RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-10-26 14:33:02 +02:00
Code Issues Projects Releases Wiki Activity
69,107 Commits 4 Branches 5 Tags 704 MiB
569cd41943
Commit Graph

6862 Commits

Author SHA1 Message Date
Chris Lattner
24ba201941 implement aliases for div/idiv that have an explicit A register operand, 
implementing rdar://8431864

llvm-svn: 118364
 2010-11-06 22:41:18 +00:00
Chris Lattner
25f31dcc8f add aliases for movs between seg registers and mem. There are multiple 
different forms of this instruction (movw/movl/movq) which we reported
as being ambiguous.  Since they all do the same thing, gas just picks the
one with the shortest encoding.  Follow its lead here.

This implements rdar://8208615

llvm-svn: 118362
 2010-11-06 22:35:34 +00:00
Chris Lattner
5cd42cb6f4 move the "sh[lr]d op,op" -> "shld $1, op,op" aliases to the .td file. 
llvm-svn: 118361
 2010-11-06 22:25:39 +00:00
Chris Lattner
95add4d79a work-in-progress 
llvm-svn: 118358
 2010-11-06 22:05:43 +00:00
Chris Lattner
0d6d870628 go to great lengths to work around a GAS bug my previous patch 
exposed:

GAS doesn't accept "fcomip %st(1)", it requires "fcomip %st(1), %st(0)"
even though st(0) is implicit in all other fp stack instructions.

Fortunately, there is an alias for fcomip named "fcompi" and gas does
accept the default argument for the alias (boggle!).

As such, switch the canonical form of this instruction to "pi" instead
of "ip".  This makes the code generator and disassembler generate pi,
avoiding the gas bug.

llvm-svn: 118356
 2010-11-06 21:37:06 +00:00
Chris Lattner
bbd318b76a rework the rotate-by-1 instructions to be defined like the 
shift-by-1 instructions, where the asmstring doesn't contain
the implicit 1.  It turns out that a bunch of these rotate
instructions were completely broken because they used 1 
instead of $1.

This fixes assembly mismatches on "rclb	$1, %bl" and friends,
where we used to generate the 3 byte form, we now generate the
proper 2-byte form.

llvm-svn: 118355
 2010-11-06 21:23:40 +00:00
Chris Lattner
af1f50772a change the fp comparison instructions to not have %st0 explicitly 
listed in its asm string, for consistency with the other similar
instructions.

llvm-svn: 118354
 2010-11-06 20:55:09 +00:00
Chris Lattner
b64f8e10ab move the plethora of fp stack aliases to the .td file. 
llvm-svn: 118353
 2010-11-06 20:47:38 +00:00
Chris Lattner
4834890f0a add (and document) the ability for alias results to have 
fixed physical registers.  Start moving fp comparison
aliases to the .td file (which default to using %st1 if
nothing is specified).

llvm-svn: 118352
 2010-11-06 19:57:21 +00:00
Chris Lattner
c0e756dc47 generalize alias support to allow the result of an alias to 
add fixed immediate values.  Move the aad and aam aliases to
use this, and document it.

llvm-svn: 118350
 2010-11-06 19:25:43 +00:00
Chris Lattner
a7c7312e55 move fnstsw aliases to .td file, fix typo 
llvm-svn: 118349
 2010-11-06 18:58:32 +00:00
Chris Lattner
b686fb93c7 move in/out aliases to the .td files. 
llvm-svn: 118348
 2010-11-06 18:52:40 +00:00
Chris Lattner
6b905e3a10 move sldt, imul, and movabsq aliases from c++ to .td file. 
llvm-svn: 118347
 2010-11-06 18:44:26 +00:00
Chris Lattner
1beb2b3fc5 correct suffix matching to search for s/l/t suffixes on 
floating point stack instructions instead of looking for b/w/l/q.

This fixes issues where we'd accidentally match fistp to fistpl,
when it is in fact an ambiguous instruction.

This changes the behavior of llvm-mc to reject fstp, which was the
correct fix for rdar://8456389:
t.s:1:1: error: ambiguous instructions require an explicit suffix (could be 'fstps', 'fstpl', or 'fstpt')
fstp	(%rax)

it also causes us to correctly reject fistp and fist, which addresses
PR8528:

t.s:2:1: error: ambiguous instructions require an explicit suffix (could be 'fistps', or 'fistpl')
fistp (%rax)
^
t.s:3:1: error: ambiguous instructions require an explicit suffix (could be 'fists', or 'fistl')
fist (%rax)
^

Thanks to Ismail Donmez for tracking down the issue here!

llvm-svn: 118346
 2010-11-06 18:28:02 +00:00
Chris Lattner
637f4910c3 fix a bug where we had an implicit assumption that the 
result instruction operand numbering matched the result pattern.

Fixing this allows us to move the xchg/test aliases to the .td file.

llvm-svn: 118334
 2010-11-06 08:20:59 +00:00
Chris Lattner
4fd3b4ad94 move the lcall/ljmp aliases to the .td file. 
llvm-svn: 118332
 2010-11-06 07:48:45 +00:00
Chris Lattner
66c868f386 move the "movsd -> movsl" alias to the .td files, 
tidy up the movsx and movzx aliases.

llvm-svn: 118331
 2010-11-06 07:34:58 +00:00
Chris Lattner
ec0ef3385d fix some bugs in the alias support, unblocking changing of "clr" aliases 
from c++ hacks to proper .td InstAlias definitions.  Change them!

llvm-svn: 118330
 2010-11-06 07:31:43 +00:00
Chris Lattner
5d1361e9ed Reimplement BuildResultOperands to be in terms of the result instruction's 
operand list instead of the operand list redundantly declared on the alias
or instruction.

With this change, we finally remove the ins/outs list on the alias.  Before:
  def : InstAlias<(outs GR16:$dst), (ins GR8 :$src),
                  "movsx $src, $dst",
                  (MOVSX16rr8W GR16:$dst, GR8:$src)>;
After:
  def : InstAlias<"movsx $src, $dst",
                  (MOVSX16rr8W GR16:$dst, GR8:$src)>;

This also makes the alias mechanism more general and powerful, which will
be exploited in subsequent patches.

llvm-svn: 118329
 2010-11-06 07:14:44 +00:00
Jim Grosbach
a85416eb77 Allow targets to specify the MachO CPUType/CPUSubtype information. 
llvm-svn: 118288
 2010-11-05 18:48:58 +00:00
Duncan Sands
1af5c00d3c When passing a huge parameter using the byval mechanism, a long 
sequence of loads and stores was being generated to perform the
copy on the x86 targets if the parameter was less than 4 byte
aligned, causing llc to use up vast amounts of memory and time.
Use a "rep movs" form instead.  PR7170.

llvm-svn: 118260
 2010-11-04 21:16:46 +00:00
Duncan Sands
3bf2a701a5 In the calling convention logic, ValVT is always a legal type, 
and as such can be represented by an MVT - the more complicated
EVT is not needed.  Use MVT for ValVT everywhere.

llvm-svn: 118245
 2010-11-04 10:49:57 +00:00
Duncan Sands
41edf30895 Simplify uses of MVT and EVT. An MVT can be compared directly 
with a SimpleValueType, while an EVT supports equality and
inequality comparisons with SimpleValueType.

llvm-svn: 118169
 2010-11-03 12:17:33 +00:00
Duncan Sands
f6e5e02c9b Inside the calling convention logic LocVT is always a simple 
value type, so there is no point in passing it around using
an EVT.  Use the simpler MVT everywhere.  Rather than trying
to propagate this information maximally in all the code that
using the calling convention stuff, I chose to do a mainly
low impact change instead.

llvm-svn: 118167
 2010-11-03 11:35:31 +00:00
Chris Lattner
f94c2b782a rearrange a bit. 
llvm-svn: 117967
 2010-11-01 23:07:52 +00:00
Chris Lattner
b0ac93cf15 use our fancy new MnemonicAlias mechanism to remove a bunch of hacks 
from X86AsmParser.cpp

llvm-svn: 117952
 2010-11-01 21:06:34 +00:00
Chris Lattner
d595d1f4d7 "mov[zs]x (mem), GR16" are not ambiguous: the mem 
must be 8 bits.  Support this memory form.

llvm-svn: 117902
 2010-11-01 05:41:10 +00:00
Chris Lattner
81d051481e Implement enough of the missing instalias support to get 
aliases installed and working.  They now work when the
matched pattern and the result instruction have exactly
the same operand list.

This is now enough for us to define proper aliases for
movzx and movsx, implementing rdar://8017633 and PR7459.

Note that we do not accept instructions like:
  movzx 0(%rsp), %rsi

GAS accepts this instruction, but it doesn't make any
sense because we don't know the size of the memory
operand.  It could be 8/16/32 bits.

llvm-svn: 117901
 2010-11-01 05:34:34 +00:00
Chris Lattner
0a4807eefc make the asm matcher emitter reject instructions that have comments 
in their asmstring.  Fix the two x86 "NOREX" instructions that have them.
If these comments are important, the instlowering stuff can print them.

llvm-svn: 117897
 2010-11-01 04:44:29 +00:00
Chris Lattner
9da275f86b reject instructions that contain a \n in their asmstring. Mark 
various X86 and ARM instructions that are bitten by this as isCodeGenOnly,
as they are.

llvm-svn: 117884
 2010-11-01 00:46:16 +00:00
Chris Lattner
a4c36d0efe fix the !eq operator in tblgen to return a bit instead of an int. 
Use this to make the X86 and ARM targets set isCodeGenOnly=1 
automatically for their instructions that have Format=Pseudo,
resolving a hack in tblgen.

llvm-svn: 117862
 2010-10-31 19:22:57 +00:00
Chris Lattner
5d088218e5 two changes: make the asmmatcher generator ignore ARM pseudos properly, 
and make it a hard error for instructions to not have an asm string.
These instructions should be marked isCodeGenOnly.

llvm-svn: 117861
 2010-10-31 19:15:18 +00:00
Chris Lattner
8aaac91ca4 sketch out the planned instruction alias mechanism, add some comments about 
how the push/pop mnemonic aliases are wrong.

llvm-svn: 117857
 2010-10-31 18:43:46 +00:00
Duncan Sands
92f33ea784 Factorize the duplicated logic for choosing the right argument 
calling convention out of the fast and normal ISel files, and
into the calling convention TD file.

llvm-svn: 117856
 2010-10-31 13:21:44 +00:00
Duncan Sands
0f49c49476 Remove CCAssignFnForRet from X86 FastISel in favour of RetCC_X86, 
which has the same logic specified in the CallingConv TD file.
This brings FastISel in line with the standard X86 ISel.

llvm-svn: 117855
 2010-10-31 13:02:38 +00:00
Chris Lattner
49227ad505 Resolve a terrible hack in tblgen: instead of hardcoding 
"In32BitMode" and "In64BitMode" into tblgen, allow any
predicate that inherits from AssemblerPredicate.

llvm-svn: 117831
 2010-10-30 19:38:20 +00:00
Chris Lattner
15e92ddd01 Implement (and document!) support for MnemonicAlias's to have Requires 
directives, allowing things like this:

def : MnemonicAlias<"pop", "popl">, Requires<[In32BitMode]>;
def : MnemonicAlias<"pop", "popq">, Requires<[In64BitMode]>;

Move the rest of the X86 MnemonicAliases over to the .td file.

llvm-svn: 117830
 2010-10-30 19:23:13 +00:00
Chris Lattner
441672d7cb really zap alias. 
llvm-svn: 117824
 2010-10-30 18:23:25 +00:00
Chris Lattner
ba9271be5b move fcompi alias to .td file and zap some useless code. 
llvm-svn: 117823
 2010-10-30 18:22:53 +00:00
Chris Lattner
1b3b2e113f move rep aliases to td file 
llvm-svn: 117822
 2010-10-30 18:17:33 +00:00
Chris Lattner
49f977366f move sal aliases to .td file. 
llvm-svn: 117821
 2010-10-30 18:14:54 +00:00
Chris Lattner
54892b4d8d fix an encoding mismatch where "sal %eax, 1" was not using the short encoding 
for shl.  Caught by inspection.

llvm-svn: 117820
 2010-10-30 18:13:10 +00:00
Chris Lattner
38179edecd move a bunch more aliases from .cpp -> .td file. 
llvm-svn: 117819
 2010-10-30 18:07:17 +00:00
Chris Lattner
ac7f4957b1 move cmov aliases to .td file. 
llvm-svn: 117818
 2010-10-30 17:56:50 +00:00
Chris Lattner
acec7b7d46 move setcc and jcc aliases from .cpp to .td 
llvm-svn: 117817
 2010-10-30 17:51:45 +00:00
Chris Lattner
e90ff8dcb6 move some code. 
llvm-svn: 117816
 2010-10-30 17:38:55 +00:00
Chris Lattner
7c61e4bca2 implement (and document!) the first kind of MC assembler alias, which 
just remaps one mnemonic to another.  Convert a few of the X86 aliases
from .cpp to .td code.

llvm-svn: 117815
 2010-10-30 17:36:36 +00:00
Jim Grosbach
951667ac39 Clean up comments. 
llvm-svn: 117785
 2010-10-30 13:48:28 +00:00
Chris Lattner
45c679522d stay out of the reserved namespace 
llvm-svn: 117773
 2010-10-30 04:57:14 +00:00
John Thompson
6115a7f1d4 Inline asm multiple alternative constraints development phase 2 - improved basic logic, added initial platform support. 
llvm-svn: 117667
 2010-10-29 17:29:13 +00:00
Dale Johannesen
b78530f9b0 Fix pastos in handling of AVX cvttsd2si, PR8491. 
Bruno, please review, but I'm pretty sure this is right.
Patch by Alex Mac!

llvm-svn: 117514
 2010-10-28 00:35:54 +00:00
Kevin Enderby
a53cc6a764 Added the x86 instruction ud2b (2nd official undefined instruction). 
llvm-svn: 117485
 2010-10-27 20:46:49 +00:00
Michael J. Spencer
5518dda87e x86-Win32: Switch ftol2 calling convention from stdcall to C. 
llvm-svn: 117474
 2010-10-27 18:52:38 +00:00
Kevin Enderby
74a2614673 Yet another tweak to X86 instructions to add ud2a as an alias to ud2 
(still to add ud2b).

llvm-svn: 117435
 2010-10-27 03:01:02 +00:00
Kevin Enderby
d22f3b9de7 Another tweak to X86 instructions to add the missing flex instruction (without 
the wait prefix).

llvm-svn: 117434
 2010-10-27 02:53:04 +00:00
Kevin Enderby
e812b356cc Tweaks to X86 instructions to allow the 'w' suffix in places it makes 
sense, when the instruction takes the 16-bit ax register or m16 memory
location.  These changes to llvm-mc matches what the darwin assembler
allows for these instructions.  Done differently than in r117031 that
caused a valgrind error which was later reverted.

llvm-svn: 117433
 2010-10-27 02:32:19 +00:00
Kevin Enderby
d5235bb45c Added some aliases to the fcomip and fucompi Intel instructions. So that llvm-mc 
will accept versions that the darwin assembler allows.  Forms ending in "pi" and
forms without all the operands.

llvm-svn: 117427
 2010-10-27 00:59:28 +00:00
Dale Johannesen
e7f07349e4 Use a MemIntrinsicSDNode for ISD::PREFETCH, which touches 
memory, so a MachineMemOperand is useful (not propagated
into the MachineInstr yet).  No functional change except
for dump output.

llvm-svn: 117413
 2010-10-26 23:11:10 +00:00
Rafael Espindola
adaf2ea5c6 handle X86::EH_RETURN64 and X86::EH_RETURN. 
llvm-svn: 117378
 2010-10-26 18:09:55 +00:00
Rafael Espindola
e38790534e Implement some relaxations for arithmetic instructions. The limitation 
on RIP relative relocations looks artificial, but this is a superset of
what we were able to do before.

llvm-svn: 117364
 2010-10-26 14:09:12 +00:00
Dale Johannesen
2566d39d07 An stdcall function calling a non-stdcall function 
cannot use tailcall.  PR 8461.

llvm-svn: 117322
 2010-10-25 22:17:05 +00:00
Rafael Espindola
c6e7790c4b Add X86::reloc_global_offset_table and use it to have a single place where 
we check for _GLOBAL_OFFSET_TABLE_.

llvm-svn: 117241
 2010-10-24 17:35:42 +00:00
Benjamin Kramer
ed68a7877c Make the disassembler tables const so they end up in read-only memory. 
llvm-svn: 117206
 2010-10-23 09:10:44 +00:00
Michael J. Spencer
87c8212d41 X86: Emit _fltused instead of __fltused on Windows x64. 
llvm-svn: 117205
 2010-10-23 09:06:59 +00:00
Chandler Carruth
ea67aaa252 Remove a define which is never referenced. 
llvm-svn: 117202
 2010-10-23 08:25:16 +00:00
Chandler Carruth
7799a873bb Switch attribute macros to use 'LLVM_' as a prefix. We retain the old names 
until other LLVM projects using these are cleaned up.

llvm-svn: 117200
 2010-10-23 08:10:43 +00:00
Wesley Peck
488027efa3 Making the e_machine configurable by the target backend in ELFObjectWriter. 
llvm-svn: 117099
 2010-10-22 15:52:49 +00:00
Andrew Trick
b9e8700beb Reverting r117031 to cleanup valgrind errors. 
It doesn't look like anything is wrong with the checkin,
but the new test cases expose a mem bug in AsmParser.

llvm-svn: 117087
 2010-10-22 03:58:29 +00:00
Sean Callanan
9a40554de4 Fixed handling of immediate operand sizes, which 
weren't properly reflecting the OperandSize attribute
of the instruction leading to improper decoding of
certain instructions with the 66H prefix.  Also added
a test case for this.

llvm-svn: 117084
 2010-10-22 01:24:11 +00:00
Kevin Enderby
c63105a064 More tweaks to X86 instructions to allow the 'w' suffix in places it makes 
sense, when the instruction takes the 16-bit ax register or m16 memory
location.  These changes to llvm-mc matches what the darwin assembler allows
for these instructions.  Also added the missing flex (without the wait prefix)
and ud2a as an alias to ud2 (still to add ud2b).

llvm-svn: 117031
 2010-10-21 17:16:46 +00:00
Duncan Sands
0c166168a0 Remove some variables that are never really used 
(gcc-4.6 warns about these).

llvm-svn: 117021
 2010-10-21 16:03:28 +00:00
Duncan Sands
8c5f243aa0 Add parentheses to pacify gcc, which warns otherwise. 
llvm-svn: 117020
 2010-10-21 16:02:12 +00:00
Michael J. Spencer
5a68d7ce94 X86: Add alloca probing to dynamic alloca on Windows. Fixes PR8424. 
llvm-svn: 116984
 2010-10-21 01:41:01 +00:00
Michael J. Spencer
b9cffadc06 CodeGen-Windows: Only emit _fltused if a VarArg function is called with floating point args. 
This should be the minimum set of functions that could possibly need it.

llvm-svn: 116978
 2010-10-21 00:08:21 +00:00
Michael J. Spencer
54b462089f Fix Whitespace. 
llvm-svn: 116972
 2010-10-20 23:40:27 +00:00
Dale Johannesen
58fe3193a6 Remove Synthesizable from the Type system; as MMX vector 
types are no longer Legal on X86, we don't need it.
No functional change.  8499854.

llvm-svn: 116947
 2010-10-20 21:32:10 +00:00
Rafael Espindola
0a50bbaee1 Handle _GLOBAL_OFFSET_TABLE_ correctly. 
llvm-svn: 116932
 2010-10-20 16:46:08 +00:00
Evan Cheng
1c8dafd12a Re-enable register pressure aware machine licm with fixes. Hoist() may have 
erased the instruction during LICM so UpdateRegPressureAfter() should not
reference it afterwards.

llvm-svn: 116845
 2010-10-19 18:58:51 +00:00
Daniel Dunbar
6ff550c84d Revert r116781 "- Add a hook for target to determine whether an instruction def 
is", which breaks some nightly tests.

llvm-svn: 116816
 2010-10-19 17:14:24 +00:00
Michael J. Spencer
e528f2588f X86: Add MS-CRT libcalls. 
llvm-svn: 116801
 2010-10-19 07:32:52 +00:00
Michael J. Spencer
cd6be63d05 Fix Whitespace. 
llvm-svn: 116800
 2010-10-19 07:32:42 +00:00
Evan Cheng
9c3f6f486e - Add a hook for target to determine whether an instruction def is 
"long latency" enough to hoist even if it may increase spilling. Reloading
  a value from spill slot is often cheaper than performing an expensive
  computation in the loop. For X86, that means machine LICM will hoist
  SQRT, DIV, etc. ARM will be somewhat aggressive with VFP and NEON
  instructions.
- Enable register pressure aware machine LICM by default.

llvm-svn: 116781
 2010-10-19 00:55:07 +00:00
Kevin Enderby
8a3030c6af Added a few tweaks to the Intel Descriptor-table support instructions to allow 
word forms and suffixed versions to match the darwin assembler in 32-bit and
64-bit modes.  This is again for use just with assembly source for llvm-mc .

llvm-svn: 116773
 2010-10-19 00:01:44 +00:00
Rafael Espindola
7cc236c87f Produce ELF::R_386_GOTPC relocations. 
llvm-svn: 116728
 2010-10-18 18:36:12 +00:00
Kevin Enderby
cb18d38b76 Added a handful of x86-32 instructions that were missing so that llvm-mc would 
be more complete.  These are only expected to be used by llvm-mc with assembly
source so there is no pattern, [], in the .td files.  Most are being added to
X86InstrInfo.td as Chris suggested and only comments about register uses are
added.  Suggestions welcome on the .td changes as I'm not sure on every detail
of the x86 records.  More missing instructions will be coming.

llvm-svn: 116716
 2010-10-18 17:04:36 +00:00
Rafael Espindola
be5c52d2dc Add a MCObjectFormat class so that code common to all targets that use a 
single object format can be shared.

This also adds support for

mov zed+(bar-foo), %eax

on ELF and COFF targets.

llvm-svn: 116675
 2010-10-16 18:23:53 +00:00
Michael J. Spencer
e57b670425 X86-Windows: Emit an undefined global __fltused symbol when targeting Windows 
if any floating point arguments are passed to an external function.

llvm-svn: 116665
 2010-10-16 08:25:41 +00:00
Rafael Espindola
b1ae74bd73 Fix another case where we were preferring instructions with large 
immediates instead of 8 bits ones.

llvm-svn: 116410
 2010-10-13 17:14:25 +00:00
Rafael Espindola
ff7f11c151 Fix PR8365 by adding a more specialized Pat that checks if an 'and' with 
8 bit constants can be used.

llvm-svn: 116403
 2010-10-13 13:31:20 +00:00
Evan Cheng
eefee972cf Turn some fp stackifier assertion into errors to avoid silently generating bad code when assertions are off. rdar://8540457. 
llvm-svn: 116368
 2010-10-12 23:19:28 +00:00
Eric Christopher
303692631e Combine these together - should probably have some text associated 
that says what why what we just asserted is wrong.

llvm-svn: 116333
 2010-10-12 19:44:17 +00:00
Nick Lewycky
a6815ae877 Mark variable 'NoImplicitFloatOps' used only in an assert as used. 
llvm-svn: 116323
 2010-10-12 18:18:03 +00:00
Dan Gohman
d904add908 Initial va_arg support for x86-64. Patch by David Meyer! 
llvm-svn: 116319
 2010-10-12 18:00:49 +00:00
Jakob Stoklund Olesen
499fe39d23 Remove the x86 MOV{32,64}{rr,rm,mr}_TC instructions. 
The reg-reg copies were no longer being generated since copyPhysReg copies
physical registers only.

The loads and stores are not necessary - The TC constraint is imposed by the
TAILJMP and TCRETURN instructions, there should be no need for constrained loads
and stores.

llvm-svn: 116314
 2010-10-12 17:15:00 +00:00
Andrew Trick
5704a15e36 Fixes bug 8297: i386 cmpxchg8b, missing MachineMemOperand 
llvm-svn: 116214
 2010-10-11 19:02:04 +00:00
Michael J. Spencer
08b96a9209 X86: Call ulldiv and ftol2 on Windows instead of their libgcc eqivilents. 
llvm-svn: 116188
 2010-10-11 05:29:15 +00:00
Michael J. Spencer
8f7251a0e9 X86: MinGW should always use libgcc on Windows. 
llvm-svn: 116177
 2010-10-10 23:11:06 +00:00
Michael J. Spencer
b13cb4dd72 X86: Call _alldiv instead of __divdi3 on Windows (excluding cygwin). 
llvm-svn: 116174
 2010-10-10 22:04:34 +00:00
Michael J. Spencer
ae56350c76 Fix Whitespace. 
llvm-svn: 116173
 2010-10-10 22:04:20 +00:00
Michael J. Spencer
1c60bd155b Fix Whitespace. 
llvm-svn: 116149
 2010-10-09 11:00:50 +00:00
Chris Lattner
68ff81a2e6 add jit support for the new psuedo instructions I added for 
the add/or xform.  The JIT isn't mcized yet, boo.

This fixes Olden/voronoi, bh and a ton of other stuff that
uses the jit.

llvm-svn: 116125
 2010-10-08 23:59:27 +00:00
Chris Lattner
2176ca6377 machine a mutable machineinstr down into emitInstruction. 
llvm-svn: 116124
 2010-10-08 23:54:01 +00:00
Cameron Esfahani
664317d6cd Recommit 116056, now with the missing file... 
llvm-svn: 116083
 2010-10-08 19:24:18 +00:00
Andrew Trick
0d8a3d67e0 reverting 116056: win64_params.ll may need to be conditionalized? 
llvm-svn: 116063
 2010-10-08 17:22:42 +00:00
Cameron Esfahani
a9f8bb1356 Small patch to restore home register stack space allocation for the Win64 case. Add test case. This code eventually needs to be tighter, since it's always allocating it, even in leaf routines. 
llvm-svn: 116056
 2010-10-08 10:31:30 +00:00
Chris Lattner
244e13f439 fix a subtle bug I introduced in my refactoring, where we stopped preferring 
the i8 versions of instructions in some cases.  In test6, we started 
generating:

	cmpq	$0, -8(%rsp)            ## encoding: [0x48,0x81,0x7c,0x24,0xf8,0x00,0x00,0x00,0x00]
                                        ## <MCInst #478 CMP64mi32
                                        ##  <MCOperand Reg:114>
                                        ##  <MCOperand Imm:1>
                                        ##  <MCOperand Reg:0>
                                        ##  <MCOperand Imm:-8>
                                        ##  <MCOperand Reg:0>
                                        ##  <MCOperand Imm:0>>

instead of:

	cmpq	$0, -8(%rsp)            ## encoding: [0x48,0x83,0x7c,0x24,0xf8,0x00]
                                        ## <MCInst #479 CMP64mi8
                                        ##  <MCOperand Reg:114>
                                        ##  <MCOperand Imm:1>
                                        ##  <MCOperand Reg:0>
                                        ##  <MCOperand Imm:-8>
                                        ##  <MCOperand Reg:0>
                                        ##  <MCOperand Imm:0>>

Fix this and add some comments.

llvm-svn: 116053
 2010-10-08 05:12:14 +00:00
Chris Lattner
82ce325f16 reapply: Use the new TB_NOT_REVERSABLE flag instead of special 
reapply: reimplement the second half of the or/add optimization.  We should now

with no changes.  Turns out that one missing "Defs = [EFLAGS]" can upset things
a bit.

llvm-svn: 116040
 2010-10-08 03:57:25 +00:00
Chris Lattner
fbdd285dd6 reapply the patch reverted in r116033: 
"Reimplement (part of) the or -> add optimization.  Matching 'or' into 'add'"

With a critical fix: the add pseudos clobber EFLAGS.

llvm-svn: 116039
 2010-10-08 03:54:52 +00:00
Daniel Dunbar
d3b6b8bf2b Revert "Reimplement (part of) the or -> add optimization. Matching 'or' into 
'add'", which seems to have broken just about everything.

llvm-svn: 116033
 2010-10-08 02:07:32 +00:00
Daniel Dunbar
59848f6703 Revert "Use the new TB_NOT_REVERSABLE flag instead of special ", which depends 
on r116007, which I am about to revert.

llvm-svn: 116032
 2010-10-08 02:07:29 +00:00
Daniel Dunbar
983fae5a86 Revert "reimplement the second half of the or/add optimization. We should now", 
which depends on r116007, which I am about to revert.

llvm-svn: 116031
 2010-10-08 02:07:26 +00:00
Chris Lattner
7577cb7b49 reimplement the second half of the or/add optimization. We should now 
only end up emitting LEA instead of OR.  If we aren't able to promote
something into an LEA, we should never be emitting it as an ADD.

Add some testcases that we emit "or" in cases where we used to produce
an "add".

llvm-svn: 116026
 2010-10-08 01:05:10 +00:00
Chris Lattner
d62e94b465 Use the new TB_NOT_REVERSABLE flag instead of special 
casing FsMOVAPDrr/FsMOVAPSrr.

llvm-svn: 116016
 2010-10-08 00:03:02 +00:00
Chris Lattner
72e7e84c3f simplify some map operations. 
llvm-svn: 116014
 2010-10-07 23:57:02 +00:00
Chris Lattner
d8f05bf65e Reimplement (part of) the or -> add optimization. Matching 'or' into 'add' 
is general goodness because it allows ORs to be converted to LEA to avoid
inserting copies.  However, this is bad because it makes the generated .s
file less obvious and gives valgrind heartburn (tons of false positives in
bitfield code).

While the general fix should be in valgrind, we can at least try to avoid
emitting ADD instructions that *don't* get promoted to LEA.  This is more
work because it requires introducing pseudo instructions to represents
"add that knows the bits are disjoint", but hey, people really love valgrind.

This fixes this testcase:
https://bugs.kde.org/show_bug.cgi?id=242137#c20

the add r/i cases are coming next.

llvm-svn: 116007
 2010-10-07 23:36:18 +00:00
Chris Lattner
2212441ff7 Reduce casting in various tables by defining the table 
with the right types.

llvm-svn: 116001
 2010-10-07 23:08:41 +00:00
Chris Lattner
17850b2677 simplify code: don't build up vector only to assert it is empty. 
llvm-svn: 115997
 2010-10-07 22:26:19 +00:00
Chris Lattner
16aa36001d convert test to use the existing classes that the multipatterns 
use.  Since TEST is completely different than all other binops,
don't define a multipattern for it.

This completes factorization of binops.

llvm-svn: 115982
 2010-10-07 21:31:03 +00:00
Chris Lattner
bd89233607 convert cmp to use a multipattern 
llvm-svn: 115978
 2010-10-07 20:56:25 +00:00
Evan Cheng
7c89d70f27 Canonicalize X86ISD::MOVDDUP nodes to v2f64 to make sure all cases match. Also eliminate unneeded isel patterns. rdar://8520311 
llvm-svn: 115977
 2010-10-07 20:50:20 +00:00
Chris Lattner
4b9e6fd743 reduce redundancy between pattern copies. 
llvm-svn: 115968
 2010-10-07 20:14:23 +00:00
Chris Lattner
41b3817cbd the opcode for BinOpMI/BinOpMI8 is always the same, remove the argument. 
llvm-svn: 115967
 2010-10-07 20:06:24 +00:00
Chris Lattner
d7f7dc04f5 convert adc/sbb to a multipattern. Because the adde/sube nodes 
are not defined as returning EFLAGS (like add_flag and friends),
the entire multipattern and several of the subclasses need to be
cloned.

This could be handled through better instantiation support in tblgen,
but it isn't meta enough.

llvm-svn: 115964
 2010-10-07 20:01:55 +00:00
Jakob Stoklund Olesen
7f38b44415 Fix obvious mistake pointed out by Michael Spencer. 
llvm-svn: 115952
 2010-10-07 18:47:10 +00:00
Chris Lattner
f0d9834e08 add support for isConvertibleToThreeAddress to ArithBinOpEFLAGS, 
allowing us to convert ADD over.  deletes 160 lines of .td file.

llvm-svn: 115897
 2010-10-07 01:37:01 +00:00
Chris Lattner
dea1395cfe Fix a few issues in ArithBinOpEFLAGS that made it specific to and. 
Start using ArithBinOpEFLAGS for OR, XOR, and SUB.

This removes 500 lines from the .td file.  Now AND/OR/XOR/SUB are all
defined exactly the same way instead of being close relatives.

llvm-svn: 115896
 2010-10-07 01:26:27 +00:00
Chris Lattner
f1f5df212b Convert 'and' to single instance of a multipattern 
which instantiates the 34 versions of and all in one
swoop.  The BaseOpc/BaseOpc2/BaseOpc4 stuff should not
be required, but tblgen's feeble brain explodes when I
use Or4<BaseOpc>.V in the multipattern.

No change in the generated .inc files.

llvm-svn: 115893
 2010-10-07 01:10:20 +00:00
Chris Lattner
a2b51bb341 add a new BinOpAI class to represent the immediate form that directly acts on EAX. 
This does change the generated .inc files to include the implicit use/def of eax.
Since these instructions are only generated by the assembler and disassembler it
doesn't actually matter though.

llvm-svn: 115885
 2010-10-07 00:43:39 +00:00
Chris Lattner
0df71280d1 add a bunch of classes for other common patterns. 
As usual, no change in generated .inc files.

llvm-svn: 115882
 2010-10-07 00:35:28 +00:00
Chris Lattner
9b3a494cdb Define a new BinOpRI8 class and use it to define the imm8 versions of and. 
llvm-svn: 115880
 2010-10-07 00:12:45 +00:00
Jakob Stoklund Olesen
5e329859cd Constrain the offset register to a *_NOSP register class when inserting LEA 
instructions.

This unbreaks the machine code verifier and fixes PR8317.

llvm-svn: 115879
 2010-10-07 00:07:26 +00:00
Chris Lattner
b520abade4 add the pattern operator to match to X86TypeInfo, use this to 
convert AND64ri32 to use BinOpRI.

llvm-svn: 115878
 2010-10-07 00:01:39 +00:00
Jakob Stoklund Olesen
8c1eafb4cb Properly handle GR32_NOSP in X86RegisterInfo::getMatchingSuperRegClass. 
This function looks like it is about ready to be generated by TebleGen.

llvm-svn: 115876
 2010-10-06 23:56:46 +00:00
Chris Lattner
30c9a175a3 enhance X86TypeInfo to include information about the encoding and 
operand kind for immediates.  Use these to define a new BinOpRI
class and switch AND8/16/32ri over to it.  AND64ri32 needs some
more refactoring before it can make the switcheroo.

llvm-svn: 115752
 2010-10-06 05:55:42 +00:00
Chris Lattner
e9e2d51853 add a class for _REV nodes. 
llvm-svn: 115748
 2010-10-06 05:35:22 +00:00
Chris Lattner
88ccb4e6b5 sink more intelligence into the ITy base class. Now it knows 
that i8 operations are even and i16,i32,i64 operations have a
low opcode bit set (they are odd).

llvm-svn: 115747
 2010-10-06 05:28:38 +00:00
Chris Lattner
96dd50c055 refactor things a bit, now the REX_W and OpSize prefix bytes are inferred from the type info. 
llvm-svn: 115745
 2010-10-06 05:20:57 +00:00
Chris Lattner
7956e6b995 with tblgen suitably extended, we can now get the load node from typeinfo. 
llvm-svn: 115744
 2010-10-06 04:58:43 +00:00
Chris Lattner
5b3e46bd6b lets go all meta and define new X86 type wrappers that declare the associated 
gunk that goes along with an MVT (e.g. reg class, preferred load operation,
memory operand)

llvm-svn: 115727
 2010-10-06 00:45:24 +00:00
Chris Lattner
e979874e3f introduce a new BinOpRM class and use it to factor AND*rm. This points out 
that I need a heavier handed approach to get ultimate factorization.

llvm-svn: 115726
 2010-10-06 00:30:49 +00:00
Chris Lattner
84846b71af remove the !nameconcat tblgen feature. It "shorthand" and only used in 4 places 
where !cast is just as short.

llvm-svn: 115722
 2010-10-06 00:19:21 +00:00
Chris Lattner
12274b9845 allow !strconcat to take more than two operands to eliminate 
!strconcat(!strconcat(!strconcat(!strconcat

Simplify some x86 td files to use it.

llvm-svn: 115719
 2010-10-05 23:58:18 +00:00
Chris Lattner
dd8227d488 associate the instruction suffix letter with the integer gpr 
register class, and use this to simplify use of BinOpRR.

llvm-svn: 115716
 2010-10-05 23:43:04 +00:00
Chris Lattner
dd4c597e38 introduce a new BinOpRR class, and convert 4 and instructions to use it. 
llvm-svn: 115715
 2010-10-05 23:32:05 +00:00
Chris Lattner
ef2e024af8 Move cmov pseudo instructions to InstrCompiler, 
convert all the rest of the cmovs to the multiclass,
with good results:

 X86InstrCMovSetCC.td |  598 +--------------------------------------------------
 X86InstrCompiler.td  |   61 +++++
 2 files changed, 77 insertions(+), 582 deletions(-)

llvm-svn: 115707
 2010-10-05 23:09:10 +00:00
Chris Lattner
195a9c3877 Use #NAME# to have the CMOV multiclass define things with the same names as before 
(e.g. CMOVBE16rr instead of CMOVBErr16).

llvm-svn: 115705
 2010-10-05 23:00:14 +00:00
Chris Lattner
3357066875 enhance tblgen to support anonymous defm's, use this to 
simplify the X86 CMOVmr's.

llvm-svn: 115702
 2010-10-05 22:51:56 +00:00
Chris Lattner
b2ac22f0a4 convert cmov mr patterns to use a multipattern. Death to redundancy 
and verbosity

llvm-svn: 115701
 2010-10-05 22:42:54 +00:00
Chris Lattner
c3c03dfeff switch CMOVBE to the multipattern: 
21 insertions(+), 53 deletions(-)

Moar change coming before I switch the rest.

llvm-svn: 115697
 2010-10-05 22:23:58 +00:00
Chris Lattner
5673933e30 fix a bug I introduced in r115669, which ended up with MOV64mr_TC 
not getting marked as mayStore.  This fixes llvm-gcc bootstrap.

llvm-svn: 115693
 2010-10-05 22:16:48 +00:00
Chris Lattner
a5c35bed7b add a multiclass for cmov's, but don't start using it yet. 
llvm-svn: 115692
 2010-10-05 22:01:02 +00:00
Chris Lattner
7065387c35 use a multipattern to define setcc instructions: 
X86InstrCMovSetCC.td |  200 ++++++---------------------------------------------
 1 file changed, 27 insertions(+), 173 deletions(-)

llvm-svn: 115689
 2010-10-05 21:34:29 +00:00
Chris Lattner
fa6f058b70 move SETB pseudos into the same place in InstrCompiler.td 
llvm-svn: 115686
 2010-10-05 21:18:04 +00:00
Chris Lattner
ca34143ddd Replace a gross hack (the MOV64ri_alt instruction) with a slightly less 
gross hack (having the asmmatcher handle the alias).

llvm-svn: 115685
 2010-10-05 21:09:45 +00:00
Chris Lattner
5d7d5a81eb distribute the rest of the contents of X86Instr64bit.td out to 
the right places.  X86Instr64bit.td now dies, long live x86-64!

llvm-svn: 115669
 2010-10-05 20:49:15 +00:00
Chris Lattner
c06e348ff8 move the rest of the simple 64-bit arithmetic into InstrArithmetic.td 
llvm-svn: 115663
 2010-10-05 20:35:37 +00:00
Chris Lattner
80ffcbd80a continue moving 64-bit stuff into X86InstrArithmetic.td 
llvm-svn: 115660
 2010-10-05 20:23:31 +00:00
Chris Lattner
aae0403342 move 64-bit add and adc to InstrArithmetic. 
llvm-svn: 115632
 2010-10-05 16:59:08 +00:00
Chris Lattner
f5f0742885 rewrote two addr constraints so that they are only set, not set and then nestedly cleared. 
llvm-svn: 115631
 2010-10-05 16:52:25 +00:00
Chris Lattner
cdf60fcc21 split the 32-bit integer arithmetic instructions out to their own file. 
llvm-svn: 115627
 2010-10-05 16:39:12 +00:00
Chris Lattner
4cf3aff9c1 integrate the 64-bit shifts into X86InstrShiftRotate.td. Enough for tonight. 
llvm-svn: 115608
 2010-10-05 07:13:35 +00:00
Chris Lattner
3d1c1ff4c2 move 32-bit shift and rotates out to their own file. 
llvm-svn: 115607
 2010-10-05 07:00:12 +00:00
Chris Lattner
40954218dd add new file 
llvm-svn: 115606
 2010-10-05 06:52:35 +00:00
Chris Lattner
1c0cbe9571 move sign and zero extensions out to their own file. 
llvm-svn: 115605
 2010-10-05 06:52:26 +00:00
Chris Lattner
12a0f5c3bd move some instructions from Instr64Bit -> InstrInfo. 
bswap32 doesn't read eflags.

llvm-svn: 115604
 2010-10-05 06:47:35 +00:00
Chris Lattner
9317bf2ed5 move CMOV_FR32 and friends to InstrCompiler, since they are 
pseudo instructions.

Move POPCNT to InstrSSE since they are SSE4 instructions.

llvm-svn: 115603
 2010-10-05 06:41:40 +00:00
Chris Lattner
d96f3fe646 move various pattern matching support goop out of X86Instr64Bit, to live 
with the 32-bit stuff.

llvm-svn: 115602
 2010-10-05 06:37:31 +00:00
Chris Lattner
7451cc0f59 split conditional moves and setcc's out to their own file. 
llvm-svn: 115601
 2010-10-05 06:33:16 +00:00
Chris Lattner
7114e187e1 move string pseudo instructions to InstrCompiler consolidate 64-bit and 32-bit together. 
llvm-svn: 115600
 2010-10-05 06:27:48 +00:00
Chris Lattner
e63d763713 move the atomic pseudo instructions out to X86InstrCompiler.td 
llvm-svn: 115599
 2010-10-05 06:22:35 +00:00
Chris Lattner
a2e5444bb4 move more pseudo instructions out to X86InstrCompiler.td 
llvm-svn: 115598
 2010-10-05 06:10:16 +00:00
Chris Lattner
383d15c9d8 move VMX instructions out to their own file. 
llvm-svn: 115597
 2010-10-05 06:06:53 +00:00
Chris Lattner
5f59acddbc continue moving stuff out to X86InstrSystem.td. Move 
control flow stuff out to X86InstrControl.td.  Move
some compiler pseudo instructions and Pat<> patterns
out to X86InstrCompiler.td

llvm-svn: 115596
 2010-10-05 06:04:14 +00:00
Chris Lattner
db65ba5acf refactor .td files a bit, moving system instructions out to X86InstrSystem.td 
llvm-svn: 115591
 2010-10-05 05:32:15 +00:00
Bill Wendling
b94ade249a The pshufw instruction came about in MMX2 when SSE was introduced. Don't place 
it in with the SSSE3 instructions.

Steward! Could you place this chair by the aft sun deck? I'm trying to get away
from the Astors. They are such boors!

llvm-svn: 115552
 2010-10-04 20:24:01 +00:00
Anton Korobeynikov
f1acea8615 va_args support for Win64. 
Patch by Cameron!

llvm-svn: 115480
 2010-10-03 22:52:07 +00:00
Anton Korobeynikov
31b3b2ca41 Properly emit stack probe on win64 (for non-mingw targets). 
Based on the patch by Cameron Esfahani!

llvm-svn: 115479
 2010-10-03 22:02:38 +00:00
Eli Friedman
35432e5685 Add 3DNowA instructions. 
llvm-svn: 115477
 2010-10-03 20:23:13 +00:00
Chris Lattner
d03783eaf2 the immediate field of pshufw is actually an 8-bit field, not a 8-bit field that is sign extended. This fixes PR8288 
llvm-svn: 115473
 2010-10-03 19:09:13 +00:00
Rafael Espindola
49d508fd19 Jim Asked us to move DataLayout on ARM back to the most specialized classes. Do 
so and also change X86 for consistency.

Investigating if this can be improved a bit.

llvm-svn: 115469
 2010-10-03 18:59:45 +00:00
Chris Lattner
4599fd89fd add support for the prefetch/prefetchw instructions, move femms into 
the right file.  The assembler supports all the 3dnow instructions now,
but not the "3dnowa" ones.

llvm-svn: 115468
 2010-10-03 18:42:30 +00:00
Chris Lattner
6d6f84f99c what the heck, add support for the rest of the 3dNow! binary operations. 
llvm-svn: 115467
 2010-10-03 18:24:18 +00:00
Chris Lattner
8174253484 Implement support for the bizarre 3DNow! encoding (which is unlike anything 
else in X86), and add support for pavgusb.  This is apparently the
only instruction (other than movsx) that is preventing ffmpeg from building
with clang.

If someone else is interested in banging out the rest of the 3DNow! 
instructions, it should be quite easy now.

llvm-svn: 115466
 2010-10-03 18:08:05 +00:00
Chris Lattner
3d148e7e31 stub out a header to put 3dNow! instructions into. 
llvm-svn: 115429
 2010-10-02 23:06:23 +00:00
Chris Lattner
3c29a2b776 fix a regression introduced in r115243, in which the instruction 
backing int_x86_ssse3_pshuf_w got removed.  This caused PR8280.

llvm-svn: 115422
 2010-10-02 21:32:15 +00:00
Jim Grosbach
2143c9d321 Rename the AsmPrinter directory to InstPrinter for those targets that have 
been MC-ized for assembly printing. MSP430 is mostly so, but still has the
asm printer and lowering code in the printer subdir for the moment.

llvm-svn: 115360
 2010-10-01 22:39:28 +00:00
Benjamin Kramer
65131b20b4 Delete token *after* reading from it. 
llvm-svn: 115311
 2010-10-01 12:25:27 +00:00
Dale Johannesen
c14a1eda84 Massive rewrite of MMX: 
The x86_mmx type is used for MMX intrinsics, parameters and
return values where these use MMX registers, and is also
supported in load, store, and bitcast.

Only the above operations generate MMX instructions, and optimizations
do not operate on or produce MMX intrinsics. 

MMX-sized vectors <2 x i32> etc. are lowered to XMM or split into
smaller pieces.  Optimizations may occur on these forms and the
result casted back to x86_mmx, provided the result feeds into a
previous existing x86_mmx operation.

The point of all this is prevent optimizations from introducing
MMX operations, which is unsafe due to the EMMS problem.

llvm-svn: 115243
 2010-09-30 23:57:10 +00:00
Jim Grosbach
45de3f6747 Clean up asm writer usage for x86 and msp430 to flag that the writer should 
use MC instructions in the printInstruction() method via the tablegen flag
for it rather than a #define prior to including the autogenerated bits.

llvm-svn: 115238
 2010-09-30 23:40:25 +00:00
Chris Lattner
df2f5c0a40 preemptively add the rest of the non-n fpstack instructions. 
llvm-svn: 115168
 2010-09-30 17:11:29 +00:00
Chris Lattner
8e7a4b7b57 implement support for finit, PR8258 
llvm-svn: 115156
 2010-09-30 16:42:53 +00:00
Chris Lattner
fecf3a7717 add support for fstcw, PR8259 
llvm-svn: 115154
 2010-09-30 16:39:29 +00:00
Kevin Enderby
dd3306fcb5 Adds getPointerSize() to the AsmBackend which will be needed by the final patch 
for the dwarf .loc support to emit dwarf line number tables.

llvm-svn: 115153
 2010-09-30 16:38:07 +00:00
Rafael Espindola
480ee577ad Correctly produce R_X86_64_32 or R_X86_64_32S. 
With this patch in

movq    $foo, foo(%rip)
foo:
.long   foo

We produce a R_X86_64_32S for the first relocation and R_X86_64_32 for the
second one.

llvm-svn: 115134
 2010-09-30 03:11:42 +00:00
Eric Christopher
80d620fb38 Noticed by inspection when looking for other cmov bits. 
llvm-svn: 115100
 2010-09-29 23:00:29 +00:00
Nick Lewycky
a533bd63e6 Add parens to fix GCC warning: 
lib/Target/X86/X86MCCodeEmitter.cpp: 190: error: suggest parentheses around '&&' within '||'

llvm-svn: 115064
 2010-09-29 18:56:57 +00:00
Chris Lattner
7f466d63e0 implement rdar://8491845 - Gas supports commuted forms of non-commutable instructions. 
llvm-svn: 115061
 2010-09-29 18:39:16 +00:00
Chris Lattner
9c58de2dc4 fix rdar://8490728 - llvm-mc rejects gpr64 form of 'movmskpd' 
llvm-svn: 115029
 2010-09-29 05:05:03 +00:00
Chris Lattner
890c21a20a add assembler support for the cvtsd2sil/cvtsd2siq mnemonics, rdar://8456382 
llvm-svn: 115027
 2010-09-29 04:55:40 +00:00
Chris Lattner
54939ddf1f make the x86 mccode emitter emit the 0x67 and 0x66 prefix bytes in the same 
order as cctools for diffability.

llvm-svn: 115022
 2010-09-29 03:43:43 +00:00