OSDN Git Service

Improve libffi comments
[pf3gnuchains/gcc-fork.git] / libffi / ChangeLog
1 2010-07-07   Neil Roberts <neil@linux.intel.com>
2
3         * src/x86/sysv.S (ffi_call_SYSV): Align the stack pointer to
4         16-bytes.
5
6 2010-07-02  Jakub Jelinek  <jakub@redhat.com>
7
8         * Makefile.am (AM_MAKEFLAGS): Pass also mandir to submakes.
9         * Makefile.in: Regenerated.
10
11 2010-05-19  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
12
13         * configure.ac (libffi_cv_as_x86_pcrel): Check for illegal in as
14         output, too.
15         (libffi_cv_as_ascii_pseudo_op): Check for .ascii.
16         (libffi_cv_as_string_pseudo_op): Check for .string.
17         * configure: Regenerate.
18         * fficonfig.h.in: Regenerate.
19         * src/x86/sysv.S (.eh_frame): Use .ascii, .string or error.
20
21 2010-04-07  Jakub Jelinek  <jakub@redhat.com>
22
23         * regex.c (byte_re_match_2_internal): Avoid set but not used
24         warning.
25
26 2010-04-02  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
27
28         * Makefile.in: Regenerate.
29         * aclocal.m4: Regenerate.
30         * include/Makefile.in: Regenerate.
31         * man/Makefile.in: Regenerate.
32         * testsuite/Makefile.in: Regenerate.
33
34 2010-03-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
35
36         * configure.ac (libffi_cv_as_x86_64_unwind_section_type): New test.
37         * configure: Regenerate.
38         * fficonfig.h.in: Regenerate.
39         * libffi/src/x86/unix64.S (.eh_frame)
40         [HAVE_AS_X86_64_UNWIND_SECTION_TYPE]: Use @unwind section type.
41
42 2010-03-14  Matthias Klose  <doko@ubuntu.com>
43
44         * src/x86/ffi64.c: Fix typo in comment.
45         * src/x86/ffi.c: Use /* ... */ comment style.
46
47 2010-02-24  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
48
49         * doc/libffi.texi (The Closure API): Fix typo.
50         * doc/libffi.info: Remove.
51
52 2010-02-15  Matthias Klose  <doko@ubuntu.com>
53
54         * src/arm/sysv.S (__ARM_ARCH__): Define for processor
55         __ARM_ARCH_7EM__.
56
57 2010-01-15  Anthony Green  <green@redhat.com>
58
59         * README: Add notes on building with Microsoft Visual C++.
60
61 2010-01-15  Daniel Witte  <dwitte@mozilla.com>
62
63         * msvcc.sh: New file.
64
65         * src/x86/win32.S: Port assembly routines to MSVC and #ifdef.
66         * src/x86/ffi.c: Tweak function declaration and remove excess
67         parens.
68         * include/ffi.h.in: Add __declspec(align(8)) to typedef struct
69         ffi_closure.
70
71         * src/x86/ffi.c: Merge ffi_call_SYSV and ffi_call_STDCALL into new
72         function ffi_call_win32 on X86_WIN32.
73         * src/x86/win32.S (ffi_call_SYSV): Rename to ffi_call_win32.
74         (ffi_call_STDCALL): Remove.
75
76         * src/prep_cif.c (ffi_prep_cif): Move stack space allocation code
77         to ffi_prep_cif_machdep for x86.
78         * src/x86/ffi.c (ffi_prep_cif_machdep): To here.
79
80 2010-01-15  Oliver Kiddle  <okiddle@yahoo.co.uk>
81
82         * src/x86/ffitarget.h (ffi_abi): Check for __i386 and __amd64 for
83         Sun Studio compiler compatibility. 
84
85 2010-01-12  Conrad Irwin <conrad.irwin@gmail.com>
86
87         * doc/libffi.texi: Add closure example.
88
89 2010-01-07  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
90
91         PR libffi/40701
92         * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRIdLL,
93         PRIuLL, PRId64, PRIu64, PRIuPTR): Define.
94         * testsuite/libffi.call/cls_align_sint64.c: Add -Wno-format on
95         alpha*-dec-osf*.
96         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
97         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
98         * testsuite/libffi.call/return_ll1.c: Likewise.
99         * testsuite/libffi.call/stret_medium2.c: Likewise.
100         * testsuite/libffi.special/ffitestcxx.h (allocate_mmap): Cast
101         MAP_FAILED to char *.
102
103 2010-01-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
104
105         * src/mips/n32.S: Use .abicalls and .eh_frame with __GNUC__.
106
107 2009-12-31  Anthony Green  <green@redhat.com>
108
109         * README: Update for libffi 3.0.9.
110
111 2009-12-27  Matthias Klose  <doko@ubuntu.com>
112
113         * configure.ac (HAVE_LONG_DOUBLE): Define for mips when
114         appropriate.
115         * configure: Rebuilt.
116
117 2009-12-26  Anthony Green  <green@redhat.com>
118
119         * testsuite/libffi.call/cls_longdouble_va.c: Mark as xfail for
120         avr32*-*-*.
121         * testsuite/libffi.call/cls_double_va.c: Ditto.
122
123 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
124
125         * testsuite/libffi.call/ffitest.h: Conditionally include stdint.h
126         and inttypes.h.
127         * testsuite/libffi.special/unwindtest.cc: Ditto.
128
129 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
130
131         * configure.ac: Add amd64-*-openbsd*.
132         * configure: Rebuilt.
133         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Link
134         openbsd programs with -lpthread.
135
136 2009-12-26  Anthony Green  <green@redhat.com>
137
138         * testsuite/libffi.call/cls_double_va.c,
139         testsuite/libffi.call/cls_longdouble.c,
140         testsuite/libffi.call/cls_longdouble_va.c,
141         testsuite/libffi.call/cls_pointer.c,
142         testsuite/libffi.call/cls_pointer_stack.c: Remove xfail for
143         mips*-*-* and arm*-*-*.
144         * testsuite/libffi.call/cls_align_longdouble_split.c,
145         testsuite/libffi.call/cls_align_longdouble_split2.c,
146         testsuite/libffi.call/stret_medium2.c,
147         testsuite/libffi.call/stret_medium.c,
148         testsuite/libffi.call/stret_large.c,
149         testsuite/libffi.call/stret_large2.c: Remove xfail for arm*-*-*.
150
151 2009-12-31  Kay Tietz  <ktietz70@googlemail.com>
152
153         * testsuite/libffi.call/ffitest.h,
154         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRuLL): Fix
155         definitions.
156
157 2009-12-31  Carlo Bramini  <carlo.bramix@libero.it>
158
159         * configure.ac (AM_LTLDFLAGS): Define for windows hosts.
160         * Makefile.am (libffi_la_LDFLAGS): Add AM_LTLDFLAGS.
161         * configure: Rebuilt.
162         * Makefile.in: Rebuilt.
163
164 2009-12-31  Anthony Green  <green@redhat.com>
165             Blake Chaffin.
166
167         * testsuite/libffi.call/huge_struct.c: New test case from Blake
168         Chaffin @ Apple.
169
170 2009-12-28  David Edelsohn  <edelsohn@gnu.org>
171
172         * src/powerpc/ffi_darwin.c (ffi_prep_args): Copy abi and nargs to
173         local variables.
174         (aix_adjust_aggregate_sizes): New function.
175         (ffi_prep_cif_machdep): Call it.
176
177 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
178
179         * configure.ac: Define FFI_MMAP_EXEC_WRIT for the given targets.
180         * configure: Regenerate.
181         * fficonfig.h.in: Likewise.
182         * src/closures.c: Remove the FFI_MMAP_EXEC_WRIT definition for
183         Solaris/x86.
184
185 2009-12-26  Andreas Schwab  <schwab@linux-m68k.org>
186
187         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Advance intarg_count
188         when a float arguments is passed in memory.
189         (ffi_closure_helper_SYSV): Mark general registers as used up when
190         a 64bit or soft-float long double argument is passed in memory.
191
192 2009-12-25  Matthias Klose  <doko@ubuntu.com>
193
194         * man/ffi_call.3: Fix #include in examples.
195         * doc/libffi.texi: Add dircategory.
196
197 2009-12-25  Frank Everdij <f.p.x.everdij@tudelft.nl>
198
199         * include/ffi.h.in: Placed '__GNUC__' ifdef around
200         '__attribute__((aligned(8)))' in ffi_closure, fixes compile for
201         IRIX MIPSPro c99.
202         * include/ffi_common.h: Added '__sgi' define to non
203         '__attribute__((__mode__()))' integer typedefs.
204         * src/mips/ffi.c (ffi_call, ffi_closure_mips_inner_O32,
205         ffi_closure_mips_inner_N32): Added 'defined(_MIPSEB)' to BE check.
206         (ffi_closure_mips_inner_O32, ffi_closure_mips_inner_N32): Added
207         FFI_LONGDOUBLE support and alignment(N32 only).
208         * src/mips/ffitarget.h: Corrected '#include <sgidefs.h>' for IRIX and
209         fixed non '__attribute__((__mode__()))' integer typedefs.
210         * src/mips/n32.S: Put '#ifdef linux' around '.abicalls' and '.eh_frame'
211         since they are Linux/GNU Assembler specific.
212
213 2009-12-25  Bradley Smith  <brad@brad-smith.co.uk>
214
215         * configure.ac, Makefile.am, src/avr32/ffi.c,
216         src/avr32/ffitarget.h,
217         src/avr32/sysv.S: Add AVR32 port.
218         * configure, Makefile.in: Rebuilt.
219
220 2009-12-21  Andreas Tobler  <a.tobler@schweiz.org>
221
222         * configure.ac: Make i?86 build on FreeBSD and OpenBSD.
223         * configure: Regenerate.
224
225 2009-12-15  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
226
227         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on PA HP-UX.
228
229 2009-12-13  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
230
231         * src/pa/ffi.c (ffi_closure_inner_pa32): Handle FFI_TYPE_LONGDOUBLE
232         type on HP-UX.
233
234 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
235
236         * src/sparc/ffi.c (ffi_closure_sparc_inner_v9): Properly align 'long
237         double' arguments.
238
239 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
240
241         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on Solaris < 10.
242
243 2009-12-10  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
244
245         PR libffi/40700
246         * src/closures.c [X86_64 && __sun__ && __svr4__]
247         (FFI_MMAP_EXEC_WRIT): Define.
248
249 2009-12-08  David Daney  <ddaney@caviumnetworks.com>
250
251         * testsuite/libffi.call/stret_medium.c: Remove xfail for mips*-*-*
252         * testsuite/libffi.call/cls_align_longdouble_split2.c: Same.
253         * testsuite/libffi.call/stret_large.c: Same.
254         * testsuite/libffi.call/cls_align_longdouble_split.c: Same.
255         * testsuite/libffi.call/stret_large2.c: Same.
256         * testsuite/libffi.call/stret_medium2.c: Same.
257
258 2009-12-07  David Edelsohn  <edelsohn@gnu.org>
259
260         * src/powerpc/aix_closure.S (libffi_closure_ASM): Fix tablejump
261         typo.
262
263 2009-12-05  David Edelsohn  <edelsohn@gnu.org>
264
265         * src/powerpc/aix.S: Update AIX32 code to be consistent with AIX64
266         code.
267         * src/powerpc/aix_closure.S: Same.
268
269 2009-12-05  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
270
271         * Makefile.in: Regenerate.
272         * configure: Regenerate.
273         * include/Makefile.in: Regenerate.
274         * man/Makefile.in: Regenerate.
275         * testsuite/Makefile.in: Regenerate.
276
277 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
278
279         * src/powerpc/aix_closure.S: Reorganize 64-bit code to match
280         linux64_closure.S.
281
282 2009-12-04  Uros Bizjak  <ubizjak@gmail.com>
283
284         PR libffi/41908
285         * src/x86/ffi64.c (classify_argument): Update from
286         gcc/config/i386/i386.c.
287         (ffi_closure_unix64_inner): Do not use the address of two consecutive
288         SSE registers directly.
289         * testsuite/libffi.call/cls_dbls_struct.c (main): Remove xfail
290         for x86_64 linux targets.
291
292 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
293
294         * src/powerpc/ffi_darwin.c (ffi_closure_helper_DARWIN): Increment
295         pfr for long double split between fpr13 and stack.
296
297 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
298
299         * src/powerpc/ffi_darwin.c (ffi_prep_args): Increment next_arg and
300         fparg_count twice for long double.
301
302 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
303
304         PR libffi/42243
305         * src/powerpc/ffi_darwin.c (ffi_prep_args): Remove extra parentheses.
306
307 2009-12-03  Uros Bizjak  <ubizjak@gmail.com>
308
309         * testsuite/libffi.call/cls_longdouble_va.c (main): Fix format string.
310         Remove xfails for x86 linux targets.
311
312 2009-12-02  David Edelsohn  <edelsohn@gnu.org>
313
314         * src/powerpc/ffi_darwin.c (ffi_prep_args): Fix typo in INT64
315         case.
316
317 2009-12-01  David Edelsohn  <edelsohn@gnu.org>
318
319         * src/powerpc/aix.S (ffi_call_AIX): Convert to more standard
320         register usage.  Call ffi_prep_args directly.  Add long double
321         return value support.
322         * src/powerpc/ffi_darwin.c (ffi_prep_args): Double arg increment
323         applies to FFI_TYPE_DOUBLE.  Correct fpr_base increment typo.
324         Separate FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases.
325         (ffi_prep_cif_machdep): Only 16 byte stack alignment in 64 bit
326         mode.
327         (ffi_closure_helper_DARWIN): Remove nf and ng counters.  Move temp
328         into case.
329         * src/powerpc/aix_closure.S: Maintain 16 byte stack alignment.
330         Allocate result area between params and FPRs.
331
332 2009-11-30  David Edelsohn  <edelsohn@gnu.org>
333
334         PR target/35484
335         * src/powerpc/ffitarget.h (POWERPC64): Define for PPC64 Linux and
336         AIX64.
337         * src/powerpc/aix.S: Implement AIX64 version.
338         * src/powerpc/aix_closure.S: Implement AIX64 version.
339         (ffi_closure_ASM): Use extsb, lha and displament addresses.
340         * src/powerpc/ffi_darwin.c (ffi_prep_args): Implement AIX64
341         support.
342         (ffi_prep_cif_machdep): Same.
343         (ffi_call): Same.
344         (ffi_closure_helper_DARWIN): Same.
345
346 2009-11-02  Andreas Tobler  <a.tobler@schweiz.org>
347
348         PR libffi/41908
349         * testsuite/libffi.call/testclosure.c: New test.
350
351 2009-09-28  Kai Tietz  <kai.tietz@onevision.com>
352
353         * src/x86/win64.S (_ffi_call_win64 stack): Remove for gnu
354         assembly version use of ___chkstk.
355
356 2009-09-23  Matthias Klose  <doko@ubuntu.com>
357
358         PR libffi/40242, PR libffi/41443
359         * src/arm/sysv.S (__ARM_ARCH__): Define for processors
360         __ARM_ARCH_6T2__, __ARM_ARCH_6M__, __ARM_ARCH_7__,
361         __ARM_ARCH_7A__, __ARM_ARCH_7R__, __ARM_ARCH_7M__.
362         Change the conditionals to __SOFTFP__ || __ARM_EABI__
363         for -mfloat-abi=softfp to work.
364
365 2009-09-17  Loren J. Rittle  <ljrittle@acm.org>
366
367         PR testsuite/32843 (strikes again)
368         * src/x86/ffi.c (ffi_prep_cif_machdep): Add X86_FREEBSD to
369         enable proper extension on char and short.
370
371 2009-09-15  David Daney  <ddaney@caviumnetworks.com>
372
373         * src/java_raw_api.c (ffi_java_raw_to_rvalue): Remove special
374         handling for FFI_TYPE_POINTER.
375         * src/mips/ffitarget.h (FFI_TYPE_STRUCT_D_SOFT,
376         FFI_TYPE_STRUCT_F_SOFT, FFI_TYPE_STRUCT_DD_SOFT,
377         FFI_TYPE_STRUCT_FF_SOFT, FFI_TYPE_STRUCT_FD_SOFT,
378         FFI_TYPE_STRUCT_DF_SOFT, FFI_TYPE_STRUCT_SOFT): New defines.
379         (FFI_N32_SOFT_FLOAT, FFI_N64_SOFT_FLOAT): New ffi_abi enumerations.
380         (enum ffi_abi): Set FFI_DEFAULT_ABI for soft-float.
381         * src/mips/n32.S (ffi_call_N32): Add handling for soft-float
382         structure and pointer returns.
383         (ffi_closure_N32): Add handling for pointer returns.
384         * src/mips/ffi.c (ffi_prep_args, calc_n32_struct_flags,
385         calc_n32_return_struct_flags): Handle soft-float.
386         (ffi_prep_cif_machdep):  Handle soft-float, fix pointer handling.
387         (ffi_call_N32): Declare proper argument types.
388         (ffi_call, copy_struct_N32, ffi_closure_mips_inner_N32): Handle
389         soft-float.
390
391 2009-08-24  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
392
393         * configure.ac (AC_PREREQ): Bump to 2.64.
394
395 2009-08-22  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
396
397         * Makefile.am (install-html, install-pdf): Remove.
398         * Makefile.in: Regenerate.
399
400         * Makefile.in: Regenerate.
401         * aclocal.m4: Regenerate.
402         * configure: Regenerate.
403         * fficonfig.h.in: Regenerate.
404         * include/Makefile.in: Regenerate.
405         * man/Makefile.in: Regenerate.
406         * testsuite/Makefile.in: Regenerate.
407
408 2009-07-30  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
409
410         * configure.ac (_AC_ARG_VAR_PRECIOUS): Use m4_rename_force.
411
412 2009-07-24  Dave Korn  <dave.korn.cygwin@gmail.com>
413
414         PR libffi/40807
415         * src/x86/ffi.c (ffi_prep_cif_machdep): Also use sign/zero-extending
416         return types for X86_WIN32.
417         * src/x86/win32.S (_ffi_call_SYSV): Handle omitted return types.
418         (_ffi_call_STDCALL, _ffi_closure_SYSV, _ffi_closure_raw_SYSV,
419         _ffi_closure_STDCALL): Likewise.
420
421         * src/closures.c (is_selinux_enabled): Define to const 0 for Cygwin.
422         (dlmmap, dlmunmap): Also use these functions on Cygwin.
423
424 2009-07-11  Richard Sandiford  <rdsandiford@googlemail.com>
425
426         PR testsuite/40699
427         PR testsuite/40707
428         PR testsuite/40709
429         * testsuite/lib/libffi-dg.exp: Revert 2009-07-02, 2009-07-01 and
430         2009-06-30 commits.
431
432 2009-07-01  Richard Sandiford  <r.sandiford@uk.ibm.com>
433
434         * testsuite/lib/libffi-dg.exp (libffi-init): Set ld_library_path
435         to "" before adding paths.  (This reinstates an assignment that
436         was removed by my 2009-06-30 commit, but changes the initial
437         value from "." to "".)
438
439 2009-07-01  H.J. Lu  <hongjiu.lu@intel.com>
440
441         PR testsuite/40601
442         * testsuite/lib/libffi-dg.exp (libffi-init): Properly set
443         gccdir.  Adjust ld_library_path for gcc only if gccdir isn't
444         empty.
445
446 2009-06-30  Richard Sandiford  <r.sandiford@uk.ibm.com>
447
448         * testsuite/lib/libffi-dg.exp (libffi-init): Don't add "."
449         to ld_library_path.  Use add_path.  Add just find_libgcc_s
450         to ld_library_path, not every libgcc multilib directory.
451
452 2009-06-16  Wim Lewis  <wiml@hhhh.org>
453
454         * src/powerpc/ffi.c: Avoid clobbering cr3 and cr4, which are
455         supposed to be callee-saved.
456         * src/powerpc/sysv.S (small_struct_return_value): Fix overrun of
457         return buffer for odd-size structs.
458
459 2009-06-16  Andreas Tobler  <a.tobler@schweiz.org>
460
461         PR libffi/40444
462         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Add
463         allow_stack_execute for Darwin.
464
465 2009-06-16  Andrew Haley  <aph@redhat.com>
466
467         * configure.ac (TARGETDIR): Add missing blank lines.
468         * configure: Regenerate.
469
470 2009-06-16  Andrew Haley  <aph@redhat.com>
471
472         * testsuite/libffi.call/cls_align_sint64.c,
473         testsuite/libffi.call/cls_align_uint64.c,
474         testsuite/libffi.call/cls_longdouble_va.c,
475         testsuite/libffi.call/cls_ulonglong.c,
476         testsuite/libffi.call/return_ll1.c,
477         testsuite/libffi.call/stret_medium2.c: Fix printf format
478         specifiers.
479         * testsuite/libffi.call/ffitest.h,
480         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRIuLL): Define.
481
482 2009-06-15  Andrew Haley  <aph@redhat.com>
483
484         * testsuite/libffi.call/err_bad_typedef.c: xfail everywhere.
485         * testsuite/libffi.call/err_bad_abi.c: Likewise.
486
487 2009-06-12  Andrew Haley  <aph@redhat.com>
488
489         * Makefile.am: Remove info_TEXINFOS.
490
491 2009-06-12  Andrew Haley  <aph@redhat.com>
492
493         * ChangeLog.libffi: testsuite/libffi.call/cls_align_sint64.c,
494         testsuite/libffi.call/cls_align_uint64.c,
495         testsuite/libffi.call/cls_ulonglong.c,
496         testsuite/libffi.call/return_ll1.c,
497         testsuite/libffi.call/stret_medium2.c: Fix printf format
498         specifiers.
499         testsuite/libffi.special/unwindtest.cc: include stdint.h.
500
501 2009-06-11  Timothy Wall  <twall@users.sf.net>
502
503         * Makefile.am,
504         configure.ac,
505         include/ffi.h.in,
506         include/ffi_common.h,
507         src/closures.c,
508         src/dlmalloc.c,
509         src/x86/ffi.c,
510         src/x86/ffitarget.h,
511         src/x86/win64.S (new),
512         README: Added win64 support (mingw or MSVC)
513         * Makefile.in,
514         include/Makefile.in,
515         man/Makefile.in,
516         testsuite/Makefile.in,
517         configure,
518         aclocal.m4: Regenerated
519         * ltcf-c.sh: properly escape cygwin/w32 path
520         * man/ffi_call.3: Clarify size requirements for return value.
521         * src/x86/ffi64.c: Fix filename in comment.
522         * src/x86/win32.S: Remove unused extern.
523
524         * testsuite/libffi.call/closure_fn0.c,
525         testsuite/libffi.call/closure_fn1.c,
526         testsuite/libffi.call/closure_fn2.c,
527         testsuite/libffi.call/closure_fn3.c,
528         testsuite/libffi.call/closure_fn4.c,
529         testsuite/libffi.call/closure_fn5.c,
530         testsuite/libffi.call/closure_fn6.c,
531         testsuite/libffi.call/closure_stdcall.c,
532         testsuite/libffi.call/cls_12byte.c,
533         testsuite/libffi.call/cls_16byte.c,
534         testsuite/libffi.call/cls_18byte.c,
535         testsuite/libffi.call/cls_19byte.c,
536         testsuite/libffi.call/cls_1_1byte.c,
537         testsuite/libffi.call/cls_20byte.c,
538         testsuite/libffi.call/cls_20byte1.c,
539         testsuite/libffi.call/cls_24byte.c,
540         testsuite/libffi.call/cls_2byte.c,
541         testsuite/libffi.call/cls_3_1byte.c,
542         testsuite/libffi.call/cls_3byte1.c,
543         testsuite/libffi.call/cls_3byte2.c,
544         testsuite/libffi.call/cls_4_1byte.c,
545         testsuite/libffi.call/cls_4byte.c,
546         testsuite/libffi.call/cls_5_1_byte.c,
547         testsuite/libffi.call/cls_5byte.c,
548         testsuite/libffi.call/cls_64byte.c,
549         testsuite/libffi.call/cls_6_1_byte.c,
550         testsuite/libffi.call/cls_6byte.c,
551         testsuite/libffi.call/cls_7_1_byte.c,
552         testsuite/libffi.call/cls_7byte.c,
553         testsuite/libffi.call/cls_8byte.c,
554         testsuite/libffi.call/cls_9byte1.c,
555         testsuite/libffi.call/cls_9byte2.c,
556         testsuite/libffi.call/cls_align_double.c,
557         testsuite/libffi.call/cls_align_float.c,
558         testsuite/libffi.call/cls_align_longdouble.c,
559         testsuite/libffi.call/cls_align_longdouble_split.c,
560         testsuite/libffi.call/cls_align_longdouble_split2.c,
561         testsuite/libffi.call/cls_align_pointer.c,
562         testsuite/libffi.call/cls_align_sint16.c,
563         testsuite/libffi.call/cls_align_sint32.c,
564         testsuite/libffi.call/cls_align_sint64.c,
565         testsuite/libffi.call/cls_align_uint16.c,
566         testsuite/libffi.call/cls_align_uint32.c,
567         testsuite/libffi.call/cls_align_uint64.c,
568         testsuite/libffi.call/cls_dbls_struct.c,
569         testsuite/libffi.call/cls_double.c,
570         testsuite/libffi.call/cls_double_va.c,
571         testsuite/libffi.call/cls_float.c,
572         testsuite/libffi.call/cls_longdouble.c,
573         testsuite/libffi.call/cls_longdouble_va.c,
574         testsuite/libffi.call/cls_multi_schar.c,
575         testsuite/libffi.call/cls_multi_sshort.c,
576         testsuite/libffi.call/cls_multi_sshortchar.c,
577         testsuite/libffi.call/cls_multi_uchar.c,
578         testsuite/libffi.call/cls_multi_ushort.c,
579         testsuite/libffi.call/cls_multi_ushortchar.c,
580         testsuite/libffi.call/cls_pointer.c,
581         testsuite/libffi.call/cls_pointer_stack.c,
582         testsuite/libffi.call/cls_schar.c,
583         testsuite/libffi.call/cls_sint.c,
584         testsuite/libffi.call/cls_sshort.c,
585         testsuite/libffi.call/cls_uchar.c,
586         testsuite/libffi.call/cls_uint.c,
587         testsuite/libffi.call/cls_ulonglong.c,
588         testsuite/libffi.call/cls_ushort.c,
589         testsuite/libffi.call/err_bad_abi.c,
590         testsuite/libffi.call/err_bad_typedef.c,
591         testsuite/libffi.call/float2.c,
592         testsuite/libffi.call/huge_struct.c,
593         testsuite/libffi.call/nested_struct.c,
594         testsuite/libffi.call/nested_struct1.c,
595         testsuite/libffi.call/nested_struct10.c,
596         testsuite/libffi.call/nested_struct2.c,
597         testsuite/libffi.call/nested_struct3.c,
598         testsuite/libffi.call/nested_struct4.c,
599         testsuite/libffi.call/nested_struct5.c,
600         testsuite/libffi.call/nested_struct6.c,
601         testsuite/libffi.call/nested_struct7.c,
602         testsuite/libffi.call/nested_struct8.c,
603         testsuite/libffi.call/nested_struct9.c,
604         testsuite/libffi.call/problem1.c,
605         testsuite/libffi.call/return_ldl.c,
606         testsuite/libffi.call/return_ll1.c,
607         testsuite/libffi.call/stret_large.c,
608         testsuite/libffi.call/stret_large2.c,
609         testsuite/libffi.call/stret_medium.c,
610         testsuite/libffi.call/stret_medium2.c,
611         testsuite/libffi.special/unwindtest.cc: use ffi_closure_alloc instead
612         of checking for MMAP.  Use intptr_t instead of long casts.
613
614 2009-06-11  Kaz Kojima  <kkojima@gcc.gnu.org>
615
616         * testsuite/libffi.call/cls_longdouble_va.c: Add xfail sh*-*-linux-*.
617         * testsuite/libffi.call/err_bad_abi.c: Add xfail sh*-*-*.
618         * testsuite/libffi.call/err_bad_typedef.c: Likewise.
619
620 2009-06-09  Andrew Haley  <aph@redhat.com>
621
622         * src/x86/freebsd.S: Add missing file.
623
624 2009-06-08  Andrew Haley  <aph@redhat.com>
625
626         Import from libffi 3.0.8:
627
628         * doc/libffi.texi: New file.
629         * doc/libffi.info: Likewise.
630         * doc/stamp-vti: Likewise.
631         * man/Makefile.am: New file.
632         * man/ffi_call.3: New file.
633
634         * Makefile.am (EXTRA_DIST): Add src/x86/darwin64.S,
635         src/dlmalloc.c.
636         (nodist_libffi_la_SOURCES): Add X86_FREEBSD.
637
638         * configure.ac: Bump version to 3.0.8.
639         parisc*-*-linux*: Add.
640         i386-*-freebsd* | i386-*-openbsd*: Add.
641         powerpc-*-beos*: Add.
642         AM_CONDITIONAL X86_FREEBSD: Add.
643         AC_CONFIG_FILES: Add man/Makefile.
644
645         * include/ffi.h.in (FFI_FN): Change void (*)() to void (*)(void).
646
647 2009-06-08  Andrew Haley  <aph@redhat.com>
648
649         * README: Import from libffi 3.0.8.
650
651 2009-06-08  Andrew Haley  <aph@redhat.com>
652
653         * testsuite/libffi.call/err_bad_abi.c: Add xfails.
654         * testsuite/libffi.call/cls_longdouble_va.c: Add xfails.
655         * testsuite/libffi.call/cls_dbls_struct.c: Add xfail x86_64-*-linux-*.
656         * testsuite/libffi.call/err_bad_typedef.c: Add xfails.
657
658         * testsuite/libffi.call/stret_medium2.c: Add __UNUSED__ to args.
659         * testsuite/libffi.call/stret_medium.c: Likewise.
660         * testsuite/libffi.call/stret_large2.c: Likewise.
661         * testsuite/libffi.call/stret_large.c:  Likewise.
662
663 2008-12-26  Timothy Wall  <twall@users.sf.net>
664
665         * testsuite/libffi.call/cls_longdouble.c,
666         testsuite/libffi.call/cls_longdouble_va.c,
667         testsuite/libffi.call/cls_align_longdouble.c,
668         testsuite/libffi.call/cls_align_longdouble_split.c,
669         testsuite/libffi.call/cls_align_longdouble_split2.c: mark expected
670         failures on x86_64 cygwin/mingw.
671
672 2008-12-22  Timothy Wall  <twall@users.sf.net>
673
674         * testsuite/libffi.call/closure_fn0.c,
675         testsuite/libffi.call/closure_fn1.c,
676         testsuite/libffi.call/closure_fn2.c,
677         testsuite/libffi.call/closure_fn3.c,
678         testsuite/libffi.call/closure_fn4.c,
679         testsuite/libffi.call/closure_fn5.c,
680         testsuite/libffi.call/closure_fn6.c,
681         testsuite/libffi.call/closure_loc_fn0.c,
682         testsuite/libffi.call/closure_stdcall.c,
683         testsuite/libffi.call/cls_align_pointer.c,
684         testsuite/libffi.call/cls_pointer.c,
685         testsuite/libffi.call/cls_pointer_stack.c: use portable cast from
686         pointer to integer (intptr_t).
687         * testsuite/libffi.call/cls_longdouble.c: disable for win64.
688
689 2008-07-24  Anthony Green  <green@redhat.com>
690
691         * testsuite/libffi.call/cls_dbls_struct.c,
692         testsuite/libffi.call/cls_double_va.c,
693         testsuite/libffi.call/cls_longdouble.c,
694         testsuite/libffi.call/cls_longdouble_va.c,
695         testsuite/libffi.call/cls_pointer.c,
696         testsuite/libffi.call/cls_pointer_stack.c,
697         testsuite/libffi.call/err_bad_abi.c: Clean up failures from
698         compiler warnings.
699
700 2008-03-04  Anthony Green  <green@redhat.com>
701             Blake Chaffin
702             hos@tamanegi.org
703
704         * testsuite/libffi.call/cls_align_longdouble_split2.c
705           testsuite/libffi.call/cls_align_longdouble_split.c
706           testsuite/libffi.call/cls_dbls_struct.c
707           testsuite/libffi.call/cls_double_va.c
708           testsuite/libffi.call/cls_longdouble.c
709           testsuite/libffi.call/cls_longdouble_va.c
710           testsuite/libffi.call/cls_pointer.c
711           testsuite/libffi.call/cls_pointer_stack.c
712           testsuite/libffi.call/err_bad_abi.c
713           testsuite/libffi.call/err_bad_typedef.c
714           testsuite/libffi.call/stret_large2.c
715           testsuite/libffi.call/stret_large.c
716           testsuite/libffi.call/stret_medium2.c
717           testsuite/libffi.call/stret_medium.c: New tests from Apple.
718
719 2009-06-05  Andrew Haley  <aph@redhat.com>
720
721         * src/x86/ffitarget.h, src/x86/ffi.c: Merge stdcall changes from
722         libffi.
723
724 2009-06-04  Andrew Haley  <aph@redhat.com>
725
726         * src/x86/ffitarget.h, src/x86/win32.S, src/x86/ffi.c: Back out
727         stdcall changes.
728
729 2008-02-26  Anthony Green  <green@redhat.com>
730             Thomas Heller  <theller@ctypes.org>
731
732         * src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C
733         comment.
734
735 2008-02-03  Timothy Wall  <twall@users.sf.net>
736
737         * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return
738           offset based on code pointer, not data pointer.
739
740 2008-01-31  Timothy Wall <twall@users.sf.net>
741
742         * testsuite/libffi.call/closure_stdcall.c: Add test for stdcall
743         closures.
744         * src/x86/ffitarget.h: Increase size of trampoline for stdcall
745         closures.
746         * src/x86/win32.S: Add assembly for stdcall closure.
747         * src/x86/ffi.c: Initialize stdcall closure trampoline.
748
749 2009-06-04  Andrew Haley  <aph@redhat.com>
750
751         * include/ffi.h.in: Change void (*)() to void (*)(void).
752         * src/x86/ffi.c: Likewise.
753
754 2009-06-04  Andrew Haley  <aph@redhat.com>
755
756         * src/powerpc/ppc_closure.S: Insert licence header.
757         * src/powerpc/linux64_closure.S: Likewise.
758         * src/m68k/sysv.S: Likewise.
759
760         * src/sh64/ffi.c: Change void (*)() to void (*)(void).
761         * src/powerpc/ffi.c: Likewise.
762         * src/powerpc/ffi_darwin.c: Likewise.
763         * src/m32r/ffi.c: Likewise.
764         * src/sh64/ffi.c: Likewise.
765         * src/x86/ffi64.c: Likewise.
766         * src/alpha/ffi.c: Likewise.
767         * src/alpha/osf.S: Likewise.
768         * src/frv/ffi.c: Likewise.
769         * src/s390/ffi.c: Likewise.
770         * src/pa/ffi.c: Likewise.
771         * src/pa/hpux32.S: Likewise.
772         * src/ia64/unix.S: Likewise.
773         * src/ia64/ffi.c: Likewise.
774         * src/sparc/ffi.c: Likewise.
775         * src/mips/ffi.c: Likewise.
776         * src/sh/ffi.c: Likewise.
777
778 2008-02-15  David Daney  <ddaney@avtrex.com>
779
780         * src/mips/ffi.c (USE__BUILTIN___CLEAR_CACHE):
781         Define (conditionally), and use it to include cachectl.h.
782         (ffi_prep_closure_loc): Fix cache flushing.
783         * src/mips/ffitarget.h (_ABIN32, _ABI64, _ABIO32): Define.
784
785 2009-06-04  Andrew Haley  <aph@redhat.com>
786
787         include/ffi.h.in,
788         src/arm/ffitarget.h,
789         src/arm/ffi.c,
790         src/arm/sysv.S,
791         src/powerpc/ffitarget.h,
792         src/closures.c,
793         src/sh64/ffitarget.h,
794         src/sh64/ffi.c,
795         src/sh64/sysv.S,
796         src/types.c,
797         src/x86/ffi64.c,
798         src/x86/ffitarget.h,
799         src/x86/win32.S,
800         src/x86/darwin.S,
801         src/x86/ffi.c,
802         src/x86/sysv.S,
803         src/x86/unix64.S,
804         src/alpha/ffitarget.h,
805         src/alpha/ffi.c,
806         src/alpha/osf.S,
807         src/m68k/ffitarget.h,
808         src/frv/ffitarget.h,
809         src/frv/ffi.c,
810         src/s390/ffitarget.h,
811         src/s390/sysv.S,
812         src/cris/ffitarget.h,
813         src/pa/linux.S,
814         src/pa/ffitarget.h,
815         src/pa/ffi.c,
816         src/raw_api.c,
817         src/ia64/ffitarget.h,
818         src/ia64/unix.S,
819         src/ia64/ffi.c,
820         src/ia64/ia64_flags.h,
821         src/java_raw_api.c,
822         src/debug.c,
823         src/sparc/v9.S,
824         src/sparc/ffitarget.h,
825         src/sparc/ffi.c,
826         src/sparc/v8.S,
827         src/mips/ffitarget.h,
828         src/mips/n32.S,
829         src/mips/o32.S,
830         src/mips/ffi.c,
831         src/prep_cif.c,
832         src/sh/ffitarget.h,
833         src/sh/ffi.c,
834         src/sh/sysv.S: Update license text.
835
836 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
837
838         * src/x86/win32.S (_ffi_closure_STDCALL):  New function.
839         (.eh_frame):  Add FDE for it.
840
841 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
842
843         * configure.ac:  Also check if assembler supports pc-relative
844         relocs on X86_WIN32 targets.
845         * configure:  Regenerate.
846         * src/x86/win32.S (ffi_prep_args):  Declare extern, not global.
847         (_ffi_call_SYSV):  Add missing function type symbol .def and
848         add EH markup labels.
849         (_ffi_call_STDCALL):  Likewise.
850         (_ffi_closure_SYSV):  Likewise.
851         (_ffi_closure_raw_SYSV):  Likewise.
852         (.eh_frame):  Add hand-crafted EH data.
853
854 2009-04-09  Jakub Jelinek  <jakub@redhat.com>
855
856         * testsuite/lib/libffi-dg.exp: Change copyright header to refer to
857         version 3 of the GNU General Public License and to point readers
858         at the COPYING3 file and the FSF's license web page.
859         * testsuite/libffi.call/call.exp: Likewise.
860         * testsuite/libffi.special/special.exp: Likewise.
861
862 2009-03-01  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
863
864         * configure: Regenerate.
865
866 2008-12-18  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
867
868         PR libffi/26048
869         * configure.ac (HAVE_AS_X86_PCREL): New test.
870         * configure: Regenerate.
871         * fficonfig.h.in: Regenerate.
872         * src/x86/sysv.S [!FFI_NO_RAW_API]: Precalculate
873         RAW_CLOSURE_CIF_OFFSET, RAW_CLOSURE_FUN_OFFSET,
874         RAW_CLOSURE_USER_DATA_OFFSET for the Solaris 10/x86 assembler.
875         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
876         * src/x86/unix64.S (.Lstore_table): Move to .text section.
877         (.Lload_table): Likewise.
878         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
879
880 2008-12-18  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
881
882         * configure: Regenerate.
883
884 2008-11-21  Eric Botcazou  <ebotcazou@adacore.com>
885
886         * src/sparc/ffi.c (ffi_prep_cif_machdep): Add support for
887         signed/unsigned int8/16 return values.
888         * src/sparc/v8.S (ffi_call_v8): Likewise.
889         (ffi_closure_v8): Likewise.
890
891 2008-09-26  Peter O'Gorman  <pogma@thewrittenword.com>
892             Steve Ellcey  <sje@cup.hp.com>
893
894         * configure: Regenerate for new libtool.
895         * Makefile.in: Ditto.
896         * include/Makefile.in: Ditto.
897         * aclocal.m4: Ditto.
898
899 2008-08-25  Andreas Tobler  <a.tobler@schweiz.org>
900
901         * src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and
902         FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum.
903         Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT.
904         Adjust copyright notice.
905         * src/powerpc/ffi.c: Add two new flags to indicate if we have one
906         register or two register to use for FFI_SYSV structs.
907         (ffi_prep_cif_machdep): Pass the right register flag introduced above.
908         (ffi_closure_helper_SYSV): Fix the return type for
909         FFI_SYSV_TYPE_SMALL_STRUCT. Comment.
910         Adjust copyright notice.
911
912 2008-07-16  Kaz Kojima  <kkojima@gcc.gnu.org>
913
914         * src/sh/ffi.c (ffi_prep_closure_loc): Turn INSN into an unsigned
915         int.
916
917 2008-06-17  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
918
919         * configure: Regenerate.
920         * include/Makefile.in: Regenerate.
921         * testsuite/Makefile.in: Regenerate.
922
923 2008-06-07  Joseph Myers  <joseph@codesourcery.com>
924
925         * configure.ac (parisc*-*-linux*, powerpc-*-sysv*,
926         powerpc-*-beos*): Remove.
927         * configure: Regenerate.
928
929 2008-05-09  Julian Brown  <julian@codesourcery.com>
930
931         * Makefile.am (LTLDFLAGS): New.
932         (libffi_la_LDFLAGS): Use above.
933         * Makefile.in: Regenerate.
934
935 2008-04-18  Paolo Bonzini  <bonzini@gnu.org>
936
937         PR bootstrap/35457
938         * aclocal.m4: Regenerate.
939         * configure: Regenerate.
940
941 2008-03-26  Kaz Kojima  <kkojima@gcc.gnu.org>
942
943         * src/sh/sysv.S: Add .note.GNU-stack on Linux.
944         * src/sh64/sysv.S: Likewise.
945
946 2008-03-26  Daniel Jacobowitz  <dan@debian.org>
947
948         * src/arm/sysv.S: Fix ARM comment marker.
949
950 2008-03-26  Jakub Jelinek  <jakub@redhat.com>
951
952         * src/alpha/osf.S: Add .note.GNU-stack on Linux.
953         * src/s390/sysv.S: Likewise.
954         * src/powerpc/ppc_closure.S: Likewise.
955         * src/powerpc/sysv.S: Likewise.
956         * src/x86/unix64.S: Likewise.
957         * src/x86/sysv.S: Likewise.
958         * src/sparc/v8.S: Likewise.
959         * src/sparc/v9.S: Likewise.
960         * src/m68k/sysv.S: Likewise.
961         * src/arm/sysv.S: Likewise.
962
963 2008-03-16  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
964
965         * aclocal.m4: Regenerate.
966         * configure: Likewise.
967         * Makefile.in: Likewise.
968         * include/Makefile.in: Likewise.
969         * testsuite/Makefile.in: Likewise.
970
971 2008-02-12  Bjoern Koenig  <bkoenig@alpha-tierchen.de>
972             Andreas Tobler  <a.tobler@schweiz.org>
973
974         * configure.ac: Add amd64-*-freebsd* target.
975         * configure: Regenerate.
976
977 2008-01-30  H.J. Lu  <hongjiu.lu@intel.com>
978
979         PR libffi/34612
980         * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when
981         returning struct.
982
983         * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer"
984         tests.
985
986 2008-01-24  David Edelsohn  <edelsohn@gnu.org>
987
988         * configure: Regenerate.
989
990 2008-01-06  Andreas Tobler  <a.tobler@schweiz.org>
991
992         * src/x86/ffi.c (ffi_prep_cif_machdep): Fix thinko.
993
994 2008-01-05  Andreas Tobler  <a.tobler@schweiz.org>
995
996         PR testsuite/32843
997         * src/x86/ffi.c (ffi_prep_cif_machdep): Add code for
998         signed/unsigned int8/16 for X86_DARWIN.
999         Updated copyright info.
1000         Handle one and two byte structs with special cif->flags.
1001         * src/x86/ffitarget.h: Add special types for one and two byte structs.
1002         Updated copyright info.
1003         * src/x86/darwin.S (ffi_call_SYSV): Rewrite to use a jump table like
1004         sysv.S
1005         Remove code to pop args from the stack after call.
1006         Special-case signed/unsigned for int8/16, one and two byte structs.
1007         (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1008         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1009         FFI_TYPE_SINT32.
1010         Updated copyright info.
1011
1012 2007-12-08  David Daney  <ddaney@avtrex.com>
1013
1014         * src/mips/n32.S (ffi_call_N32):  Replace dadd with ADDU, dsub with
1015         SUBU, add with ADDU and use smaller code sequences.
1016
1017 2007-12-07  David Daney  <ddaney@avtrex.com>
1018
1019         * src/mips/ffi.c (ffi_prep_cif_machdep): Handle long double return
1020         type.
1021
1022 2007-12-06  David Daney  <ddaney@avtrex.com>
1023
1024         * include/ffi.h.in (FFI_SIZEOF_JAVA_RAW): Define if not already
1025         defined.
1026         (ffi_java_raw): New typedef.
1027         (ffi_java_raw_call, ffi_java_ptrarray_to_raw,
1028         ffi_java_raw_to_ptrarray): Change parameter types from ffi_raw to
1029         ffi_java_raw.
1030         (ffi_java_raw_closure) : Same.
1031         (ffi_prep_java_raw_closure, ffi_prep_java_raw_closure_loc): Change
1032         parameter types.
1033         * src/java_raw_api.c (ffi_java_raw_size):  Replace FFI_SIZEOF_ARG with
1034         FFI_SIZEOF_JAVA_RAW.
1035         (ffi_java_raw_to_ptrarray): Change type of raw to ffi_java_raw.
1036         Replace FFI_SIZEOF_ARG with FFI_SIZEOF_JAVA_RAW. Use
1037         sizeof(ffi_java_raw) for alignment calculations.
1038         (ffi_java_ptrarray_to_raw): Same.
1039         (ffi_java_rvalue_to_raw): Add special handling for FFI_TYPE_POINTER
1040         if FFI_SIZEOF_JAVA_RAW == 4.
1041         (ffi_java_raw_to_rvalue): Same.
1042         (ffi_java_raw_call): Change type of raw to ffi_java_raw.
1043         (ffi_java_translate_args): Same.
1044         (ffi_prep_java_raw_closure_loc, ffi_prep_java_raw_closure): Change
1045         parameter types.
1046         * src/mips/ffitarget.h (FFI_SIZEOF_JAVA_RAW): Define for N32 ABI.
1047
1048 2007-12-06  David Daney  <ddaney@avtrex.com>
1049
1050         * src/mips/n32.S (ffi_closure_N32): Use 64-bit add instruction on
1051         pointer values.
1052
1053 2007-12-01  Andreas Tobler  <a.tobler@schweiz.org>
1054
1055         PR libffi/31937
1056         * src/powerpc/ffitarget.h: Introduce new ABI FFI_LINUX_SOFT_FLOAT.
1057         Add local FFI_TYPE_UINT128 to handle soft-float long-double-128.
1058         * src/powerpc/ffi.c: Distinguish between __NO_FPRS__ and not and
1059         set the NUM_FPR_ARG_REGISTERS according to.
1060         Add support for potential soft-float support under hard-float
1061         architecture.
1062         (ffi_prep_args_SYSV): Set NUM_FPR_ARG_REGISTERS to 0 in case of
1063         FFI_LINUX_SOFT_FLOAT, handle float, doubles and long-doubles according
1064         to the FFI_LINUX_SOFT_FLOAT ABI.
1065         (ffi_prep_cif_machdep): Likewise.
1066         (ffi_closure_helper_SYSV): Likewise.
1067         * src/powerpc/ppc_closure.S: Make sure not to store float/double
1068         on archs where __NO_FPRS__ is true.
1069         Add FFI_TYPE_UINT128 support.
1070         * src/powerpc/sysv.S: Add support for soft-float long-double-128.
1071         Adjust copyright notice.
1072
1073 2007-11-25  Andreas Tobler  <a.tobler@schweiz.org>
1074
1075         * src/closures.c: Move defintion of MAYBE_UNUSED from here to ...
1076         * include/ffi_common.h: ... here.
1077         Update copyright.
1078
1079 2007-11-17  Andreas Tobler  <a.tobler@schweiz.org>
1080
1081         * src/powerpc/sysv.S: Load correct cr to compare if we have long double.
1082         * src/powerpc/linux64.S: Likewise.
1083         * src/powerpc/ffi.c: Add a comment to show which part goes into cr6.
1084         * testsuite/libffi.call/return_ldl.c: New test.
1085
1086 2007-09-04    <aph@redhat.com>
1087
1088         * src/arm/sysv.S (UNWIND): New.
1089         (Whole file): Conditionally compile unwinder directives.
1090         * src/arm/sysv.S: Add unwinder directives.
1091
1092         * src/arm/ffi.c (ffi_prep_args): Align structs by at least 4 bytes.
1093         Only treat r0 as a struct address if we're actually returning a
1094         struct by address.
1095         Only copy the bytes that are actually within a struct.
1096         (ffi_prep_cif_machdep): A Composite Type not larger than 4 bytes
1097         is returned in r0, not passed by address.
1098         (ffi_call): Allocate a word-sized temporary for the case where
1099         a composite is returned in r0.
1100         (ffi_prep_incoming_args_SYSV): Align as necessary.
1101
1102 2007-08-05  Steven Newbury  <s_j_newbury@yahoo.co.uk>
1103
1104         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Use __clear_cache instead of
1105         directly using the sys_cacheflush syscall.
1106
1107 2007-07-27  Andrew Haley  <aph@redhat.com>
1108
1109         * src/arm/sysv.S (ffi_closure_SYSV): Add soft-float.
1110
1111 2007-09-03  Maciej W. Rozycki  <macro@linux-mips.org>
1112
1113         * Makefile.am: Unify MIPS_IRIX and MIPS_LINUX into MIPS.
1114         * configure.ac: Likewise.
1115         * Makefile.in: Regenerate.
1116         * include/Makefile.in: Likewise.
1117         * testsuite/Makefile.in: Likewise.
1118         * configure: Likewise.
1119
1120 2007-08-24  David Daney  <ddaney@avtrex.com>
1121
1122         * testsuite/libffi.call/return_sl.c: New test.
1123
1124 2007-08-10  David Daney  <ddaney@avtrex.com>
1125
1126         * testsuite/libffi.call/cls_multi_ushort.c,
1127         testsuite/libffi.call/cls_align_uint16.c,
1128         testsuite/libffi.call/nested_struct1.c,
1129         testsuite/libffi.call/nested_struct3.c,
1130         testsuite/libffi.call/cls_7_1_byte.c,
1131         testsuite/libffi.call/nested_struct5.c,
1132         testsuite/libffi.call/cls_double.c,
1133         testsuite/libffi.call/nested_struct7.c,
1134         testsuite/libffi.call/cls_sint.c,
1135         testsuite/libffi.call/nested_struct9.c,
1136         testsuite/libffi.call/cls_20byte1.c,
1137         testsuite/libffi.call/cls_multi_sshortchar.c,
1138         testsuite/libffi.call/cls_align_sint64.c,
1139         testsuite/libffi.call/cls_3byte2.c,
1140         testsuite/libffi.call/cls_multi_schar.c,
1141         testsuite/libffi.call/cls_multi_uchar.c,
1142         testsuite/libffi.call/cls_19byte.c,
1143         testsuite/libffi.call/cls_9byte1.c,
1144         testsuite/libffi.call/cls_align_float.c,
1145         testsuite/libffi.call/closure_fn1.c,
1146         testsuite/libffi.call/problem1.c,
1147         testsuite/libffi.call/closure_fn3.c,
1148         testsuite/libffi.call/cls_sshort.c,
1149         testsuite/libffi.call/closure_fn5.c,
1150         testsuite/libffi.call/cls_align_double.c,
1151         testsuite/libffi.call/nested_struct.c,
1152         testsuite/libffi.call/cls_2byte.c,
1153         testsuite/libffi.call/nested_struct10.c,
1154         testsuite/libffi.call/cls_4byte.c,
1155         testsuite/libffi.call/cls_6byte.c,
1156         testsuite/libffi.call/cls_8byte.c,
1157         testsuite/libffi.call/cls_multi_sshort.c,
1158         testsuite/libffi.call/cls_align_sint16.c,
1159         testsuite/libffi.call/cls_align_uint32.c,
1160         testsuite/libffi.call/cls_20byte.c,
1161         testsuite/libffi.call/cls_float.c,
1162         testsuite/libffi.call/nested_struct2.c,
1163         testsuite/libffi.call/cls_5_1_byte.c,
1164         testsuite/libffi.call/nested_struct4.c,
1165         testsuite/libffi.call/cls_24byte.c,
1166         testsuite/libffi.call/nested_struct6.c,
1167         testsuite/libffi.call/cls_64byte.c,
1168         testsuite/libffi.call/nested_struct8.c,
1169         testsuite/libffi.call/cls_uint.c,
1170         testsuite/libffi.call/cls_multi_ushortchar.c,
1171         testsuite/libffi.call/cls_schar.c,
1172         testsuite/libffi.call/cls_uchar.c,
1173         testsuite/libffi.call/cls_align_uint64.c,
1174         testsuite/libffi.call/cls_ulonglong.c,
1175         testsuite/libffi.call/cls_align_longdouble.c,
1176         testsuite/libffi.call/cls_1_1byte.c,
1177         testsuite/libffi.call/cls_12byte.c,
1178         testsuite/libffi.call/cls_3_1byte.c,
1179         testsuite/libffi.call/cls_3byte1.c,
1180         testsuite/libffi.call/cls_4_1byte.c,
1181         testsuite/libffi.call/cls_6_1_byte.c,
1182         testsuite/libffi.call/cls_16byte.c,
1183         testsuite/libffi.call/cls_18byte.c,
1184         testsuite/libffi.call/closure_fn0.c,
1185         testsuite/libffi.call/cls_9byte2.c,
1186         testsuite/libffi.call/closure_fn2.c,
1187         testsuite/libffi.call/closure_fn4.c,
1188         testsuite/libffi.call/cls_ushort.c,
1189         testsuite/libffi.call/closure_fn6.c,
1190         testsuite/libffi.call/cls_5byte.c,
1191         testsuite/libffi.call/cls_align_pointer.c,
1192         testsuite/libffi.call/cls_7byte.c,
1193         testsuite/libffi.call/cls_align_sint32.c,
1194         testsuite/libffi.special/unwindtest_ffi_call.cc,
1195         testsuite/libffi.special/unwindtest.cc: Remove xfail for mips64*-*-*.
1196
1197 2007-08-10  David Daney  <ddaney@avtrex.com>
1198
1199         PR libffi/28313
1200         * configure.ac: Don't treat mips64 as a special case.
1201         * Makefile.am (nodist_libffi_la_SOURCES): Add n32.S.
1202         * configure: Regenerate
1203         * Makefile.in: Ditto.
1204         * fficonfig.h.in: Ditto.
1205         * src/mips/ffitarget.h (REG_L, REG_S, SUBU, ADDU, SRL, LI): Indent.
1206         (LA, EH_FRAME_ALIGN, FDE_ADDR_BYTES): New preprocessor macros.
1207         (FFI_DEFAULT_ABI): Set for n64 case.
1208         (FFI_CLOSURES, FFI_TRAMPOLINE_SIZE): Define for n32 and n64 cases.
1209         * src/mips/n32.S (ffi_call_N32): Add debug macros and labels for FDE.
1210         (ffi_closure_N32): New function.
1211         (.eh_frame): New section
1212         * src/mips/o32.S: Clean up comments.
1213         (ffi_closure_O32): Pass ffi_closure parameter in $12.
1214         * src/mips/ffi.c: Use FFI_MIPS_N32 instead of
1215         _MIPS_SIM == _ABIN32 throughout.
1216         (FFI_MIPS_STOP_HERE): New, use in place of
1217         ffi_stop_here.
1218         (ffi_prep_args): Use unsigned long to hold pointer values.  Rewrite
1219         to support n32/n64 ABIs.
1220         (calc_n32_struct_flags): Rewrite.
1221         (calc_n32_return_struct_flags): Remove unused variable.  Reverse
1222         position of flag bits.
1223         (ffi_prep_cif_machdep): Rewrite n32 portion.
1224         (ffi_call): Enable for n64.  Add special handling for small structure
1225         return values.
1226         (ffi_prep_closure_loc): Add n32 and n64 support.
1227         (ffi_closure_mips_inner_O32): Add cast to silence warning.
1228         (copy_struct_N32, ffi_closure_mips_inner_N32): New functions.
1229
1230 2007-08-08  David Daney  <ddaney@avtrex.com>
1231
1232         * testsuite/libffi.call/ffitest.h (ffi_type_mylong): Remove definition.
1233         * testsuite/libffi.call/cls_align_uint16.c (main): Use correct type
1234         specifiers.
1235         * testsuite/libffi.call/nested_struct1.c (main): Ditto.
1236         * testsuite/libffi.call/cls_sint.c (main): Ditto.
1237         * testsuite/libffi.call/nested_struct9.c (main): Ditto.
1238         * testsuite/libffi.call/cls_20byte1.c (main): Ditto.
1239         * testsuite/libffi.call/cls_9byte1.c (main): Ditto.
1240         * testsuite/libffi.call/closure_fn1.c (main): Ditto.
1241         * testsuite/libffi.call/closure_fn3.c (main): Ditto.
1242         * testsuite/libffi.call/return_dbl2.c (main): Ditto.
1243         * testsuite/libffi.call/cls_sshort.c (main): Ditto.
1244         * testsuite/libffi.call/return_fl3.c (main): Ditto.
1245         * testsuite/libffi.call/closure_fn5.c (main): Ditto.
1246         * testsuite/libffi.call/nested_struct.c (main): Ditto.
1247         * testsuite/libffi.call/nested_struct10.c (main): Ditto.
1248         * testsuite/libffi.call/return_ll1.c (main): Ditto.
1249         * testsuite/libffi.call/cls_8byte.c (main): Ditto.
1250         * testsuite/libffi.call/cls_align_uint32.c (main): Ditto.
1251         * testsuite/libffi.call/cls_align_sint16.c (main): Ditto.
1252         * testsuite/libffi.call/cls_20byte.c (main): Ditto.
1253         * testsuite/libffi.call/nested_struct2.c (main): Ditto.
1254         * testsuite/libffi.call/cls_24byte.c (main): Ditto.
1255         * testsuite/libffi.call/nested_struct6.c (main): Ditto.
1256         * testsuite/libffi.call/cls_uint.c (main): Ditto.
1257         * testsuite/libffi.call/cls_12byte.c (main): Ditto.
1258         * testsuite/libffi.call/cls_16byte.c (main): Ditto.
1259         * testsuite/libffi.call/closure_fn0.c (main): Ditto.
1260         * testsuite/libffi.call/cls_9byte2.c (main): Ditto.
1261         * testsuite/libffi.call/closure_fn2.c (main): Ditto.
1262         * testsuite/libffi.call/return_dbl1.c (main): Ditto.
1263         * testsuite/libffi.call/closure_fn4.c (main): Ditto.
1264         * testsuite/libffi.call/closure_fn6.c (main): Ditto.
1265         * testsuite/libffi.call/cls_align_sint32.c (main): Ditto.
1266
1267 2007-08-07  Andrew Haley  <aph@redhat.com>
1268
1269         * src/x86/sysv.S (ffi_closure_raw_SYSV): Fix typo in previous
1270         checkin.
1271
1272 2007-08-06  Andrew Haley  <aph@redhat.com>
1273
1274         PR testsuite/32843
1275         * src/x86/sysv.S (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1276         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1277         FFI_TYPE_SINT32.
1278
1279 2007-08-02  David Daney  <ddaney@avtrex.com>
1280
1281         * testsuite/libffi.call/return_ul.c (main): Define return type as
1282         ffi_arg.  Use proper printf conversion specifier.
1283
1284 2007-07-30  Andrew Haley  <aph@redhat.com>
1285
1286         PR testsuite/32843
1287         * src/x86/ffi.c (ffi_prep_cif_machdep): in x86 case, add code for
1288         signed/unsigned int8/16.
1289         * src/x86/sysv.S (ffi_call_SYSV): Rewrite to:
1290         Use a jump table.
1291         Remove code to pop args from the stack after call.
1292         Special-case signed/unsigned int8/16.
1293         * testsuite/libffi.call/return_sc.c (main): Revert.
1294
1295 2007-07-26  Richard Guenther  <rguenther@suse.de>
1296
1297         PR testsuite/32843
1298         * testsuite/libffi.call/return_sc.c (main): Verify call
1299         result as signed char, not ffi_arg.
1300
1301 2007-07-16  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1302
1303         * configure.ac (i?86-*-solaris2.1[0-9]): Set TARGET to X86_64.
1304         * configure: Regenerate.
1305
1306 2007-07-11  David Daney  <ddaney@avtrex.com>
1307
1308         * src/mips/ffi.c: Don't include sys/cachectl.h.
1309         (ffi_prep_closure_loc): Use __builtin___clear_cache() instead of
1310         cacheflush().
1311
1312 2007-05-18  Aurelien Jarno  <aurelien@aurel32.net>
1313
1314         * src/arm/ffi.c (ffi_prep_closure_loc): Renamed and ajusted
1315         from (ffi_prep_closure): ... this.
1316         (FFI_INIT_TRAMPOLINE): Adjust.
1317
1318 2005-12-31  Phil Blundell  <pb@reciva.com>
1319
1320         * src/arm/ffi.c (ffi_prep_incoming_args_SYSV,
1321         ffi_closure_SYSV_inner, ffi_prep_closure): New, add closure support.
1322         * src/arm/sysv.S(ffi_closure_SYSV): Likewise.
1323         * src/arm/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1324         (FFI_CLOSURES): Enable closure support.
1325
1326 2007-07-03  Andrew Haley  <aph@hedges.billgatliff.com>
1327
1328         * testsuite/libffi.call/cls_multi_ushort.c,
1329         testsuite/libffi.call/cls_align_uint16.c,
1330         testsuite/libffi.call/nested_struct1.c,
1331         testsuite/libffi.call/nested_struct3.c,
1332         testsuite/libffi.call/cls_7_1_byte.c,
1333         testsuite/libffi.call/cls_double.c,
1334         testsuite/libffi.call/nested_struct5.c,
1335         testsuite/libffi.call/nested_struct7.c,
1336         testsuite/libffi.call/cls_sint.c,
1337         testsuite/libffi.call/nested_struct9.c,
1338         testsuite/libffi.call/cls_20byte1.c,
1339         testsuite/libffi.call/cls_multi_sshortchar.c,
1340         testsuite/libffi.call/cls_align_sint64.c,
1341         testsuite/libffi.call/cls_3byte2.c,
1342         testsuite/libffi.call/cls_multi_schar.c,
1343         testsuite/libffi.call/cls_multi_uchar.c,
1344         testsuite/libffi.call/cls_19byte.c,
1345         testsuite/libffi.call/cls_9byte1.c,
1346         testsuite/libffi.call/cls_align_float.c,
1347         testsuite/libffi.call/closure_fn1.c,
1348         testsuite/libffi.call/problem1.c,
1349         testsuite/libffi.call/closure_fn3.c,
1350         testsuite/libffi.call/cls_sshort.c,
1351         testsuite/libffi.call/closure_fn5.c,
1352         testsuite/libffi.call/cls_align_double.c,
1353         testsuite/libffi.call/cls_2byte.c,
1354         testsuite/libffi.call/nested_struct.c,
1355         testsuite/libffi.call/nested_struct10.c,
1356         testsuite/libffi.call/cls_4byte.c,
1357         testsuite/libffi.call/cls_6byte.c,
1358         testsuite/libffi.call/cls_8byte.c,
1359         testsuite/libffi.call/cls_multi_sshort.c,
1360         testsuite/libffi.call/cls_align_uint32.c,
1361         testsuite/libffi.call/cls_align_sint16.c,
1362         testsuite/libffi.call/cls_float.c,
1363         testsuite/libffi.call/cls_20byte.c,
1364         testsuite/libffi.call/cls_5_1_byte.c,
1365         testsuite/libffi.call/nested_struct2.c,
1366         testsuite/libffi.call/cls_24byte.c,
1367         testsuite/libffi.call/nested_struct4.c,
1368         testsuite/libffi.call/nested_struct6.c,
1369         testsuite/libffi.call/cls_64byte.c,
1370         testsuite/libffi.call/nested_struct8.c,
1371         testsuite/libffi.call/cls_uint.c,
1372         testsuite/libffi.call/cls_multi_ushortchar.c,
1373         testsuite/libffi.call/cls_schar.c,
1374         testsuite/libffi.call/cls_uchar.c,
1375         testsuite/libffi.call/cls_align_uint64.c,
1376         testsuite/libffi.call/cls_ulonglong.c,
1377         testsuite/libffi.call/cls_align_longdouble.c,
1378         testsuite/libffi.call/cls_1_1byte.c,
1379         testsuite/libffi.call/cls_12byte.c,
1380         testsuite/libffi.call/cls_3_1byte.c,
1381         testsuite/libffi.call/cls_3byte1.c,
1382         testsuite/libffi.call/cls_4_1byte.c,
1383         testsuite/libffi.call/cls_6_1_byte.c,
1384         testsuite/libffi.call/cls_16byte.c,
1385         testsuite/libffi.call/cls_18byte.c,
1386         testsuite/libffi.call/closure_fn0.c,
1387         testsuite/libffi.call/cls_9byte2.c,
1388         testsuite/libffi.call/closure_fn2.c,
1389         testsuite/libffi.call/closure_fn4.c,
1390         testsuite/libffi.call/cls_ushort.c,
1391         testsuite/libffi.call/closure_fn6.c,
1392         testsuite/libffi.call/cls_5byte.c,
1393         testsuite/libffi.call/cls_align_pointer.c,
1394         testsuite/libffi.call/cls_7byte.c,
1395         testsuite/libffi.call/cls_align_sint32.c,
1396         testsuite/libffi.special/unwindtest_ffi_call.cc,
1397         testsuite/libffi.special/unwindtest.cc: Enable for ARM.
1398
1399 2007-07-05  H.J. Lu  <hongjiu.lu@intel.com>
1400
1401         * aclocal.m4: Regenerated.
1402
1403 2007-06-02  Paolo Bonzini  <bonzini@gnu.org>
1404
1405         * configure: Regenerate.
1406
1407 2007-05-23  Steve Ellcey  <sje@cup.hp.com>
1408
1409         * Makefile.in: Regenerate.
1410         * configure: Regenerate.
1411         * aclocal.m4: Regenerate.
1412         * include/Makefile.in: Regenerate.
1413         * testsuite/Makefile.in: Regenerate.
1414
1415 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1416
1417         * src/m68k/ffi.c (ffi_prep_incoming_args_SYSV,
1418         ffi_closure_SYSV_inner,ffi_prep_closure): New, add closure support.
1419         * src/m68k/sysv.S(ffi_closure_SYSV,ffi_closure_struct_SYSV): Likewise.
1420         * src/m68k/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1421         (FFI_CLOSURES): Enable closure support.
1422
1423 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1424
1425         * configure.ac (HAVE_AS_CFI_PSEUDO_OP): New test.
1426         * configure: Regenerate.
1427         * fficonfig.h.in: Regenerate.
1428         * src/m68k/sysv.S (CFI_STARTPROC,CFI_ENDPROC,
1429         CFI_OFFSET,CFI_DEF_CFA): New macros.
1430         (ffi_call_SYSV): Add callframe annotation.
1431
1432 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1433
1434         * src/m68k/ffi.c (ffi_prep_args,ffi_prep_cif_machdep): Fix
1435         numerous test suite failures.
1436         * src/m68k/sysv.S (ffi_call_SYSV): Likewise.
1437
1438 2007-04-11  Paolo Bonzini  <bonzini@gnu.org>
1439
1440         * Makefile.am (EXTRA_DIST): Bring up to date.
1441         * Makefile.in: Regenerate.
1442         * src/frv/eabi.S: Remove RCS keyword.
1443
1444 2007-04-06  Richard Henderson  <rth@redhat.com>
1445
1446         * configure.ac: Tidy target case.
1447         (HAVE_LONG_DOUBLE): Allow the target to override.
1448         * configure: Regenerate.
1449         * include/ffi.h.in: Don't define ffi_type_foo if
1450         LIBFFI_HIDE_BASIC_TYPES is defined.
1451         (ffi_type_longdouble): If not HAVE_LONG_DOUBLE, define
1452         to ffi_type_double.
1453         * types.c (LIBFFI_HIDE_BASIC_TYPES): Define.
1454         (FFI_TYPEDEF, ffi_type_void): Mark the data const.
1455         (ffi_type_longdouble): Special case for Alpha.  Don't define
1456         if long double == double.
1457
1458         * src/alpha/ffi.c (FFI_TYPE_LONGDOUBLE): Assert unique value.
1459         (ffi_prep_cif_machdep): Handle it as the 128-bit type.
1460         (ffi_call, ffi_closure_osf_inner): Likewise.
1461         (ffi_closure_osf_inner): Likewise.  Mark hidden.
1462         (ffi_call_osf, ffi_closure_osf): Mark hidden.
1463         * src/alpha/ffitarget.h (FFI_LAST_ABI): Tidy definition.
1464         * src/alpha/osf.S (ffi_call_osf, ffi_closure_osf): Mark hidden.
1465         (load_table): Handle 128-bit long double.
1466
1467         * testsuite/libffi.call/float4.c: Add -mieee for alpha.
1468
1469 2007-04-06  Tom Tromey  <tromey@redhat.com>
1470
1471         PR libffi/31491:
1472         * README: Fixed bug in example.
1473
1474 2007-04-03  Jakub Jelinek  <jakub@redhat.com>
1475
1476         * src/closures.c: Include sys/statfs.h.
1477         (_GNU_SOURCE): Define on Linux.
1478         (FFI_MMAP_EXEC_SELINUX): Define.
1479         (selinux_enabled): New variable.
1480         (selinux_enabled_check): New function.
1481         (is_selinux_enabled): Define.
1482         (dlmmap): Use it.
1483
1484 2007-03-24  Uros Bizjak  <ubizjak@gmail.com>
1485
1486         * testsuite/libffi.call/return_fl2.c (return_fl): Mark as static.
1487         Use 'volatile float sum' to create sum of floats to avoid false
1488         negative due to excess precision on ix86 targets.
1489         (main): Ditto.
1490
1491 2007-03-08  Alexandre Oliva  <aoliva@redhat.com>
1492
1493         * src/powerpc/ffi.c (flush_icache): Fix left-over from previous
1494         patch.
1495         (ffi_prep_closure_loc): Remove unneeded casts.  Add needed ones.
1496
1497 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1498
1499         * include/ffi.h.in (ffi_closure_alloc, ffi_closure_free): New.
1500         (ffi_prep_closure_loc): New.
1501         (ffi_prep_raw_closure_loc): New.
1502         (ffi_prep_java_raw_closure_loc): New.
1503         * src/closures.c: New file.
1504         * src/dlmalloc.c [FFI_MMAP_EXEC_WRIT] (struct malloc_segment):
1505         Replace sflags with exec_offset.
1506         [FFI_MMAP_EXEC_WRIT] (mmap_exec_offset, add_segment_exec_offset,
1507         sub_segment_exec_offset): New macros.
1508         (get_segment_flags, set_segment_flags, check_segment_merge): New
1509         macros.
1510         (is_mmapped_segment, is_extern_segment): Use get_segment_flags.
1511         (add_segment, sys_alloc, create_mspace, create_mspace_with_base,
1512         destroy_mspace): Use new macros.
1513         (sys_alloc): Silence warning.
1514         * Makefile.am (libffi_la_SOURCES): Add src/closures.c.
1515         * Makefile.in: Rebuilt.
1516         * src/prep_cif [FFI_CLOSURES] (ffi_prep_closure): Implement in
1517         terms of ffi_prep_closure_loc.
1518         * src/raw_api.c (ffi_prep_raw_closure_loc): Renamed and adjusted
1519         from...
1520         (ffi_prep_raw_closure): ... this.  Re-implement in terms of the
1521         renamed version.
1522         * src/java_raw_api (ffi_prep_java_raw_closure_loc): Renamed and
1523         adjusted from...
1524         (ffi_prep_java_raw_closure): ... this.  Re-implement in terms of
1525         the renamed version.
1526         * src/alpha/ffi.c (ffi_prep_closure_loc): Renamed from
1527         (ffi_prep_closure): ... this.
1528         * src/pa/ffi.c: Likewise.
1529         * src/cris/ffi.c: Likewise.  Adjust.
1530         * src/frv/ffi.c: Likewise.
1531         * src/ia64/ffi.c: Likewise.
1532         * src/mips/ffi.c: Likewise.
1533         * src/powerpc/ffi_darwin.c: Likewise.
1534         * src/s390/ffi.c: Likewise.
1535         * src/sh/ffi.c: Likewise.
1536         * src/sh64/ffi.c: Likewise.
1537         * src/sparc/ffi.c: Likewise.
1538         * src/x86/ffi64.c: Likewise.
1539         * src/x86/ffi.c: Likewise.
1540         (FFI_INIT_TRAMPOLINE): Adjust.
1541         (ffi_prep_raw_closure_loc): Renamed and adjusted from...
1542         (ffi_prep_raw_closure): ... this.
1543         * src/powerpc/ffi.c (ffi_prep_closure_loc): Renamed from
1544         (ffi_prep_closure): ... this.
1545         (flush_icache): Adjust.
1546
1547 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1548
1549         * src/dlmalloc.c: New file, imported version 2.8.3 of Doug
1550         Lea's malloc.
1551
1552 2007-03-01  Brooks Moses  <brooks.moses@codesourcery.com>
1553
1554         * Makefile.am: Add dummy install-pdf target.
1555         * Makefile.in: Regenerate
1556
1557 2007-02-13  Andreas Krebbel  <krebbel1@de.ibm.com>
1558
1559         * src/s390/ffi.c (ffi_prep_args, ffi_prep_cif_machdep,
1560         ffi_closure_helper_SYSV): Add long double handling.
1561
1562 2007-02-02  Jakub Jelinek  <jakub@redhat.com>
1563
1564         * src/powerpc/linux64.S (ffi_call_LINUX64): Move restore of r2
1565         immediately after bctrl instruction.
1566
1567 2007-01-18  Alexandre Oliva  <aoliva@redhat.com>
1568
1569         * Makefile.am (all-recursive, install-recursive,
1570         mostlyclean-recursive, clean-recursive, distclean-recursive,
1571         maintainer-clean-recursive): Add missing targets.
1572         * Makefile.in: Rebuilt.
1573
1574 2006-12-14  Andreas Tobler  <a.tobler@schweiz.org>
1575
1576         * configure.ac: Add TARGET for x86_64-*-darwin*.
1577         * Makefile.am (nodist_libffi_la_SOURCES): Add rules for 64-bit sources
1578         for X86_DARWIN.
1579         * src/x86/ffitarget.h: Set trampoline size for x86_64-*-darwin*.
1580         * src/x86/darwin64.S: New file for x86_64-*-darwin* support.
1581         * configure: Regenerate.
1582         * Makefile.in: Regenerate.
1583         * include/Makefile.in: Regenerate.
1584         * testsuite/Makefile.in: Regenerate.
1585         * testsuite/libffi.special/unwindtest_ffi_call.cc: New test case for
1586         ffi_call only.
1587
1588 2006-12-13  Andreas Tobler <a.tobler@schweiz.org>
1589
1590         * aclocal.m4: Regenerate with aclocal -I .. as written in the
1591         Makefile.am.
1592
1593 2006-10-31  Geoffrey Keating  <geoffk@apple.com>
1594
1595         * src/powerpc/ffi_darwin.c (darwin_adjust_aggregate_sizes): New.
1596         (ffi_prep_cif_machdep): Call darwin_adjust_aggregate_sizes for
1597         Darwin.
1598         * testsuite/libffi.call/nested_struct4.c: Remove Darwin XFAIL.
1599         * testsuite/libffi.call/nested_struct6.c: Remove Darwin XFAIL.
1600
1601 2006-10-10  Paolo Bonzini  <bonzini@gnu.org>
1602             Sandro Tolaini  <tolaini@libero.it>
1603
1604         * configure.ac [i*86-*-darwin*]: Set X86_DARWIN symbol and
1605         conditional.
1606         * configure: Regenerated.
1607         * Makefile.am (nodist_libffi_la_SOURCES) [X86_DARWIN]: New case.
1608         (EXTRA_DIST): Add src/x86/darwin.S.
1609         * Makefile.in: Regenerated.
1610         * include/Makefile.in: Regenerated.
1611         * testsuite/Makefile.in: Regenerated.
1612
1613         * src/x86/ffi.c (ffi_prep_cif_machdep) [X86_DARWIN]: Treat like
1614         X86_WIN32, and additionally align stack to 16 bytes.
1615         * src/x86/darwin.S: New, based on sysv.S.
1616         * src/prep_cif.c (ffi_prep_cif) [X86_DARWIN]: Align > 8-byte structs.
1617
1618 2006-09-12  David Daney  <ddaney@avtrex.com>
1619
1620         PR libffi/23935
1621         * include/Makefile.am: Install both ffi.h and ffitarget.h in
1622         $(libdir)/gcc/$(target_alias)/$(gcc_version)/include.
1623         * aclocal.m4: Regenerated for automake 1.9.6.
1624         * Makefile.in: Regenerated.
1625         * include/Makefile.in: Regenerated.
1626         * testsuite/Makefile.in: Regenerated.
1627
1628 2006-08-17  Andreas Tobler  <a.tobler@schweiz.ch>
1629
1630         * include/ffi_common.h (struct): Revert accidental commit.
1631
1632 2006-08-15  Andreas Tobler  <a.tobler@schweiz.ch>
1633
1634         * include/ffi_common.h: Remove lint directives.
1635         * include/ffi.h.in: Likewise.
1636
1637 2006-07-25  Torsten Schoenfeld  <kaffeetisch@gmx.de>
1638
1639         * include/ffi.h.in (ffi_type_ulong, ffi_type_slong): Define correctly
1640         for 32-bit architectures.
1641         * testsuite/libffi.call/return_ul.c: New test case.
1642
1643 2006-07-19  David Daney  <ddaney@avtrex.com>
1644
1645         * testsuite/libffi.call/closure_fn6.c: Remove xfail for mips,
1646         xfail remains for mips64.
1647
1648 2006-05-23  Carlos O'Donell  <carlos@codesourcery.com>
1649
1650         * Makefile.am: Add install-html target. Add install-html to .PHONY
1651         * Makefile.in: Regenerate.
1652         * aclocal.m4: Regenerate.
1653         * include/Makefile.in: Regenerate.
1654         * testsuite/Makefile.in: Regenerate.
1655
1656 2006-05-18  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1657
1658         * pa/ffi.c (ffi_prep_args_pa32): Load floating point arguments from
1659         stack slot.
1660
1661 2006-04-22  Andreas Tobler  <a.tobler@schweiz.ch>
1662
1663         * README: Remove notice about 'Crazy Comments'.
1664         * src/debug.c: Remove lint directives. Cleanup white spaces.
1665         * src/java_raw_api.c: Likewise.
1666         * src/prep_cif.c: Likewise.
1667         * src/raw_api.c: Likewise.
1668         * src/ffitest.c: Delete. No longer needed, all test cases migrated
1669         to the testsuite.
1670         * src/arm/ffi.c: Remove lint directives.
1671         * src/m32r/ffi.c: Likewise.
1672         * src/pa/ffi.c: Likewise.
1673         * src/powerpc/ffi.c: Likewise.
1674         * src/powerpc/ffi_darwin.c: Likewise.
1675         * src/sh/ffi.c: Likewise.
1676         * src/sh64/ffi.c: Likewise.
1677         * src/x86/ffi.c: Likewise.
1678         * testsuite/libffi.call/float2.c: Likewise.
1679         * testsuite/libffi.call/promotion.c: Likewise.
1680         * testsuite/libffi.call/struct1.c: Likewise.
1681
1682 2006-04-13  Andreas Tobler  <a.tobler@schweiz.ch>
1683
1684         * src/pa/hpux32.S: Correct unwind offset calculation for
1685         ffi_closure_pa32.
1686         * src/pa/linux.S: Likewise.
1687
1688 2006-04-12  James E Wilson  <wilson@specifix.com>
1689
1690         PR libgcj/26483
1691         * src/ia64/ffi.c (stf_spill, ldf_fill): Rewrite as macros.
1692         (hfa_type_load): Call stf_spill.
1693         (hfa_type_store): Call ldf_fill.
1694         (ffi_call): Adjust calls to above routines.  Add local temps for
1695         macro result.
1696
1697 2006-04-10  Matthias Klose  <doko@debian.org>
1698
1699         * testsuite/lib/libffi-dg.exp (libffi-init): Recognize multilib
1700         directory names containing underscores.
1701
1702 2006-04-07  James E Wilson  <wilson@specifix.com>
1703
1704         * testsuite/libffi.call/float4.c: New testcase.
1705
1706 2006-04-05  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1707             Andreas Tobler  <a.tobler@schweiz.ch>
1708
1709         * Makefile.am: Add PA_HPUX port.
1710         * Makefile.in: Regenerate.
1711         * include/Makefile.in: Likewise.
1712         * testsuite/Makefile.in: Likewise.
1713         * configure.ac: Add PA_HPUX rules.
1714         * configure: Regenerate.
1715         * src/pa/ffitarget.h: Rename linux target to PA_LINUX.
1716         Add PA_HPUX and PA64_HPUX.
1717         Rename FFI_LINUX ABI to FFI_PA32 ABI.
1718         (FFI_TRAMPOLINE_SIZE): Define for 32-bit HP-UX targets.
1719         (FFI_TYPE_SMALL_STRUCT2): Define.
1720         (FFI_TYPE_SMALL_STRUCT4): Likewise.
1721         (FFI_TYPE_SMALL_STRUCT8): Likewise.
1722         (FFI_TYPE_SMALL_STRUCT3): Redefine.
1723         (FFI_TYPE_SMALL_STRUCT5): Likewise.
1724         (FFI_TYPE_SMALL_STRUCT6): Likewise.
1725         (FFI_TYPE_SMALL_STRUCT7): Likewise.
1726         * src/pa/ffi.c (ROUND_DOWN): Delete.
1727         (fldw, fstw, fldd, fstd): Use '__asm__'.
1728         (ffi_struct_type): Add support for FFI_TYPE_SMALL_STRUCT2,
1729         FFI_TYPE_SMALL_STRUCT4 and FFI_TYPE_SMALL_STRUCT8.
1730         (ffi_prep_args_LINUX): Rename to ffi_prep_args_pa32. Update comment.
1731         Simplify incrementing of stack slot variable. Change type of local
1732         'n' to unsigned int.
1733         (ffi_size_stack_LINUX): Rename to ffi_size_stack_pa32. Handle long
1734         double on PA_HPUX.
1735         (ffi_prep_cif_machdep): Likewise.
1736         (ffi_call): Likewise.
1737         (ffi_closure_inner_LINUX): Rename to ffi_closure_inner_pa32. Change
1738         return type to ffi_status. Simplify incrementing of stack slot
1739         variable. Only copy floating point argument registers when PA_LINUX
1740         is true. Reformat debug statement.
1741         Add support for FFI_TYPE_SMALL_STRUCT2, FFI_TYPE_SMALL_STRUCT4 and
1742         FFI_TYPE_SMALL_STRUCT8.
1743         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Add 'extern' to
1744         declaration.
1745         (ffi_prep_closure): Make linux trampoline conditional on PA_LINUX.
1746         Add nops to cache flush.  Add trampoline for PA_HPUX.
1747         * src/pa/hpux32.S: New file.
1748         * src/pa/linux.S (ffi_call_LINUX): Rename to ffi_call_pa32. Rename
1749         ffi_prep_args_LINUX to ffi_prep_args_pa32.
1750         Localize labels. Add support for 2, 4 and 8-byte small structs. Handle
1751         unaligned destinations in 3, 5, 6 and 7-byte small structs. Order
1752         argument type checks so that common argument types appear first.
1753         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Rename
1754         ffi_closure_inner_LINUX to ffi_closure_inner_pa32.
1755
1756 2006-03-24  Alan Modra  <amodra@bigpond.net.au>
1757
1758         * src/powerpc/ffitarget.h (enum ffi_abi): Add FFI_LINUX.  Default
1759         for 32-bit using IBM extended double format.  Fix FFI_LAST_ABI.
1760         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Handle linux variant of
1761         FFI_TYPE_LONGDOUBLE.
1762         (ffi_prep_args64): Assert using IBM extended double.
1763         (ffi_prep_cif_machdep): Don't munge FFI_TYPE_LONGDOUBLE type.
1764         Handle FFI_LINUX FFI_TYPE_LONGDOUBLE return and args.
1765         (ffi_call): Handle FFI_LINUX.
1766         (ffi_closure_helper_SYSV): Non FFI_LINUX long double return needs
1767         gpr3 return pointer as for struct return.  Handle FFI_LINUX
1768         FFI_TYPE_LONGDOUBLE return and args.  Don't increment "nf"
1769         unnecessarily.
1770         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Load both f1 and f2
1771         for FFI_TYPE_LONGDOUBLE.  Move epilogue insns into case table.
1772         Don't use r6 as pointer to results, instead use sp offset.  Don't
1773         make a special call to load lr with case table address, instead
1774         use offset from previous call.
1775         * src/powerpc/sysv.S (ffi_call_SYSV): Save long double return.
1776         * src/powerpc/linux64.S (ffi_call_LINUX64): Simplify long double
1777         return.
1778
1779 2006-03-15  Kaz Kojima  <kkojima@gcc.gnu.org>
1780
1781         * src/sh64/ffi.c (ffi_prep_cif_machdep): Handle float arguments
1782         passed with FP registers correctly.
1783         (ffi_closure_helper_SYSV): Likewise.
1784         * src/sh64/sysv.S: Likewise.
1785
1786 2006-03-01  Andreas Tobler  <a.tobler@schweiz.ch>
1787
1788         * testsuite/libffi.special/unwindtest.cc (closure_test_fn): Mark cif,
1789         args and userdata unused.
1790         (closure_test_fn1): Mark cif and userdata unused.
1791         (main): Remove unused res.
1792
1793 2006-02-28  Andreas Tobler  <a.tobler@schweiz.ch>
1794
1795         * testsuite/libffi.call/call.exp: Adjust FSF address. Add test runs for
1796         -O2, -O3, -Os and the warning flags -W -Wall.
1797         * testsuite/libffi.special/special.exp: Likewise.
1798         * testsuite/libffi.call/ffitest.h: Add an __UNUSED__ macro to mark
1799         unused parameter unused for gcc or else do nothing.
1800         * testsuite/libffi.special/ffitestcxx.h: Likewise.
1801         * testsuite/libffi.call/cls_12byte.c (cls_struct_12byte_gn): Mark cif
1802         and userdata unused.
1803         * testsuite/libffi.call/cls_16byte.c (cls_struct_16byte_gn): Likewise.
1804         * testsuite/libffi.call/cls_18byte.c (cls_struct_18byte_gn): Likewise.
1805         * testsuite/libffi.call/cls_19byte.c (cls_struct_19byte_gn): Likewise.
1806         * testsuite/libffi.call/cls_1_1byte.c (cls_struct_1_1byte_gn): Likewise.
1807         * testsuite/libffi.call/cls_20byte.c (cls_struct_20byte_gn): Likewise.
1808         * testsuite/libffi.call/cls_20byte1.c (cls_struct_20byte_gn): Likewise.
1809         * testsuite/libffi.call/cls_24byte.c (cls_struct_24byte_gn): Likewise.
1810         * testsuite/libffi.call/cls_2byte.c (cls_struct_2byte_gn): Likewise.
1811         * testsuite/libffi.call/cls_3_1byte.c (cls_struct_3_1byte_gn): Likewise.
1812         * testsuite/libffi.call/cls_3byte1.c (cls_struct_3byte_gn): Likewise.
1813         * testsuite/libffi.call/cls_3byte2.c (cls_struct_3byte_gn1): Likewise.
1814         * testsuite/libffi.call/cls_4_1byte.c (cls_struct_4_1byte_gn): Likewise.
1815         * testsuite/libffi.call/cls_4byte.c (cls_struct_4byte_gn): Likewise.
1816         * testsuite/libffi.call/cls_5_1_byte.c (cls_struct_5byte_gn): Likewise.
1817         * testsuite/libffi.call/cls_5byte.c (cls_struct_5byte_gn): Likewise.
1818         * testsuite/libffi.call/cls_64byte.c (cls_struct_64byte_gn): Likewise.
1819         * testsuite/libffi.call/cls_6_1_byte.c (cls_struct_6byte_gn): Likewise.
1820         * testsuite/libffi.call/cls_6byte.c (cls_struct_6byte_gn): Likewise.
1821         * testsuite/libffi.call/cls_7_1_byte.c (cls_struct_7byte_gn): Likewise.
1822         * testsuite/libffi.call/cls_7byte.c (cls_struct_7byte_gn): Likewise.
1823         * testsuite/libffi.call/cls_8byte.c (cls_struct_8byte_gn): Likewise.
1824         * testsuite/libffi.call/cls_9byte1.c (cls_struct_9byte_gn): Likewise.
1825         * testsuite/libffi.call/cls_9byte2.c (cls_struct_9byte_gn): Likewise.
1826         * testsuite/libffi.call/cls_align_double.c (cls_struct_align_gn):
1827         Likewise.
1828         * testsuite/libffi.call/cls_align_float.c (cls_struct_align_gn):
1829         Likewise.
1830         * testsuite/libffi.call/cls_align_longdouble.c (cls_struct_align_gn):
1831         Likewise.
1832         * testsuite/libffi.call/cls_align_pointer.c (cls_struct_align_fn): Cast
1833         void* to avoid compiler warning.
1834         (main): Likewise.
1835         (cls_struct_align_gn): Mark cif and userdata unused.
1836         * testsuite/libffi.call/cls_align_sint16.c (cls_struct_align_gn):
1837         Likewise.
1838         * testsuite/libffi.call/cls_align_sint32.c (cls_struct_align_gn):
1839         Likewise.
1840         * testsuite/libffi.call/cls_align_sint64.c (cls_struct_align_gn):
1841         Likewise.
1842         * testsuite/libffi.call/cls_align_uint16.c (cls_struct_align_gn):
1843         Likewise.
1844         * testsuite/libffi.call/cls_align_uint32.c (cls_struct_align_gn):
1845         Likewise.
1846         * testsuite/libffi.call/cls_double.c (cls_ret_double_fn): Likewise.
1847         * testsuite/libffi.call/cls_float.c (cls_ret_float_fn): Likewise.
1848         * testsuite/libffi.call/cls_multi_schar.c (test_func_gn): Mark cif and
1849         data unused.
1850         (main): Cast res_call to silence gcc.
1851         * testsuite/libffi.call/cls_multi_sshort.c (test_func_gn): Mark cif and
1852         data unused.
1853         (main): Cast res_call to silence gcc.
1854         * testsuite/libffi.call/cls_multi_sshortchar.c (test_func_gn): Mark cif
1855         and data unused.
1856         (main): Cast res_call to silence gcc.
1857         * testsuite/libffi.call/cls_multi_uchar.c (test_func_gn): Mark cif and
1858         data unused.
1859         (main): Cast res_call to silence gcc.
1860         * testsuite/libffi.call/cls_multi_ushort.c (test_func_gn): Mark cif and
1861         data unused.
1862         (main): Cast res_call to silence gcc.
1863         * testsuite/libffi.call/cls_multi_ushortchar.c (test_func_gn): Mark cif
1864         and data unused.
1865         (main): Cast res_call to silence gcc.
1866         * testsuite/libffi.call/cls_schar.c (cls_ret_schar_fn): Mark cif and
1867         userdata unused.
1868         (cls_ret_schar_fn): Cast printf parameter to silence gcc.
1869         * testsuite/libffi.call/cls_sint.c (cls_ret_sint_fn): Mark cif and
1870         userdata unused.
1871         (cls_ret_sint_fn): Cast printf parameter to silence gcc.
1872         * testsuite/libffi.call/cls_sshort.c (cls_ret_sshort_fn): Mark cif and
1873         userdata unused.
1874         (cls_ret_sshort_fn): Cast printf parameter to silence gcc.
1875         * testsuite/libffi.call/cls_uchar.c (cls_ret_uchar_fn):  Mark cif and
1876         userdata unused.
1877         (cls_ret_uchar_fn): Cast printf parameter to silence gcc.
1878         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Mark cif and
1879         userdata unused.
1880         (cls_ret_uint_fn): Cast printf parameter to silence gcc.
1881         * testsuite/libffi.call/cls_ulonglong.c (cls_ret_ulonglong_fn): Mark cif
1882         and userdata unused.
1883         * testsuite/libffi.call/cls_ushort.c (cls_ret_ushort_fn): Mark cif and
1884         userdata unused.
1885         (cls_ret_ushort_fn): Cast printf parameter to silence gcc.
1886         * testsuite/libffi.call/float.c (floating): Remove unused parameter e.
1887         * testsuite/libffi.call/float1.c (main): Remove unused variable i.
1888         Cleanup white spaces.
1889         * testsuite/libffi.call/negint.c (checking): Remove unused variable i.
1890         * testsuite/libffi.call/nested_struct.c (cls_struct_combined_gn): Mark
1891         cif and userdata unused.
1892         * testsuite/libffi.call/nested_struct1.c (cls_struct_combined_gn):
1893         Likewise.
1894         * testsuite/libffi.call/nested_struct10.c (B_gn): Likewise.
1895         * testsuite/libffi.call/nested_struct2.c (B_fn): Adjust printf
1896         formatters to silence gcc.
1897         (B_gn): Mark cif and userdata unused.
1898         * testsuite/libffi.call/nested_struct3.c (B_gn): Mark cif and userdata
1899         unused.
1900         * testsuite/libffi.call/nested_struct4.c: Mention related PR.
1901         (B_gn): Mark cif and userdata unused.
1902         * testsuite/libffi.call/nested_struct5.c (B_gn): Mark cif and userdata
1903         unused.
1904         * testsuite/libffi.call/nested_struct6.c: Mention related PR.
1905         (B_gn): Mark cif and userdata unused.
1906         * testsuite/libffi.call/nested_struct7.c (B_gn): Mark cif and userdata
1907         unused.
1908         * testsuite/libffi.call/nested_struct8.c (B_gn): Likewise.
1909         * testsuite/libffi.call/nested_struct9.c (B_gn): Likewise.
1910         * testsuite/libffi.call/problem1.c (stub): Likewise.
1911         * testsuite/libffi.call/pyobjc-tc.c (main): Cast the result to silence
1912         gcc.
1913         * testsuite/libffi.call/return_fl2.c (return_fl): Add the note mentioned
1914         in the last commit for this test case in the test case itself.
1915         * testsuite/libffi.call/closure_fn0.c (closure_test_fn0): Mark cif as
1916         unused.
1917         * testsuite/libffi.call/closure_fn1.c (closure_test_fn1): Likewise.
1918         * testsuite/libffi.call/closure_fn2.c (closure_test_fn2): Likewise.
1919         * testsuite/libffi.call/closure_fn3.c (closure_test_fn3): Likewise.
1920         * testsuite/libffi.call/closure_fn4.c (closure_test_fn0): Likewise.
1921         * testsuite/libffi.call/closure_fn5.c (closure_test_fn5): Likewise.
1922         * testsuite/libffi.call/closure_fn6.c (closure_test_fn0): Likewise.
1923
1924 2006-02-22  Kaz Kojima  <kkojima@gcc.gnu.org>
1925
1926         * src/sh/sysv.S: Fix register numbers in the FDE for
1927         ffi_closure_SYSV.
1928
1929 2006-02-20  Andreas Tobler  <a.tobler@schweiz.ch>
1930
1931         * testsuite/libffi.call/return_fl2.c (return_fl): Remove static
1932         declaration to avoid a false negative on ix86. See PR323.
1933
1934 2006-02-18  Kaz Kojima  <kkojima@gcc.gnu.org>
1935
1936         * src/sh/ffi.c (ffi_closure_helper_SYSV): Remove unused variable
1937         and cast integer to void * if needed.  Update the pointer to
1938         the FP register saved area correctly.
1939
1940 2006-02-17  Andreas Tobler  <a.tobler@schweiz.ch>
1941
1942         * testsuite/libffi.call/nested_struct6.c: XFAIL this test until PR25630
1943         is fixed.
1944         * testsuite/libffi.call/nested_struct4.c: Likewise.
1945
1946 2006-02-16  Andreas Tobler  <a.tobler@schweiz.ch>
1947
1948         * testsuite/libffi.call/return_dbl.c: New test case.
1949         * testsuite/libffi.call/return_dbl1.c: Likewise.
1950         * testsuite/libffi.call/return_dbl2.c: Likewise.
1951         * testsuite/libffi.call/return_fl.c: Likewise.
1952         * testsuite/libffi.call/return_fl1.c: Likewise.
1953         * testsuite/libffi.call/return_fl2.c: Likewise.
1954         * testsuite/libffi.call/return_fl3.c: Likewise.
1955         * testsuite/libffi.call/closure_fn6.c: Likewise.
1956
1957         * testsuite/libffi.call/nested_struct2.c: Remove ffi_type_mylong
1958         definition.
1959         * testsuite/libffi.call/ffitest.h: Add ffi_type_mylong definition
1960         here to be used by other test cases too.
1961
1962         * testsuite/libffi.call/nested_struct10.c: New test case.
1963         * testsuite/libffi.call/nested_struct9.c: Likewise.
1964         * testsuite/libffi.call/nested_struct8.c: Likewise.
1965         * testsuite/libffi.call/nested_struct7.c: Likewise.
1966         * testsuite/libffi.call/nested_struct6.c: Likewise.
1967         * testsuite/libffi.call/nested_struct5.c: Likewise.
1968         * testsuite/libffi.call/nested_struct4.c: Likewise.
1969
1970 2006-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
1971
1972         * configure.ac: Enable libffi for sparc64-*-freebsd*.
1973         * configure: Rebuilt.
1974
1975 2006-01-18  Jakub Jelinek  <jakub@redhat.com>
1976
1977         * src/powerpc/sysv.S (smst_two_register): Don't call __ashldi3,
1978         instead do the shifting inline.
1979         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't compute %r5
1980         shift count unconditionally.  Simplify load sequences for 1, 2, 3, 4
1981         and 8 byte structs, for the remaining struct sizes don't call
1982         __lshrdi3, instead do the shifting inline.
1983
1984 2005-12-07  Thiemo Seufer  <ths@networkno.de>
1985
1986         * src/mips/ffitarget.h: Remove obsolete sgidefs.h include. Add
1987         missing parentheses.
1988         * src/mips/o32.S (ffi_call_O32): Code formatting. Define
1989         and use A3_OFF, FP_OFF, RA_OFF. Micro-optimizations.
1990         (ffi_closure_O32): Likewise, but with newly defined A3_OFF2,
1991         A2_OFF2, A1_OFF2, A0_OFF2, RA_OFF2, FP_OFF2, S0_OFF2, GP_OFF2,
1992         V1_OFF2, V0_OFF2, FA_1_1_OFF2, FA_1_0_OFF2, FA_0_1_OFF2,
1993         FA_0_0_OFF2.
1994         * src/mips/ffi.c (ffi_prep_args): Code formatting. Fix
1995         endianness bugs.
1996         (ffi_prep_closure): Improve trampoline instruction scheduling.
1997         (ffi_closure_mips_inner_O32): Fix endianness bugs.
1998
1999 2005-12-03  Alan Modra  <amodra@bigpond.net.au>
2000
2001         * src/powerpc/ffi.c: Formatting.
2002         (ffi_prep_args_SYSV): Avoid possible aliasing problems by using unions.
2003         (ffi_prep_args64): Likewise.
2004
2005 2005-09-30  Geoffrey Keating  <geoffk@apple.com>
2006
2007         * testsuite/lib/libffi-dg.exp (libffi_target_compile): For
2008         darwin, use -shared-libgcc not -lgcc_s, and explain why.
2009
2010 2005-09-26  Tom Tromey  <tromey@redhat.com>
2011
2012         * testsuite/libffi.call/float1.c (value_type): New typedef.
2013         (CANARY): New define.
2014         (main): Check for result buffer overflow.
2015         * src/powerpc/linux64.S: Handle linux64 long double returns.
2016         * src/powerpc/ffi.c (FLAG_RETURNS_128BITS): New constant.
2017         (ffi_prep_cif_machdep): Handle linux64 long double returns.
2018
2019 2005-08-25  Alan Modra  <amodra@bigpond.net.au>
2020
2021         PR target/23404
2022         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Correct placement of stack
2023         homed fp args.
2024         (ffi_status ffi_prep_cif_machdep): Correct stack sizing for same.
2025
2026 2005-08-11  Jakub Jelinek  <jakub@redhat.com>
2027
2028         * configure.ac (HAVE_HIDDEN_VISIBILITY_ATTRIBUTE): New test.
2029         (AH_BOTTOM): Add FFI_HIDDEN definition.
2030         * configure: Rebuilt.
2031         * fficonfig.h.in: Rebuilt.
2032         * src/powerpc/ffi.c (hidden): Remove.
2033         (ffi_closure_LINUX64, ffi_prep_args64, ffi_call_LINUX64,
2034         ffi_closure_helper_LINUX64): Use FFI_HIDDEN instead of hidden.
2035         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64,
2036         .ffi_closure_LINUX64): Use FFI_HIDDEN instead of .hidden.
2037         * src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV): Remove,
2038         add FFI_HIDDEN to its prototype.
2039         (ffi_closure_SYSV_inner): New.
2040         * src/x86/sysv.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2041         * src/x86/win32.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2042
2043 2005-08-10  Alfred M. Szmidt  <ams@gnu.org>
2044
2045         PR libffi/21819:
2046         * configure: Rebuilt.
2047         * configure.ac: Handle i*86-*-gnu*.
2048
2049 2005-08-09  Jakub Jelinek  <jakub@redhat.com>
2050
2051         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Use
2052         DW_CFA_offset_extended_sf rather than
2053         DW_CFA_GNU_negative_offset_extended.
2054         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise.
2055
2056 2005-07-22  SUGIOKA Toshinobu  <sugioka@itonet.co.jp>
2057
2058         * src/sh/sysv.S (ffi_call_SYSV): Stop argument popping correctly
2059         on sh3.
2060         (ffi_closure_SYSV): Change the stack layout for sh3 struct argument.
2061         * src/sh/ffi.c (ffi_prep_args): Fix sh3 argument copy, when it is
2062         partially on register.
2063         (ffi_closure_helper_SYSV): Likewise.
2064         (ffi_prep_cif_machdep): Don't set too many cif->flags.
2065
2066 2005-07-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2067
2068         * src/sh/ffi.c (ffi_call): Handle small structures correctly.
2069         Remove empty line.
2070         * src/sh64/ffi.c (simple_type): Remove.
2071         (return_type): Handle small structures correctly.
2072         (ffi_prep_args): Likewise.
2073         (ffi_call): Likewise.
2074         (ffi_closure_helper_SYSV): Likewise.
2075         * src/sh64/sysv.S (ffi_call_SYSV): Handle 1, 2 and 4-byte return.
2076         Emit position independent code if PIC and remove wrong datalabel
2077         prefixes from EH data.
2078
2079 2005-07-19  Andreas Tobler  <a.tobler@schweiz.ch>
2080
2081         * Makefile.am (nodist_libffi_la_SOURCES): Add POWERPC_FREEBSD.
2082         * Makefile.in: Regenerate.
2083         * include/Makefile.in: Likewise.
2084         * testsuite/Makefile.in: Likewise.
2085         * configure.ac: Add POWERPC_FREEBSD rules.
2086         * configure: Regenerate.
2087         * src/powerpc/ffitarget.h: Add POWERPC_FREEBSD rules.
2088         (FFI_SYSV_TYPE_SMALL_STRUCT): Define.
2089         * src/powerpc/ffi.c: Add flags to handle small structure returns
2090         in ffi_call_SYSV.
2091         (ffi_prep_cif_machdep): Handle small structures for SYSV 4 ABI.
2092         Aka FFI_SYSV.
2093         (ffi_closure_helper_SYSV): Likewise.
2094         * src/powerpc/ppc_closure.S: Add return types for small structures.
2095         * src/powerpc/sysv.S: Add bits to handle small structures for
2096         final SYSV 4 ABI.
2097
2098 2005-07-10  Andreas Tobler  <a.tobler@schweiz.ch>
2099
2100         * testsuite/libffi.call/cls_5_1_byte.c: New test file.
2101         * testsuite/libffi.call/cls_6_1_byte.c: Likewise.
2102         * testsuite/libffi.call/cls_7_1_byte.c: Likewise.
2103
2104 2005-07-05  Randolph Chung  <tausq@debian.org>
2105
2106         * src/pa/ffi.c (ffi_struct_type): Rename FFI_TYPE_SMALL_STRUCT1
2107         as FFI_TYPE_SMALL_STRUCT3.  Break out handling for 5-7 byte
2108         structures.  Kill compilation warnings.
2109         (ffi_closure_inner_LINUX): Print return values as hex in debug
2110         message.  Rename FFI_TYPE_SMALL_STRUCT1 as FFI_TYPE_SMALL_STRUCT3.
2111         Properly handle 5-7 byte structure returns.
2112         * src/pa/ffitarget.h (FFI_TYPE_SMALL_STRUCT1)
2113         (FFI_TYPE_SMALL_STRUCT2): Remove.
2114         (FFI_TYPE_SMALL_STRUCT3, FFI_TYPE_SMALL_STRUCT5)
2115         (FFI_TYPE_SMALL_STRUCT6, FFI_TYPE_SMALL_STRUCT7): Define.
2116         * src/pa/linux.S: Mark source file as using PA1.1 assembly.
2117         (checksmst1, checksmst2): Remove.
2118         (checksmst3): Optimize handling of 3-byte struct returns.
2119         (checksmst567): Properly handle 5-7 byte struct returns.
2120
2121 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2122
2123         PR libgcj/21943
2124         * src/mips/n32.S: Enforce PIC code.
2125         * src/mips/o32.S: Likewise.
2126
2127 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2128
2129         * configure.ac: Treat i*86-*-solaris2.10 and up as X86_64.
2130         * configure: Regenerate.
2131
2132 2005-06-01  Alan Modra  <amodra@bigpond.net.au>
2133
2134         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't use JUMPTARGET
2135         to call ffi_closure_helper_SYSV.  Append @local instead.
2136         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise for ffi_prep_args_SYSV.
2137
2138 2005-05-17  Kelley Cook  <kcook@gcc.gnu.org>
2139
2140         * configure.ac: Use AC_C_BIGENDIAN instead of AC_C_BIGENDIAN_CROSS.
2141         Use AC_CHECK_SIZEOF instead of AC_COMPILE_CHECK_SIZEOF.
2142         * Makefile.am (ACLOCAL_AMFLAGS): Remove -I ../config.
2143         * aclocal.m4, configure, fficonfig.h.in, Makefile.in,
2144         include/Makefile.in, testsuite/Makefile.in: Regenerate.
2145
2146 2005-05-09  Mike Stump  <mrs@apple.com>
2147
2148         * configure: Regenerate.
2149
2150 2005-05-08  Richard Henderson  <rth@redhat.com>
2151
2152         PR libffi/21285
2153         * src/alpha/osf.S: Update unwind into to match code.
2154
2155 2005-05-04  Andreas Degert <ad@papyrus-gmbh.de>
2156             Richard Henderson  <rth@redhat.com>
2157
2158         * src/x86/ffi64.c (ffi_prep_cif_machdep): Save sse-used flag in
2159         bit 11 of flags.
2160         (ffi_call): Mask return type field.  Pass ssecount to ffi_call_unix64.
2161         (ffi_prep_closure): Set carry bit if sse-used flag set.
2162         * src/x86/unix64.S (ffi_call_unix64): Add ssecount argument.
2163         Only load sse registers if ssecount non-zero.
2164         (ffi_closure_unix64): Only save sse registers if carry set on entry.
2165
2166 2005-04-29  Ralf Corsepius  <ralf.corsepius@rtems.org>
2167
2168         * configure.ac: Add i*86-*-rtems*, sparc*-*-rtems*,
2169         powerpc-*rtems*, arm*-*-rtems*, sh-*-rtems*.
2170         * configure: Regenerate.
2171
2172 2005-04-20  Hans-Peter Nilsson  <hp@axis.com>
2173
2174         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): In regsub use,
2175         have Tcl8.3-compatible intermediate variable.
2176
2177 2005-04-18  Simon Posnjak <simon.posnjak@siol.net>
2178             Hans-Peter Nilsson  <hp@axis.com>
2179
2180         * Makefile.am: Add CRIS support.
2181         * configure.ac: Likewise.
2182         * Makefile.in, configure, testsuite/Makefile.in,
2183         include/Makefile.in: Regenerate.
2184         * src/cris: New directory.
2185         * src/cris/ffi.c, src/cris/sysv.S, src/cris/ffitarget.h: New files.
2186         * src/prep_cif.c (ffi_prep_cif): Wrap in #ifndef __CRIS__.
2187
2188         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): Replace \n with
2189         \r?\n in output tests.
2190
2191 2005-04-12  Mike Stump  <mrs@apple.com>
2192
2193         * configure: Regenerate.
2194
2195 2005-03-30  Hans Boehm  <Hans.Boehm@hp.com>
2196
2197         * src/ia64/ffitarget.h (ffi_arg): Use long long instead of DI.
2198
2199 2005-03-30  Steve Ellcey  <sje@cup.hp.com>
2200
2201         * src/ia64/ffitarget.h (ffi_arg) ADD DI attribute.
2202         (ffi_sarg) Ditto.
2203         * src/ia64/unix.S (ffi_closure_unix): Extend gp
2204         to 64 bits in ILP32 mode.
2205         Load 64 bits even for short data.
2206
2207 2005-03-23  Mike Stump  <mrs@apple.com>
2208
2209         * src/powerpc/darwin.S: Update for -m64 multilib.
2210         * src/powerpc/darwin_closure.S: Likewise.
2211
2212 2005-03-21  Zack Weinberg  <zack@codesourcery.com>
2213
2214         * configure.ac: Do not invoke TL_AC_GCC_VERSION.
2215         Do not set tool_include_dir.
2216         * aclocal.m4, configure, Makefile.in, testsuite/Makefile.in:
2217         Regenerate.
2218         * include/Makefile.am: Set gcc_version and toollibffidir.
2219         * include/Makefile.in: Regenerate.
2220
2221 2005-02-22  Andrew Haley  <aph@redhat.com>
2222
2223         * src/powerpc/ffi.c (ffi_prep_cif_machdep): Bump alignment to
2224         odd-numbered register pairs for 64-bit integer types.
2225
2226 2005-02-23  Andreas Tobler  <a.tobler@schweiz.ch>
2227
2228         PR libffi/20104
2229         * testsuite/libffi.call/return_ll1.c: New test case.
2230
2231 2005-02-11  Janis Johnson  <janis187@us.ibm.com>
2232
2233         * testsuite/libffi.call/cls_align_longdouble.c: Remove dg-options.
2234         * testsuite/libffi.call/float.c: Ditto.
2235         * testsuite/libffi.call/float2.c: Ditto.
2236         * testsuite/libffi.call/float3.c: Ditto.
2237
2238 2005-02-08  Andreas Tobler  <a.tobler@schweiz.ch>
2239
2240         * src/frv/ffitarget.h: Remove PPC stuff which does not belong to frv.
2241
2242 2005-01-12  Eric Botcazou  <ebotcazou@libertysurf.fr>
2243
2244         * testsuite/libffi.special/special.exp (cxx_options): Add
2245         -shared-libgcc.
2246
2247 2004-12-31  Richard Henderson  <rth@redhat.com>
2248
2249         * src/types.c (FFI_AGGREGATE_TYPEDEF): Remove.
2250         (FFI_TYPEDEF): Rename from FFI_INTEGRAL_TYPEDEF.  Replace size and
2251         offset parameters with a type parameter; deduce size and structure
2252         alignment.  Update all users.
2253
2254 2004-12-31  Richard Henderson  <rth@redhat.com>
2255
2256         * src/types.c (FFI_TYPE_POINTER): Define with sizeof.
2257         (FFI_TYPE_LONGDOUBLE): Fix for ia64.
2258         * src/ia64/ffitarget.h (struct ffi_ia64_trampoline_struct): Move
2259         into ffi_prep_closure.
2260         * src/ia64/ia64_flags.h, src/ia64/ffi.c, src/ia64/unix.S: Rewrite
2261         from scratch.
2262
2263 2004-12-27  Richard Henderson  <rth@redhat.com>
2264
2265         * src/x86/unix64.S: Fix typo in unwind info.
2266
2267 2004-12-25  Richard Henderson  <rth@redhat.com>
2268
2269         * src/x86/ffi64.c (struct register_args): Rename from stackLayout.
2270         (enum x86_64_reg_class): Add X86_64_COMPLEX_X87_CLASS.
2271         (merge_classes): Check for it.
2272         (SSE_CLASS_P): New.
2273         (classify_argument): Pass byte_offset by value; perform all updates
2274         inside struct case.
2275         (examine_argument): Add classes argument; handle
2276         X86_64_COMPLEX_X87_CLASS.
2277         (ffi_prep_args): Merge into ...
2278         (ffi_call): ... here.  Share stack frame with ffi_call_unix64.
2279         (ffi_prep_cif_machdep): Setup cif->flags for proper structure return.
2280         (ffi_fill_return_value): Remove.
2281         (ffi_prep_closure): Remove dead assert.
2282         (ffi_closure_unix64_inner): Rename from ffi_closure_UNIX64_inner.
2283         Rewrite to use struct register_args instead of va_list.  Create
2284         flags for handling structure returns.
2285         * src/x86/unix64.S: Remove dead strings.
2286         (ffi_call_unix64): Rename from ffi_call_UNIX64.  Rewrite to share
2287         stack frame with ffi_call.  Handle structure returns properly.
2288         (float2sse, floatfloat2sse, double2sse): Remove.
2289         (sse2float, sse2double, sse2floatfloat): Remove.
2290         (ffi_closure_unix64): Rename from ffi_closure_UNIX64.  Rewrite
2291         to handle structure returns properly.
2292
2293 2004-12-08  David Edelsohn  <edelsohn@gnu.org>
2294
2295         * Makefile.am (AM_MAKEFLAGS): Remove duplicate LIBCFLAGS and
2296         PICFLAG.
2297         * Makefile.in: Regenerated.
2298
2299 2004-12-02  Richard Sandiford  <rsandifo@redhat.com>
2300
2301         * configure.ac: Use TL_AC_GCC_VERSION to set gcc_version.
2302         * configure, aclocal.m4, Makefile.in: Regenerate.
2303         * include/Makefile.in, testsuite/Makefile.in: Regenerate.
2304
2305 2004-11-29  Kelley Cook  <kcook@gcc.gnu.org>
2306
2307         * configure: Regenerate for libtool change.
2308
2309 2004-11-25  Kelley Cook  <kcook@gcc.gnu.org>
2310
2311         * configure: Regenerate for libtool reversion.
2312
2313 2004-11-24  Kelley Cook  <kcook@gcc.gnu.org>
2314
2315         * configure: Regenerate for libtool change.
2316
2317 2004-11-23  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2318
2319         * testsuite/lib/libffi-dg.exp: Use new procs in target-libpath.exp.
2320
2321 2004-11-23  Richard Sandiford  <rsandifo@redhat.com>
2322
2323         * src/mips/o32.S (ffi_call_O32, ffi_closure_O32): Use jalr instead
2324         of jal.  Use an absolute encoding for the frame information.
2325
2326 2004-11-23  Kelley Cook  <kcook@gcc.gnu.org>
2327
2328         * Makefile.am: Remove no-dependencies.  Add ACLOCAL_AMFLAGS.
2329         * acinclude.m4: Delete logic for sincludes.
2330         * aclocal.m4, Makefile.in, configure: Regenerate.
2331         * include/Makefile: Likewise.
2332         * testsuite/Makefile: Likewise.
2333
2334 2004-11-22  Eric Botcazou  <ebotcazou@libertysurf.fr>
2335
2336         * src/sparc/ffi.c (ffi_prep_closure): Align doubles and 64-bit integers
2337         on a 8-byte boundary.
2338         * src/sparc/v8.S (ffi_closure_v8): Reserve frame space for arguments.
2339
2340 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2341
2342         * src/arm/ffi.c (ffi_prep_cif_machdep): Handle functions that return
2343         long long values.  Round stack allocation to a multiple of 8 bytes
2344         for ATPCS compatibility.
2345         * src/arm/sysv.S (ffi_call_SYSV): Rework to avoid use of APCS register
2346         names.  Handle returning long long types.  Add Thumb and interworking
2347         support.  Improve soft-float code.
2348
2349 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2350
2351         * testsuite/lib/libffi-db.exp (load_gcc_lib): New function.
2352         (libffi_exit): New function.
2353         (libffi_init): Build the testglue wrapper if needed.
2354
2355 2004-10-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2356
2357         PR other/18138
2358         * testsuite/lib/libffi-dg.exp: Accept more than one multilib libgcc.
2359
2360 2004-10-25  Kazuhiro Inaoka <inaoka.kazuhiro@renesas.com>
2361
2362         * src/m32r/libffitarget.h (FFI_CLOSURES): Set to 0.
2363
2364 2004-10-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2365
2366         * src/sh/sysv.S (ffi_call_SYSV): Don't align for double data.
2367         * testsuite/libffi.call/float3.c: New test case.
2368
2369 2004-10-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2370
2371         * src/sh/ffi.c (ffi_prep_closure): Set T bit in trampoline for
2372         the function returning a structure pointed with R2.
2373         * src/sh/sysv.S (ffi_closure_SYSV): Use R2 as the pointer to
2374         the structure return value if T bit set.  Emit position
2375         independent code and EH data if PIC.
2376
2377 2004-10-13  Kazuhiro Inaoka  <inaoka.kazuhiro@renesas.com>
2378
2379         * Makefile.am: Add m32r support.
2380         * configure.ac: Likewise.
2381         * Makefile.in: Regenerate.
2382         * confiugre: Regenerate.
2383         * src/types.c: Add m32r port to FFI_INTERNAL_TYPEDEF
2384         (uint64, sint64, double, longdouble)
2385         * src/m32r: New directory.
2386         * src/m32r/ffi.c: New file.
2387         * src/m32r/sysv.S: Likewise.
2388         * src/m32r/ffitarget.h: Likewise.
2389
2390 2004-10-02  Kaz Kojima  <kkojima@gcc.gnu.org>
2391
2392         * testsuite/libffi.call/negint.c: New test case.
2393
2394 2004-09-14  H.J. Lu  <hongjiu.lu@intel.com>
2395
2396         PR libgcj/17465
2397         * testsuite/lib/libffi-dg.exp: Don't use global ld_library_path.
2398         Set up LD_LIBRARY_PATH, SHLIB_PATH, LD_LIBRARYN32_PATH,
2399         LD_LIBRARY64_PATH, LD_LIBRARY_PATH_32, LD_LIBRARY_PATH_64 and
2400         DYLD_LIBRARY_PATH.
2401
2402 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2403
2404         * testsuite/libffi.call/many_win32.c: Remove whitespaces.
2405         * testsuite/libffi.call/promotion.c: Likewise.
2406         * testsuite/libffi.call/return_ll.c: Remove unused var. Cleanup
2407         whitespaces.
2408         * testsuite/libffi.call/return_sc.c: Likewise.
2409         * testsuite/libffi.call/return_uc.c: Likewise.
2410
2411 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2412
2413         * src/powerpc/darwin.S: Fix comments and identation.
2414         * src/powerpc/darwin_closure.S: Likewise.
2415
2416 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2417
2418         * src/powerpc/ffi_darwin.c: Add flag for longdouble return values.
2419         (ffi_prep_args): Handle longdouble arguments.
2420         (ffi_prep_cif_machdep): Set flags for longdouble. Calculate space for
2421         longdouble.
2422         (ffi_closure_helper_DARWIN): Add closure handling for longdouble.
2423         * src/powerpc/darwin.S (_ffi_call_DARWIN): Add handling of longdouble
2424         values.
2425         * src/powerpc/darwin_closure.S (_ffi_closure_ASM): Likewise.
2426         * src/types.c: Defined longdouble size and alignment for darwin.
2427
2428 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2429
2430         * src/powerpc/aix.S: Remove whitespaces.
2431         * src/powerpc/aix_closure.S: Likewise.
2432         * src/powerpc/asm.h: Likewise.
2433         * src/powerpc/ffi.c: Likewise.
2434         * src/powerpc/ffitarget.h: Likewise.
2435         * src/powerpc/linux64.S: Likewise.
2436         * src/powerpc/linux64_closure.S: Likewise.
2437         * src/powerpc/ppc_closure.S: Likewise.
2438         * src/powerpc/sysv.S: Likewise.
2439
2440 2004-08-30  Anthony Green  <green@redhat.com>
2441
2442         * Makefile.am: Add frv support.
2443         * Makefile.in, testsuite/Makefile.in: Rebuilt.
2444         * configure.ac: Read configure.host.
2445         * configure.in: Read configure.host.
2446         * configure.host: New file.  frv-elf needs libgloss.
2447         * include/ffi.h.in: Force ffi_closure to have a nice big (8)
2448         alignment.  This is needed to frv and shouldn't harm the others.
2449         * include/ffi_common.h (ALIGN_DOWN): New macro.
2450         * src/frv/ffi.c, src/frv/ffitarget.h, src/frv/eabi.S: New files.
2451
2452 2004-08-24  David Daney  <daney@avtrex.com>
2453
2454         * testsuite/libffi.call/closure_fn0.c: Xfail mips64* instead of mips*.
2455         * testsuite/libffi.call/closure_fn1.c: Likewise.
2456         * testsuite/libffi.call/closure_fn2.c  Likewise.
2457         * testsuite/libffi.call/closure_fn3.c: Likewise.
2458         * testsuite/libffi.call/closure_fn4.c: Likewise.
2459         * testsuite/libffi.call/closure_fn5.c: Likewise.
2460         * testsuite/libffi.call/cls_18byte.c: Likewise.
2461         * testsuite/libffi.call/cls_19byte.c: Likewise.
2462         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2463         * testsuite/libffi.call/cls_20byte.c: Likewise.
2464         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2465         * testsuite/libffi.call/cls_24byte.c: Likewise.
2466         * testsuite/libffi.call/cls_2byte.c: Likewise.
2467         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2468         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2469         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2470         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2471         * testsuite/libffi.call/cls_4byte.c: Likewise.
2472         * testsuite/libffi.call/cls_64byte.c: Likewise.
2473         * testsuite/libffi.call/cls_6byte.c: Likewise.
2474         * testsuite/libffi.call/cls_7byte.c: Likewise.
2475         * testsuite/libffi.call/cls_8byte.c: Likewise.
2476         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2477         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2478         * testsuite/libffi.call/cls_align_double.c: Likewise.
2479         * testsuite/libffi.call/cls_align_float.c: Likewise.
2480         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2481         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2482         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2483         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2484         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2485         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2486         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2487         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
2488         * testsuite/libffi.call/cls_double.c: Likewise.
2489         * testsuite/libffi.call/cls_float.c: Likewise.
2490         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2491         * testsuite/libffi.call/cls_multi_sshort.c: Likewise.
2492         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2493         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2494         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2495         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2496         * testsuite/libffi.call/cls_schar.c: Likewise.
2497         * testsuite/libffi.call/cls_sint.c: Likewise.
2498         * testsuite/libffi.call/cls_sshort.c: Likewise.
2499         * testsuite/libffi.call/cls_uchar.c: Likewise.
2500         * testsuite/libffi.call/cls_uint.c: Likewise.
2501         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2502         * testsuite/libffi.call/cls_ushort.c: Likewise.
2503         * testsuite/libffi.call/nested_struct.c: Likewise.
2504         * testsuite/libffi.call/nested_struct1.c: Likewise.
2505         * testsuite/libffi.call/nested_struct2.c: Likewise.
2506         * testsuite/libffi.call/nested_struct3.c: Likewise.
2507         * testsuite/libffi.call/problem1.c: Likewise.
2508         * testsuite/libffi.special/unwindtest.cc: Likewise.
2509         * testsuite/libffi.call/cls_12byte.c: Likewise and set return value
2510         to zero.
2511         * testsuite/libffi.call/cls_16byte.c: Likewise.
2512         * testsuite/libffi.call/cls_5byte.c: Likewise.
2513
2514 2004-08-23  David Daney <daney@avtrex.com>
2515
2516         PR libgcj/13141
2517         * src/mips/ffitarget.h (FFI_O32_SOFT_FLOAT): New ABI.
2518         * src/mips/ffi.c (ffi_prep_args): Fix alignment calculation.
2519         (ffi_prep_cif_machdep): Handle FFI_O32_SOFT_FLOAT floating point
2520         parameters and return types.
2521         (ffi_call): Handle FFI_O32_SOFT_FLOAT ABI.
2522         (ffi_prep_closure): Ditto.
2523         (ffi_closure_mips_inner_O32): Handle FFI_O32_SOFT_FLOAT ABI, fix
2524         alignment calculations.
2525         * src/mips/o32.S (ffi_closure_O32): Don't use floating point
2526         instructions if FFI_O32_SOFT_FLOAT, make stack frame ABI compliant.
2527
2528 2004-08-14  Casey Marshall <csm@gnu.org>
2529
2530         * src/mips/ffi.c (ffi_pref_cif_machdep): set `cif->flags' to
2531         contain `FFI_TYPE_UINT64' as return type for any 64-bit
2532         integer (O32 ABI only).
2533         (ffi_prep_closure): new function.
2534         (ffi_closure_mips_inner_O32): new function.
2535         * src/mips/ffitarget.h: Define `FFI_CLOSURES' and
2536         `FFI_TRAMPOLINE_SIZE' appropriately if the ABI is o32.
2537         * src/mips/o32.S (ffi_call_O32): add labels for .eh_frame. Return
2538         64 bit integers correctly.
2539         (ffi_closure_O32): new function.
2540         Added DWARF-2 unwind info for both functions.
2541
2542 2004-08-10  Andrew Haley  <aph@redhat.com>
2543
2544         * src/x86/ffi64.c (ffi_prep_args ): 8-align all stack arguments.
2545
2546 2004-08-01  Robert Millan  <robertmh@gnu.org>
2547
2548         * configure.ac: Detect knetbsd-gnu and kfreebsd-gnu.
2549         * configure: Regenerate.
2550
2551 2004-07-30  Maciej W. Rozycki  <macro@linux-mips.org>
2552
2553         * acinclude.m4 (AC_FUNC_MMAP_BLACKLIST): Check for <sys/mman.h>
2554         and mmap() explicitly instead of relying on preset autoconf cache
2555         variables.
2556         * aclocal.m4: Regenerate.
2557         * configure: Regenerate.
2558
2559 2004-07-11  Ulrich Weigand  <uweigand@de.ibm.com>
2560
2561         * src/s390/ffi.c (ffi_prep_args): Fix C aliasing violation.
2562         (ffi_check_float_struct): Remove unused prototype.
2563
2564 2004-06-30  Geoffrey Keating  <geoffk@apple.com>
2565
2566         * src/powerpc/ffi_darwin.c (flush_icache): ';' is a comment
2567         character on Darwin, use '\n\t' instead.
2568
2569 2004-06-26  Matthias Klose  <doko@debian.org>
2570
2571         * libtool-version: Fix typo in revision/age.
2572
2573 2004-06-17  Matthias Klose  <doko@debian.org>
2574
2575         * libtool-version: New.
2576         * Makefile.am (libffi_la_LDFLAGS): Use -version-info for soname.
2577         * Makefile.in: Regenerate.
2578
2579 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2580
2581         * Makefile.am: Remove useless multilib rules.
2582         * Makefile.in: Regenerate.
2583         * aclocal.m4: Regenerate with automake 1.8.5.
2584         * configure.ac: Remove useless multilib configury.
2585         * configure: Regenerate.
2586
2587 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2588
2589         * .cvsignore: New file.
2590
2591 2004-06-10  Jakub Jelinek  <jakub@redhat.com>
2592
2593         * src/ia64/unix.S (ffi_call_unix): Insert group barrier break
2594         fp_done.
2595         (ffi_closure_UNIX): Fix f14/f15 adjustment if FLOAT_SZ is ever
2596         changed from 8.
2597
2598 2004-06-06  Sean McNeil  <sean@mcneil.com>
2599
2600         * configure.ac: Add x86_64-*-freebsd* support.
2601         * configure: Regenerate.
2602
2603 2004-04-26  Joe Buck <jbuck@welsh-buck.org>
2604
2605         Bug 15093
2606         * configure.ac: Test for existence of mmap and sys/mman.h before
2607         checking blacklist.  Fix suggested by Jim Wilson.
2608         * configure: Regenerate.
2609
2610 2004-04-26  Matt Austern  <austern@apple.com>
2611
2612         * src/powerpc/darwin.S: Go through a non-lazy pointer for initial
2613         FDE location.
2614         * src/powerpc/darwin_closure.S: Likewise.
2615
2616 2004-04-24  Andreas Tobler  <a.tobler@schweiz.ch>
2617
2618         * testsuite/libffi.call/cls_multi_schar.c (main): Fix initialization
2619         error. Reported by Thomas Heller <theller@python.net>.
2620         * testsuite/libffi.call/cls_multi_sshort.c (main): Likewise.
2621         * testsuite/libffi.call/cls_multi_ushort.c (main): Likewise.
2622
2623 2004-03-20  Matthias Klose  <doko@debian.org>
2624
2625         * src/pa/linux.S: Fix typo.
2626
2627 2004-03-19  Matthias Klose  <doko@debian.org>
2628
2629         * Makefile.am: Update.
2630         * Makefile.in: Regenerate.
2631         * src/pa/ffi.h.in: Remove.
2632         * src/pa/ffitarget.h: New file.
2633
2634 2004-02-10  Randolph Chung  <tausq@debian.org>
2635
2636         * Makefile.am: Add PA support.
2637         * Makefile.in: Regenerate.
2638         * include/Makefile.in: Regenerate.
2639         * configure.ac: Add PA target.
2640         * configure: Regenerate.
2641         * src/pa/ffi.c: New file.
2642         * src/pa/ffi.h.in: Add PA support.
2643         * src/pa/linux.S: New file.
2644         * prep_cif.c: Add PA support.
2645
2646 2004-03-16  Hosaka Yuji  <hos@tamanegi.org>
2647
2648         * src/types.c: Fix alignment size of X86_WIN32 case int64 and
2649         double.
2650         * src/x86/ffi.c (ffi_prep_args): Replace ecif->cif->rtype->type
2651         with ecif->cif->flags.
2652         (ffi_call, ffi_prep_incoming_args_SYSV): Replace cif->rtype->type
2653         with cif->flags.
2654         (ffi_prep_cif_machdep): Add X86_WIN32 struct case.
2655         (ffi_closure_SYSV): Add 1 or 2-bytes struct case for X86_WIN32.
2656         * src/x86/win32.S (retstruct1b, retstruct2b, sc_retstruct1b,
2657         sc_retstruct2b): Add for 1 or 2-bytes struct case.
2658
2659 2004-03-15 Kelley Cook <kcook@gcc.gnu.org>
2660
2661         * configure.in: Rename file to ...
2662         * configure.ac: ... this.
2663         * fficonfig.h.in: Regenerate.
2664         * Makefile.in: Regenerate.
2665         * include/Makefile.in: Regenerate.
2666         * testsuite/Makefile.in: Regenerate.
2667
2668 2004-03-12  Matt Austern  <austern@apple.com>
2669
2670         * src/powerpc/darwin.S: Fix EH information so it corresponds to
2671         changes in EH format resulting from addition of linkonce support.
2672         * src/powerpc/darwin_closure.S: Likewise.
2673
2674 2004-03-11  Andreas Tobler  <a.tobler@schweiz.ch>
2675             Paolo Bonzini  <bonzini@gnu.org>
2676
2677         * Makefile.am (AUTOMAKE_OPTIONS): Set them.
2678         Remove VPATH. Remove rules for object files. Remove multilib support.
2679         (AM_CCASFLAGS): Add.
2680         * configure.in (AC_CONFIG_HEADERS): Relace AM_CONFIG_HEADER.
2681         (AC_PREREQ): Bump version to 2.59.
2682         (AC_INIT): Fill with version info and bug address.
2683         (ORIGINAL_LD_FOR_MULTILIBS): Remove.
2684         (AM_ENABLE_MULTILIB): Use this instead of AC_ARG_ENABLE.
2685         De-precious CC so that the right flags are passed down to multilibs.
2686         (AC_MSG_ERROR): Replace obsolete macro AC_ERROR.
2687         (AC_CONFIG_FILES): Replace obsolete macro AC_LINK_FILES.
2688         (AC_OUTPUT): Reorganize the output with AC_CONFIG_COMMANDS.
2689         * configure: Rebuilt.
2690         * aclocal.m4: Likewise.
2691         * Makefile.in, include/Makefile.in, testsuite/Makefile.in: Likewise.
2692         * fficonfig.h.in: Likewise.
2693
2694 2004-03-11  Andreas Schwab  <schwab@suse.de>
2695
2696         * src/ia64/ffi.c (ffi_prep_incoming_args_UNIX): Get floating point
2697         arguments from fp registers only for the first 8 parameter slots.
2698         Don't convert a float parameter when passed in memory.
2699
2700 2004-03-09  Hans-Peter Nilsson  <hp@axis.com>
2701
2702         * configure: Regenerate for config/accross.m4 correction.
2703
2704 2004-02-25  Matt Kraai  <kraai@alumni.cmu.edu>
2705
2706         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Change
2707         ecif->cif->bytes to bytes.
2708         (ffi_prep_cif_machdep): Add braces around nested if statement.
2709
2710 2004-02-09  Alan Modra  <amodra@bigpond.net.au>
2711
2712         * src/types.c (pointer): POWERPC64 has 8 byte pointers.
2713
2714         * src/powerpc/ffi.c (ffi_prep_args64): Correct long double handling.
2715         (ffi_closure_helper_LINUX64): Fix typo.
2716         * testsuite/libffi.call/cls_align_longdouble.c: Pass -mlong-double-128
2717         for powerpc64-*-*.
2718         * testsuite/libffi.call/float.c: Likewise.
2719         * testsuite/libffi.call/float2.c: Likewise.
2720
2721 2004-02-08  Alan Modra  <amodra@bigpond.net.au>
2722
2723         * src/powerpc/ffi.c (ffi_prep_cif_machdep <FFI_LINUX64>): Correct
2724         long double function return and long double arg handling.
2725         (ffi_closure_helper_LINUX64): Formatting.  Delete unused "ng" var.
2726         Use "end_pfr" instead of "nf".  Correct long double handling.
2727         Localise "temp".
2728         * src/powerpc/linux64.S (ffi_call_LINUX64): Save f2 long double
2729         return value.
2730         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Allocate
2731         space for long double return value.  Adjust stack frame and offsets.
2732         Load f2 long double return.
2733
2734 2004-02-07  Alan Modra  <amodra@bigpond.net.au>
2735
2736         * src/types.c: Use 16 byte long double for POWERPC64.
2737
2738 2004-01-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2739
2740         * src/sparc/ffi.c (ffi_prep_args_v9): Shift the parameter array
2741         when the structure return address is passed in %o0.
2742         (ffi_V9_return_struct): Rename into ffi_v9_layout_struct.
2743         (ffi_v9_layout_struct): Align the field following a nested structure
2744         on a word boundary.  Use memmove instead of memcpy.
2745         (ffi_call): Update call to ffi_V9_return_struct.
2746         (ffi_prep_closure): Define 'ctx' only for V8.
2747         (ffi_closure_sparc_inner): Clone into ffi_closure_sparc_inner_v8
2748         and ffi_closure_sparc_inner_v9.
2749         (ffi_closure_sparc_inner_v8): Return long doubles by reference.
2750         Always skip the structure return address.  For structures and long
2751         doubles, copy the argument directly.
2752         (ffi_closure_sparc_inner_v9): Skip the structure return address only
2753         if required.  Shift the maximum floating-point slot accordingly.  For
2754         big structures, copy the argument directly; otherwise, left-justify the
2755         argument and call ffi_v9_layout_struct to lay out the structure on
2756         the stack.
2757         * src/sparc/v8.S: Undef STACKFRAME before defining it.
2758         (ffi_closure_v8): Pass the structure return address.  Update call to
2759         ffi_closure_sparc_inner_v8.  Short-circuit FFI_TYPE_INT handling.
2760         Skip the 'unimp' insn when returning long doubles and structures.
2761         * src/sparc/v9.S: Undef STACKFRAME before defining it.
2762         (ffi_closure_v9): Increase the frame size by 2 words.  Short-circuit
2763         FFI_TYPE_INT handling.  Load structures both in integers and
2764         floating-point registers on return.
2765         * README: Update status of the SPARC port.
2766
2767 2004-01-24  Andreas Tobler  <a.tobler@schweiz.ch>
2768
2769         * testsuite/libffi.call/pyobjc-tc.c (main): Treat result value
2770         as of type ffi_arg.
2771         * testsuite/libffi.call/struct3.c (main): Fix CHECK.
2772
2773 2004-01-22  Ulrich Weigand  <uweigand@de.ibm.com>
2774
2775         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Treat result
2776         value as of type ffi_arg, not unsigned int.
2777
2778 2004-01-21  Michael Ritzert  <ritzert@t-online.de>
2779
2780         * ffi64.c (ffi_prep_args): Cast the RHS of an assignment instead
2781         of the LHS.
2782
2783 2004-01-12  Andreas Tobler  <a.tobler@schweiz.ch>
2784
2785         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_32 for
2786         Solaris.
2787
2788 2004-01-08  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2789
2790         * testsuite/libffi.call/ffitest.h (allocate_mmap): Cast MAP_FAILED
2791         to void *.
2792
2793 2003-12-10  Richard Henderson  <rth@redhat.com>
2794
2795         * testsuite/libffi.call/cls_align_pointer.c: Cast pointers to
2796         size_t instead of int.
2797
2798 2003-12-04  Hosaka Yuji  <hos@tamanegi.org>
2799
2800         * testsuite/libffi.call/many_win32.c: Include <float.h>.
2801         * testsuite/libffi.call/many_win32.c (main): Replace variable
2802         int i with unsigned long ul.
2803
2804         * testsuite/libffi.call/cls_align_uint64.c: New test case.
2805         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2806         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2807         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2808         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2809         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2810         * testsuite/libffi.call/cls_align_float.c: Likewise.
2811         * testsuite/libffi.call/cls_align_double.c: Likewise.
2812         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2813         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2814
2815 2003-12-02  Hosaka Yuji  <hos@tamanegi.org>
2816
2817         PR other/13221
2818         * src/x86/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV):
2819         Align arguments to 32 bits.
2820
2821 2003-12-01  Andreas Tobler  <a.tobler@schweiz.ch>
2822
2823         PR other/13221
2824         * testsuite/libffi.call/cls_multi_sshort.c: New test case.
2825         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2826         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2827         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2828         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2829         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2830
2831         * testsuite/libffi.special/unwindtest.cc: Cosmetics.
2832
2833 2003-11-26  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
2834
2835         * testsuite/libffi.call/ffitest.h: Include <fcntl.h>.
2836         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2837
2838 2003-11-22  Andreas Tobler  <a.tobler@schweiz.ch>
2839
2840         * Makefile.in: Rebuilt.
2841         * configure: Likewise.
2842         * testsuite/libffi.special/unwindtest.cc: Convert the mmap to
2843         the right type.
2844
2845 2003-11-21  Andreas Jaeger  <aj@suse.de>
2846             Andreas Tobler  <a.tobler@schweiz.ch>
2847
2848         * acinclude.m4: Add AC_FUNC_MMAP_BLACKLIST.
2849         * configure.in: Call AC_FUNC_MMAP_BLACKLIST.
2850         * Makefile.in: Rebuilt.
2851         * aclocal.m4: Likewise.
2852         * configure: Likewise.
2853         * fficonfig.h.in: Likewise.
2854         * testsuite/lib/libffi-dg.exp: Add include dir.
2855         * testsuite/libffi.call/ffitest.h: Add MMAP definitions.
2856         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2857         * testsuite/libffi.call/closure_fn0.c: Use MMAP functionality
2858         for ffi_closure if available.
2859         * testsuite/libffi.call/closure_fn1.c: Likewise.
2860         * testsuite/libffi.call/closure_fn2.c: Likewise.
2861         * testsuite/libffi.call/closure_fn3.c: Likewise.
2862         * testsuite/libffi.call/closure_fn4.c: Likewise.
2863         * testsuite/libffi.call/closure_fn5.c: Likewise.
2864         * testsuite/libffi.call/cls_12byte.c: Likewise.
2865         * testsuite/libffi.call/cls_16byte.c: Likewise.
2866         * testsuite/libffi.call/cls_18byte.c: Likewise.
2867         * testsuite/libffi.call/cls_19byte.c: Likewise.
2868         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2869         * testsuite/libffi.call/cls_20byte.c: Likewise.
2870         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2871         * testsuite/libffi.call/cls_24byte.c: Likewise.
2872         * testsuite/libffi.call/cls_2byte.c: Likewise.
2873         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2874         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2875         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2876         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2877         * testsuite/libffi.call/cls_4byte.c: Likewise.
2878         * testsuite/libffi.call/cls_5byte.c: Likewise.
2879         * testsuite/libffi.call/cls_64byte.c: Likewise.
2880         * testsuite/libffi.call/cls_6byte.c: Likewise.
2881         * testsuite/libffi.call/cls_7byte.c: Likewise.
2882         * testsuite/libffi.call/cls_8byte.c: Likewise.
2883         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2884         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2885         * testsuite/libffi.call/cls_double.c: Likewise.
2886         * testsuite/libffi.call/cls_float.c: Likewise.
2887         * testsuite/libffi.call/cls_schar.c: Likewise.
2888         * testsuite/libffi.call/cls_sint.c: Likewise.
2889         * testsuite/libffi.call/cls_sshort.c: Likewise.
2890         * testsuite/libffi.call/cls_uchar.c: Likewise.
2891         * testsuite/libffi.call/cls_uint.c: Likewise.
2892         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2893         * testsuite/libffi.call/cls_ushort.c: Likewise.
2894         * testsuite/libffi.call/nested_struct.c: Likewise.
2895         * testsuite/libffi.call/nested_struct1.c: Likewise.
2896         * testsuite/libffi.call/nested_struct2.c: Likewise.
2897         * testsuite/libffi.call/nested_struct3.c: Likewise.
2898         * testsuite/libffi.call/problem1.c: Likewise.
2899         * testsuite/libffi.special/unwindtest.cc: Likewise.
2900
2901 2003-11-20  Andreas Tobler  <a.tobler@schweiz.ch>
2902
2903         * testsuite/lib/libffi-dg.exp: Make the -lgcc_s conditional.
2904
2905 2003-11-19  Andreas Tobler  <a.tobler@schweiz.ch>
2906
2907         * testsuite/lib/libffi-dg.exp: Add DYLD_LIBRARY_PATH for darwin.
2908         Add -lgcc_s to additional flags.
2909
2910 2003-11-12  Andreas Tobler  <a.tobler@schweiz.ch>
2911
2912         * configure.in, include/Makefile.am: PR libgcj/11147, install
2913         the ffitarget.h header file in a gcc versioned and target
2914         dependent place.
2915         * configure: Regenerated.
2916         * Makefile.in, include/Makefile.in: Likewise.
2917         * testsuite/Makefile.in: Likewise.
2918
2919 2003-11-09  Andreas Tobler  <a.tobler@schweiz.ch>
2920
2921         * testsuite/libffi.call/closure_fn0.c: Print result and check
2922         with dg-output to make debugging easier.
2923         * testsuite/libffi.call/closure_fn1.c: Likewise.
2924         * testsuite/libffi.call/closure_fn2.c: Likewise.
2925         * testsuite/libffi.call/closure_fn3.c: Likewise.
2926         * testsuite/libffi.call/closure_fn4.c: Likewise.
2927         * testsuite/libffi.call/closure_fn5.c: Likewise.
2928         * testsuite/libffi.call/cls_12byte.c: Likewise.
2929         * testsuite/libffi.call/cls_16byte.c: Likewise.
2930         * testsuite/libffi.call/cls_18byte.c: Likewise.
2931         * testsuite/libffi.call/cls_19byte.c: Likewise.
2932         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2933         * testsuite/libffi.call/cls_20byte.c: Likewise.
2934         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2935         * testsuite/libffi.call/cls_24byte.c: Likewise.
2936         * testsuite/libffi.call/cls_2byte.c: Likewise.
2937         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2938         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2939         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2940         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2941         * testsuite/libffi.call/cls_4byte.c: Likewise.
2942         * testsuite/libffi.call/cls_5byte.c: Likewise.
2943         * testsuite/libffi.call/cls_64byte.c: Likewise.
2944         * testsuite/libffi.call/cls_6byte.c: Likewise.
2945         * testsuite/libffi.call/cls_7byte.c: Likewise.
2946         * testsuite/libffi.call/cls_8byte.c: Likewise.
2947         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2948         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2949         * testsuite/libffi.call/cls_double.c: Likewise.
2950         * testsuite/libffi.call/cls_float.c: Likewise.
2951         * testsuite/libffi.call/cls_schar.c: Likewise.
2952         * testsuite/libffi.call/cls_sint.c: Likewise.
2953         * testsuite/libffi.call/cls_sshort.c: Likewise.
2954         * testsuite/libffi.call/cls_uchar.c: Likewise.
2955         * testsuite/libffi.call/cls_uint.c: Likewise.
2956         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2957         * testsuite/libffi.call/cls_ushort.c: Likewise.
2958         * testsuite/libffi.call/problem1.c: Likewise.
2959
2960         * testsuite/libffi.special/unwindtest.cc: Make ffi_closure
2961         static.
2962
2963 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
2964
2965         * testsuite/libffi.call/cls_9byte2.c: New test case.
2966         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2967         * testsuite/libffi.call/cls_64byte.c: Likewise.
2968         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2969         * testsuite/libffi.call/cls_19byte.c: Likewise.
2970         * testsuite/libffi.call/cls_18byte.c: Likewise.
2971         * testsuite/libffi.call/closure_fn4.c: Likewise.
2972         * testsuite/libffi.call/closure_fn5.c: Likewise.
2973         * testsuite/libffi.call/cls_schar.c: Likewise.
2974         * testsuite/libffi.call/cls_sint.c: Likewise.
2975         * testsuite/libffi.call/cls_sshort.c: Likewise.
2976         * testsuite/libffi.call/nested_struct2.c: Likewise.
2977         * testsuite/libffi.call/nested_struct3.c: Likewise.
2978
2979 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
2980
2981         * testsuite/libffi.call/cls_double.c: Do a check on the result.
2982         * testsuite/libffi.call/cls_uchar.c: Likewise.
2983         * testsuite/libffi.call/cls_uint.c: Likewise.
2984         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2985         * testsuite/libffi.call/cls_ushort.c: Likewise.
2986         * testsuite/libffi.call/return_sc.c: Cleanup whitespaces.
2987
2988 2003-11-06  Andreas Tobler  <a.tobler@schweiz.ch>
2989
2990         * src/prep_cif.c (ffi_prep_cif): Move the validity check after
2991         the initialization.
2992
2993 2003-10-23  Andreas Tobler  <a.tobler@schweiz.ch>
2994
2995         * src/java_raw_api.c (ffi_java_ptrarray_to_raw): Replace
2996         FFI_ASSERT(FALSE) with FFI_ASSERT(0).
2997
2998 2003-10-22  David Daney  <ddaney@avtrex.com>
2999
3000         * src/mips/ffitarget.h: Replace undefined UINT32 and friends with
3001         __attribute__((__mode__(__SI__))) and friends.
3002
3003 2003-10-22  Andreas Schwab  <schwab@suse.de>
3004
3005         * src/ia64/ffi.c: Replace FALSE/TRUE with false/true.
3006
3007 2003-10-21  Andreas Tobler  <a.tobler@schweiz.ch>
3008
3009         * configure.in: AC_LINK_FILES(ffitarget.h).
3010         * configure: Regenerate.
3011         * Makefile.in: Likewise.
3012         * include/Makefile.in: Likewise.
3013         * testsuite/Makefile.in: Likewise.
3014         * fficonfig.h.in: Likewise.
3015
3016 2003-10-21  Paolo Bonzini  <bonzini@gnu.org>
3017             Richard Henderson  <rth@redhat.com>
3018
3019         Avoid that ffi.h includes fficonfig.h.
3020
3021         * Makefile.am (EXTRA_DIST): Include ffitarget.h files
3022         (TARGET_SRC_MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3023         (TARGET_SRC_MIPS_SGI): Removed.
3024         (MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3025         (MIPS_SGI): Removed.
3026         (CLEANFILES): Removed.
3027         (mostlyclean-am, clean-am, mostlyclean-sub, clean-sub): New
3028         targets.
3029         * acconfig.h: Removed.
3030         * configure.in: Compute sizeofs only for double and long double.
3031         Use them to define and subst HAVE_LONG_DOUBLE.  Include comments
3032         into AC_DEFINE instead of using acconfig.h.  Create
3033         include/ffitarget.h instead of include/fficonfig.h.  Rename
3034         MIPS_GCC to MIPS_IRIX, drop MIPS_SGI since we are in gcc's tree.
3035         AC_DEFINE EH_FRAME_FLAGS.
3036         * include/Makefile.am (DISTCLEANFILES): New automake macro.
3037         (hack_DATA): Add ffitarget.h.
3038         * include/ffi.h.in: Remove all system specific definitions.
3039         Declare raw API even if it is not installed, why bother?
3040         Use limits.h instead of SIZEOF_* to define ffi_type_*.  Do
3041         not define EH_FRAME_FLAGS, it is in fficonfig.h now.  Include
3042         ffitarget.h instead of fficonfig.h.  Remove ALIGN macro.
3043         (UINT_ARG, INT_ARG): Removed, use ffi_arg and ffi_sarg instead.
3044         * include/ffi_common.h (bool): Do not define.
3045         (ffi_assert): Accept failed assertion.
3046         (ffi_type_test): Return void and accept file/line.
3047         (FFI_ASSERT): Pass stringized failed assertion.
3048         (FFI_ASSERT_AT): New macro.
3049         (FFI_ASSERT_VALID_TYPE): New macro.
3050         (UINT8, SINT8, UINT16, SINT16, UINT32, SINT32,
3051         UINT64, SINT64): Define here with gcc's __attribute__ macro
3052         instead of in ffi.h
3053         (FLOAT32, ALIGN): Define here instead of in ffi.h
3054         * include/ffi-mips.h: Removed.  Its content moved to
3055         src/mips/ffitarget.h after separating assembly and C sections.
3056         * src/alpha/ffi.c, src/alpha/ffi.c, src/java_raw_api.c
3057         src/prep_cif.c, src/raw_api.c, src/ia64/ffi.c,
3058         src/mips/ffi.c, src/mips/n32.S, src/mips/o32.S,
3059         src/mips/ffitarget.h, src/sparc/ffi.c, src/x86/ffi64.c:
3060         SIZEOF_ARG -> FFI_SIZEOF_ARG.
3061         * src/ia64/ffi.c: Include stdbool.h (provided by GCC 2.95+).
3062         * src/debug.c (ffi_assert): Accept stringized failed assertion.
3063         (ffi_type_test): Rewritten.
3064         * src/prep-cif.c (initialize_aggregate, ffi_prep_cif): Call
3065         FFI_ASSERT_VALID_TYPE.
3066         * src/alpha/ffitarget.h, src/arm/ffitarget.h,
3067         src/ia64/ffitarget.h, src/m68k/ffitarget.h,
3068         src/mips/ffitarget.h, src/powerpc/ffitarget.h,
3069         src/s390/ffitarget.h, src/sh/ffitarget.h,
3070         src/sh64/ffitarget.h, src/sparc/ffitarget.h,
3071         src/x86/ffitarget.h: New files.
3072         * src/alpha/osf.S, src/arm/sysv.S, src/ia64/unix.S,
3073         src/m68k/sysv.S, src/mips/n32.S, src/mips/o32.S,
3074         src/powerpc/aix.S, src/powerpc/darwin.S,
3075         src/powerpc/ffi_darwin.c, src/powerpc/linux64.S,
3076         src/powerpc/linux64_closure.S, src/powerpc/ppc_closure.S,
3077         src/powerpc/sysv.S, src/s390/sysv.S, src/sh/sysv.S,
3078         src/sh64/sysv.S, src/sparc/v8.S, src/sparc/v9.S,
3079         src/x86/sysv.S, src/x86/unix64.S, src/x86/win32.S:
3080         include fficonfig.h
3081
3082 2003-10-20  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3083
3084         * src/mips/ffi.c: Use _ABIN32, _ABIO32 instead of external
3085         _MIPS_SIM_NABI32, _MIPS_SIM_ABI32.
3086
3087 2003-10-19  Andreas Tobler  <a.tobler@schweiz.ch>
3088
3089         * src/powerpc/ffi_darwin.c (ffi_prep_args): Declare bytes again.
3090         Used when FFI_DEBUG = 1.
3091
3092 2003-10-14  Alan Modra  <amodra@bigpond.net.au>
3093
3094         * src/types.c (double, longdouble): Default POWERPC64 to 8 byte size
3095         and align.
3096
3097 2003-10-06  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3098
3099         * include/ffi_mips.h: Define FFI_MIPS_N32 for N32/N64 ABIs,
3100         FFI_MIPS_O32 for O32 ABI.
3101
3102 2003-10-01  Andreas Tobler  <a.tobler@schweiz.ch>
3103
3104         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_64 for
3105         SPARC64. Cleanup whitespaces.
3106
3107 2003-09-19  Andreas Tobler  <a.tobler@schweiz.ch>
3108
3109         * testsuite/libffi.call/closure_fn0.c: Xfail mips, arm,
3110         strongarm, xscale. Cleanup whitespaces.
3111         * testsuite/libffi.call/closure_fn1.c: Likewise.
3112         * testsuite/libffi.call/closure_fn2.c: Likewise.
3113         * testsuite/libffi.call/closure_fn3.c: Likewise.
3114         * testsuite/libffi.call/cls_12byte.c: Likewise.
3115         * testsuite/libffi.call/cls_16byte.c: Likewise.
3116         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3117         * testsuite/libffi.call/cls_20byte.c: Likewise.
3118         * testsuite/libffi.call/cls_24byte.c: Likewise.
3119         * testsuite/libffi.call/cls_2byte.c: Likewise.
3120         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3121         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3122         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3123         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3124         * testsuite/libffi.call/cls_4byte.c: Likewise.
3125         * testsuite/libffi.call/cls_5byte.c: Likewise.
3126         * testsuite/libffi.call/cls_6byte.c: Likewise.
3127         * testsuite/libffi.call/cls_7byte.c: Likewise.
3128         * testsuite/libffi.call/cls_8byte.c: Likewise.
3129         * testsuite/libffi.call/cls_double.c: Likewise.
3130         * testsuite/libffi.call/cls_float.c: Likewise.
3131         * testsuite/libffi.call/cls_uchar.c: Likewise.
3132         * testsuite/libffi.call/cls_uint.c: Likewise.
3133         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3134         * testsuite/libffi.call/cls_ushort.c: Likewise.
3135         * testsuite/libffi.call/nested_struct.c: Likewise.
3136         * testsuite/libffi.call/nested_struct1.c: Likewise.
3137         * testsuite/libffi.call/problem1.c: Likewise.
3138         * testsuite/libffi.special/unwindtest.cc: Likewise.
3139         * testsuite/libffi.call/pyobjc-tc.c: Cleanup whitespaces.
3140
3141 2003-09-18  David Edelsohn  <edelsohn@gnu.org>
3142
3143         * src/powerpc/aix.S: Cleanup whitespaces.
3144         * src/powerpc/aix_closure.S: Likewise.
3145
3146 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3147
3148         * src/powerpc/darwin.S: Cleanup whitespaces, comment formatting.
3149         * src/powerpc/darwin_closure.S: Likewise.
3150         * src/powerpc/ffi_darwin.c: Likewise.
3151
3152 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3153             David Edelsohn  <edelsohn@gnu.org>
3154
3155         * src/types.c (double): Add AIX and Darwin to the right TYPEDEF.
3156         * src/powerpc/aix_closure.S: Remove the pointer to the outgoing
3157         parameter stack.
3158         * src/powerpc/darwin_closure.S: Likewise.
3159         * src/powerpc/ffi_darwin.c (ffi_prep_args): Handle structures
3160         according to the Darwin/AIX ABI.
3161         (ffi_prep_cif_machdep): Likewise.
3162         (ffi_closure_helper_DARWIN): Likewise.
3163         Remove the outgoing parameter stack logic. Simplify the evaluation
3164         of the different CASE types.
3165         (ffi_prep_clousure): Avoid the casts on lvalues. Change the branch
3166         statement in the trampoline code.
3167
3168 2003-09-18  Kaz Kojima  <kkojima@gcc.gnu.org>
3169
3170         * src/sh/ffi.c (ffi_prep_args): Take account into the alignement
3171         for the register size.
3172         (ffi_closure_helper_SYSV): Handle the structure return value
3173         address correctly.
3174         (ffi_closure_helper_SYSV): Return the appropriate type when
3175         the registers are used for the structure return value.
3176         * src/sh/sysv.S (ffi_closure_SYSV): Fix the stack layout for
3177         the 64-bit return value.  Update copyright years.
3178
3179 2003-09-17  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3180
3181         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Search in
3182         srcdir for ffi_mips.h.
3183
3184 2003-09-12  Alan Modra  <amodra@bigpond.net.au>
3185
3186         * src/prep_cif.c (initialize_aggregate): Include tail padding in
3187         structure size.
3188         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Correct
3189         placement of float result.
3190         * testsuite/libffi.special/unwindtest.cc (closure_test_fn1): Correct
3191         cast of "resp" for big-endian 64 bit machines.
3192
3193 2003-09-11  Alan Modra  <amodra@bigpond.net.au>
3194
3195         * src/types.c (double, longdouble): Merge identical SH and ARM
3196         typedefs, and add POWERPC64.
3197         * src/powerpc/ffi.c (ffi_prep_args64): Correct next_arg calc for
3198         struct split over gpr and rest.
3199         (ffi_prep_cif_machdep): Correct intarg_count for structures.
3200         * src/powerpc/linux64.S (ffi_call_LINUX64): Fix gpr offsets.
3201
3202 2003-09-09  Andreas Tobler  <a.tobler@schweiz.ch>
3203
3204         * src/powerpc/ffi.c (ffi_closure_helper_SYSV) Handle struct
3205         passing correctly.
3206
3207 2003-09-09  Alan Modra  <amodra@bigpond.net.au>
3208
3209         * configure: Regenerate.
3210
3211 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3212
3213         * Makefile.am: Remove build rules for ffitest.
3214         * Makefile.in: Rebuilt.
3215
3216 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3217
3218         * src/java_raw_api.c: Include <stdlib.h> to fix compiler warning
3219         about implicit declaration of abort().
3220
3221 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3222
3223         * Makefile.am: Add dejagnu test framework. Fixes PR other/11411.
3224         * Makefile.in: Rebuilt.
3225         * configure.in: Add dejagnu test framework.
3226         * configure: Rebuilt.
3227
3228         * testsuite/Makefile.am: New file.
3229         * testsuite/Makefile.in: Built
3230         * testsuite/lib/libffi-dg.exp: New file.
3231         * testsuite/config/default.exp: Likewise.
3232         * testsuite/libffi.call/call.exp: Likewise.
3233         * testsuite/libffi.call/ffitest.h: Likewise.
3234         * testsuite/libffi.call/closure_fn0.c: Likewise.
3235         * testsuite/libffi.call/closure_fn1.c: Likewise.
3236         * testsuite/libffi.call/closure_fn2.c: Likewise.
3237         * testsuite/libffi.call/closure_fn3.c: Likewise.
3238         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3239         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3240         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3241         * testsuite/libffi.call/cls_2byte.c: Likewise.
3242         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3243         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3244         * testsuite/libffi.call/cls_4byte.c: Likewise.
3245         * testsuite/libffi.call/cls_5byte.c: Likewise.
3246         * testsuite/libffi.call/cls_6byte.c: Likewise.
3247         * testsuite/libffi.call/cls_7byte.c: Likewise.
3248         * testsuite/libffi.call/cls_8byte.c: Likewise.
3249         * testsuite/libffi.call/cls_12byte.c: Likewise.
3250         * testsuite/libffi.call/cls_16byte.c: Likewise.
3251         * testsuite/libffi.call/cls_20byte.c: Likewise.
3252         * testsuite/libffi.call/cls_24byte.c: Likewise.
3253         * testsuite/libffi.call/cls_double.c: Likewise.
3254         * testsuite/libffi.call/cls_float.c: Likewise.
3255         * testsuite/libffi.call/cls_uchar.c: Likewise.
3256         * testsuite/libffi.call/cls_uint.c: Likewise.
3257         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3258         * testsuite/libffi.call/cls_ushort.c: Likewise.
3259         * testsuite/libffi.call/float.c: Likewise.
3260         * testsuite/libffi.call/float1.c: Likewise.
3261         * testsuite/libffi.call/float2.c: Likewise.
3262         * testsuite/libffi.call/many.c: Likewise.
3263         * testsuite/libffi.call/many_win32.c: Likewise.
3264         * testsuite/libffi.call/nested_struct.c: Likewise.
3265         * testsuite/libffi.call/nested_struct1.c: Likewise.
3266         * testsuite/libffi.call/pyobjc-tc.c: Likewise.
3267         * testsuite/libffi.call/problem1.c: Likewise.
3268         * testsuite/libffi.call/promotion.c: Likewise.
3269         * testsuite/libffi.call/return_ll.c: Likewise.
3270         * testsuite/libffi.call/return_sc.c: Likewise.
3271         * testsuite/libffi.call/return_uc.c: Likewise.
3272         * testsuite/libffi.call/strlen.c: Likewise.
3273         * testsuite/libffi.call/strlen_win32.c: Likewise.
3274         * testsuite/libffi.call/struct1.c: Likewise.
3275         * testsuite/libffi.call/struct2.c: Likewise.
3276         * testsuite/libffi.call/struct3.c: Likewise.
3277         * testsuite/libffi.call/struct4.c: Likewise.
3278         * testsuite/libffi.call/struct5.c: Likewise.
3279         * testsuite/libffi.call/struct6.c: Likewise.
3280         * testsuite/libffi.call/struct7.c: Likewise.
3281         * testsuite/libffi.call/struct8.c: Likewise.
3282         * testsuite/libffi.call/struct9.c: Likewise.
3283         * testsuite/libffi.special/special.exp: New file.
3284         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3285         * testsuite/libffi.special/unwindtest.cc: Likewise.
3286
3287
3288 2003-08-13  Kaz Kojima  <kkojima@gcc.gnu.org>
3289
3290         * src/sh/ffi.c (OFS_INT16): Set 0 for little endian case.  Update
3291         copyright years.
3292
3293 2003-08-02  Alan Modra  <amodra@bigpond.net.au>
3294
3295         * src/powerpc/ffi.c (ffi_prep_args64): Modify for changed gcc
3296         structure passing.
3297         (ffi_closure_helper_LINUX64): Likewise.
3298         * src/powerpc/linux64.S: Remove code writing to parm save area.
3299         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Use return
3300         address in lr from ffi_closure_helper_LINUX64 call to calculate
3301         table address.  Optimize function tail.
3302
3303 2003-07-28  Andreas Tobler  <a.tobler@schweiz.ch>
3304
3305         * src/sparc/ffi.c: Handle all floating point registers.
3306         * src/sparc/v9.S: Likewise. Fixes second part of PR target/11410.
3307
3308 2003-07-11  Gerald Pfeifer  <pfeifer@dbai.tuwien.ac.at>
3309
3310         * README: Note that libffi is not part of GCC.  Update the project
3311         URL and status.
3312
3313 2003-06-19  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3314
3315         * src/powerpc/ppc_closure.S: Include ffi.h.
3316
3317 2003-06-13  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3318
3319         * src/x86/sysv.S: Avoid gas-only .uleb128/.sleb128 directives.
3320         Use C style comments.
3321
3322 2003-06-13  Kaz Kojima  <kkojima@rr.iij4u.or.jp>
3323
3324         * Makefile.am: Add SHmedia support.  Fix a typo of SH support.
3325         * Makefile.in: Regenerate.
3326         * configure.in (sh64-*-linux*, sh5*-*-linux*): Add target.
3327         * configure: Regenerate.
3328         * include/ffi.h.in: Add SHmedia support.
3329         * src/sh64/ffi.c: New file.
3330         * src/sh64/sysv.S: New file.
3331
3332 2003-05-16  Jakub Jelinek  <jakub@redhat.com>
3333
3334         * configure.in (HAVE_RO_EH_FRAME): Check whether .eh_frame section
3335         should be read-only.
3336         * configure: Rebuilt.
3337         * fficonfig.h.in: Rebuilt.
3338         * include/ffi.h.in (EH_FRAME_FLAGS): Define.
3339         * src/alpha/osf.S: Use EH_FRAME_FLAGS.
3340         * src/powerpc/linux64.S: Likewise.
3341         * src/powerpc/linux64_closure.S: Likewise.  Include ffi.h.
3342         * src/powerpc/sysv.S: Use EH_FRAME_FLAGS.  Use pcrel encoding
3343         if -fpic/-fPIC/-mrelocatable.
3344         * src/powerpc/powerpc_closure.S: Likewise.
3345         * src/sparc/v8.S: If HAVE_RO_EH_FRAME is defined, don't include
3346         #write in .eh_frame flags.
3347         * src/sparc/v9.S: Likewise.
3348         * src/x86/unix64.S: Use EH_FRAME_FLAGS.
3349         * src/x86/sysv.S: Likewise.  Use pcrel encoding if -fpic/-fPIC.
3350         * src/s390/sysv.S: Use EH_FRAME_FLAGS.  Include ffi.h.
3351
3352 2003-05-07  Jeff Sturm  <jsturm@one-point.com>
3353
3354         Fixes PR bootstrap/10656
3355         * configure.in (HAVE_AS_REGISTER_PSEUDO_OP): Test assembler
3356         support for .register pseudo-op.
3357         * src/sparc/v8.S: Use it.
3358         * fficonfig.h.in: Rebuilt.
3359         * configure: Rebuilt.
3360
3361 2003-04-18  Jakub Jelinek  <jakub@redhat.com>
3362
3363         * include/ffi.h.in (POWERPC64): Define if 64-bit.
3364         (enum ffi_abi): Add FFI_LINUX64 on POWERPC.
3365         Make it the default on POWERPC64.
3366         (FFI_TRAMPOLINE_SIZE): Define to 24 on POWERPC64.
3367         * configure.in: Change powerpc-*-linux* into powerpc*-*-linux*.
3368         * configure: Rebuilt.
3369         * src/powerpc/ffi.c (hidden): Define.
3370         (ffi_prep_args_SYSV): Renamed from
3371         ffi_prep_args.  Cast pointers to unsigned long to shut up warnings.
3372         (NUM_GPR_ARG_REGISTERS64, NUM_FPR_ARG_REGISTERS64,
3373         ASM_NEEDS_REGISTERS64): New.
3374         (ffi_prep_args64): New function.
3375         (ffi_prep_cif_machdep): Handle FFI_LINUX64 ABI.
3376         (ffi_call): Likewise.
3377         (ffi_prep_closure): Likewise.
3378         (flush_icache): Surround by #ifndef POWERPC64.
3379         (ffi_dblfl): New union type.
3380         (ffi_closure_helper_SYSV): Use it to avoid aliasing problems.
3381         (ffi_closure_helper_LINUX64): New function.
3382         * src/powerpc/ppc_closure.S: Surround whole file by #ifndef
3383         __powerpc64__.
3384         * src/powerpc/sysv.S: Likewise.
3385         (ffi_call_SYSV): Rename ffi_prep_args to ffi_prep_args_SYSV.
3386         * src/powerpc/linux64.S: New file.
3387         * src/powerpc/linux64_closure.S: New file.
3388         * Makefile.am (EXTRA_DIST): Add src/powerpc/linux64.S and
3389         src/powerpc/linux64_closure.S.
3390         (TARGET_SRC_POWERPC): Likewise.
3391
3392         * src/ffitest.c (closure_test_fn, closure_test_fn1, closure_test_fn2,
3393         closure_test_fn3): Fix result printing on big-endian 64-bit
3394         machines.
3395         (main): Print tst2_arg instead of uninitialized tst2_result.
3396
3397         * src/ffitest.c (main): Hide what closure pointer really points to
3398         from the compiler.
3399
3400 2003-04-16  Richard Earnshaw  <rearnsha@arm.com>
3401
3402         * configure.in (arm-*-netbsdelf*): Add configuration.
3403         (configure): Regenerated.
3404
3405 2003-04-04  Loren J. Rittle  <ljrittle@acm.org>
3406
3407         * include/Makefile.in: Regenerate.
3408
3409 2003-03-21  Zdenek Dvorak  <rakdver@atrey.karlin.mff.cuni.cz>
3410
3411         * libffi/include/ffi.h.in: Define X86 instead of X86_64 in 32
3412         bit mode.
3413         * libffi/src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV):
3414         Receive closure pointer through parameter, read args using
3415         __builtin_dwarf_cfa.
3416         (FFI_INIT_TRAMPOLINE): Send closure reference through eax.
3417
3418 2003-03-12  Andreas Schwab  <schwab@suse.de>
3419
3420         * configure.in: Avoid trailing /. in toolexeclibdir.
3421         * configure: Rebuilt.
3422
3423 2003-03-03  Andreas Tobler <a.tobler@schweiz.ch>
3424
3425         * src/powerpc/darwin_closure.S: Recode to fit dynamic libraries.
3426
3427 2003-02-06  Andreas Tobler <a.tobler@schweiz.ch>
3428
3429         * libffi/src/powerpc/darwin_closure.S:
3430         Fix alignement bug, allocate 8 bytes for the result.
3431         * libffi/src/powerpc/aix_closure.S:
3432         Likewise.
3433         * libffi/src/powerpc/ffi_darwin.c:
3434         Update stackframe description for aix/darwin_closure.S.
3435
3436 2003-02-06  Jakub Jelinek  <jakub@redhat.com>
3437
3438         * src/s390/ffi.c (ffi_closure_helper_SYSV): Add&n