med
[libffi.git] / ChangeLog
1 2012-10-31  Alan Modra  <amodra@gmail.com>
2
3         * src/powerpc/linux64_closure.S: Add new ABI support.
4         * src/powerpc/linux64.S: Likewise.
5
6 2012-10-30  Magnus Granberg  <zorry@gentoo.org>
7             Pavel Labushev  <pavel.labushev@runbox.ru>
8
9         * configure.ac: New options pax_emutramp
10         * configure, fficonfig.h.in: Regenerated
11         * src/closures.c: New function emutramp_enabled_check() and
12         checks.
13
14 2012-10-30  Frederick Cheung  <frederick.cheung@gmail.com>
15
16         * configure.ac: Enable FFI_MAP_EXEC_WRIT for Darwin 12 (mountain
17         lion) and future version.
18         * configure: Rebuild.
19
20 2012-10-30  James Greenhalgh  <james.greenhalgh at arm.com>
21             Marcus Shawcroft  <marcus.shawcroft at arm.com>
22
23         * README: Add details of aarch64 port.
24         * src/aarch64/ffi.c: New.
25         * src/aarch64/ffitarget.h: Likewise.
26         * src/aarch64/sysv.S: Likewise.
27         * Makefile.am: Support aarch64.
28         * configure.ac: Support aarch64.
29         * Makefile.in, configure: Rebuilt.
30
31 2012-10-30  James Greenhalgh  <james.greenhalgh at arm.com>
32             Marcus Shawcroft  <marcus.shawcroft at arm.com>
33
34         * testsuite/lib/libffi.exp: Add support for aarch64.
35         * testsuite/libffi.call/cls_struct_va1.c: New.
36         * testsuite/libffi.call/cls_uchar_va.c: Likewise.
37         * testsuite/libffi.call/cls_uint_va.c: Likewise.
38         * testsuite/libffi.call/cls_ulong_va.c: Likewise.
39         * testsuite/libffi.call/cls_ushort_va.c: Likewise.
40         * testsuite/libffi.call/nested_struct11.c: Likewise.
41         * testsuite/libffi.call/uninitialized.c: Likewise.
42         * testsuite/libffi.call/va_1.c: Likewise.
43         * testsuite/libffi.call/va_struct1.c: Likewise.
44         * testsuite/libffi.call/va_struct2.c: Likewise.
45         * testsuite/libffi.call/va_struct3.c: Likewise.
46
47 2012-10-12  Walter Lee  <walt@tilera.com>
48
49         * Makefile.am: Add TILE-Gx/TILEPro support.
50         * configure.ac: Likewise.
51         * Makefile.in: Regenerate.
52         * configure: Likewise.
53         * src/prep_cif.c (ffi_prep_cif_core): Handle TILE-Gx/TILEPro.
54         * src/tile: New directory.
55         * src/tile/ffi.c: New file.
56         * src/tile/ffitarget.h: Ditto.
57         * src/tile/tile.S: Ditto.
58
59 2012-10-12  Matthias Klose  <doko@ubuntu.com>
60
61         * generate-osx-source-and-headers.py: Normalize whitespace.
62
63 2012-09-14  David Edelsohn  <dje.gcc@gmail.com>
64
65         * configure: Regenerated.
66
67 2012-08-26  Andrew Pinski  <apinski@cavium.com>
68
69         PR libffi/53014
70         * src/mips/ffi.c (ffi_prep_closure_loc): Allow n32 with soft-float and n64 with
71         soft-float.
72
73 2012-08-08  Uros Bizjak  <ubizjak@gmail.com>
74
75         * src/s390/ffi.c (ffi_prep_closure_loc): Don't ASSERT ABI test,
76         just return FFI_BAD_ABI when things are wrong.
77
78 2012-07-18  H.J. Lu  <hongjiu.lu@intel.com>
79
80         PR libffi/53982
81         PR libffi/53973
82         * src/x86/ffitarget.h: Check __ILP32__ instead of __LP64__ for x32.
83         (FFI_SIZEOF_JAVA_RAW): Defined to 4 for x32.
84
85 2012-05-16  H.J. Lu  <hongjiu.lu@intel.com>
86
87         * configure: Regenerated.
88
89 2012-05-05  Nicolas Lelong
90
91         * libffi.xcodeproj/project.pbxproj: Fixes.
92         * README: Update for iOS builds.
93
94 2012-04-23  Alexandre Keunecke I. de Mendonca <alexandre.keunecke@gmail.com>
95
96         * configure.ac: Add Blackfin/sysv support
97         * Makefile.am: Add Blackfin/sysv support
98         * src/bfin/ffi.c:  Add Blackfin/sysv support
99         * src/bfin/ffitarget.h: Add Blackfin/sysv support
100
101 2012-04-11  Anthony Green  <green@moxielogic.com>
102
103         * Makefile.am (EXTRA_DIST): Add new script.
104         * Makefile.in: Rebuilt.
105
106 2012-04-11  Zachary Waldowski  <zwaldowski@gmail.com>
107
108         * generate-ios-source-and-headers.py,
109         libffi.xcodeproj/project.pbxproj: Support a Mac static library via
110         Xcode. Set iOS compatibility to 4.0.  Move iOS trampoline
111         generation into an Xcode "run script" phase.  Include both as
112         Xcode build scripts. Don't always regenerate config files.
113
114 2012-04-10  Anthony Green  <green@moxielogic.com>
115
116         * src/powerpc/ffi_darwin.c (ffi_prep_args): Add missing semicolon.
117
118 2012-04-06  Anthony Green  <green@moxielogic.com>
119
120         * Makefile.am (EXTRA_DIST): Add new iOS/xcode files.
121         * Makefile.in: Rebuilt.
122
123 2012-04-06  Mike Lewis  <mikelikespie@gmail.com>
124
125         * generate-ios-source-and-headers.py: New file.
126         * libffi.xcodeproj/project.pbxproj: New file.
127         * README: Update instructions on building iOS binary.
128         * build-ios.sh: Delete.
129
130 2012-04-06  Anthony Green  <green@moxielogic.com>
131
132         * src/x86/ffi64.c (UINT128): Define differently for Intel and GNU
133         compilers, then use it.
134
135 2012-04-06  H.J. Lu  <hongjiu.lu@intel.com>
136
137         * m4/libtool.m4 (_LT_ENABLE_LOCK): Support x32.
138
139 2012-04-06  Anthony Green  <green@moxielogic.com>
140
141         * testsuite/Makefile.am (EXTRA_DIST): Add missing test cases.
142         * testsuite/Makefile.in: Rebuilt.
143
144 2012-04-05  Zachary Waldowski  <zwaldowski@gmail.com>
145
146         * include/ffi.h.in: Add missing trampoline table fields.
147         * src/arm/sysv.S: Fix ENTRY definition, and wrap symbol references
148         in CNAME.
149         * src/x86/ffi.c: Wrap Windows specific code in ifdefs.
150
151 2012-04-02  Peter Bergner  <bergner@vnet.ibm.com>
152
153         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Declare double_tmp.
154         Silence casting pointer to integer of different size warning.
155         Delete goto to previously deleted label.
156         (ffi_call): Silence possibly undefined warning.
157         (ffi_closure_helper_SYSV): Declare variable type.
158
159 2012-04-02  Peter Rosin  <peda@lysator.liu.se>
160
161         * src/x86/win32.S (ffi_call_win32): Sign/zero extend the return
162         value in the Intel version as is already done for the AT&T version.
163         (ffi_closure_SYSV): Likewise.
164         (ffi_closure_raw_SYSV): Likewise.
165         (ffi_closure_STDCALL): Likewise.
166
167 2012-03-29  Peter Rosin  <peda@lysator.liu.se>
168
169         * src/x86/win32.S (ffi_closure_raw_THISCALL): Unify the frame
170         generation, fix the ENDP label and remove the surplus third arg
171         from the 'lea' insn.
172
173 2012-03-29  Peter Rosin  <peda@lysator.liu.se>
174
175         * src/x86/win32.S (ffi_closure_raw_SYSV): Make the 'stubraw' label
176         visible outside the PROC, so that ffi_closure_raw_THISCALL can see
177         it.  Also instruct the assembler to add a frame to the function.
178
179 2012-03-23  Peter Rosin  <peda@lysator.liu.se>
180
181         * Makefile.am (AM_CPPFLAGS): Add -DFFI_BUILDING.
182         * Makefile.in: Rebuilt.
183         * include/ffi.h.in [MSVC]: Add __declspec(dllimport) decorations
184         to all data exports, when compiling libffi clients using MSVC.
185
186 2012-03-29  Peter Rosin  <peda@lysator.liu.se>
187
188         * src/x86/ffitarget.h (ffi_abi): Add new ABI FFI_MS_CDECL and
189         make it the default for MSVC.
190         (FFI_TYPE_MS_STRUCT): New structure return convention.
191         * src/x86/ffi.c (ffi_prep_cif_machdep): Tweak the structure
192         return convention for FFI_MS_CDECL to be FFI_TYPE_MS_STRUCT
193         instead of an ordinary FFI_TYPE_STRUCT.
194         (ffi_prep_args): Treat FFI_TYPE_MS_STRUCT as FFI_TYPE_STRUCT.
195         (ffi_call): Likewise.
196         (ffi_prep_incoming_args_SYSV): Likewise.
197         (ffi_raw_call): Likewise.
198         (ffi_prep_closure_loc): Treat FFI_MS_CDECL as FFI_SYSV.
199         * src/x86/win32.S (ffi_closure_SYSV): For FFI_TYPE_MS_STRUCT,
200         return a pointer to the result structure in eax and don't pop
201         that pointer from the stack, the caller takes care of it.
202         (ffi_call_win32): Treat FFI_TYPE_MS_STRUCT as FFI_TYPE_STRUCT.
203         (ffi_closure_raw_SYSV): Likewise.
204
205 2012-03-22  Peter Rosin  <peda@lysator.liu.se>
206
207         * testsuite/libffi.call/closure_stdcall.c [MSVC]: Add inline
208         assembly version with Intel syntax.
209         * testsuite/libffi.call/closure_thiscall.c [MSVC]: Likewise.
210
211 2012-03-23  Peter Rosin  <peda@lysator.liu.se>
212
213         * testsuite/libffi.call/ffitest.h: Provide abstration of
214         __attribute__((fastcall)) in the form of a __FASTCALL__
215         define.  Define it to __fastcall for MSVC.
216         * testsuite/libffi.call/fastthis1_win32.c: Use the above.
217         * testsuite/libffi.call/fastthis2_win32.c: Likewise.
218         * testsuite/libffi.call/fastthis3_win32.c: Likewise.
219         * testsuite/libffi.call/strlen2_win32.c: Likewise.
220         * testsuite/libffi.call/struct1_win32.c: Likewise.
221         * testsuite/libffi.call/struct2_win32.c: Likewise.
222
223 2012-03-22  Peter Rosin  <peda@lysator.liu.se>
224
225         * src/x86/win32.S [MSVC] (ffi_closure_THISCALL): Remove the manual
226         frame on function entry, MASM adds one automatically.
227
228 2012-03-22  Peter Rosin  <peda@lysator.liu.se>
229
230         * testsuite/libffi.call/ffitest.h [MSVC]: Add kludge for missing
231         bits in the MSVC headers.
232
233 2012-03-22  Peter Rosin  <peda@lysator.liu.se>
234
235         * testsuite/libffi.call/cls_12byte.c: Adjust to the C89 style
236         with no declarations after statements.
237         * testsuite/libffi.call/cls_16byte.c: Likewise.
238         * testsuite/libffi.call/cls_18byte.c: Likewise.
239         * testsuite/libffi.call/cls_19byte.c: Likewise.
240         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
241         * testsuite/libffi.call/cls_20byte.c: Likewise.
242         * testsuite/libffi.call/cls_20byte1.c: Likewise.
243         * testsuite/libffi.call/cls_24byte.c: Likewise.
244         * testsuite/libffi.call/cls_2byte.c: Likewise.
245         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
246         * testsuite/libffi.call/cls_3byte1.c: Likewise.
247         * testsuite/libffi.call/cls_3byte2.c: Likewise.
248         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
249         * testsuite/libffi.call/cls_4byte.c: Likewise.
250         * testsuite/libffi.call/cls_5_1_byte.c: Likewise.
251         * testsuite/libffi.call/cls_5byte.c: Likewise.
252         * testsuite/libffi.call/cls_64byte.c: Likewise.
253         * testsuite/libffi.call/cls_6_1_byte.c: Likewise.
254         * testsuite/libffi.call/cls_6byte.c: Likewise.
255         * testsuite/libffi.call/cls_7_1_byte.c: Likewise.
256         * testsuite/libffi.call/cls_7byte.c: Likewise.
257         * testsuite/libffi.call/cls_8byte.c: Likewise.
258         * testsuite/libffi.call/cls_9byte1.c: Likewise.
259         * testsuite/libffi.call/cls_9byte2.c: Likewise.
260         * testsuite/libffi.call/cls_align_double.c: Likewise.
261         * testsuite/libffi.call/cls_align_float.c: Likewise.
262         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
263         * testsuite/libffi.call/cls_align_longdouble_split.c: Likewise.
264         * testsuite/libffi.call/cls_align_longdouble_split2.c: Likewise.
265         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
266         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
267         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
268         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
269         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
270         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
271         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
272         * testsuite/libffi.call/cls_dbls_struct.c: Likewise.
273         * testsuite/libffi.call/cls_pointer_stack.c: Likewise.
274         * testsuite/libffi.call/err_bad_typedef.c: Likewise.
275         * testsuite/libffi.call/huge_struct.c: Likewise.
276         * testsuite/libffi.call/nested_struct.c: Likewise.
277         * testsuite/libffi.call/nested_struct1.c: Likewise.
278         * testsuite/libffi.call/nested_struct10.c: Likewise.
279         * testsuite/libffi.call/nested_struct2.c: Likewise.
280         * testsuite/libffi.call/nested_struct3.c: Likewise.
281         * testsuite/libffi.call/nested_struct4.c: Likewise.
282         * testsuite/libffi.call/nested_struct5.c: Likewise.
283         * testsuite/libffi.call/nested_struct6.c: Likewise.
284         * testsuite/libffi.call/nested_struct7.c: Likewise.
285         * testsuite/libffi.call/nested_struct8.c: Likewise.
286         * testsuite/libffi.call/nested_struct9.c: Likewise.
287         * testsuite/libffi.call/stret_large.c: Likewise.
288         * testsuite/libffi.call/stret_large2.c: Likewise.
289         * testsuite/libffi.call/stret_medium.c: Likewise.
290         * testsuite/libffi.call/stret_medium2.c: Likewise.
291         * testsuite/libffi.call/struct1.c: Likewise.
292         * testsuite/libffi.call/struct1_win32.c: Likewise.
293         * testsuite/libffi.call/struct2.c: Likewise.
294         * testsuite/libffi.call/struct2_win32.c: Likewise.
295         * testsuite/libffi.call/struct3.c: Likewise.
296         * testsuite/libffi.call/struct4.c: Likewise.
297         * testsuite/libffi.call/struct5.c: Likewise.
298         * testsuite/libffi.call/struct6.c: Likewise.
299         * testsuite/libffi.call/struct7.c: Likewise.
300         * testsuite/libffi.call/struct8.c: Likewise.
301         * testsuite/libffi.call/struct9.c: Likewise.
302         * testsuite/libffi.call/testclosure.c: Likewise.
303
304 2012-03-21  Peter Rosin  <peda@lysator.liu.se>
305
306         * testsuite/libffi.call/float_va.c (float_va_fn): Use %f when
307         printing doubles (%lf is for long doubles).
308         (main): Likewise.
309
310 2012-03-21  Peter Rosin  <peda@lysator.liu.se>
311
312         * testsuite/lib/target-libpath.exp [*-*-cygwin*, *-*-mingw*]
313         (set_ld_library_path_env_vars): Add the library search dir to PATH
314         (and save PATH for later).
315         (restore_ld_library_path_env_vars): Restore PATH.
316
317 2012-03-20  Peter Rosin  <peda@lysator.liu.se>
318
319         * testsuite/libffi.call/strlen2_win32.c (main): Remove bug.
320         * src/x86/win32.S [MSVC] (ffi_closure_SYSV): Make the 'stub' label
321         visible outside the PROC, so that ffi_closure_THISCALL can see it.
322
323 2012-03-20  Peter Rosin  <peda@lysator.liu.se>
324
325         * testsuite/libffi.call/strlen2_win32.c (main): Remove bug.
326         * src/x86/win32.S [MSVC] (ffi_closure_SYSV): Make the 'stub' label
327         visible outside the PROC, so that ffi_closure_THISCALL can see it.
328
329 2012-03-19  Alan Hourihane  <alanh@fairlite.co.uk>
330
331         * src/m68k/ffi.c: Add MINT support.
332         * src/m68k/sysv.S: Ditto.
333
334 2012-03-06  Chung-Lin Tang  <cltang@codesourcery.com>
335
336         * src/arm/ffi.c (ffi_call): Add __ARM_EABI__ guard around call to
337         ffi_call_VFP().
338         (ffi_prep_closure_loc): Add __ARM_EABI__ guard around use of
339         ffi_closure_VFP.
340         * src/arm/sysv.S: Add __ARM_EABI__ guard around VFP code.
341
342 2012-03-19  chennam  <csit@axway.com>
343
344         * src/powerpc/ffi_darwin.c (ffi_prep_closure_loc): Fix AIX closure
345         support.
346
347 2012-03-13  Kaz Kojima  <kkojima@gcc.gnu.org>
348
349         * src/sh/ffi.c (ffi_prep_closure_loc): Don't ASSERT ABI test,
350         just return FFI_BAD_ABI when things are wrong.
351         * src/sh64/ffi.c (ffi_prep_closure_loc): Ditto.
352
353 2012-03-09  David Edelsohn  <dje.gcc@gmail.com>
354
355         * src/powerpc/aix_closure.S (ffi_closure_ASM): Adjust for Darwin64
356         change to return value of ffi_closure_helper_DARWIN and load type
357         from return type.
358
359 2012-03-03  H.J. Lu  <hongjiu.lu@intel.com>
360
361         * src/x86/ffi64.c (ffi_call): Cast the return value to unsigned
362         long.
363         (ffi_prep_closure_loc): Cast to 64bit address in trampoline.
364         (ffi_closure_unix64_inner): Cast return pointer to unsigned long
365         first.
366
367         * src/x86/ffitarget.h (FFI_SIZEOF_ARG): Defined to 8 for x32.
368         (ffi_arg): Set to unsigned long long for x32.
369         (ffi_sarg): Set to long long for x32.
370
371 2012-03-03  H.J. Lu  <hongjiu.lu@intel.com>
372
373         * src/prep_cif.c (ffi_prep_cif_core): Properly check bad ABI.
374
375 2012-03-03  Andoni Morales Alastruey  <ylatuya@gmail.com>
376
377         * configure.ac: Add -no-undefined for both 32- and 64-bit x86
378         windows-like hosts.
379         * configure: Rebuilt.
380
381 2012-02-27  Mikael Pettersson  <mikpe@it.uu.se>
382
383         PR libffi/52223
384         * Makefile.am (FLAGS_TO_PASS): Define.
385         * Makefile.in: Regenerate.
386
387 2012-02-23  Anthony Green  <green@moxielogic.com>
388
389         * src/*/ffitarget.h: Ensure that users never include ffitarget.h
390         directly.
391
392 2012-02-23  Kai Tietz  <ktietz@redhat.com>
393
394         PR libffi/52221
395         * src/x86/ffi.c (ffi_closure_raw_THISCALL): New
396         prototype.
397         (ffi_prep_raw_closure_loc): Use ffi_closure_raw_THISCALL for
398         thiscall-convention.
399         (ffi_raw_call): Use ffi_prep_args_raw.
400         * src/x86/win32.S (ffi_closure_raw_THISCALL): Add
401         implementation for stub.
402
403 2012-02-10  Kai Tietz  <ktietz@redhat.com>
404
405         * configure.ac (AM_LTLDFLAGS): Add -no-undefine for x64
406         windows target.
407         * configure: Regenerated.
408
409 2012-02-08  Kai Tietz  <ktietz@redhat.com>
410
411         * src/prep_cif.c (ffi_prep_cif): Allow for X86_WIN32
412         also FFI_THISCALL.
413         * src/x86/ffi.c (ffi_closure_THISCALL): Add prototype.
414         (FFI_INIT_TRAMPOLINE_THISCALL): New trampoline code.
415         (ffi_prep_closure_loc): Add FFI_THISCALL support.
416         * src/x86/ffitarget.h (FFI_TRAMPOLINE_SIZE): Adjust size.
417         * src/x86/win32.S (ffi_closure_THISCALL): New closure code
418         for thiscall-calling convention.
419         * testsuite/libffi.call/closure_thiscall.c: New test.
420
421 2012-01-28  Kai Tietz  <ktietz@redhat.com>
422
423         * src/libffi/src/x86/ffi.c (ffi_call_win32): Add new
424         argument to prototype for specify calling-convention.
425         (ffi_call): Add support for stdcall/thiscall convention.
426         (ffi_prep_args): Likewise.
427         (ffi_raw_call): Likewise.
428         * src/x86/ffitarget.h (ffi_abi): Add FFI_THISCALL and
429         FFI_FASTCALL.
430         * src/x86/win32.S (_ffi_call_win32): Add support for
431         fastcall/thiscall calling-convention calls.
432         * testsuite/libffi.call/fastthis1_win32.c: New test.
433         * testsuite/libffi.call/fastthis2_win32.c: New test.
434         * testsuite/libffi.call/fastthis3_win32.c: New test.
435         * testsuite/libffi.call/strlen2_win32.c: New test.
436         * testsuite/libffi.call/many2_win32.c: New test.
437         * testsuite/libffi.call/struct1_win32.c: New test.
438         * testsuite/libffi.call/struct2_win32.c: New test.
439
440 2012-01-23  Uros Bizjak  <ubizjak@gmail.com>
441
442         * src/alpha/ffi.c (ffi_prep_closure_loc): Check for bad ABI.
443
444 2012-01-23  Anthony Green  <green@moxielogic.com>
445             Chris Young  <cdyoung@ntlworld.com>
446
447         * configure.ac: Add Amiga support.
448         * configure: Rebuilt.
449
450 2012-01-23  Dmitry Nadezhin  <dmitry.nadezhin@gmail.com>
451
452         * include/ffi_common.h (LIKELY, UNLIKELY): Fix definitions.
453
454 2012-01-23  Andreas Schwab  <schwab@linux-m68k.org>
455
456         * src/m68k/sysv.S (ffi_call_SYSV): Properly test for plain
457         mc68000.  Test for __HAVE_68881__ in addition to __MC68881__.
458
459 2012-01-19  Jakub Jelinek  <jakub@redhat.com>
460
461         PR rtl-optimization/48496
462         * src/ia64/ffi.c (ffi_call): Fix up aliasing violations.
463
464 2012-01-09  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
465
466         * configure.ac (i?86-*-*): Set TARGET to X86_64.
467         * configure: Regenerate.
468
469 2011-12-07  Andrew Pinski  <apinski@cavium.com>
470
471         PR libffi/50051
472         * src/mips/n32.S: Add ".set mips4".
473
474 2011-11-21  Andreas Tobler  <andreast@fgznet.ch>
475
476         * configure: Regenerate.
477
478 2011-11-12  David Gilbert <david.gilbert@linaro.org>
479
480         * doc/libffi.texi, include/ffi.h.in, include/ffi_common.h,
481         man/Makefile.am, man/ffi.3, man/ffi_prep_cif.3,
482         man/ffi_prep_cif_var.3, src/arm/ffi.c, src/arm/ffitarget.h,
483         src/cris/ffi.c, src/prep_cif.c,
484         testsuite/libffi.call/cls_double_va.c,
485         testsuite/libffi.call/cls_longdouble_va.c,
486         testsuite/libffi.call/float_va.c: Many changes to support variadic
487         function calls.
488
489 2011-11-12  Kyle Moffett <Kyle.D.Moffett@boeing.com>
490
491         * src/powerpc/ffi.c, src/powerpc/ffitarget.h,
492         src/powerpc/ppc_closure.S, src/powerpc/sysv.S: Many changes for
493         softfloat powerpc variants.
494
495 2011-11-12  Petr Salinger <Petr.Salinger@seznam.cz>
496
497         * configure.ac (FFI_EXEC_TRAMPOLINE_TABLE): Fix kfreebsd support.
498         * configure: Rebuilt.
499
500 2011-11-12  Timothy Wall  <twall@users.sf.net>
501
502         * src/arm/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV): Max
503         alignment of 4 for wince on ARM.
504
505 2011-11-12  Kyle Moffett <Kyle.D.Moffett@boeing.com>
506             Anthony Green <green@moxielogic.com>
507
508         * src/ppc/sysv.S, src/ppc/ffi.c: Remove use of ppc string
509         instructions (not available on some cores, like the PPC440).
510
511 2011-11-12  Kimura Wataru  <kimuraw@i.nifty.jp>
512
513         * m4/ax_enable_builddir: Change from string comparison to numeric
514         comparison for wc output.
515         * configure.ac: Enable FFI_MMAP_EXEC_WRIT for darwin11 aka Mac OS
516         X 10.7.
517         * configure: Rebuilt.
518
519 2011-11-12  Anthony Green  <green@moxielogic.com>
520
521         * Makefile.am (AM_CCASFLAGS): Add -g option to build assembly
522         files with debug info.
523         * Makefile.in: Rebuilt.
524
525 2011-11-12  Jasper Lievisse Adriaanse <jasper@openbsd.org>
526
527         * README: Update list of supported OpenBSD systems.
528
529 2011-11-12  Anthony Green  <green@moxielogic.com>
530
531         * libtool-version: Update.
532         * Makefile.am (nodist_libffi_la_SOURCES): Add src/debug.c if
533         FFI_DEBUG.
534         (libffi_la_SOURCES): Remove src/debug.c
535         (EXTRA_DIST): Add src/debug.c
536         * Makefile.in: Rebuilt.
537         * README: Update for 3.0.11.
538
539 2011-11-10  Richard Henderson  <rth@redhat.com>
540
541         * configure.ac (GCC_AS_CFI_PSEUDO_OP): Use it instead of inline check.
542         * configure, aclocal.m4: Rebuild.
543
544 2011-09-04  Iain Sandoe  <iains@gcc.gnu.org>
545
546         PR libffi/49594
547         * src/powerpc/darwin_closure.S (stubs): Make the stub binding
548         helper reference track the architecture pointer size.
549
550 2011-08-25  Andrew Haley  <aph@redhat.com>
551
552         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Remove hard-coded assembly
553         instructions.
554         * src/arm/sysv.S (ffi_arm_trampoline): Put them here instead.
555
556 2011-07-11  Andrew Haley  <aph@redhat.com>
557
558         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Clear icache.
559
560 2011-06-29  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
561
562         * testsuite/libffi.call/cls_double_va.c: Move PR number to comment.
563         * testsuite/libffi.call/cls_longdouble_va.c: Likewise.
564
565 2011-06-29  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
566
567         PR libffi/46660
568         * testsuite/libffi.call/cls_double_va.c: xfail dg-output on
569         mips-sgi-irix6*.
570         * testsuite/libffi.call/cls_longdouble_va.c: Likewise.
571
572 2011-06-14  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
573
574         * testsuite/libffi.call/huge_struct.c (test_large_fn): Use PRIu8,
575         PRId8 instead of %hhu, %hhd.
576         * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRId8,
577         PRIu8): Define.
578         [__sgi__] (PRId8, PRIu8): Define.
579
580 2011-04-29  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
581
582         * src/alpha/osf.S (UA_SI, FDE_ENCODING, FDE_ENCODE, FDE_ARANGE):
583         Define.
584         Use them to handle ELF vs. ECOFF differences.
585         [__osf__] (_GLOBAL__F_ffi_call_osf): Define.
586
587 2011-03-30  Timothy Wall  <twall@users.sf.net>
588
589         * src/powerpc/darwin.S: Fix unknown FDE encoding.
590         * src/powerpc/darwin_closure.S: ditto.
591
592 2011-02-25  Anthony Green  <green@moxielogic.com>
593
594         * src/powerpc/ffi.c (ffi_prep_closure_loc): Allow for more
595         32-bit ABIs.
596
597 2011-02-15  Anthony Green  <green@moxielogic.com>
598
599         * m4/ax_cc_maxopt.m4: Don't -malign-double or use -ffast-math.
600         * configure: Rebuilt.
601
602 2011-02-13  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
603
604         * configure: Regenerate.
605
606 2011-02-13  Anthony Green  <green@moxielogic.com>
607
608         * include/ffi_common.h (UNLIKELY, LIKELY): Define.
609         * src/x86/ffi64.c (UNLIKELY, LIKELY): Remove definition.
610         * src/prep_cif.c (UNLIKELY, LIKELY): Remove definition.
611
612         * src/prep_cif.c (initialize_aggregate): Convert assertion into
613         FFI_BAD_TYPEDEF return.  Initialize arg size and alignment to 0.
614
615         * src/pa/ffi.c (ffi_prep_closure_loc): Don't ASSERT ABI test,
616         just return FFI_BAD_ABI when things are wrong.
617         * src/arm/ffi.c (ffi_prep_closure_loc): Ditto.
618         * src/powerpc/ffi.c (ffi_prep_closure_loc): Ditto.
619         * src/mips/ffi.c (ffi_prep_closure_loc): Ditto.
620         * src/ia64/ffi.c (ffi_prep_closure_loc): Ditto.
621         * src/avr32/ffi.c (ffi_prep_closure_loc): Ditto.
622
623 2011-02-11  Anthony Green  <green@moxielogic.com>
624
625         * src/sparc/ffi.c (ffi_prep_closure_loc): Don't ASSERT ABI test,
626         just return FFI_BAD_ABI when things are wrong.
627
628 2012-02-11  Eric Botcazou  <ebotcazou@adacore.com>
629
630         * src/sparc/v9.S (STACKFRAME): Bump to 176.
631
632 2011-02-09  Stuart Shelton  <srcshelton@gmail.com>
633
634         http://bugs.gentoo.org/show_bug.cgi?id=286911
635         * src/mips/ffitarget.h: Clean up error messages.
636         * src/java_raw_api.c (ffi_java_translate_args): Cast raw arg to
637         ffi_raw*.
638         * include/ffi.h.in: Add pragma for SGI compiler.
639
640 2011-02-09  Anthony Green  <green@moxielogic.com>
641
642         * configure.ac: Add powerpc64-*-darwin* support.
643
644 2011-02-09  Anthony Green <green@moxielogic.com>
645
646         * README: Mention Interix.
647
648 2011-02-09  Jonathan Callen  <abcd@gentoo.org>
649
650         * configure.ac: Add Interix to win32/cygwin/mingw case.
651         * configure: Ditto.
652         * src/closures.c: Treat Interix like Cygwin, instead of as a
653         generic win32.
654
655 2011-02-09  Anthony Green <green@moxielogic.com>
656
657         * testsuite/libffi.call/err_bad_typedef.c: Remove xfail.
658         * testsuite/libffi.call/err_bad_abi.c: Remove xfail.
659         * src/x86/ffi64.c (UNLIKELY, LIKELY): Define.
660         (ffi_prep_closure_loc): Check for bad ABI.
661         * src/prep_cif.c (UNLIKELY, LIKELY): Define.
662         (initialize_aggregate): Check for bad types.
663
664 2011-02-09  Landon Fuller <landonf@plausible.coop>
665
666         * Makefile.am (EXTRA_DIST): Add build-ios.sh, src/arm/gentramp.sh,
667         src/arm/trampoline.S.
668         (nodist_libffi_la_SOURCES): Add src/arc/trampoline.S.
669         * configure.ac (FFI_EXEC_TRAMPOLINE_TABLE): Define.
670         * src/arm/ffi.c (ffi_trampoline_table)
671         (ffi_closure_trampoline_table_page, ffi_trampoline_table_entry)
672         (FFI_TRAMPOLINE_CODELOC_CONFIG, FFI_TRAMPOLINE_CONFIG_PAGE_OFFSET)
673         (FFI_TRAMPOLINE_COUNT, ffi_trampoline_lock, ffi_trampoline_tables)
674         (ffi_trampoline_table_alloc, ffi_closure_alloc, ffi_closure_free):
675         Define for FFI_EXEC_TRAMPOLINE_TABLE case (iOS).
676         (ffi_prep_closure_loc): Handl FFI_EXEC_TRAMPOLINE_TABLE case
677         separately.
678         * src/arm/sysv.S: Handle Apple iOS host.
679         * src/closures.c: Handle FFI_EXEC_TRAMPOLINE_TABLE case.
680         * build-ios.sh: New file.
681         * fficonfig.h.in, configure, Makefile.in: Rebuilt.
682         * README: Mention ARM iOS.
683
684 2011-02-08  Oren Held  <orenhe@il.ibm.com>
685
686         * src/dlmalloc.c (_STRUCT_MALLINFO): Define in order to avoid
687         redefinition of mallinfo on HP-UX.
688
689 2011-02-08  Ginn Chen  <ginn.chen@oracle.com>
690
691         * src/sparc/ffi.c (ffi_call): Make compatible with Solaris Studio
692         aggregate return ABI.  Flush cache.
693         (ffi_prep_closure_loc): Flush cache.
694
695 2011-02-11  Anthony Green  <green@moxielogic.com>
696
697         From Tom Honermann <tom.honermann@oracle.com>:
698         * src/powerpc/aix.S (ffi_call_AIX): Support for xlc toolchain on
699         AIX.  Declare .ffi_prep_args.  Insert nops after branch
700         instructions so that the AIX linker can insert TOC reload
701         instructions.
702         * src/powerpc/aix_closure.S: Declare .ffi_closure_helper_DARWIN.
703
704 2011-02-08  Ed  <ed@kdtc.net>
705
706         * src/powerpc/asm.h: Fix grammar nit in comment.
707
708 2011-02-08  Uli Link  <ul.mcamafia@linkitup.de>
709
710         * include/ffi.h.in (FFI_64_BIT_MAX): Define and use.
711
712 2011-02-09  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
713
714         PR libffi/46661
715         * testsuite/libffi.call/cls_pointer.c (main): Cast void * to
716         uintptr_t first.
717         * testsuite/libffi.call/cls_pointer_stack.c (main): Likewise.
718
719 2011-02-08  Rafael Avila de Espindola  <respindola@mozilla.com>
720
721         * configure.ac: Fix x86 test for pc related relocs.
722         * configure: Rebuilt.
723
724 2011-02-07  Joel Sherrill <joel.sherrill@oarcorp.com>
725
726         * libffi/src/m68k/ffi.c: Add RTEMS support for cache flushing.
727         Handle case when CPU variant does not have long double support.
728         * libffi/src/m68k/sysv.S: Add support for mc68000, Coldfire,
729         and cores with soft floating point.
730
731 2011-02-07  Joel Sherrill <joel.sherrill@oarcorp.com>
732
733         * configure.ac: Add mips*-*-rtems* support.
734         * configure: Regenerate.
735         * src/mips/ffitarget.h: Ensure needed constants are available
736         for targets which do not have sgidefs.h.
737
738 2011-01-26  Dave Korn  <dave.korn.cygwin@gmail.com>
739
740         PR target/40125
741         * configure.ac (AM_LTLDFLAGS): Add -bindir option for windows DLLs.
742         * configure: Regenerate.
743
744 2010-12-18  Iain Sandoe  <iains@gcc.gnu.org>
745
746         PR libffi/29152
747         PR libffi/42378
748         * src/powerpc/darwin_closure.S: Provide Darwin64 implementation,
749         update comments.
750         * src/powerpc/ffitarget.h (POWERPC_DARWIN64): New,
751         (FFI_TRAMPOLINE_SIZE): Update for Darwin64.
752         * src/powerpc/darwin.S: Provide Darwin64 implementation,
753         update comments.
754         * src/powerpc/ffi_darwin.c: Likewise.
755
756 2010-12-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
757
758         * configure.ac (libffi_cv_as_ascii_pseudo_op): Use double
759         backslashes.
760         (libffi_cv_as_string_pseudo_op): Likewise.
761         * configure: Regenerate.
762
763 2010-12-03  Chung-Lin Tang  <cltang@codesourcery.com>
764
765         * src/arm/sysv.S (ffi_closure_SYSV): Add UNWIND to .pad directive.
766         (ffi_closure_VFP): Same.
767         (ffi_call_VFP): Move down to before ffi_closure_VFP. Add '.fpu vfp'
768         directive.
769
770 2010-12-01  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
771
772         * testsuite/libffi.call/ffitest.h [__sgi] (PRId64, PRIu64): Define.
773         (PRIuPTR): Define.
774
775 2010-11-29  Richard Henderson  <rth@redhat.com>
776             Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
777
778         * src/x86/sysv.S (FDE_ENCODING, FDE_ENCODE): Define.
779         (.eh_frame): Use FDE_ENCODING.
780         (.LASFDE1, .LASFDE2, LASFDE3): Simplify with FDE_ENCODE.
781
782 2010-11-22  Jacek Caban <jacek@codeweavers.com>
783
784         * configure.ac: Check for symbol underscores on mingw-w64.
785         * configure: Rebuilt.
786         * src/x86/win64.S: Correctly access extern symbols in respect to
787         underscores.
788
789 2010-11-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
790
791         * testsuite/lib/libffi-dg.exp: Rename ...
792         * testsuite/lib/libffi.exp: ... to this.
793         * libffi/testsuite/libffi.call/call.exp: Don't load libffi-dg.exp.
794         * libffi/testsuite/libffi.special/special.exp: Likewise.
795
796 2010-10-28  Chung-Lin Tang  <cltang@codesourcery.com>
797
798         * src/arm/ffi.c (ffi_prep_args): Add VFP register argument handling
799         code, new parameter, and return value. Update comments.
800         (ffi_prep_cif_machdep): Add case for VFP struct return values. Add
801         call to layout_vfp_args().
802         (ffi_call_SYSV): Update declaration.
803         (ffi_call_VFP): New declaration.
804         (ffi_call): Add VFP struct return conditions. Call ffi_call_VFP()
805         when ABI is FFI_VFP.
806         (ffi_closure_VFP): New declaration.
807         (ffi_closure_SYSV_inner): Add new vfp_args parameter, update call to
808         ffi_prep_incoming_args_SYSV().
809         (ffi_prep_incoming_args_SYSV): Update parameters. Add VFP argument
810         case handling.
811         (ffi_prep_closure_loc): Pass ffi_closure_VFP to trampoline
812         construction under VFP hard-float.
813         (rec_vfp_type_p): New function.
814         (vfp_type_p): Same.
815         (place_vfp_arg): Same.
816         (layout_vfp_args): Same.
817         * src/arm/ffitarget.h (ffi_abi): Add FFI_VFP. Define FFI_DEFAULT_ABI
818         based on __ARM_PCS_VFP.
819         (FFI_EXTRA_CIF_FIELDS): Define for adding VFP hard-float specific
820         fields.
821         (FFI_TYPE_STRUCT_VFP_FLOAT): Define internally used type code.
822         (FFI_TYPE_STRUCT_VFP_DOUBLE): Same.
823         * src/arm/sysv.S (ffi_call_SYSV): Change call of ffi_prep_args() to
824         direct call. Move function pointer load upwards.
825         (ffi_call_VFP): New function.
826         (ffi_closure_VFP): Same.
827
828         * testsuite/lib/libffi-dg.exp (check-flags): New function.
829         (dg-skip-if): New function.
830         * testsuite/libffi.call/cls_double_va.c: Skip if target is arm*-*-*
831         and compiler options include -mfloat-abi=hard.
832         * testsuite/libffi.call/cls_longdouble_va.c: Same.
833
834 2010-10-01  Jakub Jelinek  <jakub@redhat.com>
835
836         PR libffi/45677
837         * src/x86/ffi64.c (ffi_prep_cif_machdep): Ensure cif->bytes is
838         a multiple of 8.
839         * testsuite/libffi.call/many2.c: New test.
840
841 2010-08-20  Mark Wielaard  <mjw@redhat.com>
842
843         * src/closures.c (open_temp_exec_file_mnt): Check if getmntent_r
844         returns NULL.
845
846 2010-08-09  Andreas Tobler  <andreast@fgznet.ch>
847
848         * configure.ac: Add target powerpc64-*-freebsd*.
849         * configure: Regenerate.
850         * testsuite/libffi.call/cls_align_longdouble_split.c: Pass
851         -mlong-double-128 only to linux targets.
852         * testsuite/libffi.call/cls_align_longdouble_split2.c: Likewise.
853         * testsuite/libffi.call/cls_longdouble.c: Likewise.
854         * testsuite/libffi.call/huge_struct.c: Likewise.
855
856 2010-08-05  Dan Witte  <dwitte@mozilla.com>
857
858         * Makefile.am: Pass FFI_DEBUG define to msvcc.sh for linking to the
859         debug CRT when --enable-debug is given.
860         * configure.ac: Define it.
861         * msvcc.sh: Translate -g and -DFFI_DEBUG appropriately.
862
863 2010-08-04  Dan Witte  <dwitte@mozilla.com>
864
865         * src/x86/ffitarget.h: Add X86_ANY define for all x86/x86_64
866         platforms.
867         * src/x86/ffi.c: Remove redundant ifdef checks.
868         * src/prep_cif.c: Push stack space computation into src/x86/ffi.c
869         for X86_ANY so return value space doesn't get added twice.
870
871 2010-08-03  Neil Rashbrooke <neil@parkwaycc.co.uk>
872
873         * msvcc.sh: Don't pass -safeseh to ml64 because behavior is buggy.
874
875 2010-07-22  Dan Witte  <dwitte@mozilla.com>
876
877         * src/*/ffitarget.h: Make FFI_LAST_ABI one past the last valid ABI.
878         * src/prep_cif.c: Fix ABI assertion.
879         * src/cris/ffi.c: Ditto.
880
881 2010-07-10  Evan Phoenix  <evan@fallingsnow.net>
882
883         * src/closures.c (selinux_enabled_check): Fix strncmp usage bug.
884
885 2010-07-07  Dan HorĂ¡k <dan@danny.cz>
886
887         * include/ffi.h.in: Protect #define with #ifndef.
888         * src/powerpc/ffitarget.h: Ditto.
889         * src/s390/ffitarget.h: Ditto.
890         * src/sparc/ffitarget.h: Ditto.
891
892 2010-07-07   Neil Roberts <neil@linux.intel.com>
893
894         * src/x86/sysv.S (ffi_call_SYSV): Align the stack pointer to
895         16-bytes.
896
897 2010-07-02  Jakub Jelinek  <jakub@redhat.com>
898
899         * Makefile.am (AM_MAKEFLAGS): Pass also mandir to submakes.
900         * Makefile.in: Regenerated.
901
902 2010-05-19  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
903
904         * configure.ac (libffi_cv_as_x86_pcrel): Check for illegal in as
905         output, too.
906         (libffi_cv_as_ascii_pseudo_op): Check for .ascii.
907         (libffi_cv_as_string_pseudo_op): Check for .string.
908         * configure: Regenerate.
909         * fficonfig.h.in: Regenerate.
910         * src/x86/sysv.S (.eh_frame): Use .ascii, .string or error.
911
912 2010-05-11  Dan Witte  <dwitte@mozilla.com>
913
914         * doc/libffi.tex: Document previous change.
915
916 2010-05-11  Makoto Kato <m_kato@ga2.so-net.ne.jp>
917
918         * src/x86/ffi.c (ffi_call): Don't copy structs passed by value.
919
920 2010-05-05  Michael Kohler <michaelkohler@live.com>
921
922         * src/dlmalloc.c (dlfree): Fix spelling.
923         * src/ia64/ffi.c (ffi_prep_cif_machdep): Ditto.
924         * configure.ac: Ditto.
925         * configure: Rebuilt.
926
927 2010-04-13  Dan Witte  <dwitte@mozilla.com>
928
929         * msvcc.sh: Build with -W3 instead of -Wall.
930         * src/powerpc/ffi_darwin.c: Remove build warnings.
931         * src/x86/ffi.c: Ditto.
932         * src/x86/ffitarget.h: Ditto.
933
934 2010-04-12  Dan Witte  <dwitte@mozilla.com>
935             Walter Meinl <wuno@lsvw.de>
936
937         * configure.ac: Add OS/2 support.
938         * configure: Rebuilt.
939         * src/closures.c: Ditto.
940         * src/dlmalloc.c: Ditto.
941         * src/x86/win32.S: Ditto.
942
943 2010-04-07  Jakub Jelinek  <jakub@redhat.com>
944
945         * testsuite/libffi.call/err_bad_abi.c: Remove unused args variable.
946
947 2010-04-02  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
948
949         * Makefile.in: Regenerate.
950         * aclocal.m4: Regenerate.
951         * include/Makefile.in: Regenerate.
952         * man/Makefile.in: Regenerate.
953         * testsuite/Makefile.in: Regenerate.
954
955 2010-03-30  Dan Witte  <dwitte@mozilla.com>
956
957         * msvcc.sh: Disable build warnings.
958         * README (tested): Clarify windows build procedure.
959
960 2010-03-15  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
961
962         * configure.ac (libffi_cv_as_x86_64_unwind_section_type): New test.
963         * configure: Regenerate.
964         * fficonfig.h.in: Regenerate.
965         * libffi/src/x86/unix64.S (.eh_frame)
966         [HAVE_AS_X86_64_UNWIND_SECTION_TYPE]: Use @unwind section type.
967
968 2010-03-14  Matthias Klose  <doko@ubuntu.com>
969
970         * src/x86/ffi64.c: Fix typo in comment.
971         * src/x86/ffi.c: Use /* ... */ comment style.
972
973 2010-02-24  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
974
975         * doc/libffi.texi (The Closure API): Fix typo.
976         * doc/libffi.info: Remove.
977
978 2010-02-15  Matthias Klose  <doko@ubuntu.com>
979
980         * src/arm/sysv.S (__ARM_ARCH__): Define for processor
981         __ARM_ARCH_7EM__.
982
983 2010-01-15  Anthony Green  <green@redhat.com>
984
985         * README: Add notes on building with Microsoft Visual C++.
986
987 2010-01-15  Daniel Witte  <dwitte@mozilla.com>
988
989         * msvcc.sh: New file.
990
991         * src/x86/win32.S: Port assembly routines to MSVC and #ifdef.
992         * src/x86/ffi.c: Tweak function declaration and remove excess
993         parens.
994         * include/ffi.h.in: Add __declspec(align(8)) to typedef struct
995         ffi_closure.
996
997         * src/x86/ffi.c: Merge ffi_call_SYSV and ffi_call_STDCALL into new
998         function ffi_call_win32 on X86_WIN32.
999         * src/x86/win32.S (ffi_call_SYSV): Rename to ffi_call_win32.
1000         (ffi_call_STDCALL): Remove.
1001
1002         * src/prep_cif.c (ffi_prep_cif): Move stack space allocation code
1003         to ffi_prep_cif_machdep for x86.
1004         * src/x86/ffi.c (ffi_prep_cif_machdep): To here.
1005
1006 2010-01-15  Oliver Kiddle  <okiddle@yahoo.co.uk>
1007
1008         * src/x86/ffitarget.h (ffi_abi): Check for __i386 and __amd64 for
1009         Sun Studio compiler compatibility.
1010
1011 2010-01-12  Conrad Irwin <conrad.irwin@gmail.com>
1012
1013         * doc/libffi.texi: Add closure example.
1014
1015 2010-01-07  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
1016
1017         PR libffi/40701
1018         * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRIdLL,
1019         PRIuLL, PRId64, PRIu64, PRIuPTR): Define.
1020         * testsuite/libffi.call/cls_align_sint64.c: Add -Wno-format on
1021         alpha*-dec-osf*.
1022         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
1023         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
1024         * testsuite/libffi.call/return_ll1.c: Likewise.
1025         * testsuite/libffi.call/stret_medium2.c: Likewise.
1026         * testsuite/libffi.special/ffitestcxx.h (allocate_mmap): Cast
1027         MAP_FAILED to char *.
1028
1029 2010-01-06  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
1030
1031         * src/mips/n32.S: Use .abicalls and .eh_frame with __GNUC__.
1032
1033 2009-12-31  Anthony Green  <green@redhat.com>
1034
1035         * README: Update for libffi 3.0.9.
1036
1037 2009-12-27  Matthias Klose  <doko@ubuntu.com>
1038
1039         * configure.ac (HAVE_LONG_DOUBLE): Define for mips when
1040         appropriate.
1041         * configure: Rebuilt.
1042
1043 2009-12-26  Anthony Green  <green@redhat.com>
1044
1045         * testsuite/libffi.call/cls_longdouble_va.c: Mark as xfail for
1046         avr32*-*-*.
1047         * testsuite/libffi.call/cls_double_va.c: Ditto.
1048
1049 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
1050
1051         * testsuite/libffi.call/ffitest.h: Conditionally include stdint.h
1052         and inttypes.h.
1053         * testsuite/libffi.special/unwindtest.cc: Ditto.
1054
1055 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
1056
1057         * configure.ac: Add amd64-*-openbsd*.
1058         * configure: Rebuilt.
1059         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Link
1060         openbsd programs with -lpthread.
1061
1062 2009-12-26  Anthony Green  <green@redhat.com>
1063
1064         * testsuite/libffi.call/cls_double_va.c,
1065         testsuite/libffi.call/cls_longdouble.c,
1066         testsuite/libffi.call/cls_longdouble_va.c,
1067         testsuite/libffi.call/cls_pointer.c,
1068         testsuite/libffi.call/cls_pointer_stack.c: Remove xfail for
1069         mips*-*-* and arm*-*-*.
1070         * testsuite/libffi.call/cls_align_longdouble_split.c,
1071         testsuite/libffi.call/cls_align_longdouble_split2.c,
1072         testsuite/libffi.call/stret_medium2.c,
1073         testsuite/libffi.call/stret_medium.c,
1074         testsuite/libffi.call/stret_large.c,
1075         testsuite/libffi.call/stret_large2.c: Remove xfail for arm*-*-*.
1076
1077 2009-12-31  Kay Tietz  <ktietz70@googlemail.com>
1078
1079         * testsuite/libffi.call/ffitest.h,
1080         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRuLL): Fix
1081         definitions.
1082
1083 2009-12-31  Carlo Bramini  <carlo.bramix@libero.it>
1084
1085         * configure.ac (AM_LTLDFLAGS): Define for windows hosts.
1086         * Makefile.am (libffi_la_LDFLAGS): Add AM_LTLDFLAGS.
1087         * configure: Rebuilt.
1088         * Makefile.in: Rebuilt.
1089
1090 2009-12-31  Anthony Green  <green@redhat.com>
1091             Blake Chaffin.
1092
1093         * testsuite/libffi.call/huge_struct.c: New test case from Blake
1094         Chaffin @ Apple.
1095
1096 2009-12-28  David Edelsohn  <edelsohn@gnu.org>
1097
1098         * src/powerpc/ffi_darwin.c (ffi_prep_args): Copy abi and nargs to
1099         local variables.
1100         (aix_adjust_aggregate_sizes): New function.
1101         (ffi_prep_cif_machdep): Call it.
1102
1103 2009-12-26  Andreas Tobler  <a.tobler@schweiz.org>
1104
1105         * configure.ac: Define FFI_MMAP_EXEC_WRIT for the given targets.
1106         * configure: Regenerate.
1107         * fficonfig.h.in: Likewise.
1108         * src/closures.c: Remove the FFI_MMAP_EXEC_WRIT definition for
1109         Solaris/x86.
1110
1111 2009-12-26  Andreas Schwab  <schwab@linux-m68k.org>
1112
1113         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Advance intarg_count
1114         when a float arguments is passed in memory.
1115         (ffi_closure_helper_SYSV): Mark general registers as used up when
1116         a 64bit or soft-float long double argument is passed in memory.
1117
1118 2009-12-25  Matthias Klose  <doko@ubuntu.com>
1119
1120         * man/ffi_call.3: Fix #include in examples.
1121         * doc/libffi.texi: Add dircategory.
1122
1123 2009-12-25  Frank Everdij <f.p.x.everdij@tudelft.nl>
1124
1125         * include/ffi.h.in: Placed '__GNUC__' ifdef around
1126         '__attribute__((aligned(8)))' in ffi_closure, fixes compile for
1127         IRIX MIPSPro c99.
1128         * include/ffi_common.h: Added '__sgi' define to non
1129         '__attribute__((__mode__()))' integer typedefs.
1130         * src/mips/ffi.c (ffi_call, ffi_closure_mips_inner_O32,
1131         ffi_closure_mips_inner_N32): Added 'defined(_MIPSEB)' to BE check.
1132         (ffi_closure_mips_inner_O32, ffi_closure_mips_inner_N32): Added
1133         FFI_LONGDOUBLE support and alignment(N32 only).
1134         * src/mips/ffitarget.h: Corrected '#include <sgidefs.h>' for IRIX and
1135         fixed non '__attribute__((__mode__()))' integer typedefs.
1136         * src/mips/n32.S: Put '#ifdef linux' around '.abicalls' and '.eh_frame'
1137         since they are Linux/GNU Assembler specific.
1138
1139 2009-12-25  Bradley Smith  <brad@brad-smith.co.uk>
1140
1141         * configure.ac, Makefile.am, src/avr32/ffi.c,
1142         src/avr32/ffitarget.h,
1143         src/avr32/sysv.S: Add AVR32 port.
1144         * configure, Makefile.in: Rebuilt.
1145
1146 2009-12-21  Andreas Tobler  <a.tobler@schweiz.org>
1147
1148         * configure.ac: Make i?86 build on FreeBSD and OpenBSD.
1149         * configure: Regenerate.
1150
1151 2009-12-15  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1152
1153         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on PA HP-UX.
1154
1155 2009-12-13  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
1156
1157         * src/pa/ffi.c (ffi_closure_inner_pa32): Handle FFI_TYPE_LONGDOUBLE
1158         type on HP-UX.
1159
1160 2012-02-13  Kai Tietz  <ktietz@redhat.com>
1161
1162         PR libffi/52221
1163         * src/x86/ffi.c (ffi_prep_raw_closure_loc): Add thiscall
1164         support for X86_WIN32.
1165         (FFI_INIT_TRAMPOLINE_THISCALL): Fix displacement.
1166
1167 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
1168
1169         * src/sparc/ffi.c (ffi_closure_sparc_inner_v9): Properly align 'long
1170         double' arguments.
1171
1172 2009-12-11  Eric Botcazou  <ebotcazou@adacore.com>
1173
1174         * testsuite/libffi.call/ffitest.h: Define PRIuPTR on Solaris < 10.
1175
1176 2009-12-10  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
1177
1178         PR libffi/40700
1179         * src/closures.c [X86_64 && __sun__ && __svr4__]
1180         (FFI_MMAP_EXEC_WRIT): Define.
1181
1182 2009-12-08  David Daney  <ddaney@caviumnetworks.com>
1183
1184         * testsuite/libffi.call/stret_medium.c: Remove xfail for mips*-*-*
1185         * testsuite/libffi.call/cls_align_longdouble_split2.c: Same.
1186         * testsuite/libffi.call/stret_large.c: Same.
1187         * testsuite/libffi.call/cls_align_longdouble_split.c: Same.
1188         * testsuite/libffi.call/stret_large2.c: Same.
1189         * testsuite/libffi.call/stret_medium2.c: Same.
1190
1191 2009-12-07  David Edelsohn  <edelsohn@gnu.org>
1192
1193         * src/powerpc/aix_closure.S (libffi_closure_ASM): Fix tablejump
1194         typo.
1195
1196 2009-12-05  David Edelsohn  <edelsohn@gnu.org>
1197
1198         * src/powerpc/aix.S: Update AIX32 code to be consistent with AIX64
1199         code.
1200         * src/powerpc/aix_closure.S: Same.
1201
1202 2009-12-05  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1203
1204         * Makefile.in: Regenerate.
1205         * configure: Regenerate.
1206         * include/Makefile.in: Regenerate.
1207         * man/Makefile.in: Regenerate.
1208         * testsuite/Makefile.in: Regenerate.
1209
1210 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
1211
1212         * src/powerpc/aix_closure.S: Reorganize 64-bit code to match
1213         linux64_closure.S.
1214
1215 2009-12-04  Uros Bizjak  <ubizjak@gmail.com>
1216
1217         PR libffi/41908
1218         * src/x86/ffi64.c (classify_argument): Update from
1219         gcc/config/i386/i386.c.
1220         (ffi_closure_unix64_inner): Do not use the address of two consecutive
1221         SSE registers directly.
1222         * testsuite/libffi.call/cls_dbls_struct.c (main): Remove xfail
1223         for x86_64 linux targets.
1224
1225 2009-12-04  David Edelsohn  <edelsohn@gnu.org>
1226
1227         * src/powerpc/ffi_darwin.c (ffi_closure_helper_DARWIN): Increment
1228         pfr for long double split between fpr13 and stack.
1229
1230 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
1231
1232         * src/powerpc/ffi_darwin.c (ffi_prep_args): Increment next_arg and
1233         fparg_count twice for long double.
1234
1235 2009-12-03  David Edelsohn  <edelsohn@gnu.org>
1236
1237         PR libffi/42243
1238         * src/powerpc/ffi_darwin.c (ffi_prep_args): Remove extra parentheses.
1239
1240 2009-12-03  Uros Bizjak  <ubizjak@gmail.com>
1241
1242         * testsuite/libffi.call/cls_longdouble_va.c (main): Fix format string.
1243         Remove xfails for x86 linux targets.
1244
1245 2009-12-02  David Edelsohn  <edelsohn@gnu.org>
1246
1247         * src/powerpc/ffi_darwin.c (ffi_prep_args): Fix typo in INT64
1248         case.
1249
1250 2009-12-01  David Edelsohn  <edelsohn@gnu.org>
1251
1252         * src/powerpc/aix.S (ffi_call_AIX): Convert to more standard
1253         register usage.  Call ffi_prep_args directly.  Add long double
1254         return value support.
1255         * src/powerpc/ffi_darwin.c (ffi_prep_args): Double arg increment
1256         applies to FFI_TYPE_DOUBLE.  Correct fpr_base increment typo.
1257         Separate FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases.
1258         (ffi_prep_cif_machdep): Only 16 byte stack alignment in 64 bit
1259         mode.
1260         (ffi_closure_helper_DARWIN): Remove nf and ng counters.  Move temp
1261         into case.
1262         * src/powerpc/aix_closure.S: Maintain 16 byte stack alignment.
1263         Allocate result area between params and FPRs.
1264
1265 2009-11-30  David Edelsohn  <edelsohn@gnu.org>
1266
1267         PR target/35484
1268         * src/powerpc/ffitarget.h (POWERPC64): Define for PPC64 Linux and
1269         AIX64.
1270         * src/powerpc/aix.S: Implement AIX64 version.
1271         * src/powerpc/aix_closure.S: Implement AIX64 version.
1272         (ffi_closure_ASM): Use extsb, lha and displament addresses.
1273         * src/powerpc/ffi_darwin.c (ffi_prep_args): Implement AIX64
1274         support.
1275         (ffi_prep_cif_machdep): Same.
1276         (ffi_call): Same.
1277         (ffi_closure_helper_DARWIN): Same.
1278
1279 2009-11-02  Andreas Tobler  <a.tobler@schweiz.org>
1280
1281         PR libffi/41908
1282         * testsuite/libffi.call/testclosure.c: New test.
1283
1284 2009-09-28  Kai Tietz  <kai.tietz@onevision.com>
1285
1286         * src/x86/win64.S (_ffi_call_win64 stack): Remove for gnu
1287         assembly version use of ___chkstk.
1288
1289 2009-09-23  Matthias Klose  <doko@ubuntu.com>
1290
1291         PR libffi/40242, PR libffi/41443
1292         * src/arm/sysv.S (__ARM_ARCH__): Define for processors
1293         __ARM_ARCH_6T2__, __ARM_ARCH_6M__, __ARM_ARCH_7__,
1294         __ARM_ARCH_7A__, __ARM_ARCH_7R__, __ARM_ARCH_7M__.
1295         Change the conditionals to __SOFTFP__ || __ARM_EABI__
1296         for -mfloat-abi=softfp to work.
1297
1298 2009-09-17  Loren J. Rittle  <ljrittle@acm.org>
1299
1300         PR testsuite/32843 (strikes again)
1301         * src/x86/ffi.c (ffi_prep_cif_machdep): Add X86_FREEBSD to
1302         enable proper extension on char and short.
1303
1304 2009-09-15  David Daney  <ddaney@caviumnetworks.com>
1305
1306         * src/java_raw_api.c (ffi_java_raw_to_rvalue): Remove special
1307         handling for FFI_TYPE_POINTER.
1308         * src/mips/ffitarget.h (FFI_TYPE_STRUCT_D_SOFT,
1309         FFI_TYPE_STRUCT_F_SOFT, FFI_TYPE_STRUCT_DD_SOFT,
1310         FFI_TYPE_STRUCT_FF_SOFT, FFI_TYPE_STRUCT_FD_SOFT,
1311         FFI_TYPE_STRUCT_DF_SOFT, FFI_TYPE_STRUCT_SOFT): New defines.
1312         (FFI_N32_SOFT_FLOAT, FFI_N64_SOFT_FLOAT): New ffi_abi enumerations.
1313         (enum ffi_abi): Set FFI_DEFAULT_ABI for soft-float.
1314         * src/mips/n32.S (ffi_call_N32): Add handling for soft-float
1315         structure and pointer returns.
1316         (ffi_closure_N32): Add handling for pointer returns.
1317         * src/mips/ffi.c (ffi_prep_args, calc_n32_struct_flags,
1318         calc_n32_return_struct_flags): Handle soft-float.
1319         (ffi_prep_cif_machdep):  Handle soft-float, fix pointer handling.
1320         (ffi_call_N32): Declare proper argument types.
1321         (ffi_call, copy_struct_N32, ffi_closure_mips_inner_N32): Handle
1322         soft-float.
1323
1324 2009-08-24  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1325
1326         * configure.ac (AC_PREREQ): Bump to 2.64.
1327
1328 2009-08-22  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1329
1330         * Makefile.am (install-html, install-pdf): Remove.
1331         * Makefile.in: Regenerate.
1332
1333         * Makefile.in: Regenerate.
1334         * aclocal.m4: Regenerate.
1335         * configure: Regenerate.
1336         * fficonfig.h.in: Regenerate.
1337         * include/Makefile.in: Regenerate.
1338         * man/Makefile.in: Regenerate.
1339         * testsuite/Makefile.in: Regenerate.
1340
1341 2011-08-22  Jasper Lievisse Adriaanse <jasper@openbsd.org>
1342
1343         * configure.ac: Add OpenBSD/hppa and OpenBSD/powerpc support.
1344         * configure: Rebuilt.
1345
1346 2009-07-30  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1347
1348         * configure.ac (_AC_ARG_VAR_PRECIOUS): Use m4_rename_force.
1349
1350 2009-07-24  Dave Korn  <dave.korn.cygwin@gmail.com>
1351
1352         PR libffi/40807
1353         * src/x86/ffi.c (ffi_prep_cif_machdep): Also use sign/zero-extending
1354         return types for X86_WIN32.
1355         * src/x86/win32.S (_ffi_call_SYSV): Handle omitted return types.
1356         (_ffi_call_STDCALL, _ffi_closure_SYSV, _ffi_closure_raw_SYSV,
1357         _ffi_closure_STDCALL): Likewise.
1358
1359         * src/closures.c (is_selinux_enabled): Define to const 0 for Cygwin.
1360         (dlmmap, dlmunmap): Also use these functions on Cygwin.
1361
1362 2009-07-11  Richard Sandiford  <rdsandiford@googlemail.com>
1363
1364         PR testsuite/40699
1365         PR testsuite/40707
1366         PR testsuite/40709
1367         * testsuite/lib/libffi-dg.exp: Revert 2009-07-02, 2009-07-01 and
1368         2009-06-30 commits.
1369
1370 2009-07-01  Richard Sandiford  <r.sandiford@uk.ibm.com>
1371
1372         * testsuite/lib/libffi-dg.exp (libffi-init): Set ld_library_path
1373         to "" before adding paths.  (This reinstates an assignment that
1374         was removed by my 2009-06-30 commit, but changes the initial
1375         value from "." to "".)
1376
1377 2009-07-01  H.J. Lu  <hongjiu.lu@intel.com>
1378
1379         PR testsuite/40601
1380         * testsuite/lib/libffi-dg.exp (libffi-init): Properly set
1381         gccdir.  Adjust ld_library_path for gcc only if gccdir isn't
1382         empty.
1383
1384 2009-06-30  Richard Sandiford  <r.sandiford@uk.ibm.com>
1385
1386         * testsuite/lib/libffi-dg.exp (libffi-init): Don't add "."
1387         to ld_library_path.  Use add_path.  Add just find_libgcc_s
1388         to ld_library_path, not every libgcc multilib directory.
1389
1390 2009-06-16  Wim Lewis  <wiml@hhhh.org>
1391
1392         * src/powerpc/ffi.c: Avoid clobbering cr3 and cr4, which are
1393         supposed to be callee-saved.
1394         * src/powerpc/sysv.S (small_struct_return_value): Fix overrun of
1395         return buffer for odd-size structs.
1396
1397 2009-06-16  Andreas Tobler  <a.tobler@schweiz.org>
1398
1399         PR libffi/40444
1400         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Add
1401         allow_stack_execute for Darwin.
1402
1403 2009-06-16  Andrew Haley  <aph@redhat.com>
1404
1405         * configure.ac (TARGETDIR): Add missing blank lines.
1406         * configure: Regenerate.
1407
1408 2009-06-16  Andrew Haley  <aph@redhat.com>
1409
1410         * testsuite/libffi.call/cls_align_sint64.c,
1411         testsuite/libffi.call/cls_align_uint64.c,
1412         testsuite/libffi.call/cls_longdouble_va.c,
1413         testsuite/libffi.call/cls_ulonglong.c,
1414         testsuite/libffi.call/return_ll1.c,
1415         testsuite/libffi.call/stret_medium2.c: Fix printf format
1416         specifiers.
1417         * testsuite/libffi.call/ffitest.h,
1418         testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRIuLL): Define.
1419
1420 2009-06-15  Andrew Haley  <aph@redhat.com>
1421
1422         * testsuite/libffi.call/err_bad_typedef.c: xfail everywhere.
1423         * testsuite/libffi.call/err_bad_abi.c: Likewise.
1424
1425 2009-06-12  Andrew Haley  <aph@redhat.com>
1426
1427         * Makefile.am: Remove info_TEXINFOS.
1428
1429 2009-06-12  Andrew Haley  <aph@redhat.com>
1430
1431         * ChangeLog.libffi: testsuite/libffi.call/cls_align_sint64.c,
1432         testsuite/libffi.call/cls_align_uint64.c,
1433         testsuite/libffi.call/cls_ulonglong.c,
1434         testsuite/libffi.call/return_ll1.c,
1435         testsuite/libffi.call/stret_medium2.c: Fix printf format
1436         specifiers.
1437         testsuite/libffi.special/unwindtest.cc: include stdint.h.
1438
1439 2009-06-11  Timothy Wall  <twall@users.sf.net>
1440
1441         * Makefile.am,
1442         configure.ac,
1443         include/ffi.h.in,
1444         include/ffi_common.h,
1445         src/closures.c,
1446         src/dlmalloc.c,
1447         src/x86/ffi.c,
1448         src/x86/ffitarget.h,
1449         src/x86/win64.S (new),
1450         README: Added win64 support (mingw or MSVC)
1451         * Makefile.in,
1452         include/Makefile.in,
1453         man/Makefile.in,
1454         testsuite/Makefile.in,
1455         configure,
1456         aclocal.m4: Regenerated
1457         * ltcf-c.sh: properly escape cygwin/w32 path
1458         * man/ffi_call.3: Clarify size requirements for return value.
1459         * src/x86/ffi64.c: Fix filename in comment.
1460         * src/x86/win32.S: Remove unused extern.
1461
1462         * testsuite/libffi.call/closure_fn0.c,
1463         testsuite/libffi.call/closure_fn1.c,
1464         testsuite/libffi.call/closure_fn2.c,
1465         testsuite/libffi.call/closure_fn3.c,
1466         testsuite/libffi.call/closure_fn4.c,
1467         testsuite/libffi.call/closure_fn5.c,
1468         testsuite/libffi.call/closure_fn6.c,
1469         testsuite/libffi.call/closure_stdcall.c,
1470         testsuite/libffi.call/cls_12byte.c,
1471         testsuite/libffi.call/cls_16byte.c,
1472         testsuite/libffi.call/cls_18byte.c,
1473         testsuite/libffi.call/cls_19byte.c,
1474         testsuite/libffi.call/cls_1_1byte.c,
1475         testsuite/libffi.call/cls_20byte.c,
1476         testsuite/libffi.call/cls_20byte1.c,
1477         testsuite/libffi.call/cls_24byte.c,
1478         testsuite/libffi.call/cls_2byte.c,
1479         testsuite/libffi.call/cls_3_1byte.c,
1480         testsuite/libffi.call/cls_3byte1.c,
1481         testsuite/libffi.call/cls_3byte2.c,
1482         testsuite/libffi.call/cls_4_1byte.c,
1483         testsuite/libffi.call/cls_4byte.c,
1484         testsuite/libffi.call/cls_5_1_byte.c,
1485         testsuite/libffi.call/cls_5byte.c,
1486         testsuite/libffi.call/cls_64byte.c,
1487         testsuite/libffi.call/cls_6_1_byte.c,
1488         testsuite/libffi.call/cls_6byte.c,
1489         testsuite/libffi.call/cls_7_1_byte.c,
1490         testsuite/libffi.call/cls_7byte.c,
1491         testsuite/libffi.call/cls_8byte.c,
1492         testsuite/libffi.call/cls_9byte1.c,
1493         testsuite/libffi.call/cls_9byte2.c,
1494         testsuite/libffi.call/cls_align_double.c,
1495         testsuite/libffi.call/cls_align_float.c,
1496         testsuite/libffi.call/cls_align_longdouble.c,
1497         testsuite/libffi.call/cls_align_longdouble_split.c,
1498         testsuite/libffi.call/cls_align_longdouble_split2.c,
1499         testsuite/libffi.call/cls_align_pointer.c,
1500         testsuite/libffi.call/cls_align_sint16.c,
1501         testsuite/libffi.call/cls_align_sint32.c,
1502         testsuite/libffi.call/cls_align_sint64.c,
1503         testsuite/libffi.call/cls_align_uint16.c,
1504         testsuite/libffi.call/cls_align_uint32.c,
1505         testsuite/libffi.call/cls_align_uint64.c,
1506         testsuite/libffi.call/cls_dbls_struct.c,
1507         testsuite/libffi.call/cls_double.c,
1508         testsuite/libffi.call/cls_double_va.c,
1509         testsuite/libffi.call/cls_float.c,
1510         testsuite/libffi.call/cls_longdouble.c,
1511         testsuite/libffi.call/cls_longdouble_va.c,
1512         testsuite/libffi.call/cls_multi_schar.c,
1513         testsuite/libffi.call/cls_multi_sshort.c,
1514         testsuite/libffi.call/cls_multi_sshortchar.c,
1515         testsuite/libffi.call/cls_multi_uchar.c,
1516         testsuite/libffi.call/cls_multi_ushort.c,
1517         testsuite/libffi.call/cls_multi_ushortchar.c,
1518         testsuite/libffi.call/cls_pointer.c,
1519         testsuite/libffi.call/cls_pointer_stack.c,
1520         testsuite/libffi.call/cls_schar.c,
1521         testsuite/libffi.call/cls_sint.c,
1522         testsuite/libffi.call/cls_sshort.c,
1523         testsuite/libffi.call/cls_uchar.c,
1524         testsuite/libffi.call/cls_uint.c,
1525         testsuite/libffi.call/cls_ulonglong.c,
1526         testsuite/libffi.call/cls_ushort.c,
1527         testsuite/libffi.call/err_bad_abi.c,
1528         testsuite/libffi.call/err_bad_typedef.c,
1529         testsuite/libffi.call/float2.c,
1530         testsuite/libffi.call/huge_struct.c,
1531         testsuite/libffi.call/nested_struct.c,
1532         testsuite/libffi.call/nested_struct1.c,
1533         testsuite/libffi.call/nested_struct10.c,
1534         testsuite/libffi.call/nested_struct2.c,
1535         testsuite/libffi.call/nested_struct3.c,
1536         testsuite/libffi.call/nested_struct4.c,
1537         testsuite/libffi.call/nested_struct5.c,
1538         testsuite/libffi.call/nested_struct6.c,
1539         testsuite/libffi.call/nested_struct7.c,
1540         testsuite/libffi.call/nested_struct8.c,
1541         testsuite/libffi.call/nested_struct9.c,
1542         testsuite/libffi.call/problem1.c,
1543         testsuite/libffi.call/return_ldl.c,
1544         testsuite/libffi.call/return_ll1.c,
1545         testsuite/libffi.call/stret_large.c,
1546         testsuite/libffi.call/stret_large2.c,
1547         testsuite/libffi.call/stret_medium.c,
1548         testsuite/libffi.call/stret_medium2.c,
1549         testsuite/libffi.special/unwindtest.cc: use ffi_closure_alloc instead
1550         of checking for MMAP.  Use intptr_t instead of long casts.
1551
1552 2009-06-11  Kaz Kojima  <kkojima@gcc.gnu.org>
1553
1554         * testsuite/libffi.call/cls_longdouble_va.c: Add xfail sh*-*-linux-*.
1555         * testsuite/libffi.call/err_bad_abi.c: Add xfail sh*-*-*.
1556         * testsuite/libffi.call/err_bad_typedef.c: Likewise.
1557
1558 2009-06-09  Andrew Haley  <aph@redhat.com>
1559
1560         * src/x86/freebsd.S: Add missing file.
1561
1562 2009-06-08  Andrew Haley  <aph@redhat.com>
1563
1564         Import from libffi 3.0.8:
1565
1566         * doc/libffi.texi: New file.
1567         * doc/libffi.info: Likewise.
1568         * doc/stamp-vti: Likewise.
1569         * man/Makefile.am: New file.
1570         * man/ffi_call.3: New file.
1571
1572         * Makefile.am (EXTRA_DIST): Add src/x86/darwin64.S,
1573         src/dlmalloc.c.
1574         (nodist_libffi_la_SOURCES): Add X86_FREEBSD.
1575
1576         * configure.ac: Bump version to 3.0.8.
1577         parisc*-*-linux*: Add.
1578         i386-*-freebsd* | i386-*-openbsd*: Add.
1579         powerpc-*-beos*: Add.
1580         AM_CONDITIONAL X86_FREEBSD: Add.
1581         AC_CONFIG_FILES: Add man/Makefile.
1582
1583         * include/ffi.h.in (FFI_FN): Change void (*)() to void (*)(void).
1584
1585 2009-06-08  Andrew Haley  <aph@redhat.com>
1586
1587         * README: Import from libffi 3.0.8.
1588
1589 2009-06-08  Andrew Haley  <aph@redhat.com>
1590
1591         * testsuite/libffi.call/err_bad_abi.c: Add xfails.
1592         * testsuite/libffi.call/cls_longdouble_va.c: Add xfails.
1593         * testsuite/libffi.call/cls_dbls_struct.c: Add xfail x86_64-*-linux-*.
1594         * testsuite/libffi.call/err_bad_typedef.c: Add xfails.
1595
1596         * testsuite/libffi.call/stret_medium2.c: Add __UNUSED__ to args.
1597         * testsuite/libffi.call/stret_medium.c: Likewise.
1598         * testsuite/libffi.call/stret_large2.c: Likewise.
1599         * testsuite/libffi.call/stret_large.c:  Likewise.
1600
1601 2008-12-26  Timothy Wall  <twall@users.sf.net>
1602
1603         * testsuite/libffi.call/cls_longdouble.c,
1604         testsuite/libffi.call/cls_longdouble_va.c,
1605         testsuite/libffi.call/cls_align_longdouble.c,
1606         testsuite/libffi.call/cls_align_longdouble_split.c,
1607         testsuite/libffi.call/cls_align_longdouble_split2.c: mark expected
1608         failures on x86_64 cygwin/mingw.
1609
1610 2008-12-22  Timothy Wall  <twall@users.sf.net>
1611
1612         * testsuite/libffi.call/closure_fn0.c,
1613         testsuite/libffi.call/closure_fn1.c,
1614         testsuite/libffi.call/closure_fn2.c,
1615         testsuite/libffi.call/closure_fn3.c,
1616         testsuite/libffi.call/closure_fn4.c,
1617         testsuite/libffi.call/closure_fn5.c,
1618         testsuite/libffi.call/closure_fn6.c,
1619         testsuite/libffi.call/closure_loc_fn0.c,
1620         testsuite/libffi.call/closure_stdcall.c,
1621         testsuite/libffi.call/cls_align_pointer.c,
1622         testsuite/libffi.call/cls_pointer.c,
1623         testsuite/libffi.call/cls_pointer_stack.c: use portable cast from
1624         pointer to integer (intptr_t).
1625         * testsuite/libffi.call/cls_longdouble.c: disable for win64.
1626
1627 2008-07-24  Anthony Green  <green@redhat.com>
1628
1629         * testsuite/libffi.call/cls_dbls_struct.c,
1630         testsuite/libffi.call/cls_double_va.c,
1631         testsuite/libffi.call/cls_longdouble.c,
1632         testsuite/libffi.call/cls_longdouble_va.c,
1633         testsuite/libffi.call/cls_pointer.c,
1634         testsuite/libffi.call/cls_pointer_stack.c,
1635         testsuite/libffi.call/err_bad_abi.c: Clean up failures from
1636         compiler warnings.
1637
1638 2008-03-04  Anthony Green  <green@redhat.com>
1639             Blake Chaffin
1640             hos@tamanegi.org
1641
1642         * testsuite/libffi.call/cls_align_longdouble_split2.c
1643           testsuite/libffi.call/cls_align_longdouble_split.c
1644           testsuite/libffi.call/cls_dbls_struct.c
1645           testsuite/libffi.call/cls_double_va.c
1646           testsuite/libffi.call/cls_longdouble.c
1647           testsuite/libffi.call/cls_longdouble_va.c
1648           testsuite/libffi.call/cls_pointer.c
1649           testsuite/libffi.call/cls_pointer_stack.c
1650           testsuite/libffi.call/err_bad_abi.c
1651           testsuite/libffi.call/err_bad_typedef.c
1652           testsuite/libffi.call/stret_large2.c
1653           testsuite/libffi.call/stret_large.c
1654           testsuite/libffi.call/stret_medium2.c
1655           testsuite/libffi.call/stret_medium.c: New tests from Apple.
1656
1657 2009-06-05  Andrew Haley  <aph@redhat.com>
1658
1659         * src/x86/ffitarget.h, src/x86/ffi.c: Merge stdcall changes from
1660         libffi.
1661
1662 2009-06-04  Andrew Haley  <aph@redhat.com>
1663
1664         * src/x86/ffitarget.h, src/x86/win32.S, src/x86/ffi.c: Back out
1665         stdcall changes.
1666
1667 2008-02-26  Anthony Green  <green@redhat.com>
1668             Thomas Heller  <theller@ctypes.org>
1669
1670         * src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C
1671         comment.
1672
1673 2008-02-03  Timothy Wall  <twall@users.sf.net>
1674
1675         * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return
1676           offset based on code pointer, not data pointer.
1677
1678 2008-01-31  Timothy Wall <twall@users.sf.net>
1679
1680         * testsuite/libffi.call/closure_stdcall.c: Add test for stdcall
1681         closures.
1682         * src/x86/ffitarget.h: Increase size of trampoline for stdcall
1683         closures.
1684         * src/x86/win32.S: Add assembly for stdcall closure.
1685         * src/x86/ffi.c: Initialize stdcall closure trampoline.
1686
1687 2009-06-04  Andrew Haley  <aph@redhat.com>
1688
1689         * include/ffi.h.in: Change void (*)() to void (*)(void).
1690         * src/x86/ffi.c: Likewise.
1691
1692 2009-06-04  Andrew Haley  <aph@redhat.com>
1693
1694         * src/powerpc/ppc_closure.S: Insert licence header.
1695         * src/powerpc/linux64_closure.S: Likewise.
1696         * src/m68k/sysv.S: Likewise.
1697
1698         * src/sh64/ffi.c: Change void (*)() to void (*)(void).
1699         * src/powerpc/ffi.c: Likewise.
1700         * src/powerpc/ffi_darwin.c: Likewise.
1701         * src/m32r/ffi.c: Likewise.
1702         * src/sh64/ffi.c: Likewise.
1703         * src/x86/ffi64.c: Likewise.
1704         * src/alpha/ffi.c: Likewise.
1705         * src/alpha/osf.S: Likewise.
1706         * src/frv/ffi.c: Likewise.
1707         * src/s390/ffi.c: Likewise.
1708         * src/pa/ffi.c: Likewise.
1709         * src/pa/hpux32.S: Likewise.
1710         * src/ia64/unix.S: Likewise.
1711         * src/ia64/ffi.c: Likewise.
1712         * src/sparc/ffi.c: Likewise.
1713         * src/mips/ffi.c: Likewise.
1714         * src/sh/ffi.c: Likewise.
1715
1716 2008-02-15  David Daney  <ddaney@avtrex.com>
1717
1718         * src/mips/ffi.c (USE__BUILTIN___CLEAR_CACHE):
1719         Define (conditionally), and use it to include cachectl.h.
1720         (ffi_prep_closure_loc): Fix cache flushing.
1721         * src/mips/ffitarget.h (_ABIN32, _ABI64, _ABIO32): Define.
1722
1723 2009-06-04  Andrew Haley  <aph@redhat.com>
1724
1725         include/ffi.h.in,
1726         src/arm/ffitarget.h,
1727         src/arm/ffi.c,
1728         src/arm/sysv.S,
1729         src/powerpc/ffitarget.h,
1730         src/closures.c,
1731         src/sh64/ffitarget.h,
1732         src/sh64/ffi.c,
1733         src/sh64/sysv.S,
1734         src/types.c,
1735         src/x86/ffi64.c,
1736         src/x86/ffitarget.h,
1737         src/x86/win32.S,
1738         src/x86/darwin.S,
1739         src/x86/ffi.c,
1740         src/x86/sysv.S,
1741         src/x86/unix64.S,
1742         src/alpha/ffitarget.h,
1743         src/alpha/ffi.c,
1744         src/alpha/osf.S,
1745         src/m68k/ffitarget.h,
1746         src/frv/ffitarget.h,
1747         src/frv/ffi.c,
1748         src/s390/ffitarget.h,
1749         src/s390/sysv.S,
1750         src/cris/ffitarget.h,
1751         src/pa/linux.S,
1752         src/pa/ffitarget.h,
1753         src/pa/ffi.c,
1754         src/raw_api.c,
1755         src/ia64/ffitarget.h,
1756         src/ia64/unix.S,
1757         src/ia64/ffi.c,
1758         src/ia64/ia64_flags.h,
1759         src/java_raw_api.c,
1760         src/debug.c,
1761         src/sparc/v9.S,
1762         src/sparc/ffitarget.h,
1763         src/sparc/ffi.c,
1764         src/sparc/v8.S,
1765         src/mips/ffitarget.h,
1766         src/mips/n32.S,
1767         src/mips/o32.S,
1768         src/mips/ffi.c,
1769         src/prep_cif.c,
1770         src/sh/ffitarget.h,
1771         src/sh/ffi.c,
1772         src/sh/sysv.S: Update license text.
1773
1774 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
1775
1776         * src/x86/win32.S (_ffi_closure_STDCALL):  New function.
1777         (.eh_frame):  Add FDE for it.
1778
1779 2009-05-22  Dave Korn  <dave.korn.cygwin@gmail.com>
1780
1781         * configure.ac:  Also check if assembler supports pc-relative
1782         relocs on X86_WIN32 targets.
1783         * configure:  Regenerate.
1784         * src/x86/win32.S (ffi_prep_args):  Declare extern, not global.
1785         (_ffi_call_SYSV):  Add missing function type symbol .def and
1786         add EH markup labels.
1787         (_ffi_call_STDCALL):  Likewise.
1788         (_ffi_closure_SYSV):  Likewise.
1789         (_ffi_closure_raw_SYSV):  Likewise.
1790         (.eh_frame):  Add hand-crafted EH data.
1791
1792 2009-04-09  Jakub Jelinek  <jakub@redhat.com>
1793
1794         * testsuite/lib/libffi-dg.exp: Change copyright header to refer to
1795         version 3 of the GNU General Public License and to point readers
1796         at the COPYING3 file and the FSF's license web page.
1797         * testsuite/libffi.call/call.exp: Likewise.
1798         * testsuite/libffi.special/special.exp: Likewise.
1799
1800 2009-03-01  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1801
1802         * configure: Regenerate.
1803
1804 2008-12-18  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
1805
1806         PR libffi/26048
1807         * configure.ac (HAVE_AS_X86_PCREL): New test.
1808         * configure: Regenerate.
1809         * fficonfig.h.in: Regenerate.
1810         * src/x86/sysv.S [!FFI_NO_RAW_API]: Precalculate
1811         RAW_CLOSURE_CIF_OFFSET, RAW_CLOSURE_FUN_OFFSET,
1812         RAW_CLOSURE_USER_DATA_OFFSET for the Solaris 10/x86 assembler.
1813         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
1814         * src/x86/unix64.S (.Lstore_table): Move to .text section.
1815         (.Lload_table): Likewise.
1816         (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL.
1817
1818 2008-12-18  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1819
1820         * configure: Regenerate.
1821
1822 2008-11-21  Eric Botcazou  <ebotcazou@adacore.com>
1823
1824         * src/sparc/ffi.c (ffi_prep_cif_machdep): Add support for
1825         signed/unsigned int8/16 return values.
1826         * src/sparc/v8.S (ffi_call_v8): Likewise.
1827         (ffi_closure_v8): Likewise.
1828
1829 2008-09-26  Peter O'Gorman  <pogma@thewrittenword.com>
1830             Steve Ellcey  <sje@cup.hp.com>
1831
1832         * configure: Regenerate for new libtool.
1833         * Makefile.in: Ditto.
1834         * include/Makefile.in: Ditto.
1835         * aclocal.m4: Ditto.
1836
1837 2008-08-25  Andreas Tobler  <a.tobler@schweiz.org>
1838
1839         * src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and
1840         FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum.
1841         Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT.
1842         Adjust copyright notice.
1843         * src/powerpc/ffi.c: Add two new flags to indicate if we have one
1844         register or two register to use for FFI_SYSV structs.
1845         (ffi_prep_cif_machdep): Pass the right register flag introduced above.
1846         (ffi_closure_helper_SYSV): Fix the return type for
1847         FFI_SYSV_TYPE_SMALL_STRUCT. Comment.
1848         Adjust copyright notice.
1849
1850 2008-07-16  Kaz Kojima  <kkojima@gcc.gnu.org>
1851
1852         * src/sh/ffi.c (ffi_prep_closure_loc): Turn INSN into an unsigned
1853         int.
1854
1855 2008-06-17  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1856
1857         * configure: Regenerate.
1858         * include/Makefile.in: Regenerate.
1859         * testsuite/Makefile.in: Regenerate.
1860
1861 2008-06-07  Joseph Myers  <joseph@codesourcery.com>
1862
1863         * configure.ac (parisc*-*-linux*, powerpc-*-sysv*,
1864         powerpc-*-beos*): Remove.
1865         * configure: Regenerate.
1866
1867 2008-05-09  Julian Brown  <julian@codesourcery.com>
1868
1869         * Makefile.am (LTLDFLAGS): New.
1870         (libffi_la_LDFLAGS): Use above.
1871         * Makefile.in: Regenerate.
1872
1873 2008-04-18  Paolo Bonzini  <bonzini@gnu.org>
1874
1875         PR bootstrap/35457
1876         * aclocal.m4: Regenerate.
1877         * configure: Regenerate.
1878
1879 2008-03-26  Kaz Kojima  <kkojima@gcc.gnu.org>
1880
1881         * src/sh/sysv.S: Add .note.GNU-stack on Linux.
1882         * src/sh64/sysv.S: Likewise.
1883
1884 2008-03-26  Daniel Jacobowitz  <dan@debian.org>
1885
1886         * src/arm/sysv.S: Fix ARM comment marker.
1887
1888 2008-03-26  Jakub Jelinek  <jakub@redhat.com>
1889
1890         * src/alpha/osf.S: Add .note.GNU-stack on Linux.
1891         * src/s390/sysv.S: Likewise.
1892         * src/powerpc/ppc_closure.S: Likewise.
1893         * src/powerpc/sysv.S: Likewise.
1894         * src/x86/unix64.S: Likewise.
1895         * src/x86/sysv.S: Likewise.
1896         * src/sparc/v8.S: Likewise.
1897         * src/sparc/v9.S: Likewise.
1898         * src/m68k/sysv.S: Likewise.
1899         * src/arm/sysv.S: Likewise.
1900
1901 2008-03-16  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
1902
1903         * aclocal.m4: Regenerate.
1904         * configure: Likewise.
1905         * Makefile.in: Likewise.
1906         * include/Makefile.in: Likewise.
1907         * testsuite/Makefile.in: Likewise.
1908
1909 2008-02-12  Bjoern Koenig  <bkoenig@alpha-tierchen.de>
1910             Andreas Tobler  <a.tobler@schweiz.org>
1911
1912         * configure.ac: Add amd64-*-freebsd* target.
1913         * configure: Regenerate.
1914
1915 2008-01-30  H.J. Lu  <hongjiu.lu@intel.com>
1916
1917         PR libffi/34612
1918         * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when
1919         returning struct.
1920
1921         * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer"
1922         tests.
1923
1924 2008-01-24  David Edelsohn  <edelsohn@gnu.org>
1925
1926         * configure: Regenerate.
1927
1928 2008-01-06  Andreas Tobler  <a.tobler@schweiz.org>
1929
1930         * src/x86/ffi.c (ffi_prep_cif_machdep): Fix thinko.
1931
1932 2008-01-05  Andreas Tobler  <a.tobler@schweiz.org>
1933
1934         PR testsuite/32843
1935         * src/x86/ffi.c (ffi_prep_cif_machdep): Add code for
1936         signed/unsigned int8/16 for X86_DARWIN.
1937         Updated copyright info.
1938         Handle one and two byte structs with special cif->flags.
1939         * src/x86/ffitarget.h: Add special types for one and two byte structs.
1940         Updated copyright info.
1941         * src/x86/darwin.S (ffi_call_SYSV): Rewrite to use a jump table like
1942         sysv.S
1943         Remove code to pop args from the stack after call.
1944         Special-case signed/unsigned for int8/16, one and two byte structs.
1945         (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
1946         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
1947         FFI_TYPE_SINT32.
1948         Updated copyright info.
1949
1950 2007-12-08  David Daney  <ddaney@avtrex.com>
1951
1952         * src/mips/n32.S (ffi_call_N32):  Replace dadd with ADDU, dsub with
1953         SUBU, add with ADDU and use smaller code sequences.
1954
1955 2007-12-07  David Daney  <ddaney@avtrex.com>
1956
1957         * src/mips/ffi.c (ffi_prep_cif_machdep): Handle long double return
1958         type.
1959
1960 2007-12-06  David Daney  <ddaney@avtrex.com>
1961
1962         * include/ffi.h.in (FFI_SIZEOF_JAVA_RAW): Define if not already
1963         defined.
1964         (ffi_java_raw): New typedef.
1965         (ffi_java_raw_call, ffi_java_ptrarray_to_raw,
1966         ffi_java_raw_to_ptrarray): Change parameter types from ffi_raw to
1967         ffi_java_raw.
1968         (ffi_java_raw_closure) : Same.
1969         (ffi_prep_java_raw_closure, ffi_prep_java_raw_closure_loc): Change
1970         parameter types.
1971         * src/java_raw_api.c (ffi_java_raw_size):  Replace FFI_SIZEOF_ARG with
1972         FFI_SIZEOF_JAVA_RAW.
1973         (ffi_java_raw_to_ptrarray): Change type of raw to ffi_java_raw.
1974         Replace FFI_SIZEOF_ARG with FFI_SIZEOF_JAVA_RAW. Use
1975         sizeof(ffi_java_raw) for alignment calculations.
1976         (ffi_java_ptrarray_to_raw): Same.
1977         (ffi_java_rvalue_to_raw): Add special handling for FFI_TYPE_POINTER
1978         if FFI_SIZEOF_JAVA_RAW == 4.
1979         (ffi_java_raw_to_rvalue): Same.
1980         (ffi_java_raw_call): Change type of raw to ffi_java_raw.
1981         (ffi_java_translate_args): Same.
1982         (ffi_prep_java_raw_closure_loc, ffi_prep_java_raw_closure): Change
1983         parameter types.
1984         * src/mips/ffitarget.h (FFI_SIZEOF_JAVA_RAW): Define for N32 ABI.
1985
1986 2007-12-06  David Daney  <ddaney@avtrex.com>
1987
1988         * src/mips/n32.S (ffi_closure_N32): Use 64-bit add instruction on
1989         pointer values.
1990
1991 2007-12-01  Andreas Tobler  <a.tobler@schweiz.org>
1992
1993         PR libffi/31937
1994         * src/powerpc/ffitarget.h: Introduce new ABI FFI_LINUX_SOFT_FLOAT.
1995         Add local FFI_TYPE_UINT128 to handle soft-float long-double-128.
1996         * src/powerpc/ffi.c: Distinguish between __NO_FPRS__ and not and
1997         set the NUM_FPR_ARG_REGISTERS according to.
1998         Add support for potential soft-float support under hard-float
1999         architecture.
2000         (ffi_prep_args_SYSV): Set NUM_FPR_ARG_REGISTERS to 0 in case of
2001         FFI_LINUX_SOFT_FLOAT, handle float, doubles and long-doubles according
2002         to the FFI_LINUX_SOFT_FLOAT ABI.
2003         (ffi_prep_cif_machdep): Likewise.
2004         (ffi_closure_helper_SYSV): Likewise.
2005         * src/powerpc/ppc_closure.S: Make sure not to store float/double
2006         on archs where __NO_FPRS__ is true.
2007         Add FFI_TYPE_UINT128 support.
2008         * src/powerpc/sysv.S: Add support for soft-float long-double-128.
2009         Adjust copyright notice.
2010
2011 2007-11-25  Andreas Tobler  <a.tobler@schweiz.org>
2012
2013         * src/closures.c: Move defintion of MAYBE_UNUSED from here to ...
2014         * include/ffi_common.h: ... here.
2015         Update copyright.
2016
2017 2007-11-17  Andreas Tobler  <a.tobler@schweiz.org>
2018
2019         * src/powerpc/sysv.S: Load correct cr to compare if we have long double.
2020         * src/powerpc/linux64.S: Likewise.
2021         * src/powerpc/ffi.c: Add a comment to show which part goes into cr6.
2022         * testsuite/libffi.call/return_ldl.c: New test.
2023
2024 2007-09-04    <aph@redhat.com>
2025
2026         * src/arm/sysv.S (UNWIND): New.
2027         (Whole file): Conditionally compile unwinder directives.
2028         * src/arm/sysv.S: Add unwinder directives.
2029
2030         * src/arm/ffi.c (ffi_prep_args): Align structs by at least 4 bytes.
2031         Only treat r0 as a struct address if we're actually returning a
2032         struct by address.
2033         Only copy the bytes that are actually within a struct.
2034         (ffi_prep_cif_machdep): A Composite Type not larger than 4 bytes
2035         is returned in r0, not passed by address.
2036         (ffi_call): Allocate a word-sized temporary for the case where
2037         a composite is returned in r0.
2038         (ffi_prep_incoming_args_SYSV): Align as necessary.
2039
2040 2007-08-05  Steven Newbury  <s_j_newbury@yahoo.co.uk>
2041
2042         * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Use __clear_cache instead of
2043         directly using the sys_cacheflush syscall.
2044
2045 2007-07-27  Andrew Haley  <aph@redhat.com>
2046
2047         * src/arm/sysv.S (ffi_closure_SYSV): Add soft-float.
2048
2049 2007-09-03  Maciej W. Rozycki  <macro@linux-mips.org>
2050
2051         * Makefile.am: Unify MIPS_IRIX and MIPS_LINUX into MIPS.
2052         * configure.ac: Likewise.
2053         * Makefile.in: Regenerate.
2054         * include/Makefile.in: Likewise.
2055         * testsuite/Makefile.in: Likewise.
2056         * configure: Likewise.
2057
2058 2007-08-24  David Daney  <ddaney@avtrex.com>
2059
2060         * testsuite/libffi.call/return_sl.c: New test.
2061
2062 2007-08-10  David Daney  <ddaney@avtrex.com>
2063
2064         * testsuite/libffi.call/cls_multi_ushort.c,
2065         testsuite/libffi.call/cls_align_uint16.c,
2066         testsuite/libffi.call/nested_struct1.c,
2067         testsuite/libffi.call/nested_struct3.c,
2068         testsuite/libffi.call/cls_7_1_byte.c,
2069         testsuite/libffi.call/nested_struct5.c,
2070         testsuite/libffi.call/cls_double.c,
2071         testsuite/libffi.call/nested_struct7.c,
2072         testsuite/libffi.call/cls_sint.c,
2073         testsuite/libffi.call/nested_struct9.c,
2074         testsuite/libffi.call/cls_20byte1.c,
2075         testsuite/libffi.call/cls_multi_sshortchar.c,
2076         testsuite/libffi.call/cls_align_sint64.c,
2077         testsuite/libffi.call/cls_3byte2.c,
2078         testsuite/libffi.call/cls_multi_schar.c,
2079         testsuite/libffi.call/cls_multi_uchar.c,
2080         testsuite/libffi.call/cls_19byte.c,
2081         testsuite/libffi.call/cls_9byte1.c,
2082         testsuite/libffi.call/cls_align_float.c,
2083         testsuite/libffi.call/closure_fn1.c,
2084         testsuite/libffi.call/problem1.c,
2085         testsuite/libffi.call/closure_fn3.c,
2086         testsuite/libffi.call/cls_sshort.c,
2087         testsuite/libffi.call/closure_fn5.c,
2088         testsuite/libffi.call/cls_align_double.c,
2089         testsuite/libffi.call/nested_struct.c,
2090         testsuite/libffi.call/cls_2byte.c,
2091         testsuite/libffi.call/nested_struct10.c,
2092         testsuite/libffi.call/cls_4byte.c,
2093         testsuite/libffi.call/cls_6byte.c,
2094         testsuite/libffi.call/cls_8byte.c,
2095         testsuite/libffi.call/cls_multi_sshort.c,
2096         testsuite/libffi.call/cls_align_sint16.c,
2097         testsuite/libffi.call/cls_align_uint32.c,
2098         testsuite/libffi.call/cls_20byte.c,
2099         testsuite/libffi.call/cls_float.c,
2100         testsuite/libffi.call/nested_struct2.c,
2101         testsuite/libffi.call/cls_5_1_byte.c,
2102         testsuite/libffi.call/nested_struct4.c,
2103         testsuite/libffi.call/cls_24byte.c,
2104         testsuite/libffi.call/nested_struct6.c,
2105         testsuite/libffi.call/cls_64byte.c,
2106         testsuite/libffi.call/nested_struct8.c,
2107         testsuite/libffi.call/cls_uint.c,
2108         testsuite/libffi.call/cls_multi_ushortchar.c,
2109         testsuite/libffi.call/cls_schar.c,
2110         testsuite/libffi.call/cls_uchar.c,
2111         testsuite/libffi.call/cls_align_uint64.c,
2112         testsuite/libffi.call/cls_ulonglong.c,
2113         testsuite/libffi.call/cls_align_longdouble.c,
2114         testsuite/libffi.call/cls_1_1byte.c,
2115         testsuite/libffi.call/cls_12byte.c,
2116         testsuite/libffi.call/cls_3_1byte.c,
2117         testsuite/libffi.call/cls_3byte1.c,
2118         testsuite/libffi.call/cls_4_1byte.c,
2119         testsuite/libffi.call/cls_6_1_byte.c,
2120         testsuite/libffi.call/cls_16byte.c,
2121         testsuite/libffi.call/cls_18byte.c,
2122         testsuite/libffi.call/closure_fn0.c,
2123         testsuite/libffi.call/cls_9byte2.c,
2124         testsuite/libffi.call/closure_fn2.c,
2125         testsuite/libffi.call/closure_fn4.c,
2126         testsuite/libffi.call/cls_ushort.c,
2127         testsuite/libffi.call/closure_fn6.c,
2128         testsuite/libffi.call/cls_5byte.c,
2129         testsuite/libffi.call/cls_align_pointer.c,
2130         testsuite/libffi.call/cls_7byte.c,
2131         testsuite/libffi.call/cls_align_sint32.c,
2132         testsuite/libffi.special/unwindtest_ffi_call.cc,
2133         testsuite/libffi.special/unwindtest.cc: Remove xfail for mips64*-*-*.
2134
2135 2007-08-10  David Daney  <ddaney@avtrex.com>
2136
2137         PR libffi/28313
2138         * configure.ac: Don't treat mips64 as a special case.
2139         * Makefile.am (nodist_libffi_la_SOURCES): Add n32.S.
2140         * configure: Regenerate
2141         * Makefile.in: Ditto.
2142         * fficonfig.h.in: Ditto.
2143         * src/mips/ffitarget.h (REG_L, REG_S, SUBU, ADDU, SRL, LI): Indent.
2144         (LA, EH_FRAME_ALIGN, FDE_ADDR_BYTES): New preprocessor macros.
2145         (FFI_DEFAULT_ABI): Set for n64 case.
2146         (FFI_CLOSURES, FFI_TRAMPOLINE_SIZE): Define for n32 and n64 cases.
2147         * src/mips/n32.S (ffi_call_N32): Add debug macros and labels for FDE.
2148         (ffi_closure_N32): New function.
2149         (.eh_frame): New section
2150         * src/mips/o32.S: Clean up comments.
2151         (ffi_closure_O32): Pass ffi_closure parameter in $12.
2152         * src/mips/ffi.c: Use FFI_MIPS_N32 instead of
2153         _MIPS_SIM == _ABIN32 throughout.
2154         (FFI_MIPS_STOP_HERE): New, use in place of
2155         ffi_stop_here.
2156         (ffi_prep_args): Use unsigned long to hold pointer values.  Rewrite
2157         to support n32/n64 ABIs.
2158         (calc_n32_struct_flags): Rewrite.
2159         (calc_n32_return_struct_flags): Remove unused variable.  Reverse
2160         position of flag bits.
2161         (ffi_prep_cif_machdep): Rewrite n32 portion.
2162         (ffi_call): Enable for n64.  Add special handling for small structure
2163         return values.
2164         (ffi_prep_closure_loc): Add n32 and n64 support.
2165         (ffi_closure_mips_inner_O32): Add cast to silence warning.
2166         (copy_struct_N32, ffi_closure_mips_inner_N32): New functions.
2167
2168 2007-08-08  David Daney  <ddaney@avtrex.com>
2169
2170         * testsuite/libffi.call/ffitest.h (ffi_type_mylong): Remove definition.
2171         * testsuite/libffi.call/cls_align_uint16.c (main): Use correct type
2172         specifiers.
2173         * testsuite/libffi.call/nested_struct1.c (main): Ditto.
2174         * testsuite/libffi.call/cls_sint.c (main): Ditto.
2175         * testsuite/libffi.call/nested_struct9.c (main): Ditto.
2176         * testsuite/libffi.call/cls_20byte1.c (main): Ditto.
2177         * testsuite/libffi.call/cls_9byte1.c (main): Ditto.
2178         * testsuite/libffi.call/closure_fn1.c (main): Ditto.
2179         * testsuite/libffi.call/closure_fn3.c (main): Ditto.
2180         * testsuite/libffi.call/return_dbl2.c (main): Ditto.
2181         * testsuite/libffi.call/cls_sshort.c (main): Ditto.
2182         * testsuite/libffi.call/return_fl3.c (main): Ditto.
2183         * testsuite/libffi.call/closure_fn5.c (main): Ditto.
2184         * testsuite/libffi.call/nested_struct.c (main): Ditto.
2185         * testsuite/libffi.call/nested_struct10.c (main): Ditto.
2186         * testsuite/libffi.call/return_ll1.c (main): Ditto.
2187         * testsuite/libffi.call/cls_8byte.c (main): Ditto.
2188         * testsuite/libffi.call/cls_align_uint32.c (main): Ditto.
2189         * testsuite/libffi.call/cls_align_sint16.c (main): Ditto.
2190         * testsuite/libffi.call/cls_20byte.c (main): Ditto.
2191         * testsuite/libffi.call/nested_struct2.c (main): Ditto.
2192         * testsuite/libffi.call/cls_24byte.c (main): Ditto.
2193         * testsuite/libffi.call/nested_struct6.c (main): Ditto.
2194         * testsuite/libffi.call/cls_uint.c (main): Ditto.
2195         * testsuite/libffi.call/cls_12byte.c (main): Ditto.
2196         * testsuite/libffi.call/cls_16byte.c (main): Ditto.
2197         * testsuite/libffi.call/closure_fn0.c (main): Ditto.
2198         * testsuite/libffi.call/cls_9byte2.c (main): Ditto.
2199         * testsuite/libffi.call/closure_fn2.c (main): Ditto.
2200         * testsuite/libffi.call/return_dbl1.c (main): Ditto.
2201         * testsuite/libffi.call/closure_fn4.c (main): Ditto.
2202         * testsuite/libffi.call/closure_fn6.c (main): Ditto.
2203         * testsuite/libffi.call/cls_align_sint32.c (main): Ditto.
2204
2205 2007-08-07  Andrew Haley  <aph@redhat.com>
2206
2207         * src/x86/sysv.S (ffi_closure_raw_SYSV): Fix typo in previous
2208         checkin.
2209
2210 2007-08-06  Andrew Haley  <aph@redhat.com>
2211
2212         PR testsuite/32843
2213         * src/x86/sysv.S (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
2214         FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
2215         FFI_TYPE_SINT32.
2216
2217 2007-08-02  David Daney  <ddaney@avtrex.com>
2218
2219         * testsuite/libffi.call/return_ul.c (main): Define return type as
2220         ffi_arg.  Use proper printf conversion specifier.
2221
2222 2007-07-30  Andrew Haley  <aph@redhat.com>
2223
2224         PR testsuite/32843
2225         * src/x86/ffi.c (ffi_prep_cif_machdep): in x86 case, add code for
2226         signed/unsigned int8/16.
2227         * src/x86/sysv.S (ffi_call_SYSV): Rewrite to:
2228         Use a jump table.
2229         Remove code to pop args from the stack after call.
2230         Special-case signed/unsigned int8/16.
2231         * testsuite/libffi.call/return_sc.c (main): Revert.
2232
2233 2007-07-26  Richard Guenther  <rguenther@suse.de>
2234
2235         PR testsuite/32843
2236         * testsuite/libffi.call/return_sc.c (main): Verify call
2237         result as signed char, not ffi_arg.
2238
2239 2007-07-16  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
2240
2241         * configure.ac (i?86-*-solaris2.1[0-9]): Set TARGET to X86_64.
2242         * configure: Regenerate.
2243
2244 2007-07-11  David Daney  <ddaney@avtrex.com>
2245
2246         * src/mips/ffi.c: Don't include sys/cachectl.h.
2247         (ffi_prep_closure_loc): Use __builtin___clear_cache() instead of
2248         cacheflush().
2249
2250 2007-05-18  Aurelien Jarno  <aurelien@aurel32.net>
2251
2252         * src/arm/ffi.c (ffi_prep_closure_loc): Renamed and ajusted
2253         from (ffi_prep_closure): ... this.
2254         (FFI_INIT_TRAMPOLINE): Adjust.
2255
2256 2005-12-31  Phil Blundell  <pb@reciva.com>
2257
2258         * src/arm/ffi.c (ffi_prep_incoming_args_SYSV,
2259         ffi_closure_SYSV_inner, ffi_prep_closure): New, add closure support.
2260         * src/arm/sysv.S(ffi_closure_SYSV): Likewise.
2261         * src/arm/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
2262         (FFI_CLOSURES): Enable closure support.
2263
2264 2007-07-03  Andrew Haley  <aph@hedges.billgatliff.com>
2265
2266         * testsuite/libffi.call/cls_multi_ushort.c,
2267         testsuite/libffi.call/cls_align_uint16.c,
2268         testsuite/libffi.call/nested_struct1.c,
2269         testsuite/libffi.call/nested_struct3.c,
2270         testsuite/libffi.call/cls_7_1_byte.c,
2271         testsuite/libffi.call/cls_double.c,
2272         testsuite/libffi.call/nested_struct5.c,
2273         testsuite/libffi.call/nested_struct7.c,
2274         testsuite/libffi.call/cls_sint.c,
2275         testsuite/libffi.call/nested_struct9.c,
2276         testsuite/libffi.call/cls_20byte1.c,
2277         testsuite/libffi.call/cls_multi_sshortchar.c,
2278         testsuite/libffi.call/cls_align_sint64.c,
2279         testsuite/libffi.call/cls_3byte2.c,
2280         testsuite/libffi.call/cls_multi_schar.c,
2281         testsuite/libffi.call/cls_multi_uchar.c,
2282         testsuite/libffi.call/cls_19byte.c,
2283         testsuite/libffi.call/cls_9byte1.c,
2284         testsuite/libffi.call/cls_align_float.c,
2285         testsuite/libffi.call/closure_fn1.c,
2286         testsuite/libffi.call/problem1.c,
2287         testsuite/libffi.call/closure_fn3.c,
2288         testsuite/libffi.call/cls_sshort.c,
2289         testsuite/libffi.call/closure_fn5.c,
2290         testsuite/libffi.call/cls_align_double.c,
2291         testsuite/libffi.call/cls_2byte.c,
2292         testsuite/libffi.call/nested_struct.c,
2293         testsuite/libffi.call/nested_struct10.c,
2294         testsuite/libffi.call/cls_4byte.c,
2295         testsuite/libffi.call/cls_6byte.c,
2296         testsuite/libffi.call/cls_8byte.c,
2297         testsuite/libffi.call/cls_multi_sshort.c,
2298         testsuite/libffi.call/cls_align_uint32.c,
2299         testsuite/libffi.call/cls_align_sint16.c,
2300         testsuite/libffi.call/cls_float.c,
2301         testsuite/libffi.call/cls_20byte.c,
2302         testsuite/libffi.call/cls_5_1_byte.c,
2303         testsuite/libffi.call/nested_struct2.c,
2304         testsuite/libffi.call/cls_24byte.c,
2305         testsuite/libffi.call/nested_struct4.c,
2306         testsuite/libffi.call/nested_struct6.c,
2307         testsuite/libffi.call/cls_64byte.c,
2308         testsuite/libffi.call/nested_struct8.c,
2309         testsuite/libffi.call/cls_uint.c,
2310         testsuite/libffi.call/cls_multi_ushortchar.c,
2311         testsuite/libffi.call/cls_schar.c,
2312         testsuite/libffi.call/cls_uchar.c,
2313         testsuite/libffi.call/cls_align_uint64.c,
2314         testsuite/libffi.call/cls_ulonglong.c,
2315         testsuite/libffi.call/cls_align_longdouble.c,
2316         testsuite/libffi.call/cls_1_1byte.c,
2317         testsuite/libffi.call/cls_12byte.c,
2318         testsuite/libffi.call/cls_3_1byte.c,
2319         testsuite/libffi.call/cls_3byte1.c,
2320         testsuite/libffi.call/cls_4_1byte.c,
2321         testsuite/libffi.call/cls_6_1_byte.c,
2322         testsuite/libffi.call/cls_16byte.c,
2323         testsuite/libffi.call/cls_18byte.c,
2324         testsuite/libffi.call/closure_fn0.c,
2325         testsuite/libffi.call/cls_9byte2.c,
2326         testsuite/libffi.call/closure_fn2.c,
2327         testsuite/libffi.call/closure_fn4.c,
2328         testsuite/libffi.call/cls_ushort.c,
2329         testsuite/libffi.call/closure_fn6.c,
2330         testsuite/libffi.call/cls_5byte.c,
2331         testsuite/libffi.call/cls_align_pointer.c,
2332         testsuite/libffi.call/cls_7byte.c,
2333         testsuite/libffi.call/cls_align_sint32.c,
2334         testsuite/libffi.special/unwindtest_ffi_call.cc,
2335         testsuite/libffi.special/unwindtest.cc: Enable for ARM.
2336
2337 2007-07-05  H.J. Lu  <hongjiu.lu@intel.com>
2338
2339         * aclocal.m4: Regenerated.
2340
2341 2007-06-02  Paolo Bonzini  <bonzini@gnu.org>
2342
2343         * configure: Regenerate.
2344
2345 2007-05-23  Steve Ellcey  <sje@cup.hp.com>
2346
2347         * Makefile.in: Regenerate.
2348         * configure: Regenerate.
2349         * aclocal.m4: Regenerate.
2350         * include/Makefile.in: Regenerate.
2351         * testsuite/Makefile.in: Regenerate.
2352
2353 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
2354
2355         * src/m68k/ffi.c (ffi_prep_incoming_args_SYSV,
2356         ffi_closure_SYSV_inner,ffi_prep_closure): New, add closure support.
2357         * src/m68k/sysv.S(ffi_closure_SYSV,ffi_closure_struct_SYSV): Likewise.
2358         * src/m68k/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
2359         (FFI_CLOSURES): Enable closure support.
2360
2361 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
2362
2363         * configure.ac (HAVE_AS_CFI_PSEUDO_OP): New test.
2364         * configure: Regenerate.
2365         * fficonfig.h.in: Regenerate.
2366         * src/m68k/sysv.S (CFI_STARTPROC,CFI_ENDPROC,
2367         CFI_OFFSET,CFI_DEF_CFA): New macros.
2368         (ffi_call_SYSV): Add callframe annotation.
2369
2370 2007-05-10  Roman Zippel <zippel@linux-m68k.org>
2371
2372         * src/m68k/ffi.c (ffi_prep_args,ffi_prep_cif_machdep): Fix
2373         numerous test suite failures.
2374         * src/m68k/sysv.S (ffi_call_SYSV): Likewise.
2375
2376 2007-04-11  Paolo Bonzini  <bonzini@gnu.org>
2377
2378         * Makefile.am (EXTRA_DIST): Bring up to date.
2379         * Makefile.in: Regenerate.
2380         * src/frv/eabi.S: Remove RCS keyword.
2381
2382 2007-04-06  Richard Henderson  <rth@redhat.com>
2383
2384         * configure.ac: Tidy target case.
2385         (HAVE_LONG_DOUBLE): Allow the target to override.
2386         * configure: Regenerate.
2387         * include/ffi.h.in: Don't define ffi_type_foo if
2388         LIBFFI_HIDE_BASIC_TYPES is defined.
2389         (ffi_type_longdouble): If not HAVE_LONG_DOUBLE, define
2390         to ffi_type_double.
2391         * types.c (LIBFFI_HIDE_BASIC_TYPES): Define.
2392         (FFI_TYPEDEF, ffi_type_void): Mark the data const.
2393         (ffi_type_longdouble): Special case for Alpha.  Don't define
2394         if long double == double.
2395
2396         * src/alpha/ffi.c (FFI_TYPE_LONGDOUBLE): Assert unique value.
2397         (ffi_prep_cif_machdep): Handle it as the 128-bit type.
2398         (ffi_call, ffi_closure_osf_inner): Likewise.
2399         (ffi_closure_osf_inner): Likewise.  Mark hidden.
2400         (ffi_call_osf, ffi_closure_osf): Mark hidden.
2401         * src/alpha/ffitarget.h (FFI_LAST_ABI): Tidy definition.
2402         * src/alpha/osf.S (ffi_call_osf, ffi_closure_osf): Mark hidden.
2403         (load_table): Handle 128-bit long double.
2404
2405         * testsuite/libffi.call/float4.c: Add -mieee for alpha.
2406
2407 2007-04-06  Tom Tromey  <tromey@redhat.com>
2408
2409         PR libffi/31491:
2410         * README: Fixed bug in example.
2411
2412 2007-04-03  Jakub Jelinek  <jakub@redhat.com>
2413
2414         * src/closures.c: Include sys/statfs.h.
2415         (_GNU_SOURCE): Define on Linux.
2416         (FFI_MMAP_EXEC_SELINUX): Define.
2417         (selinux_enabled): New variable.
2418         (selinux_enabled_check): New function.
2419         (is_selinux_enabled): Define.
2420         (dlmmap): Use it.
2421
2422 2007-03-24  Uros Bizjak  <ubizjak@gmail.com>
2423
2424         * testsuite/libffi.call/return_fl2.c (return_fl): Mark as static.
2425         Use 'volatile float sum' to create sum of floats to avoid false
2426         negative due to excess precision on ix86 targets.
2427         (main): Ditto.
2428
2429 2007-03-08  Alexandre Oliva  <aoliva@redhat.com>
2430
2431         * src/powerpc/ffi.c (flush_icache): Fix left-over from previous
2432         patch.
2433         (ffi_prep_closure_loc): Remove unneeded casts.  Add needed ones.
2434
2435 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
2436
2437         * include/ffi.h.in (ffi_closure_alloc, ffi_closure_free): New.
2438         (ffi_prep_closure_loc): New.
2439         (ffi_prep_raw_closure_loc): New.
2440         (ffi_prep_java_raw_closure_loc): New.
2441         * src/closures.c: New file.
2442         * src/dlmalloc.c [FFI_MMAP_EXEC_WRIT] (struct malloc_segment):
2443         Replace sflags with exec_offset.
2444         [FFI_MMAP_EXEC_WRIT] (mmap_exec_offset, add_segment_exec_offset,
2445         sub_segment_exec_offset): New macros.
2446         (get_segment_flags, set_segment_flags, check_segment_merge): New
2447         macros.
2448         (is_mmapped_segment, is_extern_segment): Use get_segment_flags.
2449         (add_segment, sys_alloc, create_mspace, create_mspace_with_base,
2450         destroy_mspace): Use new macros.
2451         (sys_alloc): Silence warning.
2452         * Makefile.am (libffi_la_SOURCES): Add src/closures.c.
2453         * Makefile.in: Rebuilt.
2454         * src/prep_cif [FFI_CLOSURES] (ffi_prep_closure): Implement in
2455         terms of ffi_prep_closure_loc.
2456         * src/raw_api.c (ffi_prep_raw_closure_loc): Renamed and adjusted
2457         from...
2458         (ffi_prep_raw_closure): ... this.  Re-implement in terms of the
2459         renamed version.
2460         * src/java_raw_api (ffi_prep_java_raw_closure_loc): Renamed and
2461         adjusted from...
2462         (ffi_prep_java_raw_closure): ... this.  Re-implement in terms of
2463         the renamed version.
2464         * src/alpha/ffi.c (ffi_prep_closure_loc): Renamed from
2465         (ffi_prep_closure): ... this.
2466         * src/pa/ffi.c: Likewise.
2467         * src/cris/ffi.c: Likewise.  Adjust.
2468         * src/frv/ffi.c: Likewise.
2469         * src/ia64/ffi.c: Likewise.
2470         * src/mips/ffi.c: Likewise.
2471         * src/powerpc/ffi_darwin.c: Likewise.
2472         * src/s390/ffi.c: Likewise.
2473         * src/sh/ffi.c: Likewise.
2474         * src/sh64/ffi.c: Likewise.
2475         * src/sparc/ffi.c: Likewise.
2476         * src/x86/ffi64.c: Likewise.
2477         * src/x86/ffi.c: Likewise.
2478         (FFI_INIT_TRAMPOLINE): Adjust.
2479         (ffi_prep_raw_closure_loc): Renamed and adjusted from...
2480         (ffi_prep_raw_closure): ... this.
2481         * src/powerpc/ffi.c (ffi_prep_closure_loc): Renamed from
2482         (ffi_prep_closure): ... this.
2483         (flush_icache): Adjust.
2484
2485 2007-03-07  Alexandre Oliva  <aoliva@redhat.com>
2486
2487         * src/dlmalloc.c: New file, imported version 2.8.3 of Doug
2488         Lea's malloc.
2489
2490 2007-03-01  Brooks Moses  <brooks.moses@codesourcery.com>
2491
2492         * Makefile.am: Add dummy install-pdf target.
2493         * Makefile.in: Regenerate
2494
2495 2007-02-13  Andreas Krebbel  <krebbel1@de.ibm.com>
2496
2497         * src/s390/ffi.c (ffi_prep_args, ffi_prep_cif_machdep,
2498         ffi_closure_helper_SYSV): Add long double handling.
2499
2500 2007-02-02  Jakub Jelinek  <jakub@redhat.com>
2501
2502         * src/powerpc/linux64.S (ffi_call_LINUX64): Move restore of r2
2503         immediately after bctrl instruction.
2504
2505 2007-01-18  Alexandre Oliva  <aoliva@redhat.com>
2506
2507         * Makefile.am (all-recursive, install-recursive,
2508         mostlyclean-recursive, clean-recursive, distclean-recursive,
2509         maintainer-clean-recursive): Add missing targets.
2510         * Makefile.in: Rebuilt.
2511
2512 2006-12-14  Andreas Tobler  <a.tobler@schweiz.org>
2513
2514         * configure.ac: Add TARGET for x86_64-*-darwin*.
2515         * Makefile.am (nodist_libffi_la_SOURCES): Add rules for 64-bit sources
2516         for X86_DARWIN.
2517         * src/x86/ffitarget.h: Set trampoline size for x86_64-*-darwin*.
2518         * src/x86/darwin64.S: New file for x86_64-*-darwin* support.
2519         * configure: Regenerate.
2520         * Makefile.in: Regenerate.
2521         * include/Makefile.in: Regenerate.
2522         * testsuite/Makefile.in: Regenerate.
2523         * testsuite/libffi.special/unwindtest_ffi_call.cc: New test case for
2524         ffi_call only.
2525
2526 2006-12-13  Andreas Tobler <a.tobler@schweiz.org>
2527
2528         * aclocal.m4: Regenerate with aclocal -I .. as written in the
2529         Makefile.am.
2530
2531 2006-10-31  Geoffrey Keating  <geoffk@apple.com>
2532
2533         * src/powerpc/ffi_darwin.c (darwin_adjust_aggregate_sizes): New.
2534         (ffi_prep_cif_machdep): Call darwin_adjust_aggregate_sizes for
2535         Darwin.
2536         * testsuite/libffi.call/nested_struct4.c: Remove Darwin XFAIL.
2537         * testsuite/libffi.call/nested_struct6.c: Remove Darwin XFAIL.
2538
2539 2006-10-10  Paolo Bonzini  <bonzini@gnu.org>
2540             Sandro Tolaini  <tolaini@libero.it>
2541
2542         * configure.ac [i*86-*-darwin*]: Set X86_DARWIN symbol and
2543         conditional.
2544         * configure: Regenerated.
2545         * Makefile.am (nodist_libffi_la_SOURCES) [X86_DARWIN]: New case.
2546         (EXTRA_DIST): Add src/x86/darwin.S.
2547         * Makefile.in: Regenerated.
2548         * include/Makefile.in: Regenerated.
2549         * testsuite/Makefile.in: Regenerated.
2550
2551         * src/x86/ffi.c (ffi_prep_cif_machdep) [X86_DARWIN]: Treat like
2552         X86_WIN32, and additionally align stack to 16 bytes.
2553         * src/x86/darwin.S: New, based on sysv.S.
2554         * src/prep_cif.c (ffi_prep_cif) [X86_DARWIN]: Align > 8-byte structs.
2555
2556 2006-09-12  David Daney  <ddaney@avtrex.com>
2557
2558         PR libffi/23935
2559         * include/Makefile.am: Install both ffi.h and ffitarget.h in
2560         $(libdir)/gcc/$(target_alias)/$(gcc_version)/include.
2561         * aclocal.m4: Regenerated for automake 1.9.6.
2562         * Makefile.in: Regenerated.
2563         * include/Makefile.in: Regenerated.
2564         * testsuite/Makefile.in: Regenerated.
2565
2566 2006-08-17  Andreas Tobler  <a.tobler@schweiz.ch>
2567
2568         * include/ffi_common.h (struct): Revert accidental commit.
2569
2570 2006-08-15  Andreas Tobler  <a.tobler@schweiz.ch>
2571
2572         * include/ffi_common.h: Remove lint directives.
2573         * include/ffi.h.in: Likewise.
2574
2575 2006-07-25  Torsten Schoenfeld  <kaffeetisch@gmx.de>
2576
2577         * include/ffi.h.in (ffi_type_ulong, ffi_type_slong): Define correctly
2578         for 32-bit architectures.
2579         * testsuite/libffi.call/return_ul.c: New test case.
2580
2581 2006-07-19  David Daney  <ddaney@avtrex.com>
2582
2583         * testsuite/libffi.call/closure_fn6.c: Remove xfail for mips,
2584         xfail remains for mips64.
2585
2586 2006-05-23  Carlos O'Donell  <carlos@codesourcery.com>
2587
2588         * Makefile.am: Add install-html target. Add install-html to .PHONY
2589         * Makefile.in: Regenerate.
2590         * aclocal.m4: Regenerate.
2591         * include/Makefile.in: Regenerate.
2592         * testsuite/Makefile.in: Regenerate.
2593
2594 2006-05-18  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2595
2596         * pa/ffi.c (ffi_prep_args_pa32): Load floating point arguments from
2597         stack slot.
2598
2599 2006-04-22  Andreas Tobler  <a.tobler@schweiz.ch>
2600
2601         * README: Remove notice about 'Crazy Comments'.
2602         * src/debug.c: Remove lint directives. Cleanup white spaces.
2603         * src/java_raw_api.c: Likewise.
2604         * src/prep_cif.c: Likewise.
2605         * src/raw_api.c: Likewise.
2606         * src/ffitest.c: Delete. No longer needed, all test cases migrated
2607         to the testsuite.
2608         * src/arm/ffi.c: Remove lint directives.
2609         * src/m32r/ffi.c: Likewise.
2610         * src/pa/ffi.c: Likewise.
2611         * src/powerpc/ffi.c: Likewise.
2612         * src/powerpc/ffi_darwin.c: Likewise.
2613         * src/sh/ffi.c: Likewise.
2614         * src/sh64/ffi.c: Likewise.
2615         * src/x86/ffi.c: Likewise.
2616         * testsuite/libffi.call/float2.c: Likewise.
2617         * testsuite/libffi.call/promotion.c: Likewise.
2618         * testsuite/libffi.call/struct1.c: Likewise.
2619
2620 2006-04-13  Andreas Tobler  <a.tobler@schweiz.ch>
2621
2622         * src/pa/hpux32.S: Correct unwind offset calculation for
2623         ffi_closure_pa32.
2624         * src/pa/linux.S: Likewise.
2625
2626 2006-04-12  James E Wilson  <wilson@specifix.com>
2627
2628         PR libgcj/26483
2629         * src/ia64/ffi.c (stf_spill, ldf_fill): Rewrite as macros.
2630         (hfa_type_load): Call stf_spill.
2631         (hfa_type_store): Call ldf_fill.
2632         (ffi_call): Adjust calls to above routines.  Add local temps for
2633         macro result.
2634
2635 2006-04-10  Matthias Klose  <doko@debian.org>
2636
2637         * testsuite/lib/libffi-dg.exp (libffi-init): Recognize multilib
2638         directory names containing underscores.
2639
2640 2006-04-07  James E Wilson  <wilson@specifix.com>
2641
2642         * testsuite/libffi.call/float4.c: New testcase.
2643
2644 2006-04-05  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
2645             Andreas Tobler  <a.tobler@schweiz.ch>
2646
2647         * Makefile.am: Add PA_HPUX port.
2648         * Makefile.in: Regenerate.
2649         * include/Makefile.in: Likewise.
2650         * testsuite/Makefile.in: Likewise.
2651         * configure.ac: Add PA_HPUX rules.
2652         * configure: Regenerate.
2653         * src/pa/ffitarget.h: Rename linux target to PA_LINUX.
2654         Add PA_HPUX and PA64_HPUX.
2655         Rename FFI_LINUX ABI to FFI_PA32 ABI.
2656         (FFI_TRAMPOLINE_SIZE): Define for 32-bit HP-UX targets.
2657         (FFI_TYPE_SMALL_STRUCT2): Define.
2658         (FFI_TYPE_SMALL_STRUCT4): Likewise.
2659         (FFI_TYPE_SMALL_STRUCT8): Likewise.
2660         (FFI_TYPE_SMALL_STRUCT3): Redefine.
2661         (FFI_TYPE_SMALL_STRUCT5): Likewise.
2662         (FFI_TYPE_SMALL_STRUCT6): Likewise.
2663         (FFI_TYPE_SMALL_STRUCT7): Likewise.
2664         * src/pa/ffi.c (ROUND_DOWN): Delete.
2665         (fldw, fstw, fldd, fstd): Use '__asm__'.
2666         (ffi_struct_type): Add support for FFI_TYPE_SMALL_STRUCT2,
2667         FFI_TYPE_SMALL_STRUCT4 and FFI_TYPE_SMALL_STRUCT8.
2668         (ffi_prep_args_LINUX): Rename to ffi_prep_args_pa32. Update comment.
2669         Simplify incrementing of stack slot variable. Change type of local
2670         'n' to unsigned int.
2671         (ffi_size_stack_LINUX): Rename to ffi_size_stack_pa32. Handle long
2672         double on PA_HPUX.
2673         (ffi_prep_cif_machdep): Likewise.
2674         (ffi_call): Likewise.
2675         (ffi_closure_inner_LINUX): Rename to ffi_closure_inner_pa32. Change
2676         return type to ffi_status. Simplify incrementing of stack slot
2677         variable. Only copy floating point argument registers when PA_LINUX
2678         is true. Reformat debug statement.
2679         Add support for FFI_TYPE_SMALL_STRUCT2, FFI_TYPE_SMALL_STRUCT4 and
2680         FFI_TYPE_SMALL_STRUCT8.
2681         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Add 'extern' to
2682         declaration.
2683         (ffi_prep_closure): Make linux trampoline conditional on PA_LINUX.
2684         Add nops to cache flush.  Add trampoline for PA_HPUX.
2685         * src/pa/hpux32.S: New file.
2686         * src/pa/linux.S (ffi_call_LINUX): Rename to ffi_call_pa32. Rename
2687         ffi_prep_args_LINUX to ffi_prep_args_pa32.
2688         Localize labels. Add support for 2, 4 and 8-byte small structs. Handle
2689         unaligned destinations in 3, 5, 6 and 7-byte small structs. Order
2690         argument type checks so that common argument types appear first.
2691         (ffi_closure_LINUX): Rename to ffi_closure_pa32. Rename
2692         ffi_closure_inner_LINUX to ffi_closure_inner_pa32.
2693
2694 2006-03-24  Alan Modra  <amodra@bigpond.net.au>
2695
2696         * src/powerpc/ffitarget.h (enum ffi_abi): Add FFI_LINUX.  Default
2697         for 32-bit using IBM extended double format.  Fix FFI_LAST_ABI.
2698         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Handle linux variant of
2699         FFI_TYPE_LONGDOUBLE.
2700         (ffi_prep_args64): Assert using IBM extended double.
2701         (ffi_prep_cif_machdep): Don't munge FFI_TYPE_LONGDOUBLE type.
2702         Handle FFI_LINUX FFI_TYPE_LONGDOUBLE return and args.
2703         (ffi_call): Handle FFI_LINUX.
2704         (ffi_closure_helper_SYSV): Non FFI_LINUX long double return needs
2705         gpr3 return pointer as for struct return.  Handle FFI_LINUX
2706         FFI_TYPE_LONGDOUBLE return and args.  Don't increment "nf"
2707         unnecessarily.
2708         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Load both f1 and f2
2709         for FFI_TYPE_LONGDOUBLE.  Move epilogue insns into case table.
2710         Don't use r6 as pointer to results, instead use sp offset.  Don't
2711         make a special call to load lr with case table address, instead
2712         use offset from previous call.
2713         * src/powerpc/sysv.S (ffi_call_SYSV): Save long double return.
2714         * src/powerpc/linux64.S (ffi_call_LINUX64): Simplify long double
2715         return.
2716
2717 2006-03-15  Kaz Kojima  <kkojima@gcc.gnu.org>
2718
2719         * src/sh64/ffi.c (ffi_prep_cif_machdep): Handle float arguments
2720         passed with FP registers correctly.
2721         (ffi_closure_helper_SYSV): Likewise.
2722         * src/sh64/sysv.S: Likewise.
2723
2724 2006-03-01  Andreas Tobler  <a.tobler@schweiz.ch>
2725
2726         * testsuite/libffi.special/unwindtest.cc (closure_test_fn): Mark cif,
2727         args and userdata unused.
2728         (closure_test_fn1): Mark cif and userdata unused.
2729         (main): Remove unused res.
2730
2731 2006-02-28  Andreas Tobler  <a.tobler@schweiz.ch>
2732
2733         * testsuite/libffi.call/call.exp: Adjust FSF address. Add test runs for
2734         -O2, -O3, -Os and the warning flags -W -Wall.
2735         * testsuite/libffi.special/special.exp: Likewise.
2736         * testsuite/libffi.call/ffitest.h: Add an __UNUSED__ macro to mark
2737         unused parameter unused for gcc or else do nothing.
2738         * testsuite/libffi.special/ffitestcxx.h: Likewise.
2739         * testsuite/libffi.call/cls_12byte.c (cls_struct_12byte_gn): Mark cif
2740         and userdata unused.
2741         * testsuite/libffi.call/cls_16byte.c (cls_struct_16byte_gn): Likewise.
2742         * testsuite/libffi.call/cls_18byte.c (cls_struct_18byte_gn): Likewise.
2743         * testsuite/libffi.call/cls_19byte.c (cls_struct_19byte_gn): Likewise.
2744         * testsuite/libffi.call/cls_1_1byte.c (cls_struct_1_1byte_gn): Likewise.
2745         * testsuite/libffi.call/cls_20byte.c (cls_struct_20byte_gn): Likewise.
2746         * testsuite/libffi.call/cls_20byte1.c (cls_struct_20byte_gn): Likewise.
2747         * testsuite/libffi.call/cls_24byte.c (cls_struct_24byte_gn): Likewise.
2748         * testsuite/libffi.call/cls_2byte.c (cls_struct_2byte_gn): Likewise.
2749         * testsuite/libffi.call/cls_3_1byte.c (cls_struct_3_1byte_gn): Likewise.
2750         * testsuite/libffi.call/cls_3byte1.c (cls_struct_3byte_gn): Likewise.
2751         * testsuite/libffi.call/cls_3byte2.c (cls_struct_3byte_gn1): Likewise.
2752         * testsuite/libffi.call/cls_4_1byte.c (cls_struct_4_1byte_gn): Likewise.
2753         * testsuite/libffi.call/cls_4byte.c (cls_struct_4byte_gn): Likewise.
2754         * testsuite/libffi.call/cls_5_1_byte.c (cls_struct_5byte_gn): Likewise.
2755         * testsuite/libffi.call/cls_5byte.c (cls_struct_5byte_gn): Likewise.
2756         * testsuite/libffi.call/cls_64byte.c (cls_struct_64byte_gn): Likewise.
2757         * testsuite/libffi.call/cls_6_1_byte.c (cls_struct_6byte_gn): Likewise.
2758         * testsuite/libffi.call/cls_6byte.c (cls_struct_6byte_gn): Likewise.
2759         * testsuite/libffi.call/cls_7_1_byte.c (cls_struct_7byte_gn): Likewise.
2760         * testsuite/libffi.call/cls_7byte.c (cls_struct_7byte_gn): Likewise.
2761         * testsuite/libffi.call/cls_8byte.c (cls_struct_8byte_gn): Likewise.
2762         * testsuite/libffi.call/cls_9byte1.c (cls_struct_9byte_gn): Likewise.
2763         * testsuite/libffi.call/cls_9byte2.c (cls_struct_9byte_gn): Likewise.
2764         * testsuite/libffi.call/cls_align_double.c (cls_struct_align_gn):
2765         Likewise.
2766         * testsuite/libffi.call/cls_align_float.c (cls_struct_align_gn):
2767         Likewise.
2768         * testsuite/libffi.call/cls_align_longdouble.c (cls_struct_align_gn):
2769         Likewise.
2770         * testsuite/libffi.call/cls_align_pointer.c (cls_struct_align_fn): Cast
2771         void* to avoid compiler warning.
2772         (main): Likewise.
2773         (cls_struct_align_gn): Mark cif and userdata unused.
2774         * testsuite/libffi.call/cls_align_sint16.c (cls_struct_align_gn):
2775         Likewise.
2776         * testsuite/libffi.call/cls_align_sint32.c (cls_struct_align_gn):
2777         Likewise.
2778         * testsuite/libffi.call/cls_align_sint64.c (cls_struct_align_gn):
2779         Likewise.
2780         * testsuite/libffi.call/cls_align_uint16.c (cls_struct_align_gn):
2781         Likewise.
2782         * testsuite/libffi.call/cls_align_uint32.c (cls_struct_align_gn):
2783         Likewise.
2784         * testsuite/libffi.call/cls_double.c (cls_ret_double_fn): Likewise.
2785         * testsuite/libffi.call/cls_float.c (cls_ret_float_fn): Likewise.
2786         * testsuite/libffi.call/cls_multi_schar.c (test_func_gn): Mark cif and
2787         data unused.
2788         (main): Cast res_call to silence gcc.
2789         * testsuite/libffi.call/cls_multi_sshort.c (test_func_gn): Mark cif and
2790         data unused.
2791         (main): Cast res_call to silence gcc.
2792         * testsuite/libffi.call/cls_multi_sshortchar.c (test_func_gn): Mark cif
2793         and data unused.
2794         (main): Cast res_call to silence gcc.
2795         * testsuite/libffi.call/cls_multi_uchar.c (test_func_gn): Mark cif and
2796         data unused.
2797         (main): Cast res_call to silence gcc.
2798         * testsuite/libffi.call/cls_multi_ushort.c (test_func_gn): Mark cif and
2799         data unused.
2800         (main): Cast res_call to silence gcc.
2801         * testsuite/libffi.call/cls_multi_ushortchar.c (test_func_gn): Mark cif
2802         and data unused.
2803         (main): Cast res_call to silence gcc.
2804         * testsuite/libffi.call/cls_schar.c (cls_ret_schar_fn): Mark cif and
2805         userdata unused.
2806         (cls_ret_schar_fn): Cast printf parameter to silence gcc.
2807         * testsuite/libffi.call/cls_sint.c (cls_ret_sint_fn): Mark cif and
2808         userdata unused.
2809         (cls_ret_sint_fn): Cast printf parameter to silence gcc.
2810         * testsuite/libffi.call/cls_sshort.c (cls_ret_sshort_fn): Mark cif and
2811         userdata unused.
2812         (cls_ret_sshort_fn): Cast printf parameter to silence gcc.
2813         * testsuite/libffi.call/cls_uchar.c (cls_ret_uchar_fn):  Mark cif and
2814         userdata unused.
2815         (cls_ret_uchar_fn): Cast printf parameter to silence gcc.
2816         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Mark cif and
2817         userdata unused.
2818         (cls_ret_uint_fn): Cast printf parameter to silence gcc.
2819         * testsuite/libffi.call/cls_ulonglong.c (cls_ret_ulonglong_fn): Mark cif
2820         and userdata unused.
2821         * testsuite/libffi.call/cls_ushort.c (cls_ret_ushort_fn): Mark cif and
2822         userdata unused.
2823         (cls_ret_ushort_fn): Cast printf parameter to silence gcc.
2824         * testsuite/libffi.call/float.c (floating): Remove unused parameter e.
2825         * testsuite/libffi.call/float1.c (main): Remove unused variable i.
2826         Cleanup white spaces.
2827         * testsuite/libffi.call/negint.c (checking): Remove unused variable i.
2828         * testsuite/libffi.call/nested_struct.c (cls_struct_combined_gn): Mark
2829         cif and userdata unused.
2830         * testsuite/libffi.call/nested_struct1.c (cls_struct_combined_gn):
2831         Likewise.
2832         * testsuite/libffi.call/nested_struct10.c (B_gn): Likewise.
2833         * testsuite/libffi.call/nested_struct2.c (B_fn): Adjust printf
2834         formatters to silence gcc.
2835         (B_gn): Mark cif and userdata unused.
2836         * testsuite/libffi.call/nested_struct3.c (B_gn): Mark cif and userdata
2837         unused.
2838         * testsuite/libffi.call/nested_struct4.c: Mention related PR.
2839         (B_gn): Mark cif and userdata unused.
2840         * testsuite/libffi.call/nested_struct5.c (B_gn): Mark cif and userdata
2841         unused.
2842         * testsuite/libffi.call/nested_struct6.c: Mention related PR.
2843         (B_gn): Mark cif and userdata unused.
2844         * testsuite/libffi.call/nested_struct7.c (B_gn): Mark cif and userdata
2845         unused.
2846         * testsuite/libffi.call/nested_struct8.c (B_gn): Likewise.
2847         * testsuite/libffi.call/nested_struct9.c (B_gn): Likewise.
2848         * testsuite/libffi.call/problem1.c (stub): Likewise.
2849         * testsuite/libffi.call/pyobjc-tc.c (main): Cast the result to silence
2850         gcc.
2851         * testsuite/libffi.call/return_fl2.c (return_fl): Add the note mentioned
2852         in the last commit for this test case in the test case itself.
2853         * testsuite/libffi.call/closure_fn0.c (closure_test_fn0): Mark cif as
2854         unused.
2855         * testsuite/libffi.call/closure_fn1.c (closure_test_fn1): Likewise.
2856         * testsuite/libffi.call/closure_fn2.c (closure_test_fn2): Likewise.
2857         * testsuite/libffi.call/closure_fn3.c (closure_test_fn3): Likewise.
2858         * testsuite/libffi.call/closure_fn4.c (closure_test_fn0): Likewise.
2859         * testsuite/libffi.call/closure_fn5.c (closure_test_fn5): Likewise.
2860         * testsuite/libffi.call/closure_fn6.c (closure_test_fn0): Likewise.
2861
2862 2006-02-22  Kaz Kojima  <kkojima@gcc.gnu.org>
2863
2864         * src/sh/sysv.S: Fix register numbers in the FDE for
2865         ffi_closure_SYSV.
2866
2867 2006-02-20  Andreas Tobler  <a.tobler@schweiz.ch>
2868
2869         * testsuite/libffi.call/return_fl2.c (return_fl): Remove static
2870         declaration to avoid a false negative on ix86. See PR323.
2871
2872 2006-02-18  Kaz Kojima  <kkojima@gcc.gnu.org>
2873
2874         * src/sh/ffi.c (ffi_closure_helper_SYSV): Remove unused variable
2875         and cast integer to void * if needed.  Update the pointer to
2876         the FP register saved area correctly.
2877
2878 2006-02-17  Andreas Tobler  <a.tobler@schweiz.ch>
2879
2880         * testsuite/libffi.call/nested_struct6.c: XFAIL this test until PR25630
2881         is fixed.
2882         * testsuite/libffi.call/nested_struct4.c: Likewise.
2883
2884 2006-02-16  Andreas Tobler  <a.tobler@schweiz.ch>
2885
2886         * testsuite/libffi.call/return_dbl.c: New test case.
2887         * testsuite/libffi.call/return_dbl1.c: Likewise.
2888         * testsuite/libffi.call/return_dbl2.c: Likewise.
2889         * testsuite/libffi.call/return_fl.c: Likewise.
2890         * testsuite/libffi.call/return_fl1.c: Likewise.
2891         * testsuite/libffi.call/return_fl2.c: Likewise.
2892         * testsuite/libffi.call/return_fl3.c: Likewise.
2893         * testsuite/libffi.call/closure_fn6.c: Likewise.
2894
2895         * testsuite/libffi.call/nested_struct2.c: Remove ffi_type_mylong
2896         definition.
2897         * testsuite/libffi.call/ffitest.h: Add ffi_type_mylong definition
2898         here to be used by other test cases too.
2899
2900         * testsuite/libffi.call/nested_struct10.c: New test case.
2901         * testsuite/libffi.call/nested_struct9.c: Likewise.
2902         * testsuite/libffi.call/nested_struct8.c: Likewise.
2903         * testsuite/libffi.call/nested_struct7.c: Likewise.
2904         * testsuite/libffi.call/nested_struct6.c: Likewise.
2905         * testsuite/libffi.call/nested_struct5.c: Likewise.
2906         * testsuite/libffi.call/nested_struct4.c: Likewise.
2907
2908 2006-01-21  Andreas Tobler  <a.tobler@schweiz.ch>
2909
2910         * configure.ac: Enable libffi for sparc64-*-freebsd*.
2911         * configure: Rebuilt.
2912
2913 2006-01-18  Jakub Jelinek  <jakub@redhat.com>
2914
2915         * src/powerpc/sysv.S (smst_two_register): Don't call __ashldi3,
2916         instead do the shifting inline.
2917         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't compute %r5
2918         shift count unconditionally.  Simplify load sequences for 1, 2, 3, 4
2919         and 8 byte structs, for the remaining struct sizes don't call
2920         __lshrdi3, instead do the shifting inline.
2921
2922 2005-12-07  Thiemo Seufer  <ths@networkno.de>
2923
2924         * src/mips/ffitarget.h: Remove obsolete sgidefs.h include. Add
2925         missing parentheses.
2926         * src/mips/o32.S (ffi_call_O32): Code formatting. Define
2927         and use A3_OFF, FP_OFF, RA_OFF. Micro-optimizations.
2928         (ffi_closure_O32): Likewise, but with newly defined A3_OFF2,
2929         A2_OFF2, A1_OFF2, A0_OFF2, RA_OFF2, FP_OFF2, S0_OFF2, GP_OFF2,
2930         V1_OFF2, V0_OFF2, FA_1_1_OFF2, FA_1_0_OFF2, FA_0_1_OFF2,
2931         FA_0_0_OFF2.
2932         * src/mips/ffi.c (ffi_prep_args): Code formatting. Fix
2933         endianness bugs.
2934         (ffi_prep_closure): Improve trampoline instruction scheduling.
2935         (ffi_closure_mips_inner_O32): Fix endianness bugs.
2936
2937 2005-12-03  Alan Modra  <amodra@bigpond.net.au>
2938
2939         * src/powerpc/ffi.c: Formatting.
2940         (ffi_prep_args_SYSV): Avoid possible aliasing problems by using unions.
2941         (ffi_prep_args64): Likewise.
2942
2943 2005-09-30  Geoffrey Keating  <geoffk@apple.com>
2944
2945         * testsuite/lib/libffi-dg.exp (libffi_target_compile): For
2946         darwin, use -shared-libgcc not -lgcc_s, and explain why.
2947
2948 2005-09-26  Tom Tromey  <tromey@redhat.com>
2949
2950         * testsuite/libffi.call/float1.c (value_type): New typedef.
2951         (CANARY): New define.
2952         (main): Check for result buffer overflow.
2953         * src/powerpc/linux64.S: Handle linux64 long double returns.
2954         * src/powerpc/ffi.c (FLAG_RETURNS_128BITS): New constant.
2955         (ffi_prep_cif_machdep): Handle linux64 long double returns.
2956
2957 2005-08-25  Alan Modra  <amodra@bigpond.net.au>
2958
2959         PR target/23404
2960         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Correct placement of stack
2961         homed fp args.
2962         (ffi_status ffi_prep_cif_machdep): Correct stack sizing for same.
2963
2964 2005-08-11  Jakub Jelinek  <jakub@redhat.com>
2965
2966         * configure.ac (HAVE_HIDDEN_VISIBILITY_ATTRIBUTE): New test.
2967         (AH_BOTTOM): Add FFI_HIDDEN definition.
2968         * configure: Rebuilt.
2969         * fficonfig.h.in: Rebuilt.
2970         * src/powerpc/ffi.c (hidden): Remove.
2971         (ffi_closure_LINUX64, ffi_prep_args64, ffi_call_LINUX64,
2972         ffi_closure_helper_LINUX64): Use FFI_HIDDEN instead of hidden.
2973         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64,
2974         .ffi_closure_LINUX64): Use FFI_HIDDEN instead of .hidden.
2975         * src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV): Remove,
2976         add FFI_HIDDEN to its prototype.
2977         (ffi_closure_SYSV_inner): New.
2978         * src/x86/sysv.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2979         * src/x86/win32.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
2980
2981 2005-08-10  Alfred M. Szmidt  <ams@gnu.org>
2982
2983         PR libffi/21819:
2984         * configure: Rebuilt.
2985         * configure.ac: Handle i*86-*-gnu*.
2986
2987 2005-08-09  Jakub Jelinek  <jakub@redhat.com>
2988
2989         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Use
2990         DW_CFA_offset_extended_sf rather than
2991         DW_CFA_GNU_negative_offset_extended.
2992         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise.
2993
2994 2005-07-22  SUGIOKA Toshinobu  <sugioka@itonet.co.jp>
2995
2996         * src/sh/sysv.S (ffi_call_SYSV): Stop argument popping correctly
2997         on sh3.
2998         (ffi_closure_SYSV): Change the stack layout for sh3 struct argument.
2999         * src/sh/ffi.c (ffi_prep_args): Fix sh3 argument copy, when it is
3000         partially on register.
3001         (ffi_closure_helper_SYSV): Likewise.
3002         (ffi_prep_cif_machdep): Don't set too many cif->flags.
3003
3004 2005-07-20  Kaz Kojima  <kkojima@gcc.gnu.org>
3005
3006         * src/sh/ffi.c (ffi_call): Handle small structures correctly.
3007         Remove empty line.
3008         * src/sh64/ffi.c (simple_type): Remove.
3009         (return_type): Handle small structures correctly.
3010         (ffi_prep_args): Likewise.
3011         (ffi_call): Likewise.
3012         (ffi_closure_helper_SYSV): Likewise.
3013         * src/sh64/sysv.S (ffi_call_SYSV): Handle 1, 2 and 4-byte return.
3014         Emit position independent code if PIC and remove wrong datalabel
3015         prefixes from EH data.
3016
3017 2005-07-19  Andreas Tobler  <a.tobler@schweiz.ch>
3018
3019         * Makefile.am (nodist_libffi_la_SOURCES): Add POWERPC_FREEBSD.
3020         * Makefile.in: Regenerate.
3021         * include/Makefile.in: Likewise.
3022         * testsuite/Makefile.in: Likewise.
3023         * configure.ac: Add POWERPC_FREEBSD rules.
3024         * configure: Regenerate.
3025         * src/powerpc/ffitarget.h: Add POWERPC_FREEBSD rules.
3026         (FFI_SYSV_TYPE_SMALL_STRUCT): Define.
3027         * src/powerpc/ffi.c: Add flags to handle small structure returns
3028         in ffi_call_SYSV.
3029         (ffi_prep_cif_machdep): Handle small structures for SYSV 4 ABI.
3030         Aka FFI_SYSV.
3031         (ffi_closure_helper_SYSV): Likewise.
3032         * src/powerpc/ppc_closure.S: Add return types for small structures.
3033         * src/powerpc/sysv.S: Add bits to handle small structures for
3034         final SYSV 4 ABI.
3035
3036 2005-07-10  Andreas Tobler  <a.tobler@schweiz.ch>
3037
3038         * testsuite/libffi.call/cls_5_1_byte.c: New test file.
3039         * testsuite/libffi.call/cls_6_1_byte.c: Likewise.
3040         * testsuite/libffi.call/cls_7_1_byte.c: Likewise.
3041
3042 2005-07-05  Randolph Chung  <tausq@debian.org>
3043
3044         * src/pa/ffi.c (ffi_struct_type): Rename FFI_TYPE_SMALL_STRUCT1
3045         as FFI_TYPE_SMALL_STRUCT3.  Break out handling for 5-7 byte
3046         structures.  Kill compilation warnings.
3047         (ffi_closure_inner_LINUX): Print return values as hex in debug
3048         message.  Rename FFI_TYPE_SMALL_STRUCT1 as FFI_TYPE_SMALL_STRUCT3.
3049         Properly handle 5-7 byte structure returns.
3050         * src/pa/ffitarget.h (FFI_TYPE_SMALL_STRUCT1)
3051         (FFI_TYPE_SMALL_STRUCT2): Remove.
3052         (FFI_TYPE_SMALL_STRUCT3, FFI_TYPE_SMALL_STRUCT5)
3053         (FFI_TYPE_SMALL_STRUCT6, FFI_TYPE_SMALL_STRUCT7): Define.
3054         * src/pa/linux.S: Mark source file as using PA1.1 assembly.
3055         (checksmst1, checksmst2): Remove.
3056         (checksmst3): Optimize handling of 3-byte struct returns.
3057         (checksmst567): Properly handle 5-7 byte struct returns.
3058
3059 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3060
3061         PR libgcj/21943
3062         * src/mips/n32.S: Enforce PIC code.
3063         * src/mips/o32.S: Likewise.
3064
3065 2005-06-15  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3066
3067         * configure.ac: Treat i*86-*-solaris2.10 and up as X86_64.
3068         * configure: Regenerate.
3069
3070 2005-06-01  Alan Modra  <amodra@bigpond.net.au>
3071
3072         * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't use JUMPTARGET
3073         to call ffi_closure_helper_SYSV.  Append @local instead.
3074         * src/powerpc/sysv.S (ffi_call_SYSV): Likewise for ffi_prep_args_SYSV.
3075
3076 2005-05-17  Kelley Cook  <kcook@gcc.gnu.org>
3077
3078         * configure.ac: Use AC_C_BIGENDIAN instead of AC_C_BIGENDIAN_CROSS.
3079         Use AC_CHECK_SIZEOF instead of AC_COMPILE_CHECK_SIZEOF.
3080         * Makefile.am (ACLOCAL_AMFLAGS): Remove -I ../config.
3081         * aclocal.m4, configure, fficonfig.h.in, Makefile.in,
3082         include/Makefile.in, testsuite/Makefile.in: Regenerate.
3083
3084 2005-05-09  Mike Stump  <mrs@apple.com>
3085
3086         * configure: Regenerate.
3087
3088 2005-05-08  Richard Henderson  <rth@redhat.com>
3089
3090         PR libffi/21285
3091         * src/alpha/osf.S: Update unwind into to match code.
3092
3093 2005-05-04  Andreas Degert <ad@papyrus-gmbh.de>
3094             Richard Henderson  <rth@redhat.com>
3095
3096         * src/x86/ffi64.c (ffi_prep_cif_machdep): Save sse-used flag in
3097         bit 11 of flags.
3098         (ffi_call): Mask return type field.  Pass ssecount to ffi_call_unix64.
3099         (ffi_prep_closure): Set carry bit if sse-used flag set.
3100         * src/x86/unix64.S (ffi_call_unix64): Add ssecount argument.
3101         Only load sse registers if ssecount non-zero.
3102         (ffi_closure_unix64): Only save sse registers if carry set on entry.
3103
3104 2005-04-29  Ralf Corsepius  <ralf.corsepius@rtems.org>
3105
3106         * configure.ac: Add i*86-*-rtems*, sparc*-*-rtems*,
3107         powerpc-*rtems*, arm*-*-rtems*, sh-*-rtems*.
3108         * configure: Regenerate.
3109
3110 2005-04-20  Hans-Peter Nilsson  <hp@axis.com>
3111
3112         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): In regsub use,
3113         have Tcl8.3-compatible intermediate variable.
3114
3115 2005-04-18  Simon Posnjak <simon.posnjak@siol.net>
3116             Hans-Peter Nilsson  <hp@axis.com>
3117
3118         * Makefile.am: Add CRIS support.
3119         * configure.ac: Likewise.
3120         * Makefile.in, configure, testsuite/Makefile.in,
3121         include/Makefile.in: Regenerate.
3122         * src/cris: New directory.
3123         * src/cris/ffi.c, src/cris/sysv.S, src/cris/ffitarget.h: New files.
3124         * src/prep_cif.c (ffi_prep_cif): Wrap in #ifndef __CRIS__.
3125
3126         * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): Replace \n with
3127         \r?\n in output tests.
3128
3129 2005-04-12  Mike Stump  <mrs@apple.com>
3130
3131         * configure: Regenerate.
3132
3133 2005-03-30  Hans Boehm  <Hans.Boehm@hp.com>
3134
3135         * src/ia64/ffitarget.h (ffi_arg): Use long long instead of DI.
3136
3137 2005-03-30  Steve Ellcey  <sje@cup.hp.com>
3138
3139         * src/ia64/ffitarget.h (ffi_arg) ADD DI attribute.
3140         (ffi_sarg) Ditto.
3141         * src/ia64/unix.S (ffi_closure_unix): Extend gp
3142         to 64 bits in ILP32 mode.
3143         Load 64 bits even for short data.
3144
3145 2005-03-23  Mike Stump  <mrs@apple.com>
3146
3147         * src/powerpc/darwin.S: Update for -m64 multilib.
3148         * src/powerpc/darwin_closure.S: Likewise.
3149
3150 2005-03-21  Zack Weinberg  <zack@codesourcery.com>
3151
3152         * configure.ac: Do not invoke TL_AC_GCC_VERSION.
3153         Do not set tool_include_dir.
3154         * aclocal.m4, configure, Makefile.in, testsuite/Makefile.in:
3155         Regenerate.
3156         * include/Makefile.am: Set gcc_version and toollibffidir.
3157         * include/Makefile.in: Regenerate.
3158
3159 2005-02-22  Andrew Haley  <aph@redhat.com>
3160
3161         * src/powerpc/ffi.c (ffi_prep_cif_machdep): Bump alignment to
3162         odd-numbered register pairs for 64-bit integer types.
3163
3164 2005-02-23  Andreas Tobler  <a.tobler@schweiz.ch>
3165
3166         PR libffi/20104
3167         * testsuite/libffi.call/return_ll1.c: New test case.
3168
3169 2005-02-11  Janis Johnson  <janis187@us.ibm.com>
3170
3171         * testsuite/libffi.call/cls_align_longdouble.c: Remove dg-options.
3172         * testsuite/libffi.call/float.c: Ditto.
3173         * testsuite/libffi.call/float2.c: Ditto.
3174         * testsuite/libffi.call/float3.c: Ditto.
3175
3176 2005-02-08  Andreas Tobler  <a.tobler@schweiz.ch>
3177
3178         * src/frv/ffitarget.h: Remove PPC stuff which does not belong to frv.
3179
3180 2005-01-12  Eric Botcazou  <ebotcazou@libertysurf.fr>
3181
3182         * testsuite/libffi.special/special.exp (cxx_options): Add
3183         -shared-libgcc.
3184
3185 2004-12-31  Richard Henderson  <rth@redhat.com>
3186
3187         * src/types.c (FFI_AGGREGATE_TYPEDEF): Remove.
3188         (FFI_TYPEDEF): Rename from FFI_INTEGRAL_TYPEDEF.  Replace size and
3189         offset parameters with a type parameter; deduce size and structure
3190         alignment.  Update all users.
3191
3192 2004-12-31  Richard Henderson  <rth@redhat.com>
3193
3194         * src/types.c (FFI_TYPE_POINTER): Define with sizeof.
3195         (FFI_TYPE_LONGDOUBLE): Fix for ia64.
3196         * src/ia64/ffitarget.h (struct ffi_ia64_trampoline_struct): Move
3197         into ffi_prep_closure.
3198         * src/ia64/ia64_flags.h, src/ia64/ffi.c, src/ia64/unix.S: Rewrite
3199         from scratch.
3200
3201 2004-12-27  Richard Henderson  <rth@redhat.com>
3202
3203         * src/x86/unix64.S: Fix typo in unwind info.
3204
3205 2004-12-25  Richard Henderson  <rth@redhat.com>
3206
3207         * src/x86/ffi64.c (struct register_args): Rename from stackLayout.
3208         (enum x86_64_reg_class): Add X86_64_COMPLEX_X87_CLASS.
3209         (merge_classes): Check for it.
3210         (SSE_CLASS_P): New.
3211         (classify_argument): Pass byte_offset by value; perform all updates
3212         inside struct case.
3213         (examine_argument): Add classes argument; handle
3214         X86_64_COMPLEX_X87_CLASS.
3215         (ffi_prep_args): Merge into ...
3216         (ffi_call): ... here.  Share stack frame with ffi_call_unix64.
3217         (ffi_prep_cif_machdep): Setup cif->flags for proper structure return.
3218         (ffi_fill_return_value): Remove.
3219         (ffi_prep_closure): Remove dead assert.
3220         (ffi_closure_unix64_inner): Rename from ffi_closure_UNIX64_inner.
3221         Rewrite to use struct register_args instead of va_list.  Create
3222         flags for handling structure returns.
3223         * src/x86/unix64.S: Remove dead strings.
3224         (ffi_call_unix64): Rename from ffi_call_UNIX64.  Rewrite to share
3225         stack frame with ffi_call.  Handle structure returns properly.
3226         (float2sse, floatfloat2sse, double2sse): Remove.
3227         (sse2float, sse2double, sse2floatfloat): Remove.
3228         (ffi_closure_unix64): Rename from ffi_closure_UNIX64.  Rewrite
3229         to handle structure returns properly.
3230
3231 2004-12-08  David Edelsohn  <edelsohn@gnu.org>
3232
3233         * Makefile.am (AM_MAKEFLAGS): Remove duplicate LIBCFLAGS and
3234         PICFLAG.
3235         * Makefile.in: Regenerated.
3236
3237 2004-12-02  Richard Sandiford  <rsandifo@redhat.com>
3238
3239         * configure.ac: Use TL_AC_GCC_VERSION to set gcc_version.
3240         * configure, aclocal.m4, Makefile.in: Regenerate.
3241         * include/Makefile.in, testsuite/Makefile.in: Regenerate.
3242
3243 2004-11-29  Kelley Cook  <kcook@gcc.gnu.org>
3244
3245         * configure: Regenerate for libtool change.
3246
3247 2004-11-25  Kelley Cook  <kcook@gcc.gnu.org>
3248
3249         * configure: Regenerate for libtool reversion.
3250
3251 2004-11-24  Kelley Cook  <kcook@gcc.gnu.org>
3252
3253         * configure: Regenerate for libtool change.
3254
3255 2004-11-23  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
3256
3257         * testsuite/lib/libffi-dg.exp: Use new procs in target-libpath.exp.
3258
3259 2004-11-23  Richard Sandiford  <rsandifo@redhat.com>
3260
3261         * src/mips/o32.S (ffi_call_O32, ffi_closure_O32): Use jalr instead
3262         of jal.  Use an absolute encoding for the frame information.
3263
3264 2004-11-23  Kelley Cook  <kcook@gcc.gnu.org>
3265
3266         * Makefile.am: Remove no-dependencies.  Add ACLOCAL_AMFLAGS.
3267         * acinclude.m4: Delete logic for sincludes.
3268         * aclocal.m4, Makefile.in, configure: Regenerate.
3269         * include/Makefile: Likewise.
3270         * testsuite/Makefile: Likewise.
3271
3272 2004-11-22  Eric Botcazou  <ebotcazou@libertysurf.fr>
3273
3274         * src/sparc/ffi.c (ffi_prep_closure): Align doubles and 64-bit integers
3275         on a 8-byte boundary.
3276         * src/sparc/v8.S (ffi_closure_v8): Reserve frame space for arguments.
3277
3278 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
3279
3280         * src/arm/ffi.c (ffi_prep_cif_machdep): Handle functions that return
3281         long long values.  Round stack allocation to a multiple of 8 bytes
3282         for ATPCS compatibility.
3283         * src/arm/sysv.S (ffi_call_SYSV): Rework to avoid use of APCS register
3284         names.  Handle returning long long types.  Add Thumb and interworking
3285         support.  Improve soft-float code.
3286
3287 2004-10-27  Richard Earnshaw  <rearnsha@arm.com>
3288
3289         * testsuite/lib/libffi-db.exp (load_gcc_lib): New function.
3290         (libffi_exit): New function.
3291         (libffi_init): Build the testglue wrapper if needed.
3292
3293 2004-10-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
3294
3295         PR other/18138
3296         * testsuite/lib/libffi-dg.exp: Accept more than one multilib libgcc.
3297
3298 2004-10-25  Kazuhiro Inaoka <inaoka.kazuhiro@renesas.com>
3299
3300         * src/m32r/libffitarget.h (FFI_CLOSURES): Set to 0.
3301
3302 2004-10-20  Kaz Kojima  <kkojima@gcc.gnu.org>
3303
3304         * src/sh/sysv.S (ffi_call_SYSV): Don't align for double data.
3305         * testsuite/libffi.call/float3.c: New test case.
3306
3307 2004-10-18  Kaz Kojima  <kkojima@gcc.gnu.org>
3308
3309         * src/sh/ffi.c (ffi_prep_closure): Set T bit in trampoline for
3310         the function returning a structure pointed with R2.
3311         * src/sh/sysv.S (ffi_closure_SYSV): Use R2 as the pointer to
3312         the structure return value if T bit set.  Emit position
3313         independent code and EH data if PIC.
3314
3315 2004-10-13  Kazuhiro Inaoka  <inaoka.kazuhiro@renesas.com>
3316
3317         * Makefile.am: Add m32r support.
3318         * configure.ac: Likewise.
3319         * Makefile.in: Regenerate.
3320         * confiugre: Regenerate.
3321         * src/types.c: Add m32r port to FFI_INTERNAL_TYPEDEF
3322         (uint64, sint64, double, longdouble)
3323         * src/m32r: New directory.
3324         * src/m32r/ffi.c: New file.
3325         * src/m32r/sysv.S: Likewise.
3326         * src/m32r/ffitarget.h: Likewise.
3327
3328 2004-10-02  Kaz Kojima  <kkojima@gcc.gnu.org>
3329
3330         * testsuite/libffi.call/negint.c: New test case.
3331
3332 2004-09-14  H.J. Lu  <hongjiu.lu@intel.com>
3333
3334         PR libgcj/17465
3335         * testsuite/lib/libffi-dg.exp: Don't use global ld_library_path.
3336         Set up LD_LIBRARY_PATH, SHLIB_PATH, LD_LIBRARYN32_PATH,
3337         LD_LIBRARY64_PATH, LD_LIBRARY_PATH_32, LD_LIBRARY_PATH_64 and
3338         DYLD_LIBRARY_PATH.
3339
3340 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
3341
3342         * testsuite/libffi.call/many_win32.c: Remove whitespaces.
3343         * testsuite/libffi.call/promotion.c: Likewise.
3344         * testsuite/libffi.call/return_ll.c: Remove unused var. Cleanup
3345         whitespaces.
3346         * testsuite/libffi.call/return_sc.c: Likewise.
3347         * testsuite/libffi.call/return_uc.c: Likewise.
3348
3349 2004-09-05  Andreas Tobler  <a.tobler@schweiz.ch>
3350
3351         * src/powerpc/darwin.S: Fix comments and identation.
3352         * src/powerpc/darwin_closure.S: Likewise.
3353
3354 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
3355
3356         * src/powerpc/ffi_darwin.c: Add flag for longdouble return values.
3357         (ffi_prep_args): Handle longdouble arguments.
3358         (ffi_prep_cif_machdep): Set flags for longdouble. Calculate space for
3359         longdouble.
3360         (ffi_closure_helper_DARWIN): Add closure handling for longdouble.
3361         * src/powerpc/darwin.S (_ffi_call_DARWIN): Add handling of longdouble
3362         values.
3363         * src/powerpc/darwin_closure.S (_ffi_closure_ASM): Likewise.
3364         * src/types.c: Defined longdouble size and alignment for darwin.
3365
3366 2004-09-02  Andreas Tobler  <a.tobler@schweiz.ch>
3367
3368         * src/powerpc/aix.S: Remove whitespaces.
3369         * src/powerpc/aix_closure.S: Likewise.
3370         * src/powerpc/asm.h: Likewise.
3371         * src/powerpc/ffi.c: Likewise.
3372         * src/powerpc/ffitarget.h: Likewise.
3373         * src/powerpc/linux64.S: Likewise.
3374         * src/powerpc/linux64_closure.S: Likewise.
3375         * src/powerpc/ppc_closure.S: Likewise.
3376         * src/powerpc/sysv.S: Likewise.
3377
3378 2004-08-30  Anthony Green  <green@redhat.com>
3379
3380         * Makefile.am: Add frv support.
3381         * Makefile.in, testsuite/Makefile.in: Rebuilt.
3382         * configure.ac: Read configure.host.
3383         * configure.in: Read configure.host.
3384         * configure.host: New file.  frv-elf needs libgloss.
3385         * include/ffi.h.in: Force ffi_closure to have a nice big (8)
3386         alignment.  This is needed to frv and shouldn't harm the others.
3387         * include/ffi_common.h (ALIGN_DOWN): New macro.
3388         * src/frv/ffi.c, src/frv/ffitarget.h, src/frv/eabi.S: New files.
3389
3390 2004-08-24  David Daney  <daney@avtrex.com>
3391
3392         * testsuite/libffi.call/closure_fn0.c: Xfail mips64* instead of mips*.
3393         * testsuite/libffi.call/closure_fn1.c: Likewise.
3394         * testsuite/libffi.call/closure_fn2.c  Likewise.
3395         * testsuite/libffi.call/closure_fn3.c: Likewise.
3396         * testsuite/libffi.call/closure_fn4.c: Likewise.
3397         * testsuite/libffi.call/closure_fn5.c: Likewise.
3398         * testsuite/libffi.call/cls_18byte.c: Likewise.
3399         * testsuite/libffi.call/cls_19byte.c: Likewise.
3400         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3401         * testsuite/libffi.call/cls_20byte.c: Likewise.
3402         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3403         * testsuite/libffi.call/cls_24byte.c: Likewise.
3404         * testsuite/libffi.call/cls_2byte.c: Likewise.
3405         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3406         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3407         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3408         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3409         * testsuite/libffi.call/cls_4byte.c: Likewise.
3410         * testsuite/libffi.call/cls_64byte.c: Likewise.
3411         * testsuite/libffi.call/cls_6byte.c: Likewise.
3412         * testsuite/libffi.call/cls_7byte.c: Likewise.
3413         * testsuite/libffi.call/cls_8byte.c: Likewise.
3414         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3415         * testsuite/libffi.call/cls_9byte2.c: Likewise.
3416         * testsuite/libffi.call/cls_align_double.c: Likewise.
3417         * testsuite/libffi.call/cls_align_float.c: Likewise.
3418         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
3419         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
3420         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
3421         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
3422         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
3423         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
3424         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
3425         * testsuite/libffi.call/cls_align_uint64.c: Likewise.
3426         * testsuite/libffi.call/cls_double.c: Likewise.
3427         * testsuite/libffi.call/cls_float.c: Likewise.
3428         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
3429         * testsuite/libffi.call/cls_multi_sshort.c: Likewise.
3430         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
3431         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
3432         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
3433         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
3434         * testsuite/libffi.call/cls_schar.c: Likewise.
3435         * testsuite/libffi.call/cls_sint.c: Likewise.
3436         * testsuite/libffi.call/cls_sshort.c: Likewise.
3437         * testsuite/libffi.call/cls_uchar.c: Likewise.
3438         * testsuite/libffi.call/cls_uint.c: Likewise.
3439         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3440         * testsuite/libffi.call/cls_ushort.c: Likewise.
3441         * testsuite/libffi.call/nested_struct.c: Likewise.
3442         * testsuite/libffi.call/nested_struct1.c: Likewise.
3443         * testsuite/libffi.call/nested_struct2.c: Likewise.
3444         * testsuite/libffi.call/nested_struct3.c: Likewise.
3445         * testsuite/libffi.call/problem1.c: Likewise.
3446         * testsuite/libffi.special/unwindtest.cc: Likewise.
3447         * testsuite/libffi.call/cls_12byte.c: Likewise and set return value
3448         to zero.
3449         * testsuite/libffi.call/cls_16byte.c: Likewise.
3450         * testsuite/libffi.call/cls_5byte.c: Likewise.
3451
3452 2004-08-23  David Daney <daney@avtrex.com>
3453
3454         PR libgcj/13141
3455         * src/mips/ffitarget.h (FFI_O32_SOFT_FLOAT): New ABI.
3456         * src/mips/ffi.c (ffi_prep_args): Fix alignment calculation.
3457         (ffi_prep_cif_machdep): Handle FFI_O32_SOFT_FLOAT floating point
3458         parameters and return types.
3459         (ffi_call): Handle FFI_O32_SOFT_FLOAT ABI.
3460         (ffi_prep_closure): Ditto.
3461         (ffi_closure_mips_inner_O32): Handle FFI_O32_SOFT_FLOAT ABI, fix
3462         alignment calculations.
3463         * src/mips/o32.S (ffi_closure_O32): Don't use floating point
3464         instructions if FFI_O32_SOFT_FLOAT, make stack frame ABI compliant.
3465
3466 2004-08-14  Casey Marshall <csm@gnu.org>
3467
3468         * src/mips/ffi.c (ffi_pref_cif_machdep): set `cif->flags' to
3469         contain `FFI_TYPE_UINT64' as return type for any 64-bit
3470         integer (O32 ABI only).
3471         (ffi_prep_closure): new function.
3472         (ffi_closure_mips_inner_O32): new function.
3473         * src/mips/ffitarget.h: Define `FFI_CLOSURES' and
3474         `FFI_TRAMPOLINE_SIZE' appropriately if the ABI is o32.
3475         * src/mips/o32.S (ffi_call_O32): add labels for .eh_frame. Return
3476         64 bit integers correctly.
3477         (ffi_closure_O32): new function.
3478         Added DWARF-2 unwind info for both functions.
3479
3480 2004-08-10  Andrew Haley  <aph@redhat.com>
3481
3482         * src/x86/ffi64.c (ffi_prep_args ): 8-align all stack arguments.
3483
3484 2004-08-01  Robert Millan  <robertmh@gnu.org>
3485
3486         * configure.ac: Detect knetbsd-gnu and kfreebsd-gnu.
3487         * configure: Regenerate.
3488
3489 2004-07-30  Maciej W. Rozycki  <macro@linux-mips.org>
3490
3491         * acinclude.m4 (AC_FUNC_MMAP_BLACKLIST): Check for <sys/mman.h>
3492         and mmap() explicitly instead of relying on preset autoconf cache
3493         variables.
3494         * aclocal.m4: Regenerate.
3495         * configure: Regenerate.
3496
3497 2004-07-11  Ulrich Weigand  <uweigand@de.ibm.com>
3498
3499         * src/s390/ffi.c (ffi_prep_args): Fix C aliasing violation.
3500         (ffi_check_float_struct): Remove unused prototype.
3501
3502 2004-06-30  Geoffrey Keating  <geoffk@apple.com>
3503
3504         * src/powerpc/ffi_darwin.c (flush_icache): ';' is a comment
3505         character on Darwin, use '\n\t' instead.
3506
3507 2004-06-26  Matthias Klose  <doko@debian.org>
3508
3509         * libtool-version: Fix typo in revision/age.
3510
3511 2004-06-17  Matthias Klose  <doko@debian.org>
3512
3513         * libtool-version: New.
3514         * Makefile.am (libffi_la_LDFLAGS): Use -version-info for soname.
3515         * Makefile.in: Regenerate.
3516
3517 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
3518
3519         * Makefile.am: Remove useless multilib rules.
3520         * Makefile.in: Regenerate.
3521         * aclocal.m4: Regenerate with automake 1.8.5.
3522         * configure.ac: Remove useless multilib configury.
3523         * configure: Regenerate.
3524
3525 2004-06-15  Paolo Bonzini  <bonzini@gnu.org>
3526
3527         * .cvsignore: New file.
3528
3529 2004-06-10  Jakub Jelinek  <jakub@redhat.com>
3530
3531         * src/ia64/unix.S (ffi_call_unix): Insert group barrier break
3532         fp_done.
3533         (ffi_closure_UNIX): Fix f14/f15 adjustment if FLOAT_SZ is ever
3534         changed from 8.
3535
3536 2004-06-06  Sean McNeil  <sean@mcneil.com>
3537
3538         * configure.ac: Add x86_64-*-freebsd* support.
3539         * configure: Regenerate.
3540
3541 2004-04-26  Joe Buck <jbuck@welsh-buck.org>
3542
3543         Bug 15093
3544         * configure.ac: Test for existence of mmap and sys/mman.h before
3545         checking blacklist.  Fix suggested by Jim Wilson.
3546         * configure: Regenerate.
3547
3548 2004-04-26  Matt Austern  <austern@apple.com>
3549
3550         * src/powerpc/darwin.S: Go through a non-lazy pointer for initial
3551         FDE location.
3552         * src/powerpc/darwin_closure.S: Likewise.
3553
3554 2004-04-24  Andreas Tobler  <a.tobler@schweiz.ch>
3555
3556         * testsuite/libffi.call/cls_multi_schar.c (main): Fix initialization
3557         error. Reported by Thomas Heller <theller@python.net>.
3558         * testsuite/libffi.call/cls_multi_sshort.c (main): Likewise.
3559         * testsuite/libffi.call/cls_multi_ushort.c (main): Likewise.
3560
3561 2004-03-20  Matthias Klose  <doko@debian.org>
3562
3563         * src/pa/linux.S: Fix typo.
3564
3565 2004-03-19  Matthias Klose  <doko@debian.org>
3566
3567         * Makefile.am: Update.
3568         * Makefile.in: Regenerate.
3569         * src/pa/ffi.h.in: Remove.
3570         * src/pa/ffitarget.h: New file.
3571
3572 2004-02-10  Randolph Chung  <tausq@debian.org>
3573
3574         * Makefile.am: Add PA support.
3575         * Makefile.in: Regenerate.
3576         * include/Makefile.in: Regenerate.
3577         * configure.ac: Add PA target.
3578         * configure: Regenerate.
3579         * src/pa/ffi.c: New file.
3580         * src/pa/ffi.h.in: Add PA support.
3581         * src/pa/linux.S: New file.
3582         * prep_cif.c: Add PA support.
3583
3584 2004-03-16  Hosaka Yuji  <hos@tamanegi.org>
3585
3586         * src/types.c: Fix alignment size of X86_WIN32 case int64 and
3587         double.
3588         * src/x86/ffi.c (ffi_prep_args): Replace ecif->cif->rtype->type
3589         with ecif->cif->flags.
3590         (ffi_call, ffi_prep_incoming_args_SYSV): Replace cif->rtype->type
3591         with cif->flags.
3592         (ffi_prep_cif_machdep): Add X86_WIN32 struct case.
3593         (ffi_closure_SYSV): Add 1 or 2-bytes struct case for X86_WIN32.
3594         * src/x86/win32.S (retstruct1b, retstruct2b, sc_retstruct1b,
3595         sc_retstruct2b): Add for 1 or 2-bytes struct case.
3596
3597 2004-03-15 Kelley Cook <kcook@gcc.gnu.org>
3598
3599         * configure.in: Rename file to ...
3600         * configure.ac: ... this.
3601         * fficonfig.h.in: Regenerate.
3602         * Makefile.in: Regenerate.
3603         * include/Makefile.in: Regenerate.
3604         * testsuite/Makefile.in: Regenerate.
3605
3606 2004-03-12  Matt Austern  <austern@apple.com>
3607
3608         * src/powerpc/darwin.S: Fix EH information so it corresponds to
3609         changes in EH format resulting from addition of linkonce support.
3610         * src/powerpc/darwin_closure.S: Likewise.
3611
3612 2004-03-11  Andreas Tobler  <a.tobler@schweiz.ch>
3613             Paolo Bonzini  <bonzini@gnu.org>
3614
3615         * Makefile.am (AUTOMAKE_OPTIONS): Set them.
3616         Remove VPATH. Remove rules for object files. Remove multilib support.
3617         (AM_CCASFLAGS): Add.
3618         * configure.in (AC_CONFIG_HEADERS): Relace AM_CONFIG_HEADER.
3619         (AC_PREREQ): Bump version to 2.59.
3620         (AC_INIT): Fill with version info and bug address.
3621         (ORIGINAL_LD_FOR_MULTILIBS): Remove.
3622         (AM_ENABLE_MULTILIB): Use this instead of AC_ARG_ENABLE.
3623         De-precious CC so that the right flags are passed down to multilibs.
3624         (AC_MSG_ERROR): Replace obsolete macro AC_ERROR.
3625         (AC_CONFIG_FILES): Replace obsolete macro AC_LINK_FILES.
3626         (AC_OUTPUT): Reorganize the output with AC_CONFIG_COMMANDS.
3627         * configure: Rebuilt.
3628         * aclocal.m4: Likewise.
3629         * Makefile.in, include/Makefile.in, testsuite/Makefile.in: Likewise.
3630         * fficonfig.h.in: Likewise.
3631
3632 2004-03-11  Andreas Schwab  <schwab@suse.de>
3633
3634         * src/ia64/ffi.c (ffi_prep_incoming_args_UNIX): Get floating point
3635         arguments from fp registers only for the first 8 parameter slots.
3636         Don't convert a float parameter when passed in memory.
3637
3638 2004-03-09  Hans-Peter Nilsson  <hp@axis.com>
3639
3640         * configure: Regenerate for config/accross.m4 correction.
3641
3642 2004-02-25  Matt Kraai  <kraai@alumni.cmu.edu>
3643
3644         * src/powerpc/ffi.c (ffi_prep_args_SYSV): Change
3645         ecif->cif->bytes to bytes.
3646         (ffi_prep_cif_machdep): Add braces around nested if statement.
3647
3648 2004-02-09  Alan Modra  <amodra@bigpond.net.au>
3649
3650         * src/types.c (pointer): POWERPC64 has 8 byte pointers.
3651
3652         * src/powerpc/ffi.c (ffi_prep_args64): Correct long double handling.
3653         (ffi_closure_helper_LINUX64): Fix typo.
3654         * testsuite/libffi.call/cls_align_longdouble.c: Pass -mlong-double-128
3655         for powerpc64-*-*.
3656         * testsuite/libffi.call/float.c: Likewise.
3657         * testsuite/libffi.call/float2.c: Likewise.
3658
3659 2004-02-08  Alan Modra  <amodra@bigpond.net.au>
3660
3661         * src/powerpc/ffi.c (ffi_prep_cif_machdep <FFI_LINUX64>): Correct
3662         long double function return and long double arg handling.
3663         (ffi_closure_helper_LINUX64): Formatting.  Delete unused "ng" var.
3664         Use "end_pfr" instead of "nf".  Correct long double handling.
3665         Localise "temp".
3666         * src/powerpc/linux64.S (ffi_call_LINUX64): Save f2 long double
3667         return value.
3668         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Allocate
3669         space for long double return value.  Adjust stack frame and offsets.
3670         Load f2 long double return.
3671
3672 2004-02-07  Alan Modra  <amodra@bigpond.net.au>
3673
3674         * src/types.c: Use 16 byte long double for POWERPC64.
3675
3676 2004-01-25  Eric Botcazou  <ebotcazou@libertysurf.fr>
3677
3678         * src/sparc/ffi.c (ffi_prep_args_v9): Shift the parameter array
3679         when the structure return address is passed in %o0.
3680         (ffi_V9_return_struct): Rename into ffi_v9_layout_struct.
3681         (ffi_v9_layout_struct): Align the field following a nested structure
3682         on a word boundary.  Use memmove instead of memcpy.
3683         (ffi_call): Update call to ffi_V9_return_struct.
3684         (ffi_prep_closure): Define 'ctx' only for V8.
3685         (ffi_closure_sparc_inner): Clone into ffi_closure_sparc_inner_v8
3686         and ffi_closure_sparc_inner_v9.
3687         (ffi_closure_sparc_inner_v8): Return long doubles by reference.
3688         Always skip the structure return address.  For structures and long
3689         doubles, copy the argument directly.
3690         (ffi_closure_sparc_inner_v9): Skip the structure return address only
3691         if required.  Shift the maximum floating-point slot accordingly.  For
3692         big structures, copy the argument directly; otherwise, left-justify the
3693         argument and call ffi_v9_layout_struct to lay out the structure on
3694         the stack.
3695         * src/sparc/v8.S: Undef STACKFRAME before defining it.
3696         (ffi_closure_v8): Pass the structure return address.  Update call to
3697         ffi_closure_sparc_inner_v8.  Short-circuit FFI_TYPE_INT handling.
3698         Skip the 'unimp' insn when returning long doubles and structures.
3699         * src/sparc/v9.S: Undef STACKFRAME before defining it.
3700         (ffi_closure_v9): Increase the frame size by 2 words.  Short-circuit
3701         FFI_TYPE_INT handling.  Load structures both in integers and
3702         floating-point registers on return.
3703         * README: Update status of the SPARC port.
3704
3705 2004-01-24  Andreas Tobler  <a.tobler@schweiz.ch>
3706
3707         * testsuite/libffi.call/pyobjc-tc.c (main): Treat result value
3708         as of type ffi_arg.
3709         * testsuite/libffi.call/struct3.c (main): Fix CHECK.
3710
3711 2004-01-22  Ulrich Weigand  <uweigand@de.ibm.com>
3712
3713         * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Treat result
3714         value as of type ffi_arg, not unsigned int.
3715
3716 2004-01-21  Michael Ritzert  <ritzert@t-online.de>
3717
3718         * ffi64.c (ffi_prep_args): Cast the RHS of an assignment instead
3719         of the LHS.
3720
3721 2004-01-12  Andreas Tobler  <a.tobler@schweiz.ch>
3722
3723         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_32 for
3724         Solaris.
3725
3726 2004-01-08  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
3727
3728         * testsuite/libffi.call/ffitest.h (allocate_mmap): Cast MAP_FAILED
3729         to void *.
3730
3731 2003-12-10  Richard Henderson  <rth@redhat.com>
3732
3733         * testsuite/libffi.call/cls_align_pointer.c: Cast pointers to
3734         size_t instead of int.
3735
3736 2003-12-04  Hosaka Yuji  <hos@tamanegi.org>
3737
3738         * testsuite/libffi.call/many_win32.c: Include <float.h>.
3739         * testsuite/libffi.call/many_win32.c (main): Replace variable
3740         int i with unsigned long ul.
3741
3742         * testsuite/libffi.call/cls_align_uint64.c: New test case.
3743         * testsuite/libffi.call/cls_align_sint64.c: Likewise.
3744         * testsuite/libffi.call/cls_align_uint32.c: Likewise.
3745         * testsuite/libffi.call/cls_align_sint32.c: Likewise.
3746         * testsuite/libffi.call/cls_align_uint16.c: Likewise.
3747         * testsuite/libffi.call/cls_align_sint16.c: Likewise.
3748         * testsuite/libffi.call/cls_align_float.c: Likewise.
3749         * testsuite/libffi.call/cls_align_double.c: Likewise.
3750         * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
3751         * testsuite/libffi.call/cls_align_pointer.c: Likewise.
3752
3753 2003-12-02  Hosaka Yuji  <hos@tamanegi.org>
3754
3755         PR other/13221
3756         * src/x86/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV):
3757         Align arguments to 32 bits.
3758
3759 2003-12-01  Andreas Tobler  <a.tobler@schweiz.ch>
3760
3761         PR other/13221
3762         * testsuite/libffi.call/cls_multi_sshort.c: New test case.
3763         * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
3764         * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
3765         * testsuite/libffi.call/cls_multi_schar.c: Likewise.
3766         * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
3767         * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
3768
3769         * testsuite/libffi.special/unwindtest.cc: Cosmetics.
3770
3771 2003-11-26  Kaveh R. Ghazi  <ghazi@caip.rutgers.edu>
3772
3773         * testsuite/libffi.call/ffitest.h: Include <fcntl.h>.
3774         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3775
3776 2003-11-22  Andreas Tobler  <a.tobler@schweiz.ch>
3777
3778         * Makefile.in: Rebuilt.
3779         * configure: Likewise.
3780         * testsuite/libffi.special/unwindtest.cc: Convert the mmap to
3781         the right type.
3782
3783 2003-11-21  Andreas Jaeger  <aj@suse.de>
3784             Andreas Tobler  <a.tobler@schweiz.ch>
3785
3786         * acinclude.m4: Add AC_FUNC_MMAP_BLACKLIST.
3787         * configure.in: Call AC_FUNC_MMAP_BLACKLIST.
3788         * Makefile.in: Rebuilt.
3789         * aclocal.m4: Likewise.
3790         * configure: Likewise.
3791         * fficonfig.h.in: Likewise.
3792         * testsuite/lib/libffi-dg.exp: Add include dir.
3793         * testsuite/libffi.call/ffitest.h: Add MMAP definitions.
3794         * testsuite/libffi.special/ffitestcxx.h: Likewise.
3795         * testsuite/libffi.call/closure_fn0.c: Use MMAP functionality
3796         for ffi_closure if available.
3797         * testsuite/libffi.call/closure_fn1.c: Likewise.
3798         * testsuite/libffi.call/closure_fn2.c: Likewise.
3799         * testsuite/libffi.call/closure_fn3.c: Likewise.
3800         * testsuite/libffi.call/closure_fn4.c: Likewise.
3801         * testsuite/libffi.call/closure_fn5.c: Likewise.
3802         * testsuite/libffi.call/cls_12byte.c: Likewise.
3803         * testsuite/libffi.call/cls_16byte.c: Likewise.
3804         * testsuite/libffi.call/cls_18byte.c: Likewise.
3805         * testsuite/libffi.call/cls_19byte.c: Likewise.
3806         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3807         * testsuite/libffi.call/cls_20byte.c: Likewise.
3808         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3809         * testsuite/libffi.call/cls_24byte.c: Likewise.
3810         * testsuite/libffi.call/cls_2byte.c: Likewise.
3811         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3812         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3813         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3814         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3815         * testsuite/libffi.call/cls_4byte.c: Likewise.
3816         * testsuite/libffi.call/cls_5byte.c: Likewise.
3817         * testsuite/libffi.call/cls_64byte.c: Likewise.
3818         * testsuite/libffi.call/cls_6byte.c: Likewise.
3819         * testsuite/libffi.call/cls_7byte.c: Likewise.
3820         * testsuite/libffi.call/cls_8byte.c: Likewise.
3821         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3822         * testsuite/libffi.call/cls_9byte2.c: Likewise.
3823         * testsuite/libffi.call/cls_double.c: Likewise.
3824         * testsuite/libffi.call/cls_float.c: Likewise.
3825         * testsuite/libffi.call/cls_schar.c: Likewise.
3826         * testsuite/libffi.call/cls_sint.c: Likewise.
3827         * testsuite/libffi.call/cls_sshort.c: Likewise.
3828         * testsuite/libffi.call/cls_uchar.c: Likewise.
3829         * testsuite/libffi.call/cls_uint.c: Likewise.
3830         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3831         * testsuite/libffi.call/cls_ushort.c: Likewise.
3832         * testsuite/libffi.call/nested_struct.c: Likewise.
3833         * testsuite/libffi.call/nested_struct1.c: Likewise.
3834         * testsuite/libffi.call/nested_struct2.c: Likewise.
3835         * testsuite/libffi.call/nested_struct3.c: Likewise.
3836         * testsuite/libffi.call/problem1.c: Likewise.
3837         * testsuite/libffi.special/unwindtest.cc: Likewise.
3838
3839 2003-11-20  Andreas Tobler  <a.tobler@schweiz.ch>
3840
3841         * testsuite/lib/libffi-dg.exp: Make the -lgcc_s conditional.
3842
3843 2003-11-19  Andreas Tobler  <a.tobler@schweiz.ch>
3844
3845         * testsuite/lib/libffi-dg.exp: Add DYLD_LIBRARY_PATH for darwin.
3846         Add -lgcc_s to additional flags.
3847
3848 2003-11-12  Andreas Tobler  <a.tobler@schweiz.ch>
3849
3850         * configure.in, include/Makefile.am: PR libgcj/11147, install
3851         the ffitarget.h header file in a gcc versioned and target
3852         dependent place.
3853         * configure: Regenerated.
3854         * Makefile.in, include/Makefile.in: Likewise.
3855         * testsuite/Makefile.in: Likewise.
3856
3857 2003-11-09  Andreas Tobler  <a.tobler@schweiz.ch>
3858
3859         * testsuite/libffi.call/closure_fn0.c: Print result and check
3860         with dg-output to make debugging easier.
3861         * testsuite/libffi.call/closure_fn1.c: Likewise.
3862         * testsuite/libffi.call/closure_fn2.c: Likewise.
3863         * testsuite/libffi.call/closure_fn3.c: Likewise.
3864         * testsuite/libffi.call/closure_fn4.c: Likewise.
3865         * testsuite/libffi.call/closure_fn5.c: Likewise.
3866         * testsuite/libffi.call/cls_12byte.c: Likewise.
3867         * testsuite/libffi.call/cls_16byte.c: Likewise.
3868         * testsuite/libffi.call/cls_18byte.c: Likewise.
3869         * testsuite/libffi.call/cls_19byte.c: Likewise.
3870         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
3871         * testsuite/libffi.call/cls_20byte.c: Likewise.
3872         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3873         * testsuite/libffi.call/cls_24byte.c: Likewise.
3874         * testsuite/libffi.call/cls_2byte.c: Likewise.
3875         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
3876         * testsuite/libffi.call/cls_3byte1.c: Likewise.
3877         * testsuite/libffi.call/cls_3byte2.c: Likewise.
3878         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
3879         * testsuite/libffi.call/cls_4byte.c: Likewise.
3880         * testsuite/libffi.call/cls_5byte.c: Likewise.
3881         * testsuite/libffi.call/cls_64byte.c: Likewise.
3882         * testsuite/libffi.call/cls_6byte.c: Likewise.
3883         * testsuite/libffi.call/cls_7byte.c: Likewise.
3884         * testsuite/libffi.call/cls_8byte.c: Likewise.
3885         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3886         * testsuite/libffi.call/cls_9byte2.c: Likewise.
3887         * testsuite/libffi.call/cls_double.c: Likewise.
3888         * testsuite/libffi.call/cls_float.c: Likewise.
3889         * testsuite/libffi.call/cls_schar.c: Likewise.
3890         * testsuite/libffi.call/cls_sint.c: Likewise.
3891         * testsuite/libffi.call/cls_sshort.c: Likewise.
3892         * testsuite/libffi.call/cls_uchar.c: Likewise.
3893         * testsuite/libffi.call/cls_uint.c: Likewise.
3894         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3895         * testsuite/libffi.call/cls_ushort.c: Likewise.
3896         * testsuite/libffi.call/problem1.c: Likewise.
3897
3898         * testsuite/libffi.special/unwindtest.cc: Make ffi_closure
3899         static.
3900
3901 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
3902
3903         * testsuite/libffi.call/cls_9byte2.c: New test case.
3904         * testsuite/libffi.call/cls_9byte1.c: Likewise.
3905         * testsuite/libffi.call/cls_64byte.c: Likewise.
3906         * testsuite/libffi.call/cls_20byte1.c: Likewise.
3907         * testsuite/libffi.call/cls_19byte.c: Likewise.
3908         * testsuite/libffi.call/cls_18byte.c: Likewise.
3909         * testsuite/libffi.call/closure_fn4.c: Likewise.
3910         * testsuite/libffi.call/closure_fn5.c: Likewise.
3911         * testsuite/libffi.call/cls_schar.c: Likewise.
3912         * testsuite/libffi.call/cls_sint.c: Likewise.
3913         * testsuite/libffi.call/cls_sshort.c: Likewise.
3914         * testsuite/libffi.call/nested_struct2.c: Likewise.
3915         * testsuite/libffi.call/nested_struct3.c: Likewise.
3916
3917 2003-11-08  Andreas Tobler  <a.tobler@schweiz.ch>
3918
3919         * testsuite/libffi.call/cls_double.c: Do a check on the result.
3920         * testsuite/libffi.call/cls_uchar.c: Likewise.
3921         * testsuite/libffi.call/cls_uint.c: Likewise.
3922         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
3923         * testsuite/libffi.call/cls_ushort.c: Likewise.
3924         * testsuite/libffi.call/return_sc.c: Cleanup whitespaces.
3925
3926 2003-11-06  Andreas Tobler  <a.tobler@schweiz.ch>
3927
3928         * src/prep_cif.c (ffi_prep_cif): Move the validity check after
3929         the initialization.
3930
3931 2003-10-23  Andreas Tobler  <a.tobler@schweiz.ch>
3932
3933         * src/java_raw_api.c (ffi_java_ptrarray_to_raw): Replace
3934         FFI_ASSERT(FALSE) with FFI_ASSERT(0).
3935
3936 2003-10-22  David Daney  <ddaney@avtrex.com>
3937
3938         * src/mips/ffitarget.h: Replace undefined UINT32 and friends with
3939         __attribute__((__mode__(__SI__))) and friends.
3940
3941 2003-10-22  Andreas Schwab  <schwab@suse.de>
3942
3943         * src/ia64/ffi.c: Replace FALSE/TRUE with false/true.
3944
3945 2003-10-21  Andreas Tobler  <a.tobler@schweiz.ch>
3946
3947         * configure.in: AC_LINK_FILES(ffitarget.h).
3948         * configure: Regenerate.
3949         * Makefile.in: Likewise.
3950         * include/Makefile.in: Likewise.
3951         * testsuite/Makefile.in: Likewise.
3952         * fficonfig.h.in: Likewise.
3953
3954 2003-10-21  Paolo Bonzini  <bonzini@gnu.org>
3955             Richard Henderson  <rth@redhat.com>
3956
3957         Avoid that ffi.h includes fficonfig.h.
3958
3959         * Makefile.am (EXTRA_DIST): Include ffitarget.h files
3960         (TARGET_SRC_MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3961         (TARGET_SRC_MIPS_SGI): Removed.
3962         (MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
3963         (MIPS_SGI): Removed.
3964         (CLEANFILES): Removed.
3965         (mostlyclean-am, clean-am, mostlyclean-sub, clean-sub): New
3966         targets.
3967         * acconfig.h: Removed.
3968         * configure.in: Compute sizeofs only for double and long double.
3969         Use them to define and subst HAVE_LONG_DOUBLE.  Include comments
3970         into AC_DEFINE instead of using acconfig.h.  Create
3971         include/ffitarget.h instead of include/fficonfig.h.  Rename
3972         MIPS_GCC to MIPS_IRIX, drop MIPS_SGI since we are in gcc's tree.
3973         AC_DEFINE EH_FRAME_FLAGS.
3974         * include/Makefile.am (DISTCLEANFILES): New automake macro.
3975         (hack_DATA): Add ffitarget.h.
3976         * include/ffi.h.in: Remove all system specific definitions.
3977         Declare raw API even if it is not installed, why bother?
3978         Use limits.h instead of SIZEOF_* to define ffi_type_*.  Do
3979         not define EH_FRAME_FLAGS, it is in fficonfig.h now.  Include
3980         ffitarget.h instead of fficonfig.h.  Remove ALIGN macro.
3981         (UINT_ARG, INT_ARG): Removed, use ffi_arg and ffi_sarg instead.
3982         * include/ffi_common.h (bool): Do not define.
3983         (ffi_assert): Accept failed assertion.
3984         (ffi_type_test): Return void and accept file/line.
3985         (FFI_ASSERT): Pass stringized failed assertion.
3986         (FFI_ASSERT_AT): New macro.
3987         (FFI_ASSERT_VALID_TYPE): New macro.
3988         (UINT8, SINT8, UINT16, SINT16, UINT32, SINT32,
3989         UINT64, SINT64): Define here with gcc's __attribute__ macro
3990         instead of in ffi.h
3991         (FLOAT32, ALIGN): Define here instead of in ffi.h
3992         * include/ffi-mips.h: Removed.  Its content moved to
3993         src/mips/ffitarget.h after separating assembly and C sections.
3994         * src/alpha/ffi.c, src/alpha/ffi.c, src/java_raw_api.c
3995         src/prep_cif.c, src/raw_api.c, src/ia64/ffi.c,
3996         src/mips/ffi.c, src/mips/n32.S, src/mips/o32.S,
3997         src/mips/ffitarget.h, src/sparc/ffi.c, src/x86/ffi64.c:
3998         SIZEOF_ARG -> FFI_SIZEOF_ARG.
3999         * src/ia64/ffi.c: Include stdbool.h (provided by GCC 2.95+).
4000         * src/debug.c (ffi_assert): Accept stringized failed assertion.
4001         (ffi_type_test): Rewritten.
4002         * src/prep-cif.c (initialize_aggregate, ffi_prep_cif): Call
4003         FFI_ASSERT_VALID_TYPE.
4004         * src/alpha/ffitarget.h, src/arm/ffitarget.h,
4005         src/ia64/ffitarget.h, src/m68k/ffitarget.h,
4006         src/mips/ffitarget.h, src/powerpc/ffitarget.h,
4007         src/s390/ffitarget.h, src/sh/ffitarget.h,
4008         src/sh64/ffitarget.h, src/sparc/ffitarget.h,
4009         src/x86/ffitarget.h: New files.
4010         * src/alpha/osf.S, src/arm/sysv.S, src/ia64/unix.S,
4011         src/m68k/sysv.S, src/mips/n32.S, src/mips/o32.S,
4012         src/powerpc/aix.S, src/powerpc/darwin.S,
4013         src/powerpc/ffi_darwin.c, src/powerpc/linux64.S,
4014         src/powerpc/linux64_closure.S, src/powerpc/ppc_closure.S,
4015         src/powerpc/sysv.S, src/s390/sysv.S, src/sh/sysv.S,
4016         src/sh64/sysv.S, src/sparc/v8.S, src/sparc/v9.S,
4017         src/x86/sysv.S, src/x86/unix64.S, src/x86/win32.S:
4018         include fficonfig.h
4019
4020 2003-10-20  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
4021
4022         * src/mips/ffi.c: Use _ABIN32, _ABIO32 instead of external
4023         _MIPS_SIM_NABI32, _MIPS_SIM_ABI32.
4024
4025 2003-10-19  Andreas Tobler  <a.tobler@schweiz.ch>
4026
4027         * src/powerpc/ffi_darwin.c (ffi_prep_args): Declare bytes again.
4028         Used when FFI_DEBUG = 1.
4029
4030 2003-10-14  Alan Modra  <amodra@bigpond.net.au>
4031
4032         * src/types.c (double, longdouble): Default POWERPC64 to 8 byte size
4033         and align.
4034
4035 2003-10-06  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
4036
4037         * include/ffi_mips.h: Define FFI_MIPS_N32 for N32/N64 ABIs,
4038         FFI_MIPS_O32 for O32 ABI.
4039
4040 2003-10-01  Andreas Tobler  <a.tobler@schweiz.ch>
4041
4042         * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_64 for
4043         SPARC64. Cleanup whitespaces.
4044
4045 2003-09-19  Andreas Tobler  <a.tobler@schweiz.ch>
4046
4047         * testsuite/libffi.call/closure_fn0.c: Xfail mips, arm,
4048         strongarm, xscale. Cleanup whitespaces.
4049         * testsuite/libffi.call/closure_fn1.c: Likewise.
4050         * testsuite/libffi.call/closure_fn2.c: Likewise.
4051         * testsuite/libffi.call/closure_fn3.c: Likewise.
4052         * testsuite/libffi.call/cls_12byte.c: Likewise.
4053         * testsuite/libffi.call/cls_16byte.c: Likewise.
4054         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
4055         * testsuite/libffi.call/cls_20byte.c: Likewise.
4056         * testsuite/libffi.call/cls_24byte.c: Likewise.
4057         * testsuite/libffi.call/cls_2byte.c: Likewise.
4058         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
4059         * testsuite/libffi.call/cls_3byte1.c: Likewise.
4060         * testsuite/libffi.call/cls_3byte2.c: Likewise.
4061         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
4062         * testsuite/libffi.call/cls_4byte.c: Likewise.
4063         * testsuite/libffi.call/cls_5byte.c: Likewise.
4064         * testsuite/libffi.call/cls_6byte.c: Likewise.
4065         * testsuite/libffi.call/cls_7byte.c: Likewise.
4066         * testsuite/libffi.call/cls_8byte.c: Likewise.
4067         * testsuite/libffi.call/cls_double.c: Likewise.
4068         * testsuite/libffi.call/cls_float.c: Likewise.
4069         * testsuite/libffi.call/cls_uchar.c: Likewise.
4070         * testsuite/libffi.call/cls_uint.c: Likewise.
4071         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
4072         * testsuite/libffi.call/cls_ushort.c: Likewise.
4073         * testsuite/libffi.call/nested_struct.c: Likewise.
4074         * testsuite/libffi.call/nested_struct1.c: Likewise.
4075         * testsuite/libffi.call/problem1.c: Likewise.
4076         * testsuite/libffi.special/unwindtest.cc: Likewise.
4077         * testsuite/libffi.call/pyobjc-tc.c: Cleanup whitespaces.
4078
4079 2003-09-18  David Edelsohn  <edelsohn@gnu.org>
4080
4081         * src/powerpc/aix.S: Cleanup whitespaces.
4082         * src/powerpc/aix_closure.S: Likewise.
4083
4084 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
4085
4086         * src/powerpc/darwin.S: Cleanup whitespaces, comment formatting.
4087         * src/powerpc/darwin_closure.S: Likewise.
4088         * src/powerpc/ffi_darwin.c: Likewise.
4089
4090 2003-09-18  Andreas Tobler  <a.tobler@schweiz.ch>
4091             David Edelsohn  <edelsohn@gnu.org>
4092
4093         * src/types.c (double): Add AIX and Darwin to the right TYPEDEF.
4094         * src/powerpc/aix_closure.S: Remove the pointer to the outgoing
4095         parameter stack.
4096         * src/powerpc/darwin_closure.S: Likewise.
4097         * src/powerpc/ffi_darwin.c (ffi_prep_args): Handle structures
4098         according to the Darwin/AIX ABI.
4099         (ffi_prep_cif_machdep): Likewise.
4100         (ffi_closure_helper_DARWIN): Likewise.
4101         Remove the outgoing parameter stack logic. Simplify the evaluation
4102         of the different CASE types.
4103         (ffi_prep_clousure): Avoid the casts on lvalues. Change the branch
4104         statement in the trampoline code.
4105
4106 2003-09-18  Kaz Kojima  <kkojima@gcc.gnu.org>
4107
4108         * src/sh/ffi.c (ffi_prep_args): Take account into the alignement
4109         for the register size.
4110         (ffi_closure_helper_SYSV): Handle the structure return value
4111         address correctly.
4112         (ffi_closure_helper_SYSV): Return the appropriate type when
4113         the registers are used for the structure return value.
4114         * src/sh/sysv.S (ffi_closure_SYSV): Fix the stack layout for
4115         the 64-bit return value.  Update copyright years.
4116
4117 2003-09-17  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
4118
4119         * testsuite/lib/libffi-dg.exp (libffi_target_compile): Search in
4120         srcdir for ffi_mips.h.
4121
4122 2003-09-12  Alan Modra  <amodra@bigpond.net.au>
4123
4124         * src/prep_cif.c (initialize_aggregate): Include tail padding in
4125         structure size.
4126         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Correct
4127         placement of float result.
4128         * testsuite/libffi.special/unwindtest.cc (closure_test_fn1): Correct
4129         cast of "resp" for big-endian 64 bit machines.
4130
4131 2003-09-11  Alan Modra  <amodra@bigpond.net.au>
4132
4133         * src/types.c (double, longdouble): Merge identical SH and ARM
4134         typedefs, and add POWERPC64.
4135         * src/powerpc/ffi.c (ffi_prep_args64): Correct next_arg calc for
4136         struct split over gpr and rest.
4137         (ffi_prep_cif_machdep): Correct intarg_count for structures.
4138         * src/powerpc/linux64.S (ffi_call_LINUX64): Fix gpr offsets.
4139
4140 2003-09-09  Andreas Tobler  <a.tobler@schweiz.ch>
4141
4142         * src/powerpc/ffi.c (ffi_closure_helper_SYSV) Handle struct
4143         passing correctly.
4144
4145 2003-09-09  Alan Modra  <amodra@bigpond.net.au>
4146
4147         * configure: Regenerate.
4148
4149 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
4150
4151         * Makefile.am: Remove build rules for ffitest.
4152         * Makefile.in: Rebuilt.
4153
4154 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
4155
4156         * src/java_raw_api.c: Include <stdlib.h> to fix compiler warning
4157         about implicit declaration of abort().
4158
4159 2003-09-04  Andreas Tobler  <a.tobler@schweiz.ch>
4160
4161         * Makefile.am: Add dejagnu test framework. Fixes PR other/11411.
4162         * Makefile.in: Rebuilt.
4163         * configure.in: Add dejagnu test framework.
4164         * configure: Rebuilt.
4165
4166         * testsuite/Makefile.am: New file.
4167         * testsuite/Makefile.in: Built
4168         * testsuite/lib/libffi-dg.exp: New file.
4169         * testsuite/config/default.exp: Likewise.
4170         * testsuite/libffi.call/call.exp: Likewise.
4171         * testsuite/libffi.call/ffitest.h: Likewise.
4172         * testsuite/libffi.call/closure_fn0.c: Likewise.
4173         * testsuite/libffi.call/closure_fn1.c: Likewise.
4174         * testsuite/libffi.call/closure_fn2.c: Likewise.
4175         * testsuite/libffi.call/closure_fn3.c: Likewise.
4176         * testsuite/libffi.call/cls_1_1byte.c: Likewise.
4177         * testsuite/libffi.call/cls_3_1byte.c: Likewise.
4178         * testsuite/libffi.call/cls_4_1byte.c: Likewise.
4179         * testsuite/libffi.call/cls_2byte.c: Likewise.
4180         * testsuite/libffi.call/cls_3byte1.c: Likewise.
4181         * testsuite/libffi.call/cls_3byte2.c: Likewise.
4182         * testsuite/libffi.call/cls_4byte.c: Likewise.
4183         * testsuite/libffi.call/cls_5byte.c: Likewise.
4184         * testsuite/libffi.call/cls_6byte.c: Likewise.
4185         * testsuite/libffi.call/cls_7byte.c: Likewise.
4186         * testsuite/libffi.call/cls_8byte.c: Likewise.
4187         * testsuite/libffi.call/cls_12byte.c: Likewise.
4188         * testsuite/libffi.call/cls_16byte.c: Likewise.
4189         * testsuite/libffi.call/cls_20byte.c: Likewise.
4190         * testsuite/libffi.call/cls_24byte.c: Likewise.
4191         * testsuite/libffi.call/cls_double.c: Likewise.
4192         * testsuite/libffi.call/cls_float.c: Likewise.
4193         * testsuite/libffi.call/cls_uchar.c: Likewise.
4194         * testsuite/libffi.call/cls_uint.c: Likewise.
4195         * testsuite/libffi.call/cls_ulonglong.c: Likewise.
4196         * testsuite/libffi.call/cls_ushort.c: Likewise.
4197         * testsuite/libffi.call/float.c: Likewise.
4198         * testsuite/libffi.call/float1.c: Likewise.
4199         * testsuite/libffi.call/float2.c: Likewise.
4200         * testsuite/libffi.call/many.c: Likewise.
4201         * testsuite/libffi.call/many_win32.c: Likewise.
4202         * testsuite/libffi.call/nested_struct.c: Likewise.
4203         * testsuite/libffi.call/nested_struct1.c: Likewise.
4204         * testsuite/libffi.call/pyobjc-tc.c: Likewise.
4205         * testsuite/libffi.call/problem1.c: Likewise.
4206         * testsuite/libffi.call/promotion.c: Likewise.
4207         * testsuite/libffi.call/return_ll.c: Likewise.
4208         * testsuite/libffi.call/return_sc.c: Likewise.
4209         * testsuite/libffi.call/return_uc.c: Likewise.
4210         * testsuite/libffi.call/strlen.c: Likewise.
4211         * testsuite/libffi.call/strlen_win32.c: Likewise.
4212         * testsuite/libffi.call/struct1.c: Likewise.
4213         * testsuite/libffi.call/struct2.c: Likewise.
4214         * testsuite/libffi.call/struct3.c: Likewise.
4215         * testsuite/libffi.call/struct4.c: Likewise.
4216         * testsuite/libffi.call/struct5.c: Likewise.
4217         * testsuite/libffi.call/struct6.c: Likewise.
4218         * testsuite/libffi.call/struct7.c: Likewise.
4219         * testsuite/libffi.call/struct8.c: Likewise.
4220         * testsuite/libffi.call/struct9.c: Likewise.
4221         * testsuite/libffi.special/special.exp: New file.
4222         * testsuite/libffi.special/ffitestcxx.h: Likewise.
4223         * testsuite/libffi.special/unwindtest.cc: Likewise.
4224
4225
4226 2003-08-13  Kaz Kojima  <kkojima@gcc.gnu.org>
4227
4228         * src/sh/ffi.c (OFS_INT16): Set 0 for little endian case.  Update
4229         copyright years.
4230
4231 2003-08-02  Alan Modra  <amodra@bigpond.net.au>
4232
4233         * src/powerpc/ffi.c (ffi_prep_args64): Modify for changed gcc
4234         structure passing.
4235         (ffi_closure_helper_LINUX64): Likewise.
4236         * src/powerpc/linux64.S: Remove code writing to parm save area.
4237         * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Use return
4238         address in lr from ffi_closure_helper_LINUX64 call to calculate
4239         table address.  Optimize function tail.
4240
4241 2003-07-28  Andreas Tobler  <a.tobler@schweiz.ch>
4242
4243         * src/sparc/ffi.c: Handle all floating point registers.
4244         * src/sparc/v9.S: Likewise. Fixes second part of PR target/11410.
4245
4246 2003-07-11  Gerald Pfeifer  <pfeifer@dbai.tuwien.ac.at>
4247
4248         * README: Note that libffi is not part of GCC.  Update the project
4249         URL and status.
4250
4251 2003-06-19  Franz Sirl  <Franz.Sirl-kernel@lauterbach.com>
4252
4253         * src/powerpc/ppc_closure.S: Include ffi.h.
4254
4255 2003-06-13  Rainer Orth  <ro@TechFak.Uni-Bielefeld.DE>
4256
4257         * src/x86/sysv.S: Avoid gas-only .uleb128/.sleb128 directives.
4258         Use C style comments.
4259
4260 2003-06-13  Kaz Kojima  <kkojima@rr.iij4u.or.jp>
4261
4262         * Makefile.am: Add SHmedia support.  Fix a typo of SH support.
4263         * Makefile.in: Regenerate.
4264         * configure.in (sh64-*-linux*, sh5*-*-linux*): Add target.
4265         * configure: Regenerate.
4266         * include/ffi.h.in: Add SHmedia support.
4267         * src/sh64/ffi.c: New file.
4268         * src/sh64/sysv.S: New file.
4269
4270 2003-05-16  Jakub Jelinek  <jakub@redhat.com>
4271
4272         * configure.in (HAVE_RO_EH_FRAME): Check whether .eh_frame section
4273         should be read-only.
4274         * configure: Rebuilt.
4275         * fficonfig.h.in: Rebuilt.
4276         * include/ffi.h.in (EH_FRAME_FLAGS): Define.
4277         * src/alpha/osf.S: Use EH_FRAME_FLAGS.
4278         * src/powerpc/linux64.S: Likewise.
4279         * src/powerpc/linux64_closure.S: Likewise.  Include ffi.h.
4280         * src/powerpc/sysv.S: Use EH_FRAME_FLAGS.  Use pcrel encoding
4281         if -fpic/-fPIC/-mrelocatable.
4282         * src/powerpc/powerpc_closure.S: Likewise.
4283         * src/sparc/v8.S: If HAVE_RO_EH_FRAME is defined, don't include
4284         #write in .eh_frame flags.
4285         * src/sparc/v9.S: Likewise.
4286         * src/x86/unix64.S: Use EH_FRAME_FLAGS.
4287         * src/x86/sysv.S: Likewise.  Use pcrel encoding if -fpic/-fPIC.
4288         * src/s390/sysv.S: Use EH_FRAME_FLAGS.  Include ffi.h.
4289
4290 2003-05-07  Jeff Sturm  <jsturm@one-point.com>
4291
4292         Fixes PR bootstrap/10656
4293         * configure.in (HAVE_AS_REGISTER_PSEUDO_OP): Test assembler
4294         support for .register pseudo-op.
4295         * src/sparc/v8.S: Use it.
4296         * fficonfig.h.in: Rebuilt.
4297         * configure: Rebuilt.
4298
4299 2003-04-18  Jakub Jelinek  <jakub@redhat.com>
4300
4301         * include/ffi.h.in (POWERPC64): Define if 64-bit.
4302         (enum ffi_abi): Add FFI_LINUX64 on POWERPC.
4303         Make it the default on POWERPC64.
4304         (FFI_TRAMPOLINE_SIZE): Define to 24 on POWERPC64.
4305         * configure.in: Change powerpc-*-linux* into powerpc*-*-linux*.
4306         * configure: Rebuilt.
4307         * src/powerpc/ffi.c (hidden): Define.
4308         (ffi_prep_args_SYSV): Renamed from
4309         ffi_prep_args.  Cast pointers to unsigned long to shut up warnings.
4310         (NUM_GPR_ARG_REGISTERS64, NUM_FPR_ARG_REGISTERS64,
4311         ASM_NEEDS_REGISTERS64): New.
4312         (ffi_prep_args64): New function.
4313         (ffi_prep_cif_machdep): Handle FFI_LINUX64 ABI.
4314         (ffi_call): Likewise.
4315         (ffi_prep_closure): Likewise.
4316         (flush_icache): Surround by #ifndef POWERPC64.
4317         (ffi_dblfl): New union type.
4318         (ffi_closure_helper_SYSV): Use it to avoid aliasing problems.
4319         (ffi_closure_helper_LINUX64): New function.
4320         * src/powerpc/ppc_closure.S: Surround whole file by #ifndef
4321         __powerpc64__.
4322         * src/powerpc/sysv.S: Likewise.
4323         (ffi_call_SYSV): Rename ffi_prep_args to ffi_prep_args_SYSV.
4324         * src/powerpc/linux64.S: New file.
4325         * src/powerpc/linux64_closure.S: New file.
4326         * Makefile.am (EXTRA_DIST): Add src/powerpc/linux64.S and</