Many new patches
[libffi.git] / .pc / more-openbsd / ChangeLog
1 2011-07-11  Andrew Haley  <aph@redhat.com>
2
3         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Clear icache.
4
5 2011-06-29  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
6
7         * testsuite/libffi.call/cls_double_va.c: Move PR number to comment.
8         * testsuite/libffi.call/cls_longdouble_va.c: Likewise.
9
10 2011-06-29  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
11
12         PR libffi/46660
13         * testsuite/libffi.call/cls_double_va.c: xfail dg-output on
14         mips-sgi-irix6*.
15         * testsuite/libffi.call/cls_longdouble_va.c: Likewise.
16
17 2011-06-14  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
18
19         * testsuite/libffi.call/huge_struct.c (test_large_fn): Use PRIu8,
20         PRId8 instead of %hhu, %hhd.
21         * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRId8,
22         PRIu8): Define.
23         [__sgi__] (PRId8, PRIu8): Define.
24
25 2011-04-29  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
26
27         * src/alpha/osf.S (UA_SI, FDE_ENCODING, FDE_ENCODE, FDE_ARANGE):
28         Define.
29         Use them to handle ELF vs. ECOFF differences.
30         [__osf__] (_GLOBAL__F_ffi_call_osf): Define.
31
32 2011-03-30  Timothy Wall  <twall@users.sf.net>
33
34         * src/powerpc/darwin.S: Fix unknown FDE encoding.
35         * src/powerpc/darwin_closure.S: ditto.
36
37 2011-02-25  Anthony Green  <green@moxielogic.com>
38
39         * src/powerpc/ffi.c (ffi_prep_closure_loc): Allow for more
40         32-bit ABIs.
41
42 2011-02-15  Anthony Green  <green@moxielogic.com>
43
44         * m4/ax_cc_maxopt.m4: Don't -malign-double or use -ffast-math.
45         * configure: Rebuilt.
46
47 2011-02-13  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
48
49         * configure: Regenerate.
50
51 2011-02-13  Anthony Green  <green@moxielogic.com>
52
53         * include/ffi_common.h (UNLIKELY, LIKELY): Define.
54         * src/x86/ffi64.c (UNLIKELY, LIKELY): Remove definition.
55         * src/prep_cif.c (UNLIKELY, LIKELY): Remove definition.
56
57         * src/prep_cif.c (initialize_aggregate): Convert assertion into
58         FFI_BAD_TYPEDEF return.  Initialize arg size and alignment to 0.
59
60         * src/pa/ffi.c (ffi_prep_closure_loc): Don't ASSERT ABI test,
61         just return FFI_BAD_ABI when things are wrong.
62         * src/arm/ffi.c (ffi_prep_closure_loc): Ditto.
63         * src/powerpc/ffi.c (ffi_prep_closure_loc): Ditto.
64         * src/mips/ffi.c (ffi_prep_closure_loc): Ditto.
65         * src/ia64/ffi.c (ffi_prep_closure_loc): Ditto.
66         * src/avr32/ffi.c (ffi_prep_closure_loc): Ditto.
67
68 2011-02-11  Anthony Green  <green@moxielogic.com>
69
70         * src/sparc/ffi.c (ffi_prep_closure_loc): Don't ASSERT ABI test,
71         just return FFI_BAD_ABI when things are wrong.
72
73 2011-02-09  Stuart Shelton  <srcshelton@gmail.com>
74
75         http://bugs.gentoo.org/show_bug.cgi?id=286911
76         * src/mips/ffitarget.h: Clean up error messages.
77         * src/java_raw_api.c (ffi_java_translate_args): Cast raw arg to
78         ffi_raw*.
79         * include/ffi.h.in: Add pragma for SGI compiler.
80
81 2011-02-09  Anthony Green  <green@moxielogic.com>
82
83         * configure.ac: Add powerpc64-*-darwin* support.
84
85 2011-02-09  Anthony Green <green@moxielogic.com>
86
87         * README: Mention Interix.
88
89 2011-02-09  Jonathan Callen  <abcd@gentoo.org>
90
91         * configure.ac: Add Interix to win32/cygwin/mingw case.
92         * configure: Ditto.
93         * src/closures.c: Treat Interix like Cygwin, instead of as a
94         generic win32.
95
96 2011-02-09  Anthony Green <green@moxielogic.com>
97
98         * testsuite/libffi.call/err_bad_typedef.c: Remove xfail.
99         * testsuite/libffi.call/err_bad_abi.c: Remove xfail.
100         * src/x86/ffi64.c (UNLIKELY, LIKELY): Define.
101         (ffi_prep_closure_loc): Check for bad ABI.
102         * src/prep_cif.c (UNLIKELY, LIKELY): Define.
103         (initialize_aggregate): Check for bad types.
104
105 2011-02-09  Landon Fuller <landonf@plausible.coop>
106
107         * Makefile.am (EXTRA_DIST): Add build-ios.sh, src/arm/gentramp.sh,
108         src/arm/trampoline.S.
109         (nodist_libffi_la_SOURCES): Add src/arc/trampoline.S.
110         * configure.ac (FFI_EXEC_TRAMPOLINE_TABLE): Define.
111         * src/arm/ffi.c (ffi_trampoline_table)
112         (ffi_closure_trampoline_table_page, ffi_trampoline_table_entry)
113         (FFI_TRAMPOLINE_CODELOC_CONFIG, FFI_TRAMPOLINE_CONFIG_PAGE_OFFSET)
114         (FFI_TRAMPOLINE_COUNT, ffi_trampoline_lock, ffi_trampoline_tables)
115         (ffi_trampoline_table_alloc, ffi_closure_alloc, ffi_closure_free):
116         Define for FFI_EXEC_TRAMPOLINE_TABLE case (iOS).
117         (ffi_prep_closure_loc): Handl FFI_EXEC_TRAMPOLINE_TABLE case
118         separately.
119         * src/arm/sysv.S: Handle Apple iOS host.
120         * src/closures.c: Handle FFI_EXEC_TRAMPOLINE_TABLE case.
121         * build-ios.sh: New file.
122         * fficonfig.h.in, configure, Makefile.in: Rebuilt.
123         * README: Mention ARM iOS.
124
125 2011-02-08  Oren Held  <orenhe@il.ibm.com>
126
127         * src/dlmalloc.c (_STRUCT_MALLINFO): Define in order to avoid
128         redefinition of mallinfo on HP-UX.
129
130 2011-02-08  Ginn Chen  <ginn.chen@oracle.com>
131
132         * src/sparc/ffi.c (ffi_call): Make compatible with Solaris Studio
133         aggregate return ABI.  Flush cache.
134         (ffi_prep_closure_loc): Flush cache.
135
136 2011-02-11  Anthony Green  <green@moxielogic.com>
137
138         From Tom Honermann <tom.honermann@oracle.com>:
139         * src/powerpc/aix.S (ffi_call_AIX): Support for xlc toolchain on
140         AIX.  Declare .ffi_prep_args.  Insert nops after branch
141         instructions so that the AIX linker can insert TOC reload
142         instructions.
143         * src/powerpc/aix_closure.S: Declare .ffi_closure_helper_DARWIN.
144
145 2011-02-08  Ed  <ed@kdtc.net>
146
147         * src/powerpc/asm.h: Fix grammar nit in comment.
148
149 2011-02-08  Uli Link  <ul.mcamafia@linkitup.de>
150
151         * include/ffi.h.in (FFI_64_BIT_MAX): Define and use.
152
153 2011-02-09  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
154
155         PR libffi/46661
156         * testsuite/libffi.call/cls_pointer.c (main): Cast void * to
157         uintptr_t first.
158         * testsuite/libffi.call/cls_pointer_stack.c (main): Likewise.
159
160 2011-02-08  Rafael Avila de Espindola  <respindola@mozilla.com>
161
162         * configure.ac: Fix x86 test for pc related relocs.
163         * configure: Rebuilt.
164
165 2011-02-07  Joel Sherrill <joel.sherrill@oarcorp.com>
166
167         * libffi/src/m68k/ffi.c: Add RTEMS support for cache flushing.
168         Handle case when CPU variant does not have long double support.
169         * libffi/src/m68k/sysv.S: Add support for mc68000, Coldfire,
170         and cores with soft floating point.
171
172 2011-02-07  Joel Sherrill <joel.sherrill@oarcorp.com>
173
174         * configure.ac: Add mips*-*-rtems* support.
175         * configure: Regenerate.
176         * src/mips/ffitarget.h: Ensure needed constants are available
177         for targets which do not have sgidefs.h.
178
179 2011-01-26  Dave Korn  <dave.korn.cygwin@gmail.com>
180
181         PR target/40125
182         * configure.ac (AM_LTLDFLAGS): Add -bindir option for windows DLLs.
183         * configure: Regenerate.
184
185 2010-12-18  Iain Sandoe  <iains@gcc.gnu.org>
186
187         PR libffi/29152
188         PR libffi/42378
189         * src/powerpc/darwin_closure.S: Provide Darwin64 implementation,
190         update comments.
191         * src/powerpc/ffitarget.h (POWERPC_DARWIN64): New,
192         (FFI_TRAMPOLINE_SIZE): Update for Darwin64.
193         * src/powerpc/darwin.S: Provide Darwin64 implementation,
194         update comments.
195         * src/powerpc/ffi_darwin.c: Likewise.
196
197 2010-12-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
198
199         * configure.ac (libffi_cv_as_ascii_pseudo_op): Use double
200         backslashes.
201         (libffi_cv_as_string_pseudo_op): Likewise.
202         * configure: Regenerate.
203
204 2010-12-03  Chung-Lin Tang  <cltang@codesourcery.com>
205
206         * src/arm/sysv.S (ffi_closure_SYSV): Add UNWIND to .pad directive.
207         (ffi_closure_VFP): Same.
208         (ffi_call_VFP): Move down to before ffi_closure_VFP. Add '.fpu vfp'
209         directive.
210
211 2010-12-01  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
212
213         * testsuite/libffi.call/ffitest.h [__sgi] (PRId64, PRIu64): Define.
214         (PRIuPTR): Define.
215
216 2010-11-29  Richard Henderson  <rth@redhat.com>
217             Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
218
219         * src/x86/sysv.S (FDE_ENCODING, FDE_ENCODE): Define.
220         (.eh_frame): Use FDE_ENCODING.
221         (.LASFDE1, .LASFDE2, LASFDE3): Simplify with FDE_ENCODE.
222
223 2010-11-22  Jacek Caban <jacek@codeweavers.com>
224
225         * configure.ac: Check for symbol underscores on mingw-w64.
226         * configure: Rebuilt.
227         * src/x86/win64.S: Correctly access extern symbols in respect to
228         underscores.
229
230 2010-11-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
231
232         * testsuite/lib/libffi-dg.exp: Rename ...
233         * testsuite/lib/libffi.exp: ... to this.
234         * libffi/testsuite/libffi.call/call.exp: Don't load libffi-dg.exp.
235         * libffi/testsuite/libffi.special/special.exp: Likewise.
236
237 2010-10-28  Chung-Lin Tang  <cltang@codesourcery.com>
238
239         * src/arm/ffi.c (ffi_prep_args): Add VFP register argument handling
240         code, new parameter, and return value. Update comments.
241         (ffi_prep_cif_machdep): Add case for VFP struct return values. Add
242         call to layout_vfp_args().
243         (ffi_call_SYSV): Update declaration.
244         (ffi_call_VFP): New declaration.
245         (ffi_call): Add VFP struct return conditions. Call ffi_call_VFP()
246         when ABI is FFI_VFP.
247         (ffi_closure_VFP): New declaration.
248         (ffi_closure_SYSV_inner): Add new vfp_args parameter, update call to
249         ffi_prep_incoming_args_SYSV().
250         (ffi_prep_incoming_args_SYSV): Update parameters. Add VFP argument
251         case handling.
252         (ffi_prep_closure_loc): Pass ffi_closure_VFP to trampoline
253         construction under VFP hard-float.
254         (rec_vfp_type_p): New function.
255         (vfp_type_p): Same.
256         (place_vfp_arg): Same.
257         (layout_vfp_args): Same.
258         * src/arm/ffitarget.h (ffi_abi): Add FFI_VFP. Define FFI_DEFAULT_ABI
259         based on __ARM_PCS_VFP.
260         (FFI_EXTRA_CIF_FIELDS): Define for adding VFP hard-float specific
261         fields.
262         (FFI_TYPE_STRUCT_VFP_FLOAT): Define internally used type code.
263         (FFI_TYPE_STRUCT_VFP_DOUBLE): Same.
264         * src/arm/sysv.S (ffi_call_SYSV): Change call of ffi_prep_args() to
265         direct call. Move function pointer load upwards.
266         (ffi_call_VFP): New function.
267         (ffi_closure_VFP): Same.
268
269         * testsuite/lib/libffi-dg.exp (check-flags): New function.
270         (dg-skip-if): New function.
271         * testsuite/libffi.call/cls_double_va.c: Skip if target is arm*-*-*
272         and compiler options include -mfloat-abi=hard.
273         * testsuite/libffi.call/cls_longdouble_va.c: Same.
274
275 2010-10-01  Jakub Jelinek  <jakub@redhat.com>
276
277         PR libffi/45677
278         * src/x86/ffi64.c (ffi_prep_cif_machdep): Ensure cif->bytes is
279         a multiple of 8.
280         * testsuite/libffi.call/many2.c: New test.
281
282 2010-08-20  Mark Wielaard  <mjw@redhat.com>
283
284         * src/closures.c (open_temp_exec_file_mnt): Check if getmntent_r
285         returns NULL.
286
287 2010-08-09  Andreas Tobler  <andreast@fgznet.ch>
288
289         * configure.ac: Add target powerpc64-*-freebsd*.
290         * configure: Regenerate.
291         * testsuite/libffi.call/cls_align_longdouble_split.c: Pass
292         -mlong-double-128 only to linux targets.
293         * testsuite/libffi.call/cls_align_longdouble_split2.c: Likewise.
294         * testsuite/libffi.call/cls_longdouble.c: Likewise.
295         * testsuite/libffi.call/huge_struct.c: Likewise.
296
297 2010-08-05  Dan Witte  <dwitte@mozilla.com>
298
299         * Makefile.am: Pass FFI_DEBUG define to msvcc.sh for linking to the
300         debug CRT when --enable-debug is given.
301         * configure.ac: Define it.
302         * msvcc.sh: Translate -g and -DFFI_DEBUG appropriately.
303
304 2010-08-04  Dan Witte  <dwitte@mozilla.com>
305
306         * src/x86/ffitarget.h: Add X86_ANY define for all x86/x86_64
307         platforms.
308         * src/x86/ffi.c: Remove redundant ifdef checks.
309         * src/prep_cif.c: Push stack space computation into src/x86/ffi.c
310         for X86_ANY so return value space doesn't get added twice.
311
312 2010-08-03  Neil Rashbrooke <neil@parkwaycc.co.uk>
313
314         * msvcc.sh: Don't pass -safeseh to ml64 because behavior is buggy.
315
316 2010-07-22  Dan Witte  <dwitte@mozilla.com>
317
318         * src/*/ffitarget.h: Make FFI_LAST_ABI one past the last valid ABI.
319         * src/prep_cif.c: Fix ABI assertion.
320         * src/cris/ffi.c: Ditto.
321
322 2010-07-10  Evan Phoenix  <evan@fallingsnow.net>
323
324         * src/closures.c (selinux_enabled_check): Fix strncmp usage bug.
325
326 2010-07-07  Dan HorĂ¡k <dan@danny.cz>
327
328         * include/ffi.h.in: Protect #define with #ifndef.
329         * src/powerpc/ffitarget.h: Ditto.
330         * src/s390/ffitarget.h: Ditto.
331         * src/sparc/ffitarget.h: Ditto.
332
333 2010-07-07   Neil Roberts <neil@linux.intel.com>
334
335         * src/x86/sysv.S (ffi_call_SYSV): Align the stack pointer to
336         16-bytes.
337
338 2010-07-02  Jakub Jelinek  <jakub@redhat.com>
339
340         * Makefile.am (AM_MAKEFLAGS): Pass also mandir to submakes.
341         * Makefile.in: Regenerated.
342
343 2010-05-19  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
344
345         * configure.ac (libffi_cv_as_x86_pcrel): Check for illegal in as
346         output, too.
347         (libffi_cv_as_ascii_pseudo_op): Check for .ascii.
348         (libffi_cv_as_string_pseudo_op): Check for .string.
349         * configure: Regenerate.
350         * fficonfig.h.in: Regenerate.
351         * src/x86/sysv.S (.eh_frame): Use .ascii, .string or error.
352
353 2010-05-11  Dan Witte  <dwitte@mozilla.com>
354
355         * doc/libffi.tex: Document previous change.
356
357 2010-05-11  Makoto Kato <m_kato@ga2.so-net.ne.jp>
358
359         * src/x86/ffi.c (ffi_call): Don't copy structs passed by value.
360
361 2010-05-05  Michael Kohler <michaelkohler@live.com>
362
363         * src/dlmalloc.c (dlfree): Fix spelling.
364         * src/ia64/ffi.c (ffi_prep_cif_machdep): Ditto.
365         * configure.ac: Ditto.
366         * configure: Rebuilt.
367
368 2010-04-13  Dan Witte  <dwitte@mozilla.com>
369
370         * msvcc.sh: Build with -W3 instead of -Wall.
371         * src/powerpc/ffi_darwin.c: Remove build warnings.
372         * src/x86/ffi.c: Ditto.
373         * src/x86/ffitarget.h: Ditto.
374
375 2010-04-12  Dan Witte  <dwitte@mozilla.com>
376             Walter Meinl <wuno@lsvw.de>
377
378         * configure.ac: Add OS/2 support.
379         * configure: Rebuilt.
380         * src/closures.c: Ditto.
381         * src/dlmalloc.c: Ditto.
382         * src/x86/win32.S: Ditto.
383
384 2010-04-07  Jakub Jelinek  <jakub@redhat.com>
385
386         * testsuite/libffi.call/err_bad_abi.c: Remove unused args variable.
387
388 2010-04-02  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
389
390         * Makefile.in: Regenerate.
391         * aclocal.m4: Regenerate.
392         * include/Makefile.in: Regenerate.
393         * man/Makefile.in: Regenerate.
394         * testsuite/Makefile.in: Regenerate.
395
396 2010-03-30  Dan Witte  <dwitte@mozilla.com>
397
398         * msvcc.sh: Disable build warnings.
399         * README (tested): Clarify windows build procedure.
400
401 2010-03-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
402
403         * configure.ac (libffi_cv_as_x86_64_unwind_section_type): New test.
404         * configure: Regenerate.
405         * fficonfig.h.in: Regenerate.
406         * libffi/src/x86/unix64.S (.eh_frame)
407         [HAVE_AS_X86_64_UNWIND_SECTION_TYPE]: Use @unwind section type.
408
409 2010-03-14  Matthias Klose  <doko@ubuntu.com>
410
411         * src/x86/ffi64.c: Fix typo in comment.
412         * src/x86/ffi.c: Use /* ... */ comment style.
413
414 2010-02-24  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
415
416         * doc/libffi.texi (The Closure API): Fix typo.
417         * doc/libffi.info: Remove.
418
419 2010-02-15  Matthias Klose  <doko@ubuntu.com>
420
421         * src/arm/sysv.S (__ARM_ARCH__): Define for processor
422         __ARM_ARCH_7EM__.
423
424 2010-01-15  Anthony Green  <green@redhat.com>
425
426         * README: Add notes on building with Microsoft Visual C++.
427
428 2010-01-15  Daniel Witte  <dwitte@mozilla.com>
429
430         * msvcc.sh: New file.
431
432         * src/x86/win32.S: Port assembly routines to MSVC and #ifdef.
433         * src/x86/ffi.c: Tweak function declaration and remove excess
434         parens.
435         * include/ffi.h.in: Add __declspec(align(8)) to typedef struct
436         ffi_closure.
437
438         * src/x86/ffi.c: Merge ffi_call_SYSV and ffi_call_STDCALL into new
439         function ffi_call_win32 on X86_WIN32.
440         * src/x86/win32.S (ffi_call_SYSV): Rename to ffi_call_win32.
441         (ffi_call_STDCALL): Remove.
442
443         * src/prep_cif.c (ffi_prep_cif): Move stack space allocation code
444         to ffi_prep_cif_machdep for x86.
445         * src/x86/ffi.c (ffi_prep_cif_machdep): To here.
446
447 2010-01-15  Oliver Kiddle  <okiddle@yahoo.co.uk>
448
449         * src/x86/ffitarget.h (ffi_abi): Check for __i386 and __amd64 for
450         Sun Studio compiler compatibility.
451
452 2010-01-12  Conrad Irwin <conrad.irwin@gmail.com>
453
454         * doc/libffi.texi: Add closure example.
455
456 2010-01-07  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
457
458         PR libffi/40701
459         * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRIdLL,
460         PRIuLL, PRId64, PRIu64, PRIuPTR): Define.
461         * testsuite/libffi.call/cls_align_sint64.c: Add -Wno-format on
462         alpha*-dec-osf*.
463         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
464         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
465         * testsuite/libffi.call/return_ll1.c: Likewise.
466         * testsuite/libffi.call/stret_medium2.c: Likewise.
467         * testsuite/libffi.special/ffitestcxx.h (allocate_mmap): Cast
468         MAP_FAILED to char *.
469
470 2010-01-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
471
472         * src/mips/n32.S: Use .abicalls and .eh_frame with __GNUC__.
473
474 2009-12-31  Anthony Green  <green@redhat.com>
475
476         * README: Update for libffi 3.0.9.
477
478 2009-12-27  Matthias Klose  <doko@ubuntu.com>
479
480         * configure.ac (HAVE_LONG_DOUBLE): Define for mips when
481         appropriate.
482         * configure: Rebuilt.
483
484 2009-12-26  Anthony Green  <green@redhat.com>
485
486         * testsuite/libffi.call/cls_longdouble_va.c: Mark as xfail for
487         avr32*-*-*.
488         * testsuite/libffi.call/cls_double_va.c: Ditto.
489
490 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
491
492         * testsuite/libffi.call/ffitest.h: Conditionally include stdint.h
493         and inttypes.h.
494         * testsuite/libffi.special/unwindtest.cc: Ditto.
495
496 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
497
498         * configure.ac: Add amd64-*-openbsd*.
499         * configure: Rebuilt.
500         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Link
501         openbsd programs with -lpthread.
502
503 2009-12-26  Anthony Green  <green@redhat.com>
504
505         * testsuite/libffi.call/cls_double_va.c,
506         testsuite/libffi.call/cls_longdouble.c,
507         testsuite/libffi.call/cls_longdouble_va.c,
508         testsuite/libffi.call/cls_pointer.c,
509         testsuite/libffi.call/cls_pointer_stack.c: Remove xfail for
510         mips*-*-* and arm*-*-*.
511         * testsuite/libffi.call/cls_align_longdouble_split.c,
512         testsuite/libffi.call/cls_align_longdouble_split2.c,
513         testsuite/libffi.call/stret_medium2.c,
514         testsuite/libffi.call/stret_medium.c,
515         testsuite/libffi.call/stret_large.c,
516         testsuite/libffi.call/stret_large2.c: Remove xfail for arm*-*-*.
517
518 2009-12-31  Kay Tietz  <ktietz70@googlemail.com>
519
520         * testsuite/libffi.call/ffitest.h,
521         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRuLL): Fix
522         definitions.
523
524 2009-12-31  Carlo Bramini  <carlo.bramix@libero.it>
525
526         * configure.ac (AM_LTLDFLAGS): Define for windows hosts.
527         * Makefile.am (libffi_la_LDFLAGS): Add AM_LTLDFLAGS.
528         * configure: Rebuilt.
529         * Makefile.in: Rebuilt.
530
531 2009-12-31  Anthony Green  <green@redhat.com>
532             Blake Chaffin.
533
534         * testsuite/libffi.call/huge_struct.c: New test case from Blake
535         Chaffin @ Apple.
536
537 2009-12-28  David Edelsohn  <edelsohn@gnu.org>
538
539         * src/powerpc/ffi_darwin.c (ffi_prep_args): Copy abi and nargs to
540         local variables.
541         (aix_adjust_aggregate_sizes): New function.
542         (ffi_prep_cif_machdep): Call it.
543
544 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
545
546         * configure.ac: Define FFI_MMAP_EXEC_WRIT for the given targets.
547         * configure: Regenerate.
548         * fficonfig.h.in: Likewise.
549         * src/closures.c: Remove the FFI_MMAP_EXEC_WRIT definition for
550         Solaris/x86.
551
552 2009-12-26  Andreas Schwab  <schwab@linux-m68k.org>
553
554         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Advance intarg_count
555         when a float arguments is passed in memory.
556         (ffi_closure_helper_SYSV): Mark general registers as used up when
557         a 64bit or soft-float long double argument is passed in memory.
558
559 2009-12-25  Matthias Klose  <doko@ubuntu.com>
560
561         * man/ffi_call.3: Fix #include in examples.
562         * doc/libffi.texi: Add dircategory.
563
564 2009-12-25  Frank Everdij <f.p.x.everdij@tudelft.nl>
565
566         * include/ffi.h.in: Placed '__GNUC__' ifdef around
567         '__attribute__((aligned(8)))' in ffi_closure, fixes compile for
568         IRIX MIPSPro c99.
569         * include/ffi_common.h: Added '__sgi' define to non
570         '__attribute__((__mode__()))' integer typedefs.
571         * src/mips/ffi.c (ffi_call, ffi_closure_mips_inner_O32,
572         ffi_closure_mips_inner_N32): Added 'defined(_MIPSEB)' to BE check.
573         (ffi_closure_mips_inner_O32, ffi_closure_mips_inner_N32): Added
574         FFI_LONGDOUBLE support and alignment(N32 only).
575         * src/mips/ffitarget.h: Corrected '#include <sgidefs.h>' for IRIX and
576         fixed non '__attribute__((__mode__()))' integer typedefs.
577         * src/mips/n32.S: Put '#ifdef linux' around '.abicalls' and '.eh_frame'
578         since they are Linux/GNU Assembler specific.
579
580 2009-12-25  Bradley Smith  <brad@brad-smith.co.uk>
581
582         * configure.ac, Makefile.am, src/avr32/ffi.c,
583         src/avr32/ffitarget.h,
584         src/avr32/sysv.S: Add AVR32 port.
585         * configure, Makefile.in: Rebuilt.
586
587 2009-12-21  Andreas Tobler  <a.tobler@schweiz.org>
588
589         * configure.ac: Make i?86 build on FreeBSD and OpenBSD.
590         * configure: Regenerate.
591
592 2009-12-15  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
593
594         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on PA HP-UX.
595
596 2009-12-13  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
597
598         * src/pa/ffi.c (ffi_closure_inner_pa32): Handle FFI_TYPE_LONGDOUBLE
599         type on HP-UX.
600
601 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
602
603         * src/sparc/ffi.c (ffi_closure_sparc_inner_v9): Properly align 'long
604         double' arguments.
605
606 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
607
608         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on Solaris < 10.
609
610 2009-12-10  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
611
612         PR libffi/40700
613         * src/closures.c [X86_64 && __sun__ && __svr4__]
614         (FFI_MMAP_EXEC_WRIT): Define.
615
616 2009-12-08  David Daney  <ddaney@caviumnetworks.com>
617
618         * testsuite/libffi.call/stret_medium.c: Remove xfail for mips*-*-*
619         * testsuite/libffi.call/cls_align_longdouble_split2.c: Same.
620         * testsuite/libffi.call/stret_large.c: Same.
621         * testsuite/libffi.call/cls_align_longdouble_split.c: Same.
622         * testsuite/libffi.call/stret_large2.c: Same.
623         * testsuite/libffi.call/stret_medium2.c: Same.
624
625 2009-12-07  David Edelsohn  <edelsohn@gnu.org>
626
627         * src/powerpc/aix_closure.S (libffi_closure_ASM): Fix tablejump
628         typo.
629
630 2009-12-05  David Edelsohn  <edelsohn@gnu.org>
631
632         * src/powerpc/aix.S: Update AIX32 code to be consistent with AIX64
633         code.
634         * src/powerpc/aix_closure.S: Same.
635
636 2009-12-05  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
637
638         * Makefile.in: Regenerate.
639         * configure: Regenerate.
640         * include/Makefile.in: Regenerate.
641         * man/Makefile.in: Regenerate.
642         * testsuite/Makefile.in: Regenerate.
643
644 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
645
646         * src/powerpc/aix_closure.S: Reorganize 64-bit code to match
647         linux64_closure.S.
648
649 2009-12-04  Uros Bizjak  <ubizjak@gmail.com>
650
651         PR libffi/41908
652         * src/x86/ffi64.c (classify_argument): Update from
653         gcc/config/i386/i386.c.
654         (ffi_closure_unix64_inner): Do not use the address of two consecutive
655         SSE registers directly.
656         * testsuite/libffi.call/cls_dbls_struct.c (main): Remove xfail
657         for x86_64 linux targets.
658
659 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
660
661         * src/powerpc/ffi_darwin.c (ffi_closure_helper_DARWIN): Increment
662         pfr for long double split between fpr13 and stack.
663
664 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
665
666         * src/powerpc/ffi_darwin.c (ffi_prep_args): Increment next_arg and
667         fparg_count twice for long double.
668
669 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
670
671         PR libffi/42243
672         * src/powerpc/ffi_darwin.c (ffi_prep_args): Remove extra parentheses.
673
674 2009-12-03  Uros Bizjak  <ubizjak@gmail.com>
675
676         * testsuite/libffi.call/cls_longdouble_va.c (main): Fix format string.
677         Remove xfails for x86 linux targets.
678
679 2009-12-02  David Edelsohn  <edelsohn@gnu.org>
680
681         * src/powerpc/ffi_darwin.c (ffi_prep_args): Fix typo in INT64
682         case.
683
684 2009-12-01  David Edelsohn  <edelsohn@gnu.org>
685
686         * src/powerpc/aix.S (ffi_call_AIX): Convert to more standard
687         register usage.  Call ffi_prep_args directly.  Add long double
688         return value support.
689         * src/powerpc/ffi_darwin.c (ffi_prep_args): Double arg increment
690         applies to FFI_TYPE_DOUBLE.  Correct fpr_base increment typo.
691         Separate FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases.
692         (ffi_prep_cif_machdep): Only 16 byte stack alignment in 64 bit
693         mode.
694         (ffi_closure_helper_DARWIN): Remove nf and ng counters.  Move temp
695         into case.
696         * src/powerpc/aix_closure.S: Maintain 16 byte stack alignment.
697         Allocate result area between params and FPRs.
698
699 2009-11-30  David Edelsohn  <edelsohn@gnu.org>
700
701         PR target/35484
702         * src/powerpc/ffitarget.h (POWERPC64): Define for PPC64 Linux and
703         AIX64.
704         * src/powerpc/aix.S: Implement AIX64 version.
705         * src/powerpc/aix_closure.S: Implement AIX64 version.
706         (ffi_closure_ASM): Use extsb, lha and displament addresses.
707         * src/powerpc/ffi_darwin.c (ffi_prep_args): Implement AIX64
708         support.
709         (ffi_prep_cif_machdep): Same.
710         (ffi_call): Same.
711         (ffi_closure_helper_DARWIN): Same.
712
713 2009-11-02  Andreas Tobler  <a.tobler@schweiz.org>
714
715         PR libffi/41908
716         * testsuite/libffi.call/testclosure.c: New test.
717
718 2009-09-28  Kai Tietz  <kai.tietz@onevision.com>
719
720         * src/x86/win64.S (_ffi_call_win64 stack): Remove for gnu
721         assembly version use of ___chkstk.
722
723 2009-09-23  Matthias Klose  <doko@ubuntu.com>
724
725         PR libffi/40242, PR libffi/41443
726         * src/arm/sysv.S (__ARM_ARCH__): Define for processors
727         __ARM_ARCH_6T2__, __ARM_ARCH_6M__, __ARM_ARCH_7__,
728         __ARM_ARCH_7A__, __ARM_ARCH_7R__, __ARM_ARCH_7M__.
729         Change the conditionals to __SOFTFP__ || __ARM_EABI__
730         for -mfloat-abi=softfp to work.
731
732 2009-09-17  Loren J. Rittle  <ljrittle@acm.org>
733
734         PR testsuite/32843 (strikes again)
735         * src/x86/ffi.c (ffi_prep_cif_machdep): Add X86_FREEBSD to
736         enable proper extension on char and short.
737
738 2009-09-15  David Daney  <ddaney@caviumnetworks.com>
739
740         * src/java_raw_api.c (ffi_java_raw_to_rvalue): Remove special
741         handling for FFI_TYPE_POINTER.
742         * src/mips/ffitarget.h (FFI_TYPE_STRUCT_D_SOFT,
743         FFI_TYPE_STRUCT_F_SOFT, FFI_TYPE_STRUCT_DD_SOFT,
744         FFI_TYPE_STRUCT_FF_SOFT, FFI_TYPE_STRUCT_FD_SOFT,
745         FFI_TYPE_STRUCT_DF_SOFT, FFI_TYPE_STRUCT_SOFT): New defines.
746         (FFI_N32_SOFT_FLOAT, FFI_N64_SOFT_FLOAT): New ffi_abi enumerations.
747         (enum ffi_abi): Set FFI_DEFAULT_ABI for soft-float.
748         * src/mips/n32.S (ffi_call_N32): Add handling for soft-float
749         structure and pointer returns.
750         (ffi_closure_N32): Add handling for pointer returns.
751         * src/mips/ffi.c (ffi_prep_args, calc_n32_struct_flags,
752         calc_n32_return_struct_flags): Handle soft-float.
753         (ffi_prep_cif_machdep):  Handle soft-float, fix pointer handling.
754         (ffi_call_N32): Declare proper argument types.
755         (ffi_call, copy_struct_N32, ffi_closure_mips_inner_N32): Handle
756         soft-float.
757
758 2009-08-24  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
759
760         * configure.ac (AC_PREREQ): Bump to 2.64.
761
762 2009-08-22  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
763
764         * Makefile.am (install-html, install-pdf): Remove.
765         * Makefile.in: Regenerate.
766
767         * Makefile.in: Regenerate.
768         * aclocal.m4: Regenerate.
769         * configure: Regenerate.
770         * fficonfig.h.in: Regenerate.
771         * include/Makefile.in: Regenerate.
772         * man/Makefile.in: Regenerate.
773         * testsuite/Makefile.in: Regenerate.
774
775 2009-07-30  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
776
777         * configure.ac (_AC_ARG_VAR_PRECIOUS): Use m4_rename_force.
778
779 2009-07-24  Dave Korn  <dave.korn.cygwin@gmail.com>
780
781         PR libffi/40807
782         * src/x86/ffi.c (ffi_prep_cif_machdep): Also use sign/zero-extending
783         return types for X86_WIN32.
784         * src/x86/win32.S (_ffi_call_SYSV): Handle omitted return types.
785         (_ffi_call_STDCALL, _ffi_closure_SYSV, _ffi_closure_raw_SYSV,
786         _ffi_closure_STDCALL): Likewise.
787
788         * src/closures.c (is_selinux_enabled): Define to const 0 for Cygwin.
789         (dlmmap, dlmunmap): Also use these functions on Cygwin.
790
791 2009-07-11  Richard Sandiford  <rdsandiford@googlemail.com>
792
793         PR testsuite/40699
794         PR testsuite/40707
795         PR testsuite/40709
796         * testsuite/lib/libffi-dg.exp: Revert 2009-07-02, 2009-07-01 and
797         2009-06-30 commits.
798
799 2009-07-01  Richard Sandiford  <r.sandiford@uk.ibm.com>
800
801         * testsuite/lib/libffi-dg.exp (libffi-init): Set ld_library_path
802         to "" before adding paths.  (This reinstates an assignment that
803         was removed by my 2009-06-30 commit, but changes the initial
804         value from "." to "".)
805
806 2009-07-01  H.J. Lu  <hongjiu.lu@intel.com>
807
808         PR testsuite/40601
809         * testsuite/lib/libffi-dg.exp (libffi-init): Properly set
810         gccdir.  Adjust ld_library_path for gcc only if gccdir isn't
811         empty.
812
813 2009-06-30  Richard Sandiford  <r.sandiford@uk.ibm.com>
814
815         * testsuite/lib/libffi-dg.exp (libffi-init): Don't add "."
816         to ld_library_path.  Use add_path.  Add just find_libgcc_s
817         to ld_library_path, not every libgcc multilib directory.
818
819 2009-06-16  Wim Lewis  <wiml@hhhh.org>
820
821         * src/powerpc/ffi.c: Avoid clobbering cr3 and cr4, which are
822         supposed to be callee-saved.
823         * src/powerpc/sysv.S (small_struct_return_value): Fix overrun of
824         return buffer for odd-size structs.
825
826 2009-06-16  Andreas Tobler  <a.tobler@schweiz.org>
827
828         PR libffi/40444
829         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Add
830         allow_stack_execute for Darwin.
831
832 2009-06-16  Andrew Haley  <aph@redhat.com>
833
834         * configure.ac (TARGETDIR): Add missing blank lines.
835         * configure: Regenerate.
836
837 2009-06-16  Andrew Haley  <aph@redhat.com>
838
839         * testsuite/libffi.call/cls_align_sint64.c,
840         testsuite/libffi.call/cls_align_uint64.c,
841         testsuite/libffi.call/cls_longdouble_va.c,
842         testsuite/libffi.call/cls_ulonglong.c,
843         testsuite/libffi.call/return_ll1.c,
844         testsuite/libffi.call/stret_medium2.c: Fix printf format
845         specifiers.
846         * testsuite/libffi.call/ffitest.h,
847         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRIuLL): Define.
848
849 2009-06-15  Andrew Haley  <aph@redhat.com>
850
851         * testsuite/libffi.call/err_bad_typedef.c: xfail everywhere.
852         * testsuite/libffi.call/err_bad_abi.c: Likewise.
853
854 2009-06-12  Andrew Haley  <aph@redhat.com>
855
856         * Makefile.am: Remove info_TEXINFOS.
857
858 2009-06-12  Andrew Haley  <aph@redhat.com>
859
860         * ChangeLog.libffi: testsuite/libffi.call/cls_align_sint64.c,
861         testsuite/libffi.call/cls_align_uint64.c,
862         testsuite/libffi.call/cls_ulonglong.c,
863         testsuite/libffi.call/return_ll1.c,
864         testsuite/libffi.call/stret_medium2.c: Fix printf format
865         specifiers.
866         testsuite/libffi.special/unwindtest.cc: include stdint.h.
867
868 2009-06-11  Timothy Wall  <twall@users.sf.net>
869
870         * Makefile.am,
871         configure.ac,
872         include/ffi.h.in,
873         include/ffi_common.h,
874         src/closures.c,
875         src/dlmalloc.c,
876         src/x86/ffi.c,
877         src/x86/ffitarget.h,
878         src/x86/win64.S (new),
879         README: Added win64 support (mingw or MSVC)
880         * Makefile.in,
881         include/Makefile.in,
882         man/Makefile.in,
883         testsuite/Makefile.in,
884         configure,
885         aclocal.m4: Regenerated
886         * ltcf-c.sh: properly escape cygwin/w32 path
887         * man/ffi_call.3: Clarify size requirements for return value.
888         * src/x86/ffi64.c: Fix filename in comment.
889         * src/x86/win32.S: Remove unused extern.
890
891         * testsuite/libffi.call/closure_fn0.c,
892         testsuite/libffi.call/closure_fn1.c,
893         testsuite/libffi.call/closure_fn2.c,
894         testsuite/libffi.call/closure_fn3.c,
895         testsuite/libffi.call/closure_fn4.c,
896         testsuite/libffi.call/closure_fn5.c,
897         testsuite/libffi.call/closure_fn6.c,
898         testsuite/libffi.call/closure_stdcall.c,
899         testsuite/libffi.call/cls_12byte.c,
900         testsuite/libffi.call/cls_16byte.c,
901         testsuite/libffi.call/cls_18byte.c,
902         testsuite/libffi.call/cls_19byte.c,
903         testsuite/libffi.call/cls_1_1byte.c,
904         testsuite/libffi.call/cls_20byte.c,
905         testsuite/libffi.call/cls_20byte1.c,
906         testsuite/libffi.call/cls_24byte.c,
907         testsuite/libffi.call/cls_2byte.c,
908         testsuite/libffi.call/cls_3_1byte.c,
909         testsuite/libffi.call/cls_3byte1.c,
910         testsuite/libffi.call/cls_3byte2.c,
911         testsuite/libffi.call/cls_4_1byte.c,
912         testsuite/libffi.call/cls_4byte.c,
913         testsuite/libffi.call/cls_5_1_byte.c,
914         testsuite/libffi.call/cls_5byte.c,
915         testsuite/libffi.call/cls_64byte.c,
916         testsuite/libffi.call/cls_6_1_byte.c,
917         testsuite/libffi.call/cls_6byte.c,
918         testsuite/libffi.call/cls_7_1_byte.c,
919         testsuite/libffi.call/cls_7byte.c,
920         testsuite/libffi.call/cls_8byte.c,
921         testsuite/libffi.call/cls_9byte1.c,
922         testsuite/libffi.call/cls_9byte2.c,
923         testsuite/libffi.call/cls_align_double.c,
924         testsuite/libffi.call/cls_align_float.c,
925         testsuite/libffi.call/cls_align_longdouble.c,
926         testsuite/libffi.call/cls_align_longdouble_split.c,
927         testsuite/libffi.call/cls_align_longdouble_split2.c,
928         testsuite/libffi.call/cls_align_pointer.c,
929         testsuite/libffi.call/cls_align_sint16.c,
930         testsuite/libffi.call/cls_align_sint32.c,
931         testsuite/libffi.call/cls_align_sint64.c,
932         testsuite/libffi.call/cls_align_uint16.c,
933         testsuite/libffi.call/cls_align_uint32.c,
934         testsuite/libffi.call/cls_align_uint64.c,
935         testsuite/libffi.call/cls_dbls_struct.c,
936         testsuite/libffi.call/cls_double.c,
937         testsuite/libffi.call/cls_double_va.c,
938         testsuite/libffi.call/cls_float.c,
939         testsuite/libffi.call/cls_longdouble.c,
940         testsuite/libffi.call/cls_longdouble_va.c,
941         testsuite/libffi.call/cls_multi_schar.c,
942         testsuite/libffi.call/cls_multi_sshort.c,
943         testsuite/libffi.call/cls_multi_sshortchar.c,
944         testsuite/libffi.call/cls_multi_uchar.c,
945         testsuite/libffi.call/cls_multi_ushort.c,
946         testsuite/libffi.call/cls_multi_ushortchar.c,
947         testsuite/libffi.call/cls_pointer.c,
948         testsuite/libffi.call/cls_pointer_stack.c,
949         testsuite/libffi.call/cls_schar.c,
950         testsuite/libffi.call/cls_sint.c,
951         testsuite/libffi.call/cls_sshort.c,
952         testsuite/libffi.call/cls_uchar.c,
953         testsuite/libffi.call/cls_uint.c,
954         testsuite/libffi.call/cls_ulonglong.c,
955         testsuite/libffi.call/cls_ushort.c,
956         testsuite/libffi.call/err_bad_abi.c,
957         testsuite/libffi.call/err_bad_typedef.c,
958         testsuite/libffi.call/float2.c,
959         testsuite/libffi.call/huge_struct.c,
960         testsuite/libffi.call/nested_struct.c,
961         testsuite/libffi.call/nested_struct1.c,
962         testsuite/libffi.call/nested_struct10.c,
963         testsuite/libffi.call/nested_struct2.c,
964         testsuite/libffi.call/nested_struct3.c,
965         testsuite/libffi.call/nested_struct4.c,
966         testsuite/libffi.call/nested_struct5.c,
967         testsuite/libffi.call/nested_struct6.c,
968         testsuite/libffi.call/nested_struct7.c,
969         testsuite/libffi.call/nested_struct8.c,
970         testsuite/libffi.call/nested_struct9.c,
971         testsuite/libffi.call/problem1.c,
972         testsuite/libffi.call/return_ldl.c,
973         testsuite/libffi.call/return_ll1.c,
974         testsuite/libffi.call/stret_large.c,
975         testsuite/libffi.call/stret_large2.c,
976         testsuite/libffi.call/stret_medium.c,
977         testsuite/libffi.call/stret_medium2.c,
978         testsuite/libffi.special/unwindtest.cc: use ffi_closure_alloc instead
979         of checking for MMAP.  Use intptr_t instead of long casts.
980
981 2009-06-11  Kaz Kojima  <kkojima@gcc.gnu.org>
982
983         * testsuite/libffi.call/cls_longdouble_va.c: Add xfail sh*-*-linux-*.
984         * testsuite/libffi.call/err_bad_abi.c: Add xfail sh*-*-*.
985         * testsuite/libffi.call/err_bad_typedef.c: Likewise.
986
987 2009-06-09  Andrew Haley  <aph@redhat.com>
988
989         * src/x86/freebsd.S: Add missing file.
990
991 2009-06-08  Andrew Haley  <aph@redhat.com>
992
993         Import from libffi 3.0.8:
994
995         * doc/libffi.texi: New file.
996         * doc/libffi.info: Likewise.
997         * doc/stamp-vti: Likewise.
998         * man/Makefile.am: New file.
999         * man/ffi_call.3: New file.
1000
1001         * Makefile.am (EXTRA_DIST): Add src/x86/darwin64.S,
1002         src/dlmalloc.c.
1003         (nodist_libffi_la_SOURCES): Add X86_FREEBSD.
1004
1005         * configure.ac: Bump version to 3.0.8.
1006         parisc*-*-linux*: Add.
1007         i386-*-freebsd* | i386-*-openbsd*: Add.
1008         powerpc-*-beos*: Add.
1009         AM_CONDITIONAL X86_FREEBSD: Add.
1010         AC_CONFIG_FILES: Add man/Makefile.
1011
1012         * include/ffi.h.in (FFI_FN): Change void (*)() to void (*)(void).
1013
1014 2009-06-08  Andrew Haley  <aph@redhat.com>
1015
1016         * README: Import from libffi 3.0.8.
1017
1018 2009-06-08  Andrew Haley  <aph@redhat.com>
1019
1020         * testsuite/libffi.call/err_bad_abi.c: Add xfails.
1021         * testsuite/libffi.call/cls_longdouble_va.c: Add xfails.
1022         * testsuite/libffi.call/cls_dbls_struct.c: Add xfail x86_64-*-linux-*.
1023         * testsuite/libffi.call/err_bad_typedef.c: Add xfails.
1024
1025         * testsuite/libffi.call/stret_medium2.c: Add __UNUSED__ to args.
1026         * testsuite/libffi.call/stret_medium.c: Likewise.
1027         * testsuite/libffi.call/stret_large2.c: Likewise.
1028         * testsuite/libffi.call/stret_large.c:  Likewise.
1029
1030 2008-12-26  Timothy Wall  <twall@users.sf.net>
1031
1032         * testsuite/libffi.call/cls_longdouble.c,
1033         testsuite/libffi.call/cls_longdouble_va.c,
1034         testsuite/libffi.call/cls_align_longdouble.c,
1035         testsuite/libffi.call/cls_align_longdouble_split.c,
1036         testsuite/libffi.call/cls_align_longdouble_split2.c: mark expected
1037         failures on x86_64 cygwin/mingw.
1038
1039 2008-12-22  Timothy Wall  <twall@users.sf.net>
1040
1041         * testsuite/libffi.call/closure_fn0.c,
1042         testsuite/libffi.call/closure_fn1.c,
1043         testsuite/libffi.call/closure_fn2.c,
1044         testsuite/libffi.call/closure_fn3.c,
1045         testsuite/libffi.call/closure_fn4.c,
1046         testsuite/libffi.call/closure_fn5.c,
1047         testsuite/libffi.call/closure_fn6.c,
1048         testsuite/libffi.call/closure_loc_fn0.c,
1049         testsuite/libffi.call/closure_stdcall.c,
1050         testsuite/libffi.call/cls_align_pointer.c,
1051         testsuite/libffi.call/cls_pointer.c,
1052         testsuite/libffi.call/cls_pointer_stack.c: use portable cast from
1053         pointer to integer (intptr_t).
1054         * testsuite/libffi.call/cls_longdouble.c: disable for win64.
1055
1056 2008-07-24  Anthony Green  <green@redhat.com>
1057
1058         * testsuite/libffi.call/cls_dbls_struct.c,
1059         testsuite/libffi.call/cls_double_va.c,
1060         testsuite/libffi.call/cls_longdouble.c,
1061         testsuite/libffi.call/cls_longdouble_va.c,
1062         testsuite/libffi.call/cls_pointer.c,
1063         testsuite/libffi.call/cls_pointer_stack.c,
1064         testsuite/libffi.call/err_bad_abi.c: Clean up failures from
1065         compiler warnings.
1066
1067 2008-03-04  Anthony Green  <green@redhat.com>
1068             Blake Chaffin
1069             hos@tamanegi.org
1070
1071         * testsuite/libffi.call/cls_align_longdouble_split2.c
1072           testsuite/libffi.call/cls_align_longdouble_split.c
1073           testsuite/libffi.call/cls_dbls_struct.c
1074           testsuite/libffi.call/cls_double_va.c
1075           testsuite/libffi.call/cls_longdouble.c
1076           testsuite/libffi.call/cls_longdouble_va.c
1077           testsuite/libffi.call/cls_pointer.c
1078           testsuite/libffi.call/cls_pointer_stack.c
1079           testsuite/libffi.call/err_bad_abi.c
1080           testsuite/libffi.call/err_bad_typedef.c
1081           testsuite/libffi.call/stret_large2.c
1082           testsuite/libffi.call/stret_large.c
1083           testsuite/libffi.call/stret_medium2.c
1084           testsuite/libffi.call/stret_medium.c: New tests from Apple.
1085
1086 2009-06-05  Andrew Haley  <aph@redhat.com>
1087
1088         * src/x86/ffitarget.h, src/x86/ffi.c: Merge stdcall changes from
1089         libffi.
1090
1091 2009-06-04  Andrew Haley  <aph@redhat.com>
1092
1093         * src/x86/ffitarget.h, src/x86/win32.S, src/x86/ffi.c: Back out
1094         stdcall changes.
1095
1096 2008-02-26  Anthony Green  <green@redhat.com>
1097             Thomas Heller  <theller@ctypes.org>
1098
1099         * src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C
1100         comment.
1101
1102 2008-02-03  Timothy Wall  <twall@users.sf.net>
1103
1104         * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return
1105           offset based on code pointer, not data pointer.
1106
1107 2008-01-31  Timothy Wall <twall@users.sf.net>
1108
1109         * testsuite/libffi.call/closure_stdcall.c: Add test for stdcall
1110         closures.
1111         * src/x86/ffitarget.h: Increase size of trampoline for stdcall
1112         closures.
1113         * src/x86/win32.S: Add assembly for stdcall closure.
1114         * src/x86/ffi.c: Initialize stdcall closure trampoline.
1115
1116 2009-06-04  Andrew Haley  <aph@redhat.com>
1117
1118         * include/ffi.h.in: Change void (*)() to void (*)(void).
1119         * src/x86/ffi.c: Likewise.
1120
1121 2009-06-04  Andrew Haley  <aph@redhat.com>
1122
1123         * src/powerpc/ppc_closure.S: Insert licence header.
1124         * src/powerpc/linux64_closure.S: Likewise.
1125         * src/m68k/sysv.S: Likewise.
1126
1127         * src/sh64/ffi.c: Change void (*)() to void (*)(void).
1128         * src/powerpc/ffi.c: Likewise.
1129         * src/powerpc/ffi_darwin.c: Likewise.
1130         * src/m32r/ffi.c: Likewise.
1131         * src/sh64/ffi.c: Likewise.
1132         * src/x86/ffi64.c: Likewise.
1133         * src/alpha/ffi.c: Likewise.
1134         * src/alpha/osf.S: Likewise.
1135         * src/frv/ffi.c: Likewise.
1136         * src/s390/ffi.c: Likewise.
1137         * src/pa/ffi.c: Likewise.
1138         * src/pa/hpux32.S: Likewise.
1139         * src/ia64/unix.S: Likewise.
1140         * src/ia64/ffi.c: Likewise.
1141         * src/sparc/ffi.c: Likewise.
1142         * src/mips/ffi.c: Likewise.
1143         * src/sh/ffi.c: Likewise.
1144
1145 2008-02-15  David Daney  <ddaney@avtrex.com>
1146
1147         * src/mips/ffi.c (USE__BUILTIN___CLEAR_CACHE):
1148         Define (conditionally), and use it to include cachectl.h.
1149         (ffi_prep_closure_loc): Fix cache flushing.
1150         * src/mips/ffitarget.h (_ABIN32, _ABI64, _ABIO32): Define.
1151
1152 2009-06-04  Andrew Haley  <aph@redhat.com>
1153
1154         include/ffi.h.in,
1155         src/arm/ffitarget.h,
1156         src/arm/ffi.c,
1157         src/arm/sysv.S,
1158         src/powerpc/ffitarget.h,
1159         src/closures.c,
1160         src/sh64/ffitarget.h,
1161         src/sh64/ffi.c,
1162         src/sh64/sysv.S,
1163         src/types.c,
1164         src/x86/ffi64.c,
1165         src/x86/ffitarget.h,
1166         src/x86/win32.S,
1167         src/x86/darwin.S,
1168         src/x86/ffi.c,
1169         src/x86/sysv.S,
1170         src/x86/unix64.S,
1171         src/alpha/ffitarget.h,
1172         src/alpha/ffi.c,
1173         src/alpha/osf.S,
1174         src/m68k/ffitarget.h,
1175         src/frv/ffitarget.h,
1176         src/frv/ffi.c,
1177         src/s390/ffitarget.h,
1178         src/s390/sysv.S,
1179         src/cris/ffitarget.h,
1180         src/pa/linux.S,
1181         src/pa/ffitarget.h,
1182         src/pa/ffi.c,
1183         src/raw_api.c,
1184         src/ia64/ffitarget.h,
1185         src/ia64/unix.S,
1186         src/ia64/ffi.c,
1187         src/ia64/ia64_flags.h,
1188         src/java_raw_api.c,
1189         src/debug.c,
1190         src/sparc/v9.S,
1191         src/sparc/ffitarget.h,
1192         src/sparc/ffi.c,
1193         src/sparc/v8.S,
1194         src/mips/ffitarget.h,
1195         src/mips/n32.S,
1196         src/mips/o32.S,
1197         src/mips/ffi.c,
1198         src/prep_cif.c,
1199         src/sh/ffitarget.h,
1200         src/sh/ffi.c,
1201         src/sh/sysv.S: Update license text.
1202
1203 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
1204
1205         * src/x86/win32.S (_ffi_closure_STDCALL):  New function.
1206         (.eh_frame):  Add FDE for it.
1207
1208 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
1209
1210         * configure.ac:  Also check if assembler supports pc-relative
1211         relocs on X86_WIN32 targets.
1212         * configure:  Regenerate.
1213         * src/x86/win32.S (ffi_prep_args):  Declare extern, not global.
1214         (_ffi_call_SYSV):  Add missing function type symbol .def and
1215         add EH markup labels.
1216         (_ffi_call_STDCALL):  Likewise.
1217         (_ffi_closure_SYSV):  Likewise.
1218         (_ffi_closure_raw_SYSV):  Likewise.
1219         (.eh_frame):  Add hand-crafted EH data.
1220
1221 2009-04-09  Jakub Jelinek  <jakub@redhat.com>
1222
1223         * testsuite/lib/libffi-dg.exp: Change copyright header to refer to
1224         version 3 of the GNU General Public License and to point readers
1225         at the COPYING3 file and the FSF's license web page.
1226         * testsuite/libffi.call/call.exp: Likewise.
1227         * testsuite/libffi.special/special.exp: Likewise.
1228
1229 2009-03-01  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1230
1231         * configure: Regenerate.
1232
1233 2008-12-18  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1234
1235         PR libffi/26048
1236         * configure.ac (HAVE_AS_X86_PCREL): New test.
1237         * configure: Regenerate.
1238         * fficonfig.h.in: Regenerate.
1239         * src/x86/sysv.S [!FFI_NO_RAW_API]: Precalculate
1240         RAW_CLOSURE_CIF_OFFSET, RAW_CLOSURE_FUN_OFFSET,
1241         RAW_CLOSURE_USER_DATA_OFFSET for the Solaris 10/x86 assembler.
1242         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
1243         * src/x86/unix64.S (.Lstore_table): Move to .text section.
1244         (.Lload_table): Likewise.
1245         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
1246
1247 2008-12-18  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1248
1249         * configure: Regenerate.
1250
1251 2008-11-21  Eric Botcazou  <ebotcazou@adacore.com>
1252
1253         * src/sparc/ffi.c (ffi_prep_cif_machdep): Add support for
1254         signed/unsigned int8/16 return values.
1255         * src/sparc/v8.S (ffi_call_v8): Likewise.
1256         (ffi_closure_v8): Likewise.
1257
1258 2008-09-26  Peter O'Gorman  <pogma@thewrittenword.com>
1259             Steve Ellcey  <sje@cup.hp.com>
1260
1261         * configure: Regenerate for new libtool.
1262         * Makefile.in: Ditto.
1263         * include/Makefile.in: Ditto.
1264         * aclocal.m4: Ditto.
1265
1266 2008-08-25  Andreas Tobler  <a.tobler@schweiz.org>
1267
1268         * src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and
1269         FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum.
1270         Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT.
1271         Adjust copyright notice.
1272         * src/powerpc/ffi.c: Add two new flags to indicate if we have one
1273         register or two register to use for FFI_SYSV structs.
1274         (ffi_prep_cif_machdep): Pass the right register flag introduced above.
1275         (ffi_closure_helper_SYSV): Fix the return type for
1276         FFI_SYSV_TYPE_SMALL_STRUCT. Comment.
1277         Adjust copyright notice.
1278
1279 2008-07-16  Kaz Kojima  <kkojima@gcc.gnu.org>
1280
1281         * src/sh/ffi.c (ffi_prep_closure_loc): Turn INSN into an unsigned
1282         int.
1283
1284 2008-06-17  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1285
1286         * configure: Regenerate.
1287         * include/Makefile.in: Regenerate.
1288         * testsuite/Makefile.in: Regenerate.
1289
1290 2008-06-07  Joseph Myers  <joseph@codesourcery.com>
1291
1292         * configure.ac (parisc*-*-linux*, powerpc-*-sysv*,
1293         powerpc-*-beos*): Remove.
1294         * configure: Regenerate.
1295
1296 2008-05-09  Julian Brown  <julian@codesourcery.com>
1297
1298         * Makefile.am (LTLDFLAGS): New.
1299         (libffi_la_LDFLAGS): Use above.
1300         * Makefile.in: Regenerate.
1301
1302 2008-04-18  Paolo Bonzini  <bonzini@gnu.org>
1303
1304         PR bootstrap/35457
1305         * aclocal.m4: Regenerate.
1306         * configure: Regenerate.
1307
1308 2008-03-26  Kaz Kojima  <kkojima@gcc.gnu.org>
1309
1310         * src/sh/sysv.S: Add .note.GNU-stack on Linux.
1311         * src/sh64/sysv.S: Likewise.
1312
1313 2008-03-26  Daniel Jacobowitz  <dan@debian.org>
1314
1315         * src/arm/sysv.S: Fix ARM comment marker.
1316
1317 2008-03-26  Jakub Jelinek  <jakub@redhat.com>
1318
1319         * src/alpha/osf.S: Add .note.GNU-stack on Linux.
1320         * src/s390/sysv.S: Likewise.
1321         * src/powerpc/ppc_closure.S: Likewise.
1322         * src/powerpc/sysv.S: Likewise.
1323         * src/x86/unix64.S: Likewise.
1324         * src/x86/sysv.S: Likewise.
1325         * src/sparc/v8.S: Likewise.
1326         * src/sparc/v9.S: Likewise.
1327         * src/m68k/sysv.S: Likewise.
1328         * src/arm/sysv.S: Likewise.
1329
1330 2008-03-16  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1331
1332         * aclocal.m4: Regenerate.
1333         * configure: Likewise.
1334         * Makefile.in: Likewise.
1335         * include/Makefile.in: Likewise.
1336         * testsuite/Makefile.in: Likewise.
1337
1338 2008-02-12  Bjoern Koenig  <bkoenig@alpha-tierchen.de>
1339             Andreas Tobler  <a.tobler@schweiz.org>
1340
1341         * configure.ac: Add amd64-*-freebsd* target.
1342         * configure: Regenerate.
1343
1344 2008-01-30  H.J. Lu  <hongjiu.lu@intel.com>
1345
1346         PR libffi/34612
1347         * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when
1348         returning struct.
1349
1350         * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer"
1351         tests.
1352
1353 2008-01-24  David Edelsohn  <edelsohn@gnu.org>
1354
1355         * configure: Regenerate.
1356
1357 2008-01-06  Andreas Tobler  <a.tobler@schweiz.org>
1358
1359         * src/x86/ffi.c (ffi_prep_cif_machdep): Fix thinko.
1360
1361 2008-01-05  Andreas Tobler  <a.tobler@schweiz.org>
1362
1363         PR testsuite/32843
1364         * src/x86/ffi.c (ffi_prep_cif_machdep): Add code for
1365         signed/unsigned int8/16 for X86_DARWIN.
1366         Updated copyright info.
1367         Handle one and two byte structs with special cif->flags.
1368         * src/x86/ffitarget.h: Add special types for one and two byte structs.
1369         Updated copyright info.
1370         * src/x86/darwin.S (ffi_call_SYSV): Rewrite to use a jump table like
1371         sysv.S
1372         Remove code to pop args from the stack after call.
1373         Special-case signed/unsigned for int8/16, one and two byte structs.
1374         (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1375         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1376         FFI_TYPE_SINT32.
1377         Updated copyright info.
1378
1379 2007-12-08  David Daney  <ddaney@avtrex.com>
1380
1381         * src/mips/n32.S (ffi_call_N32):  Replace dadd with ADDU, dsub with
1382         SUBU, add with ADDU and use smaller code sequences.
1383
1384 2007-12-07  David Daney  <ddaney@avtrex.com>
1385
1386         * src/mips/ffi.c (ffi_prep_cif_machdep): Handle long double return
1387         type.
1388
1389 2007-12-06  David Daney  <ddaney@avtrex.com>
1390
1391         * include/ffi.h.in (FFI_SIZEOF_JAVA_RAW): Define if not already
1392         defined.
1393         (ffi_java_raw): New typedef.
1394         (ffi_java_raw_call, ffi_java_ptrarray_to_raw,
1395         ffi_java_raw_to_ptrarray): Change parameter types from ffi_raw to
1396         ffi_java_raw.
1397         (ffi_java_raw_closure) : Same.
1398         (ffi_prep_java_raw_closure, ffi_prep_java_raw_closure_loc): Change
1399         parameter types.
1400         * src/java_raw_api.c (ffi_java_raw_size):  Replace FFI_SIZEOF_ARG with
1401         FFI_SIZEOF_JAVA_RAW.
1402         (ffi_java_raw_to_ptrarray): Change type of raw to ffi_java_raw.
1403         Replace FFI_SIZEOF_ARG with FFI_SIZEOF_JAVA_RAW. Use
1404         sizeof(ffi_java_raw) for alignment calculations.
1405         (ffi_java_ptrarray_to_raw): Same.
1406         (ffi_java_rvalue_to_raw): Add special handling for FFI_TYPE_POINTER
1407         if FFI_SIZEOF_JAVA_RAW == 4.
1408         (ffi_java_raw_to_rvalue): Same.
1409         (ffi_java_raw_call): Change type of raw to ffi_java_raw.
1410         (ffi_java_translate_args): Same.
1411         (ffi_prep_java_raw_closure_loc, ffi_prep_java_raw_closure): Change
1412         parameter types.
1413         * src/mips/ffitarget.h (FFI_SIZEOF_JAVA_RAW): Define for N32 ABI.
1414
1415 2007-12-06  David Daney  <ddaney@avtrex.com>
1416
1417         * src/mips/n32.S (ffi_closure_N32): Use 64-bit add instruction on
1418         pointer values.
1419
1420 2007-12-01  Andreas Tobler  <a.tobler@schweiz.org>
1421
1422         PR libffi/31937
1423         * src/powerpc/ffitarget.h: Introduce new ABI FFI_LINUX_SOFT_FLOAT.
1424         Add local FFI_TYPE_UINT128 to handle soft-float long-double-128.
1425         * src/powerpc/ffi.c: Distinguish between __NO_FPRS__ and not and
1426         set the NUM_FPR_ARG_REGISTERS according to.
1427         Add support for potential soft-float support under hard-float
1428         architecture.
1429         (ffi_prep_args_SYSV): Set NUM_FPR_ARG_REGISTERS to 0 in case of
1430         FFI_LINUX_SOFT_FLOAT, handle float, doubles and long-doubles according
1431         to the FFI_LINUX_SOFT_FLOAT ABI.
1432         (ffi_prep_cif_machdep): Likewise.
1433         (ffi_closure_helper_SYSV): Likewise.
1434         * src/powerpc/ppc_closure.S: Make sure not to store float/double
1435         on archs where __NO_FPRS__ is true.
1436         Add FFI_TYPE_UINT128 support.
1437         * src/powerpc/sysv.S: Add support for soft-float long-double-128.
1438         Adjust copyright notice.
1439
1440 2007-11-25  Andreas Tobler  <a.tobler@schweiz.org>
1441
1442         * src/closures.c: Move defintion of MAYBE_UNUSED from here to ...
1443         * include/ffi_common.h: ... here.
1444         Update copyright.
1445
1446 2007-11-17  Andreas Tobler  <a.tobler@schweiz.org>
1447
1448         * src/powerpc/sysv.S: Load correct cr to compare if we have long double.
1449         * src/powerpc/linux64.S: Likewise.
1450         * src/powerpc/ffi.c: Add a comment to show which part goes into cr6.
1451         * testsuite/libffi.call/return_ldl.c: New test.
1452
1453 2007-09-04    <aph@redhat.com>
1454
1455         * src/arm/sysv.S (UNWIND): New.
1456         (Whole file): Conditionally compile unwinder directives.
1457         * src/arm/sysv.S: Add unwinder directives.
1458
1459         * src/arm/ffi.c (ffi_prep_args): Align structs by at least 4 bytes.
1460         Only treat r0 as a struct address if we're actually returning a
1461         struct by address.
1462         Only copy the bytes that are actually within a struct.
1463         (ffi_prep_cif_machdep): A Composite Type not larger than 4 bytes
1464         is returned in r0, not passed by address.
1465         (ffi_call): Allocate a word-sized temporary for the case where
1466         a composite is returned in r0.
1467         (ffi_prep_incoming_args_SYSV): Align as necessary.
1468
1469 2007-08-05  Steven Newbury  <s_j_newbury@yahoo.co.uk>
1470
1471         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Use __clear_cache instead of
1472         directly using the sys_cacheflush syscall.
1473
1474 2007-07-27  Andrew Haley  <aph@redhat.com>
1475
1476         * src/arm/sysv.S (ffi_closure_SYSV): Add soft-float.
1477
1478 2007-09-03  Maciej W. Rozycki  <macro@linux-mips.org>
1479
1480         * Makefile.am: Unify MIPS_IRIX and MIPS_LINUX into MIPS.
1481         * configure.ac: Likewise.
1482         * Makefile.in: Regenerate.
1483         * include/Makefile.in: Likewise.
1484         * testsuite/Makefile.in: Likewise.
1485         * configure: Likewise.
1486
1487 2007-08-24  David Daney  <ddaney@avtrex.com>
1488
1489         * testsuite/libffi.call/return_sl.c: New test.
1490
1491 2007-08-10  David Daney  <ddaney@avtrex.com>
1492
1493         * testsuite/libffi.call/cls_multi_ushort.c,
1494         testsuite/libffi.call/cls_align_uint16.c,
1495         testsuite/libffi.call/nested_struct1.c,
1496         testsuite/libffi.call/nested_struct3.c,
1497         testsuite/libffi.call/cls_7_1_byte.c,
1498         testsuite/libffi.call/nested_struct5.c,
1499         testsuite/libffi.call/cls_double.c,
1500         testsuite/libffi.call/nested_struct7.c,
1501         testsuite/libffi.call/cls_sint.c,
1502         testsuite/libffi.call/nested_struct9.c,
1503         testsuite/libffi.call/cls_20byte1.c,
1504         testsuite/libffi.call/cls_multi_sshortchar.c,
1505         testsuite/libffi.call/cls_align_sint64.c,
1506         testsuite/libffi.call/cls_3byte2.c,
1507         testsuite/libffi.call/cls_multi_schar.c,
1508         testsuite/libffi.call/cls_multi_uchar.c,
1509         testsuite/libffi.call/cls_19byte.c,
1510         testsuite/libffi.call/cls_9byte1.c,
1511         testsuite/libffi.call/cls_align_float.c,
1512         testsuite/libffi.call/closure_fn1.c,
1513         testsuite/libffi.call/problem1.c,
1514         testsuite/libffi.call/closure_fn3.c,
1515         testsuite/libffi.call/cls_sshort.c,
1516         testsuite/libffi.call/closure_fn5.c,
1517         testsuite/libffi.call/cls_align_double.c,
1518         testsuite/libffi.call/nested_struct.c,
1519         testsuite/libffi.call/cls_2byte.c,
1520         testsuite/libffi.call/nested_struct10.c,
1521         testsuite/libffi.call/cls_4byte.c,
1522         testsuite/libffi.call/cls_6byte.c,
1523         testsuite/libffi.call/cls_8byte.c,
1524         testsuite/libffi.call/cls_multi_sshort.c,
1525         testsuite/libffi.call/cls_align_sint16.c,
1526         testsuite/libffi.call/cls_align_uint32.c,
1527         testsuite/libffi.call/cls_20byte.c,
1528         testsuite/libffi.call/cls_float.c,
1529         testsuite/libffi.call/nested_struct2.c,
1530         testsuite/libffi.call/cls_5_1_byte.c,
1531         testsuite/libffi.call/nested_struct4.c,
1532         testsuite/libffi.call/cls_24byte.c,
1533         testsuite/libffi.call/nested_struct6.c,
1534         testsuite/libffi.call/cls_64byte.c,
1535         testsuite/libffi.call/nested_struct8.c,
1536         testsuite/libffi.call/cls_uint.c,
1537         testsuite/libffi.call/cls_multi_ushortchar.c,
1538         testsuite/libffi.call/cls_schar.c,
1539         testsuite/libffi.call/cls_uchar.c,
1540         testsuite/libffi.call/cls_align_uint64.c,
1541         testsuite/libffi.call/cls_ulonglong.c,
1542         testsuite/libffi.call/cls_align_longdouble.c,
1543         testsuite/libffi.call/cls_1_1byte.c,
1544         testsuite/libffi.call/cls_12byte.c,
1545         testsuite/libffi.call/cls_3_1byte.c,
1546         testsuite/libffi.call/cls_3byte1.c,
1547         testsuite/libffi.call/cls_4_1byte.c,
1548         testsuite/libffi.call/cls_6_1_byte.c,
1549         testsuite/libffi.call/cls_16byte.c,
1550         testsuite/libffi.call/cls_18byte.c,
1551         testsuite/libffi.call/closure_fn0.c,
1552         testsuite/libffi.call/cls_9byte2.c,
1553         testsuite/libffi.call/closure_fn2.c,
1554         testsuite/libffi.call/closure_fn4.c,
1555         testsuite/libffi.call/cls_ushort.c,
1556         testsuite/libffi.call/closure_fn6.c,
1557         testsuite/libffi.call/cls_5byte.c,
1558         testsuite/libffi.call/cls_align_pointer.c,
1559         testsuite/libffi.call/cls_7byte.c,
1560         testsuite/libffi.call/cls_align_sint32.c,
1561         testsuite/libffi.special/unwindtest_ffi_call.cc,
1562         testsuite/libffi.special/unwindtest.cc: Remove xfail for mips64*-*-*.
1563
1564 2007-08-10  David Daney  <ddaney@avtrex.com>
1565
1566         PR libffi/28313
1567         * configure.ac: Don't treat mips64 as a special case.
1568         * Makefile.am (nodist_libffi_la_SOURCES): Add n32.S.
1569         * configure: Regenerate
1570         * Makefile.in: Ditto.
1571         * fficonfig.h.in: Ditto.
1572         * src/mips/ffitarget.h (REG_L, REG_S, SUBU, ADDU, SRL, LI): Indent.
1573         (LA, EH_FRAME_ALIGN, FDE_ADDR_BYTES): New preprocessor macros.
1574         (FFI_DEFAULT_ABI): Set for n64 case.
1575         (FFI_CLOSURES, FFI_TRAMPOLINE_SIZE): Define for n32 and n64 cases.
1576         * src/mips/n32.S (ffi_call_N32): Add debug macros and labels for FDE.
1577         (ffi_closure_N32): New function.
1578         (.eh_frame): New section
1579         * src/mips/o32.S: Clean up comments.
1580         (ffi_closure_O32): Pass ffi_closure parameter in $12.
1581         * src/mips/ffi.c: Use FFI_MIPS_N32 instead of
1582         _MIPS_SIM == _ABIN32 throughout.
1583         (FFI_MIPS_STOP_HERE): New, use in place of
1584         ffi_stop_here.
1585         (ffi_prep_args): Use unsigned long to hold pointer values.  Rewrite
1586         to support n32/n64 ABIs.
1587         (calc_n32_struct_flags): Rewrite.
1588         (calc_n32_return_struct_flags): Remove unused variable.  Reverse
1589         position of flag bits.
1590         (ffi_prep_cif_machdep): Rewrite n32 portion.
1591         (ffi_call): Enable for n64.  Add special handling for small structure
1592         return values.
1593         (ffi_prep_closure_loc): Add n32 and n64 support.
1594         (ffi_closure_mips_inner_O32): Add cast to silence warning.
1595         (copy_struct_N32, ffi_closure_mips_inner_N32): New functions.
1596
1597 2007-08-08  David Daney  <ddaney@avtrex.com>
1598
1599         * testsuite/libffi.call/ffitest.h (ffi_type_mylong): Remove definition.
1600         * testsuite/libffi.call/cls_align_uint16.c (main): Use correct type
1601         specifiers.
1602         * testsuite/libffi.call/nested_struct1.c (main): Ditto.
1603         * testsuite/libffi.call/cls_sint.c (main): Ditto.
1604         * testsuite/libffi.call/nested_struct9.c (main): Ditto.
1605         * testsuite/libffi.call/cls_20byte1.c (main): Ditto.
1606         * testsuite/libffi.call/cls_9byte1.c (main): Ditto.
1607         * testsuite/libffi.call/closure_fn1.c (main): Ditto.
1608         * testsuite/libffi.call/closure_fn3.c (main): Ditto.
1609         * testsuite/libffi.call/return_dbl2.c (main): Ditto.
1610         * testsuite/libffi.call/cls_sshort.c (main): Ditto.
1611         * testsuite/libffi.call/return_fl3.c (main): Ditto.
1612         * testsuite/libffi.call/closure_fn5.c (main): Ditto.
1613         * testsuite/libffi.call/nested_struct.c (main): Ditto.
1614         * testsuite/libffi.call/nested_struct10.c (main): Ditto.
1615         * testsuite/libffi.call/return_ll1.c (main): Ditto.
1616         * testsuite/libffi.call/cls_8byte.c (main): Ditto.
1617         * testsuite/libffi.call/cls_align_uint32.c (main): Ditto.
1618         * testsuite/libffi.call/cls_align_sint16.c (main): Ditto.
1619         * testsuite/libffi.call/cls_20byte.c (main): Ditto.
1620         * testsuite/libffi.call/nested_struct2.c (main): Ditto.
1621         * testsuite/libffi.call/cls_24byte.c (main): Ditto.
1622         * testsuite/libffi.call/nested_struct6.c (main): Ditto.
1623         * testsuite/libffi.call/cls_uint.c (main): Ditto.
1624         * testsuite/libffi.call/cls_12byte.c (main): Ditto.
1625         * testsuite/libffi.call/cls_16byte.c (main): Ditto.
1626         * testsuite/libffi.call/closure_fn0.c (main): Ditto.
1627         * testsuite/libffi.call/cls_9byte2.c (main): Ditto.
1628         * testsuite/libffi.call/closure_fn2.c (main): Ditto.
1629         * testsuite/libffi.call/return_dbl1.c (main): Ditto.
1630         * testsuite/libffi.call/closure_fn4.c (main): Ditto.
1631         * testsuite/libffi.call/closure_fn6.c (main): Ditto.
1632         * testsuite/libffi.call/cls_align_sint32.c (main): Ditto.
1633
1634 2007-08-07  Andrew Haley  <aph@redhat.com>
1635
1636         * src/x86/sysv.S (ffi_closure_raw_SYSV): Fix typo in previous
1637         checkin.
1638
1639 2007-08-06  Andrew Haley  <aph@redhat.com>
1640
1641         PR testsuite/32843
1642         * src/x86/sysv.S (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1643         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1644         FFI_TYPE_SINT32.
1645
1646 2007-08-02  David Daney  <ddaney@avtrex.com>
1647
1648         * testsuite/libffi.call/return_ul.c (main): Define return type as
1649         ffi_arg.  Use proper printf conversion specifier.
1650
1651 2007-07-30  Andrew Haley  <aph@redhat.com>
1652
1653         PR testsuite/32843
1654         * src/x86/ffi.c (ffi_prep_cif_machdep): in x86 case, add code for
1655         signed/unsigned int8/16.
1656         * src/x86/sysv.S (ffi_call_SYSV): Rewrite to:
1657         Use a jump table.
1658         Remove code to pop args from the stack after call.
1659         Special-case signed/unsigned int8/16.
1660         * testsuite/libffi.call/return_sc.c (main): Revert.
1661
1662 2007-07-26  Richard Guenther  <rguenther@suse.de>
1663
1664         PR testsuite/32843
1665         * testsuite/libffi.call/return_sc.c (main): Verify call
1666         result as signed char, not ffi_arg.
1667
1668 2007-07-16  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1669
1670         * configure.ac (i?86-*-solaris2.1[0-9]): Set TARGET to X86_64.
1671         * configure: Regenerate.
1672
1673 2007-07-11  David Daney  <ddaney@avtrex.com>
1674
1675         * src/mips/ffi.c: Don't include sys/cachectl.h.
1676         (ffi_prep_closure_loc): Use __builtin___clear_cache() instead of
1677         cacheflush().
1678
1679 2007-05-18  Aurelien Jarno  <aurelien@aurel32.net>
1680
1681         * src/arm/ffi.c (ffi_prep_closure_loc): Renamed and ajusted
1682         from (ffi_prep_closure): ... this.
1683         (FFI_INIT_TRAMPOLINE): Adjust.
1684
1685 2005-12-31  Phil Blundell  <pb@reciva.com>
1686
1687         * src/arm/ffi.c (ffi_prep_incoming_args_SYSV,
1688         ffi_closure_SYSV_inner, ffi_prep_closure): New, add closure support.
1689         * src/arm/sysv.S(ffi_closure_SYSV): Likewise.
1690         * src/arm/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1691         (FFI_CLOSURES): Enable closure support.
1692
1693 2007-07-03  Andrew Haley  <aph@hedges.billgatliff.com>
1694
1695         * testsuite/libffi.call/cls_multi_ushort.c,
1696         testsuite/libffi.call/cls_align_uint16.c,
1697         testsuite/libffi.call/nested_struct1.c,
1698         testsuite/libffi.call/nested_struct3.c,
1699         testsuite/libffi.call/cls_7_1_byte.c,
1700         testsuite/libffi.call/cls_double.c,
1701         testsuite/libffi.call/nested_struct5.c,
1702         testsuite/libffi.call/nested_struct7.c,
1703         testsuite/libffi.call/cls_sint.c,
1704         testsuite/libffi.call/nested_struct9.c,
1705         testsuite/libffi.call/cls_20byte1.c,
1706         testsuite/libffi.call/cls_multi_sshortchar.c,
1707         testsuite/libffi.call/cls_align_sint64.c,
1708         testsuite/libffi.call/cls_3byte2.c,
1709         testsuite/libffi.call/cls_multi_schar.c,
1710         testsuite/libffi.call/cls_multi_uchar.c,
1711         testsuite/libffi.call/cls_19byte.c,
1712         testsuite/libffi.call/cls_9byte1.c,
1713         testsuite/libffi.call/cls_align_float.c,
1714         testsuite/libffi.call/closure_fn1.c,
1715         testsuite/libffi.call/problem1.c,
1716         testsuite/libffi.call/closure_fn3.c,
1717         testsuite/libffi.call/cls_sshort.c,
1718         testsuite/libffi.call/closure_fn5.c,
1719         testsuite/libffi.call/cls_align_double.c,
1720         testsuite/libffi.call/cls_2byte.c,
1721         testsuite/libffi.call/nested_struct.c,
1722         testsuite/libffi.call/nested_struct10.c,
1723         testsuite/libffi.call/cls_4byte.c,
1724         testsuite/libffi.call/cls_6byte.c,
1725         testsuite/libffi.call/cls_8byte.c,
1726         testsuite/libffi.call/cls_multi_sshort.c,
1727         testsuite/libffi.call/cls_align_uint32.c,
1728         testsuite/libffi.call/cls_align_sint16.c,
1729         testsuite/libffi.call/cls_float.c,
1730         testsuite/libffi.call/cls_20byte.c,
1731         testsuite/libffi.call/cls_5_1_byte.c,
1732         testsuite/libffi.call/nested_struct2.c,
1733         testsuite/libffi.call/cls_24byte.c,
1734         testsuite/libffi.call/nested_struct4.c,
1735         testsuite/libffi.call/nested_struct6.c,
1736         testsuite/libffi.call/cls_64byte.c,
1737         testsuite/libffi.call/nested_struct8.c,
1738         testsuite/libffi.call/cls_uint.c,
1739         testsuite/libffi.call/cls_multi_ushortchar.c,
1740         testsuite/libffi.call/cls_schar.c,
1741         testsuite/libffi.call/cls_uchar.c,
1742         testsuite/libffi.call/cls_align_uint64.c,
1743         testsuite/libffi.call/cls_ulonglong.c,
1744         testsuite/libffi.call/cls_align_longdouble.c,
1745         testsuite/libffi.call/cls_1_1byte.c,
1746         testsuite/libffi.call/cls_12byte.c,
1747         testsuite/libffi.call/cls_3_1byte.c,
1748         testsuite/libffi.call/cls_3byte1.c,
1749         testsuite/libffi.call/cls_4_1byte.c,
1750         testsuite/libffi.call/cls_6_1_byte.c,
1751         testsuite/libffi.call/cls_16byte.c,
1752         testsuite/libffi.call/cls_18byte.c,
1753         testsuite/libffi.call/closure_fn0.c,
1754         testsuite/libffi.call/cls_9byte2.c,
1755         testsuite/libffi.call/closure_fn2.c,
1756         testsuite/libffi.call/closure_fn4.c,
1757         testsuite/libffi.call/cls_ushort.c,
1758         testsuite/libffi.call/closure_fn6.c,
1759         testsuite/libffi.call/cls_5byte.c,
1760         testsuite/libffi.call/cls_align_pointer.c,
1761         testsuite/libffi.call/cls_7byte.c,
1762         testsuite/libffi.call/cls_align_sint32.c,
1763         testsuite/libffi.special/unwindtest_ffi_call.cc,
1764         testsuite/libffi.special/unwindtest.cc: Enable for ARM.
1765
1766 2007-07-05  H.J. Lu  <hongjiu.lu@intel.com>
1767
1768         * aclocal.m4: Regenerated.
1769
1770 2007-06-02  Paolo Bonzini  <bonzini@gnu.org>
1771
1772         * configure: Regenerate.
1773
1774 2007-05-23  Steve Ellcey  <sje@cup.hp.com>
1775
1776         * Makefile.in: Regenerate.
1777         * configure: Regenerate.
1778         * aclocal.m4: Regenerate.
1779         * include/Makefile.in: Regenerate.
1780         * testsuite/Makefile.in: Regenerate.
1781
1782 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1783
1784         * src/m68k/ffi.c (ffi_prep_incoming_args_SYSV,
1785         ffi_closure_SYSV_inner,ffi_prep_closure): New, add closure support.
1786         * src/m68k/sysv.S(ffi_closure_SYSV,ffi_closure_struct_SYSV): Likewise.
1787         * src/m68k/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1788         (FFI_CLOSURES): Enable closure support.
1789
1790 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1791
1792         * configure.ac (HAVE_AS_CFI_PSEUDO_OP): New test.
1793         * configure: Regenerate.
1794         * fficonfig.h.in: Regenerate.
1795         * src/m68k/sysv.S (CFI_STARTPROC,CFI_ENDPROC,
1796         CFI_OFFSET,CFI_DEF_CFA): New macros.
1797         (ffi_call_SYSV): Add callframe annotation.
1798
1799 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1800
1801         * src/m68k/ffi.c (ffi_prep_args,ffi_prep_cif_machdep): Fix
1802         numerous test suite failures.
1803         * src/m68k/sysv.S (ffi_call_SYSV): Likewise.
1804
1805 2007-04-11  Paolo Bonzini  <bonzini@gnu.org>
1806
1807         * Makefile.am (EXTRA_DIST): Bring up to date.
1808         * Makefile.in: Regenerate.
1809         * src/frv/eabi.S: Remove RCS keyword.
1810
1811 2007-04-06  Richard Henderson  <rth@redhat.com>
1812
1813         * configure.ac: Tidy target case.
1814         (HAVE_LONG_DOUBLE): Allow the target to override.
1815         * configure: Regenerate.
1816         * include/ffi.h.in: Don't define ffi_type_foo if
1817         LIBFFI_HIDE_BASIC_TYPES is defined.
1818         (ffi_type_longdouble): If not HAVE_LONG_DOUBLE, define
1819         to ffi_type_double.
1820         * types.c (LIBFFI_HIDE_BASIC_TYPES): Define.
1821         (FFI_TYPEDEF, ffi_type_void): Mark the data const.
1822         (ffi_type_longdouble): Special case for Alpha.  Don't define
1823         if long double == double.
1824
1825         * src/alpha/ffi.c (FFI_TYPE_LONGDOUBLE): Assert unique value.
1826         (ffi_prep_cif_machdep): Handle it as the 128-bit type.
1827         (ffi_call, ffi_closure_osf_inner): Likewise.
1828         (ffi_closure_osf_inner): Likewise.  Mark hidden.
1829         (ffi_call_osf, ffi_closure_osf): Mark hidden.
1830         * src/alpha/ffitarget.h (FFI_LAST_ABI): Tidy definition.
1831         * src/alpha/osf.S (ffi_call_osf, ffi_closure_osf): Mark hidden.
1832         (load_table): Handle 128-bit long double.
1833
1834         * testsuite/libffi.call/float4.c: Add -mieee for alpha.
1835
1836 2007-04-06  Tom Tromey  <tromey@redhat.com>
1837
1838         PR libffi/31491:
1839         * README: Fixed bug in example.
1840
1841 2007-04-03  Jakub Jelinek  <jakub@redhat.com>
1842
1843         * src/closures.c: Include sys/statfs.h.
1844         (_GNU_SOURCE): Define on Linux.
1845         (FFI_MMAP_EXEC_SELINUX): Define.
1846         (selinux_enabled): New variable.
1847         (selinux_enabled_check): New function.
1848         (is_selinux_enabled): Define.
1849         (dlmmap): Use it.
1850
1851 2007-03-24  Uros Bizjak  <ubizjak@gmail.com>
1852
1853         * testsuite/libffi.call/return_fl2.c (return_fl): Mark as static.
1854         Use 'volatile float sum' to create sum of floats to avoid false
1855         negative due to excess precision on ix86 targets.
1856         (main): Ditto.
1857
1858 2007-03-08  Alexandre Oliva  <aoliva@redhat.com>
1859
1860         * src/powerpc/ffi.c (flush_icache): Fix left-over from previous
1861         patch.
1862         (ffi_prep_closure_loc): Remove unneeded casts.  Add needed ones.
1863
1864 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1865
1866         * include/ffi.h.in (ffi_closure_alloc, ffi_closure_free): New.
1867         (ffi_prep_closure_loc): New.
1868         (ffi_prep_raw_closure_loc): New.
1869         (ffi_prep_java_raw_closure_loc): New.
1870         * src/closures.c: New file.
1871         * src/dlmalloc.c [FFI_MMAP_EXEC_WRIT] (struct malloc_segment):
1872         Replace sflags with exec_offset.
1873         [FFI_MMAP_EXEC_WRIT] (mmap_exec_offset, add_segment_exec_offset,
1874         sub_segment_exec_offset): New macros.
1875         (get_segment_flags, set_segment_flags, check_segment_merge): New
1876         macros.
1877         (is_mmapped_segment, is_extern_segment): Use get_segment_flags.
1878         (add_segment, sys_alloc, create_mspace, create_mspace_with_base,
1879         destroy_mspace): Use new macros.
1880         (sys_alloc): Silence warning.
1881         * Makefile.am (libffi_la_SOURCES): Add src/closures.c.
1882         * Makefile.in: Rebuilt.
1883         * src/prep_cif [FFI_CLOSURES] (ffi_prep_closure): Implement in
1884         terms of ffi_prep_closure_loc.
1885         * src/raw_api.c (ffi_prep_raw_closure_loc): Renamed and adjusted
1886         from...
1887         (ffi_prep_raw_closure): ... this.  Re-implement in terms of the
1888         renamed version.
1889         * src/java_raw_api (ffi_prep_java_raw_closure_loc): Renamed and
1890         adjusted from...
1891         (ffi_prep_java_raw_closure): ... this.  Re-implement in terms of
1892         the renamed version.
1893         * src/alpha/ffi.c (ffi_prep_closure_loc): Renamed from
1894         (ffi_prep_closure): ... this.
1895         * src/pa/ffi.c: Likewise.
1896         * src/cris/ffi.c: Likewise.  Adjust.
1897         * src/frv/ffi.c: Likewise.
1898         * src/ia64/ffi.c: Likewise.
1899         * src/mips/ffi.c: Likewise.
1900         * src/powerpc/ffi_darwin.c: Likewise.
1901         * src/s390/ffi.c: Likewise.
1902         * src/sh/ffi.c: Likewise.
1903         * src/sh64/ffi.c: Likewise.
1904         * src/sparc/ffi.c: Likewise.
1905         * src/x86/ffi64.c: Likewise.
1906         * src/x86/ffi.c: Likewise.
1907         (FFI_INIT_TRAMPOLINE): Adjust.
1908         (ffi_prep_raw_closure_loc): Renamed and adjusted from...
1909         (ffi_prep_raw_closure): ... this.
1910         * src/powerpc/ffi.c (ffi_prep_closure_loc): Renamed from
1911         (ffi_prep_closure): ... this.
1912         (flush_icache): Adjust.
1913
1914 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1915
1916         * src/dlmalloc.c: New file, imported version 2.8.3 of Doug
1917         Lea's malloc.
1918
1919 2007-03-01  Brooks Moses  <brooks.moses@codesourcery.com>
1920
1921         * Makefile.am: Add dummy install-pdf target.
1922         * Makefile.in: Regenerate
1923
1924 2007-02-13  Andreas Krebbel  <krebbel1@de.ibm.com>
1925
1926         * src/s390/ffi.c (ffi_prep_args, ffi_prep_cif_machdep,
1927         ffi_closure_helper_SYSV): Add long double handling.
1928
1929 2007-02-02  Jakub Jelinek  <jakub@redhat.com>
1930
1931         * src/powerpc/linux64.S (ffi_call_LINUX64): Move restore of r2
1932         immediately after bctrl instruction.
1933
1934 2007-01-18  Alexandre Oliva  <aoliva@redhat.com>
1935
1936         * Makefile.am (all-recursive, install-recursive,
1937         mostlyclean-recursive, clean-recursive, distclean-recursive,
1938         maintainer-clean-recursive): Add missing targets.
1939         * Makefile.in: Rebuilt.
1940
1941 2006-12-14  Andreas Tobler  <a.tobler@schweiz.org>
1942
1943         * configure.ac: Add TARGET for x86_64-*-darwin*.
1944         * Makefile.am (nodist_libffi_la_SOURCES): Add rules for 64-bit sources
1945         for X86_DARWIN.
1946         * src/x86/ffitarget.h: Set trampoline size for x86_64-*-darwin*.
1947         * src/x86/darwin64.S: New file for x86_64-*-darwin* support.
1948         * configure: Regenerate.
1949         * Makefile.in: Regenerate.
1950         * include/Makefile.in: Regenerate.
1951         * testsuite/Makefile.in: Regenerate.
1952         * testsuite/libffi.special/unwindtest_ffi_call.cc: New test case for
1953         ffi_call only.
1954
1955 2006-12-13  Andreas Tobler <a.tobler@schweiz.org>
1956
1957         * aclocal.m4: Regenerate with aclocal -I .. as written in the
1958         Makefile.am.
1959
1960 2006-10-31  Geoffrey Keating  <geoffk@apple.com>
1961
1962         * src/powerpc/ffi_darwin.c (darwin_adjust_aggregate_sizes): New.
1963         (ffi_prep_cif_machdep): Call darwin_adjust_aggregate_sizes for
1964         Darwin.
1965         * testsuite/libffi.call/nested_struct4.c: Remove Darwin XFAIL.
1966         * testsuite/libffi.call/nested_struct6.c: Remove Darwin XFAIL.
1967
1968 2006-10-10  Paolo Bonzini  <bonzini@gnu.org>
1969             Sandro Tolaini  <tolaini@libero.it>
1970
1971         * configure.ac [i*86-*-darwin*]: Set X86_DARWIN symbol and
1972         conditional.
1973         * configure: Regenerated.
1974         * Makefile.am (nodist_libffi_la_SOURCES) [X86_DARWIN]: New case.
1975         (EXTRA_DIST): Add src/x86/darwin.S.
1976         * Makefile.in: Regenerated.
1977         * include/Makefile.in: Regenerated.
1978         * testsuite/Makefile.in: Regenerated.
1979
1980         * src/x86/ffi.c (ffi_prep_cif_machdep) [X86_DARWIN]: Treat like
1981         X86_WIN32, and additionally align stack to 16 bytes.
1982         * src/x86/darwin.S: New, based on sysv.S.
1983         * src/prep_cif.c (ffi_prep_cif) [X86_DARWIN]: Align > 8-byte structs.
1984
1985 2006-09-12  David Daney  <ddaney@avtrex.com>
1986
1987         PR libffi/23935
1988         * include/Makefile.am: Install both ffi.h and ffitarget.h in
1989         $(libdir)/gcc/$(target_alias)/$(gcc_version)/include.
1990         * aclocal.m4: Regenerated for automake 1.9.6.
1991         * Makefile.in: Regenerated.
1992         * include/Makefile.in: Regenerated.
1993         * testsuite/Makefile.in: Regenerated.
1994
1995 2006-08-17  Andreas Tobler  <a.tobler@schweiz.ch>
1996
1997         * include/ffi_common.h (struct): Revert accidental commit.
1998
1999 2006-08-15  Andreas Tobler  <a.tobler@schweiz.ch>
2000
2001         * include/ffi_common.h: Remove lint directives.
2002         * include/ffi.h.in: Likewise.
2003
2004 2006-07-25  Torsten Schoenfeld  <kaffeetisch@gmx.de>
2005
2006         * include/ffi.h.in (ffi_type_ulong, ffi_type_slong): Define correctly
2007         for 32-bit architectures.
2008         * testsuite/libffi.call/return_ul.c: New test case.
2009
2010 2006-07-19  David Daney  <ddaney@avtrex.com>
2011
2012         * testsuite/libffi.call/closure_fn6.c: Remove xfail for mips,
2013         xfail remains for mips64.
2014
2015 2006-05-23  Carlos O'Donell  <carlos@codesourcery.com>
2016
2017         * Makefile.am: Add install-html target. Add install-html to .PHONY
2018         * Makefile.in: Regenerate.
2019         * aclocal.m4: Regenerate.
2020         * include/Makefile.in: Regenerate.
2021         * testsuite/Makefile.in: Regenerate.
2022
2023 2006-05-18  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2024
2025         * pa/ffi.c (ffi_prep_args_pa32): Load floating point arguments from
2026         stack slot.
2027
2028 2006-04-22  Andreas Tobler  <a.tobler@schweiz.ch>
2029
2030         * README: Remove notice about 'Crazy Comments'.
2031         * src/debug.c: Remove lint directives. Cleanup white spaces.
2032         * src/java_raw_api.c: Likewise.
2033         * src/prep_cif.c: Likewise.
2034         * src/raw_api.c: Likewise.
2035         * src/ffitest.c: Delete. No longer needed, all test cases migrated
2036         to the testsuite.
2037         * src/arm/ffi.c: Remove lint directives.
2038         * src/m32r/ffi.c: Likewise.
2039         * src/pa/ffi.c: Likewise.
2040         * src/powerpc/ffi.c: Likewise.
2041         * src/powerpc/ffi_darwin.c: Likewise.
2042         * src/sh/ffi.c: Likewise.
2043         * src/sh64/ffi.c: Likewise.
2044         * src/x86/ffi.c: Likewise.
2045         * testsuite/libffi.call/float2.c: Likewise.
2046         * testsuite/libffi.call/promotion.c: Likewise.
2047         * testsuite/libffi.call/struct1.c: Likewise.
2048
2049 2006-04-13  Andreas Tobler  <a.tobler@schweiz.ch>
2050
2051         * src/pa/hpux32.S: Correct unwind offset calculation for
2052         ffi_closure_pa32.
2053         * src/pa/linux.S: Likewise.
2054
2055 2006-04-12  James E Wilson  <wilson@specifix.com>
2056
2057         PR libgcj/26483
2058         * src/ia64/ffi.c (stf_spill, ldf_fill): Rewrite as macros.
2059         (hfa_type_load): Call stf_spill.
2060         (hfa_type_store): Call ldf_fill.
2061         (ffi_call): Adjust calls to above routines.  Add local temps for
2062         macro result.
2063
2064 2006-04-10  Matthias Klose  <doko@debian.org>
2065
2066         * testsuite/lib/libffi-dg.exp (libffi-init): Recognize multilib
2067         directory names containing underscores.
2068
2069 2006-04-07  James E Wilson  <wilson@specifix.com>
2070
2071         * testsuite/libffi.call/float4.c: New testcase.
2072
2073 2006-04-05  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2074             Andreas Tobler  <a.tobler@schweiz.ch>
2075
2076         * Makefile.am: Add PA_HPUX port.
2077         * Makefile.in: Regenerate.
2078         * include/Makefile.in: Likewise.
2079         * testsuite/Makefile.in: Likewise.
2080         * configure.ac: Add PA_HPUX rules.
2081         * configure: Regenerate.
2082         * src/pa/ffitarget.h: Rename linux target to PA_LINUX.
2083         Add PA_HPUX and PA64_HPUX.
2084         Rename FFI_LINUX ABI to FFI_PA32 ABI.
2085         (FFI_TRAMPOLINE_SIZE): Define for 32-bit HP-UX targets.
2086         (FFI_TYPE_SMALL_STRUCT2): Define.
2087         (FFI_TYPE_SMALL_STRUCT4): Likewise.
2088         (FFI_TYPE_SMALL_STRUCT8): Likewise.
2089         (FFI_TYPE_SMALL_STRUCT3): Redefine.
2090         (FFI_TYPE_SMALL_STRUCT5): Likewise.
2091         (FFI_TYPE_SMALL_STRUCT6): Likewise.
2092         (FFI_TYPE_SMALL_STRUCT7): Likewise.
2093         * src/pa/ffi.c (ROUND_DOWN): Delete.
2094         (fldw, fstw, fldd, fstd): Use '__asm__'.
2095         (ffi_struct_type): Add support for FFI_TYPE_SMALL_STRUCT2,
2096         FFI_TYPE_SMALL_STRUCT4 and FFI_TYPE_SMALL_STRUCT8.
2097         (ffi_prep_args_LINUX): Rename to ffi_prep_args_pa32. Update comment.
2098         Simplify incrementing of stack slot variable. Change type of local
2099         'n' to unsigned int.
2100         (ffi_size_stack_LINUX): Rename to ffi_size_stack_pa32. Handle long
2101         double on PA_HPUX.
2102         (ffi_prep_cif_machdep): Likewise.
2103         (ffi_call): Likewise.
2104         (ffi_closure_inner_LINUX): Rename to ffi_closure_inner_pa32. Change
2105         return type to ffi_status. Simplify incrementing of stack slot
2106         variable. Only copy floating point argument registers when PA_LINUX
2107         is true. Reformat debug statement.
2108         Add support for FFI_TYPE_SMALL_STRUCT2, FFI_TYPE_SMALL_STRUCT4 and
2109         FFI_TYPE_SMALL_STRUCT8.
2110         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Add 'extern' to
2111         declaration.
2112         (ffi_prep_closure): Make linux trampoline conditional on PA_LINUX.
2113         Add nops to cache flush.  Add trampoline for PA_HPUX.
2114         * src/pa/hpux32.S: New file.
2115         * src/pa/linux.S (ffi_call_LINUX): Rename to ffi_call_pa32. Rename
2116         ffi_prep_args_LINUX to ffi_prep_args_pa32.
2117         Localize labels. Add support for 2, 4 and 8-byte small structs. Handle
2118         unaligned destinations in 3, 5, 6 and 7-byte small structs. Order
2119         argument type checks so that common argument types appear first.
2120         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Rename
2121         ffi_closure_inner_LINUX to ffi_closure_inner_pa32.
2122
2123 2006-03-24  Alan Modra  <amodra@bigpond.net.au>
2124
2125         * src/powerpc/ffitarget.h (enum ffi_abi): Add FFI_LINUX.  Default
2126         for 32-bit using IBM extended double format.  Fix FFI_LAST_ABI.
2127         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Handle linux variant of
2128         FFI_TYPE_LONGDOUBLE.
2129         (ffi_prep_args64): Assert using IBM extended double.
2130         (ffi_prep_cif_machdep): Don't munge FFI_TYPE_LONGDOUBLE type.
2131         Handle FFI_LINUX FFI_TYPE_LONGDOUBLE return and args.
2132         (ffi_call): Handle FFI_LINUX.
2133         (ffi_closure_helper_SYSV): Non FFI_LINUX long double return needs
2134         gpr3 return pointer as for struct return.  Handle FFI_LINUX
2135         FFI_TYPE_LONGDOUBLE return and args.  Don't increment "nf"
2136         unnecessarily.
2137         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Load both f1 and f2
2138         for FFI_TYPE_LONGDOUBLE.  Move epilogue insns into case table.
2139         Don't use r6 as pointer to results, instead use sp offset.  Don't
2140         make a special call to load lr with case table address, instead
2141         use offset from previous call.
2142         * src/powerpc/sysv.S (ffi_call_SYSV): Save long double return.
2143         * src/powerpc/linux64.S (ffi_call_LINUX64): Simplify long double
2144         return.
2145
2146 2006-03-15  Kaz Kojima  <kkojima@gcc.gnu.org>
2147
2148         * src/sh64/ffi.c (ffi_prep_cif_machdep): Handle float arguments
2149         passed with FP registers correctly.
2150         (ffi_closure_helper_SYSV): Likewise.
2151         * src/sh64/sysv.S: Likewise.
2152
2153 2006-03-01  Andreas Tobler  <a.tobler@schweiz.ch>
2154
2155         * testsuite/libffi.special/unwindtest.cc (closure_test_fn): Mark cif,
2156         args and userdata unused.
2157         (closure_test_fn1): Mark cif and userdata unused.
2158         (main): Remove unused res.
2159
2160 2006-02-28  Andreas Tobler  <a.tobler@schweiz.ch>
2161
2162         * testsuite/libffi.call/call.exp: Adjust FSF address. Add test runs for
2163         -O2, -O3, -Os and the warning flags -W -Wall.
2164         * testsuite/libffi.special/special.exp: Likewise.
2165         * testsuite/libffi.call/ffitest.h: Add an __UNUSED__ macro to mark
2166         unused parameter unused for gcc or else do nothing.
2167         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2168         * testsuite/libffi.call/cls_12byte.c (cls_struct_12byte_gn): Mark cif
2169         and userdata unused.
2170         * testsuite/libffi.call/cls_16byte.c (cls_struct_16byte_gn): Likewise.
2171         * testsuite/libffi.call/cls_18byte.c (cls_struct_18byte_gn): Likewise.
2172         * testsuite/libffi.call/cls_19byte.c (cls_struct_19byte_gn): Likewise.
2173         * testsuite/libffi.call/cls_1_1byte.c (cls_struct_1_1byte_gn): Likewise.
2174         * testsuite/libffi.call/cls_20byte.c (cls_struct_20byte_gn): Likewise.
2175         * testsuite/libffi.call/cls_20byte1.c (cls_struct_20byte_gn): Likewise.
2176         * testsuite/libffi.call/cls_24byte.c (cls_struct_24byte_gn): Likewise.
2177         * testsuite/libffi.call/cls_2byte.c (cls_struct_2byte_gn): Likewise.
2178         * testsuite/libffi.call/cls_3_1byte.c (cls_struct_3_1byte_gn): Likewise.
2179         * testsuite/libffi.call/cls_3byte1.c (cls_struct_3byte_gn): Likewise.
2180         * testsuite/libffi.call/cls_3byte2.c (cls_struct_3byte_gn1): Likewise.
2181         * testsuite/libffi.call/cls_4_1byte.c (cls_struct_4_1byte_gn): Likewise.
2182         * testsuite/libffi.call/cls_4byte.c (cls_struct_4byte_gn): Likewise.
2183         * testsuite/libffi.call/cls_5_1_byte.c (cls_struct_5byte_gn): Likewise.
2184         * testsuite/libffi.call/cls_5byte.c (cls_struct_5byte_gn): Likewise.
2185         * testsuite/libffi.call/cls_64byte.c (cls_struct_64byte_gn): Likewise.
2186         * testsuite/libffi.call/cls_6_1_byte.c (cls_struct_6byte_gn): Likewise.
2187         * testsuite/libffi.call/cls_6byte.c (cls_struct_6byte_gn): Likewise.
2188         * testsuite/libffi.call/cls_7_1_byte.c (cls_struct_7byte_gn): Likewise.
2189         * testsuite/libffi.call/cls_7byte.c (cls_struct_7byte_gn): Likewise.
2190         * testsuite/libffi.call/cls_8byte.c (cls_struct_8byte_gn): Likewise.
2191         * testsuite/libffi.call/cls_9byte1.c (cls_struct_9byte_gn): Likewise.
2192         * testsuite/libffi.call/cls_9byte2.c (cls_struct_9byte_gn): Likewise.
2193         * testsuite/libffi.call/cls_align_double.c (cls_struct_align_gn):
2194         Likewise.
2195         * testsuite/libffi.call/cls_align_float.c (cls_struct_align_gn):
2196         Likewise.
2197         * testsuite/libffi.call/cls_align_longdouble.c (cls_struct_align_gn):
2198         Likewise.
2199         * testsuite/libffi.call/cls_align_pointer.c (cls_struct_align_fn): Cast
2200         void* to avoid compiler warning.
2201         (main): Likewise.
2202         (cls_struct_align_gn): Mark cif and userdata unused.
2203         * testsuite/libffi.call/cls_align_sint16.c (cls_struct_align_gn):
2204         Likewise.
2205         * testsuite/libffi.call/cls_align_sint32.c (cls_struct_align_gn):
2206         Likewise.
2207         * testsuite/libffi.call/cls_align_sint64.c (cls_struct_align_gn):
2208         Likewise.
2209         * testsuite/libffi.call/cls_align_uint16.c (cls_struct_align_gn):
2210         Likewise.
2211         * testsuite/libffi.call/cls_align_uint32.c (cls_struct_align_gn):
2212         Likewise.
2213         * testsuite/libffi.call/cls_double.c (cls_ret_double_fn): Likewise.
2214         * testsuite/libffi.call/cls_float.c (cls_ret_float_fn): Likewise.
2215         * testsuite/libffi.call/cls_multi_schar.c (test_func_gn): Mark cif and
2216         data unused.
2217         (main): Cast res_call to silence gcc.
2218         * testsuite/libffi.call/cls_multi_sshort.c (test_func_gn): Mark cif and
2219         data unused.
2220         (main): Cast res_call to silence gcc.
2221         * testsuite/libffi.call/cls_multi_sshortchar.c (test_func_gn): Mark cif
2222         and data unused.
2223         (main): Cast res_call to silence gcc.
2224         * testsuite/libffi.call/cls_multi_uchar.c (test_func_gn): Mark cif and
2225         data unused.
2226         (main): Cast res_call to silence gcc.
2227         * testsuite/libffi.call/cls_multi_ushort.c (test_func_gn): Mark cif and
2228         data unused.
2229         (main): Cast res_call to silence gcc.
2230         * testsuite/libffi.call/cls_multi_ushortchar.c (test_func_gn): Mark cif
2231         and data unused.
2232         (main): Cast res_call to silence gcc.
2233         * testsuite/libffi.call/cls_schar.c (cls_ret_schar_fn): Mark cif and
2234         userdata unused.
2235         (cls_ret_schar_fn): Cast printf parameter to silence gcc.
2236         * testsuite/libffi.call/cls_sint.c (cls_ret_sint_fn): Mark cif and
2237         userdata unused.
2238         (cls_ret_sint_fn): Cast printf parameter to silence gcc.
2239         * testsuite/libffi.call/cls_sshort.c (cls_ret_sshort_fn): Mark cif and
2240         userdata unused.
2241         (cls_ret_sshort_fn): Cast printf parameter to silence gcc.
2242         * testsuite/libffi.call/cls_uchar.c (cls_ret_uchar_fn):  Mark cif and
2243         userdata unused.
2244         (cls_ret_uchar_fn): Cast printf parameter to silence gcc.
2245         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Mark cif and
2246         userdata unused.
2247         (cls_ret_uint_fn): Cast printf parameter to silence gcc.
2248         * testsuite/libffi.call/cls_ulonglong.c (cls_ret_ulonglong_fn): Mark cif
2249         and userdata unused.
2250         * testsuite/libffi.call/cls_ushort.c (cls_ret_ushort_fn): Mark cif and
2251         userdata unused.
2252         (cls_ret_ushort_fn): Cast printf parameter to silence gcc.
2253         * testsuite/libffi.call/float.c (floating): Remove unused parameter e.
2254         * testsuite/libffi.call/float1.c (main): Remove unused variable i.
2255         Cleanup white spaces.
2256         * testsuite/libffi.call/negint.c (checking): Remove unused variable i.
2257         * testsuite/libffi.call/nested_struct.c (cls_struct_combined_gn): Mark
2258         cif and userdata unused.
2259         * testsuite/libffi.call/nested_struct1.c (cls_struct_combined_gn):
2260         Likewise.
2261         * testsuite/libffi.call/nested_struct10.c (B_gn): Likewise.
2262         * testsuite/libffi.call/nested_struct2.c (B_fn): Adjust printf
2263         formatters to silence gcc.
2264         (B_gn): Mark cif and userdata unused.
2265         * testsuite/libffi.call/nested_struct3.c (B_gn): Mark cif and userdata
2266         unused.
2267         * testsuite/libffi.call/nested_struct4.c: Mention related PR.
2268         (B_gn): Mark cif and userdata unused.
2269         * testsuite/libffi.call/nested_struct5.c (B_gn): Mark cif and userdata
2270         unused.
2271         * testsuite/libffi.call/nested_struct6.c: Mention related PR.
2272         (B_gn): Mark cif and userdata unused.
2273         * testsuite/libffi.call/nested_struct7.c (B_gn): Mark cif and userdata
2274         unused.
2275         * testsuite/libffi.call/nested_struct8.c (B_gn): Likewise.
2276         * testsuite/libffi.call/nested_struct9.c (B_gn): Likewise.
2277         * testsuite/libffi.call/problem1.c (stub): Likewise.
2278         * testsuite/libffi.call/pyobjc-tc.c (main): Cast the result to silence
2279         gcc.
2280         * testsuite/libffi.call/return_fl2.c (return_fl): Add the note mentioned
2281         in the last commit for this test case in the test case itself.
2282         * testsuite/libffi.call/closure_fn0.c (closure_test_fn0): Mark cif as
2283         unused.
2284         * testsuite/libffi.call/closure_fn1.c (closure_test_fn1): Likewise.
2285         * testsuite/libffi.call/closure_fn2.c (closure_test_fn2): Likewise.
2286         * testsuite/libffi.call/closure_fn3.c (closure_test_fn3): Likewise.
2287         * testsuite/libffi.call/closure_fn4.c (closure_test_fn0): Likewise.
2288         * testsuite/libffi.call/closure_fn5.c (closure_test_fn5): Likewise.
2289         * testsuite/libffi.call/closure_fn6.c (closure_test_fn0): Likewise.
2290
2291 2006-02-22  Kaz Kojima  <kkojima@gcc.gnu.org>
2292
2293         * src/sh/sysv.S: Fix register numbers in the FDE for
2294         ffi_closure_SYSV.
2295
2296 2006-02-20  Andreas Tobler  <a.tobler@schweiz.ch>
2297
2298         * testsuite/libffi.call/return_fl2.c (return_fl): Remove static
2299         declaration to avoid a false negative on ix86. See PR323.
2300
2301 2006-02-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2302
2303         * src/sh/ffi.c (ffi_closure_helper_SYSV): Remove unused variable
2304         and cast integer to void * if needed.  Update the pointer to
2305         the FP register saved area correctly.
2306
2307 2006-02-17  Andreas Tobler  <a.tobler@schweiz.ch>
2308
2309         * testsuite/libffi.call/nested_struct6.c: XFAIL this test until PR25630
2310         is fixed.
2311         * testsuite/libffi.call/nested_struct4.c: Likewise.
2312
2313 2006-02-16  Andreas Tobler  <a.tobler@schweiz.ch>
2314
2315         * testsuite/libffi.call/return_dbl.c: New test case.
2316         * testsuite/libffi.call/return_dbl1.c: Likewise.
2317         * testsuite/libffi.call/return_dbl2.c: Likewise.
2318         * testsuite/libffi.call/return_fl.c: Likewise.
2319         * testsuite/libffi.call/return_fl1.c: Likewise.
2320         * testsuite/libffi.call/return_fl2.c: Likewise.
2321         * testsuite/libffi.call/return_fl3.c: Likewise.
2322         * testsuite/libffi.call/closure_fn6.c: Likewise.
2323
2324         * testsuite/libffi.call/nested_struct2.c: Remove ffi_type_mylong
2325         definition.
2326         * testsuite/libffi.call/ffitest.h: Add ffi_type_mylong definition
2327         here to be used by other test cases too.
2328
2329         * testsuite/libffi.call/nested_struct10.c: New test case.
2330         * testsuite/libffi.call/nested_struct9.c: Likewise.
2331         * testsuite/libffi.call/nested_struct8.c: Likewise.
2332         * testsuite/libffi.call/nested_struct7.c: Likewise.
2333         * testsuite/libffi.call/nested_struct6.c: Likewise.
2334         * testsuite/libffi.call/nested_struct5.c: Likewise.
2335         * testsuite/libffi.call/nested_struct4.c: Likewise.
2336
2337 2006-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
2338
2339         * configure.ac: Enable libffi for sparc64-*-freebsd*.
2340         * configure: Rebuilt.
2341
2342 2006-01-18  Jakub Jelinek  <jakub@redhat.com>
2343
2344         * src/powerpc/sysv.S (smst_two_register): Don't call __ashldi3,
2345         instead do the shifting inline.
2346         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't compute %r5
2347         shift count unconditionally.  Simplify load sequences for 1, 2, 3, 4
2348         and 8 byte structs, for the remaining struct sizes don't call
2349         __lshrdi3, instead do the shifting inline.
2350
2351 2005-12-07  Thiemo Seufer  <ths@networkno.de>
2352
2353         * src/mips/ffitarget.h: Remove obsolete sgidefs.h include. Add
2354         missing parentheses.
2355         * src/mips/o32.S (ffi_call_O32): Code formatting. Define
2356         and use A3_OFF, FP_OFF, RA_OFF. Micro-optimizations.
2357         (ffi_closure_O32): Likewise, but with newly defined A3_OFF2,
2358         A2_OFF2, A1_OFF2, A0_OFF2, RA_OFF2, FP_OFF2, S0_OFF2, GP_OFF2,
2359         V1_OFF2, V0_OFF2, FA_1_1_OFF2, FA_1_0_OFF2, FA_0_1_OFF2,
2360         FA_0_0_OFF2.
2361         * src/mips/ffi.c (ffi_prep_args): Code formatting. Fix
2362         endianness bugs.
2363         (ffi_prep_closure): Improve trampoline instruction scheduling.
2364         (ffi_closure_mips_inner_O32): Fix endianness bugs.
2365
2366 2005-12-03  Alan Modra  <amodra@bigpond.net.au>
2367
2368         * src/powerpc/ffi.c: Formatting.
2369         (ffi_prep_args_SYSV): Avoid possible aliasing problems by using unions.
2370         (ffi_prep_args64): Likewise.
2371
2372 2005-09-30  Geoffrey Keating  <geoffk@apple.com>
2373
2374         * testsuite/lib/libffi-dg.exp (libffi_target_compile): For
2375         darwin, use -shared-libgcc not -lgcc_s, and explain why.
2376
2377 2005-09-26  Tom Tromey  <tromey@redhat.com>
2378
2379         * testsuite/libffi.call/float1.c (value_type): New typedef.
2380         (CANARY): New define.
2381         (main): Check for result buffer overflow.
2382         * src/powerpc/linux64.S: Handle linux64 long double returns.
2383         * src/powerpc/ffi.c (FLAG_RETURNS_128BITS): New constant.
2384         (ffi_prep_cif_machdep): Handle linux64 long double returns.
2385
2386 2005-08-25  Alan Modra  <amodra@bigpond.net.au>
2387
2388         PR target/23404
2389         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Correct placement of stack
2390         homed fp args.
2391         (ffi_status ffi_prep_cif_machdep): Correct stack sizing for same.
2392
2393 2005-08-11  Jakub Jelinek  <jakub@redhat.com>
2394
2395         * configure.ac (HAVE_HIDDEN_VISIBILITY_ATTRIBUTE): New test.
2396         (AH_BOTTOM): Add FFI_HIDDEN definition.
2397         * configure: Rebuilt.
2398         * fficonfig.h.in: Rebuilt.
2399         * src/powerpc/ffi.c (hidden): Remove.
2400         (ffi_closure_LINUX64, ffi_prep_args64, ffi_call_LINUX64,
2401         ffi_closure_helper_LINUX64): Use FFI_HIDDEN instead of hidden.
2402         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64,
2403         .ffi_closure_LINUX64): Use FFI_HIDDEN instead of .hidden.
2404         * src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV): Remove,
2405         add FFI_HIDDEN to its prototype.
2406         (ffi_closure_SYSV_inner): New.
2407         * src/x86/sysv.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2408         * src/x86/win32.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2409
2410 2005-08-10  Alfred M. Szmidt  <ams@gnu.org>
2411
2412         PR libffi/21819:
2413         * configure: Rebuilt.
2414         * configure.ac: Handle i*86-*-gnu*.
2415
2416 2005-08-09  Jakub Jelinek  <jakub@redhat.com>
2417
2418         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Use
2419         DW_CFA_offset_extended_sf rather than
2420         DW_CFA_GNU_negative_offset_extended.
2421         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise.
2422
2423 2005-07-22  SUGIOKA Toshinobu  <sugioka@itonet.co.jp>
2424
2425         * src/sh/sysv.S (ffi_call_SYSV): Stop argument popping correctly
2426         on sh3.
2427         (ffi_closure_SYSV): Change the stack layout for sh3 struct argument.
2428         * src/sh/ffi.c (ffi_prep_args): Fix sh3 argument copy, when it is
2429         partially on register.
2430         (ffi_closure_helper_SYSV): Likewise.
2431         (ffi_prep_cif_machdep): Don't set too many cif->flags.
2432
2433 2005-07-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2434
2435         * src/sh/ffi.c (ffi_call): Handle small structures correctly.
2436         Remove empty line.
2437         * src/sh64/ffi.c (simple_type): Remove.
2438         (return_type): Handle small structures correctly.
2439         (ffi_prep_args): Likewise.
2440         (ffi_call): Likewise.
2441         (ffi_closure_helper_SYSV): Likewise.
2442         * src/sh64/sysv.S (ffi_call_SYSV): Handle 1, 2 and 4-byte return.
2443         Emit position independent code if PIC and remove wrong datalabel
2444         prefixes from EH data.
2445
2446 2005-07-19  Andreas Tobler  <a.tobler@schweiz.ch>
2447
2448         * Makefile.am (nodist_libffi_la_SOURCES): Add POWERPC_FREEBSD.
2449         * Makefile.in: Regenerate.
2450         * include/Makefile.in: Likewise.
2451         * testsuite/Makefile.in: Likewise.
2452         * configure.ac: Add POWERPC_FREEBSD rules.
2453         * configure: Regenerate.
2454         * src/powerpc/ffitarget.h: Add POWERPC_FREEBSD rules.
2455         (FFI_SYSV_TYPE_SMALL_STRUCT): Define.
2456         * src/powerpc/ffi.c: Add flags to handle small structure returns
2457         in ffi_call_SYSV.
2458         (ffi_prep_cif_machdep): Handle small structures for SYSV 4 ABI.
2459         Aka FFI_SYSV.
2460         (ffi_closure_helper_SYSV): Likewise.
2461         * src/powerpc/ppc_closure.S: Add return types for small structures.
2462         * src/powerpc/sysv.S: Add bits to handle small structures for
2463         final SYSV 4 ABI.
2464
2465 2005-07-10  Andreas Tobler  <a.tobler@schweiz.ch>
2466
2467         * testsuite/libffi.call/cls_5_1_byte.c: New test file.
2468         * testsuite/libffi.call/cls_6_1_byte.c: Likewise.
2469         * testsuite/libffi.call/cls_7_1_byte.c: Likewise.
2470
2471 2005-07-05  Randolph Chung  <tausq@debian.org>
2472
2473         * src/pa/ffi.c (ffi_struct_type): Rename FFI_TYPE_SMALL_STRUCT1
2474         as FFI_TYPE_SMALL_STRUCT3.  Break out handling for 5-7 byte
2475         structures.  Kill compilation warnings.
2476         (ffi_closure_inner_LINUX): Print return values as hex in debug
2477         message.  Rename FFI_TYPE_SMALL_STRUCT1 as FFI_TYPE_SMALL_STRUCT3.
2478         Properly handle 5-7 byte structure returns.
2479         * src/pa/ffitarget.h (FFI_TYPE_SMALL_STRUCT1)
2480         (FFI_TYPE_SMALL_STRUCT2): Remove.
2481         (FFI_TYPE_SMALL_STRUCT3, FFI_TYPE_SMALL_STRUCT5)
2482         (FFI_TYPE_SMALL_STRUCT6, FFI_TYPE_SMALL_STRUCT7): Define.
2483         * src/pa/linux.S: Mark source file as using PA1.1 assembly.
2484         (checksmst1, checksmst2): Remove.
2485         (checksmst3): Optimize handling of 3-byte struct returns.
2486         (checksmst567): Properly handle 5-7 byte struct returns.
2487
2488 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2489
2490         PR libgcj/21943
2491         * src/mips/n32.S: Enforce PIC code.
2492         * src/mips/o32.S: Likewise.
2493
2494 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2495
2496         * configure.ac: Treat i*86-*-solaris2.10 and up as X86_64.
2497         * configure: Regenerate.
2498
2499 2005-06-01  Alan Modra  <amodra@bigpond.net.au>
2500
2501         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't use JUMPTARGET
2502         to call ffi_closure_helper_SYSV.  Append @local instead.
2503         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise for ffi_prep_args_SYSV.
2504
2505 2005-05-17  Kelley Cook  <kcook@gcc.gnu.org>
2506
2507         * configure.ac: Use AC_C_BIGENDIAN instead of AC_C_BIGENDIAN_CROSS.
2508         Use AC_CHECK_SIZEOF instead of AC_COMPILE_CHECK_SIZEOF.
2509         * Makefile.am (ACLOCAL_AMFLAGS): Remove -I ../config.
2510         * aclocal.m4, configure, fficonfig.h.in, Makefile.in,
2511         include/Makefile.in, testsuite/Makefile.in: Regenerate.
2512
2513 2005-05-09  Mike Stump  <mrs@apple.com>
2514
2515         * configure: Regenerate.
2516
2517 2005-05-08  Richard Henderson  <rth@redhat.com>
2518
2519         PR libffi/21285
2520         * src/alpha/osf.S: Update unwind into to match code.
2521
2522 2005-05-04  Andreas Degert <ad@papyrus-gmbh.de>
2523             Richard Henderson  <rth@redhat.com>
2524
2525         * src/x86/ffi64.c (ffi_prep_cif_machdep): Save sse-used flag in
2526         bit 11 of flags.
2527         (ffi_call): Mask return type field.  Pass ssecount to ffi_call_unix64.
2528         (ffi_prep_closure): Set carry bit if sse-used flag set.
2529         * src/x86/unix64.S (ffi_call_unix64): Add ssecount argument.
2530         Only load sse registers if ssecount non-zero.
2531         (ffi_closure_unix64): Only save sse registers if carry set on entry.
2532
2533 2005-04-29  Ralf Corsepius  <ralf.corsepius@rtems.org>
2534
2535         * configure.ac: Add i*86-*-rtems*, sparc*-*-rtems*,
2536         powerpc-*rtems*, arm*-*-rtems*, sh-*-rtems*.
2537         * configure: Regenerate.
2538
2539 2005-04-20  Hans-Peter Nilsson  <hp@axis.com>
2540
2541         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): In regsub use,
2542         have Tcl8.3-compatible intermediate variable.
2543
2544 2005-04-18  Simon Posnjak <simon.posnjak@siol.net>
2545             Hans-Peter Nilsson  <hp@axis.com>
2546
2547         * Makefile.am: Add CRIS support.
2548         * configure.ac: Likewise.
2549         * Makefile.in, configure, testsuite/Makefile.in,
2550         include/Makefile.in: Regenerate.
2551         * src/cris: New directory.
2552         * src/cris/ffi.c, src/cris/sysv.S, src/cris/ffitarget.h: New files.
2553         * src/prep_cif.c (ffi_prep_cif): Wrap in #ifndef __CRIS__.
2554
2555         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): Replace \n with
2556         \r?\n in output tests.
2557
2558 2005-04-12  Mike Stump  <mrs@apple.com>
2559
2560         * configure: Regenerate.
2561
2562 2005-03-30  Hans Boehm  <Hans.Boehm@hp.com>
2563
2564         * src/ia64/ffitarget.h (ffi_arg): Use long long instead of DI.
2565
2566 2005-03-30  Steve Ellcey  <sje@cup.hp.com>
2567
2568         * src/ia64/ffitarget.h (ffi_arg) ADD DI attribute.
2569         (ffi_sarg) Ditto.
2570         * src/ia64/unix.S (ffi_closure_unix): Extend gp
2571         to 64 bits in ILP32 mode.
2572         Load 64 bits even for short data.
2573
2574 2005-03-23  Mike Stump  <mrs@apple.com>
2575
2576         * src/powerpc/darwin.S: Update for -m64 multilib.
2577         * src/powerpc/darwin_closure.S: Likewise.
2578
2579 2005-03-21  Zack Weinberg  <zack@codesourcery.com>
2580
2581         * configure.ac: Do not invoke TL_AC_GCC_VERSION.
2582         Do not set tool_include_dir.
2583         * aclocal.m4, configure, Makefile.in, testsuite/Makefile.in:
2584         Regenerate.
2585         * include/Makefile.am: Set gcc_version and toollibffidir.
2586         * include/Makefile.in: Regenerate.
2587
2588 2005-02-22  Andrew Haley  <aph@redhat.com>
2589
2590         * src/powerpc/ffi.c (ffi_prep_cif_machdep): Bump alignment to
2591         odd-numbered register pairs for 64-bit integer types.
2592
2593 2005-02-23  Andreas Tobler  <a.tobler@schweiz.ch>
2594
2595         PR libffi/20104
2596         * testsuite/libffi.call/return_ll1.c: New test case.
2597
2598 2005-02-11  Janis Johnson  <janis187@us.ibm.com>
2599
2600         * testsuite/libffi.call/cls_align_longdouble.c: Remove dg-options.
2601         * testsuite/libffi.call/float.c: Ditto.
2602         * testsuite/libffi.call/float2.c: Ditto.
2603         * testsuite/libffi.call/float3.c: Ditto.
2604
2605 2005-02-08  Andreas Tobler  <a.tobler@schweiz.ch>
2606
2607         * src/frv/ffitarget.h: Remove PPC stuff which does not belong to frv.
2608
2609 2005-01-12  Eric Botcazou  <ebotcazou@libertysurf.fr>
2610
2611         * testsuite/libffi.special/special.exp (cxx_options): Add
2612         -shared-libgcc.
2613
2614 2004-12-31  Richard Henderson  <rth@redhat.com>
2615
2616         * src/types.c (FFI_AGGREGATE_TYPEDEF): Remove.
2617         (FFI_TYPEDEF): Rename from FFI_INTEGRAL_TYPEDEF.  Replace size and
2618         offset parameters with a type parameter; deduce size and structure
2619         alignment.  Update all users.
2620
2621 2004-12-31  Richard Henderson  <rth@redhat.com>
2622
2623         * src/types.c (FFI_TYPE_POINTER): Define with sizeof.
2624         (FFI_TYPE_LONGDOUBLE): Fix for ia64.
2625         * src/ia64/ffitarget.h (struct ffi_ia64_trampoline_struct): Move
2626         into ffi_prep_closure.
2627         * src/ia64/ia64_flags.h, src/ia64/ffi.c, src/ia64/unix.S: Rewrite
2628         from scratch.
2629
2630 2004-12-27  Richard Henderson  <rth@redhat.com>
2631
2632         * src/x86/unix64.S: Fix typo in unwind info.
2633
2634 2004-12-25  Richard Henderson  <rth@redhat.com>
2635
2636         * src/x86/ffi64.c (struct register_args): Rename from stackLayout.
2637         (enum x86_64_reg_class): Add X86_64_COMPLEX_X87_CLASS.
2638         (merge_classes): Check for it.
2639         (SSE_CLASS_P): New.
2640         (classify_argument): Pass byte_offset by value; perform all updates
2641         inside struct case.
2642         (examine_argument): Add classes argument; handle
2643         X86_64_COMPLEX_X87_CLASS.
2644         (ffi_prep_args): Merge into ...
2645         (ffi_call): ... here.  Share stack frame with ffi_call_unix64.
2646         (ffi_prep_cif_machdep): Setup cif->flags for proper structure return.
2647         (ffi_fill_return_value): Remove.
2648         (ffi_prep_closure): Remove dead assert.
2649         (ffi_closure_unix64_inner): Rename from ffi_closure_UNIX64_inner.
2650         Rewrite to use struct register_args instead of va_list.  Create
2651         flags for handling structure returns.
2652         * src/x86/unix64.S: Remove dead strings.
2653         (ffi_call_unix64): Rename from ffi_call_UNIX64.  Rewrite to share
2654         stack frame with ffi_call.  Handle structure returns properly.
2655         (float2sse, floatfloat2sse, double2sse): Remove.
2656         (sse2float, sse2double, sse2floatfloat): Remove.
2657         (ffi_closure_unix64): Rename from ffi_closure_UNIX64.  Rewrite
2658         to handle structure returns properly.
2659
2660 2004-12-08  David Edelsohn  <edelsohn@gnu.org>
2661
2662         * Makefile.am (AM_MAKEFLAGS): Remove duplicate LIBCFLAGS and
2663         PICFLAG.
2664         * Makefile.in: Regenerated.
2665
2666 2004-12-02  Richard Sandiford  <rsandifo@redhat.com>
2667
2668         * configure.ac: Use TL_AC_GCC_VERSION to set gcc_version.
2669         * configure, aclocal.m4, Makefile.in: Regenerate.
2670         * include/Makefile.in, testsuite/Makefile.in: Regenerate.
2671
2672 2004-11-29  Kelley Cook  <kcook@gcc.gnu.org>
2673
2674         * configure: Regenerate for libtool change.
2675
2676 2004-11-25  Kelley Cook  <kcook@gcc.gnu.org>
2677
2678         * configure: Regenerate for libtool reversion.
2679
2680 2004-11-24  Kelley Cook  <kcook@gcc.gnu.org>
2681
2682         * configure: Regenerate for libtool change.
2683
2684 2004-11-23  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2685
2686         * testsuite/lib/libffi-dg.exp: Use new procs in target-libpath.exp.
2687
2688 2004-11-23  Richard Sandiford  <rsandifo@redhat.com>
2689
2690         * src/mips/o32.S (ffi_call_O32, ffi_closure_O32): Use jalr instead
2691         of jal.  Use an absolute encoding for the frame information.
2692
2693 2004-11-23  Kelley Cook  <kcook@gcc.gnu.org>
2694
2695         * Makefile.am: Remove no-dependencies.  Add ACLOCAL_AMFLAGS.
2696         * acinclude.m4: Delete logic for sincludes.
2697         * aclocal.m4, Makefile.in, configure: Regenerate.
2698         * include/Makefile: Likewise.
2699         * testsuite/Makefile: Likewise.
2700
2701 2004-11-22  Eric Botcazou  <ebotcazou@libertysurf.fr>
2702
2703         * src/sparc/ffi.c (ffi_prep_closure): Align doubles and 64-bit integers
2704         on a 8-byte boundary.
2705         * src/sparc/v8.S (ffi_closure_v8): Reserve frame space for arguments.
2706
2707 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2708
2709         * src/arm/ffi.c (ffi_prep_cif_machdep): Handle functions that return
2710         long long values.  Round stack allocation to a multiple of 8 bytes
2711         for ATPCS compatibility.
2712         * src/arm/sysv.S (ffi_call_SYSV): Rework to avoid use of APCS register
2713         names.  Handle returning long long types.  Add Thumb and interworking
2714         support.  Improve soft-float code.
2715
2716 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2717
2718         * testsuite/lib/libffi-db.exp (load_gcc_lib): New function.
2719         (libffi_exit): New function.
2720         (libffi_init): Build the testglue wrapper if needed.
2721
2722 2004-10-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2723
2724         PR other/18138
2725         * testsuite/lib/libffi-dg.exp: Accept more than one multilib libgcc.
2726
2727 2004-10-25  Kazuhiro Inaoka <inaoka.kazuhiro@renesas.com>
2728
2729         * src/m32r/libffitarget.h (FFI_CLOSURES): Set to 0.
2730
2731 2004-10-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2732
2733         * src/sh/sysv.S (ffi_call_SYSV): Don't align for double data.
2734         * testsuite/libffi.call/float3.c: New test case.
2735
2736 2004-10-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2737
2738         * src/sh/ffi.c (ffi_prep_closure): Set T bit in trampoline for
2739         the function returning a structure pointed with R2.
2740         * src/sh/sysv.S (ffi_closure_SYSV): Use R2 as the pointer to
2741         the structure return value if T bit set.  Emit position
2742         independent code and EH data if PIC.
2743
2744 2004-10-13  Kazuhiro Inaoka  <inaoka.kazuhiro@renesas.com>
2745
2746         * Makefile.am: Add m32r support.
2747         * configure.ac: Likewise.
2748         * Makefile.in: Regenerate.
2749         * confiugre: Regenerate.
2750         * src/types.c: Add m32r port to FFI_INTERNAL_TYPEDEF
2751         (uint64, sint64, double, longdouble)
2752         * src/m32r: New directory.
2753         * src/m32r/ffi.c: New file.
2754         * src/m32r/sysv.S: Likewise.
2755         * src/m32r/ffitarget.h: Likewise.
2756
2757 2004-10-02  Kaz Kojima  <kkojima@gcc.gnu.org>
2758
2759         * testsuite/libffi.call/negint.c: New test case.
2760
2761 2004-09-14  H.J. Lu  <hongjiu.lu@intel.com>
2762
2763         PR libgcj/17465
2764         * testsuite/lib/libffi-dg.exp: Don't use global ld_library_path.
2765         Set up LD_LIBRARY_PATH, SHLIB_PATH, LD_LIBRARYN32_PATH,
2766         LD_LIBRARY64_PATH, LD_LIBRARY_PATH_32, LD_LIBRARY_PATH_64 and
2767         DYLD_LIBRARY_PATH.
2768
2769 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2770
2771         * testsuite/libffi.call/many_win32.c: Remove whitespaces.
2772         * testsuite/libffi.call/promotion.c: Likewise.
2773         * testsuite/libffi.call/return_ll.c: Remove unused var. Cleanup
2774         whitespaces.
2775         * testsuite/libffi.call/return_sc.c: Likewise.
2776         * testsuite/libffi.call/return_uc.c: Likewise.
2777
2778 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2779
2780         * src/powerpc/darwin.S: Fix comments and identation.
2781         * src/powerpc/darwin_closure.S: Likewise.
2782
2783 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2784
2785         * src/powerpc/ffi_darwin.c: Add flag for longdouble return values.
2786         (ffi_prep_args): Handle longdouble arguments.
2787         (ffi_prep_cif_machdep): Set flags for longdouble. Calculate space for
2788         longdouble.
2789         (ffi_closure_helper_DARWIN): Add closure handling for longdouble.
2790         * src/powerpc/darwin.S (_ffi_call_DARWIN): Add handling of longdouble
2791         values.
2792         * src/powerpc/darwin_closure.S (_ffi_closure_ASM): Likewise.
2793         * src/types.c: Defined longdouble size and alignment for darwin.
2794
2795 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2796
2797         * src/powerpc/aix.S: Remove whitespaces.
2798         * src/powerpc/aix_closure.S: Likewise.
2799         * src/powerpc/asm.h: Likewise.
2800         * src/powerpc/ffi.c: Likewise.
2801         * src/powerpc/ffitarget.h: Likewise.
2802         * src/powerpc/linux64.S: Likewise.
2803         * src/powerpc/linux64_closure.S: Likewise.
2804         * src/powerpc/ppc_closure.S: Likewise.
2805         * src/powerpc/sysv.S: Likewise.
2806
2807 2004-08-30  Anthony Green  <green@redhat.com>
2808
2809         * Makefile.am: Add frv support.
2810         * Makefile.in, testsuite/Makefile.in: Rebuilt.
2811         * configure.ac: Read configure.host.
2812         * configure.in: Read configure.host.
2813         * configure.host: New file.  frv-elf needs libgloss.
2814         * include/ffi.h.in: Force ffi_closure to have a nice big (8)
2815         alignment.  This is needed to frv and shouldn't harm the others.
2816         * include/ffi_common.h (ALIGN_DOWN): New macro.
2817         * src/frv/ffi.c, src/frv/ffitarget.h, src/frv/eabi.S: New files.
2818
2819 2004-08-24  David Daney  <daney@avtrex.com>
2820
2821         * testsuite/libffi.call/closure_fn0.c: Xfail mips64* instead of mips*.
2822         * testsuite/libffi.call/closure_fn1.c: Likewise.
2823         * testsuite/libffi.call/closure_fn2.c  Likewise.
2824         * testsuite/libffi.call/closure_fn3.c: Likewise.
2825         * testsuite/libffi.call/closure_fn4.c: Likewise.
2826         * testsuite/libffi.call/closure_fn5.c: Likewise.
2827         * testsuite/libffi.call/cls_18byte.c: Likewise.
2828         * testsuite/libffi.call/cls_19byte.c: Likewise.
2829         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2830         * testsuite/libffi.call/cls_20byte.c: Likewise.
2831         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2832         * testsuite/libffi.call/cls_24byte.c: Likewise.
2833         * testsuite/libffi.call/cls_2byte.c: Likewise.
2834         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2835         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2836         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2837         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2838         * testsuite/libffi.call/cls_4byte.c: Likewise.
2839         * testsuite/libffi.call/cls_64byte.c: Likewise.
2840         * testsuite/libffi.call/cls_6byte.c: Likewise.
2841         * testsuite/libffi.call/cls_7byte.c: Likewise.
2842         * testsuite/libffi.call/cls_8byte.c: Likewise.
2843         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2844         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2845         * testsuite/libffi.call/cls_align_double.c: Likewise.
2846         * testsuite/libffi.call/cls_align_float.c: Likewise.
2847         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2848         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2849         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2850         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2851         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2852         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2853         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2854         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
2855         * testsuite/libffi.call/cls_double.c: Likewise.
2856         * testsuite/libffi.call/cls_float.c: Likewise.
2857         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2858         * testsuite/libffi.call/cls_multi_sshort.c: Likewise.
2859         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2860         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2861         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2862         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2863         * testsuite/libffi.call/cls_schar.c: Likewise.
2864         * testsuite/libffi.call/cls_sint.c: Likewise.
2865         * testsuite/libffi.call/cls_sshort.c: Likewise.
2866         * testsuite/libffi.call/cls_uchar.c: Likewise.
2867         * testsuite/libffi.call/cls_uint.c: Likewise.
2868         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2869         * testsuite/libffi.call/cls_ushort.c: Likewise.
2870         * testsuite/libffi.call/nested_struct.c: Likewise.
2871         * testsuite/libffi.call/nested_struct1.c: Likewise.
2872         * testsuite/libffi.call/nested_struct2.c: Likewise.
2873         * testsuite/libffi.call/nested_struct3.c: Likewise.
2874         * testsuite/libffi.call/problem1.c: Likewise.
2875         * testsuite/libffi.special/unwindtest.cc: Likewise.
2876         * testsuite/libffi.call/cls_12byte.c: Likewise and set return value
2877         to zero.
2878         * testsuite/libffi.call/cls_16byte.c: Likewise.
2879         * testsuite/libffi.call/cls_5byte.c: Likewise.
2880
2881 2004-08-23  David Daney <daney@avtrex.com>
2882
2883         PR libgcj/13141
2884         * src/mips/ffitarget.h (FFI_O32_SOFT_FLOAT): New ABI.
2885         * src/mips/ffi.c (ffi_prep_args): Fix alignment calculation.
2886         (ffi_prep_cif_machdep): Handle FFI_O32_SOFT_FLOAT floating point
2887         parameters and return types.
2888         (ffi_call): Handle FFI_O32_SOFT_FLOAT ABI.
2889         (ffi_prep_closure): Ditto.
2890         (ffi_closure_mips_inner_O32): Handle FFI_O32_SOFT_FLOAT ABI, fix
2891         alignment calculations.
2892         * src/mips/o32.S (ffi_closure_O32): Don't use floating point
2893         instructions if FFI_O32_SOFT_FLOAT, make stack frame ABI compliant.
2894
2895 2004-08-14  Casey Marshall <csm@gnu.org>
2896
2897         * src/mips/ffi.c (ffi_pref_cif_machdep): set `cif->flags' to
2898         contain `FFI_TYPE_UINT64' as return type for any 64-bit
2899         integer (O32 ABI only).
2900         (ffi_prep_closure): new function.
2901         (ffi_closure_mips_inner_O32): new function.
2902         * src/mips/ffitarget.h: Define `FFI_CLOSURES' and
2903         `FFI_TRAMPOLINE_SIZE' appropriately if the ABI is o32.
2904         * src/mips/o32.S (ffi_call_O32): add labels for .eh_frame. Return
2905         64 bit integers correctly.
2906         (ffi_closure_O32): new function.
2907         Added DWARF-2 unwind info for both functions.
2908
2909 2004-08-10  Andrew Haley  <aph@redhat.com>
2910
2911         * src/x86/ffi64.c (ffi_prep_args ): 8-align all stack arguments.
2912
2913 2004-08-01  Robert Millan  <robertmh@gnu.org>
2914
2915         * configure.ac: Detect knetbsd-gnu and kfreebsd-gnu.
2916         * configure: Regenerate.
2917
2918 2004-07-30  Maciej W. Rozycki  <macro@linux-mips.org>
2919
2920         * acinclude.m4 (AC_FUNC_MMAP_BLACKLIST): Check for <sys/mman.h>
2921         and mmap() explicitly instead of relying on preset autoconf cache
2922         variables.
2923         * aclocal.m4: Regenerate.
2924         * configure: Regenerate.
2925
2926 2004-07-11  Ulrich Weigand  <uweigand@de.ibm.com>
2927
2928         * src/s390/ffi.c (ffi_prep_args): Fix C aliasing violation.
2929         (ffi_check_float_struct): Remove unused prototype.
2930
2931 2004-06-30  Geoffrey Keating  <geoffk@apple.com>
2932
2933         * src/powerpc/ffi_darwin.c (flush_icache): ';' is a comment
2934         character on Darwin, use '\n\t' instead.
2935
2936 2004-06-26  Matthias Klose  <doko@debian.org>
2937
2938         * libtool-version: Fix typo in revision/age.
2939
2940 2004-06-17  Matthias Klose  <doko@debian.org>
2941
2942         * libtool-version: New.
2943         * Makefile.am (libffi_la_LDFLAGS): Use -version-info for soname.
2944         * Makefile.in: Regenerate.
2945
2946 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2947
2948         * Makefile.am: Remove useless multilib rules.
2949         * Makefile.in: Regenerate.
2950         * aclocal.m4: Regenerate with automake 1.8.5.
2951         * configure.ac: Remove useless multilib configury.
2952         * configure: Regenerate.
2953
2954 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2955
2956         * .cvsignore: New file.
2957
2958 2004-06-10  Jakub Jelinek  <jakub@redhat.com>
2959
2960         * src/ia64/unix.S (ffi_call_unix): Insert group barrier break
2961         fp_done.
2962         (ffi_closure_UNIX): Fix f14/f15 adjustment if FLOAT_SZ is ever
2963         changed from 8.
2964
2965 2004-06-06  Sean McNeil  <sean@mcneil.com>
2966
2967         * configure.ac: Add x86_64-*-freebsd* support.
2968         * configure: Regenerate.
2969
2970 2004-04-26  Joe Buck <jbuck@welsh-buck.org>
2971
2972         Bug 15093
2973         * configure.ac: Test for existence of mmap and sys/mman.h before
2974         checking blacklist.  Fix suggested by Jim Wilson.
2975         * configure: Regenerate.
2976
2977 2004-04-26  Matt Austern  <austern@apple.com>
2978
2979         * src/powerpc/darwin.S: Go through a non-lazy pointer for initial
2980         FDE location.
2981         * src/powerpc/darwin_closure.S: Likewise.
2982
2983 2004-04-24  Andreas Tobler  <a.tobler@schweiz.ch>
2984
2985         * testsuite/libffi.call/cls_multi_schar.c (main): Fix initialization
2986         error. Reported by Thomas Heller <theller@python.net>.
2987         * testsuite/libffi.call/cls_multi_sshort.c (main): Likewise.
2988         * testsuite/libffi.call/cls_multi_ushort.c (main): Likewise.
2989
2990 2004-03-20  Matthias Klose  <doko@debian.org>
2991
2992         * src/pa/linux.S: Fix typo.
2993
2994 2004-03-19  Matthias Klose  <doko@debian.org>
2995
2996         * Makefile.am: Update.
2997         * Makefile.in: Regenerate.
2998         * src/pa/ffi.h.in: Remove.
2999         * src/pa/ffitarget.h: New file.
3000
3001 2004-02-10  Randolph Chung  <tausq@debian.org>
3002
3003         * Makefile.am: Add PA support.
3004         * Makefile.in: Regenerate.
3005         * include/Makefile.in: Regenerate.
3006         * configure.ac: Add PA target.
3007         * configure: Regenerate.
3008         * src/pa/ffi.c: New file.
3009         * src/pa/ffi.h.in: Add PA support.
3010         * src/pa/linux.S: New file.
3011         * prep_cif.c: Add PA support.
3012
3013 2004-03-16  Hosaka Yuji  <hos@tamanegi.org>
3014
3015         * src/types.c: Fix alignment size of X86_WIN32 case int64 and
3016         double.
3017         * src/x86/ffi.c (ffi_prep_args): Replace ecif->cif->rtype->type
3018         with ecif->cif->flags.
3019         (ffi_call, ffi_prep_incoming_args_SYSV): Replace cif->rtype->type
3020         with cif->flags.
3021         (ffi_prep_cif_machdep): Add X86_WIN32 struct case.
3022         (ffi_closure_SYSV): Add 1 or 2-bytes struct case for X86_WIN32.
3023         * src/x86/win32.S (retstruct1b, retstruct2b, sc_retstruct1b,
3024         sc_retstruct2b): Add for 1 or 2-bytes struct case.
3025
3026 2004-03-15 Kelley Cook <kcook@gcc.gnu.org>
3027
3028         * configure.in: Rename file to ...
3029         * configure.ac: ... this.
3030         * fficonfig.h.in: Regenerate.
3031         * Makefile.in: Regenerate.
3032         * include/Makefile.in: Regenerate.
3033         * testsuite/Makefile.in: Regenerate.
3034
3035 2004-03-12  Matt Austern  <austern@apple.com>
3036
3037         * src/powerpc/darwin.S: Fix EH information so it corresponds to
3038         changes in EH format resulting from addition of linkonce support.
3039         * src/powerpc/darwin_closure.S: Likewise.
3040
3041 2004-03-11  Andreas Tobler  <a.tobler@schweiz.ch>
3042             Paolo Bonzini  <bonzini@gnu.org>
3043
3044         * Makefile.am (AUTOMAKE_OPTIONS): Set them.
3045         Remove VPATH. Remove rules for object files. Remove multilib support.
3046         (AM_CCASFLAGS): Add.
3047         * configure.in (AC_CONFIG_HEADERS): Relace AM_CONFIG_HEADER.
3048         (AC_PREREQ): Bump version to 2.59.
3049         (AC_INIT): Fill with version info and bug address.
3050         (ORIGINAL_LD_FOR_MULTILIBS): Remove.
3051         (AM_ENABLE_MULTILIB): Use this instead of AC_ARG_ENABLE.
3052         De-precious CC so that the right flags are passed down to multilibs.
3053         (AC_MSG_ERROR): Replace obsolete macro AC_ERROR.
3054         (AC_CONFIG_FILES): Replace obsolete macro AC_LINK_FILES.
3055         (AC_OUTPUT): Reorganize the output with AC_CONFIG_COMMANDS.
3056         * configure: Rebuilt.
3057         * aclocal.m4: Likewise.
3058         * Makefile.in, include/Makefile.in, testsuite/Makefile.in: Likewise.
3059         * fficonfig.h.in: Likewise.
3060
3061 2004-03-11  Andreas Schwab  <schwab@suse.de>
3062
3063         * src/ia64/ffi.c (ffi_prep_incoming_args_UNIX): Get floating point
3064         arguments from fp registers only for the first 8 parameter slots.
3065         Don't convert a float parameter when passed in memory.
3066
3067 2004-03-09  Hans-Peter Nilsson  <hp@axis.com>
3068
3069         * configure: Regenerate for config/accross.m4 correction.
3070
3071 2004-02-25  Matt Kraai  <kraai@alumni.cmu.edu>
3072
3073         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Change
3074         ecif->cif->bytes to bytes.
3075         (ffi_prep_cif_machdep): Add braces around nested if statement.
3076
3077 2004-02-09  Alan Modra  <amodra@bigpond.net.au>
3078
3079         * src/types.c (pointer): POWERPC64 has 8 byte pointers.
3080
3081         * src/powerpc/ffi.c (ffi_prep_args64): Correct long double handling.
3082         (ffi_closure_helper_LINUX64): Fix typo.
3083         * testsuite/libffi.call/cls_align_longdouble.c: Pass -mlong-double-128
3084         for powerpc64-*-*.
3085         * testsuite/libffi.call/float.c: Likewise.
3086         * testsuite/libffi.call/float2.c: Likewise.
3087
3088 2004-02-08  Alan Modra  <amodra@bigpond.net.au>
3089
3090         * src/powerpc/ffi.c (ffi_prep_cif_machdep <FFI_LINUX64>): Correct
3091         long double function return and long double arg handling.
3092         (ffi_closure_helper_LINUX64): Formatting.  Delete unused "ng" var.
3093         Use "end_pfr" instead of "nf".  Correct long double handling.
3094         Localise "temp".
3095         * src/powerpc/linux64.S (ffi_call_LINUX64): Save f2 long double
3096         return value.
3097         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Allocate
3098         space for long double return value.  Adjust stack frame and offsets.
3099         Load f2 long double return.
3100
3101 2004-02-07  Alan Modra  <amodra@bigpond.net.au>
3102
3103         * src/types.c: Use 16 byte long double for POWERPC64.
3104
3105 2004-01-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
3106
3107         * src/sparc/ffi.c (ffi_prep_args_v9): Shift the parameter array
3108         when the structure return address is passed in %o0.
3109         (ffi_V9_return_struct): Rename into ffi_v9_layout_struct.
3110         (ffi_v9_layout_struct): Align the field following a nested structure
3111         on a word boundary.  Use memmove instead of memcpy.
3112         (ffi_call): Update call to ffi_V9_return_struct.
3113         (ffi_prep_closure): Define 'ctx' only for V8.
3114         (ffi_closure_sparc_inner): Clone into ffi_closure_sparc_inner_v8
3115         and ffi_closure_sparc_inner_v9.
3116         (ffi_closure_sparc_inner_v8): Return long doubles by reference.
3117         Always skip the structure return address.  For structures and long
3118         doubles, copy the argument directly.
3119         (ffi_closure_sparc_inner_v9): Skip the structure return address only
3120         if required.  Shift the maximum floating-point slot accordingly.  For
3121         big structures, copy the argument directly; otherwise, left-justify the
3122         argument and call ffi_v9_layout_struct to lay out the structure on
3123         the stack.
3124         * src/sparc/v8.S: Undef STACKFRAME before defining it.
3125         (ffi_closure_v8): Pass the structure return address.  Update call to
3126         ffi_closure_sparc_inner_v8.  Short-circuit FFI_TYPE_INT handling.
3127         Skip the 'unimp' insn when returning long doubles and structures.
3128         * src/sparc/v9.S: Undef STACKFRAME before defining it.
3129         (ffi_closure_v9): Increase the frame size by 2 words.  Short-circuit
3130         FFI_TYPE_INT handling.  Load structures both in integers and
3131         floating-point registers on return.
3132         * README: Update status of the SPARC port.
3133
3134 2004-01-24  Andreas Tobler  <a.tobler@schweiz.ch>
3135
3136         * testsuite/libffi.call/pyobjc-tc.c (main): Treat result value
3137         as of type ffi_arg.
3138         * testsuite/libffi.call/struct3.c (main): Fix CHECK.
3139
3140 2004-01-22  Ulrich Weigand  <uweigand@de.ibm.com>
3141
3142         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Treat result
3143         value as of type ffi_arg, not unsigned int.
3144
3145 2004-01-21  Michael Ritzert  <ritzert@t-online.de>
3146
3147         * ffi64.c (ffi_prep_args): Cast the RHS of an assignment instead
3148         of the LHS.
3149
3150 2004-01-12  Andreas Tobler  <a.tobler@schweiz.ch>
3151
3152         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_32 for
3153         Solaris.
3154
3155 2004-01-08  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3156
3157         * testsuite/libffi.call/ffitest.h (allocate_mmap): Cast MAP_FAILED
3158         to void *.
3159
3160 2003-12-10  Richard Henderson  <rth@redhat.com>
3161
3162         * testsuite/libffi.call/cls_align_pointer.c: Cast pointers to
3163         size_t instead of int.
3164
3165 2003-12-04  Hosaka Yuji  <hos@tamanegi.org>
3166
3167         * testsuite/libffi.call/many_win32.c: Include <float.h>.
3168         * testsuite/libffi.call/many_win32.c (main): Replace variable
3169         int i with unsigned long ul.
3170
3171         * testsuite/libffi.call/cls_align_uint64.c: New test case.
3172         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
3173         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
3174         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
3175         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
3176         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
3177         * testsuite/libffi.call/cls_align_float.c: Likewise.
3178         * testsuite/libffi.call/cls_align_double.c: Likewise.
3179         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
3180         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
3181
3182 2003-12-02  Hosaka Yuji  <hos@tamanegi.org>
3183
3184         PR other/13221
3185         * src/x86/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV):
3186         Align arguments to 32 bits.
3187
3188 2003-12-01  Andreas Tobler  <a.tobler@schweiz.ch>
3189
3190         PR other/13221
3191         * testsuite/libffi.call/cls_multi_sshort.c: New test case.
3192         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
3193         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
3194         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
3195         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
3196         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
3197
3198         * testsuite/libffi.special/unwindtest.cc: Cosmetics.
3199
3200 2003-11-26  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
3201
3202         * testsuite/libffi.call/ffitest.h: Include <fcntl.h>.
3203         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3204
3205 2003-11-22  Andreas Tobler  <a.tobler@schweiz.ch>
3206
3207         * Makefile.in: Rebuilt.
3208         * configure: Likewise.
3209         * testsuite/libffi.special/unwindtest.cc: Convert the mmap to
3210         the right type.
3211
3212 2003-11-21  Andreas Jaeger  <aj@suse.de>
3213             Andreas Tobler  <a.tobler@schweiz.ch>
3214
3215         * acinclude.m4: Add AC_FUNC_MMAP_BLACKLIST.
3216         * configure.in: Call AC_FUNC_MMAP_BLACKLIST.
3217         * Makefile.in: Rebuilt.
3218         * aclocal.m4: Likewise.
3219         * configure: Likewise.
3220         * fficonfig.h.in: Likewise.
3221         * testsuite/lib/libffi-dg.exp: Add include dir.
3222         * testsuite/libffi.call/ffitest.h: Add MMAP definitions.
3223         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3224         * testsuite/libffi.call/closure_fn0.c: Use MMAP functionality
3225         for ffi_closure if available.
3226         * testsuite/libffi.call/closure_fn1.c: Likewise.
3227         * testsuite/libffi.call/closure_fn2.c: Likewise.
3228         * testsuite/libffi.call/closure_fn3.c: Likewise.
3229         * testsuite/libffi.call/closure_fn4.c: Likewise.
3230         * testsuite/libffi.call/closure_fn5.c: Likewise.
3231         * testsuite/libffi.call/cls_12byte.c: Likewise.
3232         * testsuite/libffi.call/cls_16byte.c: Likewise.
3233         * testsuite/libffi.call/cls_18byte.c: Likewise.
3234         * testsuite/libffi.call/cls_19byte.c: Likewise.
3235         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3236         * testsuite/libffi.call/cls_20byte.c: Likewise.
3237         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3238         * testsuite/libffi.call/cls_24byte.c: Likewise.
3239         * testsuite/libffi.call/cls_2byte.c: Likewise.
3240         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3241         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3242         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3243         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3244         * testsuite/libffi.call/cls_4byte.c: Likewise.
3245         * testsuite/libffi.call/cls_5byte.c: Likewise.
3246         * testsuite/libffi.call/cls_64byte.c: Likewise.
3247         * testsuite/libffi.call/cls_6byte.c: Likewise.
3248         * testsuite/libffi.call/cls_7byte.c: Likewise.
3249         * testsuite/libffi.call/cls_8byte.c: Likewise.
3250         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3251         * testsuite/libffi.call/cls_9byte2.c: Likewise.
3252         * testsuite/libffi.call/cls_double.c: Likewise.
3253         * testsuite/libffi.call/cls_float.c: Likewise.
3254         * testsuite/libffi.call/cls_schar.c: Likewise.
3255         * testsuite/libffi.call/cls_sint.c: Likewise.
3256         * testsuite/libffi.call/cls_sshort.c: Likewise.
3257         * testsuite/libffi.call/cls_uchar.c: Likewise.
3258         * testsuite/libffi.call/cls_uint.c: Likewise.
3259         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3260         * testsuite/libffi.call/cls_ushort.c: Likewise.
3261         * testsuite/libffi.call/nested_struct.c: Likewise.
3262         * testsuite/libffi.call/nested_struct1.c: Likewise.
3263         * testsuite/libffi.call/nested_struct2.c: Likewise.
3264         * testsuite/libffi.call/nested_struct3.c: Likewise.
3265         * testsuite/libffi.call/problem1.c: Likewise.
3266         * testsuite/libffi.special/unwindtest.cc: Likewise.
3267
3268 2003-11-20  Andreas Tobler  <a.tobler@schweiz.ch>
3269
3270         * testsuite/lib/libffi-dg.exp: Make the -lgcc_s conditional.
3271
3272 2003-11-19  Andreas Tobler  <a.tobler@schweiz.ch>
3273
3274         * testsuite/lib/libffi-dg.exp: Add DYLD_LIBRARY_PATH for darwin.
3275         Add -lgcc_s to additional flags.
3276
3277 2003-11-12  Andreas Tobler  <a.tobler@schweiz.ch>
3278
3279         * configure.in, include/Makefile.am: PR libgcj/11147, install
3280         the ffitarget.h header file in a gcc versioned and target
3281         dependent place.
3282         * configure: Regenerated.
3283         * Makefile.in, include/Makefile.in: Likewise.
3284         * testsuite/Makefile.in: Likewise.
3285
3286 2003-11-09  Andreas Tobler  <a.tobler@schweiz.ch>
3287
3288         * testsuite/libffi.call/closure_fn0.c: Print result and check
3289         with dg-output to make debugging easier.
3290         * testsuite/libffi.call/closure_fn1.c: Likewise.
3291         * testsuite/libffi.call/closure_fn2.c: Likewise.
3292         * testsuite/libffi.call/closure_fn3.c: Likewise.
3293         * testsuite/libffi.call/closure_fn4.c: Likewise.
3294         * testsuite/libffi.call/closure_fn5.c: Likewise.
3295         * testsuite/libffi.call/cls_12byte.c: Likewise.
3296         * testsuite/libffi.call/cls_16byte.c: Likewise.
3297         * testsuite/libffi.call/cls_18byte.c: Likewise.
3298         * testsuite/libffi.call/cls_19byte.c: Likewise.
3299         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3300         * testsuite/libffi.call/cls_20byte.c: Likewise.
3301         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3302         * testsuite/libffi.call/cls_24byte.c: Likewise.
3303         * testsuite/libffi.call/cls_2byte.c: Likewise.
3304         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3305         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3306         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3307         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3308         * testsuite/libffi.call/cls_4byte.c: Likewise.
3309         * testsuite/libffi.call/cls_5byte.c: Likewise.
3310         * testsuite/libffi.call/cls_64byte.c: Likewise.
3311         * testsuite/libffi.call/cls_6byte.c: Likewise.
3312         * testsuite/libffi.call/cls_7byte.c: Likewise.
3313         * testsuite/libffi.call/cls_8byte.c: Likewise.
3314         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3315         * testsuite/libffi.call/cls_9byte2.c: Likewise.
3316         * testsuite/libffi.call/cls_double.c: Likewise.
3317         * testsuite/libffi.call/cls_float.c: Likewise.
3318         * testsuite/libffi.call/cls_schar.c: Likewise.
3319         * testsuite/libffi.call/cls_sint.c: Likewise.
3320         * testsuite/libffi.call/cls_sshort.c: Likewise.
3321         * testsuite/libffi.call/cls_uchar.c: Likewise.
3322         * testsuite/libffi.call/cls_uint.c: Likewise.
3323         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3324         * testsuite/libffi.call/cls_ushort.c: Likewise.
3325         * testsuite/libffi.call/problem1.c: Likewise.
3326
3327         * testsuite/libffi.special/unwindtest.cc: Make ffi_closure
3328         static.
3329
3330 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
3331
3332         * testsuite/libffi.call/cls_9byte2.c: New test case.
3333         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3334         * testsuite/libffi.call/cls_64byte.c: Likewise.
3335         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3336         * testsuite/libffi.call/cls_19byte.c: Likewise.
3337         * testsuite/libffi.call/cls_18byte.c: Likewise.
3338         * testsuite/libffi.call/closure_fn4.c: Likewise.
3339         * testsuite/libffi.call/closure_fn5.c: Likewise.
3340         * testsuite/libffi.call/cls_schar.c: Likewise.
3341         * testsuite/libffi.call/cls_sint.c: Likewise.
3342         * testsuite/libffi.call/cls_sshort.c: Likewise.
3343         * testsuite/libffi.call/nested_struct2.c: Likewise.
3344         * testsuite/libffi.call/nested_struct3.c: Likewise.
3345
3346 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
3347
3348         * testsuite/libffi.call/cls_double.c: Do a check on the result.
3349         * testsuite/libffi.call/cls_uchar.c: Likewise.
3350         * testsuite/libffi.call/cls_uint.c: Likewise.
3351         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3352         * testsuite/libffi.call/cls_ushort.c: Likewise.
3353         * testsuite/libffi.call/return_sc.c: Cleanup whitespaces.
3354
3355 2003-11-06  Andreas Tobler  <a.tobler@schweiz.ch>
3356
3357         * src/prep_cif.c (ffi_prep_cif): Move the validity check after
3358         the initialization.
3359
3360 2003-10-23  Andreas Tobler  <a.tobler@schweiz.ch>
3361
3362         * src/java_raw_api.c (ffi_java_ptrarray_to_raw): Replace
3363         FFI_ASSERT(FALSE) with FFI_ASSERT(0).
3364
3365 2003-10-22  David Daney  <ddaney@avtrex.com>
3366
3367         * src/mips/ffitarget.h: Replace undefined UINT32 and friends with
3368         __attribute__((__mode__(__SI__))) and friends.
3369
3370 2003-10-22  Andreas Schwab  <schwab@suse.de>
3371
3372         * src/ia64/ffi.c: Replace FALSE/TRUE with false/true.
3373
3374 2003-10-21  Andreas Tobler  <a.tobler@schweiz.ch>
3375
3376         * configure.in: AC_LINK_FILES(ffitarget.h).
3377         * configure: Regenerate.
3378         * Makefile.in: Likewise.
3379         * include/Makefile.in: Likewise.
3380         * testsuite/Makefile.in: Likewise.
3381         * fficonfig.h.in: Likewise.
3382
3383 2003-10-21  Paolo Bonzini  <bonzini@gnu.org>
3384             Richard Henderson  <rth@redhat.com>
3385
3386         Avoid that ffi.h includes fficonfig.h.
3387
3388         * Makefile.am (EXTRA_DIST): Include ffitarget.h files
3389         (TARGET_SRC_MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3390         (TARGET_SRC_MIPS_SGI): Removed.
3391         (MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3392         (MIPS_SGI): Removed.
3393         (CLEANFILES): Removed.
3394         (mostlyclean-am, clean-am, mostlyclean-sub, clean-sub): New
3395         targets.
3396         * acconfig.h: Removed.
3397         * configure.in: Compute sizeofs only for double and long double.
3398         Use them to define and subst HAVE_LONG_DOUBLE.  Include comments
3399         into AC_DEFINE instead of using acconfig.h.  Create
3400         include/ffitarget.h instead of include/fficonfig.h.  Rename
3401         MIPS_GCC to MIPS_IRIX, drop MIPS_SGI since we are in gcc's tree.
3402         AC_DEFINE EH_FRAME_FLAGS.
3403         * include/Makefile.am (DISTCLEANFILES): New automake macro.
3404         (hack_DATA): Add ffitarget.h.
3405         * include/ffi.h.in: Remove all system specific definitions.
3406         Declare raw API even if it is not installed, why bother?
3407         Use limits.h instead of SIZEOF_* to define ffi_type_*.  Do
3408         not define EH_FRAME_FLAGS, it is in fficonfig.h now.  Include
3409         ffitarget.h instead of fficonfig.h.  Remove ALIGN macro.
3410         (UINT_ARG, INT_ARG): Removed, use ffi_arg and ffi_sarg instead.
3411         * include/ffi_common.h (bool): Do not define.
3412         (ffi_assert): Accept failed assertion.
3413         (ffi_type_test): Return void and accept file/line.
3414         (FFI_ASSERT): Pass stringized failed assertion.
3415         (FFI_ASSERT_AT): New macro.
3416         (FFI_ASSERT_VALID_TYPE): New macro.
3417         (UINT8, SINT8, UINT16, SINT16, UINT32, SINT32,
3418         UINT64, SINT64): Define here with gcc's __attribute__ macro
3419         instead of in ffi.h
3420         (FLOAT32, ALIGN): Define here instead of in ffi.h
3421         * include/ffi-mips.h: Removed.  Its content moved to
3422         src/mips/ffitarget.h after separating assembly and C sections.
3423         * src/alpha/ffi.c, src/alpha/ffi.c, src/java_raw_api.c
3424         src/prep_cif.c, src/raw_api.c, src/ia64/ffi.c,
3425         src/mips/ffi.c, src/mips/n32.S, src/mips/o32.S,
3426         src/mips/ffitarget.h, src/sparc/ffi.c, src/x86/ffi64.c:
3427         SIZEOF_ARG -> FFI_SIZEOF_ARG.
3428         * src/ia64/ffi.c: Include stdbool.h (provided by GCC 2.95+).
3429         * src/debug.c (ffi_assert): Accept stringized failed assertion.
3430         (ffi_type_test): Rewritten.
3431         * src/prep-cif.c (initialize_aggregate, ffi_prep_cif): Call
3432         FFI_ASSERT_VALID_TYPE.
3433         * src/alpha/ffitarget.h, src/arm/ffitarget.h,
3434         src/ia64/ffitarget.h, src/m68k/ffitarget.h,
3435         src/mips/ffitarget.h, src/powerpc/ffitarget.h,
3436         src/s390/ffitarget.h, src/sh/ffitarget.h,
3437         src/sh64/ffitarget.h, src/sparc/ffitarget.h,
3438         src/x86/ffitarget.h: New files.
3439         * src/alpha/osf.S, src/arm/sysv.S, src/ia64/unix.S,
3440         src/m68k/sysv.S, src/mips/n32.S, src/mips/o32.S,
3441         src/powerpc/aix.S, src/powerpc/darwin.S,
3442         src/powerpc/ffi_darwin.c, src/powerpc/linux64.S,
3443         src/powerpc/linux64_closure.S, src/powerpc/ppc_closure.S,
3444         src/powerpc/sysv.S, src/s390/sysv.S, src/sh/sysv.S,
3445         src/sh64/sysv.S, src/sparc/v8.S, src/sparc/v9.S,
3446         src/x86/sysv.S, src/x86/unix64.S, src/x86/win32.S:
3447         include fficonfig.h
3448
3449 2003-10-20  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3450
3451         * src/mips/ffi.c: Use _ABIN32, _ABIO32 instead of external
3452         _MIPS_SIM_NABI32, _MIPS_SIM_ABI32.
3453
3454 2003-10-19  Andreas Tobler  <a.tobler@schweiz.ch>
3455
3456         * src/powerpc/ffi_darwin.c (ffi_prep_args): Declare bytes again.
3457         Used when FFI_DEBUG = 1.
3458
3459 2003-10-14  Alan Modra  <amodra@bigpond.net.au>
3460
3461         * src/types.c (double, longdouble): Default POWERPC64 to 8 byte size
3462         and align.
3463
3464 2003-10-06  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3465
3466         * include/ffi_mips.h: Define FFI_MIPS_N32 for N32/N64 ABIs,
3467         FFI_MIPS_O32 for O32 ABI.
3468
3469 2003-10-01  Andreas Tobler  <a.tobler@schweiz.ch>
3470
3471         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_64 for
3472         SPARC64. Cleanup whitespaces.
3473
3474 2003-09-19  Andreas Tobler  <a.tobler@schweiz.ch>
3475
3476         * testsuite/libffi.call/closure_fn0.c: Xfail mips, arm,
3477         strongarm, xscale. Cleanup whitespaces.
3478         * testsuite/libffi.call/closure_fn1.c: Likewise.
3479         * testsuite/libffi.call/closure_fn2.c: Likewise.
3480         * testsuite/libffi.call/closure_fn3.c: Likewise.
3481         * testsuite/libffi.call/cls_12byte.c: Likewise.
3482         * testsuite/libffi.call/cls_16byte.c: Likewise.
3483         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3484         * testsuite/libffi.call/cls_20byte.c: Likewise.
3485         * testsuite/libffi.call/cls_24byte.c: Likewise.
3486         * testsuite/libffi.call/cls_2byte.c: Likewise.
3487         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3488         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3489         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3490         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3491         * testsuite/libffi.call/cls_4byte.c: Likewise.
3492         * testsuite/libffi.call/cls_5byte.c: Likewise.
3493         * testsuite/libffi.call/cls_6byte.c: Likewise.
3494         * testsuite/libffi.call/cls_7byte.c: Likewise.
3495         * testsuite/libffi.call/cls_8byte.c: Likewise.
3496         * testsuite/libffi.call/cls_double.c: Likewise.
3497         * testsuite/libffi.call/cls_float.c: Likewise.
3498         * testsuite/libffi.call/cls_uchar.c: Likewise.
3499         * testsuite/libffi.call/cls_uint.c: Likewise.
3500         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3501         * testsuite/libffi.call/cls_ushort.c: Likewise.
3502         * testsuite/libffi.call/nested_struct.c: Likewise.
3503         * testsuite/libffi.call/nested_struct1.c: Likewise.
3504         * testsuite/libffi.call/problem1.c: Likewise.
3505         * testsuite/libffi.special/unwindtest.cc: Likewise.
3506         * testsuite/libffi.call/pyobjc-tc.c: Cleanup whitespaces.
3507
3508 2003-09-18  David Edelsohn  <edelsohn@gnu.org>
3509
3510         * src/powerpc/aix.S: Cleanup whitespaces.
3511         * src/powerpc/aix_closure.S: Likewise.
3512
3513 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3514
3515         * src/powerpc/darwin.S: Cleanup whitespaces, comment formatting.
3516         * src/powerpc/darwin_closure.S: Likewise.
3517         * src/powerpc/ffi_darwin.c: Likewise.
3518
3519 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3520             David Edelsohn  <edelsohn@gnu.org>
3521
3522         * src/types.c (double): Add AIX and Darwin to the right TYPEDEF.
3523         * src/powerpc/aix_closure.S: Remove the pointer to the outgoing
3524         parameter stack.
3525         * src/powerpc/darwin_closure.S: Likewise.
3526         * src/powerpc/ffi_darwin.c (ffi_prep_args): Handle structures
3527         according to the Darwin/AIX ABI.
3528         (ffi_prep_cif_machdep): Likewise.
3529         (ffi_closure_helper_DARWIN): Likewise.
3530         Remove the outgoing parameter stack logic. Simplify the evaluation
3531         of the different CASE types.
3532         (ffi_prep_clousure): Avoid the casts on lvalues. Change the branch
3533         statement in the trampoline code.
3534
3535 2003-09-18  Kaz Kojima  <kkojima@gcc.gnu.org>
3536
3537         * src/sh/ffi.c (ffi_prep_args): Take account into the alignement
3538         for the register size.
3539         (ffi_closure_helper_SYSV): Handle the structure return value
3540         address correctly.
3541         (ffi_closure_helper_SYSV): Return the appropriate type when
3542         the registers are used for the structure return value.
3543         * src/sh/sysv.S (ffi_closure_SYSV): Fix the stack layout for
3544         the 64-bit return value.  Update copyright years.
3545
3546 2003-09-17  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3547
3548         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Search in
3549         srcdir for ffi_mips.h.
3550
3551 2003-09-12  Alan Modra  <amodra@bigpond.net.au>
3552
3553         * src/prep_cif.c (initialize_aggregate): Include tail padding in
3554         structure size.
3555         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Correct
3556         placement of float result.
3557         * testsuite/libffi.special/unwindtest.cc (closure_test_fn1): Correct
3558         cast of "resp" for big-endian 64 bit machines.
3559
3560 2003-09-11  Alan Modra  <amodra@bigpond.net.au>
3561
3562         * src/types.c (double, longdouble): Merge identical SH and ARM
3563         typedefs, and add POWERPC64.
3564         * src/powerpc/ffi.c (ffi_prep_args64): Correct next_arg calc for
3565         struct split over gpr and rest.
3566         (ffi_prep_cif_machdep): Correct intarg_count for structures.
3567         * src/powerpc/linux64.S (ffi_call_LINUX64): Fix gpr offsets.
3568
3569 2003-09-09  Andreas Tobler  <a.tobler@schweiz.ch>
3570
3571         * src/powerpc/ffi.c (ffi_closure_helper_SYSV) Handle struct
3572         passing correctly.
3573
3574 2003-09-09  Alan Modra  <amodra@bigpond.net.au>
3575
3576         * configure: Regenerate.
3577
3578 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3579
3580         * Makefile.am: Remove build rules for ffitest.
3581         * Makefile.in: Rebuilt.
3582
3583 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3584
3585         * src/java_raw_api.c: Include <stdlib.h> to fix compiler warning
3586         about implicit declaration of abort().
3587
3588 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3589
3590         * Makefile.am: Add dejagnu test framework. Fixes PR other/11411.
3591         * Makefile.in: Rebuilt.
3592         * configure.in: Add dejagnu test framework.
3593         * configure: Rebuilt.
3594
3595         * testsuite/Makefile.am: New file.
3596         * testsuite/Makefile.in: Built
3597         * testsuite/lib/libffi-dg.exp: New file.
3598         * testsuite/config/default.exp: Likewise.
3599         * testsuite/libffi.call/call.exp: Likewise.
3600         * testsuite/libffi.call/ffitest.h: Likewise.
3601         * testsuite/libffi.call/closure_fn0.c: Likewise.
3602         * testsuite/libffi.call/closure_fn1.c: Likewise.
3603         * testsuite/libffi.call/closure_fn2.c: Likewise.
3604         * testsuite/libffi.call/closure_fn3.c: Likewise.
3605         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3606         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3607         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3608         * testsuite/libffi.call/cls_2byte.c: Likewise.
3609         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3610         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3611         * testsuite/libffi.call/cls_4byte.c: Likewise.
3612         * testsuite/libffi.call/cls_5byte.c: Likewise.
3613         * testsuite/libffi.call/cls_6byte.c: Likewise.
3614         * testsuite/libffi.call/cls_7byte.c: Likewise.
3615         * testsuite/libffi.call/cls_8byte.c: Likewise.
3616         * testsuite/libffi.call/cls_12byte.c: Likewise.
3617         * testsuite/libffi.call/cls_16byte.c: Likewise.
3618         * testsuite/libffi.call/cls_20byte.c: Likewise.
3619         * testsuite/libffi.call/cls_24byte.c: Likewise.
3620         * testsuite/libffi.call/cls_double.c: Likewise.
3621         * testsuite/libffi.call/cls_float.c: Likewise.
3622         * testsuite/libffi.call/cls_uchar.c: Likewise.
3623         * testsuite/libffi.call/cls_uint.c: Likewise.
3624         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3625         * testsuite/libffi.call/cls_ushort.c: Likewise.
3626         * testsuite/libffi.call/float.c: Likewise.
3627         * testsuite/libffi.call/float1.c: Likewise.
3628         * testsuite/libffi.call/float2.c: Likewise.
3629         * testsuite/libffi.call/many.c: Likewise.
3630         * testsuite/libffi.call/many_win32.c: Likewise.
3631         * testsuite/libffi.call/nested_struct.c: Likewise.
3632         * testsuite/libffi.call/nested_struct1.c: Likewise.
3633         * testsuite/libffi.call/pyobjc-tc.c: Likewise.
3634         * testsuite/libffi.call/problem1.c: Likewise.
3635         * testsuite/libffi.call/promotion.c: Likewise.
3636         * testsuite/libffi.call/return_ll.c: Likewise.
3637         * testsuite/libffi.call/return_sc.c: Likewise.
3638         * testsuite/libffi.call/return_uc.c: Likewise.
3639         * testsuite/libffi.call/strlen.c: Likewise.
3640         * testsuite/libffi.call/strlen_win32.c: Likewise.
3641         * testsuite/libffi.call/struct1.c: Likewise.
3642         * testsuite/libffi.call/struct2.c: Likewise.
3643         * testsuite/libffi.call/struct3.c: Likewise.
3644         * testsuite/libffi.call/struct4.c: Likewise.
3645         * testsuite/libffi.call/struct5.c: Likewise.
3646         * testsuite/libffi.call/struct6.c: Likewise.
3647         * testsuite/libffi.call/struct7.c: Likewise.
3648         * testsuite/libffi.call/struct8.c: Likewise.
3649         * testsuite/libffi.call/struct9.c: Likewise.
3650         * testsuite/libffi.special/special.exp: New file.
3651         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3652         * testsuite/libffi.special/unwindtest.cc: Likewise.
3653
3654
3655 2003-08-13  Kaz Kojima  <kkojima@gcc.gnu.org>
3656
3657         * src/sh/ffi.c (OFS_INT16): Set 0 for little endian case.  Update
3658         copyright years.
3659
3660 2003-08-02  Alan Modra  <amodra@bigpond.net.au>
3661
3662         * src/powerpc/ffi.c (ffi_prep_args64): Modify for changed gcc
3663         structure passing.
3664         (ffi_closure_helper_LINUX64): Likewise.
3665         * src/powerpc/linux64.S: Remove code writing to parm save area.
3666         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Use return
3667         address in lr from ffi_closure_helper_LINUX64 call to calculate
3668         table address.  Optimize function tail.
3669
3670 2003-07-28  Andreas Tobler  <a.tobler@schweiz.ch>
3671
3672         * src/sparc/ffi.c: Handle all floating point registers.
3673         * src/sparc/v9.S: Likewise. Fixes second part of PR target/11410.
3674
3675 2003-07-11  Gerald Pfeifer  <pfeifer@dbai.tuwien.ac.at>
3676
3677         * README: Note that libffi is not part of GCC.  Update the project
3678         URL and status.
3679
3680 2003-06-19  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3681
3682         * src/powerpc/ppc_closure.S: Include ffi.h.
3683
3684 2003-06-13  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3685
3686         * src/x86/sysv.S: Avoid gas-only .uleb128/.sleb128 directives.
3687         Use C style comments.
3688
3689 2003-06-13  Kaz Kojima  <kkojima@rr.iij4u.or.jp>
3690
3691         * Makefile.am: Add SHmedia support.  Fix a typo of SH support.
3692         * Makefile.in: Regenerate.
3693         * configure.in (sh64-*-linux*, sh5*-*-linux*): Add target.
3694         * configure: Regenerate.
3695         * include/ffi.h.in: Add SHmedia support.
3696         * src/sh64/ffi.c: New file.
3697         * src/sh64/sysv.S: New file.
3698
3699 2003-05-16  Jakub Jelinek  <jakub@redhat.com>
3700
3701         * configure.in (HAVE_RO_EH_FRAME): Check whether .eh_frame section
3702         should be read-only.
3703         * configure: Rebuilt.
3704         * fficonfig.h.in: Rebuilt.
3705         * include/ffi.h.in (EH_FRAME_FLAGS): Define.
3706         * src/alpha/osf.S: Use EH_FRAME_FLAGS.
3707         * src/powerpc/linux64.S: Likewise.
3708         * src/powerpc/linux64_closure.S: Likewise.  Include ffi.h.
3709         * src/powerpc/sysv.S: Use EH_FRAME_FLAGS.  Use pcrel encoding
3710         if -fpic/-fPIC/-mrelocatable.
3711         * src/powerpc/powerpc_closure.S: Likewise.
3712         * src/sparc/v8.S: If HAVE_RO_EH_FRAME is defined, don't include
3713         #write in .eh_frame flags.
3714         * src/sparc/v9.S: Likewise.
3715         * src/x86/unix64.S: Use EH_FRAME_FLAGS.
3716         * src/x86/sysv.S: Likewise.  Use pcrel encoding if -fpic/-fPIC.
3717         * src/s390/sysv.S: Use EH_FRAME_FLAGS.  Include ffi.h.
3718
3719 2003-05-07  Jeff Sturm  <jsturm@one-point.com>
3720
3721         Fixes PR bootstrap/10656
3722         * configure.in (HAVE_AS_REGISTER_PSEUDO_OP): Test assembler
3723         support for .register pseudo-op.
3724         * src/sparc/v8.S: Use it.
3725         * fficonfig.h.in: Rebuilt.
3726         * configure: Rebuilt.
3727
3728 2003-04-18  Jakub Jelinek  <jakub@redhat.com>
3729
3730         * include/ffi.h.in (POWERPC64): Define if 64-bit.
3731         (enum ffi_abi): Add FFI_LINUX64 on POWERPC.
3732         Make it the default on POWERPC64.
3733         (FFI_TRAMPOLINE_SIZE): Define to 24 on POWERPC64.
3734         * configure.in: Change powerpc-*-linux* into powerpc*-*-linux*.
3735         * configure: Rebuilt.
3736         * src/powerpc/ffi.c (hidden): Define.
3737         (ffi_prep_args_SYSV): Renamed from
3738         ffi_prep_args.  Cast pointers to unsigned long to shut up warnings.
3739         (NUM_GPR_ARG_REGISTERS64, NUM_FPR_ARG_REGISTERS64,
3740         ASM_NEEDS_REGISTERS64): New.
3741         (ffi_prep_args64): New function.
3742         (ffi_prep_cif_machdep): Handle FFI_LINUX64 ABI.
3743         (ffi_call): Likewise.
3744         (ffi_prep_closure): Likewise.
3745         (flush_icache): Surround by #ifndef POWERPC64.
3746         (ffi_dblfl): New union type.
3747         (ffi_closure_helper_SYSV): Use it to avoid aliasing problems.
3748         (ffi_closure_helper_LINUX64): New function.
3749         * src/powerpc/ppc_closure.S: Surround whole file by #ifndef
3750         __powerpc64__.
3751         * src/powerpc/sysv.S: Likewise.
3752         (ffi_call_SYSV): Rename ffi_prep_args to ffi_prep_args_SYSV.
3753         * src/powerpc/linux64.S: New file.
3754         * src/powerpc/linux64_closure.S: New file.
3755         * Makefile.am (EXTRA_DIST): Add src/powerpc/linux64.S and
3756         src/powerpc/linux64_closure.S.
3757         (TARGET_SRC_POWERPC): Likewise.
3758
3759         * src/ffitest.c (closure_test_fn, closure_test_fn1, closure_test_fn2,
3760         closure_test_fn3): Fix result printing on big-endian 64-bit
3761         machines.
3762         (main): Print tst2_arg instead of uninitialized tst2_result.
3763
3764         * src/ffitest.c (main): Hide what closure pointer really points to
3765         from the compiler.
3766
3767 2003-04-16  Richard Earnshaw  <rearnsha@arm.com>
3768
3769         * configure.in (arm-*-netbsdelf*): Add configuration.
3770         (configure): Regenerated.
3771
3772 2003-04-04  Loren J. Rittle  <ljrittle@acm.org>
3773
3774         * include/Makefile.in: Regenerate.
3775
3776 2003-03-21  Zdenek Dvorak  <rakdver@atrey.karlin.mff.cuni.cz>
3777
3778         * libffi/include/ffi.h.in: Define X86 instead of X86_64 in 32
3779         bit mode.
3780         * libffi/src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV):
3781         Receive closure pointer through parameter, read args using
3782         __builtin_dwarf_cfa.
3783         (FFI_INIT_TRAMPOLINE): Send closure reference through eax.
3784
3785 2003-03-12  Andreas Schwab  <schwab@suse.de>
3786
3787         * configure.in: Avoid trailing /. in toolexeclibdir.
3788         * configure: Rebuilt.
3789
3790 2003-03-03  Andreas Tobler <a.tobler@schweiz.ch>
3791
3792         * src/powerpc/darwin_closure.S: Recode to fit dynamic libraries.
3793
3794 2003-02-06  Andreas Tobler <a.tobler@schweiz.ch>
3795
3796         * libffi/src/powerpc/darwin_closure.S:
3797         Fix alignement bug, allocate 8 bytes for the result.
3798         * libffi/src/powerpc/aix_closure.S:
3799         Likewise.
3800         * libffi/src/powerpc/ffi_darwin.c:
3801         Update stackframe description for aix/darwin_closure.S.
3802
3803 2003-02-06  Jakub Jelinek  <jakub@redhat.com>
3804
3805         * src/s390/ffi.c (ffi_closure_helper_SYSV): Add hidden visibility
3806         attribute.
3807
3808 2003-01-31  Christian Cornelssen  <ccorn@cs.tu-berlin.de>,
3809             Andreas Schwab  <schwab@suse.de>
3810
3811         * configure.in: Adjust command to source config-ml.in to account
3812         for changes to the libffi_basedir definition.
3813         (libffi_basedir): Remove ${srcdir} from value and include trailing
3814         slash if nonempty.
3815
3816         * configure: Regenerate.
3817
3818 2003-01-29  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3819
3820         * src/powerpc/ppc_closure.S: Recode to fit shared libs.
3821
3822 2003-01-28  Andrew Haley  <aph@redhat.com>
3823
3824         * include/ffi.h.in: Enable FFI_CLOSURES for x86_64.
3825         * src/x86/ffi64.c (ffi_prep_closure): New.
3826         (ffi_closure_UNIX64_inner): New.
3827         * src/x86/unix64.S (ffi_closure_UNIX64): New.
3828
3829 2003-01-27  Alexandre Oliva  <aoliva@redhat.com>
3830
3831         * configure.in (toolexecdir, toolexeclibdir): Set and AC_SUBST.
3832         Remove USE_LIBDIR conditional.
3833         * Makefile.am (toolexecdir, toolexeclibdir): Don't override.
3834         * Makefile.in, configure: Rebuilt.
3835
3836 2003-01027  David Edelsohn  <edelsohn@gnu.org>
3837
3838         * Makefile.am (TARGET_SRC_POWERPC_AIX): Fix typo.
3839         * Makefile.in: Regenerate.
3840
3841 2003-01-22  Andrew Haley  <aph@redhat.com>
3842
3843         * src/powerpc/darwin.S (_ffi_call_AIX): Add Augmentation size to
3844         unwind info.
3845
3846 2003-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
3847
3848         * src/powerpc/darwin.S: Add unwind info.
3849         * src/powerpc/darwin_closure.S: Likewise.
3850
3851 2003-01-14  Andrew Haley  <aph@redhat.com>
3852
3853         * src/x86/ffi64.c (ffi_prep_args): Check for void retval.
3854         (ffi_prep_cif_machdep): Likewise.
3855         * src/x86/unix64.S: Add unwind info.
3856
3857 2003-01-14  Andreas Jaeger  <aj@suse.de>
3858
3859         * src/ffitest.c (main): Only use ffi_closures if those are
3860         supported.
3861
3862 2003-01-13 Andreas Tobler <a.tobler@schweiz.ch>
3863
3864         * libffi/src/ffitest.c
3865          add closure testcases
3866
3867 2003-01-13 Kevin B. Hendricks <khendricks@ivey.uwo.ca>
3868
3869         * libffi/src/powerpc/ffi.c
3870          fix alignment bug for float (4 byte aligned iso 8 byte)
3871
3872 2003-01-09  Geoffrey Keating  <geoffk@apple.com>
3873
3874         * src/powerpc/ffi_darwin.c: Remove RCS version string.
3875         * src/powerpc/darwin.S: Remove RCS version string.
3876
3877 2003-01-03  Jeff Sturm  <jsturm@one-point.com>
3878
3879         * include/ffi.h.in: Add closure defines for SPARC, SPARC64.
3880         * src/ffitest.c (main): Use static storage for closure.
3881         * src/sparc/ffi.c (ffi_prep_closure, ffi_closure_sparc_inner): New.
3882         * src/sparc/v8.S (ffi_closure_v8): New.
3883         * src/sparc/v9.S (ffi_closure_v9): New.
3884
3885 2002-11-10  Ranjit Mathew <rmathew@hotmail.com>
3886
3887         * include/ffi.h.in: Added FFI_STDCALL ffi_type
3888           enumeration for X86_WIN32.
3889         * src/x86/win32.S: Added ffi_call_STDCALL function
3890           definition.
3891         * src/x86/ffi.c (ffi_call/ffi_raw_call): Added
3892           switch cases for recognising FFI_STDCALL and
3893           calling ffi_call_STDCALL if target is X86_WIN32.
3894         * src/ffitest.c (my_stdcall_strlen/stdcall_many):
3895           stdcall versions of the "my_strlen" and "many"
3896           test functions (for X86_WIN32).
3897           Added test cases to test stdcall invocation using
3898           these functions.
3899
3900 2002-12-02  Kaz Kojima  <kkojima@gcc.gnu.org>
3901
3902         * src/sh/sysv.S: Add DWARF2 unwind info.
3903
3904 2002-11-27  Ulrich Weigand  <uweigand@de.ibm.com>
3905
3906         * src/s390/sysv.S (.eh_frame section): Make section read-only.
3907
3908 2002-11-26  Jim Wilson  <wilson@redhat.com>
3909
3910         * src/types.c (FFI_TYPE_POINTER): Has size 8 on IA64.
3911
3912 2002-11-23  H.J. Lu <hjl@gnu.org>
3913
3914         * acinclude.m4: Add dummy AM_PROG_LIBTOOL.
3915         Include ../config/accross.m4.
3916         * aclocal.m4; Rebuild.
3917         * configure: Likewise.
3918
3919 2002-11-15  Ulrich Weigand  <uweigand@de.ibm.com>
3920
3921         * src/s390/sysv.S (.eh_frame section): Adapt to pcrel FDE encoding.
3922
3923 2002-11-11  DJ Delorie  <dj@redhat.com>
3924
3925         * configure.in: Look for common files in the right place.
3926
3927 2002-10-08  Ulrich Weigand  <uweigand@de.ibm.com>
3928
3929         * src/java_raw_api.c (ffi_java_raw_to_ptrarray): Interpret
3930         raw data as _Jv_word values, not ffi_raw.
3931         (ffi_java_ptrarray_to_raw): Likewise.
3932         (ffi_java_rvalue_to_raw): New function.
3933         (ffi_java_raw_call): Call it.
3934         (ffi_java_raw_to_rvalue): New function.
3935         (ffi_java_translate_args): Call it.
3936         * src/ffitest.c (closure_test_fn): Interpret return value
3937         as ffi_arg, not int.
3938         * src/s390/ffi.c (ffi_prep_cif_machdep): Add missing
3939         FFI_TYPE_POINTER case.
3940         (ffi_closure_helper_SYSV): Likewise.  Also, assume return
3941         values extended to word size.
3942
3943 2002-10-02  Andreas Jaeger  <aj@suse.de>
3944
3945         * src/x86/ffi64.c (ffi_prep_cif_machdep): Remove debug output.
3946
3947 2002-10-01  Bo Thorsen  <bo@smetana.suse.de>
3948
3949         * include/ffi.h.in: Fix i386 win32 compilation.
3950
3951 2002-09-30  Ulrich Weigand  <uweigand@de.ibm.com>
3952
3953         * configure.in: Add s390x-*-linux-* target.
3954         * configure: Regenerate.
3955         * include/ffi.h.in: Define S390X for s390x targets.
3956         (FFI_CLOSURES): Define for s390/s390x.
3957         (FFI_TRAMPOLINE_SIZE): Likewise.
3958         (FFI_NATIVE_RAW_API): Likewise.
3959         * src/prep_cif.c (ffi_prep_cif): Do not compute stack space for s390.
3960         * src/types.c (FFI_TYPE_POINTER): Use 8-byte pointers on s390x.
3961         * src/s390/ffi.c: Major rework of existing code.  Add support for
3962         s390x targets.  Add closure support.
3963         * src/s390/sysv.S: Likewise.
3964
3965 2002-09-29  Richard Earnshaw  <rearnsha@arm.com>
3966
3967         * src/arm/sysv.S: Fix typo.
3968
3969 2002-09-28  Richard Earnshaw  <rearnsha@arm.com>
3970
3971         * src/arm/sysv.S: If we don't have machine/asm.h and the pre-processor
3972         has defined __USER_LABEL_PREFIX__, then use it in CNAME.
3973         (ffi_call_SYSV): Handle soft-float.
3974
3975 2002-09-27  Bo Thorsen  <bo@suse.de>
3976
3977         * include/ffi.h.in: Fix multilib x86-64 support.
3978
3979 2002-09-22  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
3980
3981         * Makefile.am (all-multi): Fix multilib parallel build.
3982
3983 2002-07-19  Kaz Kojima  <kkojima@gcc.gnu.org>
3984
3985         * configure.in (sh[34]*-*-linux*): Add brackets.
3986         * configure: Regenerate.
3987
3988 2002-07-18  Kaz Kojima  <kkojima@gcc.gnu.org>
3989
3990         * Makefile.am: Add SH support.
3991         * Makefile.in: Regenerate.
3992         * configure.in (sh-*-linux*, sh[34]*-*-linux*): Add target.
3993         * configure: Regenerate.
3994         * include/ffi.h.in: Add SH support.
3995         * src/sh/ffi.c: New file.
3996         * src/sh/sysv.S: New file.
3997         * src/types.c: Add SH support.
3998
3999 2002-07-16  Bo Thorsen  <bo@suse.de>
4000
4001         * src/x86/ffi64.c: New file that adds x86-64 support.
4002         * src/x86/unix64.S: New file that handles argument setup for
4003         x86-64.
4004         * src/x86/sysv.S: Don't use this on x86-64.
4005         * src/x86/ffi.c: Don't use this on x86-64.
4006         Remove unused vars.
4007         * src/prep_cif.c (ffi_prep_cif): Don't do stack size calculation
4008         for x86-64.
4009         * src/ffitest.c (struct6): New test that tests a special case in
4010         the x86-64 ABI.
4011         (struct7): Likewise.
4012         (struct8): Likewise.
4013         (struct9): Likewise.
4014         (closure_test_fn): Silence warning about this when it's not used.
4015         (main): Add the new tests.
4016         (main): Fix a couple of wrong casts and silence some compiler warnings.
4017         * include/ffi.h.in: Add x86-64 ABI definition.
4018         * fficonfig.h.in: Regenerate.
4019         * Makefile.am: Add x86-64 support.
4020         * configure.in: Likewise.
4021         * Makefile.in: Regenerate.
4022         * configure: Likewise.
4023
4024 2002-06-24  Bo Thorsen  <bo@suse.de>
4025
4026         * src/types.c: Merge settings for similar architectures.
4027         Add x86-64 sizes and alignments.
4028
4029 2002-06-23  Bo Thorsen  <bo@suse.de>
4030
4031         * src/arm/ffi.c (ffi_prep_args): Remove unused vars.
4032         * src/sparc/ffi.c (ffi_prep_args_v8): Likewise.
4033         * src/mips/ffi.c (ffi_prep_args): Likewise.
4034         * src/m68k/ffi.c (ffi_prep_args): Likewise.
4035
4036 2002-07-18  H.J. Lu  (hjl@gnu.org)
4037
4038         * Makefile.am (TARGET_SRC_MIPS_LINUX): New.
4039         (libffi_la_SOURCES): Support MIPS_LINUX.
4040         (libffi_convenience_la_SOURCES): Likewise.
4041         * Makefile.in: Regenerated.
4042