OSDN Git Service

2009-11-02 Andreas Tobler <a.tobler@schweiz.org>
[pf3gnuchains/gcc-fork.git] / libffi / ChangeLog
1 2009-11-02  Andreas Tobler  <a.tobler@schweiz.org>
2
3         PR libffi/41908
4         * testsuite/libffi.call/testclosure.c: New test.
5
6 2009-09-28  Kai Tietz  <kai.tietz@onevision.com>
7
8         * src/x86/win64.S (_ffi_call_win64 stack): Remove for gnu
9         assembly version use of ___chkstk.
10
11 2009-09-23  Matthias Klose  <doko@ubuntu.com>
12
13         PR libffi/40242, PR libffi/41443
14         * src/arm/sysv.S (__ARM_ARCH__): Define for processors
15         __ARM_ARCH_6T2__, __ARM_ARCH_6M__, __ARM_ARCH_7__,
16         __ARM_ARCH_7A__, __ARM_ARCH_7R__, __ARM_ARCH_7M__.
17         Change the conditionals to __SOFTFP__ || __ARM_EABI__
18         for -mfloat-abi=softfp to work.
19
20 2009-09-17  Loren J. Rittle  <ljrittle@acm.org>
21
22         PR testsuite/32843 (strikes again)
23         * src/x86/ffi.c (ffi_prep_cif_machdep): Add X86_FREEBSD to
24         enable proper extension on char and short.
25
26 2009-09-15  David Daney  <ddaney@caviumnetworks.com>
27
28         * src/java_raw_api.c (ffi_java_raw_to_rvalue): Remove special
29         handling for FFI_TYPE_POINTER.
30         * src/mips/ffitarget.h (FFI_TYPE_STRUCT_D_SOFT,
31         FFI_TYPE_STRUCT_F_SOFT, FFI_TYPE_STRUCT_DD_SOFT,
32         FFI_TYPE_STRUCT_FF_SOFT, FFI_TYPE_STRUCT_FD_SOFT,
33         FFI_TYPE_STRUCT_DF_SOFT, FFI_TYPE_STRUCT_SOFT): New defines.
34         (FFI_N32_SOFT_FLOAT, FFI_N64_SOFT_FLOAT): New ffi_abi enumerations.
35         (enum ffi_abi): Set FFI_DEFAULT_ABI for soft-float.
36         * src/mips/n32.S (ffi_call_N32): Add handling for soft-float
37         structure and pointer returns.
38         (ffi_closure_N32): Add handling for pointer returns.
39         * src/mips/ffi.c (ffi_prep_args, calc_n32_struct_flags,
40         calc_n32_return_struct_flags): Handle soft-float.
41         (ffi_prep_cif_machdep):  Handle soft-float, fix pointer handling.
42         (ffi_call_N32): Declare proper argument types.
43         (ffi_call, copy_struct_N32, ffi_closure_mips_inner_N32): Handle
44         soft-float.
45
46 2009-08-24  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
47
48         * configure.ac (AC_PREREQ): Bump to 2.64.
49
50 2009-08-22  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
51
52         * Makefile.am (install-html, install-pdf): Remove.
53         * Makefile.in: Regenerate.
54
55         * Makefile.in: Regenerate.
56         * aclocal.m4: Regenerate.
57         * configure: Regenerate.
58         * fficonfig.h.in: Regenerate.
59         * include/Makefile.in: Regenerate.
60         * man/Makefile.in: Regenerate.
61         * testsuite/Makefile.in: Regenerate.
62
63 2009-07-30  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
64
65         * configure.ac (_AC_ARG_VAR_PRECIOUS): Use m4_rename_force.
66
67 2009-07-24  Dave Korn  <dave.korn.cygwin@gmail.com>
68
69         PR libffi/40807
70         * src/x86/ffi.c (ffi_prep_cif_machdep): Also use sign/zero-extending
71         return types for X86_WIN32.
72         * src/x86/win32.S (_ffi_call_SYSV): Handle omitted return types.
73         (_ffi_call_STDCALL, _ffi_closure_SYSV, _ffi_closure_raw_SYSV,
74         _ffi_closure_STDCALL): Likewise.
75
76         * src/closures.c (is_selinux_enabled): Define to const 0 for Cygwin.
77         (dlmmap, dlmunmap): Also use these functions on Cygwin.
78
79 2009-07-11  Richard Sandiford  <rdsandiford@googlemail.com>
80
81         PR testsuite/40699
82         PR testsuite/40707
83         PR testsuite/40709
84         * testsuite/lib/libffi-dg.exp: Revert 2009-07-02, 2009-07-01 and
85         2009-06-30 commits.
86
87 2009-07-01  Richard Sandiford  <r.sandiford@uk.ibm.com>
88
89         * testsuite/lib/libffi-dg.exp (libffi-init): Set ld_library_path
90         to "" before adding paths.  (This reinstates an assignment that
91         was removed by my 2009-06-30 commit, but changes the initial
92         value from "." to "".)
93
94 2009-07-01  H.J. Lu  <hongjiu.lu@intel.com>
95
96         PR testsuite/40601
97         * testsuite/lib/libffi-dg.exp (libffi-init): Properly set
98         gccdir.  Adjust ld_library_path for gcc only if gccdir isn't
99         empty.
100
101 2009-06-30  Richard Sandiford  <r.sandiford@uk.ibm.com>
102
103         * testsuite/lib/libffi-dg.exp (libffi-init): Don't add "."
104         to ld_library_path.  Use add_path.  Add just find_libgcc_s
105         to ld_library_path, not every libgcc multilib directory.
106
107 2009-06-16  Wim Lewis  <wiml@hhhh.org>
108
109         * src/powerpc/ffi.c: Avoid clobbering cr3 and cr4, which are
110         supposed to be callee-saved.
111         * src/powerpc/sysv.S (small_struct_return_value): Fix overrun of
112         return buffer for odd-size structs.
113
114 2009-06-16  Andreas Tobler  <a.tobler@schweiz.org>
115
116         PR libffi/40444
117         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Add
118         allow_stack_execute for Darwin.
119
120 2009-06-16  Andrew Haley  <aph@redhat.com>
121
122         * configure.ac (TARGETDIR): Add missing blank lines.
123         * configure: Regenerate.
124
125 2009-06-16  Andrew Haley  <aph@redhat.com>
126
127         * testsuite/libffi.call/cls_align_sint64.c,
128         testsuite/libffi.call/cls_align_uint64.c,
129         testsuite/libffi.call/cls_longdouble_va.c,
130         testsuite/libffi.call/cls_ulonglong.c,
131         testsuite/libffi.call/return_ll1.c,
132         testsuite/libffi.call/stret_medium2.c: Fix printf format
133         specifiers.
134         * testsuite/libffi.call/ffitest.h,
135         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRIuLL): Define.
136
137 2009-06-15  Andrew Haley  <aph@redhat.com>
138
139         * testsuite/libffi.call/err_bad_typedef.c: xfail everywhere.
140         * testsuite/libffi.call/err_bad_abi.c: Likewise.
141         
142 2009-06-12  Andrew Haley  <aph@redhat.com>
143
144         * Makefile.am: Remove info_TEXINFOS.
145
146 2009-06-12  Andrew Haley  <aph@redhat.com>
147
148         * ChangeLog.libffi: testsuite/libffi.call/cls_align_sint64.c,
149         testsuite/libffi.call/cls_align_uint64.c,
150         testsuite/libffi.call/cls_ulonglong.c,
151         testsuite/libffi.call/return_ll1.c,
152         testsuite/libffi.call/stret_medium2.c: Fix printf format
153         specifiers.
154         testsuite/libffi.special/unwindtest.cc: include stdint.h.
155
156 2009-06-11  Timothy Wall  <twall@users.sf.net>
157
158         * Makefile.am,
159         configure.ac,
160         include/ffi.h.in,
161         include/ffi_common.h,
162         src/closures.c,
163         src/dlmalloc.c,
164         src/x86/ffi.c,
165         src/x86/ffitarget.h,
166         src/x86/win64.S (new),
167         README: Added win64 support (mingw or MSVC)
168         * Makefile.in,
169         include/Makefile.in,
170         man/Makefile.in,
171         testsuite/Makefile.in,
172         configure,
173         aclocal.m4: Regenerated
174         * ltcf-c.sh: properly escape cygwin/w32 path
175         * man/ffi_call.3: Clarify size requirements for return value.
176         * src/x86/ffi64.c: Fix filename in comment.
177         * src/x86/win32.S: Remove unused extern.
178
179         * testsuite/libffi.call/closure_fn0.c,
180         testsuite/libffi.call/closure_fn1.c,
181         testsuite/libffi.call/closure_fn2.c,
182         testsuite/libffi.call/closure_fn3.c,
183         testsuite/libffi.call/closure_fn4.c,
184         testsuite/libffi.call/closure_fn5.c,
185         testsuite/libffi.call/closure_fn6.c,
186         testsuite/libffi.call/closure_stdcall.c,
187         testsuite/libffi.call/cls_12byte.c,
188         testsuite/libffi.call/cls_16byte.c,
189         testsuite/libffi.call/cls_18byte.c,
190         testsuite/libffi.call/cls_19byte.c,
191         testsuite/libffi.call/cls_1_1byte.c,
192         testsuite/libffi.call/cls_20byte.c,
193         testsuite/libffi.call/cls_20byte1.c,
194         testsuite/libffi.call/cls_24byte.c,
195         testsuite/libffi.call/cls_2byte.c,
196         testsuite/libffi.call/cls_3_1byte.c,
197         testsuite/libffi.call/cls_3byte1.c,
198         testsuite/libffi.call/cls_3byte2.c,
199         testsuite/libffi.call/cls_4_1byte.c,
200         testsuite/libffi.call/cls_4byte.c,
201         testsuite/libffi.call/cls_5_1_byte.c,
202         testsuite/libffi.call/cls_5byte.c,
203         testsuite/libffi.call/cls_64byte.c,
204         testsuite/libffi.call/cls_6_1_byte.c,
205         testsuite/libffi.call/cls_6byte.c,
206         testsuite/libffi.call/cls_7_1_byte.c,
207         testsuite/libffi.call/cls_7byte.c,
208         testsuite/libffi.call/cls_8byte.c,
209         testsuite/libffi.call/cls_9byte1.c,
210         testsuite/libffi.call/cls_9byte2.c,
211         testsuite/libffi.call/cls_align_double.c,
212         testsuite/libffi.call/cls_align_float.c,
213         testsuite/libffi.call/cls_align_longdouble.c,
214         testsuite/libffi.call/cls_align_longdouble_split.c,
215         testsuite/libffi.call/cls_align_longdouble_split2.c,
216         testsuite/libffi.call/cls_align_pointer.c,
217         testsuite/libffi.call/cls_align_sint16.c,
218         testsuite/libffi.call/cls_align_sint32.c,
219         testsuite/libffi.call/cls_align_sint64.c,
220         testsuite/libffi.call/cls_align_uint16.c,
221         testsuite/libffi.call/cls_align_uint32.c,
222         testsuite/libffi.call/cls_align_uint64.c,
223         testsuite/libffi.call/cls_dbls_struct.c,
224         testsuite/libffi.call/cls_double.c,
225         testsuite/libffi.call/cls_double_va.c,
226         testsuite/libffi.call/cls_float.c,
227         testsuite/libffi.call/cls_longdouble.c,
228         testsuite/libffi.call/cls_longdouble_va.c,
229         testsuite/libffi.call/cls_multi_schar.c,
230         testsuite/libffi.call/cls_multi_sshort.c,
231         testsuite/libffi.call/cls_multi_sshortchar.c,
232         testsuite/libffi.call/cls_multi_uchar.c,
233         testsuite/libffi.call/cls_multi_ushort.c,
234         testsuite/libffi.call/cls_multi_ushortchar.c,
235         testsuite/libffi.call/cls_pointer.c,
236         testsuite/libffi.call/cls_pointer_stack.c,
237         testsuite/libffi.call/cls_schar.c,
238         testsuite/libffi.call/cls_sint.c,
239         testsuite/libffi.call/cls_sshort.c,
240         testsuite/libffi.call/cls_uchar.c,
241         testsuite/libffi.call/cls_uint.c,
242         testsuite/libffi.call/cls_ulonglong.c,
243         testsuite/libffi.call/cls_ushort.c,
244         testsuite/libffi.call/err_bad_abi.c,
245         testsuite/libffi.call/err_bad_typedef.c,
246         testsuite/libffi.call/float2.c,
247         testsuite/libffi.call/huge_struct.c,
248         testsuite/libffi.call/nested_struct.c,
249         testsuite/libffi.call/nested_struct1.c,
250         testsuite/libffi.call/nested_struct10.c,
251         testsuite/libffi.call/nested_struct2.c,
252         testsuite/libffi.call/nested_struct3.c,
253         testsuite/libffi.call/nested_struct4.c,
254         testsuite/libffi.call/nested_struct5.c,
255         testsuite/libffi.call/nested_struct6.c,
256         testsuite/libffi.call/nested_struct7.c,
257         testsuite/libffi.call/nested_struct8.c,
258         testsuite/libffi.call/nested_struct9.c,
259         testsuite/libffi.call/problem1.c,
260         testsuite/libffi.call/return_ldl.c,
261         testsuite/libffi.call/return_ll1.c,
262         testsuite/libffi.call/stret_large.c,
263         testsuite/libffi.call/stret_large2.c,
264         testsuite/libffi.call/stret_medium.c,
265         testsuite/libffi.call/stret_medium2.c,
266         testsuite/libffi.special/unwindtest.cc: use ffi_closure_alloc instead
267         of checking for MMAP.  Use intptr_t instead of long casts.
268
269 2009-06-11  Kaz Kojima  <kkojima@gcc.gnu.org>
270
271         * testsuite/libffi.call/cls_longdouble_va.c: Add xfail sh*-*-linux-*.
272         * testsuite/libffi.call/err_bad_abi.c: Add xfail sh*-*-*.
273         * testsuite/libffi.call/err_bad_typedef.c: Likewise.
274
275 2009-06-09  Andrew Haley  <aph@redhat.com>
276
277         * src/x86/freebsd.S: Add missing file.
278
279 2009-06-08  Andrew Haley  <aph@redhat.com>
280
281         Import from libffi 3.0.8:
282
283         * doc/libffi.texi: New file.
284         * doc/libffi.info: Likewise.
285         * doc/stamp-vti: Likewise.
286         * man/Makefile.am: New file.
287         * man/ffi_call.3: New file.
288
289         * Makefile.am (EXTRA_DIST): Add src/x86/darwin64.S,
290         src/dlmalloc.c.
291         (nodist_libffi_la_SOURCES): Add X86_FREEBSD.
292
293         * configure.ac: Bump version to 3.0.8.
294         parisc*-*-linux*: Add.
295         i386-*-freebsd* | i386-*-openbsd*: Add.
296         powerpc-*-beos*: Add.
297         AM_CONDITIONAL X86_FREEBSD: Add.
298         AC_CONFIG_FILES: Add man/Makefile.
299
300         * include/ffi.h.in (FFI_FN): Change void (*)() to void (*)(void).
301
302 2009-06-08  Andrew Haley  <aph@redhat.com>
303
304         * README: Import from libffi 3.0.8.
305
306 2009-06-08  Andrew Haley  <aph@redhat.com>
307
308         * testsuite/libffi.call/err_bad_abi.c: Add xfails.
309         * testsuite/libffi.call/cls_longdouble_va.c: Add xfails.
310         * testsuite/libffi.call/cls_dbls_struct.c: Add xfail x86_64-*-linux-*.
311         * testsuite/libffi.call/err_bad_typedef.c: Add xfails.
312
313         * testsuite/libffi.call/stret_medium2.c: Add __UNUSED__ to args.
314         * testsuite/libffi.call/stret_medium.c: Likewise.
315         * testsuite/libffi.call/stret_large2.c: Likewise.
316         * testsuite/libffi.call/stret_large.c:  Likewise.
317
318 2008-12-26  Timothy Wall  <twall@users.sf.net>
319
320         * testsuite/libffi.call/cls_longdouble.c,
321         testsuite/libffi.call/cls_longdouble_va.c,
322         testsuite/libffi.call/cls_align_longdouble.c,
323         testsuite/libffi.call/cls_align_longdouble_split.c,
324         testsuite/libffi.call/cls_align_longdouble_split2.c: mark expected
325         failures on x86_64 cygwin/mingw.
326
327 2008-12-22  Timothy Wall  <twall@users.sf.net>
328
329         * testsuite/libffi.call/closure_fn0.c,
330         testsuite/libffi.call/closure_fn1.c,
331         testsuite/libffi.call/closure_fn2.c,
332         testsuite/libffi.call/closure_fn3.c,
333         testsuite/libffi.call/closure_fn4.c,
334         testsuite/libffi.call/closure_fn5.c,
335         testsuite/libffi.call/closure_fn6.c,
336         testsuite/libffi.call/closure_loc_fn0.c,
337         testsuite/libffi.call/closure_stdcall.c,
338         testsuite/libffi.call/cls_align_pointer.c,
339         testsuite/libffi.call/cls_pointer.c,
340         testsuite/libffi.call/cls_pointer_stack.c: use portable cast from
341         pointer to integer (intptr_t).
342         * testsuite/libffi.call/cls_longdouble.c: disable for win64.
343
344 2008-07-24  Anthony Green  <green@redhat.com>
345
346         * testsuite/libffi.call/cls_dbls_struct.c,
347         testsuite/libffi.call/cls_double_va.c,
348         testsuite/libffi.call/cls_longdouble.c,
349         testsuite/libffi.call/cls_longdouble_va.c,
350         testsuite/libffi.call/cls_pointer.c,
351         testsuite/libffi.call/cls_pointer_stack.c,
352         testsuite/libffi.call/err_bad_abi.c: Clean up failures from
353         compiler warnings.
354
355 2008-03-04  Anthony Green  <green@redhat.com>
356             Blake Chaffin
357             hos@tamanegi.org
358
359         * testsuite/libffi.call/cls_align_longdouble_split2.c
360           testsuite/libffi.call/cls_align_longdouble_split.c
361           testsuite/libffi.call/cls_dbls_struct.c
362           testsuite/libffi.call/cls_double_va.c
363           testsuite/libffi.call/cls_longdouble.c
364           testsuite/libffi.call/cls_longdouble_va.c
365           testsuite/libffi.call/cls_pointer.c
366           testsuite/libffi.call/cls_pointer_stack.c
367           testsuite/libffi.call/err_bad_abi.c
368           testsuite/libffi.call/err_bad_typedef.c
369           testsuite/libffi.call/stret_large2.c
370           testsuite/libffi.call/stret_large.c
371           testsuite/libffi.call/stret_medium2.c
372           testsuite/libffi.call/stret_medium.c: New tests from Apple.
373
374 2009-06-05  Andrew Haley  <aph@redhat.com>
375
376         * src/x86/ffitarget.h, src/x86/ffi.c: Merge stdcall changes from
377         libffi.
378
379 2009-06-04  Andrew Haley  <aph@redhat.com>
380
381         * src/x86/ffitarget.h, src/x86/win32.S, src/x86/ffi.c: Back out
382         stdcall changes.
383
384 2008-02-26  Anthony Green  <green@redhat.com>
385             Thomas Heller  <theller@ctypes.org>
386
387         * src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C
388         comment.
389
390 2008-02-03  Timothy Wall  <twall@users.sf.net>
391
392         * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return
393           offset based on code pointer, not data pointer.
394
395 2008-01-31  Timothy Wall <twall@users.sf.net>
396
397         * testsuite/libffi.call/closure_stdcall.c: Add test for stdcall
398         closures.
399         * src/x86/ffitarget.h: Increase size of trampoline for stdcall
400         closures.
401         * src/x86/win32.S: Add assembly for stdcall closure.
402         * src/x86/ffi.c: Initialize stdcall closure trampoline.
403
404 2009-06-04  Andrew Haley  <aph@redhat.com>
405
406         * include/ffi.h.in: Change void (*)() to void (*)(void).
407         * src/x86/ffi.c: Likewise.
408
409 2009-06-04  Andrew Haley  <aph@redhat.com>
410
411         * src/powerpc/ppc_closure.S: Insert licence header.
412         * src/powerpc/linux64_closure.S: Likewise.
413         * src/m68k/sysv.S: Likewise.
414
415         * src/sh64/ffi.c: Change void (*)() to void (*)(void).
416         * src/powerpc/ffi.c: Likewise.
417         * src/powerpc/ffi_darwin.c: Likewise.
418         * src/m32r/ffi.c: Likewise.
419         * src/sh64/ffi.c: Likewise.
420         * src/x86/ffi64.c: Likewise.
421         * src/alpha/ffi.c: Likewise.
422         * src/alpha/osf.S: Likewise.
423         * src/frv/ffi.c: Likewise.
424         * src/s390/ffi.c: Likewise.
425         * src/pa/ffi.c: Likewise.
426         * src/pa/hpux32.S: Likewise.
427         * src/ia64/unix.S: Likewise.
428         * src/ia64/ffi.c: Likewise.
429         * src/sparc/ffi.c: Likewise.
430         * src/mips/ffi.c: Likewise.
431         * src/sh/ffi.c: Likewise.
432
433 2008-02-15  David Daney  <ddaney@avtrex.com>
434
435         * src/mips/ffi.c (USE__BUILTIN___CLEAR_CACHE):
436         Define (conditionally), and use it to include cachectl.h.
437         (ffi_prep_closure_loc): Fix cache flushing.
438         * src/mips/ffitarget.h (_ABIN32, _ABI64, _ABIO32): Define.
439
440 2009-06-04  Andrew Haley  <aph@redhat.com>
441
442         include/ffi.h.in,
443         src/arm/ffitarget.h,
444         src/arm/ffi.c,
445         src/arm/sysv.S,
446         src/powerpc/ffitarget.h,
447         src/closures.c,
448         src/sh64/ffitarget.h,
449         src/sh64/ffi.c,
450         src/sh64/sysv.S,
451         src/types.c,
452         src/x86/ffi64.c,
453         src/x86/ffitarget.h,
454         src/x86/win32.S,
455         src/x86/darwin.S,
456         src/x86/ffi.c,
457         src/x86/sysv.S,
458         src/x86/unix64.S,
459         src/alpha/ffitarget.h,
460         src/alpha/ffi.c,
461         src/alpha/osf.S,
462         src/m68k/ffitarget.h,
463         src/frv/ffitarget.h,
464         src/frv/ffi.c,
465         src/s390/ffitarget.h,
466         src/s390/sysv.S,
467         src/cris/ffitarget.h,
468         src/pa/linux.S,
469         src/pa/ffitarget.h,
470         src/pa/ffi.c,
471         src/raw_api.c,
472         src/ia64/ffitarget.h,
473         src/ia64/unix.S,
474         src/ia64/ffi.c,
475         src/ia64/ia64_flags.h,
476         src/java_raw_api.c,
477         src/debug.c,
478         src/sparc/v9.S,
479         src/sparc/ffitarget.h,
480         src/sparc/ffi.c,
481         src/sparc/v8.S,
482         src/mips/ffitarget.h,
483         src/mips/n32.S,
484         src/mips/o32.S,
485         src/mips/ffi.c,
486         src/prep_cif.c,
487         src/sh/ffitarget.h,
488         src/sh/ffi.c,
489         src/sh/sysv.S: Update license text.
490
491 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
492
493         * src/x86/win32.S (_ffi_closure_STDCALL):  New function.
494         (.eh_frame):  Add FDE for it.
495
496 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
497
498         * configure.ac:  Also check if assembler supports pc-relative
499         relocs on X86_WIN32 targets.
500         * configure:  Regenerate.
501         * src/x86/win32.S (ffi_prep_args):  Declare extern, not global.
502         (_ffi_call_SYSV):  Add missing function type symbol .def and
503         add EH markup labels.
504         (_ffi_call_STDCALL):  Likewise.
505         (_ffi_closure_SYSV):  Likewise.
506         (_ffi_closure_raw_SYSV):  Likewise.
507         (.eh_frame):  Add hand-crafted EH data.
508
509 2009-04-09  Jakub Jelinek  <jakub@redhat.com>
510
511         * testsuite/lib/libffi-dg.exp: Change copyright header to refer to
512         version 3 of the GNU General Public License and to point readers
513         at the COPYING3 file and the FSF's license web page.
514         * testsuite/libffi.call/call.exp: Likewise.
515         * testsuite/libffi.special/special.exp: Likewise.
516
517 2009-03-01  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
518
519         * configure: Regenerate.
520
521 2008-12-18  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
522
523         PR libffi/26048
524         * configure.ac (HAVE_AS_X86_PCREL): New test.
525         * configure: Regenerate.
526         * fficonfig.h.in: Regenerate.
527         * src/x86/sysv.S [!FFI_NO_RAW_API]: Precalculate
528         RAW_CLOSURE_CIF_OFFSET, RAW_CLOSURE_FUN_OFFSET,
529         RAW_CLOSURE_USER_DATA_OFFSET for the Solaris 10/x86 assembler.
530         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
531         * src/x86/unix64.S (.Lstore_table): Move to .text section.
532         (.Lload_table): Likewise.
533         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
534
535 2008-12-18  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
536
537         * configure: Regenerate.
538
539 2008-11-21  Eric Botcazou  <ebotcazou@adacore.com>
540
541         * src/sparc/ffi.c (ffi_prep_cif_machdep): Add support for
542         signed/unsigned int8/16 return values.
543         * src/sparc/v8.S (ffi_call_v8): Likewise.
544         (ffi_closure_v8): Likewise.
545
546 2008-09-26  Peter O'Gorman  <pogma@thewrittenword.com>
547             Steve Ellcey  <sje@cup.hp.com>
548
549         * configure: Regenerate for new libtool.
550         * Makefile.in: Ditto.
551         * include/Makefile.in: Ditto.
552         * aclocal.m4: Ditto.
553
554 2008-08-25  Andreas Tobler  <a.tobler@schweiz.org>
555
556         * src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and
557         FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum.
558         Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT.
559         Adjust copyright notice.
560         * src/powerpc/ffi.c: Add two new flags to indicate if we have one
561         register or two register to use for FFI_SYSV structs.
562         (ffi_prep_cif_machdep): Pass the right register flag introduced above.
563         (ffi_closure_helper_SYSV): Fix the return type for
564         FFI_SYSV_TYPE_SMALL_STRUCT. Comment.
565         Adjust copyright notice.
566
567 2008-07-16  Kaz Kojima  <kkojima@gcc.gnu.org>
568
569         * src/sh/ffi.c (ffi_prep_closure_loc): Turn INSN into an unsigned
570         int.
571
572 2008-06-17  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
573
574         * configure: Regenerate.
575         * include/Makefile.in: Regenerate.
576         * testsuite/Makefile.in: Regenerate.
577
578 2008-06-07  Joseph Myers  <joseph@codesourcery.com>
579
580         * configure.ac (parisc*-*-linux*, powerpc-*-sysv*,
581         powerpc-*-beos*): Remove.
582         * configure: Regenerate.
583
584 2008-05-09  Julian Brown  <julian@codesourcery.com>
585
586         * Makefile.am (LTLDFLAGS): New.
587         (libffi_la_LDFLAGS): Use above.
588         * Makefile.in: Regenerate.
589
590 2008-04-18  Paolo Bonzini  <bonzini@gnu.org>
591
592         PR bootstrap/35457
593         * aclocal.m4: Regenerate.
594         * configure: Regenerate.
595
596 2008-03-26  Kaz Kojima  <kkojima@gcc.gnu.org>
597
598         * src/sh/sysv.S: Add .note.GNU-stack on Linux.
599         * src/sh64/sysv.S: Likewise.
600
601 2008-03-26  Daniel Jacobowitz  <dan@debian.org>
602
603         * src/arm/sysv.S: Fix ARM comment marker.
604
605 2008-03-26  Jakub Jelinek  <jakub@redhat.com>
606
607         * src/alpha/osf.S: Add .note.GNU-stack on Linux.
608         * src/s390/sysv.S: Likewise.
609         * src/powerpc/ppc_closure.S: Likewise.
610         * src/powerpc/sysv.S: Likewise.
611         * src/x86/unix64.S: Likewise.
612         * src/x86/sysv.S: Likewise.
613         * src/sparc/v8.S: Likewise.
614         * src/sparc/v9.S: Likewise.
615         * src/m68k/sysv.S: Likewise.
616         * src/arm/sysv.S: Likewise.
617
618 2008-03-16  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
619
620         * aclocal.m4: Regenerate.
621         * configure: Likewise.
622         * Makefile.in: Likewise.
623         * include/Makefile.in: Likewise.
624         * testsuite/Makefile.in: Likewise.
625
626 2008-02-12  Bjoern Koenig  <bkoenig@alpha-tierchen.de>
627             Andreas Tobler  <a.tobler@schweiz.org>
628
629         * configure.ac: Add amd64-*-freebsd* target.
630         * configure: Regenerate.
631
632 2008-01-30  H.J. Lu  <hongjiu.lu@intel.com>
633
634         PR libffi/34612
635         * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when
636         returning struct.
637
638         * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer"
639         tests.
640
641 2008-01-24  David Edelsohn  <edelsohn@gnu.org>
642
643         * configure: Regenerate.
644
645 2008-01-06  Andreas Tobler  <a.tobler@schweiz.org>
646
647         * src/x86/ffi.c (ffi_prep_cif_machdep): Fix thinko.
648
649 2008-01-05  Andreas Tobler  <a.tobler@schweiz.org>
650
651         PR testsuite/32843
652         * src/x86/ffi.c (ffi_prep_cif_machdep): Add code for
653         signed/unsigned int8/16 for X86_DARWIN.
654         Updated copyright info.
655         Handle one and two byte structs with special cif->flags.
656         * src/x86/ffitarget.h: Add special types for one and two byte structs.
657         Updated copyright info.
658         * src/x86/darwin.S (ffi_call_SYSV): Rewrite to use a jump table like
659         sysv.S
660         Remove code to pop args from the stack after call.
661         Special-case signed/unsigned for int8/16, one and two byte structs.
662         (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
663         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
664         FFI_TYPE_SINT32.
665         Updated copyright info.
666
667 2007-12-08  David Daney  <ddaney@avtrex.com>
668
669         * src/mips/n32.S (ffi_call_N32):  Replace dadd with ADDU, dsub with 
670         SUBU, add with ADDU and use smaller code sequences.
671
672 2007-12-07  David Daney  <ddaney@avtrex.com>
673
674         * src/mips/ffi.c (ffi_prep_cif_machdep): Handle long double return
675         type.
676
677 2007-12-06  David Daney  <ddaney@avtrex.com>
678
679         * include/ffi.h.in (FFI_SIZEOF_JAVA_RAW): Define if not already
680         defined.
681         (ffi_java_raw): New typedef.
682         (ffi_java_raw_call, ffi_java_ptrarray_to_raw,
683         ffi_java_raw_to_ptrarray): Change parameter types from ffi_raw to
684         ffi_java_raw.
685         (ffi_java_raw_closure) : Same.
686         (ffi_prep_java_raw_closure, ffi_prep_java_raw_closure_loc): Change
687         parameter types.
688         * src/java_raw_api.c (ffi_java_raw_size):  Replace FFI_SIZEOF_ARG with
689         FFI_SIZEOF_JAVA_RAW.
690         (ffi_java_raw_to_ptrarray): Change type of raw to ffi_java_raw.
691         Replace FFI_SIZEOF_ARG with FFI_SIZEOF_JAVA_RAW. Use
692         sizeof(ffi_java_raw) for alignment calculations.
693         (ffi_java_ptrarray_to_raw): Same.
694         (ffi_java_rvalue_to_raw): Add special handling for FFI_TYPE_POINTER
695         if FFI_SIZEOF_JAVA_RAW == 4.
696         (ffi_java_raw_to_rvalue): Same.
697         (ffi_java_raw_call): Change type of raw to ffi_java_raw.
698         (ffi_java_translate_args): Same.
699         (ffi_prep_java_raw_closure_loc, ffi_prep_java_raw_closure): Change
700         parameter types.
701         * src/mips/ffitarget.h (FFI_SIZEOF_JAVA_RAW): Define for N32 ABI.
702
703 2007-12-06  David Daney  <ddaney@avtrex.com>
704
705         * src/mips/n32.S (ffi_closure_N32): Use 64-bit add instruction on 
706         pointer values.
707
708 2007-12-01  Andreas Tobler  <a.tobler@schweiz.org>
709
710         PR libffi/31937
711         * src/powerpc/ffitarget.h: Introduce new ABI FFI_LINUX_SOFT_FLOAT.
712         Add local FFI_TYPE_UINT128 to handle soft-float long-double-128.
713         * src/powerpc/ffi.c: Distinguish between __NO_FPRS__ and not and
714         set the NUM_FPR_ARG_REGISTERS according to.
715         Add support for potential soft-float support under hard-float
716         architecture.
717         (ffi_prep_args_SYSV): Set NUM_FPR_ARG_REGISTERS to 0 in case of
718         FFI_LINUX_SOFT_FLOAT, handle float, doubles and long-doubles according
719         to the FFI_LINUX_SOFT_FLOAT ABI.
720         (ffi_prep_cif_machdep): Likewise.
721         (ffi_closure_helper_SYSV): Likewise.
722         * src/powerpc/ppc_closure.S: Make sure not to store float/double
723         on archs where __NO_FPRS__ is true.
724         Add FFI_TYPE_UINT128 support.
725         * src/powerpc/sysv.S: Add support for soft-float long-double-128.
726         Adjust copyright notice.
727
728 2007-11-25  Andreas Tobler  <a.tobler@schweiz.org>
729
730         * src/closures.c: Move defintion of MAYBE_UNUSED from here to ...
731         * include/ffi_common.h: ... here.
732         Update copyright.
733
734 2007-11-17  Andreas Tobler  <a.tobler@schweiz.org>
735
736         * src/powerpc/sysv.S: Load correct cr to compare if we have long double.
737         * src/powerpc/linux64.S: Likewise.
738         * src/powerpc/ffi.c: Add a comment to show which part goes into cr6.
739         * testsuite/libffi.call/return_ldl.c: New test.
740
741 2007-09-04    <aph@redhat.com>
742
743         * src/arm/sysv.S (UNWIND): New.
744         (Whole file): Conditionally compile unwinder directives.
745         * src/arm/sysv.S: Add unwinder directives.
746
747         * src/arm/ffi.c (ffi_prep_args): Align structs by at least 4 bytes.
748         Only treat r0 as a struct address if we're actually returning a
749         struct by address.
750         Only copy the bytes that are actually within a struct.
751         (ffi_prep_cif_machdep): A Composite Type not larger than 4 bytes
752         is returned in r0, not passed by address.
753         (ffi_call): Allocate a word-sized temporary for the case where
754         a composite is returned in r0.
755         (ffi_prep_incoming_args_SYSV): Align as necessary.
756
757 2007-08-05  Steven Newbury  <s_j_newbury@yahoo.co.uk>
758
759         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Use __clear_cache instead of 
760         directly using the sys_cacheflush syscall.
761
762 2007-07-27  Andrew Haley  <aph@redhat.com>
763
764         * src/arm/sysv.S (ffi_closure_SYSV): Add soft-float.
765
766 2007-09-03  Maciej W. Rozycki  <macro@linux-mips.org>
767
768         * Makefile.am: Unify MIPS_IRIX and MIPS_LINUX into MIPS.
769         * configure.ac: Likewise.
770         * Makefile.in: Regenerate.
771         * include/Makefile.in: Likewise.
772         * testsuite/Makefile.in: Likewise.
773         * configure: Likewise.
774
775 2007-08-24  David Daney  <ddaney@avtrex.com>
776
777         * testsuite/libffi.call/return_sl.c: New test.
778
779 2007-08-10  David Daney  <ddaney@avtrex.com>
780
781         * testsuite/libffi.call/cls_multi_ushort.c,
782         testsuite/libffi.call/cls_align_uint16.c,
783         testsuite/libffi.call/nested_struct1.c,
784         testsuite/libffi.call/nested_struct3.c,
785         testsuite/libffi.call/cls_7_1_byte.c,
786         testsuite/libffi.call/nested_struct5.c,
787         testsuite/libffi.call/cls_double.c,
788         testsuite/libffi.call/nested_struct7.c,
789         testsuite/libffi.call/cls_sint.c,
790         testsuite/libffi.call/nested_struct9.c,
791         testsuite/libffi.call/cls_20byte1.c,
792         testsuite/libffi.call/cls_multi_sshortchar.c,
793         testsuite/libffi.call/cls_align_sint64.c,
794         testsuite/libffi.call/cls_3byte2.c,
795         testsuite/libffi.call/cls_multi_schar.c,
796         testsuite/libffi.call/cls_multi_uchar.c,
797         testsuite/libffi.call/cls_19byte.c,
798         testsuite/libffi.call/cls_9byte1.c,
799         testsuite/libffi.call/cls_align_float.c,
800         testsuite/libffi.call/closure_fn1.c,
801         testsuite/libffi.call/problem1.c,
802         testsuite/libffi.call/closure_fn3.c,
803         testsuite/libffi.call/cls_sshort.c,
804         testsuite/libffi.call/closure_fn5.c,
805         testsuite/libffi.call/cls_align_double.c,
806         testsuite/libffi.call/nested_struct.c,
807         testsuite/libffi.call/cls_2byte.c,
808         testsuite/libffi.call/nested_struct10.c,
809         testsuite/libffi.call/cls_4byte.c,
810         testsuite/libffi.call/cls_6byte.c,
811         testsuite/libffi.call/cls_8byte.c,
812         testsuite/libffi.call/cls_multi_sshort.c,
813         testsuite/libffi.call/cls_align_sint16.c,
814         testsuite/libffi.call/cls_align_uint32.c,
815         testsuite/libffi.call/cls_20byte.c,
816         testsuite/libffi.call/cls_float.c,
817         testsuite/libffi.call/nested_struct2.c,
818         testsuite/libffi.call/cls_5_1_byte.c,
819         testsuite/libffi.call/nested_struct4.c,
820         testsuite/libffi.call/cls_24byte.c,
821         testsuite/libffi.call/nested_struct6.c,
822         testsuite/libffi.call/cls_64byte.c,
823         testsuite/libffi.call/nested_struct8.c,
824         testsuite/libffi.call/cls_uint.c,
825         testsuite/libffi.call/cls_multi_ushortchar.c,
826         testsuite/libffi.call/cls_schar.c,
827         testsuite/libffi.call/cls_uchar.c,
828         testsuite/libffi.call/cls_align_uint64.c,
829         testsuite/libffi.call/cls_ulonglong.c,
830         testsuite/libffi.call/cls_align_longdouble.c,
831         testsuite/libffi.call/cls_1_1byte.c,
832         testsuite/libffi.call/cls_12byte.c,
833         testsuite/libffi.call/cls_3_1byte.c,
834         testsuite/libffi.call/cls_3byte1.c,
835         testsuite/libffi.call/cls_4_1byte.c,
836         testsuite/libffi.call/cls_6_1_byte.c,
837         testsuite/libffi.call/cls_16byte.c,
838         testsuite/libffi.call/cls_18byte.c,
839         testsuite/libffi.call/closure_fn0.c,
840         testsuite/libffi.call/cls_9byte2.c,
841         testsuite/libffi.call/closure_fn2.c,
842         testsuite/libffi.call/closure_fn4.c,
843         testsuite/libffi.call/cls_ushort.c,
844         testsuite/libffi.call/closure_fn6.c,
845         testsuite/libffi.call/cls_5byte.c,
846         testsuite/libffi.call/cls_align_pointer.c,
847         testsuite/libffi.call/cls_7byte.c,
848         testsuite/libffi.call/cls_align_sint32.c,
849         testsuite/libffi.special/unwindtest_ffi_call.cc,
850         testsuite/libffi.special/unwindtest.cc: Remove xfail for mips64*-*-*.
851
852 2007-08-10  David Daney  <ddaney@avtrex.com>
853
854         PR libffi/28313
855         * configure.ac: Don't treat mips64 as a special case.
856         * Makefile.am (nodist_libffi_la_SOURCES): Add n32.S.
857         * configure: Regenerate
858         * Makefile.in: Ditto.
859         * fficonfig.h.in: Ditto.
860         * src/mips/ffitarget.h (REG_L, REG_S, SUBU, ADDU, SRL, LI): Indent.
861         (LA, EH_FRAME_ALIGN, FDE_ADDR_BYTES): New preprocessor macros.
862         (FFI_DEFAULT_ABI): Set for n64 case.
863         (FFI_CLOSURES, FFI_TRAMPOLINE_SIZE): Define for n32 and n64 cases.
864         * src/mips/n32.S (ffi_call_N32): Add debug macros and labels for FDE.
865         (ffi_closure_N32): New function.
866         (.eh_frame): New section
867         * src/mips/o32.S: Clean up comments.
868         (ffi_closure_O32): Pass ffi_closure parameter in $12.
869         * src/mips/ffi.c: Use FFI_MIPS_N32 instead of
870         _MIPS_SIM == _ABIN32 throughout.
871         (FFI_MIPS_STOP_HERE): New, use in place of
872         ffi_stop_here.
873         (ffi_prep_args): Use unsigned long to hold pointer values.  Rewrite
874         to support n32/n64 ABIs.
875         (calc_n32_struct_flags): Rewrite.
876         (calc_n32_return_struct_flags): Remove unused variable.  Reverse
877         position of flag bits.
878         (ffi_prep_cif_machdep): Rewrite n32 portion.
879         (ffi_call): Enable for n64.  Add special handling for small structure
880         return values.
881         (ffi_prep_closure_loc): Add n32 and n64 support.
882         (ffi_closure_mips_inner_O32): Add cast to silence warning.
883         (copy_struct_N32, ffi_closure_mips_inner_N32): New functions.
884
885 2007-08-08  David Daney  <ddaney@avtrex.com>
886
887         * testsuite/libffi.call/ffitest.h (ffi_type_mylong): Remove definition.
888         * testsuite/libffi.call/cls_align_uint16.c (main): Use correct type
889         specifiers.
890         * testsuite/libffi.call/nested_struct1.c (main): Ditto.
891         * testsuite/libffi.call/cls_sint.c (main): Ditto.
892         * testsuite/libffi.call/nested_struct9.c (main): Ditto.
893         * testsuite/libffi.call/cls_20byte1.c (main): Ditto.
894         * testsuite/libffi.call/cls_9byte1.c (main): Ditto.
895         * testsuite/libffi.call/closure_fn1.c (main): Ditto.
896         * testsuite/libffi.call/closure_fn3.c (main): Ditto.
897         * testsuite/libffi.call/return_dbl2.c (main): Ditto.
898         * testsuite/libffi.call/cls_sshort.c (main): Ditto.
899         * testsuite/libffi.call/return_fl3.c (main): Ditto.
900         * testsuite/libffi.call/closure_fn5.c (main): Ditto.
901         * testsuite/libffi.call/nested_struct.c (main): Ditto.
902         * testsuite/libffi.call/nested_struct10.c (main): Ditto.
903         * testsuite/libffi.call/return_ll1.c (main): Ditto.
904         * testsuite/libffi.call/cls_8byte.c (main): Ditto.
905         * testsuite/libffi.call/cls_align_uint32.c (main): Ditto.
906         * testsuite/libffi.call/cls_align_sint16.c (main): Ditto.
907         * testsuite/libffi.call/cls_20byte.c (main): Ditto.
908         * testsuite/libffi.call/nested_struct2.c (main): Ditto.
909         * testsuite/libffi.call/cls_24byte.c (main): Ditto.
910         * testsuite/libffi.call/nested_struct6.c (main): Ditto.
911         * testsuite/libffi.call/cls_uint.c (main): Ditto.
912         * testsuite/libffi.call/cls_12byte.c (main): Ditto.
913         * testsuite/libffi.call/cls_16byte.c (main): Ditto.
914         * testsuite/libffi.call/closure_fn0.c (main): Ditto.
915         * testsuite/libffi.call/cls_9byte2.c (main): Ditto.
916         * testsuite/libffi.call/closure_fn2.c (main): Ditto.
917         * testsuite/libffi.call/return_dbl1.c (main): Ditto.
918         * testsuite/libffi.call/closure_fn4.c (main): Ditto.
919         * testsuite/libffi.call/closure_fn6.c (main): Ditto.
920         * testsuite/libffi.call/cls_align_sint32.c (main): Ditto.
921
922 2007-08-07  Andrew Haley  <aph@redhat.com>
923
924         * src/x86/sysv.S (ffi_closure_raw_SYSV): Fix typo in previous
925         checkin.
926
927 2007-08-06  Andrew Haley  <aph@redhat.com>
928
929         PR testsuite/32843
930         * src/x86/sysv.S (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
931         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
932         FFI_TYPE_SINT32.
933
934 2007-08-02  David Daney  <ddaney@avtrex.com>
935
936         * testsuite/libffi.call/return_ul.c (main): Define return type as
937         ffi_arg.  Use proper printf conversion specifier.
938         
939 2007-07-30  Andrew Haley  <aph@redhat.com>
940
941         PR testsuite/32843
942         * src/x86/ffi.c (ffi_prep_cif_machdep): in x86 case, add code for
943         signed/unsigned int8/16.
944         * src/x86/sysv.S (ffi_call_SYSV): Rewrite to:
945         Use a jump table.
946         Remove code to pop args from the stack after call.
947         Special-case signed/unsigned int8/16.
948         * testsuite/libffi.call/return_sc.c (main): Revert.
949
950 2007-07-26  Richard Guenther  <rguenther@suse.de>
951
952         PR testsuite/32843
953         * testsuite/libffi.call/return_sc.c (main): Verify call
954         result as signed char, not ffi_arg.
955
956 2007-07-16  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
957
958         * configure.ac (i?86-*-solaris2.1[0-9]): Set TARGET to X86_64.
959         * configure: Regenerate.
960
961 2007-07-11  David Daney  <ddaney@avtrex.com>
962
963         * src/mips/ffi.c: Don't include sys/cachectl.h.
964         (ffi_prep_closure_loc): Use __builtin___clear_cache() instead of
965         cacheflush().
966
967 2007-05-18  Aurelien Jarno  <aurelien@aurel32.net>
968
969         * src/arm/ffi.c (ffi_prep_closure_loc): Renamed and ajusted
970         from (ffi_prep_closure): ... this.
971         (FFI_INIT_TRAMPOLINE): Adjust.
972
973 2005-12-31  Phil Blundell  <pb@reciva.com>
974
975         * src/arm/ffi.c (ffi_prep_incoming_args_SYSV, 
976         ffi_closure_SYSV_inner, ffi_prep_closure): New, add closure support.
977         * src/arm/sysv.S(ffi_closure_SYSV): Likewise.
978         * src/arm/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
979         (FFI_CLOSURES): Enable closure support.
980
981 2007-07-03  Andrew Haley  <aph@hedges.billgatliff.com>
982
983         * testsuite/libffi.call/cls_multi_ushort.c,
984         testsuite/libffi.call/cls_align_uint16.c,
985         testsuite/libffi.call/nested_struct1.c,
986         testsuite/libffi.call/nested_struct3.c,
987         testsuite/libffi.call/cls_7_1_byte.c,
988         testsuite/libffi.call/cls_double.c,
989         testsuite/libffi.call/nested_struct5.c,
990         testsuite/libffi.call/nested_struct7.c,
991         testsuite/libffi.call/cls_sint.c,
992         testsuite/libffi.call/nested_struct9.c,
993         testsuite/libffi.call/cls_20byte1.c,
994         testsuite/libffi.call/cls_multi_sshortchar.c,
995         testsuite/libffi.call/cls_align_sint64.c,
996         testsuite/libffi.call/cls_3byte2.c,
997         testsuite/libffi.call/cls_multi_schar.c,
998         testsuite/libffi.call/cls_multi_uchar.c,
999         testsuite/libffi.call/cls_19byte.c,
1000         testsuite/libffi.call/cls_9byte1.c,
1001         testsuite/libffi.call/cls_align_float.c,
1002         testsuite/libffi.call/closure_fn1.c,
1003         testsuite/libffi.call/problem1.c,
1004         testsuite/libffi.call/closure_fn3.c,
1005         testsuite/libffi.call/cls_sshort.c,
1006         testsuite/libffi.call/closure_fn5.c,
1007         testsuite/libffi.call/cls_align_double.c,
1008         testsuite/libffi.call/cls_2byte.c,
1009         testsuite/libffi.call/nested_struct.c,
1010         testsuite/libffi.call/nested_struct10.c,
1011         testsuite/libffi.call/cls_4byte.c,
1012         testsuite/libffi.call/cls_6byte.c,
1013         testsuite/libffi.call/cls_8byte.c,
1014         testsuite/libffi.call/cls_multi_sshort.c,
1015         testsuite/libffi.call/cls_align_uint32.c,
1016         testsuite/libffi.call/cls_align_sint16.c,
1017         testsuite/libffi.call/cls_float.c,
1018         testsuite/libffi.call/cls_20byte.c,
1019         testsuite/libffi.call/cls_5_1_byte.c,
1020         testsuite/libffi.call/nested_struct2.c,
1021         testsuite/libffi.call/cls_24byte.c,
1022         testsuite/libffi.call/nested_struct4.c,
1023         testsuite/libffi.call/nested_struct6.c,
1024         testsuite/libffi.call/cls_64byte.c,
1025         testsuite/libffi.call/nested_struct8.c,
1026         testsuite/libffi.call/cls_uint.c,
1027         testsuite/libffi.call/cls_multi_ushortchar.c,
1028         testsuite/libffi.call/cls_schar.c,
1029         testsuite/libffi.call/cls_uchar.c,
1030         testsuite/libffi.call/cls_align_uint64.c,
1031         testsuite/libffi.call/cls_ulonglong.c,
1032         testsuite/libffi.call/cls_align_longdouble.c,
1033         testsuite/libffi.call/cls_1_1byte.c,
1034         testsuite/libffi.call/cls_12byte.c,
1035         testsuite/libffi.call/cls_3_1byte.c,
1036         testsuite/libffi.call/cls_3byte1.c,
1037         testsuite/libffi.call/cls_4_1byte.c,
1038         testsuite/libffi.call/cls_6_1_byte.c,
1039         testsuite/libffi.call/cls_16byte.c,
1040         testsuite/libffi.call/cls_18byte.c,
1041         testsuite/libffi.call/closure_fn0.c,
1042         testsuite/libffi.call/cls_9byte2.c,
1043         testsuite/libffi.call/closure_fn2.c,
1044         testsuite/libffi.call/closure_fn4.c,
1045         testsuite/libffi.call/cls_ushort.c,
1046         testsuite/libffi.call/closure_fn6.c,
1047         testsuite/libffi.call/cls_5byte.c,
1048         testsuite/libffi.call/cls_align_pointer.c,
1049         testsuite/libffi.call/cls_7byte.c,
1050         testsuite/libffi.call/cls_align_sint32.c,
1051         testsuite/libffi.special/unwindtest_ffi_call.cc,
1052         testsuite/libffi.special/unwindtest.cc: Enable for ARM.
1053
1054 2007-07-05  H.J. Lu  <hongjiu.lu@intel.com>
1055
1056         * aclocal.m4: Regenerated.
1057
1058 2007-06-02  Paolo Bonzini  <bonzini@gnu.org>
1059
1060         * configure: Regenerate.
1061
1062 2007-05-23  Steve Ellcey  <sje@cup.hp.com>
1063
1064         * Makefile.in: Regenerate.
1065         * configure: Regenerate.
1066         * aclocal.m4: Regenerate.
1067         * include/Makefile.in: Regenerate.
1068         * testsuite/Makefile.in: Regenerate.
1069
1070 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1071
1072         * src/m68k/ffi.c (ffi_prep_incoming_args_SYSV,
1073         ffi_closure_SYSV_inner,ffi_prep_closure): New, add closure support.
1074         * src/m68k/sysv.S(ffi_closure_SYSV,ffi_closure_struct_SYSV): Likewise.
1075         * src/m68k/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1076         (FFI_CLOSURES): Enable closure support.
1077
1078 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1079
1080         * configure.ac (HAVE_AS_CFI_PSEUDO_OP): New test.
1081         * configure: Regenerate.
1082         * fficonfig.h.in: Regenerate.
1083         * src/m68k/sysv.S (CFI_STARTPROC,CFI_ENDPROC,
1084         CFI_OFFSET,CFI_DEF_CFA): New macros.
1085         (ffi_call_SYSV): Add callframe annotation.
1086
1087 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1088
1089         * src/m68k/ffi.c (ffi_prep_args,ffi_prep_cif_machdep): Fix
1090         numerous test suite failures.
1091         * src/m68k/sysv.S (ffi_call_SYSV): Likewise.
1092
1093 2007-04-11  Paolo Bonzini  <bonzini@gnu.org>
1094
1095         * Makefile.am (EXTRA_DIST): Bring up to date.
1096         * Makefile.in: Regenerate.
1097         * src/frv/eabi.S: Remove RCS keyword.
1098
1099 2007-04-06  Richard Henderson  <rth@redhat.com>
1100
1101         * configure.ac: Tidy target case.
1102         (HAVE_LONG_DOUBLE): Allow the target to override.
1103         * configure: Regenerate.
1104         * include/ffi.h.in: Don't define ffi_type_foo if
1105         LIBFFI_HIDE_BASIC_TYPES is defined.
1106         (ffi_type_longdouble): If not HAVE_LONG_DOUBLE, define
1107         to ffi_type_double.
1108         * types.c (LIBFFI_HIDE_BASIC_TYPES): Define.
1109         (FFI_TYPEDEF, ffi_type_void): Mark the data const.
1110         (ffi_type_longdouble): Special case for Alpha.  Don't define
1111         if long double == double.
1112
1113         * src/alpha/ffi.c (FFI_TYPE_LONGDOUBLE): Assert unique value.
1114         (ffi_prep_cif_machdep): Handle it as the 128-bit type.
1115         (ffi_call, ffi_closure_osf_inner): Likewise.
1116         (ffi_closure_osf_inner): Likewise.  Mark hidden.
1117         (ffi_call_osf, ffi_closure_osf): Mark hidden.
1118         * src/alpha/ffitarget.h (FFI_LAST_ABI): Tidy definition.
1119         * src/alpha/osf.S (ffi_call_osf, ffi_closure_osf): Mark hidden.
1120         (load_table): Handle 128-bit long double.
1121
1122         * testsuite/libffi.call/float4.c: Add -mieee for alpha.
1123
1124 2007-04-06  Tom Tromey  <tromey@redhat.com>
1125
1126         PR libffi/31491:
1127         * README: Fixed bug in example.
1128
1129 2007-04-03  Jakub Jelinek  <jakub@redhat.com>
1130
1131         * src/closures.c: Include sys/statfs.h.
1132         (_GNU_SOURCE): Define on Linux.
1133         (FFI_MMAP_EXEC_SELINUX): Define.
1134         (selinux_enabled): New variable.
1135         (selinux_enabled_check): New function.
1136         (is_selinux_enabled): Define.
1137         (dlmmap): Use it.
1138
1139 2007-03-24  Uros Bizjak  <ubizjak@gmail.com>
1140
1141         * testsuite/libffi.call/return_fl2.c (return_fl): Mark as static.
1142         Use 'volatile float sum' to create sum of floats to avoid false
1143         negative due to excess precision on ix86 targets.
1144         (main): Ditto. 
1145
1146 2007-03-08  Alexandre Oliva  <aoliva@redhat.com>
1147
1148         * src/powerpc/ffi.c (flush_icache): Fix left-over from previous
1149         patch.
1150         (ffi_prep_closure_loc): Remove unneeded casts.  Add needed ones.
1151
1152 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1153
1154         * include/ffi.h.in (ffi_closure_alloc, ffi_closure_free): New.
1155         (ffi_prep_closure_loc): New.
1156         (ffi_prep_raw_closure_loc): New.
1157         (ffi_prep_java_raw_closure_loc): New.
1158         * src/closures.c: New file.
1159         * src/dlmalloc.c [FFI_MMAP_EXEC_WRIT] (struct malloc_segment):
1160         Replace sflags with exec_offset.
1161         [FFI_MMAP_EXEC_WRIT] (mmap_exec_offset, add_segment_exec_offset,
1162         sub_segment_exec_offset): New macros.
1163         (get_segment_flags, set_segment_flags, check_segment_merge): New
1164         macros.
1165         (is_mmapped_segment, is_extern_segment): Use get_segment_flags.
1166         (add_segment, sys_alloc, create_mspace, create_mspace_with_base,
1167         destroy_mspace): Use new macros.
1168         (sys_alloc): Silence warning.
1169         * Makefile.am (libffi_la_SOURCES): Add src/closures.c.
1170         * Makefile.in: Rebuilt.
1171         * src/prep_cif [FFI_CLOSURES] (ffi_prep_closure): Implement in
1172         terms of ffi_prep_closure_loc.
1173         * src/raw_api.c (ffi_prep_raw_closure_loc): Renamed and adjusted
1174         from...
1175         (ffi_prep_raw_closure): ... this.  Re-implement in terms of the
1176         renamed version.
1177         * src/java_raw_api (ffi_prep_java_raw_closure_loc): Renamed and
1178         adjusted from...
1179         (ffi_prep_java_raw_closure): ... this.  Re-implement in terms of
1180         the renamed version.
1181         * src/alpha/ffi.c (ffi_prep_closure_loc): Renamed from
1182         (ffi_prep_closure): ... this.
1183         * src/pa/ffi.c: Likewise.
1184         * src/cris/ffi.c: Likewise.  Adjust.
1185         * src/frv/ffi.c: Likewise.
1186         * src/ia64/ffi.c: Likewise.
1187         * src/mips/ffi.c: Likewise.
1188         * src/powerpc/ffi_darwin.c: Likewise.
1189         * src/s390/ffi.c: Likewise.
1190         * src/sh/ffi.c: Likewise.
1191         * src/sh64/ffi.c: Likewise.
1192         * src/sparc/ffi.c: Likewise.
1193         * src/x86/ffi64.c: Likewise.
1194         * src/x86/ffi.c: Likewise.
1195         (FFI_INIT_TRAMPOLINE): Adjust.
1196         (ffi_prep_raw_closure_loc): Renamed and adjusted from...
1197         (ffi_prep_raw_closure): ... this.
1198         * src/powerpc/ffi.c (ffi_prep_closure_loc): Renamed from
1199         (ffi_prep_closure): ... this.
1200         (flush_icache): Adjust.
1201
1202 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1203
1204         * src/dlmalloc.c: New file, imported version 2.8.3 of Doug
1205         Lea's malloc.
1206
1207 2007-03-01  Brooks Moses  <brooks.moses@codesourcery.com>
1208
1209         * Makefile.am: Add dummy install-pdf target.
1210         * Makefile.in: Regenerate
1211
1212 2007-02-13  Andreas Krebbel  <krebbel1@de.ibm.com>
1213
1214         * src/s390/ffi.c (ffi_prep_args, ffi_prep_cif_machdep,
1215         ffi_closure_helper_SYSV): Add long double handling.
1216
1217 2007-02-02  Jakub Jelinek  <jakub@redhat.com>
1218
1219         * src/powerpc/linux64.S (ffi_call_LINUX64): Move restore of r2
1220         immediately after bctrl instruction.
1221
1222 2007-01-18  Alexandre Oliva  <aoliva@redhat.com>
1223
1224         * Makefile.am (all-recursive, install-recursive,
1225         mostlyclean-recursive, clean-recursive, distclean-recursive,
1226         maintainer-clean-recursive): Add missing targets.
1227         * Makefile.in: Rebuilt.
1228
1229 2006-12-14  Andreas Tobler  <a.tobler@schweiz.org>
1230
1231         * configure.ac: Add TARGET for x86_64-*-darwin*.
1232         * Makefile.am (nodist_libffi_la_SOURCES): Add rules for 64-bit sources
1233         for X86_DARWIN.
1234         * src/x86/ffitarget.h: Set trampoline size for x86_64-*-darwin*.
1235         * src/x86/darwin64.S: New file for x86_64-*-darwin* support.
1236         * configure: Regenerate.
1237         * Makefile.in: Regenerate.
1238         * include/Makefile.in: Regenerate.
1239         * testsuite/Makefile.in: Regenerate.
1240         * testsuite/libffi.special/unwindtest_ffi_call.cc: New test case for
1241         ffi_call only.
1242
1243 2006-12-13  Andreas Tobler <a.tobler@schweiz.org>
1244
1245         * aclocal.m4: Regenerate with aclocal -I .. as written in the
1246         Makefile.am.
1247
1248 2006-10-31  Geoffrey Keating  <geoffk@apple.com>
1249
1250         * src/powerpc/ffi_darwin.c (darwin_adjust_aggregate_sizes): New.
1251         (ffi_prep_cif_machdep): Call darwin_adjust_aggregate_sizes for
1252         Darwin.
1253         * testsuite/libffi.call/nested_struct4.c: Remove Darwin XFAIL.
1254         * testsuite/libffi.call/nested_struct6.c: Remove Darwin XFAIL.
1255
1256 2006-10-10  Paolo Bonzini  <bonzini@gnu.org>
1257             Sandro Tolaini  <tolaini@libero.it>
1258
1259         * configure.ac [i*86-*-darwin*]: Set X86_DARWIN symbol and 
1260         conditional.
1261         * configure: Regenerated.
1262         * Makefile.am (nodist_libffi_la_SOURCES) [X86_DARWIN]: New case.
1263         (EXTRA_DIST): Add src/x86/darwin.S.
1264         * Makefile.in: Regenerated.
1265         * include/Makefile.in: Regenerated.
1266         * testsuite/Makefile.in: Regenerated.
1267
1268         * src/x86/ffi.c (ffi_prep_cif_machdep) [X86_DARWIN]: Treat like
1269         X86_WIN32, and additionally align stack to 16 bytes.
1270         * src/x86/darwin.S: New, based on sysv.S.
1271         * src/prep_cif.c (ffi_prep_cif) [X86_DARWIN]: Align > 8-byte structs.
1272
1273 2006-09-12  David Daney  <ddaney@avtrex.com>
1274
1275         PR libffi/23935
1276         * include/Makefile.am: Install both ffi.h and ffitarget.h in
1277         $(libdir)/gcc/$(target_alias)/$(gcc_version)/include.
1278         * aclocal.m4: Regenerated for automake 1.9.6.
1279         * Makefile.in: Regenerated.
1280         * include/Makefile.in: Regenerated.
1281         * testsuite/Makefile.in: Regenerated.
1282
1283 2006-08-17  Andreas Tobler  <a.tobler@schweiz.ch>
1284
1285         * include/ffi_common.h (struct): Revert accidental commit.
1286
1287 2006-08-15  Andreas Tobler  <a.tobler@schweiz.ch>
1288
1289         * include/ffi_common.h: Remove lint directives.
1290         * include/ffi.h.in: Likewise.
1291
1292 2006-07-25  Torsten Schoenfeld  <kaffeetisch@gmx.de>
1293
1294         * include/ffi.h.in (ffi_type_ulong, ffi_type_slong): Define correctly
1295         for 32-bit architectures.
1296         * testsuite/libffi.call/return_ul.c: New test case.
1297
1298 2006-07-19  David Daney  <ddaney@avtrex.com>
1299
1300         * testsuite/libffi.call/closure_fn6.c: Remove xfail for mips,
1301         xfail remains for mips64.
1302
1303 2006-05-23  Carlos O'Donell  <carlos@codesourcery.com>
1304
1305         * Makefile.am: Add install-html target. Add install-html to .PHONY
1306         * Makefile.in: Regenerate.
1307         * aclocal.m4: Regenerate.
1308         * include/Makefile.in: Regenerate.
1309         * testsuite/Makefile.in: Regenerate.
1310
1311 2006-05-18  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1312
1313         * pa/ffi.c (ffi_prep_args_pa32): Load floating point arguments from
1314         stack slot.
1315
1316 2006-04-22  Andreas Tobler  <a.tobler@schweiz.ch>
1317
1318         * README: Remove notice about 'Crazy Comments'.
1319         * src/debug.c: Remove lint directives. Cleanup white spaces.
1320         * src/java_raw_api.c: Likewise.
1321         * src/prep_cif.c: Likewise.
1322         * src/raw_api.c: Likewise.
1323         * src/ffitest.c: Delete. No longer needed, all test cases migrated
1324         to the testsuite.
1325         * src/arm/ffi.c: Remove lint directives.
1326         * src/m32r/ffi.c: Likewise.
1327         * src/pa/ffi.c: Likewise.
1328         * src/powerpc/ffi.c: Likewise.
1329         * src/powerpc/ffi_darwin.c: Likewise.
1330         * src/sh/ffi.c: Likewise.
1331         * src/sh64/ffi.c: Likewise.
1332         * src/x86/ffi.c: Likewise.
1333         * testsuite/libffi.call/float2.c: Likewise.
1334         * testsuite/libffi.call/promotion.c: Likewise.
1335         * testsuite/libffi.call/struct1.c: Likewise.
1336
1337 2006-04-13  Andreas Tobler  <a.tobler@schweiz.ch>
1338
1339         * src/pa/hpux32.S: Correct unwind offset calculation for
1340         ffi_closure_pa32.
1341         * src/pa/linux.S: Likewise.
1342
1343 2006-04-12  James E Wilson  <wilson@specifix.com>
1344
1345         PR libgcj/26483
1346         * src/ia64/ffi.c (stf_spill, ldf_fill): Rewrite as macros.
1347         (hfa_type_load): Call stf_spill.
1348         (hfa_type_store): Call ldf_fill.
1349         (ffi_call): Adjust calls to above routines.  Add local temps for
1350         macro result.
1351
1352 2006-04-10  Matthias Klose  <doko@debian.org>
1353
1354         * testsuite/lib/libffi-dg.exp (libffi-init): Recognize multilib
1355         directory names containing underscores.
1356
1357 2006-04-07  James E Wilson  <wilson@specifix.com>
1358
1359         * testsuite/libffi.call/float4.c: New testcase.
1360
1361 2006-04-05  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1362             Andreas Tobler  <a.tobler@schweiz.ch>
1363
1364         * Makefile.am: Add PA_HPUX port.
1365         * Makefile.in: Regenerate.
1366         * include/Makefile.in: Likewise.
1367         * testsuite/Makefile.in: Likewise.
1368         * configure.ac: Add PA_HPUX rules.
1369         * configure: Regenerate.
1370         * src/pa/ffitarget.h: Rename linux target to PA_LINUX.
1371         Add PA_HPUX and PA64_HPUX.
1372         Rename FFI_LINUX ABI to FFI_PA32 ABI.
1373         (FFI_TRAMPOLINE_SIZE): Define for 32-bit HP-UX targets.
1374         (FFI_TYPE_SMALL_STRUCT2): Define.
1375         (FFI_TYPE_SMALL_STRUCT4): Likewise.
1376         (FFI_TYPE_SMALL_STRUCT8): Likewise.
1377         (FFI_TYPE_SMALL_STRUCT3): Redefine.
1378         (FFI_TYPE_SMALL_STRUCT5): Likewise.
1379         (FFI_TYPE_SMALL_STRUCT6): Likewise.
1380         (FFI_TYPE_SMALL_STRUCT7): Likewise.
1381         * src/pa/ffi.c (ROUND_DOWN): Delete.
1382         (fldw, fstw, fldd, fstd): Use '__asm__'.
1383         (ffi_struct_type): Add support for FFI_TYPE_SMALL_STRUCT2,
1384         FFI_TYPE_SMALL_STRUCT4 and FFI_TYPE_SMALL_STRUCT8.
1385         (ffi_prep_args_LINUX): Rename to ffi_prep_args_pa32. Update comment.
1386         Simplify incrementing of stack slot variable. Change type of local
1387         'n' to unsigned int.
1388         (ffi_size_stack_LINUX): Rename to ffi_size_stack_pa32. Handle long
1389         double on PA_HPUX.
1390         (ffi_prep_cif_machdep): Likewise.
1391         (ffi_call): Likewise.
1392         (ffi_closure_inner_LINUX): Rename to ffi_closure_inner_pa32. Change
1393         return type to ffi_status. Simplify incrementing of stack slot
1394         variable. Only copy floating point argument registers when PA_LINUX
1395         is true. Reformat debug statement.
1396         Add support for FFI_TYPE_SMALL_STRUCT2, FFI_TYPE_SMALL_STRUCT4 and
1397         FFI_TYPE_SMALL_STRUCT8.
1398         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Add 'extern' to
1399         declaration.
1400         (ffi_prep_closure): Make linux trampoline conditional on PA_LINUX.
1401         Add nops to cache flush.  Add trampoline for PA_HPUX.
1402         * src/pa/hpux32.S: New file.
1403         * src/pa/linux.S (ffi_call_LINUX): Rename to ffi_call_pa32. Rename
1404         ffi_prep_args_LINUX to ffi_prep_args_pa32.
1405         Localize labels. Add support for 2, 4 and 8-byte small structs. Handle
1406         unaligned destinations in 3, 5, 6 and 7-byte small structs. Order
1407         argument type checks so that common argument types appear first.
1408         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Rename
1409         ffi_closure_inner_LINUX to ffi_closure_inner_pa32.
1410
1411 2006-03-24  Alan Modra  <amodra@bigpond.net.au>
1412
1413         * src/powerpc/ffitarget.h (enum ffi_abi): Add FFI_LINUX.  Default
1414         for 32-bit using IBM extended double format.  Fix FFI_LAST_ABI.
1415         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Handle linux variant of
1416         FFI_TYPE_LONGDOUBLE.
1417         (ffi_prep_args64): Assert using IBM extended double.
1418         (ffi_prep_cif_machdep): Don't munge FFI_TYPE_LONGDOUBLE type.
1419         Handle FFI_LINUX FFI_TYPE_LONGDOUBLE return and args.
1420         (ffi_call): Handle FFI_LINUX.
1421         (ffi_closure_helper_SYSV): Non FFI_LINUX long double return needs
1422         gpr3 return pointer as for struct return.  Handle FFI_LINUX
1423         FFI_TYPE_LONGDOUBLE return and args.  Don't increment "nf"
1424         unnecessarily.
1425         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Load both f1 and f2
1426         for FFI_TYPE_LONGDOUBLE.  Move epilogue insns into case table.
1427         Don't use r6 as pointer to results, instead use sp offset.  Don't
1428         make a special call to load lr with case table address, instead
1429         use offset from previous call.
1430         * src/powerpc/sysv.S (ffi_call_SYSV): Save long double return.
1431         * src/powerpc/linux64.S (ffi_call_LINUX64): Simplify long double
1432         return.
1433
1434 2006-03-15  Kaz Kojima  <kkojima@gcc.gnu.org>
1435
1436         * src/sh64/ffi.c (ffi_prep_cif_machdep): Handle float arguments
1437         passed with FP registers correctly.
1438         (ffi_closure_helper_SYSV): Likewise.
1439         * src/sh64/sysv.S: Likewise.
1440
1441 2006-03-01  Andreas Tobler  <a.tobler@schweiz.ch>
1442
1443         * testsuite/libffi.special/unwindtest.cc (closure_test_fn): Mark cif,
1444         args and userdata unused.
1445         (closure_test_fn1): Mark cif and userdata unused.
1446         (main): Remove unused res.
1447
1448 2006-02-28  Andreas Tobler  <a.tobler@schweiz.ch>
1449
1450         * testsuite/libffi.call/call.exp: Adjust FSF address. Add test runs for
1451         -O2, -O3, -Os and the warning flags -W -Wall.
1452         * testsuite/libffi.special/special.exp: Likewise.
1453         * testsuite/libffi.call/ffitest.h: Add an __UNUSED__ macro to mark
1454         unused parameter unused for gcc or else do nothing.
1455         * testsuite/libffi.special/ffitestcxx.h: Likewise.
1456         * testsuite/libffi.call/cls_12byte.c (cls_struct_12byte_gn): Mark cif
1457         and userdata unused.
1458         * testsuite/libffi.call/cls_16byte.c (cls_struct_16byte_gn): Likewise.
1459         * testsuite/libffi.call/cls_18byte.c (cls_struct_18byte_gn): Likewise.
1460         * testsuite/libffi.call/cls_19byte.c (cls_struct_19byte_gn): Likewise.
1461         * testsuite/libffi.call/cls_1_1byte.c (cls_struct_1_1byte_gn): Likewise.
1462         * testsuite/libffi.call/cls_20byte.c (cls_struct_20byte_gn): Likewise.
1463         * testsuite/libffi.call/cls_20byte1.c (cls_struct_20byte_gn): Likewise.
1464         * testsuite/libffi.call/cls_24byte.c (cls_struct_24byte_gn): Likewise.
1465         * testsuite/libffi.call/cls_2byte.c (cls_struct_2byte_gn): Likewise.
1466         * testsuite/libffi.call/cls_3_1byte.c (cls_struct_3_1byte_gn): Likewise.
1467         * testsuite/libffi.call/cls_3byte1.c (cls_struct_3byte_gn): Likewise.
1468         * testsuite/libffi.call/cls_3byte2.c (cls_struct_3byte_gn1): Likewise.
1469         * testsuite/libffi.call/cls_4_1byte.c (cls_struct_4_1byte_gn): Likewise.
1470         * testsuite/libffi.call/cls_4byte.c (cls_struct_4byte_gn): Likewise.
1471         * testsuite/libffi.call/cls_5_1_byte.c (cls_struct_5byte_gn): Likewise.
1472         * testsuite/libffi.call/cls_5byte.c (cls_struct_5byte_gn): Likewise.
1473         * testsuite/libffi.call/cls_64byte.c (cls_struct_64byte_gn): Likewise.
1474         * testsuite/libffi.call/cls_6_1_byte.c (cls_struct_6byte_gn): Likewise.
1475         * testsuite/libffi.call/cls_6byte.c (cls_struct_6byte_gn): Likewise.
1476         * testsuite/libffi.call/cls_7_1_byte.c (cls_struct_7byte_gn): Likewise.
1477         * testsuite/libffi.call/cls_7byte.c (cls_struct_7byte_gn): Likewise.
1478         * testsuite/libffi.call/cls_8byte.c (cls_struct_8byte_gn): Likewise.
1479         * testsuite/libffi.call/cls_9byte1.c (cls_struct_9byte_gn): Likewise.
1480         * testsuite/libffi.call/cls_9byte2.c (cls_struct_9byte_gn): Likewise.
1481         * testsuite/libffi.call/cls_align_double.c (cls_struct_align_gn):
1482         Likewise.
1483         * testsuite/libffi.call/cls_align_float.c (cls_struct_align_gn):
1484         Likewise.
1485         * testsuite/libffi.call/cls_align_longdouble.c (cls_struct_align_gn):
1486         Likewise.
1487         * testsuite/libffi.call/cls_align_pointer.c (cls_struct_align_fn): Cast
1488         void* to avoid compiler warning.
1489         (main): Likewise.
1490         (cls_struct_align_gn): Mark cif and userdata unused.
1491         * testsuite/libffi.call/cls_align_sint16.c (cls_struct_align_gn):
1492         Likewise.
1493         * testsuite/libffi.call/cls_align_sint32.c (cls_struct_align_gn):
1494         Likewise.
1495         * testsuite/libffi.call/cls_align_sint64.c (cls_struct_align_gn):
1496         Likewise.
1497         * testsuite/libffi.call/cls_align_uint16.c (cls_struct_align_gn):
1498         Likewise.
1499         * testsuite/libffi.call/cls_align_uint32.c (cls_struct_align_gn):
1500         Likewise.
1501         * testsuite/libffi.call/cls_double.c (cls_ret_double_fn): Likewise.
1502         * testsuite/libffi.call/cls_float.c (cls_ret_float_fn): Likewise.
1503         * testsuite/libffi.call/cls_multi_schar.c (test_func_gn): Mark cif and
1504         data unused.
1505         (main): Cast res_call to silence gcc.
1506         * testsuite/libffi.call/cls_multi_sshort.c (test_func_gn): Mark cif and
1507         data unused.
1508         (main): Cast res_call to silence gcc.
1509         * testsuite/libffi.call/cls_multi_sshortchar.c (test_func_gn): Mark cif
1510         and data unused.
1511         (main): Cast res_call to silence gcc.
1512         * testsuite/libffi.call/cls_multi_uchar.c (test_func_gn): Mark cif and
1513         data unused.
1514         (main): Cast res_call to silence gcc.
1515         * testsuite/libffi.call/cls_multi_ushort.c (test_func_gn): Mark cif and
1516         data unused.
1517         (main): Cast res_call to silence gcc.
1518         * testsuite/libffi.call/cls_multi_ushortchar.c (test_func_gn): Mark cif
1519         and data unused.
1520         (main): Cast res_call to silence gcc.
1521         * testsuite/libffi.call/cls_schar.c (cls_ret_schar_fn): Mark cif and
1522         userdata unused.
1523         (cls_ret_schar_fn): Cast printf parameter to silence gcc.
1524         * testsuite/libffi.call/cls_sint.c (cls_ret_sint_fn): Mark cif and
1525         userdata unused.
1526         (cls_ret_sint_fn): Cast printf parameter to silence gcc.
1527         * testsuite/libffi.call/cls_sshort.c (cls_ret_sshort_fn): Mark cif and
1528         userdata unused.
1529         (cls_ret_sshort_fn): Cast printf parameter to silence gcc.
1530         * testsuite/libffi.call/cls_uchar.c (cls_ret_uchar_fn):  Mark cif and
1531         userdata unused.
1532         (cls_ret_uchar_fn): Cast printf parameter to silence gcc.
1533         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Mark cif and
1534         userdata unused.
1535         (cls_ret_uint_fn): Cast printf parameter to silence gcc.
1536         * testsuite/libffi.call/cls_ulonglong.c (cls_ret_ulonglong_fn): Mark cif
1537         and userdata unused.
1538         * testsuite/libffi.call/cls_ushort.c (cls_ret_ushort_fn): Mark cif and
1539         userdata unused.
1540         (cls_ret_ushort_fn): Cast printf parameter to silence gcc.
1541         * testsuite/libffi.call/float.c (floating): Remove unused parameter e.
1542         * testsuite/libffi.call/float1.c (main): Remove unused variable i.
1543         Cleanup white spaces.
1544         * testsuite/libffi.call/negint.c (checking): Remove unused variable i.
1545         * testsuite/libffi.call/nested_struct.c (cls_struct_combined_gn): Mark
1546         cif and userdata unused.
1547         * testsuite/libffi.call/nested_struct1.c (cls_struct_combined_gn):
1548         Likewise.
1549         * testsuite/libffi.call/nested_struct10.c (B_gn): Likewise.
1550         * testsuite/libffi.call/nested_struct2.c (B_fn): Adjust printf
1551         formatters to silence gcc.
1552         (B_gn): Mark cif and userdata unused.
1553         * testsuite/libffi.call/nested_struct3.c (B_gn): Mark cif and userdata
1554         unused.
1555         * testsuite/libffi.call/nested_struct4.c: Mention related PR.
1556         (B_gn): Mark cif and userdata unused.
1557         * testsuite/libffi.call/nested_struct5.c (B_gn): Mark cif and userdata
1558         unused.
1559         * testsuite/libffi.call/nested_struct6.c: Mention related PR.
1560         (B_gn): Mark cif and userdata unused.
1561         * testsuite/libffi.call/nested_struct7.c (B_gn): Mark cif and userdata
1562         unused.
1563         * testsuite/libffi.call/nested_struct8.c (B_gn): Likewise.
1564         * testsuite/libffi.call/nested_struct9.c (B_gn): Likewise.
1565         * testsuite/libffi.call/problem1.c (stub): Likewise.
1566         * testsuite/libffi.call/pyobjc-tc.c (main): Cast the result to silence
1567         gcc.
1568         * testsuite/libffi.call/return_fl2.c (return_fl): Add the note mentioned
1569         in the last commit for this test case in the test case itself.
1570         * testsuite/libffi.call/closure_fn0.c (closure_test_fn0): Mark cif as
1571         unused.
1572         * testsuite/libffi.call/closure_fn1.c (closure_test_fn1): Likewise.
1573         * testsuite/libffi.call/closure_fn2.c (closure_test_fn2): Likewise.
1574         * testsuite/libffi.call/closure_fn3.c (closure_test_fn3): Likewise.
1575         * testsuite/libffi.call/closure_fn4.c (closure_test_fn0): Likewise.
1576         * testsuite/libffi.call/closure_fn5.c (closure_test_fn5): Likewise.
1577         * testsuite/libffi.call/closure_fn6.c (closure_test_fn0): Likewise.
1578
1579 2006-02-22  Kaz Kojima  <kkojima@gcc.gnu.org>
1580
1581         * src/sh/sysv.S: Fix register numbers in the FDE for
1582         ffi_closure_SYSV.
1583
1584 2006-02-20  Andreas Tobler  <a.tobler@schweiz.ch>
1585
1586         * testsuite/libffi.call/return_fl2.c (return_fl): Remove static
1587         declaration to avoid a false negative on ix86. See PR323.
1588
1589 2006-02-18  Kaz Kojima  <kkojima@gcc.gnu.org>
1590
1591         * src/sh/ffi.c (ffi_closure_helper_SYSV): Remove unused variable
1592         and cast integer to void * if needed.  Update the pointer to
1593         the FP register saved area correctly.
1594         
1595 2006-02-17  Andreas Tobler  <a.tobler@schweiz.ch>
1596
1597         * testsuite/libffi.call/nested_struct6.c: XFAIL this test until PR25630
1598         is fixed.
1599         * testsuite/libffi.call/nested_struct4.c: Likewise.
1600
1601 2006-02-16  Andreas Tobler  <a.tobler@schweiz.ch>
1602
1603         * testsuite/libffi.call/return_dbl.c: New test case.
1604         * testsuite/libffi.call/return_dbl1.c: Likewise.
1605         * testsuite/libffi.call/return_dbl2.c: Likewise.
1606         * testsuite/libffi.call/return_fl.c: Likewise.
1607         * testsuite/libffi.call/return_fl1.c: Likewise.
1608         * testsuite/libffi.call/return_fl2.c: Likewise.
1609         * testsuite/libffi.call/return_fl3.c: Likewise.
1610         * testsuite/libffi.call/closure_fn6.c: Likewise.
1611
1612         * testsuite/libffi.call/nested_struct2.c: Remove ffi_type_mylong
1613         definition.
1614         * testsuite/libffi.call/ffitest.h: Add ffi_type_mylong definition
1615         here to be used by other test cases too.
1616
1617         * testsuite/libffi.call/nested_struct10.c: New test case.
1618         * testsuite/libffi.call/nested_struct9.c: Likewise.
1619         * testsuite/libffi.call/nested_struct8.c: Likewise.
1620         * testsuite/libffi.call/nested_struct7.c: Likewise.
1621         * testsuite/libffi.call/nested_struct6.c: Likewise.
1622         * testsuite/libffi.call/nested_struct5.c: Likewise.
1623         * testsuite/libffi.call/nested_struct4.c: Likewise.
1624
1625 2006-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
1626
1627         * configure.ac: Enable libffi for sparc64-*-freebsd*.
1628         * configure: Rebuilt.
1629
1630 2006-01-18  Jakub Jelinek  <jakub@redhat.com>
1631
1632         * src/powerpc/sysv.S (smst_two_register): Don't call __ashldi3,
1633         instead do the shifting inline.
1634         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't compute %r5
1635         shift count unconditionally.  Simplify load sequences for 1, 2, 3, 4
1636         and 8 byte structs, for the remaining struct sizes don't call
1637         __lshrdi3, instead do the shifting inline.
1638
1639 2005-12-07  Thiemo Seufer  <ths@networkno.de>
1640
1641         * src/mips/ffitarget.h: Remove obsolete sgidefs.h include. Add
1642         missing parentheses.
1643         * src/mips/o32.S (ffi_call_O32): Code formatting. Define
1644         and use A3_OFF, FP_OFF, RA_OFF. Micro-optimizations.
1645         (ffi_closure_O32): Likewise, but with newly defined A3_OFF2,
1646         A2_OFF2, A1_OFF2, A0_OFF2, RA_OFF2, FP_OFF2, S0_OFF2, GP_OFF2,
1647         V1_OFF2, V0_OFF2, FA_1_1_OFF2, FA_1_0_OFF2, FA_0_1_OFF2,
1648         FA_0_0_OFF2.
1649         * src/mips/ffi.c (ffi_prep_args): Code formatting. Fix
1650         endianness bugs.
1651         (ffi_prep_closure): Improve trampoline instruction scheduling.
1652         (ffi_closure_mips_inner_O32): Fix endianness bugs.
1653
1654 2005-12-03  Alan Modra  <amodra@bigpond.net.au>
1655
1656         * src/powerpc/ffi.c: Formatting.
1657         (ffi_prep_args_SYSV): Avoid possible aliasing problems by using unions.
1658         (ffi_prep_args64): Likewise.
1659
1660 2005-09-30  Geoffrey Keating  <geoffk@apple.com>
1661
1662         * testsuite/lib/libffi-dg.exp (libffi_target_compile): For
1663         darwin, use -shared-libgcc not -lgcc_s, and explain why.
1664
1665 2005-09-26  Tom Tromey  <tromey@redhat.com>
1666
1667         * testsuite/libffi.call/float1.c (value_type): New typedef.
1668         (CANARY): New define.
1669         (main): Check for result buffer overflow.
1670         * src/powerpc/linux64.S: Handle linux64 long double returns.
1671         * src/powerpc/ffi.c (FLAG_RETURNS_128BITS): New constant.
1672         (ffi_prep_cif_machdep): Handle linux64 long double returns.
1673
1674 2005-08-25  Alan Modra  <amodra@bigpond.net.au>
1675
1676         PR target/23404
1677         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Correct placement of stack
1678         homed fp args.
1679         (ffi_status ffi_prep_cif_machdep): Correct stack sizing for same.
1680
1681 2005-08-11  Jakub Jelinek  <jakub@redhat.com>
1682
1683         * configure.ac (HAVE_HIDDEN_VISIBILITY_ATTRIBUTE): New test.
1684         (AH_BOTTOM): Add FFI_HIDDEN definition.
1685         * configure: Rebuilt.
1686         * fficonfig.h.in: Rebuilt.
1687         * src/powerpc/ffi.c (hidden): Remove.
1688         (ffi_closure_LINUX64, ffi_prep_args64, ffi_call_LINUX64,
1689         ffi_closure_helper_LINUX64): Use FFI_HIDDEN instead of hidden.
1690         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64,
1691         .ffi_closure_LINUX64): Use FFI_HIDDEN instead of .hidden.
1692         * src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV): Remove,
1693         add FFI_HIDDEN to its prototype.
1694         (ffi_closure_SYSV_inner): New.
1695         * src/x86/sysv.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
1696         * src/x86/win32.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
1697
1698 2005-08-10  Alfred M. Szmidt  <ams@gnu.org>
1699
1700         PR libffi/21819:
1701         * configure: Rebuilt.
1702         * configure.ac: Handle i*86-*-gnu*.
1703
1704 2005-08-09  Jakub Jelinek  <jakub@redhat.com>
1705
1706         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Use
1707         DW_CFA_offset_extended_sf rather than
1708         DW_CFA_GNU_negative_offset_extended.
1709         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise.
1710
1711 2005-07-22  SUGIOKA Toshinobu  <sugioka@itonet.co.jp>
1712
1713         * src/sh/sysv.S (ffi_call_SYSV): Stop argument popping correctly
1714         on sh3.
1715         (ffi_closure_SYSV): Change the stack layout for sh3 struct argument.
1716         * src/sh/ffi.c (ffi_prep_args): Fix sh3 argument copy, when it is
1717         partially on register.
1718         (ffi_closure_helper_SYSV): Likewise.
1719         (ffi_prep_cif_machdep): Don't set too many cif->flags.
1720
1721 2005-07-20  Kaz Kojima  <kkojima@gcc.gnu.org>
1722
1723         * src/sh/ffi.c (ffi_call): Handle small structures correctly.
1724         Remove empty line.
1725         * src/sh64/ffi.c (simple_type): Remove.
1726         (return_type): Handle small structures correctly.
1727         (ffi_prep_args): Likewise.
1728         (ffi_call): Likewise.
1729         (ffi_closure_helper_SYSV): Likewise.
1730         * src/sh64/sysv.S (ffi_call_SYSV): Handle 1, 2 and 4-byte return.
1731         Emit position independent code if PIC and remove wrong datalabel
1732         prefixes from EH data.
1733
1734 2005-07-19  Andreas Tobler  <a.tobler@schweiz.ch>
1735
1736         * Makefile.am (nodist_libffi_la_SOURCES): Add POWERPC_FREEBSD.
1737         * Makefile.in: Regenerate.
1738         * include/Makefile.in: Likewise.
1739         * testsuite/Makefile.in: Likewise.
1740         * configure.ac: Add POWERPC_FREEBSD rules.
1741         * configure: Regenerate.
1742         * src/powerpc/ffitarget.h: Add POWERPC_FREEBSD rules.
1743         (FFI_SYSV_TYPE_SMALL_STRUCT): Define.
1744         * src/powerpc/ffi.c: Add flags to handle small structure returns
1745         in ffi_call_SYSV.
1746         (ffi_prep_cif_machdep): Handle small structures for SYSV 4 ABI.
1747         Aka FFI_SYSV.
1748         (ffi_closure_helper_SYSV): Likewise.
1749         * src/powerpc/ppc_closure.S: Add return types for small structures.
1750         * src/powerpc/sysv.S: Add bits to handle small structures for
1751         final SYSV 4 ABI.
1752
1753 2005-07-10  Andreas Tobler  <a.tobler@schweiz.ch>
1754
1755         * testsuite/libffi.call/cls_5_1_byte.c: New test file.
1756         * testsuite/libffi.call/cls_6_1_byte.c: Likewise.
1757         * testsuite/libffi.call/cls_7_1_byte.c: Likewise.
1758
1759 2005-07-05  Randolph Chung  <tausq@debian.org>
1760
1761         * src/pa/ffi.c (ffi_struct_type): Rename FFI_TYPE_SMALL_STRUCT1
1762         as FFI_TYPE_SMALL_STRUCT3.  Break out handling for 5-7 byte
1763         structures.  Kill compilation warnings.
1764         (ffi_closure_inner_LINUX): Print return values as hex in debug
1765         message.  Rename FFI_TYPE_SMALL_STRUCT1 as FFI_TYPE_SMALL_STRUCT3.
1766         Properly handle 5-7 byte structure returns.
1767         * src/pa/ffitarget.h (FFI_TYPE_SMALL_STRUCT1)
1768         (FFI_TYPE_SMALL_STRUCT2): Remove.
1769         (FFI_TYPE_SMALL_STRUCT3, FFI_TYPE_SMALL_STRUCT5)
1770         (FFI_TYPE_SMALL_STRUCT6, FFI_TYPE_SMALL_STRUCT7): Define.
1771         * src/pa/linux.S: Mark source file as using PA1.1 assembly.
1772         (checksmst1, checksmst2): Remove.
1773         (checksmst3): Optimize handling of 3-byte struct returns.
1774         (checksmst567): Properly handle 5-7 byte struct returns.
1775
1776 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1777
1778         PR libgcj/21943
1779         * src/mips/n32.S: Enforce PIC code.
1780         * src/mips/o32.S: Likewise.
1781
1782 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1783
1784         * configure.ac: Treat i*86-*-solaris2.10 and up as X86_64.
1785         * configure: Regenerate.
1786         
1787 2005-06-01  Alan Modra  <amodra@bigpond.net.au>
1788
1789         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't use JUMPTARGET
1790         to call ffi_closure_helper_SYSV.  Append @local instead.
1791         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise for ffi_prep_args_SYSV.
1792
1793 2005-05-17  Kelley Cook  <kcook@gcc.gnu.org>
1794
1795         * configure.ac: Use AC_C_BIGENDIAN instead of AC_C_BIGENDIAN_CROSS.
1796         Use AC_CHECK_SIZEOF instead of AC_COMPILE_CHECK_SIZEOF.
1797         * Makefile.am (ACLOCAL_AMFLAGS): Remove -I ../config.
1798         * aclocal.m4, configure, fficonfig.h.in, Makefile.in,
1799         include/Makefile.in, testsuite/Makefile.in: Regenerate.
1800         
1801 2005-05-09  Mike Stump  <mrs@apple.com>
1802
1803         * configure: Regenerate.
1804
1805 2005-05-08  Richard Henderson  <rth@redhat.com>
1806
1807         PR libffi/21285
1808         * src/alpha/osf.S: Update unwind into to match code.
1809
1810 2005-05-04  Andreas Degert <ad@papyrus-gmbh.de>
1811             Richard Henderson  <rth@redhat.com>
1812
1813         * src/x86/ffi64.c (ffi_prep_cif_machdep): Save sse-used flag in
1814         bit 11 of flags.
1815         (ffi_call): Mask return type field.  Pass ssecount to ffi_call_unix64.
1816         (ffi_prep_closure): Set carry bit if sse-used flag set.
1817         * src/x86/unix64.S (ffi_call_unix64): Add ssecount argument.
1818         Only load sse registers if ssecount non-zero.
1819         (ffi_closure_unix64): Only save sse registers if carry set on entry.
1820
1821 2005-04-29  Ralf Corsepius  <ralf.corsepius@rtems.org>
1822
1823         * configure.ac: Add i*86-*-rtems*, sparc*-*-rtems*,
1824         powerpc-*rtems*, arm*-*-rtems*, sh-*-rtems*.
1825         * configure: Regenerate.
1826
1827 2005-04-20  Hans-Peter Nilsson  <hp@axis.com>
1828
1829         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): In regsub use,
1830         have Tcl8.3-compatible intermediate variable.
1831
1832 2005-04-18  Simon Posnjak <simon.posnjak@siol.net> 
1833             Hans-Peter Nilsson  <hp@axis.com>
1834
1835         * Makefile.am: Add CRIS support.
1836         * configure.ac: Likewise.
1837         * Makefile.in, configure, testsuite/Makefile.in,
1838         include/Makefile.in: Regenerate.
1839         * src/cris: New directory.
1840         * src/cris/ffi.c, src/cris/sysv.S, src/cris/ffitarget.h: New files.
1841         * src/prep_cif.c (ffi_prep_cif): Wrap in #ifndef __CRIS__.
1842
1843         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): Replace \n with
1844         \r?\n in output tests.
1845
1846 2005-04-12  Mike Stump  <mrs@apple.com>
1847
1848         * configure: Regenerate.
1849
1850 2005-03-30  Hans Boehm  <Hans.Boehm@hp.com>
1851
1852         * src/ia64/ffitarget.h (ffi_arg): Use long long instead of DI.
1853         
1854 2005-03-30  Steve Ellcey  <sje@cup.hp.com>
1855
1856         * src/ia64/ffitarget.h (ffi_arg) ADD DI attribute.
1857         (ffi_sarg) Ditto.
1858         * src/ia64/unix.S (ffi_closure_unix): Extend gp
1859         to 64 bits in ILP32 mode.
1860         Load 64 bits even for short data.
1861
1862 2005-03-23  Mike Stump  <mrs@apple.com>
1863
1864         * src/powerpc/darwin.S: Update for -m64 multilib.
1865         * src/powerpc/darwin_closure.S: Likewise.
1866
1867 2005-03-21  Zack Weinberg  <zack@codesourcery.com>
1868
1869         * configure.ac: Do not invoke TL_AC_GCC_VERSION.
1870         Do not set tool_include_dir.
1871         * aclocal.m4, configure, Makefile.in, testsuite/Makefile.in:
1872         Regenerate.
1873         * include/Makefile.am: Set gcc_version and toollibffidir.
1874         * include/Makefile.in: Regenerate.
1875
1876 2005-02-22  Andrew Haley  <aph@redhat.com>
1877
1878         * src/powerpc/ffi.c (ffi_prep_cif_machdep): Bump alignment to
1879         odd-numbered register pairs for 64-bit integer types.
1880
1881 2005-02-23  Andreas Tobler  <a.tobler@schweiz.ch>
1882
1883         PR libffi/20104
1884         * testsuite/libffi.call/return_ll1.c: New test case.
1885
1886 2005-02-11  Janis Johnson  <janis187@us.ibm.com>
1887
1888         * testsuite/libffi.call/cls_align_longdouble.c: Remove dg-options.
1889         * testsuite/libffi.call/float.c: Ditto.
1890         * testsuite/libffi.call/float2.c: Ditto.
1891         * testsuite/libffi.call/float3.c: Ditto.
1892
1893 2005-02-08  Andreas Tobler  <a.tobler@schweiz.ch>
1894
1895         * src/frv/ffitarget.h: Remove PPC stuff which does not belong to frv.
1896
1897 2005-01-12  Eric Botcazou  <ebotcazou@libertysurf.fr>
1898
1899         * testsuite/libffi.special/special.exp (cxx_options): Add
1900         -shared-libgcc.
1901
1902 2004-12-31  Richard Henderson  <rth@redhat.com>
1903
1904         * src/types.c (FFI_AGGREGATE_TYPEDEF): Remove.
1905         (FFI_TYPEDEF): Rename from FFI_INTEGRAL_TYPEDEF.  Replace size and
1906         offset parameters with a type parameter; deduce size and structure
1907         alignment.  Update all users.
1908
1909 2004-12-31  Richard Henderson  <rth@redhat.com>
1910
1911         * src/types.c (FFI_TYPE_POINTER): Define with sizeof.
1912         (FFI_TYPE_LONGDOUBLE): Fix for ia64.
1913         * src/ia64/ffitarget.h (struct ffi_ia64_trampoline_struct): Move
1914         into ffi_prep_closure.
1915         * src/ia64/ia64_flags.h, src/ia64/ffi.c, src/ia64/unix.S: Rewrite
1916         from scratch.
1917
1918 2004-12-27  Richard Henderson  <rth@redhat.com>
1919
1920         * src/x86/unix64.S: Fix typo in unwind info.
1921
1922 2004-12-25  Richard Henderson  <rth@redhat.com>
1923
1924         * src/x86/ffi64.c (struct register_args): Rename from stackLayout.
1925         (enum x86_64_reg_class): Add X86_64_COMPLEX_X87_CLASS.
1926         (merge_classes): Check for it.
1927         (SSE_CLASS_P): New.
1928         (classify_argument): Pass byte_offset by value; perform all updates
1929         inside struct case.
1930         (examine_argument): Add classes argument; handle
1931         X86_64_COMPLEX_X87_CLASS.
1932         (ffi_prep_args): Merge into ...
1933         (ffi_call): ... here.  Share stack frame with ffi_call_unix64.
1934         (ffi_prep_cif_machdep): Setup cif->flags for proper structure return.
1935         (ffi_fill_return_value): Remove.
1936         (ffi_prep_closure): Remove dead assert.
1937         (ffi_closure_unix64_inner): Rename from ffi_closure_UNIX64_inner.
1938         Rewrite to use struct register_args instead of va_list.  Create
1939         flags for handling structure returns.
1940         * src/x86/unix64.S: Remove dead strings.
1941         (ffi_call_unix64): Rename from ffi_call_UNIX64.  Rewrite to share
1942         stack frame with ffi_call.  Handle structure returns properly.
1943         (float2sse, floatfloat2sse, double2sse): Remove.
1944         (sse2float, sse2double, sse2floatfloat): Remove.
1945         (ffi_closure_unix64): Rename from ffi_closure_UNIX64.  Rewrite
1946         to handle structure returns properly.
1947
1948 2004-12-08  David Edelsohn  <edelsohn@gnu.org>
1949
1950         * Makefile.am (AM_MAKEFLAGS): Remove duplicate LIBCFLAGS and
1951         PICFLAG.
1952         * Makefile.in: Regenerated.
1953
1954 2004-12-02  Richard Sandiford  <rsandifo@redhat.com>
1955
1956         * configure.ac: Use TL_AC_GCC_VERSION to set gcc_version.
1957         * configure, aclocal.m4, Makefile.in: Regenerate.
1958         * include/Makefile.in, testsuite/Makefile.in: Regenerate.
1959
1960 2004-11-29  Kelley Cook  <kcook@gcc.gnu.org>
1961
1962         * configure: Regenerate for libtool change.
1963
1964 2004-11-25  Kelley Cook  <kcook@gcc.gnu.org>
1965
1966         * configure: Regenerate for libtool reversion.
1967
1968 2004-11-24  Kelley Cook  <kcook@gcc.gnu.org>
1969
1970         * configure: Regenerate for libtool change.
1971
1972 2004-11-23  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1973
1974         * testsuite/lib/libffi-dg.exp: Use new procs in target-libpath.exp.
1975
1976 2004-11-23  Richard Sandiford  <rsandifo@redhat.com>
1977
1978         * src/mips/o32.S (ffi_call_O32, ffi_closure_O32): Use jalr instead
1979         of jal.  Use an absolute encoding for the frame information.
1980
1981 2004-11-23  Kelley Cook  <kcook@gcc.gnu.org>
1982
1983         * Makefile.am: Remove no-dependencies.  Add ACLOCAL_AMFLAGS.
1984         * acinclude.m4: Delete logic for sincludes.
1985         * aclocal.m4, Makefile.in, configure: Regenerate.
1986         * include/Makefile: Likewise.
1987         * testsuite/Makefile: Likewise.
1988
1989 2004-11-22  Eric Botcazou  <ebotcazou@libertysurf.fr>
1990
1991         * src/sparc/ffi.c (ffi_prep_closure): Align doubles and 64-bit integers
1992         on a 8-byte boundary.
1993         * src/sparc/v8.S (ffi_closure_v8): Reserve frame space for arguments.
1994
1995 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
1996
1997         * src/arm/ffi.c (ffi_prep_cif_machdep): Handle functions that return
1998         long long values.  Round stack allocation to a multiple of 8 bytes
1999         for ATPCS compatibility.
2000         * src/arm/sysv.S (ffi_call_SYSV): Rework to avoid use of APCS register
2001         names.  Handle returning long long types.  Add Thumb and interworking
2002         support.  Improve soft-float code.
2003
2004 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2005
2006         * testsuite/lib/libffi-db.exp (load_gcc_lib): New function.
2007         (libffi_exit): New function.
2008         (libffi_init): Build the testglue wrapper if needed.
2009
2010 2004-10-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2011
2012         PR other/18138
2013         * testsuite/lib/libffi-dg.exp: Accept more than one multilib libgcc.
2014
2015 2004-10-25  Kazuhiro Inaoka <inaoka.kazuhiro@renesas.com>
2016
2017         * src/m32r/libffitarget.h (FFI_CLOSURES): Set to 0.
2018
2019 2004-10-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2020
2021         * src/sh/sysv.S (ffi_call_SYSV): Don't align for double data.
2022         * testsuite/libffi.call/float3.c: New test case.
2023
2024 2004-10-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2025
2026         * src/sh/ffi.c (ffi_prep_closure): Set T bit in trampoline for
2027         the function returning a structure pointed with R2.
2028         * src/sh/sysv.S (ffi_closure_SYSV): Use R2 as the pointer to
2029         the structure return value if T bit set.  Emit position
2030         independent code and EH data if PIC.
2031
2032 2004-10-13  Kazuhiro Inaoka  <inaoka.kazuhiro@renesas.com>
2033
2034         * Makefile.am: Add m32r support.
2035         * configure.ac: Likewise.
2036         * Makefile.in: Regenerate.
2037         * confiugre: Regenerate.
2038         * src/types.c: Add m32r port to FFI_INTERNAL_TYPEDEF
2039         (uint64, sint64, double, longdouble)
2040         * src/m32r: New directory.
2041         * src/m32r/ffi.c: New file.
2042         * src/m32r/sysv.S: Likewise.
2043         * src/m32r/ffitarget.h: Likewise.
2044
2045 2004-10-02  Kaz Kojima  <kkojima@gcc.gnu.org>
2046
2047         * testsuite/libffi.call/negint.c: New test case.
2048
2049 2004-09-14  H.J. Lu  <hongjiu.lu@intel.com>
2050
2051         PR libgcj/17465
2052         * testsuite/lib/libffi-dg.exp: Don't use global ld_library_path.
2053         Set up LD_LIBRARY_PATH, SHLIB_PATH, LD_LIBRARYN32_PATH,
2054         LD_LIBRARY64_PATH, LD_LIBRARY_PATH_32, LD_LIBRARY_PATH_64 and
2055         DYLD_LIBRARY_PATH.
2056
2057 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2058
2059         * testsuite/libffi.call/many_win32.c: Remove whitespaces.
2060         * testsuite/libffi.call/promotion.c: Likewise.
2061         * testsuite/libffi.call/return_ll.c: Remove unused var. Cleanup
2062         whitespaces.
2063         * testsuite/libffi.call/return_sc.c: Likewise.
2064         * testsuite/libffi.call/return_uc.c: Likewise.
2065
2066 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2067
2068         * src/powerpc/darwin.S: Fix comments and identation.
2069         * src/powerpc/darwin_closure.S: Likewise.
2070
2071 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2072
2073         * src/powerpc/ffi_darwin.c: Add flag for longdouble return values.
2074         (ffi_prep_args): Handle longdouble arguments.
2075         (ffi_prep_cif_machdep): Set flags for longdouble. Calculate space for
2076         longdouble.
2077         (ffi_closure_helper_DARWIN): Add closure handling for longdouble.
2078         * src/powerpc/darwin.S (_ffi_call_DARWIN): Add handling of longdouble
2079         values.
2080         * src/powerpc/darwin_closure.S (_ffi_closure_ASM): Likewise.
2081         * src/types.c: Defined longdouble size and alignment for darwin.
2082
2083 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2084
2085         * src/powerpc/aix.S: Remove whitespaces.
2086         * src/powerpc/aix_closure.S: Likewise.
2087         * src/powerpc/asm.h: Likewise.
2088         * src/powerpc/ffi.c: Likewise.
2089         * src/powerpc/ffitarget.h: Likewise.
2090         * src/powerpc/linux64.S: Likewise.
2091         * src/powerpc/linux64_closure.S: Likewise.
2092         * src/powerpc/ppc_closure.S: Likewise.
2093         * src/powerpc/sysv.S: Likewise.
2094
2095 2004-08-30  Anthony Green  <green@redhat.com>
2096
2097         * Makefile.am: Add frv support.
2098         * Makefile.in, testsuite/Makefile.in: Rebuilt.
2099         * configure.ac: Read configure.host.
2100         * configure.in: Read configure.host.
2101         * configure.host: New file.  frv-elf needs libgloss.
2102         * include/ffi.h.in: Force ffi_closure to have a nice big (8)
2103         alignment.  This is needed to frv and shouldn't harm the others.
2104         * include/ffi_common.h (ALIGN_DOWN): New macro.
2105         * src/frv/ffi.c, src/frv/ffitarget.h, src/frv/eabi.S: New files.
2106
2107 2004-08-24  David Daney  <daney@avtrex.com>
2108
2109         * testsuite/libffi.call/closure_fn0.c: Xfail mips64* instead of mips*.
2110         * testsuite/libffi.call/closure_fn1.c: Likewise.
2111         * testsuite/libffi.call/closure_fn2.c  Likewise.
2112         * testsuite/libffi.call/closure_fn3.c: Likewise.
2113         * testsuite/libffi.call/closure_fn4.c: Likewise.
2114         * testsuite/libffi.call/closure_fn5.c: Likewise.
2115         * testsuite/libffi.call/cls_18byte.c: Likewise.
2116         * testsuite/libffi.call/cls_19byte.c: Likewise.
2117         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2118         * testsuite/libffi.call/cls_20byte.c: Likewise.
2119         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2120         * testsuite/libffi.call/cls_24byte.c: Likewise.
2121         * testsuite/libffi.call/cls_2byte.c: Likewise.
2122         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2123         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2124         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2125         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2126         * testsuite/libffi.call/cls_4byte.c: Likewise.
2127         * testsuite/libffi.call/cls_64byte.c: Likewise.
2128         * testsuite/libffi.call/cls_6byte.c: Likewise.
2129         * testsuite/libffi.call/cls_7byte.c: Likewise.
2130         * testsuite/libffi.call/cls_8byte.c: Likewise.
2131         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2132         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2133         * testsuite/libffi.call/cls_align_double.c: Likewise.
2134         * testsuite/libffi.call/cls_align_float.c: Likewise.
2135         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2136         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2137         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2138         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2139         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2140         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2141         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2142         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
2143         * testsuite/libffi.call/cls_double.c: Likewise.
2144         * testsuite/libffi.call/cls_float.c: Likewise.
2145         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2146         * testsuite/libffi.call/cls_multi_sshort.c: Likewise.
2147         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2148         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2149         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2150         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2151         * testsuite/libffi.call/cls_schar.c: Likewise.
2152         * testsuite/libffi.call/cls_sint.c: Likewise.
2153         * testsuite/libffi.call/cls_sshort.c: Likewise.
2154         * testsuite/libffi.call/cls_uchar.c: Likewise.
2155         * testsuite/libffi.call/cls_uint.c: Likewise.
2156         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2157         * testsuite/libffi.call/cls_ushort.c: Likewise.
2158         * testsuite/libffi.call/nested_struct.c: Likewise.
2159         * testsuite/libffi.call/nested_struct1.c: Likewise.
2160         * testsuite/libffi.call/nested_struct2.c: Likewise.
2161         * testsuite/libffi.call/nested_struct3.c: Likewise.
2162         * testsuite/libffi.call/problem1.c: Likewise.
2163         * testsuite/libffi.special/unwindtest.cc: Likewise.
2164         * testsuite/libffi.call/cls_12byte.c: Likewise and set return value
2165         to zero.
2166         * testsuite/libffi.call/cls_16byte.c: Likewise.
2167         * testsuite/libffi.call/cls_5byte.c: Likewise.
2168
2169 2004-08-23  David Daney <daney@avtrex.com>
2170
2171         PR libgcj/13141
2172         * src/mips/ffitarget.h (FFI_O32_SOFT_FLOAT): New ABI.
2173         * src/mips/ffi.c (ffi_prep_args): Fix alignment calculation.
2174         (ffi_prep_cif_machdep): Handle FFI_O32_SOFT_FLOAT floating point
2175         parameters and return types.
2176         (ffi_call): Handle FFI_O32_SOFT_FLOAT ABI.
2177         (ffi_prep_closure): Ditto.
2178         (ffi_closure_mips_inner_O32): Handle FFI_O32_SOFT_FLOAT ABI, fix
2179         alignment calculations.
2180         * src/mips/o32.S (ffi_closure_O32): Don't use floating point
2181         instructions if FFI_O32_SOFT_FLOAT, make stack frame ABI compliant.
2182
2183 2004-08-14  Casey Marshall <csm@gnu.org>
2184
2185         * src/mips/ffi.c (ffi_pref_cif_machdep): set `cif->flags' to
2186         contain `FFI_TYPE_UINT64' as return type for any 64-bit
2187         integer (O32 ABI only).
2188         (ffi_prep_closure): new function.
2189         (ffi_closure_mips_inner_O32): new function.
2190         * src/mips/ffitarget.h: Define `FFI_CLOSURES' and
2191         `FFI_TRAMPOLINE_SIZE' appropriately if the ABI is o32.
2192         * src/mips/o32.S (ffi_call_O32): add labels for .eh_frame. Return
2193         64 bit integers correctly.
2194         (ffi_closure_O32): new function.
2195         Added DWARF-2 unwind info for both functions.
2196
2197 2004-08-10  Andrew Haley  <aph@redhat.com>
2198
2199         * src/x86/ffi64.c (ffi_prep_args ): 8-align all stack arguments.
2200
2201 2004-08-01  Robert Millan  <robertmh@gnu.org>
2202
2203         * configure.ac: Detect knetbsd-gnu and kfreebsd-gnu.
2204         * configure: Regenerate.
2205
2206 2004-07-30  Maciej W. Rozycki  <macro@linux-mips.org>
2207
2208         * acinclude.m4 (AC_FUNC_MMAP_BLACKLIST): Check for <sys/mman.h>
2209         and mmap() explicitly instead of relying on preset autoconf cache
2210         variables.
2211         * aclocal.m4: Regenerate.
2212         * configure: Regenerate.
2213
2214 2004-07-11  Ulrich Weigand  <uweigand@de.ibm.com>
2215
2216         * src/s390/ffi.c (ffi_prep_args): Fix C aliasing violation.
2217         (ffi_check_float_struct): Remove unused prototype.
2218
2219 2004-06-30  Geoffrey Keating  <geoffk@apple.com>
2220
2221         * src/powerpc/ffi_darwin.c (flush_icache): ';' is a comment
2222         character on Darwin, use '\n\t' instead.
2223
2224 2004-06-26  Matthias Klose  <doko@debian.org>
2225
2226         * libtool-version: Fix typo in revision/age.
2227
2228 2004-06-17  Matthias Klose  <doko@debian.org>
2229
2230         * libtool-version: New.
2231         * Makefile.am (libffi_la_LDFLAGS): Use -version-info for soname.
2232         * Makefile.in: Regenerate.
2233
2234 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2235
2236         * Makefile.am: Remove useless multilib rules.
2237         * Makefile.in: Regenerate.
2238         * aclocal.m4: Regenerate with automake 1.8.5.
2239         * configure.ac: Remove useless multilib configury.
2240         * configure: Regenerate.
2241
2242 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2243
2244         * .cvsignore: New file.
2245
2246 2004-06-10  Jakub Jelinek  <jakub@redhat.com>
2247
2248         * src/ia64/unix.S (ffi_call_unix): Insert group barrier break
2249         fp_done.
2250         (ffi_closure_UNIX): Fix f14/f15 adjustment if FLOAT_SZ is ever
2251         changed from 8.
2252
2253 2004-06-06  Sean McNeil  <sean@mcneil.com>
2254
2255         * configure.ac: Add x86_64-*-freebsd* support.
2256         * configure: Regenerate.
2257
2258 2004-04-26  Joe Buck <jbuck@welsh-buck.org>
2259
2260         Bug 15093
2261         * configure.ac: Test for existence of mmap and sys/mman.h before
2262         checking blacklist.  Fix suggested by Jim Wilson.
2263         * configure: Regenerate.
2264
2265 2004-04-26  Matt Austern  <austern@apple.com>
2266
2267         * src/powerpc/darwin.S: Go through a non-lazy pointer for initial
2268         FDE location.
2269         * src/powerpc/darwin_closure.S: Likewise.
2270
2271 2004-04-24  Andreas Tobler  <a.tobler@schweiz.ch>
2272
2273         * testsuite/libffi.call/cls_multi_schar.c (main): Fix initialization
2274         error. Reported by Thomas Heller <theller@python.net>.
2275         * testsuite/libffi.call/cls_multi_sshort.c (main): Likewise.
2276         * testsuite/libffi.call/cls_multi_ushort.c (main): Likewise.
2277
2278 2004-03-20  Matthias Klose  <doko@debian.org>
2279
2280         * src/pa/linux.S: Fix typo.
2281
2282 2004-03-19  Matthias Klose  <doko@debian.org>
2283
2284         * Makefile.am: Update.
2285         * Makefile.in: Regenerate.
2286         * src/pa/ffi.h.in: Remove.
2287         * src/pa/ffitarget.h: New file.
2288
2289 2004-02-10  Randolph Chung  <tausq@debian.org>
2290
2291         * Makefile.am: Add PA support.
2292         * Makefile.in: Regenerate.
2293         * include/Makefile.in: Regenerate.
2294         * configure.ac: Add PA target.
2295         * configure: Regenerate.
2296         * src/pa/ffi.c: New file.
2297         * src/pa/ffi.h.in: Add PA support.
2298         * src/pa/linux.S: New file.
2299         * prep_cif.c: Add PA support.
2300
2301 2004-03-16  Hosaka Yuji  <hos@tamanegi.org>
2302
2303         * src/types.c: Fix alignment size of X86_WIN32 case int64 and
2304         double.
2305         * src/x86/ffi.c (ffi_prep_args): Replace ecif->cif->rtype->type
2306         with ecif->cif->flags.
2307         (ffi_call, ffi_prep_incoming_args_SYSV): Replace cif->rtype->type
2308         with cif->flags.
2309         (ffi_prep_cif_machdep): Add X86_WIN32 struct case.
2310         (ffi_closure_SYSV): Add 1 or 2-bytes struct case for X86_WIN32.
2311         * src/x86/win32.S (retstruct1b, retstruct2b, sc_retstruct1b,
2312         sc_retstruct2b): Add for 1 or 2-bytes struct case.
2313
2314 2004-03-15 Kelley Cook <kcook@gcc.gnu.org>
2315
2316         * configure.in: Rename file to ...
2317         * configure.ac: ... this.
2318         * fficonfig.h.in: Regenerate.
2319         * Makefile.in: Regenerate.
2320         * include/Makefile.in: Regenerate.
2321         * testsuite/Makefile.in: Regenerate.
2322
2323 2004-03-12  Matt Austern  <austern@apple.com>
2324
2325         * src/powerpc/darwin.S: Fix EH information so it corresponds to
2326         changes in EH format resulting from addition of linkonce support.
2327         * src/powerpc/darwin_closure.S: Likewise.
2328
2329 2004-03-11  Andreas Tobler  <a.tobler@schweiz.ch>
2330             Paolo Bonzini  <bonzini@gnu.org>
2331
2332         * Makefile.am (AUTOMAKE_OPTIONS): Set them.
2333         Remove VPATH. Remove rules for object files. Remove multilib support.
2334         (AM_CCASFLAGS): Add.
2335         * configure.in (AC_CONFIG_HEADERS): Relace AM_CONFIG_HEADER.
2336         (AC_PREREQ): Bump version to 2.59.
2337         (AC_INIT): Fill with version info and bug address.
2338         (ORIGINAL_LD_FOR_MULTILIBS): Remove.
2339         (AM_ENABLE_MULTILIB): Use this instead of AC_ARG_ENABLE.
2340         De-precious CC so that the right flags are passed down to multilibs.
2341         (AC_MSG_ERROR): Replace obsolete macro AC_ERROR.
2342         (AC_CONFIG_FILES): Replace obsolete macro AC_LINK_FILES.
2343         (AC_OUTPUT): Reorganize the output with AC_CONFIG_COMMANDS.
2344         * configure: Rebuilt.
2345         * aclocal.m4: Likewise.
2346         * Makefile.in, include/Makefile.in, testsuite/Makefile.in: Likewise.
2347         * fficonfig.h.in: Likewise.
2348
2349 2004-03-11  Andreas Schwab  <schwab@suse.de>
2350
2351         * src/ia64/ffi.c (ffi_prep_incoming_args_UNIX): Get floating point
2352         arguments from fp registers only for the first 8 parameter slots.
2353         Don't convert a float parameter when passed in memory.
2354
2355 2004-03-09  Hans-Peter Nilsson  <hp@axis.com>
2356
2357         * configure: Regenerate for config/accross.m4 correction.
2358
2359 2004-02-25  Matt Kraai  <kraai@alumni.cmu.edu>
2360
2361         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Change
2362         ecif->cif->bytes to bytes.
2363         (ffi_prep_cif_machdep): Add braces around nested if statement.
2364
2365 2004-02-09  Alan Modra  <amodra@bigpond.net.au>
2366
2367         * src/types.c (pointer): POWERPC64 has 8 byte pointers.
2368
2369         * src/powerpc/ffi.c (ffi_prep_args64): Correct long double handling.
2370         (ffi_closure_helper_LINUX64): Fix typo.
2371         * testsuite/libffi.call/cls_align_longdouble.c: Pass -mlong-double-128
2372         for powerpc64-*-*.
2373         * testsuite/libffi.call/float.c: Likewise.
2374         * testsuite/libffi.call/float2.c: Likewise.
2375
2376 2004-02-08  Alan Modra  <amodra@bigpond.net.au>
2377
2378         * src/powerpc/ffi.c (ffi_prep_cif_machdep <FFI_LINUX64>): Correct
2379         long double function return and long double arg handling.
2380         (ffi_closure_helper_LINUX64): Formatting.  Delete unused "ng" var.
2381         Use "end_pfr" instead of "nf".  Correct long double handling.
2382         Localise "temp".
2383         * src/powerpc/linux64.S (ffi_call_LINUX64): Save f2 long double
2384         return value.
2385         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Allocate
2386         space for long double return value.  Adjust stack frame and offsets.
2387         Load f2 long double return.
2388
2389 2004-02-07  Alan Modra  <amodra@bigpond.net.au>
2390
2391         * src/types.c: Use 16 byte long double for POWERPC64.
2392
2393 2004-01-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2394
2395         * src/sparc/ffi.c (ffi_prep_args_v9): Shift the parameter array
2396         when the structure return address is passed in %o0.
2397         (ffi_V9_return_struct): Rename into ffi_v9_layout_struct.
2398         (ffi_v9_layout_struct): Align the field following a nested structure
2399         on a word boundary.  Use memmove instead of memcpy.
2400         (ffi_call): Update call to ffi_V9_return_struct.
2401         (ffi_prep_closure): Define 'ctx' only for V8.
2402         (ffi_closure_sparc_inner): Clone into ffi_closure_sparc_inner_v8
2403         and ffi_closure_sparc_inner_v9.
2404         (ffi_closure_sparc_inner_v8): Return long doubles by reference.
2405         Always skip the structure return address.  For structures and long
2406         doubles, copy the argument directly.
2407         (ffi_closure_sparc_inner_v9): Skip the structure return address only
2408         if required.  Shift the maximum floating-point slot accordingly.  For
2409         big structures, copy the argument directly; otherwise, left-justify the
2410         argument and call ffi_v9_layout_struct to lay out the structure on
2411         the stack.
2412         * src/sparc/v8.S: Undef STACKFRAME before defining it.
2413         (ffi_closure_v8): Pass the structure return address.  Update call to
2414         ffi_closure_sparc_inner_v8.  Short-circuit FFI_TYPE_INT handling.
2415         Skip the 'unimp' insn when returning long doubles and structures.
2416         * src/sparc/v9.S: Undef STACKFRAME before defining it.
2417         (ffi_closure_v9): Increase the frame size by 2 words.  Short-circuit
2418         FFI_TYPE_INT handling.  Load structures both in integers and
2419         floating-point registers on return.
2420         * README: Update status of the SPARC port.
2421
2422 2004-01-24  Andreas Tobler  <a.tobler@schweiz.ch>
2423
2424         * testsuite/libffi.call/pyobjc-tc.c (main): Treat result value
2425         as of type ffi_arg.
2426         * testsuite/libffi.call/struct3.c (main): Fix CHECK.
2427
2428 2004-01-22  Ulrich Weigand  <uweigand@de.ibm.com>
2429
2430         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Treat result
2431         value as of type ffi_arg, not unsigned int.
2432
2433 2004-01-21  Michael Ritzert  <ritzert@t-online.de>
2434
2435         * ffi64.c (ffi_prep_args): Cast the RHS of an assignment instead
2436         of the LHS.
2437
2438 2004-01-12  Andreas Tobler  <a.tobler@schweiz.ch>
2439
2440         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_32 for
2441         Solaris.
2442
2443 2004-01-08  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2444
2445         * testsuite/libffi.call/ffitest.h (allocate_mmap): Cast MAP_FAILED
2446         to void *.
2447
2448 2003-12-10  Richard Henderson  <rth@redhat.com>
2449
2450         * testsuite/libffi.call/cls_align_pointer.c: Cast pointers to
2451         size_t instead of int.
2452
2453 2003-12-04  Hosaka Yuji  <hos@tamanegi.org>
2454
2455         * testsuite/libffi.call/many_win32.c: Include <float.h>.
2456         * testsuite/libffi.call/many_win32.c (main): Replace variable
2457         int i with unsigned long ul.
2458
2459         * testsuite/libffi.call/cls_align_uint64.c: New test case.
2460         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2461         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2462         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2463         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2464         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2465         * testsuite/libffi.call/cls_align_float.c: Likewise.
2466         * testsuite/libffi.call/cls_align_double.c: Likewise.
2467         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2468         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2469
2470 2003-12-02  Hosaka Yuji  <hos@tamanegi.org>
2471
2472         PR other/13221
2473         * src/x86/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV):
2474         Align arguments to 32 bits.
2475
2476 2003-12-01  Andreas Tobler  <a.tobler@schweiz.ch>
2477
2478         PR other/13221
2479         * testsuite/libffi.call/cls_multi_sshort.c: New test case.
2480         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2481         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2482         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2483         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2484         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2485
2486         * testsuite/libffi.special/unwindtest.cc: Cosmetics.
2487
2488 2003-11-26  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
2489
2490         * testsuite/libffi.call/ffitest.h: Include <fcntl.h>.
2491         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2492
2493 2003-11-22  Andreas Tobler  <a.tobler@schweiz.ch>
2494
2495         * Makefile.in: Rebuilt.
2496         * configure: Likewise.
2497         * testsuite/libffi.special/unwindtest.cc: Convert the mmap to
2498         the right type.
2499
2500 2003-11-21  Andreas Jaeger  <aj@suse.de>
2501             Andreas Tobler  <a.tobler@schweiz.ch>
2502
2503         * acinclude.m4: Add AC_FUNC_MMAP_BLACKLIST.
2504         * configure.in: Call AC_FUNC_MMAP_BLACKLIST.
2505         * Makefile.in: Rebuilt.
2506         * aclocal.m4: Likewise.
2507         * configure: Likewise.
2508         * fficonfig.h.in: Likewise.
2509         * testsuite/lib/libffi-dg.exp: Add include dir.
2510         * testsuite/libffi.call/ffitest.h: Add MMAP definitions.
2511         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2512         * testsuite/libffi.call/closure_fn0.c: Use MMAP functionality
2513         for ffi_closure if available.
2514         * testsuite/libffi.call/closure_fn1.c: Likewise.
2515         * testsuite/libffi.call/closure_fn2.c: Likewise.
2516         * testsuite/libffi.call/closure_fn3.c: Likewise.
2517         * testsuite/libffi.call/closure_fn4.c: Likewise.
2518         * testsuite/libffi.call/closure_fn5.c: Likewise.
2519         * testsuite/libffi.call/cls_12byte.c: Likewise.
2520         * testsuite/libffi.call/cls_16byte.c: Likewise.
2521         * testsuite/libffi.call/cls_18byte.c: Likewise.
2522         * testsuite/libffi.call/cls_19byte.c: Likewise.
2523         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2524         * testsuite/libffi.call/cls_20byte.c: Likewise.
2525         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2526         * testsuite/libffi.call/cls_24byte.c: Likewise.
2527         * testsuite/libffi.call/cls_2byte.c: Likewise.
2528         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2529         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2530         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2531         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2532         * testsuite/libffi.call/cls_4byte.c: Likewise.
2533         * testsuite/libffi.call/cls_5byte.c: Likewise.
2534         * testsuite/libffi.call/cls_64byte.c: Likewise.
2535         * testsuite/libffi.call/cls_6byte.c: Likewise.
2536         * testsuite/libffi.call/cls_7byte.c: Likewise.
2537         * testsuite/libffi.call/cls_8byte.c: Likewise.
2538         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2539         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2540         * testsuite/libffi.call/cls_double.c: Likewise.
2541         * testsuite/libffi.call/cls_float.c: Likewise.
2542         * testsuite/libffi.call/cls_schar.c: Likewise.
2543         * testsuite/libffi.call/cls_sint.c: Likewise.
2544         * testsuite/libffi.call/cls_sshort.c: Likewise.
2545         * testsuite/libffi.call/cls_uchar.c: Likewise.
2546         * testsuite/libffi.call/cls_uint.c: Likewise.
2547         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2548         * testsuite/libffi.call/cls_ushort.c: Likewise.
2549         * testsuite/libffi.call/nested_struct.c: Likewise.
2550         * testsuite/libffi.call/nested_struct1.c: Likewise.
2551         * testsuite/libffi.call/nested_struct2.c: Likewise.
2552         * testsuite/libffi.call/nested_struct3.c: Likewise.
2553         * testsuite/libffi.call/problem1.c: Likewise.
2554         * testsuite/libffi.special/unwindtest.cc: Likewise.
2555
2556 2003-11-20  Andreas Tobler  <a.tobler@schweiz.ch>
2557
2558         * testsuite/lib/libffi-dg.exp: Make the -lgcc_s conditional.
2559
2560 2003-11-19  Andreas Tobler  <a.tobler@schweiz.ch>
2561
2562         * testsuite/lib/libffi-dg.exp: Add DYLD_LIBRARY_PATH for darwin.
2563         Add -lgcc_s to additional flags.
2564
2565 2003-11-12  Andreas Tobler  <a.tobler@schweiz.ch>
2566
2567         * configure.in, include/Makefile.am: PR libgcj/11147, install
2568         the ffitarget.h header file in a gcc versioned and target
2569         dependent place.
2570         * configure: Regenerated.
2571         * Makefile.in, include/Makefile.in: Likewise.
2572         * testsuite/Makefile.in: Likewise.
2573
2574 2003-11-09  Andreas Tobler  <a.tobler@schweiz.ch>
2575
2576         * testsuite/libffi.call/closure_fn0.c: Print result and check
2577         with dg-output to make debugging easier.
2578         * testsuite/libffi.call/closure_fn1.c: Likewise.
2579         * testsuite/libffi.call/closure_fn2.c: Likewise.
2580         * testsuite/libffi.call/closure_fn3.c: Likewise.
2581         * testsuite/libffi.call/closure_fn4.c: Likewise.
2582         * testsuite/libffi.call/closure_fn5.c: Likewise.
2583         * testsuite/libffi.call/cls_12byte.c: Likewise.
2584         * testsuite/libffi.call/cls_16byte.c: Likewise.
2585         * testsuite/libffi.call/cls_18byte.c: Likewise.
2586         * testsuite/libffi.call/cls_19byte.c: Likewise.
2587         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2588         * testsuite/libffi.call/cls_20byte.c: Likewise.
2589         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2590         * testsuite/libffi.call/cls_24byte.c: Likewise.
2591         * testsuite/libffi.call/cls_2byte.c: Likewise.
2592         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2593         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2594         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2595         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2596         * testsuite/libffi.call/cls_4byte.c: Likewise.
2597         * testsuite/libffi.call/cls_5byte.c: Likewise.
2598         * testsuite/libffi.call/cls_64byte.c: Likewise.
2599         * testsuite/libffi.call/cls_6byte.c: Likewise.
2600         * testsuite/libffi.call/cls_7byte.c: Likewise.
2601         * testsuite/libffi.call/cls_8byte.c: Likewise.
2602         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2603         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2604         * testsuite/libffi.call/cls_double.c: Likewise.
2605         * testsuite/libffi.call/cls_float.c: Likewise.
2606         * testsuite/libffi.call/cls_schar.c: Likewise.
2607         * testsuite/libffi.call/cls_sint.c: Likewise.
2608         * testsuite/libffi.call/cls_sshort.c: Likewise.
2609         * testsuite/libffi.call/cls_uchar.c: Likewise.
2610         * testsuite/libffi.call/cls_uint.c: Likewise.
2611         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2612         * testsuite/libffi.call/cls_ushort.c: Likewise.
2613         * testsuite/libffi.call/problem1.c: Likewise.
2614
2615         * testsuite/libffi.special/unwindtest.cc: Make ffi_closure
2616         static.
2617
2618 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
2619
2620         * testsuite/libffi.call/cls_9byte2.c: New test case.
2621         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2622         * testsuite/libffi.call/cls_64byte.c: Likewise.
2623         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2624         * testsuite/libffi.call/cls_19byte.c: Likewise.
2625         * testsuite/libffi.call/cls_18byte.c: Likewise.
2626         * testsuite/libffi.call/closure_fn4.c: Likewise.
2627         * testsuite/libffi.call/closure_fn5.c: Likewise.
2628         * testsuite/libffi.call/cls_schar.c: Likewise.
2629         * testsuite/libffi.call/cls_sint.c: Likewise.
2630         * testsuite/libffi.call/cls_sshort.c: Likewise.
2631         * testsuite/libffi.call/nested_struct2.c: Likewise.
2632         * testsuite/libffi.call/nested_struct3.c: Likewise.
2633
2634 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
2635
2636         * testsuite/libffi.call/cls_double.c: Do a check on the result.
2637         * testsuite/libffi.call/cls_uchar.c: Likewise.
2638         * testsuite/libffi.call/cls_uint.c: Likewise.
2639         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2640         * testsuite/libffi.call/cls_ushort.c: Likewise.
2641         * testsuite/libffi.call/return_sc.c: Cleanup whitespaces.
2642
2643 2003-11-06  Andreas Tobler  <a.tobler@schweiz.ch>
2644
2645         * src/prep_cif.c (ffi_prep_cif): Move the validity check after
2646         the initialization.
2647
2648 2003-10-23  Andreas Tobler  <a.tobler@schweiz.ch>
2649
2650         * src/java_raw_api.c (ffi_java_ptrarray_to_raw): Replace
2651         FFI_ASSERT(FALSE) with FFI_ASSERT(0).
2652
2653 2003-10-22  David Daney  <ddaney@avtrex.com>
2654
2655         * src/mips/ffitarget.h: Replace undefined UINT32 and friends with
2656         __attribute__((__mode__(__SI__))) and friends.
2657
2658 2003-10-22  Andreas Schwab  <schwab@suse.de>
2659
2660         * src/ia64/ffi.c: Replace FALSE/TRUE with false/true.
2661
2662 2003-10-21  Andreas Tobler  <a.tobler@schweiz.ch>
2663
2664         * configure.in: AC_LINK_FILES(ffitarget.h).
2665         * configure: Regenerate.
2666         * Makefile.in: Likewise.
2667         * include/Makefile.in: Likewise.
2668         * testsuite/Makefile.in: Likewise.
2669         * fficonfig.h.in: Likewise.
2670
2671 2003-10-21  Paolo Bonzini  <bonzini@gnu.org>
2672             Richard Henderson  <rth@redhat.com>
2673
2674         Avoid that ffi.h includes fficonfig.h.
2675
2676         * Makefile.am (EXTRA_DIST): Include ffitarget.h files
2677         (TARGET_SRC_MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
2678         (TARGET_SRC_MIPS_SGI): Removed.
2679         (MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
2680         (MIPS_SGI): Removed.
2681         (CLEANFILES): Removed.
2682         (mostlyclean-am, clean-am, mostlyclean-sub, clean-sub): New
2683         targets.
2684         * acconfig.h: Removed.
2685         * configure.in: Compute sizeofs only for double and long double.
2686         Use them to define and subst HAVE_LONG_DOUBLE.  Include comments
2687         into AC_DEFINE instead of using acconfig.h.  Create
2688         include/ffitarget.h instead of include/fficonfig.h.  Rename
2689         MIPS_GCC to MIPS_IRIX, drop MIPS_SGI since we are in gcc's tree.
2690         AC_DEFINE EH_FRAME_FLAGS.
2691         * include/Makefile.am (DISTCLEANFILES): New automake macro.
2692         (hack_DATA): Add ffitarget.h.
2693         * include/ffi.h.in: Remove all system specific definitions.
2694         Declare raw API even if it is not installed, why bother?
2695         Use limits.h instead of SIZEOF_* to define ffi_type_*.  Do
2696         not define EH_FRAME_FLAGS, it is in fficonfig.h now.  Include
2697         ffitarget.h instead of fficonfig.h.  Remove ALIGN macro.
2698         (UINT_ARG, INT_ARG): Removed, use ffi_arg and ffi_sarg instead.
2699         * include/ffi_common.h (bool): Do not define.
2700         (ffi_assert): Accept failed assertion.
2701         (ffi_type_test): Return void and accept file/line.
2702         (FFI_ASSERT): Pass stringized failed assertion.
2703         (FFI_ASSERT_AT): New macro.
2704         (FFI_ASSERT_VALID_TYPE): New macro.
2705         (UINT8, SINT8, UINT16, SINT16, UINT32, SINT32,
2706         UINT64, SINT64): Define here with gcc's __attribute__ macro
2707         instead of in ffi.h
2708         (FLOAT32, ALIGN): Define here instead of in ffi.h
2709         * include/ffi-mips.h: Removed.  Its content moved to
2710         src/mips/ffitarget.h after separating assembly and C sections.
2711         * src/alpha/ffi.c, src/alpha/ffi.c, src/java_raw_api.c
2712         src/prep_cif.c, src/raw_api.c, src/ia64/ffi.c,
2713         src/mips/ffi.c, src/mips/n32.S, src/mips/o32.S,
2714         src/mips/ffitarget.h, src/sparc/ffi.c, src/x86/ffi64.c:
2715         SIZEOF_ARG -> FFI_SIZEOF_ARG.
2716         * src/ia64/ffi.c: Include stdbool.h (provided by GCC 2.95+).
2717         * src/debug.c (ffi_assert): Accept stringized failed assertion.
2718         (ffi_type_test): Rewritten.
2719         * src/prep-cif.c (initialize_aggregate, ffi_prep_cif): Call
2720         FFI_ASSERT_VALID_TYPE.
2721         * src/alpha/ffitarget.h, src/arm/ffitarget.h,
2722         src/ia64/ffitarget.h, src/m68k/ffitarget.h,
2723         src/mips/ffitarget.h, src/powerpc/ffitarget.h,
2724         src/s390/ffitarget.h, src/sh/ffitarget.h,
2725         src/sh64/ffitarget.h, src/sparc/ffitarget.h,
2726         src/x86/ffitarget.h: New files.
2727         * src/alpha/osf.S, src/arm/sysv.S, src/ia64/unix.S,
2728         src/m68k/sysv.S, src/mips/n32.S, src/mips/o32.S,
2729         src/powerpc/aix.S, src/powerpc/darwin.S,
2730         src/powerpc/ffi_darwin.c, src/powerpc/linux64.S,
2731         src/powerpc/linux64_closure.S, src/powerpc/ppc_closure.S,
2732         src/powerpc/sysv.S, src/s390/sysv.S, src/sh/sysv.S,
2733         src/sh64/sysv.S, src/sparc/v8.S, src/sparc/v9.S,
2734         src/x86/sysv.S, src/x86/unix64.S, src/x86/win32.S:
2735         include fficonfig.h
2736
2737 2003-10-20  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2738
2739         * src/mips/ffi.c: Use _ABIN32, _ABIO32 instead of external
2740         _MIPS_SIM_NABI32, _MIPS_SIM_ABI32.
2741
2742 2003-10-19  Andreas Tobler  <a.tobler@schweiz.ch>
2743
2744         * src/powerpc/ffi_darwin.c (ffi_prep_args): Declare bytes again.
2745         Used when FFI_DEBUG = 1.
2746
2747 2003-10-14  Alan Modra  <amodra@bigpond.net.au>
2748
2749         * src/types.c (double, longdouble): Default POWERPC64 to 8 byte size
2750         and align.
2751
2752 2003-10-06  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2753
2754         * include/ffi_mips.h: Define FFI_MIPS_N32 for N32/N64 ABIs,
2755         FFI_MIPS_O32 for O32 ABI.
2756
2757 2003-10-01  Andreas Tobler  <a.tobler@schweiz.ch>
2758
2759         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_64 for
2760         SPARC64. Cleanup whitespaces.
2761
2762 2003-09-19  Andreas Tobler  <a.tobler@schweiz.ch>
2763
2764         * testsuite/libffi.call/closure_fn0.c: Xfail mips, arm,
2765         strongarm, xscale. Cleanup whitespaces.
2766         * testsuite/libffi.call/closure_fn1.c: Likewise.
2767         * testsuite/libffi.call/closure_fn2.c: Likewise.
2768         * testsuite/libffi.call/closure_fn3.c: Likewise.
2769         * testsuite/libffi.call/cls_12byte.c: Likewise.
2770         * testsuite/libffi.call/cls_16byte.c: Likewise.
2771         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2772         * testsuite/libffi.call/cls_20byte.c: Likewise.
2773         * testsuite/libffi.call/cls_24byte.c: Likewise.
2774         * testsuite/libffi.call/cls_2byte.c: Likewise.
2775         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2776         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2777         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2778         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2779         * testsuite/libffi.call/cls_4byte.c: Likewise.
2780         * testsuite/libffi.call/cls_5byte.c: Likewise.
2781         * testsuite/libffi.call/cls_6byte.c: Likewise.
2782         * testsuite/libffi.call/cls_7byte.c: Likewise.
2783         * testsuite/libffi.call/cls_8byte.c: Likewise.
2784         * testsuite/libffi.call/cls_double.c: Likewise.
2785         * testsuite/libffi.call/cls_float.c: Likewise.
2786         * testsuite/libffi.call/cls_uchar.c: Likewise.
2787         * testsuite/libffi.call/cls_uint.c: Likewise.
2788         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2789         * testsuite/libffi.call/cls_ushort.c: Likewise.
2790         * testsuite/libffi.call/nested_struct.c: Likewise.
2791         * testsuite/libffi.call/nested_struct1.c: Likewise.
2792         * testsuite/libffi.call/problem1.c: Likewise.
2793         * testsuite/libffi.special/unwindtest.cc: Likewise.
2794         * testsuite/libffi.call/pyobjc-tc.c: Cleanup whitespaces.
2795
2796 2003-09-18  David Edelsohn  <edelsohn@gnu.org>
2797
2798         * src/powerpc/aix.S: Cleanup whitespaces.
2799         * src/powerpc/aix_closure.S: Likewise.
2800
2801 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
2802
2803         * src/powerpc/darwin.S: Cleanup whitespaces, comment formatting.
2804         * src/powerpc/darwin_closure.S: Likewise.
2805         * src/powerpc/ffi_darwin.c: Likewise.
2806
2807 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
2808             David Edelsohn  <edelsohn@gnu.org>
2809
2810         * src/types.c (double): Add AIX and Darwin to the right TYPEDEF.
2811         * src/powerpc/aix_closure.S: Remove the pointer to the outgoing
2812         parameter stack.
2813         * src/powerpc/darwin_closure.S: Likewise.
2814         * src/powerpc/ffi_darwin.c (ffi_prep_args): Handle structures
2815         according to the Darwin/AIX ABI.
2816         (ffi_prep_cif_machdep): Likewise.
2817         (ffi_closure_helper_DARWIN): Likewise.
2818         Remove the outgoing parameter stack logic. Simplify the evaluation
2819         of the different CASE types.
2820         (ffi_prep_clousure): Avoid the casts on lvalues. Change the branch
2821         statement in the trampoline code.
2822
2823 2003-09-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2824
2825         * src/sh/ffi.c (ffi_prep_args): Take account into the alignement
2826         for the register size.
2827         (ffi_closure_helper_SYSV): Handle the structure return value
2828         address correctly.
2829         (ffi_closure_helper_SYSV): Return the appropriate type when
2830         the registers are used for the structure return value.
2831         * src/sh/sysv.S (ffi_closure_SYSV): Fix the stack layout for
2832         the 64-bit return value.  Update copyright years.
2833
2834 2003-09-17  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2835
2836         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Search in
2837         srcdir for ffi_mips.h.
2838
2839 2003-09-12  Alan Modra  <amodra@bigpond.net.au>
2840
2841         * src/prep_cif.c (initialize_aggregate): Include tail padding in
2842         structure size.
2843         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Correct
2844         placement of float result.
2845         * testsuite/libffi.special/unwindtest.cc (closure_test_fn1): Correct
2846         cast of "resp" for big-endian 64 bit machines.
2847
2848 2003-09-11  Alan Modra  <amodra@bigpond.net.au>
2849
2850         * src/types.c (double, longdouble): Merge identical SH and ARM
2851         typedefs, and add POWERPC64.
2852         * src/powerpc/ffi.c (ffi_prep_args64): Correct next_arg calc for
2853         struct split over gpr and rest.
2854         (ffi_prep_cif_machdep): Correct intarg_count for structures.
2855         * src/powerpc/linux64.S (ffi_call_LINUX64): Fix gpr offsets.
2856
2857 2003-09-09  Andreas Tobler  <a.tobler@schweiz.ch>
2858
2859         * src/powerpc/ffi.c (ffi_closure_helper_SYSV) Handle struct
2860         passing correctly.
2861
2862 2003-09-09  Alan Modra  <amodra@bigpond.net.au>
2863
2864         * configure: Regenerate.
2865
2866 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
2867
2868         * Makefile.am: Remove build rules for ffitest.
2869         * Makefile.in: Rebuilt.
2870
2871 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
2872
2873         * src/java_raw_api.c: Include <stdlib.h> to fix compiler warning
2874         about implicit declaration of abort().
2875
2876 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
2877
2878         * Makefile.am: Add dejagnu test framework. Fixes PR other/11411.
2879         * Makefile.in: Rebuilt.
2880         * configure.in: Add dejagnu test framework.
2881         * configure: Rebuilt.
2882
2883         * testsuite/Makefile.am: New file.
2884         * testsuite/Makefile.in: Built
2885         * testsuite/lib/libffi-dg.exp: New file.
2886         * testsuite/config/default.exp: Likewise.
2887         * testsuite/libffi.call/call.exp: Likewise.
2888         * testsuite/libffi.call/ffitest.h: Likewise.
2889         * testsuite/libffi.call/closure_fn0.c: Likewise.
2890         * testsuite/libffi.call/closure_fn1.c: Likewise.
2891         * testsuite/libffi.call/closure_fn2.c: Likewise.
2892         * testsuite/libffi.call/closure_fn3.c: Likewise.
2893         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2894         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2895         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2896         * testsuite/libffi.call/cls_2byte.c: Likewise.
2897         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2898         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2899         * testsuite/libffi.call/cls_4byte.c: Likewise.
2900         * testsuite/libffi.call/cls_5byte.c: Likewise.
2901         * testsuite/libffi.call/cls_6byte.c: Likewise.
2902         * testsuite/libffi.call/cls_7byte.c: Likewise.
2903         * testsuite/libffi.call/cls_8byte.c: Likewise.
2904         * testsuite/libffi.call/cls_12byte.c: Likewise.
2905         * testsuite/libffi.call/cls_16byte.c: Likewise.
2906         * testsuite/libffi.call/cls_20byte.c: Likewise.
2907         * testsuite/libffi.call/cls_24byte.c: Likewise.
2908         * testsuite/libffi.call/cls_double.c: Likewise.
2909         * testsuite/libffi.call/cls_float.c: Likewise.
2910         * testsuite/libffi.call/cls_uchar.c: Likewise.
2911         * testsuite/libffi.call/cls_uint.c: Likewise.
2912         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2913         * testsuite/libffi.call/cls_ushort.c: Likewise.
2914         * testsuite/libffi.call/float.c: Likewise.
2915         * testsuite/libffi.call/float1.c: Likewise.
2916         * testsuite/libffi.call/float2.c: Likewise.
2917         * testsuite/libffi.call/many.c: Likewise.
2918         * testsuite/libffi.call/many_win32.c: Likewise.
2919         * testsuite/libffi.call/nested_struct.c: Likewise.
2920         * testsuite/libffi.call/nested_struct1.c: Likewise.
2921         * testsuite/libffi.call/pyobjc-tc.c: Likewise.
2922         * testsuite/libffi.call/problem1.c: Likewise.
2923         * testsuite/libffi.call/promotion.c: Likewise.
2924         * testsuite/libffi.call/return_ll.c: Likewise.
2925         * testsuite/libffi.call/return_sc.c: Likewise.
2926         * testsuite/libffi.call/return_uc.c: Likewise.
2927         * testsuite/libffi.call/strlen.c: Likewise.
2928         * testsuite/libffi.call/strlen_win32.c: Likewise.
2929         * testsuite/libffi.call/struct1.c: Likewise.
2930         * testsuite/libffi.call/struct2.c: Likewise.
2931         * testsuite/libffi.call/struct3.c: Likewise.
2932         * testsuite/libffi.call/struct4.c: Likewise.
2933         * testsuite/libffi.call/struct5.c: Likewise.
2934         * testsuite/libffi.call/struct6.c: Likewise.
2935         * testsuite/libffi.call/struct7.c: Likewise.
2936         * testsuite/libffi.call/struct8.c: Likewise.
2937         * testsuite/libffi.call/struct9.c: Likewise.
2938         * testsuite/libffi.special/special.exp: New file.
2939         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2940         * testsuite/libffi.special/unwindtest.cc: Likewise.
2941
2942
2943 2003-08-13  Kaz Kojima  <kkojima@gcc.gnu.org>
2944
2945         * src/sh/ffi.c (OFS_INT16): Set 0 for little endian case.  Update
2946         copyright years.
2947
2948 2003-08-02  Alan Modra  <amodra@bigpond.net.au>
2949
2950         * src/powerpc/ffi.c (ffi_prep_args64): Modify for changed gcc
2951         structure passing.
2952         (ffi_closure_helper_LINUX64): Likewise.
2953         * src/powerpc/linux64.S: Remove code writing to parm save area.
2954         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Use return
2955         address in lr from ffi_closure_helper_LINUX64 call to calculate
2956         table address.  Optimize function tail.
2957
2958 2003-07-28  Andreas Tobler  <a.tobler@schweiz.ch>
2959
2960         * src/sparc/ffi.c: Handle all floating point registers.
2961         * src/sparc/v9.S: Likewise. Fixes second part of PR target/11410.
2962
2963 2003-07-11  Gerald Pfeifer  <pfeifer@dbai.tuwien.ac.at>
2964
2965         * README: Note that libffi is not part of GCC.  Update the project
2966         URL and status.
2967
2968 2003-06-19  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
2969
2970         * src/powerpc/ppc_closure.S: Include ffi.h.
2971
2972 2003-06-13  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2973
2974         * src/x86/sysv.S: Avoid gas-only .uleb128/.sleb128 directives.
2975         Use C style comments.
2976
2977 2003-06-13  Kaz Kojima  <kkojima@rr.iij4u.or.jp>
2978
2979         * Makefile.am: Add SHmedia support.  Fix a typo of SH support.
2980         * Makefile.in: Regenerate.
2981         * configure.in (sh64-*-linux*, sh5*-*-linux*): Add target.
2982         * configure: Regenerate.
2983         * include/ffi.h.in: Add SHmedia support.
2984         * src/sh64/ffi.c: New file.
2985         * src/sh64/sysv.S: New file.
2986
2987 2003-05-16  Jakub Jelinek  <jakub@redhat.com>
2988
2989         * configure.in (HAVE_RO_EH_FRAME): Check whether .eh_frame section
2990         should be read-only.
2991         * configure: Rebuilt.
2992         * fficonfig.h.in: Rebuilt.
2993         * include/ffi.h.in (EH_FRAME_FLAGS): Define.
2994         * src/alpha/osf.S: Use EH_FRAME_FLAGS.
2995         * src/powerpc/linux64.S: Likewise.
2996         * src/powerpc/linux64_closure.S: Likewise.  Include ffi.h.
2997         * src/powerpc/sysv.S: Use EH_FRAME_FLAGS.  Use pcrel encoding
2998         if -fpic/-fPIC/-mrelocatable.
2999         * src/powerpc/powerpc_closure.S: Likewise.
3000         * src/sparc/v8.S: If HAVE_RO_EH_FRAME is defined, don't include
3001         #write in .eh_frame flags.
3002         * src/sparc/v9.S: Likewise.
3003         * src/x86/unix64.S: Use EH_FRAME_FLAGS.
3004         * src/x86/sysv.S: Likewise.  Use pcrel encoding if -fpic/-fPIC.
3005         * src/s390/sysv.S: Use EH_FRAME_FLAGS.  Include ffi.h.
3006
3007 2003-05-07  Jeff Sturm  <jsturm@one-point.com>
3008
3009         Fixes PR bootstrap/10656
3010         * configure.in (HAVE_AS_REGISTER_PSEUDO_OP): Test assembler
3011         support for .register pseudo-op.
3012         * src/sparc/v8.S: Use it.
3013         * fficonfig.h.in: Rebuilt.
3014         * configure: Rebuilt.
3015
3016 2003-04-18  Jakub Jelinek  <jakub@redhat.com>
3017
3018         * include/ffi.h.in (POWERPC64): Define if 64-bit.
3019         (enum ffi_abi): Add FFI_LINUX64 on POWERPC.
3020         Make it the default on POWERPC64.
3021         (FFI_TRAMPOLINE_SIZE): Define to 24 on POWERPC64.
3022         * configure.in: Change powerpc-*-linux* into powerpc*-*-linux*.
3023         * configure: Rebuilt.
3024         * src/powerpc/ffi.c (hidden): Define.
3025         (ffi_prep_args_SYSV): Renamed from
3026         ffi_prep_args.  Cast pointers to unsigned long to shut up warnings.
3027         (NUM_GPR_ARG_REGISTERS64, NUM_FPR_ARG_REGISTERS64,
3028         ASM_NEEDS_REGISTERS64): New.
3029         (ffi_prep_args64): New function.
3030         (ffi_prep_cif_machdep): Handle FFI_LINUX64 ABI.
3031         (ffi_call): Likewise.
3032         (ffi_prep_closure): Likewise.
3033         (flush_icache): Surround by #ifndef POWERPC64.
3034         (ffi_dblfl): New union type.
3035         (ffi_closure_helper_SYSV): Use it to avoid aliasing problems.
3036         (ffi_closure_helper_LINUX64): New function.
3037         * src/powerpc/ppc_closure.S: Surround whole file by #ifndef
3038         __powerpc64__.
3039         * src/powerpc/sysv.S: Likewise.
3040         (ffi_call_SYSV): Rename ffi_prep_args to ffi_prep_args_SYSV.
3041         * src/powerpc/linux64.S: New file.
3042         * src/powerpc/linux64_closure.S: New file.
3043         * Makefile.am (EXTRA_DIST): Add src/powerpc/linux64.S and
3044         src/powerpc/linux64_closure.S.
3045         (TARGET_SRC_POWERPC): Likewise.
3046
3047         * src/ffitest.c (closure_test_fn, closure_test_fn1, closure_test_fn2,
3048         closure_test_fn3): Fix result printing on big-endian 64-bit
3049         machines.
3050         (main): Print tst2_arg instead of uninitialized tst2_result.
3051
3052         * src/ffitest.c (main): Hide what closure pointer really points to
3053         from the compiler.
3054
3055 2003-04-16  Richard Earnshaw  <rearnsha@arm.com>
3056
3057         * configure.in (arm-*-netbsdelf*): Add configuration.
3058         (configure): Regenerated.
3059
3060 2003-04-04  Loren J. Rittle  <ljrittle@acm.org>
3061
3062         * include/Makefile.in: Regenerate.
3063
3064 2003-03-21  Zdenek Dvorak  <rakdver@atrey.karlin.mff.cuni.cz>
3065
3066         * libffi/include/ffi.h.in: Define X86 instead of X86_64 in 32
3067         bit mode.
3068         * libffi/src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV):
3069         Receive closure pointer through parameter, read args using
3070         __builtin_dwarf_cfa.
3071         (FFI_INIT_TRAMPOLINE): Send closure reference through eax.
3072
3073 2003-03-12  Andreas Schwab  <schwab@suse.de>
3074
3075         * configure.in: Avoid trailing /. in toolexeclibdir.
3076         * configure: Rebuilt.
3077
3078 2003-03-03  Andreas Tobler <a.tobler@schweiz.ch>
3079
3080         * src/powerpc/darwin_closure.S: Recode to fit dynamic libraries.
3081
3082 2003-02-06  Andreas Tobler <a.tobler@schweiz.ch>
3083
3084         * libffi/src/powerpc/darwin_closure.S:
3085         Fix alignement bug, allocate 8 bytes for the result.
3086         * libffi/src/powerpc/aix_closure.S:
3087         Likewise.
3088         * libffi/src/powerpc/ffi_darwin.c:
3089         Update stackframe description for aix/darwin_closure.S.
3090
3091 2003-02-06  Jakub Jelinek  <jakub@redhat.com>
3092
3093         * src/s390/ffi.c (ffi_closure_helper_SYSV): Add hidden visibility
3094         attribute.
3095
3096 2003-01-31  Christian Cornelssen  <ccorn@cs.tu-berlin.de>,
3097             Andreas Schwab  <schwab@suse.de>
3098
3099         * configure.in: Adjust command to source config-ml.in to account
3100         for changes to the libffi_basedir definition.
3101         (libffi_basedir): Remove ${srcdir} from value and include trailing
3102         slash if nonempty.
3103
3104         * configure: Regenerate.
3105
3106 2003-01-29  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3107
3108         * src/powerpc/ppc_closure.S: Recode to fit shared libs.
3109
3110 2003-01-28  Andrew Haley  <aph@redhat.com>
3111
3112         * include/ffi.h.in: Enable FFI_CLOSURES for x86_64.
3113         * src/x86/ffi64.c (ffi_prep_closure): New.
3114         (ffi_closure_UNIX64_inner): New.
3115         * src/x86/unix64.S (ffi_closure_UNIX64): New.
3116
3117 2003-01-27  Alexandre Oliva  <aoliva@redhat.com>
3118
3119         * configure.in (toolexecdir, toolexeclibdir): Set and AC_SUBST.
3120         Remove USE_LIBDIR conditional.
3121         * Makefile.am (toolexecdir, toolexeclibdir): Don't override.
3122         * Makefile.in, configure: Rebuilt.
3123
3124 2003-01027  David Edelsohn  <edelsohn@gnu.org>
3125
3126         * Makefile.am (TARGET_SRC_POWERPC_AIX): Fix typo.
3127         * Makefile.in: Regenerate.
3128
3129 2003-01-22  Andrew Haley  <aph@redhat.com>
3130
3131         * src/powerpc/darwin.S (_ffi_call_AIX): Add Augmentation size to
3132         unwind info.
3133
3134 2003-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
3135
3136         * src/powerpc/darwin.S: Add unwind info.
3137         * src/powerpc/darwin_closure.S: Likewise.
3138
3139 2003-01-14  Andrew Haley  <aph@redhat.com>
3140
3141         * src/x86/ffi64.c (ffi_prep_args): Check for void retval.
3142         (ffi_prep_cif_machdep): Likewise.
3143         * src/x86/unix64.S: Add unwind info.
3144
3145 2003-01-14  Andreas Jaeger  <aj@suse.de>
3146
3147         * src/ffitest.c (main): Only use ffi_closures if those are
3148         supported.
3149
3150 2003-01-13 Andreas Tobler <a.tobler@schweiz.ch>
3151
3152         * libffi/src/ffitest.c
3153          add closure testcases
3154
3155 2003-01-13 Kevin B. Hendricks <khendricks@ivey.uwo.ca>
3156
3157         * libffi/src/powerpc/ffi.c
3158          fix alignment bug for float (4 byte aligned iso 8 byte)
3159
3160 2003-01-09  Geoffrey Keating  <geoffk@apple.com>
3161
3162         * src/powerpc/ffi_darwin.c: Remove RCS version string.
3163         * src/powerpc/darwin.S: Remove RCS version string.
3164
3165 2003-01-03  Jeff Sturm  <jsturm@one-point.com>
3166
3167         * include/ffi.h.in: Add closure defines for SPARC, SPARC64.
3168         * src/ffitest.c (main): Use static storage for closure.
3169         * src/sparc/ffi.c (ffi_prep_closure, ffi_closure_sparc_inner): New.
3170         * src/sparc/v8.S (ffi_closure_v8): New.
3171         * src/sparc/v9.S (ffi_closure_v9): New.
3172
3173 2002-11-10  Ranjit Mathew <rmathew@hotmail.com>
3174
3175         * include/ffi.h.in: Added FFI_STDCALL ffi_type
3176           enumeration for X86_WIN32.
3177         * src/x86/win32.S: Added ffi_call_STDCALL function
3178           definition.
3179         * src/x86/ffi.c (ffi_call/ffi_raw_call): Added
3180           switch cases for recognising FFI_STDCALL and
3181           calling ffi_call_STDCALL if target is X86_WIN32.
3182         * src/ffitest.c (my_stdcall_strlen/stdcall_many):
3183           stdcall versions of the "my_strlen" and "many"
3184           test functions (for X86_WIN32).
3185           Added test cases to test stdcall invocation using
3186           these functions.
3187
3188 2002-12-02  Kaz Kojima  <kkojima@gcc.gnu.org>
3189
3190         * src/sh/sysv.S: Add DWARF2 unwind info.
3191
3192 2002-11-27  Ulrich Weigand  <uweigand@de.ibm.com>
3193
3194         * src/s390/sysv.S (.eh_frame section): Make section read-only.
3195
3196 2002-11-26  Jim Wilson  <wilson@redhat.com>
3197
3198         * src/types.c (FFI_TYPE_POINTER): Has size 8 on IA64.
3199
3200 2002-11-23  H.J. Lu <hjl@gnu.org>
3201
3202         * acinclude.m4: Add dummy AM_PROG_LIBTOOL.
3203         Include ../config/accross.m4.
3204         * aclocal.m4; Rebuild.
3205         * configure: Likewise.
3206
3207 2002-11-15  Ulrich Weigand  <uweigand@de.ibm.com>
3208
3209         * src/s390/sysv.S (.eh_frame section): Adapt to pcrel FDE encoding.
3210
3211 2002-11-11  DJ Delorie  <dj@redhat.com>
3212
3213         * configure.in: Look for common files in the right place.
3214
3215 2002-10-08  Ulrich Weigand  <uweigand@de.ibm.com>
3216
3217         * src/java_raw_api.c (ffi_java_raw_to_ptrarray): Interpret
3218         raw data as _Jv_word values, not ffi_raw.
3219         (ffi_java_ptrarray_to_raw): Likewise.
3220         (ffi_java_rvalue_to_raw): New function.
3221         (ffi_java_raw_call): Call it.
3222         (ffi_java_raw_to_rvalue): New function.
3223         (ffi_java_translate_args): Call it.
3224         * src/ffitest.c (closure_test_fn): Interpret return value
3225         as ffi_arg, not int.
3226         * src/s390/ffi.c (ffi_prep_cif_machdep): Add missing
3227         FFI_TYPE_POINTER case.
3228         (ffi_closure_helper_SYSV): Likewise.  Also, assume return
3229         values extended to word size.
3230
3231 2002-10-02  Andreas Jaeger  <aj@suse.de>
3232
3233         * src/x86/ffi64.c (ffi_prep_cif_machdep): Remove debug output.
3234
3235 2002-10-01  Bo Thorsen  <bo@smetana.suse.de>
3236
3237         * include/ffi.h.in: Fix i386 win32 compilation.
3238
3239 2002-09-30  Ulrich Weigand  <uweigand@de.ibm.com>
3240
3241         * configure.in: Add s390x-*-linux-* target.
3242         * configure: Regenerate.
3243         * include/ffi.h.in: Define S390X for s390x targets.
3244         (FFI_CLOSURES): Define for s390/s390x.
3245         (FFI_TRAMPOLINE_SIZE): Likewise.
3246         (FFI_NATIVE_RAW_API): Likewise.
3247         * src/prep_cif.c (ffi_prep_cif): Do not compute stack space for s390.
3248         * src/types.c (FFI_TYPE_POINTER): Use 8-byte pointers on s390x.
3249         * src/s390/ffi.c: Major rework of existing code.  Add support for
3250         s390x targets.  Add closure support.
3251         * src/s390/sysv.S: Likewise.
3252
3253 2002-09-29  Richard Earnshaw  <rearnsha@arm.com>
3254
3255         * src/arm/sysv.S: Fix typo.
3256
3257 2002-09-28  Richard Earnshaw  <rearnsha@arm.com>
3258
3259         * src/arm/sysv.S: If we don't have machine/asm.h and the pre-processor
3260         has defined __USER_LABEL_PREFIX__, then use it in CNAME.
3261         (ffi_call_SYSV): Handle soft-float.
3262
3263 2002-09-27  Bo Thorsen  <bo@suse.de>
3264
3265         * include/ffi.h.in: Fix multilib x86-64 support.
3266
3267 2002-09-22  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
3268
3269         * Makefile.am (all-multi): Fix multilib parallel build.
3270
3271 2002-07-19  Kaz Kojima  <kkojima@gcc.gnu.org>
3272
3273         * configure.in (sh[34]*-*-linux*): Add brackets.
3274         * configure: Regenerate.
3275
3276 2002-07-18  Kaz Kojima  <kkojima@gcc.gnu.org>
3277
3278         * Makefile.am: Add SH support.
3279         * Makefile.in: Regenerate.
3280         * configure.in (sh-*-linux*, sh[34]*-*-linux*): Add target.
3281         * configure: Regenerate.
3282         * include/ffi.h.in: Add SH support.
3283         * src/sh/ffi.c: New file.
3284         * src/sh/sysv.S: New file.
3285         * src/types.c: Add SH support.
3286
3287 2002-07-16  Bo Thorsen  <bo@suse.de>
3288
3289         * src/x86/ffi64.c: New file that adds x86-64 support.
3290         * src/x86/unix64.S: New file that handles argument setup for
3291         x86-64.
3292         * src/x86/sysv.S: Don't use this on x86-64.
3293         * src/x86/ffi.c: Don't use this on x86-64.
3294         Remove unused vars.
3295         * src/prep_cif.c (ffi_prep_cif): Don't do stack size calculation
3296         for x86-64.
3297         * src/ffitest.c (struct6): New test that tests a special case in
3298         the x86-64 ABI.
3299         (struct7): Likewise.
3300         (struct8): Likewise.
3301         (struct9): Likewise.
3302         (closure_test_fn): Silence warning about this when it's not used.
3303         (main): Add the new tests.
3304         (main): Fix a couple of wrong casts and silence some compiler warnings.
3305         * include/ffi.h.in: Add x86-64 ABI definition.
3306         * fficonfig.h.in: Regenerate.
3307         * Makefile.am: Add x86-64 support.
3308         * configure.in: Likewise.
3309         * Makefile.in: Regenerate.
3310         * configure: Likewise.
3311
3312 2002-06-24  Bo Thorsen  <bo@suse.de>
3313
3314         * src/types.c: Merge settings for similar architectures.
3315         Add x86-64 sizes and alignments.
3316
3317 2002-06-23  Bo Thorsen  <bo@suse.de>
3318
3319         * src/arm/ffi.c (ffi_prep_args): Remove unused vars.
3320         * src/sparc/ffi.c (ffi_prep_args_v8): Likewise.
3321         * src/mips/ffi.c (ffi_prep_args): Likewise.
3322         * src/m68k/ffi.c (ffi_prep_args): Likewise.
3323
3324 2002-07-18  H.J. Lu  (hjl@gnu.org)
3325
3326         * Makefile.am (TARGET_SRC_MIPS_LINUX): New.
3327         (libffi_la_SOURCES): Support MIPS_LINUX.
3328         (libffi_convenience_la_SOURCES): Likewise.
3329         * Makefile.in: Regenerated.
3330
3331         * configure.in (mips64*-*): Skip.
3332         (mips*-*-linux*): New.
3333         * configure: Regenerated.
3334
3335         * src/mips/ffi.c: Include <sgidefs.h>.
3336
3337 2002-06-06  Ulrich Weigand  <uweigand@de.ibm.com>
3338
3339         * src/s390/sysv.S: Save/restore %r6.  Add DWARF-2 unwind info.
3340
3341 2002-05-27  Roger Sayle  <roger@eyesopen.com>
3342
3343         * src/x86/ffi.c (ffi_prep_args): Remove reference to avn.
3344
3345 2002-05-27  Bo Thorsen  <bo@suse.de>
3346
3347         * src/x86/ffi.c (ffi_prep_args): Remove unused variable and
3348         fix formatting.
3349
3350 2002-05-13  Andreas Tobler  <a.tobler@schweiz.ch>
3351
3352         * src/powerpc/ffi_darwin.c (ffi_prep_closure): Declare fd at
3353         beginning of function (for older apple cc).
3354
3355 2002-05-08  Alexandre Oliva  <aoliva@redhat.com>
3356
3357         * configure.in (ORIGINAL_LD_FOR_MULTILIBS): Preserve LD at
3358         script entry, and set LD to it when configuring multilibs.
3359         * configure: Rebuilt.
3360
3361 2002-05-05  Jason Thorpe  <thorpej@wasabisystems.com>
3362
3363         * configure.in (sparc64-*-netbsd*): Add target.
3364         (sparc-*-netbsdelf*): Likewise.
3365         * configure: Regenerate.
3366
3367 2002-04-28  David S. Miller  <davem@redhat.com>
3368
3369         * configure.in, configure: Fix SPARC test in previous change.
3370
3371 2002-04-29  Gerhard Tonn  <GerhardTonn@swol.de>
3372
3373         * Makefile.am: Add Linux for S/390 support.
3374         * Makefile.in: Regenerate.
3375         * configure.in: Add Linux for S/390 support.
3376         * configure: Regenerate.
3377         * include/ffi.h.in: Add Linux for S/390 support.
3378         * src/s390/ffi.c: New file from libffi CVS tree.
3379         * src/s390/sysv.S: New file from libffi CVS tree.
3380
3381 2002-04-28  Jakub Jelinek  <jakub@redhat.com>
3382
3383         * configure.in (HAVE_AS_SPARC_UA_PCREL): Check for working
3384         %r_disp32().
3385         * src/sparc/v8.S: Use it.
3386         * src/sparc/v9.S: Likewise.
3387         * fficonfig.h.in: Rebuilt.
3388         * configure: Rebuilt.
3389
3390 2002-04-08  Hans Boehm  <Hans_Boehm@hp.com>
3391
3392         * src/java_raw_api.c (ffi_java_raw_size): Handle FFI_TYPE_DOUBLE
3393         correctly.
3394         * src/ia64/unix.S: Add unwind information. Fix comments.
3395         Save sp in a way that's compatible with unwind info.
3396         (ffi_call_unix): Correctly restore sp in all cases.
3397         * src/ia64/ffi.c: Add, fix comments.
3398
3399 2002-04-08  Jakub Jelinek  <jakub@redhat.com>
3400
3401         * src/sparc/v8.S: Make .eh_frame dependent on target word size.
3402
3403 2002-04-06  Jason Thorpe  <thorpej@wasabisystems.com>
3404
3405         * configure.in (alpha*-*-netbsd*): Add target.
3406         * configure: Regenerate.
3407
3408 2002-04-04  Jeff Sturm  <jsturm@one-point.com>
3409
3410         * src/sparc/v8.S: Add unwind info.
3411         * src/sparc/v9.S: Likewise.
3412
3413 2002-03-30  Krister Walfridsson  <cato@df.lth.se>
3414
3415         * configure.in: Enable i*86-*-netbsdelf*.
3416         * configure: Rebuilt.
3417
3418 2002-03-29  David Billinghurst <David.Billinghurst@riotinto.com>
3419
3420         PR other/2620
3421         * src/mips/n32.s: Delete
3422         * src/mips/o32.s: Delete
3423
3424 2002-03-21  Loren J. Rittle  <ljrittle@acm.org>
3425
3426         * configure.in: Enable alpha*-*-freebsd*.
3427         * configure: Rebuilt.
3428
3429 2002-03-17  Bryce McKinlay  <bryce@waitaki.otago.ac.nz>
3430
3431         * Makefile.am: libfficonvenience -> libffi_convenience.
3432         * Makefile.in: Rebuilt.
3433
3434         * Makefile.am: Define ffitest_OBJECTS.
3435         * Makefile.in: Rebuilt.
3436
3437 2002-03-07  Andreas Tobler  <toa@pop.agri.ch>
3438             David Edelsohn  <edelsohn@gnu.org>
3439