llvm-svn: 335178

This fixes an extremely subtle non-determinism that can only be
triggered by an unfortunate alignment of passes. In my case:

- JumpThreading does large dominator tree updates
- CorrelatedValuePropagation preserves domtree now
- LICM codegen depends on the order of children on domtree nodes

The last part is non-deterministic if the update was stored in a set.
But it turns out that the set is completely unnecessary, updates are
deduplicated at an earlier stage so we can just use a vector, which is
both more efficient and doesn't destroy the input ordering.

I didn't manage to get the 240 MB IR file reduced enough, triggering
this bug requires a lot of jump threading, so landing this without a
test case.

Differential Revision: https://reviews.llvm.org/D48392

llvm-svn: 335176

Summary: Make the MemorySSA verify also check that all Phi incoming blocks are block predecessors.

Reviewers: george.burgess.iv

Subscribers: sanjoy, jlebar, Prazek, llvm-commits

Differential Revision: https://reviews.llvm.org/D48333

llvm-svn: 335174

I don't believe there is any real reason to have separate X86 specific opcodes for vector compares. Setcc has the same behavior just uses a different encoding for the condition code.

I had to change the CondCodeAction for SETLT and SETLE to prevent some transforms from changing SETGT lowering.

Differential Revision: https://reviews.llvm.org/D43608

llvm-svn: 335173

As described in D48359, this patch pushes InstructionsState down the BoUpSLP call hierarchy instead of the corresponding raw OpValue. This makes it easier to track the alternate opcode etc. and avoids us having to call getAltOpcode which makes it difficult to support more than one alternate opcode.

Differential Revision: https://reviews.llvm.org/D48382

llvm-svn: 335170

Previously this folding was done only if select is a first operand.
However, for non-commutative operations constant may go before
select.

Differential Revision: https://reviews.llvm.org/D48223

llvm-svn: 335167

llvm-svn: 335165

Additionally, correct the definition of the rdhwr instruction.

Reviewers: atanasyan, abeserminji, smaksimovic

Differential Revision: https://reviews.llvm.org/D48216

llvm-svn: 335162

llvm-svn: 335161

llvm-svn: 335160

llvm-svn: 335158

There are more existing potential users of this,
but I've limited this patch to the first couple
that I found to minimize typo risk.

llvm-svn: 335157

The idea of partial unswitching is to take a *part* of a branch's
condition that is loop invariant and just unswitching that part. This
primarily makes sense with i1 conditions of branches as opposed to
switches. When dealing with i1 conditions, we can easily extract loop
invariant inputs to a a branch and unswitch them to test them entirely
outside the loop.

As part of this, we now create much more significant cruft in the loop
body, so this relies on adding cleanup passes to the loop pipeline and
revisiting unswitched loops to do that cleanup before continuing to
process them.

This already appears to be more powerful at unswitching than the old
loop unswitch pass, and so I'd appreciate pretty careful review in case
I'm just missing some correctness checks. The `LIV-loop-condition` test
case is not unswitched by the old unswitch pass, but is with this pass.

Thanks to Sanjoy and Fedor for the review!

Differential Revision: https://reviews.llvm.org/D46706

llvm-svn: 335156

These instructions were renamed in version 2.2 of the user-level ISA spec, but 
the old name should also be accepted by standard tools.

llvm-svn: 335154

The outliner emits debug info. Add a test that outlines a function
and uses llvm-dwarfdump to check the emitted DWARF for correctness.

llvm-svn: 335153

This utility should operate on Values, not Instructions. While I'm here,
I've also made it possible to skip emitting replacement dbg.values for
certain debug users (by having RewriteExpr return nullptr).

llvm-svn: 335152

These represent the most basic requested transform - a matching
operand and 2 constant operands.

llvm-svn: 335151

Using OrderedInstructions::dominates as comparator for instructions in
BBs without dominance relation can cause a non-deterministic order
between such instructions. That in turn can cause us to materialize
copies in a non-deterministic order. While this does not effect
correctness, it causes some minor non-determinism in the final generated
code, because values have slightly different labels.

Without this patch, running -print-predicateinfo on a reasonably large
module produces slightly different output on each run.

This patch uses the dominator trees DFSInNum to order instruction from
different BBs, which should enforce a deterministic ordering and
guarantee that dominated instructions come after the instructions that
dominate them.

Reviewers: dberlin, efriedma, davide

Reviewed By: efriedma

Differential Revision: https://reviews.llvm.org/D48230

llvm-svn: 335150

llvm-svn: 335146

Summary:
Due to uniqueing of DICompositeTypes, it's possible for a type from one
module to be loaded into another earlier module without being renamed.
Then when the defining module is being IRMoved, the type can be used as
a Mapping destination before being loaded, such that when it's requested
using TypeMapTy::get() it will fail with an assertion that the type is a
source type when it's actually a type in both the source and
destination modules. Correctly handle that case by allowing a non-opaque
non-literal struct type be present in both modules.

Fix for PR37684.

Reviewers: pcc, tejohnson

Reviewed By: pcc, tejohnson

Subscribers: tobiasvk, mehdi_amini, steven_wu, llvm-commits, kcc

Differential Revision: https://reviews.llvm.org/D47898

llvm-svn: 335145

The purpose of this utility is to make it easier for optimizations to
insert replacement dbg.values for instructions they are deleting. This
is useful in situations where salvageDebugInfo is inapplicable, say,
because the new dbg.value cannot refer to an operand of the dying value.

The utility is called insertReplacementDbgValues.

It assumes that the instruction 'From' is going to be deleted, and
inserts replacement dbg.values for each debug user of 'From'. The
newly-inserted dbg.values refer to 'To' instead of 'From'. Each
replacement dbg.value has the same location and variable as the debug
user it replaces, has a DIExpression determined by the result of
'RewriteExpr' applied to an old debug user of 'From', and is placed
before 'InsertBefore'.

This should simplify future patches, like D48331.

llvm-svn: 335144

llvm-svn: 335143

A future patch will have isOneOf use InstructionsState.

llvm-svn: 335142

Summary:
Found some regressions (infinite loop in DAGTypeLegalizer::RemapId)
after r334880. This patch makes sure that we do map a TableId to
itself.

Reviewers: niravd

Reviewed By: niravd

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D48364

llvm-svn: 335141

Summary:
Check that and masks are strictly smaller than implicit mask from
narrowed load.

Fixes PR37820.

Reviewers: samparker, RKSimon, nemanjai

Subscribers: hiraditya, llvm-commits

Differential Revision: https://reviews.llvm.org/D48335

llvm-svn: 335137

Differential revision: https://reviews.llvm.org/D48350

llvm-svn: 335136

Summary:
The waterfall no longer builds .s files and no longers uses
the wasm-o when it builds object files.

Subscribers: dschuff, jgravelle-google, aheejin, sunfish, llvm-commits

Differential Revision: https://reviews.llvm.org/D48371

llvm-svn: 335135

This is part of a move towards generalizing the alternate opcode mechanism and not just supporting (F)Add/(F)Sub counterparts.

The patch embeds the AltOpcode in the InstructionsState instead of calling getAltOpcode so often.

I'm hoping to eventually remove all uses of getAltOpcode and handle alternate opcode selection entirely within getSameOpcode, that will require us to use InstructionsState throughout the BoUpSLP call hierarchy (similar to some of the changes in D28907), which I will begin in future patches.

Differential Revision: https://reviews.llvm.org/D48359

llvm-svn: 335134

llvm-svn: 335131

D47985 saw the old SK_Alternate 'alternating' shuffle mask replaced with the SK_Select mask which accepts either input operand for each lane, equivalent to a vector select with a constant condition operand.

This patch updates SLPVectorizer to make full use of this SK_Select shuffle pattern by removing the 'isOdd()' limitation.

The AArch64 regression will be fixed by D48172.

Differential Revision: https://reviews.llvm.org/D48174

llvm-svn: 335130

For both operands are unsigned, the following optimizations are valid, and missing:

   1. X > Y && X != 0 --> X > Y
   2. X > Y || X != 0 --> X != 0
   3. X <= Y || X != 0 --> true
   4. X <= Y || X == 0 --> X <= Y
   5. X > Y && X == 0 --> false

unsigned foo(unsigned x, unsigned y) { return x > y && x != 0; }
should fold to x > y, but I found we haven't done it right now.
besides, unsigned foo(unsigned x, unsigned y) { return x < y && y != 0; }
Has been folded to x < y, so there may be a bug.

Patch by: Li Jia He!

Differential Revision: https://reviews.llvm.org/D47922

llvm-svn: 335129

These are the baseline tests for the functional change in D47922.

Patch by Li Jia He!

Differential Revision: https://reviews.llvm.org/D48000

llvm-svn: 335128

These are produced by GCC and supported by GAS, but not currently contained in 
the pseudoinstruction listing in the RISC-V ISA manual.

llvm-svn: 335127

llvm-svn: 335126

Summary:
This is useful when hash collisions are unlikely and acceptable, e.g. in clangd
completion ranking.

Reviewers: ioeric

Subscribers: ilya-biryukov, llvm-commits

Differential Revision: https://reviews.llvm.org/D48361

llvm-svn: 335125

Patch by Sumanth Gundapaneni.

llvm-svn: 335124

Reviewers: atanasyan, abeserminji, smaksimovic

Differential Revision: https://reviews.llvm.org/D48166

llvm-svn: 335122

This is similar to:
rL335083

Fixes::
https://bugs.llvm.org/show_bug.cgi?id=37726

llvm-svn: 335121

These are produced by GCC and supported by GAS, but not currently contained in 
the pseudoinstruction listing in the RISC-V ISA manual.

llvm-svn: 335120

Thumb has more 16-bit encoding space dedicated to ADD than ORR, allowing both a
3-address encoding and a wider range of immediates. So, particularly when
optimizing for code size (but it doesn't make things worse elsewhere) it's
beneficial to select an OR operation to an ADD if we know overflow won't occur.

This is made even better by LLVM's penchant for putting operations in canonical
form by converting the other way.

llvm-svn: 335119