OSDN Git Service

PR libffi/46661
[pf3gnuchains/gcc-fork.git] / libffi / ChangeLog
1 2011-02-09  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
2
3         PR libffi/46661
4         * testsuite/libffi.call/cls_pointer.c (main): Cast void * to
5         uintptr_t first.
6         * testsuite/libffi.call/cls_pointer_stack.c (main): Likewise.
7
8 2011-02-07  Joel Sherrill <joel.sherrill@oarcorp.com>
9
10         * libffi/src/m68k/ffi.c: Add RTEMS support for cache flushing.
11         Handle case when CPU variant does not have long double support.
12         * libffi/src/m68k/sysv.S: Add support for mc68000, Coldfire,
13         and cores with soft floating point.
14
15 2011-02-07  Joel Sherrill <joel.sherrill@oarcorp.com>
16
17         * configure.ac: Add mips*-*-rtems* support.
18         * configure: Regenerate.
19         * src/mips/ffitarget.h: Ensure needed constants are available
20         for targets which do not have sgidefs.h.
21
22 2011-01-26  Dave Korn  <dave.korn.cygwin@gmail.com>
23
24         PR target/40125
25         * configure.ac (AM_LTLDFLAGS): Add -bindir option for windows DLLs.
26         * configure: Regenerate.
27
28 2010-12-18  Iain Sandoe  <iains@gcc.gnu.org>
29
30         PR libffi/29152
31         PR libffi/42378
32         * src/powerpc/darwin_closure.S: Provide Darwin64 implementation,
33         update comments.
34         * src/powerpc/ffitarget.h (POWERPC_DARWIN64): New,
35         (FFI_TRAMPOLINE_SIZE): Update for Darwin64.
36         * src/powerpc/darwin.S: Provide Darwin64 implementation,
37         update comments.
38         * src/powerpc/ffi_darwin.c: Likewise.
39
40 2010-12-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
41
42         * configure.ac (libffi_cv_as_ascii_pseudo_op): Use double
43         backslashes.
44         (libffi_cv_as_string_pseudo_op): Likewise.
45         * configure: Regenerate.
46
47 2010-12-03  Chung-Lin Tang  <cltang@codesourcery.com>
48
49         * src/arm/sysv.S (ffi_closure_SYSV): Add UNWIND to .pad directive.
50         (ffi_closure_VFP): Same.
51         (ffi_call_VFP): Move down to before ffi_closure_VFP. Add '.fpu vfp'
52         directive.
53
54 2010-12-01  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
55
56         * testsuite/libffi.call/ffitest.h [__sgi] (PRId64, PRIu64): Define.
57         (PRIuPTR): Define.
58
59 2010-11-29  Richard Henderson  <rth@redhat.com>
60             Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
61
62         * src/x86/sysv.S (FDE_ENCODING, FDE_ENCODE): Define.
63         (.eh_frame): Use FDE_ENCODING.
64         (.LASFDE1, .LASFDE2, LASFDE3): Simplify with FDE_ENCODE.
65
66 2010-11-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
67
68         * testsuite/lib/libffi-dg.exp: Rename ...
69         * testsuite/lib/libffi.exp: ... to this.
70         * libffi/testsuite/libffi.call/call.exp: Don't load libffi-dg.exp.
71         * libffi/testsuite/libffi.special/special.exp: Likewise.
72
73 2010-10-28  Chung-Lin Tang  <cltang@codesourcery.com>
74
75         * src/arm/ffi.c (ffi_prep_args): Add VFP register argument handling
76         code, new parameter, and return value. Update comments.
77         (ffi_prep_cif_machdep): Add case for VFP struct return values. Add
78         call to layout_vfp_args().
79         (ffi_call_SYSV): Update declaration.
80         (ffi_call_VFP): New declaration.
81         (ffi_call): Add VFP struct return conditions. Call ffi_call_VFP()
82         when ABI is FFI_VFP.
83         (ffi_closure_VFP): New declaration.
84         (ffi_closure_SYSV_inner): Add new vfp_args parameter, update call to
85         ffi_prep_incoming_args_SYSV().
86         (ffi_prep_incoming_args_SYSV): Update parameters. Add VFP argument
87         case handling.
88         (ffi_prep_closure_loc): Pass ffi_closure_VFP to trampoline
89         construction under VFP hard-float.
90         (rec_vfp_type_p): New function.
91         (vfp_type_p): Same.
92         (place_vfp_arg): Same.
93         (layout_vfp_args): Same.
94         * src/arm/ffitarget.h (ffi_abi): Add FFI_VFP. Define FFI_DEFAULT_ABI
95         based on __ARM_PCS_VFP.
96         (FFI_EXTRA_CIF_FIELDS): Define for adding VFP hard-float specific
97         fields.
98         (FFI_TYPE_STRUCT_VFP_FLOAT): Define internally used type code.
99         (FFI_TYPE_STRUCT_VFP_DOUBLE): Same.
100         * src/arm/sysv.S (ffi_call_SYSV): Change call of ffi_prep_args() to
101         direct call. Move function pointer load upwards.
102         (ffi_call_VFP): New function.
103         (ffi_closure_VFP): Same.
104
105         * testsuite/lib/libffi-dg.exp (check-flags): New function.
106         (dg-skip-if): New function.
107         * testsuite/libffi.call/cls_double_va.c: Skip if target is arm*-*-*
108         and compiler options include -mfloat-abi=hard.
109         * testsuite/libffi.call/cls_longdouble_va.c: Same.
110
111 2010-10-01  Jakub Jelinek  <jakub@redhat.com>
112
113         PR libffi/45677
114         * src/x86/ffi64.c (ffi_prep_cif_machdep): Ensure cif->bytes is
115         a multiple of 8.
116         * testsuite/libffi.call/many2.c: New test.
117
118 2010-08-20  Mark Wielaard  <mjw@redhat.com>
119
120         * src/closures.c (open_temp_exec_file_mnt): Check if getmntent_r
121         returns NULL.
122
123 2010-08-09  Andreas Tobler  <andreast@fgznet.ch>
124
125         * configure.ac: Add target powerpc64-*-freebsd*.
126         * configure: Regenerate.
127         * testsuite/libffi.call/cls_align_longdouble_split.c: Pass
128         -mlong-double-128 only to linux targets.
129         * testsuite/libffi.call/cls_align_longdouble_split2.c: Likewise.
130         * testsuite/libffi.call/cls_longdouble.c: Likewise.
131         * testsuite/libffi.call/huge_struct.c: Likewise.
132
133 2010-07-10  Evan Phoenix  <evan@fallingsnow.net>
134
135         * src/closures.c (selinux_enabled_check): Fix strncmp usage bug.
136
137 2010-07-07  Dan HorĂ¡k <dan@danny.cz>
138
139         * include/ffi.h.in: Protect #define with #ifndef.
140         * src/powerpc/ffitarget.h: Ditto.
141         * src/s390/ffitarget.h: Ditto.
142         * src/sparc/ffitarget.h: Ditto.
143
144 2010-07-07   Neil Roberts <neil@linux.intel.com>
145
146         * src/x86/sysv.S (ffi_call_SYSV): Align the stack pointer to
147         16-bytes.
148
149 2010-07-02  Jakub Jelinek  <jakub@redhat.com>
150
151         * Makefile.am (AM_MAKEFLAGS): Pass also mandir to submakes.
152         * Makefile.in: Regenerated.
153
154 2010-05-19  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
155
156         * configure.ac (libffi_cv_as_x86_pcrel): Check for illegal in as
157         output, too.
158         (libffi_cv_as_ascii_pseudo_op): Check for .ascii.
159         (libffi_cv_as_string_pseudo_op): Check for .string.
160         * configure: Regenerate.
161         * fficonfig.h.in: Regenerate.
162         * src/x86/sysv.S (.eh_frame): Use .ascii, .string or error.
163
164 2010-05-05  Michael Kohler <michaelkohler@live.com>
165
166         * src/dlmalloc.c (dlfree): Fix spelling.
167         * src/ia64/ffi.c (ffi_prep_cif_machdep): Ditto.
168         * configure.ac: Ditto.
169         * configure: Rebuilt.
170
171 2010-04-13  Dan Witte  <dwitte@mozilla.com>
172
173         * msvcc.sh: Build with -W3 instead of -Wall.
174         * src/powerpc/ffi_darwin.c: Remove build warnings.
175         * src/x86/ffi.c: Ditto.
176         * src/x86/ffitarget.h: Ditto.
177
178 2010-04-12  Dan Witte  <dwitte@mozilla.com>
179             Walter Meinl <wuno@lsvw.de>
180
181         * configure.ac: Add OS/2 support.
182         * configure: Rebuilt.
183         * src/closures.c: Ditto.
184         * src/dlmalloc.c: Ditto.
185         * src/x86/win32.S: Ditto.
186
187 2010-04-07  Jakub Jelinek  <jakub@redhat.com>
188
189         * testsuite/libffi.call/err_bad_abi.c: Remove unused args variable.
190
191 2010-04-02  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
192
193         * Makefile.in: Regenerate.
194         * aclocal.m4: Regenerate.
195         * include/Makefile.in: Regenerate.
196         * man/Makefile.in: Regenerate.
197         * testsuite/Makefile.in: Regenerate.
198
199 2010-03-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
200
201         * configure.ac (libffi_cv_as_x86_64_unwind_section_type): New test.
202         * configure: Regenerate.
203         * fficonfig.h.in: Regenerate.
204         * libffi/src/x86/unix64.S (.eh_frame)
205         [HAVE_AS_X86_64_UNWIND_SECTION_TYPE]: Use @unwind section type.
206
207 2010-03-14  Matthias Klose  <doko@ubuntu.com>
208
209         * src/x86/ffi64.c: Fix typo in comment.
210         * src/x86/ffi.c: Use /* ... */ comment style.
211
212 2010-02-24  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
213
214         * doc/libffi.texi (The Closure API): Fix typo.
215         * doc/libffi.info: Remove.
216
217 2010-02-15  Matthias Klose  <doko@ubuntu.com>
218
219         * src/arm/sysv.S (__ARM_ARCH__): Define for processor
220         __ARM_ARCH_7EM__.
221
222 2010-01-15  Anthony Green  <green@redhat.com>
223
224         * README: Add notes on building with Microsoft Visual C++.
225
226 2010-01-15  Daniel Witte  <dwitte@mozilla.com>
227
228         * msvcc.sh: New file.
229
230         * src/x86/win32.S: Port assembly routines to MSVC and #ifdef.
231         * src/x86/ffi.c: Tweak function declaration and remove excess
232         parens.
233         * include/ffi.h.in: Add __declspec(align(8)) to typedef struct
234         ffi_closure.
235
236         * src/x86/ffi.c: Merge ffi_call_SYSV and ffi_call_STDCALL into new
237         function ffi_call_win32 on X86_WIN32.
238         * src/x86/win32.S (ffi_call_SYSV): Rename to ffi_call_win32.
239         (ffi_call_STDCALL): Remove.
240
241         * src/prep_cif.c (ffi_prep_cif): Move stack space allocation code
242         to ffi_prep_cif_machdep for x86.
243         * src/x86/ffi.c (ffi_prep_cif_machdep): To here.
244
245 2010-01-15  Oliver Kiddle  <okiddle@yahoo.co.uk>
246
247         * src/x86/ffitarget.h (ffi_abi): Check for __i386 and __amd64 for
248         Sun Studio compiler compatibility.
249
250 2010-01-12  Conrad Irwin <conrad.irwin@gmail.com>
251
252         * doc/libffi.texi: Add closure example.
253
254 2010-01-07  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
255
256         PR libffi/40701
257         * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRIdLL,
258         PRIuLL, PRId64, PRIu64, PRIuPTR): Define.
259         * testsuite/libffi.call/cls_align_sint64.c: Add -Wno-format on
260         alpha*-dec-osf*.
261         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
262         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
263         * testsuite/libffi.call/return_ll1.c: Likewise.
264         * testsuite/libffi.call/stret_medium2.c: Likewise.
265         * testsuite/libffi.special/ffitestcxx.h (allocate_mmap): Cast
266         MAP_FAILED to char *.
267
268 2010-01-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
269
270         * src/mips/n32.S: Use .abicalls and .eh_frame with __GNUC__.
271
272 2009-12-31  Anthony Green  <green@redhat.com>
273
274         * README: Update for libffi 3.0.9.
275
276 2009-12-27  Matthias Klose  <doko@ubuntu.com>
277
278         * configure.ac (HAVE_LONG_DOUBLE): Define for mips when
279         appropriate.
280         * configure: Rebuilt.
281
282 2009-12-26  Anthony Green  <green@redhat.com>
283
284         * testsuite/libffi.call/cls_longdouble_va.c: Mark as xfail for
285         avr32*-*-*.
286         * testsuite/libffi.call/cls_double_va.c: Ditto.
287
288 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
289
290         * testsuite/libffi.call/ffitest.h: Conditionally include stdint.h
291         and inttypes.h.
292         * testsuite/libffi.special/unwindtest.cc: Ditto.
293
294 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
295
296         * configure.ac: Add amd64-*-openbsd*.
297         * configure: Rebuilt.
298         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Link
299         openbsd programs with -lpthread.
300
301 2009-12-26  Anthony Green  <green@redhat.com>
302
303         * testsuite/libffi.call/cls_double_va.c,
304         testsuite/libffi.call/cls_longdouble.c,
305         testsuite/libffi.call/cls_longdouble_va.c,
306         testsuite/libffi.call/cls_pointer.c,
307         testsuite/libffi.call/cls_pointer_stack.c: Remove xfail for
308         mips*-*-* and arm*-*-*.
309         * testsuite/libffi.call/cls_align_longdouble_split.c,
310         testsuite/libffi.call/cls_align_longdouble_split2.c,
311         testsuite/libffi.call/stret_medium2.c,
312         testsuite/libffi.call/stret_medium.c,
313         testsuite/libffi.call/stret_large.c,
314         testsuite/libffi.call/stret_large2.c: Remove xfail for arm*-*-*.
315
316 2009-12-31  Kay Tietz  <ktietz70@googlemail.com>
317
318         * testsuite/libffi.call/ffitest.h,
319         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRuLL): Fix
320         definitions.
321
322 2009-12-31  Carlo Bramini  <carlo.bramix@libero.it>
323
324         * configure.ac (AM_LTLDFLAGS): Define for windows hosts.
325         * Makefile.am (libffi_la_LDFLAGS): Add AM_LTLDFLAGS.
326         * configure: Rebuilt.
327         * Makefile.in: Rebuilt.
328
329 2009-12-31  Anthony Green  <green@redhat.com>
330             Blake Chaffin.
331
332         * testsuite/libffi.call/huge_struct.c: New test case from Blake
333         Chaffin @ Apple.
334
335 2009-12-28  David Edelsohn  <edelsohn@gnu.org>
336
337         * src/powerpc/ffi_darwin.c (ffi_prep_args): Copy abi and nargs to
338         local variables.
339         (aix_adjust_aggregate_sizes): New function.
340         (ffi_prep_cif_machdep): Call it.
341
342 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
343
344         * configure.ac: Define FFI_MMAP_EXEC_WRIT for the given targets.
345         * configure: Regenerate.
346         * fficonfig.h.in: Likewise.
347         * src/closures.c: Remove the FFI_MMAP_EXEC_WRIT definition for
348         Solaris/x86.
349
350 2009-12-26  Andreas Schwab  <schwab@linux-m68k.org>
351
352         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Advance intarg_count
353         when a float arguments is passed in memory.
354         (ffi_closure_helper_SYSV): Mark general registers as used up when
355         a 64bit or soft-float long double argument is passed in memory.
356
357 2009-12-25  Matthias Klose  <doko@ubuntu.com>
358
359         * man/ffi_call.3: Fix #include in examples.
360         * doc/libffi.texi: Add dircategory.
361
362 2009-12-25  Frank Everdij <f.p.x.everdij@tudelft.nl>
363
364         * include/ffi.h.in: Placed '__GNUC__' ifdef around
365         '__attribute__((aligned(8)))' in ffi_closure, fixes compile for
366         IRIX MIPSPro c99.
367         * include/ffi_common.h: Added '__sgi' define to non
368         '__attribute__((__mode__()))' integer typedefs.
369         * src/mips/ffi.c (ffi_call, ffi_closure_mips_inner_O32,
370         ffi_closure_mips_inner_N32): Added 'defined(_MIPSEB)' to BE check.
371         (ffi_closure_mips_inner_O32, ffi_closure_mips_inner_N32): Added
372         FFI_LONGDOUBLE support and alignment(N32 only).
373         * src/mips/ffitarget.h: Corrected '#include <sgidefs.h>' for IRIX and
374         fixed non '__attribute__((__mode__()))' integer typedefs.
375         * src/mips/n32.S: Put '#ifdef linux' around '.abicalls' and '.eh_frame'
376         since they are Linux/GNU Assembler specific.
377
378 2009-12-25  Bradley Smith  <brad@brad-smith.co.uk>
379
380         * configure.ac, Makefile.am, src/avr32/ffi.c,
381         src/avr32/ffitarget.h,
382         src/avr32/sysv.S: Add AVR32 port.
383         * configure, Makefile.in: Rebuilt.
384
385 2009-12-21  Andreas Tobler  <a.tobler@schweiz.org>
386
387         * configure.ac: Make i?86 build on FreeBSD and OpenBSD.
388         * configure: Regenerate.
389
390 2009-12-15  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
391
392         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on PA HP-UX.
393
394 2009-12-13  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
395
396         * src/pa/ffi.c (ffi_closure_inner_pa32): Handle FFI_TYPE_LONGDOUBLE
397         type on HP-UX.
398
399 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
400
401         * src/sparc/ffi.c (ffi_closure_sparc_inner_v9): Properly align 'long
402         double' arguments.
403
404 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
405
406         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on Solaris < 10.
407
408 2009-12-10  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
409
410         PR libffi/40700
411         * src/closures.c [X86_64 && __sun__ && __svr4__]
412         (FFI_MMAP_EXEC_WRIT): Define.
413
414 2009-12-08  David Daney  <ddaney@caviumnetworks.com>
415
416         * testsuite/libffi.call/stret_medium.c: Remove xfail for mips*-*-*
417         * testsuite/libffi.call/cls_align_longdouble_split2.c: Same.
418         * testsuite/libffi.call/stret_large.c: Same.
419         * testsuite/libffi.call/cls_align_longdouble_split.c: Same.
420         * testsuite/libffi.call/stret_large2.c: Same.
421         * testsuite/libffi.call/stret_medium2.c: Same.
422
423 2009-12-07  David Edelsohn  <edelsohn@gnu.org>
424
425         * src/powerpc/aix_closure.S (libffi_closure_ASM): Fix tablejump
426         typo.
427
428 2009-12-05  David Edelsohn  <edelsohn@gnu.org>
429
430         * src/powerpc/aix.S: Update AIX32 code to be consistent with AIX64
431         code.
432         * src/powerpc/aix_closure.S: Same.
433
434 2009-12-05  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
435
436         * Makefile.in: Regenerate.
437         * configure: Regenerate.
438         * include/Makefile.in: Regenerate.
439         * man/Makefile.in: Regenerate.
440         * testsuite/Makefile.in: Regenerate.
441
442 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
443
444         * src/powerpc/aix_closure.S: Reorganize 64-bit code to match
445         linux64_closure.S.
446
447 2009-12-04  Uros Bizjak  <ubizjak@gmail.com>
448
449         PR libffi/41908
450         * src/x86/ffi64.c (classify_argument): Update from
451         gcc/config/i386/i386.c.
452         (ffi_closure_unix64_inner): Do not use the address of two consecutive
453         SSE registers directly.
454         * testsuite/libffi.call/cls_dbls_struct.c (main): Remove xfail
455         for x86_64 linux targets.
456
457 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
458
459         * src/powerpc/ffi_darwin.c (ffi_closure_helper_DARWIN): Increment
460         pfr for long double split between fpr13 and stack.
461
462 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
463
464         * src/powerpc/ffi_darwin.c (ffi_prep_args): Increment next_arg and
465         fparg_count twice for long double.
466
467 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
468
469         PR libffi/42243
470         * src/powerpc/ffi_darwin.c (ffi_prep_args): Remove extra parentheses.
471
472 2009-12-03  Uros Bizjak  <ubizjak@gmail.com>
473
474         * testsuite/libffi.call/cls_longdouble_va.c (main): Fix format string.
475         Remove xfails for x86 linux targets.
476
477 2009-12-02  David Edelsohn  <edelsohn@gnu.org>
478
479         * src/powerpc/ffi_darwin.c (ffi_prep_args): Fix typo in INT64
480         case.
481
482 2009-12-01  David Edelsohn  <edelsohn@gnu.org>
483
484         * src/powerpc/aix.S (ffi_call_AIX): Convert to more standard
485         register usage.  Call ffi_prep_args directly.  Add long double
486         return value support.
487         * src/powerpc/ffi_darwin.c (ffi_prep_args): Double arg increment
488         applies to FFI_TYPE_DOUBLE.  Correct fpr_base increment typo.
489         Separate FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases.
490         (ffi_prep_cif_machdep): Only 16 byte stack alignment in 64 bit
491         mode.
492         (ffi_closure_helper_DARWIN): Remove nf and ng counters.  Move temp
493         into case.
494         * src/powerpc/aix_closure.S: Maintain 16 byte stack alignment.
495         Allocate result area between params and FPRs.
496
497 2009-11-30  David Edelsohn  <edelsohn@gnu.org>
498
499         PR target/35484
500         * src/powerpc/ffitarget.h (POWERPC64): Define for PPC64 Linux and
501         AIX64.
502         * src/powerpc/aix.S: Implement AIX64 version.
503         * src/powerpc/aix_closure.S: Implement AIX64 version.
504         (ffi_closure_ASM): Use extsb, lha and displament addresses.
505         * src/powerpc/ffi_darwin.c (ffi_prep_args): Implement AIX64
506         support.
507         (ffi_prep_cif_machdep): Same.
508         (ffi_call): Same.
509         (ffi_closure_helper_DARWIN): Same.
510
511 2009-11-02  Andreas Tobler  <a.tobler@schweiz.org>
512
513         PR libffi/41908
514         * testsuite/libffi.call/testclosure.c: New test.
515
516 2009-09-28  Kai Tietz  <kai.tietz@onevision.com>
517
518         * src/x86/win64.S (_ffi_call_win64 stack): Remove for gnu
519         assembly version use of ___chkstk.
520
521 2009-09-23  Matthias Klose  <doko@ubuntu.com>
522
523         PR libffi/40242, PR libffi/41443
524         * src/arm/sysv.S (__ARM_ARCH__): Define for processors
525         __ARM_ARCH_6T2__, __ARM_ARCH_6M__, __ARM_ARCH_7__,
526         __ARM_ARCH_7A__, __ARM_ARCH_7R__, __ARM_ARCH_7M__.
527         Change the conditionals to __SOFTFP__ || __ARM_EABI__
528         for -mfloat-abi=softfp to work.
529
530 2009-09-17  Loren J. Rittle  <ljrittle@acm.org>
531
532         PR testsuite/32843 (strikes again)
533         * src/x86/ffi.c (ffi_prep_cif_machdep): Add X86_FREEBSD to
534         enable proper extension on char and short.
535
536 2009-09-15  David Daney  <ddaney@caviumnetworks.com>
537
538         * src/java_raw_api.c (ffi_java_raw_to_rvalue): Remove special
539         handling for FFI_TYPE_POINTER.
540         * src/mips/ffitarget.h (FFI_TYPE_STRUCT_D_SOFT,
541         FFI_TYPE_STRUCT_F_SOFT, FFI_TYPE_STRUCT_DD_SOFT,
542         FFI_TYPE_STRUCT_FF_SOFT, FFI_TYPE_STRUCT_FD_SOFT,
543         FFI_TYPE_STRUCT_DF_SOFT, FFI_TYPE_STRUCT_SOFT): New defines.
544         (FFI_N32_SOFT_FLOAT, FFI_N64_SOFT_FLOAT): New ffi_abi enumerations.
545         (enum ffi_abi): Set FFI_DEFAULT_ABI for soft-float.
546         * src/mips/n32.S (ffi_call_N32): Add handling for soft-float
547         structure and pointer returns.
548         (ffi_closure_N32): Add handling for pointer returns.
549         * src/mips/ffi.c (ffi_prep_args, calc_n32_struct_flags,
550         calc_n32_return_struct_flags): Handle soft-float.
551         (ffi_prep_cif_machdep):  Handle soft-float, fix pointer handling.
552         (ffi_call_N32): Declare proper argument types.
553         (ffi_call, copy_struct_N32, ffi_closure_mips_inner_N32): Handle
554         soft-float.
555
556 2009-08-24  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
557
558         * configure.ac (AC_PREREQ): Bump to 2.64.
559
560 2009-08-22  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
561
562         * Makefile.am (install-html, install-pdf): Remove.
563         * Makefile.in: Regenerate.
564
565         * Makefile.in: Regenerate.
566         * aclocal.m4: Regenerate.
567         * configure: Regenerate.
568         * fficonfig.h.in: Regenerate.
569         * include/Makefile.in: Regenerate.
570         * man/Makefile.in: Regenerate.
571         * testsuite/Makefile.in: Regenerate.
572
573 2009-07-30  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
574
575         * configure.ac (_AC_ARG_VAR_PRECIOUS): Use m4_rename_force.
576
577 2009-07-24  Dave Korn  <dave.korn.cygwin@gmail.com>
578
579         PR libffi/40807
580         * src/x86/ffi.c (ffi_prep_cif_machdep): Also use sign/zero-extending
581         return types for X86_WIN32.
582         * src/x86/win32.S (_ffi_call_SYSV): Handle omitted return types.
583         (_ffi_call_STDCALL, _ffi_closure_SYSV, _ffi_closure_raw_SYSV,
584         _ffi_closure_STDCALL): Likewise.
585
586         * src/closures.c (is_selinux_enabled): Define to const 0 for Cygwin.
587         (dlmmap, dlmunmap): Also use these functions on Cygwin.
588
589 2009-07-11  Richard Sandiford  <rdsandiford@googlemail.com>
590
591         PR testsuite/40699
592         PR testsuite/40707
593         PR testsuite/40709
594         * testsuite/lib/libffi-dg.exp: Revert 2009-07-02, 2009-07-01 and
595         2009-06-30 commits.
596
597 2009-07-01  Richard Sandiford  <r.sandiford@uk.ibm.com>
598
599         * testsuite/lib/libffi-dg.exp (libffi-init): Set ld_library_path
600         to "" before adding paths.  (This reinstates an assignment that
601         was removed by my 2009-06-30 commit, but changes the initial
602         value from "." to "".)
603
604 2009-07-01  H.J. Lu  <hongjiu.lu@intel.com>
605
606         PR testsuite/40601
607         * testsuite/lib/libffi-dg.exp (libffi-init): Properly set
608         gccdir.  Adjust ld_library_path for gcc only if gccdir isn't
609         empty.
610
611 2009-06-30  Richard Sandiford  <r.sandiford@uk.ibm.com>
612
613         * testsuite/lib/libffi-dg.exp (libffi-init): Don't add "."
614         to ld_library_path.  Use add_path.  Add just find_libgcc_s
615         to ld_library_path, not every libgcc multilib directory.
616
617 2009-06-16  Wim Lewis  <wiml@hhhh.org>
618
619         * src/powerpc/ffi.c: Avoid clobbering cr3 and cr4, which are
620         supposed to be callee-saved.
621         * src/powerpc/sysv.S (small_struct_return_value): Fix overrun of
622         return buffer for odd-size structs.
623
624 2009-06-16  Andreas Tobler  <a.tobler@schweiz.org>
625
626         PR libffi/40444
627         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Add
628         allow_stack_execute for Darwin.
629
630 2009-06-16  Andrew Haley  <aph@redhat.com>
631
632         * configure.ac (TARGETDIR): Add missing blank lines.
633         * configure: Regenerate.
634
635 2009-06-16  Andrew Haley  <aph@redhat.com>
636
637         * testsuite/libffi.call/cls_align_sint64.c,
638         testsuite/libffi.call/cls_align_uint64.c,
639         testsuite/libffi.call/cls_longdouble_va.c,
640         testsuite/libffi.call/cls_ulonglong.c,
641         testsuite/libffi.call/return_ll1.c,
642         testsuite/libffi.call/stret_medium2.c: Fix printf format
643         specifiers.
644         * testsuite/libffi.call/ffitest.h,
645         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRIuLL): Define.
646
647 2009-06-15  Andrew Haley  <aph@redhat.com>
648
649         * testsuite/libffi.call/err_bad_typedef.c: xfail everywhere.
650         * testsuite/libffi.call/err_bad_abi.c: Likewise.
651
652 2009-06-12  Andrew Haley  <aph@redhat.com>
653
654         * Makefile.am: Remove info_TEXINFOS.
655
656 2009-06-12  Andrew Haley  <aph@redhat.com>
657
658         * ChangeLog.libffi: testsuite/libffi.call/cls_align_sint64.c,
659         testsuite/libffi.call/cls_align_uint64.c,
660         testsuite/libffi.call/cls_ulonglong.c,
661         testsuite/libffi.call/return_ll1.c,
662         testsuite/libffi.call/stret_medium2.c: Fix printf format
663         specifiers.
664         testsuite/libffi.special/unwindtest.cc: include stdint.h.
665
666 2009-06-11  Timothy Wall  <twall@users.sf.net>
667
668         * Makefile.am,
669         configure.ac,
670         include/ffi.h.in,
671         include/ffi_common.h,
672         src/closures.c,
673         src/dlmalloc.c,
674         src/x86/ffi.c,
675         src/x86/ffitarget.h,
676         src/x86/win64.S (new),
677         README: Added win64 support (mingw or MSVC)
678         * Makefile.in,
679         include/Makefile.in,
680         man/Makefile.in,
681         testsuite/Makefile.in,
682         configure,
683         aclocal.m4: Regenerated
684         * ltcf-c.sh: properly escape cygwin/w32 path
685         * man/ffi_call.3: Clarify size requirements for return value.
686         * src/x86/ffi64.c: Fix filename in comment.
687         * src/x86/win32.S: Remove unused extern.
688
689         * testsuite/libffi.call/closure_fn0.c,
690         testsuite/libffi.call/closure_fn1.c,
691         testsuite/libffi.call/closure_fn2.c,
692         testsuite/libffi.call/closure_fn3.c,
693         testsuite/libffi.call/closure_fn4.c,
694         testsuite/libffi.call/closure_fn5.c,
695         testsuite/libffi.call/closure_fn6.c,
696         testsuite/libffi.call/closure_stdcall.c,
697         testsuite/libffi.call/cls_12byte.c,
698         testsuite/libffi.call/cls_16byte.c,
699         testsuite/libffi.call/cls_18byte.c,
700         testsuite/libffi.call/cls_19byte.c,
701         testsuite/libffi.call/cls_1_1byte.c,
702         testsuite/libffi.call/cls_20byte.c,
703         testsuite/libffi.call/cls_20byte1.c,
704         testsuite/libffi.call/cls_24byte.c,
705         testsuite/libffi.call/cls_2byte.c,
706         testsuite/libffi.call/cls_3_1byte.c,
707         testsuite/libffi.call/cls_3byte1.c,
708         testsuite/libffi.call/cls_3byte2.c,
709         testsuite/libffi.call/cls_4_1byte.c,
710         testsuite/libffi.call/cls_4byte.c,
711         testsuite/libffi.call/cls_5_1_byte.c,
712         testsuite/libffi.call/cls_5byte.c,
713         testsuite/libffi.call/cls_64byte.c,
714         testsuite/libffi.call/cls_6_1_byte.c,
715         testsuite/libffi.call/cls_6byte.c,
716         testsuite/libffi.call/cls_7_1_byte.c,
717         testsuite/libffi.call/cls_7byte.c,
718         testsuite/libffi.call/cls_8byte.c,
719         testsuite/libffi.call/cls_9byte1.c,
720         testsuite/libffi.call/cls_9byte2.c,
721         testsuite/libffi.call/cls_align_double.c,
722         testsuite/libffi.call/cls_align_float.c,
723         testsuite/libffi.call/cls_align_longdouble.c,
724         testsuite/libffi.call/cls_align_longdouble_split.c,
725         testsuite/libffi.call/cls_align_longdouble_split2.c,
726         testsuite/libffi.call/cls_align_pointer.c,
727         testsuite/libffi.call/cls_align_sint16.c,
728         testsuite/libffi.call/cls_align_sint32.c,
729         testsuite/libffi.call/cls_align_sint64.c,
730         testsuite/libffi.call/cls_align_uint16.c,
731         testsuite/libffi.call/cls_align_uint32.c,
732         testsuite/libffi.call/cls_align_uint64.c,
733         testsuite/libffi.call/cls_dbls_struct.c,
734         testsuite/libffi.call/cls_double.c,
735         testsuite/libffi.call/cls_double_va.c,
736         testsuite/libffi.call/cls_float.c,
737         testsuite/libffi.call/cls_longdouble.c,
738         testsuite/libffi.call/cls_longdouble_va.c,
739         testsuite/libffi.call/cls_multi_schar.c,
740         testsuite/libffi.call/cls_multi_sshort.c,
741         testsuite/libffi.call/cls_multi_sshortchar.c,
742         testsuite/libffi.call/cls_multi_uchar.c,
743         testsuite/libffi.call/cls_multi_ushort.c,
744         testsuite/libffi.call/cls_multi_ushortchar.c,
745         testsuite/libffi.call/cls_pointer.c,
746         testsuite/libffi.call/cls_pointer_stack.c,
747         testsuite/libffi.call/cls_schar.c,
748         testsuite/libffi.call/cls_sint.c,
749         testsuite/libffi.call/cls_sshort.c,
750         testsuite/libffi.call/cls_uchar.c,
751         testsuite/libffi.call/cls_uint.c,
752         testsuite/libffi.call/cls_ulonglong.c,
753         testsuite/libffi.call/cls_ushort.c,
754         testsuite/libffi.call/err_bad_abi.c,
755         testsuite/libffi.call/err_bad_typedef.c,
756         testsuite/libffi.call/float2.c,
757         testsuite/libffi.call/huge_struct.c,
758         testsuite/libffi.call/nested_struct.c,
759         testsuite/libffi.call/nested_struct1.c,
760         testsuite/libffi.call/nested_struct10.c,
761         testsuite/libffi.call/nested_struct2.c,
762         testsuite/libffi.call/nested_struct3.c,
763         testsuite/libffi.call/nested_struct4.c,
764         testsuite/libffi.call/nested_struct5.c,
765         testsuite/libffi.call/nested_struct6.c,
766         testsuite/libffi.call/nested_struct7.c,
767         testsuite/libffi.call/nested_struct8.c,
768         testsuite/libffi.call/nested_struct9.c,
769         testsuite/libffi.call/problem1.c,
770         testsuite/libffi.call/return_ldl.c,
771         testsuite/libffi.call/return_ll1.c,
772         testsuite/libffi.call/stret_large.c,
773         testsuite/libffi.call/stret_large2.c,
774         testsuite/libffi.call/stret_medium.c,
775         testsuite/libffi.call/stret_medium2.c,
776         testsuite/libffi.special/unwindtest.cc: use ffi_closure_alloc instead
777         of checking for MMAP.  Use intptr_t instead of long casts.
778
779 2009-06-11  Kaz Kojima  <kkojima@gcc.gnu.org>
780
781         * testsuite/libffi.call/cls_longdouble_va.c: Add xfail sh*-*-linux-*.
782         * testsuite/libffi.call/err_bad_abi.c: Add xfail sh*-*-*.
783         * testsuite/libffi.call/err_bad_typedef.c: Likewise.
784
785 2009-06-09  Andrew Haley  <aph@redhat.com>
786
787         * src/x86/freebsd.S: Add missing file.
788
789 2009-06-08  Andrew Haley  <aph@redhat.com>
790
791         Import from libffi 3.0.8:
792
793         * doc/libffi.texi: New file.
794         * doc/libffi.info: Likewise.
795         * doc/stamp-vti: Likewise.
796         * man/Makefile.am: New file.
797         * man/ffi_call.3: New file.
798
799         * Makefile.am (EXTRA_DIST): Add src/x86/darwin64.S,
800         src/dlmalloc.c.
801         (nodist_libffi_la_SOURCES): Add X86_FREEBSD.
802
803         * configure.ac: Bump version to 3.0.8.
804         parisc*-*-linux*: Add.
805         i386-*-freebsd* | i386-*-openbsd*: Add.
806         powerpc-*-beos*: Add.
807         AM_CONDITIONAL X86_FREEBSD: Add.
808         AC_CONFIG_FILES: Add man/Makefile.
809
810         * include/ffi.h.in (FFI_FN): Change void (*)() to void (*)(void).
811
812 2009-06-08  Andrew Haley  <aph@redhat.com>
813
814         * README: Import from libffi 3.0.8.
815
816 2009-06-08  Andrew Haley  <aph@redhat.com>
817
818         * testsuite/libffi.call/err_bad_abi.c: Add xfails.
819         * testsuite/libffi.call/cls_longdouble_va.c: Add xfails.
820         * testsuite/libffi.call/cls_dbls_struct.c: Add xfail x86_64-*-linux-*.
821         * testsuite/libffi.call/err_bad_typedef.c: Add xfails.
822
823         * testsuite/libffi.call/stret_medium2.c: Add __UNUSED__ to args.
824         * testsuite/libffi.call/stret_medium.c: Likewise.
825         * testsuite/libffi.call/stret_large2.c: Likewise.
826         * testsuite/libffi.call/stret_large.c:  Likewise.
827
828 2008-12-26  Timothy Wall  <twall@users.sf.net>
829
830         * testsuite/libffi.call/cls_longdouble.c,
831         testsuite/libffi.call/cls_longdouble_va.c,
832         testsuite/libffi.call/cls_align_longdouble.c,
833         testsuite/libffi.call/cls_align_longdouble_split.c,
834         testsuite/libffi.call/cls_align_longdouble_split2.c: mark expected
835         failures on x86_64 cygwin/mingw.
836
837 2008-12-22  Timothy Wall  <twall@users.sf.net>
838
839         * testsuite/libffi.call/closure_fn0.c,
840         testsuite/libffi.call/closure_fn1.c,
841         testsuite/libffi.call/closure_fn2.c,
842         testsuite/libffi.call/closure_fn3.c,
843         testsuite/libffi.call/closure_fn4.c,
844         testsuite/libffi.call/closure_fn5.c,
845         testsuite/libffi.call/closure_fn6.c,
846         testsuite/libffi.call/closure_loc_fn0.c,
847         testsuite/libffi.call/closure_stdcall.c,
848         testsuite/libffi.call/cls_align_pointer.c,
849         testsuite/libffi.call/cls_pointer.c,
850         testsuite/libffi.call/cls_pointer_stack.c: use portable cast from
851         pointer to integer (intptr_t).
852         * testsuite/libffi.call/cls_longdouble.c: disable for win64.
853
854 2008-07-24  Anthony Green  <green@redhat.com>
855
856         * testsuite/libffi.call/cls_dbls_struct.c,
857         testsuite/libffi.call/cls_double_va.c,
858         testsuite/libffi.call/cls_longdouble.c,
859         testsuite/libffi.call/cls_longdouble_va.c,
860         testsuite/libffi.call/cls_pointer.c,
861         testsuite/libffi.call/cls_pointer_stack.c,
862         testsuite/libffi.call/err_bad_abi.c: Clean up failures from
863         compiler warnings.
864
865 2008-03-04  Anthony Green  <green@redhat.com>
866             Blake Chaffin
867             hos@tamanegi.org
868
869         * testsuite/libffi.call/cls_align_longdouble_split2.c
870           testsuite/libffi.call/cls_align_longdouble_split.c
871           testsuite/libffi.call/cls_dbls_struct.c
872           testsuite/libffi.call/cls_double_va.c
873           testsuite/libffi.call/cls_longdouble.c
874           testsuite/libffi.call/cls_longdouble_va.c
875           testsuite/libffi.call/cls_pointer.c
876           testsuite/libffi.call/cls_pointer_stack.c
877           testsuite/libffi.call/err_bad_abi.c
878           testsuite/libffi.call/err_bad_typedef.c
879           testsuite/libffi.call/stret_large2.c
880           testsuite/libffi.call/stret_large.c
881           testsuite/libffi.call/stret_medium2.c
882           testsuite/libffi.call/stret_medium.c: New tests from Apple.
883
884 2009-06-05  Andrew Haley  <aph@redhat.com>
885
886         * src/x86/ffitarget.h, src/x86/ffi.c: Merge stdcall changes from
887         libffi.
888
889 2009-06-04  Andrew Haley  <aph@redhat.com>
890
891         * src/x86/ffitarget.h, src/x86/win32.S, src/x86/ffi.c: Back out
892         stdcall changes.
893
894 2008-02-26  Anthony Green  <green@redhat.com>
895             Thomas Heller  <theller@ctypes.org>
896
897         * src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C
898         comment.
899
900 2008-02-03  Timothy Wall  <twall@users.sf.net>
901
902         * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return
903           offset based on code pointer, not data pointer.
904
905 2008-01-31  Timothy Wall <twall@users.sf.net>
906
907         * testsuite/libffi.call/closure_stdcall.c: Add test for stdcall
908         closures.
909         * src/x86/ffitarget.h: Increase size of trampoline for stdcall
910         closures.
911         * src/x86/win32.S: Add assembly for stdcall closure.
912         * src/x86/ffi.c: Initialize stdcall closure trampoline.
913
914 2009-06-04  Andrew Haley  <aph@redhat.com>
915
916         * include/ffi.h.in: Change void (*)() to void (*)(void).
917         * src/x86/ffi.c: Likewise.
918
919 2009-06-04  Andrew Haley  <aph@redhat.com>
920
921         * src/powerpc/ppc_closure.S: Insert licence header.
922         * src/powerpc/linux64_closure.S: Likewise.
923         * src/m68k/sysv.S: Likewise.
924
925         * src/sh64/ffi.c: Change void (*)() to void (*)(void).
926         * src/powerpc/ffi.c: Likewise.
927         * src/powerpc/ffi_darwin.c: Likewise.
928         * src/m32r/ffi.c: Likewise.
929         * src/sh64/ffi.c: Likewise.
930         * src/x86/ffi64.c: Likewise.
931         * src/alpha/ffi.c: Likewise.
932         * src/alpha/osf.S: Likewise.
933         * src/frv/ffi.c: Likewise.
934         * src/s390/ffi.c: Likewise.
935         * src/pa/ffi.c: Likewise.
936         * src/pa/hpux32.S: Likewise.
937         * src/ia64/unix.S: Likewise.
938         * src/ia64/ffi.c: Likewise.
939         * src/sparc/ffi.c: Likewise.
940         * src/mips/ffi.c: Likewise.
941         * src/sh/ffi.c: Likewise.
942
943 2008-02-15  David Daney  <ddaney@avtrex.com>
944
945         * src/mips/ffi.c (USE__BUILTIN___CLEAR_CACHE):
946         Define (conditionally), and use it to include cachectl.h.
947         (ffi_prep_closure_loc): Fix cache flushing.
948         * src/mips/ffitarget.h (_ABIN32, _ABI64, _ABIO32): Define.
949
950 2009-06-04  Andrew Haley  <aph@redhat.com>
951
952         include/ffi.h.in,
953         src/arm/ffitarget.h,
954         src/arm/ffi.c,
955         src/arm/sysv.S,
956         src/powerpc/ffitarget.h,
957         src/closures.c,
958         src/sh64/ffitarget.h,
959         src/sh64/ffi.c,
960         src/sh64/sysv.S,
961         src/types.c,
962         src/x86/ffi64.c,
963         src/x86/ffitarget.h,
964         src/x86/win32.S,
965         src/x86/darwin.S,
966         src/x86/ffi.c,
967         src/x86/sysv.S,
968         src/x86/unix64.S,
969         src/alpha/ffitarget.h,
970         src/alpha/ffi.c,
971         src/alpha/osf.S,
972         src/m68k/ffitarget.h,
973         src/frv/ffitarget.h,
974         src/frv/ffi.c,
975         src/s390/ffitarget.h,
976         src/s390/sysv.S,
977         src/cris/ffitarget.h,
978         src/pa/linux.S,
979         src/pa/ffitarget.h,
980         src/pa/ffi.c,
981         src/raw_api.c,
982         src/ia64/ffitarget.h,
983         src/ia64/unix.S,
984         src/ia64/ffi.c,
985         src/ia64/ia64_flags.h,
986         src/java_raw_api.c,
987         src/debug.c,
988         src/sparc/v9.S,
989         src/sparc/ffitarget.h,
990         src/sparc/ffi.c,
991         src/sparc/v8.S,
992         src/mips/ffitarget.h,
993         src/mips/n32.S,
994         src/mips/o32.S,
995         src/mips/ffi.c,
996         src/prep_cif.c,
997         src/sh/ffitarget.h,
998         src/sh/ffi.c,
999         src/sh/sysv.S: Update license text.
1000
1001 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
1002
1003         * src/x86/win32.S (_ffi_closure_STDCALL):  New function.
1004         (.eh_frame):  Add FDE for it.
1005
1006 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
1007
1008         * configure.ac:  Also check if assembler supports pc-relative
1009         relocs on X86_WIN32 targets.
1010         * configure:  Regenerate.
1011         * src/x86/win32.S (ffi_prep_args):  Declare extern, not global.
1012         (_ffi_call_SYSV):  Add missing function type symbol .def and
1013         add EH markup labels.
1014         (_ffi_call_STDCALL):  Likewise.
1015         (_ffi_closure_SYSV):  Likewise.
1016         (_ffi_closure_raw_SYSV):  Likewise.
1017         (.eh_frame):  Add hand-crafted EH data.
1018
1019 2009-04-09  Jakub Jelinek  <jakub@redhat.com>
1020
1021         * testsuite/lib/libffi-dg.exp: Change copyright header to refer to
1022         version 3 of the GNU General Public License and to point readers
1023         at the COPYING3 file and the FSF's license web page.
1024         * testsuite/libffi.call/call.exp: Likewise.
1025         * testsuite/libffi.special/special.exp: Likewise.
1026
1027 2009-03-01  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1028
1029         * configure: Regenerate.
1030
1031 2008-12-18  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1032
1033         PR libffi/26048
1034         * configure.ac (HAVE_AS_X86_PCREL): New test.
1035         * configure: Regenerate.
1036         * fficonfig.h.in: Regenerate.
1037         * src/x86/sysv.S [!FFI_NO_RAW_API]: Precalculate
1038         RAW_CLOSURE_CIF_OFFSET, RAW_CLOSURE_FUN_OFFSET,
1039         RAW_CLOSURE_USER_DATA_OFFSET for the Solaris 10/x86 assembler.
1040         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
1041         * src/x86/unix64.S (.Lstore_table): Move to .text section.
1042         (.Lload_table): Likewise.
1043         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
1044
1045 2008-12-18  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1046
1047         * configure: Regenerate.
1048
1049 2008-11-21  Eric Botcazou  <ebotcazou@adacore.com>
1050
1051         * src/sparc/ffi.c (ffi_prep_cif_machdep): Add support for
1052         signed/unsigned int8/16 return values.
1053         * src/sparc/v8.S (ffi_call_v8): Likewise.
1054         (ffi_closure_v8): Likewise.
1055
1056 2008-09-26  Peter O'Gorman  <pogma@thewrittenword.com>
1057             Steve Ellcey  <sje@cup.hp.com>
1058
1059         * configure: Regenerate for new libtool.
1060         * Makefile.in: Ditto.
1061         * include/Makefile.in: Ditto.
1062         * aclocal.m4: Ditto.
1063
1064 2008-08-25  Andreas Tobler  <a.tobler@schweiz.org>
1065
1066         * src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and
1067         FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum.
1068         Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT.
1069         Adjust copyright notice.
1070         * src/powerpc/ffi.c: Add two new flags to indicate if we have one
1071         register or two register to use for FFI_SYSV structs.
1072         (ffi_prep_cif_machdep): Pass the right register flag introduced above.
1073         (ffi_closure_helper_SYSV): Fix the return type for
1074         FFI_SYSV_TYPE_SMALL_STRUCT. Comment.
1075         Adjust copyright notice.
1076
1077 2008-07-16  Kaz Kojima  <kkojima@gcc.gnu.org>
1078
1079         * src/sh/ffi.c (ffi_prep_closure_loc): Turn INSN into an unsigned
1080         int.
1081
1082 2008-06-17  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1083
1084         * configure: Regenerate.
1085         * include/Makefile.in: Regenerate.
1086         * testsuite/Makefile.in: Regenerate.
1087
1088 2008-06-07  Joseph Myers  <joseph@codesourcery.com>
1089
1090         * configure.ac (parisc*-*-linux*, powerpc-*-sysv*,
1091         powerpc-*-beos*): Remove.
1092         * configure: Regenerate.
1093
1094 2008-05-09  Julian Brown  <julian@codesourcery.com>
1095
1096         * Makefile.am (LTLDFLAGS): New.
1097         (libffi_la_LDFLAGS): Use above.
1098         * Makefile.in: Regenerate.
1099
1100 2008-04-18  Paolo Bonzini  <bonzini@gnu.org>
1101
1102         PR bootstrap/35457
1103         * aclocal.m4: Regenerate.
1104         * configure: Regenerate.
1105
1106 2008-03-26  Kaz Kojima  <kkojima@gcc.gnu.org>
1107
1108         * src/sh/sysv.S: Add .note.GNU-stack on Linux.
1109         * src/sh64/sysv.S: Likewise.
1110
1111 2008-03-26  Daniel Jacobowitz  <dan@debian.org>
1112
1113         * src/arm/sysv.S: Fix ARM comment marker.
1114
1115 2008-03-26  Jakub Jelinek  <jakub@redhat.com>
1116
1117         * src/alpha/osf.S: Add .note.GNU-stack on Linux.
1118         * src/s390/sysv.S: Likewise.
1119         * src/powerpc/ppc_closure.S: Likewise.
1120         * src/powerpc/sysv.S: Likewise.
1121         * src/x86/unix64.S: Likewise.
1122         * src/x86/sysv.S: Likewise.
1123         * src/sparc/v8.S: Likewise.
1124         * src/sparc/v9.S: Likewise.
1125         * src/m68k/sysv.S: Likewise.
1126         * src/arm/sysv.S: Likewise.
1127
1128 2008-03-16  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1129
1130         * aclocal.m4: Regenerate.
1131         * configure: Likewise.
1132         * Makefile.in: Likewise.
1133         * include/Makefile.in: Likewise.
1134         * testsuite/Makefile.in: Likewise.
1135
1136 2008-02-12  Bjoern Koenig  <bkoenig@alpha-tierchen.de>
1137             Andreas Tobler  <a.tobler@schweiz.org>
1138
1139         * configure.ac: Add amd64-*-freebsd* target.
1140         * configure: Regenerate.
1141
1142 2008-01-30  H.J. Lu  <hongjiu.lu@intel.com>
1143
1144         PR libffi/34612
1145         * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when
1146         returning struct.
1147
1148         * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer"
1149         tests.
1150
1151 2008-01-24  David Edelsohn  <edelsohn@gnu.org>
1152
1153         * configure: Regenerate.
1154
1155 2008-01-06  Andreas Tobler  <a.tobler@schweiz.org>
1156
1157         * src/x86/ffi.c (ffi_prep_cif_machdep): Fix thinko.
1158
1159 2008-01-05  Andreas Tobler  <a.tobler@schweiz.org>
1160
1161         PR testsuite/32843
1162         * src/x86/ffi.c (ffi_prep_cif_machdep): Add code for
1163         signed/unsigned int8/16 for X86_DARWIN.
1164         Updated copyright info.
1165         Handle one and two byte structs with special cif->flags.
1166         * src/x86/ffitarget.h: Add special types for one and two byte structs.
1167         Updated copyright info.
1168         * src/x86/darwin.S (ffi_call_SYSV): Rewrite to use a jump table like
1169         sysv.S
1170         Remove code to pop args from the stack after call.
1171         Special-case signed/unsigned for int8/16, one and two byte structs.
1172         (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1173         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1174         FFI_TYPE_SINT32.
1175         Updated copyright info.
1176
1177 2007-12-08  David Daney  <ddaney@avtrex.com>
1178
1179         * src/mips/n32.S (ffi_call_N32):  Replace dadd with ADDU, dsub with
1180         SUBU, add with ADDU and use smaller code sequences.
1181
1182 2007-12-07  David Daney  <ddaney@avtrex.com>
1183
1184         * src/mips/ffi.c (ffi_prep_cif_machdep): Handle long double return
1185         type.
1186
1187 2007-12-06  David Daney  <ddaney@avtrex.com>
1188
1189         * include/ffi.h.in (FFI_SIZEOF_JAVA_RAW): Define if not already
1190         defined.
1191         (ffi_java_raw): New typedef.
1192         (ffi_java_raw_call, ffi_java_ptrarray_to_raw,
1193         ffi_java_raw_to_ptrarray): Change parameter types from ffi_raw to
1194         ffi_java_raw.
1195         (ffi_java_raw_closure) : Same.
1196         (ffi_prep_java_raw_closure, ffi_prep_java_raw_closure_loc): Change
1197         parameter types.
1198         * src/java_raw_api.c (ffi_java_raw_size):  Replace FFI_SIZEOF_ARG with
1199         FFI_SIZEOF_JAVA_RAW.
1200         (ffi_java_raw_to_ptrarray): Change type of raw to ffi_java_raw.
1201         Replace FFI_SIZEOF_ARG with FFI_SIZEOF_JAVA_RAW. Use
1202         sizeof(ffi_java_raw) for alignment calculations.
1203         (ffi_java_ptrarray_to_raw): Same.
1204         (ffi_java_rvalue_to_raw): Add special handling for FFI_TYPE_POINTER
1205         if FFI_SIZEOF_JAVA_RAW == 4.
1206         (ffi_java_raw_to_rvalue): Same.
1207         (ffi_java_raw_call): Change type of raw to ffi_java_raw.
1208         (ffi_java_translate_args): Same.
1209         (ffi_prep_java_raw_closure_loc, ffi_prep_java_raw_closure): Change
1210         parameter types.
1211         * src/mips/ffitarget.h (FFI_SIZEOF_JAVA_RAW): Define for N32 ABI.
1212
1213 2007-12-06  David Daney  <ddaney@avtrex.com>
1214
1215         * src/mips/n32.S (ffi_closure_N32): Use 64-bit add instruction on
1216         pointer values.
1217
1218 2007-12-01  Andreas Tobler  <a.tobler@schweiz.org>
1219
1220         PR libffi/31937
1221         * src/powerpc/ffitarget.h: Introduce new ABI FFI_LINUX_SOFT_FLOAT.
1222         Add local FFI_TYPE_UINT128 to handle soft-float long-double-128.
1223         * src/powerpc/ffi.c: Distinguish between __NO_FPRS__ and not and
1224         set the NUM_FPR_ARG_REGISTERS according to.
1225         Add support for potential soft-float support under hard-float
1226         architecture.
1227         (ffi_prep_args_SYSV): Set NUM_FPR_ARG_REGISTERS to 0 in case of
1228         FFI_LINUX_SOFT_FLOAT, handle float, doubles and long-doubles according
1229         to the FFI_LINUX_SOFT_FLOAT ABI.
1230         (ffi_prep_cif_machdep): Likewise.
1231         (ffi_closure_helper_SYSV): Likewise.
1232         * src/powerpc/ppc_closure.S: Make sure not to store float/double
1233         on archs where __NO_FPRS__ is true.
1234         Add FFI_TYPE_UINT128 support.
1235         * src/powerpc/sysv.S: Add support for soft-float long-double-128.
1236         Adjust copyright notice.
1237
1238 2007-11-25  Andreas Tobler  <a.tobler@schweiz.org>
1239
1240         * src/closures.c: Move defintion of MAYBE_UNUSED from here to ...
1241         * include/ffi_common.h: ... here.
1242         Update copyright.
1243
1244 2007-11-17  Andreas Tobler  <a.tobler@schweiz.org>
1245
1246         * src/powerpc/sysv.S: Load correct cr to compare if we have long double.
1247         * src/powerpc/linux64.S: Likewise.
1248         * src/powerpc/ffi.c: Add a comment to show which part goes into cr6.
1249         * testsuite/libffi.call/return_ldl.c: New test.
1250
1251 2007-09-04    <aph@redhat.com>
1252
1253         * src/arm/sysv.S (UNWIND): New.
1254         (Whole file): Conditionally compile unwinder directives.
1255         * src/arm/sysv.S: Add unwinder directives.
1256
1257         * src/arm/ffi.c (ffi_prep_args): Align structs by at least 4 bytes.
1258         Only treat r0 as a struct address if we're actually returning a
1259         struct by address.
1260         Only copy the bytes that are actually within a struct.
1261         (ffi_prep_cif_machdep): A Composite Type not larger than 4 bytes
1262         is returned in r0, not passed by address.
1263         (ffi_call): Allocate a word-sized temporary for the case where
1264         a composite is returned in r0.
1265         (ffi_prep_incoming_args_SYSV): Align as necessary.
1266
1267 2007-08-05  Steven Newbury  <s_j_newbury@yahoo.co.uk>
1268
1269         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Use __clear_cache instead of
1270         directly using the sys_cacheflush syscall.
1271
1272 2007-07-27  Andrew Haley  <aph@redhat.com>
1273
1274         * src/arm/sysv.S (ffi_closure_SYSV): Add soft-float.
1275
1276 2007-09-03  Maciej W. Rozycki  <macro@linux-mips.org>
1277
1278         * Makefile.am: Unify MIPS_IRIX and MIPS_LINUX into MIPS.
1279         * configure.ac: Likewise.
1280         * Makefile.in: Regenerate.
1281         * include/Makefile.in: Likewise.
1282         * testsuite/Makefile.in: Likewise.
1283         * configure: Likewise.
1284
1285 2007-08-24  David Daney  <ddaney@avtrex.com>
1286
1287         * testsuite/libffi.call/return_sl.c: New test.
1288
1289 2007-08-10  David Daney  <ddaney@avtrex.com>
1290
1291         * testsuite/libffi.call/cls_multi_ushort.c,
1292         testsuite/libffi.call/cls_align_uint16.c,
1293         testsuite/libffi.call/nested_struct1.c,
1294         testsuite/libffi.call/nested_struct3.c,
1295         testsuite/libffi.call/cls_7_1_byte.c,
1296         testsuite/libffi.call/nested_struct5.c,
1297         testsuite/libffi.call/cls_double.c,
1298         testsuite/libffi.call/nested_struct7.c,
1299         testsuite/libffi.call/cls_sint.c,
1300         testsuite/libffi.call/nested_struct9.c,
1301         testsuite/libffi.call/cls_20byte1.c,
1302         testsuite/libffi.call/cls_multi_sshortchar.c,
1303         testsuite/libffi.call/cls_align_sint64.c,
1304         testsuite/libffi.call/cls_3byte2.c,
1305         testsuite/libffi.call/cls_multi_schar.c,
1306         testsuite/libffi.call/cls_multi_uchar.c,
1307         testsuite/libffi.call/cls_19byte.c,
1308         testsuite/libffi.call/cls_9byte1.c,
1309         testsuite/libffi.call/cls_align_float.c,
1310         testsuite/libffi.call/closure_fn1.c,
1311         testsuite/libffi.call/problem1.c,
1312         testsuite/libffi.call/closure_fn3.c,
1313         testsuite/libffi.call/cls_sshort.c,
1314         testsuite/libffi.call/closure_fn5.c,
1315         testsuite/libffi.call/cls_align_double.c,
1316         testsuite/libffi.call/nested_struct.c,
1317         testsuite/libffi.call/cls_2byte.c,
1318         testsuite/libffi.call/nested_struct10.c,
1319         testsuite/libffi.call/cls_4byte.c,
1320         testsuite/libffi.call/cls_6byte.c,
1321         testsuite/libffi.call/cls_8byte.c,
1322         testsuite/libffi.call/cls_multi_sshort.c,
1323         testsuite/libffi.call/cls_align_sint16.c,
1324         testsuite/libffi.call/cls_align_uint32.c,
1325         testsuite/libffi.call/cls_20byte.c,
1326         testsuite/libffi.call/cls_float.c,
1327         testsuite/libffi.call/nested_struct2.c,
1328         testsuite/libffi.call/cls_5_1_byte.c,
1329         testsuite/libffi.call/nested_struct4.c,
1330         testsuite/libffi.call/cls_24byte.c,
1331         testsuite/libffi.call/nested_struct6.c,
1332         testsuite/libffi.call/cls_64byte.c,
1333         testsuite/libffi.call/nested_struct8.c,
1334         testsuite/libffi.call/cls_uint.c,
1335         testsuite/libffi.call/cls_multi_ushortchar.c,
1336         testsuite/libffi.call/cls_schar.c,
1337         testsuite/libffi.call/cls_uchar.c,
1338         testsuite/libffi.call/cls_align_uint64.c,
1339         testsuite/libffi.call/cls_ulonglong.c,
1340         testsuite/libffi.call/cls_align_longdouble.c,
1341         testsuite/libffi.call/cls_1_1byte.c,
1342         testsuite/libffi.call/cls_12byte.c,
1343         testsuite/libffi.call/cls_3_1byte.c,
1344         testsuite/libffi.call/cls_3byte1.c,
1345         testsuite/libffi.call/cls_4_1byte.c,
1346         testsuite/libffi.call/cls_6_1_byte.c,
1347         testsuite/libffi.call/cls_16byte.c,
1348         testsuite/libffi.call/cls_18byte.c,
1349         testsuite/libffi.call/closure_fn0.c,
1350         testsuite/libffi.call/cls_9byte2.c,
1351         testsuite/libffi.call/closure_fn2.c,
1352         testsuite/libffi.call/closure_fn4.c,
1353         testsuite/libffi.call/cls_ushort.c,
1354         testsuite/libffi.call/closure_fn6.c,
1355         testsuite/libffi.call/cls_5byte.c,
1356         testsuite/libffi.call/cls_align_pointer.c,
1357         testsuite/libffi.call/cls_7byte.c,
1358         testsuite/libffi.call/cls_align_sint32.c,
1359         testsuite/libffi.special/unwindtest_ffi_call.cc,
1360         testsuite/libffi.special/unwindtest.cc: Remove xfail for mips64*-*-*.
1361
1362 2007-08-10  David Daney  <ddaney@avtrex.com>
1363
1364         PR libffi/28313
1365         * configure.ac: Don't treat mips64 as a special case.
1366         * Makefile.am (nodist_libffi_la_SOURCES): Add n32.S.
1367         * configure: Regenerate
1368         * Makefile.in: Ditto.
1369         * fficonfig.h.in: Ditto.
1370         * src/mips/ffitarget.h (REG_L, REG_S, SUBU, ADDU, SRL, LI): Indent.
1371         (LA, EH_FRAME_ALIGN, FDE_ADDR_BYTES): New preprocessor macros.
1372         (FFI_DEFAULT_ABI): Set for n64 case.
1373         (FFI_CLOSURES, FFI_TRAMPOLINE_SIZE): Define for n32 and n64 cases.
1374         * src/mips/n32.S (ffi_call_N32): Add debug macros and labels for FDE.
1375         (ffi_closure_N32): New function.
1376         (.eh_frame): New section
1377         * src/mips/o32.S: Clean up comments.
1378         (ffi_closure_O32): Pass ffi_closure parameter in $12.
1379         * src/mips/ffi.c: Use FFI_MIPS_N32 instead of
1380         _MIPS_SIM == _ABIN32 throughout.
1381         (FFI_MIPS_STOP_HERE): New, use in place of
1382         ffi_stop_here.
1383         (ffi_prep_args): Use unsigned long to hold pointer values.  Rewrite
1384         to support n32/n64 ABIs.
1385         (calc_n32_struct_flags): Rewrite.
1386         (calc_n32_return_struct_flags): Remove unused variable.  Reverse
1387         position of flag bits.
1388         (ffi_prep_cif_machdep): Rewrite n32 portion.
1389         (ffi_call): Enable for n64.  Add special handling for small structure
1390         return values.
1391         (ffi_prep_closure_loc): Add n32 and n64 support.
1392         (ffi_closure_mips_inner_O32): Add cast to silence warning.
1393         (copy_struct_N32, ffi_closure_mips_inner_N32): New functions.
1394
1395 2007-08-08  David Daney  <ddaney@avtrex.com>
1396
1397         * testsuite/libffi.call/ffitest.h (ffi_type_mylong): Remove definition.
1398         * testsuite/libffi.call/cls_align_uint16.c (main): Use correct type
1399         specifiers.
1400         * testsuite/libffi.call/nested_struct1.c (main): Ditto.
1401         * testsuite/libffi.call/cls_sint.c (main): Ditto.
1402         * testsuite/libffi.call/nested_struct9.c (main): Ditto.
1403         * testsuite/libffi.call/cls_20byte1.c (main): Ditto.
1404         * testsuite/libffi.call/cls_9byte1.c (main): Ditto.
1405         * testsuite/libffi.call/closure_fn1.c (main): Ditto.
1406         * testsuite/libffi.call/closure_fn3.c (main): Ditto.
1407         * testsuite/libffi.call/return_dbl2.c (main): Ditto.
1408         * testsuite/libffi.call/cls_sshort.c (main): Ditto.
1409         * testsuite/libffi.call/return_fl3.c (main): Ditto.
1410         * testsuite/libffi.call/closure_fn5.c (main): Ditto.
1411         * testsuite/libffi.call/nested_struct.c (main): Ditto.
1412         * testsuite/libffi.call/nested_struct10.c (main): Ditto.
1413         * testsuite/libffi.call/return_ll1.c (main): Ditto.
1414         * testsuite/libffi.call/cls_8byte.c (main): Ditto.
1415         * testsuite/libffi.call/cls_align_uint32.c (main): Ditto.
1416         * testsuite/libffi.call/cls_align_sint16.c (main): Ditto.
1417         * testsuite/libffi.call/cls_20byte.c (main): Ditto.
1418         * testsuite/libffi.call/nested_struct2.c (main): Ditto.
1419         * testsuite/libffi.call/cls_24byte.c (main): Ditto.
1420         * testsuite/libffi.call/nested_struct6.c (main): Ditto.
1421         * testsuite/libffi.call/cls_uint.c (main): Ditto.
1422         * testsuite/libffi.call/cls_12byte.c (main): Ditto.
1423         * testsuite/libffi.call/cls_16byte.c (main): Ditto.
1424         * testsuite/libffi.call/closure_fn0.c (main): Ditto.
1425         * testsuite/libffi.call/cls_9byte2.c (main): Ditto.
1426         * testsuite/libffi.call/closure_fn2.c (main): Ditto.
1427         * testsuite/libffi.call/return_dbl1.c (main): Ditto.
1428         * testsuite/libffi.call/closure_fn4.c (main): Ditto.
1429         * testsuite/libffi.call/closure_fn6.c (main): Ditto.
1430         * testsuite/libffi.call/cls_align_sint32.c (main): Ditto.
1431
1432 2007-08-07  Andrew Haley  <aph@redhat.com>
1433
1434         * src/x86/sysv.S (ffi_closure_raw_SYSV): Fix typo in previous
1435         checkin.
1436
1437 2007-08-06  Andrew Haley  <aph@redhat.com>
1438
1439         PR testsuite/32843
1440         * src/x86/sysv.S (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1441         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1442         FFI_TYPE_SINT32.
1443
1444 2007-08-02  David Daney  <ddaney@avtrex.com>
1445
1446         * testsuite/libffi.call/return_ul.c (main): Define return type as
1447         ffi_arg.  Use proper printf conversion specifier.
1448
1449 2007-07-30  Andrew Haley  <aph@redhat.com>
1450
1451         PR testsuite/32843
1452         * src/x86/ffi.c (ffi_prep_cif_machdep): in x86 case, add code for
1453         signed/unsigned int8/16.
1454         * src/x86/sysv.S (ffi_call_SYSV): Rewrite to:
1455         Use a jump table.
1456         Remove code to pop args from the stack after call.
1457         Special-case signed/unsigned int8/16.
1458         * testsuite/libffi.call/return_sc.c (main): Revert.
1459
1460 2007-07-26  Richard Guenther  <rguenther@suse.de>
1461
1462         PR testsuite/32843
1463         * testsuite/libffi.call/return_sc.c (main): Verify call
1464         result as signed char, not ffi_arg.
1465
1466 2007-07-16  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1467
1468         * configure.ac (i?86-*-solaris2.1[0-9]): Set TARGET to X86_64.
1469         * configure: Regenerate.
1470
1471 2007-07-11  David Daney  <ddaney@avtrex.com>
1472
1473         * src/mips/ffi.c: Don't include sys/cachectl.h.
1474         (ffi_prep_closure_loc): Use __builtin___clear_cache() instead of
1475         cacheflush().
1476
1477 2007-05-18  Aurelien Jarno  <aurelien@aurel32.net>
1478
1479         * src/arm/ffi.c (ffi_prep_closure_loc): Renamed and ajusted
1480         from (ffi_prep_closure): ... this.
1481         (FFI_INIT_TRAMPOLINE): Adjust.
1482
1483 2005-12-31  Phil Blundell  <pb@reciva.com>
1484
1485         * src/arm/ffi.c (ffi_prep_incoming_args_SYSV,
1486         ffi_closure_SYSV_inner, ffi_prep_closure): New, add closure support.
1487         * src/arm/sysv.S(ffi_closure_SYSV): Likewise.
1488         * src/arm/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1489         (FFI_CLOSURES): Enable closure support.
1490
1491 2007-07-03  Andrew Haley  <aph@hedges.billgatliff.com>
1492
1493         * testsuite/libffi.call/cls_multi_ushort.c,
1494         testsuite/libffi.call/cls_align_uint16.c,
1495         testsuite/libffi.call/nested_struct1.c,
1496         testsuite/libffi.call/nested_struct3.c,
1497         testsuite/libffi.call/cls_7_1_byte.c,
1498         testsuite/libffi.call/cls_double.c,
1499         testsuite/libffi.call/nested_struct5.c,
1500         testsuite/libffi.call/nested_struct7.c,
1501         testsuite/libffi.call/cls_sint.c,
1502         testsuite/libffi.call/nested_struct9.c,
1503         testsuite/libffi.call/cls_20byte1.c,
1504         testsuite/libffi.call/cls_multi_sshortchar.c,
1505         testsuite/libffi.call/cls_align_sint64.c,
1506         testsuite/libffi.call/cls_3byte2.c,
1507         testsuite/libffi.call/cls_multi_schar.c,
1508         testsuite/libffi.call/cls_multi_uchar.c,
1509         testsuite/libffi.call/cls_19byte.c,
1510         testsuite/libffi.call/cls_9byte1.c,
1511         testsuite/libffi.call/cls_align_float.c,
1512         testsuite/libffi.call/closure_fn1.c,
1513         testsuite/libffi.call/problem1.c,
1514         testsuite/libffi.call/closure_fn3.c,
1515         testsuite/libffi.call/cls_sshort.c,
1516         testsuite/libffi.call/closure_fn5.c,
1517         testsuite/libffi.call/cls_align_double.c,
1518         testsuite/libffi.call/cls_2byte.c,
1519         testsuite/libffi.call/nested_struct.c,
1520         testsuite/libffi.call/nested_struct10.c,
1521         testsuite/libffi.call/cls_4byte.c,
1522         testsuite/libffi.call/cls_6byte.c,
1523         testsuite/libffi.call/cls_8byte.c,
1524         testsuite/libffi.call/cls_multi_sshort.c,
1525         testsuite/libffi.call/cls_align_uint32.c,
1526         testsuite/libffi.call/cls_align_sint16.c,
1527         testsuite/libffi.call/cls_float.c,
1528         testsuite/libffi.call/cls_20byte.c,
1529         testsuite/libffi.call/cls_5_1_byte.c,
1530         testsuite/libffi.call/nested_struct2.c,
1531         testsuite/libffi.call/cls_24byte.c,
1532         testsuite/libffi.call/nested_struct4.c,
1533         testsuite/libffi.call/nested_struct6.c,
1534         testsuite/libffi.call/cls_64byte.c,
1535         testsuite/libffi.call/nested_struct8.c,
1536         testsuite/libffi.call/cls_uint.c,
1537         testsuite/libffi.call/cls_multi_ushortchar.c,
1538         testsuite/libffi.call/cls_schar.c,
1539         testsuite/libffi.call/cls_uchar.c,
1540         testsuite/libffi.call/cls_align_uint64.c,
1541         testsuite/libffi.call/cls_ulonglong.c,
1542         testsuite/libffi.call/cls_align_longdouble.c,
1543         testsuite/libffi.call/cls_1_1byte.c,
1544         testsuite/libffi.call/cls_12byte.c,
1545         testsuite/libffi.call/cls_3_1byte.c,
1546         testsuite/libffi.call/cls_3byte1.c,
1547         testsuite/libffi.call/cls_4_1byte.c,
1548         testsuite/libffi.call/cls_6_1_byte.c,
1549         testsuite/libffi.call/cls_16byte.c,
1550         testsuite/libffi.call/cls_18byte.c,
1551         testsuite/libffi.call/closure_fn0.c,
1552         testsuite/libffi.call/cls_9byte2.c,
1553         testsuite/libffi.call/closure_fn2.c,
1554         testsuite/libffi.call/closure_fn4.c,
1555         testsuite/libffi.call/cls_ushort.c,
1556         testsuite/libffi.call/closure_fn6.c,
1557         testsuite/libffi.call/cls_5byte.c,
1558         testsuite/libffi.call/cls_align_pointer.c,
1559         testsuite/libffi.call/cls_7byte.c,
1560         testsuite/libffi.call/cls_align_sint32.c,
1561         testsuite/libffi.special/unwindtest_ffi_call.cc,
1562         testsuite/libffi.special/unwindtest.cc: Enable for ARM.
1563
1564 2007-07-05  H.J. Lu  <hongjiu.lu@intel.com>
1565
1566         * aclocal.m4: Regenerated.
1567
1568 2007-06-02  Paolo Bonzini  <bonzini@gnu.org>
1569
1570         * configure: Regenerate.
1571
1572 2007-05-23  Steve Ellcey  <sje@cup.hp.com>
1573
1574         * Makefile.in: Regenerate.
1575         * configure: Regenerate.
1576         * aclocal.m4: Regenerate.
1577         * include/Makefile.in: Regenerate.
1578         * testsuite/Makefile.in: Regenerate.
1579
1580 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1581
1582         * src/m68k/ffi.c (ffi_prep_incoming_args_SYSV,
1583         ffi_closure_SYSV_inner,ffi_prep_closure): New, add closure support.
1584         * src/m68k/sysv.S(ffi_closure_SYSV,ffi_closure_struct_SYSV): Likewise.
1585         * src/m68k/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1586         (FFI_CLOSURES): Enable closure support.
1587
1588 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1589
1590         * configure.ac (HAVE_AS_CFI_PSEUDO_OP): New test.
1591         * configure: Regenerate.
1592         * fficonfig.h.in: Regenerate.
1593         * src/m68k/sysv.S (CFI_STARTPROC,CFI_ENDPROC,
1594         CFI_OFFSET,CFI_DEF_CFA): New macros.
1595         (ffi_call_SYSV): Add callframe annotation.
1596
1597 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1598
1599         * src/m68k/ffi.c (ffi_prep_args,ffi_prep_cif_machdep): Fix
1600         numerous test suite failures.
1601         * src/m68k/sysv.S (ffi_call_SYSV): Likewise.
1602
1603 2007-04-11  Paolo Bonzini  <bonzini@gnu.org>
1604
1605         * Makefile.am (EXTRA_DIST): Bring up to date.
1606         * Makefile.in: Regenerate.
1607         * src/frv/eabi.S: Remove RCS keyword.
1608
1609 2007-04-06  Richard Henderson  <rth@redhat.com>
1610
1611         * configure.ac: Tidy target case.
1612         (HAVE_LONG_DOUBLE): Allow the target to override.
1613         * configure: Regenerate.
1614         * include/ffi.h.in: Don't define ffi_type_foo if
1615         LIBFFI_HIDE_BASIC_TYPES is defined.
1616         (ffi_type_longdouble): If not HAVE_LONG_DOUBLE, define
1617         to ffi_type_double.
1618         * types.c (LIBFFI_HIDE_BASIC_TYPES): Define.
1619         (FFI_TYPEDEF, ffi_type_void): Mark the data const.
1620         (ffi_type_longdouble): Special case for Alpha.  Don't define
1621         if long double == double.
1622
1623         * src/alpha/ffi.c (FFI_TYPE_LONGDOUBLE): Assert unique value.
1624         (ffi_prep_cif_machdep): Handle it as the 128-bit type.
1625         (ffi_call, ffi_closure_osf_inner): Likewise.
1626         (ffi_closure_osf_inner): Likewise.  Mark hidden.
1627         (ffi_call_osf, ffi_closure_osf): Mark hidden.
1628         * src/alpha/ffitarget.h (FFI_LAST_ABI): Tidy definition.
1629         * src/alpha/osf.S (ffi_call_osf, ffi_closure_osf): Mark hidden.
1630         (load_table): Handle 128-bit long double.
1631
1632         * testsuite/libffi.call/float4.c: Add -mieee for alpha.
1633
1634 2007-04-06  Tom Tromey  <tromey@redhat.com>
1635
1636         PR libffi/31491:
1637         * README: Fixed bug in example.
1638
1639 2007-04-03  Jakub Jelinek  <jakub@redhat.com>
1640
1641         * src/closures.c: Include sys/statfs.h.
1642         (_GNU_SOURCE): Define on Linux.
1643         (FFI_MMAP_EXEC_SELINUX): Define.
1644         (selinux_enabled): New variable.
1645         (selinux_enabled_check): New function.
1646         (is_selinux_enabled): Define.
1647         (dlmmap): Use it.
1648
1649 2007-03-24  Uros Bizjak  <ubizjak@gmail.com>
1650
1651         * testsuite/libffi.call/return_fl2.c (return_fl): Mark as static.
1652         Use 'volatile float sum' to create sum of floats to avoid false
1653         negative due to excess precision on ix86 targets.
1654         (main): Ditto.
1655
1656 2007-03-08  Alexandre Oliva  <aoliva@redhat.com>
1657
1658         * src/powerpc/ffi.c (flush_icache): Fix left-over from previous
1659         patch.
1660         (ffi_prep_closure_loc): Remove unneeded casts.  Add needed ones.
1661
1662 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1663
1664         * include/ffi.h.in (ffi_closure_alloc, ffi_closure_free): New.
1665         (ffi_prep_closure_loc): New.
1666         (ffi_prep_raw_closure_loc): New.
1667         (ffi_prep_java_raw_closure_loc): New.
1668         * src/closures.c: New file.
1669         * src/dlmalloc.c [FFI_MMAP_EXEC_WRIT] (struct malloc_segment):
1670         Replace sflags with exec_offset.
1671         [FFI_MMAP_EXEC_WRIT] (mmap_exec_offset, add_segment_exec_offset,
1672         sub_segment_exec_offset): New macros.
1673         (get_segment_flags, set_segment_flags, check_segment_merge): New
1674         macros.
1675         (is_mmapped_segment, is_extern_segment): Use get_segment_flags.
1676         (add_segment, sys_alloc, create_mspace, create_mspace_with_base,
1677         destroy_mspace): Use new macros.
1678         (sys_alloc): Silence warning.
1679         * Makefile.am (libffi_la_SOURCES): Add src/closures.c.
1680         * Makefile.in: Rebuilt.
1681         * src/prep_cif [FFI_CLOSURES] (ffi_prep_closure): Implement in
1682         terms of ffi_prep_closure_loc.
1683         * src/raw_api.c (ffi_prep_raw_closure_loc): Renamed and adjusted
1684         from...
1685         (ffi_prep_raw_closure): ... this.  Re-implement in terms of the
1686         renamed version.
1687         * src/java_raw_api (ffi_prep_java_raw_closure_loc): Renamed and
1688         adjusted from...
1689         (ffi_prep_java_raw_closure): ... this.  Re-implement in terms of
1690         the renamed version.
1691         * src/alpha/ffi.c (ffi_prep_closure_loc): Renamed from
1692         (ffi_prep_closure): ... this.
1693         * src/pa/ffi.c: Likewise.
1694         * src/cris/ffi.c: Likewise.  Adjust.
1695         * src/frv/ffi.c: Likewise.
1696         * src/ia64/ffi.c: Likewise.
1697         * src/mips/ffi.c: Likewise.
1698         * src/powerpc/ffi_darwin.c: Likewise.
1699         * src/s390/ffi.c: Likewise.
1700         * src/sh/ffi.c: Likewise.
1701         * src/sh64/ffi.c: Likewise.
1702         * src/sparc/ffi.c: Likewise.
1703         * src/x86/ffi64.c: Likewise.
1704         * src/x86/ffi.c: Likewise.
1705         (FFI_INIT_TRAMPOLINE): Adjust.
1706         (ffi_prep_raw_closure_loc): Renamed and adjusted from...
1707         (ffi_prep_raw_closure): ... this.
1708         * src/powerpc/ffi.c (ffi_prep_closure_loc): Renamed from
1709         (ffi_prep_closure): ... this.
1710         (flush_icache): Adjust.
1711
1712 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1713
1714         * src/dlmalloc.c: New file, imported version 2.8.3 of Doug
1715         Lea's malloc.
1716
1717 2007-03-01  Brooks Moses  <brooks.moses@codesourcery.com>
1718
1719         * Makefile.am: Add dummy install-pdf target.
1720         * Makefile.in: Regenerate
1721
1722 2007-02-13  Andreas Krebbel  <krebbel1@de.ibm.com>
1723
1724         * src/s390/ffi.c (ffi_prep_args, ffi_prep_cif_machdep,
1725         ffi_closure_helper_SYSV): Add long double handling.
1726
1727 2007-02-02  Jakub Jelinek  <jakub@redhat.com>
1728
1729         * src/powerpc/linux64.S (ffi_call_LINUX64): Move restore of r2
1730         immediately after bctrl instruction.
1731
1732 2007-01-18  Alexandre Oliva  <aoliva@redhat.com>
1733
1734         * Makefile.am (all-recursive, install-recursive,
1735         mostlyclean-recursive, clean-recursive, distclean-recursive,
1736         maintainer-clean-recursive): Add missing targets.
1737         * Makefile.in: Rebuilt.
1738
1739 2006-12-14  Andreas Tobler  <a.tobler@schweiz.org>
1740
1741         * configure.ac: Add TARGET for x86_64-*-darwin*.
1742         * Makefile.am (nodist_libffi_la_SOURCES): Add rules for 64-bit sources
1743         for X86_DARWIN.
1744         * src/x86/ffitarget.h: Set trampoline size for x86_64-*-darwin*.
1745         * src/x86/darwin64.S: New file for x86_64-*-darwin* support.
1746         * configure: Regenerate.
1747         * Makefile.in: Regenerate.
1748         * include/Makefile.in: Regenerate.
1749         * testsuite/Makefile.in: Regenerate.
1750         * testsuite/libffi.special/unwindtest_ffi_call.cc: New test case for
1751         ffi_call only.
1752
1753 2006-12-13  Andreas Tobler <a.tobler@schweiz.org>
1754
1755         * aclocal.m4: Regenerate with aclocal -I .. as written in the
1756         Makefile.am.
1757
1758 2006-10-31  Geoffrey Keating  <geoffk@apple.com>
1759
1760         * src/powerpc/ffi_darwin.c (darwin_adjust_aggregate_sizes): New.
1761         (ffi_prep_cif_machdep): Call darwin_adjust_aggregate_sizes for
1762         Darwin.
1763         * testsuite/libffi.call/nested_struct4.c: Remove Darwin XFAIL.
1764         * testsuite/libffi.call/nested_struct6.c: Remove Darwin XFAIL.
1765
1766 2006-10-10  Paolo Bonzini  <bonzini@gnu.org>
1767             Sandro Tolaini  <tolaini@libero.it>
1768
1769         * configure.ac [i*86-*-darwin*]: Set X86_DARWIN symbol and
1770         conditional.
1771         * configure: Regenerated.
1772         * Makefile.am (nodist_libffi_la_SOURCES) [X86_DARWIN]: New case.
1773         (EXTRA_DIST): Add src/x86/darwin.S.
1774         * Makefile.in: Regenerated.
1775         * include/Makefile.in: Regenerated.
1776         * testsuite/Makefile.in: Regenerated.
1777
1778         * src/x86/ffi.c (ffi_prep_cif_machdep) [X86_DARWIN]: Treat like
1779         X86_WIN32, and additionally align stack to 16 bytes.
1780         * src/x86/darwin.S: New, based on sysv.S.
1781         * src/prep_cif.c (ffi_prep_cif) [X86_DARWIN]: Align > 8-byte structs.
1782
1783 2006-09-12  David Daney  <ddaney@avtrex.com>
1784
1785         PR libffi/23935
1786         * include/Makefile.am: Install both ffi.h and ffitarget.h in
1787         $(libdir)/gcc/$(target_alias)/$(gcc_version)/include.
1788         * aclocal.m4: Regenerated for automake 1.9.6.
1789         * Makefile.in: Regenerated.
1790         * include/Makefile.in: Regenerated.
1791         * testsuite/Makefile.in: Regenerated.
1792
1793 2006-08-17  Andreas Tobler  <a.tobler@schweiz.ch>
1794
1795         * include/ffi_common.h (struct): Revert accidental commit.
1796
1797 2006-08-15  Andreas Tobler  <a.tobler@schweiz.ch>
1798
1799         * include/ffi_common.h: Remove lint directives.
1800         * include/ffi.h.in: Likewise.
1801
1802 2006-07-25  Torsten Schoenfeld  <kaffeetisch@gmx.de>
1803
1804         * include/ffi.h.in (ffi_type_ulong, ffi_type_slong): Define correctly
1805         for 32-bit architectures.
1806         * testsuite/libffi.call/return_ul.c: New test case.
1807
1808 2006-07-19  David Daney  <ddaney@avtrex.com>
1809
1810         * testsuite/libffi.call/closure_fn6.c: Remove xfail for mips,
1811         xfail remains for mips64.
1812
1813 2006-05-23  Carlos O'Donell  <carlos@codesourcery.com>
1814
1815         * Makefile.am: Add install-html target. Add install-html to .PHONY
1816         * Makefile.in: Regenerate.
1817         * aclocal.m4: Regenerate.
1818         * include/Makefile.in: Regenerate.
1819         * testsuite/Makefile.in: Regenerate.
1820
1821 2006-05-18  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1822
1823         * pa/ffi.c (ffi_prep_args_pa32): Load floating point arguments from
1824         stack slot.
1825
1826 2006-04-22  Andreas Tobler  <a.tobler@schweiz.ch>
1827
1828         * README: Remove notice about 'Crazy Comments'.
1829         * src/debug.c: Remove lint directives. Cleanup white spaces.
1830         * src/java_raw_api.c: Likewise.
1831         * src/prep_cif.c: Likewise.
1832         * src/raw_api.c: Likewise.
1833         * src/ffitest.c: Delete. No longer needed, all test cases migrated
1834         to the testsuite.
1835         * src/arm/ffi.c: Remove lint directives.
1836         * src/m32r/ffi.c: Likewise.
1837         * src/pa/ffi.c: Likewise.
1838         * src/powerpc/ffi.c: Likewise.
1839         * src/powerpc/ffi_darwin.c: Likewise.
1840         * src/sh/ffi.c: Likewise.
1841         * src/sh64/ffi.c: Likewise.
1842         * src/x86/ffi.c: Likewise.
1843         * testsuite/libffi.call/float2.c: Likewise.
1844         * testsuite/libffi.call/promotion.c: Likewise.
1845         * testsuite/libffi.call/struct1.c: Likewise.
1846
1847 2006-04-13  Andreas Tobler  <a.tobler@schweiz.ch>
1848
1849         * src/pa/hpux32.S: Correct unwind offset calculation for
1850         ffi_closure_pa32.
1851         * src/pa/linux.S: Likewise.
1852
1853 2006-04-12  James E Wilson  <wilson@specifix.com>
1854
1855         PR libgcj/26483
1856         * src/ia64/ffi.c (stf_spill, ldf_fill): Rewrite as macros.
1857         (hfa_type_load): Call stf_spill.
1858         (hfa_type_store): Call ldf_fill.
1859         (ffi_call): Adjust calls to above routines.  Add local temps for
1860         macro result.
1861
1862 2006-04-10  Matthias Klose  <doko@debian.org>
1863
1864         * testsuite/lib/libffi-dg.exp (libffi-init): Recognize multilib
1865         directory names containing underscores.
1866
1867 2006-04-07  James E Wilson  <wilson@specifix.com>
1868
1869         * testsuite/libffi.call/float4.c: New testcase.
1870
1871 2006-04-05  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1872             Andreas Tobler  <a.tobler@schweiz.ch>
1873
1874         * Makefile.am: Add PA_HPUX port.
1875         * Makefile.in: Regenerate.
1876         * include/Makefile.in: Likewise.
1877         * testsuite/Makefile.in: Likewise.
1878         * configure.ac: Add PA_HPUX rules.
1879         * configure: Regenerate.
1880         * src/pa/ffitarget.h: Rename linux target to PA_LINUX.
1881         Add PA_HPUX and PA64_HPUX.
1882         Rename FFI_LINUX ABI to FFI_PA32 ABI.
1883         (FFI_TRAMPOLINE_SIZE): Define for 32-bit HP-UX targets.
1884         (FFI_TYPE_SMALL_STRUCT2): Define.
1885         (FFI_TYPE_SMALL_STRUCT4): Likewise.
1886         (FFI_TYPE_SMALL_STRUCT8): Likewise.
1887         (FFI_TYPE_SMALL_STRUCT3): Redefine.
1888         (FFI_TYPE_SMALL_STRUCT5): Likewise.
1889         (FFI_TYPE_SMALL_STRUCT6): Likewise.
1890         (FFI_TYPE_SMALL_STRUCT7): Likewise.
1891         * src/pa/ffi.c (ROUND_DOWN): Delete.
1892         (fldw, fstw, fldd, fstd): Use '__asm__'.
1893         (ffi_struct_type): Add support for FFI_TYPE_SMALL_STRUCT2,
1894         FFI_TYPE_SMALL_STRUCT4 and FFI_TYPE_SMALL_STRUCT8.
1895         (ffi_prep_args_LINUX): Rename to ffi_prep_args_pa32. Update comment.
1896         Simplify incrementing of stack slot variable. Change type of local
1897         'n' to unsigned int.
1898         (ffi_size_stack_LINUX): Rename to ffi_size_stack_pa32. Handle long
1899         double on PA_HPUX.
1900         (ffi_prep_cif_machdep): Likewise.
1901         (ffi_call): Likewise.
1902         (ffi_closure_inner_LINUX): Rename to ffi_closure_inner_pa32. Change
1903         return type to ffi_status. Simplify incrementing of stack slot
1904         variable. Only copy floating point argument registers when PA_LINUX
1905         is true. Reformat debug statement.
1906         Add support for FFI_TYPE_SMALL_STRUCT2, FFI_TYPE_SMALL_STRUCT4 and
1907         FFI_TYPE_SMALL_STRUCT8.
1908         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Add 'extern' to
1909         declaration.
1910         (ffi_prep_closure): Make linux trampoline conditional on PA_LINUX.
1911         Add nops to cache flush.  Add trampoline for PA_HPUX.
1912         * src/pa/hpux32.S: New file.
1913         * src/pa/linux.S (ffi_call_LINUX): Rename to ffi_call_pa32. Rename
1914         ffi_prep_args_LINUX to ffi_prep_args_pa32.
1915         Localize labels. Add support for 2, 4 and 8-byte small structs. Handle
1916         unaligned destinations in 3, 5, 6 and 7-byte small structs. Order
1917         argument type checks so that common argument types appear first.
1918         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Rename
1919         ffi_closure_inner_LINUX to ffi_closure_inner_pa32.
1920
1921 2006-03-24  Alan Modra  <amodra@bigpond.net.au>
1922
1923         * src/powerpc/ffitarget.h (enum ffi_abi): Add FFI_LINUX.  Default
1924         for 32-bit using IBM extended double format.  Fix FFI_LAST_ABI.
1925         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Handle linux variant of
1926         FFI_TYPE_LONGDOUBLE.
1927         (ffi_prep_args64): Assert using IBM extended double.
1928         (ffi_prep_cif_machdep): Don't munge FFI_TYPE_LONGDOUBLE type.
1929         Handle FFI_LINUX FFI_TYPE_LONGDOUBLE return and args.
1930         (ffi_call): Handle FFI_LINUX.
1931         (ffi_closure_helper_SYSV): Non FFI_LINUX long double return needs
1932         gpr3 return pointer as for struct return.  Handle FFI_LINUX
1933         FFI_TYPE_LONGDOUBLE return and args.  Don't increment "nf"
1934         unnecessarily.
1935         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Load both f1 and f2
1936         for FFI_TYPE_LONGDOUBLE.  Move epilogue insns into case table.
1937         Don't use r6 as pointer to results, instead use sp offset.  Don't
1938         make a special call to load lr with case table address, instead
1939         use offset from previous call.
1940         * src/powerpc/sysv.S (ffi_call_SYSV): Save long double return.
1941         * src/powerpc/linux64.S (ffi_call_LINUX64): Simplify long double
1942         return.
1943
1944 2006-03-15  Kaz Kojima  <kkojima@gcc.gnu.org>
1945
1946         * src/sh64/ffi.c (ffi_prep_cif_machdep): Handle float arguments
1947         passed with FP registers correctly.
1948         (ffi_closure_helper_SYSV): Likewise.
1949         * src/sh64/sysv.S: Likewise.
1950
1951 2006-03-01  Andreas Tobler  <a.tobler@schweiz.ch>
1952
1953         * testsuite/libffi.special/unwindtest.cc (closure_test_fn): Mark cif,
1954         args and userdata unused.
1955         (closure_test_fn1): Mark cif and userdata unused.
1956         (main): Remove unused res.
1957
1958 2006-02-28  Andreas Tobler  <a.tobler@schweiz.ch>
1959
1960         * testsuite/libffi.call/call.exp: Adjust FSF address. Add test runs for
1961         -O2, -O3, -Os and the warning flags -W -Wall.
1962         * testsuite/libffi.special/special.exp: Likewise.
1963         * testsuite/libffi.call/ffitest.h: Add an __UNUSED__ macro to mark
1964         unused parameter unused for gcc or else do nothing.
1965         * testsuite/libffi.special/ffitestcxx.h: Likewise.
1966         * testsuite/libffi.call/cls_12byte.c (cls_struct_12byte_gn): Mark cif
1967         and userdata unused.
1968         * testsuite/libffi.call/cls_16byte.c (cls_struct_16byte_gn): Likewise.
1969         * testsuite/libffi.call/cls_18byte.c (cls_struct_18byte_gn): Likewise.
1970         * testsuite/libffi.call/cls_19byte.c (cls_struct_19byte_gn): Likewise.
1971         * testsuite/libffi.call/cls_1_1byte.c (cls_struct_1_1byte_gn): Likewise.
1972         * testsuite/libffi.call/cls_20byte.c (cls_struct_20byte_gn): Likewise.
1973         * testsuite/libffi.call/cls_20byte1.c (cls_struct_20byte_gn): Likewise.
1974         * testsuite/libffi.call/cls_24byte.c (cls_struct_24byte_gn): Likewise.
1975         * testsuite/libffi.call/cls_2byte.c (cls_struct_2byte_gn): Likewise.
1976         * testsuite/libffi.call/cls_3_1byte.c (cls_struct_3_1byte_gn): Likewise.
1977         * testsuite/libffi.call/cls_3byte1.c (cls_struct_3byte_gn): Likewise.
1978         * testsuite/libffi.call/cls_3byte2.c (cls_struct_3byte_gn1): Likewise.
1979         * testsuite/libffi.call/cls_4_1byte.c (cls_struct_4_1byte_gn): Likewise.
1980         * testsuite/libffi.call/cls_4byte.c (cls_struct_4byte_gn): Likewise.
1981         * testsuite/libffi.call/cls_5_1_byte.c (cls_struct_5byte_gn): Likewise.
1982         * testsuite/libffi.call/cls_5byte.c (cls_struct_5byte_gn): Likewise.
1983         * testsuite/libffi.call/cls_64byte.c (cls_struct_64byte_gn): Likewise.
1984         * testsuite/libffi.call/cls_6_1_byte.c (cls_struct_6byte_gn): Likewise.
1985         * testsuite/libffi.call/cls_6byte.c (cls_struct_6byte_gn): Likewise.
1986         * testsuite/libffi.call/cls_7_1_byte.c (cls_struct_7byte_gn): Likewise.
1987         * testsuite/libffi.call/cls_7byte.c (cls_struct_7byte_gn): Likewise.
1988         * testsuite/libffi.call/cls_8byte.c (cls_struct_8byte_gn): Likewise.
1989         * testsuite/libffi.call/cls_9byte1.c (cls_struct_9byte_gn): Likewise.
1990         * testsuite/libffi.call/cls_9byte2.c (cls_struct_9byte_gn): Likewise.
1991         * testsuite/libffi.call/cls_align_double.c (cls_struct_align_gn):
1992         Likewise.
1993         * testsuite/libffi.call/cls_align_float.c (cls_struct_align_gn):
1994         Likewise.
1995         * testsuite/libffi.call/cls_align_longdouble.c (cls_struct_align_gn):
1996         Likewise.
1997         * testsuite/libffi.call/cls_align_pointer.c (cls_struct_align_fn): Cast
1998         void* to avoid compiler warning.
1999         (main): Likewise.
2000         (cls_struct_align_gn): Mark cif and userdata unused.
2001         * testsuite/libffi.call/cls_align_sint16.c (cls_struct_align_gn):
2002         Likewise.
2003         * testsuite/libffi.call/cls_align_sint32.c (cls_struct_align_gn):
2004         Likewise.
2005         * testsuite/libffi.call/cls_align_sint64.c (cls_struct_align_gn):
2006         Likewise.
2007         * testsuite/libffi.call/cls_align_uint16.c (cls_struct_align_gn):
2008         Likewise.
2009         * testsuite/libffi.call/cls_align_uint32.c (cls_struct_align_gn):
2010         Likewise.
2011         * testsuite/libffi.call/cls_double.c (cls_ret_double_fn): Likewise.
2012         * testsuite/libffi.call/cls_float.c (cls_ret_float_fn): Likewise.
2013         * testsuite/libffi.call/cls_multi_schar.c (test_func_gn): Mark cif and
2014         data unused.
2015         (main): Cast res_call to silence gcc.
2016         * testsuite/libffi.call/cls_multi_sshort.c (test_func_gn): Mark cif and
2017         data unused.
2018         (main): Cast res_call to silence gcc.
2019         * testsuite/libffi.call/cls_multi_sshortchar.c (test_func_gn): Mark cif
2020         and data unused.
2021         (main): Cast res_call to silence gcc.
2022         * testsuite/libffi.call/cls_multi_uchar.c (test_func_gn): Mark cif and
2023         data unused.
2024         (main): Cast res_call to silence gcc.
2025         * testsuite/libffi.call/cls_multi_ushort.c (test_func_gn): Mark cif and
2026         data unused.
2027         (main): Cast res_call to silence gcc.
2028         * testsuite/libffi.call/cls_multi_ushortchar.c (test_func_gn): Mark cif
2029         and data unused.
2030         (main): Cast res_call to silence gcc.
2031         * testsuite/libffi.call/cls_schar.c (cls_ret_schar_fn): Mark cif and
2032         userdata unused.
2033         (cls_ret_schar_fn): Cast printf parameter to silence gcc.
2034         * testsuite/libffi.call/cls_sint.c (cls_ret_sint_fn): Mark cif and
2035         userdata unused.
2036         (cls_ret_sint_fn): Cast printf parameter to silence gcc.
2037         * testsuite/libffi.call/cls_sshort.c (cls_ret_sshort_fn): Mark cif and
2038         userdata unused.
2039         (cls_ret_sshort_fn): Cast printf parameter to silence gcc.
2040         * testsuite/libffi.call/cls_uchar.c (cls_ret_uchar_fn):  Mark cif and
2041         userdata unused.
2042         (cls_ret_uchar_fn): Cast printf parameter to silence gcc.
2043         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Mark cif and
2044         userdata unused.
2045         (cls_ret_uint_fn): Cast printf parameter to silence gcc.
2046         * testsuite/libffi.call/cls_ulonglong.c (cls_ret_ulonglong_fn): Mark cif
2047         and userdata unused.
2048         * testsuite/libffi.call/cls_ushort.c (cls_ret_ushort_fn): Mark cif and
2049         userdata unused.
2050         (cls_ret_ushort_fn): Cast printf parameter to silence gcc.
2051         * testsuite/libffi.call/float.c (floating): Remove unused parameter e.
2052         * testsuite/libffi.call/float1.c (main): Remove unused variable i.
2053         Cleanup white spaces.
2054         * testsuite/libffi.call/negint.c (checking): Remove unused variable i.
2055         * testsuite/libffi.call/nested_struct.c (cls_struct_combined_gn): Mark
2056         cif and userdata unused.
2057         * testsuite/libffi.call/nested_struct1.c (cls_struct_combined_gn):
2058         Likewise.
2059         * testsuite/libffi.call/nested_struct10.c (B_gn): Likewise.
2060         * testsuite/libffi.call/nested_struct2.c (B_fn): Adjust printf
2061         formatters to silence gcc.
2062         (B_gn): Mark cif and userdata unused.
2063         * testsuite/libffi.call/nested_struct3.c (B_gn): Mark cif and userdata
2064         unused.
2065         * testsuite/libffi.call/nested_struct4.c: Mention related PR.
2066         (B_gn): Mark cif and userdata unused.
2067         * testsuite/libffi.call/nested_struct5.c (B_gn): Mark cif and userdata
2068         unused.
2069         * testsuite/libffi.call/nested_struct6.c: Mention related PR.
2070         (B_gn): Mark cif and userdata unused.
2071         * testsuite/libffi.call/nested_struct7.c (B_gn): Mark cif and userdata
2072         unused.
2073         * testsuite/libffi.call/nested_struct8.c (B_gn): Likewise.
2074         * testsuite/libffi.call/nested_struct9.c (B_gn): Likewise.
2075         * testsuite/libffi.call/problem1.c (stub): Likewise.
2076         * testsuite/libffi.call/pyobjc-tc.c (main): Cast the result to silence
2077         gcc.
2078         * testsuite/libffi.call/return_fl2.c (return_fl): Add the note mentioned
2079         in the last commit for this test case in the test case itself.
2080         * testsuite/libffi.call/closure_fn0.c (closure_test_fn0): Mark cif as
2081         unused.
2082         * testsuite/libffi.call/closure_fn1.c (closure_test_fn1): Likewise.
2083         * testsuite/libffi.call/closure_fn2.c (closure_test_fn2): Likewise.
2084         * testsuite/libffi.call/closure_fn3.c (closure_test_fn3): Likewise.
2085         * testsuite/libffi.call/closure_fn4.c (closure_test_fn0): Likewise.
2086         * testsuite/libffi.call/closure_fn5.c (closure_test_fn5): Likewise.
2087         * testsuite/libffi.call/closure_fn6.c (closure_test_fn0): Likewise.
2088
2089 2006-02-22  Kaz Kojima  <kkojima@gcc.gnu.org>
2090
2091         * src/sh/sysv.S: Fix register numbers in the FDE for
2092         ffi_closure_SYSV.
2093
2094 2006-02-20  Andreas Tobler  <a.tobler@schweiz.ch>
2095
2096         * testsuite/libffi.call/return_fl2.c (return_fl): Remove static
2097         declaration to avoid a false negative on ix86. See PR323.
2098
2099 2006-02-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2100
2101         * src/sh/ffi.c (ffi_closure_helper_SYSV): Remove unused variable
2102         and cast integer to void * if needed.  Update the pointer to
2103         the FP register saved area correctly.
2104
2105 2006-02-17  Andreas Tobler  <a.tobler@schweiz.ch>
2106
2107         * testsuite/libffi.call/nested_struct6.c: XFAIL this test until PR25630
2108         is fixed.
2109         * testsuite/libffi.call/nested_struct4.c: Likewise.
2110
2111 2006-02-16  Andreas Tobler  <a.tobler@schweiz.ch>
2112
2113         * testsuite/libffi.call/return_dbl.c: New test case.
2114         * testsuite/libffi.call/return_dbl1.c: Likewise.
2115         * testsuite/libffi.call/return_dbl2.c: Likewise.
2116         * testsuite/libffi.call/return_fl.c: Likewise.
2117         * testsuite/libffi.call/return_fl1.c: Likewise.
2118         * testsuite/libffi.call/return_fl2.c: Likewise.
2119         * testsuite/libffi.call/return_fl3.c: Likewise.
2120         * testsuite/libffi.call/closure_fn6.c: Likewise.
2121
2122         * testsuite/libffi.call/nested_struct2.c: Remove ffi_type_mylong
2123         definition.
2124         * testsuite/libffi.call/ffitest.h: Add ffi_type_mylong definition
2125         here to be used by other test cases too.
2126
2127         * testsuite/libffi.call/nested_struct10.c: New test case.
2128         * testsuite/libffi.call/nested_struct9.c: Likewise.
2129         * testsuite/libffi.call/nested_struct8.c: Likewise.
2130         * testsuite/libffi.call/nested_struct7.c: Likewise.
2131         * testsuite/libffi.call/nested_struct6.c: Likewise.
2132         * testsuite/libffi.call/nested_struct5.c: Likewise.
2133         * testsuite/libffi.call/nested_struct4.c: Likewise.
2134
2135 2006-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
2136
2137         * configure.ac: Enable libffi for sparc64-*-freebsd*.
2138         * configure: Rebuilt.
2139
2140 2006-01-18  Jakub Jelinek  <jakub@redhat.com>
2141
2142         * src/powerpc/sysv.S (smst_two_register): Don't call __ashldi3,
2143         instead do the shifting inline.
2144         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't compute %r5
2145         shift count unconditionally.  Simplify load sequences for 1, 2, 3, 4
2146         and 8 byte structs, for the remaining struct sizes don't call
2147         __lshrdi3, instead do the shifting inline.
2148
2149 2005-12-07  Thiemo Seufer  <ths@networkno.de>
2150
2151         * src/mips/ffitarget.h: Remove obsolete sgidefs.h include. Add
2152         missing parentheses.
2153         * src/mips/o32.S (ffi_call_O32): Code formatting. Define
2154         and use A3_OFF, FP_OFF, RA_OFF. Micro-optimizations.
2155         (ffi_closure_O32): Likewise, but with newly defined A3_OFF2,
2156         A2_OFF2, A1_OFF2, A0_OFF2, RA_OFF2, FP_OFF2, S0_OFF2, GP_OFF2,
2157         V1_OFF2, V0_OFF2, FA_1_1_OFF2, FA_1_0_OFF2, FA_0_1_OFF2,
2158         FA_0_0_OFF2.
2159         * src/mips/ffi.c (ffi_prep_args): Code formatting. Fix
2160         endianness bugs.
2161         (ffi_prep_closure): Improve trampoline instruction scheduling.
2162         (ffi_closure_mips_inner_O32): Fix endianness bugs.
2163
2164 2005-12-03  Alan Modra  <amodra@bigpond.net.au>
2165
2166         * src/powerpc/ffi.c: Formatting.
2167         (ffi_prep_args_SYSV): Avoid possible aliasing problems by using unions.
2168         (ffi_prep_args64): Likewise.
2169
2170 2005-09-30  Geoffrey Keating  <geoffk@apple.com>
2171
2172         * testsuite/lib/libffi-dg.exp (libffi_target_compile): For
2173         darwin, use -shared-libgcc not -lgcc_s, and explain why.
2174
2175 2005-09-26  Tom Tromey  <tromey@redhat.com>
2176
2177         * testsuite/libffi.call/float1.c (value_type): New typedef.
2178         (CANARY): New define.
2179         (main): Check for result buffer overflow.
2180         * src/powerpc/linux64.S: Handle linux64 long double returns.
2181         * src/powerpc/ffi.c (FLAG_RETURNS_128BITS): New constant.
2182         (ffi_prep_cif_machdep): Handle linux64 long double returns.
2183
2184 2005-08-25  Alan Modra  <amodra@bigpond.net.au>
2185
2186         PR target/23404
2187         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Correct placement of stack
2188         homed fp args.
2189         (ffi_status ffi_prep_cif_machdep): Correct stack sizing for same.
2190
2191 2005-08-11  Jakub Jelinek  <jakub@redhat.com>
2192
2193         * configure.ac (HAVE_HIDDEN_VISIBILITY_ATTRIBUTE): New test.
2194         (AH_BOTTOM): Add FFI_HIDDEN definition.
2195         * configure: Rebuilt.
2196         * fficonfig.h.in: Rebuilt.
2197         * src/powerpc/ffi.c (hidden): Remove.
2198         (ffi_closure_LINUX64, ffi_prep_args64, ffi_call_LINUX64,
2199         ffi_closure_helper_LINUX64): Use FFI_HIDDEN instead of hidden.
2200         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64,
2201         .ffi_closure_LINUX64): Use FFI_HIDDEN instead of .hidden.
2202         * src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV): Remove,
2203         add FFI_HIDDEN to its prototype.
2204         (ffi_closure_SYSV_inner): New.
2205         * src/x86/sysv.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2206         * src/x86/win32.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2207
2208 2005-08-10  Alfred M. Szmidt  <ams@gnu.org>
2209
2210         PR libffi/21819:
2211         * configure: Rebuilt.
2212         * configure.ac: Handle i*86-*-gnu*.
2213
2214 2005-08-09  Jakub Jelinek  <jakub@redhat.com>
2215
2216         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Use
2217         DW_CFA_offset_extended_sf rather than
2218         DW_CFA_GNU_negative_offset_extended.
2219         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise.
2220
2221 2005-07-22  SUGIOKA Toshinobu  <sugioka@itonet.co.jp>
2222
2223         * src/sh/sysv.S (ffi_call_SYSV): Stop argument popping correctly
2224         on sh3.
2225         (ffi_closure_SYSV): Change the stack layout for sh3 struct argument.
2226         * src/sh/ffi.c (ffi_prep_args): Fix sh3 argument copy, when it is
2227         partially on register.
2228         (ffi_closure_helper_SYSV): Likewise.
2229         (ffi_prep_cif_machdep): Don't set too many cif->flags.
2230
2231 2005-07-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2232
2233         * src/sh/ffi.c (ffi_call): Handle small structures correctly.
2234         Remove empty line.
2235         * src/sh64/ffi.c (simple_type): Remove.
2236         (return_type): Handle small structures correctly.
2237         (ffi_prep_args): Likewise.
2238         (ffi_call): Likewise.
2239         (ffi_closure_helper_SYSV): Likewise.
2240         * src/sh64/sysv.S (ffi_call_SYSV): Handle 1, 2 and 4-byte return.
2241         Emit position independent code if PIC and remove wrong datalabel
2242         prefixes from EH data.
2243
2244 2005-07-19  Andreas Tobler  <a.tobler@schweiz.ch>
2245
2246         * Makefile.am (nodist_libffi_la_SOURCES): Add POWERPC_FREEBSD.
2247         * Makefile.in: Regenerate.
2248         * include/Makefile.in: Likewise.
2249         * testsuite/Makefile.in: Likewise.
2250         * configure.ac: Add POWERPC_FREEBSD rules.
2251         * configure: Regenerate.
2252         * src/powerpc/ffitarget.h: Add POWERPC_FREEBSD rules.
2253         (FFI_SYSV_TYPE_SMALL_STRUCT): Define.
2254         * src/powerpc/ffi.c: Add flags to handle small structure returns
2255         in ffi_call_SYSV.
2256         (ffi_prep_cif_machdep): Handle small structures for SYSV 4 ABI.
2257         Aka FFI_SYSV.
2258         (ffi_closure_helper_SYSV): Likewise.
2259         * src/powerpc/ppc_closure.S: Add return types for small structures.
2260         * src/powerpc/sysv.S: Add bits to handle small structures for
2261         final SYSV 4 ABI.
2262
2263 2005-07-10  Andreas Tobler  <a.tobler@schweiz.ch>
2264
2265         * testsuite/libffi.call/cls_5_1_byte.c: New test file.
2266         * testsuite/libffi.call/cls_6_1_byte.c: Likewise.
2267         * testsuite/libffi.call/cls_7_1_byte.c: Likewise.
2268
2269 2005-07-05  Randolph Chung  <tausq@debian.org>
2270
2271         * src/pa/ffi.c (ffi_struct_type): Rename FFI_TYPE_SMALL_STRUCT1
2272         as FFI_TYPE_SMALL_STRUCT3.  Break out handling for 5-7 byte
2273         structures.  Kill compilation warnings.
2274         (ffi_closure_inner_LINUX): Print return values as hex in debug
2275         message.  Rename FFI_TYPE_SMALL_STRUCT1 as FFI_TYPE_SMALL_STRUCT3.
2276         Properly handle 5-7 byte structure returns.
2277         * src/pa/ffitarget.h (FFI_TYPE_SMALL_STRUCT1)
2278         (FFI_TYPE_SMALL_STRUCT2): Remove.
2279         (FFI_TYPE_SMALL_STRUCT3, FFI_TYPE_SMALL_STRUCT5)
2280         (FFI_TYPE_SMALL_STRUCT6, FFI_TYPE_SMALL_STRUCT7): Define.
2281         * src/pa/linux.S: Mark source file as using PA1.1 assembly.
2282         (checksmst1, checksmst2): Remove.
2283         (checksmst3): Optimize handling of 3-byte struct returns.
2284         (checksmst567): Properly handle 5-7 byte struct returns.
2285
2286 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2287
2288         PR libgcj/21943
2289         * src/mips/n32.S: Enforce PIC code.
2290         * src/mips/o32.S: Likewise.
2291
2292 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2293
2294         * configure.ac: Treat i*86-*-solaris2.10 and up as X86_64.
2295         * configure: Regenerate.
2296
2297 2005-06-01  Alan Modra  <amodra@bigpond.net.au>
2298
2299         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't use JUMPTARGET
2300         to call ffi_closure_helper_SYSV.  Append @local instead.
2301         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise for ffi_prep_args_SYSV.
2302
2303 2005-05-17  Kelley Cook  <kcook@gcc.gnu.org>
2304
2305         * configure.ac: Use AC_C_BIGENDIAN instead of AC_C_BIGENDIAN_CROSS.
2306         Use AC_CHECK_SIZEOF instead of AC_COMPILE_CHECK_SIZEOF.
2307         * Makefile.am (ACLOCAL_AMFLAGS): Remove -I ../config.
2308         * aclocal.m4, configure, fficonfig.h.in, Makefile.in,
2309         include/Makefile.in, testsuite/Makefile.in: Regenerate.
2310
2311 2005-05-09  Mike Stump  <mrs@apple.com>
2312
2313         * configure: Regenerate.
2314
2315 2005-05-08  Richard Henderson  <rth@redhat.com>
2316
2317         PR libffi/21285
2318         * src/alpha/osf.S: Update unwind into to match code.
2319
2320 2005-05-04  Andreas Degert <ad@papyrus-gmbh.de>
2321             Richard Henderson  <rth@redhat.com>
2322
2323         * src/x86/ffi64.c (ffi_prep_cif_machdep): Save sse-used flag in
2324         bit 11 of flags.
2325         (ffi_call): Mask return type field.  Pass ssecount to ffi_call_unix64.
2326         (ffi_prep_closure): Set carry bit if sse-used flag set.
2327         * src/x86/unix64.S (ffi_call_unix64): Add ssecount argument.
2328         Only load sse registers if ssecount non-zero.
2329         (ffi_closure_unix64): Only save sse registers if carry set on entry.
2330
2331 2005-04-29  Ralf Corsepius  <ralf.corsepius@rtems.org>
2332
2333         * configure.ac: Add i*86-*-rtems*, sparc*-*-rtems*,
2334         powerpc-*rtems*, arm*-*-rtems*, sh-*-rtems*.
2335         * configure: Regenerate.
2336
2337 2005-04-20  Hans-Peter Nilsson  <hp@axis.com>
2338
2339         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): In regsub use,
2340         have Tcl8.3-compatible intermediate variable.
2341
2342 2005-04-18  Simon Posnjak <simon.posnjak@siol.net>
2343             Hans-Peter Nilsson  <hp@axis.com>
2344
2345         * Makefile.am: Add CRIS support.
2346         * configure.ac: Likewise.
2347         * Makefile.in, configure, testsuite/Makefile.in,
2348         include/Makefile.in: Regenerate.
2349         * src/cris: New directory.
2350         * src/cris/ffi.c, src/cris/sysv.S, src/cris/ffitarget.h: New files.
2351         * src/prep_cif.c (ffi_prep_cif): Wrap in #ifndef __CRIS__.
2352
2353         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): Replace \n with
2354         \r?\n in output tests.
2355
2356 2005-04-12  Mike Stump  <mrs@apple.com>
2357
2358         * configure: Regenerate.
2359
2360 2005-03-30  Hans Boehm  <Hans.Boehm@hp.com>
2361
2362         * src/ia64/ffitarget.h (ffi_arg): Use long long instead of DI.
2363
2364 2005-03-30  Steve Ellcey  <sje@cup.hp.com>
2365
2366         * src/ia64/ffitarget.h (ffi_arg) ADD DI attribute.
2367         (ffi_sarg) Ditto.
2368         * src/ia64/unix.S (ffi_closure_unix): Extend gp
2369         to 64 bits in ILP32 mode.
2370         Load 64 bits even for short data.
2371
2372 2005-03-23  Mike Stump  <mrs@apple.com>
2373
2374         * src/powerpc/darwin.S: Update for -m64 multilib.
2375         * src/powerpc/darwin_closure.S: Likewise.
2376
2377 2005-03-21  Zack Weinberg  <zack@codesourcery.com>
2378
2379         * configure.ac: Do not invoke TL_AC_GCC_VERSION.
2380         Do not set tool_include_dir.
2381         * aclocal.m4, configure, Makefile.in, testsuite/Makefile.in:
2382         Regenerate.
2383         * include/Makefile.am: Set gcc_version and toollibffidir.
2384         * include/Makefile.in: Regenerate.
2385
2386 2005-02-22  Andrew Haley  <aph@redhat.com>
2387
2388         * src/powerpc/ffi.c (ffi_prep_cif_machdep): Bump alignment to
2389         odd-numbered register pairs for 64-bit integer types.
2390
2391 2005-02-23  Andreas Tobler  <a.tobler@schweiz.ch>
2392
2393         PR libffi/20104
2394         * testsuite/libffi.call/return_ll1.c: New test case.
2395
2396 2005-02-11  Janis Johnson  <janis187@us.ibm.com>
2397
2398         * testsuite/libffi.call/cls_align_longdouble.c: Remove dg-options.
2399         * testsuite/libffi.call/float.c: Ditto.
2400         * testsuite/libffi.call/float2.c: Ditto.
2401         * testsuite/libffi.call/float3.c: Ditto.
2402
2403 2005-02-08  Andreas Tobler  <a.tobler@schweiz.ch>
2404
2405         * src/frv/ffitarget.h: Remove PPC stuff which does not belong to frv.
2406
2407 2005-01-12  Eric Botcazou  <ebotcazou@libertysurf.fr>
2408
2409         * testsuite/libffi.special/special.exp (cxx_options): Add
2410         -shared-libgcc.
2411
2412 2004-12-31  Richard Henderson  <rth@redhat.com>
2413
2414         * src/types.c (FFI_AGGREGATE_TYPEDEF): Remove.
2415         (FFI_TYPEDEF): Rename from FFI_INTEGRAL_TYPEDEF.  Replace size and
2416         offset parameters with a type parameter; deduce size and structure
2417         alignment.  Update all users.
2418
2419 2004-12-31  Richard Henderson  <rth@redhat.com>
2420
2421         * src/types.c (FFI_TYPE_POINTER): Define with sizeof.
2422         (FFI_TYPE_LONGDOUBLE): Fix for ia64.
2423         * src/ia64/ffitarget.h (struct ffi_ia64_trampoline_struct): Move
2424         into ffi_prep_closure.
2425         * src/ia64/ia64_flags.h, src/ia64/ffi.c, src/ia64/unix.S: Rewrite
2426         from scratch.
2427
2428 2004-12-27  Richard Henderson  <rth@redhat.com>
2429
2430         * src/x86/unix64.S: Fix typo in unwind info.
2431
2432 2004-12-25  Richard Henderson  <rth@redhat.com>
2433
2434         * src/x86/ffi64.c (struct register_args): Rename from stackLayout.
2435         (enum x86_64_reg_class): Add X86_64_COMPLEX_X87_CLASS.
2436         (merge_classes): Check for it.
2437         (SSE_CLASS_P): New.
2438         (classify_argument): Pass byte_offset by value; perform all updates
2439         inside struct case.
2440         (examine_argument): Add classes argument; handle
2441         X86_64_COMPLEX_X87_CLASS.
2442         (ffi_prep_args): Merge into ...
2443         (ffi_call): ... here.  Share stack frame with ffi_call_unix64.
2444         (ffi_prep_cif_machdep): Setup cif->flags for proper structure return.
2445         (ffi_fill_return_value): Remove.
2446         (ffi_prep_closure): Remove dead assert.
2447         (ffi_closure_unix64_inner): Rename from ffi_closure_UNIX64_inner.
2448         Rewrite to use struct register_args instead of va_list.  Create
2449         flags for handling structure returns.
2450         * src/x86/unix64.S: Remove dead strings.
2451         (ffi_call_unix64): Rename from ffi_call_UNIX64.  Rewrite to share
2452         stack frame with ffi_call.  Handle structure returns properly.
2453         (float2sse, floatfloat2sse, double2sse): Remove.
2454         (sse2float, sse2double, sse2floatfloat): Remove.
2455         (ffi_closure_unix64): Rename from ffi_closure_UNIX64.  Rewrite
2456         to handle structure returns properly.
2457
2458 2004-12-08  David Edelsohn  <edelsohn@gnu.org>
2459
2460         * Makefile.am (AM_MAKEFLAGS): Remove duplicate LIBCFLAGS and
2461         PICFLAG.
2462         * Makefile.in: Regenerated.
2463
2464 2004-12-02  Richard Sandiford  <rsandifo@redhat.com>
2465
2466         * configure.ac: Use TL_AC_GCC_VERSION to set gcc_version.
2467         * configure, aclocal.m4, Makefile.in: Regenerate.
2468         * include/Makefile.in, testsuite/Makefile.in: Regenerate.
2469
2470 2004-11-29  Kelley Cook  <kcook@gcc.gnu.org>
2471
2472         * configure: Regenerate for libtool change.
2473
2474 2004-11-25  Kelley Cook  <kcook@gcc.gnu.org>
2475
2476         * configure: Regenerate for libtool reversion.
2477
2478 2004-11-24  Kelley Cook  <kcook@gcc.gnu.org>
2479
2480         * configure: Regenerate for libtool change.
2481
2482 2004-11-23  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2483
2484         * testsuite/lib/libffi-dg.exp: Use new procs in target-libpath.exp.
2485
2486 2004-11-23  Richard Sandiford  <rsandifo@redhat.com>
2487
2488         * src/mips/o32.S (ffi_call_O32, ffi_closure_O32): Use jalr instead
2489         of jal.  Use an absolute encoding for the frame information.
2490
2491 2004-11-23  Kelley Cook  <kcook@gcc.gnu.org>
2492
2493         * Makefile.am: Remove no-dependencies.  Add ACLOCAL_AMFLAGS.
2494         * acinclude.m4: Delete logic for sincludes.
2495         * aclocal.m4, Makefile.in, configure: Regenerate.
2496         * include/Makefile: Likewise.
2497         * testsuite/Makefile: Likewise.
2498
2499 2004-11-22  Eric Botcazou  <ebotcazou@libertysurf.fr>
2500
2501         * src/sparc/ffi.c (ffi_prep_closure): Align doubles and 64-bit integers
2502         on a 8-byte boundary.
2503         * src/sparc/v8.S (ffi_closure_v8): Reserve frame space for arguments.
2504
2505 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2506
2507         * src/arm/ffi.c (ffi_prep_cif_machdep): Handle functions that return
2508         long long values.  Round stack allocation to a multiple of 8 bytes
2509         for ATPCS compatibility.
2510         * src/arm/sysv.S (ffi_call_SYSV): Rework to avoid use of APCS register
2511         names.  Handle returning long long types.  Add Thumb and interworking
2512         support.  Improve soft-float code.
2513
2514 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2515
2516         * testsuite/lib/libffi-db.exp (load_gcc_lib): New function.
2517         (libffi_exit): New function.
2518         (libffi_init): Build the testglue wrapper if needed.
2519
2520 2004-10-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2521
2522         PR other/18138
2523         * testsuite/lib/libffi-dg.exp: Accept more than one multilib libgcc.
2524
2525 2004-10-25  Kazuhiro Inaoka <inaoka.kazuhiro@renesas.com>
2526
2527         * src/m32r/libffitarget.h (FFI_CLOSURES): Set to 0.
2528
2529 2004-10-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2530
2531         * src/sh/sysv.S (ffi_call_SYSV): Don't align for double data.
2532         * testsuite/libffi.call/float3.c: New test case.
2533
2534 2004-10-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2535
2536         * src/sh/ffi.c (ffi_prep_closure): Set T bit in trampoline for
2537         the function returning a structure pointed with R2.
2538         * src/sh/sysv.S (ffi_closure_SYSV): Use R2 as the pointer to
2539         the structure return value if T bit set.  Emit position
2540         independent code and EH data if PIC.
2541
2542 2004-10-13  Kazuhiro Inaoka  <inaoka.kazuhiro@renesas.com>
2543
2544         * Makefile.am: Add m32r support.
2545         * configure.ac: Likewise.
2546         * Makefile.in: Regenerate.
2547         * confiugre: Regenerate.
2548         * src/types.c: Add m32r port to FFI_INTERNAL_TYPEDEF
2549         (uint64, sint64, double, longdouble)
2550         * src/m32r: New directory.
2551         * src/m32r/ffi.c: New file.
2552         * src/m32r/sysv.S: Likewise.
2553         * src/m32r/ffitarget.h: Likewise.
2554
2555 2004-10-02  Kaz Kojima  <kkojima@gcc.gnu.org>
2556
2557         * testsuite/libffi.call/negint.c: New test case.
2558
2559 2004-09-14  H.J. Lu  <hongjiu.lu@intel.com>
2560
2561         PR libgcj/17465
2562         * testsuite/lib/libffi-dg.exp: Don't use global ld_library_path.
2563         Set up LD_LIBRARY_PATH, SHLIB_PATH, LD_LIBRARYN32_PATH,
2564         LD_LIBRARY64_PATH, LD_LIBRARY_PATH_32, LD_LIBRARY_PATH_64 and
2565         DYLD_LIBRARY_PATH.
2566
2567 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2568
2569         * testsuite/libffi.call/many_win32.c: Remove whitespaces.
2570         * testsuite/libffi.call/promotion.c: Likewise.
2571         * testsuite/libffi.call/return_ll.c: Remove unused var. Cleanup
2572         whitespaces.
2573         * testsuite/libffi.call/return_sc.c: Likewise.
2574         * testsuite/libffi.call/return_uc.c: Likewise.
2575
2576 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2577
2578         * src/powerpc/darwin.S: Fix comments and identation.
2579         * src/powerpc/darwin_closure.S: Likewise.
2580
2581 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2582
2583         * src/powerpc/ffi_darwin.c: Add flag for longdouble return values.
2584         (ffi_prep_args): Handle longdouble arguments.
2585         (ffi_prep_cif_machdep): Set flags for longdouble. Calculate space for
2586         longdouble.
2587         (ffi_closure_helper_DARWIN): Add closure handling for longdouble.
2588         * src/powerpc/darwin.S (_ffi_call_DARWIN): Add handling of longdouble
2589         values.
2590         * src/powerpc/darwin_closure.S (_ffi_closure_ASM): Likewise.
2591         * src/types.c: Defined longdouble size and alignment for darwin.
2592
2593 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2594
2595         * src/powerpc/aix.S: Remove whitespaces.
2596         * src/powerpc/aix_closure.S: Likewise.
2597         * src/powerpc/asm.h: Likewise.
2598         * src/powerpc/ffi.c: Likewise.
2599         * src/powerpc/ffitarget.h: Likewise.
2600         * src/powerpc/linux64.S: Likewise.
2601         * src/powerpc/linux64_closure.S: Likewise.
2602         * src/powerpc/ppc_closure.S: Likewise.
2603         * src/powerpc/sysv.S: Likewise.
2604
2605 2004-08-30  Anthony Green  <green@redhat.com>
2606
2607         * Makefile.am: Add frv support.
2608         * Makefile.in, testsuite/Makefile.in: Rebuilt.
2609         * configure.ac: Read configure.host.
2610         * configure.in: Read configure.host.
2611         * configure.host: New file.  frv-elf needs libgloss.
2612         * include/ffi.h.in: Force ffi_closure to have a nice big (8)
2613         alignment.  This is needed to frv and shouldn't harm the others.
2614         * include/ffi_common.h (ALIGN_DOWN): New macro.
2615         * src/frv/ffi.c, src/frv/ffitarget.h, src/frv/eabi.S: New files.
2616
2617 2004-08-24  David Daney  <daney@avtrex.com>
2618
2619         * testsuite/libffi.call/closure_fn0.c: Xfail mips64* instead of mips*.
2620         * testsuite/libffi.call/closure_fn1.c: Likewise.
2621         * testsuite/libffi.call/closure_fn2.c  Likewise.
2622         * testsuite/libffi.call/closure_fn3.c: Likewise.
2623         * testsuite/libffi.call/closure_fn4.c: Likewise.
2624         * testsuite/libffi.call/closure_fn5.c: Likewise.
2625         * testsuite/libffi.call/cls_18byte.c: Likewise.
2626         * testsuite/libffi.call/cls_19byte.c: Likewise.
2627         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2628         * testsuite/libffi.call/cls_20byte.c: Likewise.
2629         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2630         * testsuite/libffi.call/cls_24byte.c: Likewise.
2631         * testsuite/libffi.call/cls_2byte.c: Likewise.
2632         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2633         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2634         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2635         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2636         * testsuite/libffi.call/cls_4byte.c: Likewise.
2637         * testsuite/libffi.call/cls_64byte.c: Likewise.
2638         * testsuite/libffi.call/cls_6byte.c: Likewise.
2639         * testsuite/libffi.call/cls_7byte.c: Likewise.
2640         * testsuite/libffi.call/cls_8byte.c: Likewise.
2641         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2642         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2643         * testsuite/libffi.call/cls_align_double.c: Likewise.
2644         * testsuite/libffi.call/cls_align_float.c: Likewise.
2645         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2646         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2647         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2648         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2649         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2650         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2651         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2652         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
2653         * testsuite/libffi.call/cls_double.c: Likewise.
2654         * testsuite/libffi.call/cls_float.c: Likewise.
2655         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2656         * testsuite/libffi.call/cls_multi_sshort.c: Likewise.
2657         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2658         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2659         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2660         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2661         * testsuite/libffi.call/cls_schar.c: Likewise.
2662         * testsuite/libffi.call/cls_sint.c: Likewise.
2663         * testsuite/libffi.call/cls_sshort.c: Likewise.
2664         * testsuite/libffi.call/cls_uchar.c: Likewise.
2665         * testsuite/libffi.call/cls_uint.c: Likewise.
2666         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2667         * testsuite/libffi.call/cls_ushort.c: Likewise.
2668         * testsuite/libffi.call/nested_struct.c: Likewise.
2669         * testsuite/libffi.call/nested_struct1.c: Likewise.
2670         * testsuite/libffi.call/nested_struct2.c: Likewise.
2671         * testsuite/libffi.call/nested_struct3.c: Likewise.
2672         * testsuite/libffi.call/problem1.c: Likewise.
2673         * testsuite/libffi.special/unwindtest.cc: Likewise.
2674         * testsuite/libffi.call/cls_12byte.c: Likewise and set return value
2675         to zero.
2676         * testsuite/libffi.call/cls_16byte.c: Likewise.
2677         * testsuite/libffi.call/cls_5byte.c: Likewise.
2678
2679 2004-08-23  David Daney <daney@avtrex.com>
2680
2681         PR libgcj/13141
2682         * src/mips/ffitarget.h (FFI_O32_SOFT_FLOAT): New ABI.
2683         * src/mips/ffi.c (ffi_prep_args): Fix alignment calculation.
2684         (ffi_prep_cif_machdep): Handle FFI_O32_SOFT_FLOAT floating point
2685         parameters and return types.
2686         (ffi_call): Handle FFI_O32_SOFT_FLOAT ABI.
2687         (ffi_prep_closure): Ditto.
2688         (ffi_closure_mips_inner_O32): Handle FFI_O32_SOFT_FLOAT ABI, fix
2689         alignment calculations.
2690         * src/mips/o32.S (ffi_closure_O32): Don't use floating point
2691         instructions if FFI_O32_SOFT_FLOAT, make stack frame ABI compliant.
2692
2693 2004-08-14  Casey Marshall <csm@gnu.org>
2694
2695         * src/mips/ffi.c (ffi_pref_cif_machdep): set `cif->flags' to
2696         contain `FFI_TYPE_UINT64' as return type for any 64-bit
2697         integer (O32 ABI only).
2698         (ffi_prep_closure): new function.
2699         (ffi_closure_mips_inner_O32): new function.
2700         * src/mips/ffitarget.h: Define `FFI_CLOSURES' and
2701         `FFI_TRAMPOLINE_SIZE' appropriately if the ABI is o32.
2702         * src/mips/o32.S (ffi_call_O32): add labels for .eh_frame. Return
2703         64 bit integers correctly.
2704         (ffi_closure_O32): new function.
2705         Added DWARF-2 unwind info for both functions.
2706
2707 2004-08-10  Andrew Haley  <aph@redhat.com>
2708
2709         * src/x86/ffi64.c (ffi_prep_args ): 8-align all stack arguments.
2710
2711 2004-08-01  Robert Millan  <robertmh@gnu.org>
2712
2713         * configure.ac: Detect knetbsd-gnu and kfreebsd-gnu.
2714         * configure: Regenerate.
2715
2716 2004-07-30  Maciej W. Rozycki  <macro@linux-mips.org>
2717
2718         * acinclude.m4 (AC_FUNC_MMAP_BLACKLIST): Check for <sys/mman.h>
2719         and mmap() explicitly instead of relying on preset autoconf cache
2720         variables.
2721         * aclocal.m4: Regenerate.
2722         * configure: Regenerate.
2723
2724 2004-07-11  Ulrich Weigand  <uweigand@de.ibm.com>
2725
2726         * src/s390/ffi.c (ffi_prep_args): Fix C aliasing violation.
2727         (ffi_check_float_struct): Remove unused prototype.
2728
2729 2004-06-30  Geoffrey Keating  <geoffk@apple.com>
2730
2731         * src/powerpc/ffi_darwin.c (flush_icache): ';' is a comment
2732         character on Darwin, use '\n\t' instead.
2733
2734 2004-06-26  Matthias Klose  <doko@debian.org>
2735
2736         * libtool-version: Fix typo in revision/age.
2737
2738 2004-06-17  Matthias Klose  <doko@debian.org>
2739
2740         * libtool-version: New.
2741         * Makefile.am (libffi_la_LDFLAGS): Use -version-info for soname.
2742         * Makefile.in: Regenerate.
2743
2744 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2745
2746         * Makefile.am: Remove useless multilib rules.
2747         * Makefile.in: Regenerate.
2748         * aclocal.m4: Regenerate with automake 1.8.5.
2749         * configure.ac: Remove useless multilib configury.
2750         * configure: Regenerate.
2751
2752 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2753
2754         * .cvsignore: New file.
2755
2756 2004-06-10  Jakub Jelinek  <jakub@redhat.com>
2757
2758         * src/ia64/unix.S (ffi_call_unix): Insert group barrier break
2759         fp_done.
2760         (ffi_closure_UNIX): Fix f14/f15 adjustment if FLOAT_SZ is ever
2761         changed from 8.
2762
2763 2004-06-06  Sean McNeil  <sean@mcneil.com>
2764
2765         * configure.ac: Add x86_64-*-freebsd* support.
2766         * configure: Regenerate.
2767
2768 2004-04-26  Joe Buck <jbuck@welsh-buck.org>
2769
2770         Bug 15093
2771         * configure.ac: Test for existence of mmap and sys/mman.h before
2772         checking blacklist.  Fix suggested by Jim Wilson.
2773         * configure: Regenerate.
2774
2775 2004-04-26  Matt Austern  <austern@apple.com>
2776
2777         * src/powerpc/darwin.S: Go through a non-lazy pointer for initial
2778         FDE location.
2779         * src/powerpc/darwin_closure.S: Likewise.
2780
2781 2004-04-24  Andreas Tobler  <a.tobler@schweiz.ch>
2782
2783         * testsuite/libffi.call/cls_multi_schar.c (main): Fix initialization
2784         error. Reported by Thomas Heller <theller@python.net>.
2785         * testsuite/libffi.call/cls_multi_sshort.c (main): Likewise.
2786         * testsuite/libffi.call/cls_multi_ushort.c (main): Likewise.
2787
2788 2004-03-20  Matthias Klose  <doko@debian.org>
2789
2790         * src/pa/linux.S: Fix typo.
2791
2792 2004-03-19  Matthias Klose  <doko@debian.org>
2793
2794         * Makefile.am: Update.
2795         * Makefile.in: Regenerate.
2796         * src/pa/ffi.h.in: Remove.
2797         * src/pa/ffitarget.h: New file.
2798
2799 2004-02-10  Randolph Chung  <tausq@debian.org>
2800
2801         * Makefile.am: Add PA support.
2802         * Makefile.in: Regenerate.
2803         * include/Makefile.in: Regenerate.
2804         * configure.ac: Add PA target.
2805         * configure: Regenerate.
2806         * src/pa/ffi.c: New file.
2807         * src/pa/ffi.h.in: Add PA support.
2808         * src/pa/linux.S: New file.
2809         * prep_cif.c: Add PA support.
2810
2811 2004-03-16  Hosaka Yuji  <hos@tamanegi.org>
2812
2813         * src/types.c: Fix alignment size of X86_WIN32 case int64 and
2814         double.
2815         * src/x86/ffi.c (ffi_prep_args): Replace ecif->cif->rtype->type
2816         with ecif->cif->flags.
2817         (ffi_call, ffi_prep_incoming_args_SYSV): Replace cif->rtype->type
2818         with cif->flags.
2819         (ffi_prep_cif_machdep): Add X86_WIN32 struct case.
2820         (ffi_closure_SYSV): Add 1 or 2-bytes struct case for X86_WIN32.
2821         * src/x86/win32.S (retstruct1b, retstruct2b, sc_retstruct1b,
2822         sc_retstruct2b): Add for 1 or 2-bytes struct case.
2823
2824 2004-03-15 Kelley Cook <kcook@gcc.gnu.org>
2825
2826         * configure.in: Rename file to ...
2827         * configure.ac: ... this.
2828         * fficonfig.h.in: Regenerate.
2829         * Makefile.in: Regenerate.
2830         * include/Makefile.in: Regenerate.
2831         * testsuite/Makefile.in: Regenerate.
2832
2833 2004-03-12  Matt Austern  <austern@apple.com>
2834
2835         * src/powerpc/darwin.S: Fix EH information so it corresponds to
2836         changes in EH format resulting from addition of linkonce support.
2837         * src/powerpc/darwin_closure.S: Likewise.
2838
2839 2004-03-11  Andreas Tobler  <a.tobler@schweiz.ch>
2840             Paolo Bonzini  <bonzini@gnu.org>
2841
2842         * Makefile.am (AUTOMAKE_OPTIONS): Set them.
2843         Remove VPATH. Remove rules for object files. Remove multilib support.
2844         (AM_CCASFLAGS): Add.
2845         * configure.in (AC_CONFIG_HEADERS): Relace AM_CONFIG_HEADER.
2846         (AC_PREREQ): Bump version to 2.59.
2847         (AC_INIT): Fill with version info and bug address.
2848         (ORIGINAL_LD_FOR_MULTILIBS): Remove.
2849         (AM_ENABLE_MULTILIB): Use this instead of AC_ARG_ENABLE.
2850         De-precious CC so that the right flags are passed down to multilibs.
2851         (AC_MSG_ERROR): Replace obsolete macro AC_ERROR.
2852         (AC_CONFIG_FILES): Replace obsolete macro AC_LINK_FILES.
2853         (AC_OUTPUT): Reorganize the output with AC_CONFIG_COMMANDS.
2854         * configure: Rebuilt.
2855         * aclocal.m4: Likewise.
2856         * Makefile.in, include/Makefile.in, testsuite/Makefile.in: Likewise.
2857         * fficonfig.h.in: Likewise.
2858
2859 2004-03-11  Andreas Schwab  <schwab@suse.de>
2860
2861         * src/ia64/ffi.c (ffi_prep_incoming_args_UNIX): Get floating point
2862         arguments from fp registers only for the first 8 parameter slots.
2863         Don't convert a float parameter when passed in memory.
2864
2865 2004-03-09  Hans-Peter Nilsson  <hp@axis.com>
2866
2867         * configure: Regenerate for config/accross.m4 correction.
2868
2869 2004-02-25  Matt Kraai  <kraai@alumni.cmu.edu>
2870
2871         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Change
2872         ecif->cif->bytes to bytes.
2873         (ffi_prep_cif_machdep): Add braces around nested if statement.
2874
2875 2004-02-09  Alan Modra  <amodra@bigpond.net.au>
2876
2877         * src/types.c (pointer): POWERPC64 has 8 byte pointers.
2878
2879         * src/powerpc/ffi.c (ffi_prep_args64): Correct long double handling.
2880         (ffi_closure_helper_LINUX64): Fix typo.
2881         * testsuite/libffi.call/cls_align_longdouble.c: Pass -mlong-double-128
2882         for powerpc64-*-*.
2883         * testsuite/libffi.call/float.c: Likewise.
2884         * testsuite/libffi.call/float2.c: Likewise.
2885
2886 2004-02-08  Alan Modra  <amodra@bigpond.net.au>
2887
2888         * src/powerpc/ffi.c (ffi_prep_cif_machdep <FFI_LINUX64>): Correct
2889         long double function return and long double arg handling.
2890         (ffi_closure_helper_LINUX64): Formatting.  Delete unused "ng" var.
2891         Use "end_pfr" instead of "nf".  Correct long double handling.
2892         Localise "temp".
2893         * src/powerpc/linux64.S (ffi_call_LINUX64): Save f2 long double
2894         return value.
2895         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Allocate
2896         space for long double return value.  Adjust stack frame and offsets.
2897         Load f2 long double return.
2898
2899 2004-02-07  Alan Modra  <amodra@bigpond.net.au>
2900
2901         * src/types.c: Use 16 byte long double for POWERPC64.
2902
2903 2004-01-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2904
2905         * src/sparc/ffi.c (ffi_prep_args_v9): Shift the parameter array
2906         when the structure return address is passed in %o0.
2907         (ffi_V9_return_struct): Rename into ffi_v9_layout_struct.
2908         (ffi_v9_layout_struct): Align the field following a nested structure
2909         on a word boundary.  Use memmove instead of memcpy.
2910         (ffi_call): Update call to ffi_V9_return_struct.
2911         (ffi_prep_closure): Define 'ctx' only for V8.
2912         (ffi_closure_sparc_inner): Clone into ffi_closure_sparc_inner_v8
2913         and ffi_closure_sparc_inner_v9.
2914         (ffi_closure_sparc_inner_v8): Return long doubles by reference.
2915         Always skip the structure return address.  For structures and long
2916         doubles, copy the argument directly.
2917         (ffi_closure_sparc_inner_v9): Skip the structure return address only
2918         if required.  Shift the maximum floating-point slot accordingly.  For
2919         big structures, copy the argument directly; otherwise, left-justify the
2920         argument and call ffi_v9_layout_struct to lay out the structure on
2921         the stack.
2922         * src/sparc/v8.S: Undef STACKFRAME before defining it.
2923         (ffi_closure_v8): Pass the structure return address.  Update call to
2924         ffi_closure_sparc_inner_v8.  Short-circuit FFI_TYPE_INT handling.
2925         Skip the 'unimp' insn when returning long doubles and structures.
2926         * src/sparc/v9.S: Undef STACKFRAME before defining it.
2927         (ffi_closure_v9): Increase the frame size by 2 words.  Short-circuit
2928         FFI_TYPE_INT handling.  Load structures both in integers and
2929         floating-point registers on return.
2930         * README: Update status of the SPARC port.
2931
2932 2004-01-24  Andreas Tobler  <a.tobler@schweiz.ch>
2933
2934         * testsuite/libffi.call/pyobjc-tc.c (main): Treat result value
2935         as of type ffi_arg.
2936         * testsuite/libffi.call/struct3.c (main): Fix CHECK.
2937
2938 2004-01-22  Ulrich Weigand  <uweigand@de.ibm.com>
2939
2940         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Treat result
2941         value as of type ffi_arg, not unsigned int.
2942
2943 2004-01-21  Michael Ritzert  <ritzert@t-online.de>
2944
2945         * ffi64.c (ffi_prep_args): Cast the RHS of an assignment instead
2946         of the LHS.
2947
2948 2004-01-12  Andreas Tobler  <a.tobler@schweiz.ch>
2949
2950         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_32 for
2951         Solaris.
2952
2953 2004-01-08  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2954
2955         * testsuite/libffi.call/ffitest.h (allocate_mmap): Cast MAP_FAILED
2956         to void *.
2957
2958 2003-12-10  Richard Henderson  <rth@redhat.com>
2959
2960         * testsuite/libffi.call/cls_align_pointer.c: Cast pointers to
2961         size_t instead of int.
2962
2963 2003-12-04  Hosaka Yuji  <hos@tamanegi.org>
2964
2965         * testsuite/libffi.call/many_win32.c: Include <float.h>.
2966         * testsuite/libffi.call/many_win32.c (main): Replace variable
2967         int i with unsigned long ul.
2968
2969         * testsuite/libffi.call/cls_align_uint64.c: New test case.
2970         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2971         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2972         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2973         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2974         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2975         * testsuite/libffi.call/cls_align_float.c: Likewise.
2976         * testsuite/libffi.call/cls_align_double.c: Likewise.
2977         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2978         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2979
2980 2003-12-02  Hosaka Yuji  <hos@tamanegi.org>
2981
2982         PR other/13221
2983         * src/x86/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV):
2984         Align arguments to 32 bits.
2985
2986 2003-12-01  Andreas Tobler  <a.tobler@schweiz.ch>
2987
2988         PR other/13221
2989         * testsuite/libffi.call/cls_multi_sshort.c: New test case.
2990         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2991         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2992         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2993         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2994         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2995
2996         * testsuite/libffi.special/unwindtest.cc: Cosmetics.
2997
2998 2003-11-26  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
2999
3000         * testsuite/libffi.call/ffitest.h: Include <fcntl.h>.
3001         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3002
3003 2003-11-22  Andreas Tobler  <a.tobler@schweiz.ch>
3004
3005         * Makefile.in: Rebuilt.
3006         * configure: Likewise.
3007         * testsuite/libffi.special/unwindtest.cc: Convert the mmap to
3008         the right type.
3009
3010 2003-11-21  Andreas Jaeger  <aj@suse.de>
3011             Andreas Tobler  <a.tobler@schweiz.ch>
3012
3013         * acinclude.m4: Add AC_FUNC_MMAP_BLACKLIST.
3014         * configure.in: Call AC_FUNC_MMAP_BLACKLIST.
3015         * Makefile.in: Rebuilt.
3016         * aclocal.m4: Likewise.
3017         * configure: Likewise.
3018         * fficonfig.h.in: Likewise.
3019         * testsuite/lib/libffi-dg.exp: Add include dir.
3020         * testsuite/libffi.call/ffitest.h: Add MMAP definitions.
3021         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3022         * testsuite/libffi.call/closure_fn0.c: Use MMAP functionality
3023         for ffi_closure if available.
3024         * testsuite/libffi.call/closure_fn1.c: Likewise.
3025         * testsuite/libffi.call/closure_fn2.c: Likewise.
3026         * testsuite/libffi.call/closure_fn3.c: Likewise.
3027         * testsuite/libffi.call/closure_fn4.c: Likewise.
3028         * testsuite/libffi.call/closure_fn5.c: Likewise.
3029         * testsuite/libffi.call/cls_12byte.c: Likewise.
3030         * testsuite/libffi.call/cls_16byte.c: Likewise.
3031         * testsuite/libffi.call/cls_18byte.c: Likewise.
3032         * testsuite/libffi.call/cls_19byte.c: Likewise.
3033         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3034         * testsuite/libffi.call/cls_20byte.c: Likewise.
3035         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3036         * testsuite/libffi.call/cls_24byte.c: Likewise.
3037         * testsuite/libffi.call/cls_2byte.c: Likewise.
3038         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3039         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3040         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3041         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3042         * testsuite/libffi.call/cls_4byte.c: Likewise.
3043         * testsuite/libffi.call/cls_5byte.c: Likewise.
3044         * testsuite/libffi.call/cls_64byte.c: Likewise.
3045         * testsuite/libffi.call/cls_6byte.c: Likewise.
3046         * testsuite/libffi.call/cls_7byte.c: Likewise.
3047         * testsuite/libffi.call/cls_8byte.c: Likewise.
3048         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3049         * testsuite/libffi.call/cls_9byte2.c: Likewise.
3050         * testsuite/libffi.call/cls_double.c: Likewise.
3051         * testsuite/libffi.call/cls_float.c: Likewise.
3052         * testsuite/libffi.call/cls_schar.c: Likewise.
3053         * testsuite/libffi.call/cls_sint.c: Likewise.
3054         * testsuite/libffi.call/cls_sshort.c: Likewise.
3055         * testsuite/libffi.call/cls_uchar.c: Likewise.
3056         * testsuite/libffi.call/cls_uint.c: Likewise.
3057         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3058         * testsuite/libffi.call/cls_ushort.c: Likewise.
3059         * testsuite/libffi.call/nested_struct.c: Likewise.
3060         * testsuite/libffi.call/nested_struct1.c: Likewise.
3061         * testsuite/libffi.call/nested_struct2.c: Likewise.
3062         * testsuite/libffi.call/nested_struct3.c: Likewise.
3063         * testsuite/libffi.call/problem1.c: Likewise.
3064         * testsuite/libffi.special/unwindtest.cc: Likewise.
3065
3066 2003-11-20  Andreas Tobler  <a.tobler@schweiz.ch>
3067
3068         * testsuite/lib/libffi-dg.exp: Make the -lgcc_s conditional.
3069
3070 2003-11-19  Andreas Tobler  <a.tobler@schweiz.ch>
3071
3072         * testsuite/lib/libffi-dg.exp: Add DYLD_LIBRARY_PATH for darwin.
3073         Add -lgcc_s to additional flags.
3074
3075 2003-11-12  Andreas Tobler  <a.tobler@schweiz.ch>
3076
3077         * configure.in, include/Makefile.am: PR libgcj/11147, install
3078         the ffitarget.h header file in a gcc versioned and target
3079         dependent place.
3080         * configure: Regenerated.
3081         * Makefile.in, include/Makefile.in: Likewise.
3082         * testsuite/Makefile.in: Likewise.
3083
3084 2003-11-09  Andreas Tobler  <a.tobler@schweiz.ch>
3085
3086         * testsuite/libffi.call/closure_fn0.c: Print result and check
3087         with dg-output to make debugging easier.
3088         * testsuite/libffi.call/closure_fn1.c: Likewise.
3089         * testsuite/libffi.call/closure_fn2.c: Likewise.
3090         * testsuite/libffi.call/closure_fn3.c: Likewise.
3091         * testsuite/libffi.call/closure_fn4.c: Likewise.
3092         * testsuite/libffi.call/closure_fn5.c: Likewise.
3093         * testsuite/libffi.call/cls_12byte.c: Likewise.
3094         * testsuite/libffi.call/cls_16byte.c: Likewise.
3095         * testsuite/libffi.call/cls_18byte.c: Likewise.
3096         * testsuite/libffi.call/cls_19byte.c: Likewise.
3097         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3098         * testsuite/libffi.call/cls_20byte.c: Likewise.
3099         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3100         * testsuite/libffi.call/cls_24byte.c: Likewise.
3101         * testsuite/libffi.call/cls_2byte.c: Likewise.
3102         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3103         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3104         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3105         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3106         * testsuite/libffi.call/cls_4byte.c: Likewise.
3107         * testsuite/libffi.call/cls_5byte.c: Likewise.
3108         * testsuite/libffi.call/cls_64byte.c: Likewise.
3109         * testsuite/libffi.call/cls_6byte.c: Likewise.
3110         * testsuite/libffi.call/cls_7byte.c: Likewise.
3111         * testsuite/libffi.call/cls_8byte.c: Likewise.
3112         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3113         * testsuite/libffi.call/cls_9byte2.c: Likewise.
3114         * testsuite/libffi.call/cls_double.c: Likewise.
3115         * testsuite/libffi.call/cls_float.c: Likewise.
3116         * testsuite/libffi.call/cls_schar.c: Likewise.
3117         * testsuite/libffi.call/cls_sint.c: Likewise.
3118         * testsuite/libffi.call/cls_sshort.c: Likewise.
3119         * testsuite/libffi.call/cls_uchar.c: Likewise.
3120         * testsuite/libffi.call/cls_uint.c: Likewise.
3121         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3122         * testsuite/libffi.call/cls_ushort.c: Likewise.
3123         * testsuite/libffi.call/problem1.c: Likewise.
3124
3125         * testsuite/libffi.special/unwindtest.cc: Make ffi_closure
3126         static.
3127
3128 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
3129
3130         * testsuite/libffi.call/cls_9byte2.c: New test case.
3131         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3132         * testsuite/libffi.call/cls_64byte.c: Likewise.
3133         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3134         * testsuite/libffi.call/cls_19byte.c: Likewise.
3135         * testsuite/libffi.call/cls_18byte.c: Likewise.
3136         * testsuite/libffi.call/closure_fn4.c: Likewise.
3137         * testsuite/libffi.call/closure_fn5.c: Likewise.
3138         * testsuite/libffi.call/cls_schar.c: Likewise.
3139         * testsuite/libffi.call/cls_sint.c: Likewise.
3140         * testsuite/libffi.call/cls_sshort.c: Likewise.
3141         * testsuite/libffi.call/nested_struct2.c: Likewise.
3142         * testsuite/libffi.call/nested_struct3.c: Likewise.
3143
3144 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
3145
3146         * testsuite/libffi.call/cls_double.c: Do a check on the result.
3147         * testsuite/libffi.call/cls_uchar.c: Likewise.
3148         * testsuite/libffi.call/cls_uint.c: Likewise.
3149         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3150         * testsuite/libffi.call/cls_ushort.c: Likewise.
3151         * testsuite/libffi.call/return_sc.c: Cleanup whitespaces.
3152
3153 2003-11-06  Andreas Tobler  <a.tobler@schweiz.ch>
3154
3155         * src/prep_cif.c (ffi_prep_cif): Move the validity check after
3156         the initialization.
3157
3158 2003-10-23  Andreas Tobler  <a.tobler@schweiz.ch>
3159
3160         * src/java_raw_api.c (ffi_java_ptrarray_to_raw): Replace
3161         FFI_ASSERT(FALSE) with FFI_ASSERT(0).
3162
3163 2003-10-22  David Daney  <ddaney@avtrex.com>
3164
3165         * src/mips/ffitarget.h: Replace undefined UINT32 and friends with
3166         __attribute__((__mode__(__SI__))) and friends.
3167
3168 2003-10-22  Andreas Schwab  <schwab@suse.de>
3169
3170         * src/ia64/ffi.c: Replace FALSE/TRUE with false/true.
3171
3172 2003-10-21  Andreas Tobler  <a.tobler@schweiz.ch>
3173
3174         * configure.in: AC_LINK_FILES(ffitarget.h).
3175         * configure: Regenerate.
3176         * Makefile.in: Likewise.
3177         * include/Makefile.in: Likewise.
3178         * testsuite/Makefile.in: Likewise.
3179         * fficonfig.h.in: Likewise.
3180
3181 2003-10-21  Paolo Bonzini  <bonzini@gnu.org>
3182             Richard Henderson  <rth@redhat.com>
3183
3184         Avoid that ffi.h includes fficonfig.h.
3185
3186         * Makefile.am (EXTRA_DIST): Include ffitarget.h files
3187         (TARGET_SRC_MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3188         (TARGET_SRC_MIPS_SGI): Removed.
3189         (MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3190         (MIPS_SGI): Removed.
3191         (CLEANFILES): Removed.
3192         (mostlyclean-am, clean-am, mostlyclean-sub, clean-sub): New
3193         targets.
3194         * acconfig.h: Removed.
3195         * configure.in: Compute sizeofs only for double and long double.
3196         Use them to define and subst HAVE_LONG_DOUBLE.  Include comments
3197         into AC_DEFINE instead of using acconfig.h.  Create
3198         include/ffitarget.h instead of include/fficonfig.h.  Rename
3199         MIPS_GCC to MIPS_IRIX, drop MIPS_SGI since we are in gcc's tree.
3200         AC_DEFINE EH_FRAME_FLAGS.
3201         * include/Makefile.am (DISTCLEANFILES): New automake macro.
3202         (hack_DATA): Add ffitarget.h.
3203         * include/ffi.h.in: Remove all system specific definitions.
3204         Declare raw API even if it is not installed, why bother?
3205         Use limits.h instead of SIZEOF_* to define ffi_type_*.  Do
3206         not define EH_FRAME_FLAGS, it is in fficonfig.h now.  Include
3207         ffitarget.h instead of fficonfig.h.  Remove ALIGN macro.
3208         (UINT_ARG, INT_ARG): Removed, use ffi_arg and ffi_sarg instead.
3209         * include/ffi_common.h (bool): Do not define.
3210         (ffi_assert): Accept failed assertion.
3211         (ffi_type_test): Return void and accept file/line.
3212         (FFI_ASSERT): Pass stringized failed assertion.
3213         (FFI_ASSERT_AT): New macro.
3214         (FFI_ASSERT_VALID_TYPE): New macro.
3215         (UINT8, SINT8, UINT16, SINT16, UINT32, SINT32,
3216         UINT64, SINT64): Define here with gcc's __attribute__ macro
3217         instead of in ffi.h
3218         (FLOAT32, ALIGN): Define here instead of in ffi.h
3219         * include/ffi-mips.h: Removed.  Its content moved to
3220         src/mips/ffitarget.h after separating assembly and C sections.
3221         * src/alpha/ffi.c, src/alpha/ffi.c, src/java_raw_api.c
3222         src/prep_cif.c, src/raw_api.c, src/ia64/ffi.c,
3223         src/mips/ffi.c, src/mips/n32.S, src/mips/o32.S,
3224         src/mips/ffitarget.h, src/sparc/ffi.c, src/x86/ffi64.c:
3225         SIZEOF_ARG -> FFI_SIZEOF_ARG.
3226         * src/ia64/ffi.c: Include stdbool.h (provided by GCC 2.95+).
3227         * src/debug.c (ffi_assert): Accept stringized failed assertion.
3228         (ffi_type_test): Rewritten.
3229         * src/prep-cif.c (initialize_aggregate, ffi_prep_cif): Call
3230         FFI_ASSERT_VALID_TYPE.
3231         * src/alpha/ffitarget.h, src/arm/ffitarget.h,
3232         src/ia64/ffitarget.h, src/m68k/ffitarget.h,
3233         src/mips/ffitarget.h, src/powerpc/ffitarget.h,
3234         src/s390/ffitarget.h, src/sh/ffitarget.h,
3235         src/sh64/ffitarget.h, src/sparc/ffitarget.h,
3236         src/x86/ffitarget.h: New files.
3237         * src/alpha/osf.S, src/arm/sysv.S, src/ia64/unix.S,
3238         src/m68k/sysv.S, src/mips/n32.S, src/mips/o32.S,
3239         src/powerpc/aix.S, src/powerpc/darwin.S,
3240         src/powerpc/ffi_darwin.c, src/powerpc/linux64.S,
3241         src/powerpc/linux64_closure.S, src/powerpc/ppc_closure.S,
3242         src/powerpc/sysv.S, src/s390/sysv.S, src/sh/sysv.S,
3243         src/sh64/sysv.S, src/sparc/v8.S, src/sparc/v9.S,
3244         src/x86/sysv.S, src/x86/unix64.S, src/x86/win32.S:
3245         include fficonfig.h
3246
3247 2003-10-20  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3248
3249         * src/mips/ffi.c: Use _ABIN32, _ABIO32 instead of external
3250         _MIPS_SIM_NABI32, _MIPS_SIM_ABI32.
3251
3252 2003-10-19  Andreas Tobler  <a.tobler@schweiz.ch>
3253
3254         * src/powerpc/ffi_darwin.c (ffi_prep_args): Declare bytes again.
3255         Used when FFI_DEBUG = 1.
3256
3257 2003-10-14  Alan Modra  <amodra@bigpond.net.au>
3258
3259         * src/types.c (double, longdouble): Default POWERPC64 to 8 byte size
3260         and align.
3261
3262 2003-10-06  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3263
3264         * include/ffi_mips.h: Define FFI_MIPS_N32 for N32/N64 ABIs,
3265         FFI_MIPS_O32 for O32 ABI.
3266
3267 2003-10-01  Andreas Tobler  <a.tobler@schweiz.ch>
3268
3269         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_64 for
3270         SPARC64. Cleanup whitespaces.
3271
3272 2003-09-19  Andreas Tobler  <a.tobler@schweiz.ch>
3273
3274         * testsuite/libffi.call/closure_fn0.c: Xfail mips, arm,
3275         strongarm, xscale. Cleanup whitespaces.
3276         * testsuite/libffi.call/closure_fn1.c: Likewise.
3277         * testsuite/libffi.call/closure_fn2.c: Likewise.
3278         * testsuite/libffi.call/closure_fn3.c: Likewise.
3279         * testsuite/libffi.call/cls_12byte.c: Likewise.
3280         * testsuite/libffi.call/cls_16byte.c: Likewise.
3281         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3282         * testsuite/libffi.call/cls_20byte.c: Likewise.
3283         * testsuite/libffi.call/cls_24byte.c: Likewise.
3284         * testsuite/libffi.call/cls_2byte.c: Likewise.
3285         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3286         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3287         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3288         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3289         * testsuite/libffi.call/cls_4byte.c: Likewise.
3290         * testsuite/libffi.call/cls_5byte.c: Likewise.
3291         * testsuite/libffi.call/cls_6byte.c: Likewise.
3292         * testsuite/libffi.call/cls_7byte.c: Likewise.
3293         * testsuite/libffi.call/cls_8byte.c: Likewise.
3294         * testsuite/libffi.call/cls_double.c: Likewise.
3295         * testsuite/libffi.call/cls_float.c: Likewise.
3296         * testsuite/libffi.call/cls_uchar.c: Likewise.
3297         * testsuite/libffi.call/cls_uint.c: Likewise.
3298         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3299         * testsuite/libffi.call/cls_ushort.c: Likewise.
3300         * testsuite/libffi.call/nested_struct.c: Likewise.
3301         * testsuite/libffi.call/nested_struct1.c: Likewise.
3302         * testsuite/libffi.call/problem1.c: Likewise.
3303         * testsuite/libffi.special/unwindtest.cc: Likewise.
3304         * testsuite/libffi.call/pyobjc-tc.c: Cleanup whitespaces.
3305
3306 2003-09-18  David Edelsohn  <edelsohn@gnu.org>
3307
3308         * src/powerpc/aix.S: Cleanup whitespaces.
3309         * src/powerpc/aix_closure.S: Likewise.
3310
3311 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3312
3313         * src/powerpc/darwin.S: Cleanup whitespaces, comment formatting.
3314         * src/powerpc/darwin_closure.S: Likewise.
3315         * src/powerpc/ffi_darwin.c: Likewise.
3316
3317 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3318             David Edelsohn  <edelsohn@gnu.org>
3319
3320         * src/types.c (double): Add AIX and Darwin to the right TYPEDEF.
3321         * src/powerpc/aix_closure.S: Remove the pointer to the outgoing
3322         parameter stack.
3323         * src/powerpc/darwin_closure.S: Likewise.
3324         * src/powerpc/ffi_darwin.c (ffi_prep_args): Handle structures
3325         according to the Darwin/AIX ABI.
3326         (ffi_prep_cif_machdep): Likewise.
3327         (ffi_closure_helper_DARWIN): Likewise.
3328         Remove the outgoing parameter stack logic. Simplify the evaluation
3329         of the different CASE types.
3330         (ffi_prep_clousure): Avoid the casts on lvalues. Change the branch
3331         statement in the trampoline code.
3332
3333 2003-09-18  Kaz Kojima  <kkojima@gcc.gnu.org>
3334
3335         * src/sh/ffi.c (ffi_prep_args): Take account into the alignement
3336         for the register size.
3337         (ffi_closure_helper_SYSV): Handle the structure return value
3338         address correctly.
3339         (ffi_closure_helper_SYSV): Return the appropriate type when
3340         the registers are used for the structure return value.
3341         * src/sh/sysv.S (ffi_closure_SYSV): Fix the stack layout for
3342         the 64-bit return value.  Update copyright years.
3343
3344 2003-09-17  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3345
3346         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Search in
3347         srcdir for ffi_mips.h.
3348
3349 2003-09-12  Alan Modra  <amodra@bigpond.net.au>
3350
3351         * src/prep_cif.c (initialize_aggregate): Include tail padding in
3352         structure size.
3353         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Correct
3354         placement of float result.
3355         * testsuite/libffi.special/unwindtest.cc (closure_test_fn1): Correct
3356         cast of "resp" for big-endian 64 bit machines.
3357
3358 2003-09-11  Alan Modra  <amodra@bigpond.net.au>
3359
3360         * src/types.c (double, longdouble): Merge identical SH and ARM
3361         typedefs, and add POWERPC64.
3362         * src/powerpc/ffi.c (ffi_prep_args64): Correct next_arg calc for
3363         struct split over gpr and rest.
3364         (ffi_prep_cif_machdep): Correct intarg_count for structures.
3365         * src/powerpc/linux64.S (ffi_call_LINUX64): Fix gpr offsets.
3366
3367 2003-09-09  Andreas Tobler  <a.tobler@schweiz.ch>
3368
3369         * src/powerpc/ffi.c (ffi_closure_helper_SYSV) Handle struct
3370         passing correctly.
3371
3372 2003-09-09  Alan Modra  <amodra@bigpond.net.au>
3373
3374         * configure: Regenerate.
3375
3376 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3377
3378         * Makefile.am: Remove build rules for ffitest.
3379         * Makefile.in: Rebuilt.
3380
3381 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3382
3383         * src/java_raw_api.c: Include <stdlib.h> to fix compiler warning
3384         about implicit declaration of abort().
3385
3386 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3387
3388         * Makefile.am: Add dejagnu test framework. Fixes PR other/11411.
3389         * Makefile.in: Rebuilt.
3390         * configure.in: Add dejagnu test framework.
3391         * configure: Rebuilt.
3392
3393         * testsuite/Makefile.am: New file.
3394         * testsuite/Makefile.in: Built
3395         * testsuite/lib/libffi-dg.exp: New file.
3396         * testsuite/config/default.exp: Likewise.
3397         * testsuite/libffi.call/call.exp: Likewise.
3398         * testsuite/libffi.call/ffitest.h: Likewise.
3399         * testsuite/libffi.call/closure_fn0.c: Likewise.
3400         * testsuite/libffi.call/closure_fn1.c: Likewise.
3401         * testsuite/libffi.call/closure_fn2.c: Likewise.
3402         * testsuite/libffi.call/closure_fn3.c: Likewise.
3403         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3404         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3405         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3406         * testsuite/libffi.call/cls_2byte.c: Likewise.
3407         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3408         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3409         * testsuite/libffi.call/cls_4byte.c: Likewise.
3410         * testsuite/libffi.call/cls_5byte.c: Likewise.
3411         * testsuite/libffi.call/cls_6byte.c: Likewise.
3412         * testsuite/libffi.call/cls_7byte.c: Likewise.
3413         * testsuite/libffi.call/cls_8byte.c: Likewise.
3414         * testsuite/libffi.call/cls_12byte.c: Likewise.
3415         * testsuite/libffi.call/cls_16byte.c: Likewise.
3416         * testsuite/libffi.call/cls_20byte.c: Likewise.
3417         * testsuite/libffi.call/cls_24byte.c: Likewise.
3418         * testsuite/libffi.call/cls_double.c: Likewise.
3419         * testsuite/libffi.call/cls_float.c: Likewise.
3420         * testsuite/libffi.call/cls_uchar.c: Likewise.
3421         * testsuite/libffi.call/cls_uint.c: Likewise.
3422         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3423         * testsuite/libffi.call/cls_ushort.c: Likewise.
3424         * testsuite/libffi.call/float.c: Likewise.
3425         * testsuite/libffi.call/float1.c: Likewise.
3426         * testsuite/libffi.call/float2.c: Likewise.
3427         * testsuite/libffi.call/many.c: Likewise.
3428         * testsuite/libffi.call/many_win32.c: Likewise.
3429         * testsuite/libffi.call/nested_struct.c: Likewise.
3430         * testsuite/libffi.call/nested_struct1.c: Likewise.
3431         * testsuite/libffi.call/pyobjc-tc.c: Likewise.
3432         * testsuite/libffi.call/problem1.c: Likewise.
3433         * testsuite/libffi.call/promotion.c: Likewise.
3434         * testsuite/libffi.call/return_ll.c: Likewise.
3435         * testsuite/libffi.call/return_sc.c: Likewise.
3436         * testsuite/libffi.call/return_uc.c: Likewise.
3437         * testsuite/libffi.call/strlen.c: Likewise.
3438         * testsuite/libffi.call/strlen_win32.c: Likewise.
3439         * testsuite/libffi.call/struct1.c: Likewise.