Ensure that users don't include ffitarget.h directly
[libffi.git] / .pc / fix-xfails / ChangeLog
1 2012-02-10  Kai Tietz  <ktietz@redhat.com>
2
3         * configure.ac (AM_LTLDFLAGS): Add -no-undefine for x64
4         windows target.
5         * configure: Regenerated.
6         
7 2012-02-08  Kai Tietz  <ktietz@redhat.com>
8
9         * src/prep_cif.c (ffi_prep_cif): Allow for X86_WIN32
10         also FFI_THISCALL.
11         * src/x86/ffi.c (ffi_closure_THISCALL): Add prototype.
12         (FFI_INIT_TRAMPOLINE_THISCALL): New trampoline code.
13         (ffi_prep_closure_loc): Add FFI_THISCALL support.
14         * src/x86/ffitarget.h (FFI_TRAMPOLINE_SIZE): Adjust size.
15         * src/x86/win32.S (ffi_closure_THISCALL): New closure code
16         for thiscall-calling convention.
17         * testsuite/libffi.call/closure_thiscall.c: New test.
18
19 2012-01-28  Kai Tietz  <ktietz@redhat.com>
20
21         * src/libffi/src/x86/ffi.c (ffi_call_win32): Add new
22         argument to prototype for specify calling-convention.
23         (ffi_call): Add support for stdcall/thiscall convention.
24         (ffi_prep_args): Likewise.
25         (ffi_raw_call): Likewise.
26         * src/x86/ffitarget.h (ffi_abi): Add FFI_THISCALL and
27         FFI_FASTCALL.
28         * src/x86/win32.S (_ffi_call_win32): Add support for
29         fastcall/thiscall calling-convention calls.
30         * testsuite/libffi.call/fastthis1_win32.c: New test.
31         * testsuite/libffi.call/fastthis2_win32.c: New test.
32         * testsuite/libffi.call/fastthis3_win32.c: New test.
33         * testsuite/libffi.call/strlen2_win32.c: New test.
34         * testsuite/libffi.call/many2_win32.c: New test.
35         * testsuite/libffi.call/struct1_win32.c: New test.
36         * testsuite/libffi.call/struct2_win32.c: New test.
37
38 2012-01-23  Andreas Schwab  <schwab@linux-m68k.org>
39
40         * src/m68k/sysv.S (ffi_call_SYSV): Properly test for plain
41         mc68000.  Test for __HAVE_68881__ in addition to __MC68881__.
42
43 2012-01-19  Jakub Jelinek  <jakub@redhat.com>
44
45         PR rtl-optimization/48496
46         * src/ia64/ffi.c (ffi_call): Fix up aliasing violations.
47
48 2012-01-09  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
49
50         * configure.ac (i?86-*-*): Set TARGET to X86_64.
51         * configure: Regenerate.
52
53 2011-12-07  Andrew Pinski  <apinski@cavium.com>
54
55         PR libffi/50051
56         * src/mips/n32.S: Add ".set mips4".
57
58 2011-11-21  Andreas Tobler  <andreast@fgznet.ch>
59
60         * configure: Regenerate.
61
62 2011-11-10  Richard Henderson  <rth@redhat.com>
63
64         * configure.ac (GCC_AS_CFI_PSEUDO_OP): Use it instead of inline check.
65         * configure, aclocal.m4: Rebuild.
66
67 2011-09-04  Iain Sandoe  <iains@gcc.gnu.org>
68
69         PR libffi/49594
70         * src/powerpc/darwin_closure.S (stubs): Make the stub binding
71         helper reference track the architecture pointer size.
72
73 2011-08-25  Andrew Haley  <aph@redhat.com>
74
75         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Remove hard-coded assembly
76         instructions.
77         * src/arm/sysv.S (ffi_arm_trampoline): Put them here instead.
78
79 2011-07-11  Andrew Haley  <aph@redhat.com>
80
81         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Clear icache.
82
83 2011-06-29  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
84
85         * testsuite/libffi.call/cls_double_va.c: Move PR number to comment.
86         * testsuite/libffi.call/cls_longdouble_va.c: Likewise.
87
88 2011-06-29  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
89
90         PR libffi/46660
91         * testsuite/libffi.call/cls_double_va.c: xfail dg-output on
92         mips-sgi-irix6*.
93         * testsuite/libffi.call/cls_longdouble_va.c: Likewise.
94
95 2011-06-14  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
96
97         * testsuite/libffi.call/huge_struct.c (test_large_fn): Use PRIu8,
98         PRId8 instead of %hhu, %hhd.
99         * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRId8,
100         PRIu8): Define.
101         [__sgi__] (PRId8, PRIu8): Define.
102
103 2011-04-29  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
104
105         * src/alpha/osf.S (UA_SI, FDE_ENCODING, FDE_ENCODE, FDE_ARANGE):
106         Define.
107         Use them to handle ELF vs. ECOFF differences.
108         [__osf__] (_GLOBAL__F_ffi_call_osf): Define.
109
110 2011-02-13  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
111
112         * configure: Regenerate.
113
114 2011-02-09  Anthony Green <green@moxielogic.com>
115
116         * README: Mention ARM iOS.
117
118 2011-02-09  Landon Fuller <landonf@macports.org>
119
120         * Makefile.am (EXTRA_DIST): Add build-ios.sh, src/arm/gentramp.sh,
121         src/arm/trampoline.S.
122         (nodist_libffi_la_SOURCES): Add src/arc/trampoline.S.
123         * configure.ac (FFI_EXEC_TRAMPOLINE_TABLE): Define.
124         * src/arm/ffi.c (ffi_trampoline_table)
125         (ffi_closure_trampoline_table_page, ffi_trampoline_table_entry)
126         (FFI_TRAMPOLINE_CODELOC_CONFIG, FFI_TRAMPOLINE_CONFIG_PAGE_OFFSET)
127         (FFI_TRAMPOLINE_COUNT, ffi_trampoline_lock, ffi_trampoline_tables)
128         (ffi_trampoline_table_alloc, ffi_closure_alloc, ffi_closure_free):
129         Define for FFI_EXEC_TRAMPOLINE_TABLE case (iOS).
130         (ffi_prep_closure_loc): Handl FFI_EXEC_TRAMPOLINE_TABLE case
131         separately.
132         * src/arm/sysv.S: Handle Apple iOS host.
133         * src/closures.c: Handle FFI_EXEC_TRAMPOLINE_TABLE case.
134         * build-ios.sh: New file.
135         * fficonfig.h.in, configure, Makefile.in: Rebuilt.
136
137 2011-02-08  Oren Held  <orenhe@il.ibm.com>
138
139         * src/dlmalloc.c (_STRUCT_MALLINFO): Define in order to avoid
140         redefinition of mallinfo on HP-UX.
141
142 2011-02-08  Ginn Chen  <ginn.chen@oracle.com>
143
144         * src/sparc/ffi.c (ffi_call): Make compatible with Solaris Studio
145         aggregate return ABI.  Flush cache.
146         (ffi_prep_closure_loc): Flush cache.
147
148 2011-02-11  Anthony Green  <green@moxielogic.com>
149
150         From Tom Honermann <tom.honermann@oracle.com>:
151         * src/powerpc/aix.S (ffi_call_AIX): Support for xlc toolchain on
152         AIX.  Declare .ffi_prep_args.  Insert nops after branch
153         instructions so that the AIX linker can insert TOC reload
154         instructions.
155         * src/powerpc/aix_closure.S: Declare .ffi_closure_helper_DARWIN.
156
157 2011-02-08  Ed  <ed@kdtc.net>
158
159         * src/powerpc/asm.h: Fix grammar nit in comment.
160
161 2011-02-08  Uli Link  <ul.mcamafia@linkitup.de>
162
163         * include/ffi.h.in (FFI_64_BIT_MAX): Define and use.
164
165 2011-02-09  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
166
167         PR libffi/46661
168         * testsuite/libffi.call/cls_pointer.c (main): Cast void * to
169         uintptr_t first.
170         * testsuite/libffi.call/cls_pointer_stack.c (main): Likewise.
171
172 2011-02-08  Rafael Avila de Espindola  <respindola@mozilla.com>
173
174         * configure.ac: Fix x86 test for pc related relocs.
175         * configure: Rebuilt.
176
177 2011-02-07  Joel Sherrill <joel.sherrill@oarcorp.com>
178
179         * libffi/src/m68k/ffi.c: Add RTEMS support for cache flushing.
180         Handle case when CPU variant does not have long double support.
181         * libffi/src/m68k/sysv.S: Add support for mc68000, Coldfire,
182         and cores with soft floating point.
183
184 2011-02-07  Joel Sherrill <joel.sherrill@oarcorp.com>
185
186         * configure.ac: Add mips*-*-rtems* support.
187         * configure: Regenerate.
188         * src/mips/ffitarget.h: Ensure needed constants are available
189         for targets which do not have sgidefs.h.
190
191 2011-01-26  Dave Korn  <dave.korn.cygwin@gmail.com>
192
193         PR target/40125
194         * configure.ac (AM_LTLDFLAGS): Add -bindir option for windows DLLs.
195         * configure: Regenerate.
196
197 2010-12-18  Iain Sandoe  <iains@gcc.gnu.org>
198
199         PR libffi/29152
200         PR libffi/42378
201         * src/powerpc/darwin_closure.S: Provide Darwin64 implementation,
202         update comments.
203         * src/powerpc/ffitarget.h (POWERPC_DARWIN64): New,
204         (FFI_TRAMPOLINE_SIZE): Update for Darwin64.
205         * src/powerpc/darwin.S: Provide Darwin64 implementation,
206         update comments.
207         * src/powerpc/ffi_darwin.c: Likewise.
208
209 2010-12-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
210
211         * configure.ac (libffi_cv_as_ascii_pseudo_op): Use double
212         backslashes.
213         (libffi_cv_as_string_pseudo_op): Likewise.
214         * configure: Regenerate.
215
216 2010-12-03  Chung-Lin Tang  <cltang@codesourcery.com>
217
218         * src/arm/sysv.S (ffi_closure_SYSV): Add UNWIND to .pad directive.
219         (ffi_closure_VFP): Same.
220         (ffi_call_VFP): Move down to before ffi_closure_VFP. Add '.fpu vfp'
221         directive.
222
223 2010-12-01  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
224
225         * testsuite/libffi.call/ffitest.h [__sgi] (PRId64, PRIu64): Define.
226         (PRIuPTR): Define.
227
228 2010-11-29  Richard Henderson  <rth@redhat.com>
229             Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
230
231         * src/x86/sysv.S (FDE_ENCODING, FDE_ENCODE): Define.
232         (.eh_frame): Use FDE_ENCODING.
233         (.LASFDE1, .LASFDE2, LASFDE3): Simplify with FDE_ENCODE.
234
235 2010-11-22  Jacek Caban <jacek@codeweavers.com>
236
237         * configure.ac: Check for symbol underscores on mingw-w64.
238         * configure: Rebuilt.
239         * src/x86/win64.S: Correctly access extern symbols in respect to
240         underscores.
241
242 2010-11-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
243
244         * testsuite/lib/libffi-dg.exp: Rename ...
245         * testsuite/lib/libffi.exp: ... to this.
246         * libffi/testsuite/libffi.call/call.exp: Don't load libffi-dg.exp.
247         * libffi/testsuite/libffi.special/special.exp: Likewise.
248
249 2010-10-28  Chung-Lin Tang  <cltang@codesourcery.com>
250
251         * src/arm/ffi.c (ffi_prep_args): Add VFP register argument handling
252         code, new parameter, and return value. Update comments.
253         (ffi_prep_cif_machdep): Add case for VFP struct return values. Add
254         call to layout_vfp_args().
255         (ffi_call_SYSV): Update declaration.
256         (ffi_call_VFP): New declaration.
257         (ffi_call): Add VFP struct return conditions. Call ffi_call_VFP()
258         when ABI is FFI_VFP.
259         (ffi_closure_VFP): New declaration.
260         (ffi_closure_SYSV_inner): Add new vfp_args parameter, update call to
261         ffi_prep_incoming_args_SYSV().
262         (ffi_prep_incoming_args_SYSV): Update parameters. Add VFP argument
263         case handling.
264         (ffi_prep_closure_loc): Pass ffi_closure_VFP to trampoline
265         construction under VFP hard-float.
266         (rec_vfp_type_p): New function.
267         (vfp_type_p): Same.
268         (place_vfp_arg): Same.
269         (layout_vfp_args): Same.
270         * src/arm/ffitarget.h (ffi_abi): Add FFI_VFP. Define FFI_DEFAULT_ABI
271         based on __ARM_PCS_VFP.
272         (FFI_EXTRA_CIF_FIELDS): Define for adding VFP hard-float specific
273         fields.
274         (FFI_TYPE_STRUCT_VFP_FLOAT): Define internally used type code.
275         (FFI_TYPE_STRUCT_VFP_DOUBLE): Same.
276         * src/arm/sysv.S (ffi_call_SYSV): Change call of ffi_prep_args() to
277         direct call. Move function pointer load upwards.
278         (ffi_call_VFP): New function.
279         (ffi_closure_VFP): Same.
280
281         * testsuite/lib/libffi-dg.exp (check-flags): New function.
282         (dg-skip-if): New function.
283         * testsuite/libffi.call/cls_double_va.c: Skip if target is arm*-*-*
284         and compiler options include -mfloat-abi=hard.
285         * testsuite/libffi.call/cls_longdouble_va.c: Same.
286
287 2010-10-01  Jakub Jelinek  <jakub@redhat.com>
288
289         PR libffi/45677
290         * src/x86/ffi64.c (ffi_prep_cif_machdep): Ensure cif->bytes is
291         a multiple of 8.
292         * testsuite/libffi.call/many2.c: New test.
293
294 2010-08-20  Mark Wielaard  <mjw@redhat.com>
295
296         * src/closures.c (open_temp_exec_file_mnt): Check if getmntent_r
297         returns NULL.
298
299 2010-08-09  Andreas Tobler  <andreast@fgznet.ch>
300
301         * configure.ac: Add target powerpc64-*-freebsd*.
302         * configure: Regenerate.
303         * testsuite/libffi.call/cls_align_longdouble_split.c: Pass
304         -mlong-double-128 only to linux targets.
305         * testsuite/libffi.call/cls_align_longdouble_split2.c: Likewise.
306         * testsuite/libffi.call/cls_longdouble.c: Likewise.
307         * testsuite/libffi.call/huge_struct.c: Likewise.
308
309 2010-08-05  Dan Witte  <dwitte@mozilla.com>
310
311         * Makefile.am: Pass FFI_DEBUG define to msvcc.sh for linking to the
312         debug CRT when --enable-debug is given.
313         * configure.ac: Define it.
314         * msvcc.sh: Translate -g and -DFFI_DEBUG appropriately.
315
316 2010-08-04  Dan Witte  <dwitte@mozilla.com>
317
318         * src/x86/ffitarget.h: Add X86_ANY define for all x86/x86_64
319         platforms.
320         * src/x86/ffi.c: Remove redundant ifdef checks.
321         * src/prep_cif.c: Push stack space computation into src/x86/ffi.c
322         for X86_ANY so return value space doesn't get added twice.
323
324 2010-08-03  Neil Rashbrooke <neil@parkwaycc.co.uk>
325
326         * msvcc.sh: Don't pass -safeseh to ml64 because behavior is buggy.
327
328 2010-07-22  Dan Witte  <dwitte@mozilla.com>
329
330         * src/*/ffitarget.h: Make FFI_LAST_ABI one past the last valid ABI.
331         * src/prep_cif.c: Fix ABI assertion.
332         * src/cris/ffi.c: Ditto.
333
334 2010-07-10  Evan Phoenix  <evan@fallingsnow.net>
335
336         * src/closures.c (selinux_enabled_check): Fix strncmp usage bug.
337
338 2010-07-07  Dan HorĂ¡k <dan@danny.cz>
339
340         * include/ffi.h.in: Protect #define with #ifndef.
341         * src/powerpc/ffitarget.h: Ditto.
342         * src/s390/ffitarget.h: Ditto.
343         * src/sparc/ffitarget.h: Ditto.
344
345 2010-07-07   Neil Roberts <neil@linux.intel.com>
346
347         * src/x86/sysv.S (ffi_call_SYSV): Align the stack pointer to
348         16-bytes.
349
350 2010-07-02  Jakub Jelinek  <jakub@redhat.com>
351
352         * Makefile.am (AM_MAKEFLAGS): Pass also mandir to submakes.
353         * Makefile.in: Regenerated.
354
355 2010-05-19  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
356
357         * configure.ac (libffi_cv_as_x86_pcrel): Check for illegal in as
358         output, too.
359         (libffi_cv_as_ascii_pseudo_op): Check for .ascii.
360         (libffi_cv_as_string_pseudo_op): Check for .string.
361         * configure: Regenerate.
362         * fficonfig.h.in: Regenerate.
363         * src/x86/sysv.S (.eh_frame): Use .ascii, .string or error.
364
365 2010-05-11  Dan Witte  <dwitte@mozilla.com>
366
367         * doc/libffi.tex: Document previous change.
368
369 2010-05-11  Makoto Kato <m_kato@ga2.so-net.ne.jp>
370
371         * src/x86/ffi.c (ffi_call): Don't copy structs passed by value.
372
373 2010-05-05  Michael Kohler <michaelkohler@live.com>
374
375         * src/dlmalloc.c (dlfree): Fix spelling.
376         * src/ia64/ffi.c (ffi_prep_cif_machdep): Ditto.
377         * configure.ac: Ditto.
378         * configure: Rebuilt.
379
380 2010-04-13  Dan Witte  <dwitte@mozilla.com>
381
382         * msvcc.sh: Build with -W3 instead of -Wall.
383         * src/powerpc/ffi_darwin.c: Remove build warnings.
384         * src/x86/ffi.c: Ditto.
385         * src/x86/ffitarget.h: Ditto.
386
387 2010-04-12  Dan Witte  <dwitte@mozilla.com>
388             Walter Meinl <wuno@lsvw.de>
389
390         * configure.ac: Add OS/2 support.
391         * configure: Rebuilt.
392         * src/closures.c: Ditto.
393         * src/dlmalloc.c: Ditto.
394         * src/x86/win32.S: Ditto.
395
396 2010-04-07  Jakub Jelinek  <jakub@redhat.com>
397
398         * testsuite/libffi.call/err_bad_abi.c: Remove unused args variable.
399
400 2010-04-02  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
401
402         * Makefile.in: Regenerate.
403         * aclocal.m4: Regenerate.
404         * include/Makefile.in: Regenerate.
405         * man/Makefile.in: Regenerate.
406         * testsuite/Makefile.in: Regenerate.
407
408 2010-03-30  Dan Witte  <dwitte@mozilla.com>
409
410         * msvcc.sh: Disable build warnings.
411         * README (tested): Clarify windows build procedure.
412
413 2010-03-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
414
415         * configure.ac (libffi_cv_as_x86_64_unwind_section_type): New test.
416         * configure: Regenerate.
417         * fficonfig.h.in: Regenerate.
418         * libffi/src/x86/unix64.S (.eh_frame)
419         [HAVE_AS_X86_64_UNWIND_SECTION_TYPE]: Use @unwind section type.
420
421 2010-03-14  Matthias Klose  <doko@ubuntu.com>
422
423         * src/x86/ffi64.c: Fix typo in comment.
424         * src/x86/ffi.c: Use /* ... */ comment style.
425
426 2010-02-24  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
427
428         * doc/libffi.texi (The Closure API): Fix typo.
429         * doc/libffi.info: Remove.
430
431 2010-02-15  Matthias Klose  <doko@ubuntu.com>
432
433         * src/arm/sysv.S (__ARM_ARCH__): Define for processor
434         __ARM_ARCH_7EM__.
435
436 2010-01-15  Anthony Green  <green@redhat.com>
437
438         * README: Add notes on building with Microsoft Visual C++.
439
440 2010-01-15  Daniel Witte  <dwitte@mozilla.com>
441
442         * msvcc.sh: New file.
443
444         * src/x86/win32.S: Port assembly routines to MSVC and #ifdef.
445         * src/x86/ffi.c: Tweak function declaration and remove excess
446         parens.
447         * include/ffi.h.in: Add __declspec(align(8)) to typedef struct
448         ffi_closure.
449
450         * src/x86/ffi.c: Merge ffi_call_SYSV and ffi_call_STDCALL into new
451         function ffi_call_win32 on X86_WIN32.
452         * src/x86/win32.S (ffi_call_SYSV): Rename to ffi_call_win32.
453         (ffi_call_STDCALL): Remove.
454
455         * src/prep_cif.c (ffi_prep_cif): Move stack space allocation code
456         to ffi_prep_cif_machdep for x86.
457         * src/x86/ffi.c (ffi_prep_cif_machdep): To here.
458
459 2010-01-15  Oliver Kiddle  <okiddle@yahoo.co.uk>
460
461         * src/x86/ffitarget.h (ffi_abi): Check for __i386 and __amd64 for
462         Sun Studio compiler compatibility.
463
464 2010-01-12  Conrad Irwin <conrad.irwin@gmail.com>
465
466         * doc/libffi.texi: Add closure example.
467
468 2010-01-07  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
469
470         PR libffi/40701
471         * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRIdLL,
472         PRIuLL, PRId64, PRIu64, PRIuPTR): Define.
473         * testsuite/libffi.call/cls_align_sint64.c: Add -Wno-format on
474         alpha*-dec-osf*.
475         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
476         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
477         * testsuite/libffi.call/return_ll1.c: Likewise.
478         * testsuite/libffi.call/stret_medium2.c: Likewise.
479         * testsuite/libffi.special/ffitestcxx.h (allocate_mmap): Cast
480         MAP_FAILED to char *.
481
482 2010-01-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
483
484         * src/mips/n32.S: Use .abicalls and .eh_frame with __GNUC__.
485
486 2009-12-31  Anthony Green  <green@redhat.com>
487
488         * README: Update for libffi 3.0.9.
489
490 2009-12-27  Matthias Klose  <doko@ubuntu.com>
491
492         * configure.ac (HAVE_LONG_DOUBLE): Define for mips when
493         appropriate.
494         * configure: Rebuilt.
495
496 2009-12-26  Anthony Green  <green@redhat.com>
497
498         * testsuite/libffi.call/cls_longdouble_va.c: Mark as xfail for
499         avr32*-*-*.
500         * testsuite/libffi.call/cls_double_va.c: Ditto.
501
502 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
503
504         * testsuite/libffi.call/ffitest.h: Conditionally include stdint.h
505         and inttypes.h.
506         * testsuite/libffi.special/unwindtest.cc: Ditto.
507
508 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
509
510         * configure.ac: Add amd64-*-openbsd*.
511         * configure: Rebuilt.
512         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Link
513         openbsd programs with -lpthread.
514
515 2009-12-26  Anthony Green  <green@redhat.com>
516
517         * testsuite/libffi.call/cls_double_va.c,
518         testsuite/libffi.call/cls_longdouble.c,
519         testsuite/libffi.call/cls_longdouble_va.c,
520         testsuite/libffi.call/cls_pointer.c,
521         testsuite/libffi.call/cls_pointer_stack.c: Remove xfail for
522         mips*-*-* and arm*-*-*.
523         * testsuite/libffi.call/cls_align_longdouble_split.c,
524         testsuite/libffi.call/cls_align_longdouble_split2.c,
525         testsuite/libffi.call/stret_medium2.c,
526         testsuite/libffi.call/stret_medium.c,
527         testsuite/libffi.call/stret_large.c,
528         testsuite/libffi.call/stret_large2.c: Remove xfail for arm*-*-*.
529
530 2009-12-31  Kay Tietz  <ktietz70@googlemail.com>
531
532         * testsuite/libffi.call/ffitest.h,
533         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRuLL): Fix
534         definitions.
535
536 2009-12-31  Carlo Bramini  <carlo.bramix@libero.it>
537
538         * configure.ac (AM_LTLDFLAGS): Define for windows hosts.
539         * Makefile.am (libffi_la_LDFLAGS): Add AM_LTLDFLAGS.
540         * configure: Rebuilt.
541         * Makefile.in: Rebuilt.
542
543 2009-12-31  Anthony Green  <green@redhat.com>
544             Blake Chaffin.
545
546         * testsuite/libffi.call/huge_struct.c: New test case from Blake
547         Chaffin @ Apple.
548
549 2009-12-28  David Edelsohn  <edelsohn@gnu.org>
550
551         * src/powerpc/ffi_darwin.c (ffi_prep_args): Copy abi and nargs to
552         local variables.
553         (aix_adjust_aggregate_sizes): New function.
554         (ffi_prep_cif_machdep): Call it.
555
556 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
557
558         * configure.ac: Define FFI_MMAP_EXEC_WRIT for the given targets.
559         * configure: Regenerate.
560         * fficonfig.h.in: Likewise.
561         * src/closures.c: Remove the FFI_MMAP_EXEC_WRIT definition for
562         Solaris/x86.
563
564 2009-12-26  Andreas Schwab  <schwab@linux-m68k.org>
565
566         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Advance intarg_count
567         when a float arguments is passed in memory.
568         (ffi_closure_helper_SYSV): Mark general registers as used up when
569         a 64bit or soft-float long double argument is passed in memory.
570
571 2009-12-25  Matthias Klose  <doko@ubuntu.com>
572
573         * man/ffi_call.3: Fix #include in examples.
574         * doc/libffi.texi: Add dircategory.
575
576 2009-12-25  Frank Everdij <f.p.x.everdij@tudelft.nl>
577
578         * include/ffi.h.in: Placed '__GNUC__' ifdef around
579         '__attribute__((aligned(8)))' in ffi_closure, fixes compile for
580         IRIX MIPSPro c99.
581         * include/ffi_common.h: Added '__sgi' define to non
582         '__attribute__((__mode__()))' integer typedefs.
583         * src/mips/ffi.c (ffi_call, ffi_closure_mips_inner_O32,
584         ffi_closure_mips_inner_N32): Added 'defined(_MIPSEB)' to BE check.
585         (ffi_closure_mips_inner_O32, ffi_closure_mips_inner_N32): Added
586         FFI_LONGDOUBLE support and alignment(N32 only).
587         * src/mips/ffitarget.h: Corrected '#include <sgidefs.h>' for IRIX and
588         fixed non '__attribute__((__mode__()))' integer typedefs.
589         * src/mips/n32.S: Put '#ifdef linux' around '.abicalls' and '.eh_frame'
590         since they are Linux/GNU Assembler specific.
591
592 2009-12-25  Bradley Smith  <brad@brad-smith.co.uk>
593
594         * configure.ac, Makefile.am, src/avr32/ffi.c,
595         src/avr32/ffitarget.h,
596         src/avr32/sysv.S: Add AVR32 port.
597         * configure, Makefile.in: Rebuilt.
598
599 2009-12-21  Andreas Tobler  <a.tobler@schweiz.org>
600
601         * configure.ac: Make i?86 build on FreeBSD and OpenBSD.
602         * configure: Regenerate.
603
604 2009-12-15  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
605
606         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on PA HP-UX.
607
608 2009-12-13  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
609
610         * src/pa/ffi.c (ffi_closure_inner_pa32): Handle FFI_TYPE_LONGDOUBLE
611         type on HP-UX.
612
613 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
614
615         * src/sparc/ffi.c (ffi_closure_sparc_inner_v9): Properly align 'long
616         double' arguments.
617
618 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
619
620         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on Solaris < 10.
621
622 2009-12-10  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
623
624         PR libffi/40700
625         * src/closures.c [X86_64 && __sun__ && __svr4__]
626         (FFI_MMAP_EXEC_WRIT): Define.
627
628 2009-12-08  David Daney  <ddaney@caviumnetworks.com>
629
630         * testsuite/libffi.call/stret_medium.c: Remove xfail for mips*-*-*
631         * testsuite/libffi.call/cls_align_longdouble_split2.c: Same.
632         * testsuite/libffi.call/stret_large.c: Same.
633         * testsuite/libffi.call/cls_align_longdouble_split.c: Same.
634         * testsuite/libffi.call/stret_large2.c: Same.
635         * testsuite/libffi.call/stret_medium2.c: Same.
636
637 2009-12-07  David Edelsohn  <edelsohn@gnu.org>
638
639         * src/powerpc/aix_closure.S (libffi_closure_ASM): Fix tablejump
640         typo.
641
642 2009-12-05  David Edelsohn  <edelsohn@gnu.org>
643
644         * src/powerpc/aix.S: Update AIX32 code to be consistent with AIX64
645         code.
646         * src/powerpc/aix_closure.S: Same.
647
648 2009-12-05  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
649
650         * Makefile.in: Regenerate.
651         * configure: Regenerate.
652         * include/Makefile.in: Regenerate.
653         * man/Makefile.in: Regenerate.
654         * testsuite/Makefile.in: Regenerate.
655
656 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
657
658         * src/powerpc/aix_closure.S: Reorganize 64-bit code to match
659         linux64_closure.S.
660
661 2009-12-04  Uros Bizjak  <ubizjak@gmail.com>
662
663         PR libffi/41908
664         * src/x86/ffi64.c (classify_argument): Update from
665         gcc/config/i386/i386.c.
666         (ffi_closure_unix64_inner): Do not use the address of two consecutive
667         SSE registers directly.
668         * testsuite/libffi.call/cls_dbls_struct.c (main): Remove xfail
669         for x86_64 linux targets.
670
671 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
672
673         * src/powerpc/ffi_darwin.c (ffi_closure_helper_DARWIN): Increment
674         pfr for long double split between fpr13 and stack.
675
676 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
677
678         * src/powerpc/ffi_darwin.c (ffi_prep_args): Increment next_arg and
679         fparg_count twice for long double.
680
681 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
682
683         PR libffi/42243
684         * src/powerpc/ffi_darwin.c (ffi_prep_args): Remove extra parentheses.
685
686 2009-12-03  Uros Bizjak  <ubizjak@gmail.com>
687
688         * testsuite/libffi.call/cls_longdouble_va.c (main): Fix format string.
689         Remove xfails for x86 linux targets.
690
691 2009-12-02  David Edelsohn  <edelsohn@gnu.org>
692
693         * src/powerpc/ffi_darwin.c (ffi_prep_args): Fix typo in INT64
694         case.
695
696 2009-12-01  David Edelsohn  <edelsohn@gnu.org>
697
698         * src/powerpc/aix.S (ffi_call_AIX): Convert to more standard
699         register usage.  Call ffi_prep_args directly.  Add long double
700         return value support.
701         * src/powerpc/ffi_darwin.c (ffi_prep_args): Double arg increment
702         applies to FFI_TYPE_DOUBLE.  Correct fpr_base increment typo.
703         Separate FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases.
704         (ffi_prep_cif_machdep): Only 16 byte stack alignment in 64 bit
705         mode.
706         (ffi_closure_helper_DARWIN): Remove nf and ng counters.  Move temp
707         into case.
708         * src/powerpc/aix_closure.S: Maintain 16 byte stack alignment.
709         Allocate result area between params and FPRs.
710
711 2009-11-30  David Edelsohn  <edelsohn@gnu.org>
712
713         PR target/35484
714         * src/powerpc/ffitarget.h (POWERPC64): Define for PPC64 Linux and
715         AIX64.
716         * src/powerpc/aix.S: Implement AIX64 version.
717         * src/powerpc/aix_closure.S: Implement AIX64 version.
718         (ffi_closure_ASM): Use extsb, lha and displament addresses.
719         * src/powerpc/ffi_darwin.c (ffi_prep_args): Implement AIX64
720         support.
721         (ffi_prep_cif_machdep): Same.
722         (ffi_call): Same.
723         (ffi_closure_helper_DARWIN): Same.
724
725 2009-11-02  Andreas Tobler  <a.tobler@schweiz.org>
726
727         PR libffi/41908
728         * testsuite/libffi.call/testclosure.c: New test.
729
730 2009-09-28  Kai Tietz  <kai.tietz@onevision.com>
731
732         * src/x86/win64.S (_ffi_call_win64 stack): Remove for gnu
733         assembly version use of ___chkstk.
734
735 2009-09-23  Matthias Klose  <doko@ubuntu.com>
736
737         PR libffi/40242, PR libffi/41443
738         * src/arm/sysv.S (__ARM_ARCH__): Define for processors
739         __ARM_ARCH_6T2__, __ARM_ARCH_6M__, __ARM_ARCH_7__,
740         __ARM_ARCH_7A__, __ARM_ARCH_7R__, __ARM_ARCH_7M__.
741         Change the conditionals to __SOFTFP__ || __ARM_EABI__
742         for -mfloat-abi=softfp to work.
743
744 2009-09-17  Loren J. Rittle  <ljrittle@acm.org>
745
746         PR testsuite/32843 (strikes again)
747         * src/x86/ffi.c (ffi_prep_cif_machdep): Add X86_FREEBSD to
748         enable proper extension on char and short.
749
750 2009-09-15  David Daney  <ddaney@caviumnetworks.com>
751
752         * src/java_raw_api.c (ffi_java_raw_to_rvalue): Remove special
753         handling for FFI_TYPE_POINTER.
754         * src/mips/ffitarget.h (FFI_TYPE_STRUCT_D_SOFT,
755         FFI_TYPE_STRUCT_F_SOFT, FFI_TYPE_STRUCT_DD_SOFT,
756         FFI_TYPE_STRUCT_FF_SOFT, FFI_TYPE_STRUCT_FD_SOFT,
757         FFI_TYPE_STRUCT_DF_SOFT, FFI_TYPE_STRUCT_SOFT): New defines.
758         (FFI_N32_SOFT_FLOAT, FFI_N64_SOFT_FLOAT): New ffi_abi enumerations.
759         (enum ffi_abi): Set FFI_DEFAULT_ABI for soft-float.
760         * src/mips/n32.S (ffi_call_N32): Add handling for soft-float
761         structure and pointer returns.
762         (ffi_closure_N32): Add handling for pointer returns.
763         * src/mips/ffi.c (ffi_prep_args, calc_n32_struct_flags,
764         calc_n32_return_struct_flags): Handle soft-float.
765         (ffi_prep_cif_machdep):  Handle soft-float, fix pointer handling.
766         (ffi_call_N32): Declare proper argument types.
767         (ffi_call, copy_struct_N32, ffi_closure_mips_inner_N32): Handle
768         soft-float.
769
770 2009-08-24  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
771
772         * configure.ac (AC_PREREQ): Bump to 2.64.
773
774 2009-08-22  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
775
776         * Makefile.am (install-html, install-pdf): Remove.
777         * Makefile.in: Regenerate.
778
779         * Makefile.in: Regenerate.
780         * aclocal.m4: Regenerate.
781         * configure: Regenerate.
782         * fficonfig.h.in: Regenerate.
783         * include/Makefile.in: Regenerate.
784         * man/Makefile.in: Regenerate.
785         * testsuite/Makefile.in: Regenerate.
786
787 2009-07-30  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
788
789         * configure.ac (_AC_ARG_VAR_PRECIOUS): Use m4_rename_force.
790
791 2009-07-24  Dave Korn  <dave.korn.cygwin@gmail.com>
792
793         PR libffi/40807
794         * src/x86/ffi.c (ffi_prep_cif_machdep): Also use sign/zero-extending
795         return types for X86_WIN32.
796         * src/x86/win32.S (_ffi_call_SYSV): Handle omitted return types.
797         (_ffi_call_STDCALL, _ffi_closure_SYSV, _ffi_closure_raw_SYSV,
798         _ffi_closure_STDCALL): Likewise.
799
800         * src/closures.c (is_selinux_enabled): Define to const 0 for Cygwin.
801         (dlmmap, dlmunmap): Also use these functions on Cygwin.
802
803 2009-07-11  Richard Sandiford  <rdsandiford@googlemail.com>
804
805         PR testsuite/40699
806         PR testsuite/40707
807         PR testsuite/40709
808         * testsuite/lib/libffi-dg.exp: Revert 2009-07-02, 2009-07-01 and
809         2009-06-30 commits.
810
811 2009-07-01  Richard Sandiford  <r.sandiford@uk.ibm.com>
812
813         * testsuite/lib/libffi-dg.exp (libffi-init): Set ld_library_path
814         to "" before adding paths.  (This reinstates an assignment that
815         was removed by my 2009-06-30 commit, but changes the initial
816         value from "." to "".)
817
818 2009-07-01  H.J. Lu  <hongjiu.lu@intel.com>
819
820         PR testsuite/40601
821         * testsuite/lib/libffi-dg.exp (libffi-init): Properly set
822         gccdir.  Adjust ld_library_path for gcc only if gccdir isn't
823         empty.
824
825 2009-06-30  Richard Sandiford  <r.sandiford@uk.ibm.com>
826
827         * testsuite/lib/libffi-dg.exp (libffi-init): Don't add "."
828         to ld_library_path.  Use add_path.  Add just find_libgcc_s
829         to ld_library_path, not every libgcc multilib directory.
830
831 2009-06-16  Wim Lewis  <wiml@hhhh.org>
832
833         * src/powerpc/ffi.c: Avoid clobbering cr3 and cr4, which are
834         supposed to be callee-saved.
835         * src/powerpc/sysv.S (small_struct_return_value): Fix overrun of
836         return buffer for odd-size structs.
837
838 2009-06-16  Andreas Tobler  <a.tobler@schweiz.org>
839
840         PR libffi/40444
841         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Add
842         allow_stack_execute for Darwin.
843
844 2009-06-16  Andrew Haley  <aph@redhat.com>
845
846         * configure.ac (TARGETDIR): Add missing blank lines.
847         * configure: Regenerate.
848
849 2009-06-16  Andrew Haley  <aph@redhat.com>
850
851         * testsuite/libffi.call/cls_align_sint64.c,
852         testsuite/libffi.call/cls_align_uint64.c,
853         testsuite/libffi.call/cls_longdouble_va.c,
854         testsuite/libffi.call/cls_ulonglong.c,
855         testsuite/libffi.call/return_ll1.c,
856         testsuite/libffi.call/stret_medium2.c: Fix printf format
857         specifiers.
858         * testsuite/libffi.call/ffitest.h,
859         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRIuLL): Define.
860
861 2009-06-15  Andrew Haley  <aph@redhat.com>
862
863         * testsuite/libffi.call/err_bad_typedef.c: xfail everywhere.
864         * testsuite/libffi.call/err_bad_abi.c: Likewise.
865
866 2009-06-12  Andrew Haley  <aph@redhat.com>
867
868         * Makefile.am: Remove info_TEXINFOS.
869
870 2009-06-12  Andrew Haley  <aph@redhat.com>
871
872         * ChangeLog.libffi: testsuite/libffi.call/cls_align_sint64.c,
873         testsuite/libffi.call/cls_align_uint64.c,
874         testsuite/libffi.call/cls_ulonglong.c,
875         testsuite/libffi.call/return_ll1.c,
876         testsuite/libffi.call/stret_medium2.c: Fix printf format
877         specifiers.
878         testsuite/libffi.special/unwindtest.cc: include stdint.h.
879
880 2009-06-11  Timothy Wall  <twall@users.sf.net>
881
882         * Makefile.am,
883         configure.ac,
884         include/ffi.h.in,
885         include/ffi_common.h,
886         src/closures.c,
887         src/dlmalloc.c,
888         src/x86/ffi.c,
889         src/x86/ffitarget.h,
890         src/x86/win64.S (new),
891         README: Added win64 support (mingw or MSVC)
892         * Makefile.in,
893         include/Makefile.in,
894         man/Makefile.in,
895         testsuite/Makefile.in,
896         configure,
897         aclocal.m4: Regenerated
898         * ltcf-c.sh: properly escape cygwin/w32 path
899         * man/ffi_call.3: Clarify size requirements for return value.
900         * src/x86/ffi64.c: Fix filename in comment.
901         * src/x86/win32.S: Remove unused extern.
902
903         * testsuite/libffi.call/closure_fn0.c,
904         testsuite/libffi.call/closure_fn1.c,
905         testsuite/libffi.call/closure_fn2.c,
906         testsuite/libffi.call/closure_fn3.c,
907         testsuite/libffi.call/closure_fn4.c,
908         testsuite/libffi.call/closure_fn5.c,
909         testsuite/libffi.call/closure_fn6.c,
910         testsuite/libffi.call/closure_stdcall.c,
911         testsuite/libffi.call/cls_12byte.c,
912         testsuite/libffi.call/cls_16byte.c,
913         testsuite/libffi.call/cls_18byte.c,
914         testsuite/libffi.call/cls_19byte.c,
915         testsuite/libffi.call/cls_1_1byte.c,
916         testsuite/libffi.call/cls_20byte.c,
917         testsuite/libffi.call/cls_20byte1.c,
918         testsuite/libffi.call/cls_24byte.c,
919         testsuite/libffi.call/cls_2byte.c,
920         testsuite/libffi.call/cls_3_1byte.c,
921         testsuite/libffi.call/cls_3byte1.c,
922         testsuite/libffi.call/cls_3byte2.c,
923         testsuite/libffi.call/cls_4_1byte.c,
924         testsuite/libffi.call/cls_4byte.c,
925         testsuite/libffi.call/cls_5_1_byte.c,
926         testsuite/libffi.call/cls_5byte.c,
927         testsuite/libffi.call/cls_64byte.c,
928         testsuite/libffi.call/cls_6_1_byte.c,
929         testsuite/libffi.call/cls_6byte.c,
930         testsuite/libffi.call/cls_7_1_byte.c,
931         testsuite/libffi.call/cls_7byte.c,
932         testsuite/libffi.call/cls_8byte.c,
933         testsuite/libffi.call/cls_9byte1.c,
934         testsuite/libffi.call/cls_9byte2.c,
935         testsuite/libffi.call/cls_align_double.c,
936         testsuite/libffi.call/cls_align_float.c,
937         testsuite/libffi.call/cls_align_longdouble.c,
938         testsuite/libffi.call/cls_align_longdouble_split.c,
939         testsuite/libffi.call/cls_align_longdouble_split2.c,
940         testsuite/libffi.call/cls_align_pointer.c,
941         testsuite/libffi.call/cls_align_sint16.c,
942         testsuite/libffi.call/cls_align_sint32.c,
943         testsuite/libffi.call/cls_align_sint64.c,
944         testsuite/libffi.call/cls_align_uint16.c,
945         testsuite/libffi.call/cls_align_uint32.c,
946         testsuite/libffi.call/cls_align_uint64.c,
947         testsuite/libffi.call/cls_dbls_struct.c,
948         testsuite/libffi.call/cls_double.c,
949         testsuite/libffi.call/cls_double_va.c,
950         testsuite/libffi.call/cls_float.c,
951         testsuite/libffi.call/cls_longdouble.c,
952         testsuite/libffi.call/cls_longdouble_va.c,
953         testsuite/libffi.call/cls_multi_schar.c,
954         testsuite/libffi.call/cls_multi_sshort.c,
955         testsuite/libffi.call/cls_multi_sshortchar.c,
956         testsuite/libffi.call/cls_multi_uchar.c,
957         testsuite/libffi.call/cls_multi_ushort.c,
958         testsuite/libffi.call/cls_multi_ushortchar.c,
959         testsuite/libffi.call/cls_pointer.c,
960         testsuite/libffi.call/cls_pointer_stack.c,
961         testsuite/libffi.call/cls_schar.c,
962         testsuite/libffi.call/cls_sint.c,
963         testsuite/libffi.call/cls_sshort.c,
964         testsuite/libffi.call/cls_uchar.c,
965         testsuite/libffi.call/cls_uint.c,
966         testsuite/libffi.call/cls_ulonglong.c,
967         testsuite/libffi.call/cls_ushort.c,
968         testsuite/libffi.call/err_bad_abi.c,
969         testsuite/libffi.call/err_bad_typedef.c,
970         testsuite/libffi.call/float2.c,
971         testsuite/libffi.call/huge_struct.c,
972         testsuite/libffi.call/nested_struct.c,
973         testsuite/libffi.call/nested_struct1.c,
974         testsuite/libffi.call/nested_struct10.c,
975         testsuite/libffi.call/nested_struct2.c,
976         testsuite/libffi.call/nested_struct3.c,
977         testsuite/libffi.call/nested_struct4.c,
978         testsuite/libffi.call/nested_struct5.c,
979         testsuite/libffi.call/nested_struct6.c,
980         testsuite/libffi.call/nested_struct7.c,
981         testsuite/libffi.call/nested_struct8.c,
982         testsuite/libffi.call/nested_struct9.c,
983         testsuite/libffi.call/problem1.c,
984         testsuite/libffi.call/return_ldl.c,
985         testsuite/libffi.call/return_ll1.c,
986         testsuite/libffi.call/stret_large.c,
987         testsuite/libffi.call/stret_large2.c,
988         testsuite/libffi.call/stret_medium.c,
989         testsuite/libffi.call/stret_medium2.c,
990         testsuite/libffi.special/unwindtest.cc: use ffi_closure_alloc instead
991         of checking for MMAP.  Use intptr_t instead of long casts.
992
993 2009-06-11  Kaz Kojima  <kkojima@gcc.gnu.org>
994
995         * testsuite/libffi.call/cls_longdouble_va.c: Add xfail sh*-*-linux-*.
996         * testsuite/libffi.call/err_bad_abi.c: Add xfail sh*-*-*.
997         * testsuite/libffi.call/err_bad_typedef.c: Likewise.
998
999 2009-06-09  Andrew Haley  <aph@redhat.com>
1000
1001         * src/x86/freebsd.S: Add missing file.
1002
1003 2009-06-08  Andrew Haley  <aph@redhat.com>
1004
1005         Import from libffi 3.0.8:
1006
1007         * doc/libffi.texi: New file.
1008         * doc/libffi.info: Likewise.
1009         * doc/stamp-vti: Likewise.
1010         * man/Makefile.am: New file.
1011         * man/ffi_call.3: New file.
1012
1013         * Makefile.am (EXTRA_DIST): Add src/x86/darwin64.S,
1014         src/dlmalloc.c.
1015         (nodist_libffi_la_SOURCES): Add X86_FREEBSD.
1016
1017         * configure.ac: Bump version to 3.0.8.
1018         parisc*-*-linux*: Add.
1019         i386-*-freebsd* | i386-*-openbsd*: Add.
1020         powerpc-*-beos*: Add.
1021         AM_CONDITIONAL X86_FREEBSD: Add.
1022         AC_CONFIG_FILES: Add man/Makefile.
1023
1024         * include/ffi.h.in (FFI_FN): Change void (*)() to void (*)(void).
1025
1026 2009-06-08  Andrew Haley  <aph@redhat.com>
1027
1028         * README: Import from libffi 3.0.8.
1029
1030 2009-06-08  Andrew Haley  <aph@redhat.com>
1031
1032         * testsuite/libffi.call/err_bad_abi.c: Add xfails.
1033         * testsuite/libffi.call/cls_longdouble_va.c: Add xfails.
1034         * testsuite/libffi.call/cls_dbls_struct.c: Add xfail x86_64-*-linux-*.
1035         * testsuite/libffi.call/err_bad_typedef.c: Add xfails.
1036
1037         * testsuite/libffi.call/stret_medium2.c: Add __UNUSED__ to args.
1038         * testsuite/libffi.call/stret_medium.c: Likewise.
1039         * testsuite/libffi.call/stret_large2.c: Likewise.
1040         * testsuite/libffi.call/stret_large.c:  Likewise.
1041
1042 2008-12-26  Timothy Wall  <twall@users.sf.net>
1043
1044         * testsuite/libffi.call/cls_longdouble.c,
1045         testsuite/libffi.call/cls_longdouble_va.c,
1046         testsuite/libffi.call/cls_align_longdouble.c,
1047         testsuite/libffi.call/cls_align_longdouble_split.c,
1048         testsuite/libffi.call/cls_align_longdouble_split2.c: mark expected
1049         failures on x86_64 cygwin/mingw.
1050
1051 2008-12-22  Timothy Wall  <twall@users.sf.net>
1052
1053         * testsuite/libffi.call/closure_fn0.c,
1054         testsuite/libffi.call/closure_fn1.c,
1055         testsuite/libffi.call/closure_fn2.c,
1056         testsuite/libffi.call/closure_fn3.c,
1057         testsuite/libffi.call/closure_fn4.c,
1058         testsuite/libffi.call/closure_fn5.c,
1059         testsuite/libffi.call/closure_fn6.c,
1060         testsuite/libffi.call/closure_loc_fn0.c,
1061         testsuite/libffi.call/closure_stdcall.c,
1062         testsuite/libffi.call/cls_align_pointer.c,
1063         testsuite/libffi.call/cls_pointer.c,
1064         testsuite/libffi.call/cls_pointer_stack.c: use portable cast from
1065         pointer to integer (intptr_t).
1066         * testsuite/libffi.call/cls_longdouble.c: disable for win64.
1067
1068 2008-07-24  Anthony Green  <green@redhat.com>
1069
1070         * testsuite/libffi.call/cls_dbls_struct.c,
1071         testsuite/libffi.call/cls_double_va.c,
1072         testsuite/libffi.call/cls_longdouble.c,
1073         testsuite/libffi.call/cls_longdouble_va.c,
1074         testsuite/libffi.call/cls_pointer.c,
1075         testsuite/libffi.call/cls_pointer_stack.c,
1076         testsuite/libffi.call/err_bad_abi.c: Clean up failures from
1077         compiler warnings.
1078
1079 2008-03-04  Anthony Green  <green@redhat.com>
1080             Blake Chaffin
1081             hos@tamanegi.org
1082
1083         * testsuite/libffi.call/cls_align_longdouble_split2.c
1084           testsuite/libffi.call/cls_align_longdouble_split.c
1085           testsuite/libffi.call/cls_dbls_struct.c
1086           testsuite/libffi.call/cls_double_va.c
1087           testsuite/libffi.call/cls_longdouble.c
1088           testsuite/libffi.call/cls_longdouble_va.c
1089           testsuite/libffi.call/cls_pointer.c
1090           testsuite/libffi.call/cls_pointer_stack.c
1091           testsuite/libffi.call/err_bad_abi.c
1092           testsuite/libffi.call/err_bad_typedef.c
1093           testsuite/libffi.call/stret_large2.c
1094           testsuite/libffi.call/stret_large.c
1095           testsuite/libffi.call/stret_medium2.c
1096           testsuite/libffi.call/stret_medium.c: New tests from Apple.
1097
1098 2009-06-05  Andrew Haley  <aph@redhat.com>
1099
1100         * src/x86/ffitarget.h, src/x86/ffi.c: Merge stdcall changes from
1101         libffi.
1102
1103 2009-06-04  Andrew Haley  <aph@redhat.com>
1104
1105         * src/x86/ffitarget.h, src/x86/win32.S, src/x86/ffi.c: Back out
1106         stdcall changes.
1107
1108 2008-02-26  Anthony Green  <green@redhat.com>
1109             Thomas Heller  <theller@ctypes.org>
1110
1111         * src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C
1112         comment.
1113
1114 2008-02-03  Timothy Wall  <twall@users.sf.net>
1115
1116         * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return
1117           offset based on code pointer, not data pointer.
1118
1119 2008-01-31  Timothy Wall <twall@users.sf.net>
1120
1121         * testsuite/libffi.call/closure_stdcall.c: Add test for stdcall
1122         closures.
1123         * src/x86/ffitarget.h: Increase size of trampoline for stdcall
1124         closures.
1125         * src/x86/win32.S: Add assembly for stdcall closure.
1126         * src/x86/ffi.c: Initialize stdcall closure trampoline.
1127
1128 2009-06-04  Andrew Haley  <aph@redhat.com>
1129
1130         * include/ffi.h.in: Change void (*)() to void (*)(void).
1131         * src/x86/ffi.c: Likewise.
1132
1133 2009-06-04  Andrew Haley  <aph@redhat.com>
1134
1135         * src/powerpc/ppc_closure.S: Insert licence header.
1136         * src/powerpc/linux64_closure.S: Likewise.
1137         * src/m68k/sysv.S: Likewise.
1138
1139         * src/sh64/ffi.c: Change void (*)() to void (*)(void).
1140         * src/powerpc/ffi.c: Likewise.
1141         * src/powerpc/ffi_darwin.c: Likewise.
1142         * src/m32r/ffi.c: Likewise.
1143         * src/sh64/ffi.c: Likewise.
1144         * src/x86/ffi64.c: Likewise.
1145         * src/alpha/ffi.c: Likewise.
1146         * src/alpha/osf.S: Likewise.
1147         * src/frv/ffi.c: Likewise.
1148         * src/s390/ffi.c: Likewise.
1149         * src/pa/ffi.c: Likewise.
1150         * src/pa/hpux32.S: Likewise.
1151         * src/ia64/unix.S: Likewise.
1152         * src/ia64/ffi.c: Likewise.
1153         * src/sparc/ffi.c: Likewise.
1154         * src/mips/ffi.c: Likewise.
1155         * src/sh/ffi.c: Likewise.
1156
1157 2008-02-15  David Daney  <ddaney@avtrex.com>
1158
1159         * src/mips/ffi.c (USE__BUILTIN___CLEAR_CACHE):
1160         Define (conditionally), and use it to include cachectl.h.
1161         (ffi_prep_closure_loc): Fix cache flushing.
1162         * src/mips/ffitarget.h (_ABIN32, _ABI64, _ABIO32): Define.
1163
1164 2009-06-04  Andrew Haley  <aph@redhat.com>
1165
1166         include/ffi.h.in,
1167         src/arm/ffitarget.h,
1168         src/arm/ffi.c,
1169         src/arm/sysv.S,
1170         src/powerpc/ffitarget.h,
1171         src/closures.c,
1172         src/sh64/ffitarget.h,
1173         src/sh64/ffi.c,
1174         src/sh64/sysv.S,
1175         src/types.c,
1176         src/x86/ffi64.c,
1177         src/x86/ffitarget.h,
1178         src/x86/win32.S,
1179         src/x86/darwin.S,
1180         src/x86/ffi.c,
1181         src/x86/sysv.S,
1182         src/x86/unix64.S,
1183         src/alpha/ffitarget.h,
1184         src/alpha/ffi.c,
1185         src/alpha/osf.S,
1186         src/m68k/ffitarget.h,
1187         src/frv/ffitarget.h,
1188         src/frv/ffi.c,
1189         src/s390/ffitarget.h,
1190         src/s390/sysv.S,
1191         src/cris/ffitarget.h,
1192         src/pa/linux.S,
1193         src/pa/ffitarget.h,
1194         src/pa/ffi.c,
1195         src/raw_api.c,
1196         src/ia64/ffitarget.h,
1197         src/ia64/unix.S,
1198         src/ia64/ffi.c,
1199         src/ia64/ia64_flags.h,
1200         src/java_raw_api.c,
1201         src/debug.c,
1202         src/sparc/v9.S,
1203         src/sparc/ffitarget.h,
1204         src/sparc/ffi.c,
1205         src/sparc/v8.S,
1206         src/mips/ffitarget.h,
1207         src/mips/n32.S,
1208         src/mips/o32.S,
1209         src/mips/ffi.c,
1210         src/prep_cif.c,
1211         src/sh/ffitarget.h,
1212         src/sh/ffi.c,
1213         src/sh/sysv.S: Update license text.
1214
1215 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
1216
1217         * src/x86/win32.S (_ffi_closure_STDCALL):  New function.
1218         (.eh_frame):  Add FDE for it.
1219
1220 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
1221
1222         * configure.ac:  Also check if assembler supports pc-relative
1223         relocs on X86_WIN32 targets.
1224         * configure:  Regenerate.
1225         * src/x86/win32.S (ffi_prep_args):  Declare extern, not global.
1226         (_ffi_call_SYSV):  Add missing function type symbol .def and
1227         add EH markup labels.
1228         (_ffi_call_STDCALL):  Likewise.
1229         (_ffi_closure_SYSV):  Likewise.
1230         (_ffi_closure_raw_SYSV):  Likewise.
1231         (.eh_frame):  Add hand-crafted EH data.
1232
1233 2009-04-09  Jakub Jelinek  <jakub@redhat.com>
1234
1235         * testsuite/lib/libffi-dg.exp: Change copyright header to refer to
1236         version 3 of the GNU General Public License and to point readers
1237         at the COPYING3 file and the FSF's license web page.
1238         * testsuite/libffi.call/call.exp: Likewise.
1239         * testsuite/libffi.special/special.exp: Likewise.
1240
1241 2009-03-01  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1242
1243         * configure: Regenerate.
1244
1245 2008-12-18  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1246
1247         PR libffi/26048
1248         * configure.ac (HAVE_AS_X86_PCREL): New test.
1249         * configure: Regenerate.
1250         * fficonfig.h.in: Regenerate.
1251         * src/x86/sysv.S [!FFI_NO_RAW_API]: Precalculate
1252         RAW_CLOSURE_CIF_OFFSET, RAW_CLOSURE_FUN_OFFSET,
1253         RAW_CLOSURE_USER_DATA_OFFSET for the Solaris 10/x86 assembler.
1254         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
1255         * src/x86/unix64.S (.Lstore_table): Move to .text section.
1256         (.Lload_table): Likewise.
1257         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
1258
1259 2008-12-18  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1260
1261         * configure: Regenerate.
1262
1263 2008-11-21  Eric Botcazou  <ebotcazou@adacore.com>
1264
1265         * src/sparc/ffi.c (ffi_prep_cif_machdep): Add support for
1266         signed/unsigned int8/16 return values.
1267         * src/sparc/v8.S (ffi_call_v8): Likewise.
1268         (ffi_closure_v8): Likewise.
1269
1270 2008-09-26  Peter O'Gorman  <pogma@thewrittenword.com>
1271             Steve Ellcey  <sje@cup.hp.com>
1272
1273         * configure: Regenerate for new libtool.
1274         * Makefile.in: Ditto.
1275         * include/Makefile.in: Ditto.
1276         * aclocal.m4: Ditto.
1277
1278 2008-08-25  Andreas Tobler  <a.tobler@schweiz.org>
1279
1280         * src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and
1281         FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum.
1282         Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT.
1283         Adjust copyright notice.
1284         * src/powerpc/ffi.c: Add two new flags to indicate if we have one
1285         register or two register to use for FFI_SYSV structs.
1286         (ffi_prep_cif_machdep): Pass the right register flag introduced above.
1287         (ffi_closure_helper_SYSV): Fix the return type for
1288         FFI_SYSV_TYPE_SMALL_STRUCT. Comment.
1289         Adjust copyright notice.
1290
1291 2008-07-16  Kaz Kojima  <kkojima@gcc.gnu.org>
1292
1293         * src/sh/ffi.c (ffi_prep_closure_loc): Turn INSN into an unsigned
1294         int.
1295
1296 2008-06-17  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1297
1298         * configure: Regenerate.
1299         * include/Makefile.in: Regenerate.
1300         * testsuite/Makefile.in: Regenerate.
1301
1302 2008-06-07  Joseph Myers  <joseph@codesourcery.com>
1303
1304         * configure.ac (parisc*-*-linux*, powerpc-*-sysv*,
1305         powerpc-*-beos*): Remove.
1306         * configure: Regenerate.
1307
1308 2008-05-09  Julian Brown  <julian@codesourcery.com>
1309
1310         * Makefile.am (LTLDFLAGS): New.
1311         (libffi_la_LDFLAGS): Use above.
1312         * Makefile.in: Regenerate.
1313
1314 2008-04-18  Paolo Bonzini  <bonzini@gnu.org>
1315
1316         PR bootstrap/35457
1317         * aclocal.m4: Regenerate.
1318         * configure: Regenerate.
1319
1320 2008-03-26  Kaz Kojima  <kkojima@gcc.gnu.org>
1321
1322         * src/sh/sysv.S: Add .note.GNU-stack on Linux.
1323         * src/sh64/sysv.S: Likewise.
1324
1325 2008-03-26  Daniel Jacobowitz  <dan@debian.org>
1326
1327         * src/arm/sysv.S: Fix ARM comment marker.
1328
1329 2008-03-26  Jakub Jelinek  <jakub@redhat.com>
1330
1331         * src/alpha/osf.S: Add .note.GNU-stack on Linux.
1332         * src/s390/sysv.S: Likewise.
1333         * src/powerpc/ppc_closure.S: Likewise.
1334         * src/powerpc/sysv.S: Likewise.
1335         * src/x86/unix64.S: Likewise.
1336         * src/x86/sysv.S: Likewise.
1337         * src/sparc/v8.S: Likewise.
1338         * src/sparc/v9.S: Likewise.
1339         * src/m68k/sysv.S: Likewise.
1340         * src/arm/sysv.S: Likewise.
1341
1342 2008-03-16  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1343
1344         * aclocal.m4: Regenerate.
1345         * configure: Likewise.
1346         * Makefile.in: Likewise.
1347         * include/Makefile.in: Likewise.
1348         * testsuite/Makefile.in: Likewise.
1349
1350 2008-02-12  Bjoern Koenig  <bkoenig@alpha-tierchen.de>
1351             Andreas Tobler  <a.tobler@schweiz.org>
1352
1353         * configure.ac: Add amd64-*-freebsd* target.
1354         * configure: Regenerate.
1355
1356 2008-01-30  H.J. Lu  <hongjiu.lu@intel.com>
1357
1358         PR libffi/34612
1359         * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when
1360         returning struct.
1361
1362         * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer"
1363         tests.
1364
1365 2008-01-24  David Edelsohn  <edelsohn@gnu.org>
1366
1367         * configure: Regenerate.
1368
1369 2008-01-06  Andreas Tobler  <a.tobler@schweiz.org>
1370
1371         * src/x86/ffi.c (ffi_prep_cif_machdep): Fix thinko.
1372
1373 2008-01-05  Andreas Tobler  <a.tobler@schweiz.org>
1374
1375         PR testsuite/32843
1376         * src/x86/ffi.c (ffi_prep_cif_machdep): Add code for
1377         signed/unsigned int8/16 for X86_DARWIN.
1378         Updated copyright info.
1379         Handle one and two byte structs with special cif->flags.
1380         * src/x86/ffitarget.h: Add special types for one and two byte structs.
1381         Updated copyright info.
1382         * src/x86/darwin.S (ffi_call_SYSV): Rewrite to use a jump table like
1383         sysv.S
1384         Remove code to pop args from the stack after call.
1385         Special-case signed/unsigned for int8/16, one and two byte structs.
1386         (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1387         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1388         FFI_TYPE_SINT32.
1389         Updated copyright info.
1390
1391 2007-12-08  David Daney  <ddaney@avtrex.com>
1392
1393         * src/mips/n32.S (ffi_call_N32):  Replace dadd with ADDU, dsub with
1394         SUBU, add with ADDU and use smaller code sequences.
1395
1396 2007-12-07  David Daney  <ddaney@avtrex.com>
1397
1398         * src/mips/ffi.c (ffi_prep_cif_machdep): Handle long double return
1399         type.
1400
1401 2007-12-06  David Daney  <ddaney@avtrex.com>
1402
1403         * include/ffi.h.in (FFI_SIZEOF_JAVA_RAW): Define if not already
1404         defined.
1405         (ffi_java_raw): New typedef.
1406         (ffi_java_raw_call, ffi_java_ptrarray_to_raw,
1407         ffi_java_raw_to_ptrarray): Change parameter types from ffi_raw to
1408         ffi_java_raw.
1409         (ffi_java_raw_closure) : Same.
1410         (ffi_prep_java_raw_closure, ffi_prep_java_raw_closure_loc): Change
1411         parameter types.
1412         * src/java_raw_api.c (ffi_java_raw_size):  Replace FFI_SIZEOF_ARG with
1413         FFI_SIZEOF_JAVA_RAW.
1414         (ffi_java_raw_to_ptrarray): Change type of raw to ffi_java_raw.
1415         Replace FFI_SIZEOF_ARG with FFI_SIZEOF_JAVA_RAW. Use
1416         sizeof(ffi_java_raw) for alignment calculations.
1417         (ffi_java_ptrarray_to_raw): Same.
1418         (ffi_java_rvalue_to_raw): Add special handling for FFI_TYPE_POINTER
1419         if FFI_SIZEOF_JAVA_RAW == 4.
1420         (ffi_java_raw_to_rvalue): Same.
1421         (ffi_java_raw_call): Change type of raw to ffi_java_raw.
1422         (ffi_java_translate_args): Same.
1423         (ffi_prep_java_raw_closure_loc, ffi_prep_java_raw_closure): Change
1424         parameter types.
1425         * src/mips/ffitarget.h (FFI_SIZEOF_JAVA_RAW): Define for N32 ABI.
1426
1427 2007-12-06  David Daney  <ddaney@avtrex.com>
1428
1429         * src/mips/n32.S (ffi_closure_N32): Use 64-bit add instruction on
1430         pointer values.
1431
1432 2007-12-01  Andreas Tobler  <a.tobler@schweiz.org>
1433
1434         PR libffi/31937
1435         * src/powerpc/ffitarget.h: Introduce new ABI FFI_LINUX_SOFT_FLOAT.
1436         Add local FFI_TYPE_UINT128 to handle soft-float long-double-128.
1437         * src/powerpc/ffi.c: Distinguish between __NO_FPRS__ and not and
1438         set the NUM_FPR_ARG_REGISTERS according to.
1439         Add support for potential soft-float support under hard-float
1440         architecture.
1441         (ffi_prep_args_SYSV): Set NUM_FPR_ARG_REGISTERS to 0 in case of
1442         FFI_LINUX_SOFT_FLOAT, handle float, doubles and long-doubles according
1443         to the FFI_LINUX_SOFT_FLOAT ABI.
1444         (ffi_prep_cif_machdep): Likewise.
1445         (ffi_closure_helper_SYSV): Likewise.
1446         * src/powerpc/ppc_closure.S: Make sure not to store float/double
1447         on archs where __NO_FPRS__ is true.
1448         Add FFI_TYPE_UINT128 support.
1449         * src/powerpc/sysv.S: Add support for soft-float long-double-128.
1450         Adjust copyright notice.
1451
1452 2007-11-25  Andreas Tobler  <a.tobler@schweiz.org>
1453
1454         * src/closures.c: Move defintion of MAYBE_UNUSED from here to ...
1455         * include/ffi_common.h: ... here.
1456         Update copyright.
1457
1458 2007-11-17  Andreas Tobler  <a.tobler@schweiz.org>
1459
1460         * src/powerpc/sysv.S: Load correct cr to compare if we have long double.
1461         * src/powerpc/linux64.S: Likewise.
1462         * src/powerpc/ffi.c: Add a comment to show which part goes into cr6.
1463         * testsuite/libffi.call/return_ldl.c: New test.
1464
1465 2007-09-04    <aph@redhat.com>
1466
1467         * src/arm/sysv.S (UNWIND): New.
1468         (Whole file): Conditionally compile unwinder directives.
1469         * src/arm/sysv.S: Add unwinder directives.
1470
1471         * src/arm/ffi.c (ffi_prep_args): Align structs by at least 4 bytes.
1472         Only treat r0 as a struct address if we're actually returning a
1473         struct by address.
1474         Only copy the bytes that are actually within a struct.
1475         (ffi_prep_cif_machdep): A Composite Type not larger than 4 bytes
1476         is returned in r0, not passed by address.
1477         (ffi_call): Allocate a word-sized temporary for the case where
1478         a composite is returned in r0.
1479         (ffi_prep_incoming_args_SYSV): Align as necessary.
1480
1481 2007-08-05  Steven Newbury  <s_j_newbury@yahoo.co.uk>
1482
1483         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Use __clear_cache instead of
1484         directly using the sys_cacheflush syscall.
1485
1486 2007-07-27  Andrew Haley  <aph@redhat.com>
1487
1488         * src/arm/sysv.S (ffi_closure_SYSV): Add soft-float.
1489
1490 2007-09-03  Maciej W. Rozycki  <macro@linux-mips.org>
1491
1492         * Makefile.am: Unify MIPS_IRIX and MIPS_LINUX into MIPS.
1493         * configure.ac: Likewise.
1494         * Makefile.in: Regenerate.
1495         * include/Makefile.in: Likewise.
1496         * testsuite/Makefile.in: Likewise.
1497         * configure: Likewise.
1498
1499 2007-08-24  David Daney  <ddaney@avtrex.com>
1500
1501         * testsuite/libffi.call/return_sl.c: New test.
1502
1503 2007-08-10  David Daney  <ddaney@avtrex.com>
1504
1505         * testsuite/libffi.call/cls_multi_ushort.c,
1506         testsuite/libffi.call/cls_align_uint16.c,
1507         testsuite/libffi.call/nested_struct1.c,
1508         testsuite/libffi.call/nested_struct3.c,
1509         testsuite/libffi.call/cls_7_1_byte.c,
1510         testsuite/libffi.call/nested_struct5.c,
1511         testsuite/libffi.call/cls_double.c,
1512         testsuite/libffi.call/nested_struct7.c,
1513         testsuite/libffi.call/cls_sint.c,
1514         testsuite/libffi.call/nested_struct9.c,
1515         testsuite/libffi.call/cls_20byte1.c,
1516         testsuite/libffi.call/cls_multi_sshortchar.c,
1517         testsuite/libffi.call/cls_align_sint64.c,
1518         testsuite/libffi.call/cls_3byte2.c,
1519         testsuite/libffi.call/cls_multi_schar.c,
1520         testsuite/libffi.call/cls_multi_uchar.c,
1521         testsuite/libffi.call/cls_19byte.c,
1522         testsuite/libffi.call/cls_9byte1.c,
1523         testsuite/libffi.call/cls_align_float.c,
1524         testsuite/libffi.call/closure_fn1.c,
1525         testsuite/libffi.call/problem1.c,
1526         testsuite/libffi.call/closure_fn3.c,
1527         testsuite/libffi.call/cls_sshort.c,
1528         testsuite/libffi.call/closure_fn5.c,
1529         testsuite/libffi.call/cls_align_double.c,
1530         testsuite/libffi.call/nested_struct.c,
1531         testsuite/libffi.call/cls_2byte.c,
1532         testsuite/libffi.call/nested_struct10.c,
1533         testsuite/libffi.call/cls_4byte.c,
1534         testsuite/libffi.call/cls_6byte.c,
1535         testsuite/libffi.call/cls_8byte.c,
1536         testsuite/libffi.call/cls_multi_sshort.c,
1537         testsuite/libffi.call/cls_align_sint16.c,
1538         testsuite/libffi.call/cls_align_uint32.c,
1539         testsuite/libffi.call/cls_20byte.c,
1540         testsuite/libffi.call/cls_float.c,
1541         testsuite/libffi.call/nested_struct2.c,
1542         testsuite/libffi.call/cls_5_1_byte.c,
1543         testsuite/libffi.call/nested_struct4.c,
1544         testsuite/libffi.call/cls_24byte.c,
1545         testsuite/libffi.call/nested_struct6.c,
1546         testsuite/libffi.call/cls_64byte.c,
1547         testsuite/libffi.call/nested_struct8.c,
1548         testsuite/libffi.call/cls_uint.c,
1549         testsuite/libffi.call/cls_multi_ushortchar.c,
1550         testsuite/libffi.call/cls_schar.c,
1551         testsuite/libffi.call/cls_uchar.c,
1552         testsuite/libffi.call/cls_align_uint64.c,
1553         testsuite/libffi.call/cls_ulonglong.c,
1554         testsuite/libffi.call/cls_align_longdouble.c,
1555         testsuite/libffi.call/cls_1_1byte.c,
1556         testsuite/libffi.call/cls_12byte.c,
1557         testsuite/libffi.call/cls_3_1byte.c,
1558         testsuite/libffi.call/cls_3byte1.c,
1559         testsuite/libffi.call/cls_4_1byte.c,
1560         testsuite/libffi.call/cls_6_1_byte.c,
1561         testsuite/libffi.call/cls_16byte.c,
1562         testsuite/libffi.call/cls_18byte.c,
1563         testsuite/libffi.call/closure_fn0.c,
1564         testsuite/libffi.call/cls_9byte2.c,
1565         testsuite/libffi.call/closure_fn2.c,
1566         testsuite/libffi.call/closure_fn4.c,
1567         testsuite/libffi.call/cls_ushort.c,
1568         testsuite/libffi.call/closure_fn6.c,
1569         testsuite/libffi.call/cls_5byte.c,
1570         testsuite/libffi.call/cls_align_pointer.c,
1571         testsuite/libffi.call/cls_7byte.c,
1572         testsuite/libffi.call/cls_align_sint32.c,
1573         testsuite/libffi.special/unwindtest_ffi_call.cc,
1574         testsuite/libffi.special/unwindtest.cc: Remove xfail for mips64*-*-*.
1575
1576 2007-08-10  David Daney  <ddaney@avtrex.com>
1577
1578         PR libffi/28313
1579         * configure.ac: Don't treat mips64 as a special case.
1580         * Makefile.am (nodist_libffi_la_SOURCES): Add n32.S.
1581         * configure: Regenerate
1582         * Makefile.in: Ditto.
1583         * fficonfig.h.in: Ditto.
1584         * src/mips/ffitarget.h (REG_L, REG_S, SUBU, ADDU, SRL, LI): Indent.
1585         (LA, EH_FRAME_ALIGN, FDE_ADDR_BYTES): New preprocessor macros.
1586         (FFI_DEFAULT_ABI): Set for n64 case.
1587         (FFI_CLOSURES, FFI_TRAMPOLINE_SIZE): Define for n32 and n64 cases.
1588         * src/mips/n32.S (ffi_call_N32): Add debug macros and labels for FDE.
1589         (ffi_closure_N32): New function.
1590         (.eh_frame): New section
1591         * src/mips/o32.S: Clean up comments.
1592         (ffi_closure_O32): Pass ffi_closure parameter in $12.
1593         * src/mips/ffi.c: Use FFI_MIPS_N32 instead of
1594         _MIPS_SIM == _ABIN32 throughout.
1595         (FFI_MIPS_STOP_HERE): New, use in place of
1596         ffi_stop_here.
1597         (ffi_prep_args): Use unsigned long to hold pointer values.  Rewrite
1598         to support n32/n64 ABIs.
1599         (calc_n32_struct_flags): Rewrite.
1600         (calc_n32_return_struct_flags): Remove unused variable.  Reverse
1601         position of flag bits.
1602         (ffi_prep_cif_machdep): Rewrite n32 portion.
1603         (ffi_call): Enable for n64.  Add special handling for small structure
1604         return values.
1605         (ffi_prep_closure_loc): Add n32 and n64 support.
1606         (ffi_closure_mips_inner_O32): Add cast to silence warning.
1607         (copy_struct_N32, ffi_closure_mips_inner_N32): New functions.
1608
1609 2007-08-08  David Daney  <ddaney@avtrex.com>
1610
1611         * testsuite/libffi.call/ffitest.h (ffi_type_mylong): Remove definition.
1612         * testsuite/libffi.call/cls_align_uint16.c (main): Use correct type
1613         specifiers.
1614         * testsuite/libffi.call/nested_struct1.c (main): Ditto.
1615         * testsuite/libffi.call/cls_sint.c (main): Ditto.
1616         * testsuite/libffi.call/nested_struct9.c (main): Ditto.
1617         * testsuite/libffi.call/cls_20byte1.c (main): Ditto.
1618         * testsuite/libffi.call/cls_9byte1.c (main): Ditto.
1619         * testsuite/libffi.call/closure_fn1.c (main): Ditto.
1620         * testsuite/libffi.call/closure_fn3.c (main): Ditto.
1621         * testsuite/libffi.call/return_dbl2.c (main): Ditto.
1622         * testsuite/libffi.call/cls_sshort.c (main): Ditto.
1623         * testsuite/libffi.call/return_fl3.c (main): Ditto.
1624         * testsuite/libffi.call/closure_fn5.c (main): Ditto.
1625         * testsuite/libffi.call/nested_struct.c (main): Ditto.
1626         * testsuite/libffi.call/nested_struct10.c (main): Ditto.
1627         * testsuite/libffi.call/return_ll1.c (main): Ditto.
1628         * testsuite/libffi.call/cls_8byte.c (main): Ditto.
1629         * testsuite/libffi.call/cls_align_uint32.c (main): Ditto.
1630         * testsuite/libffi.call/cls_align_sint16.c (main): Ditto.
1631         * testsuite/libffi.call/cls_20byte.c (main): Ditto.
1632         * testsuite/libffi.call/nested_struct2.c (main): Ditto.
1633         * testsuite/libffi.call/cls_24byte.c (main): Ditto.
1634         * testsuite/libffi.call/nested_struct6.c (main): Ditto.
1635         * testsuite/libffi.call/cls_uint.c (main): Ditto.
1636         * testsuite/libffi.call/cls_12byte.c (main): Ditto.
1637         * testsuite/libffi.call/cls_16byte.c (main): Ditto.
1638         * testsuite/libffi.call/closure_fn0.c (main): Ditto.
1639         * testsuite/libffi.call/cls_9byte2.c (main): Ditto.
1640         * testsuite/libffi.call/closure_fn2.c (main): Ditto.
1641         * testsuite/libffi.call/return_dbl1.c (main): Ditto.
1642         * testsuite/libffi.call/closure_fn4.c (main): Ditto.
1643         * testsuite/libffi.call/closure_fn6.c (main): Ditto.
1644         * testsuite/libffi.call/cls_align_sint32.c (main): Ditto.
1645
1646 2007-08-07  Andrew Haley  <aph@redhat.com>
1647
1648         * src/x86/sysv.S (ffi_closure_raw_SYSV): Fix typo in previous
1649         checkin.
1650
1651 2007-08-06  Andrew Haley  <aph@redhat.com>
1652
1653         PR testsuite/32843
1654         * src/x86/sysv.S (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1655         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1656         FFI_TYPE_SINT32.
1657
1658 2007-08-02  David Daney  <ddaney@avtrex.com>
1659
1660         * testsuite/libffi.call/return_ul.c (main): Define return type as
1661         ffi_arg.  Use proper printf conversion specifier.
1662
1663 2007-07-30  Andrew Haley  <aph@redhat.com>
1664
1665         PR testsuite/32843
1666         * src/x86/ffi.c (ffi_prep_cif_machdep): in x86 case, add code for
1667         signed/unsigned int8/16.
1668         * src/x86/sysv.S (ffi_call_SYSV): Rewrite to:
1669         Use a jump table.
1670         Remove code to pop args from the stack after call.
1671         Special-case signed/unsigned int8/16.
1672         * testsuite/libffi.call/return_sc.c (main): Revert.
1673
1674 2007-07-26  Richard Guenther  <rguenther@suse.de>
1675
1676         PR testsuite/32843
1677         * testsuite/libffi.call/return_sc.c (main): Verify call
1678         result as signed char, not ffi_arg.
1679
1680 2007-07-16  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1681
1682         * configure.ac (i?86-*-solaris2.1[0-9]): Set TARGET to X86_64.
1683         * configure: Regenerate.
1684
1685 2007-07-11  David Daney  <ddaney@avtrex.com>
1686
1687         * src/mips/ffi.c: Don't include sys/cachectl.h.
1688         (ffi_prep_closure_loc): Use __builtin___clear_cache() instead of
1689         cacheflush().
1690
1691 2007-05-18  Aurelien Jarno  <aurelien@aurel32.net>
1692
1693         * src/arm/ffi.c (ffi_prep_closure_loc): Renamed and ajusted
1694         from (ffi_prep_closure): ... this.
1695         (FFI_INIT_TRAMPOLINE): Adjust.
1696
1697 2005-12-31  Phil Blundell  <pb@reciva.com>
1698
1699         * src/arm/ffi.c (ffi_prep_incoming_args_SYSV,
1700         ffi_closure_SYSV_inner, ffi_prep_closure): New, add closure support.
1701         * src/arm/sysv.S(ffi_closure_SYSV): Likewise.
1702         * src/arm/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1703         (FFI_CLOSURES): Enable closure support.
1704
1705 2007-07-03  Andrew Haley  <aph@hedges.billgatliff.com>
1706
1707         * testsuite/libffi.call/cls_multi_ushort.c,
1708         testsuite/libffi.call/cls_align_uint16.c,
1709         testsuite/libffi.call/nested_struct1.c,
1710         testsuite/libffi.call/nested_struct3.c,
1711         testsuite/libffi.call/cls_7_1_byte.c,
1712         testsuite/libffi.call/cls_double.c,
1713         testsuite/libffi.call/nested_struct5.c,
1714         testsuite/libffi.call/nested_struct7.c,
1715         testsuite/libffi.call/cls_sint.c,
1716         testsuite/libffi.call/nested_struct9.c,
1717         testsuite/libffi.call/cls_20byte1.c,
1718         testsuite/libffi.call/cls_multi_sshortchar.c,
1719         testsuite/libffi.call/cls_align_sint64.c,
1720         testsuite/libffi.call/cls_3byte2.c,
1721         testsuite/libffi.call/cls_multi_schar.c,
1722         testsuite/libffi.call/cls_multi_uchar.c,
1723         testsuite/libffi.call/cls_19byte.c,
1724         testsuite/libffi.call/cls_9byte1.c,
1725         testsuite/libffi.call/cls_align_float.c,
1726         testsuite/libffi.call/closure_fn1.c,
1727         testsuite/libffi.call/problem1.c,
1728         testsuite/libffi.call/closure_fn3.c,
1729         testsuite/libffi.call/cls_sshort.c,
1730         testsuite/libffi.call/closure_fn5.c,
1731         testsuite/libffi.call/cls_align_double.c,
1732         testsuite/libffi.call/cls_2byte.c,
1733         testsuite/libffi.call/nested_struct.c,
1734         testsuite/libffi.call/nested_struct10.c,
1735         testsuite/libffi.call/cls_4byte.c,
1736         testsuite/libffi.call/cls_6byte.c,
1737         testsuite/libffi.call/cls_8byte.c,
1738         testsuite/libffi.call/cls_multi_sshort.c,
1739         testsuite/libffi.call/cls_align_uint32.c,
1740         testsuite/libffi.call/cls_align_sint16.c,
1741         testsuite/libffi.call/cls_float.c,
1742         testsuite/libffi.call/cls_20byte.c,
1743         testsuite/libffi.call/cls_5_1_byte.c,
1744         testsuite/libffi.call/nested_struct2.c,
1745         testsuite/libffi.call/cls_24byte.c,
1746         testsuite/libffi.call/nested_struct4.c,
1747         testsuite/libffi.call/nested_struct6.c,
1748         testsuite/libffi.call/cls_64byte.c,
1749         testsuite/libffi.call/nested_struct8.c,
1750         testsuite/libffi.call/cls_uint.c,
1751         testsuite/libffi.call/cls_multi_ushortchar.c,
1752         testsuite/libffi.call/cls_schar.c,
1753         testsuite/libffi.call/cls_uchar.c,
1754         testsuite/libffi.call/cls_align_uint64.c,
1755         testsuite/libffi.call/cls_ulonglong.c,
1756         testsuite/libffi.call/cls_align_longdouble.c,
1757         testsuite/libffi.call/cls_1_1byte.c,
1758         testsuite/libffi.call/cls_12byte.c,
1759         testsuite/libffi.call/cls_3_1byte.c,
1760         testsuite/libffi.call/cls_3byte1.c,
1761         testsuite/libffi.call/cls_4_1byte.c,
1762         testsuite/libffi.call/cls_6_1_byte.c,
1763         testsuite/libffi.call/cls_16byte.c,
1764         testsuite/libffi.call/cls_18byte.c,
1765         testsuite/libffi.call/closure_fn0.c,
1766         testsuite/libffi.call/cls_9byte2.c,
1767         testsuite/libffi.call/closure_fn2.c,
1768         testsuite/libffi.call/closure_fn4.c,
1769         testsuite/libffi.call/cls_ushort.c,
1770         testsuite/libffi.call/closure_fn6.c,
1771         testsuite/libffi.call/cls_5byte.c,
1772         testsuite/libffi.call/cls_align_pointer.c,
1773         testsuite/libffi.call/cls_7byte.c,
1774         testsuite/libffi.call/cls_align_sint32.c,
1775         testsuite/libffi.special/unwindtest_ffi_call.cc,
1776         testsuite/libffi.special/unwindtest.cc: Enable for ARM.
1777
1778 2007-07-05  H.J. Lu  <hongjiu.lu@intel.com>
1779
1780         * aclocal.m4: Regenerated.
1781
1782 2007-06-02  Paolo Bonzini  <bonzini@gnu.org>
1783
1784         * configure: Regenerate.
1785
1786 2007-05-23  Steve Ellcey  <sje@cup.hp.com>
1787
1788         * Makefile.in: Regenerate.
1789         * configure: Regenerate.
1790         * aclocal.m4: Regenerate.
1791         * include/Makefile.in: Regenerate.
1792         * testsuite/Makefile.in: Regenerate.
1793
1794 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1795
1796         * src/m68k/ffi.c (ffi_prep_incoming_args_SYSV,
1797         ffi_closure_SYSV_inner,ffi_prep_closure): New, add closure support.
1798         * src/m68k/sysv.S(ffi_closure_SYSV,ffi_closure_struct_SYSV): Likewise.
1799         * src/m68k/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1800         (FFI_CLOSURES): Enable closure support.
1801
1802 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1803
1804         * configure.ac (HAVE_AS_CFI_PSEUDO_OP): New test.
1805         * configure: Regenerate.
1806         * fficonfig.h.in: Regenerate.
1807         * src/m68k/sysv.S (CFI_STARTPROC,CFI_ENDPROC,
1808         CFI_OFFSET,CFI_DEF_CFA): New macros.
1809         (ffi_call_SYSV): Add callframe annotation.
1810
1811 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1812
1813         * src/m68k/ffi.c (ffi_prep_args,ffi_prep_cif_machdep): Fix
1814         numerous test suite failures.
1815         * src/m68k/sysv.S (ffi_call_SYSV): Likewise.
1816
1817 2007-04-11  Paolo Bonzini  <bonzini@gnu.org>
1818
1819         * Makefile.am (EXTRA_DIST): Bring up to date.
1820         * Makefile.in: Regenerate.
1821         * src/frv/eabi.S: Remove RCS keyword.
1822
1823 2007-04-06  Richard Henderson  <rth@redhat.com>
1824
1825         * configure.ac: Tidy target case.
1826         (HAVE_LONG_DOUBLE): Allow the target to override.
1827         * configure: Regenerate.
1828         * include/ffi.h.in: Don't define ffi_type_foo if
1829         LIBFFI_HIDE_BASIC_TYPES is defined.
1830         (ffi_type_longdouble): If not HAVE_LONG_DOUBLE, define
1831         to ffi_type_double.
1832         * types.c (LIBFFI_HIDE_BASIC_TYPES): Define.
1833         (FFI_TYPEDEF, ffi_type_void): Mark the data const.
1834         (ffi_type_longdouble): Special case for Alpha.  Don't define
1835         if long double == double.
1836
1837         * src/alpha/ffi.c (FFI_TYPE_LONGDOUBLE): Assert unique value.
1838         (ffi_prep_cif_machdep): Handle it as the 128-bit type.
1839         (ffi_call, ffi_closure_osf_inner): Likewise.
1840         (ffi_closure_osf_inner): Likewise.  Mark hidden.
1841         (ffi_call_osf, ffi_closure_osf): Mark hidden.
1842         * src/alpha/ffitarget.h (FFI_LAST_ABI): Tidy definition.
1843         * src/alpha/osf.S (ffi_call_osf, ffi_closure_osf): Mark hidden.
1844         (load_table): Handle 128-bit long double.
1845
1846         * testsuite/libffi.call/float4.c: Add -mieee for alpha.
1847
1848 2007-04-06  Tom Tromey  <tromey@redhat.com>
1849
1850         PR libffi/31491:
1851         * README: Fixed bug in example.
1852
1853 2007-04-03  Jakub Jelinek  <jakub@redhat.com>
1854
1855         * src/closures.c: Include sys/statfs.h.
1856         (_GNU_SOURCE): Define on Linux.
1857         (FFI_MMAP_EXEC_SELINUX): Define.
1858         (selinux_enabled): New variable.
1859         (selinux_enabled_check): New function.
1860         (is_selinux_enabled): Define.
1861         (dlmmap): Use it.
1862
1863 2007-03-24  Uros Bizjak  <ubizjak@gmail.com>
1864
1865         * testsuite/libffi.call/return_fl2.c (return_fl): Mark as static.
1866         Use 'volatile float sum' to create sum of floats to avoid false
1867         negative due to excess precision on ix86 targets.
1868         (main): Ditto.
1869
1870 2007-03-08  Alexandre Oliva  <aoliva@redhat.com>
1871
1872         * src/powerpc/ffi.c (flush_icache): Fix left-over from previous
1873         patch.
1874         (ffi_prep_closure_loc): Remove unneeded casts.  Add needed ones.
1875
1876 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1877
1878         * include/ffi.h.in (ffi_closure_alloc, ffi_closure_free): New.
1879         (ffi_prep_closure_loc): New.
1880         (ffi_prep_raw_closure_loc): New.
1881         (ffi_prep_java_raw_closure_loc): New.
1882         * src/closures.c: New file.
1883         * src/dlmalloc.c [FFI_MMAP_EXEC_WRIT] (struct malloc_segment):
1884         Replace sflags with exec_offset.
1885         [FFI_MMAP_EXEC_WRIT] (mmap_exec_offset, add_segment_exec_offset,
1886         sub_segment_exec_offset): New macros.
1887         (get_segment_flags, set_segment_flags, check_segment_merge): New
1888         macros.
1889         (is_mmapped_segment, is_extern_segment): Use get_segment_flags.
1890         (add_segment, sys_alloc, create_mspace, create_mspace_with_base,
1891         destroy_mspace): Use new macros.
1892         (sys_alloc): Silence warning.
1893         * Makefile.am (libffi_la_SOURCES): Add src/closures.c.
1894         * Makefile.in: Rebuilt.
1895         * src/prep_cif [FFI_CLOSURES] (ffi_prep_closure): Implement in
1896         terms of ffi_prep_closure_loc.
1897         * src/raw_api.c (ffi_prep_raw_closure_loc): Renamed and adjusted
1898         from...
1899         (ffi_prep_raw_closure): ... this.  Re-implement in terms of the
1900         renamed version.
1901         * src/java_raw_api (ffi_prep_java_raw_closure_loc): Renamed and
1902         adjusted from...
1903         (ffi_prep_java_raw_closure): ... this.  Re-implement in terms of
1904         the renamed version.
1905         * src/alpha/ffi.c (ffi_prep_closure_loc): Renamed from
1906         (ffi_prep_closure): ... this.
1907         * src/pa/ffi.c: Likewise.
1908         * src/cris/ffi.c: Likewise.  Adjust.
1909         * src/frv/ffi.c: Likewise.
1910         * src/ia64/ffi.c: Likewise.
1911         * src/mips/ffi.c: Likewise.
1912         * src/powerpc/ffi_darwin.c: Likewise.
1913         * src/s390/ffi.c: Likewise.
1914         * src/sh/ffi.c: Likewise.
1915         * src/sh64/ffi.c: Likewise.
1916         * src/sparc/ffi.c: Likewise.
1917         * src/x86/ffi64.c: Likewise.
1918         * src/x86/ffi.c: Likewise.
1919         (FFI_INIT_TRAMPOLINE): Adjust.
1920         (ffi_prep_raw_closure_loc): Renamed and adjusted from...
1921         (ffi_prep_raw_closure): ... this.
1922         * src/powerpc/ffi.c (ffi_prep_closure_loc): Renamed from
1923         (ffi_prep_closure): ... this.
1924         (flush_icache): Adjust.
1925
1926 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1927
1928         * src/dlmalloc.c: New file, imported version 2.8.3 of Doug
1929         Lea's malloc.
1930
1931 2007-03-01  Brooks Moses  <brooks.moses@codesourcery.com>
1932
1933         * Makefile.am: Add dummy install-pdf target.
1934         * Makefile.in: Regenerate
1935
1936 2007-02-13  Andreas Krebbel  <krebbel1@de.ibm.com>
1937
1938         * src/s390/ffi.c (ffi_prep_args, ffi_prep_cif_machdep,
1939         ffi_closure_helper_SYSV): Add long double handling.
1940
1941 2007-02-02  Jakub Jelinek  <jakub@redhat.com>
1942
1943         * src/powerpc/linux64.S (ffi_call_LINUX64): Move restore of r2
1944         immediately after bctrl instruction.
1945
1946 2007-01-18  Alexandre Oliva  <aoliva@redhat.com>
1947
1948         * Makefile.am (all-recursive, install-recursive,
1949         mostlyclean-recursive, clean-recursive, distclean-recursive,
1950         maintainer-clean-recursive): Add missing targets.
1951         * Makefile.in: Rebuilt.
1952
1953 2006-12-14  Andreas Tobler  <a.tobler@schweiz.org>
1954
1955         * configure.ac: Add TARGET for x86_64-*-darwin*.
1956         * Makefile.am (nodist_libffi_la_SOURCES): Add rules for 64-bit sources
1957         for X86_DARWIN.
1958         * src/x86/ffitarget.h: Set trampoline size for x86_64-*-darwin*.
1959         * src/x86/darwin64.S: New file for x86_64-*-darwin* support.
1960         * configure: Regenerate.
1961         * Makefile.in: Regenerate.
1962         * include/Makefile.in: Regenerate.
1963         * testsuite/Makefile.in: Regenerate.
1964         * testsuite/libffi.special/unwindtest_ffi_call.cc: New test case for
1965         ffi_call only.
1966
1967 2006-12-13  Andreas Tobler <a.tobler@schweiz.org>
1968
1969         * aclocal.m4: Regenerate with aclocal -I .. as written in the
1970         Makefile.am.
1971
1972 2006-10-31  Geoffrey Keating  <geoffk@apple.com>
1973
1974         * src/powerpc/ffi_darwin.c (darwin_adjust_aggregate_sizes): New.
1975         (ffi_prep_cif_machdep): Call darwin_adjust_aggregate_sizes for
1976         Darwin.
1977         * testsuite/libffi.call/nested_struct4.c: Remove Darwin XFAIL.
1978         * testsuite/libffi.call/nested_struct6.c: Remove Darwin XFAIL.
1979
1980 2006-10-10  Paolo Bonzini  <bonzini@gnu.org>
1981             Sandro Tolaini  <tolaini@libero.it>
1982
1983         * configure.ac [i*86-*-darwin*]: Set X86_DARWIN symbol and
1984         conditional.
1985         * configure: Regenerated.
1986         * Makefile.am (nodist_libffi_la_SOURCES) [X86_DARWIN]: New case.
1987         (EXTRA_DIST): Add src/x86/darwin.S.
1988         * Makefile.in: Regenerated.
1989         * include/Makefile.in: Regenerated.
1990         * testsuite/Makefile.in: Regenerated.
1991
1992         * src/x86/ffi.c (ffi_prep_cif_machdep) [X86_DARWIN]: Treat like
1993         X86_WIN32, and additionally align stack to 16 bytes.
1994         * src/x86/darwin.S: New, based on sysv.S.
1995         * src/prep_cif.c (ffi_prep_cif) [X86_DARWIN]: Align > 8-byte structs.
1996
1997 2006-09-12  David Daney  <ddaney@avtrex.com>
1998
1999         PR libffi/23935
2000         * include/Makefile.am: Install both ffi.h and ffitarget.h in
2001         $(libdir)/gcc/$(target_alias)/$(gcc_version)/include.
2002         * aclocal.m4: Regenerated for automake 1.9.6.
2003         * Makefile.in: Regenerated.
2004         * include/Makefile.in: Regenerated.
2005         * testsuite/Makefile.in: Regenerated.
2006
2007 2006-08-17  Andreas Tobler  <a.tobler@schweiz.ch>
2008
2009         * include/ffi_common.h (struct): Revert accidental commit.
2010
2011 2006-08-15  Andreas Tobler  <a.tobler@schweiz.ch>
2012
2013         * include/ffi_common.h: Remove lint directives.
2014         * include/ffi.h.in: Likewise.
2015
2016 2006-07-25  Torsten Schoenfeld  <kaffeetisch@gmx.de>
2017
2018         * include/ffi.h.in (ffi_type_ulong, ffi_type_slong): Define correctly
2019         for 32-bit architectures.
2020         * testsuite/libffi.call/return_ul.c: New test case.
2021
2022 2006-07-19  David Daney  <ddaney@avtrex.com>
2023
2024         * testsuite/libffi.call/closure_fn6.c: Remove xfail for mips,
2025         xfail remains for mips64.
2026
2027 2006-05-23  Carlos O'Donell  <carlos@codesourcery.com>
2028
2029         * Makefile.am: Add install-html target. Add install-html to .PHONY
2030         * Makefile.in: Regenerate.
2031         * aclocal.m4: Regenerate.
2032         * include/Makefile.in: Regenerate.
2033         * testsuite/Makefile.in: Regenerate.
2034
2035 2006-05-18  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2036
2037         * pa/ffi.c (ffi_prep_args_pa32): Load floating point arguments from
2038         stack slot.
2039
2040 2006-04-22  Andreas Tobler  <a.tobler@schweiz.ch>
2041
2042         * README: Remove notice about 'Crazy Comments'.
2043         * src/debug.c: Remove lint directives. Cleanup white spaces.
2044         * src/java_raw_api.c: Likewise.
2045         * src/prep_cif.c: Likewise.
2046         * src/raw_api.c: Likewise.
2047         * src/ffitest.c: Delete. No longer needed, all test cases migrated
2048         to the testsuite.
2049         * src/arm/ffi.c: Remove lint directives.
2050         * src/m32r/ffi.c: Likewise.
2051         * src/pa/ffi.c: Likewise.
2052         * src/powerpc/ffi.c: Likewise.
2053         * src/powerpc/ffi_darwin.c: Likewise.
2054         * src/sh/ffi.c: Likewise.
2055         * src/sh64/ffi.c: Likewise.
2056         * src/x86/ffi.c: Likewise.
2057         * testsuite/libffi.call/float2.c: Likewise.
2058         * testsuite/libffi.call/promotion.c: Likewise.
2059         * testsuite/libffi.call/struct1.c: Likewise.
2060
2061 2006-04-13  Andreas Tobler  <a.tobler@schweiz.ch>
2062
2063         * src/pa/hpux32.S: Correct unwind offset calculation for
2064         ffi_closure_pa32.
2065         * src/pa/linux.S: Likewise.
2066
2067 2006-04-12  James E Wilson  <wilson@specifix.com>
2068
2069         PR libgcj/26483
2070         * src/ia64/ffi.c (stf_spill, ldf_fill): Rewrite as macros.
2071         (hfa_type_load): Call stf_spill.
2072         (hfa_type_store): Call ldf_fill.
2073         (ffi_call): Adjust calls to above routines.  Add local temps for
2074         macro result.
2075
2076 2006-04-10  Matthias Klose  <doko@debian.org>
2077
2078         * testsuite/lib/libffi-dg.exp (libffi-init): Recognize multilib
2079         directory names containing underscores.
2080
2081 2006-04-07  James E Wilson  <wilson@specifix.com>
2082
2083         * testsuite/libffi.call/float4.c: New testcase.
2084
2085 2006-04-05  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2086             Andreas Tobler  <a.tobler@schweiz.ch>
2087
2088         * Makefile.am: Add PA_HPUX port.
2089         * Makefile.in: Regenerate.
2090         * include/Makefile.in: Likewise.
2091         * testsuite/Makefile.in: Likewise.
2092         * configure.ac: Add PA_HPUX rules.
2093         * configure: Regenerate.
2094         * src/pa/ffitarget.h: Rename linux target to PA_LINUX.
2095         Add PA_HPUX and PA64_HPUX.
2096         Rename FFI_LINUX ABI to FFI_PA32 ABI.
2097         (FFI_TRAMPOLINE_SIZE): Define for 32-bit HP-UX targets.
2098         (FFI_TYPE_SMALL_STRUCT2): Define.
2099         (FFI_TYPE_SMALL_STRUCT4): Likewise.
2100         (FFI_TYPE_SMALL_STRUCT8): Likewise.
2101         (FFI_TYPE_SMALL_STRUCT3): Redefine.
2102         (FFI_TYPE_SMALL_STRUCT5): Likewise.
2103         (FFI_TYPE_SMALL_STRUCT6): Likewise.
2104         (FFI_TYPE_SMALL_STRUCT7): Likewise.
2105         * src/pa/ffi.c (ROUND_DOWN): Delete.
2106         (fldw, fstw, fldd, fstd): Use '__asm__'.
2107         (ffi_struct_type): Add support for FFI_TYPE_SMALL_STRUCT2,
2108         FFI_TYPE_SMALL_STRUCT4 and FFI_TYPE_SMALL_STRUCT8.
2109         (ffi_prep_args_LINUX): Rename to ffi_prep_args_pa32. Update comment.
2110         Simplify incrementing of stack slot variable. Change type of local
2111         'n' to unsigned int.
2112         (ffi_size_stack_LINUX): Rename to ffi_size_stack_pa32. Handle long
2113         double on PA_HPUX.
2114         (ffi_prep_cif_machdep): Likewise.
2115         (ffi_call): Likewise.
2116         (ffi_closure_inner_LINUX): Rename to ffi_closure_inner_pa32. Change
2117         return type to ffi_status. Simplify incrementing of stack slot
2118         variable. Only copy floating point argument registers when PA_LINUX
2119         is true. Reformat debug statement.
2120         Add support for FFI_TYPE_SMALL_STRUCT2, FFI_TYPE_SMALL_STRUCT4 and
2121         FFI_TYPE_SMALL_STRUCT8.
2122         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Add 'extern' to
2123         declaration.
2124         (ffi_prep_closure): Make linux trampoline conditional on PA_LINUX.
2125         Add nops to cache flush.  Add trampoline for PA_HPUX.
2126         * src/pa/hpux32.S: New file.
2127         * src/pa/linux.S (ffi_call_LINUX): Rename to ffi_call_pa32. Rename
2128         ffi_prep_args_LINUX to ffi_prep_args_pa32.
2129         Localize labels. Add support for 2, 4 and 8-byte small structs. Handle
2130         unaligned destinations in 3, 5, 6 and 7-byte small structs. Order
2131         argument type checks so that common argument types appear first.
2132         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Rename
2133         ffi_closure_inner_LINUX to ffi_closure_inner_pa32.
2134
2135 2006-03-24  Alan Modra  <amodra@bigpond.net.au>
2136
2137         * src/powerpc/ffitarget.h (enum ffi_abi): Add FFI_LINUX.  Default
2138         for 32-bit using IBM extended double format.  Fix FFI_LAST_ABI.
2139         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Handle linux variant of
2140         FFI_TYPE_LONGDOUBLE.
2141         (ffi_prep_args64): Assert using IBM extended double.
2142         (ffi_prep_cif_machdep): Don't munge FFI_TYPE_LONGDOUBLE type.
2143         Handle FFI_LINUX FFI_TYPE_LONGDOUBLE return and args.
2144         (ffi_call): Handle FFI_LINUX.
2145         (ffi_closure_helper_SYSV): Non FFI_LINUX long double return needs
2146         gpr3 return pointer as for struct return.  Handle FFI_LINUX
2147         FFI_TYPE_LONGDOUBLE return and args.  Don't increment "nf"
2148         unnecessarily.
2149         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Load both f1 and f2
2150         for FFI_TYPE_LONGDOUBLE.  Move epilogue insns into case table.
2151         Don't use r6 as pointer to results, instead use sp offset.  Don't
2152         make a special call to load lr with case table address, instead
2153         use offset from previous call.
2154         * src/powerpc/sysv.S (ffi_call_SYSV): Save long double return.
2155         * src/powerpc/linux64.S (ffi_call_LINUX64): Simplify long double
2156         return.
2157
2158 2006-03-15  Kaz Kojima  <kkojima@gcc.gnu.org>
2159
2160         * src/sh64/ffi.c (ffi_prep_cif_machdep): Handle float arguments
2161         passed with FP registers correctly.
2162         (ffi_closure_helper_SYSV): Likewise.
2163         * src/sh64/sysv.S: Likewise.
2164
2165 2006-03-01  Andreas Tobler  <a.tobler@schweiz.ch>
2166
2167         * testsuite/libffi.special/unwindtest.cc (closure_test_fn): Mark cif,
2168         args and userdata unused.
2169         (closure_test_fn1): Mark cif and userdata unused.
2170         (main): Remove unused res.
2171
2172 2006-02-28  Andreas Tobler  <a.tobler@schweiz.ch>
2173
2174         * testsuite/libffi.call/call.exp: Adjust FSF address. Add test runs for
2175         -O2, -O3, -Os and the warning flags -W -Wall.
2176         * testsuite/libffi.special/special.exp: Likewise.
2177         * testsuite/libffi.call/ffitest.h: Add an __UNUSED__ macro to mark
2178         unused parameter unused for gcc or else do nothing.
2179         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2180         * testsuite/libffi.call/cls_12byte.c (cls_struct_12byte_gn): Mark cif
2181         and userdata unused.
2182         * testsuite/libffi.call/cls_16byte.c (cls_struct_16byte_gn): Likewise.
2183         * testsuite/libffi.call/cls_18byte.c (cls_struct_18byte_gn): Likewise.
2184         * testsuite/libffi.call/cls_19byte.c (cls_struct_19byte_gn): Likewise.
2185         * testsuite/libffi.call/cls_1_1byte.c (cls_struct_1_1byte_gn): Likewise.
2186         * testsuite/libffi.call/cls_20byte.c (cls_struct_20byte_gn): Likewise.
2187         * testsuite/libffi.call/cls_20byte1.c (cls_struct_20byte_gn): Likewise.
2188         * testsuite/libffi.call/cls_24byte.c (cls_struct_24byte_gn): Likewise.
2189         * testsuite/libffi.call/cls_2byte.c (cls_struct_2byte_gn): Likewise.
2190         * testsuite/libffi.call/cls_3_1byte.c (cls_struct_3_1byte_gn): Likewise.
2191         * testsuite/libffi.call/cls_3byte1.c (cls_struct_3byte_gn): Likewise.
2192         * testsuite/libffi.call/cls_3byte2.c (cls_struct_3byte_gn1): Likewise.
2193         * testsuite/libffi.call/cls_4_1byte.c (cls_struct_4_1byte_gn): Likewise.
2194         * testsuite/libffi.call/cls_4byte.c (cls_struct_4byte_gn): Likewise.
2195         * testsuite/libffi.call/cls_5_1_byte.c (cls_struct_5byte_gn): Likewise.
2196         * testsuite/libffi.call/cls_5byte.c (cls_struct_5byte_gn): Likewise.
2197         * testsuite/libffi.call/cls_64byte.c (cls_struct_64byte_gn): Likewise.
2198         * testsuite/libffi.call/cls_6_1_byte.c (cls_struct_6byte_gn): Likewise.
2199         * testsuite/libffi.call/cls_6byte.c (cls_struct_6byte_gn): Likewise.
2200         * testsuite/libffi.call/cls_7_1_byte.c (cls_struct_7byte_gn): Likewise.
2201         * testsuite/libffi.call/cls_7byte.c (cls_struct_7byte_gn): Likewise.
2202         * testsuite/libffi.call/cls_8byte.c (cls_struct_8byte_gn): Likewise.
2203         * testsuite/libffi.call/cls_9byte1.c (cls_struct_9byte_gn): Likewise.
2204         * testsuite/libffi.call/cls_9byte2.c (cls_struct_9byte_gn): Likewise.
2205         * testsuite/libffi.call/cls_align_double.c (cls_struct_align_gn):
2206         Likewise.
2207         * testsuite/libffi.call/cls_align_float.c (cls_struct_align_gn):
2208         Likewise.
2209         * testsuite/libffi.call/cls_align_longdouble.c (cls_struct_align_gn):
2210         Likewise.
2211         * testsuite/libffi.call/cls_align_pointer.c (cls_struct_align_fn): Cast
2212         void* to avoid compiler warning.
2213         (main): Likewise.
2214         (cls_struct_align_gn): Mark cif and userdata unused.
2215         * testsuite/libffi.call/cls_align_sint16.c (cls_struct_align_gn):
2216         Likewise.
2217         * testsuite/libffi.call/cls_align_sint32.c (cls_struct_align_gn):
2218         Likewise.
2219         * testsuite/libffi.call/cls_align_sint64.c (cls_struct_align_gn):
2220         Likewise.
2221         * testsuite/libffi.call/cls_align_uint16.c (cls_struct_align_gn):
2222         Likewise.
2223         * testsuite/libffi.call/cls_align_uint32.c (cls_struct_align_gn):
2224         Likewise.
2225         * testsuite/libffi.call/cls_double.c (cls_ret_double_fn): Likewise.
2226         * testsuite/libffi.call/cls_float.c (cls_ret_float_fn): Likewise.
2227         * testsuite/libffi.call/cls_multi_schar.c (test_func_gn): Mark cif and
2228         data unused.
2229         (main): Cast res_call to silence gcc.
2230         * testsuite/libffi.call/cls_multi_sshort.c (test_func_gn): Mark cif and
2231         data unused.
2232         (main): Cast res_call to silence gcc.
2233         * testsuite/libffi.call/cls_multi_sshortchar.c (test_func_gn): Mark cif
2234         and data unused.
2235         (main): Cast res_call to silence gcc.
2236         * testsuite/libffi.call/cls_multi_uchar.c (test_func_gn): Mark cif and
2237         data unused.
2238         (main): Cast res_call to silence gcc.
2239         * testsuite/libffi.call/cls_multi_ushort.c (test_func_gn): Mark cif and
2240         data unused.
2241         (main): Cast res_call to silence gcc.
2242         * testsuite/libffi.call/cls_multi_ushortchar.c (test_func_gn): Mark cif
2243         and data unused.
2244         (main): Cast res_call to silence gcc.
2245         * testsuite/libffi.call/cls_schar.c (cls_ret_schar_fn): Mark cif and
2246         userdata unused.
2247         (cls_ret_schar_fn): Cast printf parameter to silence gcc.
2248         * testsuite/libffi.call/cls_sint.c (cls_ret_sint_fn): Mark cif and
2249         userdata unused.
2250         (cls_ret_sint_fn): Cast printf parameter to silence gcc.
2251         * testsuite/libffi.call/cls_sshort.c (cls_ret_sshort_fn): Mark cif and
2252         userdata unused.
2253         (cls_ret_sshort_fn): Cast printf parameter to silence gcc.
2254         * testsuite/libffi.call/cls_uchar.c (cls_ret_uchar_fn):  Mark cif and
2255         userdata unused.
2256         (cls_ret_uchar_fn): Cast printf parameter to silence gcc.
2257         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Mark cif and
2258         userdata unused.
2259         (cls_ret_uint_fn): Cast printf parameter to silence gcc.
2260         * testsuite/libffi.call/cls_ulonglong.c (cls_ret_ulonglong_fn): Mark cif
2261         and userdata unused.
2262         * testsuite/libffi.call/cls_ushort.c (cls_ret_ushort_fn): Mark cif and
2263         userdata unused.
2264         (cls_ret_ushort_fn): Cast printf parameter to silence gcc.
2265         * testsuite/libffi.call/float.c (floating): Remove unused parameter e.
2266         * testsuite/libffi.call/float1.c (main): Remove unused variable i.
2267         Cleanup white spaces.
2268         * testsuite/libffi.call/negint.c (checking): Remove unused variable i.
2269         * testsuite/libffi.call/nested_struct.c (cls_struct_combined_gn): Mark
2270         cif and userdata unused.
2271         * testsuite/libffi.call/nested_struct1.c (cls_struct_combined_gn):
2272         Likewise.
2273         * testsuite/libffi.call/nested_struct10.c (B_gn): Likewise.
2274         * testsuite/libffi.call/nested_struct2.c (B_fn): Adjust printf
2275         formatters to silence gcc.
2276         (B_gn): Mark cif and userdata unused.
2277         * testsuite/libffi.call/nested_struct3.c (B_gn): Mark cif and userdata
2278         unused.
2279         * testsuite/libffi.call/nested_struct4.c: Mention related PR.
2280         (B_gn): Mark cif and userdata unused.
2281         * testsuite/libffi.call/nested_struct5.c (B_gn): Mark cif and userdata
2282         unused.
2283         * testsuite/libffi.call/nested_struct6.c: Mention related PR.
2284         (B_gn): Mark cif and userdata unused.
2285         * testsuite/libffi.call/nested_struct7.c (B_gn): Mark cif and userdata
2286         unused.
2287         * testsuite/libffi.call/nested_struct8.c (B_gn): Likewise.
2288         * testsuite/libffi.call/nested_struct9.c (B_gn): Likewise.
2289         * testsuite/libffi.call/problem1.c (stub): Likewise.
2290         * testsuite/libffi.call/pyobjc-tc.c (main): Cast the result to silence
2291         gcc.
2292         * testsuite/libffi.call/return_fl2.c (return_fl): Add the note mentioned
2293         in the last commit for this test case in the test case itself.
2294         * testsuite/libffi.call/closure_fn0.c (closure_test_fn0): Mark cif as
2295         unused.
2296         * testsuite/libffi.call/closure_fn1.c (closure_test_fn1): Likewise.
2297         * testsuite/libffi.call/closure_fn2.c (closure_test_fn2): Likewise.
2298         * testsuite/libffi.call/closure_fn3.c (closure_test_fn3): Likewise.
2299         * testsuite/libffi.call/closure_fn4.c (closure_test_fn0): Likewise.
2300         * testsuite/libffi.call/closure_fn5.c (closure_test_fn5): Likewise.
2301         * testsuite/libffi.call/closure_fn6.c (closure_test_fn0): Likewise.
2302
2303 2006-02-22  Kaz Kojima  <kkojima@gcc.gnu.org>
2304
2305         * src/sh/sysv.S: Fix register numbers in the FDE for
2306         ffi_closure_SYSV.
2307
2308 2006-02-20  Andreas Tobler  <a.tobler@schweiz.ch>
2309
2310         * testsuite/libffi.call/return_fl2.c (return_fl): Remove static
2311         declaration to avoid a false negative on ix86. See PR323.
2312
2313 2006-02-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2314
2315         * src/sh/ffi.c (ffi_closure_helper_SYSV): Remove unused variable
2316         and cast integer to void * if needed.  Update the pointer to
2317         the FP register saved area correctly.
2318
2319 2006-02-17  Andreas Tobler  <a.tobler@schweiz.ch>
2320
2321         * testsuite/libffi.call/nested_struct6.c: XFAIL this test until PR25630
2322         is fixed.
2323         * testsuite/libffi.call/nested_struct4.c: Likewise.
2324
2325 2006-02-16  Andreas Tobler  <a.tobler@schweiz.ch>
2326
2327         * testsuite/libffi.call/return_dbl.c: New test case.
2328         * testsuite/libffi.call/return_dbl1.c: Likewise.
2329         * testsuite/libffi.call/return_dbl2.c: Likewise.
2330         * testsuite/libffi.call/return_fl.c: Likewise.
2331         * testsuite/libffi.call/return_fl1.c: Likewise.
2332         * testsuite/libffi.call/return_fl2.c: Likewise.
2333         * testsuite/libffi.call/return_fl3.c: Likewise.
2334         * testsuite/libffi.call/closure_fn6.c: Likewise.
2335
2336         * testsuite/libffi.call/nested_struct2.c: Remove ffi_type_mylong
2337         definition.
2338         * testsuite/libffi.call/ffitest.h: Add ffi_type_mylong definition
2339         here to be used by other test cases too.
2340
2341         * testsuite/libffi.call/nested_struct10.c: New test case.
2342         * testsuite/libffi.call/nested_struct9.c: Likewise.
2343         * testsuite/libffi.call/nested_struct8.c: Likewise.
2344         * testsuite/libffi.call/nested_struct7.c: Likewise.
2345         * testsuite/libffi.call/nested_struct6.c: Likewise.
2346         * testsuite/libffi.call/nested_struct5.c: Likewise.
2347         * testsuite/libffi.call/nested_struct4.c: Likewise.
2348
2349 2006-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
2350
2351         * configure.ac: Enable libffi for sparc64-*-freebsd*.
2352         * configure: Rebuilt.
2353
2354 2006-01-18  Jakub Jelinek  <jakub@redhat.com>
2355
2356         * src/powerpc/sysv.S (smst_two_register): Don't call __ashldi3,
2357         instead do the shifting inline.
2358         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't compute %r5
2359         shift count unconditionally.  Simplify load sequences for 1, 2, 3, 4
2360         and 8 byte structs, for the remaining struct sizes don't call
2361         __lshrdi3, instead do the shifting inline.
2362
2363 2005-12-07  Thiemo Seufer  <ths@networkno.de>
2364
2365         * src/mips/ffitarget.h: Remove obsolete sgidefs.h include. Add
2366         missing parentheses.
2367         * src/mips/o32.S (ffi_call_O32): Code formatting. Define
2368         and use A3_OFF, FP_OFF, RA_OFF. Micro-optimizations.
2369         (ffi_closure_O32): Likewise, but with newly defined A3_OFF2,
2370         A2_OFF2, A1_OFF2, A0_OFF2, RA_OFF2, FP_OFF2, S0_OFF2, GP_OFF2,
2371         V1_OFF2, V0_OFF2, FA_1_1_OFF2, FA_1_0_OFF2, FA_0_1_OFF2,
2372         FA_0_0_OFF2.
2373         * src/mips/ffi.c (ffi_prep_args): Code formatting. Fix
2374         endianness bugs.
2375         (ffi_prep_closure): Improve trampoline instruction scheduling.
2376         (ffi_closure_mips_inner_O32): Fix endianness bugs.
2377
2378 2005-12-03  Alan Modra  <amodra@bigpond.net.au>
2379
2380         * src/powerpc/ffi.c: Formatting.
2381         (ffi_prep_args_SYSV): Avoid possible aliasing problems by using unions.
2382         (ffi_prep_args64): Likewise.
2383
2384 2005-09-30  Geoffrey Keating  <geoffk@apple.com>
2385
2386         * testsuite/lib/libffi-dg.exp (libffi_target_compile): For
2387         darwin, use -shared-libgcc not -lgcc_s, and explain why.
2388
2389 2005-09-26  Tom Tromey  <tromey@redhat.com>
2390
2391         * testsuite/libffi.call/float1.c (value_type): New typedef.
2392         (CANARY): New define.
2393         (main): Check for result buffer overflow.
2394         * src/powerpc/linux64.S: Handle linux64 long double returns.
2395         * src/powerpc/ffi.c (FLAG_RETURNS_128BITS): New constant.
2396         (ffi_prep_cif_machdep): Handle linux64 long double returns.
2397
2398 2005-08-25  Alan Modra  <amodra@bigpond.net.au>
2399
2400         PR target/23404
2401         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Correct placement of stack
2402         homed fp args.
2403         (ffi_status ffi_prep_cif_machdep): Correct stack sizing for same.
2404
2405 2005-08-11  Jakub Jelinek  <jakub@redhat.com>
2406
2407         * configure.ac (HAVE_HIDDEN_VISIBILITY_ATTRIBUTE): New test.
2408         (AH_BOTTOM): Add FFI_HIDDEN definition.
2409         * configure: Rebuilt.
2410         * fficonfig.h.in: Rebuilt.
2411         * src/powerpc/ffi.c (hidden): Remove.
2412         (ffi_closure_LINUX64, ffi_prep_args64, ffi_call_LINUX64,
2413         ffi_closure_helper_LINUX64): Use FFI_HIDDEN instead of hidden.
2414         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64,
2415         .ffi_closure_LINUX64): Use FFI_HIDDEN instead of .hidden.
2416         * src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV): Remove,
2417         add FFI_HIDDEN to its prototype.
2418         (ffi_closure_SYSV_inner): New.
2419         * src/x86/sysv.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2420         * src/x86/win32.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2421
2422 2005-08-10  Alfred M. Szmidt  <ams@gnu.org>
2423
2424         PR libffi/21819:
2425         * configure: Rebuilt.
2426         * configure.ac: Handle i*86-*-gnu*.
2427
2428 2005-08-09  Jakub Jelinek  <jakub@redhat.com>
2429
2430         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Use
2431         DW_CFA_offset_extended_sf rather than
2432         DW_CFA_GNU_negative_offset_extended.
2433         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise.
2434
2435 2005-07-22  SUGIOKA Toshinobu  <sugioka@itonet.co.jp>
2436
2437         * src/sh/sysv.S (ffi_call_SYSV): Stop argument popping correctly
2438         on sh3.
2439         (ffi_closure_SYSV): Change the stack layout for sh3 struct argument.
2440         * src/sh/ffi.c (ffi_prep_args): Fix sh3 argument copy, when it is
2441         partially on register.
2442         (ffi_closure_helper_SYSV): Likewise.
2443         (ffi_prep_cif_machdep): Don't set too many cif->flags.
2444
2445 2005-07-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2446
2447         * src/sh/ffi.c (ffi_call): Handle small structures correctly.
2448         Remove empty line.
2449         * src/sh64/ffi.c (simple_type): Remove.
2450         (return_type): Handle small structures correctly.
2451         (ffi_prep_args): Likewise.
2452         (ffi_call): Likewise.
2453         (ffi_closure_helper_SYSV): Likewise.
2454         * src/sh64/sysv.S (ffi_call_SYSV): Handle 1, 2 and 4-byte return.
2455         Emit position independent code if PIC and remove wrong datalabel
2456         prefixes from EH data.
2457
2458 2005-07-19  Andreas Tobler  <a.tobler@schweiz.ch>
2459
2460         * Makefile.am (nodist_libffi_la_SOURCES): Add POWERPC_FREEBSD.
2461         * Makefile.in: Regenerate.
2462         * include/Makefile.in: Likewise.
2463         * testsuite/Makefile.in: Likewise.
2464         * configure.ac: Add POWERPC_FREEBSD rules.
2465         * configure: Regenerate.
2466         * src/powerpc/ffitarget.h: Add POWERPC_FREEBSD rules.
2467         (FFI_SYSV_TYPE_SMALL_STRUCT): Define.
2468         * src/powerpc/ffi.c: Add flags to handle small structure returns
2469         in ffi_call_SYSV.
2470         (ffi_prep_cif_machdep): Handle small structures for SYSV 4 ABI.
2471         Aka FFI_SYSV.
2472         (ffi_closure_helper_SYSV): Likewise.
2473         * src/powerpc/ppc_closure.S: Add return types for small structures.
2474         * src/powerpc/sysv.S: Add bits to handle small structures for
2475         final SYSV 4 ABI.
2476
2477 2005-07-10  Andreas Tobler  <a.tobler@schweiz.ch>
2478
2479         * testsuite/libffi.call/cls_5_1_byte.c: New test file.
2480         * testsuite/libffi.call/cls_6_1_byte.c: Likewise.
2481         * testsuite/libffi.call/cls_7_1_byte.c: Likewise.
2482
2483 2005-07-05  Randolph Chung  <tausq@debian.org>
2484
2485         * src/pa/ffi.c (ffi_struct_type): Rename FFI_TYPE_SMALL_STRUCT1
2486         as FFI_TYPE_SMALL_STRUCT3.  Break out handling for 5-7 byte
2487         structures.  Kill compilation warnings.
2488         (ffi_closure_inner_LINUX): Print return values as hex in debug
2489         message.  Rename FFI_TYPE_SMALL_STRUCT1 as FFI_TYPE_SMALL_STRUCT3.
2490         Properly handle 5-7 byte structure returns.
2491         * src/pa/ffitarget.h (FFI_TYPE_SMALL_STRUCT1)
2492         (FFI_TYPE_SMALL_STRUCT2): Remove.
2493         (FFI_TYPE_SMALL_STRUCT3, FFI_TYPE_SMALL_STRUCT5)
2494         (FFI_TYPE_SMALL_STRUCT6, FFI_TYPE_SMALL_STRUCT7): Define.
2495         * src/pa/linux.S: Mark source file as using PA1.1 assembly.
2496         (checksmst1, checksmst2): Remove.
2497         (checksmst3): Optimize handling of 3-byte struct returns.
2498         (checksmst567): Properly handle 5-7 byte struct returns.
2499
2500 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2501
2502         PR libgcj/21943
2503         * src/mips/n32.S: Enforce PIC code.
2504         * src/mips/o32.S: Likewise.
2505
2506 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2507
2508         * configure.ac: Treat i*86-*-solaris2.10 and up as X86_64.
2509         * configure: Regenerate.
2510
2511 2005-06-01  Alan Modra  <amodra@bigpond.net.au>
2512
2513         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't use JUMPTARGET
2514         to call ffi_closure_helper_SYSV.  Append @local instead.
2515         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise for ffi_prep_args_SYSV.
2516
2517 2005-05-17  Kelley Cook  <kcook@gcc.gnu.org>
2518
2519         * configure.ac: Use AC_C_BIGENDIAN instead of AC_C_BIGENDIAN_CROSS.
2520         Use AC_CHECK_SIZEOF instead of AC_COMPILE_CHECK_SIZEOF.
2521         * Makefile.am (ACLOCAL_AMFLAGS): Remove -I ../config.
2522         * aclocal.m4, configure, fficonfig.h.in, Makefile.in,
2523         include/Makefile.in, testsuite/Makefile.in: Regenerate.
2524
2525 2005-05-09  Mike Stump  <mrs@apple.com>
2526
2527         * configure: Regenerate.
2528
2529 2005-05-08  Richard Henderson  <rth@redhat.com>
2530
2531         PR libffi/21285
2532         * src/alpha/osf.S: Update unwind into to match code.
2533
2534 2005-05-04  Andreas Degert <ad@papyrus-gmbh.de>
2535             Richard Henderson  <rth@redhat.com>
2536
2537         * src/x86/ffi64.c (ffi_prep_cif_machdep): Save sse-used flag in
2538         bit 11 of flags.
2539         (ffi_call): Mask return type field.  Pass ssecount to ffi_call_unix64.
2540         (ffi_prep_closure): Set carry bit if sse-used flag set.
2541         * src/x86/unix64.S (ffi_call_unix64): Add ssecount argument.
2542         Only load sse registers if ssecount non-zero.
2543         (ffi_closure_unix64): Only save sse registers if carry set on entry.
2544
2545 2005-04-29  Ralf Corsepius  <ralf.corsepius@rtems.org>
2546
2547         * configure.ac: Add i*86-*-rtems*, sparc*-*-rtems*,
2548         powerpc-*rtems*, arm*-*-rtems*, sh-*-rtems*.
2549         * configure: Regenerate.
2550
2551 2005-04-20  Hans-Peter Nilsson  <hp@axis.com>
2552
2553         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): In regsub use,
2554         have Tcl8.3-compatible intermediate variable.
2555
2556 2005-04-18  Simon Posnjak <simon.posnjak@siol.net>
2557             Hans-Peter Nilsson  <hp@axis.com>
2558
2559         * Makefile.am: Add CRIS support.
2560         * configure.ac: Likewise.
2561         * Makefile.in, configure, testsuite/Makefile.in,
2562         include/Makefile.in: Regenerate.
2563         * src/cris: New directory.
2564         * src/cris/ffi.c, src/cris/sysv.S, src/cris/ffitarget.h: New files.
2565         * src/prep_cif.c (ffi_prep_cif): Wrap in #ifndef __CRIS__.
2566
2567         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): Replace \n with
2568         \r?\n in output tests.
2569
2570 2005-04-12  Mike Stump  <mrs@apple.com>
2571
2572         * configure: Regenerate.
2573
2574 2005-03-30  Hans Boehm  <Hans.Boehm@hp.com>
2575
2576         * src/ia64/ffitarget.h (ffi_arg): Use long long instead of DI.
2577
2578 2005-03-30  Steve Ellcey  <sje@cup.hp.com>
2579
2580         * src/ia64/ffitarget.h (ffi_arg) ADD DI attribute.
2581         (ffi_sarg) Ditto.
2582         * src/ia64/unix.S (ffi_closure_unix): Extend gp
2583         to 64 bits in ILP32 mode.
2584         Load 64 bits even for short data.
2585
2586 2005-03-23  Mike Stump  <mrs@apple.com>
2587
2588         * src/powerpc/darwin.S: Update for -m64 multilib.
2589         * src/powerpc/darwin_closure.S: Likewise.
2590
2591 2005-03-21  Zack Weinberg  <zack@codesourcery.com>
2592
2593         * configure.ac: Do not invoke TL_AC_GCC_VERSION.
2594         Do not set tool_include_dir.
2595         * aclocal.m4, configure, Makefile.in, testsuite/Makefile.in:
2596         Regenerate.
2597         * include/Makefile.am: Set gcc_version and toollibffidir.
2598         * include/Makefile.in: Regenerate.
2599
2600 2005-02-22  Andrew Haley  <aph@redhat.com>
2601
2602         * src/powerpc/ffi.c (ffi_prep_cif_machdep): Bump alignment to
2603         odd-numbered register pairs for 64-bit integer types.
2604
2605 2005-02-23  Andreas Tobler  <a.tobler@schweiz.ch>
2606
2607         PR libffi/20104
2608         * testsuite/libffi.call/return_ll1.c: New test case.
2609
2610 2005-02-11  Janis Johnson  <janis187@us.ibm.com>
2611
2612         * testsuite/libffi.call/cls_align_longdouble.c: Remove dg-options.
2613         * testsuite/libffi.call/float.c: Ditto.
2614         * testsuite/libffi.call/float2.c: Ditto.
2615         * testsuite/libffi.call/float3.c: Ditto.
2616
2617 2005-02-08  Andreas Tobler  <a.tobler@schweiz.ch>
2618
2619         * src/frv/ffitarget.h: Remove PPC stuff which does not belong to frv.
2620
2621 2005-01-12  Eric Botcazou  <ebotcazou@libertysurf.fr>
2622
2623         * testsuite/libffi.special/special.exp (cxx_options): Add
2624         -shared-libgcc.
2625
2626 2004-12-31  Richard Henderson  <rth@redhat.com>
2627
2628         * src/types.c (FFI_AGGREGATE_TYPEDEF): Remove.
2629         (FFI_TYPEDEF): Rename from FFI_INTEGRAL_TYPEDEF.  Replace size and
2630         offset parameters with a type parameter; deduce size and structure
2631         alignment.  Update all users.
2632
2633 2004-12-31  Richard Henderson  <rth@redhat.com>
2634
2635         * src/types.c (FFI_TYPE_POINTER): Define with sizeof.
2636         (FFI_TYPE_LONGDOUBLE): Fix for ia64.
2637         * src/ia64/ffitarget.h (struct ffi_ia64_trampoline_struct): Move
2638         into ffi_prep_closure.
2639         * src/ia64/ia64_flags.h, src/ia64/ffi.c, src/ia64/unix.S: Rewrite
2640         from scratch.
2641
2642 2004-12-27  Richard Henderson  <rth@redhat.com>
2643
2644         * src/x86/unix64.S: Fix typo in unwind info.
2645
2646 2004-12-25  Richard Henderson  <rth@redhat.com>
2647
2648         * src/x86/ffi64.c (struct register_args): Rename from stackLayout.
2649         (enum x86_64_reg_class): Add X86_64_COMPLEX_X87_CLASS.
2650         (merge_classes): Check for it.
2651         (SSE_CLASS_P): New.
2652         (classify_argument): Pass byte_offset by value; perform all updates
2653         inside struct case.
2654         (examine_argument): Add classes argument; handle
2655         X86_64_COMPLEX_X87_CLASS.
2656         (ffi_prep_args): Merge into ...
2657         (ffi_call): ... here.  Share stack frame with ffi_call_unix64.
2658         (ffi_prep_cif_machdep): Setup cif->flags for proper structure return.
2659         (ffi_fill_return_value): Remove.
2660         (ffi_prep_closure): Remove dead assert.
2661         (ffi_closure_unix64_inner): Rename from ffi_closure_UNIX64_inner.
2662         Rewrite to use struct register_args instead of va_list.  Create
2663         flags for handling structure returns.
2664         * src/x86/unix64.S: Remove dead strings.
2665         (ffi_call_unix64): Rename from ffi_call_UNIX64.  Rewrite to share
2666         stack frame with ffi_call.  Handle structure returns properly.
2667         (float2sse, floatfloat2sse, double2sse): Remove.
2668         (sse2float, sse2double, sse2floatfloat): Remove.
2669         (ffi_closure_unix64): Rename from ffi_closure_UNIX64.  Rewrite
2670         to handle structure returns properly.
2671
2672 2004-12-08  David Edelsohn  <edelsohn@gnu.org>
2673
2674         * Makefile.am (AM_MAKEFLAGS): Remove duplicate LIBCFLAGS and
2675         PICFLAG.
2676         * Makefile.in: Regenerated.
2677
2678 2004-12-02  Richard Sandiford  <rsandifo@redhat.com>
2679
2680         * configure.ac: Use TL_AC_GCC_VERSION to set gcc_version.
2681         * configure, aclocal.m4, Makefile.in: Regenerate.
2682         * include/Makefile.in, testsuite/Makefile.in: Regenerate.
2683
2684 2004-11-29  Kelley Cook  <kcook@gcc.gnu.org>
2685
2686         * configure: Regenerate for libtool change.
2687
2688 2004-11-25  Kelley Cook  <kcook@gcc.gnu.org>
2689
2690         * configure: Regenerate for libtool reversion.
2691
2692 2004-11-24  Kelley Cook  <kcook@gcc.gnu.org>
2693
2694         * configure: Regenerate for libtool change.
2695
2696 2004-11-23  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2697
2698         * testsuite/lib/libffi-dg.exp: Use new procs in target-libpath.exp.
2699
2700 2004-11-23  Richard Sandiford  <rsandifo@redhat.com>
2701
2702         * src/mips/o32.S (ffi_call_O32, ffi_closure_O32): Use jalr instead
2703         of jal.  Use an absolute encoding for the frame information.
2704
2705 2004-11-23  Kelley Cook  <kcook@gcc.gnu.org>
2706
2707         * Makefile.am: Remove no-dependencies.  Add ACLOCAL_AMFLAGS.
2708         * acinclude.m4: Delete logic for sincludes.
2709         * aclocal.m4, Makefile.in, configure: Regenerate.
2710         * include/Makefile: Likewise.
2711         * testsuite/Makefile: Likewise.
2712
2713 2004-11-22  Eric Botcazou  <ebotcazou@libertysurf.fr>
2714
2715         * src/sparc/ffi.c (ffi_prep_closure): Align doubles and 64-bit integers
2716         on a 8-byte boundary.
2717         * src/sparc/v8.S (ffi_closure_v8): Reserve frame space for arguments.
2718
2719 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2720
2721         * src/arm/ffi.c (ffi_prep_cif_machdep): Handle functions that return
2722         long long values.  Round stack allocation to a multiple of 8 bytes
2723         for ATPCS compatibility.
2724         * src/arm/sysv.S (ffi_call_SYSV): Rework to avoid use of APCS register
2725         names.  Handle returning long long types.  Add Thumb and interworking
2726         support.  Improve soft-float code.
2727
2728 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2729
2730         * testsuite/lib/libffi-db.exp (load_gcc_lib): New function.
2731         (libffi_exit): New function.
2732         (libffi_init): Build the testglue wrapper if needed.
2733
2734 2004-10-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2735
2736         PR other/18138
2737         * testsuite/lib/libffi-dg.exp: Accept more than one multilib libgcc.
2738
2739 2004-10-25  Kazuhiro Inaoka <inaoka.kazuhiro@renesas.com>
2740
2741         * src/m32r/libffitarget.h (FFI_CLOSURES): Set to 0.
2742
2743 2004-10-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2744
2745         * src/sh/sysv.S (ffi_call_SYSV): Don't align for double data.
2746         * testsuite/libffi.call/float3.c: New test case.
2747
2748 2004-10-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2749
2750         * src/sh/ffi.c (ffi_prep_closure): Set T bit in trampoline for
2751         the function returning a structure pointed with R2.
2752         * src/sh/sysv.S (ffi_closure_SYSV): Use R2 as the pointer to
2753         the structure return value if T bit set.  Emit position
2754         independent code and EH data if PIC.
2755
2756 2004-10-13  Kazuhiro Inaoka  <inaoka.kazuhiro@renesas.com>
2757
2758         * Makefile.am: Add m32r support.
2759         * configure.ac: Likewise.
2760         * Makefile.in: Regenerate.
2761         * confiugre: Regenerate.
2762         * src/types.c: Add m32r port to FFI_INTERNAL_TYPEDEF
2763         (uint64, sint64, double, longdouble)
2764         * src/m32r: New directory.
2765         * src/m32r/ffi.c: New file.
2766         * src/m32r/sysv.S: Likewise.
2767         * src/m32r/ffitarget.h: Likewise.
2768
2769 2004-10-02  Kaz Kojima  <kkojima@gcc.gnu.org>
2770
2771         * testsuite/libffi.call/negint.c: New test case.
2772
2773 2004-09-14  H.J. Lu  <hongjiu.lu@intel.com>
2774
2775         PR libgcj/17465
2776         * testsuite/lib/libffi-dg.exp: Don't use global ld_library_path.
2777         Set up LD_LIBRARY_PATH, SHLIB_PATH, LD_LIBRARYN32_PATH,
2778         LD_LIBRARY64_PATH, LD_LIBRARY_PATH_32, LD_LIBRARY_PATH_64 and
2779         DYLD_LIBRARY_PATH.
2780
2781 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2782
2783         * testsuite/libffi.call/many_win32.c: Remove whitespaces.
2784         * testsuite/libffi.call/promotion.c: Likewise.
2785         * testsuite/libffi.call/return_ll.c: Remove unused var. Cleanup
2786         whitespaces.
2787         * testsuite/libffi.call/return_sc.c: Likewise.
2788         * testsuite/libffi.call/return_uc.c: Likewise.
2789
2790 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2791
2792         * src/powerpc/darwin.S: Fix comments and identation.
2793         * src/powerpc/darwin_closure.S: Likewise.
2794
2795 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2796
2797         * src/powerpc/ffi_darwin.c: Add flag for longdouble return values.
2798         (ffi_prep_args): Handle longdouble arguments.
2799         (ffi_prep_cif_machdep): Set flags for longdouble. Calculate space for
2800         longdouble.
2801         (ffi_closure_helper_DARWIN): Add closure handling for longdouble.
2802         * src/powerpc/darwin.S (_ffi_call_DARWIN): Add handling of longdouble
2803         values.
2804         * src/powerpc/darwin_closure.S (_ffi_closure_ASM): Likewise.
2805         * src/types.c: Defined longdouble size and alignment for darwin.
2806
2807 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2808
2809         * src/powerpc/aix.S: Remove whitespaces.
2810         * src/powerpc/aix_closure.S: Likewise.
2811         * src/powerpc/asm.h: Likewise.
2812         * src/powerpc/ffi.c: Likewise.
2813         * src/powerpc/ffitarget.h: Likewise.
2814         * src/powerpc/linux64.S: Likewise.
2815         * src/powerpc/linux64_closure.S: Likewise.
2816         * src/powerpc/ppc_closure.S: Likewise.
2817         * src/powerpc/sysv.S: Likewise.
2818
2819 2004-08-30  Anthony Green  <green@redhat.com>
2820
2821         * Makefile.am: Add frv support.
2822         * Makefile.in, testsuite/Makefile.in: Rebuilt.
2823         * configure.ac: Read configure.host.
2824         * configure.in: Read configure.host.
2825         * configure.host: New file.  frv-elf needs libgloss.
2826         * include/ffi.h.in: Force ffi_closure to have a nice big (8)
2827         alignment.  This is needed to frv and shouldn't harm the others.
2828         * include/ffi_common.h (ALIGN_DOWN): New macro.
2829         * src/frv/ffi.c, src/frv/ffitarget.h, src/frv/eabi.S: New files.
2830
2831 2004-08-24  David Daney  <daney@avtrex.com>
2832
2833         * testsuite/libffi.call/closure_fn0.c: Xfail mips64* instead of mips*.
2834         * testsuite/libffi.call/closure_fn1.c: Likewise.
2835         * testsuite/libffi.call/closure_fn2.c  Likewise.
2836         * testsuite/libffi.call/closure_fn3.c: Likewise.
2837         * testsuite/libffi.call/closure_fn4.c: Likewise.
2838         * testsuite/libffi.call/closure_fn5.c: Likewise.
2839         * testsuite/libffi.call/cls_18byte.c: Likewise.
2840         * testsuite/libffi.call/cls_19byte.c: Likewise.
2841         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2842         * testsuite/libffi.call/cls_20byte.c: Likewise.
2843         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2844         * testsuite/libffi.call/cls_24byte.c: Likewise.
2845         * testsuite/libffi.call/cls_2byte.c: Likewise.
2846         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2847         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2848         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2849         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2850         * testsuite/libffi.call/cls_4byte.c: Likewise.
2851         * testsuite/libffi.call/cls_64byte.c: Likewise.
2852         * testsuite/libffi.call/cls_6byte.c: Likewise.
2853         * testsuite/libffi.call/cls_7byte.c: Likewise.
2854         * testsuite/libffi.call/cls_8byte.c: Likewise.
2855         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2856         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2857         * testsuite/libffi.call/cls_align_double.c: Likewise.
2858         * testsuite/libffi.call/cls_align_float.c: Likewise.
2859         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2860         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2861         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2862         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2863         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2864         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2865         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2866         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
2867         * testsuite/libffi.call/cls_double.c: Likewise.
2868         * testsuite/libffi.call/cls_float.c: Likewise.
2869         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2870         * testsuite/libffi.call/cls_multi_sshort.c: Likewise.
2871         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2872         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2873         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2874         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2875         * testsuite/libffi.call/cls_schar.c: Likewise.
2876         * testsuite/libffi.call/cls_sint.c: Likewise.
2877         * testsuite/libffi.call/cls_sshort.c: Likewise.
2878         * testsuite/libffi.call/cls_uchar.c: Likewise.
2879         * testsuite/libffi.call/cls_uint.c: Likewise.
2880         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2881         * testsuite/libffi.call/cls_ushort.c: Likewise.
2882         * testsuite/libffi.call/nested_struct.c: Likewise.
2883         * testsuite/libffi.call/nested_struct1.c: Likewise.
2884         * testsuite/libffi.call/nested_struct2.c: Likewise.
2885         * testsuite/libffi.call/nested_struct3.c: Likewise.
2886         * testsuite/libffi.call/problem1.c: Likewise.
2887         * testsuite/libffi.special/unwindtest.cc: Likewise.
2888         * testsuite/libffi.call/cls_12byte.c: Likewise and set return value
2889         to zero.
2890         * testsuite/libffi.call/cls_16byte.c: Likewise.
2891         * testsuite/libffi.call/cls_5byte.c: Likewise.
2892
2893 2004-08-23  David Daney <daney@avtrex.com>
2894
2895         PR libgcj/13141
2896         * src/mips/ffitarget.h (FFI_O32_SOFT_FLOAT): New ABI.
2897         * src/mips/ffi.c (ffi_prep_args): Fix alignment calculation.
2898         (ffi_prep_cif_machdep): Handle FFI_O32_SOFT_FLOAT floating point
2899         parameters and return types.
2900         (ffi_call): Handle FFI_O32_SOFT_FLOAT ABI.
2901         (ffi_prep_closure): Ditto.
2902         (ffi_closure_mips_inner_O32): Handle FFI_O32_SOFT_FLOAT ABI, fix
2903         alignment calculations.
2904         * src/mips/o32.S (ffi_closure_O32): Don't use floating point
2905         instructions if FFI_O32_SOFT_FLOAT, make stack frame ABI compliant.
2906
2907 2004-08-14  Casey Marshall <csm@gnu.org>
2908
2909         * src/mips/ffi.c (ffi_pref_cif_machdep): set `cif->flags' to
2910         contain `FFI_TYPE_UINT64' as return type for any 64-bit
2911         integer (O32 ABI only).
2912         (ffi_prep_closure): new function.
2913         (ffi_closure_mips_inner_O32): new function.
2914         * src/mips/ffitarget.h: Define `FFI_CLOSURES' and
2915         `FFI_TRAMPOLINE_SIZE' appropriately if the ABI is o32.
2916         * src/mips/o32.S (ffi_call_O32): add labels for .eh_frame. Return
2917         64 bit integers correctly.
2918         (ffi_closure_O32): new function.
2919         Added DWARF-2 unwind info for both functions.
2920
2921 2004-08-10  Andrew Haley  <aph@redhat.com>
2922
2923         * src/x86/ffi64.c (ffi_prep_args ): 8-align all stack arguments.
2924
2925 2004-08-01  Robert Millan  <robertmh@gnu.org>
2926
2927         * configure.ac: Detect knetbsd-gnu and kfreebsd-gnu.
2928         * configure: Regenerate.
2929
2930 2004-07-30  Maciej W. Rozycki  <macro@linux-mips.org>
2931
2932         * acinclude.m4 (AC_FUNC_MMAP_BLACKLIST): Check for <sys/mman.h>
2933         and mmap() explicitly instead of relying on preset autoconf cache
2934         variables.
2935         * aclocal.m4: Regenerate.
2936         * configure: Regenerate.
2937
2938 2004-07-11  Ulrich Weigand  <uweigand@de.ibm.com>
2939
2940         * src/s390/ffi.c (ffi_prep_args): Fix C aliasing violation.
2941         (ffi_check_float_struct): Remove unused prototype.
2942
2943 2004-06-30  Geoffrey Keating  <geoffk@apple.com>
2944
2945         * src/powerpc/ffi_darwin.c (flush_icache): ';' is a comment
2946         character on Darwin, use '\n\t' instead.
2947
2948 2004-06-26  Matthias Klose  <doko@debian.org>
2949
2950         * libtool-version: Fix typo in revision/age.
2951
2952 2004-06-17  Matthias Klose  <doko@debian.org>
2953
2954         * libtool-version: New.
2955         * Makefile.am (libffi_la_LDFLAGS): Use -version-info for soname.
2956         * Makefile.in: Regenerate.
2957
2958 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2959
2960         * Makefile.am: Remove useless multilib rules.
2961         * Makefile.in: Regenerate.
2962         * aclocal.m4: Regenerate with automake 1.8.5.
2963         * configure.ac: Remove useless multilib configury.
2964         * configure: Regenerate.
2965
2966 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2967
2968         * .cvsignore: New file.
2969
2970 2004-06-10  Jakub Jelinek  <jakub@redhat.com>
2971
2972         * src/ia64/unix.S (ffi_call_unix): Insert group barrier break
2973         fp_done.
2974         (ffi_closure_UNIX): Fix f14/f15 adjustment if FLOAT_SZ is ever
2975         changed from 8.
2976
2977 2004-06-06  Sean McNeil  <sean@mcneil.com>
2978
2979         * configure.ac: Add x86_64-*-freebsd* support.
2980         * configure: Regenerate.
2981
2982 2004-04-26  Joe Buck <jbuck@welsh-buck.org>
2983
2984         Bug 15093
2985         * configure.ac: Test for existence of mmap and sys/mman.h before
2986         checking blacklist.  Fix suggested by Jim Wilson.
2987         * configure: Regenerate.
2988
2989 2004-04-26  Matt Austern  <austern@apple.com>
2990
2991         * src/powerpc/darwin.S: Go through a non-lazy pointer for initial
2992         FDE location.
2993         * src/powerpc/darwin_closure.S: Likewise.
2994
2995 2004-04-24  Andreas Tobler  <a.tobler@schweiz.ch>
2996
2997         * testsuite/libffi.call/cls_multi_schar.c (main): Fix initialization
2998         error. Reported by Thomas Heller <theller@python.net>.
2999         * testsuite/libffi.call/cls_multi_sshort.c (main): Likewise.
3000         * testsuite/libffi.call/cls_multi_ushort.c (main): Likewise.
3001
3002 2004-03-20  Matthias Klose  <doko@debian.org>
3003
3004         * src/pa/linux.S: Fix typo.
3005
3006 2004-03-19  Matthias Klose  <doko@debian.org>
3007
3008         * Makefile.am: Update.
3009         * Makefile.in: Regenerate.
3010         * src/pa/ffi.h.in: Remove.
3011         * src/pa/ffitarget.h: New file.
3012
3013 2004-02-10  Randolph Chung  <tausq@debian.org>
3014
3015         * Makefile.am: Add PA support.
3016         * Makefile.in: Regenerate.
3017         * include/Makefile.in: Regenerate.
3018         * configure.ac: Add PA target.
3019         * configure: Regenerate.
3020         * src/pa/ffi.c: New file.
3021         * src/pa/ffi.h.in: Add PA support.
3022         * src/pa/linux.S: New file.
3023         * prep_cif.c: Add PA support.
3024
3025 2004-03-16  Hosaka Yuji  <hos@tamanegi.org>
3026
3027         * src/types.c: Fix alignment size of X86_WIN32 case int64 and
3028         double.
3029         * src/x86/ffi.c (ffi_prep_args): Replace ecif->cif->rtype->type
3030         with ecif->cif->flags.
3031         (ffi_call, ffi_prep_incoming_args_SYSV): Replace cif->rtype->type
3032         with cif->flags.
3033         (ffi_prep_cif_machdep): Add X86_WIN32 struct case.
3034         (ffi_closure_SYSV): Add 1 or 2-bytes struct case for X86_WIN32.
3035         * src/x86/win32.S (retstruct1b, retstruct2b, sc_retstruct1b,
3036         sc_retstruct2b): Add for 1 or 2-bytes struct case.
3037
3038 2004-03-15 Kelley Cook <kcook@gcc.gnu.org>
3039
3040         * configure.in: Rename file to ...
3041         * configure.ac: ... this.
3042         * fficonfig.h.in: Regenerate.
3043         * Makefile.in: Regenerate.
3044         * include/Makefile.in: Regenerate.
3045         * testsuite/Makefile.in: Regenerate.
3046
3047 2004-03-12  Matt Austern  <austern@apple.com>
3048
3049         * src/powerpc/darwin.S: Fix EH information so it corresponds to
3050         changes in EH format resulting from addition of linkonce support.
3051         * src/powerpc/darwin_closure.S: Likewise.
3052
3053 2004-03-11  Andreas Tobler  <a.tobler@schweiz.ch>
3054             Paolo Bonzini  <bonzini@gnu.org>
3055
3056         * Makefile.am (AUTOMAKE_OPTIONS): Set them.
3057         Remove VPATH. Remove rules for object files. Remove multilib support.
3058         (AM_CCASFLAGS): Add.
3059         * configure.in (AC_CONFIG_HEADERS): Relace AM_CONFIG_HEADER.
3060         (AC_PREREQ): Bump version to 2.59.
3061         (AC_INIT): Fill with version info and bug address.
3062         (ORIGINAL_LD_FOR_MULTILIBS): Remove.
3063         (AM_ENABLE_MULTILIB): Use this instead of AC_ARG_ENABLE.
3064         De-precious CC so that the right flags are passed down to multilibs.
3065         (AC_MSG_ERROR): Replace obsolete macro AC_ERROR.
3066         (AC_CONFIG_FILES): Replace obsolete macro AC_LINK_FILES.
3067         (AC_OUTPUT): Reorganize the output with AC_CONFIG_COMMANDS.
3068         * configure: Rebuilt.
3069         * aclocal.m4: Likewise.
3070         * Makefile.in, include/Makefile.in, testsuite/Makefile.in: Likewise.
3071         * fficonfig.h.in: Likewise.
3072
3073 2004-03-11  Andreas Schwab  <schwab@suse.de>
3074
3075         * src/ia64/ffi.c (ffi_prep_incoming_args_UNIX): Get floating point
3076         arguments from fp registers only for the first 8 parameter slots.
3077         Don't convert a float parameter when passed in memory.
3078
3079 2004-03-09  Hans-Peter Nilsson  <hp@axis.com>
3080
3081         * configure: Regenerate for config/accross.m4 correction.
3082
3083 2004-02-25  Matt Kraai  <kraai@alumni.cmu.edu>
3084
3085         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Change
3086         ecif->cif->bytes to bytes.
3087         (ffi_prep_cif_machdep): Add braces around nested if statement.
3088
3089 2004-02-09  Alan Modra  <amodra@bigpond.net.au>
3090
3091         * src/types.c (pointer): POWERPC64 has 8 byte pointers.
3092
3093         * src/powerpc/ffi.c (ffi_prep_args64): Correct long double handling.
3094         (ffi_closure_helper_LINUX64): Fix typo.
3095         * testsuite/libffi.call/cls_align_longdouble.c: Pass -mlong-double-128
3096         for powerpc64-*-*.
3097         * testsuite/libffi.call/float.c: Likewise.
3098         * testsuite/libffi.call/float2.c: Likewise.
3099
3100 2004-02-08  Alan Modra  <amodra@bigpond.net.au>
3101
3102         * src/powerpc/ffi.c (ffi_prep_cif_machdep <FFI_LINUX64>): Correct
3103         long double function return and long double arg handling.
3104         (ffi_closure_helper_LINUX64): Formatting.  Delete unused "ng" var.
3105         Use "end_pfr" instead of "nf".  Correct long double handling.
3106         Localise "temp".
3107         * src/powerpc/linux64.S (ffi_call_LINUX64): Save f2 long double
3108         return value.
3109         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Allocate
3110         space for long double return value.  Adjust stack frame and offsets.
3111         Load f2 long double return.
3112
3113 2004-02-07  Alan Modra  <amodra@bigpond.net.au>
3114
3115         * src/types.c: Use 16 byte long double for POWERPC64.
3116
3117 2004-01-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
3118
3119         * src/sparc/ffi.c (ffi_prep_args_v9): Shift the parameter array
3120         when the structure return address is passed in %o0.
3121         (ffi_V9_return_struct): Rename into ffi_v9_layout_struct.
3122         (ffi_v9_layout_struct): Align the field following a nested structure
3123         on a word boundary.  Use memmove instead of memcpy.
3124         (ffi_call): Update call to ffi_V9_return_struct.
3125         (ffi_prep_closure): Define 'ctx' only for V8.
3126         (ffi_closure_sparc_inner): Clone into ffi_closure_sparc_inner_v8
3127         and ffi_closure_sparc_inner_v9.
3128         (ffi_closure_sparc_inner_v8): Return long doubles by reference.
3129         Always skip the structure return address.  For structures and long
3130         doubles, copy the argument directly.
3131         (ffi_closure_sparc_inner_v9): Skip the structure return address only
3132         if required.  Shift the maximum floating-point slot accordingly.  For
3133         big structures, copy the argument directly; otherwise, left-justify the
3134         argument and call ffi_v9_layout_struct to lay out the structure on
3135         the stack.
3136         * src/sparc/v8.S: Undef STACKFRAME before defining it.
3137         (ffi_closure_v8): Pass the structure return address.  Update call to
3138         ffi_closure_sparc_inner_v8.  Short-circuit FFI_TYPE_INT handling.
3139         Skip the 'unimp' insn when returning long doubles and structures.
3140         * src/sparc/v9.S: Undef STACKFRAME before defining it.
3141         (ffi_closure_v9): Increase the frame size by 2 words.  Short-circuit
3142         FFI_TYPE_INT handling.  Load structures both in integers and
3143         floating-point registers on return.
3144         * README: Update status of the SPARC port.
3145
3146 2004-01-24  Andreas Tobler  <a.tobler@schweiz.ch>
3147
3148         * testsuite/libffi.call/pyobjc-tc.c (main): Treat result value
3149         as of type ffi_arg.
3150         * testsuite/libffi.call/struct3.c (main): Fix CHECK.
3151
3152 2004-01-22  Ulrich Weigand  <uweigand@de.ibm.com>
3153
3154         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Treat result
3155         value as of type ffi_arg, not unsigned int.
3156
3157 2004-01-21  Michael Ritzert  <ritzert@t-online.de>
3158
3159         * ffi64.c (ffi_prep_args): Cast the RHS of an assignment instead
3160         of the LHS.
3161
3162 2004-01-12  Andreas Tobler  <a.tobler@schweiz.ch>
3163
3164         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_32 for
3165         Solaris.
3166
3167 2004-01-08  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3168
3169         * testsuite/libffi.call/ffitest.h (allocate_mmap): Cast MAP_FAILED
3170         to void *.
3171
3172 2003-12-10  Richard Henderson  <rth@redhat.com>
3173
3174         * testsuite/libffi.call/cls_align_pointer.c: Cast pointers to
3175         size_t instead of int.
3176
3177 2003-12-04  Hosaka Yuji  <hos@tamanegi.org>
3178
3179         * testsuite/libffi.call/many_win32.c: Include <float.h>.
3180         * testsuite/libffi.call/many_win32.c (main): Replace variable
3181         int i with unsigned long ul.
3182
3183         * testsuite/libffi.call/cls_align_uint64.c: New test case.
3184         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
3185         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
3186         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
3187         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
3188         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
3189         * testsuite/libffi.call/cls_align_float.c: Likewise.
3190         * testsuite/libffi.call/cls_align_double.c: Likewise.
3191         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
3192         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
3193
3194 2003-12-02  Hosaka Yuji  <hos@tamanegi.org>
3195
3196         PR other/13221
3197         * src/x86/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV):
3198         Align arguments to 32 bits.
3199
3200 2003-12-01  Andreas Tobler  <a.tobler@schweiz.ch>
3201
3202         PR other/13221
3203         * testsuite/libffi.call/cls_multi_sshort.c: New test case.
3204         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
3205         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
3206         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
3207         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
3208         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
3209
3210         * testsuite/libffi.special/unwindtest.cc: Cosmetics.
3211
3212 2003-11-26  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
3213
3214         * testsuite/libffi.call/ffitest.h: Include <fcntl.h>.
3215         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3216
3217 2003-11-22  Andreas Tobler  <a.tobler@schweiz.ch>
3218
3219         * Makefile.in: Rebuilt.
3220         * configure: Likewise.
3221         * testsuite/libffi.special/unwindtest.cc: Convert the mmap to
3222         the right type.
3223
3224 2003-11-21  Andreas Jaeger  <aj@suse.de>
3225             Andreas Tobler  <a.tobler@schweiz.ch>
3226
3227         * acinclude.m4: Add AC_FUNC_MMAP_BLACKLIST.
3228         * configure.in: Call AC_FUNC_MMAP_BLACKLIST.
3229         * Makefile.in: Rebuilt.
3230         * aclocal.m4: Likewise.
3231         * configure: Likewise.
3232         * fficonfig.h.in: Likewise.
3233         * testsuite/lib/libffi-dg.exp: Add include dir.
3234         * testsuite/libffi.call/ffitest.h: Add MMAP definitions.
3235         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3236         * testsuite/libffi.call/closure_fn0.c: Use MMAP functionality
3237         for ffi_closure if available.
3238         * testsuite/libffi.call/closure_fn1.c: Likewise.
3239         * testsuite/libffi.call/closure_fn2.c: Likewise.
3240         * testsuite/libffi.call/closure_fn3.c: Likewise.
3241         * testsuite/libffi.call/closure_fn4.c: Likewise.
3242         * testsuite/libffi.call/closure_fn5.c: Likewise.
3243         * testsuite/libffi.call/cls_12byte.c: Likewise.
3244         * testsuite/libffi.call/cls_16byte.c: Likewise.
3245         * testsuite/libffi.call/cls_18byte.c: Likewise.
3246         * testsuite/libffi.call/cls_19byte.c: Likewise.
3247         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3248         * testsuite/libffi.call/cls_20byte.c: Likewise.
3249         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3250         * testsuite/libffi.call/cls_24byte.c: Likewise.
3251         * testsuite/libffi.call/cls_2byte.c: Likewise.
3252         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3253         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3254         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3255         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3256         * testsuite/libffi.call/cls_4byte.c: Likewise.
3257         * testsuite/libffi.call/cls_5byte.c: Likewise.
3258         * testsuite/libffi.call/cls_64byte.c: Likewise.
3259         * testsuite/libffi.call/cls_6byte.c: Likewise.
3260         * testsuite/libffi.call/cls_7byte.c: Likewise.
3261         * testsuite/libffi.call/cls_8byte.c: Likewise.
3262         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3263         * testsuite/libffi.call/cls_9byte2.c: Likewise.
3264         * testsuite/libffi.call/cls_double.c: Likewise.
3265         * testsuite/libffi.call/cls_float.c: Likewise.
3266         * testsuite/libffi.call/cls_schar.c: Likewise.
3267         * testsuite/libffi.call/cls_sint.c: Likewise.
3268         * testsuite/libffi.call/cls_sshort.c: Likewise.
3269         * testsuite/libffi.call/cls_uchar.c: Likewise.
3270         * testsuite/libffi.call/cls_uint.c: Likewise.
3271         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3272         * testsuite/libffi.call/cls_ushort.c: Likewise.
3273         * testsuite/libffi.call/nested_struct.c: Likewise.
3274         * testsuite/libffi.call/nested_struct1.c: Likewise.
3275         * testsuite/libffi.call/nested_struct2.c: Likewise.
3276         * testsuite/libffi.call/nested_struct3.c: Likewise.
3277         * testsuite/libffi.call/problem1.c: Likewise.
3278         * testsuite/libffi.special/unwindtest.cc: Likewise.
3279
3280 2003-11-20  Andreas Tobler  <a.tobler@schweiz.ch>
3281
3282         * testsuite/lib/libffi-dg.exp: Make the -lgcc_s conditional.
3283
3284 2003-11-19  Andreas Tobler  <a.tobler@schweiz.ch>
3285
3286         * testsuite/lib/libffi-dg.exp: Add DYLD_LIBRARY_PATH for darwin.
3287         Add -lgcc_s to additional flags.
3288
3289 2003-11-12  Andreas Tobler  <a.tobler@schweiz.ch>
3290
3291         * configure.in, include/Makefile.am: PR libgcj/11147, install
3292         the ffitarget.h header file in a gcc versioned and target
3293         dependent place.
3294         * configure: Regenerated.
3295         * Makefile.in, include/Makefile.in: Likewise.
3296         * testsuite/Makefile.in: Likewise.
3297
3298 2003-11-09  Andreas Tobler  <a.tobler@schweiz.ch>
3299
3300         * testsuite/libffi.call/closure_fn0.c: Print result and check
3301         with dg-output to make debugging easier.
3302         * testsuite/libffi.call/closure_fn1.c: Likewise.
3303         * testsuite/libffi.call/closure_fn2.c: Likewise.
3304         * testsuite/libffi.call/closure_fn3.c: Likewise.
3305         * testsuite/libffi.call/closure_fn4.c: Likewise.
3306         * testsuite/libffi.call/closure_fn5.c: Likewise.
3307         * testsuite/libffi.call/cls_12byte.c: Likewise.
3308         * testsuite/libffi.call/cls_16byte.c: Likewise.
3309         * testsuite/libffi.call/cls_18byte.c: Likewise.
3310         * testsuite/libffi.call/cls_19byte.c: Likewise.
3311         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3312         * testsuite/libffi.call/cls_20byte.c: Likewise.
3313         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3314         * testsuite/libffi.call/cls_24byte.c: Likewise.
3315         * testsuite/libffi.call/cls_2byte.c: Likewise.
3316         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3317         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3318         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3319         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3320         * testsuite/libffi.call/cls_4byte.c: Likewise.
3321         * testsuite/libffi.call/cls_5byte.c: Likewise.
3322         * testsuite/libffi.call/cls_64byte.c: Likewise.
3323         * testsuite/libffi.call/cls_6byte.c: Likewise.
3324         * testsuite/libffi.call/cls_7byte.c: Likewise.
3325         * testsuite/libffi.call/cls_8byte.c: Likewise.
3326         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3327         * testsuite/libffi.call/cls_9byte2.c: Likewise.
3328         * testsuite/libffi.call/cls_double.c: Likewise.
3329         * testsuite/libffi.call/cls_float.c: Likewise.
3330         * testsuite/libffi.call/cls_schar.c: Likewise.
3331         * testsuite/libffi.call/cls_sint.c: Likewise.
3332         * testsuite/libffi.call/cls_sshort.c: Likewise.
3333         * testsuite/libffi.call/cls_uchar.c: Likewise.
3334         * testsuite/libffi.call/cls_uint.c: Likewise.
3335         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3336         * testsuite/libffi.call/cls_ushort.c: Likewise.
3337         * testsuite/libffi.call/problem1.c: Likewise.
3338
3339         * testsuite/libffi.special/unwindtest.cc: Make ffi_closure
3340         static.
3341
3342 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
3343
3344         * testsuite/libffi.call/cls_9byte2.c: New test case.
3345         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3346         * testsuite/libffi.call/cls_64byte.c: Likewise.
3347         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3348         * testsuite/libffi.call/cls_19byte.c: Likewise.
3349         * testsuite/libffi.call/cls_18byte.c: Likewise.
3350         * testsuite/libffi.call/closure_fn4.c: Likewise.
3351         * testsuite/libffi.call/closure_fn5.c: Likewise.
3352         * testsuite/libffi.call/cls_schar.c: Likewise.
3353         * testsuite/libffi.call/cls_sint.c: Likewise.
3354         * testsuite/libffi.call/cls_sshort.c: Likewise.
3355         * testsuite/libffi.call/nested_struct2.c: Likewise.
3356         * testsuite/libffi.call/nested_struct3.c: Likewise.
3357
3358 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
3359
3360         * testsuite/libffi.call/cls_double.c: Do a check on the result.
3361         * testsuite/libffi.call/cls_uchar.c: Likewise.
3362         * testsuite/libffi.call/cls_uint.c: Likewise.
3363         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3364         * testsuite/libffi.call/cls_ushort.c: Likewise.
3365         * testsuite/libffi.call/return_sc.c: Cleanup whitespaces.
3366
3367 2003-11-06  Andreas Tobler  <a.tobler@schweiz.ch>
3368
3369         * src/prep_cif.c (ffi_prep_cif): Move the validity check after
3370         the initialization.
3371
3372 2003-10-23  Andreas Tobler  <a.tobler@schweiz.ch>
3373
3374         * src/java_raw_api.c (ffi_java_ptrarray_to_raw): Replace
3375         FFI_ASSERT(FALSE) with FFI_ASSERT(0).
3376
3377 2003-10-22  David Daney  <ddaney@avtrex.com>
3378
3379         * src/mips/ffitarget.h: Replace undefined UINT32 and friends with
3380         __attribute__((__mode__(__SI__))) and friends.
3381
3382 2003-10-22  Andreas Schwab  <schwab@suse.de>
3383
3384         * src/ia64/ffi.c: Replace FALSE/TRUE with false/true.
3385
3386 2003-10-21  Andreas Tobler  <a.tobler@schweiz.ch>
3387
3388         * configure.in: AC_LINK_FILES(ffitarget.h).
3389         * configure: Regenerate.
3390         * Makefile.in: Likewise.
3391         * include/Makefile.in: Likewise.
3392         * testsuite/Makefile.in: Likewise.
3393         * fficonfig.h.in: Likewise.
3394
3395 2003-10-21  Paolo Bonzini  <bonzini@gnu.org>
3396             Richard Henderson  <rth@redhat.com>
3397
3398         Avoid that ffi.h includes fficonfig.h.
3399
3400         * Makefile.am (EXTRA_DIST): Include ffitarget.h files
3401         (TARGET_SRC_MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3402         (TARGET_SRC_MIPS_SGI): Removed.
3403         (MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3404         (MIPS_SGI): Removed.
3405         (CLEANFILES): Removed.
3406         (mostlyclean-am, clean-am, mostlyclean-sub, clean-sub): New
3407         targets.
3408         * acconfig.h: Removed.
3409         * configure.in: Compute sizeofs only for double and long double.
3410         Use them to define and subst HAVE_LONG_DOUBLE.  Include comments
3411         into AC_DEFINE instead of using acconfig.h.  Create
3412         include/ffitarget.h instead of include/fficonfig.h.  Rename
3413         MIPS_GCC to MIPS_IRIX, drop MIPS_SGI since we are in gcc's tree.
3414         AC_DEFINE EH_FRAME_FLAGS.
3415         * include/Makefile.am (DISTCLEANFILES): New automake macro.
3416         (hack_DATA): Add ffitarget.h.
3417         * include/ffi.h.in: Remove all system specific definitions.
3418         Declare raw API even if it is not installed, why bother?
3419         Use limits.h instead of SIZEOF_* to define ffi_type_*.  Do
3420         not define EH_FRAME_FLAGS, it is in fficonfig.h now.  Include
3421         ffitarget.h instead of fficonfig.h.  Remove ALIGN macro.
3422         (UINT_ARG, INT_ARG): Removed, use ffi_arg and ffi_sarg instead.
3423         * include/ffi_common.h (bool): Do not define.
3424         (ffi_assert): Accept failed assertion.
3425         (ffi_type_test): Return void and accept file/line.
3426         (FFI_ASSERT): Pass stringized failed assertion.
3427         (FFI_ASSERT_AT): New macro.
3428         (FFI_ASSERT_VALID_TYPE): New macro.
3429         (UINT8, SINT8, UINT16, SINT16, UINT32, SINT32,
3430         UINT64, SINT64): Define here with gcc's __attribute__ macro
3431         instead of in ffi.h
3432         (FLOAT32, ALIGN): Define here instead of in ffi.h
3433         * include/ffi-mips.h: Removed.  Its content moved to
3434         src/mips/ffitarget.h after separating assembly and C sections.
3435         * src/alpha/ffi.c, src/alpha/ffi.c, src/java_raw_api.c
3436         src/prep_cif.c, src/raw_api.c, src/ia64/ffi.c,
3437         src/mips/ffi.c, src/mips/n32.S, src/mips/o32.S,
3438         src/mips/ffitarget.h, src/sparc/ffi.c, src/x86/ffi64.c:
3439         SIZEOF_ARG -> FFI_SIZEOF_ARG.
3440         * src/ia64/ffi.c: Include stdbool.h (provided by GCC 2.95+).
3441         * src/debug.c (ffi_assert): Accept stringized failed assertion.
3442         (ffi_type_test): Rewritten.
3443         * src/prep-cif.c (initialize_aggregate, ffi_prep_cif): Call
3444         FFI_ASSERT_VALID_TYPE.
3445         * src/alpha/ffitarget.h, src/arm/ffitarget.h,
3446         src/ia64/ffitarget.h, src/m68k/ffitarget.h,
3447         src/mips/ffitarget.h, src/powerpc/ffitarget.h,
3448         src/s390/ffitarget.h, src/sh/ffitarget.h,
3449         src/sh64/ffitarget.h, src/sparc/ffitarget.h,
3450         src/x86/ffitarget.h: New files.
3451         * src/alpha/osf.S, src/arm/sysv.S, src/ia64/unix.S,
3452         src/m68k/sysv.S, src/mips/n32.S, src/mips/o32.S,
3453         src/powerpc/aix.S, src/powerpc/darwin.S,
3454         src/powerpc/ffi_darwin.c, src/powerpc/linux64.S,
3455         src/powerpc/linux64_closure.S, src/powerpc/ppc_closure.S,
3456         src/powerpc/sysv.S, src/s390/sysv.S, src/sh/sysv.S,
3457         src/sh64/sysv.S, src/sparc/v8.S, src/sparc/v9.S,
3458         src/x86/sysv.S, src/x86/unix64.S, src/x86/win32.S:
3459         include fficonfig.h
3460
3461 2003-10-20  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3462
3463         * src/mips/ffi.c: Use _ABIN32, _ABIO32 instead of external
3464         _MIPS_SIM_NABI32, _MIPS_SIM_ABI32.
3465
3466 2003-10-19  Andreas Tobler  <a.tobler@schweiz.ch>
3467
3468         * src/powerpc/ffi_darwin.c (ffi_prep_args): Declare bytes again.
3469         Used when FFI_DEBUG = 1.
3470
3471 2003-10-14  Alan Modra  <amodra@bigpond.net.au>
3472
3473         * src/types.c (double, longdouble): Default POWERPC64 to 8 byte size
3474         and align.
3475
3476 2003-10-06  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3477
3478         * include/ffi_mips.h: Define FFI_MIPS_N32 for N32/N64 ABIs,
3479         FFI_MIPS_O32 for O32 ABI.
3480
3481 2003-10-01  Andreas Tobler  <a.tobler@schweiz.ch>
3482
3483         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_64 for
3484         SPARC64. Cleanup whitespaces.
3485
3486 2003-09-19  Andreas Tobler  <a.tobler@schweiz.ch>
3487
3488         * testsuite/libffi.call/closure_fn0.c: Xfail mips, arm,
3489         strongarm, xscale. Cleanup whitespaces.
3490         * testsuite/libffi.call/closure_fn1.c: Likewise.
3491         * testsuite/libffi.call/closure_fn2.c: Likewise.
3492         * testsuite/libffi.call/closure_fn3.c: Likewise.
3493         * testsuite/libffi.call/cls_12byte.c: Likewise.
3494         * testsuite/libffi.call/cls_16byte.c: Likewise.
3495         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3496         * testsuite/libffi.call/cls_20byte.c: Likewise.
3497         * testsuite/libffi.call/cls_24byte.c: Likewise.
3498         * testsuite/libffi.call/cls_2byte.c: Likewise.
3499         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3500         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3501         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3502         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3503         * testsuite/libffi.call/cls_4byte.c: Likewise.
3504         * testsuite/libffi.call/cls_5byte.c: Likewise.
3505         * testsuite/libffi.call/cls_6byte.c: Likewise.
3506         * testsuite/libffi.call/cls_7byte.c: Likewise.
3507         * testsuite/libffi.call/cls_8byte.c: Likewise.
3508         * testsuite/libffi.call/cls_double.c: Likewise.
3509         * testsuite/libffi.call/cls_float.c: Likewise.
3510         * testsuite/libffi.call/cls_uchar.c: Likewise.
3511         * testsuite/libffi.call/cls_uint.c: Likewise.
3512         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3513         * testsuite/libffi.call/cls_ushort.c: Likewise.
3514         * testsuite/libffi.call/nested_struct.c: Likewise.
3515         * testsuite/libffi.call/nested_struct1.c: Likewise.
3516         * testsuite/libffi.call/problem1.c: Likewise.
3517         * testsuite/libffi.special/unwindtest.cc: Likewise.
3518         * testsuite/libffi.call/pyobjc-tc.c: Cleanup whitespaces.
3519
3520 2003-09-18  David Edelsohn  <edelsohn@gnu.org>
3521
3522         * src/powerpc/aix.S: Cleanup whitespaces.
3523         * src/powerpc/aix_closure.S: Likewise.
3524
3525 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3526
3527         * src/powerpc/darwin.S: Cleanup whitespaces, comment formatting.
3528         * src/powerpc/darwin_closure.S: Likewise.
3529         * src/powerpc/ffi_darwin.c: Likewise.
3530
3531 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3532             David Edelsohn  <edelsohn@gnu.org>
3533
3534         * src/types.c (double): Add AIX and Darwin to the right TYPEDEF.
3535         * src/powerpc/aix_closure.S: Remove the pointer to the outgoing
3536         parameter stack.
3537         * src/powerpc/darwin_closure.S: Likewise.
3538         * src/powerpc/ffi_darwin.c (ffi_prep_args): Handle structures
3539         according to the Darwin/AIX ABI.
3540         (ffi_prep_cif_machdep): Likewise.
3541         (ffi_closure_helper_DARWIN): Likewise.
3542         Remove the outgoing parameter stack logic. Simplify the evaluation
3543         of the different CASE types.
3544         (ffi_prep_clousure): Avoid the casts on lvalues. Change the branch
3545         statement in the trampoline code.
3546
3547 2003-09-18  Kaz Kojima  <kkojima@gcc.gnu.org>
3548
3549         * src/sh/ffi.c (ffi_prep_args): Take account into the alignement
3550         for the register size.
3551         (ffi_closure_helper_SYSV): Handle the structure return value
3552         address correctly.
3553         (ffi_closure_helper_SYSV): Return the appropriate type when
3554         the registers are used for the structure return value.
3555         * src/sh/sysv.S (ffi_closure_SYSV): Fix the stack layout for
3556         the 64-bit return value.  Update copyright years.
3557
3558 2003-09-17  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3559
3560         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Search in
3561         srcdir for ffi_mips.h.
3562
3563 2003-09-12  Alan Modra  <amodra@bigpond.net.au>
3564
3565         * src/prep_cif.c (initialize_aggregate): Include tail padding in
3566         structure size.
3567         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Correct
3568         placement of float result.
3569         * testsuite/libffi.special/unwindtest.cc (closure_test_fn1): Correct
3570         cast of "resp" for big-endian 64 bit machines.
3571
3572 2003-09-11  Alan Modra  <amodra@bigpond.net.au>
3573
3574         * src/types.c (double, longdouble): Merge identical SH and ARM
3575         typedefs, and add POWERPC64.
3576         * src/powerpc/ffi.c (ffi_prep_args64): Correct next_arg calc for
3577         struct split over gpr and rest.
3578         (ffi_prep_cif_machdep): Correct intarg_count for structures.
3579         * src/powerpc/linux64.S (ffi_call_LINUX64): Fix gpr offsets.
3580
3581 2003-09-09  Andreas Tobler  <a.tobler@schweiz.ch>
3582
3583         * src/powerpc/ffi.c (ffi_closure_helper_SYSV) Handle struct
3584         passing correctly.
3585
3586 2003-09-09  Alan Modra  <amodra@bigpond.net.au>
3587
3588         * configure: Regenerate.
3589
3590 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3591
3592         * Makefile.am: Remove build rules for ffitest.
3593         * Makefile.in: Rebuilt.
3594
3595 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3596
3597         * src/java_raw_api.c: Include <stdlib.h> to fix compiler warning
3598         about implicit declaration of abort().
3599
3600 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3601
3602         * Makefile.am: Add dejagnu test framework. Fixes PR other/11411.
3603         * Makefile.in: Rebuilt.
3604         * configure.in: Add dejagnu test framework.
3605         * configure: Rebuilt.
3606
3607         * testsuite/Makefile.am: New file.
3608         * testsuite/Makefile.in: Built
3609         * testsuite/lib/libffi-dg.exp: New file.
3610         * testsuite/config/default.exp: Likewise.
3611         * testsuite/libffi.call/call.exp: Likewise.
3612         * testsuite/libffi.call/ffitest.h: Likewise.
3613         * testsuite/libffi.call/closure_fn0.c: Likewise.
3614         * testsuite/libffi.call/closure_fn1.c: Likewise.
3615         * testsuite/libffi.call/closure_fn2.c: Likewise.
3616         * testsuite/libffi.call/closure_fn3.c: Likewise.
3617         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3618         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3619         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3620         * testsuite/libffi.call/cls_2byte.c: Likewise.
3621         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3622         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3623         * testsuite/libffi.call/cls_4byte.c: Likewise.
3624         * testsuite/libffi.call/cls_5byte.c: Likewise.
3625         * testsuite/libffi.call/cls_6byte.c: Likewise.
3626         * testsuite/libffi.call/cls_7byte.c: Likewise.
3627         * testsuite/libffi.call/cls_8byte.c: Likewise.
3628         * testsuite/libffi.call/cls_12byte.c: Likewise.
3629         * testsuite/libffi.call/cls_16byte.c: Likewise.
3630         * testsuite/libffi.call/cls_20byte.c: Likewise.
3631         * testsuite/libffi.call/cls_24byte.c: Likewise.
3632         * testsuite/libffi.call/cls_double.c: Likewise.
3633         * testsuite/libffi.call/cls_float.c: Likewise.
3634         * testsuite/libffi.call/cls_uchar.c: Likewise.
3635         * testsuite/libffi.call/cls_uint.c: Likewise.
3636         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3637         * testsuite/libffi.call/cls_ushort.c: Likewise.
3638         * testsuite/libffi.call/float.c: Likewise.
3639         * testsuite/libffi.call/float1.c: Likewise.
3640         * testsuite/libffi.call/float2.c: Likewise.
3641         * testsuite/libffi.call/many.c: Likewise.
3642         * testsuite/libffi.call/many_win32.c: Likewise.
3643         * testsuite/libffi.call/nested_struct.c: Likewise.
3644         * testsuite/libffi.call/nested_struct1.c: Likewise.
3645         * testsuite/libffi.call/pyobjc-tc.c: Likewise.
3646         * testsuite/libffi.call/problem1.c: Likewise.
3647         * testsuite/libffi.call/promotion.c: Likewise.
3648         * testsuite/libffi.call/return_ll.c: Likewise.
3649         * testsuite/libffi.call/return_sc.c: Likewise.
3650         * testsuite/libffi.call/return_uc.c: Likewise.
3651         * testsuite/libffi.call/strlen.c: Likewise.
3652         * testsuite/libffi.call/strlen_win32.c: Likewise.
3653         * testsuite/libffi.call/struct1.c: Likewise.
3654         * testsuite/libffi.call/struct2.c: Likewise.
3655         * testsuite/libffi.call/struct3.c: Likewise.
3656         * testsuite/libffi.call/struct4.c: Likewise.
3657         * testsuite/libffi.call/struct5.c: Likewise.
3658         * testsuite/libffi.call/struct6.c: Likewise.
3659         * testsuite/libffi.call/struct7.c: Likewise.
3660         * testsuite/libffi.call/struct8.c: Likewise.
3661         * testsuite/libffi.call/struct9.c: Likewise.
3662         * testsuite/libffi.special/special.exp: New file.
3663         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3664         * testsuite/libffi.special/unwindtest.cc: Likewise.
3665
3666
3667 2003-08-13  Kaz Kojima  <kkojima@gcc.gnu.org>
3668
3669         * src/sh/ffi.c (OFS_INT16): Set 0 for little endian case.  Update
3670         copyright years.
3671
3672 2003-08-02  Alan Modra  <amodra@bigpond.net.au>
3673
3674         * src/powerpc/ffi.c (ffi_prep_args64): Modify for changed gcc
3675         structure passing.
3676         (ffi_closure_helper_LINUX64): Likewise.
3677         * src/powerpc/linux64.S: Remove code writing to parm save area.
3678         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Use return
3679         address in lr from ffi_closure_helper_LINUX64 call to calculate
3680         table address.  Optimize function tail.
3681
3682 2003-07-28  Andreas Tobler  <a.tobler@schweiz.ch>
3683
3684         * src/sparc/ffi.c: Handle all floating point registers.
3685         * src/sparc/v9.S: Likewise. Fixes second part of PR target/11410.
3686
3687 2003-07-11  Gerald Pfeifer  <pfeifer@dbai.tuwien.ac.at>
3688
3689         * README: Note that libffi is not part of GCC.  Update the project
3690         URL and status.
3691
3692 2003-06-19  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3693
3694         * src/powerpc/ppc_closure.S: Include ffi.h.
3695
3696 2003-06-13  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3697
3698         * src/x86/sysv.S: Avoid gas-only .uleb128/.sleb128 directives.
3699         Use C style comments.
3700
3701 2003-06-13  Kaz Kojima  <kkojima@rr.iij4u.or.jp>
3702
3703         * Makefile.am: Add SHmedia support.  Fix a typo of SH support.
3704         * Makefile.in: Regenerate.
3705         * configure.in (sh64-*-linux*, sh5*-*-linux*): Add target.
3706         * configure: Regenerate.
3707         * include/ffi.h.in: Add SHmedia support.
3708         * src/sh64/ffi.c: New file.
3709         * src/sh64/sysv.S: New file.
3710
3711 2003-05-16  Jakub Jelinek  <jakub@redhat.com>
3712
3713         * configure.in (HAVE_RO_EH_FRAME): Check whether .eh_frame section
3714         should be read-only.
3715         * configure: Rebuilt.
3716         * fficonfig.h.in: Rebuilt.
3717         * include/ffi.h.in (EH_FRAME_FLAGS): Define.
3718         * src/alpha/osf.S: Use EH_FRAME_FLAGS.
3719         * src/powerpc/linux64.S: Likewise.
3720         * src/powerpc/linux64_closure.S: Likewise.  Include ffi.h.
3721         * src/powerpc/sysv.S: Use EH_FRAME_FLAGS.  Use pcrel encoding
3722         if -fpic/-fPIC/-mrelocatable.
3723         * src/powerpc/powerpc_closure.S: Likewise.
3724         * src/sparc/v8.S: If HAVE_RO_EH_FRAME is defined, don't include
3725         #write in .eh_frame flags.
3726         * src/sparc/v9.S: Likewise.
3727         * src/x86/unix64.S: Use EH_FRAME_FLAGS.
3728         * src/x86/sysv.S: Likewise.  Use pcrel encoding if -fpic/-fPIC.
3729         * src/s390/sysv.S: Use EH_FRAME_FLAGS.  Include ffi.h.
3730
3731 2003-05-07  Jeff Sturm  <jsturm@one-point.com>
3732
3733         Fixes PR bootstrap/10656
3734         * configure.in (HAVE_AS_REGISTER_PSEUDO_OP): Test assembler
3735         support for .register pseudo-op.
3736         * src/sparc/v8.S: Use it.
3737         * fficonfig.h.in: Rebuilt.
3738         * configure: Rebuilt.
3739
3740 2003-04-18  Jakub Jelinek  <jakub@redhat.com>
3741
3742         * include/ffi.h.in (POWERPC64): Define if 64-bit.
3743         (enum ffi_abi): Add FFI_LINUX64 on POWERPC.
3744         Make it the default on POWERPC64.
3745         (FFI_TRAMPOLINE_SIZE): Define to 24 on POWERPC64.
3746         * configure.in: Change powerpc-*-linux* into powerpc*-*-linux*.
3747         * configure: Rebuilt.
3748         * src/powerpc/ffi.c (hidden): Define.
3749         (ffi_prep_args_SYSV): Renamed from
3750         ffi_prep_args.  Cast pointers to unsigned long to shut up warnings.
3751         (NUM_GPR_ARG_REGISTERS64, NUM_FPR_ARG_REGISTERS64,
3752         ASM_NEEDS_REGISTERS64): New.
3753         (ffi_prep_args64): New function.
3754         (ffi_prep_cif_machdep): Handle FFI_LINUX64 ABI.
3755         (ffi_call): Likewise.
3756         (ffi_prep_closure): Likewise.
3757         (flush_icache): Surround by #ifndef POWERPC64.
3758         (ffi_dblfl): New union type.
3759         (ffi_closure_helper_SYSV): Use it to avoid aliasing problems.
3760         (ffi_closure_helper_LINUX64): New function.
3761         * src/powerpc/ppc_closure.S: Surround whole file by #ifndef
3762         __powerpc64__.
3763         * src/powerpc/sysv.S: Likewise.
3764         (ffi_call_SYSV): Rename ffi_prep_args to ffi_prep_args_SYSV.
3765         * src/powerpc/linux64.S: New file.
3766         * src/powerpc/linux64_closure.S: New file.
3767         * Makefile.am (EXTRA_DIST): Add src/powerpc/linux64.S and
3768         src/powerpc/linux64_closure.S.
3769         (TARGET_SRC_POWERPC): Likewise.
3770
3771         * src/ffitest.c (closure_test_fn, closure_test_fn1, closure_test_fn2,
3772         closure_test_fn3): Fix result printing on big-endian 64-bit
3773         machines.
3774         (main): Print tst2_arg instead of uninitialized tst2_result.
3775
3776         * src/ffitest.c (main): Hide what closure pointer really points to
3777         from the compiler.
3778
3779 2003-04-16  Richard Earnshaw  <rearnsha@arm.com>
3780
3781         * configure.in (arm-*-netbsdelf*): Add configuration.
3782         (configure): Regenerated.
3783
3784 2003-04-04  Loren J. Rittle  <ljrittle@acm.org>
3785
3786         * include/Makefile.in: Regenerate.
3787
3788 2003-03-21  Zdenek Dvorak  <rakdver@atrey.karlin.mff.cuni.cz>
3789
3790         * libffi/include/ffi.h.in: Define X86 instead of X86_64 in 32
3791         bit mode.
3792         * libffi/src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV):
3793         Receive closure pointer through parameter, read args using
3794         __builtin_dwarf_cfa.
3795         (FFI_INIT_TRAMPOLINE): Send closure reference through eax.
3796
3797 2003-03-12  Andreas Schwab  <schwab@suse.de>
3798
3799         * configure.in: Avoid trailing /. in toolexeclibdir.
3800         * configure: Rebuilt.
3801
3802 2003-03-03  Andreas Tobler <a.tobler@schweiz.ch>
3803
3804         * src/powerpc/darwin_closure.S: Recode to fit dynamic libraries.
3805
3806 2003-02-06  Andreas Tobler <a.tobler@schweiz.ch>
3807
3808         * libffi/src/powerpc/darwin_closure.S:
3809         Fix alignement bug, allocate 8 bytes for the result.
3810         * libffi/src/powerpc/aix_closure.S:
3811         Likewise.
3812         * libffi/src/powerpc/ffi_darwin.c:
3813         Update stackframe description for aix/darwin_closure.S.
3814
3815 2003-02-06  Jakub Jelinek  <jakub@redhat.com>
3816
3817         * src/s390/ffi.c (ffi_closure_helper_SYSV): Add hidden visibility
3818         attribute.
3819
3820 2003-01-31  Christian Cornelssen  <ccorn@cs.tu-berlin.de>,
3821             Andreas Schwab  <schwab@suse.de>
3822
3823         * configure.in: Adjust command to source config-ml.in to account
3824         for changes to the libffi_basedir definition.
3825         (libffi_basedir): Remove ${srcdir} from value and include trailing
3826         slash if nonempty.
3827
3828         * configure: Regenerate.
3829
3830 2003-01-29  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3831
3832         * src/powerpc/ppc_closure.S: Recode to fit shared libs.
3833
3834 2003-01-28  Andrew Haley  <aph@redhat.com>
3835
3836         * include/ffi.h.in: Enable FFI_CLOSURES for x86_64.
3837         * src/x86/ffi64.c (ffi_prep_closure): New.
3838         (ffi_closure_UNIX64_inner): New.
3839         * src/x86/unix64.S (ffi_closure_UNIX64): New.
3840
3841 2003-01-27  Alexandre Oliva  <aoliva@redhat.com>
3842
3843         * configure.in (toolexecdir, toolexeclibdir): Set and AC_SUBST.
3844         Remove USE_LIBDIR conditional.
3845         * Makefile.am (toolexecdir, toolexeclibdir): Don't override.
3846         * Makefile.in, configure: Rebuilt.
3847
3848 2003-01027  David Edelsohn  <edelsohn@gnu.org>
3849
3850         * Makefile.am (TARGET_SRC_POWERPC_AIX): Fix typo.
3851         * Makefile.in: Regenerate.
3852
3853 2003-01-22  Andrew Haley  <aph@redhat.com>
3854
3855         * src/powerpc/darwin.S (_ffi_call_AIX): Add Augmentation size to
3856         unwind info.
3857
3858 2003-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
3859
3860         * src/powerpc/darwin.S: Add unwind info.
3861         * src/powerpc/darwin_closure.S: Likewise.
3862
3863 2003-01-14  Andrew Haley  <aph@redhat.com>
3864
3865         * src/x86/ffi64.c (ffi_prep_args): Check for void retval.
3866         (ffi_prep_cif_machdep): Likewise.
3867         * src/x86/unix64.S: Add unwind info.
3868
3869 2003-01-14  Andreas Jaeger  <aj@suse.de>
3870
3871         * src/ffitest.c (main): Only use ffi_closures if those are
3872         supported.
3873
3874 2003-01-13 Andreas Tobler <a.tobler@schweiz.ch>
3875
3876         * libffi/src/ffitest.c
3877          add closure testcases
3878
3879 2003-01-13 Kevin B. Hendricks <khendricks@ivey.uwo.ca>
3880
3881         * libffi/src/powerpc/ffi.c
3882          fix alignment bug for float (4 byte aligned iso 8 byte)
3883
3884 2003-01-09  Geoffrey Keating  <geoffk@apple.com>
3885
3886         * src/powerpc/ffi_darwin.c: Remove RCS version string.
3887         * src/powerpc/darwin.S: Remove RCS version string.
3888
3889 2003-01-03  Jeff Sturm  <jsturm@one-point.com>
3890
3891         * include/ffi.h.in: Add closure defines for SPARC, SPARC64.
3892         * src/ffitest.c (main): Use static storage for closure.
3893         * src/sparc/ffi.c (ffi_prep_closure, ffi_closure_sparc_inner): New.
3894         * src/sparc/v8.S (ffi_closure_v8): New.
3895         * src/sparc/v9.S (ffi_closure_v9): New.
3896
3897 2002-11-10  Ranjit Mathew <rmathew@hotmail.com>
3898
3899         * include/ffi.h.in: Added FFI_STDCALL ffi_type
3900           enumeration for X86_WIN32.
3901         * src/x86/win32.S: Added ffi_call_STDCALL function
3902           definition.
3903         * src/x86/ffi.c (ffi_call/ffi_raw_call): Added
3904           switch cases for recognising FFI_STDCALL and
3905           calling ffi_call_STDCALL if target is X86_WIN32.
3906         * src/ffitest.c (my_stdcall_strlen/stdcall_many):
3907           stdcall versions of the "my_strlen" and "many"
3908           test functions (for X86_WIN32).
3909           Added test cases to test stdcall invocation using
3910           these functions.
3911
3912 2002-12-02  Kaz Kojima  <kkojima@gcc.gnu.org>
3913
3914         * src/sh/sysv.S: Add DWARF2 unwind info.
3915
3916 2002-11-27  Ulrich Weigand  <uweigand@de.ibm.com>
3917
3918         * src/s390/sysv.S (.eh_frame section): Make section read-only.
3919
3920 2002-11-26  Jim Wilson  <wilson@redhat.com>
3921
3922         * src/types.c (FFI_TYPE_POINTER): Has size 8 on IA64.
3923
3924 2002-11-23  H.J. Lu <hjl@gnu.org>
3925
3926         * acinclude.m4: Add dummy AM_PROG_LIBTOOL.
3927         Include ../config/accross.m4.
3928         * aclocal.m4; Rebuild.
3929         * configure: Likewise.
3930
3931 2002-11-15  Ulrich Weigand  <uweigand@de.ibm.com>
3932
3933         * src/s390/sysv.S (.eh_frame section): Adapt to pcrel FDE encoding.
3934
3935 2002-11-11  DJ Delorie  <dj@redhat.com>
3936
3937         * configure.in: Look for common files in the right place.
3938
3939 2002-10-08  Ulrich Weigand  <uweigand@de.ibm.com>
3940
3941         * src/java_raw_api.c (ffi_java_raw_to_ptrarray): Interpret
3942         raw data as _Jv_word values, not ffi_raw.
3943         (ffi_java_ptrarray_to_raw): Likewise.
3944         (ffi_java_rvalue_to_raw): New function.
3945         (ffi_java_raw_call): Call it.
3946         (ffi_java_raw_to_rvalue): New function.
3947         (ffi_java_translate_args): Call it.
3948         * src/ffitest.c (closure_test_fn): Interpret return value
3949         as ffi_arg, not int.
3950         * src/s390/ffi.c (ffi_prep_cif_machdep): Add missing
3951         FFI_TYPE_POINTER case.
3952         (ffi_closure_helper_SYSV): Likewise.  Also, assume return
3953         values extended to word size.
3954
3955 2002-10-02  Andreas Jaeger  <aj@suse.de>
3956
3957         * src/x86/ffi64.c (ffi_prep_cif_machdep): Remove debug output.
3958
3959 2002-10-01  Bo Thorsen  <bo@smetana.suse.de>
3960
3961         * include/ffi.h.in: Fix i386 win32 compilation.
3962
3963 2002-09-30  Ulrich Weigand  <uweigand@de.ibm.com>
3964
3965         * configure.in: Add s390x-*-linux-* target.
3966         * configure: Regenerate.
3967         * include/ffi.h.in: Define S390X for s390x targets.
3968         (FFI_CLOSURES): Define for s390/s390x.
3969         (FFI_TRAMPOLINE_SIZE): Likewise.
3970         (FFI_NATIVE_RAW_API): Likewise.
3971         * src/prep_cif.c (ffi_prep_cif): Do not compute stack space for s390.
3972         * src/types.c (FFI_TYPE_POINTER): Use 8-byte pointers on s390x.
3973         * src/s390/ffi.c: Major rework of existing code.  Add support for
3974         s390x targets.  Add closure support.
3975         * src/s390/sysv.S: Likewise.
3976
3977 2002-09-29  Richard Earnshaw  <rearnsha@arm.com>
3978
3979         * src/arm/sysv.S: Fix typo.
3980
3981 2002-09-28  Richard Earnshaw  <rearnsha@arm.com>
3982
3983         * src/arm/sysv.S: If we don't have machine/asm.h and the pre-processor
3984         has defined __USER_LABEL_PREFIX__, then use it in CNAME.
3985         (ffi_call_SYSV): Handle soft-float.
3986
3987 2002-09-27  Bo Thorsen  <bo@suse.de>
3988
3989         * include/ffi.h.in: Fix multilib x86-64 support.
3990
3991 2002-09-22  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
3992
3993         * Makefile.am (all-multi): Fix multilib parallel build.
3994
3995 2002-07-19  Kaz Kojima  <kkojima@gcc.gnu.org>
3996
3997         * configure.in (sh[34]*-*-linux*): Add brackets.
3998         * configure: Regenerate.
3999
4000 2002-07-18  Kaz Kojima  <kkojima@gcc.gnu.org>
4001
4002         * Makefile.am: Add SH support.
4003         * Makefile.in: Regenerate.
4004         * configure.in (sh-*-linux*, sh[34]*-*-linux*): Add target.
4005         * configure: Regenerate.
4006         * include/ffi.h.in: Add SH support.
4007         * src/sh/ffi.c: New file.
4008         * src/sh/sysv.S: New file.
4009         * src/types.c: Add SH support.
4010
4011 2002-07-16  Bo Thorsen  <bo@suse.de>
4012
4013         * src/x86/ffi64.c: New file that adds x86-64 support.
4014         * src/x86/unix64.S: New file that handles argument setup for
4015         x86-64.
4016         * src/x86/sysv.S: Don't use this on x86-64.
4017         * src/x86/ffi.c: Don't use this on x86-64.
4018         Remove unused vars.
4019         * src/prep_cif.c (ffi_prep_cif): Don't do stack size calculation
4020         for x86-64.
4021         * src/ffitest.c (struct6): New test that tests a special case in
4022         the x86-64 ABI.
4023         (struct7): Likewise.
4024         (struct8): Likewise.
4025         (struct9): Likewise.
4026         (closure_test_fn): Silence warning about this when it's not used.
4027         (main): Add the new tests.
4028         (main): Fix a couple of wrong casts and silence some compiler warnings.
4029         * include/ffi.h.in: Add x86-64 ABI definition.
4030         * fficonfig.h.in: Regenerate.
4031         * Makefile.am: Add x86-64 support.
4032         * configure.in: Likewise.
4033         * Makefile.in: Regenerate.
4034         * configure: Likewise.
4035
4036 2002-06-24  Bo Thorsen  <bo@suse.de>
4037
4038         * src/types.c: Merge settings for similar architectures.
4039         Add x86-64 sizes and alignments.
4040
4041 2002-06-23  Bo Thorsen  <bo@suse.de>
4042
4043         * src/arm/ffi.c (ffi_prep_args): Remove unused vars.
4044         * src/sparc/ffi.c (ffi_prep_args_v8): Likewise.
4045         * src/mips/ffi.c (ffi_prep_args): Likewise.
4046         * src/m68k/ffi.c (ffi_prep_args): Likewise.
4047
4048 2002-07-18  H.J. Lu  (hjl@gnu.org)
4049
4050         * Makefile.am (TARGET_SRC_MIPS_LINUX): New.
4051         (libffi_la_SOURCES): Support MIPS_LINUX.
4052         (libffi_convenience_la_SOURCES): Likewise.
4053         * Makefile.in: Regenerated.
4054
4055         * configure.in (mips64*-*): Skip.
4056         (mips*-*-linux*): New.
4057         * configure: Regenerated.
4058
4059         * src/mips/ffi.c: Include <sgidefs.h>.
4060
4061 2002-06-06  Ulrich Weigand  <uweigand@de.ibm.com>
4062
4063         * src/s390/sysv.S: Save/restore %r6.  Add DWARF-2 unwind info.
4064
4065 2002-05-27  Roger Sayle  <roger@eyesopen.com>
4066
4067         * src/x86/ffi.c (ffi_prep_args): Remove reference to avn.
4068
4069 2002-05-27  Bo Thorsen  <bo@suse.de>
4070
4071         * src/x86/ffi.c (ffi_prep_args): Remove unused variable and
4072         fix formatting.
4073
4074 2002-05-13  Andreas Tobler  <a.tobler@schweiz.ch>
4075
4076         * src/powerpc/ffi_darwin.c (ffi_prep_closure): Declare fd at