Nikita Popov [Sun, 29 Aug 2021 19:07:13 +0000 (21:07 +0200)]
[SCEVExpander] Reuse removePointerBase() for canonical addrecs
ExposePointerBase() in SCEVExpander implements basically the same
functionality as removePointerBase() in SCEV, so reuse it.
The SCEVExpander code assumes that the pointer operand on adds is
the last one -- I'm not sure that always holds. As such this might
not be strictly NFC.
Fangrui Song [Sun, 29 Aug 2021 19:07:56 +0000 (12:07 -0700)]
[ELF] Remove unused processRelocAux argument. NFC
Nikita Popov [Sun, 29 Aug 2021 18:46:24 +0000 (20:46 +0200)]
[SCEVExpander] Remove unnecessary mul/udiv check (NFC)
Pointer-typed SCEV expressions can no longer be mul or udiv, so
we do not need to specially handle them here.
Nikita Popov [Sun, 29 Aug 2021 18:29:04 +0000 (20:29 +0200)]
[SCEVExpander] Assert single pointer op in add (NFC)
There can only be one pointer operand in an add expression, and
we have sorted operands to guarantee that it is the first. As
such, the pointer check for other operands is dead code.
Nikita Popov [Sun, 29 Aug 2021 17:58:56 +0000 (19:58 +0200)]
[SCEV] Assert unique pointer base (NFC)
Add expressions can contain at most one pointer operand nowadays,
assert that in getPointerBase() and removePointerBase().
Vince Bridgers [Fri, 27 Aug 2021 22:57:44 +0000 (17:57 -0500)]
[X86] Remove X86LowerAMXType::getRowFromCol from X86LowerAMXType.cpp
Remove method X86LowerAMXType::getRowFromCol since it's not used, and
it's causing a warning.
Reviewed By: LuoYuanke
Differential Revision: https://reviews.llvm.org/D108862
Kazu Hirata [Sun, 29 Aug 2021 16:14:47 +0000 (09:14 -0700)]
[Support] Remove redundant calls to str() and c_str() (NFC)
Identified with readability-redundant-string-cstr.
Lei Zhang [Sun, 29 Aug 2021 16:01:19 +0000 (12:01 -0400)]
[mlir][spirv] Use type dyn_cast when scanning spv.GlobalVariable
This avoids crashes when there are spv.GlobalVariable without
pointer type.
Sylvestre Ledru [Sun, 29 Aug 2021 15:44:58 +0000 (17:44 +0200)]
test-release.sh: remove an unused reference to the old svn repo
Roman Podoliaka [Sun, 29 Aug 2021 15:26:15 +0000 (17:26 +0200)]
[lldb] Allow to register frame recognizers applied beyond the first instruction
It is currently possible to register a frame recognizer, but it will be applied if and only if the frame's PC points to the very first instruction of the specified function, which limits usability of this feature.
The implementation already supports changing this behaviour by passing an additional flag, but it's not possible to set it via the command interface. Fix that.
Reviewed By: jingham
Differential Revision: https://reviews.llvm.org/D108510
Nikita Popov [Sat, 28 Aug 2021 10:25:58 +0000 (12:25 +0200)]
[NewPM] Add missing LTO ArgPromotion pass
This is a followup to D96780 to add one more pass missing from the
NewPM LTO pipeline. The missing ArgPromotion run is inserted at
the same position as in the LegacyPM, resolving the already
present FIXME:
https://github.com/llvm/llvm-project/blob/
16086d47c0d0cd08ffae8e69a69c88653e654d01/llvm/lib/Transforms/IPO/PassManagerBuilder.cpp#L1096-L1098
The compile-time impact is minimal with ~0.1% geomean regression
on CTMark.
Differential Revision: https://reviews.llvm.org/D108866
Yonghong Song [Tue, 20 Jul 2021 17:59:19 +0000 (10:59 -0700)]
[BPF] support btf_tag attribute in .BTF section
A new kind BTF_KIND_TAG is added to .BTF to encode
btf_tag attributes. The format looks like
CommonType.name : attribute string
CommonType.type : attached to a struct/union/func/var.
CommonType.info : encoding BTF_KIND_TAG
kflag == 1 to indicate the attribute is
for CommonType.type, or kflag == 0
for struct/union member or func argument.
one uint32_t : to encode which member/argument starting from 0.
If one particular type or member/argument has more than one attribute,
multiple BTF_KIND_TAG will be generated.
Differential Revision: https://reviews.llvm.org/D106622
Michael Kruse [Sun, 29 Aug 2021 03:50:07 +0000 (22:50 -0500)]
[Polly][test] Add dependency to count.
Polly does not use the count program itself, but somewhere in lit it is
expected to exists. Otherwise, the following error occurs:
llvm-lit: llvm-project/llvm/utils/lit/lit/llvm/subst.py:133: fatal: Did not find count in ./bin
Michael Kruse [Sun, 29 Aug 2021 03:38:52 +0000 (22:38 -0500)]
[Polly] Fix dumpfunction.ll test.
Steven Wan [Sat, 28 Aug 2021 23:47:31 +0000 (19:47 -0400)]
TypeInfo records more information about align requirement
Extend the information preserved in `TypeInfo` by replacing the `AlignIsRequired` bool flag with a three-valued enum, the enum also indicates where the alignment attribute come from, which could be helpful in determining whether the attribute should overrule.
Reviewed By: rjmccall
Differential Revision: https://reviews.llvm.org/D108858
Shilei Tian [Sat, 28 Aug 2021 22:08:03 +0000 (18:08 -0400)]
[OpenMP][NVPTX] Fixed missing variables for CUDA free compilation in NVPTX plugin
`CU_EVENT_DEFAULT` is defined in CUDA header. It should be added to
`openmp/libomptarget/plugins/cuda/dynamic_cuda/cuda.h` for CUDA free build.
Reviewed By: ronlieb
Differential Revision: https://reviews.llvm.org/D108878
Michael Kruse [Sat, 28 Aug 2021 21:48:28 +0000 (16:48 -0500)]
[Polly] Don't prune non-external function itself from dump.
David Carlier [Sat, 28 Aug 2021 21:53:16 +0000 (22:53 +0100)]
[Sanitizers] PGO implements enable/disable SIGKILL on callers on FreeBSD.
Reviewed By: phosek
Differential Revision: https://reviews.llvm.org/D108867
Fangrui Song [Sat, 28 Aug 2021 20:56:31 +0000 (13:56 -0700)]
[Linker] Replace comdat based bool LinkFromSrc with enum class LinkFrom and improve nodeduplicate tests. NFC
This is different from symbol resolution based LinkFromSrc. Rename to be
clearer.
In the future we may support a new enum member 'Both' for nodeduplicate. This is
feasible (by renaming to a private linkage GlobalValue), but we need to be
careful not to break InstrProfiling.cpp's expectation of parallel profd/profc.
The challenge is that current LTO symbol resolution only allows to mark one
profc as prevailing: the other profc in another comdat nodeduplicate may be
discarded while its associated profd isn't.
Nikita Popov [Sat, 28 Aug 2021 20:36:37 +0000 (22:36 +0200)]
[llc] Initialize context for parsing options
This will allow using -force-opaque-pointers in codegen tests.
Shilei Tian [Sat, 28 Aug 2021 20:24:06 +0000 (16:24 -0400)]
[OpenMP][Offloading] Add support for event related interfaces
This patch adds the support form event related interfaces, which will be used
later to fix data race. See D104418 for more details.
Reviewed By: jdoerfert, ye-luo
Differential Revision: https://reviews.llvm.org/D108528
Fangrui Song [Sat, 28 Aug 2021 17:11:08 +0000 (10:11 -0700)]
[test] Reorganize test/Linker/comdat*.ll
Aart Bik [Sat, 28 Aug 2021 16:22:15 +0000 (09:22 -0700)]
[mlir] add missing include, fix broken build
Reviewed By: mehdi_amini
Differential Revision: https://reviews.llvm.org/D108873
George Rokos [Sat, 28 Aug 2021 14:35:42 +0000 (07:35 -0700)]
[libomptarget][NFC] Fixed tests which checked for obsolete string "getOrAllocTgtPtr"
Markus Böck [Sat, 28 Aug 2021 13:34:34 +0000 (15:34 +0200)]
[mlir][NFC] Fully qualify default value of Attributes `getStorageType()` in files generated by mlir-tblgen
Kazu Hirata [Sat, 28 Aug 2021 13:37:03 +0000 (06:37 -0700)]
[llvm] Fix typos in documentation (NFC)
Kazu Hirata [Sat, 28 Aug 2021 13:37:01 +0000 (06:37 -0700)]
[Analysis] Fix a "set but not used" warning
Nikita Popov [Fri, 27 Aug 2021 20:18:07 +0000 (22:18 +0200)]
[WebAssembly] Fix FastISel of condition in different block (PR51651)
If the icmp is in a different block, then the register for the icmp
operand may not be initialized, as it nominally does not have
cross-block uses. Add a check that the icmp is in the same block
as the branch, which should be the common case.
This matches what X86 FastISel does:
https://github.com/llvm/llvm-project/blob/
5b6b090cf2129228f05d7d0f504676b67f7524cf/llvm/lib/Target/X86/X86FastISel.cpp#L1648
The "not" transform that could have a similar issue is dropped
entirely, because it is currently dead: The incoming value is
a branch or select condition of type i1, but this code requires
an i32 to trigger.
Fixes https://bugs.llvm.org/show_bug.cgi?id=51651.
Differential Revision: https://reviews.llvm.org/D108840
Uday Bondhugula [Fri, 20 Aug 2021 11:46:25 +0000 (17:16 +0530)]
[MLIR][GPU] Drop mgpuMemHostRegisterMemRef's dependence on LLVM Support
Drop mgpuMemHostRegisterMemRef's dependence on LLVM Support. This
method is the only one in CUDA runtime wrappers library that creates
a dependence on libLLVMSupport due to its use of SmallVector and
ArrayRef. The code can be as easily/compactly written without those ADT.
The dependence on LLVMSupport adds a significant amount of additional
complexity for external things that want to link this library in (both
statically or as a shared object) since libLLVMSupport includes numerous
other objects that are sensitive to C++ compiler version and ABI.
Differential Revision: https://reviews.llvm.org/D108684
luxufan [Fri, 27 Aug 2021 14:17:06 +0000 (22:17 +0800)]
[JITLink][RISCV] Support GOT/PLT relocations
This patch add the R_RISCV_GOT_HI20 and R_RISCV_CALL_PLT relocation support. And the basic got/plt was implemented. Because of riscv32 and riscv64 has different pointer size, the got entry size and instructions of plt entry is different. This patch is the basic support, the optimization pass at preFixup stage has not been implemented.
Reviewed By: lhames
Differential Revision: https://reviews.llvm.org/D107688
Nico Weber [Fri, 27 Aug 2021 23:20:29 +0000 (19:20 -0400)]
[lld/mac] Don't assert on -dead_strip + arm64 range extension thunks
The assert is harmless and thinks worked fine in builds with asserts enabled,
but it's still nice to fix the assert.
Differential Revision: https://reviews.llvm.org/D108853
Ben Barham [Sat, 28 Aug 2021 02:38:46 +0000 (19:38 -0700)]
[Modules] Change result of reading AST block to llvm::Error instead
Reading the AST block can never fail with a recoverable error as modules
cannot be removed during this phase. Change the return type of these
functions to return an llvm::Error instead, ie. either success or
failure.
NFC other than the wording of some of the errors.
Differential Revision: https://reviews.llvm.org/D108268
Mehdi Amini [Sat, 28 Aug 2021 01:09:00 +0000 (01:09 +0000)]
Remove `const` from `const T &&` in debugString() helper to make it a universal reference (NFC)
It broke lvalue arguments otherwise:
Mehdi Amini [Sat, 28 Aug 2021 00:40:28 +0000 (00:40 +0000)]
Use a universal reference (&& instead of const &) for `debugString()` helper (NFC)
Some classes like mlir::Operation have a non-const print() method.
David Blaikie [Sat, 28 Aug 2021 00:12:42 +0000 (17:12 -0700)]
TypePrinter:printTo remove redundant parameter (IsPack/SkipBrackets)
Mehdi Amini [Sat, 28 Aug 2021 00:09:27 +0000 (00:09 +0000)]
Specify argument to be `const` for `debugString()` helper (NFC)
This allows using this helper with rvalues.
David Carlier [Fri, 27 Aug 2021 22:58:06 +0000 (23:58 +0100)]
[Sanitizers] wordexp interceptor build fix on FreeBSD.
Reviewers: vitalybuka
Reviewed-By: vitalybuka
Differential Revision: https://reviews.llvm.org/D108838
Michael Jones [Fri, 27 Aug 2021 22:54:02 +0000 (22:54 +0000)]
[compiler-rt][obvious] fix typo, delete paren
I should've deleted a parenthesis on line 4222 in
https://reviews.llvm.org/D108843 and this patch fixes it.
Differential Revision: https://reviews.llvm.org/D108852
Michael Jones [Fri, 27 Aug 2021 20:54:44 +0000 (20:54 +0000)]
[compiler-rt] fix real strlens that were missed
In reviews.llvm.org/D108316 I missed a few instances of REAL(strlen) and
this change fixes that, as well as restoring one that seems like it
shouldn't have been changed.
Reviewed By: hctim, vitalybuka
Differential Revision: https://reviews.llvm.org/D108843
Nick Desaulniers [Fri, 27 Aug 2021 22:15:32 +0000 (15:15 -0700)]
[MipsISelLowering] avoid emitting libcalls to __mulodi4()
__has_builtin(__builtin_mul_overflow) returns true for 32b MIPS targets,
but Clang is deferring to compiler RT when encountering `long long`
types. This breaks sanitizer builds of the Linux kernel that are using
__builtin_mul_overflow with these types for these targets.
If the semantics of __has_builtin mean "the compiler resolves these,
always" then we shouldn't conditionally emit a libcall.
This will still need to be worked around in the Linux kernel in order to
continue to support malta_defconfig builds of the Linux kernel for this
target with older releases of clang.
Link: https://bugs.llvm.org/show_bug.cgi?id=28629
Link: https://github.com/ClangBuiltLinux/linux/issues/1438
Reviewed By: rengolin
Differential Revision: https://reviews.llvm.org/D108844
Nick Desaulniers [Fri, 27 Aug 2021 22:06:08 +0000 (15:06 -0700)]
[ARMISelLowering] avoid emitting libcalls to __mulodi4()
__has_builtin(__builtin_mul_overflow) returns true for 32b ARM targets,
but Clang is deferring to compiler RT when encountering `long long`
types. This breaks sanitizer builds of the Linux kernel that are using
__builtin_mul_overflow with these types for these targets.
If the semantics of __has_builtin mean "the compiler resolves these,
always" then we shouldn't conditionally emit a libcall.
This will still need to be worked around in the Linux kernel in order to
continue to support allmodconfig builds of the Linux kernel for this
target with older releases of clang.
Link: https://bugs.llvm.org/show_bug.cgi?id=28629
Link: https://github.com/ClangBuiltLinux/linux/issues/1438
Reviewed By: rengolin
Differential Revision: https://reviews.llvm.org/D108842
Aart Bik [Thu, 26 Aug 2021 05:23:46 +0000 (22:23 -0700)]
[mlir][sparse] fully implement sparse tensor to sparse tensor conversions
with rigorous integration test
Reviewed By: bixia
Differential Revision: https://reviews.llvm.org/D108721
Andrew Litteken [Fri, 27 Aug 2021 21:01:43 +0000 (14:01 -0700)]
[IRSim][IROutliner] Canonicalizing commutative value numbering between similarity sections.
When the initial relationship between two pairs of values between
similar sections is ambiguous to commutativity, arguments to the
outlined functions can be passed in such that the order is incorrect,
causing miscompilations. This adds a canonical mapping to each
similarity section, so that we can maintain the relationship of global
value numbering from one section to another.
Added Tests:
Transforms/IROutliner/outlining-commutative-operands-opposite-order.ll
unittests/Analysis/IRSimilarityIdentifierTest.cpp - IRSimilarityCandidate:CanonicalNumbering
Reviewers: jroelofs, jpaquette, yroux
Differential Revision: https://reviews.llvm.org/D104143
owenca [Fri, 27 Aug 2021 21:32:30 +0000 (14:32 -0700)]
[clang-format] Fix AllowAllConstructorInitializersOnNextLine value
The default value of the now deprecated
AllowAllConstructorInitializersOnNextLine was always true
regardless of the value of BasedOnStyle. This patch fixes the typo
in the code that handles the related backward compatibility.
Johannes Doerfert [Fri, 27 Aug 2021 21:37:51 +0000 (16:37 -0500)]
[Attributor][NFC] Silence unused variable warning
Nikita Popov [Thu, 26 Aug 2021 19:23:58 +0000 (21:23 +0200)]
[MergeICmps] Ignore clobbering instructions before the loads
This is another followup to D106591. Even if there is an
instruction that clobbers one of the loads, this doesn't matter if
it happens before the loads. Those instructions aren't affected by
the transform at all.
The gep-references-bb.ll is modified to preserve the spirit of the
test, as the store to @g no longer impacts the transform.
Differential Revision: https://reviews.llvm.org/D108782
owenca [Fri, 27 Aug 2021 21:12:01 +0000 (14:12 -0700)]
[clang-format] [NFC] Fix the coding style of unit tests header file
Philip Reames [Fri, 27 Aug 2021 21:18:38 +0000 (14:18 -0700)]
[LoopDeletion] Use max trip count to break backedge in addition to exact one
We'd added support a while back from breaking the backedge if SCEV can prove the trip count is zero. However, we used the exact trip count which requires *all* exits be analyzeable. I noticed while writing test cases for another patch that this disallows cases where one exit is provably taken paired with another which is unknown. This patch adds the upper bound case.
We could use a symbolic max trip count here instead, but we use an isKnownNonZero filter (presumably for compile time?) for the first-iteration reasoning. I decided this was a more obvious incremental step, and we could go back and untangle the schemes separately.
Differential Revision: https://reviews.llvm.org/D108833
Pirama Arumuga Nainar [Wed, 18 Aug 2021 20:52:14 +0000 (13:52 -0700)]
[lld/test/ELF] Test fetch from archive to resolve undefined symbols in shared libs
Add missing test coverage uncovered in review of D108006.
Differential Revision: https://reviews.llvm.org/D108328
Roland McGrath [Fri, 27 Aug 2021 21:11:31 +0000 (14:11 -0700)]
[libc][NFC] Fix onre more -Wconversion warning in strtoul test code.
The last change missed one spot.
Differential Revision: https://reviews.llvm.org/D108845
Roland McGrath [Fri, 27 Aug 2021 18:04:19 +0000 (11:04 -0700)]
[libc] Fix various -Wconversion warnings in strto*l test code.
The Fuchsia build compiles the libc and test code with lots
of warnings enabled, including all the integer conversion warnings.
There was some sloppy type usage here that triggered some of those.
Reviewed By: michaelrj
Differential Revision: https://reviews.llvm.org/D108800
Evgenii Stepanov [Fri, 27 Aug 2021 20:36:38 +0000 (13:36 -0700)]
Support LLVM_ENABLE_PER_TARGET_RUNTIME_DIR in the sanitizer symbolizer build.
In this mode libc++ headers end up in two directories:
* include/<triple>/c++/v1 for the site config header
* include/c++/v1 for everything else
Also switch from -I to -isystem.
Differential Revision: https://reviews.llvm.org/D108841
Arthur Eubanks [Fri, 27 Aug 2021 20:39:43 +0000 (13:39 -0700)]
[test] Precommit test for D108837
Valentin Churavy [Fri, 27 Aug 2021 15:20:14 +0000 (17:20 +0200)]
[MergeICmps] Don't merge icmps derived from pointers with addressspaces
IIUC we can't emit `memcmp` between pointers in addressspaces,
doing so will trigger an assertion since the signature of the memcmp
will not match it's arguments (https://bugs.llvm.org/show_bug.cgi?id=48661).
This PR disables the attempt to merge icmps,
when the pointer is in an addressspace.
Reviewed By: #julialang, vtjnash
Differential Revision: https://reviews.llvm.org/D94813
Zequan Wu [Tue, 17 Aug 2021 21:18:52 +0000 (14:18 -0700)]
[Profile] Support __llvm_profile_set_file_object in continuous mode.
Replace D107203, because __llvm_profile_set_file_object is usually used when the
process doesn't have permission to open/create file. That patch trying to copy
from old profile to new profile contradicts with the usage.
Differential Revision: https://reviews.llvm.org/D108242
Akira Hatanaka [Wed, 25 Aug 2021 20:11:19 +0000 (13:11 -0700)]
Use std::less instead of operator < in less_first and less_second
According to the standard, if p1 and p2 are both pointers, p1 < p2 and
p2 < p1 can both be false in theory in some cases:
https://eel.is/c++draft/expr.rel#4.3
std::less<void> yields a implementation-defined strict total order over
pointers:
https://eel.is/c++draft/comparisons.general
Differential Revision: https://reviews.llvm.org/D108733
Craig Topper [Fri, 27 Aug 2021 19:30:31 +0000 (12:30 -0700)]
[RISCV] Use ~0ULL instead of ~0U when checking for invalid ErrorInfo.
ErrorInfo is a uint64_t and is initialized to all 1s.
Not sure how to test this. Noticed while working on .insn support.
Michał Górny [Sat, 7 Aug 2021 14:40:09 +0000 (16:40 +0200)]
[llvm] [lit] Support forcing lexical test order
Add a new --order option to choose between available test orders:
the default "smart" order, predictable "lexical" order or "random"
order. Default to using lexical order and one job in the lit test
suite.
Differential Revision: https://reviews.llvm.org/D107695
Arthur Eubanks [Fri, 27 Aug 2021 18:16:47 +0000 (11:16 -0700)]
[test] More test precommits for D108734
Haowei Wu [Fri, 20 Aug 2021 00:39:36 +0000 (17:39 -0700)]
[ifs] Add option to hide undefined symbols
This change add an option to llvm-ifs to hide undefined symbols from
its output.
Differential Revision: https://reviews.llvm.org/D108428
mydeveloperday [Fri, 27 Aug 2021 18:10:45 +0000 (19:10 +0100)]
[clang-format] [PR51640] - New AfterEnum brace wrapping changes have cause C# behaviour to change
LLVM 13.0.0-rc2 shows change of behaviour in enum and interface BraceWrapping (likely before we simply didn't wrap) but may be related to {D99840}
Logged as https://bugs.llvm.org/show_bug.cgi?id=51640
This change ensure AfterEnum works for
`internal|public|protected|private enum A {` in the same way as it works for `enum A {` in C++
A similar issue was also observed with `interface` in C#
Reviewed By: krasimir, owenpan
Differential Revision: https://reviews.llvm.org/D108810
Johannes Doerfert [Thu, 26 Aug 2021 18:04:29 +0000 (13:04 -0500)]
[OpenMP][FIX] Allow declare variant to work with reference types
Reference types in the return or parameter position did cause the OpenMP
declare variant overload reasoning to give up. We should allow them as
we allow any other type.
This should fix the bug reported on the mailing list:
https://lists.llvm.org/pipermail/openmp-dev/2021-August/004094.html
Reviewed By: ABataev, pdhaliwal
Differential Revision: https://reviews.llvm.org/D108774
Johannes Doerfert [Tue, 17 Aug 2021 06:37:29 +0000 (01:37 -0500)]
[Attributor][FIX] Recursion via memory needs to be tracked explicitly
Recursion can happen when we see a PHI use the second time or when we
look at a store value operand use again. We already visited the
potential copies and doing so again will just cause endless looping.
Reviewed By: kuter
Differential Revision: https://reviews.llvm.org/D108190
Johannes Doerfert [Mon, 16 Aug 2021 16:04:09 +0000 (11:04 -0500)]
[Attributor][FIX] Do not treat byval args as local memory (for now)
For now we do should not treat byval arguments as local copies performed
on the call edge, though, in general we should. To make that happen we
need to teach various passes, e.g., DSE, about the copy effect of a
byval. That would also allow us to mark functions only accessing byval
arguments as readnone again, atguably their acceses have no effect
outside of the function, like accesses to allocas.
Reviewed By: kuter
Differential Revision: https://reviews.llvm.org/D108140
Jason Liu [Fri, 27 Aug 2021 17:37:34 +0000 (13:37 -0400)]
Fix assertion when passing function into inline asm's input operand
This seem to be a regression caused by this change:
https://reviews.llvm.org/D60943.
Since we delayed report the error, we would run into some invalid
state in clang and llvm.
Without this fix, clang would assert when passing function into
inline asm's input operand.
Differential Revision: https://reviews.llvm.org/D107941
LLVM GN Syncbot [Fri, 27 Aug 2021 17:29:43 +0000 (17:29 +0000)]
[gn build] Port
54e8cae56529
Philip Reames [Fri, 27 Aug 2021 17:12:15 +0000 (10:12 -0700)]
Special case common branch patterns in breakLoopBackedge (try 2)
Changes since aec08e:
* Adjust placement of a closing brace so that the general case actually runs. Turns out we had *no* coverage of the switch case. I added one in eae90fd.
* Drop .llvm.loop.* metadata from the new branch as there is no longer a loop to annotate.
Original commit message:
This special cases an unconditional latch and a conditional branch latch exit to improve codegen and test readability. I am hoping to reuse this function in the runtime unroll code, but without this change, the test diffs are far too complex to assess.
Roman Lebedev [Fri, 27 Aug 2021 17:15:45 +0000 (20:15 +0300)]
[Codegen][X86] EltsFromConsecutiveLoads(): if only have AVX1, ensure that the "load" is actually foldable (PR51615)
This fixes another reproducer from https://bugs.llvm.org/show_bug.cgi?id=51615
And again, the fix lies not in the code added in D105390
In this case, we completely don't check that the "broadcast-from-mem" we create
can actually fold the load. In this case, it's operand was not a load at all:
```
Combining: t16: v8i32 = vector_shuffle<0,u,u,u,0,u,u,u> t14, undef:v8i32
Creating new node: t29: i32 = undef
RepeatLoad:
t8: i32 = truncate t7
t7: i64 = extract_vector_elt t5, Constant:i64<0>
t5: v2i64,ch = load<(load (s128) from %ir.arg)> t0, t2, undef:i64
t2: i64,ch = CopyFromReg t0, Register:i64 %0
t1: i64 = Register %0
t4: i64 = undef
t3: i64 = Constant<0>
Combining: t15: v8i32 = undef
```
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D108821
Philipp Krones [Fri, 27 Aug 2021 16:55:52 +0000 (17:55 +0100)]
[MC][RISCV] Add RISCV MCObjectFileInfo
This makes sure, that the text section will have a 2-byte alignment, if
the +c extension is enabled.
Reviewed By: MaskRay, luismarques
Differential Revision: https://reviews.llvm.org/D102052
Craig Topper [Fri, 27 Aug 2021 16:51:05 +0000 (09:51 -0700)]
[RISCV] Add -riscv-v-fixed-length-vector-elen-max to limit the ELEN used for fixed length vectorization.
This adds an ELEN limit for fixed length vectors. This will scalarize
any elements larger than this. It will also disable some fractional
LMULs. For example, if ELEN=32 then mf8 becomes illegal, i32/f32
vectors can't use any fractional LMULs, i16/f16 can only use mf2,
and i8 can use mf2 and mf4.
We may also need something for the scalable vectors, but that has
interactions with the intrinsics and we can't scalarize a scalable
vector.
Longer term this should come from one of the Zve* features
Sizhe Zhao [Thu, 26 Aug 2021 20:11:54 +0000 (23:11 +0300)]
[libcxx] Use GetSystemTimePreciseAsFileTime() if available
We will try to use GetSystemTimePreciseAsFileTime if possible.
Reference: https://sourceforge.net/p/mingw-w64/mingw-w64/ci/
59195b2d7fe26549f70969b0dd487293819f023e/.
Reviewed By: compnerd, #libc, mstorsjo, ldionne
Differential Revision: https://reviews.llvm.org/D104987
Philip Reames [Fri, 27 Aug 2021 17:09:51 +0000 (10:09 -0700)]
[test] exercise breakLoopBackedge with a switch latch cond
This was reduced from a test case which triggered a revert to my recent change to same function. It turns out we didn't have *any* coverage of the non-branch latch and my patch was blatantly broken.
LLVM GN Syncbot [Fri, 27 Aug 2021 16:46:52 +0000 (16:46 +0000)]
[gn build] Port
c8b14c03ec74
Louis Dionne [Fri, 27 Aug 2021 14:36:04 +0000 (10:36 -0400)]
[libc++][NFC] Fix include guard for decay_copy.h and remove underscores from the header
We don't use double underscores for private header names when they are
in a subdirectory with double underscores already.
Differential Revision: https://reviews.llvm.org/D108820
Louis Dionne [Thu, 26 Aug 2021 18:54:07 +0000 (14:54 -0400)]
[libc++][NFC] Remove useless _LIBCPP_PUSH_MACROS
Only files that actually use min/max are required to do this dance.
Differential Revision: https://reviews.llvm.org/D108778
Walter Erquinigo [Fri, 27 Aug 2021 16:31:41 +0000 (09:31 -0700)]
[trace] [intel pt] Create a "process trace save" command
added new command "process trace save -d <directory>".
-it saves a JSON file as <directory>/trace.json, with the main properties of the trace session.
-it saves binary Intel-pt trace as <directory>/thread_id.trace; each file saves each thread.
-it saves modules to the directory <directory>/modules .
-it only works for live process and it only support Intel-pt right now.
Example:
```
b main
run
process trace start
n
process trace save -d /tmp/mytrace
```
A file named trace.json and xxx.trace should be generated in /tmp/mytrace. To load the trace that was just saved:
```
trace load /tmp/mytrace
thread trace dump instructions
```
You should see the instructions of the trace got printed.
To run a test:
```
cd ~/llvm-sand/build/Release/fbcode-x86_64/toolchain
ninja lldb-dotest
./bin/lldb-dotest -p TestTraceSave
```
Reviewed By: wallace
Differential Revision: https://reviews.llvm.org/D107669
Siva Chandra Reddy [Fri, 27 Aug 2021 15:48:55 +0000 (15:48 +0000)]
[libc][Obvious] Add header guards for the generated linux syscall header file.
Fangrui Song [Fri, 27 Aug 2021 15:53:55 +0000 (08:53 -0700)]
[MC] Change ELFOSABI_NONE to ELFOSABI_GNU for STB_GNU_UNIQUE
Similar to D97976.
On Linux, most GCC installations are configured with
`--enable-gnu-unique-object` and such GCC emits `@gnu_unique_object` assembly.
The feature is highly controversial and disliked by many folks.
(On glibc DF_1_NODELETE is implicitly enabled and makes dlclose a no-op).
In llvm-project STB_GNU_UNIQUE is assembly only. Clang does not use STB_GNU_UNIQUE.
Use ELFOSABI_GNU to match GNU as behavior and avoid collision with other
OSABI binding values.
Reviewed By: jrtc27
Differential Revision: https://reviews.llvm.org/D107861
Arthur Eubanks [Wed, 25 Aug 2021 22:24:49 +0000 (15:24 -0700)]
[gn build] Don't copy xray includes
The gn build doesn't support xray, so there's no reason to make the xray
headers available. Some CMake checks check if xray includes are
available to determine if xray is usable. Since we don't build the xray
runtime, there are link errors.
Reviewed By: thakis
Differential Revision: https://reviews.llvm.org/D108737
Louis Dionne [Fri, 27 Aug 2021 15:47:27 +0000 (11:47 -0400)]
[libc++][NFC] Remove unused helper function in the test suite
Fanbo Meng [Fri, 27 Aug 2021 14:26:51 +0000 (10:26 -0400)]
[MCParser][z/OS] Mark test as unsupported for the z/OS Target
Marking test as unsupported for the same reason as https://reviews.llvm.org/D105204
Reviewed By: abhina.sreeskantharajan
Differential Revision: https://reviews.llvm.org/D108819
Kazu Hirata [Fri, 27 Aug 2021 15:42:57 +0000 (08:42 -0700)]
[IR] Remove getWithOperandReplaced (NFC)
The function hasn't been used for at least 10 years.
Dmitry Preobrazhensky [Fri, 27 Aug 2021 14:16:22 +0000 (17:16 +0300)]
[AMDGPU][MC][NFC][DOC] Updated AMD GPU assembler syntax description.
Summary of changes:
- Added f16 omod modifier (bug 51386).
- Corrected names of data types (bug 48638).
- Enabled a16 with most GFX10 MIMG opcodes (see https://reviews.llvm.org/D102231).
- Corrected description of integer operands (bug 51130).
- Corrected description of 8-bit DS offsets (bug 51536).
- Improved PERMLANE op_sel description.
- Corrected *SAD* opcode types.
Joe Loser [Fri, 27 Aug 2021 14:08:11 +0000 (10:08 -0400)]
[libc++][NFC] Remove extra __ranges/take_view.h entry in CMakeLists.txt
Differential Revision: https://reviews.llvm.org/D108802
Louis Dionne [Fri, 27 Aug 2021 14:01:29 +0000 (10:01 -0400)]
Revert "[CMake] Enable LLVM_ENABLE_PER_TARGET_RUNTIME_DIR by default on Linux"
This reverts commit
abb956370ee71d018e9a88ae196f039f6c4e0dae, which broke
the libc++ CI on Linux.
owenca [Wed, 25 Aug 2021 23:22:02 +0000 (16:22 -0700)]
[clang-format] Group options that pack constructor initializers
Add a new option PackConstructorInitializers and deprecate the
related options ConstructorInitializerAllOnOneLineOrOnePerLine and
AllowAllConstructorInitializersOnNextLine. Below is the mapping:
PackConstructorInitializers ConstructorInitializer... AllowAll...
Never - -
BinPack false -
CurrentLine true false
NextLine true true
The option value Never fixes PR50549 by always placing each
constructor initializer on its own line.
Differential Revision: https://reviews.llvm.org/D108752
Matt Arsenault [Fri, 27 Aug 2021 13:18:26 +0000 (09:18 -0400)]
GlobalISel: Remove check for empty functions as these are invalid IR
Nico Weber [Fri, 27 Aug 2021 02:03:26 +0000 (22:03 -0400)]
[lld/COFF] Ignore /LTCG, /LTCG:, /LTCGOUT:, /ILK: flags
We currently complain "could not open /LTCG: no such file or directory",
which isn't very useful. We could emit a warning when we see this flag, but
just ignoring it seems fine.
Final missing part of PR38799.
Differential Revision: https://reviews.llvm.org/D108799
Nico Weber [Fri, 27 Aug 2021 02:01:00 +0000 (22:01 -0400)]
[lld/COFF] Use P_priv more
P_priv does the same as the old QF further down. Standardize on P_priv.
No behavior change.
Differential Revision: https://reviews.llvm.org/D108798
Balazs Benics [Fri, 27 Aug 2021 12:41:26 +0000 (14:41 +0200)]
[analyzer] MallocOverflow should consider comparisons only preceding malloc
MallocOverflow works in two phases:
1) Collects suspicious malloc calls, whose argument is a multiplication
2) Filters the aggregated list of suspicious malloc calls by iterating
over the BasicBlocks of the CFG looking for comparison binary
operators over the variable constituting in any suspicious malloc.
Consequently, it suppressed true-positive cases when the comparison
check was after the malloc call.
In this patch the checker will consider the relative position of the
relation check to the malloc call.
E.g.:
```lang=C++
void *check_after_malloc(int n, int x) {
int *p = NULL;
if (x == 42)
p = malloc(n * sizeof(int)); // Previously **no** warning, now it
// warns about this.
// The check is after the allocation!
if (n > 10) {
// Do something conditionally.
}
return p;
}
```
Reviewed By: martong
Differential Revision: https://reviews.llvm.org/D107804
Sanjay Patel [Fri, 27 Aug 2021 12:09:28 +0000 (08:09 -0400)]
[GlobalOpt] don't hoist constant expressions that can trap
We try to forward a stored-once-constant-value from one global access
to another, but that's not safe if the constant value is an expression
that can trap.
The tests are reduced from the miscompile examples in:
https://llvm.org/PR47578
Differential Revision: https://reviews.llvm.org/D108771
Jun Ma [Wed, 25 Aug 2021 11:25:38 +0000 (19:25 +0800)]
[AArch64][SVE] Optimize ptrue predicate pattern with known sve register width.
For vectors that are exactly equal to getMaxSVEVectorSizeInBits, just use
AArch64SVEPredPattern::all, which can enable the use of unpredicated ptrue when available.
TestPlan: check-llvm
Differential Revision: https://reviews.llvm.org/D108706
Jun Ma [Wed, 25 Aug 2021 09:25:39 +0000 (17:25 +0800)]
[AArch64][SVE] Add API for conversion between SVE predicate pattern and element number. NFC
This patch solely moves convert operation between SVE predicate pattern
and element number into two small functions. It's pre-commit patch for optimize
pture with known sve register width.
Differential Revision: https://reviews.llvm.org/D108705
Jun Ma [Wed, 25 Aug 2021 07:43:18 +0000 (15:43 +0800)]
[AArch64][SVE] Use getPTrue uniformly.NFC.
Andrea Di Biagio [Fri, 27 Aug 2021 11:48:30 +0000 (12:48 +0100)]
[MCA][NFC] Removed unused method, and fixed a coverity issue.
The coverity issue was reported agaist class MCAOperand
due to the lack of proper initialization for field Index.
No functional change intended.
Jon Chesterfield [Fri, 27 Aug 2021 11:34:02 +0000 (12:34 +0100)]
[openmp][amdgpu] Initial gfx10 offloading implementation
Lets wavefront size be 32 for amdgpu openmp, as well as 64.
Fixes up as little as possible to pass that through the libraries. This change
is end to end, as opposed to updating clang/devicertl/plugin separately. It can
be broken up for review/commit if preferred. Posting as-is so that others with
a gfx10 can try it out. It works roughly as well as gfx9 for me, but there are
probably bugs remaining as well as the todo: for letting grid values vary more.
Reviewed By: ronlieb
Differential Revision: https://reviews.llvm.org/D108708
Serge Pavlov [Fri, 13 Aug 2021 09:52:29 +0000 (16:52 +0700)]
[X86] Implement llvm.isnan(x86_fp80) as unordered comparison
x86_fp80 format allows values that do not fit any of IEEE-754 category.
Previously they were recognized by intrinsic __builtin_isnan as NaNs.
Now this intrinsic is implemented using instruction FXAM, which
distinguish between NaNs and unsupported values. It can make some
programs behave differently.
As a solution, this fix changes lowering of the intrinsic. If floating
point exceptions are ignored, llvm.isnan is lowered into unordered
comparison, as __buildtin_isnan was implemented earlier. In strictfp
functions the intrinsic is lowered using FXAM, which does not raise
exceptions even for signaling NaN, as required by IEEE-754 and C
standards.
Differential Revision: https://reviews.llvm.org/D108037
Nathan Sidwell [Thu, 26 Aug 2021 11:05:25 +0000 (04:05 -0700)]
[NFC][X86] Sret return register cleanup
There are no paths into LowerFormalParms that have already specified
the sret register. We always materialize a virtual and then assign it
to the physical reg at the point of the return.
Differential Revision: https://reviews.llvm.org/D108762
Carl Ritson [Fri, 27 Aug 2021 10:08:10 +0000 (19:08 +0900)]
[DAGCombine] Allow FMA combine with both FMA and FMAD
Without this change only the preferred fusion opcode is tested
when attempting to combine FMA operations.
If both FMA and FMAD are available then FMA ops formed prior to
legalization will not be merged post legalization as FMAD becomes
the preferred fusion opcode.
Reviewed By: foad
Differential Revision: https://reviews.llvm.org/D108619
Ricky Taylor [Thu, 26 Aug 2021 21:13:28 +0000 (22:13 +0100)]
[M68k] Update pointer data layout
Fixes PR51626.
The M68k requires that all instruction, word and long word reads are
aligned to word boundaries. From the 68020 onwards, there is a
performance benefit from aligning long words to long word boundaries.
The M68k uses the same data layout for pointers and integers.
In line with this, this commit updates the pointer data layout to
match the layout already set for 32-bit integers: 32:16:32.
Differential Revision: https://reviews.llvm.org/D108792