Anton Korobeynikov [Fri, 19 Aug 2016 14:18:34 +0000 (14:18 +0000)]
Revert r279242 - it's failing the tests
llvm-svn: 279247
Krzysztof Parzyszek [Fri, 19 Aug 2016 14:14:09 +0000 (14:14 +0000)]
[Hexagon] Handle J2_jumptpt and J2_jumpfpt instructions
llvm-svn: 279246
Krzysztof Parzyszek [Fri, 19 Aug 2016 14:12:51 +0000 (14:12 +0000)]
[Hexagon] Fix indentation, NFC
llvm-svn: 279245
Krzysztof Parzyszek [Fri, 19 Aug 2016 14:10:57 +0000 (14:10 +0000)]
[Hexagon] Remove unnecessary llvm::, NFC
llvm-svn: 279244
Krzysztof Parzyszek [Fri, 19 Aug 2016 14:09:47 +0000 (14:09 +0000)]
[Hexagon] Rename the HEXAGON_MC namespace to Hexagon_MC, NFC
llvm-svn: 279243
Anton Korobeynikov [Fri, 19 Aug 2016 14:07:10 +0000 (14:07 +0000)]
Fix PR27500: on MSP430 the branch destination offset is measured in words, not bytes.
In addition, the branch instructions will have proper BB destinations, not offsets, like before.
Patch by Vadzim Dambrouski!
Differential Revision: https://reviews.llvm.org/D20162
llvm-svn: 279242
Krzysztof Parzyszek [Fri, 19 Aug 2016 14:04:45 +0000 (14:04 +0000)]
[Hexagon] Mark PS_jumpret as pseudo-instruction, expand it into J2_jumpr
llvm-svn: 279241
Andrey Bokhanko [Fri, 19 Aug 2016 13:36:31 +0000 (13:36 +0000)]
Fixed a typo (compilation_commands.json --> compile_commands.json).
llvm-svn: 279240
Krzysztof Parzyszek [Fri, 19 Aug 2016 13:34:31 +0000 (13:34 +0000)]
[Hexagon] Improvements to handling and generation of FP instructions
Improved handling of fma, floating point min/max, additional load/store
instructions for floating point types.
Patch by Jyotsna Verma.
llvm-svn: 279239
Pavel Labath [Fri, 19 Aug 2016 13:14:13 +0000 (13:14 +0000)]
Fix 32-bit builds after r279232
GetByteSize() of a DataBuffer returns a uint64_t (it probably shouldn't), which isn't implicitly
convertible to size_t.
llvm-svn: 279238
Filipe Cabecinhas [Fri, 19 Aug 2016 13:07:23 +0000 (13:07 +0000)]
Cleanup: Move the *AddressDescription printing code to Print() members inside those structs.
Summary:
The Print() members might take optional access_size and bug_type
parameters to still be able to provide the same information
Reviewers: kcc, samsonov
Subscribers: kubabrecka, llvm-commits
Differential Revision: https://reviews.llvm.org/D23658
llvm-svn: 279237
George Rimar [Fri, 19 Aug 2016 12:57:45 +0000 (12:57 +0000)]
[ELF] - Removed excessive llvm-readobj call from linkerscript-ouputformat.s. NFC.
llvm-svn: 279236
Valery Pykhtin [Fri, 19 Aug 2016 12:54:31 +0000 (12:54 +0000)]
[AMDGPU] add s_incperflevel/s_decperflevel builtins
Differential revision: https://reviews.llvm.org/D23668
llvm-svn: 279235
Dimitar Vlahovski [Fri, 19 Aug 2016 12:44:53 +0000 (12:44 +0000)]
Fixing a Darwing test thats failing on windows
The pexpect import should be make after the skip-if-not-darwin part
because pexpect is not available on Windows
llvm-svn: 279234
Benjamin Kramer [Fri, 19 Aug 2016 12:44:24 +0000 (12:44 +0000)]
[LoopVectorize] Don't copy std::vector in for-range loop.
llvm-svn: 279233
Pavel Labath [Fri, 19 Aug 2016 12:31:49 +0000 (12:31 +0000)]
Remove the last manually constructed packet from gdb-remote register context + small refactor
Summary:
The tricky part here was that the exisiting implementation of WriteAllRegisters was expecting
hex-encoded data (as that was what the first implementation I replaced was using, but here we had
binary data to begin with. I thought the read/write register functions would be more useful if
they handled the hex-encoding themselves (all the other client functions provide the responses in
a more-or-less digested form). The read functions return a DataBuffer, so they can allocate as
much memory as they need to, while the write functions functions take an llvm::ArrayRef, as that
can be constructed from pretty much anything.
Reviewers: clayborg
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D23659
llvm-svn: 279232
Chandler Carruth [Fri, 19 Aug 2016 10:51:55 +0000 (10:51 +0000)]
[PM] Revert r279227 and r279228 until I can find someone to help me
solve completely opaque MSVC build errors. It complains about lots of
stuff with this change without givin nearly enough information to even
try to fix.
llvm-svn: 279231
Simon Pilgrim [Fri, 19 Aug 2016 10:31:53 +0000 (10:31 +0000)]
[X86][SSE] Add support for matching commuted insertps patterns
INSERTPS doesn't fit well with our shuffle mask canonicalization, so we need to attempt both the original mask and the commuted mask to more likely get a match
llvm-svn: 279230
James Molloy [Fri, 19 Aug 2016 10:10:27 +0000 (10:10 +0000)]
[SimplifyCFG] Rewrite SinkThenElseCodeToEnd
The new version has several advantages:
1) IMSHO it's more readable and neater
2) It handles loads and stores properly
3) It can handle any number of incoming blocks rather than just two. I'll be taking advantage of this in a followup patch.
With this change we can now finally sink load-modify-store idioms such as:
if (a)
return *b += 3;
else
return *b += 4;
=>
%z = load i32, i32* %y
%.sink = select i1 %a, i32 5, i32 7
%b = add i32 %z, %.sink
store i32 %b, i32* %y
ret i32 %b
When this works for switches it'll be even more powerful.
llvm-svn: 279229
Chandler Carruth [Fri, 19 Aug 2016 09:53:10 +0000 (09:53 +0000)]
[PM] Fix a compile error with GCC. NFC.
llvm-svn: 279228
Chandler Carruth [Fri, 19 Aug 2016 09:45:16 +0000 (09:45 +0000)]
[PM] Make the the new pass manager support fully generic extra arguments
to run methods, both for transform passes and analysis passes.
This also allows the analysis manager to use a different set of extra
arguments from the pass manager where useful. Consider passes over
analysis produced units of IR like SCCs of the call graph or loops.
Passes of this nature will often want to refer to the analysis result
that was used to compute their IR units (the call graph or LoopInfo).
And for transformations, they may want to communicate special update
information to the outer pass manager. With this change, it becomes
possible to have a run method for a loop pass that looks more like:
PreservedAnalyses run(Loop &L, AnalysisManager<Loop, LoopInfo> &AM,
LoopInfo &LI, LoopUpdateRecord &UR);
And to query the analysis manager like:
AM.getResult<MyLoopAnalysis>(L, LI);
This makes accessing the known-available analyses convenient and clear,
and it makes passing customized data structures around easy.
My initial use case is going to be in updating the pass manager layers
when the analysis units of IR change. But there are more use cases here
such as having a layer that lets inner passes signal whether certain
additional passes should be run because of particular simplifications
made. Two desires for this have come up in the past: triggering
additional optimization after successfully unrolling loops, and
triggering additional inlining after collapsing indirect calls to direct
calls.
Despite adding this layer of generic extensibility, the *only* change to
existing, simple usage are for places where we forward declare the
AnalysisManager template. We really shouldn't be doing this because of
the fragility exposed here, but currently it makes coping with the
legacy PM code easier.
Differential Revision: http://reviews.llvm.org/D21462
llvm-svn: 279227
Kirill Bobyrev [Fri, 19 Aug 2016 09:36:14 +0000 (09:36 +0000)]
minor header guards and help messages cleanup
llvm-svn: 279226
Chandler Carruth [Fri, 19 Aug 2016 09:26:00 +0000 (09:26 +0000)]
[PM] Try to work-around what appears to be an MSVC SFINAE issue with
r279217 where it fails to select the path that other compilers select.
The workaround won't be as careful to produce an error when an analysis
result is incorrect, but we can rely on non-MSVC builds to catch such
errors it seems and MSVC doesn't seem to support the alternative
techniques.
Hoping this brings the windows bots back to life. If not, will have to
revert all of this.
llvm-svn: 279225
James Molloy [Fri, 19 Aug 2016 08:38:50 +0000 (08:38 +0000)]
[CodeGen] Fix a trivial type conversion bug dating back to pre-2008
The heuristic above this code is incredibly suspect, but disregarding that it mutates the cast opcode so we need to check the *mutated* opcode later to see if we need to emit an AssertSext or AssertZext node.
Fixes PR29041.
llvm-svn: 279223
Vitaly Buka [Fri, 19 Aug 2016 08:33:53 +0000 (08:33 +0000)]
[asan] Fix size of shadow incorrectly calculated in r279178
Summary: r279178 generates 8 times more stores than necessary.
Reviewers: eugenis
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D23708
llvm-svn: 279222
Chandler Carruth [Fri, 19 Aug 2016 08:31:47 +0000 (08:31 +0000)]
[PM] NFC refactoring: remove the AnalysisManagerBase class, folding it
into the AnalysisManager class template.
Back when I first added this base class there were separate analysis
managers and some plausible reason why it would be a useful factoring of
common code between them. However, after a lot of refactoring cleaning,
we now have *entirely* shared code. The base class was just an arbitrary
division between code in one class template and a separate class
template. It didn't add anything and forced lots of indirection through
"derived_this" for no real gain.
We can always factor a base CRTP class out with common code if there is
ever some *other* analysis manager that wants to share a subset of
logic. But for now, folding things into the primary template is
a non-trivial simplification with no down sides I see. It shortens the
code considerably, removes an unhelpful abstraction, and will make
subsequent patches *dramatically* less complex which enhance the
analysis manager infrastructure to effectively cope with invalidation.
llvm-svn: 279221
George Rimar [Fri, 19 Aug 2016 08:31:02 +0000 (08:31 +0000)]
[ELF] - Do not change binding of symbols when creating relocatable output.
Spec says "A hidden symbol contained in a relocatable object must be either
removed or converted to STB_LOCAL binding by the link-editor when the
relocatable object is included in an executable file or shared object".
But we previously converted symbols to STB_LOCAL even when -r was specified.
Broken binary was produced, this is PR28967, patch fixes the issue.
Differential revision: https://reviews.llvm.org/D23514
llvm-svn: 279220
Vassil Vassilev [Fri, 19 Aug 2016 08:30:42 +0000 (08:30 +0000)]
[modules] Add missing include.
llvm-svn: 279219
Jonas Hahnfeld [Fri, 19 Aug 2016 08:03:26 +0000 (08:03 +0000)]
Revert "[CMake] Fix ASM building in llvm/runtimes"
This reverts the TSAN parts of commit r279215.
llvm-svn: 279218
Chandler Carruth [Fri, 19 Aug 2016 07:49:23 +0000 (07:49 +0000)]
[PM] Redesign how the new PM detects whether an analysis result provides
its own invalidate method.
Previously, the technique would assume that if a result didn't have an
invalidate method that didn't exactly match the expected signature it
didn't have one at all. This is in fact not the case. And we had
analyses with incorrect signatures for the invalidate method in the
tree that would be erroneously invalidated in certain cases! Yikes.
Moreover a result might legitimately want to have multiple overloads for
the invalidate method, and if one changes or a new one is needed we
again really want a compiler error. For example in the tree we had not
added the overload for a *function* IR unit to the invalidate routine
for TLI. Doh.
So a new techique for the SFINAE detection here: if the result has *any*
member spelled "invalidate" we turn off the synthesis of a default
version. We don't care if it is a member function or a member variable
or how many overloads there are. Once a result has something by that
name it must provide suitable overloads for the contexts in which it is
used. This seems much more resilient and durable.
Huge props to Richard Smith who helped me figure out how on earth we
could even do this in C++. It took quite some doing. The technique is
remarkably clean however, and merely requires that the analysis results
are not *final* classes. I think that's a requirement we can live with
even if it is a bit odd.
I've fixed the two bad in-tree analysis results. And this will make my
next change which changes the API for invalidate much easier to
validate as correct.
llvm-svn: 279217
Chandler Carruth [Fri, 19 Aug 2016 07:49:19 +0000 (07:49 +0000)]
[PM] Rework the new PM support for building the ModuleSummaryIndex to
directly produce the index as the value type result.
This requires making the index movable which is straightforward. It
greatly simplifies things by allowing us to completely avoid the builder
API and the layers of abstraction inherent there. Instead both pass
managers can directly construct these when run by value. They still
won't be constructed truly eagerly thanks to the optional in the legacy
PM. The code that directly builds the index can also just share a direct
function.
A notable change here is that the result type of the analysis for the
new PM is no longer a reference type. This was really problematic when
making changes to how we handle result types to make our interface
requirements *much* more strict and precise. But I think this is an
overall improvement.
Differential Revision: https://reviews.llvm.org/D23701
llvm-svn: 279216
Jonas Hahnfeld [Fri, 19 Aug 2016 06:46:00 +0000 (06:46 +0000)]
[CMake] Fix ASM building in llvm/runtimes
When compiler-rt's CMake is not directly invoked, it will currently not call
project() and thus ASM will not be enabled.
We also don't need to put the .S files through the C compiler then.
Differential Revision: https://reviews.llvm.org/D23656
llvm-svn: 279215
NAKAMURA Takumi [Fri, 19 Aug 2016 06:44:44 +0000 (06:44 +0000)]
Fix tests in llvm/test/tools/gold/X86 to satisfy r279014.
They would unexpectedly pass if test/tools/gold/X86/Output had outputs of previous tests.
llvm-svn: 279214
Xinliang David Li [Fri, 19 Aug 2016 06:31:45 +0000 (06:31 +0000)]
[Profile] Fix edge count read bug
Use uint64_t to avoid value truncation before scaling.
llvm-svn: 279213
Mehdi Amini [Fri, 19 Aug 2016 06:10:03 +0000 (06:10 +0000)]
[LTO] Move callback member from base class to the derived where it is used (NFC)
llvm-svn: 279212
Mehdi Amini [Fri, 19 Aug 2016 06:06:18 +0000 (06:06 +0000)]
Constify some path in the bitcode writer (NFC)
llvm-svn: 279211
Mehdi Amini [Fri, 19 Aug 2016 05:56:37 +0000 (05:56 +0000)]
[LTO] Add a move to inialize member in ctor initialization list (NFC)
llvm-svn: 279210
Xinliang David Li [Fri, 19 Aug 2016 05:31:33 +0000 (05:31 +0000)]
[Profile] Simple code refactoring for reuse /NFC
llvm-svn: 279209
Todd Fiala [Fri, 19 Aug 2016 05:26:40 +0000 (05:26 +0000)]
add all missing gtest tests to the Xcode lldb-gtest target
Exclusions:
ModuleCacheTest: https://llvm.org/bugs/show_bug.cgi?id=29045
SymbolFilePDBTests: https://llvm.org/bugs/show_bug.cgi?id=29046
llvm-svn: 279208
NAKAMURA Takumi [Fri, 19 Aug 2016 05:25:26 +0000 (05:25 +0000)]
clang/test/Modules/prebuilt-module.m: Prune "-triple" to fix a configuration mismatch.
llvm-svn: 279207
Yaxun Liu [Fri, 19 Aug 2016 05:17:25 +0000 (05:17 +0000)]
Re-commit [OpenCL] AMDGCN: Fix size_t type
There was a premature cast to pointer type in emitPointerArithmetic which caused assertion in tests with assertion enabled.
llvm-svn: 279206
Todd Fiala [Fri, 19 Aug 2016 04:55:12 +0000 (04:55 +0000)]
fixup macOS gtest running
Needed to explicitly turn off RTTI (since lldb-core builds
with -fno-rtti) and specify -DGTEST_HAVE_RTTI=0.
llvm-svn: 279205
Dean Michael Berris [Fri, 19 Aug 2016 04:44:30 +0000 (04:44 +0000)]
[XRay] Synthesize a reference to the xray_instr_map
Without the synthesized reference to a symbol in the xray_instr_map,
linker section garbage collection will helpfully remove the whole
xray_instr_map section from the final executable (or archive). This will
cause the runtime to not be able to identify the sleds and hot-patch the
calls/jumps into the runtime trampolines.
This change adds a reference from the text section at the end of the
function to keep around the associated xray_instr_map section as well.
We also make sure that we catch this reference in the test.
Reviewers: chandlerc, echristo, majnemer, mehdi_amini
Subscribers: mehdi_amini, llvm-commits, dberris
Differential Revision: https://reviews.llvm.org/D23398
llvm-svn: 279204
Todd Fiala [Fri, 19 Aug 2016 04:38:44 +0000 (04:38 +0000)]
fix broken gdb-remote gtest
This change adds the Process/gdb-remote gtests to the Xcode
build. It also adds a virtual method impl to the continuation
delegate that I added with the StructuredDataPlugin change.
llvm-svn: 279203
Todd Fiala [Fri, 19 Aug 2016 04:21:48 +0000 (04:21 +0000)]
Add StructuredData plugin type; showcase with new DarwinLog feature
Take 2, with missing cmake line fixed. Build tested on
Ubuntu 14.04 with clang-3.6.
See docs/structured_data/StructuredDataPlugins.md for details.
differential review: https://reviews.llvm.org/D22976
reviewers: clayborg, jingham
llvm-svn: 279202
Lang Hames [Fri, 19 Aug 2016 03:12:16 +0000 (03:12 +0000)]
[RuntimeDyld][MCJIT] Un-XFAIL some tests that were fixed by r279182.
llvm-svn: 279201
Todd Fiala [Fri, 19 Aug 2016 03:03:58 +0000 (03:03 +0000)]
Revert "Add StructuredData plugin type; showcase with new DarwinLog feature"
This reverts commit
1d885845d1451e7b232f53fba2e36be67aadabd8.
llvm-svn: 279200
Matthias Braun [Fri, 19 Aug 2016 03:03:24 +0000 (03:03 +0000)]
Revert "RegScavenging: Add scavengeRegisterBackwards()"
The ppc64 multistage bot fails on this.
This reverts commit r279124.
Also Revert "CodeGen: Add/Factor out LiveRegUnits class; NFCI" because it depends on the previous change
This reverts commit r279171.
llvm-svn: 279199
Todd Fiala [Fri, 19 Aug 2016 02:52:07 +0000 (02:52 +0000)]
Add StructuredData plugin type; showcase with new DarwinLog feature
See docs/structured_data/StructuredDataPlugins.md for details.
differential review: https://reviews.llvm.org/D22976
reviewers: clayborg, jingham
llvm-svn: 279198
Chandler Carruth [Fri, 19 Aug 2016 02:07:51 +0000 (02:07 +0000)]
[ADT] Add the worlds simplest STL extra. Or at least close to it.
This is a little class template that just builds an inheritance chain of
empty classes. Despite how simple this is, it can be used to really
nicely create ranked overload sets. I've added a unittest as much to
document this as test it. You can pass an object of this type as an
argument to a function overload set an it will call the first viable and
enabled candidate at or below the rank of the object.
I'm planning to use this in a subsequent commit to more clearly rank
overload candidates used for SFINAE. All credit for this technique and
both lines of code here to Richard Smith who was helping me rewrite the
SFINAE check in question to much more effectively capture the intended
set of checks.
llvm-svn: 279197
Richard Smith [Fri, 19 Aug 2016 01:43:06 +0000 (01:43 +0000)]
C++ Modules TS: support parsing the 'module' declaration (including extensions
from p0273r0 approved by EWG). We'll eventually need to handle this from the
lexer as well, in order to disallow preprocessor directives preceding the
module declaration and to support macro import.
llvm-svn: 279196
Devin Coughlin [Fri, 19 Aug 2016 01:22:50 +0000 (01:22 +0000)]
[www] Add analyzer FAQ about not releasing ivars in -dealloc.
llvm-svn: 279183
Lang Hames [Fri, 19 Aug 2016 01:15:39 +0000 (01:15 +0000)]
[RuntimeDyld] Add support for ELF R_ARM_REL32 and R_ARM_GOT_PREL.
Patch by William Dillon. Thanks William!
This patch adds support for the R_ARM_REL32 and R_ARM_GOT_PREL ELF ARM
relocations to RuntimeDyld, which should allow JITing of code that
produces these relocations.
No test case: Unfortunately RuntimeDyldELF's GOT building mechanism (which
uses a separate section for GOT entries) isn't compatible with
RuntimeDyldChecker. The correct fix for this is to fix RuntimeDyldELF's GOT
support (it's fundamentally broken at the moment: separate sections aren't
guaranteed to be in range of a GOT entry load), but that's a non-trivial job.
llvm-svn: 279182
Devin Coughlin [Fri, 19 Aug 2016 01:05:31 +0000 (01:05 +0000)]
[analyzer] Weaken assertion in trackNullOrUndefValue()
We should ignore paren casts when making sure that the semantic expression
in a PseudoObjectExpr for an ObjC getter is a message send.
This has no other intended functionality change.
Adding a test for this exposed an interesting issue in another test case
that only manifests under ARC. trackNullOrUndefValue() is not properly
suppressing for nil values that are the result of nil propagation from a nil
receiver when the nil is returned from a function. I've added a FIXME for that
missing suppression.
rdar://problem/
27290568
llvm-svn: 279181
Saleem Abdulrasool [Thu, 18 Aug 2016 23:59:12 +0000 (23:59 +0000)]
EHABI: cover switch once more
When making WMMX support optional, we uncovered the switch. Add the missing
entries. Since the entry is a break leading to a dead path, it should get
optimized out yet retain the switch overage.
llvm-svn: 279180
Vitaly Buka [Thu, 18 Aug 2016 23:51:15 +0000 (23:51 +0000)]
[asan] Optimize store size in FunctionStackPoisoner::poisonRedZones
Summary: Reduce store size to avoid leading and trailing zeros.
Reviewers: kcc, eugenis
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D23648
llvm-svn: 279178
Andrew Kaylor [Thu, 18 Aug 2016 22:49:51 +0000 (22:49 +0000)]
Include X86CallFrameOptimization in the opt-bisect process.
Differential Revision: https://reviews.llvm.org/D23683
llvm-svn: 279175
Saleem Abdulrasool [Thu, 18 Aug 2016 22:35:06 +0000 (22:35 +0000)]
AArch64: remove extraneous padding
The structs BarrierOp, PrefetchOp, PSBHintOp are in AArch64AsmParser.cpp
(inside anonymous namespace). This diff changes the order of fields and
removes the excessive padding (8 bytes).
Patch by Alexander Shaposhnikov!
llvm-svn: 279173
Chris Bieneman [Thu, 18 Aug 2016 22:18:11 +0000 (22:18 +0000)]
[CMake] Add variables for tracking which runtimes are included
This allows sub-projects to have conditionals based on the presence of other projects.
llvm-svn: 279172
Matthias Braun [Thu, 18 Aug 2016 22:11:28 +0000 (22:11 +0000)]
CodeGen: Add/Factor out LiveRegUnits class; NFCI
This is a set of register units intended to track register liveness, it
is similar in spirit to LivePhysRegs.
You can also think of this as the liveness tracking parts of the
RegisterScavenger factored out into an own class.
This was proposed in http://llvm.org/PR27609
Differential Revision: http://reviews.llvm.org/D21916
llvm-svn: 279171
Eugene Zelenko [Thu, 18 Aug 2016 22:11:27 +0000 (22:11 +0000)]
[Documentation] Fixed style in modernize-use-emplace.rst.
llvm-svn: 279170
Jacques Pienaar [Thu, 18 Aug 2016 22:10:06 +0000 (22:10 +0000)]
Fix link quotes on AArch64's CompilerWriterInfo section.
Reviewers: t.p.northover
Subscribers: t.p.northover, aemerson, rengolin
Differential Revision: https://reviews.llvm.org/D23697
llvm-svn: 279169
Kyle Butt [Thu, 18 Aug 2016 22:09:27 +0000 (22:09 +0000)]
CodeGen: If Convert blocks that would form a diamond when tail-merged.
The following function currently relies on tail-merging for if
conversion to succeed. The common tail of cond_true and cond_false is
extracted, and this then forms a diamond pattern that can be
successfully if converted.
If this block does not get extracted, either because tail-merging is
disabled or the threshold is higher, we should still recognize this
pattern and if-convert it.
Fixed a regression in the original commit. Need to un-reverse branches after
reversing them, or other conversions go awry.
Regression on self-hosting bots with no obvious explanation. Tidied up range
handling to be more obviously correct, but there was no smoking gun.
define i32 @t2(i32 %a, i32 %b) nounwind {
entry:
%tmp1434 = icmp eq i32 %a, %b ; <i1> [#uses=1]
br i1 %tmp1434, label %bb17, label %bb.outer
bb.outer: ; preds = %cond_false, %entry
%b_addr.021.0.ph = phi i32 [ %b, %entry ], [ %tmp10, %cond_false ]
%a_addr.026.0.ph = phi i32 [ %a, %entry ], [ %a_addr.026.0, %cond_false ]
br label %bb
bb: ; preds = %cond_true, %bb.outer
%indvar = phi i32 [ 0, %bb.outer ], [ %indvar.next, %cond_true ]
%tmp. = sub i32 0, %b_addr.021.0.ph
%tmp.40 = mul i32 %indvar, %tmp.
%a_addr.026.0 = add i32 %tmp.40, %a_addr.026.0.ph
%tmp3 = icmp sgt i32 %a_addr.026.0, %b_addr.021.0.ph
br i1 %tmp3, label %cond_true, label %cond_false
cond_true: ; preds = %bb
%tmp7 = sub i32 %a_addr.026.0, %b_addr.021.0.ph
%tmp1437 = icmp eq i32 %tmp7, %b_addr.021.0.ph
%indvar.next = add i32 %indvar, 1
br i1 %tmp1437, label %bb17, label %bb
cond_false: ; preds = %bb
%tmp10 = sub i32 %b_addr.021.0.ph, %a_addr.026.0
%tmp14 = icmp eq i32 %a_addr.026.0, %tmp10
br i1 %tmp14, label %bb17, label %bb.outer
bb17: ; preds = %cond_false, %cond_true, %entry
%a_addr.026.1 = phi i32 [ %a, %entry ], [ %tmp7, %cond_true ], [ %a_addr.026.0, %cond_false ]
ret i32 %a_addr.026.1
}
Without tail-merging or diamond-tail if conversion:
LBB1_1: @ %bb
@ =>This Inner Loop Header: Depth=1
cmp r0, r1
ble LBB1_3
@ BB#2: @ %cond_true
@ in Loop: Header=BB1_1 Depth=1
subs r0, r0, r1
cmp r1, r0
it ne
cmpne r0, r1
bgt LBB1_4
LBB1_3: @ %cond_false
@ in Loop: Header=BB1_1 Depth=1
subs r1, r1, r0
cmp r1, r0
bne LBB1_1
LBB1_4: @ %bb17
bx lr
With diamond-tail if conversion, but without tail-merging:
@ BB#0: @ %entry
cmp r0, r1
it eq
bxeq lr
LBB1_1: @ %bb
@ =>This Inner Loop Header: Depth=1
cmp r0, r1
ite le
suble r1, r1, r0
subgt r0, r0, r1
cmp r1, r0
bne LBB1_1
@ BB#2: @ %bb17
bx lr
llvm-svn: 279168
Kyle Butt [Thu, 18 Aug 2016 22:09:25 +0000 (22:09 +0000)]
IfConversion: Rescan diamonds.
The cost of predicating a diamond is only the instructions that are not shared
between the two branches. Additionally If a predicate clobbering instruction
occurs in the shared portion of the branches (e.g. a cond move), it may still
be possible to if convert the sub-cfg. This change handles these two facts by
rescanning the non-shared portion of a diamond sub-cfg to recalculate both the
predication cost and whether both blocks are pred-clobbering.
llvm-svn: 279167
Kyle Butt [Thu, 18 Aug 2016 22:09:23 +0000 (22:09 +0000)]
IfConversion: Handle inclusive ranges more carefully.
This may affect calculations for thresholds, but is not a significant change
in behavior.
The problem was that an inclusive range must have an additonal flag to showr
that it is empty, because otherwise begin == end implies that the range has one
element, and it may not be possible to move past on either side.
llvm-svn: 279166
Changpeng Fang [Thu, 18 Aug 2016 22:04:54 +0000 (22:04 +0000)]
AMDGPU: Add clang builtin for ds_swizzle.
Summary:
int __builtin_amdgcn_ds_swizzle (int a, int imm);
while imm is a constant.
Differential Revision:
http://reviews.llvm.org/D23682
llvm-svn: 279165
Vassil Vassilev [Thu, 18 Aug 2016 22:01:25 +0000 (22:01 +0000)]
PR28794: Don't try to instantiate function templates which are not visible.
Reviewed by Richard Smith.
llvm-svn: 279164
Richard Smith [Thu, 18 Aug 2016 21:59:42 +0000 (21:59 +0000)]
C++ Modules TS: Add parsing support for module import declaration.
llvm-svn: 279163
Petr Hosek [Thu, 18 Aug 2016 21:55:23 +0000 (21:55 +0000)]
[ELF] Improve error reporting for relocations
We should always include symbol name when reporting relocations
error to simplify debugging of these issues. Without symbol names
users have to manually investigate which of the libraries contain
invalid relocations which can be cumbersome when linking multiple
libraries.
Differential Revision: https://reviews.llvm.org/D23690
llvm-svn: 279162
Hemant Kulkarni [Thu, 18 Aug 2016 21:50:13 +0000 (21:50 +0000)]
llvm-objdump: Add Hexagon printer changes for -S/-l options
Differential Revision: https://reviews.llvm.org/D23521
llvm-svn: 279161
Chris Bieneman [Thu, 18 Aug 2016 21:47:18 +0000 (21:47 +0000)]
[CMake] Create convenience targets for runtime projects
Each runtime project has a top-level target that is the name of the runtime (minus the "lib" prefix if applicable). This creates top-level targets mapping to runtime projects.
llvm-svn: 279160
Justin Bogner [Thu, 18 Aug 2016 21:46:54 +0000 (21:46 +0000)]
CodeGen: Rename a variable to better fit LLVM style. NFC
llvm-svn: 279159
Zhan Jun Liau [Thu, 18 Aug 2016 21:44:15 +0000 (21:44 +0000)]
[SystemZ] Use valid base/index regs for inline asm
Summary:
Inline asm memory constraints can have the base or index register be assigned
to %r0 right now. Make sure that we assign only ADDR64 registers to the base
and index.
Reviewers: uweigand
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D23367
llvm-svn: 279157
Tim Shen [Thu, 18 Aug 2016 21:41:24 +0000 (21:41 +0000)]
[Analysis] Change several Analysis pieces to use NodeRef. NFC.
Reviewers: dblaikie, grosser
Subscribers: mzolotukhin, llvm-commits
Differential Revision: https://reviews.llvm.org/D23625
llvm-svn: 279156
Chris Bieneman [Thu, 18 Aug 2016 21:41:21 +0000 (21:41 +0000)]
[CMake] Make llvm-config implicit dependency for subprojects
The subproject interface being used for runtime libraries expects that llvm-config is passed into the subproject for consumption. We currently do this for every subproject, so we should expect that all LLVM ExternalProjects depend on llvm-config for the time being.
Eventually I'd like to see the sub-projects using LLVMConfig.cmake instead of the llvm-config binary, but that will take time to roll out.
llvm-svn: 279155
Saleem Abdulrasool [Thu, 18 Aug 2016 21:40:06 +0000 (21:40 +0000)]
CodeGen: use range based for loop, NFC
llvm-svn: 279154
Chris Bieneman [Thu, 18 Aug 2016 21:36:36 +0000 (21:36 +0000)]
[CMake] Minor fix to regex in r279152
The third version component is optional in Xcode's version spew, so we need to make it optional in the regex.
llvm-svn: 279153
Chris Bieneman [Thu, 18 Aug 2016 21:32:48 +0000 (21:32 +0000)]
[CMake] Support for generating Xcode 8 compatible toolchains
Xcode 8 requires toolchain compatibility version 2. This allows us to select the correct compatibility version based on the installed version of Xcode.
llvm-svn: 279152
Chris Bieneman [Thu, 18 Aug 2016 21:31:51 +0000 (21:31 +0000)]
[CMake] Get libcxx building under LLVM/runtimes
Summary:
The new LLVM runtimes build directory requires some basic conventions across the runtime projects. These changes make libcxx build under the runtimes subdirectory. The general idea of the changes is that the runtimes subdirectory requires some conventions to be consistent across runtime projects.
I expect to have a few more small patches that build on this to tie up check targets and other things useful in development workflows.
Summary of changes in this patch:
* Renamed variable LLVM_CONFIG -> LLVM_CONFIG_PATH
* Renamed variable LIBCXX_BUILT_STANDALONE -> LIBCXX_STANDALONE_BUILD
* Add an include of AddLLVM in the tests subdirectory for add_lit_testsuite.
Reviewers: EricWF
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D23696
llvm-svn: 279151
Sanjay Patel [Thu, 18 Aug 2016 21:28:30 +0000 (21:28 +0000)]
[InstCombine] add helper function for folds of icmp (shl 1, Y), C; NFCI
Clean up the existing code by:
1. Renaming variables
2. Adding local variables
3. Making it vector-safe
This is still guarded by a ConstantInt check, so no functional change is intended.
But this should be ready to go: if we move the ConstantInt check down, all of
these folds should do the right thing for vector types.
llvm-svn: 279150
Jacques Pienaar [Thu, 18 Aug 2016 21:25:17 +0000 (21:25 +0000)]
[lanai] Add ISA document to CompilerWritersInfo
Summary: Add Lanai ISA document to CompilerWritersInfo.
Reviewers: eliben
Subscribers: aemerson, llvm-commits
Differential Revision: https://reviews.llvm.org/D23693
llvm-svn: 279149
Tom Stellard [Thu, 18 Aug 2016 21:21:53 +0000 (21:21 +0000)]
AMDGPU/SI: Fix a test in wqm.ll to always use s_cbranch_vcc*
Summary:
We need to use floating-point compares to ensure that s_cbranch_vcc*
instructions are always generated. With integer compares, future
optimizations could cause s_cbranch_scc* to be generated instead.
Reviewers: arsenm, nhaehnle
Subscribers: llvm-commits, kzhuravl
Differential Revision: https://reviews.llvm.org/D23401
llvm-svn: 279148
Cameron Desrochers [Thu, 18 Aug 2016 20:56:48 +0000 (20:56 +0000)]
Fixed more signed/unsigned mismatch warnings introduced in my change at r279076
llvm-svn: 279145
Kostya Serebryany [Thu, 18 Aug 2016 20:52:52 +0000 (20:52 +0000)]
[libFuzzer] add more __attribute__((visibility("default")))
llvm-svn: 279143
Amaury Sechet [Thu, 18 Aug 2016 20:43:50 +0000 (20:43 +0000)]
Make cltz and cttz zero undef when the operand cannot be zero in InstCombine
Summary: Also add popcount(n) == bitsize(n) -> n == -1 transformation.
Reviewers: majnemer, spatel
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D23134
llvm-svn: 279141
Justin Lebar [Thu, 18 Aug 2016 20:43:13 +0000 (20:43 +0000)]
[CUDA] Improve handling of math functions.
Summary:
A bunch of related changes here to our CUDA math headers.
- The second arg to nexttoward is a double (well, technically, long
double, but we don't have that), not a float.
- Add a forward-declare of llround(float), which is defined in the CUDA
headers. We need this for the same reason we need most of the other
forward-declares: To prevent a constexpr function in our standard
library from becoming host+device.
- Add nexttowardf implementation.
- Pull "foobarf" functions defined by the CUDA headers in the global
namespace into namespace std. This lets you do e.g. std::sinf.
- Add overloads for math functions accepting integer types. This lets
you do e.g. std::sin(0) without having an ambiguity between the
overload that takes a float and the one that takes a double.
With these changes, we pass testcases derived from libc++ for cmath and
math.h. We can check these testcases in to the test-suite once support
for CUDA lands there.
Reviewers: tra
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D23627
llvm-svn: 279140
Sanjay Patel [Thu, 18 Aug 2016 20:28:54 +0000 (20:28 +0000)]
[InstCombine] use m_APInt to allow icmp (trunc X, Y), C folds for splat constant vectors
This is a sibling of:
https://reviews.llvm.org/rL278859
https://reviews.llvm.org/rL278935
https://reviews.llvm.org/rL278945
https://reviews.llvm.org/rL279066
https://reviews.llvm.org/rL279077
https://reviews.llvm.org/rL279101
llvm-svn: 279133
Sanjay Patel [Thu, 18 Aug 2016 20:25:16 +0000 (20:25 +0000)]
[InstCombine] clean up foldICmpTruncConstant(); NFCI
1. Fix variable names
2. Add local variables to reduce code
llvm-svn: 279132
Michael Kuperstein [Thu, 18 Aug 2016 20:08:15 +0000 (20:08 +0000)]
[SelectionDAG] Rename fextend -> fpextend, fround -> fpround, frnd -> fround
The names of the tablegen defs now match the names of the ISD nodes.
This makes the world a slightly saner place, as previously "fround" matched
ISD::FP_ROUND and not ISD::FROUND.
Differential Revision: https://reviews.llvm.org/D23597
llvm-svn: 279129
Yaxun Liu [Thu, 18 Aug 2016 20:01:06 +0000 (20:01 +0000)]
Revert [OpenCL] AMDGCN: Fix size_t type
due to regressions in test/CodeGen/exprs.c on certain platforms.
llvm-svn: 279127
Wei Ding [Thu, 18 Aug 2016 19:51:14 +0000 (19:51 +0000)]
AMDGPU : Fix QSAD and MQSAD instructions' incorrect data type.
Differential Revision: http://reviews.llvm.org/D23689
llvm-svn: 279126
Matthew Simpson [Thu, 18 Aug 2016 19:50:32 +0000 (19:50 +0000)]
[SLP] Initialize VectorizedValue when gathering
We abort building vectorizable trees in some cases (e.g., if the maximum
recursion depth is reached, if the region size is too large, etc.). If this
happens for a reduction, we can be left with a root entry that needs to be
gathered. For these cases, we need make sure we actually set VectorizedValue to
the resulting vector.
This patch ensures we properly set VectorizedValue, and it also ensures the
insertelement sequence generated for the gathers is inserted at the correct
location.
Reference: https://llvm.org/bugs/show_bug.cgi?id=28330
Differential Revison: https://reviews.llvm.org/D23410
llvm-svn: 279125
Matthias Braun [Thu, 18 Aug 2016 19:47:59 +0000 (19:47 +0000)]
RegScavenging: Add scavengeRegisterBackwards()
Re-apply r276044 with off-by-1 instruction fix for the reload placement.
This is a variant of scavengeRegister() that works for
enterBasicBlockEnd()/backward(). The benefit of the backward mode is
that it is not affected by incomplete kill flags.
This patch also changes
PrologEpilogInserter::doScavengeFrameVirtualRegs() to use the register
scavenger in backwards mode.
Differential Revision: http://reviews.llvm.org/D21885
llvm-svn: 279124
Zachary Turner [Thu, 18 Aug 2016 19:42:00 +0000 (19:42 +0000)]
Fix json compilation database syntax on non-Windows.
llvm-svn: 279122
Yaxun Liu [Thu, 18 Aug 2016 19:34:04 +0000 (19:34 +0000)]
[OpenCL] AMDGCN: Fix size_t type
Pointers of certain GPUs in AMDGCN target in private address space is 32 bit but pointers in other address spaces are 64 bit. size_t type should be defined as 64 bit for these GPUs so that it could hold pointers in all address spaces. Also fixed issues in pointer arithmetic codegen by using pointer specific intptr type.
Differential Revision: https://reviews.llvm.org/D23361
llvm-svn: 279121
Zachary Turner [Thu, 18 Aug 2016 19:31:48 +0000 (19:31 +0000)]
Resubmit "[Tooling] Parse compilation database command lines on Windows."
This patch introduced the ability to decide at runtime whether to parse
JSON compilation database command lines using Gnu syntax or Windows
syntax. However, there were many existing unit tests written that
hardcoded Gnu-specific paths. These tests were now failing because
the auto-detection logic was choosing to parse them using Windows
rules.
This resubmission of the patch fixes this by introducing an enum
which defines the syntax mode, which defaults to auto-detect, but
for which the unit tests force Gnu style parsing.
Reviewed By: alexfh
Differential Revision: https://reviews.llvm.org/D23628
llvm-svn: 279120
Simon Atanasyan [Thu, 18 Aug 2016 19:08:41 +0000 (19:08 +0000)]
[ELF][MIPS] Support R_MIPS_GOT_HI16 / R_MIPS_GOT_LO16 relocations calculation
llvm-svn: 279119
Simon Atanasyan [Thu, 18 Aug 2016 19:08:36 +0000 (19:08 +0000)]
[ELF][MIPS] Support R_MIPS_CALL_HI16 / R_MIPS_CALL_LO16 relocations calculation
llvm-svn: 279118
Kyle Butt [Thu, 18 Aug 2016 18:57:29 +0000 (18:57 +0000)]
Branch Folding: Accept explicit threshold for tail merge size.
This is prep work for allowing the threshold to be different during layout,
and to enforce a single threshold between merging and duplicating during
layout. No observable change intended.
llvm-svn: 279117
Reid Kleckner [Thu, 18 Aug 2016 18:45:07 +0000 (18:45 +0000)]
[MS] Silence -Wextern-init on const selectany variables
In C, 'extern' is typically used to avoid tentative definitions when
declaring variables in headers, but adding an intializer makes it a
defintion. This is somewhat confusing, so GCC and Clang both warn on it.
In C++, 'extern' is often used to give implictly static 'const'
variables external linkage, so don't warn in that case. If selectany is
present, this might be header code intended for C and C++ inclusion, so
apply the C++ rules.
llvm-svn: 279116