[ConstExpr] Don't create div/rem expressions

This removes creation of udiv/sdiv/urem/srem constant expressions,
in preparation for their removal. I've added a
ConstantExpr::isDesirableBinOp() predic

[ConstExpr] Don't create div/rem expressions

This removes creation of udiv/sdiv/urem/srem constant expressions,
in preparation for their removal. I've added a
ConstantExpr::isDesirableBinOp() predicate to determine whether
an expression should be created for a certain operator.

With this patch, div/rem expressions can still be created through
explicit IR/bitcode, forbidding them entirely will be the next step.

Differential Revision: https://reviews.llvm.org/D128820

show more ...

[IR] Export ConstantFold.h header (NFC)

This is in preparation for https://discourse.llvm.org/t/rfc-remove-most-constant-expressions/63179.
As part of that change, we'll want to invoke some of these

[IR] Export ConstantFold.h header (NFC)

This is in preparation for https://discourse.llvm.org/t/rfc-remove-most-constant-expressions/63179.
As part of that change, we'll want to invoke some of these constant
folding APIs explicitly, as it won't happen as part of
ConstantExpr::getXYZ() anymore.

Ideally, we'd merge these with the DL-aware constant folding APIs
and only call those, but this is not easily possible for some
current usages (most important IRBuilder, which uses DL-independent
constant folding by default, and some major layering changes would
be needed to change that).

This is basically a reboot of D115035 with different motivation.

Differential Revision: https://reviews.llvm.org/D128213

show more ...

Don't use Optional::hasValue (NFC)

[ConstantFold] Use getFltSemantics instead of manually checking the type

Simplifies the code and makes fpext/fptrunc constant folding not crash
when the result is bf16.

[IR] Allow constant folding (insertelement <vscale x 2 x i32> zeroinitializer, i32 0, i32 i32 0.

Most of insertelement constant folding is blocked if the vector type
is scalable. I believe we can ma

[IR] Allow constant folding (insertelement <vscale x 2 x i32> zeroinitializer, i32 0, i32 i32 0.

Most of insertelement constant folding is blocked if the vector type
is scalable. I believe we can make an exception for inserting null
into an all zeros vector.

Reviewed By: nikic

Differential Revision: https://reviews.llvm.org/D123413

show more ...

[ConstantFold] Fold zero-index GEPs with opaque pointers

With opaque pointers, we can eliminate zero-index GEPs even if
they have multiple indices, as this no longer impacts the result
type of the G

[ConstantFold] Fold zero-index GEPs with opaque pointers

With opaque pointers, we can eliminate zero-index GEPs even if
they have multiple indices, as this no longer impacts the result
type of the GEP.

This optimization is already done for instructions in InstSimplify,
but we were missing the corresponding constant expression handling.

The constexpr transform is a bit more powerful, because it can
produce a vector splat constant and also handles undef values --
it is an extension of an existing single-index transform.

show more ...

[ConstantFold] Fix scalable shufflevector fold with all-undef mask

If the input is scalable, we should not be returning a fixed-width
vector as a result.

Cleanup header dependencies in LLVMCore

Based on the output of include-what-you-use.

This is a big chunk of changes. It is very likely to break downstream code
unless they took a lot of care in avo

Cleanup header dependencies in LLVMCore

Based on the output of include-what-you-use.

This is a big chunk of changes. It is very likely to break downstream code
unless they took a lot of care in avoiding hidden ehader dependencies, something
the LLVM codebase doesn't do that well :-/

I've tried to summarize the biggest change below:

- llvm/include/llvm-c/Core.h: no longer includes llvm-c/ErrorHandling.h
- llvm/IR/DIBuilder.h no longer includes llvm/IR/DebugInfo.h
- llvm/IR/IRBuilder.h no longer includes llvm/IR/IntrinsicInst.h
- llvm/IR/LLVMRemarkStreamer.h no longer includes llvm/Support/ToolOutputFile.h
- llvm/IR/LegacyPassManager.h no longer include llvm/Pass.h
- llvm/IR/Type.h no longer includes llvm/ADT/SmallPtrSet.h
- llvm/IR/PassManager.h no longer includes llvm/Pass.h nor llvm/Support/Debug.h

And the usual count of preprocessed lines:
$ clang++ -E -Iinclude -I../llvm/include ../llvm/lib/IR/*.cpp -std=c++14 -fno-rtti -fno-exceptions | wc -l
before: 6400831
after: 6189948

200k lines less to process is no that bad ;-)

Discourse thread on the topic: https://llvm.discourse.group/t/include-what-you-use-include-cleanup

Differential Revision: https://reviews.llvm.org/D118652

show more ...

[ConstantFold] Disable gep of array bitcast fold with opaque pointers

Once again, this fold is meaningless with opaque pointers, as there
is no pointer element type to canonicalize. At some point, w

[ConstantFold] Disable gep of array bitcast fold with opaque pointers

Once again, this fold is meaningless with opaque pointers, as there
is no pointer element type to canonicalize. At some point, we may
want to do GEP type canonicalizations.

show more ...

[NFC] Remove uses of PointerType::getElementType()

Instead use either Type::getPointerElementType() or
Type::getNonOpaquePointerElementType().

This is part of D117885, in preparation for deprecatin

[NFC] Remove uses of PointerType::getElementType()

Instead use either Type::getPointerElementType() or
Type::getNonOpaquePointerElementType().

This is part of D117885, in preparation for deprecating the API.

show more ...

[ConstFold] Add missing check for inbounds gep

If the gep is not inbounds, then the gep might compute a null
value even if the base pointer is non-null.

[ConstantFold] Remove another incorrect icmp of gep fold

This folded (null + X) == g to false, but of course this is
incorrect if X == g.

Possibly this got confused with the null == g case, which i

[ConstantFold] Remove another incorrect icmp of gep fold

This folded (null + X) == g to false, but of course this is
incorrect if X == g.

Possibly this got confused with the null == g case, which is
already handled elsewhere.

show more ...

[ConstantFold] Remove unnecessary bounded index restriction

The fold for merging a GEP of GEP into a single GEP currently bails
if doing so would result in notional overindexing. The justification
g

[ConstantFold] Remove unnecessary bounded index restriction

The fold for merging a GEP of GEP into a single GEP currently bails
if doing so would result in notional overindexing. The justification
given in the comment above this check is dangerously incorrect: GEPs
with notional overindexing are perfectly fine, and if some code
treats them incorrectly, then that code is broken, not the GEP.
Such a GEP might legally appear in source IR, so only preventing
its creation cannot be sufficient. (The constant folder also ends
up canonicalizing the GEP to remove the notional overindexing, but
that's neither here nor there.)

This check dates back to
https://github.com/llvm/llvm-project/commit/bd4fef4a8939db18f39b108e19097b25e2c7c47a,
and as far as I can tell the original issue this was trying to
patch around has since been resolved.

Differential Revision: https://reviews.llvm.org/D116587

show more ...

[ConstFold] Slightly clean up icmp of two geps fold (NFC)

As we're only dealing with one type of constant expression here,
try to directly cast to GEPOperator.

[ConstantFold] Remove another incorrect icmp of GEP fold

This fold is not correct, because indices might evaluate to zero
even if they are not a literal zero integer. Additionally, this
fold would b

[ConstantFold] Remove another incorrect icmp of GEP fold

This fold is not correct, because indices might evaluate to zero
even if they are not a literal zero integer. Additionally, this
fold would be wrong (in the general case) for non-i8 types as well,
due to index overflow.

Drop this fold and instead let the target-dependent constant
folder compute the actual offset and fold the comparison based
on that.

show more ...

[ConstantFold] Drop unused function (NFC)

isMaybeZeroSizeType() is no longer used after
5afbfe33e7d6ce40af0ad6d99421b443c45b351b.

[ConstantFold] Make icmp of gep fold offset based

We can fold an equality or unsigned icmp between base+offset1 and
base+offset2 with inbounds offsets by comparing the offsets directly.

This replac

[ConstantFold] Make icmp of gep fold offset based

We can fold an equality or unsigned icmp between base+offset1 and
base+offset2 with inbounds offsets by comparing the offsets directly.

This replaces a pair of specialized folds that tried to reason
based on the GEP structure instead. One of those folds was plain
wrong (because it does not account for negative offsets), while
the other is unnecessarily complicated and limited (e.g. it will
fail with bitcasts involved).

The disadvantage of this change is that it requires data layout,
so the fold is no longer performed by datalayout-independent
constant folding. I don't think this is a loss in practice, but
it does regress the ConstantExprFold.ll test, which checks folding
without running any passes.

Differential Revision: https://reviews.llvm.org/D116332

show more ...

[ConstantFolding] Use ICmpInst::Predicate instead of plain integer

The function `ConstantFoldCompareInstruction` uses `unsigned short` to
represent compare predicate, although all usesrs of the resp

[ConstantFolding] Use ICmpInst::Predicate instead of plain integer

The function `ConstantFoldCompareInstruction` uses `unsigned short` to
represent compare predicate, although all usesrs of the respective
include file use definition of CmpInst also. This change replaces
predicate argument type in this function to `ICmpInst::Predicate`,
which allows to make code a bit clearer and simpler.

No functional changes.

Differential Revision: https://reviews.llvm.org/D116379

show more ...

[ConstFold] Don't fold signed comparison of gep of global

An inbounds GEP may still cross the sign boundary, so signed icmps
cannot be folded (https://alive2.llvm.org/ce/z/XSgi4D). This was
previous

[ConstFold] Don't fold signed comparison of gep of global

An inbounds GEP may still cross the sign boundary, so signed icmps
cannot be folded (https://alive2.llvm.org/ce/z/XSgi4D). This was
previously fixed for other folds in this function, but this one
was missed.

show more ...

[NFC] Method for evaluation of FCmpInst for constant operands

New method `FCmpInst::compare` is added, which evaluates the given
compare predicate for constant operands. Interface is made similar to

[NFC] Method for evaluation of FCmpInst for constant operands

New method `FCmpInst::compare` is added, which evaluates the given
compare predicate for constant operands. Interface is made similar to
`ICmpInst::compare`.

Differential Revision: https://reviews.llvm.org/D116168

show more ...

[llvm] Use range-based for loops (NFC)

[IR] In ConstantFoldShuffleVectorInstruction use zeroinitializer for splats of 0

When creating a splat of 0 for scalable vectors we tend to create them
with using a combination of shufflevector and

[IR] In ConstantFoldShuffleVectorInstruction use zeroinitializer for splats of 0

When creating a splat of 0 for scalable vectors we tend to create them
with using a combination of shufflevector and insertelement, i.e.

shufflevector (<vscale x 4 x i32> insertelement (<vscale x 4 x i32> poison, i32 0, i32 0),
<vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer)

However, for the case of a zero splat we can actually just replace the
above with zeroinitializer instead. This makes the IR a lot simpler and
easier to read. I have changed ConstantFoldShuffleVectorInstruction to
use zeroinitializer when creating a splat of integer 0 or FP +0.0 values.

Differential Revision: https://reviews.llvm.org/D113394

show more ...

[NFCI] Introduce `ICmpInst::compare()` and use it where appropriate

As noted in https://reviews.llvm.org/D90924#inline-1076197
apparently this is a pretty common pattern,
let's not repeat it yet aga

[NFCI] Introduce `ICmpInst::compare()` and use it where appropriate

As noted in https://reviews.llvm.org/D90924#inline-1076197
apparently this is a pretty common pattern,
let's not repeat it yet again, but have it in a common place.

There may be some more places where it could be used,
but these are the most obvious ones.

show more ...

[APInt] Stop using soft-deprecated constructors and methods in llvm. NFC.

Stop using APInt constructors and methods that were soft-deprecated in
D109483. This fixes all the uses I found in llvm, exc

[APInt] Stop using soft-deprecated constructors and methods in llvm. NFC.

Stop using APInt constructors and methods that were soft-deprecated in
D109483. This fixes all the uses I found in llvm, except for the APInt
unit tests which should still test the deprecated methods.

Differential Revision: https://reviews.llvm.org/D110807

show more ...

[ConstantFold] ConstantFoldGetElementPtr - use APInt::isNegative() instead of getSExtValue() to support big ints

Fixes fuzz test: https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=39197