Commit Graph

62099 Commits

Author SHA1 Message Date
Chris Lattner
11b3d1621d another v2f32 case, in this case showing poor codegen.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107614 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-05 05:52:56 +00:00
Chris Lattner
c06cbad141 some notes about suboptimal insertps's
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107613 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-05 05:48:41 +00:00
Chris Lattner
598751ed25 random tidying
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107612 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-05 05:36:21 +00:00
Chris Lattner
a5b412581c rip out even more sporadic v2f32 support.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107610 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-05 04:38:33 +00:00
Chris Lattner
39aa20a981 rip out the various v2f32 "mmx" handling logic, now that
v2f32 is illegal on x86.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107609 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-05 04:36:27 +00:00
Chris Lattner
4fd1ab3bec fix test on non-x86 hosts.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107608 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-05 03:56:55 +00:00
Jakob Stoklund Olesen
b1e11455d1 Print symbolic subreg indices on REG_SEQUENCE and INSERT_SUBREG.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107602 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-04 23:24:23 +00:00
Chris Lattner
f172ecd964 Just rip v2f32 support completely out of the X86 backend. In
the example in the testcase, we now generate:

_test1:                                 ## @test1
	movss	4(%esp), %xmm0
	addss	8(%esp), %xmm0
	movl	12(%esp), %eax
	movss	%xmm0, (%eax)
	ret

instead of:

_test1:                                                     ## @test1
	subl	$20, %esp
	movl	24(%esp), %eax
	movq	%mm0, (%esp)
	movq	%mm0, 8(%esp)
	movss	(%esp), %xmm0
	addss	12(%esp), %xmm0
	movss	%xmm0, (%eax)
	addl	$20, %esp
	ret

v2f32 support did not work reliably because most of the X86
backend didn't know it was legal.  It was apparently only added
to support returning source-level v2f32 values in MMX registers
in x86-32 mode.  If ABI compatibility is important on this
GCC-extended-vector type for some reason, then the frontend
should generate IR that returns v2i32 instead of v2f32.  However,
we generally don't try very hard to be abi compatible on gcc
extended vectors. 



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107601 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-04 23:07:25 +00:00
Chris Lattner
e35d9842f7 fix PR7518 - terrible codegen of <2 x float>, by only marking
v2f32 as legal in 32-bit mode.  It is just as terrible there,
but I just care about x86-64 and noone claims it is valuable
in 64-bit mode.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107600 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-04 22:57:10 +00:00
Chris Lattner
9d19989fe3 indentation
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107599 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-04 22:56:10 +00:00
Evan Cheng
f2f4903687 Infer alignments of fixed frame objects when they are constructed. This ensures remat'ed loads from fixed slots have the right alignments.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107591 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-04 18:52:05 +00:00
Chris Lattner
aa6c855c8e indmemrem and lowerallocs are gone.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107587 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-04 17:36:43 +00:00
Bill Wendling
f43f6bc3ec Revert r107583. I no longer think that this is the way to solve the problem.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107585 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-04 09:16:57 +00:00
Bill Wendling
4a991a6fa6 Mark sse_load_f32 and sse_load_f64 as having memory operands
(SDNPMemOperand). This way when they're morphed the memory operands will be
copied as well.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107583 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-04 08:59:55 +00:00
Gabor Greif
69439cb919 back out r107580, I have seen what I wanted
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107582 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-04 08:59:25 +00:00
Bill Wendling
d9cb7ca388 Proper indentation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107581 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-04 08:58:43 +00:00
Gabor Greif
4de8ba8940 [another test commit, just to tickle the selfhost buildbots; I'll back out in a few minutes]
second round of low-level interface squeeze-out:
making all of CallInst's low-level operand accessors
private

If you get compile errors I strongly urge you to
update your code.


I tried to write the necessary clues into the
header where the compiler may point to, but no
guarantees. It works for my GCC.

You have several options to update your code:

- you can use the v2.8 ArgOperand accessors
- you can go via a temporary CallSite
- you can upcast to, say, User and call its
  low-level accessors if your code is definitely
  operand-order agnostic.

If you run into serious problems, please
comment in below thread (and back out this
revision only if absolutely necessary):

<http://groups.google.com/group/llvm-dev/browse_thread/thread/64650cf343b28271>


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107580 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-04 07:42:31 +00:00
Eli Friedman
b482829abb Minor amendment to switch-lowering improvement.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107569 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-03 08:43:32 +00:00
Eli Friedman
b4a74c1d82 Note switch-lowering inefficiency.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107565 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-03 07:38:12 +00:00
Bruno Cardoso Lopes
68b559e5f3 Add AVX SSE4.1 blend, mpsadbw and vdp
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107560 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-03 01:37:03 +00:00
Bruno Cardoso Lopes
4a544be3a8 Add AVX SSE4.1 binop (some forms of packed max,min,mul,pack,cmp) instructions
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107558 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-03 01:15:47 +00:00
Eric Christopher
00e840f469 Fix typo.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107556 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-03 01:09:18 +00:00
Bruno Cardoso Lopes
c607570563 Add AVX SSE4.1 Horizontal Minimum and Position instruction
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107552 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-03 00:49:21 +00:00
Evan Cheng
ed2ae136d2 Remove isSS argument from CreateFixedObject. Fixed objects cannot be spill slots so it's always false.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107550 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-03 00:40:23 +00:00
Bruno Cardoso Lopes
2c70d4ad35 Add AVX SSE4.1 round instructions
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107549 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-03 00:37:44 +00:00
Jakob Stoklund Olesen
273f7e4299 Detect and handle COPY in many places.
This code is transitional, it will soon be possible to eliminate
isExtractSubreg, isInsertSubreg, and isMoveInstr in most places.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107547 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-03 00:04:37 +00:00
Dan Gohman
43b8fd728b Document that BasicAA respects noalias, while InterproceduralBasicAA
does not.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107546 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-02 23:52:36 +00:00
Dan Gohman
efca7f9681 Describe which dependencies are relevant to noalias, and document how
AliasAnalyses have the choice to respect noalias.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107544 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-02 23:46:54 +00:00
Bruno Cardoso Lopes
03560600b4 Simple refactoring of SSE4.1 instructions, making room for the AVX forms
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107540 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-02 23:27:59 +00:00
Dan Gohman
ff23535df6 Add some html anchors, to allow attributes to be linked to directly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107538 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-02 23:18:08 +00:00
Eric Christopher
cd075a4fb3 80-col fixup.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107537 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-02 23:17:38 +00:00
Jakob Stoklund Olesen
a4e1ba53dd Add a new target independent COPY instruction and code to lower it.
The COPY instruction is intended to replace the target specific copy
instructions for virtual registers as well as the EXTRACT_SUBREG and
INSERT_SUBREG instructions in MachineFunctions. It won't we used in a selection
DAG.

COPY is lowered to native register copies by LowerSubregs.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107529 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-02 22:29:50 +00:00
Bruno Cardoso Lopes
f5cd8c51e3 - Add support for the rest of AVX SSE3 instructions
- Fix VEX prefix to be emitted with 3 bytes whenever VEX_5M
represents a REX equivalent two byte leading opcode



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107523 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-02 22:06:54 +00:00
Jakob Stoklund Olesen
65766ce7df Clean up TargetOpcodes.h a bit, and limit the number of places where the full
list of predefined instructions appear. Add some consistency checks.

Ideally, TargetOpcodes.h should be produced by TableGen from Target.td, but it
is hardly worth the effort.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107520 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-02 21:44:22 +00:00
Jim Grosbach
6627ac040a Custom inserters (e.g., conditional moves in Thumb1 can introduce
new basic blocks, and if used as a function argument, that can cause call frame
setup / destroy pairs to be split across a basic block boundary. That prevents
us from doing a simple assertion to check that the pairs match and alloc/
dealloc the same amount of space. Modify the assertion to only check the
amount allocated when there are matching pairs in the same basic block.

rdar://8022442

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107517 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-02 21:23:37 +00:00
Devang Patel
1192d4d0a5 MDString is already checked earlier.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107516 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-02 21:13:23 +00:00
Evan Cheng
dca653951c Remove early IT block formation. It's not used.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107513 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-02 21:07:09 +00:00
Evan Cheng
98ec91ea80 - Two-address pass should not assume unfolding is always successful.
- X86 unfolding should check if the instructions being unfolded has memoperands.
  If there is no memoperands, then it must assume conservative alignment. If this
  would introduce an expensive sse unaligned load / store, then unfoldMemoryOperand
  etc. should not unfold the instruction.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107509 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-02 20:36:18 +00:00
Dan Gohman
541481f34c Remove an unused enum.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107507 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-02 20:20:50 +00:00
Dale Johannesen
f1e309eb48 Propagate the AlignStack bit in InlineAsm's to the
PrologEpilog code, and use it to determine whether
the asm forces stack alignment or not.  gcc consistently
does not do this for GCC-style asms; Apple gcc inconsistently
sometimes does it for asm blocks.  There is no
convenient place to put a bit in either the SDNode or
the MachineInstr form, so I've added an extra operand
to each; unlovely, but it does allow for expansion for
more bits, should we need it.  PR 5125.  Some
existing testcases are affected.
The operand lists of the SDNode and MachineInstr forms
are indexed with awesome mnemonics, like "2"; I may
fix this someday, but not now.  I'm not making it any
worse.  If anyone is inspired I think you can find all
the right places from this patch.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107506 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-02 20:16:09 +00:00
Jakob Stoklund Olesen
6bae55017c Remove invalid assert
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107505 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-02 19:54:47 +00:00
Jakob Stoklund Olesen
ec8c3e8ccc Handle unindexed instructions in SlotIndices.
SlotIndexes::insertMachineInstrInMaps would crash when trying to insert an
instruction imediately after an unmapped debug value.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107504 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-02 19:54:45 +00:00
Jakob Stoklund Olesen
3b9c7ebc44 Properly handle debug values during inline spilling.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107503 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-02 19:54:40 +00:00
Gabor Greif
135d7fe9bb beautify output
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107500 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-02 19:26:28 +00:00
Gabor Greif
53ba550df1 use ArgOperand API
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107498 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-02 19:08:46 +00:00
Dan Gohman
8c1ffae068 Don't claim to preserve AliasAnalysis. First, this is doesn't actually
have any effect, and second, deleting stores can potentially invalidate
an AliasAnalysis, and there's currently no notification for this.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107496 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-02 18:43:05 +00:00
Dan Gohman
1e109621f1 Reword the definition of the noalias attribute. The intention is for
the noalias argument on function attributes be usable to model the
C99 restrict keyword on arguments, and to allow AliasAnalysis to
consider a noalias-attributed argument to be an "identified object".

To support this, refactor a new "based on" concept out of the current
pointer aliasing "associated" concept. This "based on" concept is very
similar to (though it is not identical with) the "based on" concept
in C99.

Also, reword the definition of NoAlias to more closely describe the
concept that the optimizer uses.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107495 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-02 18:41:32 +00:00
Jakob Stoklund Olesen
8de3b1eb86 Rematerialize as much as possible before inserting spills and reloads.
This allows us to recognize the common case where all uses could be
rematerialized, and no stack slot allocation is necessary.

If some values could be fully rematerialized, remove them from the live range
before allocating a stack slot for the rest.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107492 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-02 17:44:57 +00:00
Jim Grosbach
6e9926108a 80-column and trailing whitespace cleanup.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107490 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-02 17:41:59 +00:00
Jim Grosbach
ee7f8b5f05 grammar tweaks
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107489 91177308-0d34-0410-b5e6-96231b3b80d8
2010-07-02 17:38:34 +00:00