OSDN Git Service

2010-12-03 Chung-Lin Tang <cltang@codesourcery.com>
[pf3gnuchains/gcc-fork.git] / libffi / ChangeLog
1 2010-12-03  Chung-Lin Tang  <cltang@codesourcery.com>
2
3         * src/arm/sysv.S (ffi_closure_SYSV): Add UNWIND to .pad directive.
4         (ffi_closure_VFP): Same.
5         (ffi_call_VFP): Move down to before ffi_closure_VFP. Add '.fpu vfp'
6         directive.
7
8 2010-12-01  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
9
10         * testsuite/libffi.call/ffitest.h [__sgi] (PRId64, PRIu64): Define.
11         (PRIuPTR): Define.
12
13 2010-11-29  Richard Henderson  <rth@redhat.com>
14             Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
15
16         * src/x86/sysv.S (FDE_ENCODING, FDE_ENCODE): Define.
17         (.eh_frame): Use FDE_ENCODING.
18         (.LASFDE1, .LASFDE2, LASFDE3): Simplify with FDE_ENCODE.
19
20 2010-11-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
21
22         * testsuite/lib/libffi-dg.exp: Rename ...
23         * testsuite/lib/libffi.exp: ... to this.
24         * libffi/testsuite/libffi.call/call.exp: Don't load libffi-dg.exp.
25         * libffi/testsuite/libffi.special/special.exp: Likewise.
26
27 2010-10-28  Chung-Lin Tang  <cltang@codesourcery.com>
28
29         * src/arm/ffi.c (ffi_prep_args): Add VFP register argument handling
30         code, new parameter, and return value. Update comments.
31         (ffi_prep_cif_machdep): Add case for VFP struct return values. Add
32         call to layout_vfp_args().
33         (ffi_call_SYSV): Update declaration.
34         (ffi_call_VFP): New declaration.
35         (ffi_call): Add VFP struct return conditions. Call ffi_call_VFP()
36         when ABI is FFI_VFP.
37         (ffi_closure_VFP): New declaration.
38         (ffi_closure_SYSV_inner): Add new vfp_args parameter, update call to
39         ffi_prep_incoming_args_SYSV().
40         (ffi_prep_incoming_args_SYSV): Update parameters. Add VFP argument
41         case handling.
42         (ffi_prep_closure_loc): Pass ffi_closure_VFP to trampoline
43         construction under VFP hard-float.
44         (rec_vfp_type_p): New function.
45         (vfp_type_p): Same.
46         (place_vfp_arg): Same.
47         (layout_vfp_args): Same.
48         * src/arm/ffitarget.h (ffi_abi): Add FFI_VFP. Define FFI_DEFAULT_ABI
49         based on __ARM_PCS_VFP.
50         (FFI_EXTRA_CIF_FIELDS): Define for adding VFP hard-float specific
51         fields.
52         (FFI_TYPE_STRUCT_VFP_FLOAT): Define internally used type code.
53         (FFI_TYPE_STRUCT_VFP_DOUBLE): Same.
54         * src/arm/sysv.S (ffi_call_SYSV): Change call of ffi_prep_args() to
55         direct call. Move function pointer load upwards.
56         (ffi_call_VFP): New function.
57         (ffi_closure_VFP): Same.
58
59         * testsuite/lib/libffi-dg.exp (check-flags): New function.
60         (dg-skip-if): New function.
61         * testsuite/libffi.call/cls_double_va.c: Skip if target is arm*-*-*
62         and compiler options include -mfloat-abi=hard.
63         * testsuite/libffi.call/cls_longdouble_va.c: Same.
64
65 2010-10-01  Jakub Jelinek  <jakub@redhat.com>
66
67         PR libffi/45677
68         * src/x86/ffi64.c (ffi_prep_cif_machdep): Ensure cif->bytes is
69         a multiple of 8.
70         * testsuite/libffi.call/many2.c: New test.
71
72 2010-08-20  Mark Wielaard  <mjw@redhat.com>
73
74         * src/closures.c (open_temp_exec_file_mnt): Check if getmntent_r
75         returns NULL.
76
77 2010-08-09  Andreas Tobler  <andreast@fgznet.ch>
78
79         * configure.ac: Add target powerpc64-*-freebsd*.
80         * configure: Regenerate.
81         * testsuite/libffi.call/cls_align_longdouble_split.c: Pass
82         -mlong-double-128 only to linux targets.
83         * testsuite/libffi.call/cls_align_longdouble_split2.c: Likewise.
84         * testsuite/libffi.call/cls_longdouble.c: Likewise.
85         * testsuite/libffi.call/huge_struct.c: Likewise.
86
87 2010-07-10  Evan Phoenix  <evan@fallingsnow.net>
88
89         * src/closures.c (selinux_enabled_check): Fix strncmp usage bug.
90
91 2010-07-07  Dan HorĂ¡k <dan@danny.cz>
92
93         * include/ffi.h.in: Protect #define with #ifndef.
94         * src/powerpc/ffitarget.h: Ditto.
95         * src/s390/ffitarget.h: Ditto.
96         * src/sparc/ffitarget.h: Ditto.
97
98 2010-07-07   Neil Roberts <neil@linux.intel.com>
99
100         * src/x86/sysv.S (ffi_call_SYSV): Align the stack pointer to
101         16-bytes.
102
103 2010-07-02  Jakub Jelinek  <jakub@redhat.com>
104
105         * Makefile.am (AM_MAKEFLAGS): Pass also mandir to submakes.
106         * Makefile.in: Regenerated.
107
108 2010-05-19  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
109
110         * configure.ac (libffi_cv_as_x86_pcrel): Check for illegal in as
111         output, too.
112         (libffi_cv_as_ascii_pseudo_op): Check for .ascii.
113         (libffi_cv_as_string_pseudo_op): Check for .string.
114         * configure: Regenerate.
115         * fficonfig.h.in: Regenerate.
116         * src/x86/sysv.S (.eh_frame): Use .ascii, .string or error.
117
118 2010-05-05  Michael Kohler <michaelkohler@live.com>
119
120         * src/dlmalloc.c (dlfree): Fix spelling.
121         * src/ia64/ffi.c (ffi_prep_cif_machdep): Ditto.
122         * configure.ac: Ditto.
123         * configure: Rebuilt.
124
125 2010-04-13  Dan Witte  <dwitte@mozilla.com>
126
127         * msvcc.sh: Build with -W3 instead of -Wall.
128         * src/powerpc/ffi_darwin.c: Remove build warnings.
129         * src/x86/ffi.c: Ditto.
130         * src/x86/ffitarget.h: Ditto.
131
132 2010-04-12  Dan Witte  <dwitte@mozilla.com>
133             Walter Meinl <wuno@lsvw.de>
134
135         * configure.ac: Add OS/2 support.
136         * configure: Rebuilt.
137         * src/closures.c: Ditto.
138         * src/dlmalloc.c: Ditto.
139         * src/x86/win32.S: Ditto.
140
141 2010-04-07  Jakub Jelinek  <jakub@redhat.com>
142
143         * testsuite/libffi.call/err_bad_abi.c: Remove unused args variable.
144
145 2010-04-02  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
146
147         * Makefile.in: Regenerate.
148         * aclocal.m4: Regenerate.
149         * include/Makefile.in: Regenerate.
150         * man/Makefile.in: Regenerate.
151         * testsuite/Makefile.in: Regenerate.
152
153 2010-03-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
154
155         * configure.ac (libffi_cv_as_x86_64_unwind_section_type): New test.
156         * configure: Regenerate.
157         * fficonfig.h.in: Regenerate.
158         * libffi/src/x86/unix64.S (.eh_frame)
159         [HAVE_AS_X86_64_UNWIND_SECTION_TYPE]: Use @unwind section type.
160
161 2010-03-14  Matthias Klose  <doko@ubuntu.com>
162
163         * src/x86/ffi64.c: Fix typo in comment.
164         * src/x86/ffi.c: Use /* ... */ comment style.
165
166 2010-02-24  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
167
168         * doc/libffi.texi (The Closure API): Fix typo.
169         * doc/libffi.info: Remove.
170
171 2010-02-15  Matthias Klose  <doko@ubuntu.com>
172
173         * src/arm/sysv.S (__ARM_ARCH__): Define for processor
174         __ARM_ARCH_7EM__.
175
176 2010-01-15  Anthony Green  <green@redhat.com>
177
178         * README: Add notes on building with Microsoft Visual C++.
179
180 2010-01-15  Daniel Witte  <dwitte@mozilla.com>
181
182         * msvcc.sh: New file.
183
184         * src/x86/win32.S: Port assembly routines to MSVC and #ifdef.
185         * src/x86/ffi.c: Tweak function declaration and remove excess
186         parens.
187         * include/ffi.h.in: Add __declspec(align(8)) to typedef struct
188         ffi_closure.
189
190         * src/x86/ffi.c: Merge ffi_call_SYSV and ffi_call_STDCALL into new
191         function ffi_call_win32 on X86_WIN32.
192         * src/x86/win32.S (ffi_call_SYSV): Rename to ffi_call_win32.
193         (ffi_call_STDCALL): Remove.
194
195         * src/prep_cif.c (ffi_prep_cif): Move stack space allocation code
196         to ffi_prep_cif_machdep for x86.
197         * src/x86/ffi.c (ffi_prep_cif_machdep): To here.
198
199 2010-01-15  Oliver Kiddle  <okiddle@yahoo.co.uk>
200
201         * src/x86/ffitarget.h (ffi_abi): Check for __i386 and __amd64 for
202         Sun Studio compiler compatibility. 
203
204 2010-01-12  Conrad Irwin <conrad.irwin@gmail.com>
205
206         * doc/libffi.texi: Add closure example.
207
208 2010-01-07  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
209
210         PR libffi/40701
211         * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRIdLL,
212         PRIuLL, PRId64, PRIu64, PRIuPTR): Define.
213         * testsuite/libffi.call/cls_align_sint64.c: Add -Wno-format on
214         alpha*-dec-osf*.
215         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
216         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
217         * testsuite/libffi.call/return_ll1.c: Likewise.
218         * testsuite/libffi.call/stret_medium2.c: Likewise.
219         * testsuite/libffi.special/ffitestcxx.h (allocate_mmap): Cast
220         MAP_FAILED to char *.
221
222 2010-01-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
223
224         * src/mips/n32.S: Use .abicalls and .eh_frame with __GNUC__.
225
226 2009-12-31  Anthony Green  <green@redhat.com>
227
228         * README: Update for libffi 3.0.9.
229
230 2009-12-27  Matthias Klose  <doko@ubuntu.com>
231
232         * configure.ac (HAVE_LONG_DOUBLE): Define for mips when
233         appropriate.
234         * configure: Rebuilt.
235
236 2009-12-26  Anthony Green  <green@redhat.com>
237
238         * testsuite/libffi.call/cls_longdouble_va.c: Mark as xfail for
239         avr32*-*-*.
240         * testsuite/libffi.call/cls_double_va.c: Ditto.
241
242 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
243
244         * testsuite/libffi.call/ffitest.h: Conditionally include stdint.h
245         and inttypes.h.
246         * testsuite/libffi.special/unwindtest.cc: Ditto.
247
248 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
249
250         * configure.ac: Add amd64-*-openbsd*.
251         * configure: Rebuilt.
252         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Link
253         openbsd programs with -lpthread.
254
255 2009-12-26  Anthony Green  <green@redhat.com>
256
257         * testsuite/libffi.call/cls_double_va.c,
258         testsuite/libffi.call/cls_longdouble.c,
259         testsuite/libffi.call/cls_longdouble_va.c,
260         testsuite/libffi.call/cls_pointer.c,
261         testsuite/libffi.call/cls_pointer_stack.c: Remove xfail for
262         mips*-*-* and arm*-*-*.
263         * testsuite/libffi.call/cls_align_longdouble_split.c,
264         testsuite/libffi.call/cls_align_longdouble_split2.c,
265         testsuite/libffi.call/stret_medium2.c,
266         testsuite/libffi.call/stret_medium.c,
267         testsuite/libffi.call/stret_large.c,
268         testsuite/libffi.call/stret_large2.c: Remove xfail for arm*-*-*.
269
270 2009-12-31  Kay Tietz  <ktietz70@googlemail.com>
271
272         * testsuite/libffi.call/ffitest.h,
273         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRuLL): Fix
274         definitions.
275
276 2009-12-31  Carlo Bramini  <carlo.bramix@libero.it>
277
278         * configure.ac (AM_LTLDFLAGS): Define for windows hosts.
279         * Makefile.am (libffi_la_LDFLAGS): Add AM_LTLDFLAGS.
280         * configure: Rebuilt.
281         * Makefile.in: Rebuilt.
282
283 2009-12-31  Anthony Green  <green@redhat.com>
284             Blake Chaffin.
285
286         * testsuite/libffi.call/huge_struct.c: New test case from Blake
287         Chaffin @ Apple.
288
289 2009-12-28  David Edelsohn  <edelsohn@gnu.org>
290
291         * src/powerpc/ffi_darwin.c (ffi_prep_args): Copy abi and nargs to
292         local variables.
293         (aix_adjust_aggregate_sizes): New function.
294         (ffi_prep_cif_machdep): Call it.
295
296 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
297
298         * configure.ac: Define FFI_MMAP_EXEC_WRIT for the given targets.
299         * configure: Regenerate.
300         * fficonfig.h.in: Likewise.
301         * src/closures.c: Remove the FFI_MMAP_EXEC_WRIT definition for
302         Solaris/x86.
303
304 2009-12-26  Andreas Schwab  <schwab@linux-m68k.org>
305
306         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Advance intarg_count
307         when a float arguments is passed in memory.
308         (ffi_closure_helper_SYSV): Mark general registers as used up when
309         a 64bit or soft-float long double argument is passed in memory.
310
311 2009-12-25  Matthias Klose  <doko@ubuntu.com>
312
313         * man/ffi_call.3: Fix #include in examples.
314         * doc/libffi.texi: Add dircategory.
315
316 2009-12-25  Frank Everdij <f.p.x.everdij@tudelft.nl>
317
318         * include/ffi.h.in: Placed '__GNUC__' ifdef around
319         '__attribute__((aligned(8)))' in ffi_closure, fixes compile for
320         IRIX MIPSPro c99.
321         * include/ffi_common.h: Added '__sgi' define to non
322         '__attribute__((__mode__()))' integer typedefs.
323         * src/mips/ffi.c (ffi_call, ffi_closure_mips_inner_O32,
324         ffi_closure_mips_inner_N32): Added 'defined(_MIPSEB)' to BE check.
325         (ffi_closure_mips_inner_O32, ffi_closure_mips_inner_N32): Added
326         FFI_LONGDOUBLE support and alignment(N32 only).
327         * src/mips/ffitarget.h: Corrected '#include <sgidefs.h>' for IRIX and
328         fixed non '__attribute__((__mode__()))' integer typedefs.
329         * src/mips/n32.S: Put '#ifdef linux' around '.abicalls' and '.eh_frame'
330         since they are Linux/GNU Assembler specific.
331
332 2009-12-25  Bradley Smith  <brad@brad-smith.co.uk>
333
334         * configure.ac, Makefile.am, src/avr32/ffi.c,
335         src/avr32/ffitarget.h,
336         src/avr32/sysv.S: Add AVR32 port.
337         * configure, Makefile.in: Rebuilt.
338
339 2009-12-21  Andreas Tobler  <a.tobler@schweiz.org>
340
341         * configure.ac: Make i?86 build on FreeBSD and OpenBSD.
342         * configure: Regenerate.
343
344 2009-12-15  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
345
346         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on PA HP-UX.
347
348 2009-12-13  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
349
350         * src/pa/ffi.c (ffi_closure_inner_pa32): Handle FFI_TYPE_LONGDOUBLE
351         type on HP-UX.
352
353 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
354
355         * src/sparc/ffi.c (ffi_closure_sparc_inner_v9): Properly align 'long
356         double' arguments.
357
358 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
359
360         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on Solaris < 10.
361
362 2009-12-10  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
363
364         PR libffi/40700
365         * src/closures.c [X86_64 && __sun__ && __svr4__]
366         (FFI_MMAP_EXEC_WRIT): Define.
367
368 2009-12-08  David Daney  <ddaney@caviumnetworks.com>
369
370         * testsuite/libffi.call/stret_medium.c: Remove xfail for mips*-*-*
371         * testsuite/libffi.call/cls_align_longdouble_split2.c: Same.
372         * testsuite/libffi.call/stret_large.c: Same.
373         * testsuite/libffi.call/cls_align_longdouble_split.c: Same.
374         * testsuite/libffi.call/stret_large2.c: Same.
375         * testsuite/libffi.call/stret_medium2.c: Same.
376
377 2009-12-07  David Edelsohn  <edelsohn@gnu.org>
378
379         * src/powerpc/aix_closure.S (libffi_closure_ASM): Fix tablejump
380         typo.
381
382 2009-12-05  David Edelsohn  <edelsohn@gnu.org>
383
384         * src/powerpc/aix.S: Update AIX32 code to be consistent with AIX64
385         code.
386         * src/powerpc/aix_closure.S: Same.
387
388 2009-12-05  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
389
390         * Makefile.in: Regenerate.
391         * configure: Regenerate.
392         * include/Makefile.in: Regenerate.
393         * man/Makefile.in: Regenerate.
394         * testsuite/Makefile.in: Regenerate.
395
396 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
397
398         * src/powerpc/aix_closure.S: Reorganize 64-bit code to match
399         linux64_closure.S.
400
401 2009-12-04  Uros Bizjak  <ubizjak@gmail.com>
402
403         PR libffi/41908
404         * src/x86/ffi64.c (classify_argument): Update from
405         gcc/config/i386/i386.c.
406         (ffi_closure_unix64_inner): Do not use the address of two consecutive
407         SSE registers directly.
408         * testsuite/libffi.call/cls_dbls_struct.c (main): Remove xfail
409         for x86_64 linux targets.
410
411 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
412
413         * src/powerpc/ffi_darwin.c (ffi_closure_helper_DARWIN): Increment
414         pfr for long double split between fpr13 and stack.
415
416 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
417
418         * src/powerpc/ffi_darwin.c (ffi_prep_args): Increment next_arg and
419         fparg_count twice for long double.
420
421 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
422
423         PR libffi/42243
424         * src/powerpc/ffi_darwin.c (ffi_prep_args): Remove extra parentheses.
425
426 2009-12-03  Uros Bizjak  <ubizjak@gmail.com>
427
428         * testsuite/libffi.call/cls_longdouble_va.c (main): Fix format string.
429         Remove xfails for x86 linux targets.
430
431 2009-12-02  David Edelsohn  <edelsohn@gnu.org>
432
433         * src/powerpc/ffi_darwin.c (ffi_prep_args): Fix typo in INT64
434         case.
435
436 2009-12-01  David Edelsohn  <edelsohn@gnu.org>
437
438         * src/powerpc/aix.S (ffi_call_AIX): Convert to more standard
439         register usage.  Call ffi_prep_args directly.  Add long double
440         return value support.
441         * src/powerpc/ffi_darwin.c (ffi_prep_args): Double arg increment
442         applies to FFI_TYPE_DOUBLE.  Correct fpr_base increment typo.
443         Separate FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases.
444         (ffi_prep_cif_machdep): Only 16 byte stack alignment in 64 bit
445         mode.
446         (ffi_closure_helper_DARWIN): Remove nf and ng counters.  Move temp
447         into case.
448         * src/powerpc/aix_closure.S: Maintain 16 byte stack alignment.
449         Allocate result area between params and FPRs.
450
451 2009-11-30  David Edelsohn  <edelsohn@gnu.org>
452
453         PR target/35484
454         * src/powerpc/ffitarget.h (POWERPC64): Define for PPC64 Linux and
455         AIX64.
456         * src/powerpc/aix.S: Implement AIX64 version.
457         * src/powerpc/aix_closure.S: Implement AIX64 version.
458         (ffi_closure_ASM): Use extsb, lha and displament addresses.
459         * src/powerpc/ffi_darwin.c (ffi_prep_args): Implement AIX64
460         support.
461         (ffi_prep_cif_machdep): Same.
462         (ffi_call): Same.
463         (ffi_closure_helper_DARWIN): Same.
464
465 2009-11-02  Andreas Tobler  <a.tobler@schweiz.org>
466
467         PR libffi/41908
468         * testsuite/libffi.call/testclosure.c: New test.
469
470 2009-09-28  Kai Tietz  <kai.tietz@onevision.com>
471
472         * src/x86/win64.S (_ffi_call_win64 stack): Remove for gnu
473         assembly version use of ___chkstk.
474
475 2009-09-23  Matthias Klose  <doko@ubuntu.com>
476
477         PR libffi/40242, PR libffi/41443
478         * src/arm/sysv.S (__ARM_ARCH__): Define for processors
479         __ARM_ARCH_6T2__, __ARM_ARCH_6M__, __ARM_ARCH_7__,
480         __ARM_ARCH_7A__, __ARM_ARCH_7R__, __ARM_ARCH_7M__.
481         Change the conditionals to __SOFTFP__ || __ARM_EABI__
482         for -mfloat-abi=softfp to work.
483
484 2009-09-17  Loren J. Rittle  <ljrittle@acm.org>
485
486         PR testsuite/32843 (strikes again)
487         * src/x86/ffi.c (ffi_prep_cif_machdep): Add X86_FREEBSD to
488         enable proper extension on char and short.
489
490 2009-09-15  David Daney  <ddaney@caviumnetworks.com>
491
492         * src/java_raw_api.c (ffi_java_raw_to_rvalue): Remove special
493         handling for FFI_TYPE_POINTER.
494         * src/mips/ffitarget.h (FFI_TYPE_STRUCT_D_SOFT,
495         FFI_TYPE_STRUCT_F_SOFT, FFI_TYPE_STRUCT_DD_SOFT,
496         FFI_TYPE_STRUCT_FF_SOFT, FFI_TYPE_STRUCT_FD_SOFT,
497         FFI_TYPE_STRUCT_DF_SOFT, FFI_TYPE_STRUCT_SOFT): New defines.
498         (FFI_N32_SOFT_FLOAT, FFI_N64_SOFT_FLOAT): New ffi_abi enumerations.
499         (enum ffi_abi): Set FFI_DEFAULT_ABI for soft-float.
500         * src/mips/n32.S (ffi_call_N32): Add handling for soft-float
501         structure and pointer returns.
502         (ffi_closure_N32): Add handling for pointer returns.
503         * src/mips/ffi.c (ffi_prep_args, calc_n32_struct_flags,
504         calc_n32_return_struct_flags): Handle soft-float.
505         (ffi_prep_cif_machdep):  Handle soft-float, fix pointer handling.
506         (ffi_call_N32): Declare proper argument types.
507         (ffi_call, copy_struct_N32, ffi_closure_mips_inner_N32): Handle
508         soft-float.
509
510 2009-08-24  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
511
512         * configure.ac (AC_PREREQ): Bump to 2.64.
513
514 2009-08-22  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
515
516         * Makefile.am (install-html, install-pdf): Remove.
517         * Makefile.in: Regenerate.
518
519         * Makefile.in: Regenerate.
520         * aclocal.m4: Regenerate.
521         * configure: Regenerate.
522         * fficonfig.h.in: Regenerate.
523         * include/Makefile.in: Regenerate.
524         * man/Makefile.in: Regenerate.
525         * testsuite/Makefile.in: Regenerate.
526
527 2009-07-30  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
528
529         * configure.ac (_AC_ARG_VAR_PRECIOUS): Use m4_rename_force.
530
531 2009-07-24  Dave Korn  <dave.korn.cygwin@gmail.com>
532
533         PR libffi/40807
534         * src/x86/ffi.c (ffi_prep_cif_machdep): Also use sign/zero-extending
535         return types for X86_WIN32.
536         * src/x86/win32.S (_ffi_call_SYSV): Handle omitted return types.
537         (_ffi_call_STDCALL, _ffi_closure_SYSV, _ffi_closure_raw_SYSV,
538         _ffi_closure_STDCALL): Likewise.
539
540         * src/closures.c (is_selinux_enabled): Define to const 0 for Cygwin.
541         (dlmmap, dlmunmap): Also use these functions on Cygwin.
542
543 2009-07-11  Richard Sandiford  <rdsandiford@googlemail.com>
544
545         PR testsuite/40699
546         PR testsuite/40707
547         PR testsuite/40709
548         * testsuite/lib/libffi-dg.exp: Revert 2009-07-02, 2009-07-01 and
549         2009-06-30 commits.
550
551 2009-07-01  Richard Sandiford  <r.sandiford@uk.ibm.com>
552
553         * testsuite/lib/libffi-dg.exp (libffi-init): Set ld_library_path
554         to "" before adding paths.  (This reinstates an assignment that
555         was removed by my 2009-06-30 commit, but changes the initial
556         value from "." to "".)
557
558 2009-07-01  H.J. Lu  <hongjiu.lu@intel.com>
559
560         PR testsuite/40601
561         * testsuite/lib/libffi-dg.exp (libffi-init): Properly set
562         gccdir.  Adjust ld_library_path for gcc only if gccdir isn't
563         empty.
564
565 2009-06-30  Richard Sandiford  <r.sandiford@uk.ibm.com>
566
567         * testsuite/lib/libffi-dg.exp (libffi-init): Don't add "."
568         to ld_library_path.  Use add_path.  Add just find_libgcc_s
569         to ld_library_path, not every libgcc multilib directory.
570
571 2009-06-16  Wim Lewis  <wiml@hhhh.org>
572
573         * src/powerpc/ffi.c: Avoid clobbering cr3 and cr4, which are
574         supposed to be callee-saved.
575         * src/powerpc/sysv.S (small_struct_return_value): Fix overrun of
576         return buffer for odd-size structs.
577
578 2009-06-16  Andreas Tobler  <a.tobler@schweiz.org>
579
580         PR libffi/40444
581         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Add
582         allow_stack_execute for Darwin.
583
584 2009-06-16  Andrew Haley  <aph@redhat.com>
585
586         * configure.ac (TARGETDIR): Add missing blank lines.
587         * configure: Regenerate.
588
589 2009-06-16  Andrew Haley  <aph@redhat.com>
590
591         * testsuite/libffi.call/cls_align_sint64.c,
592         testsuite/libffi.call/cls_align_uint64.c,
593         testsuite/libffi.call/cls_longdouble_va.c,
594         testsuite/libffi.call/cls_ulonglong.c,
595         testsuite/libffi.call/return_ll1.c,
596         testsuite/libffi.call/stret_medium2.c: Fix printf format
597         specifiers.
598         * testsuite/libffi.call/ffitest.h,
599         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRIuLL): Define.
600
601 2009-06-15  Andrew Haley  <aph@redhat.com>
602
603         * testsuite/libffi.call/err_bad_typedef.c: xfail everywhere.
604         * testsuite/libffi.call/err_bad_abi.c: Likewise.
605
606 2009-06-12  Andrew Haley  <aph@redhat.com>
607
608         * Makefile.am: Remove info_TEXINFOS.
609
610 2009-06-12  Andrew Haley  <aph@redhat.com>
611
612         * ChangeLog.libffi: testsuite/libffi.call/cls_align_sint64.c,
613         testsuite/libffi.call/cls_align_uint64.c,
614         testsuite/libffi.call/cls_ulonglong.c,
615         testsuite/libffi.call/return_ll1.c,
616         testsuite/libffi.call/stret_medium2.c: Fix printf format
617         specifiers.
618         testsuite/libffi.special/unwindtest.cc: include stdint.h.
619
620 2009-06-11  Timothy Wall  <twall@users.sf.net>
621
622         * Makefile.am,
623         configure.ac,
624         include/ffi.h.in,
625         include/ffi_common.h,
626         src/closures.c,
627         src/dlmalloc.c,
628         src/x86/ffi.c,
629         src/x86/ffitarget.h,
630         src/x86/win64.S (new),
631         README: Added win64 support (mingw or MSVC)
632         * Makefile.in,
633         include/Makefile.in,
634         man/Makefile.in,
635         testsuite/Makefile.in,
636         configure,
637         aclocal.m4: Regenerated
638         * ltcf-c.sh: properly escape cygwin/w32 path
639         * man/ffi_call.3: Clarify size requirements for return value.
640         * src/x86/ffi64.c: Fix filename in comment.
641         * src/x86/win32.S: Remove unused extern.
642
643         * testsuite/libffi.call/closure_fn0.c,
644         testsuite/libffi.call/closure_fn1.c,
645         testsuite/libffi.call/closure_fn2.c,
646         testsuite/libffi.call/closure_fn3.c,
647         testsuite/libffi.call/closure_fn4.c,
648         testsuite/libffi.call/closure_fn5.c,
649         testsuite/libffi.call/closure_fn6.c,
650         testsuite/libffi.call/closure_stdcall.c,
651         testsuite/libffi.call/cls_12byte.c,
652         testsuite/libffi.call/cls_16byte.c,
653         testsuite/libffi.call/cls_18byte.c,
654         testsuite/libffi.call/cls_19byte.c,
655         testsuite/libffi.call/cls_1_1byte.c,
656         testsuite/libffi.call/cls_20byte.c,
657         testsuite/libffi.call/cls_20byte1.c,
658         testsuite/libffi.call/cls_24byte.c,
659         testsuite/libffi.call/cls_2byte.c,
660         testsuite/libffi.call/cls_3_1byte.c,
661         testsuite/libffi.call/cls_3byte1.c,
662         testsuite/libffi.call/cls_3byte2.c,
663         testsuite/libffi.call/cls_4_1byte.c,
664         testsuite/libffi.call/cls_4byte.c,
665         testsuite/libffi.call/cls_5_1_byte.c,
666         testsuite/libffi.call/cls_5byte.c,
667         testsuite/libffi.call/cls_64byte.c,
668         testsuite/libffi.call/cls_6_1_byte.c,
669         testsuite/libffi.call/cls_6byte.c,
670         testsuite/libffi.call/cls_7_1_byte.c,
671         testsuite/libffi.call/cls_7byte.c,
672         testsuite/libffi.call/cls_8byte.c,
673         testsuite/libffi.call/cls_9byte1.c,
674         testsuite/libffi.call/cls_9byte2.c,
675         testsuite/libffi.call/cls_align_double.c,
676         testsuite/libffi.call/cls_align_float.c,
677         testsuite/libffi.call/cls_align_longdouble.c,
678         testsuite/libffi.call/cls_align_longdouble_split.c,
679         testsuite/libffi.call/cls_align_longdouble_split2.c,
680         testsuite/libffi.call/cls_align_pointer.c,
681         testsuite/libffi.call/cls_align_sint16.c,
682         testsuite/libffi.call/cls_align_sint32.c,
683         testsuite/libffi.call/cls_align_sint64.c,
684         testsuite/libffi.call/cls_align_uint16.c,
685         testsuite/libffi.call/cls_align_uint32.c,
686         testsuite/libffi.call/cls_align_uint64.c,
687         testsuite/libffi.call/cls_dbls_struct.c,
688         testsuite/libffi.call/cls_double.c,
689         testsuite/libffi.call/cls_double_va.c,
690         testsuite/libffi.call/cls_float.c,
691         testsuite/libffi.call/cls_longdouble.c,
692         testsuite/libffi.call/cls_longdouble_va.c,
693         testsuite/libffi.call/cls_multi_schar.c,
694         testsuite/libffi.call/cls_multi_sshort.c,
695         testsuite/libffi.call/cls_multi_sshortchar.c,
696         testsuite/libffi.call/cls_multi_uchar.c,
697         testsuite/libffi.call/cls_multi_ushort.c,
698         testsuite/libffi.call/cls_multi_ushortchar.c,
699         testsuite/libffi.call/cls_pointer.c,
700         testsuite/libffi.call/cls_pointer_stack.c,
701         testsuite/libffi.call/cls_schar.c,
702         testsuite/libffi.call/cls_sint.c,
703         testsuite/libffi.call/cls_sshort.c,
704         testsuite/libffi.call/cls_uchar.c,
705         testsuite/libffi.call/cls_uint.c,
706         testsuite/libffi.call/cls_ulonglong.c,
707         testsuite/libffi.call/cls_ushort.c,
708         testsuite/libffi.call/err_bad_abi.c,
709         testsuite/libffi.call/err_bad_typedef.c,
710         testsuite/libffi.call/float2.c,
711         testsuite/libffi.call/huge_struct.c,
712         testsuite/libffi.call/nested_struct.c,
713         testsuite/libffi.call/nested_struct1.c,
714         testsuite/libffi.call/nested_struct10.c,
715         testsuite/libffi.call/nested_struct2.c,
716         testsuite/libffi.call/nested_struct3.c,
717         testsuite/libffi.call/nested_struct4.c,
718         testsuite/libffi.call/nested_struct5.c,
719         testsuite/libffi.call/nested_struct6.c,
720         testsuite/libffi.call/nested_struct7.c,
721         testsuite/libffi.call/nested_struct8.c,
722         testsuite/libffi.call/nested_struct9.c,
723         testsuite/libffi.call/problem1.c,
724         testsuite/libffi.call/return_ldl.c,
725         testsuite/libffi.call/return_ll1.c,
726         testsuite/libffi.call/stret_large.c,
727         testsuite/libffi.call/stret_large2.c,
728         testsuite/libffi.call/stret_medium.c,
729         testsuite/libffi.call/stret_medium2.c,
730         testsuite/libffi.special/unwindtest.cc: use ffi_closure_alloc instead
731         of checking for MMAP.  Use intptr_t instead of long casts.
732
733 2009-06-11  Kaz Kojima  <kkojima@gcc.gnu.org>
734
735         * testsuite/libffi.call/cls_longdouble_va.c: Add xfail sh*-*-linux-*.
736         * testsuite/libffi.call/err_bad_abi.c: Add xfail sh*-*-*.
737         * testsuite/libffi.call/err_bad_typedef.c: Likewise.
738
739 2009-06-09  Andrew Haley  <aph@redhat.com>
740
741         * src/x86/freebsd.S: Add missing file.
742
743 2009-06-08  Andrew Haley  <aph@redhat.com>
744
745         Import from libffi 3.0.8:
746
747         * doc/libffi.texi: New file.
748         * doc/libffi.info: Likewise.
749         * doc/stamp-vti: Likewise.
750         * man/Makefile.am: New file.
751         * man/ffi_call.3: New file.
752
753         * Makefile.am (EXTRA_DIST): Add src/x86/darwin64.S,
754         src/dlmalloc.c.
755         (nodist_libffi_la_SOURCES): Add X86_FREEBSD.
756
757         * configure.ac: Bump version to 3.0.8.
758         parisc*-*-linux*: Add.
759         i386-*-freebsd* | i386-*-openbsd*: Add.
760         powerpc-*-beos*: Add.
761         AM_CONDITIONAL X86_FREEBSD: Add.
762         AC_CONFIG_FILES: Add man/Makefile.
763
764         * include/ffi.h.in (FFI_FN): Change void (*)() to void (*)(void).
765
766 2009-06-08  Andrew Haley  <aph@redhat.com>
767
768         * README: Import from libffi 3.0.8.
769
770 2009-06-08  Andrew Haley  <aph@redhat.com>
771
772         * testsuite/libffi.call/err_bad_abi.c: Add xfails.
773         * testsuite/libffi.call/cls_longdouble_va.c: Add xfails.
774         * testsuite/libffi.call/cls_dbls_struct.c: Add xfail x86_64-*-linux-*.
775         * testsuite/libffi.call/err_bad_typedef.c: Add xfails.
776
777         * testsuite/libffi.call/stret_medium2.c: Add __UNUSED__ to args.
778         * testsuite/libffi.call/stret_medium.c: Likewise.
779         * testsuite/libffi.call/stret_large2.c: Likewise.
780         * testsuite/libffi.call/stret_large.c:  Likewise.
781
782 2008-12-26  Timothy Wall  <twall@users.sf.net>
783
784         * testsuite/libffi.call/cls_longdouble.c,
785         testsuite/libffi.call/cls_longdouble_va.c,
786         testsuite/libffi.call/cls_align_longdouble.c,
787         testsuite/libffi.call/cls_align_longdouble_split.c,
788         testsuite/libffi.call/cls_align_longdouble_split2.c: mark expected
789         failures on x86_64 cygwin/mingw.
790
791 2008-12-22  Timothy Wall  <twall@users.sf.net>
792
793         * testsuite/libffi.call/closure_fn0.c,
794         testsuite/libffi.call/closure_fn1.c,
795         testsuite/libffi.call/closure_fn2.c,
796         testsuite/libffi.call/closure_fn3.c,
797         testsuite/libffi.call/closure_fn4.c,
798         testsuite/libffi.call/closure_fn5.c,
799         testsuite/libffi.call/closure_fn6.c,
800         testsuite/libffi.call/closure_loc_fn0.c,
801         testsuite/libffi.call/closure_stdcall.c,
802         testsuite/libffi.call/cls_align_pointer.c,
803         testsuite/libffi.call/cls_pointer.c,
804         testsuite/libffi.call/cls_pointer_stack.c: use portable cast from
805         pointer to integer (intptr_t).
806         * testsuite/libffi.call/cls_longdouble.c: disable for win64.
807
808 2008-07-24  Anthony Green  <green@redhat.com>
809
810         * testsuite/libffi.call/cls_dbls_struct.c,
811         testsuite/libffi.call/cls_double_va.c,
812         testsuite/libffi.call/cls_longdouble.c,
813         testsuite/libffi.call/cls_longdouble_va.c,
814         testsuite/libffi.call/cls_pointer.c,
815         testsuite/libffi.call/cls_pointer_stack.c,
816         testsuite/libffi.call/err_bad_abi.c: Clean up failures from
817         compiler warnings.
818
819 2008-03-04  Anthony Green  <green@redhat.com>
820             Blake Chaffin
821             hos@tamanegi.org
822
823         * testsuite/libffi.call/cls_align_longdouble_split2.c
824           testsuite/libffi.call/cls_align_longdouble_split.c
825           testsuite/libffi.call/cls_dbls_struct.c
826           testsuite/libffi.call/cls_double_va.c
827           testsuite/libffi.call/cls_longdouble.c
828           testsuite/libffi.call/cls_longdouble_va.c
829           testsuite/libffi.call/cls_pointer.c
830           testsuite/libffi.call/cls_pointer_stack.c
831           testsuite/libffi.call/err_bad_abi.c
832           testsuite/libffi.call/err_bad_typedef.c
833           testsuite/libffi.call/stret_large2.c
834           testsuite/libffi.call/stret_large.c
835           testsuite/libffi.call/stret_medium2.c
836           testsuite/libffi.call/stret_medium.c: New tests from Apple.
837
838 2009-06-05  Andrew Haley  <aph@redhat.com>
839
840         * src/x86/ffitarget.h, src/x86/ffi.c: Merge stdcall changes from
841         libffi.
842
843 2009-06-04  Andrew Haley  <aph@redhat.com>
844
845         * src/x86/ffitarget.h, src/x86/win32.S, src/x86/ffi.c: Back out
846         stdcall changes.
847
848 2008-02-26  Anthony Green  <green@redhat.com>
849             Thomas Heller  <theller@ctypes.org>
850
851         * src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C
852         comment.
853
854 2008-02-03  Timothy Wall  <twall@users.sf.net>
855
856         * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return
857           offset based on code pointer, not data pointer.
858
859 2008-01-31  Timothy Wall <twall@users.sf.net>
860
861         * testsuite/libffi.call/closure_stdcall.c: Add test for stdcall
862         closures.
863         * src/x86/ffitarget.h: Increase size of trampoline for stdcall
864         closures.
865         * src/x86/win32.S: Add assembly for stdcall closure.
866         * src/x86/ffi.c: Initialize stdcall closure trampoline.
867
868 2009-06-04  Andrew Haley  <aph@redhat.com>
869
870         * include/ffi.h.in: Change void (*)() to void (*)(void).
871         * src/x86/ffi.c: Likewise.
872
873 2009-06-04  Andrew Haley  <aph@redhat.com>
874
875         * src/powerpc/ppc_closure.S: Insert licence header.
876         * src/powerpc/linux64_closure.S: Likewise.
877         * src/m68k/sysv.S: Likewise.
878
879         * src/sh64/ffi.c: Change void (*)() to void (*)(void).
880         * src/powerpc/ffi.c: Likewise.
881         * src/powerpc/ffi_darwin.c: Likewise.
882         * src/m32r/ffi.c: Likewise.
883         * src/sh64/ffi.c: Likewise.
884         * src/x86/ffi64.c: Likewise.
885         * src/alpha/ffi.c: Likewise.
886         * src/alpha/osf.S: Likewise.
887         * src/frv/ffi.c: Likewise.
888         * src/s390/ffi.c: Likewise.
889         * src/pa/ffi.c: Likewise.
890         * src/pa/hpux32.S: Likewise.
891         * src/ia64/unix.S: Likewise.
892         * src/ia64/ffi.c: Likewise.
893         * src/sparc/ffi.c: Likewise.
894         * src/mips/ffi.c: Likewise.
895         * src/sh/ffi.c: Likewise.
896
897 2008-02-15  David Daney  <ddaney@avtrex.com>
898
899         * src/mips/ffi.c (USE__BUILTIN___CLEAR_CACHE):
900         Define (conditionally), and use it to include cachectl.h.
901         (ffi_prep_closure_loc): Fix cache flushing.
902         * src/mips/ffitarget.h (_ABIN32, _ABI64, _ABIO32): Define.
903
904 2009-06-04  Andrew Haley  <aph@redhat.com>
905
906         include/ffi.h.in,
907         src/arm/ffitarget.h,
908         src/arm/ffi.c,
909         src/arm/sysv.S,
910         src/powerpc/ffitarget.h,
911         src/closures.c,
912         src/sh64/ffitarget.h,
913         src/sh64/ffi.c,
914         src/sh64/sysv.S,
915         src/types.c,
916         src/x86/ffi64.c,
917         src/x86/ffitarget.h,
918         src/x86/win32.S,
919         src/x86/darwin.S,
920         src/x86/ffi.c,
921         src/x86/sysv.S,
922         src/x86/unix64.S,
923         src/alpha/ffitarget.h,
924         src/alpha/ffi.c,
925         src/alpha/osf.S,
926         src/m68k/ffitarget.h,
927         src/frv/ffitarget.h,
928         src/frv/ffi.c,
929         src/s390/ffitarget.h,
930         src/s390/sysv.S,
931         src/cris/ffitarget.h,
932         src/pa/linux.S,
933         src/pa/ffitarget.h,
934         src/pa/ffi.c,
935         src/raw_api.c,
936         src/ia64/ffitarget.h,
937         src/ia64/unix.S,
938         src/ia64/ffi.c,
939         src/ia64/ia64_flags.h,
940         src/java_raw_api.c,
941         src/debug.c,
942         src/sparc/v9.S,
943         src/sparc/ffitarget.h,
944         src/sparc/ffi.c,
945         src/sparc/v8.S,
946         src/mips/ffitarget.h,
947         src/mips/n32.S,
948         src/mips/o32.S,
949         src/mips/ffi.c,
950         src/prep_cif.c,
951         src/sh/ffitarget.h,
952         src/sh/ffi.c,
953         src/sh/sysv.S: Update license text.
954
955 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
956
957         * src/x86/win32.S (_ffi_closure_STDCALL):  New function.
958         (.eh_frame):  Add FDE for it.
959
960 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
961
962         * configure.ac:  Also check if assembler supports pc-relative
963         relocs on X86_WIN32 targets.
964         * configure:  Regenerate.
965         * src/x86/win32.S (ffi_prep_args):  Declare extern, not global.
966         (_ffi_call_SYSV):  Add missing function type symbol .def and
967         add EH markup labels.
968         (_ffi_call_STDCALL):  Likewise.
969         (_ffi_closure_SYSV):  Likewise.
970         (_ffi_closure_raw_SYSV):  Likewise.
971         (.eh_frame):  Add hand-crafted EH data.
972
973 2009-04-09  Jakub Jelinek  <jakub@redhat.com>
974
975         * testsuite/lib/libffi-dg.exp: Change copyright header to refer to
976         version 3 of the GNU General Public License and to point readers
977         at the COPYING3 file and the FSF's license web page.
978         * testsuite/libffi.call/call.exp: Likewise.
979         * testsuite/libffi.special/special.exp: Likewise.
980
981 2009-03-01  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
982
983         * configure: Regenerate.
984
985 2008-12-18  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
986
987         PR libffi/26048
988         * configure.ac (HAVE_AS_X86_PCREL): New test.
989         * configure: Regenerate.
990         * fficonfig.h.in: Regenerate.
991         * src/x86/sysv.S [!FFI_NO_RAW_API]: Precalculate
992         RAW_CLOSURE_CIF_OFFSET, RAW_CLOSURE_FUN_OFFSET,
993         RAW_CLOSURE_USER_DATA_OFFSET for the Solaris 10/x86 assembler.
994         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
995         * src/x86/unix64.S (.Lstore_table): Move to .text section.
996         (.Lload_table): Likewise.
997         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
998
999 2008-12-18  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1000
1001         * configure: Regenerate.
1002
1003 2008-11-21  Eric Botcazou  <ebotcazou@adacore.com>
1004
1005         * src/sparc/ffi.c (ffi_prep_cif_machdep): Add support for
1006         signed/unsigned int8/16 return values.
1007         * src/sparc/v8.S (ffi_call_v8): Likewise.
1008         (ffi_closure_v8): Likewise.
1009
1010 2008-09-26  Peter O'Gorman  <pogma@thewrittenword.com>
1011             Steve Ellcey  <sje@cup.hp.com>
1012
1013         * configure: Regenerate for new libtool.
1014         * Makefile.in: Ditto.
1015         * include/Makefile.in: Ditto.
1016         * aclocal.m4: Ditto.
1017
1018 2008-08-25  Andreas Tobler  <a.tobler@schweiz.org>
1019
1020         * src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and
1021         FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum.
1022         Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT.
1023         Adjust copyright notice.
1024         * src/powerpc/ffi.c: Add two new flags to indicate if we have one
1025         register or two register to use for FFI_SYSV structs.
1026         (ffi_prep_cif_machdep): Pass the right register flag introduced above.
1027         (ffi_closure_helper_SYSV): Fix the return type for
1028         FFI_SYSV_TYPE_SMALL_STRUCT. Comment.
1029         Adjust copyright notice.
1030
1031 2008-07-16  Kaz Kojima  <kkojima@gcc.gnu.org>
1032
1033         * src/sh/ffi.c (ffi_prep_closure_loc): Turn INSN into an unsigned
1034         int.
1035
1036 2008-06-17  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1037
1038         * configure: Regenerate.
1039         * include/Makefile.in: Regenerate.
1040         * testsuite/Makefile.in: Regenerate.
1041
1042 2008-06-07  Joseph Myers  <joseph@codesourcery.com>
1043
1044         * configure.ac (parisc*-*-linux*, powerpc-*-sysv*,
1045         powerpc-*-beos*): Remove.
1046         * configure: Regenerate.
1047
1048 2008-05-09  Julian Brown  <julian@codesourcery.com>
1049
1050         * Makefile.am (LTLDFLAGS): New.
1051         (libffi_la_LDFLAGS): Use above.
1052         * Makefile.in: Regenerate.
1053
1054 2008-04-18  Paolo Bonzini  <bonzini@gnu.org>
1055
1056         PR bootstrap/35457
1057         * aclocal.m4: Regenerate.
1058         * configure: Regenerate.
1059
1060 2008-03-26  Kaz Kojima  <kkojima@gcc.gnu.org>
1061
1062         * src/sh/sysv.S: Add .note.GNU-stack on Linux.
1063         * src/sh64/sysv.S: Likewise.
1064
1065 2008-03-26  Daniel Jacobowitz  <dan@debian.org>
1066
1067         * src/arm/sysv.S: Fix ARM comment marker.
1068
1069 2008-03-26  Jakub Jelinek  <jakub@redhat.com>
1070
1071         * src/alpha/osf.S: Add .note.GNU-stack on Linux.
1072         * src/s390/sysv.S: Likewise.
1073         * src/powerpc/ppc_closure.S: Likewise.
1074         * src/powerpc/sysv.S: Likewise.
1075         * src/x86/unix64.S: Likewise.
1076         * src/x86/sysv.S: Likewise.
1077         * src/sparc/v8.S: Likewise.
1078         * src/sparc/v9.S: Likewise.
1079         * src/m68k/sysv.S: Likewise.
1080         * src/arm/sysv.S: Likewise.
1081
1082 2008-03-16  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1083
1084         * aclocal.m4: Regenerate.
1085         * configure: Likewise.
1086         * Makefile.in: Likewise.
1087         * include/Makefile.in: Likewise.
1088         * testsuite/Makefile.in: Likewise.
1089
1090 2008-02-12  Bjoern Koenig  <bkoenig@alpha-tierchen.de>
1091             Andreas Tobler  <a.tobler@schweiz.org>
1092
1093         * configure.ac: Add amd64-*-freebsd* target.
1094         * configure: Regenerate.
1095
1096 2008-01-30  H.J. Lu  <hongjiu.lu@intel.com>
1097
1098         PR libffi/34612
1099         * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when
1100         returning struct.
1101
1102         * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer"
1103         tests.
1104
1105 2008-01-24  David Edelsohn  <edelsohn@gnu.org>
1106
1107         * configure: Regenerate.
1108
1109 2008-01-06  Andreas Tobler  <a.tobler@schweiz.org>
1110
1111         * src/x86/ffi.c (ffi_prep_cif_machdep): Fix thinko.
1112
1113 2008-01-05  Andreas Tobler  <a.tobler@schweiz.org>
1114
1115         PR testsuite/32843
1116         * src/x86/ffi.c (ffi_prep_cif_machdep): Add code for
1117         signed/unsigned int8/16 for X86_DARWIN.
1118         Updated copyright info.
1119         Handle one and two byte structs with special cif->flags.
1120         * src/x86/ffitarget.h: Add special types for one and two byte structs.
1121         Updated copyright info.
1122         * src/x86/darwin.S (ffi_call_SYSV): Rewrite to use a jump table like
1123         sysv.S
1124         Remove code to pop args from the stack after call.
1125         Special-case signed/unsigned for int8/16, one and two byte structs.
1126         (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1127         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1128         FFI_TYPE_SINT32.
1129         Updated copyright info.
1130
1131 2007-12-08  David Daney  <ddaney@avtrex.com>
1132
1133         * src/mips/n32.S (ffi_call_N32):  Replace dadd with ADDU, dsub with
1134         SUBU, add with ADDU and use smaller code sequences.
1135
1136 2007-12-07  David Daney  <ddaney@avtrex.com>
1137
1138         * src/mips/ffi.c (ffi_prep_cif_machdep): Handle long double return
1139         type.
1140
1141 2007-12-06  David Daney  <ddaney@avtrex.com>
1142
1143         * include/ffi.h.in (FFI_SIZEOF_JAVA_RAW): Define if not already
1144         defined.
1145         (ffi_java_raw): New typedef.
1146         (ffi_java_raw_call, ffi_java_ptrarray_to_raw,
1147         ffi_java_raw_to_ptrarray): Change parameter types from ffi_raw to
1148         ffi_java_raw.
1149         (ffi_java_raw_closure) : Same.
1150         (ffi_prep_java_raw_closure, ffi_prep_java_raw_closure_loc): Change
1151         parameter types.
1152         * src/java_raw_api.c (ffi_java_raw_size):  Replace FFI_SIZEOF_ARG with
1153         FFI_SIZEOF_JAVA_RAW.
1154         (ffi_java_raw_to_ptrarray): Change type of raw to ffi_java_raw.
1155         Replace FFI_SIZEOF_ARG with FFI_SIZEOF_JAVA_RAW. Use
1156         sizeof(ffi_java_raw) for alignment calculations.
1157         (ffi_java_ptrarray_to_raw): Same.
1158         (ffi_java_rvalue_to_raw): Add special handling for FFI_TYPE_POINTER
1159         if FFI_SIZEOF_JAVA_RAW == 4.
1160         (ffi_java_raw_to_rvalue): Same.
1161         (ffi_java_raw_call): Change type of raw to ffi_java_raw.
1162         (ffi_java_translate_args): Same.
1163         (ffi_prep_java_raw_closure_loc, ffi_prep_java_raw_closure): Change
1164         parameter types.
1165         * src/mips/ffitarget.h (FFI_SIZEOF_JAVA_RAW): Define for N32 ABI.
1166
1167 2007-12-06  David Daney  <ddaney@avtrex.com>
1168
1169         * src/mips/n32.S (ffi_closure_N32): Use 64-bit add instruction on
1170         pointer values.
1171
1172 2007-12-01  Andreas Tobler  <a.tobler@schweiz.org>
1173
1174         PR libffi/31937
1175         * src/powerpc/ffitarget.h: Introduce new ABI FFI_LINUX_SOFT_FLOAT.
1176         Add local FFI_TYPE_UINT128 to handle soft-float long-double-128.
1177         * src/powerpc/ffi.c: Distinguish between __NO_FPRS__ and not and
1178         set the NUM_FPR_ARG_REGISTERS according to.
1179         Add support for potential soft-float support under hard-float
1180         architecture.
1181         (ffi_prep_args_SYSV): Set NUM_FPR_ARG_REGISTERS to 0 in case of
1182         FFI_LINUX_SOFT_FLOAT, handle float, doubles and long-doubles according
1183         to the FFI_LINUX_SOFT_FLOAT ABI.
1184         (ffi_prep_cif_machdep): Likewise.
1185         (ffi_closure_helper_SYSV): Likewise.
1186         * src/powerpc/ppc_closure.S: Make sure not to store float/double
1187         on archs where __NO_FPRS__ is true.
1188         Add FFI_TYPE_UINT128 support.
1189         * src/powerpc/sysv.S: Add support for soft-float long-double-128.
1190         Adjust copyright notice.
1191
1192 2007-11-25  Andreas Tobler  <a.tobler@schweiz.org>
1193
1194         * src/closures.c: Move defintion of MAYBE_UNUSED from here to ...
1195         * include/ffi_common.h: ... here.
1196         Update copyright.
1197
1198 2007-11-17  Andreas Tobler  <a.tobler@schweiz.org>
1199
1200         * src/powerpc/sysv.S: Load correct cr to compare if we have long double.
1201         * src/powerpc/linux64.S: Likewise.
1202         * src/powerpc/ffi.c: Add a comment to show which part goes into cr6.
1203         * testsuite/libffi.call/return_ldl.c: New test.
1204
1205 2007-09-04    <aph@redhat.com>
1206
1207         * src/arm/sysv.S (UNWIND): New.
1208         (Whole file): Conditionally compile unwinder directives.
1209         * src/arm/sysv.S: Add unwinder directives.
1210
1211         * src/arm/ffi.c (ffi_prep_args): Align structs by at least 4 bytes.
1212         Only treat r0 as a struct address if we're actually returning a
1213         struct by address.
1214         Only copy the bytes that are actually within a struct.
1215         (ffi_prep_cif_machdep): A Composite Type not larger than 4 bytes
1216         is returned in r0, not passed by address.
1217         (ffi_call): Allocate a word-sized temporary for the case where
1218         a composite is returned in r0.
1219         (ffi_prep_incoming_args_SYSV): Align as necessary.
1220
1221 2007-08-05  Steven Newbury  <s_j_newbury@yahoo.co.uk>
1222
1223         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Use __clear_cache instead of
1224         directly using the sys_cacheflush syscall.
1225
1226 2007-07-27  Andrew Haley  <aph@redhat.com>
1227
1228         * src/arm/sysv.S (ffi_closure_SYSV): Add soft-float.
1229
1230 2007-09-03  Maciej W. Rozycki  <macro@linux-mips.org>
1231
1232         * Makefile.am: Unify MIPS_IRIX and MIPS_LINUX into MIPS.
1233         * configure.ac: Likewise.
1234         * Makefile.in: Regenerate.
1235         * include/Makefile.in: Likewise.
1236         * testsuite/Makefile.in: Likewise.
1237         * configure: Likewise.
1238
1239 2007-08-24  David Daney  <ddaney@avtrex.com>
1240
1241         * testsuite/libffi.call/return_sl.c: New test.
1242
1243 2007-08-10  David Daney  <ddaney@avtrex.com>
1244
1245         * testsuite/libffi.call/cls_multi_ushort.c,
1246         testsuite/libffi.call/cls_align_uint16.c,
1247         testsuite/libffi.call/nested_struct1.c,
1248         testsuite/libffi.call/nested_struct3.c,
1249         testsuite/libffi.call/cls_7_1_byte.c,
1250         testsuite/libffi.call/nested_struct5.c,
1251         testsuite/libffi.call/cls_double.c,
1252         testsuite/libffi.call/nested_struct7.c,
1253         testsuite/libffi.call/cls_sint.c,
1254         testsuite/libffi.call/nested_struct9.c,
1255         testsuite/libffi.call/cls_20byte1.c,
1256         testsuite/libffi.call/cls_multi_sshortchar.c,
1257         testsuite/libffi.call/cls_align_sint64.c,
1258         testsuite/libffi.call/cls_3byte2.c,
1259         testsuite/libffi.call/cls_multi_schar.c,
1260         testsuite/libffi.call/cls_multi_uchar.c,
1261         testsuite/libffi.call/cls_19byte.c,
1262         testsuite/libffi.call/cls_9byte1.c,
1263         testsuite/libffi.call/cls_align_float.c,
1264         testsuite/libffi.call/closure_fn1.c,
1265         testsuite/libffi.call/problem1.c,
1266         testsuite/libffi.call/closure_fn3.c,
1267         testsuite/libffi.call/cls_sshort.c,
1268         testsuite/libffi.call/closure_fn5.c,
1269         testsuite/libffi.call/cls_align_double.c,
1270         testsuite/libffi.call/nested_struct.c,
1271         testsuite/libffi.call/cls_2byte.c,
1272         testsuite/libffi.call/nested_struct10.c,
1273         testsuite/libffi.call/cls_4byte.c,
1274         testsuite/libffi.call/cls_6byte.c,
1275         testsuite/libffi.call/cls_8byte.c,
1276         testsuite/libffi.call/cls_multi_sshort.c,
1277         testsuite/libffi.call/cls_align_sint16.c,
1278         testsuite/libffi.call/cls_align_uint32.c,
1279         testsuite/libffi.call/cls_20byte.c,
1280         testsuite/libffi.call/cls_float.c,
1281         testsuite/libffi.call/nested_struct2.c,
1282         testsuite/libffi.call/cls_5_1_byte.c,
1283         testsuite/libffi.call/nested_struct4.c,
1284         testsuite/libffi.call/cls_24byte.c,
1285         testsuite/libffi.call/nested_struct6.c,
1286         testsuite/libffi.call/cls_64byte.c,
1287         testsuite/libffi.call/nested_struct8.c,
1288         testsuite/libffi.call/cls_uint.c,
1289         testsuite/libffi.call/cls_multi_ushortchar.c,
1290         testsuite/libffi.call/cls_schar.c,
1291         testsuite/libffi.call/cls_uchar.c,
1292         testsuite/libffi.call/cls_align_uint64.c,
1293         testsuite/libffi.call/cls_ulonglong.c,
1294         testsuite/libffi.call/cls_align_longdouble.c,
1295         testsuite/libffi.call/cls_1_1byte.c,
1296         testsuite/libffi.call/cls_12byte.c,
1297         testsuite/libffi.call/cls_3_1byte.c,
1298         testsuite/libffi.call/cls_3byte1.c,
1299         testsuite/libffi.call/cls_4_1byte.c,
1300         testsuite/libffi.call/cls_6_1_byte.c,
1301         testsuite/libffi.call/cls_16byte.c,
1302         testsuite/libffi.call/cls_18byte.c,
1303         testsuite/libffi.call/closure_fn0.c,
1304         testsuite/libffi.call/cls_9byte2.c,
1305         testsuite/libffi.call/closure_fn2.c,
1306         testsuite/libffi.call/closure_fn4.c,
1307         testsuite/libffi.call/cls_ushort.c,
1308         testsuite/libffi.call/closure_fn6.c,
1309         testsuite/libffi.call/cls_5byte.c,
1310         testsuite/libffi.call/cls_align_pointer.c,
1311         testsuite/libffi.call/cls_7byte.c,
1312         testsuite/libffi.call/cls_align_sint32.c,
1313         testsuite/libffi.special/unwindtest_ffi_call.cc,
1314         testsuite/libffi.special/unwindtest.cc: Remove xfail for mips64*-*-*.
1315
1316 2007-08-10  David Daney  <ddaney@avtrex.com>
1317
1318         PR libffi/28313
1319         * configure.ac: Don't treat mips64 as a special case.
1320         * Makefile.am (nodist_libffi_la_SOURCES): Add n32.S.
1321         * configure: Regenerate
1322         * Makefile.in: Ditto.
1323         * fficonfig.h.in: Ditto.
1324         * src/mips/ffitarget.h (REG_L, REG_S, SUBU, ADDU, SRL, LI): Indent.
1325         (LA, EH_FRAME_ALIGN, FDE_ADDR_BYTES): New preprocessor macros.
1326         (FFI_DEFAULT_ABI): Set for n64 case.
1327         (FFI_CLOSURES, FFI_TRAMPOLINE_SIZE): Define for n32 and n64 cases.
1328         * src/mips/n32.S (ffi_call_N32): Add debug macros and labels for FDE.
1329         (ffi_closure_N32): New function.
1330         (.eh_frame): New section
1331         * src/mips/o32.S: Clean up comments.
1332         (ffi_closure_O32): Pass ffi_closure parameter in $12.
1333         * src/mips/ffi.c: Use FFI_MIPS_N32 instead of
1334         _MIPS_SIM == _ABIN32 throughout.
1335         (FFI_MIPS_STOP_HERE): New, use in place of
1336         ffi_stop_here.
1337         (ffi_prep_args): Use unsigned long to hold pointer values.  Rewrite
1338         to support n32/n64 ABIs.
1339         (calc_n32_struct_flags): Rewrite.
1340         (calc_n32_return_struct_flags): Remove unused variable.  Reverse
1341         position of flag bits.
1342         (ffi_prep_cif_machdep): Rewrite n32 portion.
1343         (ffi_call): Enable for n64.  Add special handling for small structure
1344         return values.
1345         (ffi_prep_closure_loc): Add n32 and n64 support.
1346         (ffi_closure_mips_inner_O32): Add cast to silence warning.
1347         (copy_struct_N32, ffi_closure_mips_inner_N32): New functions.
1348
1349 2007-08-08  David Daney  <ddaney@avtrex.com>
1350
1351         * testsuite/libffi.call/ffitest.h (ffi_type_mylong): Remove definition.
1352         * testsuite/libffi.call/cls_align_uint16.c (main): Use correct type
1353         specifiers.
1354         * testsuite/libffi.call/nested_struct1.c (main): Ditto.
1355         * testsuite/libffi.call/cls_sint.c (main): Ditto.
1356         * testsuite/libffi.call/nested_struct9.c (main): Ditto.
1357         * testsuite/libffi.call/cls_20byte1.c (main): Ditto.
1358         * testsuite/libffi.call/cls_9byte1.c (main): Ditto.
1359         * testsuite/libffi.call/closure_fn1.c (main): Ditto.
1360         * testsuite/libffi.call/closure_fn3.c (main): Ditto.
1361         * testsuite/libffi.call/return_dbl2.c (main): Ditto.
1362         * testsuite/libffi.call/cls_sshort.c (main): Ditto.
1363         * testsuite/libffi.call/return_fl3.c (main): Ditto.
1364         * testsuite/libffi.call/closure_fn5.c (main): Ditto.
1365         * testsuite/libffi.call/nested_struct.c (main): Ditto.
1366         * testsuite/libffi.call/nested_struct10.c (main): Ditto.
1367         * testsuite/libffi.call/return_ll1.c (main): Ditto.
1368         * testsuite/libffi.call/cls_8byte.c (main): Ditto.
1369         * testsuite/libffi.call/cls_align_uint32.c (main): Ditto.
1370         * testsuite/libffi.call/cls_align_sint16.c (main): Ditto.
1371         * testsuite/libffi.call/cls_20byte.c (main): Ditto.
1372         * testsuite/libffi.call/nested_struct2.c (main): Ditto.
1373         * testsuite/libffi.call/cls_24byte.c (main): Ditto.
1374         * testsuite/libffi.call/nested_struct6.c (main): Ditto.
1375         * testsuite/libffi.call/cls_uint.c (main): Ditto.
1376         * testsuite/libffi.call/cls_12byte.c (main): Ditto.
1377         * testsuite/libffi.call/cls_16byte.c (main): Ditto.
1378         * testsuite/libffi.call/closure_fn0.c (main): Ditto.
1379         * testsuite/libffi.call/cls_9byte2.c (main): Ditto.
1380         * testsuite/libffi.call/closure_fn2.c (main): Ditto.
1381         * testsuite/libffi.call/return_dbl1.c (main): Ditto.
1382         * testsuite/libffi.call/closure_fn4.c (main): Ditto.
1383         * testsuite/libffi.call/closure_fn6.c (main): Ditto.
1384         * testsuite/libffi.call/cls_align_sint32.c (main): Ditto.
1385
1386 2007-08-07  Andrew Haley  <aph@redhat.com>
1387
1388         * src/x86/sysv.S (ffi_closure_raw_SYSV): Fix typo in previous
1389         checkin.
1390
1391 2007-08-06  Andrew Haley  <aph@redhat.com>
1392
1393         PR testsuite/32843
1394         * src/x86/sysv.S (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1395         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1396         FFI_TYPE_SINT32.
1397
1398 2007-08-02  David Daney  <ddaney@avtrex.com>
1399
1400         * testsuite/libffi.call/return_ul.c (main): Define return type as
1401         ffi_arg.  Use proper printf conversion specifier.
1402
1403 2007-07-30  Andrew Haley  <aph@redhat.com>
1404
1405         PR testsuite/32843
1406         * src/x86/ffi.c (ffi_prep_cif_machdep): in x86 case, add code for
1407         signed/unsigned int8/16.
1408         * src/x86/sysv.S (ffi_call_SYSV): Rewrite to:
1409         Use a jump table.
1410         Remove code to pop args from the stack after call.
1411         Special-case signed/unsigned int8/16.
1412         * testsuite/libffi.call/return_sc.c (main): Revert.
1413
1414 2007-07-26  Richard Guenther  <rguenther@suse.de>
1415
1416         PR testsuite/32843
1417         * testsuite/libffi.call/return_sc.c (main): Verify call
1418         result as signed char, not ffi_arg.
1419
1420 2007-07-16  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1421
1422         * configure.ac (i?86-*-solaris2.1[0-9]): Set TARGET to X86_64.
1423         * configure: Regenerate.
1424
1425 2007-07-11  David Daney  <ddaney@avtrex.com>
1426
1427         * src/mips/ffi.c: Don't include sys/cachectl.h.
1428         (ffi_prep_closure_loc): Use __builtin___clear_cache() instead of
1429         cacheflush().
1430
1431 2007-05-18  Aurelien Jarno  <aurelien@aurel32.net>
1432
1433         * src/arm/ffi.c (ffi_prep_closure_loc): Renamed and ajusted
1434         from (ffi_prep_closure): ... this.
1435         (FFI_INIT_TRAMPOLINE): Adjust.
1436
1437 2005-12-31  Phil Blundell  <pb@reciva.com>
1438
1439         * src/arm/ffi.c (ffi_prep_incoming_args_SYSV,
1440         ffi_closure_SYSV_inner, ffi_prep_closure): New, add closure support.
1441         * src/arm/sysv.S(ffi_closure_SYSV): Likewise.
1442         * src/arm/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1443         (FFI_CLOSURES): Enable closure support.
1444
1445 2007-07-03  Andrew Haley  <aph@hedges.billgatliff.com>
1446
1447         * testsuite/libffi.call/cls_multi_ushort.c,
1448         testsuite/libffi.call/cls_align_uint16.c,
1449         testsuite/libffi.call/nested_struct1.c,
1450         testsuite/libffi.call/nested_struct3.c,
1451         testsuite/libffi.call/cls_7_1_byte.c,
1452         testsuite/libffi.call/cls_double.c,
1453         testsuite/libffi.call/nested_struct5.c,
1454         testsuite/libffi.call/nested_struct7.c,
1455         testsuite/libffi.call/cls_sint.c,
1456         testsuite/libffi.call/nested_struct9.c,
1457         testsuite/libffi.call/cls_20byte1.c,
1458         testsuite/libffi.call/cls_multi_sshortchar.c,
1459         testsuite/libffi.call/cls_align_sint64.c,
1460         testsuite/libffi.call/cls_3byte2.c,
1461         testsuite/libffi.call/cls_multi_schar.c,
1462         testsuite/libffi.call/cls_multi_uchar.c,
1463         testsuite/libffi.call/cls_19byte.c,
1464         testsuite/libffi.call/cls_9byte1.c,
1465         testsuite/libffi.call/cls_align_float.c,
1466         testsuite/libffi.call/closure_fn1.c,
1467         testsuite/libffi.call/problem1.c,
1468         testsuite/libffi.call/closure_fn3.c,
1469         testsuite/libffi.call/cls_sshort.c,
1470         testsuite/libffi.call/closure_fn5.c,
1471         testsuite/libffi.call/cls_align_double.c,
1472         testsuite/libffi.call/cls_2byte.c,
1473         testsuite/libffi.call/nested_struct.c,
1474         testsuite/libffi.call/nested_struct10.c,
1475         testsuite/libffi.call/cls_4byte.c,
1476         testsuite/libffi.call/cls_6byte.c,
1477         testsuite/libffi.call/cls_8byte.c,
1478         testsuite/libffi.call/cls_multi_sshort.c,
1479         testsuite/libffi.call/cls_align_uint32.c,
1480         testsuite/libffi.call/cls_align_sint16.c,
1481         testsuite/libffi.call/cls_float.c,
1482         testsuite/libffi.call/cls_20byte.c,
1483         testsuite/libffi.call/cls_5_1_byte.c,
1484         testsuite/libffi.call/nested_struct2.c,
1485         testsuite/libffi.call/cls_24byte.c,
1486         testsuite/libffi.call/nested_struct4.c,
1487         testsuite/libffi.call/nested_struct6.c,
1488         testsuite/libffi.call/cls_64byte.c,
1489         testsuite/libffi.call/nested_struct8.c,
1490         testsuite/libffi.call/cls_uint.c,
1491         testsuite/libffi.call/cls_multi_ushortchar.c,
1492         testsuite/libffi.call/cls_schar.c,
1493         testsuite/libffi.call/cls_uchar.c,
1494         testsuite/libffi.call/cls_align_uint64.c,
1495         testsuite/libffi.call/cls_ulonglong.c,
1496         testsuite/libffi.call/cls_align_longdouble.c,
1497         testsuite/libffi.call/cls_1_1byte.c,
1498         testsuite/libffi.call/cls_12byte.c,
1499         testsuite/libffi.call/cls_3_1byte.c,
1500         testsuite/libffi.call/cls_3byte1.c,
1501         testsuite/libffi.call/cls_4_1byte.c,
1502         testsuite/libffi.call/cls_6_1_byte.c,
1503         testsuite/libffi.call/cls_16byte.c,
1504         testsuite/libffi.call/cls_18byte.c,
1505         testsuite/libffi.call/closure_fn0.c,
1506         testsuite/libffi.call/cls_9byte2.c,
1507         testsuite/libffi.call/closure_fn2.c,
1508         testsuite/libffi.call/closure_fn4.c,
1509         testsuite/libffi.call/cls_ushort.c,
1510         testsuite/libffi.call/closure_fn6.c,
1511         testsuite/libffi.call/cls_5byte.c,
1512         testsuite/libffi.call/cls_align_pointer.c,
1513         testsuite/libffi.call/cls_7byte.c,
1514         testsuite/libffi.call/cls_align_sint32.c,
1515         testsuite/libffi.special/unwindtest_ffi_call.cc,
1516         testsuite/libffi.special/unwindtest.cc: Enable for ARM.
1517
1518 2007-07-05  H.J. Lu  <hongjiu.lu@intel.com>
1519
1520         * aclocal.m4: Regenerated.
1521
1522 2007-06-02  Paolo Bonzini  <bonzini@gnu.org>
1523
1524         * configure: Regenerate.
1525
1526 2007-05-23  Steve Ellcey  <sje@cup.hp.com>
1527
1528         * Makefile.in: Regenerate.
1529         * configure: Regenerate.
1530         * aclocal.m4: Regenerate.
1531         * include/Makefile.in: Regenerate.
1532         * testsuite/Makefile.in: Regenerate.
1533
1534 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1535
1536         * src/m68k/ffi.c (ffi_prep_incoming_args_SYSV,
1537         ffi_closure_SYSV_inner,ffi_prep_closure): New, add closure support.
1538         * src/m68k/sysv.S(ffi_closure_SYSV,ffi_closure_struct_SYSV): Likewise.
1539         * src/m68k/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1540         (FFI_CLOSURES): Enable closure support.
1541
1542 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1543
1544         * configure.ac (HAVE_AS_CFI_PSEUDO_OP): New test.
1545         * configure: Regenerate.
1546         * fficonfig.h.in: Regenerate.
1547         * src/m68k/sysv.S (CFI_STARTPROC,CFI_ENDPROC,
1548         CFI_OFFSET,CFI_DEF_CFA): New macros.
1549         (ffi_call_SYSV): Add callframe annotation.
1550
1551 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1552
1553         * src/m68k/ffi.c (ffi_prep_args,ffi_prep_cif_machdep): Fix
1554         numerous test suite failures.
1555         * src/m68k/sysv.S (ffi_call_SYSV): Likewise.
1556
1557 2007-04-11  Paolo Bonzini  <bonzini@gnu.org>
1558
1559         * Makefile.am (EXTRA_DIST): Bring up to date.
1560         * Makefile.in: Regenerate.
1561         * src/frv/eabi.S: Remove RCS keyword.
1562
1563 2007-04-06  Richard Henderson  <rth@redhat.com>
1564
1565         * configure.ac: Tidy target case.
1566         (HAVE_LONG_DOUBLE): Allow the target to override.
1567         * configure: Regenerate.
1568         * include/ffi.h.in: Don't define ffi_type_foo if
1569         LIBFFI_HIDE_BASIC_TYPES is defined.
1570         (ffi_type_longdouble): If not HAVE_LONG_DOUBLE, define
1571         to ffi_type_double.
1572         * types.c (LIBFFI_HIDE_BASIC_TYPES): Define.
1573         (FFI_TYPEDEF, ffi_type_void): Mark the data const.
1574         (ffi_type_longdouble): Special case for Alpha.  Don't define
1575         if long double == double.
1576
1577         * src/alpha/ffi.c (FFI_TYPE_LONGDOUBLE): Assert unique value.
1578         (ffi_prep_cif_machdep): Handle it as the 128-bit type.
1579         (ffi_call, ffi_closure_osf_inner): Likewise.
1580         (ffi_closure_osf_inner): Likewise.  Mark hidden.
1581         (ffi_call_osf, ffi_closure_osf): Mark hidden.
1582         * src/alpha/ffitarget.h (FFI_LAST_ABI): Tidy definition.
1583         * src/alpha/osf.S (ffi_call_osf, ffi_closure_osf): Mark hidden.
1584         (load_table): Handle 128-bit long double.
1585
1586         * testsuite/libffi.call/float4.c: Add -mieee for alpha.
1587
1588 2007-04-06  Tom Tromey  <tromey@redhat.com>
1589
1590         PR libffi/31491:
1591         * README: Fixed bug in example.
1592
1593 2007-04-03  Jakub Jelinek  <jakub@redhat.com>
1594
1595         * src/closures.c: Include sys/statfs.h.
1596         (_GNU_SOURCE): Define on Linux.
1597         (FFI_MMAP_EXEC_SELINUX): Define.
1598         (selinux_enabled): New variable.
1599         (selinux_enabled_check): New function.
1600         (is_selinux_enabled): Define.
1601         (dlmmap): Use it.
1602
1603 2007-03-24  Uros Bizjak  <ubizjak@gmail.com>
1604
1605         * testsuite/libffi.call/return_fl2.c (return_fl): Mark as static.
1606         Use 'volatile float sum' to create sum of floats to avoid false
1607         negative due to excess precision on ix86 targets.
1608         (main): Ditto.
1609
1610 2007-03-08  Alexandre Oliva  <aoliva@redhat.com>
1611
1612         * src/powerpc/ffi.c (flush_icache): Fix left-over from previous
1613         patch.
1614         (ffi_prep_closure_loc): Remove unneeded casts.  Add needed ones.
1615
1616 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1617
1618         * include/ffi.h.in (ffi_closure_alloc, ffi_closure_free): New.
1619         (ffi_prep_closure_loc): New.
1620         (ffi_prep_raw_closure_loc): New.
1621         (ffi_prep_java_raw_closure_loc): New.
1622         * src/closures.c: New file.
1623         * src/dlmalloc.c [FFI_MMAP_EXEC_WRIT] (struct malloc_segment):
1624         Replace sflags with exec_offset.
1625         [FFI_MMAP_EXEC_WRIT] (mmap_exec_offset, add_segment_exec_offset,
1626         sub_segment_exec_offset): New macros.
1627         (get_segment_flags, set_segment_flags, check_segment_merge): New
1628         macros.
1629         (is_mmapped_segment, is_extern_segment): Use get_segment_flags.
1630         (add_segment, sys_alloc, create_mspace, create_mspace_with_base,
1631         destroy_mspace): Use new macros.
1632         (sys_alloc): Silence warning.
1633         * Makefile.am (libffi_la_SOURCES): Add src/closures.c.
1634         * Makefile.in: Rebuilt.
1635         * src/prep_cif [FFI_CLOSURES] (ffi_prep_closure): Implement in
1636         terms of ffi_prep_closure_loc.
1637         * src/raw_api.c (ffi_prep_raw_closure_loc): Renamed and adjusted
1638         from...
1639         (ffi_prep_raw_closure): ... this.  Re-implement in terms of the
1640         renamed version.
1641         * src/java_raw_api (ffi_prep_java_raw_closure_loc): Renamed and
1642         adjusted from...
1643         (ffi_prep_java_raw_closure): ... this.  Re-implement in terms of
1644         the renamed version.
1645         * src/alpha/ffi.c (ffi_prep_closure_loc): Renamed from
1646         (ffi_prep_closure): ... this.
1647         * src/pa/ffi.c: Likewise.
1648         * src/cris/ffi.c: Likewise.  Adjust.
1649         * src/frv/ffi.c: Likewise.
1650         * src/ia64/ffi.c: Likewise.
1651         * src/mips/ffi.c: Likewise.
1652         * src/powerpc/ffi_darwin.c: Likewise.
1653         * src/s390/ffi.c: Likewise.
1654         * src/sh/ffi.c: Likewise.
1655         * src/sh64/ffi.c: Likewise.
1656         * src/sparc/ffi.c: Likewise.
1657         * src/x86/ffi64.c: Likewise.
1658         * src/x86/ffi.c: Likewise.
1659         (FFI_INIT_TRAMPOLINE): Adjust.
1660         (ffi_prep_raw_closure_loc): Renamed and adjusted from...
1661         (ffi_prep_raw_closure): ... this.
1662         * src/powerpc/ffi.c (ffi_prep_closure_loc): Renamed from
1663         (ffi_prep_closure): ... this.
1664         (flush_icache): Adjust.
1665
1666 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1667
1668         * src/dlmalloc.c: New file, imported version 2.8.3 of Doug
1669         Lea's malloc.
1670
1671 2007-03-01  Brooks Moses  <brooks.moses@codesourcery.com>
1672
1673         * Makefile.am: Add dummy install-pdf target.
1674         * Makefile.in: Regenerate
1675
1676 2007-02-13  Andreas Krebbel  <krebbel1@de.ibm.com>
1677
1678         * src/s390/ffi.c (ffi_prep_args, ffi_prep_cif_machdep,
1679         ffi_closure_helper_SYSV): Add long double handling.
1680
1681 2007-02-02  Jakub Jelinek  <jakub@redhat.com>
1682
1683         * src/powerpc/linux64.S (ffi_call_LINUX64): Move restore of r2
1684         immediately after bctrl instruction.
1685
1686 2007-01-18  Alexandre Oliva  <aoliva@redhat.com>
1687
1688         * Makefile.am (all-recursive, install-recursive,
1689         mostlyclean-recursive, clean-recursive, distclean-recursive,
1690         maintainer-clean-recursive): Add missing targets.
1691         * Makefile.in: Rebuilt.
1692
1693 2006-12-14  Andreas Tobler  <a.tobler@schweiz.org>
1694
1695         * configure.ac: Add TARGET for x86_64-*-darwin*.
1696         * Makefile.am (nodist_libffi_la_SOURCES): Add rules for 64-bit sources
1697         for X86_DARWIN.
1698         * src/x86/ffitarget.h: Set trampoline size for x86_64-*-darwin*.
1699         * src/x86/darwin64.S: New file for x86_64-*-darwin* support.
1700         * configure: Regenerate.
1701         * Makefile.in: Regenerate.
1702         * include/Makefile.in: Regenerate.
1703         * testsuite/Makefile.in: Regenerate.
1704         * testsuite/libffi.special/unwindtest_ffi_call.cc: New test case for
1705         ffi_call only.
1706
1707 2006-12-13  Andreas Tobler <a.tobler@schweiz.org>
1708
1709         * aclocal.m4: Regenerate with aclocal -I .. as written in the
1710         Makefile.am.
1711
1712 2006-10-31  Geoffrey Keating  <geoffk@apple.com>
1713
1714         * src/powerpc/ffi_darwin.c (darwin_adjust_aggregate_sizes): New.
1715         (ffi_prep_cif_machdep): Call darwin_adjust_aggregate_sizes for
1716         Darwin.
1717         * testsuite/libffi.call/nested_struct4.c: Remove Darwin XFAIL.
1718         * testsuite/libffi.call/nested_struct6.c: Remove Darwin XFAIL.
1719
1720 2006-10-10  Paolo Bonzini  <bonzini@gnu.org>
1721             Sandro Tolaini  <tolaini@libero.it>
1722
1723         * configure.ac [i*86-*-darwin*]: Set X86_DARWIN symbol and
1724         conditional.
1725         * configure: Regenerated.
1726         * Makefile.am (nodist_libffi_la_SOURCES) [X86_DARWIN]: New case.
1727         (EXTRA_DIST): Add src/x86/darwin.S.
1728         * Makefile.in: Regenerated.
1729         * include/Makefile.in: Regenerated.
1730         * testsuite/Makefile.in: Regenerated.
1731
1732         * src/x86/ffi.c (ffi_prep_cif_machdep) [X86_DARWIN]: Treat like
1733         X86_WIN32, and additionally align stack to 16 bytes.
1734         * src/x86/darwin.S: New, based on sysv.S.
1735         * src/prep_cif.c (ffi_prep_cif) [X86_DARWIN]: Align > 8-byte structs.
1736
1737 2006-09-12  David Daney  <ddaney@avtrex.com>
1738
1739         PR libffi/23935
1740         * include/Makefile.am: Install both ffi.h and ffitarget.h in
1741         $(libdir)/gcc/$(target_alias)/$(gcc_version)/include.
1742         * aclocal.m4: Regenerated for automake 1.9.6.
1743         * Makefile.in: Regenerated.
1744         * include/Makefile.in: Regenerated.
1745         * testsuite/Makefile.in: Regenerated.
1746
1747 2006-08-17  Andreas Tobler  <a.tobler@schweiz.ch>
1748
1749         * include/ffi_common.h (struct): Revert accidental commit.
1750
1751 2006-08-15  Andreas Tobler  <a.tobler@schweiz.ch>
1752
1753         * include/ffi_common.h: Remove lint directives.
1754         * include/ffi.h.in: Likewise.
1755
1756 2006-07-25  Torsten Schoenfeld  <kaffeetisch@gmx.de>
1757
1758         * include/ffi.h.in (ffi_type_ulong, ffi_type_slong): Define correctly
1759         for 32-bit architectures.
1760         * testsuite/libffi.call/return_ul.c: New test case.
1761
1762 2006-07-19  David Daney  <ddaney@avtrex.com>
1763
1764         * testsuite/libffi.call/closure_fn6.c: Remove xfail for mips,
1765         xfail remains for mips64.
1766
1767 2006-05-23  Carlos O'Donell  <carlos@codesourcery.com>
1768
1769         * Makefile.am: Add install-html target. Add install-html to .PHONY
1770         * Makefile.in: Regenerate.
1771         * aclocal.m4: Regenerate.
1772         * include/Makefile.in: Regenerate.
1773         * testsuite/Makefile.in: Regenerate.
1774
1775 2006-05-18  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1776
1777         * pa/ffi.c (ffi_prep_args_pa32): Load floating point arguments from
1778         stack slot.
1779
1780 2006-04-22  Andreas Tobler  <a.tobler@schweiz.ch>
1781
1782         * README: Remove notice about 'Crazy Comments'.
1783         * src/debug.c: Remove lint directives. Cleanup white spaces.
1784         * src/java_raw_api.c: Likewise.
1785         * src/prep_cif.c: Likewise.
1786         * src/raw_api.c: Likewise.
1787         * src/ffitest.c: Delete. No longer needed, all test cases migrated
1788         to the testsuite.
1789         * src/arm/ffi.c: Remove lint directives.
1790         * src/m32r/ffi.c: Likewise.
1791         * src/pa/ffi.c: Likewise.
1792         * src/powerpc/ffi.c: Likewise.
1793         * src/powerpc/ffi_darwin.c: Likewise.
1794         * src/sh/ffi.c: Likewise.
1795         * src/sh64/ffi.c: Likewise.
1796         * src/x86/ffi.c: Likewise.
1797         * testsuite/libffi.call/float2.c: Likewise.
1798         * testsuite/libffi.call/promotion.c: Likewise.
1799         * testsuite/libffi.call/struct1.c: Likewise.
1800
1801 2006-04-13  Andreas Tobler  <a.tobler@schweiz.ch>
1802
1803         * src/pa/hpux32.S: Correct unwind offset calculation for
1804         ffi_closure_pa32.
1805         * src/pa/linux.S: Likewise.
1806
1807 2006-04-12  James E Wilson  <wilson@specifix.com>
1808
1809         PR libgcj/26483
1810         * src/ia64/ffi.c (stf_spill, ldf_fill): Rewrite as macros.
1811         (hfa_type_load): Call stf_spill.
1812         (hfa_type_store): Call ldf_fill.
1813         (ffi_call): Adjust calls to above routines.  Add local temps for
1814         macro result.
1815
1816 2006-04-10  Matthias Klose  <doko@debian.org>
1817
1818         * testsuite/lib/libffi-dg.exp (libffi-init): Recognize multilib
1819         directory names containing underscores.
1820
1821 2006-04-07  James E Wilson  <wilson@specifix.com>
1822
1823         * testsuite/libffi.call/float4.c: New testcase.
1824
1825 2006-04-05  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1826             Andreas Tobler  <a.tobler@schweiz.ch>
1827
1828         * Makefile.am: Add PA_HPUX port.
1829         * Makefile.in: Regenerate.
1830         * include/Makefile.in: Likewise.
1831         * testsuite/Makefile.in: Likewise.
1832         * configure.ac: Add PA_HPUX rules.
1833         * configure: Regenerate.
1834         * src/pa/ffitarget.h: Rename linux target to PA_LINUX.
1835         Add PA_HPUX and PA64_HPUX.
1836         Rename FFI_LINUX ABI to FFI_PA32 ABI.
1837         (FFI_TRAMPOLINE_SIZE): Define for 32-bit HP-UX targets.
1838         (FFI_TYPE_SMALL_STRUCT2): Define.
1839         (FFI_TYPE_SMALL_STRUCT4): Likewise.
1840         (FFI_TYPE_SMALL_STRUCT8): Likewise.
1841         (FFI_TYPE_SMALL_STRUCT3): Redefine.
1842         (FFI_TYPE_SMALL_STRUCT5): Likewise.
1843         (FFI_TYPE_SMALL_STRUCT6): Likewise.
1844         (FFI_TYPE_SMALL_STRUCT7): Likewise.
1845         * src/pa/ffi.c (ROUND_DOWN): Delete.
1846         (fldw, fstw, fldd, fstd): Use '__asm__'.
1847         (ffi_struct_type): Add support for FFI_TYPE_SMALL_STRUCT2,
1848         FFI_TYPE_SMALL_STRUCT4 and FFI_TYPE_SMALL_STRUCT8.
1849         (ffi_prep_args_LINUX): Rename to ffi_prep_args_pa32. Update comment.
1850         Simplify incrementing of stack slot variable. Change type of local
1851         'n' to unsigned int.
1852         (ffi_size_stack_LINUX): Rename to ffi_size_stack_pa32. Handle long
1853         double on PA_HPUX.
1854         (ffi_prep_cif_machdep): Likewise.
1855         (ffi_call): Likewise.
1856         (ffi_closure_inner_LINUX): Rename to ffi_closure_inner_pa32. Change
1857         return type to ffi_status. Simplify incrementing of stack slot
1858         variable. Only copy floating point argument registers when PA_LINUX
1859         is true. Reformat debug statement.
1860         Add support for FFI_TYPE_SMALL_STRUCT2, FFI_TYPE_SMALL_STRUCT4 and
1861         FFI_TYPE_SMALL_STRUCT8.
1862         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Add 'extern' to
1863         declaration.
1864         (ffi_prep_closure): Make linux trampoline conditional on PA_LINUX.
1865         Add nops to cache flush.  Add trampoline for PA_HPUX.
1866         * src/pa/hpux32.S: New file.
1867         * src/pa/linux.S (ffi_call_LINUX): Rename to ffi_call_pa32. Rename
1868         ffi_prep_args_LINUX to ffi_prep_args_pa32.
1869         Localize labels. Add support for 2, 4 and 8-byte small structs. Handle
1870         unaligned destinations in 3, 5, 6 and 7-byte small structs. Order
1871         argument type checks so that common argument types appear first.
1872         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Rename
1873         ffi_closure_inner_LINUX to ffi_closure_inner_pa32.
1874
1875 2006-03-24  Alan Modra  <amodra@bigpond.net.au>
1876
1877         * src/powerpc/ffitarget.h (enum ffi_abi): Add FFI_LINUX.  Default
1878         for 32-bit using IBM extended double format.  Fix FFI_LAST_ABI.
1879         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Handle linux variant of
1880         FFI_TYPE_LONGDOUBLE.
1881         (ffi_prep_args64): Assert using IBM extended double.
1882         (ffi_prep_cif_machdep): Don't munge FFI_TYPE_LONGDOUBLE type.
1883         Handle FFI_LINUX FFI_TYPE_LONGDOUBLE return and args.
1884         (ffi_call): Handle FFI_LINUX.
1885         (ffi_closure_helper_SYSV): Non FFI_LINUX long double return needs
1886         gpr3 return pointer as for struct return.  Handle FFI_LINUX
1887         FFI_TYPE_LONGDOUBLE return and args.  Don't increment "nf"
1888         unnecessarily.
1889         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Load both f1 and f2
1890         for FFI_TYPE_LONGDOUBLE.  Move epilogue insns into case table.
1891         Don't use r6 as pointer to results, instead use sp offset.  Don't
1892         make a special call to load lr with case table address, instead
1893         use offset from previous call.
1894         * src/powerpc/sysv.S (ffi_call_SYSV): Save long double return.
1895         * src/powerpc/linux64.S (ffi_call_LINUX64): Simplify long double
1896         return.
1897
1898 2006-03-15  Kaz Kojima  <kkojima@gcc.gnu.org>
1899
1900         * src/sh64/ffi.c (ffi_prep_cif_machdep): Handle float arguments
1901         passed with FP registers correctly.
1902         (ffi_closure_helper_SYSV): Likewise.
1903         * src/sh64/sysv.S: Likewise.
1904
1905 2006-03-01  Andreas Tobler  <a.tobler@schweiz.ch>
1906
1907         * testsuite/libffi.special/unwindtest.cc (closure_test_fn): Mark cif,
1908         args and userdata unused.
1909         (closure_test_fn1): Mark cif and userdata unused.
1910         (main): Remove unused res.
1911
1912 2006-02-28  Andreas Tobler  <a.tobler@schweiz.ch>
1913
1914         * testsuite/libffi.call/call.exp: Adjust FSF address. Add test runs for
1915         -O2, -O3, -Os and the warning flags -W -Wall.
1916         * testsuite/libffi.special/special.exp: Likewise.
1917         * testsuite/libffi.call/ffitest.h: Add an __UNUSED__ macro to mark
1918         unused parameter unused for gcc or else do nothing.
1919         * testsuite/libffi.special/ffitestcxx.h: Likewise.
1920         * testsuite/libffi.call/cls_12byte.c (cls_struct_12byte_gn): Mark cif
1921         and userdata unused.
1922         * testsuite/libffi.call/cls_16byte.c (cls_struct_16byte_gn): Likewise.
1923         * testsuite/libffi.call/cls_18byte.c (cls_struct_18byte_gn): Likewise.
1924         * testsuite/libffi.call/cls_19byte.c (cls_struct_19byte_gn): Likewise.
1925         * testsuite/libffi.call/cls_1_1byte.c (cls_struct_1_1byte_gn): Likewise.
1926         * testsuite/libffi.call/cls_20byte.c (cls_struct_20byte_gn): Likewise.
1927         * testsuite/libffi.call/cls_20byte1.c (cls_struct_20byte_gn): Likewise.
1928         * testsuite/libffi.call/cls_24byte.c (cls_struct_24byte_gn): Likewise.
1929         * testsuite/libffi.call/cls_2byte.c (cls_struct_2byte_gn): Likewise.
1930         * testsuite/libffi.call/cls_3_1byte.c (cls_struct_3_1byte_gn): Likewise.
1931         * testsuite/libffi.call/cls_3byte1.c (cls_struct_3byte_gn): Likewise.
1932         * testsuite/libffi.call/cls_3byte2.c (cls_struct_3byte_gn1): Likewise.
1933         * testsuite/libffi.call/cls_4_1byte.c (cls_struct_4_1byte_gn): Likewise.
1934         * testsuite/libffi.call/cls_4byte.c (cls_struct_4byte_gn): Likewise.
1935         * testsuite/libffi.call/cls_5_1_byte.c (cls_struct_5byte_gn): Likewise.
1936         * testsuite/libffi.call/cls_5byte.c (cls_struct_5byte_gn): Likewise.
1937         * testsuite/libffi.call/cls_64byte.c (cls_struct_64byte_gn): Likewise.
1938         * testsuite/libffi.call/cls_6_1_byte.c (cls_struct_6byte_gn): Likewise.
1939         * testsuite/libffi.call/cls_6byte.c (cls_struct_6byte_gn): Likewise.
1940         * testsuite/libffi.call/cls_7_1_byte.c (cls_struct_7byte_gn): Likewise.
1941         * testsuite/libffi.call/cls_7byte.c (cls_struct_7byte_gn): Likewise.
1942         * testsuite/libffi.call/cls_8byte.c (cls_struct_8byte_gn): Likewise.
1943         * testsuite/libffi.call/cls_9byte1.c (cls_struct_9byte_gn): Likewise.
1944         * testsuite/libffi.call/cls_9byte2.c (cls_struct_9byte_gn): Likewise.
1945         * testsuite/libffi.call/cls_align_double.c (cls_struct_align_gn):
1946         Likewise.
1947         * testsuite/libffi.call/cls_align_float.c (cls_struct_align_gn):
1948         Likewise.
1949         * testsuite/libffi.call/cls_align_longdouble.c (cls_struct_align_gn):
1950         Likewise.
1951         * testsuite/libffi.call/cls_align_pointer.c (cls_struct_align_fn): Cast
1952         void* to avoid compiler warning.
1953         (main): Likewise.
1954         (cls_struct_align_gn): Mark cif and userdata unused.
1955         * testsuite/libffi.call/cls_align_sint16.c (cls_struct_align_gn):
1956         Likewise.
1957         * testsuite/libffi.call/cls_align_sint32.c (cls_struct_align_gn):
1958         Likewise.
1959         * testsuite/libffi.call/cls_align_sint64.c (cls_struct_align_gn):
1960         Likewise.
1961         * testsuite/libffi.call/cls_align_uint16.c (cls_struct_align_gn):
1962         Likewise.
1963         * testsuite/libffi.call/cls_align_uint32.c (cls_struct_align_gn):
1964         Likewise.
1965         * testsuite/libffi.call/cls_double.c (cls_ret_double_fn): Likewise.
1966         * testsuite/libffi.call/cls_float.c (cls_ret_float_fn): Likewise.
1967         * testsuite/libffi.call/cls_multi_schar.c (test_func_gn): Mark cif and
1968         data unused.
1969         (main): Cast res_call to silence gcc.
1970         * testsuite/libffi.call/cls_multi_sshort.c (test_func_gn): Mark cif and
1971         data unused.
1972         (main): Cast res_call to silence gcc.
1973         * testsuite/libffi.call/cls_multi_sshortchar.c (test_func_gn): Mark cif
1974         and data unused.
1975         (main): Cast res_call to silence gcc.
1976         * testsuite/libffi.call/cls_multi_uchar.c (test_func_gn): Mark cif and
1977         data unused.
1978         (main): Cast res_call to silence gcc.
1979         * testsuite/libffi.call/cls_multi_ushort.c (test_func_gn): Mark cif and
1980         data unused.
1981         (main): Cast res_call to silence gcc.
1982         * testsuite/libffi.call/cls_multi_ushortchar.c (test_func_gn): Mark cif
1983         and data unused.
1984         (main): Cast res_call to silence gcc.
1985         * testsuite/libffi.call/cls_schar.c (cls_ret_schar_fn): Mark cif and
1986         userdata unused.
1987         (cls_ret_schar_fn): Cast printf parameter to silence gcc.
1988         * testsuite/libffi.call/cls_sint.c (cls_ret_sint_fn): Mark cif and
1989         userdata unused.
1990         (cls_ret_sint_fn): Cast printf parameter to silence gcc.
1991         * testsuite/libffi.call/cls_sshort.c (cls_ret_sshort_fn): Mark cif and
1992         userdata unused.
1993         (cls_ret_sshort_fn): Cast printf parameter to silence gcc.
1994         * testsuite/libffi.call/cls_uchar.c (cls_ret_uchar_fn):  Mark cif and
1995         userdata unused.
1996         (cls_ret_uchar_fn): Cast printf parameter to silence gcc.
1997         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Mark cif and
1998         userdata unused.
1999         (cls_ret_uint_fn): Cast printf parameter to silence gcc.
2000         * testsuite/libffi.call/cls_ulonglong.c (cls_ret_ulonglong_fn): Mark cif
2001         and userdata unused.
2002         * testsuite/libffi.call/cls_ushort.c (cls_ret_ushort_fn): Mark cif and
2003         userdata unused.
2004         (cls_ret_ushort_fn): Cast printf parameter to silence gcc.
2005         * testsuite/libffi.call/float.c (floating): Remove unused parameter e.
2006         * testsuite/libffi.call/float1.c (main): Remove unused variable i.
2007         Cleanup white spaces.
2008         * testsuite/libffi.call/negint.c (checking): Remove unused variable i.
2009         * testsuite/libffi.call/nested_struct.c (cls_struct_combined_gn): Mark
2010         cif and userdata unused.
2011         * testsuite/libffi.call/nested_struct1.c (cls_struct_combined_gn):
2012         Likewise.
2013         * testsuite/libffi.call/nested_struct10.c (B_gn): Likewise.
2014         * testsuite/libffi.call/nested_struct2.c (B_fn): Adjust printf
2015         formatters to silence gcc.
2016         (B_gn): Mark cif and userdata unused.
2017         * testsuite/libffi.call/nested_struct3.c (B_gn): Mark cif and userdata
2018         unused.
2019         * testsuite/libffi.call/nested_struct4.c: Mention related PR.
2020         (B_gn): Mark cif and userdata unused.
2021         * testsuite/libffi.call/nested_struct5.c (B_gn): Mark cif and userdata
2022         unused.
2023         * testsuite/libffi.call/nested_struct6.c: Mention related PR.
2024         (B_gn): Mark cif and userdata unused.
2025         * testsuite/libffi.call/nested_struct7.c (B_gn): Mark cif and userdata
2026         unused.
2027         * testsuite/libffi.call/nested_struct8.c (B_gn): Likewise.
2028         * testsuite/libffi.call/nested_struct9.c (B_gn): Likewise.
2029         * testsuite/libffi.call/problem1.c (stub): Likewise.
2030         * testsuite/libffi.call/pyobjc-tc.c (main): Cast the result to silence
2031         gcc.
2032         * testsuite/libffi.call/return_fl2.c (return_fl): Add the note mentioned
2033         in the last commit for this test case in the test case itself.
2034         * testsuite/libffi.call/closure_fn0.c (closure_test_fn0): Mark cif as
2035         unused.
2036         * testsuite/libffi.call/closure_fn1.c (closure_test_fn1): Likewise.
2037         * testsuite/libffi.call/closure_fn2.c (closure_test_fn2): Likewise.
2038         * testsuite/libffi.call/closure_fn3.c (closure_test_fn3): Likewise.
2039         * testsuite/libffi.call/closure_fn4.c (closure_test_fn0): Likewise.
2040         * testsuite/libffi.call/closure_fn5.c (closure_test_fn5): Likewise.
2041         * testsuite/libffi.call/closure_fn6.c (closure_test_fn0): Likewise.
2042
2043 2006-02-22  Kaz Kojima  <kkojima@gcc.gnu.org>
2044
2045         * src/sh/sysv.S: Fix register numbers in the FDE for
2046         ffi_closure_SYSV.
2047
2048 2006-02-20  Andreas Tobler  <a.tobler@schweiz.ch>
2049
2050         * testsuite/libffi.call/return_fl2.c (return_fl): Remove static
2051         declaration to avoid a false negative on ix86. See PR323.
2052
2053 2006-02-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2054
2055         * src/sh/ffi.c (ffi_closure_helper_SYSV): Remove unused variable
2056         and cast integer to void * if needed.  Update the pointer to
2057         the FP register saved area correctly.
2058
2059 2006-02-17  Andreas Tobler  <a.tobler@schweiz.ch>
2060
2061         * testsuite/libffi.call/nested_struct6.c: XFAIL this test until PR25630
2062         is fixed.
2063         * testsuite/libffi.call/nested_struct4.c: Likewise.
2064
2065 2006-02-16  Andreas Tobler  <a.tobler@schweiz.ch>
2066
2067         * testsuite/libffi.call/return_dbl.c: New test case.
2068         * testsuite/libffi.call/return_dbl1.c: Likewise.
2069         * testsuite/libffi.call/return_dbl2.c: Likewise.
2070         * testsuite/libffi.call/return_fl.c: Likewise.
2071         * testsuite/libffi.call/return_fl1.c: Likewise.
2072         * testsuite/libffi.call/return_fl2.c: Likewise.
2073         * testsuite/libffi.call/return_fl3.c: Likewise.
2074         * testsuite/libffi.call/closure_fn6.c: Likewise.
2075
2076         * testsuite/libffi.call/nested_struct2.c: Remove ffi_type_mylong
2077         definition.
2078         * testsuite/libffi.call/ffitest.h: Add ffi_type_mylong definition
2079         here to be used by other test cases too.
2080
2081         * testsuite/libffi.call/nested_struct10.c: New test case.
2082         * testsuite/libffi.call/nested_struct9.c: Likewise.
2083         * testsuite/libffi.call/nested_struct8.c: Likewise.
2084         * testsuite/libffi.call/nested_struct7.c: Likewise.
2085         * testsuite/libffi.call/nested_struct6.c: Likewise.
2086         * testsuite/libffi.call/nested_struct5.c: Likewise.
2087         * testsuite/libffi.call/nested_struct4.c: Likewise.
2088
2089 2006-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
2090
2091         * configure.ac: Enable libffi for sparc64-*-freebsd*.
2092         * configure: Rebuilt.
2093
2094 2006-01-18  Jakub Jelinek  <jakub@redhat.com>
2095
2096         * src/powerpc/sysv.S (smst_two_register): Don't call __ashldi3,
2097         instead do the shifting inline.
2098         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't compute %r5
2099         shift count unconditionally.  Simplify load sequences for 1, 2, 3, 4
2100         and 8 byte structs, for the remaining struct sizes don't call
2101         __lshrdi3, instead do the shifting inline.
2102
2103 2005-12-07  Thiemo Seufer  <ths@networkno.de>
2104
2105         * src/mips/ffitarget.h: Remove obsolete sgidefs.h include. Add
2106         missing parentheses.
2107         * src/mips/o32.S (ffi_call_O32): Code formatting. Define
2108         and use A3_OFF, FP_OFF, RA_OFF. Micro-optimizations.
2109         (ffi_closure_O32): Likewise, but with newly defined A3_OFF2,
2110         A2_OFF2, A1_OFF2, A0_OFF2, RA_OFF2, FP_OFF2, S0_OFF2, GP_OFF2,
2111         V1_OFF2, V0_OFF2, FA_1_1_OFF2, FA_1_0_OFF2, FA_0_1_OFF2,
2112         FA_0_0_OFF2.
2113         * src/mips/ffi.c (ffi_prep_args): Code formatting. Fix
2114         endianness bugs.
2115         (ffi_prep_closure): Improve trampoline instruction scheduling.
2116         (ffi_closure_mips_inner_O32): Fix endianness bugs.
2117
2118 2005-12-03  Alan Modra  <amodra@bigpond.net.au>
2119
2120         * src/powerpc/ffi.c: Formatting.
2121         (ffi_prep_args_SYSV): Avoid possible aliasing problems by using unions.
2122         (ffi_prep_args64): Likewise.
2123
2124 2005-09-30  Geoffrey Keating  <geoffk@apple.com>
2125
2126         * testsuite/lib/libffi-dg.exp (libffi_target_compile): For
2127         darwin, use -shared-libgcc not -lgcc_s, and explain why.
2128
2129 2005-09-26  Tom Tromey  <tromey@redhat.com>
2130
2131         * testsuite/libffi.call/float1.c (value_type): New typedef.
2132         (CANARY): New define.
2133         (main): Check for result buffer overflow.
2134         * src/powerpc/linux64.S: Handle linux64 long double returns.
2135         * src/powerpc/ffi.c (FLAG_RETURNS_128BITS): New constant.
2136         (ffi_prep_cif_machdep): Handle linux64 long double returns.
2137
2138 2005-08-25  Alan Modra  <amodra@bigpond.net.au>
2139
2140         PR target/23404
2141         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Correct placement of stack
2142         homed fp args.
2143         (ffi_status ffi_prep_cif_machdep): Correct stack sizing for same.
2144
2145 2005-08-11  Jakub Jelinek  <jakub@redhat.com>
2146
2147         * configure.ac (HAVE_HIDDEN_VISIBILITY_ATTRIBUTE): New test.
2148         (AH_BOTTOM): Add FFI_HIDDEN definition.
2149         * configure: Rebuilt.
2150         * fficonfig.h.in: Rebuilt.
2151         * src/powerpc/ffi.c (hidden): Remove.
2152         (ffi_closure_LINUX64, ffi_prep_args64, ffi_call_LINUX64,
2153         ffi_closure_helper_LINUX64): Use FFI_HIDDEN instead of hidden.
2154         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64,
2155         .ffi_closure_LINUX64): Use FFI_HIDDEN instead of .hidden.
2156         * src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV): Remove,
2157         add FFI_HIDDEN to its prototype.
2158         (ffi_closure_SYSV_inner): New.
2159         * src/x86/sysv.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2160         * src/x86/win32.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2161
2162 2005-08-10  Alfred M. Szmidt  <ams@gnu.org>
2163
2164         PR libffi/21819:
2165         * configure: Rebuilt.
2166         * configure.ac: Handle i*86-*-gnu*.
2167
2168 2005-08-09  Jakub Jelinek  <jakub@redhat.com>
2169
2170         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Use
2171         DW_CFA_offset_extended_sf rather than
2172         DW_CFA_GNU_negative_offset_extended.
2173         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise.
2174
2175 2005-07-22  SUGIOKA Toshinobu  <sugioka@itonet.co.jp>
2176
2177         * src/sh/sysv.S (ffi_call_SYSV): Stop argument popping correctly
2178         on sh3.
2179         (ffi_closure_SYSV): Change the stack layout for sh3 struct argument.
2180         * src/sh/ffi.c (ffi_prep_args): Fix sh3 argument copy, when it is
2181         partially on register.
2182         (ffi_closure_helper_SYSV): Likewise.
2183         (ffi_prep_cif_machdep): Don't set too many cif->flags.
2184
2185 2005-07-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2186
2187         * src/sh/ffi.c (ffi_call): Handle small structures correctly.
2188         Remove empty line.
2189         * src/sh64/ffi.c (simple_type): Remove.
2190         (return_type): Handle small structures correctly.
2191         (ffi_prep_args): Likewise.
2192         (ffi_call): Likewise.
2193         (ffi_closure_helper_SYSV): Likewise.
2194         * src/sh64/sysv.S (ffi_call_SYSV): Handle 1, 2 and 4-byte return.
2195         Emit position independent code if PIC and remove wrong datalabel
2196         prefixes from EH data.
2197
2198 2005-07-19  Andreas Tobler  <a.tobler@schweiz.ch>
2199
2200         * Makefile.am (nodist_libffi_la_SOURCES): Add POWERPC_FREEBSD.
2201         * Makefile.in: Regenerate.
2202         * include/Makefile.in: Likewise.
2203         * testsuite/Makefile.in: Likewise.
2204         * configure.ac: Add POWERPC_FREEBSD rules.
2205         * configure: Regenerate.
2206         * src/powerpc/ffitarget.h: Add POWERPC_FREEBSD rules.
2207         (FFI_SYSV_TYPE_SMALL_STRUCT): Define.
2208         * src/powerpc/ffi.c: Add flags to handle small structure returns
2209         in ffi_call_SYSV.
2210         (ffi_prep_cif_machdep): Handle small structures for SYSV 4 ABI.
2211         Aka FFI_SYSV.
2212         (ffi_closure_helper_SYSV): Likewise.
2213         * src/powerpc/ppc_closure.S: Add return types for small structures.
2214         * src/powerpc/sysv.S: Add bits to handle small structures for
2215         final SYSV 4 ABI.
2216
2217 2005-07-10  Andreas Tobler  <a.tobler@schweiz.ch>
2218
2219         * testsuite/libffi.call/cls_5_1_byte.c: New test file.
2220         * testsuite/libffi.call/cls_6_1_byte.c: Likewise.
2221         * testsuite/libffi.call/cls_7_1_byte.c: Likewise.
2222
2223 2005-07-05  Randolph Chung  <tausq@debian.org>
2224
2225         * src/pa/ffi.c (ffi_struct_type): Rename FFI_TYPE_SMALL_STRUCT1
2226         as FFI_TYPE_SMALL_STRUCT3.  Break out handling for 5-7 byte
2227         structures.  Kill compilation warnings.
2228         (ffi_closure_inner_LINUX): Print return values as hex in debug
2229         message.  Rename FFI_TYPE_SMALL_STRUCT1 as FFI_TYPE_SMALL_STRUCT3.
2230         Properly handle 5-7 byte structure returns.
2231         * src/pa/ffitarget.h (FFI_TYPE_SMALL_STRUCT1)
2232         (FFI_TYPE_SMALL_STRUCT2): Remove.
2233         (FFI_TYPE_SMALL_STRUCT3, FFI_TYPE_SMALL_STRUCT5)
2234         (FFI_TYPE_SMALL_STRUCT6, FFI_TYPE_SMALL_STRUCT7): Define.
2235         * src/pa/linux.S: Mark source file as using PA1.1 assembly.
2236         (checksmst1, checksmst2): Remove.
2237         (checksmst3): Optimize handling of 3-byte struct returns.
2238         (checksmst567): Properly handle 5-7 byte struct returns.
2239
2240 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2241
2242         PR libgcj/21943
2243         * src/mips/n32.S: Enforce PIC code.
2244         * src/mips/o32.S: Likewise.
2245
2246 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2247
2248         * configure.ac: Treat i*86-*-solaris2.10 and up as X86_64.
2249         * configure: Regenerate.
2250
2251 2005-06-01  Alan Modra  <amodra@bigpond.net.au>
2252
2253         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't use JUMPTARGET
2254         to call ffi_closure_helper_SYSV.  Append @local instead.
2255         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise for ffi_prep_args_SYSV.
2256
2257 2005-05-17  Kelley Cook  <kcook@gcc.gnu.org>
2258
2259         * configure.ac: Use AC_C_BIGENDIAN instead of AC_C_BIGENDIAN_CROSS.
2260         Use AC_CHECK_SIZEOF instead of AC_COMPILE_CHECK_SIZEOF.
2261         * Makefile.am (ACLOCAL_AMFLAGS): Remove -I ../config.
2262         * aclocal.m4, configure, fficonfig.h.in, Makefile.in,
2263         include/Makefile.in, testsuite/Makefile.in: Regenerate.
2264
2265 2005-05-09  Mike Stump  <mrs@apple.com>
2266
2267         * configure: Regenerate.
2268
2269 2005-05-08  Richard Henderson  <rth@redhat.com>
2270
2271         PR libffi/21285
2272         * src/alpha/osf.S: Update unwind into to match code.
2273
2274 2005-05-04  Andreas Degert <ad@papyrus-gmbh.de>
2275             Richard Henderson  <rth@redhat.com>
2276
2277         * src/x86/ffi64.c (ffi_prep_cif_machdep): Save sse-used flag in
2278         bit 11 of flags.
2279         (ffi_call): Mask return type field.  Pass ssecount to ffi_call_unix64.
2280         (ffi_prep_closure): Set carry bit if sse-used flag set.
2281         * src/x86/unix64.S (ffi_call_unix64): Add ssecount argument.
2282         Only load sse registers if ssecount non-zero.
2283         (ffi_closure_unix64): Only save sse registers if carry set on entry.
2284
2285 2005-04-29  Ralf Corsepius  <ralf.corsepius@rtems.org>
2286
2287         * configure.ac: Add i*86-*-rtems*, sparc*-*-rtems*,
2288         powerpc-*rtems*, arm*-*-rtems*, sh-*-rtems*.
2289         * configure: Regenerate.
2290
2291 2005-04-20  Hans-Peter Nilsson  <hp@axis.com>
2292
2293         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): In regsub use,
2294         have Tcl8.3-compatible intermediate variable.
2295
2296 2005-04-18  Simon Posnjak <simon.posnjak@siol.net>
2297             Hans-Peter Nilsson  <hp@axis.com>
2298
2299         * Makefile.am: Add CRIS support.
2300         * configure.ac: Likewise.
2301         * Makefile.in, configure, testsuite/Makefile.in,
2302         include/Makefile.in: Regenerate.
2303         * src/cris: New directory.
2304         * src/cris/ffi.c, src/cris/sysv.S, src/cris/ffitarget.h: New files.
2305         * src/prep_cif.c (ffi_prep_cif): Wrap in #ifndef __CRIS__.
2306
2307         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): Replace \n with
2308         \r?\n in output tests.
2309
2310 2005-04-12  Mike Stump  <mrs@apple.com>
2311
2312         * configure: Regenerate.
2313
2314 2005-03-30  Hans Boehm  <Hans.Boehm@hp.com>
2315
2316         * src/ia64/ffitarget.h (ffi_arg): Use long long instead of DI.
2317
2318 2005-03-30  Steve Ellcey  <sje@cup.hp.com>
2319
2320         * src/ia64/ffitarget.h (ffi_arg) ADD DI attribute.
2321         (ffi_sarg) Ditto.
2322         * src/ia64/unix.S (ffi_closure_unix): Extend gp
2323         to 64 bits in ILP32 mode.
2324         Load 64 bits even for short data.
2325
2326 2005-03-23  Mike Stump  <mrs@apple.com>
2327
2328         * src/powerpc/darwin.S: Update for -m64 multilib.
2329         * src/powerpc/darwin_closure.S: Likewise.
2330
2331 2005-03-21  Zack Weinberg  <zack@codesourcery.com>
2332
2333         * configure.ac: Do not invoke TL_AC_GCC_VERSION.
2334         Do not set tool_include_dir.
2335         * aclocal.m4, configure, Makefile.in, testsuite/Makefile.in:
2336         Regenerate.
2337         * include/Makefile.am: Set gcc_version and toollibffidir.
2338         * include/Makefile.in: Regenerate.
2339
2340 2005-02-22  Andrew Haley  <aph@redhat.com>
2341
2342         * src/powerpc/ffi.c (ffi_prep_cif_machdep): Bump alignment to
2343         odd-numbered register pairs for 64-bit integer types.
2344
2345 2005-02-23  Andreas Tobler  <a.tobler@schweiz.ch>
2346
2347         PR libffi/20104
2348         * testsuite/libffi.call/return_ll1.c: New test case.
2349
2350 2005-02-11  Janis Johnson  <janis187@us.ibm.com>
2351
2352         * testsuite/libffi.call/cls_align_longdouble.c: Remove dg-options.
2353         * testsuite/libffi.call/float.c: Ditto.
2354         * testsuite/libffi.call/float2.c: Ditto.
2355         * testsuite/libffi.call/float3.c: Ditto.
2356
2357 2005-02-08  Andreas Tobler  <a.tobler@schweiz.ch>
2358
2359         * src/frv/ffitarget.h: Remove PPC stuff which does not belong to frv.
2360
2361 2005-01-12  Eric Botcazou  <ebotcazou@libertysurf.fr>
2362
2363         * testsuite/libffi.special/special.exp (cxx_options): Add
2364         -shared-libgcc.
2365
2366 2004-12-31  Richard Henderson  <rth@redhat.com>
2367
2368         * src/types.c (FFI_AGGREGATE_TYPEDEF): Remove.
2369         (FFI_TYPEDEF): Rename from FFI_INTEGRAL_TYPEDEF.  Replace size and
2370         offset parameters with a type parameter; deduce size and structure
2371         alignment.  Update all users.
2372
2373 2004-12-31  Richard Henderson  <rth@redhat.com>
2374
2375         * src/types.c (FFI_TYPE_POINTER): Define with sizeof.
2376         (FFI_TYPE_LONGDOUBLE): Fix for ia64.
2377         * src/ia64/ffitarget.h (struct ffi_ia64_trampoline_struct): Move
2378         into ffi_prep_closure.
2379         * src/ia64/ia64_flags.h, src/ia64/ffi.c, src/ia64/unix.S: Rewrite
2380         from scratch.
2381
2382 2004-12-27  Richard Henderson  <rth@redhat.com>
2383
2384         * src/x86/unix64.S: Fix typo in unwind info.
2385
2386 2004-12-25  Richard Henderson  <rth@redhat.com>
2387
2388         * src/x86/ffi64.c (struct register_args): Rename from stackLayout.
2389         (enum x86_64_reg_class): Add X86_64_COMPLEX_X87_CLASS.
2390         (merge_classes): Check for it.
2391         (SSE_CLASS_P): New.
2392         (classify_argument): Pass byte_offset by value; perform all updates
2393         inside struct case.
2394         (examine_argument): Add classes argument; handle
2395         X86_64_COMPLEX_X87_CLASS.
2396         (ffi_prep_args): Merge into ...
2397         (ffi_call): ... here.  Share stack frame with ffi_call_unix64.
2398         (ffi_prep_cif_machdep): Setup cif->flags for proper structure return.
2399         (ffi_fill_return_value): Remove.
2400         (ffi_prep_closure): Remove dead assert.
2401         (ffi_closure_unix64_inner): Rename from ffi_closure_UNIX64_inner.
2402         Rewrite to use struct register_args instead of va_list.  Create
2403         flags for handling structure returns.
2404         * src/x86/unix64.S: Remove dead strings.
2405         (ffi_call_unix64): Rename from ffi_call_UNIX64.  Rewrite to share
2406         stack frame with ffi_call.  Handle structure returns properly.
2407         (float2sse, floatfloat2sse, double2sse): Remove.
2408         (sse2float, sse2double, sse2floatfloat): Remove.
2409         (ffi_closure_unix64): Rename from ffi_closure_UNIX64.  Rewrite
2410         to handle structure returns properly.
2411
2412 2004-12-08  David Edelsohn  <edelsohn@gnu.org>
2413
2414         * Makefile.am (AM_MAKEFLAGS): Remove duplicate LIBCFLAGS and
2415         PICFLAG.
2416         * Makefile.in: Regenerated.
2417
2418 2004-12-02  Richard Sandiford  <rsandifo@redhat.com>
2419
2420         * configure.ac: Use TL_AC_GCC_VERSION to set gcc_version.
2421         * configure, aclocal.m4, Makefile.in: Regenerate.
2422         * include/Makefile.in, testsuite/Makefile.in: Regenerate.
2423
2424 2004-11-29  Kelley Cook  <kcook@gcc.gnu.org>
2425
2426         * configure: Regenerate for libtool change.
2427
2428 2004-11-25  Kelley Cook  <kcook@gcc.gnu.org>
2429
2430         * configure: Regenerate for libtool reversion.
2431
2432 2004-11-24  Kelley Cook  <kcook@gcc.gnu.org>
2433
2434         * configure: Regenerate for libtool change.
2435
2436 2004-11-23  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2437
2438         * testsuite/lib/libffi-dg.exp: Use new procs in target-libpath.exp.
2439
2440 2004-11-23  Richard Sandiford  <rsandifo@redhat.com>
2441
2442         * src/mips/o32.S (ffi_call_O32, ffi_closure_O32): Use jalr instead
2443         of jal.  Use an absolute encoding for the frame information.
2444
2445 2004-11-23  Kelley Cook  <kcook@gcc.gnu.org>
2446
2447         * Makefile.am: Remove no-dependencies.  Add ACLOCAL_AMFLAGS.
2448         * acinclude.m4: Delete logic for sincludes.
2449         * aclocal.m4, Makefile.in, configure: Regenerate.
2450         * include/Makefile: Likewise.
2451         * testsuite/Makefile: Likewise.
2452
2453 2004-11-22  Eric Botcazou  <ebotcazou@libertysurf.fr>
2454
2455         * src/sparc/ffi.c (ffi_prep_closure): Align doubles and 64-bit integers
2456         on a 8-byte boundary.
2457         * src/sparc/v8.S (ffi_closure_v8): Reserve frame space for arguments.
2458
2459 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2460
2461         * src/arm/ffi.c (ffi_prep_cif_machdep): Handle functions that return
2462         long long values.  Round stack allocation to a multiple of 8 bytes
2463         for ATPCS compatibility.
2464         * src/arm/sysv.S (ffi_call_SYSV): Rework to avoid use of APCS register
2465         names.  Handle returning long long types.  Add Thumb and interworking
2466         support.  Improve soft-float code.
2467
2468 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2469
2470         * testsuite/lib/libffi-db.exp (load_gcc_lib): New function.
2471         (libffi_exit): New function.
2472         (libffi_init): Build the testglue wrapper if needed.
2473
2474 2004-10-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2475
2476         PR other/18138
2477         * testsuite/lib/libffi-dg.exp: Accept more than one multilib libgcc.
2478
2479 2004-10-25  Kazuhiro Inaoka <inaoka.kazuhiro@renesas.com>
2480
2481         * src/m32r/libffitarget.h (FFI_CLOSURES): Set to 0.
2482
2483 2004-10-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2484
2485         * src/sh/sysv.S (ffi_call_SYSV): Don't align for double data.
2486         * testsuite/libffi.call/float3.c: New test case.
2487
2488 2004-10-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2489
2490         * src/sh/ffi.c (ffi_prep_closure): Set T bit in trampoline for
2491         the function returning a structure pointed with R2.
2492         * src/sh/sysv.S (ffi_closure_SYSV): Use R2 as the pointer to
2493         the structure return value if T bit set.  Emit position
2494         independent code and EH data if PIC.
2495
2496 2004-10-13  Kazuhiro Inaoka  <inaoka.kazuhiro@renesas.com>
2497
2498         * Makefile.am: Add m32r support.
2499         * configure.ac: Likewise.
2500         * Makefile.in: Regenerate.
2501         * confiugre: Regenerate.
2502         * src/types.c: Add m32r port to FFI_INTERNAL_TYPEDEF
2503         (uint64, sint64, double, longdouble)
2504         * src/m32r: New directory.
2505         * src/m32r/ffi.c: New file.
2506         * src/m32r/sysv.S: Likewise.
2507         * src/m32r/ffitarget.h: Likewise.
2508
2509 2004-10-02  Kaz Kojima  <kkojima@gcc.gnu.org>
2510
2511         * testsuite/libffi.call/negint.c: New test case.
2512
2513 2004-09-14  H.J. Lu  <hongjiu.lu@intel.com>
2514
2515         PR libgcj/17465
2516         * testsuite/lib/libffi-dg.exp: Don't use global ld_library_path.
2517         Set up LD_LIBRARY_PATH, SHLIB_PATH, LD_LIBRARYN32_PATH,
2518         LD_LIBRARY64_PATH, LD_LIBRARY_PATH_32, LD_LIBRARY_PATH_64 and
2519         DYLD_LIBRARY_PATH.
2520
2521 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2522
2523         * testsuite/libffi.call/many_win32.c: Remove whitespaces.
2524         * testsuite/libffi.call/promotion.c: Likewise.
2525         * testsuite/libffi.call/return_ll.c: Remove unused var. Cleanup
2526         whitespaces.
2527         * testsuite/libffi.call/return_sc.c: Likewise.
2528         * testsuite/libffi.call/return_uc.c: Likewise.
2529
2530 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2531
2532         * src/powerpc/darwin.S: Fix comments and identation.
2533         * src/powerpc/darwin_closure.S: Likewise.
2534
2535 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2536
2537         * src/powerpc/ffi_darwin.c: Add flag for longdouble return values.
2538         (ffi_prep_args): Handle longdouble arguments.
2539         (ffi_prep_cif_machdep): Set flags for longdouble. Calculate space for
2540         longdouble.
2541         (ffi_closure_helper_DARWIN): Add closure handling for longdouble.
2542         * src/powerpc/darwin.S (_ffi_call_DARWIN): Add handling of longdouble
2543         values.
2544         * src/powerpc/darwin_closure.S (_ffi_closure_ASM): Likewise.
2545         * src/types.c: Defined longdouble size and alignment for darwin.
2546
2547 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2548
2549         * src/powerpc/aix.S: Remove whitespaces.
2550         * src/powerpc/aix_closure.S: Likewise.
2551         * src/powerpc/asm.h: Likewise.
2552         * src/powerpc/ffi.c: Likewise.
2553         * src/powerpc/ffitarget.h: Likewise.
2554         * src/powerpc/linux64.S: Likewise.
2555         * src/powerpc/linux64_closure.S: Likewise.
2556         * src/powerpc/ppc_closure.S: Likewise.
2557         * src/powerpc/sysv.S: Likewise.
2558
2559 2004-08-30  Anthony Green  <green@redhat.com>
2560
2561         * Makefile.am: Add frv support.
2562         * Makefile.in, testsuite/Makefile.in: Rebuilt.
2563         * configure.ac: Read configure.host.
2564         * configure.in: Read configure.host.
2565         * configure.host: New file.  frv-elf needs libgloss.
2566         * include/ffi.h.in: Force ffi_closure to have a nice big (8)
2567         alignment.  This is needed to frv and shouldn't harm the others.
2568         * include/ffi_common.h (ALIGN_DOWN): New macro.
2569         * src/frv/ffi.c, src/frv/ffitarget.h, src/frv/eabi.S: New files.
2570
2571 2004-08-24  David Daney  <daney@avtrex.com>
2572
2573         * testsuite/libffi.call/closure_fn0.c: Xfail mips64* instead of mips*.
2574         * testsuite/libffi.call/closure_fn1.c: Likewise.
2575         * testsuite/libffi.call/closure_fn2.c  Likewise.
2576         * testsuite/libffi.call/closure_fn3.c: Likewise.
2577         * testsuite/libffi.call/closure_fn4.c: Likewise.
2578         * testsuite/libffi.call/closure_fn5.c: Likewise.
2579         * testsuite/libffi.call/cls_18byte.c: Likewise.
2580         * testsuite/libffi.call/cls_19byte.c: Likewise.
2581         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2582         * testsuite/libffi.call/cls_20byte.c: Likewise.
2583         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2584         * testsuite/libffi.call/cls_24byte.c: Likewise.
2585         * testsuite/libffi.call/cls_2byte.c: Likewise.
2586         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2587         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2588         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2589         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2590         * testsuite/libffi.call/cls_4byte.c: Likewise.
2591         * testsuite/libffi.call/cls_64byte.c: Likewise.
2592         * testsuite/libffi.call/cls_6byte.c: Likewise.
2593         * testsuite/libffi.call/cls_7byte.c: Likewise.
2594         * testsuite/libffi.call/cls_8byte.c: Likewise.
2595         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2596         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2597         * testsuite/libffi.call/cls_align_double.c: Likewise.
2598         * testsuite/libffi.call/cls_align_float.c: Likewise.
2599         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2600         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2601         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2602         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2603         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2604         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2605         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2606         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
2607         * testsuite/libffi.call/cls_double.c: Likewise.
2608         * testsuite/libffi.call/cls_float.c: Likewise.
2609         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2610         * testsuite/libffi.call/cls_multi_sshort.c: Likewise.
2611         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2612         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2613         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2614         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2615         * testsuite/libffi.call/cls_schar.c: Likewise.
2616         * testsuite/libffi.call/cls_sint.c: Likewise.
2617         * testsuite/libffi.call/cls_sshort.c: Likewise.
2618         * testsuite/libffi.call/cls_uchar.c: Likewise.
2619         * testsuite/libffi.call/cls_uint.c: Likewise.
2620         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2621         * testsuite/libffi.call/cls_ushort.c: Likewise.
2622         * testsuite/libffi.call/nested_struct.c: Likewise.
2623         * testsuite/libffi.call/nested_struct1.c: Likewise.
2624         * testsuite/libffi.call/nested_struct2.c: Likewise.
2625         * testsuite/libffi.call/nested_struct3.c: Likewise.
2626         * testsuite/libffi.call/problem1.c: Likewise.
2627         * testsuite/libffi.special/unwindtest.cc: Likewise.
2628         * testsuite/libffi.call/cls_12byte.c: Likewise and set return value
2629         to zero.
2630         * testsuite/libffi.call/cls_16byte.c: Likewise.
2631         * testsuite/libffi.call/cls_5byte.c: Likewise.
2632
2633 2004-08-23  David Daney <daney@avtrex.com>
2634
2635         PR libgcj/13141
2636         * src/mips/ffitarget.h (FFI_O32_SOFT_FLOAT): New ABI.
2637         * src/mips/ffi.c (ffi_prep_args): Fix alignment calculation.
2638         (ffi_prep_cif_machdep): Handle FFI_O32_SOFT_FLOAT floating point
2639         parameters and return types.
2640         (ffi_call): Handle FFI_O32_SOFT_FLOAT ABI.
2641         (ffi_prep_closure): Ditto.
2642         (ffi_closure_mips_inner_O32): Handle FFI_O32_SOFT_FLOAT ABI, fix
2643         alignment calculations.
2644         * src/mips/o32.S (ffi_closure_O32): Don't use floating point
2645         instructions if FFI_O32_SOFT_FLOAT, make stack frame ABI compliant.
2646
2647 2004-08-14  Casey Marshall <csm@gnu.org>
2648
2649         * src/mips/ffi.c (ffi_pref_cif_machdep): set `cif->flags' to
2650         contain `FFI_TYPE_UINT64' as return type for any 64-bit
2651         integer (O32 ABI only).
2652         (ffi_prep_closure): new function.
2653         (ffi_closure_mips_inner_O32): new function.
2654         * src/mips/ffitarget.h: Define `FFI_CLOSURES' and
2655         `FFI_TRAMPOLINE_SIZE' appropriately if the ABI is o32.
2656         * src/mips/o32.S (ffi_call_O32): add labels for .eh_frame. Return
2657         64 bit integers correctly.
2658         (ffi_closure_O32): new function.
2659         Added DWARF-2 unwind info for both functions.
2660
2661 2004-08-10  Andrew Haley  <aph@redhat.com>
2662
2663         * src/x86/ffi64.c (ffi_prep_args ): 8-align all stack arguments.
2664
2665 2004-08-01  Robert Millan  <robertmh@gnu.org>
2666
2667         * configure.ac: Detect knetbsd-gnu and kfreebsd-gnu.
2668         * configure: Regenerate.
2669
2670 2004-07-30  Maciej W. Rozycki  <macro@linux-mips.org>
2671
2672         * acinclude.m4 (AC_FUNC_MMAP_BLACKLIST): Check for <sys/mman.h>
2673         and mmap() explicitly instead of relying on preset autoconf cache
2674         variables.
2675         * aclocal.m4: Regenerate.
2676         * configure: Regenerate.
2677
2678 2004-07-11  Ulrich Weigand  <uweigand@de.ibm.com>
2679
2680         * src/s390/ffi.c (ffi_prep_args): Fix C aliasing violation.
2681         (ffi_check_float_struct): Remove unused prototype.
2682
2683 2004-06-30  Geoffrey Keating  <geoffk@apple.com>
2684
2685         * src/powerpc/ffi_darwin.c (flush_icache): ';' is a comment
2686         character on Darwin, use '\n\t' instead.
2687
2688 2004-06-26  Matthias Klose  <doko@debian.org>
2689
2690         * libtool-version: Fix typo in revision/age.
2691
2692 2004-06-17  Matthias Klose  <doko@debian.org>
2693
2694         * libtool-version: New.
2695         * Makefile.am (libffi_la_LDFLAGS): Use -version-info for soname.
2696         * Makefile.in: Regenerate.
2697
2698 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2699
2700         * Makefile.am: Remove useless multilib rules.
2701         * Makefile.in: Regenerate.
2702         * aclocal.m4: Regenerate with automake 1.8.5.
2703         * configure.ac: Remove useless multilib configury.
2704         * configure: Regenerate.
2705
2706 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2707
2708         * .cvsignore: New file.
2709
2710 2004-06-10  Jakub Jelinek  <jakub@redhat.com>
2711
2712         * src/ia64/unix.S (ffi_call_unix): Insert group barrier break
2713         fp_done.
2714         (ffi_closure_UNIX): Fix f14/f15 adjustment if FLOAT_SZ is ever
2715         changed from 8.
2716
2717 2004-06-06  Sean McNeil  <sean@mcneil.com>
2718
2719         * configure.ac: Add x86_64-*-freebsd* support.
2720         * configure: Regenerate.
2721
2722 2004-04-26  Joe Buck <jbuck@welsh-buck.org>
2723
2724         Bug 15093
2725         * configure.ac: Test for existence of mmap and sys/mman.h before
2726         checking blacklist.  Fix suggested by Jim Wilson.
2727         * configure: Regenerate.
2728
2729 2004-04-26  Matt Austern  <austern@apple.com>
2730
2731         * src/powerpc/darwin.S: Go through a non-lazy pointer for initial
2732         FDE location.
2733         * src/powerpc/darwin_closure.S: Likewise.
2734
2735 2004-04-24  Andreas Tobler  <a.tobler@schweiz.ch>
2736
2737         * testsuite/libffi.call/cls_multi_schar.c (main): Fix initialization
2738         error. Reported by Thomas Heller <theller@python.net>.
2739         * testsuite/libffi.call/cls_multi_sshort.c (main): Likewise.
2740         * testsuite/libffi.call/cls_multi_ushort.c (main): Likewise.
2741
2742 2004-03-20  Matthias Klose  <doko@debian.org>
2743
2744         * src/pa/linux.S: Fix typo.
2745
2746 2004-03-19  Matthias Klose  <doko@debian.org>
2747
2748         * Makefile.am: Update.
2749         * Makefile.in: Regenerate.
2750         * src/pa/ffi.h.in: Remove.
2751         * src/pa/ffitarget.h: New file.
2752
2753 2004-02-10  Randolph Chung  <tausq@debian.org>
2754
2755         * Makefile.am: Add PA support.
2756         * Makefile.in: Regenerate.
2757         * include/Makefile.in: Regenerate.
2758         * configure.ac: Add PA target.
2759         * configure: Regenerate.
2760         * src/pa/ffi.c: New file.
2761         * src/pa/ffi.h.in: Add PA support.
2762         * src/pa/linux.S: New file.
2763         * prep_cif.c: Add PA support.
2764
2765 2004-03-16  Hosaka Yuji  <hos@tamanegi.org>
2766
2767         * src/types.c: Fix alignment size of X86_WIN32 case int64 and
2768         double.
2769         * src/x86/ffi.c (ffi_prep_args): Replace ecif->cif->rtype->type
2770         with ecif->cif->flags.
2771         (ffi_call, ffi_prep_incoming_args_SYSV): Replace cif->rtype->type
2772         with cif->flags.
2773         (ffi_prep_cif_machdep): Add X86_WIN32 struct case.
2774         (ffi_closure_SYSV): Add 1 or 2-bytes struct case for X86_WIN32.
2775         * src/x86/win32.S (retstruct1b, retstruct2b, sc_retstruct1b,
2776         sc_retstruct2b): Add for 1 or 2-bytes struct case.
2777
2778 2004-03-15 Kelley Cook <kcook@gcc.gnu.org>
2779
2780         * configure.in: Rename file to ...
2781         * configure.ac: ... this.
2782         * fficonfig.h.in: Regenerate.
2783         * Makefile.in: Regenerate.
2784         * include/Makefile.in: Regenerate.
2785         * testsuite/Makefile.in: Regenerate.
2786
2787 2004-03-12  Matt Austern  <austern@apple.com>
2788
2789         * src/powerpc/darwin.S: Fix EH information so it corresponds to
2790         changes in EH format resulting from addition of linkonce support.
2791         * src/powerpc/darwin_closure.S: Likewise.
2792
2793 2004-03-11  Andreas Tobler  <a.tobler@schweiz.ch>
2794             Paolo Bonzini  <bonzini@gnu.org>
2795
2796         * Makefile.am (AUTOMAKE_OPTIONS): Set them.
2797         Remove VPATH. Remove rules for object files. Remove multilib support.
2798         (AM_CCASFLAGS): Add.
2799         * configure.in (AC_CONFIG_HEADERS): Relace AM_CONFIG_HEADER.
2800         (AC_PREREQ): Bump version to 2.59.
2801         (AC_INIT): Fill with version info and bug address.
2802         (ORIGINAL_LD_FOR_MULTILIBS): Remove.
2803         (AM_ENABLE_MULTILIB): Use this instead of AC_ARG_ENABLE.
2804         De-precious CC so that the right flags are passed down to multilibs.
2805         (AC_MSG_ERROR): Replace obsolete macro AC_ERROR.
2806         (AC_CONFIG_FILES): Replace obsolete macro AC_LINK_FILES.
2807         (AC_OUTPUT): Reorganize the output with AC_CONFIG_COMMANDS.
2808         * configure: Rebuilt.
2809         * aclocal.m4: Likewise.
2810         * Makefile.in, include/Makefile.in, testsuite/Makefile.in: Likewise.
2811         * fficonfig.h.in: Likewise.
2812
2813 2004-03-11  Andreas Schwab  <schwab@suse.de>
2814
2815         * src/ia64/ffi.c (ffi_prep_incoming_args_UNIX): Get floating point
2816         arguments from fp registers only for the first 8 parameter slots.
2817         Don't convert a float parameter when passed in memory.
2818
2819 2004-03-09  Hans-Peter Nilsson  <hp@axis.com>
2820
2821         * configure: Regenerate for config/accross.m4 correction.
2822
2823 2004-02-25  Matt Kraai  <kraai@alumni.cmu.edu>
2824
2825         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Change
2826         ecif->cif->bytes to bytes.
2827         (ffi_prep_cif_machdep): Add braces around nested if statement.
2828
2829 2004-02-09  Alan Modra  <amodra@bigpond.net.au>
2830
2831         * src/types.c (pointer): POWERPC64 has 8 byte pointers.
2832
2833         * src/powerpc/ffi.c (ffi_prep_args64): Correct long double handling.
2834         (ffi_closure_helper_LINUX64): Fix typo.
2835         * testsuite/libffi.call/cls_align_longdouble.c: Pass -mlong-double-128
2836         for powerpc64-*-*.
2837         * testsuite/libffi.call/float.c: Likewise.
2838         * testsuite/libffi.call/float2.c: Likewise.
2839
2840 2004-02-08  Alan Modra  <amodra@bigpond.net.au>
2841
2842         * src/powerpc/ffi.c (ffi_prep_cif_machdep <FFI_LINUX64>): Correct
2843         long double function return and long double arg handling.
2844         (ffi_closure_helper_LINUX64): Formatting.  Delete unused "ng" var.
2845         Use "end_pfr" instead of "nf".  Correct long double handling.
2846         Localise "temp".
2847         * src/powerpc/linux64.S (ffi_call_LINUX64): Save f2 long double
2848         return value.
2849         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Allocate
2850         space for long double return value.  Adjust stack frame and offsets.
2851         Load f2 long double return.
2852
2853 2004-02-07  Alan Modra  <amodra@bigpond.net.au>
2854
2855         * src/types.c: Use 16 byte long double for POWERPC64.
2856
2857 2004-01-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2858
2859         * src/sparc/ffi.c (ffi_prep_args_v9): Shift the parameter array
2860         when the structure return address is passed in %o0.
2861         (ffi_V9_return_struct): Rename into ffi_v9_layout_struct.
2862         (ffi_v9_layout_struct): Align the field following a nested structure
2863         on a word boundary.  Use memmove instead of memcpy.
2864         (ffi_call): Update call to ffi_V9_return_struct.
2865         (ffi_prep_closure): Define 'ctx' only for V8.
2866         (ffi_closure_sparc_inner): Clone into ffi_closure_sparc_inner_v8
2867         and ffi_closure_sparc_inner_v9.
2868         (ffi_closure_sparc_inner_v8): Return long doubles by reference.
2869         Always skip the structure return address.  For structures and long
2870         doubles, copy the argument directly.
2871         (ffi_closure_sparc_inner_v9): Skip the structure return address only
2872         if required.  Shift the maximum floating-point slot accordingly.  For
2873         big structures, copy the argument directly; otherwise, left-justify the
2874         argument and call ffi_v9_layout_struct to lay out the structure on
2875         the stack.
2876         * src/sparc/v8.S: Undef STACKFRAME before defining it.
2877         (ffi_closure_v8): Pass the structure return address.  Update call to
2878         ffi_closure_sparc_inner_v8.  Short-circuit FFI_TYPE_INT handling.
2879         Skip the 'unimp' insn when returning long doubles and structures.
2880         * src/sparc/v9.S: Undef STACKFRAME before defining it.
2881         (ffi_closure_v9): Increase the frame size by 2 words.  Short-circuit
2882         FFI_TYPE_INT handling.  Load structures both in integers and
2883         floating-point registers on return.
2884         * README: Update status of the SPARC port.
2885
2886 2004-01-24  Andreas Tobler  <a.tobler@schweiz.ch>
2887
2888         * testsuite/libffi.call/pyobjc-tc.c (main): Treat result value
2889         as of type ffi_arg.
2890         * testsuite/libffi.call/struct3.c (main): Fix CHECK.
2891
2892 2004-01-22  Ulrich Weigand  <uweigand@de.ibm.com>
2893
2894         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Treat result
2895         value as of type ffi_arg, not unsigned int.
2896
2897 2004-01-21  Michael Ritzert  <ritzert@t-online.de>
2898
2899         * ffi64.c (ffi_prep_args): Cast the RHS of an assignment instead
2900         of the LHS.
2901
2902 2004-01-12  Andreas Tobler  <a.tobler@schweiz.ch>
2903
2904         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_32 for
2905         Solaris.
2906
2907 2004-01-08  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2908
2909         * testsuite/libffi.call/ffitest.h (allocate_mmap): Cast MAP_FAILED
2910         to void *.
2911
2912 2003-12-10  Richard Henderson  <rth@redhat.com>
2913
2914         * testsuite/libffi.call/cls_align_pointer.c: Cast pointers to
2915         size_t instead of int.
2916
2917 2003-12-04  Hosaka Yuji  <hos@tamanegi.org>
2918
2919         * testsuite/libffi.call/many_win32.c: Include <float.h>.
2920         * testsuite/libffi.call/many_win32.c (main): Replace variable
2921         int i with unsigned long ul.
2922
2923         * testsuite/libffi.call/cls_align_uint64.c: New test case.
2924         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2925         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2926         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2927         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2928         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2929         * testsuite/libffi.call/cls_align_float.c: Likewise.
2930         * testsuite/libffi.call/cls_align_double.c: Likewise.
2931         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2932         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2933
2934 2003-12-02  Hosaka Yuji  <hos@tamanegi.org>
2935
2936         PR other/13221
2937         * src/x86/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV):
2938         Align arguments to 32 bits.
2939
2940 2003-12-01  Andreas Tobler  <a.tobler@schweiz.ch>
2941
2942         PR other/13221
2943         * testsuite/libffi.call/cls_multi_sshort.c: New test case.
2944         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2945         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2946         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2947         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2948         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2949
2950         * testsuite/libffi.special/unwindtest.cc: Cosmetics.
2951
2952 2003-11-26  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
2953
2954         * testsuite/libffi.call/ffitest.h: Include <fcntl.h>.
2955         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2956
2957 2003-11-22  Andreas Tobler  <a.tobler@schweiz.ch>
2958
2959         * Makefile.in: Rebuilt.
2960         * configure: Likewise.
2961         * testsuite/libffi.special/unwindtest.cc: Convert the mmap to
2962         the right type.
2963
2964 2003-11-21  Andreas Jaeger  <aj@suse.de>
2965             Andreas Tobler  <a.tobler@schweiz.ch>
2966
2967         * acinclude.m4: Add AC_FUNC_MMAP_BLACKLIST.
2968         * configure.in: Call AC_FUNC_MMAP_BLACKLIST.
2969         * Makefile.in: Rebuilt.
2970         * aclocal.m4: Likewise.
2971         * configure: Likewise.
2972         * fficonfig.h.in: Likewise.
2973         * testsuite/lib/libffi-dg.exp: Add include dir.
2974         * testsuite/libffi.call/ffitest.h: Add MMAP definitions.
2975         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2976         * testsuite/libffi.call/closure_fn0.c: Use MMAP functionality
2977         for ffi_closure if available.
2978         * testsuite/libffi.call/closure_fn1.c: Likewise.
2979         * testsuite/libffi.call/closure_fn2.c: Likewise.
2980         * testsuite/libffi.call/closure_fn3.c: Likewise.
2981         * testsuite/libffi.call/closure_fn4.c: Likewise.
2982         * testsuite/libffi.call/closure_fn5.c: Likewise.
2983         * testsuite/libffi.call/cls_12byte.c: Likewise.
2984         * testsuite/libffi.call/cls_16byte.c: Likewise.
2985         * testsuite/libffi.call/cls_18byte.c: Likewise.
2986         * testsuite/libffi.call/cls_19byte.c: Likewise.
2987         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2988         * testsuite/libffi.call/cls_20byte.c: Likewise.
2989         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2990         * testsuite/libffi.call/cls_24byte.c: Likewise.
2991         * testsuite/libffi.call/cls_2byte.c: Likewise.
2992         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2993         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2994         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2995         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2996         * testsuite/libffi.call/cls_4byte.c: Likewise.
2997         * testsuite/libffi.call/cls_5byte.c: Likewise.
2998         * testsuite/libffi.call/cls_64byte.c: Likewise.
2999         * testsuite/libffi.call/cls_6byte.c: Likewise.
3000         * testsuite/libffi.call/cls_7byte.c: Likewise.
3001         * testsuite/libffi.call/cls_8byte.c: Likewise.
3002         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3003         * testsuite/libffi.call/cls_9byte2.c: Likewise.
3004         * testsuite/libffi.call/cls_double.c: Likewise.
3005         * testsuite/libffi.call/cls_float.c: Likewise.
3006         * testsuite/libffi.call/cls_schar.c: Likewise.
3007         * testsuite/libffi.call/cls_sint.c: Likewise.
3008         * testsuite/libffi.call/cls_sshort.c: Likewise.
3009         * testsuite/libffi.call/cls_uchar.c: Likewise.
3010         * testsuite/libffi.call/cls_uint.c: Likewise.
3011         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3012         * testsuite/libffi.call/cls_ushort.c: Likewise.
3013         * testsuite/libffi.call/nested_struct.c: Likewise.
3014         * testsuite/libffi.call/nested_struct1.c: Likewise.
3015         * testsuite/libffi.call/nested_struct2.c: Likewise.
3016         * testsuite/libffi.call/nested_struct3.c: Likewise.
3017         * testsuite/libffi.call/problem1.c: Likewise.
3018         * testsuite/libffi.special/unwindtest.cc: Likewise.
3019
3020 2003-11-20  Andreas Tobler  <a.tobler@schweiz.ch>
3021
3022         * testsuite/lib/libffi-dg.exp: Make the -lgcc_s conditional.
3023
3024 2003-11-19  Andreas Tobler  <a.tobler@schweiz.ch>
3025
3026         * testsuite/lib/libffi-dg.exp: Add DYLD_LIBRARY_PATH for darwin.
3027         Add -lgcc_s to additional flags.
3028
3029 2003-11-12  Andreas Tobler  <a.tobler@schweiz.ch>
3030
3031         * configure.in, include/Makefile.am: PR libgcj/11147, install
3032         the ffitarget.h header file in a gcc versioned and target
3033         dependent place.
3034         * configure: Regenerated.
3035         * Makefile.in, include/Makefile.in: Likewise.
3036         * testsuite/Makefile.in: Likewise.
3037
3038 2003-11-09  Andreas Tobler  <a.tobler@schweiz.ch>
3039
3040         * testsuite/libffi.call/closure_fn0.c: Print result and check
3041         with dg-output to make debugging easier.
3042         * testsuite/libffi.call/closure_fn1.c: Likewise.
3043         * testsuite/libffi.call/closure_fn2.c: Likewise.
3044         * testsuite/libffi.call/closure_fn3.c: Likewise.
3045         * testsuite/libffi.call/closure_fn4.c: Likewise.
3046         * testsuite/libffi.call/closure_fn5.c: Likewise.
3047         * testsuite/libffi.call/cls_12byte.c: Likewise.
3048         * testsuite/libffi.call/cls_16byte.c: Likewise.
3049         * testsuite/libffi.call/cls_18byte.c: Likewise.
3050         * testsuite/libffi.call/cls_19byte.c: Likewise.
3051         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3052         * testsuite/libffi.call/cls_20byte.c: Likewise.
3053         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3054         * testsuite/libffi.call/cls_24byte.c: Likewise.
3055         * testsuite/libffi.call/cls_2byte.c: Likewise.
3056         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3057         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3058         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3059         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3060         * testsuite/libffi.call/cls_4byte.c: Likewise.
3061         * testsuite/libffi.call/cls_5byte.c: Likewise.
3062         * testsuite/libffi.call/cls_64byte.c: Likewise.
3063         * testsuite/libffi.call/cls_6byte.c: Likewise.
3064         * testsuite/libffi.call/cls_7byte.c: Likewise.
3065         * testsuite/libffi.call/cls_8byte.c: Likewise.
3066         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3067         * testsuite/libffi.call/cls_9byte2.c: Likewise.
3068         * testsuite/libffi.call/cls_double.c: Likewise.
3069         * testsuite/libffi.call/cls_float.c: Likewise.
3070         * testsuite/libffi.call/cls_schar.c: Likewise.
3071         * testsuite/libffi.call/cls_sint.c: Likewise.
3072         * testsuite/libffi.call/cls_sshort.c: Likewise.
3073         * testsuite/libffi.call/cls_uchar.c: Likewise.
3074         * testsuite/libffi.call/cls_uint.c: Likewise.
3075         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3076         * testsuite/libffi.call/cls_ushort.c: Likewise.
3077         * testsuite/libffi.call/problem1.c: Likewise.
3078
3079         * testsuite/libffi.special/unwindtest.cc: Make ffi_closure
3080         static.
3081
3082 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
3083
3084         * testsuite/libffi.call/cls_9byte2.c: New test case.
3085         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3086         * testsuite/libffi.call/cls_64byte.c: Likewise.
3087         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3088         * testsuite/libffi.call/cls_19byte.c: Likewise.
3089         * testsuite/libffi.call/cls_18byte.c: Likewise.
3090         * testsuite/libffi.call/closure_fn4.c: Likewise.
3091         * testsuite/libffi.call/closure_fn5.c: Likewise.
3092         * testsuite/libffi.call/cls_schar.c: Likewise.
3093         * testsuite/libffi.call/cls_sint.c: Likewise.
3094         * testsuite/libffi.call/cls_sshort.c: Likewise.
3095         * testsuite/libffi.call/nested_struct2.c: Likewise.
3096         * testsuite/libffi.call/nested_struct3.c: Likewise.
3097
3098 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
3099
3100         * testsuite/libffi.call/cls_double.c: Do a check on the result.
3101         * testsuite/libffi.call/cls_uchar.c: Likewise.
3102         * testsuite/libffi.call/cls_uint.c: Likewise.
3103         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3104         * testsuite/libffi.call/cls_ushort.c: Likewise.
3105         * testsuite/libffi.call/return_sc.c: Cleanup whitespaces.
3106
3107 2003-11-06  Andreas Tobler  <a.tobler@schweiz.ch>
3108
3109         * src/prep_cif.c (ffi_prep_cif): Move the validity check after
3110         the initialization.
3111
3112 2003-10-23  Andreas Tobler  <a.tobler@schweiz.ch>
3113
3114         * src/java_raw_api.c (ffi_java_ptrarray_to_raw): Replace
3115         FFI_ASSERT(FALSE) with FFI_ASSERT(0).
3116
3117 2003-10-22  David Daney  <ddaney@avtrex.com>
3118
3119         * src/mips/ffitarget.h: Replace undefined UINT32 and friends with
3120         __attribute__((__mode__(__SI__))) and friends.
3121
3122 2003-10-22  Andreas Schwab  <schwab@suse.de>
3123
3124         * src/ia64/ffi.c: Replace FALSE/TRUE with false/true.
3125
3126 2003-10-21  Andreas Tobler  <a.tobler@schweiz.ch>
3127
3128         * configure.in: AC_LINK_FILES(ffitarget.h).
3129         * configure: Regenerate.
3130         * Makefile.in: Likewise.
3131         * include/Makefile.in: Likewise.
3132         * testsuite/Makefile.in: Likewise.
3133         * fficonfig.h.in: Likewise.
3134
3135 2003-10-21  Paolo Bonzini  <bonzini@gnu.org>
3136             Richard Henderson  <rth@redhat.com>
3137
3138         Avoid that ffi.h includes fficonfig.h.
3139
3140         * Makefile.am (EXTRA_DIST): Include ffitarget.h files
3141         (TARGET_SRC_MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3142         (TARGET_SRC_MIPS_SGI): Removed.
3143         (MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3144         (MIPS_SGI): Removed.
3145         (CLEANFILES): Removed.
3146         (mostlyclean-am, clean-am, mostlyclean-sub, clean-sub): New
3147         targets.
3148         * acconfig.h: Removed.
3149         * configure.in: Compute sizeofs only for double and long double.
3150         Use them to define and subst HAVE_LONG_DOUBLE.  Include comments
3151         into AC_DEFINE instead of using acconfig.h.  Create
3152         include/ffitarget.h instead of include/fficonfig.h.  Rename
3153         MIPS_GCC to MIPS_IRIX, drop MIPS_SGI since we are in gcc's tree.
3154         AC_DEFINE EH_FRAME_FLAGS.
3155         * include/Makefile.am (DISTCLEANFILES): New automake macro.
3156         (hack_DATA): Add ffitarget.h.
3157         * include/ffi.h.in: Remove all system specific definitions.
3158         Declare raw API even if it is not installed, why bother?
3159         Use limits.h instead of SIZEOF_* to define ffi_type_*.  Do
3160         not define EH_FRAME_FLAGS, it is in fficonfig.h now.  Include
3161         ffitarget.h instead of fficonfig.h.  Remove ALIGN macro.
3162         (UINT_ARG, INT_ARG): Removed, use ffi_arg and ffi_sarg instead.
3163         * include/ffi_common.h (bool): Do not define.
3164         (ffi_assert): Accept failed assertion.
3165         (ffi_type_test): Return void and accept file/line.
3166         (FFI_ASSERT): Pass stringized failed assertion.
3167         (FFI_ASSERT_AT): New macro.
3168         (FFI_ASSERT_VALID_TYPE): New macro.
3169         (UINT8, SINT8, UINT16, SINT16, UINT32, SINT32,
3170         UINT64, SINT64): Define here with gcc's __attribute__ macro
3171         instead of in ffi.h
3172         (FLOAT32, ALIGN): Define here instead of in ffi.h
3173         * include/ffi-mips.h: Removed.  Its content moved to
3174         src/mips/ffitarget.h after separating assembly and C sections.
3175         * src/alpha/ffi.c, src/alpha/ffi.c, src/java_raw_api.c
3176         src/prep_cif.c, src/raw_api.c, src/ia64/ffi.c,
3177         src/mips/ffi.c, src/mips/n32.S, src/mips/o32.S,
3178         src/mips/ffitarget.h, src/sparc/ffi.c, src/x86/ffi64.c:
3179         SIZEOF_ARG -> FFI_SIZEOF_ARG.
3180         * src/ia64/ffi.c: Include stdbool.h (provided by GCC 2.95+).
3181         * src/debug.c (ffi_assert): Accept stringized failed assertion.
3182         (ffi_type_test): Rewritten.
3183         * src/prep-cif.c (initialize_aggregate, ffi_prep_cif): Call
3184         FFI_ASSERT_VALID_TYPE.
3185         * src/alpha/ffitarget.h, src/arm/ffitarget.h,
3186         src/ia64/ffitarget.h, src/m68k/ffitarget.h,
3187         src/mips/ffitarget.h, src/powerpc/ffitarget.h,
3188         src/s390/ffitarget.h, src/sh/ffitarget.h,
3189         src/sh64/ffitarget.h, src/sparc/ffitarget.h,
3190         src/x86/ffitarget.h: New files.
3191         * src/alpha/osf.S, src/arm/sysv.S, src/ia64/unix.S,
3192         src/m68k/sysv.S, src/mips/n32.S, src/mips/o32.S,
3193         src/powerpc/aix.S, src/powerpc/darwin.S,
3194         src/powerpc/ffi_darwin.c, src/powerpc/linux64.S,
3195         src/powerpc/linux64_closure.S, src/powerpc/ppc_closure.S,
3196         src/powerpc/sysv.S, src/s390/sysv.S, src/sh/sysv.S,
3197         src/sh64/sysv.S, src/sparc/v8.S, src/sparc/v9.S,
3198         src/x86/sysv.S, src/x86/unix64.S, src/x86/win32.S:
3199         include fficonfig.h
3200
3201 2003-10-20  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3202
3203         * src/mips/ffi.c: Use _ABIN32, _ABIO32 instead of external
3204         _MIPS_SIM_NABI32, _MIPS_SIM_ABI32.
3205
3206 2003-10-19  Andreas Tobler  <a.tobler@schweiz.ch>
3207
3208         * src/powerpc/ffi_darwin.c (ffi_prep_args): Declare bytes again.
3209         Used when FFI_DEBUG = 1.
3210
3211 2003-10-14  Alan Modra  <amodra@bigpond.net.au>
3212
3213         * src/types.c (double, longdouble): Default POWERPC64 to 8 byte size
3214         and align.
3215
3216 2003-10-06  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3217
3218         * include/ffi_mips.h: Define FFI_MIPS_N32 for N32/N64 ABIs,
3219         FFI_MIPS_O32 for O32 ABI.
3220
3221 2003-10-01  Andreas Tobler  <a.tobler@schweiz.ch>
3222
3223         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_64 for
3224         SPARC64. Cleanup whitespaces.
3225
3226 2003-09-19  Andreas Tobler  <a.tobler@schweiz.ch>
3227
3228         * testsuite/libffi.call/closure_fn0.c: Xfail mips, arm,
3229         strongarm, xscale. Cleanup whitespaces.
3230         * testsuite/libffi.call/closure_fn1.c: Likewise.
3231         * testsuite/libffi.call/closure_fn2.c: Likewise.
3232         * testsuite/libffi.call/closure_fn3.c: Likewise.
3233         * testsuite/libffi.call/cls_12byte.c: Likewise.
3234         * testsuite/libffi.call/cls_16byte.c: Likewise.
3235         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3236         * testsuite/libffi.call/cls_20byte.c: Likewise.
3237         * testsuite/libffi.call/cls_24byte.c: Likewise.
3238         * testsuite/libffi.call/cls_2byte.c: Likewise.
3239         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3240         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3241         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3242         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3243         * testsuite/libffi.call/cls_4byte.c: Likewise.
3244         * testsuite/libffi.call/cls_5byte.c: Likewise.
3245         * testsuite/libffi.call/cls_6byte.c: Likewise.
3246         * testsuite/libffi.call/cls_7byte.c: Likewise.
3247         * testsuite/libffi.call/cls_8byte.c: Likewise.
3248         * testsuite/libffi.call/cls_double.c: Likewise.
3249         * testsuite/libffi.call/cls_float.c: Likewise.
3250         * testsuite/libffi.call/cls_uchar.c: Likewise.
3251         * testsuite/libffi.call/cls_uint.c: Likewise.
3252         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3253         * testsuite/libffi.call/cls_ushort.c: Likewise.
3254         * testsuite/libffi.call/nested_struct.c: Likewise.
3255         * testsuite/libffi.call/nested_struct1.c: Likewise.
3256         * testsuite/libffi.call/problem1.c: Likewise.
3257         * testsuite/libffi.special/unwindtest.cc: Likewise.
3258         * testsuite/libffi.call/pyobjc-tc.c: Cleanup whitespaces.
3259
3260 2003-09-18  David Edelsohn  <edelsohn@gnu.org>
3261
3262         * src/powerpc/aix.S: Cleanup whitespaces.
3263         * src/powerpc/aix_closure.S: Likewise.
3264
3265 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3266
3267         * src/powerpc/darwin.S: Cleanup whitespaces, comment formatting.
3268         * src/powerpc/darwin_closure.S: Likewise.
3269         * src/powerpc/ffi_darwin.c: Likewise.
3270
3271 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3272             David Edelsohn  <edelsohn@gnu.org>
3273
3274         * src/types.c (double): Add AIX and Darwin to the right TYPEDEF.
3275         * src/powerpc/aix_closure.S: Remove the pointer to the outgoing
3276         parameter stack.
3277         * src/powerpc/darwin_closure.S: Likewise.
3278         * src/powerpc/ffi_darwin.c (ffi_prep_args): Handle structures
3279         according to the Darwin/AIX ABI.
3280         (ffi_prep_cif_machdep): Likewise.
3281         (ffi_closure_helper_DARWIN): Likewise.
3282         Remove the outgoing parameter stack logic. Simplify the evaluation
3283         of the different CASE types.
3284         (ffi_prep_clousure): Avoid the casts on lvalues. Change the branch
3285         statement in the trampoline code.
3286
3287 2003-09-18  Kaz Kojima  <kkojima@gcc.gnu.org>
3288
3289         * src/sh/ffi.c (ffi_prep_args): Take account into the alignement
3290         for the register size.
3291         (ffi_closure_helper_SYSV): Handle the structure return value
3292         address correctly.
3293         (ffi_closure_helper_SYSV): Return the appropriate type when
3294         the registers are used for the structure return value.
3295         * src/sh/sysv.S (ffi_closure_SYSV): Fix the stack layout for
3296         the 64-bit return value.  Update copyright years.
3297
3298 2003-09-17  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3299
3300         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Search in
3301         srcdir for ffi_mips.h.
3302
3303 2003-09-12  Alan Modra  <amodra@bigpond.net.au>
3304
3305         * src/prep_cif.c (initialize_aggregate): Include tail padding in
3306         structure size.
3307         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Correct
3308         placement of float result.
3309         * testsuite/libffi.special/unwindtest.cc (closure_test_fn1): Correct
3310         cast of "resp" for big-endian 64 bit machines.
3311
3312 2003-09-11  Alan Modra  <amodra@bigpond.net.au>
3313
3314         * src/types.c (double, longdouble): Merge identical SH and ARM
3315         typedefs, and add POWERPC64.
3316         * src/powerpc/ffi.c (ffi_prep_args64): Correct next_arg calc for
3317         struct split over gpr and rest.
3318         (ffi_prep_cif_machdep): Correct intarg_count for structures.
3319         * src/powerpc/linux64.S (ffi_call_LINUX64): Fix gpr offsets.
3320
3321 2003-09-09  Andreas Tobler  <a.tobler@schweiz.ch>
3322
3323         * src/powerpc/ffi.c (ffi_closure_helper_SYSV) Handle struct
3324         passing correctly.
3325
3326 2003-09-09  Alan Modra  <amodra@bigpond.net.au>
3327
3328         * configure: Regenerate.
3329
3330 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3331
3332         * Makefile.am: Remove build rules for ffitest.
3333         * Makefile.in: Rebuilt.
3334
3335 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3336
3337         * src/java_raw_api.c: Include <stdlib.h> to fix compiler warning
3338         about implicit declaration of abort().
3339
3340 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3341
3342         * Makefile.am: Add dejagnu test framework. Fixes PR other/11411.
3343         * Makefile.in: Rebuilt.
3344         * configure.in: Add dejagnu test framework.
3345         * configure: Rebuilt.
3346
3347         * testsuite/Makefile.am: New file.
3348         * testsuite/Makefile.in: Built
3349         * testsuite/lib/libffi-dg.exp: New file.
3350         * testsuite/config/default.exp: Likewise.
3351         * testsuite/libffi.call/call.exp: Likewise.
3352         * testsuite/libffi.call/ffitest.h: Likewise.
3353         * testsuite/libffi.call/closure_fn0.c: Likewise.
3354         * testsuite/libffi.call/closure_fn1.c: Likewise.
3355         * testsuite/libffi.call/closure_fn2.c: Likewise.
3356         * testsuite/libffi.call/closure_fn3.c: Likewise.
3357         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3358         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3359         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3360         * testsuite/libffi.call/cls_2byte.c: Likewise.
3361         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3362         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3363         * testsuite/libffi.call/cls_4byte.c: Likewise.
3364         * testsuite/libffi.call/cls_5byte.c: Likewise.
3365         * testsuite/libffi.call/cls_6byte.c: Likewise.
3366         * testsuite/libffi.call/cls_7byte.c: Likewise.
3367         * testsuite/libffi.call/cls_8byte.c: Likewise.
3368         * testsuite/libffi.call/cls_12byte.c: Likewise.
3369         * testsuite/libffi.call/cls_16byte.c: Likewise.
3370         * testsuite/libffi.call/cls_20byte.c: Likewise.
3371         * testsuite/libffi.call/cls_24byte.c: Likewise.
3372         * testsuite/libffi.call/cls_double.c: Likewise.
3373         * testsuite/libffi.call/cls_float.c: Likewise.
3374         * testsuite/libffi.call/cls_uchar.c: Likewise.
3375         * testsuite/libffi.call/cls_uint.c: Likewise.
3376         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3377         * testsuite/libffi.call/cls_ushort.c: Likewise.
3378         * testsuite/libffi.call/float.c: Likewise.
3379         * testsuite/libffi.call/float1.c: Likewise.
3380         * testsuite/libffi.call/float2.c: Likewise.
3381         * testsuite/libffi.call/many.c: Likewise.
3382         * testsuite/libffi.call/many_win32.c: Likewise.
3383         * testsuite/libffi.call/nested_struct.c: Likewise.
3384         * testsuite/libffi.call/nested_struct1.c: Likewise.
3385         * testsuite/libffi.call/pyobjc-tc.c: Likewise.
3386         * testsuite/libffi.call/problem1.c: Likewise.
3387         * testsuite/libffi.call/promotion.c: Likewise.
3388         * testsuite/libffi.call/return_ll.c: Likewise.
3389         * testsuite/libffi.call/return_sc.c: Likewise.
3390         * testsuite/libffi.call/return_uc.c: Likewise.
3391         * testsuite/libffi.call/strlen.c: Likewise.
3392         * testsuite/libffi.call/strlen_win32.c: Likewise.
3393         * testsuite/libffi.call/struct1.c: Likewise.
3394         * testsuite/libffi.call/struct2.c: Likewise.
3395         * testsuite/libffi.call/struct3.c: Likewise.
3396         * testsuite/libffi.call/struct4.c: Likewise.
3397         * testsuite/libffi.call/struct5.c: Likewise.
3398         * testsuite/libffi.call/struct6.c: Likewise.
3399         * testsuite/libffi.call/struct7.c: Likewise.
3400         * testsuite/libffi.call/struct8.c: Likewise.
3401         * testsuite/libffi.call/struct9.c: Likewise.
3402         * testsuite/libffi.special/special.exp: New file.
3403         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3404         * testsuite/libffi.special/unwindtest.cc: Likewise.
3405
3406
3407 2003-08-13  Kaz Kojima  <kkojima@gcc.gnu.org>
3408
3409         * src/sh/ffi.c (OFS_INT16): Set 0 for little endian case.  Update
3410         copyright years.
3411
3412 2003-08-02  Alan Modra  <amodra@bigpond.net.au>
3413
3414         * src/powerpc/ffi.c (ffi_prep_args64): Modify for changed gcc
3415         structure passing.
3416         (ffi_closure_helper_LINUX64): Likewise.
3417         * src/powerpc/linux64.S: Remove code writing to parm save area.
3418         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Use return
3419         address in lr from ffi_closure_helper_LINUX64 call to calculate
3420         table address.  Optimize function tail.
3421
3422 2003-07-28  Andreas Tobler  <a.tobler@schweiz.ch>
3423
3424         * src/sparc/ffi.c: Handle all floating point registers.
3425         * src/sparc/v9.S: Likewise. Fixes second part of PR target/11410.
3426
3427 2003-07-11  Gerald Pfeifer  <pfeifer@dbai.tuwien.ac.at>
3428
3429         * README: Note that libffi is not part of GCC.  Update the project
3430         URL and status.
3431
3432 2003-06-19  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3433
3434         * src/powerpc/ppc_closure.S: Include ffi.h.
3435
3436 2003-06-13  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3437
3438         * src/x86/sysv.S: Avoid gas-only .uleb128/.sleb128 directives.
3439         Use C s