libffi/ChangeLog

   1 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
   2
   3         * src/sparc/ffi.c (ffi_closure_sparc_inner_v9): Properly align 'long
   4         double' arguments.
   5
   6 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
   7
   8         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on Solaris < 10.
   9
  10 2009-12-10  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
  11
  12         PR libffi/40700
  13         * src/closures.c [X86_64 && __sun__ && __svr4__]
  14         (FFI_MMAP_EXEC_WRIT): Define.
  15
  16 2009-12-08  David Daney  <ddaney@caviumnetworks.com>
  17
  18         * testsuite/libffi.call/stret_medium.c: Remove xfail for mips*-*-*
  19         * testsuite/libffi.call/cls_align_longdouble_split2.c: Same.
  20         * testsuite/libffi.call/stret_large.c: Same.
  21         * testsuite/libffi.call/cls_align_longdouble_split.c: Same.
  22         * testsuite/libffi.call/stret_large2.c: Same.
  23         * testsuite/libffi.call/stret_medium2.c: Same.
  24
  25 2009-12-07  David Edelsohn  <edelsohn@gnu.org>
  26
  27         * src/powerpc/aix_closure.S (libffi_closure_ASM): Fix tablejump
  28         typo.
  29
  30 2009-12-05  David Edelsohn  <edelsohn@gnu.org>
  31
  32         * src/powerpc/aix.S: Update AIX32 code to be consistent with AIX64
  33         code.
  34         * src/powerpc/aix_closure.S: Same.
  35
  36 2009-12-05  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
  37
  38         * Makefile.in: Regenerate.
  39         * configure: Regenerate.
  40         * include/Makefile.in: Regenerate.
  41         * man/Makefile.in: Regenerate.
  42         * testsuite/Makefile.in: Regenerate.
  43
  44 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
  45
  46         * src/powerpc/aix_closure.S: Reorganize 64-bit code to match
  47         linux64_closure.S.
  48
  49 2009-12-04  Uros Bizjak  <ubizjak@gmail.com>
  50
  51         PR libffi/41908
  52         * src/x86/ffi64.c (classify_argument): Update from
  53         gcc/config/i386/i386.c.
  54         (ffi_closure_unix64_inner): Do not use the address of two consecutive
  55         SSE registers directly.
  56         * testsuite/libffi.call/cls_dbls_struct.c (main): Remove xfail
  57         for x86_64 linux targets.
  58
  59 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
  60
  61         * src/powerpc/ffi_darwin.c (ffi_closure_helper_DARWIN): Increment
  62         pfr for long double split between fpr13 and stack.
  63
  64 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
  65
  66         * src/powerpc/ffi_darwin.c (ffi_prep_args): Increment next_arg and
  67         fparg_count twice for long double.
  68
  69 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
  70
  71         PR libffi/42243
  72         * src/powerpc/ffi_darwin.c (ffi_prep_args): Remove extra parentheses.
  73
  74 2009-12-03  Uros Bizjak  <ubizjak@gmail.com>
  75
  76         * testsuite/libffi.call/cls_longdouble_va.c (main): Fix format string.
  77         Remove xfails for x86 linux targets.
  78
  79 2009-12-02  David Edelsohn  <edelsohn@gnu.org>
  80
  81         * src/powerpc/ffi_darwin.c (ffi_prep_args): Fix typo in INT64
  82         case.
  83
  84 2009-12-01  David Edelsohn  <edelsohn@gnu.org>
  85
  86         * src/powerpc/aix.S (ffi_call_AIX): Convert to more standard
  87         register usage.  Call ffi_prep_args directly.  Add long double
  88         return value support.
  89         * src/powerpc/ffi_darwin.c (ffi_prep_args): Double arg increment
  90         applies to FFI_TYPE_DOUBLE.  Correct fpr_base increment typo.
  91         Separate FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases.
  92         (ffi_prep_cif_machdep): Only 16 byte stack alignment in 64 bit
  93         mode.
  94         (ffi_closure_helper_DARWIN): Remove nf and ng counters.  Move temp
  95         into case.
  96         * src/powerpc/aix_closure.S: Maintain 16 byte stack alignment.
  97         Allocate result area between params and FPRs.
  98
  99 2009-11-30  David Edelsohn  <edelsohn@gnu.org>
 100
 101         PR target/35484
 102         * src/powerpc/ffitarget.h (POWERPC64): Define for PPC64 Linux and
 103         AIX64.
 104         * src/powerpc/aix.S: Implement AIX64 version.
 105         * src/powerpc/aix_closure.S: Implement AIX64 version.
 106         (ffi_closure_ASM): Use extsb, lha and displament addresses.
 107         * src/powerpc/ffi_darwin.c (ffi_prep_args): Implement AIX64
 108         support.
 109         (ffi_prep_cif_machdep): Same.
 110         (ffi_call): Same.
 111         (ffi_closure_helper_DARWIN): Same.
 112
 113 2009-11-02  Andreas Tobler  <a.tobler@schweiz.org>
 114
 115         PR libffi/41908
 116         * testsuite/libffi.call/testclosure.c: New test.
 117
 118 2009-09-28  Kai Tietz  <kai.tietz@onevision.com>
 119
 120         * src/x86/win64.S (_ffi_call_win64 stack): Remove for gnu
 121         assembly version use of ___chkstk.
 122
 123 2009-09-23  Matthias Klose  <doko@ubuntu.com>
 124
 125         PR libffi/40242, PR libffi/41443
 126         * src/arm/sysv.S (__ARM_ARCH__): Define for processors
 127         __ARM_ARCH_6T2__, __ARM_ARCH_6M__, __ARM_ARCH_7__,
 128         __ARM_ARCH_7A__, __ARM_ARCH_7R__, __ARM_ARCH_7M__.
 129         Change the conditionals to __SOFTFP__ || __ARM_EABI__
 130         for -mfloat-abi=softfp to work.
 131
 132 2009-09-17  Loren J. Rittle  <ljrittle@acm.org>
 133
 134         PR testsuite/32843 (strikes again)
 135         * src/x86/ffi.c (ffi_prep_cif_machdep): Add X86_FREEBSD to
 136         enable proper extension on char and short.
 137
 138 2009-09-15  David Daney  <ddaney@caviumnetworks.com>
 139
 140         * src/java_raw_api.c (ffi_java_raw_to_rvalue): Remove special
 141         handling for FFI_TYPE_POINTER.
 142         * src/mips/ffitarget.h (FFI_TYPE_STRUCT_D_SOFT,
 143         FFI_TYPE_STRUCT_F_SOFT, FFI_TYPE_STRUCT_DD_SOFT,
 144         FFI_TYPE_STRUCT_FF_SOFT, FFI_TYPE_STRUCT_FD_SOFT,
 145         FFI_TYPE_STRUCT_DF_SOFT, FFI_TYPE_STRUCT_SOFT): New defines.
 146         (FFI_N32_SOFT_FLOAT, FFI_N64_SOFT_FLOAT): New ffi_abi enumerations.
 147         (enum ffi_abi): Set FFI_DEFAULT_ABI for soft-float.
 148         * src/mips/n32.S (ffi_call_N32): Add handling for soft-float
 149         structure and pointer returns.
 150         (ffi_closure_N32): Add handling for pointer returns.
 151         * src/mips/ffi.c (ffi_prep_args, calc_n32_struct_flags,
 152         calc_n32_return_struct_flags): Handle soft-float.
 153         (ffi_prep_cif_machdep):  Handle soft-float, fix pointer handling.
 154         (ffi_call_N32): Declare proper argument types.
 155         (ffi_call, copy_struct_N32, ffi_closure_mips_inner_N32): Handle
 156         soft-float.
 157
 158 2009-08-24  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 159
 160         * configure.ac (AC_PREREQ): Bump to 2.64.
 161
 162 2009-08-22  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 163
 164         * Makefile.am (install-html, install-pdf): Remove.
 165         * Makefile.in: Regenerate.
 166
 167         * Makefile.in: Regenerate.
 168         * aclocal.m4: Regenerate.
 169         * configure: Regenerate.
 170         * fficonfig.h.in: Regenerate.
 171         * include/Makefile.in: Regenerate.
 172         * man/Makefile.in: Regenerate.
 173         * testsuite/Makefile.in: Regenerate.
 174
 175 2009-07-30  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 176
 177         * configure.ac (_AC_ARG_VAR_PRECIOUS): Use m4_rename_force.
 178
 179 2009-07-24  Dave Korn  <dave.korn.cygwin@gmail.com>
 180
 181         PR libffi/40807
 182         * src/x86/ffi.c (ffi_prep_cif_machdep): Also use sign/zero-extending
 183         return types for X86_WIN32.
 184         * src/x86/win32.S (_ffi_call_SYSV): Handle omitted return types.
 185         (_ffi_call_STDCALL, _ffi_closure_SYSV, _ffi_closure_raw_SYSV,
 186         _ffi_closure_STDCALL): Likewise.
 187
 188         * src/closures.c (is_selinux_enabled): Define to const 0 for Cygwin.
 189         (dlmmap, dlmunmap): Also use these functions on Cygwin.
 190
 191 2009-07-11  Richard Sandiford  <rdsandiford@googlemail.com>
 192
 193         PR testsuite/40699
 194         PR testsuite/40707
 195         PR testsuite/40709
 196         * testsuite/lib/libffi-dg.exp: Revert 2009-07-02, 2009-07-01 and
 197         2009-06-30 commits.
 198
 199 2009-07-01  Richard Sandiford  <r.sandiford@uk.ibm.com>
 200
 201         * testsuite/lib/libffi-dg.exp (libffi-init): Set ld_library_path
 202         to "" before adding paths.  (This reinstates an assignment that
 203         was removed by my 2009-06-30 commit, but changes the initial
 204         value from "." to "".)
 205
 206 2009-07-01  H.J. Lu  <hongjiu.lu@intel.com>
 207
 208         PR testsuite/40601
 209         * testsuite/lib/libffi-dg.exp (libffi-init): Properly set
 210         gccdir.  Adjust ld_library_path for gcc only if gccdir isn't
 211         empty.
 212
 213 2009-06-30  Richard Sandiford  <r.sandiford@uk.ibm.com>
 214
 215         * testsuite/lib/libffi-dg.exp (libffi-init): Don't add "."
 216         to ld_library_path.  Use add_path.  Add just find_libgcc_s
 217         to ld_library_path, not every libgcc multilib directory.
 218
 219 2009-06-16  Wim Lewis  <wiml@hhhh.org>
 220
 221         * src/powerpc/ffi.c: Avoid clobbering cr3 and cr4, which are
 222         supposed to be callee-saved.
 223         * src/powerpc/sysv.S (small_struct_return_value): Fix overrun of
 224         return buffer for odd-size structs.
 225
 226 2009-06-16  Andreas Tobler  <a.tobler@schweiz.org>
 227
 228         PR libffi/40444
 229         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Add
 230         allow_stack_execute for Darwin.
 231
 232 2009-06-16  Andrew Haley  <aph@redhat.com>
 233
 234         * configure.ac (TARGETDIR): Add missing blank lines.
 235         * configure: Regenerate.
 236
 237 2009-06-16  Andrew Haley  <aph@redhat.com>
 238
 239         * testsuite/libffi.call/cls_align_sint64.c,
 240         testsuite/libffi.call/cls_align_uint64.c,
 241         testsuite/libffi.call/cls_longdouble_va.c,
 242         testsuite/libffi.call/cls_ulonglong.c,
 243         testsuite/libffi.call/return_ll1.c,
 244         testsuite/libffi.call/stret_medium2.c: Fix printf format
 245         specifiers.
 246         * testsuite/libffi.call/ffitest.h,
 247         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRIuLL): Define.
 248
 249 2009-06-15  Andrew Haley  <aph@redhat.com>
 250
 251         * testsuite/libffi.call/err_bad_typedef.c: xfail everywhere.
 252         * testsuite/libffi.call/err_bad_abi.c: Likewise.
 253
 254 2009-06-12  Andrew Haley  <aph@redhat.com>
 255
 256         * Makefile.am: Remove info_TEXINFOS.
 257
 258 2009-06-12  Andrew Haley  <aph@redhat.com>
 259
 260         * ChangeLog.libffi: testsuite/libffi.call/cls_align_sint64.c,
 261         testsuite/libffi.call/cls_align_uint64.c,
 262         testsuite/libffi.call/cls_ulonglong.c,
 263         testsuite/libffi.call/return_ll1.c,
 264         testsuite/libffi.call/stret_medium2.c: Fix printf format
 265         specifiers.
 266         testsuite/libffi.special/unwindtest.cc: include stdint.h.
 267
 268 2009-06-11  Timothy Wall  <twall@users.sf.net>
 269
 270         * Makefile.am,
 271         configure.ac,
 272         include/ffi.h.in,
 273         include/ffi_common.h,
 274         src/closures.c,
 275         src/dlmalloc.c,
 276         src/x86/ffi.c,
 277         src/x86/ffitarget.h,
 278         src/x86/win64.S (new),
 279         README: Added win64 support (mingw or MSVC)
 280         * Makefile.in,
 281         include/Makefile.in,
 282         man/Makefile.in,
 283         testsuite/Makefile.in,
 284         configure,
 285         aclocal.m4: Regenerated
 286         * ltcf-c.sh: properly escape cygwin/w32 path
 287         * man/ffi_call.3: Clarify size requirements for return value.
 288         * src/x86/ffi64.c: Fix filename in comment.
 289         * src/x86/win32.S: Remove unused extern.
 290
 291         * testsuite/libffi.call/closure_fn0.c,
 292         testsuite/libffi.call/closure_fn1.c,
 293         testsuite/libffi.call/closure_fn2.c,
 294         testsuite/libffi.call/closure_fn3.c,
 295         testsuite/libffi.call/closure_fn4.c,
 296         testsuite/libffi.call/closure_fn5.c,
 297         testsuite/libffi.call/closure_fn6.c,
 298         testsuite/libffi.call/closure_stdcall.c,
 299         testsuite/libffi.call/cls_12byte.c,
 300         testsuite/libffi.call/cls_16byte.c,
 301         testsuite/libffi.call/cls_18byte.c,
 302         testsuite/libffi.call/cls_19byte.c,
 303         testsuite/libffi.call/cls_1_1byte.c,
 304         testsuite/libffi.call/cls_20byte.c,
 305         testsuite/libffi.call/cls_20byte1.c,
 306         testsuite/libffi.call/cls_24byte.c,
 307         testsuite/libffi.call/cls_2byte.c,
 308         testsuite/libffi.call/cls_3_1byte.c,
 309         testsuite/libffi.call/cls_3byte1.c,
 310         testsuite/libffi.call/cls_3byte2.c,
 311         testsuite/libffi.call/cls_4_1byte.c,
 312         testsuite/libffi.call/cls_4byte.c,
 313         testsuite/libffi.call/cls_5_1_byte.c,
 314         testsuite/libffi.call/cls_5byte.c,
 315         testsuite/libffi.call/cls_64byte.c,
 316         testsuite/libffi.call/cls_6_1_byte.c,
 317         testsuite/libffi.call/cls_6byte.c,
 318         testsuite/libffi.call/cls_7_1_byte.c,
 319         testsuite/libffi.call/cls_7byte.c,
 320         testsuite/libffi.call/cls_8byte.c,
 321         testsuite/libffi.call/cls_9byte1.c,
 322         testsuite/libffi.call/cls_9byte2.c,
 323         testsuite/libffi.call/cls_align_double.c,
 324         testsuite/libffi.call/cls_align_float.c,
 325         testsuite/libffi.call/cls_align_longdouble.c,
 326         testsuite/libffi.call/cls_align_longdouble_split.c,
 327         testsuite/libffi.call/cls_align_longdouble_split2.c,
 328         testsuite/libffi.call/cls_align_pointer.c,
 329         testsuite/libffi.call/cls_align_sint16.c,
 330         testsuite/libffi.call/cls_align_sint32.c,
 331         testsuite/libffi.call/cls_align_sint64.c,
 332         testsuite/libffi.call/cls_align_uint16.c,
 333         testsuite/libffi.call/cls_align_uint32.c,
 334         testsuite/libffi.call/cls_align_uint64.c,
 335         testsuite/libffi.call/cls_dbls_struct.c,
 336         testsuite/libffi.call/cls_double.c,
 337         testsuite/libffi.call/cls_double_va.c,
 338         testsuite/libffi.call/cls_float.c,
 339         testsuite/libffi.call/cls_longdouble.c,
 340         testsuite/libffi.call/cls_longdouble_va.c,
 341         testsuite/libffi.call/cls_multi_schar.c,
 342         testsuite/libffi.call/cls_multi_sshort.c,
 343         testsuite/libffi.call/cls_multi_sshortchar.c,
 344         testsuite/libffi.call/cls_multi_uchar.c,
 345         testsuite/libffi.call/cls_multi_ushort.c,
 346         testsuite/libffi.call/cls_multi_ushortchar.c,
 347         testsuite/libffi.call/cls_pointer.c,
 348         testsuite/libffi.call/cls_pointer_stack.c,
 349         testsuite/libffi.call/cls_schar.c,
 350         testsuite/libffi.call/cls_sint.c,
 351         testsuite/libffi.call/cls_sshort.c,
 352         testsuite/libffi.call/cls_uchar.c,
 353         testsuite/libffi.call/cls_uint.c,
 354         testsuite/libffi.call/cls_ulonglong.c,
 355         testsuite/libffi.call/cls_ushort.c,
 356         testsuite/libffi.call/err_bad_abi.c,
 357         testsuite/libffi.call/err_bad_typedef.c,
 358         testsuite/libffi.call/float2.c,
 359         testsuite/libffi.call/huge_struct.c,
 360         testsuite/libffi.call/nested_struct.c,
 361         testsuite/libffi.call/nested_struct1.c,
 362         testsuite/libffi.call/nested_struct10.c,
 363         testsuite/libffi.call/nested_struct2.c,
 364         testsuite/libffi.call/nested_struct3.c,
 365         testsuite/libffi.call/nested_struct4.c,
 366         testsuite/libffi.call/nested_struct5.c,
 367         testsuite/libffi.call/nested_struct6.c,
 368         testsuite/libffi.call/nested_struct7.c,
 369         testsuite/libffi.call/nested_struct8.c,
 370         testsuite/libffi.call/nested_struct9.c,
 371         testsuite/libffi.call/problem1.c,
 372         testsuite/libffi.call/return_ldl.c,
 373         testsuite/libffi.call/return_ll1.c,
 374         testsuite/libffi.call/stret_large.c,
 375         testsuite/libffi.call/stret_large2.c,
 376         testsuite/libffi.call/stret_medium.c,
 377         testsuite/libffi.call/stret_medium2.c,
 378         testsuite/libffi.special/unwindtest.cc: use ffi_closure_alloc instead
 379         of checking for MMAP.  Use intptr_t instead of long casts.
 380
 381 2009-06-11  Kaz Kojima  <kkojima@gcc.gnu.org>
 382
 383         * testsuite/libffi.call/cls_longdouble_va.c: Add xfail sh*-*-linux-*.
 384         * testsuite/libffi.call/err_bad_abi.c: Add xfail sh*-*-*.
 385         * testsuite/libffi.call/err_bad_typedef.c: Likewise.
 386
 387 2009-06-09  Andrew Haley  <aph@redhat.com>
 388
 389         * src/x86/freebsd.S: Add missing file.
 390
 391 2009-06-08  Andrew Haley  <aph@redhat.com>
 392
 393         Import from libffi 3.0.8:
 394
 395         * doc/libffi.texi: New file.
 396         * doc/libffi.info: Likewise.
 397         * doc/stamp-vti: Likewise.
 398         * man/Makefile.am: New file.
 399         * man/ffi_call.3: New file.
 400
 401         * Makefile.am (EXTRA_DIST): Add src/x86/darwin64.S,
 402         src/dlmalloc.c.
 403         (nodist_libffi_la_SOURCES): Add X86_FREEBSD.
 404
 405         * configure.ac: Bump version to 3.0.8.
 406         parisc*-*-linux*: Add.
 407         i386-*-freebsd* | i386-*-openbsd*: Add.
 408         powerpc-*-beos*: Add.
 409         AM_CONDITIONAL X86_FREEBSD: Add.
 410         AC_CONFIG_FILES: Add man/Makefile.
 411
 412         * include/ffi.h.in (FFI_FN): Change void (*)() to void (*)(void).
 413
 414 2009-06-08  Andrew Haley  <aph@redhat.com>
 415
 416         * README: Import from libffi 3.0.8.
 417
 418 2009-06-08  Andrew Haley  <aph@redhat.com>
 419
 420         * testsuite/libffi.call/err_bad_abi.c: Add xfails.
 421         * testsuite/libffi.call/cls_longdouble_va.c: Add xfails.
 422         * testsuite/libffi.call/cls_dbls_struct.c: Add xfail x86_64-*-linux-*.
 423         * testsuite/libffi.call/err_bad_typedef.c: Add xfails.
 424
 425         * testsuite/libffi.call/stret_medium2.c: Add __UNUSED__ to args.
 426         * testsuite/libffi.call/stret_medium.c: Likewise.
 427         * testsuite/libffi.call/stret_large2.c: Likewise.
 428         * testsuite/libffi.call/stret_large.c:  Likewise.
 429
 430 2008-12-26  Timothy Wall  <twall@users.sf.net>
 431
 432         * testsuite/libffi.call/cls_longdouble.c,
 433         testsuite/libffi.call/cls_longdouble_va.c,
 434         testsuite/libffi.call/cls_align_longdouble.c,
 435         testsuite/libffi.call/cls_align_longdouble_split.c,
 436         testsuite/libffi.call/cls_align_longdouble_split2.c: mark expected
 437         failures on x86_64 cygwin/mingw.
 438
 439 2008-12-22  Timothy Wall  <twall@users.sf.net>
 440
 441         * testsuite/libffi.call/closure_fn0.c,
 442         testsuite/libffi.call/closure_fn1.c,
 443         testsuite/libffi.call/closure_fn2.c,
 444         testsuite/libffi.call/closure_fn3.c,
 445         testsuite/libffi.call/closure_fn4.c,
 446         testsuite/libffi.call/closure_fn5.c,
 447         testsuite/libffi.call/closure_fn6.c,
 448         testsuite/libffi.call/closure_loc_fn0.c,
 449         testsuite/libffi.call/closure_stdcall.c,
 450         testsuite/libffi.call/cls_align_pointer.c,
 451         testsuite/libffi.call/cls_pointer.c,
 452         testsuite/libffi.call/cls_pointer_stack.c: use portable cast from
 453         pointer to integer (intptr_t).
 454         * testsuite/libffi.call/cls_longdouble.c: disable for win64.
 455
 456 2008-07-24  Anthony Green  <green@redhat.com>
 457
 458         * testsuite/libffi.call/cls_dbls_struct.c,
 459         testsuite/libffi.call/cls_double_va.c,
 460         testsuite/libffi.call/cls_longdouble.c,
 461         testsuite/libffi.call/cls_longdouble_va.c,
 462         testsuite/libffi.call/cls_pointer.c,
 463         testsuite/libffi.call/cls_pointer_stack.c,
 464         testsuite/libffi.call/err_bad_abi.c: Clean up failures from
 465         compiler warnings.
 466
 467 2008-03-04  Anthony Green  <green@redhat.com>
 468             Blake Chaffin
 469             hos@tamanegi.org
 470
 471         * testsuite/libffi.call/cls_align_longdouble_split2.c
 472           testsuite/libffi.call/cls_align_longdouble_split.c
 473           testsuite/libffi.call/cls_dbls_struct.c
 474           testsuite/libffi.call/cls_double_va.c
 475           testsuite/libffi.call/cls_longdouble.c
 476           testsuite/libffi.call/cls_longdouble_va.c
 477           testsuite/libffi.call/cls_pointer.c
 478           testsuite/libffi.call/cls_pointer_stack.c
 479           testsuite/libffi.call/err_bad_abi.c
 480           testsuite/libffi.call/err_bad_typedef.c
 481           testsuite/libffi.call/stret_large2.c
 482           testsuite/libffi.call/stret_large.c
 483           testsuite/libffi.call/stret_medium2.c
 484           testsuite/libffi.call/stret_medium.c: New tests from Apple.
 485
 486 2009-06-05  Andrew Haley  <aph@redhat.com>
 487
 488         * src/x86/ffitarget.h, src/x86/ffi.c: Merge stdcall changes from
 489         libffi.
 490
 491 2009-06-04  Andrew Haley  <aph@redhat.com>
 492
 493         * src/x86/ffitarget.h, src/x86/win32.S, src/x86/ffi.c: Back out
 494         stdcall changes.
 495
 496 2008-02-26  Anthony Green  <green@redhat.com>
 497             Thomas Heller  <theller@ctypes.org>
 498
 499         * src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C
 500         comment.
 501
 502 2008-02-03  Timothy Wall  <twall@users.sf.net>
 503
 504         * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return
 505           offset based on code pointer, not data pointer.
 506
 507 2008-01-31  Timothy Wall <twall@users.sf.net>
 508
 509         * testsuite/libffi.call/closure_stdcall.c: Add test for stdcall
 510         closures.
 511         * src/x86/ffitarget.h: Increase size of trampoline for stdcall
 512         closures.
 513         * src/x86/win32.S: Add assembly for stdcall closure.
 514         * src/x86/ffi.c: Initialize stdcall closure trampoline.
 515
 516 2009-06-04  Andrew Haley  <aph@redhat.com>
 517
 518         * include/ffi.h.in: Change void (*)() to void (*)(void).
 519         * src/x86/ffi.c: Likewise.
 520
 521 2009-06-04  Andrew Haley  <aph@redhat.com>
 522
 523         * src/powerpc/ppc_closure.S: Insert licence header.
 524         * src/powerpc/linux64_closure.S: Likewise.
 525         * src/m68k/sysv.S: Likewise.
 526
 527         * src/sh64/ffi.c: Change void (*)() to void (*)(void).
 528         * src/powerpc/ffi.c: Likewise.
 529         * src/powerpc/ffi_darwin.c: Likewise.
 530         * src/m32r/ffi.c: Likewise.
 531         * src/sh64/ffi.c: Likewise.
 532         * src/x86/ffi64.c: Likewise.
 533         * src/alpha/ffi.c: Likewise.
 534         * src/alpha/osf.S: Likewise.
 535         * src/frv/ffi.c: Likewise.
 536         * src/s390/ffi.c: Likewise.
 537         * src/pa/ffi.c: Likewise.
 538         * src/pa/hpux32.S: Likewise.
 539         * src/ia64/unix.S: Likewise.
 540         * src/ia64/ffi.c: Likewise.
 541         * src/sparc/ffi.c: Likewise.
 542         * src/mips/ffi.c: Likewise.
 543         * src/sh/ffi.c: Likewise.
 544
 545 2008-02-15  David Daney  <ddaney@avtrex.com>
 546
 547         * src/mips/ffi.c (USE__BUILTIN___CLEAR_CACHE):
 548         Define (conditionally), and use it to include cachectl.h.
 549         (ffi_prep_closure_loc): Fix cache flushing.
 550         * src/mips/ffitarget.h (_ABIN32, _ABI64, _ABIO32): Define.
 551
 552 2009-06-04  Andrew Haley  <aph@redhat.com>
 553
 554         include/ffi.h.in,
 555         src/arm/ffitarget.h,
 556         src/arm/ffi.c,
 557         src/arm/sysv.S,
 558         src/powerpc/ffitarget.h,
 559         src/closures.c,
 560         src/sh64/ffitarget.h,
 561         src/sh64/ffi.c,
 562         src/sh64/sysv.S,
 563         src/types.c,
 564         src/x86/ffi64.c,
 565         src/x86/ffitarget.h,
 566         src/x86/win32.S,
 567         src/x86/darwin.S,
 568         src/x86/ffi.c,
 569         src/x86/sysv.S,
 570         src/x86/unix64.S,
 571         src/alpha/ffitarget.h,
 572         src/alpha/ffi.c,
 573         src/alpha/osf.S,
 574         src/m68k/ffitarget.h,
 575         src/frv/ffitarget.h,
 576         src/frv/ffi.c,
 577         src/s390/ffitarget.h,
 578         src/s390/sysv.S,
 579         src/cris/ffitarget.h,
 580         src/pa/linux.S,
 581         src/pa/ffitarget.h,
 582         src/pa/ffi.c,
 583         src/raw_api.c,
 584         src/ia64/ffitarget.h,
 585         src/ia64/unix.S,
 586         src/ia64/ffi.c,
 587         src/ia64/ia64_flags.h,
 588         src/java_raw_api.c,
 589         src/debug.c,
 590         src/sparc/v9.S,
 591         src/sparc/ffitarget.h,
 592         src/sparc/ffi.c,
 593         src/sparc/v8.S,
 594         src/mips/ffitarget.h,
 595         src/mips/n32.S,
 596         src/mips/o32.S,
 597         src/mips/ffi.c,
 598         src/prep_cif.c,
 599         src/sh/ffitarget.h,
 600         src/sh/ffi.c,
 601         src/sh/sysv.S: Update license text.
 602
 603 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
 604
 605         * src/x86/win32.S (_ffi_closure_STDCALL):  New function.
 606         (.eh_frame):  Add FDE for it.
 607
 608 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
 609
 610         * configure.ac:  Also check if assembler supports pc-relative
 611         relocs on X86_WIN32 targets.
 612         * configure:  Regenerate.
 613         * src/x86/win32.S (ffi_prep_args):  Declare extern, not global.
 614         (_ffi_call_SYSV):  Add missing function type symbol .def and
 615         add EH markup labels.
 616         (_ffi_call_STDCALL):  Likewise.
 617         (_ffi_closure_SYSV):  Likewise.
 618         (_ffi_closure_raw_SYSV):  Likewise.
 619         (.eh_frame):  Add hand-crafted EH data.
 620
 621 2009-04-09  Jakub Jelinek  <jakub@redhat.com>
 622
 623         * testsuite/lib/libffi-dg.exp: Change copyright header to refer to
 624         version 3 of the GNU General Public License and to point readers
 625         at the COPYING3 file and the FSF's license web page.
 626         * testsuite/libffi.call/call.exp: Likewise.
 627         * testsuite/libffi.special/special.exp: Likewise.
 628
 629 2009-03-01  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 630
 631         * configure: Regenerate.
 632
 633 2008-12-18  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
 634
 635         PR libffi/26048
 636         * configure.ac (HAVE_AS_X86_PCREL): New test.
 637         * configure: Regenerate.
 638         * fficonfig.h.in: Regenerate.
 639         * src/x86/sysv.S [!FFI_NO_RAW_API]: Precalculate
 640         RAW_CLOSURE_CIF_OFFSET, RAW_CLOSURE_FUN_OFFSET,
 641         RAW_CLOSURE_USER_DATA_OFFSET for the Solaris 10/x86 assembler.
 642         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
 643         * src/x86/unix64.S (.Lstore_table): Move to .text section.
 644         (.Lload_table): Likewise.
 645         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
 646
 647 2008-12-18  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 648
 649         * configure: Regenerate.
 650
 651 2008-11-21  Eric Botcazou  <ebotcazou@adacore.com>
 652
 653         * src/sparc/ffi.c (ffi_prep_cif_machdep): Add support for
 654         signed/unsigned int8/16 return values.
 655         * src/sparc/v8.S (ffi_call_v8): Likewise.
 656         (ffi_closure_v8): Likewise.
 657
 658 2008-09-26  Peter O'Gorman  <pogma@thewrittenword.com>
 659             Steve Ellcey  <sje@cup.hp.com>
 660
 661         * configure: Regenerate for new libtool.
 662         * Makefile.in: Ditto.
 663         * include/Makefile.in: Ditto.
 664         * aclocal.m4: Ditto.
 665
 666 2008-08-25  Andreas Tobler  <a.tobler@schweiz.org>
 667
 668         * src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and
 669         FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum.
 670         Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT.
 671         Adjust copyright notice.
 672         * src/powerpc/ffi.c: Add two new flags to indicate if we have one
 673         register or two register to use for FFI_SYSV structs.
 674         (ffi_prep_cif_machdep): Pass the right register flag introduced above.
 675         (ffi_closure_helper_SYSV): Fix the return type for
 676         FFI_SYSV_TYPE_SMALL_STRUCT. Comment.
 677         Adjust copyright notice.
 678
 679 2008-07-16  Kaz Kojima  <kkojima@gcc.gnu.org>
 680
 681         * src/sh/ffi.c (ffi_prep_closure_loc): Turn INSN into an unsigned
 682         int.
 683
 684 2008-06-17  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 685
 686         * configure: Regenerate.
 687         * include/Makefile.in: Regenerate.
 688         * testsuite/Makefile.in: Regenerate.
 689
 690 2008-06-07  Joseph Myers  <joseph@codesourcery.com>
 691
 692         * configure.ac (parisc*-*-linux*, powerpc-*-sysv*,
 693         powerpc-*-beos*): Remove.
 694         * configure: Regenerate.
 695
 696 2008-05-09  Julian Brown  <julian@codesourcery.com>
 697
 698         * Makefile.am (LTLDFLAGS): New.
 699         (libffi_la_LDFLAGS): Use above.
 700         * Makefile.in: Regenerate.
 701
 702 2008-04-18  Paolo Bonzini  <bonzini@gnu.org>
 703
 704         PR bootstrap/35457
 705         * aclocal.m4: Regenerate.
 706         * configure: Regenerate.
 707
 708 2008-03-26  Kaz Kojima  <kkojima@gcc.gnu.org>
 709
 710         * src/sh/sysv.S: Add .note.GNU-stack on Linux.
 711         * src/sh64/sysv.S: Likewise.
 712
 713 2008-03-26  Daniel Jacobowitz  <dan@debian.org>
 714
 715         * src/arm/sysv.S: Fix ARM comment marker.
 716
 717 2008-03-26  Jakub Jelinek  <jakub@redhat.com>
 718
 719         * src/alpha/osf.S: Add .note.GNU-stack on Linux.
 720         * src/s390/sysv.S: Likewise.
 721         * src/powerpc/ppc_closure.S: Likewise.
 722         * src/powerpc/sysv.S: Likewise.
 723         * src/x86/unix64.S: Likewise.
 724         * src/x86/sysv.S: Likewise.
 725         * src/sparc/v8.S: Likewise.
 726         * src/sparc/v9.S: Likewise.
 727         * src/m68k/sysv.S: Likewise.
 728         * src/arm/sysv.S: Likewise.
 729
 730 2008-03-16  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 731
 732         * aclocal.m4: Regenerate.
 733         * configure: Likewise.
 734         * Makefile.in: Likewise.
 735         * include/Makefile.in: Likewise.
 736         * testsuite/Makefile.in: Likewise.
 737
 738 2008-02-12  Bjoern Koenig  <bkoenig@alpha-tierchen.de>
 739             Andreas Tobler  <a.tobler@schweiz.org>
 740
 741         * configure.ac: Add amd64-*-freebsd* target.
 742         * configure: Regenerate.
 743
 744 2008-01-30  H.J. Lu  <hongjiu.lu@intel.com>
 745
 746         PR libffi/34612
 747         * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when
 748         returning struct.
 749
 750         * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer"
 751         tests.
 752
 753 2008-01-24  David Edelsohn  <edelsohn@gnu.org>
 754
 755         * configure: Regenerate.
 756
 757 2008-01-06  Andreas Tobler  <a.tobler@schweiz.org>
 758
 759         * src/x86/ffi.c (ffi_prep_cif_machdep): Fix thinko.
 760
 761 2008-01-05  Andreas Tobler  <a.tobler@schweiz.org>
 762
 763         PR testsuite/32843
 764         * src/x86/ffi.c (ffi_prep_cif_machdep): Add code for
 765         signed/unsigned int8/16 for X86_DARWIN.
 766         Updated copyright info.
 767         Handle one and two byte structs with special cif->flags.
 768         * src/x86/ffitarget.h: Add special types for one and two byte structs.
 769         Updated copyright info.
 770         * src/x86/darwin.S (ffi_call_SYSV): Rewrite to use a jump table like
 771         sysv.S
 772         Remove code to pop args from the stack after call.
 773         Special-case signed/unsigned for int8/16, one and two byte structs.
 774         (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
 775         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
 776         FFI_TYPE_SINT32.
 777         Updated copyright info.
 778
 779 2007-12-08  David Daney  <ddaney@avtrex.com>
 780
 781         * src/mips/n32.S (ffi_call_N32):  Replace dadd with ADDU, dsub with
 782         SUBU, add with ADDU and use smaller code sequences.
 783
 784 2007-12-07  David Daney  <ddaney@avtrex.com>
 785
 786         * src/mips/ffi.c (ffi_prep_cif_machdep): Handle long double return
 787         type.
 788
 789 2007-12-06  David Daney  <ddaney@avtrex.com>
 790
 791         * include/ffi.h.in (FFI_SIZEOF_JAVA_RAW): Define if not already
 792         defined.
 793         (ffi_java_raw): New typedef.
 794         (ffi_java_raw_call, ffi_java_ptrarray_to_raw,
 795         ffi_java_raw_to_ptrarray): Change parameter types from ffi_raw to
 796         ffi_java_raw.
 797         (ffi_java_raw_closure) : Same.
 798         (ffi_prep_java_raw_closure, ffi_prep_java_raw_closure_loc): Change
 799         parameter types.
 800         * src/java_raw_api.c (ffi_java_raw_size):  Replace FFI_SIZEOF_ARG with
 801         FFI_SIZEOF_JAVA_RAW.
 802         (ffi_java_raw_to_ptrarray): Change type of raw to ffi_java_raw.
 803         Replace FFI_SIZEOF_ARG with FFI_SIZEOF_JAVA_RAW. Use
 804         sizeof(ffi_java_raw) for alignment calculations.
 805         (ffi_java_ptrarray_to_raw): Same.
 806         (ffi_java_rvalue_to_raw): Add special handling for FFI_TYPE_POINTER
 807         if FFI_SIZEOF_JAVA_RAW == 4.
 808         (ffi_java_raw_to_rvalue): Same.
 809         (ffi_java_raw_call): Change type of raw to ffi_java_raw.
 810         (ffi_java_translate_args): Same.
 811         (ffi_prep_java_raw_closure_loc, ffi_prep_java_raw_closure): Change
 812         parameter types.
 813         * src/mips/ffitarget.h (FFI_SIZEOF_JAVA_RAW): Define for N32 ABI.
 814
 815 2007-12-06  David Daney  <ddaney@avtrex.com>
 816
 817         * src/mips/n32.S (ffi_closure_N32): Use 64-bit add instruction on
 818         pointer values.
 819
 820 2007-12-01  Andreas Tobler  <a.tobler@schweiz.org>
 821
 822         PR libffi/31937
 823         * src/powerpc/ffitarget.h: Introduce new ABI FFI_LINUX_SOFT_FLOAT.
 824         Add local FFI_TYPE_UINT128 to handle soft-float long-double-128.
 825         * src/powerpc/ffi.c: Distinguish between __NO_FPRS__ and not and
 826         set the NUM_FPR_ARG_REGISTERS according to.
 827         Add support for potential soft-float support under hard-float
 828         architecture.
 829         (ffi_prep_args_SYSV): Set NUM_FPR_ARG_REGISTERS to 0 in case of
 830         FFI_LINUX_SOFT_FLOAT, handle float, doubles and long-doubles according
 831         to the FFI_LINUX_SOFT_FLOAT ABI.
 832         (ffi_prep_cif_machdep): Likewise.
 833         (ffi_closure_helper_SYSV): Likewise.
 834         * src/powerpc/ppc_closure.S: Make sure not to store float/double
 835         on archs where __NO_FPRS__ is true.
 836         Add FFI_TYPE_UINT128 support.
 837         * src/powerpc/sysv.S: Add support for soft-float long-double-128.
 838         Adjust copyright notice.
 839
 840 2007-11-25  Andreas Tobler  <a.tobler@schweiz.org>
 841
 842         * src/closures.c: Move defintion of MAYBE_UNUSED from here to ...
 843         * include/ffi_common.h: ... here.
 844         Update copyright.
 845
 846 2007-11-17  Andreas Tobler  <a.tobler@schweiz.org>
 847
 848         * src/powerpc/sysv.S: Load correct cr to compare if we have long double.
 849         * src/powerpc/linux64.S: Likewise.
 850         * src/powerpc/ffi.c: Add a comment to show which part goes into cr6.
 851         * testsuite/libffi.call/return_ldl.c: New test.
 852
 853 2007-09-04    <aph@redhat.com>
 854
 855         * src/arm/sysv.S (UNWIND): New.
 856         (Whole file): Conditionally compile unwinder directives.
 857         * src/arm/sysv.S: Add unwinder directives.
 858
 859         * src/arm/ffi.c (ffi_prep_args): Align structs by at least 4 bytes.
 860         Only treat r0 as a struct address if we're actually returning a
 861         struct by address.
 862         Only copy the bytes that are actually within a struct.
 863         (ffi_prep_cif_machdep): A Composite Type not larger than 4 bytes
 864         is returned in r0, not passed by address.
 865         (ffi_call): Allocate a word-sized temporary for the case where
 866         a composite is returned in r0.
 867         (ffi_prep_incoming_args_SYSV): Align as necessary.
 868
 869 2007-08-05  Steven Newbury  <s_j_newbury@yahoo.co.uk>
 870
 871         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Use __clear_cache instead of
 872         directly using the sys_cacheflush syscall.
 873
 874 2007-07-27  Andrew Haley  <aph@redhat.com>
 875
 876         * src/arm/sysv.S (ffi_closure_SYSV): Add soft-float.
 877
 878 2007-09-03  Maciej W. Rozycki  <macro@linux-mips.org>
 879
 880         * Makefile.am: Unify MIPS_IRIX and MIPS_LINUX into MIPS.
 881         * configure.ac: Likewise.
 882         * Makefile.in: Regenerate.
 883         * include/Makefile.in: Likewise.
 884         * testsuite/Makefile.in: Likewise.
 885         * configure: Likewise.
 886
 887 2007-08-24  David Daney  <ddaney@avtrex.com>
 888
 889         * testsuite/libffi.call/return_sl.c: New test.
 890
 891 2007-08-10  David Daney  <ddaney@avtrex.com>
 892
 893         * testsuite/libffi.call/cls_multi_ushort.c,
 894         testsuite/libffi.call/cls_align_uint16.c,
 895         testsuite/libffi.call/nested_struct1.c,
 896         testsuite/libffi.call/nested_struct3.c,
 897         testsuite/libffi.call/cls_7_1_byte.c,
 898         testsuite/libffi.call/nested_struct5.c,
 899         testsuite/libffi.call/cls_double.c,
 900         testsuite/libffi.call/nested_struct7.c,
 901         testsuite/libffi.call/cls_sint.c,
 902         testsuite/libffi.call/nested_struct9.c,
 903         testsuite/libffi.call/cls_20byte1.c,
 904         testsuite/libffi.call/cls_multi_sshortchar.c,
 905         testsuite/libffi.call/cls_align_sint64.c,
 906         testsuite/libffi.call/cls_3byte2.c,
 907         testsuite/libffi.call/cls_multi_schar.c,
 908         testsuite/libffi.call/cls_multi_uchar.c,
 909         testsuite/libffi.call/cls_19byte.c,
 910         testsuite/libffi.call/cls_9byte1.c,
 911         testsuite/libffi.call/cls_align_float.c,
 912         testsuite/libffi.call/closure_fn1.c,
 913         testsuite/libffi.call/problem1.c,
 914         testsuite/libffi.call/closure_fn3.c,
 915         testsuite/libffi.call/cls_sshort.c,
 916         testsuite/libffi.call/closure_fn5.c,
 917         testsuite/libffi.call/cls_align_double.c,
 918         testsuite/libffi.call/nested_struct.c,
 919         testsuite/libffi.call/cls_2byte.c,
 920         testsuite/libffi.call/nested_struct10.c,
 921         testsuite/libffi.call/cls_4byte.c,
 922         testsuite/libffi.call/cls_6byte.c,
 923         testsuite/libffi.call/cls_8byte.c,
 924         testsuite/libffi.call/cls_multi_sshort.c,
 925         testsuite/libffi.call/cls_align_sint16.c,
 926         testsuite/libffi.call/cls_align_uint32.c,
 927         testsuite/libffi.call/cls_20byte.c,
 928         testsuite/libffi.call/cls_float.c,
 929         testsuite/libffi.call/nested_struct2.c,
 930         testsuite/libffi.call/cls_5_1_byte.c,
 931         testsuite/libffi.call/nested_struct4.c,
 932         testsuite/libffi.call/cls_24byte.c,
 933         testsuite/libffi.call/nested_struct6.c,
 934         testsuite/libffi.call/cls_64byte.c,
 935         testsuite/libffi.call/nested_struct8.c,
 936         testsuite/libffi.call/cls_uint.c,
 937         testsuite/libffi.call/cls_multi_ushortchar.c,
 938         testsuite/libffi.call/cls_schar.c,
 939         testsuite/libffi.call/cls_uchar.c,
 940         testsuite/libffi.call/cls_align_uint64.c,
 941         testsuite/libffi.call/cls_ulonglong.c,
 942         testsuite/libffi.call/cls_align_longdouble.c,
 943         testsuite/libffi.call/cls_1_1byte.c,
 944         testsuite/libffi.call/cls_12byte.c,
 945         testsuite/libffi.call/cls_3_1byte.c,
 946         testsuite/libffi.call/cls_3byte1.c,
 947         testsuite/libffi.call/cls_4_1byte.c,
 948         testsuite/libffi.call/cls_6_1_byte.c,
 949         testsuite/libffi.call/cls_16byte.c,
 950         testsuite/libffi.call/cls_18byte.c,
 951         testsuite/libffi.call/closure_fn0.c,
 952         testsuite/libffi.call/cls_9byte2.c,
 953         testsuite/libffi.call/closure_fn2.c,
 954         testsuite/libffi.call/closure_fn4.c,
 955         testsuite/libffi.call/cls_ushort.c,
 956         testsuite/libffi.call/closure_fn6.c,
 957         testsuite/libffi.call/cls_5byte.c,
 958         testsuite/libffi.call/cls_align_pointer.c,
 959         testsuite/libffi.call/cls_7byte.c,
 960         testsuite/libffi.call/cls_align_sint32.c,
 961         testsuite/libffi.special/unwindtest_ffi_call.cc,
 962         testsuite/libffi.special/unwindtest.cc: Remove xfail for mips64*-*-*.
 963
 964 2007-08-10  David Daney  <ddaney@avtrex.com>
 965
 966         PR libffi/28313
 967         * configure.ac: Don't treat mips64 as a special case.
 968         * Makefile.am (nodist_libffi_la_SOURCES): Add n32.S.
 969         * configure: Regenerate
 970         * Makefile.in: Ditto.
 971         * fficonfig.h.in: Ditto.
 972         * src/mips/ffitarget.h (REG_L, REG_S, SUBU, ADDU, SRL, LI): Indent.
 973         (LA, EH_FRAME_ALIGN, FDE_ADDR_BYTES): New preprocessor macros.
 974         (FFI_DEFAULT_ABI): Set for n64 case.
 975         (FFI_CLOSURES, FFI_TRAMPOLINE_SIZE): Define for n32 and n64 cases.
 976         * src/mips/n32.S (ffi_call_N32): Add debug macros and labels for FDE.
 977         (ffi_closure_N32): New function.
 978         (.eh_frame): New section
 979         * src/mips/o32.S: Clean up comments.
 980         (ffi_closure_O32): Pass ffi_closure parameter in $12.
 981         * src/mips/ffi.c: Use FFI_MIPS_N32 instead of
 982         _MIPS_SIM == _ABIN32 throughout.
 983         (FFI_MIPS_STOP_HERE): New, use in place of
 984         ffi_stop_here.
 985         (ffi_prep_args): Use unsigned long to hold pointer values.  Rewrite
 986         to support n32/n64 ABIs.
 987         (calc_n32_struct_flags): Rewrite.
 988         (calc_n32_return_struct_flags): Remove unused variable.  Reverse
 989         position of flag bits.
 990         (ffi_prep_cif_machdep): Rewrite n32 portion.
 991         (ffi_call): Enable for n64.  Add special handling for small structure
 992         return values.
 993         (ffi_prep_closure_loc): Add n32 and n64 support.
 994         (ffi_closure_mips_inner_O32): Add cast to silence warning.
 995         (copy_struct_N32, ffi_closure_mips_inner_N32): New functions.
 996
 997 2007-08-08  David Daney  <ddaney@avtrex.com>
 998
 999         * testsuite/libffi.call/ffitest.h (ffi_type_mylong): Remove definition.
1000         * testsuite/libffi.call/cls_align_uint16.c (main): Use correct type
1001         specifiers.
1002         * testsuite/libffi.call/nested_struct1.c (main): Ditto.
1003         * testsuite/libffi.call/cls_sint.c (main): Ditto.
1004         * testsuite/libffi.call/nested_struct9.c (main): Ditto.
1005         * testsuite/libffi.call/cls_20byte1.c (main): Ditto.
1006         * testsuite/libffi.call/cls_9byte1.c (main): Ditto.
1007         * testsuite/libffi.call/closure_fn1.c (main): Ditto.
1008         * testsuite/libffi.call/closure_fn3.c (main): Ditto.
1009         * testsuite/libffi.call/return_dbl2.c (main): Ditto.
1010         * testsuite/libffi.call/cls_sshort.c (main): Ditto.
1011         * testsuite/libffi.call/return_fl3.c (main): Ditto.
1012         * testsuite/libffi.call/closure_fn5.c (main): Ditto.
1013         * testsuite/libffi.call/nested_struct.c (main): Ditto.
1014         * testsuite/libffi.call/nested_struct10.c (main): Ditto.
1015         * testsuite/libffi.call/return_ll1.c (main): Ditto.
1016         * testsuite/libffi.call/cls_8byte.c (main): Ditto.
1017         * testsuite/libffi.call/cls_align_uint32.c (main): Ditto.
1018         * testsuite/libffi.call/cls_align_sint16.c (main): Ditto.
1019         * testsuite/libffi.call/cls_20byte.c (main): Ditto.
1020         * testsuite/libffi.call/nested_struct2.c (main): Ditto.
1021         * testsuite/libffi.call/cls_24byte.c (main): Ditto.
1022         * testsuite/libffi.call/nested_struct6.c (main): Ditto.
1023         * testsuite/libffi.call/cls_uint.c (main): Ditto.
1024         * testsuite/libffi.call/cls_12byte.c (main): Ditto.
1025         * testsuite/libffi.call/cls_16byte.c (main): Ditto.
1026         * testsuite/libffi.call/closure_fn0.c (main): Ditto.
1027         * testsuite/libffi.call/cls_9byte2.c (main): Ditto.
1028         * testsuite/libffi.call/closure_fn2.c (main): Ditto.
1029         * testsuite/libffi.call/return_dbl1.c (main): Ditto.
1030         * testsuite/libffi.call/closure_fn4.c (main): Ditto.
1031         * testsuite/libffi.call/closure_fn6.c (main): Ditto.
1032         * testsuite/libffi.call/cls_align_sint32.c (main): Ditto.
1033
1034 2007-08-07  Andrew Haley  <aph@redhat.com>
1035
1036         * src/x86/sysv.S (ffi_closure_raw_SYSV): Fix typo in previous
1037         checkin.
1038
1039 2007-08-06  Andrew Haley  <aph@redhat.com>
1040
1041         PR testsuite/32843
1042         * src/x86/sysv.S (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1043         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1044         FFI_TYPE_SINT32.
1045
1046 2007-08-02  David Daney  <ddaney@avtrex.com>
1047
1048         * testsuite/libffi.call/return_ul.c (main): Define return type as
1049         ffi_arg.  Use proper printf conversion specifier.
1050
1051 2007-07-30  Andrew Haley  <aph@redhat.com>
1052
1053         PR testsuite/32843
1054         * src/x86/ffi.c (ffi_prep_cif_machdep): in x86 case, add code for
1055         signed/unsigned int8/16.
1056         * src/x86/sysv.S (ffi_call_SYSV): Rewrite to:
1057         Use a jump table.
1058         Remove code to pop args from the stack after call.
1059         Special-case signed/unsigned int8/16.
1060         * testsuite/libffi.call/return_sc.c (main): Revert.
1061
1062 2007-07-26  Richard Guenther  <rguenther@suse.de>
1063
1064         PR testsuite/32843
1065         * testsuite/libffi.call/return_sc.c (main): Verify call
1066         result as signed char, not ffi_arg.
1067
1068 2007-07-16  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1069
1070         * configure.ac (i?86-*-solaris2.1[0-9]): Set TARGET to X86_64.
1071         * configure: Regenerate.
1072
1073 2007-07-11  David Daney  <ddaney@avtrex.com>
1074
1075         * src/mips/ffi.c: Don't include sys/cachectl.h.
1076         (ffi_prep_closure_loc): Use __builtin___clear_cache() instead of
1077         cacheflush().
1078
1079 2007-05-18  Aurelien Jarno  <aurelien@aurel32.net>
1080
1081         * src/arm/ffi.c (ffi_prep_closure_loc): Renamed and ajusted
1082         from (ffi_prep_closure): ... this.
1083         (FFI_INIT_TRAMPOLINE): Adjust.
1084
1085 2005-12-31  Phil Blundell  <pb@reciva.com>
1086
1087         * src/arm/ffi.c (ffi_prep_incoming_args_SYSV,
1088         ffi_closure_SYSV_inner, ffi_prep_closure): New, add closure support.
1089         * src/arm/sysv.S(ffi_closure_SYSV): Likewise.
1090         * src/arm/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1091         (FFI_CLOSURES): Enable closure support.
1092
1093 2007-07-03  Andrew Haley  <aph@hedges.billgatliff.com>
1094
1095         * testsuite/libffi.call/cls_multi_ushort.c,
1096         testsuite/libffi.call/cls_align_uint16.c,
1097         testsuite/libffi.call/nested_struct1.c,
1098         testsuite/libffi.call/nested_struct3.c,
1099         testsuite/libffi.call/cls_7_1_byte.c,
1100         testsuite/libffi.call/cls_double.c,
1101         testsuite/libffi.call/nested_struct5.c,
1102         testsuite/libffi.call/nested_struct7.c,
1103         testsuite/libffi.call/cls_sint.c,
1104         testsuite/libffi.call/nested_struct9.c,
1105         testsuite/libffi.call/cls_20byte1.c,
1106         testsuite/libffi.call/cls_multi_sshortchar.c,
1107         testsuite/libffi.call/cls_align_sint64.c,
1108         testsuite/libffi.call/cls_3byte2.c,
1109         testsuite/libffi.call/cls_multi_schar.c,
1110         testsuite/libffi.call/cls_multi_uchar.c,
1111         testsuite/libffi.call/cls_19byte.c,
1112         testsuite/libffi.call/cls_9byte1.c,
1113         testsuite/libffi.call/cls_align_float.c,
1114         testsuite/libffi.call/closure_fn1.c,
1115         testsuite/libffi.call/problem1.c,
1116         testsuite/libffi.call/closure_fn3.c,
1117         testsuite/libffi.call/cls_sshort.c,
1118         testsuite/libffi.call/closure_fn5.c,
1119         testsuite/libffi.call/cls_align_double.c,
1120         testsuite/libffi.call/cls_2byte.c,
1121         testsuite/libffi.call/nested_struct.c,
1122         testsuite/libffi.call/nested_struct10.c,
1123         testsuite/libffi.call/cls_4byte.c,
1124         testsuite/libffi.call/cls_6byte.c,
1125         testsuite/libffi.call/cls_8byte.c,
1126         testsuite/libffi.call/cls_multi_sshort.c,
1127         testsuite/libffi.call/cls_align_uint32.c,
1128         testsuite/libffi.call/cls_align_sint16.c,
1129         testsuite/libffi.call/cls_float.c,
1130         testsuite/libffi.call/cls_20byte.c,
1131         testsuite/libffi.call/cls_5_1_byte.c,
1132         testsuite/libffi.call/nested_struct2.c,
1133         testsuite/libffi.call/cls_24byte.c,
1134         testsuite/libffi.call/nested_struct4.c,
1135         testsuite/libffi.call/nested_struct6.c,
1136         testsuite/libffi.call/cls_64byte.c,
1137         testsuite/libffi.call/nested_struct8.c,
1138         testsuite/libffi.call/cls_uint.c,
1139         testsuite/libffi.call/cls_multi_ushortchar.c,
1140         testsuite/libffi.call/cls_schar.c,
1141         testsuite/libffi.call/cls_uchar.c,
1142         testsuite/libffi.call/cls_align_uint64.c,
1143         testsuite/libffi.call/cls_ulonglong.c,
1144         testsuite/libffi.call/cls_align_longdouble.c,
1145         testsuite/libffi.call/cls_1_1byte.c,
1146         testsuite/libffi.call/cls_12byte.c,
1147         testsuite/libffi.call/cls_3_1byte.c,
1148         testsuite/libffi.call/cls_3byte1.c,
1149         testsuite/libffi.call/cls_4_1byte.c,
1150         testsuite/libffi.call/cls_6_1_byte.c,
1151         testsuite/libffi.call/cls_16byte.c,
1152         testsuite/libffi.call/cls_18byte.c,
1153         testsuite/libffi.call/closure_fn0.c,
1154         testsuite/libffi.call/cls_9byte2.c,
1155         testsuite/libffi.call/closure_fn2.c,
1156         testsuite/libffi.call/closure_fn4.c,
1157         testsuite/libffi.call/cls_ushort.c,
1158         testsuite/libffi.call/closure_fn6.c,
1159         testsuite/libffi.call/cls_5byte.c,
1160         testsuite/libffi.call/cls_align_pointer.c,
1161         testsuite/libffi.call/cls_7byte.c,
1162         testsuite/libffi.call/cls_align_sint32.c,
1163         testsuite/libffi.special/unwindtest_ffi_call.cc,
1164         testsuite/libffi.special/unwindtest.cc: Enable for ARM.
1165
1166 2007-07-05  H.J. Lu  <hongjiu.lu@intel.com>
1167
1168         * aclocal.m4: Regenerated.
1169
1170 2007-06-02  Paolo Bonzini  <bonzini@gnu.org>
1171
1172         * configure: Regenerate.
1173
1174 2007-05-23  Steve Ellcey  <sje@cup.hp.com>
1175
1176         * Makefile.in: Regenerate.
1177         * configure: Regenerate.
1178         * aclocal.m4: Regenerate.
1179         * include/Makefile.in: Regenerate.
1180         * testsuite/Makefile.in: Regenerate.
1181
1182 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1183
1184         * src/m68k/ffi.c (ffi_prep_incoming_args_SYSV,
1185         ffi_closure_SYSV_inner,ffi_prep_closure): New, add closure support.
1186         * src/m68k/sysv.S(ffi_closure_SYSV,ffi_closure_struct_SYSV): Likewise.
1187         * src/m68k/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1188         (FFI_CLOSURES): Enable closure support.
1189
1190 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1191
1192         * configure.ac (HAVE_AS_CFI_PSEUDO_OP): New test.
1193         * configure: Regenerate.
1194         * fficonfig.h.in: Regenerate.
1195         * src/m68k/sysv.S (CFI_STARTPROC,CFI_ENDPROC,
1196         CFI_OFFSET,CFI_DEF_CFA): New macros.
1197         (ffi_call_SYSV): Add callframe annotation.
1198
1199 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1200
1201         * src/m68k/ffi.c (ffi_prep_args,ffi_prep_cif_machdep): Fix
1202         numerous test suite failures.
1203         * src/m68k/sysv.S (ffi_call_SYSV): Likewise.
1204
1205 2007-04-11  Paolo Bonzini  <bonzini@gnu.org>
1206
1207         * Makefile.am (EXTRA_DIST): Bring up to date.
1208         * Makefile.in: Regenerate.
1209         * src/frv/eabi.S: Remove RCS keyword.
1210
1211 2007-04-06  Richard Henderson  <rth@redhat.com>
1212
1213         * configure.ac: Tidy target case.
1214         (HAVE_LONG_DOUBLE): Allow the target to override.
1215         * configure: Regenerate.
1216         * include/ffi.h.in: Don't define ffi_type_foo if
1217         LIBFFI_HIDE_BASIC_TYPES is defined.
1218         (ffi_type_longdouble): If not HAVE_LONG_DOUBLE, define
1219         to ffi_type_double.
1220         * types.c (LIBFFI_HIDE_BASIC_TYPES): Define.
1221         (FFI_TYPEDEF, ffi_type_void): Mark the data const.
1222         (ffi_type_longdouble): Special case for Alpha.  Don't define
1223         if long double == double.
1224
1225         * src/alpha/ffi.c (FFI_TYPE_LONGDOUBLE): Assert unique value.
1226         (ffi_prep_cif_machdep): Handle it as the 128-bit type.
1227         (ffi_call, ffi_closure_osf_inner): Likewise.
1228         (ffi_closure_osf_inner): Likewise.  Mark hidden.
1229         (ffi_call_osf, ffi_closure_osf): Mark hidden.
1230         * src/alpha/ffitarget.h (FFI_LAST_ABI): Tidy definition.
1231         * src/alpha/osf.S (ffi_call_osf, ffi_closure_osf): Mark hidden.
1232         (load_table): Handle 128-bit long double.
1233
1234         * testsuite/libffi.call/float4.c: Add -mieee for alpha.
1235
1236 2007-04-06  Tom Tromey  <tromey@redhat.com>
1237
1238         PR libffi/31491:
1239         * README: Fixed bug in example.
1240
1241 2007-04-03  Jakub Jelinek  <jakub@redhat.com>
1242
1243         * src/closures.c: Include sys/statfs.h.
1244         (_GNU_SOURCE): Define on Linux.
1245         (FFI_MMAP_EXEC_SELINUX): Define.
1246         (selinux_enabled): New variable.
1247         (selinux_enabled_check): New function.
1248         (is_selinux_enabled): Define.
1249         (dlmmap): Use it.
1250
1251 2007-03-24  Uros Bizjak  <ubizjak@gmail.com>
1252
1253         * testsuite/libffi.call/return_fl2.c (return_fl): Mark as static.
1254         Use 'volatile float sum' to create sum of floats to avoid false
1255         negative due to excess precision on ix86 targets.
1256         (main): Ditto.
1257
1258 2007-03-08  Alexandre Oliva  <aoliva@redhat.com>
1259
1260         * src/powerpc/ffi.c (flush_icache): Fix left-over from previous
1261         patch.
1262         (ffi_prep_closure_loc): Remove unneeded casts.  Add needed ones.
1263
1264 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1265
1266         * include/ffi.h.in (ffi_closure_alloc, ffi_closure_free): New.
1267         (ffi_prep_closure_loc): New.
1268         (ffi_prep_raw_closure_loc): New.
1269         (ffi_prep_java_raw_closure_loc): New.
1270         * src/closures.c: New file.
1271         * src/dlmalloc.c [FFI_MMAP_EXEC_WRIT] (struct malloc_segment):
1272         Replace sflags with exec_offset.
1273         [FFI_MMAP_EXEC_WRIT] (mmap_exec_offset, add_segment_exec_offset,
1274         sub_segment_exec_offset): New macros.
1275         (get_segment_flags, set_segment_flags, check_segment_merge): New
1276         macros.
1277         (is_mmapped_segment, is_extern_segment): Use get_segment_flags.
1278         (add_segment, sys_alloc, create_mspace, create_mspace_with_base,
1279         destroy_mspace): Use new macros.
1280         (sys_alloc): Silence warning.
1281         * Makefile.am (libffi_la_SOURCES): Add src/closures.c.
1282         * Makefile.in: Rebuilt.
1283         * src/prep_cif [FFI_CLOSURES] (ffi_prep_closure): Implement in
1284         terms of ffi_prep_closure_loc.
1285         * src/raw_api.c (ffi_prep_raw_closure_loc): Renamed and adjusted
1286         from...
1287         (ffi_prep_raw_closure): ... this.  Re-implement in terms of the
1288         renamed version.
1289         * src/java_raw_api (ffi_prep_java_raw_closure_loc): Renamed and
1290         adjusted from...
1291         (ffi_prep_java_raw_closure): ... this.  Re-implement in terms of
1292         the renamed version.
1293         * src/alpha/ffi.c (ffi_prep_closure_loc): Renamed from
1294         (ffi_prep_closure): ... this.
1295         * src/pa/ffi.c: Likewise.
1296         * src/cris/ffi.c: Likewise.  Adjust.
1297         * src/frv/ffi.c: Likewise.
1298         * src/ia64/ffi.c: Likewise.
1299         * src/mips/ffi.c: Likewise.
1300         * src/powerpc/ffi_darwin.c: Likewise.
1301         * src/s390/ffi.c: Likewise.
1302         * src/sh/ffi.c: Likewise.
1303         * src/sh64/ffi.c: Likewise.
1304         * src/sparc/ffi.c: Likewise.
1305         * src/x86/ffi64.c: Likewise.
1306         * src/x86/ffi.c: Likewise.
1307         (FFI_INIT_TRAMPOLINE): Adjust.
1308         (ffi_prep_raw_closure_loc): Renamed and adjusted from...
1309         (ffi_prep_raw_closure): ... this.
1310         * src/powerpc/ffi.c (ffi_prep_closure_loc): Renamed from
1311         (ffi_prep_closure): ... this.
1312         (flush_icache): Adjust.
1313
1314 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1315
1316         * src/dlmalloc.c: New file, imported version 2.8.3 of Doug
1317         Lea's malloc.
1318
1319 2007-03-01  Brooks Moses  <brooks.moses@codesourcery.com>
1320
1321         * Makefile.am: Add dummy install-pdf target.
1322         * Makefile.in: Regenerate
1323
1324 2007-02-13  Andreas Krebbel  <krebbel1@de.ibm.com>
1325
1326         * src/s390/ffi.c (ffi_prep_args, ffi_prep_cif_machdep,
1327         ffi_closure_helper_SYSV): Add long double handling.
1328
1329 2007-02-02  Jakub Jelinek  <jakub@redhat.com>
1330
1331         * src/powerpc/linux64.S (ffi_call_LINUX64): Move restore of r2
1332         immediately after bctrl instruction.
1333
1334 2007-01-18  Alexandre Oliva  <aoliva@redhat.com>
1335
1336         * Makefile.am (all-recursive, install-recursive,
1337         mostlyclean-recursive, clean-recursive, distclean-recursive,
1338         maintainer-clean-recursive): Add missing targets.
1339         * Makefile.in: Rebuilt.
1340
1341 2006-12-14  Andreas Tobler  <a.tobler@schweiz.org>
1342
1343         * configure.ac: Add TARGET for x86_64-*-darwin*.
1344         * Makefile.am (nodist_libffi_la_SOURCES): Add rules for 64-bit sources
1345         for X86_DARWIN.
1346         * src/x86/ffitarget.h: Set trampoline size for x86_64-*-darwin*.
1347         * src/x86/darwin64.S: New file for x86_64-*-darwin* support.
1348         * configure: Regenerate.
1349         * Makefile.in: Regenerate.
1350         * include/Makefile.in: Regenerate.
1351         * testsuite/Makefile.in: Regenerate.
1352         * testsuite/libffi.special/unwindtest_ffi_call.cc: New test case for
1353         ffi_call only.
1354
1355 2006-12-13  Andreas Tobler <a.tobler@schweiz.org>
1356
1357         * aclocal.m4: Regenerate with aclocal -I .. as written in the
1358         Makefile.am.
1359
1360 2006-10-31  Geoffrey Keating  <geoffk@apple.com>
1361
1362         * src/powerpc/ffi_darwin.c (darwin_adjust_aggregate_sizes): New.
1363         (ffi_prep_cif_machdep): Call darwin_adjust_aggregate_sizes for
1364         Darwin.
1365         * testsuite/libffi.call/nested_struct4.c: Remove Darwin XFAIL.
1366         * testsuite/libffi.call/nested_struct6.c: Remove Darwin XFAIL.
1367
1368 2006-10-10  Paolo Bonzini  <bonzini@gnu.org>
1369             Sandro Tolaini  <tolaini@libero.it>
1370
1371         * configure.ac [i*86-*-darwin*]: Set X86_DARWIN symbol and
1372         conditional.
1373         * configure: Regenerated.
1374         * Makefile.am (nodist_libffi_la_SOURCES) [X86_DARWIN]: New case.
1375         (EXTRA_DIST): Add src/x86/darwin.S.
1376         * Makefile.in: Regenerated.
1377         * include/Makefile.in: Regenerated.
1378         * testsuite/Makefile.in: Regenerated.
1379
1380         * src/x86/ffi.c (ffi_prep_cif_machdep) [X86_DARWIN]: Treat like
1381         X86_WIN32, and additionally align stack to 16 bytes.
1382         * src/x86/darwin.S: New, based on sysv.S.
1383         * src/prep_cif.c (ffi_prep_cif) [X86_DARWIN]: Align > 8-byte structs.
1384
1385 2006-09-12  David Daney  <ddaney@avtrex.com>
1386
1387         PR libffi/23935
1388         * include/Makefile.am: Install both ffi.h and ffitarget.h in
1389         $(libdir)/gcc/$(target_alias)/$(gcc_version)/include.
1390         * aclocal.m4: Regenerated for automake 1.9.6.
1391         * Makefile.in: Regenerated.
1392         * include/Makefile.in: Regenerated.
1393         * testsuite/Makefile.in: Regenerated.
1394
1395 2006-08-17  Andreas Tobler  <a.tobler@schweiz.ch>
1396
1397         * include/ffi_common.h (struct): Revert accidental commit.
1398
1399 2006-08-15  Andreas Tobler  <a.tobler@schweiz.ch>
1400
1401         * include/ffi_common.h: Remove lint directives.
1402         * include/ffi.h.in: Likewise.
1403
1404 2006-07-25  Torsten Schoenfeld  <kaffeetisch@gmx.de>
1405
1406         * include/ffi.h.in (ffi_type_ulong, ffi_type_slong): Define correctly
1407         for 32-bit architectures.
1408         * testsuite/libffi.call/return_ul.c: New test case.
1409
1410 2006-07-19  David Daney  <ddaney@avtrex.com>
1411
1412         * testsuite/libffi.call/closure_fn6.c: Remove xfail for mips,
1413         xfail remains for mips64.
1414
1415 2006-05-23  Carlos O'Donell  <carlos@codesourcery.com>
1416
1417         * Makefile.am: Add install-html target. Add install-html to .PHONY
1418         * Makefile.in: Regenerate.
1419         * aclocal.m4: Regenerate.
1420         * include/Makefile.in: Regenerate.
1421         * testsuite/Makefile.in: Regenerate.
1422
1423 2006-05-18  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1424
1425         * pa/ffi.c (ffi_prep_args_pa32): Load floating point arguments from
1426         stack slot.
1427
1428 2006-04-22  Andreas Tobler  <a.tobler@schweiz.ch>
1429
1430         * README: Remove notice about 'Crazy Comments'.
1431         * src/debug.c: Remove lint directives. Cleanup white spaces.
1432         * src/java_raw_api.c: Likewise.
1433         * src/prep_cif.c: Likewise.
1434         * src/raw_api.c: Likewise.
1435         * src/ffitest.c: Delete. No longer needed, all test cases migrated
1436         to the testsuite.
1437         * src/arm/ffi.c: Remove lint directives.
1438         * src/m32r/ffi.c: Likewise.
1439         * src/pa/ffi.c: Likewise.
1440         * src/powerpc/ffi.c: Likewise.
1441         * src/powerpc/ffi_darwin.c: Likewise.
1442         * src/sh/ffi.c: Likewise.
1443         * src/sh64/ffi.c: Likewise.
1444         * src/x86/ffi.c: Likewise.
1445         * testsuite/libffi.call/float2.c: Likewise.
1446         * testsuite/libffi.call/promotion.c: Likewise.
1447         * testsuite/libffi.call/struct1.c: Likewise.
1448
1449 2006-04-13  Andreas Tobler  <a.tobler@schweiz.ch>
1450
1451         * src/pa/hpux32.S: Correct unwind offset calculation for
1452         ffi_closure_pa32.
1453         * src/pa/linux.S: Likewise.
1454
1455 2006-04-12  James E Wilson  <wilson@specifix.com>
1456
1457         PR libgcj/26483
1458         * src/ia64/ffi.c (stf_spill, ldf_fill): Rewrite as macros.
1459         (hfa_type_load): Call stf_spill.
1460         (hfa_type_store): Call ldf_fill.
1461         (ffi_call): Adjust calls to above routines.  Add local temps for
1462         macro result.
1463
1464 2006-04-10  Matthias Klose  <doko@debian.org>
1465
1466         * testsuite/lib/libffi-dg.exp (libffi-init): Recognize multilib
1467         directory names containing underscores.
1468
1469 2006-04-07  James E Wilson  <wilson@specifix.com>
1470
1471         * testsuite/libffi.call/float4.c: New testcase.
1472
1473 2006-04-05  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1474             Andreas Tobler  <a.tobler@schweiz.ch>
1475
1476         * Makefile.am: Add PA_HPUX port.
1477         * Makefile.in: Regenerate.
1478         * include/Makefile.in: Likewise.
1479         * testsuite/Makefile.in: Likewise.
1480         * configure.ac: Add PA_HPUX rules.
1481         * configure: Regenerate.
1482         * src/pa/ffitarget.h: Rename linux target to PA_LINUX.
1483         Add PA_HPUX and PA64_HPUX.
1484         Rename FFI_LINUX ABI to FFI_PA32 ABI.
1485         (FFI_TRAMPOLINE_SIZE): Define for 32-bit HP-UX targets.
1486         (FFI_TYPE_SMALL_STRUCT2): Define.
1487         (FFI_TYPE_SMALL_STRUCT4): Likewise.
1488         (FFI_TYPE_SMALL_STRUCT8): Likewise.
1489         (FFI_TYPE_SMALL_STRUCT3): Redefine.
1490         (FFI_TYPE_SMALL_STRUCT5): Likewise.
1491         (FFI_TYPE_SMALL_STRUCT6): Likewise.
1492         (FFI_TYPE_SMALL_STRUCT7): Likewise.
1493         * src/pa/ffi.c (ROUND_DOWN): Delete.
1494         (fldw, fstw, fldd, fstd): Use '__asm__'.
1495         (ffi_struct_type): Add support for FFI_TYPE_SMALL_STRUCT2,
1496         FFI_TYPE_SMALL_STRUCT4 and FFI_TYPE_SMALL_STRUCT8.
1497         (ffi_prep_args_LINUX): Rename to ffi_prep_args_pa32. Update comment.
1498         Simplify incrementing of stack slot variable. Change type of local
1499         'n' to unsigned int.
1500         (ffi_size_stack_LINUX): Rename to ffi_size_stack_pa32. Handle long
1501         double on PA_HPUX.
1502         (ffi_prep_cif_machdep): Likewise.
1503         (ffi_call): Likewise.
1504         (ffi_closure_inner_LINUX): Rename to ffi_closure_inner_pa32. Change
1505         return type to ffi_status. Simplify incrementing of stack slot
1506         variable. Only copy floating point argument registers when PA_LINUX
1507         is true. Reformat debug statement.
1508         Add support for FFI_TYPE_SMALL_STRUCT2, FFI_TYPE_SMALL_STRUCT4 and
1509         FFI_TYPE_SMALL_STRUCT8.
1510         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Add 'extern' to
1511         declaration.
1512         (ffi_prep_closure): Make linux trampoline conditional on PA_LINUX.
1513         Add nops to cache flush.  Add trampoline for PA_HPUX.
1514         * src/pa/hpux32.S: New file.
1515         * src/pa/linux.S (ffi_call_LINUX): Rename to ffi_call_pa32. Rename
1516         ffi_prep_args_LINUX to ffi_prep_args_pa32.
1517         Localize labels. Add support for 2, 4 and 8-byte small structs. Handle
1518         unaligned destinations in 3, 5, 6 and 7-byte small structs. Order
1519         argument type checks so that common argument types appear first.
1520         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Rename
1521         ffi_closure_inner_LINUX to ffi_closure_inner_pa32.
1522
1523 2006-03-24  Alan Modra  <amodra@bigpond.net.au>
1524
1525         * src/powerpc/ffitarget.h (enum ffi_abi): Add FFI_LINUX.  Default
1526         for 32-bit using IBM extended double format.  Fix FFI_LAST_ABI.
1527         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Handle linux variant of
1528         FFI_TYPE_LONGDOUBLE.
1529         (ffi_prep_args64): Assert using IBM extended double.
1530         (ffi_prep_cif_machdep): Don't munge FFI_TYPE_LONGDOUBLE type.
1531         Handle FFI_LINUX FFI_TYPE_LONGDOUBLE return and args.
1532         (ffi_call): Handle FFI_LINUX.
1533         (ffi_closure_helper_SYSV): Non FFI_LINUX long double return needs
1534         gpr3 return pointer as for struct return.  Handle FFI_LINUX
1535         FFI_TYPE_LONGDOUBLE return and args.  Don't increment "nf"
1536         unnecessarily.
1537         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Load both f1 and f2
1538         for FFI_TYPE_LONGDOUBLE.  Move epilogue insns into case table.
1539         Don't use r6 as pointer to results, instead use sp offset.  Don't
1540         make a special call to load lr with case table address, instead
1541         use offset from previous call.
1542         * src/powerpc/sysv.S (ffi_call_SYSV): Save long double return.
1543         * src/powerpc/linux64.S (ffi_call_LINUX64): Simplify long double
1544         return.
1545
1546 2006-03-15  Kaz Kojima  <kkojima@gcc.gnu.org>
1547
1548         * src/sh64/ffi.c (ffi_prep_cif_machdep): Handle float arguments
1549         passed with FP registers correctly.
1550         (ffi_closure_helper_SYSV): Likewise.
1551         * src/sh64/sysv.S: Likewise.
1552
1553 2006-03-01  Andreas Tobler  <a.tobler@schweiz.ch>
1554
1555         * testsuite/libffi.special/unwindtest.cc (closure_test_fn): Mark cif,
1556         args and userdata unused.
1557         (closure_test_fn1): Mark cif and userdata unused.
1558         (main): Remove unused res.
1559
1560 2006-02-28  Andreas Tobler  <a.tobler@schweiz.ch>
1561
1562         * testsuite/libffi.call/call.exp: Adjust FSF address. Add test runs for
1563         -O2, -O3, -Os and the warning flags -W -Wall.
1564         * testsuite/libffi.special/special.exp: Likewise.
1565         * testsuite/libffi.call/ffitest.h: Add an __UNUSED__ macro to mark
1566         unused parameter unused for gcc or else do nothing.
1567         * testsuite/libffi.special/ffitestcxx.h: Likewise.
1568         * testsuite/libffi.call/cls_12byte.c (cls_struct_12byte_gn): Mark cif
1569         and userdata unused.
1570         * testsuite/libffi.call/cls_16byte.c (cls_struct_16byte_gn): Likewise.
1571         * testsuite/libffi.call/cls_18byte.c (cls_struct_18byte_gn): Likewise.
1572         * testsuite/libffi.call/cls_19byte.c (cls_struct_19byte_gn): Likewise.
1573         * testsuite/libffi.call/cls_1_1byte.c (cls_struct_1_1byte_gn): Likewise.
1574         * testsuite/libffi.call/cls_20byte.c (cls_struct_20byte_gn): Likewise.
1575         * testsuite/libffi.call/cls_20byte1.c (cls_struct_20byte_gn): Likewise.
1576         * testsuite/libffi.call/cls_24byte.c (cls_struct_24byte_gn): Likewise.
1577         * testsuite/libffi.call/cls_2byte.c (cls_struct_2byte_gn): Likewise.
1578         * testsuite/libffi.call/cls_3_1byte.c (cls_struct_3_1byte_gn): Likewise.
1579         * testsuite/libffi.call/cls_3byte1.c (cls_struct_3byte_gn): Likewise.
1580         * testsuite/libffi.call/cls_3byte2.c (cls_struct_3byte_gn1): Likewise.
1581         * testsuite/libffi.call/cls_4_1byte.c (cls_struct_4_1byte_gn): Likewise.
1582         * testsuite/libffi.call/cls_4byte.c (cls_struct_4byte_gn): Likewise.
1583         * testsuite/libffi.call/cls_5_1_byte.c (cls_struct_5byte_gn): Likewise.
1584         * testsuite/libffi.call/cls_5byte.c (cls_struct_5byte_gn): Likewise.
1585         * testsuite/libffi.call/cls_64byte.c (cls_struct_64byte_gn): Likewise.
1586         * testsuite/libffi.call/cls_6_1_byte.c (cls_struct_6byte_gn): Likewise.
1587         * testsuite/libffi.call/cls_6byte.c (cls_struct_6byte_gn): Likewise.
1588         * testsuite/libffi.call/cls_7_1_byte.c (cls_struct_7byte_gn): Likewise.
1589         * testsuite/libffi.call/cls_7byte.c (cls_struct_7byte_gn): Likewise.
1590         * testsuite/libffi.call/cls_8byte.c (cls_struct_8byte_gn): Likewise.
1591         * testsuite/libffi.call/cls_9byte1.c (cls_struct_9byte_gn): Likewise.
1592         * testsuite/libffi.call/cls_9byte2.c (cls_struct_9byte_gn): Likewise.
1593         * testsuite/libffi.call/cls_align_double.c (cls_struct_align_gn):
1594         Likewise.
1595         * testsuite/libffi.call/cls_align_float.c (cls_struct_align_gn):
1596         Likewise.
1597         * testsuite/libffi.call/cls_align_longdouble.c (cls_struct_align_gn):
1598         Likewise.
1599         * testsuite/libffi.call/cls_align_pointer.c (cls_struct_align_fn): Cast
1600         void* to avoid compiler warning.
1601         (main): Likewise.
1602         (cls_struct_align_gn): Mark cif and userdata unused.
1603         * testsuite/libffi.call/cls_align_sint16.c (cls_struct_align_gn):
1604         Likewise.
1605         * testsuite/libffi.call/cls_align_sint32.c (cls_struct_align_gn):
1606         Likewise.
1607         * testsuite/libffi.call/cls_align_sint64.c (cls_struct_align_gn):
1608         Likewise.
1609         * testsuite/libffi.call/cls_align_uint16.c (cls_struct_align_gn):
1610         Likewise.
1611         * testsuite/libffi.call/cls_align_uint32.c (cls_struct_align_gn):
1612         Likewise.
1613         * testsuite/libffi.call/cls_double.c (cls_ret_double_fn): Likewise.
1614         * testsuite/libffi.call/cls_float.c (cls_ret_float_fn): Likewise.
1615         * testsuite/libffi.call/cls_multi_schar.c (test_func_gn): Mark cif and
1616         data unused.
1617         (main): Cast res_call to silence gcc.
1618         * testsuite/libffi.call/cls_multi_sshort.c (test_func_gn): Mark cif and
1619         data unused.
1620         (main): Cast res_call to silence gcc.
1621         * testsuite/libffi.call/cls_multi_sshortchar.c (test_func_gn): Mark cif
1622         and data unused.
1623         (main): Cast res_call to silence gcc.
1624         * testsuite/libffi.call/cls_multi_uchar.c (test_func_gn): Mark cif and
1625         data unused.
1626         (main): Cast res_call to silence gcc.
1627         * testsuite/libffi.call/cls_multi_ushort.c (test_func_gn): Mark cif and
1628         data unused.
1629         (main): Cast res_call to silence gcc.
1630         * testsuite/libffi.call/cls_multi_ushortchar.c (test_func_gn): Mark cif
1631         and data unused.
1632         (main): Cast res_call to silence gcc.
1633         * testsuite/libffi.call/cls_schar.c (cls_ret_schar_fn): Mark cif and
1634         userdata unused.
1635         (cls_ret_schar_fn): Cast printf parameter to silence gcc.
1636         * testsuite/libffi.call/cls_sint.c (cls_ret_sint_fn): Mark cif and
1637         userdata unused.
1638         (cls_ret_sint_fn): Cast printf parameter to silence gcc.
1639         * testsuite/libffi.call/cls_sshort.c (cls_ret_sshort_fn): Mark cif and
1640         userdata unused.
1641         (cls_ret_sshort_fn): Cast printf parameter to silence gcc.
1642         * testsuite/libffi.call/cls_uchar.c (cls_ret_uchar_fn):  Mark cif and
1643         userdata unused.
1644         (cls_ret_uchar_fn): Cast printf parameter to silence gcc.
1645         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Mark cif and
1646         userdata unused.
1647         (cls_ret_uint_fn): Cast printf parameter to silence gcc.
1648         * testsuite/libffi.call/cls_ulonglong.c (cls_ret_ulonglong_fn): Mark cif
1649         and userdata unused.
1650         * testsuite/libffi.call/cls_ushort.c (cls_ret_ushort_fn): Mark cif and
1651         userdata unused.
1652         (cls_ret_ushort_fn): Cast printf parameter to silence gcc.
1653         * testsuite/libffi.call/float.c (floating): Remove unused parameter e.
1654         * testsuite/libffi.call/float1.c (main): Remove unused variable i.
1655         Cleanup white spaces.
1656         * testsuite/libffi.call/negint.c (checking): Remove unused variable i.
1657         * testsuite/libffi.call/nested_struct.c (cls_struct_combined_gn): Mark
1658         cif and userdata unused.
1659         * testsuite/libffi.call/nested_struct1.c (cls_struct_combined_gn):
1660         Likewise.
1661         * testsuite/libffi.call/nested_struct10.c (B_gn): Likewise.
1662         * testsuite/libffi.call/nested_struct2.c (B_fn): Adjust printf
1663         formatters to silence gcc.
1664         (B_gn): Mark cif and userdata unused.
1665         * testsuite/libffi.call/nested_struct3.c (B_gn): Mark cif and userdata
1666         unused.
1667         * testsuite/libffi.call/nested_struct4.c: Mention related PR.
1668         (B_gn): Mark cif and userdata unused.
1669         * testsuite/libffi.call/nested_struct5.c (B_gn): Mark cif and userdata
1670         unused.
1671         * testsuite/libffi.call/nested_struct6.c: Mention related PR.
1672         (B_gn): Mark cif and userdata unused.
1673         * testsuite/libffi.call/nested_struct7.c (B_gn): Mark cif and userdata
1674         unused.
1675         * testsuite/libffi.call/nested_struct8.c (B_gn): Likewise.
1676         * testsuite/libffi.call/nested_struct9.c (B_gn): Likewise.
1677         * testsuite/libffi.call/problem1.c (stub): Likewise.
1678         * testsuite/libffi.call/pyobjc-tc.c (main): Cast the result to silence
1679         gcc.
1680         * testsuite/libffi.call/return_fl2.c (return_fl): Add the note mentioned
1681         in the last commit for this test case in the test case itself.
1682         * testsuite/libffi.call/closure_fn0.c (closure_test_fn0): Mark cif as
1683         unused.
1684         * testsuite/libffi.call/closure_fn1.c (closure_test_fn1): Likewise.
1685         * testsuite/libffi.call/closure_fn2.c (closure_test_fn2): Likewise.
1686         * testsuite/libffi.call/closure_fn3.c (closure_test_fn3): Likewise.
1687         * testsuite/libffi.call/closure_fn4.c (closure_test_fn0): Likewise.
1688         * testsuite/libffi.call/closure_fn5.c (closure_test_fn5): Likewise.
1689         * testsuite/libffi.call/closure_fn6.c (closure_test_fn0): Likewise.
1690
1691 2006-02-22  Kaz Kojima  <kkojima@gcc.gnu.org>
1692
1693         * src/sh/sysv.S: Fix register numbers in the FDE for
1694         ffi_closure_SYSV.
1695
1696 2006-02-20  Andreas Tobler  <a.tobler@schweiz.ch>
1697
1698         * testsuite/libffi.call/return_fl2.c (return_fl): Remove static
1699         declaration to avoid a false negative on ix86. See PR323.
1700
1701 2006-02-18  Kaz Kojima  <kkojima@gcc.gnu.org>
1702
1703         * src/sh/ffi.c (ffi_closure_helper_SYSV): Remove unused variable
1704         and cast integer to void * if needed.  Update the pointer to
1705         the FP register saved area correctly.
1706
1707 2006-02-17  Andreas Tobler  <a.tobler@schweiz.ch>
1708
1709         * testsuite/libffi.call/nested_struct6.c: XFAIL this test until PR25630
1710         is fixed.
1711         * testsuite/libffi.call/nested_struct4.c: Likewise.
1712
1713 2006-02-16  Andreas Tobler  <a.tobler@schweiz.ch>
1714
1715         * testsuite/libffi.call/return_dbl.c: New test case.
1716         * testsuite/libffi.call/return_dbl1.c: Likewise.
1717         * testsuite/libffi.call/return_dbl2.c: Likewise.
1718         * testsuite/libffi.call/return_fl.c: Likewise.
1719         * testsuite/libffi.call/return_fl1.c: Likewise.
1720         * testsuite/libffi.call/return_fl2.c: Likewise.
1721         * testsuite/libffi.call/return_fl3.c: Likewise.
1722         * testsuite/libffi.call/closure_fn6.c: Likewise.
1723
1724         * testsuite/libffi.call/nested_struct2.c: Remove ffi_type_mylong
1725         definition.
1726         * testsuite/libffi.call/ffitest.h: Add ffi_type_mylong definition
1727         here to be used by other test cases too.
1728
1729         * testsuite/libffi.call/nested_struct10.c: New test case.
1730         * testsuite/libffi.call/nested_struct9.c: Likewise.
1731         * testsuite/libffi.call/nested_struct8.c: Likewise.
1732         * testsuite/libffi.call/nested_struct7.c: Likewise.
1733         * testsuite/libffi.call/nested_struct6.c: Likewise.
1734         * testsuite/libffi.call/nested_struct5.c: Likewise.
1735         * testsuite/libffi.call/nested_struct4.c: Likewise.
1736
1737 2006-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
1738
1739         * configure.ac: Enable libffi for sparc64-*-freebsd*.
1740         * configure: Rebuilt.
1741
1742 2006-01-18  Jakub Jelinek  <jakub@redhat.com>
1743
1744         * src/powerpc/sysv.S (smst_two_register): Don't call __ashldi3,
1745         instead do the shifting inline.
1746         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't compute %r5
1747         shift count unconditionally.  Simplify load sequences for 1, 2, 3, 4
1748         and 8 byte structs, for the remaining struct sizes don't call
1749         __lshrdi3, instead do the shifting inline.
1750
1751 2005-12-07  Thiemo Seufer  <ths@networkno.de>
1752
1753         * src/mips/ffitarget.h: Remove obsolete sgidefs.h include. Add
1754         missing parentheses.
1755         * src/mips/o32.S (ffi_call_O32): Code formatting. Define
1756         and use A3_OFF, FP_OFF, RA_OFF. Micro-optimizations.
1757         (ffi_closure_O32): Likewise, but with newly defined A3_OFF2,
1758         A2_OFF2, A1_OFF2, A0_OFF2, RA_OFF2, FP_OFF2, S0_OFF2, GP_OFF2,
1759         V1_OFF2, V0_OFF2, FA_1_1_OFF2, FA_1_0_OFF2, FA_0_1_OFF2,
1760         FA_0_0_OFF2.
1761         * src/mips/ffi.c (ffi_prep_args): Code formatting. Fix
1762         endianness bugs.
1763         (ffi_prep_closure): Improve trampoline instruction scheduling.
1764         (ffi_closure_mips_inner_O32): Fix endianness bugs.
1765
1766 2005-12-03  Alan Modra  <amodra@bigpond.net.au>
1767
1768         * src/powerpc/ffi.c: Formatting.
1769         (ffi_prep_args_SYSV): Avoid possible aliasing problems by using unions.
1770         (ffi_prep_args64): Likewise.
1771
1772 2005-09-30  Geoffrey Keating  <geoffk@apple.com>
1773
1774         * testsuite/lib/libffi-dg.exp (libffi_target_compile): For
1775         darwin, use -shared-libgcc not -lgcc_s, and explain why.
1776
1777 2005-09-26  Tom Tromey  <tromey@redhat.com>
1778
1779         * testsuite/libffi.call/float1.c (value_type): New typedef.
1780         (CANARY): New define.
1781         (main): Check for result buffer overflow.
1782         * src/powerpc/linux64.S: Handle linux64 long double returns.
1783         * src/powerpc/ffi.c (FLAG_RETURNS_128BITS): New constant.
1784         (ffi_prep_cif_machdep): Handle linux64 long double returns.
1785
1786 2005-08-25  Alan Modra  <amodra@bigpond.net.au>
1787
1788         PR target/23404
1789         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Correct placement of stack
1790         homed fp args.
1791         (ffi_status ffi_prep_cif_machdep): Correct stack sizing for same.
1792
1793 2005-08-11  Jakub Jelinek  <jakub@redhat.com>
1794
1795         * configure.ac (HAVE_HIDDEN_VISIBILITY_ATTRIBUTE): New test.
1796         (AH_BOTTOM): Add FFI_HIDDEN definition.
1797         * configure: Rebuilt.
1798         * fficonfig.h.in: Rebuilt.
1799         * src/powerpc/ffi.c (hidden): Remove.
1800         (ffi_closure_LINUX64, ffi_prep_args64, ffi_call_LINUX64,
1801         ffi_closure_helper_LINUX64): Use FFI_HIDDEN instead of hidden.
1802         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64,
1803         .ffi_closure_LINUX64): Use FFI_HIDDEN instead of .hidden.
1804         * src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV): Remove,
1805         add FFI_HIDDEN to its prototype.
1806         (ffi_closure_SYSV_inner): New.
1807         * src/x86/sysv.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
1808         * src/x86/win32.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
1809
1810 2005-08-10  Alfred M. Szmidt  <ams@gnu.org>
1811
1812         PR libffi/21819:
1813         * configure: Rebuilt.
1814         * configure.ac: Handle i*86-*-gnu*.
1815
1816 2005-08-09  Jakub Jelinek  <jakub@redhat.com>
1817
1818         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Use
1819         DW_CFA_offset_extended_sf rather than
1820         DW_CFA_GNU_negative_offset_extended.
1821         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise.
1822
1823 2005-07-22  SUGIOKA Toshinobu  <sugioka@itonet.co.jp>
1824
1825         * src/sh/sysv.S (ffi_call_SYSV): Stop argument popping correctly
1826         on sh3.
1827         (ffi_closure_SYSV): Change the stack layout for sh3 struct argument.
1828         * src/sh/ffi.c (ffi_prep_args): Fix sh3 argument copy, when it is
1829         partially on register.
1830         (ffi_closure_helper_SYSV): Likewise.
1831         (ffi_prep_cif_machdep): Don't set too many cif->flags.
1832
1833 2005-07-20  Kaz Kojima  <kkojima@gcc.gnu.org>
1834
1835         * src/sh/ffi.c (ffi_call): Handle small structures correctly.
1836         Remove empty line.
1837         * src/sh64/ffi.c (simple_type): Remove.
1838         (return_type): Handle small structures correctly.
1839         (ffi_prep_args): Likewise.
1840         (ffi_call): Likewise.
1841         (ffi_closure_helper_SYSV): Likewise.
1842         * src/sh64/sysv.S (ffi_call_SYSV): Handle 1, 2 and 4-byte return.
1843         Emit position independent code if PIC and remove wrong datalabel
1844         prefixes from EH data.
1845
1846 2005-07-19  Andreas Tobler  <a.tobler@schweiz.ch>
1847
1848         * Makefile.am (nodist_libffi_la_SOURCES): Add POWERPC_FREEBSD.
1849         * Makefile.in: Regenerate.
1850         * include/Makefile.in: Likewise.
1851         * testsuite/Makefile.in: Likewise.
1852         * configure.ac: Add POWERPC_FREEBSD rules.
1853         * configure: Regenerate.
1854         * src/powerpc/ffitarget.h: Add POWERPC_FREEBSD rules.
1855         (FFI_SYSV_TYPE_SMALL_STRUCT): Define.
1856         * src/powerpc/ffi.c: Add flags to handle small structure returns
1857         in ffi_call_SYSV.
1858         (ffi_prep_cif_machdep): Handle small structures for SYSV 4 ABI.
1859         Aka FFI_SYSV.
1860         (ffi_closure_helper_SYSV): Likewise.
1861         * src/powerpc/ppc_closure.S: Add return types for small structures.
1862         * src/powerpc/sysv.S: Add bits to handle small structures for
1863         final SYSV 4 ABI.
1864
1865 2005-07-10  Andreas Tobler  <a.tobler@schweiz.ch>
1866
1867         * testsuite/libffi.call/cls_5_1_byte.c: New test file.
1868         * testsuite/libffi.call/cls_6_1_byte.c: Likewise.
1869         * testsuite/libffi.call/cls_7_1_byte.c: Likewise.
1870
1871 2005-07-05  Randolph Chung  <tausq@debian.org>
1872
1873         * src/pa/ffi.c (ffi_struct_type): Rename FFI_TYPE_SMALL_STRUCT1
1874         as FFI_TYPE_SMALL_STRUCT3.  Break out handling for 5-7 byte
1875         structures.  Kill compilation warnings.
1876         (ffi_closure_inner_LINUX): Print return values as hex in debug
1877         message.  Rename FFI_TYPE_SMALL_STRUCT1 as FFI_TYPE_SMALL_STRUCT3.
1878         Properly handle 5-7 byte structure returns.
1879         * src/pa/ffitarget.h (FFI_TYPE_SMALL_STRUCT1)
1880         (FFI_TYPE_SMALL_STRUCT2): Remove.
1881         (FFI_TYPE_SMALL_STRUCT3, FFI_TYPE_SMALL_STRUCT5)
1882         (FFI_TYPE_SMALL_STRUCT6, FFI_TYPE_SMALL_STRUCT7): Define.
1883         * src/pa/linux.S: Mark source file as using PA1.1 assembly.
1884         (checksmst1, checksmst2): Remove.
1885         (checksmst3): Optimize handling of 3-byte struct returns.
1886         (checksmst567): Properly handle 5-7 byte struct returns.
1887
1888 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1889
1890         PR libgcj/21943
1891         * src/mips/n32.S: Enforce PIC code.
1892         * src/mips/o32.S: Likewise.
1893
1894 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1895
1896         * configure.ac: Treat i*86-*-solaris2.10 and up as X86_64.
1897         * configure: Regenerate.
1898
1899 2005-06-01  Alan Modra  <amodra@bigpond.net.au>
1900
1901         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't use JUMPTARGET
1902         to call ffi_closure_helper_SYSV.  Append @local instead.
1903         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise for ffi_prep_args_SYSV.
1904
1905 2005-05-17  Kelley Cook  <kcook@gcc.gnu.org>
1906
1907         * configure.ac: Use AC_C_BIGENDIAN instead of AC_C_BIGENDIAN_CROSS.
1908         Use AC_CHECK_SIZEOF instead of AC_COMPILE_CHECK_SIZEOF.
1909         * Makefile.am (ACLOCAL_AMFLAGS): Remove -I ../config.
1910         * aclocal.m4, configure, fficonfig.h.in, Makefile.in,
1911         include/Makefile.in, testsuite/Makefile.in: Regenerate.
1912
1913 2005-05-09  Mike Stump  <mrs@apple.com>
1914
1915         * configure: Regenerate.
1916
1917 2005-05-08  Richard Henderson  <rth@redhat.com>
1918
1919         PR libffi/21285
1920         * src/alpha/osf.S: Update unwind into to match code.
1921
1922 2005-05-04  Andreas Degert <ad@papyrus-gmbh.de>
1923             Richard Henderson  <rth@redhat.com>
1924
1925         * src/x86/ffi64.c (ffi_prep_cif_machdep): Save sse-used flag in
1926         bit 11 of flags.
1927         (ffi_call): Mask return type field.  Pass ssecount to ffi_call_unix64.
1928         (ffi_prep_closure): Set carry bit if sse-used flag set.
1929         * src/x86/unix64.S (ffi_call_unix64): Add ssecount argument.
1930         Only load sse registers if ssecount non-zero.
1931         (ffi_closure_unix64): Only save sse registers if carry set on entry.
1932
1933 2005-04-29  Ralf Corsepius  <ralf.corsepius@rtems.org>
1934
1935         * configure.ac: Add i*86-*-rtems*, sparc*-*-rtems*,
1936         powerpc-*rtems*, arm*-*-rtems*, sh-*-rtems*.
1937         * configure: Regenerate.
1938
1939 2005-04-20  Hans-Peter Nilsson  <hp@axis.com>
1940
1941         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): In regsub use,
1942         have Tcl8.3-compatible intermediate variable.
1943
1944 2005-04-18  Simon Posnjak <simon.posnjak@siol.net>
1945             Hans-Peter Nilsson  <hp@axis.com>
1946
1947         * Makefile.am: Add CRIS support.
1948         * configure.ac: Likewise.
1949         * Makefile.in, configure, testsuite/Makefile.in,
1950         include/Makefile.in: Regenerate.
1951         * src/cris: New directory.
1952         * src/cris/ffi.c, src/cris/sysv.S, src/cris/ffitarget.h: New files.
1953         * src/prep_cif.c (ffi_prep_cif): Wrap in #ifndef __CRIS__.
1954
1955         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): Replace \n with
1956         \r?\n in output tests.
1957
1958 2005-04-12  Mike Stump  <mrs@apple.com>
1959
1960         * configure: Regenerate.
1961
1962 2005-03-30  Hans Boehm  <Hans.Boehm@hp.com>
1963
1964         * src/ia64/ffitarget.h (ffi_arg): Use long long instead of DI.
1965
1966 2005-03-30  Steve Ellcey  <sje@cup.hp.com>
1967
1968         * src/ia64/ffitarget.h (ffi_arg) ADD DI attribute.
1969         (ffi_sarg) Ditto.
1970         * src/ia64/unix.S (ffi_closure_unix): Extend gp
1971         to 64 bits in ILP32 mode.
1972         Load 64 bits even for short data.
1973
1974 2005-03-23  Mike Stump  <mrs@apple.com>
1975
1976         * src/powerpc/darwin.S: Update for -m64 multilib.
1977         * src/powerpc/darwin_closure.S: Likewise.
1978
1979 2005-03-21  Zack Weinberg  <zack@codesourcery.com>
1980
1981         * configure.ac: Do not invoke TL_AC_GCC_VERSION.
1982         Do not set tool_include_dir.
1983         * aclocal.m4, configure, Makefile.in, testsuite/Makefile.in:
1984         Regenerate.
1985         * include/Makefile.am: Set gcc_version and toollibffidir.
1986         * include/Makefile.in: Regenerate.
1987
1988 2005-02-22  Andrew Haley  <aph@redhat.com>
1989
1990         * src/powerpc/ffi.c (ffi_prep_cif_machdep): Bump alignment to
1991         odd-numbered register pairs for 64-bit integer types.
1992
1993 2005-02-23  Andreas Tobler  <a.tobler@schweiz.ch>
1994
1995         PR libffi/20104
1996         * testsuite/libffi.call/return_ll1.c: New test case.
1997
1998 2005-02-11  Janis Johnson  <janis187@us.ibm.com>
1999
2000         * testsuite/libffi.call/cls_align_longdouble.c: Remove dg-options.
2001         * testsuite/libffi.call/float.c: Ditto.
2002         * testsuite/libffi.call/float2.c: Ditto.
2003         * testsuite/libffi.call/float3.c: Ditto.
2004
2005 2005-02-08  Andreas Tobler  <a.tobler@schweiz.ch>
2006
2007         * src/frv/ffitarget.h: Remove PPC stuff which does not belong to frv.
2008
2009 2005-01-12  Eric Botcazou  <ebotcazou@libertysurf.fr>
2010
2011         * testsuite/libffi.special/special.exp (cxx_options): Add
2012         -shared-libgcc.
2013
2014 2004-12-31  Richard Henderson  <rth@redhat.com>
2015
2016         * src/types.c (FFI_AGGREGATE_TYPEDEF): Remove.
2017         (FFI_TYPEDEF): Rename from FFI_INTEGRAL_TYPEDEF.  Replace size and
2018         offset parameters with a type parameter; deduce size and structure
2019         alignment.  Update all users.
2020
2021 2004-12-31  Richard Henderson  <rth@redhat.com>
2022
2023         * src/types.c (FFI_TYPE_POINTER): Define with sizeof.
2024         (FFI_TYPE_LONGDOUBLE): Fix for ia64.
2025         * src/ia64/ffitarget.h (struct ffi_ia64_trampoline_struct): Move
2026         into ffi_prep_closure.
2027         * src/ia64/ia64_flags.h, src/ia64/ffi.c, src/ia64/unix.S: Rewrite
2028         from scratch.
2029
2030 2004-12-27  Richard Henderson  <rth@redhat.com>
2031
2032         * src/x86/unix64.S: Fix typo in unwind info.
2033
2034 2004-12-25  Richard Henderson  <rth@redhat.com>
2035
2036         * src/x86/ffi64.c (struct register_args): Rename from stackLayout.
2037         (enum x86_64_reg_class): Add X86_64_COMPLEX_X87_CLASS.
2038         (merge_classes): Check for it.
2039         (SSE_CLASS_P): New.
2040         (classify_argument): Pass byte_offset by value; perform all updates
2041         inside struct case.
2042         (examine_argument): Add classes argument; handle
2043         X86_64_COMPLEX_X87_CLASS.
2044         (ffi_prep_args): Merge into ...
2045         (ffi_call): ... here.  Share stack frame with ffi_call_unix64.
2046         (ffi_prep_cif_machdep): Setup cif->flags for proper structure return.
2047         (ffi_fill_return_value): Remove.
2048         (ffi_prep_closure): Remove dead assert.
2049         (ffi_closure_unix64_inner): Rename from ffi_closure_UNIX64_inner.
2050         Rewrite to use struct register_args instead of va_list.  Create
2051         flags for handling structure returns.
2052         * src/x86/unix64.S: Remove dead strings.
2053         (ffi_call_unix64): Rename from ffi_call_UNIX64.  Rewrite to share
2054         stack frame with ffi_call.  Handle structure returns properly.
2055         (float2sse, floatfloat2sse, double2sse): Remove.
2056         (sse2float, sse2double, sse2floatfloat): Remove.
2057         (ffi_closure_unix64): Rename from ffi_closure_UNIX64.  Rewrite
2058         to handle structure returns properly.
2059
2060 2004-12-08  David Edelsohn  <edelsohn@gnu.org>
2061
2062         * Makefile.am (AM_MAKEFLAGS): Remove duplicate LIBCFLAGS and
2063         PICFLAG.
2064         * Makefile.in: Regenerated.
2065
2066 2004-12-02  Richard Sandiford  <rsandifo@redhat.com>
2067
2068         * configure.ac: Use TL_AC_GCC_VERSION to set gcc_version.
2069         * configure, aclocal.m4, Makefile.in: Regenerate.
2070         * include/Makefile.in, testsuite/Makefile.in: Regenerate.
2071
2072 2004-11-29  Kelley Cook  <kcook@gcc.gnu.org>
2073
2074         * configure: Regenerate for libtool change.
2075
2076 2004-11-25  Kelley Cook  <kcook@gcc.gnu.org>
2077
2078         * configure: Regenerate for libtool reversion.
2079
2080 2004-11-24  Kelley Cook  <kcook@gcc.gnu.org>
2081
2082         * configure: Regenerate for libtool change.
2083
2084 2004-11-23  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2085
2086         * testsuite/lib/libffi-dg.exp: Use new procs in target-libpath.exp.
2087
2088 2004-11-23  Richard Sandiford  <rsandifo@redhat.com>
2089
2090         * src/mips/o32.S (ffi_call_O32, ffi_closure_O32): Use jalr instead
2091         of jal.  Use an absolute encoding for the frame information.
2092
2093 2004-11-23  Kelley Cook  <kcook@gcc.gnu.org>
2094
2095         * Makefile.am: Remove no-dependencies.  Add ACLOCAL_AMFLAGS.
2096         * acinclude.m4: Delete logic for sincludes.
2097         * aclocal.m4, Makefile.in, configure: Regenerate.
2098         * include/Makefile: Likewise.
2099         * testsuite/Makefile: Likewise.
2100
2101 2004-11-22  Eric Botcazou  <ebotcazou@libertysurf.fr>
2102
2103         * src/sparc/ffi.c (ffi_prep_closure): Align doubles and 64-bit integers
2104         on a 8-byte boundary.
2105         * src/sparc/v8.S (ffi_closure_v8): Reserve frame space for arguments.
2106
2107 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2108
2109         * src/arm/ffi.c (ffi_prep_cif_machdep): Handle functions that return
2110         long long values.  Round stack allocation to a multiple of 8 bytes
2111         for ATPCS compatibility.
2112         * src/arm/sysv.S (ffi_call_SYSV): Rework to avoid use of APCS register
2113         names.  Handle returning long long types.  Add Thumb and interworking
2114         support.  Improve soft-float code.
2115
2116 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2117
2118         * testsuite/lib/libffi-db.exp (load_gcc_lib): New function.
2119         (libffi_exit): New function.
2120         (libffi_init): Build the testglue wrapper if needed.
2121
2122 2004-10-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2123
2124         PR other/18138
2125         * testsuite/lib/libffi-dg.exp: Accept more than one multilib libgcc.
2126
2127 2004-10-25  Kazuhiro Inaoka <inaoka.kazuhiro@renesas.com>
2128
2129         * src/m32r/libffitarget.h (FFI_CLOSURES): Set to 0.
2130
2131 2004-10-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2132
2133         * src/sh/sysv.S (ffi_call_SYSV): Don't align for double data.
2134         * testsuite/libffi.call/float3.c: New test case.
2135
2136 2004-10-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2137
2138         * src/sh/ffi.c (ffi_prep_closure): Set T bit in trampoline for
2139         the function returning a structure pointed with R2.
2140         * src/sh/sysv.S (ffi_closure_SYSV): Use R2 as the pointer to
2141         the structure return value if T bit set.  Emit position
2142         independent code and EH data if PIC.
2143
2144 2004-10-13  Kazuhiro Inaoka  <inaoka.kazuhiro@renesas.com>
2145
2146         * Makefile.am: Add m32r support.
2147         * configure.ac: Likewise.
2148         * Makefile.in: Regenerate.
2149         * confiugre: Regenerate.
2150         * src/types.c: Add m32r port to FFI_INTERNAL_TYPEDEF
2151         (uint64, sint64, double, longdouble)
2152         * src/m32r: New directory.
2153         * src/m32r/ffi.c: New file.
2154         * src/m32r/sysv.S: Likewise.
2155         * src/m32r/ffitarget.h: Likewise.
2156
2157 2004-10-02  Kaz Kojima  <kkojima@gcc.gnu.org>
2158
2159         * testsuite/libffi.call/negint.c: New test case.
2160
2161 2004-09-14  H.J. Lu  <hongjiu.lu@intel.com>
2162
2163         PR libgcj/17465
2164         * testsuite/lib/libffi-dg.exp: Don't use global ld_library_path.
2165         Set up LD_LIBRARY_PATH, SHLIB_PATH, LD_LIBRARYN32_PATH,
2166         LD_LIBRARY64_PATH, LD_LIBRARY_PATH_32, LD_LIBRARY_PATH_64 and
2167         DYLD_LIBRARY_PATH.
2168
2169 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2170
2171         * testsuite/libffi.call/many_win32.c: Remove whitespaces.
2172         * testsuite/libffi.call/promotion.c: Likewise.
2173         * testsuite/libffi.call/return_ll.c: Remove unused var. Cleanup
2174         whitespaces.
2175         * testsuite/libffi.call/return_sc.c: Likewise.
2176         * testsuite/libffi.call/return_uc.c: Likewise.
2177
2178 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2179
2180         * src/powerpc/darwin.S: Fix comments and identation.
2181         * src/powerpc/darwin_closure.S: Likewise.
2182
2183 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2184
2185         * src/powerpc/ffi_darwin.c: Add flag for longdouble return values.
2186         (ffi_prep_args): Handle longdouble arguments.
2187         (ffi_prep_cif_machdep): Set flags for longdouble. Calculate space for
2188         longdouble.
2189         (ffi_closure_helper_DARWIN): Add closure handling for longdouble.
2190         * src/powerpc/darwin.S (_ffi_call_DARWIN): Add handling of longdouble
2191         values.
2192         * src/powerpc/darwin_closure.S (_ffi_closure_ASM): Likewise.
2193         * src/types.c: Defined longdouble size and alignment for darwin.
2194
2195 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2196
2197         * src/powerpc/aix.S: Remove whitespaces.
2198         * src/powerpc/aix_closure.S: Likewise.
2199         * src/powerpc/asm.h: Likewise.
2200         * src/powerpc/ffi.c: Likewise.
2201         * src/powerpc/ffitarget.h: Likewise.
2202         * src/powerpc/linux64.S: Likewise.
2203         * src/powerpc/linux64_closure.S: Likewise.
2204         * src/powerpc/ppc_closure.S: Likewise.
2205         * src/powerpc/sysv.S: Likewise.
2206
2207 2004-08-30  Anthony Green  <green@redhat.com>
2208
2209         * Makefile.am: Add frv support.
2210         * Makefile.in, testsuite/Makefile.in: Rebuilt.
2211         * configure.ac: Read configure.host.
2212         * configure.in: Read configure.host.
2213         * configure.host: New file.  frv-elf needs libgloss.
2214         * include/ffi.h.in: Force ffi_closure to have a nice big (8)
2215         alignment.  This is needed to frv and shouldn't harm the others.
2216         * include/ffi_common.h (ALIGN_DOWN): New macro.
2217         * src/frv/ffi.c, src/frv/ffitarget.h, src/frv/eabi.S: New files.
2218
2219 2004-08-24  David Daney  <daney@avtrex.com>
2220
2221         * testsuite/libffi.call/closure_fn0.c: Xfail mips64* instead of mips*.
2222         * testsuite/libffi.call/closure_fn1.c: Likewise.
2223         * testsuite/libffi.call/closure_fn2.c  Likewise.
2224         * testsuite/libffi.call/closure_fn3.c: Likewise.
2225         * testsuite/libffi.call/closure_fn4.c: Likewise.
2226         * testsuite/libffi.call/closure_fn5.c: Likewise.
2227         * testsuite/libffi.call/cls_18byte.c: Likewise.
2228         * testsuite/libffi.call/cls_19byte.c: Likewise.
2229         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2230         * testsuite/libffi.call/cls_20byte.c: Likewise.
2231         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2232         * testsuite/libffi.call/cls_24byte.c: Likewise.
2233         * testsuite/libffi.call/cls_2byte.c: Likewise.
2234         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2235         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2236         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2237         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2238         * testsuite/libffi.call/cls_4byte.c: Likewise.
2239         * testsuite/libffi.call/cls_64byte.c: Likewise.
2240         * testsuite/libffi.call/cls_6byte.c: Likewise.
2241         * testsuite/libffi.call/cls_7byte.c: Likewise.
2242         * testsuite/libffi.call/cls_8byte.c: Likewise.
2243         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2244         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2245         * testsuite/libffi.call/cls_align_double.c: Likewise.
2246         * testsuite/libffi.call/cls_align_float.c: Likewise.
2247         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2248         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2249         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2250         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2251         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2252         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2253         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2254         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
2255         * testsuite/libffi.call/cls_double.c: Likewise.
2256         * testsuite/libffi.call/cls_float.c: Likewise.
2257         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2258         * testsuite/libffi.call/cls_multi_sshort.c: Likewise.
2259         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2260         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2261         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2262         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2263         * testsuite/libffi.call/cls_schar.c: Likewise.
2264         * testsuite/libffi.call/cls_sint.c: Likewise.
2265         * testsuite/libffi.call/cls_sshort.c: Likewise.
2266         * testsuite/libffi.call/cls_uchar.c: Likewise.
2267         * testsuite/libffi.call/cls_uint.c: Likewise.
2268         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2269         * testsuite/libffi.call/cls_ushort.c: Likewise.
2270         * testsuite/libffi.call/nested_struct.c: Likewise.
2271         * testsuite/libffi.call/nested_struct1.c: Likewise.
2272         * testsuite/libffi.call/nested_struct2.c: Likewise.
2273         * testsuite/libffi.call/nested_struct3.c: Likewise.
2274         * testsuite/libffi.call/problem1.c: Likewise.
2275         * testsuite/libffi.special/unwindtest.cc: Likewise.
2276         * testsuite/libffi.call/cls_12byte.c: Likewise and set return value
2277         to zero.
2278         * testsuite/libffi.call/cls_16byte.c: Likewise.
2279         * testsuite/libffi.call/cls_5byte.c: Likewise.
2280
2281 2004-08-23  David Daney <daney@avtrex.com>
2282
2283         PR libgcj/13141
2284         * src/mips/ffitarget.h (FFI_O32_SOFT_FLOAT): New ABI.
2285         * src/mips/ffi.c (ffi_prep_args): Fix alignment calculation.
2286         (ffi_prep_cif_machdep): Handle FFI_O32_SOFT_FLOAT floating point
2287         parameters and return types.
2288         (ffi_call): Handle FFI_O32_SOFT_FLOAT ABI.
2289         (ffi_prep_closure): Ditto.
2290         (ffi_closure_mips_inner_O32): Handle FFI_O32_SOFT_FLOAT ABI, fix
2291         alignment calculations.
2292         * src/mips/o32.S (ffi_closure_O32): Don't use floating point
2293         instructions if FFI_O32_SOFT_FLOAT, make stack frame ABI compliant.
2294
2295 2004-08-14  Casey Marshall <csm@gnu.org>
2296
2297         * src/mips/ffi.c (ffi_pref_cif_machdep): set `cif->flags' to
2298         contain `FFI_TYPE_UINT64' as return type for any 64-bit
2299         integer (O32 ABI only).
2300         (ffi_prep_closure): new function.
2301         (ffi_closure_mips_inner_O32): new function.
2302         * src/mips/ffitarget.h: Define `FFI_CLOSURES' and
2303         `FFI_TRAMPOLINE_SIZE' appropriately if the ABI is o32.
2304         * src/mips/o32.S (ffi_call_O32): add labels for .eh_frame. Return
2305         64 bit integers correctly.
2306         (ffi_closure_O32): new function.
2307         Added DWARF-2 unwind info for both functions.
2308
2309 2004-08-10  Andrew Haley  <aph@redhat.com>
2310
2311         * src/x86/ffi64.c (ffi_prep_args ): 8-align all stack arguments.
2312
2313 2004-08-01  Robert Millan  <robertmh@gnu.org>
2314
2315         * configure.ac: Detect knetbsd-gnu and kfreebsd-gnu.
2316         * configure: Regenerate.
2317
2318 2004-07-30  Maciej W. Rozycki  <macro@linux-mips.org>
2319
2320         * acinclude.m4 (AC_FUNC_MMAP_BLACKLIST): Check for <sys/mman.h>
2321         and mmap() explicitly instead of relying on preset autoconf cache
2322         variables.
2323         * aclocal.m4: Regenerate.
2324         * configure: Regenerate.
2325
2326 2004-07-11  Ulrich Weigand  <uweigand@de.ibm.com>
2327
2328         * src/s390/ffi.c (ffi_prep_args): Fix C aliasing violation.
2329         (ffi_check_float_struct): Remove unused prototype.
2330
2331 2004-06-30  Geoffrey Keating  <geoffk@apple.com>
2332
2333         * src/powerpc/ffi_darwin.c (flush_icache): ';' is a comment
2334         character on Darwin, use '\n\t' instead.
2335
2336 2004-06-26  Matthias Klose  <doko@debian.org>
2337
2338         * libtool-version: Fix typo in revision/age.
2339
2340 2004-06-17  Matthias Klose  <doko@debian.org>
2341
2342         * libtool-version: New.
2343         * Makefile.am (libffi_la_LDFLAGS): Use -version-info for soname.
2344         * Makefile.in: Regenerate.
2345
2346 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2347
2348         * Makefile.am: Remove useless multilib rules.
2349         * Makefile.in: Regenerate.
2350         * aclocal.m4: Regenerate with automake 1.8.5.
2351         * configure.ac: Remove useless multilib configury.
2352         * configure: Regenerate.
2353
2354 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2355
2356         * .cvsignore: New file.
2357
2358 2004-06-10  Jakub Jelinek  <jakub@redhat.com>
2359
2360         * src/ia64/unix.S (ffi_call_unix): Insert group barrier break
2361         fp_done.
2362         (ffi_closure_UNIX): Fix f14/f15 adjustment if FLOAT_SZ is ever
2363         changed from 8.
2364
2365 2004-06-06  Sean McNeil  <sean@mcneil.com>
2366
2367         * configure.ac: Add x86_64-*-freebsd* support.
2368         * configure: Regenerate.
2369
2370 2004-04-26  Joe Buck <jbuck@welsh-buck.org>
2371
2372         Bug 15093
2373         * configure.ac: Test for existence of mmap and sys/mman.h before
2374         checking blacklist.  Fix suggested by Jim Wilson.
2375         * configure: Regenerate.
2376
2377 2004-04-26  Matt Austern  <austern@apple.com>
2378
2379         * src/powerpc/darwin.S: Go through a non-lazy pointer for initial
2380         FDE location.
2381         * src/powerpc/darwin_closure.S: Likewise.
2382
2383 2004-04-24  Andreas Tobler  <a.tobler@schweiz.ch>
2384
2385         * testsuite/libffi.call/cls_multi_schar.c (main): Fix initialization
2386         error. Reported by Thomas Heller <theller@python.net>.
2387         * testsuite/libffi.call/cls_multi_sshort.c (main): Likewise.
2388         * testsuite/libffi.call/cls_multi_ushort.c (main): Likewise.
2389
2390 2004-03-20  Matthias Klose  <doko@debian.org>
2391
2392         * src/pa/linux.S: Fix typo.
2393
2394 2004-03-19  Matthias Klose  <doko@debian.org>
2395
2396         * Makefile.am: Update.
2397         * Makefile.in: Regenerate.
2398         * src/pa/ffi.h.in: Remove.
2399         * src/pa/ffitarget.h: New file.
2400
2401 2004-02-10  Randolph Chung  <tausq@debian.org>
2402
2403         * Makefile.am: Add PA support.
2404         * Makefile.in: Regenerate.
2405         * include/Makefile.in: Regenerate.
2406         * configure.ac: Add PA target.
2407         * configure: Regenerate.
2408         * src/pa/ffi.c: New file.
2409         * src/pa/ffi.h.in: Add PA support.
2410         * src/pa/linux.S: New file.
2411         * prep_cif.c: Add PA support.
2412
2413 2004-03-16  Hosaka Yuji  <hos@tamanegi.org>
2414
2415         * src/types.c: Fix alignment size of X86_WIN32 case int64 and
2416         double.
2417         * src/x86/ffi.c (ffi_prep_args): Replace ecif->cif->rtype->type
2418         with ecif->cif->flags.
2419         (ffi_call, ffi_prep_incoming_args_SYSV): Replace cif->rtype->type
2420         with cif->flags.
2421         (ffi_prep_cif_machdep): Add X86_WIN32 struct case.
2422         (ffi_closure_SYSV): Add 1 or 2-bytes struct case for X86_WIN32.
2423         * src/x86/win32.S (retstruct1b, retstruct2b, sc_retstruct1b,
2424         sc_retstruct2b): Add for 1 or 2-bytes struct case.
2425
2426 2004-03-15 Kelley Cook <kcook@gcc.gnu.org>
2427
2428         * configure.in: Rename file to ...
2429         * configure.ac: ... this.
2430         * fficonfig.h.in: Regenerate.
2431         * Makefile.in: Regenerate.
2432         * include/Makefile.in: Regenerate.
2433         * testsuite/Makefile.in: Regenerate.
2434
2435 2004-03-12  Matt Austern  <austern@apple.com>
2436
2437         * src/powerpc/darwin.S: Fix EH information so it corresponds to
2438         changes in EH format resulting from addition of linkonce support.
2439         * src/powerpc/darwin_closure.S: Likewise.
2440
2441 2004-03-11  Andreas Tobler  <a.tobler@schweiz.ch>
2442             Paolo Bonzini  <bonzini@gnu.org>
2443
2444         * Makefile.am (AUTOMAKE_OPTIONS): Set them.
2445         Remove VPATH. Remove rules for object files. Remove multilib support.
2446         (AM_CCASFLAGS): Add.
2447         * configure.in (AC_CONFIG_HEADERS): Relace AM_CONFIG_HEADER.
2448         (AC_PREREQ): Bump version to 2.59.
2449         (AC_INIT): Fill with version info and bug address.
2450         (ORIGINAL_LD_FOR_MULTILIBS): Remove.
2451         (AM_ENABLE_MULTILIB): Use this instead of AC_ARG_ENABLE.
2452         De-precious CC so that the right flags are passed down to multilibs.
2453         (AC_MSG_ERROR): Replace obsolete macro AC_ERROR.
2454         (AC_CONFIG_FILES): Replace obsolete macro AC_LINK_FILES.
2455         (AC_OUTPUT): Reorganize the output with AC_CONFIG_COMMANDS.
2456         * configure: Rebuilt.
2457         * aclocal.m4: Likewise.
2458         * Makefile.in, include/Makefile.in, testsuite/Makefile.in: Likewise.
2459         * fficonfig.h.in: Likewise.
2460
2461 2004-03-11  Andreas Schwab  <schwab@suse.de>
2462
2463         * src/ia64/ffi.c (ffi_prep_incoming_args_UNIX): Get floating point
2464         arguments from fp registers only for the first 8 parameter slots.
2465         Don't convert a float parameter when passed in memory.
2466
2467 2004-03-09  Hans-Peter Nilsson  <hp@axis.com>
2468
2469         * configure: Regenerate for config/accross.m4 correction.
2470
2471 2004-02-25  Matt Kraai  <kraai@alumni.cmu.edu>
2472
2473         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Change
2474         ecif->cif->bytes to bytes.
2475         (ffi_prep_cif_machdep): Add braces around nested if statement.
2476
2477 2004-02-09  Alan Modra  <amodra@bigpond.net.au>
2478
2479         * src/types.c (pointer): POWERPC64 has 8 byte pointers.
2480
2481         * src/powerpc/ffi.c (ffi_prep_args64): Correct long double handling.
2482         (ffi_closure_helper_LINUX64): Fix typo.
2483         * testsuite/libffi.call/cls_align_longdouble.c: Pass -mlong-double-128
2484         for powerpc64-*-*.
2485         * testsuite/libffi.call/float.c: Likewise.
2486         * testsuite/libffi.call/float2.c: Likewise.
2487
2488 2004-02-08  Alan Modra  <amodra@bigpond.net.au>
2489
2490         * src/powerpc/ffi.c (ffi_prep_cif_machdep <FFI_LINUX64>): Correct
2491         long double function return and long double arg handling.
2492         (ffi_closure_helper_LINUX64): Formatting.  Delete unused "ng" var.
2493         Use "end_pfr" instead of "nf".  Correct long double handling.
2494         Localise "temp".
2495         * src/powerpc/linux64.S (ffi_call_LINUX64): Save f2 long double
2496         return value.
2497         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Allocate
2498         space for long double return value.  Adjust stack frame and offsets.
2499         Load f2 long double return.
2500
2501 2004-02-07  Alan Modra  <amodra@bigpond.net.au>
2502
2503         * src/types.c: Use 16 byte long double for POWERPC64.
2504
2505 2004-01-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2506
2507         * src/sparc/ffi.c (ffi_prep_args_v9): Shift the parameter array
2508         when the structure return address is passed in %o0.
2509         (ffi_V9_return_struct): Rename into ffi_v9_layout_struct.
2510         (ffi_v9_layout_struct): Align the field following a nested structure
2511         on a word boundary.  Use memmove instead of memcpy.
2512         (ffi_call): Update call to ffi_V9_return_struct.
2513         (ffi_prep_closure): Define 'ctx' only for V8.
2514         (ffi_closure_sparc_inner): Clone into ffi_closure_sparc_inner_v8
2515         and ffi_closure_sparc_inner_v9.
2516         (ffi_closure_sparc_inner_v8): Return long doubles by reference.
2517         Always skip the structure return address.  For structures and long
2518         doubles, copy the argument directly.
2519         (ffi_closure_sparc_inner_v9): Skip the structure return address only
2520         if required.  Shift the maximum floating-point slot accordingly.  For
2521         big structures, copy the argument directly; otherwise, left-justify the
2522         argument and call ffi_v9_layout_struct to lay out the structure on
2523         the stack.
2524         * src/sparc/v8.S: Undef STACKFRAME before defining it.
2525         (ffi_closure_v8): Pass the structure return address.  Update call to
2526         ffi_closure_sparc_inner_v8.  Short-circuit FFI_TYPE_INT handling.
2527         Skip the 'unimp' insn when returning long doubles and structures.
2528         * src/sparc/v9.S: Undef STACKFRAME before defining it.
2529         (ffi_closure_v9): Increase the frame size by 2 words.  Short-circuit
2530         FFI_TYPE_INT handling.  Load structures both in integers and
2531         floating-point registers on return.
2532         * README: Update status of the SPARC port.
2533
2534 2004-01-24  Andreas Tobler  <a.tobler@schweiz.ch>
2535
2536         * testsuite/libffi.call/pyobjc-tc.c (main): Treat result value
2537         as of type ffi_arg.
2538         * testsuite/libffi.call/struct3.c (main): Fix CHECK.
2539
2540 2004-01-22  Ulrich Weigand  <uweigand@de.ibm.com>
2541
2542         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Treat result
2543         value as of type ffi_arg, not unsigned int.
2544
2545 2004-01-21  Michael Ritzert  <ritzert@t-online.de>
2546
2547         * ffi64.c (ffi_prep_args): Cast the RHS of an assignment instead
2548         of the LHS.
2549
2550 2004-01-12  Andreas Tobler  <a.tobler@schweiz.ch>
2551
2552         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_32 for
2553         Solaris.
2554
2555 2004-01-08  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2556
2557         * testsuite/libffi.call/ffitest.h (allocate_mmap): Cast MAP_FAILED
2558         to void *.
2559
2560 2003-12-10  Richard Henderson  <rth@redhat.com>
2561
2562         * testsuite/libffi.call/cls_align_pointer.c: Cast pointers to
2563         size_t instead of int.
2564
2565 2003-12-04  Hosaka Yuji  <hos@tamanegi.org>
2566
2567         * testsuite/libffi.call/many_win32.c: Include <float.h>.
2568         * testsuite/libffi.call/many_win32.c (main): Replace variable
2569         int i with unsigned long ul.
2570
2571         * testsuite/libffi.call/cls_align_uint64.c: New test case.
2572         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2573         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2574         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2575         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2576         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2577         * testsuite/libffi.call/cls_align_float.c: Likewise.
2578         * testsuite/libffi.call/cls_align_double.c: Likewise.
2579         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2580         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2581
2582 2003-12-02  Hosaka Yuji  <hos@tamanegi.org>
2583
2584         PR other/13221
2585         * src/x86/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV):
2586         Align arguments to 32 bits.
2587
2588 2003-12-01  Andreas Tobler  <a.tobler@schweiz.ch>
2589
2590         PR other/13221
2591         * testsuite/libffi.call/cls_multi_sshort.c: New test case.
2592         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2593         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2594         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2595         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2596         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2597
2598         * testsuite/libffi.special/unwindtest.cc: Cosmetics.
2599
2600 2003-11-26  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
2601
2602         * testsuite/libffi.call/ffitest.h: Include <fcntl.h>.
2603         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2604
2605 2003-11-22  Andreas Tobler  <a.tobler@schweiz.ch>
2606
2607         * Makefile.in: Rebuilt.
2608         * configure: Likewise.
2609         * testsuite/libffi.special/unwindtest.cc: Convert the mmap to
2610         the right type.
2611
2612 2003-11-21  Andreas Jaeger  <aj@suse.de>
2613             Andreas Tobler  <a.tobler@schweiz.ch>
2614
2615         * acinclude.m4: Add AC_FUNC_MMAP_BLACKLIST.
2616         * configure.in: Call AC_FUNC_MMAP_BLACKLIST.
2617         * Makefile.in: Rebuilt.
2618         * aclocal.m4: Likewise.
2619         * configure: Likewise.
2620         * fficonfig.h.in: Likewise.
2621         * testsuite/lib/libffi-dg.exp: Add include dir.
2622         * testsuite/libffi.call/ffitest.h: Add MMAP definitions.
2623         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2624         * testsuite/libffi.call/closure_fn0.c: Use MMAP functionality
2625         for ffi_closure if available.
2626         * testsuite/libffi.call/closure_fn1.c: Likewise.
2627         * testsuite/libffi.call/closure_fn2.c: Likewise.
2628         * testsuite/libffi.call/closure_fn3.c: Likewise.
2629         * testsuite/libffi.call/closure_fn4.c: Likewise.
2630         * testsuite/libffi.call/closure_fn5.c: Likewise.
2631         * testsuite/libffi.call/cls_12byte.c: Likewise.
2632         * testsuite/libffi.call/cls_16byte.c: Likewise.
2633         * testsuite/libffi.call/cls_18byte.c: Likewise.
2634         * testsuite/libffi.call/cls_19byte.c: Likewise.
2635         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2636         * testsuite/libffi.call/cls_20byte.c: Likewise.
2637         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2638         * testsuite/libffi.call/cls_24byte.c: Likewise.
2639         * testsuite/libffi.call/cls_2byte.c: Likewise.
2640         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2641         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2642         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2643         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2644         * testsuite/libffi.call/cls_4byte.c: Likewise.
2645         * testsuite/libffi.call/cls_5byte.c: Likewise.
2646         * testsuite/libffi.call/cls_64byte.c: Likewise.
2647         * testsuite/libffi.call/cls_6byte.c: Likewise.
2648         * testsuite/libffi.call/cls_7byte.c: Likewise.
2649         * testsuite/libffi.call/cls_8byte.c: Likewise.
2650         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2651         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2652         * testsuite/libffi.call/cls_double.c: Likewise.
2653         * testsuite/libffi.call/cls_float.c: Likewise.
2654         * testsuite/libffi.call/cls_schar.c: Likewise.
2655         * testsuite/libffi.call/cls_sint.c: Likewise.
2656         * testsuite/libffi.call/cls_sshort.c: Likewise.
2657         * testsuite/libffi.call/cls_uchar.c: Likewise.
2658         * testsuite/libffi.call/cls_uint.c: Likewise.
2659         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2660         * testsuite/libffi.call/cls_ushort.c: Likewise.
2661         * testsuite/libffi.call/nested_struct.c: Likewise.
2662         * testsuite/libffi.call/nested_struct1.c: Likewise.
2663         * testsuite/libffi.call/nested_struct2.c: Likewise.
2664         * testsuite/libffi.call/nested_struct3.c: Likewise.
2665         * testsuite/libffi.call/problem1.c: Likewise.
2666         * testsuite/libffi.special/unwindtest.cc: Likewise.
2667
2668 2003-11-20  Andreas Tobler  <a.tobler@schweiz.ch>
2669
2670         * testsuite/lib/libffi-dg.exp: Make the -lgcc_s conditional.
2671
2672 2003-11-19  Andreas Tobler  <a.tobler@schweiz.ch>
2673
2674         * testsuite/lib/libffi-dg.exp: Add DYLD_LIBRARY_PATH for darwin.
2675         Add -lgcc_s to additional flags.
2676
2677 2003-11-12  Andreas Tobler  <a.tobler@schweiz.ch>
2678
2679         * configure.in, include/Makefile.am: PR libgcj/11147, install
2680         the ffitarget.h header file in a gcc versioned and target
2681         dependent place.
2682         * configure: Regenerated.
2683         * Makefile.in, include/Makefile.in: Likewise.
2684         * testsuite/Makefile.in: Likewise.
2685
2686 2003-11-09  Andreas Tobler  <a.tobler@schweiz.ch>
2687
2688         * testsuite/libffi.call/closure_fn0.c: Print result and check
2689         with dg-output to make debugging easier.
2690         * testsuite/libffi.call/closure_fn1.c: Likewise.
2691         * testsuite/libffi.call/closure_fn2.c: Likewise.
2692         * testsuite/libffi.call/closure_fn3.c: Likewise.
2693         * testsuite/libffi.call/closure_fn4.c: Likewise.
2694         * testsuite/libffi.call/closure_fn5.c: Likewise.
2695         * testsuite/libffi.call/cls_12byte.c: Likewise.
2696         * testsuite/libffi.call/cls_16byte.c: Likewise.
2697         * testsuite/libffi.call/cls_18byte.c: Likewise.
2698         * testsuite/libffi.call/cls_19byte.c: Likewise.
2699         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2700         * testsuite/libffi.call/cls_20byte.c: Likewise.
2701         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2702         * testsuite/libffi.call/cls_24byte.c: Likewise.
2703         * testsuite/libffi.call/cls_2byte.c: Likewise.
2704         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2705         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2706         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2707         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2708         * testsuite/libffi.call/cls_4byte.c: Likewise.
2709         * testsuite/libffi.call/cls_5byte.c: Likewise.
2710         * testsuite/libffi.call/cls_64byte.c: Likewise.
2711         * testsuite/libffi.call/cls_6byte.c: Likewise.
2712         * testsuite/libffi.call/cls_7byte.c: Likewise.
2713         * testsuite/libffi.call/cls_8byte.c: Likewise.
2714         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2715         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2716         * testsuite/libffi.call/cls_double.c: Likewise.
2717         * testsuite/libffi.call/cls_float.c: Likewise.
2718         * testsuite/libffi.call/cls_schar.c: Likewise.
2719         * testsuite/libffi.call/cls_sint.c: Likewise.
2720         * testsuite/libffi.call/cls_sshort.c: Likewise.
2721         * testsuite/libffi.call/cls_uchar.c: Likewise.
2722         * testsuite/libffi.call/cls_uint.c: Likewise.
2723         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2724         * testsuite/libffi.call/cls_ushort.c: Likewise.
2725         * testsuite/libffi.call/problem1.c: Likewise.
2726
2727         * testsuite/libffi.special/unwindtest.cc: Make ffi_closure
2728         static.
2729
2730 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
2731
2732         * testsuite/libffi.call/cls_9byte2.c: New test case.
2733         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2734         * testsuite/libffi.call/cls_64byte.c: Likewise.
2735         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2736         * testsuite/libffi.call/cls_19byte.c: Likewise.
2737         * testsuite/libffi.call/cls_18byte.c: Likewise.
2738         * testsuite/libffi.call/closure_fn4.c: Likewise.
2739         * testsuite/libffi.call/closure_fn5.c: Likewise.
2740         * testsuite/libffi.call/cls_schar.c: Likewise.
2741         * testsuite/libffi.call/cls_sint.c: Likewise.
2742         * testsuite/libffi.call/cls_sshort.c: Likewise.
2743         * testsuite/libffi.call/nested_struct2.c: Likewise.
2744         * testsuite/libffi.call/nested_struct3.c: Likewise.
2745
2746 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
2747
2748         * testsuite/libffi.call/cls_double.c: Do a check on the result.
2749         * testsuite/libffi.call/cls_uchar.c: Likewise.
2750         * testsuite/libffi.call/cls_uint.c: Likewise.
2751         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2752         * testsuite/libffi.call/cls_ushort.c: Likewise.
2753         * testsuite/libffi.call/return_sc.c: Cleanup whitespaces.
2754
2755 2003-11-06  Andreas Tobler  <a.tobler@schweiz.ch>
2756
2757         * src/prep_cif.c (ffi_prep_cif): Move the validity check after
2758         the initialization.
2759
2760 2003-10-23  Andreas Tobler  <a.tobler@schweiz.ch>
2761
2762         * src/java_raw_api.c (ffi_java_ptrarray_to_raw): Replace
2763         FFI_ASSERT(FALSE) with FFI_ASSERT(0).
2764
2765 2003-10-22  David Daney  <ddaney@avtrex.com>
2766
2767         * src/mips/ffitarget.h: Replace undefined UINT32 and friends with
2768         __attribute__((__mode__(__SI__))) and friends.
2769
2770 2003-10-22  Andreas Schwab  <schwab@suse.de>
2771
2772         * src/ia64/ffi.c: Replace FALSE/TRUE with false/true.
2773
2774 2003-10-21  Andreas Tobler  <a.tobler@schweiz.ch>
2775
2776         * configure.in: AC_LINK_FILES(ffitarget.h).
2777         * configure: Regenerate.
2778         * Makefile.in: Likewise.
2779         * include/Makefile.in: Likewise.
2780         * testsuite/Makefile.in: Likewise.
2781         * fficonfig.h.in: Likewise.
2782
2783 2003-10-21  Paolo Bonzini  <bonzini@gnu.org>
2784             Richard Henderson  <rth@redhat.com>
2785
2786         Avoid that ffi.h includes fficonfig.h.
2787
2788         * Makefile.am (EXTRA_DIST): Include ffitarget.h files
2789         (TARGET_SRC_MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
2790         (TARGET_SRC_MIPS_SGI): Removed.
2791         (MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
2792         (MIPS_SGI): Removed.
2793         (CLEANFILES): Removed.
2794         (mostlyclean-am, clean-am, mostlyclean-sub, clean-sub): New
2795         targets.
2796         * acconfig.h: Removed.
2797         * configure.in: Compute sizeofs only for double and long double.
2798         Use them to define and subst HAVE_LONG_DOUBLE.  Include comments
2799         into AC_DEFINE instead of using acconfig.h.  Create
2800         include/ffitarget.h instead of include/fficonfig.h.  Rename
2801         MIPS_GCC to MIPS_IRIX, drop MIPS_SGI since we are in gcc's tree.
2802         AC_DEFINE EH_FRAME_FLAGS.
2803         * include/Makefile.am (DISTCLEANFILES): New automake macro.
2804         (hack_DATA): Add ffitarget.h.
2805         * include/ffi.h.in: Remove all system specific definitions.
2806         Declare raw API even if it is not installed, why bother?
2807         Use limits.h instead of SIZEOF_* to define ffi_type_*.  Do
2808         not define EH_FRAME_FLAGS, it is in fficonfig.h now.  Include
2809         ffitarget.h instead of fficonfig.h.  Remove ALIGN macro.
2810         (UINT_ARG, INT_ARG): Removed, use ffi_arg and ffi_sarg instead.
2811         * include/ffi_common.h (bool): Do not define.
2812         (ffi_assert): Accept failed assertion.
2813         (ffi_type_test): Return void and accept file/line.
2814         (FFI_ASSERT): Pass stringized failed assertion.
2815         (FFI_ASSERT_AT): New macro.
2816         (FFI_ASSERT_VALID_TYPE): New macro.
2817         (UINT8, SINT8, UINT16, SINT16, UINT32, SINT32,
2818         UINT64, SINT64): Define here with gcc's __attribute__ macro
2819         instead of in ffi.h
2820         (FLOAT32, ALIGN): Define here instead of in ffi.h
2821         * include/ffi-mips.h: Removed.  Its content moved to
2822         src/mips/ffitarget.h after separating assembly and C sections.
2823         * src/alpha/ffi.c, src/alpha/ffi.c, src/java_raw_api.c
2824         src/prep_cif.c, src/raw_api.c, src/ia64/ffi.c,
2825         src/mips/ffi.c, src/mips/n32.S, src/mips/o32.S,
2826         src/mips/ffitarget.h, src/sparc/ffi.c, src/x86/ffi64.c:
2827         SIZEOF_ARG -> FFI_SIZEOF_ARG.
2828         * src/ia64/ffi.c: Include stdbool.h (provided by GCC 2.95+).
2829         * src/debug.c (ffi_assert): Accept stringized failed assertion.
2830         (ffi_type_test): Rewritten.
2831         * src/prep-cif.c (initialize_aggregate, ffi_prep_cif): Call
2832         FFI_ASSERT_VALID_TYPE.
2833         * src/alpha/ffitarget.h, src/arm/ffitarget.h,
2834         src/ia64/ffitarget.h, src/m68k/ffitarget.h,
2835         src/mips/ffitarget.h, src/powerpc/ffitarget.h,
2836         src/s390/ffitarget.h, src/sh/ffitarget.h,
2837         src/sh64/ffitarget.h, src/sparc/ffitarget.h,
2838         src/x86/ffitarget.h: New files.
2839         * src/alpha/osf.S, src/arm/sysv.S, src/ia64/unix.S,
2840         src/m68k/sysv.S, src/mips/n32.S, src/mips/o32.S,
2841         src/powerpc/aix.S, src/powerpc/darwin.S,
2842         src/powerpc/ffi_darwin.c, src/powerpc/linux64.S,
2843         src/powerpc/linux64_closure.S, src/powerpc/ppc_closure.S,
2844         src/powerpc/sysv.S, src/s390/sysv.S, src/sh/sysv.S,
2845         src/sh64/sysv.S, src/sparc/v8.S, src/sparc/v9.S,
2846         src/x86/sysv.S, src/x86/unix64.S, src/x86/win32.S:
2847         include fficonfig.h
2848
2849 2003-10-20  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2850
2851         * src/mips/ffi.c: Use _ABIN32, _ABIO32 instead of external
2852         _MIPS_SIM_NABI32, _MIPS_SIM_ABI32.
2853
2854 2003-10-19  Andreas Tobler  <a.tobler@schweiz.ch>
2855
2856         * src/powerpc/ffi_darwin.c (ffi_prep_args): Declare bytes again.
2857         Used when FFI_DEBUG = 1.
2858
2859 2003-10-14  Alan Modra  <amodra@bigpond.net.au>
2860
2861         * src/types.c (double, longdouble): Default POWERPC64 to 8 byte size
2862         and align.
2863
2864 2003-10-06  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2865
2866         * include/ffi_mips.h: Define FFI_MIPS_N32 for N32/N64 ABIs,
2867         FFI_MIPS_O32 for O32 ABI.
2868
2869 2003-10-01  Andreas Tobler  <a.tobler@schweiz.ch>
2870
2871         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_64 for
2872         SPARC64. Cleanup whitespaces.
2873
2874 2003-09-19  Andreas Tobler  <a.tobler@schweiz.ch>
2875
2876         * testsuite/libffi.call/closure_fn0.c: Xfail mips, arm,
2877         strongarm, xscale. Cleanup whitespaces.
2878         * testsuite/libffi.call/closure_fn1.c: Likewise.
2879         * testsuite/libffi.call/closure_fn2.c: Likewise.
2880         * testsuite/libffi.call/closure_fn3.c: Likewise.
2881         * testsuite/libffi.call/cls_12byte.c: Likewise.
2882         * testsuite/libffi.call/cls_16byte.c: Likewise.
2883         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2884         * testsuite/libffi.call/cls_20byte.c: Likewise.
2885         * testsuite/libffi.call/cls_24byte.c: Likewise.
2886         * testsuite/libffi.call/cls_2byte.c: Likewise.
2887         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2888         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2889         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2890         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2891         * testsuite/libffi.call/cls_4byte.c: Likewise.
2892         * testsuite/libffi.call/cls_5byte.c: Likewise.
2893         * testsuite/libffi.call/cls_6byte.c: Likewise.
2894         * testsuite/libffi.call/cls_7byte.c: Likewise.
2895         * testsuite/libffi.call/cls_8byte.c: Likewise.
2896         * testsuite/libffi.call/cls_double.c: Likewise.
2897         * testsuite/libffi.call/cls_float.c: Likewise.
2898         * testsuite/libffi.call/cls_uchar.c: Likewise.
2899         * testsuite/libffi.call/cls_uint.c: Likewise.
2900         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2901         * testsuite/libffi.call/cls_ushort.c: Likewise.
2902         * testsuite/libffi.call/nested_struct.c: Likewise.
2903         * testsuite/libffi.call/nested_struct1.c: Likewise.
2904         * testsuite/libffi.call/problem1.c: Likewise.
2905         * testsuite/libffi.special/unwindtest.cc: Likewise.
2906         * testsuite/libffi.call/pyobjc-tc.c: Cleanup whitespaces.
2907
2908 2003-09-18  David Edelsohn  <edelsohn@gnu.org>
2909
2910         * src/powerpc/aix.S: Cleanup whitespaces.
2911         * src/powerpc/aix_closure.S: Likewise.
2912
2913 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
2914
2915         * src/powerpc/darwin.S: Cleanup whitespaces, comment formatting.
2916         * src/powerpc/darwin_closure.S: Likewise.
2917         * src/powerpc/ffi_darwin.c: Likewise.
2918
2919 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
2920             David Edelsohn  <edelsohn@gnu.org>
2921
2922         * src/types.c (double): Add AIX and Darwin to the right TYPEDEF.
2923         * src/powerpc/aix_closure.S: Remove the pointer to the outgoing
2924         parameter stack.
2925         * src/powerpc/darwin_closure.S: Likewise.
2926         * src/powerpc/ffi_darwin.c (ffi_prep_args): Handle structures
2927         according to the Darwin/AIX ABI.
2928         (ffi_prep_cif_machdep): Likewise.
2929         (ffi_closure_helper_DARWIN): Likewise.
2930         Remove the outgoing parameter stack logic. Simplify the evaluation
2931         of the different CASE types.
2932         (ffi_prep_clousure): Avoid the casts on lvalues. Change the branch
2933         statement in the trampoline code.
2934
2935 2003-09-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2936
2937         * src/sh/ffi.c (ffi_prep_args): Take account into the alignement
2938         for the register size.
2939         (ffi_closure_helper_SYSV): Handle the structure return value
2940         address correctly.
2941         (ffi_closure_helper_SYSV): Return the appropriate type when
2942         the registers are used for the structure return value.
2943         * src/sh/sysv.S (ffi_closure_SYSV): Fix the stack layout for
2944         the 64-bit return value.  Update copyright years.
2945
2946 2003-09-17  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2947
2948         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Search in
2949         srcdir for ffi_mips.h.
2950
2951 2003-09-12  Alan Modra  <amodra@bigpond.net.au>
2952
2953         * src/prep_cif.c (initialize_aggregate): Include tail padding in
2954         structure size.
2955         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Correct
2956         placement of float result.
2957         * testsuite/libffi.special/unwindtest.cc (closure_test_fn1): Correct
2958         cast of "resp" for big-endian 64 bit machines.
2959
2960 2003-09-11  Alan Modra  <amodra@bigpond.net.au>
2961
2962         * src/types.c (double, longdouble): Merge identical SH and ARM
2963         typedefs, and add POWERPC64.
2964         * src/powerpc/ffi.c (ffi_prep_args64): Correct next_arg calc for
2965         struct split over gpr and rest.
2966         (ffi_prep_cif_machdep): Correct intarg_count for structures.
2967         * src/powerpc/linux64.S (ffi_call_LINUX64): Fix gpr offsets.
2968
2969 2003-09-09  Andreas Tobler  <a.tobler@schweiz.ch>
2970
2971         * src/powerpc/ffi.c (ffi_closure_helper_SYSV) Handle struct
2972         passing correctly.
2973
2974 2003-09-09  Alan Modra  <amodra@bigpond.net.au>
2975
2976         * configure: Regenerate.
2977
2978 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
2979
2980         * Makefile.am: Remove build rules for ffitest.
2981         * Makefile.in: Rebuilt.
2982
2983 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
2984
2985         * src/java_raw_api.c: Include <stdlib.h> to fix compiler warning
2986         about implicit declaration of abort().
2987
2988 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
2989
2990         * Makefile.am: Add dejagnu test framework. Fixes PR other/11411.
2991         * Makefile.in: Rebuilt.
2992         * configure.in: Add dejagnu test framework.
2993         * configure: Rebuilt.
2994
2995         * testsuite/Makefile.am: New file.
2996         * testsuite/Makefile.in: Built
2997         * testsuite/lib/libffi-dg.exp: New file.
2998         * testsuite/config/default.exp: Likewise.
2999         * testsuite/libffi.call/call.exp: Likewise.
3000         * testsuite/libffi.call/ffitest.h: Likewise.
3001         * testsuite/libffi.call/closure_fn0.c: Likewise.
3002         * testsuite/libffi.call/closure_fn1.c: Likewise.
3003         * testsuite/libffi.call/closure_fn2.c: Likewise.
3004         * testsuite/libffi.call/closure_fn3.c: Likewise.
3005         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3006         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3007         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3008         * testsuite/libffi.call/cls_2byte.c: Likewise.
3009         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3010         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3011         * testsuite/libffi.call/cls_4byte.c: Likewise.
3012         * testsuite/libffi.call/cls_5byte.c: Likewise.
3013         * testsuite/libffi.call/cls_6byte.c: Likewise.
3014         * testsuite/libffi.call/cls_7byte.c: Likewise.
3015         * testsuite/libffi.call/cls_8byte.c: Likewise.
3016         * testsuite/libffi.call/cls_12byte.c: Likewise.
3017         * testsuite/libffi.call/cls_16byte.c: Likewise.
3018         * testsuite/libffi.call/cls_20byte.c: Likewise.
3019         * testsuite/libffi.call/cls_24byte.c: Likewise.
3020         * testsuite/libffi.call/cls_double.c: Likewise.
3021         * testsuite/libffi.call/cls_float.c: Likewise.
3022         * testsuite/libffi.call/cls_uchar.c: Likewise.
3023         * testsuite/libffi.call/cls_uint.c: Likewise.
3024         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3025         * testsuite/libffi.call/cls_ushort.c: Likewise.
3026         * testsuite/libffi.call/float.c: Likewise.
3027         * testsuite/libffi.call/float1.c: Likewise.
3028         * testsuite/libffi.call/float2.c: Likewise.
3029         * testsuite/libffi.call/many.c: Likewise.
3030         * testsuite/libffi.call/many_win32.c: Likewise.
3031         * testsuite/libffi.call/nested_struct.c: Likewise.
3032         * testsuite/libffi.call/nested_struct1.c: Likewise.
3033         * testsuite/libffi.call/pyobjc-tc.c: Likewise.
3034         * testsuite/libffi.call/problem1.c: Likewise.
3035         * testsuite/libffi.call/promotion.c: Likewise.
3036         * testsuite/libffi.call/return_ll.c: Likewise.
3037         * testsuite/libffi.call/return_sc.c: Likewise.
3038         * testsuite/libffi.call/return_uc.c: Likewise.
3039         * testsuite/libffi.call/strlen.c: Likewise.
3040         * testsuite/libffi.call/strlen_win32.c: Likewise.
3041         * testsuite/libffi.call/struct1.c: Likewise.
3042         * testsuite/libffi.call/struct2.c: Likewise.
3043         * testsuite/libffi.call/struct3.c: Likewise.
3044         * testsuite/libffi.call/struct4.c: Likewise.
3045         * testsuite/libffi.call/struct5.c: Likewise.
3046         * testsuite/libffi.call/struct6.c: Likewise.
3047         * testsuite/libffi.call/struct7.c: Likewise.
3048         * testsuite/libffi.call/struct8.c: Likewise.
3049         * testsuite/libffi.call/struct9.c: Likewise.
3050         * testsuite/libffi.special/special.exp: New file.
3051         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3052         * testsuite/libffi.special/unwindtest.cc: Likewise.
3053
3054
3055 2003-08-13  Kaz Kojima  <kkojima@gcc.gnu.org>
3056
3057         * src/sh/ffi.c (OFS_INT16): Set 0 for little endian case.  Update
3058         copyright years.
3059
3060 2003-08-02  Alan Modra  <amodra@bigpond.net.au>
3061
3062         * src/powerpc/ffi.c (ffi_prep_args64): Modify for changed gcc
3063         structure passing.
3064         (ffi_closure_helper_LINUX64): Likewise.
3065         * src/powerpc/linux64.S: Remove code writing to parm save area.
3066         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Use return
3067         address in lr from ffi_closure_helper_LINUX64 call to calculate
3068         table address.  Optimize function tail.
3069
3070 2003-07-28  Andreas Tobler  <a.tobler@schweiz.ch>
3071
3072         * src/sparc/ffi.c: Handle all floating point registers.
3073         * src/sparc/v9.S: Likewise. Fixes second part of PR target/11410.
3074
3075 2003-07-11  Gerald Pfeifer  <pfeifer@dbai.tuwien.ac.at>
3076
3077         * README: Note that libffi is not part of GCC.  Update the project
3078         URL and status.
3079
3080 2003-06-19  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3081
3082         * src/powerpc/ppc_closure.S: Include ffi.h.
3083
3084 2003-06-13  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3085
3086         * src/x86/sysv.S: Avoid gas-only .uleb128/.sleb128 directives.
3087         Use C style comments.
3088
3089 2003-06-13  Kaz Kojima  <kkojima@rr.iij4u.or.jp>
3090
3091         * Makefile.am: Add SHmedia support.  Fix a typo of SH support.
3092         * Makefile.in: Regenerate.
3093         * configure.in (sh64-*-linux*, sh5*-*-linux*): Add target.
3094         * configure: Regenerate.
3095         * include/ffi.h.in: Add SHmedia support.
3096         * src/sh64/ffi.c: New file.
3097         * src/sh64/sysv.S: New file.
3098
3099 2003-05-16  Jakub Jelinek  <jakub@redhat.com>
3100
3101         * configure.in (HAVE_RO_EH_FRAME): Check whether .eh_frame section
3102         should be read-only.
3103         * configure: Rebuilt.
3104         * fficonfig.h.in: Rebuilt.
3105         * include/ffi.h.in (EH_FRAME_FLAGS): Define.
3106         * src/alpha/osf.S: Use EH_FRAME_FLAGS.
3107         * src/powerpc/linux64.S: Likewise.
3108         * src/powerpc/linux64_closure.S: Likewise.  Include ffi.h.
3109         * src/powerpc/sysv.S: Use EH_FRAME_FLAGS.  Use pcrel encoding
3110         if -fpic/-fPIC/-mrelocatable.
3111         * src/powerpc/powerpc_closure.S: Likewise.
3112         * src/sparc/v8.S: If HAVE_RO_EH_FRAME is defined, don't include
3113         #write in .eh_frame flags.
3114         * src/sparc/v9.S: Likewise.
3115         * src/x86/unix64.S: Use EH_FRAME_FLAGS.
3116         * src/x86/sysv.S: Likewise.  Use pcrel encoding if -fpic/-fPIC.
3117         * src/s390/sysv.S: Use EH_FRAME_FLAGS.  Include ffi.h.
3118
3119 2003-05-07  Jeff Sturm  <jsturm@one-point.com>
3120
3121         Fixes PR bootstrap/10656
3122         * configure.in (HAVE_AS_REGISTER_PSEUDO_OP): Test assembler
3123         support for .register pseudo-op.
3124         * src/sparc/v8.S: Use it.
3125         * fficonfig.h.in: Rebuilt.
3126         * configure: Rebuilt.
3127
3128 2003-04-18  Jakub Jelinek  <jakub@redhat.com>
3129
3130         * include/ffi.h.in (POWERPC64): Define if 64-bit.
3131         (enum ffi_abi): Add FFI_LINUX64 on POWERPC.
3132         Make it the default on POWERPC64.
3133         (FFI_TRAMPOLINE_SIZE): Define to 24 on POWERPC64.
3134         * configure.in: Change powerpc-*-linux* into powerpc*-*-linux*.
3135         * configure: Rebuilt.
3136         * src/powerpc/ffi.c (hidden): Define.
3137         (ffi_prep_args_SYSV): Renamed from
3138         ffi_prep_args.  Cast pointers to unsigned long to shut up warnings.
3139         (NUM_GPR_ARG_REGISTERS64, NUM_FPR_ARG_REGISTERS64,
3140         ASM_NEEDS_REGISTERS64): New.
3141         (ffi_prep_args64): New function.
3142         (ffi_prep_cif_machdep): Handle FFI_LINUX64 ABI.
3143         (ffi_call): Likewise.
3144         (ffi_prep_closure): Likewise.
3145         (flush_icache): Surround by #ifndef POWERPC64.
3146         (ffi_dblfl): New union type.
3147         (ffi_closure_helper_SYSV): Use it to avoid aliasing problems.
3148         (ffi_closure_helper_LINUX64): New function.
3149         * src/powerpc/ppc_closure.S: Surround whole file by #ifndef
3150         __powerpc64__.
3151         * src/powerpc/sysv.S: Likewise.
3152         (ffi_call_SYSV): Rename ffi_prep_args to ffi_prep_args_SYSV.
3153         * src/powerpc/linux64.S: New file.
3154         * src/powerpc/linux64_closure.S: New file.
3155         * Makefile.am (EXTRA_DIST): Add src/powerpc/linux64.S and
3156         src/powerpc/linux64_closure.S.
3157         (TARGET_SRC_POWERPC): Likewise.
3158
3159         * src/ffitest.c (closure_test_fn, closure_test_fn1, closure_test_fn2,
3160         closure_test_fn3): Fix result printing on big-endian 64-bit
3161         machines.
3162         (main): Print tst2_arg instead of uninitialized tst2_result.
3163
3164         * src/ffitest.c (main): Hide what closure pointer really points to
3165         from the compiler.
3166
3167 2003-04-16  Richard Earnshaw  <rearnsha@arm.com>
3168
3169         * configure.in (arm-*-netbsdelf*): Add configuration.
3170         (configure): Regenerated.
3171
3172 2003-04-04  Loren J. Rittle  <ljrittle@acm.org>
3173
3174         * include/Makefile.in: Regenerate.
3175
3176 2003-03-21  Zdenek Dvorak  <rakdver@atrey.karlin.mff.cuni.cz>
3177
3178         * libffi/include/ffi.h.in: Define X86 instead of X86_64 in 32
3179         bit mode.
3180         * libffi/src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV):
3181         Receive closure pointer through parameter, read args using
3182         __builtin_dwarf_cfa.
3183         (FFI_INIT_TRAMPOLINE): Send closure reference through eax.
3184
3185 2003-03-12  Andreas Schwab  <schwab@suse.de>
3186
3187         * configure.in: Avoid trailing /. in toolexeclibdir.
3188         * configure: Rebuilt.
3189
3190 2003-03-03  Andreas Tobler <a.tobler@schweiz.ch>
3191
3192         * src/powerpc/darwin_closure.S: Recode to fit dynamic libraries.
3193
3194 2003-02-06  Andreas Tobler <a.tobler@schweiz.ch>
3195
3196         * libffi/src/powerpc/darwin_closure.S:
3197         Fix alignement bug, allocate 8 bytes for the result.
3198         * libffi/src/powerpc/aix_closure.S:
3199         Likewise.
3200         * libffi/src/powerpc/ffi_darwin.c:
3201         Update stackframe description for aix/darwin_closure.S.
3202
3203 2003-02-06  Jakub Jelinek  <jakub@redhat.com>
3204
3205         * src/s390/ffi.c (ffi_closure_helper_SYSV): Add hidden visibility
3206         attribute.
3207
3208 2003-01-31  Christian Cornelssen  <ccorn@cs.tu-berlin.de>,
3209             Andreas Schwab  <schwab@suse.de>
3210
3211         * configure.in: Adjust command to source config-ml.in to account
3212         for changes to the libffi_basedir definition.
3213         (libffi_basedir): Remove ${srcdir} from value and include trailing
3214         slash if nonempty.
3215
3216         * configure: Regenerate.
3217
3218 2003-01-29  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3219
3220         * src/powerpc/ppc_closure.S: Recode to fit shared libs.
3221
3222 2003-01-28  Andrew Haley  <aph@redhat.com>
3223
3224         * include/ffi.h.in: Enable FFI_CLOSURES for x86_64.
3225         * src/x86/ffi64.c (ffi_prep_closure): New.
3226         (ffi_closure_UNIX64_inner): New.
3227         * src/x86/unix64.S (ffi_closure_UNIX64): New.
3228
3229 2003-01-27  Alexandre Oliva  <aoliva@redhat.com>
3230
3231         * configure.in (toolexecdir, toolexeclibdir): Set and AC_SUBST.
3232         Remove USE_LIBDIR conditional.
3233         * Makefile.am (toolexecdir, toolexeclibdir): Don't override.
3234         * Makefile.in, configure: Rebuilt.
3235
3236 2003-01027  David Edelsohn  <edelsohn@gnu.org>
3237
3238         * Makefile.am (TARGET_SRC_POWERPC_AIX): Fix typo.
3239         * Makefile.in: Regenerate.
3240
3241 2003-01-22  Andrew Haley  <aph@redhat.com>
3242
3243         * src/powerpc/darwin.S (_ffi_call_AIX): Add Augmentation size to
3244         unwind info.
3245
3246 2003-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
3247
3248         * src/powerpc/darwin.S: Add unwind info.
3249         * src/powerpc/darwin_closure.S: Likewise.
3250
3251 2003-01-14  Andrew Haley  <aph@redhat.com>
3252
3253         * src/x86/ffi64.c (ffi_prep_args): Check for void retval.
3254         (ffi_prep_cif_machdep): Likewise.
3255         * src/x86/unix64.S: Add unwind info.
3256
3257 2003-01-14  Andreas Jaeger  <aj@suse.de>
3258
3259         * src/ffitest.c (main): Only use ffi_closures if those are
3260         supported.
3261
3262 2003-01-13 Andreas Tobler <a.tobler@schweiz.ch>
3263
3264         * libffi/src/ffitest.c
3265          add closure testcases
3266
3267 2003-01-13 Kevin B. Hendricks <khendricks@ivey.uwo.ca>
3268
3269         * libffi/src/powerpc/ffi.c
3270          fix alignment bug for float (4 byte aligned iso 8 byte)
3271
3272 2003-01-09  Geoffrey Keating  <geoffk@apple.com>
3273
3274         * src/powerpc/ffi_darwin.c: Remove RCS version string.
3275         * src/powerpc/darwin.S: Remove RCS version string.
3276
3277 2003-01-03  Jeff Sturm  <jsturm@one-point.com>
3278
3279         * include/ffi.h.in: Add closure defines for SPARC, SPARC64.
3280         * src/ffitest.c (main): Use static storage for closure.
3281         * src/sparc/ffi.c (ffi_prep_closure, ffi_closure_sparc_inner): New.
3282         * src/sparc/v8.S (ffi_closure_v8): New.
3283         * src/sparc/v9.S (ffi_closure_v9): New.
3284
3285 2002-11-10  Ranjit Mathew <rmathew@hotmail.com>
3286
3287         * include/ffi.h.in: Added FFI_STDCALL ffi_type
3288           enumeration for X86_WIN32.
3289         * src/x86/win32.S: Added ffi_call_STDCALL function
3290           definition.
3291         * src/x86/ffi.c (ffi_call/ffi_raw_call): Added
3292           switch cases for recognising FFI_STDCALL and
3293           calling ffi_call_STDCALL if target is X86_WIN32.
3294         * src/ffitest.c (my_stdcall_strlen/stdcall_many):
3295           stdcall versions of the "my_strlen" and "many"
3296           test functions (for X86_WIN32).
3297           Added test cases to test stdcall invocation using
3298           these functions.
3299
3300 2002-12-02  Kaz Kojima  <kkojima@gcc.gnu.org>
3301
3302         * src/sh/sysv.S: Add DWARF2 unwind info.
3303
3304 2002-11-27  Ulrich Weigand  <uweigand@de.ibm.com>
3305
3306         * src/s390/sysv.S (.eh_frame section): Make section read-only.
3307
3308 2002-11-26  Jim Wilson  <wilson@redhat.com>
3309
3310         * src/types.c (FFI_TYPE_POINTER): Has size 8 on IA64.
3311
3312 2002-11-23  H.J. Lu <hjl@gnu.org>
3313
3314         * acinclude.m4: Add dummy AM_PROG_LIBTOOL.
3315         Include ../config/accross.m4.
3316         * aclocal.m4; Rebuild.
3317         * configure: Likewise.
3318
3319 2002-11-15  Ulrich Weigand  <uweigand@de.ibm.com>
3320
3321         * src/s390/sysv.S (.eh_frame section): Adapt to pcrel FDE encoding.
3322
3323 2002-11-11  DJ Delorie  <dj@redhat.com>
3324
3325         * configure.in: Look for common files in the right place.
3326
3327 2002-10-08  Ulrich Weigand  <uweigand@de.ibm.com>
3328
3329         * src/java_raw_api.c (ffi_java_raw_to_ptrarray): Interpret
3330         raw data as _Jv_word values, not ffi_raw.
3331         (ffi_java_ptrarray_to_raw): Likewise.
3332         (ffi_java_rvalue_to_raw): New function.
3333         (ffi_java_raw_call): Call it.
3334         (ffi_java_raw_to_rvalue): New function.
3335         (ffi_java_translate_args): Call it.
3336         * src/ffitest.c (closure_test_fn): Interpret return value
3337         as ffi_arg, not int.
3338         * src/s390/ffi.c (ffi_prep_cif_machdep): Add missing
3339         FFI_TYPE_POINTER case.
3340         (ffi_closure_helper_SYSV): Likewise.  Also, assume return
3341         values extended to word size.
3342
3343 2002-10-02  Andreas Jaeger  <aj@suse.de>
3344
3345         * src/x86/ffi64.c (ffi_prep_cif_machdep): Remove debug output.
3346
3347 2002-10-01  Bo Thorsen  <bo@smetana.suse.de>
3348
3349         * include/ffi.h.in: Fix i386 win32 compilation.
3350
3351 2002-09-30  Ulrich Weigand  <uweigand@de.ibm.com>
3352
3353         * configure.in: Add s390x-*-linux-* target.
3354         * configure: Regenerate.
3355         * include/ffi.h.in: Define S390X for s390x targets.
3356         (FFI_CLOSURES): Define for s390/s390x.
3357         (FFI_TRAMPOLINE_SIZE): Likewise.
3358         (FFI_NATIVE_RAW_API): Likewise.
3359         * src/prep_cif.c (ffi_prep_cif): Do not compute stack space for s390.
3360         * src/types.c (FFI_TYPE_POINTER): Use 8-byte pointers on s390x.
3361         * src/s390/ffi.c: Major rework of existing code.  Add support for
3362         s390x targets.  Add closure support.
3363         * src/s390/sysv.S: Likewise.
3364
3365 2002-09-29  Richard Earnshaw  <rearnsha@arm.com>
3366
3367         * src/arm/sysv.S: Fix typo.
3368
3369 2002-09-28  Richard Earnshaw  <rearnsha@arm.com>
3370
3371         * src/arm/sysv.S: If we don't have machine/asm.h and the pre-processor
3372         has defined __USER_LABEL_PREFIX__, then use it in CNAME.
3373         (ffi_call_SYSV): Handle soft-float.
3374
3375 2002-09-27  Bo Thorsen  <bo@suse.de>
3376
3377         * include/ffi.h.in: Fix multilib x86-64 support.
3378
3379 2002-09-22  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
3380
3381         * Makefile.am (all-multi): Fix multilib parallel build.
3382
3383 2002-07-19  Kaz Kojima  <kkojima@gcc.gnu.org>
3384
3385         * configure.in (sh[34]*-*-linux*): Add brackets.
3386         * configure: Regenerate.
3387
3388 2002-07-18  Kaz Kojima  <kkojima@gcc.gnu.org>
3389
3390         * Makefile.am: Add SH support.
3391         * Makefile.in: Regenerate.
3392         * configure.in (sh-*-linux*, sh[34]*-*-linux*): Add target.
3393         * configure: Regenerate.
3394         * include/ffi.h.in: Add SH support.
3395         * src/sh/ffi.c: New file.
3396         * src/sh/sysv.S: New file.
3397         * src/types.c: Add SH support.
3398
3399 2002-07-16  Bo Thorsen  <bo@suse.de>
3400
3401         * src/x86/ffi64.c: New file that adds x86-64 support.
3402         * src/x86/unix64.S: New file that handles argument setup for
3403         x86-64.
3404         * src/x86/sysv.S: Don't use this on x86-64.
3405         * src/x86/ffi.c: Don't use this on x86-64.
3406         Remove unused vars.
3407         * src/prep_cif.c (ffi_prep_cif): Don't do stack size calculation
3408         for x86-64.
3409         * src/ffitest.c (struct6): New test that tests a special case in
3410         the x86-64 ABI.
3411         (struct7): Likewise.
3412         (struct8): Likewise.
3413         (struct9): Likewise.
3414         (closure_test_fn): Silence warning about this when it's not used.
3415         (main): Add the new tests.
3416         (main): Fix a couple of wrong casts and silence some compiler warnings.
3417         * include/ffi.h.in: Add x86-64 ABI definition.
3418         * fficonfig.h.in: Regenerate.
3419         * Makefile.am: Add x86-64 support.
3420         * configure.in: Likewise.
3421         * Makefile.in: Regenerate.
3422         * configure: Likewise.
3423
3424 2002-06-24  Bo Thorsen  <bo@suse.de>
3425
3426         * src/types.c: Merge settings for similar architectures.
3427         Add x86-64 sizes and alignments.
3428
3429 2002-06-23  Bo Thorsen  <bo@suse.de>
3430
3431         * src/arm/ffi.c (ffi_prep_args): Remove unused vars.
3432         * src/sparc/ffi.c (ffi_prep_args_v8): Likewise.
3433         * src/mips/ffi.c (ffi_prep_args): Likewise.
3434         * src/m68k/ffi.c (ffi_prep_args): Likewise.
3435
3436 2002-07-18  H.J. Lu  (hjl@gnu.org)