Commit Graph

1006 Commits

Author SHA1 Message Date
sergei
e11c880b1e x86 assembler improvements:
* Don't generate rex.w for "CALL|JMP|LCALL|LJMP regmem", they are 64-bit by default.
* LCALL,LJMP flagged as calljump instructions.
* LCALL,LJMP encode only far jumps and don't accept register operands.
* GAS writer: fixed writing rip-relative operands of calljump instructions.
+ test.

git-svn-id: trunk@19413 -
2011-10-08 11:34:04 +00:00
florian
111d05c68f o patch by Alexander Shishkin, resolves #20409
* eliminate warnings in compiler (i386 & i368->x86_64) and minor refactorings
    - comment out unused vars and types
    - comment out unneeded comparisons (Longword <=> 0)
    - suppress some "comparison always true|false" warnings
    - tweak visiblity sections

git-svn-id: trunk@19385 -
2011-10-05 20:11:09 +00:00
sergei
c6d382fe18 * Moved procedure genlinearlist from ti386casenode to tx86casenode. This optimization works well on x86_64, resulting in approximately twice less instructions than when using generic code.
git-svn-id: trunk@19363 -
2011-10-04 11:48:33 +00:00
florian
d7f7a9bb76 * patch by Alexander Shishkin to clean up $ifopt usage by $push/$pop, resolves #20346
git-svn-id: trunk@19256 -
2011-09-27 20:22:40 +00:00
sergei
d94460e3b3 * Made the x86-specific scalefactor optimization of tvecnode available on x86_64 as well, by moving update_reference_reg_mul method from ti386vecnode to newly introduced tx86vecnode.
git-svn-id: trunk@19245 -
2011-09-26 12:58:59 +00:00
florian
4868b09844 * patch by alexvin to replace $ifopt by $push/$pop, resolves #20332
git-svn-id: trunk@19239 -
2011-09-25 21:19:50 +00:00
florian
e8dafe4dde * fixed change information for a lot of sse instructions
git-svn-id: trunk@18787 -
2011-08-20 12:28:49 +00:00
florian
c011949765 + iretq for x86_64
git-svn-id: trunk@18273 -
2011-08-19 12:56:26 +00:00
florian
8308b46a94 + support for assembler instructions with four operands
+ support for insertq, resolves #19910

git-svn-id: trunk@18206 -
2011-08-14 16:46:35 +00:00
florian
ff5f311b34 - removed no more used constants
git-svn-id: trunk@18199 -
2011-08-13 20:54:01 +00:00
sergei
0231863fce + Added missing PMULLD instruction, part of Mantis #19910
git-svn-id: trunk@18106 -
2011-08-06 06:59:33 +00:00
pierre
cff85bdb3b * Rename ag386nsm.pas unit source to agx86nsm.pas
git-svn-id: trunk@17933 -
2011-07-04 09:13:55 +00:00
pierre
f6ab12d677 * Move ag386nsm.pas to x86 directory
git-svn-id: trunk@17932 -
2011-07-04 09:08:54 +00:00
florian
a6b2fcfe22 * fixes breakage of r17766, should resolve #19598
git-svn-id: trunk@17814 -
2011-06-23 21:30:40 +00:00
florian
b00a610aec * optimize andb/w const,reg andw/l const,reg sequences as well
git-svn-id: trunk@17766 -
2011-06-18 15:22:09 +00:00
florian
ce76ce8769 * fixed wrong tmpref usage as mentioned by Pierre on the mailing list
git-svn-id: trunk@17754 -
2011-06-14 18:48:22 +00:00
sergei
354d0520b7 + x86 assembler: fixed MOVABS instruction (it is a x86_64-only subset of MOV with 8-byte immediates/offsets) and same-form encodings of MOV instruction.
git-svn-id: trunk@17666 -
2011-06-05 17:32:18 +00:00
florian
69e7d8b02f * generate correct PIC in tcg386.a_loadaddr_ref_cgpara, resolves #19464
git-svn-id: trunk@17657 -
2011-06-04 16:57:45 +00:00
sergei
5bda700410 * External assemblers: moved some common code into the base class, 4 copies reduced to one.
git-svn-id: trunk@17561 -
2011-05-26 13:10:49 +00:00
sergei
d89aeeadaf x86 assembler:
* Adjust rip-relative offsets using values known from calcsize, instead of reinterpreting the tail of instruction. This is simpler and remains the same regardless of particular codes used in opcode table.
* Reduced maxinfolen to 8.
- 'jmp imm64' and 64-bit versions of 'in' do not exist, removed.

git-svn-id: trunk@17546 -
2011-05-23 20:43:50 +00:00
sergei
504e0c6816 x86 assembler fixes:
* mkx86ins.pp: 'regmem' operand type in x86ins.dat must be converted to OT_RM_GPR, not OT_REGMEM. The latter does not restrict register type, allowing to use e.g. xmm registers in place of regular ones. 
* Finished 'movd' and 'movq', added some tests for them to tasm2.pp.

git-svn-id: trunk@17515 -
2011-05-20 20:39:35 +00:00
sergei
1aa5c4ac0e * Fixed wrong ModRM byte in movq encoding, causing regression of test/cg/tcalext6 in r17490.
git-svn-id: trunk@17493 -
2011-05-18 14:55:26 +00:00
sergei
9e8a31193b x86 assembler:
* Optimized the opcode representation of movq and remaining 3DNow instructions
* Disallow immediates not fitting in 32 bits (Mantis #14685) + test
* Disallow push/pop with 32-bit operands in x86_64 + test

git-svn-id: trunk@17490 -
2011-05-17 20:06:12 +00:00
sergei
c28197bac3 * PSADBW second operand is xmmrm, not xmmreg
+ A couple of tests related to the recent assembler changes

git-svn-id: trunk@17460 -
2011-05-14 18:59:06 +00:00
florian
77f2d6cc0d * introduce usage of TCGInt in the code generator units
git-svn-id: trunk@17459 -
2011-05-14 17:58:23 +00:00
sergei
958954890e A big update of x86 instruction table, part 2:
- Removed remaining address-size control codes \300 and \301
- Also removed codes \323, they are no longer necessary once REX is being written on the first literal.

git-svn-id: trunk@17457 -
2011-05-14 16:36:42 +00:00
sergei
3b979fef6d * Re-commit r17437 after more testing and fixing aasmcpu.pp in r17449.
git-svn-id: trunk@17452 -
2011-05-14 11:04:52 +00:00
sergei
b257231203 * Revert r17437, it breaks builds with -O2 and builds on i386 (although -O- on x86_64 is ok).
git-svn-id: trunk@17439 -
2011-05-12 23:53:18 +00:00
sergei
1d81a1244b A big update of x86 instruction table, part 1 (mostly SIMD instructions):
* Using ot_mmxrm and ot_xmmrm operand types to match arguments, reduces number of required entries by half.
* Replaced all literal $66, $F2 and $F3 prefixes with control codes (\361, \334 and \333, respectively).
* Prefix control codes imply writing REX, so code \323 after them is no longer necessary, removed.
* Fixed technology flags (SSSE3, SSE4.1, SSE4.2)
- Removed codes \300 and \301 (intended to generate address size prefix). FPC does not support this feature (the prefix itself is generated, but process_ea rejects operands needing non-default address size). Probably we don't even need to support it. But if we do, a much simpler solution is check all operands, like today's NASM does.
* Fixed/added some instructions along the way, namely CRC32, UNPCKHPD, CMPNEQSD.

git-svn-id: trunk@17437 -
2011-05-12 19:49:19 +00:00
sergei
de8ae0f873 x86 assembler improvements:
* Simplified REX handling: instead of useless adjusting instruction length on every REX change, adjust it just once, based on the final REX value.
* Similar for omitting REX.W in certain instructions: set a flag and consider it once the entire instruction has been processed. This removes the requirement for \335 to be the last opcode.
* maybewriterex also after $F2 prefix
* Fixed behavior of codes \310 and \311 (16- and 32-bit address respectively), this is needed for a few fancy instructions like LOOP and JCXZ, which honor address-size prefix instead of REX.
+ control code \361 for $66 prefix
+ IF_SSSE3, IF_SSE41, IF_SSE42 instruction flags (with dummy values)

x86/x86ins.dat:
* replaced literal $67 prefixes with \310 or \311
* marked encodings containing code \310 as NOX86_64
* replaced literal $48 prefixes with \326

git-svn-id: trunk@17433 -
2011-05-12 13:09:51 +00:00
sergei
6739cec2b9 * Flagged with NOX86_64 instructions/encodings that are invalid in 64-bit mode.
* AESKEYGENASSIST is not ATT-specific name, it is used by Intel-style assemblers as well. Also updated tests/test/taes1.pp to reflect the change.
+ Added SCASQ, resolves #16730 (other opcodes mentioned in that report were added/fixed earlier)

git-svn-id: trunk@17431 -
2011-05-11 15:50:59 +00:00
sergei
9bb0dc6cfe * mkx86ins.pp: ot_signed flag must not be set by literal opcodes. Ignore 2- and 3-byte literal sequences, not just 1-byte ones.
* x86ins.dat: replaced codes \17 with literal zeros. They aren't necessary for FPC, and they were removed from NASM quite a while ago.

git-svn-id: trunk@17430 -
2011-05-11 12:31:18 +00:00
Jonas Maebe
a08989a76b * zero/sign extend parameter values and return values < 32 bit to 32 bit on
x86_64 (mantis #19269)
  * this is also required for darwin/i386 and was already done there for
    parameters, but not yet for return values

git-svn-id: trunk@17388 -
2011-05-01 11:33:29 +00:00
pierre
af32b57170 * increase op2strtable size to string[15] and rectify askeygen to askeygenassist for ATT
git-svn-id: trunk@17358 -
2011-04-21 13:38:20 +00:00
sergei
f97f223de6 x86_64 assembler reader improvements:
+ Added new value TAttSuffix.attsufINTdual, assigned it to movsX and movzX instructions
* Moved suffix-to-size translation tables from rax86att.pas to itcpugas.pas
+ Added x86_64 specific suffix-to-size translation, enabling BQ and WQ suffixes (LQ seems unnecessary at the moment)
* Fixed logic of tx86attreader.is_asmopcode so it only assigns dual suffix to instructions that explicitly allow it. This disambiguates cases like movsbq=movs+bq vs. cmovbq=cmovb+q
* As a net result: movz[bw]q and movs[bw]q now compile for x86_64; cmovbw and cmovbl which were incorrectly handled for i386 are now fixed.
+ Test for correct assembling of cmov.

git-svn-id: trunk@17353 -
2011-04-20 11:18:13 +00:00
florian
ee54a8b879 * Merged helper branch made by Sven Barth
-- Zusammenführen der Unterschiede zwischen Projektarchiv-URLs in ».«:
 U   rtl/inc/objc1.inc
U    rtl/inc/system.inc
U    rtl/objpas/typinfo.pp
A    tests/test/tchlp30.pp
A    tests/test/thlp35.pp
A    tests/test/tchlp3.pp
A    tests/test/thlp7.pp
A    tests/test/trhlp24.pp
A    tests/test/tchlp13.pp
A    tests/test/thlp44.pp
A    tests/test/thlp18.pp
A    tests/test/trhlp33.pp
A    tests/test/tchlp22.pp
A    tests/test/thlp27.pp
A    tests/test/trhlp16.pp
A    tests/test/tchlp31.pp
A    tests/test/thlp36.pp
A    tests/test/tchlp4.pp
A    tests/test/thlp8.pp
A    tests/test/trhlp25.pp
A    tests/test/trhlp1.pp
A    tests/test/uhlp41a.pp
A    tests/test/tchlp40.pp
A    tests/test/tchlp14.pp
A    tests/test/thlp19.pp
A    tests/test/trhlp34.pp
 U   tests/test/ttpara1.pp
A    tests/test/tchlp23.pp
A    tests/test/thlp28.pp
A    tests/test/urhlp14.pp
A    tests/test/trhlp17.pp
A    tests/test/tchlp32.pp
A    tests/test/uhlp43.pp
A    tests/test/thlp37.pp
A    tests/test/tchlp5.pp
A    tests/test/thlp9.pp
A    tests/test/trhlp26.pp
A    tests/test/trhlp2.pp
A    tests/test/uchlp12.pp
A    tests/test/tchlp41.pp
A    tests/test/uhlp41b.pp
A    tests/test/tchlp15.pp
A    tests/test/trhlp35.pp
 U   tests/test/ttpara2.pp
A    tests/test/tchlp50.pp
A    tests/test/tchlp24.pp
A    tests/test/thlp29.pp
A    tests/test/trhlp18.pp
A    tests/test/thlp10.pp
 U   tests/test/cg/ttincdec.pp
A    tests/test/tchlp33.pp
A    tests/test/thlp38.pp
A    tests/test/tchlp6.pp
A    tests/test/trhlp27.pp
A    tests/test/trhlp3.pp
A    tests/test/tchlp42.pp
A    tests/test/tchlp16.pp
A    tests/test/trhlp36.pp
A    tests/test/tchlp51.pp
A    tests/test/tchlp25.pp
A    tests/test/trhlp19.pp
A    tests/test/thlp11.pp
A    tests/test/tchlp34.pp
A    tests/test/thlp39.pp
A    tests/test/tchlp7.pp
A    tests/test/trhlp28.pp
A    tests/test/thlp20.pp
A    tests/test/trhlp4.pp
A    tests/test/tchlp43.pp
A    tests/test/tchlp17.pp
A    tests/test/trhlp37.pp
A    tests/test/tchlp52.pp
A    tests/test/tchlp26.pp
A    tests/test/thlp1.pp
A    tests/test/urhlp17.pp
A    tests/test/thlp12.pp
A    tests/test/tchlp35.pp
A    tests/test/tchlp8.pp
A    tests/test/trhlp29.pp
A    tests/test/thlp21.pp
A    tests/test/trhlp5.pp
A    tests/test/tchlp44.pp
A    tests/test/tchlp18.pp
A    tests/test/trhlp10.pp
A    tests/test/trhlp38.pp
A    tests/test/thlp30.pp
A    tests/test/tchlp53.pp
A    tests/test/tchlp27.pp
A    tests/test/thlp2.pp
A    tests/test/thlp13.pp
A    tests/test/tchlp36.pp
A    tests/test/tchlp9.pp
A    tests/test/thlp22.pp
A    tests/test/trhlp6.pp
A    tests/test/tchlp45.pp
A    tests/test/tchlp19.pp
A    tests/test/trhlp11.pp
A    tests/test/trhlp39.pp
A    tests/test/thlp31.pp
A    tests/test/tchlp54.pp
A    tests/test/tchlp28.pp
A    tests/test/uhlp39.pp
A    tests/test/thlp3.pp
A    tests/test/trhlp20.pp
A    tests/test/thlp40.pp
A    tests/test/thlp14.pp
A    tests/test/tchlp37.pp
A    tests/test/thlp23.pp
A    tests/test/trhlp7.pp
A    tests/test/tchlp46.pp
A    tests/test/trhlp12.pp
A    tests/test/thlp32.pp
A    tests/test/tchlp29.pp
A    tests/test/thlp4.pp
A    tests/test/trhlp21.pp
A    tests/test/tchlp10.pp
A    tests/test/thlp41.pp
A    tests/test/thlp15.pp
A    tests/test/tchlp38.pp
U    tests/test/trtti1.pp
A    tests/test/trhlp30.pp
A    tests/test/thlp24.pp
A    tests/test/trhlp8.pp
A    tests/test/uchlp18.pp
A    tests/test/tchlp47.pp
A    tests/test/trhlp13.pp
A    tests/test/thlp33.pp
A    tests/test/tchlp1.pp
A    tests/test/thlp5.pp
A    tests/test/trhlp22.pp
A    tests/test/tchlp11.pp
A    tests/test/thlp42.pp
A    tests/test/thlp16.pp
A    tests/test/tchlp39.pp
A    tests/test/uhlp3.pp
A    tests/test/trhlp31.pp
A    tests/test/tchlp20.pp
A    tests/test/uhlp31.pp
A    tests/test/thlp25.pp
A    tests/test/trhlp9.pp
A    tests/test/tchlp48.pp
A    tests/test/trhlp40.pp
A    tests/test/trhlp14.pp
A    tests/test/thlp34.pp
A    tests/test/tchlp2.pp
A    tests/test/thlp6.pp
A    tests/test/trhlp23.pp
A    tests/test/tchlp12.pp
A    tests/test/thlp43.pp
A    tests/test/thlp17.pp
A    tests/test/trhlp32.pp
A    tests/test/tchlp21.pp
A    tests/test/thlp26.pp
A    tests/test/tchlp49.pp
A    tests/test/trhlp41.pp
A    tests/test/trhlp15.pp
 U   tests/webtbf/tw13815.pp
 U   tests/tbf/tb0216.pp
U    compiler/msgtxt.inc
U    compiler/nld.pas
U    compiler/dbgdwarf.pas
U    compiler/sparc/cgcpu.pas
U    compiler/fmodule.pas
U    compiler/i386/cgcpu.pas
U    compiler/msgidx.inc
U    compiler/pdecsub.pas
U    compiler/ncgld.pas
U    compiler/symdef.pas
U    compiler/nobj.pas
U    compiler/nflw.pas
U    compiler/pdecobj.pas
U    compiler/objcdef.pas
U    compiler/pinline.pas
U    compiler/pexpr.pas
U    compiler/ncgrtti.pas
U    compiler/x86_64/cgcpu.pas
U    compiler/htypechk.pas
U    compiler/parser.pas
U    compiler/tokens.pas
U    compiler/symbase.pas
U    compiler/ncal.pas
U    compiler/symtable.pas
U    compiler/pp.lpi
U    compiler/m68k/cgcpu.pas
U    compiler/ppu.pas
U    compiler/rautils.pas
U    compiler/arm/cgcpu.pas
U    compiler/ptype.pas
U    compiler/symconst.pas
U    compiler/mips/cgcpu.pas
U    compiler/msg/errore.msg
U    compiler/psub.pas
U    compiler/utils/ppudump.pp
 U   compiler/utils
U    compiler/pdecvar.pas
U    compiler/ncgcal.pas
U    compiler/pdecl.pas
U    compiler/dbgstabs.pas
U    compiler/ppcgen/cgppc.pas
 U   packages/fcl-xml/tests/README_DOM.txt
 U   packages/unzip/src/unzip51g.pp
 U   packages/winunits-base/src/dwmapi.pp
 U   .

git-svn-id: trunk@17328 -
2011-04-16 14:01:44 +00:00
sergei
06311e5aad * i386-win32: Pass const variants by reference to safecall procedures, confirmed this behavior with Delphi (in general, safecall on win32 must behave as close as possible to stdcall).
git-svn-id: trunk@17323 -
2011-04-15 12:12:36 +00:00
svenbarth
35b47e491c Rebase to revision 17306
git-svn-id: branches/svenbarth/classhelpers@17314 -
2011-04-13 10:04:14 +00:00
florian
c97869bd89 * merge user sections support branch
git-svn-id: trunk@17285 -
2011-04-10 18:08:59 +00:00
sergei
c6accaba5c * tcg386.g_copyvaluepara_openarray, improved generated code:
- second calculation of EDI is obsolete, its value is never used.
  - calculation of ECX can be partially shared with first calculation of EDI.
  - multiplying ECX by 1 is also redundant (was happening due to a wrong condition).

git-svn-id: trunk@17260 -
2011-04-06 08:54:20 +00:00
florian
9279c6955e * support for SSSE3, SSE4,1, SSE4.2, AES instructions set by Emelyanov Roman, resolves #18527
+ test for aes support

git-svn-id: trunk@17256 -
2011-04-05 20:22:57 +00:00
florian
f328b6d635 + user section type
+ parsing of section directive for variables
  + section test
  + write section names in the assembler/binary writers correctly
  * allow section only after ; and for embedded targets

git-svn-id: branches/usersections@17154 -
2011-03-20 15:42:28 +00:00
svenbarth
96116a6c3a Several adjustments because virtual methods in helpers are just normal methods and a VMT isn't generated for them either.
* $CPU/cgcpu.pas: disable the generation of VMT loading code
* dbgstabs.pas, dbgdwarf.pas: treat virtual methods of helpers as normal methods
* ncgcal.pas: don't register virtual helper methods for WPO 
* ncgrtti.pas: write virtual helper methods as normal methods to RTTI
* nobj.pas: correctly handle final and override cases in helpers
* pdecvar.pas: property getters
* rautils.pas: no VMT offset in records

git-svn-id: branches/svenbarth/classhelpers@17150 -
2011-03-20 10:41:45 +00:00
svenbarth
d12b198c7f Rebase to revision 16888
git-svn-id: branches/svenbarth/classhelpers@16891 -
2011-02-07 20:30:48 +00:00
florian
e7fa076b70 * handle variables without explicit size prefix better in inline assembler, resolves #18009
git-svn-id: trunk@16859 -
2011-01-30 21:57:30 +00:00
florian
af020e8ebc * fix assembling of ljmp mem, resolves #18556
git-svn-id: trunk@16802 -
2011-01-23 12:18:14 +00:00
florian
0be59445f8 + basic i386-embedded implementation for multiboot by Jeppe Johansen, resolves #15676
git-svn-id: trunk@16801 -
2011-01-23 11:18:50 +00:00
Jonas Maebe
d10f46ec59 * skip instructions containing a reference with a segment prefix (fixes
mantis #18113)

git-svn-id: trunk@16619 -
2010-12-23 15:24:29 +00:00
sergei
34ffb184a8 * MASM and NASM assembler writers, added check for InlineLevel>0 so it suppresses entire source file processing, not just output of source lines. Without this, auto-generated instructions cause output of bogus line information. GAS writer already has similar check.
git-svn-id: trunk@16581 -
2010-12-17 18:03:56 +00:00
paul
56bf42de57 compiler: implement record methods and class methods:
- rename tprocdef._class to tprocdef.struct and change the type from tobjectdef to tabstractrecorddef because methods can belong not to classes only now but to records too
  - replace in many places use of current_objectdef to current_structdef with typcast where is needed
  - add an argument to comp_expr, expr, factor, sub_expr to notify that we are searching type only symbol to solve the problem with records,objects,classes which contains fields with the same name as previosly declared type (like:
  HWND = type Handle;
  rec = record 
    hWnd: HWND;
  end;)
  - disable check in factor_read_id which was made for object that only static fields can be accessed as TObjectType.FieldName outside the object because it makes SizeOf(TObjectType.FieldName) imposible and since the same method was extended to handle records it also breaks a52 package compilation
  - rename tcallcandidates.collect_overloads_in_class to tcallcandidates.collect_overloads_in_struct and addapt the code to handle overloads in records too
  - fix searchsym_type to search also in object ancestors if we found an object symtable
  - add pd_record, pd_notrecord flags to mark procedure modifies which can or can't be used with records. Disallow the next modifiers for records: abstract, dynamic, export, external, far, far16, final, forward, internconst, internproc, interrupt, message, near, override, public, reintroduce, virtual, weakexternal,
Allow the next modifiers for records: static

git-svn-id: branches/paul/extended_records@16526 -
2010-12-09 02:24:46 +00:00