OSDN Git Service

01a71a8cd5cd31c893c9bff82cd32e935eb383d7
[pf3gnuchains/gcc-fork.git] / libffi / ChangeLog
1 2010-05-19  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
2
3         * configure.ac (libffi_cv_as_x86_pcrel): Check for illegal in as
4         output, too.
5         (libffi_cv_as_ascii_pseudo_op): Check for .ascii.
6         (libffi_cv_as_string_pseudo_op): Check for .string.
7         * configure: Regenerate.
8         * fficonfig.h.in: Regenerate.
9         * src/x86/sysv.S (.eh_frame): Use .ascii, .string or error.
10
11 2010-04-07  Jakub Jelinek  <jakub@redhat.com>
12
13         * regex.c (byte_re_match_2_internal): Avoid set but not used
14         warning.
15
16 2010-04-02  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
17
18         * Makefile.in: Regenerate.
19         * aclocal.m4: Regenerate.
20         * include/Makefile.in: Regenerate.
21         * man/Makefile.in: Regenerate.
22         * testsuite/Makefile.in: Regenerate.
23
24 2010-03-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
25
26         * configure.ac (libffi_cv_as_x86_64_unwind_section_type): New test.
27         * configure: Regenerate.
28         * fficonfig.h.in: Regenerate.
29         * libffi/src/x86/unix64.S (.eh_frame)
30         [HAVE_AS_X86_64_UNWIND_SECTION_TYPE]: Use @unwind section type.
31
32 2010-02-24  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
33
34         * doc/libffi.texi (The Closure API): Fix typo.
35         * doc/libffi.info: Remove.
36
37 2010-02-15  Matthias Klose  <doko@ubuntu.com>
38
39         * src/arm/sysv.S (__ARM_ARCH__): Define for processor
40         __ARM_ARCH_7EM__.
41
42 2010-01-07  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
43
44         PR libffi/40701
45         * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRIdLL,
46         PRIuLL, PRId64, PRIu64, PRIuPTR): Define.
47         * testsuite/libffi.call/cls_align_sint64.c: Add -Wno-format on
48         alpha*-dec-osf*.
49         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
50         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
51         * testsuite/libffi.call/return_ll1.c: Likewise.
52         * testsuite/libffi.call/stret_medium2.c: Likewise.
53         * testsuite/libffi.special/ffitestcxx.h (allocate_mmap): Cast
54         MAP_FAILED to char *.
55
56 2010-01-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
57
58         * src/mips/n32.S: Use .abicalls and .eh_frame with __GNUC__.
59
60 2009-12-31  Anthony Green  <green@redhat.com>
61
62         * README: Update for libffi 3.0.9.
63
64 2009-12-27  Matthias Klose  <doko@ubuntu.com>
65
66         * configure.ac (HAVE_LONG_DOUBLE): Define for mips when
67         appropriate.
68         * configure: Rebuilt.
69
70 2009-12-26  Anthony Green  <green@redhat.com>
71
72         * testsuite/libffi.call/cls_longdouble_va.c: Mark as xfail for
73         avr32*-*-*.
74         * testsuite/libffi.call/cls_double_va.c: Ditto.
75
76 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
77
78         * testsuite/libffi.call/ffitest.h: Conditionally include stdint.h
79         and inttypes.h.
80         * testsuite/libffi.special/unwindtest.cc: Ditto.
81
82 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
83
84         * configure.ac: Add amd64-*-openbsd*.
85         * configure: Rebuilt.
86         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Link
87         openbsd programs with -lpthread.
88
89 2009-12-26  Anthony Green  <green@redhat.com>
90
91         * testsuite/libffi.call/cls_double_va.c,
92         testsuite/libffi.call/cls_longdouble.c,
93         testsuite/libffi.call/cls_longdouble_va.c,
94         testsuite/libffi.call/cls_pointer.c,
95         testsuite/libffi.call/cls_pointer_stack.c: Remove xfail for
96         mips*-*-* and arm*-*-*.
97         * testsuite/libffi.call/cls_align_longdouble_split.c,
98         testsuite/libffi.call/cls_align_longdouble_split2.c,
99         testsuite/libffi.call/stret_medium2.c,
100         testsuite/libffi.call/stret_medium.c,
101         testsuite/libffi.call/stret_large.c,
102         testsuite/libffi.call/stret_large2.c: Remove xfail for arm*-*-*.
103
104 2009-12-31  Kay Tietz  <ktietz70@googlemail.com>
105
106         * testsuite/libffi.call/ffitest.h,
107         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRuLL): Fix
108         definitions.
109
110 2009-12-31  Carlo Bramini  <carlo.bramix@libero.it>
111
112         * configure.ac (AM_LTLDFLAGS): Define for windows hosts.
113         * Makefile.am (libffi_la_LDFLAGS): Add AM_LTLDFLAGS.
114         * configure: Rebuilt.
115         * Makefile.in: Rebuilt.
116
117 2009-12-31  Anthony Green  <green@redhat.com>
118             Blake Chaffin.
119
120         * testsuite/libffi.call/huge_struct.c: New test case from Blake
121         Chaffin @ Apple.
122
123 2009-12-28  David Edelsohn  <edelsohn@gnu.org>
124
125         * src/powerpc/ffi_darwin.c (ffi_prep_args): Copy abi and nargs to
126         local variables.
127         (aix_adjust_aggregate_sizes): New function.
128         (ffi_prep_cif_machdep): Call it.
129
130 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
131
132         * configure.ac: Define FFI_MMAP_EXEC_WRIT for the given targets.
133         * configure: Regenerate.
134         * fficonfig.h.in: Likewise.
135         * src/closures.c: Remove the FFI_MMAP_EXEC_WRIT definition for
136         Solaris/x86.
137
138 2009-12-26  Andreas Schwab  <schwab@linux-m68k.org>
139
140         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Advance intarg_count
141         when a float arguments is passed in memory.
142         (ffi_closure_helper_SYSV): Mark general registers as used up when
143         a 64bit or soft-float long double argument is passed in memory.
144
145 2009-12-25  Matthias Klose  <doko@ubuntu.com>
146
147         * man/ffi_call.3: Fix #include in examples.
148         * doc/libffi.texi: Add dircategory.
149
150 2009-12-25  Frank Everdij <f.p.x.everdij@tudelft.nl>
151
152         * include/ffi.h.in: Placed '__GNUC__' ifdef around
153         '__attribute__((aligned(8)))' in ffi_closure, fixes compile for
154         IRIX MIPSPro c99.
155         * include/ffi_common.h: Added '__sgi' define to non
156         '__attribute__((__mode__()))' integer typedefs.
157         * src/mips/ffi.c (ffi_call, ffi_closure_mips_inner_O32,
158         ffi_closure_mips_inner_N32): Added 'defined(_MIPSEB)' to BE check.
159         (ffi_closure_mips_inner_O32, ffi_closure_mips_inner_N32): Added
160         FFI_LONGDOUBLE support and alignment(N32 only).
161         * src/mips/ffitarget.h: Corrected '#include <sgidefs.h>' for IRIX and
162         fixed non '__attribute__((__mode__()))' integer typedefs.
163         * src/mips/n32.S: Put '#ifdef linux' around '.abicalls' and '.eh_frame'
164         since they are Linux/GNU Assembler specific.
165
166 2009-12-25  Bradley Smith  <brad@brad-smith.co.uk>
167
168         * configure.ac, Makefile.am, src/avr32/ffi.c,
169         src/avr32/ffitarget.h,
170         src/avr32/sysv.S: Add AVR32 port.
171         * configure, Makefile.in: Rebuilt.
172
173 2009-12-21  Andreas Tobler  <a.tobler@schweiz.org>
174
175         * configure.ac: Make i?86 build on FreeBSD and OpenBSD.
176         * configure: Regenerate.
177
178 2009-12-15  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
179
180         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on PA HP-UX.
181
182 2009-12-13  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
183
184         * src/pa/ffi.c (ffi_closure_inner_pa32): Handle FFI_TYPE_LONGDOUBLE
185         type on HP-UX.
186
187 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
188
189         * src/sparc/ffi.c (ffi_closure_sparc_inner_v9): Properly align 'long
190         double' arguments.
191
192 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
193
194         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on Solaris < 10.
195
196 2009-12-10  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
197
198         PR libffi/40700
199         * src/closures.c [X86_64 && __sun__ && __svr4__]
200         (FFI_MMAP_EXEC_WRIT): Define.
201
202 2009-12-08  David Daney  <ddaney@caviumnetworks.com>
203
204         * testsuite/libffi.call/stret_medium.c: Remove xfail for mips*-*-*
205         * testsuite/libffi.call/cls_align_longdouble_split2.c: Same.
206         * testsuite/libffi.call/stret_large.c: Same.
207         * testsuite/libffi.call/cls_align_longdouble_split.c: Same.
208         * testsuite/libffi.call/stret_large2.c: Same.
209         * testsuite/libffi.call/stret_medium2.c: Same.
210
211 2009-12-07  David Edelsohn  <edelsohn@gnu.org>
212
213         * src/powerpc/aix_closure.S (libffi_closure_ASM): Fix tablejump
214         typo.
215
216 2009-12-05  David Edelsohn  <edelsohn@gnu.org>
217
218         * src/powerpc/aix.S: Update AIX32 code to be consistent with AIX64
219         code.
220         * src/powerpc/aix_closure.S: Same.
221
222 2009-12-05  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
223
224         * Makefile.in: Regenerate.
225         * configure: Regenerate.
226         * include/Makefile.in: Regenerate.
227         * man/Makefile.in: Regenerate.
228         * testsuite/Makefile.in: Regenerate.
229
230 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
231
232         * src/powerpc/aix_closure.S: Reorganize 64-bit code to match
233         linux64_closure.S.
234
235 2009-12-04  Uros Bizjak  <ubizjak@gmail.com>
236
237         PR libffi/41908
238         * src/x86/ffi64.c (classify_argument): Update from
239         gcc/config/i386/i386.c.
240         (ffi_closure_unix64_inner): Do not use the address of two consecutive
241         SSE registers directly.
242         * testsuite/libffi.call/cls_dbls_struct.c (main): Remove xfail
243         for x86_64 linux targets.
244
245 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
246
247         * src/powerpc/ffi_darwin.c (ffi_closure_helper_DARWIN): Increment
248         pfr for long double split between fpr13 and stack.
249
250 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
251
252         * src/powerpc/ffi_darwin.c (ffi_prep_args): Increment next_arg and
253         fparg_count twice for long double.
254
255 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
256
257         PR libffi/42243
258         * src/powerpc/ffi_darwin.c (ffi_prep_args): Remove extra parentheses.
259
260 2009-12-03  Uros Bizjak  <ubizjak@gmail.com>
261
262         * testsuite/libffi.call/cls_longdouble_va.c (main): Fix format string.
263         Remove xfails for x86 linux targets.
264
265 2009-12-02  David Edelsohn  <edelsohn@gnu.org>
266
267         * src/powerpc/ffi_darwin.c (ffi_prep_args): Fix typo in INT64
268         case.
269
270 2009-12-01  David Edelsohn  <edelsohn@gnu.org>
271
272         * src/powerpc/aix.S (ffi_call_AIX): Convert to more standard
273         register usage.  Call ffi_prep_args directly.  Add long double
274         return value support.
275         * src/powerpc/ffi_darwin.c (ffi_prep_args): Double arg increment
276         applies to FFI_TYPE_DOUBLE.  Correct fpr_base increment typo.
277         Separate FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases.
278         (ffi_prep_cif_machdep): Only 16 byte stack alignment in 64 bit
279         mode.
280         (ffi_closure_helper_DARWIN): Remove nf and ng counters.  Move temp
281         into case.
282         * src/powerpc/aix_closure.S: Maintain 16 byte stack alignment.
283         Allocate result area between params and FPRs.
284
285 2009-11-30  David Edelsohn  <edelsohn@gnu.org>
286
287         PR target/35484
288         * src/powerpc/ffitarget.h (POWERPC64): Define for PPC64 Linux and
289         AIX64.
290         * src/powerpc/aix.S: Implement AIX64 version.
291         * src/powerpc/aix_closure.S: Implement AIX64 version.
292         (ffi_closure_ASM): Use extsb, lha and displament addresses.
293         * src/powerpc/ffi_darwin.c (ffi_prep_args): Implement AIX64
294         support.
295         (ffi_prep_cif_machdep): Same.
296         (ffi_call): Same.
297         (ffi_closure_helper_DARWIN): Same.
298
299 2009-11-02  Andreas Tobler  <a.tobler@schweiz.org>
300
301         PR libffi/41908
302         * testsuite/libffi.call/testclosure.c: New test.
303
304 2009-09-28  Kai Tietz  <kai.tietz@onevision.com>
305
306         * src/x86/win64.S (_ffi_call_win64 stack): Remove for gnu
307         assembly version use of ___chkstk.
308
309 2009-09-23  Matthias Klose  <doko@ubuntu.com>
310
311         PR libffi/40242, PR libffi/41443
312         * src/arm/sysv.S (__ARM_ARCH__): Define for processors
313         __ARM_ARCH_6T2__, __ARM_ARCH_6M__, __ARM_ARCH_7__,
314         __ARM_ARCH_7A__, __ARM_ARCH_7R__, __ARM_ARCH_7M__.
315         Change the conditionals to __SOFTFP__ || __ARM_EABI__
316         for -mfloat-abi=softfp to work.
317
318 2009-09-17  Loren J. Rittle  <ljrittle@acm.org>
319
320         PR testsuite/32843 (strikes again)
321         * src/x86/ffi.c (ffi_prep_cif_machdep): Add X86_FREEBSD to
322         enable proper extension on char and short.
323
324 2009-09-15  David Daney  <ddaney@caviumnetworks.com>
325
326         * src/java_raw_api.c (ffi_java_raw_to_rvalue): Remove special
327         handling for FFI_TYPE_POINTER.
328         * src/mips/ffitarget.h (FFI_TYPE_STRUCT_D_SOFT,
329         FFI_TYPE_STRUCT_F_SOFT, FFI_TYPE_STRUCT_DD_SOFT,
330         FFI_TYPE_STRUCT_FF_SOFT, FFI_TYPE_STRUCT_FD_SOFT,
331         FFI_TYPE_STRUCT_DF_SOFT, FFI_TYPE_STRUCT_SOFT): New defines.
332         (FFI_N32_SOFT_FLOAT, FFI_N64_SOFT_FLOAT): New ffi_abi enumerations.
333         (enum ffi_abi): Set FFI_DEFAULT_ABI for soft-float.
334         * src/mips/n32.S (ffi_call_N32): Add handling for soft-float
335         structure and pointer returns.
336         (ffi_closure_N32): Add handling for pointer returns.
337         * src/mips/ffi.c (ffi_prep_args, calc_n32_struct_flags,
338         calc_n32_return_struct_flags): Handle soft-float.
339         (ffi_prep_cif_machdep):  Handle soft-float, fix pointer handling.
340         (ffi_call_N32): Declare proper argument types.
341         (ffi_call, copy_struct_N32, ffi_closure_mips_inner_N32): Handle
342         soft-float.
343
344 2009-08-24  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
345
346         * configure.ac (AC_PREREQ): Bump to 2.64.
347
348 2009-08-22  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
349
350         * Makefile.am (install-html, install-pdf): Remove.
351         * Makefile.in: Regenerate.
352
353         * Makefile.in: Regenerate.
354         * aclocal.m4: Regenerate.
355         * configure: Regenerate.
356         * fficonfig.h.in: Regenerate.
357         * include/Makefile.in: Regenerate.
358         * man/Makefile.in: Regenerate.
359         * testsuite/Makefile.in: Regenerate.
360
361 2009-07-30  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
362
363         * configure.ac (_AC_ARG_VAR_PRECIOUS): Use m4_rename_force.
364
365 2009-07-24  Dave Korn  <dave.korn.cygwin@gmail.com>
366
367         PR libffi/40807
368         * src/x86/ffi.c (ffi_prep_cif_machdep): Also use sign/zero-extending
369         return types for X86_WIN32.
370         * src/x86/win32.S (_ffi_call_SYSV): Handle omitted return types.
371         (_ffi_call_STDCALL, _ffi_closure_SYSV, _ffi_closure_raw_SYSV,
372         _ffi_closure_STDCALL): Likewise.
373
374         * src/closures.c (is_selinux_enabled): Define to const 0 for Cygwin.
375         (dlmmap, dlmunmap): Also use these functions on Cygwin.
376
377 2009-07-11  Richard Sandiford  <rdsandiford@googlemail.com>
378
379         PR testsuite/40699
380         PR testsuite/40707
381         PR testsuite/40709
382         * testsuite/lib/libffi-dg.exp: Revert 2009-07-02, 2009-07-01 and
383         2009-06-30 commits.
384
385 2009-07-01  Richard Sandiford  <r.sandiford@uk.ibm.com>
386
387         * testsuite/lib/libffi-dg.exp (libffi-init): Set ld_library_path
388         to "" before adding paths.  (This reinstates an assignment that
389         was removed by my 2009-06-30 commit, but changes the initial
390         value from "." to "".)
391
392 2009-07-01  H.J. Lu  <hongjiu.lu@intel.com>
393
394         PR testsuite/40601
395         * testsuite/lib/libffi-dg.exp (libffi-init): Properly set
396         gccdir.  Adjust ld_library_path for gcc only if gccdir isn't
397         empty.
398
399 2009-06-30  Richard Sandiford  <r.sandiford@uk.ibm.com>
400
401         * testsuite/lib/libffi-dg.exp (libffi-init): Don't add "."
402         to ld_library_path.  Use add_path.  Add just find_libgcc_s
403         to ld_library_path, not every libgcc multilib directory.
404
405 2009-06-16  Wim Lewis  <wiml@hhhh.org>
406
407         * src/powerpc/ffi.c: Avoid clobbering cr3 and cr4, which are
408         supposed to be callee-saved.
409         * src/powerpc/sysv.S (small_struct_return_value): Fix overrun of
410         return buffer for odd-size structs.
411
412 2009-06-16  Andreas Tobler  <a.tobler@schweiz.org>
413
414         PR libffi/40444
415         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Add
416         allow_stack_execute for Darwin.
417
418 2009-06-16  Andrew Haley  <aph@redhat.com>
419
420         * configure.ac (TARGETDIR): Add missing blank lines.
421         * configure: Regenerate.
422
423 2009-06-16  Andrew Haley  <aph@redhat.com>
424
425         * testsuite/libffi.call/cls_align_sint64.c,
426         testsuite/libffi.call/cls_align_uint64.c,
427         testsuite/libffi.call/cls_longdouble_va.c,
428         testsuite/libffi.call/cls_ulonglong.c,
429         testsuite/libffi.call/return_ll1.c,
430         testsuite/libffi.call/stret_medium2.c: Fix printf format
431         specifiers.
432         * testsuite/libffi.call/ffitest.h,
433         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRIuLL): Define.
434
435 2009-06-15  Andrew Haley  <aph@redhat.com>
436
437         * testsuite/libffi.call/err_bad_typedef.c: xfail everywhere.
438         * testsuite/libffi.call/err_bad_abi.c: Likewise.
439
440 2009-06-12  Andrew Haley  <aph@redhat.com>
441
442         * Makefile.am: Remove info_TEXINFOS.
443
444 2009-06-12  Andrew Haley  <aph@redhat.com>
445
446         * ChangeLog.libffi: testsuite/libffi.call/cls_align_sint64.c,
447         testsuite/libffi.call/cls_align_uint64.c,
448         testsuite/libffi.call/cls_ulonglong.c,
449         testsuite/libffi.call/return_ll1.c,
450         testsuite/libffi.call/stret_medium2.c: Fix printf format
451         specifiers.
452         testsuite/libffi.special/unwindtest.cc: include stdint.h.
453
454 2009-06-11  Timothy Wall  <twall@users.sf.net>
455
456         * Makefile.am,
457         configure.ac,
458         include/ffi.h.in,
459         include/ffi_common.h,
460         src/closures.c,
461         src/dlmalloc.c,
462         src/x86/ffi.c,
463         src/x86/ffitarget.h,
464         src/x86/win64.S (new),
465         README: Added win64 support (mingw or MSVC)
466         * Makefile.in,
467         include/Makefile.in,
468         man/Makefile.in,
469         testsuite/Makefile.in,
470         configure,
471         aclocal.m4: Regenerated
472         * ltcf-c.sh: properly escape cygwin/w32 path
473         * man/ffi_call.3: Clarify size requirements for return value.
474         * src/x86/ffi64.c: Fix filename in comment.
475         * src/x86/win32.S: Remove unused extern.
476
477         * testsuite/libffi.call/closure_fn0.c,
478         testsuite/libffi.call/closure_fn1.c,
479         testsuite/libffi.call/closure_fn2.c,
480         testsuite/libffi.call/closure_fn3.c,
481         testsuite/libffi.call/closure_fn4.c,
482         testsuite/libffi.call/closure_fn5.c,
483         testsuite/libffi.call/closure_fn6.c,
484         testsuite/libffi.call/closure_stdcall.c,
485         testsuite/libffi.call/cls_12byte.c,
486         testsuite/libffi.call/cls_16byte.c,
487         testsuite/libffi.call/cls_18byte.c,
488         testsuite/libffi.call/cls_19byte.c,
489         testsuite/libffi.call/cls_1_1byte.c,
490         testsuite/libffi.call/cls_20byte.c,
491         testsuite/libffi.call/cls_20byte1.c,
492         testsuite/libffi.call/cls_24byte.c,
493         testsuite/libffi.call/cls_2byte.c,
494         testsuite/libffi.call/cls_3_1byte.c,
495         testsuite/libffi.call/cls_3byte1.c,
496         testsuite/libffi.call/cls_3byte2.c,
497         testsuite/libffi.call/cls_4_1byte.c,
498         testsuite/libffi.call/cls_4byte.c,
499         testsuite/libffi.call/cls_5_1_byte.c,
500         testsuite/libffi.call/cls_5byte.c,
501         testsuite/libffi.call/cls_64byte.c,
502         testsuite/libffi.call/cls_6_1_byte.c,
503         testsuite/libffi.call/cls_6byte.c,
504         testsuite/libffi.call/cls_7_1_byte.c,
505         testsuite/libffi.call/cls_7byte.c,
506         testsuite/libffi.call/cls_8byte.c,
507         testsuite/libffi.call/cls_9byte1.c,
508         testsuite/libffi.call/cls_9byte2.c,
509         testsuite/libffi.call/cls_align_double.c,
510         testsuite/libffi.call/cls_align_float.c,
511         testsuite/libffi.call/cls_align_longdouble.c,
512         testsuite/libffi.call/cls_align_longdouble_split.c,
513         testsuite/libffi.call/cls_align_longdouble_split2.c,
514         testsuite/libffi.call/cls_align_pointer.c,
515         testsuite/libffi.call/cls_align_sint16.c,
516         testsuite/libffi.call/cls_align_sint32.c,
517         testsuite/libffi.call/cls_align_sint64.c,
518         testsuite/libffi.call/cls_align_uint16.c,
519         testsuite/libffi.call/cls_align_uint32.c,
520         testsuite/libffi.call/cls_align_uint64.c,
521         testsuite/libffi.call/cls_dbls_struct.c,
522         testsuite/libffi.call/cls_double.c,
523         testsuite/libffi.call/cls_double_va.c,
524         testsuite/libffi.call/cls_float.c,
525         testsuite/libffi.call/cls_longdouble.c,
526         testsuite/libffi.call/cls_longdouble_va.c,
527         testsuite/libffi.call/cls_multi_schar.c,
528         testsuite/libffi.call/cls_multi_sshort.c,
529         testsuite/libffi.call/cls_multi_sshortchar.c,
530         testsuite/libffi.call/cls_multi_uchar.c,
531         testsuite/libffi.call/cls_multi_ushort.c,
532         testsuite/libffi.call/cls_multi_ushortchar.c,
533         testsuite/libffi.call/cls_pointer.c,
534         testsuite/libffi.call/cls_pointer_stack.c,
535         testsuite/libffi.call/cls_schar.c,
536         testsuite/libffi.call/cls_sint.c,
537         testsuite/libffi.call/cls_sshort.c,
538         testsuite/libffi.call/cls_uchar.c,
539         testsuite/libffi.call/cls_uint.c,
540         testsuite/libffi.call/cls_ulonglong.c,
541         testsuite/libffi.call/cls_ushort.c,
542         testsuite/libffi.call/err_bad_abi.c,
543         testsuite/libffi.call/err_bad_typedef.c,
544         testsuite/libffi.call/float2.c,
545         testsuite/libffi.call/huge_struct.c,
546         testsuite/libffi.call/nested_struct.c,
547         testsuite/libffi.call/nested_struct1.c,
548         testsuite/libffi.call/nested_struct10.c,
549         testsuite/libffi.call/nested_struct2.c,
550         testsuite/libffi.call/nested_struct3.c,
551         testsuite/libffi.call/nested_struct4.c,
552         testsuite/libffi.call/nested_struct5.c,
553         testsuite/libffi.call/nested_struct6.c,
554         testsuite/libffi.call/nested_struct7.c,
555         testsuite/libffi.call/nested_struct8.c,
556         testsuite/libffi.call/nested_struct9.c,
557         testsuite/libffi.call/problem1.c,
558         testsuite/libffi.call/return_ldl.c,
559         testsuite/libffi.call/return_ll1.c,
560         testsuite/libffi.call/stret_large.c,
561         testsuite/libffi.call/stret_large2.c,
562         testsuite/libffi.call/stret_medium.c,
563         testsuite/libffi.call/stret_medium2.c,
564         testsuite/libffi.special/unwindtest.cc: use ffi_closure_alloc instead
565         of checking for MMAP.  Use intptr_t instead of long casts.
566
567 2009-06-11  Kaz Kojima  <kkojima@gcc.gnu.org>
568
569         * testsuite/libffi.call/cls_longdouble_va.c: Add xfail sh*-*-linux-*.
570         * testsuite/libffi.call/err_bad_abi.c: Add xfail sh*-*-*.
571         * testsuite/libffi.call/err_bad_typedef.c: Likewise.
572
573 2009-06-09  Andrew Haley  <aph@redhat.com>
574
575         * src/x86/freebsd.S: Add missing file.
576
577 2009-06-08  Andrew Haley  <aph@redhat.com>
578
579         Import from libffi 3.0.8:
580
581         * doc/libffi.texi: New file.
582         * doc/libffi.info: Likewise.
583         * doc/stamp-vti: Likewise.
584         * man/Makefile.am: New file.
585         * man/ffi_call.3: New file.
586
587         * Makefile.am (EXTRA_DIST): Add src/x86/darwin64.S,
588         src/dlmalloc.c.
589         (nodist_libffi_la_SOURCES): Add X86_FREEBSD.
590
591         * configure.ac: Bump version to 3.0.8.
592         parisc*-*-linux*: Add.
593         i386-*-freebsd* | i386-*-openbsd*: Add.
594         powerpc-*-beos*: Add.
595         AM_CONDITIONAL X86_FREEBSD: Add.
596         AC_CONFIG_FILES: Add man/Makefile.
597
598         * include/ffi.h.in (FFI_FN): Change void (*)() to void (*)(void).
599
600 2009-06-08  Andrew Haley  <aph@redhat.com>
601
602         * README: Import from libffi 3.0.8.
603
604 2009-06-08  Andrew Haley  <aph@redhat.com>
605
606         * testsuite/libffi.call/err_bad_abi.c: Add xfails.
607         * testsuite/libffi.call/cls_longdouble_va.c: Add xfails.
608         * testsuite/libffi.call/cls_dbls_struct.c: Add xfail x86_64-*-linux-*.
609         * testsuite/libffi.call/err_bad_typedef.c: Add xfails.
610
611         * testsuite/libffi.call/stret_medium2.c: Add __UNUSED__ to args.
612         * testsuite/libffi.call/stret_medium.c: Likewise.
613         * testsuite/libffi.call/stret_large2.c: Likewise.
614         * testsuite/libffi.call/stret_large.c:  Likewise.
615
616 2008-12-26  Timothy Wall  <twall@users.sf.net>
617
618         * testsuite/libffi.call/cls_longdouble.c,
619         testsuite/libffi.call/cls_longdouble_va.c,
620         testsuite/libffi.call/cls_align_longdouble.c,
621         testsuite/libffi.call/cls_align_longdouble_split.c,
622         testsuite/libffi.call/cls_align_longdouble_split2.c: mark expected
623         failures on x86_64 cygwin/mingw.
624
625 2008-12-22  Timothy Wall  <twall@users.sf.net>
626
627         * testsuite/libffi.call/closure_fn0.c,
628         testsuite/libffi.call/closure_fn1.c,
629         testsuite/libffi.call/closure_fn2.c,
630         testsuite/libffi.call/closure_fn3.c,
631         testsuite/libffi.call/closure_fn4.c,
632         testsuite/libffi.call/closure_fn5.c,
633         testsuite/libffi.call/closure_fn6.c,
634         testsuite/libffi.call/closure_loc_fn0.c,
635         testsuite/libffi.call/closure_stdcall.c,
636         testsuite/libffi.call/cls_align_pointer.c,
637         testsuite/libffi.call/cls_pointer.c,
638         testsuite/libffi.call/cls_pointer_stack.c: use portable cast from
639         pointer to integer (intptr_t).
640         * testsuite/libffi.call/cls_longdouble.c: disable for win64.
641
642 2008-07-24  Anthony Green  <green@redhat.com>
643
644         * testsuite/libffi.call/cls_dbls_struct.c,
645         testsuite/libffi.call/cls_double_va.c,
646         testsuite/libffi.call/cls_longdouble.c,
647         testsuite/libffi.call/cls_longdouble_va.c,
648         testsuite/libffi.call/cls_pointer.c,
649         testsuite/libffi.call/cls_pointer_stack.c,
650         testsuite/libffi.call/err_bad_abi.c: Clean up failures from
651         compiler warnings.
652
653 2008-03-04  Anthony Green  <green@redhat.com>
654             Blake Chaffin
655             hos@tamanegi.org
656
657         * testsuite/libffi.call/cls_align_longdouble_split2.c
658           testsuite/libffi.call/cls_align_longdouble_split.c
659           testsuite/libffi.call/cls_dbls_struct.c
660           testsuite/libffi.call/cls_double_va.c
661           testsuite/libffi.call/cls_longdouble.c
662           testsuite/libffi.call/cls_longdouble_va.c
663           testsuite/libffi.call/cls_pointer.c
664           testsuite/libffi.call/cls_pointer_stack.c
665           testsuite/libffi.call/err_bad_abi.c
666           testsuite/libffi.call/err_bad_typedef.c
667           testsuite/libffi.call/stret_large2.c
668           testsuite/libffi.call/stret_large.c
669           testsuite/libffi.call/stret_medium2.c
670           testsuite/libffi.call/stret_medium.c: New tests from Apple.
671
672 2009-06-05  Andrew Haley  <aph@redhat.com>
673
674         * src/x86/ffitarget.h, src/x86/ffi.c: Merge stdcall changes from
675         libffi.
676
677 2009-06-04  Andrew Haley  <aph@redhat.com>
678
679         * src/x86/ffitarget.h, src/x86/win32.S, src/x86/ffi.c: Back out
680         stdcall changes.
681
682 2008-02-26  Anthony Green  <green@redhat.com>
683             Thomas Heller  <theller@ctypes.org>
684
685         * src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C
686         comment.
687
688 2008-02-03  Timothy Wall  <twall@users.sf.net>
689
690         * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return
691           offset based on code pointer, not data pointer.
692
693 2008-01-31  Timothy Wall <twall@users.sf.net>
694
695         * testsuite/libffi.call/closure_stdcall.c: Add test for stdcall
696         closures.
697         * src/x86/ffitarget.h: Increase size of trampoline for stdcall
698         closures.
699         * src/x86/win32.S: Add assembly for stdcall closure.
700         * src/x86/ffi.c: Initialize stdcall closure trampoline.
701
702 2009-06-04  Andrew Haley  <aph@redhat.com>
703
704         * include/ffi.h.in: Change void (*)() to void (*)(void).
705         * src/x86/ffi.c: Likewise.
706
707 2009-06-04  Andrew Haley  <aph@redhat.com>
708
709         * src/powerpc/ppc_closure.S: Insert licence header.
710         * src/powerpc/linux64_closure.S: Likewise.
711         * src/m68k/sysv.S: Likewise.
712
713         * src/sh64/ffi.c: Change void (*)() to void (*)(void).
714         * src/powerpc/ffi.c: Likewise.
715         * src/powerpc/ffi_darwin.c: Likewise.
716         * src/m32r/ffi.c: Likewise.
717         * src/sh64/ffi.c: Likewise.
718         * src/x86/ffi64.c: Likewise.
719         * src/alpha/ffi.c: Likewise.
720         * src/alpha/osf.S: Likewise.
721         * src/frv/ffi.c: Likewise.
722         * src/s390/ffi.c: Likewise.
723         * src/pa/ffi.c: Likewise.
724         * src/pa/hpux32.S: Likewise.
725         * src/ia64/unix.S: Likewise.
726         * src/ia64/ffi.c: Likewise.
727         * src/sparc/ffi.c: Likewise.
728         * src/mips/ffi.c: Likewise.
729         * src/sh/ffi.c: Likewise.
730
731 2008-02-15  David Daney  <ddaney@avtrex.com>
732
733         * src/mips/ffi.c (USE__BUILTIN___CLEAR_CACHE):
734         Define (conditionally), and use it to include cachectl.h.
735         (ffi_prep_closure_loc): Fix cache flushing.
736         * src/mips/ffitarget.h (_ABIN32, _ABI64, _ABIO32): Define.
737
738 2009-06-04  Andrew Haley  <aph@redhat.com>
739
740         include/ffi.h.in,
741         src/arm/ffitarget.h,
742         src/arm/ffi.c,
743         src/arm/sysv.S,
744         src/powerpc/ffitarget.h,
745         src/closures.c,
746         src/sh64/ffitarget.h,
747         src/sh64/ffi.c,
748         src/sh64/sysv.S,
749         src/types.c,
750         src/x86/ffi64.c,
751         src/x86/ffitarget.h,
752         src/x86/win32.S,
753         src/x86/darwin.S,
754         src/x86/ffi.c,
755         src/x86/sysv.S,
756         src/x86/unix64.S,
757         src/alpha/ffitarget.h,
758         src/alpha/ffi.c,
759         src/alpha/osf.S,
760         src/m68k/ffitarget.h,
761         src/frv/ffitarget.h,
762         src/frv/ffi.c,
763         src/s390/ffitarget.h,
764         src/s390/sysv.S,
765         src/cris/ffitarget.h,
766         src/pa/linux.S,
767         src/pa/ffitarget.h,
768         src/pa/ffi.c,
769         src/raw_api.c,
770         src/ia64/ffitarget.h,
771         src/ia64/unix.S,
772         src/ia64/ffi.c,
773         src/ia64/ia64_flags.h,
774         src/java_raw_api.c,
775         src/debug.c,
776         src/sparc/v9.S,
777         src/sparc/ffitarget.h,
778         src/sparc/ffi.c,
779         src/sparc/v8.S,
780         src/mips/ffitarget.h,
781         src/mips/n32.S,
782         src/mips/o32.S,
783         src/mips/ffi.c,
784         src/prep_cif.c,
785         src/sh/ffitarget.h,
786         src/sh/ffi.c,
787         src/sh/sysv.S: Update license text.
788
789 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
790
791         * src/x86/win32.S (_ffi_closure_STDCALL):  New function.
792         (.eh_frame):  Add FDE for it.
793
794 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
795
796         * configure.ac:  Also check if assembler supports pc-relative
797         relocs on X86_WIN32 targets.
798         * configure:  Regenerate.
799         * src/x86/win32.S (ffi_prep_args):  Declare extern, not global.
800         (_ffi_call_SYSV):  Add missing function type symbol .def and
801         add EH markup labels.
802         (_ffi_call_STDCALL):  Likewise.
803         (_ffi_closure_SYSV):  Likewise.
804         (_ffi_closure_raw_SYSV):  Likewise.
805         (.eh_frame):  Add hand-crafted EH data.
806
807 2009-04-09  Jakub Jelinek  <jakub@redhat.com>
808
809         * testsuite/lib/libffi-dg.exp: Change copyright header to refer to
810         version 3 of the GNU General Public License and to point readers
811         at the COPYING3 file and the FSF's license web page.
812         * testsuite/libffi.call/call.exp: Likewise.
813         * testsuite/libffi.special/special.exp: Likewise.
814
815 2009-03-01  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
816
817         * configure: Regenerate.
818
819 2008-12-18  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
820
821         PR libffi/26048
822         * configure.ac (HAVE_AS_X86_PCREL): New test.
823         * configure: Regenerate.
824         * fficonfig.h.in: Regenerate.
825         * src/x86/sysv.S [!FFI_NO_RAW_API]: Precalculate
826         RAW_CLOSURE_CIF_OFFSET, RAW_CLOSURE_FUN_OFFSET,
827         RAW_CLOSURE_USER_DATA_OFFSET for the Solaris 10/x86 assembler.
828         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
829         * src/x86/unix64.S (.Lstore_table): Move to .text section.
830         (.Lload_table): Likewise.
831         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
832
833 2008-12-18  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
834
835         * configure: Regenerate.
836
837 2008-11-21  Eric Botcazou  <ebotcazou@adacore.com>
838
839         * src/sparc/ffi.c (ffi_prep_cif_machdep): Add support for
840         signed/unsigned int8/16 return values.
841         * src/sparc/v8.S (ffi_call_v8): Likewise.
842         (ffi_closure_v8): Likewise.
843
844 2008-09-26  Peter O'Gorman  <pogma@thewrittenword.com>
845             Steve Ellcey  <sje@cup.hp.com>
846
847         * configure: Regenerate for new libtool.
848         * Makefile.in: Ditto.
849         * include/Makefile.in: Ditto.
850         * aclocal.m4: Ditto.
851
852 2008-08-25  Andreas Tobler  <a.tobler@schweiz.org>
853
854         * src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and
855         FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum.
856         Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT.
857         Adjust copyright notice.
858         * src/powerpc/ffi.c: Add two new flags to indicate if we have one
859         register or two register to use for FFI_SYSV structs.
860         (ffi_prep_cif_machdep): Pass the right register flag introduced above.
861         (ffi_closure_helper_SYSV): Fix the return type for
862         FFI_SYSV_TYPE_SMALL_STRUCT. Comment.
863         Adjust copyright notice.
864
865 2008-07-16  Kaz Kojima  <kkojima@gcc.gnu.org>
866
867         * src/sh/ffi.c (ffi_prep_closure_loc): Turn INSN into an unsigned
868         int.
869
870 2008-06-17  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
871
872         * configure: Regenerate.
873         * include/Makefile.in: Regenerate.
874         * testsuite/Makefile.in: Regenerate.
875
876 2008-06-07  Joseph Myers  <joseph@codesourcery.com>
877
878         * configure.ac (parisc*-*-linux*, powerpc-*-sysv*,
879         powerpc-*-beos*): Remove.
880         * configure: Regenerate.
881
882 2008-05-09  Julian Brown  <julian@codesourcery.com>
883
884         * Makefile.am (LTLDFLAGS): New.
885         (libffi_la_LDFLAGS): Use above.
886         * Makefile.in: Regenerate.
887
888 2008-04-18  Paolo Bonzini  <bonzini@gnu.org>
889
890         PR bootstrap/35457
891         * aclocal.m4: Regenerate.
892         * configure: Regenerate.
893
894 2008-03-26  Kaz Kojima  <kkojima@gcc.gnu.org>
895
896         * src/sh/sysv.S: Add .note.GNU-stack on Linux.
897         * src/sh64/sysv.S: Likewise.
898
899 2008-03-26  Daniel Jacobowitz  <dan@debian.org>
900
901         * src/arm/sysv.S: Fix ARM comment marker.
902
903 2008-03-26  Jakub Jelinek  <jakub@redhat.com>
904
905         * src/alpha/osf.S: Add .note.GNU-stack on Linux.
906         * src/s390/sysv.S: Likewise.
907         * src/powerpc/ppc_closure.S: Likewise.
908         * src/powerpc/sysv.S: Likewise.
909         * src/x86/unix64.S: Likewise.
910         * src/x86/sysv.S: Likewise.
911         * src/sparc/v8.S: Likewise.
912         * src/sparc/v9.S: Likewise.
913         * src/m68k/sysv.S: Likewise.
914         * src/arm/sysv.S: Likewise.
915
916 2008-03-16  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
917
918         * aclocal.m4: Regenerate.
919         * configure: Likewise.
920         * Makefile.in: Likewise.
921         * include/Makefile.in: Likewise.
922         * testsuite/Makefile.in: Likewise.
923
924 2008-02-12  Bjoern Koenig  <bkoenig@alpha-tierchen.de>
925             Andreas Tobler  <a.tobler@schweiz.org>
926
927         * configure.ac: Add amd64-*-freebsd* target.
928         * configure: Regenerate.
929
930 2008-01-30  H.J. Lu  <hongjiu.lu@intel.com>
931
932         PR libffi/34612
933         * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when
934         returning struct.
935
936         * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer"
937         tests.
938
939 2008-01-24  David Edelsohn  <edelsohn@gnu.org>
940
941         * configure: Regenerate.
942
943 2008-01-06  Andreas Tobler  <a.tobler@schweiz.org>
944
945         * src/x86/ffi.c (ffi_prep_cif_machdep): Fix thinko.
946
947 2008-01-05  Andreas Tobler  <a.tobler@schweiz.org>
948
949         PR testsuite/32843
950         * src/x86/ffi.c (ffi_prep_cif_machdep): Add code for
951         signed/unsigned int8/16 for X86_DARWIN.
952         Updated copyright info.
953         Handle one and two byte structs with special cif->flags.
954         * src/x86/ffitarget.h: Add special types for one and two byte structs.
955         Updated copyright info.
956         * src/x86/darwin.S (ffi_call_SYSV): Rewrite to use a jump table like
957         sysv.S
958         Remove code to pop args from the stack after call.
959         Special-case signed/unsigned for int8/16, one and two byte structs.
960         (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
961         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
962         FFI_TYPE_SINT32.
963         Updated copyright info.
964
965 2007-12-08  David Daney  <ddaney@avtrex.com>
966
967         * src/mips/n32.S (ffi_call_N32):  Replace dadd with ADDU, dsub with
968         SUBU, add with ADDU and use smaller code sequences.
969
970 2007-12-07  David Daney  <ddaney@avtrex.com>
971
972         * src/mips/ffi.c (ffi_prep_cif_machdep): Handle long double return
973         type.
974
975 2007-12-06  David Daney  <ddaney@avtrex.com>
976
977         * include/ffi.h.in (FFI_SIZEOF_JAVA_RAW): Define if not already
978         defined.
979         (ffi_java_raw): New typedef.
980         (ffi_java_raw_call, ffi_java_ptrarray_to_raw,
981         ffi_java_raw_to_ptrarray): Change parameter types from ffi_raw to
982         ffi_java_raw.
983         (ffi_java_raw_closure) : Same.
984         (ffi_prep_java_raw_closure, ffi_prep_java_raw_closure_loc): Change
985         parameter types.
986         * src/java_raw_api.c (ffi_java_raw_size):  Replace FFI_SIZEOF_ARG with
987         FFI_SIZEOF_JAVA_RAW.
988         (ffi_java_raw_to_ptrarray): Change type of raw to ffi_java_raw.
989         Replace FFI_SIZEOF_ARG with FFI_SIZEOF_JAVA_RAW. Use
990         sizeof(ffi_java_raw) for alignment calculations.
991         (ffi_java_ptrarray_to_raw): Same.
992         (ffi_java_rvalue_to_raw): Add special handling for FFI_TYPE_POINTER
993         if FFI_SIZEOF_JAVA_RAW == 4.
994         (ffi_java_raw_to_rvalue): Same.
995         (ffi_java_raw_call): Change type of raw to ffi_java_raw.
996         (ffi_java_translate_args): Same.
997         (ffi_prep_java_raw_closure_loc, ffi_prep_java_raw_closure): Change
998         parameter types.
999         * src/mips/ffitarget.h (FFI_SIZEOF_JAVA_RAW): Define for N32 ABI.
1000
1001 2007-12-06  David Daney  <ddaney@avtrex.com>
1002
1003         * src/mips/n32.S (ffi_closure_N32): Use 64-bit add instruction on
1004         pointer values.
1005
1006 2007-12-01  Andreas Tobler  <a.tobler@schweiz.org>
1007
1008         PR libffi/31937
1009         * src/powerpc/ffitarget.h: Introduce new ABI FFI_LINUX_SOFT_FLOAT.
1010         Add local FFI_TYPE_UINT128 to handle soft-float long-double-128.
1011         * src/powerpc/ffi.c: Distinguish between __NO_FPRS__ and not and
1012         set the NUM_FPR_ARG_REGISTERS according to.
1013         Add support for potential soft-float support under hard-float
1014         architecture.
1015         (ffi_prep_args_SYSV): Set NUM_FPR_ARG_REGISTERS to 0 in case of
1016         FFI_LINUX_SOFT_FLOAT, handle float, doubles and long-doubles according
1017         to the FFI_LINUX_SOFT_FLOAT ABI.
1018         (ffi_prep_cif_machdep): Likewise.
1019         (ffi_closure_helper_SYSV): Likewise.
1020         * src/powerpc/ppc_closure.S: Make sure not to store float/double
1021         on archs where __NO_FPRS__ is true.
1022         Add FFI_TYPE_UINT128 support.
1023         * src/powerpc/sysv.S: Add support for soft-float long-double-128.
1024         Adjust copyright notice.
1025
1026 2007-11-25  Andreas Tobler  <a.tobler@schweiz.org>
1027
1028         * src/closures.c: Move defintion of MAYBE_UNUSED from here to ...
1029         * include/ffi_common.h: ... here.
1030         Update copyright.
1031
1032 2007-11-17  Andreas Tobler  <a.tobler@schweiz.org>
1033
1034         * src/powerpc/sysv.S: Load correct cr to compare if we have long double.
1035         * src/powerpc/linux64.S: Likewise.
1036         * src/powerpc/ffi.c: Add a comment to show which part goes into cr6.
1037         * testsuite/libffi.call/return_ldl.c: New test.
1038
1039 2007-09-04    <aph@redhat.com>
1040
1041         * src/arm/sysv.S (UNWIND): New.
1042         (Whole file): Conditionally compile unwinder directives.
1043         * src/arm/sysv.S: Add unwinder directives.
1044
1045         * src/arm/ffi.c (ffi_prep_args): Align structs by at least 4 bytes.
1046         Only treat r0 as a struct address if we're actually returning a
1047         struct by address.
1048         Only copy the bytes that are actually within a struct.
1049         (ffi_prep_cif_machdep): A Composite Type not larger than 4 bytes
1050         is returned in r0, not passed by address.
1051         (ffi_call): Allocate a word-sized temporary for the case where
1052         a composite is returned in r0.
1053         (ffi_prep_incoming_args_SYSV): Align as necessary.
1054
1055 2007-08-05  Steven Newbury  <s_j_newbury@yahoo.co.uk>
1056
1057         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Use __clear_cache instead of
1058         directly using the sys_cacheflush syscall.
1059
1060 2007-07-27  Andrew Haley  <aph@redhat.com>
1061
1062         * src/arm/sysv.S (ffi_closure_SYSV): Add soft-float.
1063
1064 2007-09-03  Maciej W. Rozycki  <macro@linux-mips.org>
1065
1066         * Makefile.am: Unify MIPS_IRIX and MIPS_LINUX into MIPS.
1067         * configure.ac: Likewise.
1068         * Makefile.in: Regenerate.
1069         * include/Makefile.in: Likewise.
1070         * testsuite/Makefile.in: Likewise.
1071         * configure: Likewise.
1072
1073 2007-08-24  David Daney  <ddaney@avtrex.com>
1074
1075         * testsuite/libffi.call/return_sl.c: New test.
1076
1077 2007-08-10  David Daney  <ddaney@avtrex.com>
1078
1079         * testsuite/libffi.call/cls_multi_ushort.c,
1080         testsuite/libffi.call/cls_align_uint16.c,
1081         testsuite/libffi.call/nested_struct1.c,
1082         testsuite/libffi.call/nested_struct3.c,
1083         testsuite/libffi.call/cls_7_1_byte.c,
1084         testsuite/libffi.call/nested_struct5.c,
1085         testsuite/libffi.call/cls_double.c,
1086         testsuite/libffi.call/nested_struct7.c,
1087         testsuite/libffi.call/cls_sint.c,
1088         testsuite/libffi.call/nested_struct9.c,
1089         testsuite/libffi.call/cls_20byte1.c,
1090         testsuite/libffi.call/cls_multi_sshortchar.c,
1091         testsuite/libffi.call/cls_align_sint64.c,
1092         testsuite/libffi.call/cls_3byte2.c,
1093         testsuite/libffi.call/cls_multi_schar.c,
1094         testsuite/libffi.call/cls_multi_uchar.c,
1095         testsuite/libffi.call/cls_19byte.c,
1096         testsuite/libffi.call/cls_9byte1.c,
1097         testsuite/libffi.call/cls_align_float.c,
1098         testsuite/libffi.call/closure_fn1.c,
1099         testsuite/libffi.call/problem1.c,
1100         testsuite/libffi.call/closure_fn3.c,
1101         testsuite/libffi.call/cls_sshort.c,
1102         testsuite/libffi.call/closure_fn5.c,
1103         testsuite/libffi.call/cls_align_double.c,
1104         testsuite/libffi.call/nested_struct.c,
1105         testsuite/libffi.call/cls_2byte.c,
1106         testsuite/libffi.call/nested_struct10.c,
1107         testsuite/libffi.call/cls_4byte.c,
1108         testsuite/libffi.call/cls_6byte.c,
1109         testsuite/libffi.call/cls_8byte.c,
1110         testsuite/libffi.call/cls_multi_sshort.c,
1111         testsuite/libffi.call/cls_align_sint16.c,
1112         testsuite/libffi.call/cls_align_uint32.c,
1113         testsuite/libffi.call/cls_20byte.c,
1114         testsuite/libffi.call/cls_float.c,
1115         testsuite/libffi.call/nested_struct2.c,
1116         testsuite/libffi.call/cls_5_1_byte.c,
1117         testsuite/libffi.call/nested_struct4.c,
1118         testsuite/libffi.call/cls_24byte.c,
1119         testsuite/libffi.call/nested_struct6.c,
1120         testsuite/libffi.call/cls_64byte.c,
1121         testsuite/libffi.call/nested_struct8.c,
1122         testsuite/libffi.call/cls_uint.c,
1123         testsuite/libffi.call/cls_multi_ushortchar.c,
1124         testsuite/libffi.call/cls_schar.c,
1125         testsuite/libffi.call/cls_uchar.c,
1126         testsuite/libffi.call/cls_align_uint64.c,
1127         testsuite/libffi.call/cls_ulonglong.c,
1128         testsuite/libffi.call/cls_align_longdouble.c,
1129         testsuite/libffi.call/cls_1_1byte.c,
1130         testsuite/libffi.call/cls_12byte.c,
1131         testsuite/libffi.call/cls_3_1byte.c,
1132         testsuite/libffi.call/cls_3byte1.c,
1133         testsuite/libffi.call/cls_4_1byte.c,
1134         testsuite/libffi.call/cls_6_1_byte.c,
1135         testsuite/libffi.call/cls_16byte.c,
1136         testsuite/libffi.call/cls_18byte.c,
1137         testsuite/libffi.call/closure_fn0.c,
1138         testsuite/libffi.call/cls_9byte2.c,
1139         testsuite/libffi.call/closure_fn2.c,
1140         testsuite/libffi.call/closure_fn4.c,
1141         testsuite/libffi.call/cls_ushort.c,
1142         testsuite/libffi.call/closure_fn6.c,
1143         testsuite/libffi.call/cls_5byte.c,
1144         testsuite/libffi.call/cls_align_pointer.c,
1145         testsuite/libffi.call/cls_7byte.c,
1146         testsuite/libffi.call/cls_align_sint32.c,
1147         testsuite/libffi.special/unwindtest_ffi_call.cc,
1148         testsuite/libffi.special/unwindtest.cc: Remove xfail for mips64*-*-*.
1149
1150 2007-08-10  David Daney  <ddaney@avtrex.com>
1151
1152         PR libffi/28313
1153         * configure.ac: Don't treat mips64 as a special case.
1154         * Makefile.am (nodist_libffi_la_SOURCES): Add n32.S.
1155         * configure: Regenerate
1156         * Makefile.in: Ditto.
1157         * fficonfig.h.in: Ditto.
1158         * src/mips/ffitarget.h (REG_L, REG_S, SUBU, ADDU, SRL, LI): Indent.
1159         (LA, EH_FRAME_ALIGN, FDE_ADDR_BYTES): New preprocessor macros.
1160         (FFI_DEFAULT_ABI): Set for n64 case.
1161         (FFI_CLOSURES, FFI_TRAMPOLINE_SIZE): Define for n32 and n64 cases.
1162         * src/mips/n32.S (ffi_call_N32): Add debug macros and labels for FDE.
1163         (ffi_closure_N32): New function.
1164         (.eh_frame): New section
1165         * src/mips/o32.S: Clean up comments.
1166         (ffi_closure_O32): Pass ffi_closure parameter in $12.
1167         * src/mips/ffi.c: Use FFI_MIPS_N32 instead of
1168         _MIPS_SIM == _ABIN32 throughout.
1169         (FFI_MIPS_STOP_HERE): New, use in place of
1170         ffi_stop_here.
1171         (ffi_prep_args): Use unsigned long to hold pointer values.  Rewrite
1172         to support n32/n64 ABIs.
1173         (calc_n32_struct_flags): Rewrite.
1174         (calc_n32_return_struct_flags): Remove unused variable.  Reverse
1175         position of flag bits.
1176         (ffi_prep_cif_machdep): Rewrite n32 portion.
1177         (ffi_call): Enable for n64.  Add special handling for small structure
1178         return values.
1179         (ffi_prep_closure_loc): Add n32 and n64 support.
1180         (ffi_closure_mips_inner_O32): Add cast to silence warning.
1181         (copy_struct_N32, ffi_closure_mips_inner_N32): New functions.
1182
1183 2007-08-08  David Daney  <ddaney@avtrex.com>
1184
1185         * testsuite/libffi.call/ffitest.h (ffi_type_mylong): Remove definition.
1186         * testsuite/libffi.call/cls_align_uint16.c (main): Use correct type
1187         specifiers.
1188         * testsuite/libffi.call/nested_struct1.c (main): Ditto.
1189         * testsuite/libffi.call/cls_sint.c (main): Ditto.
1190         * testsuite/libffi.call/nested_struct9.c (main): Ditto.
1191         * testsuite/libffi.call/cls_20byte1.c (main): Ditto.
1192         * testsuite/libffi.call/cls_9byte1.c (main): Ditto.
1193         * testsuite/libffi.call/closure_fn1.c (main): Ditto.
1194         * testsuite/libffi.call/closure_fn3.c (main): Ditto.
1195         * testsuite/libffi.call/return_dbl2.c (main): Ditto.
1196         * testsuite/libffi.call/cls_sshort.c (main): Ditto.
1197         * testsuite/libffi.call/return_fl3.c (main): Ditto.
1198         * testsuite/libffi.call/closure_fn5.c (main): Ditto.
1199         * testsuite/libffi.call/nested_struct.c (main): Ditto.
1200         * testsuite/libffi.call/nested_struct10.c (main): Ditto.
1201         * testsuite/libffi.call/return_ll1.c (main): Ditto.
1202         * testsuite/libffi.call/cls_8byte.c (main): Ditto.
1203         * testsuite/libffi.call/cls_align_uint32.c (main): Ditto.
1204         * testsuite/libffi.call/cls_align_sint16.c (main): Ditto.
1205         * testsuite/libffi.call/cls_20byte.c (main): Ditto.
1206         * testsuite/libffi.call/nested_struct2.c (main): Ditto.
1207         * testsuite/libffi.call/cls_24byte.c (main): Ditto.
1208         * testsuite/libffi.call/nested_struct6.c (main): Ditto.
1209         * testsuite/libffi.call/cls_uint.c (main): Ditto.
1210         * testsuite/libffi.call/cls_12byte.c (main): Ditto.
1211         * testsuite/libffi.call/cls_16byte.c (main): Ditto.
1212         * testsuite/libffi.call/closure_fn0.c (main): Ditto.
1213         * testsuite/libffi.call/cls_9byte2.c (main): Ditto.
1214         * testsuite/libffi.call/closure_fn2.c (main): Ditto.
1215         * testsuite/libffi.call/return_dbl1.c (main): Ditto.
1216         * testsuite/libffi.call/closure_fn4.c (main): Ditto.
1217         * testsuite/libffi.call/closure_fn6.c (main): Ditto.
1218         * testsuite/libffi.call/cls_align_sint32.c (main): Ditto.
1219
1220 2007-08-07  Andrew Haley  <aph@redhat.com>
1221
1222         * src/x86/sysv.S (ffi_closure_raw_SYSV): Fix typo in previous
1223         checkin.
1224
1225 2007-08-06  Andrew Haley  <aph@redhat.com>
1226
1227         PR testsuite/32843
1228         * src/x86/sysv.S (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1229         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1230         FFI_TYPE_SINT32.
1231
1232 2007-08-02  David Daney  <ddaney@avtrex.com>
1233
1234         * testsuite/libffi.call/return_ul.c (main): Define return type as
1235         ffi_arg.  Use proper printf conversion specifier.
1236
1237 2007-07-30  Andrew Haley  <aph@redhat.com>
1238
1239         PR testsuite/32843
1240         * src/x86/ffi.c (ffi_prep_cif_machdep): in x86 case, add code for
1241         signed/unsigned int8/16.
1242         * src/x86/sysv.S (ffi_call_SYSV): Rewrite to:
1243         Use a jump table.
1244         Remove code to pop args from the stack after call.
1245         Special-case signed/unsigned int8/16.
1246         * testsuite/libffi.call/return_sc.c (main): Revert.
1247
1248 2007-07-26  Richard Guenther  <rguenther@suse.de>
1249
1250         PR testsuite/32843
1251         * testsuite/libffi.call/return_sc.c (main): Verify call
1252         result as signed char, not ffi_arg.
1253
1254 2007-07-16  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1255
1256         * configure.ac (i?86-*-solaris2.1[0-9]): Set TARGET to X86_64.
1257         * configure: Regenerate.
1258
1259 2007-07-11  David Daney  <ddaney@avtrex.com>
1260
1261         * src/mips/ffi.c: Don't include sys/cachectl.h.
1262         (ffi_prep_closure_loc): Use __builtin___clear_cache() instead of
1263         cacheflush().
1264
1265 2007-05-18  Aurelien Jarno  <aurelien@aurel32.net>
1266
1267         * src/arm/ffi.c (ffi_prep_closure_loc): Renamed and ajusted
1268         from (ffi_prep_closure): ... this.
1269         (FFI_INIT_TRAMPOLINE): Adjust.
1270
1271 2005-12-31  Phil Blundell  <pb@reciva.com>
1272
1273         * src/arm/ffi.c (ffi_prep_incoming_args_SYSV,
1274         ffi_closure_SYSV_inner, ffi_prep_closure): New, add closure support.
1275         * src/arm/sysv.S(ffi_closure_SYSV): Likewise.
1276         * src/arm/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1277         (FFI_CLOSURES): Enable closure support.
1278
1279 2007-07-03  Andrew Haley  <aph@hedges.billgatliff.com>
1280
1281         * testsuite/libffi.call/cls_multi_ushort.c,
1282         testsuite/libffi.call/cls_align_uint16.c,
1283         testsuite/libffi.call/nested_struct1.c,
1284         testsuite/libffi.call/nested_struct3.c,
1285         testsuite/libffi.call/cls_7_1_byte.c,
1286         testsuite/libffi.call/cls_double.c,
1287         testsuite/libffi.call/nested_struct5.c,
1288         testsuite/libffi.call/nested_struct7.c,
1289         testsuite/libffi.call/cls_sint.c,
1290         testsuite/libffi.call/nested_struct9.c,
1291         testsuite/libffi.call/cls_20byte1.c,
1292         testsuite/libffi.call/cls_multi_sshortchar.c,
1293         testsuite/libffi.call/cls_align_sint64.c,
1294         testsuite/libffi.call/cls_3byte2.c,
1295         testsuite/libffi.call/cls_multi_schar.c,
1296         testsuite/libffi.call/cls_multi_uchar.c,
1297         testsuite/libffi.call/cls_19byte.c,
1298         testsuite/libffi.call/cls_9byte1.c,
1299         testsuite/libffi.call/cls_align_float.c,
1300         testsuite/libffi.call/closure_fn1.c,
1301         testsuite/libffi.call/problem1.c,
1302         testsuite/libffi.call/closure_fn3.c,
1303         testsuite/libffi.call/cls_sshort.c,
1304         testsuite/libffi.call/closure_fn5.c,
1305         testsuite/libffi.call/cls_align_double.c,
1306         testsuite/libffi.call/cls_2byte.c,
1307         testsuite/libffi.call/nested_struct.c,
1308         testsuite/libffi.call/nested_struct10.c,
1309         testsuite/libffi.call/cls_4byte.c,
1310         testsuite/libffi.call/cls_6byte.c,
1311         testsuite/libffi.call/cls_8byte.c,
1312         testsuite/libffi.call/cls_multi_sshort.c,
1313         testsuite/libffi.call/cls_align_uint32.c,
1314         testsuite/libffi.call/cls_align_sint16.c,
1315         testsuite/libffi.call/cls_float.c,
1316         testsuite/libffi.call/cls_20byte.c,
1317         testsuite/libffi.call/cls_5_1_byte.c,
1318         testsuite/libffi.call/nested_struct2.c,
1319         testsuite/libffi.call/cls_24byte.c,
1320         testsuite/libffi.call/nested_struct4.c,
1321         testsuite/libffi.call/nested_struct6.c,
1322         testsuite/libffi.call/cls_64byte.c,
1323         testsuite/libffi.call/nested_struct8.c,
1324         testsuite/libffi.call/cls_uint.c,
1325         testsuite/libffi.call/cls_multi_ushortchar.c,
1326         testsuite/libffi.call/cls_schar.c,
1327         testsuite/libffi.call/cls_uchar.c,
1328         testsuite/libffi.call/cls_align_uint64.c,
1329         testsuite/libffi.call/cls_ulonglong.c,
1330         testsuite/libffi.call/cls_align_longdouble.c,
1331         testsuite/libffi.call/cls_1_1byte.c,
1332         testsuite/libffi.call/cls_12byte.c,
1333         testsuite/libffi.call/cls_3_1byte.c,
1334         testsuite/libffi.call/cls_3byte1.c,
1335         testsuite/libffi.call/cls_4_1byte.c,
1336         testsuite/libffi.call/cls_6_1_byte.c,
1337         testsuite/libffi.call/cls_16byte.c,
1338         testsuite/libffi.call/cls_18byte.c,
1339         testsuite/libffi.call/closure_fn0.c,
1340         testsuite/libffi.call/cls_9byte2.c,
1341         testsuite/libffi.call/closure_fn2.c,
1342         testsuite/libffi.call/closure_fn4.c,
1343         testsuite/libffi.call/cls_ushort.c,
1344         testsuite/libffi.call/closure_fn6.c,
1345         testsuite/libffi.call/cls_5byte.c,
1346         testsuite/libffi.call/cls_align_pointer.c,
1347         testsuite/libffi.call/cls_7byte.c,
1348         testsuite/libffi.call/cls_align_sint32.c,
1349         testsuite/libffi.special/unwindtest_ffi_call.cc,
1350         testsuite/libffi.special/unwindtest.cc: Enable for ARM.
1351
1352 2007-07-05  H.J. Lu  <hongjiu.lu@intel.com>
1353
1354         * aclocal.m4: Regenerated.
1355
1356 2007-06-02  Paolo Bonzini  <bonzini@gnu.org>
1357
1358         * configure: Regenerate.
1359
1360 2007-05-23  Steve Ellcey  <sje@cup.hp.com>
1361
1362         * Makefile.in: Regenerate.
1363         * configure: Regenerate.
1364         * aclocal.m4: Regenerate.
1365         * include/Makefile.in: Regenerate.
1366         * testsuite/Makefile.in: Regenerate.
1367
1368 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1369
1370         * src/m68k/ffi.c (ffi_prep_incoming_args_SYSV,
1371         ffi_closure_SYSV_inner,ffi_prep_closure): New, add closure support.
1372         * src/m68k/sysv.S(ffi_closure_SYSV,ffi_closure_struct_SYSV): Likewise.
1373         * src/m68k/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1374         (FFI_CLOSURES): Enable closure support.
1375
1376 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1377
1378         * configure.ac (HAVE_AS_CFI_PSEUDO_OP): New test.
1379         * configure: Regenerate.
1380         * fficonfig.h.in: Regenerate.
1381         * src/m68k/sysv.S (CFI_STARTPROC,CFI_ENDPROC,
1382         CFI_OFFSET,CFI_DEF_CFA): New macros.
1383         (ffi_call_SYSV): Add callframe annotation.
1384
1385 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1386
1387         * src/m68k/ffi.c (ffi_prep_args,ffi_prep_cif_machdep): Fix
1388         numerous test suite failures.
1389         * src/m68k/sysv.S (ffi_call_SYSV): Likewise.
1390
1391 2007-04-11  Paolo Bonzini  <bonzini@gnu.org>
1392
1393         * Makefile.am (EXTRA_DIST): Bring up to date.
1394         * Makefile.in: Regenerate.
1395         * src/frv/eabi.S: Remove RCS keyword.
1396
1397 2007-04-06  Richard Henderson  <rth@redhat.com>
1398
1399         * configure.ac: Tidy target case.
1400         (HAVE_LONG_DOUBLE): Allow the target to override.
1401         * configure: Regenerate.
1402         * include/ffi.h.in: Don't define ffi_type_foo if
1403         LIBFFI_HIDE_BASIC_TYPES is defined.
1404         (ffi_type_longdouble): If not HAVE_LONG_DOUBLE, define
1405         to ffi_type_double.
1406         * types.c (LIBFFI_HIDE_BASIC_TYPES): Define.
1407         (FFI_TYPEDEF, ffi_type_void): Mark the data const.
1408         (ffi_type_longdouble): Special case for Alpha.  Don't define
1409         if long double == double.
1410
1411         * src/alpha/ffi.c (FFI_TYPE_LONGDOUBLE): Assert unique value.
1412         (ffi_prep_cif_machdep): Handle it as the 128-bit type.
1413         (ffi_call, ffi_closure_osf_inner): Likewise.
1414         (ffi_closure_osf_inner): Likewise.  Mark hidden.
1415         (ffi_call_osf, ffi_closure_osf): Mark hidden.
1416         * src/alpha/ffitarget.h (FFI_LAST_ABI): Tidy definition.
1417         * src/alpha/osf.S (ffi_call_osf, ffi_closure_osf): Mark hidden.
1418         (load_table): Handle 128-bit long double.
1419
1420         * testsuite/libffi.call/float4.c: Add -mieee for alpha.
1421
1422 2007-04-06  Tom Tromey  <tromey@redhat.com>
1423
1424         PR libffi/31491:
1425         * README: Fixed bug in example.
1426
1427 2007-04-03  Jakub Jelinek  <jakub@redhat.com>
1428
1429         * src/closures.c: Include sys/statfs.h.
1430         (_GNU_SOURCE): Define on Linux.
1431         (FFI_MMAP_EXEC_SELINUX): Define.
1432         (selinux_enabled): New variable.
1433         (selinux_enabled_check): New function.
1434         (is_selinux_enabled): Define.
1435         (dlmmap): Use it.
1436
1437 2007-03-24  Uros Bizjak  <ubizjak@gmail.com>
1438
1439         * testsuite/libffi.call/return_fl2.c (return_fl): Mark as static.
1440         Use 'volatile float sum' to create sum of floats to avoid false
1441         negative due to excess precision on ix86 targets.
1442         (main): Ditto.
1443
1444 2007-03-08  Alexandre Oliva  <aoliva@redhat.com>
1445
1446         * src/powerpc/ffi.c (flush_icache): Fix left-over from previous
1447         patch.
1448         (ffi_prep_closure_loc): Remove unneeded casts.  Add needed ones.
1449
1450 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1451
1452         * include/ffi.h.in (ffi_closure_alloc, ffi_closure_free): New.
1453         (ffi_prep_closure_loc): New.
1454         (ffi_prep_raw_closure_loc): New.
1455         (ffi_prep_java_raw_closure_loc): New.
1456         * src/closures.c: New file.
1457         * src/dlmalloc.c [FFI_MMAP_EXEC_WRIT] (struct malloc_segment):
1458         Replace sflags with exec_offset.
1459         [FFI_MMAP_EXEC_WRIT] (mmap_exec_offset, add_segment_exec_offset,
1460         sub_segment_exec_offset): New macros.
1461         (get_segment_flags, set_segment_flags, check_segment_merge): New
1462         macros.
1463         (is_mmapped_segment, is_extern_segment): Use get_segment_flags.
1464         (add_segment, sys_alloc, create_mspace, create_mspace_with_base,
1465         destroy_mspace): Use new macros.
1466         (sys_alloc): Silence warning.
1467         * Makefile.am (libffi_la_SOURCES): Add src/closures.c.
1468         * Makefile.in: Rebuilt.
1469         * src/prep_cif [FFI_CLOSURES] (ffi_prep_closure): Implement in
1470         terms of ffi_prep_closure_loc.
1471         * src/raw_api.c (ffi_prep_raw_closure_loc): Renamed and adjusted
1472         from...
1473         (ffi_prep_raw_closure): ... this.  Re-implement in terms of the
1474         renamed version.
1475         * src/java_raw_api (ffi_prep_java_raw_closure_loc): Renamed and
1476         adjusted from...
1477         (ffi_prep_java_raw_closure): ... this.  Re-implement in terms of
1478         the renamed version.
1479         * src/alpha/ffi.c (ffi_prep_closure_loc): Renamed from
1480         (ffi_prep_closure): ... this.
1481         * src/pa/ffi.c: Likewise.
1482         * src/cris/ffi.c: Likewise.  Adjust.
1483         * src/frv/ffi.c: Likewise.
1484         * src/ia64/ffi.c: Likewise.
1485         * src/mips/ffi.c: Likewise.
1486         * src/powerpc/ffi_darwin.c: Likewise.
1487         * src/s390/ffi.c: Likewise.
1488         * src/sh/ffi.c: Likewise.
1489         * src/sh64/ffi.c: Likewise.
1490         * src/sparc/ffi.c: Likewise.
1491         * src/x86/ffi64.c: Likewise.
1492         * src/x86/ffi.c: Likewise.
1493         (FFI_INIT_TRAMPOLINE): Adjust.
1494         (ffi_prep_raw_closure_loc): Renamed and adjusted from...
1495         (ffi_prep_raw_closure): ... this.
1496         * src/powerpc/ffi.c (ffi_prep_closure_loc): Renamed from
1497         (ffi_prep_closure): ... this.
1498         (flush_icache): Adjust.
1499
1500 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1501
1502         * src/dlmalloc.c: New file, imported version 2.8.3 of Doug
1503         Lea's malloc.
1504
1505 2007-03-01  Brooks Moses  <brooks.moses@codesourcery.com>
1506
1507         * Makefile.am: Add dummy install-pdf target.
1508         * Makefile.in: Regenerate
1509
1510 2007-02-13  Andreas Krebbel  <krebbel1@de.ibm.com>
1511
1512         * src/s390/ffi.c (ffi_prep_args, ffi_prep_cif_machdep,
1513         ffi_closure_helper_SYSV): Add long double handling.
1514
1515 2007-02-02  Jakub Jelinek  <jakub@redhat.com>
1516
1517         * src/powerpc/linux64.S (ffi_call_LINUX64): Move restore of r2
1518         immediately after bctrl instruction.
1519
1520 2007-01-18  Alexandre Oliva  <aoliva@redhat.com>
1521
1522         * Makefile.am (all-recursive, install-recursive,
1523         mostlyclean-recursive, clean-recursive, distclean-recursive,
1524         maintainer-clean-recursive): Add missing targets.
1525         * Makefile.in: Rebuilt.
1526
1527 2006-12-14  Andreas Tobler  <a.tobler@schweiz.org>
1528
1529         * configure.ac: Add TARGET for x86_64-*-darwin*.
1530         * Makefile.am (nodist_libffi_la_SOURCES): Add rules for 64-bit sources
1531         for X86_DARWIN.
1532         * src/x86/ffitarget.h: Set trampoline size for x86_64-*-darwin*.
1533         * src/x86/darwin64.S: New file for x86_64-*-darwin* support.
1534         * configure: Regenerate.
1535         * Makefile.in: Regenerate.
1536         * include/Makefile.in: Regenerate.
1537         * testsuite/Makefile.in: Regenerate.
1538         * testsuite/libffi.special/unwindtest_ffi_call.cc: New test case for
1539         ffi_call only.
1540
1541 2006-12-13  Andreas Tobler <a.tobler@schweiz.org>
1542
1543         * aclocal.m4: Regenerate with aclocal -I .. as written in the
1544         Makefile.am.
1545
1546 2006-10-31  Geoffrey Keating  <geoffk@apple.com>
1547
1548         * src/powerpc/ffi_darwin.c (darwin_adjust_aggregate_sizes): New.
1549         (ffi_prep_cif_machdep): Call darwin_adjust_aggregate_sizes for
1550         Darwin.
1551         * testsuite/libffi.call/nested_struct4.c: Remove Darwin XFAIL.
1552         * testsuite/libffi.call/nested_struct6.c: Remove Darwin XFAIL.
1553
1554 2006-10-10  Paolo Bonzini  <bonzini@gnu.org>
1555             Sandro Tolaini  <tolaini@libero.it>
1556
1557         * configure.ac [i*86-*-darwin*]: Set X86_DARWIN symbol and
1558         conditional.
1559         * configure: Regenerated.
1560         * Makefile.am (nodist_libffi_la_SOURCES) [X86_DARWIN]: New case.
1561         (EXTRA_DIST): Add src/x86/darwin.S.
1562         * Makefile.in: Regenerated.
1563         * include/Makefile.in: Regenerated.
1564         * testsuite/Makefile.in: Regenerated.
1565
1566         * src/x86/ffi.c (ffi_prep_cif_machdep) [X86_DARWIN]: Treat like
1567         X86_WIN32, and additionally align stack to 16 bytes.
1568         * src/x86/darwin.S: New, based on sysv.S.
1569         * src/prep_cif.c (ffi_prep_cif) [X86_DARWIN]: Align > 8-byte structs.
1570
1571 2006-09-12  David Daney  <ddaney@avtrex.com>
1572
1573         PR libffi/23935
1574         * include/Makefile.am: Install both ffi.h and ffitarget.h in
1575         $(libdir)/gcc/$(target_alias)/$(gcc_version)/include.
1576         * aclocal.m4: Regenerated for automake 1.9.6.
1577         * Makefile.in: Regenerated.
1578         * include/Makefile.in: Regenerated.
1579         * testsuite/Makefile.in: Regenerated.
1580
1581 2006-08-17  Andreas Tobler  <a.tobler@schweiz.ch>
1582
1583         * include/ffi_common.h (struct): Revert accidental commit.
1584
1585 2006-08-15  Andreas Tobler  <a.tobler@schweiz.ch>
1586
1587         * include/ffi_common.h: Remove lint directives.
1588         * include/ffi.h.in: Likewise.
1589
1590 2006-07-25  Torsten Schoenfeld  <kaffeetisch@gmx.de>
1591
1592         * include/ffi.h.in (ffi_type_ulong, ffi_type_slong): Define correctly
1593         for 32-bit architectures.
1594         * testsuite/libffi.call/return_ul.c: New test case.
1595
1596 2006-07-19  David Daney  <ddaney@avtrex.com>
1597
1598         * testsuite/libffi.call/closure_fn6.c: Remove xfail for mips,
1599         xfail remains for mips64.
1600
1601 2006-05-23  Carlos O'Donell  <carlos@codesourcery.com>
1602
1603         * Makefile.am: Add install-html target. Add install-html to .PHONY
1604         * Makefile.in: Regenerate.
1605         * aclocal.m4: Regenerate.
1606         * include/Makefile.in: Regenerate.
1607         * testsuite/Makefile.in: Regenerate.
1608
1609 2006-05-18  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1610
1611         * pa/ffi.c (ffi_prep_args_pa32): Load floating point arguments from
1612         stack slot.
1613
1614 2006-04-22  Andreas Tobler  <a.tobler@schweiz.ch>
1615
1616         * README: Remove notice about 'Crazy Comments'.
1617         * src/debug.c: Remove lint directives. Cleanup white spaces.
1618         * src/java_raw_api.c: Likewise.
1619         * src/prep_cif.c: Likewise.
1620         * src/raw_api.c: Likewise.
1621         * src/ffitest.c: Delete. No longer needed, all test cases migrated
1622         to the testsuite.
1623         * src/arm/ffi.c: Remove lint directives.
1624         * src/m32r/ffi.c: Likewise.
1625         * src/pa/ffi.c: Likewise.
1626         * src/powerpc/ffi.c: Likewise.
1627         * src/powerpc/ffi_darwin.c: Likewise.
1628         * src/sh/ffi.c: Likewise.
1629         * src/sh64/ffi.c: Likewise.
1630         * src/x86/ffi.c: Likewise.
1631         * testsuite/libffi.call/float2.c: Likewise.
1632         * testsuite/libffi.call/promotion.c: Likewise.
1633         * testsuite/libffi.call/struct1.c: Likewise.
1634
1635 2006-04-13  Andreas Tobler  <a.tobler@schweiz.ch>
1636
1637         * src/pa/hpux32.S: Correct unwind offset calculation for
1638         ffi_closure_pa32.
1639         * src/pa/linux.S: Likewise.
1640
1641 2006-04-12  James E Wilson  <wilson@specifix.com>
1642
1643         PR libgcj/26483
1644         * src/ia64/ffi.c (stf_spill, ldf_fill): Rewrite as macros.
1645         (hfa_type_load): Call stf_spill.
1646         (hfa_type_store): Call ldf_fill.
1647         (ffi_call): Adjust calls to above routines.  Add local temps for
1648         macro result.
1649
1650 2006-04-10  Matthias Klose  <doko@debian.org>
1651
1652         * testsuite/lib/libffi-dg.exp (libffi-init): Recognize multilib
1653         directory names containing underscores.
1654
1655 2006-04-07  James E Wilson  <wilson@specifix.com>
1656
1657         * testsuite/libffi.call/float4.c: New testcase.
1658
1659 2006-04-05  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1660             Andreas Tobler  <a.tobler@schweiz.ch>
1661
1662         * Makefile.am: Add PA_HPUX port.
1663         * Makefile.in: Regenerate.
1664         * include/Makefile.in: Likewise.
1665         * testsuite/Makefile.in: Likewise.
1666         * configure.ac: Add PA_HPUX rules.
1667         * configure: Regenerate.
1668         * src/pa/ffitarget.h: Rename linux target to PA_LINUX.
1669         Add PA_HPUX and PA64_HPUX.
1670         Rename FFI_LINUX ABI to FFI_PA32 ABI.
1671         (FFI_TRAMPOLINE_SIZE): Define for 32-bit HP-UX targets.
1672         (FFI_TYPE_SMALL_STRUCT2): Define.
1673         (FFI_TYPE_SMALL_STRUCT4): Likewise.
1674         (FFI_TYPE_SMALL_STRUCT8): Likewise.
1675         (FFI_TYPE_SMALL_STRUCT3): Redefine.
1676         (FFI_TYPE_SMALL_STRUCT5): Likewise.
1677         (FFI_TYPE_SMALL_STRUCT6): Likewise.
1678         (FFI_TYPE_SMALL_STRUCT7): Likewise.
1679         * src/pa/ffi.c (ROUND_DOWN): Delete.
1680         (fldw, fstw, fldd, fstd): Use '__asm__'.
1681         (ffi_struct_type): Add support for FFI_TYPE_SMALL_STRUCT2,
1682         FFI_TYPE_SMALL_STRUCT4 and FFI_TYPE_SMALL_STRUCT8.
1683         (ffi_prep_args_LINUX): Rename to ffi_prep_args_pa32. Update comment.
1684         Simplify incrementing of stack slot variable. Change type of local
1685         'n' to unsigned int.
1686         (ffi_size_stack_LINUX): Rename to ffi_size_stack_pa32. Handle long
1687         double on PA_HPUX.
1688         (ffi_prep_cif_machdep): Likewise.
1689         (ffi_call): Likewise.
1690         (ffi_closure_inner_LINUX): Rename to ffi_closure_inner_pa32. Change
1691         return type to ffi_status. Simplify incrementing of stack slot
1692         variable. Only copy floating point argument registers when PA_LINUX
1693         is true. Reformat debug statement.
1694         Add support for FFI_TYPE_SMALL_STRUCT2, FFI_TYPE_SMALL_STRUCT4 and
1695         FFI_TYPE_SMALL_STRUCT8.
1696         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Add 'extern' to
1697         declaration.
1698         (ffi_prep_closure): Make linux trampoline conditional on PA_LINUX.
1699         Add nops to cache flush.  Add trampoline for PA_HPUX.
1700         * src/pa/hpux32.S: New file.
1701         * src/pa/linux.S (ffi_call_LINUX): Rename to ffi_call_pa32. Rename
1702         ffi_prep_args_LINUX to ffi_prep_args_pa32.
1703         Localize labels. Add support for 2, 4 and 8-byte small structs. Handle
1704         unaligned destinations in 3, 5, 6 and 7-byte small structs. Order
1705         argument type checks so that common argument types appear first.
1706         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Rename
1707         ffi_closure_inner_LINUX to ffi_closure_inner_pa32.
1708
1709 2006-03-24  Alan Modra  <amodra@bigpond.net.au>
1710
1711         * src/powerpc/ffitarget.h (enum ffi_abi): Add FFI_LINUX.  Default
1712         for 32-bit using IBM extended double format.  Fix FFI_LAST_ABI.
1713         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Handle linux variant of
1714         FFI_TYPE_LONGDOUBLE.
1715         (ffi_prep_args64): Assert using IBM extended double.
1716         (ffi_prep_cif_machdep): Don't munge FFI_TYPE_LONGDOUBLE type.
1717         Handle FFI_LINUX FFI_TYPE_LONGDOUBLE return and args.
1718         (ffi_call): Handle FFI_LINUX.
1719         (ffi_closure_helper_SYSV): Non FFI_LINUX long double return needs
1720         gpr3 return pointer as for struct return.  Handle FFI_LINUX
1721         FFI_TYPE_LONGDOUBLE return and args.  Don't increment "nf"
1722         unnecessarily.
1723         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Load both f1 and f2
1724         for FFI_TYPE_LONGDOUBLE.  Move epilogue insns into case table.
1725         Don't use r6 as pointer to results, instead use sp offset.  Don't
1726         make a special call to load lr with case table address, instead
1727         use offset from previous call.
1728         * src/powerpc/sysv.S (ffi_call_SYSV): Save long double return.
1729         * src/powerpc/linux64.S (ffi_call_LINUX64): Simplify long double
1730         return.
1731
1732 2006-03-15  Kaz Kojima  <kkojima@gcc.gnu.org>
1733
1734         * src/sh64/ffi.c (ffi_prep_cif_machdep): Handle float arguments
1735         passed with FP registers correctly.
1736         (ffi_closure_helper_SYSV): Likewise.
1737         * src/sh64/sysv.S: Likewise.
1738
1739 2006-03-01  Andreas Tobler  <a.tobler@schweiz.ch>
1740
1741         * testsuite/libffi.special/unwindtest.cc (closure_test_fn): Mark cif,
1742         args and userdata unused.
1743         (closure_test_fn1): Mark cif and userdata unused.
1744         (main): Remove unused res.
1745
1746 2006-02-28  Andreas Tobler  <a.tobler@schweiz.ch>
1747
1748         * testsuite/libffi.call/call.exp: Adjust FSF address. Add test runs for
1749         -O2, -O3, -Os and the warning flags -W -Wall.
1750         * testsuite/libffi.special/special.exp: Likewise.
1751         * testsuite/libffi.call/ffitest.h: Add an __UNUSED__ macro to mark
1752         unused parameter unused for gcc or else do nothing.
1753         * testsuite/libffi.special/ffitestcxx.h: Likewise.
1754         * testsuite/libffi.call/cls_12byte.c (cls_struct_12byte_gn): Mark cif
1755         and userdata unused.
1756         * testsuite/libffi.call/cls_16byte.c (cls_struct_16byte_gn): Likewise.
1757         * testsuite/libffi.call/cls_18byte.c (cls_struct_18byte_gn): Likewise.
1758         * testsuite/libffi.call/cls_19byte.c (cls_struct_19byte_gn): Likewise.
1759         * testsuite/libffi.call/cls_1_1byte.c (cls_struct_1_1byte_gn): Likewise.
1760         * testsuite/libffi.call/cls_20byte.c (cls_struct_20byte_gn): Likewise.
1761         * testsuite/libffi.call/cls_20byte1.c (cls_struct_20byte_gn): Likewise.
1762         * testsuite/libffi.call/cls_24byte.c (cls_struct_24byte_gn): Likewise.
1763         * testsuite/libffi.call/cls_2byte.c (cls_struct_2byte_gn): Likewise.
1764         * testsuite/libffi.call/cls_3_1byte.c (cls_struct_3_1byte_gn): Likewise.
1765         * testsuite/libffi.call/cls_3byte1.c (cls_struct_3byte_gn): Likewise.
1766         * testsuite/libffi.call/cls_3byte2.c (cls_struct_3byte_gn1): Likewise.
1767         * testsuite/libffi.call/cls_4_1byte.c (cls_struct_4_1byte_gn): Likewise.
1768         * testsuite/libffi.call/cls_4byte.c (cls_struct_4byte_gn): Likewise.
1769         * testsuite/libffi.call/cls_5_1_byte.c (cls_struct_5byte_gn): Likewise.
1770         * testsuite/libffi.call/cls_5byte.c (cls_struct_5byte_gn): Likewise.
1771         * testsuite/libffi.call/cls_64byte.c (cls_struct_64byte_gn): Likewise.
1772         * testsuite/libffi.call/cls_6_1_byte.c (cls_struct_6byte_gn): Likewise.
1773         * testsuite/libffi.call/cls_6byte.c (cls_struct_6byte_gn): Likewise.
1774         * testsuite/libffi.call/cls_7_1_byte.c (cls_struct_7byte_gn): Likewise.
1775         * testsuite/libffi.call/cls_7byte.c (cls_struct_7byte_gn): Likewise.
1776         * testsuite/libffi.call/cls_8byte.c (cls_struct_8byte_gn): Likewise.
1777         * testsuite/libffi.call/cls_9byte1.c (cls_struct_9byte_gn): Likewise.
1778         * testsuite/libffi.call/cls_9byte2.c (cls_struct_9byte_gn): Likewise.
1779         * testsuite/libffi.call/cls_align_double.c (cls_struct_align_gn):
1780         Likewise.
1781         * testsuite/libffi.call/cls_align_float.c (cls_struct_align_gn):
1782         Likewise.
1783         * testsuite/libffi.call/cls_align_longdouble.c (cls_struct_align_gn):
1784         Likewise.
1785         * testsuite/libffi.call/cls_align_pointer.c (cls_struct_align_fn): Cast
1786         void* to avoid compiler warning.
1787         (main): Likewise.
1788         (cls_struct_align_gn): Mark cif and userdata unused.
1789         * testsuite/libffi.call/cls_align_sint16.c (cls_struct_align_gn):
1790         Likewise.
1791         * testsuite/libffi.call/cls_align_sint32.c (cls_struct_align_gn):
1792         Likewise.
1793         * testsuite/libffi.call/cls_align_sint64.c (cls_struct_align_gn):
1794         Likewise.
1795         * testsuite/libffi.call/cls_align_uint16.c (cls_struct_align_gn):
1796         Likewise.
1797         * testsuite/libffi.call/cls_align_uint32.c (cls_struct_align_gn):
1798         Likewise.
1799         * testsuite/libffi.call/cls_double.c (cls_ret_double_fn): Likewise.
1800         * testsuite/libffi.call/cls_float.c (cls_ret_float_fn): Likewise.
1801         * testsuite/libffi.call/cls_multi_schar.c (test_func_gn): Mark cif and
1802         data unused.
1803         (main): Cast res_call to silence gcc.
1804         * testsuite/libffi.call/cls_multi_sshort.c (test_func_gn): Mark cif and
1805         data unused.
1806         (main): Cast res_call to silence gcc.
1807         * testsuite/libffi.call/cls_multi_sshortchar.c (test_func_gn): Mark cif
1808         and data unused.
1809         (main): Cast res_call to silence gcc.
1810         * testsuite/libffi.call/cls_multi_uchar.c (test_func_gn): Mark cif and
1811         data unused.
1812         (main): Cast res_call to silence gcc.
1813         * testsuite/libffi.call/cls_multi_ushort.c (test_func_gn): Mark cif and
1814         data unused.
1815         (main): Cast res_call to silence gcc.
1816         * testsuite/libffi.call/cls_multi_ushortchar.c (test_func_gn): Mark cif
1817         and data unused.
1818         (main): Cast res_call to silence gcc.
1819         * testsuite/libffi.call/cls_schar.c (cls_ret_schar_fn): Mark cif and
1820         userdata unused.
1821         (cls_ret_schar_fn): Cast printf parameter to silence gcc.
1822         * testsuite/libffi.call/cls_sint.c (cls_ret_sint_fn): Mark cif and
1823         userdata unused.
1824         (cls_ret_sint_fn): Cast printf parameter to silence gcc.
1825         * testsuite/libffi.call/cls_sshort.c (cls_ret_sshort_fn): Mark cif and
1826         userdata unused.
1827         (cls_ret_sshort_fn): Cast printf parameter to silence gcc.
1828         * testsuite/libffi.call/cls_uchar.c (cls_ret_uchar_fn):  Mark cif and
1829         userdata unused.
1830         (cls_ret_uchar_fn): Cast printf parameter to silence gcc.
1831         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Mark cif and
1832         userdata unused.
1833         (cls_ret_uint_fn): Cast printf parameter to silence gcc.
1834         * testsuite/libffi.call/cls_ulonglong.c (cls_ret_ulonglong_fn): Mark cif
1835         and userdata unused.
1836         * testsuite/libffi.call/cls_ushort.c (cls_ret_ushort_fn): Mark cif and
1837         userdata unused.
1838         (cls_ret_ushort_fn): Cast printf parameter to silence gcc.
1839         * testsuite/libffi.call/float.c (floating): Remove unused parameter e.
1840         * testsuite/libffi.call/float1.c (main): Remove unused variable i.
1841         Cleanup white spaces.
1842         * testsuite/libffi.call/negint.c (checking): Remove unused variable i.
1843         * testsuite/libffi.call/nested_struct.c (cls_struct_combined_gn): Mark
1844         cif and userdata unused.
1845         * testsuite/libffi.call/nested_struct1.c (cls_struct_combined_gn):
1846         Likewise.
1847         * testsuite/libffi.call/nested_struct10.c (B_gn): Likewise.
1848         * testsuite/libffi.call/nested_struct2.c (B_fn): Adjust printf
1849         formatters to silence gcc.
1850         (B_gn): Mark cif and userdata unused.
1851         * testsuite/libffi.call/nested_struct3.c (B_gn): Mark cif and userdata
1852         unused.
1853         * testsuite/libffi.call/nested_struct4.c: Mention related PR.
1854         (B_gn): Mark cif and userdata unused.
1855         * testsuite/libffi.call/nested_struct5.c (B_gn): Mark cif and userdata
1856         unused.
1857         * testsuite/libffi.call/nested_struct6.c: Mention related PR.
1858         (B_gn): Mark cif and userdata unused.
1859         * testsuite/libffi.call/nested_struct7.c (B_gn): Mark cif and userdata
1860         unused.
1861         * testsuite/libffi.call/nested_struct8.c (B_gn): Likewise.
1862         * testsuite/libffi.call/nested_struct9.c (B_gn): Likewise.
1863         * testsuite/libffi.call/problem1.c (stub): Likewise.
1864         * testsuite/libffi.call/pyobjc-tc.c (main): Cast the result to silence
1865         gcc.
1866         * testsuite/libffi.call/return_fl2.c (return_fl): Add the note mentioned
1867         in the last commit for this test case in the test case itself.
1868         * testsuite/libffi.call/closure_fn0.c (closure_test_fn0): Mark cif as
1869         unused.
1870         * testsuite/libffi.call/closure_fn1.c (closure_test_fn1): Likewise.
1871         * testsuite/libffi.call/closure_fn2.c (closure_test_fn2): Likewise.
1872         * testsuite/libffi.call/closure_fn3.c (closure_test_fn3): Likewise.
1873         * testsuite/libffi.call/closure_fn4.c (closure_test_fn0): Likewise.
1874         * testsuite/libffi.call/closure_fn5.c (closure_test_fn5): Likewise.
1875         * testsuite/libffi.call/closure_fn6.c (closure_test_fn0): Likewise.
1876
1877 2006-02-22  Kaz Kojima  <kkojima@gcc.gnu.org>
1878
1879         * src/sh/sysv.S: Fix register numbers in the FDE for
1880         ffi_closure_SYSV.
1881
1882 2006-02-20  Andreas Tobler  <a.tobler@schweiz.ch>
1883
1884         * testsuite/libffi.call/return_fl2.c (return_fl): Remove static
1885         declaration to avoid a false negative on ix86. See PR323.
1886
1887 2006-02-18  Kaz Kojima  <kkojima@gcc.gnu.org>
1888
1889         * src/sh/ffi.c (ffi_closure_helper_SYSV): Remove unused variable
1890         and cast integer to void * if needed.  Update the pointer to
1891         the FP register saved area correctly.
1892
1893 2006-02-17  Andreas Tobler  <a.tobler@schweiz.ch>
1894
1895         * testsuite/libffi.call/nested_struct6.c: XFAIL this test until PR25630
1896         is fixed.
1897         * testsuite/libffi.call/nested_struct4.c: Likewise.
1898
1899 2006-02-16  Andreas Tobler  <a.tobler@schweiz.ch>
1900
1901         * testsuite/libffi.call/return_dbl.c: New test case.
1902         * testsuite/libffi.call/return_dbl1.c: Likewise.
1903         * testsuite/libffi.call/return_dbl2.c: Likewise.
1904         * testsuite/libffi.call/return_fl.c: Likewise.
1905         * testsuite/libffi.call/return_fl1.c: Likewise.
1906         * testsuite/libffi.call/return_fl2.c: Likewise.
1907         * testsuite/libffi.call/return_fl3.c: Likewise.
1908         * testsuite/libffi.call/closure_fn6.c: Likewise.
1909
1910         * testsuite/libffi.call/nested_struct2.c: Remove ffi_type_mylong
1911         definition.
1912         * testsuite/libffi.call/ffitest.h: Add ffi_type_mylong definition
1913         here to be used by other test cases too.
1914
1915         * testsuite/libffi.call/nested_struct10.c: New test case.
1916         * testsuite/libffi.call/nested_struct9.c: Likewise.
1917         * testsuite/libffi.call/nested_struct8.c: Likewise.
1918         * testsuite/libffi.call/nested_struct7.c: Likewise.
1919         * testsuite/libffi.call/nested_struct6.c: Likewise.
1920         * testsuite/libffi.call/nested_struct5.c: Likewise.
1921         * testsuite/libffi.call/nested_struct4.c: Likewise.
1922
1923 2006-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
1924
1925         * configure.ac: Enable libffi for sparc64-*-freebsd*.
1926         * configure: Rebuilt.
1927
1928 2006-01-18  Jakub Jelinek  <jakub@redhat.com>
1929
1930         * src/powerpc/sysv.S (smst_two_register): Don't call __ashldi3,
1931         instead do the shifting inline.
1932         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't compute %r5
1933         shift count unconditionally.  Simplify load sequences for 1, 2, 3, 4
1934         and 8 byte structs, for the remaining struct sizes don't call
1935         __lshrdi3, instead do the shifting inline.
1936
1937 2005-12-07  Thiemo Seufer  <ths@networkno.de>
1938
1939         * src/mips/ffitarget.h: Remove obsolete sgidefs.h include. Add
1940         missing parentheses.
1941         * src/mips/o32.S (ffi_call_O32): Code formatting. Define
1942         and use A3_OFF, FP_OFF, RA_OFF. Micro-optimizations.
1943         (ffi_closure_O32): Likewise, but with newly defined A3_OFF2,
1944         A2_OFF2, A1_OFF2, A0_OFF2, RA_OFF2, FP_OFF2, S0_OFF2, GP_OFF2,
1945         V1_OFF2, V0_OFF2, FA_1_1_OFF2, FA_1_0_OFF2, FA_0_1_OFF2,
1946         FA_0_0_OFF2.
1947         * src/mips/ffi.c (ffi_prep_args): Code formatting. Fix
1948         endianness bugs.
1949         (ffi_prep_closure): Improve trampoline instruction scheduling.
1950         (ffi_closure_mips_inner_O32): Fix endianness bugs.
1951
1952 2005-12-03  Alan Modra  <amodra@bigpond.net.au>
1953
1954         * src/powerpc/ffi.c: Formatting.
1955         (ffi_prep_args_SYSV): Avoid possible aliasing problems by using unions.
1956         (ffi_prep_args64): Likewise.
1957
1958 2005-09-30  Geoffrey Keating  <geoffk@apple.com>
1959
1960         * testsuite/lib/libffi-dg.exp (libffi_target_compile): For
1961         darwin, use -shared-libgcc not -lgcc_s, and explain why.
1962
1963 2005-09-26  Tom Tromey  <tromey@redhat.com>
1964
1965         * testsuite/libffi.call/float1.c (value_type): New typedef.
1966         (CANARY): New define.
1967         (main): Check for result buffer overflow.
1968         * src/powerpc/linux64.S: Handle linux64 long double returns.
1969         * src/powerpc/ffi.c (FLAG_RETURNS_128BITS): New constant.
1970         (ffi_prep_cif_machdep): Handle linux64 long double returns.
1971
1972 2005-08-25  Alan Modra  <amodra@bigpond.net.au>
1973
1974         PR target/23404
1975         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Correct placement of stack
1976         homed fp args.
1977         (ffi_status ffi_prep_cif_machdep): Correct stack sizing for same.
1978
1979 2005-08-11  Jakub Jelinek  <jakub@redhat.com>
1980
1981         * configure.ac (HAVE_HIDDEN_VISIBILITY_ATTRIBUTE): New test.
1982         (AH_BOTTOM): Add FFI_HIDDEN definition.
1983         * configure: Rebuilt.
1984         * fficonfig.h.in: Rebuilt.
1985         * src/powerpc/ffi.c (hidden): Remove.
1986         (ffi_closure_LINUX64, ffi_prep_args64, ffi_call_LINUX64,
1987         ffi_closure_helper_LINUX64): Use FFI_HIDDEN instead of hidden.
1988         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64,
1989         .ffi_closure_LINUX64): Use FFI_HIDDEN instead of .hidden.
1990         * src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV): Remove,
1991         add FFI_HIDDEN to its prototype.
1992         (ffi_closure_SYSV_inner): New.
1993         * src/x86/sysv.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
1994         * src/x86/win32.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
1995
1996 2005-08-10  Alfred M. Szmidt  <ams@gnu.org>
1997
1998         PR libffi/21819:
1999         * configure: Rebuilt.
2000         * configure.ac: Handle i*86-*-gnu*.
2001
2002 2005-08-09  Jakub Jelinek  <jakub@redhat.com>
2003
2004         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Use
2005         DW_CFA_offset_extended_sf rather than
2006         DW_CFA_GNU_negative_offset_extended.
2007         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise.
2008
2009 2005-07-22  SUGIOKA Toshinobu  <sugioka@itonet.co.jp>
2010
2011         * src/sh/sysv.S (ffi_call_SYSV): Stop argument popping correctly
2012         on sh3.
2013         (ffi_closure_SYSV): Change the stack layout for sh3 struct argument.
2014         * src/sh/ffi.c (ffi_prep_args): Fix sh3 argument copy, when it is
2015         partially on register.
2016         (ffi_closure_helper_SYSV): Likewise.
2017         (ffi_prep_cif_machdep): Don't set too many cif->flags.
2018
2019 2005-07-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2020
2021         * src/sh/ffi.c (ffi_call): Handle small structures correctly.
2022         Remove empty line.
2023         * src/sh64/ffi.c (simple_type): Remove.
2024         (return_type): Handle small structures correctly.
2025         (ffi_prep_args): Likewise.
2026         (ffi_call): Likewise.
2027         (ffi_closure_helper_SYSV): Likewise.
2028         * src/sh64/sysv.S (ffi_call_SYSV): Handle 1, 2 and 4-byte return.
2029         Emit position independent code if PIC and remove wrong datalabel
2030         prefixes from EH data.
2031
2032 2005-07-19  Andreas Tobler  <a.tobler@schweiz.ch>
2033
2034         * Makefile.am (nodist_libffi_la_SOURCES): Add POWERPC_FREEBSD.
2035         * Makefile.in: Regenerate.
2036         * include/Makefile.in: Likewise.
2037         * testsuite/Makefile.in: Likewise.
2038         * configure.ac: Add POWERPC_FREEBSD rules.
2039         * configure: Regenerate.
2040         * src/powerpc/ffitarget.h: Add POWERPC_FREEBSD rules.
2041         (FFI_SYSV_TYPE_SMALL_STRUCT): Define.
2042         * src/powerpc/ffi.c: Add flags to handle small structure returns
2043         in ffi_call_SYSV.
2044         (ffi_prep_cif_machdep): Handle small structures for SYSV 4 ABI.
2045         Aka FFI_SYSV.
2046         (ffi_closure_helper_SYSV): Likewise.
2047         * src/powerpc/ppc_closure.S: Add return types for small structures.
2048         * src/powerpc/sysv.S: Add bits to handle small structures for
2049         final SYSV 4 ABI.
2050
2051 2005-07-10  Andreas Tobler  <a.tobler@schweiz.ch>
2052
2053         * testsuite/libffi.call/cls_5_1_byte.c: New test file.
2054         * testsuite/libffi.call/cls_6_1_byte.c: Likewise.
2055         * testsuite/libffi.call/cls_7_1_byte.c: Likewise.
2056
2057 2005-07-05  Randolph Chung  <tausq@debian.org>
2058
2059         * src/pa/ffi.c (ffi_struct_type): Rename FFI_TYPE_SMALL_STRUCT1
2060         as FFI_TYPE_SMALL_STRUCT3.  Break out handling for 5-7 byte
2061         structures.  Kill compilation warnings.
2062         (ffi_closure_inner_LINUX): Print return values as hex in debug
2063         message.  Rename FFI_TYPE_SMALL_STRUCT1 as FFI_TYPE_SMALL_STRUCT3.
2064         Properly handle 5-7 byte structure returns.
2065         * src/pa/ffitarget.h (FFI_TYPE_SMALL_STRUCT1)
2066         (FFI_TYPE_SMALL_STRUCT2): Remove.
2067         (FFI_TYPE_SMALL_STRUCT3, FFI_TYPE_SMALL_STRUCT5)
2068         (FFI_TYPE_SMALL_STRUCT6, FFI_TYPE_SMALL_STRUCT7): Define.
2069         * src/pa/linux.S: Mark source file as using PA1.1 assembly.
2070         (checksmst1, checksmst2): Remove.
2071         (checksmst3): Optimize handling of 3-byte struct returns.
2072         (checksmst567): Properly handle 5-7 byte struct returns.
2073
2074 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2075
2076         PR libgcj/21943
2077         * src/mips/n32.S: Enforce PIC code.
2078         * src/mips/o32.S: Likewise.
2079
2080 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2081
2082         * configure.ac: Treat i*86-*-solaris2.10 and up as X86_64.
2083         * configure: Regenerate.
2084
2085 2005-06-01  Alan Modra  <amodra@bigpond.net.au>
2086
2087         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't use JUMPTARGET
2088         to call ffi_closure_helper_SYSV.  Append @local instead.
2089         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise for ffi_prep_args_SYSV.
2090
2091 2005-05-17  Kelley Cook  <kcook@gcc.gnu.org>
2092
2093         * configure.ac: Use AC_C_BIGENDIAN instead of AC_C_BIGENDIAN_CROSS.
2094         Use AC_CHECK_SIZEOF instead of AC_COMPILE_CHECK_SIZEOF.
2095         * Makefile.am (ACLOCAL_AMFLAGS): Remove -I ../config.
2096         * aclocal.m4, configure, fficonfig.h.in, Makefile.in,
2097         include/Makefile.in, testsuite/Makefile.in: Regenerate.
2098
2099 2005-05-09  Mike Stump  <mrs@apple.com>
2100
2101         * configure: Regenerate.
2102
2103 2005-05-08  Richard Henderson  <rth@redhat.com>
2104
2105         PR libffi/21285
2106         * src/alpha/osf.S: Update unwind into to match code.
2107
2108 2005-05-04  Andreas Degert <ad@papyrus-gmbh.de>
2109             Richard Henderson  <rth@redhat.com>
2110
2111         * src/x86/ffi64.c (ffi_prep_cif_machdep): Save sse-used flag in
2112         bit 11 of flags.
2113         (ffi_call): Mask return type field.  Pass ssecount to ffi_call_unix64.
2114         (ffi_prep_closure): Set carry bit if sse-used flag set.
2115         * src/x86/unix64.S (ffi_call_unix64): Add ssecount argument.
2116         Only load sse registers if ssecount non-zero.
2117         (ffi_closure_unix64): Only save sse registers if carry set on entry.
2118
2119 2005-04-29  Ralf Corsepius  <ralf.corsepius@rtems.org>
2120
2121         * configure.ac: Add i*86-*-rtems*, sparc*-*-rtems*,
2122         powerpc-*rtems*, arm*-*-rtems*, sh-*-rtems*.
2123         * configure: Regenerate.
2124
2125 2005-04-20  Hans-Peter Nilsson  <hp@axis.com>
2126
2127         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): In regsub use,
2128         have Tcl8.3-compatible intermediate variable.
2129
2130 2005-04-18  Simon Posnjak <simon.posnjak@siol.net>
2131             Hans-Peter Nilsson  <hp@axis.com>
2132
2133         * Makefile.am: Add CRIS support.
2134         * configure.ac: Likewise.
2135         * Makefile.in, configure, testsuite/Makefile.in,
2136         include/Makefile.in: Regenerate.
2137         * src/cris: New directory.
2138         * src/cris/ffi.c, src/cris/sysv.S, src/cris/ffitarget.h: New files.
2139         * src/prep_cif.c (ffi_prep_cif): Wrap in #ifndef __CRIS__.
2140
2141         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): Replace \n with
2142         \r?\n in output tests.
2143
2144 2005-04-12  Mike Stump  <mrs@apple.com>
2145
2146         * configure: Regenerate.
2147
2148 2005-03-30  Hans Boehm  <Hans.Boehm@hp.com>
2149
2150         * src/ia64/ffitarget.h (ffi_arg): Use long long instead of DI.
2151
2152 2005-03-30  Steve Ellcey  <sje@cup.hp.com>
2153
2154         * src/ia64/ffitarget.h (ffi_arg) ADD DI attribute.
2155         (ffi_sarg) Ditto.
2156         * src/ia64/unix.S (ffi_closure_unix): Extend gp
2157         to 64 bits in ILP32 mode.
2158         Load 64 bits even for short data.
2159
2160 2005-03-23  Mike Stump  <mrs@apple.com>
2161
2162         * src/powerpc/darwin.S: Update for -m64 multilib.
2163         * src/powerpc/darwin_closure.S: Likewise.
2164
2165 2005-03-21  Zack Weinberg  <zack@codesourcery.com>
2166
2167         * configure.ac: Do not invoke TL_AC_GCC_VERSION.
2168         Do not set tool_include_dir.
2169         * aclocal.m4, configure, Makefile.in, testsuite/Makefile.in:
2170         Regenerate.
2171         * include/Makefile.am: Set gcc_version and toollibffidir.
2172         * include/Makefile.in: Regenerate.
2173
2174 2005-02-22  Andrew Haley  <aph@redhat.com>
2175
2176         * src/powerpc/ffi.c (ffi_prep_cif_machdep): Bump alignment to
2177         odd-numbered register pairs for 64-bit integer types.
2178
2179 2005-02-23  Andreas Tobler  <a.tobler@schweiz.ch>
2180
2181         PR libffi/20104
2182         * testsuite/libffi.call/return_ll1.c: New test case.
2183
2184 2005-02-11  Janis Johnson  <janis187@us.ibm.com>
2185
2186         * testsuite/libffi.call/cls_align_longdouble.c: Remove dg-options.
2187         * testsuite/libffi.call/float.c: Ditto.
2188         * testsuite/libffi.call/float2.c: Ditto.
2189         * testsuite/libffi.call/float3.c: Ditto.
2190
2191 2005-02-08  Andreas Tobler  <a.tobler@schweiz.ch>
2192
2193         * src/frv/ffitarget.h: Remove PPC stuff which does not belong to frv.
2194
2195 2005-01-12  Eric Botcazou  <ebotcazou@libertysurf.fr>
2196
2197         * testsuite/libffi.special/special.exp (cxx_options): Add
2198         -shared-libgcc.
2199
2200 2004-12-31  Richard Henderson  <rth@redhat.com>
2201
2202         * src/types.c (FFI_AGGREGATE_TYPEDEF): Remove.
2203         (FFI_TYPEDEF): Rename from FFI_INTEGRAL_TYPEDEF.  Replace size and
2204         offset parameters with a type parameter; deduce size and structure
2205         alignment.  Update all users.
2206
2207 2004-12-31  Richard Henderson  <rth@redhat.com>
2208
2209         * src/types.c (FFI_TYPE_POINTER): Define with sizeof.
2210         (FFI_TYPE_LONGDOUBLE): Fix for ia64.
2211         * src/ia64/ffitarget.h (struct ffi_ia64_trampoline_struct): Move
2212         into ffi_prep_closure.
2213         * src/ia64/ia64_flags.h, src/ia64/ffi.c, src/ia64/unix.S: Rewrite
2214         from scratch.
2215
2216 2004-12-27  Richard Henderson  <rth@redhat.com>
2217
2218         * src/x86/unix64.S: Fix typo in unwind info.
2219
2220 2004-12-25  Richard Henderson  <rth@redhat.com>
2221
2222         * src/x86/ffi64.c (struct register_args): Rename from stackLayout.
2223         (enum x86_64_reg_class): Add X86_64_COMPLEX_X87_CLASS.
2224         (merge_classes): Check for it.
2225         (SSE_CLASS_P): New.
2226         (classify_argument): Pass byte_offset by value; perform all updates
2227         inside struct case.
2228         (examine_argument): Add classes argument; handle
2229         X86_64_COMPLEX_X87_CLASS.
2230         (ffi_prep_args): Merge into ...
2231         (ffi_call): ... here.  Share stack frame with ffi_call_unix64.
2232         (ffi_prep_cif_machdep): Setup cif->flags for proper structure return.
2233         (ffi_fill_return_value): Remove.
2234         (ffi_prep_closure): Remove dead assert.
2235         (ffi_closure_unix64_inner): Rename from ffi_closure_UNIX64_inner.
2236         Rewrite to use struct register_args instead of va_list.  Create
2237         flags for handling structure returns.
2238         * src/x86/unix64.S: Remove dead strings.
2239         (ffi_call_unix64): Rename from ffi_call_UNIX64.  Rewrite to share
2240         stack frame with ffi_call.  Handle structure returns properly.
2241         (float2sse, floatfloat2sse, double2sse): Remove.
2242         (sse2float, sse2double, sse2floatfloat): Remove.
2243         (ffi_closure_unix64): Rename from ffi_closure_UNIX64.  Rewrite
2244         to handle structure returns properly.
2245
2246 2004-12-08  David Edelsohn  <edelsohn@gnu.org>
2247
2248         * Makefile.am (AM_MAKEFLAGS): Remove duplicate LIBCFLAGS and
2249         PICFLAG.
2250         * Makefile.in: Regenerated.
2251
2252 2004-12-02  Richard Sandiford  <rsandifo@redhat.com>
2253
2254         * configure.ac: Use TL_AC_GCC_VERSION to set gcc_version.
2255         * configure, aclocal.m4, Makefile.in: Regenerate.
2256         * include/Makefile.in, testsuite/Makefile.in: Regenerate.
2257
2258 2004-11-29  Kelley Cook  <kcook@gcc.gnu.org>
2259
2260         * configure: Regenerate for libtool change.
2261
2262 2004-11-25  Kelley Cook  <kcook@gcc.gnu.org>
2263
2264         * configure: Regenerate for libtool reversion.
2265
2266 2004-11-24  Kelley Cook  <kcook@gcc.gnu.org>
2267
2268         * configure: Regenerate for libtool change.
2269
2270 2004-11-23  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2271
2272         * testsuite/lib/libffi-dg.exp: Use new procs in target-libpath.exp.
2273
2274 2004-11-23  Richard Sandiford  <rsandifo@redhat.com>
2275
2276         * src/mips/o32.S (ffi_call_O32, ffi_closure_O32): Use jalr instead
2277         of jal.  Use an absolute encoding for the frame information.
2278
2279 2004-11-23  Kelley Cook  <kcook@gcc.gnu.org>
2280
2281         * Makefile.am: Remove no-dependencies.  Add ACLOCAL_AMFLAGS.
2282         * acinclude.m4: Delete logic for sincludes.
2283         * aclocal.m4, Makefile.in, configure: Regenerate.
2284         * include/Makefile: Likewise.
2285         * testsuite/Makefile: Likewise.
2286
2287 2004-11-22  Eric Botcazou  <ebotcazou@libertysurf.fr>
2288
2289         * src/sparc/ffi.c (ffi_prep_closure): Align doubles and 64-bit integers
2290         on a 8-byte boundary.
2291         * src/sparc/v8.S (ffi_closure_v8): Reserve frame space for arguments.
2292
2293 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2294
2295         * src/arm/ffi.c (ffi_prep_cif_machdep): Handle functions that return
2296         long long values.  Round stack allocation to a multiple of 8 bytes
2297         for ATPCS compatibility.
2298         * src/arm/sysv.S (ffi_call_SYSV): Rework to avoid use of APCS register
2299         names.  Handle returning long long types.  Add Thumb and interworking
2300         support.  Improve soft-float code.
2301
2302 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2303
2304         * testsuite/lib/libffi-db.exp (load_gcc_lib): New function.
2305         (libffi_exit): New function.
2306         (libffi_init): Build the testglue wrapper if needed.
2307
2308 2004-10-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2309
2310         PR other/18138
2311         * testsuite/lib/libffi-dg.exp: Accept more than one multilib libgcc.
2312
2313 2004-10-25  Kazuhiro Inaoka <inaoka.kazuhiro@renesas.com>
2314
2315         * src/m32r/libffitarget.h (FFI_CLOSURES): Set to 0.
2316
2317 2004-10-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2318
2319         * src/sh/sysv.S (ffi_call_SYSV): Don't align for double data.
2320         * testsuite/libffi.call/float3.c: New test case.
2321
2322 2004-10-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2323
2324         * src/sh/ffi.c (ffi_prep_closure): Set T bit in trampoline for
2325         the function returning a structure pointed with R2.
2326         * src/sh/sysv.S (ffi_closure_SYSV): Use R2 as the pointer to
2327         the structure return value if T bit set.  Emit position
2328         independent code and EH data if PIC.
2329
2330 2004-10-13  Kazuhiro Inaoka  <inaoka.kazuhiro@renesas.com>
2331
2332         * Makefile.am: Add m32r support.
2333         * configure.ac: Likewise.
2334         * Makefile.in: Regenerate.
2335         * confiugre: Regenerate.
2336         * src/types.c: Add m32r port to FFI_INTERNAL_TYPEDEF
2337         (uint64, sint64, double, longdouble)
2338         * src/m32r: New directory.
2339         * src/m32r/ffi.c: New file.
2340         * src/m32r/sysv.S: Likewise.
2341         * src/m32r/ffitarget.h: Likewise.
2342
2343 2004-10-02  Kaz Kojima  <kkojima@gcc.gnu.org>
2344
2345         * testsuite/libffi.call/negint.c: New test case.
2346
2347 2004-09-14  H.J. Lu  <hongjiu.lu@intel.com>
2348
2349         PR libgcj/17465
2350         * testsuite/lib/libffi-dg.exp: Don't use global ld_library_path.
2351         Set up LD_LIBRARY_PATH, SHLIB_PATH, LD_LIBRARYN32_PATH,
2352         LD_LIBRARY64_PATH, LD_LIBRARY_PATH_32, LD_LIBRARY_PATH_64 and
2353         DYLD_LIBRARY_PATH.
2354
2355 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2356
2357         * testsuite/libffi.call/many_win32.c: Remove whitespaces.
2358         * testsuite/libffi.call/promotion.c: Likewise.
2359         * testsuite/libffi.call/return_ll.c: Remove unused var. Cleanup
2360         whitespaces.
2361         * testsuite/libffi.call/return_sc.c: Likewise.
2362         * testsuite/libffi.call/return_uc.c: Likewise.
2363
2364 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2365
2366         * src/powerpc/darwin.S: Fix comments and identation.
2367         * src/powerpc/darwin_closure.S: Likewise.
2368
2369 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2370
2371         * src/powerpc/ffi_darwin.c: Add flag for longdouble return values.
2372         (ffi_prep_args): Handle longdouble arguments.
2373         (ffi_prep_cif_machdep): Set flags for longdouble. Calculate space for
2374         longdouble.
2375         (ffi_closure_helper_DARWIN): Add closure handling for longdouble.
2376         * src/powerpc/darwin.S (_ffi_call_DARWIN): Add handling of longdouble
2377         values.
2378         * src/powerpc/darwin_closure.S (_ffi_closure_ASM): Likewise.
2379         * src/types.c: Defined longdouble size and alignment for darwin.
2380
2381 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2382
2383         * src/powerpc/aix.S: Remove whitespaces.
2384         * src/powerpc/aix_closure.S: Likewise.
2385         * src/powerpc/asm.h: Likewise.
2386         * src/powerpc/ffi.c: Likewise.
2387         * src/powerpc/ffitarget.h: Likewise.
2388         * src/powerpc/linux64.S: Likewise.
2389         * src/powerpc/linux64_closure.S: Likewise.
2390         * src/powerpc/ppc_closure.S: Likewise.
2391         * src/powerpc/sysv.S: Likewise.
2392
2393 2004-08-30  Anthony Green  <green@redhat.com>
2394
2395         * Makefile.am: Add frv support.
2396         * Makefile.in, testsuite/Makefile.in: Rebuilt.
2397         * configure.ac: Read configure.host.
2398         * configure.in: Read configure.host.
2399         * configure.host: New file.  frv-elf needs libgloss.
2400         * include/ffi.h.in: Force ffi_closure to have a nice big (8)
2401         alignment.  This is needed to frv and shouldn't harm the others.
2402         * include/ffi_common.h (ALIGN_DOWN): New macro.
2403         * src/frv/ffi.c, src/frv/ffitarget.h, src/frv/eabi.S: New files.
2404
2405 2004-08-24  David Daney  <daney@avtrex.com>
2406
2407         * testsuite/libffi.call/closure_fn0.c: Xfail mips64* instead of mips*.
2408         * testsuite/libffi.call/closure_fn1.c: Likewise.
2409         * testsuite/libffi.call/closure_fn2.c  Likewise.
2410         * testsuite/libffi.call/closure_fn3.c: Likewise.
2411         * testsuite/libffi.call/closure_fn4.c: Likewise.
2412         * testsuite/libffi.call/closure_fn5.c: Likewise.
2413         * testsuite/libffi.call/cls_18byte.c: Likewise.
2414         * testsuite/libffi.call/cls_19byte.c: Likewise.
2415         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2416         * testsuite/libffi.call/cls_20byte.c: Likewise.
2417         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2418         * testsuite/libffi.call/cls_24byte.c: Likewise.
2419         * testsuite/libffi.call/cls_2byte.c: Likewise.
2420         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2421         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2422         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2423         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2424         * testsuite/libffi.call/cls_4byte.c: Likewise.
2425         * testsuite/libffi.call/cls_64byte.c: Likewise.
2426         * testsuite/libffi.call/cls_6byte.c: Likewise.
2427         * testsuite/libffi.call/cls_7byte.c: Likewise.
2428         * testsuite/libffi.call/cls_8byte.c: Likewise.
2429         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2430         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2431         * testsuite/libffi.call/cls_align_double.c: Likewise.
2432         * testsuite/libffi.call/cls_align_float.c: Likewise.
2433         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2434         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2435         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2436         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2437         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2438         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2439         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2440         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
2441         * testsuite/libffi.call/cls_double.c: Likewise.
2442         * testsuite/libffi.call/cls_float.c: Likewise.
2443         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2444         * testsuite/libffi.call/cls_multi_sshort.c: Likewise.
2445         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2446         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2447         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2448         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2449         * testsuite/libffi.call/cls_schar.c: Likewise.
2450         * testsuite/libffi.call/cls_sint.c: Likewise.
2451         * testsuite/libffi.call/cls_sshort.c: Likewise.
2452         * testsuite/libffi.call/cls_uchar.c: Likewise.
2453         * testsuite/libffi.call/cls_uint.c: Likewise.
2454         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2455         * testsuite/libffi.call/cls_ushort.c: Likewise.
2456         * testsuite/libffi.call/nested_struct.c: Likewise.
2457         * testsuite/libffi.call/nested_struct1.c: Likewise.
2458         * testsuite/libffi.call/nested_struct2.c: Likewise.
2459         * testsuite/libffi.call/nested_struct3.c: Likewise.
2460         * testsuite/libffi.call/problem1.c: Likewise.
2461         * testsuite/libffi.special/unwindtest.cc: Likewise.
2462         * testsuite/libffi.call/cls_12byte.c: Likewise and set return value
2463         to zero.
2464         * testsuite/libffi.call/cls_16byte.c: Likewise.
2465         * testsuite/libffi.call/cls_5byte.c: Likewise.
2466
2467 2004-08-23  David Daney <daney@avtrex.com>
2468
2469         PR libgcj/13141
2470         * src/mips/ffitarget.h (FFI_O32_SOFT_FLOAT): New ABI.
2471         * src/mips/ffi.c (ffi_prep_args): Fix alignment calculation.
2472         (ffi_prep_cif_machdep): Handle FFI_O32_SOFT_FLOAT floating point
2473         parameters and return types.
2474         (ffi_call): Handle FFI_O32_SOFT_FLOAT ABI.
2475         (ffi_prep_closure): Ditto.
2476         (ffi_closure_mips_inner_O32): Handle FFI_O32_SOFT_FLOAT ABI, fix
2477         alignment calculations.
2478         * src/mips/o32.S (ffi_closure_O32): Don't use floating point
2479         instructions if FFI_O32_SOFT_FLOAT, make stack frame ABI compliant.
2480
2481 2004-08-14  Casey Marshall <csm@gnu.org>
2482
2483         * src/mips/ffi.c (ffi_pref_cif_machdep): set `cif->flags' to
2484         contain `FFI_TYPE_UINT64' as return type for any 64-bit
2485         integer (O32 ABI only).
2486         (ffi_prep_closure): new function.
2487         (ffi_closure_mips_inner_O32): new function.
2488         * src/mips/ffitarget.h: Define `FFI_CLOSURES' and
2489         `FFI_TRAMPOLINE_SIZE' appropriately if the ABI is o32.
2490         * src/mips/o32.S (ffi_call_O32): add labels for .eh_frame. Return
2491         64 bit integers correctly.
2492         (ffi_closure_O32): new function.
2493         Added DWARF-2 unwind info for both functions.
2494
2495 2004-08-10  Andrew Haley  <aph@redhat.com>
2496
2497         * src/x86/ffi64.c (ffi_prep_args ): 8-align all stack arguments.
2498
2499 2004-08-01  Robert Millan  <robertmh@gnu.org>
2500
2501         * configure.ac: Detect knetbsd-gnu and kfreebsd-gnu.
2502         * configure: Regenerate.
2503
2504 2004-07-30  Maciej W. Rozycki  <macro@linux-mips.org>
2505
2506         * acinclude.m4 (AC_FUNC_MMAP_BLACKLIST): Check for <sys/mman.h>
2507         and mmap() explicitly instead of relying on preset autoconf cache
2508         variables.
2509         * aclocal.m4: Regenerate.
2510         * configure: Regenerate.
2511
2512 2004-07-11  Ulrich Weigand  <uweigand@de.ibm.com>
2513
2514         * src/s390/ffi.c (ffi_prep_args): Fix C aliasing violation.
2515         (ffi_check_float_struct): Remove unused prototype.
2516
2517 2004-06-30  Geoffrey Keating  <geoffk@apple.com>
2518
2519         * src/powerpc/ffi_darwin.c (flush_icache): ';' is a comment
2520         character on Darwin, use '\n\t' instead.
2521
2522 2004-06-26  Matthias Klose  <doko@debian.org>
2523
2524         * libtool-version: Fix typo in revision/age.
2525
2526 2004-06-17  Matthias Klose  <doko@debian.org>
2527
2528         * libtool-version: New.
2529         * Makefile.am (libffi_la_LDFLAGS): Use -version-info for soname.
2530         * Makefile.in: Regenerate.
2531
2532 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2533
2534         * Makefile.am: Remove useless multilib rules.
2535         * Makefile.in: Regenerate.
2536         * aclocal.m4: Regenerate with automake 1.8.5.
2537         * configure.ac: Remove useless multilib configury.
2538         * configure: Regenerate.
2539
2540 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2541
2542         * .cvsignore: New file.
2543
2544 2004-06-10  Jakub Jelinek  <jakub@redhat.com>
2545
2546         * src/ia64/unix.S (ffi_call_unix): Insert group barrier break
2547         fp_done.
2548         (ffi_closure_UNIX): Fix f14/f15 adjustment if FLOAT_SZ is ever
2549         changed from 8.
2550
2551 2004-06-06  Sean McNeil  <sean@mcneil.com>
2552
2553         * configure.ac: Add x86_64-*-freebsd* support.
2554         * configure: Regenerate.
2555
2556 2004-04-26  Joe Buck <jbuck@welsh-buck.org>
2557
2558         Bug 15093
2559         * configure.ac: Test for existence of mmap and sys/mman.h before
2560         checking blacklist.  Fix suggested by Jim Wilson.
2561         * configure: Regenerate.
2562
2563 2004-04-26  Matt Austern  <austern@apple.com>
2564
2565         * src/powerpc/darwin.S: Go through a non-lazy pointer for initial
2566         FDE location.
2567         * src/powerpc/darwin_closure.S: Likewise.
2568
2569 2004-04-24  Andreas Tobler  <a.tobler@schweiz.ch>
2570
2571         * testsuite/libffi.call/cls_multi_schar.c (main): Fix initialization
2572         error. Reported by Thomas Heller <theller@python.net>.
2573         * testsuite/libffi.call/cls_multi_sshort.c (main): Likewise.
2574         * testsuite/libffi.call/cls_multi_ushort.c (main): Likewise.
2575
2576 2004-03-20  Matthias Klose  <doko@debian.org>
2577
2578         * src/pa/linux.S: Fix typo.
2579
2580 2004-03-19  Matthias Klose  <doko@debian.org>
2581
2582         * Makefile.am: Update.
2583         * Makefile.in: Regenerate.
2584         * src/pa/ffi.h.in: Remove.
2585         * src/pa/ffitarget.h: New file.
2586
2587 2004-02-10  Randolph Chung  <tausq@debian.org>
2588
2589         * Makefile.am: Add PA support.
2590         * Makefile.in: Regenerate.
2591         * include/Makefile.in: Regenerate.
2592         * configure.ac: Add PA target.
2593         * configure: Regenerate.
2594         * src/pa/ffi.c: New file.
2595         * src/pa/ffi.h.in: Add PA support.
2596         * src/pa/linux.S: New file.
2597         * prep_cif.c: Add PA support.
2598
2599 2004-03-16  Hosaka Yuji  <hos@tamanegi.org>
2600
2601         * src/types.c: Fix alignment size of X86_WIN32 case int64 and
2602         double.
2603         * src/x86/ffi.c (ffi_prep_args): Replace ecif->cif->rtype->type
2604         with ecif->cif->flags.
2605         (ffi_call, ffi_prep_incoming_args_SYSV): Replace cif->rtype->type
2606         with cif->flags.
2607         (ffi_prep_cif_machdep): Add X86_WIN32 struct case.
2608         (ffi_closure_SYSV): Add 1 or 2-bytes struct case for X86_WIN32.
2609         * src/x86/win32.S (retstruct1b, retstruct2b, sc_retstruct1b,
2610         sc_retstruct2b): Add for 1 or 2-bytes struct case.
2611
2612 2004-03-15 Kelley Cook <kcook@gcc.gnu.org>
2613
2614         * configure.in: Rename file to ...
2615         * configure.ac: ... this.
2616         * fficonfig.h.in: Regenerate.
2617         * Makefile.in: Regenerate.
2618         * include/Makefile.in: Regenerate.
2619         * testsuite/Makefile.in: Regenerate.
2620
2621 2004-03-12  Matt Austern  <austern@apple.com>
2622
2623         * src/powerpc/darwin.S: Fix EH information so it corresponds to
2624         changes in EH format resulting from addition of linkonce support.
2625         * src/powerpc/darwin_closure.S: Likewise.
2626
2627 2004-03-11  Andreas Tobler  <a.tobler@schweiz.ch>
2628             Paolo Bonzini  <bonzini@gnu.org>
2629
2630         * Makefile.am (AUTOMAKE_OPTIONS): Set them.
2631         Remove VPATH. Remove rules for object files. Remove multilib support.
2632         (AM_CCASFLAGS): Add.
2633         * configure.in (AC_CONFIG_HEADERS): Relace AM_CONFIG_HEADER.
2634         (AC_PREREQ): Bump version to 2.59.
2635         (AC_INIT): Fill with version info and bug address.
2636         (ORIGINAL_LD_FOR_MULTILIBS): Remove.
2637         (AM_ENABLE_MULTILIB): Use this instead of AC_ARG_ENABLE.
2638         De-precious CC so that the right flags are passed down to multilibs.
2639         (AC_MSG_ERROR): Replace obsolete macro AC_ERROR.
2640         (AC_CONFIG_FILES): Replace obsolete macro AC_LINK_FILES.
2641         (AC_OUTPUT): Reorganize the output with AC_CONFIG_COMMANDS.
2642         * configure: Rebuilt.
2643         * aclocal.m4: Likewise.
2644         * Makefile.in, include/Makefile.in, testsuite/Makefile.in: Likewise.
2645         * fficonfig.h.in: Likewise.
2646
2647 2004-03-11  Andreas Schwab  <schwab@suse.de>
2648
2649         * src/ia64/ffi.c (ffi_prep_incoming_args_UNIX): Get floating point
2650         arguments from fp registers only for the first 8 parameter slots.
2651         Don't convert a float parameter when passed in memory.
2652
2653 2004-03-09  Hans-Peter Nilsson  <hp@axis.com>
2654
2655         * configure: Regenerate for config/accross.m4 correction.
2656
2657 2004-02-25  Matt Kraai  <kraai@alumni.cmu.edu>
2658
2659         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Change
2660         ecif->cif->bytes to bytes.
2661         (ffi_prep_cif_machdep): Add braces around nested if statement.
2662
2663 2004-02-09  Alan Modra  <amodra@bigpond.net.au>
2664
2665         * src/types.c (pointer): POWERPC64 has 8 byte pointers.
2666
2667         * src/powerpc/ffi.c (ffi_prep_args64): Correct long double handling.
2668         (ffi_closure_helper_LINUX64): Fix typo.
2669         * testsuite/libffi.call/cls_align_longdouble.c: Pass -mlong-double-128
2670         for powerpc64-*-*.
2671         * testsuite/libffi.call/float.c: Likewise.
2672         * testsuite/libffi.call/float2.c: Likewise.
2673
2674 2004-02-08  Alan Modra  <amodra@bigpond.net.au>
2675
2676         * src/powerpc/ffi.c (ffi_prep_cif_machdep <FFI_LINUX64>): Correct
2677         long double function return and long double arg handling.
2678         (ffi_closure_helper_LINUX64): Formatting.  Delete unused "ng" var.
2679         Use "end_pfr" instead of "nf".  Correct long double handling.
2680         Localise "temp".
2681         * src/powerpc/linux64.S (ffi_call_LINUX64): Save f2 long double
2682         return value.
2683         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Allocate
2684         space for long double return value.  Adjust stack frame and offsets.
2685         Load f2 long double return.
2686
2687 2004-02-07  Alan Modra  <amodra@bigpond.net.au>
2688
2689         * src/types.c: Use 16 byte long double for POWERPC64.
2690
2691 2004-01-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2692
2693         * src/sparc/ffi.c (ffi_prep_args_v9): Shift the parameter array
2694         when the structure return address is passed in %o0.
2695         (ffi_V9_return_struct): Rename into ffi_v9_layout_struct.
2696         (ffi_v9_layout_struct): Align the field following a nested structure
2697         on a word boundary.  Use memmove instead of memcpy.
2698         (ffi_call): Update call to ffi_V9_return_struct.
2699         (ffi_prep_closure): Define 'ctx' only for V8.
2700         (ffi_closure_sparc_inner): Clone into ffi_closure_sparc_inner_v8
2701         and ffi_closure_sparc_inner_v9.
2702         (ffi_closure_sparc_inner_v8): Return long doubles by reference.
2703         Always skip the structure return address.  For structures and long
2704         doubles, copy the argument directly.
2705         (ffi_closure_sparc_inner_v9): Skip the structure return address only
2706         if required.  Shift the maximum floating-point slot accordingly.  For
2707         big structures, copy the argument directly; otherwise, left-justify the
2708         argument and call ffi_v9_layout_struct to lay out the structure on
2709         the stack.
2710         * src/sparc/v8.S: Undef STACKFRAME before defining it.
2711         (ffi_closure_v8): Pass the structure return address.  Update call to
2712         ffi_closure_sparc_inner_v8.  Short-circuit FFI_TYPE_INT handling.
2713         Skip the 'unimp' insn when returning long doubles and structures.
2714         * src/sparc/v9.S: Undef STACKFRAME before defining it.
2715         (ffi_closure_v9): Increase the frame size by 2 words.  Short-circuit
2716         FFI_TYPE_INT handling.  Load structures both in integers and
2717         floating-point registers on return.
2718         * README: Update status of the SPARC port.
2719
2720 2004-01-24  Andreas Tobler  <a.tobler@schweiz.ch>
2721
2722         * testsuite/libffi.call/pyobjc-tc.c (main): Treat result value
2723         as of type ffi_arg.
2724         * testsuite/libffi.call/struct3.c (main): Fix CHECK.
2725
2726 2004-01-22  Ulrich Weigand  <uweigand@de.ibm.com>
2727
2728         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Treat result
2729         value as of type ffi_arg, not unsigned int.
2730
2731 2004-01-21  Michael Ritzert  <ritzert@t-online.de>
2732
2733         * ffi64.c (ffi_prep_args): Cast the RHS of an assignment instead
2734         of the LHS.
2735
2736 2004-01-12  Andreas Tobler  <a.tobler@schweiz.ch>
2737
2738         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_32 for
2739         Solaris.
2740
2741 2004-01-08  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2742
2743         * testsuite/libffi.call/ffitest.h (allocate_mmap): Cast MAP_FAILED
2744         to void *.
2745
2746 2003-12-10  Richard Henderson  <rth@redhat.com>
2747
2748         * testsuite/libffi.call/cls_align_pointer.c: Cast pointers to
2749         size_t instead of int.
2750
2751 2003-12-04  Hosaka Yuji  <hos@tamanegi.org>
2752
2753         * testsuite/libffi.call/many_win32.c: Include <float.h>.
2754         * testsuite/libffi.call/many_win32.c (main): Replace variable
2755         int i with unsigned long ul.
2756
2757         * testsuite/libffi.call/cls_align_uint64.c: New test case.
2758         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2759         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2760         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2761         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2762         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2763         * testsuite/libffi.call/cls_align_float.c: Likewise.
2764         * testsuite/libffi.call/cls_align_double.c: Likewise.
2765         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2766         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2767
2768 2003-12-02  Hosaka Yuji  <hos@tamanegi.org>
2769
2770         PR other/13221
2771         * src/x86/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV):
2772         Align arguments to 32 bits.
2773
2774 2003-12-01  Andreas Tobler  <a.tobler@schweiz.ch>
2775
2776         PR other/13221
2777         * testsuite/libffi.call/cls_multi_sshort.c: New test case.
2778         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2779         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2780         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2781         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2782         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2783
2784         * testsuite/libffi.special/unwindtest.cc: Cosmetics.
2785
2786 2003-11-26  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
2787
2788         * testsuite/libffi.call/ffitest.h: Include <fcntl.h>.
2789         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2790
2791 2003-11-22  Andreas Tobler  <a.tobler@schweiz.ch>
2792
2793         * Makefile.in: Rebuilt.
2794         * configure: Likewise.
2795         * testsuite/libffi.special/unwindtest.cc: Convert the mmap to
2796         the right type.
2797
2798 2003-11-21  Andreas Jaeger  <aj@suse.de>
2799             Andreas Tobler  <a.tobler@schweiz.ch>
2800
2801         * acinclude.m4: Add AC_FUNC_MMAP_BLACKLIST.
2802         * configure.in: Call AC_FUNC_MMAP_BLACKLIST.
2803         * Makefile.in: Rebuilt.
2804         * aclocal.m4: Likewise.
2805         * configure: Likewise.
2806         * fficonfig.h.in: Likewise.
2807         * testsuite/lib/libffi-dg.exp: Add include dir.
2808         * testsuite/libffi.call/ffitest.h: Add MMAP definitions.
2809         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2810         * testsuite/libffi.call/closure_fn0.c: Use MMAP functionality
2811         for ffi_closure if available.
2812         * testsuite/libffi.call/closure_fn1.c: Likewise.
2813         * testsuite/libffi.call/closure_fn2.c: Likewise.
2814         * testsuite/libffi.call/closure_fn3.c: Likewise.
2815         * testsuite/libffi.call/closure_fn4.c: Likewise.
2816         * testsuite/libffi.call/closure_fn5.c: Likewise.
2817         * testsuite/libffi.call/cls_12byte.c: Likewise.
2818         * testsuite/libffi.call/cls_16byte.c: Likewise.
2819         * testsuite/libffi.call/cls_18byte.c: Likewise.
2820         * testsuite/libffi.call/cls_19byte.c: Likewise.
2821         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2822         * testsuite/libffi.call/cls_20byte.c: Likewise.
2823         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2824         * testsuite/libffi.call/cls_24byte.c: Likewise.
2825         * testsuite/libffi.call/cls_2byte.c: Likewise.
2826         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2827         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2828         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2829         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2830         * testsuite/libffi.call/cls_4byte.c: Likewise.
2831         * testsuite/libffi.call/cls_5byte.c: Likewise.
2832         * testsuite/libffi.call/cls_64byte.c: Likewise.
2833         * testsuite/libffi.call/cls_6byte.c: Likewise.
2834         * testsuite/libffi.call/cls_7byte.c: Likewise.
2835         * testsuite/libffi.call/cls_8byte.c: Likewise.
2836         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2837         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2838         * testsuite/libffi.call/cls_double.c: Likewise.
2839         * testsuite/libffi.call/cls_float.c: Likewise.
2840         * testsuite/libffi.call/cls_schar.c: Likewise.
2841         * testsuite/libffi.call/cls_sint.c: Likewise.
2842         * testsuite/libffi.call/cls_sshort.c: Likewise.
2843         * testsuite/libffi.call/cls_uchar.c: Likewise.
2844         * testsuite/libffi.call/cls_uint.c: Likewise.
2845         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2846         * testsuite/libffi.call/cls_ushort.c: Likewise.
2847         * testsuite/libffi.call/nested_struct.c: Likewise.
2848         * testsuite/libffi.call/nested_struct1.c: Likewise.
2849         * testsuite/libffi.call/nested_struct2.c: Likewise.
2850         * testsuite/libffi.call/nested_struct3.c: Likewise.
2851         * testsuite/libffi.call/problem1.c: Likewise.
2852         * testsuite/libffi.special/unwindtest.cc: Likewise.
2853
2854 2003-11-20  Andreas Tobler  <a.tobler@schweiz.ch>
2855
2856         * testsuite/lib/libffi-dg.exp: Make the -lgcc_s conditional.
2857
2858 2003-11-19  Andreas Tobler  <a.tobler@schweiz.ch>
2859
2860         * testsuite/lib/libffi-dg.exp: Add DYLD_LIBRARY_PATH for darwin.
2861         Add -lgcc_s to additional flags.
2862
2863 2003-11-12  Andreas Tobler  <a.tobler@schweiz.ch>
2864
2865         * configure.in, include/Makefile.am: PR libgcj/11147, install
2866         the ffitarget.h header file in a gcc versioned and target
2867         dependent place.
2868         * configure: Regenerated.
2869         * Makefile.in, include/Makefile.in: Likewise.
2870         * testsuite/Makefile.in: Likewise.
2871
2872 2003-11-09  Andreas Tobler  <a.tobler@schweiz.ch>
2873
2874         * testsuite/libffi.call/closure_fn0.c: Print result and check
2875         with dg-output to make debugging easier.
2876         * testsuite/libffi.call/closure_fn1.c: Likewise.
2877         * testsuite/libffi.call/closure_fn2.c: Likewise.
2878         * testsuite/libffi.call/closure_fn3.c: Likewise.
2879         * testsuite/libffi.call/closure_fn4.c: Likewise.
2880         * testsuite/libffi.call/closure_fn5.c: Likewise.
2881         * testsuite/libffi.call/cls_12byte.c: Likewise.
2882         * testsuite/libffi.call/cls_16byte.c: Likewise.
2883         * testsuite/libffi.call/cls_18byte.c: Likewise.
2884         * testsuite/libffi.call/cls_19byte.c: Likewise.
2885         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2886         * testsuite/libffi.call/cls_20byte.c: Likewise.
2887         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2888         * testsuite/libffi.call/cls_24byte.c: Likewise.
2889         * testsuite/libffi.call/cls_2byte.c: Likewise.
2890         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2891         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2892         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2893         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2894         * testsuite/libffi.call/cls_4byte.c: Likewise.
2895         * testsuite/libffi.call/cls_5byte.c: Likewise.
2896         * testsuite/libffi.call/cls_64byte.c: Likewise.
2897         * testsuite/libffi.call/cls_6byte.c: Likewise.
2898         * testsuite/libffi.call/cls_7byte.c: Likewise.
2899         * testsuite/libffi.call/cls_8byte.c: Likewise.
2900         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2901         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2902         * testsuite/libffi.call/cls_double.c: Likewise.
2903         * testsuite/libffi.call/cls_float.c: Likewise.
2904         * testsuite/libffi.call/cls_schar.c: Likewise.
2905         * testsuite/libffi.call/cls_sint.c: Likewise.
2906         * testsuite/libffi.call/cls_sshort.c: Likewise.
2907         * testsuite/libffi.call/cls_uchar.c: Likewise.
2908         * testsuite/libffi.call/cls_uint.c: Likewise.
2909         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2910         * testsuite/libffi.call/cls_ushort.c: Likewise.
2911         * testsuite/libffi.call/problem1.c: Likewise.
2912
2913         * testsuite/libffi.special/unwindtest.cc: Make ffi_closure
2914         static.
2915
2916 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
2917
2918         * testsuite/libffi.call/cls_9byte2.c: New test case.
2919         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2920         * testsuite/libffi.call/cls_64byte.c: Likewise.
2921         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2922         * testsuite/libffi.call/cls_19byte.c: Likewise.
2923         * testsuite/libffi.call/cls_18byte.c: Likewise.
2924         * testsuite/libffi.call/closure_fn4.c: Likewise.
2925         * testsuite/libffi.call/closure_fn5.c: Likewise.
2926         * testsuite/libffi.call/cls_schar.c: Likewise.
2927         * testsuite/libffi.call/cls_sint.c: Likewise.
2928         * testsuite/libffi.call/cls_sshort.c: Likewise.
2929         * testsuite/libffi.call/nested_struct2.c: Likewise.
2930         * testsuite/libffi.call/nested_struct3.c: Likewise.
2931
2932 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
2933
2934         * testsuite/libffi.call/cls_double.c: Do a check on the result.
2935         * testsuite/libffi.call/cls_uchar.c: Likewise.
2936         * testsuite/libffi.call/cls_uint.c: Likewise.
2937         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2938         * testsuite/libffi.call/cls_ushort.c: Likewise.
2939         * testsuite/libffi.call/return_sc.c: Cleanup whitespaces.
2940
2941 2003-11-06  Andreas Tobler  <a.tobler@schweiz.ch>
2942
2943         * src/prep_cif.c (ffi_prep_cif): Move the validity check after
2944         the initialization.
2945
2946 2003-10-23  Andreas Tobler  <a.tobler@schweiz.ch>
2947
2948         * src/java_raw_api.c (ffi_java_ptrarray_to_raw): Replace
2949         FFI_ASSERT(FALSE) with FFI_ASSERT(0).
2950
2951 2003-10-22  David Daney  <ddaney@avtrex.com>
2952
2953         * src/mips/ffitarget.h: Replace undefined UINT32 and friends with
2954         __attribute__((__mode__(__SI__))) and friends.
2955
2956 2003-10-22  Andreas Schwab  <schwab@suse.de>
2957
2958         * src/ia64/ffi.c: Replace FALSE/TRUE with false/true.
2959
2960 2003-10-21  Andreas Tobler  <a.tobler@schweiz.ch>
2961
2962         * configure.in: AC_LINK_FILES(ffitarget.h).
2963         * configure: Regenerate.
2964         * Makefile.in: Likewise.
2965         * include/Makefile.in: Likewise.
2966         * testsuite/Makefile.in: Likewise.
2967         * fficonfig.h.in: Likewise.
2968
2969 2003-10-21  Paolo Bonzini  <bonzini@gnu.org>
2970             Richard Henderson  <rth@redhat.com>
2971
2972         Avoid that ffi.h includes fficonfig.h.
2973
2974         * Makefile.am (EXTRA_DIST): Include ffitarget.h files
2975         (TARGET_SRC_MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
2976         (TARGET_SRC_MIPS_SGI): Removed.
2977         (MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
2978         (MIPS_SGI): Removed.
2979         (CLEANFILES): Removed.
2980         (mostlyclean-am, clean-am, mostlyclean-sub, clean-sub): New
2981         targets.
2982         * acconfig.h: Removed.
2983         * configure.in: Compute sizeofs only for double and long double.
2984         Use them to define and subst HAVE_LONG_DOUBLE.  Include comments
2985         into AC_DEFINE instead of using acconfig.h.  Create
2986         include/ffitarget.h instead of include/fficonfig.h.  Rename
2987         MIPS_GCC to MIPS_IRIX, drop MIPS_SGI since we are in gcc's tree.
2988         AC_DEFINE EH_FRAME_FLAGS.
2989         * include/Makefile.am (DISTCLEANFILES): New automake macro.
2990         (hack_DATA): Add ffitarget.h.
2991         * include/ffi.h.in: Remove all system specific definitions.
2992         Declare raw API even if it is not installed, why bother?
2993         Use limits.h instead of SIZEOF_* to define ffi_type_*.  Do
2994         not define EH_FRAME_FLAGS, it is in fficonfig.h now.  Include
2995         ffitarget.h instead of fficonfig.h.  Remove ALIGN macro.
2996         (UINT_ARG, INT_ARG): Removed, use ffi_arg and ffi_sarg instead.
2997         * include/ffi_common.h (bool): Do not define.
2998         (ffi_assert): Accept failed assertion.
2999         (ffi_type_test): Return void and accept file/line.
3000         (FFI_ASSERT): Pass stringized failed assertion.
3001         (FFI_ASSERT_AT): New macro.
3002         (FFI_ASSERT_VALID_TYPE): New macro.
3003         (UINT8, SINT8, UINT16, SINT16, UINT32, SINT32,
3004         UINT64, SINT64): Define here with gcc's __attribute__ macro
3005         instead of in ffi.h
3006         (FLOAT32, ALIGN): Define here instead of in ffi.h
3007         * include/ffi-mips.h: Removed.  Its content moved to
3008         src/mips/ffitarget.h after separating assembly and C sections.
3009         * src/alpha/ffi.c, src/alpha/ffi.c, src/java_raw_api.c
3010         src/prep_cif.c, src/raw_api.c, src/ia64/ffi.c,
3011         src/mips/ffi.c, src/mips/n32.S, src/mips/o32.S,
3012         src/mips/ffitarget.h, src/sparc/ffi.c, src/x86/ffi64.c:
3013         SIZEOF_ARG -> FFI_SIZEOF_ARG.
3014         * src/ia64/ffi.c: Include stdbool.h (provided by GCC 2.95+).
3015         * src/debug.c (ffi_assert): Accept stringized failed assertion.
3016         (ffi_type_test): Rewritten.
3017         * src/prep-cif.c (initialize_aggregate, ffi_prep_cif): Call
3018         FFI_ASSERT_VALID_TYPE.
3019         * src/alpha/ffitarget.h, src/arm/ffitarget.h,
3020         src/ia64/ffitarget.h, src/m68k/ffitarget.h,
3021         src/mips/ffitarget.h, src/powerpc/ffitarget.h,
3022         src/s390/ffitarget.h, src/sh/ffitarget.h,
3023         src/sh64/ffitarget.h, src/sparc/ffitarget.h,
3024         src/x86/ffitarget.h: New files.
3025         * src/alpha/osf.S, src/arm/sysv.S, src/ia64/unix.S,
3026         src/m68k/sysv.S, src/mips/n32.S, src/mips/o32.S,
3027         src/powerpc/aix.S, src/powerpc/darwin.S,
3028         src/powerpc/ffi_darwin.c, src/powerpc/linux64.S,
3029         src/powerpc/linux64_closure.S, src/powerpc/ppc_closure.S,
3030         src/powerpc/sysv.S, src/s390/sysv.S, src/sh/sysv.S,
3031         src/sh64/sysv.S, src/sparc/v8.S, src/sparc/v9.S,
3032         src/x86/sysv.S, src/x86/unix64.S, src/x86/win32.S:
3033         include fficonfig.h
3034
3035 2003-10-20  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3036
3037         * src/mips/ffi.c: Use _ABIN32, _ABIO32 instead of external
3038         _MIPS_SIM_NABI32, _MIPS_SIM_ABI32.
3039
3040 2003-10-19  Andreas Tobler  <a.tobler@schweiz.ch>
3041
3042         * src/powerpc/ffi_darwin.c (ffi_prep_args): Declare bytes again.
3043         Used when FFI_DEBUG = 1.
3044
3045 2003-10-14  Alan Modra  <amodra@bigpond.net.au>
3046
3047         * src/types.c (double, longdouble): Default POWERPC64 to 8 byte size
3048         and align.
3049
3050 2003-10-06  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3051
3052         * include/ffi_mips.h: Define FFI_MIPS_N32 for N32/N64 ABIs,
3053         FFI_MIPS_O32 for O32 ABI.
3054
3055 2003-10-01  Andreas Tobler  <a.tobler@schweiz.ch>
3056
3057         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_64 for
3058         SPARC64. Cleanup whitespaces.
3059
3060 2003-09-19  Andreas Tobler  <a.tobler@schweiz.ch>
3061
3062         * testsuite/libffi.call/closure_fn0.c: Xfail mips, arm,
3063         strongarm, xscale. Cleanup whitespaces.
3064         * testsuite/libffi.call/closure_fn1.c: Likewise.
3065         * testsuite/libffi.call/closure_fn2.c: Likewise.
3066         * testsuite/libffi.call/closure_fn3.c: Likewise.
3067         * testsuite/libffi.call/cls_12byte.c: Likewise.
3068         * testsuite/libffi.call/cls_16byte.c: Likewise.
3069         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3070         * testsuite/libffi.call/cls_20byte.c: Likewise.
3071         * testsuite/libffi.call/cls_24byte.c: Likewise.
3072         * testsuite/libffi.call/cls_2byte.c: Likewise.
3073         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3074         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3075         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3076         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3077         * testsuite/libffi.call/cls_4byte.c: Likewise.
3078         * testsuite/libffi.call/cls_5byte.c: Likewise.
3079         * testsuite/libffi.call/cls_6byte.c: Likewise.
3080         * testsuite/libffi.call/cls_7byte.c: Likewise.
3081         * testsuite/libffi.call/cls_8byte.c: Likewise.
3082         * testsuite/libffi.call/cls_double.c: Likewise.
3083         * testsuite/libffi.call/cls_float.c: Likewise.
3084         * testsuite/libffi.call/cls_uchar.c: Likewise.
3085         * testsuite/libffi.call/cls_uint.c: Likewise.
3086         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3087         * testsuite/libffi.call/cls_ushort.c: Likewise.
3088         * testsuite/libffi.call/nested_struct.c: Likewise.
3089         * testsuite/libffi.call/nested_struct1.c: Likewise.
3090         * testsuite/libffi.call/problem1.c: Likewise.
3091         * testsuite/libffi.special/unwindtest.cc: Likewise.
3092         * testsuite/libffi.call/pyobjc-tc.c: Cleanup whitespaces.
3093
3094 2003-09-18  David Edelsohn  <edelsohn@gnu.org>
3095
3096         * src/powerpc/aix.S: Cleanup whitespaces.
3097         * src/powerpc/aix_closure.S: Likewise.
3098
3099 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3100
3101         * src/powerpc/darwin.S: Cleanup whitespaces, comment formatting.
3102         * src/powerpc/darwin_closure.S: Likewise.
3103         * src/powerpc/ffi_darwin.c: Likewise.
3104
3105 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3106             David Edelsohn  <edelsohn@gnu.org>
3107
3108         * src/types.c (double): Add AIX and Darwin to the right TYPEDEF.
3109         * src/powerpc/aix_closure.S: Remove the pointer to the outgoing
3110         parameter stack.
3111         * src/powerpc/darwin_closure.S: Likewise.
3112         * src/powerpc/ffi_darwin.c (ffi_prep_args): Handle structures
3113         according to the Darwin/AIX ABI.
3114         (ffi_prep_cif_machdep): Likewise.
3115         (ffi_closure_helper_DARWIN): Likewise.
3116         Remove the outgoing parameter stack logic. Simplify the evaluation
3117         of the different CASE types.
3118         (ffi_prep_clousure): Avoid the casts on lvalues. Change the branch
3119         statement in the trampoline code.
3120
3121 2003-09-18  Kaz Kojima  <kkojima@gcc.gnu.org>
3122
3123         * src/sh/ffi.c (ffi_prep_args): Take account into the alignement
3124         for the register size.
3125         (ffi_closure_helper_SYSV): Handle the structure return value
3126         address correctly.
3127         (ffi_closure_helper_SYSV): Return the appropriate type when
3128         the registers are used for the structure return value.
3129         * src/sh/sysv.S (ffi_closure_SYSV): Fix the stack layout for
3130         the 64-bit return value.  Update copyright years.
3131
3132 2003-09-17  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3133
3134         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Search in
3135         srcdir for ffi_mips.h.
3136
3137 2003-09-12  Alan Modra  <amodra@bigpond.net.au>
3138
3139         * src/prep_cif.c (initialize_aggregate): Include tail padding in
3140         structure size.
3141         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Correct
3142         placement of float result.
3143         * testsuite/libffi.special/unwindtest.cc (closure_test_fn1): Correct
3144         cast of "resp" for big-endian 64 bit machines.
3145
3146 2003-09-11  Alan Modra  <amodra@bigpond.net.au>
3147
3148         * src/types.c (double, longdouble): Merge identical SH and ARM
3149         typedefs, and add POWERPC64.
3150         * src/powerpc/ffi.c (ffi_prep_args64): Correct next_arg calc for
3151         struct split over gpr and rest.
3152         (ffi_prep_cif_machdep): Correct intarg_count for structures.
3153         * src/powerpc/linux64.S (ffi_call_LINUX64): Fix gpr offsets.
3154
3155 2003-09-09  Andreas Tobler  <a.tobler@schweiz.ch>
3156
3157         * src/powerpc/ffi.c (ffi_closure_helper_SYSV) Handle struct
3158         passing correctly.
3159
3160 2003-09-09  Alan Modra  <amodra@bigpond.net.au>
3161
3162         * configure: Regenerate.
3163
3164 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3165
3166         * Makefile.am: Remove build rules for ffitest.
3167         * Makefile.in: Rebuilt.
3168
3169 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3170
3171         * src/java_raw_api.c: Include <stdlib.h> to fix compiler warning
3172         about implicit declaration of abort().
3173
3174 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3175
3176         * Makefile.am: Add dejagnu test framework. Fixes PR other/11411.
3177         * Makefile.in: Rebuilt.
3178         * configure.in: Add dejagnu test framework.
3179         * configure: Rebuilt.
3180
3181         * testsuite/Makefile.am: New file.
3182         * testsuite/Makefile.in: Built
3183         * testsuite/lib/libffi-dg.exp: New file.
3184         * testsuite/config/default.exp: Likewise.
3185         * testsuite/libffi.call/call.exp: Likewise.
3186         * testsuite/libffi.call/ffitest.h: Likewise.
3187         * testsuite/libffi.call/closure_fn0.c: Likewise.
3188         * testsuite/libffi.call/closure_fn1.c: Likewise.
3189         * testsuite/libffi.call/closure_fn2.c: Likewise.
3190         * testsuite/libffi.call/closure_fn3.c: Likewise.
3191         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3192         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3193         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3194         * testsuite/libffi.call/cls_2byte.c: Likewise.
3195         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3196         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3197         * testsuite/libffi.call/cls_4byte.c: Likewise.
3198         * testsuite/libffi.call/cls_5byte.c: Likewise.
3199         * testsuite/libffi.call/cls_6byte.c: Likewise.
3200         * testsuite/libffi.call/cls_7byte.c: Likewise.
3201         * testsuite/libffi.call/cls_8byte.c: Likewise.
3202         * testsuite/libffi.call/cls_12byte.c: Likewise.
3203         * testsuite/libffi.call/cls_16byte.c: Likewise.
3204         * testsuite/libffi.call/cls_20byte.c: Likewise.
3205         * testsuite/libffi.call/cls_24byte.c: Likewise.
3206         * testsuite/libffi.call/cls_double.c: Likewise.
3207         * testsuite/libffi.call/cls_float.c: Likewise.
3208         * testsuite/libffi.call/cls_uchar.c: Likewise.
3209         * testsuite/libffi.call/cls_uint.c: Likewise.
3210         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3211         * testsuite/libffi.call/cls_ushort.c: Likewise.
3212         * testsuite/libffi.call/float.c: Likewise.
3213         * testsuite/libffi.call/float1.c: Likewise.
3214         * testsuite/libffi.call/float2.c: Likewise.
3215         * testsuite/libffi.call/many.c: Likewise.
3216         * testsuite/libffi.call/many_win32.c: Likewise.
3217         * testsuite/libffi.call/nested_struct.c: Likewise.
3218         * testsuite/libffi.call/nested_struct1.c: Likewise.
3219         * testsuite/libffi.call/pyobjc-tc.c: Likewise.
3220         * testsuite/libffi.call/problem1.c: Likewise.
3221         * testsuite/libffi.call/promotion.c: Likewise.
3222         * testsuite/libffi.call/return_ll.c: Likewise.
3223         * testsuite/libffi.call/return_sc.c: Likewise.
3224         * testsuite/libffi.call/return_uc.c: Likewise.
3225         * testsuite/libffi.call/strlen.c: Likewise.
3226         * testsuite/libffi.call/strlen_win32.c: Likewise.
3227         * testsuite/libffi.call/struct1.c: Likewise.
3228         * testsuite/libffi.call/struct2.c: Likewise.
3229         * testsuite/libffi.call/struct3.c: Likewise.
3230         * testsuite/libffi.call/struct4.c: Likewise.
3231         * testsuite/libffi.call/struct5.c: Likewise.
3232         * testsuite/libffi.call/struct6.c: Likewise.
3233         * testsuite/libffi.call/struct7.c: Likewise.
3234         * testsuite/libffi.call/struct8.c: Likewise.
3235         * testsuite/libffi.call/struct9.c: Likewise.
3236         * testsuite/libffi.special/special.exp: New file.
3237         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3238         * testsuite/libffi.special/unwindtest.cc: Likewise.
3239
3240
3241 2003-08-13  Kaz Kojima  <kkojima@gcc.gnu.org>
3242
3243         * src/sh/ffi.c (OFS_INT16): Set 0 for little endian case.  Update
3244         copyright years.
3245
3246 2003-08-02  Alan Modra  <amodra@bigpond.net.au>
3247
3248         * src/powerpc/ffi.c (ffi_prep_args64): Modify for changed gcc
3249         structure passing.
3250         (ffi_closure_helper_LINUX64): Likewise.
3251         * src/powerpc/linux64.S: Remove code writing to parm save area.
3252         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Use return
3253         address in lr from ffi_closure_helper_LINUX64 call to calculate
3254         table address.  Optimize function tail.
3255
3256 2003-07-28  Andreas Tobler  <a.tobler@schweiz.ch>
3257
3258         * src/sparc/ffi.c: Handle all floating point registers.
3259         * src/sparc/v9.S: Likewise. Fixes second part of PR target/11410.
3260
3261 2003-07-11  Gerald Pfeifer  <pfeifer@dbai.tuwien.ac.at>
3262
3263         * README: Note that libffi is not part of GCC.  Update the project
3264         URL and status.
3265
3266 2003-06-19  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3267
3268         * src/powerpc/ppc_closure.S: Include ffi.h.
3269
3270 2003-06-13  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3271
3272         * src/x86/sysv.S: Avoid gas-only .uleb128/.sleb128 directives.
3273         Use C style comments.
3274
3275 2003-06-13  Kaz Kojima  <kkojima@rr.iij4u.or.jp>
3276
3277         * Makefile.am: Add SHmedia support.  Fix a typo of SH support.
3278         * Makefile.in: Regenerate.
3279         * configure.in (sh64-*-linux*, sh5*-*-linux*): Add target.
3280         * configure: Regenerate.
3281         * include/ffi.h.in: Add SHmedia support.
3282         * src/sh64/ffi.c: New file.
3283         * src/sh64/sysv.S: New file.
3284
3285 2003-05-16  Jakub Jelinek  <jakub@redhat.com>
3286
3287         * configure.in (HAVE_RO_EH_FRAME): Check whether .eh_frame section
3288         should be read-only.
3289         * configure: Rebuilt.
3290         * fficonfig.h.in: Rebuilt.
3291         * include/ffi.h.in (EH_FRAME_FLAGS): Define.
3292         * src/alpha/osf.S: Use EH_FRAME_FLAGS.
3293         * src/powerpc/linux64.S: Likewise.
3294         * src/powerpc/linux64_closure.S: Likewise.  Include ffi.h.
3295         * src/powerpc/sysv.S: Use EH_FRAME_FLAGS.  Use pcrel encoding
3296         if -fpic/-fPIC/-mrelocatable.
3297         * src/powerpc/powerpc_closure.S: Likewise.
3298         * src/sparc/v8.S: If HAVE_RO_EH_FRAME is defined, don't include
3299         #write in .eh_frame flags.
3300         * src/sparc/v9.S: Likewise.
3301         * src/x86/unix64.S: Use EH_FRAME_FLAGS.
3302         * src/x86/sysv.S: Likewise.  Use pcrel encoding if -fpic/-fPIC.
3303         * src/s390/sysv.S: Use EH_FRAME_FLAGS.  Include ffi.h.
3304
3305 2003-05-07  Jeff Sturm  <jsturm@one-point.com>
3306
3307         Fixes PR bootstrap/10656
3308         * configure.in (HAVE_AS_REGISTER_PSEUDO_OP): Test assembler
3309         support for .register pseudo-op.
3310         * src/sparc/v8.S: Use it.
3311         * fficonfig.h.in: Rebuilt.
3312         * configure: Rebuilt.
3313
3314 2003-04-18  Jakub Jelinek  <jakub@redhat.com>
3315
3316         * include/ffi.h.in (POWERPC64): Define if 64-bit.
3317         (enum ffi_abi): Add FFI_LINUX64 on POWERPC.
3318         Make it the default on POWERPC64.
3319         (FFI_TRAMPOLINE_SIZE): Define to 24 on POWERPC64.
3320         * configure.in: Change powerpc-*-linux* into powerpc*-*-linux*.
3321         * configure: Rebuilt.
3322         * src/powerpc/ffi.c (hidden): Define.
3323         (ffi_prep_args_SYSV): Renamed from
3324         ffi_prep_args.  Cast pointers to unsigned long to shut up warnings.
3325         (NUM_GPR_ARG_REGISTERS64, NUM_FPR_ARG_REGISTERS64,
3326         ASM_NEEDS_REGISTERS64): New.
3327         (ffi_prep_args64): New function.
3328         (ffi_prep_cif_machdep): Handle FFI_LINUX64 ABI.
3329         (ffi_call): Likewise.
3330         (ffi_prep_closure): Likewise.
3331         (flush_icache): Surround by #ifndef POWERPC64.
3332         (ffi_dblfl): New union type.
3333         (ffi_closure_helper_SYSV): Use it to avoid aliasing problems.
3334         (ffi_closure_helper_LINUX64): New function.
3335         * src/powerpc/ppc_closure.S: Surround whole file by #ifndef
3336         __powerpc64__.
3337         * src/powerpc/sysv.S: Likewise.
3338         (ffi_call_SYSV): Rename ffi_prep_args to ffi_prep_args_SYSV.
3339         * src/powerpc/linux64.S: New file.
3340         * src/powerpc/linux64_closure.S: New file.
3341         * Makefile.am (EXTRA_DIST): Add src/powerpc/linux64.S and
3342         src/powerpc/linux64_closure.S.
3343         (TARGET_SRC_POWERPC): Likewise.
3344
3345         * src/ffitest.c (closure_test_fn, closure_test_fn1, closure_test_fn2,
3346         closure_test_fn3): Fix result printing on big-endian 64-bit
3347         machines.
3348         (main): Print tst2_arg instead of uninitialized tst2_result.
3349
3350         * src/ffitest.c (main): Hide what closure pointer really points to
3351         from the compiler.
3352
3353 2003-04-16  Richard Earnshaw  <rearnsha@arm.com>
3354
3355         * configure.in (arm-*-netbsdelf*): Add configuration.
3356         (configure): Regenerated.
3357
3358 2003-04-04  Loren J. Rittle  <ljrittle@acm.org>
3359
3360         * include/Makefile.in: Regenerate.
3361
3362 2003-03-21  Zdenek Dvorak  <rakdver@atrey.karlin.mff.cuni.cz>
3363
3364         * libffi/include/ffi.h.in: Define X86 instead of X86_64 in 32
3365         bit mode.
3366         * libffi/src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV):
3367         Receive closure pointer through parameter, read args using
3368         __builtin_dwarf_cfa.
3369         (FFI_INIT_TRAMPOLINE): Send closure reference through eax.
3370
3371 2003-03-12  Andreas Schwab  <schwab@suse.de>
3372
3373         * configure.in: Avoid trailing /. in toolexeclibdir.
3374         * configure: Rebuilt.
3375
3376 2003-03-03  Andreas Tobler <a.tobler@schweiz.ch>
3377
3378         * src/powerpc/darwin_closure.S: Recode to fit dynamic libraries.
3379
3380 2003-02-06  Andreas Tobler <a.tobler@schweiz.ch>
3381
3382         * libffi/src/powerpc/darwin_closure.S:
3383         Fix alignement bug, allocate 8 bytes for the result.
3384         * libffi/src/powerpc/aix_closure.S:
3385         Likewise.
3386         * libffi/src/powerpc/ffi_darwin.c:
3387         Update stackframe description for aix/darwin_closure.S.
3388
3389 2003-02-06  Jakub Jelinek  <jakub@redhat.com>
3390
3391         * src/s390/ffi.c (ffi_closure_helper_SYSV): Add hidden visibility
3392         attribute.
3393
3394 2003-01-31  Christian Cornelssen  <ccorn@cs.tu-berlin.de>,
3395             Andreas Schwab  <schwab@suse.de>
3396
3397         * configure.in: Adjust command to source config-ml.in to account
3398         for changes to the libffi_basedir definition.
3399         (libffi_basedir): Remove ${srcdir} from value and include trailing
3400         slash if nonempty.
3401
3402         * configure: Regenerate.
3403
3404 2003-01-29  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3405
3406         * src/powerpc/ppc_closure.S: Recode to fit shared libs.
3407
3408 2003-01-28  Andrew Haley  <aph@redhat.com>
3409
3410         * include/ffi.h.in: Enable FFI_CLOSURES for x86_64.
3411         * src/x86/ffi64.c (ffi_prep_closure): New.
3412         (ffi_closure_UNIX64_inner): New.
3413         * src/x86/unix64.S (ffi_closure_UNIX64): New.
3414
3415 2003-01-27  Alexandre Oliva  <aoliva@redhat.com>
3416
3417         * configure.in (toolexecdir, toolexeclibdir): Set and AC_SUBST.
3418         Remove USE_LIBDIR conditional.
3419         * Makefile.am (toolexecdir, toolexeclibdir): Don't override.
3420         * Makefile.in, configure: Rebuilt.
3421
3422 2003-01027  David Edelsohn  <edelsohn@gnu.org>
3423
3424         * Makefile.am (TARGET_SRC_POWERPC_AIX): Fix typo.
3425         * Makefile.in: Regenerate.
3426
3427 2003-01-22  Andrew Haley  <aph@redhat.com>
3428
3429         * src/powerpc/darwin.S (_ffi_call_AIX): Add Augmentation size to
3430         unwind info.
3431
3432 2003-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
3433
3434         * src/powerpc/darwin.S: Add unwind info.
3435         * src/powerpc/darwin_closure.S: Likewise.
3436
3437 2003-01-14  Andrew Haley  <aph@redhat.com>
3438
3439         * src/x86/ffi64.c (ffi_prep_args): Check for void retval.
3440         (ffi_prep_cif_machdep): Likewise.
3441         * src/x86/unix64.S: Add unwind info.
3442