We have isel patterns that can put an IMPLICIT_DEF on one of
the sources for these instructions. So we should make sure
we break any dependencies there. This should be done by
just using one of the other sources.

llvm-svn: 373025

Reviewers:
  arsenm

Differential Revision:
  https://reviews.llvm.org/D58360

llvm-svn: 373024

This patches fixes the case when a user-defined mapper is attached to
the elements of an array, and to report error when a mapper is used for
types other than struct, class, and union.

Patch by Lingda Li <lildmh@gmail.com>

Differential Revision: https://reviews.llvm.org/D67978

llvm-svn: 373023

Fixes a regression introduced by r369999.

llvm-svn: 373022

[X86] Add CodeGenOnly instructions for (f32 (X86selects $mask, (loadf32 addr), fp32imm0) to use masked MOVSS from memory.

Similar for f64 and having a non-zero passthru value.

We were previously not trying to fold the load at all. Using
a CodeGenOnly instruction allows us to use FR32X/FR64X as the
register class to avoid a bunch of COPY_TO_REGCLASS.

llvm-svn: 373021

Revert [SLP] Fix for PR31847: Assertion failed: (isLoopInvariant(Operands[i], L) && "SCEVAddRecExpr operand is not loop-invariant!")

This reverts r372626 (git commit 6a278d90)

llvm-svn: 373019

Summary:
This patch extends the current capabilities in loop fusion to fuse guarded loops
(as defined in https://reviews.llvm.org/D63885). The patch adds the necessary
safety checks to ensure that it safe to fuse the guarded loops (control flow
equivalent, no intervening code, and same guard conditions). It also provides an
alternative method to perform the actual fusion of guarded loops. The mechanics
to fuse guarded loops are slightly different then fusing non-guarded loops, so I
opted to keep them separate methods. I will be cleaning this up in later
patches, and hope to converge on a single method to fuse both guarded and
non-guarded loops, but for now I think the review will be easier to keep them
separate.

Reviewers: jdoerfert, Meinersbur, dmgreen, etiotto, Whitney

Subscribers: hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D65464

llvm-svn: 373018

For a runtime loop if we can compute its trip count upperbound:

Don't unroll if:
1. loop is not guaranteed to run either zero or upperbound iterations; and
2. trip count upperbound is less than UnrollMaxUpperBound
Unless user or TTI asked to do so.

If unrolling, limit unroll factor to loop's trip count upperbound.

Differential Revision: https://reviews.llvm.org/D62989

Change-Id: I6083c46a9d98b2e22cd855e60523fdc5a4929c73
llvm-svn: 373017

Summary:
`lldb-vscode` concatenates a string of sourceMap entries
specified in the config, but fails to put a space between
each entry, which causes the settings command to fail.
This patch adds a space between each mapping.

Patch by Richard Howell

Differential Revision: https://reviews.llvm.org/D67569

llvm-svn: 373016

Differential Revision: https://reviews.llvm.org/D68102

llvm-svn: 373015

llvm-svn: 373014

llvm-svn: 373013

llvm-svn: 373012

Summary: As discussed in the loop group meeting. With the current
definition of loop guard, we should not allow multiple loop exiting
blocks. For loops that has multiple loop exiting blocks, we can simply
unable to find the loop guard.
When getUniqueExitBlock() obtains a vector size not equals to one, that
means there is either no exit blocks or there exists more than one
unique block the loop exit to.
If we don't disallow loop with multiple loop exit blocks, then with our
current implementation, there can exist exit blocks don't post dominated
by the non pre-header successor of the guard block.
Reviewer: reames, Meinersbur, kbarton, etiotto, bmahjour
Reviewed By: Meinersbur, kbarton
Subscribers: fhahn, hiraditya, llvm-commits
Tag: LLVM
Differential Revision: https://reviews.llvm.org/D66529

llvm-svn: 373011

declare variant.

We can use the original function if it was used/emitted already. So,
just use warnings for these cases, not errors.

llvm-svn: 373010

This patch emits the function descriptor csect for functions with definitions
under both 32-bit/64-bit mode on AIX.

Differential Revision: https://reviews.llvm.org/D66724

llvm-svn: 373009

Summary:
The syntax table was originally based on and attributed to
jasmin.el, but was rewritten in r45192, so the comment that
says the code comes from jasmin.el is no longer accurate. This
change removes the comment, shortening the code a bit.

Reviewers: MaskRay, lattner

Reviewed By: MaskRay

Subscribers: llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D68042

llvm-svn: 373008

in the dependency source minimizer

Clang allows continuations that have whitespace between the backslash and the newline.
This patch ensures that the dependency source minimizer can handle the whitespace between
the backslash and the newline when looking for a line continuation.

Differential Revision: https://reviews.llvm.org/D68052

llvm-svn: 373007

llvm-svn: 373006

This reverts commit r372940 which was an overreaction to a flaky buildbot.

llvm-svn: 373005

The only functional change here is that -coverage-notes-file is not
passed to -cc1 in some situations.

This code appears to be trying to put the gcno and gcda output next to
the final object file, but it's doing that in a really convoluted way
that needs to be re-examined. It looks for -c or -S in the original
command, and then looks at the -o argument if present in order to handle
the -fno-integrated-as case. However, this doesn't work if this is a
link command with multiple inputs. I looked into fixing this, but the
check-profile test suite has a lot of dependencies on this behavior, so
I left it all alone.

llvm-svn: 373004

Summary:
This patch converts FileSystem::Open from this prototype:

Status
Open(File &File, const FileSpec &file_spec, ...);

to this one:

llvm::Expected<std::unique_ptr<File>>
Open(const FileSpec &file_spec, ...);

This is beneficial on its own, as llvm::Expected is a more modern
and recommended error type than Status.  It is also a necessary step
towards https://reviews.llvm.org/D67891, and further developments
for lldb_private::File.

Reviewers: JDevlieghere, jasonmolenda, labath

Reviewed By: labath

Subscribers: mgorny, lldb-commits

Tags: #lldb

Differential Revision: https://reviews.llvm.org/D67996

llvm-svn: 373003

point.

Reviewed by:    andrew.w.kaylor, uweigand, efriedma
Approved by:    andrew.w.kaylor
Differential Revision:  https://reviews.llvm.org/D67839

llvm-svn: 373002

DAGISelMatcherOpt - TGParser::ParseOperation - silence static analyzer cast_or_null<CheckTypeMatcher> null dereference warning. NFCI.

The static analyzer is warning about a potential null dereference, replace with an null/isa assertion and cast<CheckTypeMatcher>.

llvm-svn: 373001

[InstCombine] Use m_Zero instead of isNullValue() when checking if a GEP index is all zeroes to prevent an infinite loop.

The test case here previously infinite looped. Only one element from the GEP is used so SimplifyDemandedVectorElts would replace the other lanes in each index with undef leading to the first index being <0, undef, undef, undef>. But there's a GEP transform that tries to replace an index into a 0 sized type with a zero index. But the zero index check only works on ConstantInt 0 or ConstantAggregateZero so it would turn the index back to zeroinitializer. Resulting in a loop.

The fix is to use m_Zero() to allow a vector of zeroes and undefs.

Differential Revision: https://reviews.llvm.org/D67977

llvm-svn: 373000

Handling backslashes in include paths in the implementation isn't
non-conforming.

llvm-svn: 372999

llvm-svn: 372998

TGParser::ParseOperation - silence static analyzer dyn_cast<TypedInit> null dereference warning. NFCI.

The static analyzer is warning about a potential null dereference, but we should be able to use cast<TypedInit> directly and if not assert will fire for us.

I've also pulled out the repeated getType() call which was the only user of the pointer.

llvm-svn: 372997

Fixes PR43461 (regression caused by D67504)

The partition field of a SECTIONS-specified section is not set after
D67504. The 0 value affects findSection() which checks if the partition
field is 1.

So `Out::initArray = findSection(".init_array")` is null, and
DT_INIT_ARRAYSZ is not set.

Reviewed By: peter.smith

Differential Revision: https://reviews.llvm.org/D68087

llvm-svn: 372996

Summary:
The following patch avoids segfaulting if the section list is empty when writing a mach-o MH_OBJECT.  I ran into this case from a more complicated example trying to dead_strip while using '-r' in lld.

I'm not sure if having empty sections is a legal mach-o, but it does seem that other llvm-binutils tools can ingest such a boring object with out issue.  Would it be better to emit an error, emit a warning, or do nothing?  It seems that adding a warning diagnostic might be helpful to users, as I did not expect to have a section-less object when the linker was done.

Reviewers: kledzik, ruiu

Subscribers: llvm-commits, jrm

Tags: #lld, #llvm

Differential Revision: https://reviews.llvm.org/D67735

llvm-svn: 372995

You can't use -fno-integrated-as for *-msvc triples because no usable
standalone assembler exists. Perhaps we could teach clang to emit a .s
and then reinvoke itself, but that's a bit silly.

Anyway, fix the test by using an Itanium ABI triple, which will become
mingw, which will assume gnu as is a usable assembler.

llvm-svn: 372994

The static analyzer is warning about a potential null dereference, but we should be able to use cast<ExtractValueInst> directly and if not assert will fire for us.

llvm-svn: 372993

ARMBaseInstrInfo getOperandLatency - silence static analyzer dyn_cast<> null dereference warnings. NFCI.

The static analyzer is warning about potential null dereferences, but we should be able to use cast<> directly and if not assert will fire for us.

llvm-svn: 372992

llvm-svn: 372991

for extreme large case.

We had a case that a single loop which has 4000 exits and the average number
of predecessors of each exit is > 1000, and we found compiling the case spent
a significant amount of time on checking whether a loop has dedicated exits.
This patch adds a limit for the iterations to the check. With the patch, the
time to compile our testcase reduced from 1000s to 200s (clang release build).

Differential Revision: https://reviews.llvm.org/D67359

llvm-svn: 372990

Summary:
FlattenCFG merges two 'if' basicblocks by inserting one basicblock
to another basicblock. The inserted basicblock can have a successor
that contains a PHI node whoes incoming basicblock is the inserted
basicblock. Since the existing code does not handle it, it becomes
a badref.

if (cond1)
  statement
if (cond2)
  statement
successor - contains PHI node whose predecessor is cond2

-->
if (cond1 || cond2)
  statement
(BB for cond2 was deleted)
successor - contains PHI node whose predecessor is cond2 --> bad ref!

Author: Jaebaek Seo

Reviewers: asbirlea, kuhar, tstellar, chandlerc, davide, dexonsmith

Reviewed By: kuhar

Subscribers: hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D68032

llvm-svn: 372989

llvm-svn: 372988

Reviewers: courbet

Subscribers: hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D67918

llvm-svn: 372987

Refined a couple of the comments in the LWZtoc expansion code based on
a post commit review comment.

llvm-svn: 372986

Summary:
This was found during review of https://reviews.llvm.org/D66050.
In the simple test of fdiv, we miss to fold
```
        fneg 2, 2
        xsmaddasp 3, 2, 0
```
to
```
        xsnmsubasp 3, 2, 0
```
We have the patterns for Double Precision and vectors, just missing
Single Precision, the patch add that.

Reviewers: #powerpc, hfinkel, nemanjai, steven.zhang

Reviewed By: #powerpc, steven.zhang

Subscribers: wuzish, hiraditya, kbarton, MaskRay, shchenz, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D67595

llvm-svn: 372985