OSDN Git Service

2010-08-09 Andreas Tobler <andreast@fgznet.ch>
[pf3gnuchains/gcc-fork.git] / libffi / ChangeLog
1 2010-08-09  Andreas Tobler  <andreast@fgznet.ch>
2
3         * configure.ac: Add target powerpc64-*-freebsd*.
4         * configure: Regenerate.
5         * testsuite/libffi.call/cls_align_longdouble_split.c: Pass
6         -mlong-double-128 only to linux targets.
7         * testsuite/libffi.call/cls_align_longdouble_split2.c: Likewise.
8         * testsuite/libffi.call/cls_longdouble.c: Likewise.
9         * testsuite/libffi.call/huge_struct.c: Likewise.
10
11 2010-07-10  Evan Phoenix  <evan@fallingsnow.net>
12
13         * src/closures.c (selinux_enabled_check): Fix strncmp usage bug.
14
15 2010-07-07  Dan HorĂ¡k <dan@danny.cz>
16
17         * include/ffi.h.in: Protect #define with #ifndef.
18         * src/powerpc/ffitarget.h: Ditto.
19         * src/s390/ffitarget.h: Ditto.
20         * src/sparc/ffitarget.h: Ditto.
21
22 2010-07-07   Neil Roberts <neil@linux.intel.com>
23
24         * src/x86/sysv.S (ffi_call_SYSV): Align the stack pointer to
25         16-bytes.
26
27 2010-07-02  Jakub Jelinek  <jakub@redhat.com>
28
29         * Makefile.am (AM_MAKEFLAGS): Pass also mandir to submakes.
30         * Makefile.in: Regenerated.
31
32 2010-05-19  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
33
34         * configure.ac (libffi_cv_as_x86_pcrel): Check for illegal in as
35         output, too.
36         (libffi_cv_as_ascii_pseudo_op): Check for .ascii.
37         (libffi_cv_as_string_pseudo_op): Check for .string.
38         * configure: Regenerate.
39         * fficonfig.h.in: Regenerate.
40         * src/x86/sysv.S (.eh_frame): Use .ascii, .string or error.
41
42 2010-05-05  Michael Kohler <michaelkohler@live.com>
43
44         * src/dlmalloc.c (dlfree): Fix spelling.
45         * src/ia64/ffi.c (ffi_prep_cif_machdep): Ditto.
46         * configure.ac: Ditto.
47         * configure: Rebuilt.
48
49 2010-04-13  Dan Witte  <dwitte@mozilla.com>
50
51         * msvcc.sh: Build with -W3 instead of -Wall.
52         * src/powerpc/ffi_darwin.c: Remove build warnings.
53         * src/x86/ffi.c: Ditto.
54         * src/x86/ffitarget.h: Ditto.
55
56 2010-04-12  Dan Witte  <dwitte@mozilla.com>
57             Walter Meinl <wuno@lsvw.de>
58
59         * configure.ac: Add OS/2 support.
60         * configure: Rebuilt.
61         * src/closures.c: Ditto.
62         * src/dlmalloc.c: Ditto.
63         * src/x86/win32.S: Ditto.
64
65 2010-04-07  Jakub Jelinek  <jakub@redhat.com>
66
67         * regex.c (byte_re_match_2_internal): Avoid set but not used
68         warning.
69
70 2010-04-02  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
71
72         * Makefile.in: Regenerate.
73         * aclocal.m4: Regenerate.
74         * include/Makefile.in: Regenerate.
75         * man/Makefile.in: Regenerate.
76         * testsuite/Makefile.in: Regenerate.
77
78 2010-03-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
79
80         * configure.ac (libffi_cv_as_x86_64_unwind_section_type): New test.
81         * configure: Regenerate.
82         * fficonfig.h.in: Regenerate.
83         * libffi/src/x86/unix64.S (.eh_frame)
84         [HAVE_AS_X86_64_UNWIND_SECTION_TYPE]: Use @unwind section type.
85
86 2010-03-14  Matthias Klose  <doko@ubuntu.com>
87
88         * src/x86/ffi64.c: Fix typo in comment.
89         * src/x86/ffi.c: Use /* ... */ comment style.
90
91 2010-02-24  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
92
93         * doc/libffi.texi (The Closure API): Fix typo.
94         * doc/libffi.info: Remove.
95
96 2010-02-15  Matthias Klose  <doko@ubuntu.com>
97
98         * src/arm/sysv.S (__ARM_ARCH__): Define for processor
99         __ARM_ARCH_7EM__.
100
101 2010-01-15  Anthony Green  <green@redhat.com>
102
103         * README: Add notes on building with Microsoft Visual C++.
104
105 2010-01-15  Daniel Witte  <dwitte@mozilla.com>
106
107         * msvcc.sh: New file.
108
109         * src/x86/win32.S: Port assembly routines to MSVC and #ifdef.
110         * src/x86/ffi.c: Tweak function declaration and remove excess
111         parens.
112         * include/ffi.h.in: Add __declspec(align(8)) to typedef struct
113         ffi_closure.
114
115         * src/x86/ffi.c: Merge ffi_call_SYSV and ffi_call_STDCALL into new
116         function ffi_call_win32 on X86_WIN32.
117         * src/x86/win32.S (ffi_call_SYSV): Rename to ffi_call_win32.
118         (ffi_call_STDCALL): Remove.
119
120         * src/prep_cif.c (ffi_prep_cif): Move stack space allocation code
121         to ffi_prep_cif_machdep for x86.
122         * src/x86/ffi.c (ffi_prep_cif_machdep): To here.
123
124 2010-01-15  Oliver Kiddle  <okiddle@yahoo.co.uk>
125
126         * src/x86/ffitarget.h (ffi_abi): Check for __i386 and __amd64 for
127         Sun Studio compiler compatibility. 
128
129 2010-01-12  Conrad Irwin <conrad.irwin@gmail.com>
130
131         * doc/libffi.texi: Add closure example.
132
133 2010-01-07  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
134
135         PR libffi/40701
136         * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRIdLL,
137         PRIuLL, PRId64, PRIu64, PRIuPTR): Define.
138         * testsuite/libffi.call/cls_align_sint64.c: Add -Wno-format on
139         alpha*-dec-osf*.
140         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
141         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
142         * testsuite/libffi.call/return_ll1.c: Likewise.
143         * testsuite/libffi.call/stret_medium2.c: Likewise.
144         * testsuite/libffi.special/ffitestcxx.h (allocate_mmap): Cast
145         MAP_FAILED to char *.
146
147 2010-01-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
148
149         * src/mips/n32.S: Use .abicalls and .eh_frame with __GNUC__.
150
151 2009-12-31  Anthony Green  <green@redhat.com>
152
153         * README: Update for libffi 3.0.9.
154
155 2009-12-27  Matthias Klose  <doko@ubuntu.com>
156
157         * configure.ac (HAVE_LONG_DOUBLE): Define for mips when
158         appropriate.
159         * configure: Rebuilt.
160
161 2009-12-26  Anthony Green  <green@redhat.com>
162
163         * testsuite/libffi.call/cls_longdouble_va.c: Mark as xfail for
164         avr32*-*-*.
165         * testsuite/libffi.call/cls_double_va.c: Ditto.
166
167 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
168
169         * testsuite/libffi.call/ffitest.h: Conditionally include stdint.h
170         and inttypes.h.
171         * testsuite/libffi.special/unwindtest.cc: Ditto.
172
173 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
174
175         * configure.ac: Add amd64-*-openbsd*.
176         * configure: Rebuilt.
177         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Link
178         openbsd programs with -lpthread.
179
180 2009-12-26  Anthony Green  <green@redhat.com>
181
182         * testsuite/libffi.call/cls_double_va.c,
183         testsuite/libffi.call/cls_longdouble.c,
184         testsuite/libffi.call/cls_longdouble_va.c,
185         testsuite/libffi.call/cls_pointer.c,
186         testsuite/libffi.call/cls_pointer_stack.c: Remove xfail for
187         mips*-*-* and arm*-*-*.
188         * testsuite/libffi.call/cls_align_longdouble_split.c,
189         testsuite/libffi.call/cls_align_longdouble_split2.c,
190         testsuite/libffi.call/stret_medium2.c,
191         testsuite/libffi.call/stret_medium.c,
192         testsuite/libffi.call/stret_large.c,
193         testsuite/libffi.call/stret_large2.c: Remove xfail for arm*-*-*.
194
195 2009-12-31  Kay Tietz  <ktietz70@googlemail.com>
196
197         * testsuite/libffi.call/ffitest.h,
198         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRuLL): Fix
199         definitions.
200
201 2009-12-31  Carlo Bramini  <carlo.bramix@libero.it>
202
203         * configure.ac (AM_LTLDFLAGS): Define for windows hosts.
204         * Makefile.am (libffi_la_LDFLAGS): Add AM_LTLDFLAGS.
205         * configure: Rebuilt.
206         * Makefile.in: Rebuilt.
207
208 2009-12-31  Anthony Green  <green@redhat.com>
209             Blake Chaffin.
210
211         * testsuite/libffi.call/huge_struct.c: New test case from Blake
212         Chaffin @ Apple.
213
214 2009-12-28  David Edelsohn  <edelsohn@gnu.org>
215
216         * src/powerpc/ffi_darwin.c (ffi_prep_args): Copy abi and nargs to
217         local variables.
218         (aix_adjust_aggregate_sizes): New function.
219         (ffi_prep_cif_machdep): Call it.
220
221 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
222
223         * configure.ac: Define FFI_MMAP_EXEC_WRIT for the given targets.
224         * configure: Regenerate.
225         * fficonfig.h.in: Likewise.
226         * src/closures.c: Remove the FFI_MMAP_EXEC_WRIT definition for
227         Solaris/x86.
228
229 2009-12-26  Andreas Schwab  <schwab@linux-m68k.org>
230
231         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Advance intarg_count
232         when a float arguments is passed in memory.
233         (ffi_closure_helper_SYSV): Mark general registers as used up when
234         a 64bit or soft-float long double argument is passed in memory.
235
236 2009-12-25  Matthias Klose  <doko@ubuntu.com>
237
238         * man/ffi_call.3: Fix #include in examples.
239         * doc/libffi.texi: Add dircategory.
240
241 2009-12-25  Frank Everdij <f.p.x.everdij@tudelft.nl>
242
243         * include/ffi.h.in: Placed '__GNUC__' ifdef around
244         '__attribute__((aligned(8)))' in ffi_closure, fixes compile for
245         IRIX MIPSPro c99.
246         * include/ffi_common.h: Added '__sgi' define to non
247         '__attribute__((__mode__()))' integer typedefs.
248         * src/mips/ffi.c (ffi_call, ffi_closure_mips_inner_O32,
249         ffi_closure_mips_inner_N32): Added 'defined(_MIPSEB)' to BE check.
250         (ffi_closure_mips_inner_O32, ffi_closure_mips_inner_N32): Added
251         FFI_LONGDOUBLE support and alignment(N32 only).
252         * src/mips/ffitarget.h: Corrected '#include <sgidefs.h>' for IRIX and
253         fixed non '__attribute__((__mode__()))' integer typedefs.
254         * src/mips/n32.S: Put '#ifdef linux' around '.abicalls' and '.eh_frame'
255         since they are Linux/GNU Assembler specific.
256
257 2009-12-25  Bradley Smith  <brad@brad-smith.co.uk>
258
259         * configure.ac, Makefile.am, src/avr32/ffi.c,
260         src/avr32/ffitarget.h,
261         src/avr32/sysv.S: Add AVR32 port.
262         * configure, Makefile.in: Rebuilt.
263
264 2009-12-21  Andreas Tobler  <a.tobler@schweiz.org>
265
266         * configure.ac: Make i?86 build on FreeBSD and OpenBSD.
267         * configure: Regenerate.
268
269 2009-12-15  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
270
271         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on PA HP-UX.
272
273 2009-12-13  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
274
275         * src/pa/ffi.c (ffi_closure_inner_pa32): Handle FFI_TYPE_LONGDOUBLE
276         type on HP-UX.
277
278 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
279
280         * src/sparc/ffi.c (ffi_closure_sparc_inner_v9): Properly align 'long
281         double' arguments.
282
283 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
284
285         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on Solaris < 10.
286
287 2009-12-10  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
288
289         PR libffi/40700
290         * src/closures.c [X86_64 && __sun__ && __svr4__]
291         (FFI_MMAP_EXEC_WRIT): Define.
292
293 2009-12-08  David Daney  <ddaney@caviumnetworks.com>
294
295         * testsuite/libffi.call/stret_medium.c: Remove xfail for mips*-*-*
296         * testsuite/libffi.call/cls_align_longdouble_split2.c: Same.
297         * testsuite/libffi.call/stret_large.c: Same.
298         * testsuite/libffi.call/cls_align_longdouble_split.c: Same.
299         * testsuite/libffi.call/stret_large2.c: Same.
300         * testsuite/libffi.call/stret_medium2.c: Same.
301
302 2009-12-07  David Edelsohn  <edelsohn@gnu.org>
303
304         * src/powerpc/aix_closure.S (libffi_closure_ASM): Fix tablejump
305         typo.
306
307 2009-12-05  David Edelsohn  <edelsohn@gnu.org>
308
309         * src/powerpc/aix.S: Update AIX32 code to be consistent with AIX64
310         code.
311         * src/powerpc/aix_closure.S: Same.
312
313 2009-12-05  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
314
315         * Makefile.in: Regenerate.
316         * configure: Regenerate.
317         * include/Makefile.in: Regenerate.
318         * man/Makefile.in: Regenerate.
319         * testsuite/Makefile.in: Regenerate.
320
321 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
322
323         * src/powerpc/aix_closure.S: Reorganize 64-bit code to match
324         linux64_closure.S.
325
326 2009-12-04  Uros Bizjak  <ubizjak@gmail.com>
327
328         PR libffi/41908
329         * src/x86/ffi64.c (classify_argument): Update from
330         gcc/config/i386/i386.c.
331         (ffi_closure_unix64_inner): Do not use the address of two consecutive
332         SSE registers directly.
333         * testsuite/libffi.call/cls_dbls_struct.c (main): Remove xfail
334         for x86_64 linux targets.
335
336 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
337
338         * src/powerpc/ffi_darwin.c (ffi_closure_helper_DARWIN): Increment
339         pfr for long double split between fpr13 and stack.
340
341 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
342
343         * src/powerpc/ffi_darwin.c (ffi_prep_args): Increment next_arg and
344         fparg_count twice for long double.
345
346 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
347
348         PR libffi/42243
349         * src/powerpc/ffi_darwin.c (ffi_prep_args): Remove extra parentheses.
350
351 2009-12-03  Uros Bizjak  <ubizjak@gmail.com>
352
353         * testsuite/libffi.call/cls_longdouble_va.c (main): Fix format string.
354         Remove xfails for x86 linux targets.
355
356 2009-12-02  David Edelsohn  <edelsohn@gnu.org>
357
358         * src/powerpc/ffi_darwin.c (ffi_prep_args): Fix typo in INT64
359         case.
360
361 2009-12-01  David Edelsohn  <edelsohn@gnu.org>
362
363         * src/powerpc/aix.S (ffi_call_AIX): Convert to more standard
364         register usage.  Call ffi_prep_args directly.  Add long double
365         return value support.
366         * src/powerpc/ffi_darwin.c (ffi_prep_args): Double arg increment
367         applies to FFI_TYPE_DOUBLE.  Correct fpr_base increment typo.
368         Separate FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases.
369         (ffi_prep_cif_machdep): Only 16 byte stack alignment in 64 bit
370         mode.
371         (ffi_closure_helper_DARWIN): Remove nf and ng counters.  Move temp
372         into case.
373         * src/powerpc/aix_closure.S: Maintain 16 byte stack alignment.
374         Allocate result area between params and FPRs.
375
376 2009-11-30  David Edelsohn  <edelsohn@gnu.org>
377
378         PR target/35484
379         * src/powerpc/ffitarget.h (POWERPC64): Define for PPC64 Linux and
380         AIX64.
381         * src/powerpc/aix.S: Implement AIX64 version.
382         * src/powerpc/aix_closure.S: Implement AIX64 version.
383         (ffi_closure_ASM): Use extsb, lha and displament addresses.
384         * src/powerpc/ffi_darwin.c (ffi_prep_args): Implement AIX64
385         support.
386         (ffi_prep_cif_machdep): Same.
387         (ffi_call): Same.
388         (ffi_closure_helper_DARWIN): Same.
389
390 2009-11-02  Andreas Tobler  <a.tobler@schweiz.org>
391
392         PR libffi/41908
393         * testsuite/libffi.call/testclosure.c: New test.
394
395 2009-09-28  Kai Tietz  <kai.tietz@onevision.com>
396
397         * src/x86/win64.S (_ffi_call_win64 stack): Remove for gnu
398         assembly version use of ___chkstk.
399
400 2009-09-23  Matthias Klose  <doko@ubuntu.com>
401
402         PR libffi/40242, PR libffi/41443
403         * src/arm/sysv.S (__ARM_ARCH__): Define for processors
404         __ARM_ARCH_6T2__, __ARM_ARCH_6M__, __ARM_ARCH_7__,
405         __ARM_ARCH_7A__, __ARM_ARCH_7R__, __ARM_ARCH_7M__.
406         Change the conditionals to __SOFTFP__ || __ARM_EABI__
407         for -mfloat-abi=softfp to work.
408
409 2009-09-17  Loren J. Rittle  <ljrittle@acm.org>
410
411         PR testsuite/32843 (strikes again)
412         * src/x86/ffi.c (ffi_prep_cif_machdep): Add X86_FREEBSD to
413         enable proper extension on char and short.
414
415 2009-09-15  David Daney  <ddaney@caviumnetworks.com>
416
417         * src/java_raw_api.c (ffi_java_raw_to_rvalue): Remove special
418         handling for FFI_TYPE_POINTER.
419         * src/mips/ffitarget.h (FFI_TYPE_STRUCT_D_SOFT,
420         FFI_TYPE_STRUCT_F_SOFT, FFI_TYPE_STRUCT_DD_SOFT,
421         FFI_TYPE_STRUCT_FF_SOFT, FFI_TYPE_STRUCT_FD_SOFT,
422         FFI_TYPE_STRUCT_DF_SOFT, FFI_TYPE_STRUCT_SOFT): New defines.
423         (FFI_N32_SOFT_FLOAT, FFI_N64_SOFT_FLOAT): New ffi_abi enumerations.
424         (enum ffi_abi): Set FFI_DEFAULT_ABI for soft-float.
425         * src/mips/n32.S (ffi_call_N32): Add handling for soft-float
426         structure and pointer returns.
427         (ffi_closure_N32): Add handling for pointer returns.
428         * src/mips/ffi.c (ffi_prep_args, calc_n32_struct_flags,
429         calc_n32_return_struct_flags): Handle soft-float.
430         (ffi_prep_cif_machdep):  Handle soft-float, fix pointer handling.
431         (ffi_call_N32): Declare proper argument types.
432         (ffi_call, copy_struct_N32, ffi_closure_mips_inner_N32): Handle
433         soft-float.
434
435 2009-08-24  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
436
437         * configure.ac (AC_PREREQ): Bump to 2.64.
438
439 2009-08-22  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
440
441         * Makefile.am (install-html, install-pdf): Remove.
442         * Makefile.in: Regenerate.
443
444         * Makefile.in: Regenerate.
445         * aclocal.m4: Regenerate.
446         * configure: Regenerate.
447         * fficonfig.h.in: Regenerate.
448         * include/Makefile.in: Regenerate.
449         * man/Makefile.in: Regenerate.
450         * testsuite/Makefile.in: Regenerate.
451
452 2009-07-30  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
453
454         * configure.ac (_AC_ARG_VAR_PRECIOUS): Use m4_rename_force.
455
456 2009-07-24  Dave Korn  <dave.korn.cygwin@gmail.com>
457
458         PR libffi/40807
459         * src/x86/ffi.c (ffi_prep_cif_machdep): Also use sign/zero-extending
460         return types for X86_WIN32.
461         * src/x86/win32.S (_ffi_call_SYSV): Handle omitted return types.
462         (_ffi_call_STDCALL, _ffi_closure_SYSV, _ffi_closure_raw_SYSV,
463         _ffi_closure_STDCALL): Likewise.
464
465         * src/closures.c (is_selinux_enabled): Define to const 0 for Cygwin.
466         (dlmmap, dlmunmap): Also use these functions on Cygwin.
467
468 2009-07-11  Richard Sandiford  <rdsandiford@googlemail.com>
469
470         PR testsuite/40699
471         PR testsuite/40707
472         PR testsuite/40709
473         * testsuite/lib/libffi-dg.exp: Revert 2009-07-02, 2009-07-01 and
474         2009-06-30 commits.
475
476 2009-07-01  Richard Sandiford  <r.sandiford@uk.ibm.com>
477
478         * testsuite/lib/libffi-dg.exp (libffi-init): Set ld_library_path
479         to "" before adding paths.  (This reinstates an assignment that
480         was removed by my 2009-06-30 commit, but changes the initial
481         value from "." to "".)
482
483 2009-07-01  H.J. Lu  <hongjiu.lu@intel.com>
484
485         PR testsuite/40601
486         * testsuite/lib/libffi-dg.exp (libffi-init): Properly set
487         gccdir.  Adjust ld_library_path for gcc only if gccdir isn't
488         empty.
489
490 2009-06-30  Richard Sandiford  <r.sandiford@uk.ibm.com>
491
492         * testsuite/lib/libffi-dg.exp (libffi-init): Don't add "."
493         to ld_library_path.  Use add_path.  Add just find_libgcc_s
494         to ld_library_path, not every libgcc multilib directory.
495
496 2009-06-16  Wim Lewis  <wiml@hhhh.org>
497
498         * src/powerpc/ffi.c: Avoid clobbering cr3 and cr4, which are
499         supposed to be callee-saved.
500         * src/powerpc/sysv.S (small_struct_return_value): Fix overrun of
501         return buffer for odd-size structs.
502
503 2009-06-16  Andreas Tobler  <a.tobler@schweiz.org>
504
505         PR libffi/40444
506         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Add
507         allow_stack_execute for Darwin.
508
509 2009-06-16  Andrew Haley  <aph@redhat.com>
510
511         * configure.ac (TARGETDIR): Add missing blank lines.
512         * configure: Regenerate.
513
514 2009-06-16  Andrew Haley  <aph@redhat.com>
515
516         * testsuite/libffi.call/cls_align_sint64.c,
517         testsuite/libffi.call/cls_align_uint64.c,
518         testsuite/libffi.call/cls_longdouble_va.c,
519         testsuite/libffi.call/cls_ulonglong.c,
520         testsuite/libffi.call/return_ll1.c,
521         testsuite/libffi.call/stret_medium2.c: Fix printf format
522         specifiers.
523         * testsuite/libffi.call/ffitest.h,
524         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRIuLL): Define.
525
526 2009-06-15  Andrew Haley  <aph@redhat.com>
527
528         * testsuite/libffi.call/err_bad_typedef.c: xfail everywhere.
529         * testsuite/libffi.call/err_bad_abi.c: Likewise.
530
531 2009-06-12  Andrew Haley  <aph@redhat.com>
532
533         * Makefile.am: Remove info_TEXINFOS.
534
535 2009-06-12  Andrew Haley  <aph@redhat.com>
536
537         * ChangeLog.libffi: testsuite/libffi.call/cls_align_sint64.c,
538         testsuite/libffi.call/cls_align_uint64.c,
539         testsuite/libffi.call/cls_ulonglong.c,
540         testsuite/libffi.call/return_ll1.c,
541         testsuite/libffi.call/stret_medium2.c: Fix printf format
542         specifiers.
543         testsuite/libffi.special/unwindtest.cc: include stdint.h.
544
545 2009-06-11  Timothy Wall  <twall@users.sf.net>
546
547         * Makefile.am,
548         configure.ac,
549         include/ffi.h.in,
550         include/ffi_common.h,
551         src/closures.c,
552         src/dlmalloc.c,
553         src/x86/ffi.c,
554         src/x86/ffitarget.h,
555         src/x86/win64.S (new),
556         README: Added win64 support (mingw or MSVC)
557         * Makefile.in,
558         include/Makefile.in,
559         man/Makefile.in,
560         testsuite/Makefile.in,
561         configure,
562         aclocal.m4: Regenerated
563         * ltcf-c.sh: properly escape cygwin/w32 path
564         * man/ffi_call.3: Clarify size requirements for return value.
565         * src/x86/ffi64.c: Fix filename in comment.
566         * src/x86/win32.S: Remove unused extern.
567
568         * testsuite/libffi.call/closure_fn0.c,
569         testsuite/libffi.call/closure_fn1.c,
570         testsuite/libffi.call/closure_fn2.c,
571         testsuite/libffi.call/closure_fn3.c,
572         testsuite/libffi.call/closure_fn4.c,
573         testsuite/libffi.call/closure_fn5.c,
574         testsuite/libffi.call/closure_fn6.c,
575         testsuite/libffi.call/closure_stdcall.c,
576         testsuite/libffi.call/cls_12byte.c,
577         testsuite/libffi.call/cls_16byte.c,
578         testsuite/libffi.call/cls_18byte.c,
579         testsuite/libffi.call/cls_19byte.c,
580         testsuite/libffi.call/cls_1_1byte.c,
581         testsuite/libffi.call/cls_20byte.c,
582         testsuite/libffi.call/cls_20byte1.c,
583         testsuite/libffi.call/cls_24byte.c,
584         testsuite/libffi.call/cls_2byte.c,
585         testsuite/libffi.call/cls_3_1byte.c,
586         testsuite/libffi.call/cls_3byte1.c,
587         testsuite/libffi.call/cls_3byte2.c,
588         testsuite/libffi.call/cls_4_1byte.c,
589         testsuite/libffi.call/cls_4byte.c,
590         testsuite/libffi.call/cls_5_1_byte.c,
591         testsuite/libffi.call/cls_5byte.c,
592         testsuite/libffi.call/cls_64byte.c,
593         testsuite/libffi.call/cls_6_1_byte.c,
594         testsuite/libffi.call/cls_6byte.c,
595         testsuite/libffi.call/cls_7_1_byte.c,
596         testsuite/libffi.call/cls_7byte.c,
597         testsuite/libffi.call/cls_8byte.c,
598         testsuite/libffi.call/cls_9byte1.c,
599         testsuite/libffi.call/cls_9byte2.c,
600         testsuite/libffi.call/cls_align_double.c,
601         testsuite/libffi.call/cls_align_float.c,
602         testsuite/libffi.call/cls_align_longdouble.c,
603         testsuite/libffi.call/cls_align_longdouble_split.c,
604         testsuite/libffi.call/cls_align_longdouble_split2.c,
605         testsuite/libffi.call/cls_align_pointer.c,
606         testsuite/libffi.call/cls_align_sint16.c,
607         testsuite/libffi.call/cls_align_sint32.c,
608         testsuite/libffi.call/cls_align_sint64.c,
609         testsuite/libffi.call/cls_align_uint16.c,
610         testsuite/libffi.call/cls_align_uint32.c,
611         testsuite/libffi.call/cls_align_uint64.c,
612         testsuite/libffi.call/cls_dbls_struct.c,
613         testsuite/libffi.call/cls_double.c,
614         testsuite/libffi.call/cls_double_va.c,
615         testsuite/libffi.call/cls_float.c,
616         testsuite/libffi.call/cls_longdouble.c,
617         testsuite/libffi.call/cls_longdouble_va.c,
618         testsuite/libffi.call/cls_multi_schar.c,
619         testsuite/libffi.call/cls_multi_sshort.c,
620         testsuite/libffi.call/cls_multi_sshortchar.c,
621         testsuite/libffi.call/cls_multi_uchar.c,
622         testsuite/libffi.call/cls_multi_ushort.c,
623         testsuite/libffi.call/cls_multi_ushortchar.c,
624         testsuite/libffi.call/cls_pointer.c,
625         testsuite/libffi.call/cls_pointer_stack.c,
626         testsuite/libffi.call/cls_schar.c,
627         testsuite/libffi.call/cls_sint.c,
628         testsuite/libffi.call/cls_sshort.c,
629         testsuite/libffi.call/cls_uchar.c,
630         testsuite/libffi.call/cls_uint.c,
631         testsuite/libffi.call/cls_ulonglong.c,
632         testsuite/libffi.call/cls_ushort.c,
633         testsuite/libffi.call/err_bad_abi.c,
634         testsuite/libffi.call/err_bad_typedef.c,
635         testsuite/libffi.call/float2.c,
636         testsuite/libffi.call/huge_struct.c,
637         testsuite/libffi.call/nested_struct.c,
638         testsuite/libffi.call/nested_struct1.c,
639         testsuite/libffi.call/nested_struct10.c,
640         testsuite/libffi.call/nested_struct2.c,
641         testsuite/libffi.call/nested_struct3.c,
642         testsuite/libffi.call/nested_struct4.c,
643         testsuite/libffi.call/nested_struct5.c,
644         testsuite/libffi.call/nested_struct6.c,
645         testsuite/libffi.call/nested_struct7.c,
646         testsuite/libffi.call/nested_struct8.c,
647         testsuite/libffi.call/nested_struct9.c,
648         testsuite/libffi.call/problem1.c,
649         testsuite/libffi.call/return_ldl.c,
650         testsuite/libffi.call/return_ll1.c,
651         testsuite/libffi.call/stret_large.c,
652         testsuite/libffi.call/stret_large2.c,
653         testsuite/libffi.call/stret_medium.c,
654         testsuite/libffi.call/stret_medium2.c,
655         testsuite/libffi.special/unwindtest.cc: use ffi_closure_alloc instead
656         of checking for MMAP.  Use intptr_t instead of long casts.
657
658 2009-06-11  Kaz Kojima  <kkojima@gcc.gnu.org>
659
660         * testsuite/libffi.call/cls_longdouble_va.c: Add xfail sh*-*-linux-*.
661         * testsuite/libffi.call/err_bad_abi.c: Add xfail sh*-*-*.
662         * testsuite/libffi.call/err_bad_typedef.c: Likewise.
663
664 2009-06-09  Andrew Haley  <aph@redhat.com>
665
666         * src/x86/freebsd.S: Add missing file.
667
668 2009-06-08  Andrew Haley  <aph@redhat.com>
669
670         Import from libffi 3.0.8:
671
672         * doc/libffi.texi: New file.
673         * doc/libffi.info: Likewise.
674         * doc/stamp-vti: Likewise.
675         * man/Makefile.am: New file.
676         * man/ffi_call.3: New file.
677
678         * Makefile.am (EXTRA_DIST): Add src/x86/darwin64.S,
679         src/dlmalloc.c.
680         (nodist_libffi_la_SOURCES): Add X86_FREEBSD.
681
682         * configure.ac: Bump version to 3.0.8.
683         parisc*-*-linux*: Add.
684         i386-*-freebsd* | i386-*-openbsd*: Add.
685         powerpc-*-beos*: Add.
686         AM_CONDITIONAL X86_FREEBSD: Add.
687         AC_CONFIG_FILES: Add man/Makefile.
688
689         * include/ffi.h.in (FFI_FN): Change void (*)() to void (*)(void).
690
691 2009-06-08  Andrew Haley  <aph@redhat.com>
692
693         * README: Import from libffi 3.0.8.
694
695 2009-06-08  Andrew Haley  <aph@redhat.com>
696
697         * testsuite/libffi.call/err_bad_abi.c: Add xfails.
698         * testsuite/libffi.call/cls_longdouble_va.c: Add xfails.
699         * testsuite/libffi.call/cls_dbls_struct.c: Add xfail x86_64-*-linux-*.
700         * testsuite/libffi.call/err_bad_typedef.c: Add xfails.
701
702         * testsuite/libffi.call/stret_medium2.c: Add __UNUSED__ to args.
703         * testsuite/libffi.call/stret_medium.c: Likewise.
704         * testsuite/libffi.call/stret_large2.c: Likewise.
705         * testsuite/libffi.call/stret_large.c:  Likewise.
706
707 2008-12-26  Timothy Wall  <twall@users.sf.net>
708
709         * testsuite/libffi.call/cls_longdouble.c,
710         testsuite/libffi.call/cls_longdouble_va.c,
711         testsuite/libffi.call/cls_align_longdouble.c,
712         testsuite/libffi.call/cls_align_longdouble_split.c,
713         testsuite/libffi.call/cls_align_longdouble_split2.c: mark expected
714         failures on x86_64 cygwin/mingw.
715
716 2008-12-22  Timothy Wall  <twall@users.sf.net>
717
718         * testsuite/libffi.call/closure_fn0.c,
719         testsuite/libffi.call/closure_fn1.c,
720         testsuite/libffi.call/closure_fn2.c,
721         testsuite/libffi.call/closure_fn3.c,
722         testsuite/libffi.call/closure_fn4.c,
723         testsuite/libffi.call/closure_fn5.c,
724         testsuite/libffi.call/closure_fn6.c,
725         testsuite/libffi.call/closure_loc_fn0.c,
726         testsuite/libffi.call/closure_stdcall.c,
727         testsuite/libffi.call/cls_align_pointer.c,
728         testsuite/libffi.call/cls_pointer.c,
729         testsuite/libffi.call/cls_pointer_stack.c: use portable cast from
730         pointer to integer (intptr_t).
731         * testsuite/libffi.call/cls_longdouble.c: disable for win64.
732
733 2008-07-24  Anthony Green  <green@redhat.com>
734
735         * testsuite/libffi.call/cls_dbls_struct.c,
736         testsuite/libffi.call/cls_double_va.c,
737         testsuite/libffi.call/cls_longdouble.c,
738         testsuite/libffi.call/cls_longdouble_va.c,
739         testsuite/libffi.call/cls_pointer.c,
740         testsuite/libffi.call/cls_pointer_stack.c,
741         testsuite/libffi.call/err_bad_abi.c: Clean up failures from
742         compiler warnings.
743
744 2008-03-04  Anthony Green  <green@redhat.com>
745             Blake Chaffin
746             hos@tamanegi.org
747
748         * testsuite/libffi.call/cls_align_longdouble_split2.c
749           testsuite/libffi.call/cls_align_longdouble_split.c
750           testsuite/libffi.call/cls_dbls_struct.c
751           testsuite/libffi.call/cls_double_va.c
752           testsuite/libffi.call/cls_longdouble.c
753           testsuite/libffi.call/cls_longdouble_va.c
754           testsuite/libffi.call/cls_pointer.c
755           testsuite/libffi.call/cls_pointer_stack.c
756           testsuite/libffi.call/err_bad_abi.c
757           testsuite/libffi.call/err_bad_typedef.c
758           testsuite/libffi.call/stret_large2.c
759           testsuite/libffi.call/stret_large.c
760           testsuite/libffi.call/stret_medium2.c
761           testsuite/libffi.call/stret_medium.c: New tests from Apple.
762
763 2009-06-05  Andrew Haley  <aph@redhat.com>
764
765         * src/x86/ffitarget.h, src/x86/ffi.c: Merge stdcall changes from
766         libffi.
767
768 2009-06-04  Andrew Haley  <aph@redhat.com>
769
770         * src/x86/ffitarget.h, src/x86/win32.S, src/x86/ffi.c: Back out
771         stdcall changes.
772
773 2008-02-26  Anthony Green  <green@redhat.com>
774             Thomas Heller  <theller@ctypes.org>
775
776         * src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C
777         comment.
778
779 2008-02-03  Timothy Wall  <twall@users.sf.net>
780
781         * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return
782           offset based on code pointer, not data pointer.
783
784 2008-01-31  Timothy Wall <twall@users.sf.net>
785
786         * testsuite/libffi.call/closure_stdcall.c: Add test for stdcall
787         closures.
788         * src/x86/ffitarget.h: Increase size of trampoline for stdcall
789         closures.
790         * src/x86/win32.S: Add assembly for stdcall closure.
791         * src/x86/ffi.c: Initialize stdcall closure trampoline.
792
793 2009-06-04  Andrew Haley  <aph@redhat.com>
794
795         * include/ffi.h.in: Change void (*)() to void (*)(void).
796         * src/x86/ffi.c: Likewise.
797
798 2009-06-04  Andrew Haley  <aph@redhat.com>
799
800         * src/powerpc/ppc_closure.S: Insert licence header.
801         * src/powerpc/linux64_closure.S: Likewise.
802         * src/m68k/sysv.S: Likewise.
803
804         * src/sh64/ffi.c: Change void (*)() to void (*)(void).
805         * src/powerpc/ffi.c: Likewise.
806         * src/powerpc/ffi_darwin.c: Likewise.
807         * src/m32r/ffi.c: Likewise.
808         * src/sh64/ffi.c: Likewise.
809         * src/x86/ffi64.c: Likewise.
810         * src/alpha/ffi.c: Likewise.
811         * src/alpha/osf.S: Likewise.
812         * src/frv/ffi.c: Likewise.
813         * src/s390/ffi.c: Likewise.
814         * src/pa/ffi.c: Likewise.
815         * src/pa/hpux32.S: Likewise.
816         * src/ia64/unix.S: Likewise.
817         * src/ia64/ffi.c: Likewise.
818         * src/sparc/ffi.c: Likewise.
819         * src/mips/ffi.c: Likewise.
820         * src/sh/ffi.c: Likewise.
821
822 2008-02-15  David Daney  <ddaney@avtrex.com>
823
824         * src/mips/ffi.c (USE__BUILTIN___CLEAR_CACHE):
825         Define (conditionally), and use it to include cachectl.h.
826         (ffi_prep_closure_loc): Fix cache flushing.
827         * src/mips/ffitarget.h (_ABIN32, _ABI64, _ABIO32): Define.
828
829 2009-06-04  Andrew Haley  <aph@redhat.com>
830
831         include/ffi.h.in,
832         src/arm/ffitarget.h,
833         src/arm/ffi.c,
834         src/arm/sysv.S,
835         src/powerpc/ffitarget.h,
836         src/closures.c,
837         src/sh64/ffitarget.h,
838         src/sh64/ffi.c,
839         src/sh64/sysv.S,
840         src/types.c,
841         src/x86/ffi64.c,
842         src/x86/ffitarget.h,
843         src/x86/win32.S,
844         src/x86/darwin.S,
845         src/x86/ffi.c,
846         src/x86/sysv.S,
847         src/x86/unix64.S,
848         src/alpha/ffitarget.h,
849         src/alpha/ffi.c,
850         src/alpha/osf.S,
851         src/m68k/ffitarget.h,
852         src/frv/ffitarget.h,
853         src/frv/ffi.c,
854         src/s390/ffitarget.h,
855         src/s390/sysv.S,
856         src/cris/ffitarget.h,
857         src/pa/linux.S,
858         src/pa/ffitarget.h,
859         src/pa/ffi.c,
860         src/raw_api.c,
861         src/ia64/ffitarget.h,
862         src/ia64/unix.S,
863         src/ia64/ffi.c,
864         src/ia64/ia64_flags.h,
865         src/java_raw_api.c,
866         src/debug.c,
867         src/sparc/v9.S,
868         src/sparc/ffitarget.h,
869         src/sparc/ffi.c,
870         src/sparc/v8.S,
871         src/mips/ffitarget.h,
872         src/mips/n32.S,
873         src/mips/o32.S,
874         src/mips/ffi.c,
875         src/prep_cif.c,
876         src/sh/ffitarget.h,
877         src/sh/ffi.c,
878         src/sh/sysv.S: Update license text.
879
880 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
881
882         * src/x86/win32.S (_ffi_closure_STDCALL):  New function.
883         (.eh_frame):  Add FDE for it.
884
885 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
886
887         * configure.ac:  Also check if assembler supports pc-relative
888         relocs on X86_WIN32 targets.
889         * configure:  Regenerate.
890         * src/x86/win32.S (ffi_prep_args):  Declare extern, not global.
891         (_ffi_call_SYSV):  Add missing function type symbol .def and
892         add EH markup labels.
893         (_ffi_call_STDCALL):  Likewise.
894         (_ffi_closure_SYSV):  Likewise.
895         (_ffi_closure_raw_SYSV):  Likewise.
896         (.eh_frame):  Add hand-crafted EH data.
897
898 2009-04-09  Jakub Jelinek  <jakub@redhat.com>
899
900         * testsuite/lib/libffi-dg.exp: Change copyright header to refer to
901         version 3 of the GNU General Public License and to point readers
902         at the COPYING3 file and the FSF's license web page.
903         * testsuite/libffi.call/call.exp: Likewise.
904         * testsuite/libffi.special/special.exp: Likewise.
905
906 2009-03-01  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
907
908         * configure: Regenerate.
909
910 2008-12-18  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
911
912         PR libffi/26048
913         * configure.ac (HAVE_AS_X86_PCREL): New test.
914         * configure: Regenerate.
915         * fficonfig.h.in: Regenerate.
916         * src/x86/sysv.S [!FFI_NO_RAW_API]: Precalculate
917         RAW_CLOSURE_CIF_OFFSET, RAW_CLOSURE_FUN_OFFSET,
918         RAW_CLOSURE_USER_DATA_OFFSET for the Solaris 10/x86 assembler.
919         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
920         * src/x86/unix64.S (.Lstore_table): Move to .text section.
921         (.Lload_table): Likewise.
922         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
923
924 2008-12-18  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
925
926         * configure: Regenerate.
927
928 2008-11-21  Eric Botcazou  <ebotcazou@adacore.com>
929
930         * src/sparc/ffi.c (ffi_prep_cif_machdep): Add support for
931         signed/unsigned int8/16 return values.
932         * src/sparc/v8.S (ffi_call_v8): Likewise.
933         (ffi_closure_v8): Likewise.
934
935 2008-09-26  Peter O'Gorman  <pogma@thewrittenword.com>
936             Steve Ellcey  <sje@cup.hp.com>
937
938         * configure: Regenerate for new libtool.
939         * Makefile.in: Ditto.
940         * include/Makefile.in: Ditto.
941         * aclocal.m4: Ditto.
942
943 2008-08-25  Andreas Tobler  <a.tobler@schweiz.org>
944
945         * src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and
946         FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum.
947         Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT.
948         Adjust copyright notice.
949         * src/powerpc/ffi.c: Add two new flags to indicate if we have one
950         register or two register to use for FFI_SYSV structs.
951         (ffi_prep_cif_machdep): Pass the right register flag introduced above.
952         (ffi_closure_helper_SYSV): Fix the return type for
953         FFI_SYSV_TYPE_SMALL_STRUCT. Comment.
954         Adjust copyright notice.
955
956 2008-07-16  Kaz Kojima  <kkojima@gcc.gnu.org>
957
958         * src/sh/ffi.c (ffi_prep_closure_loc): Turn INSN into an unsigned
959         int.
960
961 2008-06-17  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
962
963         * configure: Regenerate.
964         * include/Makefile.in: Regenerate.
965         * testsuite/Makefile.in: Regenerate.
966
967 2008-06-07  Joseph Myers  <joseph@codesourcery.com>
968
969         * configure.ac (parisc*-*-linux*, powerpc-*-sysv*,
970         powerpc-*-beos*): Remove.
971         * configure: Regenerate.
972
973 2008-05-09  Julian Brown  <julian@codesourcery.com>
974
975         * Makefile.am (LTLDFLAGS): New.
976         (libffi_la_LDFLAGS): Use above.
977         * Makefile.in: Regenerate.
978
979 2008-04-18  Paolo Bonzini  <bonzini@gnu.org>
980
981         PR bootstrap/35457
982         * aclocal.m4: Regenerate.
983         * configure: Regenerate.
984
985 2008-03-26  Kaz Kojima  <kkojima@gcc.gnu.org>
986
987         * src/sh/sysv.S: Add .note.GNU-stack on Linux.
988         * src/sh64/sysv.S: Likewise.
989
990 2008-03-26  Daniel Jacobowitz  <dan@debian.org>
991
992         * src/arm/sysv.S: Fix ARM comment marker.
993
994 2008-03-26  Jakub Jelinek  <jakub@redhat.com>
995
996         * src/alpha/osf.S: Add .note.GNU-stack on Linux.
997         * src/s390/sysv.S: Likewise.
998         * src/powerpc/ppc_closure.S: Likewise.
999         * src/powerpc/sysv.S: Likewise.
1000         * src/x86/unix64.S: Likewise.
1001         * src/x86/sysv.S: Likewise.
1002         * src/sparc/v8.S: Likewise.
1003         * src/sparc/v9.S: Likewise.
1004         * src/m68k/sysv.S: Likewise.
1005         * src/arm/sysv.S: Likewise.
1006
1007 2008-03-16  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1008
1009         * aclocal.m4: Regenerate.
1010         * configure: Likewise.
1011         * Makefile.in: Likewise.
1012         * include/Makefile.in: Likewise.
1013         * testsuite/Makefile.in: Likewise.
1014
1015 2008-02-12  Bjoern Koenig  <bkoenig@alpha-tierchen.de>
1016             Andreas Tobler  <a.tobler@schweiz.org>
1017
1018         * configure.ac: Add amd64-*-freebsd* target.
1019         * configure: Regenerate.
1020
1021 2008-01-30  H.J. Lu  <hongjiu.lu@intel.com>
1022
1023         PR libffi/34612
1024         * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when
1025         returning struct.
1026
1027         * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer"
1028         tests.
1029
1030 2008-01-24  David Edelsohn  <edelsohn@gnu.org>
1031
1032         * configure: Regenerate.
1033
1034 2008-01-06  Andreas Tobler  <a.tobler@schweiz.org>
1035
1036         * src/x86/ffi.c (ffi_prep_cif_machdep): Fix thinko.
1037
1038 2008-01-05  Andreas Tobler  <a.tobler@schweiz.org>
1039
1040         PR testsuite/32843
1041         * src/x86/ffi.c (ffi_prep_cif_machdep): Add code for
1042         signed/unsigned int8/16 for X86_DARWIN.
1043         Updated copyright info.
1044         Handle one and two byte structs with special cif->flags.
1045         * src/x86/ffitarget.h: Add special types for one and two byte structs.
1046         Updated copyright info.
1047         * src/x86/darwin.S (ffi_call_SYSV): Rewrite to use a jump table like
1048         sysv.S
1049         Remove code to pop args from the stack after call.
1050         Special-case signed/unsigned for int8/16, one and two byte structs.
1051         (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1052         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1053         FFI_TYPE_SINT32.
1054         Updated copyright info.
1055
1056 2007-12-08  David Daney  <ddaney@avtrex.com>
1057
1058         * src/mips/n32.S (ffi_call_N32):  Replace dadd with ADDU, dsub with
1059         SUBU, add with ADDU and use smaller code sequences.
1060
1061 2007-12-07  David Daney  <ddaney@avtrex.com>
1062
1063         * src/mips/ffi.c (ffi_prep_cif_machdep): Handle long double return
1064         type.
1065
1066 2007-12-06  David Daney  <ddaney@avtrex.com>
1067
1068         * include/ffi.h.in (FFI_SIZEOF_JAVA_RAW): Define if not already
1069         defined.
1070         (ffi_java_raw): New typedef.
1071         (ffi_java_raw_call, ffi_java_ptrarray_to_raw,
1072         ffi_java_raw_to_ptrarray): Change parameter types from ffi_raw to
1073         ffi_java_raw.
1074         (ffi_java_raw_closure) : Same.
1075         (ffi_prep_java_raw_closure, ffi_prep_java_raw_closure_loc): Change
1076         parameter types.
1077         * src/java_raw_api.c (ffi_java_raw_size):  Replace FFI_SIZEOF_ARG with
1078         FFI_SIZEOF_JAVA_RAW.
1079         (ffi_java_raw_to_ptrarray): Change type of raw to ffi_java_raw.
1080         Replace FFI_SIZEOF_ARG with FFI_SIZEOF_JAVA_RAW. Use
1081         sizeof(ffi_java_raw) for alignment calculations.
1082         (ffi_java_ptrarray_to_raw): Same.
1083         (ffi_java_rvalue_to_raw): Add special handling for FFI_TYPE_POINTER
1084         if FFI_SIZEOF_JAVA_RAW == 4.
1085         (ffi_java_raw_to_rvalue): Same.
1086         (ffi_java_raw_call): Change type of raw to ffi_java_raw.
1087         (ffi_java_translate_args): Same.
1088         (ffi_prep_java_raw_closure_loc, ffi_prep_java_raw_closure): Change
1089         parameter types.
1090         * src/mips/ffitarget.h (FFI_SIZEOF_JAVA_RAW): Define for N32 ABI.
1091
1092 2007-12-06  David Daney  <ddaney@avtrex.com>
1093
1094         * src/mips/n32.S (ffi_closure_N32): Use 64-bit add instruction on
1095         pointer values.
1096
1097 2007-12-01  Andreas Tobler  <a.tobler@schweiz.org>
1098
1099         PR libffi/31937
1100         * src/powerpc/ffitarget.h: Introduce new ABI FFI_LINUX_SOFT_FLOAT.
1101         Add local FFI_TYPE_UINT128 to handle soft-float long-double-128.
1102         * src/powerpc/ffi.c: Distinguish between __NO_FPRS__ and not and
1103         set the NUM_FPR_ARG_REGISTERS according to.
1104         Add support for potential soft-float support under hard-float
1105         architecture.
1106         (ffi_prep_args_SYSV): Set NUM_FPR_ARG_REGISTERS to 0 in case of
1107         FFI_LINUX_SOFT_FLOAT, handle float, doubles and long-doubles according
1108         to the FFI_LINUX_SOFT_FLOAT ABI.
1109         (ffi_prep_cif_machdep): Likewise.
1110         (ffi_closure_helper_SYSV): Likewise.
1111         * src/powerpc/ppc_closure.S: Make sure not to store float/double
1112         on archs where __NO_FPRS__ is true.
1113         Add FFI_TYPE_UINT128 support.
1114         * src/powerpc/sysv.S: Add support for soft-float long-double-128.
1115         Adjust copyright notice.
1116
1117 2007-11-25  Andreas Tobler  <a.tobler@schweiz.org>
1118
1119         * src/closures.c: Move defintion of MAYBE_UNUSED from here to ...
1120         * include/ffi_common.h: ... here.
1121         Update copyright.
1122
1123 2007-11-17  Andreas Tobler  <a.tobler@schweiz.org>
1124
1125         * src/powerpc/sysv.S: Load correct cr to compare if we have long double.
1126         * src/powerpc/linux64.S: Likewise.
1127         * src/powerpc/ffi.c: Add a comment to show which part goes into cr6.
1128         * testsuite/libffi.call/return_ldl.c: New test.
1129
1130 2007-09-04    <aph@redhat.com>
1131
1132         * src/arm/sysv.S (UNWIND): New.
1133         (Whole file): Conditionally compile unwinder directives.
1134         * src/arm/sysv.S: Add unwinder directives.
1135
1136         * src/arm/ffi.c (ffi_prep_args): Align structs by at least 4 bytes.
1137         Only treat r0 as a struct address if we're actually returning a
1138         struct by address.
1139         Only copy the bytes that are actually within a struct.
1140         (ffi_prep_cif_machdep): A Composite Type not larger than 4 bytes
1141         is returned in r0, not passed by address.
1142         (ffi_call): Allocate a word-sized temporary for the case where
1143         a composite is returned in r0.
1144         (ffi_prep_incoming_args_SYSV): Align as necessary.
1145
1146 2007-08-05  Steven Newbury  <s_j_newbury@yahoo.co.uk>
1147
1148         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Use __clear_cache instead of
1149         directly using the sys_cacheflush syscall.
1150
1151 2007-07-27  Andrew Haley  <aph@redhat.com>
1152
1153         * src/arm/sysv.S (ffi_closure_SYSV): Add soft-float.
1154
1155 2007-09-03  Maciej W. Rozycki  <macro@linux-mips.org>
1156
1157         * Makefile.am: Unify MIPS_IRIX and MIPS_LINUX into MIPS.
1158         * configure.ac: Likewise.
1159         * Makefile.in: Regenerate.
1160         * include/Makefile.in: Likewise.
1161         * testsuite/Makefile.in: Likewise.
1162         * configure: Likewise.
1163
1164 2007-08-24  David Daney  <ddaney@avtrex.com>
1165
1166         * testsuite/libffi.call/return_sl.c: New test.
1167
1168 2007-08-10  David Daney  <ddaney@avtrex.com>
1169
1170         * testsuite/libffi.call/cls_multi_ushort.c,
1171         testsuite/libffi.call/cls_align_uint16.c,
1172         testsuite/libffi.call/nested_struct1.c,
1173         testsuite/libffi.call/nested_struct3.c,
1174         testsuite/libffi.call/cls_7_1_byte.c,
1175         testsuite/libffi.call/nested_struct5.c,
1176         testsuite/libffi.call/cls_double.c,
1177         testsuite/libffi.call/nested_struct7.c,
1178         testsuite/libffi.call/cls_sint.c,
1179         testsuite/libffi.call/nested_struct9.c,
1180         testsuite/libffi.call/cls_20byte1.c,
1181         testsuite/libffi.call/cls_multi_sshortchar.c,
1182         testsuite/libffi.call/cls_align_sint64.c,
1183         testsuite/libffi.call/cls_3byte2.c,
1184         testsuite/libffi.call/cls_multi_schar.c,
1185         testsuite/libffi.call/cls_multi_uchar.c,
1186         testsuite/libffi.call/cls_19byte.c,
1187         testsuite/libffi.call/cls_9byte1.c,
1188         testsuite/libffi.call/cls_align_float.c,
1189         testsuite/libffi.call/closure_fn1.c,
1190         testsuite/libffi.call/problem1.c,
1191         testsuite/libffi.call/closure_fn3.c,
1192         testsuite/libffi.call/cls_sshort.c,
1193         testsuite/libffi.call/closure_fn5.c,
1194         testsuite/libffi.call/cls_align_double.c,
1195         testsuite/libffi.call/nested_struct.c,
1196         testsuite/libffi.call/cls_2byte.c,
1197         testsuite/libffi.call/nested_struct10.c,
1198         testsuite/libffi.call/cls_4byte.c,
1199         testsuite/libffi.call/cls_6byte.c,
1200         testsuite/libffi.call/cls_8byte.c,
1201         testsuite/libffi.call/cls_multi_sshort.c,
1202         testsuite/libffi.call/cls_align_sint16.c,
1203         testsuite/libffi.call/cls_align_uint32.c,
1204         testsuite/libffi.call/cls_20byte.c,
1205         testsuite/libffi.call/cls_float.c,
1206         testsuite/libffi.call/nested_struct2.c,
1207         testsuite/libffi.call/cls_5_1_byte.c,
1208         testsuite/libffi.call/nested_struct4.c,
1209         testsuite/libffi.call/cls_24byte.c,
1210         testsuite/libffi.call/nested_struct6.c,
1211         testsuite/libffi.call/cls_64byte.c,
1212         testsuite/libffi.call/nested_struct8.c,
1213         testsuite/libffi.call/cls_uint.c,
1214         testsuite/libffi.call/cls_multi_ushortchar.c,
1215         testsuite/libffi.call/cls_schar.c,
1216         testsuite/libffi.call/cls_uchar.c,
1217         testsuite/libffi.call/cls_align_uint64.c,
1218         testsuite/libffi.call/cls_ulonglong.c,
1219         testsuite/libffi.call/cls_align_longdouble.c,
1220         testsuite/libffi.call/cls_1_1byte.c,
1221         testsuite/libffi.call/cls_12byte.c,
1222         testsuite/libffi.call/cls_3_1byte.c,
1223         testsuite/libffi.call/cls_3byte1.c,
1224         testsuite/libffi.call/cls_4_1byte.c,
1225         testsuite/libffi.call/cls_6_1_byte.c,
1226         testsuite/libffi.call/cls_16byte.c,
1227         testsuite/libffi.call/cls_18byte.c,
1228         testsuite/libffi.call/closure_fn0.c,
1229         testsuite/libffi.call/cls_9byte2.c,
1230         testsuite/libffi.call/closure_fn2.c,
1231         testsuite/libffi.call/closure_fn4.c,
1232         testsuite/libffi.call/cls_ushort.c,
1233         testsuite/libffi.call/closure_fn6.c,
1234         testsuite/libffi.call/cls_5byte.c,
1235         testsuite/libffi.call/cls_align_pointer.c,
1236         testsuite/libffi.call/cls_7byte.c,
1237         testsuite/libffi.call/cls_align_sint32.c,
1238         testsuite/libffi.special/unwindtest_ffi_call.cc,
1239         testsuite/libffi.special/unwindtest.cc: Remove xfail for mips64*-*-*.
1240
1241 2007-08-10  David Daney  <ddaney@avtrex.com>
1242
1243         PR libffi/28313
1244         * configure.ac: Don't treat mips64 as a special case.
1245         * Makefile.am (nodist_libffi_la_SOURCES): Add n32.S.
1246         * configure: Regenerate
1247         * Makefile.in: Ditto.
1248         * fficonfig.h.in: Ditto.
1249         * src/mips/ffitarget.h (REG_L, REG_S, SUBU, ADDU, SRL, LI): Indent.
1250         (LA, EH_FRAME_ALIGN, FDE_ADDR_BYTES): New preprocessor macros.
1251         (FFI_DEFAULT_ABI): Set for n64 case.
1252         (FFI_CLOSURES, FFI_TRAMPOLINE_SIZE): Define for n32 and n64 cases.
1253         * src/mips/n32.S (ffi_call_N32): Add debug macros and labels for FDE.
1254         (ffi_closure_N32): New function.
1255         (.eh_frame): New section
1256         * src/mips/o32.S: Clean up comments.
1257         (ffi_closure_O32): Pass ffi_closure parameter in $12.
1258         * src/mips/ffi.c: Use FFI_MIPS_N32 instead of
1259         _MIPS_SIM == _ABIN32 throughout.
1260         (FFI_MIPS_STOP_HERE): New, use in place of
1261         ffi_stop_here.
1262         (ffi_prep_args): Use unsigned long to hold pointer values.  Rewrite
1263         to support n32/n64 ABIs.
1264         (calc_n32_struct_flags): Rewrite.
1265         (calc_n32_return_struct_flags): Remove unused variable.  Reverse
1266         position of flag bits.
1267         (ffi_prep_cif_machdep): Rewrite n32 portion.
1268         (ffi_call): Enable for n64.  Add special handling for small structure
1269         return values.
1270         (ffi_prep_closure_loc): Add n32 and n64 support.
1271         (ffi_closure_mips_inner_O32): Add cast to silence warning.
1272         (copy_struct_N32, ffi_closure_mips_inner_N32): New functions.
1273
1274 2007-08-08  David Daney  <ddaney@avtrex.com>
1275
1276         * testsuite/libffi.call/ffitest.h (ffi_type_mylong): Remove definition.
1277         * testsuite/libffi.call/cls_align_uint16.c (main): Use correct type
1278         specifiers.
1279         * testsuite/libffi.call/nested_struct1.c (main): Ditto.
1280         * testsuite/libffi.call/cls_sint.c (main): Ditto.
1281         * testsuite/libffi.call/nested_struct9.c (main): Ditto.
1282         * testsuite/libffi.call/cls_20byte1.c (main): Ditto.
1283         * testsuite/libffi.call/cls_9byte1.c (main): Ditto.
1284         * testsuite/libffi.call/closure_fn1.c (main): Ditto.
1285         * testsuite/libffi.call/closure_fn3.c (main): Ditto.
1286         * testsuite/libffi.call/return_dbl2.c (main): Ditto.
1287         * testsuite/libffi.call/cls_sshort.c (main): Ditto.
1288         * testsuite/libffi.call/return_fl3.c (main): Ditto.
1289         * testsuite/libffi.call/closure_fn5.c (main): Ditto.
1290         * testsuite/libffi.call/nested_struct.c (main): Ditto.
1291         * testsuite/libffi.call/nested_struct10.c (main): Ditto.
1292         * testsuite/libffi.call/return_ll1.c (main): Ditto.
1293         * testsuite/libffi.call/cls_8byte.c (main): Ditto.
1294         * testsuite/libffi.call/cls_align_uint32.c (main): Ditto.
1295         * testsuite/libffi.call/cls_align_sint16.c (main): Ditto.
1296         * testsuite/libffi.call/cls_20byte.c (main): Ditto.
1297         * testsuite/libffi.call/nested_struct2.c (main): Ditto.
1298         * testsuite/libffi.call/cls_24byte.c (main): Ditto.
1299         * testsuite/libffi.call/nested_struct6.c (main): Ditto.
1300         * testsuite/libffi.call/cls_uint.c (main): Ditto.
1301         * testsuite/libffi.call/cls_12byte.c (main): Ditto.
1302         * testsuite/libffi.call/cls_16byte.c (main): Ditto.
1303         * testsuite/libffi.call/closure_fn0.c (main): Ditto.
1304         * testsuite/libffi.call/cls_9byte2.c (main): Ditto.
1305         * testsuite/libffi.call/closure_fn2.c (main): Ditto.
1306         * testsuite/libffi.call/return_dbl1.c (main): Ditto.
1307         * testsuite/libffi.call/closure_fn4.c (main): Ditto.
1308         * testsuite/libffi.call/closure_fn6.c (main): Ditto.
1309         * testsuite/libffi.call/cls_align_sint32.c (main): Ditto.
1310
1311 2007-08-07  Andrew Haley  <aph@redhat.com>
1312
1313         * src/x86/sysv.S (ffi_closure_raw_SYSV): Fix typo in previous
1314         checkin.
1315
1316 2007-08-06  Andrew Haley  <aph@redhat.com>
1317
1318         PR testsuite/32843
1319         * src/x86/sysv.S (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1320         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1321         FFI_TYPE_SINT32.
1322
1323 2007-08-02  David Daney  <ddaney@avtrex.com>
1324
1325         * testsuite/libffi.call/return_ul.c (main): Define return type as
1326         ffi_arg.  Use proper printf conversion specifier.
1327
1328 2007-07-30  Andrew Haley  <aph@redhat.com>
1329
1330         PR testsuite/32843
1331         * src/x86/ffi.c (ffi_prep_cif_machdep): in x86 case, add code for
1332         signed/unsigned int8/16.
1333         * src/x86/sysv.S (ffi_call_SYSV): Rewrite to:
1334         Use a jump table.
1335         Remove code to pop args from the stack after call.
1336         Special-case signed/unsigned int8/16.
1337         * testsuite/libffi.call/return_sc.c (main): Revert.
1338
1339 2007-07-26  Richard Guenther  <rguenther@suse.de>
1340
1341         PR testsuite/32843
1342         * testsuite/libffi.call/return_sc.c (main): Verify call
1343         result as signed char, not ffi_arg.
1344
1345 2007-07-16  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1346
1347         * configure.ac (i?86-*-solaris2.1[0-9]): Set TARGET to X86_64.
1348         * configure: Regenerate.
1349
1350 2007-07-11  David Daney  <ddaney@avtrex.com>
1351
1352         * src/mips/ffi.c: Don't include sys/cachectl.h.
1353         (ffi_prep_closure_loc): Use __builtin___clear_cache() instead of
1354         cacheflush().
1355
1356 2007-05-18  Aurelien Jarno  <aurelien@aurel32.net>
1357
1358         * src/arm/ffi.c (ffi_prep_closure_loc): Renamed and ajusted
1359         from (ffi_prep_closure): ... this.
1360         (FFI_INIT_TRAMPOLINE): Adjust.
1361
1362 2005-12-31  Phil Blundell  <pb@reciva.com>
1363
1364         * src/arm/ffi.c (ffi_prep_incoming_args_SYSV,
1365         ffi_closure_SYSV_inner, ffi_prep_closure): New, add closure support.
1366         * src/arm/sysv.S(ffi_closure_SYSV): Likewise.
1367         * src/arm/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1368         (FFI_CLOSURES): Enable closure support.
1369
1370 2007-07-03  Andrew Haley  <aph@hedges.billgatliff.com>
1371
1372         * testsuite/libffi.call/cls_multi_ushort.c,
1373         testsuite/libffi.call/cls_align_uint16.c,
1374         testsuite/libffi.call/nested_struct1.c,
1375         testsuite/libffi.call/nested_struct3.c,
1376         testsuite/libffi.call/cls_7_1_byte.c,
1377         testsuite/libffi.call/cls_double.c,
1378         testsuite/libffi.call/nested_struct5.c,
1379         testsuite/libffi.call/nested_struct7.c,
1380         testsuite/libffi.call/cls_sint.c,
1381         testsuite/libffi.call/nested_struct9.c,
1382         testsuite/libffi.call/cls_20byte1.c,
1383         testsuite/libffi.call/cls_multi_sshortchar.c,
1384         testsuite/libffi.call/cls_align_sint64.c,
1385         testsuite/libffi.call/cls_3byte2.c,
1386         testsuite/libffi.call/cls_multi_schar.c,
1387         testsuite/libffi.call/cls_multi_uchar.c,
1388         testsuite/libffi.call/cls_19byte.c,
1389         testsuite/libffi.call/cls_9byte1.c,
1390         testsuite/libffi.call/cls_align_float.c,
1391         testsuite/libffi.call/closure_fn1.c,
1392         testsuite/libffi.call/problem1.c,
1393         testsuite/libffi.call/closure_fn3.c,
1394         testsuite/libffi.call/cls_sshort.c,
1395         testsuite/libffi.call/closure_fn5.c,
1396         testsuite/libffi.call/cls_align_double.c,
1397         testsuite/libffi.call/cls_2byte.c,
1398         testsuite/libffi.call/nested_struct.c,
1399         testsuite/libffi.call/nested_struct10.c,
1400         testsuite/libffi.call/cls_4byte.c,
1401         testsuite/libffi.call/cls_6byte.c,
1402         testsuite/libffi.call/cls_8byte.c,
1403         testsuite/libffi.call/cls_multi_sshort.c,
1404         testsuite/libffi.call/cls_align_uint32.c,
1405         testsuite/libffi.call/cls_align_sint16.c,
1406         testsuite/libffi.call/cls_float.c,
1407         testsuite/libffi.call/cls_20byte.c,
1408         testsuite/libffi.call/cls_5_1_byte.c,
1409         testsuite/libffi.call/nested_struct2.c,
1410         testsuite/libffi.call/cls_24byte.c,
1411         testsuite/libffi.call/nested_struct4.c,
1412         testsuite/libffi.call/nested_struct6.c,
1413         testsuite/libffi.call/cls_64byte.c,
1414         testsuite/libffi.call/nested_struct8.c,
1415         testsuite/libffi.call/cls_uint.c,
1416         testsuite/libffi.call/cls_multi_ushortchar.c,
1417         testsuite/libffi.call/cls_schar.c,
1418         testsuite/libffi.call/cls_uchar.c,
1419         testsuite/libffi.call/cls_align_uint64.c,
1420         testsuite/libffi.call/cls_ulonglong.c,
1421         testsuite/libffi.call/cls_align_longdouble.c,
1422         testsuite/libffi.call/cls_1_1byte.c,
1423         testsuite/libffi.call/cls_12byte.c,
1424         testsuite/libffi.call/cls_3_1byte.c,
1425         testsuite/libffi.call/cls_3byte1.c,
1426         testsuite/libffi.call/cls_4_1byte.c,
1427         testsuite/libffi.call/cls_6_1_byte.c,
1428         testsuite/libffi.call/cls_16byte.c,
1429         testsuite/libffi.call/cls_18byte.c,
1430         testsuite/libffi.call/closure_fn0.c,
1431         testsuite/libffi.call/cls_9byte2.c,
1432         testsuite/libffi.call/closure_fn2.c,
1433         testsuite/libffi.call/closure_fn4.c,
1434         testsuite/libffi.call/cls_ushort.c,
1435         testsuite/libffi.call/closure_fn6.c,
1436         testsuite/libffi.call/cls_5byte.c,
1437         testsuite/libffi.call/cls_align_pointer.c,
1438         testsuite/libffi.call/cls_7byte.c,
1439         testsuite/libffi.call/cls_align_sint32.c,
1440         testsuite/libffi.special/unwindtest_ffi_call.cc,
1441         testsuite/libffi.special/unwindtest.cc: Enable for ARM.
1442
1443 2007-07-05  H.J. Lu  <hongjiu.lu@intel.com>
1444
1445         * aclocal.m4: Regenerated.
1446
1447 2007-06-02  Paolo Bonzini  <bonzini@gnu.org>
1448
1449         * configure: Regenerate.
1450
1451 2007-05-23  Steve Ellcey  <sje@cup.hp.com>
1452
1453         * Makefile.in: Regenerate.
1454         * configure: Regenerate.
1455         * aclocal.m4: Regenerate.
1456         * include/Makefile.in: Regenerate.
1457         * testsuite/Makefile.in: Regenerate.
1458
1459 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1460
1461         * src/m68k/ffi.c (ffi_prep_incoming_args_SYSV,
1462         ffi_closure_SYSV_inner,ffi_prep_closure): New, add closure support.
1463         * src/m68k/sysv.S(ffi_closure_SYSV,ffi_closure_struct_SYSV): Likewise.
1464         * src/m68k/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1465         (FFI_CLOSURES): Enable closure support.
1466
1467 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1468
1469         * configure.ac (HAVE_AS_CFI_PSEUDO_OP): New test.
1470         * configure: Regenerate.
1471         * fficonfig.h.in: Regenerate.
1472         * src/m68k/sysv.S (CFI_STARTPROC,CFI_ENDPROC,
1473         CFI_OFFSET,CFI_DEF_CFA): New macros.
1474         (ffi_call_SYSV): Add callframe annotation.
1475
1476 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1477
1478         * src/m68k/ffi.c (ffi_prep_args,ffi_prep_cif_machdep): Fix
1479         numerous test suite failures.
1480         * src/m68k/sysv.S (ffi_call_SYSV): Likewise.
1481
1482 2007-04-11  Paolo Bonzini  <bonzini@gnu.org>
1483
1484         * Makefile.am (EXTRA_DIST): Bring up to date.
1485         * Makefile.in: Regenerate.
1486         * src/frv/eabi.S: Remove RCS keyword.
1487
1488 2007-04-06  Richard Henderson  <rth@redhat.com>
1489
1490         * configure.ac: Tidy target case.
1491         (HAVE_LONG_DOUBLE): Allow the target to override.
1492         * configure: Regenerate.
1493         * include/ffi.h.in: Don't define ffi_type_foo if
1494         LIBFFI_HIDE_BASIC_TYPES is defined.
1495         (ffi_type_longdouble): If not HAVE_LONG_DOUBLE, define
1496         to ffi_type_double.
1497         * types.c (LIBFFI_HIDE_BASIC_TYPES): Define.
1498         (FFI_TYPEDEF, ffi_type_void): Mark the data const.
1499         (ffi_type_longdouble): Special case for Alpha.  Don't define
1500         if long double == double.
1501
1502         * src/alpha/ffi.c (FFI_TYPE_LONGDOUBLE): Assert unique value.
1503         (ffi_prep_cif_machdep): Handle it as the 128-bit type.
1504         (ffi_call, ffi_closure_osf_inner): Likewise.
1505         (ffi_closure_osf_inner): Likewise.  Mark hidden.
1506         (ffi_call_osf, ffi_closure_osf): Mark hidden.
1507         * src/alpha/ffitarget.h (FFI_LAST_ABI): Tidy definition.
1508         * src/alpha/osf.S (ffi_call_osf, ffi_closure_osf): Mark hidden.
1509         (load_table): Handle 128-bit long double.
1510
1511         * testsuite/libffi.call/float4.c: Add -mieee for alpha.
1512
1513 2007-04-06  Tom Tromey  <tromey@redhat.com>
1514
1515         PR libffi/31491:
1516         * README: Fixed bug in example.
1517
1518 2007-04-03  Jakub Jelinek  <jakub@redhat.com>
1519
1520         * src/closures.c: Include sys/statfs.h.
1521         (_GNU_SOURCE): Define on Linux.
1522         (FFI_MMAP_EXEC_SELINUX): Define.
1523         (selinux_enabled): New variable.
1524         (selinux_enabled_check): New function.
1525         (is_selinux_enabled): Define.
1526         (dlmmap): Use it.
1527
1528 2007-03-24  Uros Bizjak  <ubizjak@gmail.com>
1529
1530         * testsuite/libffi.call/return_fl2.c (return_fl): Mark as static.
1531         Use 'volatile float sum' to create sum of floats to avoid false
1532         negative due to excess precision on ix86 targets.
1533         (main): Ditto.
1534
1535 2007-03-08  Alexandre Oliva  <aoliva@redhat.com>
1536
1537         * src/powerpc/ffi.c (flush_icache): Fix left-over from previous
1538         patch.
1539         (ffi_prep_closure_loc): Remove unneeded casts.  Add needed ones.
1540
1541 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1542
1543         * include/ffi.h.in (ffi_closure_alloc, ffi_closure_free): New.
1544         (ffi_prep_closure_loc): New.
1545         (ffi_prep_raw_closure_loc): New.
1546         (ffi_prep_java_raw_closure_loc): New.
1547         * src/closures.c: New file.
1548         * src/dlmalloc.c [FFI_MMAP_EXEC_WRIT] (struct malloc_segment):
1549         Replace sflags with exec_offset.
1550         [FFI_MMAP_EXEC_WRIT] (mmap_exec_offset, add_segment_exec_offset,
1551         sub_segment_exec_offset): New macros.
1552         (get_segment_flags, set_segment_flags, check_segment_merge): New
1553         macros.
1554         (is_mmapped_segment, is_extern_segment): Use get_segment_flags.
1555         (add_segment, sys_alloc, create_mspace, create_mspace_with_base,
1556         destroy_mspace): Use new macros.
1557         (sys_alloc): Silence warning.
1558         * Makefile.am (libffi_la_SOURCES): Add src/closures.c.
1559         * Makefile.in: Rebuilt.
1560         * src/prep_cif [FFI_CLOSURES] (ffi_prep_closure): Implement in
1561         terms of ffi_prep_closure_loc.
1562         * src/raw_api.c (ffi_prep_raw_closure_loc): Renamed and adjusted
1563         from...
1564         (ffi_prep_raw_closure): ... this.  Re-implement in terms of the
1565         renamed version.
1566         * src/java_raw_api (ffi_prep_java_raw_closure_loc): Renamed and
1567         adjusted from...
1568         (ffi_prep_java_raw_closure): ... this.  Re-implement in terms of
1569         the renamed version.
1570         * src/alpha/ffi.c (ffi_prep_closure_loc): Renamed from
1571         (ffi_prep_closure): ... this.
1572         * src/pa/ffi.c: Likewise.
1573         * src/cris/ffi.c: Likewise.  Adjust.
1574         * src/frv/ffi.c: Likewise.
1575         * src/ia64/ffi.c: Likewise.
1576         * src/mips/ffi.c: Likewise.
1577         * src/powerpc/ffi_darwin.c: Likewise.
1578         * src/s390/ffi.c: Likewise.
1579         * src/sh/ffi.c: Likewise.
1580         * src/sh64/ffi.c: Likewise.
1581         * src/sparc/ffi.c: Likewise.
1582         * src/x86/ffi64.c: Likewise.
1583         * src/x86/ffi.c: Likewise.
1584         (FFI_INIT_TRAMPOLINE): Adjust.
1585         (ffi_prep_raw_closure_loc): Renamed and adjusted from...
1586         (ffi_prep_raw_closure): ... this.
1587         * src/powerpc/ffi.c (ffi_prep_closure_loc): Renamed from
1588         (ffi_prep_closure): ... this.
1589         (flush_icache): Adjust.
1590
1591 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1592
1593         * src/dlmalloc.c: New file, imported version 2.8.3 of Doug
1594         Lea's malloc.
1595
1596 2007-03-01  Brooks Moses  <brooks.moses@codesourcery.com>
1597
1598         * Makefile.am: Add dummy install-pdf target.
1599         * Makefile.in: Regenerate
1600
1601 2007-02-13  Andreas Krebbel  <krebbel1@de.ibm.com>
1602
1603         * src/s390/ffi.c (ffi_prep_args, ffi_prep_cif_machdep,
1604         ffi_closure_helper_SYSV): Add long double handling.
1605
1606 2007-02-02  Jakub Jelinek  <jakub@redhat.com>
1607
1608         * src/powerpc/linux64.S (ffi_call_LINUX64): Move restore of r2
1609         immediately after bctrl instruction.
1610
1611 2007-01-18  Alexandre Oliva  <aoliva@redhat.com>
1612
1613         * Makefile.am (all-recursive, install-recursive,
1614         mostlyclean-recursive, clean-recursive, distclean-recursive,
1615         maintainer-clean-recursive): Add missing targets.
1616         * Makefile.in: Rebuilt.
1617
1618 2006-12-14  Andreas Tobler  <a.tobler@schweiz.org>
1619
1620         * configure.ac: Add TARGET for x86_64-*-darwin*.
1621         * Makefile.am (nodist_libffi_la_SOURCES): Add rules for 64-bit sources
1622         for X86_DARWIN.
1623         * src/x86/ffitarget.h: Set trampoline size for x86_64-*-darwin*.
1624         * src/x86/darwin64.S: New file for x86_64-*-darwin* support.
1625         * configure: Regenerate.
1626         * Makefile.in: Regenerate.
1627         * include/Makefile.in: Regenerate.
1628         * testsuite/Makefile.in: Regenerate.
1629         * testsuite/libffi.special/unwindtest_ffi_call.cc: New test case for
1630         ffi_call only.
1631
1632 2006-12-13  Andreas Tobler <a.tobler@schweiz.org>
1633
1634         * aclocal.m4: Regenerate with aclocal -I .. as written in the
1635         Makefile.am.
1636
1637 2006-10-31  Geoffrey Keating  <geoffk@apple.com>
1638
1639         * src/powerpc/ffi_darwin.c (darwin_adjust_aggregate_sizes): New.
1640         (ffi_prep_cif_machdep): Call darwin_adjust_aggregate_sizes for
1641         Darwin.
1642         * testsuite/libffi.call/nested_struct4.c: Remove Darwin XFAIL.
1643         * testsuite/libffi.call/nested_struct6.c: Remove Darwin XFAIL.
1644
1645 2006-10-10  Paolo Bonzini  <bonzini@gnu.org>
1646             Sandro Tolaini  <tolaini@libero.it>
1647
1648         * configure.ac [i*86-*-darwin*]: Set X86_DARWIN symbol and
1649         conditional.
1650         * configure: Regenerated.
1651         * Makefile.am (nodist_libffi_la_SOURCES) [X86_DARWIN]: New case.
1652         (EXTRA_DIST): Add src/x86/darwin.S.
1653         * Makefile.in: Regenerated.
1654         * include/Makefile.in: Regenerated.
1655         * testsuite/Makefile.in: Regenerated.
1656
1657         * src/x86/ffi.c (ffi_prep_cif_machdep) [X86_DARWIN]: Treat like
1658         X86_WIN32, and additionally align stack to 16 bytes.
1659         * src/x86/darwin.S: New, based on sysv.S.
1660         * src/prep_cif.c (ffi_prep_cif) [X86_DARWIN]: Align > 8-byte structs.
1661
1662 2006-09-12  David Daney  <ddaney@avtrex.com>
1663
1664         PR libffi/23935
1665         * include/Makefile.am: Install both ffi.h and ffitarget.h in
1666         $(libdir)/gcc/$(target_alias)/$(gcc_version)/include.
1667         * aclocal.m4: Regenerated for automake 1.9.6.
1668         * Makefile.in: Regenerated.
1669         * include/Makefile.in: Regenerated.
1670         * testsuite/Makefile.in: Regenerated.
1671
1672 2006-08-17  Andreas Tobler  <a.tobler@schweiz.ch>
1673
1674         * include/ffi_common.h (struct): Revert accidental commit.
1675
1676 2006-08-15  Andreas Tobler  <a.tobler@schweiz.ch>
1677
1678         * include/ffi_common.h: Remove lint directives.
1679         * include/ffi.h.in: Likewise.
1680
1681 2006-07-25  Torsten Schoenfeld  <kaffeetisch@gmx.de>
1682
1683         * include/ffi.h.in (ffi_type_ulong, ffi_type_slong): Define correctly
1684         for 32-bit architectures.
1685         * testsuite/libffi.call/return_ul.c: New test case.
1686
1687 2006-07-19  David Daney  <ddaney@avtrex.com>
1688
1689         * testsuite/libffi.call/closure_fn6.c: Remove xfail for mips,
1690         xfail remains for mips64.
1691
1692 2006-05-23  Carlos O'Donell  <carlos@codesourcery.com>
1693
1694         * Makefile.am: Add install-html target. Add install-html to .PHONY
1695         * Makefile.in: Regenerate.
1696         * aclocal.m4: Regenerate.
1697         * include/Makefile.in: Regenerate.
1698         * testsuite/Makefile.in: Regenerate.
1699
1700 2006-05-18  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1701
1702         * pa/ffi.c (ffi_prep_args_pa32): Load floating point arguments from
1703         stack slot.
1704
1705 2006-04-22  Andreas Tobler  <a.tobler@schweiz.ch>
1706
1707         * README: Remove notice about 'Crazy Comments'.
1708         * src/debug.c: Remove lint directives. Cleanup white spaces.
1709         * src/java_raw_api.c: Likewise.
1710         * src/prep_cif.c: Likewise.
1711         * src/raw_api.c: Likewise.
1712         * src/ffitest.c: Delete. No longer needed, all test cases migrated
1713         to the testsuite.
1714         * src/arm/ffi.c: Remove lint directives.
1715         * src/m32r/ffi.c: Likewise.
1716         * src/pa/ffi.c: Likewise.
1717         * src/powerpc/ffi.c: Likewise.
1718         * src/powerpc/ffi_darwin.c: Likewise.
1719         * src/sh/ffi.c: Likewise.
1720         * src/sh64/ffi.c: Likewise.
1721         * src/x86/ffi.c: Likewise.
1722         * testsuite/libffi.call/float2.c: Likewise.
1723         * testsuite/libffi.call/promotion.c: Likewise.
1724         * testsuite/libffi.call/struct1.c: Likewise.
1725
1726 2006-04-13  Andreas Tobler  <a.tobler@schweiz.ch>
1727
1728         * src/pa/hpux32.S: Correct unwind offset calculation for
1729         ffi_closure_pa32.
1730         * src/pa/linux.S: Likewise.
1731
1732 2006-04-12  James E Wilson  <wilson@specifix.com>
1733
1734         PR libgcj/26483
1735         * src/ia64/ffi.c (stf_spill, ldf_fill): Rewrite as macros.
1736         (hfa_type_load): Call stf_spill.
1737         (hfa_type_store): Call ldf_fill.
1738         (ffi_call): Adjust calls to above routines.  Add local temps for
1739         macro result.
1740
1741 2006-04-10  Matthias Klose  <doko@debian.org>
1742
1743         * testsuite/lib/libffi-dg.exp (libffi-init): Recognize multilib
1744         directory names containing underscores.
1745
1746 2006-04-07  James E Wilson  <wilson@specifix.com>
1747
1748         * testsuite/libffi.call/float4.c: New testcase.
1749
1750 2006-04-05  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1751             Andreas Tobler  <a.tobler@schweiz.ch>
1752
1753         * Makefile.am: Add PA_HPUX port.
1754         * Makefile.in: Regenerate.
1755         * include/Makefile.in: Likewise.
1756         * testsuite/Makefile.in: Likewise.
1757         * configure.ac: Add PA_HPUX rules.
1758         * configure: Regenerate.
1759         * src/pa/ffitarget.h: Rename linux target to PA_LINUX.
1760         Add PA_HPUX and PA64_HPUX.
1761         Rename FFI_LINUX ABI to FFI_PA32 ABI.
1762         (FFI_TRAMPOLINE_SIZE): Define for 32-bit HP-UX targets.
1763         (FFI_TYPE_SMALL_STRUCT2): Define.
1764         (FFI_TYPE_SMALL_STRUCT4): Likewise.
1765         (FFI_TYPE_SMALL_STRUCT8): Likewise.
1766         (FFI_TYPE_SMALL_STRUCT3): Redefine.
1767         (FFI_TYPE_SMALL_STRUCT5): Likewise.
1768         (FFI_TYPE_SMALL_STRUCT6): Likewise.
1769         (FFI_TYPE_SMALL_STRUCT7): Likewise.
1770         * src/pa/ffi.c (ROUND_DOWN): Delete.
1771         (fldw, fstw, fldd, fstd): Use '__asm__'.
1772         (ffi_struct_type): Add support for FFI_TYPE_SMALL_STRUCT2,
1773         FFI_TYPE_SMALL_STRUCT4 and FFI_TYPE_SMALL_STRUCT8.
1774         (ffi_prep_args_LINUX): Rename to ffi_prep_args_pa32. Update comment.
1775         Simplify incrementing of stack slot variable. Change type of local
1776         'n' to unsigned int.
1777         (ffi_size_stack_LINUX): Rename to ffi_size_stack_pa32. Handle long
1778         double on PA_HPUX.
1779         (ffi_prep_cif_machdep): Likewise.
1780         (ffi_call): Likewise.
1781         (ffi_closure_inner_LINUX): Rename to ffi_closure_inner_pa32. Change
1782         return type to ffi_status. Simplify incrementing of stack slot
1783         variable. Only copy floating point argument registers when PA_LINUX
1784         is true. Reformat debug statement.
1785         Add support for FFI_TYPE_SMALL_STRUCT2, FFI_TYPE_SMALL_STRUCT4 and
1786         FFI_TYPE_SMALL_STRUCT8.
1787         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Add 'extern' to
1788         declaration.
1789         (ffi_prep_closure): Make linux trampoline conditional on PA_LINUX.
1790         Add nops to cache flush.  Add trampoline for PA_HPUX.
1791         * src/pa/hpux32.S: New file.
1792         * src/pa/linux.S (ffi_call_LINUX): Rename to ffi_call_pa32. Rename
1793         ffi_prep_args_LINUX to ffi_prep_args_pa32.
1794         Localize labels. Add support for 2, 4 and 8-byte small structs. Handle
1795         unaligned destinations in 3, 5, 6 and 7-byte small structs. Order
1796         argument type checks so that common argument types appear first.
1797         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Rename
1798         ffi_closure_inner_LINUX to ffi_closure_inner_pa32.
1799
1800 2006-03-24  Alan Modra  <amodra@bigpond.net.au>
1801
1802         * src/powerpc/ffitarget.h (enum ffi_abi): Add FFI_LINUX.  Default
1803         for 32-bit using IBM extended double format.  Fix FFI_LAST_ABI.
1804         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Handle linux variant of
1805         FFI_TYPE_LONGDOUBLE.
1806         (ffi_prep_args64): Assert using IBM extended double.
1807         (ffi_prep_cif_machdep): Don't munge FFI_TYPE_LONGDOUBLE type.
1808         Handle FFI_LINUX FFI_TYPE_LONGDOUBLE return and args.
1809         (ffi_call): Handle FFI_LINUX.
1810         (ffi_closure_helper_SYSV): Non FFI_LINUX long double return needs
1811         gpr3 return pointer as for struct return.  Handle FFI_LINUX
1812         FFI_TYPE_LONGDOUBLE return and args.  Don't increment "nf"
1813         unnecessarily.
1814         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Load both f1 and f2
1815         for FFI_TYPE_LONGDOUBLE.  Move epilogue insns into case table.
1816         Don't use r6 as pointer to results, instead use sp offset.  Don't
1817         make a special call to load lr with case table address, instead
1818         use offset from previous call.
1819         * src/powerpc/sysv.S (ffi_call_SYSV): Save long double return.
1820         * src/powerpc/linux64.S (ffi_call_LINUX64): Simplify long double
1821         return.
1822
1823 2006-03-15  Kaz Kojima  <kkojima@gcc.gnu.org>
1824
1825         * src/sh64/ffi.c (ffi_prep_cif_machdep): Handle float arguments
1826         passed with FP registers correctly.
1827         (ffi_closure_helper_SYSV): Likewise.
1828         * src/sh64/sysv.S: Likewise.
1829
1830 2006-03-01  Andreas Tobler  <a.tobler@schweiz.ch>
1831
1832         * testsuite/libffi.special/unwindtest.cc (closure_test_fn): Mark cif,
1833         args and userdata unused.
1834         (closure_test_fn1): Mark cif and userdata unused.
1835         (main): Remove unused res.
1836
1837 2006-02-28  Andreas Tobler  <a.tobler@schweiz.ch>
1838
1839         * testsuite/libffi.call/call.exp: Adjust FSF address. Add test runs for
1840         -O2, -O3, -Os and the warning flags -W -Wall.
1841         * testsuite/libffi.special/special.exp: Likewise.
1842         * testsuite/libffi.call/ffitest.h: Add an __UNUSED__ macro to mark
1843         unused parameter unused for gcc or else do nothing.
1844         * testsuite/libffi.special/ffitestcxx.h: Likewise.
1845         * testsuite/libffi.call/cls_12byte.c (cls_struct_12byte_gn): Mark cif
1846         and userdata unused.
1847         * testsuite/libffi.call/cls_16byte.c (cls_struct_16byte_gn): Likewise.
1848         * testsuite/libffi.call/cls_18byte.c (cls_struct_18byte_gn): Likewise.
1849         * testsuite/libffi.call/cls_19byte.c (cls_struct_19byte_gn): Likewise.
1850         * testsuite/libffi.call/cls_1_1byte.c (cls_struct_1_1byte_gn): Likewise.
1851         * testsuite/libffi.call/cls_20byte.c (cls_struct_20byte_gn): Likewise.
1852         * testsuite/libffi.call/cls_20byte1.c (cls_struct_20byte_gn): Likewise.
1853         * testsuite/libffi.call/cls_24byte.c (cls_struct_24byte_gn): Likewise.
1854         * testsuite/libffi.call/cls_2byte.c (cls_struct_2byte_gn): Likewise.
1855         * testsuite/libffi.call/cls_3_1byte.c (cls_struct_3_1byte_gn): Likewise.
1856         * testsuite/libffi.call/cls_3byte1.c (cls_struct_3byte_gn): Likewise.
1857         * testsuite/libffi.call/cls_3byte2.c (cls_struct_3byte_gn1): Likewise.
1858         * testsuite/libffi.call/cls_4_1byte.c (cls_struct_4_1byte_gn): Likewise.
1859         * testsuite/libffi.call/cls_4byte.c (cls_struct_4byte_gn): Likewise.
1860         * testsuite/libffi.call/cls_5_1_byte.c (cls_struct_5byte_gn): Likewise.
1861         * testsuite/libffi.call/cls_5byte.c (cls_struct_5byte_gn): Likewise.
1862         * testsuite/libffi.call/cls_64byte.c (cls_struct_64byte_gn): Likewise.
1863         * testsuite/libffi.call/cls_6_1_byte.c (cls_struct_6byte_gn): Likewise.
1864         * testsuite/libffi.call/cls_6byte.c (cls_struct_6byte_gn): Likewise.
1865         * testsuite/libffi.call/cls_7_1_byte.c (cls_struct_7byte_gn): Likewise.
1866         * testsuite/libffi.call/cls_7byte.c (cls_struct_7byte_gn): Likewise.
1867         * testsuite/libffi.call/cls_8byte.c (cls_struct_8byte_gn): Likewise.
1868         * testsuite/libffi.call/cls_9byte1.c (cls_struct_9byte_gn): Likewise.
1869         * testsuite/libffi.call/cls_9byte2.c (cls_struct_9byte_gn): Likewise.
1870         * testsuite/libffi.call/cls_align_double.c (cls_struct_align_gn):
1871         Likewise.
1872         * testsuite/libffi.call/cls_align_float.c (cls_struct_align_gn):
1873         Likewise.
1874         * testsuite/libffi.call/cls_align_longdouble.c (cls_struct_align_gn):
1875         Likewise.
1876         * testsuite/libffi.call/cls_align_pointer.c (cls_struct_align_fn): Cast
1877         void* to avoid compiler warning.
1878         (main): Likewise.
1879         (cls_struct_align_gn): Mark cif and userdata unused.
1880         * testsuite/libffi.call/cls_align_sint16.c (cls_struct_align_gn):
1881         Likewise.
1882         * testsuite/libffi.call/cls_align_sint32.c (cls_struct_align_gn):
1883         Likewise.
1884         * testsuite/libffi.call/cls_align_sint64.c (cls_struct_align_gn):
1885         Likewise.
1886         * testsuite/libffi.call/cls_align_uint16.c (cls_struct_align_gn):
1887         Likewise.
1888         * testsuite/libffi.call/cls_align_uint32.c (cls_struct_align_gn):
1889         Likewise.
1890         * testsuite/libffi.call/cls_double.c (cls_ret_double_fn): Likewise.
1891         * testsuite/libffi.call/cls_float.c (cls_ret_float_fn): Likewise.
1892         * testsuite/libffi.call/cls_multi_schar.c (test_func_gn): Mark cif and
1893         data unused.
1894         (main): Cast res_call to silence gcc.
1895         * testsuite/libffi.call/cls_multi_sshort.c (test_func_gn): Mark cif and
1896         data unused.
1897         (main): Cast res_call to silence gcc.
1898         * testsuite/libffi.call/cls_multi_sshortchar.c (test_func_gn): Mark cif
1899         and data unused.
1900         (main): Cast res_call to silence gcc.
1901         * testsuite/libffi.call/cls_multi_uchar.c (test_func_gn): Mark cif and
1902         data unused.
1903         (main): Cast res_call to silence gcc.
1904         * testsuite/libffi.call/cls_multi_ushort.c (test_func_gn): Mark cif and
1905         data unused.
1906         (main): Cast res_call to silence gcc.
1907         * testsuite/libffi.call/cls_multi_ushortchar.c (test_func_gn): Mark cif
1908         and data unused.
1909         (main): Cast res_call to silence gcc.
1910         * testsuite/libffi.call/cls_schar.c (cls_ret_schar_fn): Mark cif and
1911         userdata unused.
1912         (cls_ret_schar_fn): Cast printf parameter to silence gcc.
1913         * testsuite/libffi.call/cls_sint.c (cls_ret_sint_fn): Mark cif and
1914         userdata unused.
1915         (cls_ret_sint_fn): Cast printf parameter to silence gcc.
1916         * testsuite/libffi.call/cls_sshort.c (cls_ret_sshort_fn): Mark cif and
1917         userdata unused.
1918         (cls_ret_sshort_fn): Cast printf parameter to silence gcc.
1919         * testsuite/libffi.call/cls_uchar.c (cls_ret_uchar_fn):  Mark cif and
1920         userdata unused.
1921         (cls_ret_uchar_fn): Cast printf parameter to silence gcc.
1922         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Mark cif and
1923         userdata unused.
1924         (cls_ret_uint_fn): Cast printf parameter to silence gcc.
1925         * testsuite/libffi.call/cls_ulonglong.c (cls_ret_ulonglong_fn): Mark cif
1926         and userdata unused.
1927         * testsuite/libffi.call/cls_ushort.c (cls_ret_ushort_fn): Mark cif and
1928         userdata unused.
1929         (cls_ret_ushort_fn): Cast printf parameter to silence gcc.
1930         * testsuite/libffi.call/float.c (floating): Remove unused parameter e.
1931         * testsuite/libffi.call/float1.c (main): Remove unused variable i.
1932         Cleanup white spaces.
1933         * testsuite/libffi.call/negint.c (checking): Remove unused variable i.
1934         * testsuite/libffi.call/nested_struct.c (cls_struct_combined_gn): Mark
1935         cif and userdata unused.
1936         * testsuite/libffi.call/nested_struct1.c (cls_struct_combined_gn):
1937         Likewise.
1938         * testsuite/libffi.call/nested_struct10.c (B_gn): Likewise.
1939         * testsuite/libffi.call/nested_struct2.c (B_fn): Adjust printf
1940         formatters to silence gcc.
1941         (B_gn): Mark cif and userdata unused.
1942         * testsuite/libffi.call/nested_struct3.c (B_gn): Mark cif and userdata
1943         unused.
1944         * testsuite/libffi.call/nested_struct4.c: Mention related PR.
1945         (B_gn): Mark cif and userdata unused.
1946         * testsuite/libffi.call/nested_struct5.c (B_gn): Mark cif and userdata
1947         unused.
1948         * testsuite/libffi.call/nested_struct6.c: Mention related PR.
1949         (B_gn): Mark cif and userdata unused.
1950         * testsuite/libffi.call/nested_struct7.c (B_gn): Mark cif and userdata
1951         unused.
1952         * testsuite/libffi.call/nested_struct8.c (B_gn): Likewise.
1953         * testsuite/libffi.call/nested_struct9.c (B_gn): Likewise.
1954         * testsuite/libffi.call/problem1.c (stub): Likewise.
1955         * testsuite/libffi.call/pyobjc-tc.c (main): Cast the result to silence
1956         gcc.
1957         * testsuite/libffi.call/return_fl2.c (return_fl): Add the note mentioned
1958         in the last commit for this test case in the test case itself.
1959         * testsuite/libffi.call/closure_fn0.c (closure_test_fn0): Mark cif as
1960         unused.
1961         * testsuite/libffi.call/closure_fn1.c (closure_test_fn1): Likewise.
1962         * testsuite/libffi.call/closure_fn2.c (closure_test_fn2): Likewise.
1963         * testsuite/libffi.call/closure_fn3.c (closure_test_fn3): Likewise.
1964         * testsuite/libffi.call/closure_fn4.c (closure_test_fn0): Likewise.
1965         * testsuite/libffi.call/closure_fn5.c (closure_test_fn5): Likewise.
1966         * testsuite/libffi.call/closure_fn6.c (closure_test_fn0): Likewise.
1967
1968 2006-02-22  Kaz Kojima  <kkojima@gcc.gnu.org>
1969
1970         * src/sh/sysv.S: Fix register numbers in the FDE for
1971         ffi_closure_SYSV.
1972
1973 2006-02-20  Andreas Tobler  <a.tobler@schweiz.ch>
1974
1975         * testsuite/libffi.call/return_fl2.c (return_fl): Remove static
1976         declaration to avoid a false negative on ix86. See PR323.
1977
1978 2006-02-18  Kaz Kojima  <kkojima@gcc.gnu.org>
1979
1980         * src/sh/ffi.c (ffi_closure_helper_SYSV): Remove unused variable
1981         and cast integer to void * if needed.  Update the pointer to
1982         the FP register saved area correctly.
1983
1984 2006-02-17  Andreas Tobler  <a.tobler@schweiz.ch>
1985
1986         * testsuite/libffi.call/nested_struct6.c: XFAIL this test until PR25630
1987         is fixed.
1988         * testsuite/libffi.call/nested_struct4.c: Likewise.
1989
1990 2006-02-16  Andreas Tobler  <a.tobler@schweiz.ch>
1991
1992         * testsuite/libffi.call/return_dbl.c: New test case.
1993         * testsuite/libffi.call/return_dbl1.c: Likewise.
1994         * testsuite/libffi.call/return_dbl2.c: Likewise.
1995         * testsuite/libffi.call/return_fl.c: Likewise.
1996         * testsuite/libffi.call/return_fl1.c: Likewise.
1997         * testsuite/libffi.call/return_fl2.c: Likewise.
1998         * testsuite/libffi.call/return_fl3.c: Likewise.
1999         * testsuite/libffi.call/closure_fn6.c: Likewise.
2000
2001         * testsuite/libffi.call/nested_struct2.c: Remove ffi_type_mylong
2002         definition.
2003         * testsuite/libffi.call/ffitest.h: Add ffi_type_mylong definition
2004         here to be used by other test cases too.
2005
2006         * testsuite/libffi.call/nested_struct10.c: New test case.
2007         * testsuite/libffi.call/nested_struct9.c: Likewise.
2008         * testsuite/libffi.call/nested_struct8.c: Likewise.
2009         * testsuite/libffi.call/nested_struct7.c: Likewise.
2010         * testsuite/libffi.call/nested_struct6.c: Likewise.
2011         * testsuite/libffi.call/nested_struct5.c: Likewise.
2012         * testsuite/libffi.call/nested_struct4.c: Likewise.
2013
2014 2006-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
2015
2016         * configure.ac: Enable libffi for sparc64-*-freebsd*.
2017         * configure: Rebuilt.
2018
2019 2006-01-18  Jakub Jelinek  <jakub@redhat.com>
2020
2021         * src/powerpc/sysv.S (smst_two_register): Don't call __ashldi3,
2022         instead do the shifting inline.
2023         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't compute %r5
2024         shift count unconditionally.  Simplify load sequences for 1, 2, 3, 4
2025         and 8 byte structs, for the remaining struct sizes don't call
2026         __lshrdi3, instead do the shifting inline.
2027
2028 2005-12-07  Thiemo Seufer  <ths@networkno.de>
2029
2030         * src/mips/ffitarget.h: Remove obsolete sgidefs.h include. Add
2031         missing parentheses.
2032         * src/mips/o32.S (ffi_call_O32): Code formatting. Define
2033         and use A3_OFF, FP_OFF, RA_OFF. Micro-optimizations.
2034         (ffi_closure_O32): Likewise, but with newly defined A3_OFF2,
2035         A2_OFF2, A1_OFF2, A0_OFF2, RA_OFF2, FP_OFF2, S0_OFF2, GP_OFF2,
2036         V1_OFF2, V0_OFF2, FA_1_1_OFF2, FA_1_0_OFF2, FA_0_1_OFF2,
2037         FA_0_0_OFF2.
2038         * src/mips/ffi.c (ffi_prep_args): Code formatting. Fix
2039         endianness bugs.
2040         (ffi_prep_closure): Improve trampoline instruction scheduling.
2041         (ffi_closure_mips_inner_O32): Fix endianness bugs.
2042
2043 2005-12-03  Alan Modra  <amodra@bigpond.net.au>
2044
2045         * src/powerpc/ffi.c: Formatting.
2046         (ffi_prep_args_SYSV): Avoid possible aliasing problems by using unions.
2047         (ffi_prep_args64): Likewise.
2048
2049 2005-09-30  Geoffrey Keating  <geoffk@apple.com>
2050
2051         * testsuite/lib/libffi-dg.exp (libffi_target_compile): For
2052         darwin, use -shared-libgcc not -lgcc_s, and explain why.
2053
2054 2005-09-26  Tom Tromey  <tromey@redhat.com>
2055
2056         * testsuite/libffi.call/float1.c (value_type): New typedef.
2057         (CANARY): New define.
2058         (main): Check for result buffer overflow.
2059         * src/powerpc/linux64.S: Handle linux64 long double returns.
2060         * src/powerpc/ffi.c (FLAG_RETURNS_128BITS): New constant.
2061         (ffi_prep_cif_machdep): Handle linux64 long double returns.
2062
2063 2005-08-25  Alan Modra  <amodra@bigpond.net.au>
2064
2065         PR target/23404
2066         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Correct placement of stack
2067         homed fp args.
2068         (ffi_status ffi_prep_cif_machdep): Correct stack sizing for same.
2069
2070 2005-08-11  Jakub Jelinek  <jakub@redhat.com>
2071
2072         * configure.ac (HAVE_HIDDEN_VISIBILITY_ATTRIBUTE): New test.
2073         (AH_BOTTOM): Add FFI_HIDDEN definition.
2074         * configure: Rebuilt.
2075         * fficonfig.h.in: Rebuilt.
2076         * src/powerpc/ffi.c (hidden): Remove.
2077         (ffi_closure_LINUX64, ffi_prep_args64, ffi_call_LINUX64,
2078         ffi_closure_helper_LINUX64): Use FFI_HIDDEN instead of hidden.
2079         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64,
2080         .ffi_closure_LINUX64): Use FFI_HIDDEN instead of .hidden.
2081         * src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV): Remove,
2082         add FFI_HIDDEN to its prototype.
2083         (ffi_closure_SYSV_inner): New.
2084         * src/x86/sysv.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2085         * src/x86/win32.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2086
2087 2005-08-10  Alfred M. Szmidt  <ams@gnu.org>
2088
2089         PR libffi/21819:
2090         * configure: Rebuilt.
2091         * configure.ac: Handle i*86-*-gnu*.
2092
2093 2005-08-09  Jakub Jelinek  <jakub@redhat.com>
2094
2095         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Use
2096         DW_CFA_offset_extended_sf rather than
2097         DW_CFA_GNU_negative_offset_extended.
2098         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise.
2099
2100 2005-07-22  SUGIOKA Toshinobu  <sugioka@itonet.co.jp>
2101
2102         * src/sh/sysv.S (ffi_call_SYSV): Stop argument popping correctly
2103         on sh3.
2104         (ffi_closure_SYSV): Change the stack layout for sh3 struct argument.
2105         * src/sh/ffi.c (ffi_prep_args): Fix sh3 argument copy, when it is
2106         partially on register.
2107         (ffi_closure_helper_SYSV): Likewise.
2108         (ffi_prep_cif_machdep): Don't set too many cif->flags.
2109
2110 2005-07-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2111
2112         * src/sh/ffi.c (ffi_call): Handle small structures correctly.
2113         Remove empty line.
2114         * src/sh64/ffi.c (simple_type): Remove.
2115         (return_type): Handle small structures correctly.
2116         (ffi_prep_args): Likewise.
2117         (ffi_call): Likewise.
2118         (ffi_closure_helper_SYSV): Likewise.
2119         * src/sh64/sysv.S (ffi_call_SYSV): Handle 1, 2 and 4-byte return.
2120         Emit position independent code if PIC and remove wrong datalabel
2121         prefixes from EH data.
2122
2123 2005-07-19  Andreas Tobler  <a.tobler@schweiz.ch>
2124
2125         * Makefile.am (nodist_libffi_la_SOURCES): Add POWERPC_FREEBSD.
2126         * Makefile.in: Regenerate.
2127         * include/Makefile.in: Likewise.
2128         * testsuite/Makefile.in: Likewise.
2129         * configure.ac: Add POWERPC_FREEBSD rules.
2130         * configure: Regenerate.
2131         * src/powerpc/ffitarget.h: Add POWERPC_FREEBSD rules.
2132         (FFI_SYSV_TYPE_SMALL_STRUCT): Define.
2133         * src/powerpc/ffi.c: Add flags to handle small structure returns
2134         in ffi_call_SYSV.
2135         (ffi_prep_cif_machdep): Handle small structures for SYSV 4 ABI.
2136         Aka FFI_SYSV.
2137         (ffi_closure_helper_SYSV): Likewise.
2138         * src/powerpc/ppc_closure.S: Add return types for small structures.
2139         * src/powerpc/sysv.S: Add bits to handle small structures for
2140         final SYSV 4 ABI.
2141
2142 2005-07-10  Andreas Tobler  <a.tobler@schweiz.ch>
2143
2144         * testsuite/libffi.call/cls_5_1_byte.c: New test file.
2145         * testsuite/libffi.call/cls_6_1_byte.c: Likewise.
2146         * testsuite/libffi.call/cls_7_1_byte.c: Likewise.
2147
2148 2005-07-05  Randolph Chung  <tausq@debian.org>
2149
2150         * src/pa/ffi.c (ffi_struct_type): Rename FFI_TYPE_SMALL_STRUCT1
2151         as FFI_TYPE_SMALL_STRUCT3.  Break out handling for 5-7 byte
2152         structures.  Kill compilation warnings.
2153         (ffi_closure_inner_LINUX): Print return values as hex in debug
2154         message.  Rename FFI_TYPE_SMALL_STRUCT1 as FFI_TYPE_SMALL_STRUCT3.
2155         Properly handle 5-7 byte structure returns.
2156         * src/pa/ffitarget.h (FFI_TYPE_SMALL_STRUCT1)
2157         (FFI_TYPE_SMALL_STRUCT2): Remove.
2158         (FFI_TYPE_SMALL_STRUCT3, FFI_TYPE_SMALL_STRUCT5)
2159         (FFI_TYPE_SMALL_STRUCT6, FFI_TYPE_SMALL_STRUCT7): Define.
2160         * src/pa/linux.S: Mark source file as using PA1.1 assembly.
2161         (checksmst1, checksmst2): Remove.
2162         (checksmst3): Optimize handling of 3-byte struct returns.
2163         (checksmst567): Properly handle 5-7 byte struct returns.
2164
2165 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2166
2167         PR libgcj/21943
2168         * src/mips/n32.S: Enforce PIC code.
2169         * src/mips/o32.S: Likewise.
2170
2171 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2172
2173         * configure.ac: Treat i*86-*-solaris2.10 and up as X86_64.
2174         * configure: Regenerate.
2175
2176 2005-06-01  Alan Modra  <amodra@bigpond.net.au>
2177
2178         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't use JUMPTARGET
2179         to call ffi_closure_helper_SYSV.  Append @local instead.
2180         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise for ffi_prep_args_SYSV.
2181
2182 2005-05-17  Kelley Cook  <kcook@gcc.gnu.org>
2183
2184         * configure.ac: Use AC_C_BIGENDIAN instead of AC_C_BIGENDIAN_CROSS.
2185         Use AC_CHECK_SIZEOF instead of AC_COMPILE_CHECK_SIZEOF.
2186         * Makefile.am (ACLOCAL_AMFLAGS): Remove -I ../config.
2187         * aclocal.m4, configure, fficonfig.h.in, Makefile.in,
2188         include/Makefile.in, testsuite/Makefile.in: Regenerate.
2189
2190 2005-05-09  Mike Stump  <mrs@apple.com>
2191
2192         * configure: Regenerate.
2193
2194 2005-05-08  Richard Henderson  <rth@redhat.com>
2195
2196         PR libffi/21285
2197         * src/alpha/osf.S: Update unwind into to match code.
2198
2199 2005-05-04  Andreas Degert <ad@papyrus-gmbh.de>
2200             Richard Henderson  <rth@redhat.com>
2201
2202         * src/x86/ffi64.c (ffi_prep_cif_machdep): Save sse-used flag in
2203         bit 11 of flags.
2204         (ffi_call): Mask return type field.  Pass ssecount to ffi_call_unix64.
2205         (ffi_prep_closure): Set carry bit if sse-used flag set.
2206         * src/x86/unix64.S (ffi_call_unix64): Add ssecount argument.
2207         Only load sse registers if ssecount non-zero.
2208         (ffi_closure_unix64): Only save sse registers if carry set on entry.
2209
2210 2005-04-29  Ralf Corsepius  <ralf.corsepius@rtems.org>
2211
2212         * configure.ac: Add i*86-*-rtems*, sparc*-*-rtems*,
2213         powerpc-*rtems*, arm*-*-rtems*, sh-*-rtems*.
2214         * configure: Regenerate.
2215
2216 2005-04-20  Hans-Peter Nilsson  <hp@axis.com>
2217
2218         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): In regsub use,
2219         have Tcl8.3-compatible intermediate variable.
2220
2221 2005-04-18  Simon Posnjak <simon.posnjak@siol.net>
2222             Hans-Peter Nilsson  <hp@axis.com>
2223
2224         * Makefile.am: Add CRIS support.
2225         * configure.ac: Likewise.
2226         * Makefile.in, configure, testsuite/Makefile.in,
2227         include/Makefile.in: Regenerate.
2228         * src/cris: New directory.
2229         * src/cris/ffi.c, src/cris/sysv.S, src/cris/ffitarget.h: New files.
2230         * src/prep_cif.c (ffi_prep_cif): Wrap in #ifndef __CRIS__.
2231
2232         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): Replace \n with
2233         \r?\n in output tests.
2234
2235 2005-04-12  Mike Stump  <mrs@apple.com>
2236
2237         * configure: Regenerate.
2238
2239 2005-03-30  Hans Boehm  <Hans.Boehm@hp.com>
2240
2241         * src/ia64/ffitarget.h (ffi_arg): Use long long instead of DI.
2242
2243 2005-03-30  Steve Ellcey  <sje@cup.hp.com>
2244
2245         * src/ia64/ffitarget.h (ffi_arg) ADD DI attribute.
2246         (ffi_sarg) Ditto.
2247         * src/ia64/unix.S (ffi_closure_unix): Extend gp
2248         to 64 bits in ILP32 mode.
2249         Load 64 bits even for short data.
2250
2251 2005-03-23  Mike Stump  <mrs@apple.com>
2252
2253         * src/powerpc/darwin.S: Update for -m64 multilib.
2254         * src/powerpc/darwin_closure.S: Likewise.
2255
2256 2005-03-21  Zack Weinberg  <zack@codesourcery.com>
2257
2258         * configure.ac: Do not invoke TL_AC_GCC_VERSION.
2259         Do not set tool_include_dir.
2260         * aclocal.m4, configure, Makefile.in, testsuite/Makefile.in:
2261         Regenerate.
2262         * include/Makefile.am: Set gcc_version and toollibffidir.
2263         * include/Makefile.in: Regenerate.
2264
2265 2005-02-22  Andrew Haley  <aph@redhat.com>
2266
2267         * src/powerpc/ffi.c (ffi_prep_cif_machdep): Bump alignment to
2268         odd-numbered register pairs for 64-bit integer types.
2269
2270 2005-02-23  Andreas Tobler  <a.tobler@schweiz.ch>
2271
2272         PR libffi/20104
2273         * testsuite/libffi.call/return_ll1.c: New test case.
2274
2275 2005-02-11  Janis Johnson  <janis187@us.ibm.com>
2276
2277         * testsuite/libffi.call/cls_align_longdouble.c: Remove dg-options.
2278         * testsuite/libffi.call/float.c: Ditto.
2279         * testsuite/libffi.call/float2.c: Ditto.
2280         * testsuite/libffi.call/float3.c: Ditto.
2281
2282 2005-02-08  Andreas Tobler  <a.tobler@schweiz.ch>
2283
2284         * src/frv/ffitarget.h: Remove PPC stuff which does not belong to frv.
2285
2286 2005-01-12  Eric Botcazou  <ebotcazou@libertysurf.fr>
2287
2288         * testsuite/libffi.special/special.exp (cxx_options): Add
2289         -shared-libgcc.
2290
2291 2004-12-31  Richard Henderson  <rth@redhat.com>
2292
2293         * src/types.c (FFI_AGGREGATE_TYPEDEF): Remove.
2294         (FFI_TYPEDEF): Rename from FFI_INTEGRAL_TYPEDEF.  Replace size and
2295         offset parameters with a type parameter; deduce size and structure
2296         alignment.  Update all users.
2297
2298 2004-12-31  Richard Henderson  <rth@redhat.com>
2299
2300         * src/types.c (FFI_TYPE_POINTER): Define with sizeof.
2301         (FFI_TYPE_LONGDOUBLE): Fix for ia64.
2302         * src/ia64/ffitarget.h (struct ffi_ia64_trampoline_struct): Move
2303         into ffi_prep_closure.
2304         * src/ia64/ia64_flags.h, src/ia64/ffi.c, src/ia64/unix.S: Rewrite
2305         from scratch.
2306
2307 2004-12-27  Richard Henderson  <rth@redhat.com>
2308
2309         * src/x86/unix64.S: Fix typo in unwind info.
2310
2311 2004-12-25  Richard Henderson  <rth@redhat.com>
2312
2313         * src/x86/ffi64.c (struct register_args): Rename from stackLayout.
2314         (enum x86_64_reg_class): Add X86_64_COMPLEX_X87_CLASS.
2315         (merge_classes): Check for it.
2316         (SSE_CLASS_P): New.
2317         (classify_argument): Pass byte_offset by value; perform all updates
2318         inside struct case.
2319         (examine_argument): Add classes argument; handle
2320         X86_64_COMPLEX_X87_CLASS.
2321         (ffi_prep_args): Merge into ...
2322         (ffi_call): ... here.  Share stack frame with ffi_call_unix64.
2323         (ffi_prep_cif_machdep): Setup cif->flags for proper structure return.
2324         (ffi_fill_return_value): Remove.
2325         (ffi_prep_closure): Remove dead assert.
2326         (ffi_closure_unix64_inner): Rename from ffi_closure_UNIX64_inner.
2327         Rewrite to use struct register_args instead of va_list.  Create
2328         flags for handling structure returns.
2329         * src/x86/unix64.S: Remove dead strings.
2330         (ffi_call_unix64): Rename from ffi_call_UNIX64.  Rewrite to share
2331         stack frame with ffi_call.  Handle structure returns properly.
2332         (float2sse, floatfloat2sse, double2sse): Remove.
2333         (sse2float, sse2double, sse2floatfloat): Remove.
2334         (ffi_closure_unix64): Rename from ffi_closure_UNIX64.  Rewrite
2335         to handle structure returns properly.
2336
2337 2004-12-08  David Edelsohn  <edelsohn@gnu.org>
2338
2339         * Makefile.am (AM_MAKEFLAGS): Remove duplicate LIBCFLAGS and
2340         PICFLAG.
2341         * Makefile.in: Regenerated.
2342
2343 2004-12-02  Richard Sandiford  <rsandifo@redhat.com>
2344
2345         * configure.ac: Use TL_AC_GCC_VERSION to set gcc_version.
2346         * configure, aclocal.m4, Makefile.in: Regenerate.
2347         * include/Makefile.in, testsuite/Makefile.in: Regenerate.
2348
2349 2004-11-29  Kelley Cook  <kcook@gcc.gnu.org>
2350
2351         * configure: Regenerate for libtool change.
2352
2353 2004-11-25  Kelley Cook  <kcook@gcc.gnu.org>
2354
2355         * configure: Regenerate for libtool reversion.
2356
2357 2004-11-24  Kelley Cook  <kcook@gcc.gnu.org>
2358
2359         * configure: Regenerate for libtool change.
2360
2361 2004-11-23  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2362
2363         * testsuite/lib/libffi-dg.exp: Use new procs in target-libpath.exp.
2364
2365 2004-11-23  Richard Sandiford  <rsandifo@redhat.com>
2366
2367         * src/mips/o32.S (ffi_call_O32, ffi_closure_O32): Use jalr instead
2368         of jal.  Use an absolute encoding for the frame information.
2369
2370 2004-11-23  Kelley Cook  <kcook@gcc.gnu.org>
2371
2372         * Makefile.am: Remove no-dependencies.  Add ACLOCAL_AMFLAGS.
2373         * acinclude.m4: Delete logic for sincludes.
2374         * aclocal.m4, Makefile.in, configure: Regenerate.
2375         * include/Makefile: Likewise.
2376         * testsuite/Makefile: Likewise.
2377
2378 2004-11-22  Eric Botcazou  <ebotcazou@libertysurf.fr>
2379
2380         * src/sparc/ffi.c (ffi_prep_closure): Align doubles and 64-bit integers
2381         on a 8-byte boundary.
2382         * src/sparc/v8.S (ffi_closure_v8): Reserve frame space for arguments.
2383
2384 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2385
2386         * src/arm/ffi.c (ffi_prep_cif_machdep): Handle functions that return
2387         long long values.  Round stack allocation to a multiple of 8 bytes
2388         for ATPCS compatibility.
2389         * src/arm/sysv.S (ffi_call_SYSV): Rework to avoid use of APCS register
2390         names.  Handle returning long long types.  Add Thumb and interworking
2391         support.  Improve soft-float code.
2392
2393 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2394
2395         * testsuite/lib/libffi-db.exp (load_gcc_lib): New function.
2396         (libffi_exit): New function.
2397         (libffi_init): Build the testglue wrapper if needed.
2398
2399 2004-10-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2400
2401         PR other/18138
2402         * testsuite/lib/libffi-dg.exp: Accept more than one multilib libgcc.
2403
2404 2004-10-25  Kazuhiro Inaoka <inaoka.kazuhiro@renesas.com>
2405
2406         * src/m32r/libffitarget.h (FFI_CLOSURES): Set to 0.
2407
2408 2004-10-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2409
2410         * src/sh/sysv.S (ffi_call_SYSV): Don't align for double data.
2411         * testsuite/libffi.call/float3.c: New test case.
2412
2413 2004-10-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2414
2415         * src/sh/ffi.c (ffi_prep_closure): Set T bit in trampoline for
2416         the function returning a structure pointed with R2.
2417         * src/sh/sysv.S (ffi_closure_SYSV): Use R2 as the pointer to
2418         the structure return value if T bit set.  Emit position
2419         independent code and EH data if PIC.
2420
2421 2004-10-13  Kazuhiro Inaoka  <inaoka.kazuhiro@renesas.com>
2422
2423         * Makefile.am: Add m32r support.
2424         * configure.ac: Likewise.
2425         * Makefile.in: Regenerate.
2426         * confiugre: Regenerate.
2427         * src/types.c: Add m32r port to FFI_INTERNAL_TYPEDEF
2428         (uint64, sint64, double, longdouble)
2429         * src/m32r: New directory.
2430         * src/m32r/ffi.c: New file.
2431         * src/m32r/sysv.S: Likewise.
2432         * src/m32r/ffitarget.h: Likewise.
2433
2434 2004-10-02  Kaz Kojima  <kkojima@gcc.gnu.org>
2435
2436         * testsuite/libffi.call/negint.c: New test case.
2437
2438 2004-09-14  H.J. Lu  <hongjiu.lu@intel.com>
2439
2440         PR libgcj/17465
2441         * testsuite/lib/libffi-dg.exp: Don't use global ld_library_path.
2442         Set up LD_LIBRARY_PATH, SHLIB_PATH, LD_LIBRARYN32_PATH,
2443         LD_LIBRARY64_PATH, LD_LIBRARY_PATH_32, LD_LIBRARY_PATH_64 and
2444         DYLD_LIBRARY_PATH.
2445
2446 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2447
2448         * testsuite/libffi.call/many_win32.c: Remove whitespaces.
2449         * testsuite/libffi.call/promotion.c: Likewise.
2450         * testsuite/libffi.call/return_ll.c: Remove unused var. Cleanup
2451         whitespaces.
2452         * testsuite/libffi.call/return_sc.c: Likewise.
2453         * testsuite/libffi.call/return_uc.c: Likewise.
2454
2455 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2456
2457         * src/powerpc/darwin.S: Fix comments and identation.
2458         * src/powerpc/darwin_closure.S: Likewise.
2459
2460 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2461
2462         * src/powerpc/ffi_darwin.c: Add flag for longdouble return values.
2463         (ffi_prep_args): Handle longdouble arguments.
2464         (ffi_prep_cif_machdep): Set flags for longdouble. Calculate space for
2465         longdouble.
2466         (ffi_closure_helper_DARWIN): Add closure handling for longdouble.
2467         * src/powerpc/darwin.S (_ffi_call_DARWIN): Add handling of longdouble
2468         values.
2469         * src/powerpc/darwin_closure.S (_ffi_closure_ASM): Likewise.
2470         * src/types.c: Defined longdouble size and alignment for darwin.
2471
2472 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2473
2474         * src/powerpc/aix.S: Remove whitespaces.
2475         * src/powerpc/aix_closure.S: Likewise.
2476         * src/powerpc/asm.h: Likewise.
2477         * src/powerpc/ffi.c: Likewise.
2478         * src/powerpc/ffitarget.h: Likewise.
2479         * src/powerpc/linux64.S: Likewise.
2480         * src/powerpc/linux64_closure.S: Likewise.
2481         * src/powerpc/ppc_closure.S: Likewise.
2482         * src/powerpc/sysv.S: Likewise.
2483
2484 2004-08-30  Anthony Green  <green@redhat.com>
2485
2486         * Makefile.am: Add frv support.
2487         * Makefile.in, testsuite/Makefile.in: Rebuilt.
2488         * configure.ac: Read configure.host.
2489         * configure.in: Read configure.host.
2490         * configure.host: New file.  frv-elf needs libgloss.
2491         * include/ffi.h.in: Force ffi_closure to have a nice big (8)
2492         alignment.  This is needed to frv and shouldn't harm the others.
2493         * include/ffi_common.h (ALIGN_DOWN): New macro.
2494         * src/frv/ffi.c, src/frv/ffitarget.h, src/frv/eabi.S: New files.
2495
2496 2004-08-24  David Daney  <daney@avtrex.com>
2497
2498         * testsuite/libffi.call/closure_fn0.c: Xfail mips64* instead of mips*.
2499         * testsuite/libffi.call/closure_fn1.c: Likewise.
2500         * testsuite/libffi.call/closure_fn2.c  Likewise.
2501         * testsuite/libffi.call/closure_fn3.c: Likewise.
2502         * testsuite/libffi.call/closure_fn4.c: Likewise.
2503         * testsuite/libffi.call/closure_fn5.c: Likewise.
2504         * testsuite/libffi.call/cls_18byte.c: Likewise.
2505         * testsuite/libffi.call/cls_19byte.c: Likewise.
2506         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2507         * testsuite/libffi.call/cls_20byte.c: Likewise.
2508         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2509         * testsuite/libffi.call/cls_24byte.c: Likewise.
2510         * testsuite/libffi.call/cls_2byte.c: Likewise.
2511         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2512         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2513         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2514         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2515         * testsuite/libffi.call/cls_4byte.c: Likewise.
2516         * testsuite/libffi.call/cls_64byte.c: Likewise.
2517         * testsuite/libffi.call/cls_6byte.c: Likewise.
2518         * testsuite/libffi.call/cls_7byte.c: Likewise.
2519         * testsuite/libffi.call/cls_8byte.c: Likewise.
2520         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2521         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2522         * testsuite/libffi.call/cls_align_double.c: Likewise.
2523         * testsuite/libffi.call/cls_align_float.c: Likewise.
2524         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2525         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2526         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2527         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2528         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2529         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2530         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2531         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
2532         * testsuite/libffi.call/cls_double.c: Likewise.
2533         * testsuite/libffi.call/cls_float.c: Likewise.
2534         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2535         * testsuite/libffi.call/cls_multi_sshort.c: Likewise.
2536         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2537         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2538         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2539         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2540         * testsuite/libffi.call/cls_schar.c: Likewise.
2541         * testsuite/libffi.call/cls_sint.c: Likewise.
2542         * testsuite/libffi.call/cls_sshort.c: Likewise.
2543         * testsuite/libffi.call/cls_uchar.c: Likewise.
2544         * testsuite/libffi.call/cls_uint.c: Likewise.
2545         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2546         * testsuite/libffi.call/cls_ushort.c: Likewise.
2547         * testsuite/libffi.call/nested_struct.c: Likewise.
2548         * testsuite/libffi.call/nested_struct1.c: Likewise.
2549         * testsuite/libffi.call/nested_struct2.c: Likewise.
2550         * testsuite/libffi.call/nested_struct3.c: Likewise.
2551         * testsuite/libffi.call/problem1.c: Likewise.
2552         * testsuite/libffi.special/unwindtest.cc: Likewise.
2553         * testsuite/libffi.call/cls_12byte.c: Likewise and set return value
2554         to zero.
2555         * testsuite/libffi.call/cls_16byte.c: Likewise.
2556         * testsuite/libffi.call/cls_5byte.c: Likewise.
2557
2558 2004-08-23  David Daney <daney@avtrex.com>
2559
2560         PR libgcj/13141
2561         * src/mips/ffitarget.h (FFI_O32_SOFT_FLOAT): New ABI.
2562         * src/mips/ffi.c (ffi_prep_args): Fix alignment calculation.
2563         (ffi_prep_cif_machdep): Handle FFI_O32_SOFT_FLOAT floating point
2564         parameters and return types.
2565         (ffi_call): Handle FFI_O32_SOFT_FLOAT ABI.
2566         (ffi_prep_closure): Ditto.
2567         (ffi_closure_mips_inner_O32): Handle FFI_O32_SOFT_FLOAT ABI, fix
2568         alignment calculations.
2569         * src/mips/o32.S (ffi_closure_O32): Don't use floating point
2570         instructions if FFI_O32_SOFT_FLOAT, make stack frame ABI compliant.
2571
2572 2004-08-14  Casey Marshall <csm@gnu.org>
2573
2574         * src/mips/ffi.c (ffi_pref_cif_machdep): set `cif->flags' to
2575         contain `FFI_TYPE_UINT64' as return type for any 64-bit
2576         integer (O32 ABI only).
2577         (ffi_prep_closure): new function.
2578         (ffi_closure_mips_inner_O32): new function.
2579         * src/mips/ffitarget.h: Define `FFI_CLOSURES' and
2580         `FFI_TRAMPOLINE_SIZE' appropriately if the ABI is o32.
2581         * src/mips/o32.S (ffi_call_O32): add labels for .eh_frame. Return
2582         64 bit integers correctly.
2583         (ffi_closure_O32): new function.
2584         Added DWARF-2 unwind info for both functions.
2585
2586 2004-08-10  Andrew Haley  <aph@redhat.com>
2587
2588         * src/x86/ffi64.c (ffi_prep_args ): 8-align all stack arguments.
2589
2590 2004-08-01  Robert Millan  <robertmh@gnu.org>
2591
2592         * configure.ac: Detect knetbsd-gnu and kfreebsd-gnu.
2593         * configure: Regenerate.
2594
2595 2004-07-30  Maciej W. Rozycki  <macro@linux-mips.org>
2596
2597         * acinclude.m4 (AC_FUNC_MMAP_BLACKLIST): Check for <sys/mman.h>
2598         and mmap() explicitly instead of relying on preset autoconf cache
2599         variables.
2600         * aclocal.m4: Regenerate.
2601         * configure: Regenerate.
2602
2603 2004-07-11  Ulrich Weigand  <uweigand@de.ibm.com>
2604
2605         * src/s390/ffi.c (ffi_prep_args): Fix C aliasing violation.
2606         (ffi_check_float_struct): Remove unused prototype.
2607
2608 2004-06-30  Geoffrey Keating  <geoffk@apple.com>
2609
2610         * src/powerpc/ffi_darwin.c (flush_icache): ';' is a comment
2611         character on Darwin, use '\n\t' instead.
2612
2613 2004-06-26  Matthias Klose  <doko@debian.org>
2614
2615         * libtool-version: Fix typo in revision/age.
2616
2617 2004-06-17  Matthias Klose  <doko@debian.org>
2618
2619         * libtool-version: New.
2620         * Makefile.am (libffi_la_LDFLAGS): Use -version-info for soname.
2621         * Makefile.in: Regenerate.
2622
2623 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2624
2625         * Makefile.am: Remove useless multilib rules.
2626         * Makefile.in: Regenerate.
2627         * aclocal.m4: Regenerate with automake 1.8.5.
2628         * configure.ac: Remove useless multilib configury.
2629         * configure: Regenerate.
2630
2631 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2632
2633         * .cvsignore: New file.
2634
2635 2004-06-10  Jakub Jelinek  <jakub@redhat.com>
2636
2637         * src/ia64/unix.S (ffi_call_unix): Insert group barrier break
2638         fp_done.
2639         (ffi_closure_UNIX): Fix f14/f15 adjustment if FLOAT_SZ is ever
2640         changed from 8.
2641
2642 2004-06-06  Sean McNeil  <sean@mcneil.com>
2643
2644         * configure.ac: Add x86_64-*-freebsd* support.
2645         * configure: Regenerate.
2646
2647 2004-04-26  Joe Buck <jbuck@welsh-buck.org>
2648
2649         Bug 15093
2650         * configure.ac: Test for existence of mmap and sys/mman.h before
2651         checking blacklist.  Fix suggested by Jim Wilson.
2652         * configure: Regenerate.
2653
2654 2004-04-26  Matt Austern  <austern@apple.com>
2655
2656         * src/powerpc/darwin.S: Go through a non-lazy pointer for initial
2657         FDE location.
2658         * src/powerpc/darwin_closure.S: Likewise.
2659
2660 2004-04-24  Andreas Tobler  <a.tobler@schweiz.ch>
2661
2662         * testsuite/libffi.call/cls_multi_schar.c (main): Fix initialization
2663         error. Reported by Thomas Heller <theller@python.net>.
2664         * testsuite/libffi.call/cls_multi_sshort.c (main): Likewise.
2665         * testsuite/libffi.call/cls_multi_ushort.c (main): Likewise.
2666
2667 2004-03-20  Matthias Klose  <doko@debian.org>
2668
2669         * src/pa/linux.S: Fix typo.
2670
2671 2004-03-19  Matthias Klose  <doko@debian.org>
2672
2673         * Makefile.am: Update.
2674         * Makefile.in: Regenerate.
2675         * src/pa/ffi.h.in: Remove.
2676         * src/pa/ffitarget.h: New file.
2677
2678 2004-02-10  Randolph Chung  <tausq@debian.org>
2679
2680         * Makefile.am: Add PA support.
2681         * Makefile.in: Regenerate.
2682         * include/Makefile.in: Regenerate.
2683         * configure.ac: Add PA target.
2684         * configure: Regenerate.
2685         * src/pa/ffi.c: New file.
2686         * src/pa/ffi.h.in: Add PA support.
2687         * src/pa/linux.S: New file.
2688         * prep_cif.c: Add PA support.
2689
2690 2004-03-16  Hosaka Yuji  <hos@tamanegi.org>
2691
2692         * src/types.c: Fix alignment size of X86_WIN32 case int64 and
2693         double.
2694         * src/x86/ffi.c (ffi_prep_args): Replace ecif->cif->rtype->type
2695         with ecif->cif->flags.
2696         (ffi_call, ffi_prep_incoming_args_SYSV): Replace cif->rtype->type
2697         with cif->flags.
2698         (ffi_prep_cif_machdep): Add X86_WIN32 struct case.
2699         (ffi_closure_SYSV): Add 1 or 2-bytes struct case for X86_WIN32.
2700         * src/x86/win32.S (retstruct1b, retstruct2b, sc_retstruct1b,
2701         sc_retstruct2b): Add for 1 or 2-bytes struct case.
2702
2703 2004-03-15 Kelley Cook <kcook@gcc.gnu.org>
2704
2705         * configure.in: Rename file to ...
2706         * configure.ac: ... this.
2707         * fficonfig.h.in: Regenerate.
2708         * Makefile.in: Regenerate.
2709         * include/Makefile.in: Regenerate.
2710         * testsuite/Makefile.in: Regenerate.
2711
2712 2004-03-12  Matt Austern  <austern@apple.com>
2713
2714         * src/powerpc/darwin.S: Fix EH information so it corresponds to
2715         changes in EH format resulting from addition of linkonce support.
2716         * src/powerpc/darwin_closure.S: Likewise.
2717
2718 2004-03-11  Andreas Tobler  <a.tobler@schweiz.ch>
2719             Paolo Bonzini  <bonzini@gnu.org>
2720
2721         * Makefile.am (AUTOMAKE_OPTIONS): Set them.
2722         Remove VPATH. Remove rules for object files. Remove multilib support.
2723         (AM_CCASFLAGS): Add.
2724         * configure.in (AC_CONFIG_HEADERS): Relace AM_CONFIG_HEADER.
2725         (AC_PREREQ): Bump version to 2.59.
2726         (AC_INIT): Fill with version info and bug address.
2727         (ORIGINAL_LD_FOR_MULTILIBS): Remove.
2728         (AM_ENABLE_MULTILIB): Use this instead of AC_ARG_ENABLE.
2729         De-precious CC so that the right flags are passed down to multilibs.
2730         (AC_MSG_ERROR): Replace obsolete macro AC_ERROR.
2731         (AC_CONFIG_FILES): Replace obsolete macro AC_LINK_FILES.
2732         (AC_OUTPUT): Reorganize the output with AC_CONFIG_COMMANDS.
2733         * configure: Rebuilt.
2734         * aclocal.m4: Likewise.
2735         * Makefile.in, include/Makefile.in, testsuite/Makefile.in: Likewise.
2736         * fficonfig.h.in: Likewise.
2737
2738 2004-03-11  Andreas Schwab  <schwab@suse.de>
2739
2740         * src/ia64/ffi.c (ffi_prep_incoming_args_UNIX): Get floating point
2741         arguments from fp registers only for the first 8 parameter slots.
2742         Don't convert a float parameter when passed in memory.
2743
2744 2004-03-09  Hans-Peter Nilsson  <hp@axis.com>
2745
2746         * configure: Regenerate for config/accross.m4 correction.
2747
2748 2004-02-25  Matt Kraai  <kraai@alumni.cmu.edu>
2749
2750         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Change
2751         ecif->cif->bytes to bytes.
2752         (ffi_prep_cif_machdep): Add braces around nested if statement.
2753
2754 2004-02-09  Alan Modra  <amodra@bigpond.net.au>
2755
2756         * src/types.c (pointer): POWERPC64 has 8 byte pointers.
2757
2758         * src/powerpc/ffi.c (ffi_prep_args64): Correct long double handling.
2759         (ffi_closure_helper_LINUX64): Fix typo.
2760         * testsuite/libffi.call/cls_align_longdouble.c: Pass -mlong-double-128
2761         for powerpc64-*-*.
2762         * testsuite/libffi.call/float.c: Likewise.
2763         * testsuite/libffi.call/float2.c: Likewise.
2764
2765 2004-02-08  Alan Modra  <amodra@bigpond.net.au>
2766
2767         * src/powerpc/ffi.c (ffi_prep_cif_machdep <FFI_LINUX64>): Correct
2768         long double function return and long double arg handling.
2769         (ffi_closure_helper_LINUX64): Formatting.  Delete unused "ng" var.
2770         Use "end_pfr" instead of "nf".  Correct long double handling.
2771         Localise "temp".
2772         * src/powerpc/linux64.S (ffi_call_LINUX64): Save f2 long double
2773         return value.
2774         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Allocate
2775         space for long double return value.  Adjust stack frame and offsets.
2776         Load f2 long double return.
2777
2778 2004-02-07  Alan Modra  <amodra@bigpond.net.au>
2779
2780         * src/types.c: Use 16 byte long double for POWERPC64.
2781
2782 2004-01-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2783
2784         * src/sparc/ffi.c (ffi_prep_args_v9): Shift the parameter array
2785         when the structure return address is passed in %o0.
2786         (ffi_V9_return_struct): Rename into ffi_v9_layout_struct.
2787         (ffi_v9_layout_struct): Align the field following a nested structure
2788         on a word boundary.  Use memmove instead of memcpy.
2789         (ffi_call): Update call to ffi_V9_return_struct.
2790         (ffi_prep_closure): Define 'ctx' only for V8.
2791         (ffi_closure_sparc_inner): Clone into ffi_closure_sparc_inner_v8
2792         and ffi_closure_sparc_inner_v9.
2793         (ffi_closure_sparc_inner_v8): Return long doubles by reference.
2794         Always skip the structure return address.  For structures and long
2795         doubles, copy the argument directly.
2796         (ffi_closure_sparc_inner_v9): Skip the structure return address only
2797         if required.  Shift the maximum floating-point slot accordingly.  For
2798         big structures, copy the argument directly; otherwise, left-justify the
2799         argument and call ffi_v9_layout_struct to lay out the structure on
2800         the stack.
2801         * src/sparc/v8.S: Undef STACKFRAME before defining it.
2802         (ffi_closure_v8): Pass the structure return address.  Update call to
2803         ffi_closure_sparc_inner_v8.  Short-circuit FFI_TYPE_INT handling.
2804         Skip the 'unimp' insn when returning long doubles and structures.
2805         * src/sparc/v9.S: Undef STACKFRAME before defining it.
2806         (ffi_closure_v9): Increase the frame size by 2 words.  Short-circuit
2807         FFI_TYPE_INT handling.  Load structures both in integers and
2808         floating-point registers on return.
2809         * README: Update status of the SPARC port.
2810
2811 2004-01-24  Andreas Tobler  <a.tobler@schweiz.ch>
2812
2813         * testsuite/libffi.call/pyobjc-tc.c (main): Treat result value
2814         as of type ffi_arg.
2815         * testsuite/libffi.call/struct3.c (main): Fix CHECK.
2816
2817 2004-01-22  Ulrich Weigand  <uweigand@de.ibm.com>
2818
2819         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Treat result
2820         value as of type ffi_arg, not unsigned int.
2821
2822 2004-01-21  Michael Ritzert  <ritzert@t-online.de>
2823
2824         * ffi64.c (ffi_prep_args): Cast the RHS of an assignment instead
2825         of the LHS.
2826
2827 2004-01-12  Andreas Tobler  <a.tobler@schweiz.ch>
2828
2829         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_32 for
2830         Solaris.
2831
2832 2004-01-08  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2833
2834         * testsuite/libffi.call/ffitest.h (allocate_mmap): Cast MAP_FAILED
2835         to void *.
2836
2837 2003-12-10  Richard Henderson  <rth@redhat.com>
2838
2839         * testsuite/libffi.call/cls_align_pointer.c: Cast pointers to
2840         size_t instead of int.
2841
2842 2003-12-04  Hosaka Yuji  <hos@tamanegi.org>
2843
2844         * testsuite/libffi.call/many_win32.c: Include <float.h>.
2845         * testsuite/libffi.call/many_win32.c (main): Replace variable
2846         int i with unsigned long ul.
2847
2848         * testsuite/libffi.call/cls_align_uint64.c: New test case.
2849         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2850         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2851         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2852         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2853         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2854         * testsuite/libffi.call/cls_align_float.c: Likewise.
2855         * testsuite/libffi.call/cls_align_double.c: Likewise.
2856         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2857         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2858
2859 2003-12-02  Hosaka Yuji  <hos@tamanegi.org>
2860
2861         PR other/13221
2862         * src/x86/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV):
2863         Align arguments to 32 bits.
2864
2865 2003-12-01  Andreas Tobler  <a.tobler@schweiz.ch>
2866
2867         PR other/13221
2868         * testsuite/libffi.call/cls_multi_sshort.c: New test case.
2869         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2870         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2871         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2872         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2873         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2874
2875         * testsuite/libffi.special/unwindtest.cc: Cosmetics.
2876
2877 2003-11-26  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
2878
2879         * testsuite/libffi.call/ffitest.h: Include <fcntl.h>.
2880         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2881
2882 2003-11-22  Andreas Tobler  <a.tobler@schweiz.ch>
2883
2884         * Makefile.in: Rebuilt.
2885         * configure: Likewise.
2886         * testsuite/libffi.special/unwindtest.cc: Convert the mmap to
2887         the right type.
2888
2889 2003-11-21  Andreas Jaeger  <aj@suse.de>
2890             Andreas Tobler  <a.tobler@schweiz.ch>
2891
2892         * acinclude.m4: Add AC_FUNC_MMAP_BLACKLIST.
2893         * configure.in: Call AC_FUNC_MMAP_BLACKLIST.
2894         * Makefile.in: Rebuilt.
2895         * aclocal.m4: Likewise.
2896         * configure: Likewise.
2897         * fficonfig.h.in: Likewise.
2898         * testsuite/lib/libffi-dg.exp: Add include dir.
2899         * testsuite/libffi.call/ffitest.h: Add MMAP definitions.
2900         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2901         * testsuite/libffi.call/closure_fn0.c: Use MMAP functionality
2902         for ffi_closure if available.
2903         * testsuite/libffi.call/closure_fn1.c: Likewise.
2904         * testsuite/libffi.call/closure_fn2.c: Likewise.
2905         * testsuite/libffi.call/closure_fn3.c: Likewise.
2906         * testsuite/libffi.call/closure_fn4.c: Likewise.
2907         * testsuite/libffi.call/closure_fn5.c: Likewise.
2908         * testsuite/libffi.call/cls_12byte.c: Likewise.
2909         * testsuite/libffi.call/cls_16byte.c: Likewise.
2910         * testsuite/libffi.call/cls_18byte.c: Likewise.
2911         * testsuite/libffi.call/cls_19byte.c: Likewise.
2912         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2913         * testsuite/libffi.call/cls_20byte.c: Likewise.
2914         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2915         * testsuite/libffi.call/cls_24byte.c: Likewise.
2916         * testsuite/libffi.call/cls_2byte.c: Likewise.
2917         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2918         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2919         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2920         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2921         * testsuite/libffi.call/cls_4byte.c: Likewise.
2922         * testsuite/libffi.call/cls_5byte.c: Likewise.
2923         * testsuite/libffi.call/cls_64byte.c: Likewise.
2924         * testsuite/libffi.call/cls_6byte.c: Likewise.
2925         * testsuite/libffi.call/cls_7byte.c: Likewise.
2926         * testsuite/libffi.call/cls_8byte.c: Likewise.
2927         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2928         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2929         * testsuite/libffi.call/cls_double.c: Likewise.
2930         * testsuite/libffi.call/cls_float.c: Likewise.
2931         * testsuite/libffi.call/cls_schar.c: Likewise.
2932         * testsuite/libffi.call/cls_sint.c: Likewise.
2933         * testsuite/libffi.call/cls_sshort.c: Likewise.
2934         * testsuite/libffi.call/cls_uchar.c: Likewise.
2935         * testsuite/libffi.call/cls_uint.c: Likewise.
2936         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2937         * testsuite/libffi.call/cls_ushort.c: Likewise.
2938         * testsuite/libffi.call/nested_struct.c: Likewise.
2939         * testsuite/libffi.call/nested_struct1.c: Likewise.
2940         * testsuite/libffi.call/nested_struct2.c: Likewise.
2941         * testsuite/libffi.call/nested_struct3.c: Likewise.
2942         * testsuite/libffi.call/problem1.c: Likewise.
2943         * testsuite/libffi.special/unwindtest.cc: Likewise.
2944
2945 2003-11-20  Andreas Tobler  <a.tobler@schweiz.ch>
2946
2947         * testsuite/lib/libffi-dg.exp: Make the -lgcc_s conditional.
2948
2949 2003-11-19  Andreas Tobler  <a.tobler@schweiz.ch>
2950
2951         * testsuite/lib/libffi-dg.exp: Add DYLD_LIBRARY_PATH for darwin.
2952         Add -lgcc_s to additional flags.
2953
2954 2003-11-12  Andreas Tobler  <a.tobler@schweiz.ch>
2955
2956         * configure.in, include/Makefile.am: PR libgcj/11147, install
2957         the ffitarget.h header file in a gcc versioned and target
2958         dependent place.
2959         * configure: Regenerated.
2960         * Makefile.in, include/Makefile.in: Likewise.
2961         * testsuite/Makefile.in: Likewise.
2962
2963 2003-11-09  Andreas Tobler  <a.tobler@schweiz.ch>
2964
2965         * testsuite/libffi.call/closure_fn0.c: Print result and check
2966         with dg-output to make debugging easier.
2967         * testsuite/libffi.call/closure_fn1.c: Likewise.
2968         * testsuite/libffi.call/closure_fn2.c: Likewise.
2969         * testsuite/libffi.call/closure_fn3.c: Likewise.
2970         * testsuite/libffi.call/closure_fn4.c: Likewise.
2971         * testsuite/libffi.call/closure_fn5.c: Likewise.
2972         * testsuite/libffi.call/cls_12byte.c: Likewise.
2973         * testsuite/libffi.call/cls_16byte.c: Likewise.
2974         * testsuite/libffi.call/cls_18byte.c: Likewise.
2975         * testsuite/libffi.call/cls_19byte.c: Likewise.
2976         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2977         * testsuite/libffi.call/cls_20byte.c: Likewise.
2978         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2979         * testsuite/libffi.call/cls_24byte.c: Likewise.
2980         * testsuite/libffi.call/cls_2byte.c: Likewise.
2981         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2982         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2983         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2984         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2985         * testsuite/libffi.call/cls_4byte.c: Likewise.
2986         * testsuite/libffi.call/cls_5byte.c: Likewise.
2987         * testsuite/libffi.call/cls_64byte.c: Likewise.
2988         * testsuite/libffi.call/cls_6byte.c: Likewise.
2989         * testsuite/libffi.call/cls_7byte.c: Likewise.
2990         * testsuite/libffi.call/cls_8byte.c: Likewise.
2991         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2992         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2993         * testsuite/libffi.call/cls_double.c: Likewise.
2994         * testsuite/libffi.call/cls_float.c: Likewise.
2995         * testsuite/libffi.call/cls_schar.c: Likewise.
2996         * testsuite/libffi.call/cls_sint.c: Likewise.
2997         * testsuite/libffi.call/cls_sshort.c: Likewise.
2998         * testsuite/libffi.call/cls_uchar.c: Likewise.
2999         * testsuite/libffi.call/cls_uint.c: Likewise.
3000         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3001         * testsuite/libffi.call/cls_ushort.c: Likewise.
3002         * testsuite/libffi.call/problem1.c: Likewise.
3003
3004         * testsuite/libffi.special/unwindtest.cc: Make ffi_closure
3005         static.
3006
3007 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
3008
3009         * testsuite/libffi.call/cls_9byte2.c: New test case.
3010         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3011         * testsuite/libffi.call/cls_64byte.c: Likewise.
3012         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3013         * testsuite/libffi.call/cls_19byte.c: Likewise.
3014         * testsuite/libffi.call/cls_18byte.c: Likewise.
3015         * testsuite/libffi.call/closure_fn4.c: Likewise.
3016         * testsuite/libffi.call/closure_fn5.c: Likewise.
3017         * testsuite/libffi.call/cls_schar.c: Likewise.
3018         * testsuite/libffi.call/cls_sint.c: Likewise.
3019         * testsuite/libffi.call/cls_sshort.c: Likewise.
3020         * testsuite/libffi.call/nested_struct2.c: Likewise.
3021         * testsuite/libffi.call/nested_struct3.c: Likewise.
3022
3023 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
3024
3025         * testsuite/libffi.call/cls_double.c: Do a check on the result.
3026         * testsuite/libffi.call/cls_uchar.c: Likewise.
3027         * testsuite/libffi.call/cls_uint.c: Likewise.
3028         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3029         * testsuite/libffi.call/cls_ushort.c: Likewise.
3030         * testsuite/libffi.call/return_sc.c: Cleanup whitespaces.
3031
3032 2003-11-06  Andreas Tobler  <a.tobler@schweiz.ch>
3033
3034         * src/prep_cif.c (ffi_prep_cif): Move the validity check after
3035         the initialization.
3036
3037 2003-10-23  Andreas Tobler  <a.tobler@schweiz.ch>
3038
3039         * src/java_raw_api.c (ffi_java_ptrarray_to_raw): Replace
3040         FFI_ASSERT(FALSE) with FFI_ASSERT(0).
3041
3042 2003-10-22  David Daney  <ddaney@avtrex.com>
3043
3044         * src/mips/ffitarget.h: Replace undefined UINT32 and friends with
3045         __attribute__((__mode__(__SI__))) and friends.
3046
3047 2003-10-22  Andreas Schwab  <schwab@suse.de>
3048
3049         * src/ia64/ffi.c: Replace FALSE/TRUE with false/true.
3050
3051 2003-10-21  Andreas Tobler  <a.tobler@schweiz.ch>
3052
3053         * configure.in: AC_LINK_FILES(ffitarget.h).
3054         * configure: Regenerate.
3055         * Makefile.in: Likewise.
3056         * include/Makefile.in: Likewise.
3057         * testsuite/Makefile.in: Likewise.
3058         * fficonfig.h.in: Likewise.
3059
3060 2003-10-21  Paolo Bonzini  <bonzini@gnu.org>
3061             Richard Henderson  <rth@redhat.com>
3062
3063         Avoid that ffi.h includes fficonfig.h.
3064
3065         * Makefile.am (EXTRA_DIST): Include ffitarget.h files
3066         (TARGET_SRC_MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3067         (TARGET_SRC_MIPS_SGI): Removed.
3068         (MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3069         (MIPS_SGI): Removed.
3070         (CLEANFILES): Removed.
3071         (mostlyclean-am, clean-am, mostlyclean-sub, clean-sub): New
3072         targets.
3073         * acconfig.h: Removed.
3074         * configure.in: Compute sizeofs only for double and long double.
3075         Use them to define and subst HAVE_LONG_DOUBLE.  Include comments
3076         into AC_DEFINE instead of using acconfig.h.  Create
3077         include/ffitarget.h instead of include/fficonfig.h.  Rename
3078         MIPS_GCC to MIPS_IRIX, drop MIPS_SGI since we are in gcc's tree.
3079         AC_DEFINE EH_FRAME_FLAGS.
3080         * include/Makefile.am (DISTCLEANFILES): New automake macro.
3081         (hack_DATA): Add ffitarget.h.
3082         * include/ffi.h.in: Remove all system specific definitions.
3083         Declare raw API even if it is not installed, why bother?
3084         Use limits.h instead of SIZEOF_* to define ffi_type_*.  Do
3085         not define EH_FRAME_FLAGS, it is in fficonfig.h now.  Include
3086         ffitarget.h instead of fficonfig.h.  Remove ALIGN macro.
3087         (UINT_ARG, INT_ARG): Removed, use ffi_arg and ffi_sarg instead.
3088         * include/ffi_common.h (bool): Do not define.
3089         (ffi_assert): Accept failed assertion.
3090         (ffi_type_test): Return void and accept file/line.
3091         (FFI_ASSERT): Pass stringized failed assertion.
3092         (FFI_ASSERT_AT): New macro.
3093         (FFI_ASSERT_VALID_TYPE): New macro.
3094         (UINT8, SINT8, UINT16, SINT16, UINT32, SINT32,
3095         UINT64, SINT64): Define here with gcc's __attribute__ macro
3096         instead of in ffi.h
3097         (FLOAT32, ALIGN): Define here instead of in ffi.h
3098         * include/ffi-mips.h: Removed.  Its content moved to
3099         src/mips/ffitarget.h after separating assembly and C sections.
3100         * src/alpha/ffi.c, src/alpha/ffi.c, src/java_raw_api.c
3101         src/prep_cif.c, src/raw_api.c, src/ia64/ffi.c,
3102         src/mips/ffi.c, src/mips/n32.S, src/mips/o32.S,
3103         src/mips/ffitarget.h, src/sparc/ffi.c, src/x86/ffi64.c:
3104         SIZEOF_ARG -> FFI_SIZEOF_ARG.
3105         * src/ia64/ffi.c: Include stdbool.h (provided by GCC 2.95+).
3106         * src/debug.c (ffi_assert): Accept stringized failed assertion.
3107         (ffi_type_test): Rewritten.
3108         * src/prep-cif.c (initialize_aggregate, ffi_prep_cif): Call
3109         FFI_ASSERT_VALID_TYPE.
3110         * src/alpha/ffitarget.h, src/arm/ffitarget.h,
3111         src/ia64/ffitarget.h, src/m68k/ffitarget.h,
3112         src/mips/ffitarget.h, src/powerpc/ffitarget.h,
3113         src/s390/ffitarget.h, src/sh/ffitarget.h,
3114         src/sh64/ffitarget.h, src/sparc/ffitarget.h,
3115         src/x86/ffitarget.h: New files.
3116         * src/alpha/osf.S, src/arm/sysv.S, src/ia64/unix.S,
3117         src/m68k/sysv.S, src/mips/n32.S, src/mips/o32.S,
3118         src/powerpc/aix.S, src/powerpc/darwin.S,
3119         src/powerpc/ffi_darwin.c, src/powerpc/linux64.S,
3120         src/powerpc/linux64_closure.S, src/powerpc/ppc_closure.S,
3121         src/powerpc/sysv.S, src/s390/sysv.S, src/sh/sysv.S,
3122         src/sh64/sysv.S, src/sparc/v8.S, src/sparc/v9.S,
3123         src/x86/sysv.S, src/x86/unix64.S, src/x86/win32.S:
3124         include fficonfig.h
3125
3126 2003-10-20  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3127
3128         * src/mips/ffi.c: Use _ABIN32, _ABIO32 instead of external
3129         _MIPS_SIM_NABI32, _MIPS_SIM_ABI32.
3130
3131 2003-10-19  Andreas Tobler  <a.tobler@schweiz.ch>
3132
3133         * src/powerpc/ffi_darwin.c (ffi_prep_args): Declare bytes again.
3134         Used when FFI_DEBUG = 1.
3135
3136 2003-10-14  Alan Modra  <amodra@bigpond.net.au>
3137
3138         * src/types.c (double, longdouble): Default POWERPC64 to 8 byte size
3139         and align.
3140
3141 2003-10-06  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3142
3143         * include/ffi_mips.h: Define FFI_MIPS_N32 for N32/N64 ABIs,
3144         FFI_MIPS_O32 for O32 ABI.
3145
3146 2003-10-01  Andreas Tobler  <a.tobler@schweiz.ch>
3147
3148         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_64 for
3149         SPARC64. Cleanup whitespaces.
3150
3151 2003-09-19  Andreas Tobler  <a.tobler@schweiz.ch>
3152
3153         * testsuite/libffi.call/closure_fn0.c: Xfail mips, arm,
3154         strongarm, xscale. Cleanup whitespaces.
3155         * testsuite/libffi.call/closure_fn1.c: Likewise.
3156         * testsuite/libffi.call/closure_fn2.c: Likewise.
3157         * testsuite/libffi.call/closure_fn3.c: Likewise.
3158         * testsuite/libffi.call/cls_12byte.c: Likewise.
3159         * testsuite/libffi.call/cls_16byte.c: Likewise.
3160         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3161         * testsuite/libffi.call/cls_20byte.c: Likewise.
3162         * testsuite/libffi.call/cls_24byte.c: Likewise.
3163         * testsuite/libffi.call/cls_2byte.c: Likewise.
3164         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3165         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3166         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3167         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3168         * testsuite/libffi.call/cls_4byte.c: Likewise.
3169         * testsuite/libffi.call/cls_5byte.c: Likewise.
3170         * testsuite/libffi.call/cls_6byte.c: Likewise.
3171         * testsuite/libffi.call/cls_7byte.c: Likewise.
3172         * testsuite/libffi.call/cls_8byte.c: Likewise.
3173         * testsuite/libffi.call/cls_double.c: Likewise.
3174         * testsuite/libffi.call/cls_float.c: Likewise.
3175         * testsuite/libffi.call/cls_uchar.c: Likewise.
3176         * testsuite/libffi.call/cls_uint.c: Likewise.
3177         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3178         * testsuite/libffi.call/cls_ushort.c: Likewise.
3179         * testsuite/libffi.call/nested_struct.c: Likewise.
3180         * testsuite/libffi.call/nested_struct1.c: Likewise.
3181         * testsuite/libffi.call/problem1.c: Likewise.
3182         * testsuite/libffi.special/unwindtest.cc: Likewise.
3183         * testsuite/libffi.call/pyobjc-tc.c: Cleanup whitespaces.
3184
3185 2003-09-18  David Edelsohn  <edelsohn@gnu.org>
3186
3187         * src/powerpc/aix.S: Cleanup whitespaces.
3188         * src/powerpc/aix_closure.S: Likewise.
3189
3190 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3191
3192         * src/powerpc/darwin.S: Cleanup whitespaces, comment formatting.
3193         * src/powerpc/darwin_closure.S: Likewise.
3194         * src/powerpc/ffi_darwin.c: Likewise.
3195
3196 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3197             David Edelsohn  <edelsohn@gnu.org>
3198
3199         * src/types.c (double): Add AIX and Darwin to the right TYPEDEF.
3200         * src/powerpc/aix_closure.S: Remove the pointer to the outgoing
3201         parameter stack.
3202         * src/powerpc/darwin_closure.S: Likewise.
3203         * src/powerpc/ffi_darwin.c (ffi_prep_args): Handle structures
3204         according to the Darwin/AIX ABI.
3205         (ffi_prep_cif_machdep): Likewise.
3206         (ffi_closure_helper_DARWIN): Likewise.
3207         Remove the outgoing parameter stack logic. Simplify the evaluation
3208         of the different CASE types.
3209         (ffi_prep_clousure): Avoid the casts on lvalues. Change the branch
3210         statement in the trampoline code.
3211
3212 2003-09-18  Kaz Kojima  <kkojima@gcc.gnu.org>
3213
3214         * src/sh/ffi.c (ffi_prep_args): Take account into the alignement
3215         for the register size.
3216         (ffi_closure_helper_SYSV): Handle the structure return value
3217         address correctly.
3218         (ffi_closure_helper_SYSV): Return the appropriate type when
3219         the registers are used for the structure return value.
3220         * src/sh/sysv.S (ffi_closure_SYSV): Fix the stack layout for
3221         the 64-bit return value.  Update copyright years.
3222
3223 2003-09-17  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3224
3225         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Search in
3226         srcdir for ffi_mips.h.
3227
3228 2003-09-12  Alan Modra  <amodra@bigpond.net.au>
3229
3230         * src/prep_cif.c (initialize_aggregate): Include tail padding in
3231         structure size.
3232         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Correct
3233         placement of float result.
3234         * testsuite/libffi.special/unwindtest.cc (closure_test_fn1): Correct
3235         cast of "resp" for big-endian 64 bit machines.
3236
3237 2003-09-11  Alan Modra  <amodra@bigpond.net.au>
3238
3239         * src/types.c (double, longdouble): Merge identical SH and ARM
3240         typedefs, and add POWERPC64.
3241         * src/powerpc/ffi.c (ffi_prep_args64): Correct next_arg calc for
3242         struct split over gpr and rest.
3243         (ffi_prep_cif_machdep): Correct intarg_count for structures.
3244         * src/powerpc/linux64.S (ffi_call_LINUX64): Fix gpr offsets.
3245
3246 2003-09-09  Andreas Tobler  <a.tobler@schweiz.ch>
3247
3248         * src/powerpc/ffi.c (ffi_closure_helper_SYSV) Handle struct
3249         passing correctly.
3250
3251 2003-09-09  Alan Modra  <amodra@bigpond.net.au>
3252
3253         * configure: Regenerate.
3254
3255 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3256
3257         * Makefile.am: Remove build rules for ffitest.
3258         * Makefile.in: Rebuilt.
3259
3260 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3261
3262         * src/java_raw_api.c: Include <stdlib.h> to fix compiler warning
3263         about implicit declaration of abort().
3264
3265 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3266
3267         * Makefile.am: Add dejagnu test framework. Fixes PR other/11411.
3268         * Makefile.in: Rebuilt.
3269         * configure.in: Add dejagnu test framework.
3270         * configure: Rebuilt.
3271
3272         * testsuite/Makefile.am: New file.
3273         * testsuite/Makefile.in: Built
3274         * testsuite/lib/libffi-dg.exp: New file.
3275         * testsuite/config/default.exp: Likewise.
3276         * testsuite/libffi.call/call.exp: Likewise.
3277         * testsuite/libffi.call/ffitest.h: Likewise.
3278         * testsuite/libffi.call/closure_fn0.c: Likewise.
3279         * testsuite/libffi.call/closure_fn1.c: Likewise.
3280         * testsuite/libffi.call/closure_fn2.c: Likewise.
3281         * testsuite/libffi.call/closure_fn3.c: Likewise.
3282         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3283         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3284         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3285         * testsuite/libffi.call/cls_2byte.c: Likewise.
3286         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3287         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3288         * testsuite/libffi.call/cls_4byte.c: Likewise.
3289         * testsuite/libffi.call/cls_5byte.c: Likewise.
3290         * testsuite/libffi.call/cls_6byte.c: Likewise.
3291         * testsuite/libffi.call/cls_7byte.c: Likewise.
3292         * testsuite/libffi.call/cls_8byte.c: Likewise.
3293         * testsuite/libffi.call/cls_12byte.c: Likewise.
3294         * testsuite/libffi.call/cls_16byte.c: Likewise.
3295         * testsuite/libffi.call/cls_20byte.c: Likewise.
3296         * testsuite/libffi.call/cls_24byte.c: Likewise.
3297         * testsuite/libffi.call/cls_double.c: Likewise.
3298         * testsuite/libffi.call/cls_float.c: Likewise.
3299         * testsuite/libffi.call/cls_uchar.c: Likewise.
3300         * testsuite/libffi.call/cls_uint.c: Likewise.
3301         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3302         * testsuite/libffi.call/cls_ushort.c: Likewise.
3303         * testsuite/libffi.call/float.c: Likewise.
3304         * testsuite/libffi.call/float1.c: Likewise.
3305         * testsuite/libffi.call/float2.c: Likewise.
3306         * testsuite/libffi.call/many.c: Likewise.
3307         * testsuite/libffi.call/many_win32.c: Likewise.
3308         * testsuite/libffi.call/nested_struct.c: Likewise.
3309         * testsuite/libffi.call/nested_struct1.c: Likewise.
3310         * testsuite/libffi.call/pyobjc-tc.c: Likewise.
3311         * testsuite/libffi.call/problem1.c: Likewise.
3312         * testsuite/libffi.call/promotion.c: Likewise.
3313         * testsuite/libffi.call/return_ll.c: Likewise.
3314         * testsuite/libffi.call/return_sc.c: Likewise.
3315         * testsuite/libffi.call/return_uc.c: Likewise.
3316         * testsuite/libffi.call/strlen.c: Likewise.
3317         * testsuite/libffi.call/strlen_win32.c: Likewise.
3318         * testsuite/libffi.call/struct1.c: Likewise.
3319         * testsuite/libffi.call/struct2.c: Likewise.
3320         * testsuite/libffi.call/struct3.c: Likewise.
3321         * testsuite/libffi.call/struct4.c: Likewise.
3322         * testsuite/libffi.call/struct5.c: Likewise.
3323         * testsuite/libffi.call/struct6.c: Likewise.
3324         * testsuite/libffi.call/struct7.c: Likewise.
3325         * testsuite/libffi.call/struct8.c: Likewise.
3326         * testsuite/libffi.call/struct9.c: Likewise.
3327         * testsuite/libffi.special/special.exp: New file.
3328         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3329         * testsuite/libffi.special/unwindtest.cc: Likewise.
3330
3331
3332 2003-08-13  Kaz Kojima  <kkojima@gcc.gnu.org>
3333
3334         * src/sh/ffi.c (OFS_INT16): Set 0 for little endian case.  Update
3335         copyright years.
3336
3337 2003-08-02  Alan Modra  <amodra@bigpond.net.au>
3338
3339         * src/powerpc/ffi.c (ffi_prep_args64): Modify for changed gcc
3340         structure passing.
3341         (ffi_closure_helper_LINUX64): Likewise.
3342         * src/powerpc/linux64.S: Remove code writing to parm save area.
3343         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Use return
3344         address in lr from ffi_closure_helper_LINUX64 call to calculate
3345         table address.  Optimize function tail.
3346
3347 2003-07-28  Andreas Tobler  <a.tobler@schweiz.ch>
3348
3349         * src/sparc/ffi.c: Handle all floating point registers.
3350         * src/sparc/v9.S: Likewise. Fixes second part of PR target/11410.
3351
3352 2003-07-11  Gerald Pfeifer  <pfeifer@dbai.tuwien.ac.at>
3353
3354         * README: Note that libffi is not part of GCC.  Update the project
3355         URL and status.
3356
3357 2003-06-19  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3358
3359         * src/powerpc/ppc_closure.S: Include ffi.h.
3360
3361 2003-06-13  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3362
3363         * src/x86/sysv.S: Avoid gas-only .uleb128/.sleb128 directives.
3364         Use C style comments.
3365
3366 2003-06-13  Kaz Kojima  <kkojima@rr.iij4u.or.jp>
3367
3368         * Makefile.am: Add SHmedia support.  Fix a typo of SH support.
3369         * Makefile.in: Regenerate.
3370         * configure.in (sh64-*-linux*, sh5*-*-linux*): Add target.
3371         * configure: Regenerate.
3372         * include/ffi.h.in: Add SHmedia support.
3373         * src/sh64/ffi.c: New file.
3374         * src/sh64/sysv.S: New file.
3375
3376 2003-05-16  Jakub Jelinek  <jakub@redhat.com>
3377
3378         * configure.in (HAVE_RO_EH_FRAME): Check whether .eh_frame section
3379         should be read-only.
3380         * configure: Rebuilt.
3381         * fficonfig.h.in: Rebuilt.
3382         * include/ffi.h.in (EH_FRAME_FLAGS): Define.
3383         * src/alpha/osf.S: Use EH_FRAME_FLAGS.
3384         * src/powerpc/linux64.S: Likewise.
3385         * src/powerpc/linux64_closure.S: Likewise.  Include ffi.h.
3386         * src/powerpc/sysv.S: Use EH_FRAME_FLAGS.  Use pcrel encoding
3387         if -fpic/-fPIC/-mrelocatable.
3388         * src/powerpc/powerpc_closure.S: Likewise.
3389         * src/sparc/v8.S: If HAVE_RO_EH_FRAME is defined, don't include
3390         #write in .eh_frame flags.
3391         * src/sparc/v9.S: Likewise.
3392         * src/x86/unix64.S: Use EH_FRAME_FLAGS.
3393         * src/x86/sysv.S: Likewise.  Use pcrel encoding if -fpic/-fPIC.
3394         * src/s390/sysv.S: Use EH_FRAME_FLAGS.  Include ffi.h.
3395
3396 2003-05-07  Jeff Sturm  <jsturm@one-point.com>
3397
3398         Fixes PR bootstrap/10656
3399         * configure.in (HAVE_AS_REGISTER_PSEUDO_OP): Test assembler
3400         support for .register pseudo-op.
3401         * src/sparc/v8.S: Use it.
3402         * fficonfig.h.in: Rebuilt.
3403         * configure: Rebuilt.
3404
3405 2003-04-18  Jakub Jelinek  <jakub@redhat.com>
3406
3407         * include/ffi.h.in (POWERPC64): Define if 64-bit.
3408         (enum ffi_abi): Add FFI_LINUX64 on POWERPC.
3409         Make it the default on POWERPC64.
3410         (FFI_TRAMPOLINE_SIZE): Define to 24 on POWERPC64.
3411         * configure.in: Change powerpc-*-linux* into powerpc*-*-linux*.
3412         * configure: Rebuilt.
3413         * src/powerpc/ffi.c (hidden): Define.
3414         (ffi_prep_args_SYSV): Renamed from
3415         ffi_prep_args.  Cast pointers to unsigned long to shut up warnings.
3416         (NUM_GPR_ARG_REGISTERS64, NUM_FPR_ARG_REGISTERS64,
3417         ASM_NEEDS_REGISTERS64): New.
3418         (ffi_prep_args64): New function.
3419         (ffi_prep_cif_machdep): Handle FFI_LINUX64 ABI.
3420         (ffi_call): Likewise.
3421         (ffi_prep_closure): Likewise.
3422         (flush_icache): Surround by #ifndef POWERPC64.
3423         (ffi_dblfl): New union type.
3424         (ffi_closure_helper_SYSV): Use it to avoid aliasing problems.
3425         (ffi_closure_helper_LINUX64): New function.
3426         * src/powerpc/ppc_closure.S: Surround whole file by #ifndef
3427         __powerpc64__.
3428         * src/powerpc/sysv.S: Likewise.
3429         (ffi_call_SYSV): Rename ffi_prep_args to ffi_prep_args_SYSV.
3430         * src/powerpc/linux64.S: New file.
3431         * src/powerpc/linux64_closure.S: New file.
3432         * Makefile.am (EXTRA_DIST): Add src/powerpc/linux64.S and
3433         src/powerpc/linux64_closure.S.
3434         (TARGET_SRC_POWERPC): Likewise.
3435
3436         * src/ffitest.c (closure_test_fn, closure_test_fn1, closure_test_fn2,
3437         closure_test_fn3): Fix result printing on big-endian 64-bit