libffi/ChangeLog

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