Many new patches
[libffi.git] / .pc / ppc64-darwin / 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 2011-02-13  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
33
34         * configure: Regenerate.
35
36 2011-02-09  Anthony Green <green@moxielogic.com>
37
38         * README: Mention Interix.
39
40 2011-02-09  Jonathan Callen  <abcd@gentoo.org>
41
42         * configure.ac: Add Interix to win32/cygwin/mingw case.
43         * configure: Ditto.
44         * src/closures.c: Treat Interix like Cygwin, instead of as a
45         generic win32.
46
47 2011-02-09  Anthony Green <green@moxielogic.com>
48
49         * testsuite/libffi.call/err_bad_typedef.c: Remove xfail.
50         * testsuite/libffi.call/err_bad_abi.c: Remove xfail.
51         * src/x86/ffi64.c (UNLIKELY, LIKELY): Define.
52         (ffi_prep_closure_loc): Check for bad ABI.
53         * src/prep_cif.c (UNLIKELY, LIKELY): Define.
54         (initialize_aggregate): Check for bad types.
55
56 2011-02-09  Landon Fuller <landonf@macports.org>
57
58         * Makefile.am (EXTRA_DIST): Add build-ios.sh, src/arm/gentramp.sh,
59         src/arm/trampoline.S.
60         (nodist_libffi_la_SOURCES): Add src/arc/trampoline.S.
61         * configure.ac (FFI_EXEC_TRAMPOLINE_TABLE): Define.
62         * src/arm/ffi.c (ffi_trampoline_table)
63         (ffi_closure_trampoline_table_page, ffi_trampoline_table_entry)
64         (FFI_TRAMPOLINE_CODELOC_CONFIG, FFI_TRAMPOLINE_CONFIG_PAGE_OFFSET)
65         (FFI_TRAMPOLINE_COUNT, ffi_trampoline_lock, ffi_trampoline_tables)
66         (ffi_trampoline_table_alloc, ffi_closure_alloc, ffi_closure_free):
67         Define for FFI_EXEC_TRAMPOLINE_TABLE case (iOS).
68         (ffi_prep_closure_loc): Handl FFI_EXEC_TRAMPOLINE_TABLE case
69         separately.
70         * src/arm/sysv.S: Handle Apple iOS host.
71         * src/closures.c: Handle FFI_EXEC_TRAMPOLINE_TABLE case.
72         * build-ios.sh: New file.
73         * fficonfig.h.in, configure, Makefile.in: Rebuilt.
74         * README: Mention ARM iOS.
75
76 2011-02-08  Oren Held  <orenhe@il.ibm.com>
77
78         * src/dlmalloc.c (_STRUCT_MALLINFO): Define in order to avoid
79         redefinition of mallinfo on HP-UX.
80
81 2011-02-08  Ginn Chen  <ginn.chen@oracle.com>
82
83         * src/sparc/ffi.c (ffi_call): Make compatible with Solaris Studio
84         aggregate return ABI.  Flush cache.
85         (ffi_prep_closure_loc): Flush cache.
86
87 2011-02-11  Anthony Green  <green@moxielogic.com>
88
89         From Tom Honermann <tom.honermann@oracle.com>:
90         * src/powerpc/aix.S (ffi_call_AIX): Support for xlc toolchain on
91         AIX.  Declare .ffi_prep_args.  Insert nops after branch
92         instructions so that the AIX linker can insert TOC reload
93         instructions.
94         * src/powerpc/aix_closure.S: Declare .ffi_closure_helper_DARWIN.
95
96 2011-02-08  Ed  <ed@kdtc.net>
97
98         * src/powerpc/asm.h: Fix grammar nit in comment.
99
100 2011-02-08  Uli Link  <ul.mcamafia@linkitup.de>
101
102         * include/ffi.h.in (FFI_64_BIT_MAX): Define and use.
103
104 2011-02-09  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
105
106         PR libffi/46661
107         * testsuite/libffi.call/cls_pointer.c (main): Cast void * to
108         uintptr_t first.
109         * testsuite/libffi.call/cls_pointer_stack.c (main): Likewise.
110
111 2011-02-08  Rafael Avila de Espindola  <respindola@mozilla.com>
112
113         * configure.ac: Fix x86 test for pc related relocs.
114         * configure: Rebuilt.
115
116 2011-02-07  Joel Sherrill <joel.sherrill@oarcorp.com>
117
118         * libffi/src/m68k/ffi.c: Add RTEMS support for cache flushing.
119         Handle case when CPU variant does not have long double support.
120         * libffi/src/m68k/sysv.S: Add support for mc68000, Coldfire,
121         and cores with soft floating point.
122
123 2011-02-07  Joel Sherrill <joel.sherrill@oarcorp.com>
124
125         * configure.ac: Add mips*-*-rtems* support.
126         * configure: Regenerate.
127         * src/mips/ffitarget.h: Ensure needed constants are available
128         for targets which do not have sgidefs.h.
129
130 2011-01-26  Dave Korn  <dave.korn.cygwin@gmail.com>
131
132         PR target/40125
133         * configure.ac (AM_LTLDFLAGS): Add -bindir option for windows DLLs.
134         * configure: Regenerate.
135
136 2010-12-18  Iain Sandoe  <iains@gcc.gnu.org>
137
138         PR libffi/29152
139         PR libffi/42378
140         * src/powerpc/darwin_closure.S: Provide Darwin64 implementation,
141         update comments.
142         * src/powerpc/ffitarget.h (POWERPC_DARWIN64): New,
143         (FFI_TRAMPOLINE_SIZE): Update for Darwin64.
144         * src/powerpc/darwin.S: Provide Darwin64 implementation,
145         update comments.
146         * src/powerpc/ffi_darwin.c: Likewise.
147
148 2010-12-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
149
150         * configure.ac (libffi_cv_as_ascii_pseudo_op): Use double
151         backslashes.
152         (libffi_cv_as_string_pseudo_op): Likewise.
153         * configure: Regenerate.
154
155 2010-12-03  Chung-Lin Tang  <cltang@codesourcery.com>
156
157         * src/arm/sysv.S (ffi_closure_SYSV): Add UNWIND to .pad directive.
158         (ffi_closure_VFP): Same.
159         (ffi_call_VFP): Move down to before ffi_closure_VFP. Add '.fpu vfp'
160         directive.
161
162 2010-12-01  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
163
164         * testsuite/libffi.call/ffitest.h [__sgi] (PRId64, PRIu64): Define.
165         (PRIuPTR): Define.
166
167 2010-11-29  Richard Henderson  <rth@redhat.com>
168             Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
169
170         * src/x86/sysv.S (FDE_ENCODING, FDE_ENCODE): Define.
171         (.eh_frame): Use FDE_ENCODING.
172         (.LASFDE1, .LASFDE2, LASFDE3): Simplify with FDE_ENCODE.
173
174 2010-11-22  Jacek Caban <jacek@codeweavers.com>
175
176         * configure.ac: Check for symbol underscores on mingw-w64.
177         * configure: Rebuilt.
178         * src/x86/win64.S: Correctly access extern symbols in respect to
179         underscores.
180
181 2010-11-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
182
183         * testsuite/lib/libffi-dg.exp: Rename ...
184         * testsuite/lib/libffi.exp: ... to this.
185         * libffi/testsuite/libffi.call/call.exp: Don't load libffi-dg.exp.
186         * libffi/testsuite/libffi.special/special.exp: Likewise.
187
188 2010-10-28  Chung-Lin Tang  <cltang@codesourcery.com>
189
190         * src/arm/ffi.c (ffi_prep_args): Add VFP register argument handling
191         code, new parameter, and return value. Update comments.
192         (ffi_prep_cif_machdep): Add case for VFP struct return values. Add
193         call to layout_vfp_args().
194         (ffi_call_SYSV): Update declaration.
195         (ffi_call_VFP): New declaration.
196         (ffi_call): Add VFP struct return conditions. Call ffi_call_VFP()
197         when ABI is FFI_VFP.
198         (ffi_closure_VFP): New declaration.
199         (ffi_closure_SYSV_inner): Add new vfp_args parameter, update call to
200         ffi_prep_incoming_args_SYSV().
201         (ffi_prep_incoming_args_SYSV): Update parameters. Add VFP argument
202         case handling.
203         (ffi_prep_closure_loc): Pass ffi_closure_VFP to trampoline
204         construction under VFP hard-float.
205         (rec_vfp_type_p): New function.
206         (vfp_type_p): Same.
207         (place_vfp_arg): Same.
208         (layout_vfp_args): Same.
209         * src/arm/ffitarget.h (ffi_abi): Add FFI_VFP. Define FFI_DEFAULT_ABI
210         based on __ARM_PCS_VFP.
211         (FFI_EXTRA_CIF_FIELDS): Define for adding VFP hard-float specific
212         fields.
213         (FFI_TYPE_STRUCT_VFP_FLOAT): Define internally used type code.
214         (FFI_TYPE_STRUCT_VFP_DOUBLE): Same.
215         * src/arm/sysv.S (ffi_call_SYSV): Change call of ffi_prep_args() to
216         direct call. Move function pointer load upwards.
217         (ffi_call_VFP): New function.
218         (ffi_closure_VFP): Same.
219
220         * testsuite/lib/libffi-dg.exp (check-flags): New function.
221         (dg-skip-if): New function.
222         * testsuite/libffi.call/cls_double_va.c: Skip if target is arm*-*-*
223         and compiler options include -mfloat-abi=hard.
224         * testsuite/libffi.call/cls_longdouble_va.c: Same.
225
226 2010-10-01  Jakub Jelinek  <jakub@redhat.com>
227
228         PR libffi/45677
229         * src/x86/ffi64.c (ffi_prep_cif_machdep): Ensure cif->bytes is
230         a multiple of 8.
231         * testsuite/libffi.call/many2.c: New test.
232
233 2010-08-20  Mark Wielaard  <mjw@redhat.com>
234
235         * src/closures.c (open_temp_exec_file_mnt): Check if getmntent_r
236         returns NULL.
237
238 2010-08-09  Andreas Tobler  <andreast@fgznet.ch>
239
240         * configure.ac: Add target powerpc64-*-freebsd*.
241         * configure: Regenerate.
242         * testsuite/libffi.call/cls_align_longdouble_split.c: Pass
243         -mlong-double-128 only to linux targets.
244         * testsuite/libffi.call/cls_align_longdouble_split2.c: Likewise.
245         * testsuite/libffi.call/cls_longdouble.c: Likewise.
246         * testsuite/libffi.call/huge_struct.c: Likewise.
247
248 2010-08-05  Dan Witte  <dwitte@mozilla.com>
249
250         * Makefile.am: Pass FFI_DEBUG define to msvcc.sh for linking to the
251         debug CRT when --enable-debug is given.
252         * configure.ac: Define it.
253         * msvcc.sh: Translate -g and -DFFI_DEBUG appropriately.
254
255 2010-08-04  Dan Witte  <dwitte@mozilla.com>
256
257         * src/x86/ffitarget.h: Add X86_ANY define for all x86/x86_64
258         platforms.
259         * src/x86/ffi.c: Remove redundant ifdef checks.
260         * src/prep_cif.c: Push stack space computation into src/x86/ffi.c
261         for X86_ANY so return value space doesn't get added twice.
262
263 2010-08-03  Neil Rashbrooke <neil@parkwaycc.co.uk>
264
265         * msvcc.sh: Don't pass -safeseh to ml64 because behavior is buggy.
266
267 2010-07-22  Dan Witte  <dwitte@mozilla.com>
268
269         * src/*/ffitarget.h: Make FFI_LAST_ABI one past the last valid ABI.
270         * src/prep_cif.c: Fix ABI assertion.
271         * src/cris/ffi.c: Ditto.
272
273 2010-07-10  Evan Phoenix  <evan@fallingsnow.net>
274
275         * src/closures.c (selinux_enabled_check): Fix strncmp usage bug.
276
277 2010-07-07  Dan HorĂ¡k <dan@danny.cz>
278
279         * include/ffi.h.in: Protect #define with #ifndef.
280         * src/powerpc/ffitarget.h: Ditto.
281         * src/s390/ffitarget.h: Ditto.
282         * src/sparc/ffitarget.h: Ditto.
283
284 2010-07-07   Neil Roberts <neil@linux.intel.com>
285
286         * src/x86/sysv.S (ffi_call_SYSV): Align the stack pointer to
287         16-bytes.
288
289 2010-07-02  Jakub Jelinek  <jakub@redhat.com>
290
291         * Makefile.am (AM_MAKEFLAGS): Pass also mandir to submakes.
292         * Makefile.in: Regenerated.
293
294 2010-05-19  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
295
296         * configure.ac (libffi_cv_as_x86_pcrel): Check for illegal in as
297         output, too.
298         (libffi_cv_as_ascii_pseudo_op): Check for .ascii.
299         (libffi_cv_as_string_pseudo_op): Check for .string.
300         * configure: Regenerate.
301         * fficonfig.h.in: Regenerate.
302         * src/x86/sysv.S (.eh_frame): Use .ascii, .string or error.
303
304 2010-05-11  Dan Witte  <dwitte@mozilla.com>
305
306         * doc/libffi.tex: Document previous change.
307
308 2010-05-11  Makoto Kato <m_kato@ga2.so-net.ne.jp>
309
310         * src/x86/ffi.c (ffi_call): Don't copy structs passed by value.
311
312 2010-05-05  Michael Kohler <michaelkohler@live.com>
313
314         * src/dlmalloc.c (dlfree): Fix spelling.
315         * src/ia64/ffi.c (ffi_prep_cif_machdep): Ditto.
316         * configure.ac: Ditto.
317         * configure: Rebuilt.
318
319 2010-04-13  Dan Witte  <dwitte@mozilla.com>
320
321         * msvcc.sh: Build with -W3 instead of -Wall.
322         * src/powerpc/ffi_darwin.c: Remove build warnings.
323         * src/x86/ffi.c: Ditto.
324         * src/x86/ffitarget.h: Ditto.
325
326 2010-04-12  Dan Witte  <dwitte@mozilla.com>
327             Walter Meinl <wuno@lsvw.de>
328
329         * configure.ac: Add OS/2 support.
330         * configure: Rebuilt.
331         * src/closures.c: Ditto.
332         * src/dlmalloc.c: Ditto.
333         * src/x86/win32.S: Ditto.
334
335 2010-04-07  Jakub Jelinek  <jakub@redhat.com>
336
337         * testsuite/libffi.call/err_bad_abi.c: Remove unused args variable.
338
339 2010-04-02  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
340
341         * Makefile.in: Regenerate.
342         * aclocal.m4: Regenerate.
343         * include/Makefile.in: Regenerate.
344         * man/Makefile.in: Regenerate.
345         * testsuite/Makefile.in: Regenerate.
346
347 2010-03-30  Dan Witte  <dwitte@mozilla.com>
348
349         * msvcc.sh: Disable build warnings.
350         * README (tested): Clarify windows build procedure.
351
352 2010-03-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
353
354         * configure.ac (libffi_cv_as_x86_64_unwind_section_type): New test.
355         * configure: Regenerate.
356         * fficonfig.h.in: Regenerate.
357         * libffi/src/x86/unix64.S (.eh_frame)
358         [HAVE_AS_X86_64_UNWIND_SECTION_TYPE]: Use @unwind section type.
359
360 2010-03-14  Matthias Klose  <doko@ubuntu.com>
361
362         * src/x86/ffi64.c: Fix typo in comment.
363         * src/x86/ffi.c: Use /* ... */ comment style.
364
365 2010-02-24  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
366
367         * doc/libffi.texi (The Closure API): Fix typo.
368         * doc/libffi.info: Remove.
369
370 2010-02-15  Matthias Klose  <doko@ubuntu.com>
371
372         * src/arm/sysv.S (__ARM_ARCH__): Define for processor
373         __ARM_ARCH_7EM__.
374
375 2010-01-15  Anthony Green  <green@redhat.com>
376
377         * README: Add notes on building with Microsoft Visual C++.
378
379 2010-01-15  Daniel Witte  <dwitte@mozilla.com>
380
381         * msvcc.sh: New file.
382
383         * src/x86/win32.S: Port assembly routines to MSVC and #ifdef.
384         * src/x86/ffi.c: Tweak function declaration and remove excess
385         parens.
386         * include/ffi.h.in: Add __declspec(align(8)) to typedef struct
387         ffi_closure.
388
389         * src/x86/ffi.c: Merge ffi_call_SYSV and ffi_call_STDCALL into new
390         function ffi_call_win32 on X86_WIN32.
391         * src/x86/win32.S (ffi_call_SYSV): Rename to ffi_call_win32.
392         (ffi_call_STDCALL): Remove.
393
394         * src/prep_cif.c (ffi_prep_cif): Move stack space allocation code
395         to ffi_prep_cif_machdep for x86.
396         * src/x86/ffi.c (ffi_prep_cif_machdep): To here.
397
398 2010-01-15  Oliver Kiddle  <okiddle@yahoo.co.uk>
399
400         * src/x86/ffitarget.h (ffi_abi): Check for __i386 and __amd64 for
401         Sun Studio compiler compatibility.
402
403 2010-01-12  Conrad Irwin <conrad.irwin@gmail.com>
404
405         * doc/libffi.texi: Add closure example.
406
407 2010-01-07  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
408
409         PR libffi/40701
410         * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRIdLL,
411         PRIuLL, PRId64, PRIu64, PRIuPTR): Define.
412         * testsuite/libffi.call/cls_align_sint64.c: Add -Wno-format on
413         alpha*-dec-osf*.
414         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
415         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
416         * testsuite/libffi.call/return_ll1.c: Likewise.
417         * testsuite/libffi.call/stret_medium2.c: Likewise.
418         * testsuite/libffi.special/ffitestcxx.h (allocate_mmap): Cast
419         MAP_FAILED to char *.
420
421 2010-01-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
422
423         * src/mips/n32.S: Use .abicalls and .eh_frame with __GNUC__.
424
425 2009-12-31  Anthony Green  <green@redhat.com>
426
427         * README: Update for libffi 3.0.9.
428
429 2009-12-27  Matthias Klose  <doko@ubuntu.com>
430
431         * configure.ac (HAVE_LONG_DOUBLE): Define for mips when
432         appropriate.
433         * configure: Rebuilt.
434
435 2009-12-26  Anthony Green  <green@redhat.com>
436
437         * testsuite/libffi.call/cls_longdouble_va.c: Mark as xfail for
438         avr32*-*-*.
439         * testsuite/libffi.call/cls_double_va.c: Ditto.
440
441 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
442
443         * testsuite/libffi.call/ffitest.h: Conditionally include stdint.h
444         and inttypes.h.
445         * testsuite/libffi.special/unwindtest.cc: Ditto.
446
447 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
448
449         * configure.ac: Add amd64-*-openbsd*.
450         * configure: Rebuilt.
451         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Link
452         openbsd programs with -lpthread.
453
454 2009-12-26  Anthony Green  <green@redhat.com>
455
456         * testsuite/libffi.call/cls_double_va.c,
457         testsuite/libffi.call/cls_longdouble.c,
458         testsuite/libffi.call/cls_longdouble_va.c,
459         testsuite/libffi.call/cls_pointer.c,
460         testsuite/libffi.call/cls_pointer_stack.c: Remove xfail for
461         mips*-*-* and arm*-*-*.
462         * testsuite/libffi.call/cls_align_longdouble_split.c,
463         testsuite/libffi.call/cls_align_longdouble_split2.c,
464         testsuite/libffi.call/stret_medium2.c,
465         testsuite/libffi.call/stret_medium.c,
466         testsuite/libffi.call/stret_large.c,
467         testsuite/libffi.call/stret_large2.c: Remove xfail for arm*-*-*.
468
469 2009-12-31  Kay Tietz  <ktietz70@googlemail.com>
470
471         * testsuite/libffi.call/ffitest.h,
472         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRuLL): Fix
473         definitions.
474
475 2009-12-31  Carlo Bramini  <carlo.bramix@libero.it>
476
477         * configure.ac (AM_LTLDFLAGS): Define for windows hosts.
478         * Makefile.am (libffi_la_LDFLAGS): Add AM_LTLDFLAGS.
479         * configure: Rebuilt.
480         * Makefile.in: Rebuilt.
481
482 2009-12-31  Anthony Green  <green@redhat.com>
483             Blake Chaffin.
484
485         * testsuite/libffi.call/huge_struct.c: New test case from Blake
486         Chaffin @ Apple.
487
488 2009-12-28  David Edelsohn  <edelsohn@gnu.org>
489
490         * src/powerpc/ffi_darwin.c (ffi_prep_args): Copy abi and nargs to
491         local variables.
492         (aix_adjust_aggregate_sizes): New function.
493         (ffi_prep_cif_machdep): Call it.
494
495 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
496
497         * configure.ac: Define FFI_MMAP_EXEC_WRIT for the given targets.
498         * configure: Regenerate.
499         * fficonfig.h.in: Likewise.
500         * src/closures.c: Remove the FFI_MMAP_EXEC_WRIT definition for
501         Solaris/x86.
502
503 2009-12-26  Andreas Schwab  <schwab@linux-m68k.org>
504
505         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Advance intarg_count
506         when a float arguments is passed in memory.
507         (ffi_closure_helper_SYSV): Mark general registers as used up when
508         a 64bit or soft-float long double argument is passed in memory.
509
510 2009-12-25  Matthias Klose  <doko@ubuntu.com>
511
512         * man/ffi_call.3: Fix #include in examples.
513         * doc/libffi.texi: Add dircategory.
514
515 2009-12-25  Frank Everdij <f.p.x.everdij@tudelft.nl>
516
517         * include/ffi.h.in: Placed '__GNUC__' ifdef around
518         '__attribute__((aligned(8)))' in ffi_closure, fixes compile for
519         IRIX MIPSPro c99.
520         * include/ffi_common.h: Added '__sgi' define to non
521         '__attribute__((__mode__()))' integer typedefs.
522         * src/mips/ffi.c (ffi_call, ffi_closure_mips_inner_O32,
523         ffi_closure_mips_inner_N32): Added 'defined(_MIPSEB)' to BE check.
524         (ffi_closure_mips_inner_O32, ffi_closure_mips_inner_N32): Added
525         FFI_LONGDOUBLE support and alignment(N32 only).
526         * src/mips/ffitarget.h: Corrected '#include <sgidefs.h>' for IRIX and
527         fixed non '__attribute__((__mode__()))' integer typedefs.
528         * src/mips/n32.S: Put '#ifdef linux' around '.abicalls' and '.eh_frame'
529         since they are Linux/GNU Assembler specific.
530
531 2009-12-25  Bradley Smith  <brad@brad-smith.co.uk>
532
533         * configure.ac, Makefile.am, src/avr32/ffi.c,
534         src/avr32/ffitarget.h,
535         src/avr32/sysv.S: Add AVR32 port.
536         * configure, Makefile.in: Rebuilt.
537
538 2009-12-21  Andreas Tobler  <a.tobler@schweiz.org>
539
540         * configure.ac: Make i?86 build on FreeBSD and OpenBSD.
541         * configure: Regenerate.
542
543 2009-12-15  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
544
545         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on PA HP-UX.
546
547 2009-12-13  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
548
549         * src/pa/ffi.c (ffi_closure_inner_pa32): Handle FFI_TYPE_LONGDOUBLE
550         type on HP-UX.
551
552 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
553
554         * src/sparc/ffi.c (ffi_closure_sparc_inner_v9): Properly align 'long
555         double' arguments.
556
557 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
558
559         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on Solaris < 10.
560
561 2009-12-10  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
562
563         PR libffi/40700
564         * src/closures.c [X86_64 && __sun__ && __svr4__]
565         (FFI_MMAP_EXEC_WRIT): Define.
566
567 2009-12-08  David Daney  <ddaney@caviumnetworks.com>
568
569         * testsuite/libffi.call/stret_medium.c: Remove xfail for mips*-*-*
570         * testsuite/libffi.call/cls_align_longdouble_split2.c: Same.
571         * testsuite/libffi.call/stret_large.c: Same.
572         * testsuite/libffi.call/cls_align_longdouble_split.c: Same.
573         * testsuite/libffi.call/stret_large2.c: Same.
574         * testsuite/libffi.call/stret_medium2.c: Same.
575
576 2009-12-07  David Edelsohn  <edelsohn@gnu.org>
577
578         * src/powerpc/aix_closure.S (libffi_closure_ASM): Fix tablejump
579         typo.
580
581 2009-12-05  David Edelsohn  <edelsohn@gnu.org>
582
583         * src/powerpc/aix.S: Update AIX32 code to be consistent with AIX64
584         code.
585         * src/powerpc/aix_closure.S: Same.
586
587 2009-12-05  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
588
589         * Makefile.in: Regenerate.
590         * configure: Regenerate.
591         * include/Makefile.in: Regenerate.
592         * man/Makefile.in: Regenerate.
593         * testsuite/Makefile.in: Regenerate.
594
595 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
596
597         * src/powerpc/aix_closure.S: Reorganize 64-bit code to match
598         linux64_closure.S.
599
600 2009-12-04  Uros Bizjak  <ubizjak@gmail.com>
601
602         PR libffi/41908
603         * src/x86/ffi64.c (classify_argument): Update from
604         gcc/config/i386/i386.c.
605         (ffi_closure_unix64_inner): Do not use the address of two consecutive
606         SSE registers directly.
607         * testsuite/libffi.call/cls_dbls_struct.c (main): Remove xfail
608         for x86_64 linux targets.
609
610 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
611
612         * src/powerpc/ffi_darwin.c (ffi_closure_helper_DARWIN): Increment
613         pfr for long double split between fpr13 and stack.
614
615 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
616
617         * src/powerpc/ffi_darwin.c (ffi_prep_args): Increment next_arg and
618         fparg_count twice for long double.
619
620 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
621
622         PR libffi/42243
623         * src/powerpc/ffi_darwin.c (ffi_prep_args): Remove extra parentheses.
624
625 2009-12-03  Uros Bizjak  <ubizjak@gmail.com>
626
627         * testsuite/libffi.call/cls_longdouble_va.c (main): Fix format string.
628         Remove xfails for x86 linux targets.
629
630 2009-12-02  David Edelsohn  <edelsohn@gnu.org>
631
632         * src/powerpc/ffi_darwin.c (ffi_prep_args): Fix typo in INT64
633         case.
634
635 2009-12-01  David Edelsohn  <edelsohn@gnu.org>
636
637         * src/powerpc/aix.S (ffi_call_AIX): Convert to more standard
638         register usage.  Call ffi_prep_args directly.  Add long double
639         return value support.
640         * src/powerpc/ffi_darwin.c (ffi_prep_args): Double arg increment
641         applies to FFI_TYPE_DOUBLE.  Correct fpr_base increment typo.
642         Separate FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases.
643         (ffi_prep_cif_machdep): Only 16 byte stack alignment in 64 bit
644         mode.
645         (ffi_closure_helper_DARWIN): Remove nf and ng counters.  Move temp
646         into case.
647         * src/powerpc/aix_closure.S: Maintain 16 byte stack alignment.
648         Allocate result area between params and FPRs.
649
650 2009-11-30  David Edelsohn  <edelsohn@gnu.org>
651
652         PR target/35484
653         * src/powerpc/ffitarget.h (POWERPC64): Define for PPC64 Linux and
654         AIX64.
655         * src/powerpc/aix.S: Implement AIX64 version.
656         * src/powerpc/aix_closure.S: Implement AIX64 version.
657         (ffi_closure_ASM): Use extsb, lha and displament addresses.
658         * src/powerpc/ffi_darwin.c (ffi_prep_args): Implement AIX64
659         support.
660         (ffi_prep_cif_machdep): Same.
661         (ffi_call): Same.
662         (ffi_closure_helper_DARWIN): Same.
663
664 2009-11-02  Andreas Tobler  <a.tobler@schweiz.org>
665
666         PR libffi/41908
667         * testsuite/libffi.call/testclosure.c: New test.
668
669 2009-09-28  Kai Tietz  <kai.tietz@onevision.com>
670
671         * src/x86/win64.S (_ffi_call_win64 stack): Remove for gnu
672         assembly version use of ___chkstk.
673
674 2009-09-23  Matthias Klose  <doko@ubuntu.com>
675
676         PR libffi/40242, PR libffi/41443
677         * src/arm/sysv.S (__ARM_ARCH__): Define for processors
678         __ARM_ARCH_6T2__, __ARM_ARCH_6M__, __ARM_ARCH_7__,
679         __ARM_ARCH_7A__, __ARM_ARCH_7R__, __ARM_ARCH_7M__.
680         Change the conditionals to __SOFTFP__ || __ARM_EABI__
681         for -mfloat-abi=softfp to work.
682
683 2009-09-17  Loren J. Rittle  <ljrittle@acm.org>
684
685         PR testsuite/32843 (strikes again)
686         * src/x86/ffi.c (ffi_prep_cif_machdep): Add X86_FREEBSD to
687         enable proper extension on char and short.
688
689 2009-09-15  David Daney  <ddaney@caviumnetworks.com>
690
691         * src/java_raw_api.c (ffi_java_raw_to_rvalue): Remove special
692         handling for FFI_TYPE_POINTER.
693         * src/mips/ffitarget.h (FFI_TYPE_STRUCT_D_SOFT,
694         FFI_TYPE_STRUCT_F_SOFT, FFI_TYPE_STRUCT_DD_SOFT,
695         FFI_TYPE_STRUCT_FF_SOFT, FFI_TYPE_STRUCT_FD_SOFT,
696         FFI_TYPE_STRUCT_DF_SOFT, FFI_TYPE_STRUCT_SOFT): New defines.
697         (FFI_N32_SOFT_FLOAT, FFI_N64_SOFT_FLOAT): New ffi_abi enumerations.
698         (enum ffi_abi): Set FFI_DEFAULT_ABI for soft-float.
699         * src/mips/n32.S (ffi_call_N32): Add handling for soft-float
700         structure and pointer returns.
701         (ffi_closure_N32): Add handling for pointer returns.
702         * src/mips/ffi.c (ffi_prep_args, calc_n32_struct_flags,
703         calc_n32_return_struct_flags): Handle soft-float.
704         (ffi_prep_cif_machdep):  Handle soft-float, fix pointer handling.
705         (ffi_call_N32): Declare proper argument types.
706         (ffi_call, copy_struct_N32, ffi_closure_mips_inner_N32): Handle
707         soft-float.
708
709 2009-08-24  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
710
711         * configure.ac (AC_PREREQ): Bump to 2.64.
712
713 2009-08-22  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
714
715         * Makefile.am (install-html, install-pdf): Remove.
716         * Makefile.in: Regenerate.
717
718         * Makefile.in: Regenerate.
719         * aclocal.m4: Regenerate.
720         * configure: Regenerate.
721         * fficonfig.h.in: Regenerate.
722         * include/Makefile.in: Regenerate.
723         * man/Makefile.in: Regenerate.
724         * testsuite/Makefile.in: Regenerate.
725
726 2009-07-30  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
727
728         * configure.ac (_AC_ARG_VAR_PRECIOUS): Use m4_rename_force.
729
730 2009-07-24  Dave Korn  <dave.korn.cygwin@gmail.com>
731
732         PR libffi/40807
733         * src/x86/ffi.c (ffi_prep_cif_machdep): Also use sign/zero-extending
734         return types for X86_WIN32.
735         * src/x86/win32.S (_ffi_call_SYSV): Handle omitted return types.
736         (_ffi_call_STDCALL, _ffi_closure_SYSV, _ffi_closure_raw_SYSV,
737         _ffi_closure_STDCALL): Likewise.
738
739         * src/closures.c (is_selinux_enabled): Define to const 0 for Cygwin.
740         (dlmmap, dlmunmap): Also use these functions on Cygwin.
741
742 2009-07-11  Richard Sandiford  <rdsandiford@googlemail.com>
743
744         PR testsuite/40699
745         PR testsuite/40707
746         PR testsuite/40709
747         * testsuite/lib/libffi-dg.exp: Revert 2009-07-02, 2009-07-01 and
748         2009-06-30 commits.
749
750 2009-07-01  Richard Sandiford  <r.sandiford@uk.ibm.com>
751
752         * testsuite/lib/libffi-dg.exp (libffi-init): Set ld_library_path
753         to "" before adding paths.  (This reinstates an assignment that
754         was removed by my 2009-06-30 commit, but changes the initial
755         value from "." to "".)
756
757 2009-07-01  H.J. Lu  <hongjiu.lu@intel.com>
758
759         PR testsuite/40601
760         * testsuite/lib/libffi-dg.exp (libffi-init): Properly set
761         gccdir.  Adjust ld_library_path for gcc only if gccdir isn't
762         empty.
763
764 2009-06-30  Richard Sandiford  <r.sandiford@uk.ibm.com>
765
766         * testsuite/lib/libffi-dg.exp (libffi-init): Don't add "."
767         to ld_library_path.  Use add_path.  Add just find_libgcc_s
768         to ld_library_path, not every libgcc multilib directory.
769
770 2009-06-16  Wim Lewis  <wiml@hhhh.org>
771
772         * src/powerpc/ffi.c: Avoid clobbering cr3 and cr4, which are
773         supposed to be callee-saved.
774         * src/powerpc/sysv.S (small_struct_return_value): Fix overrun of
775         return buffer for odd-size structs.
776
777 2009-06-16  Andreas Tobler  <a.tobler@schweiz.org>
778
779         PR libffi/40444
780         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Add
781         allow_stack_execute for Darwin.
782
783 2009-06-16  Andrew Haley  <aph@redhat.com>
784
785         * configure.ac (TARGETDIR): Add missing blank lines.
786         * configure: Regenerate.
787
788 2009-06-16  Andrew Haley  <aph@redhat.com>
789
790         * testsuite/libffi.call/cls_align_sint64.c,
791         testsuite/libffi.call/cls_align_uint64.c,
792         testsuite/libffi.call/cls_longdouble_va.c,
793         testsuite/libffi.call/cls_ulonglong.c,
794         testsuite/libffi.call/return_ll1.c,
795         testsuite/libffi.call/stret_medium2.c: Fix printf format
796         specifiers.
797         * testsuite/libffi.call/ffitest.h,
798         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRIuLL): Define.
799
800 2009-06-15  Andrew Haley  <aph@redhat.com>
801
802         * testsuite/libffi.call/err_bad_typedef.c: xfail everywhere.
803         * testsuite/libffi.call/err_bad_abi.c: Likewise.
804
805 2009-06-12  Andrew Haley  <aph@redhat.com>
806
807         * Makefile.am: Remove info_TEXINFOS.
808
809 2009-06-12  Andrew Haley  <aph@redhat.com>
810
811         * ChangeLog.libffi: testsuite/libffi.call/cls_align_sint64.c,
812         testsuite/libffi.call/cls_align_uint64.c,
813         testsuite/libffi.call/cls_ulonglong.c,
814         testsuite/libffi.call/return_ll1.c,
815         testsuite/libffi.call/stret_medium2.c: Fix printf format
816         specifiers.
817         testsuite/libffi.special/unwindtest.cc: include stdint.h.
818
819 2009-06-11  Timothy Wall  <twall@users.sf.net>
820
821         * Makefile.am,
822         configure.ac,
823         include/ffi.h.in,
824         include/ffi_common.h,
825         src/closures.c,
826         src/dlmalloc.c,
827         src/x86/ffi.c,
828         src/x86/ffitarget.h,
829         src/x86/win64.S (new),
830         README: Added win64 support (mingw or MSVC)
831         * Makefile.in,
832         include/Makefile.in,
833         man/Makefile.in,
834         testsuite/Makefile.in,
835         configure,
836         aclocal.m4: Regenerated
837         * ltcf-c.sh: properly escape cygwin/w32 path
838         * man/ffi_call.3: Clarify size requirements for return value.
839         * src/x86/ffi64.c: Fix filename in comment.
840         * src/x86/win32.S: Remove unused extern.
841
842         * testsuite/libffi.call/closure_fn0.c,
843         testsuite/libffi.call/closure_fn1.c,
844         testsuite/libffi.call/closure_fn2.c,
845         testsuite/libffi.call/closure_fn3.c,
846         testsuite/libffi.call/closure_fn4.c,
847         testsuite/libffi.call/closure_fn5.c,
848         testsuite/libffi.call/closure_fn6.c,
849         testsuite/libffi.call/closure_stdcall.c,
850         testsuite/libffi.call/cls_12byte.c,
851         testsuite/libffi.call/cls_16byte.c,
852         testsuite/libffi.call/cls_18byte.c,
853         testsuite/libffi.call/cls_19byte.c,
854         testsuite/libffi.call/cls_1_1byte.c,
855         testsuite/libffi.call/cls_20byte.c,
856         testsuite/libffi.call/cls_20byte1.c,
857         testsuite/libffi.call/cls_24byte.c,
858         testsuite/libffi.call/cls_2byte.c,
859         testsuite/libffi.call/cls_3_1byte.c,
860         testsuite/libffi.call/cls_3byte1.c,
861         testsuite/libffi.call/cls_3byte2.c,
862         testsuite/libffi.call/cls_4_1byte.c,
863         testsuite/libffi.call/cls_4byte.c,
864         testsuite/libffi.call/cls_5_1_byte.c,
865         testsuite/libffi.call/cls_5byte.c,
866         testsuite/libffi.call/cls_64byte.c,
867         testsuite/libffi.call/cls_6_1_byte.c,
868         testsuite/libffi.call/cls_6byte.c,
869         testsuite/libffi.call/cls_7_1_byte.c,
870         testsuite/libffi.call/cls_7byte.c,
871         testsuite/libffi.call/cls_8byte.c,
872         testsuite/libffi.call/cls_9byte1.c,
873         testsuite/libffi.call/cls_9byte2.c,
874         testsuite/libffi.call/cls_align_double.c,
875         testsuite/libffi.call/cls_align_float.c,
876         testsuite/libffi.call/cls_align_longdouble.c,
877         testsuite/libffi.call/cls_align_longdouble_split.c,
878         testsuite/libffi.call/cls_align_longdouble_split2.c,
879         testsuite/libffi.call/cls_align_pointer.c,
880         testsuite/libffi.call/cls_align_sint16.c,
881         testsuite/libffi.call/cls_align_sint32.c,
882         testsuite/libffi.call/cls_align_sint64.c,
883         testsuite/libffi.call/cls_align_uint16.c,
884         testsuite/libffi.call/cls_align_uint32.c,
885         testsuite/libffi.call/cls_align_uint64.c,
886         testsuite/libffi.call/cls_dbls_struct.c,
887         testsuite/libffi.call/cls_double.c,
888         testsuite/libffi.call/cls_double_va.c,
889         testsuite/libffi.call/cls_float.c,
890         testsuite/libffi.call/cls_longdouble.c,
891         testsuite/libffi.call/cls_longdouble_va.c,
892         testsuite/libffi.call/cls_multi_schar.c,
893         testsuite/libffi.call/cls_multi_sshort.c,
894         testsuite/libffi.call/cls_multi_sshortchar.c,
895         testsuite/libffi.call/cls_multi_uchar.c,
896         testsuite/libffi.call/cls_multi_ushort.c,
897         testsuite/libffi.call/cls_multi_ushortchar.c,
898         testsuite/libffi.call/cls_pointer.c,
899         testsuite/libffi.call/cls_pointer_stack.c,
900         testsuite/libffi.call/cls_schar.c,
901         testsuite/libffi.call/cls_sint.c,
902         testsuite/libffi.call/cls_sshort.c,
903         testsuite/libffi.call/cls_uchar.c,
904         testsuite/libffi.call/cls_uint.c,
905         testsuite/libffi.call/cls_ulonglong.c,
906         testsuite/libffi.call/cls_ushort.c,
907         testsuite/libffi.call/err_bad_abi.c,
908         testsuite/libffi.call/err_bad_typedef.c,
909         testsuite/libffi.call/float2.c,
910         testsuite/libffi.call/huge_struct.c,
911         testsuite/libffi.call/nested_struct.c,
912         testsuite/libffi.call/nested_struct1.c,
913         testsuite/libffi.call/nested_struct10.c,
914         testsuite/libffi.call/nested_struct2.c,
915         testsuite/libffi.call/nested_struct3.c,
916         testsuite/libffi.call/nested_struct4.c,
917         testsuite/libffi.call/nested_struct5.c,
918         testsuite/libffi.call/nested_struct6.c,
919         testsuite/libffi.call/nested_struct7.c,
920         testsuite/libffi.call/nested_struct8.c,
921         testsuite/libffi.call/nested_struct9.c,
922         testsuite/libffi.call/problem1.c,
923         testsuite/libffi.call/return_ldl.c,
924         testsuite/libffi.call/return_ll1.c,
925         testsuite/libffi.call/stret_large.c,
926         testsuite/libffi.call/stret_large2.c,
927         testsuite/libffi.call/stret_medium.c,
928         testsuite/libffi.call/stret_medium2.c,
929         testsuite/libffi.special/unwindtest.cc: use ffi_closure_alloc instead
930         of checking for MMAP.  Use intptr_t instead of long casts.
931
932 2009-06-11  Kaz Kojima  <kkojima@gcc.gnu.org>
933
934         * testsuite/libffi.call/cls_longdouble_va.c: Add xfail sh*-*-linux-*.
935         * testsuite/libffi.call/err_bad_abi.c: Add xfail sh*-*-*.
936         * testsuite/libffi.call/err_bad_typedef.c: Likewise.
937
938 2009-06-09  Andrew Haley  <aph@redhat.com>
939
940         * src/x86/freebsd.S: Add missing file.
941
942 2009-06-08  Andrew Haley  <aph@redhat.com>
943
944         Import from libffi 3.0.8:
945
946         * doc/libffi.texi: New file.
947         * doc/libffi.info: Likewise.
948         * doc/stamp-vti: Likewise.
949         * man/Makefile.am: New file.
950         * man/ffi_call.3: New file.
951
952         * Makefile.am (EXTRA_DIST): Add src/x86/darwin64.S,
953         src/dlmalloc.c.
954         (nodist_libffi_la_SOURCES): Add X86_FREEBSD.
955
956         * configure.ac: Bump version to 3.0.8.
957         parisc*-*-linux*: Add.
958         i386-*-freebsd* | i386-*-openbsd*: Add.
959         powerpc-*-beos*: Add.
960         AM_CONDITIONAL X86_FREEBSD: Add.
961         AC_CONFIG_FILES: Add man/Makefile.
962
963         * include/ffi.h.in (FFI_FN): Change void (*)() to void (*)(void).
964
965 2009-06-08  Andrew Haley  <aph@redhat.com>
966
967         * README: Import from libffi 3.0.8.
968
969 2009-06-08  Andrew Haley  <aph@redhat.com>
970
971         * testsuite/libffi.call/err_bad_abi.c: Add xfails.
972         * testsuite/libffi.call/cls_longdouble_va.c: Add xfails.
973         * testsuite/libffi.call/cls_dbls_struct.c: Add xfail x86_64-*-linux-*.
974         * testsuite/libffi.call/err_bad_typedef.c: Add xfails.
975
976         * testsuite/libffi.call/stret_medium2.c: Add __UNUSED__ to args.
977         * testsuite/libffi.call/stret_medium.c: Likewise.
978         * testsuite/libffi.call/stret_large2.c: Likewise.
979         * testsuite/libffi.call/stret_large.c:  Likewise.
980
981 2008-12-26  Timothy Wall  <twall@users.sf.net>
982
983         * testsuite/libffi.call/cls_longdouble.c,
984         testsuite/libffi.call/cls_longdouble_va.c,
985         testsuite/libffi.call/cls_align_longdouble.c,
986         testsuite/libffi.call/cls_align_longdouble_split.c,
987         testsuite/libffi.call/cls_align_longdouble_split2.c: mark expected
988         failures on x86_64 cygwin/mingw.
989
990 2008-12-22  Timothy Wall  <twall@users.sf.net>
991
992         * testsuite/libffi.call/closure_fn0.c,
993         testsuite/libffi.call/closure_fn1.c,
994         testsuite/libffi.call/closure_fn2.c,
995         testsuite/libffi.call/closure_fn3.c,
996         testsuite/libffi.call/closure_fn4.c,
997         testsuite/libffi.call/closure_fn5.c,
998         testsuite/libffi.call/closure_fn6.c,
999         testsuite/libffi.call/closure_loc_fn0.c,
1000         testsuite/libffi.call/closure_stdcall.c,
1001         testsuite/libffi.call/cls_align_pointer.c,
1002         testsuite/libffi.call/cls_pointer.c,
1003         testsuite/libffi.call/cls_pointer_stack.c: use portable cast from
1004         pointer to integer (intptr_t).
1005         * testsuite/libffi.call/cls_longdouble.c: disable for win64.
1006
1007 2008-07-24  Anthony Green  <green@redhat.com>
1008
1009         * testsuite/libffi.call/cls_dbls_struct.c,
1010         testsuite/libffi.call/cls_double_va.c,
1011         testsuite/libffi.call/cls_longdouble.c,
1012         testsuite/libffi.call/cls_longdouble_va.c,
1013         testsuite/libffi.call/cls_pointer.c,
1014         testsuite/libffi.call/cls_pointer_stack.c,
1015         testsuite/libffi.call/err_bad_abi.c: Clean up failures from
1016         compiler warnings.
1017
1018 2008-03-04  Anthony Green  <green@redhat.com>
1019             Blake Chaffin
1020             hos@tamanegi.org
1021
1022         * testsuite/libffi.call/cls_align_longdouble_split2.c
1023           testsuite/libffi.call/cls_align_longdouble_split.c
1024           testsuite/libffi.call/cls_dbls_struct.c
1025           testsuite/libffi.call/cls_double_va.c
1026           testsuite/libffi.call/cls_longdouble.c
1027           testsuite/libffi.call/cls_longdouble_va.c
1028           testsuite/libffi.call/cls_pointer.c
1029           testsuite/libffi.call/cls_pointer_stack.c
1030           testsuite/libffi.call/err_bad_abi.c
1031           testsuite/libffi.call/err_bad_typedef.c
1032           testsuite/libffi.call/stret_large2.c
1033           testsuite/libffi.call/stret_large.c
1034           testsuite/libffi.call/stret_medium2.c
1035           testsuite/libffi.call/stret_medium.c: New tests from Apple.
1036
1037 2009-06-05  Andrew Haley  <aph@redhat.com>
1038
1039         * src/x86/ffitarget.h, src/x86/ffi.c: Merge stdcall changes from
1040         libffi.
1041
1042 2009-06-04  Andrew Haley  <aph@redhat.com>
1043
1044         * src/x86/ffitarget.h, src/x86/win32.S, src/x86/ffi.c: Back out
1045         stdcall changes.
1046
1047 2008-02-26  Anthony Green  <green@redhat.com>
1048             Thomas Heller  <theller@ctypes.org>
1049
1050         * src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C
1051         comment.
1052
1053 2008-02-03  Timothy Wall  <twall@users.sf.net>
1054
1055         * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return
1056           offset based on code pointer, not data pointer.
1057
1058 2008-01-31  Timothy Wall <twall@users.sf.net>
1059
1060         * testsuite/libffi.call/closure_stdcall.c: Add test for stdcall
1061         closures.
1062         * src/x86/ffitarget.h: Increase size of trampoline for stdcall
1063         closures.
1064         * src/x86/win32.S: Add assembly for stdcall closure.
1065         * src/x86/ffi.c: Initialize stdcall closure trampoline.
1066
1067 2009-06-04  Andrew Haley  <aph@redhat.com>
1068
1069         * include/ffi.h.in: Change void (*)() to void (*)(void).
1070         * src/x86/ffi.c: Likewise.
1071
1072 2009-06-04  Andrew Haley  <aph@redhat.com>
1073
1074         * src/powerpc/ppc_closure.S: Insert licence header.
1075         * src/powerpc/linux64_closure.S: Likewise.
1076         * src/m68k/sysv.S: Likewise.
1077
1078         * src/sh64/ffi.c: Change void (*)() to void (*)(void).
1079         * src/powerpc/ffi.c: Likewise.
1080         * src/powerpc/ffi_darwin.c: Likewise.
1081         * src/m32r/ffi.c: Likewise.
1082         * src/sh64/ffi.c: Likewise.
1083         * src/x86/ffi64.c: Likewise.
1084         * src/alpha/ffi.c: Likewise.
1085         * src/alpha/osf.S: Likewise.
1086         * src/frv/ffi.c: Likewise.
1087         * src/s390/ffi.c: Likewise.
1088         * src/pa/ffi.c: Likewise.
1089         * src/pa/hpux32.S: Likewise.
1090         * src/ia64/unix.S: Likewise.
1091         * src/ia64/ffi.c: Likewise.
1092         * src/sparc/ffi.c: Likewise.
1093         * src/mips/ffi.c: Likewise.
1094         * src/sh/ffi.c: Likewise.
1095
1096 2008-02-15  David Daney  <ddaney@avtrex.com>
1097
1098         * src/mips/ffi.c (USE__BUILTIN___CLEAR_CACHE):
1099         Define (conditionally), and use it to include cachectl.h.
1100         (ffi_prep_closure_loc): Fix cache flushing.
1101         * src/mips/ffitarget.h (_ABIN32, _ABI64, _ABIO32): Define.
1102
1103 2009-06-04  Andrew Haley  <aph@redhat.com>
1104
1105         include/ffi.h.in,
1106         src/arm/ffitarget.h,
1107         src/arm/ffi.c,
1108         src/arm/sysv.S,
1109         src/powerpc/ffitarget.h,
1110         src/closures.c,
1111         src/sh64/ffitarget.h,
1112         src/sh64/ffi.c,
1113         src/sh64/sysv.S,
1114         src/types.c,
1115         src/x86/ffi64.c,
1116         src/x86/ffitarget.h,
1117         src/x86/win32.S,
1118         src/x86/darwin.S,
1119         src/x86/ffi.c,
1120         src/x86/sysv.S,
1121         src/x86/unix64.S,
1122         src/alpha/ffitarget.h,
1123         src/alpha/ffi.c,
1124         src/alpha/osf.S,
1125         src/m68k/ffitarget.h,
1126         src/frv/ffitarget.h,
1127         src/frv/ffi.c,
1128         src/s390/ffitarget.h,
1129         src/s390/sysv.S,
1130         src/cris/ffitarget.h,
1131         src/pa/linux.S,
1132         src/pa/ffitarget.h,
1133         src/pa/ffi.c,
1134         src/raw_api.c,
1135         src/ia64/ffitarget.h,
1136         src/ia64/unix.S,
1137         src/ia64/ffi.c,
1138         src/ia64/ia64_flags.h,
1139         src/java_raw_api.c,
1140         src/debug.c,
1141         src/sparc/v9.S,
1142         src/sparc/ffitarget.h,
1143         src/sparc/ffi.c,
1144         src/sparc/v8.S,
1145         src/mips/ffitarget.h,
1146         src/mips/n32.S,
1147         src/mips/o32.S,
1148         src/mips/ffi.c,
1149         src/prep_cif.c,
1150         src/sh/ffitarget.h,
1151         src/sh/ffi.c,
1152         src/sh/sysv.S: Update license text.
1153
1154 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
1155
1156         * src/x86/win32.S (_ffi_closure_STDCALL):  New function.
1157         (.eh_frame):  Add FDE for it.
1158
1159 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
1160
1161         * configure.ac:  Also check if assembler supports pc-relative
1162         relocs on X86_WIN32 targets.
1163         * configure:  Regenerate.
1164         * src/x86/win32.S (ffi_prep_args):  Declare extern, not global.
1165         (_ffi_call_SYSV):  Add missing function type symbol .def and
1166         add EH markup labels.
1167         (_ffi_call_STDCALL):  Likewise.
1168         (_ffi_closure_SYSV):  Likewise.
1169         (_ffi_closure_raw_SYSV):  Likewise.
1170         (.eh_frame):  Add hand-crafted EH data.
1171
1172 2009-04-09  Jakub Jelinek  <jakub@redhat.com>
1173
1174         * testsuite/lib/libffi-dg.exp: Change copyright header to refer to
1175         version 3 of the GNU General Public License and to point readers
1176         at the COPYING3 file and the FSF's license web page.
1177         * testsuite/libffi.call/call.exp: Likewise.
1178         * testsuite/libffi.special/special.exp: Likewise.
1179
1180 2009-03-01  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1181
1182         * configure: Regenerate.
1183
1184 2008-12-18  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1185
1186         PR libffi/26048
1187         * configure.ac (HAVE_AS_X86_PCREL): New test.
1188         * configure: Regenerate.
1189         * fficonfig.h.in: Regenerate.
1190         * src/x86/sysv.S [!FFI_NO_RAW_API]: Precalculate
1191         RAW_CLOSURE_CIF_OFFSET, RAW_CLOSURE_FUN_OFFSET,
1192         RAW_CLOSURE_USER_DATA_OFFSET for the Solaris 10/x86 assembler.
1193         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
1194         * src/x86/unix64.S (.Lstore_table): Move to .text section.
1195         (.Lload_table): Likewise.
1196         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
1197
1198 2008-12-18  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1199
1200         * configure: Regenerate.
1201
1202 2008-11-21  Eric Botcazou  <ebotcazou@adacore.com>
1203
1204         * src/sparc/ffi.c (ffi_prep_cif_machdep): Add support for
1205         signed/unsigned int8/16 return values.
1206         * src/sparc/v8.S (ffi_call_v8): Likewise.
1207         (ffi_closure_v8): Likewise.
1208
1209 2008-09-26  Peter O'Gorman  <pogma@thewrittenword.com>
1210             Steve Ellcey  <sje@cup.hp.com>
1211
1212         * configure: Regenerate for new libtool.
1213         * Makefile.in: Ditto.
1214         * include/Makefile.in: Ditto.
1215         * aclocal.m4: Ditto.
1216
1217 2008-08-25  Andreas Tobler  <a.tobler@schweiz.org>
1218
1219         * src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and
1220         FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum.
1221         Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT.
1222         Adjust copyright notice.
1223         * src/powerpc/ffi.c: Add two new flags to indicate if we have one
1224         register or two register to use for FFI_SYSV structs.
1225         (ffi_prep_cif_machdep): Pass the right register flag introduced above.
1226         (ffi_closure_helper_SYSV): Fix the return type for
1227         FFI_SYSV_TYPE_SMALL_STRUCT. Comment.
1228         Adjust copyright notice.
1229
1230 2008-07-16  Kaz Kojima  <kkojima@gcc.gnu.org>
1231
1232         * src/sh/ffi.c (ffi_prep_closure_loc): Turn INSN into an unsigned
1233         int.
1234
1235 2008-06-17  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1236
1237         * configure: Regenerate.
1238         * include/Makefile.in: Regenerate.
1239         * testsuite/Makefile.in: Regenerate.
1240
1241 2008-06-07  Joseph Myers  <joseph@codesourcery.com>
1242
1243         * configure.ac (parisc*-*-linux*, powerpc-*-sysv*,
1244         powerpc-*-beos*): Remove.
1245         * configure: Regenerate.
1246
1247 2008-05-09  Julian Brown  <julian@codesourcery.com>
1248
1249         * Makefile.am (LTLDFLAGS): New.
1250         (libffi_la_LDFLAGS): Use above.
1251         * Makefile.in: Regenerate.
1252
1253 2008-04-18  Paolo Bonzini  <bonzini@gnu.org>
1254
1255         PR bootstrap/35457
1256         * aclocal.m4: Regenerate.
1257         * configure: Regenerate.
1258
1259 2008-03-26  Kaz Kojima  <kkojima@gcc.gnu.org>
1260
1261         * src/sh/sysv.S: Add .note.GNU-stack on Linux.
1262         * src/sh64/sysv.S: Likewise.
1263
1264 2008-03-26  Daniel Jacobowitz  <dan@debian.org>
1265
1266         * src/arm/sysv.S: Fix ARM comment marker.
1267
1268 2008-03-26  Jakub Jelinek  <jakub@redhat.com>
1269
1270         * src/alpha/osf.S: Add .note.GNU-stack on Linux.
1271         * src/s390/sysv.S: Likewise.
1272         * src/powerpc/ppc_closure.S: Likewise.
1273         * src/powerpc/sysv.S: Likewise.
1274         * src/x86/unix64.S: Likewise.
1275         * src/x86/sysv.S: Likewise.
1276         * src/sparc/v8.S: Likewise.
1277         * src/sparc/v9.S: Likewise.
1278         * src/m68k/sysv.S: Likewise.
1279         * src/arm/sysv.S: Likewise.
1280
1281 2008-03-16  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1282
1283         * aclocal.m4: Regenerate.
1284         * configure: Likewise.
1285         * Makefile.in: Likewise.
1286         * include/Makefile.in: Likewise.
1287         * testsuite/Makefile.in: Likewise.
1288
1289 2008-02-12  Bjoern Koenig  <bkoenig@alpha-tierchen.de>
1290             Andreas Tobler  <a.tobler@schweiz.org>
1291
1292         * configure.ac: Add amd64-*-freebsd* target.
1293         * configure: Regenerate.
1294
1295 2008-01-30  H.J. Lu  <hongjiu.lu@intel.com>
1296
1297         PR libffi/34612
1298         * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when
1299         returning struct.
1300
1301         * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer"
1302         tests.
1303
1304 2008-01-24  David Edelsohn  <edelsohn@gnu.org>
1305
1306         * configure: Regenerate.
1307
1308 2008-01-06  Andreas Tobler  <a.tobler@schweiz.org>
1309
1310         * src/x86/ffi.c (ffi_prep_cif_machdep): Fix thinko.
1311
1312 2008-01-05  Andreas Tobler  <a.tobler@schweiz.org>
1313
1314         PR testsuite/32843
1315         * src/x86/ffi.c (ffi_prep_cif_machdep): Add code for
1316         signed/unsigned int8/16 for X86_DARWIN.
1317         Updated copyright info.
1318         Handle one and two byte structs with special cif->flags.
1319         * src/x86/ffitarget.h: Add special types for one and two byte structs.
1320         Updated copyright info.
1321         * src/x86/darwin.S (ffi_call_SYSV): Rewrite to use a jump table like
1322         sysv.S
1323         Remove code to pop args from the stack after call.
1324         Special-case signed/unsigned for int8/16, one and two byte structs.
1325         (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1326         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1327         FFI_TYPE_SINT32.
1328         Updated copyright info.
1329
1330 2007-12-08  David Daney  <ddaney@avtrex.com>
1331
1332         * src/mips/n32.S (ffi_call_N32):  Replace dadd with ADDU, dsub with
1333         SUBU, add with ADDU and use smaller code sequences.
1334
1335 2007-12-07  David Daney  <ddaney@avtrex.com>
1336
1337         * src/mips/ffi.c (ffi_prep_cif_machdep): Handle long double return
1338         type.
1339
1340 2007-12-06  David Daney  <ddaney@avtrex.com>
1341
1342         * include/ffi.h.in (FFI_SIZEOF_JAVA_RAW): Define if not already
1343         defined.
1344         (ffi_java_raw): New typedef.
1345         (ffi_java_raw_call, ffi_java_ptrarray_to_raw,
1346         ffi_java_raw_to_ptrarray): Change parameter types from ffi_raw to
1347         ffi_java_raw.
1348         (ffi_java_raw_closure) : Same.
1349         (ffi_prep_java_raw_closure, ffi_prep_java_raw_closure_loc): Change
1350         parameter types.
1351         * src/java_raw_api.c (ffi_java_raw_size):  Replace FFI_SIZEOF_ARG with
1352         FFI_SIZEOF_JAVA_RAW.
1353         (ffi_java_raw_to_ptrarray): Change type of raw to ffi_java_raw.
1354         Replace FFI_SIZEOF_ARG with FFI_SIZEOF_JAVA_RAW. Use
1355         sizeof(ffi_java_raw) for alignment calculations.
1356         (ffi_java_ptrarray_to_raw): Same.
1357         (ffi_java_rvalue_to_raw): Add special handling for FFI_TYPE_POINTER
1358         if FFI_SIZEOF_JAVA_RAW == 4.
1359         (ffi_java_raw_to_rvalue): Same.
1360         (ffi_java_raw_call): Change type of raw to ffi_java_raw.
1361         (ffi_java_translate_args): Same.
1362         (ffi_prep_java_raw_closure_loc, ffi_prep_java_raw_closure): Change
1363         parameter types.
1364         * src/mips/ffitarget.h (FFI_SIZEOF_JAVA_RAW): Define for N32 ABI.
1365
1366 2007-12-06  David Daney  <ddaney@avtrex.com>
1367
1368         * src/mips/n32.S (ffi_closure_N32): Use 64-bit add instruction on
1369         pointer values.
1370
1371 2007-12-01  Andreas Tobler  <a.tobler@schweiz.org>
1372
1373         PR libffi/31937
1374         * src/powerpc/ffitarget.h: Introduce new ABI FFI_LINUX_SOFT_FLOAT.
1375         Add local FFI_TYPE_UINT128 to handle soft-float long-double-128.
1376         * src/powerpc/ffi.c: Distinguish between __NO_FPRS__ and not and
1377         set the NUM_FPR_ARG_REGISTERS according to.
1378         Add support for potential soft-float support under hard-float
1379         architecture.
1380         (ffi_prep_args_SYSV): Set NUM_FPR_ARG_REGISTERS to 0 in case of
1381         FFI_LINUX_SOFT_FLOAT, handle float, doubles and long-doubles according
1382         to the FFI_LINUX_SOFT_FLOAT ABI.
1383         (ffi_prep_cif_machdep): Likewise.
1384         (ffi_closure_helper_SYSV): Likewise.
1385         * src/powerpc/ppc_closure.S: Make sure not to store float/double
1386         on archs where __NO_FPRS__ is true.
1387         Add FFI_TYPE_UINT128 support.
1388         * src/powerpc/sysv.S: Add support for soft-float long-double-128.
1389         Adjust copyright notice.
1390
1391 2007-11-25  Andreas Tobler  <a.tobler@schweiz.org>
1392
1393         * src/closures.c: Move defintion of MAYBE_UNUSED from here to ...
1394         * include/ffi_common.h: ... here.
1395         Update copyright.
1396
1397 2007-11-17  Andreas Tobler  <a.tobler@schweiz.org>
1398
1399         * src/powerpc/sysv.S: Load correct cr to compare if we have long double.
1400         * src/powerpc/linux64.S: Likewise.
1401         * src/powerpc/ffi.c: Add a comment to show which part goes into cr6.
1402         * testsuite/libffi.call/return_ldl.c: New test.
1403
1404 2007-09-04    <aph@redhat.com>
1405
1406         * src/arm/sysv.S (UNWIND): New.
1407         (Whole file): Conditionally compile unwinder directives.
1408         * src/arm/sysv.S: Add unwinder directives.
1409
1410         * src/arm/ffi.c (ffi_prep_args): Align structs by at least 4 bytes.
1411         Only treat r0 as a struct address if we're actually returning a
1412         struct by address.
1413         Only copy the bytes that are actually within a struct.
1414         (ffi_prep_cif_machdep): A Composite Type not larger than 4 bytes
1415         is returned in r0, not passed by address.
1416         (ffi_call): Allocate a word-sized temporary for the case where
1417         a composite is returned in r0.
1418         (ffi_prep_incoming_args_SYSV): Align as necessary.
1419
1420 2007-08-05  Steven Newbury  <s_j_newbury@yahoo.co.uk>
1421
1422         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Use __clear_cache instead of
1423         directly using the sys_cacheflush syscall.
1424
1425 2007-07-27  Andrew Haley  <aph@redhat.com>
1426
1427         * src/arm/sysv.S (ffi_closure_SYSV): Add soft-float.
1428
1429 2007-09-03  Maciej W. Rozycki  <macro@linux-mips.org>
1430
1431         * Makefile.am: Unify MIPS_IRIX and MIPS_LINUX into MIPS.
1432         * configure.ac: Likewise.
1433         * Makefile.in: Regenerate.
1434         * include/Makefile.in: Likewise.
1435         * testsuite/Makefile.in: Likewise.
1436         * configure: Likewise.
1437
1438 2007-08-24  David Daney  <ddaney@avtrex.com>
1439
1440         * testsuite/libffi.call/return_sl.c: New test.
1441
1442 2007-08-10  David Daney  <ddaney@avtrex.com>
1443
1444         * testsuite/libffi.call/cls_multi_ushort.c,
1445         testsuite/libffi.call/cls_align_uint16.c,
1446         testsuite/libffi.call/nested_struct1.c,
1447         testsuite/libffi.call/nested_struct3.c,
1448         testsuite/libffi.call/cls_7_1_byte.c,
1449         testsuite/libffi.call/nested_struct5.c,
1450         testsuite/libffi.call/cls_double.c,
1451         testsuite/libffi.call/nested_struct7.c,
1452         testsuite/libffi.call/cls_sint.c,
1453         testsuite/libffi.call/nested_struct9.c,
1454         testsuite/libffi.call/cls_20byte1.c,
1455         testsuite/libffi.call/cls_multi_sshortchar.c,
1456         testsuite/libffi.call/cls_align_sint64.c,
1457         testsuite/libffi.call/cls_3byte2.c,
1458         testsuite/libffi.call/cls_multi_schar.c,
1459         testsuite/libffi.call/cls_multi_uchar.c,
1460         testsuite/libffi.call/cls_19byte.c,
1461         testsuite/libffi.call/cls_9byte1.c,
1462         testsuite/libffi.call/cls_align_float.c,
1463         testsuite/libffi.call/closure_fn1.c,
1464         testsuite/libffi.call/problem1.c,
1465         testsuite/libffi.call/closure_fn3.c,
1466         testsuite/libffi.call/cls_sshort.c,
1467         testsuite/libffi.call/closure_fn5.c,
1468         testsuite/libffi.call/cls_align_double.c,
1469         testsuite/libffi.call/nested_struct.c,
1470         testsuite/libffi.call/cls_2byte.c,
1471         testsuite/libffi.call/nested_struct10.c,
1472         testsuite/libffi.call/cls_4byte.c,
1473         testsuite/libffi.call/cls_6byte.c,
1474         testsuite/libffi.call/cls_8byte.c,
1475         testsuite/libffi.call/cls_multi_sshort.c,
1476         testsuite/libffi.call/cls_align_sint16.c,
1477         testsuite/libffi.call/cls_align_uint32.c,
1478         testsuite/libffi.call/cls_20byte.c,
1479         testsuite/libffi.call/cls_float.c,
1480         testsuite/libffi.call/nested_struct2.c,
1481         testsuite/libffi.call/cls_5_1_byte.c,
1482         testsuite/libffi.call/nested_struct4.c,
1483         testsuite/libffi.call/cls_24byte.c,
1484         testsuite/libffi.call/nested_struct6.c,
1485         testsuite/libffi.call/cls_64byte.c,
1486         testsuite/libffi.call/nested_struct8.c,
1487         testsuite/libffi.call/cls_uint.c,
1488         testsuite/libffi.call/cls_multi_ushortchar.c,
1489         testsuite/libffi.call/cls_schar.c,
1490         testsuite/libffi.call/cls_uchar.c,
1491         testsuite/libffi.call/cls_align_uint64.c,
1492         testsuite/libffi.call/cls_ulonglong.c,
1493         testsuite/libffi.call/cls_align_longdouble.c,
1494         testsuite/libffi.call/cls_1_1byte.c,
1495         testsuite/libffi.call/cls_12byte.c,
1496         testsuite/libffi.call/cls_3_1byte.c,
1497         testsuite/libffi.call/cls_3byte1.c,
1498         testsuite/libffi.call/cls_4_1byte.c,
1499         testsuite/libffi.call/cls_6_1_byte.c,
1500         testsuite/libffi.call/cls_16byte.c,
1501         testsuite/libffi.call/cls_18byte.c,
1502         testsuite/libffi.call/closure_fn0.c,
1503         testsuite/libffi.call/cls_9byte2.c,
1504         testsuite/libffi.call/closure_fn2.c,
1505         testsuite/libffi.call/closure_fn4.c,
1506         testsuite/libffi.call/cls_ushort.c,
1507         testsuite/libffi.call/closure_fn6.c,
1508         testsuite/libffi.call/cls_5byte.c,
1509         testsuite/libffi.call/cls_align_pointer.c,
1510         testsuite/libffi.call/cls_7byte.c,
1511         testsuite/libffi.call/cls_align_sint32.c,
1512         testsuite/libffi.special/unwindtest_ffi_call.cc,
1513         testsuite/libffi.special/unwindtest.cc: Remove xfail for mips64*-*-*.
1514
1515 2007-08-10  David Daney  <ddaney@avtrex.com>
1516
1517         PR libffi/28313
1518         * configure.ac: Don't treat mips64 as a special case.
1519         * Makefile.am (nodist_libffi_la_SOURCES): Add n32.S.
1520         * configure: Regenerate
1521         * Makefile.in: Ditto.
1522         * fficonfig.h.in: Ditto.
1523         * src/mips/ffitarget.h (REG_L, REG_S, SUBU, ADDU, SRL, LI): Indent.
1524         (LA, EH_FRAME_ALIGN, FDE_ADDR_BYTES): New preprocessor macros.
1525         (FFI_DEFAULT_ABI): Set for n64 case.
1526         (FFI_CLOSURES, FFI_TRAMPOLINE_SIZE): Define for n32 and n64 cases.
1527         * src/mips/n32.S (ffi_call_N32): Add debug macros and labels for FDE.
1528         (ffi_closure_N32): New function.
1529         (.eh_frame): New section
1530         * src/mips/o32.S: Clean up comments.
1531         (ffi_closure_O32): Pass ffi_closure parameter in $12.
1532         * src/mips/ffi.c: Use FFI_MIPS_N32 instead of
1533         _MIPS_SIM == _ABIN32 throughout.
1534         (FFI_MIPS_STOP_HERE): New, use in place of
1535         ffi_stop_here.
1536         (ffi_prep_args): Use unsigned long to hold pointer values.  Rewrite
1537         to support n32/n64 ABIs.
1538         (calc_n32_struct_flags): Rewrite.
1539         (calc_n32_return_struct_flags): Remove unused variable.  Reverse
1540         position of flag bits.
1541         (ffi_prep_cif_machdep): Rewrite n32 portion.
1542         (ffi_call): Enable for n64.  Add special handling for small structure
1543         return values.
1544         (ffi_prep_closure_loc): Add n32 and n64 support.
1545         (ffi_closure_mips_inner_O32): Add cast to silence warning.
1546         (copy_struct_N32, ffi_closure_mips_inner_N32): New functions.
1547
1548 2007-08-08  David Daney  <ddaney@avtrex.com>
1549
1550         * testsuite/libffi.call/ffitest.h (ffi_type_mylong): Remove definition.
1551         * testsuite/libffi.call/cls_align_uint16.c (main): Use correct type
1552         specifiers.
1553         * testsuite/libffi.call/nested_struct1.c (main): Ditto.
1554         * testsuite/libffi.call/cls_sint.c (main): Ditto.
1555         * testsuite/libffi.call/nested_struct9.c (main): Ditto.
1556         * testsuite/libffi.call/cls_20byte1.c (main): Ditto.
1557         * testsuite/libffi.call/cls_9byte1.c (main): Ditto.
1558         * testsuite/libffi.call/closure_fn1.c (main): Ditto.
1559         * testsuite/libffi.call/closure_fn3.c (main): Ditto.
1560         * testsuite/libffi.call/return_dbl2.c (main): Ditto.
1561         * testsuite/libffi.call/cls_sshort.c (main): Ditto.
1562         * testsuite/libffi.call/return_fl3.c (main): Ditto.
1563         * testsuite/libffi.call/closure_fn5.c (main): Ditto.
1564         * testsuite/libffi.call/nested_struct.c (main): Ditto.
1565         * testsuite/libffi.call/nested_struct10.c (main): Ditto.
1566         * testsuite/libffi.call/return_ll1.c (main): Ditto.
1567         * testsuite/libffi.call/cls_8byte.c (main): Ditto.
1568         * testsuite/libffi.call/cls_align_uint32.c (main): Ditto.
1569         * testsuite/libffi.call/cls_align_sint16.c (main): Ditto.
1570         * testsuite/libffi.call/cls_20byte.c (main): Ditto.
1571         * testsuite/libffi.call/nested_struct2.c (main): Ditto.
1572         * testsuite/libffi.call/cls_24byte.c (main): Ditto.
1573         * testsuite/libffi.call/nested_struct6.c (main): Ditto.
1574         * testsuite/libffi.call/cls_uint.c (main): Ditto.
1575         * testsuite/libffi.call/cls_12byte.c (main): Ditto.
1576         * testsuite/libffi.call/cls_16byte.c (main): Ditto.
1577         * testsuite/libffi.call/closure_fn0.c (main): Ditto.
1578         * testsuite/libffi.call/cls_9byte2.c (main): Ditto.
1579         * testsuite/libffi.call/closure_fn2.c (main): Ditto.
1580         * testsuite/libffi.call/return_dbl1.c (main): Ditto.
1581         * testsuite/libffi.call/closure_fn4.c (main): Ditto.
1582         * testsuite/libffi.call/closure_fn6.c (main): Ditto.
1583         * testsuite/libffi.call/cls_align_sint32.c (main): Ditto.
1584
1585 2007-08-07  Andrew Haley  <aph@redhat.com>
1586
1587         * src/x86/sysv.S (ffi_closure_raw_SYSV): Fix typo in previous
1588         checkin.
1589
1590 2007-08-06  Andrew Haley  <aph@redhat.com>
1591
1592         PR testsuite/32843
1593         * src/x86/sysv.S (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1594         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1595         FFI_TYPE_SINT32.
1596
1597 2007-08-02  David Daney  <ddaney@avtrex.com>
1598
1599         * testsuite/libffi.call/return_ul.c (main): Define return type as
1600         ffi_arg.  Use proper printf conversion specifier.
1601
1602 2007-07-30  Andrew Haley  <aph@redhat.com>
1603
1604         PR testsuite/32843
1605         * src/x86/ffi.c (ffi_prep_cif_machdep): in x86 case, add code for
1606         signed/unsigned int8/16.
1607         * src/x86/sysv.S (ffi_call_SYSV): Rewrite to:
1608         Use a jump table.
1609         Remove code to pop args from the stack after call.
1610         Special-case signed/unsigned int8/16.
1611         * testsuite/libffi.call/return_sc.c (main): Revert.
1612
1613 2007-07-26  Richard Guenther  <rguenther@suse.de>
1614
1615         PR testsuite/32843
1616         * testsuite/libffi.call/return_sc.c (main): Verify call
1617         result as signed char, not ffi_arg.
1618
1619 2007-07-16  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1620
1621         * configure.ac (i?86-*-solaris2.1[0-9]): Set TARGET to X86_64.
1622         * configure: Regenerate.
1623
1624 2007-07-11  David Daney  <ddaney@avtrex.com>
1625
1626         * src/mips/ffi.c: Don't include sys/cachectl.h.
1627         (ffi_prep_closure_loc): Use __builtin___clear_cache() instead of
1628         cacheflush().
1629
1630 2007-05-18  Aurelien Jarno  <aurelien@aurel32.net>
1631
1632         * src/arm/ffi.c (ffi_prep_closure_loc): Renamed and ajusted
1633         from (ffi_prep_closure): ... this.
1634         (FFI_INIT_TRAMPOLINE): Adjust.
1635
1636 2005-12-31  Phil Blundell  <pb@reciva.com>
1637
1638         * src/arm/ffi.c (ffi_prep_incoming_args_SYSV,
1639         ffi_closure_SYSV_inner, ffi_prep_closure): New, add closure support.
1640         * src/arm/sysv.S(ffi_closure_SYSV): Likewise.
1641         * src/arm/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1642         (FFI_CLOSURES): Enable closure support.
1643
1644 2007-07-03  Andrew Haley  <aph@hedges.billgatliff.com>
1645
1646         * testsuite/libffi.call/cls_multi_ushort.c,
1647         testsuite/libffi.call/cls_align_uint16.c,
1648         testsuite/libffi.call/nested_struct1.c,
1649         testsuite/libffi.call/nested_struct3.c,
1650         testsuite/libffi.call/cls_7_1_byte.c,
1651         testsuite/libffi.call/cls_double.c,
1652         testsuite/libffi.call/nested_struct5.c,
1653         testsuite/libffi.call/nested_struct7.c,
1654         testsuite/libffi.call/cls_sint.c,
1655         testsuite/libffi.call/nested_struct9.c,
1656         testsuite/libffi.call/cls_20byte1.c,
1657         testsuite/libffi.call/cls_multi_sshortchar.c,
1658         testsuite/libffi.call/cls_align_sint64.c,
1659         testsuite/libffi.call/cls_3byte2.c,
1660         testsuite/libffi.call/cls_multi_schar.c,
1661         testsuite/libffi.call/cls_multi_uchar.c,
1662         testsuite/libffi.call/cls_19byte.c,
1663         testsuite/libffi.call/cls_9byte1.c,
1664         testsuite/libffi.call/cls_align_float.c,
1665         testsuite/libffi.call/closure_fn1.c,
1666         testsuite/libffi.call/problem1.c,
1667         testsuite/libffi.call/closure_fn3.c,
1668         testsuite/libffi.call/cls_sshort.c,
1669         testsuite/libffi.call/closure_fn5.c,
1670         testsuite/libffi.call/cls_align_double.c,
1671         testsuite/libffi.call/cls_2byte.c,
1672         testsuite/libffi.call/nested_struct.c,
1673         testsuite/libffi.call/nested_struct10.c,
1674         testsuite/libffi.call/cls_4byte.c,
1675         testsuite/libffi.call/cls_6byte.c,
1676         testsuite/libffi.call/cls_8byte.c,
1677         testsuite/libffi.call/cls_multi_sshort.c,
1678         testsuite/libffi.call/cls_align_uint32.c,
1679         testsuite/libffi.call/cls_align_sint16.c,
1680         testsuite/libffi.call/cls_float.c,
1681         testsuite/libffi.call/cls_20byte.c,
1682         testsuite/libffi.call/cls_5_1_byte.c,
1683         testsuite/libffi.call/nested_struct2.c,
1684         testsuite/libffi.call/cls_24byte.c,
1685         testsuite/libffi.call/nested_struct4.c,
1686         testsuite/libffi.call/nested_struct6.c,
1687         testsuite/libffi.call/cls_64byte.c,
1688         testsuite/libffi.call/nested_struct8.c,
1689         testsuite/libffi.call/cls_uint.c,
1690         testsuite/libffi.call/cls_multi_ushortchar.c,
1691         testsuite/libffi.call/cls_schar.c,
1692         testsuite/libffi.call/cls_uchar.c,
1693         testsuite/libffi.call/cls_align_uint64.c,
1694         testsuite/libffi.call/cls_ulonglong.c,
1695         testsuite/libffi.call/cls_align_longdouble.c,
1696         testsuite/libffi.call/cls_1_1byte.c,
1697         testsuite/libffi.call/cls_12byte.c,
1698         testsuite/libffi.call/cls_3_1byte.c,
1699         testsuite/libffi.call/cls_3byte1.c,
1700         testsuite/libffi.call/cls_4_1byte.c,
1701         testsuite/libffi.call/cls_6_1_byte.c,
1702         testsuite/libffi.call/cls_16byte.c,
1703         testsuite/libffi.call/cls_18byte.c,
1704         testsuite/libffi.call/closure_fn0.c,
1705         testsuite/libffi.call/cls_9byte2.c,
1706         testsuite/libffi.call/closure_fn2.c,
1707         testsuite/libffi.call/closure_fn4.c,
1708         testsuite/libffi.call/cls_ushort.c,
1709         testsuite/libffi.call/closure_fn6.c,
1710         testsuite/libffi.call/cls_5byte.c,
1711         testsuite/libffi.call/cls_align_pointer.c,
1712         testsuite/libffi.call/cls_7byte.c,
1713         testsuite/libffi.call/cls_align_sint32.c,
1714         testsuite/libffi.special/unwindtest_ffi_call.cc,
1715         testsuite/libffi.special/unwindtest.cc: Enable for ARM.
1716
1717 2007-07-05  H.J. Lu  <hongjiu.lu@intel.com>
1718
1719         * aclocal.m4: Regenerated.
1720
1721 2007-06-02  Paolo Bonzini  <bonzini@gnu.org>
1722
1723         * configure: Regenerate.
1724
1725 2007-05-23  Steve Ellcey  <sje@cup.hp.com>
1726
1727         * Makefile.in: Regenerate.
1728         * configure: Regenerate.
1729         * aclocal.m4: Regenerate.
1730         * include/Makefile.in: Regenerate.
1731         * testsuite/Makefile.in: Regenerate.
1732
1733 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1734
1735         * src/m68k/ffi.c (ffi_prep_incoming_args_SYSV,
1736         ffi_closure_SYSV_inner,ffi_prep_closure): New, add closure support.
1737         * src/m68k/sysv.S(ffi_closure_SYSV,ffi_closure_struct_SYSV): Likewise.
1738         * src/m68k/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1739         (FFI_CLOSURES): Enable closure support.
1740
1741 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1742
1743         * configure.ac (HAVE_AS_CFI_PSEUDO_OP): New test.
1744         * configure: Regenerate.
1745         * fficonfig.h.in: Regenerate.
1746         * src/m68k/sysv.S (CFI_STARTPROC,CFI_ENDPROC,
1747         CFI_OFFSET,CFI_DEF_CFA): New macros.
1748         (ffi_call_SYSV): Add callframe annotation.
1749
1750 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1751
1752         * src/m68k/ffi.c (ffi_prep_args,ffi_prep_cif_machdep): Fix
1753         numerous test suite failures.
1754         * src/m68k/sysv.S (ffi_call_SYSV): Likewise.
1755
1756 2007-04-11  Paolo Bonzini  <bonzini@gnu.org>
1757
1758         * Makefile.am (EXTRA_DIST): Bring up to date.
1759         * Makefile.in: Regenerate.
1760         * src/frv/eabi.S: Remove RCS keyword.
1761
1762 2007-04-06  Richard Henderson  <rth@redhat.com>
1763
1764         * configure.ac: Tidy target case.
1765         (HAVE_LONG_DOUBLE): Allow the target to override.
1766         * configure: Regenerate.
1767         * include/ffi.h.in: Don't define ffi_type_foo if
1768         LIBFFI_HIDE_BASIC_TYPES is defined.
1769         (ffi_type_longdouble): If not HAVE_LONG_DOUBLE, define
1770         to ffi_type_double.
1771         * types.c (LIBFFI_HIDE_BASIC_TYPES): Define.
1772         (FFI_TYPEDEF, ffi_type_void): Mark the data const.
1773         (ffi_type_longdouble): Special case for Alpha.  Don't define
1774         if long double == double.
1775
1776         * src/alpha/ffi.c (FFI_TYPE_LONGDOUBLE): Assert unique value.
1777         (ffi_prep_cif_machdep): Handle it as the 128-bit type.
1778         (ffi_call, ffi_closure_osf_inner): Likewise.
1779         (ffi_closure_osf_inner): Likewise.  Mark hidden.
1780         (ffi_call_osf, ffi_closure_osf): Mark hidden.
1781         * src/alpha/ffitarget.h (FFI_LAST_ABI): Tidy definition.
1782         * src/alpha/osf.S (ffi_call_osf, ffi_closure_osf): Mark hidden.
1783         (load_table): Handle 128-bit long double.
1784
1785         * testsuite/libffi.call/float4.c: Add -mieee for alpha.
1786
1787 2007-04-06  Tom Tromey  <tromey@redhat.com>
1788
1789         PR libffi/31491:
1790         * README: Fixed bug in example.
1791
1792 2007-04-03  Jakub Jelinek  <jakub@redhat.com>
1793
1794         * src/closures.c: Include sys/statfs.h.
1795         (_GNU_SOURCE): Define on Linux.
1796         (FFI_MMAP_EXEC_SELINUX): Define.
1797         (selinux_enabled): New variable.
1798         (selinux_enabled_check): New function.
1799         (is_selinux_enabled): Define.
1800         (dlmmap): Use it.
1801
1802 2007-03-24  Uros Bizjak  <ubizjak@gmail.com>
1803
1804         * testsuite/libffi.call/return_fl2.c (return_fl): Mark as static.
1805         Use 'volatile float sum' to create sum of floats to avoid false
1806         negative due to excess precision on ix86 targets.
1807         (main): Ditto.
1808
1809 2007-03-08  Alexandre Oliva  <aoliva@redhat.com>
1810
1811         * src/powerpc/ffi.c (flush_icache): Fix left-over from previous
1812         patch.
1813         (ffi_prep_closure_loc): Remove unneeded casts.  Add needed ones.
1814
1815 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1816
1817         * include/ffi.h.in (ffi_closure_alloc, ffi_closure_free): New.
1818         (ffi_prep_closure_loc): New.
1819         (ffi_prep_raw_closure_loc): New.
1820         (ffi_prep_java_raw_closure_loc): New.
1821         * src/closures.c: New file.
1822         * src/dlmalloc.c [FFI_MMAP_EXEC_WRIT] (struct malloc_segment):
1823         Replace sflags with exec_offset.
1824         [FFI_MMAP_EXEC_WRIT] (mmap_exec_offset, add_segment_exec_offset,
1825         sub_segment_exec_offset): New macros.
1826         (get_segment_flags, set_segment_flags, check_segment_merge): New
1827         macros.
1828         (is_mmapped_segment, is_extern_segment): Use get_segment_flags.
1829         (add_segment, sys_alloc, create_mspace, create_mspace_with_base,
1830         destroy_mspace): Use new macros.
1831         (sys_alloc): Silence warning.
1832         * Makefile.am (libffi_la_SOURCES): Add src/closures.c.
1833         * Makefile.in: Rebuilt.
1834         * src/prep_cif [FFI_CLOSURES] (ffi_prep_closure): Implement in
1835         terms of ffi_prep_closure_loc.
1836         * src/raw_api.c (ffi_prep_raw_closure_loc): Renamed and adjusted
1837         from...
1838         (ffi_prep_raw_closure): ... this.  Re-implement in terms of the
1839         renamed version.
1840         * src/java_raw_api (ffi_prep_java_raw_closure_loc): Renamed and
1841         adjusted from...
1842         (ffi_prep_java_raw_closure): ... this.  Re-implement in terms of
1843         the renamed version.
1844         * src/alpha/ffi.c (ffi_prep_closure_loc): Renamed from
1845         (ffi_prep_closure): ... this.
1846         * src/pa/ffi.c: Likewise.
1847         * src/cris/ffi.c: Likewise.  Adjust.
1848         * src/frv/ffi.c: Likewise.
1849         * src/ia64/ffi.c: Likewise.
1850         * src/mips/ffi.c: Likewise.
1851         * src/powerpc/ffi_darwin.c: Likewise.
1852         * src/s390/ffi.c: Likewise.
1853         * src/sh/ffi.c: Likewise.
1854         * src/sh64/ffi.c: Likewise.
1855         * src/sparc/ffi.c: Likewise.
1856         * src/x86/ffi64.c: Likewise.
1857         * src/x86/ffi.c: Likewise.
1858         (FFI_INIT_TRAMPOLINE): Adjust.
1859         (ffi_prep_raw_closure_loc): Renamed and adjusted from...
1860         (ffi_prep_raw_closure): ... this.
1861         * src/powerpc/ffi.c (ffi_prep_closure_loc): Renamed from
1862         (ffi_prep_closure): ... this.
1863         (flush_icache): Adjust.
1864
1865 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1866
1867         * src/dlmalloc.c: New file, imported version 2.8.3 of Doug
1868         Lea's malloc.
1869
1870 2007-03-01  Brooks Moses  <brooks.moses@codesourcery.com>
1871
1872         * Makefile.am: Add dummy install-pdf target.
1873         * Makefile.in: Regenerate
1874
1875 2007-02-13  Andreas Krebbel  <krebbel1@de.ibm.com>
1876
1877         * src/s390/ffi.c (ffi_prep_args, ffi_prep_cif_machdep,
1878         ffi_closure_helper_SYSV): Add long double handling.
1879
1880 2007-02-02  Jakub Jelinek  <jakub@redhat.com>
1881
1882         * src/powerpc/linux64.S (ffi_call_LINUX64): Move restore of r2
1883         immediately after bctrl instruction.
1884
1885 2007-01-18  Alexandre Oliva  <aoliva@redhat.com>
1886
1887         * Makefile.am (all-recursive, install-recursive,
1888         mostlyclean-recursive, clean-recursive, distclean-recursive,
1889         maintainer-clean-recursive): Add missing targets.
1890         * Makefile.in: Rebuilt.
1891
1892 2006-12-14  Andreas Tobler  <a.tobler@schweiz.org>
1893
1894         * configure.ac: Add TARGET for x86_64-*-darwin*.
1895         * Makefile.am (nodist_libffi_la_SOURCES): Add rules for 64-bit sources
1896         for X86_DARWIN.
1897         * src/x86/ffitarget.h: Set trampoline size for x86_64-*-darwin*.
1898         * src/x86/darwin64.S: New file for x86_64-*-darwin* support.
1899         * configure: Regenerate.
1900         * Makefile.in: Regenerate.
1901         * include/Makefile.in: Regenerate.
1902         * testsuite/Makefile.in: Regenerate.
1903         * testsuite/libffi.special/unwindtest_ffi_call.cc: New test case for
1904         ffi_call only.
1905
1906 2006-12-13  Andreas Tobler <a.tobler@schweiz.org>
1907
1908         * aclocal.m4: Regenerate with aclocal -I .. as written in the
1909         Makefile.am.
1910
1911 2006-10-31  Geoffrey Keating  <geoffk@apple.com>
1912
1913         * src/powerpc/ffi_darwin.c (darwin_adjust_aggregate_sizes): New.
1914         (ffi_prep_cif_machdep): Call darwin_adjust_aggregate_sizes for
1915         Darwin.
1916         * testsuite/libffi.call/nested_struct4.c: Remove Darwin XFAIL.
1917         * testsuite/libffi.call/nested_struct6.c: Remove Darwin XFAIL.
1918
1919 2006-10-10  Paolo Bonzini  <bonzini@gnu.org>
1920             Sandro Tolaini  <tolaini@libero.it>
1921
1922         * configure.ac [i*86-*-darwin*]: Set X86_DARWIN symbol and
1923         conditional.
1924         * configure: Regenerated.
1925         * Makefile.am (nodist_libffi_la_SOURCES) [X86_DARWIN]: New case.
1926         (EXTRA_DIST): Add src/x86/darwin.S.
1927         * Makefile.in: Regenerated.
1928         * include/Makefile.in: Regenerated.
1929         * testsuite/Makefile.in: Regenerated.
1930
1931         * src/x86/ffi.c (ffi_prep_cif_machdep) [X86_DARWIN]: Treat like
1932         X86_WIN32, and additionally align stack to 16 bytes.
1933         * src/x86/darwin.S: New, based on sysv.S.
1934         * src/prep_cif.c (ffi_prep_cif) [X86_DARWIN]: Align > 8-byte structs.
1935
1936 2006-09-12  David Daney  <ddaney@avtrex.com>
1937
1938         PR libffi/23935
1939         * include/Makefile.am: Install both ffi.h and ffitarget.h in
1940         $(libdir)/gcc/$(target_alias)/$(gcc_version)/include.
1941         * aclocal.m4: Regenerated for automake 1.9.6.
1942         * Makefile.in: Regenerated.
1943         * include/Makefile.in: Regenerated.
1944         * testsuite/Makefile.in: Regenerated.
1945
1946 2006-08-17  Andreas Tobler  <a.tobler@schweiz.ch>
1947
1948         * include/ffi_common.h (struct): Revert accidental commit.
1949
1950 2006-08-15  Andreas Tobler  <a.tobler@schweiz.ch>
1951
1952         * include/ffi_common.h: Remove lint directives.
1953         * include/ffi.h.in: Likewise.
1954
1955 2006-07-25  Torsten Schoenfeld  <kaffeetisch@gmx.de>
1956
1957         * include/ffi.h.in (ffi_type_ulong, ffi_type_slong): Define correctly
1958         for 32-bit architectures.
1959         * testsuite/libffi.call/return_ul.c: New test case.
1960
1961 2006-07-19  David Daney  <ddaney@avtrex.com>
1962
1963         * testsuite/libffi.call/closure_fn6.c: Remove xfail for mips,
1964         xfail remains for mips64.
1965
1966 2006-05-23  Carlos O'Donell  <carlos@codesourcery.com>
1967
1968         * Makefile.am: Add install-html target. Add install-html to .PHONY
1969         * Makefile.in: Regenerate.
1970         * aclocal.m4: Regenerate.
1971         * include/Makefile.in: Regenerate.
1972         * testsuite/Makefile.in: Regenerate.
1973
1974 2006-05-18  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1975
1976         * pa/ffi.c (ffi_prep_args_pa32): Load floating point arguments from
1977         stack slot.
1978
1979 2006-04-22  Andreas Tobler  <a.tobler@schweiz.ch>
1980
1981         * README: Remove notice about 'Crazy Comments'.
1982         * src/debug.c: Remove lint directives. Cleanup white spaces.
1983         * src/java_raw_api.c: Likewise.
1984         * src/prep_cif.c: Likewise.
1985         * src/raw_api.c: Likewise.
1986         * src/ffitest.c: Delete. No longer needed, all test cases migrated
1987         to the testsuite.
1988         * src/arm/ffi.c: Remove lint directives.
1989         * src/m32r/ffi.c: Likewise.
1990         * src/pa/ffi.c: Likewise.
1991         * src/powerpc/ffi.c: Likewise.
1992         * src/powerpc/ffi_darwin.c: Likewise.
1993         * src/sh/ffi.c: Likewise.
1994         * src/sh64/ffi.c: Likewise.
1995         * src/x86/ffi.c: Likewise.
1996         * testsuite/libffi.call/float2.c: Likewise.
1997         * testsuite/libffi.call/promotion.c: Likewise.
1998         * testsuite/libffi.call/struct1.c: Likewise.
1999
2000 2006-04-13  Andreas Tobler  <a.tobler@schweiz.ch>
2001
2002         * src/pa/hpux32.S: Correct unwind offset calculation for
2003         ffi_closure_pa32.
2004         * src/pa/linux.S: Likewise.
2005
2006 2006-04-12  James E Wilson  <wilson@specifix.com>
2007
2008         PR libgcj/26483
2009         * src/ia64/ffi.c (stf_spill, ldf_fill): Rewrite as macros.
2010         (hfa_type_load): Call stf_spill.
2011         (hfa_type_store): Call ldf_fill.
2012         (ffi_call): Adjust calls to above routines.  Add local temps for
2013         macro result.
2014
2015 2006-04-10  Matthias Klose  <doko@debian.org>
2016
2017         * testsuite/lib/libffi-dg.exp (libffi-init): Recognize multilib
2018         directory names containing underscores.
2019
2020 2006-04-07  James E Wilson  <wilson@specifix.com>
2021
2022         * testsuite/libffi.call/float4.c: New testcase.
2023
2024 2006-04-05  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2025             Andreas Tobler  <a.tobler@schweiz.ch>
2026
2027         * Makefile.am: Add PA_HPUX port.
2028         * Makefile.in: Regenerate.
2029         * include/Makefile.in: Likewise.
2030         * testsuite/Makefile.in: Likewise.
2031         * configure.ac: Add PA_HPUX rules.
2032         * configure: Regenerate.
2033         * src/pa/ffitarget.h: Rename linux target to PA_LINUX.
2034         Add PA_HPUX and PA64_HPUX.
2035         Rename FFI_LINUX ABI to FFI_PA32 ABI.
2036         (FFI_TRAMPOLINE_SIZE): Define for 32-bit HP-UX targets.
2037         (FFI_TYPE_SMALL_STRUCT2): Define.
2038         (FFI_TYPE_SMALL_STRUCT4): Likewise.
2039         (FFI_TYPE_SMALL_STRUCT8): Likewise.
2040         (FFI_TYPE_SMALL_STRUCT3): Redefine.
2041         (FFI_TYPE_SMALL_STRUCT5): Likewise.
2042         (FFI_TYPE_SMALL_STRUCT6): Likewise.
2043         (FFI_TYPE_SMALL_STRUCT7): Likewise.
2044         * src/pa/ffi.c (ROUND_DOWN): Delete.
2045         (fldw, fstw, fldd, fstd): Use '__asm__'.
2046         (ffi_struct_type): Add support for FFI_TYPE_SMALL_STRUCT2,
2047         FFI_TYPE_SMALL_STRUCT4 and FFI_TYPE_SMALL_STRUCT8.
2048         (ffi_prep_args_LINUX): Rename to ffi_prep_args_pa32. Update comment.
2049         Simplify incrementing of stack slot variable. Change type of local
2050         'n' to unsigned int.
2051         (ffi_size_stack_LINUX): Rename to ffi_size_stack_pa32. Handle long
2052         double on PA_HPUX.
2053         (ffi_prep_cif_machdep): Likewise.
2054         (ffi_call): Likewise.
2055         (ffi_closure_inner_LINUX): Rename to ffi_closure_inner_pa32. Change
2056         return type to ffi_status. Simplify incrementing of stack slot
2057         variable. Only copy floating point argument registers when PA_LINUX
2058         is true. Reformat debug statement.
2059         Add support for FFI_TYPE_SMALL_STRUCT2, FFI_TYPE_SMALL_STRUCT4 and
2060         FFI_TYPE_SMALL_STRUCT8.
2061         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Add 'extern' to
2062         declaration.
2063         (ffi_prep_closure): Make linux trampoline conditional on PA_LINUX.
2064         Add nops to cache flush.  Add trampoline for PA_HPUX.
2065         * src/pa/hpux32.S: New file.
2066         * src/pa/linux.S (ffi_call_LINUX): Rename to ffi_call_pa32. Rename
2067         ffi_prep_args_LINUX to ffi_prep_args_pa32.
2068         Localize labels. Add support for 2, 4 and 8-byte small structs. Handle
2069         unaligned destinations in 3, 5, 6 and 7-byte small structs. Order
2070         argument type checks so that common argument types appear first.
2071         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Rename
2072         ffi_closure_inner_LINUX to ffi_closure_inner_pa32.
2073
2074 2006-03-24  Alan Modra  <amodra@bigpond.net.au>
2075
2076         * src/powerpc/ffitarget.h (enum ffi_abi): Add FFI_LINUX.  Default
2077         for 32-bit using IBM extended double format.  Fix FFI_LAST_ABI.
2078         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Handle linux variant of
2079         FFI_TYPE_LONGDOUBLE.
2080         (ffi_prep_args64): Assert using IBM extended double.
2081         (ffi_prep_cif_machdep): Don't munge FFI_TYPE_LONGDOUBLE type.
2082         Handle FFI_LINUX FFI_TYPE_LONGDOUBLE return and args.
2083         (ffi_call): Handle FFI_LINUX.
2084         (ffi_closure_helper_SYSV): Non FFI_LINUX long double return needs
2085         gpr3 return pointer as for struct return.  Handle FFI_LINUX
2086         FFI_TYPE_LONGDOUBLE return and args.  Don't increment "nf"
2087         unnecessarily.
2088         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Load both f1 and f2
2089         for FFI_TYPE_LONGDOUBLE.  Move epilogue insns into case table.
2090         Don't use r6 as pointer to results, instead use sp offset.  Don't
2091         make a special call to load lr with case table address, instead
2092         use offset from previous call.
2093         * src/powerpc/sysv.S (ffi_call_SYSV): Save long double return.
2094         * src/powerpc/linux64.S (ffi_call_LINUX64): Simplify long double
2095         return.
2096
2097 2006-03-15  Kaz Kojima  <kkojima@gcc.gnu.org>
2098
2099         * src/sh64/ffi.c (ffi_prep_cif_machdep): Handle float arguments
2100         passed with FP registers correctly.
2101         (ffi_closure_helper_SYSV): Likewise.
2102         * src/sh64/sysv.S: Likewise.
2103
2104 2006-03-01  Andreas Tobler  <a.tobler@schweiz.ch>
2105
2106         * testsuite/libffi.special/unwindtest.cc (closure_test_fn): Mark cif,
2107         args and userdata unused.
2108         (closure_test_fn1): Mark cif and userdata unused.
2109         (main): Remove unused res.
2110
2111 2006-02-28  Andreas Tobler  <a.tobler@schweiz.ch>
2112
2113         * testsuite/libffi.call/call.exp: Adjust FSF address. Add test runs for
2114         -O2, -O3, -Os and the warning flags -W -Wall.
2115         * testsuite/libffi.special/special.exp: Likewise.
2116         * testsuite/libffi.call/ffitest.h: Add an __UNUSED__ macro to mark
2117         unused parameter unused for gcc or else do nothing.
2118         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2119         * testsuite/libffi.call/cls_12byte.c (cls_struct_12byte_gn): Mark cif
2120         and userdata unused.
2121         * testsuite/libffi.call/cls_16byte.c (cls_struct_16byte_gn): Likewise.
2122         * testsuite/libffi.call/cls_18byte.c (cls_struct_18byte_gn): Likewise.
2123         * testsuite/libffi.call/cls_19byte.c (cls_struct_19byte_gn): Likewise.
2124         * testsuite/libffi.call/cls_1_1byte.c (cls_struct_1_1byte_gn): Likewise.
2125         * testsuite/libffi.call/cls_20byte.c (cls_struct_20byte_gn): Likewise.
2126         * testsuite/libffi.call/cls_20byte1.c (cls_struct_20byte_gn): Likewise.
2127         * testsuite/libffi.call/cls_24byte.c (cls_struct_24byte_gn): Likewise.
2128         * testsuite/libffi.call/cls_2byte.c (cls_struct_2byte_gn): Likewise.
2129         * testsuite/libffi.call/cls_3_1byte.c (cls_struct_3_1byte_gn): Likewise.
2130         * testsuite/libffi.call/cls_3byte1.c (cls_struct_3byte_gn): Likewise.
2131         * testsuite/libffi.call/cls_3byte2.c (cls_struct_3byte_gn1): Likewise.
2132         * testsuite/libffi.call/cls_4_1byte.c (cls_struct_4_1byte_gn): Likewise.
2133         * testsuite/libffi.call/cls_4byte.c (cls_struct_4byte_gn): Likewise.
2134         * testsuite/libffi.call/cls_5_1_byte.c (cls_struct_5byte_gn): Likewise.
2135         * testsuite/libffi.call/cls_5byte.c (cls_struct_5byte_gn): Likewise.
2136         * testsuite/libffi.call/cls_64byte.c (cls_struct_64byte_gn): Likewise.
2137         * testsuite/libffi.call/cls_6_1_byte.c (cls_struct_6byte_gn): Likewise.
2138         * testsuite/libffi.call/cls_6byte.c (cls_struct_6byte_gn): Likewise.
2139         * testsuite/libffi.call/cls_7_1_byte.c (cls_struct_7byte_gn): Likewise.
2140         * testsuite/libffi.call/cls_7byte.c (cls_struct_7byte_gn): Likewise.
2141         * testsuite/libffi.call/cls_8byte.c (cls_struct_8byte_gn): Likewise.
2142         * testsuite/libffi.call/cls_9byte1.c (cls_struct_9byte_gn): Likewise.
2143         * testsuite/libffi.call/cls_9byte2.c (cls_struct_9byte_gn): Likewise.
2144         * testsuite/libffi.call/cls_align_double.c (cls_struct_align_gn):
2145         Likewise.
2146         * testsuite/libffi.call/cls_align_float.c (cls_struct_align_gn):
2147         Likewise.
2148         * testsuite/libffi.call/cls_align_longdouble.c (cls_struct_align_gn):
2149         Likewise.
2150         * testsuite/libffi.call/cls_align_pointer.c (cls_struct_align_fn): Cast
2151         void* to avoid compiler warning.
2152         (main): Likewise.
2153         (cls_struct_align_gn): Mark cif and userdata unused.
2154         * testsuite/libffi.call/cls_align_sint16.c (cls_struct_align_gn):
2155         Likewise.
2156         * testsuite/libffi.call/cls_align_sint32.c (cls_struct_align_gn):
2157         Likewise.
2158         * testsuite/libffi.call/cls_align_sint64.c (cls_struct_align_gn):
2159         Likewise.
2160         * testsuite/libffi.call/cls_align_uint16.c (cls_struct_align_gn):
2161         Likewise.
2162         * testsuite/libffi.call/cls_align_uint32.c (cls_struct_align_gn):
2163         Likewise.
2164         * testsuite/libffi.call/cls_double.c (cls_ret_double_fn): Likewise.
2165         * testsuite/libffi.call/cls_float.c (cls_ret_float_fn): Likewise.
2166         * testsuite/libffi.call/cls_multi_schar.c (test_func_gn): Mark cif and
2167         data unused.
2168         (main): Cast res_call to silence gcc.
2169         * testsuite/libffi.call/cls_multi_sshort.c (test_func_gn): Mark cif and
2170         data unused.
2171         (main): Cast res_call to silence gcc.
2172         * testsuite/libffi.call/cls_multi_sshortchar.c (test_func_gn): Mark cif
2173         and data unused.
2174         (main): Cast res_call to silence gcc.
2175         * testsuite/libffi.call/cls_multi_uchar.c (test_func_gn): Mark cif and
2176         data unused.
2177         (main): Cast res_call to silence gcc.
2178         * testsuite/libffi.call/cls_multi_ushort.c (test_func_gn): Mark cif and
2179         data unused.
2180         (main): Cast res_call to silence gcc.
2181         * testsuite/libffi.call/cls_multi_ushortchar.c (test_func_gn): Mark cif
2182         and data unused.
2183         (main): Cast res_call to silence gcc.
2184         * testsuite/libffi.call/cls_schar.c (cls_ret_schar_fn): Mark cif and
2185         userdata unused.
2186         (cls_ret_schar_fn): Cast printf parameter to silence gcc.
2187         * testsuite/libffi.call/cls_sint.c (cls_ret_sint_fn): Mark cif and
2188         userdata unused.
2189         (cls_ret_sint_fn): Cast printf parameter to silence gcc.
2190         * testsuite/libffi.call/cls_sshort.c (cls_ret_sshort_fn): Mark cif and
2191         userdata unused.
2192         (cls_ret_sshort_fn): Cast printf parameter to silence gcc.
2193         * testsuite/libffi.call/cls_uchar.c (cls_ret_uchar_fn):  Mark cif and
2194         userdata unused.
2195         (cls_ret_uchar_fn): Cast printf parameter to silence gcc.
2196         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Mark cif and
2197         userdata unused.
2198         (cls_ret_uint_fn): Cast printf parameter to silence gcc.
2199         * testsuite/libffi.call/cls_ulonglong.c (cls_ret_ulonglong_fn): Mark cif
2200         and userdata unused.
2201         * testsuite/libffi.call/cls_ushort.c (cls_ret_ushort_fn): Mark cif and
2202         userdata unused.
2203         (cls_ret_ushort_fn): Cast printf parameter to silence gcc.
2204         * testsuite/libffi.call/float.c (floating): Remove unused parameter e.
2205         * testsuite/libffi.call/float1.c (main): Remove unused variable i.
2206         Cleanup white spaces.
2207         * testsuite/libffi.call/negint.c (checking): Remove unused variable i.
2208         * testsuite/libffi.call/nested_struct.c (cls_struct_combined_gn): Mark
2209         cif and userdata unused.
2210         * testsuite/libffi.call/nested_struct1.c (cls_struct_combined_gn):
2211         Likewise.
2212         * testsuite/libffi.call/nested_struct10.c (B_gn): Likewise.
2213         * testsuite/libffi.call/nested_struct2.c (B_fn): Adjust printf
2214         formatters to silence gcc.
2215         (B_gn): Mark cif and userdata unused.
2216         * testsuite/libffi.call/nested_struct3.c (B_gn): Mark cif and userdata
2217         unused.
2218         * testsuite/libffi.call/nested_struct4.c: Mention related PR.
2219         (B_gn): Mark cif and userdata unused.
2220         * testsuite/libffi.call/nested_struct5.c (B_gn): Mark cif and userdata
2221         unused.
2222         * testsuite/libffi.call/nested_struct6.c: Mention related PR.
2223         (B_gn): Mark cif and userdata unused.
2224         * testsuite/libffi.call/nested_struct7.c (B_gn): Mark cif and userdata
2225         unused.
2226         * testsuite/libffi.call/nested_struct8.c (B_gn): Likewise.
2227         * testsuite/libffi.call/nested_struct9.c (B_gn): Likewise.
2228         * testsuite/libffi.call/problem1.c (stub): Likewise.
2229         * testsuite/libffi.call/pyobjc-tc.c (main): Cast the result to silence
2230         gcc.
2231         * testsuite/libffi.call/return_fl2.c (return_fl): Add the note mentioned
2232         in the last commit for this test case in the test case itself.
2233         * testsuite/libffi.call/closure_fn0.c (closure_test_fn0): Mark cif as
2234         unused.
2235         * testsuite/libffi.call/closure_fn1.c (closure_test_fn1): Likewise.
2236         * testsuite/libffi.call/closure_fn2.c (closure_test_fn2): Likewise.
2237         * testsuite/libffi.call/closure_fn3.c (closure_test_fn3): Likewise.
2238         * testsuite/libffi.call/closure_fn4.c (closure_test_fn0): Likewise.
2239         * testsuite/libffi.call/closure_fn5.c (closure_test_fn5): Likewise.
2240         * testsuite/libffi.call/closure_fn6.c (closure_test_fn0): Likewise.
2241
2242 2006-02-22  Kaz Kojima  <kkojima@gcc.gnu.org>
2243
2244         * src/sh/sysv.S: Fix register numbers in the FDE for
2245         ffi_closure_SYSV.
2246
2247 2006-02-20  Andreas Tobler  <a.tobler@schweiz.ch>
2248
2249         * testsuite/libffi.call/return_fl2.c (return_fl): Remove static
2250         declaration to avoid a false negative on ix86. See PR323.
2251
2252 2006-02-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2253
2254         * src/sh/ffi.c (ffi_closure_helper_SYSV): Remove unused variable
2255         and cast integer to void * if needed.  Update the pointer to
2256         the FP register saved area correctly.
2257
2258 2006-02-17  Andreas Tobler  <a.tobler@schweiz.ch>
2259
2260         * testsuite/libffi.call/nested_struct6.c: XFAIL this test until PR25630
2261         is fixed.
2262         * testsuite/libffi.call/nested_struct4.c: Likewise.
2263
2264 2006-02-16  Andreas Tobler  <a.tobler@schweiz.ch>
2265
2266         * testsuite/libffi.call/return_dbl.c: New test case.
2267         * testsuite/libffi.call/return_dbl1.c: Likewise.
2268         * testsuite/libffi.call/return_dbl2.c: Likewise.
2269         * testsuite/libffi.call/return_fl.c: Likewise.
2270         * testsuite/libffi.call/return_fl1.c: Likewise.
2271         * testsuite/libffi.call/return_fl2.c: Likewise.
2272         * testsuite/libffi.call/return_fl3.c: Likewise.
2273         * testsuite/libffi.call/closure_fn6.c: Likewise.
2274
2275         * testsuite/libffi.call/nested_struct2.c: Remove ffi_type_mylong
2276         definition.
2277         * testsuite/libffi.call/ffitest.h: Add ffi_type_mylong definition
2278         here to be used by other test cases too.
2279
2280         * testsuite/libffi.call/nested_struct10.c: New test case.
2281         * testsuite/libffi.call/nested_struct9.c: Likewise.
2282         * testsuite/libffi.call/nested_struct8.c: Likewise.
2283         * testsuite/libffi.call/nested_struct7.c: Likewise.
2284         * testsuite/libffi.call/nested_struct6.c: Likewise.
2285         * testsuite/libffi.call/nested_struct5.c: Likewise.
2286         * testsuite/libffi.call/nested_struct4.c: Likewise.
2287
2288 2006-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
2289
2290         * configure.ac: Enable libffi for sparc64-*-freebsd*.
2291         * configure: Rebuilt.
2292
2293 2006-01-18  Jakub Jelinek  <jakub@redhat.com>
2294
2295         * src/powerpc/sysv.S (smst_two_register): Don't call __ashldi3,
2296         instead do the shifting inline.
2297         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't compute %r5
2298         shift count unconditionally.  Simplify load sequences for 1, 2, 3, 4
2299         and 8 byte structs, for the remaining struct sizes don't call
2300         __lshrdi3, instead do the shifting inline.
2301
2302 2005-12-07  Thiemo Seufer  <ths@networkno.de>
2303
2304         * src/mips/ffitarget.h: Remove obsolete sgidefs.h include. Add
2305         missing parentheses.
2306         * src/mips/o32.S (ffi_call_O32): Code formatting. Define
2307         and use A3_OFF, FP_OFF, RA_OFF. Micro-optimizations.
2308         (ffi_closure_O32): Likewise, but with newly defined A3_OFF2,
2309         A2_OFF2, A1_OFF2, A0_OFF2, RA_OFF2, FP_OFF2, S0_OFF2, GP_OFF2,
2310         V1_OFF2, V0_OFF2, FA_1_1_OFF2, FA_1_0_OFF2, FA_0_1_OFF2,
2311         FA_0_0_OFF2.
2312         * src/mips/ffi.c (ffi_prep_args): Code formatting. Fix
2313         endianness bugs.
2314         (ffi_prep_closure): Improve trampoline instruction scheduling.
2315         (ffi_closure_mips_inner_O32): Fix endianness bugs.
2316
2317 2005-12-03  Alan Modra  <amodra@bigpond.net.au>
2318
2319         * src/powerpc/ffi.c: Formatting.
2320         (ffi_prep_args_SYSV): Avoid possible aliasing problems by using unions.
2321         (ffi_prep_args64): Likewise.
2322
2323 2005-09-30  Geoffrey Keating  <geoffk@apple.com>
2324
2325         * testsuite/lib/libffi-dg.exp (libffi_target_compile): For
2326         darwin, use -shared-libgcc not -lgcc_s, and explain why.
2327
2328 2005-09-26  Tom Tromey  <tromey@redhat.com>
2329
2330         * testsuite/libffi.call/float1.c (value_type): New typedef.
2331         (CANARY): New define.
2332         (main): Check for result buffer overflow.
2333         * src/powerpc/linux64.S: Handle linux64 long double returns.
2334         * src/powerpc/ffi.c (FLAG_RETURNS_128BITS): New constant.
2335         (ffi_prep_cif_machdep): Handle linux64 long double returns.
2336
2337 2005-08-25  Alan Modra  <amodra@bigpond.net.au>
2338
2339         PR target/23404
2340         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Correct placement of stack
2341         homed fp args.
2342         (ffi_status ffi_prep_cif_machdep): Correct stack sizing for same.
2343
2344 2005-08-11  Jakub Jelinek  <jakub@redhat.com>
2345
2346         * configure.ac (HAVE_HIDDEN_VISIBILITY_ATTRIBUTE): New test.
2347         (AH_BOTTOM): Add FFI_HIDDEN definition.
2348         * configure: Rebuilt.
2349         * fficonfig.h.in: Rebuilt.
2350         * src/powerpc/ffi.c (hidden): Remove.
2351         (ffi_closure_LINUX64, ffi_prep_args64, ffi_call_LINUX64,
2352         ffi_closure_helper_LINUX64): Use FFI_HIDDEN instead of hidden.
2353         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64,
2354         .ffi_closure_LINUX64): Use FFI_HIDDEN instead of .hidden.
2355         * src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV): Remove,
2356         add FFI_HIDDEN to its prototype.
2357         (ffi_closure_SYSV_inner): New.
2358         * src/x86/sysv.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2359         * src/x86/win32.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2360
2361 2005-08-10  Alfred M. Szmidt  <ams@gnu.org>
2362
2363         PR libffi/21819:
2364         * configure: Rebuilt.
2365         * configure.ac: Handle i*86-*-gnu*.
2366
2367 2005-08-09  Jakub Jelinek  <jakub@redhat.com>
2368
2369         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Use
2370         DW_CFA_offset_extended_sf rather than
2371         DW_CFA_GNU_negative_offset_extended.
2372         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise.
2373
2374 2005-07-22  SUGIOKA Toshinobu  <sugioka@itonet.co.jp>
2375
2376         * src/sh/sysv.S (ffi_call_SYSV): Stop argument popping correctly
2377         on sh3.
2378         (ffi_closure_SYSV): Change the stack layout for sh3 struct argument.
2379         * src/sh/ffi.c (ffi_prep_args): Fix sh3 argument copy, when it is
2380         partially on register.
2381         (ffi_closure_helper_SYSV): Likewise.
2382         (ffi_prep_cif_machdep): Don't set too many cif->flags.
2383
2384 2005-07-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2385
2386         * src/sh/ffi.c (ffi_call): Handle small structures correctly.
2387         Remove empty line.
2388         * src/sh64/ffi.c (simple_type): Remove.
2389         (return_type): Handle small structures correctly.
2390         (ffi_prep_args): Likewise.
2391         (ffi_call): Likewise.
2392         (ffi_closure_helper_SYSV): Likewise.
2393         * src/sh64/sysv.S (ffi_call_SYSV): Handle 1, 2 and 4-byte return.
2394         Emit position independent code if PIC and remove wrong datalabel
2395         prefixes from EH data.
2396
2397 2005-07-19  Andreas Tobler  <a.tobler@schweiz.ch>
2398
2399         * Makefile.am (nodist_libffi_la_SOURCES): Add POWERPC_FREEBSD.
2400         * Makefile.in: Regenerate.
2401         * include/Makefile.in: Likewise.
2402         * testsuite/Makefile.in: Likewise.
2403         * configure.ac: Add POWERPC_FREEBSD rules.
2404         * configure: Regenerate.
2405         * src/powerpc/ffitarget.h: Add POWERPC_FREEBSD rules.
2406         (FFI_SYSV_TYPE_SMALL_STRUCT): Define.
2407         * src/powerpc/ffi.c: Add flags to handle small structure returns
2408         in ffi_call_SYSV.
2409         (ffi_prep_cif_machdep): Handle small structures for SYSV 4 ABI.
2410         Aka FFI_SYSV.
2411         (ffi_closure_helper_SYSV): Likewise.
2412         * src/powerpc/ppc_closure.S: Add return types for small structures.
2413         * src/powerpc/sysv.S: Add bits to handle small structures for
2414         final SYSV 4 ABI.
2415
2416 2005-07-10  Andreas Tobler  <a.tobler@schweiz.ch>
2417
2418         * testsuite/libffi.call/cls_5_1_byte.c: New test file.
2419         * testsuite/libffi.call/cls_6_1_byte.c: Likewise.
2420         * testsuite/libffi.call/cls_7_1_byte.c: Likewise.
2421
2422 2005-07-05  Randolph Chung  <tausq@debian.org>
2423
2424         * src/pa/ffi.c (ffi_struct_type): Rename FFI_TYPE_SMALL_STRUCT1
2425         as FFI_TYPE_SMALL_STRUCT3.  Break out handling for 5-7 byte
2426         structures.  Kill compilation warnings.
2427         (ffi_closure_inner_LINUX): Print return values as hex in debug
2428         message.  Rename FFI_TYPE_SMALL_STRUCT1 as FFI_TYPE_SMALL_STRUCT3.
2429         Properly handle 5-7 byte structure returns.
2430         * src/pa/ffitarget.h (FFI_TYPE_SMALL_STRUCT1)
2431         (FFI_TYPE_SMALL_STRUCT2): Remove.
2432         (FFI_TYPE_SMALL_STRUCT3, FFI_TYPE_SMALL_STRUCT5)
2433         (FFI_TYPE_SMALL_STRUCT6, FFI_TYPE_SMALL_STRUCT7): Define.
2434         * src/pa/linux.S: Mark source file as using PA1.1 assembly.
2435         (checksmst1, checksmst2): Remove.
2436         (checksmst3): Optimize handling of 3-byte struct returns.
2437         (checksmst567): Properly handle 5-7 byte struct returns.
2438
2439 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2440
2441         PR libgcj/21943
2442         * src/mips/n32.S: Enforce PIC code.
2443         * src/mips/o32.S: Likewise.
2444
2445 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2446
2447         * configure.ac: Treat i*86-*-solaris2.10 and up as X86_64.
2448         * configure: Regenerate.
2449
2450 2005-06-01  Alan Modra  <amodra@bigpond.net.au>
2451
2452         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't use JUMPTARGET
2453         to call ffi_closure_helper_SYSV.  Append @local instead.
2454         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise for ffi_prep_args_SYSV.
2455
2456 2005-05-17  Kelley Cook  <kcook@gcc.gnu.org>
2457
2458         * configure.ac: Use AC_C_BIGENDIAN instead of AC_C_BIGENDIAN_CROSS.
2459         Use AC_CHECK_SIZEOF instead of AC_COMPILE_CHECK_SIZEOF.
2460         * Makefile.am (ACLOCAL_AMFLAGS): Remove -I ../config.
2461         * aclocal.m4, configure, fficonfig.h.in, Makefile.in,
2462         include/Makefile.in, testsuite/Makefile.in: Regenerate.
2463
2464 2005-05-09  Mike Stump  <mrs@apple.com>
2465
2466         * configure: Regenerate.
2467
2468 2005-05-08  Richard Henderson  <rth@redhat.com>
2469
2470         PR libffi/21285
2471         * src/alpha/osf.S: Update unwind into to match code.
2472
2473 2005-05-04  Andreas Degert <ad@papyrus-gmbh.de>
2474             Richard Henderson  <rth@redhat.com>
2475
2476         * src/x86/ffi64.c (ffi_prep_cif_machdep): Save sse-used flag in
2477         bit 11 of flags.
2478         (ffi_call): Mask return type field.  Pass ssecount to ffi_call_unix64.
2479         (ffi_prep_closure): Set carry bit if sse-used flag set.
2480         * src/x86/unix64.S (ffi_call_unix64): Add ssecount argument.
2481         Only load sse registers if ssecount non-zero.
2482         (ffi_closure_unix64): Only save sse registers if carry set on entry.
2483
2484 2005-04-29  Ralf Corsepius  <ralf.corsepius@rtems.org>
2485
2486         * configure.ac: Add i*86-*-rtems*, sparc*-*-rtems*,
2487         powerpc-*rtems*, arm*-*-rtems*, sh-*-rtems*.
2488         * configure: Regenerate.
2489
2490 2005-04-20  Hans-Peter Nilsson  <hp@axis.com>
2491
2492         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): In regsub use,
2493         have Tcl8.3-compatible intermediate variable.
2494
2495 2005-04-18  Simon Posnjak <simon.posnjak@siol.net>
2496             Hans-Peter Nilsson  <hp@axis.com>
2497
2498         * Makefile.am: Add CRIS support.
2499         * configure.ac: Likewise.
2500         * Makefile.in, configure, testsuite/Makefile.in,
2501         include/Makefile.in: Regenerate.
2502         * src/cris: New directory.
2503         * src/cris/ffi.c, src/cris/sysv.S, src/cris/ffitarget.h: New files.
2504         * src/prep_cif.c (ffi_prep_cif): Wrap in #ifndef __CRIS__.
2505
2506         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): Replace \n with
2507         \r?\n in output tests.
2508
2509 2005-04-12  Mike Stump  <mrs@apple.com>
2510
2511         * configure: Regenerate.
2512
2513 2005-03-30  Hans Boehm  <Hans.Boehm@hp.com>
2514
2515         * src/ia64/ffitarget.h (ffi_arg): Use long long instead of DI.
2516
2517 2005-03-30  Steve Ellcey  <sje@cup.hp.com>
2518
2519         * src/ia64/ffitarget.h (ffi_arg) ADD DI attribute.
2520         (ffi_sarg) Ditto.
2521         * src/ia64/unix.S (ffi_closure_unix): Extend gp
2522         to 64 bits in ILP32 mode.
2523         Load 64 bits even for short data.
2524
2525 2005-03-23  Mike Stump  <mrs@apple.com>
2526
2527         * src/powerpc/darwin.S: Update for -m64 multilib.
2528         * src/powerpc/darwin_closure.S: Likewise.
2529
2530 2005-03-21  Zack Weinberg  <zack@codesourcery.com>
2531
2532         * configure.ac: Do not invoke TL_AC_GCC_VERSION.
2533         Do not set tool_include_dir.
2534         * aclocal.m4, configure, Makefile.in, testsuite/Makefile.in:
2535         Regenerate.
2536         * include/Makefile.am: Set gcc_version and toollibffidir.
2537         * include/Makefile.in: Regenerate.
2538
2539 2005-02-22  Andrew Haley  <aph@redhat.com>
2540
2541         * src/powerpc/ffi.c (ffi_prep_cif_machdep): Bump alignment to
2542         odd-numbered register pairs for 64-bit integer types.
2543
2544 2005-02-23  Andreas Tobler  <a.tobler@schweiz.ch>
2545
2546         PR libffi/20104
2547         * testsuite/libffi.call/return_ll1.c: New test case.
2548
2549 2005-02-11  Janis Johnson  <janis187@us.ibm.com>
2550
2551         * testsuite/libffi.call/cls_align_longdouble.c: Remove dg-options.
2552         * testsuite/libffi.call/float.c: Ditto.
2553         * testsuite/libffi.call/float2.c: Ditto.
2554         * testsuite/libffi.call/float3.c: Ditto.
2555
2556 2005-02-08  Andreas Tobler  <a.tobler@schweiz.ch>
2557
2558         * src/frv/ffitarget.h: Remove PPC stuff which does not belong to frv.
2559
2560 2005-01-12  Eric Botcazou  <ebotcazou@libertysurf.fr>
2561
2562         * testsuite/libffi.special/special.exp (cxx_options): Add
2563         -shared-libgcc.
2564
2565 2004-12-31  Richard Henderson  <rth@redhat.com>
2566
2567         * src/types.c (FFI_AGGREGATE_TYPEDEF): Remove.
2568         (FFI_TYPEDEF): Rename from FFI_INTEGRAL_TYPEDEF.  Replace size and
2569         offset parameters with a type parameter; deduce size and structure
2570         alignment.  Update all users.
2571
2572 2004-12-31  Richard Henderson  <rth@redhat.com>
2573
2574         * src/types.c (FFI_TYPE_POINTER): Define with sizeof.
2575         (FFI_TYPE_LONGDOUBLE): Fix for ia64.
2576         * src/ia64/ffitarget.h (struct ffi_ia64_trampoline_struct): Move
2577         into ffi_prep_closure.
2578         * src/ia64/ia64_flags.h, src/ia64/ffi.c, src/ia64/unix.S: Rewrite
2579         from scratch.
2580
2581 2004-12-27  Richard Henderson  <rth@redhat.com>
2582
2583         * src/x86/unix64.S: Fix typo in unwind info.
2584
2585 2004-12-25  Richard Henderson  <rth@redhat.com>
2586
2587         * src/x86/ffi64.c (struct register_args): Rename from stackLayout.
2588         (enum x86_64_reg_class): Add X86_64_COMPLEX_X87_CLASS.
2589         (merge_classes): Check for it.
2590         (SSE_CLASS_P): New.
2591         (classify_argument): Pass byte_offset by value; perform all updates
2592         inside struct case.
2593         (examine_argument): Add classes argument; handle
2594         X86_64_COMPLEX_X87_CLASS.
2595         (ffi_prep_args): Merge into ...
2596         (ffi_call): ... here.  Share stack frame with ffi_call_unix64.
2597         (ffi_prep_cif_machdep): Setup cif->flags for proper structure return.
2598         (ffi_fill_return_value): Remove.
2599         (ffi_prep_closure): Remove dead assert.
2600         (ffi_closure_unix64_inner): Rename from ffi_closure_UNIX64_inner.
2601         Rewrite to use struct register_args instead of va_list.  Create
2602         flags for handling structure returns.
2603         * src/x86/unix64.S: Remove dead strings.
2604         (ffi_call_unix64): Rename from ffi_call_UNIX64.  Rewrite to share
2605         stack frame with ffi_call.  Handle structure returns properly.
2606         (float2sse, floatfloat2sse, double2sse): Remove.
2607         (sse2float, sse2double, sse2floatfloat): Remove.
2608         (ffi_closure_unix64): Rename from ffi_closure_UNIX64.  Rewrite
2609         to handle structure returns properly.
2610
2611 2004-12-08  David Edelsohn  <edelsohn@gnu.org>
2612
2613         * Makefile.am (AM_MAKEFLAGS): Remove duplicate LIBCFLAGS and
2614         PICFLAG.
2615         * Makefile.in: Regenerated.
2616
2617 2004-12-02  Richard Sandiford  <rsandifo@redhat.com>
2618
2619         * configure.ac: Use TL_AC_GCC_VERSION to set gcc_version.
2620         * configure, aclocal.m4, Makefile.in: Regenerate.
2621         * include/Makefile.in, testsuite/Makefile.in: Regenerate.
2622
2623 2004-11-29  Kelley Cook  <kcook@gcc.gnu.org>
2624
2625         * configure: Regenerate for libtool change.
2626
2627 2004-11-25  Kelley Cook  <kcook@gcc.gnu.org>
2628
2629         * configure: Regenerate for libtool reversion.
2630
2631 2004-11-24  Kelley Cook  <kcook@gcc.gnu.org>
2632
2633         * configure: Regenerate for libtool change.
2634
2635 2004-11-23  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2636
2637         * testsuite/lib/libffi-dg.exp: Use new procs in target-libpath.exp.
2638
2639 2004-11-23  Richard Sandiford  <rsandifo@redhat.com>
2640
2641         * src/mips/o32.S (ffi_call_O32, ffi_closure_O32): Use jalr instead
2642         of jal.  Use an absolute encoding for the frame information.
2643
2644 2004-11-23  Kelley Cook  <kcook@gcc.gnu.org>
2645
2646         * Makefile.am: Remove no-dependencies.  Add ACLOCAL_AMFLAGS.
2647         * acinclude.m4: Delete logic for sincludes.
2648         * aclocal.m4, Makefile.in, configure: Regenerate.
2649         * include/Makefile: Likewise.
2650         * testsuite/Makefile: Likewise.
2651
2652 2004-11-22  Eric Botcazou  <ebotcazou@libertysurf.fr>
2653
2654         * src/sparc/ffi.c (ffi_prep_closure): Align doubles and 64-bit integers
2655         on a 8-byte boundary.
2656         * src/sparc/v8.S (ffi_closure_v8): Reserve frame space for arguments.
2657
2658 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2659
2660         * src/arm/ffi.c (ffi_prep_cif_machdep): Handle functions that return
2661         long long values.  Round stack allocation to a multiple of 8 bytes
2662         for ATPCS compatibility.
2663         * src/arm/sysv.S (ffi_call_SYSV): Rework to avoid use of APCS register
2664         names.  Handle returning long long types.  Add Thumb and interworking
2665         support.  Improve soft-float code.
2666
2667 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2668
2669         * testsuite/lib/libffi-db.exp (load_gcc_lib): New function.
2670         (libffi_exit): New function.
2671         (libffi_init): Build the testglue wrapper if needed.
2672
2673 2004-10-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2674
2675         PR other/18138
2676         * testsuite/lib/libffi-dg.exp: Accept more than one multilib libgcc.
2677
2678 2004-10-25  Kazuhiro Inaoka <inaoka.kazuhiro@renesas.com>
2679
2680         * src/m32r/libffitarget.h (FFI_CLOSURES): Set to 0.
2681
2682 2004-10-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2683
2684         * src/sh/sysv.S (ffi_call_SYSV): Don't align for double data.
2685         * testsuite/libffi.call/float3.c: New test case.
2686
2687 2004-10-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2688
2689         * src/sh/ffi.c (ffi_prep_closure): Set T bit in trampoline for
2690         the function returning a structure pointed with R2.
2691         * src/sh/sysv.S (ffi_closure_SYSV): Use R2 as the pointer to
2692         the structure return value if T bit set.  Emit position
2693         independent code and EH data if PIC.
2694
2695 2004-10-13  Kazuhiro Inaoka  <inaoka.kazuhiro@renesas.com>
2696
2697         * Makefile.am: Add m32r support.
2698         * configure.ac: Likewise.
2699         * Makefile.in: Regenerate.
2700         * confiugre: Regenerate.
2701         * src/types.c: Add m32r port to FFI_INTERNAL_TYPEDEF
2702         (uint64, sint64, double, longdouble)
2703         * src/m32r: New directory.
2704         * src/m32r/ffi.c: New file.
2705         * src/m32r/sysv.S: Likewise.
2706         * src/m32r/ffitarget.h: Likewise.
2707
2708 2004-10-02  Kaz Kojima  <kkojima@gcc.gnu.org>
2709
2710         * testsuite/libffi.call/negint.c: New test case.
2711
2712 2004-09-14  H.J. Lu  <hongjiu.lu@intel.com>
2713
2714         PR libgcj/17465
2715         * testsuite/lib/libffi-dg.exp: Don't use global ld_library_path.
2716         Set up LD_LIBRARY_PATH, SHLIB_PATH, LD_LIBRARYN32_PATH,
2717         LD_LIBRARY64_PATH, LD_LIBRARY_PATH_32, LD_LIBRARY_PATH_64 and
2718         DYLD_LIBRARY_PATH.
2719
2720 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2721
2722         * testsuite/libffi.call/many_win32.c: Remove whitespaces.
2723         * testsuite/libffi.call/promotion.c: Likewise.
2724         * testsuite/libffi.call/return_ll.c: Remove unused var. Cleanup
2725         whitespaces.
2726         * testsuite/libffi.call/return_sc.c: Likewise.
2727         * testsuite/libffi.call/return_uc.c: Likewise.
2728
2729 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2730
2731         * src/powerpc/darwin.S: Fix comments and identation.
2732         * src/powerpc/darwin_closure.S: Likewise.
2733
2734 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2735
2736         * src/powerpc/ffi_darwin.c: Add flag for longdouble return values.
2737         (ffi_prep_args): Handle longdouble arguments.
2738         (ffi_prep_cif_machdep): Set flags for longdouble. Calculate space for
2739         longdouble.
2740         (ffi_closure_helper_DARWIN): Add closure handling for longdouble.
2741         * src/powerpc/darwin.S (_ffi_call_DARWIN): Add handling of longdouble
2742         values.
2743         * src/powerpc/darwin_closure.S (_ffi_closure_ASM): Likewise.
2744         * src/types.c: Defined longdouble size and alignment for darwin.
2745
2746 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2747
2748         * src/powerpc/aix.S: Remove whitespaces.
2749         * src/powerpc/aix_closure.S: Likewise.
2750         * src/powerpc/asm.h: Likewise.
2751         * src/powerpc/ffi.c: Likewise.
2752         * src/powerpc/ffitarget.h: Likewise.
2753         * src/powerpc/linux64.S: Likewise.
2754         * src/powerpc/linux64_closure.S: Likewise.
2755         * src/powerpc/ppc_closure.S: Likewise.
2756         * src/powerpc/sysv.S: Likewise.
2757
2758 2004-08-30  Anthony Green  <green@redhat.com>
2759
2760         * Makefile.am: Add frv support.
2761         * Makefile.in, testsuite/Makefile.in: Rebuilt.
2762         * configure.ac: Read configure.host.
2763         * configure.in: Read configure.host.
2764         * configure.host: New file.  frv-elf needs libgloss.
2765         * include/ffi.h.in: Force ffi_closure to have a nice big (8)
2766         alignment.  This is needed to frv and shouldn't harm the others.
2767         * include/ffi_common.h (ALIGN_DOWN): New macro.
2768         * src/frv/ffi.c, src/frv/ffitarget.h, src/frv/eabi.S: New files.
2769
2770 2004-08-24  David Daney  <daney@avtrex.com>
2771
2772         * testsuite/libffi.call/closure_fn0.c: Xfail mips64* instead of mips*.
2773         * testsuite/libffi.call/closure_fn1.c: Likewise.
2774         * testsuite/libffi.call/closure_fn2.c  Likewise.
2775         * testsuite/libffi.call/closure_fn3.c: Likewise.
2776         * testsuite/libffi.call/closure_fn4.c: Likewise.
2777         * testsuite/libffi.call/closure_fn5.c: Likewise.
2778         * testsuite/libffi.call/cls_18byte.c: Likewise.
2779         * testsuite/libffi.call/cls_19byte.c: Likewise.
2780         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2781         * testsuite/libffi.call/cls_20byte.c: Likewise.
2782         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2783         * testsuite/libffi.call/cls_24byte.c: Likewise.
2784         * testsuite/libffi.call/cls_2byte.c: Likewise.
2785         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2786         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2787         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2788         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2789         * testsuite/libffi.call/cls_4byte.c: Likewise.
2790         * testsuite/libffi.call/cls_64byte.c: Likewise.
2791         * testsuite/libffi.call/cls_6byte.c: Likewise.
2792         * testsuite/libffi.call/cls_7byte.c: Likewise.
2793         * testsuite/libffi.call/cls_8byte.c: Likewise.
2794         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2795         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2796         * testsuite/libffi.call/cls_align_double.c: Likewise.
2797         * testsuite/libffi.call/cls_align_float.c: Likewise.
2798         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2799         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2800         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2801         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2802         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2803         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2804         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2805         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
2806         * testsuite/libffi.call/cls_double.c: Likewise.
2807         * testsuite/libffi.call/cls_float.c: Likewise.
2808         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2809         * testsuite/libffi.call/cls_multi_sshort.c: Likewise.
2810         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2811         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2812         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2813         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2814         * testsuite/libffi.call/cls_schar.c: Likewise.
2815         * testsuite/libffi.call/cls_sint.c: Likewise.
2816         * testsuite/libffi.call/cls_sshort.c: Likewise.
2817         * testsuite/libffi.call/cls_uchar.c: Likewise.
2818         * testsuite/libffi.call/cls_uint.c: Likewise.
2819         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2820         * testsuite/libffi.call/cls_ushort.c: Likewise.
2821         * testsuite/libffi.call/nested_struct.c: Likewise.
2822         * testsuite/libffi.call/nested_struct1.c: Likewise.
2823         * testsuite/libffi.call/nested_struct2.c: Likewise.
2824         * testsuite/libffi.call/nested_struct3.c: Likewise.
2825         * testsuite/libffi.call/problem1.c: Likewise.
2826         * testsuite/libffi.special/unwindtest.cc: Likewise.
2827         * testsuite/libffi.call/cls_12byte.c: Likewise and set return value
2828         to zero.
2829         * testsuite/libffi.call/cls_16byte.c: Likewise.
2830         * testsuite/libffi.call/cls_5byte.c: Likewise.
2831
2832 2004-08-23  David Daney <daney@avtrex.com>
2833
2834         PR libgcj/13141
2835         * src/mips/ffitarget.h (FFI_O32_SOFT_FLOAT): New ABI.
2836         * src/mips/ffi.c (ffi_prep_args): Fix alignment calculation.
2837         (ffi_prep_cif_machdep): Handle FFI_O32_SOFT_FLOAT floating point
2838         parameters and return types.
2839         (ffi_call): Handle FFI_O32_SOFT_FLOAT ABI.
2840         (ffi_prep_closure): Ditto.
2841         (ffi_closure_mips_inner_O32): Handle FFI_O32_SOFT_FLOAT ABI, fix
2842         alignment calculations.
2843         * src/mips/o32.S (ffi_closure_O32): Don't use floating point
2844         instructions if FFI_O32_SOFT_FLOAT, make stack frame ABI compliant.
2845
2846 2004-08-14  Casey Marshall <csm@gnu.org>
2847
2848         * src/mips/ffi.c (ffi_pref_cif_machdep): set `cif->flags' to
2849         contain `FFI_TYPE_UINT64' as return type for any 64-bit
2850         integer (O32 ABI only).
2851         (ffi_prep_closure): new function.
2852         (ffi_closure_mips_inner_O32): new function.
2853         * src/mips/ffitarget.h: Define `FFI_CLOSURES' and
2854         `FFI_TRAMPOLINE_SIZE' appropriately if the ABI is o32.
2855         * src/mips/o32.S (ffi_call_O32): add labels for .eh_frame. Return
2856         64 bit integers correctly.
2857         (ffi_closure_O32): new function.
2858         Added DWARF-2 unwind info for both functions.
2859
2860 2004-08-10  Andrew Haley  <aph@redhat.com>
2861
2862         * src/x86/ffi64.c (ffi_prep_args ): 8-align all stack arguments.
2863
2864 2004-08-01  Robert Millan  <robertmh@gnu.org>
2865
2866         * configure.ac: Detect knetbsd-gnu and kfreebsd-gnu.
2867         * configure: Regenerate.
2868
2869 2004-07-30  Maciej W. Rozycki  <macro@linux-mips.org>
2870
2871         * acinclude.m4 (AC_FUNC_MMAP_BLACKLIST): Check for <sys/mman.h>
2872         and mmap() explicitly instead of relying on preset autoconf cache
2873         variables.
2874         * aclocal.m4: Regenerate.
2875         * configure: Regenerate.
2876
2877 2004-07-11  Ulrich Weigand  <uweigand@de.ibm.com>
2878
2879         * src/s390/ffi.c (ffi_prep_args): Fix C aliasing violation.
2880         (ffi_check_float_struct): Remove unused prototype.
2881
2882 2004-06-30  Geoffrey Keating  <geoffk@apple.com>
2883
2884         * src/powerpc/ffi_darwin.c (flush_icache): ';' is a comment
2885         character on Darwin, use '\n\t' instead.
2886
2887 2004-06-26  Matthias Klose  <doko@debian.org>
2888
2889         * libtool-version: Fix typo in revision/age.
2890
2891 2004-06-17  Matthias Klose  <doko@debian.org>
2892
2893         * libtool-version: New.
2894         * Makefile.am (libffi_la_LDFLAGS): Use -version-info for soname.
2895         * Makefile.in: Regenerate.
2896
2897 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2898
2899         * Makefile.am: Remove useless multilib rules.
2900         * Makefile.in: Regenerate.
2901         * aclocal.m4: Regenerate with automake 1.8.5.
2902         * configure.ac: Remove useless multilib configury.
2903         * configure: Regenerate.
2904
2905 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2906
2907         * .cvsignore: New file.
2908
2909 2004-06-10  Jakub Jelinek  <jakub@redhat.com>
2910
2911         * src/ia64/unix.S (ffi_call_unix): Insert group barrier break
2912         fp_done.
2913         (ffi_closure_UNIX): Fix f14/f15 adjustment if FLOAT_SZ is ever
2914         changed from 8.
2915
2916 2004-06-06  Sean McNeil  <sean@mcneil.com>
2917
2918         * configure.ac: Add x86_64-*-freebsd* support.
2919         * configure: Regenerate.
2920
2921 2004-04-26  Joe Buck <jbuck@welsh-buck.org>
2922
2923         Bug 15093
2924         * configure.ac: Test for existence of mmap and sys/mman.h before
2925         checking blacklist.  Fix suggested by Jim Wilson.
2926         * configure: Regenerate.
2927
2928 2004-04-26  Matt Austern  <austern@apple.com>
2929
2930         * src/powerpc/darwin.S: Go through a non-lazy pointer for initial
2931         FDE location.
2932         * src/powerpc/darwin_closure.S: Likewise.
2933
2934 2004-04-24  Andreas Tobler  <a.tobler@schweiz.ch>
2935
2936         * testsuite/libffi.call/cls_multi_schar.c (main): Fix initialization
2937         error. Reported by Thomas Heller <theller@python.net>.
2938         * testsuite/libffi.call/cls_multi_sshort.c (main): Likewise.
2939         * testsuite/libffi.call/cls_multi_ushort.c (main): Likewise.
2940
2941 2004-03-20  Matthias Klose  <doko@debian.org>
2942
2943         * src/pa/linux.S: Fix typo.
2944
2945 2004-03-19  Matthias Klose  <doko@debian.org>
2946
2947         * Makefile.am: Update.
2948         * Makefile.in: Regenerate.
2949         * src/pa/ffi.h.in: Remove.
2950         * src/pa/ffitarget.h: New file.
2951
2952 2004-02-10  Randolph Chung  <tausq@debian.org>
2953
2954         * Makefile.am: Add PA support.
2955         * Makefile.in: Regenerate.
2956         * include/Makefile.in: Regenerate.
2957         * configure.ac: Add PA target.
2958         * configure: Regenerate.
2959         * src/pa/ffi.c: New file.
2960         * src/pa/ffi.h.in: Add PA support.
2961         * src/pa/linux.S: New file.
2962         * prep_cif.c: Add PA support.
2963
2964 2004-03-16  Hosaka Yuji  <hos@tamanegi.org>
2965
2966         * src/types.c: Fix alignment size of X86_WIN32 case int64 and
2967         double.
2968         * src/x86/ffi.c (ffi_prep_args): Replace ecif->cif->rtype->type
2969         with ecif->cif->flags.
2970         (ffi_call, ffi_prep_incoming_args_SYSV): Replace cif->rtype->type
2971         with cif->flags.
2972         (ffi_prep_cif_machdep): Add X86_WIN32 struct case.
2973         (ffi_closure_SYSV): Add 1 or 2-bytes struct case for X86_WIN32.
2974         * src/x86/win32.S (retstruct1b, retstruct2b, sc_retstruct1b,
2975         sc_retstruct2b): Add for 1 or 2-bytes struct case.
2976
2977 2004-03-15 Kelley Cook <kcook@gcc.gnu.org>
2978
2979         * configure.in: Rename file to ...
2980         * configure.ac: ... this.
2981         * fficonfig.h.in: Regenerate.
2982         * Makefile.in: Regenerate.
2983         * include/Makefile.in: Regenerate.
2984         * testsuite/Makefile.in: Regenerate.
2985
2986 2004-03-12  Matt Austern  <austern@apple.com>
2987
2988         * src/powerpc/darwin.S: Fix EH information so it corresponds to
2989         changes in EH format resulting from addition of linkonce support.
2990         * src/powerpc/darwin_closure.S: Likewise.
2991
2992 2004-03-11  Andreas Tobler  <a.tobler@schweiz.ch>
2993             Paolo Bonzini  <bonzini@gnu.org>
2994
2995         * Makefile.am (AUTOMAKE_OPTIONS): Set them.
2996         Remove VPATH. Remove rules for object files. Remove multilib support.
2997         (AM_CCASFLAGS): Add.
2998         * configure.in (AC_CONFIG_HEADERS): Relace AM_CONFIG_HEADER.
2999         (AC_PREREQ): Bump version to 2.59.
3000         (AC_INIT): Fill with version info and bug address.
3001         (ORIGINAL_LD_FOR_MULTILIBS): Remove.
3002         (AM_ENABLE_MULTILIB): Use this instead of AC_ARG_ENABLE.
3003         De-precious CC so that the right flags are passed down to multilibs.
3004         (AC_MSG_ERROR): Replace obsolete macro AC_ERROR.
3005         (AC_CONFIG_FILES): Replace obsolete macro AC_LINK_FILES.
3006         (AC_OUTPUT): Reorganize the output with AC_CONFIG_COMMANDS.
3007         * configure: Rebuilt.
3008         * aclocal.m4: Likewise.
3009         * Makefile.in, include/Makefile.in, testsuite/Makefile.in: Likewise.
3010         * fficonfig.h.in: Likewise.
3011
3012 2004-03-11  Andreas Schwab  <schwab@suse.de>
3013
3014         * src/ia64/ffi.c (ffi_prep_incoming_args_UNIX): Get floating point
3015         arguments from fp registers only for the first 8 parameter slots.
3016         Don't convert a float parameter when passed in memory.
3017
3018 2004-03-09  Hans-Peter Nilsson  <hp@axis.com>
3019
3020         * configure: Regenerate for config/accross.m4 correction.
3021
3022 2004-02-25  Matt Kraai  <kraai@alumni.cmu.edu>
3023
3024         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Change
3025         ecif->cif->bytes to bytes.
3026         (ffi_prep_cif_machdep): Add braces around nested if statement.
3027
3028 2004-02-09  Alan Modra  <amodra@bigpond.net.au>
3029
3030         * src/types.c (pointer): POWERPC64 has 8 byte pointers.
3031
3032         * src/powerpc/ffi.c (ffi_prep_args64): Correct long double handling.
3033         (ffi_closure_helper_LINUX64): Fix typo.
3034         * testsuite/libffi.call/cls_align_longdouble.c: Pass -mlong-double-128
3035         for powerpc64-*-*.
3036         * testsuite/libffi.call/float.c: Likewise.
3037         * testsuite/libffi.call/float2.c: Likewise.
3038
3039 2004-02-08  Alan Modra  <amodra@bigpond.net.au>
3040
3041         * src/powerpc/ffi.c (ffi_prep_cif_machdep <FFI_LINUX64>): Correct
3042         long double function return and long double arg handling.
3043         (ffi_closure_helper_LINUX64): Formatting.  Delete unused "ng" var.
3044         Use "end_pfr" instead of "nf".  Correct long double handling.
3045         Localise "temp".
3046         * src/powerpc/linux64.S (ffi_call_LINUX64): Save f2 long double
3047         return value.
3048         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Allocate
3049         space for long double return value.  Adjust stack frame and offsets.
3050         Load f2 long double return.
3051
3052 2004-02-07  Alan Modra  <amodra@bigpond.net.au>
3053
3054         * src/types.c: Use 16 byte long double for POWERPC64.
3055
3056 2004-01-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
3057
3058         * src/sparc/ffi.c (ffi_prep_args_v9): Shift the parameter array
3059         when the structure return address is passed in %o0.
3060         (ffi_V9_return_struct): Rename into ffi_v9_layout_struct.
3061         (ffi_v9_layout_struct): Align the field following a nested structure
3062         on a word boundary.  Use memmove instead of memcpy.
3063         (ffi_call): Update call to ffi_V9_return_struct.
3064         (ffi_prep_closure): Define 'ctx' only for V8.
3065         (ffi_closure_sparc_inner): Clone into ffi_closure_sparc_inner_v8
3066         and ffi_closure_sparc_inner_v9.
3067         (ffi_closure_sparc_inner_v8): Return long doubles by reference.
3068         Always skip the structure return address.  For structures and long
3069         doubles, copy the argument directly.
3070         (ffi_closure_sparc_inner_v9): Skip the structure return address only
3071         if required.  Shift the maximum floating-point slot accordingly.  For
3072         big structures, copy the argument directly; otherwise, left-justify the
3073         argument and call ffi_v9_layout_struct to lay out the structure on
3074         the stack.
3075         * src/sparc/v8.S: Undef STACKFRAME before defining it.
3076         (ffi_closure_v8): Pass the structure return address.  Update call to
3077         ffi_closure_sparc_inner_v8.  Short-circuit FFI_TYPE_INT handling.
3078         Skip the 'unimp' insn when returning long doubles and structures.
3079         * src/sparc/v9.S: Undef STACKFRAME before defining it.
3080         (ffi_closure_v9): Increase the frame size by 2 words.  Short-circuit
3081         FFI_TYPE_INT handling.  Load structures both in integers and
3082         floating-point registers on return.
3083         * README: Update status of the SPARC port.
3084
3085 2004-01-24  Andreas Tobler  <a.tobler@schweiz.ch>
3086
3087         * testsuite/libffi.call/pyobjc-tc.c (main): Treat result value
3088         as of type ffi_arg.
3089         * testsuite/libffi.call/struct3.c (main): Fix CHECK.
3090
3091 2004-01-22  Ulrich Weigand  <uweigand@de.ibm.com>
3092
3093         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Treat result
3094         value as of type ffi_arg, not unsigned int.
3095
3096 2004-01-21  Michael Ritzert  <ritzert@t-online.de>
3097
3098         * ffi64.c (ffi_prep_args): Cast the RHS of an assignment instead
3099         of the LHS.
3100
3101 2004-01-12  Andreas Tobler  <a.tobler@schweiz.ch>
3102
3103         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_32 for
3104         Solaris.
3105
3106 2004-01-08  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3107
3108         * testsuite/libffi.call/ffitest.h (allocate_mmap): Cast MAP_FAILED
3109         to void *.
3110
3111 2003-12-10  Richard Henderson  <rth@redhat.com>
3112
3113         * testsuite/libffi.call/cls_align_pointer.c: Cast pointers to
3114         size_t instead of int.
3115
3116 2003-12-04  Hosaka Yuji  <hos@tamanegi.org>
3117
3118         * testsuite/libffi.call/many_win32.c: Include <float.h>.
3119         * testsuite/libffi.call/many_win32.c (main): Replace variable
3120         int i with unsigned long ul.
3121
3122         * testsuite/libffi.call/cls_align_uint64.c: New test case.
3123         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
3124         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
3125         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
3126         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
3127         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
3128         * testsuite/libffi.call/cls_align_float.c: Likewise.
3129         * testsuite/libffi.call/cls_align_double.c: Likewise.
3130         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
3131         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
3132
3133 2003-12-02  Hosaka Yuji  <hos@tamanegi.org>
3134
3135         PR other/13221
3136         * src/x86/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV):
3137         Align arguments to 32 bits.
3138
3139 2003-12-01  Andreas Tobler  <a.tobler@schweiz.ch>
3140
3141         PR other/13221
3142         * testsuite/libffi.call/cls_multi_sshort.c: New test case.
3143         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
3144         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
3145         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
3146         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
3147         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
3148
3149         * testsuite/libffi.special/unwindtest.cc: Cosmetics.
3150
3151 2003-11-26  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
3152
3153         * testsuite/libffi.call/ffitest.h: Include <fcntl.h>.
3154         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3155
3156 2003-11-22  Andreas Tobler  <a.tobler@schweiz.ch>
3157
3158         * Makefile.in: Rebuilt.
3159         * configure: Likewise.
3160         * testsuite/libffi.special/unwindtest.cc: Convert the mmap to
3161         the right type.
3162
3163 2003-11-21  Andreas Jaeger  <aj@suse.de>
3164             Andreas Tobler  <a.tobler@schweiz.ch>
3165
3166         * acinclude.m4: Add AC_FUNC_MMAP_BLACKLIST.
3167         * configure.in: Call AC_FUNC_MMAP_BLACKLIST.
3168         * Makefile.in: Rebuilt.
3169         * aclocal.m4: Likewise.
3170         * configure: Likewise.
3171         * fficonfig.h.in: Likewise.
3172         * testsuite/lib/libffi-dg.exp: Add include dir.
3173         * testsuite/libffi.call/ffitest.h: Add MMAP definitions.
3174         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3175         * testsuite/libffi.call/closure_fn0.c: Use MMAP functionality
3176         for ffi_closure if available.
3177         * testsuite/libffi.call/closure_fn1.c: Likewise.
3178         * testsuite/libffi.call/closure_fn2.c: Likewise.
3179         * testsuite/libffi.call/closure_fn3.c: Likewise.
3180         * testsuite/libffi.call/closure_fn4.c: Likewise.
3181         * testsuite/libffi.call/closure_fn5.c: Likewise.
3182         * testsuite/libffi.call/cls_12byte.c: Likewise.
3183         * testsuite/libffi.call/cls_16byte.c: Likewise.
3184         * testsuite/libffi.call/cls_18byte.c: Likewise.
3185         * testsuite/libffi.call/cls_19byte.c: Likewise.
3186         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3187         * testsuite/libffi.call/cls_20byte.c: Likewise.
3188         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3189         * testsuite/libffi.call/cls_24byte.c: Likewise.
3190         * testsuite/libffi.call/cls_2byte.c: Likewise.
3191         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3192         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3193         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3194         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3195         * testsuite/libffi.call/cls_4byte.c: Likewise.
3196         * testsuite/libffi.call/cls_5byte.c: Likewise.
3197         * testsuite/libffi.call/cls_64byte.c: Likewise.
3198         * testsuite/libffi.call/cls_6byte.c: Likewise.
3199         * testsuite/libffi.call/cls_7byte.c: Likewise.
3200         * testsuite/libffi.call/cls_8byte.c: Likewise.
3201         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3202         * testsuite/libffi.call/cls_9byte2.c: Likewise.
3203         * testsuite/libffi.call/cls_double.c: Likewise.
3204         * testsuite/libffi.call/cls_float.c: Likewise.
3205         * testsuite/libffi.call/cls_schar.c: Likewise.
3206         * testsuite/libffi.call/cls_sint.c: Likewise.
3207         * testsuite/libffi.call/cls_sshort.c: Likewise.
3208         * testsuite/libffi.call/cls_uchar.c: Likewise.
3209         * testsuite/libffi.call/cls_uint.c: Likewise.
3210         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3211         * testsuite/libffi.call/cls_ushort.c: Likewise.
3212         * testsuite/libffi.call/nested_struct.c: Likewise.
3213         * testsuite/libffi.call/nested_struct1.c: Likewise.
3214         * testsuite/libffi.call/nested_struct2.c: Likewise.
3215         * testsuite/libffi.call/nested_struct3.c: Likewise.
3216         * testsuite/libffi.call/problem1.c: Likewise.
3217         * testsuite/libffi.special/unwindtest.cc: Likewise.
3218
3219 2003-11-20  Andreas Tobler  <a.tobler@schweiz.ch>
3220
3221         * testsuite/lib/libffi-dg.exp: Make the -lgcc_s conditional.
3222
3223 2003-11-19  Andreas Tobler  <a.tobler@schweiz.ch>
3224
3225         * testsuite/lib/libffi-dg.exp: Add DYLD_LIBRARY_PATH for darwin.
3226         Add -lgcc_s to additional flags.
3227
3228 2003-11-12  Andreas Tobler  <a.tobler@schweiz.ch>
3229
3230         * configure.in, include/Makefile.am: PR libgcj/11147, install
3231         the ffitarget.h header file in a gcc versioned and target
3232         dependent place.
3233         * configure: Regenerated.
3234         * Makefile.in, include/Makefile.in: Likewise.
3235         * testsuite/Makefile.in: Likewise.
3236
3237 2003-11-09  Andreas Tobler  <a.tobler@schweiz.ch>
3238
3239         * testsuite/libffi.call/closure_fn0.c: Print result and check
3240         with dg-output to make debugging easier.
3241         * testsuite/libffi.call/closure_fn1.c: Likewise.
3242         * testsuite/libffi.call/closure_fn2.c: Likewise.
3243         * testsuite/libffi.call/closure_fn3.c: Likewise.
3244         * testsuite/libffi.call/closure_fn4.c: Likewise.
3245         * testsuite/libffi.call/closure_fn5.c: Likewise.
3246         * testsuite/libffi.call/cls_12byte.c: Likewise.
3247         * testsuite/libffi.call/cls_16byte.c: Likewise.
3248         * testsuite/libffi.call/cls_18byte.c: Likewise.
3249         * testsuite/libffi.call/cls_19byte.c: Likewise.
3250         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3251         * testsuite/libffi.call/cls_20byte.c: Likewise.
3252         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3253         * testsuite/libffi.call/cls_24byte.c: Likewise.
3254         * testsuite/libffi.call/cls_2byte.c: Likewise.
3255         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3256         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3257         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3258         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3259         * testsuite/libffi.call/cls_4byte.c: Likewise.
3260         * testsuite/libffi.call/cls_5byte.c: Likewise.
3261         * testsuite/libffi.call/cls_64byte.c: Likewise.
3262         * testsuite/libffi.call/cls_6byte.c: Likewise.
3263         * testsuite/libffi.call/cls_7byte.c: Likewise.
3264         * testsuite/libffi.call/cls_8byte.c: Likewise.
3265         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3266         * testsuite/libffi.call/cls_9byte2.c: Likewise.
3267         * testsuite/libffi.call/cls_double.c: Likewise.
3268         * testsuite/libffi.call/cls_float.c: Likewise.
3269         * testsuite/libffi.call/cls_schar.c: Likewise.
3270         * testsuite/libffi.call/cls_sint.c: Likewise.
3271         * testsuite/libffi.call/cls_sshort.c: Likewise.
3272         * testsuite/libffi.call/cls_uchar.c: Likewise.
3273         * testsuite/libffi.call/cls_uint.c: Likewise.
3274         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3275         * testsuite/libffi.call/cls_ushort.c: Likewise.
3276         * testsuite/libffi.call/problem1.c: Likewise.
3277
3278         * testsuite/libffi.special/unwindtest.cc: Make ffi_closure
3279         static.
3280
3281 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
3282
3283         * testsuite/libffi.call/cls_9byte2.c: New test case.
3284         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3285         * testsuite/libffi.call/cls_64byte.c: Likewise.
3286         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3287         * testsuite/libffi.call/cls_19byte.c: Likewise.
3288         * testsuite/libffi.call/cls_18byte.c: Likewise.
3289         * testsuite/libffi.call/closure_fn4.c: Likewise.
3290         * testsuite/libffi.call/closure_fn5.c: Likewise.
3291         * testsuite/libffi.call/cls_schar.c: Likewise.
3292         * testsuite/libffi.call/cls_sint.c: Likewise.
3293         * testsuite/libffi.call/cls_sshort.c: Likewise.
3294         * testsuite/libffi.call/nested_struct2.c: Likewise.
3295         * testsuite/libffi.call/nested_struct3.c: Likewise.
3296
3297 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
3298
3299         * testsuite/libffi.call/cls_double.c: Do a check on the result.
3300         * testsuite/libffi.call/cls_uchar.c: Likewise.
3301         * testsuite/libffi.call/cls_uint.c: Likewise.
3302         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3303         * testsuite/libffi.call/cls_ushort.c: Likewise.
3304         * testsuite/libffi.call/return_sc.c: Cleanup whitespaces.
3305
3306 2003-11-06  Andreas Tobler  <a.tobler@schweiz.ch>
3307
3308         * src/prep_cif.c (ffi_prep_cif): Move the validity check after
3309         the initialization.
3310
3311 2003-10-23  Andreas Tobler  <a.tobler@schweiz.ch>
3312
3313         * src/java_raw_api.c (ffi_java_ptrarray_to_raw): Replace
3314         FFI_ASSERT(FALSE) with FFI_ASSERT(0).
3315
3316 2003-10-22  David Daney  <ddaney@avtrex.com>
3317
3318         * src/mips/ffitarget.h: Replace undefined UINT32 and friends with
3319         __attribute__((__mode__(__SI__))) and friends.
3320
3321 2003-10-22  Andreas Schwab  <schwab@suse.de>
3322
3323         * src/ia64/ffi.c: Replace FALSE/TRUE with false/true.
3324
3325 2003-10-21  Andreas Tobler  <a.tobler@schweiz.ch>
3326
3327         * configure.in: AC_LINK_FILES(ffitarget.h).
3328         * configure: Regenerate.
3329         * Makefile.in: Likewise.
3330         * include/Makefile.in: Likewise.
3331         * testsuite/Makefile.in: Likewise.
3332         * fficonfig.h.in: Likewise.
3333
3334 2003-10-21  Paolo Bonzini  <bonzini@gnu.org>
3335             Richard Henderson  <rth@redhat.com>
3336
3337         Avoid that ffi.h includes fficonfig.h.
3338
3339         * Makefile.am (EXTRA_DIST): Include ffitarget.h files
3340         (TARGET_SRC_MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3341         (TARGET_SRC_MIPS_SGI): Removed.
3342         (MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3343         (MIPS_SGI): Removed.
3344         (CLEANFILES): Removed.
3345         (mostlyclean-am, clean-am, mostlyclean-sub, clean-sub): New
3346         targets.
3347         * acconfig.h: Removed.
3348         * configure.in: Compute sizeofs only for double and long double.
3349         Use them to define and subst HAVE_LONG_DOUBLE.  Include comments
3350         into AC_DEFINE instead of using acconfig.h.  Create
3351         include/ffitarget.h instead of include/fficonfig.h.  Rename
3352         MIPS_GCC to MIPS_IRIX, drop MIPS_SGI since we are in gcc's tree.
3353         AC_DEFINE EH_FRAME_FLAGS.
3354         * include/Makefile.am (DISTCLEANFILES): New automake macro.
3355         (hack_DATA): Add ffitarget.h.
3356         * include/ffi.h.in: Remove all system specific definitions.
3357         Declare raw API even if it is not installed, why bother?
3358         Use limits.h instead of SIZEOF_* to define ffi_type_*.  Do
3359         not define EH_FRAME_FLAGS, it is in fficonfig.h now.  Include
3360         ffitarget.h instead of fficonfig.h.  Remove ALIGN macro.
3361         (UINT_ARG, INT_ARG): Removed, use ffi_arg and ffi_sarg instead.
3362         * include/ffi_common.h (bool): Do not define.
3363         (ffi_assert): Accept failed assertion.
3364         (ffi_type_test): Return void and accept file/line.
3365         (FFI_ASSERT): Pass stringized failed assertion.
3366         (FFI_ASSERT_AT): New macro.
3367         (FFI_ASSERT_VALID_TYPE): New macro.
3368         (UINT8, SINT8, UINT16, SINT16, UINT32, SINT32,
3369         UINT64, SINT64): Define here with gcc's __attribute__ macro
3370         instead of in ffi.h
3371         (FLOAT32, ALIGN): Define here instead of in ffi.h
3372         * include/ffi-mips.h: Removed.  Its content moved to
3373         src/mips/ffitarget.h after separating assembly and C sections.
3374         * src/alpha/ffi.c, src/alpha/ffi.c, src/java_raw_api.c
3375         src/prep_cif.c, src/raw_api.c, src/ia64/ffi.c,
3376         src/mips/ffi.c, src/mips/n32.S, src/mips/o32.S,
3377         src/mips/ffitarget.h, src/sparc/ffi.c, src/x86/ffi64.c:
3378         SIZEOF_ARG -> FFI_SIZEOF_ARG.
3379         * src/ia64/ffi.c: Include stdbool.h (provided by GCC 2.95+).
3380         * src/debug.c (ffi_assert): Accept stringized failed assertion.
3381         (ffi_type_test): Rewritten.
3382         * src/prep-cif.c (initialize_aggregate, ffi_prep_cif): Call
3383         FFI_ASSERT_VALID_TYPE.
3384         * src/alpha/ffitarget.h, src/arm/ffitarget.h,
3385         src/ia64/ffitarget.h, src/m68k/ffitarget.h,
3386         src/mips/ffitarget.h, src/powerpc/ffitarget.h,
3387         src/s390/ffitarget.h, src/sh/ffitarget.h,
3388         src/sh64/ffitarget.h, src/sparc/ffitarget.h,
3389         src/x86/ffitarget.h: New files.
3390         * src/alpha/osf.S, src/arm/sysv.S, src/ia64/unix.S,
3391         src/m68k/sysv.S, src/mips/n32.S, src/mips/o32.S,
3392         src/powerpc/aix.S, src/powerpc/darwin.S,
3393         src/powerpc/ffi_darwin.c, src/powerpc/linux64.S,
3394         src/powerpc/linux64_closure.S, src/powerpc/ppc_closure.S,
3395         src/powerpc/sysv.S, src/s390/sysv.S, src/sh/sysv.S,
3396         src/sh64/sysv.S, src/sparc/v8.S, src/sparc/v9.S,
3397         src/x86/sysv.S, src/x86/unix64.S, src/x86/win32.S:
3398         include fficonfig.h
3399
3400 2003-10-20  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3401
3402         * src/mips/ffi.c: Use _ABIN32, _ABIO32 instead of external
3403         _MIPS_SIM_NABI32, _MIPS_SIM_ABI32.
3404
3405 2003-10-19  Andreas Tobler  <a.tobler@schweiz.ch>
3406
3407         * src/powerpc/ffi_darwin.c (ffi_prep_args): Declare bytes again.
3408         Used when FFI_DEBUG = 1.
3409
3410 2003-10-14  Alan Modra  <amodra@bigpond.net.au>
3411
3412         * src/types.c (double, longdouble): Default POWERPC64 to 8 byte size
3413         and align.
3414
3415 2003-10-06  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3416
3417         * include/ffi_mips.h: Define FFI_MIPS_N32 for N32/N64 ABIs,
3418         FFI_MIPS_O32 for O32 ABI.
3419
3420 2003-10-01  Andreas Tobler  <a.tobler@schweiz.ch>
3421
3422         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_64 for
3423         SPARC64. Cleanup whitespaces.
3424
3425 2003-09-19  Andreas Tobler  <a.tobler@schweiz.ch>
3426
3427         * testsuite/libffi.call/closure_fn0.c: Xfail mips, arm,
3428         strongarm, xscale. Cleanup whitespaces.
3429         * testsuite/libffi.call/closure_fn1.c: Likewise.
3430         * testsuite/libffi.call/closure_fn2.c: Likewise.
3431         * testsuite/libffi.call/closure_fn3.c: Likewise.
3432         * testsuite/libffi.call/cls_12byte.c: Likewise.
3433         * testsuite/libffi.call/cls_16byte.c: Likewise.
3434         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3435         * testsuite/libffi.call/cls_20byte.c: Likewise.
3436         * testsuite/libffi.call/cls_24byte.c: Likewise.
3437         * testsuite/libffi.call/cls_2byte.c: Likewise.
3438         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3439         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3440         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3441         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3442         * testsuite/libffi.call/cls_4byte.c: Likewise.
3443         * testsuite/libffi.call/cls_5byte.c: Likewise.
3444         * testsuite/libffi.call/cls_6byte.c: Likewise.
3445         * testsuite/libffi.call/cls_7byte.c: Likewise.
3446         * testsuite/libffi.call/cls_8byte.c: Likewise.
3447         * testsuite/libffi.call/cls_double.c: Likewise.
3448         * testsuite/libffi.call/cls_float.c: Likewise.
3449         * testsuite/libffi.call/cls_uchar.c: Likewise.
3450         * testsuite/libffi.call/cls_uint.c: Likewise.
3451         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3452         * testsuite/libffi.call/cls_ushort.c: Likewise.
3453         * testsuite/libffi.call/nested_struct.c: Likewise.
3454         * testsuite/libffi.call/nested_struct1.c: Likewise.
3455         * testsuite/libffi.call/problem1.c: Likewise.
3456         * testsuite/libffi.special/unwindtest.cc: Likewise.
3457         * testsuite/libffi.call/pyobjc-tc.c: Cleanup whitespaces.
3458
3459 2003-09-18  David Edelsohn  <edelsohn@gnu.org>
3460
3461         * src/powerpc/aix.S: Cleanup whitespaces.
3462         * src/powerpc/aix_closure.S: Likewise.
3463
3464 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3465
3466         * src/powerpc/darwin.S: Cleanup whitespaces, comment formatting.
3467         * src/powerpc/darwin_closure.S: Likewise.
3468         * src/powerpc/ffi_darwin.c: Likewise.
3469
3470 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3471             David Edelsohn  <edelsohn@gnu.org>
3472
3473         * src/types.c (double): Add AIX and Darwin to the right TYPEDEF.
3474         * src/powerpc/aix_closure.S: Remove the pointer to the outgoing
3475         parameter stack.
3476         * src/powerpc/darwin_closure.S: Likewise.
3477         * src/powerpc/ffi_darwin.c (ffi_prep_args): Handle structures
3478         according to the Darwin/AIX ABI.
3479         (ffi_prep_cif_machdep): Likewise.
3480         (ffi_closure_helper_DARWIN): Likewise.
3481         Remove the outgoing parameter stack logic. Simplify the evaluation
3482         of the different CASE types.
3483         (ffi_prep_clousure): Avoid the casts on lvalues. Change the branch
3484         statement in the trampoline code.
3485
3486 2003-09-18  Kaz Kojima  <kkojima@gcc.gnu.org>
3487
3488         * src/sh/ffi.c (ffi_prep_args): Take account into the alignement
3489         for the register size.
3490         (ffi_closure_helper_SYSV): Handle the structure return value
3491         address correctly.
3492         (ffi_closure_helper_SYSV): Return the appropriate type when
3493         the registers are used for the structure return value.
3494         * src/sh/sysv.S (ffi_closure_SYSV): Fix the stack layout for
3495         the 64-bit return value.  Update copyright years.
3496
3497 2003-09-17  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3498
3499         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Search in
3500         srcdir for ffi_mips.h.
3501
3502 2003-09-12  Alan Modra  <amodra@bigpond.net.au>
3503
3504         * src/prep_cif.c (initialize_aggregate): Include tail padding in
3505         structure size.
3506         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Correct
3507         placement of float result.
3508         * testsuite/libffi.special/unwindtest.cc (closure_test_fn1): Correct
3509         cast of "resp" for big-endian 64 bit machines.
3510
3511 2003-09-11  Alan Modra  <amodra@bigpond.net.au>
3512
3513         * src/types.c (double, longdouble): Merge identical SH and ARM
3514         typedefs, and add POWERPC64.
3515         * src/powerpc/ffi.c (ffi_prep_args64): Correct next_arg calc for
3516         struct split over gpr and rest.
3517         (ffi_prep_cif_machdep): Correct intarg_count for structures.
3518         * src/powerpc/linux64.S (ffi_call_LINUX64): Fix gpr offsets.
3519
3520 2003-09-09  Andreas Tobler  <a.tobler@schweiz.ch>
3521
3522         * src/powerpc/ffi.c (ffi_closure_helper_SYSV) Handle struct
3523         passing correctly.
3524
3525 2003-09-09  Alan Modra  <amodra@bigpond.net.au>
3526
3527         * configure: Regenerate.
3528
3529 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3530
3531         * Makefile.am: Remove build rules for ffitest.
3532         * Makefile.in: Rebuilt.
3533
3534 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3535
3536         * src/java_raw_api.c: Include <stdlib.h> to fix compiler warning
3537         about implicit declaration of abort().
3538
3539 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3540
3541         * Makefile.am: Add dejagnu test framework. Fixes PR other/11411.
3542         * Makefile.in: Rebuilt.
3543         * configure.in: Add dejagnu test framework.
3544         * configure: Rebuilt.
3545
3546         * testsuite/Makefile.am: New file.
3547         * testsuite/Makefile.in: Built
3548         * testsuite/lib/libffi-dg.exp: New file.
3549         * testsuite/config/default.exp: Likewise.
3550         * testsuite/libffi.call/call.exp: Likewise.
3551         * testsuite/libffi.call/ffitest.h: Likewise.
3552         * testsuite/libffi.call/closure_fn0.c: Likewise.
3553         * testsuite/libffi.call/closure_fn1.c: Likewise.
3554         * testsuite/libffi.call/closure_fn2.c: Likewise.
3555         * testsuite/libffi.call/closure_fn3.c: Likewise.
3556         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3557         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3558         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3559         * testsuite/libffi.call/cls_2byte.c: Likewise.
3560         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3561         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3562         * testsuite/libffi.call/cls_4byte.c: Likewise.
3563         * testsuite/libffi.call/cls_5byte.c: Likewise.
3564         * testsuite/libffi.call/cls_6byte.c: Likewise.
3565         * testsuite/libffi.call/cls_7byte.c: Likewise.
3566         * testsuite/libffi.call/cls_8byte.c: Likewise.
3567         * testsuite/libffi.call/cls_12byte.c: Likewise.
3568         * testsuite/libffi.call/cls_16byte.c: Likewise.
3569         * testsuite/libffi.call/cls_20byte.c: Likewise.
3570         * testsuite/libffi.call/cls_24byte.c: Likewise.
3571         * testsuite/libffi.call/cls_double.c: Likewise.
3572         * testsuite/libffi.call/cls_float.c: Likewise.
3573         * testsuite/libffi.call/cls_uchar.c: Likewise.
3574         * testsuite/libffi.call/cls_uint.c: Likewise.
3575         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3576         * testsuite/libffi.call/cls_ushort.c: Likewise.
3577         * testsuite/libffi.call/float.c: Likewise.
3578         * testsuite/libffi.call/float1.c: Likewise.
3579         * testsuite/libffi.call/float2.c: Likewise.
3580         * testsuite/libffi.call/many.c: Likewise.
3581         * testsuite/libffi.call/many_win32.c: Likewise.
3582         * testsuite/libffi.call/nested_struct.c: Likewise.
3583         * testsuite/libffi.call/nested_struct1.c: Likewise.
3584         * testsuite/libffi.call/pyobjc-tc.c: Likewise.
3585         * testsuite/libffi.call/problem1.c: Likewise.
3586         * testsuite/libffi.call/promotion.c: Likewise.
3587         * testsuite/libffi.call/return_ll.c: Likewise.
3588         * testsuite/libffi.call/return_sc.c: Likewise.
3589         * testsuite/libffi.call/return_uc.c: Likewise.
3590         * testsuite/libffi.call/strlen.c: Likewise.
3591         * testsuite/libffi.call/strlen_win32.c: Likewise.
3592         * testsuite/libffi.call/struct1.c: Likewise.
3593         * testsuite/libffi.call/struct2.c: Likewise.
3594         * testsuite/libffi.call/struct3.c: Likewise.
3595         * testsuite/libffi.call/struct4.c: Likewise.
3596         * testsuite/libffi.call/struct5.c: Likewise.
3597         * testsuite/libffi.call/struct6.c: Likewise.
3598         * testsuite/libffi.call/struct7.c: Likewise.
3599         * testsuite/libffi.call/struct8.c: Likewise.
3600         * testsuite/libffi.call/struct9.c: Likewise.
3601         * testsuite/libffi.special/special.exp: New file.
3602         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3603         * testsuite/libffi.special/unwindtest.cc: Likewise.
3604
3605
3606 2003-08-13  Kaz Kojima  <kkojima@gcc.gnu.org>
3607
3608         * src/sh/ffi.c (OFS_INT16): Set 0 for little endian case.  Update
3609         copyright years.
3610
3611 2003-08-02  Alan Modra  <amodra@bigpond.net.au>
3612
3613         * src/powerpc/ffi.c (ffi_prep_args64): Modify for changed gcc
3614         structure passing.
3615         (ffi_closure_helper_LINUX64): Likewise.
3616         * src/powerpc/linux64.S: Remove code writing to parm save area.
3617         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Use return
3618         address in lr from ffi_closure_helper_LINUX64 call to calculate
3619         table address.  Optimize function tail.
3620
3621 2003-07-28  Andreas Tobler  <a.tobler@schweiz.ch>
3622
3623         * src/sparc/ffi.c: Handle all floating point registers.
3624         * src/sparc/v9.S: Likewise. Fixes second part of PR target/11410.
3625
3626 2003-07-11  Gerald Pfeifer  <pfeifer@dbai.tuwien.ac.at>
3627
3628         * README: Note that libffi is not part of GCC.  Update the project
3629         URL and status.
3630
3631 2003-06-19  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3632
3633         * src/powerpc/ppc_closure.S: Include ffi.h.
3634
3635 2003-06-13  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3636
3637         * src/x86/sysv.S: Avoid gas-only .uleb128/.sleb128 directives.
3638         Use C style comments.
3639
3640 2003-06-13  Kaz Kojima  <kkojima@rr.iij4u.or.jp>
3641
3642         * Makefile.am: Add SHmedia support.  Fix a typo of SH support.
3643         * Makefile.in: Regenerate.
3644         * configure.in (sh64-*-linux*, sh5*-*-linux*): Add target.
3645         * configure: Regenerate.
3646         * include/ffi.h.in: Add SHmedia support.
3647         * src/sh64/ffi.c: New file.
3648         * src/sh64/sysv.S: New file.
3649
3650 2003-05-16  Jakub Jelinek  <jakub@redhat.com>
3651
3652         * configure.in (HAVE_RO_EH_FRAME): Check whether .eh_frame section
3653         should be read-only.
3654         * configure: Rebuilt.
3655         * fficonfig.h.in: Rebuilt.
3656         * include/ffi.h.in (EH_FRAME_FLAGS): Define.
3657         * src/alpha/osf.S: Use EH_FRAME_FLAGS.
3658         * src/powerpc/linux64.S: Likewise.
3659         * src/powerpc/linux64_closure.S: Likewise.  Include ffi.h.
3660         * src/powerpc/sysv.S: Use EH_FRAME_FLAGS.  Use pcrel encoding
3661         if -fpic/-fPIC/-mrelocatable.
3662         * src/powerpc/powerpc_closure.S: Likewise.
3663         * src/sparc/v8.S: If HAVE_RO_EH_FRAME is defined, don't include
3664         #write in .eh_frame flags.
3665         * src/sparc/v9.S: Likewise.
3666         * src/x86/unix64.S: Use EH_FRAME_FLAGS.
3667         * src/x86/sysv.S: Likewise.  Use pcrel encoding if -fpic/-fPIC.
3668         * src/s390/sysv.S: Use EH_FRAME_FLAGS.  Include ffi.h.
3669
3670 2003-05-07  Jeff Sturm  <jsturm@one-point.com>
3671
3672         Fixes PR bootstrap/10656
3673         * configure.in (HAVE_AS_REGISTER_PSEUDO_OP): Test assembler
3674         support for .register pseudo-op.
3675         * src/sparc/v8.S: Use it.
3676         * fficonfig.h.in: Rebuilt.
3677         * configure: Rebuilt.
3678
3679 2003-04-18  Jakub Jelinek  <jakub@redhat.com>
3680
3681         * include/ffi.h.in (POWERPC64): Define if 64-bit.
3682         (enum ffi_abi): Add FFI_LINUX64 on POWERPC.
3683         Make it the default on POWERPC64.
3684         (FFI_TRAMPOLINE_SIZE): Define to 24 on POWERPC64.
3685         * configure.in: Change powerpc-*-linux* into powerpc*-*-linux*.
3686         * configure: Rebuilt.
3687         * src/powerpc/ffi.c (hidden): Define.
3688         (ffi_prep_args_SYSV): Renamed from
3689         ffi_prep_args.  Cast pointers to unsigned long to shut up warnings.
3690         (NUM_GPR_ARG_REGISTERS64, NUM_FPR_ARG_REGISTERS64,
3691         ASM_NEEDS_REGISTERS64): New.
3692         (ffi_prep_args64): New function.
3693         (ffi_prep_cif_machdep): Handle FFI_LINUX64 ABI.
3694         (ffi_call): Likewise.
3695         (ffi_prep_closure): Likewise.
3696         (flush_icache): Surround by #ifndef POWERPC64.
3697         (ffi_dblfl): New union type.
3698         (ffi_closure_helper_SYSV): Use it to avoid aliasing problems.
3699         (ffi_closure_helper_LINUX64): New function.
3700         * src/powerpc/ppc_closure.S: Surround whole file by #ifndef
3701         __powerpc64__.
3702         * src/powerpc/sysv.S: Likewise.
3703         (ffi_call_SYSV): Rename ffi_prep_args to ffi_prep_args_SYSV.
3704         * src/powerpc/linux64.S: New file.
3705         * src/powerpc/linux64_closure.S: New file.
3706         * Makefile.am (EXTRA_DIST): Add src/powerpc/linux64.S and
3707         src/powerpc/linux64_closure.S.
3708         (TARGET_SRC_POWERPC): Likewise.
3709
3710         * src/ffitest.c (closure_test_fn, closure_test_fn1, closure_test_fn2,
3711         closure_test_fn3): Fix result printing on big-endian 64-bit
3712         machines.
3713         (main): Print tst2_arg instead of uninitialized tst2_result.
3714
3715         * src/ffitest.c (main): Hide what closure pointer really points to
3716         from the compiler.
3717
3718 2003-04-16  Richard Earnshaw  <rearnsha@arm.com>
3719
3720         * configure.in (arm-*-netbsdelf*): Add configuration.
3721         (configure): Regenerated.
3722
3723 2003-04-04  Loren J. Rittle  <ljrittle@acm.org>
3724
3725         * include/Makefile.in: Regenerate.
3726
3727 2003-03-21  Zdenek Dvorak  <rakdver@atrey.karlin.mff.cuni.cz>
3728
3729         * libffi/include/ffi.h.in: Define X86 instead of X86_64 in 32
3730         bit mode.
3731         * libffi/src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV):
3732         Receive closure pointer through parameter, read args using
3733         __builtin_dwarf_cfa.
3734         (FFI_INIT_TRAMPOLINE): Send closure reference through eax.
3735
3736 2003-03-12  Andreas Schwab  <schwab@suse.de>
3737
3738         * configure.in: Avoid trailing /. in toolexeclibdir.
3739         * configure: Rebuilt.
3740
3741 2003-03-03  Andreas Tobler <a.tobler@schweiz.ch>
3742
3743         * src/powerpc/darwin_closure.S: Recode to fit dynamic libraries.
3744
3745 2003-02-06  Andreas Tobler <a.tobler@schweiz.ch>
3746
3747         * libffi/src/powerpc/darwin_closure.S:
3748         Fix alignement bug, allocate 8 bytes for the result.
3749         * libffi/src/powerpc/aix_closure.S:
3750         Likewise.
3751         * libffi/src/powerpc/ffi_darwin.c:
3752         Update stackframe description for aix/darwin_closure.S.
3753
3754 2003-02-06  Jakub Jelinek  <jakub@redhat.com>
3755
3756         * src/s390/ffi.c (ffi_closure_helper_SYSV): Add hidden visibility
3757         attribute.
3758
3759 2003-01-31  Christian Cornelssen  <ccorn@cs.tu-berlin.de>,
3760             Andreas Schwab  <schwab@suse.de>
3761
3762         * configure.in: Adjust command to source config-ml.in to account
3763         for changes to the libffi_basedir definition.
3764         (libffi_basedir): Remove ${srcdir} from value and include trailing
3765         slash if nonempty.
3766
3767         * configure: Regenerate.
3768
3769 2003-01-29  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3770
3771         * src/powerpc/ppc_closure.S: Recode to fit shared libs.
3772
3773 2003-01-28  Andrew Haley  <aph@redhat.com>
3774
3775         * include/ffi.h.in: Enable FFI_CLOSURES for x86_64.
3776         * src/x86/ffi64.c (ffi_prep_closure): New.
3777         (ffi_closure_UNIX64_inner): New.
3778         * src/x86/unix64.S (ffi_closure_UNIX64): New.
3779
3780 2003-01-27  Alexandre Oliva  <aoliva@redhat.com>
3781
3782         * configure.in (toolexecdir, toolexeclibdir): Set and AC_SUBST.
3783         Remove USE_LIBDIR conditional.
3784         * Makefile.am (toolexecdir, toolexeclibdir): Don't override.
3785         * Makefile.in, configure: Rebuilt.
3786
3787 2003-01027  David Edelsohn  <edelsohn@gnu.org>
3788
3789         * Makefile.am (TARGET_SRC_POWERPC_AIX): Fix typo.
3790         * Makefile.in: Regenerate.
3791
3792 2003-01-22  Andrew Haley  <aph@redhat.com>
3793
3794         * src/powerpc/darwin.S (_ffi_call_AIX): Add Augmentation size to
3795         unwind info.
3796
3797 2003-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
3798
3799         * src/powerpc/darwin.S: Add unwind info.
3800         * src/powerpc/darwin_closure.S: Likewise.
3801
3802 2003-01-14  Andrew Haley  <aph@redhat.com>
3803
3804         * src/x86/ffi64.c (ffi_prep_args): Check for void retval.
3805         (ffi_prep_cif_machdep): Likewise.
3806         * src/x86/unix64.S: Add unwind info.
3807
3808 2003-01-14  Andreas Jaeger  <aj@suse.de>
3809
3810         * src/ffitest.c (main): Only use ffi_closures if those are
3811         supported.
3812
3813 2003-01-13 Andreas Tobler <a.tobler@schweiz.ch>
3814
3815         * libffi/src/ffitest.c
3816          add closure testcases
3817
3818 2003-01-13 Kevin B. Hendricks <khendricks@ivey.uwo.ca>
3819
3820         * libffi/src/powerpc/ffi.c
3821          fix alignment bug for float (4 byte aligned iso 8 byte)
3822
3823 2003-01-09  Geoffrey Keating  <geoffk@apple.com>
3824
3825         * src/powerpc/ffi_darwin.c: Remove RCS version string.
3826         * src/powerpc/darwin.S: Remove RCS version string.
3827
3828 2003-01-03  Jeff Sturm  <jsturm@one-point.com>
3829
3830         * include/ffi.h.in: Add closure defines for SPARC, SPARC64.
3831         * src/ffitest.c (main): Use static storage for closure.
3832         * src/sparc/ffi.c (ffi_prep_closure, ffi_closure_sparc_inner): New.
3833         * src/sparc/v8.S (ffi_closure_v8): New.
3834         * src/sparc/v9.S (ffi_closure_v9): New.
3835
3836 2002-11-10  Ranjit Mathew <rmathew@hotmail.com>
3837
3838         * include/ffi.h.in: Added FFI_STDCALL ffi_type
3839           enumeration for X86_WIN32.
3840         * src/x86/win32.S: Added ffi_call_STDCALL function
3841           definition.
3842         * src/x86/ffi.c (ffi_call/ffi_raw_call): Added
3843           switch cases for recognising FFI_STDCALL and
3844           calling ffi_call_STDCALL if target is X86_WIN32.
3845         * src/ffitest.c (my_stdcall_strlen/stdcall_many):
3846           stdcall versions of the "my_strlen" and "many"
3847           test functions (for X86_WIN32).
3848           Added test cases to test stdcall invocation using
3849           these functions.
3850
3851 2002-12-02  Kaz Kojima  <kkojima@gcc.gnu.org>
3852
3853         * src/sh/sysv.S: Add DWARF2 unwind info.
3854
3855 2002-11-27  Ulrich Weigand  <uweigand@de.ibm.com>
3856
3857         * src/s390/sysv.S (.eh_frame section): Make section read-only.
3858
3859 2002-11-26  Jim Wilson  <wilson@redhat.com>
3860
3861         * src/types.c (FFI_TYPE_POINTER): Has size 8 on IA64.
3862
3863 2002-11-23  H.J. Lu <hjl@gnu.org>
3864
3865         * acinclude.m4: Add dummy AM_PROG_LIBTOOL.
3866         Include ../config/accross.m4.
3867         * aclocal.m4; Rebuild.
3868         * configure: Likewise.
3869
3870 2002-11-15  Ulrich Weigand  <uweigand@de.ibm.com>
3871
3872         * src/s390/sysv.S (.eh_frame section): Adapt to pcrel FDE encoding.
3873
3874 2002-11-11  DJ Delorie  <dj@redhat.com>
3875
3876         * configure.in: Look for common files in the right place.
3877
3878 2002-10-08  Ulrich Weigand  <uweigand@de.ibm.com>
3879
3880         * src/java_raw_api.c (ffi_java_raw_to_ptrarray): Interpret
3881         raw data as _Jv_word values, not ffi_raw.
3882         (ffi_java_ptrarray_to_raw): Likewise.
3883         (ffi_java_rvalue_to_raw): New function.
3884         (ffi_java_raw_call): Call it.
3885         (ffi_java_raw_to_rvalue): New function.
3886         (ffi_java_translate_args): Call it.
3887         * src/ffitest.c (closure_test_fn): Interpret return value
3888         as ffi_arg, not int.
3889         * src/s390/ffi.c (ffi_prep_cif_machdep): Add missing
3890         FFI_TYPE_POINTER case.
3891         (ffi_closure_helper_SYSV): Likewise.  Also, assume return
3892         values extended to word size.
3893
3894 2002-10-02  Andreas Jaeger  <aj@suse.de>
3895
3896         * src/x86/ffi64.c (ffi_prep_cif_machdep): Remove debug output.
3897
3898 2002-10-01  Bo Thorsen  <bo@smetana.suse.de>
3899
3900         * include/ffi.h.in: Fix i386 win32 compilation.
3901
3902 2002-09-30  Ulrich Weigand  <uweigand@de.ibm.com>
3903
3904         * configure.in: Add s390x-*-linux-* target.
3905         * configure: Regenerate.
3906         * include/ffi.h.in: Define S390X for s390x targets.
3907         (FFI_CLOSURES): Define for s390/s390x.
3908         (FFI_TRAMPOLINE_SIZE): Likewise.
3909         (FFI_NATIVE_RAW_API): Likewise.
3910         * src/prep_cif.c (ffi_prep_cif): Do not compute stack space for s390.
3911         * src/types.c (FFI_TYPE_POINTER): Use 8-byte pointers on s390x.
3912         * src/s390/ffi.c: Major rework of existing code.  Add support for
3913         s390x targets.  Add closure support.
3914         * src/s390/sysv.S: Likewise.
3915
3916 2002-09-29  Richard Earnshaw  <rearnsha@arm.com>
3917
3918         * src/arm/sysv.S: Fix typo.
3919
3920 2002-09-28  Richard Earnshaw  <rearnsha@arm.com>
3921
3922         * src/arm/sysv.S: If we don't have machine/asm.h and the pre-processor
3923         has defined __USER_LABEL_PREFIX__, then use it in CNAME.
3924         (ffi_call_SYSV): Handle soft-float.
3925
3926 2002-09-27  Bo Thorsen  <bo@suse.de>
3927
3928         * include/ffi.h.in: Fix multilib x86-64 support.
3929
3930 2002-09-22  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
3931
3932         * Makefile.am (all-multi): Fix multilib parallel build.
3933
3934 2002-07-19  Kaz Kojima  <kkojima@gcc.gnu.org>
3935
3936         * configure.in (sh[34]*-*-linux*): Add brackets.
3937         * configure: Regenerate.
3938
3939 2002-07-18  Kaz Kojima  <kkojima@gcc.gnu.org>
3940
3941         * Makefile.am: Add SH support.
3942         * Makefile.in: Regenerate.
3943         * configure.in (sh-*-linux*, sh[34]*-*-linux*): Add target.
3944         * configure: Regenerate.
3945         * include/ffi.h.in: Add SH support.
3946         * src/sh/ffi.c: New file.
3947         * src/sh/sysv.S: New file.
3948         * src/types.c: Add SH support.
3949
3950 2002-07-16  Bo Thorsen  <bo@suse.de>
3951
3952         * src/x86/ffi64.c: New file that adds x86-64 support.
3953         * src/x86/unix64.S: New file that handles argument setup for
3954         x86-64.
3955         * src/x86/sysv.S: Don't use this on x86-64.
3956         * src/x86/ffi.c: Don't use this on x86-64.
3957         Remove unused vars.
3958         * src/prep_cif.c (ffi_prep_cif): Don't do stack size calculation
3959         for x86-64.
3960         * src/ffitest.c (struct6): New test that tests a special case in
3961         the x86-64 ABI.
3962         (struct7): Likewise.
3963         (struct8): Likewise.
3964         (struct9): Likewise.
3965         (closure_test_fn): Silence warning about this when it's not used.
3966         (main): Add the new tests.
3967         (main): Fix a couple of wrong casts and silence some compiler warnings.
3968         * include/ffi.h.in: Add x86-64 ABI definition.
3969         * fficonfig.h.in: Regenerate.
3970         * Makefile.am: Add x86-64 support.
3971         * configure.in: Likewise.
3972         * Makefile.in: Regenerate.
3973         * configure: Likewise.
3974
3975 2002-06-24  Bo Thorsen  <bo@suse.de>
3976
3977         * src/types.c: Merge settings for similar architectures.
3978         Add x86-64 sizes and alignments.
3979
3980 2002-06-23  Bo Thorsen  <bo@suse.de>
3981
3982         * src/arm/ffi.c (ffi_prep_args): Remove unused vars.
3983         * src/sparc/ffi.c (ffi_prep_args_v8): Likewise.
3984         * src/mips/ffi.c (ffi_prep_args): Likewise.
3985         * src/m68k/ffi.c (ffi_prep_args): Likewise.
3986
3987 2002-07-18  H.J. Lu  (hjl@gnu.org)
3988
3989         * Makefile.am (TARGET_SRC_MIPS_LINUX): New.
3990         (libffi_la_SOURCES): Support MIPS_LINUX.
3991         (libffi_convenience_la_SOURCES): Likewise.
3992         * Makefile.in: Regenerated.
3993
3994         * configure.in (mips64*-*): Skip.
3995         (mips*-*-linux*): New.
3996         * configure: Regenerated.
3997
3998         * src/mips/ffi.c: Include <sgidefs.h>.
3999
4000 2002-06-06  Ulrich Weigand  <uweigand@de.ibm.com>
4001
4002         * src/s390/sysv.S: Save/restore %r6.  Add DWARF-2 unwind info.
4003
4004 2002-05-27  Roger Sayle  <roger@eyesopen.com>
4005
4006         * src/x86/ffi.c (ffi_prep_args): Remove reference to avn.
4007
4008 2002-05-27  Bo Thorsen  <bo@suse.de>
4009
4010         * src/x86/ffi.c (ffi_prep_args): Remove unused variable and
4011         fix formatting.
4012
4013 2002-05-13  Andreas Tobler  <a.tobler@schweiz.ch>
4014
4015         * src/powerpc/ffi_darwin.c (ffi_prep_closure): Declare fd at
4016         beginning of function (for older apple cc).
4017
4018 2002-05-08  Alexandre Oliva  <aoliva@redhat.com>
4019
4020         * configure.in (ORIGINAL_LD_FOR_MULTILIBS): Preserve LD at
4021         script entry, and set LD to it when configuring multilibs.
4022         * configure: Rebuilt.
4023
4024 2002-05-05  Jason Thorpe  <thorpej@wasabisystems.com>
4025
4026         * configure.in (sparc64-*-netbsd*): Add target.
4027         (sparc-*-netbsdelf*): Likewise.
4028         * configure: Regenerate.
4029
4030 2002-04-28  David S. Miller  <davem@redhat.com>
4031
4032         * configure.in, configure: Fix SPARC test in previous change.
4033
4034 2002-04-29  Gerhard Tonn  <GerhardTonn@swol.de>
4035
4036         * Makefile.am: Add Linux for S/390 support.
4037         * Makefile.in: Regenerate.
4038         * configure.in: Add Linux for S/390 support.
4039         * configure: Regenerate.
4040         * include/ffi.h.in: Add Linux for S/390 support.
4041         * src/s390/ffi.c: New file from libffi CVS tree.
4042         * src/s390/sysv.S: New file from libffi CVS tree.
4043
4044 2002-04-28