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