1 2009-06-22 Zoltan Varga <vargaz@gmail.com>
3 * ppc/ppc-codegen.h: Rework the naming of the load/store macros,
4 ldr/str now handle register sized quantities, while ldptr/stptr handle
5 pointer sized quantities.
7 2009-06-20 Zoltan Varga <vargaz@gmail.com>
9 * ppc/ppc-codegen.h: Fix the last change to avoid self-assignments inside
12 * ppc/ppc-codegen.h: Add ppc_ldr/ppc_str macros to store regsize quantities.
13 Handle little endian host platforms in ppc_emit32.
15 2009-10-06 Jerry Maine <crashfourit@gmail.com>
17 Contributed under the terms of the MIT/X11 license by
18 Jerry Maine <crashfourit@gail.com>.
20 * amd64/amd64-codegen.h: Add marcos for coding two byte SIMD/SSE opcodes.
21 Added comments to help tell the different types of SSE code gen marcos appart.
23 2009-04-06 Zoltan Varga <vargaz@gmail.com>
25 * arm/arm-vfp-codegen.h: Add ARM_FSITOS/ARM_FSITOD.
27 2009-04-03 Zoltan Varga <vargaz@gmail.com>
29 * amd64/amd64-codegen.h: Add macros for decoding the SIB byte.
31 2009-04-02 Zoltan Varga <vargaz@gmail.com>
33 * arm/arm-vfp-codegen.h: Add missing VFP codegen macros.
35 2009-03-06 Zoltan Varga <vargaz@gmail.com>
37 * arm/tramp.c: Include a change from the debian patches. Avoid #include-ing
38 a file in the middle of a function.
40 2009-02-27 Zoltan Varga <vargaz@gmail.com>
42 * arm/{arm_fpamacros.h, arm_vfpmacros.h}: Remove these files, they are
45 2009-02-02 Mark Probst <mark.probst@gmail.com>
47 Contributed under the terms of the MIT/X11 license by Steven
48 Munroe <munroesj@us.ibm.com>.
50 * ppc/ppc-codegen.h: Make operand order and case consistent
51 (assembler order) for ppc_load_reg_update, ppc_load_multiple_regs,
52 ppc_store_multiple_regs, ppc_lwz, ppc_lhz, ppc_lbz,
53 ppc_stw,ppc_sth, ppc_stb, ppc_stwu, ppc_lbzu, ppc_lfdu, ppc_lfsu,
54 ppc_lfsux, ppc_lfsx, ppc_lha, ppc_lhau, ppc_lhzu, ppc_lmw,
55 ppc_lwzu, ppc_stbu, ppc_stfdu, ppc_stfsu, ppc_sthu, ppc_stmw. Use
56 "i" or "ui" instead of "d" for immediated operands to immediate
57 arthimetic and logical instructions in macros ppc_addi, ppc_addis,
58 ppc_ori, ppc_addic, ppc_addicd, ppc_andid, ppc_andisd.
59 [__mono_ppc64__]: Make operand order and case consistent
60 (assembler order) for ppc_load_multiple_regs,
61 ppc_store_multiple_regs. Simplify the DS form and make them
62 consistent with D forms for ppc_load_reg, ppc_load_reg_update,
63 ppc_store_reg, ppc_store_reg_update. ppc_ld, ppc_lwa, ppc_ldu,
64 ppc_std, ppc_stdu. Define ppc_lwax and ppc_lwaux.
66 2009-01-19 Rodrigo Kumpera <rkumpera@novell.com>
68 * x86/x86-codegen.h: Add x86_movd_xreg_membase.
70 2008-12-19 Mark Probst <mark.probst@gmail.com>
72 * ppc/ppc-codegen.h: Fixed the argument order for lwzu in
75 2008-12-12 Mark Mason <mmason@upwardaccess.com>
77 * mips/mips-codegen.h: Changes to support n32.
79 2008-12-10 Mark Probst <mark.probst@gmail.com>
81 * ppc/ppc-codegen.h: A few new macros for the final PPC/PPC64
84 2008-12-05 Mark Probst <mark.probst@gmail.com>
86 * ppc/ppc-codegen.h: ppc_load_func must use ppc_load_sequence.
87 Added ppc_compare_log.
89 2008-12-05 Mark Probst <mark.probst@gmail.com>
91 * ppc/ppc-codegen.h: Added ppc_load_func for PPC32. Added
92 ppc_load/store_multiple_regs and ppc_compare_reg_imm.
94 2008-12-02 Mark Probst <mark.probst@gmail.com>
96 * ppc/ppc-codegen.h: Opcodes for floating point conversions from
99 Code submitted by andreas.faerber@web.de at
100 https://bugzilla.novell.com/show_bug.cgi?id=324134 under the
103 2008-11-28 Mark Probst <mark.probst@gmail.com>
105 * ppc/ppc-codegen.h: #define for the maximum length of a load
108 2008-11-21 Mark Probst <mark.probst@gmail.com>
110 * ppc/ppc-codegen.h: Make ppc_is_[u]imm16() work with 64 bit
113 2008-11-20 Rodrigo Kumpera <rkumpera@novell.com>
115 * x86/x86-codegen.h: Add X86_SSE_MOVHPD_REG_MEMBASE and renamed MOVS to MOVSD.
117 2008-11-20 Rodrigo Kumpera <rkumpera@novell.com>
119 * x86/x86-codegen.h: Add X86_SSE_MOVS_ reg/membase variants.
121 * x86/x86-codegen.h: Add x86_sse_alu_pd_reg_membase_imm.
123 * x86/x86-codegen.h: Sort the x86_sse_alu_* macros decently.
125 2008-11-20 Mark Probst <mark.probst@gmail.com>
127 * ppc/ppc-codegen.h: Use ppc_load_reg instead of ppc_ld in
128 ppc_load_func to fix the 2 bit shift.
130 2008-11-20 Mark Probst <mark.probst@gmail.com>
132 * ppc/ppc-codegen.h: 64 bit division opcodes.
134 Code submitted by andreas.faerber@web.de at
135 https://bugzilla.novell.com/show_bug.cgi?id=324134 under the
138 2008-11-20 Zoltan Varga <vargaz@gmail.com>
140 * arm/Makefile.am (libmonoarch_arm_la_SOURCES): Don't build tramp.c, it is only
141 used by the interpreter.
143 2008-11-20 Rodrigo Kumpera <rkumpera@novell.com>
145 * x86/x86-codegen.h: Add PEXTR B/W/D.
147 2008-11-18 Rodrigo Kumpera <rkumpera@novell.com>
149 * x86/x86-codegen.h: Add PINSR B/W/D.
151 2008-11-18 Mark Probst <mark.probst@gmail.com>
153 * ppc/ppc-codegen.h: Macro for nop added.
155 2008-11-18 Mark Probst <mark.probst@gmail.com>
157 * ppc/ppc-codegen.h: PPC64 code generation macros.
159 Based on code submitted by andreas.faerber@web.de at
160 https://bugzilla.novell.com/show_bug.cgi?id=324134 under the
163 2008-11-18 Mark Probst <mark.probst@gmail.com>
165 * ppc/ppc-codegen.h: A few fixes and additions.
167 Based on code submitted by andreas.faerber@web.de at
168 https://bugzilla.novell.com/show_bug.cgi?id=324134 under the
171 2008-11-17 Rodrigo Kumpera <rkumpera@novell.com>
173 * x86/x86-codegen.h: Fix comment about the version of PCMPGTQ.
175 * x86/x86-codegen.h: Add movsd constant and x86_sse_alu_sd_membase_reg
178 2008-11-17 Rodrigo Kumpera <rkumpera@novell.com>
180 * x86/x86-codegen.h: Add X86_SSE_MOVHPD_MEMBASE_REG constant
181 and x86_sse_alu_pd_membase_reg/x86_sse_alu_membase_reg macros.
183 2008-11-15 Andreas Faerber <andreas.faerber@web.de>
185 * ppc/test.c: Add support for Mac OS X.
187 Code is contributed under MIT/X11 license.
189 2008-11-13 Rodrigo Kumpera <rkumpera@novell.com>
191 * x86/x86-codegen.h: Remove not used macro x86_pshufd_reg_reg.
193 2008-11-04 Rodrigo Kumpera <rkumpera@novell.com>
195 * x86/x86-codegen.h: Add store nta.
197 2008-11-03 Rodrigo Kumpera <rkumpera@novell.com>
199 * x86/x86-codegen.h: Add prefetch instruction
200 and x86_sse_alu_reg_membase macro.
202 2008-10-28 Rodrigo Kumpera <rkumpera@novell.com>
204 * x86/x86-codegen.h: Add long version of the packed integer
207 2008-10-27 Rodrigo Kumpera <rkumpera@novell.com>
209 * x86/x86-codegen.h: Add movddup.
211 2008-10-24 Rodrigo Kumpera <rkumpera@novell.com>
213 * x86/x86-codegen.h: Add signed pack with saturation.
215 2008-10-24 Rodrigo Kumpera <rkumpera@novell.com>
217 * x86/x86-codegen.h: Add signed packed mul high.
219 2008-10-23 Rodrigo Kumpera <rkumpera@novell.com>
221 * x86/x86-codegen.h: Add signed packed max, min, add/sub with saturation
224 2008-10-20 Rodrigo Kumpera <rkumpera@novell.com>
226 * x86/x86-codegen.h: Add multiply and store high.
228 2008-10-17 Rodrigo Kumpera <rkumpera@novell.com>
230 * x86/x86-codegen.h: Add packed int shuffle.
232 2008-10-16 Rodrigo Kumpera <rkumpera@novell.com>
234 * x86/x86-codegen.h: Add packed int compare equals and
237 2008-10-15 Rodrigo Kumpera <rkumpera@novell.com>
239 * x86/x86-codegen.h: Add packed int max/min/avg/shuffle and extract mask.
241 2008-10-14 Rodrigo Kumpera <rkumpera@novell.com>
243 * x86/x86-codegen.h: Add movsldup and movshdup.
245 2008-10-13 Rodrigo Kumpera <rkumpera@novell.com>
247 * x86/x86-codegen.h: Add remaining FP sse1 ops.
248 Add sse ps encoding with imm operand.
249 Add remaining sse1 ops.
251 2008-10-12 Rodrigo Kumpera <rkumpera@novell.com>
253 * x86/x86-codegen.h: Add macro for sse41 ops.
254 Add defined for pack ops, dword shifts/mul/pack.
256 2008-10-11 Zoltan Varga <vargaz@gmail.com>
258 * amd64/amd64-codegen.h (amd64_jump_code_size): Fix the 64 bit support.
260 2008-10-10 Zoltan Varga <vargaz@gmail.com>
262 * amd64/amd64-codegen.h (amd64_jump_code_size): Handle non 32-bit targets.
264 2008-10-09 Rodrigo Kumpera <rkumpera@novell.com>
266 * x86/x86-codegen.h: Add macros for sse shift, pack, unpack,
267 saturated math and packed byte/word/dword math.
269 2008-10-03 Rodrigo Kumpera <rkumpera@novell.com>
271 * x86/x86-codegen.h: Add macros for SSE instructions.
273 2008-09-27 Mark Probst <mark.probst@gmail.com>
275 * ppc/ppc-codegen.h: A few typecasts to fix compiler warnings.
277 2008-09-07 Mark Probst <mark.probst@gmail.com>
279 * ppc/ppc-codegen.h (ppc_load): Inserted cast to fix some
282 2008-06-16 Mark Probst <mark.probst@gmail.com>
284 * amd64/amd64-codegen.h: Removed extraneous parenthesis in a macro
287 2008-06-06 Zoltan Varga <vargaz@gmail.com>
289 * amd64/amd64-codegen.h (amd64_padding_size): Rewrite this to use the
290 instructions recommended by the amd64 manual.
292 2008-04-19 Zoltan Varga <vargaz@gmail.com>
294 * amd64/amd64-codegen.h (amd64_is_imm32): Use gint64 instead of glong because of
297 2008-03-13 Geoff Norton <gnorton@novell.com>
299 * arch/arm/tramp.c: Dont compile this on PLATFORM_MACOSX
301 2008-02-18 Zoltan Varga <vargaz@gmail.com>
303 * amd64/amd64-codegen.h (amd64_movsxd_reg_mem): New codegen macro.
305 2008-02-14 Zoltan Varga <vargaz@gmail.com>
307 * amd64/amd64-codegen.h (amd64_alu_membase8_imm_size): New codegen macro.
309 2008-02-08 Zoltan Varga <vargaz@gmail.com>
311 * arm/arm-codegen.h: Fix the ordering of arguments for some load/store opcodes
312 so they are consistent.
314 2008-01-24 Zoltan Varga <vargaz@gmail.com>
316 * Makefile.am (SUBDIRS): Or if INTERP_SUPPORTED is true.
318 * Makefile.am (SUBDIRS): Only set this on arm.
320 2007-11-20 Zoltan Varga <vargaz@gmail.com>
322 * amd64/amd64-codegen.h (amd64_alu_reg_imm_size): Prefer the smaller
323 instruction encoding.
325 2007-11-03 David S. Miller <davem@davemloft.net>
327 * sparc/sparc-codegen.h (sparc_set32, sparc_set): A plain sethi
328 can be used if the constant value only has the top 22 bits set.
330 2007-11-01 Geoff Norton <gnorton@novell.com>
332 * x86/Makefile.am: Only compile tramp.c if INTERP_SUPPORTED is true
333 Fixes the build on Leopard.
335 2007-11-01 Geoff Norton <gnorton@novell.com>
337 * ppc/Makefile.am: Only compile tramp.c if INTERP_SUPPORTED is true
338 Fixes the build on Leopard.
340 2007-10-26 Jonathan Chambers <joncham@gmail.com>
342 * amd64/amd64-codegen.h: Begin Win64 port. Use AMD64_ARG_REG#
343 defines to access param registers. Replace long usage with
344 gsize as sizeof(long) != sizeof(void*) on Win64.
346 Code is contributed under MIT/X11 license.
348 2007-10-09 Zoltan Varga <vargaz@gmail.com>
350 * amd64/amd64-codegen.h (amd64_jump_membase_size): Remove an unneccesary
351 rex prefix which trips up valgrind.
353 2007-07-14 Zoltan Varga <vargaz@gmail.com>
355 * amd64/amd64-codegen.h: Remove some unused rex prefixes.
357 Wed Jul 4 15:29:07 CEST 2007 Paolo Molaro <lupus@ximian.com>
359 * x86/x86-codegen.h: added minimal sse instructions currently
362 2007-06-13 Randolph Chung <tausq@debian.org>
364 * hppa/hppa-codegen.h: Update with more instructions.
365 * hppa/tramp.c: Disable for linux since we don't support the
368 2007-05-20 Zoltan Varga <vargaz@gmail.com>
370 * amd64/amd64-codegen.h (amd64_call_reg): Remove a got prefix which isn't needed.
372 2007-04-23 Zoltan Varga <vargaz@gmail.com>
374 * alpha/alpha-codegen.h: More alpha port work from
375 Sergey Tikhonov <tsv@solvo.ru>.
377 Mon Mar 12 17:07:32 CET 2007 Paolo Molaro <lupus@ximian.com>
379 * amd64/amd64-codegen.h: removed some useless size rex prefixes.
381 Wed Jan 24 21:00:40 CET 2007 Paolo Molaro <lupus@ximian.com>
383 * arm/arm-codegen.h: fixed encoding of short/byte load/store
384 instructions with negative immediate offsets.
386 Mon Nov 20 17:36:45 CET 2006 Paolo Molaro <lupus@ximian.com>
388 * arm/arm-codegen.h: added suppot for thumb interworking instructions.
390 Wed Nov 15 16:56:53 CET 2006 Paolo Molaro <lupus@ximian.com>
392 * mips/*: fixes by Mark E Mason <mark.e.mason@broadcom.com>.
394 2006-09-12 Zoltan Varga <vargaz@gmail.com>
396 * alpha/alpha-codegen.h: More alpha updates from Sergey Tikhonov <tsv@solvo.ru>.
398 Tue Sep 12 13:09:56 CEST 2006 Paolo Molaro <lupus@ximian.com>
400 * arm/*: VFP floating point format code generation support.
402 2006-09-12 Zoltan Varga <vargaz@gmail.com>
404 * ia64/ia64-codegen.h: Add xmpy_l/u pseudo ops.
406 2006-07-19 Zoltan Varga <vargaz@gmail.com>
408 * amd64/amd64-codegen.h: Fix amd64_mov_mem_reg.
410 * alpha/alpha-codegen.h alpha/test.c alpha/tramp.c: Applied patch from
411 Sergey Tikhonov <tsv@solvo.ru>. Updates to alpha support.
413 Thu Jun 15 16:59:36 CEST 2006 Paolo Molaro <lupus@ximian.com>
415 * ppc/ppc-codegen.h: reduce noisy build warnings by
416 casting to the more commonly used unsigned char type
417 (from johannes@sipsolutions.net (Johannes Berg)).
419 2006-05-14 Zoltan Varga <vargaz@gmail.com>
421 * ia64/ia64-codegen.h (ia64_fetchadd8_acq_hint_pred): Fix encoding of this
424 2006-04-21 Zoltan Varga <vargaz@gmail.com>
426 * Makefile.am (SUBDIRS): Revert the last change as arm depends on the old
429 2006-04-12 Zoltan Varga <vargaz@gmail.com>
431 * sparc/sparc-codegen.h (sparc_inst_i): New disassembly macro.
433 2006-04-04 Zoltan Varga <vargaz@gmail.com>
435 * Makefile.am (SUBDIRS): Avoid compiling subdirs needed by the
438 2005-12-22 Zoltan Varga <vargaz@gmail.com>
440 * sparc/sparc-codegen.h (sparc_membar): Add membar instruction.
442 2005-10-30 Zoltan Varga <vargaz@gmail.com>
444 * ia64/ia64-codegen.h (ia64_m17): Fix a warning.
446 2005-10-16 Zoltan Varga <vargaz@freemail.hu>
448 * amd64/amd64-codegen.h (AMD64_CALLEE_SAVED_REGS): Add %rbp.
450 2005-09-27 Raja R Harinath <rharinath@novell.com>
452 * arm/dpiops.sh, arm/fpaops.h: Output to stdout.
453 * arm/Makefile.am (arm_dpimacros.h, arm_fpamacros.h): Update. Fix
454 for srcdir != builddir.
456 2005-09-11 Zoltan Varga <vargaz@gmail.com>
458 * ia64/ia64-codegen.h (ia64_unw_pop_frames): New unwind macro.
460 2005-09-10 Zoltan Varga <vargaz@gmail.com>
462 * ia64/ia64-codegen.h: Remove 'manual' emitting of instructions.
463 Integrate emission of unwind directives into the assembly macros.
465 2005-09-04 Zoltan Varga <vargaz@gmail.com>
467 * ia64/ia64-codegen.h (ia64_no_stop): New macro.
469 2005-08-27 Zoltan Varga <vargaz@gmail.com>
471 * ia64/ia64-codegen.h: Fix some bugs.
473 * ia64/codegen.c: Update to work with latest ia64-codegen.h
475 2005-08-26 Zoltan Varga <vargaz@gmail.com>
477 * ia64/Makefile.am: Distribute ia64-codegen.h.
479 2005-08-21 Zoltan Varga <vargaz@freemail.hu>
481 * ia64/ia64-codegen.h: Improve ins scheduling and fix some bugs.
483 2005-08-17 Zoltan Varga <vargaz@freemail.hu>
485 * ia64/ia64-codegen.h: Add dependency information for all instructions.
487 2005-07-30 Zoltan Varga <vargaz@freemail.hu>
489 * ia64/ia64-codegen.h: Ongoing IA64 work.
491 Wed Jul 20 18:01:54 BST 2005 Paolo Molaro <lupus@ximian.com>
493 * arm/*: more codegen macros.
495 2005-07-18 Zoltan Varga <vargaz@freemail.hu>
497 * ia64/ia64-codegen.h (ia64_is_adds_imm): Ongoing IA64 work.
499 2005-06-23 Zoltan Varga <vargaz@freemail.hu>
501 * ia64/ia64-codegen.h: Add some new pseudo ops.
503 2005-06-19 Zoltan Varga <vargaz@freemail.hu>
505 * ia64/ia64-codegen.h: Fix encoding of ia64_fclass.
507 2005-06-12 Zoltan Varga <vargaz@freemail.hu>
509 * ia64/ia64-codegen.h: Ongoing IA64 work.
511 2005-06-09 Zoltan Varga <vargaz@freemail.hu>
513 * ia64/ia64-codegen.h: Ongoing IA64 work.
515 2005-05-30 Zoltan Varga <vargaz@freemail.hu>
517 * ia64/codegen.c: Fix it after latest changes.
519 2005-05-29 Zoltan Varga <vargaz@freemail.hu>
521 * ia64/ia64-codegen.h: Ongoing IA64 work.
523 * ia64/ia64-codegen.h: Ongoing IA64 work.
525 2005-05-28 Zoltan Varga <vargaz@freemail.hu>
527 * ia64/ia64-codegen.h: Ongoing IA64 work.
529 * ia64/ia64-codegen.h: Ongoing IA64 work.
531 * ia64/ia64-codegen.h: Ongoing IA64 work.
533 2005-05-26 Zoltan Varga <vargaz@freemail.hu>
535 * ia64/ia64-codegen.h: Ongoing IA64 work.
537 2005-05-22 Zoltan Varga <vargaz@freemail.hu>
539 * ia64/ia64-codegen.h: Ongoing IA64 work.
541 * ia64/ia64-codegen.h: Ongoing IA64 work.
543 * ia64/ia64-codegen.h: Ongoing IA64 work.
545 2005-05-21 Zoltan Varga <vargaz@freemail.hu>
547 * ia64/ia64-codegen.h: Ongoing IA64 work.
549 2005-05-19 Zoltan Varga <vargaz@freemail.hu>
551 * ia64/ia64-codegen.h ia64/codegen.c: Ongoing ia64 work.
553 2005-05-18 Zoltan Varga <vargaz@freemail.hu>
555 * ia64/ia64-codegen.h (ia64_codegen_init): Rename macro parameter.
557 2005-05-14 Zoltan Varga <vargaz@freemail.hu>
559 * Makefile.am: Only compile libmonoarch if the interpreter is compiled.
561 * ia64/ia64-codegen.h: Add IA64 code generation macros.
563 * Makefile.am: Add ia64 subdir.
565 2005-05-05 Zoltan Varga <vargaz@freemail.hu>
567 * alpha/tramp.c: Applied patch from Jakub Bogusz <qboosh@pld-linux.org>.
569 2005-03-28 Zoltan Varga <vargaz@freemail.hu>
571 * amd64/amd64-codegen.h: Avoid emitting a rex in some places.
573 2005-03-15 Zoltan Varga <vargaz@freemail.hu>
575 * amd64/amd64-codegen.h (amd64_emit_rex): Emit a rex when accessing the
578 * amd64/amd64-codegen.h (AMD64_BYTE_REGS): Add AMD64_BYTE_REGS macro.
580 2005-03-14 Zoltan Varga <vargaz@freemail.hu>
582 * amd64/amd64-codegen.h: Add missing AMD64_XMM7.
584 2005-03-13 Zoltan Varga <vargaz@freemail.hu>
586 * amd64/amd64-codegen.h: Remove some unneccesary REXes.
588 2005-03-08 Zoltan Varga <vargaz@freemail.hu>
590 * amd64/amd64-codegen.h (amd64_sse_cvtsi2sd_reg_reg_size): Add _size
591 variants to some sse2 macros.
593 * amd64/amd64-codegen.h (amd64_sse_cvtsd2si_reg_reg): Make this convert
596 2005-03-06 Zoltan Varga <vargaz@freemail.hu>
598 * amd64/amd64-codegen.h: Add some SSE2 instructions.
600 2005-02-20 Zoltan Varga <vargaz@freemail.hu>
602 * amd64/amd64-codegen.h: Add xadd instructions.
604 * amd64/amd64-codegen.h (amd64_jump_code_size): Do not emit a rex.
606 2004-11-25 Zoltan Varga <vargaz@freemail.hu>
608 * amd64/amd64-codegen.h: Updates to support the PIC changes.
610 Fri Nov 19 17:29:22 CET 2004 Paolo Molaro <lupus@ximian.com>
612 * ppc/ppc-codegen.h: counter reg decrement branch values
613 (patch by Geoff Norton <gnorton@customerdna.com>).
615 2004-11-16 Patrik Torstensson <patrik.torstensson@gmail.com>
617 * x86/x86-codegen.h: added opcodes for xadd instructions
619 2004-11-15 Zoltan Varga <vargaz@freemail.hu>
621 * amd64/x86-64-codegen.h: Get rid of this.
623 2004-08-30 Zoltan Varga <vargaz@freemail.hu>
625 * amd64/amd64-codegen.h (amd64_imul_reg_membase_size): Fix REX
628 2004-08-29 Zoltan Varga <vargaz@freemail.hu>
630 * amd64/amd64-codegen.h: More SSE work.
632 * amd64/amd64-codegen.h: Add SSE2 instructions.
634 2004-08-21 Zoltan Varga <vargaz@freemail.hu>
636 * amd64/amd64-codegen.h (X86_IS_BYTE_REG): Redefine X86_IS_BYTE_REG
637 since under amd64, all 16 registers have a low part.
639 2004-08-16 Zoltan Varga <vargaz@freemail.hu>
641 * x86/x86-codegen.h: Add macros for accessing the mod/rm byte.
643 2004-07-30 Zoltan Varga <vargaz@freemail.hu>
645 * amd64/amd64-codegen.h: Ongoing JIT work.
647 2004-07-29 Ben Maurer <bmaurer@ximian.com>
649 * x86/x86-codegen.h: Add opcodes for cmp BYTE PTR [eax], imm
651 2004-07-28 Zoltan Varga <vargaz@freemail.hu>
653 * amd64/amd64-codegen.h: Ongoing JIT work.
655 2004-07-24 Zoltan Varga <vargaz@freemail.hu>
657 * amd64/amd64-codegen.h: Ongoing JIT work.
659 2004-07-23 Zoltan Varga <vargaz@freemail.hu>
661 * amd64/amd64-codegen.h: Ongoing JIT work.
663 2004-07-23 zovarga <vargaz@freemail.hu>
665 * amd64/amd64-codegen.h: Ongoing JIT work.
667 2004-06-30 Zoltan Varga <vargaz@freemail.hu>
669 * sparc/sparc-codegen.h: Add SPARC64 support.
671 2004-05-07 Bernie Solomon <bernard@ugsolutions.com>
673 * ppc/ppc-codegen.h: remove GCC-ism in ppc_emit32
675 2004-04-29 Bernie Solomon <bernard@ugsolutions.com>
677 * ppc/tramp.c: use sizeof (stackval), fix
678 delegate tramp frame layout for Apple
680 2004-04-26 David Waite <mass@akuma.org>
682 * unknown.c: modify to have content for defined platforms (to
685 2004-03-29 Bernie Solomon <bernard@ugsolutions.com>
694 remove child from MonoInvocation as it isn't used.
696 * hppa/hppa-codegen.h:
697 fix displacements in FP instrs
699 2004-03-23 Bernie Solomon <bernard@ugsolutions.com>
701 * hppa/hppa-codegen.h: created
703 * hppa/tramp.c: changed style to be more like
706 * hppa/Makefile.am: add hppa-codegen.h
708 2004-03-16 Zoltan Varga <vargaz@freemail.hu>
710 * sparc/sparc-codegen.h: Add v9 branches with prediction.
712 2004-03-15 Bernie Solomon <bernard@ugsolutions.com>
714 * sparc/sparc-codegen.h: tweak sparc_mov_reg_reg
715 so Sun's dis command recognizes it.
717 2004-03-15 Zoltan Varga <vargaz@freemail.hu>
719 * sparc/sparc-codegen.h: Add some v9 instructions.
721 2004-03-11 Zoltan Varga <vargaz@freemail.hu>
723 * sparc/sparc-codegen.h: Ongoing sparc work.
725 2004-03-07 Duncan Mak <duncan@ximian.com>
727 * Makefile.am: Removed the reference to 'x86-64'. This was the cause
728 of the missing Mono daily tarballs, 'make dist' wasn't working.
730 We do have an 'amd64' directory, but it doesn't make it in 'make
733 2004-02-19 Zoltan Varga <vargaz@freemail.hu>
735 * sparc/sparc-codegen.h: Fix lots of opcodes + add new ones.
737 * sparc/tramp.c: Fix alignment of structures containing doubles.
739 2004-02-02 Zoltan Varga <vargaz@freemail.hu>
741 * sparc/tramp.c: Implement all floating point argument passing conventions in
742 Sparc V8. Also fix structure passing in V8.
744 Thu Nov 13 16:24:29 CET 2003 Paolo Molaro <lupus@ximian.com>
746 * ppc/ppc-codegen.h: fixed most of the incorrect macros from ct.
748 2003-10-31 Zoltan Varga <vargaz@freemail.hu>
750 * */tramp.c (mono_create_method_pointer): Rename to
751 mono_arch_create_method_pointer, move common code to a new function in
754 * */tramp.c (mono_create_trampoline): Rename to
755 mono_arch_create_trampoline for consistency.
757 2003-10-13 Bernie Solomon <bernard@ugsolutions.com>
759 * x86/tramp.c: restore EDX after memcpy call
761 2003-10-13 Bernie Solomon <bernard@ugsolutions.com>
763 * Makefile.am: add hppa subdir
765 * hppa/tramp.c: add initial implementation - this is 64 bit only
766 hppa/Makefile.am hppa/.cvsignore: added
768 2003-10-13 Bernie Solomon <bernard@ugsolutions.com>
770 * sparc/sparc-codegen.h sparc/tramp.c: add initial implementation
771 for V9 (64 bit), cover more 32 bit cases as well.
773 2003-09-03 Zoltan Varga <vargaz@freemail.hu>
775 * x86/tramp.c: Fixes from Bernie Solomon (bernard@ugsolutions.com).
777 2003-08-21 Zoltan Varga <vargaz@freemail.hu>
779 * x86/tramp.c: Fixes from Bernie Solomon (bernard@ugsolutions.com).
781 Tue Jul 1 13:03:43 CEST 2003 Paolo Molaro <lupus@ximian.com>
783 * alpha/tramp.c: update from Laramie Leavitt (lar@leavitt.us).
785 2003-05-20 Dietmar Maurer <dietmar@ximian.com>
787 * x86/x86-codegen.h (x86_set_reg): add an assertion - it does
788 not work for all registers.
790 Sat Feb 1 10:59:31 CET 2003 Paolo Molaro <lupus@ximian.com>
792 * alpha/*: update from Laramie.
794 Mon Jan 27 12:49:10 CET 2003 Paolo Molaro <lupus@ximian.com>
796 * alpha/*: start of the port to the alpha architecture by
797 Laramie Leavitt (<lar@leavitt.us>).
799 Tue Jan 21 17:29:53 CET 2003 Paolo Molaro <lupus@ximian.com>
801 * ppc/ppc-codegen.h: completed ppc native code generation by
802 Taylor Christopher P <ct@cs.clemson.edu>.
804 Fri Jan 17 21:14:18 CET 2003 Paolo Molaro <lupus@ximian.com>
806 * ppc/tramp.c: adapted to work for MacOSX (from a patch by
809 2002-11-11 Mark Crichton <crichton@gimp.org>
811 * sparc/tramp.c: Added some disassembly bits for debugging.
813 2002-10-02 Mark Crichton <crichton@gimp.org>
815 * sparc/tramp.c: More cleanup of the trampoline code. Still some
816 problems with it w.r.t. delegates.
818 2002-09-25 Mark Crichton <crichton@gimp.org>
820 * sparc/tramp.c: Off by one error. Whoops. Trampoline code should
823 2002-09-24 Mark Crichton <crichton@gimp.org>
825 * sparc/tramp.c: Works as well as the old code did. Cleanup is
826 finished. The framework now for adding more type support is in,
827 and should be a *lot* cleaner.
829 2002-09-22 Mark Crichton <crichton@gimp.org>
831 * sparc/tramp.c: Completely broke trampolining on SPARC processors.
832 The code needed a nasty cleanup, so most of it is rewritten.
835 2002-08-20 Mark Crichton <crichton@gimp.org>
837 * sparc/tramp.c (mono_create_trampoline): Now works on Sparc. Tested
838 on an Ultra 2 running Linux.
840 Mon Aug 5 19:21:19 CEST 2002 Paolo Molaro <lupus@ximian.com>
842 * x86/tramp.c: fix random memory read in mono_create_method_pointer.
844 2002-08-05 Dietmar Maurer <dietmar@ximian.com>
846 * x86/tramp.c (mono_create_trampoline): fixed stack_size bug
848 2002-08-01 Dietmar Maurer <dietmar@ximian.com>
850 * x86/tramp.c (mono_create_method_pointer): allocate space for
852 (mono_create_trampoline): also push the value type pointer for
853 methods returning value types.
854 (mono_create_method_pointer): support valuetype returns.
856 2002-07-31 Dietmar Maurer <dietmar@ximian.com>
858 * x86/tramp.c: remove code to handle PInvoke because this is no
860 (mono_create_method_pointer): return method->addr for pinvoke methods
862 Fri Jul 19 14:18:36 CEST 2002 Paolo Molaro <lupus@ximian.com>
864 * x86/tramp.c: fix float loads. Simple delegate marshaling fix.
866 2002-07-08 Radek Doulik <rodo@ximian.com>
868 * ppc/tramp.c: marshaling for SZARRAY
870 2002-07-05 Radek Doulik <rodo@ximian.com>
872 * ppc/tramp.c: removed magic hack
874 Tue Jun 18 10:21:56 CEST 2002 Paolo Molaro <lupus@ximian.com>
876 * x86/tramp.c: marshal simple arrays correctly.
878 2002-05-27 Miguel de Icaza <miguel@ximian.com>
880 * x86/x86-codegen.h: Set the standard header format.
882 2002-05-23 Dietmar Maurer <dietmar@ximian.com>
884 * x86/tramp.c (mono_create_method_pointer): removed the magic
885 trick to store the function pointer in the prolog and use the same
886 mechanism as in the jit.
888 2002-05-13 Radek Doulik <rodo@ximian.com>
890 * ppc/tramp.c (emit_save_parameters): fix I8 parameters
892 2002-05-06 Sergey Chaban <serge@wildwestsoftware.com>
894 * x86/x86-codegen.h: added missing shifts;
895 8-bit ALU operations (reg-reg);
896 macro for FPU ops with integer operand;
897 FIST macro (without pop);
899 Mon Apr 22 12:57:31 CEST 2002 Paolo Molaro <lupus@ximian.com>
901 * x86/x86-codegen.h: added loop instructions and made x86_patch fully
904 2002-04-20 Dietmar Maurer <dietmar@ximian.com>
906 * x86/tramp.c (mono_create_trampoline): support internalcall
909 Sat Apr 6 16:29:40 CEST 2002 Paolo Molaro <lupus@ximian.com>
911 * x86/tramp.c: fix advancement of argument position on the stack.
913 Sat Mar 16 19:12:57 CET 2002 Paolo Molaro <lupus@ximian.com>
915 * x86/tramp.c: increase default allocated size for trampolines
916 and assert on overflow.
918 2002-03-14 Dietmar Maurer <dietmar@ximian.com>
920 * x86/tramp.c (mono_create_trampoline): dont use fld/fst to copy
923 Mon Mar 11 16:14:29 CET 2002 Paolo Molaro <lupus@ximian.com>
925 * x86/x86-codegen.h: addex x86_clear_reg() and changed
926 x86_mov_reg_imm() to not check for imm == 0.
928 Thu Feb 28 12:34:21 CET 2002 Paolo Molaro <lupus@ximian.com>
930 * x86/tramp.c: start handling of more complex marshaling stuff.
932 2002-02-25 Sergey Chaban <serge@wildwestsoftware.com>
934 * arm: added ARM support code.
935 * Makefile.am: added arm to DIST_SUBDIRS.
937 2002-02-24 Radek Doulik <rodo@ximian.com>
939 * ppc/tramp.c (mono_create_method_pointer): basic delegates
940 implementation, it works for simple delegates now and I am already
941 pretty close to have it working for every delegates, but I am
942 going to sleep and finish it tomorrow?
943 (mono_create_method_pointer): implements delegates with parameters
946 2002-02-22 Jeffrey Stedfast <fejj@ximian.com>
948 * sparc/tramp.c (mono_create_trampoline): Much tinkering to get
949 the opcodes more correct. Still needs a lot of work.
951 2002-02-19 Radek Doulik <rodo@ximian.com>
953 * ppc/tramp.c (emit_save_parameters): don't start saving 64bit values to
955 added stack saving for most arguments
957 Tue Feb 19 20:19:38 CET 2002 Paolo Molaro <lupus@ximian.com>
959 * x86/tramp.c: avoid pointer arthmetric (pointed out by Serge).
961 2002-02-17 Radek Doulik <rodo@ximian.com>
963 * ppc/tramp.c: fixed minimal stack size, fixed string parameters,
964 fix byte and half word parameters
965 (mono_string_new_wrapper): new helper function, cut&pasted from
966 x86, modified to check for NULL text to avoid branching in
968 (calculate_sizes): updated for string retval changes
969 (emit_call_and_store_retval): updated for string retval
971 * ppc/ppc-codegen.h (ppc_mr): added lhz, lbz, sth
973 2002-02-16 Radek Doulik <rodo@ximian.com>
975 * ppc/tramp.c (emit_call_and_store_retval): support U2, I2, CHAR
977 Mon Feb 11 18:40:04 CET 2002 Paolo Molaro <lupus@ximian.com>
979 * sparc/*: sparc codegen header and some untested trampoline code.
981 Mon Feb 11 12:32:35 CET 2002 Paolo Molaro <lupus@ximian.com>
983 * x86/tramp.c: fix handling of multiple marshaleed strings.
984 * x86/x86-codegen.h: some code to patch branch displacements.
986 Fri Feb 1 16:03:00 CET 2002 Paolo Molaro <lupus@ximian.com>
988 * x86/tramp.c, ppc/tramp.c: implement mono_method_pointer_get ().
990 2002-01-23 Miguel de Icaza <miguel@ximian.com>
992 * x86/tramp.c (mono_create_trampoline): Do not try to create a
993 mono_string_new if the return value from the PInvoke code is
996 Mon Jan 14 11:50:16 CET 2002 Paolo Molaro <lupus@ximian.com>
998 * x86/x86-codegen.c: added overflow condition code and some aliases
1001 Thu Jan 10 21:01:08 CET 2002 Paolo Molaro <lupus@ximian.com>
1003 * x86/tramp.c: mono_create_trampoline (): the runtime argument is
1004 needed to handle correctly delegates, the previous change in handling
1005 the string return type broke them.
1007 Tue Jan 8 22:38:41 MST 2002 Matt Kimball <matt@kimball.net>
1009 * x86/tramp.c: handle strings returned from functions in external
1010 libraries by converting to a Mono string object after the pinvoke'd
1013 Sat Jan 5 15:51:06 CET 2002 Paolo Molaro <lupus@ximian.com>
1015 * x86/tramp.c: handle short integer return types.
1017 Thu Dec 20 20:13:07 CET 2001 Paolo Molaro <lupus@ximian.com>
1019 * x86/tramp.c: fix create_method_pointer() to pass the arguments
1020 correctly and add check for overflow.
1022 Thu Dec 13 15:56:53 CET 2001 Paolo Molaro <lupus@ximian.com>
1024 * x86/x86-codegen.h: x86_mov_memindex_imm() added.
1026 2001-11-29 Radek Doulik <rodo@ximian.com>
1028 * ppc/tramp.c: use r12 which is volatile instead of non-volatile
1031 * Makefile.am (libmonoarch_la_LIBADD): added ppc to DIST_SUBDIRS
1032 generate libmonoarch for ppc
1034 Tue Nov 27 15:24:07 CET 2001 Paolo Molaro <lupus@ximian.com>
1036 * x96/x86-codegen.c: x86_lea_memindex() added.
1038 Thu Nov 15 17:41:01 CET 2001 Paolo Molaro <lupus@ximian.com>
1040 * x86/tramp.c: handle enums with underlying type different from int32.
1042 Wed Nov 14 19:21:26 CET 2001 Paolo Molaro <lupus@ximian.com>
1044 * x86/tramp.c: handle boolean as a return value.
1045 * x96/x86-codegen.c: x86_widen_memindex() added.
1047 2001-11-07 Miguel de Icaza <miguel@ximian.com>
1049 * x86/tramp.c: Include stdlib to kill warning.
1051 Mon Oct 22 15:20:14 CEST 2001 Paolo Molaro <lupus@ximian.com>
1053 * x86/tramp.c: handle boolean, u1 and i1 as return values.
1055 Wed Oct 10 16:07:24 CEST 2001 Paolo Molaro <lupus@ximian.com>
1057 * x86/x86-codegen.c: added x86_set_{reg,mem,membase}.
1059 Mon Oct 8 16:13:07 CEST 2001 Paolo Molaro <lupus@ximian.com>
1061 * x86/tramp.c: remove mono_get_ansi_string () and use
1062 mono_string_to_utf8 () instead.
1064 Thu Oct 4 19:09:13 CEST 2001 Paolo Molaro <lupus@ximian.com>
1066 * x86/tramp.c: allow marshalling valuetypes if they are
1069 Mon Oct 1 18:48:27 CEST 2001 Paolo Molaro <lupus@ximian.com>
1071 * x86/tramp.c: fix thinko (s/SUB/ADD/) in stack adjustment
1072 and avoid a couple of unnecessary instructions.
1074 Fri Sep 28 19:43:12 CEST 2001 Paolo Molaro <lupus@ximian.com>
1076 * x86/tramp.c: marshal valuetypes that are enums.
1078 Thu Sep 27 15:34:37 CEST 2001 Paolo Molaro <lupus@ximian.com>
1080 * x86/x86-codegen.h: in memindex operand you can use X86_NOBASEREG
1083 Wed Sep 26 16:29:36 CEST 2001 Paolo Molaro <lupus@ximian.com>
1085 * x86/x86-codegen.h: added memindex addressing mode encoding
1086 (and mov to/from register opcodes).
1088 Mon Sep 24 18:49:01 CEST 2001 Paolo Molaro <lupus@ximian.com>
1090 * x86/tramp.c: don't change a MONO_TYPE_STRING to a char*
1091 when it's an argument to an internalcall.
1093 Sun Sep 23 13:44:57 CEST 2001 Paolo Molaro <lupus@ximian.com>
1095 * x86/tramp.c: handle MONO_TYPE_CLASS in trampolines.
1097 2001-09-21 Dietmar Maurer <dietmar@ximian.com>
1099 * x86/x86-codegen.h (x86_breakpoint): added.
1101 Tue Sep 18 13:23:59 CEST 2001 Paolo Molaro <lupus@ximian.com>
1103 * x86/x86-codegen.h: remove C++ comments.
1105 2001-09-17 Dietmar Maurer <dietmar@ximian.com>
1107 * x86/x86-codegen.h (x86_alu_reg_reg): replaced src/dest
1109 Mon Sep 10 17:26:06 CEST 2001 Paolo Molaro <lupus@ximian.com>
1111 * Makefile.am, x86/Makefile.am: conditional compile logic
1112 to make porting to different targets easier.
1114 Fri Sep 7 18:43:06 CEST 2001 Paolo Molaro <lupus@ximian.com>
1116 * x86/x86-codegen.h: fixes and x86_mov_membase_imm ().
1117 * x86/tramp.c: implemented mono_create_method_pointer ():
1118 creates a native pointer to a method implementation that can be
1119 used as a normal C callback.
1121 Thu Sep 6 15:38:00 CEST 2001 Paolo Molaro <lupus@ximian.com>
1123 * x86/x86-codegen.h: added x86_rdtsc() and fixes.
1124 * x86/tramp.c: create trampolines to call pinvoke methods.
1125 * x86/Makefile.am: create a libmonoarch convenience library.
1127 Mon Aug 27 09:29:00 CEST 2001 Paolo Molaro <lupus@ximian.com>
1129 * x86/x86-codegen.h: fix x86_call_code (). x86_mov_regp_reg () added.
1131 Sat Aug 18 12:40:32 CEST 2001 Paolo Molaro <lupus@ximian.com>
1133 * x86/x86-codegen.h: fix a couple of buglets and add x86_regp_emit().
1135 Wed Aug 8 15:30:05 CEST 2001 Paolo Molaro <lupus@ximian.com>
1137 * x86/x86-codegen.h, x86/test.c: added x86 code emitter with