Commit Graph

34145 Commits

Author SHA1 Message Date
Chris Lattner
3f40e23392 Implement PR5634.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@90046 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-29 00:51:17 +00:00
Nick Lewycky
d801c10de6 Teach memdep to look for memory use intrinsics during dependency queries. Fixes
PR5574.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@90045 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-28 21:27:49 +00:00
Chris Lattner
0c264b16b0 reenable load address insertion in load pre. This allows us to
handle cases like this:
void test(int N, double* G) {
  long j;
  for (j = 1; j < N - 1; j++)
      G[j+1] = G[j] + G[j+1];
}

where G[1] isn't live into the loop.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@90041 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-28 16:08:18 +00:00
Chris Lattner
dd696052f0 Enhance InsertPHITranslatedPointer to be able to return a list of newly
inserted instructions.  No functionality change until someone starts using it.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@90039 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-28 15:39:14 +00:00
Chris Lattner
ab9530ee5d implement a FIXME: limit the depth that DecomposeGEPExpression goes the same
way that getUnderlyingObject does it. 

This fixes the 'DecomposeGEPExpression and getUnderlyingObject disagree!' 
assertion on sqlite3.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@90038 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-28 15:12:41 +00:00
Chris Lattner
af7a28939e enable code to handle un-phi-translatable cases more aggressively:
if we don't have an address expression available in a predecessor,
then model this as the value being clobbered at the end of the pred
block instead of being modeled as a complete phi translation failure.
This is important for PRE of loads because we want to see that the
load is available in all but this predecessor, and complete phi
translation failure results in not getting any information about
predecessors.

This doesn't do anything until I renable code insertion since PRE
now sees that it is available in all but one predecessors, but can't
insert the addressing in the predecessor that is missing it to 
eliminate the redundancy.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@90037 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-28 14:54:10 +00:00
Chris Lattner
971fd57f7d disable value insertion for now, I need to figure out how
to inform GVN about the newly inserted values.  This fixes 
PR5631.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@90022 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-27 22:50:07 +00:00
Chris Lattner
6f7b210b25 Rework InsertPHITranslatedPointer to handle the recursive case, this
fixes PR5630 and sets the stage for the next phase of goodness (testcase
pending).


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@90019 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-27 22:05:15 +00:00
Chris Lattner
5141421e38 recursively phi translate bitcast operands too, for consistency.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@90016 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-27 20:25:30 +00:00
Nick Lewycky
807d2fa1eb Oops! Fix bug introduced in my recent cleanup change. Thanks to Tobias Grosser
for pointing this out.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@90015 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-27 19:57:53 +00:00
Chris Lattner
11c6bab704 add support for recursive phi translation and phi
translation of add with immediate.  This allows us
to optimize this function:

void test(int N, double* G) {
  long j;
  G[1] = 1;
    for (j = 1; j < N - 1; j++)
        G[j+1] = G[j] + G[j+1];
}

to only do one load every iteration of the loop.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@90013 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-27 19:11:31 +00:00
Chris Lattner
8aee8efc0c factor some logic out of instcombine into a new SimplifyAddInst method.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@90011 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-27 17:42:22 +00:00
Chris Lattner
1d159838a3 add a deadargelim note.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@90009 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-27 17:12:30 +00:00
Chris Lattner
6d94926a52 This testcase is actually only partially redundant, and requires
the FIXME I added yesterday to be implemented.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@90008 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-27 16:53:57 +00:00
Chris Lattner
8976e5950f fix PR5436 by making the 'simple' case of SRoA not promote out of range
array indexes.  The "complex" case of SRoA still handles them, and correctly.

This fixes a weirdness where we'd correctly avoid transforming A[0][42] if
the 42 was too large, but we'd only do it if it was one gep, not two separate
ones.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@90007 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-27 16:37:41 +00:00
Chris Lattner
9763487bd5 add comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@90002 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-27 08:40:14 +00:00
Chris Lattner
e95035aea4 reduce nesting, no functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@90001 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-27 08:37:22 +00:00
Chris Lattner
a650f770d4 limit the recursion depth of GetLinearExpression. This
fixes a crash analyzing consumer-lame, which had an "%X = add %X, 1"
in unreachable code.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@90000 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-27 08:32:52 +00:00
Chris Lattner
616613d7a4 teach GVN's load PRE to insert computations of the address in predecessors
where it is not available.  It's unclear how to get this inserted 
computation into GVN's scalar availability sets, Owen, help? :)


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89997 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-27 08:25:10 +00:00
Chris Lattner
62deff066c Fix phi translation in load PRE to agree with the phi
translation done by memdep, and reenable gep translation 
again.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89992 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-27 06:31:14 +00:00
Chris Lattner
b99be5beac redisable this, my bootstrap worked because it wasn't an optimized build, whoops.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89991 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-27 05:53:01 +00:00
Chris Lattner
cca130bb66 try again.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89990 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-27 05:19:56 +00:00
Chris Lattner
518c988ae9 this is causing buildbot failures, disable for now.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89985 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-27 01:52:22 +00:00
Chris Lattner
20ce7c0ded this (and probably several others) are now done.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89982 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-27 00:35:04 +00:00
Chris Lattner
e19e4baf3b teach phi translation of GEPs to simplify geps like 'gep x, 0'.
This allows us to compile the example from PR5313 into:

LBB1_2:                                                     ## %bb
	incl	%ecx
	movb	%al, (%rsi)
	movslq	%ecx, %rax
	movb	(%rdi,%rax), %al
	testb	%al, %al
	jne	LBB1_2

instead of:

LBB1_2:                                                     ## %bb
	movslq	%eax, %rcx
	incl	%eax
	movb	(%rdi,%rcx), %cl
	movb	%cl, (%rsi)
	movslq	%eax, %rcx
	cmpb	$0, (%rdi,%rcx)
	jne	LBB1_2



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89981 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-27 00:34:38 +00:00
Chris Lattner
c514c1f521 factor some instcombine simplifications for getelementptr out to a new
SimplifyGEPInst method in InstructionSimplify.h.  No functionality change.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89980 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-27 00:29:05 +00:00
Chris Lattner
304076268a teach memdep to do trivial PHI translation of GEPs. More to
come.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89979 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-27 00:07:37 +00:00
Chris Lattner
cc3d0eb483 Teach memdep to phi translate bitcasts. This allows us to compile
the example in GCC PR16799 to:

LBB1_2:                                                     ## %bb1
	movl	%eax, %eax
	subq	%rax, %rdi
	movq	%rdi, (%rcx)
	movl	(%rdi), %eax
	testl	%eax, %eax
	je	LBB1_2

instead of:

LBB1_2:                                                     ## %bb1
	movl	(%rdi), %ecx
	subq	%rcx, %rdi
	movq	%rdi, (%rax)
	cmpl	$0, (%rdi)
	je	LBB1_2



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89978 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-26 23:41:07 +00:00
Nick Lewycky
dbee4b3be5 Fix typo spotted by Gabor Greif.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89976 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-26 23:19:05 +00:00
Chris Lattner
dc59311c5c factor some code out into some helper functions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89975 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-26 23:18:49 +00:00
Nick Lewycky
2353122a3c Clean up file, no functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89974 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-26 22:54:26 +00:00
Chris Lattner
5bda3792b4 Add a hack for PR5601, a crash on obsolete syntax that we plan to
remove in LLVM 3.0


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89973 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-26 22:48:23 +00:00
Chris Lattner
998e25a143 fix crash on Transforms/InstCombine/intrinsics.ll introduced by r89970
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89972 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-26 22:08:06 +00:00
Chris Lattner
21f83a2f95 Fix PR5471 by removing an instcombine xform. Some pieces of the code
generates store to undef and some generates store to null as the idiom
for undefined behavior.  Since simplifycfg zaps both, don't remove the
undefined behavior in instcombine.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89971 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-26 22:04:42 +00:00
Chris Lattner
2bbac75e5a implement a bunch of xforms for overflow intrinsics, based on a patch
by Alastair Lynn.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89970 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-26 21:42:47 +00:00
Bob Wilson
2d521e51ac Rename new TailDuplicationPass to avoid name conflict with the old one.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89968 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-26 21:38:41 +00:00
Chris Lattner
1ce0eaa25f Implement PR1143 (at -m64) by making basicaa look through extensions. We
previously already handled it at -m32 because there were no i32->i64 
extensions for addressing.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89959 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-26 18:53:33 +00:00
Chris Lattner
b18004c13c fix two transposed lines duncan caught and add an explanatory comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89958 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-26 18:35:46 +00:00
Chris Lattner
8fe8ac725b this todo is resolved.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89957 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-26 17:14:10 +00:00
Chris Lattner
e405c64f6b move DecomposeGEPExpression out into ValueTracking.cpp
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89956 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-26 17:12:50 +00:00
Chris Lattner
fa3966881f teach GetLinearExpression to be a bit more aggressive.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89955 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-26 17:00:01 +00:00
Chris Lattner
5369250b73 resolve a fixme. I haven't figured out how to write a testcase
to exercise this though.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89954 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-26 16:52:32 +00:00
Chris Lattner
5d5261c819 Teach basicaa that x|c == x+c when the c bits of x are clear. This
allows us to compile the example in readme.txt into:

LBB1_1:                                                     ## %bb
	movl	4(%rdx,%rax), %ecx
	movl	%ecx, %esi
	imull	(%rdx,%rax), %esi
	imull	%esi, %ecx
	movl	%esi, 8(%rdx,%rax)
	imull	%ecx, %esi
	movl	%ecx, 12(%rdx,%rax)
	movl	%esi, 16(%rdx,%rax)
	imull	%ecx, %esi
	movl	%esi, 20(%rdx,%rax)
	addq	$16, %rax
	cmpq	$4000, %rax
	jne	LBB1_1

instead of:

LBB1_1: 
	movl	(%rdx,%rax), %ecx
	imull	4(%rdx,%rax), %ecx
	movl	%ecx, 8(%rdx,%rax)
	imull	4(%rdx,%rax), %ecx
	movl	%ecx, 12(%rdx,%rax)
	imull	8(%rdx,%rax), %ecx
	movl	%ecx, 16(%rdx,%rax)
	imull	12(%rdx,%rax), %ecx
	movl	%ecx, 20(%rdx,%rax)
	addq	$16, %rax
	cmpq	$4000, %rax
	jne	LBB1_1

GCC (4.2) doesn't seem to be able to eliminate the loads in this 
testcase either, it generates:

L2:
	movl	(%rdx), %eax
	imull	4(%rdx), %eax
	movl	%eax, 8(%rdx)
	imull	4(%rdx), %eax
	movl	%eax, 12(%rdx)
	imull	8(%rdx), %eax
	movl	%eax, 16(%rdx)
	imull	12(%rdx), %eax
	movl	%eax, 20(%rdx)
	addl	$4, %ecx
	addq	$16, %rdx
	cmpl	$1002, %ecx
	jne	L2




git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89952 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-26 16:26:43 +00:00
Chris Lattner
f6ac4d9dad teach basicaa that A[i] != A[i+1].
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89951 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-26 16:18:10 +00:00
Chris Lattner
d84eb911a9 Change the other half of aliasGEP (which handles GEP differencing) to use DecomposeGEPExpression. This dramatically simplifies and shrinks the code by eliminating the horrible CheckGEPInstructions method, fixes a miscompilation (@test3) and makes the code more aggressive. In particular, we now handle the @test4 case, which is reduced from the SmallPtrSet constructor. Missing this caused us to emit a variable length memset instead of a fixed size one.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89922 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-26 02:17:34 +00:00
Chris Lattner
23e2a5b2ff Generalize DecomposeGEPExpression to exactly handle what Value::getUnderlyingObject does (when TD is around). This allows us to avoid calling DecomposeGEPExpression unless the ultimate alias check we care about passes, speedup up BasicAA a bit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89920 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-26 02:14:59 +00:00
Chris Lattner
4e91ee7a2a Implement a new DecomposeGEPExpression method, which decomposes a GEP into a list of scaled offsets. Use this to eliminate some previous ad-hoc code which was subtly broken (it assumed all Constant*'s were non-zero, but strange constant express could be zero).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89915 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-26 02:13:03 +00:00
Chris Lattner
539c9b99c2 Use GEPOperator more pervasively to simplify code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89914 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-26 02:11:08 +00:00
Chris Lattner
9c6a0dcf0e update some notes slightly
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89913 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-26 01:51:18 +00:00
Chris Lattner
11fe72661d remove some redundant braces
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89912 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-26 01:50:12 +00:00