sem_ch4.adb (Analyze_Concatenation_Rest): New procedure.
[gcc.git] / libffi / ChangeLog
1 2007-12-08 David Daney <ddaney@avtrex.com>
2
3 * src/mips/n32.S (ffi_call_N32): Replace dadd with ADDU, dsub with
4 SUBU, add with ADDU and use smaller code sequences.
5
6 2007-12-07 David Daney <ddaney@avtrex.com>
7
8 * src/mips/ffi.c (ffi_prep_cif_machdep): Handle long double return
9 type.
10
11 2007-12-06 David Daney <ddaney@avtrex.com>
12
13 * include/ffi.h.in (FFI_SIZEOF_JAVA_RAW): Define if not already
14 defined.
15 (ffi_java_raw): New typedef.
16 (ffi_java_raw_call, ffi_java_ptrarray_to_raw,
17 ffi_java_raw_to_ptrarray): Change parameter types from ffi_raw to
18 ffi_java_raw.
19 (ffi_java_raw_closure) : Same.
20 (ffi_prep_java_raw_closure, ffi_prep_java_raw_closure_loc): Change
21 parameter types.
22 * src/java_raw_api.c (ffi_java_raw_size): Replace FFI_SIZEOF_ARG with
23 FFI_SIZEOF_JAVA_RAW.
24 (ffi_java_raw_to_ptrarray): Change type of raw to ffi_java_raw.
25 Replace FFI_SIZEOF_ARG with FFI_SIZEOF_JAVA_RAW. Use
26 sizeof(ffi_java_raw) for alignment calculations.
27 (ffi_java_ptrarray_to_raw): Same.
28 (ffi_java_rvalue_to_raw): Add special handling for FFI_TYPE_POINTER
29 if FFI_SIZEOF_JAVA_RAW == 4.
30 (ffi_java_raw_to_rvalue): Same.
31 (ffi_java_raw_call): Change type of raw to ffi_java_raw.
32 (ffi_java_translate_args): Same.
33 (ffi_prep_java_raw_closure_loc, ffi_prep_java_raw_closure): Change
34 parameter types.
35 * src/mips/ffitarget.h (FFI_SIZEOF_JAVA_RAW): Define for N32 ABI.
36
37 2007-12-06 David Daney <ddaney@avtrex.com>
38
39 * src/mips/n32.S (ffi_closure_N32): Use 64-bit add instruction on
40 pointer values.
41
42 2007-12-01 Andreas Tobler <a.tobler@schweiz.org>
43
44 PR libffi/31937
45 * src/powerpc/ffitarget.h: Introduce new ABI FFI_LINUX_SOFT_FLOAT.
46 Add local FFI_TYPE_UINT128 to handle soft-float long-double-128.
47 * src/powerpc/ffi.c: Distinguish between __NO_FPRS__ and not and
48 set the NUM_FPR_ARG_REGISTERS according to.
49 Add support for potential soft-float support under hard-float
50 architecture.
51 (ffi_prep_args_SYSV): Set NUM_FPR_ARG_REGISTERS to 0 in case of
52 FFI_LINUX_SOFT_FLOAT, handle float, doubles and long-doubles according
53 to the FFI_LINUX_SOFT_FLOAT ABI.
54 (ffi_prep_cif_machdep): Likewise.
55 (ffi_closure_helper_SYSV): Likewise.
56 * src/powerpc/ppc_closure.S: Make sure not to store float/double
57 on archs where __NO_FPRS__ is true.
58 Add FFI_TYPE_UINT128 support.
59 * src/powerpc/sysv.S: Add support for soft-float long-double-128.
60 Adjust copyright notice.
61
62 2007-11-25 Andreas Tobler <a.tobler@schweiz.org>
63
64 * src/closures.c: Move defintion of MAYBE_UNUSED from here to ...
65 * include/ffi_common.h: ... here.
66 Update copyright.
67
68 2007-11-17 Andreas Tobler <a.tobler@schweiz.org>
69
70 * src/powerpc/sysv.S: Load correct cr to compare if we have long double.
71 * src/powerpc/linux64.S: Likewise.
72 * src/powerpc/ffi.c: Add a comment to show which part goes into cr6.
73 * testsuite/libffi.call/return_ldl.c: New test.
74
75 2007-09-04 <aph@redhat.com>
76
77 * src/arm/sysv.S (UNWIND): New.
78 (Whole file): Conditionally compile unwinder directives.
79 * src/arm/sysv.S: Add unwinder directives.
80
81 * src/arm/ffi.c (ffi_prep_args): Align structs by at least 4 bytes.
82 Only treat r0 as a struct address if we're actually returning a
83 struct by address.
84 Only copy the bytes that are actually within a struct.
85 (ffi_prep_cif_machdep): A Composite Type not larger than 4 bytes
86 is returned in r0, not passed by address.
87 (ffi_call): Allocate a word-sized temporary for the case where
88 a composite is returned in r0.
89 (ffi_prep_incoming_args_SYSV): Align as necessary.
90
91 2007-08-05 Steven Newbury <s_j_newbury@yahoo.co.uk>
92
93 * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Use __clear_cache instead of
94 directly using the sys_cacheflush syscall.
95
96 2007-07-27 Andrew Haley <aph@redhat.com>
97
98 * src/arm/sysv.S (ffi_closure_SYSV): Add soft-float.
99
100 2007-09-03 Maciej W. Rozycki <macro@linux-mips.org>
101
102 * Makefile.am: Unify MIPS_IRIX and MIPS_LINUX into MIPS.
103 * configure.ac: Likewise.
104 * Makefile.in: Regenerate.
105 * include/Makefile.in: Likewise.
106 * testsuite/Makefile.in: Likewise.
107 * configure: Likewise.
108
109 2007-08-24 David Daney <ddaney@avtrex.com>
110
111 * testsuite/libffi.call/return_sl.c: New test.
112
113 2007-08-10 David Daney <ddaney@avtrex.com>
114
115 * testsuite/libffi.call/cls_multi_ushort.c,
116 testsuite/libffi.call/cls_align_uint16.c,
117 testsuite/libffi.call/nested_struct1.c,
118 testsuite/libffi.call/nested_struct3.c,
119 testsuite/libffi.call/cls_7_1_byte.c,
120 testsuite/libffi.call/nested_struct5.c,
121 testsuite/libffi.call/cls_double.c,
122 testsuite/libffi.call/nested_struct7.c,
123 testsuite/libffi.call/cls_sint.c,
124 testsuite/libffi.call/nested_struct9.c,
125 testsuite/libffi.call/cls_20byte1.c,
126 testsuite/libffi.call/cls_multi_sshortchar.c,
127 testsuite/libffi.call/cls_align_sint64.c,
128 testsuite/libffi.call/cls_3byte2.c,
129 testsuite/libffi.call/cls_multi_schar.c,
130 testsuite/libffi.call/cls_multi_uchar.c,
131 testsuite/libffi.call/cls_19byte.c,
132 testsuite/libffi.call/cls_9byte1.c,
133 testsuite/libffi.call/cls_align_float.c,
134 testsuite/libffi.call/closure_fn1.c,
135 testsuite/libffi.call/problem1.c,
136 testsuite/libffi.call/closure_fn3.c,
137 testsuite/libffi.call/cls_sshort.c,
138 testsuite/libffi.call/closure_fn5.c,
139 testsuite/libffi.call/cls_align_double.c,
140 testsuite/libffi.call/nested_struct.c,
141 testsuite/libffi.call/cls_2byte.c,
142 testsuite/libffi.call/nested_struct10.c,
143 testsuite/libffi.call/cls_4byte.c,
144 testsuite/libffi.call/cls_6byte.c,
145 testsuite/libffi.call/cls_8byte.c,
146 testsuite/libffi.call/cls_multi_sshort.c,
147 testsuite/libffi.call/cls_align_sint16.c,
148 testsuite/libffi.call/cls_align_uint32.c,
149 testsuite/libffi.call/cls_20byte.c,
150 testsuite/libffi.call/cls_float.c,
151 testsuite/libffi.call/nested_struct2.c,
152 testsuite/libffi.call/cls_5_1_byte.c,
153 testsuite/libffi.call/nested_struct4.c,
154 testsuite/libffi.call/cls_24byte.c,
155 testsuite/libffi.call/nested_struct6.c,
156 testsuite/libffi.call/cls_64byte.c,
157 testsuite/libffi.call/nested_struct8.c,
158 testsuite/libffi.call/cls_uint.c,
159 testsuite/libffi.call/cls_multi_ushortchar.c,
160 testsuite/libffi.call/cls_schar.c,
161 testsuite/libffi.call/cls_uchar.c,
162 testsuite/libffi.call/cls_align_uint64.c,
163 testsuite/libffi.call/cls_ulonglong.c,
164 testsuite/libffi.call/cls_align_longdouble.c,
165 testsuite/libffi.call/cls_1_1byte.c,
166 testsuite/libffi.call/cls_12byte.c,
167 testsuite/libffi.call/cls_3_1byte.c,
168 testsuite/libffi.call/cls_3byte1.c,
169 testsuite/libffi.call/cls_4_1byte.c,
170 testsuite/libffi.call/cls_6_1_byte.c,
171 testsuite/libffi.call/cls_16byte.c,
172 testsuite/libffi.call/cls_18byte.c,
173 testsuite/libffi.call/closure_fn0.c,
174 testsuite/libffi.call/cls_9byte2.c,
175 testsuite/libffi.call/closure_fn2.c,
176 testsuite/libffi.call/closure_fn4.c,
177 testsuite/libffi.call/cls_ushort.c,
178 testsuite/libffi.call/closure_fn6.c,
179 testsuite/libffi.call/cls_5byte.c,
180 testsuite/libffi.call/cls_align_pointer.c,
181 testsuite/libffi.call/cls_7byte.c,
182 testsuite/libffi.call/cls_align_sint32.c,
183 testsuite/libffi.special/unwindtest_ffi_call.cc,
184 testsuite/libffi.special/unwindtest.cc: Remove xfail for mips64*-*-*.
185
186 2007-08-10 David Daney <ddaney@avtrex.com>
187
188 PR libffi/28313
189 * configure.ac: Don't treat mips64 as a special case.
190 * Makefile.am (nodist_libffi_la_SOURCES): Add n32.S.
191 * configure: Regenerate
192 * Makefile.in: Ditto.
193 * fficonfig.h.in: Ditto.
194 * src/mips/ffitarget.h (REG_L, REG_S, SUBU, ADDU, SRL, LI): Indent.
195 (LA, EH_FRAME_ALIGN, FDE_ADDR_BYTES): New preprocessor macros.
196 (FFI_DEFAULT_ABI): Set for n64 case.
197 (FFI_CLOSURES, FFI_TRAMPOLINE_SIZE): Define for n32 and n64 cases.
198 * src/mips/n32.S (ffi_call_N32): Add debug macros and labels for FDE.
199 (ffi_closure_N32): New function.
200 (.eh_frame): New section
201 * src/mips/o32.S: Clean up comments.
202 (ffi_closure_O32): Pass ffi_closure parameter in $12.
203 * src/mips/ffi.c: Use FFI_MIPS_N32 instead of
204 _MIPS_SIM == _ABIN32 throughout.
205 (FFI_MIPS_STOP_HERE): New, use in place of
206 ffi_stop_here.
207 (ffi_prep_args): Use unsigned long to hold pointer values. Rewrite
208 to support n32/n64 ABIs.
209 (calc_n32_struct_flags): Rewrite.
210 (calc_n32_return_struct_flags): Remove unused variable. Reverse
211 position of flag bits.
212 (ffi_prep_cif_machdep): Rewrite n32 portion.
213 (ffi_call): Enable for n64. Add special handling for small structure
214 return values.
215 (ffi_prep_closure_loc): Add n32 and n64 support.
216 (ffi_closure_mips_inner_O32): Add cast to silence warning.
217 (copy_struct_N32, ffi_closure_mips_inner_N32): New functions.
218
219 2007-08-08 David Daney <ddaney@avtrex.com>
220
221 * testsuite/libffi.call/ffitest.h (ffi_type_mylong): Remove definition.
222 * testsuite/libffi.call/cls_align_uint16.c (main): Use correct type
223 specifiers.
224 * testsuite/libffi.call/nested_struct1.c (main): Ditto.
225 * testsuite/libffi.call/cls_sint.c (main): Ditto.
226 * testsuite/libffi.call/nested_struct9.c (main): Ditto.
227 * testsuite/libffi.call/cls_20byte1.c (main): Ditto.
228 * testsuite/libffi.call/cls_9byte1.c (main): Ditto.
229 * testsuite/libffi.call/closure_fn1.c (main): Ditto.
230 * testsuite/libffi.call/closure_fn3.c (main): Ditto.
231 * testsuite/libffi.call/return_dbl2.c (main): Ditto.
232 * testsuite/libffi.call/cls_sshort.c (main): Ditto.
233 * testsuite/libffi.call/return_fl3.c (main): Ditto.
234 * testsuite/libffi.call/closure_fn5.c (main): Ditto.
235 * testsuite/libffi.call/nested_struct.c (main): Ditto.
236 * testsuite/libffi.call/nested_struct10.c (main): Ditto.
237 * testsuite/libffi.call/return_ll1.c (main): Ditto.
238 * testsuite/libffi.call/cls_8byte.c (main): Ditto.
239 * testsuite/libffi.call/cls_align_uint32.c (main): Ditto.
240 * testsuite/libffi.call/cls_align_sint16.c (main): Ditto.
241 * testsuite/libffi.call/cls_20byte.c (main): Ditto.
242 * testsuite/libffi.call/nested_struct2.c (main): Ditto.
243 * testsuite/libffi.call/cls_24byte.c (main): Ditto.
244 * testsuite/libffi.call/nested_struct6.c (main): Ditto.
245 * testsuite/libffi.call/cls_uint.c (main): Ditto.
246 * testsuite/libffi.call/cls_12byte.c (main): Ditto.
247 * testsuite/libffi.call/cls_16byte.c (main): Ditto.
248 * testsuite/libffi.call/closure_fn0.c (main): Ditto.
249 * testsuite/libffi.call/cls_9byte2.c (main): Ditto.
250 * testsuite/libffi.call/closure_fn2.c (main): Ditto.
251 * testsuite/libffi.call/return_dbl1.c (main): Ditto.
252 * testsuite/libffi.call/closure_fn4.c (main): Ditto.
253 * testsuite/libffi.call/closure_fn6.c (main): Ditto.
254 * testsuite/libffi.call/cls_align_sint32.c (main): Ditto.
255
256 2007-08-07 Andrew Haley <aph@redhat.com>
257
258 * src/x86/sysv.S (ffi_closure_raw_SYSV): Fix typo in previous
259 checkin.
260
261 2007-08-06 Andrew Haley <aph@redhat.com>
262
263 PR testsuite/32843
264 * src/x86/sysv.S (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
265 FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
266 FFI_TYPE_SINT32.
267
268 2007-08-02 David Daney <ddaney@avtrex.com>
269
270 * testsuite/libffi.call/return_ul.c (main): Define return type as
271 ffi_arg. Use proper printf conversion specifier.
272
273 2007-07-30 Andrew Haley <aph@redhat.com>
274
275 PR testsuite/32843
276 * src/x86/ffi.c (ffi_prep_cif_machdep): in x86 case, add code for
277 signed/unsigned int8/16.
278 * src/x86/sysv.S (ffi_call_SYSV): Rewrite to:
279 Use a jump table.
280 Remove code to pop args from the stack after call.
281 Special-case signed/unsigned int8/16.
282 * testsuite/libffi.call/return_sc.c (main): Revert.
283
284 2007-07-26 Richard Guenther <rguenther@suse.de>
285
286 PR testsuite/32843
287 * testsuite/libffi.call/return_sc.c (main): Verify call
288 result as signed char, not ffi_arg.
289
290 2007-07-16 Rainer Orth <ro@TechFak.Uni-Bielefeld.DE>
291
292 * configure.ac (i?86-*-solaris2.1[0-9]): Set TARGET to X86_64.
293 * configure: Regenerate.
294
295 2007-07-11 David Daney <ddaney@avtrex.com>
296
297 * src/mips/ffi.c: Don't include sys/cachectl.h.
298 (ffi_prep_closure_loc): Use __builtin___clear_cache() instead of
299 cacheflush().
300
301 2007-05-18 Aurelien Jarno <aurelien@aurel32.net>
302
303 * src/arm/ffi.c (ffi_prep_closure_loc): Renamed and ajusted
304 from (ffi_prep_closure): ... this.
305 (FFI_INIT_TRAMPOLINE): Adjust.
306
307 2005-12-31 Phil Blundell <pb@reciva.com>
308
309 * src/arm/ffi.c (ffi_prep_incoming_args_SYSV,
310 ffi_closure_SYSV_inner, ffi_prep_closure): New, add closure support.
311 * src/arm/sysv.S(ffi_closure_SYSV): Likewise.
312 * src/arm/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
313 (FFI_CLOSURES): Enable closure support.
314
315 2007-07-03 Andrew Haley <aph@hedges.billgatliff.com>
316
317 * testsuite/libffi.call/cls_multi_ushort.c,
318 testsuite/libffi.call/cls_align_uint16.c,
319 testsuite/libffi.call/nested_struct1.c,
320 testsuite/libffi.call/nested_struct3.c,
321 testsuite/libffi.call/cls_7_1_byte.c,
322 testsuite/libffi.call/cls_double.c,
323 testsuite/libffi.call/nested_struct5.c,
324 testsuite/libffi.call/nested_struct7.c,
325 testsuite/libffi.call/cls_sint.c,
326 testsuite/libffi.call/nested_struct9.c,
327 testsuite/libffi.call/cls_20byte1.c,
328 testsuite/libffi.call/cls_multi_sshortchar.c,
329 testsuite/libffi.call/cls_align_sint64.c,
330 testsuite/libffi.call/cls_3byte2.c,
331 testsuite/libffi.call/cls_multi_schar.c,
332 testsuite/libffi.call/cls_multi_uchar.c,
333 testsuite/libffi.call/cls_19byte.c,
334 testsuite/libffi.call/cls_9byte1.c,
335 testsuite/libffi.call/cls_align_float.c,
336 testsuite/libffi.call/closure_fn1.c,
337 testsuite/libffi.call/problem1.c,
338 testsuite/libffi.call/closure_fn3.c,
339 testsuite/libffi.call/cls_sshort.c,
340 testsuite/libffi.call/closure_fn5.c,
341 testsuite/libffi.call/cls_align_double.c,
342 testsuite/libffi.call/cls_2byte.c,
343 testsuite/libffi.call/nested_struct.c,
344 testsuite/libffi.call/nested_struct10.c,
345 testsuite/libffi.call/cls_4byte.c,
346 testsuite/libffi.call/cls_6byte.c,
347 testsuite/libffi.call/cls_8byte.c,
348 testsuite/libffi.call/cls_multi_sshort.c,
349 testsuite/libffi.call/cls_align_uint32.c,
350 testsuite/libffi.call/cls_align_sint16.c,
351 testsuite/libffi.call/cls_float.c,
352 testsuite/libffi.call/cls_20byte.c,
353 testsuite/libffi.call/cls_5_1_byte.c,
354 testsuite/libffi.call/nested_struct2.c,
355 testsuite/libffi.call/cls_24byte.c,
356 testsuite/libffi.call/nested_struct4.c,
357 testsuite/libffi.call/nested_struct6.c,
358 testsuite/libffi.call/cls_64byte.c,
359 testsuite/libffi.call/nested_struct8.c,
360 testsuite/libffi.call/cls_uint.c,
361 testsuite/libffi.call/cls_multi_ushortchar.c,
362 testsuite/libffi.call/cls_schar.c,
363 testsuite/libffi.call/cls_uchar.c,
364 testsuite/libffi.call/cls_align_uint64.c,
365 testsuite/libffi.call/cls_ulonglong.c,
366 testsuite/libffi.call/cls_align_longdouble.c,
367 testsuite/libffi.call/cls_1_1byte.c,
368 testsuite/libffi.call/cls_12byte.c,
369 testsuite/libffi.call/cls_3_1byte.c,
370 testsuite/libffi.call/cls_3byte1.c,
371 testsuite/libffi.call/cls_4_1byte.c,
372 testsuite/libffi.call/cls_6_1_byte.c,
373 testsuite/libffi.call/cls_16byte.c,
374 testsuite/libffi.call/cls_18byte.c,
375 testsuite/libffi.call/closure_fn0.c,
376 testsuite/libffi.call/cls_9byte2.c,
377 testsuite/libffi.call/closure_fn2.c,
378 testsuite/libffi.call/closure_fn4.c,
379 testsuite/libffi.call/cls_ushort.c,
380 testsuite/libffi.call/closure_fn6.c,
381 testsuite/libffi.call/cls_5byte.c,
382 testsuite/libffi.call/cls_align_pointer.c,
383 testsuite/libffi.call/cls_7byte.c,
384 testsuite/libffi.call/cls_align_sint32.c,
385 testsuite/libffi.special/unwindtest_ffi_call.cc,
386 testsuite/libffi.special/unwindtest.cc: Enable for ARM.
387
388 2007-07-05 H.J. Lu <hongjiu.lu@intel.com>
389
390 * aclocal.m4: Regenerated.
391
392 2007-06-02 Paolo Bonzini <bonzini@gnu.org>
393
394 * configure: Regenerate.
395
396 2007-05-23 Steve Ellcey <sje@cup.hp.com>
397
398 * Makefile.in: Regenerate.
399 * configure: Regenerate.
400 * aclocal.m4: Regenerate.
401 * include/Makefile.in: Regenerate.
402 * testsuite/Makefile.in: Regenerate.
403
404 2007-05-10 Roman Zippel <zippel@linux-m68k.org>
405
406 * src/m68k/ffi.c (ffi_prep_incoming_args_SYSV,
407 ffi_closure_SYSV_inner,ffi_prep_closure): New, add closure support.
408 * src/m68k/sysv.S(ffi_closure_SYSV,ffi_closure_struct_SYSV): Likewise.
409 * src/m68k/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise.
410 (FFI_CLOSURES): Enable closure support.
411
412 2007-05-10 Roman Zippel <zippel@linux-m68k.org>
413
414 * configure.ac (HAVE_AS_CFI_PSEUDO_OP): New test.
415 * configure: Regenerate.
416 * fficonfig.h.in: Regenerate.
417 * src/m68k/sysv.S (CFI_STARTPROC,CFI_ENDPROC,
418 CFI_OFFSET,CFI_DEF_CFA): New macros.
419 (ffi_call_SYSV): Add callframe annotation.
420
421 2007-05-10 Roman Zippel <zippel@linux-m68k.org>
422
423 * src/m68k/ffi.c (ffi_prep_args,ffi_prep_cif_machdep): Fix
424 numerous test suite failures.
425 * src/m68k/sysv.S (ffi_call_SYSV): Likewise.
426
427 2007-04-11 Paolo Bonzini <bonzini@gnu.org>
428
429 * Makefile.am (EXTRA_DIST): Bring up to date.
430 * Makefile.in: Regenerate.
431 * src/frv/eabi.S: Remove RCS keyword.
432
433 2007-04-06 Richard Henderson <rth@redhat.com>
434
435 * configure.ac: Tidy target case.
436 (HAVE_LONG_DOUBLE): Allow the target to override.
437 * configure: Regenerate.
438 * include/ffi.h.in: Don't define ffi_type_foo if
439 LIBFFI_HIDE_BASIC_TYPES is defined.
440 (ffi_type_longdouble): If not HAVE_LONG_DOUBLE, define
441 to ffi_type_double.
442 * types.c (LIBFFI_HIDE_BASIC_TYPES): Define.
443 (FFI_TYPEDEF, ffi_type_void): Mark the data const.
444 (ffi_type_longdouble): Special case for Alpha. Don't define
445 if long double == double.
446
447 * src/alpha/ffi.c (FFI_TYPE_LONGDOUBLE): Assert unique value.
448 (ffi_prep_cif_machdep): Handle it as the 128-bit type.
449 (ffi_call, ffi_closure_osf_inner): Likewise.
450 (ffi_closure_osf_inner): Likewise. Mark hidden.
451 (ffi_call_osf, ffi_closure_osf): Mark hidden.
452 * src/alpha/ffitarget.h (FFI_LAST_ABI): Tidy definition.
453 * src/alpha/osf.S (ffi_call_osf, ffi_closure_osf): Mark hidden.
454 (load_table): Handle 128-bit long double.
455
456 * testsuite/libffi.call/float4.c: Add -mieee for alpha.
457
458 2007-04-06 Tom Tromey <tromey@redhat.com>
459
460 PR libffi/31491:
461 * README: Fixed bug in example.
462
463 2007-04-03 Jakub Jelinek <jakub@redhat.com>
464
465 * src/closures.c: Include sys/statfs.h.
466 (_GNU_SOURCE): Define on Linux.
467 (FFI_MMAP_EXEC_SELINUX): Define.
468 (selinux_enabled): New variable.
469 (selinux_enabled_check): New function.
470 (is_selinux_enabled): Define.
471 (dlmmap): Use it.
472
473 2007-03-24 Uros Bizjak <ubizjak@gmail.com>
474
475 * testsuite/libffi.call/return_fl2.c (return_fl): Mark as static.
476 Use 'volatile float sum' to create sum of floats to avoid false
477 negative due to excess precision on ix86 targets.
478 (main): Ditto.
479
480 2007-03-08 Alexandre Oliva <aoliva@redhat.com>
481
482 * src/powerpc/ffi.c (flush_icache): Fix left-over from previous
483 patch.
484 (ffi_prep_closure_loc): Remove unneeded casts. Add needed ones.
485
486 2007-03-07 Alexandre Oliva <aoliva@redhat.com>
487
488 * include/ffi.h.in (ffi_closure_alloc, ffi_closure_free): New.
489 (ffi_prep_closure_loc): New.
490 (ffi_prep_raw_closure_loc): New.
491 (ffi_prep_java_raw_closure_loc): New.
492 * src/closures.c: New file.
493 * src/dlmalloc.c [FFI_MMAP_EXEC_WRIT] (struct malloc_segment):
494 Replace sflags with exec_offset.
495 [FFI_MMAP_EXEC_WRIT] (mmap_exec_offset, add_segment_exec_offset,
496 sub_segment_exec_offset): New macros.
497 (get_segment_flags, set_segment_flags, check_segment_merge): New
498 macros.
499 (is_mmapped_segment, is_extern_segment): Use get_segment_flags.
500 (add_segment, sys_alloc, create_mspace, create_mspace_with_base,
501 destroy_mspace): Use new macros.
502 (sys_alloc): Silence warning.
503 * Makefile.am (libffi_la_SOURCES): Add src/closures.c.
504 * Makefile.in: Rebuilt.
505 * src/prep_cif [FFI_CLOSURES] (ffi_prep_closure): Implement in
506 terms of ffi_prep_closure_loc.
507 * src/raw_api.c (ffi_prep_raw_closure_loc): Renamed and adjusted
508 from...
509 (ffi_prep_raw_closure): ... this. Re-implement in terms of the
510 renamed version.
511 * src/java_raw_api (ffi_prep_java_raw_closure_loc): Renamed and
512 adjusted from...
513 (ffi_prep_java_raw_closure): ... this. Re-implement in terms of
514 the renamed version.
515 * src/alpha/ffi.c (ffi_prep_closure_loc): Renamed from
516 (ffi_prep_closure): ... this.
517 * src/pa/ffi.c: Likewise.
518 * src/cris/ffi.c: Likewise. Adjust.
519 * src/frv/ffi.c: Likewise.
520 * src/ia64/ffi.c: Likewise.
521 * src/mips/ffi.c: Likewise.
522 * src/powerpc/ffi_darwin.c: Likewise.
523 * src/s390/ffi.c: Likewise.
524 * src/sh/ffi.c: Likewise.
525 * src/sh64/ffi.c: Likewise.
526 * src/sparc/ffi.c: Likewise.
527 * src/x86/ffi64.c: Likewise.
528 * src/x86/ffi.c: Likewise.
529 (FFI_INIT_TRAMPOLINE): Adjust.
530 (ffi_prep_raw_closure_loc): Renamed and adjusted from...
531 (ffi_prep_raw_closure): ... this.
532 * src/powerpc/ffi.c (ffi_prep_closure_loc): Renamed from
533 (ffi_prep_closure): ... this.
534 (flush_icache): Adjust.
535
536 2007-03-07 Alexandre Oliva <aoliva@redhat.com>
537
538 * src/dlmalloc.c: New file, imported version 2.8.3 of Doug
539 Lea's malloc.
540
541 2007-03-01 Brooks Moses <brooks.moses@codesourcery.com>
542
543 * Makefile.am: Add dummy install-pdf target.
544 * Makefile.in: Regenerate
545
546 2007-02-13 Andreas Krebbel <krebbel1@de.ibm.com>
547
548 * src/s390/ffi.c (ffi_prep_args, ffi_prep_cif_machdep,
549 ffi_closure_helper_SYSV): Add long double handling.
550
551 2007-02-02 Jakub Jelinek <jakub@redhat.com>
552
553 * src/powerpc/linux64.S (ffi_call_LINUX64): Move restore of r2
554 immediately after bctrl instruction.
555
556 2007-01-18 Alexandre Oliva <aoliva@redhat.com>
557
558 * Makefile.am (all-recursive, install-recursive,
559 mostlyclean-recursive, clean-recursive, distclean-recursive,
560 maintainer-clean-recursive): Add missing targets.
561 * Makefile.in: Rebuilt.
562
563 2006-12-14 Andreas Tobler <a.tobler@schweiz.org>
564
565 * configure.ac: Add TARGET for x86_64-*-darwin*.
566 * Makefile.am (nodist_libffi_la_SOURCES): Add rules for 64-bit sources
567 for X86_DARWIN.
568 * src/x86/ffitarget.h: Set trampoline size for x86_64-*-darwin*.
569 * src/x86/darwin64.S: New file for x86_64-*-darwin* support.
570 * configure: Regenerate.
571 * Makefile.in: Regenerate.
572 * include/Makefile.in: Regenerate.
573 * testsuite/Makefile.in: Regenerate.
574 * testsuite/libffi.special/unwindtest_ffi_call.cc: New test case for
575 ffi_call only.
576
577 2006-12-13 Andreas Tobler <a.tobler@schweiz.org>
578
579 * aclocal.m4: Regenerate with aclocal -I .. as written in the
580 Makefile.am.
581
582 2006-10-31 Geoffrey Keating <geoffk@apple.com>
583
584 * src/powerpc/ffi_darwin.c (darwin_adjust_aggregate_sizes): New.
585 (ffi_prep_cif_machdep): Call darwin_adjust_aggregate_sizes for
586 Darwin.
587 * testsuite/libffi.call/nested_struct4.c: Remove Darwin XFAIL.
588 * testsuite/libffi.call/nested_struct6.c: Remove Darwin XFAIL.
589
590 2006-10-10 Paolo Bonzini <bonzini@gnu.org>
591 Sandro Tolaini <tolaini@libero.it>
592
593 * configure.ac [i*86-*-darwin*]: Set X86_DARWIN symbol and
594 conditional.
595 * configure: Regenerated.
596 * Makefile.am (nodist_libffi_la_SOURCES) [X86_DARWIN]: New case.
597 (EXTRA_DIST): Add src/x86/darwin.S.
598 * Makefile.in: Regenerated.
599 * include/Makefile.in: Regenerated.
600 * testsuite/Makefile.in: Regenerated.
601
602 * src/x86/ffi.c (ffi_prep_cif_machdep) [X86_DARWIN]: Treat like
603 X86_WIN32, and additionally align stack to 16 bytes.
604 * src/x86/darwin.S: New, based on sysv.S.
605 * src/prep_cif.c (ffi_prep_cif) [X86_DARWIN]: Align > 8-byte structs.
606
607 2006-09-12 David Daney <ddaney@avtrex.com>
608
609 PR libffi/23935
610 * include/Makefile.am: Install both ffi.h and ffitarget.h in
611 $(libdir)/gcc/$(target_alias)/$(gcc_version)/include.
612 * aclocal.m4: Regenerated for automake 1.9.6.
613 * Makefile.in: Regenerated.
614 * include/Makefile.in: Regenerated.
615 * testsuite/Makefile.in: Regenerated.
616
617 2006-08-17 Andreas Tobler <a.tobler@schweiz.ch>
618
619 * include/ffi_common.h (struct): Revert accidental commit.
620
621 2006-08-15 Andreas Tobler <a.tobler@schweiz.ch>
622
623 * include/ffi_common.h: Remove lint directives.
624 * include/ffi.h.in: Likewise.
625
626 2006-07-25 Torsten Schoenfeld <kaffeetisch@gmx.de>
627
628 * include/ffi.h.in (ffi_type_ulong, ffi_type_slong): Define correctly
629 for 32-bit architectures.
630 * testsuite/libffi.call/return_ul.c: New test case.
631
632 2006-07-19 David Daney <ddaney@avtrex.com>
633
634 * testsuite/libffi.call/closure_fn6.c: Remove xfail for mips,
635 xfail remains for mips64.
636
637 2006-05-23 Carlos O'Donell <carlos@codesourcery.com>
638
639 * Makefile.am: Add install-html target. Add install-html to .PHONY
640 * Makefile.in: Regenerate.
641 * aclocal.m4: Regenerate.
642 * include/Makefile.in: Regenerate.
643 * testsuite/Makefile.in: Regenerate.
644
645 2006-05-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
646
647 * pa/ffi.c (ffi_prep_args_pa32): Load floating point arguments from
648 stack slot.
649
650 2006-04-22 Andreas Tobler <a.tobler@schweiz.ch>
651
652 * README: Remove notice about 'Crazy Comments'.
653 * src/debug.c: Remove lint directives. Cleanup white spaces.
654 * src/java_raw_api.c: Likewise.
655 * src/prep_cif.c: Likewise.
656 * src/raw_api.c: Likewise.
657 * src/ffitest.c: Delete. No longer needed, all test cases migrated
658 to the testsuite.
659 * src/arm/ffi.c: Remove lint directives.
660 * src/m32r/ffi.c: Likewise.
661 * src/pa/ffi.c: Likewise.
662 * src/powerpc/ffi.c: Likewise.
663 * src/powerpc/ffi_darwin.c: Likewise.
664 * src/sh/ffi.c: Likewise.
665 * src/sh64/ffi.c: Likewise.
666 * src/x86/ffi.c: Likewise.
667 * testsuite/libffi.call/float2.c: Likewise.
668 * testsuite/libffi.call/promotion.c: Likewise.
669 * testsuite/libffi.call/struct1.c: Likewise.
670
671 2006-04-13 Andreas Tobler <a.tobler@schweiz.ch>
672
673 * src/pa/hpux32.S: Correct unwind offset calculation for
674 ffi_closure_pa32.
675 * src/pa/linux.S: Likewise.
676
677 2006-04-12 James E Wilson <wilson@specifix.com>
678
679 PR libgcj/26483
680 * src/ia64/ffi.c (stf_spill, ldf_fill): Rewrite as macros.
681 (hfa_type_load): Call stf_spill.
682 (hfa_type_store): Call ldf_fill.
683 (ffi_call): Adjust calls to above routines. Add local temps for
684 macro result.
685
686 2006-04-10 Matthias Klose <doko@debian.org>
687
688 * testsuite/lib/libffi-dg.exp (libffi-init): Recognize multilib
689 directory names containing underscores.
690
691 2006-04-07 James E Wilson <wilson@specifix.com>
692
693 * testsuite/libffi.call/float4.c: New testcase.
694
695 2006-04-05 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
696 Andreas Tobler <a.tobler@schweiz.ch>
697
698 * Makefile.am: Add PA_HPUX port.
699 * Makefile.in: Regenerate.
700 * include/Makefile.in: Likewise.
701 * testsuite/Makefile.in: Likewise.
702 * configure.ac: Add PA_HPUX rules.
703 * configure: Regenerate.
704 * src/pa/ffitarget.h: Rename linux target to PA_LINUX.
705 Add PA_HPUX and PA64_HPUX.
706 Rename FFI_LINUX ABI to FFI_PA32 ABI.
707 (FFI_TRAMPOLINE_SIZE): Define for 32-bit HP-UX targets.
708 (FFI_TYPE_SMALL_STRUCT2): Define.
709 (FFI_TYPE_SMALL_STRUCT4): Likewise.
710 (FFI_TYPE_SMALL_STRUCT8): Likewise.
711 (FFI_TYPE_SMALL_STRUCT3): Redefine.
712 (FFI_TYPE_SMALL_STRUCT5): Likewise.
713 (FFI_TYPE_SMALL_STRUCT6): Likewise.
714 (FFI_TYPE_SMALL_STRUCT7): Likewise.
715 * src/pa/ffi.c (ROUND_DOWN): Delete.
716 (fldw, fstw, fldd, fstd): Use '__asm__'.
717 (ffi_struct_type): Add support for FFI_TYPE_SMALL_STRUCT2,
718 FFI_TYPE_SMALL_STRUCT4 and FFI_TYPE_SMALL_STRUCT8.
719 (ffi_prep_args_LINUX): Rename to ffi_prep_args_pa32. Update comment.
720 Simplify incrementing of stack slot variable. Change type of local
721 'n' to unsigned int.
722 (ffi_size_stack_LINUX): Rename to ffi_size_stack_pa32. Handle long
723 double on PA_HPUX.
724 (ffi_prep_cif_machdep): Likewise.
725 (ffi_call): Likewise.
726 (ffi_closure_inner_LINUX): Rename to ffi_closure_inner_pa32. Change
727 return type to ffi_status. Simplify incrementing of stack slot
728 variable. Only copy floating point argument registers when PA_LINUX
729 is true. Reformat debug statement.
730 Add support for FFI_TYPE_SMALL_STRUCT2, FFI_TYPE_SMALL_STRUCT4 and
731 FFI_TYPE_SMALL_STRUCT8.
732 (ffi_closure_LINUX): Rename to ffi_closure_pa32. Add 'extern' to
733 declaration.
734 (ffi_prep_closure): Make linux trampoline conditional on PA_LINUX.
735 Add nops to cache flush. Add trampoline for PA_HPUX.
736 * src/pa/hpux32.S: New file.
737 * src/pa/linux.S (ffi_call_LINUX): Rename to ffi_call_pa32. Rename
738 ffi_prep_args_LINUX to ffi_prep_args_pa32.
739 Localize labels. Add support for 2, 4 and 8-byte small structs. Handle
740 unaligned destinations in 3, 5, 6 and 7-byte small structs. Order
741 argument type checks so that common argument types appear first.
742 (ffi_closure_LINUX): Rename to ffi_closure_pa32. Rename
743 ffi_closure_inner_LINUX to ffi_closure_inner_pa32.
744
745 2006-03-24 Alan Modra <amodra@bigpond.net.au>
746
747 * src/powerpc/ffitarget.h (enum ffi_abi): Add FFI_LINUX. Default
748 for 32-bit using IBM extended double format. Fix FFI_LAST_ABI.
749 * src/powerpc/ffi.c (ffi_prep_args_SYSV): Handle linux variant of
750 FFI_TYPE_LONGDOUBLE.
751 (ffi_prep_args64): Assert using IBM extended double.
752 (ffi_prep_cif_machdep): Don't munge FFI_TYPE_LONGDOUBLE type.
753 Handle FFI_LINUX FFI_TYPE_LONGDOUBLE return and args.
754 (ffi_call): Handle FFI_LINUX.
755 (ffi_closure_helper_SYSV): Non FFI_LINUX long double return needs
756 gpr3 return pointer as for struct return. Handle FFI_LINUX
757 FFI_TYPE_LONGDOUBLE return and args. Don't increment "nf"
758 unnecessarily.
759 * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Load both f1 and f2
760 for FFI_TYPE_LONGDOUBLE. Move epilogue insns into case table.
761 Don't use r6 as pointer to results, instead use sp offset. Don't
762 make a special call to load lr with case table address, instead
763 use offset from previous call.
764 * src/powerpc/sysv.S (ffi_call_SYSV): Save long double return.
765 * src/powerpc/linux64.S (ffi_call_LINUX64): Simplify long double
766 return.
767
768 2006-03-15 Kaz Kojima <kkojima@gcc.gnu.org>
769
770 * src/sh64/ffi.c (ffi_prep_cif_machdep): Handle float arguments
771 passed with FP registers correctly.
772 (ffi_closure_helper_SYSV): Likewise.
773 * src/sh64/sysv.S: Likewise.
774
775 2006-03-01 Andreas Tobler <a.tobler@schweiz.ch>
776
777 * testsuite/libffi.special/unwindtest.cc (closure_test_fn): Mark cif,
778 args and userdata unused.
779 (closure_test_fn1): Mark cif and userdata unused.
780 (main): Remove unused res.
781
782 2006-02-28 Andreas Tobler <a.tobler@schweiz.ch>
783
784 * testsuite/libffi.call/call.exp: Adjust FSF address. Add test runs for
785 -O2, -O3, -Os and the warning flags -W -Wall.
786 * testsuite/libffi.special/special.exp: Likewise.
787 * testsuite/libffi.call/ffitest.h: Add an __UNUSED__ macro to mark
788 unused parameter unused for gcc or else do nothing.
789 * testsuite/libffi.special/ffitestcxx.h: Likewise.
790 * testsuite/libffi.call/cls_12byte.c (cls_struct_12byte_gn): Mark cif
791 and userdata unused.
792 * testsuite/libffi.call/cls_16byte.c (cls_struct_16byte_gn): Likewise.
793 * testsuite/libffi.call/cls_18byte.c (cls_struct_18byte_gn): Likewise.
794 * testsuite/libffi.call/cls_19byte.c (cls_struct_19byte_gn): Likewise.
795 * testsuite/libffi.call/cls_1_1byte.c (cls_struct_1_1byte_gn): Likewise.
796 * testsuite/libffi.call/cls_20byte.c (cls_struct_20byte_gn): Likewise.
797 * testsuite/libffi.call/cls_20byte1.c (cls_struct_20byte_gn): Likewise.
798 * testsuite/libffi.call/cls_24byte.c (cls_struct_24byte_gn): Likewise.
799 * testsuite/libffi.call/cls_2byte.c (cls_struct_2byte_gn): Likewise.
800 * testsuite/libffi.call/cls_3_1byte.c (cls_struct_3_1byte_gn): Likewise.
801 * testsuite/libffi.call/cls_3byte1.c (cls_struct_3byte_gn): Likewise.
802 * testsuite/libffi.call/cls_3byte2.c (cls_struct_3byte_gn1): Likewise.
803 * testsuite/libffi.call/cls_4_1byte.c (cls_struct_4_1byte_gn): Likewise.
804 * testsuite/libffi.call/cls_4byte.c (cls_struct_4byte_gn): Likewise.
805 * testsuite/libffi.call/cls_5_1_byte.c (cls_struct_5byte_gn): Likewise.
806 * testsuite/libffi.call/cls_5byte.c (cls_struct_5byte_gn): Likewise.
807 * testsuite/libffi.call/cls_64byte.c (cls_struct_64byte_gn): Likewise.
808 * testsuite/libffi.call/cls_6_1_byte.c (cls_struct_6byte_gn): Likewise.
809 * testsuite/libffi.call/cls_6byte.c (cls_struct_6byte_gn): Likewise.
810 * testsuite/libffi.call/cls_7_1_byte.c (cls_struct_7byte_gn): Likewise.
811 * testsuite/libffi.call/cls_7byte.c (cls_struct_7byte_gn): Likewise.
812 * testsuite/libffi.call/cls_8byte.c (cls_struct_8byte_gn): Likewise.
813 * testsuite/libffi.call/cls_9byte1.c (cls_struct_9byte_gn): Likewise.
814 * testsuite/libffi.call/cls_9byte2.c (cls_struct_9byte_gn): Likewise.
815 * testsuite/libffi.call/cls_align_double.c (cls_struct_align_gn):
816 Likewise.
817 * testsuite/libffi.call/cls_align_float.c (cls_struct_align_gn):
818 Likewise.
819 * testsuite/libffi.call/cls_align_longdouble.c (cls_struct_align_gn):
820 Likewise.
821 * testsuite/libffi.call/cls_align_pointer.c (cls_struct_align_fn): Cast
822 void* to avoid compiler warning.
823 (main): Likewise.
824 (cls_struct_align_gn): Mark cif and userdata unused.
825 * testsuite/libffi.call/cls_align_sint16.c (cls_struct_align_gn):
826 Likewise.
827 * testsuite/libffi.call/cls_align_sint32.c (cls_struct_align_gn):
828 Likewise.
829 * testsuite/libffi.call/cls_align_sint64.c (cls_struct_align_gn):
830 Likewise.
831 * testsuite/libffi.call/cls_align_uint16.c (cls_struct_align_gn):
832 Likewise.
833 * testsuite/libffi.call/cls_align_uint32.c (cls_struct_align_gn):
834 Likewise.
835 * testsuite/libffi.call/cls_double.c (cls_ret_double_fn): Likewise.
836 * testsuite/libffi.call/cls_float.c (cls_ret_float_fn): Likewise.
837 * testsuite/libffi.call/cls_multi_schar.c (test_func_gn): Mark cif and
838 data unused.
839 (main): Cast res_call to silence gcc.
840 * testsuite/libffi.call/cls_multi_sshort.c (test_func_gn): Mark cif and
841 data unused.
842 (main): Cast res_call to silence gcc.
843 * testsuite/libffi.call/cls_multi_sshortchar.c (test_func_gn): Mark cif
844 and data unused.
845 (main): Cast res_call to silence gcc.
846 * testsuite/libffi.call/cls_multi_uchar.c (test_func_gn): Mark cif and
847 data unused.
848 (main): Cast res_call to silence gcc.
849 * testsuite/libffi.call/cls_multi_ushort.c (test_func_gn): Mark cif and
850 data unused.
851 (main): Cast res_call to silence gcc.
852 * testsuite/libffi.call/cls_multi_ushortchar.c (test_func_gn): Mark cif
853 and data unused.
854 (main): Cast res_call to silence gcc.
855 * testsuite/libffi.call/cls_schar.c (cls_ret_schar_fn): Mark cif and
856 userdata unused.
857 (cls_ret_schar_fn): Cast printf parameter to silence gcc.
858 * testsuite/libffi.call/cls_sint.c (cls_ret_sint_fn): Mark cif and
859 userdata unused.
860 (cls_ret_sint_fn): Cast printf parameter to silence gcc.
861 * testsuite/libffi.call/cls_sshort.c (cls_ret_sshort_fn): Mark cif and
862 userdata unused.
863 (cls_ret_sshort_fn): Cast printf parameter to silence gcc.
864 * testsuite/libffi.call/cls_uchar.c (cls_ret_uchar_fn): Mark cif and
865 userdata unused.
866 (cls_ret_uchar_fn): Cast printf parameter to silence gcc.
867 * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Mark cif and
868 userdata unused.
869 (cls_ret_uint_fn): Cast printf parameter to silence gcc.
870 * testsuite/libffi.call/cls_ulonglong.c (cls_ret_ulonglong_fn): Mark cif
871 and userdata unused.
872 * testsuite/libffi.call/cls_ushort.c (cls_ret_ushort_fn): Mark cif and
873 userdata unused.
874 (cls_ret_ushort_fn): Cast printf parameter to silence gcc.
875 * testsuite/libffi.call/float.c (floating): Remove unused parameter e.
876 * testsuite/libffi.call/float1.c (main): Remove unused variable i.
877 Cleanup white spaces.
878 * testsuite/libffi.call/negint.c (checking): Remove unused variable i.
879 * testsuite/libffi.call/nested_struct.c (cls_struct_combined_gn): Mark
880 cif and userdata unused.
881 * testsuite/libffi.call/nested_struct1.c (cls_struct_combined_gn):
882 Likewise.
883 * testsuite/libffi.call/nested_struct10.c (B_gn): Likewise.
884 * testsuite/libffi.call/nested_struct2.c (B_fn): Adjust printf
885 formatters to silence gcc.
886 (B_gn): Mark cif and userdata unused.
887 * testsuite/libffi.call/nested_struct3.c (B_gn): Mark cif and userdata
888 unused.
889 * testsuite/libffi.call/nested_struct4.c: Mention related PR.
890 (B_gn): Mark cif and userdata unused.
891 * testsuite/libffi.call/nested_struct5.c (B_gn): Mark cif and userdata
892 unused.
893 * testsuite/libffi.call/nested_struct6.c: Mention related PR.
894 (B_gn): Mark cif and userdata unused.
895 * testsuite/libffi.call/nested_struct7.c (B_gn): Mark cif and userdata
896 unused.
897 * testsuite/libffi.call/nested_struct8.c (B_gn): Likewise.
898 * testsuite/libffi.call/nested_struct9.c (B_gn): Likewise.
899 * testsuite/libffi.call/problem1.c (stub): Likewise.
900 * testsuite/libffi.call/pyobjc-tc.c (main): Cast the result to silence
901 gcc.
902 * testsuite/libffi.call/return_fl2.c (return_fl): Add the note mentioned
903 in the last commit for this test case in the test case itself.
904 * testsuite/libffi.call/closure_fn0.c (closure_test_fn0): Mark cif as
905 unused.
906 * testsuite/libffi.call/closure_fn1.c (closure_test_fn1): Likewise.
907 * testsuite/libffi.call/closure_fn2.c (closure_test_fn2): Likewise.
908 * testsuite/libffi.call/closure_fn3.c (closure_test_fn3): Likewise.
909 * testsuite/libffi.call/closure_fn4.c (closure_test_fn0): Likewise.
910 * testsuite/libffi.call/closure_fn5.c (closure_test_fn5): Likewise.
911 * testsuite/libffi.call/closure_fn6.c (closure_test_fn0): Likewise.
912
913 2006-02-22 Kaz Kojima <kkojima@gcc.gnu.org>
914
915 * src/sh/sysv.S: Fix register numbers in the FDE for
916 ffi_closure_SYSV.
917
918 2006-02-20 Andreas Tobler <a.tobler@schweiz.ch>
919
920 * testsuite/libffi.call/return_fl2.c (return_fl): Remove static
921 declaration to avoid a false negative on ix86. See PR323.
922
923 2006-02-18 Kaz Kojima <kkojima@gcc.gnu.org>
924
925 * src/sh/ffi.c (ffi_closure_helper_SYSV): Remove unused variable
926 and cast integer to void * if needed. Update the pointer to
927 the FP register saved area correctly.
928
929 2006-02-17 Andreas Tobler <a.tobler@schweiz.ch>
930
931 * testsuite/libffi.call/nested_struct6.c: XFAIL this test until PR25630
932 is fixed.
933 * testsuite/libffi.call/nested_struct4.c: Likewise.
934
935 2006-02-16 Andreas Tobler <a.tobler@schweiz.ch>
936
937 * testsuite/libffi.call/return_dbl.c: New test case.
938 * testsuite/libffi.call/return_dbl1.c: Likewise.
939 * testsuite/libffi.call/return_dbl2.c: Likewise.
940 * testsuite/libffi.call/return_fl.c: Likewise.
941 * testsuite/libffi.call/return_fl1.c: Likewise.
942 * testsuite/libffi.call/return_fl2.c: Likewise.
943 * testsuite/libffi.call/return_fl3.c: Likewise.
944 * testsuite/libffi.call/closure_fn6.c: Likewise.
945
946 * testsuite/libffi.call/nested_struct2.c: Remove ffi_type_mylong
947 definition.
948 * testsuite/libffi.call/ffitest.h: Add ffi_type_mylong definition
949 here to be used by other test cases too.
950
951 * testsuite/libffi.call/nested_struct10.c: New test case.
952 * testsuite/libffi.call/nested_struct9.c: Likewise.
953 * testsuite/libffi.call/nested_struct8.c: Likewise.
954 * testsuite/libffi.call/nested_struct7.c: Likewise.
955 * testsuite/libffi.call/nested_struct6.c: Likewise.
956 * testsuite/libffi.call/nested_struct5.c: Likewise.
957 * testsuite/libffi.call/nested_struct4.c: Likewise.
958
959 2006-01-21 Andreas Tobler <a.tobler@schweiz.ch>
960
961 * configure.ac: Enable libffi for sparc64-*-freebsd*.
962 * configure: Rebuilt.
963
964 2006-01-18 Jakub Jelinek <jakub@redhat.com>
965
966 * src/powerpc/sysv.S (smst_two_register): Don't call __ashldi3,
967 instead do the shifting inline.
968 * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't compute %r5
969 shift count unconditionally. Simplify load sequences for 1, 2, 3, 4
970 and 8 byte structs, for the remaining struct sizes don't call
971 __lshrdi3, instead do the shifting inline.
972
973 2005-12-07 Thiemo Seufer <ths@networkno.de>
974
975 * src/mips/ffitarget.h: Remove obsolete sgidefs.h include. Add
976 missing parentheses.
977 * src/mips/o32.S (ffi_call_O32): Code formatting. Define
978 and use A3_OFF, FP_OFF, RA_OFF. Micro-optimizations.
979 (ffi_closure_O32): Likewise, but with newly defined A3_OFF2,
980 A2_OFF2, A1_OFF2, A0_OFF2, RA_OFF2, FP_OFF2, S0_OFF2, GP_OFF2,
981 V1_OFF2, V0_OFF2, FA_1_1_OFF2, FA_1_0_OFF2, FA_0_1_OFF2,
982 FA_0_0_OFF2.
983 * src/mips/ffi.c (ffi_prep_args): Code formatting. Fix
984 endianness bugs.
985 (ffi_prep_closure): Improve trampoline instruction scheduling.
986 (ffi_closure_mips_inner_O32): Fix endianness bugs.
987
988 2005-12-03 Alan Modra <amodra@bigpond.net.au>
989
990 * src/powerpc/ffi.c: Formatting.
991 (ffi_prep_args_SYSV): Avoid possible aliasing problems by using unions.
992 (ffi_prep_args64): Likewise.
993
994 2005-09-30 Geoffrey Keating <geoffk@apple.com>
995
996 * testsuite/lib/libffi-dg.exp (libffi_target_compile): For
997 darwin, use -shared-libgcc not -lgcc_s, and explain why.
998
999 2005-09-26 Tom Tromey <tromey@redhat.com>
1000
1001 * testsuite/libffi.call/float1.c (value_type): New typedef.
1002 (CANARY): New define.
1003 (main): Check for result buffer overflow.
1004 * src/powerpc/linux64.S: Handle linux64 long double returns.
1005 * src/powerpc/ffi.c (FLAG_RETURNS_128BITS): New constant.
1006 (ffi_prep_cif_machdep): Handle linux64 long double returns.
1007
1008 2005-08-25 Alan Modra <amodra@bigpond.net.au>
1009
1010 PR target/23404
1011 * src/powerpc/ffi.c (ffi_prep_args_SYSV): Correct placement of stack
1012 homed fp args.
1013 (ffi_status ffi_prep_cif_machdep): Correct stack sizing for same.
1014
1015 2005-08-11 Jakub Jelinek <jakub@redhat.com>
1016
1017 * configure.ac (HAVE_HIDDEN_VISIBILITY_ATTRIBUTE): New test.
1018 (AH_BOTTOM): Add FFI_HIDDEN definition.
1019 * configure: Rebuilt.
1020 * fficonfig.h.in: Rebuilt.
1021 * src/powerpc/ffi.c (hidden): Remove.
1022 (ffi_closure_LINUX64, ffi_prep_args64, ffi_call_LINUX64,
1023 ffi_closure_helper_LINUX64): Use FFI_HIDDEN instead of hidden.
1024 * src/powerpc/linux64_closure.S (ffi_closure_LINUX64,
1025 .ffi_closure_LINUX64): Use FFI_HIDDEN instead of .hidden.
1026 * src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV): Remove,
1027 add FFI_HIDDEN to its prototype.
1028 (ffi_closure_SYSV_inner): New.
1029 * src/x86/sysv.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
1030 * src/x86/win32.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New.
1031
1032 2005-08-10 Alfred M. Szmidt <ams@gnu.org>
1033
1034 PR libffi/21819:
1035 * configure: Rebuilt.
1036 * configure.ac: Handle i*86-*-gnu*.
1037
1038 2005-08-09 Jakub Jelinek <jakub@redhat.com>
1039
1040 * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Use
1041 DW_CFA_offset_extended_sf rather than
1042 DW_CFA_GNU_negative_offset_extended.
1043 * src/powerpc/sysv.S (ffi_call_SYSV): Likewise.
1044
1045 2005-07-22 SUGIOKA Toshinobu <sugioka@itonet.co.jp>
1046
1047 * src/sh/sysv.S (ffi_call_SYSV): Stop argument popping correctly
1048 on sh3.
1049 (ffi_closure_SYSV): Change the stack layout for sh3 struct argument.
1050 * src/sh/ffi.c (ffi_prep_args): Fix sh3 argument copy, when it is
1051 partially on register.
1052 (ffi_closure_helper_SYSV): Likewise.
1053 (ffi_prep_cif_machdep): Don't set too many cif->flags.
1054
1055 2005-07-20 Kaz Kojima <kkojima@gcc.gnu.org>
1056
1057 * src/sh/ffi.c (ffi_call): Handle small structures correctly.
1058 Remove empty line.
1059 * src/sh64/ffi.c (simple_type): Remove.
1060 (return_type): Handle small structures correctly.
1061 (ffi_prep_args): Likewise.
1062 (ffi_call): Likewise.
1063 (ffi_closure_helper_SYSV): Likewise.
1064 * src/sh64/sysv.S (ffi_call_SYSV): Handle 1, 2 and 4-byte return.
1065 Emit position independent code if PIC and remove wrong datalabel
1066 prefixes from EH data.
1067
1068 2005-07-19 Andreas Tobler <a.tobler@schweiz.ch>
1069
1070 * Makefile.am (nodist_libffi_la_SOURCES): Add POWERPC_FREEBSD.
1071 * Makefile.in: Regenerate.
1072 * include/Makefile.in: Likewise.
1073 * testsuite/Makefile.in: Likewise.
1074 * configure.ac: Add POWERPC_FREEBSD rules.
1075 * configure: Regenerate.
1076 * src/powerpc/ffitarget.h: Add POWERPC_FREEBSD rules.
1077 (FFI_SYSV_TYPE_SMALL_STRUCT): Define.
1078 * src/powerpc/ffi.c: Add flags to handle small structure returns
1079 in ffi_call_SYSV.
1080 (ffi_prep_cif_machdep): Handle small structures for SYSV 4 ABI.
1081 Aka FFI_SYSV.
1082 (ffi_closure_helper_SYSV): Likewise.
1083 * src/powerpc/ppc_closure.S: Add return types for small structures.
1084 * src/powerpc/sysv.S: Add bits to handle small structures for
1085 final SYSV 4 ABI.
1086
1087 2005-07-10 Andreas Tobler <a.tobler@schweiz.ch>
1088
1089 * testsuite/libffi.call/cls_5_1_byte.c: New test file.
1090 * testsuite/libffi.call/cls_6_1_byte.c: Likewise.
1091 * testsuite/libffi.call/cls_7_1_byte.c: Likewise.
1092
1093 2005-07-05 Randolph Chung <tausq@debian.org>
1094
1095 * src/pa/ffi.c (ffi_struct_type): Rename FFI_TYPE_SMALL_STRUCT1
1096 as FFI_TYPE_SMALL_STRUCT3. Break out handling for 5-7 byte
1097 structures. Kill compilation warnings.
1098 (ffi_closure_inner_LINUX): Print return values as hex in debug
1099 message. Rename FFI_TYPE_SMALL_STRUCT1 as FFI_TYPE_SMALL_STRUCT3.
1100 Properly handle 5-7 byte structure returns.
1101 * src/pa/ffitarget.h (FFI_TYPE_SMALL_STRUCT1)
1102 (FFI_TYPE_SMALL_STRUCT2): Remove.
1103 (FFI_TYPE_SMALL_STRUCT3, FFI_TYPE_SMALL_STRUCT5)
1104 (FFI_TYPE_SMALL_STRUCT6, FFI_TYPE_SMALL_STRUCT7): Define.
1105 * src/pa/linux.S: Mark source file as using PA1.1 assembly.
1106 (checksmst1, checksmst2): Remove.
1107 (checksmst3): Optimize handling of 3-byte struct returns.
1108 (checksmst567): Properly handle 5-7 byte struct returns.
1109
1110 2005-06-15 Rainer Orth <ro@TechFak.Uni-Bielefeld.DE>
1111
1112 PR libgcj/21943
1113 * src/mips/n32.S: Enforce PIC code.
1114 * src/mips/o32.S: Likewise.
1115
1116 2005-06-15 Rainer Orth <ro@TechFak.Uni-Bielefeld.DE>
1117
1118 * configure.ac: Treat i*86-*-solaris2.10 and up as X86_64.
1119 * configure: Regenerate.
1120
1121 2005-06-01 Alan Modra <amodra@bigpond.net.au>
1122
1123 * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't use JUMPTARGET
1124 to call ffi_closure_helper_SYSV. Append @local instead.
1125 * src/powerpc/sysv.S (ffi_call_SYSV): Likewise for ffi_prep_args_SYSV.
1126
1127 2005-05-17 Kelley Cook <kcook@gcc.gnu.org>
1128
1129 * configure.ac: Use AC_C_BIGENDIAN instead of AC_C_BIGENDIAN_CROSS.
1130 Use AC_CHECK_SIZEOF instead of AC_COMPILE_CHECK_SIZEOF.
1131 * Makefile.am (ACLOCAL_AMFLAGS): Remove -I ../config.
1132 * aclocal.m4, configure, fficonfig.h.in, Makefile.in,
1133 include/Makefile.in, testsuite/Makefile.in: Regenerate.
1134
1135 2005-05-09 Mike Stump <mrs@apple.com>
1136
1137 * configure: Regenerate.
1138
1139 2005-05-08 Richard Henderson <rth@redhat.com>
1140
1141 PR libffi/21285
1142 * src/alpha/osf.S: Update unwind into to match code.
1143
1144 2005-05-04 Andreas Degert <ad@papyrus-gmbh.de>
1145 Richard Henderson <rth@redhat.com>
1146
1147 * src/x86/ffi64.c (ffi_prep_cif_machdep): Save sse-used flag in
1148 bit 11 of flags.
1149 (ffi_call): Mask return type field. Pass ssecount to ffi_call_unix64.
1150 (ffi_prep_closure): Set carry bit if sse-used flag set.
1151 * src/x86/unix64.S (ffi_call_unix64): Add ssecount argument.
1152 Only load sse registers if ssecount non-zero.
1153 (ffi_closure_unix64): Only save sse registers if carry set on entry.
1154
1155 2005-04-29 Ralf Corsepius <ralf.corsepius@rtems.org>
1156
1157 * configure.ac: Add i*86-*-rtems*, sparc*-*-rtems*,
1158 powerpc-*rtems*, arm*-*-rtems*, sh-*-rtems*.
1159 * configure: Regenerate.
1160
1161 2005-04-20 Hans-Peter Nilsson <hp@axis.com>
1162
1163 * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): In regsub use,
1164 have Tcl8.3-compatible intermediate variable.
1165
1166 2005-04-18 Simon Posnjak <simon.posnjak@siol.net>
1167 Hans-Peter Nilsson <hp@axis.com>
1168
1169 * Makefile.am: Add CRIS support.
1170 * configure.ac: Likewise.
1171 * Makefile.in, configure, testsuite/Makefile.in,
1172 include/Makefile.in: Regenerate.
1173 * src/cris: New directory.
1174 * src/cris/ffi.c, src/cris/sysv.S, src/cris/ffitarget.h: New files.
1175 * src/prep_cif.c (ffi_prep_cif): Wrap in #ifndef __CRIS__.
1176
1177 * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): Replace \n with
1178 \r?\n in output tests.
1179
1180 2005-04-12 Mike Stump <mrs@apple.com>
1181
1182 * configure: Regenerate.
1183
1184 2005-03-30 Hans Boehm <Hans.Boehm@hp.com>
1185
1186 * src/ia64/ffitarget.h (ffi_arg): Use long long instead of DI.
1187
1188 2005-03-30 Steve Ellcey <sje@cup.hp.com>
1189
1190 * src/ia64/ffitarget.h (ffi_arg) ADD DI attribute.
1191 (ffi_sarg) Ditto.
1192 * src/ia64/unix.S (ffi_closure_unix): Extend gp
1193 to 64 bits in ILP32 mode.
1194 Load 64 bits even for short data.
1195
1196 2005-03-23 Mike Stump <mrs@apple.com>
1197
1198 * src/powerpc/darwin.S: Update for -m64 multilib.
1199 * src/powerpc/darwin_closure.S: Likewise.
1200
1201 2005-03-21 Zack Weinberg <zack@codesourcery.com>
1202
1203 * configure.ac: Do not invoke TL_AC_GCC_VERSION.
1204 Do not set tool_include_dir.
1205 * aclocal.m4, configure, Makefile.in, testsuite/Makefile.in:
1206 Regenerate.
1207 * include/Makefile.am: Set gcc_version and toollibffidir.
1208 * include/Makefile.in: Regenerate.
1209
1210 2005-02-22 Andrew Haley <aph@redhat.com>
1211
1212 * src/powerpc/ffi.c (ffi_prep_cif_machdep): Bump alignment to
1213 odd-numbered register pairs for 64-bit integer types.
1214
1215 2005-02-23 Andreas Tobler <a.tobler@schweiz.ch>
1216
1217 PR libffi/20104
1218 * testsuite/libffi.call/return_ll1.c: New test case.
1219
1220 2005-02-11 Janis Johnson <janis187@us.ibm.com>
1221
1222 * testsuite/libffi.call/cls_align_longdouble.c: Remove dg-options.
1223 * testsuite/libffi.call/float.c: Ditto.
1224 * testsuite/libffi.call/float2.c: Ditto.
1225 * testsuite/libffi.call/float3.c: Ditto.
1226
1227 2005-02-08 Andreas Tobler <a.tobler@schweiz.ch>
1228
1229 * src/frv/ffitarget.h: Remove PPC stuff which does not belong to frv.
1230
1231 2005-01-12 Eric Botcazou <ebotcazou@libertysurf.fr>
1232
1233 * testsuite/libffi.special/special.exp (cxx_options): Add
1234 -shared-libgcc.
1235
1236 2004-12-31 Richard Henderson <rth@redhat.com>
1237
1238 * src/types.c (FFI_AGGREGATE_TYPEDEF): Remove.
1239 (FFI_TYPEDEF): Rename from FFI_INTEGRAL_TYPEDEF. Replace size and
1240 offset parameters with a type parameter; deduce size and structure
1241 alignment. Update all users.
1242
1243 2004-12-31 Richard Henderson <rth@redhat.com>
1244
1245 * src/types.c (FFI_TYPE_POINTER): Define with sizeof.
1246 (FFI_TYPE_LONGDOUBLE): Fix for ia64.
1247 * src/ia64/ffitarget.h (struct ffi_ia64_trampoline_struct): Move
1248 into ffi_prep_closure.
1249 * src/ia64/ia64_flags.h, src/ia64/ffi.c, src/ia64/unix.S: Rewrite
1250 from scratch.
1251
1252 2004-12-27 Richard Henderson <rth@redhat.com>
1253
1254 * src/x86/unix64.S: Fix typo in unwind info.
1255
1256 2004-12-25 Richard Henderson <rth@redhat.com>
1257
1258 * src/x86/ffi64.c (struct register_args): Rename from stackLayout.
1259 (enum x86_64_reg_class): Add X86_64_COMPLEX_X87_CLASS.
1260 (merge_classes): Check for it.
1261 (SSE_CLASS_P): New.
1262 (classify_argument): Pass byte_offset by value; perform all updates
1263 inside struct case.
1264 (examine_argument): Add classes argument; handle
1265 X86_64_COMPLEX_X87_CLASS.
1266 (ffi_prep_args): Merge into ...
1267 (ffi_call): ... here. Share stack frame with ffi_call_unix64.
1268 (ffi_prep_cif_machdep): Setup cif->flags for proper structure return.
1269 (ffi_fill_return_value): Remove.
1270 (ffi_prep_closure): Remove dead assert.
1271 (ffi_closure_unix64_inner): Rename from ffi_closure_UNIX64_inner.
1272 Rewrite to use struct register_args instead of va_list. Create
1273 flags for handling structure returns.
1274 * src/x86/unix64.S: Remove dead strings.
1275 (ffi_call_unix64): Rename from ffi_call_UNIX64. Rewrite to share
1276 stack frame with ffi_call. Handle structure returns properly.
1277 (float2sse, floatfloat2sse, double2sse): Remove.
1278 (sse2float, sse2double, sse2floatfloat): Remove.
1279 (ffi_closure_unix64): Rename from ffi_closure_UNIX64. Rewrite
1280 to handle structure returns properly.
1281
1282 2004-12-08 David Edelsohn <edelsohn@gnu.org>
1283
1284 * Makefile.am (AM_MAKEFLAGS): Remove duplicate LIBCFLAGS and
1285 PICFLAG.
1286 * Makefile.in: Regenerated.
1287
1288 2004-12-02 Richard Sandiford <rsandifo@redhat.com>
1289
1290 * configure.ac: Use TL_AC_GCC_VERSION to set gcc_version.
1291 * configure, aclocal.m4, Makefile.in: Regenerate.
1292 * include/Makefile.in, testsuite/Makefile.in: Regenerate.
1293
1294 2004-11-29 Kelley Cook <kcook@gcc.gnu.org>
1295
1296 * configure: Regenerate for libtool change.
1297
1298 2004-11-25 Kelley Cook <kcook@gcc.gnu.org>
1299
1300 * configure: Regenerate for libtool reversion.
1301
1302 2004-11-24 Kelley Cook <kcook@gcc.gnu.org>
1303
1304 * configure: Regenerate for libtool change.
1305
1306 2004-11-23 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
1307
1308 * testsuite/lib/libffi-dg.exp: Use new procs in target-libpath.exp.
1309
1310 2004-11-23 Richard Sandiford <rsandifo@redhat.com>
1311
1312 * src/mips/o32.S (ffi_call_O32, ffi_closure_O32): Use jalr instead
1313 of jal. Use an absolute encoding for the frame information.
1314
1315 2004-11-23 Kelley Cook <kcook@gcc.gnu.org>
1316
1317 * Makefile.am: Remove no-dependencies. Add ACLOCAL_AMFLAGS.
1318 * acinclude.m4: Delete logic for sincludes.
1319 * aclocal.m4, Makefile.in, configure: Regenerate.
1320 * include/Makefile: Likewise.
1321 * testsuite/Makefile: Likewise.
1322
1323 2004-11-22 Eric Botcazou <ebotcazou@libertysurf.fr>
1324
1325 * src/sparc/ffi.c (ffi_prep_closure): Align doubles and 64-bit integers
1326 on a 8-byte boundary.
1327 * src/sparc/v8.S (ffi_closure_v8): Reserve frame space for arguments.
1328
1329 2004-10-27 Richard Earnshaw <rearnsha@arm.com>
1330
1331 * src/arm/ffi.c (ffi_prep_cif_machdep): Handle functions that return
1332 long long values. Round stack allocation to a multiple of 8 bytes
1333 for ATPCS compatibility.
1334 * src/arm/sysv.S (ffi_call_SYSV): Rework to avoid use of APCS register
1335 names. Handle returning long long types. Add Thumb and interworking
1336 support. Improve soft-float code.
1337
1338 2004-10-27 Richard Earnshaw <rearnsha@arm.com>
1339
1340 * testsuite/lib/libffi-db.exp (load_gcc_lib): New function.
1341 (libffi_exit): New function.
1342 (libffi_init): Build the testglue wrapper if needed.
1343
1344 2004-10-25 Eric Botcazou <ebotcazou@libertysurf.fr>
1345
1346 PR other/18138
1347 * testsuite/lib/libffi-dg.exp: Accept more than one multilib libgcc.
1348
1349 2004-10-25 Kazuhiro Inaoka <inaoka.kazuhiro@renesas.com>
1350
1351 * src/m32r/libffitarget.h (FFI_CLOSURES): Set to 0.
1352
1353 2004-10-20 Kaz Kojima <kkojima@gcc.gnu.org>
1354
1355 * src/sh/sysv.S (ffi_call_SYSV): Don't align for double data.
1356 * testsuite/libffi.call/float3.c: New test case.
1357
1358 2004-10-18 Kaz Kojima <kkojima@gcc.gnu.org>
1359
1360 * src/sh/ffi.c (ffi_prep_closure): Set T bit in trampoline for
1361 the function returning a structure pointed with R2.
1362 * src/sh/sysv.S (ffi_closure_SYSV): Use R2 as the pointer to
1363 the structure return value if T bit set. Emit position
1364 independent code and EH data if PIC.
1365
1366 2004-10-13 Kazuhiro Inaoka <inaoka.kazuhiro@renesas.com>
1367
1368 * Makefile.am: Add m32r support.
1369 * configure.ac: Likewise.
1370 * Makefile.in: Regenerate.
1371 * confiugre: Regenerate.
1372 * src/types.c: Add m32r port to FFI_INTERNAL_TYPEDEF
1373 (uint64, sint64, double, longdouble)
1374 * src/m32r: New directory.
1375 * src/m32r/ffi.c: New file.
1376 * src/m32r/sysv.S: Likewise.
1377 * src/m32r/ffitarget.h: Likewise.
1378
1379 2004-10-02 Kaz Kojima <kkojima@gcc.gnu.org>
1380
1381 * testsuite/libffi.call/negint.c: New test case.
1382
1383 2004-09-14 H.J. Lu <hongjiu.lu@intel.com>
1384
1385 PR libgcj/17465
1386 * testsuite/lib/libffi-dg.exp: Don't use global ld_library_path.
1387 Set up LD_LIBRARY_PATH, SHLIB_PATH, LD_LIBRARYN32_PATH,
1388 LD_LIBRARY64_PATH, LD_LIBRARY_PATH_32, LD_LIBRARY_PATH_64 and
1389 DYLD_LIBRARY_PATH.
1390
1391 2004-09-05 Andreas Tobler <a.tobler@schweiz.ch>
1392
1393 * testsuite/libffi.call/many_win32.c: Remove whitespaces.
1394 * testsuite/libffi.call/promotion.c: Likewise.
1395 * testsuite/libffi.call/return_ll.c: Remove unused var. Cleanup
1396 whitespaces.
1397 * testsuite/libffi.call/return_sc.c: Likewise.
1398 * testsuite/libffi.call/return_uc.c: Likewise.
1399
1400 2004-09-05 Andreas Tobler <a.tobler@schweiz.ch>
1401
1402 * src/powerpc/darwin.S: Fix comments and identation.
1403 * src/powerpc/darwin_closure.S: Likewise.
1404
1405 2004-09-02 Andreas Tobler <a.tobler@schweiz.ch>
1406
1407 * src/powerpc/ffi_darwin.c: Add flag for longdouble return values.
1408 (ffi_prep_args): Handle longdouble arguments.
1409 (ffi_prep_cif_machdep): Set flags for longdouble. Calculate space for
1410 longdouble.
1411 (ffi_closure_helper_DARWIN): Add closure handling for longdouble.
1412 * src/powerpc/darwin.S (_ffi_call_DARWIN): Add handling of longdouble
1413 values.
1414 * src/powerpc/darwin_closure.S (_ffi_closure_ASM): Likewise.
1415 * src/types.c: Defined longdouble size and alignment for darwin.
1416
1417 2004-09-02 Andreas Tobler <a.tobler@schweiz.ch>
1418
1419 * src/powerpc/aix.S: Remove whitespaces.
1420 * src/powerpc/aix_closure.S: Likewise.
1421 * src/powerpc/asm.h: Likewise.
1422 * src/powerpc/ffi.c: Likewise.
1423 * src/powerpc/ffitarget.h: Likewise.
1424 * src/powerpc/linux64.S: Likewise.
1425 * src/powerpc/linux64_closure.S: Likewise.
1426 * src/powerpc/ppc_closure.S: Likewise.
1427 * src/powerpc/sysv.S: Likewise.
1428
1429 2004-08-30 Anthony Green <green@redhat.com>
1430
1431 * Makefile.am: Add frv support.
1432 * Makefile.in, testsuite/Makefile.in: Rebuilt.
1433 * configure.ac: Read configure.host.
1434 * configure.in: Read configure.host.
1435 * configure.host: New file. frv-elf needs libgloss.
1436 * include/ffi.h.in: Force ffi_closure to have a nice big (8)
1437 alignment. This is needed to frv and shouldn't harm the others.
1438 * include/ffi_common.h (ALIGN_DOWN): New macro.
1439 * src/frv/ffi.c, src/frv/ffitarget.h, src/frv/eabi.S: New files.
1440
1441 2004-08-24 David Daney <daney@avtrex.com>
1442
1443 * testsuite/libffi.call/closure_fn0.c: Xfail mips64* instead of mips*.
1444 * testsuite/libffi.call/closure_fn1.c: Likewise.
1445 * testsuite/libffi.call/closure_fn2.c Likewise.
1446 * testsuite/libffi.call/closure_fn3.c: Likewise.
1447 * testsuite/libffi.call/closure_fn4.c: Likewise.
1448 * testsuite/libffi.call/closure_fn5.c: Likewise.
1449 * testsuite/libffi.call/cls_18byte.c: Likewise.
1450 * testsuite/libffi.call/cls_19byte.c: Likewise.
1451 * testsuite/libffi.call/cls_1_1byte.c: Likewise.
1452 * testsuite/libffi.call/cls_20byte.c: Likewise.
1453 * testsuite/libffi.call/cls_20byte1.c: Likewise.
1454 * testsuite/libffi.call/cls_24byte.c: Likewise.
1455 * testsuite/libffi.call/cls_2byte.c: Likewise.
1456 * testsuite/libffi.call/cls_3_1byte.c: Likewise.
1457 * testsuite/libffi.call/cls_3byte1.c: Likewise.
1458 * testsuite/libffi.call/cls_3byte2.c: Likewise.
1459 * testsuite/libffi.call/cls_4_1byte.c: Likewise.
1460 * testsuite/libffi.call/cls_4byte.c: Likewise.
1461 * testsuite/libffi.call/cls_64byte.c: Likewise.
1462 * testsuite/libffi.call/cls_6byte.c: Likewise.
1463 * testsuite/libffi.call/cls_7byte.c: Likewise.
1464 * testsuite/libffi.call/cls_8byte.c: Likewise.
1465 * testsuite/libffi.call/cls_9byte1.c: Likewise.
1466 * testsuite/libffi.call/cls_9byte2.c: Likewise.
1467 * testsuite/libffi.call/cls_align_double.c: Likewise.
1468 * testsuite/libffi.call/cls_align_float.c: Likewise.
1469 * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
1470 * testsuite/libffi.call/cls_align_pointer.c: Likewise.
1471 * testsuite/libffi.call/cls_align_sint16.c: Likewise.
1472 * testsuite/libffi.call/cls_align_sint32.c: Likewise.
1473 * testsuite/libffi.call/cls_align_sint64.c: Likewise.
1474 * testsuite/libffi.call/cls_align_uint16.c: Likewise.
1475 * testsuite/libffi.call/cls_align_uint32.c: Likewise.
1476 * testsuite/libffi.call/cls_align_uint64.c: Likewise.
1477 * testsuite/libffi.call/cls_double.c: Likewise.
1478 * testsuite/libffi.call/cls_float.c: Likewise.
1479 * testsuite/libffi.call/cls_multi_schar.c: Likewise.
1480 * testsuite/libffi.call/cls_multi_sshort.c: Likewise.
1481 * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
1482 * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
1483 * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
1484 * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
1485 * testsuite/libffi.call/cls_schar.c: Likewise.
1486 * testsuite/libffi.call/cls_sint.c: Likewise.
1487 * testsuite/libffi.call/cls_sshort.c: Likewise.
1488 * testsuite/libffi.call/cls_uchar.c: Likewise.
1489 * testsuite/libffi.call/cls_uint.c: Likewise.
1490 * testsuite/libffi.call/cls_ulonglong.c: Likewise.
1491 * testsuite/libffi.call/cls_ushort.c: Likewise.
1492 * testsuite/libffi.call/nested_struct.c: Likewise.
1493 * testsuite/libffi.call/nested_struct1.c: Likewise.
1494 * testsuite/libffi.call/nested_struct2.c: Likewise.
1495 * testsuite/libffi.call/nested_struct3.c: Likewise.
1496 * testsuite/libffi.call/problem1.c: Likewise.
1497 * testsuite/libffi.special/unwindtest.cc: Likewise.
1498 * testsuite/libffi.call/cls_12byte.c: Likewise and set return value
1499 to zero.
1500 * testsuite/libffi.call/cls_16byte.c: Likewise.
1501 * testsuite/libffi.call/cls_5byte.c: Likewise.
1502
1503 2004-08-23 David Daney <daney@avtrex.com>
1504
1505 PR libgcj/13141
1506 * src/mips/ffitarget.h (FFI_O32_SOFT_FLOAT): New ABI.
1507 * src/mips/ffi.c (ffi_prep_args): Fix alignment calculation.
1508 (ffi_prep_cif_machdep): Handle FFI_O32_SOFT_FLOAT floating point
1509 parameters and return types.
1510 (ffi_call): Handle FFI_O32_SOFT_FLOAT ABI.
1511 (ffi_prep_closure): Ditto.
1512 (ffi_closure_mips_inner_O32): Handle FFI_O32_SOFT_FLOAT ABI, fix
1513 alignment calculations.
1514 * src/mips/o32.S (ffi_closure_O32): Don't use floating point
1515 instructions if FFI_O32_SOFT_FLOAT, make stack frame ABI compliant.
1516
1517 2004-08-14 Casey Marshall <csm@gnu.org>
1518
1519 * src/mips/ffi.c (ffi_pref_cif_machdep): set `cif->flags' to
1520 contain `FFI_TYPE_UINT64' as return type for any 64-bit
1521 integer (O32 ABI only).
1522 (ffi_prep_closure): new function.
1523 (ffi_closure_mips_inner_O32): new function.
1524 * src/mips/ffitarget.h: Define `FFI_CLOSURES' and
1525 `FFI_TRAMPOLINE_SIZE' appropriately if the ABI is o32.
1526 * src/mips/o32.S (ffi_call_O32): add labels for .eh_frame. Return
1527 64 bit integers correctly.
1528 (ffi_closure_O32): new function.
1529 Added DWARF-2 unwind info for both functions.
1530
1531 2004-08-10 Andrew Haley <aph@redhat.com>
1532
1533 * src/x86/ffi64.c (ffi_prep_args ): 8-align all stack arguments.
1534
1535 2004-08-01 Robert Millan <robertmh@gnu.org>
1536
1537 * configure.ac: Detect knetbsd-gnu and kfreebsd-gnu.
1538 * configure: Regenerate.
1539
1540 2004-07-30 Maciej W. Rozycki <macro@linux-mips.org>
1541
1542 * acinclude.m4 (AC_FUNC_MMAP_BLACKLIST): Check for <sys/mman.h>
1543 and mmap() explicitly instead of relying on preset autoconf cache
1544 variables.
1545 * aclocal.m4: Regenerate.
1546 * configure: Regenerate.
1547
1548 2004-07-11 Ulrich Weigand <uweigand@de.ibm.com>
1549
1550 * src/s390/ffi.c (ffi_prep_args): Fix C aliasing violation.
1551 (ffi_check_float_struct): Remove unused prototype.
1552
1553 2004-06-30 Geoffrey Keating <geoffk@apple.com>
1554
1555 * src/powerpc/ffi_darwin.c (flush_icache): ';' is a comment
1556 character on Darwin, use '\n\t' instead.
1557
1558 2004-06-26 Matthias Klose <doko@debian.org>
1559
1560 * libtool-version: Fix typo in revision/age.
1561
1562 2004-06-17 Matthias Klose <doko@debian.org>
1563
1564 * libtool-version: New.
1565 * Makefile.am (libffi_la_LDFLAGS): Use -version-info for soname.
1566 * Makefile.in: Regenerate.
1567
1568 2004-06-15 Paolo Bonzini <bonzini@gnu.org>
1569
1570 * Makefile.am: Remove useless multilib rules.
1571 * Makefile.in: Regenerate.
1572 * aclocal.m4: Regenerate with automake 1.8.5.
1573 * configure.ac: Remove useless multilib configury.
1574 * configure: Regenerate.
1575
1576 2004-06-15 Paolo Bonzini <bonzini@gnu.org>
1577
1578 * .cvsignore: New file.
1579
1580 2004-06-10 Jakub Jelinek <jakub@redhat.com>
1581
1582 * src/ia64/unix.S (ffi_call_unix): Insert group barrier break
1583 fp_done.
1584 (ffi_closure_UNIX): Fix f14/f15 adjustment if FLOAT_SZ is ever
1585 changed from 8.
1586
1587 2004-06-06 Sean McNeil <sean@mcneil.com>
1588
1589 * configure.ac: Add x86_64-*-freebsd* support.
1590 * configure: Regenerate.
1591
1592 2004-04-26 Joe Buck <jbuck@welsh-buck.org>
1593
1594 Bug 15093
1595 * configure.ac: Test for existence of mmap and sys/mman.h before
1596 checking blacklist. Fix suggested by Jim Wilson.
1597 * configure: Regenerate.
1598
1599 2004-04-26 Matt Austern <austern@apple.com>
1600
1601 * src/powerpc/darwin.S: Go through a non-lazy pointer for initial
1602 FDE location.
1603 * src/powerpc/darwin_closure.S: Likewise.
1604
1605 2004-04-24 Andreas Tobler <a.tobler@schweiz.ch>
1606
1607 * testsuite/libffi.call/cls_multi_schar.c (main): Fix initialization
1608 error. Reported by Thomas Heller <theller@python.net>.
1609 * testsuite/libffi.call/cls_multi_sshort.c (main): Likewise.
1610 * testsuite/libffi.call/cls_multi_ushort.c (main): Likewise.
1611
1612 2004-03-20 Matthias Klose <doko@debian.org>
1613
1614 * src/pa/linux.S: Fix typo.
1615
1616 2004-03-19 Matthias Klose <doko@debian.org>
1617
1618 * Makefile.am: Update.
1619 * Makefile.in: Regenerate.
1620 * src/pa/ffi.h.in: Remove.
1621 * src/pa/ffitarget.h: New file.
1622
1623 2004-02-10 Randolph Chung <tausq@debian.org>
1624
1625 * Makefile.am: Add PA support.
1626 * Makefile.in: Regenerate.
1627 * include/Makefile.in: Regenerate.
1628 * configure.ac: Add PA target.
1629 * configure: Regenerate.
1630 * src/pa/ffi.c: New file.
1631 * src/pa/ffi.h.in: Add PA support.
1632 * src/pa/linux.S: New file.
1633 * prep_cif.c: Add PA support.
1634
1635 2004-03-16 Hosaka Yuji <hos@tamanegi.org>
1636
1637 * src/types.c: Fix alignment size of X86_WIN32 case int64 and
1638 double.
1639 * src/x86/ffi.c (ffi_prep_args): Replace ecif->cif->rtype->type
1640 with ecif->cif->flags.
1641 (ffi_call, ffi_prep_incoming_args_SYSV): Replace cif->rtype->type
1642 with cif->flags.
1643 (ffi_prep_cif_machdep): Add X86_WIN32 struct case.
1644 (ffi_closure_SYSV): Add 1 or 2-bytes struct case for X86_WIN32.
1645 * src/x86/win32.S (retstruct1b, retstruct2b, sc_retstruct1b,
1646 sc_retstruct2b): Add for 1 or 2-bytes struct case.
1647
1648 2004-03-15 Kelley Cook <kcook@gcc.gnu.org>
1649
1650 * configure.in: Rename file to ...
1651 * configure.ac: ... this.
1652 * fficonfig.h.in: Regenerate.
1653 * Makefile.in: Regenerate.
1654 * include/Makefile.in: Regenerate.
1655 * testsuite/Makefile.in: Regenerate.
1656
1657 2004-03-12 Matt Austern <austern@apple.com>
1658
1659 * src/powerpc/darwin.S: Fix EH information so it corresponds to
1660 changes in EH format resulting from addition of linkonce support.
1661 * src/powerpc/darwin_closure.S: Likewise.
1662
1663 2004-03-11 Andreas Tobler <a.tobler@schweiz.ch>
1664 Paolo Bonzini <bonzini@gnu.org>
1665
1666 * Makefile.am (AUTOMAKE_OPTIONS): Set them.
1667 Remove VPATH. Remove rules for object files. Remove multilib support.
1668 (AM_CCASFLAGS): Add.
1669 * configure.in (AC_CONFIG_HEADERS): Relace AM_CONFIG_HEADER.
1670 (AC_PREREQ): Bump version to 2.59.
1671 (AC_INIT): Fill with version info and bug address.
1672 (ORIGINAL_LD_FOR_MULTILIBS): Remove.
1673 (AM_ENABLE_MULTILIB): Use this instead of AC_ARG_ENABLE.
1674 De-precious CC so that the right flags are passed down to multilibs.
1675 (AC_MSG_ERROR): Replace obsolete macro AC_ERROR.
1676 (AC_CONFIG_FILES): Replace obsolete macro AC_LINK_FILES.
1677 (AC_OUTPUT): Reorganize the output with AC_CONFIG_COMMANDS.
1678 * configure: Rebuilt.
1679 * aclocal.m4: Likewise.
1680 * Makefile.in, include/Makefile.in, testsuite/Makefile.in: Likewise.
1681 * fficonfig.h.in: Likewise.
1682
1683 2004-03-11 Andreas Schwab <schwab@suse.de>
1684
1685 * src/ia64/ffi.c (ffi_prep_incoming_args_UNIX): Get floating point
1686 arguments from fp registers only for the first 8 parameter slots.
1687 Don't convert a float parameter when passed in memory.
1688
1689 2004-03-09 Hans-Peter Nilsson <hp@axis.com>
1690
1691 * configure: Regenerate for config/accross.m4 correction.
1692
1693 2004-02-25 Matt Kraai <kraai@alumni.cmu.edu>
1694
1695 * src/powerpc/ffi.c (ffi_prep_args_SYSV): Change
1696 ecif->cif->bytes to bytes.
1697 (ffi_prep_cif_machdep): Add braces around nested if statement.
1698
1699 2004-02-09 Alan Modra <amodra@bigpond.net.au>
1700
1701 * src/types.c (pointer): POWERPC64 has 8 byte pointers.
1702
1703 * src/powerpc/ffi.c (ffi_prep_args64): Correct long double handling.
1704 (ffi_closure_helper_LINUX64): Fix typo.
1705 * testsuite/libffi.call/cls_align_longdouble.c: Pass -mlong-double-128
1706 for powerpc64-*-*.
1707 * testsuite/libffi.call/float.c: Likewise.
1708 * testsuite/libffi.call/float2.c: Likewise.
1709
1710 2004-02-08 Alan Modra <amodra@bigpond.net.au>
1711
1712 * src/powerpc/ffi.c (ffi_prep_cif_machdep <FFI_LINUX64>): Correct
1713 long double function return and long double arg handling.
1714 (ffi_closure_helper_LINUX64): Formatting. Delete unused "ng" var.
1715 Use "end_pfr" instead of "nf". Correct long double handling.
1716 Localise "temp".
1717 * src/powerpc/linux64.S (ffi_call_LINUX64): Save f2 long double
1718 return value.
1719 * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Allocate
1720 space for long double return value. Adjust stack frame and offsets.
1721 Load f2 long double return.
1722
1723 2004-02-07 Alan Modra <amodra@bigpond.net.au>
1724
1725 * src/types.c: Use 16 byte long double for POWERPC64.
1726
1727 2004-01-25 Eric Botcazou <ebotcazou@libertysurf.fr>
1728
1729 * src/sparc/ffi.c (ffi_prep_args_v9): Shift the parameter array
1730 when the structure return address is passed in %o0.
1731 (ffi_V9_return_struct): Rename into ffi_v9_layout_struct.
1732 (ffi_v9_layout_struct): Align the field following a nested structure
1733 on a word boundary. Use memmove instead of memcpy.
1734 (ffi_call): Update call to ffi_V9_return_struct.
1735 (ffi_prep_closure): Define 'ctx' only for V8.
1736 (ffi_closure_sparc_inner): Clone into ffi_closure_sparc_inner_v8
1737 and ffi_closure_sparc_inner_v9.
1738 (ffi_closure_sparc_inner_v8): Return long doubles by reference.
1739 Always skip the structure return address. For structures and long
1740 doubles, copy the argument directly.
1741 (ffi_closure_sparc_inner_v9): Skip the structure return address only
1742 if required. Shift the maximum floating-point slot accordingly. For
1743 big structures, copy the argument directly; otherwise, left-justify the
1744 argument and call ffi_v9_layout_struct to lay out the structure on
1745 the stack.
1746 * src/sparc/v8.S: Undef STACKFRAME before defining it.
1747 (ffi_closure_v8): Pass the structure return address. Update call to
1748 ffi_closure_sparc_inner_v8. Short-circuit FFI_TYPE_INT handling.
1749 Skip the 'unimp' insn when returning long doubles and structures.
1750 * src/sparc/v9.S: Undef STACKFRAME before defining it.
1751 (ffi_closure_v9): Increase the frame size by 2 words. Short-circuit
1752 FFI_TYPE_INT handling. Load structures both in integers and
1753 floating-point registers on return.
1754 * README: Update status of the SPARC port.
1755
1756 2004-01-24 Andreas Tobler <a.tobler@schweiz.ch>
1757
1758 * testsuite/libffi.call/pyobjc-tc.c (main): Treat result value
1759 as of type ffi_arg.
1760 * testsuite/libffi.call/struct3.c (main): Fix CHECK.
1761
1762 2004-01-22 Ulrich Weigand <uweigand@de.ibm.com>
1763
1764 * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Treat result
1765 value as of type ffi_arg, not unsigned int.
1766
1767 2004-01-21 Michael Ritzert <ritzert@t-online.de>
1768
1769 * ffi64.c (ffi_prep_args): Cast the RHS of an assignment instead
1770 of the LHS.
1771
1772 2004-01-12 Andreas Tobler <a.tobler@schweiz.ch>
1773
1774 * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_32 for
1775 Solaris.
1776
1777 2004-01-08 Rainer Orth <ro@TechFak.Uni-Bielefeld.DE>
1778
1779 * testsuite/libffi.call/ffitest.h (allocate_mmap): Cast MAP_FAILED
1780 to void *.
1781
1782 2003-12-10 Richard Henderson <rth@redhat.com>
1783
1784 * testsuite/libffi.call/cls_align_pointer.c: Cast pointers to
1785 size_t instead of int.
1786
1787 2003-12-04 Hosaka Yuji <hos@tamanegi.org>
1788
1789 * testsuite/libffi.call/many_win32.c: Include <float.h>.
1790 * testsuite/libffi.call/many_win32.c (main): Replace variable
1791 int i with unsigned long ul.
1792
1793 * testsuite/libffi.call/cls_align_uint64.c: New test case.
1794 * testsuite/libffi.call/cls_align_sint64.c: Likewise.
1795 * testsuite/libffi.call/cls_align_uint32.c: Likewise.
1796 * testsuite/libffi.call/cls_align_sint32.c: Likewise.
1797 * testsuite/libffi.call/cls_align_uint16.c: Likewise.
1798 * testsuite/libffi.call/cls_align_sint16.c: Likewise.
1799 * testsuite/libffi.call/cls_align_float.c: Likewise.
1800 * testsuite/libffi.call/cls_align_double.c: Likewise.
1801 * testsuite/libffi.call/cls_align_longdouble.c: Likewise.
1802 * testsuite/libffi.call/cls_align_pointer.c: Likewise.
1803
1804 2003-12-02 Hosaka Yuji <hos@tamanegi.org>
1805
1806 PR other/13221
1807 * src/x86/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV):
1808 Align arguments to 32 bits.
1809
1810 2003-12-01 Andreas Tobler <a.tobler@schweiz.ch>
1811
1812 PR other/13221
1813 * testsuite/libffi.call/cls_multi_sshort.c: New test case.
1814 * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise.
1815 * testsuite/libffi.call/cls_multi_uchar.c: Likewise.
1816 * testsuite/libffi.call/cls_multi_schar.c: Likewise.
1817 * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise.
1818 * testsuite/libffi.call/cls_multi_ushort.c: Likewise.
1819
1820 * testsuite/libffi.special/unwindtest.cc: Cosmetics.
1821
1822 2003-11-26 Kaveh R. Ghazi <ghazi@caip.rutgers.edu>
1823
1824 * testsuite/libffi.call/ffitest.h: Include <fcntl.h>.
1825 * testsuite/libffi.special/ffitestcxx.h: Likewise.
1826
1827 2003-11-22 Andreas Tobler <a.tobler@schweiz.ch>
1828
1829 * Makefile.in: Rebuilt.
1830 * configure: Likewise.
1831 * testsuite/libffi.special/unwindtest.cc: Convert the mmap to
1832 the right type.
1833
1834 2003-11-21 Andreas Jaeger <aj@suse.de>
1835 Andreas Tobler <a.tobler@schweiz.ch>
1836
1837 * acinclude.m4: Add AC_FUNC_MMAP_BLACKLIST.
1838 * configure.in: Call AC_FUNC_MMAP_BLACKLIST.
1839 * Makefile.in: Rebuilt.
1840 * aclocal.m4: Likewise.
1841 * configure: Likewise.
1842 * fficonfig.h.in: Likewise.
1843 * testsuite/lib/libffi-dg.exp: Add include dir.
1844 * testsuite/libffi.call/ffitest.h: Add MMAP definitions.
1845 * testsuite/libffi.special/ffitestcxx.h: Likewise.
1846 * testsuite/libffi.call/closure_fn0.c: Use MMAP functionality
1847 for ffi_closure if available.
1848 * testsuite/libffi.call/closure_fn1.c: Likewise.
1849 * testsuite/libffi.call/closure_fn2.c: Likewise.
1850 * testsuite/libffi.call/closure_fn3.c: Likewise.
1851 * testsuite/libffi.call/closure_fn4.c: Likewise.
1852 * testsuite/libffi.call/closure_fn5.c: Likewise.
1853 * testsuite/libffi.call/cls_12byte.c: Likewise.
1854 * testsuite/libffi.call/cls_16byte.c: Likewise.
1855 * testsuite/libffi.call/cls_18byte.c: Likewise.
1856 * testsuite/libffi.call/cls_19byte.c: Likewise.
1857 * testsuite/libffi.call/cls_1_1byte.c: Likewise.
1858 * testsuite/libffi.call/cls_20byte.c: Likewise.
1859 * testsuite/libffi.call/cls_20byte1.c: Likewise.
1860 * testsuite/libffi.call/cls_24byte.c: Likewise.
1861 * testsuite/libffi.call/cls_2byte.c: Likewise.
1862 * testsuite/libffi.call/cls_3_1byte.c: Likewise.
1863 * testsuite/libffi.call/cls_3byte1.c: Likewise.
1864 * testsuite/libffi.call/cls_3byte2.c: Likewise.
1865 * testsuite/libffi.call/cls_4_1byte.c: Likewise.
1866 * testsuite/libffi.call/cls_4byte.c: Likewise.
1867 * testsuite/libffi.call/cls_5byte.c: Likewise.
1868 * testsuite/libffi.call/cls_64byte.c: Likewise.
1869 * testsuite/libffi.call/cls_6byte.c: Likewise.
1870 * testsuite/libffi.call/cls_7byte.c: Likewise.
1871 * testsuite/libffi.call/cls_8byte.c: Likewise.
1872 * testsuite/libffi.call/cls_9byte1.c: Likewise.
1873 * testsuite/libffi.call/cls_9byte2.c: Likewise.
1874 * testsuite/libffi.call/cls_double.c: Likewise.
1875 * testsuite/libffi.call/cls_float.c: Likewise.
1876 * testsuite/libffi.call/cls_schar.c: Likewise.
1877 * testsuite/libffi.call/cls_sint.c: Likewise.
1878 * testsuite/libffi.call/cls_sshort.c: Likewise.
1879 * testsuite/libffi.call/cls_uchar.c: Likewise.
1880 * testsuite/libffi.call/cls_uint.c: Likewise.
1881 * testsuite/libffi.call/cls_ulonglong.c: Likewise.
1882 * testsuite/libffi.call/cls_ushort.c: Likewise.
1883 * testsuite/libffi.call/nested_struct.c: Likewise.
1884 * testsuite/libffi.call/nested_struct1.c: Likewise.
1885 * testsuite/libffi.call/nested_struct2.c: Likewise.
1886 * testsuite/libffi.call/nested_struct3.c: Likewise.
1887 * testsuite/libffi.call/problem1.c: Likewise.
1888 * testsuite/libffi.special/unwindtest.cc: Likewise.
1889
1890 2003-11-20 Andreas Tobler <a.tobler@schweiz.ch>
1891
1892 * testsuite/lib/libffi-dg.exp: Make the -lgcc_s conditional.
1893
1894 2003-11-19 Andreas Tobler <a.tobler@schweiz.ch>
1895
1896 * testsuite/lib/libffi-dg.exp: Add DYLD_LIBRARY_PATH for darwin.
1897 Add -lgcc_s to additional flags.
1898
1899 2003-11-12 Andreas Tobler <a.tobler@schweiz.ch>
1900
1901 * configure.in, include/Makefile.am: PR libgcj/11147, install
1902 the ffitarget.h header file in a gcc versioned and target
1903 dependent place.
1904 * configure: Regenerated.
1905 * Makefile.in, include/Makefile.in: Likewise.
1906 * testsuite/Makefile.in: Likewise.
1907
1908 2003-11-09 Andreas Tobler <a.tobler@schweiz.ch>
1909
1910 * testsuite/libffi.call/closure_fn0.c: Print result and check
1911 with dg-output to make debugging easier.
1912 * testsuite/libffi.call/closure_fn1.c: Likewise.
1913 * testsuite/libffi.call/closure_fn2.c: Likewise.
1914 * testsuite/libffi.call/closure_fn3.c: Likewise.
1915 * testsuite/libffi.call/closure_fn4.c: Likewise.
1916 * testsuite/libffi.call/closure_fn5.c: Likewise.
1917 * testsuite/libffi.call/cls_12byte.c: Likewise.
1918 * testsuite/libffi.call/cls_16byte.c: Likewise.
1919 * testsuite/libffi.call/cls_18byte.c: Likewise.
1920 * testsuite/libffi.call/cls_19byte.c: Likewise.
1921 * testsuite/libffi.call/cls_1_1byte.c: Likewise.
1922 * testsuite/libffi.call/cls_20byte.c: Likewise.
1923 * testsuite/libffi.call/cls_20byte1.c: Likewise.
1924 * testsuite/libffi.call/cls_24byte.c: Likewise.
1925 * testsuite/libffi.call/cls_2byte.c: Likewise.
1926 * testsuite/libffi.call/cls_3_1byte.c: Likewise.
1927 * testsuite/libffi.call/cls_3byte1.c: Likewise.
1928 * testsuite/libffi.call/cls_3byte2.c: Likewise.
1929 * testsuite/libffi.call/cls_4_1byte.c: Likewise.
1930 * testsuite/libffi.call/cls_4byte.c: Likewise.
1931 * testsuite/libffi.call/cls_5byte.c: Likewise.
1932 * testsuite/libffi.call/cls_64byte.c: Likewise.
1933 * testsuite/libffi.call/cls_6byte.c: Likewise.
1934 * testsuite/libffi.call/cls_7byte.c: Likewise.
1935 * testsuite/libffi.call/cls_8byte.c: Likewise.
1936 * testsuite/libffi.call/cls_9byte1.c: Likewise.
1937 * testsuite/libffi.call/cls_9byte2.c: Likewise.
1938 * testsuite/libffi.call/cls_double.c: Likewise.
1939 * testsuite/libffi.call/cls_float.c: Likewise.
1940 * testsuite/libffi.call/cls_schar.c: Likewise.
1941 * testsuite/libffi.call/cls_sint.c: Likewise.
1942 * testsuite/libffi.call/cls_sshort.c: Likewise.
1943 * testsuite/libffi.call/cls_uchar.c: Likewise.
1944 * testsuite/libffi.call/cls_uint.c: Likewise.
1945 * testsuite/libffi.call/cls_ulonglong.c: Likewise.
1946 * testsuite/libffi.call/cls_ushort.c: Likewise.
1947 * testsuite/libffi.call/problem1.c: Likewise.
1948
1949 * testsuite/libffi.special/unwindtest.cc: Make ffi_closure
1950 static.
1951
1952 2003-11-08 Andreas Tobler <a.tobler@schweiz.ch>
1953
1954 * testsuite/libffi.call/cls_9byte2.c: New test case.
1955 * testsuite/libffi.call/cls_9byte1.c: Likewise.
1956 * testsuite/libffi.call/cls_64byte.c: Likewise.
1957 * testsuite/libffi.call/cls_20byte1.c: Likewise.
1958 * testsuite/libffi.call/cls_19byte.c: Likewise.
1959 * testsuite/libffi.call/cls_18byte.c: Likewise.
1960 * testsuite/libffi.call/closure_fn4.c: Likewise.
1961 * testsuite/libffi.call/closure_fn5.c: Likewise.
1962 * testsuite/libffi.call/cls_schar.c: Likewise.
1963 * testsuite/libffi.call/cls_sint.c: Likewise.
1964 * testsuite/libffi.call/cls_sshort.c: Likewise.
1965 * testsuite/libffi.call/nested_struct2.c: Likewise.
1966 * testsuite/libffi.call/nested_struct3.c: Likewise.
1967
1968 2003-11-08 Andreas Tobler <a.tobler@schweiz.ch>
1969
1970 * testsuite/libffi.call/cls_double.c: Do a check on the result.
1971 * testsuite/libffi.call/cls_uchar.c: Likewise.
1972 * testsuite/libffi.call/cls_uint.c: Likewise.
1973 * testsuite/libffi.call/cls_ulonglong.c: Likewise.
1974 * testsuite/libffi.call/cls_ushort.c: Likewise.
1975 * testsuite/libffi.call/return_sc.c: Cleanup whitespaces.
1976
1977 2003-11-06 Andreas Tobler <a.tobler@schweiz.ch>
1978
1979 * src/prep_cif.c (ffi_prep_cif): Move the validity check after
1980 the initialization.
1981
1982 2003-10-23 Andreas Tobler <a.tobler@schweiz.ch>
1983
1984 * src/java_raw_api.c (ffi_java_ptrarray_to_raw): Replace
1985 FFI_ASSERT(FALSE) with FFI_ASSERT(0).
1986
1987 2003-10-22 David Daney <ddaney@avtrex.com>
1988
1989 * src/mips/ffitarget.h: Replace undefined UINT32 and friends with
1990 __attribute__((__mode__(__SI__))) and friends.
1991
1992 2003-10-22 Andreas Schwab <schwab@suse.de>
1993
1994 * src/ia64/ffi.c: Replace FALSE/TRUE with false/true.
1995
1996 2003-10-21 Andreas Tobler <a.tobler@schweiz.ch>
1997
1998 * configure.in: AC_LINK_FILES(ffitarget.h).
1999 * configure: Regenerate.
2000 * Makefile.in: Likewise.
2001 * include/Makefile.in: Likewise.
2002 * testsuite/Makefile.in: Likewise.
2003 * fficonfig.h.in: Likewise.
2004
2005 2003-10-21 Paolo Bonzini <bonzini@gnu.org>
2006 Richard Henderson <rth@redhat.com>
2007
2008 Avoid that ffi.h includes fficonfig.h.
2009
2010 * Makefile.am (EXTRA_DIST): Include ffitarget.h files
2011 (TARGET_SRC_MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
2012 (TARGET_SRC_MIPS_SGI): Removed.
2013 (MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
2014 (MIPS_SGI): Removed.
2015 (CLEANFILES): Removed.
2016 (mostlyclean-am, clean-am, mostlyclean-sub, clean-sub): New
2017 targets.
2018 * acconfig.h: Removed.
2019 * configure.in: Compute sizeofs only for double and long double.
2020 Use them to define and subst HAVE_LONG_DOUBLE. Include comments
2021 into AC_DEFINE instead of using acconfig.h. Create
2022 include/ffitarget.h instead of include/fficonfig.h. Rename
2023 MIPS_GCC to MIPS_IRIX, drop MIPS_SGI since we are in gcc's tree.
2024 AC_DEFINE EH_FRAME_FLAGS.
2025 * include/Makefile.am (DISTCLEANFILES): New automake macro.
2026 (hack_DATA): Add ffitarget.h.
2027 * include/ffi.h.in: Remove all system specific definitions.
2028 Declare raw API even if it is not installed, why bother?
2029 Use limits.h instead of SIZEOF_* to define ffi_type_*. Do
2030 not define EH_FRAME_FLAGS, it is in fficonfig.h now. Include
2031 ffitarget.h instead of fficonfig.h. Remove ALIGN macro.
2032 (UINT_ARG, INT_ARG): Removed, use ffi_arg and ffi_sarg instead.
2033 * include/ffi_common.h (bool): Do not define.
2034 (ffi_assert): Accept failed assertion.
2035 (ffi_type_test): Return void and accept file/line.
2036 (FFI_ASSERT): Pass stringized failed assertion.
2037 (FFI_ASSERT_AT): New macro.
2038 (FFI_ASSERT_VALID_TYPE): New macro.
2039 (UINT8, SINT8, UINT16, SINT16, UINT32, SINT32,
2040 UINT64, SINT64): Define here with gcc's __attribute__ macro
2041 instead of in ffi.h
2042 (FLOAT32, ALIGN): Define here instead of in ffi.h
2043 * include/ffi-mips.h: Removed. Its content moved to
2044 src/mips/ffitarget.h after separating assembly and C sections.
2045 * src/alpha/ffi.c, src/alpha/ffi.c, src/java_raw_api.c
2046 src/prep_cif.c, src/raw_api.c, src/ia64/ffi.c,
2047 src/mips/ffi.c, src/mips/n32.S, src/mips/o32.S,
2048 src/mips/ffitarget.h, src/sparc/ffi.c, src/x86/ffi64.c:
2049 SIZEOF_ARG -> FFI_SIZEOF_ARG.
2050 * src/ia64/ffi.c: Include stdbool.h (provided by GCC 2.95+).
2051 * src/debug.c (ffi_assert): Accept stringized failed assertion.
2052 (ffi_type_test): Rewritten.
2053 * src/prep-cif.c (initialize_aggregate, ffi_prep_cif): Call
2054 FFI_ASSERT_VALID_TYPE.
2055 * src/alpha/ffitarget.h, src/arm/ffitarget.h,
2056 src/ia64/ffitarget.h, src/m68k/ffitarget.h,
2057 src/mips/ffitarget.h, src/powerpc/ffitarget.h,
2058 src/s390/ffitarget.h, src/sh/ffitarget.h,
2059 src/sh64/ffitarget.h, src/sparc/ffitarget.h,
2060 src/x86/ffitarget.h: New files.
2061 * src/alpha/osf.S, src/arm/sysv.S, src/ia64/unix.S,
2062 src/m68k/sysv.S, src/mips/n32.S, src/mips/o32.S,
2063 src/powerpc/aix.S, src/powerpc/darwin.S,
2064 src/powerpc/ffi_darwin.c, src/powerpc/linux64.S,
2065 src/powerpc/linux64_closure.S, src/powerpc/ppc_closure.S,
2066 src/powerpc/sysv.S, src/s390/sysv.S, src/sh/sysv.S,
2067 src/sh64/sysv.S, src/sparc/v8.S, src/sparc/v9.S,
2068 src/x86/sysv.S, src/x86/unix64.S, src/x86/win32.S:
2069 include fficonfig.h
2070
2071 2003-10-20 Rainer Orth <ro@TechFak.Uni-Bielefeld.DE>
2072
2073 * src/mips/ffi.c: Use _ABIN32, _ABIO32 instead of external
2074 _MIPS_SIM_NABI32, _MIPS_SIM_ABI32.
2075
2076 2003-10-19 Andreas Tobler <a.tobler@schweiz.ch>
2077
2078 * src/powerpc/ffi_darwin.c (ffi_prep_args): Declare bytes again.
2079 Used when FFI_DEBUG = 1.
2080
2081 2003-10-14 Alan Modra <amodra@bigpond.net.au>
2082
2083 * src/types.c (double, longdouble): Default POWERPC64 to 8 byte size
2084 and align.
2085
2086 2003-10-06 Rainer Orth <ro@TechFak.Uni-Bielefeld.DE>
2087
2088 * include/ffi_mips.h: Define FFI_MIPS_N32 for N32/N64 ABIs,
2089 FFI_MIPS_O32 for O32 ABI.
2090
2091 2003-10-01 Andreas Tobler <a.tobler@schweiz.ch>
2092
2093 * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_64 for
2094 SPARC64. Cleanup whitespaces.
2095
2096 2003-09-19 Andreas Tobler <a.tobler@schweiz.ch>
2097
2098 * testsuite/libffi.call/closure_fn0.c: Xfail mips, arm,
2099 strongarm, xscale. Cleanup whitespaces.
2100 * testsuite/libffi.call/closure_fn1.c: Likewise.
2101 * testsuite/libffi.call/closure_fn2.c: Likewise.
2102 * testsuite/libffi.call/closure_fn3.c: Likewise.
2103 * testsuite/libffi.call/cls_12byte.c: Likewise.
2104 * testsuite/libffi.call/cls_16byte.c: Likewise.
2105 * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2106 * testsuite/libffi.call/cls_20byte.c: Likewise.
2107 * testsuite/libffi.call/cls_24byte.c: Likewise.
2108 * testsuite/libffi.call/cls_2byte.c: Likewise.
2109 * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2110 * testsuite/libffi.call/cls_3byte1.c: Likewise.
2111 * testsuite/libffi.call/cls_3byte2.c: Likewise.
2112 * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2113 * testsuite/libffi.call/cls_4byte.c: Likewise.
2114 * testsuite/libffi.call/cls_5byte.c: Likewise.
2115 * testsuite/libffi.call/cls_6byte.c: Likewise.
2116 * testsuite/libffi.call/cls_7byte.c: Likewise.
2117 * testsuite/libffi.call/cls_8byte.c: Likewise.
2118 * testsuite/libffi.call/cls_double.c: Likewise.
2119 * testsuite/libffi.call/cls_float.c: Likewise.
2120 * testsuite/libffi.call/cls_uchar.c: Likewise.
2121 * testsuite/libffi.call/cls_uint.c: Likewise.
2122 * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2123 * testsuite/libffi.call/cls_ushort.c: Likewise.
2124 * testsuite/libffi.call/nested_struct.c: Likewise.
2125 * testsuite/libffi.call/nested_struct1.c: Likewise.
2126 * testsuite/libffi.call/problem1.c: Likewise.
2127 * testsuite/libffi.special/unwindtest.cc: Likewise.
2128 * testsuite/libffi.call/pyobjc-tc.c: Cleanup whitespaces.
2129
2130 2003-09-18 David Edelsohn <edelsohn@gnu.org>
2131
2132 * src/powerpc/aix.S: Cleanup whitespaces.
2133 * src/powerpc/aix_closure.S: Likewise.
2134
2135 2003-09-18 Andreas Tobler <a.tobler@schweiz.ch>
2136
2137 * src/powerpc/darwin.S: Cleanup whitespaces, comment formatting.
2138 * src/powerpc/darwin_closure.S: Likewise.
2139 * src/powerpc/ffi_darwin.c: Likewise.
2140
2141 2003-09-18 Andreas Tobler <a.tobler@schweiz.ch>
2142 David Edelsohn <edelsohn@gnu.org>
2143
2144 * src/types.c (double): Add AIX and Darwin to the right TYPEDEF.
2145 * src/powerpc/aix_closure.S: Remove the pointer to the outgoing
2146 parameter stack.
2147 * src/powerpc/darwin_closure.S: Likewise.
2148 * src/powerpc/ffi_darwin.c (ffi_prep_args): Handle structures
2149 according to the Darwin/AIX ABI.
2150 (ffi_prep_cif_machdep): Likewise.
2151 (ffi_closure_helper_DARWIN): Likewise.
2152 Remove the outgoing parameter stack logic. Simplify the evaluation
2153 of the different CASE types.
2154 (ffi_prep_clousure): Avoid the casts on lvalues. Change the branch
2155 statement in the trampoline code.
2156
2157 2003-09-18 Kaz Kojima <kkojima@gcc.gnu.org>
2158
2159 * src/sh/ffi.c (ffi_prep_args): Take account into the alignement
2160 for the register size.
2161 (ffi_closure_helper_SYSV): Handle the structure return value
2162 address correctly.
2163 (ffi_closure_helper_SYSV): Return the appropriate type when
2164 the registers are used for the structure return value.
2165 * src/sh/sysv.S (ffi_closure_SYSV): Fix the stack layout for
2166 the 64-bit return value. Update copyright years.
2167
2168 2003-09-17 Rainer Orth <ro@TechFak.Uni-Bielefeld.DE>
2169
2170 * testsuite/lib/libffi-dg.exp (libffi_target_compile): Search in
2171 srcdir for ffi_mips.h.
2172
2173 2003-09-12 Alan Modra <amodra@bigpond.net.au>
2174
2175 * src/prep_cif.c (initialize_aggregate): Include tail padding in
2176 structure size.
2177 * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Correct
2178 placement of float result.
2179 * testsuite/libffi.special/unwindtest.cc (closure_test_fn1): Correct
2180 cast of "resp" for big-endian 64 bit machines.
2181
2182 2003-09-11 Alan Modra <amodra@bigpond.net.au>
2183
2184 * src/types.c (double, longdouble): Merge identical SH and ARM
2185 typedefs, and add POWERPC64.
2186 * src/powerpc/ffi.c (ffi_prep_args64): Correct next_arg calc for
2187 struct split over gpr and rest.
2188 (ffi_prep_cif_machdep): Correct intarg_count for structures.
2189 * src/powerpc/linux64.S (ffi_call_LINUX64): Fix gpr offsets.
2190
2191 2003-09-09 Andreas Tobler <a.tobler@schweiz.ch>
2192
2193 * src/powerpc/ffi.c (ffi_closure_helper_SYSV) Handle struct
2194 passing correctly.
2195
2196 2003-09-09 Alan Modra <amodra@bigpond.net.au>
2197
2198 * configure: Regenerate.
2199
2200 2003-09-04 Andreas Tobler <a.tobler@schweiz.ch>
2201
2202 * Makefile.am: Remove build rules for ffitest.
2203 * Makefile.in: Rebuilt.
2204
2205 2003-09-04 Andreas Tobler <a.tobler@schweiz.ch>
2206
2207 * src/java_raw_api.c: Include <stdlib.h> to fix compiler warning
2208 about implicit declaration of abort().
2209
2210 2003-09-04 Andreas Tobler <a.tobler@schweiz.ch>
2211
2212 * Makefile.am: Add dejagnu test framework. Fixes PR other/11411.
2213 * Makefile.in: Rebuilt.
2214 * configure.in: Add dejagnu test framework.
2215 * configure: Rebuilt.
2216
2217 * testsuite/Makefile.am: New file.
2218 * testsuite/Makefile.in: Built
2219 * testsuite/lib/libffi-dg.exp: New file.
2220 * testsuite/config/default.exp: Likewise.
2221 * testsuite/libffi.call/call.exp: Likewise.
2222 * testsuite/libffi.call/ffitest.h: Likewise.
2223 * testsuite/libffi.call/closure_fn0.c: Likewise.
2224 * testsuite/libffi.call/closure_fn1.c: Likewise.
2225 * testsuite/libffi.call/closure_fn2.c: Likewise.
2226 * testsuite/libffi.call/closure_fn3.c: Likewise.
2227 * testsuite/libffi.call/cls_1_1byte.c: Likewise.
2228 * testsuite/libffi.call/cls_3_1byte.c: Likewise.
2229 * testsuite/libffi.call/cls_4_1byte.c: Likewise.
2230 * testsuite/libffi.call/cls_2byte.c: Likewise.
2231 * testsuite/libffi.call/cls_3byte1.c: Likewise.
2232 * testsuite/libffi.call/cls_3byte2.c: Likewise.
2233 * testsuite/libffi.call/cls_4byte.c: Likewise.
2234 * testsuite/libffi.call/cls_5byte.c: Likewise.
2235 * testsuite/libffi.call/cls_6byte.c: Likewise.
2236 * testsuite/libffi.call/cls_7byte.c: Likewise.
2237 * testsuite/libffi.call/cls_8byte.c: Likewise.
2238 * testsuite/libffi.call/cls_12byte.c: Likewise.
2239 * testsuite/libffi.call/cls_16byte.c: Likewise.
2240 * testsuite/libffi.call/cls_20byte.c: Likewise.
2241 * testsuite/libffi.call/cls_24byte.c: Likewise.
2242 * testsuite/libffi.call/cls_double.c: Likewise.
2243 * testsuite/libffi.call/cls_float.c: Likewise.
2244 * testsuite/libffi.call/cls_uchar.c: Likewise.
2245 * testsuite/libffi.call/cls_uint.c: Likewise.
2246 * testsuite/libffi.call/cls_ulonglong.c: Likewise.
2247 * testsuite/libffi.call/cls_ushort.c: Likewise.
2248 * testsuite/libffi.call/float.c: Likewise.
2249 * testsuite/libffi.call/float1.c: Likewise.
2250 * testsuite/libffi.call/float2.c: Likewise.
2251 * testsuite/libffi.call/many.c: Likewise.
2252 * testsuite/libffi.call/many_win32.c: Likewise.
2253 * testsuite/libffi.call/nested_struct.c: Likewise.
2254 * testsuite/libffi.call/nested_struct1.c: Likewise.
2255 * testsuite/libffi.call/pyobjc-tc.c: Likewise.
2256 * testsuite/libffi.call/problem1.c: Likewise.
2257 * testsuite/libffi.call/promotion.c: Likewise.
2258 * testsuite/libffi.call/return_ll.c: Likewise.
2259 * testsuite/libffi.call/return_sc.c: Likewise.
2260 * testsuite/libffi.call/return_uc.c: Likewise.
2261 * testsuite/libffi.call/strlen.c: Likewise.
2262 * testsuite/libffi.call/strlen_win32.c: Likewise.
2263 * testsuite/libffi.call/struct1.c: Likewise.
2264 * testsuite/libffi.call/struct2.c: Likewise.
2265 * testsuite/libffi.call/struct3.c: Likewise.
2266 * testsuite/libffi.call/struct4.c: Likewise.
2267 * testsuite/libffi.call/struct5.c: Likewise.
2268 * testsuite/libffi.call/struct6.c: Likewise.
2269 * testsuite/libffi.call/struct7.c: Likewise.
2270 * testsuite/libffi.call/struct8.c: Likewise.
2271 * testsuite/libffi.call/struct9.c: Likewise.
2272 * testsuite/libffi.special/special.exp: New file.
2273 * testsuite/libffi.special/ffitestcxx.h: Likewise.
2274 * testsuite/libffi.special/unwindtest.cc: Likewise.
2275
2276
2277 2003-08-13 Kaz Kojima <kkojima@gcc.gnu.org>
2278
2279 * src/sh/ffi.c (OFS_INT16): Set 0 for little endian case. Update
2280 copyright years.
2281
2282 2003-08-02 Alan Modra <amodra@bigpond.net.au>
2283
2284 * src/powerpc/ffi.c (ffi_prep_args64): Modify for changed gcc
2285 structure passing.
2286 (ffi_closure_helper_LINUX64): Likewise.
2287 * src/powerpc/linux64.S: Remove code writing to parm save area.
2288 * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Use return
2289 address in lr from ffi_closure_helper_LINUX64 call to calculate
2290 table address. Optimize function tail.
2291
2292 2003-07-28 Andreas Tobler <a.tobler@schweiz.ch>
2293
2294 * src/sparc/ffi.c: Handle all floating point registers.
2295 * src/sparc/v9.S: Likewise. Fixes second part of PR target/11410.
2296
2297 2003-07-11 Gerald Pfeifer <pfeifer@dbai.tuwien.ac.at>
2298
2299 * README: Note that libffi is not part of GCC. Update the project
2300 URL and status.
2301
2302 2003-06-19 Franz Sirl <Franz.Sirl-kernel@lauterbach.com>
2303
2304 * src/powerpc/ppc_closure.S: Include ffi.h.
2305
2306 2003-06-13 Rainer Orth <ro@TechFak.Uni-Bielefeld.DE>
2307
2308 * src/x86/sysv.S: Avoid gas-only .uleb128/.sleb128 directives.
2309 Use C style comments.
2310
2311 2003-06-13 Kaz Kojima <kkojima@rr.iij4u.or.jp>
2312
2313 * Makefile.am: Add SHmedia support. Fix a typo of SH support.
2314 * Makefile.in: Regenerate.
2315 * configure.in (sh64-*-linux*, sh5*-*-linux*): Add target.
2316 * configure: Regenerate.
2317 * include/ffi.h.in: Add SHmedia support.
2318 * src/sh64/ffi.c: New file.
2319 * src/sh64/sysv.S: New file.
2320
2321 2003-05-16 Jakub Jelinek <jakub@redhat.com>
2322
2323 * configure.in (HAVE_RO_EH_FRAME): Check whether .eh_frame section
2324 should be read-only.
2325 * configure: Rebuilt.
2326 * fficonfig.h.in: Rebuilt.
2327 * include/ffi.h.in (EH_FRAME_FLAGS): Define.
2328 * src/alpha/osf.S: Use EH_FRAME_FLAGS.
2329 * src/powerpc/linux64.S: Likewise.
2330 * src/powerpc/linux64_closure.S: Likewise. Include ffi.h.
2331 * src/powerpc/sysv.S: Use EH_FRAME_FLAGS. Use pcrel encoding
2332 if -fpic/-fPIC/-mrelocatable.
2333 * src/powerpc/powerpc_closure.S: Likewise.
2334 * src/sparc/v8.S: If HAVE_RO_EH_FRAME is defined, don't include
2335 #write in .eh_frame flags.
2336 * src/sparc/v9.S: Likewise.
2337 * src/x86/unix64.S: Use EH_FRAME_FLAGS.
2338 * src/x86/sysv.S: Likewise. Use pcrel encoding if -fpic/-fPIC.
2339 * src/s390/sysv.S: Use EH_FRAME_FLAGS. Include ffi.h.
2340
2341 2003-05-07 Jeff Sturm <jsturm@one-point.com>
2342
2343 Fixes PR bootstrap/10656
2344 * configure.in (HAVE_AS_REGISTER_PSEUDO_OP): Test assembler
2345 support for .register pseudo-op.
2346 * src/sparc/v8.S: Use it.
2347 * fficonfig.h.in: Rebuilt.
2348 * configure: Rebuilt.
2349
2350 2003-04-18 Jakub Jelinek <jakub@redhat.com>
2351
2352 * include/ffi.h.in (POWERPC64): Define if 64-bit.
2353 (enum ffi_abi): Add FFI_LINUX64 on POWERPC.
2354 Make it the default on POWERPC64.
2355 (FFI_TRAMPOLINE_SIZE): Define to 24 on POWERPC64.
2356 * configure.in: Change powerpc-*-linux* into powerpc*-*-linux*.
2357 * configure: Rebuilt.
2358 * src/powerpc/ffi.c (hidden): Define.
2359 (ffi_prep_args_SYSV): Renamed from
2360 ffi_prep_args. Cast pointers to unsigned long to shut up warnings.
2361 (NUM_GPR_ARG_REGISTERS64, NUM_FPR_ARG_REGISTERS64,
2362 ASM_NEEDS_REGISTERS64): New.
2363 (ffi_prep_args64): New function.
2364 (ffi_prep_cif_machdep): Handle FFI_LINUX64 ABI.
2365 (ffi_call): Likewise.
2366 (ffi_prep_closure): Likewise.
2367 (flush_icache): Surround by #ifndef POWERPC64.
2368 (ffi_dblfl): New union type.
2369 (ffi_closure_helper_SYSV): Use it to avoid aliasing problems.
2370 (ffi_closure_helper_LINUX64): New function.
2371 * src/powerpc/ppc_closure.S: Surround whole file by #ifndef
2372 __powerpc64__.
2373 * src/powerpc/sysv.S: Likewise.
2374 (ffi_call_SYSV): Rename ffi_prep_args to ffi_prep_args_SYSV.
2375 * src/powerpc/linux64.S: New file.
2376 * src/powerpc/linux64_closure.S: New file.
2377 * Makefile.am (EXTRA_DIST): Add src/powerpc/linux64.S and
2378 src/powerpc/linux64_closure.S.
2379 (TARGET_SRC_POWERPC): Likewise.
2380
2381 * src/ffitest.c (closure_test_fn, closure_test_fn1, closure_test_fn2,
2382 closure_test_fn3): Fix result printing on big-endian 64-bit
2383 machines.
2384 (main): Print tst2_arg instead of uninitialized tst2_result.
2385
2386 * src/ffitest.c (main): Hide what closure pointer really points to
2387 from the compiler.
2388
2389 2003-04-16 Richard Earnshaw <rearnsha@arm.com>
2390
2391 * configure.in (arm-*-netbsdelf*): Add configuration.
2392 (configure): Regenerated.
2393
2394 2003-04-04 Loren J. Rittle <ljrittle@acm.org>
2395
2396 * include/Makefile.in: Regenerate.
2397
2398 2003-03-21 Zdenek Dvorak <rakdver@atrey.karlin.mff.cuni.cz>
2399
2400 * libffi/include/ffi.h.in: Define X86 instead of X86_64 in 32
2401 bit mode.
2402 * libffi/src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV):
2403 Receive closure pointer through parameter, read args using
2404 __builtin_dwarf_cfa.
2405 (FFI_INIT_TRAMPOLINE): Send closure reference through eax.
2406
2407 2003-03-12 Andreas Schwab <schwab@suse.de>
2408
2409 * configure.in: Avoid trailing /. in toolexeclibdir.
2410 * configure: Rebuilt.
2411
2412 2003-03-03 Andreas Tobler <a.tobler@schweiz.ch>
2413
2414 * src/powerpc/darwin_closure.S: Recode to fit dynamic libraries.
2415
2416 2003-02-06 Andreas Tobler <a.tobler@schweiz.ch>
2417
2418 * libffi/src/powerpc/darwin_closure.S:
2419 Fix alignement bug, allocate 8 bytes for the result.
2420 * libffi/src/powerpc/aix_closure.S:
2421 Likewise.
2422 * libffi/src/powerpc/ffi_darwin.c:
2423 Update stackframe description for aix/darwin_closure.S.
2424
2425 2003-02-06 Jakub Jelinek <jakub@redhat.com>
2426
2427 * src/s390/ffi.c (ffi_closure_helper_SYSV): Add hidden visibility
2428 attribute.
2429
2430 2003-01-31 Christian Cornelssen <ccorn@cs.tu-berlin.de>,
2431 Andreas Schwab <schwab@suse.de>
2432
2433 * configure.in: Adjust command to source config-ml.in to account
2434 for changes to the libffi_basedir definition.
2435 (libffi_basedir): Remove ${srcdir} from value and include trailing
2436 slash if nonempty.
2437
2438 * configure: Regenerate.
2439
2440 2003-01-29 Franz Sirl <Franz.Sirl-kernel@lauterbach.com>
2441
2442 * src/powerpc/ppc_closure.S: Recode to fit shared libs.
2443
2444 2003-01-28 Andrew Haley <aph@redhat.com>
2445
2446 * include/ffi.h.in: Enable FFI_CLOSURES for x86_64.
2447 * src/x86/ffi64.c (ffi_prep_closure): New.
2448 (ffi_closure_UNIX64_inner): New.
2449 * src/x86/unix64.S (ffi_closure_UNIX64): New.
2450
2451 2003-01-27 Alexandre Oliva <aoliva@redhat.com>
2452
2453 * configure.in (toolexecdir, toolexeclibdir): Set and AC_SUBST.
2454 Remove USE_LIBDIR conditional.
2455 * Makefile.am (toolexecdir, toolexeclibdir): Don't override.
2456 * Makefile.in, configure: Rebuilt.
2457
2458 2003-01027 David Edelsohn <edelsohn@gnu.org>
2459
2460 * Makefile.am (TARGET_SRC_POWERPC_AIX): Fix typo.
2461 * Makefile.in: Regenerate.
2462
2463 2003-01-22 Andrew Haley <aph@redhat.com>
2464
2465 * src/powerpc/darwin.S (_ffi_call_AIX): Add Augmentation size to
2466 unwind info.
2467
2468 2003-01-21 Andreas Tobler <a.tobler@schweiz.ch>
2469
2470 * src/powerpc/darwin.S: Add unwind info.
2471 * src/powerpc/darwin_closure.S: Likewise.
2472
2473 2003-01-14 Andrew Haley <aph@redhat.com>
2474
2475 * src/x86/ffi64.c (ffi_prep_args): Check for void retval.
2476 (ffi_prep_cif_machdep): Likewise.
2477 * src/x86/unix64.S: Add unwind info.
2478
2479 2003-01-14 Andreas Jaeger <aj@suse.de>
2480
2481 * src/ffitest.c (main): Only use ffi_closures if those are
2482 supported.
2483
2484 2003-01-13 Andreas Tobler <a.tobler@schweiz.ch>
2485
2486 * libffi/src/ffitest.c
2487 add closure testcases
2488
2489 2003-01-13 Kevin B. Hendricks <khendricks@ivey.uwo.ca>
2490
2491 * libffi/src/powerpc/ffi.c
2492 fix alignment bug for float (4 byte aligned iso 8 byte)
2493
2494 2003-01-09 Geoffrey Keating <geoffk@apple.com>
2495
2496 * src/powerpc/ffi_darwin.c: Remove RCS version string.
2497 * src/powerpc/darwin.S: Remove RCS version string.
2498
2499 2003-01-03 Jeff Sturm <jsturm@one-point.com>
2500
2501 * include/ffi.h.in: Add closure defines for SPARC, SPARC64.
2502 * src/ffitest.c (main): Use static storage for closure.
2503 * src/sparc/ffi.c (ffi_prep_closure, ffi_closure_sparc_inner): New.
2504 * src/sparc/v8.S (ffi_closure_v8): New.
2505 * src/sparc/v9.S (ffi_closure_v9): New.
2506
2507 2002-11-10 Ranjit Mathew <rmathew@hotmail.com>
2508
2509 * include/ffi.h.in: Added FFI_STDCALL ffi_type
2510 enumeration for X86_WIN32.
2511 * src/x86/win32.S: Added ffi_call_STDCALL function
2512 definition.
2513 * src/x86/ffi.c (ffi_call/ffi_raw_call): Added
2514 switch cases for recognising FFI_STDCALL and
2515 calling ffi_call_STDCALL if target is X86_WIN32.
2516 * src/ffitest.c (my_stdcall_strlen/stdcall_many):
2517 stdcall versions of the "my_strlen" and "many"
2518 test functions (for X86_WIN32).
2519 Added test cases to test stdcall invocation using
2520 these functions.
2521
2522 2002-12-02 Kaz Kojima <kkojima@gcc.gnu.org>
2523
2524 * src/sh/sysv.S: Add DWARF2 unwind info.
2525
2526 2002-11-27 Ulrich Weigand <uweigand@de.ibm.com>
2527
2528 * src/s390/sysv.S (.eh_frame section): Make section read-only.
2529
2530 2002-11-26 Jim Wilson <wilson@redhat.com>
2531
2532 * src/types.c (FFI_TYPE_POINTER): Has size 8 on IA64.
2533
2534 2002-11-23 H.J. Lu <hjl@gnu.org>
2535
2536 * acinclude.m4: Add dummy AM_PROG_LIBTOOL.
2537 Include ../config/accross.m4.
2538 * aclocal.m4; Rebuild.
2539 * configure: Likewise.
2540
2541 2002-11-15 Ulrich Weigand <uweigand@de.ibm.com>
2542
2543 * src/s390/sysv.S (.eh_frame section): Adapt to pcrel FDE encoding.
2544
2545 2002-11-11 DJ Delorie <dj@redhat.com>
2546
2547 * configure.in: Look for common files in the right place.
2548
2549 2002-10-08 Ulrich Weigand <uweigand@de.ibm.com>
2550
2551 * src/java_raw_api.c (ffi_java_raw_to_ptrarray): Interpret
2552 raw data as _Jv_word values, not ffi_raw.
2553 (ffi_java_ptrarray_to_raw): Likewise.
2554 (ffi_java_rvalue_to_raw): New function.
2555 (ffi_java_raw_call): Call it.
2556 (ffi_java_raw_to_rvalue): New function.
2557 (ffi_java_translate_args): Call it.
2558 * src/ffitest.c (closure_test_fn): Interpret return value
2559 as ffi_arg, not int.
2560 * src/s390/ffi.c (ffi_prep_cif_machdep): Add missing
2561 FFI_TYPE_POINTER case.
2562 (ffi_closure_helper_SYSV): Likewise. Also, assume return
2563 values extended to word size.
2564
2565 2002-10-02 Andreas Jaeger <aj@suse.de>
2566
2567 * src/x86/ffi64.c (ffi_prep_cif_machdep): Remove debug output.
2568
2569 2002-10-01 Bo Thorsen <bo@smetana.suse.de>
2570
2571 * include/ffi.h.in: Fix i386 win32 compilation.
2572
2573 2002-09-30 Ulrich Weigand <uweigand@de.ibm.com>
2574
2575 * configure.in: Add s390x-*-linux-* target.
2576 * configure: Regenerate.
2577 * include/ffi.h.in: Define S390X for s390x targets.
2578 (FFI_CLOSURES): Define for s390/s390x.
2579 (FFI_TRAMPOLINE_SIZE): Likewise.
2580 (FFI_NATIVE_RAW_API): Likewise.
2581 * src/prep_cif.c (ffi_prep_cif): Do not compute stack space for s390.
2582 * src/types.c (FFI_TYPE_POINTER): Use 8-byte pointers on s390x.
2583 * src/s390/ffi.c: Major rework of existing code. Add support for
2584 s390x targets. Add closure support.
2585 * src/s390/sysv.S: Likewise.
2586
2587 2002-09-29 Richard Earnshaw <rearnsha@arm.com>
2588
2589 * src/arm/sysv.S: Fix typo.
2590
2591 2002-09-28 Richard Earnshaw <rearnsha@arm.com>
2592
2593 * src/arm/sysv.S: If we don't have machine/asm.h and the pre-processor
2594 has defined __USER_LABEL_PREFIX__, then use it in CNAME.
2595 (ffi_call_SYSV): Handle soft-float.
2596
2597 2002-09-27 Bo Thorsen <bo@suse.de>
2598
2599 * include/ffi.h.in: Fix multilib x86-64 support.
2600
2601 2002-09-22 Kaveh R. Ghazi <ghazi@caip.rutgers.edu>
2602
2603 * Makefile.am (all-multi): Fix multilib parallel build.
2604
2605 2002-07-19 Kaz Kojima <kkojima@gcc.gnu.org>
2606
2607 * configure.in (sh[34]*-*-linux*): Add brackets.
2608 * configure: Regenerate.
2609
2610 2002-07-18 Kaz Kojima <kkojima@gcc.gnu.org>
2611
2612 * Makefile.am: Add SH support.
2613 * Makefile.in: Regenerate.
2614 * configure.in (sh-*-linux*, sh[34]*-*-linux*): Add target.
2615 * configure: Regenerate.
2616 * include/ffi.h.in: Add SH support.
2617 * src/sh/ffi.c: New file.
2618 * src/sh/sysv.S: New file.
2619 * src/types.c: Add SH support.
2620
2621 2002-07-16 Bo Thorsen <bo@suse.de>
2622
2623 * src/x86/ffi64.c: New file that adds x86-64 support.
2624 * src/x86/unix64.S: New file that handles argument setup for
2625 x86-64.
2626 * src/x86/sysv.S: Don't use this on x86-64.
2627 * src/x86/ffi.c: Don't use this on x86-64.
2628 Remove unused vars.
2629 * src/prep_cif.c (ffi_prep_cif): Don't do stack size calculation
2630 for x86-64.
2631 * src/ffitest.c (struct6): New test that tests a special case in
2632 the x86-64 ABI.
2633 (struct7): Likewise.
2634 (struct8): Likewise.
2635 (struct9): Likewise.
2636 (closure_test_fn): Silence warning about this when it's not used.
2637 (main): Add the new tests.
2638 (main): Fix a couple of wrong casts and silence some compiler warnings.
2639 * include/ffi.h.in: Add x86-64 ABI definition.
2640 * fficonfig.h.in: Regenerate.
2641 * Makefile.am: Add x86-64 support.
2642 * configure.in: Likewise.
2643 * Makefile.in: Regenerate.
2644 * configure: Likewise.
2645
2646 2002-06-24 Bo Thorsen <bo@suse.de>
2647
2648 * src/types.c: Merge settings for similar architectures.
2649 Add x86-64 sizes and alignments.
2650
2651 2002-06-23 Bo Thorsen <bo@suse.de>
2652
2653 * src/arm/ffi.c (ffi_prep_args): Remove unused vars.
2654 * src/sparc/ffi.c (ffi_prep_args_v8): Likewise.
2655 * src/mips/ffi.c (ffi_prep_args): Likewise.
2656 * src/m68k/ffi.c (ffi_prep_args): Likewise.
2657
2658 2002-07-18 H.J. Lu (hjl@gnu.org)
2659
2660 * Makefile.am (TARGET_SRC_MIPS_LINUX): New.
2661 (libffi_la_SOURCES): Support MIPS_LINUX.
2662 (libffi_convenience_la_SOURCES): Likewise.
2663 * Makefile.in: Regenerated.
2664
2665 * configure.in (mips64*-*): Skip.
2666 (mips*-*-linux*): New.
2667 * configure: Regenerated.
2668
2669 * src/mips/ffi.c: Include <sgidefs.h>.
2670
2671 2002-06-06 Ulrich Weigand <uweigand@de.ibm.com>
2672
2673 * src/s390/sysv.S: Save/restore %r6. Add DWARF-2 unwind info.
2674
2675 2002-05-27 Roger Sayle <roger@eyesopen.com>
2676
2677 * src/x86/ffi.c (ffi_prep_args): Remove reference to avn.
2678
2679 2002-05-27 Bo Thorsen <bo@suse.de>
2680
2681 * src/x86/ffi.c (ffi_prep_args): Remove unused variable and
2682 fix formatting.
2683
2684 2002-05-13 Andreas Tobler <a.tobler@schweiz.ch>
2685
2686 * src/powerpc/ffi_darwin.c (ffi_prep_closure): Declare fd at
2687 beginning of function (for older apple cc).
2688
2689 2002-05-08 Alexandre Oliva <aoliva@redhat.com>
2690
2691 * configure.in (ORIGINAL_LD_FOR_MULTILIBS): Preserve LD at
2692 script entry, and set LD to it when configuring multilibs.
2693 * configure: Rebuilt.
2694
2695 2002-05-05 Jason Thorpe <thorpej@wasabisystems.com>
2696
2697 * configure.in (sparc64-*-netbsd*): Add target.
2698 (sparc-*-netbsdelf*): Likewise.
2699 * configure: Regenerate.
2700
2701 2002-04-28 David S. Miller <davem@redhat.com>
2702
2703 * configure.in, configure: Fix SPARC test in previous change.
2704
2705 2002-04-29 Gerhard Tonn <GerhardTonn@swol.de>
2706
2707 * Makefile.am: Add Linux for S/390 support.
2708 * Makefile.in: Regenerate.
2709 * configure.in: Add Linux for S/390 support.
2710 * configure: Regenerate.
2711 * include/ffi.h.in: Add Linux for S/390 support.
2712 * src/s390/ffi.c: New file from libffi CVS tree.
2713 * src/s390/sysv.S: New file from libffi CVS tree.
2714
2715 2002-04-28 Jakub Jelinek <jakub@redhat.com>
2716
2717 * configure.in (HAVE_AS_SPARC_UA_PCREL): Check for working
2718 %r_disp32().
2719 * src/sparc/v8.S: Use it.
2720 * src/sparc/v9.S: Likewise.
2721 * fficonfig.h.in: Rebuilt.
2722 * configure: Rebuilt.
2723
2724 2002-04-08 Hans Boehm <Hans_Boehm@hp.com>
2725
2726 * src/java_raw_api.c (ffi_java_raw_size): Handle FFI_TYPE_DOUBLE
2727 correctly.
2728 * src/ia64/unix.S: Add unwind information. Fix comments.
2729 Save sp in a way that's compatible with unwind info.
2730 (ffi_call_unix): Correctly restore sp in all cases.
2731 * src/ia64/ffi.c: Add, fix comments.
2732
2733 2002-04-08 Jakub Jelinek <jakub@redhat.com>
2734
2735 * src/sparc/v8.S: Make .eh_frame dependent on target word size.
2736
2737 2002-04-06 Jason Thorpe <thorpej@wasabisystems.com>
2738
2739 * configure.in (alpha*-*-netbsd*): Add target.
2740 * configure: Regenerate.
2741
2742 2002-04-04 Jeff Sturm <jsturm@one-point.com>
2743
2744 * src/sparc/v8.S: Add unwind info.
2745 * src/sparc/v9.S: Likewise.
2746
2747 2002-03-30 Krister Walfridsson <cato@df.lth.se>
2748
2749 * configure.in: Enable i*86-*-netbsdelf*.
2750 * configure: Rebuilt.
2751
2752 2002-03-29 David Billinghurst <David.Billinghurst@riotinto.com>
2753
2754 PR other/2620
2755 * src/mips/n32.s: Delete
2756 * src/mips/o32.s: Delete
2757
2758 2002-03-21 Loren J. Rittle <ljrittle@acm.org>
2759
2760 * configure.in: Enable alpha*-*-freebsd*.
2761 * configure: Rebuilt.
2762
2763 2002-03-17 Bryce McKinlay <bryce@waitaki.otago.ac.nz>
2764
2765 * Makefile.am: libfficonvenience -> libffi_convenience.
2766 * Makefile.in: Rebuilt.
2767
2768 * Makefile.am: Define ffitest_OBJECTS.
2769 * Makefile.in: Rebuilt.
2770
2771 2002-03-07 Andreas Tobler <toa@pop.agri.ch>
2772 David Edelsohn <edelsohn@gnu.org>
2773
2774 * Makefile.am (EXTRA_DIST): Add Darwin and AIX closure files.
2775 (TARGET_SRC_POWERPC_AIX): Add aix_closure.S.
2776 (TARGET_SRC_POWERPC_DARWIN): Add darwin_closure.S.
2777 * Makefile.in: Regenerate.
2778 * include/ffi.h.in: Add AIX and Darwin closure definitions.
2779 * src/powerpc/ffi_darwin.c (ffi_prep_closure): New function.
2780 (flush_icache, flush_range): New functions.
2781 (ffi_closure_helper_DARWIN): New function.
2782 * src/powerpc/aix_closure.S: New file.
2783 * src/powerpc/darwin_closure.S: New file.
2784
2785 2002-02-24 Jeff Sturm <jsturm@one-point.com>
2786
2787 * include/ffi.h.in: Add typedef for ffi_arg.
2788 * src/ffitest.c (main): Declare rint with ffi_arg.
2789
2790 2002-02-21 Andreas Tobler <toa@pop.agri.ch>
2791
2792 * src/powerpc/ffi_darwin.c (ffi_prep_args): Skip appropriate
2793 number of GPRs for floating-point arguments.
2794
2795 2002-01-31 Anthony Green <green@redhat.com>
2796
2797 * configure: Rebuilt.
2798 * configure.in: Replace CHECK_SIZEOF and endian tests with
2799 cross-compiler friendly macros.
2800 * aclocal.m4 (AC_COMPILE_CHECK_SIZEOF, AC_C_BIGENDIAN_CROSS): New
2801 macros.
2802
2803 2002-01-18 David Edelsohn <edelsohn@gnu.org>
2804
2805 * src/powerpc/darwin.S (_ffi_call_AIX): New.
2806 * src/powerpc/aix.S (ffi_call_DARWIN): New.
2807
2808 2002-01-17 David Edelsohn <edelsohn@gnu.org>
2809
2810 * Makefile.am (EXTRA_DIST): Add Darwin and AIX files.
2811 (TARGET_SRC_POWERPC_AIX): New.
2812 (POWERPC_AIX): New stanza.
2813 * Makefile.in: Regenerate.
2814 * configure.in: Add AIX case.
2815 * configure: Regenerate.
2816 * include/ffi.h.in (ffi_abi): Add FFI_AIX.
2817 * src/powerpc/ffi_darwin.c (ffi_status): Use "long" to scale frame
2818 size. Fix "long double" support.
2819 (ffi_call): Add FFI_AIX case.
2820 * src/powerpc/aix.S: New.
2821
2822 2001-10-09 John Hornkvist <john@toastedmarshmallow.com>
2823
2824 Implement Darwin PowerPC ABI.
2825 * configure.in: Handle powerpc-*-darwin*.
2826 * Makefile.am: Set source files for POWERPC_DARWIN.
2827 * configure: Rebuilt.
2828 * Makefile.in: Rebuilt.
2829 * include/ffi.h.in: Define FFI_DARWIN and FFI_DEFAULT_ABI for
2830 POWERPC_DARWIN.
2831 * src/powerpc/darwin.S: New file.
2832 * src/powerpc/ffi_darwin.c: New file.
2833
2834 2001-10-07 Joseph S. Myers <jsm28@cam.ac.uk>
2835
2836 * src/x86/ffi.c: Fix spelling error of "separate" as "seperate".
2837
2838 2001-07-16 Rainer Orth <ro@TechFak.Uni-Bielefeld.DE>
2839
2840 * src/x86/sysv.S: Avoid gas-only .balign directive.
2841 Use C style comments.
2842
2843 2001-07-16 Rainer Orth <ro@TechFak.Uni-Bielefeld.DE>
2844
2845 * src/alpha/ffi.c (ffi_prep_closure): Avoid gas-only mnemonic.
2846 Fixes PR bootstrap/3563.
2847
2848 2001-06-26 Rainer Orth <ro@TechFak.Uni-Bielefeld.DE>
2849
2850 * src/alpha/osf.S (ffi_closure_osf): Use .rdata for ECOFF.
2851
2852 2001-06-25 Rainer Orth <ro@TechFak.Uni-Bielefeld.DE>
2853
2854 * configure.in: Recognize sparc*-sun-* host.
2855 * configure: Regenerate.
2856
2857 2001-06-06 Andrew Haley <aph@redhat.com>
2858
2859 * src/alpha/osf.S (__FRAME_BEGIN__): Conditionalize for ELF.
2860
2861 2001-06-03 Andrew Haley <aph@redhat.com>
2862
2863 * src/alpha/osf.S: Add unwind info.
2864 * src/powerpc/sysv.S: Add unwind info.
2865 * src/powerpc/ppc_closure.S: Likewise.
2866
2867 2000-05-31 Jeff Sturm <jsturm@one-point.com>
2868
2869 * configure.in: Fix AC_ARG_ENABLE usage.
2870 * configure: Rebuilt.
2871
2872 2001-05-06 Bryce McKinlay <bryce@waitaki.otago.ac.nz>
2873
2874 * configure.in: Remove warning about beta code.
2875 * configure: Rebuilt.
2876
2877 2001-04-25 Hans Boehm <Hans_Boehm@hp.com>
2878
2879 * src/ia64/unix.S: Restore stack pointer when returning from
2880 ffi_closure_UNIX.
2881 * src/ia64/ffi.c: Fix typo in comment.
2882
2883 2001-04-18 Jim Wilson <wilson@redhat.com>
2884
2885 * src/ia64/unix.S: Delete unnecessary increment and decrement of loc2
2886 to eliminate RAW DV.
2887
2888 2001-04-12 Bryce McKinlay <bryce@albatross.co.nz>
2889
2890 * Makefile.am: Make a libtool convenience library.
2891 * Makefile.in: Rebuilt.
2892
2893 2001-03-29 Bryce McKinlay <bryce@albatross.co.nz>
2894
2895 * configure.in: Use different syntax for subdirectory creation.
2896 * configure: Rebuilt.
2897
2898 2001-03-27 Jon Beniston <jon@beniston.com>
2899
2900 * configure.in: Added X86_WIN32 target (Win32, CygWin, MingW).
2901 * configure: Rebuilt.
2902 * Makefile.am: Added X86_WIN32 target support.
2903 * Makefile.in: Rebuilt.
2904
2905 * include/ffi.h.in: Added X86_WIN32 target support.
2906
2907 * src/ffitest.c: Doesn't run structure tests for X86_WIN32 targets.
2908 * src/types.c: Added X86_WIN32 target support.
2909
2910 * src/x86/win32.S: New file. Based on sysv.S, but with EH
2911 stuff removed and made to work with CygWin's gas.
2912
2913 2001-03-26 Bryce McKinlay <bryce@albatross.co.nz>
2914
2915 * configure.in: Make target subdirectory in build dir.
2916 * Makefile.am: Override suffix based rules to specify correct output
2917 subdirectory.
2918 * Makefile.in: Rebuilt.
2919 * configure: Rebuilt.
2920
2921 2001-03-23 Kevin B Hendricks <khendricks@ivey.uwo.ca>
2922
2923 * src/powerpc/ppc_closure.S: New file.
2924 * src/powerpc/ffi.c (ffi_prep_args): Fixed ABI compatibility bug
2925 involving long long and register pairs.
2926 (ffi_prep_closure): New function.
2927 (flush_icache): Likewise.
2928 (ffi_closure_helper_SYSV): Likewise.
2929 * include/ffi.h.in (FFI_CLOSURES): Define on PPC.
2930 (FFI_TRAMPOLINE_SIZE): Likewise.
2931 (FFI_NATIVE_RAW_API): Likewise.
2932 * Makefile.in: Rebuilt.
2933 * Makefile.am (EXTRA_DIST): Added src/powerpc/ppc_closure.S.
2934 (TARGET_SRC_POWERPC): Likewise.
2935
2936 2001-03-19 Tom Tromey <tromey@redhat.com>
2937
2938 * Makefile.in: Rebuilt.
2939 * Makefile.am (ffitest_LDFLAGS): New macro.
2940
2941 2001-03-02 Nick Clifton <nickc@redhat.com>
2942
2943 * include/ffi.h.in: Remove RCS ident string.
2944 * include/ffi_mips.h: Remove RCS ident string.
2945 * src/debug.c: Remove RCS ident string.
2946 * src/ffitest.c: Remove RCS ident string.
2947 * src/prep_cif.c: Remove RCS ident string.
2948 * src/types.c: Remove RCS ident string.
2949 * src/alpha/ffi.c: Remove RCS ident string.
2950 * src/alpha/osf.S: Remove RCS ident string.
2951 * src/arm/ffi.c: Remove RCS ident string.
2952 * src/arm/sysv.S: Remove RCS ident string.
2953 * src/mips/ffi.c: Remove RCS ident string.
2954 * src/mips/n32.S: Remove RCS ident string.
2955 * src/mips/o32.S: Remove RCS ident string.
2956 * src/sparc/ffi.c: Remove RCS ident string.
2957 * src/sparc/v8.S: Remove RCS ident string.
2958 * src/sparc/v9.S: Remove RCS ident string.
2959 * src/x86/ffi.c: Remove RCS ident string.
2960 * src/x86/sysv.S: Remove RCS ident string.
2961
2962 2001-02-08 Joseph S. Myers <jsm28@cam.ac.uk>
2963
2964 * include/ffi.h.in: Change sourceware.cygnus.com references to
2965 gcc.gnu.org.
2966
2967 2000-12-09 Richard Henderson <rth@redhat.com>
2968
2969 * src/alpha/ffi.c (ffi_call): Simplify struct return test.
2970 (ffi_closure_osf_inner): Index rather than increment avalue
2971 and arg_types. Give ffi_closure_osf the raw return value type.
2972 * src/alpha/osf.S (ffi_closure_osf): Handle return value type
2973 promotion.
2974
2975 2000-12-07 Richard Henderson <rth@redhat.com>
2976
2977 * src/raw_api.c (ffi_translate_args): Fix typo.
2978 (ffi_prep_closure): Likewise.
2979
2980 * include/ffi.h.in [ALPHA]: Define FFI_CLOSURES and
2981 FFI_TRAMPOLINE_SIZE.
2982 * src/alpha/ffi.c (ffi_prep_cif_machdep): Adjust minimal
2983 cif->bytes for new ffi_call_osf implementation.
2984 (ffi_prep_args): Absorb into ...
2985 (ffi_call): ... here. Do all stack allocation here and
2986 avoid a callback function.
2987 (ffi_prep_closure, ffi_closure_osf_inner): New.
2988 * src/alpha/osf.S (ffi_call_osf): Reimplement with no callback.
2989 (ffi_closure_osf): New.
2990
2991 2000-09-10 Alexandre Oliva <aoliva@redhat.com>
2992
2993 * config.guess, config.sub, install-sh: Removed.
2994 * ltconfig, ltmain.sh, missing, mkinstalldirs: Likewise.
2995 * Makefile.in: Rebuilt.
2996
2997 * acinclude.m4: Include libtool macros from the top level.
2998 * aclocal.m4, configure: Rebuilt.
2999
3000 2000-08-22 Alexandre Oliva <aoliva@redhat.com>
3001
3002 * configure.in [i*86-*-freebsd*] (TARGET, TARGETDIR): Set.
3003 * configure: Rebuilt.
3004
3005 2000-05-11 Scott Bambrough <scottb@netwinder.org>
3006
3007 * libffi/src/arm/sysv.S (ffi_call_SYSV): Doubles are not saved to
3008 memory correctly. Use conditional instructions, not branches where
3009 possible.
3010
3011 2000-05-04 Tom Tromey <tromey@cygnus.com>
3012
3013 * configure: Rebuilt.
3014 * configure.in: Match `arm*-*-linux-*'.
3015 From Chris Dornan <cdornan@arm.com>.
3016
3017 2000-04-28 Jakub Jelinek <jakub@redhat.com>
3018
3019 * Makefile.am (SUBDIRS): Define.
3020 (AM_MAKEFLAGS): Likewise.
3021 (Multilib support.): Add section.
3022 * Makefile.in: Rebuilt.
3023 * ltconfig (extra_compiler_flags, extra_compiler_flags_value):
3024 New variables. Set for gcc using -print-multi-lib. Export them
3025 to libtool.
3026 (sparc64-*-linux-gnu*): Use libsuff 64 for search paths.
3027 * ltmain.sh (B|b|V): Don't throw away gcc's -B, -b and -V options
3028 for -shared links.
3029 (extra_compiler_flags_value, extra_compiler_flags): Check these
3030 for extra compiler options which need to be passed down in
3031 compiler_flags.
3032
3033 2000-04-16 Anthony Green <green@redhat.com>
3034
3035 * configure: Rebuilt.
3036 * configure.in: Change i*86-pc-linux* to i*86-*-linux*.
3037
3038 2000-04-14 Jakub Jelinek <jakub@redhat.com>
3039
3040 * include/ffi.h.in (SPARC64): Define for 64bit SPARC builds.
3041 Set SPARC FFI_DEFAULT_ABI based on SPARC64 define.
3042 * src/sparc/ffi.c (ffi_prep_args_v8): Renamed from ffi_prep_args.
3043 Replace all void * sizeofs with sizeof(int).
3044 Only compare type with FFI_TYPE_LONGDOUBLE if LONGDOUBLE is
3045 different than DOUBLE.
3046 Remove FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases (handled elsewhere).
3047 (ffi_prep_args_v9): New function.
3048 (ffi_prep_cif_machdep): Handle V9 ABI and long long on V8.
3049 (ffi_V9_return_struct): New function.
3050 (ffi_call): Handle FFI_V9 ABI from 64bit code and FFI_V8 ABI from
3051 32bit code (not yet cross-arch calls).
3052 * src/sparc/v8.S: Add struct return delay nop.
3053 Handle long long.
3054 * src/sparc/v9.S: New file.
3055 * src/prep_cif.c (ffi_prep_cif): Return structure pointer
3056 is used on sparc64 only for structures larger than 32 bytes.
3057 Pass by reference for structures is done for structure arguments
3058 larger than 16 bytes.
3059 * src/ffitest.c (main): Use 64bit rint on sparc64.
3060 Run long long tests on sparc.
3061 * src/types.c (FFI_TYPE_POINTER): Pointer is 64bit on alpha and
3062 sparc64.
3063 (FFI_TYPE_LONGDOUBLE): long double is 128 bit aligned to 128 bits
3064 on sparc64.
3065 * configure.in (sparc-*-linux*): New supported target.
3066 (sparc64-*-linux*): Likewise.
3067 * configure: Rebuilt.
3068 * Makefile.am: Add v9.S to SPARC files.
3069 * Makefile.in: Likewise.
3070 (LINK): Surround $(CCLD) into double quotes, so that multilib
3071 compiles work correctly.
3072
3073 2000-04-04 Alexandre Petit-Bianco <apbianco@cygnus.com>
3074
3075 * configure: Rebuilt.
3076 * configure.in: (i*86-*-solaris*): New libffi target. Patch
3077 proposed by Bryce McKinlay.
3078
3079 2000-03-20 Tom Tromey <tromey@cygnus.com>
3080
3081 * Makefile.in: Hand edit for java_raw_api.lo.
3082
3083 2000-03-08 Bryce McKinlay <bryce@albatross.co.nz>
3084
3085 * config.guess, config.sub: Update from the gcc tree.
3086 Fix for PR libgcj/168.
3087
3088 2000-03-03 Tom Tromey <tromey@cygnus.com>
3089
3090 * Makefile.in: Fixed ia64 by hand.
3091
3092 * configure: Rebuilt.
3093 * configure.in (--enable-multilib): New option.
3094 (libffi_basedir): New subst.
3095 (AC_OUTPUT): Added multilib code.
3096
3097 2000-03-02 Tom Tromey <tromey@cygnus.com>
3098
3099 * Makefile.in: Rebuilt.
3100 * Makefile.am (TARGET_SRC_IA64): Use `ia64', not `alpha', as
3101 directory name.
3102
3103 2000-02-25 Hans Boehm <boehm@acm.org>
3104
3105 * src/ia64/ffi.c, src/ia64/ia64_flags.h, src/ia64/unix.S: New
3106 files.
3107 * src/raw_api.c (ffi_translate_args): Fixed typo in argument
3108 list.
3109 (ffi_prep_raw_closure): Use ffi_translate_args, not
3110 ffi_closure_translate.
3111 * src/java_raw_api.c: New file.
3112 * src/ffitest.c (closure_test_fn): New function.
3113 (main): Define `rint' as long long on IA64. Added new test when
3114 FFI_CLOSURES is defined.
3115 * include/ffi.h.in (ALIGN): Use size_t, not unsigned.
3116 (ffi_abi): Recognize IA64.
3117 (ffi_raw): Added `flt' field.
3118 Added "Java raw API" code.
3119 * configure.in: Recognize ia64.
3120 * Makefile.am (TARGET_SRC_IA64): New macro.
3121 (libffi_la_common_SOURCES): Added java_raw_api.c.
3122 (libffi_la_SOURCES): Define in IA64 case.
3123
3124 2000-01-04 Tom Tromey <tromey@cygnus.com>
3125
3126 * Makefile.in: Rebuilt with newer automake.
3127
3128 1999-12-31 Tom Tromey <tromey@cygnus.com>
3129
3130 * Makefile.am (INCLUDES): Added -I$(top_srcdir)/src.
3131
3132 1999-09-01 Tom Tromey <tromey@cygnus.com>
3133
3134 * include/ffi.h.in: Removed PACKAGE and VERSION defines and
3135 undefs.
3136 * fficonfig.h.in: Rebuilt.
3137 * configure: Rebuilt.
3138 * configure.in: Pass 3rd argument to AM_INIT_AUTOMAKE.
3139 Use AM_PROG_LIBTOOL (automake 1.4 compatibility).
3140 * acconfig.h: Don't #undef PACKAGE or VERSION.
3141
3142 1999-08-09 Anthony Green <green@cygnus.com>
3143
3144 * include/ffi.h.in: Try to work around messy header problem
3145 with PACKAGE and VERSION.
3146
3147 * configure: Rebuilt.
3148 * configure.in: Change version to 2.00-beta.
3149
3150 * fficonfig.h.in: Rebuilt.
3151 * acconfig.h (FFI_NO_STRUCTS, FFI_NO_RAW_API): Define.
3152
3153 * src/x86/ffi.c (ffi_raw_call): Rename.
3154
3155 1999-08-02 Kresten Krab Thorup <krab@dominiq.is.s.u-tokyo.ac.jp>
3156
3157 * src/x86/ffi.c (ffi_closure_SYSV): New function.
3158 (ffi_prep_incoming_args_SYSV): Ditto.
3159 (ffi_prep_closure): Ditto.
3160 (ffi_closure_raw_SYSV): Ditto.
3161 (ffi_prep_raw_closure): More ditto.
3162 (ffi_call_raw): Final ditto.
3163
3164 * include/ffi.h.in: Add definitions for closure and raw API.
3165
3166 * src/x86/ffi.c (ffi_prep_cif_machdep): Added case for
3167 FFI_TYPE_UINT64.
3168
3169 * Makefile.am (libffi_la_common_SOURCES): Added raw_api.c
3170
3171 * src/raw_api.c: New file.
3172
3173 * include/ffi.h.in (ffi_raw): New type.
3174 (UINT_ARG, SINT_ARG): New defines.
3175 (ffi_closure, ffi_raw_closure): New types.
3176 (ffi_prep_closure, ffi_prep_raw_closure): New declarations.
3177
3178 * configure.in: Add check for endianness and sizeof void*.
3179
3180 * src/x86/sysv.S (ffi_call_SYSV): Call fixup routine via argument,
3181 instead of directly.
3182
3183 * configure: Rebuilt.
3184
3185 Thu Jul 8 14:28:42 1999 Anthony Green <green@cygnus.com>
3186
3187 * configure.in: Add x86 and powerpc BeOS configurations.
3188 From Makoto Kato <m_kato@ga2.so-net.ne.jp>.
3189
3190 1999-05-09 Anthony Green <green@cygnus.com>
3191
3192 * configure.in: Add warning about this being beta code.
3193 Remove src/Makefile.am from the picture.
3194 * configure: Rebuilt.
3195
3196 * Makefile.am: Move logic from src/Makefile.am. Add changes
3197 to support libffi as a target library.
3198 * Makefile.in: Rebuilt.
3199
3200 * aclocal.m4, config.guess, config.sub, ltconfig, ltmain.sh:
3201 Upgraded to new autoconf, automake, libtool.
3202
3203 * README: Tweaks.
3204
3205 * LICENSE: Update copyright date.
3206
3207 * src/Makefile.am, src/Makefile.in: Removed.
3208
3209 1998-11-29 Anthony Green <green@cygnus.com>
3210
3211 * include/ChangeLog: Removed.
3212 * src/ChangeLog: Removed.
3213 * src/mips/ChangeLog: Removed.
3214 * src/sparc/ChangeLog: Remboved.
3215 * src/x86/ChangeLog: Removed.
3216
3217 * ChangeLog.v1: Created.