Bill Wendling [Thu, 18 Aug 2022 21:52:49 +0000 (14:52 -0700)]
[X86][AArch64][NFC] Simplify querying used argument registers
Registers used for arguments are listed as "live-ins" into the starting
basic block. This means we don't have to go through a potentially
expensive search through all possible argument registers when we only
care about used argument registers.
Differential Revision: https://reviews.llvm.org/D132181
jeff [Thu, 28 Jul 2022 18:09:14 +0000 (11:09 -0700)]
[InferAddressSpaces] [AMDGPU] Add inference for flat_atomic intrinsics
Certain address space dependent optimizations, like SeperateConstOffsetFromGEP, assume agreement between the address space of the recursive uses and the address space of the def. If this assumption is invalid, then optimizations may or may not be correct depending on properties of an address space for a given target, the address spaces of recursive uses, and the optimization being done.
This patch infers the previous address space for flat_atomic ptr arguments. As a result, the address spaces of the uses in flat_atomic cases will agree with the address space in recursive defs. If this results in non-flat address space, then isel may infer a different intrinsic. For example, if the result is a flat_atomic using global address space, then it will be lowered to the corresponding global_atomic intrinsic.
Change-Id: Ifcd981709dc2ea94d4acbcb84efe7176593ec8c7
Denis Antrushin [Fri, 19 Aug 2022 17:58:04 +0000 (00:58 +0700)]
[NFC] Precommit test for D132255.
Michael Buch [Fri, 19 Aug 2022 18:31:14 +0000 (19:31 +0100)]
Revert "[lldb] [gdb-remote] Include PID in vCont packets if multiprocess" - Part 2
This reverts commit
ccb9d4d4addc2fb2aa94cf776d43d8be35365272.
Reverts the associated tests
Simon Pilgrim [Fri, 19 Aug 2022 18:07:02 +0000 (19:07 +0100)]
[X86] Fix znver1 256-bit ALU/Logic/Blend uop counts
ymm instructions are double pumped on znver1 - noticed while trying to review size-latency costkinds numbers for D132216
Matches AMD 17h SOG / Agner / uops.info
Mats Petersson [Fri, 19 Aug 2022 12:34:48 +0000 (13:34 +0100)]
[Flang]Fix another way to crash SimplifyIntrinsics
Under some conditions, the defining op may be NULL, so
accept that rahter than try to use it and crash!
Adds test to prevent regression
Fixes github issue #57201
Reviewed By: vzakhari
Differential Revision: https://reviews.llvm.org/D132238
Chris Bieneman [Fri, 19 Aug 2022 22:21:49 +0000 (17:21 -0500)]
[NFC] Fix warning in asserts build
When asserts are enabled a warning for implicit conversion from 64-bit to
32-bit integer occurs here.
Philip Reames [Fri, 19 Aug 2022 17:37:39 +0000 (10:37 -0700)]
[RISCV] Correct costs for vector ceil/floor/trunc/round
Add vector costs for ceil/floor/trunc/round. As can be seen in the tests, the prior default costs were a significant under estimate of the actual code generated.
These costs are computed by simply generating code with the current backend, and then counting the number of instructions. I discount one vsetvli, and ignore the return.
Differential Revision: https://reviews.llvm.org/D131967
Michael Buch [Fri, 19 Aug 2022 17:05:04 +0000 (18:05 +0100)]
Revert "[lldb] [gdb-remote] Include PID in vCont packets if multiprocess"
This reverts commit
ccb9d4d4addc2fb2aa94cf776d43d8be35365272.
https://reviews.llvm.org/D131758
Differential Revision: https://reviews.llvm.org/D132250
Saleem Abdulrasool [Fri, 5 Aug 2022 16:52:03 +0000 (16:52 +0000)]
MC: make section classification a bit more thorough
This does *NOT* change the emitted section flags in any way. This only
impacts the internal classification of sections.
Extend the section classification in LLVM for ELF targets. This has one
important change: we now classify sections as text by default rather
than readonly. This matches the behaviour for GAS better.
Ensure that any section that has a writable attribute set is not treated
as readonly. We also special case any section named `.debug_` which is
reserved for DWARF as metadata. In the case none of the attributes are
set (or because no attributes were provided), consult the section name
for classification. We match the well known names and classify the
section accordingly. Any remaining section is now classified as text.
This change allows us to classify sections in the MC layer more
precisely which is needed for subsequent changes for handling target
specific behaviour.
Differential Revision: https://reviews.llvm.org/D131270
Reviewed By: @echristo
Simon Pilgrim [Fri, 19 Aug 2022 16:37:26 +0000 (17:37 +0100)]
Fix MSVC "not all control paths return a value" warning
Sanjay Patel [Fri, 19 Aug 2022 16:21:11 +0000 (12:21 -0400)]
[EarlyCSE][ConstantFolding] do not constant fold atan2(+/-0.0, +/-0.0)
These may raise an error (set errno) as discussed in the post-commit
comments for D127964, so we can't fold away the call and potentially
alter that behavior.
Adrian Prantl [Fri, 19 Aug 2022 16:22:19 +0000 (09:22 -0700)]
Revert "[lldb] [test] Add synchronization to TestContinue"
This reverts commit
7aadecae404b8d47aabdd874e9018a76fd4d1ffa.
I'm reverting this commit because it appears to break the green dragon
incremental LLDB bot.
https://reviews.llvm.org/D131758
See LLDB Incremental buildbot: https://green.lab.llvm.org/green/view/LLDB/job/lldb-cmake/46215/execution/node/70/log/
Adrian Prantl [Fri, 19 Aug 2022 16:21:55 +0000 (09:21 -0700)]
Revert "[lldb] [test] Disable new CommunicationTests on Windows"
This reverts commit
d38985a36be8b0165787f8893b3d2b0f831d1e83.
because I'm also reverting D131758.
Sanjay Patel [Fri, 19 Aug 2022 16:18:21 +0000 (12:18 -0400)]
[EarlyCSE][ConstantFolding] add tests for atan2 with zero args; NFC
Adrian Prantl [Fri, 19 Aug 2022 01:38:52 +0000 (18:38 -0700)]
Fix UB in DIExpression::appendOffset()
The absolute value of 0x8000000000000000 does not fit into an int64_t
and UBSan tells us that by crashing.
rdar://
98799670
Differential Revision: https://reviews.llvm.org/D132194
Craig Topper [Fri, 19 Aug 2022 15:59:49 +0000 (08:59 -0700)]
[RISCV] Use Triple::isRISCV/isRISCV32/isRISCV64 helps in some places. NFC
Reviewed By: reames
Differential Revision: https://reviews.llvm.org/D132197
isuckatcs [Fri, 19 Aug 2022 15:50:49 +0000 (17:50 +0200)]
[analyzer][NFC] Be more descriptive when we replay without inlining
This patch adds a ProgramPointTag to the EpsilonPoint created
before we replay a call without inlining.
Differential Revision: https://reviews.llvm.org/D132246
Denis Fatkulin [Fri, 19 Aug 2022 16:00:54 +0000 (19:00 +0300)]
[clang][index] Index unresolved member expression as reference
Unresolved member expressions aren't indexed as references.
Example code:
```
struct Foo {
template <typename T> void bar(T t);
};
template <typename T> void test(Foo F, T t) {
F.bar(t); // Not indexed
}
```
Reviewed By: hokein
Differential Revision: https://reviews.llvm.org/D131091
Craig Topper [Fri, 19 Aug 2022 15:53:43 +0000 (08:53 -0700)]
[RISCV] Add passthru operand to RISCVISD::SETCC_VL.
Use it to the fix a bug in the fceil/ffloor lowerings. We were
setting the passthru to IMPLICIT_DEF before and using a mask
agnostic policy. This means where the incoming bits in
the mask were 0 they could be anything in the outgoing mask. We
want those bits in the outgoing mask to be 0. This means we need to
pass the input mask as the passthru.
This generates worse code because we are unable to allocate the
v0 register to the output due to an earlyclobber constraint. We
probably need a special TIED pseudoinstruction and probably custom
isel since you can't use V0 twice in the input pattern.
Reviewed By: reames
Differential Revision: https://reviews.llvm.org/D132058
Craig Topper [Fri, 19 Aug 2022 15:15:55 +0000 (08:15 -0700)]
[RISCV] Prefer vnsrl.wi v8, v8, 0 over vnsrl.wx v8, v8, x0.
I have a couple data points that some microarchitectures prefer
the immediate 0 over x0. Does anyone know of microarchitectures
where the opposite is true?
Unfortunately, this is different than the vncvt.x.x.w alias
from the spec. Perhaps the alias was poorly chosen if x0 isn't
as optimal as immediate 0 on all microarchitectures.
Reviewed By: reames
Differential Revision: https://reviews.llvm.org/D132041
Alexey Bataev [Mon, 15 Aug 2022 17:26:45 +0000 (10:26 -0700)]
[SLP]Delay vectorization of postponable values for instructions with no users.
SLP vectorizer tries to find the reductions starting the operands of the
instructions with no-users/void returns/etc. But such operands can be
postponable instructions, like Cmp, InsertElement or InsertValue. Such
operands still must be postponed, vectorizer should not try to vectorize
them immediately.
Differential Revision: https://reviews.llvm.org/D131965
Guillaume Chatelet [Thu, 18 Aug 2022 09:39:47 +0000 (09:39 +0000)]
[reland][NFC][libc] rearrange aarch64 memset code to better match new implementation
Differential Revision: https://reviews.llvm.org/D132121
Christopher Bate [Tue, 16 Aug 2022 23:02:18 +0000 (17:02 -0600)]
[mlir][Affine] Add affine.delinearize_index operation
This change adds a new AffineDelinearizeIndexOp to the affine dialect.
The operation accepts an index type as well as a basis (array of index
values) representing how the index should be decomposed into a
multi-index. The decomposition obeys a canonical semantic that treats
the final basis element as "fastest varying" and the first basis element
as "slowest varying". A naive lowering of the operation using a sequence
of AffineApplyOps is given.
RFC was discussed on discourse here: https://discourse.llvm.org/t/rfc-tensor-extracting-slices-from-tensor-collapse-shape/64034
Reviewed By: bondhugula, nicolasvasilache
Differential Revision: https://reviews.llvm.org/D131997
Alexey Bataev [Thu, 2 Jun 2022 14:26:33 +0000 (07:26 -0700)]
[SLP]Cost for a constant buildvector.
In many cases constant buildvector results in a vector load from a
constant/data pool. Need to consider this cost too.
Differential Revision: https://reviews.llvm.org/D126885
Guillaume Chatelet [Fri, 19 Aug 2022 15:02:16 +0000 (15:02 +0000)]
Revert "[NFC][libc] rearrange aarch64 memset code to better match new implementation"
This reverts commit
4d931b6e1e7e6003fa6c68aec9223b06e31c982d.
Wei Yi Tee [Tue, 16 Aug 2022 12:34:42 +0000 (12:34 +0000)]
[clang][dataflow] Debug string for value kinds.
Differential Revision: https://reviews.llvm.org/D131891
Sam McCall [Fri, 19 Aug 2022 13:53:50 +0000 (15:53 +0200)]
[pseudo] Start rules are `_ := start-symbol EOF`, improve recovery.
Previously we were calling glrRecover() ad-hoc at the end of input.
Two main problems with this:
- glrRecover() on two separate code paths is inelegant
- We may have to recover several times in succession (e.g. to exit from
nested scopes), so we need a loop at end-of-file
Having an actual shift action for an EOF terminal allows us to handle
both concerns in the main shift/recover/reduce loop.
This revealed a recovery design bug where recovery could enter a loop by
repeatedly choosing the same parent to identically recover from.
Addressed this by allowing each node to be used as a recovery base once.
Differential Revision: https://reviews.llvm.org/D130550
Guillaume Chatelet [Thu, 18 Aug 2022 09:39:47 +0000 (09:39 +0000)]
[NFC][libc] rearrange aarch64 memset code to better match new implementation
Differential Revision: https://reviews.llvm.org/D132121
Muhammad Usman Shahid [Fri, 19 Aug 2022 14:46:29 +0000 (10:46 -0400)]
Missing tautological compare warnings due to unary operators
The patch mainly focuses on the no warnings for -Wtautological-compare.
It work fine for the positive numbers but doesn't for the negative
numbers. This is because the warning explicitly checks for an
IntegerLiteral AST node, but -1 is represented by a UnaryOperator with
an IntegerLiteral sub-Expr.
Fixes #42918
Differential Revision: https://reviews.llvm.org/D130510
Alexey Bataev [Fri, 19 Aug 2022 12:13:25 +0000 (05:13 -0700)]
[COST][NFC]Introduce OperandValueKind in getMemoryOpCost, NFC.
Added OperandValueKind OpdInfo parameter to getMemoryOpCost functions to
better estimate cost with immediate values.
Part of D126885.
Michał Górny [Fri, 19 Aug 2022 14:22:55 +0000 (16:22 +0200)]
[lldb] [test] Disable new CommunicationTests on Windows
Sponsored by: The FreeBSD Foundation
Timm Bäder [Fri, 19 Aug 2022 14:13:46 +0000 (16:13 +0200)]
[clang][Interp] Rename Integral::T to Integral::ReprT
Just 'T' is a bit generic and causes confusion.
David Sherwood [Fri, 19 Aug 2022 11:11:27 +0000 (12:11 +0100)]
[SVE][LoopVectorize][NFC] Tidy up some tests
Whilst writing a patch to add extra tail-folding RUN lines to
existing tests I noticed a few areas where they can be
cleaned up a little:
1. scalable-reductions.ll: fmin_fast does not mark fcmp as fast.
2. sve-inductions-unusual-types.ll: remove direct references to
SSA variable names.
3. sve-strict-fadd-cost.ll: don't force vector width so we see
costs for different VFs in one go. This will be important for
the follow-on patch.
4. sve-vector-reverse.ll,vector-reverse-mask4.ll: add noalias
keyword to simplify IR.
4. sve-widen-gep.ll,sve-widen-phi.ll: regenerate using script.
These changes will make the subsequent patch adding RUN lines much
easier to review!
Differential Revision: https://reviews.llvm.org/D132219
Simon Tatham [Fri, 19 Aug 2022 14:07:55 +0000 (15:07 +0100)]
[libcxxabi] Fix alignment of pointers returned by fallback_malloc
This aligns the ``heap[]`` array in ``fallback_malloc.cpp`` to ensure
that it can be safely cast to ``heap_node*``, and also adjusts the
allocation algorithm to ensure that every allocated block has the
alignment requested by ``__attribute__((aligned))``, by putting the
block's ``heap_node`` header 4 bytes before an aligned address.
Patch originally by Eric Fiselier: this is an updated version of
D12669, which was never landed.
Reviewed By: ldionne, #libc_abi
Differential Revision: https://reviews.llvm.org/D129842
Timm Bäder [Tue, 9 Aug 2022 06:08:48 +0000 (08:08 +0200)]
[clang][Interp] Implement inv and neg unary operations
Implement negating and inverting values. Also implement
IntegralToBoolean casts so the operations are easier to test.
Differential Revision: https://reviews.llvm.org/D132098
Max Kazantsev [Fri, 19 Aug 2022 13:51:57 +0000 (20:51 +0700)]
[Test] Add test for miscompile described in PR57247
Max Kazantsev [Fri, 19 Aug 2022 13:51:06 +0000 (20:51 +0700)]
Revert "[SCEV] Prove condition invariance via context"
This reverts commit
a3d1fb3b59b473e4f262a05f187de6474b7721e7.
Reverting until investigation of https://github.com/llvm/llvm-project/issues/57247
has concluded.
Ellis Hoag [Fri, 19 Aug 2022 13:46:22 +0000 (06:46 -0700)]
Revert "Revert "[compiler-rt] Fix check for cross-compiling""
This reverts commit
18f6b05e0a5a4197bac787f5522669576c134869.
This is a reland of https://reviews.llvm.org/D132130 which I incorrectly
thought had broken the builds.
Differential Revision: https://reviews.llvm.org/D132230
Haojian Wu [Fri, 19 Aug 2022 13:53:16 +0000 (15:53 +0200)]
[pseudo] NFC, remove redundant ;
Michał Górny [Fri, 19 Aug 2022 13:49:35 +0000 (15:49 +0200)]
[lldb] [test] Add synchronization to TestContinue
Sponsored by: The FreeBSD Foundation
Michał Górny [Fri, 19 Aug 2022 10:30:50 +0000 (12:30 +0200)]
[lldb] [Core] Harmonize Communication::Read() returns w/ thread
Harmonize the status and error values of Communication::Read() when
running with and without read thread. Prior to this change, Read()
would return eConnectionStatusSuccess if read thread was enabled
and the read timed out or reached end-of-file, rather than
the respective states that are returned if read thread was disabled.
Now, it correctly returns eConnectionStatusTimedOut
and eConnectionStatusEndOfFile, and sets the error respectively.
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.llvm.org/D132217
Nikolas Klauser [Fri, 19 Aug 2022 11:08:01 +0000 (13:08 +0200)]
[libc++][NFC] Rename the constexpr macros
This was discussed on Discord with the consensus that we should rename the macros.
Reviewed By: ldionne, Mordante, var-const, avogelsgesang, jloser, #libc
Spies: libcxx-commits
Differential Revision: https://reviews.llvm.org/D131498
Sam McCall [Wed, 13 Jul 2022 14:27:36 +0000 (16:27 +0200)]
[Sema] Tweak diagnostic logic so suppress-in-header logic works in tools too.
Certain idioms are ignored by -Wunused in header files only.
The current "is a header" check assumes that if headers are the main file, we're
building a PCH or a module or something. However in tools we may be parsing the
header in its own right, but still want to treat it as a header.
Fixes https://github.com/clangd/vscode-clangd/issues/360
Differential Revision: https://reviews.llvm.org/D129642
Sam McCall [Fri, 19 Aug 2022 13:15:37 +0000 (15:15 +0200)]
[pseudo] Changes omitted from previous commit
Sam McCall [Fri, 19 Aug 2022 13:00:59 +0000 (15:00 +0200)]
[pseudo] Perform unconstrained reduction prior to recovery.
Our GLR uses lookahead: only perform reductions that might be consumed by the
shift immediately following. However when shift fails and so reduce is followed
by recovery instead, this restriction is incorrect and leads to missing heads.
In turn this means certain recovery strategies can't be made to work. e.g.
```
ns := NAMESPACE { namespace-body } [recover=Skip]
ns-body := namespace_opt
```
When `namespace { namespace {` is parsed, we can recover the inner `ns` (using
the `Skip` strategy to ignore the missing `}`). However this `namespace` will
not be reduced to a `namespace-body` as EOF is not in the follow-set, and so we
are unable to recover the outer `ns`.
This patch fixes this by tracking which heads were produced by constrained
reduce, and discarding and rebuilding them before performing recovery.
This is a prerequisite for the `Skip` strategy mentioned above, though there are
some other limitations we need to address too.
Reviewed By: hokein
Differential Revision: https://reviews.llvm.org/D130523
Sam McCall [Fri, 19 Aug 2022 12:51:36 +0000 (14:51 +0200)]
[clangd] Support hover on __func__ etc (PredefinedExpr)
Expose these as variables as that's what the standard calls them (and D131175).
To make this work, we also fix a bug in SelectionTree: PredefinedExpr has
an implicit/invisible StringLiteral, and SelectionTree should not traverse
implicit things.
Reviewed By: ckandeler
Differential Revision: https://reviews.llvm.org/D132135
Kristof Beyls [Thu, 18 Aug 2022 12:58:05 +0000 (14:58 +0200)]
Add guidance for what to put in community calendar invites
Differential Revision: https://reviews.llvm.org/D132134
Stefan Pintilie [Tue, 16 Aug 2022 16:08:33 +0000 (11:08 -0500)]
[PowerPC] Fix bugs in sign-/zero-extension elimination
This patch fixes the following two bugs in `PPCInstrInfo::isSignOrZeroExtended` helper, which is used from sign-/zero-extension elimination in PPCMIPeephole pass.
- Registers defined by load with update (e.g. LBZU) were identified as already sign or zero-extended. But it is true only for the first def (loaded value) and not for the second def (i.e. updated pointer).
- Registers defined by ORIS/XORIS were identified as already sign-extended. But, it is not true for sign extension depending on the immediate (while it is ok for zero extension).
To handle the first case, the parameter for the helpers is changed from `MachineInstr` to a register number to distinguish first and second defs. Also, this patch moves the initialization of PPCMIPeepholePass to allow mir test case.
Reviewed By: nemanjai
Differential Revision: https://reviews.llvm.org/D40554
Caroline Concatto [Wed, 10 Aug 2022 07:13:15 +0000 (08:13 +0100)]
[Clang][AArch64] Use generic extract/insert vector for svget/svset/svcreate tuples
This patch replaces svget, svset and svcreate aarch64 intrinsics for tuple
types with the generic llvm-ir intrinsics extract/insert vector
Differential Revision: https://reviews.llvm.org/D131547
Nico Weber [Fri, 19 Aug 2022 11:45:06 +0000 (07:45 -0400)]
[gn build] Port
c74c17f37aa9
Nico Weber [Fri, 19 Aug 2022 11:40:45 +0000 (07:40 -0400)]
Pavel Labath [Thu, 4 Aug 2022 09:26:36 +0000 (11:26 +0200)]
[lldb] Use WSAEventSelect for MainLoop polling on windows
This patch switches the MainLoop class to use the WSAEventSelect
mechanism to wait for multiple sockets to become readable. The
motivation for doing that is that this allows us to wait for other kinds
of events as well (as long as they can be converted to WSAEvents). This
will allow us to avoid (abstract away) pipe-based multiplexing
mechanisms in the generic code, since pipes cannot be combined with
sockets on windows.
Since the windows implementation will now look completely different than
the posix (file descriptor-based) implementations, I have split the
MainLoop class into two (MainLoopPosix and MainLoopWindows), with the
common code going into MainLoopBase.
Differential Revision: https://reviews.llvm.org/D131159
Vaibhav Yenamandra [Fri, 19 Aug 2022 11:13:44 +0000 (07:13 -0400)]
Add support for specifying the severity of a SARIF Result
* Extend SarifResult with level property, and allow rule configuration
* Create SarifReportingConfiguration which allow configuring rules with
a default priority, severity and an enable-toggle
* Support for setting the level property[1] of a result.
If unset, it defaults to "warning", which is the result of an empty
default configuration on rules[2]
[1]: https://docs.oasis-open.org/sarif/sarif/v2.1.0/os/sarif-v2.1.0-os.html#_Toc34317648
[2]: https://docs.oasis-open.org/sarif/sarif/v2.1.0/os/sarif-v2.1.0-os.html#_Toc34317855
Differential Revision: https://reviews.llvm.org/D131084
Caroline Concatto [Fri, 19 Aug 2022 11:08:14 +0000 (12:08 +0100)]
[InstCombine] For vector extract when extract vector and insert value type is the same
This patch has implements these optimizations:
extract.vector(insert.vector(Vector, Value, Idx), Idx) --> Value
extract.vector(insert.vector(Vector, Value, InsertIndex), ExtractIndex)
--> extract.vector(Vector, ExtractIndex)
Reviewed By: sdesmalen
Differential Revision: https://reviews.llvm.org/D132137
Dmitry Preobrazhensky [Fri, 19 Aug 2022 11:01:42 +0000 (14:01 +0300)]
[AMDGPU][MC][GFX9][NFC] Split large test file
Split gfx9_dasm_all.txt by instruction encoding.
Differential Revision: https://reviews.llvm.org/D132124
Dmitry Preobrazhensky [Fri, 19 Aug 2022 10:51:55 +0000 (13:51 +0300)]
[AMDGPU][MC][GFX8][NFC] Split large test file
Split gfx8_dasm_all.txt by instruction encoding.
Differential Revision: https://reviews.llvm.org/D132126
Dominik Adamski [Fri, 19 Aug 2022 10:00:02 +0000 (05:00 -0500)]
[NFC][OpenMP] Add test for simd directive with nested loop
This test is follow up of the review: https://reviews.llvm.org/D131402
Co-Author of the patch: Kiran Chandramohan
Reviewed By: kiranchandramohan
Differential Revision: https://reviews.llvm.org/D132215
Caroline Concatto [Thu, 11 Aug 2022 11:09:58 +0000 (12:09 +0100)]
[Clang][AArch64] Replace aarch64_sve_ldN intrinsic by aarch64_sve_ldN.sret
Differential Revision: https://reviews.llvm.org/D131687
Alexander Malkov [Fri, 19 Aug 2022 09:53:27 +0000 (09:53 +0000)]
[clang,flang] Add help text for -fsyntax-only
Fix for the problem with displaying options `-fsyntax-only` in clang and flang-new in help
Fix https://github.com/llvm/llvm-project/issues/57033
Before:
``` $ clang -help | grep syntax
-objcmt-migrate-property-dot-syntax
Enable migration of setter/getter messages to property-dot syntax
```
After:
```
$ clang -help | grep syntax
-fsyntax-only Run the preprocessor, parser and semantic analysis stages
-objcmt-migrate-property-dot-syntax
Enable migration of setter/getter messages to property-dot syntax
```
Reviewed By: vzakhari, awarzynski, MaskRay, alexiprof
Differential Revision: https://reviews.llvm.org/D131808
Simon Pilgrim [Fri, 19 Aug 2022 09:39:30 +0000 (10:39 +0100)]
[CostModel][X86] Split icmp/fcmp cost kinds tests
It's going to be much easier to maintain these tests (and all the check prefix combinations) if we don't mix cost kinds in the same file.
Simon Pilgrim [Fri, 19 Aug 2022 09:28:04 +0000 (10:28 +0100)]
[CostModel][X86] Split select cost kinds tests
It's going to be much easier to maintain these tests (and all the check prefix combinations) if we don't mix cost kinds in the same file.
Simon Pilgrim [Fri, 19 Aug 2022 09:17:44 +0000 (10:17 +0100)]
[CostModel][X86] Add tremont to slm-arith-costs.ll
Tremont uses the Goldmont div/sqrt costs but we weren't testing for them
Simon Pilgrim [Fri, 19 Aug 2022 09:09:28 +0000 (10:09 +0100)]
[CostModel][X86] Cleanup arithmetic test triples
Just specify the triple inside the RUN command (to make i686 support much easier!), and consistently use x86_64-- generic triple
Martin Storsjö [Mon, 8 Aug 2022 21:48:22 +0000 (00:48 +0300)]
[libcxx] [test] Mark the libcxx/selftest/remote-substitutions.sh.cpp test as requiring bash in the executor
It's unclear to me why this wasn't tagged this way already in
87fe0709d4ad7c3128f998f14c66245ebb423e4f / D114612 where the
feature flag executor-has-no-bash was added, as this test did exist
in its current form already at that time.
Differential Revision: https://reviews.llvm.org/D131446
Martin Storsjö [Tue, 9 Aug 2022 08:56:46 +0000 (11:56 +0300)]
[libcxx] [test] Make some threading tests more robust
Increase the timeout tolerance if TEST_IS_EXECUTED_IN_A_SLOW_ENVIRONMENT
is set, similarly to how it's done in a couple other tests.
Use `std::this_thread::yield();` instead of busylooping. When multiple
threads are busylooping, it's plausible that not all threads even get
started running before the timeout runs out.
This makes the threading tests succeed if run in Windows runners on
Github Actions.
Differential Revision: https://reviews.llvm.org/D131483
Martin Storsjö [Tue, 9 Aug 2022 10:51:33 +0000 (13:51 +0300)]
[libcxx] [test] Split the TEST_HAS_SANITIZERS define into TEST_IS_EXECUTED_IN_A_SLOW_ENVIRONMENT
This makes its role clearer. It's plausible that one may want to manually
define TEST_IS_EXECUTED_IN_A_SLOW_ENVIRONMENT when running the tests in
some environments - in particular, it seems to be necessary to use the
higher tolerance timeouts if running the tests on Windows runners
on Github Actions.
Also add the descriptive comment in one file where it was missing.
Differential Revision: https://reviews.llvm.org/D131484
Martin Storsjö [Tue, 9 Aug 2022 08:36:51 +0000 (11:36 +0300)]
[libcxx] [test] Remove --env PATH from the static clangcl config
The PATH is set in order to be able to find the tested DLL at runtime.
When linking statically, it's not necessary to set the PATH.
Setting PATH in the executor has the downside that it clears the
existing path (it's not prepended/appended to it), which means
that the executed tools can't find other tools - which sets the
executor-has-no-bash flag.
By removing the unnecessary setting of PATH, we have a properly
working bash even when wrapped by the executor, which gets rid
of the executor-has-no-bash flag in this test configuration, which
makes 9 more testcases be executed.
Differential Revision: https://reviews.llvm.org/D131481
Carlos Galvez [Fri, 12 Aug 2022 14:53:31 +0000 (14:53 +0000)]
[clang-tidy] Do not trigger cppcoreguidelines-avoid-const-or-ref-data-members on lambda captures
Lambdas are implemented as regular classes internally,
and the captured variables end up as members there.
Do not diagnose those - the check should cover only
regular classes and structs.
Differential Revision: https://reviews.llvm.org/D131780
jacquesguan [Fri, 19 Aug 2022 07:59:27 +0000 (15:59 +0800)]
[LangRef][VP] Fix typo.
Differential Revision: https://reviews.llvm.org/D132210
Iain Sandoe [Thu, 18 Aug 2022 13:34:50 +0000 (14:34 +0100)]
[clang][Modules] Fix a regression in handling missing framework headers.
The commit of
af2d11b1d5c1508b506825df460656e0151cd3b0 missed a case where
the value of a suggested module needed to be reset to nullptr. Fixed thus
and added a testcase to cover the circumstance.
Archibald Elliott [Wed, 17 Aug 2022 17:11:58 +0000 (18:11 +0100)]
[AArch64][GISel] Lower llvm.prefetch
This change adds support for lowering llvm.prefetch directly using
GlobalISel. Currently, llvm.prefetch falls back to SelectionDAG.
This Change:
- Adds an AArch64-specific G_PREFETCH generic instruction, to be used
where AArch64ISD::PREFETCH is used in SelectionDAG.
- Adds the GINodeEquiv so patterns are translated over to GlobalISel
automatically.
- Corrects the AArch64Prefetch patterns to use a target immediate, which
is needed to get the patterns to translate across correctly.
- Translates the SelectionDAG legalisation of the prefetch intrinsic
into the corresponding GlobalISel legalisation.
Differential Revision: https://reviews.llvm.org/D132043
Archibald Elliott [Wed, 17 Aug 2022 17:05:12 +0000 (18:05 +0100)]
[IR] Update llvm.prefetch to match docs
The current llvm.prefetch intrinsic docs state "The rw, locality and
cache type arguments must be constant integers."
This change:
- Makes arg 3 (cache type) an ImmArg
- Improves the verifier error messages to reference the incorrect
argument.
- Fixes two tests which contradict the docs.
This is needed as the lowering to GlobalISel is different for ImmArgs
compared to other constants. The non-ImmArgs create a G_CONSTANT MIR
instruction, the for ImmArgs the constant is put directly on the
intrinsic's MIR instruction as an immediate.
Differential Revision: https://reviews.llvm.org/D132042
Michał Górny [Fri, 19 Aug 2022 08:07:37 +0000 (10:07 +0200)]
[lldb] [test] Remove test_step_multiprocess, it's unreliable
Sponsored by: The FreeBSD Foundation
Max Kazantsev [Fri, 19 Aug 2022 07:48:39 +0000 (14:48 +0700)]
[Test] Regenerate tests using update_tests.py
Michał Górny [Fri, 19 Aug 2022 07:39:56 +0000 (09:39 +0200)]
[lldb] [test] Skip step packet test on non-amd64
Sponsored by: The FreeBSD Foundation
Siva Chandra Reddy [Fri, 19 Aug 2022 07:15:08 +0000 (07:15 +0000)]
[libc] Add more headers to the linux x86_64 and aarch64 configs.
Siva Chandra Reddy [Fri, 19 Aug 2022 07:01:47 +0000 (07:01 +0000)]
[libc][NFC] Change the libc fullbuild target name to "libc".
Chuanqi Xu [Fri, 19 Aug 2022 07:02:54 +0000 (15:02 +0800)]
[NFC] Add tests for MemorySSA shouldn't contain debug instructions
Extracted from D130153
Michał Górny [Fri, 12 Aug 2022 10:37:18 +0000 (12:37 +0200)]
[lldb] [gdb-remote] Include PID in vCont packets if multiprocess
Try to always send vCont packets and include the PID in them if running
multiprocess. This is necessary to ensure that with the upcoming full
multiprocess support always resumes the correct process without having
to resort to the legacy Hc packets.
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.llvm.org/D131758
Joey Watts [Fri, 19 Aug 2022 06:57:12 +0000 (07:57 +0100)]
[clang-tidy] Improve modernize-use-emplace check
This patch improves the modernize-use-emplace check by adding support for
detecting inefficient invocations of the `push` and `push_front` methods on
STL-style containers and replacing them with their `emplace`-style equivalent.
Fixes #56996.
Reviewed By: njames93
Differential Revision: https://reviews.llvm.org/D131623
Xiang Li [Thu, 18 Aug 2022 06:40:13 +0000 (23:40 -0700)]
[DirectX] Fix crash on ShuffleVectorInst in DXILBitcodeWriter
Use onstantDataSequential::getElementType to allow vector type.
Use getShuffleMaskForBitcode to get ShuffleMask.
Use old format for alloca Align.
Reviewed By: beanz
Differential Revision: https://reviews.llvm.org/D132102
Chuanqi Xu [Fri, 19 Aug 2022 06:41:25 +0000 (14:41 +0800)]
[NFC] Add [[maybe_unused]] to avoid warning in gcc9
GCC9 may issue warning for the 'unused' parameters in if constexpr.
This commit try to fix it by adding the [[maybe_unused]] attribute.
Valentin Clement [Fri, 19 Aug 2022 06:40:09 +0000 (08:40 +0200)]
[flang] Shift argument attributes when result operand is inserted
The TargetRewrite pass convert the signature of the function.
In some cases it adds operands to the function to hanlde the result of it.
This patch makes sure the argument attributes present before the conversion
are replaced with the correct arguments after the conversion is performed.
Depends D132113
Reviewed By: vdonaldson
Differential Revision: https://reviews.llvm.org/D132114
Max Kazantsev [Fri, 19 Aug 2022 06:24:35 +0000 (13:24 +0700)]
[Test] Remove addrspace1 ptr to not confuse alive2
addrspace here is not import for the test itself.
Timm Bäder [Thu, 11 Aug 2022 09:22:00 +0000 (11:22 +0200)]
[clang] Improve diagnostics for uninitialized constexpr variables
Instead of complaining about default initialization, tell users that
constexpr variables need to be initialized by a constant expression.
Differential Revision: https://reviews.llvm.org/D131662
Fabian Parzefall [Fri, 19 Aug 2022 04:51:51 +0000 (21:51 -0700)]
[BOLT] Insert EH trampolines for multiple fragments
This patch adds exception handling trampolines when a function is split
into more than two fragments. Trampolines are tracked per-fragment, such
that they can be removed if splitting is reversed.
Reviewed By: rafauler
Differential Revision: https://reviews.llvm.org/D132048
Fabian Parzefall [Fri, 19 Aug 2022 04:51:22 +0000 (21:51 -0700)]
[BOLT] Update buildCallGraph to check for split blocks
Use isSplit() instead of isCold() when building the call graph and
update parameter names to reflect this.
Reviewed By: rafauler
Differential Revision: https://reviews.llvm.org/D132047
Fabian Parzefall [Fri, 19 Aug 2022 04:51:01 +0000 (21:51 -0700)]
[BOLT] Add randomN split strategy
This adds a strategy to split functions into a random number of
fragments at randomly chosen split points.
Reviewed By: rafauler
Differential Revision: https://reviews.llvm.org/D130647
Fabian Parzefall [Fri, 19 Aug 2022 04:50:35 +0000 (21:50 -0700)]
[BOLT] Add split all blocks strategy
This adds a function splitting strategy that splits each outlineable
basic block into its own fragment. This is exposed through a new command
line option `--split-strategy`.
Reviewed By: rafauler
Differential Revision: https://reviews.llvm.org/D129827
Fabian Parzefall [Fri, 19 Aug 2022 04:48:19 +0000 (21:48 -0700)]
[BOLT] Generate sections for multiple fragments
This patch adds support to generate any number of sections that are
assigned to fragments of functions that are split more than two-way.
With this, a function's *nth* split fragment goes into section
`.text.cold.n`.
This also changes `FunctionLayout::erase` to make sure, that there are
no empty fragments at the end of the function. This sometimes happens
when blocks are erased from the function. To avoid creating symbols
pointing to these fragments, they need to be removed.
Reviewed By: rafauler
Differential Revision: https://reviews.llvm.org/D130521
Fabian Parzefall [Fri, 19 Aug 2022 04:40:00 +0000 (21:40 -0700)]
[BOLT] Make exception handling fragment aware
This adds basic fragment awareness in the exception handling passes and
generates the necessary symbols for fragments.
Reviewed By: rafauler
Differential Revision: https://reviews.llvm.org/D130520
Fabian Parzefall [Fri, 19 Aug 2022 04:26:18 +0000 (21:26 -0700)]
[BOLT] Support passing fragments to code emission
This changes code emission such that it can emit specific function
fragments instead of scanning all basic blocks of a function and just
emitting those that are hot or cold.
To implement this, `FunctionLayout` explicitly distinguishes the "main"
fragment (i.e. the one that contains the entry block and is associated
with the original symbol) from "split" fragments. Additionally,
`BinaryFunction` receives support for multiple cold symbols - one for
each split fragment.
Reviewed By: rafauler
Differential Revision: https://reviews.llvm.org/D130052
David Blaikie [Fri, 19 Aug 2022 04:01:42 +0000 (04:01 +0000)]
Revert "[tsan] Keep thread/stack for closed FD"
Test is flaky.
This reverts commit
e9c5bde88ea2e35fadb15c5e5a1d2cb583fd0772.
David Blaikie [Fri, 19 Aug 2022 03:33:01 +0000 (03:33 +0000)]
Simplify RAV isSameMethod with constexpr if
Owing to the large number of instantiations of this function, this small
change has a small but meaningful difference on the total size of
(especially a debug) build of clang at -O0:
```
FILE SIZE VM SIZE
-------------- --------------
+0.9% +96.9Ki +0.9% +96.9Ki .data.rel.ro
+0.7% +96.7Ki +0.7% +96.7Ki .rela.dyn
+0.0% +18.3Ki +0.0% +18.3Ki .rodata
+0.0% +324 [ = ] 0 [2 Others]
-0.2% -392 -0.2% -392 .gnu.version
-0.0% -441 [ = ] 0 .debug_abbrev
-0.1% -980 -0.1% -980 .gnu.hash
-0.2% -1.53Ki -0.2% -1.53Ki .hash
-0.2% -4.59Ki -0.2% -4.59Ki .dynsym
-0.1% -10.5Ki [ = ] 0 .debug_rnglists
-0.6% -59.0Ki -0.6% -59.0Ki .dynstr
-0.2% -191Ki [ = ] 0 .debug_str_offsets
-3.0% -233Ki -3.0% -233Ki .eh_frame_hdr
-0.7% -244Ki [ = ] 0 .debug_addr
-2.9% -699Ki [ = ] 0 .symtab
-0.6% -884Ki [ = ] 0 .debug_line
-3.0% -932Ki -3.0% -932Ki .eh_frame
-1.0% -1.48Mi -1.0% -1.48Mi .text
-0.6% -2.75Mi [ = ] 0 .debug_info
-7.3% -8.61Mi [ = ] 0 .strtab
-7.3% -17.2Mi [ = ] 0 .debug_str
-2.4% -33.0Mi -0.9% -2.47Mi TOTAL
```
If anyone's got other ideas for how to reduce this further - it's not
especially important, I just came across it while investigating a debug
info size regression, but thought it was interesting enough to poke
around at.
Craig Topper [Fri, 19 Aug 2022 03:42:59 +0000 (20:42 -0700)]
[RISCV] Copy SDNodeFlags in lowerToScalableOp.
Reviewed By: arcbbb
Differential Revision: https://reviews.llvm.org/D132177
Craig Topper [Fri, 19 Aug 2022 03:42:45 +0000 (20:42 -0700)]
[RISCV] Copy SDNodeFlags in doPeepholeMaskedRVV and doPeepholeMergeVVMFold
Especially the NoFPExcept flag for FP.
Reviewed By: fakepaper56
Differential Revision: https://reviews.llvm.org/D132173
Tung D. Le [Fri, 19 Aug 2022 02:57:20 +0000 (08:27 +0530)]
[MLIR][normalize-memrefs] Non-normalizable operations with identity map layouts do not block normalization of the entire function
The current approach is convervative in which whenever there is a
non-normalizable operations in a function will the function be labelled
as non-normalizable. It means it requires that all operations must have
MemRefsNormalizable trait.
This patch relaxes the requirement that if the memref map layouts of a
non-normalizable operation are identity, this operation does not block
the normalization of the other operations in the same function.
Reviewed By: bondhugula
Differential Revision: https://reviews.llvm.org/D125854
John Ericson [Fri, 19 Aug 2022 02:44:46 +0000 (22:44 -0400)]
Revert "[cmake] Use `CMAKE_INSTALL_LIBDIR` too"
This reverts commit
f7a33090a91015836497c75f173775392ab0304d.
Unfortunately this causes a number of failures that didn't show up in my
local build.
Yonghong Song [Thu, 18 Aug 2022 15:01:45 +0000 (08:01 -0700)]
[Clang][BPF] Support record argument with direct values
Currently, record arguments are always passed by reference by allocating
space for record values in the caller. This is less efficient for
small records which may take one or two registers. For example,
for x86_64 and aarch64, for a record size up to 16 bytes, the record
values can be passed by values directly on the registers.
This patch added BPF support of record argument with direct values
for up to 16 byte record size. If record size is 0, that record
will not take any register, which is the same behavior for x86_64
and aarch64. If the record size is greater than 16 bytes, the
record argument will be passed by reference.
Differential Revision: https://reviews.llvm.org/D132144