OSDN Git Service

* gnat.dg/socket1.adb: Disable on *-*-solaris2*.
[pf3gnuchains/gcc-fork.git] / libffi / ChangeLog
1 2010-03-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
2
3         * configure.ac (libffi_cv_as_x86_64_unwind_section_type): New test.
4         * configure: Regenerate.
5         * fficonfig.h.in: Regenerate.
6         * libffi/src/x86/unix64.S (.eh_frame)
7         [HAVE_AS_X86_64_UNWIND_SECTION_TYPE]: Use @unwind section type.
8
9 2010-02-24  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
10
11         * doc/libffi.texi (The Closure API): Fix typo.
12         * doc/libffi.info: Remove.
13
14 2010-02-15  Matthias Klose  <doko@ubuntu.com>
15
16         * src/arm/sysv.S (__ARM_ARCH__): Define for processor
17         __ARM_ARCH_7EM__.
18
19 2010-01-07  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
20
21         PR libffi/40701
22         * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRIdLL,
23         PRIuLL, PRId64, PRIu64, PRIuPTR): Define.
24         * testsuite/libffi.call/cls_align_sint64.c: Add -Wno-format on
25         alpha*-dec-osf*.
26         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
27         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
28         * testsuite/libffi.call/return_ll1.c: Likewise.
29         * testsuite/libffi.call/stret_medium2.c: Likewise.
30         * testsuite/libffi.special/ffitestcxx.h (allocate_mmap): Cast
31         MAP_FAILED to char *.
32
33 2010-01-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
34
35         * src/mips/n32.S: Use .abicalls and .eh_frame with __GNUC__.
36
37 2009-12-31  Anthony Green  <green@redhat.com>
38
39         * README: Update for libffi 3.0.9.
40
41 2009-12-27  Matthias Klose  <doko@ubuntu.com>
42
43         * configure.ac (HAVE_LONG_DOUBLE): Define for mips when
44         appropriate.
45         * configure: Rebuilt.
46
47 2009-12-26  Anthony Green  <green@redhat.com>
48
49         * testsuite/libffi.call/cls_longdouble_va.c: Mark as xfail for
50         avr32*-*-*.
51         * testsuite/libffi.call/cls_double_va.c: Ditto.
52
53 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
54
55         * testsuite/libffi.call/ffitest.h: Conditionally include stdint.h
56         and inttypes.h.
57         * testsuite/libffi.special/unwindtest.cc: Ditto.
58
59 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
60
61         * configure.ac: Add amd64-*-openbsd*.
62         * configure: Rebuilt.
63         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Link
64         openbsd programs with -lpthread.
65
66 2009-12-26  Anthony Green  <green@redhat.com>
67
68         * testsuite/libffi.call/cls_double_va.c,
69         testsuite/libffi.call/cls_longdouble.c,
70         testsuite/libffi.call/cls_longdouble_va.c,
71         testsuite/libffi.call/cls_pointer.c,
72         testsuite/libffi.call/cls_pointer_stack.c: Remove xfail for
73         mips*-*-* and arm*-*-*.
74         * testsuite/libffi.call/cls_align_longdouble_split.c,
75         testsuite/libffi.call/cls_align_longdouble_split2.c,
76         testsuite/libffi.call/stret_medium2.c,
77         testsuite/libffi.call/stret_medium.c,
78         testsuite/libffi.call/stret_large.c,
79         testsuite/libffi.call/stret_large2.c: Remove xfail for arm*-*-*.
80
81 2009-12-31  Kay Tietz  <ktietz70@googlemail.com>
82
83         * testsuite/libffi.call/ffitest.h,
84         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRuLL): Fix
85         definitions.
86
87 2009-12-31  Carlo Bramini  <carlo.bramix@libero.it>
88
89         * configure.ac (AM_LTLDFLAGS): Define for windows hosts.
90         * Makefile.am (libffi_la_LDFLAGS): Add AM_LTLDFLAGS.
91         * configure: Rebuilt.
92         * Makefile.in: Rebuilt.
93
94 2009-12-31  Anthony Green  <green@redhat.com>
95             Blake Chaffin.
96
97         * testsuite/libffi.call/huge_struct.c: New test case from Blake
98         Chaffin @ Apple.
99
100 2009-12-28  David Edelsohn  <edelsohn@gnu.org>
101
102         * src/powerpc/ffi_darwin.c (ffi_prep_args): Copy abi and nargs to
103         local variables.
104         (aix_adjust_aggregate_sizes): New function.
105         (ffi_prep_cif_machdep): Call it.
106
107 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
108
109         * configure.ac: Define FFI_MMAP_EXEC_WRIT for the given targets.
110         * configure: Regenerate.
111         * fficonfig.h.in: Likewise.
112         * src/closures.c: Remove the FFI_MMAP_EXEC_WRIT definition for
113         Solaris/x86.
114
115 2009-12-26  Andreas Schwab  <schwab@linux-m68k.org>
116
117         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Advance intarg_count
118         when a float arguments is passed in memory.
119         (ffi_closure_helper_SYSV): Mark general registers as used up when
120         a 64bit or soft-float long double argument is passed in memory.
121
122 2009-12-25  Matthias Klose  <doko@ubuntu.com>
123
124         * man/ffi_call.3: Fix #include in examples.
125         * doc/libffi.texi: Add dircategory.
126
127 2009-12-25  Frank Everdij <f.p.x.everdij@tudelft.nl>
128
129         * include/ffi.h.in: Placed '__GNUC__' ifdef around
130         '__attribute__((aligned(8)))' in ffi_closure, fixes compile for
131         IRIX MIPSPro c99.
132         * include/ffi_common.h: Added '__sgi' define to non
133         '__attribute__((__mode__()))' integer typedefs.
134         * src/mips/ffi.c (ffi_call, ffi_closure_mips_inner_O32,
135         ffi_closure_mips_inner_N32): Added 'defined(_MIPSEB)' to BE check.
136         (ffi_closure_mips_inner_O32, ffi_closure_mips_inner_N32): Added
137         FFI_LONGDOUBLE support and alignment(N32 only).
138         * src/mips/ffitarget.h: Corrected '#include <sgidefs.h>' for IRIX and
139         fixed non '__attribute__((__mode__()))' integer typedefs.
140         * src/mips/n32.S: Put '#ifdef linux' around '.abicalls' and '.eh_frame'
141         since they are Linux/GNU Assembler specific.
142
143 2009-12-25  Bradley Smith  <brad@brad-smith.co.uk>
144
145         * configure.ac, Makefile.am, src/avr32/ffi.c,
146         src/avr32/ffitarget.h,
147         src/avr32/sysv.S: Add AVR32 port.
148         * configure, Makefile.in: Rebuilt.
149
150 2009-12-21  Andreas Tobler  <a.tobler@schweiz.org>
151
152         * configure.ac: Make i?86 build on FreeBSD and OpenBSD.
153         * configure: Regenerate.
154
155 2009-12-15  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
156
157         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on PA HP-UX.
158
159 2009-12-13  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
160
161         * src/pa/ffi.c (ffi_closure_inner_pa32): Handle FFI_TYPE_LONGDOUBLE
162         type on HP-UX.
163
164 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
165
166         * src/sparc/ffi.c (ffi_closure_sparc_inner_v9): Properly align 'long
167         double' arguments.
168
169 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
170
171         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on Solaris < 10.
172
173 2009-12-10  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
174
175         PR libffi/40700
176         * src/closures.c [X86_64 && __sun__ && __svr4__]
177         (FFI_MMAP_EXEC_WRIT): Define.
178
179 2009-12-08  David Daney  <ddaney@caviumnetworks.com>
180
181         * testsuite/libffi.call/stret_medium.c: Remove xfail for mips*-*-*
182         * testsuite/libffi.call/cls_align_longdouble_split2.c: Same.
183         * testsuite/libffi.call/stret_large.c: Same.
184         * testsuite/libffi.call/cls_align_longdouble_split.c: Same.
185         * testsuite/libffi.call/stret_large2.c: Same.
186         * testsuite/libffi.call/stret_medium2.c: Same.
187
188 2009-12-07  David Edelsohn  <edelsohn@gnu.org>
189
190         * src/powerpc/aix_closure.S (libffi_closure_ASM): Fix tablejump
191         typo.
192
193 2009-12-05  David Edelsohn  <edelsohn@gnu.org>
194
195         * src/powerpc/aix.S: Update AIX32 code to be consistent with AIX64
196         code.
197         * src/powerpc/aix_closure.S: Same.
198
199 2009-12-05  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
200
201         * Makefile.in: Regenerate.
202         * configure: Regenerate.
203         * include/Makefile.in: Regenerate.
204         * man/Makefile.in: Regenerate.
205         * testsuite/Makefile.in: Regenerate.
206
207 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
208
209         * src/powerpc/aix_closure.S: Reorganize 64-bit code to match
210         linux64_closure.S.
211
212 2009-12-04  Uros Bizjak  <ubizjak@gmail.com>
213
214         PR libffi/41908
215         * src/x86/ffi64.c (classify_argument): Update from
216         gcc/config/i386/i386.c.
217         (ffi_closure_unix64_inner): Do not use the address of two consecutive
218         SSE registers directly.
219         * testsuite/libffi.call/cls_dbls_struct.c (main): Remove xfail
220         for x86_64 linux targets.
221
222 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
223
224         * src/powerpc/ffi_darwin.c (ffi_closure_helper_DARWIN): Increment
225         pfr for long double split between fpr13 and stack.
226
227 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
228
229         * src/powerpc/ffi_darwin.c (ffi_prep_args): Increment next_arg and
230         fparg_count twice for long double.
231
232 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
233
234         PR libffi/42243
235         * src/powerpc/ffi_darwin.c (ffi_prep_args): Remove extra parentheses.
236
237 2009-12-03  Uros Bizjak  <ubizjak@gmail.com>
238
239         * testsuite/libffi.call/cls_longdouble_va.c (main): Fix format string.
240         Remove xfails for x86 linux targets.
241
242 2009-12-02  David Edelsohn  <edelsohn@gnu.org>
243
244         * src/powerpc/ffi_darwin.c (ffi_prep_args): Fix typo in INT64
245         case.
246
247 2009-12-01  David Edelsohn  <edelsohn@gnu.org>
248
249         * src/powerpc/aix.S (ffi_call_AIX): Convert to more standard
250         register usage.  Call ffi_prep_args directly.  Add long double
251         return value support.
252         * src/powerpc/ffi_darwin.c (ffi_prep_args): Double arg increment
253         applies to FFI_TYPE_DOUBLE.  Correct fpr_base increment typo.
254         Separate FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases.
255         (ffi_prep_cif_machdep): Only 16 byte stack alignment in 64 bit
256         mode.
257         (ffi_closure_helper_DARWIN): Remove nf and ng counters.  Move temp
258         into case.
259         * src/powerpc/aix_closure.S: Maintain 16 byte stack alignment.
260         Allocate result area between params and FPRs.
261
262 2009-11-30  David Edelsohn  <edelsohn@gnu.org>
263
264         PR target/35484
265         * src/powerpc/ffitarget.h (POWERPC64): Define for PPC64 Linux and
266         AIX64.
267         * src/powerpc/aix.S: Implement AIX64 version.
268         * src/powerpc/aix_closure.S: Implement AIX64 version.
269         (ffi_closure_ASM): Use extsb, lha and displament addresses.
270         * src/powerpc/ffi_darwin.c (ffi_prep_args): Implement AIX64
271         support.
272         (ffi_prep_cif_machdep): Same.
273         (ffi_call): Same.
274         (ffi_closure_helper_DARWIN): Same.
275
276 2009-11-02  Andreas Tobler  <a.tobler@schweiz.org>
277
278         PR libffi/41908
279         * testsuite/libffi.call/testclosure.c: New test.
280
281 2009-09-28  Kai Tietz  <kai.tietz@onevision.com>
282
283         * src/x86/win64.S (_ffi_call_win64 stack): Remove for gnu
284         assembly version use of ___chkstk.
285
286 2009-09-23  Matthias Klose  <doko@ubuntu.com>
287
288         PR libffi/40242, PR libffi/41443
289         * src/arm/sysv.S (__ARM_ARCH__): Define for processors
290         __ARM_ARCH_6T2__, __ARM_ARCH_6M__, __ARM_ARCH_7__,
291         __ARM_ARCH_7A__, __ARM_ARCH_7R__, __ARM_ARCH_7M__.
292         Change the conditionals to __SOFTFP__ || __ARM_EABI__
293         for -mfloat-abi=softfp to work.
294
295 2009-09-17  Loren J. Rittle  <ljrittle@acm.org>
296
297         PR testsuite/32843 (strikes again)
298         * src/x86/ffi.c (ffi_prep_cif_machdep): Add X86_FREEBSD to
299         enable proper extension on char and short.
300
301 2009-09-15  David Daney  <ddaney@caviumnetworks.com>
302
303         * src/java_raw_api.c (ffi_java_raw_to_rvalue): Remove special
304         handling for FFI_TYPE_POINTER.
305         * src/mips/ffitarget.h (FFI_TYPE_STRUCT_D_SOFT,
306         FFI_TYPE_STRUCT_F_SOFT, FFI_TYPE_STRUCT_DD_SOFT,
307         FFI_TYPE_STRUCT_FF_SOFT, FFI_TYPE_STRUCT_FD_SOFT,
308         FFI_TYPE_STRUCT_DF_SOFT, FFI_TYPE_STRUCT_SOFT): New defines.
309         (FFI_N32_SOFT_FLOAT, FFI_N64_SOFT_FLOAT): New ffi_abi enumerations.
310         (enum ffi_abi): Set FFI_DEFAULT_ABI for soft-float.
311         * src/mips/n32.S (ffi_call_N32): Add handling for soft-float
312         structure and pointer returns.
313         (ffi_closure_N32): Add handling for pointer returns.
314         * src/mips/ffi.c (ffi_prep_args, calc_n32_struct_flags,
315         calc_n32_return_struct_flags): Handle soft-float.
316         (ffi_prep_cif_machdep):  Handle soft-float, fix pointer handling.
317         (ffi_call_N32): Declare proper argument types.
318         (ffi_call, copy_struct_N32, ffi_closure_mips_inner_N32): Handle
319         soft-float.
320
321 2009-08-24  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
322
323         * configure.ac (AC_PREREQ): Bump to 2.64.
324
325 2009-08-22  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
326
327         * Makefile.am (install-html, install-pdf): Remove.
328         * Makefile.in: Regenerate.
329
330         * Makefile.in: Regenerate.
331         * aclocal.m4: Regenerate.
332         * configure: Regenerate.
333         * fficonfig.h.in: Regenerate.
334         * include/Makefile.in: Regenerate.
335         * man/Makefile.in: Regenerate.
336         * testsuite/Makefile.in: Regenerate.
337
338 2009-07-30  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
339
340         * configure.ac (_AC_ARG_VAR_PRECIOUS): Use m4_rename_force.
341
342 2009-07-24  Dave Korn  <dave.korn.cygwin@gmail.com>
343
344         PR libffi/40807
345         * src/x86/ffi.c (ffi_prep_cif_machdep): Also use sign/zero-extending
346         return types for X86_WIN32.
347         * src/x86/win32.S (_ffi_call_SYSV): Handle omitted return types.
348         (_ffi_call_STDCALL, _ffi_closure_SYSV, _ffi_closure_raw_SYSV,
349         _ffi_closure_STDCALL): Likewise.
350
351         * src/closures.c (is_selinux_enabled): Define to const 0 for Cygwin.
352         (dlmmap, dlmunmap): Also use these functions on Cygwin.
353
354 2009-07-11  Richard Sandiford  <rdsandiford@googlemail.com>
355
356         PR testsuite/40699
357         PR testsuite/40707
358         PR testsuite/40709
359         * testsuite/lib/libffi-dg.exp: Revert 2009-07-02, 2009-07-01 and
360         2009-06-30 commits.
361
362 2009-07-01  Richard Sandiford  <r.sandiford@uk.ibm.com>
363
364         * testsuite/lib/libffi-dg.exp (libffi-init): Set ld_library_path
365         to "" before adding paths.  (This reinstates an assignment that
366         was removed by my 2009-06-30 commit, but changes the initial
367         value from "." to "".)
368
369 2009-07-01  H.J. Lu  <hongjiu.lu@intel.com>
370
371         PR testsuite/40601
372         * testsuite/lib/libffi-dg.exp (libffi-init): Properly set
373         gccdir.  Adjust ld_library_path for gcc only if gccdir isn't
374         empty.
375
376 2009-06-30  Richard Sandiford  <r.sandiford@uk.ibm.com>
377
378         * testsuite/lib/libffi-dg.exp (libffi-init): Don't add "."
379         to ld_library_path.  Use add_path.  Add just find_libgcc_s
380         to ld_library_path, not every libgcc multilib directory.
381
382 2009-06-16  Wim Lewis  <wiml@hhhh.org>
383
384         * src/powerpc/ffi.c: Avoid clobbering cr3 and cr4, which are
385         supposed to be callee-saved.
386         * src/powerpc/sysv.S (small_struct_return_value): Fix overrun of
387         return buffer for odd-size structs.
388
389 2009-06-16  Andreas Tobler  <a.tobler@schweiz.org>
390
391         PR libffi/40444
392         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Add
393         allow_stack_execute for Darwin.
394
395 2009-06-16  Andrew Haley  <aph@redhat.com>
396
397         * configure.ac (TARGETDIR): Add missing blank lines.
398         * configure: Regenerate.
399
400 2009-06-16  Andrew Haley  <aph@redhat.com>
401
402         * testsuite/libffi.call/cls_align_sint64.c,
403         testsuite/libffi.call/cls_align_uint64.c,
404         testsuite/libffi.call/cls_longdouble_va.c,
405         testsuite/libffi.call/cls_ulonglong.c,
406         testsuite/libffi.call/return_ll1.c,
407         testsuite/libffi.call/stret_medium2.c: Fix printf format
408         specifiers.
409         * testsuite/libffi.call/ffitest.h,
410         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRIuLL): Define.
411
412 2009-06-15  Andrew Haley  <aph@redhat.com>
413
414         * testsuite/libffi.call/err_bad_typedef.c: xfail everywhere.
415         * testsuite/libffi.call/err_bad_abi.c: Likewise.
416
417 2009-06-12  Andrew Haley  <aph@redhat.com>
418
419         * Makefile.am: Remove info_TEXINFOS.
420
421 2009-06-12  Andrew Haley  <aph@redhat.com>
422
423         * ChangeLog.libffi: testsuite/libffi.call/cls_align_sint64.c,
424         testsuite/libffi.call/cls_align_uint64.c,
425         testsuite/libffi.call/cls_ulonglong.c,
426         testsuite/libffi.call/return_ll1.c,
427         testsuite/libffi.call/stret_medium2.c: Fix printf format
428         specifiers.
429         testsuite/libffi.special/unwindtest.cc: include stdint.h.
430
431 2009-06-11  Timothy Wall  <twall@users.sf.net>
432
433         * Makefile.am,
434         configure.ac,
435         include/ffi.h.in,
436         include/ffi_common.h,
437         src/closures.c,
438         src/dlmalloc.c,
439         src/x86/ffi.c,
440         src/x86/ffitarget.h,
441         src/x86/win64.S (new),
442         README: Added win64 support (mingw or MSVC)
443         * Makefile.in,
444         include/Makefile.in,
445         man/Makefile.in,
446         testsuite/Makefile.in,
447         configure,
448         aclocal.m4: Regenerated
449         * ltcf-c.sh: properly escape cygwin/w32 path
450         * man/ffi_call.3: Clarify size requirements for return value.
451         * src/x86/ffi64.c: Fix filename in comment.
452         * src/x86/win32.S: Remove unused extern.
453
454         * testsuite/libffi.call/closure_fn0.c,
455         testsuite/libffi.call/closure_fn1.c,
456         testsuite/libffi.call/closure_fn2.c,
457         testsuite/libffi.call/closure_fn3.c,
458         testsuite/libffi.call/closure_fn4.c,
459         testsuite/libffi.call/closure_fn5.c,
460         testsuite/libffi.call/closure_fn6.c,
461         testsuite/libffi.call/closure_stdcall.c,
462         testsuite/libffi.call/cls_12byte.c,
463         testsuite/libffi.call/cls_16byte.c,
464         testsuite/libffi.call/cls_18byte.c,
465         testsuite/libffi.call/cls_19byte.c,
466         testsuite/libffi.call/cls_1_1byte.c,
467         testsuite/libffi.call/cls_20byte.c,
468         testsuite/libffi.call/cls_20byte1.c,
469         testsuite/libffi.call/cls_24byte.c,
470         testsuite/libffi.call/cls_2byte.c,
471         testsuite/libffi.call/cls_3_1byte.c,
472         testsuite/libffi.call/cls_3byte1.c,
473         testsuite/libffi.call/cls_3byte2.c,
474         testsuite/libffi.call/cls_4_1byte.c,
475         testsuite/libffi.call/cls_4byte.c,
476         testsuite/libffi.call/cls_5_1_byte.c,
477         testsuite/libffi.call/cls_5byte.c,
478         testsuite/libffi.call/cls_64byte.c,
479         testsuite/libffi.call/cls_6_1_byte.c,
480         testsuite/libffi.call/cls_6byte.c,
481         testsuite/libffi.call/cls_7_1_byte.c,
482         testsuite/libffi.call/cls_7byte.c,
483         testsuite/libffi.call/cls_8byte.c,
484         testsuite/libffi.call/cls_9byte1.c,
485         testsuite/libffi.call/cls_9byte2.c,
486         testsuite/libffi.call/cls_align_double.c,
487         testsuite/libffi.call/cls_align_float.c,
488         testsuite/libffi.call/cls_align_longdouble.c,
489         testsuite/libffi.call/cls_align_longdouble_split.c,
490         testsuite/libffi.call/cls_align_longdouble_split2.c,
491         testsuite/libffi.call/cls_align_pointer.c,
492         testsuite/libffi.call/cls_align_sint16.c,
493         testsuite/libffi.call/cls_align_sint32.c,
494         testsuite/libffi.call/cls_align_sint64.c,
495         testsuite/libffi.call/cls_align_uint16.c,
496         testsuite/libffi.call/cls_align_uint32.c,
497         testsuite/libffi.call/cls_align_uint64.c,
498         testsuite/libffi.call/cls_dbls_struct.c,
499         testsuite/libffi.call/cls_double.c,
500         testsuite/libffi.call/cls_double_va.c,
501         testsuite/libffi.call/cls_float.c,
502         testsuite/libffi.call/cls_longdouble.c,
503         testsuite/libffi.call/cls_longdouble_va.c,
504         testsuite/libffi.call/cls_multi_schar.c,
505         testsuite/libffi.call/cls_multi_sshort.c,
506         testsuite/libffi.call/cls_multi_sshortchar.c,
507         testsuite/libffi.call/cls_multi_uchar.c,
508         testsuite/libffi.call/cls_multi_ushort.c,
509         testsuite/libffi.call/cls_multi_ushortchar.c,
510         testsuite/libffi.call/cls_pointer.c,
511         testsuite/libffi.call/cls_pointer_stack.c,
512         testsuite/libffi.call/cls_schar.c,
513         testsuite/libffi.call/cls_sint.c,
514         testsuite/libffi.call/cls_sshort.c,
515         testsuite/libffi.call/cls_uchar.c,
516         testsuite/libffi.call/cls_uint.c,
517         testsuite/libffi.call/cls_ulonglong.c,
518         testsuite/libffi.call/cls_ushort.c,
519         testsuite/libffi.call/err_bad_abi.c,
520         testsuite/libffi.call/err_bad_typedef.c,
521         testsuite/libffi.call/float2.c,
522         testsuite/libffi.call/huge_struct.c,
523         testsuite/libffi.call/nested_struct.c,
524         testsuite/libffi.call/nested_struct1.c,
525         testsuite/libffi.call/nested_struct10.c,
526         testsuite/libffi.call/nested_struct2.c,
527         testsuite/libffi.call/nested_struct3.c,
528         testsuite/libffi.call/nested_struct4.c,
529         testsuite/libffi.call/nested_struct5.c,
530         testsuite/libffi.call/nested_struct6.c,
531         testsuite/libffi.call/nested_struct7.c,
532         testsuite/libffi.call/nested_struct8.c,
533         testsuite/libffi.call/nested_struct9.c,
534         testsuite/libffi.call/problem1.c,
535         testsuite/libffi.call/return_ldl.c,
536         testsuite/libffi.call/return_ll1.c,
537         testsuite/libffi.call/stret_large.c,
538         testsuite/libffi.call/stret_large2.c,
539         testsuite/libffi.call/stret_medium.c,
540         testsuite/libffi.call/stret_medium2.c,
541         testsuite/libffi.special/unwindtest.cc: use ffi_closure_alloc instead
542         of checking for MMAP.  Use intptr_t instead of long casts.
543
544 2009-06-11  Kaz Kojima  <kkojima@gcc.gnu.org>
545
546         * testsuite/libffi.call/cls_longdouble_va.c: Add xfail sh*-*-linux-*.
547         * testsuite/libffi.call/err_bad_abi.c: Add xfail sh*-*-*.
548         * testsuite/libffi.call/err_bad_typedef.c: Likewise.
549
550 2009-06-09  Andrew Haley  <aph@redhat.com>
551
552         * src/x86/freebsd.S: Add missing file.
553
554 2009-06-08  Andrew Haley  <aph@redhat.com>
555
556         Import from libffi 3.0.8:
557
558         * doc/libffi.texi: New file.
559         * doc/libffi.info: Likewise.
560         * doc/stamp-vti: Likewise.
561         * man/Makefile.am: New file.
562         * man/ffi_call.3: New file.
563
564         * Makefile.am (EXTRA_DIST): Add src/x86/darwin64.S,
565         src/dlmalloc.c.
566         (nodist_libffi_la_SOURCES): Add X86_FREEBSD.
567
568         * configure.ac: Bump version to 3.0.8.
569         parisc*-*-linux*: Add.
570         i386-*-freebsd* | i386-*-openbsd*: Add.
571         powerpc-*-beos*: Add.
572         AM_CONDITIONAL X86_FREEBSD: Add.
573         AC_CONFIG_FILES: Add man/Makefile.
574
575         * include/ffi.h.in (FFI_FN): Change void (*)() to void (*)(void).
576
577 2009-06-08  Andrew Haley  <aph@redhat.com>
578
579         * README: Import from libffi 3.0.8.
580
581 2009-06-08  Andrew Haley  <aph@redhat.com>
582
583         * testsuite/libffi.call/err_bad_abi.c: Add xfails.
584         * testsuite/libffi.call/cls_longdouble_va.c: Add xfails.
585         * testsuite/libffi.call/cls_dbls_struct.c: Add xfail x86_64-*-linux-*.
586         * testsuite/libffi.call/err_bad_typedef.c: Add xfails.
587
588         * testsuite/libffi.call/stret_medium2.c: Add __UNUSED__ to args.
589         * testsuite/libffi.call/stret_medium.c: Likewise.
590         * testsuite/libffi.call/stret_large2.c: Likewise.
591         * testsuite/libffi.call/stret_large.c:  Likewise.
592
593 2008-12-26  Timothy Wall  <twall@users.sf.net>
594
595         * testsuite/libffi.call/cls_longdouble.c,
596         testsuite/libffi.call/cls_longdouble_va.c,
597         testsuite/libffi.call/cls_align_longdouble.c,
598         testsuite/libffi.call/cls_align_longdouble_split.c,
599         testsuite/libffi.call/cls_align_longdouble_split2.c: mark expected
600         failures on x86_64 cygwin/mingw.
601
602 2008-12-22  Timothy Wall  <twall@users.sf.net>
603
604         * testsuite/libffi.call/closure_fn0.c,
605         testsuite/libffi.call/closure_fn1.c,
606         testsuite/libffi.call/closure_fn2.c,
607         testsuite/libffi.call/closure_fn3.c,
608         testsuite/libffi.call/closure_fn4.c,
609         testsuite/libffi.call/closure_fn5.c,
610         testsuite/libffi.call/closure_fn6.c,
611         testsuite/libffi.call/closure_loc_fn0.c,
612         testsuite/libffi.call/closure_stdcall.c,
613         testsuite/libffi.call/cls_align_pointer.c,
614         testsuite/libffi.call/cls_pointer.c,
615         testsuite/libffi.call/cls_pointer_stack.c: use portable cast from
616         pointer to integer (intptr_t).
617         * testsuite/libffi.call/cls_longdouble.c: disable for win64.
618
619 2008-07-24  Anthony Green  <green@redhat.com>
620
621         * testsuite/libffi.call/cls_dbls_struct.c,
622         testsuite/libffi.call/cls_double_va.c,
623         testsuite/libffi.call/cls_longdouble.c,
624         testsuite/libffi.call/cls_longdouble_va.c,
625         testsuite/libffi.call/cls_pointer.c,
626         testsuite/libffi.call/cls_pointer_stack.c,
627         testsuite/libffi.call/err_bad_abi.c: Clean up failures from
628         compiler warnings.
629
630 2008-03-04  Anthony Green  <green@redhat.com>
631             Blake Chaffin
632             hos@tamanegi.org
633
634         * testsuite/libffi.call/cls_align_longdouble_split2.c
635           testsuite/libffi.call/cls_align_longdouble_split.c
636           testsuite/libffi.call/cls_dbls_struct.c
637           testsuite/libffi.call/cls_double_va.c
638           testsuite/libffi.call/cls_longdouble.c
639           testsuite/libffi.call/cls_longdouble_va.c
640           testsuite/libffi.call/cls_pointer.c
641           testsuite/libffi.call/cls_pointer_stack.c
642           testsuite/libffi.call/err_bad_abi.c
643           testsuite/libffi.call/err_bad_typedef.c
644           testsuite/libffi.call/stret_large2.c
645           testsuite/libffi.call/stret_large.c
646           testsuite/libffi.call/stret_medium2.c
647           testsuite/libffi.call/stret_medium.c: New tests from Apple.
648
649 2009-06-05  Andrew Haley  <aph@redhat.com>
650
651         * src/x86/ffitarget.h, src/x86/ffi.c: Merge stdcall changes from
652         libffi.
653
654 2009-06-04  Andrew Haley  <aph@redhat.com>
655
656         * src/x86/ffitarget.h, src/x86/win32.S, src/x86/ffi.c: Back out
657         stdcall changes.
658
659 2008-02-26  Anthony Green  <green@redhat.com>
660             Thomas Heller  <theller@ctypes.org>
661
662         * src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C
663         comment.
664
665 2008-02-03  Timothy Wall  <twall@users.sf.net>
666
667         * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return
668           offset based on code pointer, not data pointer.
669
670 2008-01-31  Timothy Wall <twall@users.sf.net>
671
672         * testsuite/libffi.call/closure_stdcall.c: Add test for stdcall
673         closures.
674         * src/x86/ffitarget.h: Increase size of trampoline for stdcall
675         closures.
676         * src/x86/win32.S: Add assembly for stdcall closure.
677         * src/x86/ffi.c: Initialize stdcall closure trampoline.
678
679 2009-06-04  Andrew Haley  <aph@redhat.com>
680
681         * include/ffi.h.in: Change void (*)() to void (*)(void).
682         * src/x86/ffi.c: Likewise.
683
684 2009-06-04  Andrew Haley  <aph@redhat.com>
685
686         * src/powerpc/ppc_closure.S: Insert licence header.
687         * src/powerpc/linux64_closure.S: Likewise.
688         * src/m68k/sysv.S: Likewise.
689
690         * src/sh64/ffi.c: Change void (*)() to void (*)(void).
691         * src/powerpc/ffi.c: Likewise.
692         * src/powerpc/ffi_darwin.c: Likewise.
693         * src/m32r/ffi.c: Likewise.
694         * src/sh64/ffi.c: Likewise.
695         * src/x86/ffi64.c: Likewise.
696         * src/alpha/ffi.c: Likewise.
697         * src/alpha/osf.S: Likewise.
698         * src/frv/ffi.c: Likewise.
699         * src/s390/ffi.c: Likewise.
700         * src/pa/ffi.c: Likewise.
701         * src/pa/hpux32.S: Likewise.
702         * src/ia64/unix.S: Likewise.
703         * src/ia64/ffi.c: Likewise.
704         * src/sparc/ffi.c: Likewise.
705         * src/mips/ffi.c: Likewise.
706         * src/sh/ffi.c: Likewise.
707
708 2008-02-15  David Daney  <ddaney@avtrex.com>
709
710         * src/mips/ffi.c (USE__BUILTIN___CLEAR_CACHE):
711         Define (conditionally), and use it to include cachectl.h.
712         (ffi_prep_closure_loc): Fix cache flushing.
713         * src/mips/ffitarget.h (_ABIN32, _ABI64, _ABIO32): Define.
714
715 2009-06-04  Andrew Haley  <aph@redhat.com>
716
717         include/ffi.h.in,
718         src/arm/ffitarget.h,
719         src/arm/ffi.c,
720         src/arm/sysv.S,
721         src/powerpc/ffitarget.h,
722         src/closures.c,
723         src/sh64/ffitarget.h,
724         src/sh64/ffi.c,
725         src/sh64/sysv.S,
726         src/types.c,
727         src/x86/ffi64.c,
728         src/x86/ffitarget.h,
729         src/x86/win32.S,
730         src/x86/darwin.S,
731         src/x86/ffi.c,
732         src/x86/sysv.S,
733         src/x86/unix64.S,
734         src/alpha/ffitarget.h,
735         src/alpha/ffi.c,
736         src/alpha/osf.S,
737         src/m68k/ffitarget.h,
738         src/frv/ffitarget.h,
739         src/frv/ffi.c,
740         src/s390/ffitarget.h,
741         src/s390/sysv.S,
742         src/cris/ffitarget.h,
743         src/pa/linux.S,
744         src/pa/ffitarget.h,
745         src/pa/ffi.c,
746         src/raw_api.c,
747         src/ia64/ffitarget.h,
748         src/ia64/unix.S,
749         src/ia64/ffi.c,
750         src/ia64/ia64_flags.h,
751         src/java_raw_api.c,
752         src/debug.c,
753         src/sparc/v9.S,
754         src/sparc/ffitarget.h,
755         src/sparc/ffi.c,
756         src/sparc/v8.S,
757         src/mips/ffitarget.h,
758         src/mips/n32.S,
759         src/mips/o32.S,
760         src/mips/ffi.c,
761         src/prep_cif.c,
762         src/sh/ffitarget.h,
763         src/sh/ffi.c,
764         src/sh/sysv.S: Update license text.
765
766 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
767
768         * src/x86/win32.S (_ffi_closure_STDCALL):  New function.
769         (.eh_frame):  Add FDE for it.
770
771 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
772
773         * configure.ac:  Also check if assembler supports pc-relative
774         relocs on X86_WIN32 targets.
775         * configure:  Regenerate.
776         * src/x86/win32.S (ffi_prep_args):  Declare extern, not global.
777         (_ffi_call_SYSV):  Add missing function type symbol .def and
778         add EH markup labels.
779         (_ffi_call_STDCALL):  Likewise.
780         (_ffi_closure_SYSV):  Likewise.
781         (_ffi_closure_raw_SYSV):  Likewise.
782         (.eh_frame):  Add hand-crafted EH data.
783
784 2009-04-09  Jakub Jelinek  <jakub@redhat.com>
785
786         * testsuite/lib/libffi-dg.exp: Change copyright header to refer to
787         version 3 of the GNU General Public License and to point readers
788         at the COPYING3 file and the FSF's license web page.
789         * testsuite/libffi.call/call.exp: Likewise.
790         * testsuite/libffi.special/special.exp: Likewise.
791
792 2009-03-01  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
793
794         * configure: Regenerate.
795
796 2008-12-18  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
797
798         PR libffi/26048
799         * configure.ac (HAVE_AS_X86_PCREL): New test.
800         * configure: Regenerate.
801         * fficonfig.h.in: Regenerate.
802         * src/x86/sysv.S [!FFI_NO_RAW_API]: Precalculate
803         RAW_CLOSURE_CIF_OFFSET, RAW_CLOSURE_FUN_OFFSET,
804         RAW_CLOSURE_USER_DATA_OFFSET for the Solaris 10/x86 assembler.
805         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
806         * src/x86/unix64.S (.Lstore_table): Move to .text section.
807         (.Lload_table): Likewise.
808         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
809
810 2008-12-18  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
811
812         * configure: Regenerate.
813
814 2008-11-21  Eric Botcazou  <ebotcazou@adacore.com>
815
816         * src/sparc/ffi.c (ffi_prep_cif_machdep): Add support for
817         signed/unsigned int8/16 return values.
818         * src/sparc/v8.S (ffi_call_v8): Likewise.
819         (ffi_closure_v8): Likewise.
820
821 2008-09-26  Peter O'Gorman  <pogma@thewrittenword.com>
822             Steve Ellcey  <sje@cup.hp.com>
823
824         * configure: Regenerate for new libtool.
825         * Makefile.in: Ditto.
826         * include/Makefile.in: Ditto.
827         * aclocal.m4: Ditto.
828
829 2008-08-25  Andreas Tobler  <a.tobler@schweiz.org>
830
831         * src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and
832         FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum.
833         Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT.
834         Adjust copyright notice.
835         * src/powerpc/ffi.c: Add two new flags to indicate if we have one
836         register or two register to use for FFI_SYSV structs.
837         (ffi_prep_cif_machdep): Pass the right register flag introduced above.
838         (ffi_closure_helper_SYSV): Fix the return type for
839         FFI_SYSV_TYPE_SMALL_STRUCT. Comment.
840         Adjust copyright notice.
841
842 2008-07-16  Kaz Kojima  <kkojima@gcc.gnu.org>
843
844         * src/sh/ffi.c (ffi_prep_closure_loc): Turn INSN into an unsigned
845         int.
846
847 2008-06-17  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
848
849         * configure: Regenerate.
850         * include/Makefile.in: Regenerate.
851         * testsuite/Makefile.in: Regenerate.
852
853 2008-06-07  Joseph Myers  <joseph@codesourcery.com>
854
855         * configure.ac (parisc*-*-linux*, powerpc-*-sysv*,
856         powerpc-*-beos*): Remove.
857         * configure: Regenerate.
858
859 2008-05-09  Julian Brown  <julian@codesourcery.com>
860
861         * Makefile.am (LTLDFLAGS): New.
862         (libffi_la_LDFLAGS): Use above.
863         * Makefile.in: Regenerate.
864
865 2008-04-18  Paolo Bonzini  <bonzini@gnu.org>
866
867         PR bootstrap/35457
868         * aclocal.m4: Regenerate.
869         * configure: Regenerate.
870
871 2008-03-26  Kaz Kojima  <kkojima@gcc.gnu.org>
872
873         * src/sh/sysv.S: Add .note.GNU-stack on Linux.
874         * src/sh64/sysv.S: Likewise.
875
876 2008-03-26  Daniel Jacobowitz  <dan@debian.org>
877
878         * src/arm/sysv.S: Fix ARM comment marker.
879
880 2008-03-26  Jakub Jelinek  <jakub@redhat.com>
881
882         * src/alpha/osf.S: Add .note.GNU-stack on Linux.
883         * src/s390/sysv.S: Likewise.
884         * src/powerpc/ppc_closure.S: Likewise.
885         * src/powerpc/sysv.S: Likewise.
886         * src/x86/unix64.S: Likewise.
887         * src/x86/sysv.S: Likewise.
888         * src/sparc/v8.S: Likewise.
889         * src/sparc/v9.S: Likewise.
890         * src/m68k/sysv.S: Likewise.
891         * src/arm/sysv.S: Likewise.
892
893 2008-03-16  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
894
895         * aclocal.m4: Regenerate.
896         * configure: Likewise.
897         * Makefile.in: Likewise.
898         * include/Makefile.in: Likewise.
899         * testsuite/Makefile.in: Likewise.
900
901 2008-02-12  Bjoern Koenig  <bkoenig@alpha-tierchen.de>
902             Andreas Tobler  <a.tobler@schweiz.org>
903
904         * configure.ac: Add amd64-*-freebsd* target.
905         * configure: Regenerate.
906
907 2008-01-30  H.J. Lu  <hongjiu.lu@intel.com>
908
909         PR libffi/34612
910         * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when
911         returning struct.
912
913         * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer"
914         tests.
915
916 2008-01-24  David Edelsohn  <edelsohn@gnu.org>
917
918         * configure: Regenerate.
919
920 2008-01-06  Andreas Tobler  <a.tobler@schweiz.org>
921
922         * src/x86/ffi.c (ffi_prep_cif_machdep): Fix thinko.
923
924 2008-01-05  Andreas Tobler  <a.tobler@schweiz.org>
925
926         PR testsuite/32843
927         * src/x86/ffi.c (ffi_prep_cif_machdep): Add code for
928         signed/unsigned int8/16 for X86_DARWIN.
929         Updated copyright info.
930         Handle one and two byte structs with special cif->flags.
931         * src/x86/ffitarget.h: Add special types for one and two byte structs.
932         Updated copyright info.
933         * src/x86/darwin.S (ffi_call_SYSV): Rewrite to use a jump table like
934         sysv.S
935         Remove code to pop args from the stack after call.
936         Special-case signed/unsigned for int8/16, one and two byte structs.
937         (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
938         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
939         FFI_TYPE_SINT32.
940         Updated copyright info.
941
942 2007-12-08  David Daney  <ddaney@avtrex.com>
943
944         * src/mips/n32.S (ffi_call_N32):  Replace dadd with ADDU, dsub with
945         SUBU, add with ADDU and use smaller code sequences.
946
947 2007-12-07  David Daney  <ddaney@avtrex.com>
948
949         * src/mips/ffi.c (ffi_prep_cif_machdep): Handle long double return
950         type.
951
952 2007-12-06  David Daney  <ddaney@avtrex.com>
953
954         * include/ffi.h.in (FFI_SIZEOF_JAVA_RAW): Define if not already
955         defined.
956         (ffi_java_raw): New typedef.
957         (ffi_java_raw_call, ffi_java_ptrarray_to_raw,
958         ffi_java_raw_to_ptrarray): Change parameter types from ffi_raw to
959         ffi_java_raw.
960         (ffi_java_raw_closure) : Same.
961         (ffi_prep_java_raw_closure, ffi_prep_java_raw_closure_loc): Change
962         parameter types.
963         * src/java_raw_api.c (ffi_java_raw_size):  Replace FFI_SIZEOF_ARG with
964         FFI_SIZEOF_JAVA_RAW.
965         (ffi_java_raw_to_ptrarray): Change type of raw to ffi_java_raw.
966         Replace FFI_SIZEOF_ARG with FFI_SIZEOF_JAVA_RAW. Use
967         sizeof(ffi_java_raw) for alignment calculations.
968         (ffi_java_ptrarray_to_raw): Same.
969         (ffi_java_rvalue_to_raw): Add special handling for FFI_TYPE_POINTER
970         if FFI_SIZEOF_JAVA_RAW == 4.
971         (ffi_java_raw_to_rvalue): Same.
972         (ffi_java_raw_call): Change type of raw to ffi_java_raw.
973         (ffi_java_translate_args): Same.
974         (ffi_prep_java_raw_closure_loc, ffi_prep_java_raw_closure): Change
975         parameter types.
976         * src/mips/ffitarget.h (FFI_SIZEOF_JAVA_RAW): Define for N32 ABI.
977
978 2007-12-06  David Daney  <ddaney@avtrex.com>
979
980         * src/mips/n32.S (ffi_closure_N32): Use 64-bit add instruction on
981         pointer values.
982
983 2007-12-01  Andreas Tobler  <a.tobler@schweiz.org>
984
985         PR libffi/31937
986         * src/powerpc/ffitarget.h: Introduce new ABI FFI_LINUX_SOFT_FLOAT.
987         Add local FFI_TYPE_UINT128 to handle soft-float long-double-128.
988         * src/powerpc/ffi.c: Distinguish between __NO_FPRS__ and not and
989         set the NUM_FPR_ARG_REGISTERS according to.
990         Add support for potential soft-float support under hard-float
991         architecture.
992         (ffi_prep_args_SYSV): Set NUM_FPR_ARG_REGISTERS to 0 in case of
993         FFI_LINUX_SOFT_FLOAT, handle float, doubles and long-doubles according
994         to the FFI_LINUX_SOFT_FLOAT ABI.
995         (ffi_prep_cif_machdep): Likewise.
996         (ffi_closure_helper_SYSV): Likewise.
997         * src/powerpc/ppc_closure.S: Make sure not to store float/double
998         on archs where __NO_FPRS__ is true.
999         Add FFI_TYPE_UINT128 support.
1000         * src/powerpc/sysv.S: Add support for soft-float long-double-128.
1001         Adjust copyright notice.
1002
1003 2007-11-25  Andreas Tobler  <a.tobler@schweiz.org>
1004
1005         * src/closures.c: Move defintion of MAYBE_UNUSED from here to ...
1006         * include/ffi_common.h: ... here.
1007         Update copyright.
1008
1009 2007-11-17  Andreas Tobler  <a.tobler@schweiz.org>
1010
1011         * src/powerpc/sysv.S: Load correct cr to compare if we have long double.
1012         * src/powerpc/linux64.S: Likewise.
1013         * src/powerpc/ffi.c: Add a comment to show which part goes into cr6.
1014         * testsuite/libffi.call/return_ldl.c: New test.
1015
1016 2007-09-04    <aph@redhat.com>
1017
1018         * src/arm/sysv.S (UNWIND): New.
1019         (Whole file): Conditionally compile unwinder directives.
1020         * src/arm/sysv.S: Add unwinder directives.
1021
1022         * src/arm/ffi.c (ffi_prep_args): Align structs by at least 4 bytes.
1023         Only treat r0 as a struct address if we're actually returning a
1024         struct by address.
1025         Only copy the bytes that are actually within a struct.
1026         (ffi_prep_cif_machdep): A Composite Type not larger than 4 bytes
1027         is returned in r0, not passed by address.
1028         (ffi_call): Allocate a word-sized temporary for the case where
1029         a composite is returned in r0.
1030         (ffi_prep_incoming_args_SYSV): Align as necessary.
1031
1032 2007-08-05  Steven Newbury  <s_j_newbury@yahoo.co.uk>
1033
1034         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Use __clear_cache instead of
1035         directly using the sys_cacheflush syscall.
1036
1037 2007-07-27  Andrew Haley  <aph@redhat.com>
1038
1039         * src/arm/sysv.S (ffi_closure_SYSV): Add soft-float.
1040
1041 2007-09-03  Maciej W. Rozycki  <macro@linux-mips.org>
1042
1043         * Makefile.am: Unify MIPS_IRIX and MIPS_LINUX into MIPS.
1044         * configure.ac: Likewise.
1045         * Makefile.in: Regenerate.
1046         * include/Makefile.in: Likewise.
1047         * testsuite/Makefile.in: Likewise.
1048         * configure: Likewise.
1049
1050 2007-08-24  David Daney  <ddaney@avtrex.com>
1051
1052         * testsuite/libffi.call/return_sl.c: New test.
1053
1054 2007-08-10  David Daney  <ddaney@avtrex.com>
1055
1056         * testsuite/libffi.call/cls_multi_ushort.c,
1057         testsuite/libffi.call/cls_align_uint16.c,
1058         testsuite/libffi.call/nested_struct1.c,
1059         testsuite/libffi.call/nested_struct3.c,
1060         testsuite/libffi.call/cls_7_1_byte.c,
1061         testsuite/libffi.call/nested_struct5.c,
1062         testsuite/libffi.call/cls_double.c,
1063         testsuite/libffi.call/nested_struct7.c,
1064         testsuite/libffi.call/cls_sint.c,
1065         testsuite/libffi.call/nested_struct9.c,
1066         testsuite/libffi.call/cls_20byte1.c,
1067         testsuite/libffi.call/cls_multi_sshortchar.c,
1068         testsuite/libffi.call/cls_align_sint64.c,
1069         testsuite/libffi.call/cls_3byte2.c,
1070         testsuite/libffi.call/cls_multi_schar.c,
1071         testsuite/libffi.call/cls_multi_uchar.c,
1072         testsuite/libffi.call/cls_19byte.c,
1073         testsuite/libffi.call/cls_9byte1.c,
1074         testsuite/libffi.call/cls_align_float.c,
1075         testsuite/libffi.call/closure_fn1.c,
1076         testsuite/libffi.call/problem1.c,
1077         testsuite/libffi.call/closure_fn3.c,
1078         testsuite/libffi.call/cls_sshort.c,
1079         testsuite/libffi.call/closure_fn5.c,
1080         testsuite/libffi.call/cls_align_double.c,
1081         testsuite/libffi.call/nested_struct.c,
1082         testsuite/libffi.call/cls_2byte.c,
1083         testsuite/libffi.call/nested_struct10.c,
1084         testsuite/libffi.call/cls_4byte.c,
1085         testsuite/libffi.call/cls_6byte.c,
1086         testsuite/libffi.call/cls_8byte.c,
1087         testsuite/libffi.call/cls_multi_sshort.c,
1088         testsuite/libffi.call/cls_align_sint16.c,
1089         testsuite/libffi.call/cls_align_uint32.c,
1090         testsuite/libffi.call/cls_20byte.c,
1091         testsuite/libffi.call/cls_float.c,
1092         testsuite/libffi.call/nested_struct2.c,
1093         testsuite/libffi.call/cls_5_1_byte.c,
1094         testsuite/libffi.call/nested_struct4.c,
1095         testsuite/libffi.call/cls_24byte.c,
1096         testsuite/libffi.call/nested_struct6.c,
1097         testsuite/libffi.call/cls_64byte.c,
1098         testsuite/libffi.call/nested_struct8.c,
1099         testsuite/libffi.call/cls_uint.c,
1100         testsuite/libffi.call/cls_multi_ushortchar.c,
1101         testsuite/libffi.call/cls_schar.c,
1102         testsuite/libffi.call/cls_uchar.c,
1103         testsuite/libffi.call/cls_align_uint64.c,
1104         testsuite/libffi.call/cls_ulonglong.c,
1105         testsuite/libffi.call/cls_align_longdouble.c,
1106         testsuite/libffi.call/cls_1_1byte.c,
1107         testsuite/libffi.call/cls_12byte.c,
1108         testsuite/libffi.call/cls_3_1byte.c,
1109         testsuite/libffi.call/cls_3byte1.c,
1110         testsuite/libffi.call/cls_4_1byte.c,
1111         testsuite/libffi.call/cls_6_1_byte.c,
1112         testsuite/libffi.call/cls_16byte.c,
1113         testsuite/libffi.call/cls_18byte.c,
1114         testsuite/libffi.call/closure_fn0.c,
1115         testsuite/libffi.call/cls_9byte2.c,
1116         testsuite/libffi.call/closure_fn2.c,
1117         testsuite/libffi.call/closure_fn4.c,
1118         testsuite/libffi.call/cls_ushort.c,
1119         testsuite/libffi.call/closure_fn6.c,
1120         testsuite/libffi.call/cls_5byte.c,
1121         testsuite/libffi.call/cls_align_pointer.c,
1122         testsuite/libffi.call/cls_7byte.c,
1123         testsuite/libffi.call/cls_align_sint32.c,
1124         testsuite/libffi.special/unwindtest_ffi_call.cc,
1125         testsuite/libffi.special/unwindtest.cc: Remove xfail for mips64*-*-*.
1126
1127 2007-08-10  David Daney  <ddaney@avtrex.com>
1128
1129         PR libffi/28313
1130         * configure.ac: Don't treat mips64 as a special case.
1131         * Makefile.am (nodist_libffi_la_SOURCES): Add n32.S.
1132         * configure: Regenerate
1133         * Makefile.in: Ditto.
1134         * fficonfig.h.in: Ditto.
1135         * src/mips/ffitarget.h (REG_L, REG_S, SUBU, ADDU, SRL, LI): Indent.
1136         (LA, EH_FRAME_ALIGN, FDE_ADDR_BYTES): New preprocessor macros.
1137         (FFI_DEFAULT_ABI): Set for n64 case.
1138         (FFI_CLOSURES, FFI_TRAMPOLINE_SIZE): Define for n32 and n64 cases.
1139         * src/mips/n32.S (ffi_call_N32): Add debug macros and labels for FDE.
1140         (ffi_closure_N32): New function.
1141         (.eh_frame): New section
1142         * src/mips/o32.S: Clean up comments.
1143         (ffi_closure_O32): Pass ffi_closure parameter in $12.
1144         * src/mips/ffi.c: Use FFI_MIPS_N32 instead of
1145         _MIPS_SIM == _ABIN32 throughout.
1146         (FFI_MIPS_STOP_HERE): New, use in place of
1147         ffi_stop_here.
1148         (ffi_prep_args): Use unsigned long to hold pointer values.  Rewrite
1149         to support n32/n64 ABIs.
1150         (calc_n32_struct_flags): Rewrite.
1151         (calc_n32_return_struct_flags): Remove unused variable.  Reverse
1152         position of flag bits.
1153         (ffi_prep_cif_machdep): Rewrite n32 portion.
1154         (ffi_call): Enable for n64.  Add special handling for small structure
1155         return values.
1156         (ffi_prep_closure_loc): Add n32 and n64 support.
1157         (ffi_closure_mips_inner_O32): Add cast to silence warning.
1158         (copy_struct_N32, ffi_closure_mips_inner_N32): New functions.
1159
1160 2007-08-08  David Daney  <ddaney@avtrex.com>
1161
1162         * testsuite/libffi.call/ffitest.h (ffi_type_mylong): Remove definition.
1163         * testsuite/libffi.call/cls_align_uint16.c (main): Use correct type
1164         specifiers.
1165         * testsuite/libffi.call/nested_struct1.c (main): Ditto.
1166         * testsuite/libffi.call/cls_sint.c (main): Ditto.
1167         * testsuite/libffi.call/nested_struct9.c (main): Ditto.
1168         * testsuite/libffi.call/cls_20byte1.c (main): Ditto.
1169         * testsuite/libffi.call/cls_9byte1.c (main): Ditto.
1170         * testsuite/libffi.call/closure_fn1.c (main): Ditto.
1171         * testsuite/libffi.call/closure_fn3.c (main): Ditto.
1172         * testsuite/libffi.call/return_dbl2.c (main): Ditto.
1173         * testsuite/libffi.call/cls_sshort.c (main): Ditto.
1174         * testsuite/libffi.call/return_fl3.c (main): Ditto.
1175         * testsuite/libffi.call/closure_fn5.c (main): Ditto.
1176         * testsuite/libffi.call/nested_struct.c (main): Ditto.
1177         * testsuite/libffi.call/nested_struct10.c (main): Ditto.
1178         * testsuite/libffi.call/return_ll1.c (main): Ditto.
1179         * testsuite/libffi.call/cls_8byte.c (main): Ditto.
1180         * testsuite/libffi.call/cls_align_uint32.c (main): Ditto.
1181         * testsuite/libffi.call/cls_align_sint16.c (main): Ditto.
1182         * testsuite/libffi.call/cls_20byte.c (main): Ditto.
1183         * testsuite/libffi.call/nested_struct2.c (main): Ditto.
1184         * testsuite/libffi.call/cls_24byte.c (main): Ditto.
1185         * testsuite/libffi.call/nested_struct6.c (main): Ditto.
1186         * testsuite/libffi.call/cls_uint.c (main): Ditto.
1187         * testsuite/libffi.call/cls_12byte.c (main): Ditto.
1188         * testsuite/libffi.call/cls_16byte.c (main): Ditto.
1189         * testsuite/libffi.call/closure_fn0.c (main): Ditto.
1190         * testsuite/libffi.call/cls_9byte2.c (main): Ditto.
1191         * testsuite/libffi.call/closure_fn2.c (main): Ditto.
1192         * testsuite/libffi.call/return_dbl1.c (main): Ditto.
1193         * testsuite/libffi.call/closure_fn4.c (main): Ditto.
1194         * testsuite/libffi.call/closure_fn6.c (main): Ditto.
1195         * testsuite/libffi.call/cls_align_sint32.c (main): Ditto.
1196
1197 2007-08-07  Andrew Haley  <aph@redhat.com>
1198
1199         * src/x86/sysv.S (ffi_closure_raw_SYSV): Fix typo in previous
1200         checkin.
1201
1202 2007-08-06  Andrew Haley  <aph@redhat.com>
1203
1204         PR testsuite/32843
1205         * src/x86/sysv.S (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1206         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1207         FFI_TYPE_SINT32.
1208
1209 2007-08-02  David Daney  <ddaney@avtrex.com>
1210
1211         * testsuite/libffi.call/return_ul.c (main): Define return type as
1212         ffi_arg.  Use proper printf conversion specifier.
1213
1214 2007-07-30  Andrew Haley  <aph@redhat.com>
1215
1216         PR testsuite/32843
1217         * src/x86/ffi.c (ffi_prep_cif_machdep): in x86 case, add code for
1218         signed/unsigned int8/16.
1219         * src/x86/sysv.S (ffi_call_SYSV): Rewrite to:
1220         Use a jump table.
1221         Remove code to pop args from the stack after call.
1222         Special-case signed/unsigned int8/16.
1223         * testsuite/libffi.call/return_sc.c (main): Revert.
1224
1225 2007-07-26  Richard Guenther  <rguenther@suse.de>
1226
1227         PR testsuite/32843
1228         * testsuite/libffi.call/return_sc.c (main): Verify call
1229         result as signed char, not ffi_arg.
1230
1231 2007-07-16  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1232
1233         * configure.ac (i?86-*-solaris2.1[0-9]): Set TARGET to X86_64.
1234         * configure: Regenerate.
1235
1236 2007-07-11  David Daney  <ddaney@avtrex.com>
1237
1238         * src/mips/ffi.c: Don't include sys/cachectl.h.
1239         (ffi_prep_closure_loc): Use __builtin___clear_cache() instead of
1240         cacheflush().
1241
1242 2007-05-18  Aurelien Jarno  <aurelien@aurel32.net>
1243
1244         * src/arm/ffi.c (ffi_prep_closure_loc): Renamed and ajusted
1245         from (ffi_prep_closure): ... this.
1246         (FFI_INIT_TRAMPOLINE): Adjust.
1247
1248 2005-12-31  Phil Blundell  <pb@reciva.com>
1249
1250         * src/arm/ffi.c (ffi_prep_incoming_args_SYSV,
1251         ffi_closure_SYSV_inner, ffi_prep_closure): New, add closure support.
1252         * src/arm/sysv.S(ffi_closure_SYSV): Likewise.
1253         * src/arm/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1254         (FFI_CLOSURES): Enable closure support.
1255
1256 2007-07-03  Andrew Haley  <aph@hedges.billgatliff.com>
1257
1258         * testsuite/libffi.call/cls_multi_ushort.c,
1259         testsuite/libffi.call/cls_align_uint16.c,
1260         testsuite/libffi.call/nested_struct1.c,
1261         testsuite/libffi.call/nested_struct3.c,
1262         testsuite/libffi.call/cls_7_1_byte.c,
1263         testsuite/libffi.call/cls_double.c,
1264         testsuite/libffi.call/nested_struct5.c,
1265         testsuite/libffi.call/nested_struct7.c,
1266         testsuite/libffi.call/cls_sint.c,
1267         testsuite/libffi.call/nested_struct9.c,
1268         testsuite/libffi.call/cls_20byte1.c,
1269         testsuite/libffi.call/cls_multi_sshortchar.c,
1270         testsuite/libffi.call/cls_align_sint64.c,
1271         testsuite/libffi.call/cls_3byte2.c,
1272         testsuite/libffi.call/cls_multi_schar.c,
1273         testsuite/libffi.call/cls_multi_uchar.c,
1274         testsuite/libffi.call/cls_19byte.c,
1275         testsuite/libffi.call/cls_9byte1.c,
1276         testsuite/libffi.call/cls_align_float.c,
1277         testsuite/libffi.call/closure_fn1.c,
1278         testsuite/libffi.call/problem1.c,
1279         testsuite/libffi.call/closure_fn3.c,
1280         testsuite/libffi.call/cls_sshort.c,
1281         testsuite/libffi.call/closure_fn5.c,
1282         testsuite/libffi.call/cls_align_double.c,
1283         testsuite/libffi.call/cls_2byte.c,
1284         testsuite/libffi.call/nested_struct.c,
1285         testsuite/libffi.call/nested_struct10.c,
1286         testsuite/libffi.call/cls_4byte.c,
1287         testsuite/libffi.call/cls_6byte.c,
1288         testsuite/libffi.call/cls_8byte.c,
1289         testsuite/libffi.call/cls_multi_sshort.c,
1290         testsuite/libffi.call/cls_align_uint32.c,
1291         testsuite/libffi.call/cls_align_sint16.c,
1292         testsuite/libffi.call/cls_float.c,
1293         testsuite/libffi.call/cls_20byte.c,
1294         testsuite/libffi.call/cls_5_1_byte.c,
1295         testsuite/libffi.call/nested_struct2.c,
1296         testsuite/libffi.call/cls_24byte.c,
1297         testsuite/libffi.call/nested_struct4.c,
1298         testsuite/libffi.call/nested_struct6.c,
1299         testsuite/libffi.call/cls_64byte.c,
1300         testsuite/libffi.call/nested_struct8.c,
1301         testsuite/libffi.call/cls_uint.c,
1302         testsuite/libffi.call/cls_multi_ushortchar.c,
1303         testsuite/libffi.call/cls_schar.c,
1304         testsuite/libffi.call/cls_uchar.c,
1305         testsuite/libffi.call/cls_align_uint64.c,
1306         testsuite/libffi.call/cls_ulonglong.c,
1307         testsuite/libffi.call/cls_align_longdouble.c,
1308         testsuite/libffi.call/cls_1_1byte.c,
1309         testsuite/libffi.call/cls_12byte.c,
1310         testsuite/libffi.call/cls_3_1byte.c,
1311         testsuite/libffi.call/cls_3byte1.c,
1312         testsuite/libffi.call/cls_4_1byte.c,
1313         testsuite/libffi.call/cls_6_1_byte.c,
1314         testsuite/libffi.call/cls_16byte.c,
1315         testsuite/libffi.call/cls_18byte.c,
1316         testsuite/libffi.call/closure_fn0.c,
1317         testsuite/libffi.call/cls_9byte2.c,
1318         testsuite/libffi.call/closure_fn2.c,
1319         testsuite/libffi.call/closure_fn4.c,
1320         testsuite/libffi.call/cls_ushort.c,
1321         testsuite/libffi.call/closure_fn6.c,
1322         testsuite/libffi.call/cls_5byte.c,
1323         testsuite/libffi.call/cls_align_pointer.c,
1324         testsuite/libffi.call/cls_7byte.c,
1325         testsuite/libffi.call/cls_align_sint32.c,
1326         testsuite/libffi.special/unwindtest_ffi_call.cc,
1327         testsuite/libffi.special/unwindtest.cc: Enable for ARM.
1328
1329 2007-07-05  H.J. Lu  <hongjiu.lu@intel.com>
1330
1331         * aclocal.m4: Regenerated.
1332
1333 2007-06-02  Paolo Bonzini  <bonzini@gnu.org>
1334
1335         * configure: Regenerate.
1336
1337 2007-05-23  Steve Ellcey  <sje@cup.hp.com>
1338
1339         * Makefile.in: Regenerate.
1340         * configure: Regenerate.
1341         * aclocal.m4: Regenerate.
1342         * include/Makefile.in: Regenerate.
1343         * testsuite/Makefile.in: Regenerate.
1344
1345 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1346
1347         * src/m68k/ffi.c (ffi_prep_incoming_args_SYSV,
1348         ffi_closure_SYSV_inner,ffi_prep_closure): New, add closure support.
1349         * src/m68k/sysv.S(ffi_closure_SYSV,ffi_closure_struct_SYSV): Likewise.
1350         * src/m68k/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1351         (FFI_CLOSURES): Enable closure support.
1352
1353 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1354
1355         * configure.ac (HAVE_AS_CFI_PSEUDO_OP): New test.
1356         * configure: Regenerate.
1357         * fficonfig.h.in: Regenerate.
1358         * src/m68k/sysv.S (CFI_STARTPROC,CFI_ENDPROC,
1359         CFI_OFFSET,CFI_DEF_CFA): New macros.
1360         (ffi_call_SYSV): Add callframe annotation.
1361
1362 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1363
1364         * src/m68k/ffi.c (ffi_prep_args,ffi_prep_cif_machdep): Fix
1365         numerous test suite failures.
1366         * src/m68k/sysv.S (ffi_call_SYSV): Likewise.
1367
1368 2007-04-11  Paolo Bonzini  <bonzini@gnu.org>
1369
1370         * Makefile.am (EXTRA_DIST): Bring up to date.
1371         * Makefile.in: Regenerate.
1372         * src/frv/eabi.S: Remove RCS keyword.
1373
1374 2007-04-06  Richard Henderson  <rth@redhat.com>
1375
1376         * configure.ac: Tidy target case.
1377         (HAVE_LONG_DOUBLE): Allow the target to override.
1378         * configure: Regenerate.
1379         * include/ffi.h.in: Don't define ffi_type_foo if
1380         LIBFFI_HIDE_BASIC_TYPES is defined.
1381         (ffi_type_longdouble): If not HAVE_LONG_DOUBLE, define
1382         to ffi_type_double.
1383         * types.c (LIBFFI_HIDE_BASIC_TYPES): Define.
1384         (FFI_TYPEDEF, ffi_type_void): Mark the data const.
1385         (ffi_type_longdouble): Special case for Alpha.  Don't define
1386         if long double == double.
1387
1388         * src/alpha/ffi.c (FFI_TYPE_LONGDOUBLE): Assert unique value.
1389         (ffi_prep_cif_machdep): Handle it as the 128-bit type.
1390         (ffi_call, ffi_closure_osf_inner): Likewise.
1391         (ffi_closure_osf_inner): Likewise.  Mark hidden.
1392         (ffi_call_osf, ffi_closure_osf): Mark hidden.
1393         * src/alpha/ffitarget.h (FFI_LAST_ABI): Tidy definition.
1394         * src/alpha/osf.S (ffi_call_osf, ffi_closure_osf): Mark hidden.
1395         (load_table): Handle 128-bit long double.
1396
1397         * testsuite/libffi.call/float4.c: Add -mieee for alpha.
1398
1399 2007-04-06  Tom Tromey  <tromey@redhat.com>
1400
1401         PR libffi/31491:
1402         * README: Fixed bug in example.
1403
1404 2007-04-03  Jakub Jelinek  <jakub@redhat.com>
1405
1406         * src/closures.c: Include sys/statfs.h.
1407         (_GNU_SOURCE): Define on Linux.
1408         (FFI_MMAP_EXEC_SELINUX): Define.
1409         (selinux_enabled): New variable.
1410         (selinux_enabled_check): New function.
1411         (is_selinux_enabled): Define.
1412         (dlmmap): Use it.
1413
1414 2007-03-24  Uros Bizjak  <ubizjak@gmail.com>
1415
1416         * testsuite/libffi.call/return_fl2.c (return_fl): Mark as static.
1417         Use 'volatile float sum' to create sum of floats to avoid false
1418         negative due to excess precision on ix86 targets.
1419         (main): Ditto.
1420
1421 2007-03-08  Alexandre Oliva  <aoliva@redhat.com>
1422
1423         * src/powerpc/ffi.c (flush_icache): Fix left-over from previous
1424         patch.
1425         (ffi_prep_closure_loc): Remove unneeded casts.  Add needed ones.
1426
1427 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1428
1429         * include/ffi.h.in (ffi_closure_alloc, ffi_closure_free): New.
1430         (ffi_prep_closure_loc): New.
1431         (ffi_prep_raw_closure_loc): New.
1432         (ffi_prep_java_raw_closure_loc): New.
1433         * src/closures.c: New file.
1434         * src/dlmalloc.c [FFI_MMAP_EXEC_WRIT] (struct malloc_segment):
1435         Replace sflags with exec_offset.
1436         [FFI_MMAP_EXEC_WRIT] (mmap_exec_offset, add_segment_exec_offset,
1437         sub_segment_exec_offset): New macros.
1438         (get_segment_flags, set_segment_flags, check_segment_merge): New
1439         macros.
1440         (is_mmapped_segment, is_extern_segment): Use get_segment_flags.
1441         (add_segment, sys_alloc, create_mspace, create_mspace_with_base,
1442         destroy_mspace): Use new macros.
1443         (sys_alloc): Silence warning.
1444         * Makefile.am (libffi_la_SOURCES): Add src/closures.c.
1445         * Makefile.in: Rebuilt.
1446         * src/prep_cif [FFI_CLOSURES] (ffi_prep_closure): Implement in
1447         terms of ffi_prep_closure_loc.
1448         * src/raw_api.c (ffi_prep_raw_closure_loc): Renamed and adjusted
1449         from...
1450         (ffi_prep_raw_closure): ... this.  Re-implement in terms of the
1451         renamed version.
1452         * src/java_raw_api (ffi_prep_java_raw_closure_loc): Renamed and
1453         adjusted from...
1454         (ffi_prep_java_raw_closure): ... this.  Re-implement in terms of
1455         the renamed version.
1456         * src/alpha/ffi.c (ffi_prep_closure_loc): Renamed from
1457         (ffi_prep_closure): ... this.
1458         * src/pa/ffi.c: Likewise.
1459         * src/cris/ffi.c: Likewise.  Adjust.
1460         * src/frv/ffi.c: Likewise.
1461         * src/ia64/ffi.c: Likewise.
1462         * src/mips/ffi.c: Likewise.
1463         * src/powerpc/ffi_darwin.c: Likewise.
1464         * src/s390/ffi.c: Likewise.
1465         * src/sh/ffi.c: Likewise.
1466         * src/sh64/ffi.c: Likewise.
1467         * src/sparc/ffi.c: Likewise.
1468         * src/x86/ffi64.c: Likewise.
1469         * src/x86/ffi.c: Likewise.
1470         (FFI_INIT_TRAMPOLINE): Adjust.
1471         (ffi_prep_raw_closure_loc): Renamed and adjusted from...
1472         (ffi_prep_raw_closure): ... this.
1473         * src/powerpc/ffi.c (ffi_prep_closure_loc): Renamed from
1474         (ffi_prep_closure): ... this.
1475         (flush_icache): Adjust.
1476
1477 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1478
1479         * src/dlmalloc.c: New file, imported version 2.8.3 of Doug
1480         Lea's malloc.
1481
1482 2007-03-01  Brooks Moses  <brooks.moses@codesourcery.com>
1483
1484         * Makefile.am: Add dummy install-pdf target.
1485         * Makefile.in: Regenerate
1486
1487 2007-02-13  Andreas Krebbel  <krebbel1@de.ibm.com>
1488
1489         * src/s390/ffi.c (ffi_prep_args, ffi_prep_cif_machdep,
1490         ffi_closure_helper_SYSV): Add long double handling.
1491
1492 2007-02-02  Jakub Jelinek  <jakub@redhat.com>
1493
1494         * src/powerpc/linux64.S (ffi_call_LINUX64): Move restore of r2
1495         immediately after bctrl instruction.
1496
1497 2007-01-18  Alexandre Oliva  <aoliva@redhat.com>
1498
1499         * Makefile.am (all-recursive, install-recursive,
1500         mostlyclean-recursive, clean-recursive, distclean-recursive,
1501         maintainer-clean-recursive): Add missing targets.
1502         * Makefile.in: Rebuilt.
1503
1504 2006-12-14  Andreas Tobler  <a.tobler@schweiz.org>
1505
1506         * configure.ac: Add TARGET for x86_64-*-darwin*.
1507         * Makefile.am (nodist_libffi_la_SOURCES): Add rules for 64-bit sources
1508         for X86_DARWIN.
1509         * src/x86/ffitarget.h: Set trampoline size for x86_64-*-darwin*.
1510         * src/x86/darwin64.S: New file for x86_64-*-darwin* support.
1511         * configure: Regenerate.
1512         * Makefile.in: Regenerate.
1513         * include/Makefile.in: Regenerate.
1514         * testsuite/Makefile.in: Regenerate.
1515         * testsuite/libffi.special/unwindtest_ffi_call.cc: New test case for
1516         ffi_call only.
1517
1518 2006-12-13  Andreas Tobler <a.tobler@schweiz.org>
1519
1520         * aclocal.m4: Regenerate with aclocal -I .. as written in the
1521         Makefile.am.
1522
1523 2006-10-31  Geoffrey Keating  <geoffk@apple.com>
1524
1525         * src/powerpc/ffi_darwin.c (darwin_adjust_aggregate_sizes): New.
1526         (ffi_prep_cif_machdep): Call darwin_adjust_aggregate_sizes for
1527         Darwin.
1528         * testsuite/libffi.call/nested_struct4.c: Remove Darwin XFAIL.
1529         * testsuite/libffi.call/nested_struct6.c: Remove Darwin XFAIL.
1530
1531 2006-10-10  Paolo Bonzini  <bonzini@gnu.org>
1532             Sandro Tolaini  <tolaini@libero.it>
1533
1534         * configure.ac [i*86-*-darwin*]: Set X86_DARWIN symbol and
1535         conditional.
1536         * configure: Regenerated.
1537         * Makefile.am (nodist_libffi_la_SOURCES) [X86_DARWIN]: New case.
1538         (EXTRA_DIST): Add src/x86/darwin.S.
1539         * Makefile.in: Regenerated.
1540         * include/Makefile.in: Regenerated.
1541         * testsuite/Makefile.in: Regenerated.
1542
1543         * src/x86/ffi.c (ffi_prep_cif_machdep) [X86_DARWIN]: Treat like
1544         X86_WIN32, and additionally align stack to 16 bytes.
1545         * src/x86/darwin.S: New, based on sysv.S.
1546         * src/prep_cif.c (ffi_prep_cif) [X86_DARWIN]: Align > 8-byte structs.
1547
1548 2006-09-12  David Daney  <ddaney@avtrex.com>
1549
1550         PR libffi/23935
1551         * include/Makefile.am: Install both ffi.h and ffitarget.h in
1552         $(libdir)/gcc/$(target_alias)/$(gcc_version)/include.
1553         * aclocal.m4: Regenerated for automake 1.9.6.
1554         * Makefile.in: Regenerated.
1555         * include/Makefile.in: Regenerated.
1556         * testsuite/Makefile.in: Regenerated.
1557
1558 2006-08-17  Andreas Tobler  <a.tobler@schweiz.ch>
1559
1560         * include/ffi_common.h (struct): Revert accidental commit.
1561
1562 2006-08-15  Andreas Tobler  <a.tobler@schweiz.ch>
1563
1564         * include/ffi_common.h: Remove lint directives.
1565         * include/ffi.h.in: Likewise.
1566
1567 2006-07-25  Torsten Schoenfeld  <kaffeetisch@gmx.de>
1568
1569         * include/ffi.h.in (ffi_type_ulong, ffi_type_slong): Define correctly
1570         for 32-bit architectures.
1571         * testsuite/libffi.call/return_ul.c: New test case.
1572
1573 2006-07-19  David Daney  <ddaney@avtrex.com>
1574
1575         * testsuite/libffi.call/closure_fn6.c: Remove xfail for mips,
1576         xfail remains for mips64.
1577
1578 2006-05-23  Carlos O'Donell  <carlos@codesourcery.com>
1579
1580         * Makefile.am: Add install-html target. Add install-html to .PHONY
1581         * Makefile.in: Regenerate.
1582         * aclocal.m4: Regenerate.
1583         * include/Makefile.in: Regenerate.
1584         * testsuite/Makefile.in: Regenerate.
1585
1586 2006-05-18  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1587
1588         * pa/ffi.c (ffi_prep_args_pa32): Load floating point arguments from
1589         stack slot.
1590
1591 2006-04-22  Andreas Tobler  <a.tobler@schweiz.ch>
1592
1593         * README: Remove notice about 'Crazy Comments'.
1594         * src/debug.c: Remove lint directives. Cleanup white spaces.
1595         * src/java_raw_api.c: Likewise.
1596         * src/prep_cif.c: Likewise.
1597         * src/raw_api.c: Likewise.
1598         * src/ffitest.c: Delete. No longer needed, all test cases migrated
1599         to the testsuite.
1600         * src/arm/ffi.c: Remove lint directives.
1601         * src/m32r/ffi.c: Likewise.
1602         * src/pa/ffi.c: Likewise.
1603         * src/powerpc/ffi.c: Likewise.
1604         * src/powerpc/ffi_darwin.c: Likewise.
1605         * src/sh/ffi.c: Likewise.
1606         * src/sh64/ffi.c: Likewise.
1607         * src/x86/ffi.c: Likewise.
1608         * testsuite/libffi.call/float2.c: Likewise.
1609         * testsuite/libffi.call/promotion.c: Likewise.
1610         * testsuite/libffi.call/struct1.c: Likewise.
1611
1612 2006-04-13  Andreas Tobler  <a.tobler@schweiz.ch>
1613
1614         * src/pa/hpux32.S: Correct unwind offset calculation for
1615         ffi_closure_pa32.
1616         * src/pa/linux.S: Likewise.
1617
1618 2006-04-12  James E Wilson  <wilson@specifix.com>
1619
1620         PR libgcj/26483
1621         * src/ia64/ffi.c (stf_spill, ldf_fill): Rewrite as macros.
1622         (hfa_type_load): Call stf_spill.
1623         (hfa_type_store): Call ldf_fill.
1624         (ffi_call): Adjust calls to above routines.  Add local temps for
1625         macro result.
1626
1627 2006-04-10  Matthias Klose  <doko@debian.org>
1628
1629         * testsuite/lib/libffi-dg.exp (libffi-init): Recognize multilib
1630         directory names containing underscores.
1631
1632 2006-04-07  James E Wilson  <wilson@specifix.com>
1633
1634         * testsuite/libffi.call/float4.c: New testcase.
1635
1636 2006-04-05  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1637             Andreas Tobler  <a.tobler@schweiz.ch>
1638
1639         * Makefile.am: Add PA_HPUX port.
1640         * Makefile.in: Regenerate.
1641         * include/Makefile.in: Likewise.
1642         * testsuite/Makefile.in: Likewise.
1643         * configure.ac: Add PA_HPUX rules.
1644         * configure: Regenerate.
1645         * src/pa/ffitarget.h: Rename linux target to PA_LINUX.
1646         Add PA_HPUX and PA64_HPUX.
1647         Rename FFI_LINUX ABI to FFI_PA32 ABI.
1648         (FFI_TRAMPOLINE_SIZE): Define for 32-bit HP-UX targets.
1649         (FFI_TYPE_SMALL_STRUCT2): Define.
1650         (FFI_TYPE_SMALL_STRUCT4): Likewise.
1651         (FFI_TYPE_SMALL_STRUCT8): Likewise.
1652         (FFI_TYPE_SMALL_STRUCT3): Redefine.
1653         (FFI_TYPE_SMALL_STRUCT5): Likewise.
1654         (FFI_TYPE_SMALL_STRUCT6): Likewise.
1655         (FFI_TYPE_SMALL_STRUCT7): Likewise.
1656         * src/pa/ffi.c (ROUND_DOWN): Delete.
1657         (fldw, fstw, fldd, fstd): Use '__asm__'.
1658         (ffi_struct_type): Add support for FFI_TYPE_SMALL_STRUCT2,
1659         FFI_TYPE_SMALL_STRUCT4 and FFI_TYPE_SMALL_STRUCT8.
1660         (ffi_prep_args_LINUX): Rename to ffi_prep_args_pa32. Update comment.
1661         Simplify incrementing of stack slot variable. Change type of local
1662         'n' to unsigned int.
1663         (ffi_size_stack_LINUX): Rename to ffi_size_stack_pa32. Handle long
1664         double on PA_HPUX.
1665         (ffi_prep_cif_machdep): Likewise.
1666         (ffi_call): Likewise.
1667         (ffi_closure_inner_LINUX): Rename to ffi_closure_inner_pa32. Change
1668         return type to ffi_status. Simplify incrementing of stack slot
1669         variable. Only copy floating point argument registers when PA_LINUX
1670         is true. Reformat debug statement.
1671         Add support for FFI_TYPE_SMALL_STRUCT2, FFI_TYPE_SMALL_STRUCT4 and
1672         FFI_TYPE_SMALL_STRUCT8.
1673         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Add 'extern' to
1674         declaration.
1675         (ffi_prep_closure): Make linux trampoline conditional on PA_LINUX.
1676         Add nops to cache flush.  Add trampoline for PA_HPUX.
1677         * src/pa/hpux32.S: New file.
1678         * src/pa/linux.S (ffi_call_LINUX): Rename to ffi_call_pa32. Rename
1679         ffi_prep_args_LINUX to ffi_prep_args_pa32.
1680         Localize labels. Add support for 2, 4 and 8-byte small structs. Handle
1681         unaligned destinations in 3, 5, 6 and 7-byte small structs. Order
1682         argument type checks so that common argument types appear first.
1683         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Rename
1684         ffi_closure_inner_LINUX to ffi_closure_inner_pa32.
1685
1686 2006-03-24  Alan Modra  <amodra@bigpond.net.au>
1687
1688         * src/powerpc/ffitarget.h (enum ffi_abi): Add FFI_LINUX.  Default
1689         for 32-bit using IBM extended double format.  Fix FFI_LAST_ABI.
1690         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Handle linux variant of
1691         FFI_TYPE_LONGDOUBLE.
1692         (ffi_prep_args64): Assert using IBM extended double.
1693         (ffi_prep_cif_machdep): Don't munge FFI_TYPE_LONGDOUBLE type.
1694         Handle FFI_LINUX FFI_TYPE_LONGDOUBLE return and args.
1695         (ffi_call): Handle FFI_LINUX.
1696         (ffi_closure_helper_SYSV): Non FFI_LINUX long double return needs
1697         gpr3 return pointer as for struct return.  Handle FFI_LINUX
1698         FFI_TYPE_LONGDOUBLE return and args.  Don't increment "nf"
1699         unnecessarily.
1700         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Load both f1 and f2
1701         for FFI_TYPE_LONGDOUBLE.  Move epilogue insns into case table.
1702         Don't use r6 as pointer to results, instead use sp offset.  Don't
1703         make a special call to load lr with case table address, instead
1704         use offset from previous call.
1705         * src/powerpc/sysv.S (ffi_call_SYSV): Save long double return.
1706         * src/powerpc/linux64.S (ffi_call_LINUX64): Simplify long double
1707         return.
1708
1709 2006-03-15  Kaz Kojima  <kkojima@gcc.gnu.org>
1710
1711         * src/sh64/ffi.c (ffi_prep_cif_machdep): Handle float arguments
1712         passed with FP registers correctly.
1713         (ffi_closure_helper_SYSV): Likewise.
1714         * src/sh64/sysv.S: Likewise.
1715
1716 2006-03-01  Andreas Tobler  <a.tobler@schweiz.ch>
1717
1718         * testsuite/libffi.special/unwindtest.cc (closure_test_fn): Mark cif,
1719         args and userdata unused.
1720         (closure_test_fn1): Mark cif and userdata unused.
1721         (main): Remove unused res.
1722
1723 2006-02-28  Andreas Tobler  <a.tobler@schweiz.ch>
1724
1725         * testsuite/libffi.call/call.exp: Adjust FSF address. Add test runs for
1726         -O2, -O3, -Os and the warning flags -W -Wall.
1727         * testsuite/libffi.special/special.exp: Likewise.
1728         * testsuite/libffi.call/ffitest.h: Add an __UNUSED__ macro to mark
1729         unused parameter unused for gcc or else do nothing.
1730         * testsuite/libffi.special/ffitestcxx.h: Likewise.
1731         * testsuite/libffi.call/cls_12byte.c (cls_struct_12byte_gn): Mark cif
1732         and userdata unused.
1733         * testsuite/libffi.call/cls_16byte.c (cls_struct_16byte_gn): Likewise.
1734         * testsuite/libffi.call/cls_18byte.c (cls_struct_18byte_gn): Likewise.
1735         * testsuite/libffi.call/cls_19byte.c (cls_struct_19byte_gn): Likewise.
1736         * testsuite/libffi.call/cls_1_1byte.c (cls_struct_1_1byte_gn): Likewise.
1737         * testsuite/libffi.call/cls_20byte.c (cls_struct_20byte_gn): Likewise.
1738         * testsuite/libffi.call/cls_20byte1.c (cls_struct_20byte_gn): Likewise.
1739         * testsuite/libffi.call/cls_24byte.c (cls_struct_24byte_gn): Likewise.
1740         * testsuite/libffi.call/cls_2byte.c (cls_struct_2byte_gn): Likewise.
1741         * testsuite/libffi.call/cls_3_1byte.c (cls_struct_3_1byte_gn): Likewise.
1742         * testsuite/libffi.call/cls_3byte1.c (cls_struct_3byte_gn): Likewise.
1743         * testsuite/libffi.call/cls_3byte2.c (cls_struct_3byte_gn1): Likewise.
1744         * testsuite/libffi.call/cls_4_1byte.c (cls_struct_4_1byte_gn): Likewise.
1745         * testsuite/libffi.call/cls_4byte.c (cls_struct_4byte_gn): Likewise.
1746         * testsuite/libffi.call/cls_5_1_byte.c (cls_struct_5byte_gn): Likewise.
1747         * testsuite/libffi.call/cls_5byte.c (cls_struct_5byte_gn): Likewise.
1748         * testsuite/libffi.call/cls_64byte.c (cls_struct_64byte_gn): Likewise.
1749         * testsuite/libffi.call/cls_6_1_byte.c (cls_struct_6byte_gn): Likewise.
1750         * testsuite/libffi.call/cls_6byte.c (cls_struct_6byte_gn): Likewise.
1751         * testsuite/libffi.call/cls_7_1_byte.c (cls_struct_7byte_gn): Likewise.
1752         * testsuite/libffi.call/cls_7byte.c (cls_struct_7byte_gn): Likewise.
1753         * testsuite/libffi.call/cls_8byte.c (cls_struct_8byte_gn): Likewise.
1754         * testsuite/libffi.call/cls_9byte1.c (cls_struct_9byte_gn): Likewise.
1755         * testsuite/libffi.call/cls_9byte2.c (cls_struct_9byte_gn): Likewise.
1756         * testsuite/libffi.call/cls_align_double.c (cls_struct_align_gn):
1757         Likewise.
1758         * testsuite/libffi.call/cls_align_float.c (cls_struct_align_gn):
1759         Likewise.
1760         * testsuite/libffi.call/cls_align_longdouble.c (cls_struct_align_gn):
1761         Likewise.
1762         * testsuite/libffi.call/cls_align_pointer.c (cls_struct_align_fn): Cast
1763         void* to avoid compiler warning.
1764         (main): Likewise.
1765         (cls_struct_align_gn): Mark cif and userdata unused.
1766         * testsuite/libffi.call/cls_align_sint16.c (cls_struct_align_gn):
1767         Likewise.
1768         * testsuite/libffi.call/cls_align_sint32.c (cls_struct_align_gn):
1769         Likewise.
1770         * testsuite/libffi.call/cls_align_sint64.c (cls_struct_align_gn):
1771         Likewise.
1772         * testsuite/libffi.call/cls_align_uint16.c (cls_struct_align_gn):
1773         Likewise.
1774         * testsuite/libffi.call/cls_align_uint32.c (cls_struct_align_gn):
1775         Likewise.
1776         * testsuite/libffi.call/cls_double.c (cls_ret_double_fn): Likewise.
1777         * testsuite/libffi.call/cls_float.c (cls_ret_float_fn): Likewise.
1778         * testsuite/libffi.call/cls_multi_schar.c (test_func_gn): Mark cif and
1779         data unused.
1780         (main): Cast res_call to silence gcc.
1781         * testsuite/libffi.call/cls_multi_sshort.c (test_func_gn): Mark cif and
1782         data unused.
1783         (main): Cast res_call to silence gcc.
1784         * testsuite/libffi.call/cls_multi_sshortchar.c (test_func_gn): Mark cif
1785         and data unused.
1786         (main): Cast res_call to silence gcc.
1787         * testsuite/libffi.call/cls_multi_uchar.c (test_func_gn): Mark cif and
1788         data unused.
1789         (main): Cast res_call to silence gcc.
1790         * testsuite/libffi.call/cls_multi_ushort.c (test_func_gn): Mark cif and
1791         data unused.
1792         (main): Cast res_call to silence gcc.
1793         * testsuite/libffi.call/cls_multi_ushortchar.c (test_func_gn): Mark cif
1794         and data unused.
1795         (main): Cast res_call to silence gcc.
1796         * testsuite/libffi.call/cls_schar.c (cls_ret_schar_fn): Mark cif and
1797         userdata unused.
1798         (cls_ret_schar_fn): Cast printf parameter to silence gcc.
1799         * testsuite/libffi.call/cls_sint.c (cls_ret_sint_fn): Mark cif and
1800         userdata unused.
1801         (cls_ret_sint_fn): Cast printf parameter to silence gcc.
1802         * testsuite/libffi.call/cls_sshort.c (cls_ret_sshort_fn): Mark cif and
1803         userdata unused.
1804         (cls_ret_sshort_fn): Cast printf parameter to silence gcc.
1805         * testsuite/libffi.call/cls_uchar.c (cls_ret_uchar_fn):  Mark cif and
1806         userdata unused.
1807         (cls_ret_uchar_fn): Cast printf parameter to silence gcc.
1808         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Mark cif and
1809         userdata unused.
1810         (cls_ret_uint_fn): Cast printf parameter to silence gcc.
1811         * testsuite/libffi.call/cls_ulonglong.c (cls_ret_ulonglong_fn): Mark cif
1812         and userdata unused.
1813         * testsuite/libffi.call/cls_ushort.c (cls_ret_ushort_fn): Mark cif and
1814         userdata unused.
1815         (cls_ret_ushort_fn): Cast printf parameter to silence gcc.
1816         * testsuite/libffi.call/float.c (floating): Remove unused parameter e.
1817         * testsuite/libffi.call/float1.c (main): Remove unused variable i.
1818         Cleanup white spaces.
1819         * testsuite/libffi.call/negint.c (checking): Remove unused variable i.
1820         * testsuite/libffi.call/nested_struct.c (cls_struct_combined_gn): Mark
1821         cif and userdata unused.
1822         * testsuite/libffi.call/nested_struct1.c (cls_struct_combined_gn):
1823         Likewise.
1824         * testsuite/libffi.call/nested_struct10.c (B_gn): Likewise.
1825         * testsuite/libffi.call/nested_struct2.c (B_fn): Adjust printf
1826         formatters to silence gcc.
1827         (B_gn): Mark cif and userdata unused.
1828         * testsuite/libffi.call/nested_struct3.c (B_gn): Mark cif and userdata
1829         unused.
1830         * testsuite/libffi.call/nested_struct4.c: Mention related PR.
1831         (B_gn): Mark cif and userdata unused.
1832         * testsuite/libffi.call/nested_struct5.c (B_gn): Mark cif and userdata
1833         unused.
1834         * testsuite/libffi.call/nested_struct6.c: Mention related PR.
1835         (B_gn): Mark cif and userdata unused.
1836         * testsuite/libffi.call/nested_struct7.c (B_gn): Mark cif and userdata
1837         unused.
1838         * testsuite/libffi.call/nested_struct8.c (B_gn): Likewise.
1839         * testsuite/libffi.call/nested_struct9.c (B_gn): Likewise.
1840         * testsuite/libffi.call/problem1.c (stub): Likewise.
1841         * testsuite/libffi.call/pyobjc-tc.c (main): Cast the result to silence
1842         gcc.
1843         * testsuite/libffi.call/return_fl2.c (return_fl): Add the note mentioned
1844         in the last commit for this test case in the test case itself.
1845         * testsuite/libffi.call/closure_fn0.c (closure_test_fn0): Mark cif as
1846         unused.
1847         * testsuite/libffi.call/closure_fn1.c (closure_test_fn1): Likewise.
1848         * testsuite/libffi.call/closure_fn2.c (closure_test_fn2): Likewise.
1849         * testsuite/libffi.call/closure_fn3.c (closure_test_fn3): Likewise.
1850         * testsuite/libffi.call/closure_fn4.c (closure_test_fn0): Likewise.
1851         * testsuite/libffi.call/closure_fn5.c (closure_test_fn5): Likewise.
1852         * testsuite/libffi.call/closure_fn6.c (closure_test_fn0): Likewise.
1853
1854 2006-02-22  Kaz Kojima  <kkojima@gcc.gnu.org>
1855
1856         * src/sh/sysv.S: Fix register numbers in the FDE for
1857         ffi_closure_SYSV.
1858
1859 2006-02-20  Andreas Tobler  <a.tobler@schweiz.ch>
1860
1861         * testsuite/libffi.call/return_fl2.c (return_fl): Remove static
1862         declaration to avoid a false negative on ix86. See PR323.
1863
1864 2006-02-18  Kaz Kojima  <kkojima@gcc.gnu.org>
1865
1866         * src/sh/ffi.c (ffi_closure_helper_SYSV): Remove unused variable
1867         and cast integer to void * if needed.  Update the pointer to
1868         the FP register saved area correctly.
1869
1870 2006-02-17  Andreas Tobler  <a.tobler@schweiz.ch>
1871
1872         * testsuite/libffi.call/nested_struct6.c: XFAIL this test until PR25630
1873         is fixed.
1874         * testsuite/libffi.call/nested_struct4.c: Likewise.
1875
1876 2006-02-16  Andreas Tobler  <a.tobler@schweiz.ch>
1877
1878         * testsuite/libffi.call/return_dbl.c: New test case.
1879         * testsuite/libffi.call/return_dbl1.c: Likewise.
1880         * testsuite/libffi.call/return_dbl2.c: Likewise.
1881         * testsuite/libffi.call/return_fl.c: Likewise.
1882         * testsuite/libffi.call/return_fl1.c: Likewise.
1883         * testsuite/libffi.call/return_fl2.c: Likewise.
1884         * testsuite/libffi.call/return_fl3.c: Likewise.
1885         * testsuite/libffi.call/closure_fn6.c: Likewise.
1886
1887         * testsuite/libffi.call/nested_struct2.c: Remove ffi_type_mylong
1888         definition.
1889         * testsuite/libffi.call/ffitest.h: Add ffi_type_mylong definition
1890         here to be used by other test cases too.
1891
1892         * testsuite/libffi.call/nested_struct10.c: New test case.
1893         * testsuite/libffi.call/nested_struct9.c: Likewise.
1894         * testsuite/libffi.call/nested_struct8.c: Likewise.
1895         * testsuite/libffi.call/nested_struct7.c: Likewise.
1896         * testsuite/libffi.call/nested_struct6.c: Likewise.
1897         * testsuite/libffi.call/nested_struct5.c: Likewise.
1898         * testsuite/libffi.call/nested_struct4.c: Likewise.
1899
1900 2006-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
1901
1902         * configure.ac: Enable libffi for sparc64-*-freebsd*.
1903         * configure: Rebuilt.
1904
1905 2006-01-18  Jakub Jelinek  <jakub@redhat.com>
1906
1907         * src/powerpc/sysv.S (smst_two_register): Don't call __ashldi3,
1908         instead do the shifting inline.
1909         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't compute %r5
1910         shift count unconditionally.  Simplify load sequences for 1, 2, 3, 4
1911         and 8 byte structs, for the remaining struct sizes don't call
1912         __lshrdi3, instead do the shifting inline.
1913
1914 2005-12-07  Thiemo Seufer  <ths@networkno.de>
1915
1916         * src/mips/ffitarget.h: Remove obsolete sgidefs.h include. Add
1917         missing parentheses.
1918         * src/mips/o32.S (ffi_call_O32): Code formatting. Define
1919         and use A3_OFF, FP_OFF, RA_OFF. Micro-optimizations.
1920         (ffi_closure_O32): Likewise, but with newly defined A3_OFF2,
1921         A2_OFF2, A1_OFF2, A0_OFF2, RA_OFF2, FP_OFF2, S0_OFF2, GP_OFF2,
1922         V1_OFF2, V0_OFF2, FA_1_1_OFF2, FA_1_0_OFF2, FA_0_1_OFF2,
1923         FA_0_0_OFF2.
1924         * src/mips/ffi.c (ffi_prep_args): Code formatting. Fix
1925         endianness bugs.
1926         (ffi_prep_closure): Improve trampoline instruction scheduling.
1927         (ffi_closure_mips_inner_O32): Fix endianness bugs.
1928
1929 2005-12-03  Alan Modra  <amodra@bigpond.net.au>
1930
1931         * src/powerpc/ffi.c: Formatting.
1932         (ffi_prep_args_SYSV): Avoid possible aliasing problems by using unions.
1933         (ffi_prep_args64): Likewise.
1934
1935 2005-09-30  Geoffrey Keating  <geoffk@apple.com>
1936
1937         * testsuite/lib/libffi-dg.exp (libffi_target_compile): For
1938         darwin, use -shared-libgcc not -lgcc_s, and explain why.
1939
1940 2005-09-26  Tom Tromey  <tromey@redhat.com>
1941
1942         * testsuite/libffi.call/float1.c (value_type): New typedef.
1943         (CANARY): New define.
1944         (main): Check for result buffer overflow.
1945         * src/powerpc/linux64.S: Handle linux64 long double returns.
1946         * src/powerpc/ffi.c (FLAG_RETURNS_128BITS): New constant.
1947         (ffi_prep_cif_machdep): Handle linux64 long double returns.
1948
1949 2005-08-25  Alan Modra  <amodra@bigpond.net.au>
1950
1951         PR target/23404
1952         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Correct placement of stack
1953         homed fp args.
1954         (ffi_status ffi_prep_cif_machdep): Correct stack sizing for same.
1955
1956 2005-08-11  Jakub Jelinek  <jakub@redhat.com>
1957
1958         * configure.ac (HAVE_HIDDEN_VISIBILITY_ATTRIBUTE): New test.
1959         (AH_BOTTOM): Add FFI_HIDDEN definition.
1960         * configure: Rebuilt.
1961         * fficonfig.h.in: Rebuilt.
1962         * src/powerpc/ffi.c (hidden): Remove.
1963         (ffi_closure_LINUX64, ffi_prep_args64, ffi_call_LINUX64,
1964         ffi_closure_helper_LINUX64): Use FFI_HIDDEN instead of hidden.
1965         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64,
1966         .ffi_closure_LINUX64): Use FFI_HIDDEN instead of .hidden.
1967         * src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV): Remove,
1968         add FFI_HIDDEN to its prototype.
1969         (ffi_closure_SYSV_inner): New.
1970         * src/x86/sysv.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
1971         * src/x86/win32.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
1972
1973 2005-08-10  Alfred M. Szmidt  <ams@gnu.org>
1974
1975         PR libffi/21819:
1976         * configure: Rebuilt.
1977         * configure.ac: Handle i*86-*-gnu*.
1978
1979 2005-08-09  Jakub Jelinek  <jakub@redhat.com>
1980
1981         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Use
1982         DW_CFA_offset_extended_sf rather than
1983         DW_CFA_GNU_negative_offset_extended.
1984         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise.
1985
1986 2005-07-22  SUGIOKA Toshinobu  <sugioka@itonet.co.jp>
1987
1988         * src/sh/sysv.S (ffi_call_SYSV): Stop argument popping correctly
1989         on sh3.
1990         (ffi_closure_SYSV): Change the stack layout for sh3 struct argument.
1991         * src/sh/ffi.c (ffi_prep_args): Fix sh3 argument copy, when it is
1992         partially on register.
1993         (ffi_closure_helper_SYSV): Likewise.
1994         (ffi_prep_cif_machdep): Don't set too many cif->flags.
1995
1996 2005-07-20  Kaz Kojima  <kkojima@gcc.gnu.org>
1997
1998         * src/sh/ffi.c (ffi_call): Handle small structures correctly.
1999         Remove empty line.
2000         * src/sh64/ffi.c (simple_type): Remove.
2001         (return_type): Handle small structures correctly.
2002         (ffi_prep_args): Likewise.
2003         (ffi_call): Likewise.
2004         (ffi_closure_helper_SYSV): Likewise.
2005         * src/sh64/sysv.S (ffi_call_SYSV): Handle 1, 2 and 4-byte return.
2006         Emit position independent code if PIC and remove wrong datalabel
2007         prefixes from EH data.
2008
2009 2005-07-19  Andreas Tobler  <a.tobler@schweiz.ch>
2010
2011         * Makefile.am (nodist_libffi_la_SOURCES): Add POWERPC_FREEBSD.
2012         * Makefile.in: Regenerate.
2013         * include/Makefile.in: Likewise.
2014         * testsuite/Makefile.in: Likewise.
2015         * configure.ac: Add POWERPC_FREEBSD rules.
2016         * configure: Regenerate.
2017         * src/powerpc/ffitarget.h: Add POWERPC_FREEBSD rules.
2018         (FFI_SYSV_TYPE_SMALL_STRUCT): Define.
2019         * src/powerpc/ffi.c: Add flags to handle small structure returns
2020         in ffi_call_SYSV.
2021         (ffi_prep_cif_machdep): Handle small structures for SYSV 4 ABI.
2022         Aka FFI_SYSV.
2023         (ffi_closure_helper_SYSV): Likewise.
2024         * src/powerpc/ppc_closure.S: Add return types for small structures.
2025         * src/powerpc/sysv.S: Add bits to handle small structures for
2026         final SYSV 4 ABI.
2027
2028 2005-07-10  Andreas Tobler  <a.tobler@schweiz.ch>
2029
2030         * testsuite/libffi.call/cls_5_1_byte.c: New test file.
2031         * testsuite/libffi.call/cls_6_1_byte.c: Likewise.
2032         * testsuite/libffi.call/cls_7_1_byte.c: Likewise.
2033
2034 2005-07-05  Randolph Chung  <tausq@debian.org>
2035
2036         * src/pa/ffi.c (ffi_struct_type): Rename FFI_TYPE_SMALL_STRUCT1
2037         as FFI_TYPE_SMALL_STRUCT3.  Break out handling for 5-7 byte
2038         structures.  Kill compilation warnings.
2039         (ffi_closure_inner_LINUX): Print return values as hex in debug
2040         message.  Rename FFI_TYPE_SMALL_STRUCT1 as FFI_TYPE_SMALL_STRUCT3.
2041         Properly handle 5-7 byte structure returns.
2042         * src/pa/ffitarget.h (FFI_TYPE_SMALL_STRUCT1)
2043         (FFI_TYPE_SMALL_STRUCT2): Remove.
2044         (FFI_TYPE_SMALL_STRUCT3, FFI_TYPE_SMALL_STRUCT5)
2045         (FFI_TYPE_SMALL_STRUCT6, FFI_TYPE_SMALL_STRUCT7): Define.
2046         * src/pa/linux.S: Mark source file as using PA1.1 assembly.
2047         (checksmst1, checksmst2): Remove.
2048         (checksmst3): Optimize handling of 3-byte struct returns.
2049         (checksmst567): Properly handle 5-7 byte struct returns.
2050
2051 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2052
2053         PR libgcj/21943
2054         * src/mips/n32.S: Enforce PIC code.
2055         * src/mips/o32.S: Likewise.
2056
2057 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2058
2059         * configure.ac: Treat i*86-*-solaris2.10 and up as X86_64.
2060         * configure: Regenerate.
2061
2062 2005-06-01  Alan Modra  <amodra@bigpond.net.au>
2063
2064         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't use JUMPTARGET
2065         to call ffi_closure_helper_SYSV.  Append @local instead.
2066         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise for ffi_prep_args_SYSV.
2067
2068 2005-05-17  Kelley Cook  <kcook@gcc.gnu.org>
2069
2070         * configure.ac: Use AC_C_BIGENDIAN instead of AC_C_BIGENDIAN_CROSS.
2071         Use AC_CHECK_SIZEOF instead of AC_COMPILE_CHECK_SIZEOF.
2072         * Makefile.am (ACLOCAL_AMFLAGS): Remove -I ../config.
2073         * aclocal.m4, configure, fficonfig.h.in, Makefile.in,
2074         include/Makefile.in, testsuite/Makefile.in: Regenerate.
2075
2076 2005-05-09  Mike Stump  <mrs@apple.com>
2077
2078         * configure: Regenerate.
2079
2080 2005-05-08  Richard Henderson  <rth@redhat.com>
2081
2082         PR libffi/21285
2083         * src/alpha/osf.S: Update unwind into to match code.
2084
2085 2005-05-04  Andreas Degert <ad@papyrus-gmbh.de>
2086             Richard Henderson  <rth@redhat.com>
2087
2088         * src/x86/ffi64.c (ffi_prep_cif_machdep): Save sse-used flag in
2089         bit 11 of flags.
2090         (ffi_call): Mask return type field.  Pass ssecount to ffi_call_unix64.
2091         (ffi_prep_closure): Set carry bit if sse-used flag set.
2092         * src/x86/unix64.S (ffi_call_unix64): Add ssecount argument.
2093         Only load sse registers if ssecount non-zero.
2094         (ffi_closure_unix64): Only save sse registers if carry set on entry.
2095
2096 2005-04-29  Ralf Corsepius  <ralf.corsepius@rtems.org>
2097
2098         * configure.ac: Add i*86-*-rtems*, sparc*-*-rtems*,
2099         powerpc-*rtems*, arm*-*-rtems*, sh-*-rtems*.
2100         * configure: Regenerate.
2101
2102 2005-04-20  Hans-Peter Nilsson  <hp@axis.com>
2103
2104         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): In regsub use,
2105         have Tcl8.3-compatible intermediate variable.
2106
2107 2005-04-18  Simon Posnjak <simon.posnjak@siol.net>
2108             Hans-Peter Nilsson  <hp@axis.com>
2109
2110         * Makefile.am: Add CRIS support.
2111         * configure.ac: Likewise.
2112         * Makefile.in, configure, testsuite/Makefile.in,
2113         include/Makefile.in: Regenerate.
2114         * src/cris: New directory.
2115         * src/cris/ffi.c, src/cris/sysv.S, src/cris/ffitarget.h: New files.
2116         * src/prep_cif.c (ffi_prep_cif): Wrap in #ifndef __CRIS__.
2117
2118         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): Replace \n with
2119         \r?\n in output tests.
2120
2121 2005-04-12  Mike Stump  <mrs@apple.com>
2122
2123         * configure: Regenerate.
2124
2125 2005-03-30  Hans Boehm  <Hans.Boehm@hp.com>
2126
2127         * src/ia64/ffitarget.h (ffi_arg): Use long long instead of DI.
2128
2129 2005-03-30  Steve Ellcey  <sje@cup.hp.com>
2130
2131         * src/ia64/ffitarget.h (ffi_arg) ADD DI attribute.
2132         (ffi_sarg) Ditto.
2133         * src/ia64/unix.S (ffi_closure_unix): Extend gp
2134         to 64 bits in ILP32 mode.
2135         Load 64 bits even for short data.
2136
2137 2005-03-23  Mike Stump  <mrs@apple.com>
2138
2139         * src/powerpc/darwin.S: Update for -m64 multilib.
2140         * src/powerpc/darwin_closure.S: Likewise.
2141
2142 2005-03-21  Zack Weinberg  <zack@codesourcery.com>
2143
2144         * configure.ac: Do not invoke TL_AC_GCC_VERSION.
2145         Do not set tool_include_dir.
2146         * aclocal.m4, configure, Makefile.in, testsuite/Makefile.in:
2147         Regenerate.
2148         * include/Makefile.am: Set gcc_version and toollibffidir.
2149         * include/Makefile.in: Regenerate.
2150
2151 2005-02-22  Andrew Haley  <aph@redhat.com>
2152
2153         * src/powerpc/ffi.c (ffi_prep_cif_machdep): Bump alignment to
2154         odd-numbered register pairs for 64-bit integer types.
2155
2156 2005-02-23  Andreas Tobler  <a.tobler@schweiz.ch>
2157
2158         PR libffi/20104
2159         * testsuite/libffi.call/return_ll1.c: New test case.
2160
2161 2005-02-11  Janis Johnson  <janis187@us.ibm.com>
2162
2163         * testsuite/libffi.call/cls_align_longdouble.c: Remove dg-options.
2164         * testsuite/libffi.call/float.c: Ditto.
2165         * testsuite/libffi.call/float2.c: Ditto.
2166         * testsuite/libffi.call/float3.c: Ditto.
2167
2168 2005-02-08  Andreas Tobler  <a.tobler@schweiz.ch>
2169
2170         * src/frv/ffitarget.h: Remove PPC stuff which does not belong to frv.
2171
2172 2005-01-12  Eric Botcazou  <ebotcazou@libertysurf.fr>
2173
2174         * testsuite/libffi.special/special.exp (cxx_options): Add
2175         -shared-libgcc.
2176
2177 2004-12-31  Richard Henderson  <rth@redhat.com>
2178
2179         * src/types.c (FFI_AGGREGATE_TYPEDEF): Remove.
2180         (FFI_TYPEDEF): Rename from FFI_INTEGRAL_TYPEDEF.  Replace size and
2181         offset parameters with a type parameter; deduce size and structure
2182         alignment.  Update all users.
2183
2184 2004-12-31  Richard Henderson  <rth@redhat.com>
2185
2186         * src/types.c (FFI_TYPE_POINTER): Define with sizeof.
2187         (FFI_TYPE_LONGDOUBLE): Fix for ia64.
2188         * src/ia64/ffitarget.h (struct ffi_ia64_trampoline_struct): Move
2189         into ffi_prep_closure.
2190         * src/ia64/ia64_flags.h, src/ia64/ffi.c, src/ia64/unix.S: Rewrite
2191         from scratch.
2192
2193 2004-12-27  Richard Henderson  <rth@redhat.com>
2194
2195         * src/x86/unix64.S: Fix typo in unwind info.
2196
2197 2004-12-25  Richard Henderson  <rth@redhat.com>
2198
2199         * src/x86/ffi64.c (struct register_args): Rename from stackLayout.
2200         (enum x86_64_reg_class): Add X86_64_COMPLEX_X87_CLASS.
2201         (merge_classes): Check for it.
2202         (SSE_CLASS_P): New.
2203         (classify_argument): Pass byte_offset by value; perform all updates
2204         inside struct case.
2205         (examine_argument): Add classes argument; handle
2206         X86_64_COMPLEX_X87_CLASS.
2207         (ffi_prep_args): Merge into ...
2208         (ffi_call): ... here.  Share stack frame with ffi_call_unix64.
2209         (ffi_prep_cif_machdep): Setup cif->flags for proper structure return.
2210         (ffi_fill_return_value): Remove.
2211         (ffi_prep_closure): Remove dead assert.
2212         (ffi_closure_unix64_inner): Rename from ffi_closure_UNIX64_inner.
2213         Rewrite to use struct register_args instead of va_list.  Create
2214         flags for handling structure returns.
2215         * src/x86/unix64.S: Remove dead strings.
2216         (ffi_call_unix64): Rename from ffi_call_UNIX64.  Rewrite to share
2217         stack frame with ffi_call.  Handle structure returns properly.
2218         (float2sse, floatfloat2sse, double2sse): Remove.
2219         (sse2float, sse2double, sse2floatfloat): Remove.
2220         (ffi_closure_unix64): Rename from ffi_closure_UNIX64.  Rewrite
2221         to handle structure returns properly.
2222
2223 2004-12-08  David Edelsohn  <edelsohn@gnu.org>
2224
2225         * Makefile.am (AM_MAKEFLAGS): Remove duplicate LIBCFLAGS and
2226         PICFLAG.
2227         * Makefile.in: Regenerated.
2228
2229 2004-12-02  Richard Sandiford  <rsandifo@redhat.com>
2230
2231         * configure.ac: Use TL_AC_GCC_VERSION to set gcc_version.
2232         * configure, aclocal.m4, Makefile.in: Regenerate.
2233         * include/Makefile.in, testsuite/Makefile.in: Regenerate.
2234
2235 2004-11-29  Kelley Cook  <kcook@gcc.gnu.org>
2236
2237         * configure: Regenerate for libtool change.
2238
2239 2004-11-25  Kelley Cook  <kcook@gcc.gnu.org>
2240
2241         * configure: Regenerate for libtool reversion.
2242
2243 2004-11-24  Kelley Cook  <kcook@gcc.gnu.org>
2244
2245         * configure: Regenerate for libtool change.
2246
2247 2004-11-23  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2248
2249         * testsuite/lib/libffi-dg.exp: Use new procs in target-libpath.exp.
2250
2251 2004-11-23  Richard Sandiford  <rsandifo@redhat.com>
2252
2253         * src/mips/o32.S (ffi_call_O32, ffi_closure_O32): Use jalr instead
2254         of jal.  Use an absolute encoding for the frame information.
2255
2256 2004-11-23  Kelley Cook  <kcook@gcc.gnu.org>
2257
2258         * Makefile.am: Remove no-dependencies.  Add ACLOCAL_AMFLAGS.
2259         * acinclude.m4: Delete logic for sincludes.
2260         * aclocal.m4, Makefile.in, configure: Regenerate.
2261         * include/Makefile: Likewise.
2262         * testsuite/Makefile: Likewise.
2263
2264 2004-11-22  Eric Botcazou  <ebotcazou@libertysurf.fr>
2265
2266         * src/sparc/ffi.c (ffi_prep_closure): Align doubles and 64-bit integers
2267         on a 8-byte boundary.
2268         * src/sparc/v8.S (ffi_closure_v8): Reserve frame space for arguments.
2269
2270 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2271
2272         * src/arm/ffi.c (ffi_prep_cif_machdep): Handle functions that return
2273         long long values.  Round stack allocation to a multiple of 8 bytes
2274         for ATPCS compatibility.
2275         * src/arm/sysv.S (ffi_call_SYSV): Rework to avoid use of APCS register
2276         names.  Handle returning long long types.  Add Thumb and interworking
2277         support.  Improve soft-float code.
2278
2279 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2280
2281         * testsuite/lib/libffi-db.exp (load_gcc_lib): New function.
2282         (libffi_exit): New function.
2283         (libffi_init): Build the testglue wrapper if needed.
2284
2285 2004-10-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2286
2287         PR other/18138
2288         * testsuite/lib/libffi-dg.exp: Accept more than one multilib libgcc.
2289
2290 2004-10-25  Kazuhiro Inaoka <inaoka.kazuhiro@renesas.com>
2291
2292         * src/m32r/libffitarget.h (FFI_CLOSURES): Set to 0.
2293
2294 2004-10-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2295
2296         * src/sh/sysv.S (ffi_call_SYSV): Don't align for double data.
2297         * testsuite/libffi.call/float3.c: New test case.
2298
2299 2004-10-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2300
2301         * src/sh/ffi.c (ffi_prep_closure): Set T bit in trampoline for
2302         the function returning a structure pointed with R2.
2303         * src/sh/sysv.S (ffi_closure_SYSV): Use R2 as the pointer to
2304         the structure return value if T bit set.  Emit position
2305         independent code and EH data if PIC.
2306
2307 2004-10-13  Kazuhiro Inaoka  <inaoka.kazuhiro@renesas.com>
2308
2309         * Makefile.am: Add m32r support.
2310         * configure.ac: Likewise.
2311         * Makefile.in: Regenerate.
2312         * confiugre: Regenerate.
2313         * src/types.c: Add m32r port to FFI_INTERNAL_TYPEDEF
2314         (uint64, sint64, double, longdouble)
2315         * src/m32r: New directory.
2316         * src/m32r/ffi.c: New file.
2317         * src/m32r/sysv.S: Likewise.
2318         * src/m32r/ffitarget.h: Likewise.
2319
2320 2004-10-02  Kaz Kojima  <kkojima@gcc.gnu.org>
2321
2322         * testsuite/libffi.call/negint.c: New test case.
2323
2324 2004-09-14  H.J. Lu  <hongjiu.lu@intel.com>
2325
2326         PR libgcj/17465
2327         * testsuite/lib/libffi-dg.exp: Don't use global ld_library_path.
2328         Set up LD_LIBRARY_PATH, SHLIB_PATH, LD_LIBRARYN32_PATH,
2329         LD_LIBRARY64_PATH, LD_LIBRARY_PATH_32, LD_LIBRARY_PATH_64 and
2330         DYLD_LIBRARY_PATH.
2331
2332 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2333
2334         * testsuite/libffi.call/many_win32.c: Remove whitespaces.
2335         * testsuite/libffi.call/promotion.c: Likewise.
2336         * testsuite/libffi.call/return_ll.c: Remove unused var. Cleanup
2337         whitespaces.
2338         * testsuite/libffi.call/return_sc.c: Likewise.
2339         * testsuite/libffi.call/return_uc.c: Likewise.
2340
2341 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2342
2343         * src/powerpc/darwin.S: Fix comments and identation.
2344         * src/powerpc/darwin_closure.S: Likewise.
2345
2346 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2347
2348         * src/powerpc/ffi_darwin.c: Add flag for longdouble return values.
2349         (ffi_prep_args): Handle longdouble arguments.
2350         (ffi_prep_cif_machdep): Set flags for longdouble. Calculate space for
2351         longdouble.
2352         (ffi_closure_helper_DARWIN): Add closure handling for longdouble.
2353         * src/powerpc/darwin.S (_ffi_call_DARWIN): Add handling of longdouble
2354         values.
2355         * src/powerpc/darwin_closure.S (_ffi_closure_ASM): Likewise.
2356         * src/types.c: Defined longdouble size and alignment for darwin.
2357
2358 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2359
2360         * src/powerpc/aix.S: Remove whitespaces.
2361         * src/powerpc/aix_closure.S: Likewise.
2362         * src/powerpc/asm.h: Likewise.
2363         * src/powerpc/ffi.c: Likewise.
2364         * src/powerpc/ffitarget.h: Likewise.
2365         * src/powerpc/linux64.S: Likewise.
2366         * src/powerpc/linux64_closure.S: Likewise.
2367         * src/powerpc/ppc_closure.S: Likewise.
2368         * src/powerpc/sysv.S: Likewise.
2369
2370 2004-08-30  Anthony Green  <green@redhat.com>
2371
2372         * Makefile.am: Add frv support.
2373         * Makefile.in, testsuite/Makefile.in: Rebuilt.
2374         * configure.ac: Read configure.host.
2375         * configure.in: Read configure.host.
2376         * configure.host: New file.  frv-elf needs libgloss.
2377         * include/ffi.h.in: Force ffi_closure to have a nice big (8)
2378         alignment.  This is needed to frv and shouldn't harm the others.
2379         * include/ffi_common.h (ALIGN_DOWN): New macro.
2380         * src/frv/ffi.c, src/frv/ffitarget.h, src/frv/eabi.S: New files.
2381
2382 2004-08-24  David Daney  <daney@avtrex.com>
2383
2384         * testsuite/libffi.call/closure_fn0.c: Xfail mips64* instead of mips*.
2385         * testsuite/libffi.call/closure_fn1.c: Likewise.
2386         * testsuite/libffi.call/closure_fn2.c  Likewise.
2387         * testsuite/libffi.call/closure_fn3.c: Likewise.
2388         * testsuite/libffi.call/closure_fn4.c: Likewise.
2389         * testsuite/libffi.call/closure_fn5.c: Likewise.
2390         * testsuite/libffi.call/cls_18byte.c: Likewise.
2391         * testsuite/libffi.call/cls_19byte.c: Likewise.
2392         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2393         * testsuite/libffi.call/cls_20byte.c: Likewise.
2394         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2395         * testsuite/libffi.call/cls_24byte.c: Likewise.
2396         * testsuite/libffi.call/cls_2byte.c: Likewise.
2397         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2398         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2399         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2400         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2401         * testsuite/libffi.call/cls_4byte.c: Likewise.
2402         * testsuite/libffi.call/cls_64byte.c: Likewise.
2403         * testsuite/libffi.call/cls_6byte.c: Likewise.
2404         * testsuite/libffi.call/cls_7byte.c: Likewise.
2405         * testsuite/libffi.call/cls_8byte.c: Likewise.
2406         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2407         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2408         * testsuite/libffi.call/cls_align_double.c: Likewise.
2409         * testsuite/libffi.call/cls_align_float.c: Likewise.
2410         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2411         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2412         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2413         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2414         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2415         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2416         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2417         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
2418         * testsuite/libffi.call/cls_double.c: Likewise.
2419         * testsuite/libffi.call/cls_float.c: Likewise.
2420         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2421         * testsuite/libffi.call/cls_multi_sshort.c: Likewise.
2422         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2423         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2424         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2425         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2426         * testsuite/libffi.call/cls_schar.c: Likewise.
2427         * testsuite/libffi.call/cls_sint.c: Likewise.
2428         * testsuite/libffi.call/cls_sshort.c: Likewise.
2429         * testsuite/libffi.call/cls_uchar.c: Likewise.
2430         * testsuite/libffi.call/cls_uint.c: Likewise.
2431         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2432         * testsuite/libffi.call/cls_ushort.c: Likewise.
2433         * testsuite/libffi.call/nested_struct.c: Likewise.
2434         * testsuite/libffi.call/nested_struct1.c: Likewise.
2435         * testsuite/libffi.call/nested_struct2.c: Likewise.
2436         * testsuite/libffi.call/nested_struct3.c: Likewise.
2437         * testsuite/libffi.call/problem1.c: Likewise.
2438         * testsuite/libffi.special/unwindtest.cc: Likewise.
2439         * testsuite/libffi.call/cls_12byte.c: Likewise and set return value
2440         to zero.
2441         * testsuite/libffi.call/cls_16byte.c: Likewise.
2442         * testsuite/libffi.call/cls_5byte.c: Likewise.
2443
2444 2004-08-23  David Daney <daney@avtrex.com>
2445
2446         PR libgcj/13141
2447         * src/mips/ffitarget.h (FFI_O32_SOFT_FLOAT): New ABI.
2448         * src/mips/ffi.c (ffi_prep_args): Fix alignment calculation.
2449         (ffi_prep_cif_machdep): Handle FFI_O32_SOFT_FLOAT floating point
2450         parameters and return types.
2451         (ffi_call): Handle FFI_O32_SOFT_FLOAT ABI.
2452         (ffi_prep_closure): Ditto.
2453         (ffi_closure_mips_inner_O32): Handle FFI_O32_SOFT_FLOAT ABI, fix
2454         alignment calculations.
2455         * src/mips/o32.S (ffi_closure_O32): Don't use floating point
2456         instructions if FFI_O32_SOFT_FLOAT, make stack frame ABI compliant.
2457
2458 2004-08-14  Casey Marshall <csm@gnu.org>
2459
2460         * src/mips/ffi.c (ffi_pref_cif_machdep): set `cif->flags' to
2461         contain `FFI_TYPE_UINT64' as return type for any 64-bit
2462         integer (O32 ABI only).
2463         (ffi_prep_closure): new function.
2464         (ffi_closure_mips_inner_O32): new function.
2465         * src/mips/ffitarget.h: Define `FFI_CLOSURES' and
2466         `FFI_TRAMPOLINE_SIZE' appropriately if the ABI is o32.
2467         * src/mips/o32.S (ffi_call_O32): add labels for .eh_frame. Return
2468         64 bit integers correctly.
2469         (ffi_closure_O32): new function.
2470         Added DWARF-2 unwind info for both functions.
2471
2472 2004-08-10  Andrew Haley  <aph@redhat.com>
2473
2474         * src/x86/ffi64.c (ffi_prep_args ): 8-align all stack arguments.
2475
2476 2004-08-01  Robert Millan  <robertmh@gnu.org>
2477
2478         * configure.ac: Detect knetbsd-gnu and kfreebsd-gnu.
2479         * configure: Regenerate.
2480
2481 2004-07-30  Maciej W. Rozycki  <macro@linux-mips.org>
2482
2483         * acinclude.m4 (AC_FUNC_MMAP_BLACKLIST): Check for <sys/mman.h>
2484         and mmap() explicitly instead of relying on preset autoconf cache
2485         variables.
2486         * aclocal.m4: Regenerate.
2487         * configure: Regenerate.
2488
2489 2004-07-11  Ulrich Weigand  <uweigand@de.ibm.com>
2490
2491         * src/s390/ffi.c (ffi_prep_args): Fix C aliasing violation.
2492         (ffi_check_float_struct): Remove unused prototype.
2493
2494 2004-06-30  Geoffrey Keating  <geoffk@apple.com>
2495
2496         * src/powerpc/ffi_darwin.c (flush_icache): ';' is a comment
2497         character on Darwin, use '\n\t' instead.
2498
2499 2004-06-26  Matthias Klose  <doko@debian.org>
2500
2501         * libtool-version: Fix typo in revision/age.
2502
2503 2004-06-17  Matthias Klose  <doko@debian.org>
2504
2505         * libtool-version: New.
2506         * Makefile.am (libffi_la_LDFLAGS): Use -version-info for soname.
2507         * Makefile.in: Regenerate.
2508
2509 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2510
2511         * Makefile.am: Remove useless multilib rules.
2512         * Makefile.in: Regenerate.
2513         * aclocal.m4: Regenerate with automake 1.8.5.
2514         * configure.ac: Remove useless multilib configury.
2515         * configure: Regenerate.
2516
2517 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2518
2519         * .cvsignore: New file.
2520
2521 2004-06-10  Jakub Jelinek  <jakub@redhat.com>
2522
2523         * src/ia64/unix.S (ffi_call_unix): Insert group barrier break
2524         fp_done.
2525         (ffi_closure_UNIX): Fix f14/f15 adjustment if FLOAT_SZ is ever
2526         changed from 8.
2527
2528 2004-06-06  Sean McNeil  <sean@mcneil.com>
2529
2530         * configure.ac: Add x86_64-*-freebsd* support.
2531         * configure: Regenerate.
2532
2533 2004-04-26  Joe Buck <jbuck@welsh-buck.org>
2534
2535         Bug 15093
2536         * configure.ac: Test for existence of mmap and sys/mman.h before
2537         checking blacklist.  Fix suggested by Jim Wilson.
2538         * configure: Regenerate.
2539
2540 2004-04-26  Matt Austern  <austern@apple.com>
2541
2542         * src/powerpc/darwin.S: Go through a non-lazy pointer for initial
2543         FDE location.
2544         * src/powerpc/darwin_closure.S: Likewise.
2545
2546 2004-04-24  Andreas Tobler  <a.tobler@schweiz.ch>
2547
2548         * testsuite/libffi.call/cls_multi_schar.c (main): Fix initialization
2549         error. Reported by Thomas Heller <theller@python.net>.
2550         * testsuite/libffi.call/cls_multi_sshort.c (main): Likewise.
2551         * testsuite/libffi.call/cls_multi_ushort.c (main): Likewise.
2552
2553 2004-03-20  Matthias Klose  <doko@debian.org>
2554
2555         * src/pa/linux.S: Fix typo.
2556
2557 2004-03-19  Matthias Klose  <doko@debian.org>
2558
2559         * Makefile.am: Update.
2560         * Makefile.in: Regenerate.
2561         * src/pa/ffi.h.in: Remove.
2562         * src/pa/ffitarget.h: New file.
2563
2564 2004-02-10  Randolph Chung  <tausq@debian.org>
2565
2566         * Makefile.am: Add PA support.
2567         * Makefile.in: Regenerate.
2568         * include/Makefile.in: Regenerate.
2569         * configure.ac: Add PA target.
2570         * configure: Regenerate.
2571         * src/pa/ffi.c: New file.
2572         * src/pa/ffi.h.in: Add PA support.
2573         * src/pa/linux.S: New file.
2574         * prep_cif.c: Add PA support.
2575
2576 2004-03-16  Hosaka Yuji  <hos@tamanegi.org>
2577
2578         * src/types.c: Fix alignment size of X86_WIN32 case int64 and
2579         double.
2580         * src/x86/ffi.c (ffi_prep_args): Replace ecif->cif->rtype->type
2581         with ecif->cif->flags.
2582         (ffi_call, ffi_prep_incoming_args_SYSV): Replace cif->rtype->type
2583         with cif->flags.
2584         (ffi_prep_cif_machdep): Add X86_WIN32 struct case.
2585         (ffi_closure_SYSV): Add 1 or 2-bytes struct case for X86_WIN32.
2586         * src/x86/win32.S (retstruct1b, retstruct2b, sc_retstruct1b,
2587         sc_retstruct2b): Add for 1 or 2-bytes struct case.
2588
2589 2004-03-15 Kelley Cook <kcook@gcc.gnu.org>
2590
2591         * configure.in: Rename file to ...
2592         * configure.ac: ... this.
2593         * fficonfig.h.in: Regenerate.
2594         * Makefile.in: Regenerate.
2595         * include/Makefile.in: Regenerate.
2596         * testsuite/Makefile.in: Regenerate.
2597
2598 2004-03-12  Matt Austern  <austern@apple.com>
2599
2600         * src/powerpc/darwin.S: Fix EH information so it corresponds to
2601         changes in EH format resulting from addition of linkonce support.
2602         * src/powerpc/darwin_closure.S: Likewise.
2603
2604 2004-03-11  Andreas Tobler  <a.tobler@schweiz.ch>
2605             Paolo Bonzini  <bonzini@gnu.org>
2606
2607         * Makefile.am (AUTOMAKE_OPTIONS): Set them.
2608         Remove VPATH. Remove rules for object files. Remove multilib support.
2609         (AM_CCASFLAGS): Add.
2610         * configure.in (AC_CONFIG_HEADERS): Relace AM_CONFIG_HEADER.
2611         (AC_PREREQ): Bump version to 2.59.
2612         (AC_INIT): Fill with version info and bug address.
2613         (ORIGINAL_LD_FOR_MULTILIBS): Remove.
2614         (AM_ENABLE_MULTILIB): Use this instead of AC_ARG_ENABLE.
2615         De-precious CC so that the right flags are passed down to multilibs.
2616         (AC_MSG_ERROR): Replace obsolete macro AC_ERROR.
2617         (AC_CONFIG_FILES): Replace obsolete macro AC_LINK_FILES.
2618         (AC_OUTPUT): Reorganize the output with AC_CONFIG_COMMANDS.
2619         * configure: Rebuilt.
2620         * aclocal.m4: Likewise.
2621         * Makefile.in, include/Makefile.in, testsuite/Makefile.in: Likewise.
2622         * fficonfig.h.in: Likewise.
2623
2624 2004-03-11  Andreas Schwab  <schwab@suse.de>
2625
2626         * src/ia64/ffi.c (ffi_prep_incoming_args_UNIX): Get floating point
2627         arguments from fp registers only for the first 8 parameter slots.
2628         Don't convert a float parameter when passed in memory.
2629
2630 2004-03-09  Hans-Peter Nilsson  <hp@axis.com>
2631
2632         * configure: Regenerate for config/accross.m4 correction.
2633
2634 2004-02-25  Matt Kraai  <kraai@alumni.cmu.edu>
2635
2636         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Change
2637         ecif->cif->bytes to bytes.
2638         (ffi_prep_cif_machdep): Add braces around nested if statement.
2639
2640 2004-02-09  Alan Modra  <amodra@bigpond.net.au>
2641
2642         * src/types.c (pointer): POWERPC64 has 8 byte pointers.
2643
2644         * src/powerpc/ffi.c (ffi_prep_args64): Correct long double handling.
2645         (ffi_closure_helper_LINUX64): Fix typo.
2646         * testsuite/libffi.call/cls_align_longdouble.c: Pass -mlong-double-128
2647         for powerpc64-*-*.
2648         * testsuite/libffi.call/float.c: Likewise.
2649         * testsuite/libffi.call/float2.c: Likewise.
2650
2651 2004-02-08  Alan Modra  <amodra@bigpond.net.au>
2652
2653         * src/powerpc/ffi.c (ffi_prep_cif_machdep <FFI_LINUX64>): Correct
2654         long double function return and long double arg handling.
2655         (ffi_closure_helper_LINUX64): Formatting.  Delete unused "ng" var.
2656         Use "end_pfr" instead of "nf".  Correct long double handling.
2657         Localise "temp".
2658         * src/powerpc/linux64.S (ffi_call_LINUX64): Save f2 long double
2659         return value.
2660         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Allocate
2661         space for long double return value.  Adjust stack frame and offsets.
2662         Load f2 long double return.
2663
2664 2004-02-07  Alan Modra  <amodra@bigpond.net.au>
2665
2666         * src/types.c: Use 16 byte long double for POWERPC64.
2667
2668 2004-01-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2669
2670         * src/sparc/ffi.c (ffi_prep_args_v9): Shift the parameter array
2671         when the structure return address is passed in %o0.
2672         (ffi_V9_return_struct): Rename into ffi_v9_layout_struct.
2673         (ffi_v9_layout_struct): Align the field following a nested structure
2674         on a word boundary.  Use memmove instead of memcpy.
2675         (ffi_call): Update call to ffi_V9_return_struct.
2676         (ffi_prep_closure): Define 'ctx' only for V8.
2677         (ffi_closure_sparc_inner): Clone into ffi_closure_sparc_inner_v8
2678         and ffi_closure_sparc_inner_v9.
2679         (ffi_closure_sparc_inner_v8): Return long doubles by reference.
2680         Always skip the structure return address.  For structures and long
2681         doubles, copy the argument directly.
2682         (ffi_closure_sparc_inner_v9): Skip the structure return address only
2683         if required.  Shift the maximum floating-point slot accordingly.  For
2684         big structures, copy the argument directly; otherwise, left-justify the
2685         argument and call ffi_v9_layout_struct to lay out the structure on
2686         the stack.
2687         * src/sparc/v8.S: Undef STACKFRAME before defining it.
2688         (ffi_closure_v8): Pass the structure return address.  Update call to
2689         ffi_closure_sparc_inner_v8.  Short-circuit FFI_TYPE_INT handling.
2690         Skip the 'unimp' insn when returning long doubles and structures.
2691         * src/sparc/v9.S: Undef STACKFRAME before defining it.
2692         (ffi_closure_v9): Increase the frame size by 2 words.  Short-circuit
2693         FFI_TYPE_INT handling.  Load structures both in integers and
2694         floating-point registers on return.
2695         * README: Update status of the SPARC port.
2696
2697 2004-01-24  Andreas Tobler  <a.tobler@schweiz.ch>
2698
2699         * testsuite/libffi.call/pyobjc-tc.c (main): Treat result value
2700         as of type ffi_arg.
2701         * testsuite/libffi.call/struct3.c (main): Fix CHECK.
2702
2703 2004-01-22  Ulrich Weigand  <uweigand@de.ibm.com>
2704
2705         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Treat result
2706         value as of type ffi_arg, not unsigned int.
2707
2708 2004-01-21  Michael Ritzert  <ritzert@t-online.de>
2709
2710         * ffi64.c (ffi_prep_args): Cast the RHS of an assignment instead
2711         of the LHS.
2712
2713 2004-01-12  Andreas Tobler  <a.tobler@schweiz.ch>
2714
2715         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_32 for
2716         Solaris.
2717
2718 2004-01-08  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2719
2720         * testsuite/libffi.call/ffitest.h (allocate_mmap): Cast MAP_FAILED
2721         to void *.
2722
2723 2003-12-10  Richard Henderson  <rth@redhat.com>
2724
2725         * testsuite/libffi.call/cls_align_pointer.c: Cast pointers to
2726         size_t instead of int.
2727
2728 2003-12-04  Hosaka Yuji  <hos@tamanegi.org>
2729
2730         * testsuite/libffi.call/many_win32.c: Include <float.h>.
2731         * testsuite/libffi.call/many_win32.c (main): Replace variable
2732         int i with unsigned long ul.
2733
2734         * testsuite/libffi.call/cls_align_uint64.c: New test case.
2735         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2736         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2737         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2738         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2739         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2740         * testsuite/libffi.call/cls_align_float.c: Likewise.
2741         * testsuite/libffi.call/cls_align_double.c: Likewise.
2742         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2743         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2744
2745 2003-12-02  Hosaka Yuji  <hos@tamanegi.org>
2746
2747         PR other/13221
2748         * src/x86/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV):
2749         Align arguments to 32 bits.
2750
2751 2003-12-01  Andreas Tobler  <a.tobler@schweiz.ch>
2752
2753         PR other/13221
2754         * testsuite/libffi.call/cls_multi_sshort.c: New test case.
2755         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2756         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2757         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2758         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2759         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2760
2761         * testsuite/libffi.special/unwindtest.cc: Cosmetics.
2762
2763 2003-11-26  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
2764
2765         * testsuite/libffi.call/ffitest.h: Include <fcntl.h>.
2766         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2767
2768 2003-11-22  Andreas Tobler  <a.tobler@schweiz.ch>
2769
2770         * Makefile.in: Rebuilt.
2771         * configure: Likewise.
2772         * testsuite/libffi.special/unwindtest.cc: Convert the mmap to
2773         the right type.
2774
2775 2003-11-21  Andreas Jaeger  <aj@suse.de>
2776             Andreas Tobler  <a.tobler@schweiz.ch>
2777
2778         * acinclude.m4: Add AC_FUNC_MMAP_BLACKLIST.
2779         * configure.in: Call AC_FUNC_MMAP_BLACKLIST.
2780         * Makefile.in: Rebuilt.
2781         * aclocal.m4: Likewise.
2782         * configure: Likewise.
2783         * fficonfig.h.in: Likewise.
2784         * testsuite/lib/libffi-dg.exp: Add include dir.
2785         * testsuite/libffi.call/ffitest.h: Add MMAP definitions.
2786         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2787         * testsuite/libffi.call/closure_fn0.c: Use MMAP functionality
2788         for ffi_closure if available.
2789         * testsuite/libffi.call/closure_fn1.c: Likewise.
2790         * testsuite/libffi.call/closure_fn2.c: Likewise.
2791         * testsuite/libffi.call/closure_fn3.c: Likewise.
2792         * testsuite/libffi.call/closure_fn4.c: Likewise.
2793         * testsuite/libffi.call/closure_fn5.c: Likewise.
2794         * testsuite/libffi.call/cls_12byte.c: Likewise.
2795         * testsuite/libffi.call/cls_16byte.c: Likewise.
2796         * testsuite/libffi.call/cls_18byte.c: Likewise.
2797         * testsuite/libffi.call/cls_19byte.c: Likewise.
2798         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2799         * testsuite/libffi.call/cls_20byte.c: Likewise.
2800         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2801         * testsuite/libffi.call/cls_24byte.c: Likewise.
2802         * testsuite/libffi.call/cls_2byte.c: Likewise.
2803         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2804         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2805         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2806         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2807         * testsuite/libffi.call/cls_4byte.c: Likewise.
2808         * testsuite/libffi.call/cls_5byte.c: Likewise.
2809         * testsuite/libffi.call/cls_64byte.c: Likewise.
2810         * testsuite/libffi.call/cls_6byte.c: Likewise.
2811         * testsuite/libffi.call/cls_7byte.c: Likewise.
2812         * testsuite/libffi.call/cls_8byte.c: Likewise.
2813         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2814         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2815         * testsuite/libffi.call/cls_double.c: Likewise.
2816         * testsuite/libffi.call/cls_float.c: Likewise.
2817         * testsuite/libffi.call/cls_schar.c: Likewise.
2818         * testsuite/libffi.call/cls_sint.c: Likewise.
2819         * testsuite/libffi.call/cls_sshort.c: Likewise.
2820         * testsuite/libffi.call/cls_uchar.c: Likewise.
2821         * testsuite/libffi.call/cls_uint.c: Likewise.
2822         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2823         * testsuite/libffi.call/cls_ushort.c: Likewise.
2824         * testsuite/libffi.call/nested_struct.c: Likewise.
2825         * testsuite/libffi.call/nested_struct1.c: Likewise.
2826         * testsuite/libffi.call/nested_struct2.c: Likewise.
2827         * testsuite/libffi.call/nested_struct3.c: Likewise.
2828         * testsuite/libffi.call/problem1.c: Likewise.
2829         * testsuite/libffi.special/unwindtest.cc: Likewise.
2830
2831 2003-11-20  Andreas Tobler  <a.tobler@schweiz.ch>
2832
2833         * testsuite/lib/libffi-dg.exp: Make the -lgcc_s conditional.
2834
2835 2003-11-19  Andreas Tobler  <a.tobler@schweiz.ch>
2836
2837         * testsuite/lib/libffi-dg.exp: Add DYLD_LIBRARY_PATH for darwin.
2838         Add -lgcc_s to additional flags.
2839
2840 2003-11-12  Andreas Tobler  <a.tobler@schweiz.ch>
2841
2842         * configure.in, include/Makefile.am: PR libgcj/11147, install
2843         the ffitarget.h header file in a gcc versioned and target
2844         dependent place.
2845         * configure: Regenerated.
2846         * Makefile.in, include/Makefile.in: Likewise.
2847         * testsuite/Makefile.in: Likewise.
2848
2849 2003-11-09  Andreas Tobler  <a.tobler@schweiz.ch>
2850
2851         * testsuite/libffi.call/closure_fn0.c: Print result and check
2852         with dg-output to make debugging easier.
2853         * testsuite/libffi.call/closure_fn1.c: Likewise.
2854         * testsuite/libffi.call/closure_fn2.c: Likewise.
2855         * testsuite/libffi.call/closure_fn3.c: Likewise.
2856         * testsuite/libffi.call/closure_fn4.c: Likewise.
2857         * testsuite/libffi.call/closure_fn5.c: Likewise.
2858         * testsuite/libffi.call/cls_12byte.c: Likewise.
2859         * testsuite/libffi.call/cls_16byte.c: Likewise.
2860         * testsuite/libffi.call/cls_18byte.c: Likewise.
2861         * testsuite/libffi.call/cls_19byte.c: Likewise.
2862         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2863         * testsuite/libffi.call/cls_20byte.c: Likewise.
2864         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2865         * testsuite/libffi.call/cls_24byte.c: Likewise.
2866         * testsuite/libffi.call/cls_2byte.c: Likewise.
2867         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2868         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2869         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2870         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2871         * testsuite/libffi.call/cls_4byte.c: Likewise.
2872         * testsuite/libffi.call/cls_5byte.c: Likewise.
2873         * testsuite/libffi.call/cls_64byte.c: Likewise.
2874         * testsuite/libffi.call/cls_6byte.c: Likewise.
2875         * testsuite/libffi.call/cls_7byte.c: Likewise.
2876         * testsuite/libffi.call/cls_8byte.c: Likewise.
2877         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2878         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2879         * testsuite/libffi.call/cls_double.c: Likewise.
2880         * testsuite/libffi.call/cls_float.c: Likewise.
2881         * testsuite/libffi.call/cls_schar.c: Likewise.
2882         * testsuite/libffi.call/cls_sint.c: Likewise.
2883         * testsuite/libffi.call/cls_sshort.c: Likewise.
2884         * testsuite/libffi.call/cls_uchar.c: Likewise.
2885         * testsuite/libffi.call/cls_uint.c: Likewise.
2886         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2887         * testsuite/libffi.call/cls_ushort.c: Likewise.
2888         * testsuite/libffi.call/problem1.c: Likewise.
2889
2890         * testsuite/libffi.special/unwindtest.cc: Make ffi_closure
2891         static.
2892
2893 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
2894
2895         * testsuite/libffi.call/cls_9byte2.c: New test case.
2896         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2897         * testsuite/libffi.call/cls_64byte.c: Likewise.
2898         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2899         * testsuite/libffi.call/cls_19byte.c: Likewise.
2900         * testsuite/libffi.call/cls_18byte.c: Likewise.
2901         * testsuite/libffi.call/closure_fn4.c: Likewise.
2902         * testsuite/libffi.call/closure_fn5.c: Likewise.
2903         * testsuite/libffi.call/cls_schar.c: Likewise.
2904         * testsuite/libffi.call/cls_sint.c: Likewise.
2905         * testsuite/libffi.call/cls_sshort.c: Likewise.
2906         * testsuite/libffi.call/nested_struct2.c: Likewise.
2907         * testsuite/libffi.call/nested_struct3.c: Likewise.
2908
2909 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
2910
2911         * testsuite/libffi.call/cls_double.c: Do a check on the result.
2912         * testsuite/libffi.call/cls_uchar.c: Likewise.
2913         * testsuite/libffi.call/cls_uint.c: Likewise.
2914         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2915         * testsuite/libffi.call/cls_ushort.c: Likewise.
2916         * testsuite/libffi.call/return_sc.c: Cleanup whitespaces.
2917
2918 2003-11-06  Andreas Tobler  <a.tobler@schweiz.ch>
2919
2920         * src/prep_cif.c (ffi_prep_cif): Move the validity check after
2921         the initialization.
2922
2923 2003-10-23  Andreas Tobler  <a.tobler@schweiz.ch>
2924
2925         * src/java_raw_api.c (ffi_java_ptrarray_to_raw): Replace
2926         FFI_ASSERT(FALSE) with FFI_ASSERT(0).
2927
2928 2003-10-22  David Daney  <ddaney@avtrex.com>
2929
2930         * src/mips/ffitarget.h: Replace undefined UINT32 and friends with
2931         __attribute__((__mode__(__SI__))) and friends.
2932
2933 2003-10-22  Andreas Schwab  <schwab@suse.de>
2934
2935         * src/ia64/ffi.c: Replace FALSE/TRUE with false/true.
2936
2937 2003-10-21  Andreas Tobler  <a.tobler@schweiz.ch>
2938
2939         * configure.in: AC_LINK_FILES(ffitarget.h).
2940         * configure: Regenerate.
2941         * Makefile.in: Likewise.
2942         * include/Makefile.in: Likewise.
2943         * testsuite/Makefile.in: Likewise.
2944         * fficonfig.h.in: Likewise.
2945
2946 2003-10-21  Paolo Bonzini  <bonzini@gnu.org>
2947             Richard Henderson  <rth@redhat.com>
2948
2949         Avoid that ffi.h includes fficonfig.h.
2950
2951         * Makefile.am (EXTRA_DIST): Include ffitarget.h files
2952         (TARGET_SRC_MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
2953         (TARGET_SRC_MIPS_SGI): Removed.
2954         (MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
2955         (MIPS_SGI): Removed.
2956         (CLEANFILES): Removed.
2957         (mostlyclean-am, clean-am, mostlyclean-sub, clean-sub): New
2958         targets.
2959         * acconfig.h: Removed.
2960         * configure.in: Compute sizeofs only for double and long double.
2961         Use them to define and subst HAVE_LONG_DOUBLE.  Include comments
2962         into AC_DEFINE instead of using acconfig.h.  Create
2963         include/ffitarget.h instead of include/fficonfig.h.  Rename
2964         MIPS_GCC to MIPS_IRIX, drop MIPS_SGI since we are in gcc's tree.
2965         AC_DEFINE EH_FRAME_FLAGS.
2966         * include/Makefile.am (DISTCLEANFILES): New automake macro.
2967         (hack_DATA): Add ffitarget.h.
2968         * include/ffi.h.in: Remove all system specific definitions.
2969         Declare raw API even if it is not installed, why bother?
2970         Use limits.h instead of SIZEOF_* to define ffi_type_*.  Do
2971         not define EH_FRAME_FLAGS, it is in fficonfig.h now.  Include
2972         ffitarget.h instead of fficonfig.h.  Remove ALIGN macro.
2973         (UINT_ARG, INT_ARG): Removed, use ffi_arg and ffi_sarg instead.
2974         * include/ffi_common.h (bool): Do not define.
2975         (ffi_assert): Accept failed assertion.
2976         (ffi_type_test): Return void and accept file/line.
2977         (FFI_ASSERT): Pass stringized failed assertion.
2978         (FFI_ASSERT_AT): New macro.
2979         (FFI_ASSERT_VALID_TYPE): New macro.
2980         (UINT8, SINT8, UINT16, SINT16, UINT32, SINT32,
2981         UINT64, SINT64): Define here with gcc's __attribute__ macro
2982         instead of in ffi.h
2983         (FLOAT32, ALIGN): Define here instead of in ffi.h
2984         * include/ffi-mips.h: Removed.  Its content moved to
2985         src/mips/ffitarget.h after separating assembly and C sections.
2986         * src/alpha/ffi.c, src/alpha/ffi.c, src/java_raw_api.c
2987         src/prep_cif.c, src/raw_api.c, src/ia64/ffi.c,
2988         src/mips/ffi.c, src/mips/n32.S, src/mips/o32.S,
2989         src/mips/ffitarget.h, src/sparc/ffi.c, src/x86/ffi64.c:
2990         SIZEOF_ARG -> FFI_SIZEOF_ARG.
2991         * src/ia64/ffi.c: Include stdbool.h (provided by GCC 2.95+).
2992         * src/debug.c (ffi_assert): Accept stringized failed assertion.
2993         (ffi_type_test): Rewritten.
2994         * src/prep-cif.c (initialize_aggregate, ffi_prep_cif): Call
2995         FFI_ASSERT_VALID_TYPE.
2996         * src/alpha/ffitarget.h, src/arm/ffitarget.h,
2997         src/ia64/ffitarget.h, src/m68k/ffitarget.h,
2998         src/mips/ffitarget.h, src/powerpc/ffitarget.h,
2999         src/s390/ffitarget.h, src/sh/ffitarget.h,
3000         src/sh64/ffitarget.h, src/sparc/ffitarget.h,
3001         src/x86/ffitarget.h: New files.
3002         * src/alpha/osf.S, src/arm/sysv.S, src/ia64/unix.S,
3003         src/m68k/sysv.S, src/mips/n32.S, src/mips/o32.S,
3004         src/powerpc/aix.S, src/powerpc/darwin.S,
3005         src/powerpc/ffi_darwin.c, src/powerpc/linux64.S,
3006         src/powerpc/linux64_closure.S, src/powerpc/ppc_closure.S,
3007         src/powerpc/sysv.S, src/s390/sysv.S, src/sh/sysv.S,
3008         src/sh64/sysv.S, src/sparc/v8.S, src/sparc/v9.S,
3009         src/x86/sysv.S, src/x86/unix64.S, src/x86/win32.S:
3010         include fficonfig.h
3011
3012 2003-10-20  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3013
3014         * src/mips/ffi.c: Use _ABIN32, _ABIO32 instead of external
3015         _MIPS_SIM_NABI32, _MIPS_SIM_ABI32.
3016
3017 2003-10-19  Andreas Tobler  <a.tobler@schweiz.ch>
3018
3019         * src/powerpc/ffi_darwin.c (ffi_prep_args): Declare bytes again.
3020         Used when FFI_DEBUG = 1.
3021
3022 2003-10-14  Alan Modra  <amodra@bigpond.net.au>
3023
3024         * src/types.c (double, longdouble): Default POWERPC64 to 8 byte size
3025         and align.
3026
3027 2003-10-06  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3028
3029         * include/ffi_mips.h: Define FFI_MIPS_N32 for N32/N64 ABIs,
3030         FFI_MIPS_O32 for O32 ABI.
3031
3032 2003-10-01  Andreas Tobler  <a.tobler@schweiz.ch>
3033
3034         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_64 for
3035         SPARC64. Cleanup whitespaces.
3036
3037 2003-09-19  Andreas Tobler  <a.tobler@schweiz.ch>
3038
3039         * testsuite/libffi.call/closure_fn0.c: Xfail mips, arm,
3040         strongarm, xscale. Cleanup whitespaces.
3041         * testsuite/libffi.call/closure_fn1.c: Likewise.
3042         * testsuite/libffi.call/closure_fn2.c: Likewise.
3043         * testsuite/libffi.call/closure_fn3.c: Likewise.
3044         * testsuite/libffi.call/cls_12byte.c: Likewise.
3045         * testsuite/libffi.call/cls_16byte.c: Likewise.
3046         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3047         * testsuite/libffi.call/cls_20byte.c: Likewise.
3048         * testsuite/libffi.call/cls_24byte.c: Likewise.
3049         * testsuite/libffi.call/cls_2byte.c: Likewise.
3050         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3051         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3052         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3053         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3054         * testsuite/libffi.call/cls_4byte.c: Likewise.
3055         * testsuite/libffi.call/cls_5byte.c: Likewise.
3056         * testsuite/libffi.call/cls_6byte.c: Likewise.
3057         * testsuite/libffi.call/cls_7byte.c: Likewise.
3058         * testsuite/libffi.call/cls_8byte.c: Likewise.
3059         * testsuite/libffi.call/cls_double.c: Likewise.
3060         * testsuite/libffi.call/cls_float.c: Likewise.
3061         * testsuite/libffi.call/cls_uchar.c: Likewise.
3062         * testsuite/libffi.call/cls_uint.c: Likewise.
3063         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3064         * testsuite/libffi.call/cls_ushort.c: Likewise.
3065         * testsuite/libffi.call/nested_struct.c: Likewise.
3066         * testsuite/libffi.call/nested_struct1.c: Likewise.
3067         * testsuite/libffi.call/problem1.c: Likewise.
3068         * testsuite/libffi.special/unwindtest.cc: Likewise.
3069         * testsuite/libffi.call/pyobjc-tc.c: Cleanup whitespaces.
3070
3071 2003-09-18  David Edelsohn  <edelsohn@gnu.org>
3072
3073         * src/powerpc/aix.S: Cleanup whitespaces.
3074         * src/powerpc/aix_closure.S: Likewise.
3075
3076 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3077
3078         * src/powerpc/darwin.S: Cleanup whitespaces, comment formatting.
3079         * src/powerpc/darwin_closure.S: Likewise.
3080         * src/powerpc/ffi_darwin.c: Likewise.
3081
3082 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3083             David Edelsohn  <edelsohn@gnu.org>
3084
3085         * src/types.c (double): Add AIX and Darwin to the right TYPEDEF.
3086         * src/powerpc/aix_closure.S: Remove the pointer to the outgoing
3087         parameter stack.
3088         * src/powerpc/darwin_closure.S: Likewise.
3089         * src/powerpc/ffi_darwin.c (ffi_prep_args): Handle structures
3090         according to the Darwin/AIX ABI.
3091         (ffi_prep_cif_machdep): Likewise.
3092         (ffi_closure_helper_DARWIN): Likewise.
3093         Remove the outgoing parameter stack logic. Simplify the evaluation
3094         of the different CASE types.
3095         (ffi_prep_clousure): Avoid the casts on lvalues. Change the branch
3096         statement in the trampoline code.
3097
3098 2003-09-18  Kaz Kojima  <kkojima@gcc.gnu.org>
3099
3100         * src/sh/ffi.c (ffi_prep_args): Take account into the alignement
3101         for the register size.
3102         (ffi_closure_helper_SYSV): Handle the structure return value
3103         address correctly.
3104         (ffi_closure_helper_SYSV): Return the appropriate type when
3105         the registers are used for the structure return value.
3106         * src/sh/sysv.S (ffi_closure_SYSV): Fix the stack layout for
3107         the 64-bit return value.  Update copyright years.
3108
3109 2003-09-17  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3110
3111         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Search in
3112         srcdir for ffi_mips.h.
3113
3114 2003-09-12  Alan Modra  <amodra@bigpond.net.au>
3115
3116         * src/prep_cif.c (initialize_aggregate): Include tail padding in
3117         structure size.
3118         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Correct
3119         placement of float result.
3120         * testsuite/libffi.special/unwindtest.cc (closure_test_fn1): Correct
3121         cast of "resp" for big-endian 64 bit machines.
3122
3123 2003-09-11  Alan Modra  <amodra@bigpond.net.au>
3124
3125         * src/types.c (double, longdouble): Merge identical SH and ARM
3126         typedefs, and add POWERPC64.
3127         * src/powerpc/ffi.c (ffi_prep_args64): Correct next_arg calc for
3128         struct split over gpr and rest.
3129         (ffi_prep_cif_machdep): Correct intarg_count for structures.
3130         * src/powerpc/linux64.S (ffi_call_LINUX64): Fix gpr offsets.
3131
3132 2003-09-09  Andreas Tobler  <a.tobler@schweiz.ch>
3133
3134         * src/powerpc/ffi.c (ffi_closure_helper_SYSV) Handle struct
3135         passing correctly.
3136
3137 2003-09-09  Alan Modra  <amodra@bigpond.net.au>
3138
3139         * configure: Regenerate.
3140
3141 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3142
3143         * Makefile.am: Remove build rules for ffitest.
3144         * Makefile.in: Rebuilt.
3145
3146 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3147
3148         * src/java_raw_api.c: Include <stdlib.h> to fix compiler warning
3149         about implicit declaration of abort().
3150
3151 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3152
3153         * Makefile.am: Add dejagnu test framework. Fixes PR other/11411.
3154         * Makefile.in: Rebuilt.
3155         * configure.in: Add dejagnu test framework.
3156         * configure: Rebuilt.
3157
3158         * testsuite/Makefile.am: New file.
3159         * testsuite/Makefile.in: Built
3160         * testsuite/lib/libffi-dg.exp: New file.
3161         * testsuite/config/default.exp: Likewise.
3162         * testsuite/libffi.call/call.exp: Likewise.
3163         * testsuite/libffi.call/ffitest.h: Likewise.
3164         * testsuite/libffi.call/closure_fn0.c: Likewise.
3165         * testsuite/libffi.call/closure_fn1.c: Likewise.
3166         * testsuite/libffi.call/closure_fn2.c: Likewise.
3167         * testsuite/libffi.call/closure_fn3.c: Likewise.
3168         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3169         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3170         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3171         * testsuite/libffi.call/cls_2byte.c: Likewise.
3172         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3173         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3174         * testsuite/libffi.call/cls_4byte.c: Likewise.
3175         * testsuite/libffi.call/cls_5byte.c: Likewise.
3176         * testsuite/libffi.call/cls_6byte.c: Likewise.
3177         * testsuite/libffi.call/cls_7byte.c: Likewise.
3178         * testsuite/libffi.call/cls_8byte.c: Likewise.
3179         * testsuite/libffi.call/cls_12byte.c: Likewise.
3180         * testsuite/libffi.call/cls_16byte.c: Likewise.
3181         * testsuite/libffi.call/cls_20byte.c: Likewise.
3182         * testsuite/libffi.call/cls_24byte.c: Likewise.
3183         * testsuite/libffi.call/cls_double.c: Likewise.
3184         * testsuite/libffi.call/cls_float.c: Likewise.
3185         * testsuite/libffi.call/cls_uchar.c: Likewise.
3186         * testsuite/libffi.call/cls_uint.c: Likewise.
3187         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3188         * testsuite/libffi.call/cls_ushort.c: Likewise.
3189         * testsuite/libffi.call/float.c: Likewise.
3190         * testsuite/libffi.call/float1.c: Likewise.
3191         * testsuite/libffi.call/float2.c: Likewise.
3192         * testsuite/libffi.call/many.c: Likewise.
3193         * testsuite/libffi.call/many_win32.c: Likewise.
3194         * testsuite/libffi.call/nested_struct.c: Likewise.
3195         * testsuite/libffi.call/nested_struct1.c: Likewise.
3196         * testsuite/libffi.call/pyobjc-tc.c: Likewise.
3197         * testsuite/libffi.call/problem1.c: Likewise.
3198         * testsuite/libffi.call/promotion.c: Likewise.
3199         * testsuite/libffi.call/return_ll.c: Likewise.
3200         * testsuite/libffi.call/return_sc.c: Likewise.
3201         * testsuite/libffi.call/return_uc.c: Likewise.
3202         * testsuite/libffi.call/strlen.c: Likewise.
3203         * testsuite/libffi.call/strlen_win32.c: Likewise.
3204         * testsuite/libffi.call/struct1.c: Likewise.
3205         * testsuite/libffi.call/struct2.c: Likewise.
3206         * testsuite/libffi.call/struct3.c: Likewise.
3207         * testsuite/libffi.call/struct4.c: Likewise.
3208         * testsuite/libffi.call/struct5.c: Likewise.
3209         * testsuite/libffi.call/struct6.c: Likewise.
3210         * testsuite/libffi.call/struct7.c: Likewise.
3211         * testsuite/libffi.call/struct8.c: Likewise.
3212         * testsuite/libffi.call/struct9.c: Likewise.
3213         * testsuite/libffi.special/special.exp: New file.
3214         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3215         * testsuite/libffi.special/unwindtest.cc: Likewise.
3216
3217
3218 2003-08-13  Kaz Kojima  <kkojima@gcc.gnu.org>
3219
3220         * src/sh/ffi.c (OFS_INT16): Set 0 for little endian case.  Update
3221         copyright years.
3222
3223 2003-08-02  Alan Modra  <amodra@bigpond.net.au>
3224
3225         * src/powerpc/ffi.c (ffi_prep_args64): Modify for changed gcc
3226         structure passing.
3227         (ffi_closure_helper_LINUX64): Likewise.
3228         * src/powerpc/linux64.S: Remove code writing to parm save area.
3229         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Use return
3230         address in lr from ffi_closure_helper_LINUX64 call to calculate
3231         table address.  Optimize function tail.
3232
3233 2003-07-28  Andreas Tobler  <a.tobler@schweiz.ch>
3234
3235         * src/sparc/ffi.c: Handle all floating point registers.
3236         * src/sparc/v9.S: Likewise. Fixes second part of PR target/11410.
3237
3238 2003-07-11  Gerald Pfeifer  <pfeifer@dbai.tuwien.ac.at>
3239
3240         * README: Note that libffi is not part of GCC.  Update the project
3241         URL and status.
3242
3243 2003-06-19  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3244
3245         * src/powerpc/ppc_closure.S: Include ffi.h.
3246
3247 2003-06-13  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3248
3249         * src/x86/sysv.S: Avoid gas-only .uleb128/.sleb128 directives.
3250         Use C style comments.
3251
3252 2003-06-13  Kaz Kojima  <kkojima@rr.iij4u.or.jp>
3253
3254         * Makefile.am: Add SHmedia support.  Fix a typo of SH support.
3255         * Makefile.in: Regenerate.
3256         * configure.in (sh64-*-linux*, sh5*-*-linux*): Add target.
3257         * configure: Regenerate.
3258         * include/ffi.h.in: Add SHmedia support.
3259         * src/sh64/ffi.c: New file.
3260         * src/sh64/sysv.S: New file.
3261
3262 2003-05-16  Jakub Jelinek  <jakub@redhat.com>
3263
3264         * configure.in (HAVE_RO_EH_FRAME): Check whether .eh_frame section
3265         should be read-only.
3266         * configure: Rebuilt.
3267         * fficonfig.h.in: Rebuilt.
3268         * include/ffi.h.in (EH_FRAME_FLAGS): Define.
3269         * src/alpha/osf.S: Use EH_FRAME_FLAGS.
3270         * src/powerpc/linux64.S: Likewise.
3271         * src/powerpc/linux64_closure.S: Likewise.  Include ffi.h.
3272         * src/powerpc/sysv.S: Use EH_FRAME_FLAGS.  Use pcrel encoding
3273         if -fpic/-fPIC/-mrelocatable.
3274         * src/powerpc/powerpc_closure.S: Likewise.
3275         * src/sparc/v8.S: If HAVE_RO_EH_FRAME is defined, don't include
3276         #write in .eh_frame flags.
3277         * src/sparc/v9.S: Likewise.
3278         * src/x86/unix64.S: Use EH_FRAME_FLAGS.
3279         * src/x86/sysv.S: Likewise.  Use pcrel encoding if -fpic/-fPIC.
3280         * src/s390/sysv.S: Use EH_FRAME_FLAGS.  Include ffi.h.
3281
3282 2003-05-07  Jeff Sturm  <jsturm@one-point.com>
3283
3284         Fixes PR bootstrap/10656
3285         * configure.in (HAVE_AS_REGISTER_PSEUDO_OP): Test assembler
3286         support for .register pseudo-op.
3287         * src/sparc/v8.S: Use it.
3288         * fficonfig.h.in: Rebuilt.
3289         * configure: Rebuilt.
3290
3291 2003-04-18  Jakub Jelinek  <jakub@redhat.com>
3292
3293         * include/ffi.h.in (POWERPC64): Define if 64-bit.
3294         (enum ffi_abi): Add FFI_LINUX64 on POWERPC.
3295         Make it the default on POWERPC64.
3296         (FFI_TRAMPOLINE_SIZE): Define to 24 on POWERPC64.
3297         * configure.in: Change powerpc-*-linux* into powerpc*-*-linux*.
3298         * configure: Rebuilt.
3299         * src/powerpc/ffi.c (hidden): Define.
3300         (ffi_prep_args_SYSV): Renamed from
3301         ffi_prep_args.  Cast pointers to unsigned long to shut up warnings.
3302         (NUM_GPR_ARG_REGISTERS64, NUM_FPR_ARG_REGISTERS64,
3303         ASM_NEEDS_REGISTERS64): New.
3304         (ffi_prep_args64): New function.
3305         (ffi_prep_cif_machdep): Handle FFI_LINUX64 ABI.
3306         (ffi_call): Likewise.
3307         (ffi_prep_closure): Likewise.
3308         (flush_icache): Surround by #ifndef POWERPC64.
3309         (ffi_dblfl): New union type.
3310         (ffi_closure_helper_SYSV): Use it to avoid aliasing problems.
3311         (ffi_closure_helper_LINUX64): New function.
3312         * src/powerpc/ppc_closure.S: Surround whole file by #ifndef
3313         __powerpc64__.
3314         * src/powerpc/sysv.S: Likewise.
3315         (ffi_call_SYSV): Rename ffi_prep_args to ffi_prep_args_SYSV.
3316         * src/powerpc/linux64.S: New file.
3317         * src/powerpc/linux64_closure.S: New file.
3318         * Makefile.am (EXTRA_DIST): Add src/powerpc/linux64.S and
3319         src/powerpc/linux64_closure.S.
3320         (TARGET_SRC_POWERPC): Likewise.
3321
3322         * src/ffitest.c (closure_test_fn, closure_test_fn1, closure_test_fn2,
3323         closure_test_fn3): Fix result printing on big-endian 64-bit
3324         machines.
3325         (main): Print tst2_arg instead of uninitialized tst2_result.
3326
3327         * src/ffitest.c (main): Hide what closure pointer really points to
3328         from the compiler.
3329
3330 2003-04-16  Richard Earnshaw  <rearnsha@arm.com>
3331
3332         * configure.in (arm-*-netbsdelf*): Add configuration.
3333         (configure): Regenerated.
3334
3335 2003-04-04  Loren J. Rittle  <ljrittle@acm.org>
3336
3337         * include/Makefile.in: Regenerate.
3338
3339 2003-03-21  Zdenek Dvorak  <rakdver@atrey.karlin.mff.cuni.cz>
3340
3341         * libffi/include/ffi.h.in: Define X86 instead of X86_64 in 32
3342         bit mode.
3343         * libffi/src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV):
3344         Receive closure pointer through parameter, read args using
3345         __builtin_dwarf_cfa.
3346         (FFI_INIT_TRAMPOLINE): Send closure reference through eax.
3347
3348 2003-03-12  Andreas Schwab  <schwab@suse.de>
3349
3350         * configure.in: Avoid trailing /. in toolexeclibdir.
3351         * configure: Rebuilt.
3352
3353 2003-03-03  Andreas Tobler <a.tobler@schweiz.ch>
3354
3355         * src/powerpc/darwin_closure.S: Recode to fit dynamic libraries.
3356
3357 2003-02-06  Andreas Tobler <a.tobler@schweiz.ch>
3358
3359         * libffi/src/powerpc/darwin_closure.S:
3360         Fix alignement bug, allocate 8 bytes for the result.
3361         * libffi/src/powerpc/aix_closure.S:
3362         Likewise.
3363         * libffi/src/powerpc/ffi_darwin.c:
3364         Update stackframe description for aix/darwin_closure.S.
3365
3366 2003-02-06  Jakub Jelinek  <jakub@redhat.com>
3367
3368         * src/s390/ffi.c (ffi_closure_helper_SYSV): Add hidden visibility
3369         attribute.
3370
3371 2003-01-31  Christian Cornelssen  <ccorn@cs.tu-berlin.de>,
3372             Andreas Schwab  <schwab@suse.de>
3373
3374         * configure.in: Adjust command to source config-ml.in to account
3375         for changes to the libffi_basedir definition.
3376         (libffi_basedir): Remove ${srcdir} from value and include trailing
3377         slash if nonempty.
3378
3379         * configure: Regenerate.
3380
3381 2003-01-29  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3382
3383         * src/powerpc/ppc_closure.S: Recode to fit shared libs.
3384
3385 2003-01-28  Andrew Haley  <aph@redhat.com>
3386
3387         * include/ffi.h.in: Enable FFI_CLOSURES for x86_64.
3388         * src/x86/ffi64.c (ffi_prep_closure): New.
3389         (ffi_closure_UNIX64_inner): New.
3390         * src/x86/unix64.S (ffi_closure_UNIX64): New.
3391
3392 2003-01-27  Alexandre Oliva  <aoliva@redhat.com>
3393
3394         * configure.in (toolexecdir, toolexeclibdir): Set and AC_SUBST.
3395         Remove USE_LIBDIR conditional.
3396         * Makefile.am (toolexecdir, toolexeclibdir): Don't override.
3397         * Makefile.in, configure: Rebuilt.
3398
3399 2003-01027  David Edelsohn  <edelsohn@gnu.org>
3400
3401         * Makefile.am (TARGET_SRC_POWERPC_AIX): Fix typo.
3402         * Makefile.in: Regenerate.
3403
3404 2003-01-22  Andrew Haley  <aph@redhat.com>
3405
3406         * src/powerpc/darwin.S (_ffi_call_AIX): Add Augmentation size to
3407         unwind info.
3408
3409 2003-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
3410
3411         * src/powerpc/darwin.S: Add unwind info.
3412         * src/powerpc/darwin_closure.S: Likewise.
3413
3414 2003-01-14  Andrew Haley  <aph@redhat.com>
3415
3416         * src/x86/ffi64.c (ffi_prep_args): Check for void retval.
3417         (ffi_prep_cif_machdep): Likewise.
3418         * src/x86/unix64.S: Add unwind info.
3419
3420 2003-01-14  Andreas Jaeger  <aj@suse.de>
3421
3422         * src/ffitest.c (main): Only use ffi_closures if those are
3423         supported.
3424
3425 2003-01-13 Andreas Tobler <a.tobler@schweiz.ch>
3426
3427         * libffi/src/ffitest.c
3428          add closure testcases
3429
3430 2003-01-13 Kevin B. Hendricks <khendricks@ivey.uwo.ca>
3431
3432         * libffi/src/powerpc/ffi.c
3433          fix alignment bug for float (4 byte aligned iso 8 byte)
3434
3435 2003-01-09  Geoffrey Keating  <geoffk@apple.com>
3436
3437         * src/powerpc/ffi_darwin.c: Remove RCS version string.
3438         * src/powerpc/darwin.S: Remove RCS version string.
3439