[InstCombine] Use combineMetadataForCSE in phi of loads fold

Use combineMetadataForCSE instead of manually enumerating known
metadata kinds. This is a typical sinking transform for which
combineMeta

[InstCombine] Use combineMetadataForCSE in phi of loads fold

Use combineMetadataForCSE instead of manually enumerating known
metadata kinds. This is a typical sinking transform for which
combineMetadataForCSE is safe to use (with DoesKMove=true).

Part of https://github.com/llvm/llvm-project/issues/121495.

show more ...

[MemProf][PGO] Prevent dropping of profile metadata during optimization (#121359)

This patch fixes a couple of places where memprof-related metadata
(!memprof and !callsite) were being dropped, and

[MemProf][PGO] Prevent dropping of profile metadata during optimization (#121359)

This patch fixes a couple of places where memprof-related metadata
(!memprof and !callsite) were being dropped, and one place where PGO
metadata (!prof) was being dropped.

All were due to instances of combineMetadata() being invoked. That
function drops all metadata not in the list provided by the client, and
also drops any not in its switch statement.

Memprof metadata needed a case in the combineMetadata switch statement.
For now we simply keep the metadata of the instruction being kept, which
doesn't retain all the profile information when two calls with
memprof metadata are being combined, but at least retains some.

For the memprof metadata being dropped during call CSE, add memprof and
callsite metadata to the list of known ids in combineMetadataForCSE.

Neither memprof nor regular prof metadata were in the list of known ids
for the callsite in MemCpyOptimizer, which was added to combine AA
metadata after optimization of byval arguments fed by memcpy
instructions, and similar types of optimizations of memcpy uses.

There is one other callsite of combineMetadata, but it is only invoked
on load instructions, which do not carry these types of metadata.

show more ...

[InstCombine] Intersect nowrap flags between geps while folding into phi

A miscompilation issue has been addressed with refined checking.

Fixes: https://github.com/llvm/llvm-project/issues/115149.

[InstCombine] Remove dead phi web (#108876)

In current visitPHINode function during InstCombine, it can remove dead
phi cycles (all phis have one use, which is another phi). However, it
cannot dea

[InstCombine] Remove dead phi web (#108876)

In current visitPHINode function during InstCombine, it can remove dead
phi cycles (all phis have one use, which is another phi). However, it
cannot deal with the case when the phis form a web (all phis have one or
more uses, and all the uses are phi). This change extends the algorithm
so that it can also deal with the dead phi web.

show more ...

[InstCombine] Don't look at ConstantData users

When looking at PHI operand for combining, only look at instructions and
arguments. The loop later iteraters over Arg's users, which is not
useful if A

[InstCombine] Don't look at ConstantData users

When looking at PHI operand for combining, only look at instructions and
arguments. The loop later iteraters over Arg's users, which is not
useful if Arg is a constant -- it's users are not meaningful and might
be in different functions, which causes problems for the dominates()
query.

Pull Request: https://github.com/llvm/llvm-project/pull/103302

show more ...

Revert "[IR][NFC] Update IRBuilder to use InsertPosition (#96497)"

Reverts the above commit, as it updates a common header function and
did not update all callsites:

https://lab.llvm.org/buildbot

Revert "[IR][NFC] Update IRBuilder to use InsertPosition (#96497)"

Reverts the above commit, as it updates a common header function and
did not update all callsites:

https://lab.llvm.org/buildbot/#/builders/29/builds/382

This reverts commit 6481dc57612671ebe77fe9c34214fba94e1b3b27.

show more ...

[IR][NFC] Update IRBuilder to use InsertPosition (#96497)

Uses the new InsertPosition class (added in #94226) to simplify some of
the IRBuilder interface, and removes the need to pass a BasicBlock

[IR][NFC] Update IRBuilder to use InsertPosition (#96497)

Uses the new InsertPosition class (added in #94226) to simplify some of
the IRBuilder interface, and removes the need to pass a BasicBlock
alongside a BasicBlock::iterator, using the fact that we can now get the
parent basic block from the iterator even if it points to the sentinel.
This patch removes the BasicBlock argument from each constructor or call
to setInsertPoint.

This has no functional effect, but later on as we look to remove the
`Instruction *InsertBefore` argument from instruction-creation
(discussed
[here](https://discourse.llvm.org/t/psa-instruction-constructors-changing-to-iterator-only-insertion/77845)),
this will simplify the process by allowing us to deprecate the
InsertPosition constructor directly and catch all the cases where we use
instructions rather than iterators.

show more ...

[InstCombine] Preserve all gep flags in dependent IV fold

[InstCombine] Preserve all flags in phi of gep fold

Preserve the intersection of all flags. Add GEPNoWrapFlags::all()
to serve as the initialization value for the intersection.

Fix typo "indicies" (#92232)

[ValueTracking] Restore isKnownNonZero parameter order. (#88873)

Prior to #85863, the required parameters of llvm::isKnownNonZero were
Value and DataLayout. After, they are Value, Depth, and Simpli

[ValueTracking] Restore isKnownNonZero parameter order. (#88873)

Prior to #85863, the required parameters of llvm::isKnownNonZero were
Value and DataLayout. After, they are Value, Depth, and SimplifyQuery,
where SimplifyQuery is implicitly constructible from DataLayout. The
change to move Depth before SimplifyQuery needed callers to be updated
unnecessarily, and as commented in #85863, we actually want Depth to be
after SimplifyQuery anyway so that it can be defaulted and the caller
does not need to specify it.

show more ...

[ValueTracking] Convert `isKnownNonZero` to use SimplifyQuery (#85863)

This patch converts `isKnownNonZero` to use SimplifyQuery. Then we can
use the context information from `DomCondCache`.

Fix

[ValueTracking] Convert `isKnownNonZero` to use SimplifyQuery (#85863)

This patch converts `isKnownNonZero` to use SimplifyQuery. Then we can
use the context information from `DomCondCache`.

Fixes https://github.com/llvm/llvm-project/issues/85823.
Alive2: https://alive2.llvm.org/ce/z/QUvHVj

show more ...

[NFC][RemoveDIs] Insert instruction using iterators in Transforms/

As part of the RemoveDIs project we need LLVM to insert instructions using
iterators wherever possible, so that the iterators can c

[NFC][RemoveDIs] Insert instruction using iterators in Transforms/

As part of the RemoveDIs project we need LLVM to insert instructions using
iterators wherever possible, so that the iterators can carry a bit of
debug-info. This commit implements some of that by updating the contents of
llvm/lib/Transforms/Utils to always use iterator-versions of instruction
constructors.

There are two general flavours of update:
* Almost all call-sites just call getIterator on an instruction
* Several make use of an existing iterator (scenarios where the code is
actually significant for debug-info)
The underlying logic is that any call to getFirstInsertionPt or similar
APIs that identify the start of a block need to have that iterator passed
directly to the insertion function, without being converted to a bare
Instruction pointer along the way.

Noteworthy changes:
* FindInsertedValue now takes an optional iterator rather than an
instruction pointer, as we need to always insert with iterators,
* I've added a few iterator-taking versions of some value-tracking and
DomTree methods -- they just unwrap the iterator. These are purely
convenience methods to avoid extra syntax in some passes.
* A few calls to getNextNode become std::next instead (to keep in the
theme of using iterators for positions),
* SeparateConstOffsetFromGEP has it's insertion-position field changed.
Noteworthy because it's not a purely localised spelling change.

All this should be NFC.

show more ...

[InstCombine] Fold dependent IVs (#81151)

Fold `iv = phi(start, iv.next = iv2.next + start)` where `iv2 =
phi(iv2.start, iv2.next = iv2 + iv2.step)`
to `iv = iv2 + start` removing one induction v

[InstCombine] Fold dependent IVs (#81151)

Fold `iv = phi(start, iv.next = iv2.next + start)` where `iv2 =
phi(iv2.start, iv2.next = iv2 + iv2.step)`
to `iv = iv2 + start` removing one induction variable from the loop.

Proof: https://alive2.llvm.org/ce/z/hfmwgf

Fixes https://github.com/llvm/llvm-project/issues/77108.

show more ...

[InstCombine] Fix phi or icmp fold with disjoint flag

We're changing the operand of the or here, such that the disjoint
flag may no longer hold. Clear it.

[InstCombinePHI] Remove dead PHI on UnaryOperator (#71386)

This patch mainly solves the problem of dead PHI on UnaryOperator

[InstCombine] Extend Phi-Icmp use to include or (#67682)

In InstCombinePHI currently the only use of PHI as an Icmp is being
checked as a requirement to reduce a value if isKnownNonZero.
However t

[InstCombine] Extend Phi-Icmp use to include or (#67682)

In InstCombinePHI currently the only use of PHI as an Icmp is being
checked as a requirement to reduce a value if isKnownNonZero.
However this can be extended to include or(icmp) . This is always true
as OR only adds bits and we are checking against 0.

show more ...

[InstCombine] Expand redundant phi cycle elimination (#67968)

There is a combine in instcombine that will look for phi cycles that only have
a single incoming value:
```
%0 = phi i64 [ %3, %exit

[InstCombine] Expand redundant phi cycle elimination (#67968)

There is a combine in instcombine that will look for phi cycles that only have
a single incoming value:
```
%0 = phi i64 [ %3, %exit ], [ %othervalue, %preheader ]
%3 = phi i64 [ %0, %body ], [ %othervalue, %body2 ]
```

This currently doesn't handle if %othervalue is a phi though, as the algorithm
will recurse into the phi and fail with multiple incoming values. This adjusts
the algorithm, not requiring the initial value to be found immediately,
allowing it to be set to the value of one of the phis that would otherwise fail
due to having multiple input values.

show more ...

[InstCombine] Avoid some uses of ConstantExpr::getZExt() (NFC)

Add helpers getLosslessUnsignedTrunc/getLosslessSignedTrunc for
this common pattern.

[InstCombine] Canonicalize phi order for newly inserted nodes

When new phi nodes are inserted at the start of the block, the
order of these does not get canonicalized, as we pick the first
phi node

[InstCombine] Canonicalize phi order for newly inserted nodes

When new phi nodes are inserted at the start of the block, the
order of these does not get canonicalized, as we pick the first
phi node to canonicalize towards (and the other phi nodes may have
already been visited). This results in phi nodes not being
deduplicated and thus a fix-point verification failure.

Fix this by remembering the predecessors of the first phi node
we encounter -- this will usually result in the same order we
picked previously. I also considered using predecessors() order
instead, but that causes substantial test fallout. Additionally
the predecessors() order tends to be the reverse of the "natural"
order.

Fixes https://github.com/llvm/llvm-project/issues/46688.

show more ...

[NFC][RemoveDIs] Use iterators over inst-pointers when using IRBuilder

This patch adds a two-argument SetInsertPoint method to IRBuilder that
takes a block/iterator instead of an instruction, and up

[NFC][RemoveDIs] Use iterators over inst-pointers when using IRBuilder

This patch adds a two-argument SetInsertPoint method to IRBuilder that
takes a block/iterator instead of an instruction, and updates many call
sites to use it. The motivating reason for doing this is given here [0],
we'd like to pass around more information about the position of debug-info
in the iterator object. That necessitates passing iterators around most of
the time.

[0] https://discourse.llvm.org/t/rfc-instruction-api-changes-needed-to-eliminate-debug-intrinsics-from-ir/68939

Differential Revision: https://reviews.llvm.org/D152468

show more ...

[NFC][RemoveDIs] Prefer iterators over inst-pointers in InstCombine

As per my proposal for how to eliminate debug intrinsics [0], for various
places in InstCombine prefer to insert using an instruct

[NFC][RemoveDIs] Prefer iterators over inst-pointers in InstCombine

As per my proposal for how to eliminate debug intrinsics [0], for various
places in InstCombine prefer to insert using an instruction iterator rather
than an instruction pointer. This is so that we can eventually pass more
information in the iterator class. These call-sites where I've changed the
spelling are those that necessary to build a stage2clang to produce an
identical binary in the coming no-debug-intrinsics mode.

[0] https://discourse.llvm.org/t/rfc-instruction-api-changes-needed-to-eliminate-debug-intrinsics-from-ir/68939

Differential Revision: https://reviews.llvm.org/D152543

show more ...

[InstCombine] Fix worklist management in foldPHIArgIntToPtrToPHI()

Make sure the old operand is added back to the worklist for DCE.

[InstCombine] use dyn_cast instead of isa+cast; NFC

[Local][InstCombine][GVN] Handle !noundef metadata in combineMetadata

Handle !noundef metadata in comhineMetadata. The behavior of violating
!noundef metadata is undefined behavior. So if K dominate

[Local][InstCombine][GVN] Handle !noundef metadata in combineMetadata

Handle !noundef metadata in comhineMetadata. The behavior of violating
!noundef metadata is undefined behavior. So if K dominates J, we can
preserve it uncontionally, otherwise, we preserve it if both K and J
have !noundef metadata been set.

This patch also makes !noundef metadata added in KnownIDs when doing
instruction combine for phi node or global value numbering for loads.

Reviewed By: nikic

Differential Revision: https://reviews.llvm.org/D142801

show more ...