Volkan Keles [Mon, 18 May 2020 19:11:46 +0000 (12:11 -0700)]
LoadStoreVectorizer: Match nested adds to prove vectorization is safe
If both OpA and OpB is an add with NSW/NUW and with the same LHS operand,
we can guarantee that the transformation is safe if we can prove that OpA
won't overflow when IdxDiff added to the RHS of OpA.
Review: https://reviews.llvm.org/D79817
Nikita Popov [Mon, 18 May 2020 17:09:03 +0000 (19:09 +0200)]
[Loads] Require Align in isSafeToLoadUnconditionally() (NFC)
Now that load/store have required alignment, accept Align here.
This also avoids uses of getPointerElementType(), which is
incompatible with opaque pointers.
George [Mon, 18 May 2020 18:40:32 +0000 (14:40 -0400)]
Fixed a typo in the comment for allocateBuffer()
Differential Revision: https://reviews.llvm.org/D80087
Jonas Devlieghere [Mon, 18 May 2020 18:34:09 +0000 (11:34 -0700)]
[dsymutil] Make sure the --help output and man page are consistent
As suggested by Adrian in D79398.
Arthur Eubanks [Mon, 18 May 2020 16:02:12 +0000 (09:02 -0700)]
Add verifier check that musttail and preallocated are not used together
Summary:
Currently they are not supported together. Supporting them will require
a LangRef change. See discussion in https://reviews.llvm.org/D77689.
Reviewers: rnk, efriedma
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D80132
Jay Foad [Mon, 18 May 2020 15:06:42 +0000 (16:06 +0100)]
[IR] Revert r2694 in BasicBlock::removePredecessor
r2694 fixed a bug where removePredecessor could create IR with a use not
dominated by its def in a self loop. But this could only happen in an
unreachable loop, and since that time the rules have been relaxed so
that defs don't have to dominate uses in unreachable code, so the fix is
unnecessary. The regression test added in r2691 still stands.
Differential Revision: https://reviews.llvm.org/D80128
Jonas Paulsson [Mon, 18 May 2020 16:20:40 +0000 (18:20 +0200)]
[SystemZ] Don't create PERMUTE nodes with an undef operand.
It's better to reuse the first source value than to use an undef second
operand, because that will make more resulting VPERMs have identical operands
and therefore MachineCSE more successful.
Review: Ulrich Weigand
Mircea Trofin [Mon, 18 May 2020 17:26:45 +0000 (10:26 -0700)]
[llvm][NFC] Fixed non-compliant style in InlineAdvisor.h
Changed OnPass{Entry|Exit} -> onPass{Entry|Exit}
Also fixed a small typo in a comment.
Jonas Devlieghere [Mon, 18 May 2020 16:55:33 +0000 (09:55 -0700)]
[lldb/Reproducers] Add skipIfReproducer to more tests
Mark more tests as unsupported with reproducers.
Vedant Kumar [Fri, 15 May 2020 20:58:40 +0000 (13:58 -0700)]
[Local] Do not ignore zexts in salvageDebugInfo, PR45923
Summary:
When salvaging a dead zext instruction, append a convert operation to
the DIExpressions of the debug uses of the instruction, to prevent the
salvaged value from being sign-extended.
I confirmed that lldb prints out the correct unsigned result for "f" in
the example from PR45923 with this changed applied.
rdar://
63246143
Reviewers: aprantl, jmorse, chrisjackson, davide
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D80034
Alexandre Rames [Mon, 18 May 2020 16:44:26 +0000 (18:44 +0200)]
[MLIR] Use `MLIR_INCLUDE_TESTS` to conditionally compile tests.
This is equivalent to what is done for other projects (e.g. clang).
Differential Revision: https://reviews.llvm.org/D80022
Tobias Gysi [Mon, 18 May 2020 16:30:39 +0000 (18:30 +0200)]
[mlir] Support optional attributes in assembly formats
Summary: This revision adds support for assembly formats with optional attributes. It elides optional attributes that are part of the syntax from the attribute dictionary.
Reviewers: ftynse, Kayjukh
Reviewed By: ftynse, Kayjukh
Subscribers: mehdi_amini, rriddle, jpienaar, shauheen, antiagainst, nicolasvasilache, arpith-jacob, mgester, lucyrfox, liufengdb, stephenneuendorffer, Joonsoo, grosul1, frgossen, jurahul, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D80113
Pierre Oechsel [Mon, 18 May 2020 16:25:23 +0000 (18:25 +0200)]
[MLIR] [Linalg] Add option to use the partial view after promotion.
For now the promoted buffer is indexed using the `full view`. The full view might be
slightly bigger than the partial view (which is accounting for boundaries).
Unfortunately this does not compose easily with other transformations when multiple buffers
with shapes related to each other are involved.
Take `linalg.matmul A B C` (with A of size MxK, B of size KxN and C of size MxN) and suppose we are:
- Tiling over M by 100
- Promoting A only
This is producing a `linalg.matmul promoted_A B subview_C` where `promoted_A` is a promoted buffer
of `A` of size (100xK) and `subview_C` is a subview of size mxK where m could be smaller than 100 due
to boundaries thus leading to a possible incorrect behavior.
We propose to:
- Add a new parameter to the tiling promotion allowing to enable the use of the full tile buffer.
- By default all promoted buffers will be indexed by the partial view.
Note that this could be considered as a breaking change in comparison to the way the tiling promotion
was working.
Differential Revision: https://reviews.llvm.org/D79927
Matt Arsenault [Sun, 17 May 2020 23:06:10 +0000 (19:06 -0400)]
AMDGPU/GlobalISel: Fix splitting wide VALU, non-vector loads
Matt Arsenault [Sun, 17 May 2020 16:37:27 +0000 (12:37 -0400)]
AMDGPU: Remove outdated comment
Nicolas Vasilache [Mon, 18 May 2020 15:51:56 +0000 (11:51 -0400)]
[mlir][Vector] Add an optional "masked" boolean array attribute to vector transfer operations
Summary:
Vector transfer ops semantic is extended to allow specifying a per-dimension `masked`
attribute. When the attribute is false on a particular dimension, lowering to LLVM emits
unmasked load and store operations.
Differential Revision: https://reviews.llvm.org/D80098
Nicolas Vasilache [Mon, 18 May 2020 15:40:59 +0000 (11:40 -0400)]
[mlir][Vector] Make minor identity permutation map optional in transfer op printing and parsing
Summary:
This revision makes the use of vector transfer operatons more idiomatic by
allowing to omit and inferring the permutation_map.
Differential Revision: https://reviews.llvm.org/D80092
David Sherwood [Wed, 6 May 2020 10:23:04 +0000 (11:23 +0100)]
[SVE] Ignore scalable vectors in InterleavedLoadCombinePass
I have changed the pass so that we ignore shuffle vectors with
scalable vector types, and replaced VectorType with FixedVectorType
in the rest of the pass. I couldn't think of an easy way to test
this change, since for scalable vectors we shouldn't be using
shufflevectors for interleaving. This change fixes up some
type size assert warnings I found in the following test:
CodeGen/AArch64/sve-intrinsics-int-arith-imm.ll
Differential Revision: https://reviews.llvm.org/D79700
Wouter van Oortmerssen [Fri, 15 May 2020 23:09:20 +0000 (16:09 -0700)]
[WebAssembly] iterate stack in DebugFixup from the top.
Differential Revision: https://reviews.llvm.org/D80045
Balázs Kéri [Mon, 18 May 2020 15:14:39 +0000 (17:14 +0200)]
[Analyzer][StreamChecker] Fixed compile error - NFC.
Balázs Kéri [Mon, 18 May 2020 13:07:32 +0000 (15:07 +0200)]
[Analyzer][StreamChecker] Changed representation of stream error state - NFC.
Summary:
State of error flags for a stream is handled by having separate flags
that allow combination of multiple error states to be described with one
error state object.
After a failed function the error state is set in the stream state
and must not be determined later based on the last failed function
like before this change. The error state can not always be determined
from the last failed function and it was not the best design.
Reviewers: Szelethus
Reviewed By: Szelethus
Subscribers: xazax.hun, baloghadamsoftware, szepet, a.sidorin, mikhail.ramalho, Szelethus, donat.nagy, dkrupp, gamesh411, Charusso, martong, ASDenysPetrov, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D80009
Dvorskiy, Mikhail [Tue, 10 Mar 2020 16:09:57 +0000 (19:09 +0300)]
[pstl] A fix for move placement-new (and destroy) allocated objects from raw memory.
https://reviews.llvm.org/D74123
The fix affects follow algorithms:
remove_if, unique, rotate, inplace_merge, partial_sort_copy, set_union, set_intersection, set_difference, set_symmetric_difference.
For "is_trivial" types there are no problems with "creating objects/clean-up"
For non-trivial types the algo results are also correct, but possible incorrect copying/moving "operator=" calls "by raw memory" within one of mentioned algo or incorrect destructor calls in the end of algo.
Nicolas Vasilache [Mon, 18 May 2020 13:41:59 +0000 (09:41 -0400)]
[mlir] Add BoolArrayAttr in Tablegen + Builder support
Differential Revision: https://reviews.llvm.org/D80090
Nicolas Vasilache [Mon, 18 May 2020 13:33:35 +0000 (09:33 -0400)]
[mlir] NFC - Fix OperationSupport.cpp::findNamedAttr
Summary:
When NamedAttrList::get is called against a StringRef and the
entry is not present, the Identifier::operator== crashes.
Interestingly there seems to be no use of the NamedAttrList::get(StringRef) in
the codebase so far. A subsequent commit will introduce such a use.
Differential Revision: https://reviews.llvm.org/D80089
AndreyChurbanov [Mon, 18 May 2020 12:32:13 +0000 (15:32 +0300)]
[openmp] Fixed hang if detached task was serialized.
The patch fixes https://bugs.llvm.org/show_bug.cgi?id=45904.
Differential Revision: https://reviews.llvm.org/D79944
serge-sans-paille [Tue, 12 May 2020 16:46:32 +0000 (18:46 +0200)]
Reorder Type fields to make various isa< > check more concise
Depending on the order of fields, some isa < > checks can be faster because of
tests that check a range of type, leading to assembly simplification.
To find a relevant ordering, I... brute-forced the permutation among the derived
types and pick the combination that resulted in the smallest libLLVM-11.so.
On my laptop (x86_64), this reduces the size of libLLVM-11.so from
127344064 bytes to
127335336,
that's 8728 bytes shaved without much effort.
Also removed obsolete comments in the process.
Differential Revision: https://reviews.llvm.org/D79996
Michael Schellenberger Costa [Mon, 18 May 2020 12:01:55 +0000 (14:01 +0200)]
[libcxx][type_traits] Add C++20 changes to common_type
Summary: This already implements the expected changes for LWG-3205
Reviewers: mclow.lists, EricWF, CaseyCarter, cjdb, #libc, ldionne
Reviewed By: #libc, ldionne
Subscribers: dexonsmith, broadwaylamb, christof, ldionne, libcxx-commits
Tags: #libc
Differential Revision: https://reviews.llvm.org/D74350
David Zarzycki [Mon, 18 May 2020 11:39:12 +0000 (07:39 -0400)]
[libcxx testing] Remove ALLOW_RETRIES from lock_guard tests
These two tests were clumsily using time measurements to determine
whether std::lock_guard was working correctly. In practice, this
approach merely verified that the underlying lock properly waits.
Now these two tests verify that lock is acquired, not dropped
prematurely, and finally, actually dropped at the end of the scope.
Hans Wennborg [Fri, 15 May 2020 15:20:27 +0000 (17:20 +0200)]
Turn -Wmax-tokens off by default
On the one hand, one might interpret the use of the max-token pragmas or
-fmax-tokens flag as an opt-in to the warning. However, in Chromium
we've found it useful to only opt in selected build configurations, even
though we have the pragmas in the code. For that reason, we think it
makes sense to turn it off by default.
Differential revision: https://reviews.llvm.org/D80014
Max Kazantsev [Mon, 18 May 2020 11:09:24 +0000 (18:09 +0700)]
[InstCombine][NFC] Simplify check in sinking
We just need to check that the only predecessor of user parent is
BB, we don't need to iterate through BB's successors for it.
Anastasia Stulova [Mon, 18 May 2020 10:02:01 +0000 (11:02 +0100)]
[OpenCL] Added addrspace_cast operator in C++ mode.
This operator is intended for casting between
pointers to objects in different address spaces
and follows similar logic as const_cast in C++.
Tags: #clang
Differential Revision: https://reviews.llvm.org/D60193
Dmitry Preobrazhensky [Mon, 18 May 2020 11:04:58 +0000 (14:04 +0300)]
[AMDGPU][MC] Corrected branch relocation handling to detect undefined labels
Fixed ELF object writer to die gracefully when an undefined label is encountered in a branch instruction.
See https://bugs.llvm.org/show_bug.cgi?id=41914.
Reviewers: arsenm, rampitec
Differential Revision: https://reviews.llvm.org/D79943
Kirill Bobyrev [Mon, 18 May 2020 10:20:11 +0000 (12:20 +0200)]
[clang-tidy] NFC: Cleanup Python scripts
Summary: Silence few PEP8 warnings.
Reviewers: hokein
Reviewed By: hokein
Subscribers: Eugene.Zelenko, xazax.hun, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D78985
Max Kazantsev [Mon, 18 May 2020 09:25:16 +0000 (16:25 +0700)]
[Test] Opportunity for sinking to unreachable in InstCombine
Hans Wennborg [Mon, 18 May 2020 09:20:42 +0000 (11:20 +0200)]
Revert
76c5f277f2 "Re-land [Debug][CodeView] Emit fully qualified names for globals"
> Before this patch, S_[L|G][THREAD32|DATA32] records were emitted with a simple name, not the fully qualified name (namespace + class scope).
>
> Differential Revision: https://reviews.llvm.org/D79447
This causes asserts in Chromium builds:
CodeViewDebug.cpp:2997: void llvm::CodeViewDebug::emitDebugInfoForUDTs(const std::vector<std::pair<std::string, const DIType *>> &):
Assertion `OriginalSize == UDTs.size()' failed.
I will follow up on the Phabricator issue.
Nathan James [Mon, 18 May 2020 09:11:22 +0000 (10:11 +0100)]
[NFC][clang-tidy] use hasOperands in place of hasEitherOperand
OCHyams [Mon, 18 May 2020 08:25:29 +0000 (09:25 +0100)]
[DebugInfo][DWARF] Emit a single location instead of a location list
for variables in nested scopes (including inlined functions) if there is a
single location which covers the entire scope and the scope is contained in a
single block.
Based on work by @jmorse.
Reviewed By: vsk, aprantl
Differential Revision: https://reviews.llvm.org/D79571
Martin Storsjö [Mon, 18 May 2020 08:04:42 +0000 (11:04 +0300)]
[compiler-rt] Don't error out on missing builtins when builting builtins themselves
This fixes bootstrapping the builtins when no previous version of
them exists after
2fe66bdb2e5fb556e15d4a07158e89ec007c587b.
Also fix a whitespace issue in that commit.
Clement Courbet [Mon, 27 Apr 2020 08:45:41 +0000 (10:45 +0200)]
[TargetLibraryInfo] memcpy and memset have optimized codegen.
Summary:
Add memcpy and memset to the list of function with optimized codegen.
hasOptimizedCodeGen() was returning false for them.
This was not an issue when working with c++ as clang typically emits an
intrinsic rather than a libfunc call, but when working with IR,
this would cause some issues with sanitizers which would not correctly
mark the function as nobuiltin in `maybeMarkSanitizerLibraryCallNoBuiltin()`.
Subscribers: llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D78991
Jim Lin [Mon, 18 May 2020 03:10:46 +0000 (11:10 +0800)]
[RISCV] Fix passing two floating-point values in complex separately by two GPRs on RV64
Summary:
This patch fixed the error of counting the remaining FPRs. Complex floating-point
values should be passed by two FPRs for the hard-float ABI. If no two FPRs are
available, it should be passed via a 64-bit GPR (fp+fp). `ArgFPRsLeft` is only
decreased one while the type is complex floating-point. It causes two floating-point
values in the complex are passed separately by two GPRs.
Reviewers: asb, luismarques, lenary
Reviewed By: asb
Subscribers: rbar, johnrusso, simoncook, sabuasal, niosHD, kito-cheng, shiva0217, jrtc27, zzheng, edward-jones, rogfer01, MartinMosbeck, brucehoult, the_o, rkruppe, PkmX, jocewei, psnobl, benna, s.egerton, pzheng, sameer.abuasal, apazos, evandro, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D79770
Alexander Shaposhnikov [Mon, 18 May 2020 03:46:17 +0000 (20:46 -0700)]
[llvm-objcopy][MachO] Fix code formatting
Apply clang-format to MachOObjcopy.cpp. NFC.
Test plan: make check-all
Chen Zheng [Mon, 18 May 2020 01:00:09 +0000 (21:00 -0400)]
[PowerPC] add more high latency opcodes for machinecombiner - NFC
Stephen Neuendorffer [Sun, 17 May 2020 22:17:04 +0000 (15:17 -0700)]
[MLIR][cmake] Fix unittest/Tablegen by special casing LLVMMLIRTableGen
unittest/Tablegen generates an executable that depends on MLIRIR and
LLVMMLIRTableGen. Avoid specifying linkage dependence on LLVM
libraries here because then everyone has to depend on those libraries.
Differential Revision: https://reviews.llvm.org/D80093
Mehdi Amini [Sun, 17 May 2020 23:16:21 +0000 (23:16 +0000)]
Fix warning "defined but not used" for debug function (NFC)
Mehdi Amini [Sun, 17 May 2020 19:25:35 +0000 (19:25 +0000)]
Replace dyn_cast with isa when the result isn't used (NFC)
Fix build warning: unused variable 'BB'
Mehdi Amini [Sun, 17 May 2020 19:24:33 +0000 (19:24 +0000)]
Remove spurious semicolon after function definition (NFC)
This fixes some GCC pedantic warnings.
Craig Topper [Sun, 17 May 2020 21:57:04 +0000 (14:57 -0700)]
ValueMapper does not preserve inline assembly dialect when remapping the type
Bug report: https://bugs.llvm.org/show_bug.cgi?id=45291
Patch by Tomasz Miąsko
Differential Revision: https://reviews.llvm.org/D80066
Stephen Neuendorffer [Sun, 17 May 2020 06:01:52 +0000 (23:01 -0700)]
[MLIR] Fix linkage for libMLIR.so
Generally:
1) don't use target_link_libraries() and add_mlir_library() on the same target, use LINK_LIBS PUBLIC instead.
2) don't use LINK_LIBS to specify LLVM libraries. Use LINK_COMPONENTS instead
3) no need to link against LLVMSupport. We pull it in by default.
Differential Revision: https://reviews.llvm.org/D80076
Stephen Neuendorffer [Sun, 17 May 2020 05:46:57 +0000 (22:46 -0700)]
[MLIR][cmake][NFC] Update linkage checker for mlir-opt
New CMakeLists.txt for MLIRStandardOpsTransforms was incorrect, but wasn't
caught by the check.
Differential Revision: https://reviews.llvm.org/D80075
Nikita Popov [Sun, 17 May 2020 20:14:42 +0000 (22:14 +0200)]
[Alignment] Remove unnecessary getValueOrABITypeAlignment calls (NFC)
Now that load/store alignment is required, we no longer need most
of them. Also switch the getLoadStoreAlignment() helper to return
Align instead of MaybeAlign.
Roman Lebedev [Sun, 17 May 2020 18:58:55 +0000 (21:58 +0300)]
[InstCombine] visitMaskedMerge(): when unfolding, sanitize undef constants (PR45955)
We can't leave undef vector element constants as-is,
it is a miscompile, so we need to sanitize them.
We have two vectors (C and ~C):
* We can't replace undef with 0 in both of them
* We can't replace undef with 0 in only one of them
* We could replace undef with -1 in both of them
* We could replace undef with -1 in only one(!) of them
* We could replace undef with -1 in one and 0 in another one of them.
Therefore, it seems best to go with the last option, since otherwise
we'd loose knowledge that C and ~C have no common bits set,
which seems more important than preserving partial undef knowledge.
Fixes https://bugs.llvm.org/show_bug.cgi?id=45955
David Blaikie [Sun, 17 May 2020 19:17:31 +0000 (12:17 -0700)]
DebugInfo: Reduce long-distance dependence on what will/won't emit a debug_addr section
This is a no-op/NFC at the moment & generally makes the code /somewhat/
cleaner/less reliant on assumptions about what will produce a debug_addr
section.
It's still a bit "spooky action at a distance" - the add ranges code
pre-emptively inserts addresses into the address pool it knows will
eventually be used by the range emission code (or low/high pc).
The 'ideal' would be either to actually compute the addresses needed for
range (& loc) emission earlier - which would mean decanonicalizing the
range/loc representation earlier to account for whether it was going to
use addrx encodings or not (which would be unfortunate, but could be
refactored to be relatively unobtrusive).
Alternatively, emitting the range/loc sections earlier would cause them
to request the needed addresses sooner - but then you endup having to
split finalizeModuleInfo because some things need to be handled there
before the ranges/locs are emitted, I think...
Nikita Popov [Wed, 25 Mar 2020 19:51:01 +0000 (20:51 +0100)]
[LVI] Don't use dominator tree in isValidAssumeForContext()
LVI and its consumers currently have quite a bit of complexity
related to dominator tree management. However, it doesn't look
like it is actually needed...
The only use of the dominator tree is inside isValidAssumeForContext().
However, due to the way LVI queries work, it is not needed:
If we query a value for some block, we will first get the edge values
from all predecessor blocks, which also includes an intersection with
assumptions that apply to the terminator of the predecessor. As such,
we will already have processed all assumptions from predecessor blocks
(this is actually stronger than what isValidAssumeForContext() does
with a DT, because this is capable of combining non-dominating
assumptions). The only additional assumptions we need to take into
account are those in the block being queried. And we don't need a
dominator tree for that.
This patch only removes the use of DT, I will drop the machinery
around it in a followup.
Differential Revision: https://reviews.llvm.org/D76797
Fedor Sergeev [Sun, 17 May 2020 19:17:56 +0000 (02:17 +0700)]
Add missing include Host.h in llvm-mc-* fuzzers. NFC.
Fixes build failure in these fuzzers.
Nathan James [Sat, 16 May 2020 01:00:33 +0000 (02:00 +0100)]
[ASTMatchers] Added BinaryOperator hasOperands matcher
Summary: Adds a matcher called `hasOperands` for `BinaryOperator`'s when you need to match both sides but the order isn't important, usually on commutative operators.
Reviewers: klimek, aaron.ballman, gribozavr2, alexfh
Reviewed By: aaron.ballman
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D80054
JinGu Kang [Sun, 17 May 2020 18:49:37 +0000 (19:49 +0100)]
test commit
Simon Pilgrim [Sun, 17 May 2020 17:51:21 +0000 (18:51 +0100)]
Revert rGca18ce1a00cd8b7cb7ce0e130440f5ae1ffe86ee "GlobPattern.h - remove unnecessary BitVector.h/StringRef.h includes. NFC"
Causes lld build errors
Simon Pilgrim [Sun, 17 May 2020 17:29:27 +0000 (18:29 +0100)]
GlobPattern.h - remove unnecessary BitVector.h/StringRef.h includes. NFC
Use forward declarations (BitVector already had one) and an headers to source file that were implicitly using them.
Simon Pilgrim [Sun, 17 May 2020 17:03:15 +0000 (18:03 +0100)]
ImmutableGraph.h - remove unused raw_ostream.h include. NFC
Fangrui Song [Sun, 17 May 2020 16:31:09 +0000 (09:31 -0700)]
[XRay] Migrate xray_naive_log=true tests to xray_mode=xray-basic
Sanjay Patel [Sun, 17 May 2020 15:56:26 +0000 (11:56 -0400)]
[x86] favor vector constant load to avoid GPR to XMM transfer
This build vector lowering pattern came up in D79886.
I've tried to limit the improvement to cases where it looks
clearly better to load, but we could remove the 'TODO'
predicates already if we are willing to overlook some
corner cases.
Differential Revision: https://reviews.llvm.org/D80013
Sanjay Patel [Sun, 17 May 2020 15:40:27 +0000 (11:40 -0400)]
[InstCombine] add tests for FP cast of cast; NFC
A fold of casts is proposed as a backend transform in D79187,
but we can also do that in IR (and that may obsolete the need
for a backend transform).
Xing GUO [Sun, 17 May 2020 14:48:40 +0000 (22:48 +0800)]
[ObjectYAML][DWARF] Take into account other debug sections in DWARFYAML::Data::isEmpty().
Dylan McKay [Sun, 17 May 2020 14:18:54 +0000 (02:18 +1200)]
[AVR] Explicitly set the address of the data section when invoking the linker
This is required to get avr-gdb correctly showing values at the right
addresses. This problem was discovered by using debug symbols in an
external program to lookup values in an AVR simulator.
Nicolas Vasilache [Sun, 17 May 2020 14:15:58 +0000 (10:15 -0400)]
[mlir] NFC - VectorTransforms use OpBuilder where relevant
Summary: This will allow using unrolling outside of only rewrite patterns.
Differential Revision: https://reviews.llvm.org/D80083
Simon Pilgrim [Sun, 17 May 2020 13:58:17 +0000 (14:58 +0100)]
[X86] Add getTargetConstantFromBasePtr helper. NFC.
Allows us to share code from LoadSDNode and MemIntrinsicSDNode constant pool loads.
Simon Pilgrim [Sun, 17 May 2020 13:39:51 +0000 (14:39 +0100)]
[X86] getTargetConstantBitsFromNode - remove unnecessary X86ISD::VBROADCAST handling.
We create X86ISD::VBROADCAST_LOAD for constant pool folds now.
Sanjay Patel [Sun, 17 May 2020 12:43:40 +0000 (08:43 -0400)]
[InstCombine] improve analysis of FP->int->FP to eliminate fpextend
This was originally in D79116.
Converting from a narrow-enough FP source value to integer and
back to FP guarantees that the conversion to FP is exact because
of UB/poison-on-overflow.
This was suggested in PR36617:
https://bugs.llvm.org/show_bug.cgi?id=36617#c19
Florian Hahn [Sun, 17 May 2020 12:55:11 +0000 (13:55 +0100)]
[LoopUnroll] Extend test case with additional loop with larger TC.
Florian Hahn [Sun, 17 May 2020 12:26:51 +0000 (13:26 +0100)]
[LoopUnroll] Precommit test for PR459393.
Christudasan Devadasan [Tue, 21 Apr 2020 09:34:33 +0000 (15:04 +0530)]
[AMDGPU] Enable base pointer.
When the callee requires a dynamic stack realignment,
it is not possible to correcty access the incoming
stack arguments using the stack pointer. We reserve a
base pointer in such cases to access the function arguments
inside the callee. The base pointer will hold the incoming
stack pointer value before any kind of delta added to it.
Reviewed By: arsenm, scott.linder
Differential Revision: https://reviews.llvm.org/D78811
Joachim Protze [Sun, 17 May 2020 10:25:02 +0000 (12:25 +0200)]
[OpenMP] Fix race condition in the completion/freeing of detached tasks
Spurious assertion failures are symptoms of a race condition for the handling
of detached tasks:
Assertion failure at kmp_tasking.cpp(3744): taskdata->td_flags.complete == 1.
Assertion failure at kmp_tasking.cpp(710): taskdata->td_flags.executing == 0.
in the case of detach=true, all accesses to taskdata in __kmp_task_finish need
to happen before (~line 873):
taskdata->td_flags.proxy = TASK_PROXY;
This assignment signals to __kmp_fulfill_event, that the task will need to be
freed there. So, conceptionally the ownership of taskdata is moved.
Reviewed By: AndreyChurbanov
Differential Revision: https://reviews.llvm.org/D79702
Fedor Sergeev [Sun, 17 May 2020 09:31:33 +0000 (16:31 +0700)]
[Inliner][NFC] silence gcc 'overloaded-virtual' warning on hiding of Pass::doInitialization
When compiling with -Werror=overloaded-virtual, gcc emits this:
====
llvm/include/llvm/Pass.h:102:16: error: ‘virtual bool llvm::Pass::doInitialization(llvm::Module&)’ was hidden [-Werror=overloaded-virtual]
virtual bool doInitialization(Module &) { return false; }
^~~~~~~~~~~~~~~~
In file included from llvm/lib/Transforms/IPO/Inliner.cpp:20:0:
llvm/include/llvm/Transforms/IPO/Inliner.h:38:8: error: by ‘virtual bool llvm::LegacyInlinerBase::doInitialization(llvm::CallGraph&)’ [-Werror=overloaded-virtual]
bool doInitialization(CallGraph &CG) override;
^~~~~~~~~~~~~~~~
====
This is an old issue which has just started biting our downstream after
a slight rearrangement of includes around Inliner.
Fixing it similar to how doFinalization was done years ago.
Dylan McKay [Sun, 17 May 2020 04:46:32 +0000 (16:46 +1200)]
[LLVM][AVR] Support for R_AVR_6 fixup
Summary: Handle the emission of `R_AVR_6` ELF relocation type.
Reviewers: dylanmckay
Reviewed By: dylanmckay
Subscribers: hiraditya, Jim, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D78721
Patch by @LemonBoy https://reviews.llvm.org/p/LemonBoy/
Dylan McKay [Sun, 17 May 2020 04:25:12 +0000 (16:25 +1200)]
[AVR] Fix I/O instructions on XMEGA
Summary:
On XMEGA, I/O address space is same as data address space - there is no 0x20 offset,
because CPU General Purpose Registers are not mapped in data address space.
From https://en.wikipedia.org/wiki/AVR_microcontrollers
> In the XMEGA variant, the working register file is not mapped into the data address space; as such, it is not possible to treat any of the XMEGA's working registers as though they were SRAM. Instead, the I/O registers are mapped into the data address space starting at the very beginning of the address space.
Reviewers: dylanmckay
Reviewed By: dylanmckay
Subscribers: hiraditya, Jim, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77207
Patch by Vlastimil Labsky.
Fangrui Song [Sun, 17 May 2020 06:32:28 +0000 (23:32 -0700)]
[Driver] Render -T for Gnu.cpp
clang -T a.lds a.c currently does not render -T.
Stephen Neuendorffer [Sun, 17 May 2020 05:42:32 +0000 (22:42 -0700)]
[MLIR][cmake] use LINK_LIBS PUBLIC for MLIRStandardOpsTransforms
Without this LLVM_LINK_LLVM_DYLIB is broken
Differential Revision: https://reviews.llvm.org/D80074
Fangrui Song [Sun, 17 May 2020 05:42:52 +0000 (22:42 -0700)]
[llvm-xray] consumeError when trying big-endian
Follow-up of rL341226.
Fixes "Expected<T> must be checked before access or destruction"
Arthur Eubanks [Fri, 15 May 2020 23:31:17 +0000 (16:31 -0700)]
[NFC] Run clang-format on ISDOpcodes.h
Subscribers: jfb, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D80050
Yi Kong [Sat, 16 May 2020 03:59:21 +0000 (11:59 +0800)]
[Compiler-rt] Emit error if builtins library cannot be found
Since setting COMPILER_RT_USE_BUILTINS_LIBRARY would remove -z,defs
flag, missing builtins library would continue to build unnoticed.
Explicitly emit an error in such case.
Differential Revision: https://reviews.llvm.org/D79470
Nico Weber [Sun, 17 May 2020 00:38:28 +0000 (20:38 -0400)]
Fix a few doc typos to cycle bots.
Nico Weber [Sun, 17 May 2020 00:32:58 +0000 (20:32 -0400)]
Try to heal bots after https://reviews.llvm.org/D79655
Craig Topper [Sat, 16 May 2020 22:48:08 +0000 (15:48 -0700)]
[LegalizeDAG] Use MachinePointerInfo::getUnknownStack in place of MachinePointerInfo() in a couple places. NFC
We know the pointer somewhere on the stack, we just don't know
exactly where since the index may be variable.
Differential Revision: https://reviews.llvm.org/D80060
Eli Friedman [Fri, 15 May 2020 20:23:14 +0000 (13:23 -0700)]
AllocaInst should store Align instead of MaybeAlign.
Along the lines of D77454 and D79968. Unlike loads and stores, the
default alignment is getPrefTypeAlign, to match the existing handling in
various places, including SelectionDAG and InstCombine.
Differential Revision: https://reviews.llvm.org/D80044
Craig Topper [Sat, 16 May 2020 19:38:34 +0000 (12:38 -0700)]
[X86] Replace selectScalarSSELoad ComplexPattern with PatFrags to handle the 3 types of loads we currently match.
This ensures we create mem operands for these instructions fixing PR45949.
Unfortunately, it increases the size of X86GenDAGISel.inc, but some dag
combine canonicalization could reduce the types of load we need to match.
Eli Friedman [Sat, 16 May 2020 21:01:54 +0000 (14:01 -0700)]
Harden IR and bitcode parsers against infinite size types.
If isSized is passed a SmallPtrSet, it uses that set to catch infinitely
recursive types (for example, a struct that has itself as a member).
Otherwise, it just crashes on such types.
faisal vali [Sat, 16 May 2020 20:12:04 +0000 (15:12 -0500)]
Revert "[nfc] test commit"
This reverts commit
0ee46e857d81ea815e5b11d266c0c118a2c2e714.
faisal vali [Sat, 16 May 2020 20:08:30 +0000 (15:08 -0500)]
[nfc] test commit
John McCall [Sat, 16 May 2020 18:44:54 +0000 (14:44 -0400)]
Expose IRGen API to add the default IR attributes to a function definition.
I've also made a stab at imposing some more order on where and how we add
attributes; this part should be NFC. I wasn't sure whether the CUDA use
case for libdevice should propagate CPU/features attributes, so there's a
bit of unnecessary duplication.
mydeveloperday [Sat, 16 May 2020 17:48:13 +0000 (18:48 +0100)]
The release notes for ObjCBreakBeforeNestedBlockParam was placed between the release note for IndentCaseBlocks and its example code
Remove other whitespace and line limit issues and double blank line issues
Sanjay Patel [Sat, 16 May 2020 17:08:01 +0000 (13:08 -0400)]
[VectorCombine] forward walk through instructions to improve chaining of transforms
This is split off from D79799 - where I was proposing to fully iterate
over a function until there are no more transforms. I suspect we are
still going to want to do something like that eventually.
But we can achieve the same gains much more efficiently on the current
set of regression tests just by reversing the order that we visit the
instructions.
This may also reduce the motivation for D79078, but we are still not
getting the optimal pattern for a reduction.
Sanjay Patel [Sat, 16 May 2020 16:50:16 +0000 (12:50 -0400)]
[PhaseOrdering] add vector reduction tests; NFC
These are based on tests originally included in:
D79078
Nikita Popov [Sat, 16 May 2020 16:46:03 +0000 (18:46 +0200)]
[InstCombine] Clean up alignment handling (NFC)
Now that load/store alignment is required, we can simplify code
in some places.
David Green [Sat, 16 May 2020 14:27:20 +0000 (15:27 +0100)]
[ARM] Patterns for VQSHRN
Given a VQMOVN(VSHR), we can fold that into a VQSHRN simply enough using
a few tablegen patterns.
Differential Revision: https://reviews.llvm.org/D77720
Sanjay Patel [Sat, 16 May 2020 16:44:00 +0000 (12:44 -0400)]
[VectorCombine] add reduction-like patterns; NFC
These are based on tests originally included in:
D79078
Jay Foad [Sat, 16 May 2020 15:00:02 +0000 (16:00 +0100)]
[AArch64] Precommit tests for D77316
Sanjay Patel [Fri, 15 May 2020 19:22:30 +0000 (15:22 -0400)]
[x86][CGP] try to hoist funnel shift above select-of-splats
This is basically the same patch as D63233, but converted to
funnel shifts rather than regular shifts. I did not see a
way to effectively share code for these 2 cases though.
This follows D79718 and D79827 to re-fix PR37426 because
that gets canonicalized to funnel shift intrinsics in IR.
I did draft an alternative patch as an enhancement to
"shouldSinkOperands()", but that was awkward because
we have to key the transform from the select, but then
look at both its users and its operands.
David Green [Sat, 16 May 2020 14:12:41 +0000 (15:12 +0100)]
[ARM] Combines for VMOVN
This adds two combines for VMOVN, one to fold
VMOVN[tb](c, VQMOVNb(a, b)) => VQMOVN[tb](c, b)
The other to perform demand bits analysis on the lanes of a VMOVN. We
know that only the bottom lanes of the second operand and the top or
bottom lanes of the Qd operand are needed in the result, depending on if
the VMOVN is bottom or top.
Differential Revision: https://reviews.llvm.org/D77718
David Green [Sat, 16 May 2020 13:54:33 +0000 (14:54 +0100)]
[ARM] MVE saturating truncates
This adds some custom lowering for VQMOVN, an instruction that can be
used to perform saturating truncates from a pair of min(max(X, -0x8000),
0x7fff), providing those constants are correct. This leaves a VQMOVNBs
which saturates the value and inserts that into the bottom lanes of an
existing vector. We then need to do something with the other lanes,
extending the value using a vmovlb.
Ideally, as will often be the case, only the bottom lane of what remains
will be demanded, allowing the vmovlb to be removed. Which should mean
the instruction is either equal or a win most of the time, and allows
some extra follow-up folding to happen.
Differential Revision: https://reviews.llvm.org/D77590
Simon Pilgrim [Sat, 16 May 2020 13:40:00 +0000 (14:40 +0100)]
DIEHash.cpp - remove headers explicitly included in DIEHash.h. NFC.
Don't duplicate module header includes.