Need to check if target allows/supports masked gathers before trying to
estimate its cost, otherwise we may fail to vectorize some of the
patterns because of too pessimistic cost model.
Part of D57059.
Differential Revision: https://reviews.llvm.org/D101297
Since googlebench builds as c++11, the change there is incorrect and breaks the
googlebench build when the STL implementation is strict about std::enable_if_t
not being available in lesser c++ versions.
partial revert of: 1bd6123b78 (https://reviews.llvm.org/D74384)
Differential Revision: https://reviews.llvm.org/D101583
In matrix type casts, we were doing bitcast when the matrices had the same size. This was incorrect and this patch fixes that.
Also added some new CodeGen tests for signed <-> usigned conversions
Reviewed By: fhahn
Differential Revision: https://reviews.llvm.org/D101754
Need to check if target allows/supports masked gathers before trying to
estimate its cost, otherwise we may fail to vectorize some of the
patterns because of too pessimistic cost model.
Part of D57059.
Differential Revision: https://reviews.llvm.org/D101297
As we gradually move more elements of LV to VPlan, we are trying to
reduce the number of places that still has to check IR of the original
loop.
This patch adjusts the code to fix cross iteration phis to get the PHIs
to fix directly from the VPlan that is executed. We still need the
original PHI to check for first-order recurrences, but we can get rid of
that once we model that explicitly in VPlan as well.
Reviewed By: Ayal
Differential Revision: https://reviews.llvm.org/D99293
This patch adds the basic functions needed for controlling auto conversion on z/OS.
Auto conversion is enabled on untagged input file to ASCII by making the assumption that all untagged files are EBCDIC encoded. Output files are auto converted to EBCDIC IBM-1047.
This change also enables conversion for stdin/stdout/stderr.
For more information on how fcntl controls codepage https://www.ibm.com/docs/en/zos/2.4.0?topic=descriptions-fcntl-bpx1fct-bpx4fct-control-open-file-descriptors
Reviewed By: anirudhp
Differential Revision: https://reviews.llvm.org/D100483
We weren't modifying the lock set when intersecting with one coming
from a break-terminated block. This is inconsistent, since break isn't a
back edge, and it leads to false negatives with scoped locks. We usually
don't warn for those when joining locksets aren't the same, we just
silently remove locks that are not in the intersection. But not warning
and not removing them isn't right.
Reviewed By: aaron.ballman
Differential Revision: https://reviews.llvm.org/D101202
The motivation here is to make it available in the base class whether a
fact is managed or not. That would have meant three flags on the base
class, so I had a look whether we really have 8 possible combinations.
It turns out we don't: asserted and declared are obviously mutually
exclusive. Managed facts are only created when we acquire a capability
through a scoped capability. Adopting an asserted or declared lock will
not (in fact can not, because Facts are immutable) make them managed.
We probably don't want to allow adopting an asserted lock (because then
the function should probably have a release attribute, and then the
assertion is pointless), but we might at some point decide to replace a
declared fact on adoption.
Reviewed By: aaron.ballman
Differential Revision: https://reviews.llvm.org/D100801
MSVC has added some new flags. Although they're not supported, this adds
parsing support for them so clang-cl doesn't treat them as filenames.
Except for /fsanitize=address which we do support. (clang-cl already
exposes the -fsanitize= option, but this allows using the
MSVC-spelling with a slash.)
Differential revision: https://reviews.llvm.org/D101439
This adds support for reading AArch64 Pointer Authentication regset
from elf-core file. Also includes a test-case for the same. Furthermore
there is also a slight refactoring of RegisterContextPOSIXCore_arm64
members and constructor. linux-aarch64-pac.core file is generated using
lldb/test/API/functionalities/postmortem/elf-core/main.c with following
clang arguments:
-march=armv8.5-a -mbranch-protection=pac-ret+leaf -nostdlib -static -g
Reviewed By: DavidSpickett
Differential Revision: https://reviews.llvm.org/D99941
Similarly to D101096, this makes sure that MMO operands get propagated
through from MVE gathers/scatters to the Machine Instructions. This
allows extra scheduling freedom, not forcing the instructions to act as
scheduling barriers. We create MMO's with an unknown size, specifying
that they can load from anywhere in memory, similar to the masked_gather
or X86 intrinsics.
Differential Revision: https://reviews.llvm.org/D101219
SITargetLowering::LowerFormalArguments asserts that none of these
features are used for graphics calling conventions, so
AnnotateKernelFeatures should not add them.
Differential Revision: https://reviews.llvm.org/D101534
The number of bytes copied in CopyAndPad should depend on the size of
the type being copied, not on its shift value (which in the case of char
is 0, leading to no bytes at all being copied).
Add unit tests for CharacterMin and CharacterMax, which exercise this
code path.
Differential Revision: https://reviews.llvm.org/D101355
The current code computes the minimum element length based on the `type`
used to create the descriptor and uses that as the element length
whenever it is greater than 0. This means that the `elem_len` parameter
is essentially ignored for any type where we can compute a minimum
element length (which includes `CFI_type_char[16|32]_t`), and we may
therefore end up with descriptors with a lower element length than
expected.
This patch fixes the issue by explicitly doing what the standard says,
i.e. it uses the given `elem_len` for character types, `CFI_type_struct`
and `CFI_type_other`, and ignores it (falls back to the minimum element
length) for everything else.
Differential Revision: https://reviews.llvm.org/D101659
It looks like CFI_type_struct was once used as the last valid CFI_type
value, but in the meantime CFI_type_char16_t and CFI_type_char32_t were
added, making that assumption no longer true. Luckily, in the meantime
we also got a define for CFI_TYPE_LAST, which we can now use to allow
CFI_establish and CFI_allocate to work with descriptors of
CFI_type_char16_t, CFI_type_char32_t and any other future types.
Differential Revision: https://reviews.llvm.org/D101658
This enables the runtime tests on amdgpu targets.
10 tests have been marked as XFAIL on amdgcn currently mostly due to
missing printf.
Reviewed By: protze.joachim
Differential Revision: https://reviews.llvm.org/D99656
Add address sanitizer instrumentation support for accesses to global
and constant address spaces in AMDGPU. It strictly avoids instrumenting
the stack and assumes x86 as the host.
Reviewed by: vitalybuka
Differential Revision: https://reviews.llvm.org/D99071
We create MMO's for the VLDn/VSTn intrinsics in ARMTargetLowering::
getTgtMemIntrinsic, but they do not currently make it ll the way through
ISel. This changes that in the various places it needs changing, making
sure that the MMO is propagate through to the final instruction. This
can help in scheduling, not treating the VLD2/VST2 as a scheduling
barrier.
Differential Revision: https://reviews.llvm.org/D101096
Setting the preffered function alignment to 16 for Cortex A53/A55
improves performance in a wide range of benchmarks. This brings it
in line with the Cortex-A53/A55 tuning that is used in GCC
(gcc/config/aarch64/aarch64.c).
Differential Revision: https://reviews.llvm.org/D101636
Change-Id: I2ce47fe7ab5e3b54f49c89038d8da4e404742de2
This allows for a much more efficient encoding for small negative
numbers by storing the sign bit first and negating the rest of
the bits. This was already being used for OPC_CheckInteger.
For every in tree target this affects, the table got smaller.
R600GenDAGISel.inc saw the largest reduction of 7K.
I did have to add a new opcode for StringIntegers used for
register class ids and subregister indices since we don't have the
integer value to encode. The enum name is emitted directly into
the table. Previously assumed the enum would expand to a positive
7-bit number. We might be able to just shift that right by 1 and
assume it is a positive 6 bit number, but that will need more
investigation.
This shrinks the immediate that isel table needs to emit for these
instructions. Hoping this allows me to change OPC_EmitInteger to
use a better variable length encoding for representing negative
numbers. Similar to what was done a few months ago for OPC_CheckInteger.
The alternative encoding uses less bytes for negative numbers, but
increases the number of bytes need to encode 64 which was a very
common number in the RISCV table due to SEW=64. By using Log2 this
becomes 6 and is no longer a problem.
As opposed to going through the Aliasee type.
For opaque pointers, we're trying to remove uses of PointerType::getElementType().
Reviewed By: dblaikie
Differential Revision: https://reviews.llvm.org/D101715
This patch introduces a helper to obtain an iterator range for the
PHI-like recipes in a block.
Reviewed By: Ayal
Differential Revision: https://reviews.llvm.org/D100101
This patch makes the builtin operand order match the C operand order
for all intrinsics. With this we can use clang_builtin_alias for
all overloaded intrinsics.
This should further reduce the test time for vector intrinsics.
Differential Revision: https://reviews.llvm.org/D101700
`test/std/ranges/range.access/range.access.cbegin/incomplete.compile.verify.cpp`
was accidentally copied (and apparently the author either forgot to
delete it or forgot to commit the deletion).
TEST=`ninja cxx && ninja check-cxx` locally
Alexey Bataev