Commit Graph

104 Commits

Author SHA1 Message Date
Dan Gohman 580b80d6d9 Make ConstantFoldConstantExpression recursively visit the entire
ConstantExpr, not just the top-level operator. This allows it to
fold many more constants.

Also, make GlobalOpt call ConstantFoldConstantExpression on
GlobalVariable initializers.

llvm-svn: 89659
2009-11-23 16:22:21 +00:00
Chris Lattner cdfb80de16 fix ConstantFoldCompareInstOperands to take the LHS/RHS as
individual operands instead of taking a temporary array

llvm-svn: 86619
2009-11-09 23:06:58 +00:00
Chris Lattner 46b5c642b9 remove a bunch of extraneous LLVMContext arguments
from various APIs, addressing PR5325.

llvm-svn: 86231
2009-11-06 04:27:31 +00:00
Chris Lattner d8e8fb4d64 Teach FoldBitCast to be able to handle bitcasts from (e.g.) i128 -> <4 x float>.
This allows us to simplify this:
union vec2d {
    double e[2];
    double v __attribute__((vector_size(16)));
};
typedef union vec2d vec2d;

static vec2d a={{1,2}}, b={{3,4}};
    
vec2d foo () {
    return (vec2d){ .v = a.v + b.v * (vec2d){{5,5}}.v };
}

down to:

define %0 @foo() nounwind ssp {
entry:
  %mrv5 = insertvalue %0 undef, double 1.600000e+01, 0 ; <%0> [#uses=1]
  %mrv6 = insertvalue %0 %mrv5, double 2.200000e+01, 1 ; <%0> [#uses=1]
  ret %0 %mrv6
}

instead of:

define %0 @foo() nounwind ssp {
entry:
  %mrv5 = insertvalue %0 undef, double extractelement (<2 x double> fadd (<2 x double> fmul (<2 x double> bitcast (<1 x i128> <i128 85174437667405312423031577302488055808> to <2 x double>), <2 x double> <double 3.000000e+00, double 4.000000e+00>), <2 x double> <double 1.000000e+00, double 2.000000e+00>), i32 0), 0 ; <%0> [#uses=1]
  %mrv6 = insertvalue %0 %mrv5, double extractelement (<2 x double> fadd (<2 x double> fmul (<2 x double> bitcast (<1 x i128> <i128 85174437667405312423031577302488055808> to <2 x double>), <2 x double> <double 3.000000e+00, double 4.000000e+00>), <2 x double> <double 1.000000e+00, double 2.000000e+00>), i32 1), 1 ; <%0> [#uses=1]
  ret %0 %mrv6
}

llvm-svn: 85040
2009-10-25 06:15:37 +00:00
Chris Lattner 9d051246eb move FoldBitCast earlier in the file, and use it instead of
ConstantExpr::getBitCast in various places.

llvm-svn: 85039
2009-10-25 06:08:26 +00:00
Chris Lattner c5fd5ad4aa refactor FoldBitCast to reduce nesting and to always return a constantexpr
instead of returning null on failure.  No functionality change.

llvm-svn: 85038
2009-10-25 06:02:57 +00:00
Chris Lattner 3db7bd27c2 various cleanups suggested by Duncan
llvm-svn: 84993
2009-10-24 05:27:19 +00:00
Chris Lattner 9e2d5b3b8e fix PR5287, a serious regression from my previous patches. Thanks to
Duncan for the nice tiny testcase.

llvm-svn: 84992
2009-10-24 05:22:15 +00:00
Chris Lattner ccf1e84779 teach libanalysis to simplify vector loads with bitcast sources. This
implements something out of Target/README.txt producing:

_foo:                                                       ## @foo
	movl	4(%esp), %eax
	movapd	LCPI1_0, %xmm0
	movapd	%xmm0, (%eax)
	ret	$4

instead of:

_foo:                                                       ## @foo
	movl	4(%esp), %eax
	movapd	_b, %xmm0
	mulpd	LCPI1_0, %xmm0
	addpd	_a, %xmm0
	movapd	%xmm0, (%eax)
	ret	$4

llvm-svn: 84942
2009-10-23 06:57:37 +00:00
Chris Lattner 59f94c01dd enhance FoldReinterpretLoadFromConstPtr to handle loads of up to 32
bytes (i256).

llvm-svn: 84941
2009-10-23 06:50:36 +00:00
Chris Lattner ed00b80bf8 teach libanalysis to fold int and fp loads from almost arbitrary
non-type-safe constant initializers.  This sort of thing happens
quite a bit for 4-byte loads out of string constants, unions, 
bitfields, and an interesting endianness check from sqlite, which
is something like this:

const int sqlite3one = 1;
# define SQLITE_BIGENDIAN    (*(char *)(&sqlite3one)==0)
# define SQLITE_LITTLEENDIAN (*(char *)(&sqlite3one)==1)
# define SQLITE_UTF16NATIVE (SQLITE_BIGENDIAN?SQLITE_UTF16BE:SQLITE_UTF16LE)

all of these macros now constant fold away.

This implements PR3152 and is based on a patch started by Eli, but heavily
modified and extended.

llvm-svn: 84936
2009-10-23 06:23:49 +00:00
Chris Lattner cf7e8947e9 move another load optimization from instcombine -> libanalysis.
llvm-svn: 84841
2009-10-22 06:44:07 +00:00
Chris Lattner 51d2f70e32 move 'loading i32 from string' optimization from instcombine
to libanalysis.  Instcombine shrinking... does this even 
make sense???

llvm-svn: 84840
2009-10-22 06:38:35 +00:00
Chris Lattner 1664a4fd86 Move some constant folding logic for loads out of instcombine into
Analysis/ConstantFolding.cpp.  This doesn't change the behavior of
instcombine but makes other clients of ConstantFoldInstruction
able to handle loads.  This was partially extracted from Eli's patch
in PR3152.

llvm-svn: 84836
2009-10-22 06:25:11 +00:00
Evan Phoenix 44e5dbcaf0 Extend ConstantFolding to understand signed overflow variants
llvm-svn: 83338
2009-10-05 22:53:52 +00:00
Dan Gohman e525d9ddc0 Remove an unnnecessary LLVMContext argument in
ConstantFoldLoadThroughGEPConstantExpr.

llvm-svn: 83311
2009-10-05 16:36:26 +00:00
Chris Lattner 59d939894b teach the optimizer how to constant fold uadd/usub intrinsics.
llvm-svn: 83295
2009-10-05 05:26:04 +00:00
Chris Lattner 351534f9b1 simplify this code a bunch.
llvm-svn: 83294
2009-10-05 05:06:24 +00:00
Chris Lattner 9ca7c0913a code simplifications.
llvm-svn: 83292
2009-10-05 05:00:35 +00:00
Chris Lattner 3045f9e57b simplify some code
llvm-svn: 81961
2009-09-16 00:08:07 +00:00
Dan Gohman 21c6216c87 Teach lib/VMCore/ConstantFold.cpp how to set the inbounds keyword and
how to fold notionally-out-of-bounds array getelementptr indices instead
of just doing these in lib/Analysis/ConstantFolding.cpp, because it can
be done in a fairly general way without TargetData, and because not all
constants are visited by lib/Analysis/ConstantFolding.cpp. This enables
more constant folding.

Also, set the "inbounds" flag when the getelementptr indices are
one-past-the-end.

llvm-svn: 81483
2009-09-11 00:04:14 +00:00
Dan Gohman 91d598de5c Give these files top-level comments that describe the current code.
llvm-svn: 81473
2009-09-10 23:07:18 +00:00
Dan Gohman e4ca02da1b Revert 80959. It isn't sufficient to solve the full problem. And it
introduced regressions in the Ocaml bindings tests.

llvm-svn: 80969
2009-09-03 23:34:49 +00:00
Dan Gohman 2a53b30f6d Remove the API for creating ConstantExprs with the nsw, nuw, inbounds,
and exact flags. Because ConstantExprs are uniqued, creating an
expression with this flag causes all expressions with the same operands
to have the same flag, which may not be safe. Add, sub, mul, and sdiv
ConstantExprs are usually folded anyway, so the main interesting flag
here is inbounds, and the constant folder already knows how to set the
inbounds flag automatically in most cases, so there isn't an urgent need
for the API support.

This can be reconsidered in the future, but for now just removing these
API bits eliminates a source of potential trouble with little downside.

llvm-svn: 80959
2009-09-03 22:17:40 +00:00
Chris Lattner 950e60ea97 fix PR4848 an infinite loop when indexing down through a recursive gep
and we get the original pointer type.  This doesn't mean that we're
at the first pointer being indexed.  Correct the predicate.

llvm-svn: 80762
2009-09-02 05:35:45 +00:00
Dan Gohman c0b967dfb4 Don't assume that the operand of an inttoptr is an pointer-sized integer.
llvm-svn: 79651
2009-08-21 18:27:26 +00:00
Dan Gohman 23e62c532c Change getelementptr folding to use APInt instead of uint64_t for
offset computations. This fixes a truncation bug on targets that
don't have 64-bit pointers.

llvm-svn: 79639
2009-08-21 16:52:54 +00:00
Dan Gohman 8a8ad7d8e8 Various comment and whitespace cleanups.
llvm-svn: 79533
2009-08-20 16:42:55 +00:00
Dan Gohman c59ba42b0f Fix a bug in the over-index constant folding. When over-indexing an
array member of a struct, it's possible to land in an arbitrary position
inside that struct, such that attempting to find further getelementptr
indices will fail. In such cases, folding cannot be done.

llvm-svn: 79485
2009-08-19 22:46:59 +00:00
Dan Gohman 81ce842ecf Canonicalize indices in a constantexpr GEP. If Indices exceed the
static extents of the static array type, it causes GlobalOpt and
other passes to be more conservative. This canonicalization also
allows the constant folder to add "inbounds" to GEPs.

llvm-svn: 79440
2009-08-19 18:18:36 +00:00
Owen Anderson 55f1c09e31 Push LLVMContexts through the IntegerType APIs.
llvm-svn: 78948
2009-08-13 21:58:54 +00:00
Owen Anderson 5a1acd9912 Move a few more APIs back to 2.5 forms. The only remaining ones left to change back are
metadata related, which I'm waiting on to avoid conflicting with Devang.

llvm-svn: 77721
2009-07-31 20:28:14 +00:00
Owen Anderson b292b8ce70 Move more code back to 2.5 APIs.
llvm-svn: 77635
2009-07-30 23:03:37 +00:00
Owen Anderson 4056ca9568 Move types back to the 2.5 API.
llvm-svn: 77516
2009-07-29 22:17:13 +00:00
Owen Anderson 487375e9a2 Move ConstantExpr to 2.5 API.
llvm-svn: 77494
2009-07-29 18:55:55 +00:00
Owen Anderson 4aa3295a65 Return ConstantVector to 2.5 API.
llvm-svn: 77366
2009-07-28 21:19:26 +00:00
Owen Anderson 69c464dec4 Move ConstantFP construction back to the 2.5-ish API.
llvm-svn: 77247
2009-07-27 20:59:43 +00:00
Daniel Dunbar ca414c7cae Remove Value::getNameLen
llvm-svn: 77148
2009-07-26 08:34:35 +00:00
Owen Anderson edb4a70325 Revert the ConstantInt constructors back to their 2.5 forms where possible, thanks to contexts-on-types. More to come.
llvm-svn: 77011
2009-07-24 23:12:02 +00:00
Owen Anderson 47db941fd3 Get rid of the Pass+Context magic.
llvm-svn: 76702
2009-07-22 00:24:57 +00:00
Torok Edwin fbcc663cbf llvm_unreachable->llvm_unreachable(0), LLVM_UNREACHABLE->llvm_unreachable.
This adds location info for all llvm_unreachable calls (which is a macro now) in
!NDEBUG builds.
In NDEBUG builds location info and the message is off (it only prints
"UREACHABLE executed").

llvm-svn: 75640
2009-07-14 16:55:14 +00:00
Torok Edwin 56d0659726 assert(0) -> LLVM_UNREACHABLE.
Make llvm_unreachable take an optional string, thus moving the cerr<< out of
line.
LLVM_UNREACHABLE is now a simple wrapper that makes the message go away for
NDEBUG builds.

llvm-svn: 75379
2009-07-11 20:10:48 +00:00
Nick Lewycky a21d3daadc Remove the vicmp and vfcmp instructions. Because we never had a release with
these instructions, no autoupgrade or backwards compatibility support is
provided.

llvm-svn: 74991
2009-07-08 03:04:38 +00:00
Owen Anderson 38264b1554 "LLVMContext* " --> "LLVMContext *"
llvm-svn: 74878
2009-07-06 23:00:19 +00:00
Owen Anderson 39f00cc1d4 Thread LLVMContext through the constant folding APIs, which touches a lot of files.
llvm-svn: 74844
2009-07-06 18:42:36 +00:00
Dan Gohman 7ccc52f131 Support vector casts in more places, fixing a variety of assertion
failures.

To support this, add some utility functions to Type to help support
vector/scalar-independent code. Change ConstantInt::get and
ConstantFP::get to support vector types, and add an overload to
ConstantInt::get that uses a static IntegerType type, for
convenience.

Introduce a new getConstant method for ScalarEvolution, to simplify
common use cases.

llvm-svn: 73431
2009-06-15 22:12:54 +00:00
Dan Gohman 136bd719f0 Change ConstantFoldConstantExpression to accept a null
TargetData pointer. The only thing it's used for are
calls to ConstantFoldCompareInstOperands and
ConstantFoldInstOperands, which both already accept a
null TargetData pointer. This makes
ConstantFoldConstantExpression easier to use in clients
where TargetData is optional.

llvm-svn: 72741
2009-06-02 21:48:15 +00:00
Jay Foad 7d0479f2c2 Use v.data() instead of &v[0] when SmallVector v might be empty.
llvm-svn: 72210
2009-05-21 09:52:38 +00:00
Duncan Sands af9eaa830a Rename PaddedSize to AllocSize, in the hope that this
will make it more obvious what it represents, and stop
it being confused with the StoreSize.

llvm-svn: 71349
2009-05-09 07:06:46 +00:00
Dan Gohman 4bb6fa23cb Revert 71165. It did more than just revert 71158 and it introduced
several regressions. The problem due to 71158 is now fixed.

llvm-svn: 71176
2009-05-07 19:46:24 +00:00