OSDN Git Service

* postreload-gcse.c (insert_expr_in_table): Replace BLOCK_NUM
[pf3gnuchains/gcc-fork.git] / libffi / ChangeLog
1 2009-12-31  Anthony Green  <green@redhat.com>
2
3         * README: Update for libffi 3.0.9.
4
5 2009-12-27  Matthias Klose  <doko@ubuntu.com>
6
7         * configure.ac (HAVE_LONG_DOUBLE): Define for mips when
8         appropriate.
9         * configure: Rebuilt.
10
11 2009-12-26  Anthony Green  <green@redhat.com>
12
13         * testsuite/libffi.call/cls_longdouble_va.c: Mark as xfail for
14         avr32*-*-*.
15         * testsuite/libffi.call/cls_double_va.c: Ditto.
16
17 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
18
19         * testsuite/libffi.call/ffitest.h: Conditionally include stdint.h
20         and inttypes.h.
21         * testsuite/libffi.special/unwindtest.cc: Ditto.
22
23 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
24
25         * configure.ac: Add amd64-*-openbsd*.
26         * configure: Rebuilt.
27         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Link
28         openbsd programs with -lpthread.
29
30 2009-12-26  Anthony Green  <green@redhat.com>
31
32         * testsuite/libffi.call/cls_double_va.c,
33         testsuite/libffi.call/cls_longdouble.c,
34         testsuite/libffi.call/cls_longdouble_va.c,
35         testsuite/libffi.call/cls_pointer.c,
36         testsuite/libffi.call/cls_pointer_stack.c: Remove xfail for
37         mips*-*-* and arm*-*-*.
38         * testsuite/libffi.call/cls_align_longdouble_split.c,
39         testsuite/libffi.call/cls_align_longdouble_split2.c,
40         testsuite/libffi.call/stret_medium2.c,
41         testsuite/libffi.call/stret_medium.c,
42         testsuite/libffi.call/stret_large.c,
43         testsuite/libffi.call/stret_large2.c: Remove xfail for arm*-*-*.
44
45 2009-12-31  Kay Tietz  <ktietz70@googlemail.com>
46
47         * testsuite/libffi.call/ffitest.h,
48         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRuLL): Fix
49         definitions.
50
51 2009-12-31  Carlo Bramini  <carlo.bramix@libero.it>
52
53         * configure.ac (AM_LTLDFLAGS): Define for windows hosts.
54         * Makefile.am (libffi_la_LDFLAGS): Add AM_LTLDFLAGS.
55         * configure: Rebuilt.
56         * Makefile.in: Rebuilt.
57
58 2009-12-31  Anthony Green  <green@redhat.com>
59             Blake Chaffin.
60
61         * testsuite/libffi.call/huge_struct.c: New test case from Blake
62         Chaffin @ Apple.
63
64 2009-12-28  David Edelsohn  <edelsohn@gnu.org>
65
66         * src/powerpc/ffi_darwin.c (ffi_prep_args): Copy abi and nargs to
67         local variables.
68         (aix_adjust_aggregate_sizes): New function.
69         (ffi_prep_cif_machdep): Call it.
70
71 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
72
73         * configure.ac: Define FFI_MMAP_EXEC_WRIT for the given targets.
74         * configure: Regenerate.
75         * fficonfig.h.in: Likewise.
76         * src/closures.c: Remove the FFI_MMAP_EXEC_WRIT definition for
77         Solaris/x86.
78
79 2009-12-26  Andreas Schwab  <schwab@linux-m68k.org>
80
81         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Advance intarg_count
82         when a float arguments is passed in memory.
83         (ffi_closure_helper_SYSV): Mark general registers as used up when
84         a 64bit or soft-float long double argument is passed in memory.
85
86 2009-12-25  Matthias Klose  <doko@ubuntu.com>
87
88         * man/ffi_call.3: Fix #include in examples.
89         * doc/libffi.texi: Add dircategory.
90
91 2009-12-25  Frank Everdij <f.p.x.everdij@tudelft.nl>
92
93         * include/ffi.h.in: Placed '__GNUC__' ifdef around
94         '__attribute__((aligned(8)))' in ffi_closure, fixes compile for
95         IRIX MIPSPro c99.
96         * include/ffi_common.h: Added '__sgi' define to non
97         '__attribute__((__mode__()))' integer typedefs.
98         * src/mips/ffi.c (ffi_call, ffi_closure_mips_inner_O32,
99         ffi_closure_mips_inner_N32): Added 'defined(_MIPSEB)' to BE check.
100         (ffi_closure_mips_inner_O32, ffi_closure_mips_inner_N32): Added
101         FFI_LONGDOUBLE support and alignment(N32 only).
102         * src/mips/ffitarget.h: Corrected '#include <sgidefs.h>' for IRIX and
103         fixed non '__attribute__((__mode__()))' integer typedefs.
104         * src/mips/n32.S: Put '#ifdef linux' around '.abicalls' and '.eh_frame'
105         since they are Linux/GNU Assembler specific.
106
107 2009-12-25  Bradley Smith  <brad@brad-smith.co.uk>
108
109         * configure.ac, Makefile.am, src/avr32/ffi.c,
110         src/avr32/ffitarget.h,
111         src/avr32/sysv.S: Add AVR32 port.
112         * configure, Makefile.in: Rebuilt.
113
114 2009-12-21  Andreas Tobler  <a.tobler@schweiz.org>
115
116         * configure.ac: Make i?86 build on FreeBSD and OpenBSD.
117         * configure: Regenerate.
118
119 2009-12-15  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
120
121         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on PA HP-UX.
122
123 2009-12-13  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
124
125         * src/pa/ffi.c (ffi_closure_inner_pa32): Handle FFI_TYPE_LONGDOUBLE
126         type on HP-UX.
127
128 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
129
130         * src/sparc/ffi.c (ffi_closure_sparc_inner_v9): Properly align 'long
131         double' arguments.
132
133 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
134
135         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on Solaris < 10.
136
137 2009-12-10  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
138
139         PR libffi/40700
140         * src/closures.c [X86_64 && __sun__ && __svr4__]
141         (FFI_MMAP_EXEC_WRIT): Define.
142
143 2009-12-08  David Daney  <ddaney@caviumnetworks.com>
144
145         * testsuite/libffi.call/stret_medium.c: Remove xfail for mips*-*-*
146         * testsuite/libffi.call/cls_align_longdouble_split2.c: Same.
147         * testsuite/libffi.call/stret_large.c: Same.
148         * testsuite/libffi.call/cls_align_longdouble_split.c: Same.
149         * testsuite/libffi.call/stret_large2.c: Same.
150         * testsuite/libffi.call/stret_medium2.c: Same.
151
152 2009-12-07  David Edelsohn  <edelsohn@gnu.org>
153
154         * src/powerpc/aix_closure.S (libffi_closure_ASM): Fix tablejump
155         typo.
156
157 2009-12-05  David Edelsohn  <edelsohn@gnu.org>
158
159         * src/powerpc/aix.S: Update AIX32 code to be consistent with AIX64
160         code.
161         * src/powerpc/aix_closure.S: Same.
162
163 2009-12-05  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
164
165         * Makefile.in: Regenerate.
166         * configure: Regenerate.
167         * include/Makefile.in: Regenerate.
168         * man/Makefile.in: Regenerate.
169         * testsuite/Makefile.in: Regenerate.
170
171 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
172
173         * src/powerpc/aix_closure.S: Reorganize 64-bit code to match
174         linux64_closure.S.
175
176 2009-12-04  Uros Bizjak  <ubizjak@gmail.com>
177
178         PR libffi/41908
179         * src/x86/ffi64.c (classify_argument): Update from
180         gcc/config/i386/i386.c.
181         (ffi_closure_unix64_inner): Do not use the address of two consecutive
182         SSE registers directly.
183         * testsuite/libffi.call/cls_dbls_struct.c (main): Remove xfail
184         for x86_64 linux targets.
185
186 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
187
188         * src/powerpc/ffi_darwin.c (ffi_closure_helper_DARWIN): Increment
189         pfr for long double split between fpr13 and stack.
190
191 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
192
193         * src/powerpc/ffi_darwin.c (ffi_prep_args): Increment next_arg and
194         fparg_count twice for long double.
195
196 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
197
198         PR libffi/42243
199         * src/powerpc/ffi_darwin.c (ffi_prep_args): Remove extra parentheses.
200
201 2009-12-03  Uros Bizjak  <ubizjak@gmail.com>
202
203         * testsuite/libffi.call/cls_longdouble_va.c (main): Fix format string.
204         Remove xfails for x86 linux targets.
205
206 2009-12-02  David Edelsohn  <edelsohn@gnu.org>
207
208         * src/powerpc/ffi_darwin.c (ffi_prep_args): Fix typo in INT64
209         case.
210
211 2009-12-01  David Edelsohn  <edelsohn@gnu.org>
212
213         * src/powerpc/aix.S (ffi_call_AIX): Convert to more standard
214         register usage.  Call ffi_prep_args directly.  Add long double
215         return value support.
216         * src/powerpc/ffi_darwin.c (ffi_prep_args): Double arg increment
217         applies to FFI_TYPE_DOUBLE.  Correct fpr_base increment typo.
218         Separate FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases.
219         (ffi_prep_cif_machdep): Only 16 byte stack alignment in 64 bit
220         mode.
221         (ffi_closure_helper_DARWIN): Remove nf and ng counters.  Move temp
222         into case.
223         * src/powerpc/aix_closure.S: Maintain 16 byte stack alignment.
224         Allocate result area between params and FPRs.
225
226 2009-11-30  David Edelsohn  <edelsohn@gnu.org>
227
228         PR target/35484
229         * src/powerpc/ffitarget.h (POWERPC64): Define for PPC64 Linux and
230         AIX64.
231         * src/powerpc/aix.S: Implement AIX64 version.
232         * src/powerpc/aix_closure.S: Implement AIX64 version.
233         (ffi_closure_ASM): Use extsb, lha and displament addresses.
234         * src/powerpc/ffi_darwin.c (ffi_prep_args): Implement AIX64
235         support.
236         (ffi_prep_cif_machdep): Same.
237         (ffi_call): Same.
238         (ffi_closure_helper_DARWIN): Same.
239
240 2009-11-02  Andreas Tobler  <a.tobler@schweiz.org>
241
242         PR libffi/41908
243         * testsuite/libffi.call/testclosure.c: New test.
244
245 2009-09-28  Kai Tietz  <kai.tietz@onevision.com>
246
247         * src/x86/win64.S (_ffi_call_win64 stack): Remove for gnu
248         assembly version use of ___chkstk.
249
250 2009-09-23  Matthias Klose  <doko@ubuntu.com>
251
252         PR libffi/40242, PR libffi/41443
253         * src/arm/sysv.S (__ARM_ARCH__): Define for processors
254         __ARM_ARCH_6T2__, __ARM_ARCH_6M__, __ARM_ARCH_7__,
255         __ARM_ARCH_7A__, __ARM_ARCH_7R__, __ARM_ARCH_7M__.
256         Change the conditionals to __SOFTFP__ || __ARM_EABI__
257         for -mfloat-abi=softfp to work.
258
259 2009-09-17  Loren J. Rittle  <ljrittle@acm.org>
260
261         PR testsuite/32843 (strikes again)
262         * src/x86/ffi.c (ffi_prep_cif_machdep): Add X86_FREEBSD to
263         enable proper extension on char and short.
264
265 2009-09-15  David Daney  <ddaney@caviumnetworks.com>
266
267         * src/java_raw_api.c (ffi_java_raw_to_rvalue): Remove special
268         handling for FFI_TYPE_POINTER.
269         * src/mips/ffitarget.h (FFI_TYPE_STRUCT_D_SOFT,
270         FFI_TYPE_STRUCT_F_SOFT, FFI_TYPE_STRUCT_DD_SOFT,
271         FFI_TYPE_STRUCT_FF_SOFT, FFI_TYPE_STRUCT_FD_SOFT,
272         FFI_TYPE_STRUCT_DF_SOFT, FFI_TYPE_STRUCT_SOFT): New defines.
273         (FFI_N32_SOFT_FLOAT, FFI_N64_SOFT_FLOAT): New ffi_abi enumerations.
274         (enum ffi_abi): Set FFI_DEFAULT_ABI for soft-float.
275         * src/mips/n32.S (ffi_call_N32): Add handling for soft-float
276         structure and pointer returns.
277         (ffi_closure_N32): Add handling for pointer returns.
278         * src/mips/ffi.c (ffi_prep_args, calc_n32_struct_flags,
279         calc_n32_return_struct_flags): Handle soft-float.
280         (ffi_prep_cif_machdep):  Handle soft-float, fix pointer handling.
281         (ffi_call_N32): Declare proper argument types.
282         (ffi_call, copy_struct_N32, ffi_closure_mips_inner_N32): Handle
283         soft-float.
284
285 2009-08-24  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
286
287         * configure.ac (AC_PREREQ): Bump to 2.64.
288
289 2009-08-22  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
290
291         * Makefile.am (install-html, install-pdf): Remove.
292         * Makefile.in: Regenerate.
293
294         * Makefile.in: Regenerate.
295         * aclocal.m4: Regenerate.
296         * configure: Regenerate.
297         * fficonfig.h.in: Regenerate.
298         * include/Makefile.in: Regenerate.
299         * man/Makefile.in: Regenerate.
300         * testsuite/Makefile.in: Regenerate.
301
302 2009-07-30  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
303
304         * configure.ac (_AC_ARG_VAR_PRECIOUS): Use m4_rename_force.
305
306 2009-07-24  Dave Korn  <dave.korn.cygwin@gmail.com>
307
308         PR libffi/40807
309         * src/x86/ffi.c (ffi_prep_cif_machdep): Also use sign/zero-extending
310         return types for X86_WIN32.
311         * src/x86/win32.S (_ffi_call_SYSV): Handle omitted return types.
312         (_ffi_call_STDCALL, _ffi_closure_SYSV, _ffi_closure_raw_SYSV,
313         _ffi_closure_STDCALL): Likewise.
314
315         * src/closures.c (is_selinux_enabled): Define to const 0 for Cygwin.
316         (dlmmap, dlmunmap): Also use these functions on Cygwin.
317
318 2009-07-11  Richard Sandiford  <rdsandiford@googlemail.com>
319
320         PR testsuite/40699
321         PR testsuite/40707
322         PR testsuite/40709
323         * testsuite/lib/libffi-dg.exp: Revert 2009-07-02, 2009-07-01 and
324         2009-06-30 commits.
325
326 2009-07-01  Richard Sandiford  <r.sandiford@uk.ibm.com>
327
328         * testsuite/lib/libffi-dg.exp (libffi-init): Set ld_library_path
329         to "" before adding paths.  (This reinstates an assignment that
330         was removed by my 2009-06-30 commit, but changes the initial
331         value from "." to "".)
332
333 2009-07-01  H.J. Lu  <hongjiu.lu@intel.com>
334
335         PR testsuite/40601
336         * testsuite/lib/libffi-dg.exp (libffi-init): Properly set
337         gccdir.  Adjust ld_library_path for gcc only if gccdir isn't
338         empty.
339
340 2009-06-30  Richard Sandiford  <r.sandiford@uk.ibm.com>
341
342         * testsuite/lib/libffi-dg.exp (libffi-init): Don't add "."
343         to ld_library_path.  Use add_path.  Add just find_libgcc_s
344         to ld_library_path, not every libgcc multilib directory.
345
346 2009-06-16  Wim Lewis  <wiml@hhhh.org>
347
348         * src/powerpc/ffi.c: Avoid clobbering cr3 and cr4, which are
349         supposed to be callee-saved.
350         * src/powerpc/sysv.S (small_struct_return_value): Fix overrun of
351         return buffer for odd-size structs.
352
353 2009-06-16  Andreas Tobler  <a.tobler@schweiz.org>
354
355         PR libffi/40444
356         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Add
357         allow_stack_execute for Darwin.
358
359 2009-06-16  Andrew Haley  <aph@redhat.com>
360
361         * configure.ac (TARGETDIR): Add missing blank lines.
362         * configure: Regenerate.
363
364 2009-06-16  Andrew Haley  <aph@redhat.com>
365
366         * testsuite/libffi.call/cls_align_sint64.c,
367         testsuite/libffi.call/cls_align_uint64.c,
368         testsuite/libffi.call/cls_longdouble_va.c,
369         testsuite/libffi.call/cls_ulonglong.c,
370         testsuite/libffi.call/return_ll1.c,
371         testsuite/libffi.call/stret_medium2.c: Fix printf format
372         specifiers.
373         * testsuite/libffi.call/ffitest.h,
374         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRIuLL): Define.
375
376 2009-06-15  Andrew Haley  <aph@redhat.com>
377
378         * testsuite/libffi.call/err_bad_typedef.c: xfail everywhere.
379         * testsuite/libffi.call/err_bad_abi.c: Likewise.
380
381 2009-06-12  Andrew Haley  <aph@redhat.com>
382
383         * Makefile.am: Remove info_TEXINFOS.
384
385 2009-06-12  Andrew Haley  <aph@redhat.com>
386
387         * ChangeLog.libffi: testsuite/libffi.call/cls_align_sint64.c,
388         testsuite/libffi.call/cls_align_uint64.c,
389         testsuite/libffi.call/cls_ulonglong.c,
390         testsuite/libffi.call/return_ll1.c,
391         testsuite/libffi.call/stret_medium2.c: Fix printf format
392         specifiers.
393         testsuite/libffi.special/unwindtest.cc: include stdint.h.
394
395 2009-06-11  Timothy Wall  <twall@users.sf.net>
396
397         * Makefile.am,
398         configure.ac,
399         include/ffi.h.in,
400         include/ffi_common.h,
401         src/closures.c,
402         src/dlmalloc.c,
403         src/x86/ffi.c,
404         src/x86/ffitarget.h,
405         src/x86/win64.S (new),
406         README: Added win64 support (mingw or MSVC)
407         * Makefile.in,
408         include/Makefile.in,
409         man/Makefile.in,
410         testsuite/Makefile.in,
411         configure,
412         aclocal.m4: Regenerated
413         * ltcf-c.sh: properly escape cygwin/w32 path
414         * man/ffi_call.3: Clarify size requirements for return value.
415         * src/x86/ffi64.c: Fix filename in comment.
416         * src/x86/win32.S: Remove unused extern.
417
418         * testsuite/libffi.call/closure_fn0.c,
419         testsuite/libffi.call/closure_fn1.c,
420         testsuite/libffi.call/closure_fn2.c,
421         testsuite/libffi.call/closure_fn3.c,
422         testsuite/libffi.call/closure_fn4.c,
423         testsuite/libffi.call/closure_fn5.c,
424         testsuite/libffi.call/closure_fn6.c,
425         testsuite/libffi.call/closure_stdcall.c,
426         testsuite/libffi.call/cls_12byte.c,
427         testsuite/libffi.call/cls_16byte.c,
428         testsuite/libffi.call/cls_18byte.c,
429         testsuite/libffi.call/cls_19byte.c,
430         testsuite/libffi.call/cls_1_1byte.c,
431         testsuite/libffi.call/cls_20byte.c,
432         testsuite/libffi.call/cls_20byte1.c,
433         testsuite/libffi.call/cls_24byte.c,
434         testsuite/libffi.call/cls_2byte.c,
435         testsuite/libffi.call/cls_3_1byte.c,
436         testsuite/libffi.call/cls_3byte1.c,
437         testsuite/libffi.call/cls_3byte2.c,
438         testsuite/libffi.call/cls_4_1byte.c,
439         testsuite/libffi.call/cls_4byte.c,
440         testsuite/libffi.call/cls_5_1_byte.c,
441         testsuite/libffi.call/cls_5byte.c,
442         testsuite/libffi.call/cls_64byte.c,
443         testsuite/libffi.call/cls_6_1_byte.c,
444         testsuite/libffi.call/cls_6byte.c,
445         testsuite/libffi.call/cls_7_1_byte.c,
446         testsuite/libffi.call/cls_7byte.c,
447         testsuite/libffi.call/cls_8byte.c,
448         testsuite/libffi.call/cls_9byte1.c,
449         testsuite/libffi.call/cls_9byte2.c,
450         testsuite/libffi.call/cls_align_double.c,
451         testsuite/libffi.call/cls_align_float.c,
452         testsuite/libffi.call/cls_align_longdouble.c,
453         testsuite/libffi.call/cls_align_longdouble_split.c,
454         testsuite/libffi.call/cls_align_longdouble_split2.c,
455         testsuite/libffi.call/cls_align_pointer.c,
456         testsuite/libffi.call/cls_align_sint16.c,
457         testsuite/libffi.call/cls_align_sint32.c,
458         testsuite/libffi.call/cls_align_sint64.c,
459         testsuite/libffi.call/cls_align_uint16.c,
460         testsuite/libffi.call/cls_align_uint32.c,
461         testsuite/libffi.call/cls_align_uint64.c,
462         testsuite/libffi.call/cls_dbls_struct.c,
463         testsuite/libffi.call/cls_double.c,
464         testsuite/libffi.call/cls_double_va.c,
465         testsuite/libffi.call/cls_float.c,
466         testsuite/libffi.call/cls_longdouble.c,
467         testsuite/libffi.call/cls_longdouble_va.c,
468         testsuite/libffi.call/cls_multi_schar.c,
469         testsuite/libffi.call/cls_multi_sshort.c,
470         testsuite/libffi.call/cls_multi_sshortchar.c,
471         testsuite/libffi.call/cls_multi_uchar.c,
472         testsuite/libffi.call/cls_multi_ushort.c,
473         testsuite/libffi.call/cls_multi_ushortchar.c,
474         testsuite/libffi.call/cls_pointer.c,
475         testsuite/libffi.call/cls_pointer_stack.c,
476         testsuite/libffi.call/cls_schar.c,
477         testsuite/libffi.call/cls_sint.c,
478         testsuite/libffi.call/cls_sshort.c,
479         testsuite/libffi.call/cls_uchar.c,
480         testsuite/libffi.call/cls_uint.c,
481         testsuite/libffi.call/cls_ulonglong.c,
482         testsuite/libffi.call/cls_ushort.c,
483         testsuite/libffi.call/err_bad_abi.c,
484         testsuite/libffi.call/err_bad_typedef.c,
485         testsuite/libffi.call/float2.c,
486         testsuite/libffi.call/huge_struct.c,
487         testsuite/libffi.call/nested_struct.c,
488         testsuite/libffi.call/nested_struct1.c,
489         testsuite/libffi.call/nested_struct10.c,
490         testsuite/libffi.call/nested_struct2.c,
491         testsuite/libffi.call/nested_struct3.c,
492         testsuite/libffi.call/nested_struct4.c,
493         testsuite/libffi.call/nested_struct5.c,
494         testsuite/libffi.call/nested_struct6.c,
495         testsuite/libffi.call/nested_struct7.c,
496         testsuite/libffi.call/nested_struct8.c,
497         testsuite/libffi.call/nested_struct9.c,
498         testsuite/libffi.call/problem1.c,
499         testsuite/libffi.call/return_ldl.c,
500         testsuite/libffi.call/return_ll1.c,
501         testsuite/libffi.call/stret_large.c,
502         testsuite/libffi.call/stret_large2.c,
503         testsuite/libffi.call/stret_medium.c,
504         testsuite/libffi.call/stret_medium2.c,
505         testsuite/libffi.special/unwindtest.cc: use ffi_closure_alloc instead
506         of checking for MMAP.  Use intptr_t instead of long casts.
507
508 2009-06-11  Kaz Kojima  <kkojima@gcc.gnu.org>
509
510         * testsuite/libffi.call/cls_longdouble_va.c: Add xfail sh*-*-linux-*.
511         * testsuite/libffi.call/err_bad_abi.c: Add xfail sh*-*-*.
512         * testsuite/libffi.call/err_bad_typedef.c: Likewise.
513
514 2009-06-09  Andrew Haley  <aph@redhat.com>
515
516         * src/x86/freebsd.S: Add missing file.
517
518 2009-06-08  Andrew Haley  <aph@redhat.com>
519
520         Import from libffi 3.0.8:
521
522         * doc/libffi.texi: New file.
523         * doc/libffi.info: Likewise.
524         * doc/stamp-vti: Likewise.
525         * man/Makefile.am: New file.
526         * man/ffi_call.3: New file.
527
528         * Makefile.am (EXTRA_DIST): Add src/x86/darwin64.S,
529         src/dlmalloc.c.
530         (nodist_libffi_la_SOURCES): Add X86_FREEBSD.
531
532         * configure.ac: Bump version to 3.0.8.
533         parisc*-*-linux*: Add.
534         i386-*-freebsd* | i386-*-openbsd*: Add.
535         powerpc-*-beos*: Add.
536         AM_CONDITIONAL X86_FREEBSD: Add.
537         AC_CONFIG_FILES: Add man/Makefile.
538
539         * include/ffi.h.in (FFI_FN): Change void (*)() to void (*)(void).
540
541 2009-06-08  Andrew Haley  <aph@redhat.com>
542
543         * README: Import from libffi 3.0.8.
544
545 2009-06-08  Andrew Haley  <aph@redhat.com>
546
547         * testsuite/libffi.call/err_bad_abi.c: Add xfails.
548         * testsuite/libffi.call/cls_longdouble_va.c: Add xfails.
549         * testsuite/libffi.call/cls_dbls_struct.c: Add xfail x86_64-*-linux-*.
550         * testsuite/libffi.call/err_bad_typedef.c: Add xfails.
551
552         * testsuite/libffi.call/stret_medium2.c: Add __UNUSED__ to args.
553         * testsuite/libffi.call/stret_medium.c: Likewise.
554         * testsuite/libffi.call/stret_large2.c: Likewise.
555         * testsuite/libffi.call/stret_large.c:  Likewise.
556
557 2008-12-26  Timothy Wall  <twall@users.sf.net>
558
559         * testsuite/libffi.call/cls_longdouble.c,
560         testsuite/libffi.call/cls_longdouble_va.c,
561         testsuite/libffi.call/cls_align_longdouble.c,
562         testsuite/libffi.call/cls_align_longdouble_split.c,
563         testsuite/libffi.call/cls_align_longdouble_split2.c: mark expected
564         failures on x86_64 cygwin/mingw.
565
566 2008-12-22  Timothy Wall  <twall@users.sf.net>
567
568         * testsuite/libffi.call/closure_fn0.c,
569         testsuite/libffi.call/closure_fn1.c,
570         testsuite/libffi.call/closure_fn2.c,
571         testsuite/libffi.call/closure_fn3.c,
572         testsuite/libffi.call/closure_fn4.c,
573         testsuite/libffi.call/closure_fn5.c,
574         testsuite/libffi.call/closure_fn6.c,
575         testsuite/libffi.call/closure_loc_fn0.c,
576         testsuite/libffi.call/closure_stdcall.c,
577         testsuite/libffi.call/cls_align_pointer.c,
578         testsuite/libffi.call/cls_pointer.c,
579         testsuite/libffi.call/cls_pointer_stack.c: use portable cast from
580         pointer to integer (intptr_t).
581         * testsuite/libffi.call/cls_longdouble.c: disable for win64.
582
583 2008-07-24  Anthony Green  <green@redhat.com>
584
585         * testsuite/libffi.call/cls_dbls_struct.c,
586         testsuite/libffi.call/cls_double_va.c,
587         testsuite/libffi.call/cls_longdouble.c,
588         testsuite/libffi.call/cls_longdouble_va.c,
589         testsuite/libffi.call/cls_pointer.c,
590         testsuite/libffi.call/cls_pointer_stack.c,
591         testsuite/libffi.call/err_bad_abi.c: Clean up failures from
592         compiler warnings.
593
594 2008-03-04  Anthony Green  <green@redhat.com>
595             Blake Chaffin
596             hos@tamanegi.org
597
598         * testsuite/libffi.call/cls_align_longdouble_split2.c
599           testsuite/libffi.call/cls_align_longdouble_split.c
600           testsuite/libffi.call/cls_dbls_struct.c
601           testsuite/libffi.call/cls_double_va.c
602           testsuite/libffi.call/cls_longdouble.c
603           testsuite/libffi.call/cls_longdouble_va.c
604           testsuite/libffi.call/cls_pointer.c
605           testsuite/libffi.call/cls_pointer_stack.c
606           testsuite/libffi.call/err_bad_abi.c
607           testsuite/libffi.call/err_bad_typedef.c
608           testsuite/libffi.call/stret_large2.c
609           testsuite/libffi.call/stret_large.c
610           testsuite/libffi.call/stret_medium2.c
611           testsuite/libffi.call/stret_medium.c: New tests from Apple.
612
613 2009-06-05  Andrew Haley  <aph@redhat.com>
614
615         * src/x86/ffitarget.h, src/x86/ffi.c: Merge stdcall changes from
616         libffi.
617
618 2009-06-04  Andrew Haley  <aph@redhat.com>
619
620         * src/x86/ffitarget.h, src/x86/win32.S, src/x86/ffi.c: Back out
621         stdcall changes.
622
623 2008-02-26  Anthony Green  <green@redhat.com>
624             Thomas Heller  <theller@ctypes.org>
625
626         * src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C
627         comment.
628
629 2008-02-03  Timothy Wall  <twall@users.sf.net>
630
631         * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return
632           offset based on code pointer, not data pointer.
633
634 2008-01-31  Timothy Wall <twall@users.sf.net>
635
636         * testsuite/libffi.call/closure_stdcall.c: Add test for stdcall
637         closures.
638         * src/x86/ffitarget.h: Increase size of trampoline for stdcall
639         closures.
640         * src/x86/win32.S: Add assembly for stdcall closure.
641         * src/x86/ffi.c: Initialize stdcall closure trampoline.
642
643 2009-06-04  Andrew Haley  <aph@redhat.com>
644
645         * include/ffi.h.in: Change void (*)() to void (*)(void).
646         * src/x86/ffi.c: Likewise.
647
648 2009-06-04  Andrew Haley  <aph@redhat.com>
649
650         * src/powerpc/ppc_closure.S: Insert licence header.
651         * src/powerpc/linux64_closure.S: Likewise.
652         * src/m68k/sysv.S: Likewise.
653
654         * src/sh64/ffi.c: Change void (*)() to void (*)(void).
655         * src/powerpc/ffi.c: Likewise.
656         * src/powerpc/ffi_darwin.c: Likewise.
657         * src/m32r/ffi.c: Likewise.
658         * src/sh64/ffi.c: Likewise.
659         * src/x86/ffi64.c: Likewise.
660         * src/alpha/ffi.c: Likewise.
661         * src/alpha/osf.S: Likewise.
662         * src/frv/ffi.c: Likewise.
663         * src/s390/ffi.c: Likewise.
664         * src/pa/ffi.c: Likewise.
665         * src/pa/hpux32.S: Likewise.
666         * src/ia64/unix.S: Likewise.
667         * src/ia64/ffi.c: Likewise.
668         * src/sparc/ffi.c: Likewise.
669         * src/mips/ffi.c: Likewise.
670         * src/sh/ffi.c: Likewise.
671
672 2008-02-15  David Daney  <ddaney@avtrex.com>
673
674         * src/mips/ffi.c (USE__BUILTIN___CLEAR_CACHE):
675         Define (conditionally), and use it to include cachectl.h.
676         (ffi_prep_closure_loc): Fix cache flushing.
677         * src/mips/ffitarget.h (_ABIN32, _ABI64, _ABIO32): Define.
678
679 2009-06-04  Andrew Haley  <aph@redhat.com>
680
681         include/ffi.h.in,
682         src/arm/ffitarget.h,
683         src/arm/ffi.c,
684         src/arm/sysv.S,
685         src/powerpc/ffitarget.h,
686         src/closures.c,
687         src/sh64/ffitarget.h,
688         src/sh64/ffi.c,
689         src/sh64/sysv.S,
690         src/types.c,
691         src/x86/ffi64.c,
692         src/x86/ffitarget.h,
693         src/x86/win32.S,
694         src/x86/darwin.S,
695         src/x86/ffi.c,
696         src/x86/sysv.S,
697         src/x86/unix64.S,
698         src/alpha/ffitarget.h,
699         src/alpha/ffi.c,
700         src/alpha/osf.S,
701         src/m68k/ffitarget.h,
702         src/frv/ffitarget.h,
703         src/frv/ffi.c,
704         src/s390/ffitarget.h,
705         src/s390/sysv.S,
706         src/cris/ffitarget.h,
707         src/pa/linux.S,
708         src/pa/ffitarget.h,
709         src/pa/ffi.c,
710         src/raw_api.c,
711         src/ia64/ffitarget.h,
712         src/ia64/unix.S,
713         src/ia64/ffi.c,
714         src/ia64/ia64_flags.h,
715         src/java_raw_api.c,
716         src/debug.c,
717         src/sparc/v9.S,
718         src/sparc/ffitarget.h,
719         src/sparc/ffi.c,
720         src/sparc/v8.S,
721         src/mips/ffitarget.h,
722         src/mips/n32.S,
723         src/mips/o32.S,
724         src/mips/ffi.c,
725         src/prep_cif.c,
726         src/sh/ffitarget.h,
727         src/sh/ffi.c,
728         src/sh/sysv.S: Update license text.
729
730 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
731
732         * src/x86/win32.S (_ffi_closure_STDCALL):  New function.
733         (.eh_frame):  Add FDE for it.
734
735 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
736
737         * configure.ac:  Also check if assembler supports pc-relative
738         relocs on X86_WIN32 targets.
739         * configure:  Regenerate.
740         * src/x86/win32.S (ffi_prep_args):  Declare extern, not global.
741         (_ffi_call_SYSV):  Add missing function type symbol .def and
742         add EH markup labels.
743         (_ffi_call_STDCALL):  Likewise.
744         (_ffi_closure_SYSV):  Likewise.
745         (_ffi_closure_raw_SYSV):  Likewise.
746         (.eh_frame):  Add hand-crafted EH data.
747
748 2009-04-09  Jakub Jelinek  <jakub@redhat.com>
749
750         * testsuite/lib/libffi-dg.exp: Change copyright header to refer to
751         version 3 of the GNU General Public License and to point readers
752         at the COPYING3 file and the FSF's license web page.
753         * testsuite/libffi.call/call.exp: Likewise.
754         * testsuite/libffi.special/special.exp: Likewise.
755
756 2009-03-01  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
757
758         * configure: Regenerate.
759
760 2008-12-18  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
761
762         PR libffi/26048
763         * configure.ac (HAVE_AS_X86_PCREL): New test.
764         * configure: Regenerate.
765         * fficonfig.h.in: Regenerate.
766         * src/x86/sysv.S [!FFI_NO_RAW_API]: Precalculate
767         RAW_CLOSURE_CIF_OFFSET, RAW_CLOSURE_FUN_OFFSET,
768         RAW_CLOSURE_USER_DATA_OFFSET for the Solaris 10/x86 assembler.
769         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
770         * src/x86/unix64.S (.Lstore_table): Move to .text section.
771         (.Lload_table): Likewise.
772         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
773
774 2008-12-18  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
775
776         * configure: Regenerate.
777
778 2008-11-21  Eric Botcazou  <ebotcazou@adacore.com>
779
780         * src/sparc/ffi.c (ffi_prep_cif_machdep): Add support for
781         signed/unsigned int8/16 return values.
782         * src/sparc/v8.S (ffi_call_v8): Likewise.
783         (ffi_closure_v8): Likewise.
784
785 2008-09-26  Peter O'Gorman  <pogma@thewrittenword.com>
786             Steve Ellcey  <sje@cup.hp.com>
787
788         * configure: Regenerate for new libtool.
789         * Makefile.in: Ditto.
790         * include/Makefile.in: Ditto.
791         * aclocal.m4: Ditto.
792
793 2008-08-25  Andreas Tobler  <a.tobler@schweiz.org>
794
795         * src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and
796         FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum.
797         Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT.
798         Adjust copyright notice.
799         * src/powerpc/ffi.c: Add two new flags to indicate if we have one
800         register or two register to use for FFI_SYSV structs.
801         (ffi_prep_cif_machdep): Pass the right register flag introduced above.
802         (ffi_closure_helper_SYSV): Fix the return type for
803         FFI_SYSV_TYPE_SMALL_STRUCT. Comment.
804         Adjust copyright notice.
805
806 2008-07-16  Kaz Kojima  <kkojima@gcc.gnu.org>
807
808         * src/sh/ffi.c (ffi_prep_closure_loc): Turn INSN into an unsigned
809         int.
810
811 2008-06-17  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
812
813         * configure: Regenerate.
814         * include/Makefile.in: Regenerate.
815         * testsuite/Makefile.in: Regenerate.
816
817 2008-06-07  Joseph Myers  <joseph@codesourcery.com>
818
819         * configure.ac (parisc*-*-linux*, powerpc-*-sysv*,
820         powerpc-*-beos*): Remove.
821         * configure: Regenerate.
822
823 2008-05-09  Julian Brown  <julian@codesourcery.com>
824
825         * Makefile.am (LTLDFLAGS): New.
826         (libffi_la_LDFLAGS): Use above.
827         * Makefile.in: Regenerate.
828
829 2008-04-18  Paolo Bonzini  <bonzini@gnu.org>
830
831         PR bootstrap/35457
832         * aclocal.m4: Regenerate.
833         * configure: Regenerate.
834
835 2008-03-26  Kaz Kojima  <kkojima@gcc.gnu.org>
836
837         * src/sh/sysv.S: Add .note.GNU-stack on Linux.
838         * src/sh64/sysv.S: Likewise.
839
840 2008-03-26  Daniel Jacobowitz  <dan@debian.org>
841
842         * src/arm/sysv.S: Fix ARM comment marker.
843
844 2008-03-26  Jakub Jelinek  <jakub@redhat.com>
845
846         * src/alpha/osf.S: Add .note.GNU-stack on Linux.
847         * src/s390/sysv.S: Likewise.
848         * src/powerpc/ppc_closure.S: Likewise.
849         * src/powerpc/sysv.S: Likewise.
850         * src/x86/unix64.S: Likewise.
851         * src/x86/sysv.S: Likewise.
852         * src/sparc/v8.S: Likewise.
853         * src/sparc/v9.S: Likewise.
854         * src/m68k/sysv.S: Likewise.
855         * src/arm/sysv.S: Likewise.
856
857 2008-03-16  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
858
859         * aclocal.m4: Regenerate.
860         * configure: Likewise.
861         * Makefile.in: Likewise.
862         * include/Makefile.in: Likewise.
863         * testsuite/Makefile.in: Likewise.
864
865 2008-02-12  Bjoern Koenig  <bkoenig@alpha-tierchen.de>
866             Andreas Tobler  <a.tobler@schweiz.org>
867
868         * configure.ac: Add amd64-*-freebsd* target.
869         * configure: Regenerate.
870
871 2008-01-30  H.J. Lu  <hongjiu.lu@intel.com>
872
873         PR libffi/34612
874         * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when
875         returning struct.
876
877         * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer"
878         tests.
879
880 2008-01-24  David Edelsohn  <edelsohn@gnu.org>
881
882         * configure: Regenerate.
883
884 2008-01-06  Andreas Tobler  <a.tobler@schweiz.org>
885
886         * src/x86/ffi.c (ffi_prep_cif_machdep): Fix thinko.
887
888 2008-01-05  Andreas Tobler  <a.tobler@schweiz.org>
889
890         PR testsuite/32843
891         * src/x86/ffi.c (ffi_prep_cif_machdep): Add code for
892         signed/unsigned int8/16 for X86_DARWIN.
893         Updated copyright info.
894         Handle one and two byte structs with special cif->flags.
895         * src/x86/ffitarget.h: Add special types for one and two byte structs.
896         Updated copyright info.
897         * src/x86/darwin.S (ffi_call_SYSV): Rewrite to use a jump table like
898         sysv.S
899         Remove code to pop args from the stack after call.
900         Special-case signed/unsigned for int8/16, one and two byte structs.
901         (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
902         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
903         FFI_TYPE_SINT32.
904         Updated copyright info.
905
906 2007-12-08  David Daney  <ddaney@avtrex.com>
907
908         * src/mips/n32.S (ffi_call_N32):  Replace dadd with ADDU, dsub with
909         SUBU, add with ADDU and use smaller code sequences.
910
911 2007-12-07  David Daney  <ddaney@avtrex.com>
912
913         * src/mips/ffi.c (ffi_prep_cif_machdep): Handle long double return
914         type.
915
916 2007-12-06  David Daney  <ddaney@avtrex.com>
917
918         * include/ffi.h.in (FFI_SIZEOF_JAVA_RAW): Define if not already
919         defined.
920         (ffi_java_raw): New typedef.
921         (ffi_java_raw_call, ffi_java_ptrarray_to_raw,
922         ffi_java_raw_to_ptrarray): Change parameter types from ffi_raw to
923         ffi_java_raw.
924         (ffi_java_raw_closure) : Same.
925         (ffi_prep_java_raw_closure, ffi_prep_java_raw_closure_loc): Change
926         parameter types.
927         * src/java_raw_api.c (ffi_java_raw_size):  Replace FFI_SIZEOF_ARG with
928         FFI_SIZEOF_JAVA_RAW.
929         (ffi_java_raw_to_ptrarray): Change type of raw to ffi_java_raw.
930         Replace FFI_SIZEOF_ARG with FFI_SIZEOF_JAVA_RAW. Use
931         sizeof(ffi_java_raw) for alignment calculations.
932         (ffi_java_ptrarray_to_raw): Same.
933         (ffi_java_rvalue_to_raw): Add special handling for FFI_TYPE_POINTER
934         if FFI_SIZEOF_JAVA_RAW == 4.
935         (ffi_java_raw_to_rvalue): Same.
936         (ffi_java_raw_call): Change type of raw to ffi_java_raw.
937         (ffi_java_translate_args): Same.
938         (ffi_prep_java_raw_closure_loc, ffi_prep_java_raw_closure): Change
939         parameter types.
940         * src/mips/ffitarget.h (FFI_SIZEOF_JAVA_RAW): Define for N32 ABI.
941
942 2007-12-06  David Daney  <ddaney@avtrex.com>
943
944         * src/mips/n32.S (ffi_closure_N32): Use 64-bit add instruction on
945         pointer values.
946
947 2007-12-01  Andreas Tobler  <a.tobler@schweiz.org>
948
949         PR libffi/31937
950         * src/powerpc/ffitarget.h: Introduce new ABI FFI_LINUX_SOFT_FLOAT.
951         Add local FFI_TYPE_UINT128 to handle soft-float long-double-128.
952         * src/powerpc/ffi.c: Distinguish between __NO_FPRS__ and not and
953         set the NUM_FPR_ARG_REGISTERS according to.
954         Add support for potential soft-float support under hard-float
955         architecture.
956         (ffi_prep_args_SYSV): Set NUM_FPR_ARG_REGISTERS to 0 in case of
957         FFI_LINUX_SOFT_FLOAT, handle float, doubles and long-doubles according
958         to the FFI_LINUX_SOFT_FLOAT ABI.
959         (ffi_prep_cif_machdep): Likewise.
960         (ffi_closure_helper_SYSV): Likewise.
961         * src/powerpc/ppc_closure.S: Make sure not to store float/double
962         on archs where __NO_FPRS__ is true.
963         Add FFI_TYPE_UINT128 support.
964         * src/powerpc/sysv.S: Add support for soft-float long-double-128.
965         Adjust copyright notice.
966
967 2007-11-25  Andreas Tobler  <a.tobler@schweiz.org>
968
969         * src/closures.c: Move defintion of MAYBE_UNUSED from here to ...
970         * include/ffi_common.h: ... here.
971         Update copyright.
972
973 2007-11-17  Andreas Tobler  <a.tobler@schweiz.org>
974
975         * src/powerpc/sysv.S: Load correct cr to compare if we have long double.
976         * src/powerpc/linux64.S: Likewise.
977         * src/powerpc/ffi.c: Add a comment to show which part goes into cr6.
978         * testsuite/libffi.call/return_ldl.c: New test.
979
980 2007-09-04    <aph@redhat.com>
981
982         * src/arm/sysv.S (UNWIND): New.
983         (Whole file): Conditionally compile unwinder directives.
984         * src/arm/sysv.S: Add unwinder directives.
985
986         * src/arm/ffi.c (ffi_prep_args): Align structs by at least 4 bytes.
987         Only treat r0 as a struct address if we're actually returning a
988         struct by address.
989         Only copy the bytes that are actually within a struct.
990         (ffi_prep_cif_machdep): A Composite Type not larger than 4 bytes
991         is returned in r0, not passed by address.
992         (ffi_call): Allocate a word-sized temporary for the case where
993         a composite is returned in r0.
994         (ffi_prep_incoming_args_SYSV): Align as necessary.
995
996 2007-08-05  Steven Newbury  <s_j_newbury@yahoo.co.uk>
997
998         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Use __clear_cache instead of
999         directly using the sys_cacheflush syscall.
1000
1001 2007-07-27  Andrew Haley  <aph@redhat.com>
1002
1003         * src/arm/sysv.S (ffi_closure_SYSV): Add soft-float.
1004
1005 2007-09-03  Maciej W. Rozycki  <macro@linux-mips.org>
1006
1007         * Makefile.am: Unify MIPS_IRIX and MIPS_LINUX into MIPS.
1008         * configure.ac: Likewise.
1009         * Makefile.in: Regenerate.
1010         * include/Makefile.in: Likewise.
1011         * testsuite/Makefile.in: Likewise.
1012         * configure: Likewise.
1013
1014 2007-08-24  David Daney  <ddaney@avtrex.com>
1015
1016         * testsuite/libffi.call/return_sl.c: New test.
1017
1018 2007-08-10  David Daney  <ddaney@avtrex.com>
1019
1020         * testsuite/libffi.call/cls_multi_ushort.c,
1021         testsuite/libffi.call/cls_align_uint16.c,
1022         testsuite/libffi.call/nested_struct1.c,
1023         testsuite/libffi.call/nested_struct3.c,
1024         testsuite/libffi.call/cls_7_1_byte.c,
1025         testsuite/libffi.call/nested_struct5.c,
1026         testsuite/libffi.call/cls_double.c,
1027         testsuite/libffi.call/nested_struct7.c,
1028         testsuite/libffi.call/cls_sint.c,
1029         testsuite/libffi.call/nested_struct9.c,
1030         testsuite/libffi.call/cls_20byte1.c,
1031         testsuite/libffi.call/cls_multi_sshortchar.c,
1032         testsuite/libffi.call/cls_align_sint64.c,
1033         testsuite/libffi.call/cls_3byte2.c,
1034         testsuite/libffi.call/cls_multi_schar.c,
1035         testsuite/libffi.call/cls_multi_uchar.c,
1036         testsuite/libffi.call/cls_19byte.c,
1037         testsuite/libffi.call/cls_9byte1.c,
1038         testsuite/libffi.call/cls_align_float.c,
1039         testsuite/libffi.call/closure_fn1.c,
1040         testsuite/libffi.call/problem1.c,
1041         testsuite/libffi.call/closure_fn3.c,
1042         testsuite/libffi.call/cls_sshort.c,
1043         testsuite/libffi.call/closure_fn5.c,
1044         testsuite/libffi.call/cls_align_double.c,
1045         testsuite/libffi.call/nested_struct.c,
1046         testsuite/libffi.call/cls_2byte.c,
1047         testsuite/libffi.call/nested_struct10.c,
1048         testsuite/libffi.call/cls_4byte.c,
1049         testsuite/libffi.call/cls_6byte.c,
1050         testsuite/libffi.call/cls_8byte.c,
1051         testsuite/libffi.call/cls_multi_sshort.c,
1052         testsuite/libffi.call/cls_align_sint16.c,
1053         testsuite/libffi.call/cls_align_uint32.c,
1054         testsuite/libffi.call/cls_20byte.c,
1055         testsuite/libffi.call/cls_float.c,
1056         testsuite/libffi.call/nested_struct2.c,
1057         testsuite/libffi.call/cls_5_1_byte.c,
1058         testsuite/libffi.call/nested_struct4.c,
1059         testsuite/libffi.call/cls_24byte.c,
1060         testsuite/libffi.call/nested_struct6.c,
1061         testsuite/libffi.call/cls_64byte.c,
1062         testsuite/libffi.call/nested_struct8.c,
1063         testsuite/libffi.call/cls_uint.c,
1064         testsuite/libffi.call/cls_multi_ushortchar.c,
1065         testsuite/libffi.call/cls_schar.c,
1066         testsuite/libffi.call/cls_uchar.c,
1067         testsuite/libffi.call/cls_align_uint64.c,
1068         testsuite/libffi.call/cls_ulonglong.c,
1069         testsuite/libffi.call/cls_align_longdouble.c,
1070         testsuite/libffi.call/cls_1_1byte.c,
1071         testsuite/libffi.call/cls_12byte.c,
1072         testsuite/libffi.call/cls_3_1byte.c,
1073         testsuite/libffi.call/cls_3byte1.c,
1074         testsuite/libffi.call/cls_4_1byte.c,
1075         testsuite/libffi.call/cls_6_1_byte.c,
1076         testsuite/libffi.call/cls_16byte.c,
1077         testsuite/libffi.call/cls_18byte.c,
1078         testsuite/libffi.call/closure_fn0.c,
1079         testsuite/libffi.call/cls_9byte2.c,
1080         testsuite/libffi.call/closure_fn2.c,
1081         testsuite/libffi.call/closure_fn4.c,
1082         testsuite/libffi.call/cls_ushort.c,
1083         testsuite/libffi.call/closure_fn6.c,
1084         testsuite/libffi.call/cls_5byte.c,
1085         testsuite/libffi.call/cls_align_pointer.c,
1086         testsuite/libffi.call/cls_7byte.c,
1087         testsuite/libffi.call/cls_align_sint32.c,
1088         testsuite/libffi.special/unwindtest_ffi_call.cc,
1089         testsuite/libffi.special/unwindtest.cc: Remove xfail for mips64*-*-*.
1090
1091 2007-08-10  David Daney  <ddaney@avtrex.com>
1092
1093         PR libffi/28313
1094         * configure.ac: Don't treat mips64 as a special case.
1095         * Makefile.am (nodist_libffi_la_SOURCES): Add n32.S.
1096         * configure: Regenerate
1097         * Makefile.in: Ditto.
1098         * fficonfig.h.in: Ditto.
1099         * src/mips/ffitarget.h (REG_L, REG_S, SUBU, ADDU, SRL, LI): Indent.
1100         (LA, EH_FRAME_ALIGN, FDE_ADDR_BYTES): New preprocessor macros.
1101         (FFI_DEFAULT_ABI): Set for n64 case.
1102         (FFI_CLOSURES, FFI_TRAMPOLINE_SIZE): Define for n32 and n64 cases.
1103         * src/mips/n32.S (ffi_call_N32): Add debug macros and labels for FDE.
1104         (ffi_closure_N32): New function.
1105         (.eh_frame): New section
1106         * src/mips/o32.S: Clean up comments.
1107         (ffi_closure_O32): Pass ffi_closure parameter in $12.
1108         * src/mips/ffi.c: Use FFI_MIPS_N32 instead of
1109         _MIPS_SIM == _ABIN32 throughout.
1110         (FFI_MIPS_STOP_HERE): New, use in place of
1111         ffi_stop_here.
1112         (ffi_prep_args): Use unsigned long to hold pointer values.  Rewrite
1113         to support n32/n64 ABIs.
1114         (calc_n32_struct_flags): Rewrite.
1115         (calc_n32_return_struct_flags): Remove unused variable.  Reverse
1116         position of flag bits.
1117         (ffi_prep_cif_machdep): Rewrite n32 portion.
1118         (ffi_call): Enable for n64.  Add special handling for small structure
1119         return values.
1120         (ffi_prep_closure_loc): Add n32 and n64 support.
1121         (ffi_closure_mips_inner_O32): Add cast to silence warning.
1122         (copy_struct_N32, ffi_closure_mips_inner_N32): New functions.
1123
1124 2007-08-08  David Daney  <ddaney@avtrex.com>
1125
1126         * testsuite/libffi.call/ffitest.h (ffi_type_mylong): Remove definition.
1127         * testsuite/libffi.call/cls_align_uint16.c (main): Use correct type
1128         specifiers.
1129         * testsuite/libffi.call/nested_struct1.c (main): Ditto.
1130         * testsuite/libffi.call/cls_sint.c (main): Ditto.
1131         * testsuite/libffi.call/nested_struct9.c (main): Ditto.
1132         * testsuite/libffi.call/cls_20byte1.c (main): Ditto.
1133         * testsuite/libffi.call/cls_9byte1.c (main): Ditto.
1134         * testsuite/libffi.call/closure_fn1.c (main): Ditto.
1135         * testsuite/libffi.call/closure_fn3.c (main): Ditto.
1136         * testsuite/libffi.call/return_dbl2.c (main): Ditto.
1137         * testsuite/libffi.call/cls_sshort.c (main): Ditto.
1138         * testsuite/libffi.call/return_fl3.c (main): Ditto.
1139         * testsuite/libffi.call/closure_fn5.c (main): Ditto.
1140         * testsuite/libffi.call/nested_struct.c (main): Ditto.
1141         * testsuite/libffi.call/nested_struct10.c (main): Ditto.
1142         * testsuite/libffi.call/return_ll1.c (main): Ditto.
1143         * testsuite/libffi.call/cls_8byte.c (main): Ditto.
1144         * testsuite/libffi.call/cls_align_uint32.c (main): Ditto.
1145         * testsuite/libffi.call/cls_align_sint16.c (main): Ditto.
1146         * testsuite/libffi.call/cls_20byte.c (main): Ditto.
1147         * testsuite/libffi.call/nested_struct2.c (main): Ditto.
1148         * testsuite/libffi.call/cls_24byte.c (main): Ditto.
1149         * testsuite/libffi.call/nested_struct6.c (main): Ditto.
1150         * testsuite/libffi.call/cls_uint.c (main): Ditto.
1151         * testsuite/libffi.call/cls_12byte.c (main): Ditto.
1152         * testsuite/libffi.call/cls_16byte.c (main): Ditto.
1153         * testsuite/libffi.call/closure_fn0.c (main): Ditto.
1154         * testsuite/libffi.call/cls_9byte2.c (main): Ditto.
1155         * testsuite/libffi.call/closure_fn2.c (main): Ditto.
1156         * testsuite/libffi.call/return_dbl1.c (main): Ditto.
1157         * testsuite/libffi.call/closure_fn4.c (main): Ditto.
1158         * testsuite/libffi.call/closure_fn6.c (main): Ditto.
1159         * testsuite/libffi.call/cls_align_sint32.c (main): Ditto.
1160
1161 2007-08-07  Andrew Haley  <aph@redhat.com>
1162
1163         * src/x86/sysv.S (ffi_closure_raw_SYSV): Fix typo in previous
1164         checkin.
1165
1166 2007-08-06  Andrew Haley  <aph@redhat.com>
1167
1168         PR testsuite/32843
1169         * src/x86/sysv.S (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1170         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1171         FFI_TYPE_SINT32.
1172
1173 2007-08-02  David Daney  <ddaney@avtrex.com>
1174
1175         * testsuite/libffi.call/return_ul.c (main): Define return type as
1176         ffi_arg.  Use proper printf conversion specifier.
1177
1178 2007-07-30  Andrew Haley  <aph@redhat.com>
1179
1180         PR testsuite/32843
1181         * src/x86/ffi.c (ffi_prep_cif_machdep): in x86 case, add code for
1182         signed/unsigned int8/16.
1183         * src/x86/sysv.S (ffi_call_SYSV): Rewrite to:
1184         Use a jump table.
1185         Remove code to pop args from the stack after call.
1186         Special-case signed/unsigned int8/16.
1187         * testsuite/libffi.call/return_sc.c (main): Revert.
1188
1189 2007-07-26  Richard Guenther  <rguenther@suse.de>
1190
1191         PR testsuite/32843
1192         * testsuite/libffi.call/return_sc.c (main): Verify call
1193         result as signed char, not ffi_arg.
1194
1195 2007-07-16  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1196
1197         * configure.ac (i?86-*-solaris2.1[0-9]): Set TARGET to X86_64.
1198         * configure: Regenerate.
1199
1200 2007-07-11  David Daney  <ddaney@avtrex.com>
1201
1202         * src/mips/ffi.c: Don't include sys/cachectl.h.
1203         (ffi_prep_closure_loc): Use __builtin___clear_cache() instead of
1204         cacheflush().
1205
1206 2007-05-18  Aurelien Jarno  <aurelien@aurel32.net>
1207
1208         * src/arm/ffi.c (ffi_prep_closure_loc): Renamed and ajusted
1209         from (ffi_prep_closure): ... this.
1210         (FFI_INIT_TRAMPOLINE): Adjust.
1211
1212 2005-12-31  Phil Blundell  <pb@reciva.com>
1213
1214         * src/arm/ffi.c (ffi_prep_incoming_args_SYSV,
1215         ffi_closure_SYSV_inner, ffi_prep_closure): New, add closure support.
1216         * src/arm/sysv.S(ffi_closure_SYSV): Likewise.
1217         * src/arm/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1218         (FFI_CLOSURES): Enable closure support.
1219
1220 2007-07-03  Andrew Haley  <aph@hedges.billgatliff.com>
1221
1222         * testsuite/libffi.call/cls_multi_ushort.c,
1223         testsuite/libffi.call/cls_align_uint16.c,
1224         testsuite/libffi.call/nested_struct1.c,
1225         testsuite/libffi.call/nested_struct3.c,
1226         testsuite/libffi.call/cls_7_1_byte.c,
1227         testsuite/libffi.call/cls_double.c,
1228         testsuite/libffi.call/nested_struct5.c,
1229         testsuite/libffi.call/nested_struct7.c,
1230         testsuite/libffi.call/cls_sint.c,
1231         testsuite/libffi.call/nested_struct9.c,
1232         testsuite/libffi.call/cls_20byte1.c,
1233         testsuite/libffi.call/cls_multi_sshortchar.c,
1234         testsuite/libffi.call/cls_align_sint64.c,
1235         testsuite/libffi.call/cls_3byte2.c,
1236         testsuite/libffi.call/cls_multi_schar.c,
1237         testsuite/libffi.call/cls_multi_uchar.c,
1238         testsuite/libffi.call/cls_19byte.c,
1239         testsuite/libffi.call/cls_9byte1.c,
1240         testsuite/libffi.call/cls_align_float.c,
1241         testsuite/libffi.call/closure_fn1.c,
1242         testsuite/libffi.call/problem1.c,
1243         testsuite/libffi.call/closure_fn3.c,
1244         testsuite/libffi.call/cls_sshort.c,
1245         testsuite/libffi.call/closure_fn5.c,
1246         testsuite/libffi.call/cls_align_double.c,
1247         testsuite/libffi.call/cls_2byte.c,
1248         testsuite/libffi.call/nested_struct.c,
1249         testsuite/libffi.call/nested_struct10.c,
1250         testsuite/libffi.call/cls_4byte.c,
1251         testsuite/libffi.call/cls_6byte.c,
1252         testsuite/libffi.call/cls_8byte.c,
1253         testsuite/libffi.call/cls_multi_sshort.c,
1254         testsuite/libffi.call/cls_align_uint32.c,
1255         testsuite/libffi.call/cls_align_sint16.c,
1256         testsuite/libffi.call/cls_float.c,
1257         testsuite/libffi.call/cls_20byte.c,
1258         testsuite/libffi.call/cls_5_1_byte.c,
1259         testsuite/libffi.call/nested_struct2.c,
1260         testsuite/libffi.call/cls_24byte.c,
1261         testsuite/libffi.call/nested_struct4.c,
1262         testsuite/libffi.call/nested_struct6.c,
1263         testsuite/libffi.call/cls_64byte.c,
1264         testsuite/libffi.call/nested_struct8.c,
1265         testsuite/libffi.call/cls_uint.c,
1266         testsuite/libffi.call/cls_multi_ushortchar.c,
1267         testsuite/libffi.call/cls_schar.c,
1268         testsuite/libffi.call/cls_uchar.c,
1269         testsuite/libffi.call/cls_align_uint64.c,
1270         testsuite/libffi.call/cls_ulonglong.c,
1271         testsuite/libffi.call/cls_align_longdouble.c,
1272         testsuite/libffi.call/cls_1_1byte.c,
1273         testsuite/libffi.call/cls_12byte.c,
1274         testsuite/libffi.call/cls_3_1byte.c,
1275         testsuite/libffi.call/cls_3byte1.c,
1276         testsuite/libffi.call/cls_4_1byte.c,
1277         testsuite/libffi.call/cls_6_1_byte.c,
1278         testsuite/libffi.call/cls_16byte.c,
1279         testsuite/libffi.call/cls_18byte.c,
1280         testsuite/libffi.call/closure_fn0.c,
1281         testsuite/libffi.call/cls_9byte2.c,
1282         testsuite/libffi.call/closure_fn2.c,
1283         testsuite/libffi.call/closure_fn4.c,
1284         testsuite/libffi.call/cls_ushort.c,
1285         testsuite/libffi.call/closure_fn6.c,
1286         testsuite/libffi.call/cls_5byte.c,
1287         testsuite/libffi.call/cls_align_pointer.c,
1288         testsuite/libffi.call/cls_7byte.c,
1289         testsuite/libffi.call/cls_align_sint32.c,
1290         testsuite/libffi.special/unwindtest_ffi_call.cc,
1291         testsuite/libffi.special/unwindtest.cc: Enable for ARM.
1292
1293 2007-07-05  H.J. Lu  <hongjiu.lu@intel.com>
1294
1295         * aclocal.m4: Regenerated.
1296
1297 2007-06-02  Paolo Bonzini  <bonzini@gnu.org>
1298
1299         * configure: Regenerate.
1300
1301 2007-05-23  Steve Ellcey  <sje@cup.hp.com>
1302
1303         * Makefile.in: Regenerate.
1304         * configure: Regenerate.
1305         * aclocal.m4: Regenerate.
1306         * include/Makefile.in: Regenerate.
1307         * testsuite/Makefile.in: Regenerate.
1308
1309 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1310
1311         * src/m68k/ffi.c (ffi_prep_incoming_args_SYSV,
1312         ffi_closure_SYSV_inner,ffi_prep_closure): New, add closure support.
1313         * src/m68k/sysv.S(ffi_closure_SYSV,ffi_closure_struct_SYSV): Likewise.
1314         * src/m68k/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1315         (FFI_CLOSURES): Enable closure support.
1316
1317 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1318
1319         * configure.ac (HAVE_AS_CFI_PSEUDO_OP): New test.
1320         * configure: Regenerate.
1321         * fficonfig.h.in: Regenerate.
1322         * src/m68k/sysv.S (CFI_STARTPROC,CFI_ENDPROC,
1323         CFI_OFFSET,CFI_DEF_CFA): New macros.
1324         (ffi_call_SYSV): Add callframe annotation.
1325
1326 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1327
1328         * src/m68k/ffi.c (ffi_prep_args,ffi_prep_cif_machdep): Fix
1329         numerous test suite failures.
1330         * src/m68k/sysv.S (ffi_call_SYSV): Likewise.
1331
1332 2007-04-11  Paolo Bonzini  <bonzini@gnu.org>
1333
1334         * Makefile.am (EXTRA_DIST): Bring up to date.
1335         * Makefile.in: Regenerate.
1336         * src/frv/eabi.S: Remove RCS keyword.
1337
1338 2007-04-06  Richard Henderson  <rth@redhat.com>
1339
1340         * configure.ac: Tidy target case.
1341         (HAVE_LONG_DOUBLE): Allow the target to override.
1342         * configure: Regenerate.
1343         * include/ffi.h.in: Don't define ffi_type_foo if
1344         LIBFFI_HIDE_BASIC_TYPES is defined.
1345         (ffi_type_longdouble): If not HAVE_LONG_DOUBLE, define
1346         to ffi_type_double.
1347         * types.c (LIBFFI_HIDE_BASIC_TYPES): Define.
1348         (FFI_TYPEDEF, ffi_type_void): Mark the data const.
1349         (ffi_type_longdouble): Special case for Alpha.  Don't define
1350         if long double == double.
1351
1352         * src/alpha/ffi.c (FFI_TYPE_LONGDOUBLE): Assert unique value.
1353         (ffi_prep_cif_machdep): Handle it as the 128-bit type.
1354         (ffi_call, ffi_closure_osf_inner): Likewise.
1355         (ffi_closure_osf_inner): Likewise.  Mark hidden.
1356         (ffi_call_osf, ffi_closure_osf): Mark hidden.
1357         * src/alpha/ffitarget.h (FFI_LAST_ABI): Tidy definition.
1358         * src/alpha/osf.S (ffi_call_osf, ffi_closure_osf): Mark hidden.
1359         (load_table): Handle 128-bit long double.
1360
1361         * testsuite/libffi.call/float4.c: Add -mieee for alpha.
1362
1363 2007-04-06  Tom Tromey  <tromey@redhat.com>
1364
1365         PR libffi/31491:
1366         * README: Fixed bug in example.
1367
1368 2007-04-03  Jakub Jelinek  <jakub@redhat.com>
1369
1370         * src/closures.c: Include sys/statfs.h.
1371         (_GNU_SOURCE): Define on Linux.
1372         (FFI_MMAP_EXEC_SELINUX): Define.
1373         (selinux_enabled): New variable.
1374         (selinux_enabled_check): New function.
1375         (is_selinux_enabled): Define.
1376         (dlmmap): Use it.
1377
1378 2007-03-24  Uros Bizjak  <ubizjak@gmail.com>
1379
1380         * testsuite/libffi.call/return_fl2.c (return_fl): Mark as static.
1381         Use 'volatile float sum' to create sum of floats to avoid false
1382         negative due to excess precision on ix86 targets.
1383         (main): Ditto.
1384
1385 2007-03-08  Alexandre Oliva  <aoliva@redhat.com>
1386
1387         * src/powerpc/ffi.c (flush_icache): Fix left-over from previous
1388         patch.
1389         (ffi_prep_closure_loc): Remove unneeded casts.  Add needed ones.
1390
1391 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1392
1393         * include/ffi.h.in (ffi_closure_alloc, ffi_closure_free): New.
1394         (ffi_prep_closure_loc): New.
1395         (ffi_prep_raw_closure_loc): New.
1396         (ffi_prep_java_raw_closure_loc): New.
1397         * src/closures.c: New file.
1398         * src/dlmalloc.c [FFI_MMAP_EXEC_WRIT] (struct malloc_segment):
1399         Replace sflags with exec_offset.
1400         [FFI_MMAP_EXEC_WRIT] (mmap_exec_offset, add_segment_exec_offset,
1401         sub_segment_exec_offset): New macros.
1402         (get_segment_flags, set_segment_flags, check_segment_merge): New
1403         macros.
1404         (is_mmapped_segment, is_extern_segment): Use get_segment_flags.
1405         (add_segment, sys_alloc, create_mspace, create_mspace_with_base,
1406         destroy_mspace): Use new macros.
1407         (sys_alloc): Silence warning.
1408         * Makefile.am (libffi_la_SOURCES): Add src/closures.c.
1409         * Makefile.in: Rebuilt.
1410         * src/prep_cif [FFI_CLOSURES] (ffi_prep_closure): Implement in
1411         terms of ffi_prep_closure_loc.
1412         * src/raw_api.c (ffi_prep_raw_closure_loc): Renamed and adjusted
1413         from...
1414         (ffi_prep_raw_closure): ... this.  Re-implement in terms of the
1415         renamed version.
1416         * src/java_raw_api (ffi_prep_java_raw_closure_loc): Renamed and
1417         adjusted from...
1418         (ffi_prep_java_raw_closure): ... this.  Re-implement in terms of
1419         the renamed version.
1420         * src/alpha/ffi.c (ffi_prep_closure_loc): Renamed from
1421         (ffi_prep_closure): ... this.
1422         * src/pa/ffi.c: Likewise.
1423         * src/cris/ffi.c: Likewise.  Adjust.
1424         * src/frv/ffi.c: Likewise.
1425         * src/ia64/ffi.c: Likewise.
1426         * src/mips/ffi.c: Likewise.
1427         * src/powerpc/ffi_darwin.c: Likewise.
1428         * src/s390/ffi.c: Likewise.
1429         * src/sh/ffi.c: Likewise.
1430         * src/sh64/ffi.c: Likewise.
1431         * src/sparc/ffi.c: Likewise.
1432         * src/x86/ffi64.c: Likewise.
1433         * src/x86/ffi.c: Likewise.
1434         (FFI_INIT_TRAMPOLINE): Adjust.
1435         (ffi_prep_raw_closure_loc): Renamed and adjusted from...
1436         (ffi_prep_raw_closure): ... this.
1437         * src/powerpc/ffi.c (ffi_prep_closure_loc): Renamed from
1438         (ffi_prep_closure): ... this.
1439         (flush_icache): Adjust.
1440
1441 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1442
1443         * src/dlmalloc.c: New file, imported version 2.8.3 of Doug
1444         Lea's malloc.
1445
1446 2007-03-01  Brooks Moses  <brooks.moses@codesourcery.com>
1447
1448         * Makefile.am: Add dummy install-pdf target.
1449         * Makefile.in: Regenerate
1450
1451 2007-02-13  Andreas Krebbel  <krebbel1@de.ibm.com>
1452
1453         * src/s390/ffi.c (ffi_prep_args, ffi_prep_cif_machdep,
1454         ffi_closure_helper_SYSV): Add long double handling.
1455
1456 2007-02-02  Jakub Jelinek  <jakub@redhat.com>
1457
1458         * src/powerpc/linux64.S (ffi_call_LINUX64): Move restore of r2
1459         immediately after bctrl instruction.
1460
1461 2007-01-18  Alexandre Oliva  <aoliva@redhat.com>
1462
1463         * Makefile.am (all-recursive, install-recursive,
1464         mostlyclean-recursive, clean-recursive, distclean-recursive,
1465         maintainer-clean-recursive): Add missing targets.
1466         * Makefile.in: Rebuilt.
1467
1468 2006-12-14  Andreas Tobler  <a.tobler@schweiz.org>
1469
1470         * configure.ac: Add TARGET for x86_64-*-darwin*.
1471         * Makefile.am (nodist_libffi_la_SOURCES): Add rules for 64-bit sources
1472         for X86_DARWIN.
1473         * src/x86/ffitarget.h: Set trampoline size for x86_64-*-darwin*.
1474         * src/x86/darwin64.S: New file for x86_64-*-darwin* support.
1475         * configure: Regenerate.
1476         * Makefile.in: Regenerate.
1477         * include/Makefile.in: Regenerate.
1478         * testsuite/Makefile.in: Regenerate.
1479         * testsuite/libffi.special/unwindtest_ffi_call.cc: New test case for
1480         ffi_call only.
1481
1482 2006-12-13  Andreas Tobler <a.tobler@schweiz.org>
1483
1484         * aclocal.m4: Regenerate with aclocal -I .. as written in the
1485         Makefile.am.
1486
1487 2006-10-31  Geoffrey Keating  <geoffk@apple.com>
1488
1489         * src/powerpc/ffi_darwin.c (darwin_adjust_aggregate_sizes): New.
1490         (ffi_prep_cif_machdep): Call darwin_adjust_aggregate_sizes for
1491         Darwin.
1492         * testsuite/libffi.call/nested_struct4.c: Remove Darwin XFAIL.
1493         * testsuite/libffi.call/nested_struct6.c: Remove Darwin XFAIL.
1494
1495 2006-10-10  Paolo Bonzini  <bonzini@gnu.org>
1496             Sandro Tolaini  <tolaini@libero.it>
1497
1498         * configure.ac [i*86-*-darwin*]: Set X86_DARWIN symbol and
1499         conditional.
1500         * configure: Regenerated.
1501         * Makefile.am (nodist_libffi_la_SOURCES) [X86_DARWIN]: New case.
1502         (EXTRA_DIST): Add src/x86/darwin.S.
1503         * Makefile.in: Regenerated.
1504         * include/Makefile.in: Regenerated.
1505         * testsuite/Makefile.in: Regenerated.
1506
1507         * src/x86/ffi.c (ffi_prep_cif_machdep) [X86_DARWIN]: Treat like
1508         X86_WIN32, and additionally align stack to 16 bytes.
1509         * src/x86/darwin.S: New, based on sysv.S.
1510         * src/prep_cif.c (ffi_prep_cif) [X86_DARWIN]: Align > 8-byte structs.
1511
1512 2006-09-12  David Daney  <ddaney@avtrex.com>
1513
1514         PR libffi/23935
1515         * include/Makefile.am: Install both ffi.h and ffitarget.h in
1516         $(libdir)/gcc/$(target_alias)/$(gcc_version)/include.
1517         * aclocal.m4: Regenerated for automake 1.9.6.
1518         * Makefile.in: Regenerated.
1519         * include/Makefile.in: Regenerated.
1520         * testsuite/Makefile.in: Regenerated.
1521
1522 2006-08-17  Andreas Tobler  <a.tobler@schweiz.ch>
1523
1524         * include/ffi_common.h (struct): Revert accidental commit.
1525
1526 2006-08-15  Andreas Tobler  <a.tobler@schweiz.ch>
1527
1528         * include/ffi_common.h: Remove lint directives.
1529         * include/ffi.h.in: Likewise.
1530
1531 2006-07-25  Torsten Schoenfeld  <kaffeetisch@gmx.de>
1532
1533         * include/ffi.h.in (ffi_type_ulong, ffi_type_slong): Define correctly
1534         for 32-bit architectures.
1535         * testsuite/libffi.call/return_ul.c: New test case.
1536
1537 2006-07-19  David Daney  <ddaney@avtrex.com>
1538
1539         * testsuite/libffi.call/closure_fn6.c: Remove xfail for mips,
1540         xfail remains for mips64.
1541
1542 2006-05-23  Carlos O'Donell  <carlos@codesourcery.com>
1543
1544         * Makefile.am: Add install-html target. Add install-html to .PHONY
1545         * Makefile.in: Regenerate.
1546         * aclocal.m4: Regenerate.
1547         * include/Makefile.in: Regenerate.
1548         * testsuite/Makefile.in: Regenerate.
1549
1550 2006-05-18  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1551
1552         * pa/ffi.c (ffi_prep_args_pa32): Load floating point arguments from
1553         stack slot.
1554
1555 2006-04-22  Andreas Tobler  <a.tobler@schweiz.ch>
1556
1557         * README: Remove notice about 'Crazy Comments'.
1558         * src/debug.c: Remove lint directives. Cleanup white spaces.
1559         * src/java_raw_api.c: Likewise.
1560         * src/prep_cif.c: Likewise.
1561         * src/raw_api.c: Likewise.
1562         * src/ffitest.c: Delete. No longer needed, all test cases migrated
1563         to the testsuite.
1564         * src/arm/ffi.c: Remove lint directives.
1565         * src/m32r/ffi.c: Likewise.
1566         * src/pa/ffi.c: Likewise.
1567         * src/powerpc/ffi.c: Likewise.
1568         * src/powerpc/ffi_darwin.c: Likewise.
1569         * src/sh/ffi.c: Likewise.
1570         * src/sh64/ffi.c: Likewise.
1571         * src/x86/ffi.c: Likewise.
1572         * testsuite/libffi.call/float2.c: Likewise.
1573         * testsuite/libffi.call/promotion.c: Likewise.
1574         * testsuite/libffi.call/struct1.c: Likewise.
1575
1576 2006-04-13  Andreas Tobler  <a.tobler@schweiz.ch>
1577
1578         * src/pa/hpux32.S: Correct unwind offset calculation for
1579         ffi_closure_pa32.
1580         * src/pa/linux.S: Likewise.
1581
1582 2006-04-12  James E Wilson  <wilson@specifix.com>
1583
1584         PR libgcj/26483
1585         * src/ia64/ffi.c (stf_spill, ldf_fill): Rewrite as macros.
1586         (hfa_type_load): Call stf_spill.
1587         (hfa_type_store): Call ldf_fill.
1588         (ffi_call): Adjust calls to above routines.  Add local temps for
1589         macro result.
1590
1591 2006-04-10  Matthias Klose  <doko@debian.org>
1592
1593         * testsuite/lib/libffi-dg.exp (libffi-init): Recognize multilib
1594         directory names containing underscores.
1595
1596 2006-04-07  James E Wilson  <wilson@specifix.com>
1597
1598         * testsuite/libffi.call/float4.c: New testcase.
1599
1600 2006-04-05  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1601             Andreas Tobler  <a.tobler@schweiz.ch>
1602
1603         * Makefile.am: Add PA_HPUX port.
1604         * Makefile.in: Regenerate.
1605         * include/Makefile.in: Likewise.
1606         * testsuite/Makefile.in: Likewise.
1607         * configure.ac: Add PA_HPUX rules.
1608         * configure: Regenerate.
1609         * src/pa/ffitarget.h: Rename linux target to PA_LINUX.
1610         Add PA_HPUX and PA64_HPUX.
1611         Rename FFI_LINUX ABI to FFI_PA32 ABI.
1612         (FFI_TRAMPOLINE_SIZE): Define for 32-bit HP-UX targets.
1613         (FFI_TYPE_SMALL_STRUCT2): Define.
1614         (FFI_TYPE_SMALL_STRUCT4): Likewise.
1615         (FFI_TYPE_SMALL_STRUCT8): Likewise.
1616         (FFI_TYPE_SMALL_STRUCT3): Redefine.
1617         (FFI_TYPE_SMALL_STRUCT5): Likewise.
1618         (FFI_TYPE_SMALL_STRUCT6): Likewise.
1619         (FFI_TYPE_SMALL_STRUCT7): Likewise.
1620         * src/pa/ffi.c (ROUND_DOWN): Delete.
1621         (fldw, fstw, fldd, fstd): Use '__asm__'.
1622         (ffi_struct_type): Add support for FFI_TYPE_SMALL_STRUCT2,
1623         FFI_TYPE_SMALL_STRUCT4 and FFI_TYPE_SMALL_STRUCT8.
1624         (ffi_prep_args_LINUX): Rename to ffi_prep_args_pa32. Update comment.
1625         Simplify incrementing of stack slot variable. Change type of local
1626         'n' to unsigned int.
1627         (ffi_size_stack_LINUX): Rename to ffi_size_stack_pa32. Handle long
1628         double on PA_HPUX.
1629         (ffi_prep_cif_machdep): Likewise.
1630         (ffi_call): Likewise.
1631         (ffi_closure_inner_LINUX): Rename to ffi_closure_inner_pa32. Change
1632         return type to ffi_status. Simplify incrementing of stack slot
1633         variable. Only copy floating point argument registers when PA_LINUX
1634         is true. Reformat debug statement.
1635         Add support for FFI_TYPE_SMALL_STRUCT2, FFI_TYPE_SMALL_STRUCT4 and
1636         FFI_TYPE_SMALL_STRUCT8.
1637         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Add 'extern' to
1638         declaration.
1639         (ffi_prep_closure): Make linux trampoline conditional on PA_LINUX.
1640         Add nops to cache flush.  Add trampoline for PA_HPUX.
1641         * src/pa/hpux32.S: New file.
1642         * src/pa/linux.S (ffi_call_LINUX): Rename to ffi_call_pa32. Rename
1643         ffi_prep_args_LINUX to ffi_prep_args_pa32.
1644         Localize labels. Add support for 2, 4 and 8-byte small structs. Handle
1645         unaligned destinations in 3, 5, 6 and 7-byte small structs. Order
1646         argument type checks so that common argument types appear first.
1647         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Rename
1648         ffi_closure_inner_LINUX to ffi_closure_inner_pa32.
1649
1650 2006-03-24  Alan Modra  <amodra@bigpond.net.au>
1651
1652         * src/powerpc/ffitarget.h (enum ffi_abi): Add FFI_LINUX.  Default
1653         for 32-bit using IBM extended double format.  Fix FFI_LAST_ABI.
1654         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Handle linux variant of
1655         FFI_TYPE_LONGDOUBLE.
1656         (ffi_prep_args64): Assert using IBM extended double.
1657         (ffi_prep_cif_machdep): Don't munge FFI_TYPE_LONGDOUBLE type.
1658         Handle FFI_LINUX FFI_TYPE_LONGDOUBLE return and args.
1659         (ffi_call): Handle FFI_LINUX.
1660         (ffi_closure_helper_SYSV): Non FFI_LINUX long double return needs
1661         gpr3 return pointer as for struct return.  Handle FFI_LINUX
1662         FFI_TYPE_LONGDOUBLE return and args.  Don't increment "nf"
1663         unnecessarily.
1664         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Load both f1 and f2
1665         for FFI_TYPE_LONGDOUBLE.  Move epilogue insns into case table.
1666         Don't use r6 as pointer to results, instead use sp offset.  Don't
1667         make a special call to load lr with case table address, instead
1668         use offset from previous call.
1669         * src/powerpc/sysv.S (ffi_call_SYSV): Save long double return.
1670         * src/powerpc/linux64.S (ffi_call_LINUX64): Simplify long double
1671         return.
1672
1673 2006-03-15  Kaz Kojima  <kkojima@gcc.gnu.org>
1674
1675         * src/sh64/ffi.c (ffi_prep_cif_machdep): Handle float arguments
1676         passed with FP registers correctly.
1677         (ffi_closure_helper_SYSV): Likewise.
1678         * src/sh64/sysv.S: Likewise.
1679
1680 2006-03-01  Andreas Tobler  <a.tobler@schweiz.ch>
1681
1682         * testsuite/libffi.special/unwindtest.cc (closure_test_fn): Mark cif,
1683         args and userdata unused.
1684         (closure_test_fn1): Mark cif and userdata unused.
1685         (main): Remove unused res.
1686
1687 2006-02-28  Andreas Tobler  <a.tobler@schweiz.ch>
1688
1689         * testsuite/libffi.call/call.exp: Adjust FSF address. Add test runs for
1690         -O2, -O3, -Os and the warning flags -W -Wall.
1691         * testsuite/libffi.special/special.exp: Likewise.
1692         * testsuite/libffi.call/ffitest.h: Add an __UNUSED__ macro to mark
1693         unused parameter unused for gcc or else do nothing.
1694         * testsuite/libffi.special/ffitestcxx.h: Likewise.
1695         * testsuite/libffi.call/cls_12byte.c (cls_struct_12byte_gn): Mark cif
1696         and userdata unused.
1697         * testsuite/libffi.call/cls_16byte.c (cls_struct_16byte_gn): Likewise.
1698         * testsuite/libffi.call/cls_18byte.c (cls_struct_18byte_gn): Likewise.
1699         * testsuite/libffi.call/cls_19byte.c (cls_struct_19byte_gn): Likewise.
1700         * testsuite/libffi.call/cls_1_1byte.c (cls_struct_1_1byte_gn): Likewise.
1701         * testsuite/libffi.call/cls_20byte.c (cls_struct_20byte_gn): Likewise.
1702         * testsuite/libffi.call/cls_20byte1.c (cls_struct_20byte_gn): Likewise.
1703         * testsuite/libffi.call/cls_24byte.c (cls_struct_24byte_gn): Likewise.
1704         * testsuite/libffi.call/cls_2byte.c (cls_struct_2byte_gn): Likewise.
1705         * testsuite/libffi.call/cls_3_1byte.c (cls_struct_3_1byte_gn): Likewise.
1706         * testsuite/libffi.call/cls_3byte1.c (cls_struct_3byte_gn): Likewise.
1707         * testsuite/libffi.call/cls_3byte2.c (cls_struct_3byte_gn1): Likewise.
1708         * testsuite/libffi.call/cls_4_1byte.c (cls_struct_4_1byte_gn): Likewise.
1709         * testsuite/libffi.call/cls_4byte.c (cls_struct_4byte_gn): Likewise.
1710         * testsuite/libffi.call/cls_5_1_byte.c (cls_struct_5byte_gn): Likewise.
1711         * testsuite/libffi.call/cls_5byte.c (cls_struct_5byte_gn): Likewise.
1712         * testsuite/libffi.call/cls_64byte.c (cls_struct_64byte_gn): Likewise.
1713         * testsuite/libffi.call/cls_6_1_byte.c (cls_struct_6byte_gn): Likewise.
1714         * testsuite/libffi.call/cls_6byte.c (cls_struct_6byte_gn): Likewise.
1715         * testsuite/libffi.call/cls_7_1_byte.c (cls_struct_7byte_gn): Likewise.
1716         * testsuite/libffi.call/cls_7byte.c (cls_struct_7byte_gn): Likewise.
1717         * testsuite/libffi.call/cls_8byte.c (cls_struct_8byte_gn): Likewise.
1718         * testsuite/libffi.call/cls_9byte1.c (cls_struct_9byte_gn): Likewise.
1719         * testsuite/libffi.call/cls_9byte2.c (cls_struct_9byte_gn): Likewise.
1720         * testsuite/libffi.call/cls_align_double.c (cls_struct_align_gn):
1721         Likewise.
1722         * testsuite/libffi.call/cls_align_float.c (cls_struct_align_gn):
1723         Likewise.
1724         * testsuite/libffi.call/cls_align_longdouble.c (cls_struct_align_gn):
1725         Likewise.
1726         * testsuite/libffi.call/cls_align_pointer.c (cls_struct_align_fn): Cast
1727         void* to avoid compiler warning.
1728         (main): Likewise.
1729         (cls_struct_align_gn): Mark cif and userdata unused.
1730         * testsuite/libffi.call/cls_align_sint16.c (cls_struct_align_gn):
1731         Likewise.
1732         * testsuite/libffi.call/cls_align_sint32.c (cls_struct_align_gn):
1733         Likewise.
1734         * testsuite/libffi.call/cls_align_sint64.c (cls_struct_align_gn):
1735         Likewise.
1736         * testsuite/libffi.call/cls_align_uint16.c (cls_struct_align_gn):
1737         Likewise.
1738         * testsuite/libffi.call/cls_align_uint32.c (cls_struct_align_gn):
1739         Likewise.
1740         * testsuite/libffi.call/cls_double.c (cls_ret_double_fn): Likewise.
1741         * testsuite/libffi.call/cls_float.c (cls_ret_float_fn): Likewise.
1742         * testsuite/libffi.call/cls_multi_schar.c (test_func_gn): Mark cif and
1743         data unused.
1744         (main): Cast res_call to silence gcc.
1745         * testsuite/libffi.call/cls_multi_sshort.c (test_func_gn): Mark cif and
1746         data unused.
1747         (main): Cast res_call to silence gcc.
1748         * testsuite/libffi.call/cls_multi_sshortchar.c (test_func_gn): Mark cif
1749         and data unused.
1750         (main): Cast res_call to silence gcc.
1751         * testsuite/libffi.call/cls_multi_uchar.c (test_func_gn): Mark cif and
1752         data unused.
1753         (main): Cast res_call to silence gcc.
1754         * testsuite/libffi.call/cls_multi_ushort.c (test_func_gn): Mark cif and
1755         data unused.
1756         (main): Cast res_call to silence gcc.
1757         * testsuite/libffi.call/cls_multi_ushortchar.c (test_func_gn): Mark cif
1758         and data unused.
1759         (main): Cast res_call to silence gcc.
1760         * testsuite/libffi.call/cls_schar.c (cls_ret_schar_fn): Mark cif and
1761         userdata unused.
1762         (cls_ret_schar_fn): Cast printf parameter to silence gcc.
1763         * testsuite/libffi.call/cls_sint.c (cls_ret_sint_fn): Mark cif and
1764         userdata unused.
1765         (cls_ret_sint_fn): Cast printf parameter to silence gcc.
1766         * testsuite/libffi.call/cls_sshort.c (cls_ret_sshort_fn): Mark cif and
1767         userdata unused.
1768         (cls_ret_sshort_fn): Cast printf parameter to silence gcc.
1769         * testsuite/libffi.call/cls_uchar.c (cls_ret_uchar_fn):  Mark cif and
1770         userdata unused.
1771         (cls_ret_uchar_fn): Cast printf parameter to silence gcc.
1772         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Mark cif and
1773         userdata unused.
1774         (cls_ret_uint_fn): Cast printf parameter to silence gcc.
1775         * testsuite/libffi.call/cls_ulonglong.c (cls_ret_ulonglong_fn): Mark cif
1776         and userdata unused.
1777         * testsuite/libffi.call/cls_ushort.c (cls_ret_ushort_fn): Mark cif and
1778         userdata unused.
1779         (cls_ret_ushort_fn): Cast printf parameter to silence gcc.
1780         * testsuite/libffi.call/float.c (floating): Remove unused parameter e.
1781         * testsuite/libffi.call/float1.c (main): Remove unused variable i.
1782         Cleanup white spaces.
1783         * testsuite/libffi.call/negint.c (checking): Remove unused variable i.
1784         * testsuite/libffi.call/nested_struct.c (cls_struct_combined_gn): Mark
1785         cif and userdata unused.
1786         * testsuite/libffi.call/nested_struct1.c (cls_struct_combined_gn):
1787         Likewise.
1788         * testsuite/libffi.call/nested_struct10.c (B_gn): Likewise.
1789         * testsuite/libffi.call/nested_struct2.c (B_fn): Adjust printf
1790         formatters to silence gcc.
1791         (B_gn): Mark cif and userdata unused.
1792         * testsuite/libffi.call/nested_struct3.c (B_gn): Mark cif and userdata
1793         unused.
1794         * testsuite/libffi.call/nested_struct4.c: Mention related PR.
1795         (B_gn): Mark cif and userdata unused.
1796         * testsuite/libffi.call/nested_struct5.c (B_gn): Mark cif and userdata
1797         unused.
1798         * testsuite/libffi.call/nested_struct6.c: Mention related PR.
1799         (B_gn): Mark cif and userdata unused.
1800         * testsuite/libffi.call/nested_struct7.c (B_gn): Mark cif and userdata
1801         unused.
1802         * testsuite/libffi.call/nested_struct8.c (B_gn): Likewise.
1803         * testsuite/libffi.call/nested_struct9.c (B_gn): Likewise.
1804         * testsuite/libffi.call/problem1.c (stub): Likewise.
1805         * testsuite/libffi.call/pyobjc-tc.c (main): Cast the result to silence
1806         gcc.
1807         * testsuite/libffi.call/return_fl2.c (return_fl): Add the note mentioned
1808         in the last commit for this test case in the test case itself.
1809         * testsuite/libffi.call/closure_fn0.c (closure_test_fn0): Mark cif as
1810         unused.
1811         * testsuite/libffi.call/closure_fn1.c (closure_test_fn1): Likewise.
1812         * testsuite/libffi.call/closure_fn2.c (closure_test_fn2): Likewise.
1813         * testsuite/libffi.call/closure_fn3.c (closure_test_fn3): Likewise.
1814         * testsuite/libffi.call/closure_fn4.c (closure_test_fn0): Likewise.
1815         * testsuite/libffi.call/closure_fn5.c (closure_test_fn5): Likewise.
1816         * testsuite/libffi.call/closure_fn6.c (closure_test_fn0): Likewise.
1817
1818 2006-02-22  Kaz Kojima  <kkojima@gcc.gnu.org>
1819
1820         * src/sh/sysv.S: Fix register numbers in the FDE for
1821         ffi_closure_SYSV.
1822
1823 2006-02-20  Andreas Tobler  <a.tobler@schweiz.ch>
1824
1825         * testsuite/libffi.call/return_fl2.c (return_fl): Remove static
1826         declaration to avoid a false negative on ix86. See PR323.
1827
1828 2006-02-18  Kaz Kojima  <kkojima@gcc.gnu.org>
1829
1830         * src/sh/ffi.c (ffi_closure_helper_SYSV): Remove unused variable
1831         and cast integer to void * if needed.  Update the pointer to
1832         the FP register saved area correctly.
1833
1834 2006-02-17  Andreas Tobler  <a.tobler@schweiz.ch>
1835
1836         * testsuite/libffi.call/nested_struct6.c: XFAIL this test until PR25630
1837         is fixed.
1838         * testsuite/libffi.call/nested_struct4.c: Likewise.
1839
1840 2006-02-16  Andreas Tobler  <a.tobler@schweiz.ch>
1841
1842         * testsuite/libffi.call/return_dbl.c: New test case.
1843         * testsuite/libffi.call/return_dbl1.c: Likewise.
1844         * testsuite/libffi.call/return_dbl2.c: Likewise.
1845         * testsuite/libffi.call/return_fl.c: Likewise.
1846         * testsuite/libffi.call/return_fl1.c: Likewise.
1847         * testsuite/libffi.call/return_fl2.c: Likewise.
1848         * testsuite/libffi.call/return_fl3.c: Likewise.
1849         * testsuite/libffi.call/closure_fn6.c: Likewise.
1850
1851         * testsuite/libffi.call/nested_struct2.c: Remove ffi_type_mylong
1852         definition.
1853         * testsuite/libffi.call/ffitest.h: Add ffi_type_mylong definition
1854         here to be used by other test cases too.
1855
1856         * testsuite/libffi.call/nested_struct10.c: New test case.
1857         * testsuite/libffi.call/nested_struct9.c: Likewise.
1858         * testsuite/libffi.call/nested_struct8.c: Likewise.
1859         * testsuite/libffi.call/nested_struct7.c: Likewise.
1860         * testsuite/libffi.call/nested_struct6.c: Likewise.
1861         * testsuite/libffi.call/nested_struct5.c: Likewise.
1862         * testsuite/libffi.call/nested_struct4.c: Likewise.
1863
1864 2006-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
1865
1866         * configure.ac: Enable libffi for sparc64-*-freebsd*.
1867         * configure: Rebuilt.
1868
1869 2006-01-18  Jakub Jelinek  <jakub@redhat.com>
1870
1871         * src/powerpc/sysv.S (smst_two_register): Don't call __ashldi3,
1872         instead do the shifting inline.
1873         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't compute %r5
1874         shift count unconditionally.  Simplify load sequences for 1, 2, 3, 4
1875         and 8 byte structs, for the remaining struct sizes don't call
1876         __lshrdi3, instead do the shifting inline.
1877
1878 2005-12-07  Thiemo Seufer  <ths@networkno.de>
1879
1880         * src/mips/ffitarget.h: Remove obsolete sgidefs.h include. Add
1881         missing parentheses.
1882         * src/mips/o32.S (ffi_call_O32): Code formatting. Define
1883         and use A3_OFF, FP_OFF, RA_OFF. Micro-optimizations.
1884         (ffi_closure_O32): Likewise, but with newly defined A3_OFF2,
1885         A2_OFF2, A1_OFF2, A0_OFF2, RA_OFF2, FP_OFF2, S0_OFF2, GP_OFF2,
1886         V1_OFF2, V0_OFF2, FA_1_1_OFF2, FA_1_0_OFF2, FA_0_1_OFF2,
1887         FA_0_0_OFF2.
1888         * src/mips/ffi.c (ffi_prep_args): Code formatting. Fix
1889         endianness bugs.
1890         (ffi_prep_closure): Improve trampoline instruction scheduling.
1891         (ffi_closure_mips_inner_O32): Fix endianness bugs.
1892
1893 2005-12-03  Alan Modra  <amodra@bigpond.net.au>
1894
1895         * src/powerpc/ffi.c: Formatting.
1896         (ffi_prep_args_SYSV): Avoid possible aliasing problems by using unions.
1897         (ffi_prep_args64): Likewise.
1898
1899 2005-09-30  Geoffrey Keating  <geoffk@apple.com>
1900
1901         * testsuite/lib/libffi-dg.exp (libffi_target_compile): For
1902         darwin, use -shared-libgcc not -lgcc_s, and explain why.
1903
1904 2005-09-26  Tom Tromey  <tromey@redhat.com>
1905
1906         * testsuite/libffi.call/float1.c (value_type): New typedef.
1907         (CANARY): New define.
1908         (main): Check for result buffer overflow.
1909         * src/powerpc/linux64.S: Handle linux64 long double returns.
1910         * src/powerpc/ffi.c (FLAG_RETURNS_128BITS): New constant.
1911         (ffi_prep_cif_machdep): Handle linux64 long double returns.
1912
1913 2005-08-25  Alan Modra  <amodra@bigpond.net.au>
1914
1915         PR target/23404
1916         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Correct placement of stack
1917         homed fp args.
1918         (ffi_status ffi_prep_cif_machdep): Correct stack sizing for same.
1919
1920 2005-08-11  Jakub Jelinek  <jakub@redhat.com>
1921
1922         * configure.ac (HAVE_HIDDEN_VISIBILITY_ATTRIBUTE): New test.
1923         (AH_BOTTOM): Add FFI_HIDDEN definition.
1924         * configure: Rebuilt.
1925         * fficonfig.h.in: Rebuilt.
1926         * src/powerpc/ffi.c (hidden): Remove.
1927         (ffi_closure_LINUX64, ffi_prep_args64, ffi_call_LINUX64,
1928         ffi_closure_helper_LINUX64): Use FFI_HIDDEN instead of hidden.
1929         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64,
1930         .ffi_closure_LINUX64): Use FFI_HIDDEN instead of .hidden.
1931         * src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV): Remove,
1932         add FFI_HIDDEN to its prototype.
1933         (ffi_closure_SYSV_inner): New.
1934         * src/x86/sysv.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
1935         * src/x86/win32.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
1936
1937 2005-08-10  Alfred M. Szmidt  <ams@gnu.org>
1938
1939         PR libffi/21819:
1940         * configure: Rebuilt.
1941         * configure.ac: Handle i*86-*-gnu*.
1942
1943 2005-08-09  Jakub Jelinek  <jakub@redhat.com>
1944
1945         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Use
1946         DW_CFA_offset_extended_sf rather than
1947         DW_CFA_GNU_negative_offset_extended.
1948         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise.
1949
1950 2005-07-22  SUGIOKA Toshinobu  <sugioka@itonet.co.jp>
1951
1952         * src/sh/sysv.S (ffi_call_SYSV): Stop argument popping correctly
1953         on sh3.
1954         (ffi_closure_SYSV): Change the stack layout for sh3 struct argument.
1955         * src/sh/ffi.c (ffi_prep_args): Fix sh3 argument copy, when it is
1956         partially on register.
1957         (ffi_closure_helper_SYSV): Likewise.
1958         (ffi_prep_cif_machdep): Don't set too many cif->flags.
1959
1960 2005-07-20  Kaz Kojima  <kkojima@gcc.gnu.org>
1961
1962         * src/sh/ffi.c (ffi_call): Handle small structures correctly.
1963         Remove empty line.
1964         * src/sh64/ffi.c (simple_type): Remove.
1965         (return_type): Handle small structures correctly.
1966         (ffi_prep_args): Likewise.
1967         (ffi_call): Likewise.
1968         (ffi_closure_helper_SYSV): Likewise.
1969         * src/sh64/sysv.S (ffi_call_SYSV): Handle 1, 2 and 4-byte return.
1970         Emit position independent code if PIC and remove wrong datalabel
1971         prefixes from EH data.
1972
1973 2005-07-19  Andreas Tobler  <a.tobler@schweiz.ch>
1974
1975         * Makefile.am (nodist_libffi_la_SOURCES): Add POWERPC_FREEBSD.
1976         * Makefile.in: Regenerate.
1977         * include/Makefile.in: Likewise.
1978         * testsuite/Makefile.in: Likewise.
1979         * configure.ac: Add POWERPC_FREEBSD rules.
1980         * configure: Regenerate.
1981         * src/powerpc/ffitarget.h: Add POWERPC_FREEBSD rules.
1982         (FFI_SYSV_TYPE_SMALL_STRUCT): Define.
1983         * src/powerpc/ffi.c: Add flags to handle small structure returns
1984         in ffi_call_SYSV.
1985         (ffi_prep_cif_machdep): Handle small structures for SYSV 4 ABI.
1986         Aka FFI_SYSV.
1987         (ffi_closure_helper_SYSV): Likewise.
1988         * src/powerpc/ppc_closure.S: Add return types for small structures.
1989         * src/powerpc/sysv.S: Add bits to handle small structures for
1990         final SYSV 4 ABI.
1991
1992 2005-07-10  Andreas Tobler  <a.tobler@schweiz.ch>
1993
1994         * testsuite/libffi.call/cls_5_1_byte.c: New test file.
1995         * testsuite/libffi.call/cls_6_1_byte.c: Likewise.
1996         * testsuite/libffi.call/cls_7_1_byte.c: Likewise.
1997
1998 2005-07-05  Randolph Chung  <tausq@debian.org>
1999
2000         * src/pa/ffi.c (ffi_struct_type): Rename FFI_TYPE_SMALL_STRUCT1
2001         as FFI_TYPE_SMALL_STRUCT3.  Break out handling for 5-7 byte
2002         structures.  Kill compilation warnings.
2003         (ffi_closure_inner_LINUX): Print return values as hex in debug
2004         message.  Rename FFI_TYPE_SMALL_STRUCT1 as FFI_TYPE_SMALL_STRUCT3.
2005         Properly handle 5-7 byte structure returns.
2006         * src/pa/ffitarget.h (FFI_TYPE_SMALL_STRUCT1)
2007         (FFI_TYPE_SMALL_STRUCT2): Remove.
2008         (FFI_TYPE_SMALL_STRUCT3, FFI_TYPE_SMALL_STRUCT5)
2009         (FFI_TYPE_SMALL_STRUCT6, FFI_TYPE_SMALL_STRUCT7): Define.
2010         * src/pa/linux.S: Mark source file as using PA1.1 assembly.
2011         (checksmst1, checksmst2): Remove.
2012         (checksmst3): Optimize handling of 3-byte struct returns.
2013         (checksmst567): Properly handle 5-7 byte struct returns.
2014
2015 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2016
2017         PR libgcj/21943
2018         * src/mips/n32.S: Enforce PIC code.
2019         * src/mips/o32.S: Likewise.
2020
2021 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2022
2023         * configure.ac: Treat i*86-*-solaris2.10 and up as X86_64.
2024         * configure: Regenerate.
2025
2026 2005-06-01  Alan Modra  <amodra@bigpond.net.au>
2027
2028         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't use JUMPTARGET
2029         to call ffi_closure_helper_SYSV.  Append @local instead.
2030         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise for ffi_prep_args_SYSV.
2031
2032 2005-05-17  Kelley Cook  <kcook@gcc.gnu.org>
2033
2034         * configure.ac: Use AC_C_BIGENDIAN instead of AC_C_BIGENDIAN_CROSS.
2035         Use AC_CHECK_SIZEOF instead of AC_COMPILE_CHECK_SIZEOF.
2036         * Makefile.am (ACLOCAL_AMFLAGS): Remove -I ../config.
2037         * aclocal.m4, configure, fficonfig.h.in, Makefile.in,
2038         include/Makefile.in, testsuite/Makefile.in: Regenerate.
2039
2040 2005-05-09  Mike Stump  <mrs@apple.com>
2041
2042         * configure: Regenerate.
2043
2044 2005-05-08  Richard Henderson  <rth@redhat.com>
2045
2046         PR libffi/21285
2047         * src/alpha/osf.S: Update unwind into to match code.
2048
2049 2005-05-04  Andreas Degert <ad@papyrus-gmbh.de>
2050             Richard Henderson  <rth@redhat.com>
2051
2052         * src/x86/ffi64.c (ffi_prep_cif_machdep): Save sse-used flag in
2053         bit 11 of flags.
2054         (ffi_call): Mask return type field.  Pass ssecount to ffi_call_unix64.
2055         (ffi_prep_closure): Set carry bit if sse-used flag set.
2056         * src/x86/unix64.S (ffi_call_unix64): Add ssecount argument.
2057         Only load sse registers if ssecount non-zero.
2058         (ffi_closure_unix64): Only save sse registers if carry set on entry.
2059
2060 2005-04-29  Ralf Corsepius  <ralf.corsepius@rtems.org>
2061
2062         * configure.ac: Add i*86-*-rtems*, sparc*-*-rtems*,
2063         powerpc-*rtems*, arm*-*-rtems*, sh-*-rtems*.
2064         * configure: Regenerate.
2065
2066 2005-04-20  Hans-Peter Nilsson  <hp@axis.com>
2067
2068         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): In regsub use,
2069         have Tcl8.3-compatible intermediate variable.
2070
2071 2005-04-18  Simon Posnjak <simon.posnjak@siol.net>
2072             Hans-Peter Nilsson  <hp@axis.com>
2073
2074         * Makefile.am: Add CRIS support.
2075         * configure.ac: Likewise.
2076         * Makefile.in, configure, testsuite/Makefile.in,
2077         include/Makefile.in: Regenerate.
2078         * src/cris: New directory.
2079         * src/cris/ffi.c, src/cris/sysv.S, src/cris/ffitarget.h: New files.
2080         * src/prep_cif.c (ffi_prep_cif): Wrap in #ifndef __CRIS__.
2081
2082         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): Replace \n with
2083         \r?\n in output tests.
2084
2085 2005-04-12  Mike Stump  <mrs@apple.com>
2086
2087         * configure: Regenerate.
2088
2089 2005-03-30  Hans Boehm  <Hans.Boehm@hp.com>
2090
2091         * src/ia64/ffitarget.h (ffi_arg): Use long long instead of DI.
2092
2093 2005-03-30  Steve Ellcey  <sje@cup.hp.com>
2094
2095         * src/ia64/ffitarget.h (ffi_arg) ADD DI attribute.
2096         (ffi_sarg) Ditto.
2097         * src/ia64/unix.S (ffi_closure_unix): Extend gp
2098         to 64 bits in ILP32 mode.
2099         Load 64 bits even for short data.
2100
2101 2005-03-23  Mike Stump  <mrs@apple.com>
2102
2103         * src/powerpc/darwin.S: Update for -m64 multilib.
2104         * src/powerpc/darwin_closure.S: Likewise.
2105
2106 2005-03-21  Zack Weinberg  <zack@codesourcery.com>
2107
2108         * configure.ac: Do not invoke TL_AC_GCC_VERSION.
2109         Do not set tool_include_dir.
2110         * aclocal.m4, configure, Makefile.in, testsuite/Makefile.in:
2111         Regenerate.
2112         * include/Makefile.am: Set gcc_version and toollibffidir.
2113         * include/Makefile.in: Regenerate.
2114
2115 2005-02-22  Andrew Haley  <aph@redhat.com>
2116
2117         * src/powerpc/ffi.c (ffi_prep_cif_machdep): Bump alignment to
2118         odd-numbered register pairs for 64-bit integer types.
2119
2120 2005-02-23  Andreas Tobler  <a.tobler@schweiz.ch>
2121
2122         PR libffi/20104
2123         * testsuite/libffi.call/return_ll1.c: New test case.
2124
2125 2005-02-11  Janis Johnson  <janis187@us.ibm.com>
2126
2127         * testsuite/libffi.call/cls_align_longdouble.c: Remove dg-options.
2128         * testsuite/libffi.call/float.c: Ditto.
2129         * testsuite/libffi.call/float2.c: Ditto.
2130         * testsuite/libffi.call/float3.c: Ditto.
2131
2132 2005-02-08  Andreas Tobler  <a.tobler@schweiz.ch>
2133
2134         * src/frv/ffitarget.h: Remove PPC stuff which does not belong to frv.
2135
2136 2005-01-12  Eric Botcazou  <ebotcazou@libertysurf.fr>
2137
2138         * testsuite/libffi.special/special.exp (cxx_options): Add
2139         -shared-libgcc.
2140
2141 2004-12-31  Richard Henderson  <rth@redhat.com>
2142
2143         * src/types.c (FFI_AGGREGATE_TYPEDEF): Remove.
2144         (FFI_TYPEDEF): Rename from FFI_INTEGRAL_TYPEDEF.  Replace size and
2145         offset parameters with a type parameter; deduce size and structure
2146         alignment.  Update all users.
2147
2148 2004-12-31  Richard Henderson  <rth@redhat.com>
2149
2150         * src/types.c (FFI_TYPE_POINTER): Define with sizeof.
2151         (FFI_TYPE_LONGDOUBLE): Fix for ia64.
2152         * src/ia64/ffitarget.h (struct ffi_ia64_trampoline_struct): Move
2153         into ffi_prep_closure.
2154         * src/ia64/ia64_flags.h, src/ia64/ffi.c, src/ia64/unix.S: Rewrite
2155         from scratch.
2156
2157 2004-12-27  Richard Henderson  <rth@redhat.com>
2158
2159         * src/x86/unix64.S: Fix typo in unwind info.
2160
2161 2004-12-25  Richard Henderson  <rth@redhat.com>
2162
2163         * src/x86/ffi64.c (struct register_args): Rename from stackLayout.
2164         (enum x86_64_reg_class): Add X86_64_COMPLEX_X87_CLASS.
2165         (merge_classes): Check for it.
2166         (SSE_CLASS_P): New.
2167         (classify_argument): Pass byte_offset by value; perform all updates
2168         inside struct case.
2169         (examine_argument): Add classes argument; handle
2170         X86_64_COMPLEX_X87_CLASS.
2171         (ffi_prep_args): Merge into ...
2172         (ffi_call): ... here.  Share stack frame with ffi_call_unix64.
2173         (ffi_prep_cif_machdep): Setup cif->flags for proper structure return.
2174         (ffi_fill_return_value): Remove.
2175         (ffi_prep_closure): Remove dead assert.
2176         (ffi_closure_unix64_inner): Rename from ffi_closure_UNIX64_inner.
2177         Rewrite to use struct register_args instead of va_list.  Create
2178         flags for handling structure returns.
2179         * src/x86/unix64.S: Remove dead strings.
2180         (ffi_call_unix64): Rename from ffi_call_UNIX64.  Rewrite to share
2181         stack frame with ffi_call.  Handle structure returns properly.
2182         (float2sse, floatfloat2sse, double2sse): Remove.
2183         (sse2float, sse2double, sse2floatfloat): Remove.
2184         (ffi_closure_unix64): Rename from ffi_closure_UNIX64.  Rewrite
2185         to handle structure returns properly.
2186
2187 2004-12-08  David Edelsohn  <edelsohn@gnu.org>
2188
2189         * Makefile.am (AM_MAKEFLAGS): Remove duplicate LIBCFLAGS and
2190         PICFLAG.
2191         * Makefile.in: Regenerated.
2192
2193 2004-12-02  Richard Sandiford  <rsandifo@redhat.com>
2194
2195         * configure.ac: Use TL_AC_GCC_VERSION to set gcc_version.
2196         * configure, aclocal.m4, Makefile.in: Regenerate.
2197         * include/Makefile.in, testsuite/Makefile.in: Regenerate.
2198
2199 2004-11-29  Kelley Cook  <kcook@gcc.gnu.org>
2200
2201         * configure: Regenerate for libtool change.
2202
2203 2004-11-25  Kelley Cook  <kcook@gcc.gnu.org>
2204
2205         * configure: Regenerate for libtool reversion.
2206
2207 2004-11-24  Kelley Cook  <kcook@gcc.gnu.org>
2208
2209         * configure: Regenerate for libtool change.
2210
2211 2004-11-23  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2212
2213         * testsuite/lib/libffi-dg.exp: Use new procs in target-libpath.exp.
2214
2215 2004-11-23  Richard Sandiford  <rsandifo@redhat.com>
2216
2217         * src/mips/o32.S (ffi_call_O32, ffi_closure_O32): Use jalr instead
2218         of jal.  Use an absolute encoding for the frame information.
2219
2220 2004-11-23  Kelley Cook  <kcook@gcc.gnu.org>
2221
2222         * Makefile.am: Remove no-dependencies.  Add ACLOCAL_AMFLAGS.
2223         * acinclude.m4: Delete logic for sincludes.
2224         * aclocal.m4, Makefile.in, configure: Regenerate.
2225         * include/Makefile: Likewise.
2226         * testsuite/Makefile: Likewise.
2227
2228 2004-11-22  Eric Botcazou  <ebotcazou@libertysurf.fr>
2229
2230         * src/sparc/ffi.c (ffi_prep_closure): Align doubles and 64-bit integers
2231         on a 8-byte boundary.
2232         * src/sparc/v8.S (ffi_closure_v8): Reserve frame space for arguments.
2233
2234 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2235
2236         * src/arm/ffi.c (ffi_prep_cif_machdep): Handle functions that return
2237         long long values.  Round stack allocation to a multiple of 8 bytes
2238         for ATPCS compatibility.
2239         * src/arm/sysv.S (ffi_call_SYSV): Rework to avoid use of APCS register
2240         names.  Handle returning long long types.  Add Thumb and interworking
2241         support.  Improve soft-float code.
2242
2243 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2244
2245         * testsuite/lib/libffi-db.exp (load_gcc_lib): New function.
2246         (libffi_exit): New function.
2247         (libffi_init): Build the testglue wrapper if needed.
2248
2249 2004-10-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2250
2251         PR other/18138
2252         * testsuite/lib/libffi-dg.exp: Accept more than one multilib libgcc.
2253
2254 2004-10-25  Kazuhiro Inaoka <inaoka.kazuhiro@renesas.com>
2255
2256         * src/m32r/libffitarget.h (FFI_CLOSURES): Set to 0.
2257
2258 2004-10-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2259
2260         * src/sh/sysv.S (ffi_call_SYSV): Don't align for double data.
2261         * testsuite/libffi.call/float3.c: New test case.
2262
2263 2004-10-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2264
2265         * src/sh/ffi.c (ffi_prep_closure): Set T bit in trampoline for
2266         the function returning a structure pointed with R2.
2267         * src/sh/sysv.S (ffi_closure_SYSV): Use R2 as the pointer to
2268         the structure return value if T bit set.  Emit position
2269         independent code and EH data if PIC.
2270
2271 2004-10-13  Kazuhiro Inaoka  <inaoka.kazuhiro@renesas.com>
2272
2273         * Makefile.am: Add m32r support.
2274         * configure.ac: Likewise.
2275         * Makefile.in: Regenerate.
2276         * confiugre: Regenerate.
2277         * src/types.c: Add m32r port to FFI_INTERNAL_TYPEDEF
2278         (uint64, sint64, double, longdouble)
2279         * src/m32r: New directory.
2280         * src/m32r/ffi.c: New file.
2281         * src/m32r/sysv.S: Likewise.
2282         * src/m32r/ffitarget.h: Likewise.
2283
2284 2004-10-02  Kaz Kojima  <kkojima@gcc.gnu.org>
2285
2286         * testsuite/libffi.call/negint.c: New test case.
2287
2288 2004-09-14  H.J. Lu  <hongjiu.lu@intel.com>
2289
2290         PR libgcj/17465
2291         * testsuite/lib/libffi-dg.exp: Don't use global ld_library_path.
2292         Set up LD_LIBRARY_PATH, SHLIB_PATH, LD_LIBRARYN32_PATH,
2293         LD_LIBRARY64_PATH, LD_LIBRARY_PATH_32, LD_LIBRARY_PATH_64 and
2294         DYLD_LIBRARY_PATH.
2295
2296 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2297
2298         * testsuite/libffi.call/many_win32.c: Remove whitespaces.
2299         * testsuite/libffi.call/promotion.c: Likewise.
2300         * testsuite/libffi.call/return_ll.c: Remove unused var. Cleanup
2301         whitespaces.
2302         * testsuite/libffi.call/return_sc.c: Likewise.
2303         * testsuite/libffi.call/return_uc.c: Likewise.
2304
2305 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2306
2307         * src/powerpc/darwin.S: Fix comments and identation.
2308         * src/powerpc/darwin_closure.S: Likewise.
2309
2310 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2311
2312         * src/powerpc/ffi_darwin.c: Add flag for longdouble return values.
2313         (ffi_prep_args): Handle longdouble arguments.
2314         (ffi_prep_cif_machdep): Set flags for longdouble. Calculate space for
2315         longdouble.
2316         (ffi_closure_helper_DARWIN): Add closure handling for longdouble.
2317         * src/powerpc/darwin.S (_ffi_call_DARWIN): Add handling of longdouble
2318         values.
2319         * src/powerpc/darwin_closure.S (_ffi_closure_ASM): Likewise.
2320         * src/types.c: Defined longdouble size and alignment for darwin.
2321
2322 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2323
2324         * src/powerpc/aix.S: Remove whitespaces.
2325         * src/powerpc/aix_closure.S: Likewise.
2326         * src/powerpc/asm.h: Likewise.
2327         * src/powerpc/ffi.c: Likewise.
2328         * src/powerpc/ffitarget.h: Likewise.
2329         * src/powerpc/linux64.S: Likewise.
2330         * src/powerpc/linux64_closure.S: Likewise.
2331         * src/powerpc/ppc_closure.S: Likewise.
2332         * src/powerpc/sysv.S: Likewise.
2333
2334 2004-08-30  Anthony Green  <green@redhat.com>
2335
2336         * Makefile.am: Add frv support.
2337         * Makefile.in, testsuite/Makefile.in: Rebuilt.
2338         * configure.ac: Read configure.host.
2339         * configure.in: Read configure.host.
2340         * configure.host: New file.  frv-elf needs libgloss.
2341         * include/ffi.h.in: Force ffi_closure to have a nice big (8)
2342         alignment.  This is needed to frv and shouldn't harm the others.
2343         * include/ffi_common.h (ALIGN_DOWN): New macro.
2344         * src/frv/ffi.c, src/frv/ffitarget.h, src/frv/eabi.S: New files.
2345
2346 2004-08-24  David Daney  <daney@avtrex.com>
2347
2348         * testsuite/libffi.call/closure_fn0.c: Xfail mips64* instead of mips*.
2349         * testsuite/libffi.call/closure_fn1.c: Likewise.
2350         * testsuite/libffi.call/closure_fn2.c  Likewise.
2351         * testsuite/libffi.call/closure_fn3.c: Likewise.
2352         * testsuite/libffi.call/closure_fn4.c: Likewise.
2353         * testsuite/libffi.call/closure_fn5.c: Likewise.
2354         * testsuite/libffi.call/cls_18byte.c: Likewise.
2355         * testsuite/libffi.call/cls_19byte.c: Likewise.
2356         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2357         * testsuite/libffi.call/cls_20byte.c: Likewise.
2358         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2359         * testsuite/libffi.call/cls_24byte.c: Likewise.
2360         * testsuite/libffi.call/cls_2byte.c: Likewise.
2361         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2362         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2363         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2364         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2365         * testsuite/libffi.call/cls_4byte.c: Likewise.
2366         * testsuite/libffi.call/cls_64byte.c: Likewise.
2367         * testsuite/libffi.call/cls_6byte.c: Likewise.
2368         * testsuite/libffi.call/cls_7byte.c: Likewise.
2369         * testsuite/libffi.call/cls_8byte.c: Likewise.
2370         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2371         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2372         * testsuite/libffi.call/cls_align_double.c: Likewise.
2373         * testsuite/libffi.call/cls_align_float.c: Likewise.
2374         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2375         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2376         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2377         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2378         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2379         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2380         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2381         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
2382         * testsuite/libffi.call/cls_double.c: Likewise.
2383         * testsuite/libffi.call/cls_float.c: Likewise.
2384         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2385         * testsuite/libffi.call/cls_multi_sshort.c: Likewise.
2386         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2387         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2388         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2389         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2390         * testsuite/libffi.call/cls_schar.c: Likewise.
2391         * testsuite/libffi.call/cls_sint.c: Likewise.
2392         * testsuite/libffi.call/cls_sshort.c: Likewise.
2393         * testsuite/libffi.call/cls_uchar.c: Likewise.
2394         * testsuite/libffi.call/cls_uint.c: Likewise.
2395         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2396         * testsuite/libffi.call/cls_ushort.c: Likewise.
2397         * testsuite/libffi.call/nested_struct.c: Likewise.
2398         * testsuite/libffi.call/nested_struct1.c: Likewise.
2399         * testsuite/libffi.call/nested_struct2.c: Likewise.
2400         * testsuite/libffi.call/nested_struct3.c: Likewise.
2401         * testsuite/libffi.call/problem1.c: Likewise.
2402         * testsuite/libffi.special/unwindtest.cc: Likewise.
2403         * testsuite/libffi.call/cls_12byte.c: Likewise and set return value
2404         to zero.
2405         * testsuite/libffi.call/cls_16byte.c: Likewise.
2406         * testsuite/libffi.call/cls_5byte.c: Likewise.
2407
2408 2004-08-23  David Daney <daney@avtrex.com>
2409
2410         PR libgcj/13141
2411         * src/mips/ffitarget.h (FFI_O32_SOFT_FLOAT): New ABI.
2412         * src/mips/ffi.c (ffi_prep_args): Fix alignment calculation.
2413         (ffi_prep_cif_machdep): Handle FFI_O32_SOFT_FLOAT floating point
2414         parameters and return types.
2415         (ffi_call): Handle FFI_O32_SOFT_FLOAT ABI.
2416         (ffi_prep_closure): Ditto.
2417         (ffi_closure_mips_inner_O32): Handle FFI_O32_SOFT_FLOAT ABI, fix
2418         alignment calculations.
2419         * src/mips/o32.S (ffi_closure_O32): Don't use floating point
2420         instructions if FFI_O32_SOFT_FLOAT, make stack frame ABI compliant.
2421
2422 2004-08-14  Casey Marshall <csm@gnu.org>
2423
2424         * src/mips/ffi.c (ffi_pref_cif_machdep): set `cif->flags' to
2425         contain `FFI_TYPE_UINT64' as return type for any 64-bit
2426         integer (O32 ABI only).
2427         (ffi_prep_closure): new function.
2428         (ffi_closure_mips_inner_O32): new function.
2429         * src/mips/ffitarget.h: Define `FFI_CLOSURES' and
2430         `FFI_TRAMPOLINE_SIZE' appropriately if the ABI is o32.
2431         * src/mips/o32.S (ffi_call_O32): add labels for .eh_frame. Return
2432         64 bit integers correctly.
2433         (ffi_closure_O32): new function.
2434         Added DWARF-2 unwind info for both functions.
2435
2436 2004-08-10  Andrew Haley  <aph@redhat.com>
2437
2438         * src/x86/ffi64.c (ffi_prep_args ): 8-align all stack arguments.
2439
2440 2004-08-01  Robert Millan  <robertmh@gnu.org>
2441
2442         * configure.ac: Detect knetbsd-gnu and kfreebsd-gnu.
2443         * configure: Regenerate.
2444
2445 2004-07-30  Maciej W. Rozycki  <macro@linux-mips.org>
2446
2447         * acinclude.m4 (AC_FUNC_MMAP_BLACKLIST): Check for <sys/mman.h>
2448         and mmap() explicitly instead of relying on preset autoconf cache
2449         variables.
2450         * aclocal.m4: Regenerate.
2451         * configure: Regenerate.
2452
2453 2004-07-11  Ulrich Weigand  <uweigand@de.ibm.com>
2454
2455         * src/s390/ffi.c (ffi_prep_args): Fix C aliasing violation.
2456         (ffi_check_float_struct): Remove unused prototype.
2457
2458 2004-06-30  Geoffrey Keating  <geoffk@apple.com>
2459
2460         * src/powerpc/ffi_darwin.c (flush_icache): ';' is a comment
2461         character on Darwin, use '\n\t' instead.
2462
2463 2004-06-26  Matthias Klose  <doko@debian.org>
2464
2465         * libtool-version: Fix typo in revision/age.
2466
2467 2004-06-17  Matthias Klose  <doko@debian.org>
2468
2469         * libtool-version: New.
2470         * Makefile.am (libffi_la_LDFLAGS): Use -version-info for soname.
2471         * Makefile.in: Regenerate.
2472
2473 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2474
2475         * Makefile.am: Remove useless multilib rules.
2476         * Makefile.in: Regenerate.
2477         * aclocal.m4: Regenerate with automake 1.8.5.
2478         * configure.ac: Remove useless multilib configury.
2479         * configure: Regenerate.
2480
2481 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2482
2483         * .cvsignore: New file.
2484
2485 2004-06-10  Jakub Jelinek  <jakub@redhat.com>
2486
2487         * src/ia64/unix.S (ffi_call_unix): Insert group barrier break
2488         fp_done.
2489         (ffi_closure_UNIX): Fix f14/f15 adjustment if FLOAT_SZ is ever
2490         changed from 8.
2491
2492 2004-06-06  Sean McNeil  <sean@mcneil.com>
2493
2494         * configure.ac: Add x86_64-*-freebsd* support.
2495         * configure: Regenerate.
2496
2497 2004-04-26  Joe Buck <jbuck@welsh-buck.org>
2498
2499         Bug 15093
2500         * configure.ac: Test for existence of mmap and sys/mman.h before
2501         checking blacklist.  Fix suggested by Jim Wilson.
2502         * configure: Regenerate.
2503
2504 2004-04-26  Matt Austern  <austern@apple.com>
2505
2506         * src/powerpc/darwin.S: Go through a non-lazy pointer for initial
2507         FDE location.
2508         * src/powerpc/darwin_closure.S: Likewise.
2509
2510 2004-04-24  Andreas Tobler  <a.tobler@schweiz.ch>
2511
2512         * testsuite/libffi.call/cls_multi_schar.c (main): Fix initialization
2513         error. Reported by Thomas Heller <theller@python.net>.
2514         * testsuite/libffi.call/cls_multi_sshort.c (main): Likewise.
2515         * testsuite/libffi.call/cls_multi_ushort.c (main): Likewise.
2516
2517 2004-03-20  Matthias Klose  <doko@debian.org>
2518
2519         * src/pa/linux.S: Fix typo.
2520
2521 2004-03-19  Matthias Klose  <doko@debian.org>
2522
2523         * Makefile.am: Update.
2524         * Makefile.in: Regenerate.
2525         * src/pa/ffi.h.in: Remove.
2526         * src/pa/ffitarget.h: New file.
2527
2528 2004-02-10  Randolph Chung  <tausq@debian.org>
2529
2530         * Makefile.am: Add PA support.
2531         * Makefile.in: Regenerate.
2532         * include/Makefile.in: Regenerate.
2533         * configure.ac: Add PA target.
2534         * configure: Regenerate.
2535         * src/pa/ffi.c: New file.
2536         * src/pa/ffi.h.in: Add PA support.
2537         * src/pa/linux.S: New file.
2538         * prep_cif.c: Add PA support.
2539
2540 2004-03-16  Hosaka Yuji  <hos@tamanegi.org>
2541
2542         * src/types.c: Fix alignment size of X86_WIN32 case int64 and
2543         double.
2544         * src/x86/ffi.c (ffi_prep_args): Replace ecif->cif->rtype->type
2545         with ecif->cif->flags.
2546         (ffi_call, ffi_prep_incoming_args_SYSV): Replace cif->rtype->type
2547         with cif->flags.
2548         (ffi_prep_cif_machdep): Add X86_WIN32 struct case.
2549         (ffi_closure_SYSV): Add 1 or 2-bytes struct case for X86_WIN32.
2550         * src/x86/win32.S (retstruct1b, retstruct2b, sc_retstruct1b,
2551         sc_retstruct2b): Add for 1 or 2-bytes struct case.
2552
2553 2004-03-15 Kelley Cook <kcook@gcc.gnu.org>
2554
2555         * configure.in: Rename file to ...
2556         * configure.ac: ... this.
2557         * fficonfig.h.in: Regenerate.
2558         * Makefile.in: Regenerate.
2559         * include/Makefile.in: Regenerate.
2560         * testsuite/Makefile.in: Regenerate.
2561
2562 2004-03-12  Matt Austern  <austern@apple.com>
2563
2564         * src/powerpc/darwin.S: Fix EH information so it corresponds to
2565         changes in EH format resulting from addition of linkonce support.
2566         * src/powerpc/darwin_closure.S: Likewise.
2567
2568 2004-03-11  Andreas Tobler  <a.tobler@schweiz.ch>
2569             Paolo Bonzini  <bonzini@gnu.org>
2570
2571         * Makefile.am (AUTOMAKE_OPTIONS): Set them.
2572         Remove VPATH. Remove rules for object files. Remove multilib support.
2573         (AM_CCASFLAGS): Add.
2574         * configure.in (AC_CONFIG_HEADERS): Relace AM_CONFIG_HEADER.
2575         (AC_PREREQ): Bump version to 2.59.
2576         (AC_INIT): Fill with version info and bug address.
2577         (ORIGINAL_LD_FOR_MULTILIBS): Remove.
2578         (AM_ENABLE_MULTILIB): Use this instead of AC_ARG_ENABLE.
2579         De-precious CC so that the right flags are passed down to multilibs.
2580         (AC_MSG_ERROR): Replace obsolete macro AC_ERROR.
2581         (AC_CONFIG_FILES): Replace obsolete macro AC_LINK_FILES.
2582         (AC_OUTPUT): Reorganize the output with AC_CONFIG_COMMANDS.
2583         * configure: Rebuilt.
2584         * aclocal.m4: Likewise.
2585         * Makefile.in, include/Makefile.in, testsuite/Makefile.in: Likewise.
2586         * fficonfig.h.in: Likewise.
2587
2588 2004-03-11  Andreas Schwab  <schwab@suse.de>
2589
2590         * src/ia64/ffi.c (ffi_prep_incoming_args_UNIX): Get floating point
2591         arguments from fp registers only for the first 8 parameter slots.
2592         Don't convert a float parameter when passed in memory.
2593
2594 2004-03-09  Hans-Peter Nilsson  <hp@axis.com>
2595
2596         * configure: Regenerate for config/accross.m4 correction.
2597
2598 2004-02-25  Matt Kraai  <kraai@alumni.cmu.edu>
2599
2600         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Change
2601         ecif->cif->bytes to bytes.
2602         (ffi_prep_cif_machdep): Add braces around nested if statement.
2603
2604 2004-02-09  Alan Modra  <amodra@bigpond.net.au>
2605
2606         * src/types.c (pointer): POWERPC64 has 8 byte pointers.
2607
2608         * src/powerpc/ffi.c (ffi_prep_args64): Correct long double handling.
2609         (ffi_closure_helper_LINUX64): Fix typo.
2610         * testsuite/libffi.call/cls_align_longdouble.c: Pass -mlong-double-128
2611         for powerpc64-*-*.
2612         * testsuite/libffi.call/float.c: Likewise.
2613         * testsuite/libffi.call/float2.c: Likewise.
2614
2615 2004-02-08  Alan Modra  <amodra@bigpond.net.au>
2616
2617         * src/powerpc/ffi.c (ffi_prep_cif_machdep <FFI_LINUX64>): Correct
2618         long double function return and long double arg handling.
2619         (ffi_closure_helper_LINUX64): Formatting.  Delete unused "ng" var.
2620         Use "end_pfr" instead of "nf".  Correct long double handling.
2621         Localise "temp".
2622         * src/powerpc/linux64.S (ffi_call_LINUX64): Save f2 long double
2623         return value.
2624         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Allocate
2625         space for long double return value.  Adjust stack frame and offsets.
2626         Load f2 long double return.
2627
2628 2004-02-07  Alan Modra  <amodra@bigpond.net.au>
2629
2630         * src/types.c: Use 16 byte long double for POWERPC64.
2631
2632 2004-01-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2633
2634         * src/sparc/ffi.c (ffi_prep_args_v9): Shift the parameter array
2635         when the structure return address is passed in %o0.
2636         (ffi_V9_return_struct): Rename into ffi_v9_layout_struct.
2637         (ffi_v9_layout_struct): Align the field following a nested structure
2638         on a word boundary.  Use memmove instead of memcpy.
2639         (ffi_call): Update call to ffi_V9_return_struct.
2640         (ffi_prep_closure): Define 'ctx' only for V8.
2641         (ffi_closure_sparc_inner): Clone into ffi_closure_sparc_inner_v8
2642         and ffi_closure_sparc_inner_v9.
2643         (ffi_closure_sparc_inner_v8): Return long doubles by reference.
2644         Always skip the structure return address.  For structures and long
2645         doubles, copy the argument directly.
2646         (ffi_closure_sparc_inner_v9): Skip the structure return address only
2647         if required.  Shift the maximum floating-point slot accordingly.  For
2648         big structures, copy the argument directly; otherwise, left-justify the
2649         argument and call ffi_v9_layout_struct to lay out the structure on
2650         the stack.
2651         * src/sparc/v8.S: Undef STACKFRAME before defining it.
2652         (ffi_closure_v8): Pass the structure return address.  Update call to
2653         ffi_closure_sparc_inner_v8.  Short-circuit FFI_TYPE_INT handling.
2654         Skip the 'unimp' insn when returning long doubles and structures.
2655         * src/sparc/v9.S: Undef STACKFRAME before defining it.
2656         (ffi_closure_v9): Increase the frame size by 2 words.  Short-circuit
2657         FFI_TYPE_INT handling.  Load structures both in integers and
2658         floating-point registers on return.
2659         * README: Update status of the SPARC port.
2660
2661 2004-01-24  Andreas Tobler  <a.tobler@schweiz.ch>
2662
2663         * testsuite/libffi.call/pyobjc-tc.c (main): Treat result value
2664         as of type ffi_arg.
2665         * testsuite/libffi.call/struct3.c (main): Fix CHECK.
2666
2667 2004-01-22  Ulrich Weigand  <uweigand@de.ibm.com>
2668
2669         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Treat result
2670         value as of type ffi_arg, not unsigned int.
2671
2672 2004-01-21  Michael Ritzert  <ritzert@t-online.de>
2673
2674         * ffi64.c (ffi_prep_args): Cast the RHS of an assignment instead
2675         of the LHS.
2676
2677 2004-01-12  Andreas Tobler  <a.tobler@schweiz.ch>
2678
2679         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_32 for
2680         Solaris.
2681
2682 2004-01-08  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2683
2684         * testsuite/libffi.call/ffitest.h (allocate_mmap): Cast MAP_FAILED
2685         to void *.
2686
2687 2003-12-10  Richard Henderson  <rth@redhat.com>
2688
2689         * testsuite/libffi.call/cls_align_pointer.c: Cast pointers to
2690         size_t instead of int.
2691
2692 2003-12-04  Hosaka Yuji  <hos@tamanegi.org>
2693
2694         * testsuite/libffi.call/many_win32.c: Include <float.h>.
2695         * testsuite/libffi.call/many_win32.c (main): Replace variable
2696         int i with unsigned long ul.
2697
2698         * testsuite/libffi.call/cls_align_uint64.c: New test case.
2699         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2700         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2701         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2702         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2703         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2704         * testsuite/libffi.call/cls_align_float.c: Likewise.
2705         * testsuite/libffi.call/cls_align_double.c: Likewise.
2706         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2707         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2708
2709 2003-12-02  Hosaka Yuji  <hos@tamanegi.org>
2710
2711         PR other/13221
2712         * src/x86/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV):
2713         Align arguments to 32 bits.
2714
2715 2003-12-01  Andreas Tobler  <a.tobler@schweiz.ch>
2716
2717         PR other/13221
2718         * testsuite/libffi.call/cls_multi_sshort.c: New test case.
2719         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2720         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2721         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2722         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2723         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2724
2725         * testsuite/libffi.special/unwindtest.cc: Cosmetics.
2726
2727 2003-11-26  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
2728
2729         * testsuite/libffi.call/ffitest.h: Include <fcntl.h>.
2730         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2731
2732 2003-11-22  Andreas Tobler  <a.tobler@schweiz.ch>
2733
2734         * Makefile.in: Rebuilt.
2735         * configure: Likewise.
2736         * testsuite/libffi.special/unwindtest.cc: Convert the mmap to
2737         the right type.
2738
2739 2003-11-21  Andreas Jaeger  <aj@suse.de>
2740             Andreas Tobler  <a.tobler@schweiz.ch>
2741
2742         * acinclude.m4: Add AC_FUNC_MMAP_BLACKLIST.
2743         * configure.in: Call AC_FUNC_MMAP_BLACKLIST.
2744         * Makefile.in: Rebuilt.
2745         * aclocal.m4: Likewise.
2746         * configure: Likewise.
2747         * fficonfig.h.in: Likewise.
2748         * testsuite/lib/libffi-dg.exp: Add include dir.
2749         * testsuite/libffi.call/ffitest.h: Add MMAP definitions.
2750         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2751         * testsuite/libffi.call/closure_fn0.c: Use MMAP functionality
2752         for ffi_closure if available.
2753         * testsuite/libffi.call/closure_fn1.c: Likewise.
2754         * testsuite/libffi.call/closure_fn2.c: Likewise.
2755         * testsuite/libffi.call/closure_fn3.c: Likewise.
2756         * testsuite/libffi.call/closure_fn4.c: Likewise.
2757         * testsuite/libffi.call/closure_fn5.c: Likewise.
2758         * testsuite/libffi.call/cls_12byte.c: Likewise.
2759         * testsuite/libffi.call/cls_16byte.c: Likewise.
2760         * testsuite/libffi.call/cls_18byte.c: Likewise.
2761         * testsuite/libffi.call/cls_19byte.c: Likewise.
2762         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2763         * testsuite/libffi.call/cls_20byte.c: Likewise.
2764         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2765         * testsuite/libffi.call/cls_24byte.c: Likewise.
2766         * testsuite/libffi.call/cls_2byte.c: Likewise.
2767         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2768         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2769         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2770         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2771         * testsuite/libffi.call/cls_4byte.c: Likewise.
2772         * testsuite/libffi.call/cls_5byte.c: Likewise.
2773         * testsuite/libffi.call/cls_64byte.c: Likewise.
2774         * testsuite/libffi.call/cls_6byte.c: Likewise.
2775         * testsuite/libffi.call/cls_7byte.c: Likewise.
2776         * testsuite/libffi.call/cls_8byte.c: Likewise.
2777         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2778         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2779         * testsuite/libffi.call/cls_double.c: Likewise.
2780         * testsuite/libffi.call/cls_float.c: Likewise.
2781         * testsuite/libffi.call/cls_schar.c: Likewise.
2782         * testsuite/libffi.call/cls_sint.c: Likewise.
2783         * testsuite/libffi.call/cls_sshort.c: Likewise.
2784         * testsuite/libffi.call/cls_uchar.c: Likewise.
2785         * testsuite/libffi.call/cls_uint.c: Likewise.
2786         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2787         * testsuite/libffi.call/cls_ushort.c: Likewise.
2788         * testsuite/libffi.call/nested_struct.c: Likewise.
2789         * testsuite/libffi.call/nested_struct1.c: Likewise.
2790         * testsuite/libffi.call/nested_struct2.c: Likewise.
2791         * testsuite/libffi.call/nested_struct3.c: Likewise.
2792         * testsuite/libffi.call/problem1.c: Likewise.
2793         * testsuite/libffi.special/unwindtest.cc: Likewise.
2794
2795 2003-11-20  Andreas Tobler  <a.tobler@schweiz.ch>
2796
2797         * testsuite/lib/libffi-dg.exp: Make the -lgcc_s conditional.
2798
2799 2003-11-19  Andreas Tobler  <a.tobler@schweiz.ch>
2800
2801         * testsuite/lib/libffi-dg.exp: Add DYLD_LIBRARY_PATH for darwin.
2802         Add -lgcc_s to additional flags.
2803
2804 2003-11-12  Andreas Tobler  <a.tobler@schweiz.ch>
2805
2806         * configure.in, include/Makefile.am: PR libgcj/11147, install
2807         the ffitarget.h header file in a gcc versioned and target
2808         dependent place.
2809         * configure: Regenerated.
2810         * Makefile.in, include/Makefile.in: Likewise.
2811         * testsuite/Makefile.in: Likewise.
2812
2813 2003-11-09  Andreas Tobler  <a.tobler@schweiz.ch>
2814
2815         * testsuite/libffi.call/closure_fn0.c: Print result and check
2816         with dg-output to make debugging easier.
2817         * testsuite/libffi.call/closure_fn1.c: Likewise.
2818         * testsuite/libffi.call/closure_fn2.c: Likewise.
2819         * testsuite/libffi.call/closure_fn3.c: Likewise.
2820         * testsuite/libffi.call/closure_fn4.c: Likewise.
2821         * testsuite/libffi.call/closure_fn5.c: Likewise.
2822         * testsuite/libffi.call/cls_12byte.c: Likewise.
2823         * testsuite/libffi.call/cls_16byte.c: Likewise.
2824         * testsuite/libffi.call/cls_18byte.c: Likewise.
2825         * testsuite/libffi.call/cls_19byte.c: Likewise.
2826         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2827         * testsuite/libffi.call/cls_20byte.c: Likewise.
2828         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2829         * testsuite/libffi.call/cls_24byte.c: Likewise.
2830         * testsuite/libffi.call/cls_2byte.c: Likewise.
2831         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2832         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2833         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2834         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2835         * testsuite/libffi.call/cls_4byte.c: Likewise.
2836         * testsuite/libffi.call/cls_5byte.c: Likewise.
2837         * testsuite/libffi.call/cls_64byte.c: Likewise.
2838         * testsuite/libffi.call/cls_6byte.c: Likewise.
2839         * testsuite/libffi.call/cls_7byte.c: Likewise.
2840         * testsuite/libffi.call/cls_8byte.c: Likewise.
2841         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2842         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2843         * testsuite/libffi.call/cls_double.c: Likewise.
2844         * testsuite/libffi.call/cls_float.c: Likewise.
2845         * testsuite/libffi.call/cls_schar.c: Likewise.
2846         * testsuite/libffi.call/cls_sint.c: Likewise.
2847         * testsuite/libffi.call/cls_sshort.c: Likewise.
2848         * testsuite/libffi.call/cls_uchar.c: Likewise.
2849         * testsuite/libffi.call/cls_uint.c: Likewise.
2850         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2851         * testsuite/libffi.call/cls_ushort.c: Likewise.
2852         * testsuite/libffi.call/problem1.c: Likewise.
2853
2854         * testsuite/libffi.special/unwindtest.cc: Make ffi_closure
2855         static.
2856
2857 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
2858
2859         * testsuite/libffi.call/cls_9byte2.c: New test case.
2860         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2861         * testsuite/libffi.call/cls_64byte.c: Likewise.
2862         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2863         * testsuite/libffi.call/cls_19byte.c: Likewise.
2864         * testsuite/libffi.call/cls_18byte.c: Likewise.
2865         * testsuite/libffi.call/closure_fn4.c: Likewise.
2866         * testsuite/libffi.call/closure_fn5.c: Likewise.
2867         * testsuite/libffi.call/cls_schar.c: Likewise.
2868         * testsuite/libffi.call/cls_sint.c: Likewise.
2869         * testsuite/libffi.call/cls_sshort.c: Likewise.
2870         * testsuite/libffi.call/nested_struct2.c: Likewise.
2871         * testsuite/libffi.call/nested_struct3.c: Likewise.
2872
2873 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
2874
2875         * testsuite/libffi.call/cls_double.c: Do a check on the result.
2876         * testsuite/libffi.call/cls_uchar.c: Likewise.
2877         * testsuite/libffi.call/cls_uint.c: Likewise.
2878         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2879         * testsuite/libffi.call/cls_ushort.c: Likewise.
2880         * testsuite/libffi.call/return_sc.c: Cleanup whitespaces.
2881
2882 2003-11-06  Andreas Tobler  <a.tobler@schweiz.ch>
2883
2884         * src/prep_cif.c (ffi_prep_cif): Move the validity check after
2885         the initialization.
2886
2887 2003-10-23  Andreas Tobler  <a.tobler@schweiz.ch>
2888
2889         * src/java_raw_api.c (ffi_java_ptrarray_to_raw): Replace
2890         FFI_ASSERT(FALSE) with FFI_ASSERT(0).
2891
2892 2003-10-22  David Daney  <ddaney@avtrex.com>
2893
2894         * src/mips/ffitarget.h: Replace undefined UINT32 and friends with
2895         __attribute__((__mode__(__SI__))) and friends.
2896
2897 2003-10-22  Andreas Schwab  <schwab@suse.de>
2898
2899         * src/ia64/ffi.c: Replace FALSE/TRUE with false/true.
2900
2901 2003-10-21  Andreas Tobler  <a.tobler@schweiz.ch>
2902
2903         * configure.in: AC_LINK_FILES(ffitarget.h).
2904         * configure: Regenerate.
2905         * Makefile.in: Likewise.
2906         * include/Makefile.in: Likewise.
2907         * testsuite/Makefile.in: Likewise.
2908         * fficonfig.h.in: Likewise.
2909
2910 2003-10-21  Paolo Bonzini  <bonzini@gnu.org>
2911             Richard Henderson  <rth@redhat.com>
2912
2913         Avoid that ffi.h includes fficonfig.h.
2914
2915         * Makefile.am (EXTRA_DIST): Include ffitarget.h files
2916         (TARGET_SRC_MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
2917         (TARGET_SRC_MIPS_SGI): Removed.
2918         (MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
2919         (MIPS_SGI): Removed.
2920         (CLEANFILES): Removed.
2921         (mostlyclean-am, clean-am, mostlyclean-sub, clean-sub): New
2922         targets.
2923         * acconfig.h: Removed.
2924         * configure.in: Compute sizeofs only for double and long double.
2925         Use them to define and subst HAVE_LONG_DOUBLE.  Include comments
2926         into AC_DEFINE instead of using acconfig.h.  Create
2927         include/ffitarget.h instead of include/fficonfig.h.  Rename
2928         MIPS_GCC to MIPS_IRIX, drop MIPS_SGI since we are in gcc's tree.
2929         AC_DEFINE EH_FRAME_FLAGS.
2930         * include/Makefile.am (DISTCLEANFILES): New automake macro.
2931         (hack_DATA): Add ffitarget.h.
2932         * include/ffi.h.in: Remove all system specific definitions.
2933         Declare raw API even if it is not installed, why bother?
2934         Use limits.h instead of SIZEOF_* to define ffi_type_*.  Do
2935         not define EH_FRAME_FLAGS, it is in fficonfig.h now.  Include
2936         ffitarget.h instead of fficonfig.h.  Remove ALIGN macro.
2937         (UINT_ARG, INT_ARG): Removed, use ffi_arg and ffi_sarg instead.
2938         * include/ffi_common.h (bool): Do not define.
2939         (ffi_assert): Accept failed assertion.
2940         (ffi_type_test): Return void and accept file/line.
2941         (FFI_ASSERT): Pass stringized failed assertion.
2942         (FFI_ASSERT_AT): New macro.
2943         (FFI_ASSERT_VALID_TYPE): New macro.
2944         (UINT8, SINT8, UINT16, SINT16, UINT32, SINT32,
2945         UINT64, SINT64): Define here with gcc's __attribute__ macro
2946         instead of in ffi.h
2947         (FLOAT32, ALIGN): Define here instead of in ffi.h
2948         * include/ffi-mips.h: Removed.  Its content moved to
2949         src/mips/ffitarget.h after separating assembly and C sections.
2950         * src/alpha/ffi.c, src/alpha/ffi.c, src/java_raw_api.c
2951         src/prep_cif.c, src/raw_api.c, src/ia64/ffi.c,
2952         src/mips/ffi.c, src/mips/n32.S, src/mips/o32.S,
2953         src/mips/ffitarget.h, src/sparc/ffi.c, src/x86/ffi64.c:
2954         SIZEOF_ARG -> FFI_SIZEOF_ARG.
2955         * src/ia64/ffi.c: Include stdbool.h (provided by GCC 2.95+).
2956         * src/debug.c (ffi_assert): Accept stringized failed assertion.
2957         (ffi_type_test): Rewritten.
2958         * src/prep-cif.c (initialize_aggregate, ffi_prep_cif): Call
2959         FFI_ASSERT_VALID_TYPE.
2960         * src/alpha/ffitarget.h, src/arm/ffitarget.h,
2961         src/ia64/ffitarget.h, src/m68k/ffitarget.h,
2962         src/mips/ffitarget.h, src/powerpc/ffitarget.h,
2963         src/s390/ffitarget.h, src/sh/ffitarget.h,
2964         src/sh64/ffitarget.h, src/sparc/ffitarget.h,
2965         src/x86/ffitarget.h: New files.
2966         * src/alpha/osf.S, src/arm/sysv.S, src/ia64/unix.S,
2967         src/m68k/sysv.S, src/mips/n32.S, src/mips/o32.S,
2968         src/powerpc/aix.S, src/powerpc/darwin.S,
2969         src/powerpc/ffi_darwin.c, src/powerpc/linux64.S,
2970         src/powerpc/linux64_closure.S, src/powerpc/ppc_closure.S,
2971         src/powerpc/sysv.S, src/s390/sysv.S, src/sh/sysv.S,
2972         src/sh64/sysv.S, src/sparc/v8.S, src/sparc/v9.S,
2973         src/x86/sysv.S, src/x86/unix64.S, src/x86/win32.S:
2974         include fficonfig.h
2975
2976 2003-10-20  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2977
2978         * src/mips/ffi.c: Use _ABIN32, _ABIO32 instead of external
2979         _MIPS_SIM_NABI32, _MIPS_SIM_ABI32.
2980
2981 2003-10-19  Andreas Tobler  <a.tobler@schweiz.ch>
2982
2983         * src/powerpc/ffi_darwin.c (ffi_prep_args): Declare bytes again.
2984         Used when FFI_DEBUG = 1.
2985
2986 2003-10-14  Alan Modra  <amodra@bigpond.net.au>
2987
2988         * src/types.c (double, longdouble): Default POWERPC64 to 8 byte size
2989         and align.
2990
2991 2003-10-06  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2992
2993         * include/ffi_mips.h: Define FFI_MIPS_N32 for N32/N64 ABIs,
2994         FFI_MIPS_O32 for O32 ABI.
2995
2996 2003-10-01  Andreas Tobler  <a.tobler@schweiz.ch>
2997
2998         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_64 for
2999         SPARC64. Cleanup whitespaces.
3000
3001 2003-09-19  Andreas Tobler  <a.tobler@schweiz.ch>
3002
3003         * testsuite/libffi.call/closure_fn0.c: Xfail mips, arm,
3004         strongarm, xscale. Cleanup whitespaces.
3005         * testsuite/libffi.call/closure_fn1.c: Likewise.
3006         * testsuite/libffi.call/closure_fn2.c: Likewise.
3007         * testsuite/libffi.call/closure_fn3.c: Likewise.
3008         * testsuite/libffi.call/cls_12byte.c: Likewise.
3009         * testsuite/libffi.call/cls_16byte.c: Likewise.
3010         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3011         * testsuite/libffi.call/cls_20byte.c: Likewise.
3012         * testsuite/libffi.call/cls_24byte.c: Likewise.
3013         * testsuite/libffi.call/cls_2byte.c: Likewise.
3014         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3015         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3016         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3017         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3018         * testsuite/libffi.call/cls_4byte.c: Likewise.
3019         * testsuite/libffi.call/cls_5byte.c: Likewise.
3020         * testsuite/libffi.call/cls_6byte.c: Likewise.
3021         * testsuite/libffi.call/cls_7byte.c: Likewise.
3022         * testsuite/libffi.call/cls_8byte.c: Likewise.
3023         * testsuite/libffi.call/cls_double.c: Likewise.
3024         * testsuite/libffi.call/cls_float.c: Likewise.
3025         * testsuite/libffi.call/cls_uchar.c: Likewise.
3026         * testsuite/libffi.call/cls_uint.c: Likewise.
3027         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3028         * testsuite/libffi.call/cls_ushort.c: Likewise.
3029         * testsuite/libffi.call/nested_struct.c: Likewise.
3030         * testsuite/libffi.call/nested_struct1.c: Likewise.
3031         * testsuite/libffi.call/problem1.c: Likewise.
3032         * testsuite/libffi.special/unwindtest.cc: Likewise.
3033         * testsuite/libffi.call/pyobjc-tc.c: Cleanup whitespaces.
3034
3035 2003-09-18  David Edelsohn  <edelsohn@gnu.org>
3036
3037         * src/powerpc/aix.S: Cleanup whitespaces.
3038         * src/powerpc/aix_closure.S: Likewise.
3039
3040 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3041
3042         * src/powerpc/darwin.S: Cleanup whitespaces, comment formatting.
3043         * src/powerpc/darwin_closure.S: Likewise.
3044         * src/powerpc/ffi_darwin.c: Likewise.
3045
3046 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
3047             David Edelsohn  <edelsohn@gnu.org>
3048
3049         * src/types.c (double): Add AIX and Darwin to the right TYPEDEF.
3050         * src/powerpc/aix_closure.S: Remove the pointer to the outgoing
3051         parameter stack.
3052         * src/powerpc/darwin_closure.S: Likewise.
3053         * src/powerpc/ffi_darwin.c (ffi_prep_args): Handle structures
3054         according to the Darwin/AIX ABI.
3055         (ffi_prep_cif_machdep): Likewise.
3056         (ffi_closure_helper_DARWIN): Likewise.
3057         Remove the outgoing parameter stack logic. Simplify the evaluation
3058         of the different CASE types.
3059         (ffi_prep_clousure): Avoid the casts on lvalues. Change the branch
3060         statement in the trampoline code.
3061
3062 2003-09-18  Kaz Kojima  <kkojima@gcc.gnu.org>
3063
3064         * src/sh/ffi.c (ffi_prep_args): Take account into the alignement
3065         for the register size.
3066         (ffi_closure_helper_SYSV): Handle the structure return value
3067         address correctly.
3068         (ffi_closure_helper_SYSV): Return the appropriate type when
3069         the registers are used for the structure return value.
3070         * src/sh/sysv.S (ffi_closure_SYSV): Fix the stack layout for
3071         the 64-bit return value.  Update copyright years.
3072
3073 2003-09-17  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3074
3075         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Search in
3076         srcdir for ffi_mips.h.
3077
3078 2003-09-12  Alan Modra  <amodra@bigpond.net.au>
3079
3080         * src/prep_cif.c (initialize_aggregate): Include tail padding in
3081         structure size.
3082         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Correct
3083         placement of float result.
3084         * testsuite/libffi.special/unwindtest.cc (closure_test_fn1): Correct
3085         cast of "resp" for big-endian 64 bit machines.
3086
3087 2003-09-11  Alan Modra  <amodra@bigpond.net.au>
3088
3089         * src/types.c (double, longdouble): Merge identical SH and ARM
3090         typedefs, and add POWERPC64.
3091         * src/powerpc/ffi.c (ffi_prep_args64): Correct next_arg calc for
3092         struct split over gpr and rest.
3093         (ffi_prep_cif_machdep): Correct intarg_count for structures.
3094         * src/powerpc/linux64.S (ffi_call_LINUX64): Fix gpr offsets.
3095
3096 2003-09-09  Andreas Tobler  <a.tobler@schweiz.ch>
3097
3098         * src/powerpc/ffi.c (ffi_closure_helper_SYSV) Handle struct
3099         passing correctly.
3100
3101 2003-09-09  Alan Modra  <amodra@bigpond.net.au>
3102
3103         * configure: Regenerate.
3104
3105 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3106
3107         * Makefile.am: Remove build rules for ffitest.
3108         * Makefile.in: Rebuilt.
3109
3110 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3111
3112         * src/java_raw_api.c: Include <stdlib.h> to fix compiler warning
3113         about implicit declaration of abort().
3114
3115 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
3116
3117         * Makefile.am: Add dejagnu test framework. Fixes PR other/11411.
3118         * Makefile.in: Rebuilt.
3119         * configure.in: Add dejagnu test framework.
3120         * configure: Rebuilt.
3121
3122         * testsuite/Makefile.am: New file.
3123         * testsuite/Makefile.in: Built
3124         * testsuite/lib/libffi-dg.exp: New file.
3125         * testsuite/config/default.exp: Likewise.
3126         * testsuite/libffi.call/call.exp: Likewise.
3127         * testsuite/libffi.call/ffitest.h: Likewise.
3128         * testsuite/libffi.call/closure_fn0.c: Likewise.
3129         * testsuite/libffi.call/closure_fn1.c: Likewise.
3130         * testsuite/libffi.call/closure_fn2.c: Likewise.
3131         * testsuite/libffi.call/closure_fn3.c: Likewise.
3132         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3133         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3134         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3135         * testsuite/libffi.call/cls_2byte.c: Likewise.
3136         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3137         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3138         * testsuite/libffi.call/cls_4byte.c: Likewise.
3139         * testsuite/libffi.call/cls_5byte.c: Likewise.
3140         * testsuite/libffi.call/cls_6byte.c: Likewise.
3141         * testsuite/libffi.call/cls_7byte.c: Likewise.
3142         * testsuite/libffi.call/cls_8byte.c: Likewise.
3143         * testsuite/libffi.call/cls_12byte.c: Likewise.
3144         * testsuite/libffi.call/cls_16byte.c: Likewise.
3145         * testsuite/libffi.call/cls_20byte.c: Likewise.
3146         * testsuite/libffi.call/cls_24byte.c: Likewise.
3147         * testsuite/libffi.call/cls_double.c: Likewise.
3148         * testsuite/libffi.call/cls_float.c: Likewise.
3149         * testsuite/libffi.call/cls_uchar.c: Likewise.
3150         * testsuite/libffi.call/cls_uint.c: Likewise.
3151         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3152         * testsuite/libffi.call/cls_ushort.c: Likewise.
3153         * testsuite/libffi.call/float.c: Likewise.
3154         * testsuite/libffi.call/float1.c: Likewise.
3155         * testsuite/libffi.call/float2.c: Likewise.
3156         * testsuite/libffi.call/many.c: Likewise.
3157         * testsuite/libffi.call/many_win32.c: Likewise.
3158         * testsuite/libffi.call/nested_struct.c: Likewise.
3159         * testsuite/libffi.call/nested_struct1.c: Likewise.
3160         * testsuite/libffi.call/pyobjc-tc.c: Likewise.
3161         * testsuite/libffi.call/problem1.c: Likewise.
3162         * testsuite/libffi.call/promotion.c: Likewise.
3163         * testsuite/libffi.call/return_ll.c: Likewise.
3164         * testsuite/libffi.call/return_sc.c: Likewise.
3165         * testsuite/libffi.call/return_uc.c: Likewise.
3166         * testsuite/libffi.call/strlen.c: Likewise.
3167         * testsuite/libffi.call/strlen_win32.c: Likewise.
3168         * testsuite/libffi.call/struct1.c: Likewise.
3169         * testsuite/libffi.call/struct2.c: Likewise.
3170         * testsuite/libffi.call/struct3.c: Likewise.
3171         * testsuite/libffi.call/struct4.c: Likewise.
3172         * testsuite/libffi.call/struct5.c: Likewise.
3173         * testsuite/libffi.call/struct6.c: Likewise.
3174         * testsuite/libffi.call/struct7.c: Likewise.
3175         * testsuite/libffi.call/struct8.c: Likewise.
3176         * testsuite/libffi.call/struct9.c: Likewise.
3177         * testsuite/libffi.special/special.exp: New file.
3178         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3179         * testsuite/libffi.special/unwindtest.cc: Likewise.
3180
3181
3182 2003-08-13  Kaz Kojima  <kkojima@gcc.gnu.org>
3183
3184         * src/sh/ffi.c (OFS_INT16): Set 0 for little endian case.  Update
3185         copyright years.
3186
3187 2003-08-02  Alan Modra  <amodra@bigpond.net.au>
3188
3189         * src/powerpc/ffi.c (ffi_prep_args64): Modify for changed gcc
3190         structure passing.
3191         (ffi_closure_helper_LINUX64): Likewise.
3192         * src/powerpc/linux64.S: Remove code writing to parm save area.
3193         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Use return
3194         address in lr from ffi_closure_helper_LINUX64 call to calculate
3195         table address.  Optimize function tail.
3196
3197 2003-07-28  Andreas Tobler  <a.tobler@schweiz.ch>
3198
3199         * src/sparc/ffi.c: Handle all floating point registers.
3200         * src/sparc/v9.S: Likewise. Fixes second part of PR target/11410.
3201
3202 2003-07-11  Gerald Pfeifer  <pfeifer@dbai.tuwien.ac.at>
3203
3204         * README: Note that libffi is not part of GCC.  Update the project
3205         URL and status.
3206
3207 2003-06-19  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3208
3209         * src/powerpc/ppc_closure.S: Include ffi.h.
3210
3211 2003-06-13  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3212
3213         * src/x86/sysv.S: Avoid gas-only .uleb128/.sleb128 directives.
3214         Use C style comments.
3215
3216 2003-06-13  Kaz Kojima  <kkojima@rr.iij4u.or.jp>
3217
3218         * Makefile.am: Add SHmedia support.  Fix a typo of SH support.
3219         * Makefile.in: Regenerate.
3220         * configure.in (sh64-*-linux*, sh5*-*-linux*): Add target.
3221         * configure: Regenerate.
3222         * include/ffi.h.in: Add SHmedia support.
3223         * src/sh64/ffi.c: New file.
3224         * src/sh64/sysv.S: New file.
3225
3226 2003-05-16  Jakub Jelinek  <jakub@redhat.com>
3227
3228         * configure.in (HAVE_RO_EH_FRAME): Check whether .eh_frame section
3229         should be read-only.
3230         * configure: Rebuilt.
3231         * fficonfig.h.in: Rebuilt.
3232         * include/ffi.h.in (EH_FRAME_FLAGS): Define.
3233         * src/alpha/osf.S: Use EH_FRAME_FLAGS.
3234         * src/powerpc/linux64.S: Likewise.
3235         * src/powerpc/linux64_closure.S: Likewise.  Include ffi.h.
3236         * src/powerpc/sysv.S: Use EH_FRAME_FLAGS.  Use pcrel encoding
3237         if -fpic/-fPIC/-mrelocatable.
3238         * src/powerpc/powerpc_closure.S: Likewise.
3239         * src/sparc/v8.S: If HAVE_RO_EH_FRAME is defined, don't include
3240         #write in .eh_frame flags.
3241         * src/sparc/v9.S: Likewise.
3242         * src/x86/unix64.S: Use EH_FRAME_FLAGS.
3243         * src/x86/sysv.S: Likewise.  Use pcrel encoding if -fpic/-fPIC.
3244         * src/s390/sysv.S: Use EH_FRAME_FLAGS.  Include ffi.h.
3245
3246 2003-05-07  Jeff Sturm  <jsturm@one-point.com>
3247
3248         Fixes PR bootstrap/10656
3249         * configure.in (HAVE_AS_REGISTER_PSEUDO_OP): Test assembler
3250         support for .register pseudo-op.
3251         * src/sparc/v8.S: Use it.
3252         * fficonfig.h.in: Rebuilt.
3253         * configure: Rebuilt.
3254
3255 2003-04-18  Jakub Jelinek  <jakub@redhat.com>
3256
3257         * include/ffi.h.in (POWERPC64): Define if 64-bit.
3258         (enum ffi_abi): Add FFI_LINUX64 on POWERPC.
3259         Make it the default on POWERPC64.
3260         (FFI_TRAMPOLINE_SIZE): Define to 24 on POWERPC64.
3261         * configure.in: Change powerpc-*-linux* into powerpc*-*-linux*.
3262         * configure: Rebuilt.
3263         * src/powerpc/ffi.c (hidden): Define.
3264         (ffi_prep_args_SYSV): Renamed from
3265         ffi_prep_args.  Cast pointers to unsigned long to shut up warnings.
3266         (NUM_GPR_ARG_REGISTERS64, NUM_FPR_ARG_REGISTERS64,
3267         ASM_NEEDS_REGISTERS64): New.
3268         (ffi_prep_args64): New function.
3269         (ffi_prep_cif_machdep): Handle FFI_LINUX64 ABI.
3270         (ffi_call): Likewise.
3271         (ffi_prep_closure): Likewise.
3272         (flush_icache): Surround by #ifndef POWERPC64.
3273         (ffi_dblfl): New union type.
3274         (ffi_closure_helper_SYSV): Use it to avoid aliasing problems.
3275         (ffi_closure_helper_LINUX64): New function.
3276         * src/powerpc/ppc_closure.S: Surround whole file by #ifndef
3277         __powerpc64__.
3278         * src/powerpc/sysv.S: Likewise.
3279         (ffi_call_SYSV): Rename ffi_prep_args to ffi_prep_args_SYSV.
3280         * src/powerpc/linux64.S: New file.
3281         * src/powerpc/linux64_closure.S: New file.
3282         * Makefile.am (EXTRA_DIST): Add src/powerpc/linux64.S and
3283         src/powerpc/linux64_closure.S.
3284         (TARGET_SRC_POWERPC): Likewise.
3285
3286         * src/ffitest.c (closure_test_fn, closure_test_fn1, closure_test_fn2,
3287         closure_test_fn3): Fix result printing on big-endian 64-bit
3288         machines.
3289         (main): Print tst2_arg instead of uninitialized tst2_result.
3290
3291         * src/ffitest.c (main): Hide what closure pointer really points to
3292         from the compiler.
3293
3294 2003-04-16  Richard Earnshaw  <rearnsha@arm.com>
3295
3296         * configure.in (arm-*-netbsdelf*): Add configuration.
3297         (configure): Regenerated.
3298
3299 2003-04-04  Loren J. Rittle  <ljrittle@acm.org>
3300
3301         * include/Makefile.in: Regenerate.
3302
3303 2003-03-21  Zdenek Dvorak  <rakdver@atrey.karlin.mff.cuni.cz>
3304
3305         * libffi/include/ffi.h.in: Define X86 instead of X86_64 in 32
3306         bit mode.
3307         * libffi/src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV):
3308         Receive closure pointer through parameter, read args using
3309         __builtin_dwarf_cfa.
3310         (FFI_INIT_TRAMPOLINE): Send closure reference through eax.
3311
3312 2003-03-12  Andreas Schwab  <schwab@suse.de>
3313
3314         * configure.in: Avoid trailing /. in toolexeclibdir.
3315         * configure: Rebuilt.
3316
3317 2003-03-03  Andreas Tobler <a.tobler@schweiz.ch>
3318
3319         * src/powerpc/darwin_closure.S: Recode to fit dynamic libraries.
3320
3321 2003-02-06  Andreas Tobler <a.tobler@schweiz.ch>
3322
3323         * libffi/src/powerpc/darwin_closure.S:
3324         Fix alignement bug, allocate 8 bytes for the result.
3325         * libffi/src/powerpc/aix_closure.S:
3326         Likewise.
3327         * libffi/src/powerpc/ffi_darwin.c:
3328         Update stackframe description for aix/darwin_closure.S.
3329
3330 2003-02-06  Jakub Jelinek  <jakub@redhat.com>
3331
3332         * src/s390/ffi.c (ffi_closure_helper_SYSV): Add hidden visibility
3333         attribute.
3334
3335 2003-01-31  Christian Cornelssen  <ccorn@cs.tu-berlin.de>,
3336             Andreas Schwab  <schwab@suse.de>
3337
3338         * configure.in: Adjust command to source config-ml.in to account
3339         for changes to the libffi_basedir definition.
3340         (libffi_basedir): Remove ${srcdir} from value and include trailing
3341         slash if nonempty.
3342
3343         * configure: Regenerate.
3344
3345 2003-01-29  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3346
3347         * src/powerpc/ppc_closure.S: Recode to fit shared libs.
3348
3349 2003-01-28  Andrew Haley  <aph@redhat.com>
3350
3351         * include/ffi.h.in: Enable FFI_CLOSURES for x86_64.
3352         * src/x86/ffi64.c (ffi_prep_closure): New.
3353         (ffi_closure_UNIX64_inner): New.
3354         * src/x86/unix64.S (ffi_closure_UNIX64): New.
3355
3356 2003-01-27  Alexandre Oliva  <aoliva@redhat.com>
3357
3358         * configure.in (toolexecdir, toolexeclibdir): Set and AC_SUBST.
3359         Remove USE_LIBDIR conditional.
3360         * Makefile.am (toolexecdir, toolexeclibdir): Don't override.
3361         * Makefile.in, configure: Rebuilt.
3362
3363 2003-01027  David Edelsohn  <edelsohn@gnu.org>
3364
3365         * Makefile.am (TARGET_SRC_POWERPC_AIX): Fix typo.
3366         * Makefile.in: Regenerate.
3367
3368 2003-01-22  Andrew Haley  <aph@redhat.com>
3369
3370         * src/powerpc/darwin.S (_ffi_call_AIX): Add Augmentation size to
3371         unwind info.
3372
3373 2003-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
3374
3375         * src/powerpc/darwin.S: Add unwind info.
3376         * src/powerpc/darwin_closure.S: Likewise.
3377
3378 2003-01-14  Andrew Haley  <aph@redhat.com>
3379
3380         * src/x86/ffi64.c (ffi_prep_args): Check for void retval.
3381         (ffi_prep_cif_machdep): Likewise.
3382         * src/x86/unix64.S: Add unwind info.
3383
3384 2003-01-14  Andreas Jaeger  <aj@suse.de>
3385
3386         * src/ffitest.c (main): Only use ffi_closures if those are
3387         supported.
3388
3389 2003-01-13 Andreas Tobler <a.tobler@schweiz.ch>
3390
3391         * libffi/src/ffitest.c
3392          add closure testcases
3393
3394 2003-01-13 Kevin B. Hendricks <khendricks@ivey.uwo.ca>
3395
3396         * libffi/src/powerpc/ffi.c
3397          fix alignment bug for float (4 byte aligned iso 8 byte)
3398
3399 2003-01-09  Geoffrey Keating  <geoffk@apple.com>
3400
3401         * src/powerpc/ffi_darwin.c: Remove RCS version string.
3402         * src/powerpc/darwin.S: Remove RCS version string.
3403
3404 2003-01-03  Jeff Sturm  <jsturm@one-point.com>
3405
3406         * include/ffi.h.in: Add closure defines for SPARC, SPARC64.
3407         * src/ffitest.c (main): Use static storage for closure.
3408         * src/sparc/ffi.c (ffi_prep_closure, ffi_closure_sparc_inner): New.
3409         * src/sparc/v8.S (ffi_closure_v8): New.
3410         * src/sparc/v9.S (ffi_closure_v9): New.
3411
3412 2002-11-10  Ranjit Mathew <rmathew@hotmail.com>
3413
3414         * include/ffi.h.in: Added FFI_STDCALL ffi_type
3415           enumeration for X86_WIN32.
3416         * src/x86/win32.S: Added ffi_call_STDCALL function
3417           definition.
3418         * src/x86/ffi.c (ffi_call/ffi_raw_call): Added
3419           switch cases for recognising FFI_STDCALL and
3420           calling ffi_call_STDCALL if target is X86_WIN32.
3421         * src/ffitest.c (my_stdcall_strlen/stdcall_many):
3422           stdcall versions of the "my_strlen" and "many"
3423           test functions (for X86_WIN32).
3424           Added test cases to test stdcall invocation using
3425           these functions.
3426
3427 2002-12-02  Kaz Kojima  <kkojima@gcc.gnu.org>
3428
3429         * src/sh/sysv.S: Add DWARF2 unwind info.
3430
3431 2002-11-27  Ulrich Weigand  <uweigand@de.ibm.com>
3432
3433         * src/s390/sysv.S (.eh_frame section): Make section read-only.
3434
3435 2002-11-26  Jim Wilson  <wilson@redhat.com>
3436
3437         * src/types.c (FFI_TYPE_POINTER): Has size 8 on IA64.
3438