64, PRIuPTR): Define.
  37         * testsuite/libffi.call/cls_align_sint64.c: Add -Wno-format on
  38         alpha*-dec-osf*.
  39         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
  40         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
  41         * testsuite/libffi.call/return_ll1.c: Likewise.
  42         * testsuite/libffi.call/stret_medium2.c: Likewise.
  43         * testsuite/libffi.special/ffitestcxx.h (allocate_mmap): Cast
  44         MAP_FAILED to char *.
  45
  46 2010-01-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
  47
  48         * src/mips/n32.S: Use .abicalls and .eh_frame with __GNUC__.
  49
  50 2009-12-31  Anthony Green  <green@redhat.com>
  51
  52         * README: Update for libffi 3.0.9.
  53
  54 2009-12-27  Matthias Klose  <doko@ubuntu.com>
  55
  56         * configure.ac (HAVE_LONG_DOUBLE): Define for mips when
  57         appropriate.
  58         * configure: Rebuilt.
  59
  60 2009-12-26  Anthony Green  <green@redhat.com>
  61
  62         * testsuite/libffi.call/cls_longdouble_va.c: Mark as xfail for
  63         avr32*-*-*.
  64         * testsuite/libffi.call/cls_double_va.c: Ditto.
  65
  66 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
  67
  68         * testsuite/libffi.call/ffitest.h: Conditionally include stdint.h
  69         and inttypes.h.
  70         * testsuite/libffi.special/unwindtest.cc: Ditto.
  71
  72 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
  73
  74         * configure.ac: Add amd64-*-openbsd*.
  75         * configure: Rebuilt.
  76         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Link
  77         openbsd programs with -lpthread.
  78
  79 2009-12-26  Anthony Green  <green@redhat.com>
  80
  81         * testsuite/libffi.call/cls_double_va.c,
  82         testsuite/libffi.call/cls_longdouble.c,
  83         testsuite/libffi.call/cls_longdouble_va.c,
  84         testsuite/libffi.call/cls_pointer.c,
  85         testsuite/libffi.call/cls_pointer_stack.c: Remove xfail for
  86         mips*-*-* and arm*-*-*.
  87         * testsuite/libffi.call/cls_align_longdouble_split.c,
  88         testsuite/libffi.call/cls_align_longdouble_split2.c,
  89         testsuite/libffi.call/stret_medium2.c,
  90         testsuite/libffi.call/stret_medium.c,
  91         testsuite/libffi.call/stret_large.c,
  92         testsuite/libffi.call/stret_large2.c: Remove xfail for arm*-*-*.
  93
  94 2009-12-31  Kay Tietz  <ktietz70@googlemail.com>
  95
  96         * testsuite/libffi.call/ffitest.h,
  97         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRuLL): Fix
  98         definitions.
  99
 100 2009-12-31  Carlo Bramini  <carlo.bramix@libero.it>
 101
 102         * configure.ac (AM_LTLDFLAGS): Define for windows hosts.
 103         * Makefile.am (libffi_la_LDFLAGS): Add AM_LTLDFLAGS.
 104         * configure: Rebuilt.
 105         * Makefile.in: Rebuilt.
 106
 107 2009-12-31  Anthony Green  <green@redhat.com>
 108             Blake Chaffin.
 109
 110         * testsuite/libffi.call/huge_struct.c: New test case from Blake
 111         Chaffin @ Apple.
 112
 113 2009-12-28  David Edelsohn  <edelsohn@gnu.org>
 114
 115         * src/powerpc/ffi_darwin.c (ffi_prep_args): Copy abi and nargs to
 116         local variables.
 117         (aix_adjust_aggregate_sizes): New function.
 118         (ffi_prep_cif_machdep): Call it.
 119
 120 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
 121
 122         * configure.ac: Define FFI_MMAP_EXEC_WRIT for the given targets.
 123         * configure: Regenerate.
 124         * fficonfig.h.in: Likewise.
 125         * src/closures.c: Remove the FFI_MMAP_EXEC_WRIT definition for
 126         Solaris/x86.
 127
 128 2009-12-26  Andreas Schwab  <schwab@linux-m68k.org>
 129
 130         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Advance intarg_count
 131         when a float arguments is passed in memory.
 132         (ffi_closure_helper_SYSV): Mark general registers as used up when
 133         a 64bit or soft-float long double argument is passed in memory.
 134
 135 2009-12-25  Matthias Klose  <doko@ubuntu.com>
 136
 137         * man/ffi_call.3: Fix #include in examples.
 138         * doc/libffi.texi: Add dircategory.
 139
 140 2009-12-25  Frank Everdij <f.p.x.everdij@tudelft.nl>
 141
 142         * include/ffi.h.in: Placed '__GNUC__' ifdef around
 143         '__attribute__((aligned(8)))' in ffi_closure, fixes compile for
 144         IRIX MIPSPro c99.
 145         * include/ffi_common.h: Added '__sgi' define to non
 146         '__attribute__((__mode__()))' integer typedefs.
 147         * src/mips/ffi.c (ffi_call, ffi_closure_mips_inner_O32,
 148         ffi_closure_mips_inner_N32): Added 'defined(_MIPSEB)' to BE check.
 149         (ffi_closure_mips_inner_O32, ffi_closure_mips_inner_N32): Added
 150         FFI_LONGDOUBLE support and alignment(N32 only).
 151         * src/mips/ffitarget.h: Corrected '#include <sgidefs.h>' for IRIX and
 152         fixed non '__attribute__((__mode__()))' integer typedefs.
 153         * src/mips/n32.S: Put '#ifdef linux' around '.abicalls' and '.eh_frame'
 154         since they are Linux/GNU Assembler specific.
 155
 156 2009-12-25  Bradley Smith  <brad@brad-smith.co.uk>
 157
 158         * configure.ac, Makefile.am, src/avr32/ffi.c,
 159         src/avr32/ffitarget.h,
 160         src/avr32/sysv.S: Add AVR32 port.
 161         * configure, Makefile.in: Rebuilt.
 162
 163 2009-12-21  Andreas Tobler  <a.tobler@schweiz.org>
 164
 165         * configure.ac: Make i?86 build on FreeBSD and OpenBSD.
 166         * configure: Regenerate.
 167
 168 2009-12-15  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
 169
 170         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on PA HP-UX.
 171
 172 2009-12-13  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
 173
 174         * src/pa/ffi.c (ffi_closure_inner_pa32): Handle FFI_TYPE_LONGDOUBLE
 175         type on HP-UX.
 176
 177 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
 178
 179         * src/sparc/ffi.c (ffi_closure_sparc_inner_v9): Properly align 'long
 180         double' arguments.
 181
 182 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
 183
 184         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on Solaris < 10.
 185
 186 2009-12-10  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 187
 188         PR libffi/40700
 189         * src/closures.c [X86_64 && __sun__ && __svr4__]
 190         (FFI_MMAP_EXEC_WRIT): Define.
 191
 192 2009-12-08  David Daney  <ddaney@caviumnetworks.com>
 193
 194         * testsuite/libffi.call/stret_medium.c: Remove xfail for mips*-*-*
 195         * testsuite/libffi.call/cls_align_longdouble_split2.c: Same.
 196         * testsuite/libffi.call/stret_large.c: Same.
 197         * testsuite/libffi.call/cls_align_longdouble_split.c: Same.
 198         * testsuite/libffi.call/stret_large2.c: Same.
 199         * testsuite/libffi.call/stret_medium2.c: Same.
 200
 201 2009-12-07  David Edelsohn  <edelsohn@gnu.org>
 202
 203         * src/powerpc/aix_closure.S (libffi_closure_ASM): Fix tablejump
 204         typo.
 205
 206 2009-12-05  David Edelsohn  <edelsohn@gnu.org>
 207
 208         * src/powerpc/aix.S: Update AIX32 code to be consistent with AIX64
 209         code.
 210         * src/powerpc/aix_closure.S: Same.
 211
 212 2009-12-05  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 213
 214         * Makefile.in: Regenerate.
 215         * configure: Regenerate.
 216         * include/Makefile.in: Regenerate.
 217         * man/Makefile.in: Regenerate.
 218         * testsuite/Makefile.in: Regenerate.
 219
 220 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
 221
 222         * src/powerpc/aix_closure.S: Reorganize 64-bit code to match
 223         linux64_closure.S.
 224
 225 2009-12-04  Uros Bizjak  <ubizjak@gmail.com>
 226
 227         PR libffi/41908
 228         * src/x86/ffi64.c (classify_argument): Update from
 229         gcc/config/i386/i386.c.
 230         (ffi_closure_unix64_inner): Do not use the address of two consecutive
 231         SSE registers directly.
 232         * testsuite/libffi.call/cls_dbls_struct.c (main): Remove xfail
 233         for x86_64 linux targets.
 234
 235 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
 236
 237         * src/powerpc/ffi_darwin.c (ffi_closure_helper_DARWIN): Increment
 238         pfr for long double split between fpr13 and stack.
 239
 240 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
 241
 242         * src/powerpc/ffi_darwin.c (ffi_prep_args): Increment next_arg and
 243         fparg_count twice for long double.
 244
 245 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
 246
 247         PR libffi/42243
 248         * src/powerpc/ffi_darwin.c (ffi_prep_args): Remove extra parentheses.
 249
 250 2009-12-03  Uros Bizjak  <ubizjak@gmail.com>
 251
 252         * testsuite/libffi.call/cls_longdouble_va.c (main): Fix format string.
 253         Remove xfails for x86 linux targets.
 254
 255 2009-12-02  David Edelsohn  <edelsohn@gnu.org>
 256
 257         * src/powerpc/ffi_darwin.c (ffi_prep_args): Fix typo in INT64
 258         case.
 259
 260 2009-12-01  David Edelsohn  <edelsohn@gnu.org>
 261
 262         * src/powerpc/aix.S (ffi_call_AIX): Convert to more standard
 263         register usage.  Call ffi_prep_args directly.  Add long double
 264         return value support.
 265         * src/powerpc/ffi_darwin.c (ffi_prep_args): Double arg increment
 266         applies to FFI_TYPE_DOUBLE.  Correct fpr_base increment typo.
 267         Separate FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases.
 268         (ffi_prep_cif_machdep): Only 16 byte stack alignment in 64 bit
 269         mode.
 270         (ffi_closure_helper_DARWIN): Remove nf and ng counters.  Move temp
 271         into case.
 272         * src/powerpc/aix_closure.S: Maintain 16 byte stack alignment.
 273         Allocate result area between params and FPRs.
 274
 275 2009-11-30  David Edelsohn  <edelsohn@gnu.org>
 276
 277         PR target/35484
 278         * src/powerpc/ffitarget.h (POWERPC64): Define for PPC64 Linux and
 279         AIX64.
 280         * src/powerpc/aix.S: Implement AIX64 version.
 281         * src/powerpc/aix_closure.S: Implement AIX64 version.
 282         (ffi_closure_ASM): Use extsb, lha and displament addresses.
 283         * src/powerpc/ffi_darwin.c (ffi_prep_args): Implement AIX64
 284         support.
 285         (ffi_prep_cif_machdep): Same.
 286         (ffi_call): Same.
 287         (ffi_closure_helper_DARWIN): Same.
 288
 289 2009-11-02  Andreas Tobler  <a.tobler@schweiz.org>
 290
 291         PR libffi/41908
 292         * testsuite/libffi.call/testclosure.c: New test.
 293
 294 2009-09-28  Kai Tietz  <kai.tietz@onevision.com>
 295
 296         * src/x86/win64.S (_ffi_call_win64 stack): Remove for gnu
 297         assembly version use of ___chkstk.
 298
 299 2009-09-23  Matthias Klose  <doko@ubuntu.com>
 300
 301         PR libffi/40242, PR libffi/41443
 302         * src/arm/sysv.S (__ARM_ARCH__): Define for processors
 303         __ARM_ARCH_6T2__, __ARM_ARCH_6M__, __ARM_ARCH_7__,
 304         __ARM_ARCH_7A__, __ARM_ARCH_7R__, __ARM_ARCH_7M__.
 305         Change the conditionals to __SOFTFP__ || __ARM_EABI__
 306         for -mfloat-abi=softfp to work.
 307
 308 2009-09-17  Loren J. Rittle  <ljrittle@acm.org>
 309
 310         PR testsuite/32843 (strikes again)
 311         * src/x86/ffi.c (ffi_prep_cif_machdep): Add X86_FREEBSD to
 312         enable proper extension on char and short.
 313
 314 2009-09-15  David Daney  <ddaney@caviumnetworks.com>
 315
 316         * src/java_raw_api.c (ffi_java_raw_to_rvalue): Remove special
 317         handling for FFI_TYPE_POINTER.
 318         * src/mips/ffitarget.h (FFI_TYPE_STRUCT_D_SOFT,
 319         FFI_TYPE_STRUCT_F_SOFT, FFI_TYPE_STRUCT_DD_SOFT,
 320         FFI_TYPE_STRUCT_FF_SOFT, FFI_TYPE_STRUCT_FD_SOFT,
 321         FFI_TYPE_STRUCT_DF_SOFT, FFI_TYPE_STRUCT_SOFT): New defines.
 322         (FFI_N32_SOFT_FLOAT, FFI_N64_SOFT_FLOAT): New ffi_abi enumerations.
 323         (enum ffi_abi): Set FFI_DEFAULT_ABI for soft-float.
 324         * src/mips/n32.S (ffi_call_N32): Add handling for soft-float
 325         structure and pointer returns.
 326         (ffi_closure_N32): Add handling for pointer returns.
 327         * src/mips/ffi.c (ffi_prep_args, calc_n32_struct_flags,
 328         calc_n32_return_struct_flags): Handle soft-float.
 329         (ffi_prep_cif_machdep):  Handle soft-float, fix pointer handling.
 330         (ffi_call_N32): Declare proper argument types.
 331         (ffi_call, copy_struct_N32, ffi_closure_mips_inner_N32): Handle
 332         soft-float.
 333
 334 2009-08-24  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 335
 336         * configure.ac (AC_PREREQ): Bump to 2.64.
 337
 338 2009-08-22  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 339
 340         * Makefile.am (install-html, install-pdf): Remove.
 341         * Makefile.in: Regenerate.
 342
 343         * Makefile.in: Regenerate.
 344         * aclocal.m4: Regenerate.
 345         * configure: Regenerate.
 346         * fficonfig.h.in: Regenerate.
 347         * include/Makefile.in: Regenerate.
 348         * man/Makefile.in: Regenerate.
 349         * testsuite/Makefile.in: Regenerate.
 350
 351 2009-07-30  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 352
 353         * configure.ac (_AC_ARG_VAR_PRECIOUS): Use m4_rename_force.
 354
 355 2009-07-24  Dave Korn  <dave.korn.cygwin@gmail.com>
 356
 357         PR libffi/40807
 358         * src/x86/ffi.c (ffi_prep_cif_machdep): Also use sign/zero-extending
 359         return types for X86_WIN32.
 360         * src/x86/win32.S (_ffi_call_SYSV): Handle omitted return types.
 361         (_ffi_call_STDCALL, _ffi_closure_SYSV, _ffi_closure_raw_SYSV,
 362         _ffi_closure_STDCALL): Likewise.
 363
 364         * src/closures.c (is_selinux_enabled): Define to const 0 for Cygwin.
 365         (dlmmap, dlmunmap): Also use these functions on Cygwin.
 366
 367 2009-07-11  Richard Sandiford  <rdsandiford@googlemail.com>
 368
 369         PR testsuite/40699
 370         PR testsuite/40707
 371         PR testsuite/40709
 372         * testsuite/lib/libffi-dg.exp: Revert 2009-07-02, 2009-07-01 and
 373         2009-06-30 commits.
 374
 375 2009-07-01  Richard Sandiford  <r.sandiford@uk.ibm.com>
 376
 377         * testsuite/lib/libffi-dg.exp (libffi-init): Set ld_library_path
 378         to "" before adding paths.  (This reinstates an assignment that
 379         was removed by my 2009-06-30 commit, but changes the initial
 380         value from "." to "".)
 381
 382 2009-07-01  H.J. Lu  <hongjiu.lu@intel.com>
 383
 384         PR testsuite/40601
 385         * testsuite/lib/libffi-dg.exp (libffi-init): Properly set
 386         gccdir.  Adjust ld_library_path for gcc only if gccdir isn't
 387         empty.
 388
 389 2009-06-30  Richard Sandiford  <r.sandiford@uk.ibm.com>
 390
 391         * testsuite/lib/libffi-dg.exp (libffi-init): Don't add "."
 392         to ld_library_path.  Use add_path.  Add just find_libgcc_s
 393         to ld_library_path, not every libgcc multilib directory.
 394
 395 2009-06-16  Wim Lewis  <wiml@hhhh.org>
 396
 397         * src/powerpc/ffi.c: Avoid clobbering cr3 and cr4, which are
 398         supposed to be callee-saved.
 399         * src/powerpc/sysv.S (small_struct_return_value): Fix overrun of
 400         return buffer for odd-size structs.
 401
 402 2009-06-16  Andreas Tobler  <a.tobler@schweiz.org>
 403
 404         PR libffi/40444
 405         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Add
 406         allow_stack_execute for Darwin.
 407
 408 2009-06-16  Andrew Haley  <aph@redhat.com>
 409
 410         * configure.ac (TARGETDIR): Add missing blank lines.
 411         * configure: Regenerate.
 412
 413 2009-06-16  Andrew Haley  <aph@redhat.com>
 414
 415         * testsuite/libffi.call/cls_align_sint64.c,
 416         testsuite/libffi.call/cls_align_uint64.c,
 417         testsuite/libffi.call/cls_longdouble_va.c,
 418         testsuite/libffi.call/cls_ulonglong.c,
 419         testsuite/libffi.call/return_ll1.c,
 420         testsuite/libffi.call/stret_medium2.c: Fix printf format
 421         specifiers.
 422         * testsuite/libffi.call/ffitest.h,
 423         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRIuLL): Define.
 424
 425 2009-06-15  Andrew Haley  <aph@redhat.com>
 426
 427         * testsuite/libffi.call/err_bad_typedef.c: xfail everywhere.
 428         * testsuite/libffi.call/err_bad_abi.c: Likewise.
 429
 430 2009-06-12  Andrew Haley  <aph@redhat.com>
 431
 432         * Makefile.am: Remove info_TEXINFOS.
 433
 434 2009-06-12  Andrew Haley  <aph@redhat.com>
 435
 436         * ChangeLog.libffi: testsuite/libffi.call/cls_align_sint64.c,
 437         testsuite/libffi.call/cls_align_uint64.c,
 438         testsuite/libffi.call/cls_ulonglong.c,
 439         testsuite/libffi.call/return_ll1.c,
 440         testsuite/libffi.call/stret_medium2.c: Fix printf format
 441         specifiers.
 442         testsuite/libffi.special/unwindtest.cc: include stdint.h.
 443
 444 2009-06-11  Timothy Wall  <twall@users.sf.net>
 445
 446         * Makefile.am,
 447         configure.ac,
 448         include/ffi.h.in,
 449         include/ffi_common.h,
 450         src/closures.c,
 451         src/dlmalloc.c,
 452         src/x86/ffi.c,
 453         src/x86/ffitarget.h,
 454         src/x86/win64.S (new),
 455         README: Added win64 support (mingw or MSVC)
 456         * Makefile.in,
 457         include/Makefile.in,
 458         man/Makefile.in,
 459         testsuite/Makefile.in,
 460         configure,
 461         aclocal.m4: Regenerated
 462         * ltcf-c.sh: properly escape cygwin/w32 path
 463         * man/ffi_call.3: Clarify size requirements for return value.
 464         * src/x86/ffi64.c: Fix filename in comment.
 465         * src/x86/win32.S: Remove unused extern.
 466
 467         * testsuite/libffi.call/closure_fn0.c,
 468         testsuite/libffi.call/closure_fn1.c,
 469         testsuite/libffi.call/closure_fn2.c,
 470         testsuite/libffi.call/closure_fn3.c,
 471         testsuite/libffi.call/closure_fn4.c,
 472         testsuite/libffi.call/closure_fn5.c,
 473         testsuite/libffi.call/closure_fn6.c,
 474         testsuite/libffi.call/closure_stdcall.c,
 475         testsuite/libffi.call/cls_12byte.c,
 476         testsuite/libffi.call/cls_16byte.c,
 477         testsuite/libffi.call/cls_18byte.c,
 478         testsuite/libffi.call/cls_19byte.c,
 479         testsuite/libffi.call/cls_1_1byte.c,
 480         testsuite/libffi.call/cls_20byte.c,
 481         testsuite/libffi.call/cls_20byte1.c,
 482         testsuite/libffi.call/cls_24byte.c,
 483         testsuite/libffi.call/cls_2byte.c,
 484         testsuite/libffi.call/cls_3_1byte.c,
 485         testsuite/libffi.call/cls_3byte1.c,
 486         testsuite/libffi.call/cls_3byte2.c,
 487         testsuite/libffi.call/cls_4_1byte.c,
 488         testsuite/libffi.call/cls_4byte.c,
 489         testsuite/libffi.call/cls_5_1_byte.c,
 490         testsuite/libffi.call/cls_5byte.c,
 491         testsuite/libffi.call/cls_64byte.c,
 492         testsuite/libffi.call/cls_6_1_byte.c,
 493         testsuite/libffi.call/cls_6byte.c,
 494         testsuite/libffi.call/cls_7_1_byte.c,
 495         testsuite/libffi.call/cls_7byte.c,
 496         testsuite/libffi.call/cls_8byte.c,
 497         testsuite/libffi.call/cls_9byte1.c,
 498         testsuite/libffi.call/cls_9byte2.c,
 499         testsuite/libffi.call/cls_align_double.c,
 500         testsuite/libffi.call/cls_align_float.c,
 501         testsuite/libffi.call/cls_align_longdouble.c,
 502         testsuite/libffi.call/cls_align_longdouble_split.c,
 503         testsuite/libffi.call/cls_align_longdouble_split2.c,
 504         testsuite/libffi.call/cls_align_pointer.c,
 505         testsuite/libffi.call/cls_align_sint16.c,
 506         testsuite/libffi.call/cls_align_sint32.c,
 507         testsuite/libffi.call/cls_align_sint64.c,
 508         testsuite/libffi.call/cls_align_uint16.c,
 509         testsuite/libffi.call/cls_align_uint32.c,
 510         testsuite/libffi.call/cls_align_uint64.c,
 511         testsuite/libffi.call/cls_dbls_struct.c,
 512         testsuite/libffi.call/cls_double.c,
 513         testsuite/libffi.call/cls_double_va.c,
 514         testsuite/libffi.call/cls_float.c,
 515         testsuite/libffi.call/cls_longdouble.c,
 516         testsuite/libffi.call/cls_longdouble_va.c,
 517         testsuite/libffi.call/cls_multi_schar.c,
 518         testsuite/libffi.call/cls_multi_sshort.c,
 519         testsuite/libffi.call/cls_multi_sshortchar.c,
 520         testsuite/libffi.call/cls_multi_uchar.c,
 521         testsuite/libffi.call/cls_multi_ushort.c,
 522         testsuite/libffi.call/cls_multi_ushortchar.c,
 523         testsuite/libffi.call/cls_pointer.c,
 524         testsuite/libffi.call/cls_pointer_stack.c,
 525         testsuite/libffi.call/cls_schar.c,
 526         testsuite/libffi.call/cls_sint.c,
 527         testsuite/libffi.call/cls_sshort.c,
 528         testsuite/libffi.call/cls_uchar.c,
 529         testsuite/libffi.call/cls_uint.c,
 530         testsuite/libffi.call/cls_ulonglong.c,
 531         testsuite/libffi.call/cls_ushort.c,
 532         testsuite/libffi.call/err_bad_abi.c,
 533         testsuite/libffi.call/err_bad_typedef.c,
 534         testsuite/libffi.call/float2.c,
 535         testsuite/libffi.call/huge_struct.c,
 536         testsuite/libffi.call/nested_struct.c,
 537         testsuite/libffi.call/nested_struct1.c,
 538         testsuite/libffi.call/nested_struct10.c,
 539         testsuite/libffi.call/nested_struct2.c,
 540         testsuite/libffi.call/nested_struct3.c,
 541         testsuite/libffi.call/nested_struct4.c,
 542         testsuite/libffi.call/nested_struct5.c,
 543         testsuite/libffi.call/nested_struct6.c,
 544         testsuite/libffi.call/nested_struct7.c,
 545         testsuite/libffi.call/nested_struct8.c,
 546         testsuite/libffi.call/nested_struct9.c,
 547         testsuite/libffi.call/problem1.c,
 548         testsuite/libffi.call/return_ldl.c,
 549         testsuite/libffi.call/return_ll1.c,
 550         testsuite/libffi.call/stret_large.c,
 551         testsuite/libffi.call/stret_large2.c,
 552         testsuite/libffi.call/stret_medium.c,
 553         testsuite/libffi.call/stret_medium2.c,
 554         testsuite/libffi.special/unwindtest.cc: use ffi_closure_alloc instead
 555         of checking for MMAP.  Use intptr_t instead of long casts.
 556
 557 2009-06-11  Kaz Kojima  <kkojima@gcc.gnu.org>
 558
 559         * testsuite/libffi.call/cls_longdouble_va.c: Add xfail sh*-*-linux-*.
 560         * testsuite/libffi.call/err_bad_abi.c: Add xfail sh*-*-*.
 561         * testsuite/libffi.call/err_bad_typedef.c: Likewise.
 562
 563 2009-06-09  Andrew Haley  <aph@redhat.com>
 564
 565         * src/x86/freebsd.S: Add missing file.
 566
 567 2009-06-08  Andrew Haley  <aph@redhat.com>
 568
 569         Import from libffi 3.0.8:
 570
 571         * doc/libffi.texi: New file.
 572         * doc/libffi.info: Likewise.
 573         * doc/stamp-vti: Likewise.
 574         * man/Makefile.am: New file.
 575         * man/ffi_call.3: New file.
 576
 577         * Makefile.am (EXTRA_DIST): Add src/x86/darwin64.S,
 578         src/dlmalloc.c.
 579         (nodist_libffi_la_SOURCES): Add X86_FREEBSD.
 580
 581         * configure.ac: Bump version to 3.0.8.
 582         parisc*-*-linux*: Add.
 583         i386-*-freebsd* | i386-*-openbsd*: Add.
 584         powerpc-*-beos*: Add.
 585         AM_CONDITIONAL X86_FREEBSD: Add.
 586         AC_CONFIG_FILES: Add man/Makefile.
 587
 588         * include/ffi.h.in (FFI_FN): Change void (*)() to void (*)(void).
 589
 590 2009-06-08  Andrew Haley  <aph@redhat.com>
 591
 592         * README: Import from libffi 3.0.8.
 593
 594 2009-06-08  Andrew Haley  <aph@redhat.com>
 595
 596         * testsuite/libffi.call/err_bad_abi.c: Add xfails.
 597         * testsuite/libffi.call/cls_longdouble_va.c: Add xfails.
 598         * testsuite/libffi.call/cls_dbls_struct.c: Add xfail x86_64-*-linux-*.
 599         * testsuite/libffi.call/err_bad_typedef.c: Add xfails.
 600
 601         * testsuite/libffi.call/stret_medium2.c: Add __UNUSED__ to args.
 602         * testsuite/libffi.call/stret_medium.c: Likewise.
 603         * testsuite/libffi.call/stret_large2.c: Likewise.
 604         * testsuite/libffi.call/stret_large.c:  Likewise.
 605
 606 2008-12-26  Timothy Wall  <twall@users.sf.net>
 607
 608         * testsuite/libffi.call/cls_longdouble.c,
 609         testsuite/libffi.call/cls_longdouble_va.c,
 610         testsuite/libffi.call/cls_align_longdouble.c,
 611         testsuite/libffi.call/cls_align_longdouble_split.c,
 612         testsuite/libffi.call/cls_align_longdouble_split2.c: mark expected
 613         failures on x86_64 cygwin/mingw.
 614
 615 2008-12-22  Timothy Wall  <twall@users.sf.net>
 616
 617         * testsuite/libffi.call/closure_fn0.c,
 618         testsuite/libffi.call/closure_fn1.c,
 619         testsuite/libffi.call/closure_fn2.c,
 620         testsuite/libffi.call/closure_fn3.c,
 621         testsuite/libffi.call/closure_fn4.c,
 622         testsuite/libffi.call/closure_fn5.c,
 623         testsuite/libffi.call/closure_fn6.c,
 624         testsuite/libffi.call/closure_loc_fn0.c,
 625         testsuite/libffi.call/closure_stdcall.c,
 626         testsuite/libffi.call/cls_align_pointer.c,
 627         testsuite/libffi.call/cls_pointer.c,
 628         testsuite/libffi.call/cls_pointer_stack.c: use portable cast from
 629         pointer to integer (intptr_t).
 630         * testsuite/libffi.call/cls_longdouble.c: disable for win64.
 631
 632 2008-07-24  Anthony Green  <green@redhat.com>
 633
 634         * testsuite/libffi.call/cls_dbls_struct.c,
 635         testsuite/libffi.call/cls_double_va.c,
 636         testsuite/libffi.call/cls_longdouble.c,
 637         testsuite/libffi.call/cls_longdouble_va.c,
 638         testsuite/libffi.call/cls_pointer.c,
 639         testsuite/libffi.call/cls_pointer_stack.c,
 640         testsuite/libffi.call/err_bad_abi.c: Clean up failures from
 641         compiler warnings.
 642
 643 2008-03-04  Anthony Green  <green@redhat.com>
 644             Blake Chaffin
 645             hos@tamanegi.org
 646
 647         * testsuite/libffi.call/cls_align_longdouble_split2.c
 648           testsuite/libffi.call/cls_align_longdouble_split.c
 649           testsuite/libffi.call/cls_dbls_struct.c
 650           testsuite/libffi.call/cls_double_va.c
 651           testsuite/libffi.call/cls_longdouble.c
 652           testsuite/libffi.call/cls_longdouble_va.c
 653           testsuite/libffi.call/cls_pointer.c
 654           testsuite/libffi.call/cls_pointer_stack.c
 655           testsuite/libffi.call/err_bad_abi.c
 656           testsuite/libffi.call/err_bad_typedef.c
 657           testsuite/libffi.call/stret_large2.c
 658           testsuite/libffi.call/stret_large.c
 659           testsuite/libffi.call/stret_medium2.c
 660           testsuite/libffi.call/stret_medium.c: New tests from Apple.
 661
 662 2009-06-05  Andrew Haley  <aph@redhat.com>
 663
 664         * src/x86/ffitarget.h, src/x86/ffi.c: Merge stdcall changes from
 665         libffi.
 666
 667 2009-06-04  Andrew Haley  <aph@redhat.com>
 668
 669         * src/x86/ffitarget.h, src/x86/win32.S, src/x86/ffi.c: Back out
 670         stdcall changes.
 671
 672 2008-02-26  Anthony Green  <green@redhat.com>
 673             Thomas Heller  <theller@ctypes.org>
 674
 675         * src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C
 676         comment.
 677
 678 2008-02-03  Timothy Wall  <twall@users.sf.net>
 679
 680         * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return
 681           offset based on code pointer, not data pointer.
 682
 683 2008-01-31  Timothy Wall <twall@users.sf.net>
 684
 685         * testsuite/libffi.call/closure_stdcall.c: Add test for stdcall
 686         closures.
 687         * src/x86/ffitarget.h: Increase size of trampoline for stdcall
 688         closures.
 689         * src/x86/win32.S: Add assembly for stdcall closure.
 690         * src/x86/ffi.c: Initialize stdcall closure trampoline.
 691
 692 2009-06-04  Andrew Haley  <aph@redhat.com>
 693
 694         * include/ffi.h.in: Change void (*)() to void (*)(void).
 695         * src/x86/ffi.c: Likewise.
 696
 697 2009-06-04  Andrew Haley  <aph@redhat.com>
 698
 699         * src/powerpc/ppc_closure.S: Insert licence header.
 700         * src/powerpc/linux64_closure.S: Likewise.
 701         * src/m68k/sysv.S: Likewise.
 702
 703         * src/sh64/ffi.c: Change void (*)() to void (*)(void).
 704         * src/powerpc/ffi.c: Likewise.
 705         * src/powerpc/ffi_darwin.c: Likewise.
 706         * src/m32r/ffi.c: Likewise.
 707         * src/sh64/ffi.c: Likewise.
 708         * src/x86/ffi64.c: Likewise.
 709         * src/alpha/ffi.c: Likewise.
 710         * src/alpha/osf.S: Likewise.
 711         * src/frv/ffi.c: Likewise.
 712         * src/s390/ffi.c: Likewise.
 713         * src/pa/ffi.c: Likewise.
 714         * src/pa/hpux32.S: Likewise.
 715         * src/ia64/unix.S: Likewise.
 716         * src/ia64/ffi.c: Likewise.
 717         * src/sparc/ffi.c: Likewise.
 718         * src/mips/ffi.c: Likewise.
 719         * src/sh/ffi.c: Likewise.
 720
 721 2008-02-15  David Daney  <ddaney@avtrex.com>
 722
 723         * src/mips/ffi.c (USE__BUILTIN___CLEAR_CACHE):
 724         Define (conditionally), and use it to include cachectl.h.
 725         (ffi_prep_closure_loc): Fix cache flushing.
 726         * src/mips/ffitarget.h (_ABIN32, _ABI64, _ABIO32): Define.
 727
 728 2009-06-04  Andrew Haley  <aph@redhat.com>
 729
 730         include/ffi.h.in,
 731         src/arm/ffitarget.h,
 732         src/arm/ffi.c,
 733         src/arm/sysv.S,
 734         src/powerpc/ffitarget.h,
 735         src/closures.c,
 736         src/sh64/ffitarget.h,
 737         src/sh64/ffi.c,
 738         src/sh64/sysv.S,
 739         src/types.c,
 740         src/x86/ffi64.c,
 741         src/x86/ffitarget.h,
 742         src/x86/win32.S,
 743         src/x86/darwin.S,
 744         src/x86/ffi.c,
 745         src/x86/sysv.S,
 746         src/x86/unix64.S,
 747         src/alpha/ffitarget.h,
 748         src/alpha/ffi.c,
 749         src/alpha/osf.S,
 750         src/m68k/ffitarget.h,
 751         src/frv/ffitarget.h,
 752         src/frv/ffi.c,
 753         src/s390/ffitarget.h,
 754         src/s390/sysv.S,
 755         src/cris/ffitarget.h,
 756         src/pa/linux.S,
 757         src/pa/ffitarget.h,
 758         src/pa/ffi.c,
 759         src/raw_api.c,
 760         src/ia64/ffitarget.h,
 761         src/ia64/unix.S,
 762         src/ia64/ffi.c,
 763         src/ia64/ia64_flags.h,
 764         src/java_raw_api.c,
 765         src/debug.c,
 766         src/sparc/v9.S,
 767         src/sparc/ffitarget.h,
 768         src/sparc/ffi.c,
 769         src/sparc/v8.S,
 770         src/mips/ffitarget.h,
 771         src/mips/n32.S,
 772         src/mips/o32.S,
 773         src/mips/ffi.c,
 774         src/prep_cif.c,
 775         src/sh/ffitarget.h,
 776         src/sh/ffi.c,
 777         src/sh/sysv.S: Update license text.
 778
 779 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
 780
 781         * src/x86/win32.S (_ffi_closure_STDCALL):  New function.
 782         (.eh_frame):  Add FDE for it.
 783
 784 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
 785
 786         * configure.ac:  Also check if assembler supports pc-relative
 787         relocs on X86_WIN32 targets.
 788         * configure:  Regenerate.
 789         * src/x86/win32.S (ffi_prep_args):  Declare extern, not global.
 790         (_ffi_call_SYSV):  Add missing function type symbol .def and
 791         add EH markup labels.
 792         (_ffi_call_STDCALL):  Likewise.
 793         (_ffi_closure_SYSV):  Likewise.
 794         (_ffi_closure_raw_SYSV):  Likewise.
 795         (.eh_frame):  Add hand-crafted EH data.
 796
 797 2009-04-09  Jakub Jelinek  <jakub@redhat.com>
 798
 799         * testsuite/lib/libffi-dg.exp: Change copyright header to refer to
 800         version 3 of the GNU General Public License and to point readers
 801         at the COPYING3 file and the FSF's license web page.
 802         * testsuite/libffi.call/call.exp: Likewise.
 803         * testsuite/libffi.special/special.exp: Likewise.
 804
 805 2009-03-01  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 806
 807         * configure: Regenerate.
 808
 809 2008-12-18  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
 810
 811         PR libffi/26048
 812         * configure.ac (HAVE_AS_X86_PCREL): New test.
 813         * configure: Regenerate.
 814         * fficonfig.h.in: Regenerate.
 815         * src/x86/sysv.S [!FFI_NO_RAW_API]: Precalculate
 816         RAW_CLOSURE_CIF_OFFSET, RAW_CLOSURE_FUN_OFFSET,
 817         RAW_CLOSURE_USER_DATA_OFFSET for the Solaris 10/x86 assembler.
 818         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
 819         * src/x86/unix64.S (.Lstore_table): Move to .text section.
 820         (.Lload_table): Likewise.
 821         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
 822
 823 2008-12-18  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 824
 825         * configure: Regenerate.
 826
 827 2008-11-21  Eric Botcazou  <ebotcazou@adacore.com>
 828
 829         * src/sparc/ffi.c (ffi_prep_cif_machdep): Add support for
 830         signed/unsigned int8/16 return values.
 831         * src/sparc/v8.S (ffi_call_v8): Likewise.
 832         (ffi_closure_v8): Likewise.
 833
 834 2008-09-26  Peter O'Gorman  <pogma@thewrittenword.com>
 835             Steve Ellcey  <sje@cup.hp.com>
 836
 837         * configure: Regenerate for new libtool.
 838         * Makefile.in: Ditto.
 839         * include/Makefile.in: Ditto.
 840         * aclocal.m4: Ditto.
 841
 842 2008-08-25  Andreas Tobler  <a.tobler@schweiz.org>
 843
 844         * src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and
 845         FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum.
 846         Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT.
 847         Adjust copyright notice.
 848         * src/powerpc/ffi.c: Add two new flags to indicate if we have one
 849         register or two register to use for FFI_SYSV structs.
 850         (ffi_prep_cif_machdep): Pass the right register flag introduced above.
 851         (ffi_closure_helper_SYSV): Fix the return type for
 852         FFI_SYSV_TYPE_SMALL_STRUCT. Comment.
 853         Adjust copyright notice.
 854
 855 2008-07-16  Kaz Kojima  <kkojima@gcc.gnu.org>
 856
 857         * src/sh/ffi.c (ffi_prep_closure_loc): Turn INSN into an unsigned
 858         int.
 859
 860 2008-06-17  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 861
 862         * configure: Regenerate.
 863         * include/Makefile.in: Regenerate.
 864         * testsuite/Makefile.in: Regenerate.
 865
 866 2008-06-07  Joseph Myers  <joseph@codesourcery.com>
 867
 868         * configure.ac (parisc*-*-linux*, powerpc-*-sysv*,
 869         powerpc-*-beos*): Remove.
 870         * configure: Regenerate.
 871
 872 2008-05-09  Julian Brown  <julian@codesourcery.com>
 873
 874         * Makefile.am (LTLDFLAGS): New.
 875         (libffi_la_LDFLAGS): Use above.
 876         * Makefile.in: Regenerate.
 877
 878 2008-04-18  Paolo Bonzini  <bonzini@gnu.org>
 879
 880         PR bootstrap/35457
 881         * aclocal.m4: Regenerate.
 882         * configure: Regenerate.
 883
 884 2008-03-26  Kaz Kojima  <kkojima@gcc.gnu.org>
 885
 886         * src/sh/sysv.S: Add .note.GNU-stack on Linux.
 887         * src/sh64/sysv.S: Likewise.
 888
 889 2008-03-26  Daniel Jacobowitz  <dan@debian.org>
 890
 891         * src/arm/sysv.S: Fix ARM comment marker.
 892
 893 2008-03-26  Jakub Jelinek  <jakub@redhat.com>
 894
 895         * src/alpha/osf.S: Add .note.GNU-stack on Linux.
 896         * src/s390/sysv.S: Likewise.
 897         * src/powerpc/ppc_closure.S: Likewise.
 898         * src/powerpc/sysv.S: Likewise.
 899         * src/x86/unix64.S: Likewise.
 900         * src/x86/sysv.S: Likewise.
 901         * src/sparc/v8.S: Likewise.
 902         * src/sparc/v9.S: Likewise.
 903         * src/m68k/sysv.S: Likewise.
 904         * src/arm/sysv.S: Likewise.
 905
 906 2008-03-16  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 907
 908         * aclocal.m4: Regenerate.
 909         * configure: Likewise.
 910         * Makefile.in: Likewise.
 911         * include/Makefile.in: Likewise.
 912         * testsuite/Makefile.in: Likewise.
 913
 914 2008-02-12  Bjoern Koenig  <bkoenig@alpha-tierchen.de>
 915             Andreas Tobler  <a.tobler@schweiz.org>
 916
 917         * configure.ac: Add amd64-*-freebsd* target.
 918         * configure: Regenerate.
 919
 920 2008-01-30  H.J. Lu  <hongjiu.lu@intel.com>
 921
 922         PR libffi/34612
 923         * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when
 924         returning struct.
 925
 926         * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer"
 927         tests.
 928
 929 2008-01-24  David Edelsohn  <edelsohn@gnu.org>
 930
 931         * configure: Regenerate.
 932
 933 2008-01-06  Andreas Tobler  <a.tobler@schweiz.org>
 934
 935         * src/x86/ffi.c (ffi_prep_cif_machdep): Fix thinko.
 936
 937 2008-01-05  Andreas Tobler  <a.tobler@schweiz.org>
 938
 939         PR testsuite/32843
 940         * src/x86/ffi.c (ffi_prep_cif_machdep): Add code for
 941         signed/unsigned int8/16 for X86_DARWIN.
 942         Updated copyright info.
 943         Handle one and two byte structs with special cif->flags.
 944         * src/x86/ffitarget.h: Add special types for one and two byte structs.
 945         Updated copyright info.
 946         * src/x86/darwin.S (ffi_call_SYSV): Rewrite to use a jump table like
 947         sysv.S
 948         Remove code to pop args from the stack after call.
 949         Special-case signed/unsigned for int8/16, one and two byte structs.
 950         (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
 951         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
 952         FFI_TYPE_SINT32.
 953         Updated copyright info.
 954
 955 2007-12-08  David Daney  <ddaney@avtrex.com>
 956
 957         * src/mips/n32.S (ffi_call_N32):  Replace dadd with ADDU, dsub with
 958         SUBU, add with ADDU and use smaller code sequences.
 959
 960 2007-12-07  David Daney  <ddaney@avtrex.com>
 961
 962         * src/mips/ffi.c (ffi_prep_cif_machdep): Handle long double return
 963         type.
 964
 965 2007-12-06  David Daney  <ddaney@avtrex.com>
 966
 967         * include/ffi.h.in (FFI_SIZEOF_JAVA_RAW): Define if not already
 968         defined.
 969         (ffi_java_raw): New typedef.
 970         (ffi_java_raw_call, ffi_java_ptrarray_to_raw,
 971         ffi_java_raw_to_ptrarray): Change parameter types from ffi_raw to
 972         ffi_java_raw.
 973         (ffi_java_raw_closure) : Same.
 974         (ffi_prep_java_raw_closure, ffi_prep_java_raw_closure_loc): Change
 975         parameter types.
 976         * src/java_raw_api.c (ffi_java_raw_size):  Replace FFI_SIZEOF_ARG with
 977         FFI_SIZEOF_JAVA_RAW.
 978         (ffi_java_raw_to_ptrarray): Change type of raw to ffi_java_raw.
 979         Replace FFI_SIZEOF_ARG with FFI_SIZEOF_JAVA_RAW. Use
 980         sizeof(ffi_java_raw) for alignment calculations.
 981         (ffi_java_ptrarray_to_raw): Same.
 982         (ffi_java_rvalue_to_raw): Add special handling for FFI_TYPE_POINTER
 983         if FFI_SIZEOF_JAVA_RAW == 4.
 984         (ffi_java_raw_to_rvalue): Same.
 985         (ffi_java_raw_call): Change type of raw to ffi_java_raw.
 986         (ffi_java_translate_args): Same.
 987         (ffi_prep_java_raw_closure_loc, ffi_prep_java_raw_closure): Change
 988         parameter types.
 989         * src/mips/ffitarget.h (FFI_SIZEOF_JAVA_RAW): Define for N32 ABI.
 990
 991 2007-12-06  David Daney  <ddaney@avtrex.com>
 992
 993         * src/mips/n32.S (ffi_closure_N32): Use 64-bit add instruction on
 994         pointer values.
 995
 996 2007-12-01  Andreas Tobler  <a.tobler@schweiz.org>
 997
 998         PR libffi/31937
 999         * src/powerpc/ffitarget.h: Introduce new ABI FFI_LINUX_SOFT_FLOAT.
1000         Add local FFI_TYPE_UINT128 to handle soft-float long-double-128.
1001         * src/powerpc/ffi.c: Distinguish between __NO_FPRS__ and not and
1002         set the NUM_FPR_ARG_REGISTERS according to.
1003         Add support for potential soft-float support under hard-float
1004         architecture.
1005         (ffi_prep_args_SYSV): Set NUM_FPR_ARG_REGISTERS to 0 in case of
1006         FFI_LINUX_SOFT_FLOAT, handle float, doubles and long-doubles according
1007         to the FFI_LINUX_SOFT_FLOAT ABI.
1008         (ffi_prep_cif_machdep): Likewise.
1009         (ffi_closure_helper_SYSV): Likewise.
1010         * src/powerpc/ppc_closure.S: Make sure not to store float/double
1011         on archs where __NO_FPRS__ is true.
1012         Add FFI_TYPE_UINT128 support.
1013         * src/powerpc/sysv.S: Add support for soft-float long-double-128.
1014         Adjust copyright notice.
1015
1016 2007-11-25  Andreas Tobler  <a.tobler@schweiz.org>
1017
1018         * src/closures.c: Move defintion of MAYBE_UNUSED from here to ...
1019         * include/ffi_common.h: ... here.
1020         Update copyright.
1021
1022 2007-11-17  Andreas Tobler  <a.tobler@schweiz.org>
1023
1024         * src/powerpc/sysv.S: Load correct cr to compare if we have long double.
1025         * src/powerpc/linux64.S: Likewise.
1026         * src/powerpc/ffi.c: Add a comment to show which part goes into cr6.
1027         * testsuite/libffi.call/return_ldl.c: New test.
1028
1029 2007-09-04    <aph@redhat.com>
1030
1031         * src/arm/sysv.S (UNWIND): New.
1032         (Whole file): Conditionally compile unwinder directives.
1033         * src/arm/sysv.S: Add unwinder directives.
1034
1035         * src/arm/ffi.c (ffi_prep_args): Align structs by at least 4 bytes.
1036         Only treat r0 as a struct address if we're actually returning a
1037         struct by address.
1038         Only copy the bytes that are actually within a struct.
1039         (ffi_prep_cif_machdep): A Composite Type not larger than 4 bytes
1040         is returned in r0, not passed by address.
1041         (ffi_call): Allocate a word-sized temporary for the case where
1042         a composite is returned in r0.
1043         (ffi_prep_incoming_args_SYSV): Align as necessary.
1044
1045 2007-08-05  Steven Newbury  <s_j_newbury@yahoo.co.uk>
1046
1047         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Use __clear_cache instead of
1048         directly using the sys_cacheflush syscall.
1049
1050 2007-07-27  Andrew Haley  <aph@redhat.com>
1051
1052         * src/arm/sysv.S (ffi_closure_SYSV): Add soft-float.
1053
1054 2007-09-03  Maciej W. Rozycki  <macro@linux-mips.org>
1055
1056         * Makefile.am: Unify MIPS_IRIX and MIPS_LINUX into MIPS.
1057         * configure.ac: Likewise.
1058         * Makefile.in: Regenerate.
1059         * include/Makefile.in: Likewise.
1060         * testsuite/Makefile.in: Likewise.
1061         * configure: Likewise.
1062
1063 2007-08-24  David Daney  <ddaney@avtrex.com>
1064
1065         * testsuite/libffi.call/return_sl.c: New test.
1066
1067 2007-08-10  David Daney  <ddaney@avtrex.com>
1068
1069         * testsuite/libffi.call/cls_multi_ushort.c,
1070         testsuite/libffi.call/cls_align_uint16.c,
1071         testsuite/libffi.call/nested_struct1.c,
1072         testsuite/libffi.call/nested_struct3.c,
1073         testsuite/libffi.call/cls_7_1_byte.c,
1074         testsuite/libffi.call/nested_struct5.c,
1075         testsuite/libffi.call/cls_double.c,
1076         testsuite/libffi.call/nested_struct7.c,
1077         testsuite/libffi.call/cls_sint.c,
1078         testsuite/libffi.call/nested_struct9.c,
1079         testsuite/libffi.call/cls_20byte1.c,
1080         testsuite/libffi.call/cls_multi_sshortchar.c,
1081         testsuite/libffi.call/cls_align_sint64.c,
1082         testsuite/libffi.call/cls_3byte2.c,
1083         testsuite/libffi.call/cls_multi_schar.c,
1084         testsuite/libffi.call/cls_multi_uchar.c,
1085         testsuite/libffi.call/cls_19byte.c,
1086         testsuite/libffi.call/cls_9byte1.c,
1087         testsuite/libffi.call/cls_align_float.c,
1088         testsuite/libffi.call/closure_fn1.c,
1089         testsuite/libffi.call/problem1.c,
1090         testsuite/libffi.call/closure_fn3.c,
1091         testsuite/libffi.call/cls_sshort.c,
1092         testsuite/libffi.call/closure_fn5.c,
1093         testsuite/libffi.call/cls_align_double.c,
1094         testsuite/libffi.call/nested_struct.c,
1095         testsuite/libffi.call/cls_2byte.c,
1096         testsuite/libffi.call/nested_struct10.c,
1097         testsuite/libffi.call/cls_4byte.c,
1098         testsuite/libffi.call/cls_6byte.c,
1099         testsuite/libffi.call/cls_8byte.c,
1100         testsuite/libffi.call/cls_multi_sshort.c,
1101         testsuite/libffi.call/cls_align_sint16.c,
1102         testsuite/libffi.call/cls_align_uint32.c,
1103         testsuite/libffi.call/cls_20byte.c,
1104         testsuite/libffi.call/cls_float.c,
1105         testsuite/libffi.call/nested_struct2.c,
1106         testsuite/libffi.call/cls_5_1_byte.c,
1107         testsuite/libffi.call/nested_struct4.c,
1108         testsuite/libffi.call/cls_24byte.c,
1109         testsuite/libffi.call/nested_struct6.c,
1110         testsuite/libffi.call/cls_64byte.c,
1111         testsuite/libffi.call/nested_struct8.c,
1112         testsuite/libffi.call/cls_uint.c,
1113         testsuite/libffi.call/cls_multi_ushortchar.c,
1114         testsuite/libffi.call/cls_schar.c,
1115         testsuite/libffi.call/cls_uchar.c,
1116         testsuite/libffi.call/cls_align_uint64.c,
1117         testsuite/libffi.call/cls_ulonglong.c,
1118         testsuite/libffi.call/cls_align_longdouble.c,
1119         testsuite/libffi.call/cls_1_1byte.c,
1120         testsuite/libffi.call/cls_12byte.c,
1121         testsuite/libffi.call/cls_3_1byte.c,
1122         testsuite/libffi.call/cls_3byte1.c,
1123         testsuite/libffi.call/cls_4_1byte.c,
1124         testsuite/libffi.call/cls_6_1_byte.c,
1125         testsuite/libffi.call/cls_16byte.c,
1126         testsuite/libffi.call/cls_18byte.c,
1127         testsuite/libffi.call/closure_fn0.c,
1128         testsuite/libffi.call/cls_9byte2.c,
1129         testsuite/libffi.call/closure_fn2.c,
1130         testsuite/libffi.call/closure_fn4.c,
1131         testsuite/libffi.call/cls_ushort.c,
1132         testsuite/libffi.call/closure_fn6.c,
1133         testsuite/libffi.call/cls_5byte.c,
1134         testsuite/libffi.call/cls_align_pointer.c,
1135         testsuite/libffi.call/cls_7byte.c,
1136         testsuite/libffi.call/cls_align_sint32.c,
1137         testsuite/libffi.special/unwindtest_ffi_call.cc,
1138         testsuite/libffi.special/unwindtest.cc: Remove xfail for mips64*-*-*.
1139
1140 2007-08-10  David Daney  <ddaney@avtrex.com>
1141
1142         PR libffi/28313
1143         * configure.ac: Don't treat mips64 as a special case.
1144         * Makefile.am (nodist_libffi_la_SOURCES): Add n32.S.
1145         * configure: Regenerate
1146         * Makefile.in: Ditto.
1147         * fficonfig.h.in: Ditto.
1148         * src/mips/ffitarget.h (REG_L, REG_S, SUBU, ADDU, SRL, LI): Indent.
1149         (LA, EH_FRAME_ALIGN, FDE_ADDR_BYTES): New preprocessor macros.
1150         (FFI_DEFAULT_ABI): Set for n64 case.
1151         (FFI_CLOSURES, FFI_TRAMPOLINE_SIZE): Define for n32 and n64 cases.
1152         * src/mips/n32.S (ffi_call_N32): Add debug macros and labels for FDE.
1153         (ffi_closure_N32): New function.
1154         (.eh_frame): New section
1155         * src/mips/o32.S: Clean up comments.
1156         (ffi_closure_O32): Pass ffi_closure parameter in $12.
1157         * src/mips/ffi.c: Use FFI_MIPS_N32 instead of
1158         _MIPS_SIM == _ABIN32 throughout.
1159         (FFI_MIPS_STOP_HERE): New, use in place of
1160         ffi_stop_here.
1161         (ffi_prep_args): Use unsigned long to hold pointer values.  Rewrite
1162         to support n32/n64 ABIs.
1163         (calc_n32_struct_flags): Rewrite.
1164         (calc_n32_return_struct_flags): Remove unused variable.  Reverse
1165         position of flag bits.
1166         (ffi_prep_cif_machdep): Rewrite n32 portion.
1167         (ffi_call): Enable for n64.  Add special handling for small structure
1168         return values.
1169         (ffi_prep_closure_loc): Add n32 and n64 support.
1170         (ffi_closure_mips_inner_O32): Add cast to silence warning.
1171         (copy_struct_N32, ffi_closure_mips_inner_N32): New functions.
1172
1173 2007-08-08  David Daney  <ddaney@avtrex.com>
1174
1175         * testsuite/libffi.call/ffitest.h (ffi_type_mylong): Remove definition.
1176         * testsuite/libffi.call/cls_align_uint16.c (main): Use correct type
1177         specifiers.
1178         * testsuite/libffi.call/nested_struct1.c (main): Ditto.
1179         * testsuite/libffi.call/cls_sint.c (main): Ditto.
1180         * testsuite/libffi.call/nested_struct9.c (main): Ditto.
1181         * testsuite/libffi.call/cls_20byte1.c (main): Ditto.
1182         * testsuite/libffi.call/cls_9byte1.c (main): Ditto.
1183         * testsuite/libffi.call/closure_fn1.c (main): Ditto.
1184         * testsuite/libffi.call/closure_fn3.c (main): Ditto.
1185         * testsuite/libffi.call/return_dbl2.c (main): Ditto.
1186         * testsuite/libffi.call/cls_sshort.c (main): Ditto.
1187         * testsuite/libffi.call/return_fl3.c (main): Ditto.
1188         * testsuite/libffi.call/closure_fn5.c (main): Ditto.
1189         * testsuite/libffi.call/nested_struct.c (main): Ditto.
1190         * testsuite/libffi.call/nested_struct10.c (main): Ditto.
1191         * testsuite/libffi.call/return_ll1.c (main): Ditto.
1192         * testsuite/libffi.call/cls_8byte.c (main): Ditto.
1193         * testsuite/libffi.call/cls_align_uint32.c (main): Ditto.
1194         * testsuite/libffi.call/cls_align_sint16.c (main): Ditto.
1195         * testsuite/libffi.call/cls_20byte.c (main): Ditto.
1196         * testsuite/libffi.call/nested_struct2.c (main): Ditto.
1197         * testsuite/libffi.call/cls_24byte.c (main): Ditto.
1198         * testsuite/libffi.call/nested_struct6.c (main): Ditto.
1199         * testsuite/libffi.call/cls_uint.c (main): Ditto.
1200         * testsuite/libffi.call/cls_12byte.c (main): Ditto.
1201         * testsuite/libffi.call/cls_16byte.c (main): Ditto.
1202         * testsuite/libffi.call/closure_fn0.c (main): Ditto.
1203         * testsuite/libffi.call/cls_9byte2.c (main): Ditto.
1204         * testsuite/libffi.call/closure_fn2.c (main): Ditto.
1205         * testsuite/libffi.call/return_dbl1.c (main): Ditto.
1206         * testsuite/libffi.call/closure_fn4.c (main): Ditto.
1207         * testsuite/libffi.call/closure_fn6.c (main): Ditto.
1208         * testsuite/libffi.call/cls_align_sint32.c (main): Ditto.
1209
1210 2007-08-07  Andrew Haley  <aph@redhat.com>
1211
1212         * src/x86/sysv.S (ffi_closure_raw_SYSV): Fix typo in previous
1213         checkin.
1214
1215 2007-08-06  Andrew Haley  <aph@redhat.com>
1216
1217         PR testsuite/32843
1218         * src/x86/sysv.S (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1219         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1220         FFI_TYPE_SINT32.
1221
1222 2007-08-02  David Daney  <ddaney@avtrex.com>
1223
1224         * testsuite/libffi.call/return_ul.c (main): Define return type as
1225         ffi_arg.  Use proper printf conversion specifier.
1226
1227 2007-07-30  Andrew Haley  <aph@redhat.com>
1228
1229         PR testsuite/32843
1230         * src/x86/ffi.c (ffi_prep_cif_machdep): in x86 case, add code for
1231         signed/unsigned int8/16.
1232         * src/x86/sysv.S (ffi_call_SYSV): Rewrite to:
1233         Use a jump table.
1234         Remove code to pop args from the stack after call.
1235         Special-case signed/unsigned int8/16.
1236         * testsuite/libffi.call/return_sc.c (main): Revert.
1237
1238 2007-07-26  Richard Guenther  <rguenther@suse.de>
1239
1240         PR testsuite/32843
1241         * testsuite/libffi.call/return_sc.c (main): Verify call
1242         result as signed char, not ffi_arg.
1243
1244 2007-07-16  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1245
1246         * configure.ac (i?86-*-solaris2.1[0-9]): Set TARGET to X86_64.
1247         * configure: Regenerate.
1248
1249 2007-07-11  David Daney  <ddaney@avtrex.com>
1250
1251         * src/mips/ffi.c: Don't include sys/cachectl.h.
1252         (ffi_prep_closure_loc): Use __builtin___clear_cache() instead of
1253         cacheflush().
1254
1255 2007-05-18  Aurelien Jarno  <aurelien@aurel32.net>
1256
1257         * src/arm/ffi.c (ffi_prep_closure_loc): Renamed and ajusted
1258         from (ffi_prep_closure): ... this.
1259         (FFI_INIT_TRAMPOLINE): Adjust.
1260
1261 2005-12-31  Phil Blundell  <pb@reciva.com>
1262
1263         * src/arm/ffi.c (ffi_prep_incoming_args_SYSV,
1264         ffi_closure_SYSV_inner, ffi_prep_closure): New, add closure support.
1265         * src/arm/sysv.S(ffi_closure_SYSV): Likewise.
1266         * src/arm/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1267         (FFI_CLOSURES): Enable closure support.
1268
1269 2007-07-03  Andrew Haley  <aph@hedges.billgatliff.com>
1270
1271         * testsuite/libffi.call/cls_multi_ushort.c,
1272         testsuite/libffi.call/cls_align_uint16.c,
1273         testsuite/libffi.call/nested_struct1.c,
1274         testsuite/libffi.call/nested_struct3.c,
1275         testsuite/libffi.call/cls_7_1_byte.c,
1276         testsuite/libffi.call/cls_double.c,
1277         testsuite/libffi.call/nested_struct5.c,
1278         testsuite/libffi.call/nested_struct7.c,
1279         testsuite/libffi.call/cls_sint.c,
1280         testsuite/libffi.call/nested_struct9.c,
1281         testsuite/libffi.call/cls_20byte1.c,
1282         testsuite/libffi.call/cls_multi_sshortchar.c,
1283         testsuite/libffi.call/cls_align_sint64.c,
1284         testsuite/libffi.call/cls_3byte2.c,
1285         testsuite/libffi.call/cls_multi_schar.c,
1286         testsuite/libffi.call/cls_multi_uchar.c,
1287         testsuite/libffi.call/cls_19byte.c,
1288         testsuite/libffi.call/cls_9byte1.c,
1289         testsuite/libffi.call/cls_align_float.c,
1290         testsuite/libffi.call/closure_fn1.c,
1291         testsuite/libffi.call/problem1.c,
1292         testsuite/libffi.call/closure_fn3.c,
1293         testsuite/libffi.call/cls_sshort.c,
1294         testsuite/libffi.call/closure_fn5.c,
1295         testsuite/libffi.call/cls_align_double.c,
1296         testsuite/libffi.call/cls_2byte.c,
1297         testsuite/libffi.call/nested_struct.c,
1298         testsuite/libffi.call/nested_struct10.c,
1299         testsuite/libffi.call/cls_4byte.c,
1300         testsuite/libffi.call/cls_6byte.c,
1301         testsuite/libffi.call/cls_8byte.c,
1302         testsuite/libffi.call/cls_multi_sshort.c,
1303         testsuite/libffi.call/cls_align_uint32.c,
1304         testsuite/libffi.call/cls_align_sint16.c,
1305         testsuite/libffi.call/cls_float.c,
1306         testsuite/libffi.call/cls_20byte.c,
1307         testsuite/libffi.call/cls_5_1_byte.c,
1308         testsuite/libffi.call/nested_struct2.c,
1309         testsuite/libffi.call/cls_24byte.c,
1310         testsuite/libffi.call/nested_struct4.c,
1311         testsuite/libffi.call/nested_struct6.c,
1312         testsuite/libffi.call/cls_64byte.c,
1313         testsuite/libffi.call/nested_struct8.c,
1314         testsuite/libffi.call/cls_uint.c,
1315         testsuite/libffi.call/cls_multi_ushortchar.c,
1316         testsuite/libffi.call/cls_schar.c,
1317         testsuite/libffi.call/cls_uchar.c,
1318         testsuite/libffi.call/cls_align_uint64.c,
1319         testsuite/libffi.call/cls_ulonglong.c,
1320         testsuite/libffi.call/cls_align_longdouble.c,
1321         testsuite/libffi.call/cls_1_1byte.c,
1322         testsuite/libffi.call/cls_12byte.c,
1323         testsuite/libffi.call/cls_3_1byte.c,
1324         testsuite/libffi.call/cls_3byte1.c,
1325         testsuite/libffi.call/cls_4_1byte.c,
1326         testsuite/libffi.call/cls_6_1_byte.c,
1327         testsuite/libffi.call/cls_16byte.c,
1328         testsuite/libffi.call/cls_18byte.c,
1329         testsuite/libffi.call/closure_fn0.c,
1330         testsuite/libffi.call/cls_9byte2.c,
1331         testsuite/libffi.call/closure_fn2.c,
1332         testsuite/libffi.call/closure_fn4.c,
1333         testsuite/libffi.call/cls_ushort.c,
1334         testsuite/libffi.call/closure_fn6.c,
1335         testsuite/libffi.call/cls_5byte.c,
1336         testsuite/libffi.call/cls_align_pointer.c,
1337         testsuite/libffi.call/cls_7byte.c,
1338         testsuite/libffi.call/cls_align_sint32.c,
1339         testsuite/libffi.special/unwindtest_ffi_call.cc,
1340         testsuite/libffi.special/unwindtest.cc: Enable for ARM.
1341
1342 2007-07-05  H.J. Lu  <hongjiu.lu@intel.com>
1343
1344         * aclocal.m4: Regenerated.
1345
1346 2007-06-02  Paolo Bonzini  <bonzini@gnu.org>
1347
1348         * configure: Regenerate.
1349
1350 2007-05-23  Steve Ellcey  <sje@cup.hp.com>
1351
1352         * Makefile.in: Regenerate.
1353         * configure: Regenerate.
1354         * aclocal.m4: Regenerate.
1355         * include/Makefile.in: Regenerate.
1356         * testsuite/Makefile.in: Regenerate.
1357
1358 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1359
1360         * src/m68k/ffi.c (ffi_prep_incoming_args_SYSV,
1361         ffi_closure_SYSV_inner,ffi_prep_closure): New, add closure support.
1362         * src/m68k/sysv.S(ffi_closure_SYSV,ffi_closure_struct_SYSV): Likewise.
1363         * src/m68k/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1364         (FFI_CLOSURES): Enable closure support.
1365
1366 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1367
1368         * configure.ac (HAVE_AS_CFI_PSEUDO_OP): New test.
1369         * configure: Regenerate.
1370         * fficonfig.h.in: Regenerate.
1371         * src/m68k/sysv.S (CFI_STARTPROC,CFI_ENDPROC,
1372         CFI_OFFSET,CFI_DEF_CFA): New macros.
1373         (ffi_call_SYSV): Add callframe annotation.
1374
1375 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1376
1377         * src/m68k/ffi.c (ffi_prep_args,ffi_prep_cif_machdep): Fix
1378         numerous test suite failures.
1379         * src/m68k/sysv.S (ffi_call_SYSV): Likewise.
1380
1381 2007-04-11  Paolo Bonzini  <bonzini@gnu.org>
1382
1383         * Makefile.am (EXTRA_DIST): Bring up to date.
1384         * Makefile.in: Regenerate.
1385         * src/frv/eabi.S: Remove RCS keyword.
1386
1387 2007-04-06  Richard Henderson  <rth@redhat.com>
1388
1389         * configure.ac: Tidy target case.
1390         (HAVE_LONG_DOUBLE): Allow the target to override.
1391         * configure: Regenerate.
1392         * include/ffi.h.in: Don't define ffi_type_foo if
1393         LIBFFI_HIDE_BASIC_TYPES is defined.
1394         (ffi_type_longdouble): If not HAVE_LONG_DOUBLE, define
1395         to ffi_type_double.
1396         * types.c (LIBFFI_HIDE_BASIC_TYPES): Define.
1397         (FFI_TYPEDEF, ffi_type_void): Mark the data const.
1398         (ffi_type_longdouble): Special case for Alpha.  Don't define
1399         if long double == double.
1400
1401         * src/alpha/ffi.c (FFI_TYPE_LONGDOUBLE): Assert unique value.
1402         (ffi_prep_cif_machdep): Handle it as the 128-bit type.
1403         (ffi_call, ffi_closure_osf_inner): Likewise.
1404         (ffi_closure_osf_inner): Likewise.  Mark hidden.
1405         (ffi_call_osf, ffi_closure_osf): Mark hidden.
1406         * src/alpha/ffitarget.h (FFI_LAST_ABI): Tidy definition.
1407         * src/alpha/osf.S (ffi_call_osf, ffi_closure_osf): Mark hidden.
1408         (load_table): Handle 128-bit long double.
1409
1410         * testsuite/libffi.call/float4.c: Add -mieee for alpha.
1411
1412 2007-04-06  Tom Tromey  <tromey@redhat.com>
1413
1414         PR libffi/31491:
1415         * README: Fixed bug in example.
1416
1417 2007-04-03  Jakub Jelinek  <jakub@redhat.com>
1418
1419         * src/closures.c: Include sys/statfs.h.
1420         (_GNU_SOURCE): Define on Linux.
1421         (FFI_MMAP_EXEC_SELINUX): Define.
1422         (selinux_enabled): New variable.
1423         (selinux_enabled_check): New function.
1424         (is_selinux_enabled): Define.
1425         (dlmmap): Use it.
1426
1427 2007-03-24  Uros Bizjak  <ubizjak@gmail.com>
1428
1429         * testsuite/libffi.call/return_fl2.c (return_fl): Mark as static.
1430         Use 'volatile float sum' to create sum of floats to avoid false
1431         negative due to excess precision on ix86 targets.
1432         (main): Ditto.
1433
1434 2007-03-08  Alexandre Oliva  <aoliva@redhat.com>
1435
1436         * src/powerpc/ffi.c (flush_icache): Fix left-over from previous
1437         patch.
1438         (ffi_prep_closure_loc): Remove unneeded casts.  Add needed ones.
1439
1440 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1441
1442         * include/ffi.h.in (ffi_closure_alloc, ffi_closure_free): New.
1443         (ffi_prep_closure_loc): New.
1444         (ffi_prep_raw_closure_loc): New.
1445         (ffi_prep_java_raw_closure_loc): New.
1446         * src/closures.c: New file.
1447         * src/dlmalloc.c [FFI_MMAP_EXEC_WRIT] (struct malloc_segment):
1448         Replace sflags with exec_offset.
1449         [FFI_MMAP_EXEC_WRIT] (mmap_exec_offset, add_segment_exec_offset,
1450         sub_segment_exec_offset): New macros.
1451         (get_segment_flags, set_segment_flags, check_segment_merge): New
1452         macros.
1453         (is_mmapped_segment, is_extern_segment): Use get_segment_flags.
1454         (add_segment, sys_alloc, create_mspace, create_mspace_with_base,
1455         destroy_mspace): Use new macros.
1456         (sys_alloc): Silence warning.
1457         * Makefile.am (libffi_la_SOURCES): Add src/closures.c.
1458         * Makefile.in: Rebuilt.
1459         * src/prep_cif [FFI_CLOSURES] (ffi_prep_closure): Implement in
1460         terms of ffi_prep_closure_loc.
1461         * src/raw_api.c (ffi_prep_raw_closure_loc): Renamed and adjusted
1462         from...
1463         (ffi_prep_raw_closure): ... this.  Re-implement in terms of the
1464         renamed version.
1465         * src/java_raw_api (ffi_prep_java_raw_closure_loc): Renamed and
1466         adjusted from...
1467         (ffi_prep_java_raw_closure): ... this.  Re-implement in terms of
1468         the renamed version.
1469         * src/alpha/ffi.c (ffi_prep_closure_loc): Renamed from
1470         (ffi_prep_closure): ... this.
1471         * src/pa/ffi.c: Likewise.
1472         * src/cris/ffi.c: Likewise.  Adjust.
1473         * src/frv/ffi.c: Likewise.
1474         * src/ia64/ffi.c: Likewise.
1475         * src/mips/ffi.c: Likewise.
1476         * src/powerpc/ffi_darwin.c: Likewise.
1477         * src/s390/ffi.c: Likewise.
1478         * src/sh/ffi.c: Likewise.
1479         * src/sh64/ffi.c: Likewise.
1480         * src/sparc/ffi.c: Likewise.
1481         * src/x86/ffi64.c: Likewise.
1482         * src/x86/ffi.c: Likewise.
1483         (FFI_INIT_TRAMPOLINE): Adjust.
1484         (ffi_prep_raw_closure_loc): Renamed and adjusted from...
1485         (ffi_prep_raw_closure): ... this.
1486         * src/powerpc/ffi.c (ffi_prep_closure_loc): Renamed from
1487         (ffi_prep_closure): ... this.
1488         (flush_icache): Adjust.
1489
1490 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1491
1492         * src/dlmalloc.c: New file, imported version 2.8.3 of Doug
1493         Lea's malloc.
1494
1495 2007-03-01  Brooks Moses  <brooks.moses@codesourcery.com>
1496
1497         * Makefile.am: Add dummy install-pdf target.
1498         * Makefile.in: Regenerate
1499
1500 2007-02-13  Andreas Krebbel  <krebbel1@de.ibm.com>
1501
1502         * src/s390/ffi.c (ffi_prep_args, ffi_prep_cif_machdep,
1503         ffi_closure_helper_SYSV): Add long double handling.
1504
1505 2007-02-02  Jakub Jelinek  <jakub@redhat.com>
1506
1507         * src/powerpc/linux64.S (ffi_call_LINUX64): Move restore of r2
1508         immediately after bctrl instruction.
1509
1510 2007-01-18  Alexandre Oliva  <aoliva@redhat.com>
1511
1512         * Makefile.am (all-recursive, install-recursive,
1513         mostlyclean-recursive, clean-recursive, distclean-recursive,
1514         maintainer-clean-recursive): Add missing targets.
1515         * Makefile.in: Rebuilt.
1516
1517 2006-12-14  Andreas Tobler  <a.tobler@schweiz.org>
1518
1519         * configure.ac: Add TARGET for x86_64-*-darwin*.
1520         * Makefile.am (nodist_libffi_la_SOURCES): Add rules for 64-bit sources
1521         for X86_DARWIN.
1522         * src/x86/ffitarget.h: Set trampoline size for x86_64-*-darwin*.
1523         * src/x86/darwin64.S: New file for x86_64-*-darwin* support.
1524         * configure: Regenerate.
1525         * Makefile.in: Regenerate.
1526         * include/Makefile.in: Regenerate.
1527         * testsuite/Makefile.in: Regenerate.
1528         * testsuite/libffi.special/unwindtest_ffi_call.cc: New test case for
1529         ffi_call only.
1530
1531 2006-12-13  Andreas Tobler <a.tobler@schweiz.org>
1532
1533         * aclocal.m4: Regenerate with aclocal -I .. as written in the
1534         Makefile.am.
1535
1536 2006-10-31  Geoffrey Keating  <geoffk@apple.com>
1537
1538         * src/powerpc/ffi_darwin.c (darwin_adjust_aggregate_sizes): New.
1539         (ffi_prep_cif_machdep): Call darwin_adjust_aggregate_sizes for
1540         Darwin.
1541         * testsuite/libffi.call/nested_struct4.c: Remove Darwin XFAIL.
1542         * testsuite/libffi.call/nested_struct6.c: Remove Darwin XFAIL.
1543
1544 2006-10-10  Paolo Bonzini  <bonzini@gnu.org>
1545             Sandro Tolaini  <tolaini@libero.it>
1546
1547         * configure.ac [i*86-*-darwin*]: Set X86_DARWIN symbol and
1548         conditional.
1549         * configure: Regenerated.
1550         * Makefile.am (nodist_libffi_la_SOURCES) [X86_DARWIN]: New case.
1551         (EXTRA_DIST): Add src/x86/darwin.S.
1552         * Makefile.in: Regenerated.
1553         * include/Makefile.in: Regenerated.
1554         * testsuite/Makefile.in: Regenerated.
1555
1556         * src/x86/ffi.c (ffi_prep_cif_machdep) [X86_DARWIN]: Treat like
1557         X86_WIN32, and additionally align stack to 16 bytes.
1558         * src/x86/darwin.S: New, based on sysv.S.
1559         * src/prep_cif.c (ffi_prep_cif) [X86_DARWIN]: Align > 8-byte structs.
1560
1561 2006-09-12  David Daney  <ddaney@avtrex.com>
1562
1563         PR libffi/23935
1564         * include/Makefile.am: Install both ffi.h and ffitarget.h in
1565         $(libdir)/gcc/$(target_alias)/$(gcc_version)/include.
1566         * aclocal.m4: Regenerated for automake 1.9.6.
1567         * Makefile.in: Regenerated.
1568         * include/Makefile.in: Regenerated.
1569         * testsuite/Makefile.in: Regenerated.
1570
1571 2006-08-17  Andreas Tobler  <a.tobler@schweiz.ch>
1572
1573         * include/ffi_common.h (struct): Revert accidental commit.
1574
1575 2006-08-15  Andreas Tobler  <a.tobler@schweiz.ch>
1576
1577         * include/ffi_common.h: Remove lint directives.
1578         * include/ffi.h.in: Likewise.
1579
1580 2006-07-25  Torsten Schoenfeld  <kaffeetisch@gmx.de>
1581
1582         * include/ffi.h.in (ffi_type_ulong, ffi_type_slong): Define correctly
1583         for 32-bit architectures.
1584         * testsuite/libffi.call/return_ul.c: New test case.
1585
1586 2006-07-19  David Daney  <ddaney@avtrex.com>
1587
1588         * testsuite/libffi.call/closure_fn6.c: Remove xfail for mips,
1589         xfail remains for mips64.
1590
1591 2006-05-23  Carlos O'Donell  <carlos@codesourcery.com>
1592
1593         * Makefile.am: Add install-html target. Add install-html to .PHONY
1594         * Makefile.in: Regenerate.
1595         * aclocal.m4: Regenerate.
1596         * include/Makefile.in: Regenerate.
1597         * testsuite/Makefile.in: Regenerate.
1598
1599 2006-05-18  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1600
1601         * pa/ffi.c (ffi_prep_args_pa32): Load floating point arguments from
1602         stack slot.
1603
1604 2006-04-22  Andreas Tobler  <a.tobler@schweiz.ch>
1605
1606         * README: Remove notice about 'Crazy Comments'.
1607         * src/debug.c: Remove lint directives. Cleanup white spaces.
1608         * src/java_raw_api.c: Likewise.
1609         * src/prep_cif.c: Likewise.
1610         * src/raw_api.c: Likewise.
1611         * src/ffitest.c: Delete. No longer needed, all test cases migrated
1612         to the testsuite.
1613         * src/arm/ffi.c: Remove lint directives.
1614         * src/m32r/ffi.c: Likewise.
1615         * src/pa/ffi.c: Likewise.
1616         * src/powerpc/ffi.c: Likewise.
1617         * src/powerpc/ffi_darwin.c: Likewise.
1618         * src/sh/ffi.c: Likewise.
1619         * src/sh64/ffi.c: Likewise.
1620         * src/x86/ffi.c: Likewise.
1621         * testsuite/libffi.call/float2.c: Likewise.
1622         * testsuite/libffi.call/promotion.c: Likewise.
1623         * testsuite/libffi.call/struct1.c: Likewise.
1624
1625 2006-04-13  Andreas Tobler  <a.tobler@schweiz.ch>
1626
1627         * src/pa/hpux32.S: Correct unwind offset calculation for
1628         ffi_closure_pa32.
1629         * src/pa/linux.S: Likewise.
1630
1631 2006-04-12  James E Wilson  <wilson@specifix.com>
1632
1633         PR libgcj/26483
1634         * src/ia64/ffi.c (stf_spill, ldf_fill): Rewrite as macros.
1635         (hfa_type_load): Call stf_spill.
1636         (hfa_type_store): Call ldf_fill.
1637         (ffi_call): Adjust calls to above routines.  Add local temps for
1638         macro result.
1639
1640 2006-04-10  Matthias Klose  <doko@debian.org>
1641
1642         * testsuite/lib/libffi-dg.exp (libffi-init): Recognize multilib
1643         directory names containing underscores.
1644
1645 2006-04-07  James E Wilson  <wilson@specifix.com>
1646
1647         * testsuite/libffi.call/float4.c: New testcase.
1648
1649 2006-04-05  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1650             Andreas Tobler  <a.tobler@schweiz.ch>
1651
1652         * Makefile.am: Add PA_HPUX port.
1653         * Makefile.in: Regenerate.
1654         * include/Makefile.in: Likewise.
1655         * testsuite/Makefile.in: Likewise.
1656         * configure.ac: Add PA_HPUX rules.
1657         * configure: Regenerate.
1658         * src/pa/ffitarget.h: Rename linux target to PA_LINUX.
1659         Add PA_HPUX and PA64_HPUX.
1660         Rename FFI_LINUX ABI to FFI_PA32 ABI.
1661         (FFI_TRAMPOLINE_SIZE): Define for 32-bit HP-UX targets.
1662         (FFI_TYPE_SMALL_STRUCT2): Define.
1663         (FFI_TYPE_SMALL_STRUCT4): Likewise.
1664         (FFI_TYPE_SMALL_STRUCT8): Likewise.
1665         (FFI_TYPE_SMALL_STRUCT3): Redefine.
1666         (FFI_TYPE_SMALL_STRUCT5): Likewise.
1667         (FFI_TYPE_SMALL_STRUCT6): Likewise.
1668         (FFI_TYPE_SMALL_STRUCT7): Likewise.
1669         * src/pa/ffi.c (ROUND_DOWN): Delete.
1670         (fldw, fstw, fldd, fstd): Use '__asm__'.
1671         (ffi_struct_type): Add support for FFI_TYPE_SMALL_STRUCT2,
1672         FFI_TYPE_SMALL_STRUCT4 and FFI_TYPE_SMALL_STRUCT8.
1673         (ffi_prep_args_LINUX): Rename to ffi_prep_args_pa32. Update comment.
1674         Simplify incrementing of stack slot variable. Change type of local
1675         'n' to unsigned int.
1676         (ffi_size_stack_LINUX): Rename to ffi_size_stack_pa32. Handle long
1677         double on PA_HPUX.
1678         (ffi_prep_cif_machdep): Likewise.
1679         (ffi_call): Likewise.
1680         (ffi_closure_inner_LINUX): Rename to ffi_closure_inner_pa32. Change
1681         return type to ffi_status. Simplify incrementing of stack slot
1682         variable. Only copy floating point argument registers when PA_LINUX
1683         is true. Reformat debug statement.
1684         Add support for FFI_TYPE_SMALL_STRUCT2, FFI_TYPE_SMALL_STRUCT4 and
1685         FFI_TYPE_SMALL_STRUCT8.
1686         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Add 'extern' to
1687         declaration.
1688         (ffi_prep_closure): Make linux trampoline conditional on PA_LINUX.
1689         Add nops to cache flush.  Add trampoline for PA_HPUX.
1690         * src/pa/hpux32.S: New file.
1691         * src/pa/linux.S (ffi_call_LINUX): Rename to ffi_call_pa32. Rename
1692         ffi_prep_args_LINUX to ffi_prep_args_pa32.
1693         Localize labels. Add support for 2, 4 and 8-byte small structs. Handle
1694         unaligned destinations in 3, 5, 6 and 7-byte small structs. Order
1695         argument type checks so that common argument types appear first.
1696         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Rename
1697         ffi_closure_inner_LINUX to ffi_closure_inner_pa32.
1698
1699 2006-03-24  Alan Modra  <amodra@bigpond.net.au>
1700
1701         * src/powerpc/ffitarget.h (enum ffi_abi): Add FFI_LINUX.  Default
1702         for 32-bit using IBM extended double format.  Fix FFI_LAST_ABI.
1703         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Handle linux variant of
1704         FFI_TYPE_LONGDOUBLE.
1705         (ffi_prep_args64): Assert using IBM extended double.
1706         (ffi_prep_cif_machdep): Don't munge FFI_TYPE_LONGDOUBLE type.
1707         Handle FFI_LINUX FFI_TYPE_LONGDOUBLE return and args.
1708         (ffi_call): Handle FFI_LINUX.
1709         (ffi_closure_helper_SYSV): Non FFI_LINUX long double return needs
1710         gpr3 return pointer as for struct return.  Handle FFI_LINUX
1711         FFI_TYPE_LONGDOUBLE return and args.  Don't increment "nf"
1712         unnecessarily.
1713         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Load both f1 and f2
1714         for FFI_TYPE_LONGDOUBLE.  Move epilogue insns into case table.
1715         Don't use r6 as pointer to results, instead use sp offset.  Don't
1716         make a special call to load lr with case table address, instead
1717         use offset from previous call.
1718         * src/powerpc/sysv.S (ffi_call_SYSV): Save long double return.
1719         * src/powerpc/linux64.S (ffi_call_LINUX64): Simplify long double
1720         return.
1721
1722 2006-03-15  Kaz Kojima  <kkojima@gcc.gnu.org>
1723
1724         * src/sh64/ffi.c (ffi_prep_cif_machdep): Handle float arguments
1725         passed with FP registers correctly.
1726         (ffi_closure_helper_SYSV): Likewise.
1727         * src/sh64/sysv.S: Likewise.
1728
1729 2006-03-01  Andreas Tobler  <a.tobler@schweiz.ch>
1730
1731         * testsuite/libffi.special/unwindtest.cc (closure_test_fn): Mark cif,
1732         args and userdata unused.
1733         (closure_test_fn1): Mark cif and userdata unused.
1734         (main): Remove unused res.
1735
1736 2006-02-28  Andreas Tobler  <a.tobler@schweiz.ch>
1737
1738         * testsuite/libffi.call/call.exp: Adjust FSF address. Add test runs for
1739         -O2, -O3, -Os and the warning flags -W -Wall.
1740         * testsuite/libffi.special/special.exp: Likewise.
1741         * testsuite/libffi.call/ffitest.h: Add an __UNUSED__ macro to mark
1742         unused parameter unused for gcc or else do nothing.
1743         * testsuite/libffi.special/ffitestcxx.h: Likewise.
1744         * testsuite/libffi.call/cls_12byte.c (cls_struct_12byte_gn): Mark cif
1745         and userdata unused.
1746         * testsuite/libffi.call/cls_16byte.c (cls_struct_16byte_gn): Likewise.
1747         * testsuite/libffi.call/cls_18byte.c (cls_struct_18byte_gn): Likewise.
1748         * testsuite/libffi.call/cls_19byte.c (cls_struct_19byte_gn): Likewise.
1749         * testsuite/libffi.call/cls_1_1byte.c (cls_struct_1_1byte_gn): Likewise.
1750         * testsuite/libffi.call/cls_20byte.c (cls_struct_20byte_gn): Likewise.
1751         * testsuite/libffi.call/cls_20byte1.c (cls_struct_20byte_gn): Likewise.
1752         * testsuite/libffi.call/cls_24byte.c (cls_struct_24byte_gn): Likewise.
1753         * testsuite/libffi.call/cls_2byte.c (cls_struct_2byte_gn): Likewise.
1754         * testsuite/libffi.call/cls_3_1byte.c (cls_struct_3_1byte_gn): Likewise.
1755         * testsuite/libffi.call/cls_3byte1.c (cls_struct_3byte_gn): Likewise.
1756         * testsuite/libffi.call/cls_3byte2.c (cls_struct_3byte_gn1): Likewise.
1757         * testsuite/libffi.call/cls_4_1byte.c (cls_struct_4_1byte_gn): Likewise.
1758         * testsuite/libffi.call/cls_4byte.c (cls_struct_4byte_gn): Likewise.
1759         * testsuite/libffi.call/cls_5_1_byte.c (cls_struct_5byte_gn): Likewise.
1760         * testsuite/libffi.call/cls_5byte.c (cls_struct_5byte_gn): Likewise.
1761         * testsuite/libffi.call/cls_64byte.c (cls_struct_64byte_gn): Likewise.
1762         * testsuite/libffi.call/cls_6_1_byte.c (cls_struct_6byte_gn): Likewise.
1763         * testsuite/libffi.call/cls_6byte.c (cls_struct_6byte_gn): Likewise.
1764         * testsuite/libffi.call/cls_7_1_byte.c (cls_struct_7byte_gn): Likewise.
1765         * testsuite/libffi.call/cls_7byte.c (cls_struct_7byte_gn): Likewise.
1766         * testsuite/libffi.call/cls_8byte.c (cls_struct_8byte_gn): Likewise.
1767         * testsuite/libffi.call/cls_9byte1.c (cls_struct_9byte_gn): Likewise.
1768         * testsuite/libffi.call/cls_9byte2.c (cls_struct_9byte_gn): Likewise.
1769         * testsuite/libffi.call/cls_align_double.c (cls_struct_align_gn):
1770         Likewise.
1771         * testsuite/libffi.call/cls_align_float.c (cls_struct_align_gn):
1772         Likewise.
1773         * testsuite/libffi.call/cls_align_longdouble.c (cls_struct_align_gn):
1774         Likewise.
1775         * testsuite/libffi.call/cls_align_pointer.c (cls_struct_align_fn): Cast
1776         void* to avoid compiler warning.
1777         (main): Likewise.
1778         (cls_struct_align_gn): Mark cif and userdata unused.
1779         * testsuite/libffi.call/cls_align_sint16.c (cls_struct_align_gn):
1780         Likewise.
1781         * testsuite/libffi.call/cls_align_sint32.c (cls_struct_align_gn):
1782         Likewise.
1783         * testsuite/libffi.call/cls_align_sint64.c (cls_struct_align_gn):
1784         Likewise.
1785         * testsuite/libffi.call/cls_align_uint16.c (cls_struct_align_gn):
1786         Likewise.
1787         * testsuite/libffi.call/cls_align_uint32.c (cls_struct_align_gn):
1788         Likewise.
1789         * testsuite/libffi.call/cls_double.c (cls_ret_double_fn): Likewise.
1790         * testsuite/libffi.call/cls_float.c (cls_ret_float_fn): Likewise.
1791         * testsuite/libffi.call/cls_multi_schar.c (test_func_gn): Mark cif and
1792         data unused.
1793         (main): Cast res_call to silence gcc.
1794         * testsuite/libffi.call/cls_multi_sshort.c (test_func_gn): Mark cif and
1795         data unused.
1796         (main): Cast res_call to silence gcc.
1797         * testsuite/libffi.call/cls_multi_sshortchar.c (test_func_gn): Mark cif
1798         and data unused.
1799         (main): Cast res_call to silence gcc.
1800         * testsuite/libffi.call/cls_multi_uchar.c (test_func_gn): Mark cif and
1801         data unused.
1802         (main): Cast res_call to silence gcc.
1803         * testsuite/libffi.call/cls_multi_ushort.c (test_func_gn): Mark cif and
1804         data unused.
1805         (main): Cast res_call to silence gcc.
1806         * testsuite/libffi.call/cls_multi_ushortchar.c (test_func_gn): Mark cif
1807         and data unused.
1808         (main): Cast res_call to silence gcc.
1809         * testsuite/libffi.call/cls_schar.c (cls_ret_schar_fn): Mark cif and
1810         userdata unused.
1811         (cls_ret_schar_fn): Cast printf parameter to silence gcc.
1812         * testsuite/libffi.call/cls_sint.c (cls_ret_sint_fn): Mark cif and
1813         userdata unused.
1814         (cls_ret_sint_fn): Cast printf parameter to silence gcc.
1815         * testsuite/libffi.call/cls_sshort.c (cls_ret_sshort_fn): Mark cif and
1816         userdata unused.
1817         (cls_ret_sshort_fn): Cast printf parameter to silence gcc.
1818         * testsuite/libffi.call/cls_uchar.c (cls_ret_uchar_fn):  Mark cif and
1819         userdata unused.
1820         (cls_ret_uchar_fn): Cast printf parameter to silence gcc.
1821         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Mark cif and
1822         userdata unused.
1823         (cls_ret_uint_fn): Cast printf parameter to silence gcc.
1824         * testsuite/libffi.call/cls_ulonglong.c (cls_ret_ulonglong_fn): Mark cif
1825         and userdata unused.
1826         * testsuite/libffi.call/cls_ushort.c (cls_ret_ushort_fn): Mark cif and
1827         userdata unused.
1828         (cls_ret_ushort_fn): Cast printf parameter to silence gcc.
1829         * testsuite/libffi.call/float.c (floating): Remove unused parameter e.
1830         * testsuite/libffi.call/float1.c (main): Remove unused variable i.
1831         Cleanup white spaces.
1832         * testsuite/libffi.call/negint.c (checking): Remove unused variable i.
1833         * testsuite/libffi.call/nested_struct.c (cls_struct_combined_gn): Mark
1834         cif and userdata unused.
1835         * testsuite/libffi.call/nested_struct1.c (cls_struct_combined_gn):
1836         Likewise.
1837         * testsuite/libffi.call/nested_struct10.c (B_gn): Likewise.
1838         * testsuite/libffi.call/nested_struct2.c (B_fn): Adjust printf
1839         formatters to silence gcc.
1840         (B_gn): Mark cif and userdata unused.
1841         * testsuite/libffi.call/nested_struct3.c (B_gn): Mark cif and userdata
1842         unused.
1843         * testsuite/libffi.call/nested_struct4.c: Mention related PR.
1844         (B_gn): Mark cif and userdata unused.
1845         * testsuite/libffi.call/nested_struct5.c (B_gn): Mark cif and userdata
1846         unused.
1847         * testsuite/libffi.call/nested_struct6.c: Mention related PR.
1848         (B_gn): Mark cif and userdata unused.
1849         * testsuite/libffi.call/nested_struct7.c (B_gn): Mark cif and userdata
1850         unused.
1851         * testsuite/libffi.call/nested_struct8.c (B_gn): Likewise.
1852         * testsuite/libffi.call/nested_struct9.c (B_gn): Likewise.
1853         * testsuite/libffi.call/problem1.c (stub): Likewise.
1854         * testsuite/libffi.call/pyobjc-tc.c (main): Cast the result to silence
1855         gcc.
1856         * testsuite/libffi.call/return_fl2.c (return_fl): Add the note mentioned
1857         in the last commit for this test case in the test case itself.
1858         * testsuite/libffi.call/closure_fn0.c (closure_test_fn0): Mark cif as
1859         unused.
1860         * testsuite/libffi.call/closure_fn1.c (closure_test_fn1): Likewise.
1861         * testsuite/libffi.call/closure_fn2.c (closure_test_fn2): Likewise.
1862         * testsuite/libffi.call/closure_fn3.c (closure_test_fn3): Likewise.
1863         * testsuite/libffi.call/closure_fn4.c (closure_test_fn0): Likewise.
1864         * testsuite/libffi.call/closure_fn5.c (closure_test_fn5): Likewise.
1865         * testsuite/libffi.call/closure_fn6.c (closure_test_fn0): Likewise.
1866
1867 2006-02-22  Kaz Kojima  <kkojima@gcc.gnu.org>
1868
1869         * src/sh/sysv.S: Fix register numbers in the FDE for
1870         ffi_closure_SYSV.
1871
1872 2006-02-20  Andreas Tobler  <a.tobler@schweiz.ch>
1873
1874         * testsuite/libffi.call/return_fl2.c (return_fl): Remove static
1875         declaration to avoid a false negative on ix86. See PR323.
1876
1877 2006-02-18  Kaz Kojima  <kkojima@gcc.gnu.org>
1878
1879         * src/sh/ffi.c (ffi_closure_helper_SYSV): Remove unused variable
1880         and cast integer to void * if needed.  Update the pointer to
1881         the FP register saved area correctly.
1882
1883 2006-02-17  Andreas Tobler  <a.tobler@schweiz.ch>
1884
1885         * testsuite/libffi.call/nested_struct6.c: XFAIL this test until PR25630
1886         is fixed.
1887         * testsuite/libffi.call/nested_struct4.c: Likewise.
1888
1889 2006-02-16  Andreas Tobler  <a.tobler@schweiz.ch>
1890
1891         * testsuite/libffi.call/return_dbl.c: New test case.
1892         * testsuite/libffi.call/return_dbl1.c: Likewise.
1893         * testsuite/libffi.call/return_dbl2.c: Likewise.
1894         * testsuite/libffi.call/return_fl.c: Likewise.
1895         * testsuite/libffi.call/return_fl1.c: Likewise.
1896         * testsuite/libffi.call/return_fl2.c: Likewise.
1897         * testsuite/libffi.call/return_fl3.c: Likewise.
1898         * testsuite/libffi.call/closure_fn6.c: Likewise.
1899
1900         * testsuite/libffi.call/nested_struct2.c: Remove ffi_type_mylong
1901         definition.
1902         * testsuite/libffi.call/ffitest.h: Add ffi_type_mylong definition
1903         here to be used by other test cases too.
1904
1905         * testsuite/libffi.call/nested_struct10.c: New test case.
1906         * testsuite/libffi.call/nested_struct9.c: Likewise.
1907         * testsuite/libffi.call/nested_struct8.c: Likewise.
1908         * testsuite/libffi.call/nested_struct7.c: Likewise.
1909         * testsuite/libffi.call/nested_struct6.c: Likewise.
1910         * testsuite/libffi.call/nested_struct5.c: Likewise.
1911         * testsuite/libffi.call/nested_struct4.c: Likewise.
1912
1913 2006-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
1914
1915         * configure.ac: Enable libffi for sparc64-*-freebsd*.
1916         * configure: Rebuilt.
1917
1918 2006-01-18  Jakub Jelinek  <jakub@redhat.com>
1919
1920         * src/powerpc/sysv.S (smst_two_register): Don't call __ashldi3,
1921         instead do the shifting inline.
1922         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't compute %r5
1923         shift count unconditionally.  Simplify load sequences for 1, 2, 3, 4
1924         and 8 byte structs, for the remaining struct sizes don't call
1925         __lshrdi3, instead do the shifting inline.
1926
1927 2005-12-07  Thiemo Seufer  <ths@networkno.de>
1928
1929         * src/mips/ffitarget.h: Remove obsolete sgidefs.h include. Add
1930         missing parentheses.
1931         * src/mips/o32.S (ffi_call_O32): Code formatting. Define
1932         and use A3_OFF, FP_OFF, RA_OFF. Micro-optimizations.
1933         (ffi_closure_O32): Likewise, but with newly defined A3_OFF2,
1934         A2_OFF2, A1_OFF2, A0_OFF2, RA_OFF2, FP_OFF2, S0_OFF2, GP_OFF2,
1935         V1_OFF2, V0_OFF2, FA_1_1_OFF2, FA_1_0_OFF2, FA_0_1_OFF2,
1936         FA_0_0_OFF2.
1937         * src/mips/ffi.c (ffi_prep_args): Code formatting. Fix
1938         endianness bugs.
1939         (ffi_prep_closure): Improve trampoline instruction scheduling.
1940         (ffi_closure_mips_inner_O32): Fix endianness bugs.
1941
1942 2005-12-03  Alan Modra  <amodra@bigpond.net.au>
1943
1944         * src/powerpc/ffi.c: Formatting.
1945         (ffi_prep_args_SYSV): Avoid possible aliasing problems by using unions.
1946         (ffi_prep_args64): Likewise.
1947
1948 2005-09-30  Geoffrey Keating  <geoffk@apple.com>
1949
1950         * testsuite/lib/libffi-dg.exp (libffi_target_compile): For
1951         darwin, use -shared-libgcc not -lgcc_s, and explain why.
1952
1953 2005-09-26  Tom Tromey  <tromey@redhat.com>
1954
1955         * testsuite/libffi.call/float1.c (value_type): New typedef.
1956         (CANARY): New define.
1957         (main): Check for result buffer overflow.
1958         * src/powerpc/linux64.S: Handle linux64 long double returns.
1959         * src/powerpc/ffi.c (FLAG_RETURNS_128BITS): New constant.
1960         (ffi_prep_cif_machdep): Handle linux64 long double returns.
1961
1962 2005-08-25  Alan Modra  <amodra@bigpond.net.au>
1963
1964         PR target/23404
1965         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Correct placement of stack
1966         homed fp args.
1967         (ffi_status ffi_prep_cif_machdep): Correct stack sizing for same.
1968
1969 2005-08-11  Jakub Jelinek  <jakub@redhat.com>
1970
1971         * configure.ac (HAVE_HIDDEN_VISIBILITY_ATTRIBUTE): New test.
1972         (AH_BOTTOM): Add FFI_HIDDEN definition.
1973         * configure: Rebuilt.
1974         * fficonfig.h.in: Rebuilt.
1975         * src/powerpc/ffi.c (hidden): Remove.
1976         (ffi_closure_LINUX64, ffi_prep_args64, ffi_call_LINUX64,
1977         ffi_closure_helper_LINUX64): Use FFI_HIDDEN instead of hidden.
1978         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64,
1979         .ffi_closure_LINUX64): Use FFI_HIDDEN instead of .hidden.
1980         * src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV): Remove,
1981         add FFI_HIDDEN to its prototype.
1982         (ffi_closure_SYSV_inner): New.
1983         * src/x86/sysv.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
1984         * src/x86/win32.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
1985
1986 2005-08-10  Alfred M. Szmidt  <ams@gnu.org>
1987
1988         PR libffi/21819:
1989         * configure: Rebuilt.
1990         * configure.ac: Handle i*86-*-gnu*.
1991
1992 2005-08-09  Jakub Jelinek  <jakub@redhat.com>
1993
1994         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Use
1995         DW_CFA_offset_extended_sf rather than
1996         DW_CFA_GNU_negative_offset_extended.
1997         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise.
1998
1999 2005-07-22  SUGIOKA Toshinobu  <sugioka@itonet.co.jp>
2000
2001         * src/sh/sysv.S (ffi_call_SYSV): Stop argument popping correctly
2002         on sh3.
2003         (ffi_closure_SYSV): Change the stack layout for sh3 struct argument.
2004         * src/sh/ffi.c (ffi_prep_args): Fix sh3 argument copy, when it is
2005         partially on register.
2006         (ffi_closure_helper_SYSV): Likewise.
2007         (ffi_prep_cif_machdep): Don't set too many cif->flags.
2008
2009 2005-07-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2010
2011         * src/sh/ffi.c (ffi_call): Handle small structures correctly.
2012         Remove empty line.
2013         * src/sh64/ffi.c (simple_type): Remove.
2014         (return_type): Handle small structures correctly.
2015         (ffi_prep_args): Likewise.
2016         (ffi_call): Likewise.
2017         (ffi_closure_helper_SYSV): Likewise.
2018         * src/sh64/sysv.S (ffi_call_SYSV): Handle 1, 2 and 4-byte return.
2019         Emit position independent code if PIC and remove wrong datalabel
2020         prefixes from EH data.
2021
2022 2005-07-19  Andreas Tobler  <a.tobler@schweiz.ch>
2023
2024         * Makefile.am (nodist_libffi_la_SOURCES): Add POWERPC_FREEBSD.
2025         * Makefile.in: Regenerate.
2026         * include/Makefile.in: Likewise.
2027         * testsuite/Makefile.in: Likewise.
2028         * configure.ac: Add POWERPC_FREEBSD rules.
2029         * configure: Regenerate.
2030         * src/powerpc/ffitarget.h: Add POWERPC_FREEBSD rules.
2031         (FFI_SYSV_TYPE_SMALL_STRUCT): Define.
2032         * src/powerpc/ffi.c: Add flags to handle small structure returns
2033         in ffi_call_SYSV.
2034         (ffi_prep_cif_machdep): Handle small structures for SYSV 4 ABI.
2035         Aka FFI_SYSV.
2036         (ffi_closure_helper_SYSV): Likewise.
2037         * src/powerpc/ppc_closure.S: Add return types for small structures.
2038         * src/powerpc/sysv.S: Add bits to handle small structures for
2039         final SYSV 4 ABI.
2040
2041 2005-07-10  Andreas Tobler  <a.tobler@schweiz.ch>
2042
2043         * testsuite/libffi.call/cls_5_1_byte.c: New test file.
2044         * testsuite/libffi.call/cls_6_1_byte.c: Likewise.
2045         * testsuite/libffi.call/cls_7_1_byte.c: Likewise.
2046
2047 2005-07-05  Randolph Chung  <tausq@debian.org>
2048
2049         * src/pa/ffi.c (ffi_struct_type): Rename FFI_TYPE_SMALL_STRUCT1
2050         as FFI_TYPE_SMALL_STRUCT3.  Break out handling for 5-7 byte
2051         structures.  Kill compilation warnings.
2052         (ffi_closure_inner_LINUX): Print return values as hex in debug
2053         message.  Rename FFI_TYPE_SMALL_STRUCT1 as FFI_TYPE_SMALL_STRUCT3.
2054         Properly handle 5-7 byte structure returns.
2055         * src/pa/ffitarget.h (FFI_TYPE_SMALL_STRUCT1)
2056         (FFI_TYPE_SMALL_STRUCT2): Remove.
2057         (FFI_TYPE_SMALL_STRUCT3, FFI_TYPE_SMALL_STRUCT5)
2058         (FFI_TYPE_SMALL_STRUCT6, FFI_TYPE_SMALL_STRUCT7): Define.
2059         * src/pa/linux.S: Mark source file as using PA1.1 assembly.
2060         (checksmst1, checksmst2): Remove.
2061         (checksmst3): Optimize handling of 3-byte struct returns.
2062         (checksmst567): Properly handle 5-7 byte struct returns.
2063
2064 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2065
2066         PR libgcj/21943
2067         * src/mips/n32.S: Enforce PIC code.
2068         * src/mips/o32.S: Likewise.
2069
2070 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2071
2072         * configure.ac: Treat i*86-*-solaris2.10 and up as X86_64.
2073         * configure: Regenerate.
2074
2075 2005-06-01  Alan Modra  <amodra@bigpond.net.au>
2076
2077         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't use JUMPTARGET
2078         to call ffi_closure_helper_SYSV.  Append @local instead.
2079         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise for ffi_prep_args_SYSV.
2080
2081 2005-05-17  Kelley Cook  <kcook@gcc.gnu.org>
2082
2083         * configure.ac: Use AC_C_BIGENDIAN instead of AC_C_BIGENDIAN_CROSS.
2084         Use AC_CHECK_SIZEOF instead of AC_COMPILE_CHECK_SIZEOF.
2085         * Makefile.am (ACLOCAL_AMFLAGS): Remove -I ../config.
2086         * aclocal.m4, configure, fficonfig.h.in, Makefile.in,
2087         include/Makefile.in, testsuite/Makefile.in: Regenerate.
2088
2089 2005-05-09  Mike Stump  <mrs@apple.com>
2090
2091         * configure: Regenerate.
2092
2093 2005-05-08  Richard Henderson  <rth@redhat.com>
2094
2095         PR libffi/21285
2096         * src/alpha/osf.S: Update unwind into to match code.
2097
2098 2005-05-04  Andreas Degert <ad@papyrus-gmbh.de>
2099             Richard Henderson  <rth@redhat.com>
2100
2101         * src/x86/ffi64.c (ffi_prep_cif_machdep): Save sse-used flag in
2102         bit 11 of flags.
2103         (ffi_call): Mask return type field.  Pass ssecount to ffi_call_unix64.
2104         (ffi_prep_closure): Set carry bit if sse-used flag set.
2105         * src/x86/unix64.S (ffi_call_unix64): Add ssecount argument.
2106         Only load sse registers if ssecount non-zero.
2107         (ffi_closure_unix64): Only save sse registers if carry set on entry.
2108
2109 2005-04-29  Ralf Corsepius  <ralf.corsepius@rtems.org>
2110
2111         * configure.ac: Add i*86-*-rtems*, sparc*-*-rtems*,
2112         powerpc-*rtems*, arm*-*-rtems*, sh-*-rtems*.
2113         * configure: Regenerate.
2114
2115 2005-04-20  Hans-Peter Nilsson  <hp@axis.com>
2116
2117         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): In regsub use,
2118         have Tcl8.3-compatible intermediate variable.
2119
2120 2005-04-18  Simon Posnjak <simon.posnjak@siol.net>
2121             Hans-Peter Nilsson  <hp@axis.com>
2122
2123         * Makefile.am: Add CRIS support.
2124         * configure.ac: Likewise.
2125         * Makefile.in, configure, testsuite/Makefile.in,
2126         include/Makefile.in: Regenerate.
2127         * src/cris: New directory.
2128         * src/cris/ffi.c, src/cris/sysv.S, src/cris/ffitarget.h: New files.
2129         * src/prep_cif.c (ffi_prep_cif): Wrap in #ifndef __CRIS__.
2130
2131         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): Replace \n with
2132         \r?\n in output tests.
2133
2134 2005-04-12  Mike Stump  <mrs@apple.com>
2135
2136         * configure: Regenerate.
2137
2138 2005-03-30  Hans Boehm  <Hans.Boehm@hp.com>
2139
2140         * src/ia64/ffitarget.h (ffi_arg): Use long long instead of DI.
2141
2142 2005-03-30  Steve Ellcey  <sje@cup.hp.com>
2143
2144         * src/ia64/ffitarget.h (ffi_arg) ADD DI attribute.
2145         (ffi_sarg) Ditto.
2146         * src/ia64/unix.S (ffi_closure_unix): Extend gp
2147         to 64 bits in ILP32 mode.
2148         Load 64 bits even for short data.
2149
2150 2005-03-23  Mike Stump  <mrs@apple.com>
2151
2152         * src/powerpc/darwin.S: Update for -m64 multilib.
2153         * src/powerpc/darwin_closure.S: Likewise.
2154
2155 2005-03-21  Zack Weinberg  <zack@codesourcery.com>
2156
2157         * configure.ac: Do not invoke TL_AC_GCC_VERSION.
2158         Do not set tool_include_dir.
2159         * aclocal.m4, configure, Makefile.in, testsuite/Makefile.in:
2160         Regenerate.
2161         * include/Makefile.am: Set gcc_version and toollibffidir.
2162         * include/Makefile.in: Regenerate.
2163
2164 2005-02-22  Andrew Haley  <aph@redhat.com>
2165
2166         * src/powerpc/ffi.c (ffi_prep_cif_machdep): Bump alignment to
2167         odd-numbered register pairs for 64-bit integer types.
2168
2169 2005-02-23  Andreas Tobler  <a.tobler@schweiz.ch>
2170
2171         PR libffi/20104
2172         * testsuite/libffi.call/return_ll1.c: New test case.
2173
2174 2005-02-11  Janis Johnson  <janis187@us.ibm.com>
2175
2176         * testsuite/libffi.call/cls_align_longdouble.c: Remove dg-options.
2177         * testsuite/libffi.call/float.c: Ditto.
2178         * testsuite/libffi.call/float2.c: Ditto.
2179         * testsuite/libffi.call/float3.c: Ditto.
2180
2181 2005-02-08  Andreas Tobler  <a.tobler@schweiz.ch>
2182
2183         * src/frv/ffitarget.h: Remove PPC stuff which does not belong to frv.
2184
2185 2005-01-12  Eric Botcazou  <ebotcazou@libertysurf.fr>
2186
2187         * testsuite/libffi.special/special.exp (cxx_options): Add
2188         -shared-libgcc.
2189
2190 2004-12-31  Richard Henderson  <rth@redhat.com>
2191
2192         * src/types.c (FFI_AGGREGATE_TYPEDEF): Remove.
2193         (FFI_TYPEDEF): Rename from FFI_INTEGRAL_TYPEDEF.  Replace size and
2194         offset parameters with a type parameter; deduce size and structure
2195         alignment.  Update all users.
2196
2197 2004-12-31  Richard Henderson  <rth@redhat.com>
2198
2199         * src/types.c (FFI_TYPE_POINTER): Define with sizeof.
2200         (FFI_TYPE_LONGDOUBLE): Fix for ia64.
2201         * src/ia64/ffitarget.h (struct ffi_ia64_trampoline_struct): Move
2202         into ffi_prep_closure.
2203         * src/ia64/ia64_flags.h, src/ia64/ffi.c, src/ia64/unix.S: Rewrite
2204         from scratch.
2205
2206 2004-12-27  Richard Henderson  <rth@redhat.com>
2207
2208         * src/x86/unix64.S: Fix typo in unwind info.
2209
2210 2004-12-25  Richard Henderson  <rth@redhat.com>
2211
2212         * src/x86/ffi64.c (struct register_args): Rename from stackLayout.
2213         (enum x86_64_reg_class): Add X86_64_COMPLEX_X87_CLASS.
2214         (merge_classes): Check for it.
2215         (SSE_CLASS_P): New.
2216         (classify_argument): Pass byte_offset by value; perform all updates
2217         inside struct case.
2218         (examine_argument): Add classes argument; handle
2219         X86_64_COMPLEX_X87_CLASS.
2220         (ffi_prep_args): Merge into ...
2221         (ffi_call): ... here.  Share stack frame with ffi_call_unix64.
2222         (ffi_prep_cif_machdep): Setup cif->flags for proper structure return.
2223         (ffi_fill_return_value): Remove.
2224         (ffi_prep_closure): Remove dead assert.
2225         (ffi_closure_unix64_inner): Rename from ffi_closure_UNIX64_inner.
2226         Rewrite to use struct register_args instead of va_list.  Create
2227         flags for handling structure returns.
2228         * src/x86/unix64.S: Remove dead strings.
2229         (ffi_call_unix64): Rename from ffi_call_UNIX64.  Rewrite to share
2230         stack frame with ffi_call.  Handle structure returns properly.
2231         (float2sse, floatfloat2sse, double2sse): Remove.
2232         (sse2float, sse2double, sse2floatfloat): Remove.
2233         (ffi_closure_unix64): Rename from ffi_closure_UNIX64.  Rewrite
2234         to handle structure returns properly.
2235
2236 2004-12-08  David Edelsohn  <edelsohn@gnu.org>
2237
2238         * Makefile.am (AM_MAKEFLAGS): Remove duplicate LIBCFLAGS and
2239         PICFLAG.
2240         * Makefile.in: Regenerated.
2241
2242 2004-12-02  Richard Sandiford  <rsandifo@redhat.com>
2243
2244         * configure.ac: Use TL_AC_GCC_VERSION to set gcc_version.
2245         * configure, aclocal.m4, Makefile.in: Regenerate.
2246         * include/Makefile.in, testsuite/Makefile.in: Regenerate.
2247
2248 2004-11-29  Kelley Cook  <kcook@gcc.gnu.org>
2249
2250         * configure: Regenerate for libtool change.
2251
2252 2004-11-25  Kelley Cook  <kcook@gcc.gnu.org>
2253
2254         * configure: Regenerate for libtool reversion.
2255
2256 2004-11-24  Kelley Cook  <kcook@gcc.gnu.org>
2257
2258         * configure: Regenerate for libtool change.
2259
2260 2004-11-23  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2261
2262         * testsuite/lib/libffi-dg.exp: Use new procs in target-libpath.exp.
2263
2264 2004-11-23  Richard Sandiford  <rsandifo@redhat.com>
2265
2266         * src/mips/o32.S (ffi_call_O32, ffi_closure_O32): Use jalr instead
2267         of jal.  Use an absolute encoding for the frame information.
2268
2269 2004-11-23  Kelley Cook  <kcook@gcc.gnu.org>
2270
2271         * Makefile.am: Remove no-dependencies.  Add ACLOCAL_AMFLAGS.
2272         * acinclude.m4: Delete logic for sincludes.
2273         * aclocal.m4, Makefile.in, configure: Regenerate.
2274         * include/Makefile: Likewise.
2275         * testsuite/Makefile: Likewise.
2276
2277 2004-11-22  Eric Botcazou  <ebotcazou@libertysurf.fr>
2278
2279         * src/sparc/ffi.c (ffi_prep_closure): Align doubles and 64-bit integers
2280         on a 8-byte boundary.
2281         * src/sparc/v8.S (ffi_closure_v8): Reserve frame space for arguments.
2282
2283 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2284
2285         * src/arm/ffi.c (ffi_prep_cif_machdep): Handle functions that return
2286         long long values.  Round stack allocation to a multiple of 8 bytes
2287         for ATPCS compatibility.
2288         * src/arm/sysv.S (ffi_call_SYSV): Rework to avoid use of APCS register
2289         names.  Handle returning long long types.  Add Thumb and interworking
2290         support.  Improve soft-float code.
2291
2292 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2293
2294         * testsuite/lib/libffi-db.exp (load_gcc_lib): New function.
2295         (libffi_exit): New function.
2296         (libffi_init): Build the testglue wrapper if needed.
2297
2298 2004-10-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2299
2300         PR other/18138
2301         * testsuite/lib/libffi-dg.exp: Accept more than one multilib libgcc.
2302
2303 2004-10-25  Kazuhiro Inaoka <inaoka.kazuhiro@renesas.com>
2304
2305         * src/m32r/libffitarget.h (FFI_CLOSURES): Set to 0.
2306
2307 2004-10-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2308
2309         * src/sh/sysv.S (ffi_call_SYSV): Don't align for double data.
2310         * testsuite/libffi.call/float3.c: New test case.
2311
2312 2004-10-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2313
2314         * src/sh/ffi.c (ffi_prep_closure): Set T bit in trampoline for
2315         the function returning a structure pointed with R2.
2316         * src/sh/sysv.S (ffi_closure_SYSV): Use R2 as the pointer to
2317         the structure return value if T bit set.  Emit position
2318         independent code and EH data if PIC.
2319
2320 2004-10-13  Kazuhiro Inaoka  <inaoka.kazuhiro@renesas.com>
2321
2322         * Makefile.am: Add m32r support.
2323         * configure.ac: Likewise.
2324         * Makefile.in: Regenerate.
2325         * confiugre: Regenerate.
2326         * src/types.c: Add m32r port to FFI_INTERNAL_TYPEDEF
2327         (uint64, sint64, double, longdouble)
2328         * src/m32r: New directory.
2329         * src/m32r/ffi.c: New file.
2330         * src/m32r/sysv.S: Likewise.
2331         * src/m32r/ffitarget.h: Likewise.
2332
2333 2004-10-02  Kaz Kojima  <kkojima@gcc.gnu.org>
2334
2335         * testsuite/libffi.call/negint.c: New test case.
2336
2337 2004-09-14  H.J. Lu  <hongjiu.lu@intel.com>
2338
2339         PR libgcj/17465
2340         * testsuite/lib/libffi-dg.exp: Don't use global ld_library_path.
2341         Set up LD_LIBRARY_PATH, SHLIB_PATH, LD_LIBRARYN32_PATH,
2342         LD_LIBRARY64_PATH, LD_LIBRARY_PATH_32, LD_LIBRARY_PATH_64 and
2343         DYLD_LIBRARY_PATH.
2344
2345 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2346
2347         * testsuite/libffi.call/many_win32.c: Remove whitespaces.
2348         * testsuite/libffi.call/promotion.c: Likewise.
2349         * testsuite/libffi.call/return_ll.c: Remove unused var. Cleanup
2350         whitespaces.
2351         * testsuite/libffi.call/return_sc.c: Likewise.
2352         * testsuite/libffi.call/return_uc.c: Likewise.
2353
2354 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2355
2356         * src/powerpc/darwin.S: Fix comments and identation.
2357         * src/powerpc/darwin_closure.S: Likewise.
2358
2359 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2360
2361         * src/powerpc/ffi_darwin.c: Add flag for longdouble return values.
2362         (ffi_prep_args): Handle longdouble arguments.
2363         (ffi_prep_cif_machdep): Set flags for longdouble. Calculate space for
2364         longdouble.
2365         (ffi_closure_helper_DARWIN): Add closure handling for longdouble.
2366         * src/powerpc/darwin.S (_ffi_call_DARWIN): Add handling of longdouble
2367         values.
2368         * src/powerpc/darwin_closure.S (_ffi_closure_ASM): Likewise.
2369         * src/types.c: Defined longdouble size and alignment for darwin.
2370
2371 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2372
2373         * src/powerpc/aix.S: Remove whitespaces.
2374         * src/powerpc/aix_closure.S: Likewise.
2375         * src/powerpc/asm.h: Likewise.
2376         * src/powerpc/ffi.c: Likewise.
2377         * src/powerpc/ffitarget.h: Likewise.
2378         * src/powerpc/linux64.S: Likewise.
2379         * src/powerpc/linux64_closure.S: Likewise.
2380         * src/powerpc/ppc_closure.S: Likewise.
2381         * src/powerpc/sysv.S: Likewise.
2382
2383 2004-08-30  Anthony Green  <green@redhat.com>
2384
2385         * Makefile.am: Add frv support.
2386         * Makefile.in, testsuite/Makefile.in: Rebuilt.
2387         * configure.ac: Read configure.host.
2388         * configure.in: Read configure.host.
2389         * configure.host: New file.  frv-elf needs libgloss.
2390         * include/ffi.h.in: Force ffi_closure to have a nice big (8)
2391         alignment.  This is needed to frv and shouldn't harm the others.
2392         * include/ffi_common.h (ALIGN_DOWN): New macro.
2393         * src/frv/ffi.c, src/frv/ffitarget.h, src/frv/eabi.S: New files.
2394
2395 2004-08-24  David Daney  <daney@avtrex.com>
2396
2397         * testsuite/libffi.call/closure_fn0.c: Xfail mips64* instead of mips*.
2398         * testsuite/libffi.call/closure_fn1.c: Likewise.
2399         * testsuite/libffi.call/closure_fn2.c  Likewise.
2400         * testsuite/libffi.call/closure_fn3.c: Likewise.
2401         * testsuite/libffi.call/closure_fn4.c: Likewise.
2402         * testsuite/libffi.call/closure_fn5.c: Likewise.
2403         * testsuite/libffi.call/cls_18byte.c: Likewise.
2404         * testsuite/libffi.call/cls_19byte.c: Likewise.
2405         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2406         * testsuite/libffi.call/cls_20byte.c: Likewise.
2407         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2408         * testsuite/libffi.call/cls_24byte.c: Likewise.
2409         * testsuite/libffi.call/cls_2byte.c: Likewise.
2410         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2411         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2412         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2413         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2414         * testsuite/libffi.call/cls_4byte.c: Likewise.
2415         * testsuite/libffi.call/cls_64byte.c: Likewise.
2416         * testsuite/libffi.call/cls_6byte.c: Likewise.
2417         * testsuite/libffi.call/cls_7byte.c: Likewise.
2418         * testsuite/libffi.call/cls_8byte.c: Likewise.
2419         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2420         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2421         * testsuite/libffi.call/cls_align_double.c: Likewise.
2422         * testsuite/libffi.call/cls_align_float.c: Likewise.
2423         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2424         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2425         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2426         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2427         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2428         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2429         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2430         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
2431         * testsuite/libffi.call/cls_double.c: Likewise.
2432         * testsuite/libffi.call/cls_float.c: Likewise.
2433         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2434         * testsuite/libffi.call/cls_multi_sshort.c: Likewise.
2435         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2436         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2437         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2438         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2439         * testsuite/libffi.call/cls_schar.c: Likewise.
2440         * testsuite/libffi.call/cls_sint.c: Likewise.
2441         * testsuite/libffi.call/cls_sshort.c: Likewise.
2442         * testsuite/libffi.call/cls_uchar.c: Likewise.
2443         * testsuite/libffi.call/cls_uint.c: Likewise.
2444         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2445         * testsuite/libffi.call/cls_ushort.c: Likewise.
2446         * testsuite/libffi.call/nested_struct.c: Likewise.
2447         * testsuite/libffi.call/nested_struct1.c: Likewise.
2448         * testsuite/libffi.call/nested_struct2.c: Likewise.
2449         * testsuite/libffi.call/nested_struct3.c: Likewise.
2450         * testsuite/libffi.call/problem1.c: Likewise.
2451         * testsuite/libffi.special/unwindtest.cc: Likewise.
2452         * testsuite/libffi.call/cls_12byte.c: Likewise and set return value
2453         to zero.
2454         * testsuite/libffi.call/cls_16byte.c: Likewise.
2455         * testsuite/libffi.call/cls_5byte.c: Likewise.
2456
2457 2004-08-23  David Daney <daney@avtrex.com>
2458
2459         PR libgcj/13141
2460         * src/mips/ffitarget.h (FFI_O32_SOFT_FLOAT): New ABI.
2461         * src/mips/ffi.c (ffi_prep_args): Fix alignment calculation.
2462         (ffi_prep_cif_machdep): Handle FFI_O32_SOFT_FLOAT floating point
2463         parameters and return types.
2464         (ffi_call): Handle FFI_O32_SOFT_FLOAT ABI.
2465         (ffi_prep_closure): Ditto.
2466         (ffi_closure_mips_inner_O32): Handle FFI_O32_SOFT_FLOAT ABI, fix
2467         alignment calculations.
2468         * src/mips/o32.S (ffi_closure_O32): Don't use floating point
2469         instructions if FFI_O32_SOFT_FLOAT, make stack frame ABI compliant.
2470
2471 2004-08-14  Casey Marshall <csm@gnu.org>
2472
2473         * src/mips/ffi.c (ffi_pref_cif_machdep): set `cif->flags' to
2474         contain `FFI_TYPE_UINT64' as return type for any 64-bit
2475         integer (O32 ABI only).
2476         (ffi_prep_closure): new function.
2477         (ffi_closure_mips_inner_O32): new function.
2478         * src/mips/ffitarget.h: Define `FFI_CLOSURES' and
2479         `FFI_TRAMPOLINE_SIZE' appropriately if the ABI is o32.
2480         * src/mips/o32.S (ffi_call_O32): add labels for .eh_frame. Return
2481         64 bit integers correctly.
2482         (ffi_closure_O32): new function.
2483         Added DWARF-2 unwind info for both functions.
2484
2485 2004-08-10  Andrew Haley  <aph@redhat.com>
2486
2487         * src/x86/ffi64.c (ffi_prep_args ): 8-align all stack arguments.
2488
2489 2004-08-01  Robert Millan  <robertmh@gnu.org>
2490
2491         * configure.ac: Detect knetbsd-gnu and kfreebsd-gnu.
2492         * configure: Regenerate.
2493
2494 2004-07-30  Maciej W. Rozycki  <macro@linux-mips.org>
2495
2496         * acinclude.m4 (AC_FUNC_MMAP_BLACKLIST): Check for <sys/mman.h>
2497         and mmap() explicitly instead of relying on preset autoconf cache
2498         variables.
2499         * aclocal.m4: Regenerate.
2500         * configure: Regenerate.
2501
2502 2004-07-11  Ulrich Weigand  <uweigand@de.ibm.com>
2503
2504         * src/s390/ffi.c (ffi_prep_args): Fix C aliasing violation.
2505         (ffi_check_float_struct): Remove unused prototype.
2506
2507 2004-06-30  Geoffrey Keating  <geoffk@apple.com>
2508
2509         * src/powerpc/ffi_darwin.c (flush_icache): ';' is a comment
2510         character on Darwin, use '\n\t' instead.
2511
2512 2004-06-26  Matthias Klose  <doko@debian.org>
2513
2514         * libtool-version: Fix typo in revision/age.
2515
2516 2004-06-17  Matthias Klose  <doko@debian.org>
2517
2518         * libtool-version: New.
2519         * Makefile.am (libffi_la_LDFLAGS): Use -version-info for soname.
2520         * Makefile.in: Regenerate.
2521
2522 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2523
2524         * Makefile.am: Remove useless multilib rules.
2525         * Makefile.in: Regenerate.
2526         * aclocal.m4: Regenerate with automake 1.8.5.
2527         * configure.ac: Remove useless multilib configury.
2528         * configure: Regenerate.
2529
2530 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2531
2532         * .cvsignore: New file.
2533
2534 2004-06-10  Jakub Jelinek  <jakub@redhat.com>
2535
2536         * src/ia64/unix.S (ffi_call_unix): Insert group barrier break
2537         fp_done.
2538         (ffi_closure_UNIX): Fix f14/f15 adjustment if FLOAT_SZ is ever
2539         changed from 8.
2540
2541 2004-06-06  Sean McNeil  <sean@mcneil.com>
2542
2543         * configure.ac: Add x86_64-*-freebsd* support.
2544         * configure: Regenerate.
2545
2546 2004-04-26  Joe Buck <jbuck@welsh-buck.org>
2547
2548         Bug 15093
2549         * configure.ac: Test for existence of mmap and sys/mman.h before
2550         checking blacklist.  Fix suggested by Jim Wilson.
2551         * configure: Regenerate.
2552
2553 2004-04-26  Matt Austern  <austern@apple.com>
2554
2555         * src/powerpc/darwin.S: Go through a non-lazy pointer for initial
2556         FDE location.
2557         * src/powerpc/darwin_closure.S: Likewise.
2558
2559 2004-04-24  Andreas Tobler  <a.tobler@schweiz.ch>
2560
2561         * testsuite/libffi.call/cls_multi_schar.c (main): Fix initialization
2562         error. Reported by Thomas Heller <theller@python.net>.
2563         * testsuite/libffi.call/cls_multi_sshort.c (main): Likewise.
2564         * testsuite/libffi.call/cls_multi_ushort.c (main): Likewise.
2565
2566 2004-03-20  Matthias Klose  <doko@debian.org>
2567
2568         * src/pa/linux.S: Fix typo.
2569
2570 2004-03-19  Matthias Klose  <doko@debian.org>
2571
2572         * Makefile.am: Update.
2573         * Makefile.in: Regenerate.
2574         * src/pa/ffi.h.in: Remove.
2575         * src/pa/ffitarget.h: New file.
2576
2577 2004-02-10  Randolph Chung  <tausq@debian.org>
2578
2579         * Makefile.am: Add PA support.
2580         * Makefile.in: Regenerate.
2581         * include/Makefile.in: Regenerate.
2582         * configure.ac: Add PA target.
2583         * configure: Regenerate.
2584         * src/pa/ffi.c: New file.
2585         * src/pa/ffi.h.in: Add PA support.
2586         * src/pa/linux.S: New file.
2587         * prep_cif.c: Add PA support.
2588
2589 2004-03-16  Hosaka Yuji  <hos@tamanegi.org>
2590
2591         * src/types.c: Fix alignment size of X86_WIN32 case int64 and
2592         double.
2593         * src/x86/ffi.c (ffi_prep_args): Replace ecif->cif->rtype->type
2594         with ecif->cif->flags.
2595         (ffi_call, ffi_prep_incoming_args_SYSV): Replace cif->rtype->type
2596         with cif->flags.
2597         (ffi_prep_cif_machdep): Add X86_WIN32 struct case.
2598         (ffi_closure_SYSV): Add 1 or 2-bytes struct case for X86_WIN32.
2599         * src/x86/win32.S (retstruct1b, retstruct2b, sc_retstruct1b,
2600         sc_retstruct2b): Add for 1 or 2-bytes struct case.
2601
2602 2004-03-15 Kelley Cook <kcook@gcc.gnu.org>
2603
2604         * configure.in: Rename file to ...
2605         * configure.ac: ... this.
2606         * fficonfig.h.in: Regenerate.
2607         * Makefile.in: Regenerate.
2608         * include/Makefile.in: Regenerate.
2609         * testsuite/Makefile.in: Regenerate.
2610
2611 2004-03-12  Matt Austern  <austern@apple.com>
2612
2613         * src/powerpc/darwin.S: Fix EH information so it corresponds to
2614         changes in EH format resulting from addition of linkonce support.
2615         * src/powerpc/darwin_closure.S: Likewise.
2616
2617 2004-03-11  Andreas Tobler  <a.tobler@schweiz.ch>
2618             Paolo Bonzini  <bonzini@gnu.org>
2619
2620         * Makefile.am (AUTOMAKE_OPTIONS): Set them.
2621         Remove VPATH. Remove rules for object files. Remove multilib support.
2622         (AM_CCASFLAGS): Add.
2623         * configure.in (AC_CONFIG_HEADERS): Relace AM_CONFIG_HEADER.
2624         (AC_PREREQ): Bump version to 2.59.
2625         (AC_INIT): Fill with version info and bug address.
2626         (ORIGINAL_LD_FOR_MULTILIBS): Remove.
2627         (AM_ENABLE_MULTILIB): Use this instead of AC_ARG_ENABLE.
2628         De-precious CC so that the right flags are passed down to multilibs.
2629         (AC_MSG_ERROR): Replace obsolete macro AC_ERROR.
2630         (AC_CONFIG_FILES): Replace obsolete macro AC_LINK_FILES.
2631         (AC_OUTPUT): Reorganize the output with AC_CONFIG_COMMANDS.
2632         * configure: Rebuilt.
2633         * aclocal.m4: Likewise.
2634         * Makefile.in, include/Makefile.in, testsuite/Makefile.in: Likewise.
2635         * fficonfig.h.in: Likewise.
2636
2637 2004-03-11  Andreas Schwab  <schwab@suse.de>
2638
2639         * src/ia64/ffi.c (ffi_prep_incoming_args_UNIX): Get floating point
2640         arguments from fp registers only for the first 8 parameter slots.
2641         Don't convert a float parameter when passed in memory.
2642
2643 2004-03-09  Hans-Peter Nilsson  <hp@axis.com>
2644
2645         * configure: Regenerate for config/accross.m4 correction.
2646
2647 2004-02-25  Matt Kraai  <kraai@alumni.cmu.edu>
2648
2649         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Change
2650         ecif->cif->bytes to bytes.
2651         (ffi_prep_cif_machdep): Add braces around nested if statement.
2652
2653 2004-02-09  Alan Modra  <amodra@bigpond.net.au>
2654
2655         * src/types.c (pointer): POWERPC64 has 8 byte pointers.
2656
2657         * src/powerpc/ffi.c (ffi_prep_args64): Correct long double handling.
2658         (ffi_closure_helper_LINUX64): Fix typo.
2659         * testsuite/libffi.call/cls_align_longdouble.c: Pass -mlong-double-128
2660         for powerpc64-*-*.
2661         * testsuite/libffi.call/float.c: Likewise.
2662         * testsuite/libffi.call/float2.c: Likewise.
2663
2664 2004-02-08  Alan Modra  <amodra@bigpond.net.au>
2665
2666         * src/powerpc/ffi.c (ffi_prep_cif_machdep <FFI_LINUX64>): Correct
2667         long double function return and long double arg handling.
2668         (ffi_closure_helper_LINUX64): Formatting.  Delete unused "ng" var.
2669         Use "end_pfr" instead of "nf".  Correct long double handling.
2670         Localise "temp".
2671         * src/powerpc/linux64.S (ffi_call_LINUX64): Save f2 long double
2672         return value.
2673         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Allocate
2674         space for long double return value.  Adjust stack frame and offsets.
2675         Load f2 long double return.
2676
2677 2004-02-07  Alan Modra  <amodra@bigpond.net.au>
2678
2679         * src/types.c: Use 16 byte long double for POWERPC64.
2680
2681 2004-01-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2682
2683         * src/sparc/ffi.c (ffi_prep_args_v9): Shift the parameter array
2684         when the structure return address is passed in %o0.
2685         (ffi_V9_return_struct): Rename into ffi_v9_layout_struct.
2686         (ffi_v9_layout_struct): Align the field following a nested structure
2687         on a word boundary.  Use memmove instead of memcpy.
2688         (ffi_call): Update call to ffi_V9_return_struct.
2689         (ffi_prep_closure): Define 'ctx' only for V8.
2690         (ffi_closure_sparc_inner): Clone into ffi_closure_sparc_inner_v8
2691         and ffi_closure_sparc_inner_v9.
2692         (ffi_closure_sparc_inner_v8): Return long doubles by reference.
2693         Always skip the structure return address.  For structures and long
2694         doubles, copy the argument directly.
2695         (ffi_closure_sparc_inner_v9): Skip the structure return address only
2696         if required.  Shift the maximum floating-point slot accordingly.  For
2697         big structures, copy the argument directly; otherwise, left-justify the
2698         argument and call ffi_v9_layout_struct to lay out the structure on
2699         the stack.
2700         * src/sparc/v8.S: Undef STACKFRAME before defining it.
2701         (ffi_closure_v8): Pass the structure return address.  Update call to
2702         ffi_closure_sparc_inner_v8.  Short-circuit FFI_TYPE_INT handling.
2703         Skip the 'unimp' insn when returning long doubles and structures.
2704         * src/sparc/v9.S: Undef STACKFRAME before defining it.
2705         (ffi_closure_v9): Increase the frame size by 2 words.  Short-circuit
2706         FFI_TYPE_INT handling.  Load structures both in integers and
2707         floating-point registers on return.
2708         * README: Update status of the SPARC port.
2709
2710 2004-01-24  Andreas Tobler  <a.tobler@schweiz.ch>
2711
2712         * testsuite/libffi.call/pyobjc-tc.c (main): Treat result value
2713         as of type ffi_arg.
2714         * testsuite/libffi.call/struct3.c (main): Fix CHECK.
2715
2716 2004-01-22  Ulrich Weigand  <uweigand@de.ibm.com>
2717
2718         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Treat result
2719         value as of type ffi_arg, not unsigned int.
2720
2721 2004-01-21  Michael Ritzert  <ritzert@t-online.de>
2722
2723         * ffi64.c (ffi_prep_args): Cast the RHS of an assignment instead
2724         of the LHS.
2725
2726 2004-01-12  Andreas Tobler  <a.tobler@schweiz.ch>
2727
2728         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_32 for
2729         Solaris.
2730
2731 2004-01-08  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2732
2733         * testsuite/libffi.call/ffitest.h (allocate_mmap): Cast MAP_FAILED
2734         to void *.
2735
2736 2003-12-10  Richard Henderson  <rth@redhat.com>
2737
2738         * testsuite/libffi.call/cls_align_pointer.c: Cast pointers to
2739         size_t instead of int.
2740
2741 2003-12-04  Hosaka Yuji  <hos@tamanegi.org>
2742
2743         * testsuite/libffi.call/many_win32.c: Include <float.h>.
2744         * testsuite/libffi.call/many_win32.c (main): Replace variable
2745         int i with unsigned long ul.
2746
2747         * testsuite/libffi.call/cls_align_uint64.c: New test case.
2748         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2749         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2750         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2751         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2752         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2753         * testsuite/libffi.call/cls_align_float.c: Likewise.
2754         * testsuite/libffi.call/cls_align_double.c: Likewise.
2755         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2756         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2757
2758 2003-12-02  Hosaka Yuji  <hos@tamanegi.org>
2759
2760         PR other/13221
2761         * src/x86/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV):
2762         Align arguments to 32 bits.
2763
2764 2003-12-01  Andreas Tobler  <a.tobler@schweiz.ch>
2765
2766         PR other/13221
2767         * testsuite/libffi.call/cls_multi_sshort.c: New test case.
2768         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2769         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2770         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2771         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2772         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2773
2774         * testsuite/libffi.special/unwindtest.cc: Cosmetics.
2775
2776 2003-11-26  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
2777
2778         * testsuite/libffi.call/ffitest.h: Include <fcntl.h>.
2779         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2780
2781 2003-11-22  Andreas Tobler  <a.tobler@schweiz.ch>
2782
2783         * Makefile.in: Rebuilt.
2784         * configure: Likewise.
2785         * testsuite/libffi.special/unwindtest.cc: Convert the mmap to
2786         the right type.
2787
2788 2003-11-21  Andreas Jaeger  <aj@suse.de>
2789             Andreas Tobler  <a.tobler@schweiz.ch>
2790
2791         * acinclude.m4: Add AC_FUNC_MMAP_BLACKLIST.
2792         * configure.in: Call AC_FUNC_MMAP_BLACKLIST.
2793         * Makefile.in: Rebuilt.
2794         * aclocal.m4: Likewise.
2795         * configure: Likewise.
2796         * fficonfig.h.in: Likewise.
2797         * testsuite/lib/libffi-dg.exp: Add include dir.
2798         * testsuite/libffi.call/ffitest.h: Add MMAP definitions.
2799         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2800         * testsuite/libffi.call/closure_fn0.c: Use MMAP functionality
2801         for ffi_closure if available.
2802         * testsuite/libffi.call/closure_fn1.c: Likewise.
2803         * testsuite/libffi.call/closure_fn2.c: Likewise.
2804         * testsuite/libffi.call/closure_fn3.c: Likewise.
2805         * testsuite/libffi.call/closure_fn4.c: Likewise.
2806         * testsuite/libffi.call/closure_fn5.c: Likewise.
2807         * testsuite/libffi.call/cls_12byte.c: Likewise.
2808         * testsuite/libffi.call/cls_16byte.c: Likewise.
2809         * testsuite/libffi.call/cls_18byte.c: Likewise.
2810         * testsuite/libffi.call/cls_19byte.c: Likewise.
2811         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2812         * testsuite/libffi.call/cls_20byte.c: Likewise.
2813         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2814         * testsuite/libffi.call/cls_24byte.c: Likewise.
2815         * testsuite/libffi.call/cls_2byte.c: Likewise.
2816         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2817         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2818         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2819         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2820         * testsuite/libffi.call/cls_4byte.c: Likewise.
2821         * testsuite/libffi.call/cls_5byte.c: Likewise.
2822         * testsuite/libffi.call/cls_64byte.c: Likewise.
2823         * testsuite/libffi.call/cls_6byte.c: Likewise.
2824         * testsuite/libffi.call/cls_7byte.c: Likewise.
2825         * testsuite/libffi.call/cls_8byte.c: Likewise.
2826         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2827         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2828         * testsuite/libffi.call/cls_double.c: Likewise.
2829         * testsuite/libffi.call/cls_float.c: Likewise.
2830         * testsuite/libffi.call/cls_schar.c: Likewise.
2831         * testsuite/libffi.call/cls_sint.c: Likewise.
2832         * testsuite/libffi.call/cls_sshort.c: Likewise.
2833         * testsuite/libffi.call/cls_uchar.c: Likewise.
2834         * testsuite/libffi.call/cls_uint.c: Likewise.
2835         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2836         * testsuite/libffi.call/cls_ushort.c: Likewise.
2837         * testsuite/libffi.call/nested_struct.c: Likewise.
2838         * testsuite/libffi.call/nested_struct1.c: Likewise.
2839         * testsuite/libffi.call/nested_struct2.c: Likewise.
2840         * testsuite/libffi.call/nested_struct3.c: Likewise.
2841         * testsuite/libffi.call/problem1.c: Likewise.
2842         * testsuite/libffi.special/unwindtest.cc: Likewise.
2843
2844 2003-11-20  Andreas Tobler  <a.tobler@schweiz.ch>
2845
2846         * testsuite/lib/libffi-dg.exp: Make the -lgcc_s conditional.
2847
2848 2003-11-19  Andreas Tobler  <a.tobler@schweiz.ch>
2849
2850         * testsuite/lib/libffi-dg.exp: Add DYLD_LIBRARY_PATH for darwin.
2851         Add -lgcc_s to additional flags.
2852
2853 2003-11-12  Andreas Tobler  <a.tobler@schweiz.ch>
2854
2855         * configure.in, include/Makefile.am: PR libgcj/11147, install
2856         the ffitarget.h header file in a gcc versioned and target
2857         dependent place.
2858         * configure: Regenerated.
2859         * Makefile.in, include/Makefile.in: Likewise.
2860         * testsuite/Makefile.in: Likewise.
2861
2862 2003-11-09  Andreas Tobler  <a.tobler@schweiz.ch>
2863
2864         * testsuite/libffi.call/closure_fn0.c: Print result and check
2865         with dg-output to make debugging easier.
2866         * testsuite/libffi.call/closure_fn1.c: Likewise.
2867         * testsuite/libffi.call/closure_fn2.c: Likewise.
2868         * testsuite/libffi.call/closure_fn3.c: Likewise.
2869         * testsuite/libffi.call/closure_fn4.c: Likewise.
2870         * testsuite/libffi.call/closure_fn5.c: Likewise.
2871         * testsuite/libffi.call/cls_12byte.c: Likewise.
2872         * testsuite/libffi.call/cls_16byte.c: Likewise.
2873         * testsuite/libffi.call/cls_18byte.c: Likewise.
2874         * testsuite/libffi.call/cls_19byte.c: Likewise.
2875         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2876         * testsuite/libffi.call/cls_20byte.c: Likewise.
2877         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2878         * testsuite/libffi.call/cls_24byte.c: Likewise.
2879         * testsuite/libffi.call/cls_2byte.c: Likewise.
2880         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2881         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2882         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2883         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2884         * testsuite/libffi.call/cls_4byte.c: Likewise.
2885         * testsuite/libffi.call/cls_5byte.c: Likewise.
2886         * testsuite/libffi.call/cls_64byte.c: Likewise.
2887         * testsuite/libffi.call/cls_6byte.c: Likewise.
2888         * testsuite/libffi.call/cls_7byte.c: Likewise.
2889         * testsuite/libffi.call/cls_8byte.c: Likewise.
2890         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2891         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2892         * testsuite/libffi.call/cls_double.c: Likewise.
2893         * testsuite/libffi.call/cls_float.c: Likewise.
2894         * testsuite/libffi.call/cls_schar.c: Likewise.
2895         * testsuite/libffi.call/cls_sint.c: Likewise.
2896         * testsuite/libffi.call/cls_sshort.c: Likewise.
2897         * testsuite/libffi.call/cls_uchar.c: Likewise.
2898         * testsuite/libffi.call/cls_uint.c: Likewise.
2899         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2900         * testsuite/libffi.call/cls_ushort.c: Likewise.
2901         * testsuite/libffi.call/problem1.c: Likewise.
2902
2903         * testsuite/libffi.special/unwindtest.cc: Make ffi_closure
2904         static.
2905
2906 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
2907
2908         * testsuite/libffi.call/cls_9byte2.c: New test case.
2909         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2910         * testsuite/libffi.call/cls_64byte.c: Likewise.
2911         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2912         * testsuite/libffi.call/cls_19byte.c: Likewise.
2913         * testsuite/libffi.call/cls_18byte.c: Likewise.
2914         * testsuite/libffi.call/closure_fn4.c: Likewise.
2915         * testsuite/libffi.call/closure_fn5.c: Likewise.
2916         * testsuite/libffi.call/cls_schar.c: Likewise.
2917         * testsuite/libffi.call/cls_sint.c: Likewise.
2918         * testsuite/libffi.call/cls_sshort.c: Likewise.
2919         * testsuite/libffi.call/nested_struct2.c: Likewise.
2920         * testsuite/libffi.call/nested_struct3.c: Likewise.
2921
2922 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
2923
2924         * testsuite/libffi.call/cls_double.c: Do a check on the result.
2925         * testsuite/libffi.call/cls_uchar.c: Likewise.
2926         * testsuite/libffi.call/cls_uint.c: Likewise.
2927         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2928         * testsuite/libffi.call/cls_ushort.c: Likewise.
2929         * testsuite/libffi.call/return_sc.c: Cleanup whitespaces.
2930
2931 2003-11-06  Andreas Tobler  <a.tobler@schweiz.ch>
2932
2933         * src/prep_cif.c (ffi_prep_cif): Move the validity check after
2934         the initialization.
2935
2936 2003-10-23  Andreas Tobler  <a.tobler@schweiz.ch>
2937
2938         * src/java_raw_api.c (ffi_java_ptrarray_to_raw): Replace
2939         FFI_ASSERT(FALSE) with FFI_ASSERT(0).
2940
2941 2003-10-22  David Daney  <ddaney@avtrex.com>
2942
2943         * src/mips/ffitarget.h: Replace undefined UINT32 and friends with
2944         __attribute__((__mode__(__SI__))) and friends.
2945
2946 2003-10-22  Andreas Schwab  <schwab@suse.de>
2947
2948         * src/ia64/ffi.c: Replace FALSE/TRUE with false/true.
2949
2950 2003-10-21  Andreas Tobler  <a.tobler@schweiz.ch>
2951
2952         * configure.in: AC_LINK_FILES(ffitarget.h).
2953         * configure: Regenerate.
2954         * Makefile.in: Likewise.
2955         * include/Makefile.in: Likewise.
2956         * testsuite/Makefile.in: Likewise.
2957         * fficonfig.h.in: Likewise.
2958
2959 2003-10-21  Paolo Bonzini  <bonzini@gnu.org>
2960             Richard Henderson  <rth@redhat.com>
2961
2962         Avoid that ffi.h includes fficonfig.h.
2963
2964         * Makefile.am (EXTRA_DIST): Include ffitarget.h files
2965         (TARGET_SRC_MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
2966         (TARGET_SRC_MIPS_SGI): Removed.
2967         (MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
2968         (MIPS_SGI): Removed.
2969         (CLEANFILES): Removed.
2970         (mostlyclean-am, clean-am, mostlyclean-sub, clean-sub): New
2971         targets.
2972         * acconfig.h: Removed.
2973         * configure.in: Compute sizeofs only for double and long double.
2974         Use them to define and subst HAVE_LONG_DOUBLE.  Include comments
2975         into AC_DEFINE instead of using acconfig.h.  Create
2976         include/ffitarget.h instead of include/fficonfig.h.  Rename
2977         MIPS_GCC to MIPS_IRIX, drop MIPS_SGI since we are in gcc's tree.
2978         AC_DEFINE EH_FRAME_FLAGS.
2979         * include/Makefile.am (DISTCLEANFILES): New automake macro.
2980         (hack_DATA): Add ffitarget.h.
2981         * include/ffi.h.in: Remove all system specific definitions.
2982         Declare raw API even if it is not installed, why bother?
2983         Use limits.h instead of SIZEOF_* to define ffi_type_*.  Do
2984         not define EH_FRAME_FLAGS, it is in fficonfig.h now.  Include
2985         ffitarget.h instead of fficonfig.h.  Remove ALIGN macro.
2986         (UINT_ARG, INT_ARG): Removed, use ffi_arg and ffi_sarg instead.
2987         * include/ffi_common.h (bool): Do not define.
2988         (ffi_assert): Accept failed assertion.
2989         (ffi_type_test): Return void and accept file/line.
2990         (FFI_ASSERT): Pass stringized failed assertion.
2991         (FFI_ASSERT_AT): New macro.
2992         (FFI_ASSERT_VALID_TYPE): New macro.
2993         (UINT8, SINT8, UINT16, SINT16, UINT32, SINT32,
2994         UINT64, SINT64): Define here with gcc's __attribute__ macro
2995         instead of in ffi.h
2996         (FLOAT32, ALIGN): Define here instead of in ffi.h
2997         * include/ffi-mips.h: Removed.  Its content moved to
2998         src/mips/ffitarget.h after separating assembly and C sections.
2999         * src/alpha/ffi.c, src/alpha/ffi.c, src/java_raw_api.c
3000         src/prep_cif.c, src/raw_api.c, src/ia64/ffi.c,
3001         src/mips/ffi.c, src/mips/n32.S, src/mips/o32.S,
3002         src/mips/ffitarget.h, src/sparc/ffi.c, src/x86/ffi64.c:
3003         SIZEOF_ARG -> FFI_SIZEOF_ARG.
3004         * src/ia64/ffi.c: Include stdbool.h (provided by GCC 2.95+).
3005         * src/debug.c (ffi_assert): Accept stringized failed assertion.
3006         (ffi_type_test): Rewritten.
3007         * src/prep-cif.c (initialize_aggregate, ffi_prep_cif): Call
3008         FFI_ASSERT_VALID_TYPE.
3009         * src/alpha/ffitarget.h, src/arm/ffitarget.h,
3010         src/ia64/ffitarget.h, src/m68k/ffitarget.h,
3011         src/mips/ffitarget.h, src/powerpc/ffitarget.h,
3012         src/s390/ffitarget.h, src/sh/ffitarget.h,
3013         src/sh64/ffitarget.h, src/sparc/ffitarget.h,
3014         src/x86/ffitarget.h: New files.
3015         * src/alpha/osf.S, src/arm/sysv.S, src/ia64/unix.S,
3016         src/m68k/sysv.S, src/mips/n32.S, src/mips/o32.S,
3017         src/powerpc/aix.S, src/powerpc/darwin.S,
3018         src/powerpc/ffi_darwin.c, src/powerpc/linux64.S,
3019         src/powerpc/linux64_closure.S, src/powerpc/ppc_closure.S,
3020         src/powerpc/sysv.S, src/s390/sysv.S, src/sh/sysv.S,
3021         src/sh64/sysv.S, src/sparc/v8.S, src/sparc/v9.S,
3022         src/x86/sysv.S, src/x86/unix64.S, src/x86/win32.S:
3023         include fficonfig.h
3024
3025 2003-10-20  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3026
3027         * src/mips/ffi.c: Use _ABIN32, _ABIO32 instead of external
3028         _MIPS_SIM_NABI32, _MIPS_SIM_ABI32.
3029
3030 2003-10-19  Andreas Tobler  <a.tobler@schweiz.ch>
3031
3032         * src/powerpc/ffi_darwin.c (ffi_prep_args): Declare bytes again.
3033         Used when FFI_DEBUG = 1.
3034
3035 2003-10-14  Alan Modra  <amodra@bigpond.net.au>
3036
3037         * src/types.c (double, longdouble): Default POWERPC64 to 8 byte size
3038         and align.
3039
3040 2003-10-06  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3041
3042         * include/ffi_mips.h: Define FFI_MIPS_N32 for N32/N64 ABIs,
3043         FFI_MIPS_O32 for O32 ABI.
3044
3045 2003-10-01  Andreas Tobler  <a.tobler@schweiz.ch>
3046
3047         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_64 for
3048         SPARC64. Cleanup whitespaces.
3049
3050 2003-09-19  Andreas Tobler  <a.tobler@schweiz.ch>
3051
3052         * testsuite/libffi.call/closure_fn0.c: Xfail mips, arm,
3053         strongarm, xscale. Cleanup whitespaces.
3054         * testsuite/libffi.call/closure_fn1.c: Likewise.
3055         * testsuite/libffi.call/closure_fn2.c: Likewise.
3056         * testsuite/libffi.call/closure_fn3.c: Likewise.
3057         * testsuite/libffi.call/cls_12byte.c: Likewise.
3058         * testsuite/libffi.call/cls_16byte.c: Likewise.
3059         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3060         * testsuite/libffi.call/cls_20byte.c: Likewise.
3061         * testsuite/libffi.call/cls_24byte.c: Likewise.
3062         * testsuite/libffi.call/cls_2byte.c: Likewise.
3063         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3064         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3065         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3066         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3067         * testsuite/libffi.call/cls_4byte.c: Likewise.
3068         * testsuite/libffi.call/cls_5byte.c: Likewise.
3069         * testsuite/libffi.call/cls_6byte.c: Likewise.
3070         * testsuite/libffi.call/cls_7byte.c: Likewise.
3071         * testsuite/libffi.call/cls_8byte.c: Likewise.
3072         * testsuite/libffi.call/cls_double.c: Likewise.
3073         * testsuite/libffi.call/cls_float.c: Likewise.
3074         * testsuite/libffi.call/cls_uchar.c: Likewise.
3075         * testsuite/libffi.call/cls_uint.c: Likewise.
3076         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3077         * testsuite/libffi.call/cls_ushort.c: Likewise.
3078         * testsuite/libffi.call/nested_struct.c: Likewise.
3079         * testsuite/libffi.call/nested_struct1.c: Likewise.
3080         * testsuite/libffi.call/problem1.c: Likewise.
3081         * testsuite/libffi.special/unwindtest.cc: Likewise.
3082         * testsuite/libffi.call/pyobjc-tc.c: Cleanup whitespaces.
3083
3084 2003-09-18  David Edelsohn  <edelsohn@gnu.org>
3085
3086         * src/powerpc/aix.S: Cleanup whitespaces.
3087         * src/powerpc/aix_closure.S: Likewise.
3088
3089 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3090
3091         * src/powerpc/darwin.S: Cleanup whitespaces, comment formatting.
3092         * src/powerpc/darwin_closure.S: Likewise.
3093         * src/powerpc/ffi_darwin.c: Likewise.
3094
3095 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3096             David Edelsohn  <edelsohn@gnu.org>
3097
3098         * src/types.c (double): Add AIX and Darwin to the right TYPEDEF.
3099         * src/powerpc/aix_closure.S: Remove the pointer to the outgoing
3100         parameter stack.
3101         * src/powerpc/darwin_closure.S: Likewise.
3102         * src/powerpc/ffi_darwin.c (ffi_prep_args): Handle structures
3103         according to the Darwin/AIX ABI.
3104         (ffi_prep_cif_machdep): Likewise.
3105         (ffi_closure_helper_DARWIN): Likewise.
3106         Remove the outgoing parameter stack logic. Simplify the evaluation
3107         of the different CASE types.
3108         (ffi_prep_clousure): Avoid the casts on lvalues. Change the branch
3109         statement in the trampoline code.
3110
3111 2003-09-18  Kaz Kojima  <kkojima@gcc.gnu.org>
3112
3113         * src/sh/ffi.c (ffi_prep_args): Take account into the alignement
3114         for the register size.
3115         (ffi_closure_helper_SYSV): Handle the structure return value
3116         address correctly.
3117         (ffi_closure_helper_SYSV): Return the appropriate type when
3118         the registers are used for the structure return value.
3119         * src/sh/sysv.S (ffi_closure_SYSV): Fix the stack layout for
3120         the 64-bit return value.  Update copyright years.
3121
3122 2003-09-17  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3123
3124         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Search in
3125         srcdir for ffi_mips.h.
3126
3127 2003-09-12  Alan Modra  <amodra@bigpond.net.au>
3128
3129         * src/prep_cif.c (initialize_aggregate): Include tail padding in
3130         structure size.
3131         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Correct
3132         placement of float result.
3133         * testsuite/libffi.special/unwindtest.cc (closure_test_fn1): Correct
3134         cast of "resp" for big-endian 64 bit machines.
3135
3136 2003-09-11  Alan Modra  <amodra@bigpond.net.au>
3137
3138         * src/types.c (double, longdouble): Merge identical SH and ARM
3139         typedefs, and add POWERPC64.
3140         * src/powerpc/ffi.c (ffi_prep_args64): Correct next_arg calc for
3141         struct split over gpr and rest.
3142         (ffi_prep_cif_machdep): Correct intarg_count for structures.
3143         * src/powerpc/linux64.S (ffi_call_LINUX64): Fix gpr offsets.
3144
3145 2003-09-09  Andreas Tobler  <a.tobler@schweiz.ch>
3146
3147         * src/powerpc/ffi.c (ffi_closure_helper_SYSV) Handle struct
3148         passing correctly.
3149
3150 2003-09-09  Alan Modra  <amodra@bigpond.net.au>
3151
3152         * configure: Regenerate.
3153
3154 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3155
3156         * Makefile.am: Remove build rules for ffitest.
3157         * Makefile.in: Rebuilt.
3158
3159 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3160
3161         * src/java_raw_api.c: Include <stdlib.h> to fix compiler warning
3162         about implicit declaration of abort().
3163
3164 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3165
3166         * Makefile.am: Add dejagnu test framework. Fixes PR other/11411.
3167         * Makefile.in: Rebuilt.
3168         * configure.in: Add dejagnu test framework.
3169         * configure: Rebuilt.
3170
3171         * testsuite/Makefile.am: New file.
3172         * testsuite/Makefile.in: Built
3173         * testsuite/lib/libffi-dg.exp: New file.
3174         * testsuite/config/default.exp: Likewise.
3175         * testsuite/libffi.call/call.exp: Likewise.
3176         * testsuite/libffi.call/ffitest.h: Likewise.
3177         * testsuite/libffi.call/closure_fn0.c: Likewise.
3178         * testsuite/libffi.call/closure_fn1.c: Likewise.
3179         * testsuite/libffi.call/closure_fn2.c: Likewise.
3180         * testsuite/libffi.call/closure_fn3.c: Likewise.
3181         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3182         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3183         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3184         * testsuite/libffi.call/cls_2byte.c: Likewise.
3185         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3186         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3187         * testsuite/libffi.call/cls_4byte.c: Likewise.
3188         * testsuite/libffi.call/cls_5byte.c: Likewise.
3189         * testsuite/libffi.call/cls_6byte.c: Likewise.
3190         * testsuite/libffi.call/cls_7byte.c: Likewise.
3191         * testsuite/libffi.call/cls_8byte.c: Likewise.
3192         * testsuite/libffi.call/cls_12byte.c: Likewise.
3193         * testsuite/libffi.call/cls_16byte.c: Likewise.
3194         * testsuite/libffi.call/cls_20byte.c: Likewise.
3195         * testsuite/libffi.call/cls_24byte.c: Likewise.
3196         * testsuite/libffi.call/cls_double.c: Likewise.
3197         * testsuite/libffi.call/cls_float.c: Likewise.
3198         * testsuite/libffi.call/cls_uchar.c: Likewise.
3199         * testsuite/libffi.call/cls_uint.c: Likewise.
3200         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3201         * testsuite/libffi.call/cls_ushort.c: Likewise.
3202         * testsuite/libffi.call/float.c: Likewise.
3203         * testsuite/libffi.call/float1.c: Likewise.
3204         * testsuite/libffi.call/float2.c: Likewise.
3205         * testsuite/libffi.call/many.c: Likewise.
3206         * testsuite/libffi.call/many_win32.c: Likewise.
3207         * testsuite/libffi.call/nested_struct.c: Likewise.
3208         * testsuite/libffi.call/nested_struct1.c: Likewise.
3209         * testsuite/libffi.call/pyobjc-tc.c: Likewise.
3210         * testsuite/libffi.call/problem1.c: Likewise.
3211         * testsuite/libffi.call/promotion.c: Likewise.
3212         * testsuite/libffi.call/return_ll.c: Likewise.
3213         * testsuite/libffi.call/return_sc.c: Likewise.
3214         * testsuite/libffi.call/return_uc.c: Likewise.
3215         * testsuite/libffi.call/strlen.c: Likewise.
3216         * testsuite/libffi.call/strlen_win32.c: Likewise.
3217         * testsuite/libffi.call/struct1.c: Likewise.
3218         * testsuite/libffi.call/struct2.c: Likewise.
3219         * testsuite/libffi.call/struct3.c: Likewise.
3220         * testsuite/libffi.call/struct4.c: Likewise.
3221         * testsuite/libffi.call/struct5.c: Likewise.
3222         * testsuite/libffi.call/struct6.c: Likewise.
3223         * testsuite/libffi.call/struct7.c: Likewise.
3224         * testsuite/libffi.call/struct8.c: Likewise.
3225         * testsuite/libffi.call/struct9.c: Likewise.
3226         * testsuite/libffi.special/special.exp: New file.
3227         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3228         * testsuite/libffi.special/unwindtest.cc: Likewise.
3229
3230
3231 2003-08-13  Kaz Kojima  <kkojima@gcc.gnu.org>
3232
3233         * src/sh/ffi.c (OFS_INT16): Set 0 for little endian case.  Update
3234         copyright years.
3235
3236 2003-08-02  Alan Modra  <amodra@bigpond.net.au>
3237
3238         * src/powerpc/ffi.c (ffi_prep_args64): Modify for changed gcc
3239         structure passing.
3240         (ffi_closure_helper_LINUX64): Likewise.
3241         * src/powerpc/linux64.S: Remove code writing to parm save area.
3242         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Use return
3243         address in lr from ffi_closure_helper_LINUX64 call to calculate
3244         table address.  Optimize function tail.
3245
3246 2003-07-28  Andreas Tobler  <a.tobler@schweiz.ch>
3247
3248         * src/sparc/ffi.c: Handle all floating point registers.
3249         * src/sparc/v9.S: Likewise. Fixes second part of PR target/11410.
3250
3251 2003-07-11  Gerald Pfeifer  <pfeifer@dbai.tuwien.ac.at>
3252
3253         * README: Note that libffi is not part of GCC.  Update the project
3254         URL and status.
3255
3256 2003-06-19  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3257
3258         * src/powerpc/ppc_closure.S: Include ffi.h.
3259
3260 2003-06-13  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3261
3262         * src/x86/sysv.S: Avoid gas-only .uleb128/.sleb128 directives.
3263         Use C style comments.
3264
3265 2003-06-13  Kaz Kojima  <kkojima@rr.iij4u.or.jp>
3266
3267         * Makefile.am: Add SHmedia support.  Fix a typo of SH support.
3268         * Makefile.in: Regenerate.
3269         * configure.in (sh64-*-linux*, sh5*-*-linux*): Add target.
3270         * configure: Regenerate.
3271         * include/ffi.h.in: Add SHmedia support.
3272         * src/sh64/ffi.c: New file.
3273         * src/sh64/sysv.S: New file.
3274
3275 2003-05-16  Jakub Jelinek  <jakub@redhat.com>
3276
3277         * configure.in (HAVE_RO_EH_FRAME): Check whether .eh_frame section
3278         should be read-only.
3279         * configure: Rebuilt.
3280         * fficonfig.h.in: Rebuilt.
3281         * include/ffi.h.in (EH_FRAME_FLAGS): Define.
3282         * src/alpha/osf.S: Use EH_FRAME_FLAGS.
3283         * src/powerpc/linux64.S: Likewise.
3284         * src/powerpc/linux64_closure.S: Likewise.  Include ffi.h.
3285         * src/powerpc/sysv.S: Use EH_FRAME_FLAGS.  Use pcrel encoding
3286         if -fpic/-fPIC/-mrelocatable.
3287         * src/powerpc/powerpc_closure.S: Likewise.
3288         * src/sparc/v8.S: If HAVE_RO_EH_FRAME is defined, don't include
3289         #write in .eh_frame flags.
3290         * src/sparc/v9.S: Likewise.
3291         * src/x86/unix64.S: Use EH_FRAME_FLAGS.
3292         * src/x86/sysv.S: Likewise.  Use pcrel encoding if -fpic/-fPIC.
3293         * src/s390/sysv.S: Use EH_FRAME_FLAGS.  Include ffi.h.
3294
3295 2003-05-07  Jeff Sturm  <jsturm@one-point.com>
3296
3297         Fixes PR bootstrap/10656
3298         * configure.in (HAVE_AS_REGISTER_PSEUDO_OP): Test assembler
3299         support for .register pseudo-op.
3300         * src/sparc/v8.S: Use it.
3301         * fficonfig.h.in: Rebuilt.
3302         * configure: Rebuilt.
3303
3304 2003-04-18  Jakub Jelinek  <jakub@redhat.com>
3305
3306         * include/ffi.h.in (POWERPC64): Define if 64-bit.
3307         (enum ffi_abi): Add FFI_LINUX64 on POWERPC.
3308         Make it the default on POWERPC64.
3309         (FFI_TRAMPOLINE_SIZE): Define to 24 on POWERPC64.
3310         * configure.in: Change powerpc-*-linux* into powerpc*-*-linux*.
3311         * configure: Rebuilt.
3312         * src/powerpc/ffi.c (hidden): Define.
3313         (ffi_prep_args_SYSV): Renamed from
3314         ffi_prep_args.  Cast pointers to unsigned long to shut up warnings.
3315         (NUM_GPR_ARG_REGISTERS64, NUM_FPR_ARG_REGISTERS64,
3316         ASM_NEEDS_REGISTERS64): New.
3317         (ffi_prep_args64): New function.
3318         (ffi_prep_cif_machdep): Handle FFI_LINUX64 ABI.
3319         (ffi_call): Likewise.
3320         (ffi_prep_closure): Likewise.
3321         (flush_icache): Surround by #ifndef POWERPC64.
3322         (ffi_dblfl): New union type.
3323         (ffi_closure_helper_SYSV): Use it to avoid aliasing problems.
3324         (ffi_closure_helper_LINUX64): New function.
3325         * src/powerpc/ppc_closure.S: Surround whole file by #ifndef
3326         __powerpc64__.
3327         * src/powerpc/sysv.S: Likewise.
3328         (ffi_call_SYSV): Rename ffi_prep_args to ffi_prep_args_SYSV.
3329         * src/powerpc/linux64.S: New file.
3330         * src/powerpc/linux64_closure.S: New file.
3331         * Makefile.am (EXTRA_DIST): Add src/powerpc/linux64.S and
3332         src/powerpc/linux64_closure.S.
3333         (TARGET_SRC_POWERPC): Likewise.
3334
3335         * src/ffitest.c (closure_test_fn, closure_test_fn1, closure_test_fn2,
3336         closure_test_fn3): Fix result printing on big-endian 64-bit
3337         machines.
3338         (main): Print tst2_arg instead of uninitialized tst2_result.
3339
3340         * src/ffitest.c (main): Hide what closure pointer really points to
3341         from the compiler.
3342
3343 2003-04-16  Richard Earnshaw  <rearnsha@arm.com>
3344
3345         * configure.in (arm-*-netbsdelf*): Add configuration.
3346         (configure): Regenerated.
3347
3348 2003-04-04  Loren J. Rittle  <ljrittle@acm.org>
3349
3350         * include/Makefile.in: Regenerate.
3351
3352 2003-03-21  Zdenek Dvorak  <rakdver@atrey.karlin.mff.cuni.cz>
3353
3354         * libffi/include/ffi.h.in: Define X86 instead of X86_64 in 32
3355         bit mode.
3356         * libffi/src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV):
3357         Receive closure pointer through parameter, read args using
3358         __builtin_dwarf_cfa.
3359         (FFI_INIT_TRAMPOLINE): Send closure reference through eax.
3360
3361 2003-03-12  Andreas Schwab  <schwab@suse.de>
3362
3363         * configure.in: Avoid trailing /. in toolexeclibdir.
3364         * configure: Rebuilt.
3365
3366 2003-03-03  Andreas Tobler <a.tobler@schweiz.ch>
3367
3368         * src/powerpc/darwin_closure.S: Recode to fit dynamic libraries.
3369
3370 2003-02-06  Andreas Tobler <a.tobler@schweiz.ch>
3371
3372         * libffi/src/powerpc/darwin_closure.S:
3373         Fix alignement bug, allocate 8 bytes for the result.
3374         * libffi/src/powerpc/aix_closure.S:
3375         Likewise.
3376         * libffi/src/powerpc/ffi_darwin.c:
3377         Update stackframe description for aix/darwin_closure.S.
3378
3379 2003-02-06  Jakub Jelinek  <jakub@redhat.com>
3380
3381         * src/s390/ffi.c (ffi_closure_helper_SYSV): Add hidden visibility
3382         attribute.
3383
3384 2003-01-31  Christian Cornelssen  <ccorn@cs.tu-berlin.de>,
3385             Andreas Schwab  <schwab@suse.de>
3386
3387         * configure.in: Adjust command to source config-ml.in to account
3388         for changes to the libffi_basedir definition.
3389         (libffi_basedir): Remove ${srcdir} from value and include trailing
3390         slash if nonempty.
3391
3392         * configure: Regenerate.
3393
3394 2003-01-29  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3395
3396         * src/powerpc/ppc_closure.S: Recode to fit shared libs.
3397
3398 2003-01-28  Andrew Haley  <aph@redhat.com>
3399
3400         * include/ffi.h.in: Enable FFI_CLOSURES for x86_64.
3401         * src/x86/ffi64.c (ffi_prep_closure): New.
3402         (ffi_closure_UNIX64_inner): New.
3403         * src/x86/unix64.S (ffi_closure_UNIX64): New.
3404
3405 2003-01-27  Alexandre Oliva  <aoliva@redhat.com>
3406
3407         * configure.in (toolexecdir, toolexeclibdir): Set and AC_SUBST.
3408         Remove USE_LIBDIR conditional.
3409         * Makefile.am (toolexecdir, toolexeclibdir): Don't override.
3410         * Makefile.in, configure: Rebuilt.
3411
3412 2003-01027  David Edelsohn  <edelsohn@gnu.org>
3413
3414         * Makefile.am (TARGET_SRC_POWERPC_AIX): Fix typo.
3415         * Makefile.in: Regenerate.
3416
3417 2003-01-22  Andrew Haley  <aph@redhat.com>
3418
3419         * src/powerpc/darwin.S (_ffi_call_AIX): Add Augmentation size to
3420         unwind info.
3421
3422 2003-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
3423
3424         * src/powerpc/darwin.S: Add unwind info.
3425         * src/powerpc/darwin_closure.S: Likewise.
3426
3427 2003-01-14  Andrew Haley  <aph@redhat.com>
3428
3429         * src/x86/ffi64.c (ffi_prep_args): Check for void retval.
3430         (ffi_prep_cif_machdep): Likewise.
3431         * src/x86/unix64.S: Add unwind info.
3432
3433 2003-01-14  Andreas Jaeger  <aj@suse.de>
3434
3435         * src/ffitest.c (main): Only use ffi_closures if those are
3436         supported.
3437
3438 2003-01-13 Andreas Tobler <a.tobler@schweiz.ch>
3439