Commit Graph

288 Commits

Author SHA1 Message Date
Chris Lattner
61c70e98ac remove unions from LLVM IR. They are severely buggy and not
being actively maintained, improved, or extended.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112356 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-28 04:09:24 +00:00
Dan Gohman
0dd3549edc Constant fold x == undef to undef.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107074 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-28 21:30:07 +00:00
Dan Gohman
403a8cdda5 Use A.append(...) instead of A.insert(A.end(), ...) when A is a
SmallVector, and other SmallVector simplifications.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106452 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-21 19:47:52 +00:00
Dan Gohman
fdbef823ad getFoldedOffsetOf no longer does anything special with vector types.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@105514 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-05 00:47:34 +00:00
Nick Lewycky
5bf7f88ea2 Make the 'icmp pred trunc(ext(X)), CST --> icmp pred X, ext(trunc(CST))'
transformation much more careful. Truncating binary '01' to '1' sounds like it's
safe until you realize that it switched from positive to negative under a signed
interpretation, and that depends on the icmp predicate.

Also a few miscellaneous cleanups.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@97721 91177308-0d34-0410-b5e6-96231b3b80d8
2010-03-04 06:54:10 +00:00
Chris Lattner
c8e14b3d37 fix incorrect folding of icmp with undef, PR6481.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@97659 91177308-0d34-0410-b5e6-96231b3b80d8
2010-03-03 19:46:03 +00:00
Dan Gohman
9ee0d8a917 Teach the constant folder about union types.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@97142 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-25 16:45:19 +00:00
Dan Gohman
2440cf1c6f Remove code which assumes it knows how vectors are stored in memory.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@97141 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-25 16:05:33 +00:00
Dan Gohman
0e488b3d1c Canonicalize ConstantInts to the right operand of commutative
operators.

The test difference is just due to the multiplication operands
being commuted (and thus requiring a more elaborate match). In
optimized code, that expression would be folded.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96816 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-22 22:43:23 +00:00
Dan Gohman
d97439d1e5 Use Instruction::isCommutative instead of duplicating it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96807 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-22 22:05:18 +00:00
Duncan Sands
47c5188789 Introduce isOpaqueTy and use it rather than isa<OpaqueType>. Also, move some
methods to try to have the type predicates be more logically positioned.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96349 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-16 14:50:09 +00:00
Duncan Sands
1df9859c40 There are two ways of checking for a given type, for example isa<PointerType>(T)
and T->isPointerTy().  Convert most instances of the first form to the second form.
Requested by Chris.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96344 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-16 11:11:14 +00:00
Duncan Sands
b0bc6c361d Uniformize the names of type predicates: rather than having isFloatTy and
isInteger, we now have isFloatTy and isIntegerTy.  Requested by Chris!


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96223 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-15 16:12:20 +00:00
Chris Lattner
fdfeb6976f Add support for a union type in LLVM IR. Patch by Talin!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96011 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-12 20:49:41 +00:00
Dan Gohman
a84ffedafc Canonicalize sizeof and alignof on pointer types to a canonical
pointer type.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@95769 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-10 06:13:07 +00:00
Dan Gohman
6acb86dcfa Factor out alignof expression folding into a separate function and
generalize it to handle more cases.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@95045 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-02 01:41:39 +00:00
Chris Lattner
b29d596072 eliminate a bunch of pointless LLVMContext arguments.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@95001 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-01 20:48:08 +00:00
Duncan Sands
a419b56d89 Fix typo "of" -> "or" and change the way a line was formatted to fit
into 80 columns to match my artistic preferences.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@95000 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-01 20:42:02 +00:00
Chris Lattner
6304b0dd63 fix PR6195, a bug constant folding scalar -> vector compares.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@94997 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-01 20:04:40 +00:00
Chris Lattner
b63127d435 fix PR6197 - infinite recursion in ipsccp due to block addresses
evaluateICmpRelation wasn't handling blockaddress.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@94993 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-01 19:35:08 +00:00
Dan Gohman
4f8eea82d8 Generalize target-independent folding rules for sizeof to handle more
cases, and implement target-independent folding rules for alignof and
offsetof. Also, reassociate reassociative operators when it leads to
more folding.

Generalize ScalarEvolution's isOffsetOf to recognize offsetof on
arrays. Rename getAllocSizeExpr to getSizeOfExpr, and getFieldOffsetExpr
to getOffsetOfExpr, for consistency with analagous ConstantExpr routines.

Make the target-dependent folder promote GEP array indices to
pointer-sized integers, to make implicit casting explicit and exposed
to subsequent folding.

And add a bunch of testcases for this new functionality, and a bunch
of related existing functionality.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@94987 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-01 18:27:38 +00:00
Dan Gohman
8ca83b4111 Remove the folding rule
getelementptr (i8* inttoptr (i64 1 to i8*), i32 -1) 
  to
  inttoptr (i64 0 to i8*)
from the VMCore constant folder. It didn't handle sign-extension properly
in the case where the source integer is smaller than a pointer size. And,
it relied on an assumption about sizeof(i8).

The Analysis constant folder still folds these kinds of things; it has
access to TargetData, so it can do them right.

Add a testcase which tests that the VMCore constant folder doesn't
miscompile this, and that the Analysis folder does fold it.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@94750 91177308-0d34-0410-b5e6-96231b3b80d8
2010-01-28 18:08:26 +00:00
Dan Gohman
0f5efe5625 Remove SCEVAllocSizeExpr and SCEVFieldOffsetExpr, and in their place
use plain SCEVUnknowns with ConstantExpr::getSizeOf and
ConstantExpr::getOffsetOf constants. This eliminates a bunch of
special-case code.

Also add code for pattern-matching these expressions, for clients that
want to recognize them.

Move ScalarEvolution's logic for expanding array and vector sizeof
expressions into an element count times the element size, to expose
the multiplication to subsequent folding, into the regular constant
folder.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@94737 91177308-0d34-0410-b5e6-96231b3b80d8
2010-01-28 02:15:55 +00:00
Nick Lewycky
401f3258f6 Fix a crasher trying to fold each element in a comparison between two vectors
if one of the vectors didn't have elements (such as undef). Fixes PR 6096.

Fix an issue in the constant folder where fcmp (<2 x %ty>, <2 x %ty>) would
have <2 x i1> type if constant folding was successful and i1 type if it wasn't.
This exposed a related issue in the bitcode reader.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@94069 91177308-0d34-0410-b5e6-96231b3b80d8
2010-01-21 07:03:21 +00:00
Benjamin Kramer
8c65f6e71c Move remaining stuff to the isInteger predicate.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@92771 91177308-0d34-0410-b5e6-96231b3b80d8
2010-01-05 21:05:54 +00:00
Benjamin Kramer
11acaa374c Convert a ton of simple integer type equality tests to the new predicate.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@92760 91177308-0d34-0410-b5e6-96231b3b80d8
2010-01-05 20:07:06 +00:00
Eli Friedman
e8e17833be Aggressively flip compare constant expressions where appropriate; constant
folding in particular expects null to be on the RHS.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@91587 91177308-0d34-0410-b5e6-96231b3b80d8
2009-12-17 06:07:04 +00:00
Chris Lattner
0eeb913aa1 Previously, all operands to Constant were themselves constant.
In the new world order, BlockAddress can have a BasicBlock operand.
This doesn't permute much, because if you have a ConstantExpr (or
anything more specific than Constant) we still know the operand has
to be a Constant.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@85375 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-28 05:14:34 +00:00
Chris Lattner
59c4eba416 Teach vm core to more aggressively fold 'trunc' constantexprs,
allowing it to simplify the crazy constantexprs in the testcases
down to something sensible.  This allows -std-compile-opts to
completely "devirtualize" the pointers to member functions in
the testcase from PR5176.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84368 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-17 21:53:27 +00:00
Chris Lattner
cf0fe8d813 strength reduce a ton of type equality tests to check the typeid (Through
the new predicates I added) instead of going through a context and doing a
pointer comparison.  Besides being cheaper, this allows a smart compiler
to turn the if sequence into a switch.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83297 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-05 05:54:46 +00:00
Nick Lewycky
149cbc2a24 Peer through zext and sext to eliminate them when it is safe to do so.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@82389 91177308-0d34-0410-b5e6-96231b3b80d8
2009-09-20 07:31:25 +00:00
Nick Lewycky
85958b059d Fold 'icmp eq (icmp), true' into an xor(icmp).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@82386 91177308-0d34-0410-b5e6-96231b3b80d8
2009-09-20 07:21:39 +00:00
Nick Lewycky
b13efdafec Correct the comment; this applies to fcmp too.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@82380 91177308-0d34-0410-b5e6-96231b3b80d8
2009-09-20 06:27:35 +00:00
Nick Lewycky
8577e27dd3 Remove tab, again.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@82379 91177308-0d34-0410-b5e6-96231b3b80d8
2009-09-20 06:26:34 +00:00
Nick Lewycky
3105ebfaf7 Teach the constant folder how to not a cmpinst.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@82378 91177308-0d34-0410-b5e6-96231b3b80d8
2009-09-20 06:24:51 +00:00
Nick Lewycky
55a722bf6a Try turning icmp(bitcast(x), bitcast(y)) into icmp(bitcast(bitcast(x)), y) in
the hopes that the two bitcasts will merge.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@82371 91177308-0d34-0410-b5e6-96231b3b80d8
2009-09-20 05:48:50 +00:00
Nick Lewycky
8de003663d Remove tabs I added.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@82369 91177308-0d34-0410-b5e6-96231b3b80d8
2009-09-20 05:47:45 +00:00
Nick Lewycky
3892baac42 Clean up the usage of evaluateICmpRelation's return value.
Add another line to the ConstantExprFold test to demonstrate the GEPs may not
wrap around in either the signed or unsigned senses.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@82361 91177308-0d34-0410-b5e6-96231b3b80d8
2009-09-20 04:27:06 +00:00
Nick Lewycky
d43737bd09 Remove dead store by taking a guess at what Chris meant. I wasn't able to
design a testcase that would tickle this behaviour.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@82357 91177308-0d34-0410-b5e6-96231b3b80d8
2009-09-20 03:48:46 +00:00
Nick Lewycky
a80e9598ca Delete dead code. sext and zext can not turn integers into pointers. Further,
the optimization described in the comment is only valid with target data.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@82353 91177308-0d34-0410-b5e6-96231b3b80d8
2009-09-20 02:11:47 +00:00
Nick Lewycky
33c06adcf1 Value* were never meant to be const. Removing constness from the constant
folder removes a lot of const_casting and requires no changes to clang or
llvm-gcc.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@82349 91177308-0d34-0410-b5e6-96231b3b80d8
2009-09-20 01:35:59 +00:00
Nick Lewycky
f4d1882743 Teach the constant folder how to handle a few simple i1 cases.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@82340 91177308-0d34-0410-b5e6-96231b3b80d8
2009-09-20 00:04:02 +00:00
Dan Gohman
f672ab89ec Fix an accidental inversion of the inbounds flag.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@81862 91177308-0d34-0410-b5e6-96231b3b80d8
2009-09-15 16:00:30 +00:00
Chris Lattner
dd4238e04d fix PR4963: folding insertvalue would sometimes turn a packed struct into
an unpacked one.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@81845 91177308-0d34-0410-b5e6-96231b3b80d8
2009-09-15 06:28:12 +00:00
Dan Gohman
3bfbc4587a 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@81483 91177308-0d34-0410-b5e6-96231b3b80d8
2009-09-11 00:04:14 +00:00
Dan Gohman
e6992f728a Factor out the code for checking that all indices in a getelementptr are
within the notional bounds of the static type of the getelementptr (which
is not the same as "inbounds") from GlobalOpt into a utility routine,
and use it in ConstantFold.cpp to check whether there are any mis-behaved
indices.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@81478 91177308-0d34-0410-b5e6-96231b3b80d8
2009-09-10 23:37:55 +00:00
Dan Gohman
6e7ad95868 Revert 80959. It isn't sufficient to solve the full problem. And it
introduced regressions in the Ocaml bindings tests.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@80969 91177308-0d34-0410-b5e6-96231b3b80d8
2009-09-03 23:34:49 +00:00
Dan Gohman
e56a94ef91 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.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@80959 91177308-0d34-0410-b5e6-96231b3b80d8
2009-09-03 22:17:40 +00:00
Dan Gohman
f8a87e8343 Cleanup whitespace and indentation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@80451 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-29 23:35:16 +00:00
Owen Anderson
1d0be15f89 Push LLVMContexts through the IntegerType APIs.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78948 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-13 21:58:54 +00:00