OSDN Git Service

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