The dependency matrix is only empty if no conflicting load/store
instructions have been found. In that case, it is safe to interchange.

For the LLVM test-suite, after this change around 1900 loops are
interchanged, whereas it is 15 before this change. On cortex-a57,
this gives an improvement of -0.57% on the geomean execution
time of SPEC2006, SPEC2000 and the test-suite. There are a
few small perf regressions, but I think we can improve on those
by making the cost model better.

Reviewers: karthikthecool, mcrosier

Reviewed by: karthikthecool

Differential Revision: https://reviews.llvm.org/D43236

llvm-svn: 326077

Differential Revision: https://reviews.llvm.org/D41278

llvm-svn: 326074

Also un-revert (isl_pw_*_alloc: add missing check for compatible spaces, Wed Sep
6 12:18:04 2017 +0200).

This patch is a proposed fix to avoid asserts due to stricter space checking
within isl, which resulted in failures when converting a schedule tree to
a schedule map.

llvm-svn: 326073

The patch introduces the new function in ScalarEvolution to get
all loops used in specified SCEV.

This is a preparation for re-writing isKnownPredicate utility as
described in https://reviews.llvm.org/D42417.

Reviewers: sanjoy, mkazantsev, reames
Reviewed By: sanjoy
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D43504

llvm-svn: 326072

The patch introduces the SCEVPostIncRewriter rewriter which
is similar to SCEVInitRewriter but rewrites AddRec with post increment
value of this AddRec.

This is a preparation for re-writing isKnownPredicate utility as
described in https://reviews.llvm.org/D42417.

Reviewers: sanjoy, mkazantsev, reames
Reviewed By: sanjoy
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D43499

llvm-svn: 326071

[clangd] don't insert new includes if either original header or canonical header is already included.

Summary:
Changes:
o Store both the original header and the canonical header in LSP command.
o Also check that both original and canonical headers are not already included
by comparing both resolved header path and written literal includes.

This addresses the use case where private IWYU pragma is defined in a private
header while it would still be preferrable to include the private header, in the
internal implementation file. If we have seen that the priviate header is already
included, we don't try to insert the canonical include.

Reviewers: sammccall

Reviewed By: sammccall

Subscribers: klimek, ilya-biryukov, jkorous-apple, cfe-commits

Differential Revision: https://reviews.llvm.org/D43510

llvm-svn: 326070

Enable multiple COPY hints to eliminate more COPYs during register allocation.

Note that this is something all targets should do, see
https://reviews.llvm.org/D38128.

Review: Robert Lytton
llvm-svn: 326069

llvm-svn: 326068

The patch introduces an additional parameter IgnoreOtherLoops to SCEVInitRewriter.
if it is equal to true then rewriter will not invalidate result in case
SCEV depends on other loops then specified during creation.

The patch does not change the default behavior.
This is a preparation for re-writing isKnownPredicate utility as
described in https://reviews.llvm.org/D42417.

Reviewers: sanjoy, mkazantsev, reames
Reviewed By: sanjoy
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D43498

llvm-svn: 326067

llvm-svn: 326066

llvm-svn: 326065

This code seemed to try to widen to 128, 256, or 512 bit vectors, but we only create X86ISD::AVG with a power of 2 number of elements. This means the only nodes that need to be legalized are less than 128-bits and need to be widened up to 128 bits.

llvm-svn: 326064

Which types are considered 'simple' is a function of the requirements of all targets that LLVM supports. That shouldn't directly affect what types we are able to handle. The remainder of this code checks that the number of elements is a power of 2 and takes care of splitting down to a legal size.

llvm-svn: 326063

The video is now uploaded to YouTube.

llvm-svn: 326062

It is redundant with the implementation in TypedInit.

Change-Id: I8ab1fb5c77e4923f7eb3ffae5889f0f8af6093b4

Reviewers: arsenm, craig.topper, tra, MartinO

Subscribers: wdng, llvm-commits

Differential Revision: https://reviews.llvm.org/D43678

llvm-svn: 326061

Summary:
FieldInit will just rely on the standardized resolving mechanism to give
us DefInits for folding, thus simplifying the code.

Unlike the removal of resolveListElementReference, this shouldn't have
performance implications, because DefInits do not recurse inside their
record.

Change-Id: Id4544c774c9d9ee92f293615af6ecff706453f21

Reviewers: arsenm, craig.topper, tra, MartinO

Subscribers: wdng, llvm-commits

Differential Revision: https://reviews.llvm.org/D43563

llvm-svn: 326060

Summary:
Resolving a VarListElementInit should just resolve the list and then
take its element. This eliminates a lot of duplicated logic and
simplifies the next steps of refactoring resolveReferences.

This does potentially cause sub-elements of the entire list to be
resolved resulting in more work, but I didn't notice a measurable
change in performance, and a later patch adds a caching mechanism that
covers at least the common case of `var[i]` in a more generic way.

Change-Id: I7b59185b855c7368585c329c31e5be38c5749dac

Reviewers: arsenm, craig.topper, tra, MartinO

Subscribers: wdng, llvm-commits

Differential Revision: https://reviews.llvm.org/D43562

llvm-svn: 326059

When diagnosing the arguments to alloc_size, report the failing argument using a 1-based index instead of a 0-based index for consistency.

Patch by Joel Denny.

llvm-svn: 326058

Fix a failing assertion with the pointer_with_type_tag attribute when the function the attribute appertains to is variadic.

Patch by Joel Denny.

llvm-svn: 326057

Summary: This fixes failure in DebugInfo/X86/multiple-aranges.ll uncovered by D39245.

Reviewers: rafael, echristo, probinson

Reviewed By: probinson

Subscribers: probinson, llvm-commits, JDevlieghere

Tags: #debug-info

Differential Revision: https://reviews.llvm.org/D39950

llvm-svn: 326056

Switch the default behavior of the Clang<> spelling to opt-in to the C2x attribute spellings. NFC to the attribute spellings themselves.

The Clang<> spelling helper generates a spelling for C++11, GNU, and C2x attribute spellings. Previously, users had to manually opt in to the C2x spelling while we cautiously added attributes to that spelling. Now that majority of attributes are exposed in C2x, we can switch the default.

llvm-svn: 326055

Document why the consumed attributes (consumable, callable_when, et al) are not exposed with a C2x spelling. NFC.

llvm-svn: 326054

Add a C2x spelling for the external_source_symbol and internal_linkage attributes in the clang vendor namespace.

Both of these attributes have existing meaning in C code, so there was no reason to exclude them from using the new spelling.

llvm-svn: 326053

Add a C++11 and C2x spelling for the type safety attribute (argument_with_type_tag, pointer_with_type_tag, and type_tag_for_datatype) in the clang vendor namespace.

The TypeTagForDatatype attribute had custom parsing rules that previously prevented it from being supported with square bracket notation. The ArgumentWithTypeTag attribute previously had unnecessary custom parsing that could be handled declaratively.

llvm-svn: 326052

* Address a FIXME by warning the user that both -run-synchronously and -j X are
  passed.
* Fix a comment to suppress clang-tidy warning by passing the correct argument
  name.

Reviewers: ioeric

Subscribers: ilya-biryukov, jkorous-apple, cfe-commits

Differential Revision: https://reviews.llvm.org/D43671

llvm-svn: 326051

llvm-svn: 326050

llvm-svn: 326049

llvm-svn: 326048

Summary:
The current Objective-C global variable declaration check restricts naming that is permitted by the Google Objective-C style guide.

The Objective-C style guide states the following:
"Global and file scope constants should have an appropriate prefix. [...] Constants may use a lowercase k prefix when appropriate"
http://google.github.io/styleguide/objcguide#constants

This change fixes the check to allow two or more capital letters as an appropriate prefix. This change intentionally avoids making a decision regarding whether to flag constants that use a two letter prefix (two letter prefixes are reserved by Apple¹ but many projects seem to violate this guideline).

This change eliminates an important category of false positives (constants prefixed with '[A-Z]{2,}') at the cost of introducing a less important category of false negatives (constants prefixed with only '[A-Z]'). The false positives are observed in standard recommended code while the false negatives occur in non-standard unrecommended code. The number of eliminated false positives is expected to be significantly larger than the number of exposed false negatives.

❧

(1)
"Two-letter prefixes like these are reserved by Apple for use in framework classes."
https://developer.apple.com/library/content/documentation/Cocoa/Conceptual/ProgrammingWithObjectiveC/Conventions/Conventions.html

Reviewers: aaron.ballman, Wizard, hokein, benhamilton

Reviewed By: aaron.ballman, Wizard

Subscribers: aaron.ballman, cfe-commits

Differential Revision: https://reviews.llvm.org/D43581

llvm-svn: 326046

Summary:
If the flag -fforce-enable-int128 is passed, it will enable support for __int128_t and __uint128_t types.
This flag can then be used to build compiler-rt for RISCV32.

Reviewers: asb, kito-cheng, apazos, efriedma

Reviewed By: asb, efriedma

Subscribers: shiva0217, efriedma, jfb, dschuff, sdardis, sbc100, jgravelle-google, aheejin, rbar, johnrusso, simoncook, jordy.potman.lists, sabuasal, niosHD, cfe-commits

Differential Revision: https://reviews.llvm.org/D43105

llvm-svn: 326045

llvm-svn: 326044

llvm-svn: 326043

llvm-svn: 326042

[X86] Allow int_x86_sse2_cvtps2dq and int_x86_avx_cvt_ps2dq_256 to select EVEX encoded instructions.

llvm-svn: 326041

llvm-svn: 326040

[X86] Remove some masked cvt builtins that can be replaced with legacy sse/avx buiiltins and a select.

llvm-svn: 326039

This attribute has custom parsing rules that previously prevented it from being supported with square bracket notation.

llvm-svn: 326038

This reverts commit r325881.

Breaks many bots

llvm-svn: 326037

This attribute has custom parsing rules that previously prevented it from being supported with square bracket notation. Rework the clang attribute argument parsing to be more easily extended for other custom-parsed attributes.

llvm-svn: 326036

llvm-svn: 326035