I've split the git archive generation into three steps:
1. generate pure tarball
2. append top-level cmake directory to all tarballs
3. compress the archive
This was inspired by D118252 and can be considered an alternative
approach for all projects to have access to the shared cmake
directory when building in standalone mode.
When generating source tarballs on my local laptop it takes 9 minutes and 45 seconds WITH this patch applied. When this patch is not applied, it takes 9minutes and 38 seconds. That means, this patch introduces a slowdown of 7 seconds, which seems fair.
Reviewed By: tstellar
Differential Revision: https://reviews.llvm.org/D118481
The pointer is always dereferenced, so assert the cast is correct (which it should be as we just created that ScalableVectorType) instead of returning nullptr
Add the atomic overloads for the `global` and `local` address spaces,
which are new in OpenCL 3.0. Ensure the preexisting `generic`
overloads are guarded by the generic address space feature macro.
Ensure a subset of the atomic builtins are guarded by the
`__opencl_c_atomic_order_seq_cst` and `__opencl_c_atomic_scope_device`
feature macros, and enable those macros for SPIR/SPIR-V targets in
`opencl-c-base.h`.
Also guard the `cl_ext_float_atomics` builtins with the atomic order
and scope feature macros.
Differential Revision: https://reviews.llvm.org/D119420
When deciding where to split a block to insert stack guard checks, we should
move past any debug instructions we see that might (e.g.) be separating a tail
call from its frame wrangling.
SCCP requires that the load/store type and global type are the
same (it does not support bitcasts of tracked globals). With
typed pointers this was implicitly enforced.
`CallDescriptions` for builtin functions relaxes the match rules
somewhat, so that the `CallDescription` will match for calls that have
some prefix or suffix. This was achieved by doing a `StringRef::contains()`.
However, this is somewhat problematic for builtins that are substrings
of each other.
Consider the following:
`CallDescription{ builtin, "memcpy"}` will match for
`__builtin_wmemcpy()` calls, which is unfortunate.
This patch addresses/works around the issue by checking if the
characters around the function's name are not part of the 'name'
semantically. In other words, to accept a match for `"memcpy"` the call
should not have alphanumeric (`[a-zA-Z]`) characters around the 'match'.
So, `CallDescription{ builtin, "memcpy"}` will not match on:
- `__builtin_wmemcpy: there is a `w` alphanumeric character before the match.
- `__builtin_memcpyFOoBar_inline`: there is a `F` character after the match.
- `__builtin_memcpyX_inline`: there is an `X` character after the match.
But it will still match for:
- `memcpy`: exact match
- `__builtin_memcpy`: there is an _ before the match
- `__builtin_memcpy_inline`: there is an _ after the match
- `memcpy_inline_builtinFooBar`: there is an _ after the match
Reviewed By: NoQ
Differential Revision: https://reviews.llvm.org/D118388
As usual with that header cleanup series, some implicit dependencies now need to
be explicit:
llvm/MC/MCParser/MCAsmParser.h no longer includes llvm/MC/MCParser/MCAsmLexer.h
Preprocessed lines to build llvm on my setup:
after: 1068185081
before: 1068324320
So no compile time benefit to expect, but we still get the looser coupling
between files which is great.
Discourse thread: https://discourse.llvm.org/t/include-what-you-use-include-cleanup
Differential Revision: https://reviews.llvm.org/D119359
This patch simply adds an optional garbage collector attribute to LLVMFuncOp which maps 1:1 to the "gc" property of functions in LLVM.
Differential Revision: https://reviews.llvm.org/D119492
In Clang we can attach TBAA metadata based on the load/store intrinsics
based on the operation's element type.
This also contains changes to InstCombine where the AArch64-specific
intrinsics are transformed into generic LLVM load/store operations,
to ensure that all metadata is transferred to the new instruction.
There will be some further work after this patch to also emit TBAA
metadata for SVE's gather/scatter- and struct load/store intrinsics.
Reviewed By: paulwalker-arm
Differential Revision: https://reviews.llvm.org/D119319
This allows operations to control the block ids used by the printer in nested regions.
Reviewed By: Mogball
Differential Revision: https://reviews.llvm.org/D115849
Auto-upgrades that rely on the pointer element type do not work in
opaque pointer mode. The idea behind this patch is that we can
instead work with type IDs, for which we can retain the pointer
element type. For typed pointer bitcode, we will have a distinct
type ID for pointers with distinct element type, even if there will
only be a single corresponding opaque pointer type.
The disclaimer here is that this is only the first step of the change,
and there are still more getPointerElementType() calls to remove.
I expect that two more patches will be needed:
1. Track all "contained" type IDs, which will allow us to handle
function params (which are contained in the function type) and GEPs
(which may use vectors of pointers)
2. Track type IDs for values, which is e.g. necessary to handle loads.
Differential Revision: https://reviews.llvm.org/D118694
In addition to the self-recursion check, also check whether there
is more than one node in the SCC, which implies that there is a
larger cycle. I believe checking SCC structure (rather than
something like norecurse) is the right thing to do here, because
this is specifically about preventing infinite loops over the SCC.
Fixes https://github.com/llvm/llvm-project/issues/42028.
Differential Revision: https://reviews.llvm.org/D119418
Previously, OpDSL did not support rank polymorphism, which required a separate implementation of linalg.fill. This revision extends OpDSL to support rank polymorphism for a limited class of operations that access only scalars and tensors of rank zero. At operation instantiation time, it scales these scalar computations to multi-dimensional pointwise computations by replacing the empty indexing maps with identity index maps. The revision does not change the DSL itself, instead it adapts the Python emitter and the YAML generator to generate different indexing maps and and iterators depending on the rank of the first output.
Additionally, the revision introduces a `linalg.fill_tensor` operation that in a future revision shall replace the current handwritten `linalg.fill` operation. `linalg.fill_tensor` is thus only temporarily available and will be renamed to `linalg.fill`.
Reviewed By: nicolasvasilache, stellaraccident
Differential Revision: https://reviews.llvm.org/D119003
libc++ -DLLVM_ENABLE_MODULES=on build needs the complete type. This fixes
error: invalid application of 'sizeof' to an incomplete type 'llvm::MemoryBuffer'
An impilt used of C++ module without prebuild path may cause crash.
For example:
```
// ./dir1/C.cppm
export module C;
// ./dir2/B.cppm
export module B;
import C;
// ./A.cpp
import B;
import C;
```
When we compile A.cpp without the prebuild path of C.pcm, the compiler
will crash.
```
clang++ -std=c++20 --precompile -c ./dir1/C.cppm -o ./dir1/C.pcm
clang++ -std=c++20 --precompile -fprebuilt-module-path=./dir2 -c
./dir2/B.cppm -o ./dir2/B.pcm
clang++ -std=c++20 -fprebuilt-module-path=./dir2 A.cpp
```
The prebuilt path of module C is cached when import module B, and in the
function HeaderSearch::getCachedModuleFileName, the compiler try to get
the filename by modulemap without check if modulemap exists, and there
is no modulemap in C++ module.
Reviewed By: ChuanqiXu
Differential review: https://reviews.llvm.org/D119426
`{back,front}_insert_iterator` and `ostream{,buf}_iterator` effectively
fully implement the One Ranges Proposal already, so mark them as done:
- the change to `difference_type` was made by D103273;
- default constructors and the associated default member initializers
were removed by wg21.link/P2325 (implemented by D102468).
Also fix a stale template signature in the `<iterator>` synopsis.
None of these tests are really intended to test the file offset as much
as to test the structure. Making the regex allows this test to work
even if the file is checked out with CRLF line endings.
Reviewed By: aaron.ballman
Differential Revision: https://reviews.llvm.org/D119362
The last use of this file was removed in late 2013 in ea56494625.
The last use was in PathProfiling.cpp, which had an overview comment
of the overall approach.
Similar functionality lives in the slight more cryptically named
CFGMST.h in this same directory. A similar overview comment is
in PGOInstrumentation.cpp.
No behavior change.
We have a mix of substituted lld (`%lld`) and hard-coded lld (`ld64.lld`) commands.
When testing with different versions of LLD, this would require going into every place
where lld is hard-coded and changing that. If we centralize it, this'll only require us
to modify it in only one place and will make it easy to run the same test suite. Plus,
this will make it be consistent with how we write other tests.
Reviewed By: #lld-macho, int3, oontvoo
Differential Revision: https://reviews.llvm.org/D119394
Add new operations to the gpu dialect to represent device side
asynchronous copies. This also add the lowering of those operations to
nvvm dialect.
Those ops are meant to be low level and map directly to llvm dialects
like nvvm or rocdl.
We can further add higher level of abstraction by building on top of
those operations.
This has been discuss here:
https://discourse.llvm.org/t/modeling-gpu-async-copy-ampere-feature/4924
Differential Revision: https://reviews.llvm.org/D119191
Note that this doesn't actually cause the top level predicate to become a non-union just yet.
The * above comes from a case in the LoopVectorizer where a predicate which is later proven no longer blocks vectorization due to a change from checking if predicates exists to whether the predicate is possibly false.
The superclass method handles a bunch of useful things. For example
it applies flags such as `-fnew-alignment` which doesn't work without
this patch.
Differential Revision: https://reviews.llvm.org/D118573
Due to the way type units work, this would lead to a declaration in a
type unit of a local type in a CU - which is ambiguous. Rather than
trying to resolve that relative to the CU that references the type unit,
let's just not try to simplify these names.
Longer term this should be fixed by not putting the template
instantiation in a type unit to begin with - since it references an
internal linkage type, it can't legitimately be duplicated/in more than
one translation unit, so skip the type unit overhead. (but the right fix
for that is to move type unit management into a DICompositeType flag
(dropping the "identifier" field is not a perfect solution since it
breaks LLVM IR linking decl/def merging during IR linking))
`opt -analyze` is legacy PM-specific. Show better ways of doing the same
thing, generally with some sort of `-passes=print<foo>`.
Reviewed By: asbirlea
Differential Revision: https://reviews.llvm.org/D119486
`bug49334.cpp` directly uses `!=` to compare two floating point values,
which is almost wrong.
Reviewed By: jhuber6
Differential Revision: https://reviews.llvm.org/D119485
Konrad Kleine