Truncates the APInt if the bit width is greater than the width specified,
otherwise do nothing
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D91445
D71880 makes this dependency redundant and we can safely remove it. Tested for
both shared lib build and static lib build.
Reviewed By: hokein
Differential Revision: https://reviews.llvm.org/D91951
Previously, lax conversions were only allowed between SVE vector-length
agnostic types and vector-length specific types. This meant that code
such as the following:
#include <arm_sve.h>
#define N __ARM_FEATURE_SVE_BITS
#define FIXED_ATTR __attribute__ ((vector_size (N/8)))
typedef float fixed_float32_t FIXED_ATTR;
void foo() {
fixed_float32_t fs32;
svfloat64_t s64;
fs32 = s64;
}
was not allowed.
This patch makes a minor change to areLaxCompatibleSveTypes to allow for
lax conversions to be performed between SVE vector-length agnostic types
and GNU vectors.
Differential Revision: https://reviews.llvm.org/D91696
This patch introduces new canonicalization rules which are used for AST-based
rename in Clangd. By comparing two canonical declarations of inspected nodes,
Clangd determines whether both of them belong to the same entity user would
like to rename. Such functionality is relatively concise compared to the
Clang-Rename API that is used right now. It also helps to overcome the
limitations that Clang-Rename originally had and helps to eliminate several
classes of bugs.
Clangd AST-based rename currently relies on Clang-Rename which has design
limitations and also lacks some features. This patch breaks this dependency and
significantly reduces the amount of code to maintain (Clang-Rename is ~2000 LOC,
this patch is just <30 LOC of replacement code).
We eliminate technical debt by simultaneously
* Maintaining feature parity and ensuring no regressions
* Opening a straightforward path to improving existing rename bugs
* Making it possible to add more capabilities to rename feature which would not
be possible with Clang-Rename
Reviewed By: hokein
Differential Revision: https://reviews.llvm.org/D71880
This revision refactors code used in various Linalg transformations and makes it a first class citizen to the LinalgStructureOpInterface. This is in preparation to allowing more advanced Linalg behavior but is otherwise NFC.
Differential revision: https://reviews.llvm.org/D91863
The dependency mechanism for C has been implemented, and we have rolled out
this to all internal users, didn't see crashy issues, we consider it is stable
enough.
Differential Revision: https://reviews.llvm.org/D89046
It is possible to trigger a crash/misbehavior when the st_name field of
the signature symbol goes past the end of the string table.
This patch fixes it.
Differential revision: https://reviews.llvm.org/D91943
SCEV makes a logical mistake when handling EitherMayExit in
case when both conditions must be met to exit the loop. The
mistake looks like follows: "if condition `A` fails within at most `X` first
iterations, and `B` fails within at most `Y` first iterations, then `A & B`
fails at most within `min (X, Y)` first iterations". This is wrong, because
both of them must fail at the same time.
Simple example illustrating this is following: we have an IV with step 1,
condition `A` = "IV is even", condition `B` = "IV is odd". Both `A` and `B`
will fail within first two iterations. But it doesn't mean that both of them
will fail within first two first iterations at the same time, which would mean
that IV is neither even nor odd at the same time within first 2 iterations.
We can only do so for known exact BE counts, but not for max.
Differential Revision: https://reviews.llvm.org/D91942
Reviewed By: nikic
This patch moves the selection of the style used to emit the numbers
(DW_OP_implicit_value vs. DW_OP_const+DW_OP_stack_value) into
DwarfExpression::addUnsignedConstant. This logic is not FP-specific, and
it will be needed for large integers too.
The refactor also makes DW_OP_implicit_value (DW_OP_stack_value worked
already) be used for floating point constants other than float and
double, so I've added a _Float16 test for it.
Split off from D90916.
Differential Revision: https://reviews.llvm.org/D91058
I disabled the widening in fa5cb4b because it run in an assert, which was
related to replacing values with different types. I forgot that an extend could
also be a zero-extend, which I have added now. This means that the approach now
is to create and insert a trunc value of the outerloop for each user, and use
that to replace IV values.
Differential Revision: https://reviews.llvm.org/D91690
Add a 'can_connect' parameter to Process plugin initialization, and use
it to filter plugins to these capable of remote connections. This is
used to prevent 'process connect' from picking up a plugin that can only
be used locally, e.g. the legacy FreeBSD plugin.
Differential Revision: https://reviews.llvm.org/D91810
I saw this crash in our internal production, but unfortunately didn't get
reproduced testcase, we likely hit this crash when the AST is ill-formed
(e.g. broken code).
Reviewed By: gribozavr2
Differential Revision: https://reviews.llvm.org/D91614
Restore Linux-alike regset names for AVX/MPX registers
as TestLldbGdbServer seems to depend on them. At the same time, fix
TestRegisters to be aware that they are not available on FreeBSD
and NetBSD, at least until we figure out a better way of reporting
unsupported register sets.
Differential Revision: https://reviews.llvm.org/D91923
Fix qRegisterInfo tests to handle Exx error response when querying
registers that are not supported on the platform in question. This
is how FreeBSD and NetBSD platforms reporting missing registers right
now, and there certainly is value from verifying the remaining
registers.
This change fixes the test for FreeBSD but NetBSD has other regressions
that still need to be researched.
Differential Revision: https://reviews.llvm.org/D91922
It is possible to trigger reading past the EOF by breaking fields like
DT_PLTRELSZ, DT_RELSZ or DT_RELASZ
This patch adds a validation in `DynRegionInfo` helper class.
Differential revision: https://reviews.llvm.org/D91787
Handling of `and` and `or` vastly uses copy-paste. Factored out into
a helper function as preparation step for further fix (see PR48225).
Differential Revision: https://reviews.llvm.org/D91864
Reviewed By: nikic
The function was introduced on Jan 23, 2019 in commit
73078ecd38.
Its definition was removed on Oct 27, 2020 in commit
0930763b4b, leaving the declaration
unused.
- Fixes bug 48242 point 3 crash.
- Makes the improvments from points 1 & 2.
https://bugs.llvm.org/show_bug.cgi?id=48262
```
def RTLValueType : Type<CPred<"isRTLValueType($_self)">, "Type"> {
string cppType = "::mlir::Type";
}
```
Works now, but merely by happenstance. Parameters expects a `TypeParameter` class def or a string representing a c++ type but doesn't enforce it.
Reviewed By: lattner
Differential Revision: https://reviews.llvm.org/D91939
It'd be nicer if there was a group target that forwarded either to
//clang-tools-extra/clangd/index/remote or
//clangd/index/remote/unimplemented based on if remote index is enabled,
but for now it's never enabled in the gn build.
* Makes `pip install pybind11` do the right thing with no further config.
* Since we now require a version of pybind11 greater than many LTS OS installs (>=2.6), a more convenient way to get a recent version is preferable.
* Also adds the version spec to find_package so it will skip older versions that may be lying around.
* Tested the full matrix of old system install, no system install, pip install and no pip install.
Differential Revision: https://reviews.llvm.org/D91903
Put project-aware-index between command-line specified static index and
ClangdServer indexes.
This also moves remote-index dependency from clangDaemon to ClangdMain
in an attempt to prevent cyclic dependency between clangDaemon and
remote-index-marshalling.
Differential Revision: https://reviews.llvm.org/D91860
An index implementation that can dispatch to a variety of indexes
depending on the file path. Enables clangd to work with multiple indexes in the
same instance, configured via config files.
Depends on D90749, D90746
Differential Revision: https://reviews.llvm.org/D90750
Compilation logic for External blocks. A few of the high level points:
- Requires exactly one-of File/Server at a time:
- Server is ignored in case of both, with a warning.
- Having none is an error, would render ExternalBlock void.
- Ensures mountpoint is an absolute path:
- Interprets it as relative to FragmentDirectory.
- Defaults to FragmentDirectory when empty.
- Marks Background as Skip.
Depends on D90748.
Differential Revision: https://reviews.llvm.org/D90749
Enable configuration of remote and static indexes through config files
in addition to command line arguments.
Differential Revision: https://reviews.llvm.org/D90748
Currently the tarballs contain superfluous metadata, like the user name
of the packager and via Pax headers even the PID of the tar process that
packaged the files. We build the monorepo projects directly from the git
repo using "git archive" and for the test-suite we add some flags as
recommended by https://reproducible-builds.org/docs/archives/. We don't
use numeric owners though to be compatible with "git archive".
The advantage of "git archive" is that the releaser doesn't have to
download the tar ball and extract it, rather the archive is built
directly from the repository. This is probably what GitHub uses
internally to produce the tarballs, so I wouldn't expect a difference.
Reviewed By: tstellar
Differential Revision: https://reviews.llvm.org/D91494
Technically 'noexcept' isn't a qualifier, so this should be a separate conversion.
Also make the test a pure frontend test.
Reviewed By: rsmith
Differential Revision: https://reviews.llvm.org/D67112
This might be a regression for some ARM targets, but that should
be changed in the target-specific overrides.
There is apparently still no default lowering for these nodes,
so I am assuming these intrinsics are not in common use.
X86, PowerPC, and RISC-V for example, just crash given the most
basic IR.
Test a few more variations:
* NoAlias with different strides
* MustAlias without loop
* MustAlias with same stride
* MustAlias base pointers with different stride
Semantic check added to check and restrict the value of the parameter in the COLLAPSE or ORDERED clause
if it is larger than the number of nested loops following the construct.
Test Cases:
omp-do-collapse-positivecases.f90
omp-do-collapse.f90
omp-do-ordered-positivecases.f90
omp-do-ordered.f90
Reviewed by: Kiran Chandramohan @kiranchandramohan , Valentin Clement @clementval
Differential Revision: https://reviews.llvm.org/D89860
This is part of the discussion on D91876 about trying to reduce custom lowering of MIN/MAX ops on older SSE targets - if we can improve generic vector expansion we should be able to relax the limitations in SelectionDAGBuilder when it will let MIN/MAX ops be generated, and avoid having to flag so many ops as 'custom'.
Some of the buildbots were failing due to what seems to be them using a non c++14 compilant std::string implementation.
Since c++14 std::basic_string::append(const basic_string, size_t, size_t) has a defaulted 3rd paramater, but some of the build bots were reporting that it wasn't defaulted in their implementation.
Kerry McLaughlin