3e182f92b344694869f4a256655153d8e45d2672
[libffi.git] / .pc / more-openbsd / ChangeLog
1 2012-01-28  Kai Tietz  <ktietz@redhat.com>
2
3         * src/libffi/src/x86/ffi.c (ffi_call_win32): Add new
4         argument to prototype for specify calling-convention.
5         (ffi_call): Add support for stdcall/thiscall convention.
6         (ffi_prep_args): Likewise.
7         (ffi_raw_call): Likewise.
8         * src/x86/ffitarget.h (ffi_abi): Add FFI_THISCALL and
9         FFI_FASTCALL.
10         * src/x86/win32.S (_ffi_call_win32): Add support for
11         fastcall/thiscall calling-convention calls.
12         * testsuite/libffi.call/fastthis1_win32.c: New test.
13         * testsuite/libffi.call/fastthis2_win32.c: New test.
14         * testsuite/libffi.call/fastthis3_win32.c: New test.
15         * testsuite/libffi.call/strlen2_win32.c: New test.
16         * testsuite/libffi.call/many2_win32.c: New test.
17         * testsuite/libffi.call/struct1_win32.c: New test.
18         * testsuite/libffi.call/struct2_win32.c: New test.
19
20 2012-01-23  Andreas Schwab  <schwab@linux-m68k.org>
21
22         * src/m68k/sysv.S (ffi_call_SYSV): Properly test for plain
23         mc68000.  Test for __HAVE_68881__ in addition to __MC68881__.
24
25 2012-01-19  Jakub Jelinek  <jakub@redhat.com>
26
27         PR rtl-optimization/48496
28         * src/ia64/ffi.c (ffi_call): Fix up aliasing violations.
29
30 2012-01-09  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
31
32         * configure.ac (i?86-*-*): Set TARGET to X86_64.
33         * configure: Regenerate.
34
35 2011-12-07  Andrew Pinski  <apinski@cavium.com>
36
37         PR libffi/50051
38         * src/mips/n32.S: Add ".set mips4".
39
40 2011-11-21  Andreas Tobler  <andreast@fgznet.ch>
41
42         * configure: Regenerate.
43
44 2011-11-10  Richard Henderson  <rth@redhat.com>
45
46         * configure.ac (GCC_AS_CFI_PSEUDO_OP): Use it instead of inline check.
47         * configure, aclocal.m4: Rebuild.
48
49 2011-09-04  Iain Sandoe  <iains@gcc.gnu.org>
50
51         PR libffi/49594
52         * src/powerpc/darwin_closure.S (stubs): Make the stub binding
53         helper reference track the architecture pointer size.
54
55 2011-08-25  Andrew Haley  <aph@redhat.com>
56
57         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Remove hard-coded assembly
58         instructions.
59         * src/arm/sysv.S (ffi_arm_trampoline): Put them here instead.
60
61 2011-07-11  Andrew Haley  <aph@redhat.com>
62
63         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Clear icache.
64
65 2011-06-29  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
66
67         * testsuite/libffi.call/cls_double_va.c: Move PR number to comment.
68         * testsuite/libffi.call/cls_longdouble_va.c: Likewise.
69
70 2011-06-29  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
71
72         PR libffi/46660
73         * testsuite/libffi.call/cls_double_va.c: xfail dg-output on
74         mips-sgi-irix6*.
75         * testsuite/libffi.call/cls_longdouble_va.c: Likewise.
76
77 2011-06-14  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
78
79         * testsuite/libffi.call/huge_struct.c (test_large_fn): Use PRIu8,
80         PRId8 instead of %hhu, %hhd.
81         * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRId8,
82         PRIu8): Define.
83         [__sgi__] (PRId8, PRIu8): Define.
84
85 2011-04-29  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
86
87         * src/alpha/osf.S (UA_SI, FDE_ENCODING, FDE_ENCODE, FDE_ARANGE):
88         Define.
89         Use them to handle ELF vs. ECOFF differences.
90         [__osf__] (_GLOBAL__F_ffi_call_osf): Define.
91
92 2011-03-30  Timothy Wall  <twall@users.sf.net>
93
94         * src/powerpc/darwin.S: Fix unknown FDE encoding.
95         * src/powerpc/darwin_closure.S: ditto.
96
97 2011-02-25  Anthony Green  <green@moxielogic.com>
98
99         * src/powerpc/ffi.c (ffi_prep_closure_loc): Allow for more
100         32-bit ABIs.
101
102 2011-02-15  Anthony Green  <green@moxielogic.com>
103
104         * m4/ax_cc_maxopt.m4: Don't -malign-double or use -ffast-math.
105         * configure: Rebuilt.
106
107 2011-02-13  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
108
109         * configure: Regenerate.
110
111 2011-02-13  Anthony Green  <green@moxielogic.com>
112
113         * include/ffi_common.h (UNLIKELY, LIKELY): Define.
114         * src/x86/ffi64.c (UNLIKELY, LIKELY): Remove definition.
115         * src/prep_cif.c (UNLIKELY, LIKELY): Remove definition.
116
117         * src/prep_cif.c (initialize_aggregate): Convert assertion into
118         FFI_BAD_TYPEDEF return.  Initialize arg size and alignment to 0.
119
120         * src/pa/ffi.c (ffi_prep_closure_loc): Don't ASSERT ABI test,
121         just return FFI_BAD_ABI when things are wrong.
122         * src/arm/ffi.c (ffi_prep_closure_loc): Ditto.
123         * src/powerpc/ffi.c (ffi_prep_closure_loc): Ditto.
124         * src/mips/ffi.c (ffi_prep_closure_loc): Ditto.
125         * src/ia64/ffi.c (ffi_prep_closure_loc): Ditto.
126         * src/avr32/ffi.c (ffi_prep_closure_loc): Ditto.
127
128 2011-02-11  Anthony Green  <green@moxielogic.com>
129
130         * src/sparc/ffi.c (ffi_prep_closure_loc): Don't ASSERT ABI test,
131         just return FFI_BAD_ABI when things are wrong.
132
133 2011-02-09  Stuart Shelton  <srcshelton@gmail.com>
134
135         http://bugs.gentoo.org/show_bug.cgi?id=286911
136         * src/mips/ffitarget.h: Clean up error messages.
137         * src/java_raw_api.c (ffi_java_translate_args): Cast raw arg to
138         ffi_raw*.
139         * include/ffi.h.in: Add pragma for SGI compiler.
140
141 2011-02-09  Anthony Green  <green@moxielogic.com>
142
143         * configure.ac: Add powerpc64-*-darwin* support.
144
145 2011-02-09  Anthony Green <green@moxielogic.com>
146
147         * README: Mention Interix.
148
149 2011-02-09  Jonathan Callen  <abcd@gentoo.org>
150
151         * configure.ac: Add Interix to win32/cygwin/mingw case.
152         * configure: Ditto.
153         * src/closures.c: Treat Interix like Cygwin, instead of as a
154         generic win32.
155
156 2011-02-09  Anthony Green <green@moxielogic.com>
157
158         * testsuite/libffi.call/err_bad_typedef.c: Remove xfail.
159         * testsuite/libffi.call/err_bad_abi.c: Remove xfail.
160         * src/x86/ffi64.c (UNLIKELY, LIKELY): Define.
161         (ffi_prep_closure_loc): Check for bad ABI.
162         * src/prep_cif.c (UNLIKELY, LIKELY): Define.
163         (initialize_aggregate): Check for bad types.
164
165 2011-02-09  Landon Fuller <landonf@plausible.coop>
166
167         * Makefile.am (EXTRA_DIST): Add build-ios.sh, src/arm/gentramp.sh,
168         src/arm/trampoline.S.
169         (nodist_libffi_la_SOURCES): Add src/arc/trampoline.S.
170         * configure.ac (FFI_EXEC_TRAMPOLINE_TABLE): Define.
171         * src/arm/ffi.c (ffi_trampoline_table)
172         (ffi_closure_trampoline_table_page, ffi_trampoline_table_entry)
173         (FFI_TRAMPOLINE_CODELOC_CONFIG, FFI_TRAMPOLINE_CONFIG_PAGE_OFFSET)
174         (FFI_TRAMPOLINE_COUNT, ffi_trampoline_lock, ffi_trampoline_tables)
175         (ffi_trampoline_table_alloc, ffi_closure_alloc, ffi_closure_free):
176         Define for FFI_EXEC_TRAMPOLINE_TABLE case (iOS).
177         (ffi_prep_closure_loc): Handl FFI_EXEC_TRAMPOLINE_TABLE case
178         separately.
179         * src/arm/sysv.S: Handle Apple iOS host.
180         * src/closures.c: Handle FFI_EXEC_TRAMPOLINE_TABLE case.
181         * build-ios.sh: New file.
182         * fficonfig.h.in, configure, Makefile.in: Rebuilt.
183         * README: Mention ARM iOS.
184
185 2011-02-08  Oren Held  <orenhe@il.ibm.com>
186
187         * src/dlmalloc.c (_STRUCT_MALLINFO): Define in order to avoid
188         redefinition of mallinfo on HP-UX.
189
190 2011-02-08  Ginn Chen  <ginn.chen@oracle.com>
191
192         * src/sparc/ffi.c (ffi_call): Make compatible with Solaris Studio
193         aggregate return ABI.  Flush cache.
194         (ffi_prep_closure_loc): Flush cache.
195
196 2011-02-11  Anthony Green  <green@moxielogic.com>
197
198         From Tom Honermann <tom.honermann@oracle.com>:
199         * src/powerpc/aix.S (ffi_call_AIX): Support for xlc toolchain on
200         AIX.  Declare .ffi_prep_args.  Insert nops after branch
201         instructions so that the AIX linker can insert TOC reload
202         instructions.
203         * src/powerpc/aix_closure.S: Declare .ffi_closure_helper_DARWIN.
204
205 2011-02-08  Ed  <ed@kdtc.net>
206
207         * src/powerpc/asm.h: Fix grammar nit in comment.
208
209 2011-02-08  Uli Link  <ul.mcamafia@linkitup.de>
210
211         * include/ffi.h.in (FFI_64_BIT_MAX): Define and use.
212
213 2011-02-09  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
214
215         PR libffi/46661
216         * testsuite/libffi.call/cls_pointer.c (main): Cast void * to
217         uintptr_t first.
218         * testsuite/libffi.call/cls_pointer_stack.c (main): Likewise.
219
220 2011-02-08  Rafael Avila de Espindola  <respindola@mozilla.com>
221
222         * configure.ac: Fix x86 test for pc related relocs.
223         * configure: Rebuilt.
224
225 2011-02-07  Joel Sherrill <joel.sherrill@oarcorp.com>
226
227         * libffi/src/m68k/ffi.c: Add RTEMS support for cache flushing.
228         Handle case when CPU variant does not have long double support.
229         * libffi/src/m68k/sysv.S: Add support for mc68000, Coldfire,
230         and cores with soft floating point.
231
232 2011-02-07  Joel Sherrill <joel.sherrill@oarcorp.com>
233
234         * configure.ac: Add mips*-*-rtems* support.
235         * configure: Regenerate.
236         * src/mips/ffitarget.h: Ensure needed constants are available
237         for targets which do not have sgidefs.h.
238
239 2011-01-26  Dave Korn  <dave.korn.cygwin@gmail.com>
240
241         PR target/40125
242         * configure.ac (AM_LTLDFLAGS): Add -bindir option for windows DLLs.
243         * configure: Regenerate.
244
245 2010-12-18  Iain Sandoe  <iains@gcc.gnu.org>
246
247         PR libffi/29152
248         PR libffi/42378
249         * src/powerpc/darwin_closure.S: Provide Darwin64 implementation,
250         update comments.
251         * src/powerpc/ffitarget.h (POWERPC_DARWIN64): New,
252         (FFI_TRAMPOLINE_SIZE): Update for Darwin64.
253         * src/powerpc/darwin.S: Provide Darwin64 implementation,
254         update comments.
255         * src/powerpc/ffi_darwin.c: Likewise.
256
257 2010-12-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
258
259         * configure.ac (libffi_cv_as_ascii_pseudo_op): Use double
260         backslashes.
261         (libffi_cv_as_string_pseudo_op): Likewise.
262         * configure: Regenerate.
263
264 2010-12-03  Chung-Lin Tang  <cltang@codesourcery.com>
265
266         * src/arm/sysv.S (ffi_closure_SYSV): Add UNWIND to .pad directive.
267         (ffi_closure_VFP): Same.
268         (ffi_call_VFP): Move down to before ffi_closure_VFP. Add '.fpu vfp'
269         directive.
270
271 2010-12-01  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
272
273         * testsuite/libffi.call/ffitest.h [__sgi] (PRId64, PRIu64): Define.
274         (PRIuPTR): Define.
275
276 2010-11-29  Richard Henderson  <rth@redhat.com>
277             Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
278
279         * src/x86/sysv.S (FDE_ENCODING, FDE_ENCODE): Define.
280         (.eh_frame): Use FDE_ENCODING.
281         (.LASFDE1, .LASFDE2, LASFDE3): Simplify with FDE_ENCODE.
282
283 2010-11-22  Jacek Caban <jacek@codeweavers.com>
284
285         * configure.ac: Check for symbol underscores on mingw-w64.
286         * configure: Rebuilt.
287         * src/x86/win64.S: Correctly access extern symbols in respect to
288         underscores.
289
290 2010-11-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
291
292         * testsuite/lib/libffi-dg.exp: Rename ...
293         * testsuite/lib/libffi.exp: ... to this.
294         * libffi/testsuite/libffi.call/call.exp: Don't load libffi-dg.exp.
295         * libffi/testsuite/libffi.special/special.exp: Likewise.
296
297 2010-10-28  Chung-Lin Tang  <cltang@codesourcery.com>
298
299         * src/arm/ffi.c (ffi_prep_args): Add VFP register argument handling
300         code, new parameter, and return value. Update comments.
301         (ffi_prep_cif_machdep): Add case for VFP struct return values. Add
302         call to layout_vfp_args().
303         (ffi_call_SYSV): Update declaration.
304         (ffi_call_VFP): New declaration.
305         (ffi_call): Add VFP struct return conditions. Call ffi_call_VFP()
306         when ABI is FFI_VFP.
307         (ffi_closure_VFP): New declaration.
308         (ffi_closure_SYSV_inner): Add new vfp_args parameter, update call to
309         ffi_prep_incoming_args_SYSV().
310         (ffi_prep_incoming_args_SYSV): Update parameters. Add VFP argument
311         case handling.
312         (ffi_prep_closure_loc): Pass ffi_closure_VFP to trampoline
313         construction under VFP hard-float.
314         (rec_vfp_type_p): New function.
315         (vfp_type_p): Same.
316         (place_vfp_arg): Same.
317         (layout_vfp_args): Same.
318         * src/arm/ffitarget.h (ffi_abi): Add FFI_VFP. Define FFI_DEFAULT_ABI
319         based on __ARM_PCS_VFP.
320         (FFI_EXTRA_CIF_FIELDS): Define for adding VFP hard-float specific
321         fields.
322         (FFI_TYPE_STRUCT_VFP_FLOAT): Define internally used type code.
323         (FFI_TYPE_STRUCT_VFP_DOUBLE): Same.
324         * src/arm/sysv.S (ffi_call_SYSV): Change call of ffi_prep_args() to
325         direct call. Move function pointer load upwards.
326         (ffi_call_VFP): New function.
327         (ffi_closure_VFP): Same.
328
329         * testsuite/lib/libffi-dg.exp (check-flags): New function.
330         (dg-skip-if): New function.
331         * testsuite/libffi.call/cls_double_va.c: Skip if target is arm*-*-*
332         and compiler options include -mfloat-abi=hard.
333         * testsuite/libffi.call/cls_longdouble_va.c: Same.
334
335 2010-10-01  Jakub Jelinek  <jakub@redhat.com>
336
337         PR libffi/45677
338         * src/x86/ffi64.c (ffi_prep_cif_machdep): Ensure cif->bytes is
339         a multiple of 8.
340         * testsuite/libffi.call/many2.c: New test.
341
342 2010-08-20  Mark Wielaard  <mjw@redhat.com>
343
344         * src/closures.c (open_temp_exec_file_mnt): Check if getmntent_r
345         returns NULL.
346
347 2010-08-09  Andreas Tobler  <andreast@fgznet.ch>
348
349         * configure.ac: Add target powerpc64-*-freebsd*.
350         * configure: Regenerate.
351         * testsuite/libffi.call/cls_align_longdouble_split.c: Pass
352         -mlong-double-128 only to linux targets.
353         * testsuite/libffi.call/cls_align_longdouble_split2.c: Likewise.
354         * testsuite/libffi.call/cls_longdouble.c: Likewise.
355         * testsuite/libffi.call/huge_struct.c: Likewise.
356
357 2010-08-05  Dan Witte  <dwitte@mozilla.com>
358
359         * Makefile.am: Pass FFI_DEBUG define to msvcc.sh for linking to the
360         debug CRT when --enable-debug is given.
361         * configure.ac: Define it.
362         * msvcc.sh: Translate -g and -DFFI_DEBUG appropriately.
363
364 2010-08-04  Dan Witte  <dwitte@mozilla.com>
365
366         * src/x86/ffitarget.h: Add X86_ANY define for all x86/x86_64
367         platforms.
368         * src/x86/ffi.c: Remove redundant ifdef checks.
369         * src/prep_cif.c: Push stack space computation into src/x86/ffi.c
370         for X86_ANY so return value space doesn't get added twice.
371
372 2010-08-03  Neil Rashbrooke <neil@parkwaycc.co.uk>
373
374         * msvcc.sh: Don't pass -safeseh to ml64 because behavior is buggy.
375
376 2010-07-22  Dan Witte  <dwitte@mozilla.com>
377
378         * src/*/ffitarget.h: Make FFI_LAST_ABI one past the last valid ABI.
379         * src/prep_cif.c: Fix ABI assertion.
380         * src/cris/ffi.c: Ditto.
381
382 2010-07-10  Evan Phoenix  <evan@fallingsnow.net>
383
384         * src/closures.c (selinux_enabled_check): Fix strncmp usage bug.
385
386 2010-07-07  Dan HorĂ¡k <dan@danny.cz>
387
388         * include/ffi.h.in: Protect #define with #ifndef.
389         * src/powerpc/ffitarget.h: Ditto.
390         * src/s390/ffitarget.h: Ditto.
391         * src/sparc/ffitarget.h: Ditto.
392
393 2010-07-07   Neil Roberts <neil@linux.intel.com>
394
395         * src/x86/sysv.S (ffi_call_SYSV): Align the stack pointer to
396         16-bytes.
397
398 2010-07-02  Jakub Jelinek  <jakub@redhat.com>
399
400         * Makefile.am (AM_MAKEFLAGS): Pass also mandir to submakes.
401         * Makefile.in: Regenerated.
402
403 2010-05-19  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
404
405         * configure.ac (libffi_cv_as_x86_pcrel): Check for illegal in as
406         output, too.
407         (libffi_cv_as_ascii_pseudo_op): Check for .ascii.
408         (libffi_cv_as_string_pseudo_op): Check for .string.
409         * configure: Regenerate.
410         * fficonfig.h.in: Regenerate.
411         * src/x86/sysv.S (.eh_frame): Use .ascii, .string or error.
412
413 2010-05-11  Dan Witte  <dwitte@mozilla.com>
414
415         * doc/libffi.tex: Document previous change.
416
417 2010-05-11  Makoto Kato <m_kato@ga2.so-net.ne.jp>
418
419         * src/x86/ffi.c (ffi_call): Don't copy structs passed by value.
420
421 2010-05-05  Michael Kohler <michaelkohler@live.com>
422
423         * src/dlmalloc.c (dlfree): Fix spelling.
424         * src/ia64/ffi.c (ffi_prep_cif_machdep): Ditto.
425         * configure.ac: Ditto.
426         * configure: Rebuilt.
427
428 2010-04-13  Dan Witte  <dwitte@mozilla.com>
429
430         * msvcc.sh: Build with -W3 instead of -Wall.
431         * src/powerpc/ffi_darwin.c: Remove build warnings.
432         * src/x86/ffi.c: Ditto.
433         * src/x86/ffitarget.h: Ditto.
434
435 2010-04-12  Dan Witte  <dwitte@mozilla.com>
436             Walter Meinl <wuno@lsvw.de>
437
438         * configure.ac: Add OS/2 support.
439         * configure: Rebuilt.
440         * src/closures.c: Ditto.
441         * src/dlmalloc.c: Ditto.
442         * src/x86/win32.S: Ditto.
443
444 2010-04-07  Jakub Jelinek  <jakub@redhat.com>
445
446         * testsuite/libffi.call/err_bad_abi.c: Remove unused args variable.
447
448 2010-04-02  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
449
450         * Makefile.in: Regenerate.
451         * aclocal.m4: Regenerate.
452         * include/Makefile.in: Regenerate.
453         * man/Makefile.in: Regenerate.
454         * testsuite/Makefile.in: Regenerate.
455
456 2010-03-30  Dan Witte  <dwitte@mozilla.com>
457
458         * msvcc.sh: Disable build warnings.
459         * README (tested): Clarify windows build procedure.
460
461 2010-03-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
462
463         * configure.ac (libffi_cv_as_x86_64_unwind_section_type): New test.
464         * configure: Regenerate.
465         * fficonfig.h.in: Regenerate.
466         * libffi/src/x86/unix64.S (.eh_frame)
467         [HAVE_AS_X86_64_UNWIND_SECTION_TYPE]: Use @unwind section type.
468
469 2010-03-14  Matthias Klose  <doko@ubuntu.com>
470
471         * src/x86/ffi64.c: Fix typo in comment.
472         * src/x86/ffi.c: Use /* ... */ comment style.
473
474 2010-02-24  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
475
476         * doc/libffi.texi (The Closure API): Fix typo.
477         * doc/libffi.info: Remove.
478
479 2010-02-15  Matthias Klose  <doko@ubuntu.com>
480
481         * src/arm/sysv.S (__ARM_ARCH__): Define for processor
482         __ARM_ARCH_7EM__.
483
484 2010-01-15  Anthony Green  <green@redhat.com>
485
486         * README: Add notes on building with Microsoft Visual C++.
487
488 2010-01-15  Daniel Witte  <dwitte@mozilla.com>
489
490         * msvcc.sh: New file.
491
492         * src/x86/win32.S: Port assembly routines to MSVC and #ifdef.
493         * src/x86/ffi.c: Tweak function declaration and remove excess
494         parens.
495         * include/ffi.h.in: Add __declspec(align(8)) to typedef struct
496         ffi_closure.
497
498         * src/x86/ffi.c: Merge ffi_call_SYSV and ffi_call_STDCALL into new
499         function ffi_call_win32 on X86_WIN32.
500         * src/x86/win32.S (ffi_call_SYSV): Rename to ffi_call_win32.
501         (ffi_call_STDCALL): Remove.
502
503         * src/prep_cif.c (ffi_prep_cif): Move stack space allocation code
504         to ffi_prep_cif_machdep for x86.
505         * src/x86/ffi.c (ffi_prep_cif_machdep): To here.
506
507 2010-01-15  Oliver Kiddle  <okiddle@yahoo.co.uk>
508
509         * src/x86/ffitarget.h (ffi_abi): Check for __i386 and __amd64 for
510         Sun Studio compiler compatibility.
511
512 2010-01-12  Conrad Irwin <conrad.irwin@gmail.com>
513
514         * doc/libffi.texi: Add closure example.
515
516 2010-01-07  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
517
518         PR libffi/40701
519         * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRIdLL,
520         PRIuLL, PRId64, PRIu64, PRIuPTR): Define.
521         * testsuite/libffi.call/cls_align_sint64.c: Add -Wno-format on
522         alpha*-dec-osf*.
523         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
524         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
525         * testsuite/libffi.call/return_ll1.c: Likewise.
526         * testsuite/libffi.call/stret_medium2.c: Likewise.
527         * testsuite/libffi.special/ffitestcxx.h (allocate_mmap): Cast
528         MAP_FAILED to char *.
529
530 2010-01-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
531
532         * src/mips/n32.S: Use .abicalls and .eh_frame with __GNUC__.
533
534 2009-12-31  Anthony Green  <green@redhat.com>
535
536         * README: Update for libffi 3.0.9.
537
538 2009-12-27  Matthias Klose  <doko@ubuntu.com>
539
540         * configure.ac (HAVE_LONG_DOUBLE): Define for mips when
541         appropriate.
542         * configure: Rebuilt.
543
544 2009-12-26  Anthony Green  <green@redhat.com>
545
546         * testsuite/libffi.call/cls_longdouble_va.c: Mark as xfail for
547         avr32*-*-*.
548         * testsuite/libffi.call/cls_double_va.c: Ditto.
549
550 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
551
552         * testsuite/libffi.call/ffitest.h: Conditionally include stdint.h
553         and inttypes.h.
554         * testsuite/libffi.special/unwindtest.cc: Ditto.
555
556 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
557
558         * configure.ac: Add amd64-*-openbsd*.
559         * configure: Rebuilt.
560         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Link
561         openbsd programs with -lpthread.
562
563 2009-12-26  Anthony Green  <green@redhat.com>
564
565         * testsuite/libffi.call/cls_double_va.c,
566         testsuite/libffi.call/cls_longdouble.c,
567         testsuite/libffi.call/cls_longdouble_va.c,
568         testsuite/libffi.call/cls_pointer.c,
569         testsuite/libffi.call/cls_pointer_stack.c: Remove xfail for
570         mips*-*-* and arm*-*-*.
571         * testsuite/libffi.call/cls_align_longdouble_split.c,
572         testsuite/libffi.call/cls_align_longdouble_split2.c,
573         testsuite/libffi.call/stret_medium2.c,
574         testsuite/libffi.call/stret_medium.c,
575         testsuite/libffi.call/stret_large.c,
576         testsuite/libffi.call/stret_large2.c: Remove xfail for arm*-*-*.
577
578 2009-12-31  Kay Tietz  <ktietz70@googlemail.com>
579
580         * testsuite/libffi.call/ffitest.h,
581         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRuLL): Fix
582         definitions.
583
584 2009-12-31  Carlo Bramini  <carlo.bramix@libero.it>
585
586         * configure.ac (AM_LTLDFLAGS): Define for windows hosts.
587         * Makefile.am (libffi_la_LDFLAGS): Add AM_LTLDFLAGS.
588         * configure: Rebuilt.
589         * Makefile.in: Rebuilt.
590
591 2009-12-31  Anthony Green  <green@redhat.com>
592             Blake Chaffin.
593
594         * testsuite/libffi.call/huge_struct.c: New test case from Blake
595         Chaffin @ Apple.
596
597 2009-12-28  David Edelsohn  <edelsohn@gnu.org>
598
599         * src/powerpc/ffi_darwin.c (ffi_prep_args): Copy abi and nargs to
600         local variables.
601         (aix_adjust_aggregate_sizes): New function.
602         (ffi_prep_cif_machdep): Call it.
603
604 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
605
606         * configure.ac: Define FFI_MMAP_EXEC_WRIT for the given targets.
607         * configure: Regenerate.
608         * fficonfig.h.in: Likewise.
609         * src/closures.c: Remove the FFI_MMAP_EXEC_WRIT definition for
610         Solaris/x86.
611
612 2009-12-26  Andreas Schwab  <schwab@linux-m68k.org>
613
614         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Advance intarg_count
615         when a float arguments is passed in memory.
616         (ffi_closure_helper_SYSV): Mark general registers as used up when
617         a 64bit or soft-float long double argument is passed in memory.
618
619 2009-12-25  Matthias Klose  <doko@ubuntu.com>
620
621         * man/ffi_call.3: Fix #include in examples.
622         * doc/libffi.texi: Add dircategory.
623
624 2009-12-25  Frank Everdij <f.p.x.everdij@tudelft.nl>
625
626         * include/ffi.h.in: Placed '__GNUC__' ifdef around
627         '__attribute__((aligned(8)))' in ffi_closure, fixes compile for
628         IRIX MIPSPro c99.
629         * include/ffi_common.h: Added '__sgi' define to non
630         '__attribute__((__mode__()))' integer typedefs.
631         * src/mips/ffi.c (ffi_call, ffi_closure_mips_inner_O32,
632         ffi_closure_mips_inner_N32): Added 'defined(_MIPSEB)' to BE check.
633         (ffi_closure_mips_inner_O32, ffi_closure_mips_inner_N32): Added
634         FFI_LONGDOUBLE support and alignment(N32 only).
635         * src/mips/ffitarget.h: Corrected '#include <sgidefs.h>' for IRIX and
636         fixed non '__attribute__((__mode__()))' integer typedefs.
637         * src/mips/n32.S: Put '#ifdef linux' around '.abicalls' and '.eh_frame'
638         since they are Linux/GNU Assembler specific.
639
640 2009-12-25  Bradley Smith  <brad@brad-smith.co.uk>
641
642         * configure.ac, Makefile.am, src/avr32/ffi.c,
643         src/avr32/ffitarget.h,
644         src/avr32/sysv.S: Add AVR32 port.
645         * configure, Makefile.in: Rebuilt.
646
647 2009-12-21  Andreas Tobler  <a.tobler@schweiz.org>
648
649         * configure.ac: Make i?86 build on FreeBSD and OpenBSD.
650         * configure: Regenerate.
651
652 2009-12-15  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
653
654         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on PA HP-UX.
655
656 2009-12-13  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
657
658         * src/pa/ffi.c (ffi_closure_inner_pa32): Handle FFI_TYPE_LONGDOUBLE
659         type on HP-UX.
660
661 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
662
663         * src/sparc/ffi.c (ffi_closure_sparc_inner_v9): Properly align 'long
664         double' arguments.
665
666 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
667
668         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on Solaris < 10.
669
670 2009-12-10  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
671
672         PR libffi/40700
673         * src/closures.c [X86_64 && __sun__ && __svr4__]
674         (FFI_MMAP_EXEC_WRIT): Define.
675
676 2009-12-08  David Daney  <ddaney@caviumnetworks.com>
677
678         * testsuite/libffi.call/stret_medium.c: Remove xfail for mips*-*-*
679         * testsuite/libffi.call/cls_align_longdouble_split2.c: Same.
680         * testsuite/libffi.call/stret_large.c: Same.
681         * testsuite/libffi.call/cls_align_longdouble_split.c: Same.
682         * testsuite/libffi.call/stret_large2.c: Same.
683         * testsuite/libffi.call/stret_medium2.c: Same.
684
685 2009-12-07  David Edelsohn  <edelsohn@gnu.org>
686
687         * src/powerpc/aix_closure.S (libffi_closure_ASM): Fix tablejump
688         typo.
689
690 2009-12-05  David Edelsohn  <edelsohn@gnu.org>
691
692         * src/powerpc/aix.S: Update AIX32 code to be consistent with AIX64
693         code.
694         * src/powerpc/aix_closure.S: Same.
695
696 2009-12-05  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
697
698         * Makefile.in: Regenerate.
699         * configure: Regenerate.
700         * include/Makefile.in: Regenerate.
701         * man/Makefile.in: Regenerate.
702         * testsuite/Makefile.in: Regenerate.
703
704 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
705
706         * src/powerpc/aix_closure.S: Reorganize 64-bit code to match
707         linux64_closure.S.
708
709 2009-12-04  Uros Bizjak  <ubizjak@gmail.com>
710
711         PR libffi/41908
712         * src/x86/ffi64.c (classify_argument): Update from
713         gcc/config/i386/i386.c.
714         (ffi_closure_unix64_inner): Do not use the address of two consecutive
715         SSE registers directly.
716         * testsuite/libffi.call/cls_dbls_struct.c (main): Remove xfail
717         for x86_64 linux targets.
718
719 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
720
721         * src/powerpc/ffi_darwin.c (ffi_closure_helper_DARWIN): Increment
722         pfr for long double split between fpr13 and stack.
723
724 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
725
726         * src/powerpc/ffi_darwin.c (ffi_prep_args): Increment next_arg and
727         fparg_count twice for long double.
728
729 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
730
731         PR libffi/42243
732         * src/powerpc/ffi_darwin.c (ffi_prep_args): Remove extra parentheses.
733
734 2009-12-03  Uros Bizjak  <ubizjak@gmail.com>
735
736         * testsuite/libffi.call/cls_longdouble_va.c (main): Fix format string.
737         Remove xfails for x86 linux targets.
738
739 2009-12-02  David Edelsohn  <edelsohn@gnu.org>
740
741         * src/powerpc/ffi_darwin.c (ffi_prep_args): Fix typo in INT64
742         case.
743
744 2009-12-01  David Edelsohn  <edelsohn@gnu.org>
745
746         * src/powerpc/aix.S (ffi_call_AIX): Convert to more standard
747         register usage.  Call ffi_prep_args directly.  Add long double
748         return value support.
749         * src/powerpc/ffi_darwin.c (ffi_prep_args): Double arg increment
750         applies to FFI_TYPE_DOUBLE.  Correct fpr_base increment typo.
751         Separate FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases.
752         (ffi_prep_cif_machdep): Only 16 byte stack alignment in 64 bit
753         mode.
754         (ffi_closure_helper_DARWIN): Remove nf and ng counters.  Move temp
755         into case.
756         * src/powerpc/aix_closure.S: Maintain 16 byte stack alignment.
757         Allocate result area between params and FPRs.
758
759 2009-11-30  David Edelsohn  <edelsohn@gnu.org>
760
761         PR target/35484
762         * src/powerpc/ffitarget.h (POWERPC64): Define for PPC64 Linux and
763         AIX64.
764         * src/powerpc/aix.S: Implement AIX64 version.
765         * src/powerpc/aix_closure.S: Implement AIX64 version.
766         (ffi_closure_ASM): Use extsb, lha and displament addresses.
767         * src/powerpc/ffi_darwin.c (ffi_prep_args): Implement AIX64
768         support.
769         (ffi_prep_cif_machdep): Same.
770         (ffi_call): Same.
771         (ffi_closure_helper_DARWIN): Same.
772
773 2009-11-02  Andreas Tobler  <a.tobler@schweiz.org>
774
775         PR libffi/41908
776         * testsuite/libffi.call/testclosure.c: New test.
777
778 2009-09-28  Kai Tietz  <kai.tietz@onevision.com>
779
780         * src/x86/win64.S (_ffi_call_win64 stack): Remove for gnu
781         assembly version use of ___chkstk.
782
783 2009-09-23  Matthias Klose  <doko@ubuntu.com>
784
785         PR libffi/40242, PR libffi/41443
786         * src/arm/sysv.S (__ARM_ARCH__): Define for processors
787         __ARM_ARCH_6T2__, __ARM_ARCH_6M__, __ARM_ARCH_7__,
788         __ARM_ARCH_7A__, __ARM_ARCH_7R__, __ARM_ARCH_7M__.
789         Change the conditionals to __SOFTFP__ || __ARM_EABI__
790         for -mfloat-abi=softfp to work.
791
792 2009-09-17  Loren J. Rittle  <ljrittle@acm.org>
793
794         PR testsuite/32843 (strikes again)
795         * src/x86/ffi.c (ffi_prep_cif_machdep): Add X86_FREEBSD to
796         enable proper extension on char and short.
797
798 2009-09-15  David Daney  <ddaney@caviumnetworks.com>
799
800         * src/java_raw_api.c (ffi_java_raw_to_rvalue): Remove special
801         handling for FFI_TYPE_POINTER.
802         * src/mips/ffitarget.h (FFI_TYPE_STRUCT_D_SOFT,
803         FFI_TYPE_STRUCT_F_SOFT, FFI_TYPE_STRUCT_DD_SOFT,
804         FFI_TYPE_STRUCT_FF_SOFT, FFI_TYPE_STRUCT_FD_SOFT,
805         FFI_TYPE_STRUCT_DF_SOFT, FFI_TYPE_STRUCT_SOFT): New defines.
806         (FFI_N32_SOFT_FLOAT, FFI_N64_SOFT_FLOAT): New ffi_abi enumerations.
807         (enum ffi_abi): Set FFI_DEFAULT_ABI for soft-float.
808         * src/mips/n32.S (ffi_call_N32): Add handling for soft-float
809         structure and pointer returns.
810         (ffi_closure_N32): Add handling for pointer returns.
811         * src/mips/ffi.c (ffi_prep_args, calc_n32_struct_flags,
812         calc_n32_return_struct_flags): Handle soft-float.
813         (ffi_prep_cif_machdep):  Handle soft-float, fix pointer handling.
814         (ffi_call_N32): Declare proper argument types.
815         (ffi_call, copy_struct_N32, ffi_closure_mips_inner_N32): Handle
816         soft-float.
817
818 2009-08-24  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
819
820         * configure.ac (AC_PREREQ): Bump to 2.64.
821
822 2009-08-22  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
823
824         * Makefile.am (install-html, install-pdf): Remove.
825         * Makefile.in: Regenerate.
826
827         * Makefile.in: Regenerate.
828         * aclocal.m4: Regenerate.
829         * configure: Regenerate.
830         * fficonfig.h.in: Regenerate.
831         * include/Makefile.in: Regenerate.
832         * man/Makefile.in: Regenerate.
833         * testsuite/Makefile.in: Regenerate.
834
835 2009-07-30  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
836
837         * configure.ac (_AC_ARG_VAR_PRECIOUS): Use m4_rename_force.
838
839 2009-07-24  Dave Korn  <dave.korn.cygwin@gmail.com>
840
841         PR libffi/40807
842         * src/x86/ffi.c (ffi_prep_cif_machdep): Also use sign/zero-extending
843         return types for X86_WIN32.
844         * src/x86/win32.S (_ffi_call_SYSV): Handle omitted return types.
845         (_ffi_call_STDCALL, _ffi_closure_SYSV, _ffi_closure_raw_SYSV,
846         _ffi_closure_STDCALL): Likewise.
847
848         * src/closures.c (is_selinux_enabled): Define to const 0 for Cygwin.
849         (dlmmap, dlmunmap): Also use these functions on Cygwin.
850
851 2009-07-11  Richard Sandiford  <rdsandiford@googlemail.com>
852
853         PR testsuite/40699
854         PR testsuite/40707
855         PR testsuite/40709
856         * testsuite/lib/libffi-dg.exp: Revert 2009-07-02, 2009-07-01 and
857         2009-06-30 commits.
858
859 2009-07-01  Richard Sandiford  <r.sandiford@uk.ibm.com>
860
861         * testsuite/lib/libffi-dg.exp (libffi-init): Set ld_library_path
862         to "" before adding paths.  (This reinstates an assignment that
863         was removed by my 2009-06-30 commit, but changes the initial
864         value from "." to "".)
865
866 2009-07-01  H.J. Lu  <hongjiu.lu@intel.com>
867
868         PR testsuite/40601
869         * testsuite/lib/libffi-dg.exp (libffi-init): Properly set
870         gccdir.  Adjust ld_library_path for gcc only if gccdir isn't
871         empty.
872
873 2009-06-30  Richard Sandiford  <r.sandiford@uk.ibm.com>
874
875         * testsuite/lib/libffi-dg.exp (libffi-init): Don't add "."
876         to ld_library_path.  Use add_path.  Add just find_libgcc_s
877         to ld_library_path, not every libgcc multilib directory.
878
879 2009-06-16  Wim Lewis  <wiml@hhhh.org>
880
881         * src/powerpc/ffi.c: Avoid clobbering cr3 and cr4, which are
882         supposed to be callee-saved.
883         * src/powerpc/sysv.S (small_struct_return_value): Fix overrun of
884         return buffer for odd-size structs.
885
886 2009-06-16  Andreas Tobler  <a.tobler@schweiz.org>
887
888         PR libffi/40444
889         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Add
890         allow_stack_execute for Darwin.
891
892 2009-06-16  Andrew Haley  <aph@redhat.com>
893
894         * configure.ac (TARGETDIR): Add missing blank lines.
895         * configure: Regenerate.
896
897 2009-06-16  Andrew Haley  <aph@redhat.com>
898
899         * testsuite/libffi.call/cls_align_sint64.c,
900         testsuite/libffi.call/cls_align_uint64.c,
901         testsuite/libffi.call/cls_longdouble_va.c,
902         testsuite/libffi.call/cls_ulonglong.c,
903         testsuite/libffi.call/return_ll1.c,
904         testsuite/libffi.call/stret_medium2.c: Fix printf format
905         specifiers.
906         * testsuite/libffi.call/ffitest.h,
907         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRIuLL): Define.
908
909 2009-06-15  Andrew Haley  <aph@redhat.com>
910
911         * testsuite/libffi.call/err_bad_typedef.c: xfail everywhere.
912         * testsuite/libffi.call/err_bad_abi.c: Likewise.
913
914 2009-06-12  Andrew Haley  <aph@redhat.com>
915
916         * Makefile.am: Remove info_TEXINFOS.
917
918 2009-06-12  Andrew Haley  <aph@redhat.com>
919
920         * ChangeLog.libffi: testsuite/libffi.call/cls_align_sint64.c,
921         testsuite/libffi.call/cls_align_uint64.c,
922         testsuite/libffi.call/cls_ulonglong.c,
923         testsuite/libffi.call/return_ll1.c,
924         testsuite/libffi.call/stret_medium2.c: Fix printf format
925         specifiers.
926         testsuite/libffi.special/unwindtest.cc: include stdint.h.
927
928 2009-06-11  Timothy Wall  <twall@users.sf.net>
929
930         * Makefile.am,
931         configure.ac,
932         include/ffi.h.in,
933         include/ffi_common.h,
934         src/closures.c,
935         src/dlmalloc.c,
936         src/x86/ffi.c,
937         src/x86/ffitarget.h,
938         src/x86/win64.S (new),
939         README: Added win64 support (mingw or MSVC)
940         * Makefile.in,
941         include/Makefile.in,
942         man/Makefile.in,
943         testsuite/Makefile.in,
944         configure,
945         aclocal.m4: Regenerated
946         * ltcf-c.sh: properly escape cygwin/w32 path
947         * man/ffi_call.3: Clarify size requirements for return value.
948         * src/x86/ffi64.c: Fix filename in comment.
949         * src/x86/win32.S: Remove unused extern.
950
951         * testsuite/libffi.call/closure_fn0.c,
952         testsuite/libffi.call/closure_fn1.c,
953         testsuite/libffi.call/closure_fn2.c,
954         testsuite/libffi.call/closure_fn3.c,
955         testsuite/libffi.call/closure_fn4.c,
956         testsuite/libffi.call/closure_fn5.c,
957         testsuite/libffi.call/closure_fn6.c,
958         testsuite/libffi.call/closure_stdcall.c,
959         testsuite/libffi.call/cls_12byte.c,
960         testsuite/libffi.call/cls_16byte.c,
961         testsuite/libffi.call/cls_18byte.c,
962         testsuite/libffi.call/cls_19byte.c,
963         testsuite/libffi.call/cls_1_1byte.c,
964         testsuite/libffi.call/cls_20byte.c,
965         testsuite/libffi.call/cls_20byte1.c,
966         testsuite/libffi.call/cls_24byte.c,
967         testsuite/libffi.call/cls_2byte.c,
968         testsuite/libffi.call/cls_3_1byte.c,
969         testsuite/libffi.call/cls_3byte1.c,
970         testsuite/libffi.call/cls_3byte2.c,
971         testsuite/libffi.call/cls_4_1byte.c,
972         testsuite/libffi.call/cls_4byte.c,
973         testsuite/libffi.call/cls_5_1_byte.c,
974         testsuite/libffi.call/cls_5byte.c,
975         testsuite/libffi.call/cls_64byte.c,
976         testsuite/libffi.call/cls_6_1_byte.c,
977         testsuite/libffi.call/cls_6byte.c,
978         testsuite/libffi.call/cls_7_1_byte.c,
979         testsuite/libffi.call/cls_7byte.c,
980         testsuite/libffi.call/cls_8byte.c,
981         testsuite/libffi.call/cls_9byte1.c,
982         testsuite/libffi.call/cls_9byte2.c,
983         testsuite/libffi.call/cls_align_double.c,
984         testsuite/libffi.call/cls_align_float.c,
985         testsuite/libffi.call/cls_align_longdouble.c,
986         testsuite/libffi.call/cls_align_longdouble_split.c,
987         testsuite/libffi.call/cls_align_longdouble_split2.c,
988         testsuite/libffi.call/cls_align_pointer.c,
989         testsuite/libffi.call/cls_align_sint16.c,
990         testsuite/libffi.call/cls_align_sint32.c,
991         testsuite/libffi.call/cls_align_sint64.c,
992         testsuite/libffi.call/cls_align_uint16.c,
993         testsuite/libffi.call/cls_align_uint32.c,
994         testsuite/libffi.call/cls_align_uint64.c,
995         testsuite/libffi.call/cls_dbls_struct.c,
996         testsuite/libffi.call/cls_double.c,
997         testsuite/libffi.call/cls_double_va.c,
998         testsuite/libffi.call/cls_float.c,
999         testsuite/libffi.call/cls_longdouble.c,
1000         testsuite/libffi.call/cls_longdouble_va.c,
1001         testsuite/libffi.call/cls_multi_schar.c,
1002         testsuite/libffi.call/cls_multi_sshort.c,
1003         testsuite/libffi.call/cls_multi_sshortchar.c,
1004         testsuite/libffi.call/cls_multi_uchar.c,
1005         testsuite/libffi.call/cls_multi_ushort.c,
1006         testsuite/libffi.call/cls_multi_ushortchar.c,
1007         testsuite/libffi.call/cls_pointer.c,
1008         testsuite/libffi.call/cls_pointer_stack.c,
1009         testsuite/libffi.call/cls_schar.c,
1010         testsuite/libffi.call/cls_sint.c,
1011         testsuite/libffi.call/cls_sshort.c,
1012         testsuite/libffi.call/cls_uchar.c,
1013         testsuite/libffi.call/cls_uint.c,
1014         testsuite/libffi.call/cls_ulonglong.c,
1015         testsuite/libffi.call/cls_ushort.c,
1016         testsuite/libffi.call/err_bad_abi.c,
1017         testsuite/libffi.call/err_bad_typedef.c,
1018         testsuite/libffi.call/float2.c,
1019         testsuite/libffi.call/huge_struct.c,
1020         testsuite/libffi.call/nested_struct.c,
1021         testsuite/libffi.call/nested_struct1.c,
1022         testsuite/libffi.call/nested_struct10.c,
1023         testsuite/libffi.call/nested_struct2.c,
1024         testsuite/libffi.call/nested_struct3.c,
1025         testsuite/libffi.call/nested_struct4.c,
1026         testsuite/libffi.call/nested_struct5.c,
1027         testsuite/libffi.call/nested_struct6.c,
1028         testsuite/libffi.call/nested_struct7.c,
1029         testsuite/libffi.call/nested_struct8.c,
1030         testsuite/libffi.call/nested_struct9.c,
1031         testsuite/libffi.call/problem1.c,
1032         testsuite/libffi.call/return_ldl.c,
1033         testsuite/libffi.call/return_ll1.c,
1034         testsuite/libffi.call/stret_large.c,
1035         testsuite/libffi.call/stret_large2.c,
1036         testsuite/libffi.call/stret_medium.c,
1037         testsuite/libffi.call/stret_medium2.c,
1038         testsuite/libffi.special/unwindtest.cc: use ffi_closure_alloc instead
1039         of checking for MMAP.  Use intptr_t instead of long casts.
1040
1041 2009-06-11  Kaz Kojima  <kkojima@gcc.gnu.org>
1042
1043         * testsuite/libffi.call/cls_longdouble_va.c: Add xfail sh*-*-linux-*.
1044         * testsuite/libffi.call/err_bad_abi.c: Add xfail sh*-*-*.
1045         * testsuite/libffi.call/err_bad_typedef.c: Likewise.
1046
1047 2009-06-09  Andrew Haley  <aph@redhat.com>
1048
1049         * src/x86/freebsd.S: Add missing file.
1050
1051 2009-06-08  Andrew Haley  <aph@redhat.com>
1052
1053         Import from libffi 3.0.8:
1054
1055         * doc/libffi.texi: New file.
1056         * doc/libffi.info: Likewise.
1057         * doc/stamp-vti: Likewise.
1058         * man/Makefile.am: New file.
1059         * man/ffi_call.3: New file.
1060
1061         * Makefile.am (EXTRA_DIST): Add src/x86/darwin64.S,
1062         src/dlmalloc.c.
1063         (nodist_libffi_la_SOURCES): Add X86_FREEBSD.
1064
1065         * configure.ac: Bump version to 3.0.8.
1066         parisc*-*-linux*: Add.
1067         i386-*-freebsd* | i386-*-openbsd*: Add.
1068         powerpc-*-beos*: Add.
1069         AM_CONDITIONAL X86_FREEBSD: Add.
1070         AC_CONFIG_FILES: Add man/Makefile.
1071
1072         * include/ffi.h.in (FFI_FN): Change void (*)() to void (*)(void).
1073
1074 2009-06-08  Andrew Haley  <aph@redhat.com>
1075
1076         * README: Import from libffi 3.0.8.
1077
1078 2009-06-08  Andrew Haley  <aph@redhat.com>
1079
1080         * testsuite/libffi.call/err_bad_abi.c: Add xfails.
1081         * testsuite/libffi.call/cls_longdouble_va.c: Add xfails.
1082         * testsuite/libffi.call/cls_dbls_struct.c: Add xfail x86_64-*-linux-*.
1083         * testsuite/libffi.call/err_bad_typedef.c: Add xfails.
1084
1085         * testsuite/libffi.call/stret_medium2.c: Add __UNUSED__ to args.
1086         * testsuite/libffi.call/stret_medium.c: Likewise.
1087         * testsuite/libffi.call/stret_large2.c: Likewise.
1088         * testsuite/libffi.call/stret_large.c:  Likewise.
1089
1090 2008-12-26  Timothy Wall  <twall@users.sf.net>
1091
1092         * testsuite/libffi.call/cls_longdouble.c,
1093         testsuite/libffi.call/cls_longdouble_va.c,
1094         testsuite/libffi.call/cls_align_longdouble.c,
1095         testsuite/libffi.call/cls_align_longdouble_split.c,
1096         testsuite/libffi.call/cls_align_longdouble_split2.c: mark expected
1097         failures on x86_64 cygwin/mingw.
1098
1099 2008-12-22  Timothy Wall  <twall@users.sf.net>
1100
1101         * testsuite/libffi.call/closure_fn0.c,
1102         testsuite/libffi.call/closure_fn1.c,
1103         testsuite/libffi.call/closure_fn2.c,
1104         testsuite/libffi.call/closure_fn3.c,
1105         testsuite/libffi.call/closure_fn4.c,
1106         testsuite/libffi.call/closure_fn5.c,
1107         testsuite/libffi.call/closure_fn6.c,
1108         testsuite/libffi.call/closure_loc_fn0.c,
1109         testsuite/libffi.call/closure_stdcall.c,
1110         testsuite/libffi.call/cls_align_pointer.c,
1111         testsuite/libffi.call/cls_pointer.c,
1112         testsuite/libffi.call/cls_pointer_stack.c: use portable cast from
1113         pointer to integer (intptr_t).
1114         * testsuite/libffi.call/cls_longdouble.c: disable for win64.
1115
1116 2008-07-24  Anthony Green  <green@redhat.com>
1117
1118         * testsuite/libffi.call/cls_dbls_struct.c,
1119         testsuite/libffi.call/cls_double_va.c,
1120         testsuite/libffi.call/cls_longdouble.c,
1121         testsuite/libffi.call/cls_longdouble_va.c,
1122         testsuite/libffi.call/cls_pointer.c,
1123         testsuite/libffi.call/cls_pointer_stack.c,
1124         testsuite/libffi.call/err_bad_abi.c: Clean up failures from
1125         compiler warnings.
1126
1127 2008-03-04  Anthony Green  <green@redhat.com>
1128             Blake Chaffin
1129             hos@tamanegi.org
1130
1131         * testsuite/libffi.call/cls_align_longdouble_split2.c
1132           testsuite/libffi.call/cls_align_longdouble_split.c
1133           testsuite/libffi.call/cls_dbls_struct.c
1134           testsuite/libffi.call/cls_double_va.c
1135           testsuite/libffi.call/cls_longdouble.c
1136           testsuite/libffi.call/cls_longdouble_va.c
1137           testsuite/libffi.call/cls_pointer.c
1138           testsuite/libffi.call/cls_pointer_stack.c
1139           testsuite/libffi.call/err_bad_abi.c
1140           testsuite/libffi.call/err_bad_typedef.c
1141           testsuite/libffi.call/stret_large2.c
1142           testsuite/libffi.call/stret_large.c
1143           testsuite/libffi.call/stret_medium2.c
1144           testsuite/libffi.call/stret_medium.c: New tests from Apple.
1145
1146 2009-06-05  Andrew Haley  <aph@redhat.com>
1147
1148         * src/x86/ffitarget.h, src/x86/ffi.c: Merge stdcall changes from
1149         libffi.
1150
1151 2009-06-04  Andrew Haley  <aph@redhat.com>
1152
1153         * src/x86/ffitarget.h, src/x86/win32.S, src/x86/ffi.c: Back out
1154         stdcall changes.
1155
1156 2008-02-26  Anthony Green  <green@redhat.com>
1157             Thomas Heller  <theller@ctypes.org>
1158
1159         * src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C
1160         comment.
1161
1162 2008-02-03  Timothy Wall  <twall@users.sf.net>
1163
1164         * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return
1165           offset based on code pointer, not data pointer.
1166
1167 2008-01-31  Timothy Wall <twall@users.sf.net>
1168
1169         * testsuite/libffi.call/closure_stdcall.c: Add test for stdcall
1170         closures.
1171         * src/x86/ffitarget.h: Increase size of trampoline for stdcall
1172         closures.
1173         * src/x86/win32.S: Add assembly for stdcall closure.
1174         * src/x86/ffi.c: Initialize stdcall closure trampoline.
1175
1176 2009-06-04  Andrew Haley  <aph@redhat.com>
1177
1178         * include/ffi.h.in: Change void (*)() to void (*)(void).
1179         * src/x86/ffi.c: Likewise.
1180
1181 2009-06-04  Andrew Haley  <aph@redhat.com>
1182
1183         * src/powerpc/ppc_closure.S: Insert licence header.
1184         * src/powerpc/linux64_closure.S: Likewise.
1185         * src/m68k/sysv.S: Likewise.
1186
1187         * src/sh64/ffi.c: Change void (*)() to void (*)(void).
1188         * src/powerpc/ffi.c: Likewise.
1189         * src/powerpc/ffi_darwin.c: Likewise.
1190         * src/m32r/ffi.c: Likewise.
1191         * src/sh64/ffi.c: Likewise.
1192         * src/x86/ffi64.c: Likewise.
1193         * src/alpha/ffi.c: Likewise.
1194         * src/alpha/osf.S: Likewise.
1195         * src/frv/ffi.c: Likewise.
1196         * src/s390/ffi.c: Likewise.
1197         * src/pa/ffi.c: Likewise.
1198         * src/pa/hpux32.S: Likewise.
1199         * src/ia64/unix.S: Likewise.
1200         * src/ia64/ffi.c: Likewise.
1201         * src/sparc/ffi.c: Likewise.
1202         * src/mips/ffi.c: Likewise.
1203         * src/sh/ffi.c: Likewise.
1204
1205 2008-02-15  David Daney  <ddaney@avtrex.com>
1206
1207         * src/mips/ffi.c (USE__BUILTIN___CLEAR_CACHE):
1208         Define (conditionally), and use it to include cachectl.h.
1209         (ffi_prep_closure_loc): Fix cache flushing.
1210         * src/mips/ffitarget.h (_ABIN32, _ABI64, _ABIO32): Define.
1211
1212 2009-06-04  Andrew Haley  <aph@redhat.com>
1213
1214         include/ffi.h.in,
1215         src/arm/ffitarget.h,
1216         src/arm/ffi.c,
1217         src/arm/sysv.S,
1218         src/powerpc/ffitarget.h,
1219         src/closures.c,
1220         src/sh64/ffitarget.h,
1221         src/sh64/ffi.c,
1222         src/sh64/sysv.S,
1223         src/types.c,
1224         src/x86/ffi64.c,
1225         src/x86/ffitarget.h,
1226         src/x86/win32.S,
1227         src/x86/darwin.S,
1228         src/x86/ffi.c,
1229         src/x86/sysv.S,
1230         src/x86/unix64.S,
1231         src/alpha/ffitarget.h,
1232         src/alpha/ffi.c,
1233         src/alpha/osf.S,
1234         src/m68k/ffitarget.h,
1235         src/frv/ffitarget.h,
1236         src/frv/ffi.c,
1237         src/s390/ffitarget.h,
1238         src/s390/sysv.S,
1239         src/cris/ffitarget.h,
1240         src/pa/linux.S,
1241         src/pa/ffitarget.h,
1242         src/pa/ffi.c,
1243         src/raw_api.c,
1244         src/ia64/ffitarget.h,
1245         src/ia64/unix.S,
1246         src/ia64/ffi.c,
1247         src/ia64/ia64_flags.h,
1248         src/java_raw_api.c,
1249         src/debug.c,
1250         src/sparc/v9.S,
1251         src/sparc/ffitarget.h,
1252         src/sparc/ffi.c,
1253         src/sparc/v8.S,
1254         src/mips/ffitarget.h,
1255         src/mips/n32.S,
1256         src/mips/o32.S,
1257         src/mips/ffi.c,
1258         src/prep_cif.c,
1259         src/sh/ffitarget.h,
1260         src/sh/ffi.c,
1261         src/sh/sysv.S: Update license text.
1262
1263 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
1264
1265         * src/x86/win32.S (_ffi_closure_STDCALL):  New function.
1266         (.eh_frame):  Add FDE for it.
1267
1268 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
1269
1270         * configure.ac:  Also check if assembler supports pc-relative
1271         relocs on X86_WIN32 targets.
1272         * configure:  Regenerate.
1273         * src/x86/win32.S (ffi_prep_args):  Declare extern, not global.
1274         (_ffi_call_SYSV):  Add missing function type symbol .def and
1275         add EH markup labels.
1276         (_ffi_call_STDCALL):  Likewise.
1277         (_ffi_closure_SYSV):  Likewise.
1278         (_ffi_closure_raw_SYSV):  Likewise.
1279         (.eh_frame):  Add hand-crafted EH data.
1280
1281 2009-04-09  Jakub Jelinek  <jakub@redhat.com>
1282
1283         * testsuite/lib/libffi-dg.exp: Change copyright header to refer to
1284         version 3 of the GNU General Public License and to point readers
1285         at the COPYING3 file and the FSF's license web page.
1286         * testsuite/libffi.call/call.exp: Likewise.
1287         * testsuite/libffi.special/special.exp: Likewise.
1288
1289 2009-03-01  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1290
1291         * configure: Regenerate.
1292
1293 2008-12-18  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1294
1295         PR libffi/26048
1296         * configure.ac (HAVE_AS_X86_PCREL): New test.
1297         * configure: Regenerate.
1298         * fficonfig.h.in: Regenerate.
1299         * src/x86/sysv.S [!FFI_NO_RAW_API]: Precalculate
1300         RAW_CLOSURE_CIF_OFFSET, RAW_CLOSURE_FUN_OFFSET,
1301         RAW_CLOSURE_USER_DATA_OFFSET for the Solaris 10/x86 assembler.
1302         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
1303         * src/x86/unix64.S (.Lstore_table): Move to .text section.
1304         (.Lload_table): Likewise.
1305         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
1306
1307 2008-12-18  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1308
1309         * configure: Regenerate.
1310
1311 2008-11-21  Eric Botcazou  <ebotcazou@adacore.com>
1312
1313         * src/sparc/ffi.c (ffi_prep_cif_machdep): Add support for
1314         signed/unsigned int8/16 return values.
1315         * src/sparc/v8.S (ffi_call_v8): Likewise.
1316         (ffi_closure_v8): Likewise.
1317
1318 2008-09-26  Peter O'Gorman  <pogma@thewrittenword.com>
1319             Steve Ellcey  <sje@cup.hp.com>
1320
1321         * configure: Regenerate for new libtool.
1322         * Makefile.in: Ditto.
1323         * include/Makefile.in: Ditto.
1324         * aclocal.m4: Ditto.
1325
1326 2008-08-25  Andreas Tobler  <a.tobler@schweiz.org>
1327
1328         * src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and
1329         FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum.
1330         Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT.
1331         Adjust copyright notice.
1332         * src/powerpc/ffi.c: Add two new flags to indicate if we have one
1333         register or two register to use for FFI_SYSV structs.
1334         (ffi_prep_cif_machdep): Pass the right register flag introduced above.
1335         (ffi_closure_helper_SYSV): Fix the return type for
1336         FFI_SYSV_TYPE_SMALL_STRUCT. Comment.
1337         Adjust copyright notice.
1338
1339 2008-07-16  Kaz Kojima  <kkojima@gcc.gnu.org>
1340
1341         * src/sh/ffi.c (ffi_prep_closure_loc): Turn INSN into an unsigned
1342         int.
1343
1344 2008-06-17  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1345
1346         * configure: Regenerate.
1347         * include/Makefile.in: Regenerate.
1348         * testsuite/Makefile.in: Regenerate.
1349
1350 2008-06-07  Joseph Myers  <joseph@codesourcery.com>
1351
1352         * configure.ac (parisc*-*-linux*, powerpc-*-sysv*,
1353         powerpc-*-beos*): Remove.
1354         * configure: Regenerate.
1355
1356 2008-05-09  Julian Brown  <julian@codesourcery.com>
1357
1358         * Makefile.am (LTLDFLAGS): New.
1359         (libffi_la_LDFLAGS): Use above.
1360         * Makefile.in: Regenerate.
1361
1362 2008-04-18  Paolo Bonzini  <bonzini@gnu.org>
1363
1364         PR bootstrap/35457
1365         * aclocal.m4: Regenerate.
1366         * configure: Regenerate.
1367
1368 2008-03-26  Kaz Kojima  <kkojima@gcc.gnu.org>
1369
1370         * src/sh/sysv.S: Add .note.GNU-stack on Linux.
1371         * src/sh64/sysv.S: Likewise.
1372
1373 2008-03-26  Daniel Jacobowitz  <dan@debian.org>
1374
1375         * src/arm/sysv.S: Fix ARM comment marker.
1376
1377 2008-03-26  Jakub Jelinek  <jakub@redhat.com>
1378
1379         * src/alpha/osf.S: Add .note.GNU-stack on Linux.
1380         * src/s390/sysv.S: Likewise.
1381         * src/powerpc/ppc_closure.S: Likewise.
1382         * src/powerpc/sysv.S: Likewise.
1383         * src/x86/unix64.S: Likewise.
1384         * src/x86/sysv.S: Likewise.
1385         * src/sparc/v8.S: Likewise.
1386         * src/sparc/v9.S: Likewise.
1387         * src/m68k/sysv.S: Likewise.
1388         * src/arm/sysv.S: Likewise.
1389
1390 2008-03-16  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1391
1392         * aclocal.m4: Regenerate.
1393         * configure: Likewise.
1394         * Makefile.in: Likewise.
1395         * include/Makefile.in: Likewise.
1396         * testsuite/Makefile.in: Likewise.
1397
1398 2008-02-12  Bjoern Koenig  <bkoenig@alpha-tierchen.de>
1399             Andreas Tobler  <a.tobler@schweiz.org>
1400
1401         * configure.ac: Add amd64-*-freebsd* target.
1402         * configure: Regenerate.
1403
1404 2008-01-30  H.J. Lu  <hongjiu.lu@intel.com>
1405
1406         PR libffi/34612
1407         * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when
1408         returning struct.
1409
1410         * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer"
1411         tests.
1412
1413 2008-01-24  David Edelsohn  <edelsohn@gnu.org>
1414
1415         * configure: Regenerate.
1416
1417 2008-01-06  Andreas Tobler  <a.tobler@schweiz.org>
1418
1419         * src/x86/ffi.c (ffi_prep_cif_machdep): Fix thinko.
1420
1421 2008-01-05  Andreas Tobler  <a.tobler@schweiz.org>
1422
1423         PR testsuite/32843
1424         * src/x86/ffi.c (ffi_prep_cif_machdep): Add code for
1425         signed/unsigned int8/16 for X86_DARWIN.
1426         Updated copyright info.
1427         Handle one and two byte structs with special cif->flags.
1428         * src/x86/ffitarget.h: Add special types for one and two byte structs.
1429         Updated copyright info.
1430         * src/x86/darwin.S (ffi_call_SYSV): Rewrite to use a jump table like
1431         sysv.S
1432         Remove code to pop args from the stack after call.
1433         Special-case signed/unsigned for int8/16, one and two byte structs.
1434         (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1435         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1436         FFI_TYPE_SINT32.
1437         Updated copyright info.
1438
1439 2007-12-08  David Daney  <ddaney@avtrex.com>
1440
1441         * src/mips/n32.S (ffi_call_N32):  Replace dadd with ADDU, dsub with
1442         SUBU, add with ADDU and use smaller code sequences.
1443
1444 2007-12-07  David Daney  <ddaney@avtrex.com>
1445
1446         * src/mips/ffi.c (ffi_prep_cif_machdep): Handle long double return
1447         type.
1448
1449 2007-12-06  David Daney  <ddaney@avtrex.com>
1450
1451         * include/ffi.h.in (FFI_SIZEOF_JAVA_RAW): Define if not already
1452         defined.
1453         (ffi_java_raw): New typedef.
1454         (ffi_java_raw_call, ffi_java_ptrarray_to_raw,
1455         ffi_java_raw_to_ptrarray): Change parameter types from ffi_raw to
1456         ffi_java_raw.
1457         (ffi_java_raw_closure) : Same.
1458         (ffi_prep_java_raw_closure, ffi_prep_java_raw_closure_loc): Change
1459         parameter types.
1460         * src/java_raw_api.c (ffi_java_raw_size):  Replace FFI_SIZEOF_ARG with
1461         FFI_SIZEOF_JAVA_RAW.
1462         (ffi_java_raw_to_ptrarray): Change type of raw to ffi_java_raw.
1463         Replace FFI_SIZEOF_ARG with FFI_SIZEOF_JAVA_RAW. Use
1464         sizeof(ffi_java_raw) for alignment calculations.
1465         (ffi_java_ptrarray_to_raw): Same.
1466         (ffi_java_rvalue_to_raw): Add special handling for FFI_TYPE_POINTER
1467         if FFI_SIZEOF_JAVA_RAW == 4.
1468         (ffi_java_raw_to_rvalue): Same.
1469         (ffi_java_raw_call): Change type of raw to ffi_java_raw.
1470         (ffi_java_translate_args): Same.
1471         (ffi_prep_java_raw_closure_loc, ffi_prep_java_raw_closure): Change
1472         parameter types.
1473         * src/mips/ffitarget.h (FFI_SIZEOF_JAVA_RAW): Define for N32 ABI.
1474
1475 2007-12-06  David Daney  <ddaney@avtrex.com>
1476
1477         * src/mips/n32.S (ffi_closure_N32): Use 64-bit add instruction on
1478         pointer values.
1479
1480 2007-12-01  Andreas Tobler  <a.tobler@schweiz.org>
1481
1482         PR libffi/31937
1483         * src/powerpc/ffitarget.h: Introduce new ABI FFI_LINUX_SOFT_FLOAT.
1484         Add local FFI_TYPE_UINT128 to handle soft-float long-double-128.
1485         * src/powerpc/ffi.c: Distinguish between __NO_FPRS__ and not and
1486         set the NUM_FPR_ARG_REGISTERS according to.
1487         Add support for potential soft-float support under hard-float
1488         architecture.
1489         (ffi_prep_args_SYSV): Set NUM_FPR_ARG_REGISTERS to 0 in case of
1490         FFI_LINUX_SOFT_FLOAT, handle float, doubles and long-doubles according
1491         to the FFI_LINUX_SOFT_FLOAT ABI.
1492         (ffi_prep_cif_machdep): Likewise.
1493         (ffi_closure_helper_SYSV): Likewise.
1494         * src/powerpc/ppc_closure.S: Make sure not to store float/double
1495         on archs where __NO_FPRS__ is true.
1496         Add FFI_TYPE_UINT128 support.
1497         * src/powerpc/sysv.S: Add support for soft-float long-double-128.
1498         Adjust copyright notice.
1499
1500 2007-11-25  Andreas Tobler  <a.tobler@schweiz.org>
1501
1502         * src/closures.c: Move defintion of MAYBE_UNUSED from here to ...
1503         * include/ffi_common.h: ... here.
1504         Update copyright.
1505
1506 2007-11-17  Andreas Tobler  <a.tobler@schweiz.org>
1507
1508         * src/powerpc/sysv.S: Load correct cr to compare if we have long double.
1509         * src/powerpc/linux64.S: Likewise.
1510         * src/powerpc/ffi.c: Add a comment to show which part goes into cr6.
1511         * testsuite/libffi.call/return_ldl.c: New test.
1512
1513 2007-09-04    <aph@redhat.com>
1514
1515         * src/arm/sysv.S (UNWIND): New.
1516         (Whole file): Conditionally compile unwinder directives.
1517         * src/arm/sysv.S: Add unwinder directives.
1518
1519         * src/arm/ffi.c (ffi_prep_args): Align structs by at least 4 bytes.
1520         Only treat r0 as a struct address if we're actually returning a
1521         struct by address.
1522         Only copy the bytes that are actually within a struct.
1523         (ffi_prep_cif_machdep): A Composite Type not larger than 4 bytes
1524         is returned in r0, not passed by address.
1525         (ffi_call): Allocate a word-sized temporary for the case where
1526         a composite is returned in r0.
1527         (ffi_prep_incoming_args_SYSV): Align as necessary.
1528
1529 2007-08-05  Steven Newbury  <s_j_newbury@yahoo.co.uk>
1530
1531         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Use __clear_cache instead of
1532         directly using the sys_cacheflush syscall.
1533
1534 2007-07-27  Andrew Haley  <aph@redhat.com>
1535
1536         * src/arm/sysv.S (ffi_closure_SYSV): Add soft-float.
1537
1538 2007-09-03  Maciej W. Rozycki  <macro@linux-mips.org>
1539
1540         * Makefile.am: Unify MIPS_IRIX and MIPS_LINUX into MIPS.
1541         * configure.ac: Likewise.
1542         * Makefile.in: Regenerate.
1543         * include/Makefile.in: Likewise.
1544         * testsuite/Makefile.in: Likewise.
1545         * configure: Likewise.
1546
1547 2007-08-24  David Daney  <ddaney@avtrex.com>
1548
1549         * testsuite/libffi.call/return_sl.c: New test.
1550
1551 2007-08-10  David Daney  <ddaney@avtrex.com>
1552
1553         * testsuite/libffi.call/cls_multi_ushort.c,
1554         testsuite/libffi.call/cls_align_uint16.c,
1555         testsuite/libffi.call/nested_struct1.c,
1556         testsuite/libffi.call/nested_struct3.c,
1557         testsuite/libffi.call/cls_7_1_byte.c,
1558         testsuite/libffi.call/nested_struct5.c,
1559         testsuite/libffi.call/cls_double.c,
1560         testsuite/libffi.call/nested_struct7.c,
1561         testsuite/libffi.call/cls_sint.c,
1562         testsuite/libffi.call/nested_struct9.c,
1563         testsuite/libffi.call/cls_20byte1.c,
1564         testsuite/libffi.call/cls_multi_sshortchar.c,
1565         testsuite/libffi.call/cls_align_sint64.c,
1566         testsuite/libffi.call/cls_3byte2.c,
1567         testsuite/libffi.call/cls_multi_schar.c,
1568         testsuite/libffi.call/cls_multi_uchar.c,
1569         testsuite/libffi.call/cls_19byte.c,
1570         testsuite/libffi.call/cls_9byte1.c,
1571         testsuite/libffi.call/cls_align_float.c,
1572         testsuite/libffi.call/closure_fn1.c,
1573         testsuite/libffi.call/problem1.c,
1574         testsuite/libffi.call/closure_fn3.c,
1575         testsuite/libffi.call/cls_sshort.c,
1576         testsuite/libffi.call/closure_fn5.c,
1577         testsuite/libffi.call/cls_align_double.c,
1578         testsuite/libffi.call/nested_struct.c,
1579         testsuite/libffi.call/cls_2byte.c,
1580         testsuite/libffi.call/nested_struct10.c,
1581         testsuite/libffi.call/cls_4byte.c,
1582         testsuite/libffi.call/cls_6byte.c,
1583         testsuite/libffi.call/cls_8byte.c,
1584         testsuite/libffi.call/cls_multi_sshort.c,
1585         testsuite/libffi.call/cls_align_sint16.c,
1586         testsuite/libffi.call/cls_align_uint32.c,
1587         testsuite/libffi.call/cls_20byte.c,
1588         testsuite/libffi.call/cls_float.c,
1589         testsuite/libffi.call/nested_struct2.c,
1590         testsuite/libffi.call/cls_5_1_byte.c,
1591         testsuite/libffi.call/nested_struct4.c,
1592         testsuite/libffi.call/cls_24byte.c,
1593         testsuite/libffi.call/nested_struct6.c,
1594         testsuite/libffi.call/cls_64byte.c,
1595         testsuite/libffi.call/nested_struct8.c,
1596         testsuite/libffi.call/cls_uint.c,
1597         testsuite/libffi.call/cls_multi_ushortchar.c,
1598         testsuite/libffi.call/cls_schar.c,
1599         testsuite/libffi.call/cls_uchar.c,
1600         testsuite/libffi.call/cls_align_uint64.c,
1601         testsuite/libffi.call/cls_ulonglong.c,
1602         testsuite/libffi.call/cls_align_longdouble.c,
1603         testsuite/libffi.call/cls_1_1byte.c,
1604         testsuite/libffi.call/cls_12byte.c,
1605         testsuite/libffi.call/cls_3_1byte.c,
1606         testsuite/libffi.call/cls_3byte1.c,
1607         testsuite/libffi.call/cls_4_1byte.c,
1608         testsuite/libffi.call/cls_6_1_byte.c,
1609         testsuite/libffi.call/cls_16byte.c,
1610         testsuite/libffi.call/cls_18byte.c,
1611         testsuite/libffi.call/closure_fn0.c,
1612         testsuite/libffi.call/cls_9byte2.c,
1613         testsuite/libffi.call/closure_fn2.c,
1614         testsuite/libffi.call/closure_fn4.c,
1615         testsuite/libffi.call/cls_ushort.c,
1616         testsuite/libffi.call/closure_fn6.c,
1617         testsuite/libffi.call/cls_5byte.c,
1618         testsuite/libffi.call/cls_align_pointer.c,
1619         testsuite/libffi.call/cls_7byte.c,
1620         testsuite/libffi.call/cls_align_sint32.c,
1621         testsuite/libffi.special/unwindtest_ffi_call.cc,
1622         testsuite/libffi.special/unwindtest.cc: Remove xfail for mips64*-*-*.
1623
1624 2007-08-10  David Daney  <ddaney@avtrex.com>
1625
1626         PR libffi/28313
1627         * configure.ac: Don't treat mips64 as a special case.
1628         * Makefile.am (nodist_libffi_la_SOURCES): Add n32.S.
1629         * configure: Regenerate
1630         * Makefile.in: Ditto.
1631         * fficonfig.h.in: Ditto.
1632         * src/mips/ffitarget.h (REG_L, REG_S, SUBU, ADDU, SRL, LI): Indent.
1633         (LA, EH_FRAME_ALIGN, FDE_ADDR_BYTES): New preprocessor macros.
1634         (FFI_DEFAULT_ABI): Set for n64 case.
1635         (FFI_CLOSURES, FFI_TRAMPOLINE_SIZE): Define for n32 and n64 cases.
1636         * src/mips/n32.S (ffi_call_N32): Add debug macros and labels for FDE.
1637         (ffi_closure_N32): New function.
1638         (.eh_frame): New section
1639         * src/mips/o32.S: Clean up comments.
1640         (ffi_closure_O32): Pass ffi_closure parameter in $12.
1641         * src/mips/ffi.c: Use FFI_MIPS_N32 instead of
1642         _MIPS_SIM == _ABIN32 throughout.
1643         (FFI_MIPS_STOP_HERE): New, use in place of
1644         ffi_stop_here.
1645         (ffi_prep_args): Use unsigned long to hold pointer values.  Rewrite
1646         to support n32/n64 ABIs.
1647         (calc_n32_struct_flags): Rewrite.
1648         (calc_n32_return_struct_flags): Remove unused variable.  Reverse
1649         position of flag bits.
1650         (ffi_prep_cif_machdep): Rewrite n32 portion.
1651         (ffi_call): Enable for n64.  Add special handling for small structure
1652         return values.
1653         (ffi_prep_closure_loc): Add n32 and n64 support.
1654         (ffi_closure_mips_inner_O32): Add cast to silence warning.
1655         (copy_struct_N32, ffi_closure_mips_inner_N32): New functions.
1656
1657 2007-08-08  David Daney  <ddaney@avtrex.com>
1658
1659         * testsuite/libffi.call/ffitest.h (ffi_type_mylong): Remove definition.
1660         * testsuite/libffi.call/cls_align_uint16.c (main): Use correct type
1661         specifiers.
1662         * testsuite/libffi.call/nested_struct1.c (main): Ditto.
1663         * testsuite/libffi.call/cls_sint.c (main): Ditto.
1664         * testsuite/libffi.call/nested_struct9.c (main): Ditto.
1665         * testsuite/libffi.call/cls_20byte1.c (main): Ditto.
1666         * testsuite/libffi.call/cls_9byte1.c (main): Ditto.
1667         * testsuite/libffi.call/closure_fn1.c (main): Ditto.
1668         * testsuite/libffi.call/closure_fn3.c (main): Ditto.
1669         * testsuite/libffi.call/return_dbl2.c (main): Ditto.
1670         * testsuite/libffi.call/cls_sshort.c (main): Ditto.
1671         * testsuite/libffi.call/return_fl3.c (main): Ditto.
1672         * testsuite/libffi.call/closure_fn5.c (main): Ditto.
1673         * testsuite/libffi.call/nested_struct.c (main): Ditto.
1674         * testsuite/libffi.call/nested_struct10.c (main): Ditto.
1675         * testsuite/libffi.call/return_ll1.c (main): Ditto.
1676         * testsuite/libffi.call/cls_8byte.c (main): Ditto.
1677         * testsuite/libffi.call/cls_align_uint32.c (main): Ditto.
1678         * testsuite/libffi.call/cls_align_sint16.c (main): Ditto.
1679         * testsuite/libffi.call/cls_20byte.c (main): Ditto.
1680         * testsuite/libffi.call/nested_struct2.c (main): Ditto.
1681         * testsuite/libffi.call/cls_24byte.c (main): Ditto.
1682         * testsuite/libffi.call/nested_struct6.c (main): Ditto.
1683         * testsuite/libffi.call/cls_uint.c (main): Ditto.
1684         * testsuite/libffi.call/cls_12byte.c (main): Ditto.
1685         * testsuite/libffi.call/cls_16byte.c (main): Ditto.
1686         * testsuite/libffi.call/closure_fn0.c (main): Ditto.
1687         * testsuite/libffi.call/cls_9byte2.c (main): Ditto.
1688         * testsuite/libffi.call/closure_fn2.c (main): Ditto.
1689         * testsuite/libffi.call/return_dbl1.c (main): Ditto.
1690         * testsuite/libffi.call/closure_fn4.c (main): Ditto.
1691         * testsuite/libffi.call/closure_fn6.c (main): Ditto.
1692         * testsuite/libffi.call/cls_align_sint32.c (main): Ditto.
1693
1694 2007-08-07  Andrew Haley  <aph@redhat.com>
1695
1696         * src/x86/sysv.S (ffi_closure_raw_SYSV): Fix typo in previous
1697         checkin.
1698
1699 2007-08-06  Andrew Haley  <aph@redhat.com>
1700
1701         PR testsuite/32843
1702         * src/x86/sysv.S (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1703         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1704         FFI_TYPE_SINT32.
1705
1706 2007-08-02  David Daney  <ddaney@avtrex.com>
1707
1708         * testsuite/libffi.call/return_ul.c (main): Define return type as
1709         ffi_arg.  Use proper printf conversion specifier.
1710
1711 2007-07-30  Andrew Haley  <aph@redhat.com>
1712
1713         PR testsuite/32843
1714         * src/x86/ffi.c (ffi_prep_cif_machdep): in x86 case, add code for
1715         signed/unsigned int8/16.
1716         * src/x86/sysv.S (ffi_call_SYSV): Rewrite to:
1717         Use a jump table.
1718         Remove code to pop args from the stack after call.
1719         Special-case signed/unsigned int8/16.
1720         * testsuite/libffi.call/return_sc.c (main): Revert.
1721
1722 2007-07-26  Richard Guenther  <rguenther@suse.de>
1723
1724         PR testsuite/32843
1725         * testsuite/libffi.call/return_sc.c (main): Verify call
1726         result as signed char, not ffi_arg.
1727
1728 2007-07-16  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1729
1730         * configure.ac (i?86-*-solaris2.1[0-9]): Set TARGET to X86_64.
1731         * configure: Regenerate.
1732
1733 2007-07-11  David Daney  <ddaney@avtrex.com>
1734
1735         * src/mips/ffi.c: Don't include sys/cachectl.h.
1736         (ffi_prep_closure_loc): Use __builtin___clear_cache() instead of
1737         cacheflush().
1738
1739 2007-05-18  Aurelien Jarno  <aurelien@aurel32.net>
1740
1741         * src/arm/ffi.c (ffi_prep_closure_loc): Renamed and ajusted
1742         from (ffi_prep_closure): ... this.
1743         (FFI_INIT_TRAMPOLINE): Adjust.
1744
1745 2005-12-31  Phil Blundell  <pb@reciva.com>
1746
1747         * src/arm/ffi.c (ffi_prep_incoming_args_SYSV,
1748         ffi_closure_SYSV_inner, ffi_prep_closure): New, add closure support.
1749         * src/arm/sysv.S(ffi_closure_SYSV): Likewise.
1750         * src/arm/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1751         (FFI_CLOSURES): Enable closure support.
1752
1753 2007-07-03  Andrew Haley  <aph@hedges.billgatliff.com>
1754
1755         * testsuite/libffi.call/cls_multi_ushort.c,
1756         testsuite/libffi.call/cls_align_uint16.c,
1757         testsuite/libffi.call/nested_struct1.c,
1758         testsuite/libffi.call/nested_struct3.c,
1759         testsuite/libffi.call/cls_7_1_byte.c,
1760         testsuite/libffi.call/cls_double.c,
1761         testsuite/libffi.call/nested_struct5.c,
1762         testsuite/libffi.call/nested_struct7.c,
1763         testsuite/libffi.call/cls_sint.c,
1764         testsuite/libffi.call/nested_struct9.c,
1765         testsuite/libffi.call/cls_20byte1.c,
1766         testsuite/libffi.call/cls_multi_sshortchar.c,
1767         testsuite/libffi.call/cls_align_sint64.c,
1768         testsuite/libffi.call/cls_3byte2.c,
1769         testsuite/libffi.call/cls_multi_schar.c,
1770         testsuite/libffi.call/cls_multi_uchar.c,
1771         testsuite/libffi.call/cls_19byte.c,
1772         testsuite/libffi.call/cls_9byte1.c,
1773         testsuite/libffi.call/cls_align_float.c,
1774         testsuite/libffi.call/closure_fn1.c,
1775         testsuite/libffi.call/problem1.c,
1776         testsuite/libffi.call/closure_fn3.c,
1777         testsuite/libffi.call/cls_sshort.c,
1778         testsuite/libffi.call/closure_fn5.c,
1779         testsuite/libffi.call/cls_align_double.c,
1780         testsuite/libffi.call/cls_2byte.c,
1781         testsuite/libffi.call/nested_struct.c,
1782         testsuite/libffi.call/nested_struct10.c,
1783         testsuite/libffi.call/cls_4byte.c,
1784         testsuite/libffi.call/cls_6byte.c,
1785         testsuite/libffi.call/cls_8byte.c,
1786         testsuite/libffi.call/cls_multi_sshort.c,
1787         testsuite/libffi.call/cls_align_uint32.c,
1788         testsuite/libffi.call/cls_align_sint16.c,
1789         testsuite/libffi.call/cls_float.c,
1790         testsuite/libffi.call/cls_20byte.c,
1791         testsuite/libffi.call/cls_5_1_byte.c,
1792         testsuite/libffi.call/nested_struct2.c,
1793         testsuite/libffi.call/cls_24byte.c,
1794         testsuite/libffi.call/nested_struct4.c,
1795         testsuite/libffi.call/nested_struct6.c,
1796         testsuite/libffi.call/cls_64byte.c,
1797         testsuite/libffi.call/nested_struct8.c,
1798         testsuite/libffi.call/cls_uint.c,
1799         testsuite/libffi.call/cls_multi_ushortchar.c,
1800         testsuite/libffi.call/cls_schar.c,
1801         testsuite/libffi.call/cls_uchar.c,
1802         testsuite/libffi.call/cls_align_uint64.c,
1803         testsuite/libffi.call/cls_ulonglong.c,
1804         testsuite/libffi.call/cls_align_longdouble.c,
1805         testsuite/libffi.call/cls_1_1byte.c,
1806         testsuite/libffi.call/cls_12byte.c,
1807         testsuite/libffi.call/cls_3_1byte.c,
1808         testsuite/libffi.call/cls_3byte1.c,
1809         testsuite/libffi.call/cls_4_1byte.c,
1810         testsuite/libffi.call/cls_6_1_byte.c,
1811         testsuite/libffi.call/cls_16byte.c,
1812         testsuite/libffi.call/cls_18byte.c,
1813         testsuite/libffi.call/closure_fn0.c,
1814         testsuite/libffi.call/cls_9byte2.c,
1815         testsuite/libffi.call/closure_fn2.c,
1816         testsuite/libffi.call/closure_fn4.c,
1817         testsuite/libffi.call/cls_ushort.c,
1818         testsuite/libffi.call/closure_fn6.c,
1819         testsuite/libffi.call/cls_5byte.c,
1820         testsuite/libffi.call/cls_align_pointer.c,
1821         testsuite/libffi.call/cls_7byte.c,
1822         testsuite/libffi.call/cls_align_sint32.c,
1823         testsuite/libffi.special/unwindtest_ffi_call.cc,
1824         testsuite/libffi.special/unwindtest.cc: Enable for ARM.
1825
1826 2007-07-05  H.J. Lu  <hongjiu.lu@intel.com>
1827
1828         * aclocal.m4: Regenerated.
1829
1830 2007-06-02  Paolo Bonzini  <bonzini@gnu.org>
1831
1832         * configure: Regenerate.
1833
1834 2007-05-23  Steve Ellcey  <sje@cup.hp.com>
1835
1836         * Makefile.in: Regenerate.
1837         * configure: Regenerate.
1838         * aclocal.m4: Regenerate.
1839         * include/Makefile.in: Regenerate.
1840         * testsuite/Makefile.in: Regenerate.
1841
1842 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1843
1844         * src/m68k/ffi.c (ffi_prep_incoming_args_SYSV,
1845         ffi_closure_SYSV_inner,ffi_prep_closure): New, add closure support.
1846         * src/m68k/sysv.S(ffi_closure_SYSV,ffi_closure_struct_SYSV): Likewise.
1847         * src/m68k/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1848         (FFI_CLOSURES): Enable closure support.
1849
1850 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1851
1852         * configure.ac (HAVE_AS_CFI_PSEUDO_OP): New test.
1853         * configure: Regenerate.
1854         * fficonfig.h.in: Regenerate.
1855         * src/m68k/sysv.S (CFI_STARTPROC,CFI_ENDPROC,
1856         CFI_OFFSET,CFI_DEF_CFA): New macros.
1857         (ffi_call_SYSV): Add callframe annotation.
1858
1859 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1860
1861         * src/m68k/ffi.c (ffi_prep_args,ffi_prep_cif_machdep): Fix
1862         numerous test suite failures.
1863         * src/m68k/sysv.S (ffi_call_SYSV): Likewise.
1864
1865 2007-04-11  Paolo Bonzini  <bonzini@gnu.org>
1866
1867         * Makefile.am (EXTRA_DIST): Bring up to date.
1868         * Makefile.in: Regenerate.
1869         * src/frv/eabi.S: Remove RCS keyword.
1870
1871 2007-04-06  Richard Henderson  <rth@redhat.com>
1872
1873         * configure.ac: Tidy target case.
1874         (HAVE_LONG_DOUBLE): Allow the target to override.
1875         * configure: Regenerate.
1876         * include/ffi.h.in: Don't define ffi_type_foo if
1877         LIBFFI_HIDE_BASIC_TYPES is defined.
1878         (ffi_type_longdouble): If not HAVE_LONG_DOUBLE, define
1879         to ffi_type_double.
1880         * types.c (LIBFFI_HIDE_BASIC_TYPES): Define.
1881         (FFI_TYPEDEF, ffi_type_void): Mark the data const.
1882         (ffi_type_longdouble): Special case for Alpha.  Don't define
1883         if long double == double.
1884
1885         * src/alpha/ffi.c (FFI_TYPE_LONGDOUBLE): Assert unique value.
1886         (ffi_prep_cif_machdep): Handle it as the 128-bit type.
1887         (ffi_call, ffi_closure_osf_inner): Likewise.
1888         (ffi_closure_osf_inner): Likewise.  Mark hidden.
1889         (ffi_call_osf, ffi_closure_osf): Mark hidden.
1890         * src/alpha/ffitarget.h (FFI_LAST_ABI): Tidy definition.
1891         * src/alpha/osf.S (ffi_call_osf, ffi_closure_osf): Mark hidden.
1892         (load_table): Handle 128-bit long double.
1893
1894         * testsuite/libffi.call/float4.c: Add -mieee for alpha.
1895
1896 2007-04-06  Tom Tromey  <tromey@redhat.com>
1897
1898         PR libffi/31491:
1899         * README: Fixed bug in example.
1900
1901 2007-04-03  Jakub Jelinek  <jakub@redhat.com>
1902
1903         * src/closures.c: Include sys/statfs.h.
1904         (_GNU_SOURCE): Define on Linux.
1905         (FFI_MMAP_EXEC_SELINUX): Define.
1906         (selinux_enabled): New variable.
1907         (selinux_enabled_check): New function.
1908         (is_selinux_enabled): Define.
1909         (dlmmap): Use it.
1910
1911 2007-03-24  Uros Bizjak  <ubizjak@gmail.com>
1912
1913         * testsuite/libffi.call/return_fl2.c (return_fl): Mark as static.
1914         Use 'volatile float sum' to create sum of floats to avoid false
1915         negative due to excess precision on ix86 targets.
1916         (main): Ditto.
1917
1918 2007-03-08  Alexandre Oliva  <aoliva@redhat.com>
1919
1920         * src/powerpc/ffi.c (flush_icache): Fix left-over from previous
1921         patch.
1922         (ffi_prep_closure_loc): Remove unneeded casts.  Add needed ones.
1923
1924 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1925
1926         * include/ffi.h.in (ffi_closure_alloc, ffi_closure_free): New.
1927         (ffi_prep_closure_loc): New.
1928         (ffi_prep_raw_closure_loc): New.
1929         (ffi_prep_java_raw_closure_loc): New.
1930         * src/closures.c: New file.
1931         * src/dlmalloc.c [FFI_MMAP_EXEC_WRIT] (struct malloc_segment):
1932         Replace sflags with exec_offset.
1933         [FFI_MMAP_EXEC_WRIT] (mmap_exec_offset, add_segment_exec_offset,
1934         sub_segment_exec_offset): New macros.
1935         (get_segment_flags, set_segment_flags, check_segment_merge): New
1936         macros.
1937         (is_mmapped_segment, is_extern_segment): Use get_segment_flags.
1938         (add_segment, sys_alloc, create_mspace, create_mspace_with_base,
1939         destroy_mspace): Use new macros.
1940         (sys_alloc): Silence warning.
1941         * Makefile.am (libffi_la_SOURCES): Add src/closures.c.
1942         * Makefile.in: Rebuilt.
1943         * src/prep_cif [FFI_CLOSURES] (ffi_prep_closure): Implement in
1944         terms of ffi_prep_closure_loc.
1945         * src/raw_api.c (ffi_prep_raw_closure_loc): Renamed and adjusted
1946         from...
1947         (ffi_prep_raw_closure): ... this.  Re-implement in terms of the
1948         renamed version.
1949         * src/java_raw_api (ffi_prep_java_raw_closure_loc): Renamed and
1950         adjusted from...
1951         (ffi_prep_java_raw_closure): ... this.  Re-implement in terms of
1952         the renamed version.
1953         * src/alpha/ffi.c (ffi_prep_closure_loc): Renamed from
1954         (ffi_prep_closure): ... this.
1955         * src/pa/ffi.c: Likewise.
1956         * src/cris/ffi.c: Likewise.  Adjust.
1957         * src/frv/ffi.c: Likewise.
1958         * src/ia64/ffi.c: Likewise.
1959         * src/mips/ffi.c: Likewise.
1960         * src/powerpc/ffi_darwin.c: Likewise.
1961         * src/s390/ffi.c: Likewise.
1962         * src/sh/ffi.c: Likewise.
1963         * src/sh64/ffi.c: Likewise.
1964         * src/sparc/ffi.c: Likewise.
1965         * src/x86/ffi64.c: Likewise.
1966         * src/x86/ffi.c: Likewise.
1967         (FFI_INIT_TRAMPOLINE): Adjust.
1968         (ffi_prep_raw_closure_loc): Renamed and adjusted from...
1969         (ffi_prep_raw_closure): ... this.
1970         * src/powerpc/ffi.c (ffi_prep_closure_loc): Renamed from
1971         (ffi_prep_closure): ... this.
1972         (flush_icache): Adjust.
1973
1974 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1975
1976         * src/dlmalloc.c: New file, imported version 2.8.3 of Doug
1977         Lea's malloc.
1978
1979 2007-03-01  Brooks Moses  <brooks.moses@codesourcery.com>
1980
1981         * Makefile.am: Add dummy install-pdf target.
1982         * Makefile.in: Regenerate
1983
1984 2007-02-13  Andreas Krebbel  <krebbel1@de.ibm.com>
1985
1986         * src/s390/ffi.c (ffi_prep_args, ffi_prep_cif_machdep,
1987         ffi_closure_helper_SYSV): Add long double handling.
1988
1989 2007-02-02  Jakub Jelinek  <jakub@redhat.com>
1990
1991         * src/powerpc/linux64.S (ffi_call_LINUX64): Move restore of r2
1992         immediately after bctrl instruction.
1993
1994 2007-01-18  Alexandre Oliva  <aoliva@redhat.com>
1995
1996         * Makefile.am (all-recursive, install-recursive,
1997         mostlyclean-recursive, clean-recursive, distclean-recursive,
1998         maintainer-clean-recursive): Add missing targets.
1999         * Makefile.in: Rebuilt.
2000
2001 2006-12-14  Andreas Tobler  <a.tobler@schweiz.org>
2002
2003         * configure.ac: Add TARGET for x86_64-*-darwin*.
2004         * Makefile.am (nodist_libffi_la_SOURCES): Add rules for 64-bit sources
2005         for X86_DARWIN.
2006         * src/x86/ffitarget.h: Set trampoline size for x86_64-*-darwin*.
2007         * src/x86/darwin64.S: New file for x86_64-*-darwin* support.
2008         * configure: Regenerate.
2009         * Makefile.in: Regenerate.
2010         * include/Makefile.in: Regenerate.
2011         * testsuite/Makefile.in: Regenerate.
2012         * testsuite/libffi.special/unwindtest_ffi_call.cc: New test case for
2013         ffi_call only.
2014
2015 2006-12-13  Andreas Tobler <a.tobler@schweiz.org>
2016
2017         * aclocal.m4: Regenerate with aclocal -I .. as written in the
2018         Makefile.am.
2019
2020 2006-10-31  Geoffrey Keating  <geoffk@apple.com>
2021
2022         * src/powerpc/ffi_darwin.c (darwin_adjust_aggregate_sizes): New.
2023         (ffi_prep_cif_machdep): Call darwin_adjust_aggregate_sizes for
2024         Darwin.
2025         * testsuite/libffi.call/nested_struct4.c: Remove Darwin XFAIL.
2026         * testsuite/libffi.call/nested_struct6.c: Remove Darwin XFAIL.
2027
2028 2006-10-10  Paolo Bonzini  <bonzini@gnu.org>
2029             Sandro Tolaini  <tolaini@libero.it>
2030
2031         * configure.ac [i*86-*-darwin*]: Set X86_DARWIN symbol and
2032         conditional.
2033         * configure: Regenerated.
2034         * Makefile.am (nodist_libffi_la_SOURCES) [X86_DARWIN]: New case.
2035         (EXTRA_DIST): Add src/x86/darwin.S.
2036         * Makefile.in: Regenerated.
2037         * include/Makefile.in: Regenerated.
2038         * testsuite/Makefile.in: Regenerated.
2039
2040         * src/x86/ffi.c (ffi_prep_cif_machdep) [X86_DARWIN]: Treat like
2041         X86_WIN32, and additionally align stack to 16 bytes.
2042         * src/x86/darwin.S: New, based on sysv.S.
2043         * src/prep_cif.c (ffi_prep_cif) [X86_DARWIN]: Align > 8-byte structs.
2044
2045 2006-09-12  David Daney  <ddaney@avtrex.com>
2046
2047         PR libffi/23935
2048         * include/Makefile.am: Install both ffi.h and ffitarget.h in
2049         $(libdir)/gcc/$(target_alias)/$(gcc_version)/include.
2050         * aclocal.m4: Regenerated for automake 1.9.6.
2051         * Makefile.in: Regenerated.
2052         * include/Makefile.in: Regenerated.
2053         * testsuite/Makefile.in: Regenerated.
2054
2055 2006-08-17  Andreas Tobler  <a.tobler@schweiz.ch>
2056
2057         * include/ffi_common.h (struct): Revert accidental commit.
2058
2059 2006-08-15  Andreas Tobler  <a.tobler@schweiz.ch>
2060
2061         * include/ffi_common.h: Remove lint directives.
2062         * include/ffi.h.in: Likewise.
2063
2064 2006-07-25  Torsten Schoenfeld  <kaffeetisch@gmx.de>
2065
2066         * include/ffi.h.in (ffi_type_ulong, ffi_type_slong): Define correctly
2067         for 32-bit architectures.
2068         * testsuite/libffi.call/return_ul.c: New test case.
2069
2070 2006-07-19  David Daney  <ddaney@avtrex.com>
2071
2072         * testsuite/libffi.call/closure_fn6.c: Remove xfail for mips,
2073         xfail remains for mips64.
2074
2075 2006-05-23  Carlos O'Donell  <carlos@codesourcery.com>
2076
2077         * Makefile.am: Add install-html target. Add install-html to .PHONY
2078         * Makefile.in: Regenerate.
2079         * aclocal.m4: Regenerate.
2080         * include/Makefile.in: Regenerate.
2081         * testsuite/Makefile.in: Regenerate.
2082
2083 2006-05-18  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2084
2085         * pa/ffi.c (ffi_prep_args_pa32): Load floating point arguments from
2086         stack slot.
2087
2088 2006-04-22  Andreas Tobler  <a.tobler@schweiz.ch>
2089
2090         * README: Remove notice about 'Crazy Comments'.
2091         * src/debug.c: Remove lint directives. Cleanup white spaces.
2092         * src/java_raw_api.c: Likewise.
2093         * src/prep_cif.c: Likewise.
2094         * src/raw_api.c: Likewise.
2095         * src/ffitest.c: Delete. No longer needed, all test cases migrated
2096         to the testsuite.
2097         * src/arm/ffi.c: Remove lint directives.
2098         * src/m32r/ffi.c: Likewise.
2099         * src/pa/ffi.c: Likewise.
2100         * src/powerpc/ffi.c: Likewise.
2101         * src/powerpc/ffi_darwin.c: Likewise.
2102         * src/sh/ffi.c: Likewise.
2103         * src/sh64/ffi.c: Likewise.
2104         * src/x86/ffi.c: Likewise.
2105         * testsuite/libffi.call/float2.c: Likewise.
2106         * testsuite/libffi.call/promotion.c: Likewise.
2107         * testsuite/libffi.call/struct1.c: Likewise.
2108
2109 2006-04-13  Andreas Tobler  <a.tobler@schweiz.ch>
2110
2111         * src/pa/hpux32.S: Correct unwind offset calculation for
2112         ffi_closure_pa32.
2113         * src/pa/linux.S: Likewise.
2114
2115 2006-04-12  James E Wilson  <wilson@specifix.com>
2116
2117         PR libgcj/26483
2118         * src/ia64/ffi.c (stf_spill, ldf_fill): Rewrite as macros.
2119         (hfa_type_load): Call stf_spill.
2120         (hfa_type_store): Call ldf_fill.
2121         (ffi_call): Adjust calls to above routines.  Add local temps for
2122         macro result.
2123
2124 2006-04-10  Matthias Klose  <doko@debian.org>
2125
2126         * testsuite/lib/libffi-dg.exp (libffi-init): Recognize multilib
2127         directory names containing underscores.
2128
2129 2006-04-07  James E Wilson  <wilson@specifix.com>
2130
2131         * testsuite/libffi.call/float4.c: New testcase.
2132
2133 2006-04-05  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2134             Andreas Tobler  <a.tobler@schweiz.ch>
2135
2136         * Makefile.am: Add PA_HPUX port.
2137         * Makefile.in: Regenerate.
2138         * include/Makefile.in: Likewise.
2139         * testsuite/Makefile.in: Likewise.
2140         * configure.ac: Add PA_HPUX rules.
2141         * configure: Regenerate.
2142         * src/pa/ffitarget.h: Rename linux target to PA_LINUX.
2143         Add PA_HPUX and PA64_HPUX.
2144         Rename FFI_LINUX ABI to FFI_PA32 ABI.
2145         (FFI_TRAMPOLINE_SIZE): Define for 32-bit HP-UX targets.
2146         (FFI_TYPE_SMALL_STRUCT2): Define.
2147         (FFI_TYPE_SMALL_STRUCT4): Likewise.
2148         (FFI_TYPE_SMALL_STRUCT8): Likewise.
2149         (FFI_TYPE_SMALL_STRUCT3): Redefine.
2150         (FFI_TYPE_SMALL_STRUCT5): Likewise.
2151         (FFI_TYPE_SMALL_STRUCT6): Likewise.
2152         (FFI_TYPE_SMALL_STRUCT7): Likewise.
2153         * src/pa/ffi.c (ROUND_DOWN): Delete.
2154         (fldw, fstw, fldd, fstd): Use '__asm__'.
2155         (ffi_struct_type): Add support for FFI_TYPE_SMALL_STRUCT2,
2156         FFI_TYPE_SMALL_STRUCT4 and FFI_TYPE_SMALL_STRUCT8.
2157         (ffi_prep_args_LINUX): Rename to ffi_prep_args_pa32. Update comment.
2158         Simplify incrementing of stack slot variable. Change type of local
2159         'n' to unsigned int.
2160         (ffi_size_stack_LINUX): Rename to ffi_size_stack_pa32. Handle long
2161         double on PA_HPUX.
2162         (ffi_prep_cif_machdep): Likewise.
2163         (ffi_call): Likewise.
2164         (ffi_closure_inner_LINUX): Rename to ffi_closure_inner_pa32. Change
2165         return type to ffi_status. Simplify incrementing of stack slot
2166         variable. Only copy floating point argument registers when PA_LINUX
2167         is true. Reformat debug statement.
2168         Add support for FFI_TYPE_SMALL_STRUCT2, FFI_TYPE_SMALL_STRUCT4 and
2169         FFI_TYPE_SMALL_STRUCT8.
2170         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Add 'extern' to
2171         declaration.
2172         (ffi_prep_closure): Make linux trampoline conditional on PA_LINUX.
2173         Add nops to cache flush.  Add trampoline for PA_HPUX.
2174         * src/pa/hpux32.S: New file.
2175         * src/pa/linux.S (ffi_call_LINUX): Rename to ffi_call_pa32. Rename
2176         ffi_prep_args_LINUX to ffi_prep_args_pa32.
2177         Localize labels. Add support for 2, 4 and 8-byte small structs. Handle
2178         unaligned destinations in 3, 5, 6 and 7-byte small structs. Order
2179         argument type checks so that common argument types appear first.
2180         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Rename
2181         ffi_closure_inner_LINUX to ffi_closure_inner_pa32.
2182
2183 2006-03-24  Alan Modra  <amodra@bigpond.net.au>
2184
2185         * src/powerpc/ffitarget.h (enum ffi_abi): Add FFI_LINUX.  Default
2186         for 32-bit using IBM extended double format.  Fix FFI_LAST_ABI.
2187         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Handle linux variant of
2188         FFI_TYPE_LONGDOUBLE.
2189         (ffi_prep_args64): Assert using IBM extended double.
2190         (ffi_prep_cif_machdep): Don't munge FFI_TYPE_LONGDOUBLE type.
2191         Handle FFI_LINUX FFI_TYPE_LONGDOUBLE return and args.
2192         (ffi_call): Handle FFI_LINUX.
2193         (ffi_closure_helper_SYSV): Non FFI_LINUX long double return needs
2194         gpr3 return pointer as for struct return.  Handle FFI_LINUX
2195         FFI_TYPE_LONGDOUBLE return and args.  Don't increment "nf"
2196         unnecessarily.
2197         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Load both f1 and f2
2198         for FFI_TYPE_LONGDOUBLE.  Move epilogue insns into case table.
2199         Don't use r6 as pointer to results, instead use sp offset.  Don't
2200         make a special call to load lr with case table address, instead
2201         use offset from previous call.
2202         * src/powerpc/sysv.S (ffi_call_SYSV): Save long double return.
2203         * src/powerpc/linux64.S (ffi_call_LINUX64): Simplify long double
2204         return.
2205
2206 2006-03-15  Kaz Kojima  <kkojima@gcc.gnu.org>
2207
2208         * src/sh64/ffi.c (ffi_prep_cif_machdep): Handle float arguments
2209         passed with FP registers correctly.
2210         (ffi_closure_helper_SYSV): Likewise.
2211         * src/sh64/sysv.S: Likewise.
2212
2213 2006-03-01  Andreas Tobler  <a.tobler@schweiz.ch>
2214
2215         * testsuite/libffi.special/unwindtest.cc (closure_test_fn): Mark cif,
2216         args and userdata unused.
2217         (closure_test_fn1): Mark cif and userdata unused.
2218         (main): Remove unused res.
2219
2220 2006-02-28  Andreas Tobler  <a.tobler@schweiz.ch>
2221
2222         * testsuite/libffi.call/call.exp: Adjust FSF address. Add test runs for
2223         -O2, -O3, -Os and the warning flags -W -Wall.
2224         * testsuite/libffi.special/special.exp: Likewise.
2225         * testsuite/libffi.call/ffitest.h: Add an __UNUSED__ macro to mark
2226         unused parameter unused for gcc or else do nothing.
2227         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2228         * testsuite/libffi.call/cls_12byte.c (cls_struct_12byte_gn): Mark cif
2229         and userdata unused.
2230         * testsuite/libffi.call/cls_16byte.c (cls_struct_16byte_gn): Likewise.
2231         * testsuite/libffi.call/cls_18byte.c (cls_struct_18byte_gn): Likewise.
2232         * testsuite/libffi.call/cls_19byte.c (cls_struct_19byte_gn): Likewise.
2233         * testsuite/libffi.call/cls_1_1byte.c (cls_struct_1_1byte_gn): Likewise.
2234         * testsuite/libffi.call/cls_20byte.c (cls_struct_20byte_gn): Likewise.
2235         * testsuite/libffi.call/cls_20byte1.c (cls_struct_20byte_gn): Likewise.
2236         * testsuite/libffi.call/cls_24byte.c (cls_struct_24byte_gn): Likewise.
2237         * testsuite/libffi.call/cls_2byte.c (cls_struct_2byte_gn): Likewise.
2238         * testsuite/libffi.call/cls_3_1byte.c (cls_struct_3_1byte_gn): Likewise.
2239         * testsuite/libffi.call/cls_3byte1.c (cls_struct_3byte_gn): Likewise.
2240         * testsuite/libffi.call/cls_3byte2.c (cls_struct_3byte_gn1): Likewise.
2241         * testsuite/libffi.call/cls_4_1byte.c (cls_struct_4_1byte_gn): Likewise.
2242         * testsuite/libffi.call/cls_4byte.c (cls_struct_4byte_gn): Likewise.
2243         * testsuite/libffi.call/cls_5_1_byte.c (cls_struct_5byte_gn): Likewise.
2244         * testsuite/libffi.call/cls_5byte.c (cls_struct_5byte_gn): Likewise.
2245         * testsuite/libffi.call/cls_64byte.c (cls_struct_64byte_gn): Likewise.
2246         * testsuite/libffi.call/cls_6_1_byte.c (cls_struct_6byte_gn): Likewise.
2247         * testsuite/libffi.call/cls_6byte.c (cls_struct_6byte_gn): Likewise.
2248         * testsuite/libffi.call/cls_7_1_byte.c (cls_struct_7byte_gn): Likewise.
2249         * testsuite/libffi.call/cls_7byte.c (cls_struct_7byte_gn): Likewise.
2250         * testsuite/libffi.call/cls_8byte.c (cls_struct_8byte_gn): Likewise.
2251         * testsuite/libffi.call/cls_9byte1.c (cls_struct_9byte_gn): Likewise.
2252         * testsuite/libffi.call/cls_9byte2.c (cls_struct_9byte_gn): Likewise.
2253         * testsuite/libffi.call/cls_align_double.c (cls_struct_align_gn):
2254         Likewise.
2255         * testsuite/libffi.call/cls_align_float.c (cls_struct_align_gn):
2256         Likewise.
2257         * testsuite/libffi.call/cls_align_longdouble.c (cls_struct_align_gn):
2258         Likewise.
2259         * testsuite/libffi.call/cls_align_pointer.c (cls_struct_align_fn): Cast
2260         void* to avoid compiler warning.
2261         (main): Likewise.
2262         (cls_struct_align_gn): Mark cif and userdata unused.
2263         * testsuite/libffi.call/cls_align_sint16.c (cls_struct_align_gn):
2264         Likewise.
2265         * testsuite/libffi.call/cls_align_sint32.c (cls_struct_align_gn):
2266         Likewise.
2267         * testsuite/libffi.call/cls_align_sint64.c (cls_struct_align_gn):
2268         Likewise.
2269         * testsuite/libffi.call/cls_align_uint16.c (cls_struct_align_gn):
2270         Likewise.
2271         * testsuite/libffi.call/cls_align_uint32.c (cls_struct_align_gn):
2272         Likewise.
2273         * testsuite/libffi.call/cls_double.c (cls_ret_double_fn): Likewise.
2274         * testsuite/libffi.call/cls_float.c (cls_ret_float_fn): Likewise.
2275         * testsuite/libffi.call/cls_multi_schar.c (test_func_gn): Mark cif and
2276         data unused.
2277         (main): Cast res_call to silence gcc.
2278         * testsuite/libffi.call/cls_multi_sshort.c (test_func_gn): Mark cif and
2279         data unused.
2280         (main): Cast res_call to silence gcc.
2281         * testsuite/libffi.call/cls_multi_sshortchar.c (test_func_gn): Mark cif
2282         and data unused.
2283         (main): Cast res_call to silence gcc.
2284         * testsuite/libffi.call/cls_multi_uchar.c (test_func_gn): Mark cif and
2285         data unused.
2286         (main): Cast res_call to silence gcc.
2287         * testsuite/libffi.call/cls_multi_ushort.c (test_func_gn): Mark cif and
2288         data unused.
2289         (main): Cast res_call to silence gcc.
2290         * testsuite/libffi.call/cls_multi_ushortchar.c (test_func_gn): Mark cif
2291         and data unused.
2292         (main): Cast res_call to silence gcc.
2293         * testsuite/libffi.call/cls_schar.c (cls_ret_schar_fn): Mark cif and
2294         userdata unused.
2295         (cls_ret_schar_fn): Cast printf parameter to silence gcc.
2296         * testsuite/libffi.call/cls_sint.c (cls_ret_sint_fn): Mark cif and
2297         userdata unused.
2298         (cls_ret_sint_fn): Cast printf parameter to silence gcc.
2299         * testsuite/libffi.call/cls_sshort.c (cls_ret_sshort_fn): Mark cif and
2300         userdata unused.
2301         (cls_ret_sshort_fn): Cast printf parameter to silence gcc.
2302         * testsuite/libffi.call/cls_uchar.c (cls_ret_uchar_fn):  Mark cif and
2303         userdata unused.
2304         (cls_ret_uchar_fn): Cast printf parameter to silence gcc.
2305         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Mark cif and
2306         userdata unused.
2307         (cls_ret_uint_fn): Cast printf parameter to silence gcc.
2308         * testsuite/libffi.call/cls_ulonglong.c (cls_ret_ulonglong_fn): Mark cif
2309         and userdata unused.
2310         * testsuite/libffi.call/cls_ushort.c (cls_ret_ushort_fn): Mark cif and
2311         userdata unused.
2312         (cls_ret_ushort_fn): Cast printf parameter to silence gcc.
2313         * testsuite/libffi.call/float.c (floating): Remove unused parameter e.
2314         * testsuite/libffi.call/float1.c (main): Remove unused variable i.
2315         Cleanup white spaces.
2316         * testsuite/libffi.call/negint.c (checking): Remove unused variable i.
2317         * testsuite/libffi.call/nested_struct.c (cls_struct_combined_gn): Mark
2318         cif and userdata unused.
2319         * testsuite/libffi.call/nested_struct1.c (cls_struct_combined_gn):
2320         Likewise.
2321         * testsuite/libffi.call/nested_struct10.c (B_gn): Likewise.
2322         * testsuite/libffi.call/nested_struct2.c (B_fn): Adjust printf
2323         formatters to silence gcc.
2324         (B_gn): Mark cif and userdata unused.
2325         * testsuite/libffi.call/nested_struct3.c (B_gn): Mark cif and userdata
2326         unused.
2327         * testsuite/libffi.call/nested_struct4.c: Mention related PR.
2328         (B_gn): Mark cif and userdata unused.
2329         * testsuite/libffi.call/nested_struct5.c (B_gn): Mark cif and userdata
2330         unused.
2331         * testsuite/libffi.call/nested_struct6.c: Mention related PR.
2332         (B_gn): Mark cif and userdata unused.
2333         * testsuite/libffi.call/nested_struct7.c (B_gn): Mark cif and userdata
2334         unused.
2335         * testsuite/libffi.call/nested_struct8.c (B_gn): Likewise.
2336         * testsuite/libffi.call/nested_struct9.c (B_gn): Likewise.
2337         * testsuite/libffi.call/problem1.c (stub): Likewise.
2338         * testsuite/libffi.call/pyobjc-tc.c (main): Cast the result to silence
2339         gcc.
2340         * testsuite/libffi.call/return_fl2.c (return_fl): Add the note mentioned
2341         in the last commit for this test case in the test case itself.
2342         * testsuite/libffi.call/closure_fn0.c (closure_test_fn0): Mark cif as
2343         unused.
2344         * testsuite/libffi.call/closure_fn1.c (closure_test_fn1): Likewise.
2345         * testsuite/libffi.call/closure_fn2.c (closure_test_fn2): Likewise.
2346         * testsuite/libffi.call/closure_fn3.c (closure_test_fn3): Likewise.
2347         * testsuite/libffi.call/closure_fn4.c (closure_test_fn0): Likewise.
2348         * testsuite/libffi.call/closure_fn5.c (closure_test_fn5): Likewise.
2349         * testsuite/libffi.call/closure_fn6.c (closure_test_fn0): Likewise.
2350
2351 2006-02-22  Kaz Kojima  <kkojima@gcc.gnu.org>
2352
2353         * src/sh/sysv.S: Fix register numbers in the FDE for
2354         ffi_closure_SYSV.
2355
2356 2006-02-20  Andreas Tobler  <a.tobler@schweiz.ch>
2357
2358         * testsuite/libffi.call/return_fl2.c (return_fl): Remove static
2359         declaration to avoid a false negative on ix86. See PR323.
2360
2361 2006-02-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2362
2363         * src/sh/ffi.c (ffi_closure_helper_SYSV): Remove unused variable
2364         and cast integer to void * if needed.  Update the pointer to
2365         the FP register saved area correctly.
2366
2367 2006-02-17  Andreas Tobler  <a.tobler@schweiz.ch>
2368
2369         * testsuite/libffi.call/nested_struct6.c: XFAIL this test until PR25630
2370         is fixed.
2371         * testsuite/libffi.call/nested_struct4.c: Likewise.
2372
2373 2006-02-16  Andreas Tobler  <a.tobler@schweiz.ch>
2374
2375         * testsuite/libffi.call/return_dbl.c: New test case.
2376         * testsuite/libffi.call/return_dbl1.c: Likewise.
2377         * testsuite/libffi.call/return_dbl2.c: Likewise.
2378         * testsuite/libffi.call/return_fl.c: Likewise.
2379         * testsuite/libffi.call/return_fl1.c: Likewise.
2380         * testsuite/libffi.call/return_fl2.c: Likewise.
2381         * testsuite/libffi.call/return_fl3.c: Likewise.
2382         * testsuite/libffi.call/closure_fn6.c: Likewise.
2383
2384         * testsuite/libffi.call/nested_struct2.c: Remove ffi_type_mylong
2385         definition.
2386         * testsuite/libffi.call/ffitest.h: Add ffi_type_mylong definition
2387         here to be used by other test cases too.
2388
2389         * testsuite/libffi.call/nested_struct10.c: New test case.
2390         * testsuite/libffi.call/nested_struct9.c: Likewise.
2391         * testsuite/libffi.call/nested_struct8.c: Likewise.
2392         * testsuite/libffi.call/nested_struct7.c: Likewise.
2393         * testsuite/libffi.call/nested_struct6.c: Likewise.
2394         * testsuite/libffi.call/nested_struct5.c: Likewise.
2395         * testsuite/libffi.call/nested_struct4.c: Likewise.
2396
2397 2006-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
2398
2399         * configure.ac: Enable libffi for sparc64-*-freebsd*.
2400         * configure: Rebuilt.
2401
2402 2006-01-18  Jakub Jelinek  <jakub@redhat.com>
2403
2404         * src/powerpc/sysv.S (smst_two_register): Don't call __ashldi3,
2405         instead do the shifting inline.
2406         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't compute %r5
2407         shift count unconditionally.  Simplify load sequences for 1, 2, 3, 4
2408         and 8 byte structs, for the remaining struct sizes don't call
2409         __lshrdi3, instead do the shifting inline.
2410
2411 2005-12-07  Thiemo Seufer  <ths@networkno.de>
2412
2413         * src/mips/ffitarget.h: Remove obsolete sgidefs.h include. Add
2414         missing parentheses.
2415         * src/mips/o32.S (ffi_call_O32): Code formatting. Define
2416         and use A3_OFF, FP_OFF, RA_OFF. Micro-optimizations.
2417         (ffi_closure_O32): Likewise, but with newly defined A3_OFF2,
2418         A2_OFF2, A1_OFF2, A0_OFF2, RA_OFF2, FP_OFF2, S0_OFF2, GP_OFF2,
2419         V1_OFF2, V0_OFF2, FA_1_1_OFF2, FA_1_0_OFF2, FA_0_1_OFF2,
2420         FA_0_0_OFF2.
2421         * src/mips/ffi.c (ffi_prep_args): Code formatting. Fix
2422         endianness bugs.
2423         (ffi_prep_closure): Improve trampoline instruction scheduling.
2424         (ffi_closure_mips_inner_O32): Fix endianness bugs.
2425
2426 2005-12-03  Alan Modra  <amodra@bigpond.net.au>
2427
2428         * src/powerpc/ffi.c: Formatting.
2429         (ffi_prep_args_SYSV): Avoid possible aliasing problems by using unions.
2430         (ffi_prep_args64): Likewise.
2431
2432 2005-09-30  Geoffrey Keating  <geoffk@apple.com>
2433
2434         * testsuite/lib/libffi-dg.exp (libffi_target_compile): For
2435         darwin, use -shared-libgcc not -lgcc_s, and explain why.
2436
2437 2005-09-26  Tom Tromey  <tromey@redhat.com>
2438
2439         * testsuite/libffi.call/float1.c (value_type): New typedef.
2440         (CANARY): New define.
2441         (main): Check for result buffer overflow.
2442         * src/powerpc/linux64.S: Handle linux64 long double returns.
2443         * src/powerpc/ffi.c (FLAG_RETURNS_128BITS): New constant.
2444         (ffi_prep_cif_machdep): Handle linux64 long double returns.
2445
2446 2005-08-25  Alan Modra  <amodra@bigpond.net.au>
2447
2448         PR target/23404
2449         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Correct placement of stack
2450         homed fp args.
2451         (ffi_status ffi_prep_cif_machdep): Correct stack sizing for same.
2452
2453 2005-08-11  Jakub Jelinek  <jakub@redhat.com>
2454
2455         * configure.ac (HAVE_HIDDEN_VISIBILITY_ATTRIBUTE): New test.
2456         (AH_BOTTOM): Add FFI_HIDDEN definition.
2457         * configure: Rebuilt.
2458         * fficonfig.h.in: Rebuilt.
2459         * src/powerpc/ffi.c (hidden): Remove.
2460         (ffi_closure_LINUX64, ffi_prep_args64, ffi_call_LINUX64,
2461         ffi_closure_helper_LINUX64): Use FFI_HIDDEN instead of hidden.
2462         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64,
2463         .ffi_closure_LINUX64): Use FFI_HIDDEN instead of .hidden.
2464         * src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV): Remove,
2465         add FFI_HIDDEN to its prototype.
2466         (ffi_closure_SYSV_inner): New.
2467         * src/x86/sysv.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2468         * src/x86/win32.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2469
2470 2005-08-10  Alfred M. Szmidt  <ams@gnu.org>
2471
2472         PR libffi/21819:
2473         * configure: Rebuilt.
2474         * configure.ac: Handle i*86-*-gnu*.
2475
2476 2005-08-09  Jakub Jelinek  <jakub@redhat.com>
2477
2478         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Use
2479         DW_CFA_offset_extended_sf rather than
2480         DW_CFA_GNU_negative_offset_extended.
2481         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise.
2482
2483 2005-07-22  SUGIOKA Toshinobu  <sugioka@itonet.co.jp>
2484
2485         * src/sh/sysv.S (ffi_call_SYSV): Stop argument popping correctly
2486         on sh3.
2487         (ffi_closure_SYSV): Change the stack layout for sh3 struct argument.
2488         * src/sh/ffi.c (ffi_prep_args): Fix sh3 argument copy, when it is
2489         partially on register.
2490         (ffi_closure_helper_SYSV): Likewise.
2491         (ffi_prep_cif_machdep): Don't set too many cif->flags.
2492
2493 2005-07-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2494
2495         * src/sh/ffi.c (ffi_call): Handle small structures correctly.
2496         Remove empty line.
2497         * src/sh64/ffi.c (simple_type): Remove.
2498         (return_type): Handle small structures correctly.
2499         (ffi_prep_args): Likewise.
2500         (ffi_call): Likewise.
2501         (ffi_closure_helper_SYSV): Likewise.
2502         * src/sh64/sysv.S (ffi_call_SYSV): Handle 1, 2 and 4-byte return.
2503         Emit position independent code if PIC and remove wrong datalabel
2504         prefixes from EH data.
2505
2506 2005-07-19  Andreas Tobler  <a.tobler@schweiz.ch>
2507
2508         * Makefile.am (nodist_libffi_la_SOURCES): Add POWERPC_FREEBSD.
2509         * Makefile.in: Regenerate.
2510         * include/Makefile.in: Likewise.
2511         * testsuite/Makefile.in: Likewise.
2512         * configure.ac: Add POWERPC_FREEBSD rules.
2513         * configure: Regenerate.
2514         * src/powerpc/ffitarget.h: Add POWERPC_FREEBSD rules.
2515         (FFI_SYSV_TYPE_SMALL_STRUCT): Define.
2516         * src/powerpc/ffi.c: Add flags to handle small structure returns
2517         in ffi_call_SYSV.
2518         (ffi_prep_cif_machdep): Handle small structures for SYSV 4 ABI.
2519         Aka FFI_SYSV.
2520         (ffi_closure_helper_SYSV): Likewise.
2521         * src/powerpc/ppc_closure.S: Add return types for small structures.
2522         * src/powerpc/sysv.S: Add bits to handle small structures for
2523         final SYSV 4 ABI.
2524
2525 2005-07-10  Andreas Tobler  <a.tobler@schweiz.ch>
2526
2527         * testsuite/libffi.call/cls_5_1_byte.c: New test file.
2528         * testsuite/libffi.call/cls_6_1_byte.c: Likewise.
2529         * testsuite/libffi.call/cls_7_1_byte.c: Likewise.
2530
2531 2005-07-05  Randolph Chung  <tausq@debian.org>
2532
2533         * src/pa/ffi.c (ffi_struct_type): Rename FFI_TYPE_SMALL_STRUCT1
2534         as FFI_TYPE_SMALL_STRUCT3.  Break out handling for 5-7 byte
2535         structures.  Kill compilation warnings.
2536         (ffi_closure_inner_LINUX): Print return values as hex in debug
2537         message.  Rename FFI_TYPE_SMALL_STRUCT1 as FFI_TYPE_SMALL_STRUCT3.
2538         Properly handle 5-7 byte structure returns.
2539         * src/pa/ffitarget.h (FFI_TYPE_SMALL_STRUCT1)
2540         (FFI_TYPE_SMALL_STRUCT2): Remove.
2541         (FFI_TYPE_SMALL_STRUCT3, FFI_TYPE_SMALL_STRUCT5)
2542         (FFI_TYPE_SMALL_STRUCT6, FFI_TYPE_SMALL_STRUCT7): Define.
2543         * src/pa/linux.S: Mark source file as using PA1.1 assembly.
2544         (checksmst1, checksmst2): Remove.
2545         (checksmst3): Optimize handling of 3-byte struct returns.
2546         (checksmst567): Properly handle 5-7 byte struct returns.
2547
2548 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2549
2550         PR libgcj/21943
2551         * src/mips/n32.S: Enforce PIC code.
2552         * src/mips/o32.S: Likewise.
2553
2554 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2555
2556         * configure.ac: Treat i*86-*-solaris2.10 and up as X86_64.
2557         * configure: Regenerate.
2558
2559 2005-06-01  Alan Modra  <amodra@bigpond.net.au>
2560
2561         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't use JUMPTARGET
2562         to call ffi_closure_helper_SYSV.  Append @local instead.
2563         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise for ffi_prep_args_SYSV.
2564
2565 2005-05-17  Kelley Cook  <kcook@gcc.gnu.org>
2566
2567         * configure.ac: Use AC_C_BIGENDIAN instead of AC_C_BIGENDIAN_CROSS.
2568         Use AC_CHECK_SIZEOF instead of AC_COMPILE_CHECK_SIZEOF.
2569         * Makefile.am (ACLOCAL_AMFLAGS): Remove -I ../config.
2570         * aclocal.m4, configure, fficonfig.h.in, Makefile.in,
2571         include/Makefile.in, testsuite/Makefile.in: Regenerate.
2572
2573 2005-05-09  Mike Stump  <mrs@apple.com>
2574
2575         * configure: Regenerate.
2576
2577 2005-05-08  Richard Henderson  <rth@redhat.com>
2578
2579         PR libffi/21285
2580         * src/alpha/osf.S: Update unwind into to match code.
2581
2582 2005-05-04  Andreas Degert <ad@papyrus-gmbh.de>
2583             Richard Henderson  <rth@redhat.com>
2584
2585         * src/x86/ffi64.c (ffi_prep_cif_machdep): Save sse-used flag in
2586         bit 11 of flags.
2587         (ffi_call): Mask return type field.  Pass ssecount to ffi_call_unix64.
2588         (ffi_prep_closure): Set carry bit if sse-used flag set.
2589         * src/x86/unix64.S (ffi_call_unix64): Add ssecount argument.
2590         Only load sse registers if ssecount non-zero.
2591         (ffi_closure_unix64): Only save sse registers if carry set on entry.
2592
2593 2005-04-29  Ralf Corsepius  <ralf.corsepius@rtems.org>
2594
2595         * configure.ac: Add i*86-*-rtems*, sparc*-*-rtems*,
2596         powerpc-*rtems*, arm*-*-rtems*, sh-*-rtems*.
2597         * configure: Regenerate.
2598
2599 2005-04-20  Hans-Peter Nilsson  <hp@axis.com>
2600
2601         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): In regsub use,
2602         have Tcl8.3-compatible intermediate variable.
2603
2604 2005-04-18  Simon Posnjak <simon.posnjak@siol.net>
2605             Hans-Peter Nilsson  <hp@axis.com>
2606
2607         * Makefile.am: Add CRIS support.
2608         * configure.ac: Likewise.
2609         * Makefile.in, configure, testsuite/Makefile.in,
2610         include/Makefile.in: Regenerate.
2611         * src/cris: New directory.
2612         * src/cris/ffi.c, src/cris/sysv.S, src/cris/ffitarget.h: New files.
2613         * src/prep_cif.c (ffi_prep_cif): Wrap in #ifndef __CRIS__.
2614
2615         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): Replace \n with
2616         \r?\n in output tests.
2617
2618 2005-04-12  Mike Stump  <mrs@apple.com>
2619
2620         * configure: Regenerate.
2621
2622 2005-03-30  Hans Boehm  <Hans.Boehm@hp.com>
2623
2624         * src/ia64/ffitarget.h (ffi_arg): Use long long instead of DI.
2625
2626 2005-03-30  Steve Ellcey  <sje@cup.hp.com>
2627
2628         * src/ia64/ffitarget.h (ffi_arg) ADD DI attribute.
2629         (ffi_sarg) Ditto.
2630         * src/ia64/unix.S (ffi_closure_unix): Extend gp
2631         to 64 bits in ILP32 mode.
2632         Load 64 bits even for short data.
2633
2634 2005-03-23  Mike Stump  <mrs@apple.com>
2635
2636         * src/powerpc/darwin.S: Update for -m64 multilib.
2637         * src/powerpc/darwin_closure.S: Likewise.
2638
2639 2005-03-21  Zack Weinberg  <zack@codesourcery.com>
2640
2641         * configure.ac: Do not invoke TL_AC_GCC_VERSION.
2642         Do not set tool_include_dir.
2643         * aclocal.m4, configure, Makefile.in, testsuite/Makefile.in:
2644         Regenerate.
2645         * include/Makefile.am: Set gcc_version and toollibffidir.
2646         * include/Makefile.in: Regenerate.
2647
2648 2005-02-22  Andrew Haley  <aph@redhat.com>
2649
2650         * src/powerpc/ffi.c (ffi_prep_cif_machdep): Bump alignment to
2651         odd-numbered register pairs for 64-bit integer types.
2652
2653 2005-02-23  Andreas Tobler  <a.tobler@schweiz.ch>
2654
2655         PR libffi/20104
2656         * testsuite/libffi.call/return_ll1.c: New test case.
2657
2658 2005-02-11  Janis Johnson  <janis187@us.ibm.com>
2659
2660         * testsuite/libffi.call/cls_align_longdouble.c: Remove dg-options.
2661         * testsuite/libffi.call/float.c: Ditto.
2662         * testsuite/libffi.call/float2.c: Ditto.
2663         * testsuite/libffi.call/float3.c: Ditto.
2664
2665 2005-02-08  Andreas Tobler  <a.tobler@schweiz.ch>
2666
2667         * src/frv/ffitarget.h: Remove PPC stuff which does not belong to frv.
2668
2669 2005-01-12  Eric Botcazou  <ebotcazou@libertysurf.fr>
2670
2671         * testsuite/libffi.special/special.exp (cxx_options): Add
2672         -shared-libgcc.
2673
2674 2004-12-31  Richard Henderson  <rth@redhat.com>
2675
2676         * src/types.c (FFI_AGGREGATE_TYPEDEF): Remove.
2677         (FFI_TYPEDEF): Rename from FFI_INTEGRAL_TYPEDEF.  Replace size and
2678         offset parameters with a type parameter; deduce size and structure
2679         alignment.  Update all users.
2680
2681 2004-12-31  Richard Henderson  <rth@redhat.com>
2682
2683         * src/types.c (FFI_TYPE_POINTER): Define with sizeof.
2684         (FFI_TYPE_LONGDOUBLE): Fix for ia64.
2685         * src/ia64/ffitarget.h (struct ffi_ia64_trampoline_struct): Move
2686         into ffi_prep_closure.
2687         * src/ia64/ia64_flags.h, src/ia64/ffi.c, src/ia64/unix.S: Rewrite
2688         from scratch.
2689
2690 2004-12-27  Richard Henderson  <rth@redhat.com>
2691
2692         * src/x86/unix64.S: Fix typo in unwind info.
2693
2694 2004-12-25  Richard Henderson  <rth@redhat.com>
2695
2696         * src/x86/ffi64.c (struct register_args): Rename from stackLayout.
2697         (enum x86_64_reg_class): Add X86_64_COMPLEX_X87_CLASS.
2698         (merge_classes): Check for it.
2699         (SSE_CLASS_P): New.
2700         (classify_argument): Pass byte_offset by value; perform all updates
2701         inside struct case.
2702         (examine_argument): Add classes argument; handle
2703         X86_64_COMPLEX_X87_CLASS.
2704         (ffi_prep_args): Merge into ...
2705         (ffi_call): ... here.  Share stack frame with ffi_call_unix64.
2706         (ffi_prep_cif_machdep): Setup cif->flags for proper structure return.
2707         (ffi_fill_return_value): Remove.
2708         (ffi_prep_closure): Remove dead assert.
2709         (ffi_closure_unix64_inner): Rename from ffi_closure_UNIX64_inner.
2710         Rewrite to use struct register_args instead of va_list.  Create
2711         flags for handling structure returns.
2712         * src/x86/unix64.S: Remove dead strings.
2713         (ffi_call_unix64): Rename from ffi_call_UNIX64.  Rewrite to share
2714         stack frame with ffi_call.  Handle structure returns properly.
2715         (float2sse, floatfloat2sse, double2sse): Remove.
2716         (sse2float, sse2double, sse2floatfloat): Remove.
2717         (ffi_closure_unix64): Rename from ffi_closure_UNIX64.  Rewrite
2718         to handle structure returns properly.
2719
2720 2004-12-08  David Edelsohn  <edelsohn@gnu.org>
2721
2722         * Makefile.am (AM_MAKEFLAGS): Remove duplicate LIBCFLAGS and
2723         PICFLAG.
2724         * Makefile.in: Regenerated.
2725
2726 2004-12-02  Richard Sandiford  <rsandifo@redhat.com>
2727
2728         * configure.ac: Use TL_AC_GCC_VERSION to set gcc_version.
2729         * configure, aclocal.m4, Makefile.in: Regenerate.
2730         * include/Makefile.in, testsuite/Makefile.in: Regenerate.
2731
2732 2004-11-29  Kelley Cook  <kcook@gcc.gnu.org>
2733
2734         * configure: Regenerate for libtool change.
2735
2736 2004-11-25  Kelley Cook  <kcook@gcc.gnu.org>
2737
2738         * configure: Regenerate for libtool reversion.
2739
2740 2004-11-24  Kelley Cook  <kcook@gcc.gnu.org>
2741
2742         * configure: Regenerate for libtool change.
2743
2744 2004-11-23  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2745
2746         * testsuite/lib/libffi-dg.exp: Use new procs in target-libpath.exp.
2747
2748 2004-11-23  Richard Sandiford  <rsandifo@redhat.com>
2749
2750         * src/mips/o32.S (ffi_call_O32, ffi_closure_O32): Use jalr instead
2751         of jal.  Use an absolute encoding for the frame information.
2752
2753 2004-11-23  Kelley Cook  <kcook@gcc.gnu.org>
2754
2755         * Makefile.am: Remove no-dependencies.  Add ACLOCAL_AMFLAGS.
2756         * acinclude.m4: Delete logic for sincludes.
2757         * aclocal.m4, Makefile.in, configure: Regenerate.
2758         * include/Makefile: Likewise.
2759         * testsuite/Makefile: Likewise.
2760
2761 2004-11-22  Eric Botcazou  <ebotcazou@libertysurf.fr>
2762
2763         * src/sparc/ffi.c (ffi_prep_closure): Align doubles and 64-bit integers
2764         on a 8-byte boundary.
2765         * src/sparc/v8.S (ffi_closure_v8): Reserve frame space for arguments.
2766
2767 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2768
2769         * src/arm/ffi.c (ffi_prep_cif_machdep): Handle functions that return
2770         long long values.  Round stack allocation to a multiple of 8 bytes
2771         for ATPCS compatibility.
2772         * src/arm/sysv.S (ffi_call_SYSV): Rework to avoid use of APCS register
2773         names.  Handle returning long long types.  Add Thumb and interworking
2774         support.  Improve soft-float code.
2775
2776 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2777
2778         * testsuite/lib/libffi-db.exp (load_gcc_lib): New function.
2779         (libffi_exit): New function.
2780         (libffi_init): Build the testglue wrapper if needed.
2781
2782 2004-10-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2783
2784         PR other/18138
2785         * testsuite/lib/libffi-dg.exp: Accept more than one multilib libgcc.
2786
2787 2004-10-25  Kazuhiro Inaoka <inaoka.kazuhiro@renesas.com>
2788
2789         * src/m32r/libffitarget.h (FFI_CLOSURES): Set to 0.
2790
2791 2004-10-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2792
2793         * src/sh/sysv.S (ffi_call_SYSV): Don't align for double data.
2794         * testsuite/libffi.call/float3.c: New test case.
2795
2796 2004-10-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2797
2798         * src/sh/ffi.c (ffi_prep_closure): Set T bit in trampoline for
2799         the function returning a structure pointed with R2.
2800         * src/sh/sysv.S (ffi_closure_SYSV): Use R2 as the pointer to
2801         the structure return value if T bit set.  Emit position
2802         independent code and EH data if PIC.
2803
2804 2004-10-13  Kazuhiro Inaoka  <inaoka.kazuhiro@renesas.com>
2805
2806         * Makefile.am: Add m32r support.
2807         * configure.ac: Likewise.
2808         * Makefile.in: Regenerate.
2809         * confiugre: Regenerate.
2810         * src/types.c: Add m32r port to FFI_INTERNAL_TYPEDEF
2811         (uint64, sint64, double, longdouble)
2812         * src/m32r: New directory.
2813         * src/m32r/ffi.c: New file.
2814         * src/m32r/sysv.S: Likewise.
2815         * src/m32r/ffitarget.h: Likewise.
2816
2817 2004-10-02  Kaz Kojima  <kkojima@gcc.gnu.org>
2818
2819         * testsuite/libffi.call/negint.c: New test case.
2820
2821 2004-09-14  H.J. Lu  <hongjiu.lu@intel.com>
2822
2823         PR libgcj/17465
2824         * testsuite/lib/libffi-dg.exp: Don't use global ld_library_path.
2825         Set up LD_LIBRARY_PATH, SHLIB_PATH, LD_LIBRARYN32_PATH,
2826         LD_LIBRARY64_PATH, LD_LIBRARY_PATH_32, LD_LIBRARY_PATH_64 and
2827         DYLD_LIBRARY_PATH.
2828
2829 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2830
2831         * testsuite/libffi.call/many_win32.c: Remove whitespaces.
2832         * testsuite/libffi.call/promotion.c: Likewise.
2833         * testsuite/libffi.call/return_ll.c: Remove unused var. Cleanup
2834         whitespaces.
2835         * testsuite/libffi.call/return_sc.c: Likewise.
2836         * testsuite/libffi.call/return_uc.c: Likewise.
2837
2838 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2839
2840         * src/powerpc/darwin.S: Fix comments and identation.
2841         * src/powerpc/darwin_closure.S: Likewise.
2842
2843 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2844
2845         * src/powerpc/ffi_darwin.c: Add flag for longdouble return values.
2846         (ffi_prep_args): Handle longdouble arguments.
2847         (ffi_prep_cif_machdep): Set flags for longdouble. Calculate space for
2848         longdouble.
2849         (ffi_closure_helper_DARWIN): Add closure handling for longdouble.
2850         * src/powerpc/darwin.S (_ffi_call_DARWIN): Add handling of longdouble
2851         values.
2852         * src/powerpc/darwin_closure.S (_ffi_closure_ASM): Likewise.
2853         * src/types.c: Defined longdouble size and alignment for darwin.
2854
2855 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2856
2857         * src/powerpc/aix.S: Remove whitespaces.
2858         * src/powerpc/aix_closure.S: Likewise.
2859         * src/powerpc/asm.h: Likewise.
2860         * src/powerpc/ffi.c: Likewise.
2861         * src/powerpc/ffitarget.h: Likewise.
2862         * src/powerpc/linux64.S: Likewise.
2863         * src/powerpc/linux64_closure.S: Likewise.
2864         * src/powerpc/ppc_closure.S: Likewise.
2865         * src/powerpc/sysv.S: Likewise.
2866
2867 2004-08-30  Anthony Green  <green@redhat.com>
2868
2869         * Makefile.am: Add frv support.
2870         * Makefile.in, testsuite/Makefile.in: Rebuilt.
2871         * configure.ac: Read configure.host.
2872         * configure.in: Read configure.host.
2873         * configure.host: New file.  frv-elf needs libgloss.
2874         * include/ffi.h.in: Force ffi_closure to have a nice big (8)
2875         alignment.  This is needed to frv and shouldn't harm the others.
2876         * include/ffi_common.h (ALIGN_DOWN): New macro.
2877         * src/frv/ffi.c, src/frv/ffitarget.h, src/frv/eabi.S: New files.
2878
2879 2004-08-24  David Daney  <daney@avtrex.com>
2880
2881         * testsuite/libffi.call/closure_fn0.c: Xfail mips64* instead of mips*.
2882         * testsuite/libffi.call/closure_fn1.c: Likewise.
2883         * testsuite/libffi.call/closure_fn2.c  Likewise.
2884         * testsuite/libffi.call/closure_fn3.c: Likewise.
2885         * testsuite/libffi.call/closure_fn4.c: Likewise.
2886         * testsuite/libffi.call/closure_fn5.c: Likewise.
2887         * testsuite/libffi.call/cls_18byte.c: Likewise.
2888         * testsuite/libffi.call/cls_19byte.c: Likewise.
2889         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2890         * testsuite/libffi.call/cls_20byte.c: Likewise.
2891         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2892         * testsuite/libffi.call/cls_24byte.c: Likewise.
2893         * testsuite/libffi.call/cls_2byte.c: Likewise.
2894         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2895         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2896         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2897         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2898         * testsuite/libffi.call/cls_4byte.c: Likewise.
2899         * testsuite/libffi.call/cls_64byte.c: Likewise.
2900         * testsuite/libffi.call/cls_6byte.c: Likewise.
2901         * testsuite/libffi.call/cls_7byte.c: Likewise.
2902         * testsuite/libffi.call/cls_8byte.c: Likewise.
2903         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2904         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2905         * testsuite/libffi.call/cls_align_double.c: Likewise.
2906         * testsuite/libffi.call/cls_align_float.c: Likewise.
2907         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2908         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2909         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2910         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2911         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2912         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2913         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2914         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
2915         * testsuite/libffi.call/cls_double.c: Likewise.
2916         * testsuite/libffi.call/cls_float.c: Likewise.
2917         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2918         * testsuite/libffi.call/cls_multi_sshort.c: Likewise.
2919         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2920         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2921         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2922         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2923         * testsuite/libffi.call/cls_schar.c: Likewise.
2924         * testsuite/libffi.call/cls_sint.c: Likewise.
2925         * testsuite/libffi.call/cls_sshort.c: Likewise.
2926         * testsuite/libffi.call/cls_uchar.c: Likewise.
2927         * testsuite/libffi.call/cls_uint.c: Likewise.
2928         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2929         * testsuite/libffi.call/cls_ushort.c: Likewise.
2930         * testsuite/libffi.call/nested_struct.c: Likewise.
2931         * testsuite/libffi.call/nested_struct1.c: Likewise.
2932         * testsuite/libffi.call/nested_struct2.c: Likewise.
2933         * testsuite/libffi.call/nested_struct3.c: Likewise.
2934         * testsuite/libffi.call/problem1.c: Likewise.
2935         * testsuite/libffi.special/unwindtest.cc: Likewise.
2936         * testsuite/libffi.call/cls_12byte.c: Likewise and set return value
2937         to zero.
2938         * testsuite/libffi.call/cls_16byte.c: Likewise.
2939         * testsuite/libffi.call/cls_5byte.c: Likewise.
2940
2941 2004-08-23  David Daney <daney@avtrex.com>
2942
2943         PR libgcj/13141
2944         * src/mips/ffitarget.h (FFI_O32_SOFT_FLOAT): New ABI.
2945         * src/mips/ffi.c (ffi_prep_args): Fix alignment calculation.
2946         (ffi_prep_cif_machdep): Handle FFI_O32_SOFT_FLOAT floating point
2947         parameters and return types.
2948         (ffi_call): Handle FFI_O32_SOFT_FLOAT ABI.
2949         (ffi_prep_closure): Ditto.
2950         (ffi_closure_mips_inner_O32): Handle FFI_O32_SOFT_FLOAT ABI, fix
2951         alignment calculations.
2952         * src/mips/o32.S (ffi_closure_O32): Don't use floating point
2953         instructions if FFI_O32_SOFT_FLOAT, make stack frame ABI compliant.
2954
2955 2004-08-14  Casey Marshall <csm@gnu.org>
2956
2957         * src/mips/ffi.c (ffi_pref_cif_machdep): set `cif->flags' to
2958         contain `FFI_TYPE_UINT64' as return type for any 64-bit
2959         integer (O32 ABI only).
2960         (ffi_prep_closure): new function.
2961         (ffi_closure_mips_inner_O32): new function.
2962         * src/mips/ffitarget.h: Define `FFI_CLOSURES' and
2963         `FFI_TRAMPOLINE_SIZE' appropriately if the ABI is o32.
2964         * src/mips/o32.S (ffi_call_O32): add labels for .eh_frame. Return
2965         64 bit integers correctly.
2966         (ffi_closure_O32): new function.
2967         Added DWARF-2 unwind info for both functions.
2968
2969 2004-08-10  Andrew Haley  <aph@redhat.com>
2970
2971         * src/x86/ffi64.c (ffi_prep_args ): 8-align all stack arguments.
2972
2973 2004-08-01  Robert Millan  <robertmh@gnu.org>
2974
2975         * configure.ac: Detect knetbsd-gnu and kfreebsd-gnu.
2976         * configure: Regenerate.
2977
2978 2004-07-30  Maciej W. Rozycki  <macro@linux-mips.org>
2979
2980         * acinclude.m4 (AC_FUNC_MMAP_BLACKLIST): Check for <sys/mman.h>
2981         and mmap() explicitly instead of relying on preset autoconf cache
2982         variables.
2983         * aclocal.m4: Regenerate.
2984         * configure: Regenerate.
2985
2986 2004-07-11  Ulrich Weigand  <uweigand@de.ibm.com>
2987
2988         * src/s390/ffi.c (ffi_prep_args): Fix C aliasing violation.
2989         (ffi_check_float_struct): Remove unused prototype.
2990
2991 2004-06-30  Geoffrey Keating  <geoffk@apple.com>
2992
2993         * src/powerpc/ffi_darwin.c (flush_icache): ';' is a comment
2994         character on Darwin, use '\n\t' instead.
2995
2996 2004-06-26  Matthias Klose  <doko@debian.org>
2997
2998         * libtool-version: Fix typo in revision/age.
2999
3000 2004-06-17  Matthias Klose  <doko@debian.org>
3001
3002         * libtool-version: New.
3003         * Makefile.am (libffi_la_LDFLAGS): Use -version-info for soname.
3004         * Makefile.in: Regenerate.
3005
3006 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
3007
3008         * Makefile.am: Remove useless multilib rules.
3009         * Makefile.in: Regenerate.
3010         * aclocal.m4: Regenerate with automake 1.8.5.
3011         * configure.ac: Remove useless multilib configury.
3012         * configure: Regenerate.
3013
3014 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
3015
3016         * .cvsignore: New file.
3017
3018 2004-06-10  Jakub Jelinek  <jakub@redhat.com>
3019
3020         * src/ia64/unix.S (ffi_call_unix): Insert group barrier break
3021         fp_done.
3022         (ffi_closure_UNIX): Fix f14/f15 adjustment if FLOAT_SZ is ever
3023         changed from 8.
3024
3025 2004-06-06  Sean McNeil  <sean@mcneil.com>
3026
3027         * configure.ac: Add x86_64-*-freebsd* support.
3028         * configure: Regenerate.
3029
3030 2004-04-26  Joe Buck <jbuck@welsh-buck.org>
3031
3032         Bug 15093
3033         * configure.ac: Test for existence of mmap and sys/mman.h before
3034         checking blacklist.  Fix suggested by Jim Wilson.
3035         * configure: Regenerate.
3036
3037 2004-04-26  Matt Austern  <austern@apple.com>
3038
3039         * src/powerpc/darwin.S: Go through a non-lazy pointer for initial
3040         FDE location.
3041         * src/powerpc/darwin_closure.S: Likewise.
3042
3043 2004-04-24  Andreas Tobler  <a.tobler@schweiz.ch>
3044
3045         * testsuite/libffi.call/cls_multi_schar.c (main): Fix initialization
3046         error. Reported by Thomas Heller <theller@python.net>.
3047         * testsuite/libffi.call/cls_multi_sshort.c (main): Likewise.
3048         * testsuite/libffi.call/cls_multi_ushort.c (main): Likewise.
3049
3050 2004-03-20  Matthias Klose  <doko@debian.org>
3051
3052         * src/pa/linux.S: Fix typo.
3053
3054 2004-03-19  Matthias Klose  <doko@debian.org>
3055
3056         * Makefile.am: Update.
3057         * Makefile.in: Regenerate.
3058         * src/pa/ffi.h.in: Remove.
3059         * src/pa/ffitarget.h: New file.
3060
3061 2004-02-10  Randolph Chung  <tausq@debian.org>
3062
3063         * Makefile.am: Add PA support.
3064         * Makefile.in: Regenerate.
3065         * include/Makefile.in: Regenerate.
3066         * configure.ac: Add PA target.
3067         * configure: Regenerate.
3068         * src/pa/ffi.c: New file.
3069         * src/pa/ffi.h.in: Add PA support.
3070         * src/pa/linux.S: New file.
3071         * prep_cif.c: Add PA support.
3072
3073 2004-03-16  Hosaka Yuji  <hos@tamanegi.org>
3074
3075         * src/types.c: Fix alignment size of X86_WIN32 case int64 and
3076         double.
3077         * src/x86/ffi.c (ffi_prep_args): Replace ecif->cif->rtype->type
3078         with ecif->cif->flags.
3079         (ffi_call, ffi_prep_incoming_args_SYSV): Replace cif->rtype->type
3080         with cif->flags.
3081         (ffi_prep_cif_machdep): Add X86_WIN32 struct case.
3082         (ffi_closure_SYSV): Add 1 or 2-bytes struct case for X86_WIN32.
3083         * src/x86/win32.S (retstruct1b, retstruct2b, sc_retstruct1b,
3084         sc_retstruct2b): Add for 1 or 2-bytes struct case.
3085
3086 2004-03-15 Kelley Cook <kcook@gcc.gnu.org>
3087
3088         * configure.in: Rename file to ...
3089         * configure.ac: ... this.
3090         * fficonfig.h.in: Regenerate.
3091         * Makefile.in: Regenerate.
3092         * include/Makefile.in: Regenerate.
3093         * testsuite/Makefile.in: Regenerate.
3094
3095 2004-03-12  Matt Austern  <austern@apple.com>
3096
3097         * src/powerpc/darwin.S: Fix EH information so it corresponds to
3098         changes in EH format resulting from addition of linkonce support.
3099         * src/powerpc/darwin_closure.S: Likewise.
3100
3101 2004-03-11  Andreas Tobler  <a.tobler@schweiz.ch>
3102             Paolo Bonzini  <bonzini@gnu.org>
3103
3104         * Makefile.am (AUTOMAKE_OPTIONS): Set them.
3105         Remove VPATH. Remove rules for object files. Remove multilib support.
3106         (AM_CCASFLAGS): Add.
3107         * configure.in (AC_CONFIG_HEADERS): Relace AM_CONFIG_HEADER.
3108         (AC_PREREQ): Bump version to 2.59.
3109         (AC_INIT): Fill with version info and bug address.
3110         (ORIGINAL_LD_FOR_MULTILIBS): Remove.
3111         (AM_ENABLE_MULTILIB): Use this instead of AC_ARG_ENABLE.
3112         De-precious CC so that the right flags are passed down to multilibs.
3113         (AC_MSG_ERROR): Replace obsolete macro AC_ERROR.
3114         (AC_CONFIG_FILES): Replace obsolete macro AC_LINK_FILES.
3115         (AC_OUTPUT): Reorganize the output with AC_CONFIG_COMMANDS.
3116         * configure: Rebuilt.
3117         * aclocal.m4: Likewise.
3118         * Makefile.in, include/Makefile.in, testsuite/Makefile.in: Likewise.
3119         * fficonfig.h.in: Likewise.
3120
3121 2004-03-11  Andreas Schwab  <schwab@suse.de>
3122
3123         * src/ia64/ffi.c (ffi_prep_incoming_args_UNIX): Get floating point
3124         arguments from fp registers only for the first 8 parameter slots.
3125         Don't convert a float parameter when passed in memory.
3126
3127 2004-03-09  Hans-Peter Nilsson  <hp@axis.com>
3128
3129         * configure: Regenerate for config/accross.m4 correction.
3130
3131 2004-02-25  Matt Kraai  <kraai@alumni.cmu.edu>
3132
3133         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Change
3134         ecif->cif->bytes to bytes.
3135         (ffi_prep_cif_machdep): Add braces around nested if statement.
3136
3137 2004-02-09  Alan Modra  <amodra@bigpond.net.au>
3138
3139         * src/types.c (pointer): POWERPC64 has 8 byte pointers.
3140
3141         * src/powerpc/ffi.c (ffi_prep_args64): Correct long double handling.
3142         (ffi_closure_helper_LINUX64): Fix typo.
3143         * testsuite/libffi.call/cls_align_longdouble.c: Pass -mlong-double-128
3144         for powerpc64-*-*.
3145         * testsuite/libffi.call/float.c: Likewise.
3146         * testsuite/libffi.call/float2.c: Likewise.
3147
3148 2004-02-08  Alan Modra  <amodra@bigpond.net.au>
3149
3150         * src/powerpc/ffi.c (ffi_prep_cif_machdep <FFI_LINUX64>): Correct
3151         long double function return and long double arg handling.
3152         (ffi_closure_helper_LINUX64): Formatting.  Delete unused "ng" var.
3153         Use "end_pfr" instead of "nf".  Correct long double handling.
3154         Localise "temp".
3155         * src/powerpc/linux64.S (ffi_call_LINUX64): Save f2 long double
3156         return value.
3157         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Allocate
3158         space for long double return value.  Adjust stack frame and offsets.
3159         Load f2 long double return.
3160
3161 2004-02-07  Alan Modra  <amodra@bigpond.net.au>
3162
3163         * src/types.c: Use 16 byte long double for POWERPC64.
3164
3165 2004-01-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
3166
3167         * src/sparc/ffi.c (ffi_prep_args_v9): Shift the parameter array
3168         when the structure return address is passed in %o0.
3169         (ffi_V9_return_struct): Rename into ffi_v9_layout_struct.
3170         (ffi_v9_layout_struct): Align the field following a nested structure
3171         on a word boundary.  Use memmove instead of memcpy.
3172         (ffi_call): Update call to ffi_V9_return_struct.
3173         (ffi_prep_closure): Define 'ctx' only for V8.
3174         (ffi_closure_sparc_inner): Clone into ffi_closure_sparc_inner_v8
3175         and ffi_closure_sparc_inner_v9.
3176         (ffi_closure_sparc_inner_v8): Return long doubles by reference.
3177         Always skip the structure return address.  For structures and long
3178         doubles, copy the argument directly.
3179         (ffi_closure_sparc_inner_v9): Skip the structure return address only
3180         if required.  Shift the maximum floating-point slot accordingly.  For
3181         big structures, copy the argument directly; otherwise, left-justify the
3182         argument and call ffi_v9_layout_struct to lay out the structure on
3183         the stack.
3184         * src/sparc/v8.S: Undef STACKFRAME before defining it.
3185         (ffi_closure_v8): Pass the structure return address.  Update call to
3186         ffi_closure_sparc_inner_v8.  Short-circuit FFI_TYPE_INT handling.
3187         Skip the 'unimp' insn when returning long doubles and structures.
3188         * src/sparc/v9.S: Undef STACKFRAME before defining it.
3189         (ffi_closure_v9): Increase the frame size by 2 words.  Short-circuit
3190         FFI_TYPE_INT handling.  Load structures both in integers and
3191         floating-point registers on return.
3192         * README: Update status of the SPARC port.
3193
3194 2004-01-24  Andreas Tobler  <a.tobler@schweiz.ch>
3195
3196         * testsuite/libffi.call/pyobjc-tc.c (main): Treat result value
3197         as of type ffi_arg.
3198         * testsuite/libffi.call/struct3.c (main): Fix CHECK.
3199
3200 2004-01-22  Ulrich Weigand  <uweigand@de.ibm.com>
3201
3202         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Treat result
3203         value as of type ffi_arg, not unsigned int.
3204
3205 2004-01-21  Michael Ritzert  <ritzert@t-online.de>
3206
3207         * ffi64.c (ffi_prep_args): Cast the RHS of an assignment instead
3208         of the LHS.
3209
3210 2004-01-12  Andreas Tobler  <a.tobler@schweiz.ch>
3211
3212         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_32 for
3213         Solaris.
3214
3215 2004-01-08  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3216
3217         * testsuite/libffi.call/ffitest.h (allocate_mmap): Cast MAP_FAILED
3218         to void *.
3219
3220 2003-12-10  Richard Henderson  <rth@redhat.com>
3221
3222         * testsuite/libffi.call/cls_align_pointer.c: Cast pointers to
3223         size_t instead of int.
3224
3225 2003-12-04  Hosaka Yuji  <hos@tamanegi.org>
3226
3227         * testsuite/libffi.call/many_win32.c: Include <float.h>.
3228         * testsuite/libffi.call/many_win32.c (main): Replace variable
3229         int i with unsigned long ul.
3230
3231         * testsuite/libffi.call/cls_align_uint64.c: New test case.
3232         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
3233         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
3234         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
3235         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
3236         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
3237         * testsuite/libffi.call/cls_align_float.c: Likewise.
3238         * testsuite/libffi.call/cls_align_double.c: Likewise.
3239         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
3240         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
3241
3242 2003-12-02  Hosaka Yuji  <hos@tamanegi.org>
3243
3244         PR other/13221
3245         * src/x86/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV):
3246         Align arguments to 32 bits.
3247
3248 2003-12-01  Andreas Tobler  <a.tobler@schweiz.ch>
3249
3250         PR other/13221
3251         * testsuite/libffi.call/cls_multi_sshort.c: New test case.
3252         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
3253         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
3254         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
3255         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
3256         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
3257
3258         * testsuite/libffi.special/unwindtest.cc: Cosmetics.
3259
3260 2003-11-26  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
3261
3262         * testsuite/libffi.call/ffitest.h: Include <fcntl.h>.
3263         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3264
3265 2003-11-22  Andreas Tobler  <a.tobler@schweiz.ch>
3266
3267         * Makefile.in: Rebuilt.
3268         * configure: Likewise.
3269         * testsuite/libffi.special/unwindtest.cc: Convert the mmap to
3270         the right type.
3271
3272 2003-11-21  Andreas Jaeger  <aj@suse.de>
3273             Andreas Tobler  <a.tobler@schweiz.ch>
3274
3275         * acinclude.m4: Add AC_FUNC_MMAP_BLACKLIST.
3276         * configure.in: Call AC_FUNC_MMAP_BLACKLIST.
3277         * Makefile.in: Rebuilt.
3278         * aclocal.m4: Likewise.
3279         * configure: Likewise.
3280         * fficonfig.h.in: Likewise.
3281         * testsuite/lib/libffi-dg.exp: Add include dir.
3282         * testsuite/libffi.call/ffitest.h: Add MMAP definitions.
3283         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3284         * testsuite/libffi.call/closure_fn0.c: Use MMAP functionality
3285         for ffi_closure if available.
3286         * testsuite/libffi.call/closure_fn1.c: Likewise.
3287         * testsuite/libffi.call/closure_fn2.c: Likewise.
3288         * testsuite/libffi.call/closure_fn3.c: Likewise.
3289         * testsuite/libffi.call/closure_fn4.c: Likewise.
3290         * testsuite/libffi.call/closure_fn5.c: Likewise.
3291         * testsuite/libffi.call/cls_12byte.c: Likewise.
3292         * testsuite/libffi.call/cls_16byte.c: Likewise.
3293         * testsuite/libffi.call/cls_18byte.c: Likewise.
3294         * testsuite/libffi.call/cls_19byte.c: Likewise.
3295         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3296         * testsuite/libffi.call/cls_20byte.c: Likewise.
3297         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3298         * testsuite/libffi.call/cls_24byte.c: Likewise.
3299         * testsuite/libffi.call/cls_2byte.c: Likewise.
3300         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3301         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3302         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3303         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3304         * testsuite/libffi.call/cls_4byte.c: Likewise.
3305         * testsuite/libffi.call/cls_5byte.c: Likewise.
3306         * testsuite/libffi.call/cls_64byte.c: Likewise.
3307         * testsuite/libffi.call/cls_6byte.c: Likewise.
3308         * testsuite/libffi.call/cls_7byte.c: Likewise.
3309         * testsuite/libffi.call/cls_8byte.c: Likewise.
3310         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3311         * testsuite/libffi.call/cls_9byte2.c: Likewise.
3312         * testsuite/libffi.call/cls_double.c: Likewise.
3313         * testsuite/libffi.call/cls_float.c: Likewise.
3314         * testsuite/libffi.call/cls_schar.c: Likewise.
3315         * testsuite/libffi.call/cls_sint.c: Likewise.
3316         * testsuite/libffi.call/cls_sshort.c: Likewise.
3317         * testsuite/libffi.call/cls_uchar.c: Likewise.
3318         * testsuite/libffi.call/cls_uint.c: Likewise.
3319         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3320         * testsuite/libffi.call/cls_ushort.c: Likewise.
3321         * testsuite/libffi.call/nested_struct.c: Likewise.
3322         * testsuite/libffi.call/nested_struct1.c: Likewise.
3323         * testsuite/libffi.call/nested_struct2.c: Likewise.
3324         * testsuite/libffi.call/nested_struct3.c: Likewise.
3325         * testsuite/libffi.call/problem1.c: Likewise.
3326         * testsuite/libffi.special/unwindtest.cc: Likewise.
3327
3328 2003-11-20  Andreas Tobler  <a.tobler@schweiz.ch>
3329
3330         * testsuite/lib/libffi-dg.exp: Make the -lgcc_s conditional.
3331
3332 2003-11-19  Andreas Tobler  <a.tobler@schweiz.ch>
3333
3334         * testsuite/lib/libffi-dg.exp: Add DYLD_LIBRARY_PATH for darwin.
3335         Add -lgcc_s to additional flags.
3336
3337 2003-11-12  Andreas Tobler  <a.tobler@schweiz.ch>
3338
3339         * configure.in, include/Makefile.am: PR libgcj/11147, install
3340         the ffitarget.h header file in a gcc versioned and target
3341         dependent place.
3342         * configure: Regenerated.
3343         * Makefile.in, include/Makefile.in: Likewise.
3344         * testsuite/Makefile.in: Likewise.
3345
3346 2003-11-09  Andreas Tobler  <a.tobler@schweiz.ch>
3347
3348         * testsuite/libffi.call/closure_fn0.c: Print result and check
3349         with dg-output to make debugging easier.
3350         * testsuite/libffi.call/closure_fn1.c: Likewise.
3351         * testsuite/libffi.call/closure_fn2.c: Likewise.
3352         * testsuite/libffi.call/closure_fn3.c: Likewise.
3353         * testsuite/libffi.call/closure_fn4.c: Likewise.
3354         * testsuite/libffi.call/closure_fn5.c: Likewise.
3355         * testsuite/libffi.call/cls_12byte.c: Likewise.
3356         * testsuite/libffi.call/cls_16byte.c: Likewise.
3357         * testsuite/libffi.call/cls_18byte.c: Likewise.
3358         * testsuite/libffi.call/cls_19byte.c: Likewise.
3359         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3360         * testsuite/libffi.call/cls_20byte.c: Likewise.
3361         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3362         * testsuite/libffi.call/cls_24byte.c: Likewise.
3363         * testsuite/libffi.call/cls_2byte.c: Likewise.
3364         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3365         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3366         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3367         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3368         * testsuite/libffi.call/cls_4byte.c: Likewise.
3369         * testsuite/libffi.call/cls_5byte.c: Likewise.
3370         * testsuite/libffi.call/cls_64byte.c: Likewise.
3371         * testsuite/libffi.call/cls_6byte.c: Likewise.
3372         * testsuite/libffi.call/cls_7byte.c: Likewise.
3373         * testsuite/libffi.call/cls_8byte.c: Likewise.
3374         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3375         * testsuite/libffi.call/cls_9byte2.c: Likewise.
3376         * testsuite/libffi.call/cls_double.c: Likewise.
3377         * testsuite/libffi.call/cls_float.c: Likewise.
3378         * testsuite/libffi.call/cls_schar.c: Likewise.
3379         * testsuite/libffi.call/cls_sint.c: Likewise.
3380         * testsuite/libffi.call/cls_sshort.c: Likewise.
3381         * testsuite/libffi.call/cls_uchar.c: Likewise.
3382         * testsuite/libffi.call/cls_uint.c: Likewise.
3383         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3384         * testsuite/libffi.call/cls_ushort.c: Likewise.
3385         * testsuite/libffi.call/problem1.c: Likewise.
3386
3387         * testsuite/libffi.special/unwindtest.cc: Make ffi_closure
3388         static.
3389
3390 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
3391
3392         * testsuite/libffi.call/cls_9byte2.c: New test case.
3393         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3394         * testsuite/libffi.call/cls_64byte.c: Likewise.
3395         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3396         * testsuite/libffi.call/cls_19byte.c: Likewise.
3397         * testsuite/libffi.call/cls_18byte.c: Likewise.
3398         * testsuite/libffi.call/closure_fn4.c: Likewise.
3399         * testsuite/libffi.call/closure_fn5.c: Likewise.
3400         * testsuite/libffi.call/cls_schar.c: Likewise.
3401         * testsuite/libffi.call/cls_sint.c: Likewise.
3402         * testsuite/libffi.call/cls_sshort.c: Likewise.
3403         * testsuite/libffi.call/nested_struct2.c: Likewise.
3404         * testsuite/libffi.call/nested_struct3.c: Likewise.
3405
3406 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
3407
3408         * testsuite/libffi.call/cls_double.c: Do a check on the result.
3409         * testsuite/libffi.call/cls_uchar.c: Likewise.
3410         * testsuite/libffi.call/cls_uint.c: Likewise.
3411         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3412         * testsuite/libffi.call/cls_ushort.c: Likewise.
3413         * testsuite/libffi.call/return_sc.c: Cleanup whitespaces.
3414
3415 2003-11-06  Andreas Tobler  <a.tobler@schweiz.ch>
3416
3417         * src/prep_cif.c (ffi_prep_cif): Move the validity check after
3418         the initialization.
3419
3420 2003-10-23  Andreas Tobler  <a.tobler@schweiz.ch>
3421
3422         * src/java_raw_api.c (ffi_java_ptrarray_to_raw): Replace
3423         FFI_ASSERT(FALSE) with FFI_ASSERT(0).
3424
3425 2003-10-22  David Daney  <ddaney@avtrex.com>
3426
3427         * src/mips/ffitarget.h: Replace undefined UINT32 and friends with
3428         __attribute__((__mode__(__SI__))) and friends.
3429
3430 2003-10-22  Andreas Schwab  <schwab@suse.de>
3431
3432         * src/ia64/ffi.c: Replace FALSE/TRUE with false/true.
3433
3434 2003-10-21  Andreas Tobler  <a.tobler@schweiz.ch>
3435
3436         * configure.in: AC_LINK_FILES(ffitarget.h).
3437         * configure: Regenerate.
3438         * Makefile.in: Likewise.
3439         * include/Makefile.in: Likewise.
3440         * testsuite/Makefile.in: Likewise.
3441         * fficonfig.h.in: Likewise.
3442
3443 2003-10-21  Paolo Bonzini  <bonzini@gnu.org>
3444             Richard Henderson  <rth@redhat.com>
3445
3446         Avoid that ffi.h includes fficonfig.h.
3447
3448         * Makefile.am (EXTRA_DIST): Include ffitarget.h files
3449         (TARGET_SRC_MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3450         (TARGET_SRC_MIPS_SGI): Removed.
3451         (MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3452         (MIPS_SGI): Removed.
3453         (CLEANFILES): Removed.
3454         (mostlyclean-am, clean-am, mostlyclean-sub, clean-sub): New
3455         targets.
3456         * acconfig.h: Removed.
3457         * configure.in: Compute sizeofs only for double and long double.
3458         Use them to define and subst HAVE_LONG_DOUBLE.  Include comments
3459         into AC_DEFINE instead of using acconfig.h.  Create
3460         include/ffitarget.h instead of include/fficonfig.h.  Rename
3461         MIPS_GCC to MIPS_IRIX, drop MIPS_SGI since we are in gcc's tree.
3462         AC_DEFINE EH_FRAME_FLAGS.
3463         * include/Makefile.am (DISTCLEANFILES): New automake macro.
3464         (hack_DATA): Add ffitarget.h.
3465         * include/ffi.h.in: Remove all system specific definitions.
3466         Declare raw API even if it is not installed, why bother?
3467         Use limits.h instead of SIZEOF_* to define ffi_type_*.  Do
3468         not define EH_FRAME_FLAGS, it is in fficonfig.h now.  Include
3469         ffitarget.h instead of fficonfig.h.  Remove ALIGN macro.
3470         (UINT_ARG, INT_ARG): Removed, use ffi_arg and ffi_sarg instead.
3471         * include/ffi_common.h (bool): Do not define.
3472         (ffi_assert): Accept failed assertion.
3473         (ffi_type_test): Return void and accept file/line.
3474         (FFI_ASSERT): Pass stringized failed assertion.
3475         (FFI_ASSERT_AT): New macro.
3476         (FFI_ASSERT_VALID_TYPE): New macro.
3477         (UINT8, SINT8, UINT16, SINT16, UINT32, SINT32,
3478         UINT64, SINT64): Define here with gcc's __attribute__ macro
3479         instead of in ffi.h
3480         (FLOAT32, ALIGN): Define here instead of in ffi.h
3481         * include/ffi-mips.h: Removed.  Its content moved to
3482         src/mips/ffitarget.h after separating assembly and C sections.
3483         * src/alpha/ffi.c, src/alpha/ffi.c, src/java_raw_api.c
3484         src/prep_cif.c, src/raw_api.c, src/ia64/ffi.c,
3485         src/mips/ffi.c, src/mips/n32.S, src/mips/o32.S,
3486         src/mips/ffitarget.h, src/sparc/ffi.c, src/x86/ffi64.c:
3487         SIZEOF_ARG -> FFI_SIZEOF_ARG.
3488         * src/ia64/ffi.c: Include stdbool.h (provided by GCC 2.95+).
3489         * src/debug.c (ffi_assert): Accept stringized failed assertion.
3490         (ffi_type_test): Rewritten.
3491         * src/prep-cif.c (initialize_aggregate, ffi_prep_cif): Call
3492         FFI_ASSERT_VALID_TYPE.
3493         * src/alpha/ffitarget.h, src/arm/ffitarget.h,
3494         src/ia64/ffitarget.h, src/m68k/ffitarget.h,
3495         src/mips/ffitarget.h, src/powerpc/ffitarget.h,
3496         src/s390/ffitarget.h, src/sh/ffitarget.h,
3497         src/sh64/ffitarget.h, src/sparc/ffitarget.h,
3498         src/x86/ffitarget.h: New files.
3499         * src/alpha/osf.S, src/arm/sysv.S, src/ia64/unix.S,
3500         src/m68k/sysv.S, src/mips/n32.S, src/mips/o32.S,
3501         src/powerpc/aix.S, src/powerpc/darwin.S,
3502         src/powerpc/ffi_darwin.c, src/powerpc/linux64.S,
3503         src/powerpc/linux64_closure.S, src/powerpc/ppc_closure.S,
3504         src/powerpc/sysv.S, src/s390/sysv.S, src/sh/sysv.S,
3505         src/sh64/sysv.S, src/sparc/v8.S, src/sparc/v9.S,
3506         src/x86/sysv.S, src/x86/unix64.S, src/x86/win32.S:
3507         include fficonfig.h
3508
3509 2003-10-20  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3510
3511         * src/mips/ffi.c: Use _ABIN32, _ABIO32 instead of external
3512         _MIPS_SIM_NABI32, _MIPS_SIM_ABI32.
3513
3514 2003-10-19  Andreas Tobler  <a.tobler@schweiz.ch>
3515
3516         * src/powerpc/ffi_darwin.c (ffi_prep_args): Declare bytes again.
3517         Used when FFI_DEBUG = 1.
3518
3519 2003-10-14  Alan Modra  <amodra@bigpond.net.au>
3520
3521         * src/types.c (double, longdouble): Default POWERPC64 to 8 byte size
3522         and align.
3523
3524 2003-10-06  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3525
3526         * include/ffi_mips.h: Define FFI_MIPS_N32 for N32/N64 ABIs,
3527         FFI_MIPS_O32 for O32 ABI.
3528
3529 2003-10-01  Andreas Tobler  <a.tobler@schweiz.ch>
3530
3531         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_64 for
3532         SPARC64. Cleanup whitespaces.
3533
3534 2003-09-19  Andreas Tobler  <a.tobler@schweiz.ch>
3535
3536         * testsuite/libffi.call/closure_fn0.c: Xfail mips, arm,
3537         strongarm, xscale. Cleanup whitespaces.
3538         * testsuite/libffi.call/closure_fn1.c: Likewise.
3539         * testsuite/libffi.call/closure_fn2.c: Likewise.
3540         * testsuite/libffi.call/closure_fn3.c: Likewise.
3541         * testsuite/libffi.call/cls_12byte.c: Likewise.
3542         * testsuite/libffi.call/cls_16byte.c: Likewise.
3543         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3544         * testsuite/libffi.call/cls_20byte.c: Likewise.
3545         * testsuite/libffi.call/cls_24byte.c: Likewise.
3546         * testsuite/libffi.call/cls_2byte.c: Likewise.
3547         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3548         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3549         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3550         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3551         * testsuite/libffi.call/cls_4byte.c: Likewise.
3552         * testsuite/libffi.call/cls_5byte.c: Likewise.
3553         * testsuite/libffi.call/cls_6byte.c: Likewise.
3554         * testsuite/libffi.call/cls_7byte.c: Likewise.
3555         * testsuite/libffi.call/cls_8byte.c: Likewise.
3556         * testsuite/libffi.call/cls_double.c: Likewise.
3557         * testsuite/libffi.call/cls_float.c: Likewise.
3558         * testsuite/libffi.call/cls_uchar.c: Likewise.
3559         * testsuite/libffi.call/cls_uint.c: Likewise.
3560         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3561         * testsuite/libffi.call/cls_ushort.c: Likewise.
3562         * testsuite/libffi.call/nested_struct.c: Likewise.
3563         * testsuite/libffi.call/nested_struct1.c: Likewise.
3564         * testsuite/libffi.call/problem1.c: Likewise.
3565         * testsuite/libffi.special/unwindtest.cc: Likewise.
3566         * testsuite/libffi.call/pyobjc-tc.c: Cleanup whitespaces.
3567
3568 2003-09-18  David Edelsohn  <edelsohn@gnu.org>
3569
3570         * src/powerpc/aix.S: Cleanup whitespaces.
3571         * src/powerpc/aix_closure.S: Likewise.
3572
3573 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3574
3575         * src/powerpc/darwin.S: Cleanup whitespaces, comment formatting.
3576         * src/powerpc/darwin_closure.S: Likewise.
3577         * src/powerpc/ffi_darwin.c: Likewise.
3578
3579 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3580             David Edelsohn  <edelsohn@gnu.org>
3581
3582         * src/types.c (double): Add AIX and Darwin to the right TYPEDEF.
3583         * src/powerpc/aix_closure.S: Remove the pointer to the outgoing
3584         parameter stack.
3585         * src/powerpc/darwin_closure.S: Likewise.
3586         * src/powerpc/ffi_darwin.c (ffi_prep_args): Handle structures
3587         according to the Darwin/AIX ABI.
3588         (ffi_prep_cif_machdep): Likewise.
3589         (ffi_closure_helper_DARWIN): Likewise.
3590         Remove the outgoing parameter stack logic. Simplify the evaluation
3591         of the different CASE types.
3592         (ffi_prep_clousure): Avoid the casts on lvalues. Change the branch
3593         statement in the trampoline code.
3594
3595 2003-09-18  Kaz Kojima  <kkojima@gcc.gnu.org>
3596
3597         * src/sh/ffi.c (ffi_prep_args): Take account into the alignement
3598         for the register size.
3599         (ffi_closure_helper_SYSV): Handle the structure return value
3600         address correctly.
3601         (ffi_closure_helper_SYSV): Return the appropriate type when
3602         the registers are used for the structure return value.
3603         * src/sh/sysv.S (ffi_closure_SYSV): Fix the stack layout for
3604         the 64-bit return value.  Update copyright years.
3605
3606 2003-09-17  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3607
3608         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Search in
3609         srcdir for ffi_mips.h.
3610
3611 2003-09-12  Alan Modra  <amodra@bigpond.net.au>
3612
3613         * src/prep_cif.c (initialize_aggregate): Include tail padding in
3614         structure size.
3615         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Correct
3616         placement of float result.
3617         * testsuite/libffi.special/unwindtest.cc (closure_test_fn1): Correct
3618         cast of "resp" for big-endian 64 bit machines.
3619
3620 2003-09-11  Alan Modra  <amodra@bigpond.net.au>
3621
3622         * src/types.c (double, longdouble): Merge identical SH and ARM
3623         typedefs, and add POWERPC64.
3624         * src/powerpc/ffi.c (ffi_prep_args64): Correct next_arg calc for
3625         struct split over gpr and rest.
3626         (ffi_prep_cif_machdep): Correct intarg_count for structures.
3627         * src/powerpc/linux64.S (ffi_call_LINUX64): Fix gpr offsets.
3628
3629 2003-09-09  Andreas Tobler  <a.tobler@schweiz.ch>
3630
3631         * src/powerpc/ffi.c (ffi_closure_helper_SYSV) Handle struct
3632         passing correctly.
3633
3634 2003-09-09  Alan Modra  <amodra@bigpond.net.au>
3635
3636         * configure: Regenerate.
3637
3638 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3639
3640         * Makefile.am: Remove build rules for ffitest.
3641         * Makefile.in: Rebuilt.
3642
3643 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3644
3645         * src/java_raw_api.c: Include <stdlib.h> to fix compiler warning
3646         about implicit declaration of abort().
3647
3648 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3649
3650         * Makefile.am: Add dejagnu test framework. Fixes PR other/11411.
3651         * Makefile.in: Rebuilt.
3652         * configure.in: Add dejagnu test framework.
3653         * configure: Rebuilt.
3654
3655         * testsuite/Makefile.am: New file.
3656         * testsuite/Makefile.in: Built
3657         * testsuite/lib/libffi-dg.exp: New file.
3658         * testsuite/config/default.exp: Likewise.
3659         * testsuite/libffi.call/call.exp: Likewise.
3660         * testsuite/libffi.call/ffitest.h: Likewise.
3661         * testsuite/libffi.call/closure_fn0.c: Likewise.
3662         * testsuite/libffi.call/closure_fn1.c: Likewise.
3663         * testsuite/libffi.call/closure_fn2.c: Likewise.
3664         * testsuite/libffi.call/closure_fn3.c: Likewise.
3665         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3666         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3667         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3668         * testsuite/libffi.call/cls_2byte.c: Likewise.
3669         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3670         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3671         * testsuite/libffi.call/cls_4byte.c: Likewise.
3672         * testsuite/libffi.call/cls_5byte.c: Likewise.
3673         * testsuite/libffi.call/cls_6byte.c: Likewise.
3674         * testsuite/libffi.call/cls_7byte.c: Likewise.
3675         * testsuite/libffi.call/cls_8byte.c: Likewise.
3676         * testsuite/libffi.call/cls_12byte.c: Likewise.
3677         * testsuite/libffi.call/cls_16byte.c: Likewise.
3678         * testsuite/libffi.call/cls_20byte.c: Likewise.
3679         * testsuite/libffi.call/cls_24byte.c: Likewise.
3680         * testsuite/libffi.call/cls_double.c: Likewise.
3681         * testsuite/libffi.call/cls_float.c: Likewise.
3682         * testsuite/libffi.call/cls_uchar.c: Likewise.
3683         * testsuite/libffi.call/cls_uint.c: Likewise.
3684         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3685         * testsuite/libffi.call/cls_ushort.c: Likewise.
3686         * testsuite/libffi.call/float.c: Likewise.
3687         * testsuite/libffi.call/float1.c: Likewise.
3688         * testsuite/libffi.call/float2.c: Likewise.
3689         * testsuite/libffi.call/many.c: Likewise.
3690         * testsuite/libffi.call/many_win32.c: Likewise.
3691         * testsuite/libffi.call/nested_struct.c: Likewise.
3692         * testsuite/libffi.call/nested_struct1.c: Likewise.
3693         * testsuite/libffi.call/pyobjc-tc.c: Likewise.
3694         * testsuite/libffi.call/problem1.c: Likewise.
3695         * testsuite/libffi.call/promotion.c: Likewise.
3696         * testsuite/libffi.call/return_ll.c: Likewise.
3697         * testsuite/libffi.call/return_sc.c: Likewise.
3698         * testsuite/libffi.call/return_uc.c: Likewise.
3699         * testsuite/libffi.call/strlen.c: Likewise.
3700         * testsuite/libffi.call/strlen_win32.c: Likewise.
3701         * testsuite/libffi.call/struct1.c: Likewise.
3702         * testsuite/libffi.call/struct2.c: Likewise.
3703         * testsuite/libffi.call/struct3.c: Likewise.
3704         * testsuite/libffi.call/struct4.c: Likewise.
3705         * testsuite/libffi.call/struct5.c: Likewise.
3706         * testsuite/libffi.call/struct6.c: Likewise.
3707         * testsuite/libffi.call/struct7.c: Likewise.
3708         * testsuite/libffi.call/struct8.c: Likewise.
3709         * testsuite/libffi.call/struct9.c: Likewise.
3710         * testsuite/libffi.special/special.exp: New file.
3711         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3712         * testsuite/libffi.special/unwindtest.cc: Likewise.
3713
3714
3715 2003-08-13  Kaz Kojima  <kkojima@gcc.gnu.org>
3716
3717         * src/sh/ffi.c (OFS_INT16): Set 0 for little endian case.  Update
3718         copyright years.
3719
3720 2003-08-02  Alan Modra  <amodra@bigpond.net.au>
3721
3722         * src/powerpc/ffi.c (ffi_prep_args64): Modify for changed gcc
3723         structure passing.
3724         (ffi_closure_helper_LINUX64): Likewise.
3725         * src/powerpc/linux64.S: Remove code writing to parm save area.
3726         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Use return
3727         address in lr from ffi_closure_helper_LINUX64 call to calculate
3728         table address.  Optimize function tail.
3729
3730 2003-07-28  Andreas Tobler  <a.tobler@schweiz.ch>
3731
3732         * src/sparc/ffi.c: Handle all floating point registers.
3733         * src/sparc/v9.S: Likewise. Fixes second part of PR target/11410.
3734
3735 2003-07-11  Gerald Pfeifer  <pfeifer@dbai.tuwien.ac.at>
3736
3737         * README: Note that libffi is not part of GCC.  Update the project
3738         URL and status.
3739
3740 2003-06-19  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3741
3742         * src/powerpc/ppc_closure.S: Include ffi.h.
3743
3744 2003-06-13  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3745
3746         * src/x86/sysv.S: Avoid gas-only .uleb128/.sleb128 directives.
3747         Use C style comments.
3748
3749 2003-06-13  Kaz Kojima  <kkojima@rr.iij4u.or.jp>
3750
3751         * Makefile.am: Add SHmedia support.  Fix a typo of SH support.
3752         * Makefile.in: Regenerate.
3753         * configure.in (sh64-*-linux*, sh5*-*-linux*): Add target.
3754         * configure: Regenerate.
3755         * include/ffi.h.in: Add SHmedia support.
3756         * src/sh64/ffi.c: New file.
3757         * src/sh64/sysv.S: New file.
3758
3759 2003-05-16  Jakub Jelinek  <jakub@redhat.com>
3760
3761         * configure.in (HAVE_RO_EH_FRAME): Check whether .eh_frame section
3762         should be read-only.
3763         * configure: Rebuilt.
3764         * fficonfig.h.in: Rebuilt.
3765         * include/ffi.h.in (EH_FRAME_FLAGS): Define.
3766         * src/alpha/osf.S: Use EH_FRAME_FLAGS.
3767         * src/powerpc/linux64.S: Likewise.
3768         * src/powerpc/linux64_closure.S: Likewise.  Include ffi.h.
3769         * src/powerpc/sysv.S: Use EH_FRAME_FLAGS.  Use pcrel encoding
3770         if -fpic/-fPIC/-mrelocatable.
3771         * src/powerpc/powerpc_closure.S: Likewise.
3772         * src/sparc/v8.S: If HAVE_RO_EH_FRAME is defined, don't include
3773         #write in .eh_frame flags.
3774         * src/sparc/v9.S: Likewise.
3775         * src/x86/unix64.S: Use EH_FRAME_FLAGS.
3776         * src/x86/sysv.S: Likewise.  Use pcrel encoding if -fpic/-fPIC.
3777         * src/s390/sysv.S: Use EH_FRAME_FLAGS.  Include ffi.h.
3778
3779 2003-05-07  Jeff Sturm  <jsturm@one-point.com>
3780
3781         Fixes PR bootstrap/10656
3782         * configure.in (HAVE_AS_REGISTER_PSEUDO_OP): Test assembler
3783         support for .register pseudo-op.
3784         * src/sparc/v8.S: Use it.
3785         * fficonfig.h.in: Rebuilt.
3786         * configure: Rebuilt.
3787
3788 2003-04-18  Jakub Jelinek  <jakub@redhat.com>
3789
3790         * include/ffi.h.in (POWERPC64): Define if 64-bit.
3791         (enum ffi_abi): Add FFI_LINUX64 on POWERPC.
3792         Make it the default on POWERPC64.
3793         (FFI_TRAMPOLINE_SIZE): Define to 24 on POWERPC64.
3794         * configure.in: Change powerpc-*-linux* into powerpc*-*-linux*.
3795         * configure: Rebuilt.
3796         * src/powerpc/ffi.c (hidden): Define.
3797         (ffi_prep_args_SYSV): Renamed from
3798         ffi_prep_args.  Cast pointers to unsigned long to shut up warnings.
3799         (NUM_GPR_ARG_REGISTERS64, NUM_FPR_ARG_REGISTERS64,
3800         ASM_NEEDS_REGISTERS64): New.
3801         (ffi_prep_args64): New function.
3802         (ffi_prep_cif_machdep): Handle FFI_LINUX64 ABI.
3803         (ffi_call): Likewise.
3804         (ffi_prep_closure): Likewise.
3805         (flush_icache): Surround by #ifndef POWERPC64.
3806         (ffi_dblfl): New union type.
3807         (ffi_closure_helper_SYSV): Use it to avoid aliasing problems.
3808         (ffi_closure_helper_LINUX64): New function.
3809         * src/powerpc/ppc_closure.S: Surround whole file by #ifndef
3810         __powerpc64__.
3811         * src/powerpc/sysv.S: Likewise.
3812         (ffi_call_SYSV): Rename ffi_prep_args to ffi_prep_args_SYSV.
3813         * src/powerpc/linux64.S: New file.
3814         * src/powerpc/linux64_closure.S: New file.
3815         * Makefile.am (EXTRA_DIST): Add src/powerpc/linux64.S and
3816         src/powerpc/linux64_closure.S.
3817         (TARGET_SRC_POWERPC): Likewise.
3818
3819         * src/ffitest.c (closure_test_fn, closure_test_fn1, closure_test_fn2,
3820         closure_test_fn3): Fix result printing on big-endian 64-bit
3821         machines.
3822         (main): Print tst2_arg instead of uninitialized tst2_result.
3823
3824         * src/ffitest.c (main): Hide what closure pointer really points to
3825         from the compiler.
3826
3827 2003-04-16  Richard Earnshaw  <rearnsha@arm.com>
3828
3829         * configure.in (arm-*-netbsdelf*): Add configuration.
3830         (configure): Regenerated.
3831
3832 2003-04-04  Loren J. Rittle  <ljrittle@acm.org>
3833
3834         * include/Makefile.in: Regenerate.
3835
3836 2003-03-21  Zdenek Dvorak  <rakdver@atrey.karlin.mff.cuni.cz>
3837
3838         * libffi/include/ffi.h.in: Define X86 instead of X86_64 in 32
3839         bit mode.
3840         * libffi/src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV):
3841         Receive closure pointer through parameter, read args using
3842         __builtin_dwarf_cfa.
3843         (FFI_INIT_TRAMPOLINE): Send closure reference through eax.
3844
3845 2003-03-12  Andreas Schwab  <schwab@suse.de>
3846
3847         * configure.in: Avoid trailing /. in toolexeclibdir.
3848         * configure: Rebuilt.
3849
3850 2003-03-03  Andreas Tobler <a.tobler@schweiz.ch>
3851
3852         * src/powerpc/darwin_closure.S: Recode to fit dynamic libraries.
3853
3854 2003-02-06  Andreas Tobler <a.tobler@schweiz.ch>
3855
3856         * libffi/src/powerpc/darwin_closure.S:
3857         Fix alignement bug, allocate 8 bytes for the result.
3858         * libffi/src/powerpc/aix_closure.S:
3859         Likewise.
3860         * libffi/src/powerpc/ffi_darwin.c:
3861         Update stackframe description for aix/darwin_closure.S.
3862
3863 2003-02-06  Jakub Jelinek  <jakub@redhat.com>
3864
3865         * src/s390/ffi.c (ffi_closure_helper_SYSV): Add hidden visibility
3866         attribute.
3867
3868 2003-01-31  Christian Cornelssen  <ccorn@cs.tu-berlin.de>,
3869             Andreas Schwab  <schwab@suse.de>
3870
3871         * configure.in: Adjust command to source config-ml.in to account
3872         for changes to the libffi_basedir definition.
3873         (libffi_basedir): Remove ${srcdir} from value and include trailing
3874         slash if nonempty.
3875
3876         * configure: Regenerate.
3877
3878 2003-01-29  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3879
3880         * src/powerpc/ppc_closure.S: Recode to fit shared libs.
3881
3882 2003-01-28  Andrew Haley  <aph@redhat.com>
3883
3884         * include/ffi.h.in: Enable FFI_CLOSURES for x86_64.
3885         * src/x86/ffi64.c (ffi_prep_closure): New.
3886         (ffi_closure_UNIX64_inner): New.
3887         * src/x86/unix64.S (ffi_closure_UNIX64): New.
3888
3889 2003-01-27  Alexandre Oliva  <aoliva@redhat.com>
3890
3891         * configure.in (toolexecdir, toolexeclibdir): Set and AC_SUBST.
3892         Remove USE_LIBDIR conditional.
3893         * Makefile.am (toolexecdir, toolexeclibdir): Don't override.
3894         * Makefile.in, configure: Rebuilt.
3895
3896 2003-01027  David Edelsohn  <edelsohn@gnu.org>
3897
3898         * Makefile.am (TARGET_SRC_POWERPC_AIX): Fix typo.
3899         * Makefile.in: Regenerate.
3900
3901 2003-01-22  Andrew Haley  <aph@redhat.com>
3902
3903         * src/powerpc/darwin.S (_ffi_call_AIX): Add Augmentation size to
3904         unwind info.
3905
3906 2003-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
3907
3908         * src/powerpc/darwin.S: Add unwind info.
3909         * src/powerpc/darwin_closure.S: Likewise.
3910
3911 2003-01-14  Andrew Haley  <aph@redhat.com>
3912
3913         * src/x86/ffi64.c (ffi_prep_args): Check for void retval.
3914         (ffi_prep_cif_machdep): Likewise.
3915         * src/x86/unix64.S: Add unwind info.
3916
3917 2003-01-14  Andreas Jaeger  <aj@suse.de>
3918
3919         * src/ffitest.c (main): Only use ffi_closures if those are
3920         supported.
3921
3922 2003-01-13 Andreas Tobler <a.tobler@schweiz.ch>
3923
3924         * libffi/src/ffitest.c
3925          add closure testcases
3926
3927 2003-01-13 Kevin B. Hendricks <khendricks@ivey.uwo.ca>
3928
3929         * libffi/src/powerpc/ffi.c
3930          fix alignment bug for float (4 byte aligned iso 8 byte)
3931
3932 2003-01-09  Geoffrey Keating  <geoffk@apple.com>
3933
3934         * src/powerpc/ffi_darwin.c: Remove RCS version string.
3935         * src/powerpc/darwin.S: Remove RCS version string.
3936
3937 2003-01-03  Jeff Sturm  <jsturm@one-point.com>
3938
3939         * include/ffi.h.in: Add closure defines for SPARC, SPARC64.
3940         * src/ffitest.c (main): Use static storage for closure.
3941         * src/sparc/ffi.c (ffi_prep_closure, ffi_closure_sparc_inner): New.
3942         * src/sparc/v8.S (ffi_closure_v8): New.
3943         * src/sparc/v9.S (ffi_closure_v9): New.
3944
3945 2002-11-10  Ranjit Mathew <rmathew@hotmail.com>
3946
3947         * include/ffi.h.in: Added FFI_STDCALL ffi_type
3948           enumeration for X86_WIN32.
3949         * src/x86/win32.S: Added ffi_call_STDCALL function
3950           definition.
3951         * src/x86/ffi.c (ffi_call/ffi_raw_call): Added
3952           switch cases for recognising FFI_STDCALL and
3953           calling ffi_call_STDCALL if target is X86_WIN32.
3954         * src/ffitest.c (my_stdcall_strlen/stdcall_many):
3955           stdcall versions of the "my_strlen" and "many"
3956           test functions (for X86_WIN32).
3957           Added test cases to test stdcall invocation using
3958           these functions.
3959
3960 2002-12-02  Kaz Kojima  <kkojima@gcc.gnu.org>
3961
3962         * src/sh/sysv.S: Add DWARF2 unwind info.
3963
3964 2002-11-27  Ulrich Weigand  <uweigand@de.ibm.com>
3965
3966         * src/s390/sysv.S (.eh_frame section): Make section read-only.
3967
3968 2002-11-26  Jim Wilson  <wilson@redhat.com>
3969
3970         * src/types.c (FFI_TYPE_POINTER): Has size 8 on IA64.
3971
3972 2002-11-23  H.J. Lu <hjl@gnu.org>
3973
3974         * acinclude.m4: Add dummy AM_PROG_LIBTOOL.
3975         Include ../config/accross.m4.
3976         * aclocal.m4; Rebuild.
3977         * configure: Likewise.
3978
3979 2002-11-15  Ulrich Weigand  <uweigand@de.ibm.com>
3980
3981         * src/s390/sysv.S (.eh_frame section): Adapt to pcrel FDE encoding.
3982
3983 2002-11-11  DJ Delorie  <dj@redhat.com>
3984
3985         * configure.in: Look for common files in the right place.
3986
3987 2002-10-08  Ulrich Weigand  <uweigand@de.ibm.com>
3988
3989         * src/java_raw_api.c (ffi_java_raw_to_ptrarray): Interpret
3990         raw data as _Jv_word values, not ffi_raw.
3991         (ffi_java_ptrarray_to_raw): Likewise.
3992         (ffi_java_rvalue_to_raw): New function.
3993         (ffi_java_raw_call): Call it.
3994         (ffi_java_raw_to_rvalue): New function.
3995         (ffi_java_translate_args): Call it.
3996         * src/ffitest.c (closure_test_fn): Interpret return value
3997         as ffi_arg, not int.
3998         * src/s390/ffi.c (ffi_prep_cif_machdep): Add missing
3999         FFI_TYPE_POINTER case.
4000         (ffi_closure_helper_SYSV): Likewise.  Also, assume return
4001         values extended to word size.
4002
4003 2002-10-02  Andreas Jaeger  <aj@suse.de>
4004
4005         * src/x86/ffi64.c (ffi_prep_cif_machdep): Remove debug output.
4006
4007 2002-10-01  Bo Thorsen  <bo@smetana.suse.de>
4008
4009         * include/ffi.h.in: Fix i386 win32 compilation.
4010
4011 2002-09-30  Ulrich Weigand  <uweigand@de.ibm.com>
4012
4013         * configure.in: Add s390x-*-linux-* target.
4014         * configure: Regenerate.
4015         * include/ffi.h.in: Define S390X for s390x targets.
4016         (FFI_CLOSURES): Define for s390/s390x.
4017         (FFI_TRAMPOLINE_SIZE): Likewise.
4018         (FFI_NATIVE_RAW_API): Likewise.
4019         * src/prep_cif.c (ffi_prep_cif): Do not compute stack space for s390.
4020         * src/types.c (FFI_TYPE_POINTER): Use 8-byte pointers on s390x.
4021         * src/s390/ffi.c: Major rework of existing code.  Add support for
4022         s390x targets.  Add closure support.
4023         * src/s390/sysv.S: Likewise.
4024
4025 2002-09-29  Richard Earnshaw  <rearnsha@arm.com>
4026
4027         * src/arm/sysv.S: Fix typo.
4028
4029 2002-09-28  Richard Earnshaw  <rearnsha@arm.com>
4030
4031         * src/arm/sysv.S: If we don't have machine/asm.h and the pre-processor