rldimi is 64-bit instruction, due to backward compatibility, it needs to
be expanded into series of rotate and masking in 32-bit environment. In
the future, we may improve bit permutation selector and remove such
direct codegen.
Currently, the builtins used for implementing `va_list` handling
unconditionally take their arguments as unqualified `ptr`s i.e. pointers
to AS 0. This does not work for targets where the default AS is not 0 or
AS 0 is not a viable AS (for example, a target might choose 0 to
represent the constant address space). This patch changes the builtins'
signature to take generic `anyptr` args, which corrects this issue. It
is noisy due to the number of tests affected. A test for an upstream
target which does not use 0 as its default AS (SPIRV for HIP device
compilations) is added as well.
rldimi is 64-bit instruction, so the corresponding builtin should not
be available in 32-bit mode. Rotate amount should be in range and
cases when mask is zero needs special handling.
This change also swaps the first and second operands of rldimi/rlwimi
to match previous behavior. For masks not ending at bit 63-SH,
rotation will be inserted before rldimi.
These builtins are already there in Clang, however current codegen may
produce suboptimal results due to their complex behavior. Implement them
as intrinsics to ensure expected instructions are emitted.
In the beginning, Clang only emitted atomic IR for operations it knew
the
underlying microarch had instructions for, meaning it required
significant
knowledge of the target. Later, the backend acquired the ability to
lower
IR to libcalls. To avoid duplicating logic and improve logic locality,
we'd like to move as much as possible to the backend.
There are many ways to describe this change. For example, this change
reduces the variables Clang uses to decide whether to emit libcalls or
IR, down to only the atomic's size.
Moved from https://reviews.llvm.org/D126302
The current behaviour with these three options is quite undesirable:
-mno-altivec -mvsx allows VSX to override no Altivec, thereby turning on
both
-msoft-float -maltivec causes a crash if an actual Altivec instruction
is required because soft float turns of Altivec
-msoft-float -mvsx is also accepted with both Altivec and VSX turned off
(potentially causing crashes as above)
This patch diagnoses these impossible combinations in the driver so the
user does not end up with surprises in terms of their options being
ignored or silently overridden.
Fixes https://github.com/llvm/llvm-project/issues/55556
---------
Co-authored-by: Nemanja Ivanovic <nemanja.i.ibm@gmail.com>
This patch canonicalizes getelementptr instructions with constant
indices to use the `i8` source element type. This makes it easier for
optimizations to recognize that two GEPs are identical, because they
don't need to see past many different ways to express the same offset.
This is a first step towards
https://discourse.llvm.org/t/rfc-replacing-getelementptr-with-ptradd/68699.
This is limited to constant GEPs only for now, as they have a clear
canonical form, while we're not yet sure how exactly to deal with
variable indices.
The test llvm/test/Transforms/PhaseOrdering/switch_with_geps.ll gives
two representative examples of the kind of optimization improvement we
expect from this change. In the first test SimplifyCFG can now realize
that all switch branches are actually the same. In the second test it
can convert it into simple arithmetic. These are representative of
common optimization failures we see in Rust.
Fixes https://github.com/llvm/llvm-project/issues/69841.
Set the writable and dead_on_unwind attributes for sret arguments. These
indicate that the argument points to writable memory (and it's legal to
introduce spurious writes to it on entry to the function) and that the
argument memory will not be used if the call unwinds.
This enables additional MemCpyOpt/DSE/LICM optimizations.
When constructing vectors from elements, use poison instead of
undef as the base value. These literals always initialize all
elements (padding the remainder with zero), so that the choice
of base value does not affect semantics.
smmintrin.h uses __builtin_mffs, __builtin_mffsl, __builtin_mtfsf and
__builtin_set_fpscr_rn. This patch replaces the uses with ppc prefix
and implement the missing ones.
Functions with these attributes will be automatically called before
main() or after main() exits gracefully. In glibc environments, the
constructor function is passed the same arguments as main(), so that
signature is allowed. In all other environments, we require the function
to accept no arguments and either return `void` or `int`. The functions
must use the C calling convention. In C++ language modes, the functions
cannot be a nonstatic member function, or a consteval function.
Additionally, these reuse the same priority logic as the init_priority
attribute which explicitly reserved priorty values <= 100 or > 65535. So
we now diagnose use of reserved priorities the same as we do for the
init_priority attribute, but we downgrade the error to be a warning
which defaults to an error to ease use for implementers like compiler-rt
or libc.
This relands a633a3761f with fixes.
Functions with these attributes will be automatically called before
main() or after main() exits gracefully. In glibc environments, the
constructor function is passed the same arguments as main(), so that
signature is allowed. In all other environments, we require the function
to accept no arguments and either return `void` or `int`. The functions
must use the C calling convention. In C++ language modes, the functions
cannot be a nonstatic member function, or a consteval function.
Additionally, these reuse the same priority logic as the init_priority
attribute which explicitly reserved priorty values <= 100 or > 65535. So
we now diagnose use of reserved priorities the same as we do for the
init_priority attribute, but we downgrade the error to be a warning
which defaults to an error to ease use for implementers like compiler-rt
or libc.
This relands b4435104ca with fixes.
This reverts commit b4435104ca.
This caused too many disruptions in compiler-rt where reserved
priorities were being used. Reverting to investigate appropriate
solutions.
Functions with these attributes will be automatically called before
`main()` or after `main()` exits gracefully, which means the functions
should not accept arguments or have a returned value (nothing can
provide an argument to the call in these cases, and nothing can use the
returned value), nor should they be allowed on a non-static member
function or consteval function in C++. We allow 'int' as a return type for
the function due to finding a significant amount of historical code using
`int(void)` as a signature.
Additionally, these reuse the same priority logic as the init_priority
attribute which explicitly reserved priorty values <= 100 or > 65535. So
we now diagnose use of reserved priorities the same as we do for the
init_priority attribute.
Add __builtin_bcopy to the list of GNU builtins. This was causing a
series of test failures in glibc.
Adjust the tests to reflect the changes in codegen.
Fixes#51409.
Fixes#63065.
mffsl is available since ISA 3.0. The builtin is named with ppc prefix
to follow our convention. For targets earlier than power9, GCC generates
extra code to support the functionality, while this patch does not
implement such behavior.
Reviewed By: nemanjai, tuliom
Differential Revision: https://reviews.llvm.org/D158065
This patch renames the `OpenMPIRBuilderConfig` flags to reduce confusion over
their meaning. `IsTargetCodegen` becomes `IsGPU`, whereas `IsEmbedded` becomes
`IsTargetDevice`. The `-fopenmp-is-device` compiler option is also renamed to
`-fopenmp-is-target-device` and the `omp.is_device` MLIR attribute is renamed
to `omp.is_target_device`. Getters and setters of all these renamed properties
are also updated accordingly. Many unit tests have been updated to use the new
names, but an alias for the `-fopenmp-is-device` option is created so that
external programs do not stop working after the name change.
`IsGPU` is set when the target triple is AMDGCN or NVIDIA PTX, and it is only
valid if `IsTargetDevice` is specified as well. `IsTargetDevice` is set by the
`-fopenmp-is-target-device` compiler frontend option, which is only added to
the OpenMP device invocation for offloading-enabled programs.
Differential Revision: https://reviews.llvm.org/D154591
This patch relaxes the front end AIX diagnostics added in D102070 to accept the
local-exec TLS model, as we plan to support this model in a series of future patches.
The diagnostics are relaxed when local-exec is used as a compiler option to
`-ftls-model=*` and in the `__attribute__((tls_model("local-exec")))` attribute.
Differential Revision: https://reviews.llvm.org/D149596
This is the consolidation of D151644 and D151943 moved from
InstCombine to FunctionAttrs. This is based on discussion in the above
patches as well as D152081 (Attributor). This patch was written in a
way so it can have an immediate impact in currently active passes
(FunctionAttrs), but should be easy to port elsewhere (Attributor or
Inliner) if that makes more sense later on.
Some function attributes imply the attribute for all/some instructions
in the function. These attributes can be safely propagated to
callsites within the function that are missing the attribute. This can
be useful when 1) analyzing individual instructions in a function
and 2) if the original caller is later inlined, as if the attributes are
not propagated, they will be lost.
This patch implements propagation in a new class/file
`InferCallsiteAttrs` which can hypothetically be included elsewhere.
At the moment this patch infers the following:
Function Attributes:
- mustprogress
- nofree
- willreturn
- All memory attributes (readnone, readonly, writeonly, argmem,
etc...)
- The memory attributes are only propagated IFF the set of
pointers available to the callsite is the same as the set
available outside the caller (i.e no local memory arguments
from alloca or local malloc like functions).
Argument Attributes:
- noundef
- nonnull
- nofree
- readnone
- readonly
- writeonly
- nocapture
- nocapture is only propagated IFF the set of pointers
available to the callsite is the same as the set available
outside the caller and its guranteed that between the
callsite and function return, the state of any capture
pointers will not change (so the nocaptured gurantee of the
caller has been met by the instruction preceding the
callsite and will not changed).
Argument are only propagated to callsite arguments that are also function
arguments, but not derived values.
Return Attributes:
- noundef
- nonnull
Return attributes are only propagated if the callsite's return value
is used as the caller's return and execution is guranteed to pass from
callsite to return.
The compile time hit of this for -O3 and -O3+thinLTO is ~[.02, .37]%
regression. Proper LTO, however, has more significant regressions (up
to 3.92%):
https://llvm-compile-time-tracker.com/compare.php?from=94407e1bba9807193afde61c56b6125c0fc0b1d1&to=79feb6e78b818e33ec69abdc58c5f713d691554f&stat=instructions:u
Differential Revision: https://reviews.llvm.org/D152226
Add more builtins for stdio functions as in GCC, along with their
mutations under IEEE float128 ABI.
Reviewed By: tuliom
Differential Revision: https://reviews.llvm.org/D150087
This patch adds clang options `-mxcoff-roptr` and `-mno-xcoff-roptr` to specify storage locations for constant pointers on AIX.
When the `-mxcoff-roptr` option is in effect, constant pointers, virtual function tables, and virtual type tables are placed in read-only storage. When the `-mno-xcoff-roptr` option is in effect, pointers, virtual function tables, and virtual type tables are placed are placed in read/write storage.
This patch depends on https://reviews.llvm.org/D144189.
Reviewed By: hubert.reinterpretcast, stephenpeckham
Differential Revision: https://reviews.llvm.org/D144190
Clang has mechanism to specify required target features of a built-in
function. This patch adds such definitions to Altivec, VSX, HTM,
PairedVec and MMA builtins.
This will help frontend to detect incompatible target features of
bulitin when using target attribute syntax.
Reviewed By: nemanjai, kamaub
Differential Revision: https://reviews.llvm.org/D143467
With this patch an undefined mask in a shufflevector will be printed as poison.
This change is done to support the new shufflevector semantics
for undefined mask elements.
Differential Revision: https://reviews.llvm.org/D149210
This patch prefixes omp outlined helpers and reduction funcs
with the original function's name.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D140722
This patch attempts to prefix omp outlined helpers and reduction funcs
with the original function's name.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D140722
This patch is to fix some implicit castings for emulated intrinsics
so that there are no lax-vector-conversions errors and warnings.
Reviewed By: nemanjai, #powerpc
Differential Revision: https://reviews.llvm.org/D144293
Instcombine prefers this canonical form (see getPreferredVectorIndex),
as does IRBuilder when passing the index as an integer so we may as
well use the prefered form from creation.
NOTE: All test changes are mechanical with nothing else expected
beyond a change of index type from i32 to i64.
Differential Revision: https://reviews.llvm.org/D140983
