OSDN Git Service

Backported from mainline
[pf3gnuchains/gcc-fork.git] / libffi / ChangeLog
1 2013-04-11  Release Manager
2
3         * GCC 4.7.3 released.
4
5 2012-09-20  Jakub Jelinek  <jakub@redhat.com>
6
7         PR other/43620
8         * configure.ac (AM_INIT_AUTOMAKE): Add no-dist.
9         * Makefile.in: Regenerated.
10         * include/Makefile.in: Regenerated.
11         * man/Makefile.in: Regenerated.
12         * testsuite/Makefile.in: Regenerated.
13
14 2012-09-20  Release Manager
15
16         * GCC 4.7.2 released.
17
18 2012-06-14  Release Manager
19
20         * GCC 4.7.1 released.
21
22 2012-03-22  David Edelsohn  <dje.gcc@gmail.com>
23
24         Backport from mainline:
25         2012-03-09  David Edelsohn  <dje.gcc@gmail.com>
26
27         * src/powerpc/aix_closure.S (ffi_closure_ASM): Adjust for Darwin64
28         change to return value of ffi_closure_helper_DARWIN and load type
29         from return type.
30
31         From Tom Honermann <tom.honermann@oracle.com>:
32         * src/powerpc/aix.S: Declare .ffi_prep_args.  Insert nops after
33         branch instructions.
34         * src/powerpc/aix_closure.S: Declare .ffi_closure_helper_DARWIN.
35
36 2012-03-22  Release Manager
37
38         * GCC 4.7.0 released.
39
40 2012-02-27  Mikael Pettersson  <mikpe@it.uu.se>
41
42         PR libffi/52223
43         * Makefile.am (FLAGS_TO_PASS): Define.
44         * Makefile.in: Regenerate.
45
46 2012-02-23  Kai Tietz  <ktietz@redhat.com>
47
48         PR libffi/52221
49         * src/x86/ffi.c (ffi_closure_raw_THISCALL): New
50         prototype.
51         (ffi_prep_raw_closure_loc): Use ffi_closure_raw_THISCALL for
52         thiscall-convention.
53         (ffi_raw_call): Use ffi_prep_args_raw.
54         * src/x86/win32.S (ffi_closure_raw_THISCALL): Add
55         implementation for stub.
56
57 2012-02-13  Kai Tietz  <ktietz@redhat.com>
58
59         PR libffi/52221
60         * src/x86/ffi.c (ffi_prep_raw_closure_loc): Add thiscall
61         support for X86_WIN32.
62         (FFI_INIT_TRAMPOLINE_THISCALL): Fix displacement.
63
64 2012-02-11  Eric Botcazou  <ebotcazou@adacore.com>
65
66         * src/sparc/v9.S (STACKFRAME): Bump to 176.
67
68 2012-02-10  Kai Tietz  <ktietz@redhat.com>
69
70         * configure.ac (AM_LTLDFLAGS): Add -no-undefine for x64
71         windows target.
72         * configure: Regenerated.
73         
74 2012-02-08  Kai Tietz  <ktietz@redhat.com>
75
76         * src/prep_cif.c (ffi_prep_cif): Allow for X86_WIN32
77         also FFI_THISCALL.
78         * src/x86/ffi.c (ffi_closure_THISCALL): Add prototype.
79         (FFI_INIT_TRAMPOLINE_THISCALL): New trampoline code.
80         (ffi_prep_closure_loc): Add FFI_THISCALL support.
81         * src/x86/ffitarget.h (FFI_TRAMPOLINE_SIZE): Adjust size.
82         * src/x86/win32.S (ffi_closure_THISCALL): New closure code
83         for thiscall-calling convention.
84         * testsuite/libffi.call/closure_thiscall.c: New test.
85
86 2012-01-28  Kai Tietz  <ktietz@redhat.com>
87
88         * src/libffi/src/x86/ffi.c (ffi_call_win32): Add new
89         argument to prototype for specify calling-convention.
90         (ffi_call): Add support for stdcall/thiscall convention.
91         (ffi_prep_args): Likewise.
92         (ffi_raw_call): Likewise.
93         * src/x86/ffitarget.h (ffi_abi): Add FFI_THISCALL and
94         FFI_FASTCALL.
95         * src/x86/win32.S (_ffi_call_win32): Add support for
96         fastcall/thiscall calling-convention calls.
97         * testsuite/libffi.call/fastthis1_win32.c: New test.
98         * testsuite/libffi.call/fastthis2_win32.c: New test.
99         * testsuite/libffi.call/fastthis3_win32.c: New test.
100         * testsuite/libffi.call/strlen2_win32.c: New test.
101         * testsuite/libffi.call/many2_win32.c: New test.
102         * testsuite/libffi.call/struct1_win32.c: New test.
103         * testsuite/libffi.call/struct2_win32.c: New test.
104
105 2012-01-23  Andreas Schwab  <schwab@linux-m68k.org>
106
107         * src/m68k/sysv.S (ffi_call_SYSV): Properly test for plain
108         mc68000.  Test for __HAVE_68881__ in addition to __MC68881__.
109
110 2012-01-19  Jakub Jelinek  <jakub@redhat.com>
111
112         PR rtl-optimization/48496
113         * src/ia64/ffi.c (ffi_call): Fix up aliasing violations.
114
115 2012-01-09  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
116
117         * configure.ac (i?86-*-*): Set TARGET to X86_64.
118         * configure: Regenerate.
119
120 2011-12-07  Andrew Pinski  <apinski@cavium.com>
121
122         PR libffi/50051
123         * src/mips/n32.S: Add ".set mips4".
124
125 2011-11-21  Andreas Tobler  <andreast@fgznet.ch>
126
127         * configure: Regenerate.
128
129 2011-11-10  Richard Henderson  <rth@redhat.com>
130
131         * configure.ac (GCC_AS_CFI_PSEUDO_OP): Use it instead of inline check.
132         * configure, aclocal.m4: Rebuild.
133
134 2011-09-04  Iain Sandoe  <iains@gcc.gnu.org>
135
136         PR libffi/49594
137         * src/powerpc/darwin_closure.S (stubs): Make the stub binding
138         helper reference track the architecture pointer size.
139
140 2011-08-25  Andrew Haley  <aph@redhat.com>
141
142         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Remove hard-coded assembly
143         instructions.
144         * src/arm/sysv.S (ffi_arm_trampoline): Put them here instead.
145
146 2011-07-11  Andrew Haley  <aph@redhat.com>
147
148         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Clear icache.
149
150 2011-06-29  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
151
152         * testsuite/libffi.call/cls_double_va.c: Move PR number to comment.
153         * testsuite/libffi.call/cls_longdouble_va.c: Likewise.
154
155 2011-06-29  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
156
157         PR libffi/46660
158         * testsuite/libffi.call/cls_double_va.c: xfail dg-output on
159         mips-sgi-irix6*.
160         * testsuite/libffi.call/cls_longdouble_va.c: Likewise.
161
162 2011-06-14  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
163
164         * testsuite/libffi.call/huge_struct.c (test_large_fn): Use PRIu8,
165         PRId8 instead of %hhu, %hhd.
166         * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRId8,
167         PRIu8): Define.
168         [__sgi__] (PRId8, PRIu8): Define.
169
170 2011-04-29  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
171
172         * src/alpha/osf.S (UA_SI, FDE_ENCODING, FDE_ENCODE, FDE_ARANGE):
173         Define.
174         Use them to handle ELF vs. ECOFF differences.
175         [__osf__] (_GLOBAL__F_ffi_call_osf): Define.
176
177 2011-02-13  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
178
179         * configure: Regenerate.
180
181 2011-02-09  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
182
183         PR libffi/46661
184         * testsuite/libffi.call/cls_pointer.c (main): Cast void * to
185         uintptr_t first.
186         * testsuite/libffi.call/cls_pointer_stack.c (main): Likewise.
187
188 2011-02-07  Joel Sherrill <joel.sherrill@oarcorp.com>
189
190         * libffi/src/m68k/ffi.c: Add RTEMS support for cache flushing.
191         Handle case when CPU variant does not have long double support.
192         * libffi/src/m68k/sysv.S: Add support for mc68000, Coldfire,
193         and cores with soft floating point.
194
195 2011-02-07  Joel Sherrill <joel.sherrill@oarcorp.com>
196
197         * configure.ac: Add mips*-*-rtems* support.
198         * configure: Regenerate.
199         * src/mips/ffitarget.h: Ensure needed constants are available
200         for targets which do not have sgidefs.h.
201
202 2011-01-26  Dave Korn  <dave.korn.cygwin@gmail.com>
203
204         PR target/40125
205         * configure.ac (AM_LTLDFLAGS): Add -bindir option for windows DLLs.
206         * configure: Regenerate.
207
208 2010-12-18  Iain Sandoe  <iains@gcc.gnu.org>
209
210         PR libffi/29152
211         PR libffi/42378
212         * src/powerpc/darwin_closure.S: Provide Darwin64 implementation,
213         update comments.
214         * src/powerpc/ffitarget.h (POWERPC_DARWIN64): New,
215         (FFI_TRAMPOLINE_SIZE): Update for Darwin64.
216         * src/powerpc/darwin.S: Provide Darwin64 implementation,
217         update comments.
218         * src/powerpc/ffi_darwin.c: Likewise.
219
220 2010-12-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
221
222         * configure.ac (libffi_cv_as_ascii_pseudo_op): Use double
223         backslashes.
224         (libffi_cv_as_string_pseudo_op): Likewise.
225         * configure: Regenerate.
226
227 2010-12-03  Chung-Lin Tang  <cltang@codesourcery.com>
228
229         * src/arm/sysv.S (ffi_closure_SYSV): Add UNWIND to .pad directive.
230         (ffi_closure_VFP): Same.
231         (ffi_call_VFP): Move down to before ffi_closure_VFP. Add '.fpu vfp'
232         directive.
233
234 2010-12-01  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
235
236         * testsuite/libffi.call/ffitest.h [__sgi] (PRId64, PRIu64): Define.
237         (PRIuPTR): Define.
238
239 2010-11-29  Richard Henderson  <rth@redhat.com>
240             Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
241
242         * src/x86/sysv.S (FDE_ENCODING, FDE_ENCODE): Define.
243         (.eh_frame): Use FDE_ENCODING.
244         (.LASFDE1, .LASFDE2, LASFDE3): Simplify with FDE_ENCODE.
245
246 2010-11-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
247
248         * testsuite/lib/libffi-dg.exp: Rename ...
249         * testsuite/lib/libffi.exp: ... to this.
250         * libffi/testsuite/libffi.call/call.exp: Don't load libffi-dg.exp.
251         * libffi/testsuite/libffi.special/special.exp: Likewise.
252
253 2010-10-28  Chung-Lin Tang  <cltang@codesourcery.com>
254
255         * src/arm/ffi.c (ffi_prep_args): Add VFP register argument handling
256         code, new parameter, and return value. Update comments.
257         (ffi_prep_cif_machdep): Add case for VFP struct return values. Add
258         call to layout_vfp_args().
259         (ffi_call_SYSV): Update declaration.
260         (ffi_call_VFP): New declaration.
261         (ffi_call): Add VFP struct return conditions. Call ffi_call_VFP()
262         when ABI is FFI_VFP.
263         (ffi_closure_VFP): New declaration.
264         (ffi_closure_SYSV_inner): Add new vfp_args parameter, update call to
265         ffi_prep_incoming_args_SYSV().
266         (ffi_prep_incoming_args_SYSV): Update parameters. Add VFP argument
267         case handling.
268         (ffi_prep_closure_loc): Pass ffi_closure_VFP to trampoline
269         construction under VFP hard-float.
270         (rec_vfp_type_p): New function.
271         (vfp_type_p): Same.
272         (place_vfp_arg): Same.
273         (layout_vfp_args): Same.
274         * src/arm/ffitarget.h (ffi_abi): Add FFI_VFP. Define FFI_DEFAULT_ABI
275         based on __ARM_PCS_VFP.
276         (FFI_EXTRA_CIF_FIELDS): Define for adding VFP hard-float specific
277         fields.
278         (FFI_TYPE_STRUCT_VFP_FLOAT): Define internally used type code.
279         (FFI_TYPE_STRUCT_VFP_DOUBLE): Same.
280         * src/arm/sysv.S (ffi_call_SYSV): Change call of ffi_prep_args() to
281         direct call. Move function pointer load upwards.
282         (ffi_call_VFP): New function.
283         (ffi_closure_VFP): Same.
284
285         * testsuite/lib/libffi-dg.exp (check-flags): New function.
286         (dg-skip-if): New function.
287         * testsuite/libffi.call/cls_double_va.c: Skip if target is arm*-*-*
288         and compiler options include -mfloat-abi=hard.
289         * testsuite/libffi.call/cls_longdouble_va.c: Same.
290
291 2010-10-01  Jakub Jelinek  <jakub@redhat.com>
292
293         PR libffi/45677
294         * src/x86/ffi64.c (ffi_prep_cif_machdep): Ensure cif->bytes is
295         a multiple of 8.
296         * testsuite/libffi.call/many2.c: New test.
297
298 2010-08-20  Mark Wielaard  <mjw@redhat.com>
299
300         * src/closures.c (open_temp_exec_file_mnt): Check if getmntent_r
301         returns NULL.
302
303 2010-08-09  Andreas Tobler  <andreast@fgznet.ch>
304
305         * configure.ac: Add target powerpc64-*-freebsd*.
306         * configure: Regenerate.
307         * testsuite/libffi.call/cls_align_longdouble_split.c: Pass
308         -mlong-double-128 only to linux targets.
309         * testsuite/libffi.call/cls_align_longdouble_split2.c: Likewise.
310         * testsuite/libffi.call/cls_longdouble.c: Likewise.
311         * testsuite/libffi.call/huge_struct.c: Likewise.
312
313 2010-07-10  Evan Phoenix  <evan@fallingsnow.net>
314
315         * src/closures.c (selinux_enabled_check): Fix strncmp usage bug.
316
317 2010-07-07  Dan HorĂ¡k <dan@danny.cz>
318
319         * include/ffi.h.in: Protect #define with #ifndef.
320         * src/powerpc/ffitarget.h: Ditto.
321         * src/s390/ffitarget.h: Ditto.
322         * src/sparc/ffitarget.h: Ditto.
323
324 2010-07-07   Neil Roberts <neil@linux.intel.com>
325
326         * src/x86/sysv.S (ffi_call_SYSV): Align the stack pointer to
327         16-bytes.
328
329 2010-07-02  Jakub Jelinek  <jakub@redhat.com>
330
331         * Makefile.am (AM_MAKEFLAGS): Pass also mandir to submakes.
332         * Makefile.in: Regenerated.
333
334 2010-05-19  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
335
336         * configure.ac (libffi_cv_as_x86_pcrel): Check for illegal in as
337         output, too.
338         (libffi_cv_as_ascii_pseudo_op): Check for .ascii.
339         (libffi_cv_as_string_pseudo_op): Check for .string.
340         * configure: Regenerate.
341         * fficonfig.h.in: Regenerate.
342         * src/x86/sysv.S (.eh_frame): Use .ascii, .string or error.
343
344 2010-05-05  Michael Kohler <michaelkohler@live.com>
345
346         * src/dlmalloc.c (dlfree): Fix spelling.
347         * src/ia64/ffi.c (ffi_prep_cif_machdep): Ditto.
348         * configure.ac: Ditto.
349         * configure: Rebuilt.
350
351 2010-04-13  Dan Witte  <dwitte@mozilla.com>
352
353         * msvcc.sh: Build with -W3 instead of -Wall.
354         * src/powerpc/ffi_darwin.c: Remove build warnings.
355         * src/x86/ffi.c: Ditto.
356         * src/x86/ffitarget.h: Ditto.
357
358 2010-04-12  Dan Witte  <dwitte@mozilla.com>
359             Walter Meinl <wuno@lsvw.de>
360
361         * configure.ac: Add OS/2 support.
362         * configure: Rebuilt.
363         * src/closures.c: Ditto.
364         * src/dlmalloc.c: Ditto.
365         * src/x86/win32.S: Ditto.
366
367 2010-04-07  Jakub Jelinek  <jakub@redhat.com>
368
369         * testsuite/libffi.call/err_bad_abi.c: Remove unused args variable.
370
371 2010-04-02  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
372
373         * Makefile.in: Regenerate.
374         * aclocal.m4: Regenerate.
375         * include/Makefile.in: Regenerate.
376         * man/Makefile.in: Regenerate.
377         * testsuite/Makefile.in: Regenerate.
378
379 2010-03-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
380
381         * configure.ac (libffi_cv_as_x86_64_unwind_section_type): New test.
382         * configure: Regenerate.
383         * fficonfig.h.in: Regenerate.
384         * libffi/src/x86/unix64.S (.eh_frame)
385         [HAVE_AS_X86_64_UNWIND_SECTION_TYPE]: Use @unwind section type.
386
387 2010-03-14  Matthias Klose  <doko@ubuntu.com>
388
389         * src/x86/ffi64.c: Fix typo in comment.
390         * src/x86/ffi.c: Use /* ... */ comment style.
391
392 2010-02-24  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
393
394         * doc/libffi.texi (The Closure API): Fix typo.
395         * doc/libffi.info: Remove.
396
397 2010-02-15  Matthias Klose  <doko@ubuntu.com>
398
399         * src/arm/sysv.S (__ARM_ARCH__): Define for processor
400         __ARM_ARCH_7EM__.
401
402 2010-01-15  Anthony Green  <green@redhat.com>
403
404         * README: Add notes on building with Microsoft Visual C++.
405
406 2010-01-15  Daniel Witte  <dwitte@mozilla.com>
407
408         * msvcc.sh: New file.
409
410         * src/x86/win32.S: Port assembly routines to MSVC and #ifdef.
411         * src/x86/ffi.c: Tweak function declaration and remove excess
412         parens.
413         * include/ffi.h.in: Add __declspec(align(8)) to typedef struct
414         ffi_closure.
415
416         * src/x86/ffi.c: Merge ffi_call_SYSV and ffi_call_STDCALL into new
417         function ffi_call_win32 on X86_WIN32.
418         * src/x86/win32.S (ffi_call_SYSV): Rename to ffi_call_win32.
419         (ffi_call_STDCALL): Remove.
420
421         * src/prep_cif.c (ffi_prep_cif): Move stack space allocation code
422         to ffi_prep_cif_machdep for x86.
423         * src/x86/ffi.c (ffi_prep_cif_machdep): To here.
424
425 2010-01-15  Oliver Kiddle  <okiddle@yahoo.co.uk>
426
427         * src/x86/ffitarget.h (ffi_abi): Check for __i386 and __amd64 for
428         Sun Studio compiler compatibility.
429
430 2010-01-12  Conrad Irwin <conrad.irwin@gmail.com>
431
432         * doc/libffi.texi: Add closure example.
433
434 2010-01-07  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
435
436         PR libffi/40701
437         * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRIdLL,
438         PRIuLL, PRId64, PRIu64, PRIuPTR): Define.
439         * testsuite/libffi.call/cls_align_sint64.c: Add -Wno-format on
440         alpha*-dec-osf*.
441         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
442         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
443         * testsuite/libffi.call/return_ll1.c: Likewise.
444         * testsuite/libffi.call/stret_medium2.c: Likewise.
445         * testsuite/libffi.special/ffitestcxx.h (allocate_mmap): Cast
446         MAP_FAILED to char *.
447
448 2010-01-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
449
450         * src/mips/n32.S: Use .abicalls and .eh_frame with __GNUC__.
451
452 2009-12-31  Anthony Green  <green@redhat.com>
453
454         * README: Update for libffi 3.0.9.
455
456 2009-12-27  Matthias Klose  <doko@ubuntu.com>
457
458         * configure.ac (HAVE_LONG_DOUBLE): Define for mips when
459         appropriate.
460         * configure: Rebuilt.
461
462 2009-12-26  Anthony Green  <green@redhat.com>
463
464         * testsuite/libffi.call/cls_longdouble_va.c: Mark as xfail for
465         avr32*-*-*.
466         * testsuite/libffi.call/cls_double_va.c: Ditto.
467
468 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
469
470         * testsuite/libffi.call/ffitest.h: Conditionally include stdint.h
471         and inttypes.h.
472         * testsuite/libffi.special/unwindtest.cc: Ditto.
473
474 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
475
476         * configure.ac: Add amd64-*-openbsd*.
477         * configure: Rebuilt.
478         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Link
479         openbsd programs with -lpthread.
480
481 2009-12-26  Anthony Green  <green@redhat.com>
482
483         * testsuite/libffi.call/cls_double_va.c,
484         testsuite/libffi.call/cls_longdouble.c,
485         testsuite/libffi.call/cls_longdouble_va.c,
486         testsuite/libffi.call/cls_pointer.c,
487         testsuite/libffi.call/cls_pointer_stack.c: Remove xfail for
488         mips*-*-* and arm*-*-*.
489         * testsuite/libffi.call/cls_align_longdouble_split.c,
490         testsuite/libffi.call/cls_align_longdouble_split2.c,
491         testsuite/libffi.call/stret_medium2.c,
492         testsuite/libffi.call/stret_medium.c,
493         testsuite/libffi.call/stret_large.c,
494         testsuite/libffi.call/stret_large2.c: Remove xfail for arm*-*-*.
495
496 2009-12-31  Kay Tietz  <ktietz70@googlemail.com>
497
498         * testsuite/libffi.call/ffitest.h,
499         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRuLL): Fix
500         definitions.
501
502 2009-12-31  Carlo Bramini  <carlo.bramix@libero.it>
503
504         * configure.ac (AM_LTLDFLAGS): Define for windows hosts.
505         * Makefile.am (libffi_la_LDFLAGS): Add AM_LTLDFLAGS.
506         * configure: Rebuilt.
507         * Makefile.in: Rebuilt.
508
509 2009-12-31  Anthony Green  <green@redhat.com>
510             Blake Chaffin.
511
512         * testsuite/libffi.call/huge_struct.c: New test case from Blake
513         Chaffin @ Apple.
514
515 2009-12-28  David Edelsohn  <edelsohn@gnu.org>
516
517         * src/powerpc/ffi_darwin.c (ffi_prep_args): Copy abi and nargs to
518         local variables.
519         (aix_adjust_aggregate_sizes): New function.
520         (ffi_prep_cif_machdep): Call it.
521
522 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
523
524         * configure.ac: Define FFI_MMAP_EXEC_WRIT for the given targets.
525         * configure: Regenerate.
526         * fficonfig.h.in: Likewise.
527         * src/closures.c: Remove the FFI_MMAP_EXEC_WRIT definition for
528         Solaris/x86.
529
530 2009-12-26  Andreas Schwab  <schwab@linux-m68k.org>
531
532         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Advance intarg_count
533         when a float arguments is passed in memory.
534         (ffi_closure_helper_SYSV): Mark general registers as used up when
535         a 64bit or soft-float long double argument is passed in memory.
536
537 2009-12-25  Matthias Klose  <doko@ubuntu.com>
538
539         * man/ffi_call.3: Fix #include in examples.
540         * doc/libffi.texi: Add dircategory.
541
542 2009-12-25  Frank Everdij <f.p.x.everdij@tudelft.nl>
543
544         * include/ffi.h.in: Placed '__GNUC__' ifdef around
545         '__attribute__((aligned(8)))' in ffi_closure, fixes compile for
546         IRIX MIPSPro c99.
547         * include/ffi_common.h: Added '__sgi' define to non
548         '__attribute__((__mode__()))' integer typedefs.
549         * src/mips/ffi.c (ffi_call, ffi_closure_mips_inner_O32,
550         ffi_closure_mips_inner_N32): Added 'defined(_MIPSEB)' to BE check.
551         (ffi_closure_mips_inner_O32, ffi_closure_mips_inner_N32): Added
552         FFI_LONGDOUBLE support and alignment(N32 only).
553         * src/mips/ffitarget.h: Corrected '#include <sgidefs.h>' for IRIX and
554         fixed non '__attribute__((__mode__()))' integer typedefs.
555         * src/mips/n32.S: Put '#ifdef linux' around '.abicalls' and '.eh_frame'
556         since they are Linux/GNU Assembler specific.
557
558 2009-12-25  Bradley Smith  <brad@brad-smith.co.uk>
559
560         * configure.ac, Makefile.am, src/avr32/ffi.c,
561         src/avr32/ffitarget.h,
562         src/avr32/sysv.S: Add AVR32 port.
563         * configure, Makefile.in: Rebuilt.
564
565 2009-12-21  Andreas Tobler  <a.tobler@schweiz.org>
566
567         * configure.ac: Make i?86 build on FreeBSD and OpenBSD.
568         * configure: Regenerate.
569
570 2009-12-15  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
571
572         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on PA HP-UX.
573
574 2009-12-13  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
575
576         * src/pa/ffi.c (ffi_closure_inner_pa32): Handle FFI_TYPE_LONGDOUBLE
577         type on HP-UX.
578
579 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
580
581         * src/sparc/ffi.c (ffi_closure_sparc_inner_v9): Properly align 'long
582         double' arguments.
583
584 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
585
586         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on Solaris < 10.
587
588 2009-12-10  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
589
590         PR libffi/40700
591         * src/closures.c [X86_64 && __sun__ && __svr4__]
592         (FFI_MMAP_EXEC_WRIT): Define.
593
594 2009-12-08  David Daney  <ddaney@caviumnetworks.com>
595
596         * testsuite/libffi.call/stret_medium.c: Remove xfail for mips*-*-*
597         * testsuite/libffi.call/cls_align_longdouble_split2.c: Same.
598         * testsuite/libffi.call/stret_large.c: Same.
599         * testsuite/libffi.call/cls_align_longdouble_split.c: Same.
600         * testsuite/libffi.call/stret_large2.c: Same.
601         * testsuite/libffi.call/stret_medium2.c: Same.
602
603 2009-12-07  David Edelsohn  <edelsohn@gnu.org>
604
605         * src/powerpc/aix_closure.S (libffi_closure_ASM): Fix tablejump
606         typo.
607
608 2009-12-05  David Edelsohn  <edelsohn@gnu.org>
609
610         * src/powerpc/aix.S: Update AIX32 code to be consistent with AIX64
611         code.
612         * src/powerpc/aix_closure.S: Same.
613
614 2009-12-05  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
615
616         * Makefile.in: Regenerate.
617         * configure: Regenerate.
618         * include/Makefile.in: Regenerate.
619         * man/Makefile.in: Regenerate.
620         * testsuite/Makefile.in: Regenerate.
621
622 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
623
624         * src/powerpc/aix_closure.S: Reorganize 64-bit code to match
625         linux64_closure.S.
626
627 2009-12-04  Uros Bizjak  <ubizjak@gmail.com>
628
629         PR libffi/41908
630         * src/x86/ffi64.c (classify_argument): Update from
631         gcc/config/i386/i386.c.
632         (ffi_closure_unix64_inner): Do not use the address of two consecutive
633         SSE registers directly.
634         * testsuite/libffi.call/cls_dbls_struct.c (main): Remove xfail
635         for x86_64 linux targets.
636
637 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
638
639         * src/powerpc/ffi_darwin.c (ffi_closure_helper_DARWIN): Increment
640         pfr for long double split between fpr13 and stack.
641
642 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
643
644         * src/powerpc/ffi_darwin.c (ffi_prep_args): Increment next_arg and
645         fparg_count twice for long double.
646
647 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
648
649         PR libffi/42243
650         * src/powerpc/ffi_darwin.c (ffi_prep_args): Remove extra parentheses.
651
652 2009-12-03  Uros Bizjak  <ubizjak@gmail.com>
653
654         * testsuite/libffi.call/cls_longdouble_va.c (main): Fix format string.
655         Remove xfails for x86 linux targets.
656
657 2009-12-02  David Edelsohn  <edelsohn@gnu.org>
658
659         * src/powerpc/ffi_darwin.c (ffi_prep_args): Fix typo in INT64
660         case.
661
662 2009-12-01  David Edelsohn  <edelsohn@gnu.org>
663
664         * src/powerpc/aix.S (ffi_call_AIX): Convert to more standard
665         register usage.  Call ffi_prep_args directly.  Add long double
666         return value support.
667         * src/powerpc/ffi_darwin.c (ffi_prep_args): Double arg increment
668         applies to FFI_TYPE_DOUBLE.  Correct fpr_base increment typo.
669         Separate FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases.
670         (ffi_prep_cif_machdep): Only 16 byte stack alignment in 64 bit
671         mode.
672         (ffi_closure_helper_DARWIN): Remove nf and ng counters.  Move temp
673         into case.
674         * src/powerpc/aix_closure.S: Maintain 16 byte stack alignment.
675         Allocate result area between params and FPRs.
676
677 2009-11-30  David Edelsohn  <edelsohn@gnu.org>
678
679         PR target/35484
680         * src/powerpc/ffitarget.h (POWERPC64): Define for PPC64 Linux and
681         AIX64.
682         * src/powerpc/aix.S: Implement AIX64 version.
683         * src/powerpc/aix_closure.S: Implement AIX64 version.
684         (ffi_closure_ASM): Use extsb, lha and displament addresses.
685         * src/powerpc/ffi_darwin.c (ffi_prep_args): Implement AIX64
686         support.
687         (ffi_prep_cif_machdep): Same.
688         (ffi_call): Same.
689         (ffi_closure_helper_DARWIN): Same.
690
691 2009-11-02  Andreas Tobler  <a.tobler@schweiz.org>
692
693         PR libffi/41908
694         * testsuite/libffi.call/testclosure.c: New test.
695
696 2009-09-28  Kai Tietz  <kai.tietz@onevision.com>
697
698         * src/x86/win64.S (_ffi_call_win64 stack): Remove for gnu
699         assembly version use of ___chkstk.
700
701 2009-09-23  Matthias Klose  <doko@ubuntu.com>
702
703         PR libffi/40242, PR libffi/41443
704         * src/arm/sysv.S (__ARM_ARCH__): Define for processors
705         __ARM_ARCH_6T2__, __ARM_ARCH_6M__, __ARM_ARCH_7__,
706         __ARM_ARCH_7A__, __ARM_ARCH_7R__, __ARM_ARCH_7M__.
707         Change the conditionals to __SOFTFP__ || __ARM_EABI__
708         for -mfloat-abi=softfp to work.
709
710 2009-09-17  Loren J. Rittle  <ljrittle@acm.org>
711
712         PR testsuite/32843 (strikes again)
713         * src/x86/ffi.c (ffi_prep_cif_machdep): Add X86_FREEBSD to
714         enable proper extension on char and short.
715
716 2009-09-15  David Daney  <ddaney@caviumnetworks.com>
717
718         * src/java_raw_api.c (ffi_java_raw_to_rvalue): Remove special
719         handling for FFI_TYPE_POINTER.
720         * src/mips/ffitarget.h (FFI_TYPE_STRUCT_D_SOFT,
721         FFI_TYPE_STRUCT_F_SOFT, FFI_TYPE_STRUCT_DD_SOFT,
722         FFI_TYPE_STRUCT_FF_SOFT, FFI_TYPE_STRUCT_FD_SOFT,
723         FFI_TYPE_STRUCT_DF_SOFT, FFI_TYPE_STRUCT_SOFT): New defines.
724         (FFI_N32_SOFT_FLOAT, FFI_N64_SOFT_FLOAT): New ffi_abi enumerations.
725         (enum ffi_abi): Set FFI_DEFAULT_ABI for soft-float.
726         * src/mips/n32.S (ffi_call_N32): Add handling for soft-float
727         structure and pointer returns.
728         (ffi_closure_N32): Add handling for pointer returns.
729         * src/mips/ffi.c (ffi_prep_args, calc_n32_struct_flags,
730         calc_n32_return_struct_flags): Handle soft-float.
731         (ffi_prep_cif_machdep):  Handle soft-float, fix pointer handling.
732         (ffi_call_N32): Declare proper argument types.
733         (ffi_call, copy_struct_N32, ffi_closure_mips_inner_N32): Handle
734         soft-float.
735
736 2009-08-24  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
737
738         * configure.ac (AC_PREREQ): Bump to 2.64.
739
740 2009-08-22  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
741
742         * Makefile.am (install-html, install-pdf): Remove.
743         * Makefile.in: Regenerate.
744
745         * Makefile.in: Regenerate.
746         * aclocal.m4: Regenerate.
747         * configure: Regenerate.
748         * fficonfig.h.in: Regenerate.
749         * include/Makefile.in: Regenerate.
750         * man/Makefile.in: Regenerate.
751         * testsuite/Makefile.in: Regenerate.
752
753 2009-07-30  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
754
755         * configure.ac (_AC_ARG_VAR_PRECIOUS): Use m4_rename_force.
756
757 2009-07-24  Dave Korn  <dave.korn.cygwin@gmail.com>
758
759         PR libffi/40807
760         * src/x86/ffi.c (ffi_prep_cif_machdep): Also use sign/zero-extending
761         return types for X86_WIN32.
762         * src/x86/win32.S (_ffi_call_SYSV): Handle omitted return types.
763         (_ffi_call_STDCALL, _ffi_closure_SYSV, _ffi_closure_raw_SYSV,
764         _ffi_closure_STDCALL): Likewise.
765
766         * src/closures.c (is_selinux_enabled): Define to const 0 for Cygwin.
767         (dlmmap, dlmunmap): Also use these functions on Cygwin.
768
769 2009-07-11  Richard Sandiford  <rdsandiford@googlemail.com>
770
771         PR testsuite/40699
772         PR testsuite/40707
773         PR testsuite/40709
774         * testsuite/lib/libffi-dg.exp: Revert 2009-07-02, 2009-07-01 and
775         2009-06-30 commits.
776
777 2009-07-01  Richard Sandiford  <r.sandiford@uk.ibm.com>
778
779         * testsuite/lib/libffi-dg.exp (libffi-init): Set ld_library_path
780         to "" before adding paths.  (This reinstates an assignment that
781         was removed by my 2009-06-30 commit, but changes the initial
782         value from "." to "".)
783
784 2009-07-01  H.J. Lu  <hongjiu.lu@intel.com>
785
786         PR testsuite/40601
787         * testsuite/lib/libffi-dg.exp (libffi-init): Properly set
788         gccdir.  Adjust ld_library_path for gcc only if gccdir isn't
789         empty.
790
791 2009-06-30  Richard Sandiford  <r.sandiford@uk.ibm.com>
792
793         * testsuite/lib/libffi-dg.exp (libffi-init): Don't add "."
794         to ld_library_path.  Use add_path.  Add just find_libgcc_s
795         to ld_library_path, not every libgcc multilib directory.
796
797 2009-06-16  Wim Lewis  <wiml@hhhh.org>
798
799         * src/powerpc/ffi.c: Avoid clobbering cr3 and cr4, which are
800         supposed to be callee-saved.
801         * src/powerpc/sysv.S (small_struct_return_value): Fix overrun of
802         return buffer for odd-size structs.
803
804 2009-06-16  Andreas Tobler  <a.tobler@schweiz.org>
805
806         PR libffi/40444
807         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Add
808         allow_stack_execute for Darwin.
809
810 2009-06-16  Andrew Haley  <aph@redhat.com>
811
812         * configure.ac (TARGETDIR): Add missing blank lines.
813         * configure: Regenerate.
814
815 2009-06-16  Andrew Haley  <aph@redhat.com>
816
817         * testsuite/libffi.call/cls_align_sint64.c,
818         testsuite/libffi.call/cls_align_uint64.c,
819         testsuite/libffi.call/cls_longdouble_va.c,
820         testsuite/libffi.call/cls_ulonglong.c,
821         testsuite/libffi.call/return_ll1.c,
822         testsuite/libffi.call/stret_medium2.c: Fix printf format
823         specifiers.
824         * testsuite/libffi.call/ffitest.h,
825         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRIuLL): Define.
826
827 2009-06-15  Andrew Haley  <aph@redhat.com>
828
829         * testsuite/libffi.call/err_bad_typedef.c: xfail everywhere.
830         * testsuite/libffi.call/err_bad_abi.c: Likewise.
831
832 2009-06-12  Andrew Haley  <aph@redhat.com>
833
834         * Makefile.am: Remove info_TEXINFOS.
835
836 2009-06-12  Andrew Haley  <aph@redhat.com>
837
838         * ChangeLog.libffi: testsuite/libffi.call/cls_align_sint64.c,
839         testsuite/libffi.call/cls_align_uint64.c,
840         testsuite/libffi.call/cls_ulonglong.c,
841         testsuite/libffi.call/return_ll1.c,
842         testsuite/libffi.call/stret_medium2.c: Fix printf format
843         specifiers.
844         testsuite/libffi.special/unwindtest.cc: include stdint.h.
845
846 2009-06-11  Timothy Wall  <twall@users.sf.net>
847
848         * Makefile.am,
849         configure.ac,
850         include/ffi.h.in,
851         include/ffi_common.h,
852         src/closures.c,
853         src/dlmalloc.c,
854         src/x86/ffi.c,
855         src/x86/ffitarget.h,
856         src/x86/win64.S (new),
857         README: Added win64 support (mingw or MSVC)
858         * Makefile.in,
859         include/Makefile.in,
860         man/Makefile.in,
861         testsuite/Makefile.in,
862         configure,
863         aclocal.m4: Regenerated
864         * ltcf-c.sh: properly escape cygwin/w32 path
865         * man/ffi_call.3: Clarify size requirements for return value.
866         * src/x86/ffi64.c: Fix filename in comment.
867         * src/x86/win32.S: Remove unused extern.
868
869         * testsuite/libffi.call/closure_fn0.c,
870         testsuite/libffi.call/closure_fn1.c,
871         testsuite/libffi.call/closure_fn2.c,
872         testsuite/libffi.call/closure_fn3.c,
873         testsuite/libffi.call/closure_fn4.c,
874         testsuite/libffi.call/closure_fn5.c,
875         testsuite/libffi.call/closure_fn6.c,
876         testsuite/libffi.call/closure_stdcall.c,
877         testsuite/libffi.call/cls_12byte.c,
878         testsuite/libffi.call/cls_16byte.c,
879         testsuite/libffi.call/cls_18byte.c,
880         testsuite/libffi.call/cls_19byte.c,
881         testsuite/libffi.call/cls_1_1byte.c,
882         testsuite/libffi.call/cls_20byte.c,
883         testsuite/libffi.call/cls_20byte1.c,
884         testsuite/libffi.call/cls_24byte.c,
885         testsuite/libffi.call/cls_2byte.c,
886         testsuite/libffi.call/cls_3_1byte.c,
887         testsuite/libffi.call/cls_3byte1.c,
888         testsuite/libffi.call/cls_3byte2.c,
889         testsuite/libffi.call/cls_4_1byte.c,
890         testsuite/libffi.call/cls_4byte.c,
891         testsuite/libffi.call/cls_5_1_byte.c,
892         testsuite/libffi.call/cls_5byte.c,
893         testsuite/libffi.call/cls_64byte.c,
894         testsuite/libffi.call/cls_6_1_byte.c,
895         testsuite/libffi.call/cls_6byte.c,
896         testsuite/libffi.call/cls_7_1_byte.c,
897         testsuite/libffi.call/cls_7byte.c,
898         testsuite/libffi.call/cls_8byte.c,
899         testsuite/libffi.call/cls_9byte1.c,
900         testsuite/libffi.call/cls_9byte2.c,
901         testsuite/libffi.call/cls_align_double.c,
902         testsuite/libffi.call/cls_align_float.c,
903         testsuite/libffi.call/cls_align_longdouble.c,
904         testsuite/libffi.call/cls_align_longdouble_split.c,
905         testsuite/libffi.call/cls_align_longdouble_split2.c,
906         testsuite/libffi.call/cls_align_pointer.c,
907         testsuite/libffi.call/cls_align_sint16.c,
908         testsuite/libffi.call/cls_align_sint32.c,
909         testsuite/libffi.call/cls_align_sint64.c,
910         testsuite/libffi.call/cls_align_uint16.c,
911         testsuite/libffi.call/cls_align_uint32.c,
912         testsuite/libffi.call/cls_align_uint64.c,
913         testsuite/libffi.call/cls_dbls_struct.c,
914         testsuite/libffi.call/cls_double.c,
915         testsuite/libffi.call/cls_double_va.c,
916         testsuite/libffi.call/cls_float.c,
917         testsuite/libffi.call/cls_longdouble.c,
918         testsuite/libffi.call/cls_longdouble_va.c,
919         testsuite/libffi.call/cls_multi_schar.c,
920         testsuite/libffi.call/cls_multi_sshort.c,
921         testsuite/libffi.call/cls_multi_sshortchar.c,
922         testsuite/libffi.call/cls_multi_uchar.c,
923         testsuite/libffi.call/cls_multi_ushort.c,
924         testsuite/libffi.call/cls_multi_ushortchar.c,
925         testsuite/libffi.call/cls_pointer.c,
926         testsuite/libffi.call/cls_pointer_stack.c,
927         testsuite/libffi.call/cls_schar.c,
928         testsuite/libffi.call/cls_sint.c,
929         testsuite/libffi.call/cls_sshort.c,
930         testsuite/libffi.call/cls_uchar.c,
931         testsuite/libffi.call/cls_uint.c,
932         testsuite/libffi.call/cls_ulonglong.c,
933         testsuite/libffi.call/cls_ushort.c,
934         testsuite/libffi.call/err_bad_abi.c,
935         testsuite/libffi.call/err_bad_typedef.c,
936         testsuite/libffi.call/float2.c,
937         testsuite/libffi.call/huge_struct.c,
938         testsuite/libffi.call/nested_struct.c,
939         testsuite/libffi.call/nested_struct1.c,
940         testsuite/libffi.call/nested_struct10.c,
941         testsuite/libffi.call/nested_struct2.c,
942         testsuite/libffi.call/nested_struct3.c,
943         testsuite/libffi.call/nested_struct4.c,
944         testsuite/libffi.call/nested_struct5.c,
945         testsuite/libffi.call/nested_struct6.c,
946         testsuite/libffi.call/nested_struct7.c,
947         testsuite/libffi.call/nested_struct8.c,
948         testsuite/libffi.call/nested_struct9.c,
949         testsuite/libffi.call/problem1.c,
950         testsuite/libffi.call/return_ldl.c,
951         testsuite/libffi.call/return_ll1.c,
952         testsuite/libffi.call/stret_large.c,
953         testsuite/libffi.call/stret_large2.c,
954         testsuite/libffi.call/stret_medium.c,
955         testsuite/libffi.call/stret_medium2.c,
956         testsuite/libffi.special/unwindtest.cc: use ffi_closure_alloc instead
957         of checking for MMAP.  Use intptr_t instead of long casts.
958
959 2009-06-11  Kaz Kojima  <kkojima@gcc.gnu.org>
960
961         * testsuite/libffi.call/cls_longdouble_va.c: Add xfail sh*-*-linux-*.
962         * testsuite/libffi.call/err_bad_abi.c: Add xfail sh*-*-*.
963         * testsuite/libffi.call/err_bad_typedef.c: Likewise.
964
965 2009-06-09  Andrew Haley  <aph@redhat.com>
966
967         * src/x86/freebsd.S: Add missing file.
968
969 2009-06-08  Andrew Haley  <aph@redhat.com>
970
971         Import from libffi 3.0.8:
972
973         * doc/libffi.texi: New file.
974         * doc/libffi.info: Likewise.
975         * doc/stamp-vti: Likewise.
976         * man/Makefile.am: New file.
977         * man/ffi_call.3: New file.
978
979         * Makefile.am (EXTRA_DIST): Add src/x86/darwin64.S,
980         src/dlmalloc.c.
981         (nodist_libffi_la_SOURCES): Add X86_FREEBSD.
982
983         * configure.ac: Bump version to 3.0.8.
984         parisc*-*-linux*: Add.
985         i386-*-freebsd* | i386-*-openbsd*: Add.
986         powerpc-*-beos*: Add.
987         AM_CONDITIONAL X86_FREEBSD: Add.
988         AC_CONFIG_FILES: Add man/Makefile.
989
990         * include/ffi.h.in (FFI_FN): Change void (*)() to void (*)(void).
991
992 2009-06-08  Andrew Haley  <aph@redhat.com>
993
994         * README: Import from libffi 3.0.8.
995
996 2009-06-08  Andrew Haley  <aph@redhat.com>
997
998         * testsuite/libffi.call/err_bad_abi.c: Add xfails.
999         * testsuite/libffi.call/cls_longdouble_va.c: Add xfails.
1000         * testsuite/libffi.call/cls_dbls_struct.c: Add xfail x86_64-*-linux-*.
1001         * testsuite/libffi.call/err_bad_typedef.c: Add xfails.
1002
1003         * testsuite/libffi.call/stret_medium2.c: Add __UNUSED__ to args.
1004         * testsuite/libffi.call/stret_medium.c: Likewise.
1005         * testsuite/libffi.call/stret_large2.c: Likewise.
1006         * testsuite/libffi.call/stret_large.c:  Likewise.
1007
1008 2008-12-26  Timothy Wall  <twall@users.sf.net>
1009
1010         * testsuite/libffi.call/cls_longdouble.c,
1011         testsuite/libffi.call/cls_longdouble_va.c,
1012         testsuite/libffi.call/cls_align_longdouble.c,
1013         testsuite/libffi.call/cls_align_longdouble_split.c,
1014         testsuite/libffi.call/cls_align_longdouble_split2.c: mark expected
1015         failures on x86_64 cygwin/mingw.
1016
1017 2008-12-22  Timothy Wall  <twall@users.sf.net>
1018
1019         * testsuite/libffi.call/closure_fn0.c,
1020         testsuite/libffi.call/closure_fn1.c,
1021         testsuite/libffi.call/closure_fn2.c,
1022         testsuite/libffi.call/closure_fn3.c,
1023         testsuite/libffi.call/closure_fn4.c,
1024         testsuite/libffi.call/closure_fn5.c,
1025         testsuite/libffi.call/closure_fn6.c,
1026         testsuite/libffi.call/closure_loc_fn0.c,
1027         testsuite/libffi.call/closure_stdcall.c,
1028         testsuite/libffi.call/cls_align_pointer.c,
1029         testsuite/libffi.call/cls_pointer.c,
1030         testsuite/libffi.call/cls_pointer_stack.c: use portable cast from
1031         pointer to integer (intptr_t).
1032         * testsuite/libffi.call/cls_longdouble.c: disable for win64.
1033
1034 2008-07-24  Anthony Green  <green@redhat.com>
1035
1036         * testsuite/libffi.call/cls_dbls_struct.c,
1037         testsuite/libffi.call/cls_double_va.c,
1038         testsuite/libffi.call/cls_longdouble.c,
1039         testsuite/libffi.call/cls_longdouble_va.c,
1040         testsuite/libffi.call/cls_pointer.c,
1041         testsuite/libffi.call/cls_pointer_stack.c,
1042         testsuite/libffi.call/err_bad_abi.c: Clean up failures from
1043         compiler warnings.
1044
1045 2008-03-04  Anthony Green  <green@redhat.com>
1046             Blake Chaffin
1047             hos@tamanegi.org
1048
1049         * testsuite/libffi.call/cls_align_longdouble_split2.c
1050           testsuite/libffi.call/cls_align_longdouble_split.c
1051           testsuite/libffi.call/cls_dbls_struct.c
1052           testsuite/libffi.call/cls_double_va.c
1053           testsuite/libffi.call/cls_longdouble.c
1054           testsuite/libffi.call/cls_longdouble_va.c
1055           testsuite/libffi.call/cls_pointer.c
1056           testsuite/libffi.call/cls_pointer_stack.c
1057           testsuite/libffi.call/err_bad_abi.c
1058           testsuite/libffi.call/err_bad_typedef.c
1059           testsuite/libffi.call/stret_large2.c
1060           testsuite/libffi.call/stret_large.c
1061           testsuite/libffi.call/stret_medium2.c
1062           testsuite/libffi.call/stret_medium.c: New tests from Apple.
1063
1064 2009-06-05  Andrew Haley  <aph@redhat.com>
1065
1066         * src/x86/ffitarget.h, src/x86/ffi.c: Merge stdcall changes from
1067         libffi.
1068
1069 2009-06-04  Andrew Haley  <aph@redhat.com>
1070
1071         * src/x86/ffitarget.h, src/x86/win32.S, src/x86/ffi.c: Back out
1072         stdcall changes.
1073
1074 2008-02-26  Anthony Green  <green@redhat.com>
1075             Thomas Heller  <theller@ctypes.org>
1076
1077         * src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C
1078         comment.
1079
1080 2008-02-03  Timothy Wall  <twall@users.sf.net>
1081
1082         * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return
1083           offset based on code pointer, not data pointer.
1084
1085 2008-01-31  Timothy Wall <twall@users.sf.net>
1086
1087         * testsuite/libffi.call/closure_stdcall.c: Add test for stdcall
1088         closures.
1089         * src/x86/ffitarget.h: Increase size of trampoline for stdcall
1090         closures.
1091         * src/x86/win32.S: Add assembly for stdcall closure.
1092         * src/x86/ffi.c: Initialize stdcall closure trampoline.
1093
1094 2009-06-04  Andrew Haley  <aph@redhat.com>
1095
1096         * include/ffi.h.in: Change void (*)() to void (*)(void).
1097         * src/x86/ffi.c: Likewise.
1098
1099 2009-06-04  Andrew Haley  <aph@redhat.com>
1100
1101         * src/powerpc/ppc_closure.S: Insert licence header.
1102         * src/powerpc/linux64_closure.S: Likewise.
1103         * src/m68k/sysv.S: Likewise.
1104
1105         * src/sh64/ffi.c: Change void (*)() to void (*)(void).
1106         * src/powerpc/ffi.c: Likewise.
1107         * src/powerpc/ffi_darwin.c: Likewise.
1108         * src/m32r/ffi.c: Likewise.
1109         * src/sh64/ffi.c: Likewise.
1110         * src/x86/ffi64.c: Likewise.
1111         * src/alpha/ffi.c: Likewise.
1112         * src/alpha/osf.S: Likewise.
1113         * src/frv/ffi.c: Likewise.
1114         * src/s390/ffi.c: Likewise.
1115         * src/pa/ffi.c: Likewise.
1116         * src/pa/hpux32.S: Likewise.
1117         * src/ia64/unix.S: Likewise.
1118         * src/ia64/ffi.c: Likewise.
1119         * src/sparc/ffi.c: Likewise.
1120         * src/mips/ffi.c: Likewise.
1121         * src/sh/ffi.c: Likewise.
1122
1123 2008-02-15  David Daney  <ddaney@avtrex.com>
1124
1125         * src/mips/ffi.c (USE__BUILTIN___CLEAR_CACHE):
1126         Define (conditionally), and use it to include cachectl.h.
1127         (ffi_prep_closure_loc): Fix cache flushing.
1128         * src/mips/ffitarget.h (_ABIN32, _ABI64, _ABIO32): Define.
1129
1130 2009-06-04  Andrew Haley  <aph@redhat.com>
1131
1132         include/ffi.h.in,
1133         src/arm/ffitarget.h,
1134         src/arm/ffi.c,
1135         src/arm/sysv.S,
1136         src/powerpc/ffitarget.h,
1137         src/closures.c,
1138         src/sh64/ffitarget.h,
1139         src/sh64/ffi.c,
1140         src/sh64/sysv.S,
1141         src/types.c,
1142         src/x86/ffi64.c,
1143         src/x86/ffitarget.h,
1144         src/x86/win32.S,
1145         src/x86/darwin.S,
1146         src/x86/ffi.c,
1147         src/x86/sysv.S,
1148         src/x86/unix64.S,
1149         src/alpha/ffitarget.h,
1150         src/alpha/ffi.c,
1151         src/alpha/osf.S,
1152         src/m68k/ffitarget.h,
1153         src/frv/ffitarget.h,
1154         src/frv/ffi.c,
1155         src/s390/ffitarget.h,
1156         src/s390/sysv.S,
1157         src/cris/ffitarget.h,
1158         src/pa/linux.S,
1159         src/pa/ffitarget.h,
1160         src/pa/ffi.c,
1161         src/raw_api.c,
1162         src/ia64/ffitarget.h,
1163         src/ia64/unix.S,
1164         src/ia64/ffi.c,
1165         src/ia64/ia64_flags.h,
1166         src/java_raw_api.c,
1167         src/debug.c,
1168         src/sparc/v9.S,
1169         src/sparc/ffitarget.h,
1170         src/sparc/ffi.c,
1171         src/sparc/v8.S,
1172         src/mips/ffitarget.h,
1173         src/mips/n32.S,
1174         src/mips/o32.S,
1175         src/mips/ffi.c,
1176         src/prep_cif.c,
1177         src/sh/ffitarget.h,
1178         src/sh/ffi.c,
1179         src/sh/sysv.S: Update license text.
1180
1181 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
1182
1183         * src/x86/win32.S (_ffi_closure_STDCALL):  New function.
1184         (.eh_frame):  Add FDE for it.
1185
1186 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
1187
1188         * configure.ac:  Also check if assembler supports pc-relative
1189         relocs on X86_WIN32 targets.
1190         * configure:  Regenerate.
1191         * src/x86/win32.S (ffi_prep_args):  Declare extern, not global.
1192         (_ffi_call_SYSV):  Add missing function type symbol .def and
1193         add EH markup labels.
1194         (_ffi_call_STDCALL):  Likewise.
1195         (_ffi_closure_SYSV):  Likewise.
1196         (_ffi_closure_raw_SYSV):  Likewise.
1197         (.eh_frame):  Add hand-crafted EH data.
1198
1199 2009-04-09  Jakub Jelinek  <jakub@redhat.com>
1200
1201         * testsuite/lib/libffi-dg.exp: Change copyright header to refer to
1202         version 3 of the GNU General Public License and to point readers
1203         at the COPYING3 file and the FSF's license web page.
1204         * testsuite/libffi.call/call.exp: Likewise.
1205         * testsuite/libffi.special/special.exp: Likewise.
1206
1207 2009-03-01  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1208
1209         * configure: Regenerate.
1210
1211 2008-12-18  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1212
1213         PR libffi/26048
1214         * configure.ac (HAVE_AS_X86_PCREL): New test.
1215         * configure: Regenerate.
1216         * fficonfig.h.in: Regenerate.
1217         * src/x86/sysv.S [!FFI_NO_RAW_API]: Precalculate
1218         RAW_CLOSURE_CIF_OFFSET, RAW_CLOSURE_FUN_OFFSET,
1219         RAW_CLOSURE_USER_DATA_OFFSET for the Solaris 10/x86 assembler.
1220         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
1221         * src/x86/unix64.S (.Lstore_table): Move to .text section.
1222         (.Lload_table): Likewise.
1223         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
1224
1225 2008-12-18  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1226
1227         * configure: Regenerate.
1228
1229 2008-11-21  Eric Botcazou  <ebotcazou@adacore.com>
1230
1231         * src/sparc/ffi.c (ffi_prep_cif_machdep): Add support for
1232         signed/unsigned int8/16 return values.
1233         * src/sparc/v8.S (ffi_call_v8): Likewise.
1234         (ffi_closure_v8): Likewise.
1235
1236 2008-09-26  Peter O'Gorman  <pogma@thewrittenword.com>
1237             Steve Ellcey  <sje@cup.hp.com>
1238
1239         * configure: Regenerate for new libtool.
1240         * Makefile.in: Ditto.
1241         * include/Makefile.in: Ditto.
1242         * aclocal.m4: Ditto.
1243
1244 2008-08-25  Andreas Tobler  <a.tobler@schweiz.org>
1245
1246         * src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and
1247         FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum.
1248         Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT.
1249         Adjust copyright notice.
1250         * src/powerpc/ffi.c: Add two new flags to indicate if we have one
1251         register or two register to use for FFI_SYSV structs.
1252         (ffi_prep_cif_machdep): Pass the right register flag introduced above.
1253         (ffi_closure_helper_SYSV): Fix the return type for
1254         FFI_SYSV_TYPE_SMALL_STRUCT. Comment.
1255         Adjust copyright notice.
1256
1257 2008-07-16  Kaz Kojima  <kkojima@gcc.gnu.org>
1258
1259         * src/sh/ffi.c (ffi_prep_closure_loc): Turn INSN into an unsigned
1260         int.
1261
1262 2008-06-17  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1263
1264         * configure: Regenerate.
1265         * include/Makefile.in: Regenerate.
1266         * testsuite/Makefile.in: Regenerate.
1267
1268 2008-06-07  Joseph Myers  <joseph@codesourcery.com>
1269
1270         * configure.ac (parisc*-*-linux*, powerpc-*-sysv*,
1271         powerpc-*-beos*): Remove.
1272         * configure: Regenerate.
1273
1274 2008-05-09  Julian Brown  <julian@codesourcery.com>
1275
1276         * Makefile.am (LTLDFLAGS): New.
1277         (libffi_la_LDFLAGS): Use above.
1278         * Makefile.in: Regenerate.
1279
1280 2008-04-18  Paolo Bonzini  <bonzini@gnu.org>
1281
1282         PR bootstrap/35457
1283         * aclocal.m4: Regenerate.
1284         * configure: Regenerate.
1285
1286 2008-03-26  Kaz Kojima  <kkojima@gcc.gnu.org>
1287
1288         * src/sh/sysv.S: Add .note.GNU-stack on Linux.
1289         * src/sh64/sysv.S: Likewise.
1290
1291 2008-03-26  Daniel Jacobowitz  <dan@debian.org>
1292
1293         * src/arm/sysv.S: Fix ARM comment marker.
1294
1295 2008-03-26  Jakub Jelinek  <jakub@redhat.com>
1296
1297         * src/alpha/osf.S: Add .note.GNU-stack on Linux.
1298         * src/s390/sysv.S: Likewise.
1299         * src/powerpc/ppc_closure.S: Likewise.
1300         * src/powerpc/sysv.S: Likewise.
1301         * src/x86/unix64.S: Likewise.
1302         * src/x86/sysv.S: Likewise.
1303         * src/sparc/v8.S: Likewise.
1304         * src/sparc/v9.S: Likewise.
1305         * src/m68k/sysv.S: Likewise.
1306         * src/arm/sysv.S: Likewise.
1307
1308 2008-03-16  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1309
1310         * aclocal.m4: Regenerate.
1311         * configure: Likewise.
1312         * Makefile.in: Likewise.
1313         * include/Makefile.in: Likewise.
1314         * testsuite/Makefile.in: Likewise.
1315
1316 2008-02-12  Bjoern Koenig  <bkoenig@alpha-tierchen.de>
1317             Andreas Tobler  <a.tobler@schweiz.org>
1318
1319         * configure.ac: Add amd64-*-freebsd* target.
1320         * configure: Regenerate.
1321
1322 2008-01-30  H.J. Lu  <hongjiu.lu@intel.com>
1323
1324         PR libffi/34612
1325         * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when
1326         returning struct.
1327
1328         * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer"
1329         tests.
1330
1331 2008-01-24  David Edelsohn  <edelsohn@gnu.org>
1332
1333         * configure: Regenerate.
1334
1335 2008-01-06  Andreas Tobler  <a.tobler@schweiz.org>
1336
1337         * src/x86/ffi.c (ffi_prep_cif_machdep): Fix thinko.
1338
1339 2008-01-05  Andreas Tobler  <a.tobler@schweiz.org>
1340
1341         PR testsuite/32843
1342         * src/x86/ffi.c (ffi_prep_cif_machdep): Add code for
1343         signed/unsigned int8/16 for X86_DARWIN.
1344         Updated copyright info.
1345         Handle one and two byte structs with special cif->flags.
1346         * src/x86/ffitarget.h: Add special types for one and two byte structs.
1347         Updated copyright info.
1348         * src/x86/darwin.S (ffi_call_SYSV): Rewrite to use a jump table like
1349         sysv.S
1350         Remove code to pop args from the stack after call.
1351         Special-case signed/unsigned for int8/16, one and two byte structs.
1352         (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1353         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1354         FFI_TYPE_SINT32.
1355         Updated copyright info.
1356
1357 2007-12-08  David Daney  <ddaney@avtrex.com>
1358
1359         * src/mips/n32.S (ffi_call_N32):  Replace dadd with ADDU, dsub with
1360         SUBU, add with ADDU and use smaller code sequences.
1361
1362 2007-12-07  David Daney  <ddaney@avtrex.com>
1363
1364         * src/mips/ffi.c (ffi_prep_cif_machdep): Handle long double return
1365         type.
1366
1367 2007-12-06  David Daney  <ddaney@avtrex.com>
1368
1369         * include/ffi.h.in (FFI_SIZEOF_JAVA_RAW): Define if not already
1370         defined.
1371         (ffi_java_raw): New typedef.
1372         (ffi_java_raw_call, ffi_java_ptrarray_to_raw,
1373         ffi_java_raw_to_ptrarray): Change parameter types from ffi_raw to
1374         ffi_java_raw.
1375         (ffi_java_raw_closure) : Same.
1376         (ffi_prep_java_raw_closure, ffi_prep_java_raw_closure_loc): Change
1377         parameter types.
1378         * src/java_raw_api.c (ffi_java_raw_size):  Replace FFI_SIZEOF_ARG with
1379         FFI_SIZEOF_JAVA_RAW.
1380         (ffi_java_raw_to_ptrarray): Change type of raw to ffi_java_raw.
1381         Replace FFI_SIZEOF_ARG with FFI_SIZEOF_JAVA_RAW. Use
1382         sizeof(ffi_java_raw) for alignment calculations.
1383         (ffi_java_ptrarray_to_raw): Same.
1384         (ffi_java_rvalue_to_raw): Add special handling for FFI_TYPE_POINTER
1385         if FFI_SIZEOF_JAVA_RAW == 4.
1386         (ffi_java_raw_to_rvalue): Same.
1387         (ffi_java_raw_call): Change type of raw to ffi_java_raw.
1388         (ffi_java_translate_args): Same.
1389         (ffi_prep_java_raw_closure_loc, ffi_prep_java_raw_closure): Change
1390         parameter types.
1391         * src/mips/ffitarget.h (FFI_SIZEOF_JAVA_RAW): Define for N32 ABI.
1392
1393 2007-12-06  David Daney  <ddaney@avtrex.com>
1394
1395         * src/mips/n32.S (ffi_closure_N32): Use 64-bit add instruction on
1396         pointer values.
1397
1398 2007-12-01  Andreas Tobler  <a.tobler@schweiz.org>
1399
1400         PR libffi/31937
1401         * src/powerpc/ffitarget.h: Introduce new ABI FFI_LINUX_SOFT_FLOAT.
1402         Add local FFI_TYPE_UINT128 to handle soft-float long-double-128.
1403         * src/powerpc/ffi.c: Distinguish between __NO_FPRS__ and not and
1404         set the NUM_FPR_ARG_REGISTERS according to.
1405         Add support for potential soft-float support under hard-float
1406         architecture.
1407         (ffi_prep_args_SYSV): Set NUM_FPR_ARG_REGISTERS to 0 in case of
1408         FFI_LINUX_SOFT_FLOAT, handle float, doubles and long-doubles according
1409         to the FFI_LINUX_SOFT_FLOAT ABI.
1410         (ffi_prep_cif_machdep): Likewise.
1411         (ffi_closure_helper_SYSV): Likewise.
1412         * src/powerpc/ppc_closure.S: Make sure not to store float/double
1413         on archs where __NO_FPRS__ is true.
1414         Add FFI_TYPE_UINT128 support.
1415         * src/powerpc/sysv.S: Add support for soft-float long-double-128.
1416         Adjust copyright notice.
1417
1418 2007-11-25  Andreas Tobler  <a.tobler@schweiz.org>
1419
1420         * src/closures.c: Move defintion of MAYBE_UNUSED from here to ...
1421         * include/ffi_common.h: ... here.
1422         Update copyright.
1423
1424 2007-11-17  Andreas Tobler  <a.tobler@schweiz.org>
1425
1426         * src/powerpc/sysv.S: Load correct cr to compare if we have long double.
1427         * src/powerpc/linux64.S: Likewise.
1428         * src/powerpc/ffi.c: Add a comment to show which part goes into cr6.
1429         * testsuite/libffi.call/return_ldl.c: New test.
1430
1431 2007-09-04    <aph@redhat.com>
1432
1433         * src/arm/sysv.S (UNWIND): New.
1434         (Whole file): Conditionally compile unwinder directives.
1435         * src/arm/sysv.S: Add unwinder directives.
1436
1437         * src/arm/ffi.c (ffi_prep_args): Align structs by at least 4 bytes.
1438         Only treat r0 as a struct address if we're actually returning a
1439         struct by address.
1440         Only copy the bytes that are actually within a struct.
1441         (ffi_prep_cif_machdep): A Composite Type not larger than 4 bytes
1442         is returned in r0, not passed by address.
1443         (ffi_call): Allocate a word-sized temporary for the case where
1444         a composite is returned in r0.
1445         (ffi_prep_incoming_args_SYSV): Align as necessary.
1446
1447 2007-08-05  Steven Newbury  <s_j_newbury@yahoo.co.uk>
1448
1449         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Use __clear_cache instead of
1450         directly using the sys_cacheflush syscall.
1451
1452 2007-07-27  Andrew Haley  <aph@redhat.com>
1453
1454         * src/arm/sysv.S (ffi_closure_SYSV): Add soft-float.
1455
1456 2007-09-03  Maciej W. Rozycki  <macro@linux-mips.org>
1457
1458         * Makefile.am: Unify MIPS_IRIX and MIPS_LINUX into MIPS.
1459         * configure.ac: Likewise.
1460         * Makefile.in: Regenerate.
1461         * include/Makefile.in: Likewise.
1462         * testsuite/Makefile.in: Likewise.
1463         * configure: Likewise.
1464
1465 2007-08-24  David Daney  <ddaney@avtrex.com>
1466
1467         * testsuite/libffi.call/return_sl.c: New test.
1468
1469 2007-08-10  David Daney  <ddaney@avtrex.com>
1470
1471         * testsuite/libffi.call/cls_multi_ushort.c,
1472         testsuite/libffi.call/cls_align_uint16.c,
1473         testsuite/libffi.call/nested_struct1.c,
1474         testsuite/libffi.call/nested_struct3.c,
1475         testsuite/libffi.call/cls_7_1_byte.c,
1476         testsuite/libffi.call/nested_struct5.c,
1477         testsuite/libffi.call/cls_double.c,
1478         testsuite/libffi.call/nested_struct7.c,
1479         testsuite/libffi.call/cls_sint.c,
1480         testsuite/libffi.call/nested_struct9.c,
1481         testsuite/libffi.call/cls_20byte1.c,
1482         testsuite/libffi.call/cls_multi_sshortchar.c,
1483         testsuite/libffi.call/cls_align_sint64.c,
1484         testsuite/libffi.call/cls_3byte2.c,
1485         testsuite/libffi.call/cls_multi_schar.c,
1486         testsuite/libffi.call/cls_multi_uchar.c,
1487         testsuite/libffi.call/cls_19byte.c,
1488         testsuite/libffi.call/cls_9byte1.c,
1489         testsuite/libffi.call/cls_align_float.c,
1490         testsuite/libffi.call/closure_fn1.c,
1491         testsuite/libffi.call/problem1.c,
1492         testsuite/libffi.call/closure_fn3.c,
1493         testsuite/libffi.call/cls_sshort.c,
1494         testsuite/libffi.call/closure_fn5.c,
1495         testsuite/libffi.call/cls_align_double.c,
1496         testsuite/libffi.call/nested_struct.c,
1497         testsuite/libffi.call/cls_2byte.c,
1498         testsuite/libffi.call/nested_struct10.c,
1499         testsuite/libffi.call/cls_4byte.c,
1500         testsuite/libffi.call/cls_6byte.c,
1501         testsuite/libffi.call/cls_8byte.c,
1502         testsuite/libffi.call/cls_multi_sshort.c,
1503         testsuite/libffi.call/cls_align_sint16.c,
1504         testsuite/libffi.call/cls_align_uint32.c,
1505         testsuite/libffi.call/cls_20byte.c,
1506         testsuite/libffi.call/cls_float.c,
1507         testsuite/libffi.call/nested_struct2.c,
1508         testsuite/libffi.call/cls_5_1_byte.c,
1509         testsuite/libffi.call/nested_struct4.c,
1510         testsuite/libffi.call/cls_24byte.c,
1511         testsuite/libffi.call/nested_struct6.c,
1512         testsuite/libffi.call/cls_64byte.c,
1513         testsuite/libffi.call/nested_struct8.c,
1514         testsuite/libffi.call/cls_uint.c,
1515         testsuite/libffi.call/cls_multi_ushortchar.c,
1516         testsuite/libffi.call/cls_schar.c,
1517         testsuite/libffi.call/cls_uchar.c,
1518         testsuite/libffi.call/cls_align_uint64.c,
1519         testsuite/libffi.call/cls_ulonglong.c,
1520         testsuite/libffi.call/cls_align_longdouble.c,
1521         testsuite/libffi.call/cls_1_1byte.c,
1522         testsuite/libffi.call/cls_12byte.c,
1523         testsuite/libffi.call/cls_3_1byte.c,
1524         testsuite/libffi.call/cls_3byte1.c,
1525         testsuite/libffi.call/cls_4_1byte.c,
1526         testsuite/libffi.call/cls_6_1_byte.c,
1527         testsuite/libffi.call/cls_16byte.c,
1528         testsuite/libffi.call/cls_18byte.c,
1529         testsuite/libffi.call/closure_fn0.c,
1530         testsuite/libffi.call/cls_9byte2.c,
1531         testsuite/libffi.call/closure_fn2.c,
1532         testsuite/libffi.call/closure_fn4.c,
1533         testsuite/libffi.call/cls_ushort.c,
1534         testsuite/libffi.call/closure_fn6.c,
1535         testsuite/libffi.call/cls_5byte.c,
1536         testsuite/libffi.call/cls_align_pointer.c,
1537         testsuite/libffi.call/cls_7byte.c,
1538         testsuite/libffi.call/cls_align_sint32.c,
1539         testsuite/libffi.special/unwindtest_ffi_call.cc,
1540         testsuite/libffi.special/unwindtest.cc: Remove xfail for mips64*-*-*.
1541
1542 2007-08-10  David Daney  <ddaney@avtrex.com>
1543
1544         PR libffi/28313
1545         * configure.ac: Don't treat mips64 as a special case.
1546         * Makefile.am (nodist_libffi_la_SOURCES): Add n32.S.
1547         * configure: Regenerate
1548         * Makefile.in: Ditto.
1549         * fficonfig.h.in: Ditto.
1550         * src/mips/ffitarget.h (REG_L, REG_S, SUBU, ADDU, SRL, LI): Indent.
1551         (LA, EH_FRAME_ALIGN, FDE_ADDR_BYTES): New preprocessor macros.
1552         (FFI_DEFAULT_ABI): Set for n64 case.
1553         (FFI_CLOSURES, FFI_TRAMPOLINE_SIZE): Define for n32 and n64 cases.
1554         * src/mips/n32.S (ffi_call_N32): Add debug macros and labels for FDE.
1555         (ffi_closure_N32): New function.
1556         (.eh_frame): New section
1557         * src/mips/o32.S: Clean up comments.
1558         (ffi_closure_O32): Pass ffi_closure parameter in $12.
1559         * src/mips/ffi.c: Use FFI_MIPS_N32 instead of
1560         _MIPS_SIM == _ABIN32 throughout.
1561         (FFI_MIPS_STOP_HERE): New, use in place of
1562         ffi_stop_here.
1563         (ffi_prep_args): Use unsigned long to hold pointer values.  Rewrite
1564         to support n32/n64 ABIs.
1565         (calc_n32_struct_flags): Rewrite.
1566         (calc_n32_return_struct_flags): Remove unused variable.  Reverse
1567         position of flag bits.
1568         (ffi_prep_cif_machdep): Rewrite n32 portion.
1569         (ffi_call): Enable for n64.  Add special handling for small structure
1570         return values.
1571         (ffi_prep_closure_loc): Add n32 and n64 support.
1572         (ffi_closure_mips_inner_O32): Add cast to silence warning.
1573         (copy_struct_N32, ffi_closure_mips_inner_N32): New functions.
1574
1575 2007-08-08  David Daney  <ddaney@avtrex.com>
1576
1577         * testsuite/libffi.call/ffitest.h (ffi_type_mylong): Remove definition.
1578         * testsuite/libffi.call/cls_align_uint16.c (main): Use correct type
1579         specifiers.
1580         * testsuite/libffi.call/nested_struct1.c (main): Ditto.
1581         * testsuite/libffi.call/cls_sint.c (main): Ditto.
1582         * testsuite/libffi.call/nested_struct9.c (main): Ditto.
1583         * testsuite/libffi.call/cls_20byte1.c (main): Ditto.
1584         * testsuite/libffi.call/cls_9byte1.c (main): Ditto.
1585         * testsuite/libffi.call/closure_fn1.c (main): Ditto.
1586         * testsuite/libffi.call/closure_fn3.c (main): Ditto.
1587         * testsuite/libffi.call/return_dbl2.c (main): Ditto.
1588         * testsuite/libffi.call/cls_sshort.c (main): Ditto.
1589         * testsuite/libffi.call/return_fl3.c (main): Ditto.
1590         * testsuite/libffi.call/closure_fn5.c (main): Ditto.
1591         * testsuite/libffi.call/nested_struct.c (main): Ditto.
1592         * testsuite/libffi.call/nested_struct10.c (main): Ditto.
1593         * testsuite/libffi.call/return_ll1.c (main): Ditto.
1594         * testsuite/libffi.call/cls_8byte.c (main): Ditto.
1595         * testsuite/libffi.call/cls_align_uint32.c (main): Ditto.
1596         * testsuite/libffi.call/cls_align_sint16.c (main): Ditto.
1597         * testsuite/libffi.call/cls_20byte.c (main): Ditto.
1598         * testsuite/libffi.call/nested_struct2.c (main): Ditto.
1599         * testsuite/libffi.call/cls_24byte.c (main): Ditto.
1600         * testsuite/libffi.call/nested_struct6.c (main): Ditto.
1601         * testsuite/libffi.call/cls_uint.c (main): Ditto.
1602         * testsuite/libffi.call/cls_12byte.c (main): Ditto.
1603         * testsuite/libffi.call/cls_16byte.c (main): Ditto.
1604         * testsuite/libffi.call/closure_fn0.c (main): Ditto.
1605         * testsuite/libffi.call/cls_9byte2.c (main): Ditto.
1606         * testsuite/libffi.call/closure_fn2.c (main): Ditto.
1607         * testsuite/libffi.call/return_dbl1.c (main): Ditto.
1608         * testsuite/libffi.call/closure_fn4.c (main): Ditto.
1609         * testsuite/libffi.call/closure_fn6.c (main): Ditto.
1610         * testsuite/libffi.call/cls_align_sint32.c (main): Ditto.
1611
1612 2007-08-07  Andrew Haley  <aph@redhat.com>
1613
1614         * src/x86/sysv.S (ffi_closure_raw_SYSV): Fix typo in previous
1615         checkin.
1616
1617 2007-08-06  Andrew Haley  <aph@redhat.com>
1618
1619         PR testsuite/32843
1620         * src/x86/sysv.S (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1621         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1622         FFI_TYPE_SINT32.
1623
1624 2007-08-02  David Daney  <ddaney@avtrex.com>
1625
1626         * testsuite/libffi.call/return_ul.c (main): Define return type as
1627         ffi_arg.  Use proper printf conversion specifier.
1628
1629 2007-07-30  Andrew Haley  <aph@redhat.com>
1630
1631         PR testsuite/32843
1632         * src/x86/ffi.c (ffi_prep_cif_machdep): in x86 case, add code for
1633         signed/unsigned int8/16.
1634         * src/x86/sysv.S (ffi_call_SYSV): Rewrite to:
1635         Use a jump table.
1636         Remove code to pop args from the stack after call.
1637         Special-case signed/unsigned int8/16.
1638         * testsuite/libffi.call/return_sc.c (main): Revert.
1639
1640 2007-07-26  Richard Guenther  <rguenther@suse.de>
1641
1642         PR testsuite/32843
1643         * testsuite/libffi.call/return_sc.c (main): Verify call
1644         result as signed char, not ffi_arg.
1645
1646 2007-07-16  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1647
1648         * configure.ac (i?86-*-solaris2.1[0-9]): Set TARGET to X86_64.
1649         * configure: Regenerate.
1650
1651 2007-07-11  David Daney  <ddaney@avtrex.com>
1652
1653         * src/mips/ffi.c: Don't include sys/cachectl.h.
1654         (ffi_prep_closure_loc): Use __builtin___clear_cache() instead of
1655         cacheflush().
1656
1657 2007-05-18  Aurelien Jarno  <aurelien@aurel32.net>
1658
1659         * src/arm/ffi.c (ffi_prep_closure_loc): Renamed and ajusted
1660         from (ffi_prep_closure): ... this.
1661         (FFI_INIT_TRAMPOLINE): Adjust.
1662
1663 2005-12-31  Phil Blundell  <pb@reciva.com>
1664
1665         * src/arm/ffi.c (ffi_prep_incoming_args_SYSV,
1666         ffi_closure_SYSV_inner, ffi_prep_closure): New, add closure support.
1667         * src/arm/sysv.S(ffi_closure_SYSV): Likewise.
1668         * src/arm/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1669         (FFI_CLOSURES): Enable closure support.
1670
1671 2007-07-03  Andrew Haley  <aph@hedges.billgatliff.com>
1672
1673         * testsuite/libffi.call/cls_multi_ushort.c,
1674         testsuite/libffi.call/cls_align_uint16.c,
1675         testsuite/libffi.call/nested_struct1.c,
1676         testsuite/libffi.call/nested_struct3.c,
1677         testsuite/libffi.call/cls_7_1_byte.c,
1678         testsuite/libffi.call/cls_double.c,
1679         testsuite/libffi.call/nested_struct5.c,
1680         testsuite/libffi.call/nested_struct7.c,
1681         testsuite/libffi.call/cls_sint.c,
1682         testsuite/libffi.call/nested_struct9.c,
1683         testsuite/libffi.call/cls_20byte1.c,
1684         testsuite/libffi.call/cls_multi_sshortchar.c,
1685         testsuite/libffi.call/cls_align_sint64.c,
1686         testsuite/libffi.call/cls_3byte2.c,
1687         testsuite/libffi.call/cls_multi_schar.c,
1688         testsuite/libffi.call/cls_multi_uchar.c,
1689         testsuite/libffi.call/cls_19byte.c,
1690         testsuite/libffi.call/cls_9byte1.c,
1691         testsuite/libffi.call/cls_align_float.c,
1692         testsuite/libffi.call/closure_fn1.c,
1693         testsuite/libffi.call/problem1.c,
1694         testsuite/libffi.call/closure_fn3.c,
1695         testsuite/libffi.call/cls_sshort.c,
1696         testsuite/libffi.call/closure_fn5.c,
1697         testsuite/libffi.call/cls_align_double.c,
1698         testsuite/libffi.call/cls_2byte.c,
1699         testsuite/libffi.call/nested_struct.c,
1700         testsuite/libffi.call/nested_struct10.c,
1701         testsuite/libffi.call/cls_4byte.c,
1702         testsuite/libffi.call/cls_6byte.c,
1703         testsuite/libffi.call/cls_8byte.c,
1704         testsuite/libffi.call/cls_multi_sshort.c,
1705         testsuite/libffi.call/cls_align_uint32.c,
1706         testsuite/libffi.call/cls_align_sint16.c,
1707         testsuite/libffi.call/cls_float.c,
1708         testsuite/libffi.call/cls_20byte.c,
1709         testsuite/libffi.call/cls_5_1_byte.c,
1710         testsuite/libffi.call/nested_struct2.c,
1711         testsuite/libffi.call/cls_24byte.c,
1712         testsuite/libffi.call/nested_struct4.c,
1713         testsuite/libffi.call/nested_struct6.c,
1714         testsuite/libffi.call/cls_64byte.c,
1715         testsuite/libffi.call/nested_struct8.c,
1716         testsuite/libffi.call/cls_uint.c,
1717         testsuite/libffi.call/cls_multi_ushortchar.c,
1718         testsuite/libffi.call/cls_schar.c,
1719         testsuite/libffi.call/cls_uchar.c,
1720         testsuite/libffi.call/cls_align_uint64.c,
1721         testsuite/libffi.call/cls_ulonglong.c,
1722         testsuite/libffi.call/cls_align_longdouble.c,
1723         testsuite/libffi.call/cls_1_1byte.c,
1724         testsuite/libffi.call/cls_12byte.c,
1725         testsuite/libffi.call/cls_3_1byte.c,
1726         testsuite/libffi.call/cls_3byte1.c,
1727         testsuite/libffi.call/cls_4_1byte.c,
1728         testsuite/libffi.call/cls_6_1_byte.c,
1729         testsuite/libffi.call/cls_16byte.c,
1730         testsuite/libffi.call/cls_18byte.c,
1731         testsuite/libffi.call/closure_fn0.c,
1732         testsuite/libffi.call/cls_9byte2.c,
1733         testsuite/libffi.call/closure_fn2.c,
1734         testsuite/libffi.call/closure_fn4.c,
1735         testsuite/libffi.call/cls_ushort.c,
1736         testsuite/libffi.call/closure_fn6.c,
1737         testsuite/libffi.call/cls_5byte.c,
1738         testsuite/libffi.call/cls_align_pointer.c,
1739         testsuite/libffi.call/cls_7byte.c,
1740         testsuite/libffi.call/cls_align_sint32.c,
1741         testsuite/libffi.special/unwindtest_ffi_call.cc,
1742         testsuite/libffi.special/unwindtest.cc: Enable for ARM.
1743
1744 2007-07-05  H.J. Lu  <hongjiu.lu@intel.com>
1745
1746         * aclocal.m4: Regenerated.
1747
1748 2007-06-02  Paolo Bonzini  <bonzini@gnu.org>
1749
1750         * configure: Regenerate.
1751
1752 2007-05-23  Steve Ellcey  <sje@cup.hp.com>
1753
1754         * Makefile.in: Regenerate.
1755         * configure: Regenerate.
1756         * aclocal.m4: Regenerate.
1757         * include/Makefile.in: Regenerate.
1758         * testsuite/Makefile.in: Regenerate.
1759
1760 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1761
1762         * src/m68k/ffi.c (ffi_prep_incoming_args_SYSV,
1763         ffi_closure_SYSV_inner,ffi_prep_closure): New, add closure support.
1764         * src/m68k/sysv.S(ffi_closure_SYSV,ffi_closure_struct_SYSV): Likewise.
1765         * src/m68k/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1766         (FFI_CLOSURES): Enable closure support.
1767
1768 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1769
1770         * configure.ac (HAVE_AS_CFI_PSEUDO_OP): New test.
1771         * configure: Regenerate.
1772         * fficonfig.h.in: Regenerate.
1773         * src/m68k/sysv.S (CFI_STARTPROC,CFI_ENDPROC,
1774         CFI_OFFSET,CFI_DEF_CFA): New macros.
1775         (ffi_call_SYSV): Add callframe annotation.
1776
1777 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1778
1779         * src/m68k/ffi.c (ffi_prep_args,ffi_prep_cif_machdep): Fix
1780         numerous test suite failures.
1781         * src/m68k/sysv.S (ffi_call_SYSV): Likewise.
1782
1783 2007-04-11  Paolo Bonzini  <bonzini@gnu.org>
1784
1785         * Makefile.am (EXTRA_DIST): Bring up to date.
1786         * Makefile.in: Regenerate.
1787         * src/frv/eabi.S: Remove RCS keyword.
1788
1789 2007-04-06  Richard Henderson  <rth@redhat.com>
1790
1791         * configure.ac: Tidy target case.
1792         (HAVE_LONG_DOUBLE): Allow the target to override.
1793         * configure: Regenerate.
1794         * include/ffi.h.in: Don't define ffi_type_foo if
1795         LIBFFI_HIDE_BASIC_TYPES is defined.
1796         (ffi_type_longdouble): If not HAVE_LONG_DOUBLE, define
1797         to ffi_type_double.
1798         * types.c (LIBFFI_HIDE_BASIC_TYPES): Define.
1799         (FFI_TYPEDEF, ffi_type_void): Mark the data const.
1800         (ffi_type_longdouble): Special case for Alpha.  Don't define
1801         if long double == double.
1802
1803         * src/alpha/ffi.c (FFI_TYPE_LONGDOUBLE): Assert unique value.
1804         (ffi_prep_cif_machdep): Handle it as the 128-bit type.
1805         (ffi_call, ffi_closure_osf_inner): Likewise.
1806         (ffi_closure_osf_inner): Likewise.  Mark hidden.
1807         (ffi_call_osf, ffi_closure_osf): Mark hidden.
1808         * src/alpha/ffitarget.h (FFI_LAST_ABI): Tidy definition.
1809         * src/alpha/osf.S (ffi_call_osf, ffi_closure_osf): Mark hidden.
1810         (load_table): Handle 128-bit long double.
1811
1812         * testsuite/libffi.call/float4.c: Add -mieee for alpha.
1813
1814 2007-04-06  Tom Tromey  <tromey@redhat.com>
1815
1816         PR libffi/31491:
1817         * README: Fixed bug in example.
1818
1819 2007-04-03  Jakub Jelinek  <jakub@redhat.com>
1820
1821         * src/closures.c: Include sys/statfs.h.
1822         (_GNU_SOURCE): Define on Linux.
1823         (FFI_MMAP_EXEC_SELINUX): Define.
1824         (selinux_enabled): New variable.
1825         (selinux_enabled_check): New function.
1826         (is_selinux_enabled): Define.
1827         (dlmmap): Use it.
1828
1829 2007-03-24  Uros Bizjak  <ubizjak@gmail.com>
1830
1831         * testsuite/libffi.call/return_fl2.c (return_fl): Mark as static.
1832         Use 'volatile float sum' to create sum of floats to avoid false
1833         negative due to excess precision on ix86 targets.
1834         (main): Ditto.
1835
1836 2007-03-08  Alexandre Oliva  <aoliva@redhat.com>
1837
1838         * src/powerpc/ffi.c (flush_icache): Fix left-over from previous
1839         patch.
1840         (ffi_prep_closure_loc): Remove unneeded casts.  Add needed ones.
1841
1842 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1843
1844         * include/ffi.h.in (ffi_closure_alloc, ffi_closure_free): New.
1845         (ffi_prep_closure_loc): New.
1846         (ffi_prep_raw_closure_loc): New.
1847         (ffi_prep_java_raw_closure_loc): New.
1848         * src/closures.c: New file.
1849         * src/dlmalloc.c [FFI_MMAP_EXEC_WRIT] (struct malloc_segment):
1850         Replace sflags with exec_offset.
1851         [FFI_MMAP_EXEC_WRIT] (mmap_exec_offset, add_segment_exec_offset,
1852         sub_segment_exec_offset): New macros.
1853         (get_segment_flags, set_segment_flags, check_segment_merge): New
1854         macros.
1855         (is_mmapped_segment, is_extern_segment): Use get_segment_flags.
1856         (add_segment, sys_alloc, create_mspace, create_mspace_with_base,
1857         destroy_mspace): Use new macros.
1858         (sys_alloc): Silence warning.
1859         * Makefile.am (libffi_la_SOURCES): Add src/closures.c.
1860         * Makefile.in: Rebuilt.
1861         * src/prep_cif [FFI_CLOSURES] (ffi_prep_closure): Implement in
1862         terms of ffi_prep_closure_loc.
1863         * src/raw_api.c (ffi_prep_raw_closure_loc): Renamed and adjusted
1864         from...
1865         (ffi_prep_raw_closure): ... this.  Re-implement in terms of the
1866         renamed version.
1867         * src/java_raw_api (ffi_prep_java_raw_closure_loc): Renamed and
1868         adjusted from...
1869         (ffi_prep_java_raw_closure): ... this.  Re-implement in terms of
1870         the renamed version.
1871         * src/alpha/ffi.c (ffi_prep_closure_loc): Renamed from
1872         (ffi_prep_closure): ... this.
1873         * src/pa/ffi.c: Likewise.
1874         * src/cris/ffi.c: Likewise.  Adjust.
1875         * src/frv/ffi.c: Likewise.
1876         * src/ia64/ffi.c: Likewise.
1877         * src/mips/ffi.c: Likewise.
1878         * src/powerpc/ffi_darwin.c: Likewise.
1879         * src/s390/ffi.c: Likewise.
1880         * src/sh/ffi.c: Likewise.
1881         * src/sh64/ffi.c: Likewise.
1882         * src/sparc/ffi.c: Likewise.
1883         * src/x86/ffi64.c: Likewise.
1884         * src/x86/ffi.c: Likewise.
1885         (FFI_INIT_TRAMPOLINE): Adjust.
1886         (ffi_prep_raw_closure_loc): Renamed and adjusted from...
1887         (ffi_prep_raw_closure): ... this.
1888         * src/powerpc/ffi.c (ffi_prep_closure_loc): Renamed from
1889         (ffi_prep_closure): ... this.
1890         (flush_icache): Adjust.
1891
1892 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1893
1894         * src/dlmalloc.c: New file, imported version 2.8.3 of Doug
1895         Lea's malloc.
1896
1897 2007-03-01  Brooks Moses  <brooks.moses@codesourcery.com>
1898
1899         * Makefile.am: Add dummy install-pdf target.
1900         * Makefile.in: Regenerate
1901
1902 2007-02-13  Andreas Krebbel  <krebbel1@de.ibm.com>
1903
1904         * src/s390/ffi.c (ffi_prep_args, ffi_prep_cif_machdep,
1905         ffi_closure_helper_SYSV): Add long double handling.
1906
1907 2007-02-02  Jakub Jelinek  <jakub@redhat.com>
1908
1909         * src/powerpc/linux64.S (ffi_call_LINUX64): Move restore of r2
1910         immediately after bctrl instruction.
1911
1912 2007-01-18  Alexandre Oliva  <aoliva@redhat.com>
1913
1914         * Makefile.am (all-recursive, install-recursive,
1915         mostlyclean-recursive, clean-recursive, distclean-recursive,
1916         maintainer-clean-recursive): Add missing targets.
1917         * Makefile.in: Rebuilt.
1918
1919 2006-12-14  Andreas Tobler  <a.tobler@schweiz.org>
1920
1921         * configure.ac: Add TARGET for x86_64-*-darwin*.
1922         * Makefile.am (nodist_libffi_la_SOURCES): Add rules for 64-bit sources
1923         for X86_DARWIN.
1924         * src/x86/ffitarget.h: Set trampoline size for x86_64-*-darwin*.
1925         * src/x86/darwin64.S: New file for x86_64-*-darwin* support.
1926         * configure: Regenerate.
1927         * Makefile.in: Regenerate.
1928         * include/Makefile.in: Regenerate.
1929         * testsuite/Makefile.in: Regenerate.
1930         * testsuite/libffi.special/unwindtest_ffi_call.cc: New test case for
1931         ffi_call only.
1932
1933 2006-12-13  Andreas Tobler <a.tobler@schweiz.org>
1934
1935         * aclocal.m4: Regenerate with aclocal -I .. as written in the
1936         Makefile.am.
1937
1938 2006-10-31  Geoffrey Keating  <geoffk@apple.com>
1939
1940         * src/powerpc/ffi_darwin.c (darwin_adjust_aggregate_sizes): New.
1941         (ffi_prep_cif_machdep): Call darwin_adjust_aggregate_sizes for
1942         Darwin.
1943         * testsuite/libffi.call/nested_struct4.c: Remove Darwin XFAIL.
1944         * testsuite/libffi.call/nested_struct6.c: Remove Darwin XFAIL.
1945
1946 2006-10-10  Paolo Bonzini  <bonzini@gnu.org>
1947             Sandro Tolaini  <tolaini@libero.it>
1948
1949         * configure.ac [i*86-*-darwin*]: Set X86_DARWIN symbol and
1950         conditional.
1951         * configure: Regenerated.
1952         * Makefile.am (nodist_libffi_la_SOURCES) [X86_DARWIN]: New case.
1953         (EXTRA_DIST): Add src/x86/darwin.S.
1954         * Makefile.in: Regenerated.
1955         * include/Makefile.in: Regenerated.
1956         * testsuite/Makefile.in: Regenerated.
1957
1958         * src/x86/ffi.c (ffi_prep_cif_machdep) [X86_DARWIN]: Treat like
1959         X86_WIN32, and additionally align stack to 16 bytes.
1960         * src/x86/darwin.S: New, based on sysv.S.
1961         * src/prep_cif.c (ffi_prep_cif) [X86_DARWIN]: Align > 8-byte structs.
1962
1963 2006-09-12  David Daney  <ddaney@avtrex.com>
1964
1965         PR libffi/23935
1966         * include/Makefile.am: Install both ffi.h and ffitarget.h in
1967         $(libdir)/gcc/$(target_alias)/$(gcc_version)/include.
1968         * aclocal.m4: Regenerated for automake 1.9.6.
1969         * Makefile.in: Regenerated.
1970         * include/Makefile.in: Regenerated.
1971         * testsuite/Makefile.in: Regenerated.
1972
1973 2006-08-17  Andreas Tobler  <a.tobler@schweiz.ch>
1974
1975         * include/ffi_common.h (struct): Revert accidental commit.
1976
1977 2006-08-15  Andreas Tobler  <a.tobler@schweiz.ch>
1978
1979         * include/ffi_common.h: Remove lint directives.
1980         * include/ffi.h.in: Likewise.
1981
1982 2006-07-25  Torsten Schoenfeld  <kaffeetisch@gmx.de>
1983
1984         * include/ffi.h.in (ffi_type_ulong, ffi_type_slong): Define correctly
1985         for 32-bit architectures.
1986         * testsuite/libffi.call/return_ul.c: New test case.
1987
1988 2006-07-19  David Daney  <ddaney@avtrex.com>
1989
1990         * testsuite/libffi.call/closure_fn6.c: Remove xfail for mips,
1991         xfail remains for mips64.
1992
1993 2006-05-23  Carlos O'Donell  <carlos@codesourcery.com>
1994
1995         * Makefile.am: Add install-html target. Add install-html to .PHONY
1996         * Makefile.in: Regenerate.
1997         * aclocal.m4: Regenerate.
1998         * include/Makefile.in: Regenerate.
1999         * testsuite/Makefile.in: Regenerate.
2000
2001 2006-05-18  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2002
2003         * pa/ffi.c (ffi_prep_args_pa32): Load floating point arguments from
2004         stack slot.
2005
2006 2006-04-22  Andreas Tobler  <a.tobler@schweiz.ch>
2007
2008         * README: Remove notice about 'Crazy Comments'.
2009         * src/debug.c: Remove lint directives. Cleanup white spaces.
2010         * src/java_raw_api.c: Likewise.
2011         * src/prep_cif.c: Likewise.
2012         * src/raw_api.c: Likewise.
2013         * src/ffitest.c: Delete. No longer needed, all test cases migrated
2014         to the testsuite.
2015         * src/arm/ffi.c: Remove lint directives.
2016         * src/m32r/ffi.c: Likewise.
2017         * src/pa/ffi.c: Likewise.
2018         * src/powerpc/ffi.c: Likewise.
2019         * src/powerpc/ffi_darwin.c: Likewise.
2020         * src/sh/ffi.c: Likewise.
2021         * src/sh64/ffi.c: Likewise.
2022         * src/x86/ffi.c: Likewise.
2023         * testsuite/libffi.call/float2.c: Likewise.
2024         * testsuite/libffi.call/promotion.c: Likewise.
2025         * testsuite/libffi.call/struct1.c: Likewise.
2026
2027 2006-04-13  Andreas Tobler  <a.tobler@schweiz.ch>
2028
2029         * src/pa/hpux32.S: Correct unwind offset calculation for
2030         ffi_closure_pa32.
2031         * src/pa/linux.S: Likewise.
2032
2033 2006-04-12  James E Wilson  <wilson@specifix.com>
2034
2035         PR libgcj/26483
2036         * src/ia64/ffi.c (stf_spill, ldf_fill): Rewrite as macros.
2037         (hfa_type_load): Call stf_spill.
2038         (hfa_type_store): Call ldf_fill.
2039         (ffi_call): Adjust calls to above routines.  Add local temps for
2040         macro result.
2041
2042 2006-04-10  Matthias Klose  <doko@debian.org>
2043
2044         * testsuite/lib/libffi-dg.exp (libffi-init): Recognize multilib
2045         directory names containing underscores.
2046
2047 2006-04-07  James E Wilson  <wilson@specifix.com>
2048
2049         * testsuite/libffi.call/float4.c: New testcase.
2050
2051 2006-04-05  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2052             Andreas Tobler  <a.tobler@schweiz.ch>
2053
2054         * Makefile.am: Add PA_HPUX port.
2055         * Makefile.in: Regenerate.
2056         * include/Makefile.in: Likewise.
2057         * testsuite/Makefile.in: Likewise.
2058         * configure.ac: Add PA_HPUX rules.
2059         * configure: Regenerate.
2060         * src/pa/ffitarget.h: Rename linux target to PA_LINUX.
2061         Add PA_HPUX and PA64_HPUX.
2062         Rename FFI_LINUX ABI to FFI_PA32 ABI.
2063         (FFI_TRAMPOLINE_SIZE): Define for 32-bit HP-UX targets.
2064         (FFI_TYPE_SMALL_STRUCT2): Define.
2065         (FFI_TYPE_SMALL_STRUCT4): Likewise.
2066         (FFI_TYPE_SMALL_STRUCT8): Likewise.
2067         (FFI_TYPE_SMALL_STRUCT3): Redefine.
2068         (FFI_TYPE_SMALL_STRUCT5): Likewise.
2069         (FFI_TYPE_SMALL_STRUCT6): Likewise.
2070         (FFI_TYPE_SMALL_STRUCT7): Likewise.
2071         * src/pa/ffi.c (ROUND_DOWN): Delete.
2072         (fldw, fstw, fldd, fstd): Use '__asm__'.
2073         (ffi_struct_type): Add support for FFI_TYPE_SMALL_STRUCT2,
2074         FFI_TYPE_SMALL_STRUCT4 and FFI_TYPE_SMALL_STRUCT8.
2075         (ffi_prep_args_LINUX): Rename to ffi_prep_args_pa32. Update comment.
2076         Simplify incrementing of stack slot variable. Change type of local
2077         'n' to unsigned int.
2078         (ffi_size_stack_LINUX): Rename to ffi_size_stack_pa32. Handle long
2079         double on PA_HPUX.
2080         (ffi_prep_cif_machdep): Likewise.
2081         (ffi_call): Likewise.
2082         (ffi_closure_inner_LINUX): Rename to ffi_closure_inner_pa32. Change
2083         return type to ffi_status. Simplify incrementing of stack slot
2084         variable. Only copy floating point argument registers when PA_LINUX
2085         is true. Reformat debug statement.
2086         Add support for FFI_TYPE_SMALL_STRUCT2, FFI_TYPE_SMALL_STRUCT4 and
2087         FFI_TYPE_SMALL_STRUCT8.
2088         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Add 'extern' to
2089         declaration.
2090         (ffi_prep_closure): Make linux trampoline conditional on PA_LINUX.
2091         Add nops to cache flush.  Add trampoline for PA_HPUX.
2092         * src/pa/hpux32.S: New file.
2093         * src/pa/linux.S (ffi_call_LINUX): Rename to ffi_call_pa32. Rename
2094         ffi_prep_args_LINUX to ffi_prep_args_pa32.
2095         Localize labels. Add support for 2, 4 and 8-byte small structs. Handle
2096         unaligned destinations in 3, 5, 6 and 7-byte small structs. Order
2097         argument type checks so that common argument types appear first.
2098         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Rename
2099         ffi_closure_inner_LINUX to ffi_closure_inner_pa32.
2100
2101 2006-03-24  Alan Modra  <amodra@bigpond.net.au>
2102
2103         * src/powerpc/ffitarget.h (enum ffi_abi): Add FFI_LINUX.  Default
2104         for 32-bit using IBM extended double format.  Fix FFI_LAST_ABI.
2105         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Handle linux variant of
2106         FFI_TYPE_LONGDOUBLE.
2107         (ffi_prep_args64): Assert using IBM extended double.
2108         (ffi_prep_cif_machdep): Don't munge FFI_TYPE_LONGDOUBLE type.
2109         Handle FFI_LINUX FFI_TYPE_LONGDOUBLE return and args.
2110         (ffi_call): Handle FFI_LINUX.
2111         (ffi_closure_helper_SYSV): Non FFI_LINUX long double return needs
2112         gpr3 return pointer as for struct return.  Handle FFI_LINUX
2113         FFI_TYPE_LONGDOUBLE return and args.  Don't increment "nf"
2114         unnecessarily.
2115         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Load both f1 and f2
2116         for FFI_TYPE_LONGDOUBLE.  Move epilogue insns into case table.
2117         Don't use r6 as pointer to results, instead use sp offset.  Don't
2118         make a special call to load lr with case table address, instead
2119         use offset from previous call.
2120         * src/powerpc/sysv.S (ffi_call_SYSV): Save long double return.
2121         * src/powerpc/linux64.S (ffi_call_LINUX64): Simplify long double
2122         return.
2123
2124 2006-03-15  Kaz Kojima  <kkojima@gcc.gnu.org>
2125
2126         * src/sh64/ffi.c (ffi_prep_cif_machdep): Handle float arguments
2127         passed with FP registers correctly.
2128         (ffi_closure_helper_SYSV): Likewise.
2129         * src/sh64/sysv.S: Likewise.
2130
2131 2006-03-01  Andreas Tobler  <a.tobler@schweiz.ch>
2132
2133         * testsuite/libffi.special/unwindtest.cc (closure_test_fn): Mark cif,
2134         args and userdata unused.
2135         (closure_test_fn1): Mark cif and userdata unused.
2136         (main): Remove unused res.
2137
2138 2006-02-28  Andreas Tobler  <a.tobler@schweiz.ch>
2139
2140         * testsuite/libffi.call/call.exp: Adjust FSF address. Add test runs for
2141         -O2, -O3, -Os and the warning flags -W -Wall.
2142         * testsuite/libffi.special/special.exp: Likewise.
2143         * testsuite/libffi.call/ffitest.h: Add an __UNUSED__ macro to mark
2144         unused parameter unused for gcc or else do nothing.
2145         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2146         * testsuite/libffi.call/cls_12byte.c (cls_struct_12byte_gn): Mark cif
2147         and userdata unused.
2148         * testsuite/libffi.call/cls_16byte.c (cls_struct_16byte_gn): Likewise.
2149         * testsuite/libffi.call/cls_18byte.c (cls_struct_18byte_gn): Likewise.
2150         * testsuite/libffi.call/cls_19byte.c (cls_struct_19byte_gn): Likewise.
2151         * testsuite/libffi.call/cls_1_1byte.c (cls_struct_1_1byte_gn): Likewise.
2152         * testsuite/libffi.call/cls_20byte.c (cls_struct_20byte_gn): Likewise.
2153         * testsuite/libffi.call/cls_20byte1.c (cls_struct_20byte_gn): Likewise.
2154         * testsuite/libffi.call/cls_24byte.c (cls_struct_24byte_gn): Likewise.
2155         * testsuite/libffi.call/cls_2byte.c (cls_struct_2byte_gn): Likewise.
2156         * testsuite/libffi.call/cls_3_1byte.c (cls_struct_3_1byte_gn): Likewise.
2157         * testsuite/libffi.call/cls_3byte1.c (cls_struct_3byte_gn): Likewise.
2158         * testsuite/libffi.call/cls_3byte2.c (cls_struct_3byte_gn1): Likewise.
2159         * testsuite/libffi.call/cls_4_1byte.c (cls_struct_4_1byte_gn): Likewise.
2160         * testsuite/libffi.call/cls_4byte.c (cls_struct_4byte_gn): Likewise.
2161         * testsuite/libffi.call/cls_5_1_byte.c (cls_struct_5byte_gn): Likewise.
2162         * testsuite/libffi.call/cls_5byte.c (cls_struct_5byte_gn): Likewise.
2163         * testsuite/libffi.call/cls_64byte.c (cls_struct_64byte_gn): Likewise.
2164         * testsuite/libffi.call/cls_6_1_byte.c (cls_struct_6byte_gn): Likewise.
2165         * testsuite/libffi.call/cls_6byte.c (cls_struct_6byte_gn): Likewise.
2166         * testsuite/libffi.call/cls_7_1_byte.c (cls_struct_7byte_gn): Likewise.
2167         * testsuite/libffi.call/cls_7byte.c (cls_struct_7byte_gn): Likewise.
2168         * testsuite/libffi.call/cls_8byte.c (cls_struct_8byte_gn): Likewise.
2169         * testsuite/libffi.call/cls_9byte1.c (cls_struct_9byte_gn): Likewise.
2170         * testsuite/libffi.call/cls_9byte2.c (cls_struct_9byte_gn): Likewise.
2171         * testsuite/libffi.call/cls_align_double.c (cls_struct_align_gn):
2172         Likewise.
2173         * testsuite/libffi.call/cls_align_float.c (cls_struct_align_gn):
2174         Likewise.
2175         * testsuite/libffi.call/cls_align_longdouble.c (cls_struct_align_gn):
2176         Likewise.
2177         * testsuite/libffi.call/cls_align_pointer.c (cls_struct_align_fn): Cast
2178         void* to avoid compiler warning.
2179         (main): Likewise.
2180         (cls_struct_align_gn): Mark cif and userdata unused.
2181         * testsuite/libffi.call/cls_align_sint16.c (cls_struct_align_gn):
2182         Likewise.
2183         * testsuite/libffi.call/cls_align_sint32.c (cls_struct_align_gn):
2184         Likewise.
2185         * testsuite/libffi.call/cls_align_sint64.c (cls_struct_align_gn):
2186         Likewise.
2187         * testsuite/libffi.call/cls_align_uint16.c (cls_struct_align_gn):
2188         Likewise.
2189         * testsuite/libffi.call/cls_align_uint32.c (cls_struct_align_gn):
2190         Likewise.
2191         * testsuite/libffi.call/cls_double.c (cls_ret_double_fn): Likewise.
2192         * testsuite/libffi.call/cls_float.c (cls_ret_float_fn): Likewise.
2193         * testsuite/libffi.call/cls_multi_schar.c (test_func_gn): Mark cif and
2194         data unused.
2195         (main): Cast res_call to silence gcc.
2196         * testsuite/libffi.call/cls_multi_sshort.c (test_func_gn): Mark cif and
2197         data unused.
2198         (main): Cast res_call to silence gcc.
2199         * testsuite/libffi.call/cls_multi_sshortchar.c (test_func_gn): Mark cif
2200         and data unused.
2201         (main): Cast res_call to silence gcc.
2202         * testsuite/libffi.call/cls_multi_uchar.c (test_func_gn): Mark cif and
2203         data unused.
2204         (main): Cast res_call to silence gcc.
2205         * testsuite/libffi.call/cls_multi_ushort.c (test_func_gn): Mark cif and
2206         data unused.
2207         (main): Cast res_call to silence gcc.
2208         * testsuite/libffi.call/cls_multi_ushortchar.c (test_func_gn): Mark cif
2209         and data unused.
2210         (main): Cast res_call to silence gcc.
2211         * testsuite/libffi.call/cls_schar.c (cls_ret_schar_fn): Mark cif and
2212         userdata unused.
2213         (cls_ret_schar_fn): Cast printf parameter to silence gcc.
2214         * testsuite/libffi.call/cls_sint.c (cls_ret_sint_fn): Mark cif and
2215         userdata unused.
2216         (cls_ret_sint_fn): Cast printf parameter to silence gcc.
2217         * testsuite/libffi.call/cls_sshort.c (cls_ret_sshort_fn): Mark cif and
2218         userdata unused.
2219         (cls_ret_sshort_fn): Cast printf parameter to silence gcc.
2220         * testsuite/libffi.call/cls_uchar.c (cls_ret_uchar_fn):  Mark cif and
2221         userdata unused.
2222         (cls_ret_uchar_fn): Cast printf parameter to silence gcc.
2223         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Mark cif and
2224         userdata unused.
2225         (cls_ret_uint_fn): Cast printf parameter to silence gcc.
2226         * testsuite/libffi.call/cls_ulonglong.c (cls_ret_ulonglong_fn): Mark cif
2227         and userdata unused.
2228         * testsuite/libffi.call/cls_ushort.c (cls_ret_ushort_fn): Mark cif and
2229         userdata unused.
2230         (cls_ret_ushort_fn): Cast printf parameter to silence gcc.
2231         * testsuite/libffi.call/float.c (floating): Remove unused parameter e.
2232         * testsuite/libffi.call/float1.c (main): Remove unused variable i.
2233         Cleanup white spaces.
2234         * testsuite/libffi.call/negint.c (checking): Remove unused variable i.
2235         * testsuite/libffi.call/nested_struct.c (cls_struct_combined_gn): Mark
2236         cif and userdata unused.
2237         * testsuite/libffi.call/nested_struct1.c (cls_struct_combined_gn):
2238         Likewise.
2239         * testsuite/libffi.call/nested_struct10.c (B_gn): Likewise.
2240         * testsuite/libffi.call/nested_struct2.c (B_fn): Adjust printf
2241         formatters to silence gcc.
2242         (B_gn): Mark cif and userdata unused.
2243         * testsuite/libffi.call/nested_struct3.c (B_gn): Mark cif and userdata
2244         unused.
2245         * testsuite/libffi.call/nested_struct4.c: Mention related PR.
2246         (B_gn): Mark cif and userdata unused.
2247         * testsuite/libffi.call/nested_struct5.c (B_gn): Mark cif and userdata
2248         unused.
2249         * testsuite/libffi.call/nested_struct6.c: Mention related PR.
2250         (B_gn): Mark cif and userdata unused.
2251         * testsuite/libffi.call/nested_struct7.c (B_gn): Mark cif and userdata
2252         unused.
2253         * testsuite/libffi.call/nested_struct8.c (B_gn): Likewise.
2254         * testsuite/libffi.call/nested_struct9.c (B_gn): Likewise.
2255         * testsuite/libffi.call/problem1.c (stub): Likewise.
2256         * testsuite/libffi.call/pyobjc-tc.c (main): Cast the result to silence
2257         gcc.
2258         * testsuite/libffi.call/return_fl2.c (return_fl): Add the note mentioned
2259         in the last commit for this test case in the test case itself.
2260         * testsuite/libffi.call/closure_fn0.c (closure_test_fn0): Mark cif as
2261         unused.
2262         * testsuite/libffi.call/closure_fn1.c (closure_test_fn1): Likewise.
2263         * testsuite/libffi.call/closure_fn2.c (closure_test_fn2): Likewise.
2264         * testsuite/libffi.call/closure_fn3.c (closure_test_fn3): Likewise.
2265         * testsuite/libffi.call/closure_fn4.c (closure_test_fn0): Likewise.
2266         * testsuite/libffi.call/closure_fn5.c (closure_test_fn5): Likewise.
2267         * testsuite/libffi.call/closure_fn6.c (closure_test_fn0): Likewise.
2268
2269 2006-02-22  Kaz Kojima  <kkojima@gcc.gnu.org>
2270
2271         * src/sh/sysv.S: Fix register numbers in the FDE for
2272         ffi_closure_SYSV.
2273
2274 2006-02-20  Andreas Tobler  <a.tobler@schweiz.ch>
2275
2276         * testsuite/libffi.call/return_fl2.c (return_fl): Remove static
2277         declaration to avoid a false negative on ix86. See PR323.
2278
2279 2006-02-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2280
2281         * src/sh/ffi.c (ffi_closure_helper_SYSV): Remove unused variable
2282         and cast integer to void * if needed.  Update the pointer to
2283         the FP register saved area correctly.
2284
2285 2006-02-17  Andreas Tobler  <a.tobler@schweiz.ch>
2286
2287         * testsuite/libffi.call/nested_struct6.c: XFAIL this test until PR25630
2288         is fixed.
2289         * testsuite/libffi.call/nested_struct4.c: Likewise.
2290
2291 2006-02-16  Andreas Tobler  <a.tobler@schweiz.ch>
2292
2293         * testsuite/libffi.call/return_dbl.c: New test case.
2294         * testsuite/libffi.call/return_dbl1.c: Likewise.
2295         * testsuite/libffi.call/return_dbl2.c: Likewise.
2296         * testsuite/libffi.call/return_fl.c: Likewise.
2297         * testsuite/libffi.call/return_fl1.c: Likewise.
2298         * testsuite/libffi.call/return_fl2.c: Likewise.
2299         * testsuite/libffi.call/return_fl3.c: Likewise.
2300         * testsuite/libffi.call/closure_fn6.c: Likewise.
2301
2302         * testsuite/libffi.call/nested_struct2.c: Remove ffi_type_mylong
2303         definition.
2304         * testsuite/libffi.call/ffitest.h: Add ffi_type_mylong definition
2305         here to be used by other test cases too.
2306
2307         * testsuite/libffi.call/nested_struct10.c: New test case.
2308         * testsuite/libffi.call/nested_struct9.c: Likewise.
2309         * testsuite/libffi.call/nested_struct8.c: Likewise.
2310         * testsuite/libffi.call/nested_struct7.c: Likewise.
2311         * testsuite/libffi.call/nested_struct6.c: Likewise.
2312         * testsuite/libffi.call/nested_struct5.c: Likewise.
2313         * testsuite/libffi.call/nested_struct4.c: Likewise.
2314
2315 2006-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
2316
2317         * configure.ac: Enable libffi for sparc64-*-freebsd*.
2318         * configure: Rebuilt.
2319
2320 2006-01-18  Jakub Jelinek  <jakub@redhat.com>
2321
2322         * src/powerpc/sysv.S (smst_two_register): Don't call __ashldi3,
2323         instead do the shifting inline.
2324         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't compute %r5
2325         shift count unconditionally.  Simplify load sequences for 1, 2, 3, 4
2326         and 8 byte structs, for the remaining struct sizes don't call
2327         __lshrdi3, instead do the shifting inline.
2328
2329 2005-12-07  Thiemo Seufer  <ths@networkno.de>
2330
2331         * src/mips/ffitarget.h: Remove obsolete sgidefs.h include. Add
2332         missing parentheses.
2333         * src/mips/o32.S (ffi_call_O32): Code formatting. Define
2334         and use A3_OFF, FP_OFF, RA_OFF. Micro-optimizations.
2335         (ffi_closure_O32): Likewise, but with newly defined A3_OFF2,
2336         A2_OFF2, A1_OFF2, A0_OFF2, RA_OFF2, FP_OFF2, S0_OFF2, GP_OFF2,
2337         V1_OFF2, V0_OFF2, FA_1_1_OFF2, FA_1_0_OFF2, FA_0_1_OFF2,
2338         FA_0_0_OFF2.
2339         * src/mips/ffi.c (ffi_prep_args): Code formatting. Fix
2340         endianness bugs.
2341         (ffi_prep_closure): Improve trampoline instruction scheduling.
2342         (ffi_closure_mips_inner_O32): Fix endianness bugs.
2343
2344 2005-12-03  Alan Modra  <amodra@bigpond.net.au>
2345
2346         * src/powerpc/ffi.c: Formatting.
2347         (ffi_prep_args_SYSV): Avoid possible aliasing problems by using unions.
2348         (ffi_prep_args64): Likewise.
2349
2350 2005-09-30  Geoffrey Keating  <geoffk@apple.com>
2351
2352         * testsuite/lib/libffi-dg.exp (libffi_target_compile): For
2353         darwin, use -shared-libgcc not -lgcc_s, and explain why.
2354
2355 2005-09-26  Tom Tromey  <tromey@redhat.com>
2356
2357         * testsuite/libffi.call/float1.c (value_type): New typedef.
2358         (CANARY): New define.
2359         (main): Check for result buffer overflow.
2360         * src/powerpc/linux64.S: Handle linux64 long double returns.
2361         * src/powerpc/ffi.c (FLAG_RETURNS_128BITS): New constant.
2362         (ffi_prep_cif_machdep): Handle linux64 long double returns.
2363
2364 2005-08-25  Alan Modra  <amodra@bigpond.net.au>
2365
2366         PR target/23404
2367         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Correct placement of stack
2368         homed fp args.
2369         (ffi_status ffi_prep_cif_machdep): Correct stack sizing for same.
2370
2371 2005-08-11  Jakub Jelinek  <jakub@redhat.com>
2372
2373         * configure.ac (HAVE_HIDDEN_VISIBILITY_ATTRIBUTE): New test.
2374         (AH_BOTTOM): Add FFI_HIDDEN definition.
2375         * configure: Rebuilt.
2376         * fficonfig.h.in: Rebuilt.
2377         * src/powerpc/ffi.c (hidden): Remove.
2378         (ffi_closure_LINUX64, ffi_prep_args64, ffi_call_LINUX64,
2379         ffi_closure_helper_LINUX64): Use FFI_HIDDEN instead of hidden.
2380         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64,
2381         .ffi_closure_LINUX64): Use FFI_HIDDEN instead of .hidden.
2382         * src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV): Remove,
2383         add FFI_HIDDEN to its prototype.
2384         (ffi_closure_SYSV_inner): New.
2385         * src/x86/sysv.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2386         * src/x86/win32.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2387
2388 2005-08-10  Alfred M. Szmidt  <ams@gnu.org>
2389
2390         PR libffi/21819:
2391         * configure: Rebuilt.
2392         * configure.ac: Handle i*86-*-gnu*.
2393
2394 2005-08-09  Jakub Jelinek  <jakub@redhat.com>
2395
2396         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Use
2397         DW_CFA_offset_extended_sf rather than
2398         DW_CFA_GNU_negative_offset_extended.
2399         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise.
2400
2401 2005-07-22  SUGIOKA Toshinobu  <sugioka@itonet.co.jp>
2402
2403         * src/sh/sysv.S (ffi_call_SYSV): Stop argument popping correctly
2404         on sh3.
2405         (ffi_closure_SYSV): Change the stack layout for sh3 struct argument.
2406         * src/sh/ffi.c (ffi_prep_args): Fix sh3 argument copy, when it is
2407         partially on register.
2408         (ffi_closure_helper_SYSV): Likewise.
2409         (ffi_prep_cif_machdep): Don't set too many cif->flags.
2410
2411 2005-07-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2412
2413         * src/sh/ffi.c (ffi_call): Handle small structures correctly.
2414         Remove empty line.
2415         * src/sh64/ffi.c (simple_type): Remove.
2416         (return_type): Handle small structures correctly.
2417         (ffi_prep_args): Likewise.
2418         (ffi_call): Likewise.
2419         (ffi_closure_helper_SYSV): Likewise.
2420         * src/sh64/sysv.S (ffi_call_SYSV): Handle 1, 2 and 4-byte return.
2421         Emit position independent code if PIC and remove wrong datalabel
2422         prefixes from EH data.
2423
2424 2005-07-19  Andreas Tobler  <a.tobler@schweiz.ch>
2425
2426         * Makefile.am (nodist_libffi_la_SOURCES): Add POWERPC_FREEBSD.
2427         * Makefile.in: Regenerate.
2428         * include/Makefile.in: Likewise.
2429         * testsuite/Makefile.in: Likewise.
2430         * configure.ac: Add POWERPC_FREEBSD rules.
2431         * configure: Regenerate.
2432         * src/powerpc/ffitarget.h: Add POWERPC_FREEBSD rules.
2433         (FFI_SYSV_TYPE_SMALL_STRUCT): Define.
2434         * src/powerpc/ffi.c: Add flags to handle small structure returns
2435         in ffi_call_SYSV.
2436         (ffi_prep_cif_machdep): Handle small structures for SYSV 4 ABI.
2437         Aka FFI_SYSV.
2438         (ffi_closure_helper_SYSV): Likewise.
2439         * src/powerpc/ppc_closure.S: Add return types for small structures.
2440         * src/powerpc/sysv.S: Add bits to handle small structures for
2441         final SYSV 4 ABI.
2442
2443 2005-07-10  Andreas Tobler  <a.tobler@schweiz.ch>
2444
2445         * testsuite/libffi.call/cls_5_1_byte.c: New test file.
2446         * testsuite/libffi.call/cls_6_1_byte.c: Likewise.
2447         * testsuite/libffi.call/cls_7_1_byte.c: Likewise.
2448
2449 2005-07-05  Randolph Chung  <tausq@debian.org>
2450
2451         * src/pa/ffi.c (ffi_struct_type): Rename FFI_TYPE_SMALL_STRUCT1
2452         as FFI_TYPE_SMALL_STRUCT3.  Break out handling for 5-7 byte
2453         structures.  Kill compilation warnings.
2454         (ffi_closure_inner_LINUX): Print return values as hex in debug
2455         message.  Rename FFI_TYPE_SMALL_STRUCT1 as FFI_TYPE_SMALL_STRUCT3.
2456         Properly handle 5-7 byte structure returns.
2457         * src/pa/ffitarget.h (FFI_TYPE_SMALL_STRUCT1)
2458         (FFI_TYPE_SMALL_STRUCT2): Remove.
2459         (FFI_TYPE_SMALL_STRUCT3, FFI_TYPE_SMALL_STRUCT5)
2460         (FFI_TYPE_SMALL_STRUCT6, FFI_TYPE_SMALL_STRUCT7): Define.
2461         * src/pa/linux.S: Mark source file as using PA1.1 assembly.
2462         (checksmst1, checksmst2): Remove.
2463         (checksmst3): Optimize handling of 3-byte struct returns.
2464         (checksmst567): Properly handle 5-7 byte struct returns.
2465
2466 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2467
2468         PR libgcj/21943
2469         * src/mips/n32.S: Enforce PIC code.
2470         * src/mips/o32.S: Likewise.
2471
2472 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2473
2474         * configure.ac: Treat i*86-*-solaris2.10 and up as X86_64.
2475         * configure: Regenerate.
2476
2477 2005-06-01  Alan Modra  <amodra@bigpond.net.au>
2478
2479         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't use JUMPTARGET
2480         to call ffi_closure_helper_SYSV.  Append @local instead.
2481         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise for ffi_prep_args_SYSV.
2482
2483 2005-05-17  Kelley Cook  <kcook@gcc.gnu.org>
2484
2485         * configure.ac: Use AC_C_BIGENDIAN instead of AC_C_BIGENDIAN_CROSS.
2486         Use AC_CHECK_SIZEOF instead of AC_COMPILE_CHECK_SIZEOF.
2487         * Makefile.am (ACLOCAL_AMFLAGS): Remove -I ../config.
2488         * aclocal.m4, configure, fficonfig.h.in, Makefile.in,
2489         include/Makefile.in, testsuite/Makefile.in: Regenerate.
2490
2491 2005-05-09  Mike Stump  <mrs@apple.com>
2492
2493         * configure: Regenerate.
2494
2495 2005-05-08  Richard Henderson  <rth@redhat.com>
2496
2497         PR libffi/21285
2498         * src/alpha/osf.S: Update unwind into to match code.
2499
2500 2005-05-04  Andreas Degert <ad@papyrus-gmbh.de>
2501             Richard Henderson  <rth@redhat.com>
2502
2503         * src/x86/ffi64.c (ffi_prep_cif_machdep): Save sse-used flag in
2504         bit 11 of flags.
2505         (ffi_call): Mask return type field.  Pass ssecount to ffi_call_unix64.
2506         (ffi_prep_closure): Set carry bit if sse-used flag set.
2507         * src/x86/unix64.S (ffi_call_unix64): Add ssecount argument.
2508         Only load sse registers if ssecount non-zero.
2509         (ffi_closure_unix64): Only save sse registers if carry set on entry.
2510
2511 2005-04-29  Ralf Corsepius  <ralf.corsepius@rtems.org>
2512
2513         * configure.ac: Add i*86-*-rtems*, sparc*-*-rtems*,
2514         powerpc-*rtems*, arm*-*-rtems*, sh-*-rtems*.
2515         * configure: Regenerate.
2516
2517 2005-04-20  Hans-Peter Nilsson  <hp@axis.com>
2518
2519         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): In regsub use,
2520         have Tcl8.3-compatible intermediate variable.
2521
2522 2005-04-18  Simon Posnjak <simon.posnjak@siol.net>
2523             Hans-Peter Nilsson  <hp@axis.com>
2524
2525         * Makefile.am: Add CRIS support.
2526         * configure.ac: Likewise.
2527         * Makefile.in, configure, testsuite/Makefile.in,
2528         include/Makefile.in: Regenerate.
2529         * src/cris: New directory.
2530         * src/cris/ffi.c, src/cris/sysv.S, src/cris/ffitarget.h: New files.
2531         * src/prep_cif.c (ffi_prep_cif): Wrap in #ifndef __CRIS__.
2532
2533         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): Replace \n with
2534         \r?\n in output tests.
2535
2536 2005-04-12  Mike Stump  <mrs@apple.com>
2537
2538         * configure: Regenerate.
2539
2540 2005-03-30  Hans Boehm  <Hans.Boehm@hp.com>
2541
2542         * src/ia64/ffitarget.h (ffi_arg): Use long long instead of DI.
2543
2544 2005-03-30  Steve Ellcey  <sje@cup.hp.com>
2545
2546         * src/ia64/ffitarget.h (ffi_arg) ADD DI attribute.
2547         (ffi_sarg) Ditto.
2548         * src/ia64/unix.S (ffi_closure_unix): Extend gp
2549         to 64 bits in ILP32 mode.
2550         Load 64 bits even for short data.
2551
2552 2005-03-23  Mike Stump  <mrs@apple.com>
2553
2554         * src/powerpc/darwin.S: Update for -m64 multilib.
2555         * src/powerpc/darwin_closure.S: Likewise.
2556
2557 2005-03-21  Zack Weinberg  <zack@codesourcery.com>
2558
2559         * configure.ac: Do not invoke TL_AC_GCC_VERSION.
2560         Do not set tool_include_dir.
2561         * aclocal.m4, configure, Makefile.in, testsuite/Makefile.in:
2562         Regenerate.
2563         * include/Makefile.am: Set gcc_version and toollibffidir.
2564         * include/Makefile.in: Regenerate.
2565
2566 2005-02-22  Andrew Haley  <aph@redhat.com>
2567
2568         * src/powerpc/ffi.c (ffi_prep_cif_machdep): Bump alignment to
2569         odd-numbered register pairs for 64-bit integer types.
2570
2571 2005-02-23  Andreas Tobler  <a.tobler@schweiz.ch>
2572
2573         PR libffi/20104
2574         * testsuite/libffi.call/return_ll1.c: New test case.
2575
2576 2005-02-11  Janis Johnson  <janis187@us.ibm.com>
2577
2578         * testsuite/libffi.call/cls_align_longdouble.c: Remove dg-options.
2579         * testsuite/libffi.call/float.c: Ditto.
2580         * testsuite/libffi.call/float2.c: Ditto.
2581         * testsuite/libffi.call/float3.c: Ditto.
2582
2583 2005-02-08  Andreas Tobler  <a.tobler@schweiz.ch>
2584
2585         * src/frv/ffitarget.h: Remove PPC stuff which does not belong to frv.
2586
2587 2005-01-12  Eric Botcazou  <ebotcazou@libertysurf.fr>
2588
2589         * testsuite/libffi.special/special.exp (cxx_options): Add
2590         -shared-libgcc.
2591
2592 2004-12-31  Richard Henderson  <rth@redhat.com>
2593
2594         * src/types.c (FFI_AGGREGATE_TYPEDEF): Remove.
2595         (FFI_TYPEDEF): Rename from FFI_INTEGRAL_TYPEDEF.  Replace size and
2596         offset parameters with a type parameter; deduce size and structure
2597         alignment.  Update all users.
2598
2599 2004-12-31  Richard Henderson  <rth@redhat.com>
2600
2601         * src/types.c (FFI_TYPE_POINTER): Define with sizeof.
2602         (FFI_TYPE_LONGDOUBLE): Fix for ia64.
2603         * src/ia64/ffitarget.h (struct ffi_ia64_trampoline_struct): Move
2604         into ffi_prep_closure.
2605         * src/ia64/ia64_flags.h, src/ia64/ffi.c, src/ia64/unix.S: Rewrite
2606         from scratch.
2607
2608 2004-12-27  Richard Henderson  <rth@redhat.com>
2609
2610         * src/x86/unix64.S: Fix typo in unwind info.
2611
2612 2004-12-25  Richard Henderson  <rth@redhat.com>
2613
2614         * src/x86/ffi64.c (struct register_args): Rename from stackLayout.
2615         (enum x86_64_reg_class): Add X86_64_COMPLEX_X87_CLASS.
2616         (merge_classes): Check for it.
2617         (SSE_CLASS_P): New.
2618         (classify_argument): Pass byte_offset by value; perform all updates
2619         inside struct case.
2620         (examine_argument): Add classes argument; handle
2621         X86_64_COMPLEX_X87_CLASS.
2622         (ffi_prep_args): Merge into ...
2623         (ffi_call): ... here.  Share stack frame with ffi_call_unix64.
2624         (ffi_prep_cif_machdep): Setup cif->flags for proper structure return.
2625         (ffi_fill_return_value): Remove.
2626         (ffi_prep_closure): Remove dead assert.
2627         (ffi_closure_unix64_inner): Rename from ffi_closure_UNIX64_inner.
2628         Rewrite to use struct register_args instead of va_list.  Create
2629         flags for handling structure returns.
2630         * src/x86/unix64.S: Remove dead strings.
2631         (ffi_call_unix64): Rename from ffi_call_UNIX64.  Rewrite to share
2632         stack frame with ffi_call.  Handle structure returns properly.
2633         (float2sse, floatfloat2sse, double2sse): Remove.
2634         (sse2float, sse2double, sse2floatfloat): Remove.
2635         (ffi_closure_unix64): Rename from ffi_closure_UNIX64.  Rewrite
2636         to handle structure returns properly.
2637
2638 2004-12-08  David Edelsohn  <edelsohn@gnu.org>
2639
2640         * Makefile.am (AM_MAKEFLAGS): Remove duplicate LIBCFLAGS and
2641         PICFLAG.
2642         * Makefile.in: Regenerated.
2643
2644 2004-12-02  Richard Sandiford  <rsandifo@redhat.com>
2645
2646         * configure.ac: Use TL_AC_GCC_VERSION to set gcc_version.
2647         * configure, aclocal.m4, Makefile.in: Regenerate.
2648         * include/Makefile.in, testsuite/Makefile.in: Regenerate.
2649
2650 2004-11-29  Kelley Cook  <kcook@gcc.gnu.org>
2651
2652         * configure: Regenerate for libtool change.
2653
2654 2004-11-25  Kelley Cook  <kcook@gcc.gnu.org>
2655
2656         * configure: Regenerate for libtool reversion.
2657
2658 2004-11-24  Kelley Cook  <kcook@gcc.gnu.org>
2659
2660         * configure: Regenerate for libtool change.
2661
2662 2004-11-23  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2663
2664         * testsuite/lib/libffi-dg.exp: Use new procs in target-libpath.exp.
2665
2666 2004-11-23  Richard Sandiford  <rsandifo@redhat.com>
2667
2668         * src/mips/o32.S (ffi_call_O32, ffi_closure_O32): Use jalr instead
2669         of jal.  Use an absolute encoding for the frame information.
2670
2671 2004-11-23  Kelley Cook  <kcook@gcc.gnu.org>
2672
2673         * Makefile.am: Remove no-dependencies.  Add ACLOCAL_AMFLAGS.
2674         * acinclude.m4: Delete logic for sincludes.
2675         * aclocal.m4, Makefile.in, configure: Regenerate.
2676         * include/Makefile: Likewise.
2677         * testsuite/Makefile: Likewise.
2678
2679 2004-11-22  Eric Botcazou  <ebotcazou@libertysurf.fr>
2680
2681         * src/sparc/ffi.c (ffi_prep_closure): Align doubles and 64-bit integers
2682         on a 8-byte boundary.
2683         * src/sparc/v8.S (ffi_closure_v8): Reserve frame space for arguments.
2684
2685 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2686
2687         * src/arm/ffi.c (ffi_prep_cif_machdep): Handle functions that return
2688         long long values.  Round stack allocation to a multiple of 8 bytes
2689         for ATPCS compatibility.
2690         * src/arm/sysv.S (ffi_call_SYSV): Rework to avoid use of APCS register
2691         names.  Handle returning long long types.  Add Thumb and interworking
2692         support.  Improve soft-float code.
2693
2694 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2695
2696         * testsuite/lib/libffi-db.exp (load_gcc_lib): New function.
2697         (libffi_exit): New function.
2698         (libffi_init): Build the testglue wrapper if needed.
2699
2700 2004-10-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2701
2702         PR other/18138
2703         * testsuite/lib/libffi-dg.exp: Accept more than one multilib libgcc.
2704
2705 2004-10-25  Kazuhiro Inaoka <inaoka.kazuhiro@renesas.com>
2706
2707         * src/m32r/libffitarget.h (FFI_CLOSURES): Set to 0.
2708
2709 2004-10-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2710
2711         * src/sh/sysv.S (ffi_call_SYSV): Don't align for double data.
2712         * testsuite/libffi.call/float3.c: New test case.
2713
2714 2004-10-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2715
2716         * src/sh/ffi.c (ffi_prep_closure): Set T bit in trampoline for
2717         the function returning a structure pointed with R2.
2718         * src/sh/sysv.S (ffi_closure_SYSV): Use R2 as the pointer to
2719         the structure return value if T bit set.  Emit position
2720         independent code and EH data if PIC.
2721
2722 2004-10-13  Kazuhiro Inaoka  <inaoka.kazuhiro@renesas.com>
2723
2724         * Makefile.am: Add m32r support.
2725         * configure.ac: Likewise.
2726         * Makefile.in: Regenerate.
2727         * confiugre: Regenerate.
2728         * src/types.c: Add m32r port to FFI_INTERNAL_TYPEDEF
2729         (uint64, sint64, double, longdouble)
2730         * src/m32r: New directory.
2731         * src/m32r/ffi.c: New file.
2732         * src/m32r/sysv.S: Likewise.
2733         * src/m32r/ffitarget.h: Likewise.
2734
2735 2004-10-02  Kaz Kojima  <kkojima@gcc.gnu.org>
2736
2737         * testsuite/libffi.call/negint.c: New test case.
2738
2739 2004-09-14  H.J. Lu  <hongjiu.lu@intel.com>
2740
2741         PR libgcj/17465
2742         * testsuite/lib/libffi-dg.exp: Don't use global ld_library_path.
2743         Set up LD_LIBRARY_PATH, SHLIB_PATH, LD_LIBRARYN32_PATH,
2744         LD_LIBRARY64_PATH, LD_LIBRARY_PATH_32, LD_LIBRARY_PATH_64 and
2745         DYLD_LIBRARY_PATH.
2746
2747 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2748
2749         * testsuite/libffi.call/many_win32.c: Remove whitespaces.
2750         * testsuite/libffi.call/promotion.c: Likewise.
2751         * testsuite/libffi.call/return_ll.c: Remove unused var. Cleanup
2752         whitespaces.
2753         * testsuite/libffi.call/return_sc.c: Likewise.
2754         * testsuite/libffi.call/return_uc.c: Likewise.
2755
2756 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2757
2758         * src/powerpc/darwin.S: Fix comments and identation.
2759         * src/powerpc/darwin_closure.S: Likewise.
2760
2761 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2762
2763         * src/powerpc/ffi_darwin.c: Add flag for longdouble return values.
2764         (ffi_prep_args): Handle longdouble arguments.
2765         (ffi_prep_cif_machdep): Set flags for longdouble. Calculate space for
2766         longdouble.
2767         (ffi_closure_helper_DARWIN): Add closure handling for longdouble.
2768         * src/powerpc/darwin.S (_ffi_call_DARWIN): Add handling of longdouble
2769         values.
2770         * src/powerpc/darwin_closure.S (_ffi_closure_ASM): Likewise.
2771         * src/types.c: Defined longdouble size and alignment for darwin.
2772
2773 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2774
2775         * src/powerpc/aix.S: Remove whitespaces.
2776         * src/powerpc/aix_closure.S: Likewise.
2777         * src/powerpc/asm.h: Likewise.
2778         * src/powerpc/ffi.c: Likewise.
2779         * src/powerpc/ffitarget.h: Likewise.
2780         * src/powerpc/linux64.S: Likewise.
2781         * src/powerpc/linux64_closure.S: Likewise.
2782         * src/powerpc/ppc_closure.S: Likewise.
2783         * src/powerpc/sysv.S: Likewise.
2784
2785 2004-08-30  Anthony Green  <green@redhat.com>
2786
2787         * Makefile.am: Add frv support.
2788         * Makefile.in, testsuite/Makefile.in: Rebuilt.
2789         * configure.ac: Read configure.host.
2790         * configure.in: Read configure.host.
2791         * configure.host: New file.  frv-elf needs libgloss.
2792         * include/ffi.h.in: Force ffi_closure to have a nice big (8)
2793         alignment.  This is needed to frv and shouldn't harm the others.
2794         * include/ffi_common.h (ALIGN_DOWN): New macro.
2795         * src/frv/ffi.c, src/frv/ffitarget.h, src/frv/eabi.S: New files.
2796
2797 2004-08-24  David Daney  <daney@avtrex.com>
2798
2799         * testsuite/libffi.call/closure_fn0.c: Xfail mips64* instead of mips*.
2800         * testsuite/libffi.call/closure_fn1.c: Likewise.
2801         * testsuite/libffi.call/closure_fn2.c  Likewise.
2802         * testsuite/libffi.call/closure_fn3.c: Likewise.
2803         * testsuite/libffi.call/closure_fn4.c: Likewise.
2804         * testsuite/libffi.call/closure_fn5.c: Likewise.
2805         * testsuite/libffi.call/cls_18byte.c: Likewise.
2806         * testsuite/libffi.call/cls_19byte.c: Likewise.
2807         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2808         * testsuite/libffi.call/cls_20byte.c: Likewise.
2809         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2810         * testsuite/libffi.call/cls_24byte.c: Likewise.
2811         * testsuite/libffi.call/cls_2byte.c: Likewise.
2812         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2813         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2814         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2815         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2816         * testsuite/libffi.call/cls_4byte.c: Likewise.
2817         * testsuite/libffi.call/cls_64byte.c: Likewise.
2818         * testsuite/libffi.call/cls_6byte.c: Likewise.
2819         * testsuite/libffi.call/cls_7byte.c: Likewise.
2820         * testsuite/libffi.call/cls_8byte.c: Likewise.
2821         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2822         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2823         * testsuite/libffi.call/cls_align_double.c: Likewise.
2824         * testsuite/libffi.call/cls_align_float.c: Likewise.
2825         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2826         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2827         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2828         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2829         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2830         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2831         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2832         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
2833         * testsuite/libffi.call/cls_double.c: Likewise.
2834         * testsuite/libffi.call/cls_float.c: Likewise.
2835         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2836         * testsuite/libffi.call/cls_multi_sshort.c: Likewise.
2837         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2838         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2839         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2840         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2841         * testsuite/libffi.call/cls_schar.c: Likewise.
2842         * testsuite/libffi.call/cls_sint.c: Likewise.
2843         * testsuite/libffi.call/cls_sshort.c: Likewise.
2844         * testsuite/libffi.call/cls_uchar.c: Likewise.
2845         * testsuite/libffi.call/cls_uint.c: Likewise.
2846         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2847         * testsuite/libffi.call/cls_ushort.c: Likewise.
2848         * testsuite/libffi.call/nested_struct.c: Likewise.
2849         * testsuite/libffi.call/nested_struct1.c: Likewise.
2850         * testsuite/libffi.call/nested_struct2.c: Likewise.
2851         * testsuite/libffi.call/nested_struct3.c: Likewise.
2852         * testsuite/libffi.call/problem1.c: Likewise.
2853         * testsuite/libffi.special/unwindtest.cc: Likewise.
2854         * testsuite/libffi.call/cls_12byte.c: Likewise and set return value
2855         to zero.
2856         * testsuite/libffi.call/cls_16byte.c: Likewise.
2857         * testsuite/libffi.call/cls_5byte.c: Likewise.
2858
2859 2004-08-23  David Daney <daney@avtrex.com>
2860
2861         PR libgcj/13141
2862         * src/mips/ffitarget.h (FFI_O32_SOFT_FLOAT): New ABI.
2863         * src/mips/ffi.c (ffi_prep_args): Fix alignment calculation.
2864         (ffi_prep_cif_machdep): Handle FFI_O32_SOFT_FLOAT floating point
2865         parameters and return types.
2866         (ffi_call): Handle FFI_O32_SOFT_FLOAT ABI.
2867         (ffi_prep_closure): Ditto.
2868         (ffi_closure_mips_inner_O32): Handle FFI_O32_SOFT_FLOAT ABI, fix
2869         alignment calculations.
2870         * src/mips/o32.S (ffi_closure_O32): Don't use floating point
2871         instructions if FFI_O32_SOFT_FLOAT, make stack frame ABI compliant.
2872
2873 2004-08-14  Casey Marshall <csm@gnu.org>
2874
2875         * src/mips/ffi.c (ffi_pref_cif_machdep): set `cif->flags' to
2876         contain `FFI_TYPE_UINT64' as return type for any 64-bit
2877         integer (O32 ABI only).
2878         (ffi_prep_closure): new function.
2879         (ffi_closure_mips_inner_O32): new function.
2880         * src/mips/ffitarget.h: Define `FFI_CLOSURES' and
2881         `FFI_TRAMPOLINE_SIZE' appropriately if the ABI is o32.
2882         * src/mips/o32.S (ffi_call_O32): add labels for .eh_frame. Return
2883         64 bit integers correctly.
2884         (ffi_closure_O32): new function.
2885         Added DWARF-2 unwind info for both functions.
2886
2887 2004-08-10  Andrew Haley  <aph@redhat.com>
2888
2889         * src/x86/ffi64.c (ffi_prep_args ): 8-align all stack arguments.
2890
2891 2004-08-01  Robert Millan  <robertmh@gnu.org>
2892
2893         * configure.ac: Detect knetbsd-gnu and kfreebsd-gnu.
2894         * configure: Regenerate.
2895
2896 2004-07-30  Maciej W. Rozycki  <macro@linux-mips.org>
2897
2898         * acinclude.m4 (AC_FUNC_MMAP_BLACKLIST): Check for <sys/mman.h>
2899         and mmap() explicitly instead of relying on preset autoconf cache
2900         variables.
2901         * aclocal.m4: Regenerate.
2902         * configure: Regenerate.
2903
2904 2004-07-11  Ulrich Weigand  <uweigand@de.ibm.com>
2905
2906         * src/s390/ffi.c (ffi_prep_args): Fix C aliasing violation.
2907         (ffi_check_float_struct): Remove unused prototype.
2908
2909 2004-06-30  Geoffrey Keating  <geoffk@apple.com>
2910
2911         * src/powerpc/ffi_darwin.c (flush_icache): ';' is a comment
2912         character on Darwin, use '\n\t' instead.
2913
2914 2004-06-26  Matthias Klose  <doko@debian.org>
2915
2916         * libtool-version: Fix typo in revision/age.
2917
2918 2004-06-17  Matthias Klose  <doko@debian.org>
2919
2920         * libtool-version: New.
2921         * Makefile.am (libffi_la_LDFLAGS): Use -version-info for soname.
2922         * Makefile.in: Regenerate.
2923
2924 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2925
2926         * Makefile.am: Remove useless multilib rules.
2927         * Makefile.in: Regenerate.
2928         * aclocal.m4: Regenerate with automake 1.8.5.
2929         * configure.ac: Remove useless multilib configury.
2930         * configure: Regenerate.
2931
2932 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2933
2934         * .cvsignore: New file.
2935
2936 2004-06-10  Jakub Jelinek  <jakub@redhat.com>
2937
2938         * src/ia64/unix.S (ffi_call_unix): Insert group barrier break
2939         fp_done.
2940         (ffi_closure_UNIX): Fix f14/f15 adjustment if FLOAT_SZ is ever
2941         changed from 8.
2942
2943 2004-06-06  Sean McNeil  <sean@mcneil.com>
2944
2945         * configure.ac: Add x86_64-*-freebsd* support.
2946         * configure: Regenerate.
2947
2948 2004-04-26  Joe Buck <jbuck@welsh-buck.org>
2949
2950         Bug 15093
2951         * configure.ac: Test for existence of mmap and sys/mman.h before
2952         checking blacklist.  Fix suggested by Jim Wilson.
2953         * configure: Regenerate.
2954
2955 2004-04-26  Matt Austern  <austern@apple.com>
2956
2957         * src/powerpc/darwin.S: Go through a non-lazy pointer for initial
2958         FDE location.
2959         * src/powerpc/darwin_closure.S: Likewise.
2960
2961 2004-04-24  Andreas Tobler  <a.tobler@schweiz.ch>
2962
2963         * testsuite/libffi.call/cls_multi_schar.c (main): Fix initialization
2964         error. Reported by Thomas Heller <theller@python.net>.
2965         * testsuite/libffi.call/cls_multi_sshort.c (main): Likewise.
2966         * testsuite/libffi.call/cls_multi_ushort.c (main): Likewise.
2967
2968 2004-03-20  Matthias Klose  <doko@debian.org>
2969
2970         * src/pa/linux.S: Fix typo.
2971
2972 2004-03-19  Matthias Klose  <doko@debian.org>
2973
2974         * Makefile.am: Update.
2975         * Makefile.in: Regenerate.
2976         * src/pa/ffi.h.in: Remove.
2977         * src/pa/ffitarget.h: New file.
2978
2979 2004-02-10  Randolph Chung  <tausq@debian.org>
2980
2981         * Makefile.am: Add PA support.
2982         * Makefile.in: Regenerate.
2983         * include/Makefile.in: Regenerate.
2984         * configure.ac: Add PA target.
2985         * configure: Regenerate.
2986         * src/pa/ffi.c: New file.
2987         * src/pa/ffi.h.in: Add PA support.
2988         * src/pa/linux.S: New file.
2989         * prep_cif.c: Add PA support.
2990
2991 2004-03-16  Hosaka Yuji  <hos@tamanegi.org>
2992
2993         * src/types.c: Fix alignment size of X86_WIN32 case int64 and
2994         double.
2995         * src/x86/ffi.c (ffi_prep_args): Replace ecif->cif->rtype->type
2996         with ecif->cif->flags.
2997         (ffi_call, ffi_prep_incoming_args_SYSV): Replace cif->rtype->type
2998         with cif->flags.
2999         (ffi_prep_cif_machdep): Add X86_WIN32 struct case.
3000         (ffi_closure_SYSV): Add 1 or 2-bytes struct case for X86_WIN32.
3001         * src/x86/win32.S (retstruct1b, retstruct2b, sc_retstruct1b,
3002         sc_retstruct2b): Add for 1 or 2-bytes struct case.
3003
3004 2004-03-15 Kelley Cook <kcook@gcc.gnu.org>
3005
3006         * configure.in: Rename file to ...
3007         * configure.ac: ... this.
3008         * fficonfig.h.in: Regenerate.
3009         * Makefile.in: Regenerate.
3010         * include/Makefile.in: Regenerate.
3011         * testsuite/Makefile.in: Regenerate.
3012
3013 2004-03-12  Matt Austern  <austern@apple.com>
3014
3015         * src/powerpc/darwin.S: Fix EH information so it corresponds to
3016         changes in EH format resulting from addition of linkonce support.
3017         * src/powerpc/darwin_closure.S: Likewise.
3018
3019 2004-03-11  Andreas Tobler  <a.tobler@schweiz.ch>
3020             Paolo Bonzini  <bonzini@gnu.org>
3021
3022         * Makefile.am (AUTOMAKE_OPTIONS): Set them.
3023         Remove VPATH. Remove rules for object files. Remove multilib support.
3024         (AM_CCASFLAGS): Add.
3025         * configure.in (AC_CONFIG_HEADERS): Relace AM_CONFIG_HEADER.
3026         (AC_PREREQ): Bump version to 2.59.
3027         (AC_INIT): Fill with version info and bug address.
3028         (ORIGINAL_LD_FOR_MULTILIBS): Remove.
3029         (AM_ENABLE_MULTILIB): Use this instead of AC_ARG_ENABLE.
3030         De-precious CC so that the right flags are passed down to multilibs.
3031         (AC_MSG_ERROR): Replace obsolete macro AC_ERROR.
3032         (AC_CONFIG_FILES): Replace obsolete macro AC_LINK_FILES.
3033         (AC_OUTPUT): Reorganize the output with AC_CONFIG_COMMANDS.
3034         * configure: Rebuilt.
3035         * aclocal.m4: Likewise.
3036         * Makefile.in, include/Makefile.in, testsuite/Makefile.in: Likewise.
3037         * fficonfig.h.in: Likewise.
3038
3039 2004-03-11  Andreas Schwab  <schwab@suse.de>
3040
3041         * src/ia64/ffi.c (ffi_prep_incoming_args_UNIX): Get floating point
3042         arguments from fp registers only for the first 8 parameter slots.
3043         Don't convert a float parameter when passed in memory.
3044
3045 2004-03-09  Hans-Peter Nilsson  <hp@axis.com>
3046
3047         * configure: Regenerate for config/accross.m4 correction.
3048
3049 2004-02-25  Matt Kraai  <kraai@alumni.cmu.edu>
3050
3051         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Change
3052         ecif->cif->bytes to bytes.
3053         (ffi_prep_cif_machdep): Add braces around nested if statement.
3054
3055 2004-02-09  Alan Modra  <amodra@bigpond.net.au>
3056
3057         * src/types.c (pointer): POWERPC64 has 8 byte pointers.
3058
3059         * src/powerpc/ffi.c (ffi_prep_args64): Correct long double handling.
3060         (ffi_closure_helper_LINUX64): Fix typo.
3061         * testsuite/libffi.call/cls_align_longdouble.c: Pass -mlong-double-128
3062         for powerpc64-*-*.
3063         * testsuite/libffi.call/float.c: Likewise.
3064         * testsuite/libffi.call/float2.c: Likewise.
3065
3066 2004-02-08  Alan Modra  <amodra@bigpond.net.au>
3067
3068         * src/powerpc/ffi.c (ffi_prep_cif_machdep <FFI_LINUX64>): Correct
3069         long double function return and long double arg handling.
3070         (ffi_closure_helper_LINUX64): Formatting.  Delete unused "ng" var.
3071         Use "end_pfr" instead of "nf".  Correct long double handling.
3072         Localise "temp".
3073         * src/powerpc/linux64.S (ffi_call_LINUX64): Save f2 long double
3074         return value.
3075         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Allocate
3076         space for long double return value.  Adjust stack frame and offsets.
3077         Load f2 long double return.
3078
3079 2004-02-07  Alan Modra  <amodra@bigpond.net.au>
3080
3081         * src/types.c: Use 16 byte long double for POWERPC64.
3082
3083 2004-01-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
3084
3085         * src/sparc/ffi.c (ffi_prep_args_v9): Shift the parameter array
3086         when the structure return address is passed in %o0.
3087         (ffi_V9_return_struct): Rename into ffi_v9_layout_struct.
3088         (ffi_v9_layout_struct): Align the field following a nested structure
3089         on a word boundary.  Use memmove instead of memcpy.
3090         (ffi_call): Update call to ffi_V9_return_struct.
3091         (ffi_prep_closure): Define 'ctx' only for V8.
3092         (ffi_closure_sparc_inner): Clone into ffi_closure_sparc_inner_v8
3093         and ffi_closure_sparc_inner_v9.
3094         (ffi_closure_sparc_inner_v8): Return long doubles by reference.
3095         Always skip the structure return address.  For structures and long
3096         doubles, copy the argument directly.
3097         (ffi_closure_sparc_inner_v9): Skip the structure return address only
3098         if required.  Shift the maximum floating-point slot accordingly.  For
3099         big structures, copy the argument directly; otherwise, left-justify the
3100         argument and call ffi_v9_layout_struct to lay out the structure on
3101         the stack.
3102         * src/sparc/v8.S: Undef STACKFRAME before defining it.
3103         (ffi_closure_v8): Pass the structure return address.  Update call to
3104         ffi_closure_sparc_inner_v8.  Short-circuit FFI_TYPE_INT handling.
3105         Skip the 'unimp' insn when returning long doubles and structures.
3106         * src/sparc/v9.S: Undef STACKFRAME before defining it.
3107         (ffi_closure_v9): Increase the frame size by 2 words.  Short-circuit
3108         FFI_TYPE_INT handling.  Load structures both in integers and
3109         floating-point registers on return.
3110         * README: Update status of the SPARC port.
3111
3112 2004-01-24  Andreas Tobler  <a.tobler@schweiz.ch>
3113
3114         * testsuite/libffi.call/pyobjc-tc.c (main): Treat result value
3115         as of type ffi_arg.
3116         * testsuite/libffi.call/struct3.c (main): Fix CHECK.
3117
3118 2004-01-22  Ulrich Weigand  <uweigand@de.ibm.com>
3119
3120         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Treat result
3121         value as of type ffi_arg, not unsigned int.
3122
3123 2004-01-21  Michael Ritzert  <ritzert@t-online.de>
3124
3125         * ffi64.c (ffi_prep_args): Cast the RHS of an assignment instead
3126         of the LHS.
3127
3128 2004-01-12  Andreas Tobler  <a.tobler@schweiz.ch>
3129
3130         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_32 for
3131         Solaris.
3132
3133 2004-01-08  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3134
3135         * testsuite/libffi.call/ffitest.h (allocate_mmap): Cast MAP_FAILED
3136         to void *.
3137
3138 2003-12-10  Richard Henderson  <rth@redhat.com>
3139
3140         * testsuite/libffi.call/cls_align_pointer.c: Cast pointers to
3141         size_t instead of int.
3142
3143 2003-12-04  Hosaka Yuji  <hos@tamanegi.org>
3144
3145         * testsuite/libffi.call/many_win32.c: Include <float.h>.
3146         * testsuite/libffi.call/many_win32.c (main): Replace variable
3147         int i with unsigned long ul.
3148
3149         * testsuite/libffi.call/cls_align_uint64.c: New test case.
3150         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
3151         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
3152         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
3153         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
3154         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
3155         * testsuite/libffi.call/cls_align_float.c: Likewise.
3156         * testsuite/libffi.call/cls_align_double.c: Likewise.
3157         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
3158         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
3159
3160 2003-12-02  Hosaka Yuji  <hos@tamanegi.org>
3161
3162         PR other/13221
3163         * src/x86/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV):
3164         Align arguments to 32 bits.
3165
3166 2003-12-01  Andreas Tobler  <a.tobler@schweiz.ch>
3167
3168         PR other/13221
3169         * testsuite/libffi.call/cls_multi_sshort.c: New test case.
3170         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
3171         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
3172         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
3173         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
3174         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
3175
3176         * testsuite/libffi.special/unwindtest.cc: Cosmetics.
3177
3178 2003-11-26  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
3179
3180         * testsuite/libffi.call/ffitest.h: Include <fcntl.h>.
3181         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3182
3183 2003-11-22  Andreas Tobler  <a.tobler@schweiz.ch>
3184
3185         * Makefile.in: Rebuilt.
3186         * configure: Likewise.
3187         * testsuite/libffi.special/unwindtest.cc: Convert the mmap to
3188         the right type.
3189
3190 2003-11-21  Andreas Jaeger  <aj@suse.de>
3191             Andreas Tobler  <a.tobler@schweiz.ch>
3192
3193         * acinclude.m4: Add AC_FUNC_MMAP_BLACKLIST.
3194         * configure.in: Call AC_FUNC_MMAP_BLACKLIST.
3195         * Makefile.in: Rebuilt.
3196         * aclocal.m4: Likewise.
3197         * configure: Likewise.
3198         * fficonfig.h.in: Likewise.
3199         * testsuite/lib/libffi-dg.exp: Add include dir.
3200         * testsuite/libffi.call/ffitest.h: Add MMAP definitions.
3201         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3202         * testsuite/libffi.call/closure_fn0.c: Use MMAP functionality
3203         for ffi_closure if available.
3204         * testsuite/libffi.call/closure_fn1.c: Likewise.
3205         * testsuite/libffi.call/closure_fn2.c: Likewise.
3206         * testsuite/libffi.call/closure_fn3.c: Likewise.
3207         * testsuite/libffi.call/closure_fn4.c: Likewise.
3208         * testsuite/libffi.call/closure_fn5.c: Likewise.
3209         * testsuite/libffi.call/cls_12byte.c: Likewise.
3210         * testsuite/libffi.call/cls_16byte.c: Likewise.
3211         * testsuite/libffi.call/cls_18byte.c: Likewise.
3212         * testsuite/libffi.call/cls_19byte.c: Likewise.
3213         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3214         * testsuite/libffi.call/cls_20byte.c: Likewise.
3215         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3216         * testsuite/libffi.call/cls_24byte.c: Likewise.
3217         * testsuite/libffi.call/cls_2byte.c: Likewise.
3218         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3219         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3220         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3221         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3222         * testsuite/libffi.call/cls_4byte.c: Likewise.
3223         * testsuite/libffi.call/cls_5byte.c: Likewise.
3224         * testsuite/libffi.call/cls_64byte.c: Likewise.
3225         * testsuite/libffi.call/cls_6byte.c: Likewise.
3226         * testsuite/libffi.call/cls_7byte.c: Likewise.
3227         * testsuite/libffi.call/cls_8byte.c: Likewise.
3228         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3229         * testsuite/libffi.call/cls_9byte2.c: Likewise.
3230         * testsuite/libffi.call/cls_double.c: Likewise.
3231         * testsuite/libffi.call/cls_float.c: Likewise.
3232         * testsuite/libffi.call/cls_schar.c: Likewise.
3233         * testsuite/libffi.call/cls_sint.c: Likewise.
3234         * testsuite/libffi.call/cls_sshort.c: Likewise.
3235         * testsuite/libffi.call/cls_uchar.c: Likewise.
3236         * testsuite/libffi.call/cls_uint.c: Likewise.
3237         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3238         * testsuite/libffi.call/cls_ushort.c: Likewise.
3239         * testsuite/libffi.call/nested_struct.c: Likewise.
3240         * testsuite/libffi.call/nested_struct1.c: Likewise.
3241         * testsuite/libffi.call/nested_struct2.c: Likewise.
3242         * testsuite/libffi.call/nested_struct3.c: Likewise.
3243         * testsuite/libffi.call/problem1.c: Likewise.
3244         * testsuite/libffi.special/unwindtest.cc: Likewise.
3245
3246 2003-11-20  Andreas Tobler  <a.tobler@schweiz.ch>
3247
3248         * testsuite/lib/libffi-dg.exp: Make the -lgcc_s conditional.
3249
3250 2003-11-19  Andreas Tobler  <a.tobler@schweiz.ch>
3251
3252         * testsuite/lib/libffi-dg.exp: Add DYLD_LIBRARY_PATH for darwin.
3253         Add -lgcc_s to additional flags.
3254
3255 2003-11-12  Andreas Tobler  <a.tobler@schweiz.ch>
3256
3257         * configure.in, include/Makefile.am: PR libgcj/11147, install
3258         the ffitarget.h header file in a gcc versioned and target
3259         dependent place.
3260         * configure: Regenerated.
3261         * Makefile.in, include/Makefile.in: Likewise.
3262         * testsuite/Makefile.in: Likewise.
3263
3264 2003-11-09  Andreas Tobler  <a.tobler@schweiz.ch>
3265
3266         * testsuite/libffi.call/closure_fn0.c: Print result and check
3267         with dg-output to make debugging easier.
3268         * testsuite/libffi.call/closure_fn1.c: Likewise.
3269         * testsuite/libffi.call/closure_fn2.c: Likewise.
3270         * testsuite/libffi.call/closure_fn3.c: Likewise.
3271         * testsuite/libffi.call/closure_fn4.c: Likewise.
3272         * testsuite/libffi.call/closure_fn5.c: Likewise.
3273         * testsuite/libffi.call/cls_12byte.c: Likewise.
3274         * testsuite/libffi.call/cls_16byte.c: Likewise.
3275         * testsuite/libffi.call/cls_18byte.c: Likewise.
3276         * testsuite/libffi.call/cls_19byte.c: Likewise.
3277         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3278         * testsuite/libffi.call/cls_20byte.c: Likewise.
3279         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3280         * testsuite/libffi.call/cls_24byte.c: Likewise.
3281         * testsuite/libffi.call/cls_2byte.c: Likewise.
3282         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3283         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3284         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3285         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3286         * testsuite/libffi.call/cls_4byte.c: Likewise.
3287         * testsuite/libffi.call/cls_5byte.c: Likewise.
3288         * testsuite/libffi.call/cls_64byte.c: Likewise.
3289         * testsuite/libffi.call/cls_6byte.c: Likewise.
3290         * testsuite/libffi.call/cls_7byte.c: Likewise.
3291         * testsuite/libffi.call/cls_8byte.c: Likewise.
3292         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3293         * testsuite/libffi.call/cls_9byte2.c: Likewise.
3294         * testsuite/libffi.call/cls_double.c: Likewise.
3295         * testsuite/libffi.call/cls_float.c: Likewise.
3296         * testsuite/libffi.call/cls_schar.c: Likewise.
3297         * testsuite/libffi.call/cls_sint.c: Likewise.
3298         * testsuite/libffi.call/cls_sshort.c: Likewise.
3299         * testsuite/libffi.call/cls_uchar.c: Likewise.
3300         * testsuite/libffi.call/cls_uint.c: Likewise.
3301         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3302         * testsuite/libffi.call/cls_ushort.c: Likewise.
3303         * testsuite/libffi.call/problem1.c: Likewise.
3304
3305         * testsuite/libffi.special/unwindtest.cc: Make ffi_closure
3306         static.
3307
3308 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
3309
3310         * testsuite/libffi.call/cls_9byte2.c: New test case.
3311         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3312         * testsuite/libffi.call/cls_64byte.c: Likewise.
3313         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3314         * testsuite/libffi.call/cls_19byte.c: Likewise.
3315         * testsuite/libffi.call/cls_18byte.c: Likewise.
3316         * testsuite/libffi.call/closure_fn4.c: Likewise.
3317         * testsuite/libffi.call/closure_fn5.c: Likewise.
3318         * testsuite/libffi.call/cls_schar.c: Likewise.
3319         * testsuite/libffi.call/cls_sint.c: Likewise.
3320         * testsuite/libffi.call/cls_sshort.c: Likewise.
3321         * testsuite/libffi.call/nested_struct2.c: Likewise.
3322         * testsuite/libffi.call/nested_struct3.c: Likewise.
3323
3324 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
3325
3326         * testsuite/libffi.call/cls_double.c: Do a check on the result.
3327         * testsuite/libffi.call/cls_uchar.c: Likewise.
3328         * testsuite/libffi.call/cls_uint.c: Likewise.
3329         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3330         * testsuite/libffi.call/cls_ushort.c: Likewise.
3331         * testsuite/libffi.call/return_sc.c: Cleanup whitespaces.
3332
3333 2003-11-06  Andreas Tobler  <a.tobler@schweiz.ch>
3334
3335         * src/prep_cif.c (ffi_prep_cif): Move the validity check after
3336         the initialization.
3337
3338 2003-10-23  Andreas Tobler  <a.tobler@schweiz.ch>
3339
3340         * src/java_raw_api.c (ffi_java_ptrarray_to_raw): Replace
3341         FFI_ASSERT(FALSE) with FFI_ASSERT(0).
3342
3343 2003-10-22  David Daney  <ddaney@avtrex.com>
3344
3345         * src/mips/ffitarget.h: Replace undefined UINT32 and friends with
3346         __attribute__((__mode__(__SI__))) and friends.
3347
3348 2003-10-22  Andreas Schwab  <schwab@suse.de>
3349
3350         * src/ia64/ffi.c: Replace FALSE/TRUE with false/true.
3351
3352 2003-10-21  Andreas Tobler  <a.tobler@schweiz.ch>
3353
3354         * configure.in: AC_LINK_FILES(ffitarget.h).
3355         * configure: Regenerate.
3356         * Makefile.in: Likewise.
3357         * include/Makefile.in: Likewise.
3358         * testsuite/Makefile.in: Likewise.
3359         * fficonfig.h.in: Likewise.
3360
3361 2003-10-21  Paolo Bonzini  <bonzini@gnu.org>
3362             Richard Henderson  <rth@redhat.com>
3363
3364         Avoid that ffi.h includes fficonfig.h.
3365
3366         * Makefile.am (EXTRA_DIST): Include ffitarget.h files
3367         (TARGET_SRC_MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3368         (TARGET_SRC_MIPS_SGI): Removed.
3369         (MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3370         (MIPS_SGI): Removed.
3371         (CLEANFILES): Removed.
3372         (mostlyclean-am, clean-am, mostlyclean-sub, clean-sub): New
3373         targets.
3374         * acconfig.h: Removed.
3375         * configure.in: Compute sizeofs only for double and long double.
3376         Use them to define and subst HAVE_LONG_DOUBLE.  Include comments
3377         into AC_DEFINE instead of using acconfig.h.  Create
3378         include/ffitarget.h instead of include/fficonfig.h.  Rename
3379         MIPS_GCC to MIPS_IRIX, drop MIPS_SGI since we are in gcc's tree.
3380         AC_DEFINE EH_FRAME_FLAGS.
3381         * include/Makefile.am (DISTCLEANFILES): New automake macro.
3382         (hack_DATA): Add ffitarget.h.
3383         * include/ffi.h.in: Remove all system specific definitions.
3384         Declare raw API even if it is not installed, why bother?
3385         Use limits.h instead of SIZEOF_* to define ffi_type_*.  Do
3386         not define EH_FRAME_FLAGS, it is in fficonfig.h now.  Include
3387         ffitarget.h instead of fficonfig.h.  Remove ALIGN macro.
3388         (UINT_ARG, INT_ARG): Removed, use ffi_arg and ffi_sarg instead.
3389         * include/ffi_common.h (bool): Do not define.
3390         (ffi_assert): Accept failed assertion.
3391         (ffi_type_test): Return void and accept file/line.
3392         (FFI_ASSERT): Pass stringized failed assertion.
3393         (FFI_ASSERT_AT): New macro.
3394         (FFI_ASSERT_VALID_TYPE): New macro.
3395         (UINT8, SINT8, UINT16, SINT16, UINT32, SINT32,
3396         UINT64, SINT64): Define here with gcc's __attribute__ macro
3397         instead of in ffi.h
3398         (FLOAT32, ALIGN): Define here instead of in ffi.h
3399         * include/ffi-mips.h: Removed.  Its content moved to
3400         src/mips/ffitarget.h after separating assembly and C sections.
3401         * src/alpha/ffi.c, src/alpha/ffi.c, src/java_raw_api.c
3402         src/prep_cif.c, src/raw_api.c, src/ia64/ffi.c,
3403         src/mips/ffi.c, src/mips/n32.S, src/mips/o32.S,
3404         src/mips/ffitarget.h, src/sparc/ffi.c, src/x86/ffi64.c:
3405         SIZEOF_ARG -> FFI_SIZEOF_ARG.
3406         * src/ia64/ffi.c: Include stdbool.h (provided by GCC 2.95+).
3407         * src/debug.c (ffi_assert): Accept stringized failed assertion.
3408         (ffi_type_test): Rewritten.
3409         * src/prep-cif.c (initialize_aggregate, ffi_prep_cif): Call
3410         FFI_ASSERT_VALID_TYPE.
3411         * src/alpha/ffitarget.h, src/arm/ffitarget.h,
3412         src/ia64/ffitarget.h, src/m68k/ffitarget.h,
3413         src/mips/ffitarget.h, src/powerpc/ffitarget.h,
3414         src/s390/ffitarget.h, src/sh/ffitarget.h,
3415         src/sh64/ffitarget.h, src/sparc/ffitarget.h,
3416         src/x86/ffitarget.h: New files.
3417         * src/alpha/osf.S, src/arm/sysv.S, src/ia64/unix.S,
3418         src/m68k/sysv.S, src/mips/n32.S, src/mips/o32.S,
3419         src/powerpc/aix.S, src/powerpc/darwin.S,
3420         src/powerpc/ffi_darwin.c, src/powerpc/linux64.S,
3421         src/powerpc/linux64_closure.S, src/powerpc/ppc_closure.S,
3422         src/powerpc/sysv.S, src/s390/sysv.S, src/sh/sysv.S,
3423         src/sh64/sysv.S, src/sparc/v8.S, src/sparc/v9.S,
3424         src/x86/sysv.S, src/x86/unix64.S, src/x86/win32.S:
3425         include fficonfig.h
3426
3427 2003-10-20  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3428
3429         * src/mips/ffi.c: Use _ABIN32, _ABIO32 instead of external
3430         _MIPS_SIM_NABI32, _MIPS_SIM_ABI32.
3431
3432 2003-10-19  Andreas Tobler  <a.tobler@schweiz.ch>
3433
3434         * src/powerpc/ffi_darwin.c (ffi_prep_args): Declare bytes again.
3435         Used when FFI_DEBUG = 1.
3436
3437 2003-10-14  Alan Modra  <amodra@bigpond.net.au>
3438
3439         * src/types.c (double, longdouble): Default POWERPC64 to 8 byte size
3440         and align.
3441
3442 2003-10-06  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3443
3444         * include/ffi_mips.h: Define FFI_MIPS_N32 for N32/N64 ABIs,
3445         FFI_MIPS_O32 for O32 ABI.
3446
3447 2003-10-01  Andreas Tobler  <a.tobler@schweiz.ch>
3448
3449         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_64 for
3450         SPARC64. Cleanup whitespaces.
3451
3452 2003-09-19  Andreas Tobler  <a.tobler@schweiz.ch>