libffi/ChangeLog

   1 2008-02-12  Bjoern Koenig  <bkoenig@alpha-tierchen.de>
   2             Andreas Tobler  <a.tobler@schweiz.org>
   3
   4         * configure.ac: Add amd64-*-freebsd* target.
   5         * configure: Regenerate.
   6
   7 2008-01-30  H.J. Lu  <hongjiu.lu@intel.com>
   8
   9         PR libffi/34612
  10         * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when
  11         returning struct.
  12
  13         * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer"
  14         tests.
  15
  16 2008-01-24  David Edelsohn  <edelsohn@gnu.org>
  17
  18         * configure: Regenerate.
  19
  20 2008-01-06  Andreas Tobler  <a.tobler@schweiz.org>
  21
  22         * src/x86/ffi.c (ffi_prep_cif_machdep): Fix thinko.
  23
  24 2008-01-05  Andreas Tobler  <a.tobler@schweiz.org>
  25
  26         PR testsuite/32843
  27         * src/x86/ffi.c (ffi_prep_cif_machdep): Add code for
  28         signed/unsigned int8/16 for X86_DARWIN.
  29         Updated copyright info.
  30         Handle one and two byte structs with special cif->flags.
  31         * src/x86/ffitarget.h: Add special types for one and two byte structs.
  32         Updated copyright info.
  33         * src/x86/darwin.S (ffi_call_SYSV): Rewrite to use a jump table like
  34         sysv.S
  35         Remove code to pop args from the stack after call.
  36         Special-case signed/unsigned for int8/16, one and two byte structs.
  37         (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
  38         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
  39         FFI_TYPE_SINT32.
  40         Updated copyright info.
  41
  42 2007-12-08  David Daney  <ddaney@avtrex.com>
  43
  44         * src/mips/n32.S (ffi_call_N32):  Replace dadd with ADDU, dsub with
  45         SUBU, add with ADDU and use smaller code sequences.
  46
  47 2007-12-07  David Daney  <ddaney@avtrex.com>
  48
  49         * src/mips/ffi.c (ffi_prep_cif_machdep): Handle long double return
  50         type.
  51
  52 2007-12-06  David Daney  <ddaney@avtrex.com>
  53
  54         * include/ffi.h.in (FFI_SIZEOF_JAVA_RAW): Define if not already
  55         defined.
  56         (ffi_java_raw): New typedef.
  57         (ffi_java_raw_call, ffi_java_ptrarray_to_raw,
  58         ffi_java_raw_to_ptrarray): Change parameter types from ffi_raw to
  59         ffi_java_raw.
  60         (ffi_java_raw_closure) : Same.
  61         (ffi_prep_java_raw_closure, ffi_prep_java_raw_closure_loc): Change
  62         parameter types.
  63         * src/java_raw_api.c (ffi_java_raw_size):  Replace FFI_SIZEOF_ARG with
  64         FFI_SIZEOF_JAVA_RAW.
  65         (ffi_java_raw_to_ptrarray): Change type of raw to ffi_java_raw.
  66         Replace FFI_SIZEOF_ARG with FFI_SIZEOF_JAVA_RAW. Use
  67         sizeof(ffi_java_raw) for alignment calculations.
  68         (ffi_java_ptrarray_to_raw): Same.
  69         (ffi_java_rvalue_to_raw): Add special handling for FFI_TYPE_POINTER
  70         if FFI_SIZEOF_JAVA_RAW == 4.
  71         (ffi_java_raw_to_rvalue): Same.
  72         (ffi_java_raw_call): Change type of raw to ffi_java_raw.
  73         (ffi_java_translate_args): Same.
  74         (ffi_prep_java_raw_closure_loc, ffi_prep_java_raw_closure): Change
  75         parameter types.
  76         * src/mips/ffitarget.h (FFI_SIZEOF_JAVA_RAW): Define for N32 ABI.
  77
  78 2007-12-06  David Daney  <ddaney@avtrex.com>
  79
  80         * src/mips/n32.S (ffi_closure_N32): Use 64-bit add instruction on
  81         pointer values.
  82
  83 2007-12-01  Andreas Tobler  <a.tobler@schweiz.org>
  84
  85         PR libffi/31937
  86         * src/powerpc/ffitarget.h: Introduce new ABI FFI_LINUX_SOFT_FLOAT.
  87         Add local FFI_TYPE_UINT128 to handle soft-float long-double-128.
  88         * src/powerpc/ffi.c: Distinguish between __NO_FPRS__ and not and
  89         set the NUM_FPR_ARG_REGISTERS according to.
  90         Add support for potential soft-float support under hard-float
  91         architecture.
  92         (ffi_prep_args_SYSV): Set NUM_FPR_ARG_REGISTERS to 0 in case of
  93         FFI_LINUX_SOFT_FLOAT, handle float, doubles and long-doubles according
  94         to the FFI_LINUX_SOFT_FLOAT ABI.
  95         (ffi_prep_cif_machdep): Likewise.
  96         (ffi_closure_helper_SYSV): Likewise.
  97         * src/powerpc/ppc_closure.S: Make sure not to store float/double
  98         on archs where __NO_FPRS__ is true.
  99         Add FFI_TYPE_UINT128 support.
 100         * src/powerpc/sysv.S: Add support for soft-float long-double-128.
 101         Adjust copyright notice.
 102
 103 2007-11-25  Andreas Tobler  <a.tobler@schweiz.org>
 104
 105         * src/closures.c: Move defintion of MAYBE_UNUSED from here to ...
 106         * include/ffi_common.h: ... here.
 107         Update copyright.
 108
 109 2007-11-17  Andreas Tobler  <a.tobler@schweiz.org>
 110
 111         * src/powerpc/sysv.S: Load correct cr to compare if we have long double.
 112         * src/powerpc/linux64.S: Likewise.
 113         * src/powerpc/ffi.c: Add a comment to show which part goes into cr6.
 114         * testsuite/libffi.call/return_ldl.c: New test.
 115
 116 2007-09-04    <aph@redhat.com>
 117
 118         * src/arm/sysv.S (UNWIND): New.
 119         (Whole file): Conditionally compile unwinder directives.
 120         * src/arm/sysv.S: Add unwinder directives.
 121
 122         * src/arm/ffi.c (ffi_prep_args): Align structs by at least 4 bytes.
 123         Only treat r0 as a struct address if we're actually returning a
 124         struct by address.
 125         Only copy the bytes that are actually within a struct.
 126         (ffi_prep_cif_machdep): A Composite Type not larger than 4 bytes
 127         is returned in r0, not passed by address.
 128         (ffi_call): Allocate a word-sized temporary for the case where
 129         a composite is returned in r0.
 130         (ffi_prep_incoming_args_SYSV): Align as necessary.
 131
 132 2007-08-05  Steven Newbury  <s_j_newbury@yahoo.co.uk>
 133
 134         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Use __clear_cache instead of
 135         directly using the sys_cacheflush syscall.
 136
 137 2007-07-27  Andrew Haley  <aph@redhat.com>
 138
 139         * src/arm/sysv.S (ffi_closure_SYSV): Add soft-float.
 140
 141 2007-09-03  Maciej W. Rozycki  <macro@linux-mips.org>
 142
 143         * Makefile.am: Unify MIPS_IRIX and MIPS_LINUX into MIPS.
 144         * configure.ac: Likewise.
 145         * Makefile.in: Regenerate.
 146         * include/Makefile.in: Likewise.
 147         * testsuite/Makefile.in: Likewise.
 148         * configure: Likewise.
 149
 150 2007-08-24  David Daney  <ddaney@avtrex.com>
 151
 152         * testsuite/libffi.call/return_sl.c: New test.
 153
 154 2007-08-10  David Daney  <ddaney@avtrex.com>
 155
 156         * testsuite/libffi.call/cls_multi_ushort.c,
 157         testsuite/libffi.call/cls_align_uint16.c,
 158         testsuite/libffi.call/nested_struct1.c,
 159         testsuite/libffi.call/nested_struct3.c,
 160         testsuite/libffi.call/cls_7_1_byte.c,
 161         testsuite/libffi.call/nested_struct5.c,
 162         testsuite/libffi.call/cls_double.c,
 163         testsuite/libffi.call/nested_struct7.c,
 164         testsuite/libffi.call/cls_sint.c,
 165         testsuite/libffi.call/nested_struct9.c,
 166         testsuite/libffi.call/cls_20byte1.c,
 167         testsuite/libffi.call/cls_multi_sshortchar.c,
 168         testsuite/libffi.call/cls_align_sint64.c,
 169         testsuite/libffi.call/cls_3byte2.c,
 170         testsuite/libffi.call/cls_multi_schar.c,
 171         testsuite/libffi.call/cls_multi_uchar.c,
 172         testsuite/libffi.call/cls_19byte.c,
 173         testsuite/libffi.call/cls_9byte1.c,
 174         testsuite/libffi.call/cls_align_float.c,
 175         testsuite/libffi.call/closure_fn1.c,
 176         testsuite/libffi.call/problem1.c,
 177         testsuite/libffi.call/closure_fn3.c,
 178         testsuite/libffi.call/cls_sshort.c,
 179         testsuite/libffi.call/closure_fn5.c,
 180         testsuite/libffi.call/cls_align_double.c,
 181         testsuite/libffi.call/nested_struct.c,
 182         testsuite/libffi.call/cls_2byte.c,
 183         testsuite/libffi.call/nested_struct10.c,
 184         testsuite/libffi.call/cls_4byte.c,
 185         testsuite/libffi.call/cls_6byte.c,
 186         testsuite/libffi.call/cls_8byte.c,
 187         testsuite/libffi.call/cls_multi_sshort.c,
 188         testsuite/libffi.call/cls_align_sint16.c,
 189         testsuite/libffi.call/cls_align_uint32.c,
 190         testsuite/libffi.call/cls_20byte.c,
 191         testsuite/libffi.call/cls_float.c,
 192         testsuite/libffi.call/nested_struct2.c,
 193         testsuite/libffi.call/cls_5_1_byte.c,
 194         testsuite/libffi.call/nested_struct4.c,
 195         testsuite/libffi.call/cls_24byte.c,
 196         testsuite/libffi.call/nested_struct6.c,
 197         testsuite/libffi.call/cls_64byte.c,
 198         testsuite/libffi.call/nested_struct8.c,
 199         testsuite/libffi.call/cls_uint.c,
 200         testsuite/libffi.call/cls_multi_ushortchar.c,
 201         testsuite/libffi.call/cls_schar.c,
 202         testsuite/libffi.call/cls_uchar.c,
 203         testsuite/libffi.call/cls_align_uint64.c,
 204         testsuite/libffi.call/cls_ulonglong.c,
 205         testsuite/libffi.call/cls_align_longdouble.c,
 206         testsuite/libffi.call/cls_1_1byte.c,
 207         testsuite/libffi.call/cls_12byte.c,
 208         testsuite/libffi.call/cls_3_1byte.c,
 209         testsuite/libffi.call/cls_3byte1.c,
 210         testsuite/libffi.call/cls_4_1byte.c,
 211         testsuite/libffi.call/cls_6_1_byte.c,
 212         testsuite/libffi.call/cls_16byte.c,
 213         testsuite/libffi.call/cls_18byte.c,
 214         testsuite/libffi.call/closure_fn0.c,
 215         testsuite/libffi.call/cls_9byte2.c,
 216         testsuite/libffi.call/closure_fn2.c,
 217         testsuite/libffi.call/closure_fn4.c,
 218         testsuite/libffi.call/cls_ushort.c,
 219         testsuite/libffi.call/closure_fn6.c,
 220         testsuite/libffi.call/cls_5byte.c,
 221         testsuite/libffi.call/cls_align_pointer.c,
 222         testsuite/libffi.call/cls_7byte.c,
 223         testsuite/libffi.call/cls_align_sint32.c,
 224         testsuite/libffi.special/unwindtest_ffi_call.cc,
 225         testsuite/libffi.special/unwindtest.cc: Remove xfail for mips64*-*-*.
 226
 227 2007-08-10  David Daney  <ddaney@avtrex.com>
 228
 229         PR libffi/28313
 230         * configure.ac: Don't treat mips64 as a special case.
 231         * Makefile.am (nodist_libffi_la_SOURCES): Add n32.S.
 232         * configure: Regenerate
 233         * Makefile.in: Ditto.
 234         * fficonfig.h.in: Ditto.
 235         * src/mips/ffitarget.h (REG_L, REG_S, SUBU, ADDU, SRL, LI): Indent.
 236         (LA, EH_FRAME_ALIGN, FDE_ADDR_BYTES): New preprocessor macros.
 237         (FFI_DEFAULT_ABI): Set for n64 case.
 238         (FFI_CLOSURES, FFI_TRAMPOLINE_SIZE): Define for n32 and n64 cases.
 239         * src/mips/n32.S (ffi_call_N32): Add debug macros and labels for FDE.
 240         (ffi_closure_N32): New function.
 241         (.eh_frame): New section
 242         * src/mips/o32.S: Clean up comments.
 243         (ffi_closure_O32): Pass ffi_closure parameter in $12.
 244         * src/mips/ffi.c: Use FFI_MIPS_N32 instead of
 245         _MIPS_SIM == _ABIN32 throughout.
 246         (FFI_MIPS_STOP_HERE): New, use in place of
 247         ffi_stop_here.
 248         (ffi_prep_args): Use unsigned long to hold pointer values.  Rewrite
 249         to support n32/n64 ABIs.
 250         (calc_n32_struct_flags): Rewrite.
 251         (calc_n32_return_struct_flags): Remove unused variable.  Reverse
 252         position of flag bits.
 253         (ffi_prep_cif_machdep): Rewrite n32 portion.
 254         (ffi_call): Enable for n64.  Add special handling for small structure
 255         return values.
 256         (ffi_prep_closure_loc): Add n32 and n64 support.
 257         (ffi_closure_mips_inner_O32): Add cast to silence warning.
 258         (copy_struct_N32, ffi_closure_mips_inner_N32): New functions.
 259
 260 2007-08-08  David Daney  <ddaney@avtrex.com>
 261
 262         * testsuite/libffi.call/ffitest.h (ffi_type_mylong): Remove definition.
 263         * testsuite/libffi.call/cls_align_uint16.c (main): Use correct type
 264         specifiers.
 265         * testsuite/libffi.call/nested_struct1.c (main): Ditto.
 266         * testsuite/libffi.call/cls_sint.c (main): Ditto.
 267         * testsuite/libffi.call/nested_struct9.c (main): Ditto.
 268         * testsuite/libffi.call/cls_20byte1.c (main): Ditto.
 269         * testsuite/libffi.call/cls_9byte1.c (main): Ditto.
 270         * testsuite/libffi.call/closure_fn1.c (main): Ditto.
 271         * testsuite/libffi.call/closure_fn3.c (main): Ditto.
 272         * testsuite/libffi.call/return_dbl2.c (main): Ditto.
 273         * testsuite/libffi.call/cls_sshort.c (main): Ditto.
 274         * testsuite/libffi.call/return_fl3.c (main): Ditto.
 275         * testsuite/libffi.call/closure_fn5.c (main): Ditto.
 276         * testsuite/libffi.call/nested_struct.c (main): Ditto.
 277         * testsuite/libffi.call/nested_struct10.c (main): Ditto.
 278         * testsuite/libffi.call/return_ll1.c (main): Ditto.
 279         * testsuite/libffi.call/cls_8byte.c (main): Ditto.
 280         * testsuite/libffi.call/cls_align_uint32.c (main): Ditto.
 281         * testsuite/libffi.call/cls_align_sint16.c (main): Ditto.
 282         * testsuite/libffi.call/cls_20byte.c (main): Ditto.
 283         * testsuite/libffi.call/nested_struct2.c (main): Ditto.
 284         * testsuite/libffi.call/cls_24byte.c (main): Ditto.
 285         * testsuite/libffi.call/nested_struct6.c (main): Ditto.
 286         * testsuite/libffi.call/cls_uint.c (main): Ditto.
 287         * testsuite/libffi.call/cls_12byte.c (main): Ditto.
 288         * testsuite/libffi.call/cls_16byte.c (main): Ditto.
 289         * testsuite/libffi.call/closure_fn0.c (main): Ditto.
 290         * testsuite/libffi.call/cls_9byte2.c (main): Ditto.
 291         * testsuite/libffi.call/closure_fn2.c (main): Ditto.
 292         * testsuite/libffi.call/return_dbl1.c (main): Ditto.
 293         * testsuite/libffi.call/closure_fn4.c (main): Ditto.
 294         * testsuite/libffi.call/closure_fn6.c (main): Ditto.
 295         * testsuite/libffi.call/cls_align_sint32.c (main): Ditto.
 296
 297 2007-08-07  Andrew Haley  <aph@redhat.com>
 298
 299         * src/x86/sysv.S (ffi_closure_raw_SYSV): Fix typo in previous
 300         checkin.
 301
 302 2007-08-06  Andrew Haley  <aph@redhat.com>
 303
 304         PR testsuite/32843
 305         * src/x86/sysv.S (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
 306         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
 307         FFI_TYPE_SINT32.
 308
 309 2007-08-02  David Daney  <ddaney@avtrex.com>
 310
 311         * testsuite/libffi.call/return_ul.c (main): Define return type as
 312         ffi_arg.  Use proper printf conversion specifier.
 313
 314 2007-07-30  Andrew Haley  <aph@redhat.com>
 315
 316         PR testsuite/32843
 317         * src/x86/ffi.c (ffi_prep_cif_machdep): in x86 case, add code for
 318         signed/unsigned int8/16.
 319         * src/x86/sysv.S (ffi_call_SYSV): Rewrite to:
 320         Use a jump table.
 321         Remove code to pop args from the stack after call.
 322         Special-case signed/unsigned int8/16.
 323         * testsuite/libffi.call/return_sc.c (main): Revert.
 324
 325 2007-07-26  Richard Guenther  <rguenther@suse.de>
 326
 327         PR testsuite/32843
 328         * testsuite/libffi.call/return_sc.c (main): Verify call
 329         result as signed char, not ffi_arg.
 330
 331 2007-07-16  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
 332
 333         * configure.ac (i?86-*-solaris2.1[0-9]): Set TARGET to X86_64.
 334         * configure: Regenerate.
 335
 336 2007-07-11  David Daney  <ddaney@avtrex.com>
 337
 338         * src/mips/ffi.c: Don't include sys/cachectl.h.
 339         (ffi_prep_closure_loc): Use __builtin___clear_cache() instead of
 340         cacheflush().
 341
 342 2007-05-18  Aurelien Jarno  <aurelien@aurel32.net>
 343
 344         * src/arm/ffi.c (ffi_prep_closure_loc): Renamed and ajusted
 345         from (ffi_prep_closure): ... this.
 346         (FFI_INIT_TRAMPOLINE): Adjust.
 347
 348 2005-12-31  Phil Blundell  <pb@reciva.com>
 349
 350         * src/arm/ffi.c (ffi_prep_incoming_args_SYSV,
 351         ffi_closure_SYSV_inner, ffi_prep_closure): New, add closure support.
 352         * src/arm/sysv.S(ffi_closure_SYSV): Likewise.
 353         * src/arm/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
 354         (FFI_CLOSURES): Enable closure support.
 355
 356 2007-07-03  Andrew Haley  <aph@hedges.billgatliff.com>
 357
 358         * testsuite/libffi.call/cls_multi_ushort.c,
 359         testsuite/libffi.call/cls_align_uint16.c,
 360         testsuite/libffi.call/nested_struct1.c,
 361         testsuite/libffi.call/nested_struct3.c,
 362         testsuite/libffi.call/cls_7_1_byte.c,
 363         testsuite/libffi.call/cls_double.c,
 364         testsuite/libffi.call/nested_struct5.c,
 365         testsuite/libffi.call/nested_struct7.c,
 366         testsuite/libffi.call/cls_sint.c,
 367         testsuite/libffi.call/nested_struct9.c,
 368         testsuite/libffi.call/cls_20byte1.c,
 369         testsuite/libffi.call/cls_multi_sshortchar.c,
 370         testsuite/libffi.call/cls_align_sint64.c,
 371         testsuite/libffi.call/cls_3byte2.c,
 372         testsuite/libffi.call/cls_multi_schar.c,
 373         testsuite/libffi.call/cls_multi_uchar.c,
 374         testsuite/libffi.call/cls_19byte.c,
 375         testsuite/libffi.call/cls_9byte1.c,
 376         testsuite/libffi.call/cls_align_float.c,
 377         testsuite/libffi.call/closure_fn1.c,
 378         testsuite/libffi.call/problem1.c,
 379         testsuite/libffi.call/closure_fn3.c,
 380         testsuite/libffi.call/cls_sshort.c,
 381         testsuite/libffi.call/closure_fn5.c,
 382         testsuite/libffi.call/cls_align_double.c,
 383         testsuite/libffi.call/cls_2byte.c,
 384         testsuite/libffi.call/nested_struct.c,
 385         testsuite/libffi.call/nested_struct10.c,
 386         testsuite/libffi.call/cls_4byte.c,
 387         testsuite/libffi.call/cls_6byte.c,
 388         testsuite/libffi.call/cls_8byte.c,
 389         testsuite/libffi.call/cls_multi_sshort.c,
 390         testsuite/libffi.call/cls_align_uint32.c,
 391         testsuite/libffi.call/cls_align_sint16.c,
 392         testsuite/libffi.call/cls_float.c,
 393         testsuite/libffi.call/cls_20byte.c,
 394         testsuite/libffi.call/cls_5_1_byte.c,
 395         testsuite/libffi.call/nested_struct2.c,
 396         testsuite/libffi.call/cls_24byte.c,
 397         testsuite/libffi.call/nested_struct4.c,
 398         testsuite/libffi.call/nested_struct6.c,
 399         testsuite/libffi.call/cls_64byte.c,
 400         testsuite/libffi.call/nested_struct8.c,
 401         testsuite/libffi.call/cls_uint.c,
 402         testsuite/libffi.call/cls_multi_ushortchar.c,
 403         testsuite/libffi.call/cls_schar.c,
 404         testsuite/libffi.call/cls_uchar.c,
 405         testsuite/libffi.call/cls_align_uint64.c,
 406         testsuite/libffi.call/cls_ulonglong.c,
 407         testsuite/libffi.call/cls_align_longdouble.c,
 408         testsuite/libffi.call/cls_1_1byte.c,
 409         testsuite/libffi.call/cls_12byte.c,
 410         testsuite/libffi.call/cls_3_1byte.c,
 411         testsuite/libffi.call/cls_3byte1.c,
 412         testsuite/libffi.call/cls_4_1byte.c,
 413         testsuite/libffi.call/cls_6_1_byte.c,
 414         testsuite/libffi.call/cls_16byte.c,
 415         testsuite/libffi.call/cls_18byte.c,
 416         testsuite/libffi.call/closure_fn0.c,
 417         testsuite/libffi.call/cls_9byte2.c,
 418         testsuite/libffi.call/closure_fn2.c,
 419         testsuite/libffi.call/closure_fn4.c,
 420         testsuite/libffi.call/cls_ushort.c,
 421         testsuite/libffi.call/closure_fn6.c,
 422         testsuite/libffi.call/cls_5byte.c,
 423         testsuite/libffi.call/cls_align_pointer.c,
 424         testsuite/libffi.call/cls_7byte.c,
 425         testsuite/libffi.call/cls_align_sint32.c,
 426         testsuite/libffi.special/unwindtest_ffi_call.cc,
 427         testsuite/libffi.special/unwindtest.cc: Enable for ARM.
 428
 429 2007-07-05  H.J. Lu  <hongjiu.lu@intel.com>
 430
 431         * aclocal.m4: Regenerated.
 432
 433 2007-06-02  Paolo Bonzini  <bonzini@gnu.org>
 434
 435         * configure: Regenerate.
 436
 437 2007-05-23  Steve Ellcey  <sje@cup.hp.com>
 438
 439         * Makefile.in: Regenerate.
 440         * configure: Regenerate.
 441         * aclocal.m4: Regenerate.
 442         * include/Makefile.in: Regenerate.
 443         * testsuite/Makefile.in: Regenerate.
 444
 445 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
 446
 447         * src/m68k/ffi.c (ffi_prep_incoming_args_SYSV,
 448         ffi_closure_SYSV_inner,ffi_prep_closure): New, add closure support.
 449         * src/m68k/sysv.S(ffi_closure_SYSV,ffi_closure_struct_SYSV): Likewise.
 450         * src/m68k/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
 451         (FFI_CLOSURES): Enable closure support.
 452
 453 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
 454
 455         * configure.ac (HAVE_AS_CFI_PSEUDO_OP): New test.
 456         * configure: Regenerate.
 457         * fficonfig.h.in: Regenerate.
 458         * src/m68k/sysv.S (CFI_STARTPROC,CFI_ENDPROC,
 459         CFI_OFFSET,CFI_DEF_CFA): New macros.
 460         (ffi_call_SYSV): Add callframe annotation.
 461
 462 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
 463
 464         * src/m68k/ffi.c (ffi_prep_args,ffi_prep_cif_machdep): Fix
 465         numerous test suite failures.
 466         * src/m68k/sysv.S (ffi_call_SYSV): Likewise.
 467
 468 2007-04-11  Paolo Bonzini  <bonzini@gnu.org>
 469
 470         * Makefile.am (EXTRA_DIST): Bring up to date.
 471         * Makefile.in: Regenerate.
 472         * src/frv/eabi.S: Remove RCS keyword.
 473
 474 2007-04-06  Richard Henderson  <rth@redhat.com>
 475
 476         * configure.ac: Tidy target case.
 477         (HAVE_LONG_DOUBLE): Allow the target to override.
 478         * configure: Regenerate.
 479         * include/ffi.h.in: Don't define ffi_type_foo if
 480         LIBFFI_HIDE_BASIC_TYPES is defined.
 481         (ffi_type_longdouble): If not HAVE_LONG_DOUBLE, define
 482         to ffi_type_double.
 483         * types.c (LIBFFI_HIDE_BASIC_TYPES): Define.
 484         (FFI_TYPEDEF, ffi_type_void): Mark the data const.
 485         (ffi_type_longdouble): Special case for Alpha.  Don't define
 486         if long double == double.
 487
 488         * src/alpha/ffi.c (FFI_TYPE_LONGDOUBLE): Assert unique value.
 489         (ffi_prep_cif_machdep): Handle it as the 128-bit type.
 490         (ffi_call, ffi_closure_osf_inner): Likewise.
 491         (ffi_closure_osf_inner): Likewise.  Mark hidden.
 492         (ffi_call_osf, ffi_closure_osf): Mark hidden.
 493         * src/alpha/ffitarget.h (FFI_LAST_ABI): Tidy definition.
 494         * src/alpha/osf.S (ffi_call_osf, ffi_closure_osf): Mark hidden.
 495         (load_table): Handle 128-bit long double.
 496
 497         * testsuite/libffi.call/float4.c: Add -mieee for alpha.
 498
 499 2007-04-06  Tom Tromey  <tromey@redhat.com>
 500
 501         PR libffi/31491:
 502         * README: Fixed bug in example.
 503
 504 2007-04-03  Jakub Jelinek  <jakub@redhat.com>
 505
 506         * src/closures.c: Include sys/statfs.h.
 507         (_GNU_SOURCE): Define on Linux.
 508         (FFI_MMAP_EXEC_SELINUX): Define.
 509         (selinux_enabled): New variable.
 510         (selinux_enabled_check): New function.
 511         (is_selinux_enabled): Define.
 512         (dlmmap): Use it.
 513
 514 2007-03-24  Uros Bizjak  <ubizjak@gmail.com>
 515
 516         * testsuite/libffi.call/return_fl2.c (return_fl): Mark as static.
 517         Use 'volatile float sum' to create sum of floats to avoid false
 518         negative due to excess precision on ix86 targets.
 519         (main): Ditto.
 520
 521 2007-03-08  Alexandre Oliva  <aoliva@redhat.com>
 522
 523         * src/powerpc/ffi.c (flush_icache): Fix left-over from previous
 524         patch.
 525         (ffi_prep_closure_loc): Remove unneeded casts.  Add needed ones.
 526
 527 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
 528
 529         * include/ffi.h.in (ffi_closure_alloc, ffi_closure_free): New.
 530         (ffi_prep_closure_loc): New.
 531         (ffi_prep_raw_closure_loc): New.
 532         (ffi_prep_java_raw_closure_loc): New.
 533         * src/closures.c: New file.
 534         * src/dlmalloc.c [FFI_MMAP_EXEC_WRIT] (struct malloc_segment):
 535         Replace sflags with exec_offset.
 536         [FFI_MMAP_EXEC_WRIT] (mmap_exec_offset, add_segment_exec_offset,
 537         sub_segment_exec_offset): New macros.
 538         (get_segment_flags, set_segment_flags, check_segment_merge): New
 539         macros.
 540         (is_mmapped_segment, is_extern_segment): Use get_segment_flags.
 541         (add_segment, sys_alloc, create_mspace, create_mspace_with_base,
 542         destroy_mspace): Use new macros.
 543         (sys_alloc): Silence warning.
 544         * Makefile.am (libffi_la_SOURCES): Add src/closures.c.
 545         * Makefile.in: Rebuilt.
 546         * src/prep_cif [FFI_CLOSURES] (ffi_prep_closure): Implement in
 547         terms of ffi_prep_closure_loc.
 548         * src/raw_api.c (ffi_prep_raw_closure_loc): Renamed and adjusted
 549         from...
 550         (ffi_prep_raw_closure): ... this.  Re-implement in terms of the
 551         renamed version.
 552         * src/java_raw_api (ffi_prep_java_raw_closure_loc): Renamed and
 553         adjusted from...
 554         (ffi_prep_java_raw_closure): ... this.  Re-implement in terms of
 555         the renamed version.
 556         * src/alpha/ffi.c (ffi_prep_closure_loc): Renamed from
 557         (ffi_prep_closure): ... this.
 558         * src/pa/ffi.c: Likewise.
 559         * src/cris/ffi.c: Likewise.  Adjust.
 560         * src/frv/ffi.c: Likewise.
 561         * src/ia64/ffi.c: Likewise.
 562         * src/mips/ffi.c: Likewise.
 563         * src/powerpc/ffi_darwin.c: Likewise.
 564         * src/s390/ffi.c: Likewise.
 565         * src/sh/ffi.c: Likewise.
 566         * src/sh64/ffi.c: Likewise.
 567         * src/sparc/ffi.c: Likewise.
 568         * src/x86/ffi64.c: Likewise.
 569         * src/x86/ffi.c: Likewise.
 570         (FFI_INIT_TRAMPOLINE): Adjust.
 571         (ffi_prep_raw_closure_loc): Renamed and adjusted from...
 572         (ffi_prep_raw_closure): ... this.
 573         * src/powerpc/ffi.c (ffi_prep_closure_loc): Renamed from
 574         (ffi_prep_closure): ... this.
 575         (flush_icache): Adjust.
 576
 577 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
 578
 579         * src/dlmalloc.c: New file, imported version 2.8.3 of Doug
 580         Lea's malloc.
 581
 582 2007-03-01  Brooks Moses  <brooks.moses@codesourcery.com>
 583
 584         * Makefile.am: Add dummy install-pdf target.
 585         * Makefile.in: Regenerate
 586
 587 2007-02-13  Andreas Krebbel  <krebbel1@de.ibm.com>
 588
 589         * src/s390/ffi.c (ffi_prep_args, ffi_prep_cif_machdep,
 590         ffi_closure_helper_SYSV): Add long double handling.
 591
 592 2007-02-02  Jakub Jelinek  <jakub@redhat.com>
 593
 594         * src/powerpc/linux64.S (ffi_call_LINUX64): Move restore of r2
 595         immediately after bctrl instruction.
 596
 597 2007-01-18  Alexandre Oliva  <aoliva@redhat.com>
 598
 599         * Makefile.am (all-recursive, install-recursive,
 600         mostlyclean-recursive, clean-recursive, distclean-recursive,
 601         maintainer-clean-recursive): Add missing targets.
 602         * Makefile.in: Rebuilt.
 603
 604 2006-12-14  Andreas Tobler  <a.tobler@schweiz.org>
 605
 606         * configure.ac: Add TARGET for x86_64-*-darwin*.
 607         * Makefile.am (nodist_libffi_la_SOURCES): Add rules for 64-bit sources
 608         for X86_DARWIN.
 609         * src/x86/ffitarget.h: Set trampoline size for x86_64-*-darwin*.
 610         * src/x86/darwin64.S: New file for x86_64-*-darwin* support.
 611         * configure: Regenerate.
 612         * Makefile.in: Regenerate.
 613         * include/Makefile.in: Regenerate.
 614         * testsuite/Makefile.in: Regenerate.
 615         * testsuite/libffi.special/unwindtest_ffi_call.cc: New test case for
 616         ffi_call only.
 617
 618 2006-12-13  Andreas Tobler <a.tobler@schweiz.org>
 619
 620         * aclocal.m4: Regenerate with aclocal -I .. as written in the
 621         Makefile.am.
 622
 623 2006-10-31  Geoffrey Keating  <geoffk@apple.com>
 624
 625         * src/powerpc/ffi_darwin.c (darwin_adjust_aggregate_sizes): New.
 626         (ffi_prep_cif_machdep): Call darwin_adjust_aggregate_sizes for
 627         Darwin.
 628         * testsuite/libffi.call/nested_struct4.c: Remove Darwin XFAIL.
 629         * testsuite/libffi.call/nested_struct6.c: Remove Darwin XFAIL.
 630
 631 2006-10-10  Paolo Bonzini  <bonzini@gnu.org>
 632             Sandro Tolaini  <tolaini@libero.it>
 633
 634         * configure.ac [i*86-*-darwin*]: Set X86_DARWIN symbol and
 635         conditional.
 636         * configure: Regenerated.
 637         * Makefile.am (nodist_libffi_la_SOURCES) [X86_DARWIN]: New case.
 638         (EXTRA_DIST): Add src/x86/darwin.S.
 639         * Makefile.in: Regenerated.
 640         * include/Makefile.in: Regenerated.
 641         * testsuite/Makefile.in: Regenerated.
 642
 643         * src/x86/ffi.c (ffi_prep_cif_machdep) [X86_DARWIN]: Treat like
 644         X86_WIN32, and additionally align stack to 16 bytes.
 645         * src/x86/darwin.S: New, based on sysv.S.
 646         * src/prep_cif.c (ffi_prep_cif) [X86_DARWIN]: Align > 8-byte structs.
 647
 648 2006-09-12  David Daney  <ddaney@avtrex.com>
 649
 650         PR libffi/23935
 651         * include/Makefile.am: Install both ffi.h and ffitarget.h in
 652         $(libdir)/gcc/$(target_alias)/$(gcc_version)/include.
 653         * aclocal.m4: Regenerated for automake 1.9.6.
 654         * Makefile.in: Regenerated.
 655         * include/Makefile.in: Regenerated.
 656         * testsuite/Makefile.in: Regenerated.
 657
 658 2006-08-17  Andreas Tobler  <a.tobler@schweiz.ch>
 659
 660         * include/ffi_common.h (struct): Revert accidental commit.
 661
 662 2006-08-15  Andreas Tobler  <a.tobler@schweiz.ch>
 663
 664         * include/ffi_common.h: Remove lint directives.
 665         * include/ffi.h.in: Likewise.
 666
 667 2006-07-25  Torsten Schoenfeld  <kaffeetisch@gmx.de>
 668
 669         * include/ffi.h.in (ffi_type_ulong, ffi_type_slong): Define correctly
 670         for 32-bit architectures.
 671         * testsuite/libffi.call/return_ul.c: New test case.
 672
 673 2006-07-19  David Daney  <ddaney@avtrex.com>
 674
 675         * testsuite/libffi.call/closure_fn6.c: Remove xfail for mips,
 676         xfail remains for mips64.
 677
 678 2006-05-23  Carlos O'Donell  <carlos@codesourcery.com>
 679
 680         * Makefile.am: Add install-html target. Add install-html to .PHONY
 681         * Makefile.in: Regenerate.
 682         * aclocal.m4: Regenerate.
 683         * include/Makefile.in: Regenerate.
 684         * testsuite/Makefile.in: Regenerate.
 685
 686 2006-05-18  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
 687
 688         * pa/ffi.c (ffi_prep_args_pa32): Load floating point arguments from
 689         stack slot.
 690
 691 2006-04-22  Andreas Tobler  <a.tobler@schweiz.ch>
 692
 693         * README: Remove notice about 'Crazy Comments'.
 694         * src/debug.c: Remove lint directives. Cleanup white spaces.
 695         * src/java_raw_api.c: Likewise.
 696         * src/prep_cif.c: Likewise.
 697         * src/raw_api.c: Likewise.
 698         * src/ffitest.c: Delete. No longer needed, all test cases migrated
 699         to the testsuite.
 700         * src/arm/ffi.c: Remove lint directives.
 701         * src/m32r/ffi.c: Likewise.
 702         * src/pa/ffi.c: Likewise.
 703         * src/powerpc/ffi.c: Likewise.
 704         * src/powerpc/ffi_darwin.c: Likewise.
 705         * src/sh/ffi.c: Likewise.
 706         * src/sh64/ffi.c: Likewise.
 707         * src/x86/ffi.c: Likewise.
 708         * testsuite/libffi.call/float2.c: Likewise.
 709         * testsuite/libffi.call/promotion.c: Likewise.
 710         * testsuite/libffi.call/struct1.c: Likewise.
 711
 712 2006-04-13  Andreas Tobler  <a.tobler@schweiz.ch>
 713
 714         * src/pa/hpux32.S: Correct unwind offset calculation for
 715         ffi_closure_pa32.
 716         * src/pa/linux.S: Likewise.
 717
 718 2006-04-12  James E Wilson  <wilson@specifix.com>
 719
 720         PR libgcj/26483
 721         * src/ia64/ffi.c (stf_spill, ldf_fill): Rewrite as macros.
 722         (hfa_type_load): Call stf_spill.
 723         (hfa_type_store): Call ldf_fill.
 724         (ffi_call): Adjust calls to above routines.  Add local temps for
 725         macro result.
 726
 727 2006-04-10  Matthias Klose  <doko@debian.org>
 728
 729         * testsuite/lib/libffi-dg.exp (libffi-init): Recognize multilib
 730         directory names containing underscores.
 731
 732 2006-04-07  James E Wilson  <wilson@specifix.com>
 733
 734         * testsuite/libffi.call/float4.c: New testcase.
 735
 736 2006-04-05  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
 737             Andreas Tobler  <a.tobler@schweiz.ch>
 738
 739         * Makefile.am: Add PA_HPUX port.
 740         * Makefile.in: Regenerate.
 741         * include/Makefile.in: Likewise.
 742         * testsuite/Makefile.in: Likewise.
 743         * configure.ac: Add PA_HPUX rules.
 744         * configure: Regenerate.
 745         * src/pa/ffitarget.h: Rename linux target to PA_LINUX.
 746         Add PA_HPUX and PA64_HPUX.
 747         Rename FFI_LINUX ABI to FFI_PA32 ABI.
 748         (FFI_TRAMPOLINE_SIZE): Define for 32-bit HP-UX targets.
 749         (FFI_TYPE_SMALL_STRUCT2): Define.
 750         (FFI_TYPE_SMALL_STRUCT4): Likewise.
 751         (FFI_TYPE_SMALL_STRUCT8): Likewise.
 752         (FFI_TYPE_SMALL_STRUCT3): Redefine.
 753         (FFI_TYPE_SMALL_STRUCT5): Likewise.
 754         (FFI_TYPE_SMALL_STRUCT6): Likewise.
 755         (FFI_TYPE_SMALL_STRUCT7): Likewise.
 756         * src/pa/ffi.c (ROUND_DOWN): Delete.
 757         (fldw, fstw, fldd, fstd): Use '__asm__'.
 758         (ffi_struct_type): Add support for FFI_TYPE_SMALL_STRUCT2,
 759         FFI_TYPE_SMALL_STRUCT4 and FFI_TYPE_SMALL_STRUCT8.
 760         (ffi_prep_args_LINUX): Rename to ffi_prep_args_pa32. Update comment.
 761         Simplify incrementing of stack slot variable. Change type of local
 762         'n' to unsigned int.
 763         (ffi_size_stack_LINUX): Rename to ffi_size_stack_pa32. Handle long
 764         double on PA_HPUX.
 765         (ffi_prep_cif_machdep): Likewise.
 766         (ffi_call): Likewise.
 767         (ffi_closure_inner_LINUX): Rename to ffi_closure_inner_pa32. Change
 768         return type to ffi_status. Simplify incrementing of stack slot
 769         variable. Only copy floating point argument registers when PA_LINUX
 770         is true. Reformat debug statement.
 771         Add support for FFI_TYPE_SMALL_STRUCT2, FFI_TYPE_SMALL_STRUCT4 and
 772         FFI_TYPE_SMALL_STRUCT8.
 773         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Add 'extern' to
 774         declaration.
 775         (ffi_prep_closure): Make linux trampoline conditional on PA_LINUX.
 776         Add nops to cache flush.  Add trampoline for PA_HPUX.
 777         * src/pa/hpux32.S: New file.
 778         * src/pa/linux.S (ffi_call_LINUX): Rename to ffi_call_pa32. Rename
 779         ffi_prep_args_LINUX to ffi_prep_args_pa32.
 780         Localize labels. Add support for 2, 4 and 8-byte small structs. Handle
 781         unaligned destinations in 3, 5, 6 and 7-byte small structs. Order
 782         argument type checks so that common argument types appear first.
 783         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Rename
 784         ffi_closure_inner_LINUX to ffi_closure_inner_pa32.
 785
 786 2006-03-24  Alan Modra  <amodra@bigpond.net.au>
 787
 788         * src/powerpc/ffitarget.h (enum ffi_abi): Add FFI_LINUX.  Default
 789         for 32-bit using IBM extended double format.  Fix FFI_LAST_ABI.
 790         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Handle linux variant of
 791         FFI_TYPE_LONGDOUBLE.
 792         (ffi_prep_args64): Assert using IBM extended double.
 793         (ffi_prep_cif_machdep): Don't munge FFI_TYPE_LONGDOUBLE type.
 794         Handle FFI_LINUX FFI_TYPE_LONGDOUBLE return and args.
 795         (ffi_call): Handle FFI_LINUX.
 796         (ffi_closure_helper_SYSV): Non FFI_LINUX long double return needs
 797         gpr3 return pointer as for struct return.  Handle FFI_LINUX
 798         FFI_TYPE_LONGDOUBLE return and args.  Don't increment "nf"
 799         unnecessarily.
 800         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Load both f1 and f2
 801         for FFI_TYPE_LONGDOUBLE.  Move epilogue insns into case table.
 802         Don't use r6 as pointer to results, instead use sp offset.  Don't
 803         make a special call to load lr with case table address, instead
 804         use offset from previous call.
 805         * src/powerpc/sysv.S (ffi_call_SYSV): Save long double return.
 806         * src/powerpc/linux64.S (ffi_call_LINUX64): Simplify long double
 807         return.
 808
 809 2006-03-15  Kaz Kojima  <kkojima@gcc.gnu.org>
 810
 811         * src/sh64/ffi.c (ffi_prep_cif_machdep): Handle float arguments
 812         passed with FP registers correctly.
 813         (ffi_closure_helper_SYSV): Likewise.
 814         * src/sh64/sysv.S: Likewise.
 815
 816 2006-03-01  Andreas Tobler  <a.tobler@schweiz.ch>
 817
 818         * testsuite/libffi.special/unwindtest.cc (closure_test_fn): Mark cif,
 819         args and userdata unused.
 820         (closure_test_fn1): Mark cif and userdata unused.
 821         (main): Remove unused res.
 822
 823 2006-02-28  Andreas Tobler  <a.tobler@schweiz.ch>
 824
 825         * testsuite/libffi.call/call.exp: Adjust FSF address. Add test runs for
 826         -O2, -O3, -Os and the warning flags -W -Wall.
 827         * testsuite/libffi.special/special.exp: Likewise.
 828         * testsuite/libffi.call/ffitest.h: Add an __UNUSED__ macro to mark
 829         unused parameter unused for gcc or else do nothing.
 830         * testsuite/libffi.special/ffitestcxx.h: Likewise.
 831         * testsuite/libffi.call/cls_12byte.c (cls_struct_12byte_gn): Mark cif
 832         and userdata unused.
 833         * testsuite/libffi.call/cls_16byte.c (cls_struct_16byte_gn): Likewise.
 834         * testsuite/libffi.call/cls_18byte.c (cls_struct_18byte_gn): Likewise.
 835         * testsuite/libffi.call/cls_19byte.c (cls_struct_19byte_gn): Likewise.
 836         * testsuite/libffi.call/cls_1_1byte.c (cls_struct_1_1byte_gn): Likewise.
 837         * testsuite/libffi.call/cls_20byte.c (cls_struct_20byte_gn): Likewise.
 838         * testsuite/libffi.call/cls_20byte1.c (cls_struct_20byte_gn): Likewise.
 839         * testsuite/libffi.call/cls_24byte.c (cls_struct_24byte_gn): Likewise.
 840         * testsuite/libffi.call/cls_2byte.c (cls_struct_2byte_gn): Likewise.
 841         * testsuite/libffi.call/cls_3_1byte.c (cls_struct_3_1byte_gn): Likewise.
 842         * testsuite/libffi.call/cls_3byte1.c (cls_struct_3byte_gn): Likewise.
 843         * testsuite/libffi.call/cls_3byte2.c (cls_struct_3byte_gn1): Likewise.
 844         * testsuite/libffi.call/cls_4_1byte.c (cls_struct_4_1byte_gn): Likewise.
 845         * testsuite/libffi.call/cls_4byte.c (cls_struct_4byte_gn): Likewise.
 846         * testsuite/libffi.call/cls_5_1_byte.c (cls_struct_5byte_gn): Likewise.
 847         * testsuite/libffi.call/cls_5byte.c (cls_struct_5byte_gn): Likewise.
 848         * testsuite/libffi.call/cls_64byte.c (cls_struct_64byte_gn): Likewise.
 849         * testsuite/libffi.call/cls_6_1_byte.c (cls_struct_6byte_gn): Likewise.
 850         * testsuite/libffi.call/cls_6byte.c (cls_struct_6byte_gn): Likewise.
 851         * testsuite/libffi.call/cls_7_1_byte.c (cls_struct_7byte_gn): Likewise.
 852         * testsuite/libffi.call/cls_7byte.c (cls_struct_7byte_gn): Likewise.
 853         * testsuite/libffi.call/cls_8byte.c (cls_struct_8byte_gn): Likewise.
 854         * testsuite/libffi.call/cls_9byte1.c (cls_struct_9byte_gn): Likewise.
 855         * testsuite/libffi.call/cls_9byte2.c (cls_struct_9byte_gn): Likewise.
 856         * testsuite/libffi.call/cls_align_double.c (cls_struct_align_gn):
 857         Likewise.
 858         * testsuite/libffi.call/cls_align_float.c (cls_struct_align_gn):
 859         Likewise.
 860         * testsuite/libffi.call/cls_align_longdouble.c (cls_struct_align_gn):
 861         Likewise.
 862         * testsuite/libffi.call/cls_align_pointer.c (cls_struct_align_fn): Cast
 863         void* to avoid compiler warning.
 864         (main): Likewise.
 865         (cls_struct_align_gn): Mark cif and userdata unused.
 866         * testsuite/libffi.call/cls_align_sint16.c (cls_struct_align_gn):
 867         Likewise.
 868         * testsuite/libffi.call/cls_align_sint32.c (cls_struct_align_gn):
 869         Likewise.
 870         * testsuite/libffi.call/cls_align_sint64.c (cls_struct_align_gn):
 871         Likewise.
 872         * testsuite/libffi.call/cls_align_uint16.c (cls_struct_align_gn):
 873         Likewise.
 874         * testsuite/libffi.call/cls_align_uint32.c (cls_struct_align_gn):
 875         Likewise.
 876         * testsuite/libffi.call/cls_double.c (cls_ret_double_fn): Likewise.
 877         * testsuite/libffi.call/cls_float.c (cls_ret_float_fn): Likewise.
 878         * testsuite/libffi.call/cls_multi_schar.c (test_func_gn): Mark cif and
 879         data unused.
 880         (main): Cast res_call to silence gcc.
 881         * testsuite/libffi.call/cls_multi_sshort.c (test_func_gn): Mark cif and
 882         data unused.
 883         (main): Cast res_call to silence gcc.
 884         * testsuite/libffi.call/cls_multi_sshortchar.c (test_func_gn): Mark cif
 885         and data unused.
 886         (main): Cast res_call to silence gcc.
 887         * testsuite/libffi.call/cls_multi_uchar.c (test_func_gn): Mark cif and
 888         data unused.
 889         (main): Cast res_call to silence gcc.
 890         * testsuite/libffi.call/cls_multi_ushort.c (test_func_gn): Mark cif and
 891         data unused.
 892         (main): Cast res_call to silence gcc.
 893         * testsuite/libffi.call/cls_multi_ushortchar.c (test_func_gn): Mark cif
 894         and data unused.
 895         (main): Cast res_call to silence gcc.
 896         * testsuite/libffi.call/cls_schar.c (cls_ret_schar_fn): Mark cif and
 897         userdata unused.
 898         (cls_ret_schar_fn): Cast printf parameter to silence gcc.
 899         * testsuite/libffi.call/cls_sint.c (cls_ret_sint_fn): Mark cif and
 900         userdata unused.
 901         (cls_ret_sint_fn): Cast printf parameter to silence gcc.
 902         * testsuite/libffi.call/cls_sshort.c (cls_ret_sshort_fn): Mark cif and
 903         userdata unused.
 904         (cls_ret_sshort_fn): Cast printf parameter to silence gcc.
 905         * testsuite/libffi.call/cls_uchar.c (cls_ret_uchar_fn):  Mark cif and
 906         userdata unused.
 907         (cls_ret_uchar_fn): Cast printf parameter to silence gcc.
 908         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Mark cif and
 909         userdata unused.
 910         (cls_ret_uint_fn): Cast printf parameter to silence gcc.
 911         * testsuite/libffi.call/cls_ulonglong.c (cls_ret_ulonglong_fn): Mark cif
 912         and userdata unused.
 913         * testsuite/libffi.call/cls_ushort.c (cls_ret_ushort_fn): Mark cif and
 914         userdata unused.
 915         (cls_ret_ushort_fn): Cast printf parameter to silence gcc.
 916         * testsuite/libffi.call/float.c (floating): Remove unused parameter e.
 917         * testsuite/libffi.call/float1.c (main): Remove unused variable i.
 918         Cleanup white spaces.
 919         * testsuite/libffi.call/negint.c (checking): Remove unused variable i.
 920         * testsuite/libffi.call/nested_struct.c (cls_struct_combined_gn): Mark
 921         cif and userdata unused.
 922         * testsuite/libffi.call/nested_struct1.c (cls_struct_combined_gn):
 923         Likewise.
 924         * testsuite/libffi.call/nested_struct10.c (B_gn): Likewise.
 925         * testsuite/libffi.call/nested_struct2.c (B_fn): Adjust printf
 926         formatters to silence gcc.
 927         (B_gn): Mark cif and userdata unused.
 928         * testsuite/libffi.call/nested_struct3.c (B_gn): Mark cif and userdata
 929         unused.
 930         * testsuite/libffi.call/nested_struct4.c: Mention related PR.
 931         (B_gn): Mark cif and userdata unused.
 932         * testsuite/libffi.call/nested_struct5.c (B_gn): Mark cif and userdata
 933         unused.
 934         * testsuite/libffi.call/nested_struct6.c: Mention related PR.
 935         (B_gn): Mark cif and userdata unused.
 936         * testsuite/libffi.call/nested_struct7.c (B_gn): Mark cif and userdata
 937         unused.
 938         * testsuite/libffi.call/nested_struct8.c (B_gn): Likewise.
 939         * testsuite/libffi.call/nested_struct9.c (B_gn): Likewise.
 940         * testsuite/libffi.call/problem1.c (stub): Likewise.
 941         * testsuite/libffi.call/pyobjc-tc.c (main): Cast the result to silence
 942         gcc.
 943         * testsuite/libffi.call/return_fl2.c (return_fl): Add the note mentioned
 944         in the last commit for this test case in the test case itself.
 945         * testsuite/libffi.call/closure_fn0.c (closure_test_fn0): Mark cif as
 946         unused.
 947         * testsuite/libffi.call/closure_fn1.c (closure_test_fn1): Likewise.
 948         * testsuite/libffi.call/closure_fn2.c (closure_test_fn2): Likewise.
 949         * testsuite/libffi.call/closure_fn3.c (closure_test_fn3): Likewise.
 950         * testsuite/libffi.call/closure_fn4.c (closure_test_fn0): Likewise.
 951         * testsuite/libffi.call/closure_fn5.c (closure_test_fn5): Likewise.
 952         * testsuite/libffi.call/closure_fn6.c (closure_test_fn0): Likewise.
 953
 954 2006-02-22  Kaz Kojima  <kkojima@gcc.gnu.org>
 955
 956         * src/sh/sysv.S: Fix register numbers in the FDE for
 957         ffi_closure_SYSV.
 958
 959 2006-02-20  Andreas Tobler  <a.tobler@schweiz.ch>
 960
 961         * testsuite/libffi.call/return_fl2.c (return_fl): Remove static
 962         declaration to avoid a false negative on ix86. See PR323.
 963
 964 2006-02-18  Kaz Kojima  <kkojima@gcc.gnu.org>
 965
 966         * src/sh/ffi.c (ffi_closure_helper_SYSV): Remove unused variable
 967         and cast integer to void * if needed.  Update the pointer to
 968         the FP register saved area correctly.
 969
 970 2006-02-17  Andreas Tobler  <a.tobler@schweiz.ch>
 971
 972         * testsuite/libffi.call/nested_struct6.c: XFAIL this test until PR25630
 973         is fixed.
 974         * testsuite/libffi.call/nested_struct4.c: Likewise.
 975
 976 2006-02-16  Andreas Tobler  <a.tobler@schweiz.ch>
 977
 978         * testsuite/libffi.call/return_dbl.c: New test case.
 979         * testsuite/libffi.call/return_dbl1.c: Likewise.
 980         * testsuite/libffi.call/return_dbl2.c: Likewise.
 981         * testsuite/libffi.call/return_fl.c: Likewise.
 982         * testsuite/libffi.call/return_fl1.c: Likewise.
 983         * testsuite/libffi.call/return_fl2.c: Likewise.
 984         * testsuite/libffi.call/return_fl3.c: Likewise.
 985         * testsuite/libffi.call/closure_fn6.c: Likewise.
 986
 987         * testsuite/libffi.call/nested_struct2.c: Remove ffi_type_mylong
 988         definition.
 989         * testsuite/libffi.call/ffitest.h: Add ffi_type_mylong definition
 990         here to be used by other test cases too.
 991
 992         * testsuite/libffi.call/nested_struct10.c: New test case.
 993         * testsuite/libffi.call/nested_struct9.c: Likewise.
 994         * testsuite/libffi.call/nested_struct8.c: Likewise.
 995         * testsuite/libffi.call/nested_struct7.c: Likewise.
 996         * testsuite/libffi.call/nested_struct6.c: Likewise.
 997         * testsuite/libffi.call/nested_struct5.c: Likewise.
 998         * testsuite/libffi.call/nested_struct4.c: Likewise.
 999
1000 2006-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
1001
1002         * configure.ac: Enable libffi for sparc64-*-freebsd*.
1003         * configure: Rebuilt.
1004
1005 2006-01-18  Jakub Jelinek  <jakub@redhat.com>
1006
1007         * src/powerpc/sysv.S (smst_two_register): Don't call __ashldi3,
1008         instead do the shifting inline.
1009         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't compute %r5
1010         shift count unconditionally.  Simplify load sequences for 1, 2, 3, 4
1011         and 8 byte structs, for the remaining struct sizes don't call
1012         __lshrdi3, instead do the shifting inline.
1013
1014 2005-12-07  Thiemo Seufer  <ths@networkno.de>
1015
1016         * src/mips/ffitarget.h: Remove obsolete sgidefs.h include. Add
1017         missing parentheses.
1018         * src/mips/o32.S (ffi_call_O32): Code formatting. Define
1019         and use A3_OFF, FP_OFF, RA_OFF. Micro-optimizations.
1020         (ffi_closure_O32): Likewise, but with newly defined A3_OFF2,
1021         A2_OFF2, A1_OFF2, A0_OFF2, RA_OFF2, FP_OFF2, S0_OFF2, GP_OFF2,
1022         V1_OFF2, V0_OFF2, FA_1_1_OFF2, FA_1_0_OFF2, FA_0_1_OFF2,
1023         FA_0_0_OFF2.
1024         * src/mips/ffi.c (ffi_prep_args): Code formatting. Fix
1025         endianness bugs.
1026         (ffi_prep_closure): Improve trampoline instruction scheduling.
1027         (ffi_closure_mips_inner_O32): Fix endianness bugs.
1028
1029 2005-12-03  Alan Modra  <amodra@bigpond.net.au>
1030
1031         * src/powerpc/ffi.c: Formatting.
1032         (ffi_prep_args_SYSV): Avoid possible aliasing problems by using unions.
1033         (ffi_prep_args64): Likewise.
1034
1035 2005-09-30  Geoffrey Keating  <geoffk@apple.com>
1036
1037         * testsuite/lib/libffi-dg.exp (libffi_target_compile): For
1038         darwin, use -shared-libgcc not -lgcc_s, and explain why.
1039
1040 2005-09-26  Tom Tromey  <tromey@redhat.com>
1041
1042         * testsuite/libffi.call/float1.c (value_type): New typedef.
1043         (CANARY): New define.
1044         (main): Check for result buffer overflow.
1045         * src/powerpc/linux64.S: Handle linux64 long double returns.
1046         * src/powerpc/ffi.c (FLAG_RETURNS_128BITS): New constant.
1047         (ffi_prep_cif_machdep): Handle linux64 long double returns.
1048
1049 2005-08-25  Alan Modra  <amodra@bigpond.net.au>
1050
1051         PR target/23404
1052         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Correct placement of stack
1053         homed fp args.
1054         (ffi_status ffi_prep_cif_machdep): Correct stack sizing for same.
1055
1056 2005-08-11  Jakub Jelinek  <jakub@redhat.com>
1057
1058         * configure.ac (HAVE_HIDDEN_VISIBILITY_ATTRIBUTE): New test.
1059         (AH_BOTTOM): Add FFI_HIDDEN definition.
1060         * configure: Rebuilt.
1061         * fficonfig.h.in: Rebuilt.
1062         * src/powerpc/ffi.c (hidden): Remove.
1063         (ffi_closure_LINUX64, ffi_prep_args64, ffi_call_LINUX64,
1064         ffi_closure_helper_LINUX64): Use FFI_HIDDEN instead of hidden.
1065         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64,
1066         .ffi_closure_LINUX64): Use FFI_HIDDEN instead of .hidden.
1067         * src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV): Remove,
1068         add FFI_HIDDEN to its prototype.
1069         (ffi_closure_SYSV_inner): New.
1070         * src/x86/sysv.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
1071         * src/x86/win32.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
1072
1073 2005-08-10  Alfred M. Szmidt  <ams@gnu.org>
1074
1075         PR libffi/21819:
1076         * configure: Rebuilt.
1077         * configure.ac: Handle i*86-*-gnu*.
1078
1079 2005-08-09  Jakub Jelinek  <jakub@redhat.com>
1080
1081         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Use
1082         DW_CFA_offset_extended_sf rather than
1083         DW_CFA_GNU_negative_offset_extended.
1084         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise.
1085
1086 2005-07-22  SUGIOKA Toshinobu  <sugioka@itonet.co.jp>
1087
1088         * src/sh/sysv.S (ffi_call_SYSV): Stop argument popping correctly
1089         on sh3.
1090         (ffi_closure_SYSV): Change the stack layout for sh3 struct argument.
1091         * src/sh/ffi.c (ffi_prep_args): Fix sh3 argument copy, when it is
1092         partially on register.
1093         (ffi_closure_helper_SYSV): Likewise.
1094         (ffi_prep_cif_machdep): Don't set too many cif->flags.
1095
1096 2005-07-20  Kaz Kojima  <kkojima@gcc.gnu.org>
1097
1098         * src/sh/ffi.c (ffi_call): Handle small structures correctly.
1099         Remove empty line.
1100         * src/sh64/ffi.c (simple_type): Remove.
1101         (return_type): Handle small structures correctly.
1102         (ffi_prep_args): Likewise.
1103         (ffi_call): Likewise.
1104         (ffi_closure_helper_SYSV): Likewise.
1105         * src/sh64/sysv.S (ffi_call_SYSV): Handle 1, 2 and 4-byte return.
1106         Emit position independent code if PIC and remove wrong datalabel
1107         prefixes from EH data.
1108
1109 2005-07-19  Andreas Tobler  <a.tobler@schweiz.ch>
1110
1111         * Makefile.am (nodist_libffi_la_SOURCES): Add POWERPC_FREEBSD.
1112         * Makefile.in: Regenerate.
1113         * include/Makefile.in: Likewise.
1114         * testsuite/Makefile.in: Likewise.
1115         * configure.ac: Add POWERPC_FREEBSD rules.
1116         * configure: Regenerate.
1117         * src/powerpc/ffitarget.h: Add POWERPC_FREEBSD rules.
1118         (FFI_SYSV_TYPE_SMALL_STRUCT): Define.
1119         * src/powerpc/ffi.c: Add flags to handle small structure returns
1120         in ffi_call_SYSV.
1121         (ffi_prep_cif_machdep): Handle small structures for SYSV 4 ABI.
1122         Aka FFI_SYSV.
1123         (ffi_closure_helper_SYSV): Likewise.
1124         * src/powerpc/ppc_closure.S: Add return types for small structures.
1125         * src/powerpc/sysv.S: Add bits to handle small structures for
1126         final SYSV 4 ABI.
1127
1128 2005-07-10  Andreas Tobler  <a.tobler@schweiz.ch>
1129
1130         * testsuite/libffi.call/cls_5_1_byte.c: New test file.
1131         * testsuite/libffi.call/cls_6_1_byte.c: Likewise.
1132         * testsuite/libffi.call/cls_7_1_byte.c: Likewise.
1133
1134 2005-07-05  Randolph Chung  <tausq@debian.org>
1135
1136         * src/pa/ffi.c (ffi_struct_type): Rename FFI_TYPE_SMALL_STRUCT1
1137         as FFI_TYPE_SMALL_STRUCT3.  Break out handling for 5-7 byte
1138         structures.  Kill compilation warnings.
1139         (ffi_closure_inner_LINUX): Print return values as hex in debug
1140         message.  Rename FFI_TYPE_SMALL_STRUCT1 as FFI_TYPE_SMALL_STRUCT3.
1141         Properly handle 5-7 byte structure returns.
1142         * src/pa/ffitarget.h (FFI_TYPE_SMALL_STRUCT1)
1143         (FFI_TYPE_SMALL_STRUCT2): Remove.
1144         (FFI_TYPE_SMALL_STRUCT3, FFI_TYPE_SMALL_STRUCT5)
1145         (FFI_TYPE_SMALL_STRUCT6, FFI_TYPE_SMALL_STRUCT7): Define.
1146         * src/pa/linux.S: Mark source file as using PA1.1 assembly.
1147         (checksmst1, checksmst2): Remove.
1148         (checksmst3): Optimize handling of 3-byte struct returns.
1149         (checksmst567): Properly handle 5-7 byte struct returns.
1150
1151 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1152
1153         PR libgcj/21943
1154         * src/mips/n32.S: Enforce PIC code.
1155         * src/mips/o32.S: Likewise.
1156
1157 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1158
1159         * configure.ac: Treat i*86-*-solaris2.10 and up as X86_64.
1160         * configure: Regenerate.
1161
1162 2005-06-01  Alan Modra  <amodra@bigpond.net.au>
1163
1164         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't use JUMPTARGET
1165         to call ffi_closure_helper_SYSV.  Append @local instead.
1166         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise for ffi_prep_args_SYSV.
1167
1168 2005-05-17  Kelley Cook  <kcook@gcc.gnu.org>
1169
1170         * configure.ac: Use AC_C_BIGENDIAN instead of AC_C_BIGENDIAN_CROSS.
1171         Use AC_CHECK_SIZEOF instead of AC_COMPILE_CHECK_SIZEOF.
1172         * Makefile.am (ACLOCAL_AMFLAGS): Remove -I ../config.
1173         * aclocal.m4, configure, fficonfig.h.in, Makefile.in,
1174         include/Makefile.in, testsuite/Makefile.in: Regenerate.
1175
1176 2005-05-09  Mike Stump  <mrs@apple.com>
1177
1178         * configure: Regenerate.
1179
1180 2005-05-08  Richard Henderson  <rth@redhat.com>
1181
1182         PR libffi/21285
1183         * src/alpha/osf.S: Update unwind into to match code.
1184
1185 2005-05-04  Andreas Degert <ad@papyrus-gmbh.de>
1186             Richard Henderson  <rth@redhat.com>
1187
1188         * src/x86/ffi64.c (ffi_prep_cif_machdep): Save sse-used flag in
1189         bit 11 of flags.
1190         (ffi_call): Mask return type field.  Pass ssecount to ffi_call_unix64.
1191         (ffi_prep_closure): Set carry bit if sse-used flag set.
1192         * src/x86/unix64.S (ffi_call_unix64): Add ssecount argument.
1193         Only load sse registers if ssecount non-zero.
1194         (ffi_closure_unix64): Only save sse registers if carry set on entry.
1195
1196 2005-04-29  Ralf Corsepius  <ralf.corsepius@rtems.org>
1197
1198         * configure.ac: Add i*86-*-rtems*, sparc*-*-rtems*,
1199         powerpc-*rtems*, arm*-*-rtems*, sh-*-rtems*.
1200         * configure: Regenerate.
1201
1202 2005-04-20  Hans-Peter Nilsson  <hp@axis.com>
1203
1204         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): In regsub use,
1205         have Tcl8.3-compatible intermediate variable.
1206
1207 2005-04-18  Simon Posnjak <simon.posnjak@siol.net>
1208             Hans-Peter Nilsson  <hp@axis.com>
1209
1210         * Makefile.am: Add CRIS support.
1211         * configure.ac: Likewise.
1212         * Makefile.in, configure, testsuite/Makefile.in,
1213         include/Makefile.in: Regenerate.
1214         * src/cris: New directory.
1215         * src/cris/ffi.c, src/cris/sysv.S, src/cris/ffitarget.h: New files.
1216         * src/prep_cif.c (ffi_prep_cif): Wrap in #ifndef __CRIS__.
1217
1218         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): Replace \n with
1219         \r?\n in output tests.
1220
1221 2005-04-12  Mike Stump  <mrs@apple.com>
1222
1223         * configure: Regenerate.
1224
1225 2005-03-30  Hans Boehm  <Hans.Boehm@hp.com>
1226
1227         * src/ia64/ffitarget.h (ffi_arg): Use long long instead of DI.
1228
1229 2005-03-30  Steve Ellcey  <sje@cup.hp.com>
1230
1231         * src/ia64/ffitarget.h (ffi_arg) ADD DI attribute.
1232         (ffi_sarg) Ditto.
1233         * src/ia64/unix.S (ffi_closure_unix): Extend gp
1234         to 64 bits in ILP32 mode.
1235         Load 64 bits even for short data.
1236
1237 2005-03-23  Mike Stump  <mrs@apple.com>
1238
1239         * src/powerpc/darwin.S: Update for -m64 multilib.
1240         * src/powerpc/darwin_closure.S: Likewise.
1241
1242 2005-03-21  Zack Weinberg  <zack@codesourcery.com>
1243
1244         * configure.ac: Do not invoke TL_AC_GCC_VERSION.
1245         Do not set tool_include_dir.
1246         * aclocal.m4, configure, Makefile.in, testsuite/Makefile.in:
1247         Regenerate.
1248         * include/Makefile.am: Set gcc_version and toollibffidir.
1249         * include/Makefile.in: Regenerate.
1250
1251 2005-02-22  Andrew Haley  <aph@redhat.com>
1252
1253         * src/powerpc/ffi.c (ffi_prep_cif_machdep): Bump alignment to
1254         odd-numbered register pairs for 64-bit integer types.
1255
1256 2005-02-23  Andreas Tobler  <a.tobler@schweiz.ch>
1257
1258         PR libffi/20104
1259         * testsuite/libffi.call/return_ll1.c: New test case.
1260
1261 2005-02-11  Janis Johnson  <janis187@us.ibm.com>
1262
1263         * testsuite/libffi.call/cls_align_longdouble.c: Remove dg-options.
1264         * testsuite/libffi.call/float.c: Ditto.
1265         * testsuite/libffi.call/float2.c: Ditto.
1266         * testsuite/libffi.call/float3.c: Ditto.
1267
1268 2005-02-08  Andreas Tobler  <a.tobler@schweiz.ch>
1269
1270         * src/frv/ffitarget.h: Remove PPC stuff which does not belong to frv.
1271
1272 2005-01-12  Eric Botcazou  <ebotcazou@libertysurf.fr>
1273
1274         * testsuite/libffi.special/special.exp (cxx_options): Add
1275         -shared-libgcc.
1276
1277 2004-12-31  Richard Henderson  <rth@redhat.com>
1278
1279         * src/types.c (FFI_AGGREGATE_TYPEDEF): Remove.
1280         (FFI_TYPEDEF): Rename from FFI_INTEGRAL_TYPEDEF.  Replace size and
1281         offset parameters with a type parameter; deduce size and structure
1282         alignment.  Update all users.
1283
1284 2004-12-31  Richard Henderson  <rth@redhat.com>
1285
1286         * src/types.c (FFI_TYPE_POINTER): Define with sizeof.
1287         (FFI_TYPE_LONGDOUBLE): Fix for ia64.
1288         * src/ia64/ffitarget.h (struct ffi_ia64_trampoline_struct): Move
1289         into ffi_prep_closure.
1290         * src/ia64/ia64_flags.h, src/ia64/ffi.c, src/ia64/unix.S: Rewrite
1291         from scratch.
1292
1293 2004-12-27  Richard Henderson  <rth@redhat.com>
1294
1295         * src/x86/unix64.S: Fix typo in unwind info.
1296
1297 2004-12-25  Richard Henderson  <rth@redhat.com>
1298
1299         * src/x86/ffi64.c (struct register_args): Rename from stackLayout.
1300         (enum x86_64_reg_class): Add X86_64_COMPLEX_X87_CLASS.
1301         (merge_classes): Check for it.
1302         (SSE_CLASS_P): New.
1303         (classify_argument): Pass byte_offset by value; perform all updates
1304         inside struct case.
1305         (examine_argument): Add classes argument; handle
1306         X86_64_COMPLEX_X87_CLASS.
1307         (ffi_prep_args): Merge into ...
1308         (ffi_call): ... here.  Share stack frame with ffi_call_unix64.
1309         (ffi_prep_cif_machdep): Setup cif->flags for proper structure return.
1310         (ffi_fill_return_value): Remove.
1311         (ffi_prep_closure): Remove dead assert.
1312         (ffi_closure_unix64_inner): Rename from ffi_closure_UNIX64_inner.
1313         Rewrite to use struct register_args instead of va_list.  Create
1314         flags for handling structure returns.
1315         * src/x86/unix64.S: Remove dead strings.
1316         (ffi_call_unix64): Rename from ffi_call_UNIX64.  Rewrite to share
1317         stack frame with ffi_call.  Handle structure returns properly.
1318         (float2sse, floatfloat2sse, double2sse): Remove.
1319         (sse2float, sse2double, sse2floatfloat): Remove.
1320         (ffi_closure_unix64): Rename from ffi_closure_UNIX64.  Rewrite
1321         to handle structure returns properly.
1322
1323 2004-12-08  David Edelsohn  <edelsohn@gnu.org>
1324
1325         * Makefile.am (AM_MAKEFLAGS): Remove duplicate LIBCFLAGS and
1326         PICFLAG.
1327         * Makefile.in: Regenerated.
1328
1329 2004-12-02  Richard Sandiford  <rsandifo@redhat.com>
1330
1331         * configure.ac: Use TL_AC_GCC_VERSION to set gcc_version.
1332         * configure, aclocal.m4, Makefile.in: Regenerate.
1333         * include/Makefile.in, testsuite/Makefile.in: Regenerate.
1334
1335 2004-11-29  Kelley Cook  <kcook@gcc.gnu.org>
1336
1337         * configure: Regenerate for libtool change.
1338
1339 2004-11-25  Kelley Cook  <kcook@gcc.gnu.org>
1340
1341         * configure: Regenerate for libtool reversion.
1342
1343 2004-11-24  Kelley Cook  <kcook@gcc.gnu.org>
1344
1345         * configure: Regenerate for libtool change.
1346
1347 2004-11-23  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1348
1349         * testsuite/lib/libffi-dg.exp: Use new procs in target-libpath.exp.
1350
1351 2004-11-23  Richard Sandiford  <rsandifo@redhat.com>
1352
1353         * src/mips/o32.S (ffi_call_O32, ffi_closure_O32): Use jalr instead
1354         of jal.  Use an absolute encoding for the frame information.
1355
1356 2004-11-23  Kelley Cook  <kcook@gcc.gnu.org>
1357
1358         * Makefile.am: Remove no-dependencies.  Add ACLOCAL_AMFLAGS.
1359         * acinclude.m4: Delete logic for sincludes.
1360         * aclocal.m4, Makefile.in, configure: Regenerate.
1361         * include/Makefile: Likewise.
1362         * testsuite/Makefile: Likewise.
1363
1364 2004-11-22  Eric Botcazou  <ebotcazou@libertysurf.fr>
1365
1366         * src/sparc/ffi.c (ffi_prep_closure): Align doubles and 64-bit integers
1367         on a 8-byte boundary.
1368         * src/sparc/v8.S (ffi_closure_v8): Reserve frame space for arguments.
1369
1370 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
1371
1372         * src/arm/ffi.c (ffi_prep_cif_machdep): Handle functions that return
1373         long long values.  Round stack allocation to a multiple of 8 bytes
1374         for ATPCS compatibility.
1375         * src/arm/sysv.S (ffi_call_SYSV): Rework to avoid use of APCS register
1376         names.  Handle returning long long types.  Add Thumb and interworking
1377         support.  Improve soft-float code.
1378
1379 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
1380
1381         * testsuite/lib/libffi-db.exp (load_gcc_lib): New function.
1382         (libffi_exit): New function.
1383         (libffi_init): Build the testglue wrapper if needed.
1384
1385 2004-10-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
1386
1387         PR other/18138
1388         * testsuite/lib/libffi-dg.exp: Accept more than one multilib libgcc.
1389
1390 2004-10-25  Kazuhiro Inaoka <inaoka.kazuhiro@renesas.com>
1391
1392         * src/m32r/libffitarget.h (FFI_CLOSURES): Set to 0.
1393
1394 2004-10-20  Kaz Kojima  <kkojima@gcc.gnu.org>
1395
1396         * src/sh/sysv.S (ffi_call_SYSV): Don't align for double data.
1397         * testsuite/libffi.call/float3.c: New test case.
1398
1399 2004-10-18  Kaz Kojima  <kkojima@gcc.gnu.org>
1400
1401         * src/sh/ffi.c (ffi_prep_closure): Set T bit in trampoline for
1402         the function returning a structure pointed with R2.
1403         * src/sh/sysv.S (ffi_closure_SYSV): Use R2 as the pointer to
1404         the structure return value if T bit set.  Emit position
1405         independent code and EH data if PIC.
1406
1407 2004-10-13  Kazuhiro Inaoka  <inaoka.kazuhiro@renesas.com>
1408
1409         * Makefile.am: Add m32r support.
1410         * configure.ac: Likewise.
1411         * Makefile.in: Regenerate.
1412         * confiugre: Regenerate.
1413         * src/types.c: Add m32r port to FFI_INTERNAL_TYPEDEF
1414         (uint64, sint64, double, longdouble)
1415         * src/m32r: New directory.
1416         * src/m32r/ffi.c: New file.
1417         * src/m32r/sysv.S: Likewise.
1418         * src/m32r/ffitarget.h: Likewise.
1419
1420 2004-10-02  Kaz Kojima  <kkojima@gcc.gnu.org>
1421
1422         * testsuite/libffi.call/negint.c: New test case.
1423
1424 2004-09-14  H.J. Lu  <hongjiu.lu@intel.com>
1425
1426         PR libgcj/17465
1427         * testsuite/lib/libffi-dg.exp: Don't use global ld_library_path.
1428         Set up LD_LIBRARY_PATH, SHLIB_PATH, LD_LIBRARYN32_PATH,
1429         LD_LIBRARY64_PATH, LD_LIBRARY_PATH_32, LD_LIBRARY_PATH_64 and
1430         DYLD_LIBRARY_PATH.
1431
1432 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
1433
1434         * testsuite/libffi.call/many_win32.c: Remove whitespaces.
1435         * testsuite/libffi.call/promotion.c: Likewise.
1436         * testsuite/libffi.call/return_ll.c: Remove unused var. Cleanup
1437         whitespaces.
1438         * testsuite/libffi.call/return_sc.c: Likewise.
1439         * testsuite/libffi.call/return_uc.c: Likewise.
1440
1441 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
1442
1443         * src/powerpc/darwin.S: Fix comments and identation.
1444         * src/powerpc/darwin_closure.S: Likewise.
1445
1446 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
1447
1448         * src/powerpc/ffi_darwin.c: Add flag for longdouble return values.
1449         (ffi_prep_args): Handle longdouble arguments.
1450         (ffi_prep_cif_machdep): Set flags for longdouble. Calculate space for
1451         longdouble.
1452         (ffi_closure_helper_DARWIN): Add closure handling for longdouble.
1453         * src/powerpc/darwin.S (_ffi_call_DARWIN): Add handling of longdouble
1454         values.
1455         * src/powerpc/darwin_closure.S (_ffi_closure_ASM): Likewise.
1456         * src/types.c: Defined longdouble size and alignment for darwin.
1457
1458 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
1459
1460         * src/powerpc/aix.S: Remove whitespaces.
1461         * src/powerpc/aix_closure.S: Likewise.
1462         * src/powerpc/asm.h: Likewise.
1463         * src/powerpc/ffi.c: Likewise.
1464         * src/powerpc/ffitarget.h: Likewise.
1465         * src/powerpc/linux64.S: Likewise.
1466         * src/powerpc/linux64_closure.S: Likewise.
1467         * src/powerpc/ppc_closure.S: Likewise.
1468         * src/powerpc/sysv.S: Likewise.
1469
1470 2004-08-30  Anthony Green  <green@redhat.com>
1471
1472         * Makefile.am: Add frv support.
1473         * Makefile.in, testsuite/Makefile.in: Rebuilt.
1474         * configure.ac: Read configure.host.
1475         * configure.in: Read configure.host.
1476         * configure.host: New file.  frv-elf needs libgloss.
1477         * include/ffi.h.in: Force ffi_closure to have a nice big (8)
1478         alignment.  This is needed to frv and shouldn't harm the others.
1479         * include/ffi_common.h (ALIGN_DOWN): New macro.
1480         * src/frv/ffi.c, src/frv/ffitarget.h, src/frv/eabi.S: New files.
1481
1482 2004-08-24  David Daney  <daney@avtrex.com>
1483
1484         * testsuite/libffi.call/closure_fn0.c: Xfail mips64* instead of mips*.
1485         * testsuite/libffi.call/closure_fn1.c: Likewise.
1486         * testsuite/libffi.call/closure_fn2.c  Likewise.
1487         * testsuite/libffi.call/closure_fn3.c: Likewise.
1488         * testsuite/libffi.call/closure_fn4.c: Likewise.
1489         * testsuite/libffi.call/closure_fn5.c: Likewise.
1490         * testsuite/libffi.call/cls_18byte.c: Likewise.
1491         * testsuite/libffi.call/cls_19byte.c: Likewise.
1492         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
1493         * testsuite/libffi.call/cls_20byte.c: Likewise.
1494         * testsuite/libffi.call/cls_20byte1.c: Likewise.
1495         * testsuite/libffi.call/cls_24byte.c: Likewise.
1496         * testsuite/libffi.call/cls_2byte.c: Likewise.
1497         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
1498         * testsuite/libffi.call/cls_3byte1.c: Likewise.
1499         * testsuite/libffi.call/cls_3byte2.c: Likewise.
1500         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
1501         * testsuite/libffi.call/cls_4byte.c: Likewise.
1502         * testsuite/libffi.call/cls_64byte.c: Likewise.
1503         * testsuite/libffi.call/cls_6byte.c: Likewise.
1504         * testsuite/libffi.call/cls_7byte.c: Likewise.
1505         * testsuite/libffi.call/cls_8byte.c: Likewise.
1506         * testsuite/libffi.call/cls_9byte1.c: Likewise.
1507         * testsuite/libffi.call/cls_9byte2.c: Likewise.
1508         * testsuite/libffi.call/cls_align_double.c: Likewise.
1509         * testsuite/libffi.call/cls_align_float.c: Likewise.
1510         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
1511         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
1512         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
1513         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
1514         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
1515         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
1516         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
1517         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
1518         * testsuite/libffi.call/cls_double.c: Likewise.
1519         * testsuite/libffi.call/cls_float.c: Likewise.
1520         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
1521         * testsuite/libffi.call/cls_multi_sshort.c: Likewise.
1522         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
1523         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
1524         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
1525         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
1526         * testsuite/libffi.call/cls_schar.c: Likewise.
1527         * testsuite/libffi.call/cls_sint.c: Likewise.
1528         * testsuite/libffi.call/cls_sshort.c: Likewise.
1529         * testsuite/libffi.call/cls_uchar.c: Likewise.
1530         * testsuite/libffi.call/cls_uint.c: Likewise.
1531         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
1532         * testsuite/libffi.call/cls_ushort.c: Likewise.
1533         * testsuite/libffi.call/nested_struct.c: Likewise.
1534         * testsuite/libffi.call/nested_struct1.c: Likewise.
1535         * testsuite/libffi.call/nested_struct2.c: Likewise.
1536         * testsuite/libffi.call/nested_struct3.c: Likewise.
1537         * testsuite/libffi.call/problem1.c: Likewise.
1538         * testsuite/libffi.special/unwindtest.cc: Likewise.
1539         * testsuite/libffi.call/cls_12byte.c: Likewise and set return value
1540         to zero.
1541         * testsuite/libffi.call/cls_16byte.c: Likewise.
1542         * testsuite/libffi.call/cls_5byte.c: Likewise.
1543
1544 2004-08-23  David Daney <daney@avtrex.com>
1545
1546         PR libgcj/13141
1547         * src/mips/ffitarget.h (FFI_O32_SOFT_FLOAT): New ABI.
1548         * src/mips/ffi.c (ffi_prep_args): Fix alignment calculation.
1549         (ffi_prep_cif_machdep): Handle FFI_O32_SOFT_FLOAT floating point
1550         parameters and return types.
1551         (ffi_call): Handle FFI_O32_SOFT_FLOAT ABI.
1552         (ffi_prep_closure): Ditto.
1553         (ffi_closure_mips_inner_O32): Handle FFI_O32_SOFT_FLOAT ABI, fix
1554         alignment calculations.
1555         * src/mips/o32.S (ffi_closure_O32): Don't use floating point
1556         instructions if FFI_O32_SOFT_FLOAT, make stack frame ABI compliant.
1557
1558 2004-08-14  Casey Marshall <csm@gnu.org>
1559
1560         * src/mips/ffi.c (ffi_pref_cif_machdep): set `cif->flags' to
1561         contain `FFI_TYPE_UINT64' as return type for any 64-bit
1562         integer (O32 ABI only).
1563         (ffi_prep_closure): new function.
1564         (ffi_closure_mips_inner_O32): new function.
1565         * src/mips/ffitarget.h: Define `FFI_CLOSURES' and
1566         `FFI_TRAMPOLINE_SIZE' appropriately if the ABI is o32.
1567         * src/mips/o32.S (ffi_call_O32): add labels for .eh_frame. Return
1568         64 bit integers correctly.
1569         (ffi_closure_O32): new function.
1570         Added DWARF-2 unwind info for both functions.
1571
1572 2004-08-10  Andrew Haley  <aph@redhat.com>
1573
1574         * src/x86/ffi64.c (ffi_prep_args ): 8-align all stack arguments.
1575
1576 2004-08-01  Robert Millan  <robertmh@gnu.org>
1577
1578         * configure.ac: Detect knetbsd-gnu and kfreebsd-gnu.
1579         * configure: Regenerate.
1580
1581 2004-07-30  Maciej W. Rozycki  <macro@linux-mips.org>
1582
1583         * acinclude.m4 (AC_FUNC_MMAP_BLACKLIST): Check for <sys/mman.h>
1584         and mmap() explicitly instead of relying on preset autoconf cache
1585         variables.
1586         * aclocal.m4: Regenerate.
1587         * configure: Regenerate.
1588
1589 2004-07-11  Ulrich Weigand  <uweigand@de.ibm.com>
1590
1591         * src/s390/ffi.c (ffi_prep_args): Fix C aliasing violation.
1592         (ffi_check_float_struct): Remove unused prototype.
1593
1594 2004-06-30  Geoffrey Keating  <geoffk@apple.com>
1595
1596         * src/powerpc/ffi_darwin.c (flush_icache): ';' is a comment
1597         character on Darwin, use '\n\t' instead.
1598
1599 2004-06-26  Matthias Klose  <doko@debian.org>
1600
1601         * libtool-version: Fix typo in revision/age.
1602
1603 2004-06-17  Matthias Klose  <doko@debian.org>
1604
1605         * libtool-version: New.
1606         * Makefile.am (libffi_la_LDFLAGS): Use -version-info for soname.
1607         * Makefile.in: Regenerate.
1608
1609 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
1610
1611         * Makefile.am: Remove useless multilib rules.
1612         * Makefile.in: Regenerate.
1613         * aclocal.m4: Regenerate with automake 1.8.5.
1614         * configure.ac: Remove useless multilib configury.
1615         * configure: Regenerate.
1616
1617 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
1618
1619         * .cvsignore: New file.
1620
1621 2004-06-10  Jakub Jelinek  <jakub@redhat.com>
1622
1623         * src/ia64/unix.S (ffi_call_unix): Insert group barrier break
1624         fp_done.
1625         (ffi_closure_UNIX): Fix f14/f15 adjustment if FLOAT_SZ is ever
1626         changed from 8.
1627
1628 2004-06-06  Sean McNeil  <sean@mcneil.com>
1629
1630         * configure.ac: Add x86_64-*-freebsd* support.
1631         * configure: Regenerate.
1632
1633 2004-04-26  Joe Buck <jbuck@welsh-buck.org>
1634
1635         Bug 15093
1636         * configure.ac: Test for existence of mmap and sys/mman.h before
1637         checking blacklist.  Fix suggested by Jim Wilson.
1638         * configure: Regenerate.
1639
1640 2004-04-26  Matt Austern  <austern@apple.com>
1641
1642         * src/powerpc/darwin.S: Go through a non-lazy pointer for initial
1643         FDE location.
1644         * src/powerpc/darwin_closure.S: Likewise.
1645
1646 2004-04-24  Andreas Tobler  <a.tobler@schweiz.ch>
1647
1648         * testsuite/libffi.call/cls_multi_schar.c (main): Fix initialization
1649         error. Reported by Thomas Heller <theller@python.net>.
1650         * testsuite/libffi.call/cls_multi_sshort.c (main): Likewise.
1651         * testsuite/libffi.call/cls_multi_ushort.c (main): Likewise.
1652
1653 2004-03-20  Matthias Klose  <doko@debian.org>
1654
1655         * src/pa/linux.S: Fix typo.
1656
1657 2004-03-19  Matthias Klose  <doko@debian.org>
1658
1659         * Makefile.am: Update.
1660         * Makefile.in: Regenerate.
1661         * src/pa/ffi.h.in: Remove.
1662         * src/pa/ffitarget.h: New file.
1663
1664 2004-02-10  Randolph Chung  <tausq@debian.org>
1665
1666         * Makefile.am: Add PA support.
1667         * Makefile.in: Regenerate.
1668         * include/Makefile.in: Regenerate.
1669         * configure.ac: Add PA target.
1670         * configure: Regenerate.
1671         * src/pa/ffi.c: New file.
1672         * src/pa/ffi.h.in: Add PA support.
1673         * src/pa/linux.S: New file.
1674         * prep_cif.c: Add PA support.
1675
1676 2004-03-16  Hosaka Yuji  <hos@tamanegi.org>
1677
1678         * src/types.c: Fix alignment size of X86_WIN32 case int64 and
1679         double.
1680         * src/x86/ffi.c (ffi_prep_args): Replace ecif->cif->rtype->type
1681         with ecif->cif->flags.
1682         (ffi_call, ffi_prep_incoming_args_SYSV): Replace cif->rtype->type
1683         with cif->flags.
1684         (ffi_prep_cif_machdep): Add X86_WIN32 struct case.
1685         (ffi_closure_SYSV): Add 1 or 2-bytes struct case for X86_WIN32.
1686         * src/x86/win32.S (retstruct1b, retstruct2b, sc_retstruct1b,
1687         sc_retstruct2b): Add for 1 or 2-bytes struct case.
1688
1689 2004-03-15 Kelley Cook <kcook@gcc.gnu.org>
1690
1691         * configure.in: Rename file to ...
1692         * configure.ac: ... this.
1693         * fficonfig.h.in: Regenerate.
1694         * Makefile.in: Regenerate.
1695         * include/Makefile.in: Regenerate.
1696         * testsuite/Makefile.in: Regenerate.
1697
1698 2004-03-12  Matt Austern  <austern@apple.com>
1699
1700         * src/powerpc/darwin.S: Fix EH information so it corresponds to
1701         changes in EH format resulting from addition of linkonce support.
1702         * src/powerpc/darwin_closure.S: Likewise.
1703
1704 2004-03-11  Andreas Tobler  <a.tobler@schweiz.ch>
1705             Paolo Bonzini  <bonzini@gnu.org>
1706
1707         * Makefile.am (AUTOMAKE_OPTIONS): Set them.
1708         Remove VPATH. Remove rules for object files. Remove multilib support.
1709         (AM_CCASFLAGS): Add.
1710         * configure.in (AC_CONFIG_HEADERS): Relace AM_CONFIG_HEADER.
1711         (AC_PREREQ): Bump version to 2.59.
1712         (AC_INIT): Fill with version info and bug address.
1713         (ORIGINAL_LD_FOR_MULTILIBS): Remove.
1714         (AM_ENABLE_MULTILIB): Use this instead of AC_ARG_ENABLE.
1715         De-precious CC so that the right flags are passed down to multilibs.
1716         (AC_MSG_ERROR): Replace obsolete macro AC_ERROR.
1717         (AC_CONFIG_FILES): Replace obsolete macro AC_LINK_FILES.
1718         (AC_OUTPUT): Reorganize the output with AC_CONFIG_COMMANDS.
1719         * configure: Rebuilt.
1720         * aclocal.m4: Likewise.
1721         * Makefile.in, include/Makefile.in, testsuite/Makefile.in: Likewise.
1722         * fficonfig.h.in: Likewise.
1723
1724 2004-03-11  Andreas Schwab  <schwab@suse.de>
1725
1726         * src/ia64/ffi.c (ffi_prep_incoming_args_UNIX): Get floating point
1727         arguments from fp registers only for the first 8 parameter slots.
1728         Don't convert a float parameter when passed in memory.
1729
1730 2004-03-09  Hans-Peter Nilsson  <hp@axis.com>
1731
1732         * configure: Regenerate for config/accross.m4 correction.
1733
1734 2004-02-25  Matt Kraai  <kraai@alumni.cmu.edu>
1735
1736         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Change
1737         ecif->cif->bytes to bytes.
1738         (ffi_prep_cif_machdep): Add braces around nested if statement.
1739
1740 2004-02-09  Alan Modra  <amodra@bigpond.net.au>
1741
1742         * src/types.c (pointer): POWERPC64 has 8 byte pointers.
1743
1744         * src/powerpc/ffi.c (ffi_prep_args64): Correct long double handling.
1745         (ffi_closure_helper_LINUX64): Fix typo.
1746         * testsuite/libffi.call/cls_align_longdouble.c: Pass -mlong-double-128
1747         for powerpc64-*-*.
1748         * testsuite/libffi.call/float.c: Likewise.
1749         * testsuite/libffi.call/float2.c: Likewise.
1750
1751 2004-02-08  Alan Modra  <amodra@bigpond.net.au>
1752
1753         * src/powerpc/ffi.c (ffi_prep_cif_machdep <FFI_LINUX64>): Correct
1754         long double function return and long double arg handling.
1755         (ffi_closure_helper_LINUX64): Formatting.  Delete unused "ng" var.
1756         Use "end_pfr" instead of "nf".  Correct long double handling.
1757         Localise "temp".
1758         * src/powerpc/linux64.S (ffi_call_LINUX64): Save f2 long double
1759         return value.
1760         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Allocate
1761         space for long double return value.  Adjust stack frame and offsets.
1762         Load f2 long double return.
1763
1764 2004-02-07  Alan Modra  <amodra@bigpond.net.au>
1765
1766         * src/types.c: Use 16 byte long double for POWERPC64.
1767
1768 2004-01-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
1769
1770         * src/sparc/ffi.c (ffi_prep_args_v9): Shift the parameter array
1771         when the structure return address is passed in %o0.
1772         (ffi_V9_return_struct): Rename into ffi_v9_layout_struct.
1773         (ffi_v9_layout_struct): Align the field following a nested structure
1774         on a word boundary.  Use memmove instead of memcpy.
1775         (ffi_call): Update call to ffi_V9_return_struct.
1776         (ffi_prep_closure): Define 'ctx' only for V8.
1777         (ffi_closure_sparc_inner): Clone into ffi_closure_sparc_inner_v8
1778         and ffi_closure_sparc_inner_v9.
1779         (ffi_closure_sparc_inner_v8): Return long doubles by reference.
1780         Always skip the structure return address.  For structures and long
1781         doubles, copy the argument directly.
1782         (ffi_closure_sparc_inner_v9): Skip the structure return address only
1783         if required.  Shift the maximum floating-point slot accordingly.  For
1784         big structures, copy the argument directly; otherwise, left-justify the
1785         argument and call ffi_v9_layout_struct to lay out the structure on
1786         the stack.
1787         * src/sparc/v8.S: Undef STACKFRAME before defining it.
1788         (ffi_closure_v8): Pass the structure return address.  Update call to
1789         ffi_closure_sparc_inner_v8.  Short-circuit FFI_TYPE_INT handling.
1790         Skip the 'unimp' insn when returning long doubles and structures.
1791         * src/sparc/v9.S: Undef STACKFRAME before defining it.
1792         (ffi_closure_v9): Increase the frame size by 2 words.  Short-circuit
1793         FFI_TYPE_INT handling.  Load structures both in integers and
1794         floating-point registers on return.
1795         * README: Update status of the SPARC port.
1796
1797 2004-01-24  Andreas Tobler  <a.tobler@schweiz.ch>
1798
1799         * testsuite/libffi.call/pyobjc-tc.c (main): Treat result value
1800         as of type ffi_arg.
1801         * testsuite/libffi.call/struct3.c (main): Fix CHECK.
1802
1803 2004-01-22  Ulrich Weigand  <uweigand@de.ibm.com>
1804
1805         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Treat result
1806         value as of type ffi_arg, not unsigned int.
1807
1808 2004-01-21  Michael Ritzert  <ritzert@t-online.de>
1809
1810         * ffi64.c (ffi_prep_args): Cast the RHS of an assignment instead
1811         of the LHS.
1812
1813 2004-01-12  Andreas Tobler  <a.tobler@schweiz.ch>
1814
1815         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_32 for
1816         Solaris.
1817
1818 2004-01-08  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1819
1820         * testsuite/libffi.call/ffitest.h (allocate_mmap): Cast MAP_FAILED
1821         to void *.
1822
1823 2003-12-10  Richard Henderson  <rth@redhat.com>
1824
1825         * testsuite/libffi.call/cls_align_pointer.c: Cast pointers to
1826         size_t instead of int.
1827
1828 2003-12-04  Hosaka Yuji  <hos@tamanegi.org>
1829
1830         * testsuite/libffi.call/many_win32.c: Include <float.h>.
1831         * testsuite/libffi.call/many_win32.c (main): Replace variable
1832         int i with unsigned long ul.
1833
1834         * testsuite/libffi.call/cls_align_uint64.c: New test case.
1835         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
1836         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
1837         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
1838         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
1839         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
1840         * testsuite/libffi.call/cls_align_float.c: Likewise.
1841         * testsuite/libffi.call/cls_align_double.c: Likewise.
1842         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
1843         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
1844
1845 2003-12-02  Hosaka Yuji  <hos@tamanegi.org>
1846
1847         PR other/13221
1848         * src/x86/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV):
1849         Align arguments to 32 bits.
1850
1851 2003-12-01  Andreas Tobler  <a.tobler@schweiz.ch>
1852
1853         PR other/13221
1854         * testsuite/libffi.call/cls_multi_sshort.c: New test case.
1855         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
1856         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
1857         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
1858         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
1859         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
1860
1861         * testsuite/libffi.special/unwindtest.cc: Cosmetics.
1862
1863 2003-11-26  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
1864
1865         * testsuite/libffi.call/ffitest.h: Include <fcntl.h>.
1866         * testsuite/libffi.special/ffitestcxx.h: Likewise.
1867
1868 2003-11-22  Andreas Tobler  <a.tobler@schweiz.ch>
1869
1870         * Makefile.in: Rebuilt.
1871         * configure: Likewise.
1872         * testsuite/libffi.special/unwindtest.cc: Convert the mmap to
1873         the right type.
1874
1875 2003-11-21  Andreas Jaeger  <aj@suse.de>
1876             Andreas Tobler  <a.tobler@schweiz.ch>
1877
1878         * acinclude.m4: Add AC_FUNC_MMAP_BLACKLIST.
1879         * configure.in: Call AC_FUNC_MMAP_BLACKLIST.
1880         * Makefile.in: Rebuilt.
1881         * aclocal.m4: Likewise.
1882         * configure: Likewise.
1883         * fficonfig.h.in: Likewise.
1884         * testsuite/lib/libffi-dg.exp: Add include dir.
1885         * testsuite/libffi.call/ffitest.h: Add MMAP definitions.
1886         * testsuite/libffi.special/ffitestcxx.h: Likewise.
1887         * testsuite/libffi.call/closure_fn0.c: Use MMAP functionality
1888         for ffi_closure if available.
1889         * testsuite/libffi.call/closure_fn1.c: Likewise.
1890         * testsuite/libffi.call/closure_fn2.c: Likewise.
1891         * testsuite/libffi.call/closure_fn3.c: Likewise.
1892         * testsuite/libffi.call/closure_fn4.c: Likewise.
1893         * testsuite/libffi.call/closure_fn5.c: Likewise.
1894         * testsuite/libffi.call/cls_12byte.c: Likewise.
1895         * testsuite/libffi.call/cls_16byte.c: Likewise.
1896         * testsuite/libffi.call/cls_18byte.c: Likewise.
1897         * testsuite/libffi.call/cls_19byte.c: Likewise.
1898         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
1899         * testsuite/libffi.call/cls_20byte.c: Likewise.
1900         * testsuite/libffi.call/cls_20byte1.c: Likewise.
1901         * testsuite/libffi.call/cls_24byte.c: Likewise.
1902         * testsuite/libffi.call/cls_2byte.c: Likewise.
1903         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
1904         * testsuite/libffi.call/cls_3byte1.c: Likewise.
1905         * testsuite/libffi.call/cls_3byte2.c: Likewise.
1906         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
1907         * testsuite/libffi.call/cls_4byte.c: Likewise.
1908         * testsuite/libffi.call/cls_5byte.c: Likewise.
1909         * testsuite/libffi.call/cls_64byte.c: Likewise.
1910         * testsuite/libffi.call/cls_6byte.c: Likewise.
1911         * testsuite/libffi.call/cls_7byte.c: Likewise.
1912         * testsuite/libffi.call/cls_8byte.c: Likewise.
1913         * testsuite/libffi.call/cls_9byte1.c: Likewise.
1914         * testsuite/libffi.call/cls_9byte2.c: Likewise.
1915         * testsuite/libffi.call/cls_double.c: Likewise.
1916         * testsuite/libffi.call/cls_float.c: Likewise.
1917         * testsuite/libffi.call/cls_schar.c: Likewise.
1918         * testsuite/libffi.call/cls_sint.c: Likewise.
1919         * testsuite/libffi.call/cls_sshort.c: Likewise.
1920         * testsuite/libffi.call/cls_uchar.c: Likewise.
1921         * testsuite/libffi.call/cls_uint.c: Likewise.
1922         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
1923         * testsuite/libffi.call/cls_ushort.c: Likewise.
1924         * testsuite/libffi.call/nested_struct.c: Likewise.
1925         * testsuite/libffi.call/nested_struct1.c: Likewise.
1926         * testsuite/libffi.call/nested_struct2.c: Likewise.
1927         * testsuite/libffi.call/nested_struct3.c: Likewise.
1928         * testsuite/libffi.call/problem1.c: Likewise.
1929         * testsuite/libffi.special/unwindtest.cc: Likewise.
1930
1931 2003-11-20  Andreas Tobler  <a.tobler@schweiz.ch>
1932
1933         * testsuite/lib/libffi-dg.exp: Make the -lgcc_s conditional.
1934
1935 2003-11-19  Andreas Tobler  <a.tobler@schweiz.ch>
1936
1937         * testsuite/lib/libffi-dg.exp: Add DYLD_LIBRARY_PATH for darwin.
1938         Add -lgcc_s to additional flags.
1939
1940 2003-11-12  Andreas Tobler  <a.tobler@schweiz.ch>
1941
1942         * configure.in, include/Makefile.am: PR libgcj/11147, install
1943         the ffitarget.h header file in a gcc versioned and target
1944         dependent place.
1945         * configure: Regenerated.
1946         * Makefile.in, include/Makefile.in: Likewise.
1947         * testsuite/Makefile.in: Likewise.
1948
1949 2003-11-09  Andreas Tobler  <a.tobler@schweiz.ch>
1950
1951         * testsuite/libffi.call/closure_fn0.c: Print result and check
1952         with dg-output to make debugging easier.
1953         * testsuite/libffi.call/closure_fn1.c: Likewise.
1954         * testsuite/libffi.call/closure_fn2.c: Likewise.
1955         * testsuite/libffi.call/closure_fn3.c: Likewise.
1956         * testsuite/libffi.call/closure_fn4.c: Likewise.
1957         * testsuite/libffi.call/closure_fn5.c: Likewise.
1958         * testsuite/libffi.call/cls_12byte.c: Likewise.
1959         * testsuite/libffi.call/cls_16byte.c: Likewise.
1960         * testsuite/libffi.call/cls_18byte.c: Likewise.
1961         * testsuite/libffi.call/cls_19byte.c: Likewise.
1962         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
1963         * testsuite/libffi.call/cls_20byte.c: Likewise.
1964         * testsuite/libffi.call/cls_20byte1.c: Likewise.
1965         * testsuite/libffi.call/cls_24byte.c: Likewise.
1966         * testsuite/libffi.call/cls_2byte.c: Likewise.
1967         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
1968         * testsuite/libffi.call/cls_3byte1.c: Likewise.
1969         * testsuite/libffi.call/cls_3byte2.c: Likewise.
1970         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
1971         * testsuite/libffi.call/cls_4byte.c: Likewise.
1972         * testsuite/libffi.call/cls_5byte.c: Likewise.
1973         * testsuite/libffi.call/cls_64byte.c: Likewise.
1974         * testsuite/libffi.call/cls_6byte.c: Likewise.
1975         * testsuite/libffi.call/cls_7byte.c: Likewise.
1976         * testsuite/libffi.call/cls_8byte.c: Likewise.
1977         * testsuite/libffi.call/cls_9byte1.c: Likewise.
1978         * testsuite/libffi.call/cls_9byte2.c: Likewise.
1979         * testsuite/libffi.call/cls_double.c: Likewise.
1980         * testsuite/libffi.call/cls_float.c: Likewise.
1981         * testsuite/libffi.call/cls_schar.c: Likewise.
1982         * testsuite/libffi.call/cls_sint.c: Likewise.
1983         * testsuite/libffi.call/cls_sshort.c: Likewise.
1984         * testsuite/libffi.call/cls_uchar.c: Likewise.
1985         * testsuite/libffi.call/cls_uint.c: Likewise.
1986         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
1987         * testsuite/libffi.call/cls_ushort.c: Likewise.
1988         * testsuite/libffi.call/problem1.c: Likewise.
1989
1990         * testsuite/libffi.special/unwindtest.cc: Make ffi_closure
1991         static.
1992
1993 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
1994
1995         * testsuite/libffi.call/cls_9byte2.c: New test case.
1996         * testsuite/libffi.call/cls_9byte1.c: Likewise.
1997         * testsuite/libffi.call/cls_64byte.c: Likewise.
1998         * testsuite/libffi.call/cls_20byte1.c: Likewise.
1999         * testsuite/libffi.call/cls_19byte.c: Likewise.
2000         * testsuite/libffi.call/cls_18byte.c: Likewise.
2001         * testsuite/libffi.call/closure_fn4.c: Likewise.
2002         * testsuite/libffi.call/closure_fn5.c: Likewise.
2003         * testsuite/libffi.call/cls_schar.c: Likewise.
2004         * testsuite/libffi.call/cls_sint.c: Likewise.
2005         * testsuite/libffi.call/cls_sshort.c: Likewise.
2006         * testsuite/libffi.call/nested_struct2.c: Likewise.
2007         * testsuite/libffi.call/nested_struct3.c: Likewise.
2008
2009 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
2010
2011         * testsuite/libffi.call/cls_double.c: Do a check on the result.
2012         * testsuite/libffi.call/cls_uchar.c: Likewise.
2013         * testsuite/libffi.call/cls_uint.c: Likewise.
2014         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2015         * testsuite/libffi.call/cls_ushort.c: Likewise.
2016         * testsuite/libffi.call/return_sc.c: Cleanup whitespaces.
2017
2018 2003-11-06  Andreas Tobler  <a.tobler@schweiz.ch>
2019
2020         * src/prep_cif.c (ffi_prep_cif): Move the validity check after
2021         the initialization.
2022
2023 2003-10-23  Andreas Tobler  <a.tobler@schweiz.ch>
2024
2025         * src/java_raw_api.c (ffi_java_ptrarray_to_raw): Replace
2026         FFI_ASSERT(FALSE) with FFI_ASSERT(0).
2027
2028 2003-10-22  David Daney  <ddaney@avtrex.com>
2029
2030         * src/mips/ffitarget.h: Replace undefined UINT32 and friends with
2031         __attribute__((__mode__(__SI__))) and friends.
2032
2033 2003-10-22  Andreas Schwab  <schwab@suse.de>
2034
2035         * src/ia64/ffi.c: Replace FALSE/TRUE with false/true.
2036
2037 2003-10-21  Andreas Tobler  <a.tobler@schweiz.ch>
2038
2039         * configure.in: AC_LINK_FILES(ffitarget.h).
2040         * configure: Regenerate.
2041         * Makefile.in: Likewise.
2042         * include/Makefile.in: Likewise.
2043         * testsuite/Makefile.in: Likewise.
2044         * fficonfig.h.in: Likewise.
2045
2046 2003-10-21  Paolo Bonzini  <bonzini@gnu.org>
2047             Richard Henderson  <rth@redhat.com>
2048
2049         Avoid that ffi.h includes fficonfig.h.
2050
2051         * Makefile.am (EXTRA_DIST): Include ffitarget.h files
2052         (TARGET_SRC_MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
2053         (TARGET_SRC_MIPS_SGI): Removed.
2054         (MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
2055         (MIPS_SGI): Removed.
2056         (CLEANFILES): Removed.
2057         (mostlyclean-am, clean-am, mostlyclean-sub, clean-sub): New
2058         targets.
2059         * acconfig.h: Removed.
2060         * configure.in: Compute sizeofs only for double and long double.
2061         Use them to define and subst HAVE_LONG_DOUBLE.  Include comments
2062         into AC_DEFINE instead of using acconfig.h.  Create
2063         include/ffitarget.h instead of include/fficonfig.h.  Rename
2064         MIPS_GCC to MIPS_IRIX, drop MIPS_SGI since we are in gcc's tree.
2065         AC_DEFINE EH_FRAME_FLAGS.
2066         * include/Makefile.am (DISTCLEANFILES): New automake macro.
2067         (hack_DATA): Add ffitarget.h.
2068         * include/ffi.h.in: Remove all system specific definitions.
2069         Declare raw API even if it is not installed, why bother?
2070         Use limits.h instead of SIZEOF_* to define ffi_type_*.  Do
2071         not define EH_FRAME_FLAGS, it is in fficonfig.h now.  Include
2072         ffitarget.h instead of fficonfig.h.  Remove ALIGN macro.
2073         (UINT_ARG, INT_ARG): Removed, use ffi_arg and ffi_sarg instead.
2074         * include/ffi_common.h (bool): Do not define.
2075         (ffi_assert): Accept failed assertion.
2076         (ffi_type_test): Return void and accept file/line.
2077         (FFI_ASSERT): Pass stringized failed assertion.
2078         (FFI_ASSERT_AT): New macro.
2079         (FFI_ASSERT_VALID_TYPE): New macro.
2080         (UINT8, SINT8, UINT16, SINT16, UINT32, SINT32,
2081         UINT64, SINT64): Define here with gcc's __attribute__ macro
2082         instead of in ffi.h
2083         (FLOAT32, ALIGN): Define here instead of in ffi.h
2084         * include/ffi-mips.h: Removed.  Its content moved to
2085         src/mips/ffitarget.h after separating assembly and C sections.
2086         * src/alpha/ffi.c, src/alpha/ffi.c, src/java_raw_api.c
2087         src/prep_cif.c, src/raw_api.c, src/ia64/ffi.c,
2088         src/mips/ffi.c, src/mips/n32.S, src/mips/o32.S,
2089         src/mips/ffitarget.h, src/sparc/ffi.c, src/x86/ffi64.c:
2090         SIZEOF_ARG -> FFI_SIZEOF_ARG.
2091         * src/ia64/ffi.c: Include stdbool.h (provided by GCC 2.95+).
2092         * src/debug.c (ffi_assert): Accept stringized failed assertion.
2093         (ffi_type_test): Rewritten.
2094         * src/prep-cif.c (initialize_aggregate, ffi_prep_cif): Call
2095         FFI_ASSERT_VALID_TYPE.
2096         * src/alpha/ffitarget.h, src/arm/ffitarget.h,
2097         src/ia64/ffitarget.h, src/m68k/ffitarget.h,
2098         src/mips/ffitarget.h, src/powerpc/ffitarget.h,
2099         src/s390/ffitarget.h, src/sh/ffitarget.h,
2100         src/sh64/ffitarget.h, src/sparc/ffitarget.h,
2101         src/x86/ffitarget.h: New files.
2102         * src/alpha/osf.S, src/arm/sysv.S, src/ia64/unix.S,
2103         src/m68k/sysv.S, src/mips/n32.S, src/mips/o32.S,
2104         src/powerpc/aix.S, src/powerpc/darwin.S,
2105         src/powerpc/ffi_darwin.c, src/powerpc/linux64.S,
2106         src/powerpc/linux64_closure.S, src/powerpc/ppc_closure.S,
2107         src/powerpc/sysv.S, src/s390/sysv.S, src/sh/sysv.S,
2108         src/sh64/sysv.S, src/sparc/v8.S, src/sparc/v9.S,
2109         src/x86/sysv.S, src/x86/unix64.S, src/x86/win32.S:
2110         include fficonfig.h
2111
2112 2003-10-20  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2113
2114         * src/mips/ffi.c: Use _ABIN32, _ABIO32 instead of external
2115         _MIPS_SIM_NABI32, _MIPS_SIM_ABI32.
2116
2117 2003-10-19  Andreas Tobler  <a.tobler@schweiz.ch>
2118
2119         * src/powerpc/ffi_darwin.c (ffi_prep_args): Declare bytes again.
2120         Used when FFI_DEBUG = 1.
2121
2122 2003-10-14  Alan Modra  <amodra@bigpond.net.au>
2123
2124         * src/types.c (double, longdouble): Default POWERPC64 to 8 byte size
2125         and align.
2126
2127 2003-10-06  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2128
2129         * include/ffi_mips.h: Define FFI_MIPS_N32 for N32/N64 ABIs,
2130         FFI_MIPS_O32 for O32 ABI.
2131
2132 2003-10-01  Andreas Tobler  <a.tobler@schweiz.ch>
2133
2134         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_64 for
2135         SPARC64. Cleanup whitespaces.
2136
2137 2003-09-19  Andreas Tobler  <a.tobler@schweiz.ch>
2138
2139         * testsuite/libffi.call/closure_fn0.c: Xfail mips, arm,
2140         strongarm, xscale. Cleanup whitespaces.
2141         * testsuite/libffi.call/closure_fn1.c: Likewise.
2142         * testsuite/libffi.call/closure_fn2.c: Likewise.
2143         * testsuite/libffi.call/closure_fn3.c: Likewise.
2144         * testsuite/libffi.call/cls_12byte.c: Likewise.
2145         * testsuite/libffi.call/cls_16byte.c: Likewise.
2146         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2147         * testsuite/libffi.call/cls_20byte.c: Likewise.
2148         * testsuite/libffi.call/cls_24byte.c: Likewise.
2149         * testsuite/libffi.call/cls_2byte.c: Likewise.
2150         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2151         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2152         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2153         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2154         * testsuite/libffi.call/cls_4byte.c: Likewise.
2155         * testsuite/libffi.call/cls_5byte.c: Likewise.
2156         * testsuite/libffi.call/cls_6byte.c: Likewise.
2157         * testsuite/libffi.call/cls_7byte.c: Likewise.
2158         * testsuite/libffi.call/cls_8byte.c: Likewise.
2159         * testsuite/libffi.call/cls_double.c: Likewise.
2160         * testsuite/libffi.call/cls_float.c: Likewise.
2161         * testsuite/libffi.call/cls_uchar.c: Likewise.
2162         * testsuite/libffi.call/cls_uint.c: Likewise.
2163         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2164         * testsuite/libffi.call/cls_ushort.c: Likewise.
2165         * testsuite/libffi.call/nested_struct.c: Likewise.
2166         * testsuite/libffi.call/nested_struct1.c: Likewise.
2167         * testsuite/libffi.call/problem1.c: Likewise.
2168         * testsuite/libffi.special/unwindtest.cc: Likewise.
2169         * testsuite/libffi.call/pyobjc-tc.c: Cleanup whitespaces.
2170
2171 2003-09-18  David Edelsohn  <edelsohn@gnu.org>
2172
2173         * src/powerpc/aix.S: Cleanup whitespaces.
2174         * src/powerpc/aix_closure.S: Likewise.
2175
2176 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
2177
2178         * src/powerpc/darwin.S: Cleanup whitespaces, comment formatting.
2179         * src/powerpc/darwin_closure.S: Likewise.
2180         * src/powerpc/ffi_darwin.c: Likewise.
2181
2182 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
2183             David Edelsohn  <edelsohn@gnu.org>
2184
2185         * src/types.c (double): Add AIX and Darwin to the right TYPEDEF.
2186         * src/powerpc/aix_closure.S: Remove the pointer to the outgoing
2187         parameter stack.
2188         * src/powerpc/darwin_closure.S: Likewise.
2189         * src/powerpc/ffi_darwin.c (ffi_prep_args): Handle structures
2190         according to the Darwin/AIX ABI.
2191         (ffi_prep_cif_machdep): Likewise.
2192         (ffi_closure_helper_DARWIN): Likewise.
2193         Remove the outgoing parameter stack logic. Simplify the evaluation
2194         of the different CASE types.
2195         (ffi_prep_clousure): Avoid the casts on lvalues. Change the branch
2196         statement in the trampoline code.
2197
2198 2003-09-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2199
2200         * src/sh/ffi.c (ffi_prep_args): Take account into the alignement
2201         for the register size.
2202         (ffi_closure_helper_SYSV): Handle the structure return value
2203         address correctly.
2204         (ffi_closure_helper_SYSV): Return the appropriate type when
2205         the registers are used for the structure return value.
2206         * src/sh/sysv.S (ffi_closure_SYSV): Fix the stack layout for
2207         the 64-bit return value.  Update copyright years.
2208
2209 2003-09-17  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2210
2211         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Search in
2212         srcdir for ffi_mips.h.
2213
2214 2003-09-12  Alan Modra  <amodra@bigpond.net.au>
2215
2216         * src/prep_cif.c (initialize_aggregate): Include tail padding in
2217         structure size.
2218         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Correct
2219         placement of float result.
2220         * testsuite/libffi.special/unwindtest.cc (closure_test_fn1): Correct
2221         cast of "resp" for big-endian 64 bit machines.
2222
2223 2003-09-11  Alan Modra  <amodra@bigpond.net.au>
2224
2225         * src/types.c (double, longdouble): Merge identical SH and ARM
2226         typedefs, and add POWERPC64.
2227         * src/powerpc/ffi.c (ffi_prep_args64): Correct next_arg calc for
2228         struct split over gpr and rest.
2229         (ffi_prep_cif_machdep): Correct intarg_count for structures.
2230         * src/powerpc/linux64.S (ffi_call_LINUX64): Fix gpr offsets.
2231
2232 2003-09-09  Andreas Tobler  <a.tobler@schweiz.ch>
2233
2234         * src/powerpc/ffi.c (ffi_closure_helper_SYSV) Handle struct
2235         passing correctly.
2236
2237 2003-09-09  Alan Modra  <amodra@bigpond.net.au>
2238
2239         * configure: Regenerate.
2240
2241 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
2242
2243         * Makefile.am: Remove build rules for ffitest.
2244         * Makefile.in: Rebuilt.
2245
2246 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
2247
2248         * src/java_raw_api.c: Include <stdlib.h> to fix compiler warning
2249         about implicit declaration of abort().
2250
2251 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
2252
2253         * Makefile.am: Add dejagnu test framework. Fixes PR other/11411.
2254         * Makefile.in: Rebuilt.
2255         * configure.in: Add dejagnu test framework.
2256         * configure: Rebuilt.
2257
2258         * testsuite/Makefile.am: New file.
2259         * testsuite/Makefile.in: Built
2260         * testsuite/lib/libffi-dg.exp: New file.
2261         * testsuite/config/default.exp: Likewise.
2262         * testsuite/libffi.call/call.exp: Likewise.
2263         * testsuite/libffi.call/ffitest.h: Likewise.
2264         * testsuite/libffi.call/closure_fn0.c: Likewise.
2265         * testsuite/libffi.call/closure_fn1.c: Likewise.
2266         * testsuite/libffi.call/closure_fn2.c: Likewise.
2267         * testsuite/libffi.call/closure_fn3.c: Likewise.
2268         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2269         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2270         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2271         * testsuite/libffi.call/cls_2byte.c: Likewise.
2272         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2273         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2274         * testsuite/libffi.call/cls_4byte.c: Likewise.
2275         * testsuite/libffi.call/cls_5byte.c: Likewise.
2276         * testsuite/libffi.call/cls_6byte.c: Likewise.
2277         * testsuite/libffi.call/cls_7byte.c: Likewise.
2278         * testsuite/libffi.call/cls_8byte.c: Likewise.
2279         * testsuite/libffi.call/cls_12byte.c: Likewise.
2280         * testsuite/libffi.call/cls_16byte.c: Likewise.
2281         * testsuite/libffi.call/cls_20byte.c: Likewise.
2282         * testsuite/libffi.call/cls_24byte.c: Likewise.
2283         * testsuite/libffi.call/cls_double.c: Likewise.
2284         * testsuite/libffi.call/cls_float.c: Likewise.
2285         * testsuite/libffi.call/cls_uchar.c: Likewise.
2286         * testsuite/libffi.call/cls_uint.c: Likewise.
2287         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2288         * testsuite/libffi.call/cls_ushort.c: Likewise.
2289         * testsuite/libffi.call/float.c: Likewise.
2290         * testsuite/libffi.call/float1.c: Likewise.
2291         * testsuite/libffi.call/float2.c: Likewise.
2292         * testsuite/libffi.call/many.c: Likewise.
2293         * testsuite/libffi.call/many_win32.c: Likewise.
2294         * testsuite/libffi.call/nested_struct.c: Likewise.
2295         * testsuite/libffi.call/nested_struct1.c: Likewise.
2296         * testsuite/libffi.call/pyobjc-tc.c: Likewise.
2297         * testsuite/libffi.call/problem1.c: Likewise.
2298         * testsuite/libffi.call/promotion.c: Likewise.
2299         * testsuite/libffi.call/return_ll.c: Likewise.
2300         * testsuite/libffi.call/return_sc.c: Likewise.
2301         * testsuite/libffi.call/return_uc.c: Likewise.
2302         * testsuite/libffi.call/strlen.c: Likewise.
2303         * testsuite/libffi.call/strlen_win32.c: Likewise.
2304         * testsuite/libffi.call/struct1.c: Likewise.
2305         * testsuite/libffi.call/struct2.c: Likewise.
2306         * testsuite/libffi.call/struct3.c: Likewise.
2307         * testsuite/libffi.call/struct4.c: Likewise.
2308         * testsuite/libffi.call/struct5.c: Likewise.
2309         * testsuite/libffi.call/struct6.c: Likewise.
2310         * testsuite/libffi.call/struct7.c: Likewise.
2311         * testsuite/libffi.call/struct8.c: Likewise.
2312         * testsuite/libffi.call/struct9.c: Likewise.
2313         * testsuite/libffi.special/special.exp: New file.
2314         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2315         * testsuite/libffi.special/unwindtest.cc: Likewise.
2316
2317
2318 2003-08-13  Kaz Kojima  <kkojima@gcc.gnu.org>
2319
2320         * src/sh/ffi.c (OFS_INT16): Set 0 for little endian case.  Update
2321         copyright years.
2322
2323 2003-08-02  Alan Modra  <amodra@bigpond.net.au>
2324
2325         * src/powerpc/ffi.c (ffi_prep_args64): Modify for changed gcc
2326         structure passing.
2327         (ffi_closure_helper_LINUX64): Likewise.
2328         * src/powerpc/linux64.S: Remove code writing to parm save area.
2329         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Use return
2330         address in lr from ffi_closure_helper_LINUX64 call to calculate
2331         table address.  Optimize function tail.
2332
2333 2003-07-28  Andreas Tobler  <a.tobler@schweiz.ch>
2334
2335         * src/sparc/ffi.c: Handle all floating point registers.
2336         * src/sparc/v9.S: Likewise. Fixes second part of PR target/11410.
2337
2338 2003-07-11  Gerald Pfeifer  <pfeifer@dbai.tuwien.ac.at>
2339
2340         * README: Note that libffi is not part of GCC.  Update the project
2341         URL and status.
2342
2343 2003-06-19  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
2344
2345         * src/powerpc/ppc_closure.S: Include ffi.h.
2346
2347 2003-06-13  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2348
2349         * src/x86/sysv.S: Avoid gas-only .uleb128/.sleb128 directives.
2350         Use C style comments.
2351
2352 2003-06-13  Kaz Kojima  <kkojima@rr.iij4u.or.jp>
2353
2354         * Makefile.am: Add SHmedia support.  Fix a typo of SH support.
2355         * Makefile.in: Regenerate.
2356         * configure.in (sh64-*-linux*, sh5*-*-linux*): Add target.
2357         * configure: Regenerate.
2358         * include/ffi.h.in: Add SHmedia support.
2359         * src/sh64/ffi.c: New file.
2360         * src/sh64/sysv.S: New file.
2361
2362 2003-05-16  Jakub Jelinek  <jakub@redhat.com>
2363
2364         * configure.in (HAVE_RO_EH_FRAME): Check whether .eh_frame section
2365         should be read-only.
2366         * configure: Rebuilt.
2367         * fficonfig.h.in: Rebuilt.
2368         * include/ffi.h.in (EH_FRAME_FLAGS): Define.
2369         * src/alpha/osf.S: Use EH_FRAME_FLAGS.
2370         * src/powerpc/linux64.S: Likewise.
2371         * src/powerpc/linux64_closure.S: Likewise.  Include ffi.h.
2372         * src/powerpc/sysv.S: Use EH_FRAME_FLAGS.  Use pcrel encoding
2373         if -fpic/-fPIC/-mrelocatable.
2374         * src/powerpc/powerpc_closure.S: Likewise.
2375         * src/sparc/v8.S: If HAVE_RO_EH_FRAME is defined, don't include
2376         #write in .eh_frame flags.
2377         * src/sparc/v9.S: Likewise.
2378         * src/x86/unix64.S: Use EH_FRAME_FLAGS.
2379         * src/x86/sysv.S: Likewise.  Use pcrel encoding if -fpic/-fPIC.
2380         * src/s390/sysv.S: Use EH_FRAME_FLAGS.  Include ffi.h.
2381
2382 2003-05-07  Jeff Sturm  <jsturm@one-point.com>
2383
2384         Fixes PR bootstrap/10656
2385         * configure.in (HAVE_AS_REGISTER_PSEUDO_OP): Test assembler
2386         support for .register pseudo-op.
2387         * src/sparc/v8.S: Use it.
2388         * fficonfig.h.in: Rebuilt.
2389         * configure: Rebuilt.
2390
2391 2003-04-18  Jakub Jelinek  <jakub@redhat.com>
2392
2393         * include/ffi.h.in (POWERPC64): Define if 64-bit.
2394         (enum ffi_abi): Add FFI_LINUX64 on POWERPC.
2395         Make it the default on POWERPC64.
2396         (FFI_TRAMPOLINE_SIZE): Define to 24 on POWERPC64.
2397         * configure.in: Change powerpc-*-linux* into powerpc*-*-linux*.
2398         * configure: Rebuilt.
2399         * src/powerpc/ffi.c (hidden): Define.
2400         (ffi_prep_args_SYSV): Renamed from
2401         ffi_prep_args.  Cast pointers to unsigned long to shut up warnings.
2402         (NUM_GPR_ARG_REGISTERS64, NUM_FPR_ARG_REGISTERS64,
2403         ASM_NEEDS_REGISTERS64): New.
2404         (ffi_prep_args64): New function.
2405         (ffi_prep_cif_machdep): Handle FFI_LINUX64 ABI.
2406         (ffi_call): Likewise.
2407         (ffi_prep_closure): Likewise.
2408         (flush_icache): Surround by #ifndef POWERPC64.
2409         (ffi_dblfl): New union type.
2410         (ffi_closure_helper_SYSV): Use it to avoid aliasing problems.
2411         (ffi_closure_helper_LINUX64): New function.
2412         * src/powerpc/ppc_closure.S: Surround whole file by #ifndef
2413         __powerpc64__.
2414         * src/powerpc/sysv.S: Likewise.
2415         (ffi_call_SYSV): Rename ffi_prep_args to ffi_prep_args_SYSV.
2416         * src/powerpc/linux64.S: New file.
2417         * src/powerpc/linux64_closure.S: New file.
2418         * Makefile.am (EXTRA_DIST): Add src/powerpc/linux64.S and
2419         src/powerpc/linux64_closure.S.
2420         (TARGET_SRC_POWERPC): Likewise.
2421
2422         * src/ffitest.c (closure_test_fn, closure_test_fn1, closure_test_fn2,
2423         closure_test_fn3): Fix result printing on big-endian 64-bit
2424         machines.
2425         (main): Print tst2_arg instead of uninitialized tst2_result.
2426
2427         * src/ffitest.c (main): Hide what closure pointer really points to
2428         from the compiler.
2429
2430 2003-04-16  Richard Earnshaw  <rearnsha@arm.com>
2431
2432         * configure.in (arm-*-netbsdelf*): Add configuration.
2433         (configure): Regenerated.
2434
2435 2003-04-04  Loren J. Rittle  <ljrittle@acm.org>
2436
2437         * include/Makefile.in: Regenerate.
2438
2439 2003-03-21  Zdenek Dvorak  <rakdver@atrey.karlin.mff.cuni.cz>
2440
2441         * libffi/include/ffi.h.in: Define X86 instead of X86_64 in 32
2442         bit mode.
2443         * libffi/src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV):
2444         Receive closure pointer through parameter, read args using
2445         __builtin_dwarf_cfa.
2446         (FFI_INIT_TRAMPOLINE): Send closure reference through eax.
2447
2448 2003-03-12  Andreas Schwab  <schwab@suse.de>
2449
2450         * configure.in: Avoid trailing /. in toolexeclibdir.
2451         * configure: Rebuilt.
2452
2453 2003-03-03  Andreas Tobler <a.tobler@schweiz.ch>
2454
2455         * src/powerpc/darwin_closure.S: Recode to fit dynamic libraries.
2456
2457 2003-02-06  Andreas Tobler <a.tobler@schweiz.ch>
2458
2459         * libffi/src/powerpc/darwin_closure.S:
2460         Fix alignement bug, allocate 8 bytes for the result.
2461         * libffi/src/powerpc/aix_closure.S:
2462         Likewise.
2463         * libffi/src/powerpc/ffi_darwin.c:
2464         Update stackframe description for aix/darwin_closure.S.
2465
2466 2003-02-06  Jakub Jelinek  <jakub@redhat.com>
2467
2468         * src/s390/ffi.c (ffi_closure_helper_SYSV): Add hidden visibility
2469         attribute.
2470
2471 2003-01-31  Christian Cornelssen  <ccorn@cs.tu-berlin.de>,
2472             Andreas Schwab  <schwab@suse.de>
2473
2474         * configure.in: Adjust command to source config-ml.in to account
2475         for changes to the libffi_basedir definition.
2476         (libffi_basedir): Remove ${srcdir} from value and include trailing
2477         slash if nonempty.
2478
2479         * configure: Regenerate.
2480
2481 2003-01-29  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
2482
2483         * src/powerpc/ppc_closure.S: Recode to fit shared libs.
2484
2485 2003-01-28  Andrew Haley  <aph@redhat.com>
2486
2487         * include/ffi.h.in: Enable FFI_CLOSURES for x86_64.
2488         * src/x86/ffi64.c (ffi_prep_closure): New.
2489         (ffi_closure_UNIX64_inner): New.
2490         * src/x86/unix64.S (ffi_closure_UNIX64): New.
2491
2492 2003-01-27  Alexandre Oliva  <aoliva@redhat.com>
2493
2494         * configure.in (toolexecdir, toolexeclibdir): Set and AC_SUBST.
2495         Remove USE_LIBDIR conditional.
2496         * Makefile.am (toolexecdir, toolexeclibdir): Don't override.
2497         * Makefile.in, configure: Rebuilt.
2498
2499 2003-01027  David Edelsohn  <edelsohn@gnu.org>
2500
2501         * Makefile.am (TARGET_SRC_POWERPC_AIX): Fix typo.
2502         * Makefile.in: Regenerate.
2503
2504 2003-01-22  Andrew Haley  <aph@redhat.com>
2505
2506         * src/powerpc/darwin.S (_ffi_call_AIX): Add Augmentation size to
2507         unwind info.
2508
2509 2003-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
2510
2511         * src/powerpc/darwin.S: Add unwind info.
2512         * src/powerpc/darwin_closure.S: Likewise.
2513
2514 2003-01-14  Andrew Haley  <aph@redhat.com>
2515
2516         * src/x86/ffi64.c (ffi_prep_args): Check for void retval.
2517         (ffi_prep_cif_machdep): Likewise.
2518         * src/x86/unix64.S: Add unwind info.
2519
2520 2003-01-14  Andreas Jaeger  <aj@suse.de>
2521
2522         * src/ffitest.c (main): Only use ffi_closures if those are
2523         supported.
2524
2525 2003-01-13 Andreas Tobler <a.tobler@schweiz.ch>
2526
2527         * libffi/src/ffitest.c
2528          add closure testcases
2529
2530 2003-01-13 Kevin B. Hendricks <khendricks@ivey.uwo.ca>
2531
2532         * libffi/src/powerpc/ffi.c
2533          fix alignment bug for float (4 byte aligned iso 8 byte)
2534
2535 2003-01-09  Geoffrey Keating  <geoffk@apple.com>
2536
2537         * src/powerpc/ffi_darwin.c: Remove RCS version string.
2538         * src/powerpc/darwin.S: Remove RCS version string.
2539
2540 2003-01-03  Jeff Sturm  <jsturm@one-point.com>
2541
2542         * include/ffi.h.in: Add closure defines for SPARC, SPARC64.
2543         * src/ffitest.c (main): Use static storage for closure.
2544         * src/sparc/ffi.c (ffi_prep_closure, ffi_closure_sparc_inner): New.
2545         * src/sparc/v8.S (ffi_closure_v8): New.
2546         * src/sparc/v9.S (ffi_closure_v9): New.
2547
2548 2002-11-10  Ranjit Mathew <rmathew@hotmail.com>
2549
2550         * include/ffi.h.in: Added FFI_STDCALL ffi_type
2551           enumeration for X86_WIN32.
2552         * src/x86/win32.S: Added ffi_call_STDCALL function
2553           definition.
2554         * src/x86/ffi.c (ffi_call/ffi_raw_call): Added
2555           switch cases for recognising FFI_STDCALL and
2556           calling ffi_call_STDCALL if target is X86_WIN32.
2557         * src/ffitest.c (my_stdcall_strlen/stdcall_many):
2558           stdcall versions of the "my_strlen" and "many"
2559           test functions (for X86_WIN32).
2560           Added test cases to test stdcall invocation using
2561           these functions.
2562
2563 2002-12-02  Kaz Kojima  <kkojima@gcc.gnu.org>
2564
2565         * src/sh/sysv.S: Add DWARF2 unwind info.
2566
2567 2002-11-27  Ulrich Weigand  <uweigand@de.ibm.com>
2568
2569         * src/s390/sysv.S (.eh_frame section): Make section read-only.
2570
2571 2002-11-26  Jim Wilson  <wilson@redhat.com>
2572
2573         * src/types.c (FFI_TYPE_POINTER): Has size 8 on IA64.
2574
2575 2002-11-23  H.J. Lu <hjl@gnu.org>
2576
2577         * acinclude.m4: Add dummy AM_PROG_LIBTOOL.
2578         Include ../config/accross.m4.
2579         * aclocal.m4; Rebuild.
2580         * configure: Likewise.
2581
2582 2002-11-15  Ulrich Weigand  <uweigand@de.ibm.com>
2583
2584         * src/s390/sysv.S (.eh_frame section): Adapt to pcrel FDE encoding.
2585
2586 2002-11-11  DJ Delorie  <dj@redhat.com>
2587
2588         * configure.in: Look for common files in the right place.
2589
2590 2002-10-08  Ulrich Weigand  <uweigand@de.ibm.com>
2591
2592         * src/java_raw_api.c (ffi_java_raw_to_ptrarray): Interpret
2593         raw data as _Jv_word values, not ffi_raw.
2594         (ffi_java_ptrarray_to_raw): Likewise.
2595         (ffi_java_rvalue_to_raw): New function.
2596         (ffi_java_raw_call): Call it.
2597         (ffi_java_raw_to_rvalue): New function.
2598         (ffi_java_translate_args): Call it.
2599         * src/ffitest.c (closure_test_fn): Interpret return value
2600         as ffi_arg, not int.
2601         * src/s390/ffi.c (ffi_prep_cif_machdep): Add missing
2602         FFI_TYPE_POINTER case.
2603         (ffi_closure_helper_SYSV): Likewise.  Also, assume return
2604         values extended to word size.
2605
2606 2002-10-02  Andreas Jaeger  <aj@suse.de>
2607
2608         * src/x86/ffi64.c (ffi_prep_cif_machdep): Remove debug output.
2609
2610 2002-10-01  Bo Thorsen  <bo@smetana.suse.de>
2611
2612         * include/ffi.h.in: Fix i386 win32 compilation.
2613
2614 2002-09-30  Ulrich Weigand  <uweigand@de.ibm.com>
2615
2616         * configure.in: Add s390x-*-linux-* target.
2617         * configure: Regenerate.
2618         * include/ffi.h.in: Define S390X for s390x targets.
2619         (FFI_CLOSURES): Define for s390/s390x.
2620         (FFI_TRAMPOLINE_SIZE): Likewise.
2621         (FFI_NATIVE_RAW_API): Likewise.
2622         * src/prep_cif.c (ffi_prep_cif): Do not compute stack space for s390.
2623         * src/types.c (FFI_TYPE_POINTER): Use 8-byte pointers on s390x.
2624         * src/s390/ffi.c: Major rework of existing code.  Add support for
2625         s390x targets.  Add closure support.
2626         * src/s390/sysv.S: Likewise.
2627
2628 2002-09-29  Richard Earnshaw  <rearnsha@arm.com>
2629
2630         * src/arm/sysv.S: Fix typo.
2631
2632 2002-09-28  Richard Earnshaw  <rearnsha@arm.com>
2633
2634         * src/arm/sysv.S: If we don't have machine/asm.h and the pre-processor
2635         has defined __USER_LABEL_PREFIX__, then use it in CNAME.
2636         (ffi_call_SYSV): Handle soft-float.
2637
2638 2002-09-27  Bo Thorsen  <bo@suse.de>
2639
2640         * include/ffi.h.in: Fix multilib x86-64 support.
2641
2642 2002-09-22  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
2643
2644         * Makefile.am (all-multi): Fix multilib parallel build.
2645
2646 2002-07-19  Kaz Kojima  <kkojima@gcc.gnu.org>
2647
2648         * configure.in (sh[34]*-*-linux*): Add brackets.
2649         * configure: Regenerate.
2650
2651 2002-07-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2652
2653         * Makefile.am: Add SH support.
2654         * Makefile.in: Regenerate.
2655         * configure.in (sh-*-linux*, sh[34]*-*-linux*): Add target.
2656         * configure: Regenerate.
2657         * include/ffi.h.in: Add SH support.
2658         * src/sh/ffi.c: New file.
2659         * src/sh/sysv.S: New file.
2660         * src/types.c: Add SH support.
2661
2662 2002-07-16  Bo Thorsen  <bo@suse.de>
2663
2664         * src/x86/ffi64.c: New file that adds x86-64 support.
2665         * src/x86/unix64.S: New file that handles argument setup for
2666         x86-64.
2667         * src/x86/sysv.S: Don't use this on x86-64.
2668         * src/x86/ffi.c: Don't use this on x86-64.
2669         Remove unused vars.
2670         * src/prep_cif.c (ffi_prep_cif): Don't do stack size calculation
2671         for x86-64.
2672         * src/ffitest.c (struct6): New test that tests a special case in
2673         the x86-64 ABI.
2674         (struct7): Likewise.
2675         (struct8): Likewise.
2676         (struct9): Likewise.
2677         (closure_test_fn): Silence warning about this when it's not used.
2678         (main): Add the new tests.
2679         (main): Fix a couple of wrong casts and silence some compiler warnings.
2680         * include/ffi.h.in: Add x86-64 ABI definition.
2681         * fficonfig.h.in: Regenerate.
2682         * Makefile.am: Add x86-64 support.
2683         * configure.in: Likewise.
2684         * Makefile.in: Regenerate.
2685         * configure: Likewise.
2686
2687 2002-06-24  Bo Thorsen  <bo@suse.de>
2688
2689         * src/types.c: Merge settings for similar architectures.
2690         Add x86-64 sizes and alignments.
2691
2692 2002-06-23  Bo Thorsen  <bo@suse.de>
2693
2694         * src/arm/ffi.c (ffi_prep_args): Remove unused vars.
2695         * src/sparc/ffi.c (ffi_prep_args_v8): Likewise.
2696         * src/mips/ffi.c (ffi_prep_args): Likewise.
2697         * src/m68k/ffi.c (ffi_prep_args): Likewise.
2698
2699 2002-07-18  H.J. Lu  (hjl@gnu.org)
2700
2701         * Makefile.am (TARGET_SRC_MIPS_LINUX): New.
2702         (libffi_la_SOURCES): Support MIPS_LINUX.
2703         (libffi_convenience_la_SOURCES): Likewise.
2704         * Makefile.in: Regenerated.
2705
2706         * configure.in (mips64*-*): Skip.
2707         (mips*-*-linux*): New.
2708         * configure: Regenerated.
2709
2710         * src/mips/ffi.c: Include <sgidefs.h>.
2711
2712 2002-06-06  Ulrich Weigand  <uweigand@de.ibm.com>
2713
2714         * src/s390/sysv.S: Save/restore %r6.  Add DWARF-2 unwind info.
2715
2716 2002-05-27  Roger Sayle  <roger@eyesopen.com>
2717
2718         * src/x86/ffi.c (ffi_prep_args): Remove reference to avn.
2719
2720 2002-05-27  Bo Thorsen  <bo@suse.de>
2721
2722         * src/x86/ffi.c (ffi_prep_args): Remove unused variable and
2723         fix formatting.
2724
2725 2002-05-13  Andreas Tobler  <a.tobler@schweiz.ch>
2726
2727         * src/powerpc/ffi_darwin.c (ffi_prep_closure): Declare fd at
2728         beginning of function (for older apple cc).
2729
2730 2002-05-08  Alexandre Oliva  <aoliva@redhat.com>
2731
2732         * configure.in (ORIGINAL_LD_FOR_MULTILIBS): Preserve LD at
2733         script entry, and set LD to it when configuring multilibs.
2734         * configure: Rebuilt.
2735
2736 2002-05-05  Jason Thorpe  <thorpej@wasabisystems.com>
2737
2738         * configure.in (sparc64-*-netbsd*): Add target.
2739         (sparc-*-netbsdelf*): Likewise.
2740         * configure: Regenerate.
2741
2742 2002-04-28  David S. Miller  <davem@redhat.com>
2743
2744         * configure.in, configure: Fix SPARC test in previous change.
2745
2746 2002-04-29  Gerhard Tonn  <GerhardTonn@swol.de>
2747
2748         * Makefile.am: Add Linux for S/390 support.
2749         * Makefile.in: Regenerate.
2750         * configure.in: Add Linux for S/390 support.
2751         * configure: Regenerate.
2752         * include/ffi.h.in: Add Linux for S/390 support.
2753         * src/s390/ffi.c: New file from libffi CVS tree.
2754         * src/s390/sysv.S: New file from libffi CVS tree.
2755
2756 2002-04-28  Jakub Jelinek  <jakub@redhat.com>
2757
2758         * configure.in (HAVE_AS_SPARC_UA_PCREL): Check for working
2759         %r_disp32().
2760         * src/sparc/v8.S: Use it.
2761         * src/sparc/v9.S: Likewise.
2762         * fficonfig.h.in: Rebuilt.
2763         * configure: Rebuilt.
2764
2765 2002-04-08  Hans Boehm  <Hans_Boehm@hp.com>
2766
2767         * src/java_raw_api.c (ffi_java_raw_size): Handle FFI_TYPE_DOUBLE
2768         correctly.
2769         * src/ia64/unix.S: Add unwind information. Fix comments.
2770         Save sp in a way that's compatible with unwind info.
2771         (ffi_call_unix): Correctly restore sp in all cases.
2772         * src/ia64/ffi.c: Add, fix comments.
2773
2774 2002-04-08  Jakub Jelinek  <jakub@redhat.com>
2775
2776         * src/sparc/v8.S: Make .eh_frame dependent on target word size.
2777
2778 2002-04-06  Jason Thorpe  <thorpej@wasabisystems.com>
2779
2780         * configure.in (alpha*-*-netbsd*): Add target.
2781         * configure: Regenerate.
2782
2783 2002-04-04  Jeff Sturm  <jsturm@one-point.com>
2784
2785         * src/sparc/v8.S: Add unwind info.
2786         * src/sparc/v9.S: Likewise.
2787
2788 2002-03-30  Krister Walfridsson  <cato@df.lth.se>
2789
2790         * configure.in: Enable i*86-*-netbsdelf*.
2791         * configure: Rebuilt.
2792
2793 2002-03-29  David Billinghurst <David.Billinghurst@riotinto.com>
2794
2795         PR other/2620
2796         * src/mips/n32.s: Delete
2797         * src/mips/o32.s: Delete
2798
2799 2002-03-21  Loren J. Rittle  <ljrittle@acm.org>
2800
2801         * configure.in: Enable alpha*-*-freebsd*.
2802         * configure: Rebuilt.
2803
2804 2002-03-17  Bryce McKinlay  <bryce@waitaki.otago.ac.nz>
2805
2806         * Makefile.am: libfficonvenience -> libffi_convenience.
2807         * Makefile.in: Rebuilt.
2808
2809         * Makefile.am: Define ffitest_OBJECTS.
2810         * Makefile.in: Rebuilt.
2811
2812 2002-03-07  Andreas Tobler  <toa@pop.agri.ch>
2813             David Edelsohn  <edelsohn@gnu.org>
2814
2815         * Makefile.am (EXTRA_DIST): Add Darwin and AIX closure files.
2816         (TARGET_SRC_POWERPC_AIX): Add aix_closure.S.
2817         (TARGET_SRC_POWERPC_DARWIN): Add darwin_closure.S.
2818         * Makefile.in: Regenerate.
2819         * include/ffi.h.in: Add AIX and Darwin closure definitions.
2820         * src/powerpc/ffi_darwin.c (ffi_prep_closure): New function.
2821         (flush_icache, flush_range): New functions.
2822         (ffi_closure_helper_DARWIN): New function.
2823         * src/powerpc/aix_closure.S: New file.
2824         * src/powerpc/darwin_closure.S: New file.
2825
2826 2002-02-24  Jeff Sturm  <jsturm@one-point.com>
2827
2828         * include/ffi.h.in: Add typedef for ffi_arg.
2829         * src/ffitest.c (main): Declare rint with ffi_arg.
2830
2831 2002-02-21  Andreas Tobler  <toa@pop.agri.ch>
2832
2833         * src/powerpc/ffi_darwin.c (ffi_prep_args): Skip appropriate
2834         number of GPRs for floating-point arguments.
2835
2836 2002-01-31  Anthony Green  <green@redhat.com>
2837
2838         * configure: Rebuilt.
2839         * configure.in: Replace CHECK_SIZEOF and endian tests with
2840         cross-compiler friendly macros.
2841         * aclocal.m4 (AC_COMPILE_CHECK_SIZEOF, AC_C_BIGENDIAN_CROSS): New
2842         macros.
2843
2844 2002-01-18  David Edelsohn  <edelsohn@gnu.org>
2845
2846         * src/powerpc/darwin.S (_ffi_call_AIX): New.
2847         * src/powerpc/aix.S (ffi_call_DARWIN): New.
2848
2849 2002-01-17  David Edelsohn  <edelsohn@gnu.org>
2850
2851         * Makefile.am (EXTRA_DIST): Add Darwin and AIX files.
2852         (TARGET_SRC_POWERPC_AIX): New.
2853         (POWERPC_AIX): New stanza.
2854         * Makefile.in: Regenerate.
2855         * configure.in: Add AIX case.
2856         * configure: Regenerate.
2857         * include/ffi.h.in (ffi_abi): Add FFI_AIX.
2858         * src/powerpc/ffi_darwin.c (ffi_status): Use "long" to scale frame
2859         size.  Fix "long double" support.
2860         (ffi_call): Add FFI_AIX case.
2861         * src/powerpc/aix.S: New.
2862
2863 2001-10-09  John Hornkvist  <john@toastedmarshmallow.com>
2864
2865         Implement Darwin PowerPC ABI.
2866         * configure.in: Handle powerpc-*-darwin*.
2867         * Makefile.am: Set source files for POWERPC_DARWIN.
2868         * configure: Rebuilt.
2869         * Makefile.in: Rebuilt.
2870         * include/ffi.h.in: Define FFI_DARWIN and FFI_DEFAULT_ABI for
2871         POWERPC_DARWIN.
2872         * src/powerpc/darwin.S: New file.
2873         * src/powerpc/ffi_darwin.c: New file.
2874
2875 2001-10-07  Joseph S. Myers  <jsm28@cam.ac.uk>
2876
2877         * src/x86/ffi.c: Fix spelling error of "separate" as "seperate".
2878
2879 2001-07-16  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2880
2881         * src/x86/sysv.S: Avoid gas-only .balign directive.
2882         Use C style comments.
2883
2884 2001-07-16  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2885
2886         * src/alpha/ffi.c (ffi_prep_closure): Avoid gas-only mnemonic.
2887         Fixes PR bootstrap/3563.
2888
2889 2001-06-26  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2890
2891         * src/alpha/osf.S (ffi_closure_osf): Use .rdata for ECOFF.
2892
2893 2001-06-25  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2894
2895         * configure.in: Recognize sparc*-sun-* host.
2896         * configure: Regenerate.
2897
2898 2001-06-06  Andrew Haley  <aph@redhat.com>
2899
2900         * src/alpha/osf.S (__FRAME_BEGIN__): Conditionalize for ELF.
2901
2902 2001-06-03  Andrew Haley  <aph@redhat.com>
2903
2904         * src/alpha/osf.S: Add unwind info.
2905         * src/powerpc/sysv.S: Add unwind info.
2906         * src/powerpc/ppc_closure.S: Likewise.
2907
2908 2000-05-31  Jeff Sturm  <jsturm@one-point.com>
2909
2910         * configure.in: Fix AC_ARG_ENABLE usage.
2911         * configure: Rebuilt.
2912
2913 2001-05-06  Bryce McKinlay  <bryce@waitaki.otago.ac.nz>
2914
2915         * configure.in: Remove warning about beta code.
2916         * configure: Rebuilt.
2917
2918 2001-04-25  Hans Boehm <Hans_Boehm@hp.com>
2919
2920         * src/ia64/unix.S: Restore stack pointer when returning from
2921         ffi_closure_UNIX.
2922         * src/ia64/ffi.c: Fix typo in comment.
2923
2924 2001-04-18  Jim Wilson  <wilson@redhat.com>
2925
2926         * src/ia64/unix.S: Delete unnecessary increment and decrement of loc2
2927         to eliminate RAW DV.
2928
2929 2001-04-12  Bryce McKinlay  <bryce@albatross.co.nz>
2930
2931         * Makefile.am: Make a libtool convenience library.
2932         * Makefile.in: Rebuilt.
2933
2934 2001-03-29  Bryce McKinlay  <bryce@albatross.co.nz>
2935
2936         * configure.in: Use different syntax for subdirectory creation.
2937         * configure: Rebuilt.
2938
2939 2001-03-27  Jon Beniston  <jon@beniston.com>
2940
2941         * configure.in: Added X86_WIN32 target (Win32, CygWin, MingW).
2942         * configure: Rebuilt.
2943         * Makefile.am: Added X86_WIN32 target support.
2944         * Makefile.in: Rebuilt.
2945
2946         * include/ffi.h.in: Added X86_WIN32 target support.
2947
2948         * src/ffitest.c: Doesn't run structure tests for X86_WIN32 targets.
2949         * src/types.c: Added X86_WIN32 target support.
2950
2951         * src/x86/win32.S: New file. Based on sysv.S, but with EH
2952         stuff removed and made to work with CygWin's gas.
2953
2954 2001-03-26  Bryce McKinlay  <bryce@albatross.co.nz>
2955
2956         * configure.in: Make target subdirectory in build dir.
2957         * Makefile.am: Override suffix based rules to specify correct output
2958         subdirectory.
2959         * Makefile.in: Rebuilt.
2960         * configure: Rebuilt.
2961
2962 2001-03-23  Kevin B Hendricks  <khendricks@ivey.uwo.ca>
2963
2964         * src/powerpc/ppc_closure.S: New file.
2965         * src/powerpc/ffi.c (ffi_prep_args): Fixed ABI compatibility bug
2966         involving long long and register pairs.
2967         (ffi_prep_closure): New function.
2968         (flush_icache): Likewise.
2969         (ffi_closure_helper_SYSV): Likewise.
2970         * include/ffi.h.in (FFI_CLOSURES): Define on PPC.
2971         (FFI_TRAMPOLINE_SIZE): Likewise.
2972         (FFI_NATIVE_RAW_API): Likewise.
2973         * Makefile.in: Rebuilt.
2974         * Makefile.am (EXTRA_DIST): Added src/powerpc/ppc_closure.S.
2975         (TARGET_SRC_POWERPC): Likewise.
2976
2977 2001-03-19  Tom Tromey  <tromey@redhat.com>
2978
2979         * Makefile.in: Rebuilt.
2980         * Makefile.am (ffitest_LDFLAGS): New macro.
2981
2982 2001-03-02  Nick Clifton  <nickc@redhat.com>
2983
2984         * include/ffi.h.in: Remove RCS ident string.
2985         * include/ffi_mips.h: Remove RCS ident string.
2986         * src/debug.c: Remove RCS ident string.
2987         * src/ffitest.c: Remove RCS ident string.
2988         * src/prep_cif.c: Remove RCS ident string.
2989         * src/types.c: Remove RCS ident string.
2990         * src/alpha/ffi.c: Remove RCS ident string.
2991         * src/alpha/osf.S: Remove RCS ident string.
2992         * src/arm/ffi.c: Remove RCS ident string.
2993         * src/arm/sysv.S: Remove RCS ident string.
2994         * src/mips/ffi.c: Remove RCS ident string.
2995         * src/mips/n32.S: Remove RCS ident string.
2996         * src/mips/o32.S: Remove RCS ident string.
2997         * src/sparc/ffi.c: Remove RCS ident string.
2998         * src/sparc/v8.S: Remove RCS ident string.
2999         * src/sparc/v9.S: Remove RCS ident string.
3000         * src/x86/ffi.c: Remove RCS ident string.
3001         * src/x86/sysv.S: Remove RCS ident string.
3002
3003 2001-02-08  Joseph S. Myers  <jsm28@cam.ac.uk>
3004
3005         * include/ffi.h.in: Change sourceware.cygnus.com references to
3006         gcc.gnu.org.
3007
3008 2000-12-09  Richard Henderson  <rth@redhat.com>
3009
3010         * src/alpha/ffi.c (ffi_call): Simplify struct return test.
3011         (ffi_closure_osf_inner): Index rather than increment avalue
3012         and arg_types.  Give ffi_closure_osf the raw return value type.
3013         * src/alpha/osf.S (ffi_closure_osf): Handle return value type
3014         promotion.
3015
3016 2000-12-07  Richard Henderson  <rth@redhat.com>
3017
3018         * src/raw_api.c (ffi_translate_args): Fix typo.
3019         (ffi_prep_closure): Likewise.
3020
3021         * include/ffi.h.in [ALPHA]: Define FFI_CLOSURES and
3022         FFI_TRAMPOLINE_SIZE.
3023         * src/alpha/ffi.c (ffi_prep_cif_machdep): Adjust minimal
3024         cif->bytes for new ffi_call_osf implementation.
3025         (ffi_prep_args): Absorb into ...
3026         (ffi_call): ... here.  Do all stack allocation here and
3027         avoid a callback function.
3028         (ffi_prep_closure, ffi_closure_osf_inner): New.
3029         * src/alpha/osf.S (ffi_call_osf): Reimplement with no callback.
3030         (ffi_closure_osf): New.
3031
3032 2000-09-10  Alexandre Oliva  <aoliva@redhat.com>
3033
3034         * config.guess, config.sub, install-sh: Removed.
3035         * ltconfig, ltmain.sh, missing, mkinstalldirs: Likewise.
3036         * Makefile.in: Rebuilt.
3037
3038         * acinclude.m4: Include libtool macros from the top level.
3039         * aclocal.m4, configure: Rebuilt.
3040
3041 2000-08-22  Alexandre Oliva  <aoliva@redhat.com>
3042
3043         * configure.in [i*86-*-freebsd*] (TARGET, TARGETDIR): Set.
3044         * configure: Rebuilt.
3045
3046 2000-05-11  Scott Bambrough  <scottb@netwinder.org>
3047
3048         * libffi/src/arm/sysv.S (ffi_call_SYSV): Doubles are not saved to
3049         memory correctly.  Use conditional instructions, not branches where
3050         possible.
3051
3052 2000-05-04  Tom Tromey  <tromey@cygnus.com>
3053
3054         * configure: Rebuilt.
3055         * configure.in: Match `arm*-*-linux-*'.
3056         From Chris Dornan <cdornan@arm.com>.
3057
3058 2000-04-28  Jakub Jelinek  <jakub@redhat.com>
3059
3060         * Makefile.am (SUBDIRS): Define.
3061         (AM_MAKEFLAGS): Likewise.
3062         (Multilib support.): Add section.
3063         * Makefile.in: Rebuilt.
3064         * ltconfig (extra_compiler_flags, extra_compiler_flags_value):
3065         New variables. Set for gcc using -print-multi-lib. Export them
3066         to libtool.
3067         (sparc64-*-linux-gnu*): Use libsuff 64 for search paths.
3068         * ltmain.sh (B|b|V): Don't throw away gcc's -B, -b and -V options
3069         for -shared links.
3070         (extra_compiler_flags_value, extra_compiler_flags): Check these
3071         for extra compiler options which need to be passed down in
3072         compiler_flags.
3073
3074 2000-04-16  Anthony Green  <green@redhat.com>
3075
3076         * configure: Rebuilt.
3077         * configure.in: Change i*86-pc-linux* to i*86-*-linux*.
3078
3079 2000-04-14  Jakub Jelinek  <jakub@redhat.com>
3080
3081         * include/ffi.h.in (SPARC64): Define for 64bit SPARC builds.
3082         Set SPARC FFI_DEFAULT_ABI based on SPARC64 define.
3083         * src/sparc/ffi.c (ffi_prep_args_v8): Renamed from ffi_prep_args.
3084         Replace all void * sizeofs with sizeof(int).
3085         Only compare type with FFI_TYPE_LONGDOUBLE if LONGDOUBLE is
3086         different than DOUBLE.
3087         Remove FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases (handled elsewhere).
3088         (ffi_prep_args_v9): New function.
3089         (ffi_prep_cif_machdep): Handle V9 ABI and long long on V8.
3090         (ffi_V9_return_struct): New function.
3091         (ffi_call): Handle FFI_V9 ABI from 64bit code and FFI_V8 ABI from
3092         32bit code (not yet cross-arch calls).
3093         * src/sparc/v8.S: Add struct return delay nop.
3094         Handle long long.
3095         * src/sparc/v9.S: New file.
3096         * src/prep_cif.c (ffi_prep_cif): Return structure pointer
3097         is used on sparc64 only for structures larger than 32 bytes.
3098         Pass by reference for structures is done for structure arguments
3099         larger than 16 bytes.
3100         * src/ffitest.c (main): Use 64bit rint on sparc64.
3101         Run long long tests on sparc.
3102         * src/types.c (FFI_TYPE_POINTER): Pointer is 64bit on alpha and
3103         sparc64.
3104         (FFI_TYPE_LONGDOUBLE): long double is 128 bit aligned to 128 bits
3105         on sparc64.
3106         * configure.in (sparc-*-linux*): New supported target.
3107         (sparc64-*-linux*): Likewise.
3108         * configure: Rebuilt.
3109         * Makefile.am: Add v9.S to SPARC files.
3110         * Makefile.in: Likewise.
3111         (LINK): Surround $(CCLD) into double quotes, so that multilib
3112         compiles work correctly.
3113
3114 2000-04-04  Alexandre Petit-Bianco  <apbianco@cygnus.com>
3115
3116         * configure: Rebuilt.
3117         * configure.in: (i*86-*-solaris*): New libffi target. Patch
3118         proposed by Bryce McKinlay.
3119
3120 2000-03-20  Tom Tromey  <tromey@cygnus.com>
3121
3122         * Makefile.in: Hand edit for java_raw_api.lo.
3123
3124 2000-03-08  Bryce McKinlay  <bryce@albatross.co.nz>
3125
3126         * config.guess, config.sub: Update from the gcc tree.
3127         Fix for PR libgcj/168.
3128
3129 2000-03-03  Tom Tromey  <tromey@cygnus.com>
3130
3131         * Makefile.in: Fixed ia64 by hand.
3132
3133         * configure: Rebuilt.
3134         * configure.in (--enable-multilib): New option.
3135         (libffi_basedir): New subst.
3136         (AC_OUTPUT): Added multilib code.
3137
3138 2000-03-02  Tom Tromey  <tromey@cygnus.com>
3139
3140         * Makefile.in: Rebuilt.
3141         * Makefile.am (TARGET_SRC_IA64): Use `ia64', not `alpha', as
3142         directory name.
3143
3144 2000-02-25  Hans Boehm <boehm@acm.org>
3145
3146         * src/ia64/ffi.c, src/ia64/ia64_flags.h, src/ia64/unix.S: New
3147         files.
3148         * src/raw_api.c (ffi_translate_args): Fixed typo in argument
3149         list.
3150         (ffi_prep_raw_closure): Use ffi_translate_args, not
3151         ffi_closure_translate.
3152         * src/java_raw_api.c: New file.
3153         * src/ffitest.c (closure_test_fn): New function.
3154         (main): Define `rint' as long long on IA64.  Added new test when
3155         FFI_CLOSURES is defined.
3156         * include/ffi.h.in (ALIGN): Use size_t, not unsigned.
3157         (ffi_abi): Recognize IA64.
3158         (ffi_raw): Added `flt' field.
3159         Added "Java raw API" code.
3160         * configure.in: Recognize ia64.
3161         * Makefile.am (TARGET_SRC_IA64): New macro.
3162         (libffi_la_common_SOURCES): Added java_raw_api.c.
3163         (libffi_la_SOURCES): Define in IA64 case.
3164
3165 2000-01-04  Tom Tromey  <tromey@cygnus.com>
3166
3167         * Makefile.in: Rebuilt with newer automake.
3168
3169 1999-12-31  Tom Tromey  <tromey@cygnus.com>
3170
3171         * Makefile.am (INCLUDES): Added -I$(top_srcdir)/src.
3172
3173 1999-09-01  Tom Tromey  <tromey@cygnus.com>
3174
3175         * include/ffi.h.in: Removed PACKAGE and VERSION defines and
3176         undefs.
3177         * fficonfig.h.in: Rebuilt.
3178         * configure: Rebuilt.
3179         * configure.in: Pass 3rd argument to AM_INIT_AUTOMAKE.
3180         Use AM_PROG_LIBTOOL (automake 1.4 compatibility).
3181         * acconfig.h: Don't #undef PACKAGE or VERSION.
3182
3183 1999-08-09  Anthony Green  <green@cygnus.com>
3184
3185         * include/ffi.h.in: Try to work around messy header problem
3186         with PACKAGE and VERSION.
3187
3188         * configure: Rebuilt.
3189         * configure.in: Change version to 2.00-beta.
3190
3191         * fficonfig.h.in: Rebuilt.
3192         * acconfig.h (FFI_NO_STRUCTS, FFI_NO_RAW_API): Define.
3193
3194         * src/x86/ffi.c (ffi_raw_call): Rename.
3195
3196 1999-08-02  Kresten Krab Thorup  <krab@dominiq.is.s.u-tokyo.ac.jp>
3197
3198         * src/x86/ffi.c (ffi_closure_SYSV): New function.
3199         (ffi_prep_incoming_args_SYSV): Ditto.
3200         (ffi_prep_closure): Ditto.
3201         (ffi_closure_raw_SYSV): Ditto.
3202         (ffi_prep_raw_closure): More ditto.
3203         (ffi_call_raw): Final ditto.
3204
3205         * include/ffi.h.in: Add definitions for closure and raw API.
3206
3207         * src/x86/ffi.c (ffi_prep_cif_machdep): Added case for
3208         FFI_TYPE_UINT64.
3209
3210         * Makefile.am (libffi_la_common_SOURCES): Added raw_api.c
3211
3212         * src/raw_api.c: New file.
3213
3214         * include/ffi.h.in (ffi_raw): New type.
3215         (UINT_ARG, SINT_ARG): New defines.
3216         (ffi_closure, ffi_raw_closure): New types.
3217         (ffi_prep_closure, ffi_prep_raw_closure): New declarations.
3218
3219         * configure.in: Add check for endianness and sizeof void*.
3220
3221         * src/x86/sysv.S (ffi_call_SYSV): Call fixup routine via argument,
3222         instead of directly.
3223
3224         * configure: Rebuilt.
3225
3226 Thu Jul  8 14:28:42 1999  Anthony Green  <green@cygnus.com>
3227
3228         * configure.in: Add x86 and powerpc BeOS configurations.
3229         From Makoto Kato <m_kato@ga2.so-net.ne.jp>.
3230
3231 1999-05-09  Anthony Green  <green@cygnus.com>
3232
3233         * configure.in: Add warning about this being beta code.
3234         Remove src/Makefile.am from the picture.
3235         * configure: Rebuilt.
3236
3237         * Makefile.am: Move logic from src/Makefile.am.  Add changes
3238         to support libffi as a target library.
3239         * Makefile.in: Rebuilt.
3240
3241         * aclocal.m4, config.guess, config.sub, ltconfig, ltmain.sh:
3242         Upgraded to new autoconf, automake, libtool.
3243
3244         * README: Tweaks.
3245
3246         * LICENSE: Update copyright date.
3247
3248         * src/Makefile.am, src/Makefile.in: Removed.
3249
3250 1998-11-29  Anthony Green  <green@cygnus.com>
3251
3252         * include/ChangeLog: Removed.
3253         * src/ChangeLog: Removed.
3254         * src/mips/ChangeLog: Removed.
3255         * src/sparc/ChangeLog: Remboved.
3256         * src/x86/ChangeLog: Removed.
3257
3258         * ChangeLog.v1: Created.