OSDN Git Service

Fix for Homebrew. I don't care about another environments. This may be reverted and...
[pf3gnuchains/gcc-fork.git] / libffi / ChangeLog
1 2012-09-20  Jakub Jelinek  <jakub@redhat.com>
2
3         PR other/43620
4         * configure.ac (AM_INIT_AUTOMAKE): Add no-dist.
5         * Makefile.in: Regenerated.
6         * include/Makefile.in: Regenerated.
7         * man/Makefile.in: Regenerated.
8         * testsuite/Makefile.in: Regenerated.
9
10 2012-03-22  David Edelsohn  <dje.gcc@gmail.com>
11
12         Backport from mainline:
13         2012-03-09  David Edelsohn  <dje.gcc@gmail.com>
14
15         * src/powerpc/aix_closure.S (ffi_closure_ASM): Adjust for Darwin64
16         change to return value of ffi_closure_helper_DARWIN and load type
17         from return type.
18
19         From Tom Honermann <tom.honermann@oracle.com>:
20         * src/powerpc/aix.S: Declare .ffi_prep_args.  Insert nops after
21         branch instructions.
22         * src/powerpc/aix_closure.S: Declare .ffi_closure_helper_DARWIN.
23
24 2012-03-01  Release Manager
25
26         * GCC 4.6.3 released.
27
28 2012-02-10  Kai Tietz  <ktietz@redhat.com>
29
30         * configure.ac (AM_LTLDFLAGS): Add -no-undefine for x64
31         windows target.
32         * configure: Regenerated
33
34 2011-11-20  Andreas Tobler  <andreast@fgznet.ch>
35
36         * configure: Regenerate.
37
38 2011-10-26  Release Manager
39
40         * GCC 4.6.2 released.
41
42 2011-10-04  Andrew Haley  <aph@redhat.com>
43
44         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Clear icache.
45
46 2011-09-04  Iain Sandoe  <iains@gcc.gnu.org>
47
48         PR libffi/49594
49         * src/powerpc/darwin_closure.S (stubs): Make the stub binding
50         helper reference track the architecture pointer size.
51
52 2011-06-29  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
53
54         * testsuite/libffi.call/cls_double_va.c: Move PR number to comment.
55         * testsuite/libffi.call/cls_longdouble_va.c: Likewise.
56
57 2011-06-29  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
58
59         PR libffi/46660
60         * testsuite/libffi.call/cls_double_va.c: xfail dg-output on
61         mips-sgi-irix6*.
62         * testsuite/libffi.call/cls_longdouble_va.c: Likewise.
63
64 2011-06-27  Release Manager
65
66         * GCC 4.6.1 released.
67
68 2011-05-02  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
69
70         Backport from mainline:
71         2011-04-29  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
72
73         * src/alpha/osf.S (UA_SI, FDE_ENCODING, FDE_ENCODE, FDE_ARANGE):
74         Define.
75         Use them to handle ELF vs. ECOFF differences.
76         [__osf__] (_GLOBAL__F_ffi_call_osf): Define.
77
78 2011-03-25  Release Manager
79
80         * GCC 4.6.0 released.
81
82 2011-02-13  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
83
84         * configure: Regenerate.
85
86 2011-02-09  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
87
88         PR libffi/46661
89         * testsuite/libffi.call/cls_pointer.c (main): Cast void * to
90         uintptr_t first.
91         * testsuite/libffi.call/cls_pointer_stack.c (main): Likewise.
92
93 2011-02-07  Joel Sherrill <joel.sherrill@oarcorp.com>
94
95         * libffi/src/m68k/ffi.c: Add RTEMS support for cache flushing.
96         Handle case when CPU variant does not have long double support.
97         * libffi/src/m68k/sysv.S: Add support for mc68000, Coldfire,
98         and cores with soft floating point.
99
100 2011-02-07  Joel Sherrill <joel.sherrill@oarcorp.com>
101
102         * configure.ac: Add mips*-*-rtems* support.
103         * configure: Regenerate.
104         * src/mips/ffitarget.h: Ensure needed constants are available
105         for targets which do not have sgidefs.h.
106
107 2011-01-26  Dave Korn  <dave.korn.cygwin@gmail.com>
108
109         PR target/40125
110         * configure.ac (AM_LTLDFLAGS): Add -bindir option for windows DLLs.
111         * configure: Regenerate.
112
113 2010-12-18  Iain Sandoe  <iains@gcc.gnu.org>
114
115         PR libffi/29152
116         PR libffi/42378
117         * src/powerpc/darwin_closure.S: Provide Darwin64 implementation,
118         update comments.
119         * src/powerpc/ffitarget.h (POWERPC_DARWIN64): New,
120         (FFI_TRAMPOLINE_SIZE): Update for Darwin64.
121         * src/powerpc/darwin.S: Provide Darwin64 implementation,
122         update comments.
123         * src/powerpc/ffi_darwin.c: Likewise.
124
125 2010-12-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
126
127         * configure.ac (libffi_cv_as_ascii_pseudo_op): Use double
128         backslashes.
129         (libffi_cv_as_string_pseudo_op): Likewise.
130         * configure: Regenerate.
131
132 2010-12-03  Chung-Lin Tang  <cltang@codesourcery.com>
133
134         * src/arm/sysv.S (ffi_closure_SYSV): Add UNWIND to .pad directive.
135         (ffi_closure_VFP): Same.
136         (ffi_call_VFP): Move down to before ffi_closure_VFP. Add '.fpu vfp'
137         directive.
138
139 2010-12-01  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
140
141         * testsuite/libffi.call/ffitest.h [__sgi] (PRId64, PRIu64): Define.
142         (PRIuPTR): Define.
143
144 2010-11-29  Richard Henderson  <rth@redhat.com>
145             Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
146
147         * src/x86/sysv.S (FDE_ENCODING, FDE_ENCODE): Define.
148         (.eh_frame): Use FDE_ENCODING.
149         (.LASFDE1, .LASFDE2, LASFDE3): Simplify with FDE_ENCODE.
150
151 2010-11-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
152
153         * testsuite/lib/libffi-dg.exp: Rename ...
154         * testsuite/lib/libffi.exp: ... to this.
155         * libffi/testsuite/libffi.call/call.exp: Don't load libffi-dg.exp.
156         * libffi/testsuite/libffi.special/special.exp: Likewise.
157
158 2010-10-28  Chung-Lin Tang  <cltang@codesourcery.com>
159
160         * src/arm/ffi.c (ffi_prep_args): Add VFP register argument handling
161         code, new parameter, and return value. Update comments.
162         (ffi_prep_cif_machdep): Add case for VFP struct return values. Add
163         call to layout_vfp_args().
164         (ffi_call_SYSV): Update declaration.
165         (ffi_call_VFP): New declaration.
166         (ffi_call): Add VFP struct return conditions. Call ffi_call_VFP()
167         when ABI is FFI_VFP.
168         (ffi_closure_VFP): New declaration.
169         (ffi_closure_SYSV_inner): Add new vfp_args parameter, update call to
170         ffi_prep_incoming_args_SYSV().
171         (ffi_prep_incoming_args_SYSV): Update parameters. Add VFP argument
172         case handling.
173         (ffi_prep_closure_loc): Pass ffi_closure_VFP to trampoline
174         construction under VFP hard-float.
175         (rec_vfp_type_p): New function.
176         (vfp_type_p): Same.
177         (place_vfp_arg): Same.
178         (layout_vfp_args): Same.
179         * src/arm/ffitarget.h (ffi_abi): Add FFI_VFP. Define FFI_DEFAULT_ABI
180         based on __ARM_PCS_VFP.
181         (FFI_EXTRA_CIF_FIELDS): Define for adding VFP hard-float specific
182         fields.
183         (FFI_TYPE_STRUCT_VFP_FLOAT): Define internally used type code.
184         (FFI_TYPE_STRUCT_VFP_DOUBLE): Same.
185         * src/arm/sysv.S (ffi_call_SYSV): Change call of ffi_prep_args() to
186         direct call. Move function pointer load upwards.
187         (ffi_call_VFP): New function.
188         (ffi_closure_VFP): Same.
189
190         * testsuite/lib/libffi-dg.exp (check-flags): New function.
191         (dg-skip-if): New function.
192         * testsuite/libffi.call/cls_double_va.c: Skip if target is arm*-*-*
193         and compiler options include -mfloat-abi=hard.
194         * testsuite/libffi.call/cls_longdouble_va.c: Same.
195
196 2010-10-01  Jakub Jelinek  <jakub@redhat.com>
197
198         PR libffi/45677
199         * src/x86/ffi64.c (ffi_prep_cif_machdep): Ensure cif->bytes is
200         a multiple of 8.
201         * testsuite/libffi.call/many2.c: New test.
202
203 2010-08-20  Mark Wielaard  <mjw@redhat.com>
204
205         * src/closures.c (open_temp_exec_file_mnt): Check if getmntent_r
206         returns NULL.
207
208 2010-08-09  Andreas Tobler  <andreast@fgznet.ch>
209
210         * configure.ac: Add target powerpc64-*-freebsd*.
211         * configure: Regenerate.
212         * testsuite/libffi.call/cls_align_longdouble_split.c: Pass
213         -mlong-double-128 only to linux targets.
214         * testsuite/libffi.call/cls_align_longdouble_split2.c: Likewise.
215         * testsuite/libffi.call/cls_longdouble.c: Likewise.
216         * testsuite/libffi.call/huge_struct.c: Likewise.
217
218 2010-07-10  Evan Phoenix  <evan@fallingsnow.net>
219
220         * src/closures.c (selinux_enabled_check): Fix strncmp usage bug.
221
222 2010-07-07  Dan HorĂ¡k <dan@danny.cz>
223
224         * include/ffi.h.in: Protect #define with #ifndef.
225         * src/powerpc/ffitarget.h: Ditto.
226         * src/s390/ffitarget.h: Ditto.
227         * src/sparc/ffitarget.h: Ditto.
228
229 2010-07-07   Neil Roberts <neil@linux.intel.com>
230
231         * src/x86/sysv.S (ffi_call_SYSV): Align the stack pointer to
232         16-bytes.
233
234 2010-07-02  Jakub Jelinek  <jakub@redhat.com>
235
236         * Makefile.am (AM_MAKEFLAGS): Pass also mandir to submakes.
237         * Makefile.in: Regenerated.
238
239 2010-05-19  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
240
241         * configure.ac (libffi_cv_as_x86_pcrel): Check for illegal in as
242         output, too.
243         (libffi_cv_as_ascii_pseudo_op): Check for .ascii.
244         (libffi_cv_as_string_pseudo_op): Check for .string.
245         * configure: Regenerate.
246         * fficonfig.h.in: Regenerate.
247         * src/x86/sysv.S (.eh_frame): Use .ascii, .string or error.
248
249 2010-05-05  Michael Kohler <michaelkohler@live.com>
250
251         * src/dlmalloc.c (dlfree): Fix spelling.
252         * src/ia64/ffi.c (ffi_prep_cif_machdep): Ditto.
253         * configure.ac: Ditto.
254         * configure: Rebuilt.
255
256 2010-04-13  Dan Witte  <dwitte@mozilla.com>
257
258         * msvcc.sh: Build with -W3 instead of -Wall.
259         * src/powerpc/ffi_darwin.c: Remove build warnings.
260         * src/x86/ffi.c: Ditto.
261         * src/x86/ffitarget.h: Ditto.
262
263 2010-04-12  Dan Witte  <dwitte@mozilla.com>
264             Walter Meinl <wuno@lsvw.de>
265
266         * configure.ac: Add OS/2 support.
267         * configure: Rebuilt.
268         * src/closures.c: Ditto.
269         * src/dlmalloc.c: Ditto.
270         * src/x86/win32.S: Ditto.
271
272 2010-04-07  Jakub Jelinek  <jakub@redhat.com>
273
274         * testsuite/libffi.call/err_bad_abi.c: Remove unused args variable.
275
276 2010-04-02  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
277
278         * Makefile.in: Regenerate.
279         * aclocal.m4: Regenerate.
280         * include/Makefile.in: Regenerate.
281         * man/Makefile.in: Regenerate.
282         * testsuite/Makefile.in: Regenerate.
283
284 2010-03-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
285
286         * configure.ac (libffi_cv_as_x86_64_unwind_section_type): New test.
287         * configure: Regenerate.
288         * fficonfig.h.in: Regenerate.
289         * libffi/src/x86/unix64.S (.eh_frame)
290         [HAVE_AS_X86_64_UNWIND_SECTION_TYPE]: Use @unwind section type.
291
292 2010-03-14  Matthias Klose  <doko@ubuntu.com>
293
294         * src/x86/ffi64.c: Fix typo in comment.
295         * src/x86/ffi.c: Use /* ... */ comment style.
296
297 2010-02-24  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
298
299         * doc/libffi.texi (The Closure API): Fix typo.
300         * doc/libffi.info: Remove.
301
302 2010-02-15  Matthias Klose  <doko@ubuntu.com>
303
304         * src/arm/sysv.S (__ARM_ARCH__): Define for processor
305         __ARM_ARCH_7EM__.
306
307 2010-01-15  Anthony Green  <green@redhat.com>
308
309         * README: Add notes on building with Microsoft Visual C++.
310
311 2010-01-15  Daniel Witte  <dwitte@mozilla.com>
312
313         * msvcc.sh: New file.
314
315         * src/x86/win32.S: Port assembly routines to MSVC and #ifdef.
316         * src/x86/ffi.c: Tweak function declaration and remove excess
317         parens.
318         * include/ffi.h.in: Add __declspec(align(8)) to typedef struct
319         ffi_closure.
320
321         * src/x86/ffi.c: Merge ffi_call_SYSV and ffi_call_STDCALL into new
322         function ffi_call_win32 on X86_WIN32.
323         * src/x86/win32.S (ffi_call_SYSV): Rename to ffi_call_win32.
324         (ffi_call_STDCALL): Remove.
325
326         * src/prep_cif.c (ffi_prep_cif): Move stack space allocation code
327         to ffi_prep_cif_machdep for x86.
328         * src/x86/ffi.c (ffi_prep_cif_machdep): To here.
329
330 2010-01-15  Oliver Kiddle  <okiddle@yahoo.co.uk>
331
332         * src/x86/ffitarget.h (ffi_abi): Check for __i386 and __amd64 for
333         Sun Studio compiler compatibility.
334
335 2010-01-12  Conrad Irwin <conrad.irwin@gmail.com>
336
337         * doc/libffi.texi: Add closure example.
338
339 2010-01-07  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
340
341         PR libffi/40701
342         * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRIdLL,
343         PRIuLL, PRId64, PRIu64, PRIuPTR): Define.
344         * testsuite/libffi.call/cls_align_sint64.c: Add -Wno-format on
345         alpha*-dec-osf*.
346         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
347         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
348         * testsuite/libffi.call/return_ll1.c: Likewise.
349         * testsuite/libffi.call/stret_medium2.c: Likewise.
350         * testsuite/libffi.special/ffitestcxx.h (allocate_mmap): Cast
351         MAP_FAILED to char *.
352
353 2010-01-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
354
355         * src/mips/n32.S: Use .abicalls and .eh_frame with __GNUC__.
356
357 2009-12-31  Anthony Green  <green@redhat.com>
358
359         * README: Update for libffi 3.0.9.
360
361 2009-12-27  Matthias Klose  <doko@ubuntu.com>
362
363         * configure.ac (HAVE_LONG_DOUBLE): Define for mips when
364         appropriate.
365         * configure: Rebuilt.
366
367 2009-12-26  Anthony Green  <green@redhat.com>
368
369         * testsuite/libffi.call/cls_longdouble_va.c: Mark as xfail for
370         avr32*-*-*.
371         * testsuite/libffi.call/cls_double_va.c: Ditto.
372
373 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
374
375         * testsuite/libffi.call/ffitest.h: Conditionally include stdint.h
376         and inttypes.h.
377         * testsuite/libffi.special/unwindtest.cc: Ditto.
378
379 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
380
381         * configure.ac: Add amd64-*-openbsd*.
382         * configure: Rebuilt.
383         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Link
384         openbsd programs with -lpthread.
385
386 2009-12-26  Anthony Green  <green@redhat.com>
387
388         * testsuite/libffi.call/cls_double_va.c,
389         testsuite/libffi.call/cls_longdouble.c,
390         testsuite/libffi.call/cls_longdouble_va.c,
391         testsuite/libffi.call/cls_pointer.c,
392         testsuite/libffi.call/cls_pointer_stack.c: Remove xfail for
393         mips*-*-* and arm*-*-*.
394         * testsuite/libffi.call/cls_align_longdouble_split.c,
395         testsuite/libffi.call/cls_align_longdouble_split2.c,
396         testsuite/libffi.call/stret_medium2.c,
397         testsuite/libffi.call/stret_medium.c,
398         testsuite/libffi.call/stret_large.c,
399         testsuite/libffi.call/stret_large2.c: Remove xfail for arm*-*-*.
400
401 2009-12-31  Kay Tietz  <ktietz70@googlemail.com>
402
403         * testsuite/libffi.call/ffitest.h,
404         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRuLL): Fix
405         definitions.
406
407 2009-12-31  Carlo Bramini  <carlo.bramix@libero.it>
408
409         * configure.ac (AM_LTLDFLAGS): Define for windows hosts.
410         * Makefile.am (libffi_la_LDFLAGS): Add AM_LTLDFLAGS.
411         * configure: Rebuilt.
412         * Makefile.in: Rebuilt.
413
414 2009-12-31  Anthony Green  <green@redhat.com>
415             Blake Chaffin.
416
417         * testsuite/libffi.call/huge_struct.c: New test case from Blake
418         Chaffin @ Apple.
419
420 2009-12-28  David Edelsohn  <edelsohn@gnu.org>
421
422         * src/powerpc/ffi_darwin.c (ffi_prep_args): Copy abi and nargs to
423         local variables.
424         (aix_adjust_aggregate_sizes): New function.
425         (ffi_prep_cif_machdep): Call it.
426
427 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
428
429         * configure.ac: Define FFI_MMAP_EXEC_WRIT for the given targets.
430         * configure: Regenerate.
431         * fficonfig.h.in: Likewise.
432         * src/closures.c: Remove the FFI_MMAP_EXEC_WRIT definition for
433         Solaris/x86.
434
435 2009-12-26  Andreas Schwab  <schwab@linux-m68k.org>
436
437         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Advance intarg_count
438         when a float arguments is passed in memory.
439         (ffi_closure_helper_SYSV): Mark general registers as used up when
440         a 64bit or soft-float long double argument is passed in memory.
441
442 2009-12-25  Matthias Klose  <doko@ubuntu.com>
443
444         * man/ffi_call.3: Fix #include in examples.
445         * doc/libffi.texi: Add dircategory.
446
447 2009-12-25  Frank Everdij <f.p.x.everdij@tudelft.nl>
448
449         * include/ffi.h.in: Placed '__GNUC__' ifdef around
450         '__attribute__((aligned(8)))' in ffi_closure, fixes compile for
451         IRIX MIPSPro c99.
452         * include/ffi_common.h: Added '__sgi' define to non
453         '__attribute__((__mode__()))' integer typedefs.
454         * src/mips/ffi.c (ffi_call, ffi_closure_mips_inner_O32,
455         ffi_closure_mips_inner_N32): Added 'defined(_MIPSEB)' to BE check.
456         (ffi_closure_mips_inner_O32, ffi_closure_mips_inner_N32): Added
457         FFI_LONGDOUBLE support and alignment(N32 only).
458         * src/mips/ffitarget.h: Corrected '#include <sgidefs.h>' for IRIX and
459         fixed non '__attribute__((__mode__()))' integer typedefs.
460         * src/mips/n32.S: Put '#ifdef linux' around '.abicalls' and '.eh_frame'
461         since they are Linux/GNU Assembler specific.
462
463 2009-12-25  Bradley Smith  <brad@brad-smith.co.uk>
464
465         * configure.ac, Makefile.am, src/avr32/ffi.c,
466         src/avr32/ffitarget.h,
467         src/avr32/sysv.S: Add AVR32 port.
468         * configure, Makefile.in: Rebuilt.
469
470 2009-12-21  Andreas Tobler  <a.tobler@schweiz.org>
471
472         * configure.ac: Make i?86 build on FreeBSD and OpenBSD.
473         * configure: Regenerate.
474
475 2009-12-15  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
476
477         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on PA HP-UX.
478
479 2009-12-13  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
480
481         * src/pa/ffi.c (ffi_closure_inner_pa32): Handle FFI_TYPE_LONGDOUBLE
482         type on HP-UX.
483
484 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
485
486         * src/sparc/ffi.c (ffi_closure_sparc_inner_v9): Properly align 'long
487         double' arguments.
488
489 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
490
491         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on Solaris < 10.
492
493 2009-12-10  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
494
495         PR libffi/40700
496         * src/closures.c [X86_64 && __sun__ && __svr4__]
497         (FFI_MMAP_EXEC_WRIT): Define.
498
499 2009-12-08  David Daney  <ddaney@caviumnetworks.com>
500
501         * testsuite/libffi.call/stret_medium.c: Remove xfail for mips*-*-*
502         * testsuite/libffi.call/cls_align_longdouble_split2.c: Same.
503         * testsuite/libffi.call/stret_large.c: Same.
504         * testsuite/libffi.call/cls_align_longdouble_split.c: Same.
505         * testsuite/libffi.call/stret_large2.c: Same.
506         * testsuite/libffi.call/stret_medium2.c: Same.
507
508 2009-12-07  David Edelsohn  <edelsohn@gnu.org>
509
510         * src/powerpc/aix_closure.S (libffi_closure_ASM): Fix tablejump
511         typo.
512
513 2009-12-05  David Edelsohn  <edelsohn@gnu.org>
514
515         * src/powerpc/aix.S: Update AIX32 code to be consistent with AIX64
516         code.
517         * src/powerpc/aix_closure.S: Same.
518
519 2009-12-05  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
520
521         * Makefile.in: Regenerate.
522         * configure: Regenerate.
523         * include/Makefile.in: Regenerate.
524         * man/Makefile.in: Regenerate.
525         * testsuite/Makefile.in: Regenerate.
526
527 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
528
529         * src/powerpc/aix_closure.S: Reorganize 64-bit code to match
530         linux64_closure.S.
531
532 2009-12-04  Uros Bizjak  <ubizjak@gmail.com>
533
534         PR libffi/41908
535         * src/x86/ffi64.c (classify_argument): Update from
536         gcc/config/i386/i386.c.
537         (ffi_closure_unix64_inner): Do not use the address of two consecutive
538         SSE registers directly.
539         * testsuite/libffi.call/cls_dbls_struct.c (main): Remove xfail
540         for x86_64 linux targets.
541
542 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
543
544         * src/powerpc/ffi_darwin.c (ffi_closure_helper_DARWIN): Increment
545         pfr for long double split between fpr13 and stack.
546
547 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
548
549         * src/powerpc/ffi_darwin.c (ffi_prep_args): Increment next_arg and
550         fparg_count twice for long double.
551
552 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
553
554         PR libffi/42243
555         * src/powerpc/ffi_darwin.c (ffi_prep_args): Remove extra parentheses.
556
557 2009-12-03  Uros Bizjak  <ubizjak@gmail.com>
558
559         * testsuite/libffi.call/cls_longdouble_va.c (main): Fix format string.
560         Remove xfails for x86 linux targets.
561
562 2009-12-02  David Edelsohn  <edelsohn@gnu.org>
563
564         * src/powerpc/ffi_darwin.c (ffi_prep_args): Fix typo in INT64
565         case.
566
567 2009-12-01  David Edelsohn  <edelsohn@gnu.org>
568
569         * src/powerpc/aix.S (ffi_call_AIX): Convert to more standard
570         register usage.  Call ffi_prep_args directly.  Add long double
571         return value support.
572         * src/powerpc/ffi_darwin.c (ffi_prep_args): Double arg increment
573         applies to FFI_TYPE_DOUBLE.  Correct fpr_base increment typo.
574         Separate FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases.
575         (ffi_prep_cif_machdep): Only 16 byte stack alignment in 64 bit
576         mode.
577         (ffi_closure_helper_DARWIN): Remove nf and ng counters.  Move temp
578         into case.
579         * src/powerpc/aix_closure.S: Maintain 16 byte stack alignment.
580         Allocate result area between params and FPRs.
581
582 2009-11-30  David Edelsohn  <edelsohn@gnu.org>
583
584         PR target/35484
585         * src/powerpc/ffitarget.h (POWERPC64): Define for PPC64 Linux and
586         AIX64.
587         * src/powerpc/aix.S: Implement AIX64 version.
588         * src/powerpc/aix_closure.S: Implement AIX64 version.
589         (ffi_closure_ASM): Use extsb, lha and displament addresses.
590         * src/powerpc/ffi_darwin.c (ffi_prep_args): Implement AIX64
591         support.
592         (ffi_prep_cif_machdep): Same.
593         (ffi_call): Same.
594         (ffi_closure_helper_DARWIN): Same.
595
596 2009-11-02  Andreas Tobler  <a.tobler@schweiz.org>
597
598         PR libffi/41908
599         * testsuite/libffi.call/testclosure.c: New test.
600
601 2009-09-28  Kai Tietz  <kai.tietz@onevision.com>
602
603         * src/x86/win64.S (_ffi_call_win64 stack): Remove for gnu
604         assembly version use of ___chkstk.
605
606 2009-09-23  Matthias Klose  <doko@ubuntu.com>
607
608         PR libffi/40242, PR libffi/41443
609         * src/arm/sysv.S (__ARM_ARCH__): Define for processors
610         __ARM_ARCH_6T2__, __ARM_ARCH_6M__, __ARM_ARCH_7__,
611         __ARM_ARCH_7A__, __ARM_ARCH_7R__, __ARM_ARCH_7M__.
612         Change the conditionals to __SOFTFP__ || __ARM_EABI__
613         for -mfloat-abi=softfp to work.
614
615 2009-09-17  Loren J. Rittle  <ljrittle@acm.org>
616
617         PR testsuite/32843 (strikes again)
618         * src/x86/ffi.c (ffi_prep_cif_machdep): Add X86_FREEBSD to
619         enable proper extension on char and short.
620
621 2009-09-15  David Daney  <ddaney@caviumnetworks.com>
622
623         * src/java_raw_api.c (ffi_java_raw_to_rvalue): Remove special
624         handling for FFI_TYPE_POINTER.
625         * src/mips/ffitarget.h (FFI_TYPE_STRUCT_D_SOFT,
626         FFI_TYPE_STRUCT_F_SOFT, FFI_TYPE_STRUCT_DD_SOFT,
627         FFI_TYPE_STRUCT_FF_SOFT, FFI_TYPE_STRUCT_FD_SOFT,
628         FFI_TYPE_STRUCT_DF_SOFT, FFI_TYPE_STRUCT_SOFT): New defines.
629         (FFI_N32_SOFT_FLOAT, FFI_N64_SOFT_FLOAT): New ffi_abi enumerations.
630         (enum ffi_abi): Set FFI_DEFAULT_ABI for soft-float.
631         * src/mips/n32.S (ffi_call_N32): Add handling for soft-float
632         structure and pointer returns.
633         (ffi_closure_N32): Add handling for pointer returns.
634         * src/mips/ffi.c (ffi_prep_args, calc_n32_struct_flags,
635         calc_n32_return_struct_flags): Handle soft-float.
636         (ffi_prep_cif_machdep):  Handle soft-float, fix pointer handling.
637         (ffi_call_N32): Declare proper argument types.
638         (ffi_call, copy_struct_N32, ffi_closure_mips_inner_N32): Handle
639         soft-float.
640
641 2009-08-24  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
642
643         * configure.ac (AC_PREREQ): Bump to 2.64.
644
645 2009-08-22  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
646
647         * Makefile.am (install-html, install-pdf): Remove.
648         * Makefile.in: Regenerate.
649
650         * Makefile.in: Regenerate.
651         * aclocal.m4: Regenerate.
652         * configure: Regenerate.
653         * fficonfig.h.in: Regenerate.
654         * include/Makefile.in: Regenerate.
655         * man/Makefile.in: Regenerate.
656         * testsuite/Makefile.in: Regenerate.
657
658 2009-07-30  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
659
660         * configure.ac (_AC_ARG_VAR_PRECIOUS): Use m4_rename_force.
661
662 2009-07-24  Dave Korn  <dave.korn.cygwin@gmail.com>
663
664         PR libffi/40807
665         * src/x86/ffi.c (ffi_prep_cif_machdep): Also use sign/zero-extending
666         return types for X86_WIN32.
667         * src/x86/win32.S (_ffi_call_SYSV): Handle omitted return types.
668         (_ffi_call_STDCALL, _ffi_closure_SYSV, _ffi_closure_raw_SYSV,
669         _ffi_closure_STDCALL): Likewise.
670
671         * src/closures.c (is_selinux_enabled): Define to const 0 for Cygwin.
672         (dlmmap, dlmunmap): Also use these functions on Cygwin.
673
674 2009-07-11  Richard Sandiford  <rdsandiford@googlemail.com>
675
676         PR testsuite/40699
677         PR testsuite/40707
678         PR testsuite/40709
679         * testsuite/lib/libffi-dg.exp: Revert 2009-07-02, 2009-07-01 and
680         2009-06-30 commits.
681
682 2009-07-01  Richard Sandiford  <r.sandiford@uk.ibm.com>
683
684         * testsuite/lib/libffi-dg.exp (libffi-init): Set ld_library_path
685         to "" before adding paths.  (This reinstates an assignment that
686         was removed by my 2009-06-30 commit, but changes the initial
687         value from "." to "".)
688
689 2009-07-01  H.J. Lu  <hongjiu.lu@intel.com>
690
691         PR testsuite/40601
692         * testsuite/lib/libffi-dg.exp (libffi-init): Properly set
693         gccdir.  Adjust ld_library_path for gcc only if gccdir isn't
694         empty.
695
696 2009-06-30  Richard Sandiford  <r.sandiford@uk.ibm.com>
697
698         * testsuite/lib/libffi-dg.exp (libffi-init): Don't add "."
699         to ld_library_path.  Use add_path.  Add just find_libgcc_s
700         to ld_library_path, not every libgcc multilib directory.
701
702 2009-06-16  Wim Lewis  <wiml@hhhh.org>
703
704         * src/powerpc/ffi.c: Avoid clobbering cr3 and cr4, which are
705         supposed to be callee-saved.
706         * src/powerpc/sysv.S (small_struct_return_value): Fix overrun of
707         return buffer for odd-size structs.
708
709 2009-06-16  Andreas Tobler  <a.tobler@schweiz.org>
710
711         PR libffi/40444
712         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Add
713         allow_stack_execute for Darwin.
714
715 2009-06-16  Andrew Haley  <aph@redhat.com>
716
717         * configure.ac (TARGETDIR): Add missing blank lines.
718         * configure: Regenerate.
719
720 2009-06-16  Andrew Haley  <aph@redhat.com>
721
722         * testsuite/libffi.call/cls_align_sint64.c,
723         testsuite/libffi.call/cls_align_uint64.c,
724         testsuite/libffi.call/cls_longdouble_va.c,
725         testsuite/libffi.call/cls_ulonglong.c,
726         testsuite/libffi.call/return_ll1.c,
727         testsuite/libffi.call/stret_medium2.c: Fix printf format
728         specifiers.
729         * testsuite/libffi.call/ffitest.h,
730         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRIuLL): Define.
731
732 2009-06-15  Andrew Haley  <aph@redhat.com>
733
734         * testsuite/libffi.call/err_bad_typedef.c: xfail everywhere.
735         * testsuite/libffi.call/err_bad_abi.c: Likewise.
736
737 2009-06-12  Andrew Haley  <aph@redhat.com>
738
739         * Makefile.am: Remove info_TEXINFOS.
740
741 2009-06-12  Andrew Haley  <aph@redhat.com>
742
743         * ChangeLog.libffi: testsuite/libffi.call/cls_align_sint64.c,
744         testsuite/libffi.call/cls_align_uint64.c,
745         testsuite/libffi.call/cls_ulonglong.c,
746         testsuite/libffi.call/return_ll1.c,
747         testsuite/libffi.call/stret_medium2.c: Fix printf format
748         specifiers.
749         testsuite/libffi.special/unwindtest.cc: include stdint.h.
750
751 2009-06-11  Timothy Wall  <twall@users.sf.net>
752
753         * Makefile.am,
754         configure.ac,
755         include/ffi.h.in,
756         include/ffi_common.h,
757         src/closures.c,
758         src/dlmalloc.c,
759         src/x86/ffi.c,
760         src/x86/ffitarget.h,
761         src/x86/win64.S (new),
762         README: Added win64 support (mingw or MSVC)
763         * Makefile.in,
764         include/Makefile.in,
765         man/Makefile.in,
766         testsuite/Makefile.in,
767         configure,
768         aclocal.m4: Regenerated
769         * ltcf-c.sh: properly escape cygwin/w32 path
770         * man/ffi_call.3: Clarify size requirements for return value.
771         * src/x86/ffi64.c: Fix filename in comment.
772         * src/x86/win32.S: Remove unused extern.
773
774         * testsuite/libffi.call/closure_fn0.c,
775         testsuite/libffi.call/closure_fn1.c,
776         testsuite/libffi.call/closure_fn2.c,
777         testsuite/libffi.call/closure_fn3.c,
778         testsuite/libffi.call/closure_fn4.c,
779         testsuite/libffi.call/closure_fn5.c,
780         testsuite/libffi.call/closure_fn6.c,
781         testsuite/libffi.call/closure_stdcall.c,
782         testsuite/libffi.call/cls_12byte.c,
783         testsuite/libffi.call/cls_16byte.c,
784         testsuite/libffi.call/cls_18byte.c,
785         testsuite/libffi.call/cls_19byte.c,
786         testsuite/libffi.call/cls_1_1byte.c,
787         testsuite/libffi.call/cls_20byte.c,
788         testsuite/libffi.call/cls_20byte1.c,
789         testsuite/libffi.call/cls_24byte.c,
790         testsuite/libffi.call/cls_2byte.c,
791         testsuite/libffi.call/cls_3_1byte.c,
792         testsuite/libffi.call/cls_3byte1.c,
793         testsuite/libffi.call/cls_3byte2.c,
794         testsuite/libffi.call/cls_4_1byte.c,
795         testsuite/libffi.call/cls_4byte.c,
796         testsuite/libffi.call/cls_5_1_byte.c,
797         testsuite/libffi.call/cls_5byte.c,
798         testsuite/libffi.call/cls_64byte.c,
799         testsuite/libffi.call/cls_6_1_byte.c,
800         testsuite/libffi.call/cls_6byte.c,
801         testsuite/libffi.call/cls_7_1_byte.c,
802         testsuite/libffi.call/cls_7byte.c,
803         testsuite/libffi.call/cls_8byte.c,
804         testsuite/libffi.call/cls_9byte1.c,
805         testsuite/libffi.call/cls_9byte2.c,
806         testsuite/libffi.call/cls_align_double.c,
807         testsuite/libffi.call/cls_align_float.c,
808         testsuite/libffi.call/cls_align_longdouble.c,
809         testsuite/libffi.call/cls_align_longdouble_split.c,
810         testsuite/libffi.call/cls_align_longdouble_split2.c,
811         testsuite/libffi.call/cls_align_pointer.c,
812         testsuite/libffi.call/cls_align_sint16.c,
813         testsuite/libffi.call/cls_align_sint32.c,
814         testsuite/libffi.call/cls_align_sint64.c,
815         testsuite/libffi.call/cls_align_uint16.c,
816         testsuite/libffi.call/cls_align_uint32.c,
817         testsuite/libffi.call/cls_align_uint64.c,
818         testsuite/libffi.call/cls_dbls_struct.c,
819         testsuite/libffi.call/cls_double.c,
820         testsuite/libffi.call/cls_double_va.c,
821         testsuite/libffi.call/cls_float.c,
822         testsuite/libffi.call/cls_longdouble.c,
823         testsuite/libffi.call/cls_longdouble_va.c,
824         testsuite/libffi.call/cls_multi_schar.c,
825         testsuite/libffi.call/cls_multi_sshort.c,
826         testsuite/libffi.call/cls_multi_sshortchar.c,
827         testsuite/libffi.call/cls_multi_uchar.c,
828         testsuite/libffi.call/cls_multi_ushort.c,
829         testsuite/libffi.call/cls_multi_ushortchar.c,
830         testsuite/libffi.call/cls_pointer.c,
831         testsuite/libffi.call/cls_pointer_stack.c,
832         testsuite/libffi.call/cls_schar.c,
833         testsuite/libffi.call/cls_sint.c,
834         testsuite/libffi.call/cls_sshort.c,
835         testsuite/libffi.call/cls_uchar.c,
836         testsuite/libffi.call/cls_uint.c,
837         testsuite/libffi.call/cls_ulonglong.c,
838         testsuite/libffi.call/cls_ushort.c,
839         testsuite/libffi.call/err_bad_abi.c,
840         testsuite/libffi.call/err_bad_typedef.c,
841         testsuite/libffi.call/float2.c,
842         testsuite/libffi.call/huge_struct.c,
843         testsuite/libffi.call/nested_struct.c,
844         testsuite/libffi.call/nested_struct1.c,
845         testsuite/libffi.call/nested_struct10.c,
846         testsuite/libffi.call/nested_struct2.c,
847         testsuite/libffi.call/nested_struct3.c,
848         testsuite/libffi.call/nested_struct4.c,
849         testsuite/libffi.call/nested_struct5.c,
850         testsuite/libffi.call/nested_struct6.c,
851         testsuite/libffi.call/nested_struct7.c,
852         testsuite/libffi.call/nested_struct8.c,
853         testsuite/libffi.call/nested_struct9.c,
854         testsuite/libffi.call/problem1.c,
855         testsuite/libffi.call/return_ldl.c,
856         testsuite/libffi.call/return_ll1.c,
857         testsuite/libffi.call/stret_large.c,
858         testsuite/libffi.call/stret_large2.c,
859         testsuite/libffi.call/stret_medium.c,
860         testsuite/libffi.call/stret_medium2.c,
861         testsuite/libffi.special/unwindtest.cc: use ffi_closure_alloc instead
862         of checking for MMAP.  Use intptr_t instead of long casts.
863
864 2009-06-11  Kaz Kojima  <kkojima@gcc.gnu.org>
865
866         * testsuite/libffi.call/cls_longdouble_va.c: Add xfail sh*-*-linux-*.
867         * testsuite/libffi.call/err_bad_abi.c: Add xfail sh*-*-*.
868         * testsuite/libffi.call/err_bad_typedef.c: Likewise.
869
870 2009-06-09  Andrew Haley  <aph@redhat.com>
871
872         * src/x86/freebsd.S: Add missing file.
873
874 2009-06-08  Andrew Haley  <aph@redhat.com>
875
876         Import from libffi 3.0.8:
877
878         * doc/libffi.texi: New file.
879         * doc/libffi.info: Likewise.
880         * doc/stamp-vti: Likewise.
881         * man/Makefile.am: New file.
882         * man/ffi_call.3: New file.
883
884         * Makefile.am (EXTRA_DIST): Add src/x86/darwin64.S,
885         src/dlmalloc.c.
886         (nodist_libffi_la_SOURCES): Add X86_FREEBSD.
887
888         * configure.ac: Bump version to 3.0.8.
889         parisc*-*-linux*: Add.
890         i386-*-freebsd* | i386-*-openbsd*: Add.
891         powerpc-*-beos*: Add.
892         AM_CONDITIONAL X86_FREEBSD: Add.
893         AC_CONFIG_FILES: Add man/Makefile.
894
895         * include/ffi.h.in (FFI_FN): Change void (*)() to void (*)(void).
896
897 2009-06-08  Andrew Haley  <aph@redhat.com>
898
899         * README: Import from libffi 3.0.8.
900
901 2009-06-08  Andrew Haley  <aph@redhat.com>
902
903         * testsuite/libffi.call/err_bad_abi.c: Add xfails.
904         * testsuite/libffi.call/cls_longdouble_va.c: Add xfails.
905         * testsuite/libffi.call/cls_dbls_struct.c: Add xfail x86_64-*-linux-*.
906         * testsuite/libffi.call/err_bad_typedef.c: Add xfails.
907
908         * testsuite/libffi.call/stret_medium2.c: Add __UNUSED__ to args.
909         * testsuite/libffi.call/stret_medium.c: Likewise.
910         * testsuite/libffi.call/stret_large2.c: Likewise.
911         * testsuite/libffi.call/stret_large.c:  Likewise.
912
913 2008-12-26  Timothy Wall  <twall@users.sf.net>
914
915         * testsuite/libffi.call/cls_longdouble.c,
916         testsuite/libffi.call/cls_longdouble_va.c,
917         testsuite/libffi.call/cls_align_longdouble.c,
918         testsuite/libffi.call/cls_align_longdouble_split.c,
919         testsuite/libffi.call/cls_align_longdouble_split2.c: mark expected
920         failures on x86_64 cygwin/mingw.
921
922 2008-12-22  Timothy Wall  <twall@users.sf.net>
923
924         * testsuite/libffi.call/closure_fn0.c,
925         testsuite/libffi.call/closure_fn1.c,
926         testsuite/libffi.call/closure_fn2.c,
927         testsuite/libffi.call/closure_fn3.c,
928         testsuite/libffi.call/closure_fn4.c,
929         testsuite/libffi.call/closure_fn5.c,
930         testsuite/libffi.call/closure_fn6.c,
931         testsuite/libffi.call/closure_loc_fn0.c,
932         testsuite/libffi.call/closure_stdcall.c,
933         testsuite/libffi.call/cls_align_pointer.c,
934         testsuite/libffi.call/cls_pointer.c,
935         testsuite/libffi.call/cls_pointer_stack.c: use portable cast from
936         pointer to integer (intptr_t).
937         * testsuite/libffi.call/cls_longdouble.c: disable for win64.
938
939 2008-07-24  Anthony Green  <green@redhat.com>
940
941         * testsuite/libffi.call/cls_dbls_struct.c,
942         testsuite/libffi.call/cls_double_va.c,
943         testsuite/libffi.call/cls_longdouble.c,
944         testsuite/libffi.call/cls_longdouble_va.c,
945         testsuite/libffi.call/cls_pointer.c,
946         testsuite/libffi.call/cls_pointer_stack.c,
947         testsuite/libffi.call/err_bad_abi.c: Clean up failures from
948         compiler warnings.
949
950 2008-03-04  Anthony Green  <green@redhat.com>
951             Blake Chaffin
952             hos@tamanegi.org
953
954         * testsuite/libffi.call/cls_align_longdouble_split2.c
955           testsuite/libffi.call/cls_align_longdouble_split.c
956           testsuite/libffi.call/cls_dbls_struct.c
957           testsuite/libffi.call/cls_double_va.c
958           testsuite/libffi.call/cls_longdouble.c
959           testsuite/libffi.call/cls_longdouble_va.c
960           testsuite/libffi.call/cls_pointer.c
961           testsuite/libffi.call/cls_pointer_stack.c
962           testsuite/libffi.call/err_bad_abi.c
963           testsuite/libffi.call/err_bad_typedef.c
964           testsuite/libffi.call/stret_large2.c
965           testsuite/libffi.call/stret_large.c
966           testsuite/libffi.call/stret_medium2.c
967           testsuite/libffi.call/stret_medium.c: New tests from Apple.
968
969 2009-06-05  Andrew Haley  <aph@redhat.com>
970
971         * src/x86/ffitarget.h, src/x86/ffi.c: Merge stdcall changes from
972         libffi.
973
974 2009-06-04  Andrew Haley  <aph@redhat.com>
975
976         * src/x86/ffitarget.h, src/x86/win32.S, src/x86/ffi.c: Back out
977         stdcall changes.
978
979 2008-02-26  Anthony Green  <green@redhat.com>
980             Thomas Heller  <theller@ctypes.org>
981
982         * src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C
983         comment.
984
985 2008-02-03  Timothy Wall  <twall@users.sf.net>
986
987         * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return
988           offset based on code pointer, not data pointer.
989
990 2008-01-31  Timothy Wall <twall@users.sf.net>
991
992         * testsuite/libffi.call/closure_stdcall.c: Add test for stdcall
993         closures.
994         * src/x86/ffitarget.h: Increase size of trampoline for stdcall
995         closures.
996         * src/x86/win32.S: Add assembly for stdcall closure.
997         * src/x86/ffi.c: Initialize stdcall closure trampoline.
998
999 2009-06-04  Andrew Haley  <aph@redhat.com>
1000
1001         * include/ffi.h.in: Change void (*)() to void (*)(void).
1002         * src/x86/ffi.c: Likewise.
1003
1004 2009-06-04  Andrew Haley  <aph@redhat.com>
1005
1006         * src/powerpc/ppc_closure.S: Insert licence header.
1007         * src/powerpc/linux64_closure.S: Likewise.
1008         * src/m68k/sysv.S: Likewise.
1009
1010         * src/sh64/ffi.c: Change void (*)() to void (*)(void).
1011         * src/powerpc/ffi.c: Likewise.
1012         * src/powerpc/ffi_darwin.c: Likewise.
1013         * src/m32r/ffi.c: Likewise.
1014         * src/sh64/ffi.c: Likewise.
1015         * src/x86/ffi64.c: Likewise.
1016         * src/alpha/ffi.c: Likewise.
1017         * src/alpha/osf.S: Likewise.
1018         * src/frv/ffi.c: Likewise.
1019         * src/s390/ffi.c: Likewise.
1020         * src/pa/ffi.c: Likewise.
1021         * src/pa/hpux32.S: Likewise.
1022         * src/ia64/unix.S: Likewise.
1023         * src/ia64/ffi.c: Likewise.
1024         * src/sparc/ffi.c: Likewise.
1025         * src/mips/ffi.c: Likewise.
1026         * src/sh/ffi.c: Likewise.
1027
1028 2008-02-15  David Daney  <ddaney@avtrex.com>
1029
1030         * src/mips/ffi.c (USE__BUILTIN___CLEAR_CACHE):
1031         Define (conditionally), and use it to include cachectl.h.
1032         (ffi_prep_closure_loc): Fix cache flushing.
1033         * src/mips/ffitarget.h (_ABIN32, _ABI64, _ABIO32): Define.
1034
1035 2009-06-04  Andrew Haley  <aph@redhat.com>
1036
1037         include/ffi.h.in,
1038         src/arm/ffitarget.h,
1039         src/arm/ffi.c,
1040         src/arm/sysv.S,
1041         src/powerpc/ffitarget.h,
1042         src/closures.c,
1043         src/sh64/ffitarget.h,
1044         src/sh64/ffi.c,
1045         src/sh64/sysv.S,
1046         src/types.c,
1047         src/x86/ffi64.c,
1048         src/x86/ffitarget.h,
1049         src/x86/win32.S,
1050         src/x86/darwin.S,
1051         src/x86/ffi.c,
1052         src/x86/sysv.S,
1053         src/x86/unix64.S,
1054         src/alpha/ffitarget.h,
1055         src/alpha/ffi.c,
1056         src/alpha/osf.S,
1057         src/m68k/ffitarget.h,
1058         src/frv/ffitarget.h,
1059         src/frv/ffi.c,
1060         src/s390/ffitarget.h,
1061         src/s390/sysv.S,
1062         src/cris/ffitarget.h,
1063         src/pa/linux.S,
1064         src/pa/ffitarget.h,
1065         src/pa/ffi.c,
1066         src/raw_api.c,
1067         src/ia64/ffitarget.h,
1068         src/ia64/unix.S,
1069         src/ia64/ffi.c,
1070         src/ia64/ia64_flags.h,
1071         src/java_raw_api.c,
1072         src/debug.c,
1073         src/sparc/v9.S,
1074         src/sparc/ffitarget.h,
1075         src/sparc/ffi.c,
1076         src/sparc/v8.S,
1077         src/mips/ffitarget.h,
1078         src/mips/n32.S,
1079         src/mips/o32.S,
1080         src/mips/ffi.c,
1081         src/prep_cif.c,
1082         src/sh/ffitarget.h,
1083         src/sh/ffi.c,
1084         src/sh/sysv.S: Update license text.
1085
1086 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
1087
1088         * src/x86/win32.S (_ffi_closure_STDCALL):  New function.
1089         (.eh_frame):  Add FDE for it.
1090
1091 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
1092
1093         * configure.ac:  Also check if assembler supports pc-relative
1094         relocs on X86_WIN32 targets.
1095         * configure:  Regenerate.
1096         * src/x86/win32.S (ffi_prep_args):  Declare extern, not global.
1097         (_ffi_call_SYSV):  Add missing function type symbol .def and
1098         add EH markup labels.
1099         (_ffi_call_STDCALL):  Likewise.
1100         (_ffi_closure_SYSV):  Likewise.
1101         (_ffi_closure_raw_SYSV):  Likewise.
1102         (.eh_frame):  Add hand-crafted EH data.
1103
1104 2009-04-09  Jakub Jelinek  <jakub@redhat.com>
1105
1106         * testsuite/lib/libffi-dg.exp: Change copyright header to refer to
1107         version 3 of the GNU General Public License and to point readers
1108         at the COPYING3 file and the FSF's license web page.
1109         * testsuite/libffi.call/call.exp: Likewise.
1110         * testsuite/libffi.special/special.exp: Likewise.
1111
1112 2009-03-01  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1113
1114         * configure: Regenerate.
1115
1116 2008-12-18  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1117
1118         PR libffi/26048
1119         * configure.ac (HAVE_AS_X86_PCREL): New test.
1120         * configure: Regenerate.
1121         * fficonfig.h.in: Regenerate.
1122         * src/x86/sysv.S [!FFI_NO_RAW_API]: Precalculate
1123         RAW_CLOSURE_CIF_OFFSET, RAW_CLOSURE_FUN_OFFSET,
1124         RAW_CLOSURE_USER_DATA_OFFSET for the Solaris 10/x86 assembler.
1125         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
1126         * src/x86/unix64.S (.Lstore_table): Move to .text section.
1127         (.Lload_table): Likewise.
1128         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
1129
1130 2008-12-18  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1131
1132         * configure: Regenerate.
1133
1134 2008-11-21  Eric Botcazou  <ebotcazou@adacore.com>
1135
1136         * src/sparc/ffi.c (ffi_prep_cif_machdep): Add support for
1137         signed/unsigned int8/16 return values.
1138         * src/sparc/v8.S (ffi_call_v8): Likewise.
1139         (ffi_closure_v8): Likewise.
1140
1141 2008-09-26  Peter O'Gorman  <pogma@thewrittenword.com>
1142             Steve Ellcey  <sje@cup.hp.com>
1143
1144         * configure: Regenerate for new libtool.
1145         * Makefile.in: Ditto.
1146         * include/Makefile.in: Ditto.
1147         * aclocal.m4: Ditto.
1148
1149 2008-08-25  Andreas Tobler  <a.tobler@schweiz.org>
1150
1151         * src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and
1152         FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum.
1153         Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT.
1154         Adjust copyright notice.
1155         * src/powerpc/ffi.c: Add two new flags to indicate if we have one
1156         register or two register to use for FFI_SYSV structs.
1157         (ffi_prep_cif_machdep): Pass the right register flag introduced above.
1158         (ffi_closure_helper_SYSV): Fix the return type for
1159         FFI_SYSV_TYPE_SMALL_STRUCT. Comment.
1160         Adjust copyright notice.
1161
1162 2008-07-16  Kaz Kojima  <kkojima@gcc.gnu.org>
1163
1164         * src/sh/ffi.c (ffi_prep_closure_loc): Turn INSN into an unsigned
1165         int.
1166
1167 2008-06-17  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1168
1169         * configure: Regenerate.
1170         * include/Makefile.in: Regenerate.
1171         * testsuite/Makefile.in: Regenerate.
1172
1173 2008-06-07  Joseph Myers  <joseph@codesourcery.com>
1174
1175         * configure.ac (parisc*-*-linux*, powerpc-*-sysv*,
1176         powerpc-*-beos*): Remove.
1177         * configure: Regenerate.
1178
1179 2008-05-09  Julian Brown  <julian@codesourcery.com>
1180
1181         * Makefile.am (LTLDFLAGS): New.
1182         (libffi_la_LDFLAGS): Use above.
1183         * Makefile.in: Regenerate.
1184
1185 2008-04-18  Paolo Bonzini  <bonzini@gnu.org>
1186
1187         PR bootstrap/35457
1188         * aclocal.m4: Regenerate.
1189         * configure: Regenerate.
1190
1191 2008-03-26  Kaz Kojima  <kkojima@gcc.gnu.org>
1192
1193         * src/sh/sysv.S: Add .note.GNU-stack on Linux.
1194         * src/sh64/sysv.S: Likewise.
1195
1196 2008-03-26  Daniel Jacobowitz  <dan@debian.org>
1197
1198         * src/arm/sysv.S: Fix ARM comment marker.
1199
1200 2008-03-26  Jakub Jelinek  <jakub@redhat.com>
1201
1202         * src/alpha/osf.S: Add .note.GNU-stack on Linux.
1203         * src/s390/sysv.S: Likewise.
1204         * src/powerpc/ppc_closure.S: Likewise.
1205         * src/powerpc/sysv.S: Likewise.
1206         * src/x86/unix64.S: Likewise.
1207         * src/x86/sysv.S: Likewise.
1208         * src/sparc/v8.S: Likewise.
1209         * src/sparc/v9.S: Likewise.
1210         * src/m68k/sysv.S: Likewise.
1211         * src/arm/sysv.S: Likewise.
1212
1213 2008-03-16  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1214
1215         * aclocal.m4: Regenerate.
1216         * configure: Likewise.
1217         * Makefile.in: Likewise.
1218         * include/Makefile.in: Likewise.
1219         * testsuite/Makefile.in: Likewise.
1220
1221 2008-02-12  Bjoern Koenig  <bkoenig@alpha-tierchen.de>
1222             Andreas Tobler  <a.tobler@schweiz.org>
1223
1224         * configure.ac: Add amd64-*-freebsd* target.
1225         * configure: Regenerate.
1226
1227 2008-01-30  H.J. Lu  <hongjiu.lu@intel.com>
1228
1229         PR libffi/34612
1230         * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when
1231         returning struct.
1232
1233         * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer"
1234         tests.
1235
1236 2008-01-24  David Edelsohn  <edelsohn@gnu.org>
1237
1238         * configure: Regenerate.
1239
1240 2008-01-06  Andreas Tobler  <a.tobler@schweiz.org>
1241
1242         * src/x86/ffi.c (ffi_prep_cif_machdep): Fix thinko.
1243
1244 2008-01-05  Andreas Tobler  <a.tobler@schweiz.org>
1245
1246         PR testsuite/32843
1247         * src/x86/ffi.c (ffi_prep_cif_machdep): Add code for
1248         signed/unsigned int8/16 for X86_DARWIN.
1249         Updated copyright info.
1250         Handle one and two byte structs with special cif->flags.
1251         * src/x86/ffitarget.h: Add special types for one and two byte structs.
1252         Updated copyright info.
1253         * src/x86/darwin.S (ffi_call_SYSV): Rewrite to use a jump table like
1254         sysv.S
1255         Remove code to pop args from the stack after call.
1256         Special-case signed/unsigned for int8/16, one and two byte structs.
1257         (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1258         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1259         FFI_TYPE_SINT32.
1260         Updated copyright info.
1261
1262 2007-12-08  David Daney  <ddaney@avtrex.com>
1263
1264         * src/mips/n32.S (ffi_call_N32):  Replace dadd with ADDU, dsub with
1265         SUBU, add with ADDU and use smaller code sequences.
1266
1267 2007-12-07  David Daney  <ddaney@avtrex.com>
1268
1269         * src/mips/ffi.c (ffi_prep_cif_machdep): Handle long double return
1270         type.
1271
1272 2007-12-06  David Daney  <ddaney@avtrex.com>
1273
1274         * include/ffi.h.in (FFI_SIZEOF_JAVA_RAW): Define if not already
1275         defined.
1276         (ffi_java_raw): New typedef.
1277         (ffi_java_raw_call, ffi_java_ptrarray_to_raw,
1278         ffi_java_raw_to_ptrarray): Change parameter types from ffi_raw to
1279         ffi_java_raw.
1280         (ffi_java_raw_closure) : Same.
1281         (ffi_prep_java_raw_closure, ffi_prep_java_raw_closure_loc): Change
1282         parameter types.
1283         * src/java_raw_api.c (ffi_java_raw_size):  Replace FFI_SIZEOF_ARG with
1284         FFI_SIZEOF_JAVA_RAW.
1285         (ffi_java_raw_to_ptrarray): Change type of raw to ffi_java_raw.
1286         Replace FFI_SIZEOF_ARG with FFI_SIZEOF_JAVA_RAW. Use
1287         sizeof(ffi_java_raw) for alignment calculations.
1288         (ffi_java_ptrarray_to_raw): Same.
1289         (ffi_java_rvalue_to_raw): Add special handling for FFI_TYPE_POINTER
1290         if FFI_SIZEOF_JAVA_RAW == 4.
1291         (ffi_java_raw_to_rvalue): Same.
1292         (ffi_java_raw_call): Change type of raw to ffi_java_raw.
1293         (ffi_java_translate_args): Same.
1294         (ffi_prep_java_raw_closure_loc, ffi_prep_java_raw_closure): Change
1295         parameter types.
1296         * src/mips/ffitarget.h (FFI_SIZEOF_JAVA_RAW): Define for N32 ABI.
1297
1298 2007-12-06  David Daney  <ddaney@avtrex.com>
1299
1300         * src/mips/n32.S (ffi_closure_N32): Use 64-bit add instruction on
1301         pointer values.
1302
1303 2007-12-01  Andreas Tobler  <a.tobler@schweiz.org>
1304
1305         PR libffi/31937
1306         * src/powerpc/ffitarget.h: Introduce new ABI FFI_LINUX_SOFT_FLOAT.
1307         Add local FFI_TYPE_UINT128 to handle soft-float long-double-128.
1308         * src/powerpc/ffi.c: Distinguish between __NO_FPRS__ and not and
1309         set the NUM_FPR_ARG_REGISTERS according to.
1310         Add support for potential soft-float support under hard-float
1311         architecture.
1312         (ffi_prep_args_SYSV): Set NUM_FPR_ARG_REGISTERS to 0 in case of
1313         FFI_LINUX_SOFT_FLOAT, handle float, doubles and long-doubles according
1314         to the FFI_LINUX_SOFT_FLOAT ABI.
1315         (ffi_prep_cif_machdep): Likewise.
1316         (ffi_closure_helper_SYSV): Likewise.
1317         * src/powerpc/ppc_closure.S: Make sure not to store float/double
1318         on archs where __NO_FPRS__ is true.
1319         Add FFI_TYPE_UINT128 support.
1320         * src/powerpc/sysv.S: Add support for soft-float long-double-128.
1321         Adjust copyright notice.
1322
1323 2007-11-25  Andreas Tobler  <a.tobler@schweiz.org>
1324
1325         * src/closures.c: Move defintion of MAYBE_UNUSED from here to ...
1326         * include/ffi_common.h: ... here.
1327         Update copyright.
1328
1329 2007-11-17  Andreas Tobler  <a.tobler@schweiz.org>
1330
1331         * src/powerpc/sysv.S: Load correct cr to compare if we have long double.
1332         * src/powerpc/linux64.S: Likewise.
1333         * src/powerpc/ffi.c: Add a comment to show which part goes into cr6.
1334         * testsuite/libffi.call/return_ldl.c: New test.
1335
1336 2007-09-04    <aph@redhat.com>
1337
1338         * src/arm/sysv.S (UNWIND): New.
1339         (Whole file): Conditionally compile unwinder directives.
1340         * src/arm/sysv.S: Add unwinder directives.
1341
1342         * src/arm/ffi.c (ffi_prep_args): Align structs by at least 4 bytes.
1343         Only treat r0 as a struct address if we're actually returning a
1344         struct by address.
1345         Only copy the bytes that are actually within a struct.
1346         (ffi_prep_cif_machdep): A Composite Type not larger than 4 bytes
1347         is returned in r0, not passed by address.
1348         (ffi_call): Allocate a word-sized temporary for the case where
1349         a composite is returned in r0.
1350         (ffi_prep_incoming_args_SYSV): Align as necessary.
1351
1352 2007-08-05  Steven Newbury  <s_j_newbury@yahoo.co.uk>
1353
1354         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Use __clear_cache instead of
1355         directly using the sys_cacheflush syscall.
1356
1357 2007-07-27  Andrew Haley  <aph@redhat.com>
1358
1359         * src/arm/sysv.S (ffi_closure_SYSV): Add soft-float.
1360
1361 2007-09-03  Maciej W. Rozycki  <macro@linux-mips.org>
1362
1363         * Makefile.am: Unify MIPS_IRIX and MIPS_LINUX into MIPS.
1364         * configure.ac: Likewise.
1365         * Makefile.in: Regenerate.
1366         * include/Makefile.in: Likewise.
1367         * testsuite/Makefile.in: Likewise.
1368         * configure: Likewise.
1369
1370 2007-08-24  David Daney  <ddaney@avtrex.com>
1371
1372         * testsuite/libffi.call/return_sl.c: New test.
1373
1374 2007-08-10  David Daney  <ddaney@avtrex.com>
1375
1376         * testsuite/libffi.call/cls_multi_ushort.c,
1377         testsuite/libffi.call/cls_align_uint16.c,
1378         testsuite/libffi.call/nested_struct1.c,
1379         testsuite/libffi.call/nested_struct3.c,
1380         testsuite/libffi.call/cls_7_1_byte.c,
1381         testsuite/libffi.call/nested_struct5.c,
1382         testsuite/libffi.call/cls_double.c,
1383         testsuite/libffi.call/nested_struct7.c,
1384         testsuite/libffi.call/cls_sint.c,
1385         testsuite/libffi.call/nested_struct9.c,
1386         testsuite/libffi.call/cls_20byte1.c,
1387         testsuite/libffi.call/cls_multi_sshortchar.c,
1388         testsuite/libffi.call/cls_align_sint64.c,
1389         testsuite/libffi.call/cls_3byte2.c,
1390         testsuite/libffi.call/cls_multi_schar.c,
1391         testsuite/libffi.call/cls_multi_uchar.c,
1392         testsuite/libffi.call/cls_19byte.c,
1393         testsuite/libffi.call/cls_9byte1.c,
1394         testsuite/libffi.call/cls_align_float.c,
1395         testsuite/libffi.call/closure_fn1.c,
1396         testsuite/libffi.call/problem1.c,
1397         testsuite/libffi.call/closure_fn3.c,
1398         testsuite/libffi.call/cls_sshort.c,
1399         testsuite/libffi.call/closure_fn5.c,
1400         testsuite/libffi.call/cls_align_double.c,
1401         testsuite/libffi.call/nested_struct.c,
1402         testsuite/libffi.call/cls_2byte.c,
1403         testsuite/libffi.call/nested_struct10.c,
1404         testsuite/libffi.call/cls_4byte.c,
1405         testsuite/libffi.call/cls_6byte.c,
1406         testsuite/libffi.call/cls_8byte.c,
1407         testsuite/libffi.call/cls_multi_sshort.c,
1408         testsuite/libffi.call/cls_align_sint16.c,
1409         testsuite/libffi.call/cls_align_uint32.c,
1410         testsuite/libffi.call/cls_20byte.c,
1411         testsuite/libffi.call/cls_float.c,
1412         testsuite/libffi.call/nested_struct2.c,
1413         testsuite/libffi.call/cls_5_1_byte.c,
1414         testsuite/libffi.call/nested_struct4.c,
1415         testsuite/libffi.call/cls_24byte.c,
1416         testsuite/libffi.call/nested_struct6.c,
1417         testsuite/libffi.call/cls_64byte.c,
1418         testsuite/libffi.call/nested_struct8.c,
1419         testsuite/libffi.call/cls_uint.c,
1420         testsuite/libffi.call/cls_multi_ushortchar.c,
1421         testsuite/libffi.call/cls_schar.c,
1422         testsuite/libffi.call/cls_uchar.c,
1423         testsuite/libffi.call/cls_align_uint64.c,
1424         testsuite/libffi.call/cls_ulonglong.c,
1425         testsuite/libffi.call/cls_align_longdouble.c,
1426         testsuite/libffi.call/cls_1_1byte.c,
1427         testsuite/libffi.call/cls_12byte.c,
1428         testsuite/libffi.call/cls_3_1byte.c,
1429         testsuite/libffi.call/cls_3byte1.c,
1430         testsuite/libffi.call/cls_4_1byte.c,
1431         testsuite/libffi.call/cls_6_1_byte.c,
1432         testsuite/libffi.call/cls_16byte.c,
1433         testsuite/libffi.call/cls_18byte.c,
1434         testsuite/libffi.call/closure_fn0.c,
1435         testsuite/libffi.call/cls_9byte2.c,
1436         testsuite/libffi.call/closure_fn2.c,
1437         testsuite/libffi.call/closure_fn4.c,
1438         testsuite/libffi.call/cls_ushort.c,
1439         testsuite/libffi.call/closure_fn6.c,
1440         testsuite/libffi.call/cls_5byte.c,
1441         testsuite/libffi.call/cls_align_pointer.c,
1442         testsuite/libffi.call/cls_7byte.c,
1443         testsuite/libffi.call/cls_align_sint32.c,
1444         testsuite/libffi.special/unwindtest_ffi_call.cc,
1445         testsuite/libffi.special/unwindtest.cc: Remove xfail for mips64*-*-*.
1446
1447 2007-08-10  David Daney  <ddaney@avtrex.com>
1448
1449         PR libffi/28313
1450         * configure.ac: Don't treat mips64 as a special case.
1451         * Makefile.am (nodist_libffi_la_SOURCES): Add n32.S.
1452         * configure: Regenerate
1453         * Makefile.in: Ditto.
1454         * fficonfig.h.in: Ditto.
1455         * src/mips/ffitarget.h (REG_L, REG_S, SUBU, ADDU, SRL, LI): Indent.
1456         (LA, EH_FRAME_ALIGN, FDE_ADDR_BYTES): New preprocessor macros.
1457         (FFI_DEFAULT_ABI): Set for n64 case.
1458         (FFI_CLOSURES, FFI_TRAMPOLINE_SIZE): Define for n32 and n64 cases.
1459         * src/mips/n32.S (ffi_call_N32): Add debug macros and labels for FDE.
1460         (ffi_closure_N32): New function.
1461         (.eh_frame): New section
1462         * src/mips/o32.S: Clean up comments.
1463         (ffi_closure_O32): Pass ffi_closure parameter in $12.
1464         * src/mips/ffi.c: Use FFI_MIPS_N32 instead of
1465         _MIPS_SIM == _ABIN32 throughout.
1466         (FFI_MIPS_STOP_HERE): New, use in place of
1467         ffi_stop_here.
1468         (ffi_prep_args): Use unsigned long to hold pointer values.  Rewrite
1469         to support n32/n64 ABIs.
1470         (calc_n32_struct_flags): Rewrite.
1471         (calc_n32_return_struct_flags): Remove unused variable.  Reverse
1472         position of flag bits.
1473         (ffi_prep_cif_machdep): Rewrite n32 portion.
1474         (ffi_call): Enable for n64.  Add special handling for small structure
1475         return values.
1476         (ffi_prep_closure_loc): Add n32 and n64 support.
1477         (ffi_closure_mips_inner_O32): Add cast to silence warning.
1478         (copy_struct_N32, ffi_closure_mips_inner_N32): New functions.
1479
1480 2007-08-08  David Daney  <ddaney@avtrex.com>
1481
1482         * testsuite/libffi.call/ffitest.h (ffi_type_mylong): Remove definition.
1483         * testsuite/libffi.call/cls_align_uint16.c (main): Use correct type
1484         specifiers.
1485         * testsuite/libffi.call/nested_struct1.c (main): Ditto.
1486         * testsuite/libffi.call/cls_sint.c (main): Ditto.
1487         * testsuite/libffi.call/nested_struct9.c (main): Ditto.
1488         * testsuite/libffi.call/cls_20byte1.c (main): Ditto.
1489         * testsuite/libffi.call/cls_9byte1.c (main): Ditto.
1490         * testsuite/libffi.call/closure_fn1.c (main): Ditto.
1491         * testsuite/libffi.call/closure_fn3.c (main): Ditto.
1492         * testsuite/libffi.call/return_dbl2.c (main): Ditto.
1493         * testsuite/libffi.call/cls_sshort.c (main): Ditto.
1494         * testsuite/libffi.call/return_fl3.c (main): Ditto.
1495         * testsuite/libffi.call/closure_fn5.c (main): Ditto.
1496         * testsuite/libffi.call/nested_struct.c (main): Ditto.
1497         * testsuite/libffi.call/nested_struct10.c (main): Ditto.
1498         * testsuite/libffi.call/return_ll1.c (main): Ditto.
1499         * testsuite/libffi.call/cls_8byte.c (main): Ditto.
1500         * testsuite/libffi.call/cls_align_uint32.c (main): Ditto.
1501         * testsuite/libffi.call/cls_align_sint16.c (main): Ditto.
1502         * testsuite/libffi.call/cls_20byte.c (main): Ditto.
1503         * testsuite/libffi.call/nested_struct2.c (main): Ditto.
1504         * testsuite/libffi.call/cls_24byte.c (main): Ditto.
1505         * testsuite/libffi.call/nested_struct6.c (main): Ditto.
1506         * testsuite/libffi.call/cls_uint.c (main): Ditto.
1507         * testsuite/libffi.call/cls_12byte.c (main): Ditto.
1508         * testsuite/libffi.call/cls_16byte.c (main): Ditto.
1509         * testsuite/libffi.call/closure_fn0.c (main): Ditto.
1510         * testsuite/libffi.call/cls_9byte2.c (main): Ditto.
1511         * testsuite/libffi.call/closure_fn2.c (main): Ditto.
1512         * testsuite/libffi.call/return_dbl1.c (main): Ditto.
1513         * testsuite/libffi.call/closure_fn4.c (main): Ditto.
1514         * testsuite/libffi.call/closure_fn6.c (main): Ditto.
1515         * testsuite/libffi.call/cls_align_sint32.c (main): Ditto.
1516
1517 2007-08-07  Andrew Haley  <aph@redhat.com>
1518
1519         * src/x86/sysv.S (ffi_closure_raw_SYSV): Fix typo in previous
1520         checkin.
1521
1522 2007-08-06  Andrew Haley  <aph@redhat.com>
1523
1524         PR testsuite/32843
1525         * src/x86/sysv.S (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1526         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1527         FFI_TYPE_SINT32.
1528
1529 2007-08-02  David Daney  <ddaney@avtrex.com>
1530
1531         * testsuite/libffi.call/return_ul.c (main): Define return type as
1532         ffi_arg.  Use proper printf conversion specifier.
1533
1534 2007-07-30  Andrew Haley  <aph@redhat.com>
1535
1536         PR testsuite/32843
1537         * src/x86/ffi.c (ffi_prep_cif_machdep): in x86 case, add code for
1538         signed/unsigned int8/16.
1539         * src/x86/sysv.S (ffi_call_SYSV): Rewrite to:
1540         Use a jump table.
1541         Remove code to pop args from the stack after call.
1542         Special-case signed/unsigned int8/16.
1543         * testsuite/libffi.call/return_sc.c (main): Revert.
1544
1545 2007-07-26  Richard Guenther  <rguenther@suse.de>
1546
1547         PR testsuite/32843
1548         * testsuite/libffi.call/return_sc.c (main): Verify call
1549         result as signed char, not ffi_arg.
1550
1551 2007-07-16  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1552
1553         * configure.ac (i?86-*-solaris2.1[0-9]): Set TARGET to X86_64.
1554         * configure: Regenerate.
1555
1556 2007-07-11  David Daney  <ddaney@avtrex.com>
1557
1558         * src/mips/ffi.c: Don't include sys/cachectl.h.
1559         (ffi_prep_closure_loc): Use __builtin___clear_cache() instead of
1560         cacheflush().
1561
1562 2007-05-18  Aurelien Jarno  <aurelien@aurel32.net>
1563
1564         * src/arm/ffi.c (ffi_prep_closure_loc): Renamed and ajusted
1565         from (ffi_prep_closure): ... this.
1566         (FFI_INIT_TRAMPOLINE): Adjust.
1567
1568 2005-12-31  Phil Blundell  <pb@reciva.com>
1569
1570         * src/arm/ffi.c (ffi_prep_incoming_args_SYSV,
1571         ffi_closure_SYSV_inner, ffi_prep_closure): New, add closure support.
1572         * src/arm/sysv.S(ffi_closure_SYSV): Likewise.
1573         * src/arm/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1574         (FFI_CLOSURES): Enable closure support.
1575
1576 2007-07-03  Andrew Haley  <aph@hedges.billgatliff.com>
1577
1578         * testsuite/libffi.call/cls_multi_ushort.c,
1579         testsuite/libffi.call/cls_align_uint16.c,
1580         testsuite/libffi.call/nested_struct1.c,
1581         testsuite/libffi.call/nested_struct3.c,
1582         testsuite/libffi.call/cls_7_1_byte.c,
1583         testsuite/libffi.call/cls_double.c,
1584         testsuite/libffi.call/nested_struct5.c,
1585         testsuite/libffi.call/nested_struct7.c,
1586         testsuite/libffi.call/cls_sint.c,
1587         testsuite/libffi.call/nested_struct9.c,
1588         testsuite/libffi.call/cls_20byte1.c,
1589         testsuite/libffi.call/cls_multi_sshortchar.c,
1590         testsuite/libffi.call/cls_align_sint64.c,
1591         testsuite/libffi.call/cls_3byte2.c,
1592         testsuite/libffi.call/cls_multi_schar.c,
1593         testsuite/libffi.call/cls_multi_uchar.c,
1594         testsuite/libffi.call/cls_19byte.c,
1595         testsuite/libffi.call/cls_9byte1.c,
1596         testsuite/libffi.call/cls_align_float.c,
1597         testsuite/libffi.call/closure_fn1.c,
1598         testsuite/libffi.call/problem1.c,
1599         testsuite/libffi.call/closure_fn3.c,
1600         testsuite/libffi.call/cls_sshort.c,
1601         testsuite/libffi.call/closure_fn5.c,
1602         testsuite/libffi.call/cls_align_double.c,
1603         testsuite/libffi.call/cls_2byte.c,
1604         testsuite/libffi.call/nested_struct.c,
1605         testsuite/libffi.call/nested_struct10.c,
1606         testsuite/libffi.call/cls_4byte.c,
1607         testsuite/libffi.call/cls_6byte.c,
1608         testsuite/libffi.call/cls_8byte.c,
1609         testsuite/libffi.call/cls_multi_sshort.c,
1610         testsuite/libffi.call/cls_align_uint32.c,
1611         testsuite/libffi.call/cls_align_sint16.c,
1612         testsuite/libffi.call/cls_float.c,
1613         testsuite/libffi.call/cls_20byte.c,
1614         testsuite/libffi.call/cls_5_1_byte.c,
1615         testsuite/libffi.call/nested_struct2.c,
1616         testsuite/libffi.call/cls_24byte.c,
1617         testsuite/libffi.call/nested_struct4.c,
1618         testsuite/libffi.call/nested_struct6.c,
1619         testsuite/libffi.call/cls_64byte.c,
1620         testsuite/libffi.call/nested_struct8.c,
1621         testsuite/libffi.call/cls_uint.c,
1622         testsuite/libffi.call/cls_multi_ushortchar.c,
1623         testsuite/libffi.call/cls_schar.c,
1624         testsuite/libffi.call/cls_uchar.c,
1625         testsuite/libffi.call/cls_align_uint64.c,
1626         testsuite/libffi.call/cls_ulonglong.c,
1627         testsuite/libffi.call/cls_align_longdouble.c,
1628         testsuite/libffi.call/cls_1_1byte.c,
1629         testsuite/libffi.call/cls_12byte.c,
1630         testsuite/libffi.call/cls_3_1byte.c,
1631         testsuite/libffi.call/cls_3byte1.c,
1632         testsuite/libffi.call/cls_4_1byte.c,
1633         testsuite/libffi.call/cls_6_1_byte.c,
1634         testsuite/libffi.call/cls_16byte.c,
1635         testsuite/libffi.call/cls_18byte.c,
1636         testsuite/libffi.call/closure_fn0.c,
1637         testsuite/libffi.call/cls_9byte2.c,
1638         testsuite/libffi.call/closure_fn2.c,
1639         testsuite/libffi.call/closure_fn4.c,
1640         testsuite/libffi.call/cls_ushort.c,
1641         testsuite/libffi.call/closure_fn6.c,
1642         testsuite/libffi.call/cls_5byte.c,
1643         testsuite/libffi.call/cls_align_pointer.c,
1644         testsuite/libffi.call/cls_7byte.c,
1645         testsuite/libffi.call/cls_align_sint32.c,
1646         testsuite/libffi.special/unwindtest_ffi_call.cc,
1647         testsuite/libffi.special/unwindtest.cc: Enable for ARM.
1648
1649 2007-07-05  H.J. Lu  <hongjiu.lu@intel.com>
1650
1651         * aclocal.m4: Regenerated.
1652
1653 2007-06-02  Paolo Bonzini  <bonzini@gnu.org>
1654
1655         * configure: Regenerate.
1656
1657 2007-05-23  Steve Ellcey  <sje@cup.hp.com>
1658
1659         * Makefile.in: Regenerate.
1660         * configure: Regenerate.
1661         * aclocal.m4: Regenerate.
1662         * include/Makefile.in: Regenerate.
1663         * testsuite/Makefile.in: Regenerate.
1664
1665 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1666
1667         * src/m68k/ffi.c (ffi_prep_incoming_args_SYSV,
1668         ffi_closure_SYSV_inner,ffi_prep_closure): New, add closure support.
1669         * src/m68k/sysv.S(ffi_closure_SYSV,ffi_closure_struct_SYSV): Likewise.
1670         * src/m68k/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1671         (FFI_CLOSURES): Enable closure support.
1672
1673 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1674
1675         * configure.ac (HAVE_AS_CFI_PSEUDO_OP): New test.
1676         * configure: Regenerate.
1677         * fficonfig.h.in: Regenerate.
1678         * src/m68k/sysv.S (CFI_STARTPROC,CFI_ENDPROC,
1679         CFI_OFFSET,CFI_DEF_CFA): New macros.
1680         (ffi_call_SYSV): Add callframe annotation.
1681
1682 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1683
1684         * src/m68k/ffi.c (ffi_prep_args,ffi_prep_cif_machdep): Fix
1685         numerous test suite failures.
1686         * src/m68k/sysv.S (ffi_call_SYSV): Likewise.
1687
1688 2007-04-11  Paolo Bonzini  <bonzini@gnu.org>
1689
1690         * Makefile.am (EXTRA_DIST): Bring up to date.
1691         * Makefile.in: Regenerate.
1692         * src/frv/eabi.S: Remove RCS keyword.
1693
1694 2007-04-06  Richard Henderson  <rth@redhat.com>
1695
1696         * configure.ac: Tidy target case.
1697         (HAVE_LONG_DOUBLE): Allow the target to override.
1698         * configure: Regenerate.
1699         * include/ffi.h.in: Don't define ffi_type_foo if
1700         LIBFFI_HIDE_BASIC_TYPES is defined.
1701         (ffi_type_longdouble): If not HAVE_LONG_DOUBLE, define
1702         to ffi_type_double.
1703         * types.c (LIBFFI_HIDE_BASIC_TYPES): Define.
1704         (FFI_TYPEDEF, ffi_type_void): Mark the data const.
1705         (ffi_type_longdouble): Special case for Alpha.  Don't define
1706         if long double == double.
1707
1708         * src/alpha/ffi.c (FFI_TYPE_LONGDOUBLE): Assert unique value.
1709         (ffi_prep_cif_machdep): Handle it as the 128-bit type.
1710         (ffi_call, ffi_closure_osf_inner): Likewise.
1711         (ffi_closure_osf_inner): Likewise.  Mark hidden.
1712         (ffi_call_osf, ffi_closure_osf): Mark hidden.
1713         * src/alpha/ffitarget.h (FFI_LAST_ABI): Tidy definition.
1714         * src/alpha/osf.S (ffi_call_osf, ffi_closure_osf): Mark hidden.
1715         (load_table): Handle 128-bit long double.
1716
1717         * testsuite/libffi.call/float4.c: Add -mieee for alpha.
1718
1719 2007-04-06  Tom Tromey  <tromey@redhat.com>
1720
1721         PR libffi/31491:
1722         * README: Fixed bug in example.
1723
1724 2007-04-03  Jakub Jelinek  <jakub@redhat.com>
1725
1726         * src/closures.c: Include sys/statfs.h.
1727         (_GNU_SOURCE): Define on Linux.
1728         (FFI_MMAP_EXEC_SELINUX): Define.
1729         (selinux_enabled): New variable.
1730         (selinux_enabled_check): New function.
1731         (is_selinux_enabled): Define.
1732         (dlmmap): Use it.
1733
1734 2007-03-24  Uros Bizjak  <ubizjak@gmail.com>
1735
1736         * testsuite/libffi.call/return_fl2.c (return_fl): Mark as static.
1737         Use 'volatile float sum' to create sum of floats to avoid false
1738         negative due to excess precision on ix86 targets.
1739         (main): Ditto.
1740
1741 2007-03-08  Alexandre Oliva  <aoliva@redhat.com>
1742
1743         * src/powerpc/ffi.c (flush_icache): Fix left-over from previous
1744         patch.
1745         (ffi_prep_closure_loc): Remove unneeded casts.  Add needed ones.
1746
1747 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1748
1749         * include/ffi.h.in (ffi_closure_alloc, ffi_closure_free): New.
1750         (ffi_prep_closure_loc): New.
1751         (ffi_prep_raw_closure_loc): New.
1752         (ffi_prep_java_raw_closure_loc): New.
1753         * src/closures.c: New file.
1754         * src/dlmalloc.c [FFI_MMAP_EXEC_WRIT] (struct malloc_segment):
1755         Replace sflags with exec_offset.
1756         [FFI_MMAP_EXEC_WRIT] (mmap_exec_offset, add_segment_exec_offset,
1757         sub_segment_exec_offset): New macros.
1758         (get_segment_flags, set_segment_flags, check_segment_merge): New
1759         macros.
1760         (is_mmapped_segment, is_extern_segment): Use get_segment_flags.
1761         (add_segment, sys_alloc, create_mspace, create_mspace_with_base,
1762         destroy_mspace): Use new macros.
1763         (sys_alloc): Silence warning.
1764         * Makefile.am (libffi_la_SOURCES): Add src/closures.c.
1765         * Makefile.in: Rebuilt.
1766         * src/prep_cif [FFI_CLOSURES] (ffi_prep_closure): Implement in
1767         terms of ffi_prep_closure_loc.
1768         * src/raw_api.c (ffi_prep_raw_closure_loc): Renamed and adjusted
1769         from...
1770         (ffi_prep_raw_closure): ... this.  Re-implement in terms of the
1771         renamed version.
1772         * src/java_raw_api (ffi_prep_java_raw_closure_loc): Renamed and
1773         adjusted from...
1774         (ffi_prep_java_raw_closure): ... this.  Re-implement in terms of
1775         the renamed version.
1776         * src/alpha/ffi.c (ffi_prep_closure_loc): Renamed from
1777         (ffi_prep_closure): ... this.
1778         * src/pa/ffi.c: Likewise.
1779         * src/cris/ffi.c: Likewise.  Adjust.
1780         * src/frv/ffi.c: Likewise.
1781         * src/ia64/ffi.c: Likewise.
1782         * src/mips/ffi.c: Likewise.
1783         * src/powerpc/ffi_darwin.c: Likewise.
1784         * src/s390/ffi.c: Likewise.
1785         * src/sh/ffi.c: Likewise.
1786         * src/sh64/ffi.c: Likewise.
1787         * src/sparc/ffi.c: Likewise.
1788         * src/x86/ffi64.c: Likewise.
1789         * src/x86/ffi.c: Likewise.
1790         (FFI_INIT_TRAMPOLINE): Adjust.
1791         (ffi_prep_raw_closure_loc): Renamed and adjusted from...
1792         (ffi_prep_raw_closure): ... this.
1793         * src/powerpc/ffi.c (ffi_prep_closure_loc): Renamed from
1794         (ffi_prep_closure): ... this.
1795         (flush_icache): Adjust.
1796
1797 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1798
1799         * src/dlmalloc.c: New file, imported version 2.8.3 of Doug
1800         Lea's malloc.
1801
1802 2007-03-01  Brooks Moses  <brooks.moses@codesourcery.com>
1803
1804         * Makefile.am: Add dummy install-pdf target.
1805         * Makefile.in: Regenerate
1806
1807 2007-02-13  Andreas Krebbel  <krebbel1@de.ibm.com>
1808
1809         * src/s390/ffi.c (ffi_prep_args, ffi_prep_cif_machdep,
1810         ffi_closure_helper_SYSV): Add long double handling.
1811
1812 2007-02-02  Jakub Jelinek  <jakub@redhat.com>
1813
1814         * src/powerpc/linux64.S (ffi_call_LINUX64): Move restore of r2
1815         immediately after bctrl instruction.
1816
1817 2007-01-18  Alexandre Oliva  <aoliva@redhat.com>
1818
1819         * Makefile.am (all-recursive, install-recursive,
1820         mostlyclean-recursive, clean-recursive, distclean-recursive,
1821         maintainer-clean-recursive): Add missing targets.
1822         * Makefile.in: Rebuilt.
1823
1824 2006-12-14  Andreas Tobler  <a.tobler@schweiz.org>
1825
1826         * configure.ac: Add TARGET for x86_64-*-darwin*.
1827         * Makefile.am (nodist_libffi_la_SOURCES): Add rules for 64-bit sources
1828         for X86_DARWIN.
1829         * src/x86/ffitarget.h: Set trampoline size for x86_64-*-darwin*.
1830         * src/x86/darwin64.S: New file for x86_64-*-darwin* support.
1831         * configure: Regenerate.
1832         * Makefile.in: Regenerate.
1833         * include/Makefile.in: Regenerate.
1834         * testsuite/Makefile.in: Regenerate.
1835         * testsuite/libffi.special/unwindtest_ffi_call.cc: New test case for
1836         ffi_call only.
1837
1838 2006-12-13  Andreas Tobler <a.tobler@schweiz.org>
1839
1840         * aclocal.m4: Regenerate with aclocal -I .. as written in the
1841         Makefile.am.
1842
1843 2006-10-31  Geoffrey Keating  <geoffk@apple.com>
1844
1845         * src/powerpc/ffi_darwin.c (darwin_adjust_aggregate_sizes): New.
1846         (ffi_prep_cif_machdep): Call darwin_adjust_aggregate_sizes for
1847         Darwin.
1848         * testsuite/libffi.call/nested_struct4.c: Remove Darwin XFAIL.
1849         * testsuite/libffi.call/nested_struct6.c: Remove Darwin XFAIL.
1850
1851 2006-10-10  Paolo Bonzini  <bonzini@gnu.org>
1852             Sandro Tolaini  <tolaini@libero.it>
1853
1854         * configure.ac [i*86-*-darwin*]: Set X86_DARWIN symbol and
1855         conditional.
1856         * configure: Regenerated.
1857         * Makefile.am (nodist_libffi_la_SOURCES) [X86_DARWIN]: New case.
1858         (EXTRA_DIST): Add src/x86/darwin.S.
1859         * Makefile.in: Regenerated.
1860         * include/Makefile.in: Regenerated.
1861         * testsuite/Makefile.in: Regenerated.
1862
1863         * src/x86/ffi.c (ffi_prep_cif_machdep) [X86_DARWIN]: Treat like
1864         X86_WIN32, and additionally align stack to 16 bytes.
1865         * src/x86/darwin.S: New, based on sysv.S.
1866         * src/prep_cif.c (ffi_prep_cif) [X86_DARWIN]: Align > 8-byte structs.
1867
1868 2006-09-12  David Daney  <ddaney@avtrex.com>
1869
1870         PR libffi/23935
1871         * include/Makefile.am: Install both ffi.h and ffitarget.h in
1872         $(libdir)/gcc/$(target_alias)/$(gcc_version)/include.
1873         * aclocal.m4: Regenerated for automake 1.9.6.
1874         * Makefile.in: Regenerated.
1875         * include/Makefile.in: Regenerated.
1876         * testsuite/Makefile.in: Regenerated.
1877
1878 2006-08-17  Andreas Tobler  <a.tobler@schweiz.ch>
1879
1880         * include/ffi_common.h (struct): Revert accidental commit.
1881
1882 2006-08-15  Andreas Tobler  <a.tobler@schweiz.ch>
1883
1884         * include/ffi_common.h: Remove lint directives.
1885         * include/ffi.h.in: Likewise.
1886
1887 2006-07-25  Torsten Schoenfeld  <kaffeetisch@gmx.de>
1888
1889         * include/ffi.h.in (ffi_type_ulong, ffi_type_slong): Define correctly
1890         for 32-bit architectures.
1891         * testsuite/libffi.call/return_ul.c: New test case.
1892
1893 2006-07-19  David Daney  <ddaney@avtrex.com>
1894
1895         * testsuite/libffi.call/closure_fn6.c: Remove xfail for mips,
1896         xfail remains for mips64.
1897
1898 2006-05-23  Carlos O'Donell  <carlos@codesourcery.com>
1899
1900         * Makefile.am: Add install-html target. Add install-html to .PHONY
1901         * Makefile.in: Regenerate.
1902         * aclocal.m4: Regenerate.
1903         * include/Makefile.in: Regenerate.
1904         * testsuite/Makefile.in: Regenerate.
1905
1906 2006-05-18  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1907
1908         * pa/ffi.c (ffi_prep_args_pa32): Load floating point arguments from
1909         stack slot.
1910
1911 2006-04-22  Andreas Tobler  <a.tobler@schweiz.ch>
1912
1913         * README: Remove notice about 'Crazy Comments'.
1914         * src/debug.c: Remove lint directives. Cleanup white spaces.
1915         * src/java_raw_api.c: Likewise.
1916         * src/prep_cif.c: Likewise.
1917         * src/raw_api.c: Likewise.
1918         * src/ffitest.c: Delete. No longer needed, all test cases migrated
1919         to the testsuite.
1920         * src/arm/ffi.c: Remove lint directives.
1921         * src/m32r/ffi.c: Likewise.
1922         * src/pa/ffi.c: Likewise.
1923         * src/powerpc/ffi.c: Likewise.
1924         * src/powerpc/ffi_darwin.c: Likewise.
1925         * src/sh/ffi.c: Likewise.
1926         * src/sh64/ffi.c: Likewise.
1927         * src/x86/ffi.c: Likewise.
1928         * testsuite/libffi.call/float2.c: Likewise.
1929         * testsuite/libffi.call/promotion.c: Likewise.
1930         * testsuite/libffi.call/struct1.c: Likewise.
1931
1932 2006-04-13  Andreas Tobler  <a.tobler@schweiz.ch>
1933
1934         * src/pa/hpux32.S: Correct unwind offset calculation for
1935         ffi_closure_pa32.
1936         * src/pa/linux.S: Likewise.
1937
1938 2006-04-12  James E Wilson  <wilson@specifix.com>
1939
1940         PR libgcj/26483
1941         * src/ia64/ffi.c (stf_spill, ldf_fill): Rewrite as macros.
1942         (hfa_type_load): Call stf_spill.
1943         (hfa_type_store): Call ldf_fill.
1944         (ffi_call): Adjust calls to above routines.  Add local temps for
1945         macro result.
1946
1947 2006-04-10  Matthias Klose  <doko@debian.org>
1948
1949         * testsuite/lib/libffi-dg.exp (libffi-init): Recognize multilib
1950         directory names containing underscores.
1951
1952 2006-04-07  James E Wilson  <wilson@specifix.com>
1953
1954         * testsuite/libffi.call/float4.c: New testcase.
1955
1956 2006-04-05  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1957             Andreas Tobler  <a.tobler@schweiz.ch>
1958
1959         * Makefile.am: Add PA_HPUX port.
1960         * Makefile.in: Regenerate.
1961         * include/Makefile.in: Likewise.
1962         * testsuite/Makefile.in: Likewise.
1963         * configure.ac: Add PA_HPUX rules.
1964         * configure: Regenerate.
1965         * src/pa/ffitarget.h: Rename linux target to PA_LINUX.
1966         Add PA_HPUX and PA64_HPUX.
1967         Rename FFI_LINUX ABI to FFI_PA32 ABI.
1968         (FFI_TRAMPOLINE_SIZE): Define for 32-bit HP-UX targets.
1969         (FFI_TYPE_SMALL_STRUCT2): Define.
1970         (FFI_TYPE_SMALL_STRUCT4): Likewise.
1971         (FFI_TYPE_SMALL_STRUCT8): Likewise.
1972         (FFI_TYPE_SMALL_STRUCT3): Redefine.
1973         (FFI_TYPE_SMALL_STRUCT5): Likewise.
1974         (FFI_TYPE_SMALL_STRUCT6): Likewise.
1975         (FFI_TYPE_SMALL_STRUCT7): Likewise.
1976         * src/pa/ffi.c (ROUND_DOWN): Delete.
1977         (fldw, fstw, fldd, fstd): Use '__asm__'.
1978         (ffi_struct_type): Add support for FFI_TYPE_SMALL_STRUCT2,
1979         FFI_TYPE_SMALL_STRUCT4 and FFI_TYPE_SMALL_STRUCT8.
1980         (ffi_prep_args_LINUX): Rename to ffi_prep_args_pa32. Update comment.
1981         Simplify incrementing of stack slot variable. Change type of local
1982         'n' to unsigned int.
1983         (ffi_size_stack_LINUX): Rename to ffi_size_stack_pa32. Handle long
1984         double on PA_HPUX.
1985         (ffi_prep_cif_machdep): Likewise.
1986         (ffi_call): Likewise.
1987         (ffi_closure_inner_LINUX): Rename to ffi_closure_inner_pa32. Change
1988         return type to ffi_status. Simplify incrementing of stack slot
1989         variable. Only copy floating point argument registers when PA_LINUX
1990         is true. Reformat debug statement.
1991         Add support for FFI_TYPE_SMALL_STRUCT2, FFI_TYPE_SMALL_STRUCT4 and
1992         FFI_TYPE_SMALL_STRUCT8.
1993         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Add 'extern' to
1994         declaration.
1995         (ffi_prep_closure): Make linux trampoline conditional on PA_LINUX.
1996         Add nops to cache flush.  Add trampoline for PA_HPUX.
1997         * src/pa/hpux32.S: New file.
1998         * src/pa/linux.S (ffi_call_LINUX): Rename to ffi_call_pa32. Rename
1999         ffi_prep_args_LINUX to ffi_prep_args_pa32.
2000         Localize labels. Add support for 2, 4 and 8-byte small structs. Handle
2001         unaligned destinations in 3, 5, 6 and 7-byte small structs. Order
2002         argument type checks so that common argument types appear first.
2003         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Rename
2004         ffi_closure_inner_LINUX to ffi_closure_inner_pa32.
2005
2006 2006-03-24  Alan Modra  <amodra@bigpond.net.au>
2007
2008         * src/powerpc/ffitarget.h (enum ffi_abi): Add FFI_LINUX.  Default
2009         for 32-bit using IBM extended double format.  Fix FFI_LAST_ABI.
2010         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Handle linux variant of
2011         FFI_TYPE_LONGDOUBLE.
2012         (ffi_prep_args64): Assert using IBM extended double.
2013         (ffi_prep_cif_machdep): Don't munge FFI_TYPE_LONGDOUBLE type.
2014         Handle FFI_LINUX FFI_TYPE_LONGDOUBLE return and args.
2015         (ffi_call): Handle FFI_LINUX.
2016         (ffi_closure_helper_SYSV): Non FFI_LINUX long double return needs
2017         gpr3 return pointer as for struct return.  Handle FFI_LINUX
2018         FFI_TYPE_LONGDOUBLE return and args.  Don't increment "nf"
2019         unnecessarily.
2020         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Load both f1 and f2
2021         for FFI_TYPE_LONGDOUBLE.  Move epilogue insns into case table.
2022         Don't use r6 as pointer to results, instead use sp offset.  Don't
2023         make a special call to load lr with case table address, instead
2024         use offset from previous call.
2025         * src/powerpc/sysv.S (ffi_call_SYSV): Save long double return.
2026         * src/powerpc/linux64.S (ffi_call_LINUX64): Simplify long double
2027         return.
2028
2029 2006-03-15  Kaz Kojima  <kkojima@gcc.gnu.org>
2030
2031         * src/sh64/ffi.c (ffi_prep_cif_machdep): Handle float arguments
2032         passed with FP registers correctly.
2033         (ffi_closure_helper_SYSV): Likewise.
2034         * src/sh64/sysv.S: Likewise.
2035
2036 2006-03-01  Andreas Tobler  <a.tobler@schweiz.ch>
2037
2038         * testsuite/libffi.special/unwindtest.cc (closure_test_fn): Mark cif,
2039         args and userdata unused.
2040         (closure_test_fn1): Mark cif and userdata unused.
2041         (main): Remove unused res.
2042
2043 2006-02-28  Andreas Tobler  <a.tobler@schweiz.ch>
2044
2045         * testsuite/libffi.call/call.exp: Adjust FSF address. Add test runs for
2046         -O2, -O3, -Os and the warning flags -W -Wall.
2047         * testsuite/libffi.special/special.exp: Likewise.
2048         * testsuite/libffi.call/ffitest.h: Add an __UNUSED__ macro to mark
2049         unused parameter unused for gcc or else do nothing.
2050         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2051         * testsuite/libffi.call/cls_12byte.c (cls_struct_12byte_gn): Mark cif
2052         and userdata unused.
2053         * testsuite/libffi.call/cls_16byte.c (cls_struct_16byte_gn): Likewise.
2054         * testsuite/libffi.call/cls_18byte.c (cls_struct_18byte_gn): Likewise.
2055         * testsuite/libffi.call/cls_19byte.c (cls_struct_19byte_gn): Likewise.
2056         * testsuite/libffi.call/cls_1_1byte.c (cls_struct_1_1byte_gn): Likewise.
2057         * testsuite/libffi.call/cls_20byte.c (cls_struct_20byte_gn): Likewise.
2058         * testsuite/libffi.call/cls_20byte1.c (cls_struct_20byte_gn): Likewise.
2059         * testsuite/libffi.call/cls_24byte.c (cls_struct_24byte_gn): Likewise.
2060         * testsuite/libffi.call/cls_2byte.c (cls_struct_2byte_gn): Likewise.
2061         * testsuite/libffi.call/cls_3_1byte.c (cls_struct_3_1byte_gn): Likewise.
2062         * testsuite/libffi.call/cls_3byte1.c (cls_struct_3byte_gn): Likewise.
2063         * testsuite/libffi.call/cls_3byte2.c (cls_struct_3byte_gn1): Likewise.
2064         * testsuite/libffi.call/cls_4_1byte.c (cls_struct_4_1byte_gn): Likewise.
2065         * testsuite/libffi.call/cls_4byte.c (cls_struct_4byte_gn): Likewise.
2066         * testsuite/libffi.call/cls_5_1_byte.c (cls_struct_5byte_gn): Likewise.
2067         * testsuite/libffi.call/cls_5byte.c (cls_struct_5byte_gn): Likewise.
2068         * testsuite/libffi.call/cls_64byte.c (cls_struct_64byte_gn): Likewise.
2069         * testsuite/libffi.call/cls_6_1_byte.c (cls_struct_6byte_gn): Likewise.
2070         * testsuite/libffi.call/cls_6byte.c (cls_struct_6byte_gn): Likewise.
2071         * testsuite/libffi.call/cls_7_1_byte.c (cls_struct_7byte_gn): Likewise.
2072         * testsuite/libffi.call/cls_7byte.c (cls_struct_7byte_gn): Likewise.
2073         * testsuite/libffi.call/cls_8byte.c (cls_struct_8byte_gn): Likewise.
2074         * testsuite/libffi.call/cls_9byte1.c (cls_struct_9byte_gn): Likewise.
2075         * testsuite/libffi.call/cls_9byte2.c (cls_struct_9byte_gn): Likewise.
2076         * testsuite/libffi.call/cls_align_double.c (cls_struct_align_gn):
2077         Likewise.
2078         * testsuite/libffi.call/cls_align_float.c (cls_struct_align_gn):
2079         Likewise.
2080         * testsuite/libffi.call/cls_align_longdouble.c (cls_struct_align_gn):
2081         Likewise.
2082         * testsuite/libffi.call/cls_align_pointer.c (cls_struct_align_fn): Cast
2083         void* to avoid compiler warning.
2084         (main): Likewise.
2085         (cls_struct_align_gn): Mark cif and userdata unused.
2086         * testsuite/libffi.call/cls_align_sint16.c (cls_struct_align_gn):
2087         Likewise.
2088         * testsuite/libffi.call/cls_align_sint32.c (cls_struct_align_gn):
2089         Likewise.
2090         * testsuite/libffi.call/cls_align_sint64.c (cls_struct_align_gn):
2091         Likewise.
2092         * testsuite/libffi.call/cls_align_uint16.c (cls_struct_align_gn):
2093         Likewise.
2094         * testsuite/libffi.call/cls_align_uint32.c (cls_struct_align_gn):
2095         Likewise.
2096         * testsuite/libffi.call/cls_double.c (cls_ret_double_fn): Likewise.
2097         * testsuite/libffi.call/cls_float.c (cls_ret_float_fn): Likewise.
2098         * testsuite/libffi.call/cls_multi_schar.c (test_func_gn): Mark cif and
2099         data unused.
2100         (main): Cast res_call to silence gcc.
2101         * testsuite/libffi.call/cls_multi_sshort.c (test_func_gn): Mark cif and
2102         data unused.
2103         (main): Cast res_call to silence gcc.
2104         * testsuite/libffi.call/cls_multi_sshortchar.c (test_func_gn): Mark cif
2105         and data unused.
2106         (main): Cast res_call to silence gcc.
2107         * testsuite/libffi.call/cls_multi_uchar.c (test_func_gn): Mark cif and
2108         data unused.
2109         (main): Cast res_call to silence gcc.
2110         * testsuite/libffi.call/cls_multi_ushort.c (test_func_gn): Mark cif and
2111         data unused.
2112         (main): Cast res_call to silence gcc.
2113         * testsuite/libffi.call/cls_multi_ushortchar.c (test_func_gn): Mark cif
2114         and data unused.
2115         (main): Cast res_call to silence gcc.
2116         * testsuite/libffi.call/cls_schar.c (cls_ret_schar_fn): Mark cif and
2117         userdata unused.
2118         (cls_ret_schar_fn): Cast printf parameter to silence gcc.
2119         * testsuite/libffi.call/cls_sint.c (cls_ret_sint_fn): Mark cif and
2120         userdata unused.
2121         (cls_ret_sint_fn): Cast printf parameter to silence gcc.
2122         * testsuite/libffi.call/cls_sshort.c (cls_ret_sshort_fn): Mark cif and
2123         userdata unused.
2124         (cls_ret_sshort_fn): Cast printf parameter to silence gcc.
2125         * testsuite/libffi.call/cls_uchar.c (cls_ret_uchar_fn):  Mark cif and
2126         userdata unused.
2127         (cls_ret_uchar_fn): Cast printf parameter to silence gcc.
2128         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Mark cif and
2129         userdata unused.
2130         (cls_ret_uint_fn): Cast printf parameter to silence gcc.
2131         * testsuite/libffi.call/cls_ulonglong.c (cls_ret_ulonglong_fn): Mark cif
2132         and userdata unused.
2133         * testsuite/libffi.call/cls_ushort.c (cls_ret_ushort_fn): Mark cif and
2134         userdata unused.
2135         (cls_ret_ushort_fn): Cast printf parameter to silence gcc.
2136         * testsuite/libffi.call/float.c (floating): Remove unused parameter e.
2137         * testsuite/libffi.call/float1.c (main): Remove unused variable i.
2138         Cleanup white spaces.
2139         * testsuite/libffi.call/negint.c (checking): Remove unused variable i.
2140         * testsuite/libffi.call/nested_struct.c (cls_struct_combined_gn): Mark
2141         cif and userdata unused.
2142         * testsuite/libffi.call/nested_struct1.c (cls_struct_combined_gn):
2143         Likewise.
2144         * testsuite/libffi.call/nested_struct10.c (B_gn): Likewise.
2145         * testsuite/libffi.call/nested_struct2.c (B_fn): Adjust printf
2146         formatters to silence gcc.
2147         (B_gn): Mark cif and userdata unused.
2148         * testsuite/libffi.call/nested_struct3.c (B_gn): Mark cif and userdata
2149         unused.
2150         * testsuite/libffi.call/nested_struct4.c: Mention related PR.
2151         (B_gn): Mark cif and userdata unused.
2152         * testsuite/libffi.call/nested_struct5.c (B_gn): Mark cif and userdata
2153         unused.
2154         * testsuite/libffi.call/nested_struct6.c: Mention related PR.
2155         (B_gn): Mark cif and userdata unused.
2156         * testsuite/libffi.call/nested_struct7.c (B_gn): Mark cif and userdata
2157         unused.
2158         * testsuite/libffi.call/nested_struct8.c (B_gn): Likewise.
2159         * testsuite/libffi.call/nested_struct9.c (B_gn): Likewise.
2160         * testsuite/libffi.call/problem1.c (stub): Likewise.
2161         * testsuite/libffi.call/pyobjc-tc.c (main): Cast the result to silence
2162         gcc.
2163         * testsuite/libffi.call/return_fl2.c (return_fl): Add the note mentioned
2164         in the last commit for this test case in the test case itself.
2165         * testsuite/libffi.call/closure_fn0.c (closure_test_fn0): Mark cif as
2166         unused.
2167         * testsuite/libffi.call/closure_fn1.c (closure_test_fn1): Likewise.
2168         * testsuite/libffi.call/closure_fn2.c (closure_test_fn2): Likewise.
2169         * testsuite/libffi.call/closure_fn3.c (closure_test_fn3): Likewise.
2170         * testsuite/libffi.call/closure_fn4.c (closure_test_fn0): Likewise.
2171         * testsuite/libffi.call/closure_fn5.c (closure_test_fn5): Likewise.
2172         * testsuite/libffi.call/closure_fn6.c (closure_test_fn0): Likewise.
2173
2174 2006-02-22  Kaz Kojima  <kkojima@gcc.gnu.org>
2175
2176         * src/sh/sysv.S: Fix register numbers in the FDE for
2177         ffi_closure_SYSV.
2178
2179 2006-02-20  Andreas Tobler  <a.tobler@schweiz.ch>
2180
2181         * testsuite/libffi.call/return_fl2.c (return_fl): Remove static
2182         declaration to avoid a false negative on ix86. See PR323.
2183
2184 2006-02-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2185
2186         * src/sh/ffi.c (ffi_closure_helper_SYSV): Remove unused variable
2187         and cast integer to void * if needed.  Update the pointer to
2188         the FP register saved area correctly.
2189
2190 2006-02-17  Andreas Tobler  <a.tobler@schweiz.ch>
2191
2192         * testsuite/libffi.call/nested_struct6.c: XFAIL this test until PR25630
2193         is fixed.
2194         * testsuite/libffi.call/nested_struct4.c: Likewise.
2195
2196 2006-02-16  Andreas Tobler  <a.tobler@schweiz.ch>
2197
2198         * testsuite/libffi.call/return_dbl.c: New test case.
2199         * testsuite/libffi.call/return_dbl1.c: Likewise.
2200         * testsuite/libffi.call/return_dbl2.c: Likewise.
2201         * testsuite/libffi.call/return_fl.c: Likewise.
2202         * testsuite/libffi.call/return_fl1.c: Likewise.
2203         * testsuite/libffi.call/return_fl2.c: Likewise.
2204         * testsuite/libffi.call/return_fl3.c: Likewise.
2205         * testsuite/libffi.call/closure_fn6.c: Likewise.
2206
2207         * testsuite/libffi.call/nested_struct2.c: Remove ffi_type_mylong
2208         definition.
2209         * testsuite/libffi.call/ffitest.h: Add ffi_type_mylong definition
2210         here to be used by other test cases too.
2211
2212         * testsuite/libffi.call/nested_struct10.c: New test case.
2213         * testsuite/libffi.call/nested_struct9.c: Likewise.
2214         * testsuite/libffi.call/nested_struct8.c: Likewise.
2215         * testsuite/libffi.call/nested_struct7.c: Likewise.
2216         * testsuite/libffi.call/nested_struct6.c: Likewise.
2217         * testsuite/libffi.call/nested_struct5.c: Likewise.
2218         * testsuite/libffi.call/nested_struct4.c: Likewise.
2219
2220 2006-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
2221
2222         * configure.ac: Enable libffi for sparc64-*-freebsd*.
2223         * configure: Rebuilt.
2224
2225 2006-01-18  Jakub Jelinek  <jakub@redhat.com>
2226
2227         * src/powerpc/sysv.S (smst_two_register): Don't call __ashldi3,
2228         instead do the shifting inline.
2229         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't compute %r5
2230         shift count unconditionally.  Simplify load sequences for 1, 2, 3, 4
2231         and 8 byte structs, for the remaining struct sizes don't call
2232         __lshrdi3, instead do the shifting inline.
2233
2234 2005-12-07  Thiemo Seufer  <ths@networkno.de>
2235
2236         * src/mips/ffitarget.h: Remove obsolete sgidefs.h include. Add
2237         missing parentheses.
2238         * src/mips/o32.S (ffi_call_O32): Code formatting. Define
2239         and use A3_OFF, FP_OFF, RA_OFF. Micro-optimizations.
2240         (ffi_closure_O32): Likewise, but with newly defined A3_OFF2,
2241         A2_OFF2, A1_OFF2, A0_OFF2, RA_OFF2, FP_OFF2, S0_OFF2, GP_OFF2,
2242         V1_OFF2, V0_OFF2, FA_1_1_OFF2, FA_1_0_OFF2, FA_0_1_OFF2,
2243         FA_0_0_OFF2.
2244         * src/mips/ffi.c (ffi_prep_args): Code formatting. Fix
2245         endianness bugs.
2246         (ffi_prep_closure): Improve trampoline instruction scheduling.
2247         (ffi_closure_mips_inner_O32): Fix endianness bugs.
2248
2249 2005-12-03  Alan Modra  <amodra@bigpond.net.au>
2250
2251         * src/powerpc/ffi.c: Formatting.
2252         (ffi_prep_args_SYSV): Avoid possible aliasing problems by using unions.
2253         (ffi_prep_args64): Likewise.
2254
2255 2005-09-30  Geoffrey Keating  <geoffk@apple.com>
2256
2257         * testsuite/lib/libffi-dg.exp (libffi_target_compile): For
2258         darwin, use -shared-libgcc not -lgcc_s, and explain why.
2259
2260 2005-09-26  Tom Tromey  <tromey@redhat.com>
2261
2262         * testsuite/libffi.call/float1.c (value_type): New typedef.
2263         (CANARY): New define.
2264         (main): Check for result buffer overflow.
2265         * src/powerpc/linux64.S: Handle linux64 long double returns.
2266         * src/powerpc/ffi.c (FLAG_RETURNS_128BITS): New constant.
2267         (ffi_prep_cif_machdep): Handle linux64 long double returns.
2268
2269 2005-08-25  Alan Modra  <amodra@bigpond.net.au>
2270
2271         PR target/23404
2272         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Correct placement of stack
2273         homed fp args.
2274         (ffi_status ffi_prep_cif_machdep): Correct stack sizing for same.
2275
2276 2005-08-11  Jakub Jelinek  <jakub@redhat.com>
2277
2278         * configure.ac (HAVE_HIDDEN_VISIBILITY_ATTRIBUTE): New test.
2279         (AH_BOTTOM): Add FFI_HIDDEN definition.
2280         * configure: Rebuilt.
2281         * fficonfig.h.in: Rebuilt.
2282         * src/powerpc/ffi.c (hidden): Remove.
2283         (ffi_closure_LINUX64, ffi_prep_args64, ffi_call_LINUX64,
2284         ffi_closure_helper_LINUX64): Use FFI_HIDDEN instead of hidden.
2285         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64,
2286         .ffi_closure_LINUX64): Use FFI_HIDDEN instead of .hidden.
2287         * src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV): Remove,
2288         add FFI_HIDDEN to its prototype.
2289         (ffi_closure_SYSV_inner): New.
2290         * src/x86/sysv.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2291         * src/x86/win32.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2292
2293 2005-08-10  Alfred M. Szmidt  <ams@gnu.org>
2294
2295         PR libffi/21819:
2296         * configure: Rebuilt.
2297         * configure.ac: Handle i*86-*-gnu*.
2298
2299 2005-08-09  Jakub Jelinek  <jakub@redhat.com>
2300
2301         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Use
2302         DW_CFA_offset_extended_sf rather than
2303         DW_CFA_GNU_negative_offset_extended.
2304         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise.
2305
2306 2005-07-22  SUGIOKA Toshinobu  <sugioka@itonet.co.jp>
2307
2308         * src/sh/sysv.S (ffi_call_SYSV): Stop argument popping correctly
2309         on sh3.
2310         (ffi_closure_SYSV): Change the stack layout for sh3 struct argument.
2311         * src/sh/ffi.c (ffi_prep_args): Fix sh3 argument copy, when it is
2312         partially on register.
2313         (ffi_closure_helper_SYSV): Likewise.
2314         (ffi_prep_cif_machdep): Don't set too many cif->flags.
2315
2316 2005-07-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2317
2318         * src/sh/ffi.c (ffi_call): Handle small structures correctly.
2319         Remove empty line.
2320         * src/sh64/ffi.c (simple_type): Remove.
2321         (return_type): Handle small structures correctly.
2322         (ffi_prep_args): Likewise.
2323         (ffi_call): Likewise.
2324         (ffi_closure_helper_SYSV): Likewise.
2325         * src/sh64/sysv.S (ffi_call_SYSV): Handle 1, 2 and 4-byte return.
2326         Emit position independent code if PIC and remove wrong datalabel
2327         prefixes from EH data.
2328
2329 2005-07-19  Andreas Tobler  <a.tobler@schweiz.ch>
2330
2331         * Makefile.am (nodist_libffi_la_SOURCES): Add POWERPC_FREEBSD.
2332         * Makefile.in: Regenerate.
2333         * include/Makefile.in: Likewise.
2334         * testsuite/Makefile.in: Likewise.
2335         * configure.ac: Add POWERPC_FREEBSD rules.
2336         * configure: Regenerate.
2337         * src/powerpc/ffitarget.h: Add POWERPC_FREEBSD rules.
2338         (FFI_SYSV_TYPE_SMALL_STRUCT): Define.
2339         * src/powerpc/ffi.c: Add flags to handle small structure returns
2340         in ffi_call_SYSV.
2341         (ffi_prep_cif_machdep): Handle small structures for SYSV 4 ABI.
2342         Aka FFI_SYSV.
2343         (ffi_closure_helper_SYSV): Likewise.
2344         * src/powerpc/ppc_closure.S: Add return types for small structures.
2345         * src/powerpc/sysv.S: Add bits to handle small structures for
2346         final SYSV 4 ABI.
2347
2348 2005-07-10  Andreas Tobler  <a.tobler@schweiz.ch>
2349
2350         * testsuite/libffi.call/cls_5_1_byte.c: New test file.
2351         * testsuite/libffi.call/cls_6_1_byte.c: Likewise.
2352         * testsuite/libffi.call/cls_7_1_byte.c: Likewise.
2353
2354 2005-07-05  Randolph Chung  <tausq@debian.org>
2355
2356         * src/pa/ffi.c (ffi_struct_type): Rename FFI_TYPE_SMALL_STRUCT1
2357         as FFI_TYPE_SMALL_STRUCT3.  Break out handling for 5-7 byte
2358         structures.  Kill compilation warnings.
2359         (ffi_closure_inner_LINUX): Print return values as hex in debug
2360         message.  Rename FFI_TYPE_SMALL_STRUCT1 as FFI_TYPE_SMALL_STRUCT3.
2361         Properly handle 5-7 byte structure returns.
2362         * src/pa/ffitarget.h (FFI_TYPE_SMALL_STRUCT1)
2363         (FFI_TYPE_SMALL_STRUCT2): Remove.
2364         (FFI_TYPE_SMALL_STRUCT3, FFI_TYPE_SMALL_STRUCT5)
2365         (FFI_TYPE_SMALL_STRUCT6, FFI_TYPE_SMALL_STRUCT7): Define.
2366         * src/pa/linux.S: Mark source file as using PA1.1 assembly.
2367         (checksmst1, checksmst2): Remove.
2368         (checksmst3): Optimize handling of 3-byte struct returns.
2369         (checksmst567): Properly handle 5-7 byte struct returns.
2370
2371 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2372
2373         PR libgcj/21943
2374         * src/mips/n32.S: Enforce PIC code.
2375         * src/mips/o32.S: Likewise.
2376
2377 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2378
2379         * configure.ac: Treat i*86-*-solaris2.10 and up as X86_64.
2380         * configure: Regenerate.
2381
2382 2005-06-01  Alan Modra  <amodra@bigpond.net.au>
2383
2384         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't use JUMPTARGET
2385         to call ffi_closure_helper_SYSV.  Append @local instead.
2386         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise for ffi_prep_args_SYSV.
2387
2388 2005-05-17  Kelley Cook  <kcook@gcc.gnu.org>
2389
2390         * configure.ac: Use AC_C_BIGENDIAN instead of AC_C_BIGENDIAN_CROSS.
2391         Use AC_CHECK_SIZEOF instead of AC_COMPILE_CHECK_SIZEOF.
2392         * Makefile.am (ACLOCAL_AMFLAGS): Remove -I ../config.
2393         * aclocal.m4, configure, fficonfig.h.in, Makefile.in,
2394         include/Makefile.in, testsuite/Makefile.in: Regenerate.
2395
2396 2005-05-09  Mike Stump  <mrs@apple.com>
2397
2398         * configure: Regenerate.
2399
2400 2005-05-08  Richard Henderson  <rth@redhat.com>
2401
2402         PR libffi/21285
2403         * src/alpha/osf.S: Update unwind into to match code.
2404
2405 2005-05-04  Andreas Degert <ad@papyrus-gmbh.de>
2406             Richard Henderson  <rth@redhat.com>
2407
2408         * src/x86/ffi64.c (ffi_prep_cif_machdep): Save sse-used flag in
2409         bit 11 of flags.
2410         (ffi_call): Mask return type field.  Pass ssecount to ffi_call_unix64.
2411         (ffi_prep_closure): Set carry bit if sse-used flag set.
2412         * src/x86/unix64.S (ffi_call_unix64): Add ssecount argument.
2413         Only load sse registers if ssecount non-zero.
2414         (ffi_closure_unix64): Only save sse registers if carry set on entry.
2415
2416 2005-04-29  Ralf Corsepius  <ralf.corsepius@rtems.org>
2417
2418         * configure.ac: Add i*86-*-rtems*, sparc*-*-rtems*,
2419         powerpc-*rtems*, arm*-*-rtems*, sh-*-rtems*.
2420         * configure: Regenerate.
2421
2422 2005-04-20  Hans-Peter Nilsson  <hp@axis.com>
2423
2424         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): In regsub use,
2425         have Tcl8.3-compatible intermediate variable.
2426
2427 2005-04-18  Simon Posnjak <simon.posnjak@siol.net>
2428             Hans-Peter Nilsson  <hp@axis.com>
2429
2430         * Makefile.am: Add CRIS support.
2431         * configure.ac: Likewise.
2432         * Makefile.in, configure, testsuite/Makefile.in,
2433         include/Makefile.in: Regenerate.
2434         * src/cris: New directory.
2435         * src/cris/ffi.c, src/cris/sysv.S, src/cris/ffitarget.h: New files.
2436         * src/prep_cif.c (ffi_prep_cif): Wrap in #ifndef __CRIS__.
2437
2438         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): Replace \n with
2439         \r?\n in output tests.
2440
2441 2005-04-12  Mike Stump  <mrs@apple.com>
2442
2443         * configure: Regenerate.
2444
2445 2005-03-30  Hans Boehm  <Hans.Boehm@hp.com>
2446
2447         * src/ia64/ffitarget.h (ffi_arg): Use long long instead of DI.
2448
2449 2005-03-30  Steve Ellcey  <sje@cup.hp.com>
2450
2451         * src/ia64/ffitarget.h (ffi_arg) ADD DI attribute.
2452         (ffi_sarg) Ditto.
2453         * src/ia64/unix.S (ffi_closure_unix): Extend gp
2454         to 64 bits in ILP32 mode.
2455         Load 64 bits even for short data.
2456
2457 2005-03-23  Mike Stump  <mrs@apple.com>
2458
2459         * src/powerpc/darwin.S: Update for -m64 multilib.
2460         * src/powerpc/darwin_closure.S: Likewise.
2461
2462 2005-03-21  Zack Weinberg  <zack@codesourcery.com>
2463
2464         * configure.ac: Do not invoke TL_AC_GCC_VERSION.
2465         Do not set tool_include_dir.
2466         * aclocal.m4, configure, Makefile.in, testsuite/Makefile.in:
2467         Regenerate.
2468         * include/Makefile.am: Set gcc_version and toollibffidir.
2469         * include/Makefile.in: Regenerate.
2470
2471 2005-02-22  Andrew Haley  <aph@redhat.com>
2472
2473         * src/powerpc/ffi.c (ffi_prep_cif_machdep): Bump alignment to
2474         odd-numbered register pairs for 64-bit integer types.
2475
2476 2005-02-23  Andreas Tobler  <a.tobler@schweiz.ch>
2477
2478         PR libffi/20104
2479         * testsuite/libffi.call/return_ll1.c: New test case.
2480
2481 2005-02-11  Janis Johnson  <janis187@us.ibm.com>
2482
2483         * testsuite/libffi.call/cls_align_longdouble.c: Remove dg-options.
2484         * testsuite/libffi.call/float.c: Ditto.
2485         * testsuite/libffi.call/float2.c: Ditto.
2486         * testsuite/libffi.call/float3.c: Ditto.
2487
2488 2005-02-08  Andreas Tobler  <a.tobler@schweiz.ch>
2489
2490         * src/frv/ffitarget.h: Remove PPC stuff which does not belong to frv.
2491
2492 2005-01-12  Eric Botcazou  <ebotcazou@libertysurf.fr>
2493
2494         * testsuite/libffi.special/special.exp (cxx_options): Add
2495         -shared-libgcc.
2496
2497 2004-12-31  Richard Henderson  <rth@redhat.com>
2498
2499         * src/types.c (FFI_AGGREGATE_TYPEDEF): Remove.
2500         (FFI_TYPEDEF): Rename from FFI_INTEGRAL_TYPEDEF.  Replace size and
2501         offset parameters with a type parameter; deduce size and structure
2502         alignment.  Update all users.
2503
2504 2004-12-31  Richard Henderson  <rth@redhat.com>
2505
2506         * src/types.c (FFI_TYPE_POINTER): Define with sizeof.
2507         (FFI_TYPE_LONGDOUBLE): Fix for ia64.
2508         * src/ia64/ffitarget.h (struct ffi_ia64_trampoline_struct): Move
2509         into ffi_prep_closure.
2510         * src/ia64/ia64_flags.h, src/ia64/ffi.c, src/ia64/unix.S: Rewrite
2511         from scratch.
2512
2513 2004-12-27  Richard Henderson  <rth@redhat.com>
2514
2515         * src/x86/unix64.S: Fix typo in unwind info.
2516
2517 2004-12-25  Richard Henderson  <rth@redhat.com>
2518
2519         * src/x86/ffi64.c (struct register_args): Rename from stackLayout.
2520         (enum x86_64_reg_class): Add X86_64_COMPLEX_X87_CLASS.
2521         (merge_classes): Check for it.
2522         (SSE_CLASS_P): New.
2523         (classify_argument): Pass byte_offset by value; perform all updates
2524         inside struct case.
2525         (examine_argument): Add classes argument; handle
2526         X86_64_COMPLEX_X87_CLASS.
2527         (ffi_prep_args): Merge into ...
2528         (ffi_call): ... here.  Share stack frame with ffi_call_unix64.
2529         (ffi_prep_cif_machdep): Setup cif->flags for proper structure return.
2530         (ffi_fill_return_value): Remove.
2531         (ffi_prep_closure): Remove dead assert.
2532         (ffi_closure_unix64_inner): Rename from ffi_closure_UNIX64_inner.
2533         Rewrite to use struct register_args instead of va_list.  Create
2534         flags for handling structure returns.
2535         * src/x86/unix64.S: Remove dead strings.
2536         (ffi_call_unix64): Rename from ffi_call_UNIX64.  Rewrite to share
2537         stack frame with ffi_call.  Handle structure returns properly.
2538         (float2sse, floatfloat2sse, double2sse): Remove.
2539         (sse2float, sse2double, sse2floatfloat): Remove.
2540         (ffi_closure_unix64): Rename from ffi_closure_UNIX64.  Rewrite
2541         to handle structure returns properly.
2542
2543 2004-12-08  David Edelsohn  <edelsohn@gnu.org>
2544
2545         * Makefile.am (AM_MAKEFLAGS): Remove duplicate LIBCFLAGS and
2546         PICFLAG.
2547         * Makefile.in: Regenerated.
2548
2549 2004-12-02  Richard Sandiford  <rsandifo@redhat.com>
2550
2551         * configure.ac: Use TL_AC_GCC_VERSION to set gcc_version.
2552         * configure, aclocal.m4, Makefile.in: Regenerate.
2553         * include/Makefile.in, testsuite/Makefile.in: Regenerate.
2554
2555 2004-11-29  Kelley Cook  <kcook@gcc.gnu.org>
2556
2557         * configure: Regenerate for libtool change.
2558
2559 2004-11-25  Kelley Cook  <kcook@gcc.gnu.org>
2560
2561         * configure: Regenerate for libtool reversion.
2562
2563 2004-11-24  Kelley Cook  <kcook@gcc.gnu.org>
2564
2565         * configure: Regenerate for libtool change.
2566
2567 2004-11-23  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2568
2569         * testsuite/lib/libffi-dg.exp: Use new procs in target-libpath.exp.
2570
2571 2004-11-23  Richard Sandiford  <rsandifo@redhat.com>
2572
2573         * src/mips/o32.S (ffi_call_O32, ffi_closure_O32): Use jalr instead
2574         of jal.  Use an absolute encoding for the frame information.
2575
2576 2004-11-23  Kelley Cook  <kcook@gcc.gnu.org>
2577
2578         * Makefile.am: Remove no-dependencies.  Add ACLOCAL_AMFLAGS.
2579         * acinclude.m4: Delete logic for sincludes.
2580         * aclocal.m4, Makefile.in, configure: Regenerate.
2581         * include/Makefile: Likewise.
2582         * testsuite/Makefile: Likewise.
2583
2584 2004-11-22  Eric Botcazou  <ebotcazou@libertysurf.fr>
2585
2586         * src/sparc/ffi.c (ffi_prep_closure): Align doubles and 64-bit integers
2587         on a 8-byte boundary.
2588         * src/sparc/v8.S (ffi_closure_v8): Reserve frame space for arguments.
2589
2590 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2591
2592         * src/arm/ffi.c (ffi_prep_cif_machdep): Handle functions that return
2593         long long values.  Round stack allocation to a multiple of 8 bytes
2594         for ATPCS compatibility.
2595         * src/arm/sysv.S (ffi_call_SYSV): Rework to avoid use of APCS register
2596         names.  Handle returning long long types.  Add Thumb and interworking
2597         support.  Improve soft-float code.
2598
2599 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2600
2601         * testsuite/lib/libffi-db.exp (load_gcc_lib): New function.
2602         (libffi_exit): New function.
2603         (libffi_init): Build the testglue wrapper if needed.
2604
2605 2004-10-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2606
2607         PR other/18138
2608         * testsuite/lib/libffi-dg.exp: Accept more than one multilib libgcc.
2609
2610 2004-10-25  Kazuhiro Inaoka <inaoka.kazuhiro@renesas.com>
2611
2612         * src/m32r/libffitarget.h (FFI_CLOSURES): Set to 0.
2613
2614 2004-10-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2615
2616         * src/sh/sysv.S (ffi_call_SYSV): Don't align for double data.
2617         * testsuite/libffi.call/float3.c: New test case.
2618
2619 2004-10-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2620
2621         * src/sh/ffi.c (ffi_prep_closure): Set T bit in trampoline for
2622         the function returning a structure pointed with R2.
2623         * src/sh/sysv.S (ffi_closure_SYSV): Use R2 as the pointer to
2624         the structure return value if T bit set.  Emit position
2625         independent code and EH data if PIC.
2626
2627 2004-10-13  Kazuhiro Inaoka  <inaoka.kazuhiro@renesas.com>
2628
2629         * Makefile.am: Add m32r support.
2630         * configure.ac: Likewise.
2631         * Makefile.in: Regenerate.
2632         * confiugre: Regenerate.
2633         * src/types.c: Add m32r port to FFI_INTERNAL_TYPEDEF
2634         (uint64, sint64, double, longdouble)
2635         * src/m32r: New directory.
2636         * src/m32r/ffi.c: New file.
2637         * src/m32r/sysv.S: Likewise.
2638         * src/m32r/ffitarget.h: Likewise.
2639
2640 2004-10-02  Kaz Kojima  <kkojima@gcc.gnu.org>
2641
2642         * testsuite/libffi.call/negint.c: New test case.
2643
2644 2004-09-14  H.J. Lu  <hongjiu.lu@intel.com>
2645
2646         PR libgcj/17465
2647         * testsuite/lib/libffi-dg.exp: Don't use global ld_library_path.
2648         Set up LD_LIBRARY_PATH, SHLIB_PATH, LD_LIBRARYN32_PATH,
2649         LD_LIBRARY64_PATH, LD_LIBRARY_PATH_32, LD_LIBRARY_PATH_64 and
2650         DYLD_LIBRARY_PATH.
2651
2652 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2653
2654         * testsuite/libffi.call/many_win32.c: Remove whitespaces.
2655         * testsuite/libffi.call/promotion.c: Likewise.
2656         * testsuite/libffi.call/return_ll.c: Remove unused var. Cleanup
2657         whitespaces.
2658         * testsuite/libffi.call/return_sc.c: Likewise.
2659         * testsuite/libffi.call/return_uc.c: Likewise.
2660
2661 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2662
2663         * src/powerpc/darwin.S: Fix comments and identation.
2664         * src/powerpc/darwin_closure.S: Likewise.
2665
2666 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2667
2668         * src/powerpc/ffi_darwin.c: Add flag for longdouble return values.
2669         (ffi_prep_args): Handle longdouble arguments.
2670         (ffi_prep_cif_machdep): Set flags for longdouble. Calculate space for
2671         longdouble.
2672         (ffi_closure_helper_DARWIN): Add closure handling for longdouble.
2673         * src/powerpc/darwin.S (_ffi_call_DARWIN): Add handling of longdouble
2674         values.
2675         * src/powerpc/darwin_closure.S (_ffi_closure_ASM): Likewise.
2676         * src/types.c: Defined longdouble size and alignment for darwin.
2677
2678 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2679
2680         * src/powerpc/aix.S: Remove whitespaces.
2681         * src/powerpc/aix_closure.S: Likewise.
2682         * src/powerpc/asm.h: Likewise.
2683         * src/powerpc/ffi.c: Likewise.
2684         * src/powerpc/ffitarget.h: Likewise.
2685         * src/powerpc/linux64.S: Likewise.
2686         * src/powerpc/linux64_closure.S: Likewise.
2687         * src/powerpc/ppc_closure.S: Likewise.
2688         * src/powerpc/sysv.S: Likewise.
2689
2690 2004-08-30  Anthony Green  <green@redhat.com>
2691
2692         * Makefile.am: Add frv support.
2693         * Makefile.in, testsuite/Makefile.in: Rebuilt.
2694         * configure.ac: Read configure.host.
2695         * configure.in: Read configure.host.
2696         * configure.host: New file.  frv-elf needs libgloss.
2697         * include/ffi.h.in: Force ffi_closure to have a nice big (8)
2698         alignment.  This is needed to frv and shouldn't harm the others.
2699         * include/ffi_common.h (ALIGN_DOWN): New macro.
2700         * src/frv/ffi.c, src/frv/ffitarget.h, src/frv/eabi.S: New files.
2701
2702 2004-08-24  David Daney  <daney@avtrex.com>
2703
2704         * testsuite/libffi.call/closure_fn0.c: Xfail mips64* instead of mips*.
2705         * testsuite/libffi.call/closure_fn1.c: Likewise.
2706         * testsuite/libffi.call/closure_fn2.c  Likewise.
2707         * testsuite/libffi.call/closure_fn3.c: Likewise.
2708         * testsuite/libffi.call/closure_fn4.c: Likewise.
2709         * testsuite/libffi.call/closure_fn5.c: Likewise.
2710         * testsuite/libffi.call/cls_18byte.c: Likewise.
2711         * testsuite/libffi.call/cls_19byte.c: Likewise.
2712         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2713         * testsuite/libffi.call/cls_20byte.c: Likewise.
2714         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2715         * testsuite/libffi.call/cls_24byte.c: Likewise.
2716         * testsuite/libffi.call/cls_2byte.c: Likewise.
2717         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2718         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2719         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2720         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2721         * testsuite/libffi.call/cls_4byte.c: Likewise.
2722         * testsuite/libffi.call/cls_64byte.c: Likewise.
2723         * testsuite/libffi.call/cls_6byte.c: Likewise.
2724         * testsuite/libffi.call/cls_7byte.c: Likewise.
2725         * testsuite/libffi.call/cls_8byte.c: Likewise.
2726         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2727         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2728         * testsuite/libffi.call/cls_align_double.c: Likewise.
2729         * testsuite/libffi.call/cls_align_float.c: Likewise.
2730         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2731         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2732         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2733         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2734         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2735         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2736         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2737         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
2738         * testsuite/libffi.call/cls_double.c: Likewise.
2739         * testsuite/libffi.call/cls_float.c: Likewise.
2740         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2741         * testsuite/libffi.call/cls_multi_sshort.c: Likewise.
2742         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2743         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2744         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2745         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2746         * testsuite/libffi.call/cls_schar.c: Likewise.
2747         * testsuite/libffi.call/cls_sint.c: Likewise.
2748         * testsuite/libffi.call/cls_sshort.c: Likewise.
2749         * testsuite/libffi.call/cls_uchar.c: Likewise.
2750         * testsuite/libffi.call/cls_uint.c: Likewise.
2751         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2752         * testsuite/libffi.call/cls_ushort.c: Likewise.
2753         * testsuite/libffi.call/nested_struct.c: Likewise.
2754         * testsuite/libffi.call/nested_struct1.c: Likewise.
2755         * testsuite/libffi.call/nested_struct2.c: Likewise.
2756         * testsuite/libffi.call/nested_struct3.c: Likewise.
2757         * testsuite/libffi.call/problem1.c: Likewise.
2758         * testsuite/libffi.special/unwindtest.cc: Likewise.
2759         * testsuite/libffi.call/cls_12byte.c: Likewise and set return value
2760         to zero.
2761         * testsuite/libffi.call/cls_16byte.c: Likewise.
2762         * testsuite/libffi.call/cls_5byte.c: Likewise.
2763
2764 2004-08-23  David Daney <daney@avtrex.com>
2765
2766         PR libgcj/13141
2767         * src/mips/ffitarget.h (FFI_O32_SOFT_FLOAT): New ABI.
2768         * src/mips/ffi.c (ffi_prep_args): Fix alignment calculation.
2769         (ffi_prep_cif_machdep): Handle FFI_O32_SOFT_FLOAT floating point
2770         parameters and return types.
2771         (ffi_call): Handle FFI_O32_SOFT_FLOAT ABI.
2772         (ffi_prep_closure): Ditto.
2773         (ffi_closure_mips_inner_O32): Handle FFI_O32_SOFT_FLOAT ABI, fix
2774         alignment calculations.
2775         * src/mips/o32.S (ffi_closure_O32): Don't use floating point
2776         instructions if FFI_O32_SOFT_FLOAT, make stack frame ABI compliant.
2777
2778 2004-08-14  Casey Marshall <csm@gnu.org>
2779
2780         * src/mips/ffi.c (ffi_pref_cif_machdep): set `cif->flags' to
2781         contain `FFI_TYPE_UINT64' as return type for any 64-bit
2782         integer (O32 ABI only).
2783         (ffi_prep_closure): new function.
2784         (ffi_closure_mips_inner_O32): new function.
2785         * src/mips/ffitarget.h: Define `FFI_CLOSURES' and
2786         `FFI_TRAMPOLINE_SIZE' appropriately if the ABI is o32.
2787         * src/mips/o32.S (ffi_call_O32): add labels for .eh_frame. Return
2788         64 bit integers correctly.
2789         (ffi_closure_O32): new function.
2790         Added DWARF-2 unwind info for both functions.
2791
2792 2004-08-10  Andrew Haley  <aph@redhat.com>
2793
2794         * src/x86/ffi64.c (ffi_prep_args ): 8-align all stack arguments.
2795
2796 2004-08-01  Robert Millan  <robertmh@gnu.org>
2797
2798         * configure.ac: Detect knetbsd-gnu and kfreebsd-gnu.
2799         * configure: Regenerate.
2800
2801 2004-07-30  Maciej W. Rozycki  <macro@linux-mips.org>
2802
2803         * acinclude.m4 (AC_FUNC_MMAP_BLACKLIST): Check for <sys/mman.h>
2804         and mmap() explicitly instead of relying on preset autoconf cache
2805         variables.
2806         * aclocal.m4: Regenerate.
2807         * configure: Regenerate.
2808
2809 2004-07-11  Ulrich Weigand  <uweigand@de.ibm.com>
2810
2811         * src/s390/ffi.c (ffi_prep_args): Fix C aliasing violation.
2812         (ffi_check_float_struct): Remove unused prototype.
2813
2814 2004-06-30  Geoffrey Keating  <geoffk@apple.com>
2815
2816         * src/powerpc/ffi_darwin.c (flush_icache): ';' is a comment
2817         character on Darwin, use '\n\t' instead.
2818
2819 2004-06-26  Matthias Klose  <doko@debian.org>
2820
2821         * libtool-version: Fix typo in revision/age.
2822
2823 2004-06-17  Matthias Klose  <doko@debian.org>
2824
2825         * libtool-version: New.
2826         * Makefile.am (libffi_la_LDFLAGS): Use -version-info for soname.
2827         * Makefile.in: Regenerate.
2828
2829 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2830
2831         * Makefile.am: Remove useless multilib rules.
2832         * Makefile.in: Regenerate.
2833         * aclocal.m4: Regenerate with automake 1.8.5.
2834         * configure.ac: Remove useless multilib configury.
2835         * configure: Regenerate.
2836
2837 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2838
2839         * .cvsignore: New file.
2840
2841 2004-06-10  Jakub Jelinek  <jakub@redhat.com>
2842
2843         * src/ia64/unix.S (ffi_call_unix): Insert group barrier break
2844         fp_done.
2845         (ffi_closure_UNIX): Fix f14/f15 adjustment if FLOAT_SZ is ever
2846         changed from 8.
2847
2848 2004-06-06  Sean McNeil  <sean@mcneil.com>
2849
2850         * configure.ac: Add x86_64-*-freebsd* support.
2851         * configure: Regenerate.
2852
2853 2004-04-26  Joe Buck <jbuck@welsh-buck.org>
2854
2855         Bug 15093
2856         * configure.ac: Test for existence of mmap and sys/mman.h before
2857         checking blacklist.  Fix suggested by Jim Wilson.
2858         * configure: Regenerate.
2859
2860 2004-04-26  Matt Austern  <austern@apple.com>
2861
2862         * src/powerpc/darwin.S: Go through a non-lazy pointer for initial
2863         FDE location.
2864         * src/powerpc/darwin_closure.S: Likewise.
2865
2866 2004-04-24  Andreas Tobler  <a.tobler@schweiz.ch>
2867
2868         * testsuite/libffi.call/cls_multi_schar.c (main): Fix initialization
2869         error. Reported by Thomas Heller <theller@python.net>.
2870         * testsuite/libffi.call/cls_multi_sshort.c (main): Likewise.
2871         * testsuite/libffi.call/cls_multi_ushort.c (main): Likewise.
2872
2873 2004-03-20  Matthias Klose  <doko@debian.org>
2874
2875         * src/pa/linux.S: Fix typo.
2876
2877 2004-03-19  Matthias Klose  <doko@debian.org>
2878
2879         * Makefile.am: Update.
2880         * Makefile.in: Regenerate.
2881         * src/pa/ffi.h.in: Remove.
2882         * src/pa/ffitarget.h: New file.
2883
2884 2004-02-10  Randolph Chung  <tausq@debian.org>
2885
2886         * Makefile.am: Add PA support.
2887         * Makefile.in: Regenerate.
2888         * include/Makefile.in: Regenerate.
2889         * configure.ac: Add PA target.
2890         * configure: Regenerate.
2891         * src/pa/ffi.c: New file.
2892         * src/pa/ffi.h.in: Add PA support.
2893         * src/pa/linux.S: New file.
2894         * prep_cif.c: Add PA support.
2895
2896 2004-03-16  Hosaka Yuji  <hos@tamanegi.org>
2897
2898         * src/types.c: Fix alignment size of X86_WIN32 case int64 and
2899         double.
2900         * src/x86/ffi.c (ffi_prep_args): Replace ecif->cif->rtype->type
2901         with ecif->cif->flags.
2902         (ffi_call, ffi_prep_incoming_args_SYSV): Replace cif->rtype->type
2903         with cif->flags.
2904         (ffi_prep_cif_machdep): Add X86_WIN32 struct case.
2905         (ffi_closure_SYSV): Add 1 or 2-bytes struct case for X86_WIN32.
2906         * src/x86/win32.S (retstruct1b, retstruct2b, sc_retstruct1b,
2907         sc_retstruct2b): Add for 1 or 2-bytes struct case.
2908
2909 2004-03-15 Kelley Cook <kcook@gcc.gnu.org>
2910
2911         * configure.in: Rename file to ...
2912         * configure.ac: ... this.
2913         * fficonfig.h.in: Regenerate.
2914         * Makefile.in: Regenerate.
2915         * include/Makefile.in: Regenerate.
2916         * testsuite/Makefile.in: Regenerate.
2917
2918 2004-03-12  Matt Austern  <austern@apple.com>
2919
2920         * src/powerpc/darwin.S: Fix EH information so it corresponds to
2921         changes in EH format resulting from addition of linkonce support.
2922         * src/powerpc/darwin_closure.S: Likewise.
2923
2924 2004-03-11  Andreas Tobler  <a.tobler@schweiz.ch>
2925             Paolo Bonzini  <bonzini@gnu.org>
2926
2927         * Makefile.am (AUTOMAKE_OPTIONS): Set them.
2928         Remove VPATH. Remove rules for object files. Remove multilib support.
2929         (AM_CCASFLAGS): Add.
2930         * configure.in (AC_CONFIG_HEADERS): Relace AM_CONFIG_HEADER.
2931         (AC_PREREQ): Bump version to 2.59.
2932         (AC_INIT): Fill with version info and bug address.
2933         (ORIGINAL_LD_FOR_MULTILIBS): Remove.
2934         (AM_ENABLE_MULTILIB): Use this instead of AC_ARG_ENABLE.
2935         De-precious CC so that the right flags are passed down to multilibs.
2936         (AC_MSG_ERROR): Replace obsolete macro AC_ERROR.
2937         (AC_CONFIG_FILES): Replace obsolete macro AC_LINK_FILES.
2938         (AC_OUTPUT): Reorganize the output with AC_CONFIG_COMMANDS.
2939         * configure: Rebuilt.
2940         * aclocal.m4: Likewise.
2941         * Makefile.in, include/Makefile.in, testsuite/Makefile.in: Likewise.
2942         * fficonfig.h.in: Likewise.
2943
2944 2004-03-11  Andreas Schwab  <schwab@suse.de>
2945
2946         * src/ia64/ffi.c (ffi_prep_incoming_args_UNIX): Get floating point
2947         arguments from fp registers only for the first 8 parameter slots.
2948         Don't convert a float parameter when passed in memory.
2949
2950 2004-03-09  Hans-Peter Nilsson  <hp@axis.com>
2951
2952         * configure: Regenerate for config/accross.m4 correction.
2953
2954 2004-02-25  Matt Kraai  <kraai@alumni.cmu.edu>
2955
2956         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Change
2957         ecif->cif->bytes to bytes.
2958         (ffi_prep_cif_machdep): Add braces around nested if statement.
2959
2960 2004-02-09  Alan Modra  <amodra@bigpond.net.au>
2961
2962         * src/types.c (pointer): POWERPC64 has 8 byte pointers.
2963
2964         * src/powerpc/ffi.c (ffi_prep_args64): Correct long double handling.
2965         (ffi_closure_helper_LINUX64): Fix typo.
2966         * testsuite/libffi.call/cls_align_longdouble.c: Pass -mlong-double-128
2967         for powerpc64-*-*.
2968         * testsuite/libffi.call/float.c: Likewise.
2969         * testsuite/libffi.call/float2.c: Likewise.
2970
2971 2004-02-08  Alan Modra  <amodra@bigpond.net.au>
2972
2973         * src/powerpc/ffi.c (ffi_prep_cif_machdep <FFI_LINUX64>): Correct
2974         long double function return and long double arg handling.
2975         (ffi_closure_helper_LINUX64): Formatting.  Delete unused "ng" var.
2976         Use "end_pfr" instead of "nf".  Correct long double handling.
2977         Localise "temp".
2978         * src/powerpc/linux64.S (ffi_call_LINUX64): Save f2 long double
2979         return value.
2980         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Allocate
2981         space for long double return value.  Adjust stack frame and offsets.
2982         Load f2 long double return.
2983
2984 2004-02-07  Alan Modra  <amodra@bigpond.net.au>
2985
2986         * src/types.c: Use 16 byte long double for POWERPC64.
2987
2988 2004-01-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2989
2990         * src/sparc/ffi.c (ffi_prep_args_v9): Shift the parameter array
2991         when the structure return address is passed in %o0.
2992         (ffi_V9_return_struct): Rename into ffi_v9_layout_struct.
2993         (ffi_v9_layout_struct): Align the field following a nested structure
2994         on a word boundary.  Use memmove instead of memcpy.
2995         (ffi_call): Update call to ffi_V9_return_struct.
2996         (ffi_prep_closure): Define 'ctx' only for V8.
2997         (ffi_closure_sparc_inner): Clone into ffi_closure_sparc_inner_v8
2998         and ffi_closure_sparc_inner_v9.
2999         (ffi_closure_sparc_inner_v8): Return long doubles by reference.
3000         Always skip the structure return address.  For structures and long
3001         doubles, copy the argument directly.
3002         (ffi_closure_sparc_inner_v9): Skip the structure return address only
3003         if required.  Shift the maximum floating-point slot accordingly.  For
3004         big structures, copy the argument directly; otherwise, left-justify the
3005         argument and call ffi_v9_layout_struct to lay out the structure on
3006         the stack.
3007         * src/sparc/v8.S: Undef STACKFRAME before defining it.
3008         (ffi_closure_v8): Pass the structure return address.  Update call to
3009         ffi_closure_sparc_inner_v8.  Short-circuit FFI_TYPE_INT handling.
3010         Skip the 'unimp' insn when returning long doubles and structures.
3011         * src/sparc/v9.S: Undef STACKFRAME before defining it.
3012         (ffi_closure_v9): Increase the frame size by 2 words.  Short-circuit
3013         FFI_TYPE_INT handling.  Load structures both in integers and
3014         floating-point registers on return.
3015         * README: Update status of the SPARC port.
3016
3017 2004-01-24  Andreas Tobler  <a.tobler@schweiz.ch>
3018
3019         * testsuite/libffi.call/pyobjc-tc.c (main): Treat result value
3020         as of type ffi_arg.
3021         * testsuite/libffi.call/struct3.c (main): Fix CHECK.
3022
3023 2004-01-22  Ulrich Weigand  <uweigand@de.ibm.com>
3024
3025         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Treat result
3026         value as of type ffi_arg, not unsigned int.
3027
3028 2004-01-21  Michael Ritzert  <ritzert@t-online.de>
3029
3030         * ffi64.c (ffi_prep_args): Cast the RHS of an assignment instead
3031         of the LHS.
3032
3033 2004-01-12  Andreas Tobler  <a.tobler@schweiz.ch>
3034
3035         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_32 for
3036         Solaris.
3037
3038 2004-01-08  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3039
3040         * testsuite/libffi.call/ffitest.h (allocate_mmap): Cast MAP_FAILED
3041         to void *.
3042
3043 2003-12-10  Richard Henderson  <rth@redhat.com>
3044
3045         * testsuite/libffi.call/cls_align_pointer.c: Cast pointers to
3046         size_t instead of int.
3047
3048 2003-12-04  Hosaka Yuji  <hos@tamanegi.org>
3049
3050         * testsuite/libffi.call/many_win32.c: Include <float.h>.
3051         * testsuite/libffi.call/many_win32.c (main): Replace variable
3052         int i with unsigned long ul.
3053
3054         * testsuite/libffi.call/cls_align_uint64.c: New test case.
3055         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
3056         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
3057         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
3058         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
3059         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
3060         * testsuite/libffi.call/cls_align_float.c: Likewise.
3061         * testsuite/libffi.call/cls_align_double.c: Likewise.
3062         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
3063         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
3064
3065 2003-12-02  Hosaka Yuji  <hos@tamanegi.org>
3066
3067         PR other/13221
3068         * src/x86/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV):
3069         Align arguments to 32 bits.
3070
3071 2003-12-01  Andreas Tobler  <a.tobler@schweiz.ch>
3072
3073         PR other/13221
3074         * testsuite/libffi.call/cls_multi_sshort.c: New test case.
3075         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
3076         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
3077         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
3078         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
3079         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
3080
3081         * testsuite/libffi.special/unwindtest.cc: Cosmetics.
3082
3083 2003-11-26  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
3084
3085         * testsuite/libffi.call/ffitest.h: Include <fcntl.h>.
3086         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3087
3088 2003-11-22  Andreas Tobler  <a.tobler@schweiz.ch>
3089
3090         * Makefile.in: Rebuilt.
3091         * configure: Likewise.
3092         * testsuite/libffi.special/unwindtest.cc: Convert the mmap to
3093         the right type.
3094
3095 2003-11-21  Andreas Jaeger  <aj@suse.de>
3096             Andreas Tobler  <a.tobler@schweiz.ch>
3097
3098         * acinclude.m4: Add AC_FUNC_MMAP_BLACKLIST.
3099         * configure.in: Call AC_FUNC_MMAP_BLACKLIST.
3100         * Makefile.in: Rebuilt.
3101         * aclocal.m4: Likewise.
3102         * configure: Likewise.
3103         * fficonfig.h.in: Likewise.
3104         * testsuite/lib/libffi-dg.exp: Add include dir.
3105         * testsuite/libffi.call/ffitest.h: Add MMAP definitions.
3106         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3107         * testsuite/libffi.call/closure_fn0.c: Use MMAP functionality
3108         for ffi_closure if available.
3109         * testsuite/libffi.call/closure_fn1.c: Likewise.
3110         * testsuite/libffi.call/closure_fn2.c: Likewise.
3111         * testsuite/libffi.call/closure_fn3.c: Likewise.
3112         * testsuite/libffi.call/closure_fn4.c: Likewise.
3113         * testsuite/libffi.call/closure_fn5.c: Likewise.
3114         * testsuite/libffi.call/cls_12byte.c: Likewise.
3115         * testsuite/libffi.call/cls_16byte.c: Likewise.
3116         * testsuite/libffi.call/cls_18byte.c: Likewise.
3117         * testsuite/libffi.call/cls_19byte.c: Likewise.
3118         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3119         * testsuite/libffi.call/cls_20byte.c: Likewise.
3120         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3121         * testsuite/libffi.call/cls_24byte.c: Likewise.
3122         * testsuite/libffi.call/cls_2byte.c: Likewise.
3123         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3124         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3125         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3126         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3127         * testsuite/libffi.call/cls_4byte.c: Likewise.
3128         * testsuite/libffi.call/cls_5byte.c: Likewise.
3129         * testsuite/libffi.call/cls_64byte.c: Likewise.
3130         * testsuite/libffi.call/cls_6byte.c: Likewise.
3131         * testsuite/libffi.call/cls_7byte.c: Likewise.
3132         * testsuite/libffi.call/cls_8byte.c: Likewise.
3133         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3134         * testsuite/libffi.call/cls_9byte2.c: Likewise.
3135         * testsuite/libffi.call/cls_double.c: Likewise.
3136         * testsuite/libffi.call/cls_float.c: Likewise.
3137         * testsuite/libffi.call/cls_schar.c: Likewise.
3138         * testsuite/libffi.call/cls_sint.c: Likewise.
3139         * testsuite/libffi.call/cls_sshort.c: Likewise.
3140         * testsuite/libffi.call/cls_uchar.c: Likewise.
3141         * testsuite/libffi.call/cls_uint.c: Likewise.
3142         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3143         * testsuite/libffi.call/cls_ushort.c: Likewise.
3144         * testsuite/libffi.call/nested_struct.c: Likewise.
3145         * testsuite/libffi.call/nested_struct1.c: Likewise.
3146         * testsuite/libffi.call/nested_struct2.c: Likewise.
3147         * testsuite/libffi.call/nested_struct3.c: Likewise.
3148         * testsuite/libffi.call/problem1.c: Likewise.
3149         * testsuite/libffi.special/unwindtest.cc: Likewise.
3150
3151 2003-11-20  Andreas Tobler  <a.tobler@schweiz.ch>
3152
3153         * testsuite/lib/libffi-dg.exp: Make the -lgcc_s conditional.
3154
3155 2003-11-19  Andreas Tobler  <a.tobler@schweiz.ch>
3156
3157         * testsuite/lib/libffi-dg.exp: Add DYLD_LIBRARY_PATH for darwin.
3158         Add -lgcc_s to additional flags.
3159
3160 2003-11-12  Andreas Tobler  <a.tobler@schweiz.ch>
3161
3162         * configure.in, include/Makefile.am: PR libgcj/11147, install
3163         the ffitarget.h header file in a gcc versioned and target
3164         dependent place.
3165         * configure: Regenerated.
3166         * Makefile.in, include/Makefile.in: Likewise.
3167         * testsuite/Makefile.in: Likewise.
3168
3169 2003-11-09  Andreas Tobler  <a.tobler@schweiz.ch>
3170
3171         * testsuite/libffi.call/closure_fn0.c: Print result and check
3172         with dg-output to make debugging easier.
3173         * testsuite/libffi.call/closure_fn1.c: Likewise.
3174         * testsuite/libffi.call/closure_fn2.c: Likewise.
3175         * testsuite/libffi.call/closure_fn3.c: Likewise.
3176         * testsuite/libffi.call/closure_fn4.c: Likewise.
3177         * testsuite/libffi.call/closure_fn5.c: Likewise.
3178         * testsuite/libffi.call/cls_12byte.c: Likewise.
3179         * testsuite/libffi.call/cls_16byte.c: Likewise.
3180         * testsuite/libffi.call/cls_18byte.c: Likewise.
3181         * testsuite/libffi.call/cls_19byte.c: Likewise.
3182         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3183         * testsuite/libffi.call/cls_20byte.c: Likewise.
3184         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3185         * testsuite/libffi.call/cls_24byte.c: Likewise.
3186         * testsuite/libffi.call/cls_2byte.c: Likewise.
3187         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3188         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3189         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3190         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3191         * testsuite/libffi.call/cls_4byte.c: Likewise.
3192         * testsuite/libffi.call/cls_5byte.c: Likewise.
3193         * testsuite/libffi.call/cls_64byte.c: Likewise.
3194         * testsuite/libffi.call/cls_6byte.c: Likewise.
3195         * testsuite/libffi.call/cls_7byte.c: Likewise.
3196         * testsuite/libffi.call/cls_8byte.c: Likewise.
3197         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3198         * testsuite/libffi.call/cls_9byte2.c: Likewise.
3199         * testsuite/libffi.call/cls_double.c: Likewise.
3200         * testsuite/libffi.call/cls_float.c: Likewise.
3201         * testsuite/libffi.call/cls_schar.c: Likewise.
3202         * testsuite/libffi.call/cls_sint.c: Likewise.
3203         * testsuite/libffi.call/cls_sshort.c: Likewise.
3204         * testsuite/libffi.call/cls_uchar.c: Likewise.
3205         * testsuite/libffi.call/cls_uint.c: Likewise.
3206         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3207         * testsuite/libffi.call/cls_ushort.c: Likewise.
3208         * testsuite/libffi.call/problem1.c: Likewise.
3209
3210         * testsuite/libffi.special/unwindtest.cc: Make ffi_closure
3211         static.
3212
3213 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
3214
3215         * testsuite/libffi.call/cls_9byte2.c: New test case.
3216         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3217         * testsuite/libffi.call/cls_64byte.c: Likewise.
3218         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3219         * testsuite/libffi.call/cls_19byte.c: Likewise.
3220         * testsuite/libffi.call/cls_18byte.c: Likewise.
3221         * testsuite/libffi.call/closure_fn4.c: Likewise.
3222         * testsuite/libffi.call/closure_fn5.c: Likewise.
3223         * testsuite/libffi.call/cls_schar.c: Likewise.
3224         * testsuite/libffi.call/cls_sint.c: Likewise.
3225         * testsuite/libffi.call/cls_sshort.c: Likewise.
3226         * testsuite/libffi.call/nested_struct2.c: Likewise.
3227         * testsuite/libffi.call/nested_struct3.c: Likewise.
3228
3229 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
3230
3231         * testsuite/libffi.call/cls_double.c: Do a check on the result.
3232         * testsuite/libffi.call/cls_uchar.c: Likewise.
3233         * testsuite/libffi.call/cls_uint.c: Likewise.
3234         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3235         * testsuite/libffi.call/cls_ushort.c: Likewise.
3236         * testsuite/libffi.call/return_sc.c: Cleanup whitespaces.
3237
3238 2003-11-06  Andreas Tobler  <a.tobler@schweiz.ch>
3239
3240         * src/prep_cif.c (ffi_prep_cif): Move the validity check after
3241         the initialization.
3242
3243 2003-10-23  Andreas Tobler  <a.tobler@schweiz.ch>
3244
3245         * src/java_raw_api.c (ffi_java_ptrarray_to_raw): Replace
3246         FFI_ASSERT(FALSE) with FFI_ASSERT(0).
3247
3248 2003-10-22  David Daney  <ddaney@avtrex.com>
3249
3250         * src/mips/ffitarget.h: Replace undefined UINT32 and friends with
3251         __attribute__((__mode__(__SI__))) and friends.
3252
3253 2003-10-22  Andreas Schwab  <schwab@suse.de>
3254
3255         * src/ia64/ffi.c: Replace FALSE/TRUE with false/true.
3256
3257 2003-10-21  Andreas Tobler  <a.tobler@schweiz.ch>
3258
3259         * configure.in: AC_LINK_FILES(ffitarget.h).
3260         * configure: Regenerate.
3261         * Makefile.in: Likewise.
3262         * include/Makefile.in: Likewise.
3263         * testsuite/Makefile.in: Likewise.
3264         * fficonfig.h.in: Likewise.
3265
3266 2003-10-21  Paolo Bonzini  <bonzini@gnu.org>
3267             Richard Henderson  <rth@redhat.com>
3268
3269         Avoid that ffi.h includes fficonfig.h.
3270
3271         * Makefile.am (EXTRA_DIST): Include ffitarget.h files
3272         (TARGET_SRC_MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3273         (TARGET_SRC_MIPS_SGI): Removed.
3274         (MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3275         (MIPS_SGI): Removed.
3276         (CLEANFILES): Removed.
3277         (mostlyclean-am, clean-am, mostlyclean-sub, clean-sub): New
3278         targets.
3279         * acconfig.h: Removed.
3280         * configure.in: Compute sizeofs only for double and long double.
3281         Use them to define and subst HAVE_LONG_DOUBLE.  Include comments
3282         into AC_DEFINE instead of using acconfig.h.  Create
3283         include/ffitarget.h instead of include/fficonfig.h.  Rename
3284         MIPS_GCC to MIPS_IRIX, drop MIPS_SGI since we are in gcc's tree.
3285         AC_DEFINE EH_FRAME_FLAGS.
3286         * include/Makefile.am (DISTCLEANFILES): New automake macro.
3287         (hack_DATA): Add ffitarget.h.
3288         * include/ffi.h.in: Remove all system specific definitions.
3289         Declare raw API even if it is not installed, why bother?
3290         Use limits.h instead of SIZEOF_* to define ffi_type_*.  Do
3291         not define EH_FRAME_FLAGS, it is in fficonfig.h now.  Include
3292         ffitarget.h instead of fficonfig.h.  Remove ALIGN macro.
3293         (UINT_ARG, INT_ARG): Removed, use ffi_arg and ffi_sarg instead.
3294         * include/ffi_common.h (bool): Do not define.
3295         (ffi_assert): Accept failed assertion.
3296         (ffi_type_test): Return void and accept file/line.
3297         (FFI_ASSERT): Pass stringized failed assertion.
3298         (FFI_ASSERT_AT): New macro.
3299         (FFI_ASSERT_VALID_TYPE): New macro.
3300         (UINT8, SINT8, UINT16, SINT16, UINT32, SINT32,
3301         UINT64, SINT64): Define here with gcc's __attribute__ macro
3302         instead of in ffi.h
3303         (FLOAT32, ALIGN): Define here instead of in ffi.h
3304         * include/ffi-mips.h: Removed.  Its content moved to
3305         src/mips/ffitarget.h after separating assembly and C sections.
3306         * src/alpha/ffi.c, src/alpha/ffi.c, src/java_raw_api.c
3307         src/prep_cif.c, src/raw_api.c, src/ia64/ffi.c,
3308         src/mips/ffi.c, src/mips/n32.S, src/mips/o32.S,
3309         src/mips/ffitarget.h, src/sparc/ffi.c, src/x86/ffi64.c:
3310         SIZEOF_ARG -> FFI_SIZEOF_ARG.
3311         * src/ia64/ffi.c: Include stdbool.h (provided by GCC 2.95+).
3312         * src/debug.c (ffi_assert): Accept stringized failed assertion.
3313         (ffi_type_test): Rewritten.
3314         * src/prep-cif.c (initialize_aggregate, ffi_prep_cif): Call
3315         FFI_ASSERT_VALID_TYPE.
3316         * src/alpha/ffitarget.h, src/arm/ffitarget.h,
3317         src/ia64/ffitarget.h, src/m68k/ffitarget.h,
3318         src/mips/ffitarget.h, src/powerpc/ffitarget.h,
3319         src/s390/ffitarget.h, src/sh/ffitarget.h,
3320         src/sh64/ffitarget.h, src/sparc/ffitarget.h,
3321         src/x86/ffitarget.h: New files.
3322         * src/alpha/osf.S, src/arm/sysv.S, src/ia64/unix.S,
3323         src/m68k/sysv.S, src/mips/n32.S, src/mips/o32.S,
3324         src/powerpc/aix.S, src/powerpc/darwin.S,
3325         src/powerpc/ffi_darwin.c, src/powerpc/linux64.S,
3326         src/powerpc/linux64_closure.S, src/powerpc/ppc_closure.S,
3327         src/powerpc/sysv.S, src/s390/sysv.S, src/sh/sysv.S,
3328         src/sh64/sysv.S, src/sparc/v8.S, src/sparc/v9.S,
3329         src/x86/sysv.S, src/x86/unix64.S, src/x86/win32.S:
3330         include fficonfig.h
3331
3332 2003-10-20  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3333
3334         * src/mips/ffi.c: Use _ABIN32, _ABIO32 instead of external
3335         _MIPS_SIM_NABI32, _MIPS_SIM_ABI32.
3336
3337 2003-10-19  Andreas Tobler  <a.tobler@schweiz.ch>
3338
3339         * src/powerpc/ffi_darwin.c (ffi_prep_args): Declare bytes again.
3340         Used when FFI_DEBUG = 1.
3341
3342 2003-10-14  Alan Modra  <amodra@bigpond.net.au>
3343
3344         * src/types.c (double, longdouble): Default POWERPC64 to 8 byte size
3345         and align.
3346
3347 2003-10-06  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3348
3349         * include/ffi_mips.h: Define FFI_MIPS_N32 for N32/N64 ABIs,
3350         FFI_MIPS_O32 for O32 ABI.
3351
3352 2003-10-01  Andreas Tobler  <a.tobler@schweiz.ch>
3353
3354         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_64 for
3355         SPARC64. Cleanup whitespaces.
3356
3357 2003-09-19  Andreas Tobler  <a.tobler@schweiz.ch>
3358
3359         * testsuite/libffi.call/closure_fn0.c: Xfail mips, arm,
3360         strongarm, xscale. Cleanup whitespaces.
3361         * testsuite/libffi.call/closure_fn1.c: Likewise.
3362         * testsuite/libffi.call/closure_fn2.c: Likewise.
3363         * testsuite/libffi.call/closure_fn3.c: Likewise.
3364         * testsuite/libffi.call/cls_12byte.c: Likewise.
3365         * testsuite/libffi.call/cls_16byte.c: Likewise.
3366         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3367         * testsuite/libffi.call/cls_20byte.c: Likewise.
3368         * testsuite/libffi.call/cls_24byte.c: Likewise.
3369         * testsuite/libffi.call/cls_2byte.c: Likewise.
3370         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3371         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3372         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3373         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3374         * testsuite/libffi.call/cls_4byte.c: Likewise.
3375         * testsuite/libffi.call/cls_5byte.c: Likewise.
3376         * testsuite/libffi.call/cls_6byte.c: Likewise.
3377         * testsuite/libffi.call/cls_7byte.c: Likewise.
3378         * testsuite/libffi.call/cls_8byte.c: Likewise.
3379         * testsuite/libffi.call/cls_double.c: Likewise.
3380         * testsuite/libffi.call/cls_float.c: Likewise.
3381         * testsuite/libffi.call/cls_uchar.c: Likewise.
3382         * testsuite/libffi.call/cls_uint.c: Likewise.
3383         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3384         * testsuite/libffi.call/cls_ushort.c: Likewise.
3385         * testsuite/libffi.call/nested_struct.c: Likewise.
3386         * testsuite/libffi.call/nested_struct1.c: Likewise.
3387         * testsuite/libffi.call/problem1.c: Likewise.
3388         * testsuite/libffi.special/unwindtest.cc: Likewise.
3389         * testsuite/libffi.call/pyobjc-tc.c: Cleanup whitespaces.
3390
3391 2003-09-18  David Edelsohn  <edelsohn@gnu.org>
3392
3393         * src/powerpc/aix.S: Cleanup whitespaces.
3394         * src/powerpc/aix_closure.S: Likewise.
3395
3396 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3397
3398         * src/powerpc/darwin.S: Cleanup whitespaces, comment formatting.
3399         * src/powerpc/darwin_closure.S: Likewise.
3400         * src/powerpc/ffi_darwin.c: Likewise.
3401
3402 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3403             David Edelsohn  <edelsohn@gnu.org>
3404
3405         * src/types.c (double): Add AIX and Darwin to the right TYPEDEF.
3406         * src/powerpc/aix_closure.S: Remove the pointer to the outgoing
3407         parameter stack.
3408         * src/powerpc/darwin_closure.S: Likewise.
3409         * src/powerpc/ffi_darwin.c (ffi_prep_args): Handle structures
3410         according to the Darwin/AIX ABI.
3411         (ffi_prep_cif_machdep): Likewise.
3412         (ffi_closure_helper_DARWIN): Likewise.
3413         Remove the outgoing parameter stack logic. Simplify the evaluation
3414         of the different CASE types.
3415         (ffi_prep_clousure): Avoid the casts on lvalues. Change the branch
3416         statement in the trampoline code.
3417
3418 2003-09-18  Kaz Kojima  <kkojima@gcc.gnu.org>
3419
3420         * src/sh/ffi.c (ffi_prep_args): Take account into the alignement
3421         for the register size.
3422         (ffi_closure_helper_SYSV): Handle the structure return value
3423         address correctly.
3424         (ffi_closure_helper_SYSV): Return the appropriate type when
3425         the registers are used for the structure return value.
3426         * src/sh/sysv.S (ffi_closure_SYSV): Fix the stack layout for
3427         the 64-bit return value.  Update copyright years.
3428
3429 2003-09-17  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3430
3431         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Search in
3432         srcdir for ffi_mips.h.
3433
3434 2003-09-12  Alan Modra  <amodra@bigpond.net.au>
3435
3436         * src/prep_cif.c (initialize_aggregate): Include tail padding in
3437         structure size.
3438         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Correct
3439         placement of float result.
3440         * testsuite/libffi.special/unwindtest.cc (closure_test_fn1): Correct
3441         cast of "resp" for big-endian 64 bit machines.
3442
3443 2003-09-11  Alan Modra  <amodra@bigpond.net.au>
3444
3445         * src/types.c (double, longdouble): Merge identical SH and ARM
3446         typedefs, and add POWERPC64.
3447         * src/powerpc/ffi.c (ffi_prep_args64): Correct next_arg calc for