OSDN Git Service

bdd86bffa74d8e85ee216774e1aff39da76f2f0f
[pf3gnuchains/gcc-fork.git] / libffi / ChangeLog
1 2010-07-02  Jakub Jelinek  <jakub@redhat.com>
2
3         * Makefile.am (AM_MAKEFLAGS): Pass also mandir to submakes.
4         * Makefile.in: Regenerated.
5
6 2010-05-19  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
7
8         * configure.ac (libffi_cv_as_x86_pcrel): Check for illegal in as
9         output, too.
10         (libffi_cv_as_ascii_pseudo_op): Check for .ascii.
11         (libffi_cv_as_string_pseudo_op): Check for .string.
12         * configure: Regenerate.
13         * fficonfig.h.in: Regenerate.
14         * src/x86/sysv.S (.eh_frame): Use .ascii, .string or error.
15
16 2010-04-07  Jakub Jelinek  <jakub@redhat.com>
17
18         * regex.c (byte_re_match_2_internal): Avoid set but not used
19         warning.
20
21 2010-04-02  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
22
23         * Makefile.in: Regenerate.
24         * aclocal.m4: Regenerate.
25         * include/Makefile.in: Regenerate.
26         * man/Makefile.in: Regenerate.
27         * testsuite/Makefile.in: Regenerate.
28
29 2010-03-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
30
31         * configure.ac (libffi_cv_as_x86_64_unwind_section_type): New test.
32         * configure: Regenerate.
33         * fficonfig.h.in: Regenerate.
34         * libffi/src/x86/unix64.S (.eh_frame)
35         [HAVE_AS_X86_64_UNWIND_SECTION_TYPE]: Use @unwind section type.
36
37 2010-02-24  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
38
39         * doc/libffi.texi (The Closure API): Fix typo.
40         * doc/libffi.info: Remove.
41
42 2010-02-15  Matthias Klose  <doko@ubuntu.com>
43
44         * src/arm/sysv.S (__ARM_ARCH__): Define for processor
45         __ARM_ARCH_7EM__.
46
47 2010-01-07  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
48
49         PR libffi/40701
50         * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRIdLL,
51         PRIuLL, PRId64, PRIu64, PRIuPTR): Define.
52         * testsuite/libffi.call/cls_align_sint64.c: Add -Wno-format on
53         alpha*-dec-osf*.
54         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
55         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
56         * testsuite/libffi.call/return_ll1.c: Likewise.
57         * testsuite/libffi.call/stret_medium2.c: Likewise.
58         * testsuite/libffi.special/ffitestcxx.h (allocate_mmap): Cast
59         MAP_FAILED to char *.
60
61 2010-01-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
62
63         * src/mips/n32.S: Use .abicalls and .eh_frame with __GNUC__.
64
65 2009-12-31  Anthony Green  <green@redhat.com>
66
67         * README: Update for libffi 3.0.9.
68
69 2009-12-27  Matthias Klose  <doko@ubuntu.com>
70
71         * configure.ac (HAVE_LONG_DOUBLE): Define for mips when
72         appropriate.
73         * configure: Rebuilt.
74
75 2009-12-26  Anthony Green  <green@redhat.com>
76
77         * testsuite/libffi.call/cls_longdouble_va.c: Mark as xfail for
78         avr32*-*-*.
79         * testsuite/libffi.call/cls_double_va.c: Ditto.
80
81 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
82
83         * testsuite/libffi.call/ffitest.h: Conditionally include stdint.h
84         and inttypes.h.
85         * testsuite/libffi.special/unwindtest.cc: Ditto.
86
87 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
88
89         * configure.ac: Add amd64-*-openbsd*.
90         * configure: Rebuilt.
91         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Link
92         openbsd programs with -lpthread.
93
94 2009-12-26  Anthony Green  <green@redhat.com>
95
96         * testsuite/libffi.call/cls_double_va.c,
97         testsuite/libffi.call/cls_longdouble.c,
98         testsuite/libffi.call/cls_longdouble_va.c,
99         testsuite/libffi.call/cls_pointer.c,
100         testsuite/libffi.call/cls_pointer_stack.c: Remove xfail for
101         mips*-*-* and arm*-*-*.
102         * testsuite/libffi.call/cls_align_longdouble_split.c,
103         testsuite/libffi.call/cls_align_longdouble_split2.c,
104         testsuite/libffi.call/stret_medium2.c,
105         testsuite/libffi.call/stret_medium.c,
106         testsuite/libffi.call/stret_large.c,
107         testsuite/libffi.call/stret_large2.c: Remove xfail for arm*-*-*.
108
109 2009-12-31  Kay Tietz  <ktietz70@googlemail.com>
110
111         * testsuite/libffi.call/ffitest.h,
112         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRuLL): Fix
113         definitions.
114
115 2009-12-31  Carlo Bramini  <carlo.bramix@libero.it>
116
117         * configure.ac (AM_LTLDFLAGS): Define for windows hosts.
118         * Makefile.am (libffi_la_LDFLAGS): Add AM_LTLDFLAGS.
119         * configure: Rebuilt.
120         * Makefile.in: Rebuilt.
121
122 2009-12-31  Anthony Green  <green@redhat.com>
123             Blake Chaffin.
124
125         * testsuite/libffi.call/huge_struct.c: New test case from Blake
126         Chaffin @ Apple.
127
128 2009-12-28  David Edelsohn  <edelsohn@gnu.org>
129
130         * src/powerpc/ffi_darwin.c (ffi_prep_args): Copy abi and nargs to
131         local variables.
132         (aix_adjust_aggregate_sizes): New function.
133         (ffi_prep_cif_machdep): Call it.
134
135 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
136
137         * configure.ac: Define FFI_MMAP_EXEC_WRIT for the given targets.
138         * configure: Regenerate.
139         * fficonfig.h.in: Likewise.
140         * src/closures.c: Remove the FFI_MMAP_EXEC_WRIT definition for
141         Solaris/x86.
142
143 2009-12-26  Andreas Schwab  <schwab@linux-m68k.org>
144
145         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Advance intarg_count
146         when a float arguments is passed in memory.
147         (ffi_closure_helper_SYSV): Mark general registers as used up when
148         a 64bit or soft-float long double argument is passed in memory.
149
150 2009-12-25  Matthias Klose  <doko@ubuntu.com>
151
152         * man/ffi_call.3: Fix #include in examples.
153         * doc/libffi.texi: Add dircategory.
154
155 2009-12-25  Frank Everdij <f.p.x.everdij@tudelft.nl>
156
157         * include/ffi.h.in: Placed '__GNUC__' ifdef around
158         '__attribute__((aligned(8)))' in ffi_closure, fixes compile for
159         IRIX MIPSPro c99.
160         * include/ffi_common.h: Added '__sgi' define to non
161         '__attribute__((__mode__()))' integer typedefs.
162         * src/mips/ffi.c (ffi_call, ffi_closure_mips_inner_O32,
163         ffi_closure_mips_inner_N32): Added 'defined(_MIPSEB)' to BE check.
164         (ffi_closure_mips_inner_O32, ffi_closure_mips_inner_N32): Added
165         FFI_LONGDOUBLE support and alignment(N32 only).
166         * src/mips/ffitarget.h: Corrected '#include <sgidefs.h>' for IRIX and
167         fixed non '__attribute__((__mode__()))' integer typedefs.
168         * src/mips/n32.S: Put '#ifdef linux' around '.abicalls' and '.eh_frame'
169         since they are Linux/GNU Assembler specific.
170
171 2009-12-25  Bradley Smith  <brad@brad-smith.co.uk>
172
173         * configure.ac, Makefile.am, src/avr32/ffi.c,
174         src/avr32/ffitarget.h,
175         src/avr32/sysv.S: Add AVR32 port.
176         * configure, Makefile.in: Rebuilt.
177
178 2009-12-21  Andreas Tobler  <a.tobler@schweiz.org>
179
180         * configure.ac: Make i?86 build on FreeBSD and OpenBSD.
181         * configure: Regenerate.
182
183 2009-12-15  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
184
185         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on PA HP-UX.
186
187 2009-12-13  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
188
189         * src/pa/ffi.c (ffi_closure_inner_pa32): Handle FFI_TYPE_LONGDOUBLE
190         type on HP-UX.
191
192 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
193
194         * src/sparc/ffi.c (ffi_closure_sparc_inner_v9): Properly align 'long
195         double' arguments.
196
197 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
198
199         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on Solaris < 10.
200
201 2009-12-10  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
202
203         PR libffi/40700
204         * src/closures.c [X86_64 && __sun__ && __svr4__]
205         (FFI_MMAP_EXEC_WRIT): Define.
206
207 2009-12-08  David Daney  <ddaney@caviumnetworks.com>
208
209         * testsuite/libffi.call/stret_medium.c: Remove xfail for mips*-*-*
210         * testsuite/libffi.call/cls_align_longdouble_split2.c: Same.
211         * testsuite/libffi.call/stret_large.c: Same.
212         * testsuite/libffi.call/cls_align_longdouble_split.c: Same.
213         * testsuite/libffi.call/stret_large2.c: Same.
214         * testsuite/libffi.call/stret_medium2.c: Same.
215
216 2009-12-07  David Edelsohn  <edelsohn@gnu.org>
217
218         * src/powerpc/aix_closure.S (libffi_closure_ASM): Fix tablejump
219         typo.
220
221 2009-12-05  David Edelsohn  <edelsohn@gnu.org>
222
223         * src/powerpc/aix.S: Update AIX32 code to be consistent with AIX64
224         code.
225         * src/powerpc/aix_closure.S: Same.
226
227 2009-12-05  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
228
229         * Makefile.in: Regenerate.
230         * configure: Regenerate.
231         * include/Makefile.in: Regenerate.
232         * man/Makefile.in: Regenerate.
233         * testsuite/Makefile.in: Regenerate.
234
235 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
236
237         * src/powerpc/aix_closure.S: Reorganize 64-bit code to match
238         linux64_closure.S.
239
240 2009-12-04  Uros Bizjak  <ubizjak@gmail.com>
241
242         PR libffi/41908
243         * src/x86/ffi64.c (classify_argument): Update from
244         gcc/config/i386/i386.c.
245         (ffi_closure_unix64_inner): Do not use the address of two consecutive
246         SSE registers directly.
247         * testsuite/libffi.call/cls_dbls_struct.c (main): Remove xfail
248         for x86_64 linux targets.
249
250 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
251
252         * src/powerpc/ffi_darwin.c (ffi_closure_helper_DARWIN): Increment
253         pfr for long double split between fpr13 and stack.
254
255 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
256
257         * src/powerpc/ffi_darwin.c (ffi_prep_args): Increment next_arg and
258         fparg_count twice for long double.
259
260 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
261
262         PR libffi/42243
263         * src/powerpc/ffi_darwin.c (ffi_prep_args): Remove extra parentheses.
264
265 2009-12-03  Uros Bizjak  <ubizjak@gmail.com>
266
267         * testsuite/libffi.call/cls_longdouble_va.c (main): Fix format string.
268         Remove xfails for x86 linux targets.
269
270 2009-12-02  David Edelsohn  <edelsohn@gnu.org>
271
272         * src/powerpc/ffi_darwin.c (ffi_prep_args): Fix typo in INT64
273         case.
274
275 2009-12-01  David Edelsohn  <edelsohn@gnu.org>
276
277         * src/powerpc/aix.S (ffi_call_AIX): Convert to more standard
278         register usage.  Call ffi_prep_args directly.  Add long double
279         return value support.
280         * src/powerpc/ffi_darwin.c (ffi_prep_args): Double arg increment
281         applies to FFI_TYPE_DOUBLE.  Correct fpr_base increment typo.
282         Separate FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases.
283         (ffi_prep_cif_machdep): Only 16 byte stack alignment in 64 bit
284         mode.
285         (ffi_closure_helper_DARWIN): Remove nf and ng counters.  Move temp
286         into case.
287         * src/powerpc/aix_closure.S: Maintain 16 byte stack alignment.
288         Allocate result area between params and FPRs.
289
290 2009-11-30  David Edelsohn  <edelsohn@gnu.org>
291
292         PR target/35484
293         * src/powerpc/ffitarget.h (POWERPC64): Define for PPC64 Linux and
294         AIX64.
295         * src/powerpc/aix.S: Implement AIX64 version.
296         * src/powerpc/aix_closure.S: Implement AIX64 version.
297         (ffi_closure_ASM): Use extsb, lha and displament addresses.
298         * src/powerpc/ffi_darwin.c (ffi_prep_args): Implement AIX64
299         support.
300         (ffi_prep_cif_machdep): Same.
301         (ffi_call): Same.
302         (ffi_closure_helper_DARWIN): Same.
303
304 2009-11-02  Andreas Tobler  <a.tobler@schweiz.org>
305
306         PR libffi/41908
307         * testsuite/libffi.call/testclosure.c: New test.
308
309 2009-09-28  Kai Tietz  <kai.tietz@onevision.com>
310
311         * src/x86/win64.S (_ffi_call_win64 stack): Remove for gnu
312         assembly version use of ___chkstk.
313
314 2009-09-23  Matthias Klose  <doko@ubuntu.com>
315
316         PR libffi/40242, PR libffi/41443
317         * src/arm/sysv.S (__ARM_ARCH__): Define for processors
318         __ARM_ARCH_6T2__, __ARM_ARCH_6M__, __ARM_ARCH_7__,
319         __ARM_ARCH_7A__, __ARM_ARCH_7R__, __ARM_ARCH_7M__.
320         Change the conditionals to __SOFTFP__ || __ARM_EABI__
321         for -mfloat-abi=softfp to work.
322
323 2009-09-17  Loren J. Rittle  <ljrittle@acm.org>
324
325         PR testsuite/32843 (strikes again)
326         * src/x86/ffi.c (ffi_prep_cif_machdep): Add X86_FREEBSD to
327         enable proper extension on char and short.
328
329 2009-09-15  David Daney  <ddaney@caviumnetworks.com>
330
331         * src/java_raw_api.c (ffi_java_raw_to_rvalue): Remove special
332         handling for FFI_TYPE_POINTER.
333         * src/mips/ffitarget.h (FFI_TYPE_STRUCT_D_SOFT,
334         FFI_TYPE_STRUCT_F_SOFT, FFI_TYPE_STRUCT_DD_SOFT,
335         FFI_TYPE_STRUCT_FF_SOFT, FFI_TYPE_STRUCT_FD_SOFT,
336         FFI_TYPE_STRUCT_DF_SOFT, FFI_TYPE_STRUCT_SOFT): New defines.
337         (FFI_N32_SOFT_FLOAT, FFI_N64_SOFT_FLOAT): New ffi_abi enumerations.
338         (enum ffi_abi): Set FFI_DEFAULT_ABI for soft-float.
339         * src/mips/n32.S (ffi_call_N32): Add handling for soft-float
340         structure and pointer returns.
341         (ffi_closure_N32): Add handling for pointer returns.
342         * src/mips/ffi.c (ffi_prep_args, calc_n32_struct_flags,
343         calc_n32_return_struct_flags): Handle soft-float.
344         (ffi_prep_cif_machdep):  Handle soft-float, fix pointer handling.
345         (ffi_call_N32): Declare proper argument types.
346         (ffi_call, copy_struct_N32, ffi_closure_mips_inner_N32): Handle
347         soft-float.
348
349 2009-08-24  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
350
351         * configure.ac (AC_PREREQ): Bump to 2.64.
352
353 2009-08-22  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
354
355         * Makefile.am (install-html, install-pdf): Remove.
356         * Makefile.in: Regenerate.
357
358         * Makefile.in: Regenerate.
359         * aclocal.m4: Regenerate.
360         * configure: Regenerate.
361         * fficonfig.h.in: Regenerate.
362         * include/Makefile.in: Regenerate.
363         * man/Makefile.in: Regenerate.
364         * testsuite/Makefile.in: Regenerate.
365
366 2009-07-30  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
367
368         * configure.ac (_AC_ARG_VAR_PRECIOUS): Use m4_rename_force.
369
370 2009-07-24  Dave Korn  <dave.korn.cygwin@gmail.com>
371
372         PR libffi/40807
373         * src/x86/ffi.c (ffi_prep_cif_machdep): Also use sign/zero-extending
374         return types for X86_WIN32.
375         * src/x86/win32.S (_ffi_call_SYSV): Handle omitted return types.
376         (_ffi_call_STDCALL, _ffi_closure_SYSV, _ffi_closure_raw_SYSV,
377         _ffi_closure_STDCALL): Likewise.
378
379         * src/closures.c (is_selinux_enabled): Define to const 0 for Cygwin.
380         (dlmmap, dlmunmap): Also use these functions on Cygwin.
381
382 2009-07-11  Richard Sandiford  <rdsandiford@googlemail.com>
383
384         PR testsuite/40699
385         PR testsuite/40707
386         PR testsuite/40709
387         * testsuite/lib/libffi-dg.exp: Revert 2009-07-02, 2009-07-01 and
388         2009-06-30 commits.
389
390 2009-07-01  Richard Sandiford  <r.sandiford@uk.ibm.com>
391
392         * testsuite/lib/libffi-dg.exp (libffi-init): Set ld_library_path
393         to "" before adding paths.  (This reinstates an assignment that
394         was removed by my 2009-06-30 commit, but changes the initial
395         value from "." to "".)
396
397 2009-07-01  H.J. Lu  <hongjiu.lu@intel.com>
398
399         PR testsuite/40601
400         * testsuite/lib/libffi-dg.exp (libffi-init): Properly set
401         gccdir.  Adjust ld_library_path for gcc only if gccdir isn't
402         empty.
403
404 2009-06-30  Richard Sandiford  <r.sandiford@uk.ibm.com>
405
406         * testsuite/lib/libffi-dg.exp (libffi-init): Don't add "."
407         to ld_library_path.  Use add_path.  Add just find_libgcc_s
408         to ld_library_path, not every libgcc multilib directory.
409
410 2009-06-16  Wim Lewis  <wiml@hhhh.org>
411
412         * src/powerpc/ffi.c: Avoid clobbering cr3 and cr4, which are
413         supposed to be callee-saved.
414         * src/powerpc/sysv.S (small_struct_return_value): Fix overrun of
415         return buffer for odd-size structs.
416
417 2009-06-16  Andreas Tobler  <a.tobler@schweiz.org>
418
419         PR libffi/40444
420         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Add
421         allow_stack_execute for Darwin.
422
423 2009-06-16  Andrew Haley  <aph@redhat.com>
424
425         * configure.ac (TARGETDIR): Add missing blank lines.
426         * configure: Regenerate.
427
428 2009-06-16  Andrew Haley  <aph@redhat.com>
429
430         * testsuite/libffi.call/cls_align_sint64.c,
431         testsuite/libffi.call/cls_align_uint64.c,
432         testsuite/libffi.call/cls_longdouble_va.c,
433         testsuite/libffi.call/cls_ulonglong.c,
434         testsuite/libffi.call/return_ll1.c,
435         testsuite/libffi.call/stret_medium2.c: Fix printf format
436         specifiers.
437         * testsuite/libffi.call/ffitest.h,
438         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRIuLL): Define.
439
440 2009-06-15  Andrew Haley  <aph@redhat.com>
441
442         * testsuite/libffi.call/err_bad_typedef.c: xfail everywhere.
443         * testsuite/libffi.call/err_bad_abi.c: Likewise.
444
445 2009-06-12  Andrew Haley  <aph@redhat.com>
446
447         * Makefile.am: Remove info_TEXINFOS.
448
449 2009-06-12  Andrew Haley  <aph@redhat.com>
450
451         * ChangeLog.libffi: testsuite/libffi.call/cls_align_sint64.c,
452         testsuite/libffi.call/cls_align_uint64.c,
453         testsuite/libffi.call/cls_ulonglong.c,
454         testsuite/libffi.call/return_ll1.c,
455         testsuite/libffi.call/stret_medium2.c: Fix printf format
456         specifiers.
457         testsuite/libffi.special/unwindtest.cc: include stdint.h.
458
459 2009-06-11  Timothy Wall  <twall@users.sf.net>
460
461         * Makefile.am,
462         configure.ac,
463         include/ffi.h.in,
464         include/ffi_common.h,
465         src/closures.c,
466         src/dlmalloc.c,
467         src/x86/ffi.c,
468         src/x86/ffitarget.h,
469         src/x86/win64.S (new),
470         README: Added win64 support (mingw or MSVC)
471         * Makefile.in,
472         include/Makefile.in,
473         man/Makefile.in,
474         testsuite/Makefile.in,
475         configure,
476         aclocal.m4: Regenerated
477         * ltcf-c.sh: properly escape cygwin/w32 path
478         * man/ffi_call.3: Clarify size requirements for return value.
479         * src/x86/ffi64.c: Fix filename in comment.
480         * src/x86/win32.S: Remove unused extern.
481
482         * testsuite/libffi.call/closure_fn0.c,
483         testsuite/libffi.call/closure_fn1.c,
484         testsuite/libffi.call/closure_fn2.c,
485         testsuite/libffi.call/closure_fn3.c,
486         testsuite/libffi.call/closure_fn4.c,
487         testsuite/libffi.call/closure_fn5.c,
488         testsuite/libffi.call/closure_fn6.c,
489         testsuite/libffi.call/closure_stdcall.c,
490         testsuite/libffi.call/cls_12byte.c,
491         testsuite/libffi.call/cls_16byte.c,
492         testsuite/libffi.call/cls_18byte.c,
493         testsuite/libffi.call/cls_19byte.c,
494         testsuite/libffi.call/cls_1_1byte.c,
495         testsuite/libffi.call/cls_20byte.c,
496         testsuite/libffi.call/cls_20byte1.c,
497         testsuite/libffi.call/cls_24byte.c,
498         testsuite/libffi.call/cls_2byte.c,
499         testsuite/libffi.call/cls_3_1byte.c,
500         testsuite/libffi.call/cls_3byte1.c,
501         testsuite/libffi.call/cls_3byte2.c,
502         testsuite/libffi.call/cls_4_1byte.c,
503         testsuite/libffi.call/cls_4byte.c,
504         testsuite/libffi.call/cls_5_1_byte.c,
505         testsuite/libffi.call/cls_5byte.c,
506         testsuite/libffi.call/cls_64byte.c,
507         testsuite/libffi.call/cls_6_1_byte.c,
508         testsuite/libffi.call/cls_6byte.c,
509         testsuite/libffi.call/cls_7_1_byte.c,
510         testsuite/libffi.call/cls_7byte.c,
511         testsuite/libffi.call/cls_8byte.c,
512         testsuite/libffi.call/cls_9byte1.c,
513         testsuite/libffi.call/cls_9byte2.c,
514         testsuite/libffi.call/cls_align_double.c,
515         testsuite/libffi.call/cls_align_float.c,
516         testsuite/libffi.call/cls_align_longdouble.c,
517         testsuite/libffi.call/cls_align_longdouble_split.c,
518         testsuite/libffi.call/cls_align_longdouble_split2.c,
519         testsuite/libffi.call/cls_align_pointer.c,
520         testsuite/libffi.call/cls_align_sint16.c,
521         testsuite/libffi.call/cls_align_sint32.c,
522         testsuite/libffi.call/cls_align_sint64.c,
523         testsuite/libffi.call/cls_align_uint16.c,
524         testsuite/libffi.call/cls_align_uint32.c,
525         testsuite/libffi.call/cls_align_uint64.c,
526         testsuite/libffi.call/cls_dbls_struct.c,
527         testsuite/libffi.call/cls_double.c,
528         testsuite/libffi.call/cls_double_va.c,
529         testsuite/libffi.call/cls_float.c,
530         testsuite/libffi.call/cls_longdouble.c,
531         testsuite/libffi.call/cls_longdouble_va.c,
532         testsuite/libffi.call/cls_multi_schar.c,
533         testsuite/libffi.call/cls_multi_sshort.c,
534         testsuite/libffi.call/cls_multi_sshortchar.c,
535         testsuite/libffi.call/cls_multi_uchar.c,
536         testsuite/libffi.call/cls_multi_ushort.c,
537         testsuite/libffi.call/cls_multi_ushortchar.c,
538         testsuite/libffi.call/cls_pointer.c,
539         testsuite/libffi.call/cls_pointer_stack.c,
540         testsuite/libffi.call/cls_schar.c,
541         testsuite/libffi.call/cls_sint.c,
542         testsuite/libffi.call/cls_sshort.c,
543         testsuite/libffi.call/cls_uchar.c,
544         testsuite/libffi.call/cls_uint.c,
545         testsuite/libffi.call/cls_ulonglong.c,
546         testsuite/libffi.call/cls_ushort.c,
547         testsuite/libffi.call/err_bad_abi.c,
548         testsuite/libffi.call/err_bad_typedef.c,
549         testsuite/libffi.call/float2.c,
550         testsuite/libffi.call/huge_struct.c,
551         testsuite/libffi.call/nested_struct.c,
552         testsuite/libffi.call/nested_struct1.c,
553         testsuite/libffi.call/nested_struct10.c,
554         testsuite/libffi.call/nested_struct2.c,
555         testsuite/libffi.call/nested_struct3.c,
556         testsuite/libffi.call/nested_struct4.c,
557         testsuite/libffi.call/nested_struct5.c,
558         testsuite/libffi.call/nested_struct6.c,
559         testsuite/libffi.call/nested_struct7.c,
560         testsuite/libffi.call/nested_struct8.c,
561         testsuite/libffi.call/nested_struct9.c,
562         testsuite/libffi.call/problem1.c,
563         testsuite/libffi.call/return_ldl.c,
564         testsuite/libffi.call/return_ll1.c,
565         testsuite/libffi.call/stret_large.c,
566         testsuite/libffi.call/stret_large2.c,
567         testsuite/libffi.call/stret_medium.c,
568         testsuite/libffi.call/stret_medium2.c,
569         testsuite/libffi.special/unwindtest.cc: use ffi_closure_alloc instead
570         of checking for MMAP.  Use intptr_t instead of long casts.
571
572 2009-06-11  Kaz Kojima  <kkojima@gcc.gnu.org>
573
574         * testsuite/libffi.call/cls_longdouble_va.c: Add xfail sh*-*-linux-*.
575         * testsuite/libffi.call/err_bad_abi.c: Add xfail sh*-*-*.
576         * testsuite/libffi.call/err_bad_typedef.c: Likewise.
577
578 2009-06-09  Andrew Haley  <aph@redhat.com>
579
580         * src/x86/freebsd.S: Add missing file.
581
582 2009-06-08  Andrew Haley  <aph@redhat.com>
583
584         Import from libffi 3.0.8:
585
586         * doc/libffi.texi: New file.
587         * doc/libffi.info: Likewise.
588         * doc/stamp-vti: Likewise.
589         * man/Makefile.am: New file.
590         * man/ffi_call.3: New file.
591
592         * Makefile.am (EXTRA_DIST): Add src/x86/darwin64.S,
593         src/dlmalloc.c.
594         (nodist_libffi_la_SOURCES): Add X86_FREEBSD.
595
596         * configure.ac: Bump version to 3.0.8.
597         parisc*-*-linux*: Add.
598         i386-*-freebsd* | i386-*-openbsd*: Add.
599         powerpc-*-beos*: Add.
600         AM_CONDITIONAL X86_FREEBSD: Add.
601         AC_CONFIG_FILES: Add man/Makefile.
602
603         * include/ffi.h.in (FFI_FN): Change void (*)() to void (*)(void).
604
605 2009-06-08  Andrew Haley  <aph@redhat.com>
606
607         * README: Import from libffi 3.0.8.
608
609 2009-06-08  Andrew Haley  <aph@redhat.com>
610
611         * testsuite/libffi.call/err_bad_abi.c: Add xfails.
612         * testsuite/libffi.call/cls_longdouble_va.c: Add xfails.
613         * testsuite/libffi.call/cls_dbls_struct.c: Add xfail x86_64-*-linux-*.
614         * testsuite/libffi.call/err_bad_typedef.c: Add xfails.
615
616         * testsuite/libffi.call/stret_medium2.c: Add __UNUSED__ to args.
617         * testsuite/libffi.call/stret_medium.c: Likewise.
618         * testsuite/libffi.call/stret_large2.c: Likewise.
619         * testsuite/libffi.call/stret_large.c:  Likewise.
620
621 2008-12-26  Timothy Wall  <twall@users.sf.net>
622
623         * testsuite/libffi.call/cls_longdouble.c,
624         testsuite/libffi.call/cls_longdouble_va.c,
625         testsuite/libffi.call/cls_align_longdouble.c,
626         testsuite/libffi.call/cls_align_longdouble_split.c,
627         testsuite/libffi.call/cls_align_longdouble_split2.c: mark expected
628         failures on x86_64 cygwin/mingw.
629
630 2008-12-22  Timothy Wall  <twall@users.sf.net>
631
632         * testsuite/libffi.call/closure_fn0.c,
633         testsuite/libffi.call/closure_fn1.c,
634         testsuite/libffi.call/closure_fn2.c,
635         testsuite/libffi.call/closure_fn3.c,
636         testsuite/libffi.call/closure_fn4.c,
637         testsuite/libffi.call/closure_fn5.c,
638         testsuite/libffi.call/closure_fn6.c,
639         testsuite/libffi.call/closure_loc_fn0.c,
640         testsuite/libffi.call/closure_stdcall.c,
641         testsuite/libffi.call/cls_align_pointer.c,
642         testsuite/libffi.call/cls_pointer.c,
643         testsuite/libffi.call/cls_pointer_stack.c: use portable cast from
644         pointer to integer (intptr_t).
645         * testsuite/libffi.call/cls_longdouble.c: disable for win64.
646
647 2008-07-24  Anthony Green  <green@redhat.com>
648
649         * testsuite/libffi.call/cls_dbls_struct.c,
650         testsuite/libffi.call/cls_double_va.c,
651         testsuite/libffi.call/cls_longdouble.c,
652         testsuite/libffi.call/cls_longdouble_va.c,
653         testsuite/libffi.call/cls_pointer.c,
654         testsuite/libffi.call/cls_pointer_stack.c,
655         testsuite/libffi.call/err_bad_abi.c: Clean up failures from
656         compiler warnings.
657
658 2008-03-04  Anthony Green  <green@redhat.com>
659             Blake Chaffin
660             hos@tamanegi.org
661
662         * testsuite/libffi.call/cls_align_longdouble_split2.c
663           testsuite/libffi.call/cls_align_longdouble_split.c
664           testsuite/libffi.call/cls_dbls_struct.c
665           testsuite/libffi.call/cls_double_va.c
666           testsuite/libffi.call/cls_longdouble.c
667           testsuite/libffi.call/cls_longdouble_va.c
668           testsuite/libffi.call/cls_pointer.c
669           testsuite/libffi.call/cls_pointer_stack.c
670           testsuite/libffi.call/err_bad_abi.c
671           testsuite/libffi.call/err_bad_typedef.c
672           testsuite/libffi.call/stret_large2.c
673           testsuite/libffi.call/stret_large.c
674           testsuite/libffi.call/stret_medium2.c
675           testsuite/libffi.call/stret_medium.c: New tests from Apple.
676
677 2009-06-05  Andrew Haley  <aph@redhat.com>
678
679         * src/x86/ffitarget.h, src/x86/ffi.c: Merge stdcall changes from
680         libffi.
681
682 2009-06-04  Andrew Haley  <aph@redhat.com>
683
684         * src/x86/ffitarget.h, src/x86/win32.S, src/x86/ffi.c: Back out
685         stdcall changes.
686
687 2008-02-26  Anthony Green  <green@redhat.com>
688             Thomas Heller  <theller@ctypes.org>
689
690         * src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C
691         comment.
692
693 2008-02-03  Timothy Wall  <twall@users.sf.net>
694
695         * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return
696           offset based on code pointer, not data pointer.
697
698 2008-01-31  Timothy Wall <twall@users.sf.net>
699
700         * testsuite/libffi.call/closure_stdcall.c: Add test for stdcall
701         closures.
702         * src/x86/ffitarget.h: Increase size of trampoline for stdcall
703         closures.
704         * src/x86/win32.S: Add assembly for stdcall closure.
705         * src/x86/ffi.c: Initialize stdcall closure trampoline.
706
707 2009-06-04  Andrew Haley  <aph@redhat.com>
708
709         * include/ffi.h.in: Change void (*)() to void (*)(void).
710         * src/x86/ffi.c: Likewise.
711
712 2009-06-04  Andrew Haley  <aph@redhat.com>
713
714         * src/powerpc/ppc_closure.S: Insert licence header.
715         * src/powerpc/linux64_closure.S: Likewise.
716         * src/m68k/sysv.S: Likewise.
717
718         * src/sh64/ffi.c: Change void (*)() to void (*)(void).
719         * src/powerpc/ffi.c: Likewise.
720         * src/powerpc/ffi_darwin.c: Likewise.
721         * src/m32r/ffi.c: Likewise.
722         * src/sh64/ffi.c: Likewise.
723         * src/x86/ffi64.c: Likewise.
724         * src/alpha/ffi.c: Likewise.
725         * src/alpha/osf.S: Likewise.
726         * src/frv/ffi.c: Likewise.
727         * src/s390/ffi.c: Likewise.
728         * src/pa/ffi.c: Likewise.
729         * src/pa/hpux32.S: Likewise.
730         * src/ia64/unix.S: Likewise.
731         * src/ia64/ffi.c: Likewise.
732         * src/sparc/ffi.c: Likewise.
733         * src/mips/ffi.c: Likewise.
734         * src/sh/ffi.c: Likewise.
735
736 2008-02-15  David Daney  <ddaney@avtrex.com>
737
738         * src/mips/ffi.c (USE__BUILTIN___CLEAR_CACHE):
739         Define (conditionally), and use it to include cachectl.h.
740         (ffi_prep_closure_loc): Fix cache flushing.
741         * src/mips/ffitarget.h (_ABIN32, _ABI64, _ABIO32): Define.
742
743 2009-06-04  Andrew Haley  <aph@redhat.com>
744
745         include/ffi.h.in,
746         src/arm/ffitarget.h,
747         src/arm/ffi.c,
748         src/arm/sysv.S,
749         src/powerpc/ffitarget.h,
750         src/closures.c,
751         src/sh64/ffitarget.h,
752         src/sh64/ffi.c,
753         src/sh64/sysv.S,
754         src/types.c,
755         src/x86/ffi64.c,
756         src/x86/ffitarget.h,
757         src/x86/win32.S,
758         src/x86/darwin.S,
759         src/x86/ffi.c,
760         src/x86/sysv.S,
761         src/x86/unix64.S,
762         src/alpha/ffitarget.h,
763         src/alpha/ffi.c,
764         src/alpha/osf.S,
765         src/m68k/ffitarget.h,
766         src/frv/ffitarget.h,
767         src/frv/ffi.c,
768         src/s390/ffitarget.h,
769         src/s390/sysv.S,
770         src/cris/ffitarget.h,
771         src/pa/linux.S,
772         src/pa/ffitarget.h,
773         src/pa/ffi.c,
774         src/raw_api.c,
775         src/ia64/ffitarget.h,
776         src/ia64/unix.S,
777         src/ia64/ffi.c,
778         src/ia64/ia64_flags.h,
779         src/java_raw_api.c,
780         src/debug.c,
781         src/sparc/v9.S,
782         src/sparc/ffitarget.h,
783         src/sparc/ffi.c,
784         src/sparc/v8.S,
785         src/mips/ffitarget.h,
786         src/mips/n32.S,
787         src/mips/o32.S,
788         src/mips/ffi.c,
789         src/prep_cif.c,
790         src/sh/ffitarget.h,
791         src/sh/ffi.c,
792         src/sh/sysv.S: Update license text.
793
794 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
795
796         * src/x86/win32.S (_ffi_closure_STDCALL):  New function.
797         (.eh_frame):  Add FDE for it.
798
799 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
800
801         * configure.ac:  Also check if assembler supports pc-relative
802         relocs on X86_WIN32 targets.
803         * configure:  Regenerate.
804         * src/x86/win32.S (ffi_prep_args):  Declare extern, not global.
805         (_ffi_call_SYSV):  Add missing function type symbol .def and
806         add EH markup labels.
807         (_ffi_call_STDCALL):  Likewise.
808         (_ffi_closure_SYSV):  Likewise.
809         (_ffi_closure_raw_SYSV):  Likewise.
810         (.eh_frame):  Add hand-crafted EH data.
811
812 2009-04-09  Jakub Jelinek  <jakub@redhat.com>
813
814         * testsuite/lib/libffi-dg.exp: Change copyright header to refer to
815         version 3 of the GNU General Public License and to point readers
816         at the COPYING3 file and the FSF's license web page.
817         * testsuite/libffi.call/call.exp: Likewise.
818         * testsuite/libffi.special/special.exp: Likewise.
819
820 2009-03-01  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
821
822         * configure: Regenerate.
823
824 2008-12-18  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
825
826         PR libffi/26048
827         * configure.ac (HAVE_AS_X86_PCREL): New test.
828         * configure: Regenerate.
829         * fficonfig.h.in: Regenerate.
830         * src/x86/sysv.S [!FFI_NO_RAW_API]: Precalculate
831         RAW_CLOSURE_CIF_OFFSET, RAW_CLOSURE_FUN_OFFSET,
832         RAW_CLOSURE_USER_DATA_OFFSET for the Solaris 10/x86 assembler.
833         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
834         * src/x86/unix64.S (.Lstore_table): Move to .text section.
835         (.Lload_table): Likewise.
836         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
837
838 2008-12-18  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
839
840         * configure: Regenerate.
841
842 2008-11-21  Eric Botcazou  <ebotcazou@adacore.com>
843
844         * src/sparc/ffi.c (ffi_prep_cif_machdep): Add support for
845         signed/unsigned int8/16 return values.
846         * src/sparc/v8.S (ffi_call_v8): Likewise.
847         (ffi_closure_v8): Likewise.
848
849 2008-09-26  Peter O'Gorman  <pogma@thewrittenword.com>
850             Steve Ellcey  <sje@cup.hp.com>
851
852         * configure: Regenerate for new libtool.
853         * Makefile.in: Ditto.
854         * include/Makefile.in: Ditto.
855         * aclocal.m4: Ditto.
856
857 2008-08-25  Andreas Tobler  <a.tobler@schweiz.org>
858
859         * src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and
860         FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum.
861         Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT.
862         Adjust copyright notice.
863         * src/powerpc/ffi.c: Add two new flags to indicate if we have one
864         register or two register to use for FFI_SYSV structs.
865         (ffi_prep_cif_machdep): Pass the right register flag introduced above.
866         (ffi_closure_helper_SYSV): Fix the return type for
867         FFI_SYSV_TYPE_SMALL_STRUCT. Comment.
868         Adjust copyright notice.
869
870 2008-07-16  Kaz Kojima  <kkojima@gcc.gnu.org>
871
872         * src/sh/ffi.c (ffi_prep_closure_loc): Turn INSN into an unsigned
873         int.
874
875 2008-06-17  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
876
877         * configure: Regenerate.
878         * include/Makefile.in: Regenerate.
879         * testsuite/Makefile.in: Regenerate.
880
881 2008-06-07  Joseph Myers  <joseph@codesourcery.com>
882
883         * configure.ac (parisc*-*-linux*, powerpc-*-sysv*,
884         powerpc-*-beos*): Remove.
885         * configure: Regenerate.
886
887 2008-05-09  Julian Brown  <julian@codesourcery.com>
888
889         * Makefile.am (LTLDFLAGS): New.
890         (libffi_la_LDFLAGS): Use above.
891         * Makefile.in: Regenerate.
892
893 2008-04-18  Paolo Bonzini  <bonzini@gnu.org>
894
895         PR bootstrap/35457
896         * aclocal.m4: Regenerate.
897         * configure: Regenerate.
898
899 2008-03-26  Kaz Kojima  <kkojima@gcc.gnu.org>
900
901         * src/sh/sysv.S: Add .note.GNU-stack on Linux.
902         * src/sh64/sysv.S: Likewise.
903
904 2008-03-26  Daniel Jacobowitz  <dan@debian.org>
905
906         * src/arm/sysv.S: Fix ARM comment marker.
907
908 2008-03-26  Jakub Jelinek  <jakub@redhat.com>
909
910         * src/alpha/osf.S: Add .note.GNU-stack on Linux.
911         * src/s390/sysv.S: Likewise.
912         * src/powerpc/ppc_closure.S: Likewise.
913         * src/powerpc/sysv.S: Likewise.
914         * src/x86/unix64.S: Likewise.
915         * src/x86/sysv.S: Likewise.
916         * src/sparc/v8.S: Likewise.
917         * src/sparc/v9.S: Likewise.
918         * src/m68k/sysv.S: Likewise.
919         * src/arm/sysv.S: Likewise.
920
921 2008-03-16  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
922
923         * aclocal.m4: Regenerate.
924         * configure: Likewise.
925         * Makefile.in: Likewise.
926         * include/Makefile.in: Likewise.
927         * testsuite/Makefile.in: Likewise.
928
929 2008-02-12  Bjoern Koenig  <bkoenig@alpha-tierchen.de>
930             Andreas Tobler  <a.tobler@schweiz.org>
931
932         * configure.ac: Add amd64-*-freebsd* target.
933         * configure: Regenerate.
934
935 2008-01-30  H.J. Lu  <hongjiu.lu@intel.com>
936
937         PR libffi/34612
938         * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when
939         returning struct.
940
941         * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer"
942         tests.
943
944 2008-01-24  David Edelsohn  <edelsohn@gnu.org>
945
946         * configure: Regenerate.
947
948 2008-01-06  Andreas Tobler  <a.tobler@schweiz.org>
949
950         * src/x86/ffi.c (ffi_prep_cif_machdep): Fix thinko.
951
952 2008-01-05  Andreas Tobler  <a.tobler@schweiz.org>
953
954         PR testsuite/32843
955         * src/x86/ffi.c (ffi_prep_cif_machdep): Add code for
956         signed/unsigned int8/16 for X86_DARWIN.
957         Updated copyright info.
958         Handle one and two byte structs with special cif->flags.
959         * src/x86/ffitarget.h: Add special types for one and two byte structs.
960         Updated copyright info.
961         * src/x86/darwin.S (ffi_call_SYSV): Rewrite to use a jump table like
962         sysv.S
963         Remove code to pop args from the stack after call.
964         Special-case signed/unsigned for int8/16, one and two byte structs.
965         (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
966         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
967         FFI_TYPE_SINT32.
968         Updated copyright info.
969
970 2007-12-08  David Daney  <ddaney@avtrex.com>
971
972         * src/mips/n32.S (ffi_call_N32):  Replace dadd with ADDU, dsub with
973         SUBU, add with ADDU and use smaller code sequences.
974
975 2007-12-07  David Daney  <ddaney@avtrex.com>
976
977         * src/mips/ffi.c (ffi_prep_cif_machdep): Handle long double return
978         type.
979
980 2007-12-06  David Daney  <ddaney@avtrex.com>
981
982         * include/ffi.h.in (FFI_SIZEOF_JAVA_RAW): Define if not already
983         defined.
984         (ffi_java_raw): New typedef.
985         (ffi_java_raw_call, ffi_java_ptrarray_to_raw,
986         ffi_java_raw_to_ptrarray): Change parameter types from ffi_raw to
987         ffi_java_raw.
988         (ffi_java_raw_closure) : Same.
989         (ffi_prep_java_raw_closure, ffi_prep_java_raw_closure_loc): Change
990         parameter types.
991         * src/java_raw_api.c (ffi_java_raw_size):  Replace FFI_SIZEOF_ARG with
992         FFI_SIZEOF_JAVA_RAW.
993         (ffi_java_raw_to_ptrarray): Change type of raw to ffi_java_raw.
994         Replace FFI_SIZEOF_ARG with FFI_SIZEOF_JAVA_RAW. Use
995         sizeof(ffi_java_raw) for alignment calculations.
996         (ffi_java_ptrarray_to_raw): Same.
997         (ffi_java_rvalue_to_raw): Add special handling for FFI_TYPE_POINTER
998         if FFI_SIZEOF_JAVA_RAW == 4.
999         (ffi_java_raw_to_rvalue): Same.
1000         (ffi_java_raw_call): Change type of raw to ffi_java_raw.
1001         (ffi_java_translate_args): Same.
1002         (ffi_prep_java_raw_closure_loc, ffi_prep_java_raw_closure): Change
1003         parameter types.
1004         * src/mips/ffitarget.h (FFI_SIZEOF_JAVA_RAW): Define for N32 ABI.
1005
1006 2007-12-06  David Daney  <ddaney@avtrex.com>
1007
1008         * src/mips/n32.S (ffi_closure_N32): Use 64-bit add instruction on
1009         pointer values.
1010
1011 2007-12-01  Andreas Tobler  <a.tobler@schweiz.org>
1012
1013         PR libffi/31937
1014         * src/powerpc/ffitarget.h: Introduce new ABI FFI_LINUX_SOFT_FLOAT.
1015         Add local FFI_TYPE_UINT128 to handle soft-float long-double-128.
1016         * src/powerpc/ffi.c: Distinguish between __NO_FPRS__ and not and
1017         set the NUM_FPR_ARG_REGISTERS according to.
1018         Add support for potential soft-float support under hard-float
1019         architecture.
1020         (ffi_prep_args_SYSV): Set NUM_FPR_ARG_REGISTERS to 0 in case of
1021         FFI_LINUX_SOFT_FLOAT, handle float, doubles and long-doubles according
1022         to the FFI_LINUX_SOFT_FLOAT ABI.
1023         (ffi_prep_cif_machdep): Likewise.
1024         (ffi_closure_helper_SYSV): Likewise.
1025         * src/powerpc/ppc_closure.S: Make sure not to store float/double
1026         on archs where __NO_FPRS__ is true.
1027         Add FFI_TYPE_UINT128 support.
1028         * src/powerpc/sysv.S: Add support for soft-float long-double-128.
1029         Adjust copyright notice.
1030
1031 2007-11-25  Andreas Tobler  <a.tobler@schweiz.org>
1032
1033         * src/closures.c: Move defintion of MAYBE_UNUSED from here to ...
1034         * include/ffi_common.h: ... here.
1035         Update copyright.
1036
1037 2007-11-17  Andreas Tobler  <a.tobler@schweiz.org>
1038
1039         * src/powerpc/sysv.S: Load correct cr to compare if we have long double.
1040         * src/powerpc/linux64.S: Likewise.
1041         * src/powerpc/ffi.c: Add a comment to show which part goes into cr6.
1042         * testsuite/libffi.call/return_ldl.c: New test.
1043
1044 2007-09-04    <aph@redhat.com>
1045
1046         * src/arm/sysv.S (UNWIND): New.
1047         (Whole file): Conditionally compile unwinder directives.
1048         * src/arm/sysv.S: Add unwinder directives.
1049
1050         * src/arm/ffi.c (ffi_prep_args): Align structs by at least 4 bytes.
1051         Only treat r0 as a struct address if we're actually returning a
1052         struct by address.
1053         Only copy the bytes that are actually within a struct.
1054         (ffi_prep_cif_machdep): A Composite Type not larger than 4 bytes
1055         is returned in r0, not passed by address.
1056         (ffi_call): Allocate a word-sized temporary for the case where
1057         a composite is returned in r0.
1058         (ffi_prep_incoming_args_SYSV): Align as necessary.
1059
1060 2007-08-05  Steven Newbury  <s_j_newbury@yahoo.co.uk>
1061
1062         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Use __clear_cache instead of
1063         directly using the sys_cacheflush syscall.
1064
1065 2007-07-27  Andrew Haley  <aph@redhat.com>
1066
1067         * src/arm/sysv.S (ffi_closure_SYSV): Add soft-float.
1068
1069 2007-09-03  Maciej W. Rozycki  <macro@linux-mips.org>
1070
1071         * Makefile.am: Unify MIPS_IRIX and MIPS_LINUX into MIPS.
1072         * configure.ac: Likewise.
1073         * Makefile.in: Regenerate.
1074         * include/Makefile.in: Likewise.
1075         * testsuite/Makefile.in: Likewise.
1076         * configure: Likewise.
1077
1078 2007-08-24  David Daney  <ddaney@avtrex.com>
1079
1080         * testsuite/libffi.call/return_sl.c: New test.
1081
1082 2007-08-10  David Daney  <ddaney@avtrex.com>
1083
1084         * testsuite/libffi.call/cls_multi_ushort.c,
1085         testsuite/libffi.call/cls_align_uint16.c,
1086         testsuite/libffi.call/nested_struct1.c,
1087         testsuite/libffi.call/nested_struct3.c,
1088         testsuite/libffi.call/cls_7_1_byte.c,
1089         testsuite/libffi.call/nested_struct5.c,
1090         testsuite/libffi.call/cls_double.c,
1091         testsuite/libffi.call/nested_struct7.c,
1092         testsuite/libffi.call/cls_sint.c,
1093         testsuite/libffi.call/nested_struct9.c,
1094         testsuite/libffi.call/cls_20byte1.c,
1095         testsuite/libffi.call/cls_multi_sshortchar.c,
1096         testsuite/libffi.call/cls_align_sint64.c,
1097         testsuite/libffi.call/cls_3byte2.c,
1098         testsuite/libffi.call/cls_multi_schar.c,
1099         testsuite/libffi.call/cls_multi_uchar.c,
1100         testsuite/libffi.call/cls_19byte.c,
1101         testsuite/libffi.call/cls_9byte1.c,
1102         testsuite/libffi.call/cls_align_float.c,
1103         testsuite/libffi.call/closure_fn1.c,
1104         testsuite/libffi.call/problem1.c,
1105         testsuite/libffi.call/closure_fn3.c,
1106         testsuite/libffi.call/cls_sshort.c,
1107         testsuite/libffi.call/closure_fn5.c,
1108         testsuite/libffi.call/cls_align_double.c,
1109         testsuite/libffi.call/nested_struct.c,
1110         testsuite/libffi.call/cls_2byte.c,
1111         testsuite/libffi.call/nested_struct10.c,
1112         testsuite/libffi.call/cls_4byte.c,
1113         testsuite/libffi.call/cls_6byte.c,
1114         testsuite/libffi.call/cls_8byte.c,
1115         testsuite/libffi.call/cls_multi_sshort.c,
1116         testsuite/libffi.call/cls_align_sint16.c,
1117         testsuite/libffi.call/cls_align_uint32.c,
1118         testsuite/libffi.call/cls_20byte.c,
1119         testsuite/libffi.call/cls_float.c,
1120         testsuite/libffi.call/nested_struct2.c,
1121         testsuite/libffi.call/cls_5_1_byte.c,
1122         testsuite/libffi.call/nested_struct4.c,
1123         testsuite/libffi.call/cls_24byte.c,
1124         testsuite/libffi.call/nested_struct6.c,
1125         testsuite/libffi.call/cls_64byte.c,
1126         testsuite/libffi.call/nested_struct8.c,
1127         testsuite/libffi.call/cls_uint.c,
1128         testsuite/libffi.call/cls_multi_ushortchar.c,
1129         testsuite/libffi.call/cls_schar.c,
1130         testsuite/libffi.call/cls_uchar.c,
1131         testsuite/libffi.call/cls_align_uint64.c,
1132         testsuite/libffi.call/cls_ulonglong.c,
1133         testsuite/libffi.call/cls_align_longdouble.c,
1134         testsuite/libffi.call/cls_1_1byte.c,
1135         testsuite/libffi.call/cls_12byte.c,
1136         testsuite/libffi.call/cls_3_1byte.c,
1137         testsuite/libffi.call/cls_3byte1.c,
1138         testsuite/libffi.call/cls_4_1byte.c,
1139         testsuite/libffi.call/cls_6_1_byte.c,
1140         testsuite/libffi.call/cls_16byte.c,
1141         testsuite/libffi.call/cls_18byte.c,
1142         testsuite/libffi.call/closure_fn0.c,
1143         testsuite/libffi.call/cls_9byte2.c,
1144         testsuite/libffi.call/closure_fn2.c,
1145         testsuite/libffi.call/closure_fn4.c,
1146         testsuite/libffi.call/cls_ushort.c,
1147         testsuite/libffi.call/closure_fn6.c,
1148         testsuite/libffi.call/cls_5byte.c,
1149         testsuite/libffi.call/cls_align_pointer.c,
1150         testsuite/libffi.call/cls_7byte.c,
1151         testsuite/libffi.call/cls_align_sint32.c,
1152         testsuite/libffi.special/unwindtest_ffi_call.cc,
1153         testsuite/libffi.special/unwindtest.cc: Remove xfail for mips64*-*-*.
1154
1155 2007-08-10  David Daney  <ddaney@avtrex.com>
1156
1157         PR libffi/28313
1158         * configure.ac: Don't treat mips64 as a special case.
1159         * Makefile.am (nodist_libffi_la_SOURCES): Add n32.S.
1160         * configure: Regenerate
1161         * Makefile.in: Ditto.
1162         * fficonfig.h.in: Ditto.
1163         * src/mips/ffitarget.h (REG_L, REG_S, SUBU, ADDU, SRL, LI): Indent.
1164         (LA, EH_FRAME_ALIGN, FDE_ADDR_BYTES): New preprocessor macros.
1165         (FFI_DEFAULT_ABI): Set for n64 case.
1166         (FFI_CLOSURES, FFI_TRAMPOLINE_SIZE): Define for n32 and n64 cases.
1167         * src/mips/n32.S (ffi_call_N32): Add debug macros and labels for FDE.
1168         (ffi_closure_N32): New function.
1169         (.eh_frame): New section
1170         * src/mips/o32.S: Clean up comments.
1171         (ffi_closure_O32): Pass ffi_closure parameter in $12.
1172         * src/mips/ffi.c: Use FFI_MIPS_N32 instead of
1173         _MIPS_SIM == _ABIN32 throughout.
1174         (FFI_MIPS_STOP_HERE): New, use in place of
1175         ffi_stop_here.
1176         (ffi_prep_args): Use unsigned long to hold pointer values.  Rewrite
1177         to support n32/n64 ABIs.
1178         (calc_n32_struct_flags): Rewrite.
1179         (calc_n32_return_struct_flags): Remove unused variable.  Reverse
1180         position of flag bits.
1181         (ffi_prep_cif_machdep): Rewrite n32 portion.
1182         (ffi_call): Enable for n64.  Add special handling for small structure
1183         return values.
1184         (ffi_prep_closure_loc): Add n32 and n64 support.
1185         (ffi_closure_mips_inner_O32): Add cast to silence warning.
1186         (copy_struct_N32, ffi_closure_mips_inner_N32): New functions.
1187
1188 2007-08-08  David Daney  <ddaney@avtrex.com>
1189
1190         * testsuite/libffi.call/ffitest.h (ffi_type_mylong): Remove definition.
1191         * testsuite/libffi.call/cls_align_uint16.c (main): Use correct type
1192         specifiers.
1193         * testsuite/libffi.call/nested_struct1.c (main): Ditto.
1194         * testsuite/libffi.call/cls_sint.c (main): Ditto.
1195         * testsuite/libffi.call/nested_struct9.c (main): Ditto.
1196         * testsuite/libffi.call/cls_20byte1.c (main): Ditto.
1197         * testsuite/libffi.call/cls_9byte1.c (main): Ditto.
1198         * testsuite/libffi.call/closure_fn1.c (main): Ditto.
1199         * testsuite/libffi.call/closure_fn3.c (main): Ditto.
1200         * testsuite/libffi.call/return_dbl2.c (main): Ditto.
1201         * testsuite/libffi.call/cls_sshort.c (main): Ditto.
1202         * testsuite/libffi.call/return_fl3.c (main): Ditto.
1203         * testsuite/libffi.call/closure_fn5.c (main): Ditto.
1204         * testsuite/libffi.call/nested_struct.c (main): Ditto.
1205         * testsuite/libffi.call/nested_struct10.c (main): Ditto.
1206         * testsuite/libffi.call/return_ll1.c (main): Ditto.
1207         * testsuite/libffi.call/cls_8byte.c (main): Ditto.
1208         * testsuite/libffi.call/cls_align_uint32.c (main): Ditto.
1209         * testsuite/libffi.call/cls_align_sint16.c (main): Ditto.
1210         * testsuite/libffi.call/cls_20byte.c (main): Ditto.
1211         * testsuite/libffi.call/nested_struct2.c (main): Ditto.
1212         * testsuite/libffi.call/cls_24byte.c (main): Ditto.
1213         * testsuite/libffi.call/nested_struct6.c (main): Ditto.
1214         * testsuite/libffi.call/cls_uint.c (main): Ditto.
1215         * testsuite/libffi.call/cls_12byte.c (main): Ditto.
1216         * testsuite/libffi.call/cls_16byte.c (main): Ditto.
1217         * testsuite/libffi.call/closure_fn0.c (main): Ditto.
1218         * testsuite/libffi.call/cls_9byte2.c (main): Ditto.
1219         * testsuite/libffi.call/closure_fn2.c (main): Ditto.
1220         * testsuite/libffi.call/return_dbl1.c (main): Ditto.
1221         * testsuite/libffi.call/closure_fn4.c (main): Ditto.
1222         * testsuite/libffi.call/closure_fn6.c (main): Ditto.
1223         * testsuite/libffi.call/cls_align_sint32.c (main): Ditto.
1224
1225 2007-08-07  Andrew Haley  <aph@redhat.com>
1226
1227         * src/x86/sysv.S (ffi_closure_raw_SYSV): Fix typo in previous
1228         checkin.
1229
1230 2007-08-06  Andrew Haley  <aph@redhat.com>
1231
1232         PR testsuite/32843
1233         * src/x86/sysv.S (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1234         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1235         FFI_TYPE_SINT32.
1236
1237 2007-08-02  David Daney  <ddaney@avtrex.com>
1238
1239         * testsuite/libffi.call/return_ul.c (main): Define return type as
1240         ffi_arg.  Use proper printf conversion specifier.
1241
1242 2007-07-30  Andrew Haley  <aph@redhat.com>
1243
1244         PR testsuite/32843
1245         * src/x86/ffi.c (ffi_prep_cif_machdep): in x86 case, add code for
1246         signed/unsigned int8/16.
1247         * src/x86/sysv.S (ffi_call_SYSV): Rewrite to:
1248         Use a jump table.
1249         Remove code to pop args from the stack after call.
1250         Special-case signed/unsigned int8/16.
1251         * testsuite/libffi.call/return_sc.c (main): Revert.
1252
1253 2007-07-26  Richard Guenther  <rguenther@suse.de>
1254
1255         PR testsuite/32843
1256         * testsuite/libffi.call/return_sc.c (main): Verify call
1257         result as signed char, not ffi_arg.
1258
1259 2007-07-16  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1260
1261         * configure.ac (i?86-*-solaris2.1[0-9]): Set TARGET to X86_64.
1262         * configure: Regenerate.
1263
1264 2007-07-11  David Daney  <ddaney@avtrex.com>
1265
1266         * src/mips/ffi.c: Don't include sys/cachectl.h.
1267         (ffi_prep_closure_loc): Use __builtin___clear_cache() instead of
1268         cacheflush().
1269
1270 2007-05-18  Aurelien Jarno  <aurelien@aurel32.net>
1271
1272         * src/arm/ffi.c (ffi_prep_closure_loc): Renamed and ajusted
1273         from (ffi_prep_closure): ... this.
1274         (FFI_INIT_TRAMPOLINE): Adjust.
1275
1276 2005-12-31  Phil Blundell  <pb@reciva.com>
1277
1278         * src/arm/ffi.c (ffi_prep_incoming_args_SYSV,
1279         ffi_closure_SYSV_inner, ffi_prep_closure): New, add closure support.
1280         * src/arm/sysv.S(ffi_closure_SYSV): Likewise.
1281         * src/arm/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1282         (FFI_CLOSURES): Enable closure support.
1283
1284 2007-07-03  Andrew Haley  <aph@hedges.billgatliff.com>
1285
1286         * testsuite/libffi.call/cls_multi_ushort.c,
1287         testsuite/libffi.call/cls_align_uint16.c,
1288         testsuite/libffi.call/nested_struct1.c,
1289         testsuite/libffi.call/nested_struct3.c,
1290         testsuite/libffi.call/cls_7_1_byte.c,
1291         testsuite/libffi.call/cls_double.c,
1292         testsuite/libffi.call/nested_struct5.c,
1293         testsuite/libffi.call/nested_struct7.c,
1294         testsuite/libffi.call/cls_sint.c,
1295         testsuite/libffi.call/nested_struct9.c,
1296         testsuite/libffi.call/cls_20byte1.c,
1297         testsuite/libffi.call/cls_multi_sshortchar.c,
1298         testsuite/libffi.call/cls_align_sint64.c,
1299         testsuite/libffi.call/cls_3byte2.c,
1300         testsuite/libffi.call/cls_multi_schar.c,
1301         testsuite/libffi.call/cls_multi_uchar.c,
1302         testsuite/libffi.call/cls_19byte.c,
1303         testsuite/libffi.call/cls_9byte1.c,
1304         testsuite/libffi.call/cls_align_float.c,
1305         testsuite/libffi.call/closure_fn1.c,
1306         testsuite/libffi.call/problem1.c,
1307         testsuite/libffi.call/closure_fn3.c,
1308         testsuite/libffi.call/cls_sshort.c,
1309         testsuite/libffi.call/closure_fn5.c,
1310         testsuite/libffi.call/cls_align_double.c,
1311         testsuite/libffi.call/cls_2byte.c,
1312         testsuite/libffi.call/nested_struct.c,
1313         testsuite/libffi.call/nested_struct10.c,
1314         testsuite/libffi.call/cls_4byte.c,
1315         testsuite/libffi.call/cls_6byte.c,
1316         testsuite/libffi.call/cls_8byte.c,
1317         testsuite/libffi.call/cls_multi_sshort.c,
1318         testsuite/libffi.call/cls_align_uint32.c,
1319         testsuite/libffi.call/cls_align_sint16.c,
1320         testsuite/libffi.call/cls_float.c,
1321         testsuite/libffi.call/cls_20byte.c,
1322         testsuite/libffi.call/cls_5_1_byte.c,
1323         testsuite/libffi.call/nested_struct2.c,
1324         testsuite/libffi.call/cls_24byte.c,
1325         testsuite/libffi.call/nested_struct4.c,
1326         testsuite/libffi.call/nested_struct6.c,
1327         testsuite/libffi.call/cls_64byte.c,
1328         testsuite/libffi.call/nested_struct8.c,
1329         testsuite/libffi.call/cls_uint.c,
1330         testsuite/libffi.call/cls_multi_ushortchar.c,
1331         testsuite/libffi.call/cls_schar.c,
1332         testsuite/libffi.call/cls_uchar.c,
1333         testsuite/libffi.call/cls_align_uint64.c,
1334         testsuite/libffi.call/cls_ulonglong.c,
1335         testsuite/libffi.call/cls_align_longdouble.c,
1336         testsuite/libffi.call/cls_1_1byte.c,
1337         testsuite/libffi.call/cls_12byte.c,
1338         testsuite/libffi.call/cls_3_1byte.c,
1339         testsuite/libffi.call/cls_3byte1.c,
1340         testsuite/libffi.call/cls_4_1byte.c,
1341         testsuite/libffi.call/cls_6_1_byte.c,
1342         testsuite/libffi.call/cls_16byte.c,
1343         testsuite/libffi.call/cls_18byte.c,
1344         testsuite/libffi.call/closure_fn0.c,
1345         testsuite/libffi.call/cls_9byte2.c,
1346         testsuite/libffi.call/closure_fn2.c,
1347         testsuite/libffi.call/closure_fn4.c,
1348         testsuite/libffi.call/cls_ushort.c,
1349         testsuite/libffi.call/closure_fn6.c,
1350         testsuite/libffi.call/cls_5byte.c,
1351         testsuite/libffi.call/cls_align_pointer.c,
1352         testsuite/libffi.call/cls_7byte.c,
1353         testsuite/libffi.call/cls_align_sint32.c,
1354         testsuite/libffi.special/unwindtest_ffi_call.cc,
1355         testsuite/libffi.special/unwindtest.cc: Enable for ARM.
1356
1357 2007-07-05  H.J. Lu  <hongjiu.lu@intel.com>
1358
1359         * aclocal.m4: Regenerated.
1360
1361 2007-06-02  Paolo Bonzini  <bonzini@gnu.org>
1362
1363         * configure: Regenerate.
1364
1365 2007-05-23  Steve Ellcey  <sje@cup.hp.com>
1366
1367         * Makefile.in: Regenerate.
1368         * configure: Regenerate.
1369         * aclocal.m4: Regenerate.
1370         * include/Makefile.in: Regenerate.
1371         * testsuite/Makefile.in: Regenerate.
1372
1373 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1374
1375         * src/m68k/ffi.c (ffi_prep_incoming_args_SYSV,
1376         ffi_closure_SYSV_inner,ffi_prep_closure): New, add closure support.
1377         * src/m68k/sysv.S(ffi_closure_SYSV,ffi_closure_struct_SYSV): Likewise.
1378         * src/m68k/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1379         (FFI_CLOSURES): Enable closure support.
1380
1381 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1382
1383         * configure.ac (HAVE_AS_CFI_PSEUDO_OP): New test.
1384         * configure: Regenerate.
1385         * fficonfig.h.in: Regenerate.
1386         * src/m68k/sysv.S (CFI_STARTPROC,CFI_ENDPROC,
1387         CFI_OFFSET,CFI_DEF_CFA): New macros.
1388         (ffi_call_SYSV): Add callframe annotation.
1389
1390 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1391
1392         * src/m68k/ffi.c (ffi_prep_args,ffi_prep_cif_machdep): Fix
1393         numerous test suite failures.
1394         * src/m68k/sysv.S (ffi_call_SYSV): Likewise.
1395
1396 2007-04-11  Paolo Bonzini  <bonzini@gnu.org>
1397
1398         * Makefile.am (EXTRA_DIST): Bring up to date.
1399         * Makefile.in: Regenerate.
1400         * src/frv/eabi.S: Remove RCS keyword.
1401
1402 2007-04-06  Richard Henderson  <rth@redhat.com>
1403
1404         * configure.ac: Tidy target case.
1405         (HAVE_LONG_DOUBLE): Allow the target to override.
1406         * configure: Regenerate.
1407         * include/ffi.h.in: Don't define ffi_type_foo if
1408         LIBFFI_HIDE_BASIC_TYPES is defined.
1409         (ffi_type_longdouble): If not HAVE_LONG_DOUBLE, define
1410         to ffi_type_double.
1411         * types.c (LIBFFI_HIDE_BASIC_TYPES): Define.
1412         (FFI_TYPEDEF, ffi_type_void): Mark the data const.
1413         (ffi_type_longdouble): Special case for Alpha.  Don't define
1414         if long double == double.
1415
1416         * src/alpha/ffi.c (FFI_TYPE_LONGDOUBLE): Assert unique value.
1417         (ffi_prep_cif_machdep): Handle it as the 128-bit type.
1418         (ffi_call, ffi_closure_osf_inner): Likewise.
1419         (ffi_closure_osf_inner): Likewise.  Mark hidden.
1420         (ffi_call_osf, ffi_closure_osf): Mark hidden.
1421         * src/alpha/ffitarget.h (FFI_LAST_ABI): Tidy definition.
1422         * src/alpha/osf.S (ffi_call_osf, ffi_closure_osf): Mark hidden.
1423         (load_table): Handle 128-bit long double.
1424
1425         * testsuite/libffi.call/float4.c: Add -mieee for alpha.
1426
1427 2007-04-06  Tom Tromey  <tromey@redhat.com>
1428
1429         PR libffi/31491:
1430         * README: Fixed bug in example.
1431
1432 2007-04-03  Jakub Jelinek  <jakub@redhat.com>
1433
1434         * src/closures.c: Include sys/statfs.h.
1435         (_GNU_SOURCE): Define on Linux.
1436         (FFI_MMAP_EXEC_SELINUX): Define.
1437         (selinux_enabled): New variable.
1438         (selinux_enabled_check): New function.
1439         (is_selinux_enabled): Define.
1440         (dlmmap): Use it.
1441
1442 2007-03-24  Uros Bizjak  <ubizjak@gmail.com>
1443
1444         * testsuite/libffi.call/return_fl2.c (return_fl): Mark as static.
1445         Use 'volatile float sum' to create sum of floats to avoid false
1446         negative due to excess precision on ix86 targets.
1447         (main): Ditto.
1448
1449 2007-03-08  Alexandre Oliva  <aoliva@redhat.com>
1450
1451         * src/powerpc/ffi.c (flush_icache): Fix left-over from previous
1452         patch.
1453         (ffi_prep_closure_loc): Remove unneeded casts.  Add needed ones.
1454
1455 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1456
1457         * include/ffi.h.in (ffi_closure_alloc, ffi_closure_free): New.
1458         (ffi_prep_closure_loc): New.
1459         (ffi_prep_raw_closure_loc): New.
1460         (ffi_prep_java_raw_closure_loc): New.
1461         * src/closures.c: New file.
1462         * src/dlmalloc.c [FFI_MMAP_EXEC_WRIT] (struct malloc_segment):
1463         Replace sflags with exec_offset.
1464         [FFI_MMAP_EXEC_WRIT] (mmap_exec_offset, add_segment_exec_offset,
1465         sub_segment_exec_offset): New macros.
1466         (get_segment_flags, set_segment_flags, check_segment_merge): New
1467         macros.
1468         (is_mmapped_segment, is_extern_segment): Use get_segment_flags.
1469         (add_segment, sys_alloc, create_mspace, create_mspace_with_base,
1470         destroy_mspace): Use new macros.
1471         (sys_alloc): Silence warning.
1472         * Makefile.am (libffi_la_SOURCES): Add src/closures.c.
1473         * Makefile.in: Rebuilt.
1474         * src/prep_cif [FFI_CLOSURES] (ffi_prep_closure): Implement in
1475         terms of ffi_prep_closure_loc.
1476         * src/raw_api.c (ffi_prep_raw_closure_loc): Renamed and adjusted
1477         from...
1478         (ffi_prep_raw_closure): ... this.  Re-implement in terms of the
1479         renamed version.
1480         * src/java_raw_api (ffi_prep_java_raw_closure_loc): Renamed and
1481         adjusted from...
1482         (ffi_prep_java_raw_closure): ... this.  Re-implement in terms of
1483         the renamed version.
1484         * src/alpha/ffi.c (ffi_prep_closure_loc): Renamed from
1485         (ffi_prep_closure): ... this.
1486         * src/pa/ffi.c: Likewise.
1487         * src/cris/ffi.c: Likewise.  Adjust.
1488         * src/frv/ffi.c: Likewise.
1489         * src/ia64/ffi.c: Likewise.
1490         * src/mips/ffi.c: Likewise.
1491         * src/powerpc/ffi_darwin.c: Likewise.
1492         * src/s390/ffi.c: Likewise.
1493         * src/sh/ffi.c: Likewise.
1494         * src/sh64/ffi.c: Likewise.
1495         * src/sparc/ffi.c: Likewise.
1496         * src/x86/ffi64.c: Likewise.
1497         * src/x86/ffi.c: Likewise.
1498         (FFI_INIT_TRAMPOLINE): Adjust.
1499         (ffi_prep_raw_closure_loc): Renamed and adjusted from...
1500         (ffi_prep_raw_closure): ... this.
1501         * src/powerpc/ffi.c (ffi_prep_closure_loc): Renamed from
1502         (ffi_prep_closure): ... this.
1503         (flush_icache): Adjust.
1504
1505 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1506
1507         * src/dlmalloc.c: New file, imported version 2.8.3 of Doug
1508         Lea's malloc.
1509
1510 2007-03-01  Brooks Moses  <brooks.moses@codesourcery.com>
1511
1512         * Makefile.am: Add dummy install-pdf target.
1513         * Makefile.in: Regenerate
1514
1515 2007-02-13  Andreas Krebbel  <krebbel1@de.ibm.com>
1516
1517         * src/s390/ffi.c (ffi_prep_args, ffi_prep_cif_machdep,
1518         ffi_closure_helper_SYSV): Add long double handling.
1519
1520 2007-02-02  Jakub Jelinek  <jakub@redhat.com>
1521
1522         * src/powerpc/linux64.S (ffi_call_LINUX64): Move restore of r2
1523         immediately after bctrl instruction.
1524
1525 2007-01-18  Alexandre Oliva  <aoliva@redhat.com>
1526
1527         * Makefile.am (all-recursive, install-recursive,
1528         mostlyclean-recursive, clean-recursive, distclean-recursive,
1529         maintainer-clean-recursive): Add missing targets.
1530         * Makefile.in: Rebuilt.
1531
1532 2006-12-14  Andreas Tobler  <a.tobler@schweiz.org>
1533
1534         * configure.ac: Add TARGET for x86_64-*-darwin*.
1535         * Makefile.am (nodist_libffi_la_SOURCES): Add rules for 64-bit sources
1536         for X86_DARWIN.
1537         * src/x86/ffitarget.h: Set trampoline size for x86_64-*-darwin*.
1538         * src/x86/darwin64.S: New file for x86_64-*-darwin* support.
1539         * configure: Regenerate.
1540         * Makefile.in: Regenerate.
1541         * include/Makefile.in: Regenerate.
1542         * testsuite/Makefile.in: Regenerate.
1543         * testsuite/libffi.special/unwindtest_ffi_call.cc: New test case for
1544         ffi_call only.
1545
1546 2006-12-13  Andreas Tobler <a.tobler@schweiz.org>
1547
1548         * aclocal.m4: Regenerate with aclocal -I .. as written in the
1549         Makefile.am.
1550
1551 2006-10-31  Geoffrey Keating  <geoffk@apple.com>
1552
1553         * src/powerpc/ffi_darwin.c (darwin_adjust_aggregate_sizes): New.
1554         (ffi_prep_cif_machdep): Call darwin_adjust_aggregate_sizes for
1555         Darwin.
1556         * testsuite/libffi.call/nested_struct4.c: Remove Darwin XFAIL.
1557         * testsuite/libffi.call/nested_struct6.c: Remove Darwin XFAIL.
1558
1559 2006-10-10  Paolo Bonzini  <bonzini@gnu.org>
1560             Sandro Tolaini  <tolaini@libero.it>
1561
1562         * configure.ac [i*86-*-darwin*]: Set X86_DARWIN symbol and
1563         conditional.
1564         * configure: Regenerated.
1565         * Makefile.am (nodist_libffi_la_SOURCES) [X86_DARWIN]: New case.
1566         (EXTRA_DIST): Add src/x86/darwin.S.
1567         * Makefile.in: Regenerated.
1568         * include/Makefile.in: Regenerated.
1569         * testsuite/Makefile.in: Regenerated.
1570
1571         * src/x86/ffi.c (ffi_prep_cif_machdep) [X86_DARWIN]: Treat like
1572         X86_WIN32, and additionally align stack to 16 bytes.
1573         * src/x86/darwin.S: New, based on sysv.S.
1574         * src/prep_cif.c (ffi_prep_cif) [X86_DARWIN]: Align > 8-byte structs.
1575
1576 2006-09-12  David Daney  <ddaney@avtrex.com>
1577
1578         PR libffi/23935
1579         * include/Makefile.am: Install both ffi.h and ffitarget.h in
1580         $(libdir)/gcc/$(target_alias)/$(gcc_version)/include.
1581         * aclocal.m4: Regenerated for automake 1.9.6.
1582         * Makefile.in: Regenerated.
1583         * include/Makefile.in: Regenerated.
1584         * testsuite/Makefile.in: Regenerated.
1585
1586 2006-08-17  Andreas Tobler  <a.tobler@schweiz.ch>
1587
1588         * include/ffi_common.h (struct): Revert accidental commit.
1589
1590 2006-08-15  Andreas Tobler  <a.tobler@schweiz.ch>
1591
1592         * include/ffi_common.h: Remove lint directives.
1593         * include/ffi.h.in: Likewise.
1594
1595 2006-07-25  Torsten Schoenfeld  <kaffeetisch@gmx.de>
1596
1597         * include/ffi.h.in (ffi_type_ulong, ffi_type_slong): Define correctly
1598         for 32-bit architectures.
1599         * testsuite/libffi.call/return_ul.c: New test case.
1600
1601 2006-07-19  David Daney  <ddaney@avtrex.com>
1602
1603         * testsuite/libffi.call/closure_fn6.c: Remove xfail for mips,
1604         xfail remains for mips64.
1605
1606 2006-05-23  Carlos O'Donell  <carlos@codesourcery.com>
1607
1608         * Makefile.am: Add install-html target. Add install-html to .PHONY
1609         * Makefile.in: Regenerate.
1610         * aclocal.m4: Regenerate.
1611         * include/Makefile.in: Regenerate.
1612         * testsuite/Makefile.in: Regenerate.
1613
1614 2006-05-18  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1615
1616         * pa/ffi.c (ffi_prep_args_pa32): Load floating point arguments from
1617         stack slot.
1618
1619 2006-04-22  Andreas Tobler  <a.tobler@schweiz.ch>
1620
1621         * README: Remove notice about 'Crazy Comments'.
1622         * src/debug.c: Remove lint directives. Cleanup white spaces.
1623         * src/java_raw_api.c: Likewise.
1624         * src/prep_cif.c: Likewise.
1625         * src/raw_api.c: Likewise.
1626         * src/ffitest.c: Delete. No longer needed, all test cases migrated
1627         to the testsuite.
1628         * src/arm/ffi.c: Remove lint directives.
1629         * src/m32r/ffi.c: Likewise.
1630         * src/pa/ffi.c: Likewise.
1631         * src/powerpc/ffi.c: Likewise.
1632         * src/powerpc/ffi_darwin.c: Likewise.
1633         * src/sh/ffi.c: Likewise.
1634         * src/sh64/ffi.c: Likewise.
1635         * src/x86/ffi.c: Likewise.
1636         * testsuite/libffi.call/float2.c: Likewise.
1637         * testsuite/libffi.call/promotion.c: Likewise.
1638         * testsuite/libffi.call/struct1.c: Likewise.
1639
1640 2006-04-13  Andreas Tobler  <a.tobler@schweiz.ch>
1641
1642         * src/pa/hpux32.S: Correct unwind offset calculation for
1643         ffi_closure_pa32.
1644         * src/pa/linux.S: Likewise.
1645
1646 2006-04-12  James E Wilson  <wilson@specifix.com>
1647
1648         PR libgcj/26483
1649         * src/ia64/ffi.c (stf_spill, ldf_fill): Rewrite as macros.
1650         (hfa_type_load): Call stf_spill.
1651         (hfa_type_store): Call ldf_fill.
1652         (ffi_call): Adjust calls to above routines.  Add local temps for
1653         macro result.
1654
1655 2006-04-10  Matthias Klose  <doko@debian.org>
1656
1657         * testsuite/lib/libffi-dg.exp (libffi-init): Recognize multilib
1658         directory names containing underscores.
1659
1660 2006-04-07  James E Wilson  <wilson@specifix.com>
1661
1662         * testsuite/libffi.call/float4.c: New testcase.
1663
1664 2006-04-05  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1665             Andreas Tobler  <a.tobler@schweiz.ch>
1666
1667         * Makefile.am: Add PA_HPUX port.
1668         * Makefile.in: Regenerate.
1669         * include/Makefile.in: Likewise.
1670         * testsuite/Makefile.in: Likewise.
1671         * configure.ac: Add PA_HPUX rules.
1672         * configure: Regenerate.
1673         * src/pa/ffitarget.h: Rename linux target to PA_LINUX.
1674         Add PA_HPUX and PA64_HPUX.
1675         Rename FFI_LINUX ABI to FFI_PA32 ABI.
1676         (FFI_TRAMPOLINE_SIZE): Define for 32-bit HP-UX targets.
1677         (FFI_TYPE_SMALL_STRUCT2): Define.
1678         (FFI_TYPE_SMALL_STRUCT4): Likewise.
1679         (FFI_TYPE_SMALL_STRUCT8): Likewise.
1680         (FFI_TYPE_SMALL_STRUCT3): Redefine.
1681         (FFI_TYPE_SMALL_STRUCT5): Likewise.
1682         (FFI_TYPE_SMALL_STRUCT6): Likewise.
1683         (FFI_TYPE_SMALL_STRUCT7): Likewise.
1684         * src/pa/ffi.c (ROUND_DOWN): Delete.
1685         (fldw, fstw, fldd, fstd): Use '__asm__'.
1686         (ffi_struct_type): Add support for FFI_TYPE_SMALL_STRUCT2,
1687         FFI_TYPE_SMALL_STRUCT4 and FFI_TYPE_SMALL_STRUCT8.
1688         (ffi_prep_args_LINUX): Rename to ffi_prep_args_pa32. Update comment.
1689         Simplify incrementing of stack slot variable. Change type of local
1690         'n' to unsigned int.
1691         (ffi_size_stack_LINUX): Rename to ffi_size_stack_pa32. Handle long
1692         double on PA_HPUX.
1693         (ffi_prep_cif_machdep): Likewise.
1694         (ffi_call): Likewise.
1695         (ffi_closure_inner_LINUX): Rename to ffi_closure_inner_pa32. Change
1696         return type to ffi_status. Simplify incrementing of stack slot
1697         variable. Only copy floating point argument registers when PA_LINUX
1698         is true. Reformat debug statement.
1699         Add support for FFI_TYPE_SMALL_STRUCT2, FFI_TYPE_SMALL_STRUCT4 and
1700         FFI_TYPE_SMALL_STRUCT8.
1701         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Add 'extern' to
1702         declaration.
1703         (ffi_prep_closure): Make linux trampoline conditional on PA_LINUX.
1704         Add nops to cache flush.  Add trampoline for PA_HPUX.
1705         * src/pa/hpux32.S: New file.
1706         * src/pa/linux.S (ffi_call_LINUX): Rename to ffi_call_pa32. Rename
1707         ffi_prep_args_LINUX to ffi_prep_args_pa32.
1708         Localize labels. Add support for 2, 4 and 8-byte small structs. Handle
1709         unaligned destinations in 3, 5, 6 and 7-byte small structs. Order
1710         argument type checks so that common argument types appear first.
1711         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Rename
1712         ffi_closure_inner_LINUX to ffi_closure_inner_pa32.
1713
1714 2006-03-24  Alan Modra  <amodra@bigpond.net.au>
1715
1716         * src/powerpc/ffitarget.h (enum ffi_abi): Add FFI_LINUX.  Default
1717         for 32-bit using IBM extended double format.  Fix FFI_LAST_ABI.
1718         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Handle linux variant of
1719         FFI_TYPE_LONGDOUBLE.
1720         (ffi_prep_args64): Assert using IBM extended double.
1721         (ffi_prep_cif_machdep): Don't munge FFI_TYPE_LONGDOUBLE type.
1722         Handle FFI_LINUX FFI_TYPE_LONGDOUBLE return and args.
1723         (ffi_call): Handle FFI_LINUX.
1724         (ffi_closure_helper_SYSV): Non FFI_LINUX long double return needs
1725         gpr3 return pointer as for struct return.  Handle FFI_LINUX
1726         FFI_TYPE_LONGDOUBLE return and args.  Don't increment "nf"
1727         unnecessarily.
1728         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Load both f1 and f2
1729         for FFI_TYPE_LONGDOUBLE.  Move epilogue insns into case table.
1730         Don't use r6 as pointer to results, instead use sp offset.  Don't
1731         make a special call to load lr with case table address, instead
1732         use offset from previous call.
1733         * src/powerpc/sysv.S (ffi_call_SYSV): Save long double return.
1734         * src/powerpc/linux64.S (ffi_call_LINUX64): Simplify long double
1735         return.
1736
1737 2006-03-15  Kaz Kojima  <kkojima@gcc.gnu.org>
1738
1739         * src/sh64/ffi.c (ffi_prep_cif_machdep): Handle float arguments
1740         passed with FP registers correctly.
1741         (ffi_closure_helper_SYSV): Likewise.
1742         * src/sh64/sysv.S: Likewise.
1743
1744 2006-03-01  Andreas Tobler  <a.tobler@schweiz.ch>
1745
1746         * testsuite/libffi.special/unwindtest.cc (closure_test_fn): Mark cif,
1747         args and userdata unused.
1748         (closure_test_fn1): Mark cif and userdata unused.
1749         (main): Remove unused res.
1750
1751 2006-02-28  Andreas Tobler  <a.tobler@schweiz.ch>
1752
1753         * testsuite/libffi.call/call.exp: Adjust FSF address. Add test runs for
1754         -O2, -O3, -Os and the warning flags -W -Wall.
1755         * testsuite/libffi.special/special.exp: Likewise.
1756         * testsuite/libffi.call/ffitest.h: Add an __UNUSED__ macro to mark
1757         unused parameter unused for gcc or else do nothing.
1758         * testsuite/libffi.special/ffitestcxx.h: Likewise.
1759         * testsuite/libffi.call/cls_12byte.c (cls_struct_12byte_gn): Mark cif
1760         and userdata unused.
1761         * testsuite/libffi.call/cls_16byte.c (cls_struct_16byte_gn): Likewise.
1762         * testsuite/libffi.call/cls_18byte.c (cls_struct_18byte_gn): Likewise.
1763         * testsuite/libffi.call/cls_19byte.c (cls_struct_19byte_gn): Likewise.
1764         * testsuite/libffi.call/cls_1_1byte.c (cls_struct_1_1byte_gn): Likewise.
1765         * testsuite/libffi.call/cls_20byte.c (cls_struct_20byte_gn): Likewise.
1766         * testsuite/libffi.call/cls_20byte1.c (cls_struct_20byte_gn): Likewise.
1767         * testsuite/libffi.call/cls_24byte.c (cls_struct_24byte_gn): Likewise.
1768         * testsuite/libffi.call/cls_2byte.c (cls_struct_2byte_gn): Likewise.
1769         * testsuite/libffi.call/cls_3_1byte.c (cls_struct_3_1byte_gn): Likewise.
1770         * testsuite/libffi.call/cls_3byte1.c (cls_struct_3byte_gn): Likewise.
1771         * testsuite/libffi.call/cls_3byte2.c (cls_struct_3byte_gn1): Likewise.
1772         * testsuite/libffi.call/cls_4_1byte.c (cls_struct_4_1byte_gn): Likewise.
1773         * testsuite/libffi.call/cls_4byte.c (cls_struct_4byte_gn): Likewise.
1774         * testsuite/libffi.call/cls_5_1_byte.c (cls_struct_5byte_gn): Likewise.
1775         * testsuite/libffi.call/cls_5byte.c (cls_struct_5byte_gn): Likewise.
1776         * testsuite/libffi.call/cls_64byte.c (cls_struct_64byte_gn): Likewise.
1777         * testsuite/libffi.call/cls_6_1_byte.c (cls_struct_6byte_gn): Likewise.
1778         * testsuite/libffi.call/cls_6byte.c (cls_struct_6byte_gn): Likewise.
1779         * testsuite/libffi.call/cls_7_1_byte.c (cls_struct_7byte_gn): Likewise.
1780         * testsuite/libffi.call/cls_7byte.c (cls_struct_7byte_gn): Likewise.
1781         * testsuite/libffi.call/cls_8byte.c (cls_struct_8byte_gn): Likewise.
1782         * testsuite/libffi.call/cls_9byte1.c (cls_struct_9byte_gn): Likewise.
1783         * testsuite/libffi.call/cls_9byte2.c (cls_struct_9byte_gn): Likewise.
1784         * testsuite/libffi.call/cls_align_double.c (cls_struct_align_gn):
1785         Likewise.
1786         * testsuite/libffi.call/cls_align_float.c (cls_struct_align_gn):
1787         Likewise.
1788         * testsuite/libffi.call/cls_align_longdouble.c (cls_struct_align_gn):
1789         Likewise.
1790         * testsuite/libffi.call/cls_align_pointer.c (cls_struct_align_fn): Cast
1791         void* to avoid compiler warning.
1792         (main): Likewise.
1793         (cls_struct_align_gn): Mark cif and userdata unused.
1794         * testsuite/libffi.call/cls_align_sint16.c (cls_struct_align_gn):
1795         Likewise.
1796         * testsuite/libffi.call/cls_align_sint32.c (cls_struct_align_gn):
1797         Likewise.
1798         * testsuite/libffi.call/cls_align_sint64.c (cls_struct_align_gn):
1799         Likewise.
1800         * testsuite/libffi.call/cls_align_uint16.c (cls_struct_align_gn):
1801         Likewise.
1802         * testsuite/libffi.call/cls_align_uint32.c (cls_struct_align_gn):
1803         Likewise.
1804         * testsuite/libffi.call/cls_double.c (cls_ret_double_fn): Likewise.
1805         * testsuite/libffi.call/cls_float.c (cls_ret_float_fn): Likewise.
1806         * testsuite/libffi.call/cls_multi_schar.c (test_func_gn): Mark cif and
1807         data unused.
1808         (main): Cast res_call to silence gcc.
1809         * testsuite/libffi.call/cls_multi_sshort.c (test_func_gn): Mark cif and
1810         data unused.
1811         (main): Cast res_call to silence gcc.
1812         * testsuite/libffi.call/cls_multi_sshortchar.c (test_func_gn): Mark cif
1813         and data unused.
1814         (main): Cast res_call to silence gcc.
1815         * testsuite/libffi.call/cls_multi_uchar.c (test_func_gn): Mark cif and
1816         data unused.
1817         (main): Cast res_call to silence gcc.
1818         * testsuite/libffi.call/cls_multi_ushort.c (test_func_gn): Mark cif and
1819         data unused.
1820         (main): Cast res_call to silence gcc.
1821         * testsuite/libffi.call/cls_multi_ushortchar.c (test_func_gn): Mark cif
1822         and data unused.
1823         (main): Cast res_call to silence gcc.
1824         * testsuite/libffi.call/cls_schar.c (cls_ret_schar_fn): Mark cif and
1825         userdata unused.
1826         (cls_ret_schar_fn): Cast printf parameter to silence gcc.
1827         * testsuite/libffi.call/cls_sint.c (cls_ret_sint_fn): Mark cif and
1828         userdata unused.
1829         (cls_ret_sint_fn): Cast printf parameter to silence gcc.
1830         * testsuite/libffi.call/cls_sshort.c (cls_ret_sshort_fn): Mark cif and
1831         userdata unused.
1832         (cls_ret_sshort_fn): Cast printf parameter to silence gcc.
1833         * testsuite/libffi.call/cls_uchar.c (cls_ret_uchar_fn):  Mark cif and
1834         userdata unused.
1835         (cls_ret_uchar_fn): Cast printf parameter to silence gcc.
1836         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Mark cif and
1837         userdata unused.
1838         (cls_ret_uint_fn): Cast printf parameter to silence gcc.
1839         * testsuite/libffi.call/cls_ulonglong.c (cls_ret_ulonglong_fn): Mark cif
1840         and userdata unused.
1841         * testsuite/libffi.call/cls_ushort.c (cls_ret_ushort_fn): Mark cif and
1842         userdata unused.
1843         (cls_ret_ushort_fn): Cast printf parameter to silence gcc.
1844         * testsuite/libffi.call/float.c (floating): Remove unused parameter e.
1845         * testsuite/libffi.call/float1.c (main): Remove unused variable i.
1846         Cleanup white spaces.
1847         * testsuite/libffi.call/negint.c (checking): Remove unused variable i.
1848         * testsuite/libffi.call/nested_struct.c (cls_struct_combined_gn): Mark
1849         cif and userdata unused.
1850         * testsuite/libffi.call/nested_struct1.c (cls_struct_combined_gn):
1851         Likewise.
1852         * testsuite/libffi.call/nested_struct10.c (B_gn): Likewise.
1853         * testsuite/libffi.call/nested_struct2.c (B_fn): Adjust printf
1854         formatters to silence gcc.
1855         (B_gn): Mark cif and userdata unused.
1856         * testsuite/libffi.call/nested_struct3.c (B_gn): Mark cif and userdata
1857         unused.
1858         * testsuite/libffi.call/nested_struct4.c: Mention related PR.
1859         (B_gn): Mark cif and userdata unused.
1860         * testsuite/libffi.call/nested_struct5.c (B_gn): Mark cif and userdata
1861         unused.
1862         * testsuite/libffi.call/nested_struct6.c: Mention related PR.
1863         (B_gn): Mark cif and userdata unused.
1864         * testsuite/libffi.call/nested_struct7.c (B_gn): Mark cif and userdata
1865         unused.
1866         * testsuite/libffi.call/nested_struct8.c (B_gn): Likewise.
1867         * testsuite/libffi.call/nested_struct9.c (B_gn): Likewise.
1868         * testsuite/libffi.call/problem1.c (stub): Likewise.
1869         * testsuite/libffi.call/pyobjc-tc.c (main): Cast the result to silence
1870         gcc.
1871         * testsuite/libffi.call/return_fl2.c (return_fl): Add the note mentioned
1872         in the last commit for this test case in the test case itself.
1873         * testsuite/libffi.call/closure_fn0.c (closure_test_fn0): Mark cif as
1874         unused.
1875         * testsuite/libffi.call/closure_fn1.c (closure_test_fn1): Likewise.
1876         * testsuite/libffi.call/closure_fn2.c (closure_test_fn2): Likewise.
1877         * testsuite/libffi.call/closure_fn3.c (closure_test_fn3): Likewise.
1878         * testsuite/libffi.call/closure_fn4.c (closure_test_fn0): Likewise.
1879         * testsuite/libffi.call/closure_fn5.c (closure_test_fn5): Likewise.
1880         * testsuite/libffi.call/closure_fn6.c (closure_test_fn0): Likewise.
1881
1882 2006-02-22  Kaz Kojima  <kkojima@gcc.gnu.org>
1883
1884         * src/sh/sysv.S: Fix register numbers in the FDE for
1885         ffi_closure_SYSV.
1886
1887 2006-02-20  Andreas Tobler  <a.tobler@schweiz.ch>
1888
1889         * testsuite/libffi.call/return_fl2.c (return_fl): Remove static
1890         declaration to avoid a false negative on ix86. See PR323.
1891
1892 2006-02-18  Kaz Kojima  <kkojima@gcc.gnu.org>
1893
1894         * src/sh/ffi.c (ffi_closure_helper_SYSV): Remove unused variable
1895         and cast integer to void * if needed.  Update the pointer to
1896         the FP register saved area correctly.
1897
1898 2006-02-17  Andreas Tobler  <a.tobler@schweiz.ch>
1899
1900         * testsuite/libffi.call/nested_struct6.c: XFAIL this test until PR25630
1901         is fixed.
1902         * testsuite/libffi.call/nested_struct4.c: Likewise.
1903
1904 2006-02-16  Andreas Tobler  <a.tobler@schweiz.ch>
1905
1906         * testsuite/libffi.call/return_dbl.c: New test case.
1907         * testsuite/libffi.call/return_dbl1.c: Likewise.
1908         * testsuite/libffi.call/return_dbl2.c: Likewise.
1909         * testsuite/libffi.call/return_fl.c: Likewise.
1910         * testsuite/libffi.call/return_fl1.c: Likewise.
1911         * testsuite/libffi.call/return_fl2.c: Likewise.
1912         * testsuite/libffi.call/return_fl3.c: Likewise.
1913         * testsuite/libffi.call/closure_fn6.c: Likewise.
1914
1915         * testsuite/libffi.call/nested_struct2.c: Remove ffi_type_mylong
1916         definition.
1917         * testsuite/libffi.call/ffitest.h: Add ffi_type_mylong definition
1918         here to be used by other test cases too.
1919
1920         * testsuite/libffi.call/nested_struct10.c: New test case.
1921         * testsuite/libffi.call/nested_struct9.c: Likewise.
1922         * testsuite/libffi.call/nested_struct8.c: Likewise.
1923         * testsuite/libffi.call/nested_struct7.c: Likewise.
1924         * testsuite/libffi.call/nested_struct6.c: Likewise.
1925         * testsuite/libffi.call/nested_struct5.c: Likewise.
1926         * testsuite/libffi.call/nested_struct4.c: Likewise.
1927
1928 2006-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
1929
1930         * configure.ac: Enable libffi for sparc64-*-freebsd*.
1931         * configure: Rebuilt.
1932
1933 2006-01-18  Jakub Jelinek  <jakub@redhat.com>
1934
1935         * src/powerpc/sysv.S (smst_two_register): Don't call __ashldi3,
1936         instead do the shifting inline.
1937         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't compute %r5
1938         shift count unconditionally.  Simplify load sequences for 1, 2, 3, 4
1939         and 8 byte structs, for the remaining struct sizes don't call
1940         __lshrdi3, instead do the shifting inline.
1941
1942 2005-12-07  Thiemo Seufer  <ths@networkno.de>
1943
1944         * src/mips/ffitarget.h: Remove obsolete sgidefs.h include. Add
1945         missing parentheses.
1946         * src/mips/o32.S (ffi_call_O32): Code formatting. Define
1947         and use A3_OFF, FP_OFF, RA_OFF. Micro-optimizations.
1948         (ffi_closure_O32): Likewise, but with newly defined A3_OFF2,
1949         A2_OFF2, A1_OFF2, A0_OFF2, RA_OFF2, FP_OFF2, S0_OFF2, GP_OFF2,
1950         V1_OFF2, V0_OFF2, FA_1_1_OFF2, FA_1_0_OFF2, FA_0_1_OFF2,
1951         FA_0_0_OFF2.
1952         * src/mips/ffi.c (ffi_prep_args): Code formatting. Fix
1953         endianness bugs.
1954         (ffi_prep_closure): Improve trampoline instruction scheduling.
1955         (ffi_closure_mips_inner_O32): Fix endianness bugs.
1956
1957 2005-12-03  Alan Modra  <amodra@bigpond.net.au>
1958
1959         * src/powerpc/ffi.c: Formatting.
1960         (ffi_prep_args_SYSV): Avoid possible aliasing problems by using unions.
1961         (ffi_prep_args64): Likewise.
1962
1963 2005-09-30  Geoffrey Keating  <geoffk@apple.com>
1964
1965         * testsuite/lib/libffi-dg.exp (libffi_target_compile): For
1966         darwin, use -shared-libgcc not -lgcc_s, and explain why.
1967
1968 2005-09-26  Tom Tromey  <tromey@redhat.com>
1969
1970         * testsuite/libffi.call/float1.c (value_type): New typedef.
1971         (CANARY): New define.
1972         (main): Check for result buffer overflow.
1973         * src/powerpc/linux64.S: Handle linux64 long double returns.
1974         * src/powerpc/ffi.c (FLAG_RETURNS_128BITS): New constant.
1975         (ffi_prep_cif_machdep): Handle linux64 long double returns.
1976
1977 2005-08-25  Alan Modra  <amodra@bigpond.net.au>
1978
1979         PR target/23404
1980         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Correct placement of stack
1981         homed fp args.
1982         (ffi_status ffi_prep_cif_machdep): Correct stack sizing for same.
1983
1984 2005-08-11  Jakub Jelinek  <jakub@redhat.com>
1985
1986         * configure.ac (HAVE_HIDDEN_VISIBILITY_ATTRIBUTE): New test.
1987         (AH_BOTTOM): Add FFI_HIDDEN definition.
1988         * configure: Rebuilt.
1989         * fficonfig.h.in: Rebuilt.
1990         * src/powerpc/ffi.c (hidden): Remove.
1991         (ffi_closure_LINUX64, ffi_prep_args64, ffi_call_LINUX64,
1992         ffi_closure_helper_LINUX64): Use FFI_HIDDEN instead of hidden.
1993         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64,
1994         .ffi_closure_LINUX64): Use FFI_HIDDEN instead of .hidden.
1995         * src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV): Remove,
1996         add FFI_HIDDEN to its prototype.
1997         (ffi_closure_SYSV_inner): New.
1998         * src/x86/sysv.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
1999         * src/x86/win32.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2000
2001 2005-08-10  Alfred M. Szmidt  <ams@gnu.org>
2002
2003         PR libffi/21819:
2004         * configure: Rebuilt.
2005         * configure.ac: Handle i*86-*-gnu*.
2006
2007 2005-08-09  Jakub Jelinek  <jakub@redhat.com>
2008
2009         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Use
2010         DW_CFA_offset_extended_sf rather than
2011         DW_CFA_GNU_negative_offset_extended.
2012         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise.
2013
2014 2005-07-22  SUGIOKA Toshinobu  <sugioka@itonet.co.jp>
2015
2016         * src/sh/sysv.S (ffi_call_SYSV): Stop argument popping correctly
2017         on sh3.
2018         (ffi_closure_SYSV): Change the stack layout for sh3 struct argument.
2019         * src/sh/ffi.c (ffi_prep_args): Fix sh3 argument copy, when it is
2020         partially on register.
2021         (ffi_closure_helper_SYSV): Likewise.
2022         (ffi_prep_cif_machdep): Don't set too many cif->flags.
2023
2024 2005-07-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2025
2026         * src/sh/ffi.c (ffi_call): Handle small structures correctly.
2027         Remove empty line.
2028         * src/sh64/ffi.c (simple_type): Remove.
2029         (return_type): Handle small structures correctly.
2030         (ffi_prep_args): Likewise.
2031         (ffi_call): Likewise.
2032         (ffi_closure_helper_SYSV): Likewise.
2033         * src/sh64/sysv.S (ffi_call_SYSV): Handle 1, 2 and 4-byte return.
2034         Emit position independent code if PIC and remove wrong datalabel
2035         prefixes from EH data.
2036
2037 2005-07-19  Andreas Tobler  <a.tobler@schweiz.ch>
2038
2039         * Makefile.am (nodist_libffi_la_SOURCES): Add POWERPC_FREEBSD.
2040         * Makefile.in: Regenerate.
2041         * include/Makefile.in: Likewise.
2042         * testsuite/Makefile.in: Likewise.
2043         * configure.ac: Add POWERPC_FREEBSD rules.
2044         * configure: Regenerate.
2045         * src/powerpc/ffitarget.h: Add POWERPC_FREEBSD rules.
2046         (FFI_SYSV_TYPE_SMALL_STRUCT): Define.
2047         * src/powerpc/ffi.c: Add flags to handle small structure returns
2048         in ffi_call_SYSV.
2049         (ffi_prep_cif_machdep): Handle small structures for SYSV 4 ABI.
2050         Aka FFI_SYSV.
2051         (ffi_closure_helper_SYSV): Likewise.
2052         * src/powerpc/ppc_closure.S: Add return types for small structures.
2053         * src/powerpc/sysv.S: Add bits to handle small structures for
2054         final SYSV 4 ABI.
2055
2056 2005-07-10  Andreas Tobler  <a.tobler@schweiz.ch>
2057
2058         * testsuite/libffi.call/cls_5_1_byte.c: New test file.
2059         * testsuite/libffi.call/cls_6_1_byte.c: Likewise.
2060         * testsuite/libffi.call/cls_7_1_byte.c: Likewise.
2061
2062 2005-07-05  Randolph Chung  <tausq@debian.org>
2063
2064         * src/pa/ffi.c (ffi_struct_type): Rename FFI_TYPE_SMALL_STRUCT1
2065         as FFI_TYPE_SMALL_STRUCT3.  Break out handling for 5-7 byte
2066         structures.  Kill compilation warnings.
2067         (ffi_closure_inner_LINUX): Print return values as hex in debug
2068         message.  Rename FFI_TYPE_SMALL_STRUCT1 as FFI_TYPE_SMALL_STRUCT3.
2069         Properly handle 5-7 byte structure returns.
2070         * src/pa/ffitarget.h (FFI_TYPE_SMALL_STRUCT1)
2071         (FFI_TYPE_SMALL_STRUCT2): Remove.
2072         (FFI_TYPE_SMALL_STRUCT3, FFI_TYPE_SMALL_STRUCT5)
2073         (FFI_TYPE_SMALL_STRUCT6, FFI_TYPE_SMALL_STRUCT7): Define.
2074         * src/pa/linux.S: Mark source file as using PA1.1 assembly.
2075         (checksmst1, checksmst2): Remove.
2076         (checksmst3): Optimize handling of 3-byte struct returns.
2077         (checksmst567): Properly handle 5-7 byte struct returns.
2078
2079 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2080
2081         PR libgcj/21943
2082         * src/mips/n32.S: Enforce PIC code.
2083         * src/mips/o32.S: Likewise.
2084
2085 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2086
2087         * configure.ac: Treat i*86-*-solaris2.10 and up as X86_64.
2088         * configure: Regenerate.
2089
2090 2005-06-01  Alan Modra  <amodra@bigpond.net.au>
2091
2092         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't use JUMPTARGET
2093         to call ffi_closure_helper_SYSV.  Append @local instead.
2094         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise for ffi_prep_args_SYSV.
2095
2096 2005-05-17  Kelley Cook  <kcook@gcc.gnu.org>
2097
2098         * configure.ac: Use AC_C_BIGENDIAN instead of AC_C_BIGENDIAN_CROSS.
2099         Use AC_CHECK_SIZEOF instead of AC_COMPILE_CHECK_SIZEOF.
2100         * Makefile.am (ACLOCAL_AMFLAGS): Remove -I ../config.
2101         * aclocal.m4, configure, fficonfig.h.in, Makefile.in,
2102         include/Makefile.in, testsuite/Makefile.in: Regenerate.
2103
2104 2005-05-09  Mike Stump  <mrs@apple.com>
2105
2106         * configure: Regenerate.
2107
2108 2005-05-08  Richard Henderson  <rth@redhat.com>
2109
2110         PR libffi/21285
2111         * src/alpha/osf.S: Update unwind into to match code.
2112
2113 2005-05-04  Andreas Degert <ad@papyrus-gmbh.de>
2114             Richard Henderson  <rth@redhat.com>
2115
2116         * src/x86/ffi64.c (ffi_prep_cif_machdep): Save sse-used flag in
2117         bit 11 of flags.
2118         (ffi_call): Mask return type field.  Pass ssecount to ffi_call_unix64.
2119         (ffi_prep_closure): Set carry bit if sse-used flag set.
2120         * src/x86/unix64.S (ffi_call_unix64): Add ssecount argument.
2121         Only load sse registers if ssecount non-zero.
2122         (ffi_closure_unix64): Only save sse registers if carry set on entry.
2123
2124 2005-04-29  Ralf Corsepius  <ralf.corsepius@rtems.org>
2125
2126         * configure.ac: Add i*86-*-rtems*, sparc*-*-rtems*,
2127         powerpc-*rtems*, arm*-*-rtems*, sh-*-rtems*.
2128         * configure: Regenerate.
2129
2130 2005-04-20  Hans-Peter Nilsson  <hp@axis.com>
2131
2132         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): In regsub use,
2133         have Tcl8.3-compatible intermediate variable.
2134
2135 2005-04-18  Simon Posnjak <simon.posnjak@siol.net>
2136             Hans-Peter Nilsson  <hp@axis.com>
2137
2138         * Makefile.am: Add CRIS support.
2139         * configure.ac: Likewise.
2140         * Makefile.in, configure, testsuite/Makefile.in,
2141         include/Makefile.in: Regenerate.
2142         * src/cris: New directory.
2143         * src/cris/ffi.c, src/cris/sysv.S, src/cris/ffitarget.h: New files.
2144         * src/prep_cif.c (ffi_prep_cif): Wrap in #ifndef __CRIS__.
2145
2146         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): Replace \n with
2147         \r?\n in output tests.
2148
2149 2005-04-12  Mike Stump  <mrs@apple.com>
2150
2151         * configure: Regenerate.
2152
2153 2005-03-30  Hans Boehm  <Hans.Boehm@hp.com>
2154
2155         * src/ia64/ffitarget.h (ffi_arg): Use long long instead of DI.
2156
2157 2005-03-30  Steve Ellcey  <sje@cup.hp.com>
2158
2159         * src/ia64/ffitarget.h (ffi_arg) ADD DI attribute.
2160         (ffi_sarg) Ditto.
2161         * src/ia64/unix.S (ffi_closure_unix): Extend gp
2162         to 64 bits in ILP32 mode.
2163         Load 64 bits even for short data.
2164
2165 2005-03-23  Mike Stump  <mrs@apple.com>
2166
2167         * src/powerpc/darwin.S: Update for -m64 multilib.
2168         * src/powerpc/darwin_closure.S: Likewise.
2169
2170 2005-03-21  Zack Weinberg  <zack@codesourcery.com>
2171
2172         * configure.ac: Do not invoke TL_AC_GCC_VERSION.
2173         Do not set tool_include_dir.
2174         * aclocal.m4, configure, Makefile.in, testsuite/Makefile.in:
2175         Regenerate.
2176         * include/Makefile.am: Set gcc_version and toollibffidir.
2177         * include/Makefile.in: Regenerate.
2178
2179 2005-02-22  Andrew Haley  <aph@redhat.com>
2180
2181         * src/powerpc/ffi.c (ffi_prep_cif_machdep): Bump alignment to
2182         odd-numbered register pairs for 64-bit integer types.
2183
2184 2005-02-23  Andreas Tobler  <a.tobler@schweiz.ch>
2185
2186         PR libffi/20104
2187         * testsuite/libffi.call/return_ll1.c: New test case.
2188
2189 2005-02-11  Janis Johnson  <janis187@us.ibm.com>
2190
2191         * testsuite/libffi.call/cls_align_longdouble.c: Remove dg-options.
2192         * testsuite/libffi.call/float.c: Ditto.
2193         * testsuite/libffi.call/float2.c: Ditto.
2194         * testsuite/libffi.call/float3.c: Ditto.
2195
2196 2005-02-08  Andreas Tobler  <a.tobler@schweiz.ch>
2197
2198         * src/frv/ffitarget.h: Remove PPC stuff which does not belong to frv.
2199
2200 2005-01-12  Eric Botcazou  <ebotcazou@libertysurf.fr>
2201
2202         * testsuite/libffi.special/special.exp (cxx_options): Add
2203         -shared-libgcc.
2204
2205 2004-12-31  Richard Henderson  <rth@redhat.com>
2206
2207         * src/types.c (FFI_AGGREGATE_TYPEDEF): Remove.
2208         (FFI_TYPEDEF): Rename from FFI_INTEGRAL_TYPEDEF.  Replace size and
2209         offset parameters with a type parameter; deduce size and structure
2210         alignment.  Update all users.
2211
2212 2004-12-31  Richard Henderson  <rth@redhat.com>
2213
2214         * src/types.c (FFI_TYPE_POINTER): Define with sizeof.
2215         (FFI_TYPE_LONGDOUBLE): Fix for ia64.
2216         * src/ia64/ffitarget.h (struct ffi_ia64_trampoline_struct): Move
2217         into ffi_prep_closure.
2218         * src/ia64/ia64_flags.h, src/ia64/ffi.c, src/ia64/unix.S: Rewrite
2219         from scratch.
2220
2221 2004-12-27  Richard Henderson  <rth@redhat.com>
2222
2223         * src/x86/unix64.S: Fix typo in unwind info.
2224
2225 2004-12-25  Richard Henderson  <rth@redhat.com>
2226
2227         * src/x86/ffi64.c (struct register_args): Rename from stackLayout.
2228         (enum x86_64_reg_class): Add X86_64_COMPLEX_X87_CLASS.
2229         (merge_classes): Check for it.
2230         (SSE_CLASS_P): New.
2231         (classify_argument): Pass byte_offset by value; perform all updates
2232         inside struct case.
2233         (examine_argument): Add classes argument; handle
2234         X86_64_COMPLEX_X87_CLASS.
2235         (ffi_prep_args): Merge into ...
2236         (ffi_call): ... here.  Share stack frame with ffi_call_unix64.
2237         (ffi_prep_cif_machdep): Setup cif->flags for proper structure return.
2238         (ffi_fill_return_value): Remove.
2239         (ffi_prep_closure): Remove dead assert.
2240         (ffi_closure_unix64_inner): Rename from ffi_closure_UNIX64_inner.
2241         Rewrite to use struct register_args instead of va_list.  Create
2242         flags for handling structure returns.
2243         * src/x86/unix64.S: Remove dead strings.
2244         (ffi_call_unix64): Rename from ffi_call_UNIX64.  Rewrite to share
2245         stack frame with ffi_call.  Handle structure returns properly.
2246         (float2sse, floatfloat2sse, double2sse): Remove.
2247         (sse2float, sse2double, sse2floatfloat): Remove.
2248         (ffi_closure_unix64): Rename from ffi_closure_UNIX64.  Rewrite
2249         to handle structure returns properly.
2250
2251 2004-12-08  David Edelsohn  <edelsohn@gnu.org>
2252
2253         * Makefile.am (AM_MAKEFLAGS): Remove duplicate LIBCFLAGS and
2254         PICFLAG.
2255         * Makefile.in: Regenerated.
2256
2257 2004-12-02  Richard Sandiford  <rsandifo@redhat.com>
2258
2259         * configure.ac: Use TL_AC_GCC_VERSION to set gcc_version.
2260         * configure, aclocal.m4, Makefile.in: Regenerate.
2261         * include/Makefile.in, testsuite/Makefile.in: Regenerate.
2262
2263 2004-11-29  Kelley Cook  <kcook@gcc.gnu.org>
2264
2265         * configure: Regenerate for libtool change.
2266
2267 2004-11-25  Kelley Cook  <kcook@gcc.gnu.org>
2268
2269         * configure: Regenerate for libtool reversion.
2270
2271 2004-11-24  Kelley Cook  <kcook@gcc.gnu.org>
2272
2273         * configure: Regenerate for libtool change.
2274
2275 2004-11-23  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2276
2277         * testsuite/lib/libffi-dg.exp: Use new procs in target-libpath.exp.
2278
2279 2004-11-23  Richard Sandiford  <rsandifo@redhat.com>
2280
2281         * src/mips/o32.S (ffi_call_O32, ffi_closure_O32): Use jalr instead
2282         of jal.  Use an absolute encoding for the frame information.
2283
2284 2004-11-23  Kelley Cook  <kcook@gcc.gnu.org>
2285
2286         * Makefile.am: Remove no-dependencies.  Add ACLOCAL_AMFLAGS.
2287         * acinclude.m4: Delete logic for sincludes.
2288         * aclocal.m4, Makefile.in, configure: Regenerate.
2289         * include/Makefile: Likewise.
2290         * testsuite/Makefile: Likewise.
2291
2292 2004-11-22  Eric Botcazou  <ebotcazou@libertysurf.fr>
2293
2294         * src/sparc/ffi.c (ffi_prep_closure): Align doubles and 64-bit integers
2295         on a 8-byte boundary.
2296         * src/sparc/v8.S (ffi_closure_v8): Reserve frame space for arguments.
2297
2298 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2299
2300         * src/arm/ffi.c (ffi_prep_cif_machdep): Handle functions that return
2301         long long values.  Round stack allocation to a multiple of 8 bytes
2302         for ATPCS compatibility.
2303         * src/arm/sysv.S (ffi_call_SYSV): Rework to avoid use of APCS register
2304         names.  Handle returning long long types.  Add Thumb and interworking
2305         support.  Improve soft-float code.
2306
2307 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2308
2309         * testsuite/lib/libffi-db.exp (load_gcc_lib): New function.
2310         (libffi_exit): New function.
2311         (libffi_init): Build the testglue wrapper if needed.
2312
2313 2004-10-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2314
2315         PR other/18138
2316         * testsuite/lib/libffi-dg.exp: Accept more than one multilib libgcc.
2317
2318 2004-10-25  Kazuhiro Inaoka <inaoka.kazuhiro@renesas.com>
2319
2320         * src/m32r/libffitarget.h (FFI_CLOSURES): Set to 0.
2321
2322 2004-10-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2323
2324         * src/sh/sysv.S (ffi_call_SYSV): Don't align for double data.
2325         * testsuite/libffi.call/float3.c: New test case.
2326
2327 2004-10-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2328
2329         * src/sh/ffi.c (ffi_prep_closure): Set T bit in trampoline for
2330         the function returning a structure pointed with R2.
2331         * src/sh/sysv.S (ffi_closure_SYSV): Use R2 as the pointer to
2332         the structure return value if T bit set.  Emit position
2333         independent code and EH data if PIC.
2334
2335 2004-10-13  Kazuhiro Inaoka  <inaoka.kazuhiro@renesas.com>
2336
2337         * Makefile.am: Add m32r support.
2338         * configure.ac: Likewise.
2339         * Makefile.in: Regenerate.
2340         * confiugre: Regenerate.
2341         * src/types.c: Add m32r port to FFI_INTERNAL_TYPEDEF
2342         (uint64, sint64, double, longdouble)
2343         * src/m32r: New directory.
2344         * src/m32r/ffi.c: New file.
2345         * src/m32r/sysv.S: Likewise.
2346         * src/m32r/ffitarget.h: Likewise.
2347
2348 2004-10-02  Kaz Kojima  <kkojima@gcc.gnu.org>
2349
2350         * testsuite/libffi.call/negint.c: New test case.
2351
2352 2004-09-14  H.J. Lu  <hongjiu.lu@intel.com>
2353
2354         PR libgcj/17465
2355         * testsuite/lib/libffi-dg.exp: Don't use global ld_library_path.
2356         Set up LD_LIBRARY_PATH, SHLIB_PATH, LD_LIBRARYN32_PATH,
2357         LD_LIBRARY64_PATH, LD_LIBRARY_PATH_32, LD_LIBRARY_PATH_64 and
2358         DYLD_LIBRARY_PATH.
2359
2360 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2361
2362         * testsuite/libffi.call/many_win32.c: Remove whitespaces.
2363         * testsuite/libffi.call/promotion.c: Likewise.
2364         * testsuite/libffi.call/return_ll.c: Remove unused var. Cleanup
2365         whitespaces.
2366         * testsuite/libffi.call/return_sc.c: Likewise.
2367         * testsuite/libffi.call/return_uc.c: Likewise.
2368
2369 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2370
2371         * src/powerpc/darwin.S: Fix comments and identation.
2372         * src/powerpc/darwin_closure.S: Likewise.
2373
2374 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2375
2376         * src/powerpc/ffi_darwin.c: Add flag for longdouble return values.
2377         (ffi_prep_args): Handle longdouble arguments.
2378         (ffi_prep_cif_machdep): Set flags for longdouble. Calculate space for
2379         longdouble.
2380         (ffi_closure_helper_DARWIN): Add closure handling for longdouble.
2381         * src/powerpc/darwin.S (_ffi_call_DARWIN): Add handling of longdouble
2382         values.
2383         * src/powerpc/darwin_closure.S (_ffi_closure_ASM): Likewise.
2384         * src/types.c: Defined longdouble size and alignment for darwin.
2385
2386 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2387
2388         * src/powerpc/aix.S: Remove whitespaces.
2389         * src/powerpc/aix_closure.S: Likewise.
2390         * src/powerpc/asm.h: Likewise.
2391         * src/powerpc/ffi.c: Likewise.
2392         * src/powerpc/ffitarget.h: Likewise.
2393         * src/powerpc/linux64.S: Likewise.
2394         * src/powerpc/linux64_closure.S: Likewise.
2395         * src/powerpc/ppc_closure.S: Likewise.
2396         * src/powerpc/sysv.S: Likewise.
2397
2398 2004-08-30  Anthony Green  <green@redhat.com>
2399
2400         * Makefile.am: Add frv support.
2401         * Makefile.in, testsuite/Makefile.in: Rebuilt.
2402         * configure.ac: Read configure.host.
2403         * configure.in: Read configure.host.
2404         * configure.host: New file.  frv-elf needs libgloss.
2405         * include/ffi.h.in: Force ffi_closure to have a nice big (8)
2406         alignment.  This is needed to frv and shouldn't harm the others.
2407         * include/ffi_common.h (ALIGN_DOWN): New macro.
2408         * src/frv/ffi.c, src/frv/ffitarget.h, src/frv/eabi.S: New files.
2409
2410 2004-08-24  David Daney  <daney@avtrex.com>
2411
2412         * testsuite/libffi.call/closure_fn0.c: Xfail mips64* instead of mips*.
2413         * testsuite/libffi.call/closure_fn1.c: Likewise.
2414         * testsuite/libffi.call/closure_fn2.c  Likewise.
2415         * testsuite/libffi.call/closure_fn3.c: Likewise.
2416         * testsuite/libffi.call/closure_fn4.c: Likewise.
2417         * testsuite/libffi.call/closure_fn5.c: Likewise.
2418         * testsuite/libffi.call/cls_18byte.c: Likewise.
2419         * testsuite/libffi.call/cls_19byte.c: Likewise.
2420         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2421         * testsuite/libffi.call/cls_20byte.c: Likewise.
2422         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2423         * testsuite/libffi.call/cls_24byte.c: Likewise.
2424         * testsuite/libffi.call/cls_2byte.c: Likewise.
2425         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2426         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2427         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2428         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2429         * testsuite/libffi.call/cls_4byte.c: Likewise.
2430         * testsuite/libffi.call/cls_64byte.c: Likewise.
2431         * testsuite/libffi.call/cls_6byte.c: Likewise.
2432         * testsuite/libffi.call/cls_7byte.c: Likewise.
2433         * testsuite/libffi.call/cls_8byte.c: Likewise.
2434         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2435         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2436         * testsuite/libffi.call/cls_align_double.c: Likewise.
2437         * testsuite/libffi.call/cls_align_float.c: Likewise.
2438         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2439         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2440         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2441         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2442         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2443         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2444         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2445         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
2446         * testsuite/libffi.call/cls_double.c: Likewise.
2447         * testsuite/libffi.call/cls_float.c: Likewise.
2448         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2449         * testsuite/libffi.call/cls_multi_sshort.c: Likewise.
2450         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2451         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2452         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2453         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2454         * testsuite/libffi.call/cls_schar.c: Likewise.
2455         * testsuite/libffi.call/cls_sint.c: Likewise.
2456         * testsuite/libffi.call/cls_sshort.c: Likewise.
2457         * testsuite/libffi.call/cls_uchar.c: Likewise.
2458         * testsuite/libffi.call/cls_uint.c: Likewise.
2459         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2460         * testsuite/libffi.call/cls_ushort.c: Likewise.
2461         * testsuite/libffi.call/nested_struct.c: Likewise.
2462         * testsuite/libffi.call/nested_struct1.c: Likewise.
2463         * testsuite/libffi.call/nested_struct2.c: Likewise.
2464         * testsuite/libffi.call/nested_struct3.c: Likewise.
2465         * testsuite/libffi.call/problem1.c: Likewise.
2466         * testsuite/libffi.special/unwindtest.cc: Likewise.
2467         * testsuite/libffi.call/cls_12byte.c: Likewise and set return value
2468         to zero.
2469         * testsuite/libffi.call/cls_16byte.c: Likewise.
2470         * testsuite/libffi.call/cls_5byte.c: Likewise.
2471
2472 2004-08-23  David Daney <daney@avtrex.com>
2473
2474         PR libgcj/13141
2475         * src/mips/ffitarget.h (FFI_O32_SOFT_FLOAT): New ABI.
2476         * src/mips/ffi.c (ffi_prep_args): Fix alignment calculation.
2477         (ffi_prep_cif_machdep): Handle FFI_O32_SOFT_FLOAT floating point
2478         parameters and return types.
2479         (ffi_call): Handle FFI_O32_SOFT_FLOAT ABI.
2480         (ffi_prep_closure): Ditto.
2481         (ffi_closure_mips_inner_O32): Handle FFI_O32_SOFT_FLOAT ABI, fix
2482         alignment calculations.
2483         * src/mips/o32.S (ffi_closure_O32): Don't use floating point
2484         instructions if FFI_O32_SOFT_FLOAT, make stack frame ABI compliant.
2485
2486 2004-08-14  Casey Marshall <csm@gnu.org>
2487
2488         * src/mips/ffi.c (ffi_pref_cif_machdep): set `cif->flags' to
2489         contain `FFI_TYPE_UINT64' as return type for any 64-bit
2490         integer (O32 ABI only).
2491         (ffi_prep_closure): new function.
2492         (ffi_closure_mips_inner_O32): new function.
2493         * src/mips/ffitarget.h: Define `FFI_CLOSURES' and
2494         `FFI_TRAMPOLINE_SIZE' appropriately if the ABI is o32.
2495         * src/mips/o32.S (ffi_call_O32): add labels for .eh_frame. Return
2496         64 bit integers correctly.
2497         (ffi_closure_O32): new function.
2498         Added DWARF-2 unwind info for both functions.
2499
2500 2004-08-10  Andrew Haley  <aph@redhat.com>
2501
2502         * src/x86/ffi64.c (ffi_prep_args ): 8-align all stack arguments.
2503
2504 2004-08-01  Robert Millan  <robertmh@gnu.org>
2505
2506         * configure.ac: Detect knetbsd-gnu and kfreebsd-gnu.
2507         * configure: Regenerate.
2508
2509 2004-07-30  Maciej W. Rozycki  <macro@linux-mips.org>
2510
2511         * acinclude.m4 (AC_FUNC_MMAP_BLACKLIST): Check for <sys/mman.h>
2512         and mmap() explicitly instead of relying on preset autoconf cache
2513         variables.
2514         * aclocal.m4: Regenerate.
2515         * configure: Regenerate.
2516
2517 2004-07-11  Ulrich Weigand  <uweigand@de.ibm.com>
2518
2519         * src/s390/ffi.c (ffi_prep_args): Fix C aliasing violation.
2520         (ffi_check_float_struct): Remove unused prototype.
2521
2522 2004-06-30  Geoffrey Keating  <geoffk@apple.com>
2523
2524         * src/powerpc/ffi_darwin.c (flush_icache): ';' is a comment
2525         character on Darwin, use '\n\t' instead.
2526
2527 2004-06-26  Matthias Klose  <doko@debian.org>
2528
2529         * libtool-version: Fix typo in revision/age.
2530
2531 2004-06-17  Matthias Klose  <doko@debian.org>
2532
2533         * libtool-version: New.
2534         * Makefile.am (libffi_la_LDFLAGS): Use -version-info for soname.
2535         * Makefile.in: Regenerate.
2536
2537 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2538
2539         * Makefile.am: Remove useless multilib rules.
2540         * Makefile.in: Regenerate.
2541         * aclocal.m4: Regenerate with automake 1.8.5.
2542         * configure.ac: Remove useless multilib configury.
2543         * configure: Regenerate.
2544
2545 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2546
2547         * .cvsignore: New file.
2548
2549 2004-06-10  Jakub Jelinek  <jakub@redhat.com>
2550
2551         * src/ia64/unix.S (ffi_call_unix): Insert group barrier break
2552         fp_done.
2553         (ffi_closure_UNIX): Fix f14/f15 adjustment if FLOAT_SZ is ever
2554         changed from 8.
2555
2556 2004-06-06  Sean McNeil  <sean@mcneil.com>
2557
2558         * configure.ac: Add x86_64-*-freebsd* support.
2559         * configure: Regenerate.
2560
2561 2004-04-26  Joe Buck <jbuck@welsh-buck.org>
2562
2563         Bug 15093
2564         * configure.ac: Test for existence of mmap and sys/mman.h before
2565         checking blacklist.  Fix suggested by Jim Wilson.
2566         * configure: Regenerate.
2567
2568 2004-04-26  Matt Austern  <austern@apple.com>
2569
2570         * src/powerpc/darwin.S: Go through a non-lazy pointer for initial
2571         FDE location.
2572         * src/powerpc/darwin_closure.S: Likewise.
2573
2574 2004-04-24  Andreas Tobler  <a.tobler@schweiz.ch>
2575
2576         * testsuite/libffi.call/cls_multi_schar.c (main): Fix initialization
2577         error. Reported by Thomas Heller <theller@python.net>.
2578         * testsuite/libffi.call/cls_multi_sshort.c (main): Likewise.
2579         * testsuite/libffi.call/cls_multi_ushort.c (main): Likewise.
2580
2581 2004-03-20  Matthias Klose  <doko@debian.org>
2582
2583         * src/pa/linux.S: Fix typo.
2584
2585 2004-03-19  Matthias Klose  <doko@debian.org>
2586
2587         * Makefile.am: Update.
2588         * Makefile.in: Regenerate.
2589         * src/pa/ffi.h.in: Remove.
2590         * src/pa/ffitarget.h: New file.
2591
2592 2004-02-10  Randolph Chung  <tausq@debian.org>
2593
2594         * Makefile.am: Add PA support.
2595         * Makefile.in: Regenerate.
2596         * include/Makefile.in: Regenerate.
2597         * configure.ac: Add PA target.
2598         * configure: Regenerate.
2599         * src/pa/ffi.c: New file.
2600         * src/pa/ffi.h.in: Add PA support.
2601         * src/pa/linux.S: New file.
2602         * prep_cif.c: Add PA support.
2603
2604 2004-03-16  Hosaka Yuji  <hos@tamanegi.org>
2605
2606         * src/types.c: Fix alignment size of X86_WIN32 case int64 and
2607         double.
2608         * src/x86/ffi.c (ffi_prep_args): Replace ecif->cif->rtype->type
2609         with ecif->cif->flags.
2610         (ffi_call, ffi_prep_incoming_args_SYSV): Replace cif->rtype->type
2611         with cif->flags.
2612         (ffi_prep_cif_machdep): Add X86_WIN32 struct case.
2613         (ffi_closure_SYSV): Add 1 or 2-bytes struct case for X86_WIN32.
2614         * src/x86/win32.S (retstruct1b, retstruct2b, sc_retstruct1b,
2615         sc_retstruct2b): Add for 1 or 2-bytes struct case.
2616
2617 2004-03-15 Kelley Cook <kcook@gcc.gnu.org>
2618
2619         * configure.in: Rename file to ...
2620         * configure.ac: ... this.
2621         * fficonfig.h.in: Regenerate.
2622         * Makefile.in: Regenerate.
2623         * include/Makefile.in: Regenerate.
2624         * testsuite/Makefile.in: Regenerate.
2625
2626 2004-03-12  Matt Austern  <austern@apple.com>
2627
2628         * src/powerpc/darwin.S: Fix EH information so it corresponds to
2629         changes in EH format resulting from addition of linkonce support.
2630         * src/powerpc/darwin_closure.S: Likewise.
2631
2632 2004-03-11  Andreas Tobler  <a.tobler@schweiz.ch>
2633             Paolo Bonzini  <bonzini@gnu.org>
2634
2635         * Makefile.am (AUTOMAKE_OPTIONS): Set them.
2636         Remove VPATH. Remove rules for object files. Remove multilib support.
2637         (AM_CCASFLAGS): Add.
2638         * configure.in (AC_CONFIG_HEADERS): Relace AM_CONFIG_HEADER.
2639         (AC_PREREQ): Bump version to 2.59.
2640         (AC_INIT): Fill with version info and bug address.
2641         (ORIGINAL_LD_FOR_MULTILIBS): Remove.
2642         (AM_ENABLE_MULTILIB): Use this instead of AC_ARG_ENABLE.
2643         De-precious CC so that the right flags are passed down to multilibs.
2644         (AC_MSG_ERROR): Replace obsolete macro AC_ERROR.
2645         (AC_CONFIG_FILES): Replace obsolete macro AC_LINK_FILES.
2646         (AC_OUTPUT): Reorganize the output with AC_CONFIG_COMMANDS.
2647         * configure: Rebuilt.
2648         * aclocal.m4: Likewise.
2649         * Makefile.in, include/Makefile.in, testsuite/Makefile.in: Likewise.
2650         * fficonfig.h.in: Likewise.
2651
2652 2004-03-11  Andreas Schwab  <schwab@suse.de>
2653
2654         * src/ia64/ffi.c (ffi_prep_incoming_args_UNIX): Get floating point
2655         arguments from fp registers only for the first 8 parameter slots.
2656         Don't convert a float parameter when passed in memory.
2657
2658 2004-03-09  Hans-Peter Nilsson  <hp@axis.com>
2659
2660         * configure: Regenerate for config/accross.m4 correction.
2661
2662 2004-02-25  Matt Kraai  <kraai@alumni.cmu.edu>
2663
2664         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Change
2665         ecif->cif->bytes to bytes.
2666         (ffi_prep_cif_machdep): Add braces around nested if statement.
2667
2668 2004-02-09  Alan Modra  <amodra@bigpond.net.au>
2669
2670         * src/types.c (pointer): POWERPC64 has 8 byte pointers.
2671
2672         * src/powerpc/ffi.c (ffi_prep_args64): Correct long double handling.
2673         (ffi_closure_helper_LINUX64): Fix typo.
2674         * testsuite/libffi.call/cls_align_longdouble.c: Pass -mlong-double-128
2675         for powerpc64-*-*.
2676         * testsuite/libffi.call/float.c: Likewise.
2677         * testsuite/libffi.call/float2.c: Likewise.
2678
2679 2004-02-08  Alan Modra  <amodra@bigpond.net.au>
2680
2681         * src/powerpc/ffi.c (ffi_prep_cif_machdep <FFI_LINUX64>): Correct
2682         long double function return and long double arg handling.
2683         (ffi_closure_helper_LINUX64): Formatting.  Delete unused "ng" var.
2684         Use "end_pfr" instead of "nf".  Correct long double handling.
2685         Localise "temp".
2686         * src/powerpc/linux64.S (ffi_call_LINUX64): Save f2 long double
2687         return value.
2688         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Allocate
2689         space for long double return value.  Adjust stack frame and offsets.
2690         Load f2 long double return.
2691
2692 2004-02-07  Alan Modra  <amodra@bigpond.net.au>
2693
2694         * src/types.c: Use 16 byte long double for POWERPC64.
2695
2696 2004-01-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2697
2698         * src/sparc/ffi.c (ffi_prep_args_v9): Shift the parameter array
2699         when the structure return address is passed in %o0.
2700         (ffi_V9_return_struct): Rename into ffi_v9_layout_struct.
2701         (ffi_v9_layout_struct): Align the field following a nested structure
2702         on a word boundary.  Use memmove instead of memcpy.
2703         (ffi_call): Update call to ffi_V9_return_struct.
2704         (ffi_prep_closure): Define 'ctx' only for V8.
2705         (ffi_closure_sparc_inner): Clone into ffi_closure_sparc_inner_v8
2706         and ffi_closure_sparc_inner_v9.
2707         (ffi_closure_sparc_inner_v8): Return long doubles by reference.
2708         Always skip the structure return address.  For structures and long
2709         doubles, copy the argument directly.
2710         (ffi_closure_sparc_inner_v9): Skip the structure return address only
2711         if required.  Shift the maximum floating-point slot accordingly.  For
2712         big structures, copy the argument directly; otherwise, left-justify the
2713         argument and call ffi_v9_layout_struct to lay out the structure on
2714         the stack.
2715         * src/sparc/v8.S: Undef STACKFRAME before defining it.
2716         (ffi_closure_v8): Pass the structure return address.  Update call to
2717         ffi_closure_sparc_inner_v8.  Short-circuit FFI_TYPE_INT handling.
2718         Skip the 'unimp' insn when returning long doubles and structures.
2719         * src/sparc/v9.S: Undef STACKFRAME before defining it.
2720         (ffi_closure_v9): Increase the frame size by 2 words.  Short-circuit
2721         FFI_TYPE_INT handling.  Load structures both in integers and
2722         floating-point registers on return.
2723         * README: Update status of the SPARC port.
2724
2725 2004-01-24  Andreas Tobler  <a.tobler@schweiz.ch>
2726
2727         * testsuite/libffi.call/pyobjc-tc.c (main): Treat result value
2728         as of type ffi_arg.
2729         * testsuite/libffi.call/struct3.c (main): Fix CHECK.
2730
2731 2004-01-22  Ulrich Weigand  <uweigand@de.ibm.com>
2732
2733         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Treat result
2734         value as of type ffi_arg, not unsigned int.
2735
2736 2004-01-21  Michael Ritzert  <ritzert@t-online.de>
2737
2738         * ffi64.c (ffi_prep_args): Cast the RHS of an assignment instead
2739         of the LHS.
2740
2741 2004-01-12  Andreas Tobler  <a.tobler@schweiz.ch>
2742
2743         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_32 for
2744         Solaris.
2745
2746 2004-01-08  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2747
2748         * testsuite/libffi.call/ffitest.h (allocate_mmap): Cast MAP_FAILED
2749         to void *.
2750
2751 2003-12-10  Richard Henderson  <rth@redhat.com>
2752
2753         * testsuite/libffi.call/cls_align_pointer.c: Cast pointers to
2754         size_t instead of int.
2755
2756 2003-12-04  Hosaka Yuji  <hos@tamanegi.org>
2757
2758         * testsuite/libffi.call/many_win32.c: Include <float.h>.
2759         * testsuite/libffi.call/many_win32.c (main): Replace variable
2760         int i with unsigned long ul.
2761
2762         * testsuite/libffi.call/cls_align_uint64.c: New test case.
2763         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2764         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2765         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2766         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2767         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2768         * testsuite/libffi.call/cls_align_float.c: Likewise.
2769         * testsuite/libffi.call/cls_align_double.c: Likewise.
2770         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2771         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2772
2773 2003-12-02  Hosaka Yuji  <hos@tamanegi.org>
2774
2775         PR other/13221
2776         * src/x86/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV):
2777         Align arguments to 32 bits.
2778
2779 2003-12-01  Andreas Tobler  <a.tobler@schweiz.ch>
2780
2781         PR other/13221
2782         * testsuite/libffi.call/cls_multi_sshort.c: New test case.
2783         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2784         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2785         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2786         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2787         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2788
2789         * testsuite/libffi.special/unwindtest.cc: Cosmetics.
2790
2791 2003-11-26  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
2792
2793         * testsuite/libffi.call/ffitest.h: Include <fcntl.h>.
2794         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2795
2796 2003-11-22  Andreas Tobler  <a.tobler@schweiz.ch>
2797
2798         * Makefile.in: Rebuilt.
2799         * configure: Likewise.
2800         * testsuite/libffi.special/unwindtest.cc: Convert the mmap to
2801         the right type.
2802
2803 2003-11-21  Andreas Jaeger  <aj@suse.de>
2804             Andreas Tobler  <a.tobler@schweiz.ch>
2805
2806         * acinclude.m4: Add AC_FUNC_MMAP_BLACKLIST.
2807         * configure.in: Call AC_FUNC_MMAP_BLACKLIST.
2808         * Makefile.in: Rebuilt.
2809         * aclocal.m4: Likewise.
2810         * configure: Likewise.
2811         * fficonfig.h.in: Likewise.
2812         * testsuite/lib/libffi-dg.exp: Add include dir.
2813         * testsuite/libffi.call/ffitest.h: Add MMAP definitions.
2814         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2815         * testsuite/libffi.call/closure_fn0.c: Use MMAP functionality
2816         for ffi_closure if available.
2817         * testsuite/libffi.call/closure_fn1.c: Likewise.
2818         * testsuite/libffi.call/closure_fn2.c: Likewise.
2819         * testsuite/libffi.call/closure_fn3.c: Likewise.
2820         * testsuite/libffi.call/closure_fn4.c: Likewise.
2821         * testsuite/libffi.call/closure_fn5.c: Likewise.
2822         * testsuite/libffi.call/cls_12byte.c: Likewise.
2823         * testsuite/libffi.call/cls_16byte.c: Likewise.
2824         * testsuite/libffi.call/cls_18byte.c: Likewise.
2825         * testsuite/libffi.call/cls_19byte.c: Likewise.
2826         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2827         * testsuite/libffi.call/cls_20byte.c: Likewise.
2828         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2829         * testsuite/libffi.call/cls_24byte.c: Likewise.
2830         * testsuite/libffi.call/cls_2byte.c: Likewise.
2831         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2832         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2833         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2834         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2835         * testsuite/libffi.call/cls_4byte.c: Likewise.
2836         * testsuite/libffi.call/cls_5byte.c: Likewise.
2837         * testsuite/libffi.call/cls_64byte.c: Likewise.
2838         * testsuite/libffi.call/cls_6byte.c: Likewise.
2839         * testsuite/libffi.call/cls_7byte.c: Likewise.
2840         * testsuite/libffi.call/cls_8byte.c: Likewise.
2841         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2842         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2843         * testsuite/libffi.call/cls_double.c: Likewise.
2844         * testsuite/libffi.call/cls_float.c: Likewise.
2845         * testsuite/libffi.call/cls_schar.c: Likewise.
2846         * testsuite/libffi.call/cls_sint.c: Likewise.
2847         * testsuite/libffi.call/cls_sshort.c: Likewise.
2848         * testsuite/libffi.call/cls_uchar.c: Likewise.
2849         * testsuite/libffi.call/cls_uint.c: Likewise.
2850         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2851         * testsuite/libffi.call/cls_ushort.c: Likewise.
2852         * testsuite/libffi.call/nested_struct.c: Likewise.
2853         * testsuite/libffi.call/nested_struct1.c: Likewise.
2854         * testsuite/libffi.call/nested_struct2.c: Likewise.
2855         * testsuite/libffi.call/nested_struct3.c: Likewise.
2856         * testsuite/libffi.call/problem1.c: Likewise.
2857         * testsuite/libffi.special/unwindtest.cc: Likewise.
2858
2859 2003-11-20  Andreas Tobler  <a.tobler@schweiz.ch>
2860
2861         * testsuite/lib/libffi-dg.exp: Make the -lgcc_s conditional.
2862
2863 2003-11-19  Andreas Tobler  <a.tobler@schweiz.ch>
2864
2865         * testsuite/lib/libffi-dg.exp: Add DYLD_LIBRARY_PATH for darwin.
2866         Add -lgcc_s to additional flags.
2867
2868 2003-11-12  Andreas Tobler  <a.tobler@schweiz.ch>
2869
2870         * configure.in, include/Makefile.am: PR libgcj/11147, install
2871         the ffitarget.h header file in a gcc versioned and target
2872         dependent place.
2873         * configure: Regenerated.
2874         * Makefile.in, include/Makefile.in: Likewise.
2875         * testsuite/Makefile.in: Likewise.
2876
2877 2003-11-09  Andreas Tobler  <a.tobler@schweiz.ch>
2878
2879         * testsuite/libffi.call/closure_fn0.c: Print result and check
2880         with dg-output to make debugging easier.
2881         * testsuite/libffi.call/closure_fn1.c: Likewise.
2882         * testsuite/libffi.call/closure_fn2.c: Likewise.
2883         * testsuite/libffi.call/closure_fn3.c: Likewise.
2884         * testsuite/libffi.call/closure_fn4.c: Likewise.
2885         * testsuite/libffi.call/closure_fn5.c: Likewise.
2886         * testsuite/libffi.call/cls_12byte.c: Likewise.
2887         * testsuite/libffi.call/cls_16byte.c: Likewise.
2888         * testsuite/libffi.call/cls_18byte.c: Likewise.
2889         * testsuite/libffi.call/cls_19byte.c: Likewise.
2890         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2891         * testsuite/libffi.call/cls_20byte.c: Likewise.
2892         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2893         * testsuite/libffi.call/cls_24byte.c: Likewise.
2894         * testsuite/libffi.call/cls_2byte.c: Likewise.
2895         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2896         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2897         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2898         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2899         * testsuite/libffi.call/cls_4byte.c: Likewise.
2900         * testsuite/libffi.call/cls_5byte.c: Likewise.
2901         * testsuite/libffi.call/cls_64byte.c: Likewise.
2902         * testsuite/libffi.call/cls_6byte.c: Likewise.
2903         * testsuite/libffi.call/cls_7byte.c: Likewise.
2904         * testsuite/libffi.call/cls_8byte.c: Likewise.
2905         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2906         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2907         * testsuite/libffi.call/cls_double.c: Likewise.
2908         * testsuite/libffi.call/cls_float.c: Likewise.
2909         * testsuite/libffi.call/cls_schar.c: Likewise.
2910         * testsuite/libffi.call/cls_sint.c: Likewise.
2911         * testsuite/libffi.call/cls_sshort.c: Likewise.
2912         * testsuite/libffi.call/cls_uchar.c: Likewise.
2913         * testsuite/libffi.call/cls_uint.c: Likewise.
2914         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2915         * testsuite/libffi.call/cls_ushort.c: Likewise.
2916         * testsuite/libffi.call/problem1.c: Likewise.
2917
2918         * testsuite/libffi.special/unwindtest.cc: Make ffi_closure
2919         static.
2920
2921 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
2922
2923         * testsuite/libffi.call/cls_9byte2.c: New test case.
2924         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2925         * testsuite/libffi.call/cls_64byte.c: Likewise.
2926         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2927         * testsuite/libffi.call/cls_19byte.c: Likewise.
2928         * testsuite/libffi.call/cls_18byte.c: Likewise.
2929         * testsuite/libffi.call/closure_fn4.c: Likewise.
2930         * testsuite/libffi.call/closure_fn5.c: Likewise.
2931         * testsuite/libffi.call/cls_schar.c: Likewise.
2932         * testsuite/libffi.call/cls_sint.c: Likewise.
2933         * testsuite/libffi.call/cls_sshort.c: Likewise.
2934         * testsuite/libffi.call/nested_struct2.c: Likewise.
2935         * testsuite/libffi.call/nested_struct3.c: Likewise.
2936
2937 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
2938
2939         * testsuite/libffi.call/cls_double.c: Do a check on the result.
2940         * testsuite/libffi.call/cls_uchar.c: Likewise.
2941         * testsuite/libffi.call/cls_uint.c: Likewise.
2942         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2943         * testsuite/libffi.call/cls_ushort.c: Likewise.
2944         * testsuite/libffi.call/return_sc.c: Cleanup whitespaces.
2945
2946 2003-11-06  Andreas Tobler  <a.tobler@schweiz.ch>
2947
2948         * src/prep_cif.c (ffi_prep_cif): Move the validity check after
2949         the initialization.
2950
2951 2003-10-23  Andreas Tobler  <a.tobler@schweiz.ch>
2952
2953         * src/java_raw_api.c (ffi_java_ptrarray_to_raw): Replace
2954         FFI_ASSERT(FALSE) with FFI_ASSERT(0).
2955
2956 2003-10-22  David Daney  <ddaney@avtrex.com>
2957
2958         * src/mips/ffitarget.h: Replace undefined UINT32 and friends with
2959         __attribute__((__mode__(__SI__))) and friends.
2960
2961 2003-10-22  Andreas Schwab  <schwab@suse.de>
2962
2963         * src/ia64/ffi.c: Replace FALSE/TRUE with false/true.
2964
2965 2003-10-21  Andreas Tobler  <a.tobler@schweiz.ch>
2966
2967         * configure.in: AC_LINK_FILES(ffitarget.h).
2968         * configure: Regenerate.
2969         * Makefile.in: Likewise.
2970         * include/Makefile.in: Likewise.
2971         * testsuite/Makefile.in: Likewise.
2972         * fficonfig.h.in: Likewise.
2973
2974 2003-10-21  Paolo Bonzini  <bonzini@gnu.org>
2975             Richard Henderson  <rth@redhat.com>
2976
2977         Avoid that ffi.h includes fficonfig.h.
2978
2979         * Makefile.am (EXTRA_DIST): Include ffitarget.h files
2980         (TARGET_SRC_MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
2981         (TARGET_SRC_MIPS_SGI): Removed.
2982         (MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
2983         (MIPS_SGI): Removed.
2984         (CLEANFILES): Removed.
2985         (mostlyclean-am, clean-am, mostlyclean-sub, clean-sub): New
2986         targets.
2987         * acconfig.h: Removed.
2988         * configure.in: Compute sizeofs only for double and long double.
2989         Use them to define and subst HAVE_LONG_DOUBLE.  Include comments
2990         into AC_DEFINE instead of using acconfig.h.  Create
2991         include/ffitarget.h instead of include/fficonfig.h.  Rename
2992         MIPS_GCC to MIPS_IRIX, drop MIPS_SGI since we are in gcc's tree.
2993         AC_DEFINE EH_FRAME_FLAGS.
2994         * include/Makefile.am (DISTCLEANFILES): New automake macro.
2995         (hack_DATA): Add ffitarget.h.
2996         * include/ffi.h.in: Remove all system specific definitions.
2997         Declare raw API even if it is not installed, why bother?
2998         Use limits.h instead of SIZEOF_* to define ffi_type_*.  Do
2999         not define EH_FRAME_FLAGS, it is in fficonfig.h now.  Include
3000         ffitarget.h instead of fficonfig.h.  Remove ALIGN macro.
3001         (UINT_ARG, INT_ARG): Removed, use ffi_arg and ffi_sarg instead.
3002         * include/ffi_common.h (bool): Do not define.
3003         (ffi_assert): Accept failed assertion.
3004         (ffi_type_test): Return void and accept file/line.
3005         (FFI_ASSERT): Pass stringized failed assertion.
3006         (FFI_ASSERT_AT): New macro.
3007         (FFI_ASSERT_VALID_TYPE): New macro.
3008         (UINT8, SINT8, UINT16, SINT16, UINT32, SINT32,
3009         UINT64, SINT64): Define here with gcc's __attribute__ macro
3010         instead of in ffi.h
3011         (FLOAT32, ALIGN): Define here instead of in ffi.h
3012         * include/ffi-mips.h: Removed.  Its content moved to
3013         src/mips/ffitarget.h after separating assembly and C sections.
3014         * src/alpha/ffi.c, src/alpha/ffi.c, src/java_raw_api.c
3015         src/prep_cif.c, src/raw_api.c, src/ia64/ffi.c,
3016         src/mips/ffi.c, src/mips/n32.S, src/mips/o32.S,
3017         src/mips/ffitarget.h, src/sparc/ffi.c, src/x86/ffi64.c:
3018         SIZEOF_ARG -> FFI_SIZEOF_ARG.
3019         * src/ia64/ffi.c: Include stdbool.h (provided by GCC 2.95+).
3020         * src/debug.c (ffi_assert): Accept stringized failed assertion.
3021         (ffi_type_test): Rewritten.
3022         * src/prep-cif.c (initialize_aggregate, ffi_prep_cif): Call
3023         FFI_ASSERT_VALID_TYPE.
3024         * src/alpha/ffitarget.h, src/arm/ffitarget.h,
3025         src/ia64/ffitarget.h, src/m68k/ffitarget.h,
3026         src/mips/ffitarget.h, src/powerpc/ffitarget.h,
3027         src/s390/ffitarget.h, src/sh/ffitarget.h,
3028         src/sh64/ffitarget.h, src/sparc/ffitarget.h,
3029         src/x86/ffitarget.h: New files.
3030         * src/alpha/osf.S, src/arm/sysv.S, src/ia64/unix.S,
3031         src/m68k/sysv.S, src/mips/n32.S, src/mips/o32.S,
3032         src/powerpc/aix.S, src/powerpc/darwin.S,
3033         src/powerpc/ffi_darwin.c, src/powerpc/linux64.S,
3034         src/powerpc/linux64_closure.S, src/powerpc/ppc_closure.S,
3035         src/powerpc/sysv.S, src/s390/sysv.S, src/sh/sysv.S,
3036         src/sh64/sysv.S, src/sparc/v8.S, src/sparc/v9.S,
3037         src/x86/sysv.S, src/x86/unix64.S, src/x86/win32.S:
3038         include fficonfig.h
3039
3040 2003-10-20  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3041
3042         * src/mips/ffi.c: Use _ABIN32, _ABIO32 instead of external
3043         _MIPS_SIM_NABI32, _MIPS_SIM_ABI32.
3044
3045 2003-10-19  Andreas Tobler  <a.tobler@schweiz.ch>
3046
3047         * src/powerpc/ffi_darwin.c (ffi_prep_args): Declare bytes again.
3048         Used when FFI_DEBUG = 1.
3049
3050 2003-10-14  Alan Modra  <amodra@bigpond.net.au>
3051
3052         * src/types.c (double, longdouble): Default POWERPC64 to 8 byte size
3053         and align.
3054
3055 2003-10-06  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3056
3057         * include/ffi_mips.h: Define FFI_MIPS_N32 for N32/N64 ABIs,
3058         FFI_MIPS_O32 for O32 ABI.
3059
3060 2003-10-01  Andreas Tobler  <a.tobler@schweiz.ch>
3061
3062         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_64 for
3063         SPARC64. Cleanup whitespaces.
3064
3065 2003-09-19  Andreas Tobler  <a.tobler@schweiz.ch>
3066
3067         * testsuite/libffi.call/closure_fn0.c: Xfail mips, arm,
3068         strongarm, xscale. Cleanup whitespaces.
3069         * testsuite/libffi.call/closure_fn1.c: Likewise.
3070         * testsuite/libffi.call/closure_fn2.c: Likewise.
3071         * testsuite/libffi.call/closure_fn3.c: Likewise.
3072         * testsuite/libffi.call/cls_12byte.c: Likewise.
3073         * testsuite/libffi.call/cls_16byte.c: Likewise.
3074         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3075         * testsuite/libffi.call/cls_20byte.c: Likewise.
3076         * testsuite/libffi.call/cls_24byte.c: Likewise.
3077         * testsuite/libffi.call/cls_2byte.c: Likewise.
3078         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3079         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3080         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3081         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3082         * testsuite/libffi.call/cls_4byte.c: Likewise.
3083         * testsuite/libffi.call/cls_5byte.c: Likewise.
3084         * testsuite/libffi.call/cls_6byte.c: Likewise.
3085         * testsuite/libffi.call/cls_7byte.c: Likewise.
3086         * testsuite/libffi.call/cls_8byte.c: Likewise.
3087         * testsuite/libffi.call/cls_double.c: Likewise.
3088         * testsuite/libffi.call/cls_float.c: Likewise.
3089         * testsuite/libffi.call/cls_uchar.c: Likewise.
3090         * testsuite/libffi.call/cls_uint.c: Likewise.
3091         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3092         * testsuite/libffi.call/cls_ushort.c: Likewise.
3093         * testsuite/libffi.call/nested_struct.c: Likewise.
3094         * testsuite/libffi.call/nested_struct1.c: Likewise.
3095         * testsuite/libffi.call/problem1.c: Likewise.
3096         * testsuite/libffi.special/unwindtest.cc: Likewise.
3097         * testsuite/libffi.call/pyobjc-tc.c: Cleanup whitespaces.
3098
3099 2003-09-18  David Edelsohn  <edelsohn@gnu.org>
3100
3101         * src/powerpc/aix.S: Cleanup whitespaces.
3102         * src/powerpc/aix_closure.S: Likewise.
3103
3104 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3105
3106         * src/powerpc/darwin.S: Cleanup whitespaces, comment formatting.
3107         * src/powerpc/darwin_closure.S: Likewise.
3108         * src/powerpc/ffi_darwin.c: Likewise.
3109
3110 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3111             David Edelsohn  <edelsohn@gnu.org>
3112
3113         * src/types.c (double): Add AIX and Darwin to the right TYPEDEF.
3114         * src/powerpc/aix_closure.S: Remove the pointer to the outgoing
3115         parameter stack.
3116         * src/powerpc/darwin_closure.S: Likewise.
3117         * src/powerpc/ffi_darwin.c (ffi_prep_args): Handle structures
3118         according to the Darwin/AIX ABI.
3119         (ffi_prep_cif_machdep): Likewise.
3120         (ffi_closure_helper_DARWIN): Likewise.
3121         Remove the outgoing parameter stack logic. Simplify the evaluation
3122         of the different CASE types.
3123         (ffi_prep_clousure): Avoid the casts on lvalues. Change the branch
3124         statement in the trampoline code.
3125
3126 2003-09-18  Kaz Kojima  <kkojima@gcc.gnu.org>
3127
3128         * src/sh/ffi.c (ffi_prep_args): Take account into the alignement
3129         for the register size.
3130         (ffi_closure_helper_SYSV): Handle the structure return value
3131         address correctly.
3132         (ffi_closure_helper_SYSV): Return the appropriate type when
3133         the registers are used for the structure return value.
3134         * src/sh/sysv.S (ffi_closure_SYSV): Fix the stack layout for
3135         the 64-bit return value.  Update copyright years.
3136
3137 2003-09-17  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3138
3139         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Search in
3140         srcdir for ffi_mips.h.
3141
3142 2003-09-12  Alan Modra  <amodra@bigpond.net.au>
3143
3144         * src/prep_cif.c (initialize_aggregate): Include tail padding in
3145         structure size.
3146         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Correct
3147         placement of float result.
3148         * testsuite/libffi.special/unwindtest.cc (closure_test_fn1): Correct
3149         cast of "resp" for big-endian 64 bit machines.
3150
3151 2003-09-11  Alan Modra  <amodra@bigpond.net.au>
3152
3153         * src/types.c (double, longdouble): Merge identical SH and ARM
3154         typedefs, and add POWERPC64.
3155         * src/powerpc/ffi.c (ffi_prep_args64): Correct next_arg calc for
3156         struct split over gpr and rest.
3157         (ffi_prep_cif_machdep): Correct intarg_count for structures.
3158         * src/powerpc/linux64.S (ffi_call_LINUX64): Fix gpr offsets.
3159
3160 2003-09-09  Andreas Tobler  <a.tobler@schweiz.ch>
3161
3162         * src/powerpc/ffi.c (ffi_closure_helper_SYSV) Handle struct
3163         passing correctly.
3164
3165 2003-09-09  Alan Modra  <amodra@bigpond.net.au>
3166
3167         * configure: Regenerate.
3168
3169 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3170
3171         * Makefile.am: Remove build rules for ffitest.
3172         * Makefile.in: Rebuilt.
3173
3174 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3175
3176         * src/java_raw_api.c: Include <stdlib.h> to fix compiler warning
3177         about implicit declaration of abort().
3178
3179 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3180
3181         * Makefile.am: Add dejagnu test framework. Fixes PR other/11411.
3182         * Makefile.in: Rebuilt.
3183         * configure.in: Add dejagnu test framework.
3184         * configure: Rebuilt.
3185
3186         * testsuite/Makefile.am: New file.
3187         * testsuite/Makefile.in: Built
3188         * testsuite/lib/libffi-dg.exp: New file.
3189         * testsuite/config/default.exp: Likewise.
3190         * testsuite/libffi.call/call.exp: Likewise.
3191         * testsuite/libffi.call/ffitest.h: Likewise.
3192         * testsuite/libffi.call/closure_fn0.c: Likewise.
3193         * testsuite/libffi.call/closure_fn1.c: Likewise.
3194         * testsuite/libffi.call/closure_fn2.c: Likewise.
3195         * testsuite/libffi.call/closure_fn3.c: Likewise.
3196         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3197         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3198         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3199         * testsuite/libffi.call/cls_2byte.c: Likewise.
3200         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3201         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3202         * testsuite/libffi.call/cls_4byte.c: Likewise.
3203         * testsuite/libffi.call/cls_5byte.c: Likewise.
3204         * testsuite/libffi.call/cls_6byte.c: Likewise.
3205         * testsuite/libffi.call/cls_7byte.c: Likewise.
3206         * testsuite/libffi.call/cls_8byte.c: Likewise.
3207         * testsuite/libffi.call/cls_12byte.c: Likewise.
3208         * testsuite/libffi.call/cls_16byte.c: Likewise.
3209         * testsuite/libffi.call/cls_20byte.c: Likewise.
3210         * testsuite/libffi.call/cls_24byte.c: Likewise.
3211         * testsuite/libffi.call/cls_double.c: Likewise.
3212         * testsuite/libffi.call/cls_float.c: Likewise.
3213         * testsuite/libffi.call/cls_uchar.c: Likewise.
3214         * testsuite/libffi.call/cls_uint.c: Likewise.
3215         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3216         * testsuite/libffi.call/cls_ushort.c: Likewise.
3217         * testsuite/libffi.call/float.c: Likewise.
3218         * testsuite/libffi.call/float1.c: Likewise.
3219         * testsuite/libffi.call/float2.c: Likewise.
3220         * testsuite/libffi.call/many.c: Likewise.
3221         * testsuite/libffi.call/many_win32.c: Likewise.
3222         * testsuite/libffi.call/nested_struct.c: Likewise.
3223         * testsuite/libffi.call/nested_struct1.c: Likewise.
3224         * testsuite/libffi.call/pyobjc-tc.c: Likewise.
3225         * testsuite/libffi.call/problem1.c: Likewise.
3226         * testsuite/libffi.call/promotion.c: Likewise.
3227         * testsuite/libffi.call/return_ll.c: Likewise.
3228         * testsuite/libffi.call/return_sc.c: Likewise.
3229         * testsuite/libffi.call/return_uc.c: Likewise.
3230         * testsuite/libffi.call/strlen.c: Likewise.
3231         * testsuite/libffi.call/strlen_win32.c: Likewise.
3232         * testsuite/libffi.call/struct1.c: Likewise.
3233         * testsuite/libffi.call/struct2.c: Likewise.
3234         * testsuite/libffi.call/struct3.c: Likewise.
3235         * testsuite/libffi.call/struct4.c: Likewise.
3236         * testsuite/libffi.call/struct5.c: Likewise.
3237         * testsuite/libffi.call/struct6.c: Likewise.
3238         * testsuite/libffi.call/struct7.c: Likewise.
3239         * testsuite/libffi.call/struct8.c: Likewise.
3240         * testsuite/libffi.call/struct9.c: Likewise.
3241         * testsuite/libffi.special/special.exp: New file.
3242         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3243         * testsuite/libffi.special/unwindtest.cc: Likewise.
3244
3245
3246 2003-08-13  Kaz Kojima  <kkojima@gcc.gnu.org>
3247
3248         * src/sh/ffi.c (OFS_INT16): Set 0 for little endian case.  Update
3249         copyright years.
3250
3251 2003-08-02  Alan Modra  <amodra@bigpond.net.au>
3252
3253         * src/powerpc/ffi.c (ffi_prep_args64): Modify for changed gcc
3254         structure passing.
3255         (ffi_closure_helper_LINUX64): Likewise.
3256         * src/powerpc/linux64.S: Remove code writing to parm save area.
3257         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Use return
3258         address in lr from ffi_closure_helper_LINUX64 call to calculate
3259         table address.  Optimize function tail.
3260
3261 2003-07-28  Andreas Tobler  <a.tobler@schweiz.ch>
3262
3263         * src/sparc/ffi.c: Handle all floating point registers.
3264         * src/sparc/v9.S: Likewise. Fixes second part of PR target/11410.
3265
3266 2003-07-11  Gerald Pfeifer  <pfeifer@dbai.tuwien.ac.at>
3267
3268         * README: Note that libffi is not part of GCC.  Update the project
3269         URL and status.
3270
3271 2003-06-19  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3272
3273         * src/powerpc/ppc_closure.S: Include ffi.h.
3274
3275 2003-06-13  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3276
3277         * src/x86/sysv.S: Avoid gas-only .uleb128/.sleb128 directives.
3278         Use C style comments.
3279
3280 2003-06-13  Kaz Kojima  <kkojima@rr.iij4u.or.jp>
3281
3282         * Makefile.am: Add SHmedia support.  Fix a typo of SH support.
3283         * Makefile.in: Regenerate.
3284         * configure.in (sh64-*-linux*, sh5*-*-linux*): Add target.
3285         * configure: Regenerate.
3286         * include/ffi.h.in: Add SHmedia support.
3287         * src/sh64/ffi.c: New file.
3288         * src/sh64/sysv.S: New file.
3289
3290 2003-05-16  Jakub Jelinek  <jakub@redhat.com>
3291
3292         * configure.in (HAVE_RO_EH_FRAME): Check whether .eh_frame section
3293         should be read-only.
3294         * configure: Rebuilt.
3295         * fficonfig.h.in: Rebuilt.
3296         * include/ffi.h.in (EH_FRAME_FLAGS): Define.
3297         * src/alpha/osf.S: Use EH_FRAME_FLAGS.
3298         * src/powerpc/linux64.S: Likewise.
3299         * src/powerpc/linux64_closure.S: Likewise.  Include ffi.h.
3300         * src/powerpc/sysv.S: Use EH_FRAME_FLAGS.  Use pcrel encoding
3301         if -fpic/-fPIC/-mrelocatable.
3302         * src/powerpc/powerpc_closure.S: Likewise.
3303         * src/sparc/v8.S: If HAVE_RO_EH_FRAME is defined, don't include
3304         #write in .eh_frame flags.
3305         * src/sparc/v9.S: Likewise.
3306         * src/x86/unix64.S: Use EH_FRAME_FLAGS.
3307         * src/x86/sysv.S: Likewise.  Use pcrel encoding if -fpic/-fPIC.
3308         * src/s390/sysv.S: Use EH_FRAME_FLAGS.  Include ffi.h.
3309
3310 2003-05-07  Jeff Sturm  <jsturm@one-point.com>
3311
3312         Fixes PR bootstrap/10656
3313         * configure.in (HAVE_AS_REGISTER_PSEUDO_OP): Test assembler
3314         support for .register pseudo-op.
3315         * src/sparc/v8.S: Use it.
3316         * fficonfig.h.in: Rebuilt.
3317         * configure: Rebuilt.
3318
3319 2003-04-18  Jakub Jelinek  <jakub@redhat.com>
3320
3321         * include/ffi.h.in (POWERPC64): Define if 64-bit.
3322         (enum ffi_abi): Add FFI_LINUX64 on POWERPC.
3323         Make it the default on POWERPC64.
3324         (FFI_TRAMPOLINE_SIZE): Define to 24 on POWERPC64.
3325         * configure.in: Change powerpc-*-linux* into powerpc*-*-linux*.
3326         * configure: Rebuilt.
3327         * src/powerpc/ffi.c (hidden): Define.
3328         (ffi_prep_args_SYSV): Renamed from
3329         ffi_prep_args.  Cast pointers to unsigned long to shut up warnings.
3330         (NUM_GPR_ARG_REGISTERS64, NUM_FPR_ARG_REGISTERS64,
3331         ASM_NEEDS_REGISTERS64): New.
3332         (ffi_prep_args64): New function.
3333         (ffi_prep_cif_machdep): Handle FFI_LINUX64 ABI.
3334         (ffi_call): Likewise.
3335         (ffi_prep_closure): Likewise.
3336         (flush_icache): Surround by #ifndef POWERPC64.
3337         (ffi_dblfl): New union type.
3338         (ffi_closure_helper_SYSV): Use it to avoid aliasing problems.
3339         (ffi_closure_helper_LINUX64): New function.
3340         * src/powerpc/ppc_closure.S: Surround whole file by #ifndef
3341         __powerpc64__.
3342         * src/powerpc/sysv.S: Likewise.
3343         (ffi_call_SYSV): Rename ffi_prep_args to ffi_prep_args_SYSV.
3344         * src/powerpc/linux64.S: New file.
3345         * src/powerpc/linux64_closure.S: New file.
3346         * Makefile.am (EXTRA_DIST): Add src/powerpc/linux64.S and
3347         src/powerpc/linux64_closure.S.
3348         (TARGET_SRC_POWERPC): Likewise.
3349
3350         * src/ffitest.c (closure_test_fn, closure_test_fn1, closure_test_fn2,
3351         closure_test_fn3): Fix result printing on big-endian 64-bit
3352         machines.
3353         (main): Print tst2_arg instead of uninitialized tst2_result.
3354
3355         * src/ffitest.c (main): Hide what closure pointer really points to
3356         from the compiler.
3357
3358 2003-04-16  Richard Earnshaw  <rearnsha@arm.com>
3359
3360         * configure.in (arm-*-netbsdelf*): Add configuration.
3361         (configure): Regenerated.
3362
3363 2003-04-04  Loren J. Rittle  <ljrittle@acm.org>
3364
3365         * include/Makefile.in: Regenerate.
3366
3367 2003-03-21  Zdenek Dvorak  <rakdver@atrey.karlin.mff.cuni.cz>
3368
3369         * libffi/include/ffi.h.in: Define X86 instead of X86_64 in 32
3370         bit mode.
3371         * libffi/src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV):
3372         Receive closure pointer through parameter, read args using
3373         __builtin_dwarf_cfa.
3374         (FFI_INIT_TRAMPOLINE): Send closure reference through eax.
3375
3376 2003-03-12  Andreas Schwab  <schwab@suse.de>
3377
3378         * configure.in: Avoid trailing /. in toolexeclibdir.
3379         * configure: Rebuilt.
3380
3381 2003-03-03  Andreas Tobler <a.tobler@schweiz.ch>
3382
3383         * src/powerpc/darwin_closure.S: Recode to fit dynamic libraries.
3384
3385 2003-02-06  Andreas Tobler <a.tobler@schweiz.ch>
3386
3387         * libffi/src/powerpc/darwin_closure.S:
3388         Fix alignement bug, allocate 8 bytes for the result.
3389         * libffi/src/powerpc/aix_closure.S:
3390         Likewise.
3391         * libffi/src/powerpc/ffi_darwin.c:
3392         Update stackframe description for aix/darwin_closure.S.
3393
3394 2003-02-06  Jakub Jelinek  <jakub@redhat.com>
3395
3396         * src/s390/ffi.c (ffi_closure_helper_SYSV): Add hidden visibility
3397         attribute.
3398
3399 2003-01-31  Christian Cornelssen  <ccorn@cs.tu-berlin.de>,
3400             Andreas Schwab  <schwab@suse.de>
3401
3402         * configure.in: Adjust command to source config-ml.in to account
3403         for changes to the libffi_basedir definition.
3404         (libffi_basedir): Remove ${srcdir} from value and include trailing
3405         slash if nonempty.
3406
3407         * configure: Regenerate.
3408
3409 2003-01-29  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3410
3411         * src/powerpc/ppc_closure.S: Recode to fit shared libs.
3412
3413 2003-01-28  Andrew Haley  <aph@redhat.com>
3414
3415         * include/ffi.h.in: Enable FFI_CLOSURES for x86_64.
3416         * src/x86/ffi64.c (ffi_prep_closure): New.
3417         (ffi_closure_UNIX64_inner): New.
3418         * src/x86/unix64.S (ffi_closure_UNIX64): New.
3419
3420 2003-01-27  Alexandre Oliva  <aoliva@redhat.com>
3421
3422         * configure.in (toolexecdir, toolexeclibdir): Set and AC_SUBST.
3423         Remove USE_LIBDIR conditional.
3424         * Makefile.am (toolexecdir, toolexeclibdir): Don't override.
3425         * Makefile.in, configure: Rebuilt.
3426
3427 2003-01027  David Edelsohn  <edelsohn@gnu.org>
3428
3429         * Makefile.am (TARGET_SRC_POWERPC_AIX): Fix typo.
3430         * Makefile.in: Regenerate.
3431
3432 2003-01-22  Andrew Haley  <aph@redhat.com>
3433
3434         * src/powerpc/darwin.S (_ffi_call_AIX): Add Augmentation size to
3435         unwind info.
3436
3437 2003-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
3438
3439         * src/powerpc/darwin.S: Add unwind info.
3440         * src/powerpc/darwin_closure.S: Likewise.
3441
3442 2003-01-14  Andrew Haley  <aph@redhat.com>