Paul Walker [Sat, 18 Jun 2022 13:55:31 +0000 (14:55 +0100)]
[NFC][SVE] Simplify SUBR_ZI isel patterns.
Differential Revision: https://reviews.llvm.org/D128199
Paul Walker [Sat, 18 Jun 2022 12:36:22 +0000 (13:36 +0100)]
[SVE] Lower "unpredicated" sabd/uabd intrinsics to ISD::ABDS/U.
This enables an existing transformation that when combined with an
add will emit saba/uaba instructions.
Differential Revision: https://reviews.llvm.org/D128198
Daniel Bertalan [Tue, 21 Jun 2022 20:40:27 +0000 (16:40 -0400)]
Reland "[lld-macho] Show source information for undefined references"
The error used to look like this:
ld64.lld: error: undefined symbol: _foo
>>> referenced by /path/to/bar.o:(symbol _baz+0x4)
If DWARF line information is available, we now show where in the source
the references are coming from:
ld64.lld: error: unreferenced symbol: _foo
>>> referenced by: bar.cpp:42 (/path/to/bar.cpp:42)
>>> /path/to/bar.o:(symbol _baz+0x4)
The reland is identical to the first time this landed. The fix was in D128294.
This reverts commit
0cc7ad417585b3185c32e395cc5e6cf082a347af.
Differential Revision: https://reviews.llvm.org/D128184
Aart Bik [Tue, 21 Jun 2022 21:13:14 +0000 (14:13 -0700)]
[mlir][sparse] refine bufferization allocation lowering
Marking bufferization allocation operation as invalid
during sparse lowering is too strict, since dense and
sparse allocation can co-exist. This revision refines
the lowering with a dynamic type check.
Reviewed By: bixia
Differential Revision: https://reviews.llvm.org/D128305
LLVM GN Syncbot [Tue, 21 Jun 2022 21:53:32 +0000 (21:53 +0000)]
[gn build] Port
79fbee3cc562
Anubhab Ghosh [Tue, 21 Jun 2022 21:31:40 +0000 (23:31 +0200)]
Re-apply "[JITLink][Orc] Add MemoryMapper interface with InProcess implementation"
[JITLink][Orc] Add MemoryMapper interface with InProcess implementation
MemoryMapper class takes care of cross-process and in-process address space
reservation, mapping, transferring content and applying protections.
Implementations of this class can support different ways to do this such
as using shared memory, transferring memory contents over EPC or just
mapping memory in the same process (InProcessMemoryMapper).
The original patch landed with commit
6ede65205073d3cf6b1ed4d101e66eae3e0fc8e6
It was reverted temporarily in commit
6a4056ab2ada0046ff97a55a5fb34c2c59504fd1
Reviewed By: sgraenitz, lhames
Differential Revision: https://reviews.llvm.org/D127491
Philip Reames [Tue, 21 Jun 2022 21:42:50 +0000 (14:42 -0700)]
[RISCV] Add cost model test coverage for loads and stores (both fixed and scalable)
Vasileios Porpodas [Thu, 12 May 2022 23:25:34 +0000 (16:25 -0700)]
[SLP][NFC] Precommit test for a followup patch that improves reordering for addsubs
Differential Revision: https://reviews.llvm.org/D126091
Florian Hahn [Tue, 21 Jun 2022 21:34:03 +0000 (23:34 +0200)]
[ConstraintElimination] Add tests for transferring info between systems.
Martin Boehme [Fri, 17 Jun 2022 20:41:53 +0000 (22:41 +0200)]
[Clang] Fix compile time regression caused by D126061.
As noted by @nikic, D126061 causes a compile time regression of about
0.5% on -O0 builds:
http://llvm-compile-time-tracker.com/compare.php?from=
7acc88be0312c721bc082ed9934e381d297f4707&to=
8c7b64b5ae2a09027c38db969a04fc9ddd0cd6bb&stat=instructions
This happens because, in a number of places, D126061 creates an
additional local variable of type `ParsedAttributes`. In the large
majority of cases, no attributes are added to this `ParsedAttributes`,
but it turns out that creating an empty `ParsedAttributes`, then
destroying it is a relatively expensive operation.
The reason for this is because `AttributePool` uses a `TinyPtrVector` as
its underlying vector class, and the technique that `TinyPtrVector`
employs to achieve its extreme memory frugality makes the `begin()` and
`end()` member functions relatively slow. The `ParsedAttributes`
destructor iterates over the attributes in its `AttributePool`, and this
is a relatively expensive operation because `TinyPtrVector`'s `begin()` and
`end()` are relatively slow.
The fix for this is to replace `TinyPtrVector` in `ParsedAttributes` and
`AttributePool` with `SmallVector`. `ParsedAttributes` and
`AttributePool` objects are only ever allocated on the stack (they're
not part of the AST), and only a small number of these objects are live
at any given time, so they don't need the extreme memory frugality of
`TinyPtrVector`.
I've confirmed with valgrind that this patch does not increase heap
memory usage, and it actually makes compiles slightly faster than they
were before D126061.
Here are instruction count measurements (obtained with callgrind)
running `clang -c MultiSource/Applications/JM/lencod/parsetcommon.c`
(a file from llvm-test-suite that exhibited a particularly large
compile-time regression):
7acc88be0312c721bc082ed9934e381d297f4707
(baseline one commit before D126061 landed)
102,280,068 instructions
8c7b64b5ae2a09027c38db969a04fc9ddd0cd6bb
(the patch that landed D126061)
103,289,454 instructions
(+0.99% relative to baseline)
This patch applied onto
8c7b64b5ae2a09027c38db969a04fc9ddd0cd6bb
101,117,584 instructions
(-1.14% relative to baseline)
Reviewed By: aaron.ballman
Differential Revision: https://reviews.llvm.org/D128097
Simon Pilgrim [Tue, 21 Jun 2022 21:07:28 +0000 (22:07 +0100)]
[DAG] visitTRUNCATE - move TRUNCATE(ADDE/ADDCARRY) folds to switch statement handling the other binops. NFC.
Craig Topper [Tue, 21 Jun 2022 20:48:12 +0000 (13:48 -0700)]
[RISCV] Move the passthru operand for RISCVISD::VRGATHER*_VL nodes. NFC
Put it before the VL instead of as the first operand. I want to add
passthru to more operands, but the commutable ones like VADD_VL
require the commutable operands to be operand 0 and 1. So we can't
have the passthru as operand 0 for those.
Philip Reames [Tue, 21 Jun 2022 20:54:53 +0000 (13:54 -0700)]
LoopVect, tests] Add some basic coverage for scalable costing of scatter/gather patterns on RISCV
This just adds some very basic vectorizer testing with both fixed and scalable vectorization enabled.
Philip Reames [Tue, 21 Jun 2022 20:22:13 +0000 (13:22 -0700)]
[LoopVect, tests] Add some basic coverage for scalable costing on RISCV
This just adds some very basic vectorizer testing with both fixed and scalable vectorization enabled. For context, I just yesterday fixed a crash in costing of the splat_ptr example - see bbf3fd.
Arthur Eubanks [Tue, 21 Jun 2022 20:26:40 +0000 (13:26 -0700)]
[polly] #include <algorithm>
For the usage of std::max in the header.
Speculative fix for
https://ci.chromium.org/ui/p/fuchsia/builders/toolchain.ci/clang-windows-x64/
b8810806780048763729/overview
reported in https://reviews.llvm.org/D125263.
Simon Pilgrim [Tue, 21 Jun 2022 20:23:04 +0000 (21:23 +0100)]
[DAG] Remove SelectionDAG::GetDemandedBits DemandedElts variant. NFC.
We're slowly removing SelectionDAG::GetDemandedBits and replacing it with SimplifyMultipleUseDemandedBits, we no longer have any uses for the vector demanded elt variant.
Xing Xue [Tue, 21 Jun 2022 20:18:17 +0000 (16:18 -0400)]
[libc++abi][AIX] Use _LIBCXXABI_FUNC_VIS for exported routines
Summary:
This patch adds _LIBCXXABI_FUNC_VIS to the definitions of the personality and helper routines for the state table based EH, now that the support of the visibility attribute is being added to AIX Clang. Currently an export list is generated in the absence of the visibility attribute support downstream.
Reviewed by: MaskRay, daltenty
Differential Revision: https://reviews.llvm.org/D126915
Fangrui Song [Tue, 21 Jun 2022 20:13:31 +0000 (13:13 -0700)]
Revert "Revert "[CMake] Enable LLVM_ENABLE_PER_TARGET_RUNTIME_DIR by default on Linux""
This reverts commit
90c9d41c8acc34fb68958f373eb23f53c1b1c840.
Keeping arm* LLVM_ENABLE_PER_TARGET_RUNTIME_DIR=on is sufficient to work around
LLVM_ENABLE_RUNTIMES not working well with arm. It is more important for
LLVM_ENABLE_PROJECTS and LLVM_ENABLE_RUNTIMES to match in the file hierarchy.
Jason Molenda [Tue, 21 Jun 2022 19:57:42 +0000 (12:57 -0700)]
Roll back Michał's changes to debugserver, not meant for there
Michał's change in https://reviews.llvm.org/D127193 did a search &
replace for a pattern that also appears in debugserver, but it
shouldn't be done there.
Florian Hahn [Tue, 21 Jun 2022 19:54:37 +0000 (21:54 +0200)]
[LV] Add new block to place recurrence splice, if needed.
In some cases, a recurrence splice instructions needs to be inserted
between to regions, for example if the regions get re-arranged during
sinking.
Fixes #56146.
Daniel Bertalan [Tue, 21 Jun 2022 19:47:45 +0000 (15:47 -0400)]
[ADT] [lld-macho] Check for end iterator deref in filter_iterator_base
If ld64.lld was supplied an object file that had a `__debug_abbrev` or
`__debug_str` section, but didn't have any compile unit DIEs in
`__debug_info`, it would dereference an iterator pointing to the empty
array of DIEs. This underlying issue started causing segmentation faults
when parsing for `__debug_info` was addded in D128184. That commit was
reverted, and this one fixes the invalid dereference to allow relanding
it.
This commit adds an assertion to `filter_iterator_base`'s dereference
operators to catch bugs like this one.
Ran check-llvm, check-clang and check-lld.
Differential Revision: https://reviews.llvm.org/D128294
Nico Weber [Tue, 21 Jun 2022 19:12:02 +0000 (15:12 -0400)]
[lld/mac] Replace while loop with for loop
No behavior change. In preparation for using a parallelFor() here.
Differential Revision: https://reviews.llvm.org/D128295
Matt Arsenault [Wed, 15 Jun 2022 17:56:15 +0000 (13:56 -0400)]
AMDGPU: Regenerate test checks
Matt Arsenault [Sat, 11 Jun 2022 15:12:36 +0000 (11:12 -0400)]
AMDGPU: Don't use branches to entry block in test
This created a weird loop making the tested registers live out of the
block, which I don't think is relevant to the purpose of the
tests. This caused regressions when the validity queries are changed
to use tests based whether the use instruction was a kill. If the
register was live out for the loop, it was still live.
I guess we could still do this in a narrow case where the value loops
back, but that's most a pointlessly complex case to handle.
Joe Nash [Mon, 20 Jun 2022 13:50:17 +0000 (09:50 -0400)]
[AMDGPU] gfx11 Remove SDWA from shuffle_vector ISel
gfx11 does not have SDWA
Reviewed By: #amdgpu, rampitec
Differential Revision: https://reviews.llvm.org/D128208
Craig Topper [Tue, 21 Jun 2022 18:54:56 +0000 (11:54 -0700)]
[RISCV] Remove true_mask patterns for VRGATHERE16..
After adding it to the table so the post-isel peephole can handle it.
Craig Topper [Tue, 21 Jun 2022 18:40:53 +0000 (11:40 -0700)]
[RISCV] Remove true_mask patterns for VRGATHER.
These can be handled by the post-isel peephole.
Nico Weber [Tue, 21 Jun 2022 17:49:26 +0000 (13:49 -0400)]
[lld/mac] On Apple systems, call CC_SHA256 from libSystem
It's in libSystem, so it doesn't bring in any new deps, and it's
currently much faster than LLVM's current SHA256 implementation.
Makes linking (arm64) Chromium Framework with ld64.lld 17% faster.
See also PR56121.
No behavior change.
Differential Revision: https://reviews.llvm.org/D128290
Heejin Ahn [Mon, 20 Jun 2022 05:48:48 +0000 (22:48 -0700)]
[DSE] Don't remove nounwind invokes
For non-mem-intrinsic and non-lifetime `CallBase`s, the current
`isRemovable` function only checks if the `CallBase` 1. has no uses 2.
will return 3. does not throw:
https://github.com/llvm/llvm-project/blob/
80fb7823367c1d105fcbc8f21b69205a0d68c859/llvm/lib/Transforms/Scalar/DeadStoreElimination.cpp#L1017
But we should also exclude invokes even in case they don't throw,
because they are terminators and thus cannot be removed. While it
doesn't seem to make much sense for `invoke`s to have an `nounwind`
target, this kind of code can be generated and is also valid bitcode.
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D128224
Haojian Wu [Tue, 21 Jun 2022 18:52:07 +0000 (20:52 +0200)]
Fix an unused-variable warning in release build, NFC.
Valentin Clement [Tue, 21 Jun 2022 18:42:43 +0000 (20:42 +0200)]
[flang] Add more diagnostics to fir.coordinate_of
Add more diagnostics to fir.coordinate_of to provide better checking
that the IR is sane.
This patch is part of the upstreaming effort from fir-dev branch.
Reviewed By: PeteSteinfeld
Differential Revision: https://reviews.llvm.org/D128255
Co-authored-by: Eric Schweitz <eschweitz@nvidia.com>
Nikolas Klauser [Tue, 21 Jun 2022 18:26:43 +0000 (20:26 +0200)]
[libc++] Pass -fno-modules to clang-tidy instead of disabling it for the modules build
Reviewed By: ldionne, #libc
Spies: libcxx-commits
Differential Revision: https://reviews.llvm.org/D128053
Nikolas Klauser [Wed, 15 Jun 2022 08:21:19 +0000 (10:21 +0200)]
[libc++][clang-tidy] Enable bugprone-use-after-move and explicitly list all used checks
Reviewed By: #libc, ldionne
Spies: aheejin, libcxx-commits, xazax.hun
Differential Revision: https://reviews.llvm.org/D127896
Michał Górny [Tue, 21 Jun 2022 18:08:06 +0000 (20:08 +0200)]
[lldb] [llgs] Attempt to fix LLGS tests on Windows
Sponsored by: The FreeBSD Foundation
Nico Weber [Tue, 21 Jun 2022 17:46:40 +0000 (13:46 -0400)]
[lld/mac] Extract a sha256() function
No behavior change.
Differential Revision: https://reviews.llvm.org/D128289
Martin Sebor [Tue, 21 Jun 2022 17:31:58 +0000 (11:31 -0600)]
[InstCombine] handle subobjects of constant aggregates
Remove the known limitation of the library function call folders to only
work with top-level arrays of characters (as per the TODO comment in
the code) and allows them to also fold calls involving subobjects of
constant aggregates such as member arrays.
Michał Górny [Wed, 8 Jun 2022 11:09:08 +0000 (13:09 +0200)]
[lldb] [llgs] Add a test for detach-all packet
Add a test verifying that plain 'D' packet correctly detaches all
processes.
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.llvm.org/D127291
Michał Górny [Wed, 8 Jun 2022 10:52:52 +0000 (12:52 +0200)]
[lldb] [llgs] Refactor fork/vfork tests, verify state
Refactor the fork and vfork tests to reuse the code better, avoid
unnecessary regexps and avoid unnecessary conversions between
hex-strings and integers.
Verify the server state after detaching. In particular, verify that
the detached process' PID/TID pair is no longer valid,
and that the correct process remains running.
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.llvm.org/D127290
Michał Górny [Mon, 20 Jun 2022 14:50:21 +0000 (16:50 +0200)]
[lldb] [llgs] Fix signo sent with fork/vfork/vforkdone events
Fix ThreadStopInfo struct to include the signal number for all events.
Since signo was not included in the details for fork, vfork
and vforkdone stops, the code incidentally referenced the wrong union
member, resulting in wrong signo being sent.
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.llvm.org/D127193
Michał Górny [Tue, 21 Jun 2022 09:55:48 +0000 (11:55 +0200)]
[lldb] [MainLoop] Support "pending callbacks", to be called once
Support adding a "pending callback" to the main loop, that will be
called once after all the pending events are processed. This can be
e.g. to defer destroying the process instance until its exit is fully
processed, as suggested in D127500.
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.llvm.org/D128253
Chris Bieneman [Tue, 14 Jun 2022 21:48:30 +0000 (16:48 -0500)]
[HLSL] Support HLSL vector initializers
In HLSL vectors are ext_vectors in all respects except that they
support a constructor style syntax for initializing vectors. This
change adds a translation of vector constructor arguments into
initializer lists.
This supports two oddities of HLSL syntax:
(1) HLSL vectors support constructor syntax
(2) HLSL vectors are expanded to constituate components in constructors
Reviewed By: aaron.ballman
Differential Revision: https://reviews.llvm.org/D127802
Mahesh Ravishankar [Tue, 21 Jun 2022 16:59:26 +0000 (16:59 +0000)]
[mlir][TilingInterface] Enable tile and fuse using TilingInterface.
This patch implements tile and fuse transformation for ops that
implement the tiling interface. To do so,
- `TilingInterface` needs a new method that generates a tiled
implementation of the operation based on the tile of the result
needed.
- A pattern is added that replaces a `tensor.extract_slice` whose
source is defined by an operation that implements the
`TilingInterface` with a tiled implementation that produces the
extracted slice in-place (using the method added to
`TilingInterface`).
- A pattern is added that takes a sequence of operations that
implement the `TilingInterface` (for now `LinalgOp`s), tiles the
consumer, and greedily fuses its producers iteratively.
Differential Revision: https://reviews.llvm.org/D127809
Alexey Bataev [Tue, 21 Jun 2022 17:04:39 +0000 (10:04 -0700)]
[SLP][NFC]Fix a warning in a comparison, NFC.
Fixed signedness warning.
Michał Górny [Tue, 12 Apr 2022 14:21:09 +0000 (16:21 +0200)]
[lldb] [llgs] Implement non-stop style stop notification packets
Implement the support for %Stop asynchronous notification packet format
in LLGS. This does not implement full support for non-stop mode for
threaded programs -- process plugins continue stopping all threads
on every event. However, it will be used to implement asynchronous
events in multiprocess debugging.
The non-stop protocol is enabled using QNonStop packet. When it is
enabled, the server uses notification protocol instead of regular stop
replies. Since all threads are always stopped, notifications are always
generated for all active threads and copied into stop notification
queue.
If the queue was empty, the initial asynchronous %Stop notification
is sent to the client immediately. The client needs to (eventually)
acknowledge the notification by sending the vStopped packet, in which
case it is popped from the queue and the stop reason for the next thread
is reported. This continues until notification queue is empty again,
in which case an OK reply is sent.
Asychronous notifications are also used for vAttach results and program
exits. The `?` packet uses a hybrid approach -- it returns the first
stop reason synchronously, and exposes the stop reasons for remaining
threads via vStopped queue.
The change includes a test case for a program generating a segfault
on 3 threads. The server is expected to generate a stop notification
for the segfaulting thread, along with the notifications for the other
running threads (with "no stop reason"). This verifies that the stop
reasons are correctly reported for all threads, and that notification
queue works.
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.llvm.org/D125575
Mark de Wever [Mon, 30 May 2022 16:38:27 +0000 (18:38 +0200)]
[libc++][CI] Updates GCC to version 12.
Reviewed By: ldionne, philnik, #libc, #libc_abi
Differential Revision: https://reviews.llvm.org/D126667
Mahesh Ravishankar [Tue, 21 Jun 2022 16:56:10 +0000 (16:56 +0000)]
Revert "[mlir][TilingInterface] Enable tile and fuse using TilingInterface."
This reverts commit
ea75511319d9dff8c38c8794c3949c40b63a38d7 due to build failures.
Argyrios Kyrtzidis [Tue, 21 Jun 2022 16:46:06 +0000 (09:46 -0700)]
[Support/BLAKE3] CMake: Check for `IS_X64` or `CMAKE_OSX_ARCHITECTURES` before adding the assembly files
This should fix `clang-ppc64-aix` builder (https://lab.llvm.org/buildbot/#/builders/214)
John Regehr [Tue, 21 Jun 2022 15:01:42 +0000 (09:01 -0600)]
in the absense of the -max-pass-iterations command line options, make
llvm-reduce run its full pass sequence up to 5 times, instead of just
once
Differential Revision: https://reviews.llvm.org/D128284
Mahesh Ravishankar [Mon, 13 Jun 2022 23:24:31 +0000 (23:24 +0000)]
[mlir][TilingInterface] Enable tile and fuse using TilingInterface.
This patch implements tile and fuse transformation for ops that
implement the tiling interface. To do so,
- `TilingInterface` needs a new method that generates a tiled
implementation of the operation based on the tile of the result
needed.
- A pattern is added that replaces a `tensor.extract_slice` whose
source is defined by an operation that implements the
`TilingInterface` with a tiled implementation that produces the
extracted slice in-place (using the method added to
`TilingInterface`).
- A pattern is added that takes a sequence of operations that
implement the `TilingInterface` (for now `LinalgOp`s), tiles the
consumer, and greedily fuses its producers iteratively.
Differential Revision: https://reviews.llvm.org/D127809
serge-sans-paille [Thu, 19 May 2022 12:27:16 +0000 (14:27 +0200)]
[Scalarizer] No need to gather a scattered extracted element
ExtractElement does not produce a vector out of a vector, so there's no need to
call a gather once done.
Fix #54469
Credits to npopov@redhat.com for the original approach.
Differential Revision: https://reviews.llvm.org/D126012
Matheus Izvekov [Sat, 18 Jun 2022 02:05:45 +0000 (04:05 +0200)]
[NFC] clang: Implement tests for PR56099
Signed-off-by: Matheus Izvekov <mizvekov@gmail.com>
Differential Revision: https://reviews.llvm.org/D128112
Arthur Eubanks [Sun, 19 Jun 2022 19:20:11 +0000 (12:20 -0700)]
Reland [GlobalOpt] Preserve CFG analyses
The only place we modify the CFG is when calling
removeUnreachableBlocks(), so insert a callback there which invalidates
analyses for that function (or recomputes DT in the legacy PM).
We may delete functions, make sure to clear analyses for those
functions. (this was missed in the original revision)
Small compile time wins across the board:
https://llvm-compile-time-tracker.com/compare.php?from=
f444ea8ce0aaaa5ec1a4129809389da15cc41396&to=
698f41f4fc26cbf1006ed5d88e9d658edfc5b749&stat=instructions
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D128145
Krzysztof Drewniak [Thu, 16 Jun 2022 14:38:56 +0000 (14:38 +0000)]
[gdb-scripts] Add to_string methods to printer implementations
Some GDB versions require all prettyprinter classes to define to_string.
This commit adds these definitions.
Reviewed By: csigg
Differential Revision: https://reviews.llvm.org/D127969
bixia1 [Fri, 17 Jun 2022 23:14:36 +0000 (16:14 -0700)]
[mlir][sparse][taco] Support f16.
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D128105
Mark de Wever [Mon, 16 May 2022 05:42:40 +0000 (07:42 +0200)]
[libc++] Improve charconv base10 algorithm.
This change is a preparation to add the 128-bit integral output.
Before
```
--------------------------------------------------------------
Benchmark Time CPU Iterations
--------------------------------------------------------------
BM_to_chars_good/2 20.1 ns 20.1 ns
35045000
BM_to_chars_good/3 117 ns 117 ns 5916000
BM_to_chars_good/4 83.7 ns 83.7 ns 8401000
BM_to_chars_good/5 70.6 ns 70.6 ns 9915000
BM_to_chars_good/6 59.9 ns 59.9 ns
11678000
BM_to_chars_good/7 53.9 ns 53.8 ns
12995000
BM_to_chars_good/8 19.0 ns 19.0 ns
37110000
BM_to_chars_good/9 45.9 ns 45.8 ns
15278000
BM_to_chars_good/10 9.24 ns 9.24 ns
75343000
BM_to_chars_good/11 42.6 ns 42.6 ns
16449000
BM_to_chars_good/12 38.8 ns 38.8 ns
18101000
BM_to_chars_good/13 38.8 ns 38.8 ns
17999000
BM_to_chars_good/14 37.7 ns 37.6 ns
18571000
BM_to_chars_good/15 35.8 ns 35.8 ns
19660000
BM_to_chars_good/16 15.4 ns 15.4 ns
46129000
BM_to_chars_good/17 32.3 ns 32.3 ns
21763000
BM_to_chars_good/18 32.8 ns 32.8 ns
21396000
BM_to_chars_good/19 33.4 ns 33.4 ns
21078000
BM_to_chars_good/20 33.3 ns 33.3 ns
21020000
BM_to_chars_good/21 32.3 ns 32.3 ns
21807000
BM_to_chars_good/22 31.6 ns 31.6 ns
22057000
BM_to_chars_good/23 30.7 ns 30.7 ns
22938000
BM_to_chars_good/24 28.3 ns 28.3 ns
24659000
BM_to_chars_good/25 28.2 ns 28.2 ns
24790000
BM_to_chars_good/26 28.4 ns 28.4 ns
24410000
BM_to_chars_good/27 28.7 ns 28.7 ns
24423000
BM_to_chars_good/28 28.9 ns 28.9 ns
24139000
BM_to_chars_good/29 28.9 ns 28.9 ns
24347000
BM_to_chars_good/30 29.2 ns 29.2 ns
24141000
BM_to_chars_good/31 29.6 ns 29.6 ns
23699000
BM_to_chars_good/32 29.5 ns 29.5 ns
23933000
BM_to_chars_good/33 28.9 ns 28.9 ns
24042000
BM_to_chars_good/34 28.7 ns 28.7 ns
24361000
BM_to_chars_good/35 28.3 ns 28.3 ns
24703000
BM_to_chars_good/36 28.1 ns 28.1 ns
24924000
BM_to_chars_bad/2 6.16 ns 6.15 ns
114101000
BM_to_chars_bad/3 14.5 ns 14.5 ns
48244000
BM_to_chars_bad/4 16.9 ns 16.9 ns
41974000
BM_to_chars_bad/5 12.5 ns 12.5 ns
56080000
BM_to_chars_bad/6 10.9 ns 10.9 ns
64036000
BM_to_chars_bad/7 14.5 ns 14.5 ns
47294000
BM_to_chars_bad/8 6.36 ns 6.35 ns
110430000
BM_to_chars_bad/9 12.4 ns 12.4 ns
56448000
BM_to_chars_bad/10 5.13 ns 5.13 ns
137596000
BM_to_chars_bad/11 9.88 ns 9.88 ns
69015000
BM_to_chars_bad/12 10.8 ns 10.8 ns
63990000
BM_to_chars_bad/13 10.7 ns 10.7 ns
65066000
BM_to_chars_bad/14 9.71 ns 9.71 ns
71775000
BM_to_chars_bad/15 9.18 ns 9.18 ns
75267000
BM_to_chars_bad/16 6.12 ns 6.12 ns
115000000
BM_to_chars_bad/17 10.7 ns 10.7 ns
65504000
BM_to_chars_bad/18 10.6 ns 10.6 ns
65685000
BM_to_chars_bad/19 9.98 ns 9.98 ns
69894000
BM_to_chars_bad/20 9.74 ns 9.74 ns
72098000
BM_to_chars_bad/21 9.25 ns 9.25 ns
75184000
BM_to_chars_bad/22 9.10 ns 9.10 ns
75602000
BM_to_chars_bad/23 9.48 ns 9.48 ns
72824000
BM_to_chars_bad/24 9.27 ns 9.27 ns
75112000
BM_to_chars_bad/25 9.61 ns 9.61 ns
72080000
BM_to_chars_bad/26 9.72 ns 9.72 ns
72178000
BM_to_chars_bad/27 10.0 ns 10.0 ns
69733000
BM_to_chars_bad/28 10.3 ns 10.3 ns
67409000
BM_to_chars_bad/29 9.97 ns 9.97 ns
69193000
BM_to_chars_bad/30 10.1 ns 10.1 ns
69007000
BM_to_chars_bad/31 9.68 ns 9.68 ns
72232000
BM_to_chars_bad/32 8.99 ns 8.99 ns
76825000
BM_to_chars_bad/33 8.82 ns 8.82 ns
79293000
BM_to_chars_bad/34 8.64 ns 8.64 ns
80441000
BM_to_chars_bad/35 8.96 ns 8.96 ns
75320000
BM_to_chars_bad/36 8.87 ns 8.87 ns
77293000
```
After
```
--------------------------------------------------------------
Benchmark Time CPU Iterations
--------------------------------------------------------------
BM_to_chars_good/2 14.7 ns 14.7 ns
47583000
BM_to_chars_good/3 101 ns 101 ns 6901000
BM_to_chars_good/4 68.4 ns 68.4 ns
10088000
BM_to_chars_good/5 58.2 ns 58.2 ns
12007000
BM_to_chars_good/6 51.1 ns 51.1 ns
13687000
BM_to_chars_good/7 45.6 ns 45.6 ns
15323000
BM_to_chars_good/8 14.6 ns 14.6 ns
47795000
BM_to_chars_good/9 40.7 ns 40.7 ns
17371000
BM_to_chars_good/10 7.48 ns 7.48 ns
90931000
BM_to_chars_good/11 37.6 ns 37.6 ns
18542000
BM_to_chars_good/12 35.2 ns 35.2 ns
19922000
BM_to_chars_good/13 34.9 ns 34.9 ns
20105000
BM_to_chars_good/14 33.5 ns 33.5 ns
20863000
BM_to_chars_good/15 31.9 ns 31.9 ns
22014000
BM_to_chars_good/16 11.7 ns 11.7 ns
60012000
BM_to_chars_good/17 28.9 ns 28.9 ns
24148000
BM_to_chars_good/18 29.0 ns 29.0 ns
24317000
BM_to_chars_good/19 28.7 ns 28.7 ns
24363000
BM_to_chars_good/20 28.1 ns 28.1 ns
24899000
BM_to_chars_good/21 27.5 ns 27.5 ns
25499000
BM_to_chars_good/22 26.9 ns 26.9 ns
25929000
BM_to_chars_good/23 26.2 ns 26.2 ns
26828000
BM_to_chars_good/24 25.1 ns 25.1 ns
27742000
BM_to_chars_good/25 25.3 ns 25.3 ns
27720000
BM_to_chars_good/26 25.2 ns 25.2 ns
27789000
BM_to_chars_good/27 25.3 ns 25.3 ns
27777000
BM_to_chars_good/28 25.3 ns 25.3 ns
27643000
BM_to_chars_good/29 25.3 ns 25.3 ns
27750000
BM_to_chars_good/30 25.4 ns 25.4 ns
27566000
BM_to_chars_good/31 25.4 ns 25.4 ns
27611000
BM_to_chars_good/32 25.8 ns 25.8 ns
27218000
BM_to_chars_good/33 25.7 ns 25.7 ns
27070000
BM_to_chars_good/34 26.1 ns 26.1 ns
26693000
BM_to_chars_good/35 26.4 ns 26.4 ns
26486000
BM_to_chars_good/36 26.3 ns 26.3 ns
26619000
BM_to_chars_bad/2 5.99 ns 5.99 ns
118787000
BM_to_chars_bad/3 14.3 ns 14.3 ns
48567000
BM_to_chars_bad/4 16.0 ns 16.0 ns
43239000
BM_to_chars_bad/5 12.6 ns 12.6 ns
55354000
BM_to_chars_bad/6 10.7 ns 10.7 ns
65491000
BM_to_chars_bad/7 14.4 ns 14.4 ns
48723000
BM_to_chars_bad/8 6.50 ns 6.50 ns
104967000
BM_to_chars_bad/9 12.0 ns 12.0 ns
56552000
BM_to_chars_bad/10 5.16 ns 5.16 ns
136380000
BM_to_chars_bad/11 10.5 ns 10.5 ns
66764000
BM_to_chars_bad/12 10.7 ns 10.7 ns
65534000
BM_to_chars_bad/13 11.0 ns 11.0 ns
63426000
BM_to_chars_bad/14 9.90 ns 9.90 ns
68575000
BM_to_chars_bad/15 9.52 ns 9.52 ns
70932000
BM_to_chars_bad/16 6.14 ns 6.14 ns
111762000
BM_to_chars_bad/17 10.6 ns 10.6 ns
65883000
BM_to_chars_bad/18 10.5 ns 10.5 ns
67606000
BM_to_chars_bad/19 9.96 ns 9.96 ns
68898000
BM_to_chars_bad/20 9.40 ns 9.41 ns
73116000
BM_to_chars_bad/21 9.12 ns 9.12 ns
78647000
BM_to_chars_bad/22 8.95 ns 8.95 ns
80211000
BM_to_chars_bad/23 9.50 ns 9.49 ns
73571000
BM_to_chars_bad/24 9.29 ns 9.29 ns
74690000
BM_to_chars_bad/25 9.65 ns 9.65 ns
72877000
BM_to_chars_bad/26 9.78 ns 9.78 ns
70171000
BM_to_chars_bad/27 10.1 ns 10.1 ns
69543000
BM_to_chars_bad/28 10.4 ns 10.4 ns
67582000
BM_to_chars_bad/29 10.00 ns 10.00 ns
70806000
BM_to_chars_bad/30 9.99 ns 9.99 ns
70340000
BM_to_chars_bad/31 9.56 ns 9.56 ns
74159000
BM_to_chars_bad/32 8.97 ns 8.97 ns
78052000
BM_to_chars_bad/33 8.86 ns 8.86 ns
78586000
BM_to_chars_bad/34 8.81 ns 8.81 ns
78562000
BM_to_chars_bad/35 8.90 ns 8.90 ns
77384000
BM_to_chars_bad/36 9.04 ns 9.04 ns
77263000
```
Reviewed By: #libc, ldionne
Differential Revision: https://reviews.llvm.org/D127764
Mark de Wever [Tue, 14 Jun 2022 17:25:42 +0000 (19:25 +0200)]
[libc++][format] Improves the handle test.
A formatter using a handle only needs to satisfy the BasicFormatter
requirements. The current test allowed more than that minimum. Changed
it to the minimum to make sure it works.
This was due to a post-commit review comment of @vitaut in D121530.
Reviewed By: ldionne, vitaut, #libc
Differential Revision: https://reviews.llvm.org/D127767
Mircea Trofin [Wed, 15 Jun 2022 23:59:47 +0000 (16:59 -0700)]
[FunctionPropertiesAnalysis] Generalize support for unreachable
Generalized support for subgraphs that get rendered unreachable, for
both `call` and `invoke` cases.
Differential Revision: https://reviews.llvm.org/D127921
David Green [Tue, 21 Jun 2022 15:07:01 +0000 (16:07 +0100)]
[AArch64] Move add_and_or_is_add pattern. NFC
This just moves the add_and_or_is_add further up in the file, so that it
can be shared with SVE as in D128159.
Nabeel Omer [Tue, 21 Jun 2022 15:05:21 +0000 (15:05 +0000)]
Revert rGe6ccb57bb3f6b761f2310e97fd6ca99eff42f73e "[SLP] Add cost model for `llvm.powi.*` intrinsics"
This reverts commit
e6ccb57bb3f6b761f2310e97fd6ca99eff42f73e.
Pavel Labath [Tue, 21 Jun 2022 14:47:08 +0000 (16:47 +0200)]
[lldb] Skip Recognizer/assert.test on linux
-> PR56144
Jonas Paulsson [Mon, 31 Aug 2020 14:35:39 +0000 (16:35 +0200)]
[SystemZ] Fix the cost function for vector zero extend.
Zero extend of a vector is done with either a single unpack or a vector
permute, and the TTI cost function should reflect this.
Review: Ulrich Weigand
Nabeel Omer [Mon, 20 Jun 2022 12:49:25 +0000 (12:49 +0000)]
[SLP] Add cost model for `llvm.powi.*` intrinsics
This patch adds handling for the llvm.powi.* intrinsics in
BasicTTIImplBase::getIntrinsicInstrCost() and improves vectorization.
Closes #53887.
Differential Revision: https://reviews.llvm.org/D128172
Jan Svoboda [Tue, 21 Jun 2022 08:20:24 +0000 (10:20 +0200)]
[llvm][vfs] Implement in-memory symlinks
This patch implements symlinks for the in-memory VFS. Original author: @erik.pilkington.
Depends on D117648 & D117649.
Reviewed By: sammccall
Differential Revision: https://reviews.llvm.org/D117650
Jan Svoboda [Tue, 21 Jun 2022 14:20:53 +0000 (16:20 +0200)]
[llvm][vfs] NFC: Promote `InMemoryDirIterator` to nested class
Jan Svoboda [Tue, 21 Jun 2022 14:16:32 +0000 (16:16 +0200)]
[llvm][vfs] NFC: Promote `lookupInMemoryNode()` to member function
Jan Svoboda [Tue, 21 Jun 2022 14:12:09 +0000 (16:12 +0200)]
[llvm][vfs] NFC: Rename `InMemoryFileSystem::addHardLink()` arguments
Alexey Bataev [Fri, 17 Jun 2022 17:23:12 +0000 (10:23 -0700)]
[SLP]Fix a crash when insert subvector is out of range.
If the OffsetBeg + InsertVecSz is greater than VecSz, need to estimate
the cost as shuffle of 2 vector, not as insert of subvector. Otherwise,
the inserted subvector is out of range and compiler may crash.
Differential Revision: https://reviews.llvm.org/D128071
Simon Pilgrim [Tue, 21 Jun 2022 14:08:39 +0000 (15:08 +0100)]
[X86] fold (and (mul x, c1), c2) -> (mul x, (and c1, c2)) iff c2 is all/no bits mask
Noticed on D128216 - if we're zeroing out vector elements of a mul/mulh result then see if we can merge the and-mask into the mul by just multiplying by zero.
Ideally we'd make this generic (similar to the existing foldSelectWithIdentityConstant?), but these cases are appearing very late, after the constants have been lowered to constant-pool loads.
LLVM GN Syncbot [Tue, 21 Jun 2022 14:01:32 +0000 (14:01 +0000)]
[gn build] Port
6a4056ab2ada
Florian Hahn [Tue, 21 Jun 2022 13:55:30 +0000 (15:55 +0200)]
[ConstraintElimination] Remove unneeded StackEntry::Condition (NFC).
The field was only used for debug printing. Print constraint from the
system instead.
Nico Weber [Tue, 21 Jun 2022 13:56:49 +0000 (09:56 -0400)]
Revert "[JITLink][Orc] Add MemoryMapper interface with InProcess implementation"
This reverts commit
6ede65205073d3cf6b1ed4d101e66eae3e0fc8e6.
Doesn't build on Windows, see https://reviews.llvm.org/D127491#3598773
Jay Foad [Wed, 6 Apr 2022 09:49:43 +0000 (10:49 +0100)]
[AMDGPU] Update SPI_SHADER_PGM_RSRC2_PS.EXTRA_LDS_SIZE for GFX11
The granularity of SPI_SHADER_PGM_RSRC2_PS.EXTRA_LDS_SIZE changed
in GFX11. It is now in units of 256 dwords instead of 128 dwords.
COMPUTE_PGM_RSRC2.LDS_SIZE is unaffected. It is still in units of
128 dwords.
Differential Revision: https://reviews.llvm.org/D128179
Benjamin Kramer [Tue, 21 Jun 2022 13:22:35 +0000 (15:22 +0200)]
[mlir][sparse] Preserve NaNs when converting float to bfloat
David Spickett [Tue, 21 Jun 2022 13:12:38 +0000 (13:12 +0000)]
Revert "[CMake] Enable LLVM_ENABLE_PER_TARGET_RUNTIME_DIR by default on Linux"
This reverts commit
311f7839602344ca347816146edb68c0ffaaa060.
Due to not working for Arm where arm and armhf configs are built.
Emre Kultursay [Tue, 21 Jun 2022 12:28:53 +0000 (14:28 +0200)]
Support expressions in the context of a reference
...type variable by dereferencing the variable before
evaluating the expression.
Reviewed By: labath
Differential Revision: https://reviews.llvm.org/D128126
Nikita Popov [Wed, 8 Jun 2022 08:00:31 +0000 (10:00 +0200)]
[GlobalsModRef] Remove check for allocator calls
As the FIXME already indicates, I don't see why this code would be
necessary. If there's a call to an allocator function, that should
get treated just like any other function call -- usually it will be
a declaration and handled conservatively based on memory attributes
only. There should be no need to explicitly force it to be modref.
No test failures either, so I think this is just dead code.
Differential Revision: https://reviews.llvm.org/D127273
Nicolas Vasilache [Tue, 21 Jun 2022 08:38:54 +0000 (01:38 -0700)]
[mlir][Linalg] Split reduction transform op
This revision separates the `LinalgSplitReduction` pattern, whose application is based on attributes,
from its implementation.
A transform dialect op extension is added to control the application of the transformation at a finer granularity.
Differential Revision: https://reviews.llvm.org/D128165
David Spickett [Tue, 21 Jun 2022 11:50:38 +0000 (11:50 +0000)]
[llvm] Disable LLVM_ENABLE_PER_TARGET_RUNTIME_DIR for Arm Linux
I did this before but the other change got reverted and relanded
recently.
arm vs armhf isn't handled correctly so this setting doesn't work
for 32 bit Arm Linux. Causing failures like:
https://lab.llvm.org/buildbot/#/builders/178/builds/2293
LLVM GN Syncbot [Tue, 21 Jun 2022 11:45:40 +0000 (11:45 +0000)]
[gn build] Port
6ede65205073
Anubhab Ghosh [Tue, 21 Jun 2022 11:44:17 +0000 (13:44 +0200)]
[JITLink][Orc] Add MemoryMapper interface with InProcess implementation
MemoryMapper class takes care of cross-process and in-process address space
reservation, mapping, transferring content and applying protections.
Implementations of this class can support different ways to do this such
as using shared memory, transferring memory contents over EPC or just
mapping memory in the same process (InProcessMemoryMapper).
Reviewed By: sgraenitz, lhames
Differential Revision: https://reviews.llvm.org/D127491
Simon Pilgrim [Tue, 21 Jun 2022 11:30:53 +0000 (12:30 +0100)]
[X86] combineAndnp - constant fold ANDNP(C,X) -> AND(~C,X)
If the LHS op has a single use then using the more general AND op is likely to allow commutation, load folding, generic folds etc.
David Green [Tue, 21 Jun 2022 11:21:37 +0000 (12:21 +0100)]
[AArch64] Remove unnecessary funnel shift sve costs.
D127680 added some unnecessary funnel shift costs for AArch64 to "match
the legacy behaviour". The default costs are closer to the correct
values and line up with the scalar/neon costs better. Remove the lines
again to clean up the code, they can be added back at a later date with
better values if needed.
Nikita Popov [Fri, 17 Jun 2022 13:19:23 +0000 (15:19 +0200)]
[SROA] Migrate tests to opaque pointers (NFC)
Tests were updated with this script:
https://gist.github.com/nikic/
98357b71fd67756b0f064c9517b62a34
However, in this case a lot of fixup was required, due to many
minor, but ultimately immaterial differences in results. In
particular, the GEP representation changes slightly in many cases,
either because we now use an i8 GEP, or because we now leave a
GEP alone, using it's original index types and (lack of) inbounds.
basictest-opaque-ptrs.ll has been dropped, because it was an
opaque pointers duplicate of basictest.ll.
Simon Pilgrim [Tue, 21 Jun 2022 10:47:51 +0000 (11:47 +0100)]
[X86] computeKnownBitsForTargetNode - add X86ISD::VBROADCAST_LOAD handling
This requires us to override the isTargetCanonicalConstantNode callback introduced in D128144, so we can recognise the various cases where a VBROADCAST_LOAD constant is being reused at different vector widths to prevent infinite loops.
Martin Storsjö [Tue, 21 Jun 2022 10:16:00 +0000 (13:16 +0300)]
Revert "[LLD] [COFF] Use StringTableBuilder to optimize the string table"
This reverts commit
9ffeaaa0ea54307db309104696a0b6cce6ddda38.
This fixes debugging large executables with lldb and gdb.
When StringTableBuilder is used, the string offsets for any string
can point anywhere in the string table - while previously, all strings
were inserted in order (without deduplication and tail merging).
For symbols, there's no complications in encoding the string offset;
the offset is encoded as a raw 32 bit binary number in half of the
symbol name field.
For sections, the string table offset is written as
"/<decimaloffset>", but if the decimal offset would be larger than
7 digits, it's instead written as "//<base64offset>". Tools that
operate on object files can handle the base64 offset format, but
apparently neither lldb nor gdb expect that syntax when locating the
debug information section. Prior to the reverted commit, all long
section names were located at the start of the string table, so
their offset never exceeded the range for the decimal syntax.
Just reverting this change for now, as the actual benefit from it
was fairly modest.
Longer term, lld could write all long section names unoptimized
at the start of the string table, followed by all the strings for
symbol names, with deduplication and tail merging. And lldb and
gdb could be fixed to handle sections with the base64 offset syntax.
This fixes https://github.com/mstorsjo/llvm-mingw/issues/289.
Nikita Popov [Tue, 21 Jun 2022 10:23:42 +0000 (12:23 +0200)]
[SROA] Regenerate test checks (NFC)
Shraiysh Vaishay [Tue, 21 Jun 2022 09:57:01 +0000 (15:27 +0530)]
[mlir][OpenMP][NFC] Parameter refers to single args and hence changing description for taskgroup allocate clause.
Florian Hahn [Tue, 21 Jun 2022 09:50:13 +0000 (11:50 +0200)]
[ConstraintElimination] Move logic to check condition to helper (NFC).
Balazs Benics [Tue, 21 Jun 2022 09:42:09 +0000 (11:42 +0200)]
[clang-tidy][docs] Fix wrong sphinx link after
d9afb8c3e8fd01a3c89ab2ddebcd44602a30a975
There was a copy-paste mistake at the embedded link:
`clang-tidy/checks/cppcoreguidelines-virtual-class-destructor`
->
`clang-tidy/checks/cppcoreguidelines/virtual-class-destructor`
Sphinx error:
/home/zbebnal/git/llvm-project/clang-tools-extra/docs/ReleaseNotes.rst:168:unknown document: clang-tidy/checks/cppcoreguidelines-virtual-class-destructor
Build bot: https://lab.llvm.org/buildbot#builders/115/builds/29805
Differential Revision: https://reviews.llvm.org/D126891
Balazs Benics [Tue, 21 Jun 2022 09:02:18 +0000 (11:02 +0200)]
[clang-tidy] cppcoreguidelines-virtual-class-destructor should ignore final classes
The `cppcoreguidelines-virtual-class-destructor` supposed to enforce
http://isocpp.github.io/CppCoreGuidelines/CppCoreGuidelines#c35-a-base-class-destructor-should-be-either-public-and-virtual-or-protected-and-non-virtual
Quote:
> A **base** class destructor should be either public and virtual, or
> protected and non-virtual
[emphasis mine]
However, this check still rules the following case:
class MostDerived final : public Base {
public:
MostDerived() = default;
~MostDerived() = default;
void func() final;
};
Even though `MostDerived` class is marked `final`, thus it should not be
considered as a **base** class. Consequently, the rule is satisfied, yet
the check still flags this code.
In this patch, I'm proposing to ignore `final` classes since they cannot
be //base// classes.
Reviewed By: whisperity
Differential Revision: https://reviews.llvm.org/D126891
David Green [Tue, 21 Jun 2022 08:51:22 +0000 (09:51 +0100)]
[AArch64] Add Extract(DUP(C)) as a canonical constant.
As a followup to D128144, this adds extract(DUP(C)) as a canonical
constant to prevent it being transformed back into a BUILD_VECTOR,
leading to an infinite loop.
Carl Ritson [Wed, 28 Jul 2021 10:57:09 +0000 (19:57 +0900)]
[AMDGPU] Set GFX11 null export target based on export attributes
If shader only has depth exports use MRTZ otherwise use MRT0.
Differential Revision: https://reviews.llvm.org/D128185
Nicolas Vasilache [Tue, 21 Jun 2022 07:41:31 +0000 (00:41 -0700)]
[mlir][SCF] Fold tensor.cast feeding into scf.foreach_thread.parallel_insert_slice
Differential Revision: https://reviews.llvm.org/D128247
Matthias Springer [Tue, 21 Jun 2022 08:07:52 +0000 (10:07 +0200)]
[mlir][memref] Fix layout map computation in inferRankReducedResultType
Differential Revision: https://reviews.llvm.org/D128160
Nicolas Vasilache [Mon, 20 Jun 2022 20:34:51 +0000 (13:34 -0700)]
[mlir][SCF] Add scf::ForeachThread canonicalization.
This revision adds the necessary plumbing for canonicalizing scf::ForeachThread with the
`AffineOpSCFCanonicalizationPattern`.
In the process the `loopMatcher` helper is updated to take OpFoldResult instead of just values.
This allows composing various scenarios without the need for an artificial builder.
Differential Revision: https://reviews.llvm.org/D128244
Balázs Kéri [Tue, 21 Jun 2022 06:29:31 +0000 (08:29 +0200)]
[clang][Analyzer] Add errno state to standard functions modeling.
This updates StdLibraryFunctionsChecker to set the state of 'errno'
by using the new errno_modeling functionality.
The errno value is set in the PostCall callback. Setting it in call::Eval
did not work for some reason and then every function should be
EvalCallAsPure which may be bad to do. Now the errno value and state
is not allowed to be checked in any PostCall checker callback because
it is unspecified if the errno was set already or will be set later
by this checker.
Reviewed By: martong, steakhal
Differential Revision: https://reviews.llvm.org/D125400
Kazu Hirata [Tue, 21 Jun 2022 06:35:53 +0000 (23:35 -0700)]
Don't use Optional::getValue (NFC)
Nikolas Klauser [Thu, 16 Jun 2022 09:48:10 +0000 (11:48 +0200)]
[libc++] Implement P0174R2 (Deprecating Vestigial Library Parts in C++17)
Reviewed By: ldionne, Mordante, #libc
Spies: jwakely, libcxx-commits
Differential Revision: https://reviews.llvm.org/D127387
Kazu Hirata [Tue, 21 Jun 2022 06:20:25 +0000 (23:20 -0700)]
[mlir] Don't use Optional::getValue (NFC)
Markus Lavin [Tue, 21 Jun 2022 06:08:06 +0000 (08:08 +0200)]
[machinesink] fix debug invariance issue
Do not include debug instructions when comparing block sizes with
thresholds.
Differential Revision: https://reviews.llvm.org/D127208