OSDN Git Service

* src/powerpc/ffi_darwin.c (ffi_closure_helper_DARWIN): Increment
[pf3gnuchains/gcc-fork.git] / libffi / ChangeLog
1 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
2
3         * src/powerpc/ffi_darwin.c (ffi_closure_helper_DARWIN): Increment
4         pfr for long double split between fpr13 and stack.
5
6 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
7
8         * src/powerpc/ffi_darwin.c (ffi_prep_args): Increment next_arg and
9         fparg_count twice for long double.
10
11 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
12
13         PR libffi/42243
14         * src/powerpc/ffi_darwin.c (ffi_prep_args): Remove extra parentheses.
15
16 2009-12-03  Uros Bizjak  <ubizjak@gmail.com>
17
18         * testsuite/libffi.call/cls_longdouble_va.c (main): Fix format string.
19         Remove xfails for x86 linux targets.
20
21 2009-12-02  David Edelsohn  <edelsohn@gnu.org>
22
23         * src/powerpc/ffi_darwin.c (ffi_prep_args): Fix typo in INT64
24         case.
25
26 2009-12-01  David Edelsohn  <edelsohn@gnu.org>
27
28         * src/powerpc/aix.S (ffi_call_AIX): Convert to more standard
29         register usage.  Call ffi_prep_args directly.  Add long double
30         return value support.
31         * src/powerpc/ffi_darwin.c (ffi_prep_args): Double arg increment
32         applies to FFI_TYPE_DOUBLE.  Correct fpr_base increment typo.
33         Separate FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases.
34         (ffi_prep_cif_machdep): Only 16 byte stack alignment in 64 bit
35         mode.
36         (ffi_closure_helper_DARWIN): Remove nf and ng counters.  Move temp
37         into case.
38         * src/powerpc/aix_closure.S: Maintain 16 byte stack alignment.
39         Allocate result area between params and FPRs.
40
41 2009-11-30  David Edelsohn  <edelsohn@gnu.org>
42
43         PR target/35484
44         * src/powerpc/ffitarget.h (POWERPC64): Define for PPC64 Linux and
45         AIX64.
46         * src/powerpc/aix.S: Implement AIX64 version.
47         * src/powerpc/aix_closure.S: Implement AIX64 version.
48         (ffi_closure_ASM): Use extsb, lha and displament addresses.
49         * src/powerpc/ffi_darwin.c (ffi_prep_args): Implement AIX64
50         support.
51         (ffi_prep_cif_machdep): Same.
52         (ffi_call): Same.
53         (ffi_closure_helper_DARWIN): Same.
54
55 2009-11-02  Andreas Tobler  <a.tobler@schweiz.org>
56
57         PR libffi/41908
58         * testsuite/libffi.call/testclosure.c: New test.
59
60 2009-09-28  Kai Tietz  <kai.tietz@onevision.com>
61
62         * src/x86/win64.S (_ffi_call_win64 stack): Remove for gnu
63         assembly version use of ___chkstk.
64
65 2009-09-23  Matthias Klose  <doko@ubuntu.com>
66
67         PR libffi/40242, PR libffi/41443
68         * src/arm/sysv.S (__ARM_ARCH__): Define for processors
69         __ARM_ARCH_6T2__, __ARM_ARCH_6M__, __ARM_ARCH_7__,
70         __ARM_ARCH_7A__, __ARM_ARCH_7R__, __ARM_ARCH_7M__.
71         Change the conditionals to __SOFTFP__ || __ARM_EABI__
72         for -mfloat-abi=softfp to work.
73
74 2009-09-17  Loren J. Rittle  <ljrittle@acm.org>
75
76         PR testsuite/32843 (strikes again)
77         * src/x86/ffi.c (ffi_prep_cif_machdep): Add X86_FREEBSD to
78         enable proper extension on char and short.
79
80 2009-09-15  David Daney  <ddaney@caviumnetworks.com>
81
82         * src/java_raw_api.c (ffi_java_raw_to_rvalue): Remove special
83         handling for FFI_TYPE_POINTER.
84         * src/mips/ffitarget.h (FFI_TYPE_STRUCT_D_SOFT,
85         FFI_TYPE_STRUCT_F_SOFT, FFI_TYPE_STRUCT_DD_SOFT,
86         FFI_TYPE_STRUCT_FF_SOFT, FFI_TYPE_STRUCT_FD_SOFT,
87         FFI_TYPE_STRUCT_DF_SOFT, FFI_TYPE_STRUCT_SOFT): New defines.
88         (FFI_N32_SOFT_FLOAT, FFI_N64_SOFT_FLOAT): New ffi_abi enumerations.
89         (enum ffi_abi): Set FFI_DEFAULT_ABI for soft-float.
90         * src/mips/n32.S (ffi_call_N32): Add handling for soft-float
91         structure and pointer returns.
92         (ffi_closure_N32): Add handling for pointer returns.
93         * src/mips/ffi.c (ffi_prep_args, calc_n32_struct_flags,
94         calc_n32_return_struct_flags): Handle soft-float.
95         (ffi_prep_cif_machdep):  Handle soft-float, fix pointer handling.
96         (ffi_call_N32): Declare proper argument types.
97         (ffi_call, copy_struct_N32, ffi_closure_mips_inner_N32): Handle
98         soft-float.
99
100 2009-08-24  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
101
102         * configure.ac (AC_PREREQ): Bump to 2.64.
103
104 2009-08-22  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
105
106         * Makefile.am (install-html, install-pdf): Remove.
107         * Makefile.in: Regenerate.
108
109         * Makefile.in: Regenerate.
110         * aclocal.m4: Regenerate.
111         * configure: Regenerate.
112         * fficonfig.h.in: Regenerate.
113         * include/Makefile.in: Regenerate.
114         * man/Makefile.in: Regenerate.
115         * testsuite/Makefile.in: Regenerate.
116
117 2009-07-30  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
118
119         * configure.ac (_AC_ARG_VAR_PRECIOUS): Use m4_rename_force.
120
121 2009-07-24  Dave Korn  <dave.korn.cygwin@gmail.com>
122
123         PR libffi/40807
124         * src/x86/ffi.c (ffi_prep_cif_machdep): Also use sign/zero-extending
125         return types for X86_WIN32.
126         * src/x86/win32.S (_ffi_call_SYSV): Handle omitted return types.
127         (_ffi_call_STDCALL, _ffi_closure_SYSV, _ffi_closure_raw_SYSV,
128         _ffi_closure_STDCALL): Likewise.
129
130         * src/closures.c (is_selinux_enabled): Define to const 0 for Cygwin.
131         (dlmmap, dlmunmap): Also use these functions on Cygwin.
132
133 2009-07-11  Richard Sandiford  <rdsandiford@googlemail.com>
134
135         PR testsuite/40699
136         PR testsuite/40707
137         PR testsuite/40709
138         * testsuite/lib/libffi-dg.exp: Revert 2009-07-02, 2009-07-01 and
139         2009-06-30 commits.
140
141 2009-07-01  Richard Sandiford  <r.sandiford@uk.ibm.com>
142
143         * testsuite/lib/libffi-dg.exp (libffi-init): Set ld_library_path
144         to "" before adding paths.  (This reinstates an assignment that
145         was removed by my 2009-06-30 commit, but changes the initial
146         value from "." to "".)
147
148 2009-07-01  H.J. Lu  <hongjiu.lu@intel.com>
149
150         PR testsuite/40601
151         * testsuite/lib/libffi-dg.exp (libffi-init): Properly set
152         gccdir.  Adjust ld_library_path for gcc only if gccdir isn't
153         empty.
154
155 2009-06-30  Richard Sandiford  <r.sandiford@uk.ibm.com>
156
157         * testsuite/lib/libffi-dg.exp (libffi-init): Don't add "."
158         to ld_library_path.  Use add_path.  Add just find_libgcc_s
159         to ld_library_path, not every libgcc multilib directory.
160
161 2009-06-16  Wim Lewis  <wiml@hhhh.org>
162
163         * src/powerpc/ffi.c: Avoid clobbering cr3 and cr4, which are
164         supposed to be callee-saved.
165         * src/powerpc/sysv.S (small_struct_return_value): Fix overrun of
166         return buffer for odd-size structs.
167
168 2009-06-16  Andreas Tobler  <a.tobler@schweiz.org>
169
170         PR libffi/40444
171         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Add
172         allow_stack_execute for Darwin.
173
174 2009-06-16  Andrew Haley  <aph@redhat.com>
175
176         * configure.ac (TARGETDIR): Add missing blank lines.
177         * configure: Regenerate.
178
179 2009-06-16  Andrew Haley  <aph@redhat.com>
180
181         * testsuite/libffi.call/cls_align_sint64.c,
182         testsuite/libffi.call/cls_align_uint64.c,
183         testsuite/libffi.call/cls_longdouble_va.c,
184         testsuite/libffi.call/cls_ulonglong.c,
185         testsuite/libffi.call/return_ll1.c,
186         testsuite/libffi.call/stret_medium2.c: Fix printf format
187         specifiers.
188         * testsuite/libffi.call/ffitest.h,
189         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRIuLL): Define.
190
191 2009-06-15  Andrew Haley  <aph@redhat.com>
192
193         * testsuite/libffi.call/err_bad_typedef.c: xfail everywhere.
194         * testsuite/libffi.call/err_bad_abi.c: Likewise.
195         
196 2009-06-12  Andrew Haley  <aph@redhat.com>
197
198         * Makefile.am: Remove info_TEXINFOS.
199
200 2009-06-12  Andrew Haley  <aph@redhat.com>
201
202         * ChangeLog.libffi: testsuite/libffi.call/cls_align_sint64.c,
203         testsuite/libffi.call/cls_align_uint64.c,
204         testsuite/libffi.call/cls_ulonglong.c,
205         testsuite/libffi.call/return_ll1.c,
206         testsuite/libffi.call/stret_medium2.c: Fix printf format
207         specifiers.
208         testsuite/libffi.special/unwindtest.cc: include stdint.h.
209
210 2009-06-11  Timothy Wall  <twall@users.sf.net>
211
212         * Makefile.am,
213         configure.ac,
214         include/ffi.h.in,
215         include/ffi_common.h,
216         src/closures.c,
217         src/dlmalloc.c,
218         src/x86/ffi.c,
219         src/x86/ffitarget.h,
220         src/x86/win64.S (new),
221         README: Added win64 support (mingw or MSVC)
222         * Makefile.in,
223         include/Makefile.in,
224         man/Makefile.in,
225         testsuite/Makefile.in,
226         configure,
227         aclocal.m4: Regenerated
228         * ltcf-c.sh: properly escape cygwin/w32 path
229         * man/ffi_call.3: Clarify size requirements for return value.
230         * src/x86/ffi64.c: Fix filename in comment.
231         * src/x86/win32.S: Remove unused extern.
232
233         * testsuite/libffi.call/closure_fn0.c,
234         testsuite/libffi.call/closure_fn1.c,
235         testsuite/libffi.call/closure_fn2.c,
236         testsuite/libffi.call/closure_fn3.c,
237         testsuite/libffi.call/closure_fn4.c,
238         testsuite/libffi.call/closure_fn5.c,
239         testsuite/libffi.call/closure_fn6.c,
240         testsuite/libffi.call/closure_stdcall.c,
241         testsuite/libffi.call/cls_12byte.c,
242         testsuite/libffi.call/cls_16byte.c,
243         testsuite/libffi.call/cls_18byte.c,
244         testsuite/libffi.call/cls_19byte.c,
245         testsuite/libffi.call/cls_1_1byte.c,
246         testsuite/libffi.call/cls_20byte.c,
247         testsuite/libffi.call/cls_20byte1.c,
248         testsuite/libffi.call/cls_24byte.c,
249         testsuite/libffi.call/cls_2byte.c,
250         testsuite/libffi.call/cls_3_1byte.c,
251         testsuite/libffi.call/cls_3byte1.c,
252         testsuite/libffi.call/cls_3byte2.c,
253         testsuite/libffi.call/cls_4_1byte.c,
254         testsuite/libffi.call/cls_4byte.c,
255         testsuite/libffi.call/cls_5_1_byte.c,
256         testsuite/libffi.call/cls_5byte.c,
257         testsuite/libffi.call/cls_64byte.c,
258         testsuite/libffi.call/cls_6_1_byte.c,
259         testsuite/libffi.call/cls_6byte.c,
260         testsuite/libffi.call/cls_7_1_byte.c,
261         testsuite/libffi.call/cls_7byte.c,
262         testsuite/libffi.call/cls_8byte.c,
263         testsuite/libffi.call/cls_9byte1.c,
264         testsuite/libffi.call/cls_9byte2.c,
265         testsuite/libffi.call/cls_align_double.c,
266         testsuite/libffi.call/cls_align_float.c,
267         testsuite/libffi.call/cls_align_longdouble.c,
268         testsuite/libffi.call/cls_align_longdouble_split.c,
269         testsuite/libffi.call/cls_align_longdouble_split2.c,
270         testsuite/libffi.call/cls_align_pointer.c,
271         testsuite/libffi.call/cls_align_sint16.c,
272         testsuite/libffi.call/cls_align_sint32.c,
273         testsuite/libffi.call/cls_align_sint64.c,
274         testsuite/libffi.call/cls_align_uint16.c,
275         testsuite/libffi.call/cls_align_uint32.c,
276         testsuite/libffi.call/cls_align_uint64.c,
277         testsuite/libffi.call/cls_dbls_struct.c,
278         testsuite/libffi.call/cls_double.c,
279         testsuite/libffi.call/cls_double_va.c,
280         testsuite/libffi.call/cls_float.c,
281         testsuite/libffi.call/cls_longdouble.c,
282         testsuite/libffi.call/cls_longdouble_va.c,
283         testsuite/libffi.call/cls_multi_schar.c,
284         testsuite/libffi.call/cls_multi_sshort.c,
285         testsuite/libffi.call/cls_multi_sshortchar.c,
286         testsuite/libffi.call/cls_multi_uchar.c,
287         testsuite/libffi.call/cls_multi_ushort.c,
288         testsuite/libffi.call/cls_multi_ushortchar.c,
289         testsuite/libffi.call/cls_pointer.c,
290         testsuite/libffi.call/cls_pointer_stack.c,
291         testsuite/libffi.call/cls_schar.c,
292         testsuite/libffi.call/cls_sint.c,
293         testsuite/libffi.call/cls_sshort.c,
294         testsuite/libffi.call/cls_uchar.c,
295         testsuite/libffi.call/cls_uint.c,
296         testsuite/libffi.call/cls_ulonglong.c,
297         testsuite/libffi.call/cls_ushort.c,
298         testsuite/libffi.call/err_bad_abi.c,
299         testsuite/libffi.call/err_bad_typedef.c,
300         testsuite/libffi.call/float2.c,
301         testsuite/libffi.call/huge_struct.c,
302         testsuite/libffi.call/nested_struct.c,
303         testsuite/libffi.call/nested_struct1.c,
304         testsuite/libffi.call/nested_struct10.c,
305         testsuite/libffi.call/nested_struct2.c,
306         testsuite/libffi.call/nested_struct3.c,
307         testsuite/libffi.call/nested_struct4.c,
308         testsuite/libffi.call/nested_struct5.c,
309         testsuite/libffi.call/nested_struct6.c,
310         testsuite/libffi.call/nested_struct7.c,
311         testsuite/libffi.call/nested_struct8.c,
312         testsuite/libffi.call/nested_struct9.c,
313         testsuite/libffi.call/problem1.c,
314         testsuite/libffi.call/return_ldl.c,
315         testsuite/libffi.call/return_ll1.c,
316         testsuite/libffi.call/stret_large.c,
317         testsuite/libffi.call/stret_large2.c,
318         testsuite/libffi.call/stret_medium.c,
319         testsuite/libffi.call/stret_medium2.c,
320         testsuite/libffi.special/unwindtest.cc: use ffi_closure_alloc instead
321         of checking for MMAP.  Use intptr_t instead of long casts.
322
323 2009-06-11  Kaz Kojima  <kkojima@gcc.gnu.org>
324
325         * testsuite/libffi.call/cls_longdouble_va.c: Add xfail sh*-*-linux-*.
326         * testsuite/libffi.call/err_bad_abi.c: Add xfail sh*-*-*.
327         * testsuite/libffi.call/err_bad_typedef.c: Likewise.
328
329 2009-06-09  Andrew Haley  <aph@redhat.com>
330
331         * src/x86/freebsd.S: Add missing file.
332
333 2009-06-08  Andrew Haley  <aph@redhat.com>
334
335         Import from libffi 3.0.8:
336
337         * doc/libffi.texi: New file.
338         * doc/libffi.info: Likewise.
339         * doc/stamp-vti: Likewise.
340         * man/Makefile.am: New file.
341         * man/ffi_call.3: New file.
342
343         * Makefile.am (EXTRA_DIST): Add src/x86/darwin64.S,
344         src/dlmalloc.c.
345         (nodist_libffi_la_SOURCES): Add X86_FREEBSD.
346
347         * configure.ac: Bump version to 3.0.8.
348         parisc*-*-linux*: Add.
349         i386-*-freebsd* | i386-*-openbsd*: Add.
350         powerpc-*-beos*: Add.
351         AM_CONDITIONAL X86_FREEBSD: Add.
352         AC_CONFIG_FILES: Add man/Makefile.
353
354         * include/ffi.h.in (FFI_FN): Change void (*)() to void (*)(void).
355
356 2009-06-08  Andrew Haley  <aph@redhat.com>
357
358         * README: Import from libffi 3.0.8.
359
360 2009-06-08  Andrew Haley  <aph@redhat.com>
361
362         * testsuite/libffi.call/err_bad_abi.c: Add xfails.
363         * testsuite/libffi.call/cls_longdouble_va.c: Add xfails.
364         * testsuite/libffi.call/cls_dbls_struct.c: Add xfail x86_64-*-linux-*.
365         * testsuite/libffi.call/err_bad_typedef.c: Add xfails.
366
367         * testsuite/libffi.call/stret_medium2.c: Add __UNUSED__ to args.
368         * testsuite/libffi.call/stret_medium.c: Likewise.
369         * testsuite/libffi.call/stret_large2.c: Likewise.
370         * testsuite/libffi.call/stret_large.c:  Likewise.
371
372 2008-12-26  Timothy Wall  <twall@users.sf.net>
373
374         * testsuite/libffi.call/cls_longdouble.c,
375         testsuite/libffi.call/cls_longdouble_va.c,
376         testsuite/libffi.call/cls_align_longdouble.c,
377         testsuite/libffi.call/cls_align_longdouble_split.c,
378         testsuite/libffi.call/cls_align_longdouble_split2.c: mark expected
379         failures on x86_64 cygwin/mingw.
380
381 2008-12-22  Timothy Wall  <twall@users.sf.net>
382
383         * testsuite/libffi.call/closure_fn0.c,
384         testsuite/libffi.call/closure_fn1.c,
385         testsuite/libffi.call/closure_fn2.c,
386         testsuite/libffi.call/closure_fn3.c,
387         testsuite/libffi.call/closure_fn4.c,
388         testsuite/libffi.call/closure_fn5.c,
389         testsuite/libffi.call/closure_fn6.c,
390         testsuite/libffi.call/closure_loc_fn0.c,
391         testsuite/libffi.call/closure_stdcall.c,
392         testsuite/libffi.call/cls_align_pointer.c,
393         testsuite/libffi.call/cls_pointer.c,
394         testsuite/libffi.call/cls_pointer_stack.c: use portable cast from
395         pointer to integer (intptr_t).
396         * testsuite/libffi.call/cls_longdouble.c: disable for win64.
397
398 2008-07-24  Anthony Green  <green@redhat.com>
399
400         * testsuite/libffi.call/cls_dbls_struct.c,
401         testsuite/libffi.call/cls_double_va.c,
402         testsuite/libffi.call/cls_longdouble.c,
403         testsuite/libffi.call/cls_longdouble_va.c,
404         testsuite/libffi.call/cls_pointer.c,
405         testsuite/libffi.call/cls_pointer_stack.c,
406         testsuite/libffi.call/err_bad_abi.c: Clean up failures from
407         compiler warnings.
408
409 2008-03-04  Anthony Green  <green@redhat.com>
410             Blake Chaffin
411             hos@tamanegi.org
412
413         * testsuite/libffi.call/cls_align_longdouble_split2.c
414           testsuite/libffi.call/cls_align_longdouble_split.c
415           testsuite/libffi.call/cls_dbls_struct.c
416           testsuite/libffi.call/cls_double_va.c
417           testsuite/libffi.call/cls_longdouble.c
418           testsuite/libffi.call/cls_longdouble_va.c
419           testsuite/libffi.call/cls_pointer.c
420           testsuite/libffi.call/cls_pointer_stack.c
421           testsuite/libffi.call/err_bad_abi.c
422           testsuite/libffi.call/err_bad_typedef.c
423           testsuite/libffi.call/stret_large2.c
424           testsuite/libffi.call/stret_large.c
425           testsuite/libffi.call/stret_medium2.c
426           testsuite/libffi.call/stret_medium.c: New tests from Apple.
427
428 2009-06-05  Andrew Haley  <aph@redhat.com>
429
430         * src/x86/ffitarget.h, src/x86/ffi.c: Merge stdcall changes from
431         libffi.
432
433 2009-06-04  Andrew Haley  <aph@redhat.com>
434
435         * src/x86/ffitarget.h, src/x86/win32.S, src/x86/ffi.c: Back out
436         stdcall changes.
437
438 2008-02-26  Anthony Green  <green@redhat.com>
439             Thomas Heller  <theller@ctypes.org>
440
441         * src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C
442         comment.
443
444 2008-02-03  Timothy Wall  <twall@users.sf.net>
445
446         * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return
447           offset based on code pointer, not data pointer.
448
449 2008-01-31  Timothy Wall <twall@users.sf.net>
450
451         * testsuite/libffi.call/closure_stdcall.c: Add test for stdcall
452         closures.
453         * src/x86/ffitarget.h: Increase size of trampoline for stdcall
454         closures.
455         * src/x86/win32.S: Add assembly for stdcall closure.
456         * src/x86/ffi.c: Initialize stdcall closure trampoline.
457
458 2009-06-04  Andrew Haley  <aph@redhat.com>
459
460         * include/ffi.h.in: Change void (*)() to void (*)(void).
461         * src/x86/ffi.c: Likewise.
462
463 2009-06-04  Andrew Haley  <aph@redhat.com>
464
465         * src/powerpc/ppc_closure.S: Insert licence header.
466         * src/powerpc/linux64_closure.S: Likewise.
467         * src/m68k/sysv.S: Likewise.
468
469         * src/sh64/ffi.c: Change void (*)() to void (*)(void).
470         * src/powerpc/ffi.c: Likewise.
471         * src/powerpc/ffi_darwin.c: Likewise.
472         * src/m32r/ffi.c: Likewise.
473         * src/sh64/ffi.c: Likewise.
474         * src/x86/ffi64.c: Likewise.
475         * src/alpha/ffi.c: Likewise.
476         * src/alpha/osf.S: Likewise.
477         * src/frv/ffi.c: Likewise.
478         * src/s390/ffi.c: Likewise.
479         * src/pa/ffi.c: Likewise.
480         * src/pa/hpux32.S: Likewise.
481         * src/ia64/unix.S: Likewise.
482         * src/ia64/ffi.c: Likewise.
483         * src/sparc/ffi.c: Likewise.
484         * src/mips/ffi.c: Likewise.
485         * src/sh/ffi.c: Likewise.
486
487 2008-02-15  David Daney  <ddaney@avtrex.com>
488
489         * src/mips/ffi.c (USE__BUILTIN___CLEAR_CACHE):
490         Define (conditionally), and use it to include cachectl.h.
491         (ffi_prep_closure_loc): Fix cache flushing.
492         * src/mips/ffitarget.h (_ABIN32, _ABI64, _ABIO32): Define.
493
494 2009-06-04  Andrew Haley  <aph@redhat.com>
495
496         include/ffi.h.in,
497         src/arm/ffitarget.h,
498         src/arm/ffi.c,
499         src/arm/sysv.S,
500         src/powerpc/ffitarget.h,
501         src/closures.c,
502         src/sh64/ffitarget.h,
503         src/sh64/ffi.c,
504         src/sh64/sysv.S,
505         src/types.c,
506         src/x86/ffi64.c,
507         src/x86/ffitarget.h,
508         src/x86/win32.S,
509         src/x86/darwin.S,
510         src/x86/ffi.c,
511         src/x86/sysv.S,
512         src/x86/unix64.S,
513         src/alpha/ffitarget.h,
514         src/alpha/ffi.c,
515         src/alpha/osf.S,
516         src/m68k/ffitarget.h,
517         src/frv/ffitarget.h,
518         src/frv/ffi.c,
519         src/s390/ffitarget.h,
520         src/s390/sysv.S,
521         src/cris/ffitarget.h,
522         src/pa/linux.S,
523         src/pa/ffitarget.h,
524         src/pa/ffi.c,
525         src/raw_api.c,
526         src/ia64/ffitarget.h,
527         src/ia64/unix.S,
528         src/ia64/ffi.c,
529         src/ia64/ia64_flags.h,
530         src/java_raw_api.c,
531         src/debug.c,
532         src/sparc/v9.S,
533         src/sparc/ffitarget.h,
534         src/sparc/ffi.c,
535         src/sparc/v8.S,
536         src/mips/ffitarget.h,
537         src/mips/n32.S,
538         src/mips/o32.S,
539         src/mips/ffi.c,
540         src/prep_cif.c,
541         src/sh/ffitarget.h,
542         src/sh/ffi.c,
543         src/sh/sysv.S: Update license text.
544
545 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
546
547         * src/x86/win32.S (_ffi_closure_STDCALL):  New function.
548         (.eh_frame):  Add FDE for it.
549
550 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
551
552         * configure.ac:  Also check if assembler supports pc-relative
553         relocs on X86_WIN32 targets.
554         * configure:  Regenerate.
555         * src/x86/win32.S (ffi_prep_args):  Declare extern, not global.
556         (_ffi_call_SYSV):  Add missing function type symbol .def and
557         add EH markup labels.
558         (_ffi_call_STDCALL):  Likewise.
559         (_ffi_closure_SYSV):  Likewise.
560         (_ffi_closure_raw_SYSV):  Likewise.
561         (.eh_frame):  Add hand-crafted EH data.
562
563 2009-04-09  Jakub Jelinek  <jakub@redhat.com>
564
565         * testsuite/lib/libffi-dg.exp: Change copyright header to refer to
566         version 3 of the GNU General Public License and to point readers
567         at the COPYING3 file and the FSF's license web page.
568         * testsuite/libffi.call/call.exp: Likewise.
569         * testsuite/libffi.special/special.exp: Likewise.
570
571 2009-03-01  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
572
573         * configure: Regenerate.
574
575 2008-12-18  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
576
577         PR libffi/26048
578         * configure.ac (HAVE_AS_X86_PCREL): New test.
579         * configure: Regenerate.
580         * fficonfig.h.in: Regenerate.
581         * src/x86/sysv.S [!FFI_NO_RAW_API]: Precalculate
582         RAW_CLOSURE_CIF_OFFSET, RAW_CLOSURE_FUN_OFFSET,
583         RAW_CLOSURE_USER_DATA_OFFSET for the Solaris 10/x86 assembler.
584         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
585         * src/x86/unix64.S (.Lstore_table): Move to .text section.
586         (.Lload_table): Likewise.
587         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
588
589 2008-12-18  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
590
591         * configure: Regenerate.
592
593 2008-11-21  Eric Botcazou  <ebotcazou@adacore.com>
594
595         * src/sparc/ffi.c (ffi_prep_cif_machdep): Add support for
596         signed/unsigned int8/16 return values.
597         * src/sparc/v8.S (ffi_call_v8): Likewise.
598         (ffi_closure_v8): Likewise.
599
600 2008-09-26  Peter O'Gorman  <pogma@thewrittenword.com>
601             Steve Ellcey  <sje@cup.hp.com>
602
603         * configure: Regenerate for new libtool.
604         * Makefile.in: Ditto.
605         * include/Makefile.in: Ditto.
606         * aclocal.m4: Ditto.
607
608 2008-08-25  Andreas Tobler  <a.tobler@schweiz.org>
609
610         * src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and
611         FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum.
612         Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT.
613         Adjust copyright notice.
614         * src/powerpc/ffi.c: Add two new flags to indicate if we have one
615         register or two register to use for FFI_SYSV structs.
616         (ffi_prep_cif_machdep): Pass the right register flag introduced above.
617         (ffi_closure_helper_SYSV): Fix the return type for
618         FFI_SYSV_TYPE_SMALL_STRUCT. Comment.
619         Adjust copyright notice.
620
621 2008-07-16  Kaz Kojima  <kkojima@gcc.gnu.org>
622
623         * src/sh/ffi.c (ffi_prep_closure_loc): Turn INSN into an unsigned
624         int.
625
626 2008-06-17  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
627
628         * configure: Regenerate.
629         * include/Makefile.in: Regenerate.
630         * testsuite/Makefile.in: Regenerate.
631
632 2008-06-07  Joseph Myers  <joseph@codesourcery.com>
633
634         * configure.ac (parisc*-*-linux*, powerpc-*-sysv*,
635         powerpc-*-beos*): Remove.
636         * configure: Regenerate.
637
638 2008-05-09  Julian Brown  <julian@codesourcery.com>
639
640         * Makefile.am (LTLDFLAGS): New.
641         (libffi_la_LDFLAGS): Use above.
642         * Makefile.in: Regenerate.
643
644 2008-04-18  Paolo Bonzini  <bonzini@gnu.org>
645
646         PR bootstrap/35457
647         * aclocal.m4: Regenerate.
648         * configure: Regenerate.
649
650 2008-03-26  Kaz Kojima  <kkojima@gcc.gnu.org>
651
652         * src/sh/sysv.S: Add .note.GNU-stack on Linux.
653         * src/sh64/sysv.S: Likewise.
654
655 2008-03-26  Daniel Jacobowitz  <dan@debian.org>
656
657         * src/arm/sysv.S: Fix ARM comment marker.
658
659 2008-03-26  Jakub Jelinek  <jakub@redhat.com>
660
661         * src/alpha/osf.S: Add .note.GNU-stack on Linux.
662         * src/s390/sysv.S: Likewise.
663         * src/powerpc/ppc_closure.S: Likewise.
664         * src/powerpc/sysv.S: Likewise.
665         * src/x86/unix64.S: Likewise.
666         * src/x86/sysv.S: Likewise.
667         * src/sparc/v8.S: Likewise.
668         * src/sparc/v9.S: Likewise.
669         * src/m68k/sysv.S: Likewise.
670         * src/arm/sysv.S: Likewise.
671
672 2008-03-16  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
673
674         * aclocal.m4: Regenerate.
675         * configure: Likewise.
676         * Makefile.in: Likewise.
677         * include/Makefile.in: Likewise.
678         * testsuite/Makefile.in: Likewise.
679
680 2008-02-12  Bjoern Koenig  <bkoenig@alpha-tierchen.de>
681             Andreas Tobler  <a.tobler@schweiz.org>
682
683         * configure.ac: Add amd64-*-freebsd* target.
684         * configure: Regenerate.
685
686 2008-01-30  H.J. Lu  <hongjiu.lu@intel.com>
687
688         PR libffi/34612
689         * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when
690         returning struct.
691
692         * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer"
693         tests.
694
695 2008-01-24  David Edelsohn  <edelsohn@gnu.org>
696
697         * configure: Regenerate.
698
699 2008-01-06  Andreas Tobler  <a.tobler@schweiz.org>
700
701         * src/x86/ffi.c (ffi_prep_cif_machdep): Fix thinko.
702
703 2008-01-05  Andreas Tobler  <a.tobler@schweiz.org>
704
705         PR testsuite/32843
706         * src/x86/ffi.c (ffi_prep_cif_machdep): Add code for
707         signed/unsigned int8/16 for X86_DARWIN.
708         Updated copyright info.
709         Handle one and two byte structs with special cif->flags.
710         * src/x86/ffitarget.h: Add special types for one and two byte structs.
711         Updated copyright info.
712         * src/x86/darwin.S (ffi_call_SYSV): Rewrite to use a jump table like
713         sysv.S
714         Remove code to pop args from the stack after call.
715         Special-case signed/unsigned for int8/16, one and two byte structs.
716         (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
717         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
718         FFI_TYPE_SINT32.
719         Updated copyright info.
720
721 2007-12-08  David Daney  <ddaney@avtrex.com>
722
723         * src/mips/n32.S (ffi_call_N32):  Replace dadd with ADDU, dsub with 
724         SUBU, add with ADDU and use smaller code sequences.
725
726 2007-12-07  David Daney  <ddaney@avtrex.com>
727
728         * src/mips/ffi.c (ffi_prep_cif_machdep): Handle long double return
729         type.
730
731 2007-12-06  David Daney  <ddaney@avtrex.com>
732
733         * include/ffi.h.in (FFI_SIZEOF_JAVA_RAW): Define if not already
734         defined.
735         (ffi_java_raw): New typedef.
736         (ffi_java_raw_call, ffi_java_ptrarray_to_raw,
737         ffi_java_raw_to_ptrarray): Change parameter types from ffi_raw to
738         ffi_java_raw.
739         (ffi_java_raw_closure) : Same.
740         (ffi_prep_java_raw_closure, ffi_prep_java_raw_closure_loc): Change
741         parameter types.
742         * src/java_raw_api.c (ffi_java_raw_size):  Replace FFI_SIZEOF_ARG with
743         FFI_SIZEOF_JAVA_RAW.
744         (ffi_java_raw_to_ptrarray): Change type of raw to ffi_java_raw.
745         Replace FFI_SIZEOF_ARG with FFI_SIZEOF_JAVA_RAW. Use
746         sizeof(ffi_java_raw) for alignment calculations.
747         (ffi_java_ptrarray_to_raw): Same.
748         (ffi_java_rvalue_to_raw): Add special handling for FFI_TYPE_POINTER
749         if FFI_SIZEOF_JAVA_RAW == 4.
750         (ffi_java_raw_to_rvalue): Same.
751         (ffi_java_raw_call): Change type of raw to ffi_java_raw.
752         (ffi_java_translate_args): Same.
753         (ffi_prep_java_raw_closure_loc, ffi_prep_java_raw_closure): Change
754         parameter types.
755         * src/mips/ffitarget.h (FFI_SIZEOF_JAVA_RAW): Define for N32 ABI.
756
757 2007-12-06  David Daney  <ddaney@avtrex.com>
758
759         * src/mips/n32.S (ffi_closure_N32): Use 64-bit add instruction on 
760         pointer values.
761
762 2007-12-01  Andreas Tobler  <a.tobler@schweiz.org>
763
764         PR libffi/31937
765         * src/powerpc/ffitarget.h: Introduce new ABI FFI_LINUX_SOFT_FLOAT.
766         Add local FFI_TYPE_UINT128 to handle soft-float long-double-128.
767         * src/powerpc/ffi.c: Distinguish between __NO_FPRS__ and not and
768         set the NUM_FPR_ARG_REGISTERS according to.
769         Add support for potential soft-float support under hard-float
770         architecture.
771         (ffi_prep_args_SYSV): Set NUM_FPR_ARG_REGISTERS to 0 in case of
772         FFI_LINUX_SOFT_FLOAT, handle float, doubles and long-doubles according
773         to the FFI_LINUX_SOFT_FLOAT ABI.
774         (ffi_prep_cif_machdep): Likewise.
775         (ffi_closure_helper_SYSV): Likewise.
776         * src/powerpc/ppc_closure.S: Make sure not to store float/double
777         on archs where __NO_FPRS__ is true.
778         Add FFI_TYPE_UINT128 support.
779         * src/powerpc/sysv.S: Add support for soft-float long-double-128.
780         Adjust copyright notice.
781
782 2007-11-25  Andreas Tobler  <a.tobler@schweiz.org>
783
784         * src/closures.c: Move defintion of MAYBE_UNUSED from here to ...
785         * include/ffi_common.h: ... here.
786         Update copyright.
787
788 2007-11-17  Andreas Tobler  <a.tobler@schweiz.org>
789
790         * src/powerpc/sysv.S: Load correct cr to compare if we have long double.
791         * src/powerpc/linux64.S: Likewise.
792         * src/powerpc/ffi.c: Add a comment to show which part goes into cr6.
793         * testsuite/libffi.call/return_ldl.c: New test.
794
795 2007-09-04    <aph@redhat.com>
796
797         * src/arm/sysv.S (UNWIND): New.
798         (Whole file): Conditionally compile unwinder directives.
799         * src/arm/sysv.S: Add unwinder directives.
800
801         * src/arm/ffi.c (ffi_prep_args): Align structs by at least 4 bytes.
802         Only treat r0 as a struct address if we're actually returning a
803         struct by address.
804         Only copy the bytes that are actually within a struct.
805         (ffi_prep_cif_machdep): A Composite Type not larger than 4 bytes
806         is returned in r0, not passed by address.
807         (ffi_call): Allocate a word-sized temporary for the case where
808         a composite is returned in r0.
809         (ffi_prep_incoming_args_SYSV): Align as necessary.
810
811 2007-08-05  Steven Newbury  <s_j_newbury@yahoo.co.uk>
812
813         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Use __clear_cache instead of 
814         directly using the sys_cacheflush syscall.
815
816 2007-07-27  Andrew Haley  <aph@redhat.com>
817
818         * src/arm/sysv.S (ffi_closure_SYSV): Add soft-float.
819
820 2007-09-03  Maciej W. Rozycki  <macro@linux-mips.org>
821
822         * Makefile.am: Unify MIPS_IRIX and MIPS_LINUX into MIPS.
823         * configure.ac: Likewise.
824         * Makefile.in: Regenerate.
825         * include/Makefile.in: Likewise.
826         * testsuite/Makefile.in: Likewise.
827         * configure: Likewise.
828
829 2007-08-24  David Daney  <ddaney@avtrex.com>
830
831         * testsuite/libffi.call/return_sl.c: New test.
832
833 2007-08-10  David Daney  <ddaney@avtrex.com>
834
835         * testsuite/libffi.call/cls_multi_ushort.c,
836         testsuite/libffi.call/cls_align_uint16.c,
837         testsuite/libffi.call/nested_struct1.c,
838         testsuite/libffi.call/nested_struct3.c,
839         testsuite/libffi.call/cls_7_1_byte.c,
840         testsuite/libffi.call/nested_struct5.c,
841         testsuite/libffi.call/cls_double.c,
842         testsuite/libffi.call/nested_struct7.c,
843         testsuite/libffi.call/cls_sint.c,
844         testsuite/libffi.call/nested_struct9.c,
845         testsuite/libffi.call/cls_20byte1.c,
846         testsuite/libffi.call/cls_multi_sshortchar.c,
847         testsuite/libffi.call/cls_align_sint64.c,
848         testsuite/libffi.call/cls_3byte2.c,
849         testsuite/libffi.call/cls_multi_schar.c,
850         testsuite/libffi.call/cls_multi_uchar.c,
851         testsuite/libffi.call/cls_19byte.c,
852         testsuite/libffi.call/cls_9byte1.c,
853         testsuite/libffi.call/cls_align_float.c,
854         testsuite/libffi.call/closure_fn1.c,
855         testsuite/libffi.call/problem1.c,
856         testsuite/libffi.call/closure_fn3.c,
857         testsuite/libffi.call/cls_sshort.c,
858         testsuite/libffi.call/closure_fn5.c,
859         testsuite/libffi.call/cls_align_double.c,
860         testsuite/libffi.call/nested_struct.c,
861         testsuite/libffi.call/cls_2byte.c,
862         testsuite/libffi.call/nested_struct10.c,
863         testsuite/libffi.call/cls_4byte.c,
864         testsuite/libffi.call/cls_6byte.c,
865         testsuite/libffi.call/cls_8byte.c,
866         testsuite/libffi.call/cls_multi_sshort.c,
867         testsuite/libffi.call/cls_align_sint16.c,
868         testsuite/libffi.call/cls_align_uint32.c,
869         testsuite/libffi.call/cls_20byte.c,
870         testsuite/libffi.call/cls_float.c,
871         testsuite/libffi.call/nested_struct2.c,
872         testsuite/libffi.call/cls_5_1_byte.c,
873         testsuite/libffi.call/nested_struct4.c,
874         testsuite/libffi.call/cls_24byte.c,
875         testsuite/libffi.call/nested_struct6.c,
876         testsuite/libffi.call/cls_64byte.c,
877         testsuite/libffi.call/nested_struct8.c,
878         testsuite/libffi.call/cls_uint.c,
879         testsuite/libffi.call/cls_multi_ushortchar.c,
880         testsuite/libffi.call/cls_schar.c,
881         testsuite/libffi.call/cls_uchar.c,
882         testsuite/libffi.call/cls_align_uint64.c,
883         testsuite/libffi.call/cls_ulonglong.c,
884         testsuite/libffi.call/cls_align_longdouble.c,
885         testsuite/libffi.call/cls_1_1byte.c,
886         testsuite/libffi.call/cls_12byte.c,
887         testsuite/libffi.call/cls_3_1byte.c,
888         testsuite/libffi.call/cls_3byte1.c,
889         testsuite/libffi.call/cls_4_1byte.c,
890         testsuite/libffi.call/cls_6_1_byte.c,
891         testsuite/libffi.call/cls_16byte.c,
892         testsuite/libffi.call/cls_18byte.c,
893         testsuite/libffi.call/closure_fn0.c,
894         testsuite/libffi.call/cls_9byte2.c,
895         testsuite/libffi.call/closure_fn2.c,
896         testsuite/libffi.call/closure_fn4.c,
897         testsuite/libffi.call/cls_ushort.c,
898         testsuite/libffi.call/closure_fn6.c,
899         testsuite/libffi.call/cls_5byte.c,
900         testsuite/libffi.call/cls_align_pointer.c,
901         testsuite/libffi.call/cls_7byte.c,
902         testsuite/libffi.call/cls_align_sint32.c,
903         testsuite/libffi.special/unwindtest_ffi_call.cc,
904         testsuite/libffi.special/unwindtest.cc: Remove xfail for mips64*-*-*.
905
906 2007-08-10  David Daney  <ddaney@avtrex.com>
907
908         PR libffi/28313
909         * configure.ac: Don't treat mips64 as a special case.
910         * Makefile.am (nodist_libffi_la_SOURCES): Add n32.S.
911         * configure: Regenerate
912         * Makefile.in: Ditto.
913         * fficonfig.h.in: Ditto.
914         * src/mips/ffitarget.h (REG_L, REG_S, SUBU, ADDU, SRL, LI): Indent.
915         (LA, EH_FRAME_ALIGN, FDE_ADDR_BYTES): New preprocessor macros.
916         (FFI_DEFAULT_ABI): Set for n64 case.
917         (FFI_CLOSURES, FFI_TRAMPOLINE_SIZE): Define for n32 and n64 cases.
918         * src/mips/n32.S (ffi_call_N32): Add debug macros and labels for FDE.
919         (ffi_closure_N32): New function.
920         (.eh_frame): New section
921         * src/mips/o32.S: Clean up comments.
922         (ffi_closure_O32): Pass ffi_closure parameter in $12.
923         * src/mips/ffi.c: Use FFI_MIPS_N32 instead of
924         _MIPS_SIM == _ABIN32 throughout.
925         (FFI_MIPS_STOP_HERE): New, use in place of
926         ffi_stop_here.
927         (ffi_prep_args): Use unsigned long to hold pointer values.  Rewrite
928         to support n32/n64 ABIs.
929         (calc_n32_struct_flags): Rewrite.
930         (calc_n32_return_struct_flags): Remove unused variable.  Reverse
931         position of flag bits.
932         (ffi_prep_cif_machdep): Rewrite n32 portion.
933         (ffi_call): Enable for n64.  Add special handling for small structure
934         return values.
935         (ffi_prep_closure_loc): Add n32 and n64 support.
936         (ffi_closure_mips_inner_O32): Add cast to silence warning.
937         (copy_struct_N32, ffi_closure_mips_inner_N32): New functions.
938
939 2007-08-08  David Daney  <ddaney@avtrex.com>
940
941         * testsuite/libffi.call/ffitest.h (ffi_type_mylong): Remove definition.
942         * testsuite/libffi.call/cls_align_uint16.c (main): Use correct type
943         specifiers.
944         * testsuite/libffi.call/nested_struct1.c (main): Ditto.
945         * testsuite/libffi.call/cls_sint.c (main): Ditto.
946         * testsuite/libffi.call/nested_struct9.c (main): Ditto.
947         * testsuite/libffi.call/cls_20byte1.c (main): Ditto.
948         * testsuite/libffi.call/cls_9byte1.c (main): Ditto.
949         * testsuite/libffi.call/closure_fn1.c (main): Ditto.
950         * testsuite/libffi.call/closure_fn3.c (main): Ditto.
951         * testsuite/libffi.call/return_dbl2.c (main): Ditto.
952         * testsuite/libffi.call/cls_sshort.c (main): Ditto.
953         * testsuite/libffi.call/return_fl3.c (main): Ditto.
954         * testsuite/libffi.call/closure_fn5.c (main): Ditto.
955         * testsuite/libffi.call/nested_struct.c (main): Ditto.
956         * testsuite/libffi.call/nested_struct10.c (main): Ditto.
957         * testsuite/libffi.call/return_ll1.c (main): Ditto.
958         * testsuite/libffi.call/cls_8byte.c (main): Ditto.
959         * testsuite/libffi.call/cls_align_uint32.c (main): Ditto.
960         * testsuite/libffi.call/cls_align_sint16.c (main): Ditto.
961         * testsuite/libffi.call/cls_20byte.c (main): Ditto.
962         * testsuite/libffi.call/nested_struct2.c (main): Ditto.
963         * testsuite/libffi.call/cls_24byte.c (main): Ditto.
964         * testsuite/libffi.call/nested_struct6.c (main): Ditto.
965         * testsuite/libffi.call/cls_uint.c (main): Ditto.
966         * testsuite/libffi.call/cls_12byte.c (main): Ditto.
967         * testsuite/libffi.call/cls_16byte.c (main): Ditto.
968         * testsuite/libffi.call/closure_fn0.c (main): Ditto.
969         * testsuite/libffi.call/cls_9byte2.c (main): Ditto.
970         * testsuite/libffi.call/closure_fn2.c (main): Ditto.
971         * testsuite/libffi.call/return_dbl1.c (main): Ditto.
972         * testsuite/libffi.call/closure_fn4.c (main): Ditto.
973         * testsuite/libffi.call/closure_fn6.c (main): Ditto.
974         * testsuite/libffi.call/cls_align_sint32.c (main): Ditto.
975
976 2007-08-07  Andrew Haley  <aph@redhat.com>
977
978         * src/x86/sysv.S (ffi_closure_raw_SYSV): Fix typo in previous
979         checkin.
980
981 2007-08-06  Andrew Haley  <aph@redhat.com>
982
983         PR testsuite/32843
984         * src/x86/sysv.S (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
985         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
986         FFI_TYPE_SINT32.
987
988 2007-08-02  David Daney  <ddaney@avtrex.com>
989
990         * testsuite/libffi.call/return_ul.c (main): Define return type as
991         ffi_arg.  Use proper printf conversion specifier.
992         
993 2007-07-30  Andrew Haley  <aph@redhat.com>
994
995         PR testsuite/32843
996         * src/x86/ffi.c (ffi_prep_cif_machdep): in x86 case, add code for
997         signed/unsigned int8/16.
998         * src/x86/sysv.S (ffi_call_SYSV): Rewrite to:
999         Use a jump table.
1000         Remove code to pop args from the stack after call.
1001         Special-case signed/unsigned int8/16.
1002         * testsuite/libffi.call/return_sc.c (main): Revert.
1003
1004 2007-07-26  Richard Guenther  <rguenther@suse.de>
1005
1006         PR testsuite/32843
1007         * testsuite/libffi.call/return_sc.c (main): Verify call
1008         result as signed char, not ffi_arg.
1009
1010 2007-07-16  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1011
1012         * configure.ac (i?86-*-solaris2.1[0-9]): Set TARGET to X86_64.
1013         * configure: Regenerate.
1014
1015 2007-07-11  David Daney  <ddaney@avtrex.com>
1016
1017         * src/mips/ffi.c: Don't include sys/cachectl.h.
1018         (ffi_prep_closure_loc): Use __builtin___clear_cache() instead of
1019         cacheflush().
1020
1021 2007-05-18  Aurelien Jarno  <aurelien@aurel32.net>
1022
1023         * src/arm/ffi.c (ffi_prep_closure_loc): Renamed and ajusted
1024         from (ffi_prep_closure): ... this.
1025         (FFI_INIT_TRAMPOLINE): Adjust.
1026
1027 2005-12-31  Phil Blundell  <pb@reciva.com>
1028
1029         * src/arm/ffi.c (ffi_prep_incoming_args_SYSV, 
1030         ffi_closure_SYSV_inner, ffi_prep_closure): New, add closure support.
1031         * src/arm/sysv.S(ffi_closure_SYSV): Likewise.
1032         * src/arm/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1033         (FFI_CLOSURES): Enable closure support.
1034
1035 2007-07-03  Andrew Haley  <aph@hedges.billgatliff.com>
1036
1037         * testsuite/libffi.call/cls_multi_ushort.c,
1038         testsuite/libffi.call/cls_align_uint16.c,
1039         testsuite/libffi.call/nested_struct1.c,
1040         testsuite/libffi.call/nested_struct3.c,
1041         testsuite/libffi.call/cls_7_1_byte.c,
1042         testsuite/libffi.call/cls_double.c,
1043         testsuite/libffi.call/nested_struct5.c,
1044         testsuite/libffi.call/nested_struct7.c,
1045         testsuite/libffi.call/cls_sint.c,
1046         testsuite/libffi.call/nested_struct9.c,
1047         testsuite/libffi.call/cls_20byte1.c,
1048         testsuite/libffi.call/cls_multi_sshortchar.c,
1049         testsuite/libffi.call/cls_align_sint64.c,
1050         testsuite/libffi.call/cls_3byte2.c,
1051         testsuite/libffi.call/cls_multi_schar.c,
1052         testsuite/libffi.call/cls_multi_uchar.c,
1053         testsuite/libffi.call/cls_19byte.c,
1054         testsuite/libffi.call/cls_9byte1.c,
1055         testsuite/libffi.call/cls_align_float.c,
1056         testsuite/libffi.call/closure_fn1.c,
1057         testsuite/libffi.call/problem1.c,
1058         testsuite/libffi.call/closure_fn3.c,
1059         testsuite/libffi.call/cls_sshort.c,
1060         testsuite/libffi.call/closure_fn5.c,
1061         testsuite/libffi.call/cls_align_double.c,
1062         testsuite/libffi.call/cls_2byte.c,
1063         testsuite/libffi.call/nested_struct.c,
1064         testsuite/libffi.call/nested_struct10.c,
1065         testsuite/libffi.call/cls_4byte.c,
1066         testsuite/libffi.call/cls_6byte.c,
1067         testsuite/libffi.call/cls_8byte.c,
1068         testsuite/libffi.call/cls_multi_sshort.c,
1069         testsuite/libffi.call/cls_align_uint32.c,
1070         testsuite/libffi.call/cls_align_sint16.c,
1071         testsuite/libffi.call/cls_float.c,
1072         testsuite/libffi.call/cls_20byte.c,
1073         testsuite/libffi.call/cls_5_1_byte.c,
1074         testsuite/libffi.call/nested_struct2.c,
1075         testsuite/libffi.call/cls_24byte.c,
1076         testsuite/libffi.call/nested_struct4.c,
1077         testsuite/libffi.call/nested_struct6.c,
1078         testsuite/libffi.call/cls_64byte.c,
1079         testsuite/libffi.call/nested_struct8.c,
1080         testsuite/libffi.call/cls_uint.c,
1081         testsuite/libffi.call/cls_multi_ushortchar.c,
1082         testsuite/libffi.call/cls_schar.c,
1083         testsuite/libffi.call/cls_uchar.c,
1084         testsuite/libffi.call/cls_align_uint64.c,
1085         testsuite/libffi.call/cls_ulonglong.c,
1086         testsuite/libffi.call/cls_align_longdouble.c,
1087         testsuite/libffi.call/cls_1_1byte.c,
1088         testsuite/libffi.call/cls_12byte.c,
1089         testsuite/libffi.call/cls_3_1byte.c,
1090         testsuite/libffi.call/cls_3byte1.c,
1091         testsuite/libffi.call/cls_4_1byte.c,
1092         testsuite/libffi.call/cls_6_1_byte.c,
1093         testsuite/libffi.call/cls_16byte.c,
1094         testsuite/libffi.call/cls_18byte.c,
1095         testsuite/libffi.call/closure_fn0.c,
1096         testsuite/libffi.call/cls_9byte2.c,
1097         testsuite/libffi.call/closure_fn2.c,
1098         testsuite/libffi.call/closure_fn4.c,
1099         testsuite/libffi.call/cls_ushort.c,
1100         testsuite/libffi.call/closure_fn6.c,
1101         testsuite/libffi.call/cls_5byte.c,
1102         testsuite/libffi.call/cls_align_pointer.c,
1103         testsuite/libffi.call/cls_7byte.c,
1104         testsuite/libffi.call/cls_align_sint32.c,
1105         testsuite/libffi.special/unwindtest_ffi_call.cc,
1106         testsuite/libffi.special/unwindtest.cc: Enable for ARM.
1107
1108 2007-07-05  H.J. Lu  <hongjiu.lu@intel.com>
1109
1110         * aclocal.m4: Regenerated.
1111
1112 2007-06-02  Paolo Bonzini  <bonzini@gnu.org>
1113
1114         * configure: Regenerate.
1115
1116 2007-05-23  Steve Ellcey  <sje@cup.hp.com>
1117
1118         * Makefile.in: Regenerate.
1119         * configure: Regenerate.
1120         * aclocal.m4: Regenerate.
1121         * include/Makefile.in: Regenerate.
1122         * testsuite/Makefile.in: Regenerate.
1123
1124 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1125
1126         * src/m68k/ffi.c (ffi_prep_incoming_args_SYSV,
1127         ffi_closure_SYSV_inner,ffi_prep_closure): New, add closure support.
1128         * src/m68k/sysv.S(ffi_closure_SYSV,ffi_closure_struct_SYSV): Likewise.
1129         * src/m68k/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
1130         (FFI_CLOSURES): Enable closure support.
1131
1132 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1133
1134         * configure.ac (HAVE_AS_CFI_PSEUDO_OP): New test.
1135         * configure: Regenerate.
1136         * fficonfig.h.in: Regenerate.
1137         * src/m68k/sysv.S (CFI_STARTPROC,CFI_ENDPROC,
1138         CFI_OFFSET,CFI_DEF_CFA): New macros.
1139         (ffi_call_SYSV): Add callframe annotation.
1140
1141 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
1142
1143         * src/m68k/ffi.c (ffi_prep_args,ffi_prep_cif_machdep): Fix
1144         numerous test suite failures.
1145         * src/m68k/sysv.S (ffi_call_SYSV): Likewise.
1146
1147 2007-04-11  Paolo Bonzini  <bonzini@gnu.org>
1148
1149         * Makefile.am (EXTRA_DIST): Bring up to date.
1150         * Makefile.in: Regenerate.
1151         * src/frv/eabi.S: Remove RCS keyword.
1152
1153 2007-04-06  Richard Henderson  <rth@redhat.com>
1154
1155         * configure.ac: Tidy target case.
1156         (HAVE_LONG_DOUBLE): Allow the target to override.
1157         * configure: Regenerate.
1158         * include/ffi.h.in: Don't define ffi_type_foo if
1159         LIBFFI_HIDE_BASIC_TYPES is defined.
1160         (ffi_type_longdouble): If not HAVE_LONG_DOUBLE, define
1161         to ffi_type_double.
1162         * types.c (LIBFFI_HIDE_BASIC_TYPES): Define.
1163         (FFI_TYPEDEF, ffi_type_void): Mark the data const.
1164         (ffi_type_longdouble): Special case for Alpha.  Don't define
1165         if long double == double.
1166
1167         * src/alpha/ffi.c (FFI_TYPE_LONGDOUBLE): Assert unique value.
1168         (ffi_prep_cif_machdep): Handle it as the 128-bit type.
1169         (ffi_call, ffi_closure_osf_inner): Likewise.
1170         (ffi_closure_osf_inner): Likewise.  Mark hidden.
1171         (ffi_call_osf, ffi_closure_osf): Mark hidden.
1172         * src/alpha/ffitarget.h (FFI_LAST_ABI): Tidy definition.
1173         * src/alpha/osf.S (ffi_call_osf, ffi_closure_osf): Mark hidden.
1174         (load_table): Handle 128-bit long double.
1175
1176         * testsuite/libffi.call/float4.c: Add -mieee for alpha.
1177
1178 2007-04-06  Tom Tromey  <tromey@redhat.com>
1179
1180         PR libffi/31491:
1181         * README: Fixed bug in example.
1182
1183 2007-04-03  Jakub Jelinek  <jakub@redhat.com>
1184
1185         * src/closures.c: Include sys/statfs.h.
1186         (_GNU_SOURCE): Define on Linux.
1187         (FFI_MMAP_EXEC_SELINUX): Define.
1188         (selinux_enabled): New variable.
1189         (selinux_enabled_check): New function.
1190         (is_selinux_enabled): Define.
1191         (dlmmap): Use it.
1192
1193 2007-03-24  Uros Bizjak  <ubizjak@gmail.com>
1194
1195         * testsuite/libffi.call/return_fl2.c (return_fl): Mark as static.
1196         Use 'volatile float sum' to create sum of floats to avoid false
1197         negative due to excess precision on ix86 targets.
1198         (main): Ditto. 
1199
1200 2007-03-08  Alexandre Oliva  <aoliva@redhat.com>
1201
1202         * src/powerpc/ffi.c (flush_icache): Fix left-over from previous
1203         patch.
1204         (ffi_prep_closure_loc): Remove unneeded casts.  Add needed ones.
1205
1206 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1207
1208         * include/ffi.h.in (ffi_closure_alloc, ffi_closure_free): New.
1209         (ffi_prep_closure_loc): New.
1210         (ffi_prep_raw_closure_loc): New.
1211         (ffi_prep_java_raw_closure_loc): New.
1212         * src/closures.c: New file.
1213         * src/dlmalloc.c [FFI_MMAP_EXEC_WRIT] (struct malloc_segment):
1214         Replace sflags with exec_offset.
1215         [FFI_MMAP_EXEC_WRIT] (mmap_exec_offset, add_segment_exec_offset,
1216         sub_segment_exec_offset): New macros.
1217         (get_segment_flags, set_segment_flags, check_segment_merge): New
1218         macros.
1219         (is_mmapped_segment, is_extern_segment): Use get_segment_flags.
1220         (add_segment, sys_alloc, create_mspace, create_mspace_with_base,
1221         destroy_mspace): Use new macros.
1222         (sys_alloc): Silence warning.
1223         * Makefile.am (libffi_la_SOURCES): Add src/closures.c.
1224         * Makefile.in: Rebuilt.
1225         * src/prep_cif [FFI_CLOSURES] (ffi_prep_closure): Implement in
1226         terms of ffi_prep_closure_loc.
1227         * src/raw_api.c (ffi_prep_raw_closure_loc): Renamed and adjusted
1228         from...
1229         (ffi_prep_raw_closure): ... this.  Re-implement in terms of the
1230         renamed version.
1231         * src/java_raw_api (ffi_prep_java_raw_closure_loc): Renamed and
1232         adjusted from...
1233         (ffi_prep_java_raw_closure): ... this.  Re-implement in terms of
1234         the renamed version.
1235         * src/alpha/ffi.c (ffi_prep_closure_loc): Renamed from
1236         (ffi_prep_closure): ... this.
1237         * src/pa/ffi.c: Likewise.
1238         * src/cris/ffi.c: Likewise.  Adjust.
1239         * src/frv/ffi.c: Likewise.
1240         * src/ia64/ffi.c: Likewise.
1241         * src/mips/ffi.c: Likewise.
1242         * src/powerpc/ffi_darwin.c: Likewise.
1243         * src/s390/ffi.c: Likewise.
1244         * src/sh/ffi.c: Likewise.
1245         * src/sh64/ffi.c: Likewise.
1246         * src/sparc/ffi.c: Likewise.
1247         * src/x86/ffi64.c: Likewise.
1248         * src/x86/ffi.c: Likewise.
1249         (FFI_INIT_TRAMPOLINE): Adjust.
1250         (ffi_prep_raw_closure_loc): Renamed and adjusted from...
1251         (ffi_prep_raw_closure): ... this.
1252         * src/powerpc/ffi.c (ffi_prep_closure_loc): Renamed from
1253         (ffi_prep_closure): ... this.
1254         (flush_icache): Adjust.
1255
1256 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
1257
1258         * src/dlmalloc.c: New file, imported version 2.8.3 of Doug
1259         Lea's malloc.
1260
1261 2007-03-01  Brooks Moses  <brooks.moses@codesourcery.com>
1262
1263         * Makefile.am: Add dummy install-pdf target.
1264         * Makefile.in: Regenerate
1265
1266 2007-02-13  Andreas Krebbel  <krebbel1@de.ibm.com>
1267
1268         * src/s390/ffi.c (ffi_prep_args, ffi_prep_cif_machdep,
1269         ffi_closure_helper_SYSV): Add long double handling.
1270
1271 2007-02-02  Jakub Jelinek  <jakub@redhat.com>
1272
1273         * src/powerpc/linux64.S (ffi_call_LINUX64): Move restore of r2
1274         immediately after bctrl instruction.
1275
1276 2007-01-18  Alexandre Oliva  <aoliva@redhat.com>
1277
1278         * Makefile.am (all-recursive, install-recursive,
1279         mostlyclean-recursive, clean-recursive, distclean-recursive,
1280         maintainer-clean-recursive): Add missing targets.
1281         * Makefile.in: Rebuilt.
1282
1283 2006-12-14  Andreas Tobler  <a.tobler@schweiz.org>
1284
1285         * configure.ac: Add TARGET for x86_64-*-darwin*.
1286         * Makefile.am (nodist_libffi_la_SOURCES): Add rules for 64-bit sources
1287         for X86_DARWIN.
1288         * src/x86/ffitarget.h: Set trampoline size for x86_64-*-darwin*.
1289         * src/x86/darwin64.S: New file for x86_64-*-darwin* support.
1290         * configure: Regenerate.
1291         * Makefile.in: Regenerate.
1292         * include/Makefile.in: Regenerate.
1293         * testsuite/Makefile.in: Regenerate.
1294         * testsuite/libffi.special/unwindtest_ffi_call.cc: New test case for
1295         ffi_call only.
1296
1297 2006-12-13  Andreas Tobler <a.tobler@schweiz.org>
1298
1299         * aclocal.m4: Regenerate with aclocal -I .. as written in the
1300         Makefile.am.
1301
1302 2006-10-31  Geoffrey Keating  <geoffk@apple.com>
1303
1304         * src/powerpc/ffi_darwin.c (darwin_adjust_aggregate_sizes): New.
1305         (ffi_prep_cif_machdep): Call darwin_adjust_aggregate_sizes for
1306         Darwin.
1307         * testsuite/libffi.call/nested_struct4.c: Remove Darwin XFAIL.
1308         * testsuite/libffi.call/nested_struct6.c: Remove Darwin XFAIL.
1309
1310 2006-10-10  Paolo Bonzini  <bonzini@gnu.org>
1311             Sandro Tolaini  <tolaini@libero.it>
1312
1313         * configure.ac [i*86-*-darwin*]: Set X86_DARWIN symbol and 
1314         conditional.
1315         * configure: Regenerated.
1316         * Makefile.am (nodist_libffi_la_SOURCES) [X86_DARWIN]: New case.
1317         (EXTRA_DIST): Add src/x86/darwin.S.
1318         * Makefile.in: Regenerated.
1319         * include/Makefile.in: Regenerated.
1320         * testsuite/Makefile.in: Regenerated.
1321
1322         * src/x86/ffi.c (ffi_prep_cif_machdep) [X86_DARWIN]: Treat like
1323         X86_WIN32, and additionally align stack to 16 bytes.
1324         * src/x86/darwin.S: New, based on sysv.S.
1325         * src/prep_cif.c (ffi_prep_cif) [X86_DARWIN]: Align > 8-byte structs.
1326
1327 2006-09-12  David Daney  <ddaney@avtrex.com>
1328
1329         PR libffi/23935
1330         * include/Makefile.am: Install both ffi.h and ffitarget.h in
1331         $(libdir)/gcc/$(target_alias)/$(gcc_version)/include.
1332         * aclocal.m4: Regenerated for automake 1.9.6.
1333         * Makefile.in: Regenerated.
1334         * include/Makefile.in: Regenerated.
1335         * testsuite/Makefile.in: Regenerated.
1336
1337 2006-08-17  Andreas Tobler  <a.tobler@schweiz.ch>
1338
1339         * include/ffi_common.h (struct): Revert accidental commit.
1340
1341 2006-08-15  Andreas Tobler  <a.tobler@schweiz.ch>
1342
1343         * include/ffi_common.h: Remove lint directives.
1344         * include/ffi.h.in: Likewise.
1345
1346 2006-07-25  Torsten Schoenfeld  <kaffeetisch@gmx.de>
1347
1348         * include/ffi.h.in (ffi_type_ulong, ffi_type_slong): Define correctly
1349         for 32-bit architectures.
1350         * testsuite/libffi.call/return_ul.c: New test case.
1351
1352 2006-07-19  David Daney  <ddaney@avtrex.com>
1353
1354         * testsuite/libffi.call/closure_fn6.c: Remove xfail for mips,
1355         xfail remains for mips64.
1356
1357 2006-05-23  Carlos O'Donell  <carlos@codesourcery.com>
1358
1359         * Makefile.am: Add install-html target. Add install-html to .PHONY
1360         * Makefile.in: Regenerate.
1361         * aclocal.m4: Regenerate.
1362         * include/Makefile.in: Regenerate.
1363         * testsuite/Makefile.in: Regenerate.
1364
1365 2006-05-18  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1366
1367         * pa/ffi.c (ffi_prep_args_pa32): Load floating point arguments from
1368         stack slot.
1369
1370 2006-04-22  Andreas Tobler  <a.tobler@schweiz.ch>
1371
1372         * README: Remove notice about 'Crazy Comments'.
1373         * src/debug.c: Remove lint directives. Cleanup white spaces.
1374         * src/java_raw_api.c: Likewise.
1375         * src/prep_cif.c: Likewise.
1376         * src/raw_api.c: Likewise.
1377         * src/ffitest.c: Delete. No longer needed, all test cases migrated
1378         to the testsuite.
1379         * src/arm/ffi.c: Remove lint directives.
1380         * src/m32r/ffi.c: Likewise.
1381         * src/pa/ffi.c: Likewise.
1382         * src/powerpc/ffi.c: Likewise.
1383         * src/powerpc/ffi_darwin.c: Likewise.
1384         * src/sh/ffi.c: Likewise.
1385         * src/sh64/ffi.c: Likewise.
1386         * src/x86/ffi.c: Likewise.
1387         * testsuite/libffi.call/float2.c: Likewise.
1388         * testsuite/libffi.call/promotion.c: Likewise.
1389         * testsuite/libffi.call/struct1.c: Likewise.
1390
1391 2006-04-13  Andreas Tobler  <a.tobler@schweiz.ch>
1392
1393         * src/pa/hpux32.S: Correct unwind offset calculation for
1394         ffi_closure_pa32.
1395         * src/pa/linux.S: Likewise.
1396
1397 2006-04-12  James E Wilson  <wilson@specifix.com>
1398
1399         PR libgcj/26483
1400         * src/ia64/ffi.c (stf_spill, ldf_fill): Rewrite as macros.
1401         (hfa_type_load): Call stf_spill.
1402         (hfa_type_store): Call ldf_fill.
1403         (ffi_call): Adjust calls to above routines.  Add local temps for
1404         macro result.
1405
1406 2006-04-10  Matthias Klose  <doko@debian.org>
1407
1408         * testsuite/lib/libffi-dg.exp (libffi-init): Recognize multilib
1409         directory names containing underscores.
1410
1411 2006-04-07  James E Wilson  <wilson@specifix.com>
1412
1413         * testsuite/libffi.call/float4.c: New testcase.
1414
1415 2006-04-05  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1416             Andreas Tobler  <a.tobler@schweiz.ch>
1417
1418         * Makefile.am: Add PA_HPUX port.
1419         * Makefile.in: Regenerate.
1420         * include/Makefile.in: Likewise.
1421         * testsuite/Makefile.in: Likewise.
1422         * configure.ac: Add PA_HPUX rules.
1423         * configure: Regenerate.
1424         * src/pa/ffitarget.h: Rename linux target to PA_LINUX.
1425         Add PA_HPUX and PA64_HPUX.
1426         Rename FFI_LINUX ABI to FFI_PA32 ABI.
1427         (FFI_TRAMPOLINE_SIZE): Define for 32-bit HP-UX targets.
1428         (FFI_TYPE_SMALL_STRUCT2): Define.
1429         (FFI_TYPE_SMALL_STRUCT4): Likewise.
1430         (FFI_TYPE_SMALL_STRUCT8): Likewise.
1431         (FFI_TYPE_SMALL_STRUCT3): Redefine.
1432         (FFI_TYPE_SMALL_STRUCT5): Likewise.
1433         (FFI_TYPE_SMALL_STRUCT6): Likewise.
1434         (FFI_TYPE_SMALL_STRUCT7): Likewise.
1435         * src/pa/ffi.c (ROUND_DOWN): Delete.
1436         (fldw, fstw, fldd, fstd): Use '__asm__'.
1437         (ffi_struct_type): Add support for FFI_TYPE_SMALL_STRUCT2,
1438         FFI_TYPE_SMALL_STRUCT4 and FFI_TYPE_SMALL_STRUCT8.
1439         (ffi_prep_args_LINUX): Rename to ffi_prep_args_pa32. Update comment.
1440         Simplify incrementing of stack slot variable. Change type of local
1441         'n' to unsigned int.
1442         (ffi_size_stack_LINUX): Rename to ffi_size_stack_pa32. Handle long
1443         double on PA_HPUX.
1444         (ffi_prep_cif_machdep): Likewise.
1445         (ffi_call): Likewise.
1446         (ffi_closure_inner_LINUX): Rename to ffi_closure_inner_pa32. Change
1447         return type to ffi_status. Simplify incrementing of stack slot
1448         variable. Only copy floating point argument registers when PA_LINUX
1449         is true. Reformat debug statement.
1450         Add support for FFI_TYPE_SMALL_STRUCT2, FFI_TYPE_SMALL_STRUCT4 and
1451         FFI_TYPE_SMALL_STRUCT8.
1452         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Add 'extern' to
1453         declaration.
1454         (ffi_prep_closure): Make linux trampoline conditional on PA_LINUX.
1455         Add nops to cache flush.  Add trampoline for PA_HPUX.
1456         * src/pa/hpux32.S: New file.
1457         * src/pa/linux.S (ffi_call_LINUX): Rename to ffi_call_pa32. Rename
1458         ffi_prep_args_LINUX to ffi_prep_args_pa32.
1459         Localize labels. Add support for 2, 4 and 8-byte small structs. Handle
1460         unaligned destinations in 3, 5, 6 and 7-byte small structs. Order
1461         argument type checks so that common argument types appear first.
1462         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Rename
1463         ffi_closure_inner_LINUX to ffi_closure_inner_pa32.
1464
1465 2006-03-24  Alan Modra  <amodra@bigpond.net.au>
1466
1467         * src/powerpc/ffitarget.h (enum ffi_abi): Add FFI_LINUX.  Default
1468         for 32-bit using IBM extended double format.  Fix FFI_LAST_ABI.
1469         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Handle linux variant of
1470         FFI_TYPE_LONGDOUBLE.
1471         (ffi_prep_args64): Assert using IBM extended double.
1472         (ffi_prep_cif_machdep): Don't munge FFI_TYPE_LONGDOUBLE type.
1473         Handle FFI_LINUX FFI_TYPE_LONGDOUBLE return and args.
1474         (ffi_call): Handle FFI_LINUX.
1475         (ffi_closure_helper_SYSV): Non FFI_LINUX long double return needs
1476         gpr3 return pointer as for struct return.  Handle FFI_LINUX
1477         FFI_TYPE_LONGDOUBLE return and args.  Don't increment "nf"
1478         unnecessarily.
1479         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Load both f1 and f2
1480         for FFI_TYPE_LONGDOUBLE.  Move epilogue insns into case table.
1481         Don't use r6 as pointer to results, instead use sp offset.  Don't
1482         make a special call to load lr with case table address, instead
1483         use offset from previous call.
1484         * src/powerpc/sysv.S (ffi_call_SYSV): Save long double return.
1485         * src/powerpc/linux64.S (ffi_call_LINUX64): Simplify long double
1486         return.
1487
1488 2006-03-15  Kaz Kojima  <kkojima@gcc.gnu.org>
1489
1490         * src/sh64/ffi.c (ffi_prep_cif_machdep): Handle float arguments
1491         passed with FP registers correctly.
1492         (ffi_closure_helper_SYSV): Likewise.
1493         * src/sh64/sysv.S: Likewise.
1494
1495 2006-03-01  Andreas Tobler  <a.tobler@schweiz.ch>
1496
1497         * testsuite/libffi.special/unwindtest.cc (closure_test_fn): Mark cif,
1498         args and userdata unused.
1499         (closure_test_fn1): Mark cif and userdata unused.
1500         (main): Remove unused res.
1501
1502 2006-02-28  Andreas Tobler  <a.tobler@schweiz.ch>
1503
1504         * testsuite/libffi.call/call.exp: Adjust FSF address. Add test runs for
1505         -O2, -O3, -Os and the warning flags -W -Wall.
1506         * testsuite/libffi.special/special.exp: Likewise.
1507         * testsuite/libffi.call/ffitest.h: Add an __UNUSED__ macro to mark
1508         unused parameter unused for gcc or else do nothing.
1509         * testsuite/libffi.special/ffitestcxx.h: Likewise.
1510         * testsuite/libffi.call/cls_12byte.c (cls_struct_12byte_gn): Mark cif
1511         and userdata unused.
1512         * testsuite/libffi.call/cls_16byte.c (cls_struct_16byte_gn): Likewise.
1513         * testsuite/libffi.call/cls_18byte.c (cls_struct_18byte_gn): Likewise.
1514         * testsuite/libffi.call/cls_19byte.c (cls_struct_19byte_gn): Likewise.
1515         * testsuite/libffi.call/cls_1_1byte.c (cls_struct_1_1byte_gn): Likewise.
1516         * testsuite/libffi.call/cls_20byte.c (cls_struct_20byte_gn): Likewise.
1517         * testsuite/libffi.call/cls_20byte1.c (cls_struct_20byte_gn): Likewise.
1518         * testsuite/libffi.call/cls_24byte.c (cls_struct_24byte_gn): Likewise.
1519         * testsuite/libffi.call/cls_2byte.c (cls_struct_2byte_gn): Likewise.
1520         * testsuite/libffi.call/cls_3_1byte.c (cls_struct_3_1byte_gn): Likewise.
1521         * testsuite/libffi.call/cls_3byte1.c (cls_struct_3byte_gn): Likewise.
1522         * testsuite/libffi.call/cls_3byte2.c (cls_struct_3byte_gn1): Likewise.
1523         * testsuite/libffi.call/cls_4_1byte.c (cls_struct_4_1byte_gn): Likewise.
1524         * testsuite/libffi.call/cls_4byte.c (cls_struct_4byte_gn): Likewise.
1525         * testsuite/libffi.call/cls_5_1_byte.c (cls_struct_5byte_gn): Likewise.
1526         * testsuite/libffi.call/cls_5byte.c (cls_struct_5byte_gn): Likewise.
1527         * testsuite/libffi.call/cls_64byte.c (cls_struct_64byte_gn): Likewise.
1528         * testsuite/libffi.call/cls_6_1_byte.c (cls_struct_6byte_gn): Likewise.
1529         * testsuite/libffi.call/cls_6byte.c (cls_struct_6byte_gn): Likewise.
1530         * testsuite/libffi.call/cls_7_1_byte.c (cls_struct_7byte_gn): Likewise.
1531         * testsuite/libffi.call/cls_7byte.c (cls_struct_7byte_gn): Likewise.
1532         * testsuite/libffi.call/cls_8byte.c (cls_struct_8byte_gn): Likewise.
1533         * testsuite/libffi.call/cls_9byte1.c (cls_struct_9byte_gn): Likewise.
1534         * testsuite/libffi.call/cls_9byte2.c (cls_struct_9byte_gn): Likewise.
1535         * testsuite/libffi.call/cls_align_double.c (cls_struct_align_gn):
1536         Likewise.
1537         * testsuite/libffi.call/cls_align_float.c (cls_struct_align_gn):
1538         Likewise.
1539         * testsuite/libffi.call/cls_align_longdouble.c (cls_struct_align_gn):
1540         Likewise.
1541         * testsuite/libffi.call/cls_align_pointer.c (cls_struct_align_fn): Cast
1542         void* to avoid compiler warning.
1543         (main): Likewise.
1544         (cls_struct_align_gn): Mark cif and userdata unused.
1545         * testsuite/libffi.call/cls_align_sint16.c (cls_struct_align_gn):
1546         Likewise.
1547         * testsuite/libffi.call/cls_align_sint32.c (cls_struct_align_gn):
1548         Likewise.
1549         * testsuite/libffi.call/cls_align_sint64.c (cls_struct_align_gn):
1550         Likewise.
1551         * testsuite/libffi.call/cls_align_uint16.c (cls_struct_align_gn):
1552         Likewise.
1553         * testsuite/libffi.call/cls_align_uint32.c (cls_struct_align_gn):
1554         Likewise.
1555         * testsuite/libffi.call/cls_double.c (cls_ret_double_fn): Likewise.
1556         * testsuite/libffi.call/cls_float.c (cls_ret_float_fn): Likewise.
1557         * testsuite/libffi.call/cls_multi_schar.c (test_func_gn): Mark cif and
1558         data unused.
1559         (main): Cast res_call to silence gcc.
1560         * testsuite/libffi.call/cls_multi_sshort.c (test_func_gn): Mark cif and
1561         data unused.
1562         (main): Cast res_call to silence gcc.
1563         * testsuite/libffi.call/cls_multi_sshortchar.c (test_func_gn): Mark cif
1564         and data unused.
1565         (main): Cast res_call to silence gcc.
1566         * testsuite/libffi.call/cls_multi_uchar.c (test_func_gn): Mark cif and
1567         data unused.
1568         (main): Cast res_call to silence gcc.
1569         * testsuite/libffi.call/cls_multi_ushort.c (test_func_gn): Mark cif and
1570         data unused.
1571         (main): Cast res_call to silence gcc.
1572         * testsuite/libffi.call/cls_multi_ushortchar.c (test_func_gn): Mark cif
1573         and data unused.
1574         (main): Cast res_call to silence gcc.
1575         * testsuite/libffi.call/cls_schar.c (cls_ret_schar_fn): Mark cif and
1576         userdata unused.
1577         (cls_ret_schar_fn): Cast printf parameter to silence gcc.
1578         * testsuite/libffi.call/cls_sint.c (cls_ret_sint_fn): Mark cif and
1579         userdata unused.
1580         (cls_ret_sint_fn): Cast printf parameter to silence gcc.
1581         * testsuite/libffi.call/cls_sshort.c (cls_ret_sshort_fn): Mark cif and
1582         userdata unused.
1583         (cls_ret_sshort_fn): Cast printf parameter to silence gcc.
1584         * testsuite/libffi.call/cls_uchar.c (cls_ret_uchar_fn):  Mark cif and
1585         userdata unused.
1586         (cls_ret_uchar_fn): Cast printf parameter to silence gcc.
1587         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Mark cif and
1588         userdata unused.
1589         (cls_ret_uint_fn): Cast printf parameter to silence gcc.
1590         * testsuite/libffi.call/cls_ulonglong.c (cls_ret_ulonglong_fn): Mark cif
1591         and userdata unused.
1592         * testsuite/libffi.call/cls_ushort.c (cls_ret_ushort_fn): Mark cif and
1593         userdata unused.
1594         (cls_ret_ushort_fn): Cast printf parameter to silence gcc.
1595         * testsuite/libffi.call/float.c (floating): Remove unused parameter e.
1596         * testsuite/libffi.call/float1.c (main): Remove unused variable i.
1597         Cleanup white spaces.
1598         * testsuite/libffi.call/negint.c (checking): Remove unused variable i.
1599         * testsuite/libffi.call/nested_struct.c (cls_struct_combined_gn): Mark
1600         cif and userdata unused.
1601         * testsuite/libffi.call/nested_struct1.c (cls_struct_combined_gn):
1602         Likewise.
1603         * testsuite/libffi.call/nested_struct10.c (B_gn): Likewise.
1604         * testsuite/libffi.call/nested_struct2.c (B_fn): Adjust printf
1605         formatters to silence gcc.
1606         (B_gn): Mark cif and userdata unused.
1607         * testsuite/libffi.call/nested_struct3.c (B_gn): Mark cif and userdata
1608         unused.
1609         * testsuite/libffi.call/nested_struct4.c: Mention related PR.
1610         (B_gn): Mark cif and userdata unused.
1611         * testsuite/libffi.call/nested_struct5.c (B_gn): Mark cif and userdata
1612         unused.
1613         * testsuite/libffi.call/nested_struct6.c: Mention related PR.
1614         (B_gn): Mark cif and userdata unused.
1615         * testsuite/libffi.call/nested_struct7.c (B_gn): Mark cif and userdata
1616         unused.
1617         * testsuite/libffi.call/nested_struct8.c (B_gn): Likewise.
1618         * testsuite/libffi.call/nested_struct9.c (B_gn): Likewise.
1619         * testsuite/libffi.call/problem1.c (stub): Likewise.
1620         * testsuite/libffi.call/pyobjc-tc.c (main): Cast the result to silence
1621         gcc.
1622         * testsuite/libffi.call/return_fl2.c (return_fl): Add the note mentioned
1623         in the last commit for this test case in the test case itself.
1624         * testsuite/libffi.call/closure_fn0.c (closure_test_fn0): Mark cif as
1625         unused.
1626         * testsuite/libffi.call/closure_fn1.c (closure_test_fn1): Likewise.
1627         * testsuite/libffi.call/closure_fn2.c (closure_test_fn2): Likewise.
1628         * testsuite/libffi.call/closure_fn3.c (closure_test_fn3): Likewise.
1629         * testsuite/libffi.call/closure_fn4.c (closure_test_fn0): Likewise.
1630         * testsuite/libffi.call/closure_fn5.c (closure_test_fn5): Likewise.
1631         * testsuite/libffi.call/closure_fn6.c (closure_test_fn0): Likewise.
1632
1633 2006-02-22  Kaz Kojima  <kkojima@gcc.gnu.org>
1634
1635         * src/sh/sysv.S: Fix register numbers in the FDE for
1636         ffi_closure_SYSV.
1637
1638 2006-02-20  Andreas Tobler  <a.tobler@schweiz.ch>
1639
1640         * testsuite/libffi.call/return_fl2.c (return_fl): Remove static
1641         declaration to avoid a false negative on ix86. See PR323.
1642
1643 2006-02-18  Kaz Kojima  <kkojima@gcc.gnu.org>
1644
1645         * src/sh/ffi.c (ffi_closure_helper_SYSV): Remove unused variable
1646         and cast integer to void * if needed.  Update the pointer to
1647         the FP register saved area correctly.
1648         
1649 2006-02-17  Andreas Tobler  <a.tobler@schweiz.ch>
1650
1651         * testsuite/libffi.call/nested_struct6.c: XFAIL this test until PR25630
1652         is fixed.
1653         * testsuite/libffi.call/nested_struct4.c: Likewise.
1654
1655 2006-02-16  Andreas Tobler  <a.tobler@schweiz.ch>
1656
1657         * testsuite/libffi.call/return_dbl.c: New test case.
1658         * testsuite/libffi.call/return_dbl1.c: Likewise.
1659         * testsuite/libffi.call/return_dbl2.c: Likewise.
1660         * testsuite/libffi.call/return_fl.c: Likewise.
1661         * testsuite/libffi.call/return_fl1.c: Likewise.
1662         * testsuite/libffi.call/return_fl2.c: Likewise.
1663         * testsuite/libffi.call/return_fl3.c: Likewise.
1664         * testsuite/libffi.call/closure_fn6.c: Likewise.
1665
1666         * testsuite/libffi.call/nested_struct2.c: Remove ffi_type_mylong
1667         definition.
1668         * testsuite/libffi.call/ffitest.h: Add ffi_type_mylong definition
1669         here to be used by other test cases too.
1670
1671         * testsuite/libffi.call/nested_struct10.c: New test case.
1672         * testsuite/libffi.call/nested_struct9.c: Likewise.
1673         * testsuite/libffi.call/nested_struct8.c: Likewise.
1674         * testsuite/libffi.call/nested_struct7.c: Likewise.
1675         * testsuite/libffi.call/nested_struct6.c: Likewise.
1676         * testsuite/libffi.call/nested_struct5.c: Likewise.
1677         * testsuite/libffi.call/nested_struct4.c: Likewise.
1678
1679 2006-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
1680
1681         * configure.ac: Enable libffi for sparc64-*-freebsd*.
1682         * configure: Rebuilt.
1683
1684 2006-01-18  Jakub Jelinek  <jakub@redhat.com>
1685
1686         * src/powerpc/sysv.S (smst_two_register): Don't call __ashldi3,
1687         instead do the shifting inline.
1688         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't compute %r5
1689         shift count unconditionally.  Simplify load sequences for 1, 2, 3, 4
1690         and 8 byte structs, for the remaining struct sizes don't call
1691         __lshrdi3, instead do the shifting inline.
1692
1693 2005-12-07  Thiemo Seufer  <ths@networkno.de>
1694
1695         * src/mips/ffitarget.h: Remove obsolete sgidefs.h include. Add
1696         missing parentheses.
1697         * src/mips/o32.S (ffi_call_O32): Code formatting. Define
1698         and use A3_OFF, FP_OFF, RA_OFF. Micro-optimizations.
1699         (ffi_closure_O32): Likewise, but with newly defined A3_OFF2,
1700         A2_OFF2, A1_OFF2, A0_OFF2, RA_OFF2, FP_OFF2, S0_OFF2, GP_OFF2,
1701         V1_OFF2, V0_OFF2, FA_1_1_OFF2, FA_1_0_OFF2, FA_0_1_OFF2,
1702         FA_0_0_OFF2.
1703         * src/mips/ffi.c (ffi_prep_args): Code formatting. Fix
1704         endianness bugs.
1705         (ffi_prep_closure): Improve trampoline instruction scheduling.
1706         (ffi_closure_mips_inner_O32): Fix endianness bugs.
1707
1708 2005-12-03  Alan Modra  <amodra@bigpond.net.au>
1709
1710         * src/powerpc/ffi.c: Formatting.
1711         (ffi_prep_args_SYSV): Avoid possible aliasing problems by using unions.
1712         (ffi_prep_args64): Likewise.
1713
1714 2005-09-30  Geoffrey Keating  <geoffk@apple.com>
1715
1716         * testsuite/lib/libffi-dg.exp (libffi_target_compile): For
1717         darwin, use -shared-libgcc not -lgcc_s, and explain why.
1718
1719 2005-09-26  Tom Tromey  <tromey@redhat.com>
1720
1721         * testsuite/libffi.call/float1.c (value_type): New typedef.
1722         (CANARY): New define.
1723         (main): Check for result buffer overflow.
1724         * src/powerpc/linux64.S: Handle linux64 long double returns.
1725         * src/powerpc/ffi.c (FLAG_RETURNS_128BITS): New constant.
1726         (ffi_prep_cif_machdep): Handle linux64 long double returns.
1727
1728 2005-08-25  Alan Modra  <amodra@bigpond.net.au>
1729
1730         PR target/23404
1731         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Correct placement of stack
1732         homed fp args.
1733         (ffi_status ffi_prep_cif_machdep): Correct stack sizing for same.
1734
1735 2005-08-11  Jakub Jelinek  <jakub@redhat.com>
1736
1737         * configure.ac (HAVE_HIDDEN_VISIBILITY_ATTRIBUTE): New test.
1738         (AH_BOTTOM): Add FFI_HIDDEN definition.
1739         * configure: Rebuilt.
1740         * fficonfig.h.in: Rebuilt.
1741         * src/powerpc/ffi.c (hidden): Remove.
1742         (ffi_closure_LINUX64, ffi_prep_args64, ffi_call_LINUX64,
1743         ffi_closure_helper_LINUX64): Use FFI_HIDDEN instead of hidden.
1744         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64,
1745         .ffi_closure_LINUX64): Use FFI_HIDDEN instead of .hidden.
1746         * src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV): Remove,
1747         add FFI_HIDDEN to its prototype.
1748         (ffi_closure_SYSV_inner): New.
1749         * src/x86/sysv.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
1750         * src/x86/win32.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
1751
1752 2005-08-10  Alfred M. Szmidt  <ams@gnu.org>
1753
1754         PR libffi/21819:
1755         * configure: Rebuilt.
1756         * configure.ac: Handle i*86-*-gnu*.
1757
1758 2005-08-09  Jakub Jelinek  <jakub@redhat.com>
1759
1760         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Use
1761         DW_CFA_offset_extended_sf rather than
1762         DW_CFA_GNU_negative_offset_extended.
1763         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise.
1764
1765 2005-07-22  SUGIOKA Toshinobu  <sugioka@itonet.co.jp>
1766
1767         * src/sh/sysv.S (ffi_call_SYSV): Stop argument popping correctly
1768         on sh3.
1769         (ffi_closure_SYSV): Change the stack layout for sh3 struct argument.
1770         * src/sh/ffi.c (ffi_prep_args): Fix sh3 argument copy, when it is
1771         partially on register.
1772         (ffi_closure_helper_SYSV): Likewise.
1773         (ffi_prep_cif_machdep): Don't set too many cif->flags.
1774
1775 2005-07-20  Kaz Kojima  <kkojima@gcc.gnu.org>
1776
1777         * src/sh/ffi.c (ffi_call): Handle small structures correctly.
1778         Remove empty line.
1779         * src/sh64/ffi.c (simple_type): Remove.
1780         (return_type): Handle small structures correctly.
1781         (ffi_prep_args): Likewise.
1782         (ffi_call): Likewise.
1783         (ffi_closure_helper_SYSV): Likewise.
1784         * src/sh64/sysv.S (ffi_call_SYSV): Handle 1, 2 and 4-byte return.
1785         Emit position independent code if PIC and remove wrong datalabel
1786         prefixes from EH data.
1787
1788 2005-07-19  Andreas Tobler  <a.tobler@schweiz.ch>
1789
1790         * Makefile.am (nodist_libffi_la_SOURCES): Add POWERPC_FREEBSD.
1791         * Makefile.in: Regenerate.
1792         * include/Makefile.in: Likewise.
1793         * testsuite/Makefile.in: Likewise.
1794         * configure.ac: Add POWERPC_FREEBSD rules.
1795         * configure: Regenerate.
1796         * src/powerpc/ffitarget.h: Add POWERPC_FREEBSD rules.
1797         (FFI_SYSV_TYPE_SMALL_STRUCT): Define.
1798         * src/powerpc/ffi.c: Add flags to handle small structure returns
1799         in ffi_call_SYSV.
1800         (ffi_prep_cif_machdep): Handle small structures for SYSV 4 ABI.
1801         Aka FFI_SYSV.
1802         (ffi_closure_helper_SYSV): Likewise.
1803         * src/powerpc/ppc_closure.S: Add return types for small structures.
1804         * src/powerpc/sysv.S: Add bits to handle small structures for
1805         final SYSV 4 ABI.
1806
1807 2005-07-10  Andreas Tobler  <a.tobler@schweiz.ch>
1808
1809         * testsuite/libffi.call/cls_5_1_byte.c: New test file.
1810         * testsuite/libffi.call/cls_6_1_byte.c: Likewise.
1811         * testsuite/libffi.call/cls_7_1_byte.c: Likewise.
1812
1813 2005-07-05  Randolph Chung  <tausq@debian.org>
1814
1815         * src/pa/ffi.c (ffi_struct_type): Rename FFI_TYPE_SMALL_STRUCT1
1816         as FFI_TYPE_SMALL_STRUCT3.  Break out handling for 5-7 byte
1817         structures.  Kill compilation warnings.
1818         (ffi_closure_inner_LINUX): Print return values as hex in debug
1819         message.  Rename FFI_TYPE_SMALL_STRUCT1 as FFI_TYPE_SMALL_STRUCT3.
1820         Properly handle 5-7 byte structure returns.
1821         * src/pa/ffitarget.h (FFI_TYPE_SMALL_STRUCT1)
1822         (FFI_TYPE_SMALL_STRUCT2): Remove.
1823         (FFI_TYPE_SMALL_STRUCT3, FFI_TYPE_SMALL_STRUCT5)
1824         (FFI_TYPE_SMALL_STRUCT6, FFI_TYPE_SMALL_STRUCT7): Define.
1825         * src/pa/linux.S: Mark source file as using PA1.1 assembly.
1826         (checksmst1, checksmst2): Remove.
1827         (checksmst3): Optimize handling of 3-byte struct returns.
1828         (checksmst567): Properly handle 5-7 byte struct returns.
1829
1830 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1831
1832         PR libgcj/21943
1833         * src/mips/n32.S: Enforce PIC code.
1834         * src/mips/o32.S: Likewise.
1835
1836 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1837
1838         * configure.ac: Treat i*86-*-solaris2.10 and up as X86_64.
1839         * configure: Regenerate.
1840         
1841 2005-06-01  Alan Modra  <amodra@bigpond.net.au>
1842
1843         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't use JUMPTARGET
1844         to call ffi_closure_helper_SYSV.  Append @local instead.
1845         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise for ffi_prep_args_SYSV.
1846
1847 2005-05-17  Kelley Cook  <kcook@gcc.gnu.org>
1848
1849         * configure.ac: Use AC_C_BIGENDIAN instead of AC_C_BIGENDIAN_CROSS.
1850         Use AC_CHECK_SIZEOF instead of AC_COMPILE_CHECK_SIZEOF.
1851         * Makefile.am (ACLOCAL_AMFLAGS): Remove -I ../config.
1852         * aclocal.m4, configure, fficonfig.h.in, Makefile.in,
1853         include/Makefile.in, testsuite/Makefile.in: Regenerate.
1854         
1855 2005-05-09  Mike Stump  <mrs@apple.com>
1856
1857         * configure: Regenerate.
1858
1859 2005-05-08  Richard Henderson  <rth@redhat.com>
1860
1861         PR libffi/21285
1862         * src/alpha/osf.S: Update unwind into to match code.
1863
1864 2005-05-04  Andreas Degert <ad@papyrus-gmbh.de>
1865             Richard Henderson  <rth@redhat.com>
1866
1867         * src/x86/ffi64.c (ffi_prep_cif_machdep): Save sse-used flag in
1868         bit 11 of flags.
1869         (ffi_call): Mask return type field.  Pass ssecount to ffi_call_unix64.
1870         (ffi_prep_closure): Set carry bit if sse-used flag set.
1871         * src/x86/unix64.S (ffi_call_unix64): Add ssecount argument.
1872         Only load sse registers if ssecount non-zero.
1873         (ffi_closure_unix64): Only save sse registers if carry set on entry.
1874
1875 2005-04-29  Ralf Corsepius  <ralf.corsepius@rtems.org>
1876
1877         * configure.ac: Add i*86-*-rtems*, sparc*-*-rtems*,
1878         powerpc-*rtems*, arm*-*-rtems*, sh-*-rtems*.
1879         * configure: Regenerate.
1880
1881 2005-04-20  Hans-Peter Nilsson  <hp@axis.com>
1882
1883         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): In regsub use,
1884         have Tcl8.3-compatible intermediate variable.
1885
1886 2005-04-18  Simon Posnjak <simon.posnjak@siol.net> 
1887             Hans-Peter Nilsson  <hp@axis.com>
1888
1889         * Makefile.am: Add CRIS support.
1890         * configure.ac: Likewise.
1891         * Makefile.in, configure, testsuite/Makefile.in,
1892         include/Makefile.in: Regenerate.
1893         * src/cris: New directory.
1894         * src/cris/ffi.c, src/cris/sysv.S, src/cris/ffitarget.h: New files.
1895         * src/prep_cif.c (ffi_prep_cif): Wrap in #ifndef __CRIS__.
1896
1897         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): Replace \n with
1898         \r?\n in output tests.
1899
1900 2005-04-12  Mike Stump  <mrs@apple.com>
1901
1902         * configure: Regenerate.
1903
1904 2005-03-30  Hans Boehm  <Hans.Boehm@hp.com>
1905
1906         * src/ia64/ffitarget.h (ffi_arg): Use long long instead of DI.
1907         
1908 2005-03-30  Steve Ellcey  <sje@cup.hp.com>
1909
1910         * src/ia64/ffitarget.h (ffi_arg) ADD DI attribute.
1911         (ffi_sarg) Ditto.
1912         * src/ia64/unix.S (ffi_closure_unix): Extend gp
1913         to 64 bits in ILP32 mode.
1914         Load 64 bits even for short data.
1915
1916 2005-03-23  Mike Stump  <mrs@apple.com>
1917
1918         * src/powerpc/darwin.S: Update for -m64 multilib.
1919         * src/powerpc/darwin_closure.S: Likewise.
1920
1921 2005-03-21  Zack Weinberg  <zack@codesourcery.com>
1922
1923         * configure.ac: Do not invoke TL_AC_GCC_VERSION.
1924         Do not set tool_include_dir.
1925         * aclocal.m4, configure, Makefile.in, testsuite/Makefile.in:
1926         Regenerate.
1927         * include/Makefile.am: Set gcc_version and toollibffidir.
1928         * include/Makefile.in: Regenerate.
1929
1930 2005-02-22  Andrew Haley  <aph@redhat.com>
1931
1932         * src/powerpc/ffi.c (ffi_prep_cif_machdep): Bump alignment to
1933         odd-numbered register pairs for 64-bit integer types.
1934
1935 2005-02-23  Andreas Tobler  <a.tobler@schweiz.ch>
1936
1937         PR libffi/20104
1938         * testsuite/libffi.call/return_ll1.c: New test case.
1939
1940 2005-02-11  Janis Johnson  <janis187@us.ibm.com>
1941
1942         * testsuite/libffi.call/cls_align_longdouble.c: Remove dg-options.
1943         * testsuite/libffi.call/float.c: Ditto.
1944         * testsuite/libffi.call/float2.c: Ditto.
1945         * testsuite/libffi.call/float3.c: Ditto.
1946
1947 2005-02-08  Andreas Tobler  <a.tobler@schweiz.ch>
1948
1949         * src/frv/ffitarget.h: Remove PPC stuff which does not belong to frv.
1950
1951 2005-01-12  Eric Botcazou  <ebotcazou@libertysurf.fr>
1952
1953         * testsuite/libffi.special/special.exp (cxx_options): Add
1954         -shared-libgcc.
1955
1956 2004-12-31  Richard Henderson  <rth@redhat.com>
1957
1958         * src/types.c (FFI_AGGREGATE_TYPEDEF): Remove.
1959         (FFI_TYPEDEF): Rename from FFI_INTEGRAL_TYPEDEF.  Replace size and
1960         offset parameters with a type parameter; deduce size and structure
1961         alignment.  Update all users.
1962
1963 2004-12-31  Richard Henderson  <rth@redhat.com>
1964
1965         * src/types.c (FFI_TYPE_POINTER): Define with sizeof.
1966         (FFI_TYPE_LONGDOUBLE): Fix for ia64.
1967         * src/ia64/ffitarget.h (struct ffi_ia64_trampoline_struct): Move
1968         into ffi_prep_closure.
1969         * src/ia64/ia64_flags.h, src/ia64/ffi.c, src/ia64/unix.S: Rewrite
1970         from scratch.
1971
1972 2004-12-27  Richard Henderson  <rth@redhat.com>
1973
1974         * src/x86/unix64.S: Fix typo in unwind info.
1975
1976 2004-12-25  Richard Henderson  <rth@redhat.com>
1977
1978         * src/x86/ffi64.c (struct register_args): Rename from stackLayout.
1979         (enum x86_64_reg_class): Add X86_64_COMPLEX_X87_CLASS.
1980         (merge_classes): Check for it.
1981         (SSE_CLASS_P): New.
1982         (classify_argument): Pass byte_offset by value; perform all updates
1983         inside struct case.
1984         (examine_argument): Add classes argument; handle
1985         X86_64_COMPLEX_X87_CLASS.
1986         (ffi_prep_args): Merge into ...
1987         (ffi_call): ... here.  Share stack frame with ffi_call_unix64.
1988         (ffi_prep_cif_machdep): Setup cif->flags for proper structure return.
1989         (ffi_fill_return_value): Remove.
1990         (ffi_prep_closure): Remove dead assert.
1991         (ffi_closure_unix64_inner): Rename from ffi_closure_UNIX64_inner.
1992         Rewrite to use struct register_args instead of va_list.  Create
1993         flags for handling structure returns.
1994         * src/x86/unix64.S: Remove dead strings.
1995         (ffi_call_unix64): Rename from ffi_call_UNIX64.  Rewrite to share
1996         stack frame with ffi_call.  Handle structure returns properly.
1997         (float2sse, floatfloat2sse, double2sse): Remove.
1998         (sse2float, sse2double, sse2floatfloat): Remove.
1999         (ffi_closure_unix64): Rename from ffi_closure_UNIX64.  Rewrite
2000         to handle structure returns properly.
2001
2002 2004-12-08  David Edelsohn  <edelsohn@gnu.org>
2003
2004         * Makefile.am (AM_MAKEFLAGS): Remove duplicate LIBCFLAGS and
2005         PICFLAG.
2006         * Makefile.in: Regenerated.
2007
2008 2004-12-02  Richard Sandiford  <rsandifo@redhat.com>
2009
2010         * configure.ac: Use TL_AC_GCC_VERSION to set gcc_version.
2011         * configure, aclocal.m4, Makefile.in: Regenerate.
2012         * include/Makefile.in, testsuite/Makefile.in: Regenerate.
2013
2014 2004-11-29  Kelley Cook  <kcook@gcc.gnu.org>
2015
2016         * configure: Regenerate for libtool change.
2017
2018 2004-11-25  Kelley Cook  <kcook@gcc.gnu.org>
2019
2020         * configure: Regenerate for libtool reversion.
2021
2022 2004-11-24  Kelley Cook  <kcook@gcc.gnu.org>
2023
2024         * configure: Regenerate for libtool change.
2025
2026 2004-11-23  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2027
2028         * testsuite/lib/libffi-dg.exp: Use new procs in target-libpath.exp.
2029
2030 2004-11-23  Richard Sandiford  <rsandifo@redhat.com>
2031
2032         * src/mips/o32.S (ffi_call_O32, ffi_closure_O32): Use jalr instead
2033         of jal.  Use an absolute encoding for the frame information.
2034
2035 2004-11-23  Kelley Cook  <kcook@gcc.gnu.org>
2036
2037         * Makefile.am: Remove no-dependencies.  Add ACLOCAL_AMFLAGS.
2038         * acinclude.m4: Delete logic for sincludes.
2039         * aclocal.m4, Makefile.in, configure: Regenerate.
2040         * include/Makefile: Likewise.
2041         * testsuite/Makefile: Likewise.
2042
2043 2004-11-22  Eric Botcazou  <ebotcazou@libertysurf.fr>
2044
2045         * src/sparc/ffi.c (ffi_prep_closure): Align doubles and 64-bit integers
2046         on a 8-byte boundary.
2047         * src/sparc/v8.S (ffi_closure_v8): Reserve frame space for arguments.
2048
2049 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2050
2051         * src/arm/ffi.c (ffi_prep_cif_machdep): Handle functions that return
2052         long long values.  Round stack allocation to a multiple of 8 bytes
2053         for ATPCS compatibility.
2054         * src/arm/sysv.S (ffi_call_SYSV): Rework to avoid use of APCS register
2055         names.  Handle returning long long types.  Add Thumb and interworking
2056         support.  Improve soft-float code.
2057
2058 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
2059
2060         * testsuite/lib/libffi-db.exp (load_gcc_lib): New function.
2061         (libffi_exit): New function.
2062         (libffi_init): Build the testglue wrapper if needed.
2063
2064 2004-10-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2065
2066         PR other/18138
2067         * testsuite/lib/libffi-dg.exp: Accept more than one multilib libgcc.
2068
2069 2004-10-25  Kazuhiro Inaoka <inaoka.kazuhiro@renesas.com>
2070
2071         * src/m32r/libffitarget.h (FFI_CLOSURES): Set to 0.
2072
2073 2004-10-20  Kaz Kojima  <kkojima@gcc.gnu.org>
2074
2075         * src/sh/sysv.S (ffi_call_SYSV): Don't align for double data.
2076         * testsuite/libffi.call/float3.c: New test case.
2077
2078 2004-10-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2079
2080         * src/sh/ffi.c (ffi_prep_closure): Set T bit in trampoline for
2081         the function returning a structure pointed with R2.
2082         * src/sh/sysv.S (ffi_closure_SYSV): Use R2 as the pointer to
2083         the structure return value if T bit set.  Emit position
2084         independent code and EH data if PIC.
2085
2086 2004-10-13  Kazuhiro Inaoka  <inaoka.kazuhiro@renesas.com>
2087
2088         * Makefile.am: Add m32r support.
2089         * configure.ac: Likewise.
2090         * Makefile.in: Regenerate.
2091         * confiugre: Regenerate.
2092         * src/types.c: Add m32r port to FFI_INTERNAL_TYPEDEF
2093         (uint64, sint64, double, longdouble)
2094         * src/m32r: New directory.
2095         * src/m32r/ffi.c: New file.
2096         * src/m32r/sysv.S: Likewise.
2097         * src/m32r/ffitarget.h: Likewise.
2098
2099 2004-10-02  Kaz Kojima  <kkojima@gcc.gnu.org>
2100
2101         * testsuite/libffi.call/negint.c: New test case.
2102
2103 2004-09-14  H.J. Lu  <hongjiu.lu@intel.com>
2104
2105         PR libgcj/17465
2106         * testsuite/lib/libffi-dg.exp: Don't use global ld_library_path.
2107         Set up LD_LIBRARY_PATH, SHLIB_PATH, LD_LIBRARYN32_PATH,
2108         LD_LIBRARY64_PATH, LD_LIBRARY_PATH_32, LD_LIBRARY_PATH_64 and
2109         DYLD_LIBRARY_PATH.
2110
2111 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2112
2113         * testsuite/libffi.call/many_win32.c: Remove whitespaces.
2114         * testsuite/libffi.call/promotion.c: Likewise.
2115         * testsuite/libffi.call/return_ll.c: Remove unused var. Cleanup
2116         whitespaces.
2117         * testsuite/libffi.call/return_sc.c: Likewise.
2118         * testsuite/libffi.call/return_uc.c: Likewise.
2119
2120 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
2121
2122         * src/powerpc/darwin.S: Fix comments and identation.
2123         * src/powerpc/darwin_closure.S: Likewise.
2124
2125 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2126
2127         * src/powerpc/ffi_darwin.c: Add flag for longdouble return values.
2128         (ffi_prep_args): Handle longdouble arguments.
2129         (ffi_prep_cif_machdep): Set flags for longdouble. Calculate space for
2130         longdouble.
2131         (ffi_closure_helper_DARWIN): Add closure handling for longdouble.
2132         * src/powerpc/darwin.S (_ffi_call_DARWIN): Add handling of longdouble
2133         values.
2134         * src/powerpc/darwin_closure.S (_ffi_closure_ASM): Likewise.
2135         * src/types.c: Defined longdouble size and alignment for darwin.
2136
2137 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
2138
2139         * src/powerpc/aix.S: Remove whitespaces.
2140         * src/powerpc/aix_closure.S: Likewise.
2141         * src/powerpc/asm.h: Likewise.
2142         * src/powerpc/ffi.c: Likewise.
2143         * src/powerpc/ffitarget.h: Likewise.
2144         * src/powerpc/linux64.S: Likewise.
2145         * src/powerpc/linux64_closure.S: Likewise.
2146         * src/powerpc/ppc_closure.S: Likewise.
2147         * src/powerpc/sysv.S: Likewise.
2148
2149 2004-08-30  Anthony Green  <green@redhat.com>
2150
2151         * Makefile.am: Add frv support.
2152         * Makefile.in, testsuite/Makefile.in: Rebuilt.
2153         * configure.ac: Read configure.host.
2154         * configure.in: Read configure.host.
2155         * configure.host: New file.  frv-elf needs libgloss.
2156         * include/ffi.h.in: Force ffi_closure to have a nice big (8)
2157         alignment.  This is needed to frv and shouldn't harm the others.
2158         * include/ffi_common.h (ALIGN_DOWN): New macro.
2159         * src/frv/ffi.c, src/frv/ffitarget.h, src/frv/eabi.S: New files.
2160
2161 2004-08-24  David Daney  <daney@avtrex.com>
2162
2163         * testsuite/libffi.call/closure_fn0.c: Xfail mips64* instead of mips*.
2164         * testsuite/libffi.call/closure_fn1.c: Likewise.
2165         * testsuite/libffi.call/closure_fn2.c  Likewise.
2166         * testsuite/libffi.call/closure_fn3.c: Likewise.
2167         * testsuite/libffi.call/closure_fn4.c: Likewise.
2168         * testsuite/libffi.call/closure_fn5.c: Likewise.
2169         * testsuite/libffi.call/cls_18byte.c: Likewise.
2170         * testsuite/libffi.call/cls_19byte.c: Likewise.
2171         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2172         * testsuite/libffi.call/cls_20byte.c: Likewise.
2173         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2174         * testsuite/libffi.call/cls_24byte.c: Likewise.
2175         * testsuite/libffi.call/cls_2byte.c: Likewise.
2176         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2177         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2178         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2179         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2180         * testsuite/libffi.call/cls_4byte.c: Likewise.
2181         * testsuite/libffi.call/cls_64byte.c: Likewise.
2182         * testsuite/libffi.call/cls_6byte.c: Likewise.
2183         * testsuite/libffi.call/cls_7byte.c: Likewise.
2184         * testsuite/libffi.call/cls_8byte.c: Likewise.
2185         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2186         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2187         * testsuite/libffi.call/cls_align_double.c: Likewise.
2188         * testsuite/libffi.call/cls_align_float.c: Likewise.
2189         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2190         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2191         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2192         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2193         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2194         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2195         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2196         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
2197         * testsuite/libffi.call/cls_double.c: Likewise.
2198         * testsuite/libffi.call/cls_float.c: Likewise.
2199         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2200         * testsuite/libffi.call/cls_multi_sshort.c: Likewise.
2201         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2202         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2203         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2204         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2205         * testsuite/libffi.call/cls_schar.c: Likewise.
2206         * testsuite/libffi.call/cls_sint.c: Likewise.
2207         * testsuite/libffi.call/cls_sshort.c: Likewise.
2208         * testsuite/libffi.call/cls_uchar.c: Likewise.
2209         * testsuite/libffi.call/cls_uint.c: Likewise.
2210         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2211         * testsuite/libffi.call/cls_ushort.c: Likewise.
2212         * testsuite/libffi.call/nested_struct.c: Likewise.
2213         * testsuite/libffi.call/nested_struct1.c: Likewise.
2214         * testsuite/libffi.call/nested_struct2.c: Likewise.
2215         * testsuite/libffi.call/nested_struct3.c: Likewise.
2216         * testsuite/libffi.call/problem1.c: Likewise.
2217         * testsuite/libffi.special/unwindtest.cc: Likewise.
2218         * testsuite/libffi.call/cls_12byte.c: Likewise and set return value
2219         to zero.
2220         * testsuite/libffi.call/cls_16byte.c: Likewise.
2221         * testsuite/libffi.call/cls_5byte.c: Likewise.
2222
2223 2004-08-23  David Daney <daney@avtrex.com>
2224
2225         PR libgcj/13141
2226         * src/mips/ffitarget.h (FFI_O32_SOFT_FLOAT): New ABI.
2227         * src/mips/ffi.c (ffi_prep_args): Fix alignment calculation.
2228         (ffi_prep_cif_machdep): Handle FFI_O32_SOFT_FLOAT floating point
2229         parameters and return types.
2230         (ffi_call): Handle FFI_O32_SOFT_FLOAT ABI.
2231         (ffi_prep_closure): Ditto.
2232         (ffi_closure_mips_inner_O32): Handle FFI_O32_SOFT_FLOAT ABI, fix
2233         alignment calculations.
2234         * src/mips/o32.S (ffi_closure_O32): Don't use floating point
2235         instructions if FFI_O32_SOFT_FLOAT, make stack frame ABI compliant.
2236
2237 2004-08-14  Casey Marshall <csm@gnu.org>
2238
2239         * src/mips/ffi.c (ffi_pref_cif_machdep): set `cif->flags' to
2240         contain `FFI_TYPE_UINT64' as return type for any 64-bit
2241         integer (O32 ABI only).
2242         (ffi_prep_closure): new function.
2243         (ffi_closure_mips_inner_O32): new function.
2244         * src/mips/ffitarget.h: Define `FFI_CLOSURES' and
2245         `FFI_TRAMPOLINE_SIZE' appropriately if the ABI is o32.
2246         * src/mips/o32.S (ffi_call_O32): add labels for .eh_frame. Return
2247         64 bit integers correctly.
2248         (ffi_closure_O32): new function.
2249         Added DWARF-2 unwind info for both functions.
2250
2251 2004-08-10  Andrew Haley  <aph@redhat.com>
2252
2253         * src/x86/ffi64.c (ffi_prep_args ): 8-align all stack arguments.
2254
2255 2004-08-01  Robert Millan  <robertmh@gnu.org>
2256
2257         * configure.ac: Detect knetbsd-gnu and kfreebsd-gnu.
2258         * configure: Regenerate.
2259
2260 2004-07-30  Maciej W. Rozycki  <macro@linux-mips.org>
2261
2262         * acinclude.m4 (AC_FUNC_MMAP_BLACKLIST): Check for <sys/mman.h>
2263         and mmap() explicitly instead of relying on preset autoconf cache
2264         variables.
2265         * aclocal.m4: Regenerate.
2266         * configure: Regenerate.
2267
2268 2004-07-11  Ulrich Weigand  <uweigand@de.ibm.com>
2269
2270         * src/s390/ffi.c (ffi_prep_args): Fix C aliasing violation.
2271         (ffi_check_float_struct): Remove unused prototype.
2272
2273 2004-06-30  Geoffrey Keating  <geoffk@apple.com>
2274
2275         * src/powerpc/ffi_darwin.c (flush_icache): ';' is a comment
2276         character on Darwin, use '\n\t' instead.
2277
2278 2004-06-26  Matthias Klose  <doko@debian.org>
2279
2280         * libtool-version: Fix typo in revision/age.
2281
2282 2004-06-17  Matthias Klose  <doko@debian.org>
2283
2284         * libtool-version: New.
2285         * Makefile.am (libffi_la_LDFLAGS): Use -version-info for soname.
2286         * Makefile.in: Regenerate.
2287
2288 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2289
2290         * Makefile.am: Remove useless multilib rules.
2291         * Makefile.in: Regenerate.
2292         * aclocal.m4: Regenerate with automake 1.8.5.
2293         * configure.ac: Remove useless multilib configury.
2294         * configure: Regenerate.
2295
2296 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
2297
2298         * .cvsignore: New file.
2299
2300 2004-06-10  Jakub Jelinek  <jakub@redhat.com>
2301
2302         * src/ia64/unix.S (ffi_call_unix): Insert group barrier break
2303         fp_done.
2304         (ffi_closure_UNIX): Fix f14/f15 adjustment if FLOAT_SZ is ever
2305         changed from 8.
2306
2307 2004-06-06  Sean McNeil  <sean@mcneil.com>
2308
2309         * configure.ac: Add x86_64-*-freebsd* support.
2310         * configure: Regenerate.
2311
2312 2004-04-26  Joe Buck <jbuck@welsh-buck.org>
2313
2314         Bug 15093
2315         * configure.ac: Test for existence of mmap and sys/mman.h before
2316         checking blacklist.  Fix suggested by Jim Wilson.
2317         * configure: Regenerate.
2318
2319 2004-04-26  Matt Austern  <austern@apple.com>
2320
2321         * src/powerpc/darwin.S: Go through a non-lazy pointer for initial
2322         FDE location.
2323         * src/powerpc/darwin_closure.S: Likewise.
2324
2325 2004-04-24  Andreas Tobler  <a.tobler@schweiz.ch>
2326
2327         * testsuite/libffi.call/cls_multi_schar.c (main): Fix initialization
2328         error. Reported by Thomas Heller <theller@python.net>.
2329         * testsuite/libffi.call/cls_multi_sshort.c (main): Likewise.
2330         * testsuite/libffi.call/cls_multi_ushort.c (main): Likewise.
2331
2332 2004-03-20  Matthias Klose  <doko@debian.org>
2333
2334         * src/pa/linux.S: Fix typo.
2335
2336 2004-03-19  Matthias Klose  <doko@debian.org>
2337
2338         * Makefile.am: Update.
2339         * Makefile.in: Regenerate.
2340         * src/pa/ffi.h.in: Remove.
2341         * src/pa/ffitarget.h: New file.
2342
2343 2004-02-10  Randolph Chung  <tausq@debian.org>
2344
2345         * Makefile.am: Add PA support.
2346         * Makefile.in: Regenerate.
2347         * include/Makefile.in: Regenerate.
2348         * configure.ac: Add PA target.
2349         * configure: Regenerate.
2350         * src/pa/ffi.c: New file.
2351         * src/pa/ffi.h.in: Add PA support.
2352         * src/pa/linux.S: New file.
2353         * prep_cif.c: Add PA support.
2354
2355 2004-03-16  Hosaka Yuji  <hos@tamanegi.org>
2356
2357         * src/types.c: Fix alignment size of X86_WIN32 case int64 and
2358         double.
2359         * src/x86/ffi.c (ffi_prep_args): Replace ecif->cif->rtype->type
2360         with ecif->cif->flags.
2361         (ffi_call, ffi_prep_incoming_args_SYSV): Replace cif->rtype->type
2362         with cif->flags.
2363         (ffi_prep_cif_machdep): Add X86_WIN32 struct case.
2364         (ffi_closure_SYSV): Add 1 or 2-bytes struct case for X86_WIN32.
2365         * src/x86/win32.S (retstruct1b, retstruct2b, sc_retstruct1b,
2366         sc_retstruct2b): Add for 1 or 2-bytes struct case.
2367
2368 2004-03-15 Kelley Cook <kcook@gcc.gnu.org>
2369
2370         * configure.in: Rename file to ...
2371         * configure.ac: ... this.
2372         * fficonfig.h.in: Regenerate.
2373         * Makefile.in: Regenerate.
2374         * include/Makefile.in: Regenerate.
2375         * testsuite/Makefile.in: Regenerate.
2376
2377 2004-03-12  Matt Austern  <austern@apple.com>
2378
2379         * src/powerpc/darwin.S: Fix EH information so it corresponds to
2380         changes in EH format resulting from addition of linkonce support.
2381         * src/powerpc/darwin_closure.S: Likewise.
2382
2383 2004-03-11  Andreas Tobler  <a.tobler@schweiz.ch>
2384             Paolo Bonzini  <bonzini@gnu.org>
2385
2386         * Makefile.am (AUTOMAKE_OPTIONS): Set them.
2387         Remove VPATH. Remove rules for object files. Remove multilib support.
2388         (AM_CCASFLAGS): Add.
2389         * configure.in (AC_CONFIG_HEADERS): Relace AM_CONFIG_HEADER.
2390         (AC_PREREQ): Bump version to 2.59.
2391         (AC_INIT): Fill with version info and bug address.
2392         (ORIGINAL_LD_FOR_MULTILIBS): Remove.
2393         (AM_ENABLE_MULTILIB): Use this instead of AC_ARG_ENABLE.
2394         De-precious CC so that the right flags are passed down to multilibs.
2395         (AC_MSG_ERROR): Replace obsolete macro AC_ERROR.
2396         (AC_CONFIG_FILES): Replace obsolete macro AC_LINK_FILES.
2397         (AC_OUTPUT): Reorganize the output with AC_CONFIG_COMMANDS.
2398         * configure: Rebuilt.
2399         * aclocal.m4: Likewise.
2400         * Makefile.in, include/Makefile.in, testsuite/Makefile.in: Likewise.
2401         * fficonfig.h.in: Likewise.
2402
2403 2004-03-11  Andreas Schwab  <schwab@suse.de>
2404
2405         * src/ia64/ffi.c (ffi_prep_incoming_args_UNIX): Get floating point
2406         arguments from fp registers only for the first 8 parameter slots.
2407         Don't convert a float parameter when passed in memory.
2408
2409 2004-03-09  Hans-Peter Nilsson  <hp@axis.com>
2410
2411         * configure: Regenerate for config/accross.m4 correction.
2412
2413 2004-02-25  Matt Kraai  <kraai@alumni.cmu.edu>
2414
2415         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Change
2416         ecif->cif->bytes to bytes.
2417         (ffi_prep_cif_machdep): Add braces around nested if statement.
2418
2419 2004-02-09  Alan Modra  <amodra@bigpond.net.au>
2420
2421         * src/types.c (pointer): POWERPC64 has 8 byte pointers.
2422
2423         * src/powerpc/ffi.c (ffi_prep_args64): Correct long double handling.
2424         (ffi_closure_helper_LINUX64): Fix typo.
2425         * testsuite/libffi.call/cls_align_longdouble.c: Pass -mlong-double-128
2426         for powerpc64-*-*.
2427         * testsuite/libffi.call/float.c: Likewise.
2428         * testsuite/libffi.call/float2.c: Likewise.
2429
2430 2004-02-08  Alan Modra  <amodra@bigpond.net.au>
2431
2432         * src/powerpc/ffi.c (ffi_prep_cif_machdep <FFI_LINUX64>): Correct
2433         long double function return and long double arg handling.
2434         (ffi_closure_helper_LINUX64): Formatting.  Delete unused "ng" var.
2435         Use "end_pfr" instead of "nf".  Correct long double handling.
2436         Localise "temp".
2437         * src/powerpc/linux64.S (ffi_call_LINUX64): Save f2 long double
2438         return value.
2439         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Allocate
2440         space for long double return value.  Adjust stack frame and offsets.
2441         Load f2 long double return.
2442
2443 2004-02-07  Alan Modra  <amodra@bigpond.net.au>
2444
2445         * src/types.c: Use 16 byte long double for POWERPC64.
2446
2447 2004-01-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
2448
2449         * src/sparc/ffi.c (ffi_prep_args_v9): Shift the parameter array
2450         when the structure return address is passed in %o0.
2451         (ffi_V9_return_struct): Rename into ffi_v9_layout_struct.
2452         (ffi_v9_layout_struct): Align the field following a nested structure
2453         on a word boundary.  Use memmove instead of memcpy.
2454         (ffi_call): Update call to ffi_V9_return_struct.
2455         (ffi_prep_closure): Define 'ctx' only for V8.
2456         (ffi_closure_sparc_inner): Clone into ffi_closure_sparc_inner_v8
2457         and ffi_closure_sparc_inner_v9.
2458         (ffi_closure_sparc_inner_v8): Return long doubles by reference.
2459         Always skip the structure return address.  For structures and long
2460         doubles, copy the argument directly.
2461         (ffi_closure_sparc_inner_v9): Skip the structure return address only
2462         if required.  Shift the maximum floating-point slot accordingly.  For
2463         big structures, copy the argument directly; otherwise, left-justify the
2464         argument and call ffi_v9_layout_struct to lay out the structure on
2465         the stack.
2466         * src/sparc/v8.S: Undef STACKFRAME before defining it.
2467         (ffi_closure_v8): Pass the structure return address.  Update call to
2468         ffi_closure_sparc_inner_v8.  Short-circuit FFI_TYPE_INT handling.
2469         Skip the 'unimp' insn when returning long doubles and structures.
2470         * src/sparc/v9.S: Undef STACKFRAME before defining it.
2471         (ffi_closure_v9): Increase the frame size by 2 words.  Short-circuit
2472         FFI_TYPE_INT handling.  Load structures both in integers and
2473         floating-point registers on return.
2474         * README: Update status of the SPARC port.
2475
2476 2004-01-24  Andreas Tobler  <a.tobler@schweiz.ch>
2477
2478         * testsuite/libffi.call/pyobjc-tc.c (main): Treat result value
2479         as of type ffi_arg.
2480         * testsuite/libffi.call/struct3.c (main): Fix CHECK.
2481
2482 2004-01-22  Ulrich Weigand  <uweigand@de.ibm.com>
2483
2484         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Treat result
2485         value as of type ffi_arg, not unsigned int.
2486
2487 2004-01-21  Michael Ritzert  <ritzert@t-online.de>
2488
2489         * ffi64.c (ffi_prep_args): Cast the RHS of an assignment instead
2490         of the LHS.
2491
2492 2004-01-12  Andreas Tobler  <a.tobler@schweiz.ch>
2493
2494         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_32 for
2495         Solaris.
2496
2497 2004-01-08  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2498
2499         * testsuite/libffi.call/ffitest.h (allocate_mmap): Cast MAP_FAILED
2500         to void *.
2501
2502 2003-12-10  Richard Henderson  <rth@redhat.com>
2503
2504         * testsuite/libffi.call/cls_align_pointer.c: Cast pointers to
2505         size_t instead of int.
2506
2507 2003-12-04  Hosaka Yuji  <hos@tamanegi.org>
2508
2509         * testsuite/libffi.call/many_win32.c: Include <float.h>.
2510         * testsuite/libffi.call/many_win32.c (main): Replace variable
2511         int i with unsigned long ul.
2512
2513         * testsuite/libffi.call/cls_align_uint64.c: New test case.
2514         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
2515         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
2516         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
2517         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
2518         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
2519         * testsuite/libffi.call/cls_align_float.c: Likewise.
2520         * testsuite/libffi.call/cls_align_double.c: Likewise.
2521         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
2522         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
2523
2524 2003-12-02  Hosaka Yuji  <hos@tamanegi.org>
2525
2526         PR other/13221
2527         * src/x86/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV):
2528         Align arguments to 32 bits.
2529
2530 2003-12-01  Andreas Tobler  <a.tobler@schweiz.ch>
2531
2532         PR other/13221
2533         * testsuite/libffi.call/cls_multi_sshort.c: New test case.
2534         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
2535         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
2536         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
2537         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
2538         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
2539
2540         * testsuite/libffi.special/unwindtest.cc: Cosmetics.
2541
2542 2003-11-26  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
2543
2544         * testsuite/libffi.call/ffitest.h: Include <fcntl.h>.
2545         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2546
2547 2003-11-22  Andreas Tobler  <a.tobler@schweiz.ch>
2548
2549         * Makefile.in: Rebuilt.
2550         * configure: Likewise.
2551         * testsuite/libffi.special/unwindtest.cc: Convert the mmap to
2552         the right type.
2553
2554 2003-11-21  Andreas Jaeger  <aj@suse.de>
2555             Andreas Tobler  <a.tobler@schweiz.ch>
2556
2557         * acinclude.m4: Add AC_FUNC_MMAP_BLACKLIST.
2558         * configure.in: Call AC_FUNC_MMAP_BLACKLIST.
2559         * Makefile.in: Rebuilt.
2560         * aclocal.m4: Likewise.
2561         * configure: Likewise.
2562         * fficonfig.h.in: Likewise.
2563         * testsuite/lib/libffi-dg.exp: Add include dir.
2564         * testsuite/libffi.call/ffitest.h: Add MMAP definitions.
2565         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2566         * testsuite/libffi.call/closure_fn0.c: Use MMAP functionality
2567         for ffi_closure if available.
2568         * testsuite/libffi.call/closure_fn1.c: Likewise.
2569         * testsuite/libffi.call/closure_fn2.c: Likewise.
2570         * testsuite/libffi.call/closure_fn3.c: Likewise.
2571         * testsuite/libffi.call/closure_fn4.c: Likewise.
2572         * testsuite/libffi.call/closure_fn5.c: Likewise.
2573         * testsuite/libffi.call/cls_12byte.c: Likewise.
2574         * testsuite/libffi.call/cls_16byte.c: Likewise.
2575         * testsuite/libffi.call/cls_18byte.c: Likewise.
2576         * testsuite/libffi.call/cls_19byte.c: Likewise.
2577         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2578         * testsuite/libffi.call/cls_20byte.c: Likewise.
2579         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2580         * testsuite/libffi.call/cls_24byte.c: Likewise.
2581         * testsuite/libffi.call/cls_2byte.c: Likewise.
2582         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2583         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2584         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2585         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2586         * testsuite/libffi.call/cls_4byte.c: Likewise.
2587         * testsuite/libffi.call/cls_5byte.c: Likewise.
2588         * testsuite/libffi.call/cls_64byte.c: Likewise.
2589         * testsuite/libffi.call/cls_6byte.c: Likewise.
2590         * testsuite/libffi.call/cls_7byte.c: Likewise.
2591         * testsuite/libffi.call/cls_8byte.c: Likewise.
2592         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2593         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2594         * testsuite/libffi.call/cls_double.c: Likewise.
2595         * testsuite/libffi.call/cls_float.c: Likewise.
2596         * testsuite/libffi.call/cls_schar.c: Likewise.
2597         * testsuite/libffi.call/cls_sint.c: Likewise.
2598         * testsuite/libffi.call/cls_sshort.c: Likewise.
2599         * testsuite/libffi.call/cls_uchar.c: Likewise.
2600         * testsuite/libffi.call/cls_uint.c: Likewise.
2601         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2602         * testsuite/libffi.call/cls_ushort.c: Likewise.
2603         * testsuite/libffi.call/nested_struct.c: Likewise.
2604         * testsuite/libffi.call/nested_struct1.c: Likewise.
2605         * testsuite/libffi.call/nested_struct2.c: Likewise.
2606         * testsuite/libffi.call/nested_struct3.c: Likewise.
2607         * testsuite/libffi.call/problem1.c: Likewise.
2608         * testsuite/libffi.special/unwindtest.cc: Likewise.
2609
2610 2003-11-20  Andreas Tobler  <a.tobler@schweiz.ch>
2611
2612         * testsuite/lib/libffi-dg.exp: Make the -lgcc_s conditional.
2613
2614 2003-11-19  Andreas Tobler  <a.tobler@schweiz.ch>
2615
2616         * testsuite/lib/libffi-dg.exp: Add DYLD_LIBRARY_PATH for darwin.
2617         Add -lgcc_s to additional flags.
2618
2619 2003-11-12  Andreas Tobler  <a.tobler@schweiz.ch>
2620
2621         * configure.in, include/Makefile.am: PR libgcj/11147, install
2622         the ffitarget.h header file in a gcc versioned and target
2623         dependent place.
2624         * configure: Regenerated.
2625         * Makefile.in, include/Makefile.in: Likewise.
2626         * testsuite/Makefile.in: Likewise.
2627
2628 2003-11-09  Andreas Tobler  <a.tobler@schweiz.ch>
2629
2630         * testsuite/libffi.call/closure_fn0.c: Print result and check
2631         with dg-output to make debugging easier.
2632         * testsuite/libffi.call/closure_fn1.c: Likewise.
2633         * testsuite/libffi.call/closure_fn2.c: Likewise.
2634         * testsuite/libffi.call/closure_fn3.c: Likewise.
2635         * testsuite/libffi.call/closure_fn4.c: Likewise.
2636         * testsuite/libffi.call/closure_fn5.c: Likewise.
2637         * testsuite/libffi.call/cls_12byte.c: Likewise.
2638         * testsuite/libffi.call/cls_16byte.c: Likewise.
2639         * testsuite/libffi.call/cls_18byte.c: Likewise.
2640         * testsuite/libffi.call/cls_19byte.c: Likewise.
2641         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2642         * testsuite/libffi.call/cls_20byte.c: Likewise.
2643         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2644         * testsuite/libffi.call/cls_24byte.c: Likewise.
2645         * testsuite/libffi.call/cls_2byte.c: Likewise.
2646         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2647         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2648         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2649         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2650         * testsuite/libffi.call/cls_4byte.c: Likewise.
2651         * testsuite/libffi.call/cls_5byte.c: Likewise.
2652         * testsuite/libffi.call/cls_64byte.c: Likewise.
2653         * testsuite/libffi.call/cls_6byte.c: Likewise.
2654         * testsuite/libffi.call/cls_7byte.c: Likewise.
2655         * testsuite/libffi.call/cls_8byte.c: Likewise.
2656         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2657         * testsuite/libffi.call/cls_9byte2.c: Likewise.
2658         * testsuite/libffi.call/cls_double.c: Likewise.
2659         * testsuite/libffi.call/cls_float.c: Likewise.
2660         * testsuite/libffi.call/cls_schar.c: Likewise.
2661         * testsuite/libffi.call/cls_sint.c: Likewise.
2662         * testsuite/libffi.call/cls_sshort.c: Likewise.
2663         * testsuite/libffi.call/cls_uchar.c: Likewise.
2664         * testsuite/libffi.call/cls_uint.c: Likewise.
2665         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2666         * testsuite/libffi.call/cls_ushort.c: Likewise.
2667         * testsuite/libffi.call/problem1.c: Likewise.
2668
2669         * testsuite/libffi.special/unwindtest.cc: Make ffi_closure
2670         static.
2671
2672 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
2673
2674         * testsuite/libffi.call/cls_9byte2.c: New test case.
2675         * testsuite/libffi.call/cls_9byte1.c: Likewise.
2676         * testsuite/libffi.call/cls_64byte.c: Likewise.
2677         * testsuite/libffi.call/cls_20byte1.c: Likewise.
2678         * testsuite/libffi.call/cls_19byte.c: Likewise.
2679         * testsuite/libffi.call/cls_18byte.c: Likewise.
2680         * testsuite/libffi.call/closure_fn4.c: Likewise.
2681         * testsuite/libffi.call/closure_fn5.c: Likewise.
2682         * testsuite/libffi.call/cls_schar.c: Likewise.
2683         * testsuite/libffi.call/cls_sint.c: Likewise.
2684         * testsuite/libffi.call/cls_sshort.c: Likewise.
2685         * testsuite/libffi.call/nested_struct2.c: Likewise.
2686         * testsuite/libffi.call/nested_struct3.c: Likewise.
2687
2688 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
2689
2690         * testsuite/libffi.call/cls_double.c: Do a check on the result.
2691         * testsuite/libffi.call/cls_uchar.c: Likewise.
2692         * testsuite/libffi.call/cls_uint.c: Likewise.
2693         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2694         * testsuite/libffi.call/cls_ushort.c: Likewise.
2695         * testsuite/libffi.call/return_sc.c: Cleanup whitespaces.
2696
2697 2003-11-06  Andreas Tobler  <a.tobler@schweiz.ch>
2698
2699         * src/prep_cif.c (ffi_prep_cif): Move the validity check after
2700         the initialization.
2701
2702 2003-10-23  Andreas Tobler  <a.tobler@schweiz.ch>
2703
2704         * src/java_raw_api.c (ffi_java_ptrarray_to_raw): Replace
2705         FFI_ASSERT(FALSE) with FFI_ASSERT(0).
2706
2707 2003-10-22  David Daney  <ddaney@avtrex.com>
2708
2709         * src/mips/ffitarget.h: Replace undefined UINT32 and friends with
2710         __attribute__((__mode__(__SI__))) and friends.
2711
2712 2003-10-22  Andreas Schwab  <schwab@suse.de>
2713
2714         * src/ia64/ffi.c: Replace FALSE/TRUE with false/true.
2715
2716 2003-10-21  Andreas Tobler  <a.tobler@schweiz.ch>
2717
2718         * configure.in: AC_LINK_FILES(ffitarget.h).
2719         * configure: Regenerate.
2720         * Makefile.in: Likewise.
2721         * include/Makefile.in: Likewise.
2722         * testsuite/Makefile.in: Likewise.
2723         * fficonfig.h.in: Likewise.
2724
2725 2003-10-21  Paolo Bonzini  <bonzini@gnu.org>
2726             Richard Henderson  <rth@redhat.com>
2727
2728         Avoid that ffi.h includes fficonfig.h.
2729
2730         * Makefile.am (EXTRA_DIST): Include ffitarget.h files
2731         (TARGET_SRC_MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
2732         (TARGET_SRC_MIPS_SGI): Removed.
2733         (MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
2734         (MIPS_SGI): Removed.
2735         (CLEANFILES): Removed.
2736         (mostlyclean-am, clean-am, mostlyclean-sub, clean-sub): New
2737         targets.
2738         * acconfig.h: Removed.
2739         * configure.in: Compute sizeofs only for double and long double.
2740         Use them to define and subst HAVE_LONG_DOUBLE.  Include comments
2741         into AC_DEFINE instead of using acconfig.h.  Create
2742         include/ffitarget.h instead of include/fficonfig.h.  Rename
2743         MIPS_GCC to MIPS_IRIX, drop MIPS_SGI since we are in gcc's tree.
2744         AC_DEFINE EH_FRAME_FLAGS.
2745         * include/Makefile.am (DISTCLEANFILES): New automake macro.
2746         (hack_DATA): Add ffitarget.h.
2747         * include/ffi.h.in: Remove all system specific definitions.
2748         Declare raw API even if it is not installed, why bother?
2749         Use limits.h instead of SIZEOF_* to define ffi_type_*.  Do
2750         not define EH_FRAME_FLAGS, it is in fficonfig.h now.  Include
2751         ffitarget.h instead of fficonfig.h.  Remove ALIGN macro.
2752         (UINT_ARG, INT_ARG): Removed, use ffi_arg and ffi_sarg instead.
2753         * include/ffi_common.h (bool): Do not define.
2754         (ffi_assert): Accept failed assertion.
2755         (ffi_type_test): Return void and accept file/line.
2756         (FFI_ASSERT): Pass stringized failed assertion.
2757         (FFI_ASSERT_AT): New macro.
2758         (FFI_ASSERT_VALID_TYPE): New macro.
2759         (UINT8, SINT8, UINT16, SINT16, UINT32, SINT32,
2760         UINT64, SINT64): Define here with gcc's __attribute__ macro
2761         instead of in ffi.h
2762         (FLOAT32, ALIGN): Define here instead of in ffi.h
2763         * include/ffi-mips.h: Removed.  Its content moved to
2764         src/mips/ffitarget.h after separating assembly and C sections.
2765         * src/alpha/ffi.c, src/alpha/ffi.c, src/java_raw_api.c
2766         src/prep_cif.c, src/raw_api.c, src/ia64/ffi.c,
2767         src/mips/ffi.c, src/mips/n32.S, src/mips/o32.S,
2768         src/mips/ffitarget.h, src/sparc/ffi.c, src/x86/ffi64.c:
2769         SIZEOF_ARG -> FFI_SIZEOF_ARG.
2770         * src/ia64/ffi.c: Include stdbool.h (provided by GCC 2.95+).
2771         * src/debug.c (ffi_assert): Accept stringized failed assertion.
2772         (ffi_type_test): Rewritten.
2773         * src/prep-cif.c (initialize_aggregate, ffi_prep_cif): Call
2774         FFI_ASSERT_VALID_TYPE.
2775         * src/alpha/ffitarget.h, src/arm/ffitarget.h,
2776         src/ia64/ffitarget.h, src/m68k/ffitarget.h,
2777         src/mips/ffitarget.h, src/powerpc/ffitarget.h,
2778         src/s390/ffitarget.h, src/sh/ffitarget.h,
2779         src/sh64/ffitarget.h, src/sparc/ffitarget.h,
2780         src/x86/ffitarget.h: New files.
2781         * src/alpha/osf.S, src/arm/sysv.S, src/ia64/unix.S,
2782         src/m68k/sysv.S, src/mips/n32.S, src/mips/o32.S,
2783         src/powerpc/aix.S, src/powerpc/darwin.S,
2784         src/powerpc/ffi_darwin.c, src/powerpc/linux64.S,
2785         src/powerpc/linux64_closure.S, src/powerpc/ppc_closure.S,
2786         src/powerpc/sysv.S, src/s390/sysv.S, src/sh/sysv.S,
2787         src/sh64/sysv.S, src/sparc/v8.S, src/sparc/v9.S,
2788         src/x86/sysv.S, src/x86/unix64.S, src/x86/win32.S:
2789         include fficonfig.h
2790
2791 2003-10-20  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2792
2793         * src/mips/ffi.c: Use _ABIN32, _ABIO32 instead of external
2794         _MIPS_SIM_NABI32, _MIPS_SIM_ABI32.
2795
2796 2003-10-19  Andreas Tobler  <a.tobler@schweiz.ch>
2797
2798         * src/powerpc/ffi_darwin.c (ffi_prep_args): Declare bytes again.
2799         Used when FFI_DEBUG = 1.
2800
2801 2003-10-14  Alan Modra  <amodra@bigpond.net.au>
2802
2803         * src/types.c (double, longdouble): Default POWERPC64 to 8 byte size
2804         and align.
2805
2806 2003-10-06  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2807
2808         * include/ffi_mips.h: Define FFI_MIPS_N32 for N32/N64 ABIs,
2809         FFI_MIPS_O32 for O32 ABI.
2810
2811 2003-10-01  Andreas Tobler  <a.tobler@schweiz.ch>
2812
2813         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_64 for
2814         SPARC64. Cleanup whitespaces.
2815
2816 2003-09-19  Andreas Tobler  <a.tobler@schweiz.ch>
2817
2818         * testsuite/libffi.call/closure_fn0.c: Xfail mips, arm,
2819         strongarm, xscale. Cleanup whitespaces.
2820         * testsuite/libffi.call/closure_fn1.c: Likewise.
2821         * testsuite/libffi.call/closure_fn2.c: Likewise.
2822         * testsuite/libffi.call/closure_fn3.c: Likewise.
2823         * testsuite/libffi.call/cls_12byte.c: Likewise.
2824         * testsuite/libffi.call/cls_16byte.c: Likewise.
2825         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2826         * testsuite/libffi.call/cls_20byte.c: Likewise.
2827         * testsuite/libffi.call/cls_24byte.c: Likewise.
2828         * testsuite/libffi.call/cls_2byte.c: Likewise.
2829         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2830         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2831         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2832         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2833         * testsuite/libffi.call/cls_4byte.c: Likewise.
2834         * testsuite/libffi.call/cls_5byte.c: Likewise.
2835         * testsuite/libffi.call/cls_6byte.c: Likewise.
2836         * testsuite/libffi.call/cls_7byte.c: Likewise.
2837         * testsuite/libffi.call/cls_8byte.c: Likewise.
2838         * testsuite/libffi.call/cls_double.c: Likewise.
2839         * testsuite/libffi.call/cls_float.c: Likewise.
2840         * testsuite/libffi.call/cls_uchar.c: Likewise.
2841         * testsuite/libffi.call/cls_uint.c: Likewise.
2842         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2843         * testsuite/libffi.call/cls_ushort.c: Likewise.
2844         * testsuite/libffi.call/nested_struct.c: Likewise.
2845         * testsuite/libffi.call/nested_struct1.c: Likewise.
2846         * testsuite/libffi.call/problem1.c: Likewise.
2847         * testsuite/libffi.special/unwindtest.cc: Likewise.
2848         * testsuite/libffi.call/pyobjc-tc.c: Cleanup whitespaces.
2849
2850 2003-09-18  David Edelsohn  <edelsohn@gnu.org>
2851
2852         * src/powerpc/aix.S: Cleanup whitespaces.
2853         * src/powerpc/aix_closure.S: Likewise.
2854
2855 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
2856
2857         * src/powerpc/darwin.S: Cleanup whitespaces, comment formatting.
2858         * src/powerpc/darwin_closure.S: Likewise.
2859         * src/powerpc/ffi_darwin.c: Likewise.
2860
2861 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
2862             David Edelsohn  <edelsohn@gnu.org>
2863
2864         * src/types.c (double): Add AIX and Darwin to the right TYPEDEF.
2865         * src/powerpc/aix_closure.S: Remove the pointer to the outgoing
2866         parameter stack.
2867         * src/powerpc/darwin_closure.S: Likewise.
2868         * src/powerpc/ffi_darwin.c (ffi_prep_args): Handle structures
2869         according to the Darwin/AIX ABI.
2870         (ffi_prep_cif_machdep): Likewise.
2871         (ffi_closure_helper_DARWIN): Likewise.
2872         Remove the outgoing parameter stack logic. Simplify the evaluation
2873         of the different CASE types.
2874         (ffi_prep_clousure): Avoid the casts on lvalues. Change the branch
2875         statement in the trampoline code.
2876
2877 2003-09-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2878
2879         * src/sh/ffi.c (ffi_prep_args): Take account into the alignement
2880         for the register size.
2881         (ffi_closure_helper_SYSV): Handle the structure return value
2882         address correctly.
2883         (ffi_closure_helper_SYSV): Return the appropriate type when
2884         the registers are used for the structure return value.
2885         * src/sh/sysv.S (ffi_closure_SYSV): Fix the stack layout for
2886         the 64-bit return value.  Update copyright years.
2887
2888 2003-09-17  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2889
2890         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Search in
2891         srcdir for ffi_mips.h.
2892
2893 2003-09-12  Alan Modra  <amodra@bigpond.net.au>
2894
2895         * src/prep_cif.c (initialize_aggregate): Include tail padding in
2896         structure size.
2897         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Correct
2898         placement of float result.
2899         * testsuite/libffi.special/unwindtest.cc (closure_test_fn1): Correct
2900         cast of "resp" for big-endian 64 bit machines.
2901
2902 2003-09-11  Alan Modra  <amodra@bigpond.net.au>
2903
2904         * src/types.c (double, longdouble): Merge identical SH and ARM
2905         typedefs, and add POWERPC64.
2906         * src/powerpc/ffi.c (ffi_prep_args64): Correct next_arg calc for
2907         struct split over gpr and rest.
2908         (ffi_prep_cif_machdep): Correct intarg_count for structures.
2909         * src/powerpc/linux64.S (ffi_call_LINUX64): Fix gpr offsets.
2910
2911 2003-09-09  Andreas Tobler  <a.tobler@schweiz.ch>
2912
2913         * src/powerpc/ffi.c (ffi_closure_helper_SYSV) Handle struct
2914         passing correctly.
2915
2916 2003-09-09  Alan Modra  <amodra@bigpond.net.au>
2917
2918         * configure: Regenerate.
2919
2920 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
2921
2922         * Makefile.am: Remove build rules for ffitest.
2923         * Makefile.in: Rebuilt.
2924
2925 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
2926
2927         * src/java_raw_api.c: Include <stdlib.h> to fix compiler warning
2928         about implicit declaration of abort().
2929
2930 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
2931
2932         * Makefile.am: Add dejagnu test framework. Fixes PR other/11411.
2933         * Makefile.in: Rebuilt.
2934         * configure.in: Add dejagnu test framework.
2935         * configure: Rebuilt.
2936
2937         * testsuite/Makefile.am: New file.
2938         * testsuite/Makefile.in: Built
2939         * testsuite/lib/libffi-dg.exp: New file.
2940         * testsuite/config/default.exp: Likewise.
2941         * testsuite/libffi.call/call.exp: Likewise.
2942         * testsuite/libffi.call/ffitest.h: Likewise.
2943         * testsuite/libffi.call/closure_fn0.c: Likewise.
2944         * testsuite/libffi.call/closure_fn1.c: Likewise.
2945         * testsuite/libffi.call/closure_fn2.c: Likewise.
2946         * testsuite/libffi.call/closure_fn3.c: Likewise.
2947         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2948         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2949         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2950         * testsuite/libffi.call/cls_2byte.c: Likewise.
2951         * testsuite/libffi.call/cls_3byte1.c: Likewise.
2952         * testsuite/libffi.call/cls_3byte2.c: Likewise.
2953         * testsuite/libffi.call/cls_4byte.c: Likewise.
2954         * testsuite/libffi.call/cls_5byte.c: Likewise.
2955         * testsuite/libffi.call/cls_6byte.c: Likewise.
2956         * testsuite/libffi.call/cls_7byte.c: Likewise.
2957         * testsuite/libffi.call/cls_8byte.c: Likewise.
2958         * testsuite/libffi.call/cls_12byte.c: Likewise.
2959         * testsuite/libffi.call/cls_16byte.c: Likewise.
2960         * testsuite/libffi.call/cls_20byte.c: Likewise.
2961         * testsuite/libffi.call/cls_24byte.c: Likewise.
2962         * testsuite/libffi.call/cls_double.c: Likewise.
2963         * testsuite/libffi.call/cls_float.c: Likewise.
2964         * testsuite/libffi.call/cls_uchar.c: Likewise.
2965         * testsuite/libffi.call/cls_uint.c: Likewise.
2966         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2967         * testsuite/libffi.call/cls_ushort.c: Likewise.
2968         * testsuite/libffi.call/float.c: Likewise.
2969         * testsuite/libffi.call/float1.c: Likewise.
2970         * testsuite/libffi.call/float2.c: Likewise.
2971         * testsuite/libffi.call/many.c: Likewise.
2972         * testsuite/libffi.call/many_win32.c: Likewise.
2973         * testsuite/libffi.call/nested_struct.c: Likewise.
2974         * testsuite/libffi.call/nested_struct1.c: Likewise.
2975         * testsuite/libffi.call/pyobjc-tc.c: Likewise.
2976         * testsuite/libffi.call/problem1.c: Likewise.
2977         * testsuite/libffi.call/promotion.c: Likewise.
2978         * testsuite/libffi.call/return_ll.c: Likewise.
2979         * testsuite/libffi.call/return_sc.c: Likewise.
2980         * testsuite/libffi.call/return_uc.c: Likewise.
2981         * testsuite/libffi.call/strlen.c: Likewise.
2982         * testsuite/libffi.call/strlen_win32.c: Likewise.
2983         * testsuite/libffi.call/struct1.c: Likewise.
2984         * testsuite/libffi.call/struct2.c: Likewise.
2985         * testsuite/libffi.call/struct3.c: Likewise.
2986         * testsuite/libffi.call/struct4.c: Likewise.
2987         * testsuite/libffi.call/struct5.c: Likewise.
2988         * testsuite/libffi.call/struct6.c: Likewise.
2989         * testsuite/libffi.call/struct7.c: Likewise.
2990         * testsuite/libffi.call/struct8.c: Likewise.
2991         * testsuite/libffi.call/struct9.c: Likewise.
2992         * testsuite/libffi.special/special.exp: New file.
2993         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2994         * testsuite/libffi.special/unwindtest.cc: Likewise.
2995
2996
2997 2003-08-13  Kaz Kojima  <kkojima@gcc.gnu.org>
2998
2999         * src/sh/ffi.c (OFS_INT16): Set 0 for little endian case.  Update
3000         copyright years.
3001
3002 2003-08-02  Alan Modra  <amodra@bigpond.net.au>
3003
3004         * src/powerpc/ffi.c (ffi_prep_args64): Modify for changed gcc
3005         structure passing.
3006         (ffi_closure_helper_LINUX64): Likewise.
3007         * src/powerpc/linux64.S: Remove code writing to parm save area.
3008         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Use return
3009         address in lr from ffi_closure_helper_LINUX64 call to calculate
3010         table address.  Optimize function tail.
3011
3012 2003-07-28  Andreas Tobler  <a.tobler@schweiz.ch>
3013
3014         * src/sparc/ffi.c: Handle all floating point registers.
3015         * src/sparc/v9.S: Likewise. Fixes second part of PR target/11410.
3016
3017 2003-07-11  Gerald Pfeifer  <pfeifer@dbai.tuwien.ac.at>
3018
3019         * README: Note that libffi is not part of GCC.  Update the project
3020         URL and status.
3021
3022 2003-06-19  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3023
3024         * src/powerpc/ppc_closure.S: Include ffi.h.
3025
3026 2003-06-13  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3027
3028         * src/x86/sysv.S: Avoid gas-only .uleb128/.sleb128 directives.
3029         Use C style comments.
3030
3031 2003-06-13  Kaz Kojima  <kkojima@rr.iij4u.or.jp>
3032
3033         * Makefile.am: Add SHmedia support.  Fix a typo of SH support.
3034         * Makefile.in: Regenerate.
3035         * configure.in (sh64-*-linux*, sh5*-*-linux*): Add target.
3036         * configure: Regenerate.
3037         * include/ffi.h.in: Add SHmedia support.
3038         * src/sh64/ffi.c: New file.
3039         * src/sh64/sysv.S: New file.
3040
3041 2003-05-16  Jakub Jelinek  <jakub@redhat.com>
3042
3043         * configure.in (HAVE_RO_EH_FRAME): Check whether .eh_frame section
3044         should be read-only.
3045         * configure: Rebuilt.
3046         * fficonfig.h.in: Rebuilt.
3047         * include/ffi.h.in (EH_FRAME_FLAGS): Define.
3048         * src/alpha/osf.S: Use EH_FRAME_FLAGS.
3049         * src/powerpc/linux64.S: Likewise.
3050         * src/powerpc/linux64_closure.S: Likewise.  Include ffi.h.
3051         * src/powerpc/sysv.S: Use EH_FRAME_FLAGS.  Use pcrel encoding
3052         if -fpic/-fPIC/-mrelocatable.
3053         * src/powerpc/powerpc_closure.S: Likewise.
3054         * src/sparc/v8.S: If HAVE_RO_EH_FRAME is defined, don't include
3055         #write in .eh_frame flags.
3056         * src/sparc/v9.S: Likewise.
3057         * src/x86/unix64.S: Use EH_FRAME_FLAGS.
3058         * src/x86/sysv.S: Likewise.  Use pcrel encoding if -fpic/-fPIC.
3059         * src/s390/sysv.S: Use EH_FRAME_FLAGS.  Include ffi.h.
3060
3061 2003-05-07  Jeff Sturm  <jsturm@one-point.com>
3062
3063         Fixes PR bootstrap/10656
3064         * configure.in (HAVE_AS_REGISTER_PSEUDO_OP): Test assembler
3065         support for .register pseudo-op.
3066         * src/sparc/v8.S: Use it.
3067         * fficonfig.h.in: Rebuilt.
3068         * configure: Rebuilt.
3069
3070 2003-04-18  Jakub Jelinek  <jakub@redhat.com>
3071
3072         * include/ffi.h.in (POWERPC64): Define if 64-bit.
3073         (enum ffi_abi): Add FFI_LINUX64 on POWERPC.
3074         Make it the default on POWERPC64.
3075         (FFI_TRAMPOLINE_SIZE): Define to 24 on POWERPC64.
3076         * configure.in: Change powerpc-*-linux* into powerpc*-*-linux*.
3077         * configure: Rebuilt.
3078         * src/powerpc/ffi.c (hidden): Define.
3079         (ffi_prep_args_SYSV): Renamed from
3080         ffi_prep_args.  Cast pointers to unsigned long to shut up warnings.
3081         (NUM_GPR_ARG_REGISTERS64, NUM_FPR_ARG_REGISTERS64,
3082         ASM_NEEDS_REGISTERS64): New.
3083         (ffi_prep_args64): New function.
3084         (ffi_prep_cif_machdep): Handle FFI_LINUX64 ABI.
3085         (ffi_call): Likewise.
3086         (ffi_prep_closure): Likewise.
3087         (flush_icache): Surround by #ifndef POWERPC64.
3088         (ffi_dblfl): New union type.
3089         (ffi_closure_helper_SYSV): Use it to avoid aliasing problems.
3090         (ffi_closure_helper_LINUX64): New function.
3091         * src/powerpc/ppc_closure.S: Surround whole file by #ifndef
3092         __powerpc64__.
3093         * src/powerpc/sysv.S: Likewise.
3094         (ffi_call_SYSV): Rename ffi_prep_args to ffi_prep_args_SYSV.
3095         * src/powerpc/linux64.S: New file.
3096         * src/powerpc/linux64_closure.S: New file.
3097         * Makefile.am (EXTRA_DIST): Add src/powerpc/linux64.S and
3098         src/powerpc/linux64_closure.S.
3099         (TARGET_SRC_POWERPC): Likewise.
3100
3101         * src/ffitest.c (closure_test_fn, closure_test_fn1, closure_test_fn2,
3102         closure_test_fn3): Fix result printing on big-endian 64-bit
3103         machines.
3104         (main): Print tst2_arg instead of uninitialized tst2_result.
3105
3106         * src/ffitest.c (main): Hide what closure pointer really points to
3107         from the compiler.
3108
3109 2003-04-16  Richard Earnshaw  <rearnsha@arm.com>
3110
3111         * configure.in (arm-*-netbsdelf*): Add configuration.
3112         (configure): Regenerated.
3113
3114 2003-04-04  Loren J. Rittle  <ljrittle@acm.org>
3115
3116         * include/Makefile.in: Regenerate.
3117
3118 2003-03-21  Zdenek Dvorak  <rakdver@atrey.karlin.mff.cuni.cz>
3119
3120         * libffi/include/ffi.h.in: Define X86 instead of X86_64 in 32
3121         bit mode.
3122         * libffi/src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV):
3123         Receive closure pointer through parameter, read args using
3124         __builtin_dwarf_cfa.
3125         (FFI_INIT_TRAMPOLINE): Send closure reference through eax.
3126
3127 2003-03-12  Andreas Schwab  <schwab@suse.de>
3128
3129         * configure.in: Avoid trailing /. in toolexeclibdir.
3130         * configure: Rebuilt.
3131
3132 2003-03-03  Andreas Tobler <a.tobler@schweiz.ch>
3133
3134         * src/powerpc/darwin_closure.S: Recode to fit dynamic libraries.
3135
3136 2003-02-06  Andreas Tobler <a.tobler@schweiz.ch>
3137
3138         * libffi/src/powerpc/darwin_closure.S:
3139         Fix alignement bug, allocate 8 bytes for the result.
3140         * libffi/src/powerpc/aix_closure.S:
3141         Likewise.
3142         * libffi/src/powerpc/ffi_darwin.c:
3143         Update stackframe description for aix/darwin_closure.S.
3144
3145 2003-02-06  Jakub Jelinek  <jakub@redhat.com>
3146
3147         * src/s390/ffi.c (ffi_closure_helper_SYSV): Add hidden visibility
3148         attribute.
3149
3150 2003-01-31  Christian Cornelssen  <ccorn@cs.tu-berlin.de>,
3151             Andreas Schwab  <schwab@suse.de>
3152
3153         * configure.in: Adjust command to source config-ml.in to account
3154         for changes to the libffi_basedir definition.
3155         (libffi_basedir): Remove ${srcdir} from value and include trailing
3156         slash if nonempty.
3157
3158         * configure: Regenerate.
3159
3160 2003-01-29  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
3161
3162         * src/powerpc/ppc_closure.S: Recode to fit shared libs.
3163
3164 2003-01-28  Andrew Haley  <aph@redhat.com>
3165
3166         * include/ffi.h.in: Enable FFI_CLOSURES for x86_64.
3167         * src/x86/ffi64.c (ffi_prep_closure): New.
3168         (ffi_closure_UNIX64_inner): New.
3169         * src/x86/unix64.S (ffi_closure_UNIX64): New.
3170
3171 2003-01-27  Alexandre Oliva  <aoliva@redhat.com>
3172
3173         * configure.in (toolexecdir, toolexeclibdir): Set and AC_SUBST.
3174         Remove USE_LIBDIR conditional.
3175         * Makefile.am (toolexecdir, toolexeclibdir): Don't override.
3176         * Makefile.in, configure: Rebuilt.
3177
3178 2003-01027  David Edelsohn  <edelsohn@gnu.org>
3179
3180         * Makefile.am (TARGET_SRC_POWERPC_AIX): Fix typo.
3181         * Makefile.in: Regenerate.
3182
3183 2003-01-22  Andrew Haley  <aph@redhat.com>
3184
3185         * src/powerpc/darwin.S (_ffi_call_AIX): Add Augmentation size to
3186         unwind info.
3187
3188 2003-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
3189
3190         * src/powerpc/darwin.S: Add unwind info.
3191         * src/powerpc/darwin_closure.S: Likewise.
3192
3193 2003-01-14  Andrew Haley  <aph@redhat.com>
3194
3195         * src/x86/ffi64.c (ffi_prep_args): Check for void retval.
3196         (ffi_prep_cif_machdep): Likewise.
3197         * src/x86/unix64.S: Add unwind info.
3198
3199 2003-01-14  Andreas Jaeger  <aj@suse.de>
3200
3201         * src/ffitest.c (main): Only use ffi_closures if those are
3202         supported.
3203
3204 2003-01-13 Andreas Tobler <a.tobler@schweiz.ch>
3205
3206         * libffi/src/ffitest.c
3207          add closure testcases
3208
3209 2003-01-13 Kevin B. Hendricks <khendricks@ivey.uwo.ca>
3210
3211         * libffi/src/powerpc/ffi.c
3212          fix alignment bug for float (4 byte aligned iso 8 byte)
3213
3214 2003-01-09  Geoffrey Keating  <geoffk@apple.com>
3215
3216         * src/powerpc/ffi_darwin.c: Remove RCS version string.
3217         * src/powerpc/darwin.S: Remove RCS version string.
3218
3219 2003-01-03  Jeff Sturm  <jsturm@one-point.com>
3220
3221         * include/ffi.h.in: Add closure defines for SPARC, SPARC64.
3222         * src/ffitest.c (main): Use static storage for closure.
3223         * src/sparc/ffi.c (ffi_prep_closure, ffi_closure_sparc_inner): New.
3224         * src/sparc/v8.S (ffi_closure_v8): New.
3225         * src/sparc/v9.S (ffi_closure_v9): New.
3226
3227 2002-11-10  Ranjit Mathew <rmathew@hotmail.com>
3228
3229         * include/ffi.h.in: Added FFI_STDCALL ffi_type
3230           enumeration for X86_WIN32.
3231         * src/x86/win32.S: Added ffi_call_STDCALL function
3232           definition.
3233         * src/x86/ffi.c (ffi_call/ffi_raw_call): Added
3234           switch cases for recognising FFI_STDCALL and
3235           calling ffi_call_STDCALL if target is X86_WIN32.
3236         * src/ffitest.c (my_stdcall_strlen/stdcall_many):
3237           stdcall versions of the "my_strlen" and "many"
3238           test functions (for X86_WIN32).
3239           Added test cases to test stdcall invocation using
3240           these functions.
3241
3242 2002-12-02  Kaz Kojima  <kkojima@gcc.gnu.org>
3243
3244         * src/sh/sysv.S: Add DWARF2 unwind info.
3245
3246 2002-11-27  Ulrich Weigand  <uweigand@de.ibm.com>
3247
3248         * src/s390/sysv.S (.eh_frame section): Make section read-only.
3249
3250 2002-11-26  Jim Wilson  <wilson@redhat.com>
3251
3252         * src/types.c (FFI_TYPE_POINTER): Has size 8 on IA64.
3253
3254 2002-11-23  H.J. Lu <hjl@gnu.org>
3255
3256         * acinclude.m4: Add dummy AM_PROG_LIBTOOL.
3257         Include ../config/accross.m4.
3258         * aclocal.m4; Rebuild.
3259         * configure: Likewise.
3260
3261 2002-11-15  Ulrich Weigand  <uweigand@de.ibm.com>
3262
3263         * src/s390/sysv.S (.eh_frame section): Adapt to pcrel FDE encoding.
3264
3265 2002-11-11  DJ Delorie  <dj@redhat.com>
3266
3267         * configure.in: Look for common files in the right place.
3268
3269 2002-10-08  Ulrich Weigand  <uweigand@de.ibm.com>
3270
3271         * src/java_raw_api.c (ffi_java_raw_to_ptrarray): Interpret
3272         raw data as _Jv_word values, not ffi_raw.
3273         (ffi_java_ptrarray_to_raw): Likewise.
3274         (ffi_java_rvalue_to_raw): New function.
3275         (ffi_java_raw_call): Call it.
3276         (ffi_java_raw_to_rvalue): New function.
3277         (ffi_java_translate_args): Call it.
3278         * src/ffitest.c (closure_test_fn): Interpret return value
3279         as ffi_arg, not int.
3280         * src/s390/ffi.c (ffi_prep_cif_machdep): Add missing
3281         FFI_TYPE_POINTER case.
3282         (ffi_closure_helper_SYSV): Likewise.  Also, assume return
3283         values extended to word size.
3284
3285 2002-10-02  Andreas Jaeger  <aj@suse.de>
3286
3287         * src/x86/ffi64.c (ffi_prep_cif_machdep): Remove debug output.
3288
3289 2002-10-01  Bo Thorsen  <bo@smetana.suse.de>
3290
3291         * include/ffi.h.in: Fix i386 win32 compilation.
3292
3293 2002-09-30  Ulrich Weigand  <uweigand@de.ibm.com>
3294
3295         * configure.in: Add s390x-*-linux-* target.
3296         * configure: Regenerate.
3297         * include/ffi.h.in: Define S390X for s390x targets.
3298         (FFI_CLOSURES): Define for s390/s390x.
3299         (FFI_TRAMPOLINE_SIZE): Likewise.
3300         (FFI_NATIVE_RAW_API): Likewise.
3301         * src/prep_cif.c (ffi_prep_cif): Do not compute stack space for s390.
3302         * src/types.c (FFI_TYPE_POINTER): Use 8-byte pointers on s390x.
3303         * src/s390/ffi.c: Major rework of existing code.  Add support for
3304         s390x targets.  Add closure support.
3305         * src/s390/sysv.S: Likewise.
3306
3307 2002-09-29  Richard Earnshaw  <rearnsha@arm.com>
3308
3309         * src/arm/sysv.S: Fix typo.
3310
3311 2002-09-28  Richard Earnshaw  <rearnsha@arm.com>
3312
3313         * src/arm/sysv.S: If we don't have machine/asm.h and the pre-processor
3314         has defined __USER_LABEL_PREFIX__, then use it in CNAME.
3315         (ffi_call_SYSV): Handle soft-float.
3316
3317 2002-09-27  Bo Thorsen  <bo@suse.de>
3318
3319         * include/ffi.h.in: Fix multilib x86-64 support.
3320
3321 2002-09-22  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
3322
3323         * Makefile.am (all-multi): Fix multilib parallel build.
3324
3325 2002-07-19  Kaz Kojima  <kkojima@gcc.gnu.org>
3326
3327         * configure.in (sh[34]*-*-linux*): Add brackets.
3328         * configure: Regenerate.
3329
3330 2002-07-18  Kaz Kojima  <kkojima@gcc.gnu.org>
3331
3332         * Makefile.am: Add SH support.
3333         * Makefile.in: Regenerate.
3334         * configure.in (sh-*-linux*, sh[34]*-*-linux*): Add target.
3335         * configure: Regenerate.
3336         * include/ffi.h.in: Add SH support.
3337         * src/sh/ffi.c: New file.
3338         * src/sh/sysv.S: New file.
3339         * src/types.c: Add SH support.
3340
3341 2002-07-16  Bo Thorsen  <bo@suse.de>
3342
3343         * src/x86/ffi64.c: New file that adds x86-64 support.
3344         * src/x86/unix64.S: New file that handles argument setup for
3345         x86-64.
3346         * src/x86/sysv.S: Don't use this on x86-64.
3347         * src/x86/ffi.c: Don't use this on x86-64.
3348         Remove unused vars.
3349         * src/prep_cif.c (ffi_prep_cif): Don't do stack size calculation
3350         for x86-64.
3351         * src/ffitest.c (struct6): New test that tests a special case in
3352         the x86-64 ABI.
3353         (struct7): Likewise.
3354         (struct8): Likewise.
3355         (struct9): Likewise.
3356         (closure_test_fn): Silence warning about this when it's not used.
3357         (main): Add the new tests.
3358         (main): Fix a couple of wrong casts and silence some compiler warnings.
3359         * include/ffi.h.in: Add x86-64 ABI definition.
3360         * fficonfig.h.in: Regenerate.
3361         * Makefile.am: Add x86-64 support.
3362         * configure.in: Likewise.
3363         * Makefile.in: Regenerate.
3364         * configure: Likewise.
3365
3366 2002-06-24  Bo Thorsen  <bo@suse.de>
3367
3368         * src/types.c: Merge settings for similar architectures.
3369         Add x86-64 sizes and alignments.
3370
3371 2002-06-23  Bo Thorsen  <bo@suse.de>
3372
3373         * src/arm/ffi.c (ffi_prep_args): Remove unused vars.
3374         * src/sparc/ffi.c (ffi_prep_args_v8): Likewise.
3375         * src/mips/ffi.c (ffi_prep_args): Likewise.
3376         * src/m68k/ffi.c (ffi_prep_args): Likewise.
3377
3378 2002-07-18  H.J. Lu  (hjl@gnu.org)
3379
3380         * Makefile.am (TARGET_SRC_MIPS_LINUX): New.
3381         (libffi_la_SOURCES): Support MIPS_LINUX.
3382         (libffi_convenience_la_SOURCES): Likewise.
3383         * Makefile.in: Regenerated.
3384
3385         * configure.in (mips64*-*): Skip.
3386         (mips*-*-linux*): New.
3387         * configure: Regenerated.
3388
3389         * src/mips/ffi.c: Include <sgidefs.h>.
3390
3391 2002-06-06  Ulrich Weigand  <uweigand@de.ibm.com>
3392
3393         * src/s390/sysv.S: Save/restore %r6.  Add DWARF-2 unwind info.
3394
3395 2002-05-27  Roger Sayle  <roger@eyesopen.com>
3396
3397         * src/x86/ffi.c (ffi_prep_args): Remove reference to avn.
3398
3399 2002-05-27  Bo Thorsen  <bo@suse.de>
3400
3401         * src/x86/ffi.c (ffi_prep_args): Remove unused variable and
3402         fix formatting.
3403
3404 2002-05-13  Andreas Tobler  <a.tobler@schweiz.ch>
3405
3406         * src/powerpc/ffi_darwin.c (ffi_prep_closure): Declare fd at
3407         beginning of function (for older apple cc).
3408
3409 2002-05-08  Alexandre Oliva  <aoliva@redhat.com>
3410
3411         * configure.in (ORIGINAL_LD_FOR_MULTILIBS): Preserve LD at
3412         script entry, and set LD to it when configuring multilibs.
3413         * configure: Rebuilt.
3414
3415 2002-05-05  Jason Thorpe  <thorpej@wasabisystems.com>
3416
3417         * configure.in (sparc64-*-netbsd*): Add target.
3418         (sparc-*-netbsdelf*): Likewise.
3419         * configure: Regenerate.
3420
3421 2002-04-28  David S. Miller  <davem@redhat.com>
3422
3423         * configure.in, configure: Fix SPARC test in previous change.
3424
3425 2002-04-29  Gerhard Tonn  <GerhardTonn@swol.de>
3426
3427         * Makefile.am: Add Linux for S/390 support.
3428         * Makefile.in: Regenerate.
3429         * configure.in: Add Linux for S/390 support.
3430         * configure: Regenerate.
3431         * include/ffi.h.in: Add Linux for S/390 support.
3432         * src/s390/ffi.c: New file from libffi CVS tree.
3433         * src/s390/sysv.S: New file from libffi CVS tree.
3434
3435 2002-04-28  Jakub Jelinek  <jakub@redhat.com>
3436
3437